6 International

Seminar on Process

editors
Fernando Valenzuela L.

Hydrometallurgy

p ro ce e di n g s
sponsor

6 International
Seminar on Process
Hydrometallurgy

6 International
Seminar on Process
Hydrometallurgy

July 23-25, 2014

Via del Mar, Chile

editor
Fernando Valenzuela L.

Copyright
Gecamin, Chile. All rights reserved.
No part of this publication may be reproduced,
stored or transmitted in any form or by any
means, electronic, mechanical, by photocopying,
recording or otherwise, without the prior
written permission from Gecamin.
Authors disclaimer
Any views and opinions presented in the articles
published in these proceedings are solely
those of the authors and do not necessarily
represent those of Gecamin. The authors take
full and exclusive responsibility for technical
content, style, language and accuracy of the
information published herein. This information
is not intended nor implied to be a substitute
for professional advice. The editors are not
responsible for any damage to property or
persons that may occur as a result of use of
the information contained in this volume.
I.S.B.N. 978-956-9393-09-9
Gecamin
Paseo Bulnes 197, Piso 6
Santiago, Chile
Postcode: 833 0336
Telephone: +56 2 2652 1500
www.gecamin.com
Online proceedings
www.gecaminpublications.com/hydroprocess2014

 Contents

Organizers

Committees

Foreword

Preface

Acknowledgements

Proceedings Sponsor

c h a pt er 1

p l en ary p res en tati o n s

Rare Earths processing routes: A summary of the key

technologies. Geysa Pereira

Hydrometallurgy exploration for secondary copper lead

smelters byproducts and copper zinc plants byproducts.

Ricardo Benavides
Metallurgical Geometallurgy. Jorge Menacho and

Gerardo Molina

c h a pt er 2 l eac h i n g s tu di es

Penetration depth and unleached core area within large

zinc ore particles. Yousef Ghorbani, Jochen Petersen and
Jean-Paul Franzidis

4D automatic quantification of ore particles in heap

leaching from X-ray micro computed tomography.
Qingyang Lin, Stephen Neethling, Katherine Dobson,
Loic Courtois and Peter Lee

Acid In Situ Leaching (ISL) of phosphor-coffinite at

K ATCO. Sergio Bustos, Diana Mango, Michel Belires,

Frank Ling, Marc Brouand and Csar Garca
Application of unconventional methods to the treatment

of leachates contaminated with heavy metals. Mara Ruiz,

Alicia Guevara and Ernesto de la Torre
Industrial results of online monitoring system for

leaching pads. Juan Larenas, Paula Lpez, Toms

Fuenzalida and Andrs Alvarado

 c h a pt er 3 c o p p er h y dro metal l u r g i c a l p r o c e s se s
Copper sulfide ore leaching in seawater: Oxidants influence.

Pa Hernndez, Mara Taboada, Tefilo Graber and

Hctor Galleguillos
Immobilization of arsenic by schwertmannite synthesized

from copper heap leach solutions. Sengpasith HoungAloune,

Naoki Hiroyoshi, Mayumi Ito and Masato Fujimura
Possible role of copper species in enhanced chalcopyrite

dissolution. Cynthia Torres, Mara Taboada, Tefilo Graber,

Osvaldo Herrero, Helen Watling, Denis Shiers and
Pa Hernndez
Tellurium precipitation from copper sulfate leach solutions.

Yusuf Bello and Christie Dorfling

In situ measurements of copper and moisture in an

operating heap using advanced geophysics. Kent Lang,

Ned Clayton, Roland Banas and Ricardo Ruano
Improved processes of extracting copper from its ores.

William Culver

c ha pt e r 4

h y dro m etal l u rg y o f b a s e m e ta l s

an d s al i n e co mp o u n d s

Design of a treatment system for leachate from uncontrolled

disposal of steel dust on soil. Mara Ruiz, Alicia Guevara and

Ernesto de la Torre
Dephosphorization of iron ore by acid leaching. 

Christie Dorfling and Eliana Boshoff

Leaching of coarse caliche ore. Javier Ordez,

Silvia Valdez, Luis Cisternas and Luis Moreno

Reducing fresh water use in the production of metals. 

Robert Bruce and Tyler Seaman

 c ha pt e r 5 g o l d an d val u ab l e m e ta l s h y d r o m e ta l l u r g y
SART to remove zinc and copper from a silver Merrill Crowe

barren leach solution. David Sanguinetti, Farzad Mohammadi

and scar Lpez
Development of the Albion Process plant to treat refractory

concentrates from the GPM Gold Project. 

Mike Hourn, Paul Voigt and Duncan Turner
Cerro Negro Process design, from the design to operation.

Cristian Ramos and Steve Dixon

c h a pt er 6

mo del i n g an d o p ti m i z i n g o f

h y dro m etal l u rg y o p e r at i o n s

Improved models for predicting heap leaching

hydrodynamics. Saman Ilankoon and Stephen Neethling

Hydro-mechanical-chemical coupled modeling of copper

heap leaching with COMSOL-PHREEQC. Orlando Silva, Albert

Nardi, Gabriela Romn-Ross, Jorge Molinero, David Arcos,
Pablo Quesada, Eduardo Ruiz and Jordi Guimer
New approaches to old problems can dramatically improve

the viability of operations. 

David Robinson, Iztok Livk, Jie Wu, Phillip Fawell, Keith
Barnard, Rueben Rajasingam and John Farrow
Characterization and in situ monitoring of large scale heap

leach fluid dynamics. Danny Santis, Michael Milczarek,

Tzung-mow Yao and Jason Keller
Phenomenological modeling of the intensive cyanidation

of gold and silver concentrates in continuous reactors. 

Moiss Miranda

 c h a pt er 7 i n n o vati o n s i n el ec t r o m e ta l l u r g y

an d h y dro m etal l u r g y

Sump box for retaining organic entrainments in

electro-winning cells. Rodrigo Villarroel, Fernando Penna,

Cristin Villaseca and Hans Tippmann
Failure mechanisms of lead electrowinning anodes. 

Abbas Mirza, Matt Burr, Timothy Ellis, Larry Webb,

Ricardo Cabrejas, Eben Lombard, Andrew Johnston and
Franois Leclercq
Steel dust leaching with organic acids for zincs recovery

using a bioelectrolitic cell. Byron Quiroga and

Ernesto de la Torre
Flowmeter validation for metallurgical balance in Minera

Michilla, using radiotracers. F rancisco Daz, scar Jimnez,

Genny Leinenweber, Abraham Backit, Armando Cceres and
Cristian Mundaca
Extraction of sulfides from aqueous effluents of the oil

industry. Joo Bosco De Araujo Paulo, Ana Karla Costa De

Oliveira, Elisa Maria Bittencourt Dutra De Sousa and Kssia
Larissa Pinheiro De Lima

c h a pt er 8

s o l v en t extracti o n s t u d i e s a n d p r o c e ss e s

Use of CuPRO MEX extractants in copper solvent

extraction plants. Mauro Torres, Brant Sudderth,

Pablo Tapia, Cristin Salgado, lvaro Arancibia and
Hctor Bravo
Experimental study of entrainment of phase in copper

solvent extraction. Jonathan Castillo, Patricio Navarro and

Cristian Vargas
Outotecs hydrometallurgical process for recovery of copper

from challenging concentrates. Kari Valkama, Marko Lampi

and Janne Karonen

 Change in extractant formulation in Minera Franke.

F rancisco Villegas, Ral Manrquez, Ral Funes,

Rodrigo Zambra and David Acevedo
Recovery of the copper content of a mine drainage by

solvent extraction using a liquid membrane emulsion

process. Fernando Valenzuela, Claudio Araneda, Carlos
Basualto and Jaime Sapag-Hagar
Improvements in the operation of SX plants of Anglo

American Copper. Juan Nez, Jessica Cueto,

Jos Fernndez, Cristian Salgado, Marco Torres and
Gabriel Zarate
Extraction of Cu(II) ions through a process coupling an

ionic liquid and supercritical CO2. Rossana Seplveda,

Jonathan Castillo, Julio Romero and Jos Snchez
Lanthanide metals extraction with extractants attached

to magnetite nanoparticles surface. Jos Gaete, Lorena

Molina, Fernando Valenzuela and Carlos Basualto

c ha pt e r 9 b i o h y dro m etal l u rgy s t u d i e s

Synthetic Biology in Extraction, Separation and

Remediation. Christopher DaCunha and Patrick Nee

Remediation of acid mine drainage by microorganism.

Pamela Gutirrez, Mara Taboada and Pamela Chvez

c ha pt e r 10

recy c l i n g o f m etal s

Solvent extraction for separation of metals from hev

Li-ion batteries nitrate leach liquor. Sravya Kosaraju,

Christian Ekberg and Stefan Allard
Solvent extraction for separation of metals from HEV

Li-ion batteries sulfate leach liquor. Sravya Kosaraju,

Christian Ekberg and Stefan Allard

Editors

Organizers

The Hydroprocess 2014 Seminar was organized by Gecamin, Chile.

Gecamin
Powering professional development
for sustainable mining
Gecamin is a Chilean company with 16 years of experience organizing
technical and international conferences for the mining industry. Our
conferences aim to inform and inspire professionals from all over the
world, fostering the exchange of best practices and innovative experiences.
Over 18,000 professionals have attended our events and have been
trained in areas fundamental to the mining industry. These areas include
Geology and Mining, Mineral Processing, Hydrometallurgy, Sustainability
and Environment, Water and Energy, Maintenance and Automation, and
Human Capital.
Gecamin seeks to contribute to the sustainable development of the
mining industry by openly addressing its most pressing concerns and by
offering a platform for knowledge exchange that aims at identifying the
most sustainable solutions.
Learn more about Gecamin conferences by visiting www.gecamin.com

Committees

Organizing Committee

Advisory Committee

Executive Committee

Pablo Amigo, Jacobs, Chile

chair

Corby Anderson, Colorado School of Mines, USA

Luis Snchez, President, BHP Billiton

Pampa Norte, Chile
co-ch air s
Ricardo Benavides, Deputy Director,
Centro de Investigacin y Desarrollo
Tecnolgico, Industrias Peoles, Mexico
Geysa Pereira, Technical Manager,
Hydrometallurgy, Vale, Brazil
Joe Pezo, Process Plant Manager, Mina Lagunas
Norte, Barrick Gold Corporation, Peru
past president
Gabriel Meruane, R&D Project Manager,
SQM, Chile
e x ecu tiv e dir ec tor
Carlos Barahona, General Manager,
Gecamin, Chile
t e c h n i c a l c o o r d i n at o r
Fernando Valenzuela, Professor,
Universidad de Chile
s e m i n a r c o o r d i n at o r
Fabiola Bustamante,
Gecamin, Chile

Directing Members
Percy Mayta, Plant Manager Oxides,
Sierra Gorda SCM, Chile
Rolando Morales, Plant Manager, SCM
Minera Lumina Copper, Caserones, Chile
Gustavo Tapia, Technological Innovation and
Process Manager, Antofagasta Minerals, Chile

Jaime Arana, Tia Mara Project, Southern Copper

Corporation, Peru
Francisco Arriagada, Arcadis, Chile
Antonio Ballester, Universidad Complutense de
Madrid, Spain
Sergio Castro, Antofagasta Minerals, Chile
Dick Celmer, Fluor, Chile
Virginia Ciminelli, Universidade Federal de Minas
Gerais, UFMG, Brazil
Clenilson Da Silva Souza Junior, Instituto Federal do
Rio de Janeiro, Brazil
Ernesto De La Torre, Escuela Politcnica
Nacional, Ecuador
Claudia Diniz, Vale, Brazil
Steve Dixon, GoldCorp Tucson, USA
Esteban Domic, Domic Ingeniera y
Construcciones, Chile
scar Ferrada, Minera Escondida, BHP Billiton, Chile
Rolando Garca, Servicios Corporativos del
Mayo, Mexico
Javier Guevara, Sociedad Minera Cerro Verde, Peru
Manuel Guzmn, Molibdenos y Metales, Chile
Norma Hoyos, Minera Escondida, BHP Billiton, Chile
Marcelo Jo, Jo y Loyola Consultores de Procesos, Chile
Gustavo Lagos, Centro de Minera, Pontificia
Universidad Catlica, Chile
Cleve Lightfoot, BHP Billiton, Chile
Jorge Menacho, De Re Metallica Ingeniera, Chile
Luis Moreno, Royal Institute of Technology, Sweden
Patricio Navarro, Department of Metallurgical
Engineering, Universidad de Santiago de Chile
Felipe Nez, Yamana, Chile
Kwadwo Osseo-Asare, Penn State University, USA
Eduardo Patio, BHP Billiton Chile
Sergio Riveras, El Salvador Division, Codelco, Chile
Dave Robinson, CSIRO, Australia

Fernando Romero, Anglo American, Chile

Jaime Tapia, Universidad Arturo Prat, Chile

Julio Tremolada, Instituto Iberoamericano de

Diego Verdejo, Minera El Tesoro, Chile

Metalurgia Extractiva, Peru

Petrus Van Staden, Mintek, South Africa

Editorial Committee

Courtney Young, Montana Tech, USA

editor

Technical Committee

Fernando Valenzuela, Universidad de Chile

Fernando Acevedo, Pontifica Universidad Catlica

reviewers

de Valparaso, Chile

Sergio Bustos, Areva Mines, France

Francisco Carranza, Universidad de Sevilla, Spain

Jess Casas, Process Consulting, Chile

Jess Casas, Process Consulting, Chile

Jonathan Castillo, Universidad de Atacama, Chile

Danny Castillo, Ausenco, Chile

Virginia Ciminelli, Universidade Federal de Minas

Manuel Chvez, Freeport McMoran Copper &

Gerais, UFMG, Brazil

Gold, Chile

Luis Cisternas, Universidad de Antofagasta, CICITEM,

Luis Cisternas, Universidad de Antofagasta,

CSIRO, Chile

CICITEM, CSIRO, Chile

Wiliam Culver, State University of New York, USA

Juan Pablo Garcs, Compaa Minera Doa Ins de

Steve Dixon, GoldCorp Tucson, USA

Collahuasi, Chile

Humberto Estay, Arcadis, Chile

Juan Carlos Gentina, Pontificia Universidad

Jordi Guimera, Amphos 21 Consulting, Chile

Catlica de Valparaso, Chile

Amado Guzmn, HydroGeoSense, USA

Tefilo Graber, Department of Chemical

Saman Iialkoon, Imperial College London, UK

Engineering, Universidad de Antofagasta, Chile

Romke Kuyvenhoven, Gecamin, Chile

Amado Guzmn, HydroGeoSense, USA

Cleve Lightfoot, BHP Billiton, Chile

Christian Hu, Independent, Chile

Abbas Mirza, RSR Anode Group and RSR

Juan Patricio Ibez, Universidad Tcnica Federico

Technologies, USA

Santa Mara, Chile

Gerardo Olivares, Tesra, Chile

Cristian Jara, Compaa Minera Teck Carmen

Rafael Padilla, Universidad de Concepcin, Chile

de Andacollo, Chile

Li Quian, Central South University, China

Juan Alberto Montes, Andina Division, Codelco, Chile

Dave Robinson, CSIRO, Australia

Gonzalo Montes-Atenas, Department of Mining

Hernn Rodrguez, Compaa Minera Teck Quebrada

Engineering, Universidad de Chile

Blanca, Chile

Rafael Padilla, Universidad de Concepcin, Chile

Mara Cristina Ruiz, Universidad de Concepcin, Chile

Carolina Paipa, Universidad de Playa Ancha, Chile

Mara Ruiz, Escuela Politcnica Nacional, Ecuador

Eduardo Robles, Hatch, Chile

David Sanguinetti, BioteQ Environmental

Vernica Rueda, SNC Lavalin, Chile

Technologies, Canada

Mara Cristina Ruiz, Universidad de

Orlando Silva, Amphos 21 Consulting, Spain

Concepcin, Chile

Gustavo Tapia, Antofagasta Minerals, Chile

Andrs Soto, Consultora Esmeralda and

Jaime Tapia, Universidad Arturo Prat, Chile

Universidad Los Andes, Chile

Mauro Torres, SNF FloMin South America, Chile

Mara Elisa Taboada, Universidad de

Gabriel Zarate, Anglo American Chile

Antofagasta, Chile

Foreword

In the context of an increasingly demanding market where innovation

and efficiency have become critical success factors, hydrometallurgy has
positioned itself as a specific and highly attractive alternative for mineral
processing. Far from the secondary role that it held decades ago, this
area of work has been gradually internalized in the industry through the
development of projects that require lower capital resources and have
low operating costs when compared to other processing methods
currently employed.
In addition, the relative water and energy efficiency, make
hydrometallurgical processes compatible options for mining efforts to
develop a sustainable activity within the environment and its surroundings.
These characteristics show that hydrometallurgy is able to help solve the
different challenges of mining, including productivity. This concept is now
high on the agenda of Chile, as part of the route to achieve development
goals. Again hydrometallurgy opens doors, thanks to the creativity,
innovation and improvement of processes that may materially impact
productivity of the mining industry, which is the main economic activity of
the country.
Besides the advantages, the development of hydrometallurgy presents
multiple challenges. One of them is to strengthen human capital, through
the training of researchers, professionals, consultants and qualified
technical personnel, required to take this area to the next level. We must
move forward in the development of methods which help us to process ores
with greater complexity, according to the characteristics of the ores that we
operate today. Thus, hydrometallurgical processes will continue to be part
of the portfolio of projects that the industry expects to develop in the
coming decades.
The efforts in scientific research and the companies operational ability to
internalize new developments, will cement part of this path. Also, we must
add the contribution of small and medium size entrepreneurs who include
hydrometallurgical processes as part of their innovative solutions. We will
keep working to strenghten hydrometallurgy as another success factor that
allows us to contribute to building a better future for our society.
Thank you for being part of this great challenge.
Luis Snchez
chair
Hydroprocess 2014~6 International Seminar
on Process Hydrometallurgy

Preface

Hydroprocess 2014, held on July23-25, 2014 at the Sheraton Miramar Hotel

in Via del Mar, is the sixth of a series of international seminars on Process
Hydrometallurgy initiated in 2006 by Gecamin, Chile. The organization of
Hydroprocess 2014 has represented a great effort and special dedication for
all of us involved in the conference, however at the same time, it has
signified an excellent opportunity to meet and work with all the researchers
and professionals from different countries that have selflessly collaborated
with us. Undoubtedly, Hydroprocess 2014 has provided an international
forum where professionals and researchers can analyze and discuss recent
innovations and developments concerning to the wet processing of ores
and materials.
The seminar was organized according to the high standards this
meeting requires. Thank you to the authors and their organizations for
submitting papers; to all the technical experts, for sharing their expertise,
dedicating valuable time correcting the articles, and for providing
insightful comments thus enhancing the quality of this publication and the
standard of the seminar.
This book contains 43 abstracts written by delegates from 18 different
countries. The conference has been organized by area of interest, including
Base metals hydrometallurgical processes, Hydrometallurgy of gold and
saline compounds, Solvent extraction and Electrometallurgical processes,
Biohydrometallurgy, Modeling and optimizing of hydrometallurgical
operations, Recycling of metals and alloys and Industrial case studies.
Unquestionably, this book will be of interest to the international mining
hydrometallurgical community highly motivated and engaged to
accomplish with the primary spirit of this version of the conference,
Contributing to clean and efficient mining. We sincerely hope all
delegates of this conference had a nice stay in Via del Mar and have
enjoyed a fruitful participation in Hydroprocess 2014.
Fernando Valenzuela
t ech nic a l coor din ator a n d edi tor
Hydroprocess 2014~6 International Seminar
on Process Hydrometallurgy

Acknowledgements

The Executive Organizing Committee acknowledges with gratitude the

efforts of all the authors for contributing a large variety of high quality,
detailed and innovative papers to the technical program. We also would
like to thank the reviewers, the employees from Gecamin, and all those
involved in the creation of these proceedings for their assistance. The
support of the Organizing, Advisory and Technical Committees has been
greatly appreciated, as has been the support of the Hydroprocess 2014
Chair, Co-Chairs and the Chairs of technical sessions.
The Executive Organizing Committee also wishes to thank the following
sponsors (as of June 26, 2014 in the order in which they were confirmed) for
their generous support:
Gold: Biohydro.cl and BASF
Silver: Outotec, RSR Anodes, TecnoComposites, Cytec, GEA Westfalia
Separator Chile and Metalex
Official Material: Arcadis and Cytec
Institutional Partners: Sociedad Nacional de Minera (SONAMI), Chile;
Consejo Minero, Chile; Servicio Nacional de Geologa y Minera
(SER NAGEOMIN), Chile; Instituto de Ingenieros de Minas del Per (IIMP),
and The Brazilian Mining Association (IBR A M)
Proceedings: BHP Billiton Pampa Norte
Official Media: AreaMinera, Chile
Media Partners: Nueva Minera & Energa, Chile; Direcmin, Chile; and Qu
Pasa Minera, Chile.
Finally, we would like to thank all the delegates who attended the
seminar and exchanged their valuable knowledge and expertise, thus
contributing to the great success of this 6th edition of the International
Seminar on Process Hydrometallurgy, Hydroprocess 2014. We are looking
forward to seeing you all again during the next version of Hydroprocess, in
the year 2015.
Executive Organizing Committee
Hydroprocess 2014~6 International Seminar
on Process Hydrometallurgy

Proceedings Sponsor

We proudly acknowledge BHP Billiton Pampa Norte, as the Proceedings Sponsor

of the 6th International Seminar on Process Hydrometallurgy.

chapter 1
Plenary presentations

Rare Earths processing routes: A summary

of the key technologies

Geysa Pereira
Vale, Brazil

The Rare Earths comprise a group of fifteen elements with atomic numbers
57 through 71, from lanthanum to lutetium (lanthanide group), and
yttrium. Ytrium is also included with this group because of its chemical
and physical similarities with the lanthanides. Due to their atomic weight
they are grouped as Light Rare Earth Elements (LREE) and Height Rare
Earth Elements (HREE).
The demand of these elements is relatively small, but they are essential
to hightechnology applications. Magnets, metal alloys for batteries and
light-weight structures, and phosphorous are essential for many current
and emerging alternative energy technologies (electric vehicles, wind
power). They are also critical for defense systems.
The Peoples Republic of China is the major world supplier of these
elements, producing more than 95% of the entire worlds annual new
supply of Rare Earths. Constraints on Chinese exports are creating
opportunities for non-chinese projects, and in the last three years, the
world is facing a strong movement towards new developments in order to
decrease this dependence.
There are a wide variety of REE-bearing ores, and based on their
mineral class, a specific beneficiation and treatment process route is
required. Despite a variety of mineral classes, only a few mineral species
such as bastnaesite, monazite, xenotime and ionic clays, have been
commercially processed.
This paper aims to present an overview of the key technologies being
applied in the current projects as well the proposed flowsheet for some
new projects.

There is no full article associated with this abstract.

Hydrometallurgy exploration for secondary copper lead

smelters byproducts and copper zinc plants byproducts

Ricardo Benavides
Centro de Investigacin y
Desarrollo Tecnolgico,
Industrias Peoles, Mexico

The natural drain for copper lead smelters are copper products obtained in
the lead smelting, copper is normally associated with arsenic and sulfur,
species as speiss and copper matte feature high in values gold and silver.
In the refining of zinc plant (RLEW) normal purge copper zinc plants,
the copper is drained and cemented in the purification processes of
electrolyte as cement copper which has poor values of realization due to the
high content of impurities, mainly zinc and cadmium.
Peoles has become a technological business challenge, to treat lead
concentrates with high (very high) contained silver and gold and arsenic,
copper and antimony being. Also in the zinc plant for processes of complex
zinc concentrates with high iron, copper, cadmium, cobalt and silver. The
drain of copper from zinc plants are the cements products of copper,
cadmium and cobalt.
Peoles CIDT undertook the task of evaluating pyro and
hydrometallurgical technology alternatives in order to meet this challenge
and discusses a hydrometallurgical alternative in this presentation. This
evaluation opens alternatives to other materials (ores and concentrates)
that could be evaluated.

There is no full article associated with this abstract.

Metallurgical geometallurgy

Jorge Menacho and

Gerardo Molina
De Re Metallica
Consultants, Chile

This presentation will give a critical analysis of the state-of-the-art in

modern geometallurgy is given. Neither a clear definition of attributes to
be measured nor appropriate measuring devices are currently available.
Lack of robust interpretation models other than statistical fitting tools is
also a reality.
Geology has been the classical starting point for geometallurgy so that
metallurgical domain responses are tried to predict starting from
geological domains, not always biunivocally related.
In a different approach metallurgical domain responses are defined from
the own metallurgical responses. Characteristic pattern responses are
properly identified assisted by the DRM.AcidLeach phenomenological
model. Domains are defined according to parameter domains rather than
response domains. Parameters for leach applications are classified into three
groups: (i) Metallurgical, (ii) rock/soil mechanics and (iii) hydrodynamics.
Application examples are given and also linked to classical
geometallurgical models. The proposed approach is 95% reliable while the
classical one is only 80%.

There is no full article associated with this abstract.

chapter 2
Leaching studies

Penetration depth and unleached core area within

large zinc ore particles

Yousef Ghorbani
Universidad Catlica del
Norte and Centro de
Investigacin Cientfico
Tecnolgico para la Minera
(cicitem), Chile
Jochen Petersen and
Jean-Paul Franzidis
University of Cape Town,
South Africa

The economics of heap leaching are strongly governed by the trade-off

between the slow rate and limited extent of leaching from large particles
and the cost of ore crushing. X-ray Computed Tomography (CT) was used to
follow the progression of Zn leaching in a number of individual large
sphalerite ore particles, which were subjected to a long-term simulated
heap bioleaching environment. The ore was prepared by two different
modes of comminution-HPGR (at 45, 90 and 120 bar) and cone crusher and
individual particles were selected from three different size fractions.
Investigation of the reacted fraction of Zn vs. distance from the center of
each particle indicated that leaching from large particles leads to near
complete conversion close to the surface, but only partial conversion in the
zones that are closer to the center of particles. The cores of the conecrushed particles show hardly any conversion at all, especially in the larger
particle sizes. The results indicated that the penetration depth (Pd) is a
function of both the comminution mode and the particle size/radius (R).
Density of cracks and particle porosity are important in increasing the
penetration depth, and these characteristics are more pronounced in
HPGR-crushed particles.

INTRODUCTION
Although both, particle scale and bulk scale, effects are important in heap bioleaching, little has
been done to systematically integrate particle scale models into bulk scale models. Most existing
bulk scale models account for the effect of particle topology using simplified models such as the
shrinking core model, applied to an average particle size. Implicit in these simplifications is an
assumption regarding the relative significance of particle scale phenomena and which kinetics
(diffusion or reaction) are limiting at the particle scale. Yet it must be understood, that a
conventional shrinking core approach would work only for ore particles that are homogeneously
porous and have mineral grains well distributed throughout (Liddell, 2005; Velardo et al., 2002;
Vegli et al., 2001). The shrinking core model has been shown to be a poor predictor of leach
kinetics based on the crushed ore size distribution (Miller, 2003). However, the model itself is a very
useful tool to analyse commercial heap data. Data is obtained from batch leaching, which is then
fitted according to the equations describing the topological model, and then the set of parameters
that gives the best fit is chosen as best representative of the system. There are no reliable criteria to
predict a priori the rate-controlling regime in which a given system will operate. Clearly, obtaining
these criteria will require a greater understanding of the controlling processes on a microscopic
level (Liddell, 2005; Velardo et al., 2002). A part of the difficulty in understanding leaching reactions
lies in their heterogeneous nature and the variability from one ore or concentrate to another.
However, another aspect relates to the fact that the conventional shrinking-core equations have
been narrowly focused on the lixiviant and the particles being leached (Liddell, 2005). Models
usually assume spherical geometry of the particles, although ore or concentrate particles are never
spherically symmetric. Deviation from the spherical shape leads to an increase in the surface-tovolume ratio (Rossi, 1990). Furthermore, particle scale effects in heap bioleaching are influenced by
other factors, including the ore porosity, grain exposure, grain and particle size and shape
distribution, the ore mineralogy, surface properties, pore size distribution and microbial
interactions (Ghorbani et al., 2011a).
The overall rate at which a mineral is dissolved from an ore particle is quite often the manifestation
of a complex network of individual phenomena, each proceeding at its own intrinsic rate. For finely
ground particles such as those encountered in tank leaching processes, leaching rates are typically
controlled by the intrinsic kinetics of mineral breakdown. In heap or dump leaching, on the other
hand, mineral grains are usually embedded within larger ore particles or solution-filled
agglomerates, and thus accessible only by diffusion through a network of pores (Ghorbani et al.,
2011a; Bouffard and Dixon, 2001). Compared with the efforts usually taken to determine and
interpret the effects of temperature, reagent concentration and other solution variables, very little
has been done to measure or detect the effect of the variables relating to the actual ore particles in
most systems. Historically this has been constrained by technical limitations to study mineralogy
and the inability to look inside particles while they were leaching. Although such limits in the
techniques for measurement of these parameters remain, an increasingly wider range of
microscopy and tomography techniques is available to provide visual evidence for suspected and
unsuspected phenomena.
In this paper one of these techniques, X-ray Computed Tomography (X-ray-CT), is used to explore
the progression of leaching inside particles. X-ray CT as an advanced diagnostic and nondestructive technique is used for measuring the presence and disappearance of mineral grains
within individual ore particles in the course of the long-term leach experiment.

MATERIALS AND METHODS

A bulk sphalerite ore (see TABLE 1 for main sulphide minerals) was obtained from the Gamsberg
Zinc mine in the Northern Cape Province, South Africa. The bulk sample was initially crushed by a
jaw crusher from 300 mm top size down to -40 mm. The jaw crusher product was representatively
split into 250 kg bags for further crushing down to -25 mm by either a Kppern HPGR unit at 90 bar
pressure or cone crusher at Mintek in Randburg, South Africa. Products from the cone crusher and
HPGR were then screened into five size fractions (-25/+23, -23/+16, -16/+14, -14/+6.75, -6.75/+5.25
mm). In this study, subsamples from the large size fraction (-25/+23mm), medium size fraction
(-16/+14mm) and small size fraction (-6.75/+5.25 mm) were subjected to a long-term bioleach study
in a specially designed circulating fluid fixed-bed reactor. Results of this study are described in
detail in a previous paper (Ghorbani et al., 2012).
Table 1 Mineralogical composition of the ore
Mineral

Amount (wt. %)

Pyrrhotite

1.2

Pyrite

33.8

Sphalerite

16.0

Galena

0.2

In the present study, we describe an analysis of the progress of leaching in selected individual
particles from different size fractions in the course of the previous bioleaching study. For this, the
reactors were stopped from time to time to investigate the progress of leaching by taking and then
analysing X-ray CT images of the same individual particles. The use of this technique to identify
both cracks and sphalerite grains within a matrix of gangue and other sulphide minerals especially
pyrite, has been discussed in a previous paper (Ghorbani et al., 2011b). This approach was used here
to collect images of particles before and at different stages of the leach experiment.
An HMXST CT scanner (X-Sight X-ray Services in Stellenbosch, South Africa), with 225 kV X-ray
source, 3 m resolution reflection target, and interchangeable Nano-tech 1 m transmission target
was employed. Further operating details of the X-ray CT conditions are given in TABLE 2.
Table 2 Experimental conditions used for X-ray CT measurements
X-ray energy

130 kV and 200 kV micro-focus source with 3 m focal spot size

X-ray intensity

120 A beam current

Integration frames

Each projection image was exposed for 1 second per frame - no integration of frames

Random movements

No random movements and continuous rotation of sample

Beam hardening filters

A software correction was used which modifies the Beer-Lambert Law (attenuation
curve) to match the sample. A pre-set correction was chosen including both linear and
quadratic curve fitting

Second radiation filters

1mm Cu

Following the reconstruction of the acquired X-ray CT data, the ore particles were analysed using
the VGStudio Max 2.1 image analysis software (Volume Graphics GmbH, Heidelberg, Germany).
The crack depth was calculated from analysing the initial X-ray CT images of the particles at the
scale of image resolution (up to 15 m). The results were further validated with those obtained
using traditional (destructive) techniques such as SEM/EDS and QEMSCAN on other particles
taken from the same reactor at the various intervals (Ghorbani et al., 2011b). Only few selected
particles from each size fraction (eight particles for the small, four for medium and two for large
size fraction) were analysed using X-ray CT, and are hence not necessarily representative of the
overall leach trends. To ensure that results from the analysis of the individual particles compare to
those of the overall leach experiments, total Zn conversions have been compiled from a succession
of X-ray CT analyses. The results indicated that Zn conversion data obtained from X-ray CT on the
solid particles follow the overall solution leach curve in trend, but not necessarily in extent and it is
lower (Ghorbani et al., 2013a). The fact is that the calculation method when considering the
individual particles is not explicit. It is by averaging the zinc grade diminution rate in the selected
particles.

RESULTS AND DISCUSSIONS

Mineral conversion from single ore particles
To investigate the progression of leaching on the inside of the individual particles, each single
selected particle was divided into a succession of shells relating to the distance from the centre of
particle, as illustrated in Figure 1. The concentration of Zn grains were determined in each shell and
each time the particle was analysed, thus allowing determination of the progression of Zn leaching
with both time and distance from the surface.

Figure 1 An illustration of the division of each single particle into different shells bases on the distance from
centre of the particle

Using VGStudio MAX, each shell is defined as a Region of Interest (ROI), and due to the irregular
shape of particles, the polygonal surface model tool was used to draw the boundary of each shell
following the shape of each particle. Defining thus the ROI in each, x, y and z, direction of the
image and extracting it with the polygonal surface extraction tool provided by the software,
analysis of each sub-zone or shell as a separate 3-D object became possible. Figure 2 shows the
average Zn grain distribution and Zn grade within selected particles from each size fraction
crushed by HPGR at 95bar.

Figure 2 Average Zn grain distribution and average of Zn grade within selected particles from each size
fraction crushed by HPGR-95bars

Zn percentage in Figure 2 refers to mass percentage of Zn grains located in a given shell. Given that
the volumes of the shells are not the same, but become smaller in the centre of the particles, similar
percentages of Zn in each shell translate into different zinc grades (or concentrations). As can be
seen, counter to a key assumption for using the shrinking core approach, the sphalerite mineral
grains are not homogeneously distributed, but tend to be concentrated in the particle centre. Figure
3 shows the average Zn conversion as a function of distance from the centre of the particle in the
selected ore particles, produced by the two different comminution devices, and in the different size
fractions.

HPGR

Cone crusher

Figure 3 Comparison of the average conversion X (Zn) over time (from start of the experiment) at different
positions within an ore particle crushed by HPGR-95bar and cone crusher for different size fractions

This analysis clearly shows that leaching from large particles approaches complete conversion near
the surface, but only partial conversion in the zones that are closer to the centre of particles. The
cores of the cone-crushed particles show hardly any conversion at all, especially in the larger
particles.

Crack depth and surface zone

The higher crack density in the particles crushed by HPGR compared to those produced by the
cone crusher (Ghorbani et al., 2013b) provides a larger initial surface zone for reaction. In the HPGR
product, the reaction zone covers both the particle surface and subsurface zone, which has access to
the surface through cracks. In some particles, even grains close to the centre of the particles
dissolved early if they were connected to particularly deep cracks. In the case of the cone crusher
product, the outer surface of the particle is the main reaction surface and reaction occurs in the
subsurface zones to not nearly the same depth as the particles crushed by HPGR. This implies the
existence of micro-cracks in these particles, which were not detectable by the visual techniques used
in the present study. Reaction near the surface of particles leads to partial overall conversion in the
coarse particles. After this initially rapid phase, leaching slows down dramatically. This is
explained by the fact that after depletion of the easy-to-leach minerals near the surface, reagents
have to migrate deeper into the particles via the pore network. This will result in a rim-leaching
effect in larger particles. All minerals near the surface are depleted first before the leach front can
migrate further into the particle. On the inside of the particle, the rate is then controlled by the rate
at which reagent (ferric iron in the present case) diffuses deeper into the particle.
Figure 3 clearly shows that HPGR-crushed ore leached more rapidly than the cone-crushed material
in all particle size classes, and to slightly greater depth. Table 3 shows the penetration depth (Pd)
estimated from Figure 3, as well as average crack depth from the surface of the selected particles of
different size fraction as a function of the crushing device.
Table 3 Estimated penetration depth (Pd) and average crack depth of the selected particles after
crushing in different size fraction (Nd denotes not detected)
Size fraction (mm)

Mean R

Average crack depth (mm)

Penetration depth (Pd) (mm)

cone crusher

HPGR

cone crusher

HPGR

Small (-6.75+5.25)

3.0

Nd

1.08

2.0

3.0

Medium (-16+14)

7.5

Nd

3.43

4.5

6.0

Large (-25+23)

12.0

Nd

6.35

6.0

8.0

It is interesting to note that measured crack and inferred penetration depths for HPGR consistently
differ to the order of around 2mm. This is an indication that cracks detectable by X-ray CT are
generally larger than those needed for reagent penetration during leaching. It is therefore safe to
say that also cone-crushed particles exhibit a network of micro-pores near the surface, even though
they are not detectable by X-ray CT. By extension, it is likely that also HPGR crushed particles have
a network of such micro-pores deep inside the particles extending beyond the rim of larger pores
detectable by X-ray CT.

The results presented in Figure 4 indicate furthermore that the penetration depth (Pd) is as function
of both the comminution mode and the particle size in terms of radius (R). It is clear that prevalence
of cracks and particle porosity play key roles in increasing the penetration depth, and as these
characteristics are more pronounced in HPGR-crushed particles, the conversion front clearly
penetrates more deeply in these. As can be seen from the trend line equations, the correlation
between the penetration depth (Pd) and particle radius (R) can be represented as a power law as
per Eq. (1):
= aR

(1)

With b = 0.7-0.8 more or less the same for the products of the two different comminution devices,
the only difference is the value of a, which is 1.4 for the particles crushed using HPGR and 0.8 for
particles crushed using cone crusher. The exponential relationship could be related to the
mechanics of the crack propagation, with the exponent b as a material related parameter (since it is
more or less the same for both sets) and the parameter a related to the comminution mode.

Figure 4 The penetration depth (Pd) vs mean particle radius in the HPGR-95 bars and cone crusher product

Three stages leaching in large particles

Zn extraction data generated in as set of bioleaching experiments in circulating flow fixed-bed
reactors extending over 11-months as described by Ghorbani et al. (2012) were analysed using
shrinking core model, reaction-diffusion model, K- model (Bouffard, 2003; Bouffard and Dixon,
2001). An extended K- model was formulated as appropriate modelling approaches to large
particle leaching. Further detail of fitting data with models has been discussed in Ghorbani et al.
(2013a).
The application of the diffusion controlled shrinking core model and surface reaction controlled
shrinking sphere model to the leaching of zinc in the leach liquors from the different reactors,
clearly showed that leaching from large particles contrary to the common assumption does not

follow either of the two models. Results indicated that in fact there are three different stages during
the leaching process. The first stage is leaching of grains at the surface of the ore particles that
extends over the first 50 days, followed by a second stage, which is leaching of the grains located in
the subsurface within the thickness of the penetration depth. A clear inflection point after which the
slope changes, lies between days 90 and 120 for the HPGR-crushed particles (with larger particles
tending to show the inflection later) and between days 120 and 150 for the cone-crushed material. It
is postulated that this point relates to the depletion of mineral grains directly accessible from the
surface of particles (via cracks); thereafter reagent has to penetrate deeper through micro-pores.
Thus in the cracked rim, there is a rapid penetration of leaching reagent through the crack network,
but slow leaching from zones between cracks. This is illustrated in Figure 5a. This mechanism
would also explain why the HPGR profiles in Figure 3 are quite flat, showing rapid penetration to
a certain depth, but only slow and quite homogeneous leaching thereafter. Cone-crushed particles,
which have few cracks (as indicated in TABLE 4) would not show this behaviour.
This second stage is a combination of diffusion and reaction, which proceeds up to a certain
penetration depth (Ghorbani et al., 2013a). After depletion of mineral grains that can be accessed by
this second stage process, a much slower mechanism governs leaching from the inner core. This is
shown schematically in Figure 5b. It should be noted that there is not enough data for this third
phase leaching as the experiments were not continued for long enough. It is thus not possible to
determine what mechanism this follows, but the trends the overall solution leach curves (Ghorbani
et al., 2012) indicate that the last stage of leaching can be approximated as linear. Given the timescale of a typical heap leaching process, a linear model to describe this slow residual leaching is
acceptable.

Figure 5 Schematic cross section of a large particle, a. partially leached particle with the effective penetration
depth (Pd), b. two different leach regime

CONCLUSION
Investigation of mineral conversion within single ore particles through X-ray CT indicated that
leaching from large particles leads to near complete conversion near the surface, but only partial
conversion in the grains that are closer to the centre of particles. After all minerals near the surface
are depleted, the leaching regime changes from the readily accessible grains on the particle surface
ore exposed to large cracks to the more inaccessible grains within particles. In the HPGR product,
the reaction zone thus covers both the particle surface and a relatively deep subsurface zone, which
is connected to the surface through cracks. In the case of the cone crusher product, the outer surface
of the particle is the main reaction surface with only limited reaction occurring in the subsurface
zones and not nearly as deep as the particles crushed by HPGR. This implies the existence of microcracks in these particles, which are not detectable by the techniques used in this study. The results
indicate that the penetration depth (Pd) is a function of both the comminution mode and the
particle size/radius (R) and can be described by a similar power-law correlation. It is clear that
prevalence of cracks and particle porosity play key roles in increasing the penetration depth, and
these characteristics are more pronounced in HPGR-crushed particles.
Application of neither the shrinking core nor the shrinking sphere models is appropriate to describe
leaching. Three different stages during the leaching process were identified. The first stage refers to
leaching of grains at the surface of the ore particles, followed by a second stage, which is leaching of
the grains located in the subsurface within the thickness of the penetration depth. A third stage
relates to very slow, linear leaching from uncracked zones within the particle, primarily the inner
core.

ACKNOWLEDGEMENTS
The authors are grateful to Paul Keanly (X-Sight X-ray Services), and the Centre for Advanced
Scanning Technologies (Department of Physic, UCT) for their advice and support. Financial
support from the South Africa Research Chair Initiative (SARChI) in Mineral Beneficiation, and a
Research Niche Area (RNA) grant from the National Research Foundation (NRF) of South Africa
are also acknowledged.

REFERENCES
Bouffard, S. C. and Dixon, D. G., 2001, Investigative study into the hydrodynamics of heap leaching processes.
Metallurgical and Materials Transactions B., 32, 763-776.
Bouffard, S.C., 2003, Understanding the Heap Biooxidation of Sulfidic Refractory Gold Ores. PhD dissertation
University of British Columbia.
Ghorbani, Y., Becker, M., Mainza, A.N., Franzidis, J-P., Petersen, J., 2011a, Large particle effects in
chemical/biochemical heap leach processes - A review, Minerals Engineering, 24, 1172-1184.
Ghorbani, Y., Becker, M., Petersen, J., Mora, S. H., Mainza, A.N, Franzidis, J-P., 2011b, Use of X-ray computed
tomography to investigate crack distribution and mineral dissemination in sphalerite ore particles,
Minerals Engineering, 24, 1249-1257.
Ghorbani, Y., Petersen, J., Harrison, S.T.L., Tupikina, O. V., Becker, M., Mainza, A. N., and Franzidis, J-P., 2012,
An experimental study of the long-term bioleaching of large sphalerite ore particles in a circulating
fluid fixed-bed reactor, Hydrometallurgy (129130), 161-171.

Ghorbani, Y., Petersen, J., Becker, M., Mainza, A.N., Franzidis, J-P., 2013a, Investigation and modelling of the
progression of zinc leaching from large sphalerite ore particles, Hydrometallurgy (131-132), 8-23.
Ghorbani, Y., Mainza, A.N., Petersen, J., Becker, M., Franzidis, J-P., and Kalala, J.T., 2013b, Investigation of
particles with high crack density produced by HPGR and its effect on the redistribution of the particle
size fraction in heaps, Minerals Engineering, Minerals Engineering (43-44), 44-51.
Liddell, K. C., 2005, Shrinking core models in hydrometallurgy: What students are not being told about the
pseudo-steady approximation, Hydrometallurgy, 79, 62-72.
Miller, G., 2003, Ore geotechnical effects on copper heap leach kinetics. TMS (The Minerals, Metals and
Materials Society), 329-342.
Rossi, G., 1990. Biohydrometallurgy. McGraw-Hill, New York.
Vegli, F., Trifoni, M., Pagnanelli, F. and Toro, L., 2001, Shrinking core model with variable activation energy:
a kinetic model of manganiferous ore leaching with sulphuric acid and lactose. Hydrometallurgy, 60,
167-179.
Velardo, A., Giona, M., Adrover, A., Pagnanelli, F. and Toro, L., 2002, Two-layer shrinking-core model:
parameter estimation for the reaction order in leaching processes. Chemical Engineering Journal, 90,
231-240.

4D automatic quantification of ore particles in heap

leaching from X-ray micro computed tomography

Qingyang Lin and

Stephen Neethling
Imperial College London,
United Kingdom
Katherine Dobson,
Loic Courtois
and Peter Lee
Rutherford Appleton
Laboratory, Didcot,
United Kingdom

Heap leaching is an important hydrometallurgical technology, which

extracts valuable metals such as gold, silver, copper, from low grade ores.
The use of X-ray micro computed tomography (XMT) to study the internal
mineralogy of ore particles in a non-destructive manner has in recent years
gained popularity. In this paper it will be shown how a small leaching
column can be repeatedly scanned at different time points to allow both
the spatial and temporal evolution of the mineralogy to be tracked as it
undergoes leaching (4D analysis).
In particular, the paper will show how the evolution of the thousands of
individual mineral grains within the ore particles can be tracked over time.
There are a number of challenges that needed to be overcome to carry out
this quantification. For instance, the individual ore particles needed to be
identified, which was easy early in the leaching cycle, but became more
challenging as deposits of, for instance, jarosite appeared between the
particles. The particles in the column also shifted and rotated over the
course of the leaching, which needed to be accounted for in the comparisons
between time steps. The same mineral grains across different images also
needed to be identified, which is complicated by the fact that they change
size and can even disappear. All of these steps have been fully automated.

INTRODUCTION
Global demand for metals has been increasing rapidly in recent years, but this has been
accompanied by a decline in the high grade ore resources. For example in the United States the
average head grade of copper ore has declined from 2.5% in 1930 to approximately 0.5% in 2000
(McKinney et al., 2007; Wills & Napier-Munn, 2006). As the grade of ores decreases, heap leaching is
becoming a more attractive alternative to conventional processing routes such as flotation followed
by smelting. The biggest disadvantages of heap leaching, though, are the long processing time and
relatively low extraction efficiencies, especially for primary sulphides. Most studies of heap
leaching have concentrated on either column experiment by studying leaching performance as
function of fluid flow around ore particles (Ilankoon & Neethling, 2013) and different feed
conditions (e.g. van Hille et al. (2010) or obtaining the surface reaction kinetics by studying the
dissolution of finely milled ore or pure mineral particles in stirred tank experiments (Crdoba et al.,
2008, 2009; Hiroyoshi et al., 2001; Koleini et al., 2011). There is a lack of information on the transport
processes within an ore particle, especially the balance between the effect of mass transport
between the particle surface and the mineral surface and the surface reaction rate.
The most common technique to estimate this information is to make an assumption about the leach
behavior, such as that it fits the shrinking core model, and then fit the macroscopic apparent leach
kinetics to the model to estimate the mass transport and surface reaction contributions to the
behavior. A more direct analysis would require measurements of the behavior at the scale of the
individual mineral grains within the ore particles. Mineral grain level analysis of these ores has
traditionally relied on thin sections, which provides a lot of information on the mineralogy, but is
destructive and therefore cannot be used to track the evolution of individual mineral grains as they
undergo leaching.
X-ray micro computed tomography (XMT) is a non-destructive and non-invasive technique that
allows the same samples to be scanned repeatedly and gives three dimensional visualization of the
internal structure (Ketcham & Carlson, 2001; Puncreobutr et al., 2012; Stock, 1999, 2008). The sample
is placed within an X-ray beam, and rotated through 360 while a series of micron level resolution
2D radiograph projections are collected (Stauber & Mller, 2008). This series of 2D radiographs are
then reconstructed into a 3D volume image. By using this technique, the column leaching process
can be imaged in situ to give 4th dimension in analysis, time, without affecting the leaching
behavior. There have been a number of minerals processing and heap leaching studies that have
made use of XMT, though most of them have focused on either average changes in different phase
volume fractions (e.g. mineral phase or porosity) or have done more qualitative analysis of for
instance textures or fracture distributions (Burlion et al., 2006; Ghorbani et al., 2011; Ketcham &
Carlson, 2001; Kodali et al., 2011; Lin & Miller, 2005). What none of these studies do is mineral grain
level tracking of the leaching behavior over a number of different time points.
The main objective of this paper is to introduce consistent algorithms, which can be fully
automated, to quantify leaching performance by using imaging techniques. The experiments on
which these algorithms are applied come from a small scale leaching column with 28mm internal
diameter and 190mm height. The leaching solution is 0.1M H2SO4 with 5g/L Fe3+. The column
leaching is carried out in a temperature controlled incubator at 60C. The column was scanned
using a Nikon Metris Custom Bay with a 1mm aluminum filter to reduce the noise level, 89kV
energy, 0.708s exposure time and 2001 projections. The detector size is 20002000 pixels which gives
a linear resolution of approximately 17 microns.

4D quantification and its automation methodology

In this section the steps required to firstly identify, match and register the ore particles between
different time steps are outlined, followed by the algorithm by which the individual mineral grains
within the ore particles are identified and tracked. Figure 1 shows the main steps and difficulties
carrying out efficient and consistent 4D image quantification, as well as the software used.

Figure 1 Main steps and difficulties in order to carry out efficient and consistent 4D image quantification

Image pre-processing
After each 3D attenuation image was produced, the data had a 333 median filter applied to
reduce the noise level. The data sets were then normalized based on the intensity of the background
peak, I1, and the intensity of the ore peak, I2 (the intensity of background peak is always smaller).
The normalized intensity for each pixel, INew can be calculated as:

I new

( I 2 I ) I Low ( I I1 ) I High
I 2 I1

(1)

where I is the original voxel intensity, ILow is the reference intensity for background peak and IHigh is
the reference intensity for ore peak.

Thresholding, segmentation and labelling of ore samples

Global thresholding based on the histogram can be used to threshold between different phases
when there are distinct peaks in the histogram. This is case for the air and the gangue rock (the
glass column has essentially the same grey value as the gangue within the ore particles), where the
peaks are quite distinct (see Figure 2). However, there is no clear peak for metal sulphide (MS)
particles as its peak value has similar counts as the noise within the ore phase. More advanced
methods need to be used, which will be discussed in later sections.

Figure 2 The histogram of a tyical scanning volume with estimated thresholding values for ore and metal
sulphides

Histogram based global thresholding using the Otsu algorithm (Otsu, 1979) was used to threshold
ore samples from background air as a consistent method, followed by fillings holes, to generate the
binary image of all ore particles (Figure 3a and Figure 3b). The connected ore particles are then
separated using a distance map and marker control based 3D watershed algorithm. This algorithm
expands the region according to a priority map, which is normally a 3D distance map, until the
regions touch the watershed lines. The algorithm also merges local maxima marker-control to avoid
over segmentation. Flood-filling is then used to assign a unique label to the connected regions in
each particle (Figure 3c) uses a flood-fill algorithm with 18 neighbors (MathWorks, 2013).

Figure 3 An example slice from a volume data showing its histogram, binary image after thresholding and
image after watershed separation

Particle tracking and registration of ore particles

To make all quantification fully automated, one of the most difficult part is particle tracking. When
scanning the same volume at different time points, the particles change location due to both the ore
particle movement as they settle over time and slightly different orientations of the column when
placed onto the sample holder. The location of each particle may be rotated and/or shifted between
time points. After labelling, the same particle in different scans may therefore have different label
numbers. Therefore, particle tracking needs to be carried out to find the same ore particle in
different scans so that it can be relabeled to have the same label number across the entire time
sequence. The centroid tracking algorithm, which is based on the calculation of the center of mass
(centroid) of each particle, is used to track each ore particle in different scans (Blair & Dufresne,
2008; Crocker & Grier, 1996). The mass centroids of individual particles in different scans are
calculated separately. By setting up an estimated maximum distance for centroids to move, the new
location of each centroid was matched up by looking for most likely correspondence in later scan.
Each tracked ore particles in different scans was then registered (or aligned) to the same orientation
using Day 0 as a reference (Studholme et al., 1999). This procedure is shown in Figure 5. Good
registration (or alignment) of the particles is the basis of the fast mineral grain tracking, which is
described later.

Figure 4 An example of ore particle tracking and relabeling of a volume in 2 different scans

Figure 5 Registration example of an ore particle after 136 days leaching

Thresholding of metal sulphide grains

From Figure 2 it can been seen that there is no distinction between the metal sulphide and gangue
peaks in the histogram and thus no obvious threshold value to choose. The main reason for no clear
peak is that the total volume of metal sulphides compared with the volume of the ore particle is
very small and the metal sulphide peak is actually hidden in the tail of ore peak. A number of
different thresholding methods were tested using the auto threshold plug-in in ImageJ (Fiji, 2013),
which are shown in Figure 6. Maximum entropy, Renyi entropy and Yens methods are developed
from the idea of the maximum entropy indicating the best separation of phases (Kapur et al., 1985;
Yen et al., 1995). According to Gonzalez et al. (2003) and MathWorks (2013), entropy is a statistical
measure of randomness that can be used to characterize the texture of the input image, which is
defined as:

h(i )

H (t )
i 0

h( j )
j 0

log

h(i )
t

h( j )

(2)

j 0

where H is the calculated entropy, t is the threshold value, h(i) is a normalized histogram and i is
the intensity value. The optimal threshold can be calculated by maximizing the sum of the black
(non-sulphide) and white (sulphide) entropies. While using a more advanced method for obtaining
the threshold, the maximum entropy methods are still global. Local thresholding algorithms (e.g.
3D watershed based) can also be used to threshold sulphides. Figure 7 shows a comparison of
Maximum entropy and 3D watershed algorithms for identifying sulphide grains.

Figure 6 Initial investigation of different global thresholding algorithm for metal sulphide grains in ImageJ

Figure 7 The comparison of different methods to threshold MS particles using Max Entropy and 3D watershed

The 3D watershed algorithm sometimes cannot pick small particles (Shown in red) and, due to its
gradient sensitivity, it will also tend to over-estimate the size of very bright particles. Local methods
are also typically computationally more expensive than global methods. On the other hand, local
methods are less sensitive to artefacts such as beam hardening and can be more easily tuned to
ignore noise in the image.

Metal sulphide grain tracking

The tracking of the metal sulphide grains is not done using the same centroid based tracking
algorithm that was used to track the overall ore particles. There are a number of reasons for not
using this algorithm:
Firstly, the sizes of each mineral grain changes over the course of the experiments due to
leaching (it is actually this change in size that is of most interest to us), which means that
the centroid will change position.
The mineral grains do not only shrink, but can also disappear. This causes problems for the
centroid based algorithm, which will typically identify another particle that has not
disappeared as the matching particle.
A final consideration is the speed of the algorithm, which needs to be fast given the very
large number of mineral grains involved.
The algorithm developed for this work starts by identifying all the metal sulphide grains in the
initial image. The connectivity of the grains is analyzed so that each isolated grain is given a unique
identifier. On subsequent images, voxels that are identified as metal sulphide grains are given the
same identifier as the voxel in at the same location in the initial image by applying the initial
identifier label as mask. Due to the variation of scanning quality and uncertainty caused by
thresholding algorithm, rims of mineral voxels that have not been unassigned an identifier can
occur at the edges of grains, especially those that have undergone little size reduction due to

leaching. This problem is resolved by assigning these rim voxels the identifier of a neighboring
metal sulphide voxel which has an identifier. The advantages of this method is that particles that
disappear can be easily identified and, where leaching causes a grain to split into a number of
smaller child grains, each of these grains will have the same identifier as the initial grain (Figure 8).
Figure 9 shows the isosurface for an example grain after tracking through time.

Figure 8 An example slice of results after metal sulphide particle tracking

Figure 9 Mineral grain tracking for an example grain during 136 leaching days

CONCLUSION
In this paper, the details of an automatic quantification algorithm for tracking particles undergoing
leaching were introduced. Each quantification step, including thresholding, tracking, labelling and
registration for both the ore particles themselves as well as the mineral grains have been introduced
and explained. The ore thresholding uses Otsus method while the metal sulphide grains
thresholding uses the Maximum Entropy algorithm. For particle tracking, ore particles in different
scans can be tracked using the centroid of each particle as input. Metal sulphide particle tracking is
achieved by using labelled initial metal sulphide identifier mask and calculate the value for
unsigned voxels.

ACKNOWLEDGEMENTS
This study was performed in the Rio Tinto Centre for Advanced Mineral Recovery at Imperial
College London. The authors gratefully acknowledge Rio Tinto for their financial support for this
project.

REFERENCES
Blair, D., Dufresne, E., (2008) The Matlab Particle Tracking Code Repository,
<http://physics.georgetown.edu/matlab/>.
Burlion, N., Bernard, D., Chen, D., (2006) X-ray microtomography: Application to microstructure analysis of a
cementitious material during leaching process, Cement and Concrete Research, vol. 36, pp. 346-357.
Crdoba, E.M., Muoz, J.A., Blzquez, M.L., Gonzlez, F., Ballester, A., (2008) Leaching of chalcopyrite with
ferric ion. Part II: Effect of redox potential, Hydrometallurgy, vol. 93, pp. 88-96.
Crdoba, E.M., Muoz, J.A., Blzquez, M.L., Gonzlez, F., Ballester, A., (2008) Passivation of chalcopyrite
during its chemical leaching with ferric ion at 68C, Minerals Engineering, vol. 22, pp. 229-235.
Crocker, J.C., Grier, D.G., (1996) Methods of Digital Video Microscopy for Colloidal Studies, Journal of Colloid
and Interface Science, vol. 179, pp. 298-310.
Fiji, 2013. Auto Threshold, <http://fiji.sc/wiki/index.php/Auto_Threshold#>.
Ghorbani, Y., Becker, M., Petersen, J., Morar, S.H., Mainza, A., Franzidis, J.P., (2011) Use of X-ray computed
tomography to investigate crack distribution and mineral dissemination in sphalerite ore particles,
Minerals Engineering, 2011, vol. 24, pp. 1249-1257.
Gonzalez, R.C., Woods, R.E., Eddins, S.L., (2003) Digital Image Processing Using MATLAB, Prentice-Hall, Inc.
Hiroyoshi, N., Miki, H., Hirajima, T., Tsunekawa, M., (2001) Enhancement of chalcopyrite leaching by ferrous
ions in acidic ferric sulfate solutions, Hydrometallurgy, vol. 60, pp. 185-197.
Ilankoon, I.M.S.K., Neethling, S.J., (2013) The effect of particle porosity on liquid holdup in heap leaching,
Minerals Engineering, vol. 45, pp. 73-80.
Kapur, J.N., Sahoo, P.K., Wong, A.K.C., (1985) A new method for gray-level picture thresholding using the
entropy of the histogram, Computer Vision, Graphics, and Image Processing, vol. 29, pp. 273-285.
Ketcham, R.A., Carlson, W.D., (2001) Acquisition, optimization and interpretation of X-ray computed
tomographic imagery: applications to the geosciences, Computers & Geosciences, vol. 27, pp. 381-400.
Kodali, P., Dhawan, N., Depci, T., Lin, C.L., Miller, J.D., (2011) Particle damage and exposure analysis in
HPGR crushing of selected copper ores for column leaching, Minerals Engineering, vol. 24, pp. 14781487.
Koleini, S.M.J., Aghazadeh, V., Sandstrm, ., (2011) Acidic sulphate leaching of chalcopyrite concentrates in
presence of pyrite, Minerals Engineering, vol. 24, pp. 381-386.
Lin, C.L., Miller, J.D., (2005) 3D characterization and analysis of particle shape using X-ray microtomography
(XMT), Powder Technology, vol. 154, pp. 61-69.
MathWorks, 2013. Image Processing Toolbox Users Guide
<http://www.mathworks.co.uk/help/pdf_doc/images/images_tb.pdf>.
McKinney, M., L., Schoch Robert, M., Yonavjak, L., (2007) Environmental Science: Systems and Solutions. Jones
and Bartlett Publishers.

Otsu, N., (1979) A threshold selection method from gray-level histograms, IEEE Transactions on Systems, Man
and Cybernetics, vol. 9, pp. 62-66.
Puncreobutr, C., Lee, P.D., Hamilton, R.W., Phillion, A.B., (2012) Quantitative 3D Characterization of
Solidification Structure and Defect Evolution in Al Alloys, JOM, vol. 64, pp. 89-95.
Stauber, M., Mller, R., (2008). Micro-Computed Tomography: A Method for the Non-Destructive Evaluation of the
Three-Dimensional Structure of Biological Specimens Osteoporosis, J.J. Humana Press, pp. 273-292.
Stock, S.R., (1999) X-ray microtomography of materials, International Materials Reviews, vol. 44, pp. 141-164.
Stock, S.R., (2008) Recent advances in X-ray microtomography applied to materials, International Materials
Reviews, vol. 53, pp. 129-181.
Studholme, C., Hill, D.L.G., Hawkes, D.J., (1999) An overlap invariant entropy measure of 3D medical image
alignment. Pattern Recognition, vol. 32, pp. 71-86.
van Hille, R.P., van Zyl, A.W., Spurr, N.R.L., Harrison, S.T.L., (2010) Investigating heap bioleaching: Effect of
feed iron concentration on bioleaching performance, Minerals Engineering, vol. 23, pp. 518-525.
Wills, B., Napier-Munn, T., (2006) Mineral Processing Technology. Seventh Edition, Elsevier Science & Technology
Books.
Yen, J.-C., Chang, F.-J., Chang, S., (1995) A new criterion for automatic multilevel thresholding, Image
Processing, IEEE Transactions on, vol. 4, pp. 370-378.

Acid In Situ Leaching (isl) of

phospho-coffinite at Katco

Sergio Bustos,
Diana Mango, Michel
Belires, Frank Ling and
Marc Brouand
Areva Mines, France
Csar Garca
Sarl sm Katco,
Republic of Kazakhstan

Katco, a mining company based in Almaty, was created in 1996 to develop

and mine the Muyunkum and Tortkuduk uranium deposits in southern
Kazakhstan. Its shareholders are 51% Areva and 49% KazAtomProm,
Kazakhstans national natural uranium producer. The selected In Situ
Leaching (ISL) technology for recovering the uranium consists in injecting
a dilute sulfuric acid solution into the deposit through boreholes and
allowing it to circulate to dissolve the uranium-bearing minerals, mainly
uraninite and calcium phospho-coffinite. Ferric ion is required to oxidize
the U(IV) mineral species present in the ore, thus allowing the UO22+ ion to
dissolve in weak sulfuric acid environments. The role of ferric ion as the
oxidant for the uranium in the leaching system makes it necessary to know
how it dissolves and how it evolves in the leaching solutions as a function
of the acid addition in order to maintain the sustainability and the
efficiency of uranium dissolution in the well field.
A concentrated parameters model approach is proposed to represent the
acid ferric leaching reaction system for the dissolution of uranium from
Katco uraninite/phospho-coffinite ores. The uranium dissolution kinetics is
assumed to be proportional to the ferric active concentration at the
ore-solution inter-phase, which results from ferric being added with the
injection solution plus ferric being dissolved from the ore.
Column leach test results as well as leaching performance results from
industrial ore blocks confirm this approach opening an opportunity for
application of the model to anticipate the uranium extraction of new blocks
in the mine production plan, provided proper ore characterization is
available and ISL operating conditions are known.

INTRODUCTION
KATCO is a joint venture between AREVA (51%) the French State owned Uranium Mining,
Processing and Enriching Company, and Kazatomprom (49%), the Kazakhstan State Corporation of
Nuclear Energy Enterprises.
It was created in 1996 to produce uranium concentrate from the exploitation of the uranium
reserves of the Muyumkum ore body located at the Shu-Sarysu geological zone in the Suzak
District, South Kazakhstan Region at about 250 Km north-west from Shimkent.
The ISL mine at the Muyumkum site (MYK) was commissioned in 2006 and at the Tortkuduk site
(TKD) in 2007, since then resulting in a steady increase in uranium concentrate production from
1058 t/year in 2007 to 1367 t/year in 2008 and to 3130 t/year in 2009. This production was expanded
to 4000 t/year in 2013 by increasing the well fields rich solution flow and the associated adsorption
capacity at both existing plants up to 3000 m3/h each.
The uranium mineralization of the Muyunkum and Tortkuduk deposits is contained in a roll-front
type ore body. Such deposits form naturally where groundwater in permeable sandstone or
conglomerate encounters the interface between oxidizing and reducing conditions; the uranium
naturally present in solution in the aquifer precipitates at this interface, commonly forming a
crescent-shaped body. Over the geological time, with the reduction front migrating in the direction
of groundwater flow, the roll-front ore body can extend for hundreds of meters.
The ISL method that is used to mine the low-grade uranium in the vast KATCO Muyunkum
deposit consists in injecting a leaching solution (in this case, dilute sulfuric acid) into the deposit
through boreholes and allowing it to circulate to dissolve the uranium. The pregnant solution is
then pumped out of the deposit and piped to the processing plant where the uranium is extracted
onto a resin. The metal is subsequently stripped from the resin, purified, concentrated and
packaged as ammonium diuranate yellow cake at a conventional ore-processing plant (Bustos et al.,
2010), (Bustos et al., 2010).
ISL techniques have evolved to the point where it is now a controllable, safe, and environmentally
benign method of mining with strict operational and regulatory controls. It results in little surface
disturbance and generates no tailings or waste rock. The aquifer volume affected by the
acidification process can be reclaimed at the end of the leaching time through conventional effluent
treatment procedures or through natural attenuation of pollutants, depending on solution
composition and/or particular local environmental regulation demands (Schmitt, Descostes &
Polak, 2013). Capital costs are low (relative to conventional mining) and it is often the most effective
method for mining low-grade uranium deep deposits. ISL is selective for the recovery of uranium
and does not create any radioactive rock stockpiles or radioactive tailings on the surface, although
relatively small volumes of naturally radioactive residues are generated.
This paper summarizes the advance on the knowledge of KATCO ISL reaction system which
provides basic information to anticipate the ore metallurgical behavior to be considered during the
elaboration of mine production plans.

ACID FERRIC LEACHING OF PHOSPHO-COFFINITE

The three KATCO well fields are geologically similar: uranium is distributed in a sandy formation
isolated between two layers of clay which constitute the upper and the lower limits of the aquifer
where is located the mineralized zone. The uranium mineralogy is mainly represented by uraninite,
coffinite and calcium phospho-coffinite. The last one, can be envisaged as a variable association of
CaO, P2O5, SiO2 and UO2. This mineral species lay mostly surrounded by clayey mineral species,
mainly represented by the smectite as shown in Figure 1.

Figure 1 Mineralogical analysis of a typical ore sample

Other mineral species are also present in variable proportions in the ore, thus leading to acid
consumption and to dissolution of metal ions such as Fe, Al, Mg, K and Ca. Acid is necessary to
dissolve Fe3+ from soluble ferric bearing species in the rock, which is the oxidant for the reduced
uranium species such as uraninite (UO2). The main reactions that take place during acid leaching
are those stated below.
UO2 + 2 Fe3+ = UO22+ + 2 Fe2+
UO2.3CaO.P2O5 + 2 Fe3+ + 3 SO42- = UO2(SO4)34- + 3 Ca2+ + 2 PO43- + 2 Fe2+
3 H2SO4 + 2 FeOOH = Fe22+ + (SO42-)3 + 4 H2O

The ISL concentrated parameters model approach

At an open pit mine operation it is possible to identify, collect, separate and/or blend the main geometallurgical units (GMU) of the ore body to produce a regular and known ore feed to the process
plant, thus ensuring the process performance. This is not possible at an ISL operation system. The
leaching reactor being the mineralized zone of the ore body affected by the leaching solution
percolating through the ore mass includes all the multiple existing ore heterogeneity (particle size
defining effective porosity and permeability, uranium ore grade, gangue composition and
reactivity, etc.). Solution concentration profiles along the solution flow paths within the ore mass
are then characteristic of this complex distributed parameters reaction system.
However, feedback from KATCO operation results suggests that at block level (considered as the
volume element of mine production plan, corresponding to about one million ton of ore), the actual
ore variability is mitigated and the phospho-coffinite-uraninite acid ferric leaching reaction system
can be represented by a simplified concentrated parameter model which assumes that the ore-

solution inter-phase can be defined by using average values for the ore and solution
characterization parameters. Under this approach, the uranium extraction evolution should depend
only on the applied operating conditions, mainly specific flow rate and acid concentration in the
injection solution.
The ISL concentrated parameter model assumes that there are relatively large sectors in the ore
body which can be considered as pertaining to a certain GMU which can be characterized through
average properties, such as the lithology, the uranium mineralogy and grade, gangue reactivity and
the oxidized/reduced character of the hoist matrix. Each GMU should show a particular
metallurgical behavior under the applied operating conditions. Therefore, the ore heterogeneity can
be overcome by considering a block of ore as the mixture of its constituent GMU and the block
behavior can be estimated as an average of behaviors of the constituent GMU.
Ferric ion is required to oxidize the UIV mineral species present in the ore, thus allowing the UO22+
ion to dissolve in weak sulfuric acid environments. The role of ferric ion as the oxidant for the
uranium in the leaching system makes it necessary to know how it dissolves and how it evolves in
the leaching solutions in order to maintain the sustainability and the efficiency of uranium
dissolution in the well field.

Ferric generation
The reaction system at the ISL uranium recovery process at KATCO well fields is characterized by
the oxidation of the uranium bearing species identified in the ore body, mainly represented by
uraninite (UO2) and phospho-coffinite (3CaO.P2O5.UO2.SiO2). The oxidant reagent is the ferric ion,
which dissolves from the ore due to the acid attack. Therefore, the ISL process at KATCO is at the
same time both uranium and ferric leaching process both enhanced by the acid attack to the ore
gangue minerals.
The uranium dissolution rate is dependent on the ferric ion concentration at the ore-solution interphase, and the leaching time is affected by the ferric concentration in the injection solution,
particularly for those ore blocks which have a relatively low ferric soluble grade. This concentration
shows a pseudo steady state level, which is the result of composting the effluent solutions of all
blocks under operation. Therefore, the anticipated knowledge of the ferric evolution during the acid
leaching of an ore block should help in better defining the leaching time of this particular block and
in maintaining the ferric concentration level in the common injection solution of the well field.
Ferric ion evolution in acid leaching systems involves a sequence of processes, which depend on the
acid concentration variation in the system. Soluble iron, mainly coming from ferric hydrated oxides
in the ore dissolves due to acid attack, but ferric ion is stable only at pH < 2. Therefore an initial
acidification stage of the aquifer is carried out before connecting the block to the leach solution
production circuit, thus avoiding Fe precipitation when introducing the injection solution.
Consequently, once the proper acid pH is reached, the rate of Fe3+ generation (r Fe3+ Gen) should be
equal to the rate of Fe3+ dissolution (r Fe3+ Diss) minus the rate of Fe3+ reduction (r Fe3+ RedU) due to
reaction with the UIV mineral species:

rFe 3Gen rFe 3 Diss rFe 3Re dU

(1)

Ferric dissolution
The ferric dissolution from the ore can be associated to the heterogeneous reaction of acid with
ferric soluble species in the ore gangue, such as limonite or goethite. For the goethite this reaction
would be:
2 FeO.OH + 3 H2SO4 = 2 Fe3+ + 3 SO4= + 4 H2O
The rate of ferric dissolution can be represented by:

rFe 3 Diss

d
j[ H ]
dt

(2)

r Fe3+ Diss corresponds to the rate at which the ferric soluble ore grade [kg/t] decays in time, and it is
assumed to be a function of the free acid concentration [H +] in the leaching solution in contact with
the ore, and of the ferric soluble ore grade .
The ferric extraction must follow a sigmoid curve, as it has a value zero at t=0, has a delay due to
the initial pH evolution up to pH 2 when ferric can start dissolving, and reaches a maximum value
* at long leaching times. Therefore,

* *

(1 exp( t ))
(1 exp( t ))

(3)

And the ore soluble ferric grade should evolve during leaching time according to:

(1 ) exp( t )
(1 exp( t ))

(4)

The rate of ferric dissolution from the ore can be then derived as:

rFe 3 Diss

d
(1 ) exp( t )
*
dt
(1 exp( t )) 2

(5)

* represents the Fe3+ soluble grade of the ore and can be obtained through laboratory standard
leaching tests. The kinetic parameters and are function of the acid concentration in the injection
solution, but also of ore lithology and of ferric soluble mineral species.

Ferric generation and Fe3+ active concentration

The rate of ferric generation rFe3+ Gen can be obtained from the Fe3+ balance in the solutions:

G [ Fe 3 ] IN G [ Fe 3 ]OUT rFe 3Gen

d [ Fe 3 ]
dt

(6)

G in m3/t is the solution volume application rate as referred to the block ore mass under leaching
and in m3/t is the solution inventory retained by the ore block. By considering the pseudo steady
state approach, replacing the previous expression for rFe3+Gen of equation (1) and rearranging:

1
1
[ Fe 3 ]OUT ( ) rFe 3 Re dU [ Fe 3 ] IN ( ) rFe 3 Diss
G
G

(7)

The left hand side of this equation is the active ferric concentration [Fe3+]ACTIVE , representing the
actual Fe3+ concentration available at the ore solution inter-phase, which determines the rate of
uranium dissolution.

1
[ Fe 3 ] ACTIVE [ Fe 3 ]OUT ( ) rFe 3 Re dU
G

(8)

The [Fe3+]ACTIVE concentration can be obtained from column leach tests by following the ferric
concentration evolution in the effluent solutions and by adding the equivalent stoichiometric ferric
concentration as calculated from the corresponding uranium concentration.
Equation (7) can be rewritten as:

1
[ Fe 3 ] ACTIVE [ Fe 3 ] IN ( ) rFe 3 Diss
G

(9)

As the ferric concentration in the injection solution can be considered constant, it is possible to
obtain the kinetic parameters and by fitting the ferric dissolution rate expression of equation (5)
to column leach test results.

Uranium dissolution
The uranium dissolution is controlled by the oxidation of UIV contained in the uraninite and
calcium phospho-coffinite by Fe3+ ions. The uranium dissolution reaction can be represented by the
global chemical equation:
UO2 + 2 Fe3+ = UO22+ + 2 Fe2+
This is a heterogeneous reaction occurring at the ore solution inter-phase where the uraninite
active ore surface allows the electron transfer for ferric reduction, thus leading uranium in solution
as uranil ion.
The rate of uranium dissolution rU [kg/td] is the rate at which the uranium ore grade [kg/t] decays
in time, and it is assumed to be a function of the ferric ion concentration in the leaching solution in
contact with the ore, i.e., the Fe3+ active concentration, and of the uranium ore grade :

rU

d
k [ Fe 3 ] ACTIVE
dt

(10)

By rearranging and integrating this expression, it is possible to obtain the uranium extraction
evolution U:

U (1 exp(k [ Fe 3 ]

ACTIVE

dt ))

(11)

MODEL VERIFICATION
Several test work programs have been performed both at the site as well as at SEPA (AREVAs test
work facility at Bessines, France) by using small laboratory columns of about 1 kg ore, to determine
the metallurgical behavior of ore samples from KATCO well fields. The experimental conditions
have been defined in order to simulate the operating conditions applied at the site particularly in
terms of injection flow and acid and Fe 3+ concentrations in the injection solution. The ore samples
have been prepared by composting selected sections of diamond drill cores which represent the ore
mass within a block to be affected by the acid ferric leach solution. Figure 2 shows a view of the
experimental set-up.
Pumping system

Column loaded with ore

Automatic sampler

Figure 2 Experimental set-up for laboratory column leach tests

Figure 3 shows the results obtained for a column charged with an ore sample assaying 894 g/t of
uranium, and irrigated at 5 L/hrm2 with an acid solution similar to that observed at one of the
KATCO well field containing 12 g/L H2SO4 and 0,1 g/L Fe3+. Typical chemical assays from irrigation
and effluent solutions from the columns have allowed assessing the uranium and ferric evolution in
time. Typical sigmoid uranium extraction evolution is observed. The ferric generation evolution
shows a relatively long period of negative generation which is associated to the initial change in
ferric concentration of the solution retained by the ore porosity plus the ferric reduction due to
reaction with uranium reduced species. At the end of the leaching time, the ore under examination
shows a positive ferric balance which should contribute to maintain the ferric concentration in the
common well field injection solution.

Acid ferric leach results

Column Test
U ext and Fe3+ Gen, %

100
80
60

U extraction, %

40
Fe3+ Generation,
%

20
0
-20

0,0 0,2 0,4 0,6 0,8 1,0

Normalized time

Figure 3 Column leaching test results

Mineralogical analysis of column leach residues, confirm significant gypsum formation during
leaching of this particular ore sample. Gypsum is normally surrounding the calcite as shown in
Figure 4; the gypsum crystallization seems to have occurred very fast producing a relatively dense
layer around the initial calcite grains, thus probably passivating the calcite solution reactive interphase. But it also occupies the inter-granular space between quartz grains initially occupied by
clays, such as smectite, thus indicating the reactivity of clayey material and their contribution to the
acid consumption and probably to ferric dissolution from beidellite, clay pertaining to the smectite
family.

Figure 4 Mineralogical analysis of column leach test residues

The resulting column leach data have been processed in order to obtain the actual ferric active
concentration by applying equation (8). Fitting of the parameters of equation (9) to this
experimental result allowed obtaining the parameters *, and . Further integration of the
[Fe3+]ACTIVE concentration and adjusting of the parameter k in equation (11) allowed obtaining the
uranium recovery evolution curve shown in Figure 5.

Acid ferric leach results

Column Test

0,600

100

0,500

80

0,400
[Fe3+]ACTIVE,
actual

0,300
0,200

[Fe3+]ACTIVE
Model

0,100

U extraction, %

[Fe3+]Active, %

Acid ferric leach results

Column Test

60

U extraction, %

40
U extraction
model, %

20
0

0,000
0,0 0,2 0,4 0,6 0,8 1,0

0,0 0,2 0,4 0,6 0,8 1,0

Normalized time

Normalized time

Figure 5 Column leach test results. [Fe3+]ACTIVE and Uranium recovery evolution

A similar verification procedure has been applied to block TSU-20 from of one of the KATCO well
fields. Once the actual [Fe3+]ACTIVE concentration evolution was obtained, the same fitting procedure
has been applied to the parameters of the ferric active equation (9) and to the parameter k in
equation (11), to obtain the corresponding ferric active and uranium evolution curves as predicted
by the model. The results of this procedure are shown in Figure 6 and are represented by the
equation:

U [%] 100 1 exp 3,53 t

1 exp(4,91 t 0,54

4,91

[Fe3+]Active
Block TSU-20
0,300
[Fe3+]ACTIVE
kg/m3

0,150

[Fe3+]ACT
Model kg/m3

0,100
0,050
0,000
0,0

0,2

0,4

0,6

Normalized time

0,8

U extraction, [%]

[Fe3+]ACT, kg/m3

0,350

0,200

1 150 exp(24,54 t

(12)

U Extraction
Block TSU-20

0,400

0,250

1 exp(24,54 t

100
90
80
70
60
50
40
30
20
10
0

U extraction %
l(1-exp(int([Fe3+]act))
U Model 2

0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7

Normalized time

Figure 6 ISL results from block TSU-20. [Fe3+]ACTIVE and Uranium recovery evolution

Similar results are found when applying the same procedure to other blocks from the same well
field, but showing different uranium initial ore grade varying from 0,150 0,400 kg/t, as shown in
Figure 7.

U extraction, %

U extraction evolution
Blocks from TSU
U rec TSU-06
Actual, kg/t

100
90
80
70
60
50
40
30
20
10
0

U rec TSU-10
Actual, kg/t
U rec TSU-20
Actual, kg/t
U rec TSU-06
Model, kg/t
U rec TSU-10
Model, kg/t
0,0

0,2

0,4

0,6

0,8

Normalized time

1,0

U rec TSU-20
Model, kg/t

Figure 7 ISL operational results of blocks from KATCO Tortkuduk South well field

The observed deviations of adjusted curves from the actual uranium recovery evolution curve
reflect the inherent ore heterogeneity through out the ore body. The parameters included in the
equations of the model which have been adjusted to fit the experimental results are related to ore
characterization parameters, such as the lithology (particle size distribution), the uranium
mineralogy, the acid consumption (grade of carbonate and of other gangue mineral species which
consume acid) uranium grade, ferric soluble grade. The fitting procedure as performed before
considers that a block is homogeneous, while it might have several GMU in different proportions,
some of them being more reactive than others, thus affecting the actual uranium extraction
evolution. However, the observed tendency is quite well reproduced, which suggest that the actual
behavior is more affected by changes in the operational conditions.
Entrainment of fine suspended solids in the leach solutions promote plugging of injection and
production screens, thus slowly reducing the flow. Cleaning of wells screens allows improving this
situation but obliges to stop and restart with the associated flow perturbation.
Acid concentration is also an operational variable which when increased can enhance the uranium
dissolution but at the cost of higher acid consumption. Acid concentration is usually adjusted
according to the acid consumption response of the ore in a particular block. Usually it is high at the
beginning to be reduced by the end of the leaching time to minimize the acid loss in the final
solution inventory retained by the ore. These acid concentration adjustments at the injection
solution produce perturbations to the uranium extraction evolution which cannot be assessed by
the model.
Other operational variables, such as re-drilling of wells within a block and normal replacement of
depleted blocks by new ones produce changes of solution composition introducing solution
inventory changes and perturbations to the solution mass balances, which are on the base of the
uranium extraction curves.
However, the uranium extraction curves obtained by fitting the model parameters to the actual
extraction results seem to reproduce well the metallurgical response of the ore under the applied
leaching conditions. A thorough ore characterization and testing program should help to assess the

ore variability, to identify the main GMU within the ore body and to determine the dependency of
model parameters on the ore characterization parameters. The model should be then applicable to
anticipate the ore behavior and to perform uranium production estimates.

CONCLUSIONS
The ISL process at KATCO is both uranium and ferric leaching process enhanced by the acid attack
to the ore gangue minerals. Feedback from industrial operating results suggests that the actual ore
heterogeneity leading to a complex distributed parameter reaction system can be mitigated when
dealing with ore blocks of about one million ton, which are considered as the volume elements for
the mine production plan estimates.
Therefore, a concentrated parameters model approach has been proposed to represent the acid
ferric leaching reaction system for the dissolution of uranium from KATCO uraninite/phosphocoffinite ores. The uranium dissolution kinetics is assumed to be proportional to the ferric active
concentration at the ore-solution inter-phase, which results from ferric being added with the
injection solution plus ferric being dissolved from the ore.
Column leach test results as well as leaching performance results from industrial ore blocks confirm
this approach opening an opportunity for application of the model to anticipate the uranium
extraction of new blocks in the mine production plan, provided proper ore characterization is
available and ISL operating conditions are known.

ACKNOWLEDGEMENTS
The authors want to thank the collaboration KATCO Mining Operation Department for their
support and collaboration in discussing the ISL operating conditions and performance.

REFERENCES
Bustos, S. Schnell, H., Rincon, M., Rey, S. (2010) Yellow cake production from In Situ Recovery Operation at
KATCO, Hydroprocess 2010, Santiago, Chile, 11-13 August 2010, Chapter IV, Article 42.
Bustos, S. Simon-Coll, Y., Schnell, H., Rincon, M., Rey, S. (2010) Uranium ionic exchange performance at
KATCO, Uranium 2010, 3rd International Uranium Conference, Saskatoon, Canada, August 15-18
2010, Vol II, pp. 39-50.
Schmitt, J.M., Descostes, M., Polak, C. (2013) Lexploitation par In Situ Recovery (ISR) des gisements
duranium de type Roll front: des interactions multiples avec les eaux souterraines, Geologues
no.179, December 2013.

Application of unconventional methods to the

treatment of leachates contaminated with
heavy metals

Mara Ruiz,
Alicia Guevara and
Ernesto de la Torre
Escuela Politcnica
Nacional, Ecuador

This paper addresses the evaluation of the application of unconventional

methods to the treatment of leachates contaminated with heavy metals,
products of the uncontrolled disposition of steel dust on soil. For the
development of the methodology presented in this manuscript, columns
of glass (6 cm diameter) were filled with 15 cm of steel dust on top of 50 cm
of soil. The columns were irrigated with constant fluxes of water
equivalents to the average precipitations of the Ecuadorian Highland
region (0.15 mL/h). To avoid the wall effect, fiberglass (6 cm diameter, 0.8
mm thickness) was located on the solid, in order to distribute the water
uniformly throughout the transversal section. Samples of the generated
leachates were collected weekly for four months to determinate the heavy
metal content. Thereby, it was found that the effluents exceeded the
maximum Ecuadorian limits for concentrations of chromium and arsenic
that may be discharged into freshwater bodies. Afterwards, the following
depuration methods of the obtained liquid were employed: (i)
Stabilization of the polluted leachate with Portland cement and lime, (ii)
depuration of the contaminated liquid with a combination of the
electrocoagulation and phytoremediation processes, and (iii)
recuperation of economically valuable attractive metals with electrolysis
prior to the leachate treatment by conventional methods. The preliminary
economic evaluation of each procedure resulted in operating costs of US$
401.74, US$ 1.57 and US$ 11.84 respectively. However, the low concentration
of heavy metals, organic matter and zinc in the leachates generated by
steel powder and percolated through soil does not justify the application
of the treatment methods considered in this work, and it is recommended
to use conventional techniques.

INTRODUCTION
Improper disposal of hazardous wastes can lead to soluble heavy metals accumulating in soils,
breaking the natural balances in the environment and polluting surface and underground aquifers.
This in turn is damaging to the food chain. Environmental oxidation of metals exposed to the outdoors leads to dissolution of the metals and the acidification of the soil, and the movement of the
unconsolidated material affects surface water and groundwater. (Bautista, 1999) Heavy metals are
toxic, usually persisting in the environment for decades, and their removal is extremely difficult
due to their strong retention in soil exchange points. Consequently, heavy metal soil contamination
is one of todays biggest environmental challenges. (Thompson & Troeh, 1988).
Steel dust is a toxic residue that has a high content of heavy metals and a tendency to generate
highly alkaline leachates. Several authors define it as a hazardous waste and warn about the possibility of percolation of its elements from the upper layers of soil to groundwater. Therefore, improper steel dust disposal threatens to contaminate soil and water, due to its relative ease of leaching heavy metals into soils. (Seas, Maiza, Marfil, & Valea, 2003)
Leachate is the liquid that results from the percolation of fluids through any permeable material,
dragging with it numerous compounds present in the original waste solids both as suspended and
dissolved matter (Fernndez, 2006). In other words, when water comes into contact with hazardous
solid waste it frees significant components and the soil becomes contaminated (Giraldo, 2003).
Leachate treatment alternatives can be categorized according to the characteristics of the materials
used in the various methods. In recent years, many researchers have studied the use of phytoremediation to treat natural resources contaminated with heavy metals. This procedure relies on plants
to remove, transfer, stabilize, concentrate and degrade contaminants. (Navarro, Aguilar, & Lpez,
2007) Encapsulation of contaminants into a solid matrix using cement is also one of the most common techniques for the treatment of potentially hazardous solid wastes. (Domnech, 1995)
Leachate produced by steel dust recycling is also suitable for recovery of valuable metals such as
zinc. Such methods applied to metal recovery are classed as electrochemical. Electrochemical processes take advantage of the oxidation-reduction reactions so that the metal of interest will be deposited on cathodes and then subjected to pyro-metallurgical processes. (Espinoza, 2012)

METHODOLOGY
Soil and steel dust samples were collected and classified. With column tests were produced, collected and characterized leachates generated by steel dust and percolated through soil. The effluent
was then stabilized with Portland cement and lime in different proportions.

Collection and characterization of samples of soil and steel dust

Steel dust and previously uncontaminated cultivable soil samples were taken in Lasso (Ecuador)
using a simple random procedure. Their physical and chemical parameters were determined: real
and apparent density, porosity, humidity, pH, conductivity and particle size. Additionally, total element determination was carried out by atomic absorption (Perkin Elmer AAnalyst 300 Spectrophotometer). Finally, the Toxicity Characteristic Leaching Procedure (TCLP EPA Method 1311) was
used in order to assess the toxicity of the steel dust.

Column experiments for the production, collection and characterization of leachates

generated by steel dust and percolated through soil
Three columns of glass (6 cm diameter, 75 cm height) were installed. At the bottom of the columns 5
cm of granite (mesh 4 6) was placed in order to filter the effluent that will flow into the collection
chamber. The collection chamber, in the form of a funnel, connected the lower part of the column
with a 250 mL beaker. In the drainage chamber, 15 cm of steel dust was placed on 50 cm of soil. The
columns were irrigated with constant flows of potable water of 0.15 mL/h, equivalent to the mean
annual precipitation of Latacunga an amount equal to 455 mm.
The leachates produced by the percolation of water through the layers of dust and soil in the columns were characterized in terms of flow rate, pH, elemental composition (Ag, Al, As, Ba, Ca, Co,
Cr, Cu, Fe, K, Mg, Na, Hg, Ni, Pb, Si, V, Zn) and total organic carbon. This data was analyzed with
the Anova method using the Statgraphics software.

Leachates stabilization with cement and lime

Leachates generated by steel dust and percolated through the soil were mixed in different amounts
(80 mL, 100 mL and 120 mL of water) with 160 g of Portland cement and 40 g of lime. Stabilized solids were characterized in terms of apparent density, moisture, pH and conductivity. In addition,
these solids were subjected to the Toxicity Characteristic Leaching Procedure (TCLP, EPA Method
1311) and a compression strength test (INEN 1485) in order to verifier if they meet the requirements
to be disposed of in landfills.

Comparison of unconventional methods for leachate treatment

Leachate treatment diagram was based on the effluents average flow. The proposed process was
economically evaluated in order to define the operating cost of the system and compare it to other
alternatives such as the depuration of the leachate with electrocoagulation and phytoremediation,
as well as the recuperation of economically valuable metals with electrolysis prior to leachate
treatment with conventional methods (precipitation, filtration and adsorption).

RESULTS AND DISCUSSION

Soil and steel dust characterization
The most relevant physical and chemical properties of the soil and the steel dust tested in this research are tabulated in Table 1.
Table 1 Initial Physical and Chemical Properties of the Soil and Steel Dust
Property
Apparent density
Real density
Porosity
Moisture
pH
Conductivity
d50 particle size
d80 particle size

Unit
g/mL
g/mL
%
%

S/cm
m
m

Soil
0.8
1.8
52.9
8.1
6.9
235
112.3
524.4

Steel dust
0.9
2.5
62.8
3.6
12.7
344
94.0
145.5

The information presented in Table 1 demonstrates that steel dust is a dangerous waste, as the diameter of at least 50 % of the particles are less than 100 m, a fineness that can cause damage to the
human respiratory system. The soil pH classifies it as neutral and soil conductivity as normal. At
the same time, the steel dust is quite alkaline and slightly saline.
The soil composition results were determined by atomic absorption spectrophotometry at the beginning (Initial Sample) and end (Final Sample) of the column tests. The results are tabulated in Table 2.
Table 2 Initial and Final Elemental Composition of the Soil
Element

Units

Na
Ca
K
Al
Fe
Mg
Pb
Zn
Cr
Cu

%
%
%
%
%
%
mg/kg
mg/kg
mg/kg
mg/kg

Concentration
Initial Sample
43.7
14.2
11.0
6.2
2.7
2.6
315.5
313.9
89.6
78.0

Final Sample
40.2
13.5
10.9
6.1
2.7
2.5
235.1
384.9
106.9
41.6

Among the elements accumulated in the soil test, sodiums concentration is the highest, followed
by calcium, potassium and aluminum concentrations. All four were slightly reduced in the final
sample. Lead, zinc, chromium and copper, all trace elements, showed mixed results in the final
sample. Higher concentrations of chromium and zinc were found in soil affected by steel dust. The
dilution of the other elements is evident. Hence, steel dust contributes significant amounts of chromium and zinc to the soil.
Table 3 shows the steel dusts elemental composition determined in the sample before (Initial Sample) and after (Final Sample) the column tests by atomic absorption spectrophotometry.
Table 3 Initial and Final Elemental Composition of the Steel Dust
Element

Units

Zn
Fe
Cd
Pb

%
%
%
mg/kg

Concentration
Initial Sample
16.4
9.8
0.8
124.1

Final Sample
15.6
8.7
0.4
106.3

Data organized in Table 3 confirms that the steel dust tested in this research is a hazardous waste
due to the presence of heavy metals such as zinc, lead, and cadmium. Zinc is the most concentrated
metal dues to its low solubility in molten steel and slag. However, steel dust is rich in zinc even after the water percolation as shown in the column tests. This finding is important because the solid
waste could be recycled in order to recover the zinc even after its improper disposal.
Results obtained for the toxicological characterization of the initial steel dust sample are presented
in Table 4.

Table 4 Leachate Concentrations with Toxicological Characterization of Steel Dust Results

Element
Pb
Cd
As
Ni
Ba
Fe
Cr

TCLP Extract Concentration (mg/L)

195.0
11.0
0.7
0.4
0.2
0.1
0.1

Maximum Permitted Concentration (mg/L)

5.0
1.0
5.0
2.0
100.0
30.0
5.0

* (EPA, 2003)

TCLP Extract Concentrations of cadmium and lead exceed the EPA maximum permitted by 11 and
39 times respectively, so the presence of steel dust is of great concern. The tested steel dust is a dangerous solid due to its high cadmium and lead content.

Chemical characterization of leachate generated by steel dust and percolated through

soil
Figure 1 illustrates the concentration of heavy metals (Fe, Cr, Cu, Zn, As, Ni) in the leachate generat-

0,1

a)

0,08

Fe

0,06
0,04
0,02
0
0

50

100

Concentration (mg/L)

Concentration (mg/L)

ed by steel dust and percolated through soil as a function of time.

b)

1,5
1
0,5
0
0

50

c)

0,06

Cu

0,04
0,02
0
0

50
Time (days)

100

Time (days)

100

Concentration (mg/L)

Concentration (mg/L)

Time (days)
0,08

Cr

0,2

d)

0,15

Zn

0,1
0,05
0
0

50
Time (days)

Figure 1 Concentration of a) Fe, b) Cr, c) Cu, d) Zn, e) As and f) Ni in the Leachates Vs Time

100

e)

As

0,1
0,05
0
0

50

100

Concentration (mg/L)

Concentration (mg/L)

0,15

0,04

f)

0,03

Ni

0,02
0,01

Time (days)

0
0

50

100

Time (days)

Figure 1 Concentration of a) Fe, b) Cr, c) Cu, d) Zn, e) As and f) Ni in the Leachates Vs Time (Continuation)

Compositions of the leachates obtained in this research needs to be analyzed under the Ecuadorian
legislation and regulations TULAS (Unified Text for Secondary Environmental Legislation), Book
VI, Annex 1, entitled Norma de calidad ambiental y de descarga de efluentes: Recurso Agua
(Quality Standard on Environment and Effluents Discharge: Water Resource). For this reason, in
Table 5 lists the maximum permitted concentrations of heavy metals in the effluents for its discharge to freshwater bodies.
Table 5 TULAS Maximum Permitted Concentrations for Effluents Discharge to Freshwater Bodies
Element
As
Cu
Cr
Fe
Ni
Zn
pH

Maximum Permitted Concentration (mg/L) *

0.1
1.0
0.5
10.0
2.0
5.0
9.0

* (Ministerio del Ambiente de Ecuador, 2003)

As can be seen in Figure 1 (b), Chromium exceeded permitted levels during the first 19 days of percolation, so the leachate generated by steel dust and percolated through soil could not at that point
be discharged to bodies of freshwater. Arsenic also exceeds the maximum limit established in TULAS, but it exceeds the maximum value of 0.1 mg/L only at one single point at the start of the column tests. As regards of the rest of heavy metals, their concentrations are less than the limits set in
the Ecuadorian regulations during the period of four months of the column tests.
Figure 2 illustrates the leachates content of organic carbon as a function of time. As can be seen, the
content of organic carbon of the leachate remains practically constant. Probably, soil poor aeration
impeded the solubilization of the organic fraction in the percolation water.
Figure 3 presents the leachates pH in the time domain and the maximum limit established by Ecuadorian regulations for the pH of effluents discharged to freshwater bodies. The leachate exceeded
the pH limit set by TULAS during the first thirty days, after which the pH did not exceed the limit.
Additionally, as the leachate pH does not exceed the value of 11, this explains the low dissolution

Total Organic Carbon (%)

rate of elements such as lead that is very soluble in pH ranges of 12 to 13. However, this property
shows a downward trend, which implies an acidification of the liquids and therefore a greater dissolution rate of elements whose solubility increases when pH decreases.

10

10

pH 8

2
0

Leachate pH
pH Limit (TULAS)

6
0

20

40

60

80

Time (days)

50

100

Time (days)

Figure 2 Leachate Total Organic Carbon Vs Time

Figure 3 Leachate pH Vs Time

Stabilization by encapsulation of leachates

The results of leachate stabilization with cement and lime are presented in TABLE 6. As can be
seen, solids stabilized with an L/S relationship equal to 0.6 do not comply with the limit set for
moisture in order to be disposed of in landfills, but there is not any problem with the other parameters. On the contrary, solids obtained with an L/S relationship equal to 0.4 and 0.5 satisfy the limits
for all properties so that they would meet the requirements to be confined in landfills containment
cells.
Table 6 Physical and Chemical Properties of Stabilized Solids
Property
Density
Moisture
pH
Conductivity
Compression Strength

Units
g/mL
%

S/cm
kPa

L/S = 0.4*
0.57
5.44
7.65
307
964.91

L/S = 0.5*
0.57
8.15
7.72
294
771.60

L/S = 0.6*
0.62
11.33
8.47
288
584.42

Limit**

<10
4 - 13
<100 000
>50

*L/S = Liquid mass/Solid mass; ** (Ruiz, 2003)

In Table 7 reports the toxicological characterization of the stabilized solids. The TCLP extract toxicity limits are included in order to determine if the samples tested are toxic. Based on the data, it can
be inferred that the solids obtained with the stabilization of leachates generated by steel dust and
percolated through soil cannot be considered toxic, since their concentration in all cases is lower
than the maximum permissible.

In conclusion, the best formulation for stabilization of leachates generated by steel dust and percolated through soil corresponds to the L/S ratio equal to 0.4.
Table 7 Leachate Concentrations from Toxicological Characterization of Solids Stabilized Results
Element
As
Cd
Cr
Fe
Ni
Ag
Pb
Se
As

TCLP Extract Concentration (mg/L)

L/S = 0.4*
L/S = 0.5*
L/S = 0.6*
0.0495
0.0986
0.0062
<0.01
<0.01
<0.01
0.03
0.38
0.48
8.27
1.48
7.77
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.10
<0.10
<0.10
0.04
0.05
0.02
0,0
0,1
0,0

Maximum Permitted Concentration (mg/L) **

5.0
1.0
5.0
30.0
2.0
5.0
5.0
1.0
5.0

*L/S = Liquid mass/Solid mass; ** (EPA, 2003)

System proposed for the leachates treatment

The leachate treatment system capacity is 2 m/month, based on the average discharge flow of the
leachates obtained in the three experimental units (0.09 mL/h) and on the steel dust production rate
in an Ecuadorian scrap metal recycling company (9 t/day). Consequently, the necessary raw materials for the leachates treatment are 4 t of Portland cement and 1 t of lime.
The process proposed starts with the collection of leachates generated by steel dust and percolated
through soil, which will be accumulated in a 2.1 m tank. Subsequently, the first day of each month
the collected liquid will be mixed for 4 min with cement and lime in a concrete mixer. Afterward
the pulp will be led to a vibro-compaction machine where the molds will be filled with the mixture
for 30 s, the auto sweeper will remove excess material on the molds and the press will compact the
mixture. Then stabilized solid blocks fall onto a board located under the vibro-compaction machine
and will dry at room temperature. Finally, stabilized blocks will be brought to the landfill.

Economic evaluation of the proposed system and comparison with alternative methods
Table 8 lists the operating costs required for the stabilization of leachates generated by steel dust
and percolated through soil assuming that the industrial plant responsible for the effluents production has all the assets and personnel necessary for the development of the process.
Table 8 Estimation of the Investment per Month in Direct Materials
Material
Cement
Lime
Water
Diesel
Electricity

Quantity
4
1
11.8
31.4
0.7091

Unit
t/month
t/month
m/month
gal/month
kWh/month

Unit value (US$)

152.80
160.00
0.31
0.91
0.07
Total

Monthly investment (US$)

611.20
160.00
3.66
28.57
0.05
803.48

The proposed process for treating the heavy metal content of leachates generated by steel dust and
percolated through soil contrasts with the following alternatives: depuration of the contaminated

liquid with a combination of the electrocoagulation and phytoremediation processes, proposed by

Guanoluisa (2012), and zinc recuperation by electrolysis prior to the effluent treatment with conventional methods, developed by Espinoza (2012).
Guanoluisa (2012) proposes a leachate treatment system consisting of electrocoagulation as the
primary treatment, phytoremediation as the secondary treatment, and activated carbon adsorption
as the tertiary treatment. All three steps are necessary in order to reduce the heavy metal and organic matter content to values accepted by the Ordinance 213 of Quito. The system consists of an
electrocoagulation reactor of 2 m which operates with aluminum electrodes (53 A/m, 20 min), a
Napier grass surface wetlands, and a column of activated carbon. This system involves operating
costs of US$ 1.57 per m of leachate.
Espinoza (2012) proposed a method for zinc recovery using electrolysis (lead anodes, aluminum
cathodes, pH between 4 and 5, 400 A/m, 8 h) and a smelting process of the cathode with ammonium chloride (450 C, 2 h) in order to recover from the leachate high purity 99% metallic zinc. Under
these conditions, the system recovers 71% of the metal with an operating cost of US$ 11.84 per m of
liquid treated.
Table 9 are summarizes the operating costs of the methods considered in this paper for the treatment of leachates generated by steel dust and percolated through soil, and the wastes that are generated with each one of the proposed systems.
Table 9 Operating Cost of the Leachate Treatment Methods
Method
Leachate stabilization by encapsulation
Leachate depuration by electrocoagulation and phytoremediation *
Recovery of zinc by electrocoagulation and fusion **

Cost (US$/m)
401.74
1.57
11.84

Treatment Residue
None
Electrolysis sludge
Electrolysis sludge

* (Guanoluisa, 2012), ** (Espinoza, 2012)

Costs presented in Table 9 suggests that the leachate stabilization with cement and lime is 250 times
more expensive than the effluent treatment with electrocoagulation, either with the aim that the
liquid comply with the Ecuadorian regulations or in order to recover a valuable material (Zn).
However, despite its cost, the stabilization results in a final disposition of the leachate without the
emergence of a residual parallel wastes that must themselves be treated at additional cost. Additionally, in the case of leachate generated by steel dust deposited on soil, the contents of zinc (0.03
mg/L on average) does not justify the application of the alternative proposed by Espinoza (2012)
and the content of total organic carbon (6.6% on average) does not support the use of the system of
Guanoluisa (2012). Finally, although the heavy metal content exceeds the limits established by the
Ecuadorian regulations, the difference is relatively low as to apply simpler conventional methods.
The technology needed for the stabilization of an effluent is less expensive than the required for the
electrocoagulation, phytoremediation and fusion, fits in the smallest space and its technical requirements are the simplest. In addition, the time needed to stabilize a leachate is less than the corresponding to the other alternatives.

CONCLUSION
Leachates generated by steel dust and percolated through soil do not comply with the limits established by the Ecuadorian regulations for discharge into freshwater bodies. Chromium, with a concentration greater than 0.5 mg/L, and zinc, with an average concentration equal to 0.03 mg/L (for
6.6% of total organic carbon) are not reduced by percolation. Stabilization with cement and lime in
a liquid mass/solid mass equal to 0.4 is an effective method to obtain blocks that can be disposed in
landfills containment cells. Zinc recovery from the waste does not justify the valorization of the effluent in order to recover this metal and the heavy metals and organic matter content does not justify the use of non-conventional techniques such as the electrocoagulation and phytoremediation.
Therefore, cement lime encapsulation is the best option for treatment of leachates generated by steel
dust deposited on soil. For Ecuador, it is among the simplest conventional techniques.

REFERENCES
Bautista, F. (1999). Introduccin al estudio de la contaminacin del suelo por metales pesados. Mrida, Yucatn,
Mxico: Editorial de la Universidad Autnoma de Yucatn.
Domnech, X. (1995). Qumica del suelo El impacto de los contaminantes (3rd ed.). Madrid, Espaa: Miraguano
Editions.
EPA, E. P. (2003). Method 1311: Toxicity Characteristic Leaching Procedure. Retrieved February 13, 2013, from
http://www.epa.gov/wastes/hazard/testmethods/sw846/pdfs/1311.pdf
Espinoza, C. (2012). Recuperacin de zinc a partir de polvos de acera. (Proyecto de Titulacin previo a la obtencin
del ttulo de Ingeniero Qumico), Escuela Politcnica Nacional, Quito, Ecuador.
Fernndez,
A.
(2006).
Contaminacin
por
lixiviados.
Retrieved
August
8,
http://www.consumer.es/web/es/medio_ambiente/urbano/2006/10/13/156373.php

2012,

from

Giraldo, E. (2003). Tratamiento de lixiviados de rellenos sanitarios. Avances recientes. Revista de ingeniera
Uniandes, 14(8), 44-55.
Guanoluisa, L. (2012). Diseo de un sistema de tratamiento de lixiviados del relleno sanitario El Inga mediante
electrocoagulacin y fitorremediacin. Escuela Politcnica Nacional, Quito, Ecuador.
Ministerio del Ambiente de Ecuador. (2003). Norma de calidad ambiental y de descarga de efluentes Recurso agua.
Retrieved February 19, 2013, from http://www.efficacitas.com/efficacitas_es/assets/Anexo%201.pdf
Navarro, J., Aguilar, I., & Lpez, J. (2007). Aspectos bioqumicos y genticos de la tolerancia y acumulacin de
metales pesados en plantas. Revista cientfica de ecologa y medio ambiente Ecosistemas, 16(2), 10-25.
Ruiz, U. (2003). Recomendaciones tcnicas generales para el diseo y construccin de celdas de confinamiento y rellenos
sanitarios.
Retrieved
November
20,
2012,
from
http://www.semarnat.gob.mx/temas/gestionambiental/Materiales%20y%20Actividades%20Riesgosas/
sitioscontaminados
Seas, L., Maiza, P., Marfil, S., & Valea, J. (2003). Impacto ambiental producido por morteros cementicios con
polvos de acera. Revista Ciencia e Ingeniera, 24(3), 47-53.
Thompson, L., & Troeh, F. (1988). Los suelos y su fertilidad (4th ed.). Madrid, Espaa: Revert Editorial.

Industrial results of online monitoring

system for leaching pads

Juan Larenas, Paula Lpez

and Toms Fuenzalida
bhp Billiton, Chile
Andrs Alvarado
Aplik Mining, Chile

At Spence, 28 operators must go over the surface of the leaching pads

(553.000 [m2]) in order to perform maintenance and inspection tasks.
These operators control the permeability of the stacked ore under
irrigation conditions and fix failures of the irrigation system.
The people in our team are exposed to: (i) Material risks (fatality)
associated to ground failure and contact with dangerous substances, (ii)
extreme levels of UV radiation.
Spence changed this approach to leaching and developed a novel
technology with APLIK, which lead to the implementation of the
Online monitoring system for leaching pads. This system eliminates
the risks associated with the inspection and maintenance tasks, since
the operator can check the ground conditions and identify
malfunctions remotely and in real time. Additionally, it allows for a
total control during night-time, something with no precedents in the
mining industry.
The system is based in advanced artificial-vision technology and can
perform a total scan of the oxide leaching pads in less than 15 minutes,
which allows for the detection and timely correction of operational
malfunctions. Without this system, it wouldnt be possible to have total
control of the process without exposing the operators at the same time, nor

observe the totality of the area under irrigation in a constant and efficient
way. The possibility of night-time control is another benefit, since in the past
only a peripheral control was possible at night due to safety concerns.
In terms of production, this project impacts recovery positively, since it
makes it possible to keep the irrigation system according to design
parameters, while it also allows for the correction of zones without
irrigation, flooded zones as a consequence of excessive irrigation, channelformation, piping disconnections, and possible ground failures. All these
problems are attacked in their origin, no matter how small they are.

There is no full article associated with this abstract.

chapter 3
Copper hydrometallurgical processes

Copper sulfide ore leaching in seawater:

Oxidants influence

Pa Hernndez, Mara
Taboada, Tefilo Graber
and Hctor Galleguillos
csiro Chile, Centre of
Excellence in Mining and
Mineral Processing;
and Universidad de
Antofagasta, Chile

The aim of this study is to determine the effect of different oxidants on

copper sulfide ore leaching and its influence on copper recovery percentage
in solution, with a vision to replace fresh water with seawater, as an
alternative to scarce of hydric resource looking for a sustainability and to
test new methods of recovering copper via hydrometallurgy process from
copper sulfide ores.
Mining activity in Chile is concentrated in the north of the country
where the weather is desert and it scarcely rains. Currently, some mines are
using seawater o desalinated water directly in their processes. The study of
the influence of quality of water (saline waters, seawater) and different
oxidants in this medium is not widely documented, nor is the information
disseminated by the scientific community. In a preliminary work, agitation
leaching tests were performed in a laboratory using ore from the
Antofagasta Region with 0.77% total Cu and chalcopyrite and bornite as
principal species of copper. This ore does not have copper oxide. Different
leaching solutions using H2SO4 with seawater and saline process water as
solvents at 45C, fine particle size and leaching time of 7 days were used.
The effect of each oxidant added is determined (nitrate, ferric/ferrous and
cupric). Samples were drawn at to different times and copper in the
solution was measured to obtain the kinetic curve. Preliminary results were
obtained for each system, with a maximum copper recovery of 80% for
H2SO4 + NaNO3 system in seven days of leaching at 45C. Ferric, ferrous and
cupric systems showed similar results on copper extraction ( 60%). Saline
medium can present good results in copper extraction using oxidants,
under certain leaching conditions.

INTRODUCTION
Currently freshwater is being replaced by seawater in the mining industries (Moreno et al. 2011;
Gonzlez et al. 2012) due to the growing scarcity of this resource (Cochilco 2012). Therefore it is
important determine the effect of seawater in the leaching system when different oxidants are
present. The seawater has a high salts concentration (Gianguzza et al. 2002) with an approximately
3.5 % of salinity and its use in mineral processes might produce changes in the behavior of the
solutions and properties. On the other hand, copper sulfide, especially chalcopyrite is the major
source of copper that exists worldwide and its treatment by hydrometallurgical route is still under
discussion. Large-scale industrial process is produced by the pyrometallurgical route producing
high contamination by liberation of gaseous SO2 to environmental (Prasad & Pandey 1998). The
alternative is hydrometallurgical route. For the chalcopyrite leaching is necessary conditions very
special because this mineral tends to passivation forming layer of material around of its surface.
This produce difficulty in mass transfer therefore the leaching kinetic is very slow. The presence of
oxidant agent like oxygen, ferric, cupric, chlorine, etc., is necessary. Many researchers have worked
in chloride medium due to formation of complex that better the solubility of the mineral and Cu(I)
is stable in this conditions. Chalcopyrite leaching is better in chloride medium than sulfate medium.
In general, an increase in the chloride concentration of leaching solution until 0.5 M shows a
positive effect in the copper extraction. It has been shown that the presence of chlorides produces a
porous layer around the particle leached allowing better diffusion between the liquid and the solid.
Several authors have studied the leaching of chalcopyrite using chloride medium (Skrobian et al.
2005; Herreros & Vials 2007; Liddicoat & Dreisinger 2007; Senanayake 2009; Velsquez-Yvenes et
al. 2010; Yvenes et al. 2010; Yoo et al. 2010; Lundstrm et al. 2012; Ibez & Velsquez 2013), nitrate,
cupric and ferric ions like oxidants (Carneiro & Leo 2007; Crdoba et al. 2008; Soki et al. 2009; Gok
& Anderson 2013; Lu & Dreisinger 2013; Watling 2013) and redox potential control (Hiroyoshi et al.
2007; Hiroyoshi et al. 2008). Torres et al. (2013) studied copper leaching from oxide ore using
seawater and process water. With the aim of continue this study, the effect of different oxidants on
acid leaching of copper sulfide ore and its influence on copper extraction percentage in solution
using raw seawater and synthetic process water (seawater concentrate in different ions) as solvents
were determined in a preliminary study where several tests were performed.

METHODOLOGY
Ore
Ore of Antofagasta Region was used. Mineralogical species were determined by optical microscopy
and X ray diffraction (XRD). Chalcopyrite was the main source of copper (1.21%) present in the ore
(see TABLE 1). Bornite, chalcocite and enargite were present too but at low percentage ( 0.5%).
Copper total was about 0.77% determined by inductively coupled plasma atomic emission
spectrometry (ICP-AES) and corroborated with atomic absorption spectrometry (AA). Other species
present were quartz, gypsum, calcite, magnetite, pyrite and silicates (gangue). The size particle
used was below 150 m obtained by crushed and milled of the ore.

Table 1 Mineralogical species of ore used obtained by XRD

Compound

Formula

wt%

Compound Formula

wt%

Cuarzo

SiO2

32.29

Clorita

(Mg5Al)(Si,Al)4O10(OH)8

0.87

Albita

NaAlSi3O8

29.36

Fe3O4

0.73

Ortoclasa

KAlSi3O8

13.84

Magnetita
Copiapita
magnsica

MgFe4(SO4)6(OH)220H2O

0.70

Yeso
Clorita
ferrfera

CaSO42H2O

7.54

Bornita

Cu5FeS4

0.54

(Mg,Fe)6(Si,Al)4O10(OH)8

5.54

Pirita

FeS2

0.50

Moscovita

(K,Na)(Al,Mg,Fe)2(Si3.1Al0.9)O10(OH)2 5.31

Saponita

Na0.3Mg3(Si,Al)4O10(OH)26H2O 0.30

Calcita

CaCO3

1.24

Covelita

CuS

0.01

Calcopirita

CuFeS2

1.21

Enargita

Cu3AsS4

0.005

Reagents
H2SO4 was used for acidification of solution. NaCl, NaNO3, CuSO45H2O, FeSO47H2O and
Fe2(SO4)3xH2O were the analytical grade reagents (Merck) used for preparation of leaching
solutions. Seawater was obtained of San Jorge Bay (Antofagasta). It was filtered until 1 m and
used as solvent without other treatment. Synthetic process water was prepared by addition to
seawater of NaCl and CuSO45H2O. In a real leaching process using seawater as solvent, different
ions present in the system will begin to concentrate and they will affect in the process. The synthetic
process water simulates this effect over time. The ions compositions of seawater and synthetic
process water are shown in TABLE 2.
Table 2 Solvents composition (mgL-1)
ID

Na+

K+

Mg+2

Ca+2

Cu2+

B+3

Cl-

HCO3-

NO3-

SO4-2

Seawater

9,480

374

1,190

386

0.072

4.6

18,765

142

2.05

2,771

Synthetic process water

17,255

374

1,190

386

2,072

4,6

30,765

142

2,05

5,794

The agitation leaching tests were performed in jacketed glass reactors equipped with caps to
prevent evaporation. Agitation was mechanical using a Teflon stir bar. The temperature control is
accomplished using a thermostatic bath which recirculated water at the desired temperature in the
reactor jacket. Temperature of work was 45C, agitation velocity was 400 rpm, solid liquid ratio was
100 g ore in 1 L solution and the leaching time was 7 days. Solution samples (10 mL) were obtained
periodically and analysed for copper by atomic absorption spectrometry. This was corroborated by
residue analysis for copper using AA.

RESULTS AND DISCUSSION

Different types of oxidants were studied in different solvents. Ferric, cupric and nitrate ions were
added to systems for determine their effects. Ferrous effect was studied too. The ratio between
copper in solution at time t and the copper initial present in the ore is the copper extraction.

Nitrate systems
When nitrate is used as oxidant, Soki et al. (2009) proposed the follow equation:
+5

+5

) +

+2 +5

+5

(1)

H2SO4 + NaNO3 in seawater and synthetic process water at different concentrations of reagents,
were studied. H2SO4 concentration was studied to 0.2, 0.3, 0.5 and 0.7 M, and NaNO 3 concentration
was studied to 0.2, 0.7 and 1 M. In this work the reagents are in excess according to the reaction (1).
It is due to presence of gangues consumers acid. Moreover, the effect of excces of nitrate will be
studied. Figure 1 shown the results obtained for those systems
100

Cu Ext/%

80
60
40
20
0
0

20

40

60

80 100
t/h

120

140

160

180

H2SO4 (0.7 M)+ NaNO3 (0.7 M) + SW

H2SO4 (0.7 M)+ NaNO3 (0.7 M) + PWS
H2SO4 (0.7 M) + NaNO3 (1 M) + SW
H2SO4 (0.7 M) + NaNO3 (1 M) + PWS
H2SO4 (0.5 M) + NaNO3 (1 M) + SW
H2SO4 (0.5 M) + NaNO3 (1 M) + PWS
H2SO4 (0.3 M) + NaNO3 (1M) + SW
H2SO4 (0.3 M) + NaNO3 (1 M) + PWS
H2SO4 (0.2 M) + NaNO3 (0.2 M) + SW
H2SO4 (0.2 M) + NaNO3 (0.2 M) + PWS

Figure 1 Nitrate and acid effect on copper extraction

Maximum copper extraction to leaching time for seawater and synthetic process water, respectively
were of 79.2% and 87.0% for H2SO4 (0.7 M) + NaNO3 (1 M) systems. A better copper extraction from
ore is obtaining when synthetic process water is used as solvent, for all systems. Cupric and
chloride ions presents in the solvent, increased the copper extraction. Decrease acid concentration
from 0.7 to 0.2 M produce a decrease on copper extraction. The same is observed when decrease
nitrate concentration but is less notorious. The oxidant power of nitrate is better when increases the
acid concentration, chloride and cupric.

Ferric and ferrous systems

Dutrizac and MacDonald (1974) proposed the next reaction when ferric is used as oxidant:
CuFeS + 4Fe

Cu

+ 5Fe

+ 2S

(2)

Moreover, Hiroyoshi et al. (2001) proposed two sequential reactions when ferrous and cupric are
present in solution for chalcopyrite leaching:
CuFeS + 3Cu + 3Fe 2Cu S + 4Fe
2Cu S + 8Fe 4Cu + 8Fe + S

(3)
(4)

Cu Ext/ %

H2SO4 + Fe2(SO4)3 and H2SO4 + FeSO4 in seawater and synthetic process water at different
concentrations of reagents, were studied. H2SO4 concentration was studied to 0.2 M, Fe2(SO4)3 and
FeSO4 concentration were studied to 0.1 and 0.2 M. According to preliminaries reactions, the
reagents are in excess. Figure 2 and 3 shown the results obtained for these systems.
70
60
50
40
30
20
10
0

H2SO4 (0.2 M) + Fe2(SO4)3 (0.1 M) + SW

H2SO4 (0.2 M) + Fe2(SO4)3 (0.1 M) + PWS
H2SO4 (0.2 M) + Fe2(SO4)3 (0.2 M) + SW
H2SO4 (0.2 M) + Fe2(SO4)3 (0.2 M) + PWS

20

40

60

80

100 120 140 160 180

t/h
Figure 2 Ferric and acid effect on copper extraction
70

H2SO4 (0.2 M) + FeSO4 (0.1 M) + SW

Cu Ext/ %

60
50

H2SO4 (0.2 M) + FeSO4 (0.1 M) + PWS

40

H2SO4 (0.2 M) + FeSO4 (0.2 M) + SW

30

H2SO4 (0.2 M) + FeSO4 (0.2 M) + PWS

20
10
0
0

20

40

60

80 100 120 140 160 180

t/h

Figure 3 Ferrous and acid effect on copper extraction

In both systems (ferric and ferrous) the best copper extraction (58.4%) to leaching time was
obtained using synthetic process and concentrations of 0.2 M Fe3+ and 0.1 M Fe2+. Note that ferric is
an oxidant and ferrous acts as an oxidant source (ferric ions) when it itself is oxidized.
For ferric systems: slightly better copper extraction was obtained when synthetic process water was
used as solvent for 0.1 M Fe3+ and for 0.2 M Fe3+, comparing the same reagent concentration but
different solvents. Comparing the same solvent, similar kinetics curves were observed using 0.1 and
0.2 M Fe3+ for both solvents.
For ferrous systems, better copper extraction was obtained using synthetic process water for H 2SO4
(0.2 M) + FeSO4 (0.1 M). To 0.2 M of ferrous, both kinetics curves (with seawater and synthetic
process water) are very similar. The same behavior was observed on kinetic curves of both systems
(0.1 and 0.2 M Fe2+) with seawater as solvent, but slightly better copper extraction was obtained
when process water was used to 0.1 M Fe2+. A high chloride concentration helps to increase the
copper extraction.

Comparing the systems with 0.1 M of Fe2+ and Fe3+ in the same solvent, the kinetic curves were very
similar. This was also observed for 0.2 M of Fe 2+ and Fe3+ in the same solvent. These results confirm
that ferrous ion is only a source of ferric and its oxidation is fast in the systems to conditions
studied in this work.

Cupric systems
When cupric is used as oxidant, Habashi (1978) proposed the follow equation:
CuFeS + 3Cu

4Cu + Fe

+ 2S

(5)

H2SO4 + CuSO4 in seawater and synthetic process water at different concentrations of reagents,
were studied. H2SO4 concentration was studied to 0.2 M, and CuSO4 concentration was studied to
0.05, 0.1, 0.2 M. According to (5) equation, the reagents are in excess. Figure 4 shown the results
obtained for these systems.
70
H2SO4 (0.2 M)+ CuSO4 (0.05 M) + SW
H2SO4 (0.2 M) + CuSO4 (0.05 M) + PWS
H2SO4 (0.2 M) + CuSO4 (0.1 M) + SW
H2SO4 (0.2 M) + CuSO4 (0.1 M) + PWS
H2SO4 (0.2 M) + CuSO4 (0.2 M) + SW
H2SO4 (0.2 M) + CuSO4 (0.2 M) + PWS

60

Cu Ext/ %

50
40
30
20
10
0
0

20

40

60

80 100 120 140 160 180

t/h

Figure 4 Copper and acid effect on copper extraction

Maximum copper extraction to leaching time for seawater and synthetic process water, respectively
were of 58.5% and 61.1% for H2SO4 (0.2 M) + CuSO4 (0.1 M) and H2SO4 (0.2 M) + CuSO4 (0.2 M)
systems. From the figure it can be observed that the kinetics of the most curves have the similar
trend in the end of the tests with minimal differences on maximum copper extraction. A slightly
better copper extraction from ore is obtaining when synthetic process water is used as solvent, for
all systems but the trends at the end, are similars.

CONCLUSION
Oxidants effect using two types of solvents were studied on leaching of copper sulfide ore with low
particle size. Maximum copper extraction was obtained using nitrate as oxidant to reach an
approximately 87% in H2SO4 (0.7 M) + NaNO3 (1 M) + saline solvent in 7 days. Copper extraction of
about 57% was achieved using ferric and ferrous, and 61% when cupric ions are added on the acid
systems.

Nitrate systems should be more studied including temperature and concentrations effects.
Impurities analysis and gases production due to reactions should be studied with more detail.
To continue the study, it is recommendable to work in column leaching and determine the effect of
the different variables involved in order to scale to industrial level.
Better copper extraction was obtained when synthetic process water (seawater with addition of
chloride and cupric ions) was used as solvents in almost all the systems studied. This confirms that
the chloride and cupric ions help to the leaching process. Use of seawater is not detrimental for the
leaching process, in accordance to studied in this work. The difficulty is pumping seawater toward
the industry and the type of material used in the facility. New clean energy technologies can be
developed over time and the use of new material for industrial plants can help solve these
problems.

ACKNOWLEDGEMENTS
The authors are grateful for the financial support provided by CSIRO-Chile, CONICYT through
Fondecyt Project N 1140169 and CICITEM R04I1001.

NOMENCLATURE
SW

seawater as solvent

PWS

synthetic process water as solvent

REFERENCES
Carneiro, M. F. C. and V. A. Leo (2007) 'The role of sodium chloride on surface properties of chalcopyrite
leached with ferric sulphate.' Hydrometallurgy, 87, (34), pp. 73-82.
Cochilco (2012). Consumo de agua en la minera del cobre 2011 de/09/2012. G. d. C. Ministerio de Minera.
Crdoba, E. M., J. A. Muoz, et al. (2008) 'Leaching of chalcopyrite with ferric ion. Part I: General aspects.'
Hydrometallurgy, 93, (34), pp. 81-87.
Dutrizac, J. E. and R. J. C. MacDonald (1974) 'Ferric ion as a leaching medium.' Miner. Sci. Eng., v. 6, no. 2, pp.
59-100, pp.
Gianguzza, A., E. Pelizzetti, et al. (2002). Chemistry of marine water and sediments, Springer.
Gok, O. and C. G. Anderson (2013) 'Dissolution of low-grade chalcopyrite concentrate in acidified nitrite
electrolyte.' Hydrometallurgy, pp.
Gonzlez, S., M. Ramrez, et al. (2012) 'The impact of corrosion on mine water supply systems.', WIM2012
Proceeding of 3rd International Congress on Water Management in the Mining Industry, Santiago, Chile, pp.
74-84
Habashi, F. (1978). Chalcopyrite. Its Chemistry and Metallurgy.
Herreros, O. and J. Vials (2007) 'Leaching of sulfide copper ore in a NaClH2SO4O2 media with acid pretreatment.' Hydrometallurgy, 89, (34), pp. 260-268.
Hiroyoshi, N., H. Kitagawa, et al. (2008) 'Effect of solution composition on the optimum redox potential for
chalcopyrite leaching in sulfuric acid solutions.' Hydrometallurgy, 91, (1), pp. 144-149.

Hiroyoshi, N., S. Kuroiwa, et al. (2007) 'Effects of coexisting metal ions on the redox potential dependence of
chalcopyrite leaching in sulfuric acid solutions.' Hydrometallurgy, 87, (12), pp. 1-10.
Hiroyoshi, N., H. Miki, et al. (2001) 'Enhancement of chalcopyrite leaching by ferrous ions in acidic ferric
sulfate solutions.' Hydrometallurgy, 60, (3), pp. 185-197.
Ibez, T. and L. Velsquez (2013) 'Lixiviacin de la calcopirita en medios clorurados.' Revista de Metalurgia, 49,
(2), pp. 131-144.
Liddicoat, J. and D. Dreisinger (2007) 'Chloride leaching of chalcopyrite.' Hydrometallurgy, 89, (34), pp. 323331.
Lu, J. and D. Dreisinger (2013) 'Copper leaching from chalcopyrite concentrate in Cu (II)/Fe (III) chloride
system.' Minerals Engineering, 45, pp. 185-190.
Lundstrm, M., J. Liipo, et al. (2012) 'Dissolution of copper and iron from sulfide concentrates in cupric
chloride solution.' International Journal of Mineral Processing, 102103, (0), pp. 13-18.
Moreno, P. A., H. Aral, et al. (2011) 'The use of seawater as process water at Las Luces coppermolybdenum
beneficiation plant in Taltal (Chile).' Minerals Engineering, 24, (8), pp. 852-858.
Prasad, S. and B. Pandey (1998) 'Alternative processes for treatment of chalcopyritea review.' Minerals
Engineering, 11, (8), pp. 763-781.
Senanayake, G. (2009) 'A review of chloride assisted copper sulfide leaching by oxygenated sulfuric acid and
mechanistic considerations.' Hydrometallurgy, 98, (1), pp. 21-32.
Skrobian, M., T. Havlik, et al. (2005) 'Effect of NaCl concentration and particle size on chalcopyrite leaching in
cupric chloride solution.' Hydrometallurgy, 77, (1), pp. 109-114.
Soki, M. D., B. Markovi, et al. (2009) 'Kinetics of chalcopyrite leaching by sodium nitrate in sulphuric acid.'
Hydrometallurgy, 95, (34), pp. 273-279.
Torres, C. M., M. E. Taboada, et al. (2013) 'Copper extraction from oxide ore using seawater', Hydroprocess2013
5th International Seminar on Process Hydrometallurgy, Santiago, Chile, pp. 69-77
Velsquez-Yvenes, L., M. Nicol, et al. (2010) 'The dissolution of chalcopyrite in chloride solutions: Part 1. The
effect of solution potential.' Hydrometallurgy, 103, (14), pp. 108-113.
Watling, H. (2013) 'Chalcopyrite hydrometallurgy at atmospheric pressure: 1. Review of acidic sulfate, sulfate
chloride and sulfatenitrate process options.' Hydrometallurgy, 140, pp. 163-180.
Yvenes, L. V., H. Miki, et al. (2010) 'The dissolution of chalcopyrite in chloride solutions: Part 2: Effect of
various parameters on the rate.' Hydrometallurgy, 103, (14), pp. 80-85.
Yoo, K., S.-k. Kim, et al. (2010) 'Effect of chloride ions on leaching rate of chalcopyrite.' Minerals Engineering, 23,
(6), pp. 471-477.

Immobilization of arsenic by
schwertmannite synthesized from
copper heap leach solutions

Sengpasith HoungAloune,
Naoki Hiroyoshi, Mayumi
Ito and Masato Fujimura
Hokkaido University, Japan

Arsenic is a very poisonous inorganic pollutant and its contamination of

effluents resulting from mining activity is a worldwide problem and
regulations for arsenic control have become stricter. Therefore, better
technologies to remove and stabilize arsenic from effluents are necessary.
Arsenic retention by schwertmannite (Fe8O8(OH)6SO4) has received great
attention in recent years due to the high sorption capacity and strong
binding affinity of schwertmannite to arsenic species. In porphyry copper
mines, heap leaching solutions generally contain ferric and sulfate ions,
components of schwertmannite. Synthesis of schwertmannite from these
solutions is of economic and environmental interest for arsenic removal
from effluents.
This study investigates the synthesis of schwertmannite from copper
heap leaching solutions, focusing on the effect of co-existing metal ions on
the synthesis pathway; and the efficiency of schwertmannite in removal
and immobilization of arsenic from aqueous solutions. Results show that
schwertmannite can be synthesized from copper heap leach solutions. The
co-existing metal ions had a significant influence on the synthesis.
Schwertmannite was generated in the presence of Cu2+ and Fe2+, while
goethite was mainly formed in the presence of other metal ions (Mg2+, Zn2+,
Co2+, Ni2+) and Fe2+; suggesting the critical role that Cu2+ and Fe2+
availability plays in controlling the synthesis of schwertmannite from
copper heap leaching solutions.
Batch sorption studies indicate that As(V) in acidic solutions (pH 34)
can be removed effectively by synthesized schwertmannite with As(V)
sorption capacity of ~1.5 mmol g-1 ; and a lower As(V) sorption capacity (~0.7
mmol g-1) was observed on products containing goethite. In addition,
arsenic(V)-sorbed schwertmannite aged in deionized water at 25C exhibits
no mineralogical phase changes even after ageing at pH 3-4 for 120 days;
and the concentration of arsenic released from the solid phase is very low,
suggesting long-term stability.

INTRODUCTION
Arsenic (As), a toxic/carcinogenic element, is one of the most dangerous inorganic pollutant that
causes environmental and health problems in several areas of the world (Smedley & Kinniburgh,
2002). Arsenic is released in large quantities as a result of mining activities, including mineral
excavations, smelting, refining, and waste/wastewater disposal around the mines, both during the
active operation and also long after mine closure (Langsch et al., 2012). Arsenic pollutions of
groundwater and soils related to mining activities are serious problem in many countries
worldwide (Choi et al., 2009).
Recently, regulations to limit the release of arsenic to the environment are becoming more stringent
with increasing public awareness of the toxicity of arsenic and better understanding of its impact on
the environment (Long et al., 2012). Therefore, arsenic control remains an important concern for the
mining and metallurgical industries. Arsenic removal by schwertmannite has received great
attention in recent years due to the high specific surface area of schwertmannite and its strong
binding affinity to arsenic (Bigham et al., 1994; Fukushi et al., 2003; Regenspurg & Peiffer, 2005).
Schwertmannite is a poorly crystalline Fe(III)-oxyhydroxysulfate mineral (Fe8O8(OH)8-2x(SO4)x; x =
11.75), which is commonly formed in acidic (pH 3.0-4.0) iron- and sulfate-rich (10003000 mg/L)
environments (Bigham et al., 1994; Bigham & Nordstrom, 2000). Schwertmannite is also known to
be a metastable mineral and over weeks or months it readily undergoes phase transformations to
more stable iron oxides such as goethite (FeOOH) under some pH and oxidizing conditions (Acero
et al., 2006; Antelo et al., 2012; Jnsson et al., 2005; Liao et al., 2011). However, it has been reported
that arsenic incorporated into schwertmannite retards or signicantl y inhibits the transformation to
goethite (Regenspurg & Peiffer, 2005; Fukushi et al., 2003). Extremely high arsenic concentrations in
acid mine drainage (AMD) have been reported to be naturally decreased to background levels by
incorporation into schwertmannite within short distance downstream of effluent outlets (Fukushi et
al. 2003).
Schwertmannite can be synthesized by several techniques, including dialysis by adding ferric
chloride/nitrate to sodium/potassium sulfate solutions (Bigham et al., 1994), oxidation of FeSO4
solutions by H2O2 (Regenspurg et al., 2004), biooxidation of FeSO4 solutions by Acidithiobacillus
ferrooxidans (Liao et al., 2009), and neutralization of acidic Fe2(SO4)3 solutions. On-site synthesis of
schwertmannite is of economic and environmental interest for arsenic removal in effluents
generated by mining activities. In porphyry copper mines, synthesis of schwertmannite from heap
leach solutions is possible because the solutions contain high Fe3+ and SO42-, the components of
schwertmannite, suggesting that these solutions can be used to synthesize schwertmannite.
However, the leaching solutions also contain a variety of metal ions, such as Fe2+, Cu2+, Mg2+, Zn2+,
Co2+, or Ni2+ as co-existing ions, and the effects of these ions on schwertmannite synthesis is not well
known. The present study is to investigate the effects of these co-existing ions on schwertmannite
synthesis; and the role of the synthesized products in arsenic removal and immobilization in acidic
aqueous solutions.

METHODOLOGY
Synthesis of schwertmannite
The synthesis of schwertmannite was carried out by neutralization of Fe2(SO4)3 solution. First, 500
mL of solution containing 10 mM H2SO4 and 50 mM Fe2(SO4)3 with or without 100 mM Fe2+ and 100
mM of MSO4 (M: Cu, Mg, Zn, Co, or Ni) was prepared. The solution was then heated to 65C in a
beaker under magnetic stirring at 350 rpm. After discontinuing the magnetic stirring, 1 M Na 2CO3
was titrated to the solution to adjust the final pH to 3 - 4 (titration rate, 1500 L/min). The
precipitates formed in the beaker were then harvested by filtering through 5A filter paper, dried at
40C in a vacuum oven for 24 hours, and subjected to X-ray diffraction (XRD) analysis.

As(V) sorption experiment

Arsenic sorption experiments were conducted in a 200-mL flask with 100 mg of dried synthetic
product and 100 ml of solutions containing known concentrations of As (V) at initial pH 3.0-4.0. The
flask was shaken for 24 hrs in a water bath at 25C at 120 rpm. Then, the solution was filtrated
through a membrane filter (pore size, 0.2-m) and analyzed for arsenic by ICP-AES. The amount of
sorbed arsenic was calculated from the differences between the initial (added) amount and final
(remaining) amount of As in the solutions.

Stability of As(V)-sorbed schwertmannite

With 100 mg of As(V)-sorbed schwertmannite (dry matter basis, prepared through As(V) sorption
experiment as described above) about ~10 wt.% of As(V) and 100 mL of deionized water was added
in a 200-mL flask. The suspension pH was adjusted to 3.0-4.0 by dropwise addition of 1 M
H2SO4/NaOH. The flask was then placed in water bath at 25C and agitated at 120 rpm. This
suspension was incubated for 120 days and at a given ageing time, the suspension was sampled
with a syringe and filtered through a 0.2-m membrane. The supernatant was analyzed for arsenic
and the solid phase was dried at 40C and examined by XRD.

Analytical methods
The solid phase was characterized by Xray diffraction (XRD) using a JEOL JDX-3500 powder
diffractometer with Cu K radiation (40 kV and 40 mA). The specimens were step scanned from 10
or 20 to 70 2 with continuous scans at a rate of 1/2 2/min. The XRD patterns were evaluated by
comparison with a schwertmannite standard with an XRD pattern identical to that of the
schwertmannite described by Bigham et al. (1990). The aqueous phase was analyzed for arsenic,
iron, and sulfur by inductively coupled plasma atomic emission spectroscopy (ICP-AES) using a
Seiko instruments SPS7800. A pH-meter and combined electrode with temperature compensation
was used to measure the pH and was calibrated regularly with standard buffer solutions of pH
1.68, 4.01, and 6.86.

RESULTS AND DISCUSSION

Effect of co-existing metal ions on schwertmannite synthesis
Figure 1 shows the XRD patterns of the synthesized products with or without Fe2+ and MSO4 (M:
Cu, Mg, Zn, Co, or Ni). The result confirms that schwertmannite is the only mineral detectable in
the product synthesized with the absence of the co-existing metal ions. At the initial Fe2+
concentration of 100 mM, goethite is the main mineral phase formed, indicating that Fe2+ has a
significant effect on the synthesis of schwertmannite. However, schwertmannite is the prominent
mineral with no goethite phase detected in the product synthesized in the presence of 100 mM Fe2+
and 100 mM Cu2+, suggesting that Cu2+ has the ability to inhibit goethite formation in the presence
of Fe2+. On the other hand, goethite is the main mineral phase formed for all the products
synthesized in the presence of 100 mM Fe2+ and 100 mM MSO4 (M: Mg, Zn, Co, or Ni), exhibiting
that these co-exisitng metal ions do not have the ability to prevent the geothite formation in the
presence of Fe2+.

Fe2+

Intensity

Fe2+ and Mg2+

Fe2+ and Zn2+

Fe2+ and Co2+

Fe2+ and Ni2+

Fe2+ and Cu2+
Without Fe2+ and MSO4

2 () CuK
Figure 1 XRD patterns for samples synthesized with or without 100 mM Fe 2+ and 100 mM of MSO4 (M: Cu, Mg,
Zn, Co, or Ni). Schwertmannite is denoted as sh and goethite as G

Goethite formation process in the presence of Fe 2+ may be assumed to occur due to the
transformation of intermediated schwertmannite during the synthesis. Such Fe 2+ is known to
catalyze the transformation of schwertmannite to goethite (Burton et al., 2010, 2008; 2007). The Fe2+catalyzed transformation of schwertmannite involves the surface complex formation between Fe2+
and schwertmannite. The electron transfer then occurs between the adsorbed Fe 2+ and structural
Fe3+; and this nascent structural Fe2+ destabilizes the schwertmannite surface layer, causing
reductive dissolution and the release of adsorbed Fe3+, structural Fe2+, SO42-, and OH-. Finally, the
desorbed Fe3+ rapidly hydrolyzes and precipitates mainly as goethite (Burton et al., 2008).
As mentioned above that goethite was not formed during the synthesis in the presence of Fe2+ and
Cu2+, but goethite was mainly formed in the presence of Fe 2+ and other metal ions (Mg2+, Zn2+, Co2+,
Ni2+). This may be assumed that Cu2+ would prevent the Fe2+adsorption onto the surface of
schwertmannite by mean of competitive adsorption. When Fe2+ (an electron donor) supplies
electrons to schwertmannite, Cu2+ (an electron acceptor) consumes some of these electrons, and the
reductive dissolution rate of schwertmannite is expected to decrease. Thus, Fe 2+-catalyzed
transformation of schwertmannite to goethite may be inhibited or retarded in the presence of Cu 2+.
On the contrary, Mg2+, Zn2+, Co2+, or Ni2+ would not prevent the Fe2+ adsorption onto the surface of
schwertmannite by mean of competitive adsorption. Thus, electron transfer occurs between the
adsorbed Fe2+ and structural Fe3+, causing the reductive dissolution of schwertmannite and goethite
is then formed. However, further investigation is required to establish the details of these
phenomena.
These results highlight the critical role that Cu2+ availability plays in controlling the synthesis of
schwertmannite from heap leach solutions, particularly in the presence of Fe2+ which acts as a
catalyst to transform schwertmannite to goethite.

As(V) sorption
As(V) removal from aqueous solution was investigated at initial As(V) concentrations of 0.5-3.0
mmol L-1 at pH 3-4. Sorption equilibration was assumed to be achieved within 24 hours as describe
in Carlson et al. (2002) and all the isotherm studies were conducted for 24 hours to allow maximum
As(V) sorption. Extensive studies have shown that As(V) is sorbed onto iron hydroxides, such as
schwertmannite, forming inner sphere surface complexes by ligand exchange with hydroxyl groups
at the mineral surface (Asta et al., 2009).
The As(V) sorption data for the synthesized products are showed in Figure 2. The results show that
only product synthesized in the presence of Cu2+ and Fe2+ has very similar As(V) sorption capacity
(~1.5 mmol g-1) to the product synthesized in the absence of co-existing metal ions. The sorption
capacity here is consistent with Antelo et al. (2012) who found that schwertmannite sorbed ~1.6
mmol g-1 at pH 4.5.
For the products synthesized with the presence of Fe2+ and MSO4 (M: Mg, Zn, Co, or Ni), the
maximum sorption capacities were found to be approximately 0.7 mmol g-1, which are consistent
with the sorption capacity of the product synthesized in the presence of Fe 2+ (goethite product). The
data here are similar to the finding reported by Lehmann et al. (2005) at pH 3 for natural goethite. A
lower As(V) sorption capacity on the products containing goethite observed here may be due to a
lesser number of available sorption sites on the formed goethite (Burton et al., 2008).

These results imply that the presence of Cu2+ in copper heap leach liquid may assist to synthesize
schwertmannite with high arsenic sorption capacity.

As(V) adsorbed (mmol g-1)

Without Fe2+ and MSO4

Fe2+ and Cu2+
Fe2+
Fe2+ and Mg2+
Fe2+ and Zn2+
Fe2+ and Co2+
Fe2+ and Ni2+

Dissolved As(V) concentration (mmol L-1)

Figure 2 Sorption isotherms of As(V) onto samples synthesized with or without 100 mM Fe2+ and 100 mM of
MSO4 (M: Cu, Mg, Zn, Co, or Ni)

Stability of As(V)-sorbed synthetic schwertmannite

As mentioned previously, schwertmannite is a metastable compound that is transformed to stable
crystalline goethite via dissolution and re-precipitation over time according to the general reaction
(Regenspurg et al., 2004):
Fe8O8 (OH)8-2x (SO 4 )x + 2xH 2 O 8FeOOH + xSO 2-4 + 2xH +

(1)

As(V), however is known to retard the slow transformation of schwertmannite to goethite under
oxic conditions (Burton et al., 2010; Fukushi et al., 2003; Regenspurg & Peiffer, 2005). In the present
study, stability of As(V)-sorbed schwertmannite was carried out at 25C and pH 3-4 for 120 days.
Here, the use of mineralogical characterization by XRD in combination with As release was applied
to observe the stability of schwertmannite.
The results show that no phases other than schwertmannite are detected (Figure 3), suggesting that
As(V)-sorbed schwertmannite exhibits no mineralogy phase change when aged at pH 3-4 for 120
days. Figure 4 shows a fraction of the As released from the solid phase during the aging period. The
results indicate that the amount of released As remains almost constant throughout the experiment;
and the highest concentration of released As is ~0.05 mmol L-1 corresponding to ~3.4% of the total
solid phase As. These results suggests the long-term stability of As(V)-sorbed schwertmannite
synthesized from copper heap leach solution.

Intensity

sh

sh
sh

sh

sh

sh
120 d

Initial
2 () CuK
Figure 3 XRD pattern for As(V)-sorbed schwertmannite stabilized at pH 3-4 for 120 days , compared to the
original pattern (initial). Schwertmannite is denoted as sh

As release (mmol L-1)

1.5

1.0

0.5

0.0

Time (days)
Figure 4 Fraction of released As with respect to time within 120 days at pH 3-4 of As(V)-sorbed
schwertmannite stability. The total amount of As in the solid (~1.5 mmol L-1) is indicated by a dashed line

CONCLUSION
This study investigates schwertmannite synthesis from copper heap leach solution, focusing on the
effect of co-existing metal ions on the synthesis pathway and its efficiency in As(V) removal from
acidic solution. Results show that schwertmannite can be synthesized from copper heap leach
solution. The co-existing metal ions had significant influence on the synthesis. Schwertmannite was
produced with the presence of Cu2+ and Fe2+, while goethite was mainly formed in the presence of
other metal ions (Mg2+, Zn2+, Co2+, Ni2+) and Fe2+; suggesting the critical role that Cu2+ availability can
play in controlling the synthesis of schwertmannite from copper heap leach solutions.
Batch sorption studies show that arsenic(V) in acidic solutions (pH 34) can be removed effectively
by synthesized schwertmannite with As(V) sorption capacity of ~1.5 mmol g-1. A lower As(V)
sorption capacity (~0.7 mmol g-1) was observed on products containing goethite. In addition, the

relative low As release occurred without any detectable mineralogical schwertmannite

transformation even after 120 days aging experiment at pH 3-4, suggesting the stability of As(V)sorbed synthetic schwertmannite in long-term.

ACKNOWLEDGMENT
This work is supported by a collaborative research program between Hokkaido University and
Sumitomo Metal Mining Co., LTD.

REFERENCES
Acero, P., Ayora, C., Torrento, C., Nieto, J.-M. (2006) The behaviour of trace elements during schwertmannite
precipitation and subsequent transformation into goethite and jarosite, Geochim. Cosmochim. Acta
70, pp. 41304139.
Antelo, J., Fiol, S., Gondar, D., Lpez, R., Arce, F. (2012) Comparison of arsenate, chromate and molybdate
binding on schwertmannite: Surface adsorption vs anion-exchange, J. Colloid Interface Sci. 386, pp.
338343.
Asta, M. P., Cama, J., Martnez, M., Gimnez, J. (2009) Arsenic removal by goethite and jarosite in acidic
conditions and its environmental implications, J. Hazard. Mater. 171, pp. 965972.
Bigham J. M. and Nordstrom D. K. (2000) Iron and aluminium hydroxysulfates from acid sulfate waters, Rev.
Miner. Geochem. 40, pp. 351403.
Bigham, J.M., Carlson, L., Murad, E. (1994) Schwertmannite, a new iron oxyhydroxysulfate from Pyhasalmi,
Finland and other localities, Mineralogical Magazine 58, pp. 641648.
Bigham, J. M., Schwertmann, U., Carlson, L., and Murad, E. (1990) A poorly crystallized oxyhydroxysulfate of
iron formed by bacterial oxidation of Fe(II) in acid mine waters, Geochim. Cosmochim. Ac., 54, pp.
27432758.
Burton, E.D., Johnston, S.G., Watling, K., Bush, R.T., Keene, A.F., Sullivan, L.A. (2010) Arsenic effects and
behavior in association with the Fe(II)-catalyzed transformation of schwertmannite, Environ. Sci.
Technol 44. pp. 2016-2021.
Burton, E.D., Bush, R.T., Sullivan, L.A., Mitchell, D.R.G. (2008) Schwertmannite transformation to goethite via
the Fe(II) pathway: reaction rates and implications for ironsulde formation, Geochim. Cosmochim.
Acta 72, pp. 45514564.
Burton, E. D., Bush, R. T., Sullivan, L. A., Mitchell, D. R. G. (2007) Reductive transformation of iron and sulfur
in schwertmannite-rich accumulations associated with acidied coastal lowlands, Geochim.
Cosmochim. Acta 71, pp. 44564473.
Carlson, L., Bigham, J.M., Schwertmann, U., Kyek, A., Wagner, F. (2002) Scavenging of As from acid mine
drainage by schwertmannite and ferrihydrite: a comparison with synthesized analogues. Environ. Sci.
and Technol. 36, pp. 17121719.
Choi, W.-H., Lee, S.-R., Park, J.-Y. (2009) Cement based solidification/stabilization of arsenic-contaminated
mine tailings. Waste Management 29, pp. 17661771.
Fukushi, K., Sasaki, M., Sato, T., Yanase, N., Amano, H., Ikeda, H. (2003) A natural attenuation of arsenic in
drainage from an abandoned arsenic mine dump, Appl. Geochem. 18, pp. 1267-1278.
Jnsson, J., Persson, P., Sjberg, S., Lvgren, L. (2005) Schwertmannite precipitated from acid mine drainage:
phase transformation, sulfate release and surface properties, Appl. Geochem. 20, pp. 179191.

Langsch, J. E., Costa, M., Moore, L., Morais, P., Bellezza, A., & Falco, S. (2012) New Technology for Arsenic
Removal from Mining Effluents, Journal of Materials Research and Technology 1, pp. 178-181.
Lehmann, M., Zouboulis, A.I., Matis, K.A., Grohmann, A. (2005) Sorption of arsenic oxyanions from a queous
solution on goethite: study of process modeling, Microchim. Acta 151, pp. 269275.
Liao, Y., Liang, J., Zhou, L. (2011) Adsorptive removal of As(III) by biogenic schwertmannite from simulated
As-contaminated groundwater, Chemosphere 83, pp. 295301.
Liao, Y., Zhou, L., Liang, J., Xiong, H. (2009) Biosynthesis of schwertmannite by Acidithiobacillus ferrooxidans
cell suspensions under different pH condition, Mater. Sci. Eng. C. 29, pp. 211215.
Long, G., Peng, Y., Bradshaw, D. (2012) A review of copperarsenic mineral removal from copper concentrates,
Miner. Eng. 36-38, pp. 179-186.
Regenspurg, S. and Peiffer, S. (2005) Arsenate and chromate incorporation in schwertmannite, Appl. Geochem
20, pp. 12261239.
Regenspurg, S., Brand, A., Peiffer, S. (2004) Formation and stability of schwertmannite in acid mining lakes,
Geochim. Cosmochim. Ac. 68, pp. 11851197.
Smedley, P. L., Kinniburgh, D. G. (2002) A review of the source, behavior and distribution of arsenic in natural
waters, Applied Geochemistry 17, pp. 517-568.

Possible role of copper species in

enhanced chalcopyrite dissolution

Cynthia Torres
Centro de Investigacin
Cientfico Tecnolgico para
la Minera (cicitem), Chile
Mara Elisa Taboada,
Tefilo Graber and
Osvaldo Herreros
Universidad de
Antofagasta, Chile
Helen Watling
and Denis Shiers
csiro Chile Centre of
Excellence in Mining and
Mineral Processing; csiro,
Australian Minerals
Research Centre
Pa Hernndez
csiro Chile Centre of
Excellence in Mining and
Mineral Processing

The results of a previous study on copper extraction from a low-grade

copper ore using acidified fresh or saline water indicated that seawater and
freshwater were equally suitable as solvents for leaching and that the rate of
copper extraction was markedly enhanced when cupric ions were added to
the system. When predicted concentrations of major copper ions were
compared with Cu extractions, it was found that the concentration of
copper monochloride ion (+1) (CuCl+) correlated directly with extraction.
This relationship is being examined more closely in the present study,
using a chalcopyrite concentrate and a matrix of solutions encompassing
five chloride concentrations, three cupric ion concentrations and two initial
solution pH set points. All tests are being conducted at 60C to obtain a
robust result in shortened leach duration. The presence of chloride and
cupric ion were important for the formation of chlorocomplex species,
especially CuCl+, the predicted concentration of which was directly
proportional to the measured copper extraction.

INTRODUCTION
Interest in chalcopyrite dissolution in chloride systems extends back decades and has generated a
number of process options, mostly involving high-concentration chloride solutions (brines) at or
near the boiling point of the solutions at atmospheric pressure. Most utilise ferric or cupric ions or
mixtures of the two as oxidants contained in sodium chloride brine (~45 M). Recent examples are
the Intec (Everett 1996) or HydroCopper (Hamalainen and Hyvarinen 1999) processes. These are
high intensity processes suitable for concentrates but not economic for low grade ores, where heap
leaching (or possibly vat leaching) provide low-intensity, lower-cost process options. The body of
work undertaken by Nicol and co-researchers (Nicol et al. 2010; Velsquez-Yvenes et al. 2010;
Yvenes et al. 2010; Miki and Nicol 2011) underpinned a recent patent for a chloride-heap leaching
process (Muller et al. 2011). In that work, it was proposed to control ORP (oxidation reduction
potential) in acidic mixed sulfate-chloride medium such that the chalcopyrite surface potential was
maintained in the range 550600 mV vs SHE. The chloride concentration was in the range 0.142.8
M, oxygen, at least 1 mg L1, was required for the reaction and the tests were conducted at 35 C.
Muller et al. (2011) asserted that iron did not play a direct role in the leaching mechanism but that
ORP determined the mixed potential at the mineral surface and controlled the mechanism of
chalcopyrite dissolution (reactions 13). Leaching at ORP <540 mV caused reduced rates of
chalcopyrite dissolution and covellite or chalcocite to form on some chalcopyrite surfaces.
CuFeS2 + 4H+ + 2SO42 CuSO4 + FeSO4 + 2H2S(aq)

(1)

CuFeS2 + H2SO4 = CuS + FeSO4 + H2S

(2)

H2S(aq) + O2 2S0 + 2H2O

(3)

(copper(II) catalysed reaction)

In a precursor to the present study (Torres et al. 2013), copper extraction from a low-grade mixed
ore (100 g ore L1 medium) was investigated using freshwater, seawater and a more-concentrated
process water acidified with H2SO4. In that study copper extraction was rapid, in the range 6570%
within an hour and similar in tap water or seawater media. In comparison, while extraction was
slow initially in simulated process water, it increased to >90% in a 20-hour leach, prompting
research to identify the key component that had caused the accelerated leaching rate. Limited data
indicated that the cupric monochloride complex ion CuCl + might play a key role (Figure 1).

Estimated Cu extraction [%]

100

90

80

70

y = 12.162x + 72.007
R = 0.9927

60
0

0.5

1.5

2.5

Cupric monochloride ion (+1) [M]

Figure 1 Relationship between predicted concentration of copper monochloride ion(+1)

with copper extraction from copper ore. (Torres et al. 2013)

In the present study, copper extraction from chalcopyrite concentrate was investigated using a suite
of solutions acidified with H2SO4 and encompassing five chloride concentrations, three cupric ion
concentrations and two initial solution pH set points. Soluble copper was measured during the
leaching experiments. The extraction data were analysed in respect of the roles of major copper
species present in initial acidified leach media.

MATERIALS AND METHODS

Materials
Analytical grade reagents: H2SO4, NaCl and CuSO4.5H2O, were used to prepare leach media. The
chalcopyrite concentrate, obtained from Mount Isa Mines (MIM), Australia, contained 24.5% Cu,
mainly as chalcopyrite (TABLES 1 & 2).
Table 1 Chemical analysis of chalcopyrite concentrate [%]
Cu

Fe

As

Si

Na

Al

Mg

Zn

Pb

24.5

25.9

28.6

0.126

5.19

0.036

<0.05

0.323

0.950

0.106

0.094

Table 2 Mineralogical analysis of chalcopyrite concentrate [%]

Mineral phase
Chalcopyrite
Melanterite
Calcite
Quartz
K-Jarosite
Pentlandite
Rmerite
Pyrite

Content
74
5
4.5
2.5
2.5
2
2
2

Ideal formula
CuFeS2
FeSO4.7H2O
CaCO3
SiO2
KFe3(SO4)2(OH)6
(Fe,Ni)9S8
Fe3(SO4)4.14H2O
FeS2

Mineral phase
Hematite
Bernalite
Magnesioferrite
Siderite
Ferrihydrite
Talc
Dwornikite
Magnesite

Content
1
1
1
0.5
0.5
0.5
0.5
0.5

Ideal formula
Fe2O3
Fe(OH)3
MgFe3+2O4
FeCO3
Fe5O3(OH)9
Mg3(Si2O5)2(OH)2
NiSO4.H2O
MgCO3

Copper extraction in stirred reactors

Tests were conducted in stirred instrumented reactors. Each reactor of 2000 cm3 volume capacity
had a glass top with multi-port lid on which a stirrer was mounted. A sample mass of 2.5 g CuFeS2
concentrate was weighed into each reactor containing 500 mL of defined media acidified with
H2SO4. Variables were Cu (0.52.5 g L1; 0.00790.0394 M), Cl (010 g L1; 00.28 M) and initial pH
(pH 1 or pH 2). Five-day leaching tests were conducted at 60 C to exploit the increased rate of
chalcopyrite dissolution at increased temperature. Solution samples were collected periodically for
analysis. Soluble Cu was measured in solution samples using atomic absorption spectroscopy.
Copper extractions were calculated after correction for the amount of copper present in the initial
solutions.

RESULTS AND DISCUSSION

Effect of Cu2+ concentration
While an initial copper concentration 0.0078 M was limiting, copper extractions were not affected
greatly at higher initial copper concentrations (Figure 2). Thus the presence of cupric ions is
necessary for accelerated chalcopyrite dissolution, but above the limiting concentration, rates of
dissolution are not enhanced for increased copper concentration. The loss of approximately 20%
copper from dilute copper solution by adsorption onto, or reaction with, the chalcopyrite surface
yielded a misleadingly-low estimate of extraction in pH 1 medium, as indicated by the adjusted
extraction data as if no adsorption had taken place. The adsorption or surface-reaction phenomenon
observed in the pH 1 low-copper medium did not occur in the pH 2 low-copper medium.

Estimated Cu extraction [%]

100
Cu 0.0078 M

80
60

Cu 0.0234 M

40
Cu 0.0391 M

20
0
-20
0

50

100

150

Cu 0.0078 M
adjusted f or
adsorption

Time [hours]

Figure 2 Initial copper concentration effect in leach medium on copper extraction.

The media were 0.2 M chloride at pH 1

A possible reaction between cupric ions and the chalcopyrite surface in the absence of chloride
could result in the formation of bornite (Equation 4). Warren et al. (1982) conducted an
electrochemical study of the dissolution of chalcopyrite in sulfuric acid solutions. On the basis that
iron was preferentially released from the chalcopyrite lattice, they proposed that a bornite (irondeficient chalcopyrite Cu1xFe1yS2z) intermediate product was formed, together with covellite
(CuS), under their experimental conditions. Mikhlin et al. (2004) also accepted that a disordered
metal-depleted phase Cu1xFe1yS2z formed on chalcopyrite surfaces but proposed a reaction that
formed Cu2+ rather than consuming it from solution (Equation 5). More recently Majuste et al. (2012)
using synchrotron small-angle X-ray diffraction and micro-Raman spectroscopy, detected bornite
and covellite (reaction 6) together with elemental sulfur on chalcopyrite surfaces leached in H 2SO4
media. In both cases, slow chalcopyrite dissolution was attributed to the formation of these
intermediate bornite-like or covellite-like phases on the chalcopyrite surface. At higher initial
copper concentrations, the adsorption-phenomenon is not discernible.
3CuFeS2 + 2Cu2+ = Cu5FeS4 + 2Fe2+ 2S0

(4)

CuFeS2 xCu2+ + yFe2+ + Cu1xFe1yS2z +zS0 + 2(x+y)e

(5)

2CuFeS2 Cu + 2Fe + CuS + 3S + 8e

(6)

2+

3+

Effect of pH
Copper extraction was faster in chloride media initially pH 1 compared with pH 2 (Figure 3). The
results are consistent with those of Velsquez-Yvenes et al. (2010) who reported that extraction
rates are almost independent of pH in the range 0.5-2.0 but that a fairly low pH is necessary to
keep iron(III) in solution.

Estimated Cu extraction [%]

100
Cu 0.0234 M pH 1

80

Cu 0.0234 M pH 2

60

Cu 0.0391 M pH 1

40

Cu 0.0391 M pH 2
20
0
0

50

100

150

Time [hours]

Figure 3 Effect of initial solution pH on copper extraction. The media were 0.143 M chloride

Effect of Cl- concentration

The presence of chloride in the initial leach medium resulted in increased copper extraction,
compared with the chloride-free medium (Figure 4). This result is consistent with comparative data
mostly obtained from ferric sulfate and ferric chloride leach media (Dutrizac 1981). However, the
rate of chalcopyrite dissolution is not greatly affected by the chloride concentration. This result is
also consistent with the data of Velsquez-Yvenes et al. (2010), who note that chloride ion plays an
important role in the auto-oxidation of ferrous and cuprous ions (Miki and Nicol 2008) (reactions 7
and 8), which is essential in a system not catalysed by microorganisms.
Estimated Cu extraction [%]

100
80
60

Cl 0 M
Cl 0.143 M

40

Cl 0.200 M
Cl 0.286 M

20
0
0

50

100

150

Time [hours]

Figure 4 Effect of chloride concentration on copper extraction from chalcopyrite.

Initial copper concentration was 0.0391 M in media at pH 1

Fe(II) + Cu(II) = Fe(III) + Cu(I)

(7)

4Cu(I) + O2 + 4H+ 4Cu(II) + H2O)

(8)

Predicted distribution of major copper species and free chloride in leach media
The main difference between this work and previous studies (Yvenes et al. 2010), is that solution
compositions were chosen to encompass a wide range of concentrations of the key components
copper and chloride.
A model of copper speciation in the experimental range predicted that, for a single copper
concentration (example data shown for an intermediate Cu concentration; Figure 5) but increased
Cl concentration, Cu+2 concentration decreased significantly while the chlorocomplex species CuCl+
concentration increased. Another important point to highlight is that as the initial copper
concentration was increased, the chlorocomplex species concentration also increased.
Chlorocomplex species concentrations were obtained using Excel with Solver, applying the
constants of formation (K1K2K3K4) from TABLE 3.

Figure 5 Predicted copper speciation [M] as a function of chloride concentration.

Initial copper concentration was 0.0234 M in media pH 1

It was proposed to investigate in greater detail conditions that maximized CuCl+ concentration in
solution and describe the relationship between CuCl+ concentration and copper extraction. In
general terms, from Table 4 it was determined that the higher the concentration of CuCl+ the greater
the rate of copper extraction.

Table 3 Formation of chlorocomplexes

G
(kJ/mol)

Cu2+ + Cl CuCl+

-2.198

K1=2.426

(9)

CuCl+ + Cl CuCl2(aq)

6.653

K2=0.068

(10)

CuCl2(aq) + Cl CuCl3

9.131

K3=0.025

(11)

CuCl3 + Cl CuCl42

13.142

K4=0.005

(12)

Reaction

Reaction

Table 4 Copper extraction [%] at pH 1 correlated with predicted copper speciation at different
chloride and copper concentrations
Total Cl
[M]

Total Cu
[M]

Cu+2
[M]

CuCl+
[M]

Extraction
[%]

0.0078

0.0078

16

0.0234

0.0234

24

0.0391

0.0391

30

0.143

0.0078

0.0058

0.0020

39

0.200

0.0078

0.0053

0.0025

55

0.286

0.0078

0.0046

0.0032

57

0.143

0.0234

0.0176

0.0058

60

0.200

0.0234

0.0159

0.0074

67

0.286

0.0234

0.0139

0.0093

67

0.143

0.0391

0.0295

0.0095

57

0.200

0.0391

0.0267

0.0122

58

0.286

0.0391

0.0234

0.0153

57

CONCLUSIONS
At low initial copper concentration and pH 1, copper was removed from solution in the first 24
hours of leaching. This could be due to adsorption of cupric ions or the formation of a disordered
bornite-like phase on the chalcopyrite surface. This phenomenon was not discerned at pH 2 or in
higher-concentration copper solutions. An increase in the initial copper concentration resulted in
increased copper extraction from chalcopyrite. Copper extraction was faster in media initially pH 1
compared with pH 2.
Copper extraction increased with increase chloride concentration but chloride itself was not the key
component. Using a thermodynamic model, it was predicted that CuCl+ was the dominant copper
species in the system at the concentrations studied. In summary, it can be concluded that, while the
chloride concentration had a minor effect on copper extraction in the ranges tested, its presence was
important for the formation of chlorocomplex species, especially CuCl +, the predicted concentration
of which was directly proportional to the measured copper extraction.

ACKNOWLEDGEMENTS
Assistance in ore preparation by the Mining Engineering and Chemical Engineering Departments
of UA is acknowledged with thanks. The authors are grateful for the financial support provided by
CICITEM R04I1001, CSIROChile International Centre of Excellence, and CONICYT through
Fondecyt Project N 1140169.

REFERENCES
Dutrizac, J. (1981) 'The dissolution of chalcopyrite in ferric sulfate and ferric chloride media.' Metallurgical
Transactions B, 12, (2), pp. 371-378.
Everett, P. K. (1996). Production of metals from minerals, U.S. Patent 5,487,819.
Hamalainen, M. and O. Hyvarinen (1999). Method for producing copper in hydrometallurgical process, U.S.
Patent 6,007,600.
Majuste, D., V. Ciminelli, et al. (2012) 'Electrochemical dissolution of chalcopyrite: Detection of bornite by
synchrotron small angle X-ray diffraction and its correlation with the hindered dissolution process.'
Hydrometallurgy, 111, pp. 114-123.
Mikhlin, Y. L., Y. V. Tomashevich, et al. (2004) 'Spectroscopic and electrochemical characterization of the
surface layers of chalcopyrite (CuFeS2) reacted in acidic solutions.' Applied Surface Science, 225, (1), pp.
395-409.
Miki, H. and M. Nicol (2008) 'Hydrometallurgy 2008 Sixth International Symposium', The kinetics of the coppercatalysed oxidation of iron (II) in chloride solutions, Phoenix, USA, pp. 971-979
Miki, H. and M. Nicol (2011) 'The dissolution of chalcopyrite in chloride solutions. IV. The kinetics of the autooxidation of copper(I).' Hydrometallurgy, 105, (34), pp. 246-250.
Muller, E. L., P. Basson, et al. (2011). Chloride heap leaching, U.S. Patent 8.070,851.
Nicol, M., H. Miki, et al. (2010) 'The dissolution of chalcopyrite in chloride solutions: Part 3. Mechanisms.'
Hydrometallurgy, 103, (14), pp. 86-95.
Torres, C. M., M. E. Taboada, et al. (2013) 'Copper extraction from oxide ore using seawater', Hydroprocess2013
5th International Seminar on Process Hydrometallurgy, Santiago, Chile, pp. 69-77
Velsquez-Yvenes, L., M. Nicol, et al. (2010) 'The dissolution of chalcopyrite in chloride solutions: Part 1. The
effect of solution potential.' Hydrometallurgy, 103, (14), pp. 108-113.
Warren, G., M. Wadsworth, et al. (1982) 'Passive and transpassive anodic behavior of chalcopyrite in acid
solutions.' Metallurgical Transactions B, 13, (4), pp. 571-579.
Yvenes, L. V., H. Miki, et al. (2010) 'The dissolution of chalcopyrite in chloride solutions: Part 2: Effect of
various parameters on the rate.' Hydrometallurgy, 103, (14), pp. 80-85.

Tellurium precipitation from copper

sulfate leach solutions

Yusuf Bello and

Christie Dorfling
University of Stellenbosch,
South Africa

The presence of selenium and tellurium in copper sulfate leach solutions is

undesirable because of its detrimental effect on electrowinning efficiency
and copper cathode quality. Conventionally, these elements are removed by
means of precipitation with sulfurous acid. This process is effective for the
removal of selenium, but the precipitation of tellurium remains
challenging. Tellurium removal has been reported to be enhanced by the
addition of metallic copper. The objective of this study was to determine at
which conditions metallic copper addition would result in increased
tellurium precipitation and to develop a better understanding of the Te
precipitation mechanism.
Batch precipitation tests were performed by bubbling SO2 through an
industrial CuSO4 leach solution in a reactor with a working volume of 1 L.
The temperature was varied between 65 and 95C for SO2 flow rates of 4.35
and 5.80 mL/min and agitation rates of 250 and 500 rpm. Copper was added
as powder or as plates.
The effects of process variables on Te precipitation were similar for the
tests performed without any copper addition and those with Cu plate
addition. For these tests, the agitation speed and SO2 flow rate had no effect
on the Te precipitation at 95C; 92% and 94% Te precipitation were
obtained on average for the tests performed with and without Cu addition,
respectively. At 65C, better precipitation was achieved by increasing the
SO2 flow rate and decreasing the agitation speed, while copper plate
addition also increased Te precipitation from 77 to 83%. When adding Cu in
the form of powder, none of the process variables affected the Te
precipitation. The addition of Cu powder resulted in 93% Te precipitation
on average. In general, the addition of Cu, either as powder or as plates,
also resulted in significantly faster precipitation kinetics.

INTRODUCTION
Impurities such as selenium and tellurium that are present in copper sulfate leach solutions have a
negative effect on electrowinning efficiency and copper cathode quality. These species therefore
have to be removed from the leach solution prior to electrowinning. One method commonly
employed for the removal of these species is precipitation of Se and Te as Cu 2Se and Cu2Te by the
addition of sulfurous acid. In general, this purification process achieves satisfactory Se
precipitation, but tellurium removal remains challenging.
Lottering, Eksteen, and Steenekamp (2012) reported that other precipitation reagents such as
thiosulfate, formaldehyde, formic acid, and thiourea did not result in improved precipitation
compared to sulfurous acid at similar operating conditions. At a temperature of 80C, none of the
reagents resulted in noticeable Te precipitation. At 115C and higher temperatures, thiourea and
sulfurous acid achieved comparable Te precipitation.
Similar results for Te precipitation by sulfurous acid addition were reported by Wang, Wesstrom,
and Fernandez (2003), who indicated that the reductive precipitation of Te from a sulfuric acid
solution by sulfurous acid addition alone is unlikely. It was observed that any Te precipitation was
likely to be because of a secondary cementation reaction between dissolved Te and metallic copper
in contact with the solution. Metallic copper can be produced by the disproportionation of cuprous
brought about by the reduction of cupric ions by SO2. The formation of cuprous is given by
Equation 1 (Crundwell et al., 2011), while the disproportionation of cuprous ions is illustrated by
Equation 2 (Wang, Wesstrom, and Fernandez, 2003).
2CuSO4 + H2SO3 + H2O Cu2SO4 + 2H2SO4
Cu2SO4 CuSO4 + Cu

(1)
(2)

It was concluded that the addition of metallic copper would be necessary to enhance tellurium
precipitation, and that temperatures above 92C are required if Te precipitation were to occur
without any copper addition. Experimental results indicated that the addition of copper chips
resulted in significantly higher tellurium precipitation when using sulfurous acid as precipitation
reagent (Wang, Wesstrom, and Fernandez, 2003). These authors proposed the precipitation
reactions between metallic copper and tellurium ions to be as given by Equations 3 and 4:
4Cu + H2TeO3 + 2H2SO4 Cu2Te + 2CuSO4 + 3H2O

(3)

5Cu + H6TeO6 + 3H2SO4 Cu2Te + 3CuSO4 + 6H2O

(4)

The enhancement of tellurium precipitation through the addition of metallic copper has also been
investigated by Jennings, Themelis, and Stratigakos (1969), Shibasaki, Abe, and Takeuchi (1992),
and Sugawara et al. (1992).
Jennings, Themelis, and Stratigakos (1969) studied the kinetics of tellurium precipitation in a
rotating drum reactor at different temperatures and with different masses of copper shots added.
The precipitation of tellurium was proposed to occur by Equation 4. It was found that temperature
plays a key role in tellurium precipitation kinetics while the effect of agitation was minimal. It was
furthermore proven that tellurium precipitation is a first order reaction with respect to Te
concentration in the solution (Jennings, Themelis, and Stratigakos, 1969).

The kinetics of tellurium precipitation has also been studied in a fixed-bed reactor filled with
copper pieces of small and medium sizes (Shibasaki, Abe, and Takeuchi, 1992). The copper sulfate
solution was heated to the desired temperature (ranging from 60C to 90C), where after it was
passed through the bed of copper pieces. This study confirmed that the rate of tellurium
precipitation has a first order dependence on the tellurium concentration. It was furthermore
observed that reaction rate was dependent on the specific surface area of the copper pieces
(Shibasaki, Abe, and Takeuchi, 1992). The effect of the specific copper surface area on precipitation
kinetics was further elucidated by Sugawara et al. (1992), who observed that the addition of copper
powder resulted in increased tellurium precipitation kinetics at 70C. It was, however, suggested
that the utilization of copper plates would allow easier operation since cuprous telluride deposits
can be removed more easily than when using small copper pieces or copper powder (in which case
the deposit on the copper surface could prevent further precipitation from taking place).
It can be concluded that the addition of metallic copper is essential for the removal of tellurium by
precipitation with sulfurous acid if the precipitation is to be conducted at temperatures below 92C.
The objective of this study was to evaluate the effect of the form and amount of copper addition on
tellurium precipitation efficiency at different temperatures, precipitation reagent addition rates, and
agitation rates.

EXPERIMENTAL
Experimental design
Experiments were conducted to compare the tellurium precipitation efficiency for following three
processes utilizing SO2 as precipitation reagent: a) no metallic copper addition, b) metallic copper
addition in the form of copper powder, and c) metallic copper addition in the form of copper plates.
All the experiments for the three respective processes were conducted according to full factorial
experimental designs. The operating conditions that were varied include the temperature (65C and
95C), SO2 flow rate (4.35 mL/min and 5.80 mL/min), and the agitation speed (250 rpm and
500 rpm). For experiments using copper as precipitation enhancing reagent, the amount of copper
powder (1 g/L and 2 g/L) and the size of the copper plates (202.5 mm2/L and 405 mm2/L) were
varied, respectively, in addition to the abovementioned process variables.

Materials
The copper sulfate leach solution was obtained from the base metal refinery of a platinum group
metal producer. The tellurium concentration was approximately three orders of magnitude lower
than the copper concentration. Sulfur dioxide (99.9 % purity) was supplied by Afrox. Copper
powder (45 m particle size and 99.9 % purity) was supplied by Sigma Aldrich. A 1.5 mm thick
copper sheet was cut into pieces with dimensions of 10 mm by 7.5 mm to be used in the tests
utilizing copper plates.

Equipment
All the experiments were performed in a Bchi Polyclave Type 3 pressure vessel with a volumetric
capacity of two liters. The reaction vessel was equipped with a programmable temperature
controller and a Cyclone 300 magnetic stirrer drive. An anchor stirrer was used for all the tests. A
variable area flow meter manufactured by Aalborg and equipped with a needle valve was used to

measure and manually control the SO2 flow rate. Sampling was done through the drain valve
situated at the bottom of the vessel.

Procedure
One liter of the process solution was introduced into the reaction vessel, where after the vessel was
sealed, the stirrer switched on, and the solution heated to the desired temperature. Once the
solution had reached the desired temperature, a sample of the solution was taken to confirm the
initial solution composition. For tests that investigated the effect of copper addition, the copper was
added to the vessel. The flow of sulfur dioxide into the vessel was initiated directly after copper
addition. The sulfur dioxide was bubbled into the solution for 3 hours via a tube connected to the
reactor cover plate and held vertically within the pressure vessel.
Samples were taken at the following intervals (in minutes) after the sulfur dioxide flow had been
initiated: 0, 10, 20, 30, 60, 120, 240, 360, and 480. All samples were analyzed using inductively
coupled plasma atomic emission spectroscopy (ICP-AES), and the percentage metal precipitation
calculated by mass balance.

RESULTS AND DISCUSSION

Effect of agitation on Te precipitation
In the case of the tests performed at 65C with SO2 only, the tests performed at lower agitation rates
achieved higher percentages Te precipitation after eight hours (80 % Te precipitation on average)
than the tests performed at the higher agitation rates (73 % Te precipitation on average). Similar
results were observed for the tests performed at 65C with Cu plate addition; the lower agitation
rate and the higher agitation rate achieved 88 % Te precipitation and 77 % Te precipitation,
respectively. The results for these tests performed at 65C and a SO2 flow rate of 4.35 mL/min are
shown in Figure 1(a) and (b) as an illustration of the typical Te precipitation behavior observed. At
95C, the agitation rate did not significantly influence the percentage Te precipitation achieved after
eight hours for either the tests utilizing SO2 only or the tests utilizing Cu plates as precipitation
enhancement reagent.
Figure 1(c) shows the effect of agitation on tellurium precipitation achieved with the addition of Cu
powder as precipitation enhancing reagent for the test performed at 65C at a SO 2 flow rate of
4.35 mL/min. No difference in the precipitation behavior was observed for the different agitation
rates. Similar observations were made for the tests performed at different temperatures, agitation
rates, and SO2 flow rates.

Effect of SO2 flow rate on Te precipitation

As was the case for the effect of the agitation rate on Te precipitation behavior, the effect of SO 2
flow rate was only significant for the tests performed with SO2 only and with Cu plate addition at a
temperature of 65C. For the tests performed at 65C with SO 2 only, the percentage tellurium
precipitation achieved after 8 hours increased from 69 % to 84 % as the SO2 flow rate was increased
from 4.35 to 5.80 mL/min. The results achieved with the different processing techniques at 65C and
with an agitation rate of 250 rpm are shown in Figure 2 for illustrative purposes. From Figure 2(a) it
can be seen that the higher SO2 flow rate also contributed to increased precipitation kinetics and

allowed the precipitation to proceed at an earlier point in time. It is likely that the increase in the
SO2 flow rate enhanced the rate of reduction of cupric ions to cuprous ions according to Equation 1.
As a result, the formation of elemental copper according to Equation 2 and, subsequently, the
precipitation of tellurium (Equations 3 and 4) proceeded at a faster rate.

80

80

% Te precipitated

100

% Te precipitated

100

60
40
20
0
0

100

200

300

400

60
40
20
0
0

500

100

Time (minutes)
250 rpm

200

300

400

500

Time (minutes)
500 rpm

250 rpm

(a)

500 rpm

(b)

% Te precipitated

100
80
60
40
20
0
0

100

200

300

Time (minutes)

250 rpm

400

500

500 rpm

(c)
Figure 1 Effect of agitation on Te recovery with the use of: a) SO2 only, b) SO2 and Cu plate (202.5 mm2/L),
and c) SO2 and Cu powder (1 g/L)

With the addition of copper plates, Te precipitation started at an earlier point in time in general, as
can be seen from Figure 2(b). The effect of the SO2 flow rate on the precipitation kinetics was less
profound than for the tests performed with SO2 only; increasing the SO2 flow rate did, however,
result in an increase from 77 % to 88 % Te precipitation achieved after 8 hours. It can thus be
concluded that, although metallic copper was added, the specific surface area of the copper plates
was not sufficiently large to eliminate the dependence of tellurium precipitation on the reduction of
cupric ions and the formation of elemental copper from the process solution.
The SO2 flow rate did not affect either of the two processes discussed above at a temperature of
95C. At this temperature, the rate of cupric reduction was sufficiently fast with both SO2 flow rates

that the tellurium precipitation occurred at similar rates. There was also no significant effect of SO2
flow rate on the Te precipitation achieved after 8 hours when using copper powder as the
precipitation enhancing reagent. With the copper powder as precipitation enhancing reagent, the
large specific surface area of the metallic copper allowed tellurium precipitation to proceed mainly
by reaction with the added metallic copper; any metallic copper formation as a result of SO2
addition did not significantly influence the tellurium precipitation.

80

80

% Te precipitated

100

% Te precipitated

100

60
40
20
0

60
40
20
0

100

200

300

Time (minutes)

4.35mL/min

400

500

100

5.80mL/min

200

300

Time (minutes)

4.35mL/min

(a)

400

500

5.80mL/min

(b)

% Te precipitated

100
80
60
40
20
0
0

100

200

300

Time (minutes)

4.35mL/min

400

500

5.80mL/min

(c)
Figure 2 Effect of SO2 addition on Te recovery with the use: a) SO2 only, b) SO2 and Cu plate (202.5 mm/L),
and c) SO2 and Cu powder (1 g/L)

Effect of temperature on Te precipitation

As discussed before, the rate of cupric reduction and the subsequent elemental copper formation
appeared to increase as the temperature was increased. As a result, the rate of tellurium
precipitation was noticeably faster at 95C than at 65C for tests utilizing SO 2 only or SO2 with
copper plate addition as the precipitation reagents. One example of the faster kinetics observed is
illustrated in Figure 3(a) and (b) for the tests performed at an agitation rate of 250 rpm and a SO2

flow rate of 4.35 mL/min. In addition, the percentage tellurium precipitation achieved after eight
hours was higher at 95C; for the tests utilizing SO2 only and SO2 with copper plate addition, the
percentage tellurium precipitation increased from 77 % to 94 % on average and from 83 % to 92 %
on average, respectively.
The precipitation tests utilizing copper powder, where the tellurium precipitation is not dependent
on the reduction of cupric and the formation of elemental copper, did not show a large dependence
on the temperature. On average, the percentage tellurium precipitation achieved after eight hours
were 94 % at high temperature conditions and 92 % at low temperature conditions. The slightly
faster precipitation kinetics observed for the higher temperature test in Figure 3(c) could be an
indication that the rate of the precipitation reaction between dissolved tellurium and metallic
copper (Equations 3 and 4) was chemical reaction controlled.
100

80

% Te precipitated

% Te precipitated

100

60
40
20

80
60
40
20
0

0
0

100

200

300

Time (minutes)
65C

400

500

100

200

300

Time (minutes)
65C

95C

(a)

400

500

95C

(b)

% Te precipitated

100
80
60
40
20
0
0

100

200

300

Time (minutes)
65C

400

500

95C

(c)
Figure 3 Effect of temperature on Te recovery with the use of: a) SO2 only, b) SO2 and Cu plate (202.50 mm2/L),
and c) SO2 and Cu powder (1 g/L)

Effect of Cu addition on Te precipitation

Considering the tests performed at 65C, as shown in Figure 4(a), the precipitation kinetics
achieved when using copper powder as the precipitation enhancing reagent is noticeably faster
than when using copper plates as the precipitation enhancing reagent; the addition of copper plates
does, however, improve the precipitation kinetics compared to the utilization of SO 2 only. This is in
agreement with the earlier discussion regarding the dependence of tellurium precipitation on the
reduction of cupric and the formation of elemental copper. The amount of copper powder did not
affect the percentage tellurium precipitation achieved after 8 hours; the precipitation kinetics did,
however, appear to be slightly faster for the tests conducted with 2 g/L Cu than for the tests
conducted with 1 g/L Cu due to the larger solid surface area. Similar observations were made for
the precipitation kinetics when using Cu plates; for these tests, however, the percentage tellurium
precipitation achieved after eight hours was also lower for the smaller number of plates.

100

100

80

80

% Te precipitated

% Te precipitated

At 95C, the extents of tellurium precipitation achieved after eight hours were comparable for all
the tests, regardless of whether copper was added as precipitation enhancing reagent or not.
Although the tests performed with copper powder continued to exhibit the fastest precipitation
kinetics, the differences in the precipitation rates obtained with the different processing techniques
were significantly smaller than at 65C. The effect of the amount of copper available was also less
noticeable for the tests performed at the higher temperature. No significant co-precipitation of
copper was noticed for any of the tests performed regardless of the methods used and the operating
conditions involved.

60
40
20
0

60
40
20
0

100

200

300

400

500

100

200

300

400

Time (minutes)

Time (minutes)

SO only

SO only

SO & 202.5 mm Cu plate

SO & 202.5 mm Cu plate

SO & 405.0 mm Cu plate

SO & 405.0 mm Cu plate

SO & 1g Cu powder

SO & 1g Cu powder

SO & 2g Cu powder

SO & 2g Cu powder

(a)

500

(b)

Figure 4 Effect of copper addition on Te recovery at: a) 65C, and b) 95C for a SO 2 flow rate of 4.35 mL/min
and an agitation speed of 250 rpm

CONCLUSIONS
The effects of agitation rate (250 rpm and 500 rpm), SO2 flow rate (4.35 mL/min and 5.8 mL/min),
temperature (65C and 95C), and copper addition on tellurium precipitation from a copper sulfate
leach solution were investigated.
The agitation rate and the SO2 flow rate only affected tellurium precipitation in the tests performed
at 65C utilizing SO2 only or SO2 with copper plates as precipitation reagents. At 95C, the
reduction of cupric ions by SO2 and the subsequent metallic copper formation and tellurium
precipitation were noticeably faster than at 65C. It was concluded that the specific surface area of
the copper plates was too small to allow the precipitation reaction between dissolved tellurium and
metallic copper to proceed at a sufficiently fast rate; as a result, the tellurium precipitation in this
process was still dependent on the cupric reduction and elemental copper formation to proceed.
With the copper powder as precipitation enhancing reagent, the large specific surface area of the
metallic copper allowed tellurium precipitation to proceed mainly by reaction with the added
metallic copper; any metallic copper formation as a result of SO2 addition did not significantly
influence the tellurium precipitation. The tellurium precipitation kinetics achieved when using
copper powder as precipitation enhancing reagent were consequently noticeably faster than for the
other precipitation processes.

ACKNOWLEDGEMENT
Material and financial support provided by Lonmin Plc is gratefully acknowledged.

REFERENCES
Crundwell, F., Moats, M., Ramachandran, V., Robinson, T., and Davenport, W., (2011), Extractive metallurgy of
nickel, cobalt and platinum group metals, Elsevier Science Limited.
Jennings, P.H., Themelis, N.J., and Stratigakos, E.S., (1969), A continuous-flow reactor for the precipitation of
tellurium, Canadian Metallurgical Quarterly, 8(3):281-286.
Lottering, C., Eksteen J.J., and Steenekamp, N., (2012), Precipitation of rhodium from a copper sulphate leach
solution in the selenium/tellurium removal section of a base metal refinery, Journal of the Southern
African Institute of Mining and Metallurgy, 112(4):287-294.
Mokmeli, M., Wassink, B., and Dreisinger, D.B., (2012), Equilibrium cuprous concentrations in copper
sulphate-sulphuric acid solutions containing 50-110 g/L Cu2+ and 10-200 g/L H2SO4 at 50-95C,
Hydrometallurgy, 121-124:100-106.
Shibasaki, T., Abe, K., and Takeuchi, H., (1992), Recovery of tellurium from decopperising leach solution of
copper refinery slimes by a fixed bed reactor, Hydrometallurgy, 29(1-3):399-412.
Sugawara, Y., Hayashi, M., Konishi, J., and Hayashi, S., (1992), Process for recovering tellurium from copper
electrolysis slime, United States Patent, Patent number 5 160 588.
Wang, S., Wesstrom, B., and Fernandez, J., (2003), A novel process for recovery of Te and Se from copper
slimes autoclave leach solution, Journal of Minerals & Materials Characterization & Engineering, 2(1):5364.

In situ measurements of copper and

moisture in an operating heap using
advanced geophysics

Kent Lang and

Ned Clayton
Schlumberger, usa
Roland Banas
Schlumberger, Mexico
Ricardo Ruano
Schlumberger, Chile

A test section of the upper-most lift and underlying lifts of a large scale
copper heap leach operation was instrumented and monitored with
advanced wireline geophysics, fiber optics and resistivity electrode arrays
as well as in situ sampling. Elemental spectroscopy sonde copper
measurements, true resistivity measurements from an array induction
imager (AIT) tool, moisture and saturation measurement from a
combination of advanced measurements, electrical resistivity tomography
(ERT) and distributed temperature sensing (DTS) with fiber optics provided
integrated near real time data about the performance of the heap.
During the leaching cycle, capture spectroscopy provided continuous
vertical measurements of copper in the upper most and underlying lifts in
the test section of the heap and was used to generate a copper recovery
curve. ERT inversion modeling provided high resolution 3D resistivity
measurements, which were used to evaluate moisture distribution. DTS
measurements of temperature were used to track the pregnant leaching
solution (PLS) front and calculate velocity. AIT induction provided highly
precise bulk resistivity measurements. In situ samplers provided PLS
chemistry at multiple points inside the heap.
Properly integrated monitoring data provided the heap leach operator
with information that was used to evaluate the hydrodynamics and
metallurgy of the leaching process in near real time. Problems in the heap,
such as pooling, channeling, longer lag and breakthrough times, excess
acid consumption, increased solution in inventory, toe saturation and slope
stability were able to be detected in near real time with the in situ
measurements. These evaluations resulted in a better understanding of the
performance of the heap leaching process and resulted in an ability to
make changes to the process to improve leaching efficiency.

INTRODUCTION
For the past three years, Schlumberger has been working with the mining industry to develop and
conduct proof of concept testing of monitoring applications that can provide in-situ measurements
of the hydrometallurgical processes in a heap under leach. The objective of this work is to provide
operators with information that can be used to increase production, reduce costs and minimize
risks. The key objectives established at the start of proof of concept testing included the following:
Measure copper-in-solution production from the uppermost lift of a multi-lift heap or
stockpile in near real-time.
Track copper inventory in near real-time.
Track PLS grade front as it moves through the heap/stockpile.
Track copper extraction from a given block of ore in near real time.
Measure solution and gas permeability and moisture content at multiple representative
locations within the heap.
This article provides the results of measurements made in an engineered multi-lift heap with a
combination of conventional and advanced geophysics applications and focuses on the distribution
of copper, resistivity, moisture, saturation, and temperature. The owner and location of the heap are
confidential.

METHODOLOGY
A test section of the upper most lift and underlying lifts of an engineered heap under leach was
specially instrumented with advanced wireline (WL) geophysics monitoring ports, fiber optics and
resistivity electrode arrays. ECS elemental cspectroscopy sonde measurements of copper, true
resistivity measurements from an AIT array induction imager tool, electrical resistivity tomography
(ERT) and fiber optics distributed temperature sensing (DTS) were used, in part, to make in-situ
measurements and evaluate the hydrodynamics of the leaching process in near real time.
Additional advanced geophysics WL logs were acquired using the accelerator porosity sonde
(APS1), which measures volumetric water content, the Triple Detector Litho-Density (TLD ) tool,
which measures bulk density, and the combinable magnetic resonance (CMR) tool, which measures
total as well as moveable and bound volumetric water content. Porosity, saturation and
permeability were also determined through joint processing of these logs. A diagram that shows
the design and instrumentation layout for the test section is displayed in Figure 1.
ERT arrays and DTS cable were installed at the bottom (before ore stacking) and top (after ore
stacking) of the uppermost lift. A total of nine sonic borings were drilled through the upper most
and underlying lifts to a total depth of 150 feet, representing a total of eight lifts. Three borings were
cased with fiberglass and used for WL logging. Fiber optics cables and ERT electrode arrays were
installed together in three of the borings, and ERT electrode arrays only in three of the borings, all
to total depths of approximately 140 feet. Although other instruments were installed and used to
take measurements in the test section, their use and results are not within the scope of this article.

Mark of Sclhumberger

Figure 1 Test Section Design and Layout of Instrumentation

Raffinate was applied to the test section with drip emitters and monitoring was conducted for the
first 90 days of the leach cycle. Frequency of measurements is summarized in TABLE 1.
Table 1 Monitoring Frequency by Method
Method

Month 1

Month 2 through 3

ECS
AIT
APS
TLD
ERT
DTS

2 to 4/week
2 to 4/week
2 to 4/week
2 to 4/week
2/day
140/day

1/month
1/month
1/month
1/month
1/week
1/week

The ERT instrumentation monitoring and data processing was performed by HydroGEOPHYSICS
(HGI). Advanced geophysics Wireline logs from ECS, AIT, APS, TLD, CMR, and DTS data were
processed by Schlumberger with proprietary software and methods.

RESULTS AND DISCUSSION

Wireline geophysics and electrical resistivity tomography
As shown in Figure 2, the ECS logs collected during the initial month of leaching provided
continuous vertical measurements of copper in the upper most and underlying lifts in the test
section of the heap. Over time, the measurements clearly show a decrease in the copper content of
the ore in the upper most lift as the leaching cycle progresses. On day one, the ECS measured an
average of 0.45 percent copper (dry weight basis) in the fresh lift, which is in agreement with
analysed ore samples collected from drilled boreholes. By day 102, the last logging event, the
average copper in the upper most lift had decreased to 0.13 to 0.14 percent.

Figure 2 Total Copper Measured by ECS

The small changes from one day to another was attributed to a combination of the statistical errors
of the nuclear measurement (low copper contribution to the composite gamma ray spectrum) and
to the copper rich PLS as it percolated down through multiple lifts.
The AIT logs provided highly precise, continuous bulk resistivity measurements at a radius of 90
inches around the boring. The high precision and sensitivity of the AIT measurements detected
small changes in resistivity with a high degree of precision and accuracy. AIT measurements were
also effective to calibrate and ground-truth ERT measurements.
Elemental Log Analysis (ELAN) software was used to perform integrated log analysis and calculate
volumetric moisture and saturation from APS and TLD logs. Figure 3 shows volumetric moisture
and saturation over time for one of the WL logging ports prior to leaching (baseline reading) and
for two logging events during the leaching of the test section.

The logs in Figure 3 show variability in water content (left log) and saturation (right log) through
the leaching cycle and indicate areas of higher water content and saturation above drier zones at
certain depths, attributed primarily to compaction and associated permeability barriers. In-situ WL
measurements of copper also indicated there was copper left behind in these zones (refer to Figure
2).
The configuration of ERT arrays in the test section provided a high resolution 3D resistivity
distribution. Figure 4 provides an overlay of temperature corrected ERT and AIT resistivity from
the test section. As can be seen in Figure 4, agreement between the two measurements is quite good
to a depth of approximately 80 feet, but at greater depths, ERT resistivity is not as accurate or
reliable as AIT bulk resistivity measurements and also lacks the high depth resolution of the AIT.

Figure 3 Volumetric Water Content and Saturation from APS and TLD

Depth (ft)
Figure 4 Overlay of ERT/AIT Resistivity (ohm/m) for Day 5

Distributed temperature profiling

Monitoring the temperature profile inside the test section of the heap provided valuable
information about the leaching process, which included but was not limited to the following:
Tracking the leaching solution front.
Estimate PLS front velocity and hydraulic conductivity.
Identification of permeability/heterogeneity problems.
Estimate exothermic/endothermic chemical reactions activity.
The temperature inside the heap was driven by the temperature of the raffinate. The contribution of
exothermal chemical reactions was small in comparison. Figure 5 provides 3D temperature profiles
with depth and time for one of the monitored borings in the test section.
The PLS front was tracked by monitoring changes in temperature, prior to and during the heap test
section leach period, from which fluid velocity was calculated. Temperature data were also used to
correct ERT and AIT resistivity measurements.

Figure 5 Temperature Profiles from one borehole with Depth and Time

ECS and column testing

The ECS measured total copper around the borehole. The amount of copper removed from the
fresh lift and moved to underlying lifts was calculated for each day of logging. These in-situ ECS
copper measurements and calculations are effectively the same as in column tests, where recovered
copper is computed from the copper content and the flow rate of the PLS at the bottom of the
column.
Four column tests were run. Two columns with belt samples of ore (before agglomeration) and two
with agglomerated ore collected from the pilot area during stacking. The column test results
indicated belt samples of ore had higher ultimate copper recovery than pilot site ore samples.

Figure 6 Copper Recovery Curve Computed with the ECS versus Column Testing

Figure 6 shows the copper recovery curve computed from the average ECS measurement for all
three monitoring ports. The dash curve is an attempt to match fit recovery curve data with an
exponential function. The green curve is for the column tests using agglomerated ore samples
collected form the pilot area.
The curve computed from elemental spectroscopy logs indicates a faster rate of recovery with
earlier break-through and the final recovery (about 70%) compared to recovery from the column
test (especially the test using the pilot site ore).
To confirm the validity of the elemental spectroscopy recovery curves, a portable drill was used to
collect samples to a depth of 6 m (19.7 feet), which were analyzed in the laboratory. The copper
recoveries computed from this analysis (using the assay from the drilled boreholes as a starting
point) are plotted in Figure 6 as purple triangles and agree well with the wireline measurements.

Integration of data sets

Integration of multiple data sets provides distinct benefits and advantages over applications that
provide single measurements, which include the following:
Optimizes data collection and results and provides a more robust interpretation and
understanding of in-situ leaching processes.
DTS measurements not only allow tracking the PLS front and velocity but can also be used
to correct temperature sensitive ERT and AIT resistivity measurements.
Multiple measurements of resistivity, moisture and temperature facilitate the quality
assurance and quality control of monitoring data and interpretations though comparisons
and cross checking of results by mass and energy balance accounting.
Integrated WL log analysis enables quantitative evaluation of key hydrodynamic and
matrix properties including porosity, fluid saturation and electrical conductivity,
permeability, and mineralogy. In addition, these properties can be used to better
understand and infer important conclusions about the ERT and DTS measurements.

CONCLUSION
A test section of the upper most lift and underlying lifts of an engineered heap under leach was
instrumented and monitored with advanced wireline geophysics, fiber optics, and electrical
resistivity applications. These applications successfully and accurately measured copper,
temperature, induction and distributed resistivity, volumetric moisture, and saturation in the upper
most and underlying lifts in the heap. Measurements provided valuable information on the
hydrodynamics of the leaching process in near real-time.
ECS provided continuous vertical measurements of copper in the upper most and underlying lifts
in the test section of the heap and was used to generate a copper recovery curve. The AIT logs
provided highly precise, continuous and very high resolution bulk resistivity measurements. The
high precision and sensitivity of the AIT measurements detected small changes in resistivity with a
high degree of accuracy. ERT arrays in the test section provided high resolution 3D resistivity
distribution. AIT measurements were effective to ground truth ERT measurements, which
improved reliability.

Temperature profiling with fiber optics combined with AIT resistivity made it possible to track the
PLS front and calculate velocity and hydraulic conductivity. Temperature data were also used to
correct ERT and AIT resistivity measurements, which further increased the accuracy and reliability
of measurements for evaluating moisture, saturation, and PLS movement.
Integration of multiple data sets provided distinct benefits and advantages over single
measurements by providing a more robust interpretation and understanding of in-situ leaching
processes.
Problems in the heap with pooling, channeling, longer lag and breakthrough times, increased
solution in inventory, toe saturation and slope stability can be detected in near real-time with the
in-situ measurements evaluated. Properly integrated data sets can provide the heap leach operator
with information in near real-time that can be used to increase production, reduce costs, and
minimize risks.

Improved processes of extracting copper from its ores

William Culver
State University of
New York, usa

This article examines the chemistry of the Hunt and Douglas Copper
Process in the context of 1860s science and engineering. It was a humid
process, developed at Quebecs Harvey Hill Mine. In the archaic language
of the original patent the process involved reactions between
protochloride of iron and the oxyds of copper. The original equation was
in a notation system common before modern theory-based chemistry. The
article includes how the process worked, how the patent holders believed it
worked, its engineering, and the engineering breakthroughs that were still
ahead those required for a business using such a process to earn a profit.
Chemically, the Hunt and Douglas Process succeeded, but not the original
company using it. There is a moment for any technology thats critically
important when it moves from an area of scientific interest to one that
companies are founded on. Patent No 86,754 made that transition in 1870
when the Compaa de Minas de la Invernada was organized in Valparaiso.
This was Chiles first joint-stock mining company, and perhaps the first
copper mining company anywhere formed around a processing technology
rather than a new mineral claim. As a joint-stock company, Invernada
spread the risk of investing in the revolutionary but untested Hunt and
Douglas Process. It was the first commercial hydrometallurgical process for
copper it involved vat leaching and used scrap iron for precipitation. To
help get the new plant operating, James Douglas travelled to Tiltil, Chile, in
1871 as a consultant. Bornite (4-5% Cu5FeS4) was the principal ore, and the
initial plant was to process several tons a day. After Invernada,
hydrometallurgy in Chile became a questioned alternative. In Chile, as
worldwide, decades of blind alleys would pass before hydrotechnology
would finally come into its own.

INTRODUCTION
In popular memory nineteenth century metal-mining is replete with colorful prospectors and
stubborn miners, both painted with anecdotes of sudden wealth. The Devils decoy is how James
Douglas termed the latter accounts of miners who refused to quit and, even though out of money,
borrowed from family and friends to dig just a few feet deeper only to uncover a rich vein
(Douglas, 1884). This brand of the mine discovery myth is skilfully narrated by writers such as
Mark Twain (Roughing It, 1872), Jotabeche (Sketches of Life in Chile, 1841-1851) and Benjamn
Vicua Mackenna (Libro de Plata, 1882; Libro de Cobre, 1883). Or take Alberto Blest Gana (Martn
Rivas, 1862) where an eventual rich silver discovery is a tragic back-story. All four writers leave
their urban readers with positive images of nineteenth century mining life. Whatever may have
been the reality for laborers struggling for better working conditions, or investors unwisely risking
their savings amid dreams of easy riches, overall the public thought well of underground metalmining.
This benign feeling is gone today. In country after country new mine projects are denounced as a
public danger. Anti-mining campaigns inevitably make their case using suggestive language such
as toxic mining and through photographs posted on a Website. The uglier an open pit mine, or
the more multi-coloured an evaporation pond, the better. Almost without exception strong
opposition hinges on perceived threats to scarce water resources, combined with distrust over
possible long-term contamination of soil and ground-water. Moreover, there is a belief that limited
liability operating-companies are created as a tactic to avoid responsibility in case something goes
wrong. This modern reaction is disconnected from the earlier romantic notions of mining.
Whatever else happened over the years to change general public opinion about mining, worry
about chemical contamination has become widespread.
In the twentieth-first century the mining industry gropes for how to respond to these decentralized
assaults attacks that are complicated to track and answer. The situation is even worse in mining
countries where the regulatory state is considered weak, corrupt, and thus ineffective. In countries
where environmental standards are more carefully enforced, opposition is more often technical, but
nonetheless insistent. Overall, rejection of new projects everywhere is expressed in terms of a
Pandoras box of poisonous chemicals. Rational debate over where society is to obtain its metals
over the coming decades is missing. As most of todays recoverable non-ferrous metal ores are only
found in very low-grades, and are nearly all processed via hydrotechnology,1 missing from the
conversation is an appreciation of the role chemical knowledge has played in ore processing
technology all along.
This paper is a reminder that ore processing by hydrotechnology, based on an ever improving
understanding of the elements, is one of the pillars of industrial society. The Hunt and Douglas
Copper Process was the first commercial process crossing the bridge from the chemistry laboratory
to an industrial application, from a patented scientific insight to an operating plant. While there
were other hydrometallurgical efforts started earlier, the Hunt and Douglas patent was the first
successful chemical process intended to produce metal, as opposed to metal as a by-product. When
the history of hydrometallurgy is written, the experiments carried out by James Douglas and T.
Sterry Hunt will be near the start of it all. In the nineteenth century, before the word chemical
Hydrotechnology is used here to refer to copper, silver and gold processing where ore is ground into
particles for placement in a solution. From there the mineral compounds may be floated to the surface as a
concentrate (froth flotation), precipitated as a metal through a chemical reaction or extracted via electrolysis.
The latter two are the domain of hydrometallurgy.
1

became a negative in popular culture, research chemists sought to reduce substances to their core
elements. New elements were discovered year after year. By mid-nineteenth century accumulated
laboratory observations and the demonstrated power of common salt in making copper pyrites
soluble combined with the need at more and more mines for a low-grade metal ore extraction
process. Thus in North American and Europe there was a flurry of metal extraction experiments
based on aqueous chemical solutions (Longmaid, 1914; Anonymous, 1875). Much of the interest
focused on ferrous chloride, observed more than chemically understood. The objective was a
commercial process useful for ores under 5%, in the instance of copper.
The much lower sulfur content of Western Hemisphere copper ores when compared with those of
Spain and Portugal made the oldest hydrometallurgical method, the Longmaid-Henderson pyrite
process, impractical. A process was needed at the mine transporting all but high-grade ores to a
central plant was too expensive. What the copper industry required was someone somewhere to
solve the problem of how to do in the field, at an industrial level, what was already possible in the
laboratory - a process to produce copper metal, not just as by-product. The first such documented
effort is the Whelpley and Storer Copper Process, developed in Boston, but it never reached the
operational stage due to a plant fire at the Harvey Hill Mine, located in the Eastern Townships of
Quebec, where it was undergoing its first test (Culver, 2012). In the mid-1860s the Harvey Hill
owners, the Douglas family, turned in desperation to the Whelpley and Storer Process - they faced
financial ruin. Their higher grades of ore exhausted; they were spending heavily on development
work with the hope of finding further high-grade ore. Local networking through the Literary and
Historical Society of Quebec led James Douglas to consult with T. Sterry Hunt, a geochemist
working for the Canadian Geological Survey. Hunt knew of a revolutionary process just developed
in Boston. He had helped his friend James Whelpley work out the formula for the chemical
reactions Whelpley and Jacob Storer observed. The Boston team stumbled upon unexpected
reactions when they employed a bath of calcium chloride, and into it threw oxidized copper ore
and sulphurous acid from a shaft furnace (Culver, 2012). The experiment was to develop
applications for their line of ore pulverizing machinery and blowers. The process promised a way
to extract copper metal from ore below the 5% threshold - both efficiently and economically. While
working out the engineering of the Whelpley and Storer design, Douglas noticed some additional
chemical reactions. Under Hunts guidance, Douglas pursued a better understanding of his
observations and conducted further experiments. The result was the Hunt and Douglas Copper
Process, U. S. patent number 86,754. While it was developed at Harvey Hill, the first plant built to
try it was in Tiltil, Chile by the Invernada Mine Company (Culver, 2005).

METHODOLOGY
This paper is based on a wide reading of period sources, including thirteen U. S. patents filed
between 1869 and 1900 by Hunt and/or Douglas. The Hunt and Douglas patents in Canada, the
U.K., Bolivia, Mexico and Chile were also examined. There are dozens of articles about the process,
many written by Hunt or Douglas, but also by other metallurgists. The Mining Journal (London),
the Engineering and Mining Journal (New York), the Mining and Scientific Press (San Francisco,
U.S.A.), and the Boletn Minero (Santiago, Chile) were read, as were numerous English language
science and chemical journals from the 1810s through the 1880s. At times the article draws broadly
on the authors earlier work (Culver & Reinhart, 1989). Chemistry textbooks from 1818 until 1900
were another source. The objective in all of this is to better appreciate the chemical context of the
period, and learn how Hunt and Douglas understood their chemical observations. This research is

part of the authors forthcoming book: A Beautiful Process: James Douglass Hydrotechnology, Chile and
Nineteenth Century Copper Capitalism.

RESULTS AND DISCUSSION

The Hunt and Douglas Copper Process began as a simple vat leaching method for chloridizing
copper oxides, taking advantage of chlorines high reactivity and strong oxidizing capacity. T.
Sterry Hunt and James Douglas engineered a way to remove copper from its ore using a watery
solution of ferrous chloride (FeCl2) and sodium chloride (NaCl - common salt). The original 1869
patent described what today we recognize as a spontaneous double-replacement reaction between
cupric oxide (CuO) and ferrous chloride (FeCl2) yielding as a product cupric chloride (CuCl2) and
ferrous oxide (FeCl2). In turn a further chloride replacement reaction between metallic iron and the
cuprous chloride precipitated out the metallic copper (Hunt and Douglas, 1869). Hunt and Douglas,
or rather more Douglas than Hunt, continuously introduced innovations in their procedures
between the years 1869-1900. With each complication they identified, there was a corresponding
variation in the specified steps of the next patent.2
In the 1950s George Van Arsdale characterized the original chemical reactions observed by James
Douglas at Harvey Hill mine as somewhat complicated, but Van Arsdale wrote out a simplified
version of the equation. Van Arsdale described the 1869 process as a neutral leaching method
making use of the solvent action of hot ferrous chloride brine on oxidized or roasted copper ore.
He suggests the following molecular formulas for the original process (Van Arsdale. 1953).
CuO + FeCl2 + H2O CuCl2 + Fe(OH)2
Then he noted that the ferrous hydroxide, Fe(OH)2, readily oxidizes when exposed to air to form
Fe(OH)3. He goes on to suspect that cuprous chloride was formed, probably in a greater amount
than the cupric chloride. The cautious wording suggests he had never actually witnessed the
process in operation. His final balanced equation is below, with added ionic charges and the nature
of the substance:
3Cu2+O2-(s) + 2Fe2+Cl2-(aq) + 3H2O 2CuCl (aq) + Cu2+Cl2-(aq) + 2Fe(OH)3(aq)
Van Arsdale concluded that the cuprous chloride was held in solution by excess of either
ferrous chloride or sodium chloride brine. The copper was then leached by filtration and
precipitation, ...thereby regenerating the solvent, ferrous chloride (Van Arsdale. 1953).
Precipitation was the main extraction method until the last two patents - both emphasized
extraction by electrolysis. This 1953 discussion is perhaps the fullest modern statement on the
process. Douglas was much admired by Van Arsdale, whose career began in Arizona where he
worked under Douglas direction. Van Arsdale investigated the underlying principles of the Rio
Tinto pyrite process so they could be applied at Bisbees Sacramento Hill. He dedicated his career
culminating book to the memory of James Douglas, Pioneer U.S. Hydrometallurgist, and to J.V.N.
Dorr, without whose inventions modern leaching would not be possible (Van Arsdale. 1953).

There are conflicting statements as to the scientific experiments of others that guided Hunt and Douglas but
we can assume Hunt knew the literature in French, German and English.
2

This dedication to Dorr reminds us that the engineering for Hunt and Douglas Process, and all
hydrometallurgy, lagged behind the scientific insights. By the time James Douglas received the final
Hunt and Douglas patent in 1900 the copper industry had relegated the many humid alternatives
to a niche corner in the metallurgists toolbox. After decades of experimentation and field trials
ferrous chloride seemed nothing more than a blind alley Van Arsdale called it an indifferent
solvent for copper (Van Arsdale. 1953). As Douglas and Hunt gained experience and solved
problems they patented each new variation in their procedure. Their chemical descriptions changed
even as the original chloridizing idea remained constant. They began to give attention to the precise
temperature of the brine, but never seemed to vary from their specification for a neutral solution.
Notwithstanding their improvements, the necessary benefits of non-selective mining methods and
large-scale operation lay ahead, as were precision grinding and sorting methods (Van Arsdale,
1953). After the turn of the century, copper hydrometallurgy was overshadowed by the other
hydrotechnology froth flotation. With flotation concentration methods perfected, the urgency for
a hydrometallurgical method faded.

Language and atomic weights

Language is the greatest obstacle to our precise understanding of the Hunt and Douglas Process.
The patentees wrote many articles about their process and necessarily relied on the chemical
nomenclature of the day. However, this was a time of rapid scientific development. Amid the
discovery of new elementary substances and competing classification schemes, each year brought
new understandings of the characteristics of known elements and their compounds. Terminology
evolved and words once commonplace began to lose meaning. With time, older terms faded with
the older generation and they ceased to be used. Today, unless we take the time to study the history
of chemistry, it is if the older terms never existed. Over the last half of the nineteenth-century a
degree of scientific bilingualism was necessary amid false cognates and chemically obscure usages
in the decades before sub-atomic knowledge. Much of the intellectual struggle in Hunt and
Douglas time concerned developing a consensus over the terminology to describe observed
reactions.
Of even more consequence, formulas and equations could not avoid errors by todays standards
until the matter of atomic weights was sorted out. It was only in 1869 that Dimitri Mendeleev
published his ideas for systematically organizing the elements in a table based on atomic weight
(Scerri, 2007). There were at least five other chemists conceiving their own periodic tables and
putting forth their own improved terminology. Scientific debates of the 1860s included
clarification of the notions atom and molecule. In turn this was the key to sorting out the ratios
in which elements combined. Measuring the combining power of elements took the form of
equivalent weights for the early nineteenth-century. Subsequent attention to atomic weights
became the focus. Hunt was already an internationally recognized scientist when their patent was
granted. His was the name that gave their process credibility and visibility. By 1869 Hunt had a
long list of publications on inorganic chemistry and geology perhaps over 200. Hunts position in
the debate over the fundamentals was to question the very notion of a smallest particle that is,
Daltons atomic theory. Only today is Hunts case for continuous matter coming back as a way of
thinking (Tong, 2012).
Prior to 1860 only a minority of chemists avoided an error in calculating atomic weights by a factor
of two, an error traced back to John Daltons original calculations. Most chemists derived atomic
weights from relative combining weights and equivalents (Levere, 2001 & Scerri, 2007). See TABLE

1. Lack of agreement about atomic weights translated into errors in formulas. As we now know,
period chemists had yet to recognize sub-atomic structures. Molecular formulas were easily misled
by diatomic molecules. Dalton thought diatomic molecules could not exist and that view became
orthodoxy this due to the great respect with which his overall insights were held. Yet the
accumulated work on atomic weights often led to accuracy in many formulas. As for Hunt and
Douglas, they merely noted what they observed, and their patents or scientific articles sought to
accurately communicate the results of their latest work as best they could. Their patents contained
only descriptions, with no equations or formulas. Only in publications did they translate their
observations into explicit chemical notations. While Cannizzaro published his corrected weights in
1858, and set others including Mendeleev on a new path, there is no evidence Hunt gave
Cannizzaros article any importance (Cannizzaro, 1858). It was Hunt who following the scientific
journals.
Table 1 Published Atomic Weights with the Hunt & Douglas Process
1858
Fownes

1858
Cannizzaro

1869
Medeleev

1876
Hunt &
Douglas

Modern

Cu

31.71

31.7

63

63.4

31.75

63.55

Fe

27.18

28

----

56

28

55.85

Cl

35.47

35.5

35.5

35.5

35.5

35.45

8.01

16

16

16

Element

1840
Graham

While there was scientific agreement about the atomic weight of chlorine, in the early years of the
Hunt and Douglas Process, Hunt and Douglas did not modernize their other weights and thus
their formulas, as shown in TABLE 2. Still, by the 1860s recognition of two types of copper oxide
was long established copper was at the time one of the most studied elements. Todays cuprous
(copper I Cu2O) and cupric (copper II - CuO) were recognized, but by other names. In 1869 Hunt
and Douglas referred to copper oxides as protoxides (CuO) and suboxides (Cu 2O). And it becomes
even more confusing. The older rule was to use protoxide for the oxide containing the least
oxygen (Cu2O), and to call the next binoxide (CuO), then tritoxide, and so on. But by 1860 this old
rule was breaking down, and protoxide was used for the oxide where the basic characteristics
were most marked this is the usage in the original Hunt and Douglas Process (Thomson, 1818;
Graham, 1840 & Fownes, 1858). Any compound containing what was considered a smaller
proportion of oxygen then became a suboxide. Hunt and Douglas considered protoxide as a
compound of two atoms of copper and two of oxygen whereas today we see it as one atom of
each. See TABLE 2. Great caution is required while reading this periods literature on copper in
order to avoid misunderstandings. Most articles on copper chemistry after 1860 will mention
protoxide for what was once binoxide, and the former protoxide is now degraded to suboxide.
Then by the 1880s, protoxide becomes cupric oxide.
In 1876 Hunt and Douglas still referred to protoxide of copper, but had moved on to using dinoxyd
for copper (I) oxide (Cu2O) rather than suboxide. See TABLE 2. The molecular formula they used
for copper (I) oxide was accurate, but was wrong for copper (II) oxide. Yet it seems they understood
the changing consensus over weights following the 1869 publication of Mendeleevs periodic table.

In their 1876 booklet on their process Hunt and Douglas note, the reactions between protochlorid
of iron and the oxyds of copper are thus expressed as chemical symbols, using the older notation,
in which Cu=31.75, Fe=28, Cl=35.5, and O=8 (emphasis added) (Hunt & Douglas, 1876). In the end
their audience was made up of metallurgists, and they made a decision that it was better to
communicate in terms their non-scientific audience were accustomed reading.
One more point needs to be considered before looking more closely at the process itself. It is
important to emphasize that chemically the process was a success, that is, the reactions they
describe happened and could be replicated most of the time. The problem with their process, and
all humid processes, was commercial not scientific. The one and only justification for moving from
heat-based smelting to a wet or humid process was to lower production costs. The expense of fuel
for processing copper ore under 5% was, at the time, more than the value of the metal. The Hunt
and Douglas Process provided a hope for profitably working low-grade copper, which is the reason
most North American observers considered it revolutionary. In Chile the implications were grasped
more rapidly than anywhere else in the Americas or Europe. In an 1871 letter to Chilean President
Jos Joaqun Prez, Juan Stewart Jackson, President of the Invernada Mining Company, made a case
for the importance of the Hunt and Douglas Process by pointing out the huge quantities of
copper tailings of between 4 to 5% in Chile that until now have been considered without value, and
that there are immense mountains of low-grade minerals that presently left unexploited for lack of
a low-cost method to benefit them (Jackson, 1871). When Douglas and Jackson met face-to-face for
the first time in February 1871, Jackson assured Douglas that if the experiments at the mine bore out
his promises, he, Douglas, would go home with a fortune (Douglas, 1871).
Table 2 Formula and Terminology Comparison Chart
Modern
Molecular
/Ionic
Formulas

Hunt &
Douglas
(1876)
Formula

Cu2O/
Cu2+O2

Cu2O

CuO/
Cu2+O2-

Cu2O2

CuCl/
Cu+Cl-

Cu2Cl

CuCl2/
Cu2+Cl2-

Hunt (1869)
Scientific
Copper
Article

Other Hunt &

Douglas
(1869)
Terminology

Hunt & Douglas

(1876)
Promotional
Booklet

Latin
System
Names

Stock System
Names

oxides of
copper/red

cuprous oxyd

suboxide

dinoxyd of copper

cuprous oxide

copper (I) oxide

oxides of
copper/black

oxidum
cuprosum

protoxide

protoxyd of copper

cupric oxide

copper (II) oxide

cuprous chlorid

protochlorid of
copper

cuprous
chloride

copper (I)
chloride

dichloride of
copper

cupric chlorid

dichloride of copper
protochloride of iron

ferrous
chloride

iron (II)
chloride

hydrated ferrous
oxide

ferrous oxide

iron (II) oxide

ferric oxide

iron (III) oxide

Hunt &
Douglas
(1869) Patent

FeCl2/
Fe2+Cl2-

FeCl

protochloride of
iron

ferrous chlorid

FeO/
Fe2+O2-

FeO

ferrous oxyd

ferrous oxyd

Fe2O3/
Fe22+O32-

F2O3

ferric oxyd

ferric oxyd

FeOCl/
Fe3+O2-Cl-

iron oxychloride

NaCl/
Na+Cl-

chlorid of sodium

iron dichloride

oxychloride of
iron
chlorid of
sodium

common salt

iron (III)
oxychloride
chlorid of sodium

sodium
chloride

sodium chloride

The 1869 Patent: improved processes of extracting copper from its ores
On January 11, 1869 Hunt and Douglas filed their first United States patent application from
Montreal, Canada. By current standards the patent award came rapidly, less than a month later, on
February 9. They termed their humid process an improvement, as it was a variation on existing
laboratory and industrial knowledge. They claimed as their main invention the use of ferrous
chloride - calling it protochloride of iron - as a solvent to convert copper oxides into soluble cupric
chloride. They considered their invention a regenerative process; the double replacement reaction
not only created cupric chloride, but oxychloride of iron. They claimed in their patent the use of
sulphurous acid to decompose the oxychloride of iron and to restore the original ferrous chloride.
Their original patent lacked formulas or an equation for the reactions they described. Rather they
conveyed their process with the chemical nomenclature of the day using terms with roots in the
eighteenth-century Lavoisier system. Today, reading their descriptions and mineral equations it
easy to be adrift as to what exactly they were doing. They clearly observed the same reactions we
see today, but as ionic compounds did not exist for them they had no way to explain with precision
the outcomes. The following selection is from the second paragraph of the 1869 patent refer to
TABLE 2 for assistance in following their language. The bracketed formulas are in the modern
formulas.
For the extraction of copper ores by this process it is necessary that it should be in the
state of protoxide [CuO], or suboxide [Cu2O], or some compound of these oxides The
pulverized and naturally or artificially oxidized ore are then digested with a watery
solution of protosalt of iron, with or without the addition of an earthy or alkaline
chloride. We prefer the neutral protochloride of iron [FeCl2] By the action of a solution
of protochloride of iron [FeCl2] on the oxides of copper [CuO & Cu2O] these are
converted into dichloride of copper [CuCl2], which is readily soluble in concentrated
solutions of earthy or alkaline chlorides such as common salt [NaCl]. At the same time
the iron separates from the solution as an insoluble oxychloride [FeOCl] (Hunt &
Douglas, 1869).
The power of these reactions at an industrial level could not help but excite anyone who anticipated
access to the worlds huge deposits of low-grade ore, which at the time had no value for lack of a
commercial process.

Hunts 1869 Contributions to the Chemistry of Copper3

The experiments behind the patent found an outlet in Hunts 1869 Contributions paper presented
in Salem, Massachusetts at the 1869 meeting of the American Association for the Advancement of
Science. Hunt never mentions the patent as he describes his broader scientific observations relevant
to copper (Hunt, 1869). Interestingly, he provides the same explanation as in the patent, but with
language variations. Now he wrote with scientific precision rather than use popular terms
aimed at the mining public. Again, the bracketed formulas are the modern formulas.
When ferrous chlorid [FeCl2] in solution with chlorid of sodium [NaCl] is heated with a
sufficient quantity of cuprous oxyd [Cu2O], the whole of the iron is precipitated as ferric
oxyd [F2O3], mingled with metallic copper [Cu], while cuprous chlorid [CuCl] remains in
solution. Experiments made with an excess of ferrous chlorid [FeCl2] show that one third
This article was first published in the American Chemist, and then picked up by the Engineering and Mining
Journal with attribution to the American Chemist.
3

of the copper is reduced, while two thirds are dissolved as dichloride [CuCl2]. This
reduction may be effected directly by ferrous oxyd [FeO]; if to a solution of cuprous
chlorid [CuCl] in chlorid of sodium [NaCl], we add hydrated ferrous oxyd
(FeO(OH)H2O) recently precipitated by an alkaline base and still suspended in the
liquid, it is at once converted into ferric oxyd [Fe2O3], with precipitation of metallic
copper.
The above paragraph is intended to illustrate Hunts use of scientific language. The following
are Hunts original equations with his formulas albeit with an error for cuprous chloride which
is CuCl:
First stage reactions: Cu2O + FeCl = Cu2Cl + FeO
Second stage reactions: Cu2Cl + 3FeO = Cu2 + FeCl + Fe2O3
Final Result: 3Cu2O + 2FeCl = 2Cu2Cl + 2Cu + Fe2O3
At this stage of hydrometallurgy Hunt and Douglas experiments were trail blazing. In the
original process, and in Hunts AAAS paper, pains were taken to explain how to make ferrous
chloride there was no supplier to order from. One benefit of the original process was that it
left arsenic undissolved and out of the metallic copper.

Subsequent patented improvements

In May 1895 Douglas delivered four Society of Arts (London) Cantor Lectures on Recent American
Methods and Appliances Employed in the Metallurgy of Copper, Lead, Gold and Silver. His final
lecture in this prestigious annual series covered The Humid Metallurgy of Copper and Metallurgy
of Lead. He starts with his observation that humid methods had not much been practiced in the
United States as there were no deposits of pyrites such as in Spain and Portugal. When he turned to
the Hunt and Douglas Process he writes, In our first method, we took advantage of the reaction
pointed out by Meyer, in 1862, between ferrous chloride and cupric oxide, resulting in the
production of cuprous and cupric chloride. The insolubility of the latter necessitated the addition to
the chloride of iron of a solvent of the cuprous chloride, such as common salt (Douglas, 1895). The
way he spoke of his process revealed that he understood it was a step forward, but that it did not
fulfil the revolutionary impact they hoped for back in the late 1860s.

Table 3: Hunt and Douglas Copper Process Improvements & Patents

Year

Patent
#

Name
Patentee

Title of Patent
(solvent/product)

Extraction
Method

1869

86.754

H&D

Improvement in Processes of Extracting Copper from Its Ores

(protochloride of iron & common salt /cuprous chloride)

iron precipitate

1874

151,763

D, H, S

Improvement in Extracting Silver, Gold, and Other Metals from Their Ores
(protochloride of iron & common salt/chloride of copper )

iron precipitate

1880

227,902

H&D

Extracting Copper from Ores

(chloride & sulphurous acid/dichloride of copper)

iron precipitation
or filter

1887

364,174

H&D

Process of Extracting Copper from Its Ores

(ferrous chloride & common salt/cuprous & cupric chloride)

iron precipitation
or filter

1893

506,985

Methods of Separating and Recovering Copper

(chloride of iron & sulphurous acid/sub-chloride of copper)

iron precipitation
or filter

1896

563,144

Process of Extracting Copper from Ores

(chlorid salt & sulphurous-acid gas/cuprous chloride)

Electrolysis

1900

664,537

Process of Extracting Copper

(chlorid/cuprous chloride)

Electrolysis

Patentees: H = T. Sterry Hunt; D = James Douglas; S = James O. Stewart. (Hunt died in 1892.)

He made clear his view that one reason for its limited applicability is the difficulty of separating
silver from the mixed solution (Douglas, 1895). The trouble with the first process was its retention
of any silver present in the ores. In the 1874 improvement silver is left undissolved in the residue
from the ore, but now the solution was vulnerable to contamination with arsenic. The 1880
improvements sought to isolate other metals from the cuprous chloride. By 1887 the aim was to
better deal with both silver and arsenic. Douglas dedication to experimentation and the search for a
commercial humid process culminated in his 1896 and 1990 patents for extracting copper from a
solution by means of electrolysis. He had worked on such methods since the late 1870s, and finally
found a procedure that satisfied him. He still sought cuprous chloride as a product.

CONCLUSIONS
Douglas leading reputation as a metallurgist was built on three decades of pioneering
hydrometallurgical experimentation. He had no way to know what he did not know. Yet he
obviously knew what was revealed by the experiments he carried out with Hunt. What Douglas
and Hunt lacked was underlying theory a limitation tied to the chemical knowledge of the time.
Douglas put what he and Hunt learned into patents as a business matter. All through his career he
advocated open sharing of technology accompanied by payment of fees when a patented
technology was adopted (Douglas, 1904). Hunt, for his part, used their research to publish scientific
papers. Today, Hunt would be considered a geochemist. Douglas example inspired subsequent
generations of researchers trying to fulfil the dream of a process that was both scientifically perfect
and profitable. A handful of metallurgists kept at it seeking to engineer what still seemed as a
good idea (Greenawalt, 1912; Hofman, 1914; & Van Arsdale, 1953).
Today, Douglas is remembered by the AIME James Douglas Gold Medal established in 1922 to
recognize distinguished achievement in nonferrous metallurgy, including both the beneficiation of
ores and the alloying and utilization of nonferrous metals. Dr. Douglas, founder of Phelps Dodge

Corporation and twice President of AIME, was also an AIME Honorary Member. He was an
industrialist, a mining engineer, a metallurgical engineer and a noted inventor of metallurgical
equipment. The only omission in AIME citation is Douglas parallel career as a writer, and perhaps
a more explicit mention of hydrometallurgy. Finally, in considering the Hunt and Douglas Copper
Process as one of the earliest efforts to solve the low-grade copper ore problem, we must necessarily
take note of the daring visionaries of modern copper mining who built the worlds first commercial
hydrometallurgical plant up the Maritatas Canyon in Tiltil, Chile.

REFERENCES
Anonymous, (1875) Wet Processes, Encyclopaedia Britannica, London.
Cannizzaro, S. (1858) Sketch of a Course of Chemical Philosophy, Edinburgh: The Alembic Club.
Culver, W.W. (2005), Lost Mines and Missing Technology: The Application of the Hunt and Douglas Wet
Method for Smelting Copper - Chile 1871, Proceedings of the Taller Internacional de Hidrometalurgia del
Cobre/HydroCopper 2005, Santiago, Chile, 24-25 November.
Culver, W.W. (2012), Innovations in Copper Hydrometallurgy: The Earliest Years and the Whelpley & Storer
Process, Proceedings of the 4th International Seminar on Process Hydrometallurgy, Jess M. Casas, et al.,
editors, Santiago, Chile: Gecamin, pp. 426-437.
Culver, W.W., & Reinhart, C. J. (1989) Capitalist Dreams: Chiles Response to Nineteenth-Century World
Copper Competition, Comparative Studies in Society and History, 31: 4, pp. 722-744.
Douglas, J. (1871) Letter to Naomi Douglas in Quebec City from the Invernada Mine, February 22, Douglas
Papers, Special Collections, University of Arizona, Call No. AZ 290, Folder 1. Letterbook, 1871, James
Douglas to wife, sister, & others, p. 9.
Douglas, J., Jr. (1884), Chili, Her Mines and Miners II, Engineering and Mining Journal, 38:4, July 26, p. 56.
Douglas, J. (1895) Lecture IV. The Humid Metallurgy of Copper and the Metallurgy of Lead. Wet Methods of
Copper Extraction Cantor Lectures: Recent American Methods and Appliances Employed in the
Metallurgy of Copper, Lead, Gold and Silver, delivered May 12, 1895, Journal of the Society of Arts
(London) XLIII: 2,323, pp. 845-854.
Douglas, J. (1904) Untechnical Addresses on Technical Subjects, New York: John Wiley & Sons.
Fownes, G. (1868), A Manual of Elementary Chemistry, Theoretical and Practical, London.
Graham, T. (1840) Elements of Chemistry, Including the Applications of the Science in the Arts, Part Fourth:
Containing The Metals of the Alkalies, Alkaline Earths, and Earths Proper, with their Salts. London: H.
Bailliere.
Greenawalt, W. E. (1912) The Hydrometallurgy of Copper, New York: McGraw-Hill Book Company.
Hofman, H. O. (1914) Metallurgy of Copper, New York: McGraw-Hill Book Company.
Hunt, T. S. (1870) Contributions to the Chemistry of Copper, Part I, American Journal of Science, XLIX, pp. 153157.
Hunt, T. S., and James Douglas, Jr. (1869) Improvement in Processes of Extracting Copper from Its Ores, U.S.
Patent 86,754 (issued February 9, 1869)
Hunt, T. S., James Douglas, James Oscar Stewart (1876) The Hunt and Douglas Process for Extracting Copper from
Its Ores, Boston: Press of A.A. Kingman.
Jackson, J. S. (1871) Letter from Jackson in Valparaso to Jos Joaqun Prez, President of Chile, July 12, 1871.
(Translation by author.)

Longmaid, J. (1914) The Longmaid Process of Copper Extraction, Engineering and Mining Journal, pp. 177-178.
Mendeleev, D. I. (1871) Chemical Notices from Foreign Sources, Chemical News (London) 23, p. 252. (This was
the first announcement in English of Mendeleevs mature periodic system, originally published in
Russian (1869).
Scerri, E. R. (2007) The Periodic Table: Its Story and Its Significance, New York: Oxford University Press.
Thomson, T. (1818) System of Chemistry in Four Volumes, Philadelphia: Abraham Small.
Tong, D. (2012) Is Quantum Really Analog After All?, Scientific American, 307: 6 December 01, p. 47.
Van Arsdale, G. D. (1953) Hydrometallurgy of Base Metals, New York: McGraw-Hill Book Company. Inc.

chapter 4
Water supply

Design of a treatment system for leachate from

uncontrolled disposal of steel dust on soil

Mara Ruiz,
Alicia Guevara and
Ernesto de la Torre
Escuela Politcnica
Nacional, Ecuador

This paper addresses the design of a treatment system for leachate

generated by the uncontrolled disposal of steel dust on agricultural land.
For the development of the methodology presented in this manuscript,
columns of glass (6 cm diameter), in which 15 cm of steel dust was used to
cover 50 cm of soil, were installed. The columns were irrigated with
constant fluxes of water, equivalent to the average precipitation of the
Ecuadorian Highland region (0.05 mL/h.cm2). To avoid the wall effect,
fiberglass (6 cm diameter, 0.8 mm thickness) was used to cover the solids in
order to allow for uniform distribution of the water throughout the column.
Leachate was collected every three days during four months and
characterized by atomic absorption spectrophotometry. Thus, it was found
that the effluents exceeded the maximum permitted concentration of
chromium and arsenic in Ecuador that can be discharged into freshwater
bodies (0.5 mg/L y 0.1 mg/L respectively). Subsequently, the effluent was
stabilized with cement and lime in liquid/solid ratio of 0.4, 0.5 and 0.6; it
was demonstrated that the liquid/solid ratio of 0.4 and 0.5 allow for
obtaining blocks that contain the recommended amount of moisture to be
confined in landfills (10%). Finally, a 2 m3/month capacity treatment system
of was designed based on the results. In this system, the following process
was proposed: Mixing of the leachate in liquid/solid ratios of 0.4 (equivalent
to 0.5 relative to cement and 2.0 relative to lime), molding and compacting
of the mixture, curing and drying of the blocks every three days and
transportation of the blocks to the landfill. This system requires an initial
investment of US$ 38,782 and has an operating cost of US$ 803.48.

INTRODUCTION
Global and national steel demand rises significantly, especially due to the continued growth of the
emerging economies (Gerdau S.A., 2011). World Steel Association (2012) registered a world production of 1 518 million tons for 2012 despite the economic crisis in Europe that severely affected the
construction sector.
The steel dust generation rate is between 12 and 14 kg per ton of steel produced (Sandoval, 2010), so
the global production of this waste exceeds 17 million tons according to the data reported for 2012
by the World Steel Association (2012). In Ecuador, Acera del Ecuador (2011) generates 1 000 t annually of this waste while in Novacero (2010) more than 3 200 t of steel dust is produced each year.
Steel dust is a toxic residue that has a high content of heavy metals and a tendency to generate
highly alkaline leachates. Several authors define it as a hazardous waste and have expressed concern about the possibility of dissolution and subsequent percolation of its toxic elements from the
upper layers of the soil downwards, contaminating groundwater. Therefore, the disposal of this
hazardous waste is a possible source of contamination of soil and water, considering the relatively
high ease of leaching heavy metals that may percolate. (Seas, Maiza, Marfil, & Valea, 2003)
In Lasso, an Ecuadorian agricultural region, steel dust generated in certain scrap metal recycling
plants is used as filling in ravines. However, this is an improper practice since rainwater or acids
generated by microorganisms can leach heavy metals compounds from soil into the phreatic mantle
of the zone and drainage areas (Sandoval, 2010). Additionally, waste is exposed to open air so that
dust could easily be transported by wind up to the surface of the agricultural soils of the area.
Ecuadorian environmental legislation severely controls practices in the steel industry and punishes
companies that cause environmental damage with very strong sanctions. Constitution of the Republic of Ecuador clearly states in article 396 that each one of the players in the processes of production, distribution, marketing and use of goods or services shall accept direct responsibility for
preventing any environmental impact, for mitigating and repairing the damages caused, and for
maintaining an ongoing environmental monitoring system. (Georgetown University Political
Database of the Americas, 2011). Therefore, the design of a treatment system of the leachates that is
generated by the improper disposal of steel dust on agricultural soil provides a solution to inadequate environmental practices in the steel industry.

METHODOLOGY
For the development of the methodology, soil and steel dust samples were collected and characterized, column tests were employed in order to produce, collect and characterize the leachates generated by steel dust and percolated through soil, and the effluent was stabilized with Portland cement
and lime in different proportions.

Collection and characterization of samples of soil and steel dust

Steel dust and unaffected cultivable soil samples were taken in Lasso (Ecuador) with a simple random procedure. The following physical and chemical parameters were determined: real and apparent density, porosity, humidity, pH, conductivity and particle size. In addition, total chemical composition was carried out by atomic absorption (Perkin Elmer AAnalyst 300 Spectrophotometer). Finally, the Toxicity Characteristic Leaching Procedure (TCLP EPA Method 1311) was used in order
to assess the toxicity of the steel dust.

Column experiments for the production, collection and characterization of leachates

generated by steel dust and percolated through soil
Three columns of glass (6 cm diameter, 75 cm height) were installed. At the bottom of the columns,
5 cm of granite (mesh 4 6) was placed in order to filter the effluent that will be collected into the
collection chamber; the collection chamber consisted of a funnel that connected the lower part of the
column with a 250 mL beaker. In the drainage chamber, 15 cm of steel dust was placed on 50 cm of
soil. The columns were irrigated with constant flows of potable water of 0.05 mL/h per cm, equivalent to mean annual precipitation of Latacunga equals to 455 mm.
The leachates produced by the percolation of water through the layers of dust and soil in the columns were characterized in terms of flow rate, pH, chemical composition (Ag, Al, As, Ba, Ca, Co,
Cr, Cu, Fe, K, Mg, Na, Hg, Ni, Pb, Si, V, Zn) and total organic carbon content. These data were analyzed with the Anova method using the Statgraphics software.

Leachates stabilization with cement and lime

Leachates generated by steel dust and subsequently percolated through the soil were mixed in different amounts (80 mL, 100 mL and 120 mL of water) with 160 g of Portland cement and 40 g of
lime. Stabilized solids were characterized in terms of apparent density, moisture, pH and conductivity. In addition, these solids were subjected to the Toxicity Characteristic Leaching Procedure
(TCLP, EPA Method 1311) and a compression strength test (INEN 1485) in order to verifier if they
meet the requirements for being disposed as landfills.

Design of the system proposed for the leachates treatment

The treatment system and the equipment required for the leachate encapsulation were designed
based on the leachate flow average and the optimum mixture of the effluents to be stabilized with
cement and lime. The capacity of the treatment system was determined by scaling-up the area from
the lysimeters flow section (28.27 cm) to the area of the agricultural soil affected with the steel dust
(86.74 m), and is equal to 2 m/month.

RESULTS AND DISCUSSION

Soil and steel dust characterization
The most relevant physical and chemical properties of the soil and the steel dust tested in this research are shown in Table 1.
Table 1 Initial Physical and Chemical Properties of the Soil and Steel Dust
Property
Apparent density
Real density
Porosity
Moisture
pH
Conductivity
d50 particle size
d80 particle size

Unit
g/mL
g/mL
%
%

S/cm
m
m

Soil
0.8
1.8
52.9
8.1
6.9
235
112.3
524.4

Steel dust
0.9
2.5
62.8
3.6
12.7
344
94.0
145.5

The information presented in Table 1 evidences that steel dust is a dangerous waste, considering
that the diameter of at least 50 % of particles is less than 100 m, so they can cause damage to the
respiratory system. The soil pH classifies it as neutral and soil conductivity as normal. In a similar
manner, it can be seen that the steel dust is too alkaline and slightly saline.
The soil composition determined by atomic absorption spectrophotometry at the beginning (Initial
Sample) and end (Final Sample) of the column tests is shown in Table 2.
Table 2 Initial and Final Elemental Composition of the Soil
Element

Units

Na
Ca
K
Al
Fe
Pb
Zn
Cr
Cu

%
%
%
%
%
mg/kg
mg/kg
mg/kg
mg/kg

Concentration
Initial Sample
43.7
14.2
11.0
6.2
2.7
315.5
313.9
89.6
78.0

Final Sample
40.2
13.5
10.9
6.1
2.7
235.1
384.9
106.9
41.6

Among the elements accumulated in the tested soil, sodium concentration is the highest, followed
by calcium, potassium and aluminum concentrations. On the contrary, lead, zinc, chromium and
copper are trace elements. In addition, higher concentrations of chromium and zinc are found in
soil affected by steel dust as a result of the improper disposal of this hazardous material. Simultaneously, the dilution of the other elements is evident. Hence, steel dust contributes with a significant amount of chromium and zinc to the soil contamination.
In Table 3 elemental composition by atomic absorption spectrophotometry of the steel dust as determined in the original and final sample is shown, from the column tests.
Table 3 Initial and Final Elemental Composition of the Steel Dust
Element

Units

Zn
Fe
Cd
Pb

%
%
%
mg/kg

Concentration
Initial Sample
16.4
9.8
0.8
124.1

Final Sample
15.6
8.7
0.4
106.3

Data organized in Table 3 confirm that the steel dust tested in this research is a hazardous waste
due to the presence of heavy metals such as zinc, lead, and cadmium. Zinc is the most concentrated
specie, for it is a metal with low solubility in molten steel and slag. However, steel dust is rich in
zinc even after the water percolation in the column tests. This is important because the solid waste
could be recycled in order to recover zinc after its improper disposal.
Results obtained for the toxicological characterization of the initial steel dust sample are presented
in Table 4. As can be seen, TCLP Extract Concentration of cadmium and lead exceeds the EPA maximum permitted levels by 11 and 39 times respectively, so steel dust is a toxic waste. Therefore,
tested steel dust is a toxic solid due to its high cadmium and lead content.

Table 4 Leachate Concentrations from Toxicological Characterization of Steel Dust Results

Element
Pb
Cd
As
Ni
Ba
Fe
Cr

TCLP Extract Concentration (mg/L)

195.0
11.0
0.7
0.4
0.2
0.1
0.1

Maximum Permitted Concentration (mg/L)

5.0
1.0
5.0
2.0
100.0
30.0
5.0

* (EPA, 2003)

Chemical characterization of leachate generated by steel dust and percolated through

soil
Figure 1 graphically illustrates the concentration profile of heavy metals that exceeded the maxi-

a)

1,5

Cr

1
0,5
0
0

50
Time (days)

100

Concentration (mg/L)

Concentration (mg/L)

mum concentrations permitted by the Ecuadorian legislation TULAS (Unified Text of the Secondary Environmental Legislation), Book VI, Annex 1, entitled Quality Standard on Environment and
Effluents Discharge: Water Resource, for the effluents discharged into freshwater bodies. These
metals are chromium and arsenic, whose concentrations are greater than 0.5 mg/L and 0.1 mg/L respectively at certain points. (Ministerio del Ambiente de Ecuador, 2003)

0,12
0,1
0,08
0,06
0,04
0,02
0

b)

As

50

100

Time (days)

Figure 1 Concentration of a) Cr and b) As in the Leachates Vs Time

It can be observed in Figure 1 (b) that chromium surpasses the concentrations permitted by the Ecuadorian legislation (TULAS, Book VI, Annex 1) during the first 19 days of percolation, so the
leachate generated by steel dust and percolation through soil cannot be discharged into freshwater
bodies. Arsenic also exceeds the maximum limit established in TULAS, but it surpasses the maximum value of 0.1 mg/L only at one single point at the start of the column tests. It should be noted
that the concentrations of the rest of heavy metals are less than the limits set in the Ecuadorian legislation during the period of four months of the column tests.
Figure 2 illustrates the organic carbon content in the leachates as a function of time. As can be seen,
the content of organic carbon in the leachate remains practically constant. Probably, poor soil aeration impeded the solubilization of the organic fraction in the percolation water.

Total Organic Carbon (%)

10
8
6
4
2
0
0

20

40

60

80

Time (days)

Figure 2 Total Organic Carbon Content in Leachate Vs Time

Figure 3 presents the leachates pH in the time domain and the maximum limit established by Ecuadorian legislation TULAS with respect to the pH of effluents discharged to freshwater bodies
(Ministerio del Ambiente de Ecuador, 2003). It can be seen the leachate exceeded the limit set by
TULAS for pH of effluents discharged into freshwater bodies during the first thirty days, after
which pH did not exceed the limit. In addition, leachate pH does not exceed the value of 11, which
explains the low dissolution rate of elements such as lead that is very soluble in pH ranges of 12 to
13. However, this property shows a downward trend, which implies an acidification of the liquids
and therefore a greater dissolution rate of elements whose solubility increases when pH decreases.

10

pH

9
8
Leachate pH
pH Limit (TULAS)

7
6
0

20

40

60

80

100

Time (days)

Figure 3 Leachate pH Vs Time

Stabilization by encapsulation of leachates

The results of leachate stabilization with cement and lime are presented in Table 5. As can be seen,
solid stabilized with an L/S relationship equal to 0.6 does not comply with the limit set for moisture
content in order to be disposed of in landfills, but there is not any problem with the other parameters. On the contrary, solids obtained with an L/S relationship equal to 0.4 and 0.5 satisfy the limits
for all properties meeting the requirements to be confined in landfills containment cells.

Table 5 Physical and Chemical Properties of Stabilized Solids

Property
Density
Moisture
pH
Conductivity
Compression Strength

Units
g/mL
%

S/cm
kPa

L/S = 0.4*
0.57
5.44
7.65
307
964.91

L/S = 0.5*
0.57
8.15
7.72
294
771.60

L/S = 0.6*
0.62
11.33
8.47
288
584.42

Limit*

<10
4 - 13
<100 000
>50

*L/S = Liquid mass/Solid mass; ** (Ruiz, 2003)

Table 6 reports the toxicological characterization of the stabilized solids. The TCLP extract toxicity
limits are also included, in order to determine if the samples tested are toxic. Based on the data, it
can be inferred that the solids obtained with the stabilization of leachates generated by steel dust
and percolated through soil cannot be considered toxic.
Table 6 Leachate Concentrations from Toxicological Characterization of Solids Stabilized Results
Element
As
Cd
Cr
Fe
Ni
Pb
Se

TCLP Extract Concentration (mg/L)

L/S = 0.4*
L/S = 0.5*
L/S = 0.6*
0.0495
0.0986
0.0062
<0.01
<0.01
<0.01
0.03
0.38
0.48
8.27
1.48
7.77
<0.01
<0.01
<0.01
<0.10
<0.10
<0.10
0.04
0.05
0.02

Maximum Permitted Concentration (mg/L) **

5.0
1.0
5.0
30.0
2.0
5.0
1.0

*L/S = Liquid mass/Solid mass; ** (EPA, 2003)

In conclusion, the best formulation for stabilization of leachates generated by steel dust and percolated through soil corresponding to the L/S ratio equal to 0.4.

System proposed for the leachates treatment

The leachate treatment system capacity is 2 m/month, based on the average discharge flow of the
leachates obtained in the three experimental units (0.09 mL/h) and on the steel dust production rate
in an Ecuadorian scrap metal recycling company (9 t/day). Furthermore, the best formulation for
stabilization of leachates generated by steel dust that percolates through soil corresponds to the L/S
ratio equal to 0.4. Consequently, the required raw materials for the leachate treatment are 4 t of
Portland cement and 1 t of lime.
The proposed process starts with the collection of leachates generated by steel dust that afterwards
percolated through soil, which will be accumulated in a 2.1 m tank. Subsequently, the first day of
each month the collected liquid will be mixed for 4 min with cement and lime in the concrete mixer
DasWell, model JZR350DW (Daswell Maquinaria Co. Ltda., 2012). Afterward the pulp will be led to
the vibro-compaction machine Menegotti, model MBM-3, (Ingenieria VDM Ltda., 2008) where the
molds will be filling with the mixture for 30 s, the auto sweeper will remove excess material on the
molds and the press will compact the mixture. Then stabilized solid blocks fall onto a board located
under the vibro-compaction machine and will dry at ambient temperature during three days. Finally, stabilized blocks will be brought to the landfill.

Figure 4 shows the block diagram of the proposed system for the treatment of the leachate.

Leachate
2,67 L/h
Storage

(Ambient temperature)

Leachate
2 m/month
Cement
4 t/month

Mixing and Stirring

(Ambient temperature)

Lime
1 t/month

Mixture
7 t Total
Molding and
compaction
Fresh block
Drying

Stabilized
solid

(Ambient temperature)

Transportation to
landfill

Stabilized
solid

Confinement in
landfills cells

Figure 4 Block Diagram of the System Proposed For the Leachates Treatment

Collection drains such as presented in Figure 5 are required to collect the leachates. Those consist of
half slotted pipes of high-density polyethylene. The non-slotted half will be placed down and covered with bentonite while the slotted half will be packed with gravel. The floor will be covered with
high-density polyethylene geomembranes and at a 1 % slope. (Ruiz, 2003)

Geomembrane
Gravel
HDPE pipe
Bentonite

Figure 5 Leachate Collection System

Economic evaluation of the proposed system

Table 7 shows the required investment for the construction and implementation of a treatment
plant based in the proposed system. Since samples of steel dust and soil tested in this research came
from Lasso (Ecuador), the considered values were based on the prices of this region.

Table 7 Fixed Investment Estimation

Fixed investment
Land
Constructions
Machinery and equipment
Other assets
Total

Value (US$)
6 300
10 500
12 282
9 700
38 782

Table 8 presents the operating costs that were estimated for the stabilization of the leachates, assuming that the industrial plant responsible for the effluents production has the required personnel
for the development of the process.
Table 8 Estimation of the Operating Costs per Month in Direct Materials
Material
Cement
Lime
Water
Diesel
Electricity

Quantity
4
1
11.8
31.4
0.7091

Unit
t/month
t/month
m3/month
gal/month
kWh/month

Unit value (US$)

152.80
160.00
0.31
0.91
0.07
Total

Monthly cost (US$)

611.20
160.00
3.66
28.57
0.05
803.48

The objective of the system proposed in this research is not the production of certain assets or service, but rather is the remediation of an environmental liability. Therefore, the industrial sites that
implement this process should not expect additional income, but they will prevent the payment of
fines related to violations of environmental laws or regulations. In Ecuador, non-compliance with
environmental standards is punished with a fine of between 20 and 200 unified basic wages
(Ministerio del Ambiente de Ecuador, 2003) depending on the infraction level and duration, as well
as with the restoration of the affected natural resource. In Ecuador, the unified basic wage is equal
to US$ 318.00 (Ministerio de Relaciones Laborales de Ecuador, 2013), so the amount that must be
paid is at least $6 360.00 per audit because of the non-treatment of leachates generated by steel
powder and percolated thought soil.

CONCLUSION
Leachates generated by steel dust that afterwards percolate through soil do not comply with the
limits established by the Ecuadorian legislation TULAS for being discharged into freshwater bodies, because of its chromium concentration (greater than 0.5 mg/L). Stabilization with cement and
lime in a liquid mass/solid mass rate equal to 0.4 is an effective alternative to obtain blocks that can
be disposed in landfills containment cells.
For the monthly treatment of 2 m of leachate, the following process is proposed: 1) collection and
storage of leachates for a month, 2) mixture of leachates with 4 t of cement and 1 t of lime during 4
min, 3) molding and compaction of the mixture for 30 s, 4) drying of the blocks at ambient temperature, and 5) transportation of the stabilized solids to the landfill. The construction and commissioning of the treatment plant requires a capital investment of US$ 38 782 and has an operation cost of

$803.48 per month. This process will allow the industrial site that produces steel dust to avoid penalties due to non-compliance with environmental regulations.

REFERENCES
Acera

del Ecuador C.A. (2011). Tratamiento de humos. Retrieved

http://www.adelca.com/sitio/esp/reciclaje_emisiondegases.php

Alvarez,

E. (2010). Hasta el humo se reutiliza. Retrieved noviembre

http://www.vistazo.com/ea/especiales/imprimir.php?Vistazo.com&id=4376

November

20,
20,

2012,
2012,

from
from

Daswell Maquinaria Co. Ltda. (2012). Introduccin de Mezcladora de concreto con motor diesel. Retrieved October
25, 2013, from http://www.mezcladora-de-concreto.es/product/p12.html
EPA, E. P. (2003). Method 1311: Toxicity Characteristic Leaching Procedure. Retrieved February 13, 2013, from
http://www.epa.gov/wastes/hazard/testmethods/sw846/pdfs/1311.pdf
Georgetown University Political Database of the Americas. (2011). Constitution of the Republic of Ecuador.
Retrieved
November
20,
2012,
from
http://pdba.georgetown.edu/Constitutions/Ecuador/english08.html
Gerdau S.A. (2011). Demanda mundial de acero tocar rcord el 2012. Retrieved November 20, 2012, from
http://www.gerdau.cl/demanda-mundial-de-acero-tocara-record-el-2012/
Ingenieria VDM Ltda. (2008). Mquina para bloques MBM - 3. Retrieved October 23, 2013, from
http://www.vdm.cl/Productos/Bloqueras.html
Ministerio de Relaciones Laborales de Ecuador. (2013). Tablas de incremento para la remuneracin mnima sectorial
y tarifas 2013. Retrieved November 29, 2013, from http://www.relacioneslaborales.gob.ec/tablas-deincremento-para-la-remuneracion-minima-sectorial-y-tarifas/
Ministerio del Ambiente de Ecuador. (2003). Norma de calidad ambiental y de descarga de efluentes Recurso agua.
Retrieved February 19, 2013, from http://www.efficacitas.com/efficacitas_es/assets/Anexo%201.pdf
Ruiz, U. (2003). Recomendaciones tcnicas generales para el diseo y construccin de celdas de confinamiento y rellenos
sanitarios.
Retrieved
November
20,
2012,
from
http://www.semarnat.gob.mx/temas/gestionambiental/Materiales%20y%20Actividades%20Riesgosas/
sitioscontaminados
Sandoval, D. (2010). Remediacin por lixiviacin dinmica de polvos de acera usando disoluciones de cidos orgnicos.
Universidad Simn Bolvar, Sartenejas, Colombia.
Seas, L., Maiza, P., Marfil, S., & Valea, J. (2003). Impacto ambiental producido por morteros cementicios con
polvos de acera. Revista Ciencia e Ingeniera, 24(3), 47-53.
World

Steel Association. (2012). Crude steel production. Retrieved

http://www.worldsteel.org/statistics/crude-steel-production.html

November

20,

2012,

from

Dephosphorization of iron ore

by acid leaching

Christie Dorfling and

Eliana Boshoff
University of Stellenbosch,
South Africa

The availability of high-grade ore at an iron ore mine in the Limpopo

province of South Africa is gradually decreasing. The usability of a portion
of the high-grade ore is limited by the fact that it contains phosphor to a
level of 0.5% on average which is ten times the clients specification.
Physical beneficiation does not allow the removal of the phosphorous
mineral as it is finely ingrained in the hematite. A means thus has to be
found that would allow dephosphorization of the ore. The objective of this
study was to determine the feasibility of acid leaching as a processing route
to decrease the phosphor content of the iron ore and to monitor the
associated iron losses.
Laboratory scale batch leaching experiments were performed to evaluate
the effects of solid-to-liquid ratio (1:1 and 1:4 kg ore per liter acid), acid type
(sulfuric acid, nitric acid, and hydrochloric acid), acid concentration (0.2
mol/L, 0.8 mol/L, and 2 mol/L), and leaching temperature (25 and 60C) on
the rate and extent of phosphor removal. A full factorial experimental
design was used.
The solid-to-liquid ratio had a significant effect on the percentage
phosphor removal, with the percentage dephosphorization increasing from
35% to 51% on average when decreasing the solid-to-liquid ratio from 1:1 to
1:4. On average, the percentage dephosphorization achieved at acid
concentrations of 0.2 mol/L, 0.8 mol/L, and 2 mol/L were 28.6%, 47.5%, and
53.8%, respectively. The maximum percentage dephosphorization (89.2%)
was achieved when using 0.8 mol/L HCl at 60C and a solid-to-liquid ratio
of 1:4. At the same temperature and solid-to-liquid ratio, 76.1% and 82.6%
dephosphorisation was achieved with 0.8 mol/L and 2 mol/L HNO3,
respectively. It was not possible to remove more than 58% of the phosphor
when using H2SO4.

INTRODUCTION
One of the main challenges faced by iron ore producers worldwide is the decrease in the
availability of high-grade ore. A specific iron ore mine situated in the Limpopo province of South
Africa typically produces high-grade hematite ore with an iron content exceeding 62 %. The
phosphor content of the ore needs to be maintained at levels below 0.06 % to be suitable for iron
and steel production purposes and to meet clients specifications. Williams (2008) presented an
overview of the negative effects that elements such as phosphor present in iron ore have on the
steel manufacturing process; one example is the formation of cracks in furnaces refractory
linings. Previously, ore that did not meet these specifications were blended with high quality highgrade ore. However, an increasing portion of the available ore is characterised by a phosphor
content exceeding 0.06 % with levels of 0.5 % on average. As a result, an alternative to blending has
to be found to effectively utilise the ore with the high phosphor content. For this particular ore
body, the phosphorous mineral apatite is finely ingrained in the hematite; physical beneficiation is
thus not a feasible option for the removal of the apatite.
Several studies have investigated acid leaching as a processing route for the dephosphorisation of
iron ore originating from different locations. The general reaction describing the leaching of apatite
is presented in Equation 1:
Ca5(PO4)3(OH, F, Cl) + 10H+ 3H3PO4 + H(OH, F, Cl) + 5Ca2+

(1)

Williams (2008) investigated a range of acids including acetic acid, citric acid, hydrochloric acid,
lactic acid, oxalic acid, propionic acid, and sulfuric acid for the dephosphorisation of iron ore
originating from the Sishen iron ore mine in the Northern Cape province of South Africa. Sulfuric
acid achieved the highest percentage phosphor removal (27.6 %), but also resulted in 1.14 % iron
loss. Jin et al. (2006) reported 91.6 % phosphorous removal from Changde iron ore with less than
0.25 % iron loss when using sulfuric acid as leaching agent. Hydrochloric acid and nitric acid
achieved 85.0 % and 89.4 % dephosphorisation for this particular ore, respectively. Cheng et al.
(1999) treated Western Australian iron ore. Because the predominant iron phase in this ore was
goethite, roasting was used to convert goethite to hematite prior to acid leaching for
dephosphorisation. Only sulfuric acid was considered as leaching agent; on average, 67 % and 80 %
phosphor was removed from the roasted lump ore when using 0.1 mol/L and 0.5 mol/L sulfuric
acid, respectively.
Zhang and Muhammed (1989) considered nitric acid as leaching agent for the dephosphorisation of
a Swedish iron ore with magnetite as the main iron phase. It was predicted that acid consumption
and iron losses would be unacceptably high when using HCl due to the reactivity of this acid with
magnetite. The formation of calcium sulfate precipitate when using sulfuric acid as leaching agent
was furthermore considered to be unacceptable. Nitric acid was hence selected as leaching agent;
this acid leached in excess of 95 % of the phosphor and resulted in less than 0.5 % iron loss.
Chime, Onukwuli, and Ujam (2012) reported on the use of hydrochloric acid for the successful
removal of phosphor (typically in excess of 95 %) from Agbaja iron ore mined in the Kogi state of
Nigeria. Obiorah, Menkiti, and Nnuka (2011) did, however, report similar results when using
sulfuric acid for dephosphorisation of this ore body.
Leaching temperature (Williams, 2008; Xia, Ren, and Goa, 2011), acid type (Williams, 2008), acid
concentration (Cheng et al., 1999; Jin et al., 2006; Williams, 2008; Obiorah, Menkiti, and Nnuka, 2011;
Xia, Ren, and Goa, 2011; Chime, Onukwuli, and Ujam, 2012), particle size distribution of the ore

(Cheng et al., 1999; Obiorah, Menkiti, and Nnuka, 2011; Xia, Ren, and Goa, 2011; Chime, Onukwuli,
and Ujam, 2012), and solid-to-liquid ratio (Obiorah, Menkiti, and Nnuka, 2011; Xia, Ren, and Goa,
2011) have been identified as variables with significant effects on leaching performance.
It can be concluded that sulfuric acid, nitric acid, and hydrochloric acid are most commonly
considered as leaching agents for the dephosphorisation of iron ore. As discussed, the extent of
phosphor removal depends on numerous operating conditions; the mineralogy of the iron ore is,
however, one of the most important factors to consider when evaluating acid leaching for
dephosphorisation. The aim of this project was hence to evaluate the feasibility of using acid
leaching as a dephosphorisation technique for the iron ore mined in the Limpopo province of South
Africa. The effects of the key process variables on the percentage dephosphorisation were evaluated
in order to propose conditions that could potentially be used to decrease the phosphor content of
the ore to acceptable levels by means of acid leaching.

EXPERIMENTAL
The iron ore was obtained from the mines fine-ore plant. Although it is expected that faster
leaching kinetics would be achieved when using smaller ore particles, additional size reduction
units would result in significant additional processing costs; the ore was thus used as received with
all particle sizes equal to or less than 2.25 mm. A rotary splitter was used to divide the ore sample
into smaller samples of appropriate size to be used in the leaching tests. AR grade hydrochloric acid
(32 %), sulfuric acid (98 %), and nitric acid (55 %) were supplied by KIMIX South Africa.
The batch leaching experiments were conducted in a covered 1 L glass beaker placed in a
temperature controlled water bath. An overhead stirrer was used for agitation. The tests for each of
the three acids were conducted according to a full factorial experimental design. The values of the
respective operating variables are summarised in TABLE 1.
Table 1 Summary of experimental conditions
Level

Acid concentration

Solid-to-liquid ratio

Temperature

Low

0.2 mol/L

1 kg ore : 4 L acid

25C

Base

0.8 mol/L

High

2 mol/L

1 kg ore : 1 L acid

60C

The experimental procedure involved diluting the acid to the desired concentration before adding
700 mL of diluted acid to the reaction beaker. The acid was heated and, once the desired
temperature had been reached, the amount of ore required to obtain the desired solid-to-liquid ratio
was added to the leach solution. Samples of the leach solutions were taken at the following time
intervals (in minutes): 0, 5, 10, 15, 25, 35, 45, 60, 90, 120, 150, and 180. The samples were filtered
using 0.45 m nylon syringe filters, diluted, and analysed using inductively coupled plasma atomic
emission spectroscopy (ICP-AES). The percentage leaching of the respective metals were calculated
by mass balance.

RESULTS AND DISCUSSION

Effect of variables on HCl leaching
From Figure 1 it is evident that the percentage dephosphorisation increased as the solid-to-liquid
ratio was decreased, regardless of the acid concentration and leaching temperature. On average,
31.2 % and 54.2 % dephosphorisation were achieved at the high and low solid-to-liquid ratios,
respectively. The corresponding iron losses were 0.18 % and 0.57 %, respectively.
The effect of acid concentration on the leaching was dependent on the leaching temperature. At
25C, the percentage dephosphorisation increased as the acid concentration was increased;
increasing the acid concentration from 0.8 mol/L to 2 mol/L resulted in a larger increase in
dephosphorisation and iron leaching than when increasing the acid concentration from 0.2 mol/L to
0.8 mol/L. However, when leaching is performed at 60C, an optimum acid concentration appears
to exist. At this temperature, the maximum percentage dephosphorisation for both solid-to-liquid
ratios was achieved at an acid concentration of 0.8 mol/L although the undesirable iron dissolution
continued to increase as the acid concentration was increased further. Although Obiorah, Menkiti,
and Nnuka (2011) used sulfuric acid as leaching agent, they also reported the existence of an
optimum acid concentration where maximum dephosphorisation is achieved.
Of the other species present in the ore, potassium and manganese could potentially have a
detrimental effect on the downstream blast furnace operation due to the accumulation of these
species in the furnace. It is hence desirable to decrease the concentration of these species in the ore
during the acid leaching as well. The percentage dissolution achieved with HCl for species other
than Fe and P is summarized in TABLE 2. At the conditions where the highest percentage
dephosphorisation was achieved, 34.68 % Mn and 4.22 % K were also removed. The average
percentage manganese removal achieved with HCl (11.64 %) was higher than with HNO3 (3.16 %,
shown in Table 3) and H2SO4 (5.19 %, shown in Table 4). Potassium removal was low in general, but
higher for HCl (2.28 % on average) than for HNO3 (1.65 %) or H2SO4 (1.60 %).

2,0

80

% Fe dissolution

% P dissolution

100

60
40
20

1,5
1,0
0,5
0,0

0
0,2
0,8
2
Acid concentration (mol/L)
25C, 1kg:1L
25C, 1kg:4L
60C, 1kg:1L
60C, 1kg:4L
(a)

0,2
0,8
2
Acid concentration (mol/L)
25C, 1kg:1L
25C, 1kg:4L
60C, 1kg:1L
60C, 1kg:4L
(b)

Figure 1 Percentage dissolution achieved with hydrochloric acid at the different experimental conditions for a)
phosphor and b) iron

Table 2 Percentage dissolution achieved with hydrochloric acid at the different experimental
conditions for species other than Fe and P

Temperature

Solid-toliquid ratio

Acid
concentration

Al

Mg

Mn

Si

25C

1 kg:1L

0.2 mol/L

0.28

0.70

2.27

1.26

0.13

60C

1 kg:1L

0.2 mol/L

0.31

0.46

2.75

1.11

0.17

25C

1 kg:4L

0.2 mol/L

0.38

1.10

3.61

1.69

0.20

60C

1 kg:4L

0.2 mol/L

1.10

1.80

10.11

4.06

0.60

25C

1 kg:1L

0.8 mol/L

0.45

0.94

3.15

2.86

0.18

60C

1 kg:1L

0.8 mol/L

2.56

2.46

13.56

14.68

0.52

25C

1 kg:4L

0.8 mol/L

0.84

2.04

4.96

5.87

0.34

60C

1 kg:4L

0.8 mol/L

7.19

4.22

40.49

34.68

0.96

25C

1 kg:1L

2.0 mol/L

1.68

2.11

8.72

7.67

0.16

60C

1 kg:1L

2.0 mol/L

2.29

1.92

10.75

10.48

0.30

25C

1 kg:4L

2.0 mol/L

3.59

4.58

18.92

19.49

0.48

60C

1 kg:4L

2.0 mol/L

7.48

5.00

36.64

35.87

0.85

Effect of variables on HNO3 leaching

0,4

80

% Fe dissolution

% P dissolution

100

60
40
20

0,3
0,2
0,1
0,0

0
0,2
0,8
2
Acid concentration (mol/L)
25C, 1kg:1L
25C, 1kg:4L
60C, 1kg:1L
60C, 1kg:4L
(a)

0,2
0,8
2
Acid concentration (mol/L)
25C, 1kg:1L
25C, 1kg:4L
60C, 1kg:1L
60C, 1kg:4L
(b)

Figure 2 Percentage dissolution achieved with nitric acid at the different experimental conditions for a)
phosphor and b) iron

The effects of solid-to-liquid ratio on dephosphorisation and iron loss when using nitric acid as
leaching agent were similar to the effects observed for hydrochloric acid. It was once again
observed that, for leaching performed at 60C, the increase in acid concentration from 0.2 mol/L to
0.8 mol/L resulted in a more significant increase in dephosphorisation than the increase from
0.8 mol/L to 2 mol/L; iron dissolution does, however, continued to increase as the acid
concentration was increased. Similar results were reported by Zhang and Muhammed (1989), who
achieved less dephosphorisation but higher iron losses as the nitric acid concentration was
increased above 4 mol/L during the leaching of a Swedish iron ore.
Table 3 Percentage dissolution achieved with nitric acid at the different experimental conditions
for species other than Fe and P

Temperature

Solid-toliquid ratio

Acid
concentration

Al

Mg

Mn

Si

25C

1 kg:1L

0.2 mol/L

0.66

0.98

3.63

1.05

0.33

60C

1 kg:1L

0.2 mol/L

1.15

1.34

6.01

1.34

0.57

25C

1 kg:4L

0.2 mol/L

0.64

2.10

4.41

1.73

0.29

60C

1 kg:4L

0.2 mol/L

2.13

2.37

12.21

3.09

0.94

25C

1 kg:1L

0.8 mol/L

0.41

0.79

3.70

1.05

0.16

60C

1 kg:1L

0.8 mol/L

2.18

1.49

10.08

2.91

0.55

25C

1 kg:4L

0.8 mol/L

0.94

1.75

7.63

3.00

0.38

60C

1 kg:4L

0.8 mol/L

4.69

2.81

26.32

6.97

1.07

25C

1 kg:1L

2.0 mol/L

0.67

0.89

3.93

1.40

0.25

60C

1 kg:1L

2.0 mol/L

2.35

1.38

10.20

3.60

0.35

25C

1 kg:4L

2.0 mol/L

1.25

1.18

6.95

2.72

0.42

60C

1 kg:4L

2.0 mol/L

5.93

2.79

24.78

9.00

0.99

Effect of variables on H2SO4 leaching

The effects of different variables on the sulfuric acid leaching process were less profound, as shown
in Figure 3. A regression analysis considering the percentage phosphor dissolution as the response
variable indicated that the temperature and the solid-to-liquid ratio did not have statistically
significant effects on the percentage dephosphorisation. Similar results were obtained when
considering the percentage iron as the response variable.
Similar to the effects observed for hydrochloric acid leaching and nitric acid leaching, the effect of
the acid concentration was temperature dependent. At 25C, the percentage dephosphorisation
increased as the acid concentration increased; at 60C, however, the maximum percentage
dephosphorisation was achieved at an acid concentration of 0.8 mol/L. It is not clear why the iron
concentration decreased when the acid concentration was increased from 0.8 mol/L to 2 mol/L at
60C.

0,6

60

0,5

% Fe dissolution

% P dissolution

70
50
40
30
20
10
0

0,4
0,3
0,2
0,1
0,0

0,2
0,8
2
Acid concentration (mol/L)

0,2
0,8
2
Acid concentration (mol/L)

25C, 1kg:1L
60C, 1kg:1L

25C, 1kg:1L
60C, 1kg:1L

25C, 1kg:4L
60C, 1kg:4L
(a)

25C, 1kg:4L
60C, 1kg:4L
(b)

Figure 3 Percentage dissolution achieved with sulfuric acid at the different experimental conditions for a)
phosphor and b) iron

Table 4 Percentage dissolution achieved with sulfuric acid at the different experimental conditions
for species other than Fe and P

Temperature

Solid-toliquid ratio

Acid
concentration

Al

Mg

Mn

Si

25C

1 kg:1L

0.2 mol/L

0.50

0.73

2.98

1.23

0.22

60C

1 kg:1L

0.2 mol/L

1.25

1.18

6.50

2.02

0.51

25C

1 kg:4L

0.2 mol/L

0.64

1.16

3.81

1.80

0.32

60C

1 kg:4L

0.2 mol/L

2.22

2.12

13.19

3.96

1.05

25C

1 kg:1L

0.8 mol/L

1.06

1.41

5.18

2.53

0.25

60C

1 kg:1L

0.8 mol/L

2.54

1.40

12.03

5.44

0.64

25C

1 kg:4L

0.8 mol/L

1.08

1.72

5.76

3.33

0.47

60C

1 kg:4L

0.8 mol/L

3.75

2.23

23.63

8.60

0.81

25C

1 kg:1L

2.0 mol/L

5.13

2.40

21.07

13.62

0.05

60C

1 kg:1L

2.0 mol/L

1.05

1.08

6.05

3.09

0.18

25C

1 kg:4L

2.0 mol/L

4.17

2.68

23.14

13.44

0.26

60C

1 kg:4L

2.0 mol/L

1.04

1.10

6.64

3.20

0.34

Comparison of leaching agents

Summaries of the average percentage dephosphorisation and iron dissolution with different
leaching agents and at different operating conditions are summarised in TABLE 5 and TABLE 6,
respectively. These average values further support the observations regarding the effects of process
variables on the leaching behaviour for the different acids. What is furthermore evident from the
data presented in Table 5 is the much larger dependence on acid concentration in the low
concentration range when using HCl compared to HNO3 and H2SO4.
One of the main factors to consider when investigating the use of acid leaching for
dephosphorisation of iron ore is the associated iron loss. The iron loss exceeded 0.06 % only in one
of the 12 tests utilising an acid concentration of 0.2 mol/L, namely when using H2SO4 at 60C and a
solid-to-liquid ratio of 1 kg ore per 4 L acid. On average, the iron loss observed for H2SO4 (0.05 %)
were slightly larger than for HNO3 (0.03 %) and HCl (0.02 %). At higher acid concentrations,
leaching with hydrochloric acid resulted in more than double the iron loss observed for sulfuric
acid and more than four times the iron loss observed for nitric acid. This observation is in
agreement with the statement of Zhang and Muhammed (1989) that hydrochloric acid has the
highest reactivity towards iron oxides and would hence result in the larger iron losses compared to
sulfuric acid and nitric acid.
Table 5 Percentage dephosphorisation averaged for different operating conditions
Acid

Solid-to-liquid
(kg ore : L acid)
1:1
1:4

Temperature
(C)
25
60

Concentration
(mol/L)
0.2
0.8

H2SO4

44.6

42.5

41.1

46.1

36.0

46.4

48.4

HCl

31.2

54.2

35.0

50.4

17.9

50.9

59.4

HNO3

30.0

57.4

37.2

50.2

32.0

45.3

53.7

Table 6 Percentage iron dissolution averaged for different operating conditions

Acid

Solid-to-liquid
(kg ore : L acid)
1:1
1:4

Temperature
(C)
25
60

Concentration
(Mol/L)
0.2
0.8

H2SO4

0.16

0.18

0.21

0.14

0.05

0.16

0.31

HCl

0.18

0.57

0.19

0.56

0.02

0.38

0.72

HNO3

0.05

0.13

0.03

0.14

0.03

0.09

0.15

The maximum percentage dephosphorisation achieved was 89.2 % when using 0.8 mol/L HCl at a
temperature of 60C and a solid-to-liquid ratio of 1 kg ore per 4 L acid. This extent of
dephosphorisation provides a leached iron ore with a phosphor content that meets clients
specifications, and would allow the leached ore to be sold directly without blending with highgrade ore. However, the associated iron dissolution was high with 0.98 % of iron dissolving.
Leaching with nitric acid at the abovementioned conditions also resulted in comparatively good
results, with 76.1 % dephosphorisation and 0.21 % iron loss. This would allow leached ore and

high-grade ore to be blended in a ratio of 1 ton leached ore to 2 tons high-grade ore. The only set of
conditions at which nitric acid leaching achieved a higher percentage dephosphorisation (82.6 %)
entailed leaching with 2 mol/L acid at 60C and a solid-to-liquid ratio of 1 kg ore per 4 L acid. The
associated iron loss was 0.36 %. The corresponding blending ratio that would still ensure that
clients specifications in terms of phosphor content are met is 2 tons leached ore per 1 ton highgrade ore.
The highest percentage dephosphorisation achieved with leaching processes performed at 25C was
78.7 % with 0.69 % associated iron losses when using 2 mol/L hydrochloric acid at a solid-to-liquid
ratio of 1 kg ore per 4 L acid. Sulfuric acid was not successfully used as a leaching agent for
dephosphorisation of iron ore, achieving maximum dephosphorisation of only 58.0 %.
It is clear that there are several combinations of operating conditions that might result in effective
dephosphorisation of the particular iron ore to the extent that the leached ore may be sold directly
or by blending it with high-grade ore in a higher ratio than currently possible. The final decision
regarding the optimum operating conditions can only be made once an economic analysis has been
done for an integrated process flow sheet taking into account factors such as water consumption
and reuse, effluent generation and treatment, acid consumption, and the availability of ore with the
respective grades, amongst other factors. Zhang and Muhammed (1989) did propose a conceptual
flow sheet for the dephosphorisation of iron ore by acid leaching. The treatment of the leach
solution (typically by solvent extraction) for the recovery of phosphoric acid as by-product and the
regeneration of leaching agents were identified as key factors determining the overall economic
viability of these types of processes as well as the extent to which process streams can be recycled
and effluent streams minimised. These aspects of an integrated process flow sheet do, however,
necessitate further detailed investigations.

CONCLUSIONS
The objective of this study was to evaluate the feasibility of using acid leaching as a
dephosphorisation technique for iron ore mined in the Limpopo province of South Africa. It has
been proven that dephosphorisation in excess of 75 % is readily achievable using either HCl or
HNO3 as leaching agent. In general, it has been observed that the percentage dephosphorisation
increased as the solid-to-liquid ratio decreased. When leaching with HCl or HNO3 at a temperature
of 60C, an optimum acid concentration appeared to exist; increasing the acid concentration above
this value resulted in significantly higher iron losses without noticeable improvement in the
dephosphorisation.
If the phosphorous content of the ore can be reduced by 75 %, it can be blended with high-grade ore
in a significantly higher ratio than currently possible while delivering iron ore with a grade that
meets the clients specifications. Although several combinations of operating conditions could allow
the effective utilisation of the phosphor containing ore, economic analyses are required to be able to
determine the most optimum process solution.

REFERENCES
Cheng, C.Y., Misra, V.N., Clough, J., and Mun, R., (1999), Dephosphorisation of Western Australian iron ore by
hydrometallurgical process, Minerals Engineering, 12(9):1083-1092.

Chime, T.O., Onukwuli, O.D., and Ujam, A.J., (2012), Hydrometallurgical process of dephosphorisation of
Agbaja iron ore, Journal of Asian Scientific Research, 2(12):840-855.
Jin, Y., Jiang, T., Yang, Y., Li, Q., Li, G., and Guo, Y., (2006), Removal of phosphorus from iron ores by chemical
leaching, Journal of Central South University of Technology, 13(6):673-677.
Obiorah, S.M.O., Menkiti, M.C., and Nnuka, E.E., (2011), Beneficiation processing of Agbaja iron ore by
chemical leaching technique, New York Science Journal, 4(5):22-27.
Williams, P.J., (2008), The use of Aspergillus Niger for the removal of potassium and phosphorous from the iron
ore of the Sishen iron ore mine, South Africa, PhD Dissertation, University of Pretoria, South Africa.
Xia, W., Ren, Z., and Gao, Y., (2011), Removal of phosphorus from high phosphorus iron ores by selective HCl
leaching method, Journal of Iron and Steel Research, 18(5):01-04.
Zhang, Y., and Muhammed, M., (1989), The removal of phosphorus from iron ore by leaching with nitric acid,
Hydrometallurgy, 21:255-275.

Leaching of coarse caliche ore

Javier Ordez,
Silvia Valdez and
Luis Cisternas
Universidad de
Antofagasta, Chile
Luis Moreno
Royal Institute of
Technology, Sweden

Heap leaching of caliche is carried out using particles of different sizes,

from fine material up to particles with several decimeters. Lab-scale
experiments are usually performed using small particles, but due to the fact
that particle size is a critical variable in heap leaching, experiments with
larger particles were carried out in a transparent column with a square base.
The empty space was filled with small glass-stones and the column was
irrigated initially with an average rate of 0.004m3/m2/h, which was reduced
to half after a while. Outlet concentrations for several ions were determined.
The maximum values were lower than those observed for leaching
experiments using small particles. However, the concentrations were
maintained high during longer times.
For determining if liquid flow takes place through the caliche or only on
the particle surface or as a combination of both, prior to loading the caliche
in the column, some stones were cut to obtain a plane face that was put
against one of the column walls to observe the liquid flow in the stone. The
upper border was sealed to avoid liquid flow between the stone and the
wall. The observations showed that the liquid is introduced into the stone
by capillary forces, but no flow is observed through the caliche stone.
Additional experiments were performed to physically characterize the
caliche, which showed that capillary forces are large, but that the
permeability is very small. This implies that flow through caliche stones
would be very small.

INTRODUCTION
Caliche is a mineral composed of a high proportion of water-soluble species that is naturally
present in northern Chile, in the regions of Tarapac and Antofagasta. The most important products
obtained from this mineral are nitrates and iodine. Natural nitrates (also named Saltpeter) are used
for the manufacture of fertilizers and as industrial salts, while iodine is employed in medical and
technological applications (Glvez et al., 2012; Valencia et al., 2008; Wisniak and Garcs, 2001). Chile
is the world's largest producer of iodine and natural nitrates (USGS, 2013).
One of the stages in the production of iodine and nitrates from caliche is the leaching, which takes
place in vats or heaps depending on the grade and granulometry of ore. The heaps are frequently
constructed by accumulation of 600 to 900 thousand tons of ROM caliche, which means that a wide
spectrum of particle sizes composes each heap. The heap leaching for caliche operations has gained
great importance since 80s due to the decreasing of ore grades and the increment of processing
costs.
Heaps of caliche are irrigated with water or recirculated solutions at the top and the liquid obtained
from the bottom is sent to the iodine extraction plant, where the iodate is transformed to metallic
iodine. The remaining solution is subsequently conducted to evaporation ponds in order to obtain
sodium nitrate and other salts (SQM, 2013; Wheeler, 2012).
The leaching of soluble species of caliche is based on solubility differences, whereas for metallic
minerals (such as copper ores) the solubilization is governed by chemical reactions. Other
important difference is that caliche has a high fraction of soluble minerals formed by different
species that are dissolved simultaneously at different rates, which makes the modeling of caliche
heap leaching a complex task (Glvez et al., 2012). The high fraction of soluble species causes a
gradual decreasing in the size of the particles and, therefore, a compaction of the entire heap.
The size of the particles is an important factor affecting the performance of heap leaching, since the
magnitude of the radius has direct relation with the resistances to material transport within the
solid matrix. In previous works (Ordez et al, 2013; Ordez et al., 2014), column leaching
experiments at pilot scale using fine material were performed. In the present study, coarser
particles were leached, in order to get empirical information closer to industrial conditions, where
the particles can reach up to several decimeters.
On other hand, one of the causes for poor performance of heap leaching is the inadequate
understanding of the mineralogy and the physical properties related with it. Therefore, experiments
to physically characterize the material were done, such as permeability and capillary suction tests.

METHODOLOGY
Column leaching experiments
The particles of caliche used in the column experiment and in the physical characterization tests
were obtained from Northern Chile and the soluble fraction of the mineral was chemically analyzed
(TABLE 1). Nitrate was determined by UV molecular absorption spectroscopy. Sulfate was
quantified by gravimetry, while sodium, potassium, magnesium and calcium were measured by
atomic absorption spectroscopy. Acid-Base volumetry was used for boron determination. Chloride

and iodine were determined by oxide-reduction volumetry. For perchlorate analysis, a specificelectrode was used.
Table 1 Composition of soluble species in caliche. Concentration in wt%
NO3-

IO3-

Cl-

SO42-

ClO4-

BO33-

K+

Ca2+

Mg2+

Na+

4.65

0.01

8.45

9.50

0.03

0.09

1.18

0.82

0.62

8.56

According to mineralogical XRD analysis (Siemens D5000 Diffractometer), the caliche ore used in
the experiments is rich in soluble species. The most abundant water-soluble species is halite that is
the unique source of chloride, meanwhile nitrate is shared between Humberstonite and Nitratine,
both very soluble minerals. Sulfate is widely spread in caliche and it is associated with sodium,
potassium, magnesium and calcium. The insoluble species are present as silicates such as Quartz
and Albite.
A square base column of acrylic was used for the leaching experiment, with sides of 28 cm and a
height of 80 cm. Particles with an average size of 14 cm were placed in the column, some of them
were flattened on one side and attached to the walls of the column, using silicone on the upper part
of the perimeter to avoid liquid flow between the caliche stone and the wall. The remaining spaces
were filled with inert glass particles (Figure 1).
The caliche was leached using demineralized water, with an initial irrigation rate of 4 L/h/m2, which
was reduced to the half after a while. The leaching was performed with an intermittent irrigation
regime of 8 hours a day for a total period of 71 days. During this time, samples of liquid were taken
every day and analyze them chemically. Once finalized the leaching, the column was dismantled
and the remaining rocks were removed and weighed. Samples of residues were also analyzed. The
recovery of each ion was calculated by a material balance, taking into account the composition of
liquid samples and the residues in relation to the initial content of caliche (relative recovery).

Figure 1 Column leaching experiment with large particles

Physical characterization of caliche ores

Four physical properties of the mineral were determined: a) capillary suction, b) apparent density,
c) open porosity and d) permeability of caliche. These characteristics give useful information to
understand the heap leaching phenomena associated to the mineral.
The capillary suction was measured in terms of the capillary coefficient, k c , which is determined
calculating the mass variation per cross sectional area and per the square root of the elapsed time
during solution uptake. For determining the capillary coefficient, cube-shaped caliche rocks were
put in contact with kerosene, maintaining an immersion level of 3 mm at room temperature during
5 hours. Kerosene was chosen because caliche does not dissolve in it. Measurements of mass
variations along time were done, at short time intervals at the beginning and longer ones later. In
each time, the sample was rapidly removed from the support, weighed, and returned to the
container. Since the test is based on kerosene that has different characteristics than those of the
water, the obtained values were transformed in terms of water properties by using the ratios
between the densities and the capillarities.
Regarding to the determination of apparent mineral density,

, caliche rocks with different sizes

and previously dried and weighed, were submerged in a glass beaker filled with kerosene for a
very short time to avoid the penetration of kerosene into the caliche pores. The displaced volume of
kerosene was quantified and the caliche density was calculated by dividing the mass of the sample
by the displaced volume, Vs . This is the apparent density since the pores (open and closed) are
included in the apparent volume.
To estimate the open porosity of the same caliche rocks, , dried samples were put in a desiccator
that contains kerosene at the bottom. The gas in the inner pores was removed using a vacuum
pump for 2 hours. After that, the caliche was submerged in kerosene, the vacuum pump was

stopped and the valve was open to recover the atmospheric pressure onto the submerged sample,
determining the saturated mass.
The pore volume was determined using the mass of the sample in the saturation, M sat , the initial
mass, M s , the volume of sample, Vs and the density of kerosene,

k .

Open porosity was

calculated using the Equation 1.

M sat M s
Vs k

(1)

Permeability, K , is the property of a solid that lets fluids to flow through it. In the essays, cubic
samples with cross sectional areas of 50 cm2 and predrilled in the middle with a hole of 1 cm of
diameter were used. The top and the bottom side of samples were sealed with silicone and bottom
was adhered to a smooth surface. The top of rocks was covered with an acrylic plate with a hole
connected with the perforation of the rock. Once samples were sealed they were saturated in
vacuum with kerosene for 24 hours and subsequently, in the hole, a hose filled with kerosene was
connected (Figure 2). The total height of the experiment from the solid sample to the initial level of
kerosene in the hose was 3.5 m.

Figure 2 Layout of permeability test. H=3.5 m

The variation of height level of kerosene in the hose was measured and used to calculate the
permeability according to the Equation 2.

r
ln e Q
r
K i
k g h 2 Z

(2)

Where re is the external radius of the sample, ri corresponds to the hole radius, Q is the flow,
is the dynamic viscosity of kerosene,

g is the gravity, k is the kerosene density and Z represents

the height of the sample. Due to equation 2 is formulated for cylindrical geometries and the used
samples have cubic shapes, a weighted radius defined as 90% of average radius was considered for
re . This approach gave the best agreement for the different ratios re / ri tested.

RESULTS AND DISCUSSION

Column leaching experiments

1.0

1.0

0.8

0.8

0.8

NO3

0.6

Cl
Ca

0.4
0.2

Relative recovery

1.0

Relative recovery

Relative recovery

The recovery of all ions was determined by mass balance as previously mentioned. Nitrate,
perchlorate and iodine reached the highest final recoveries, about 99%. On other hand, the species
formed by calcium (Glauberite and Polyhalite) and boron were the least leached, since just about
50% of these ions were depleted from the column. The results confirm the exponential recovery
behaviors of the different ions (Figure 3).

0.6
0.4

ClO4
IO3

0.2

0.6
0.4
Na

0.2

Mg

SO4

0.0

0.0
0

250
Time, h

500

0.0
0

250
Time, h

500

250

500

Time, h

Figure 3 Recovery curves for analyzed ions of large particle column leaching

In order to analyze the leaching performance, the elapsed time used for the experiment was
adjusted due to the irrigation intermittences. For this reason, an effective time was calculated that
considers the time that leachant was added to the column. Significant increments of recovery were
achieved up to 200 h, after that the extraction changed to a slower kinetic. According to the leachant
consumption, it was noticed that the volume of leachant used for experiment was sufficient because
the recoveries of ions were relatively high at the final for highly water-soluble species and even for
less soluble ions such as sulfate that was close to 80%. The use of glass stones favors the channeling
in the column and a significant fraction of leachant pass directly through the column without be
used for leaching. Therefore, similar recovery levels could be achieved with lower water
requirements.
The effect of the particle size on recovery was evaluated, by comparing the results of coarse
leaching (CL experiment) with another leaching experiment previously performed (Ordez et al.,
2013), which used fine particles (<1.4 cm of diameter) and tap water as dissolution agent (FL

1.0

1.0

0.8

0.8

Relative recovery

Relative recovery

experiment). As shown in Figure 4, the final recoveries of all ions were similar between both
experiments; however the increment was smoother for CL experiment. The particle size is
determinant in this result, due to larger particles present higher resistance to inner transport that
slows the depleting of ions from caliche. The leachant preferably flowed by the periphery of rocks;
however at the end, all particles were completely wetted and with different levels of water
saturation. The ratio of leachant volume over mass loaded that for FL experiment was 2.4 L/kg,
while for CL was equal to 3.1 L/kg. The last one consumed more solution and for this, same level of
recoveries would be achieved with less solution.

0.6
0.4
NO3_FL

0.2

NO3_CL

0.0
0

250

0.6
0.4
IO3_FL

0.2

IO3_CL

0.0
500

Time, h

250

500

Time, h
b)

1.0

1.0

0.8

0.8

Relative recovery

Relative recovery

a)

0.6
0.4
Cl_FL

0.2

Cl_CL

0.0
0

250

0.4
SO4_FL

0.2

SO4_CL

0.0
500

Time, h
c)

0.6

250

500

Time, h
d)

Figure 4 Comparison of recoveries between column leaching using small and large particle. a) Nitrate, b)
iodine, c) sulfate and d) chloride. FL: Fine particle leaching and CL: Coarse particle leaching

Physical characterization
The porosity of rocks and their densities were determined in the same experiments and the results
are summarized in Table 2. Regarding the open porosity determinations, it was observed that the
open porosity of the caliche range between 2.4% and 16.3%, with an average value of 9.5% for the

14 measured samples. When the porosity was determined by simple immersion without treatment
with vacuum, the porosity was significantly less.
In the density measures, no big differences were observed between the rocks. The average density
is 2 300 kg/m3, ranging from 2 102 to 2 468 kg/m3. These values are notably greater that those
reported in the literature for caliche ores (Wheeler, 2012). The density reported was 1 900 kg/m3,
and corresponds to measurements in situ including intra-particle spaces and fractures, which
diminish the density value. A correlation between density and porosity was not observed for the
performed tests, since the used specimens of caliche have high heterogeneity in their compositions.
Table 2 Density and porosity determinations. They correspond to the statistics for 14 samples
Entity

Density, kg/m3

Porosity, %
(Vacuum + immersion)

Porosity, %
(Only immersion),

Lowest value

2 102

2.4

2.2

Highest value

2 468

16.3

12.4

Average value

2 309

9.5

6.9

131

5.5

4.0

SD

The implication of these porosity values in the leaching performance is mainly related with the flow
of the leaching solution that may takes place around the particle or through the caliche. Apparently
the solution is transported on the particle surface and the inner transport through the particle is not
the most important mechanism by which the solution flows along the heap. Complementary
information to study the importance of inner transport is obtained from the subsequent
experiments of capillary suction and permeability.
From laboratory observations that were confirmed in the column experiment, when the lower part
of the particles were dripped in kerosene, two different regions are observed within the rocks: a
saturated layer of material in the periphery and material slightly humid in the core of the rocks. The
capillarity of caliche was high and could be dependent of other characteristics of rocks such as
chemical and mineral composition in addition to porosity and pore size. In the capillarity tests, the
velocity of kerosene absorption was initially faster and gradually decayed until to reach a
stationary state, which was different for each sample (Figure 5).

Specific solution mass, kg /m2

14
C9

12

C7

10

C4

8
6
4
2
0
0.0

0.4
time,

0.8

1.2

h0.5

Figure 5 Kerosene absorption by caliche during capillarity tests for 3 different samples (C4, C7 and C9)

In order to compare the different values of capillarity obtained in the experiments, the capillary
absorption coefficient was determined by each piece (TABLE 3). The values vary between 0.5 and
8.7 kg/m2/h0.5; these large differences are attributed to the heterogeneities in their compositions.
Table 3 Capillary absorption coefficients of caliche specimens
C1

C2

C3

C4

C5

C6

C7

C8

C9

Mean

SD

k c , kg/m2/h0.5 0.5

1.1

1.5

2.7

3.6

3.7

4.5

6.8

8.7

3.6

2.7

Sample

Regarding permeability coefficients of the caliche samples, these were small (TABLE 4), about
5*10-17 m2. This makes sense to observe, in general, the low porosities of the mineral. Note that low
permeability values are obtained not only in formations with not connected pores but also with
small pores.
Table 4 Permeability coefficient ( K ) determined for caliche samples
Sample

P1

P2

P3

Mean

SD

K , m2

7.4*10-18

5.7*10-17

8.1*10-17

4.9*10-17

3.8*10-17

CONCLUSIONS
A column leaching experiment was performed, using large particles of caliche. Although the
particle size has a significant effect on the recovery behavior of the leaching, for caliche its
importance appears to be no critical since high recoveries may be reached if enough leachant is
added.

Caliche is a mineral that has a wide variety of constituents, for this reason the determination of
characteristic properties results complex. Range of values for apparent density, total porosity,
capillary and permeability coefficients were proposed in this work, which enriches the knowledge
about physical parameters of caliche that are related with the leaching performance.
Density values with low variation were obtained, around 2 300 kg/m3, which is consistent with
other reference calculated for bed of caliche. One of the most important parameters that affect other
properties is the porosity, which resulted with a wide variation. For its estimation, two saturation
methods were proved, being the vacuum saturation the best one. Porosities between 3-16% were
reached. Capillary coefficients were determined for caliche and between 0.5-9 kg/m2/h0.5. The values
suggest that this material can retain high fractions of liquid during leaching, but the inner transport
of solution is not high. Low permeability and total porosity values were obtained for the samples.
These results are associated with the high resistances of pass of liquid to cross the solid matrix. For
this reason and despite 2 mechanisms of liquid transport into column have been described (intra
and inter-particle), the outer movement of solution around particles appears to be the main
mechanism that determines the caliche leaching.
The physical characterization of solid samples performed in this work is valid for untreated
specimens (or initial rocks previous to leach), because after leaching the pore structure changes
considerably. In further works, the variation of porosity and related parameters might be evaluated
during column leaching.

ACKNOWLEDGEMENTS
The authors wish to thank CONICYT and FIC-R Antofagasta for the supporting through projects
MEL 81105010 and Anillo ACT 1201. J. Ordez thanks CONICYT for the PhD. scholarship.

REFERENCES
Glvez, E.D., Moreno, L., Mellado, M.E., Ordez, J.I., Cisternas, L.A., 2012. Heap leaching of Caliche minerals:
Phenomenological and analytical models Some comparisons. Minerals Engineering 33, 46-53.
Ordez, J.I., Moreno, L., Glvez, E.D., Cisternas, L.A., 2013. Seawater leaching of caliche mineral in column
experiments. Hydrometallurgy 139, 79-87.
Ordez, J.I., Moreno, L., Mellado, M.E., Cisternas, L.A., 2014. Modeling validation of caliche ore leaching
using seawater. The International Journal of Mineral Processing.
SQM webpage. Available at August 2013 in http://www.sqm.com.
USGS, United States Geological Survey webpage. Available at October 2013 in http://www.usgs.gov.
Valencia, J.A., Mndez, D.A., Cueto, J.Y., Cisternas, L.A., 2008. Saltpeter extraction and modelling of caliche
mineral heap leaching. Hydrometallurgy 90, 103114.
Wheeler, A., 2010. Technical report on the Aguas Blancas property, Chile. Internal report, pp 1-57.
Wisniak, J., Garcs, I., 2001. The rise and fall of the salitre (sodium nitrate) industry. Industrial Journal of
Chemical Technology 8, 427-438.

Reducing fresh water use in the

production of metals

Robert Bruce and

Tyler Seaman
Teck Resources Limited,
Canada

Water supply plays an important role in all mining ventures, particularly

with projects in water scarce regions. A comprehensive water supply
strategy for mining projects is important to ensure that fair allocation of
water is maintained for municipal, agricultural and industrial users in the
region. The consumption of fresh water for processing copper concentrates
can be substantially reduced through the adoption of hydrometallurgical
processes that incorporate salt water; particularly when compared to
pyrometallurgical processes that require fresh water.
Developed by Teck, the CESL hydrometallurgical process allows for the
incorporation of salt water to reduce fresh water requirements in the
processing of copper concentrates. Further, it should be noted that water
additions to the process are already minimized through reclaiming process
water by use of a closed-circuit evaporator. Bench-scale testwork is
presented which demonstrates the opportunities related to salt water
additions on metal recovery and overall metallurgical performance. An
average copper extraction of 97.3% was achieved while fresh water
consumption was decreased from 1.63 m3/t concentrate to 0.67 m3/t
concentrate through salt water substitution. Conceptual economic benefits
of salt water addition to the flowsheet are presented. Results from fully
integrated pilot scale trials are discussed which show no negative
metallurgical impacts of salt water integration on settling characteristics
or residue stability.

INTRODUCTION
Availability of fresh water for mining projects, particularly in arid regions, is an important
consideration for all stakeholders. There are several recent examples in the resource industry where
resource development projects have been slowed or halted based on issues relating to water supply
and management. Water use in mining/mineral concentration and concentrate processing/refining
is divided among several main operating steps. Largest among these is the process to produce
mineral concentrates. Water losses from this stage of the process include moisture contained within
concentrate, evaporation, and water lost to tailings. There are several common methods used to
reduce the amount of fresh water used in mineral concentration. These include:

Water Control and Recycling

- Adequate measurement and control of water inputs to the plant
- Use of high efficiency thickeners to decrease water losses to tailings
- Recovery of water infiltration from tailings for use in the process
- Recycling of tailings runoff to the process
Water Substitution
- Use of wastewater/grey water in the process
- Use of desalination plants for water supply
- Direct use of sea water

Of the methods listed, water control and recycling has been employed successfully at mine sites to
decrease the volume of fresh water required. For areas where availability and access to fresh water
is limited, water substitution is evaluated on a case-by-case basis and often fulfills the fresh water
needs of the project. Freeport-McMorans Cerro Verde project is constructing a wastewater
treatment plant to meet site water requirements as well as those of the nearby communities
[International Council on Mining and Metals (ICMM), 2012].
Two examples in Chiles arid northern region where design considerations for reduced water
consumption have been adopted are the application of desalination at BHPs Escondida mine
[Brantes & Olivares, 2008] and the direct use of sea water at Antofagastas Esperanza mine [ICMM,
2012]. In both of these cases, the mine/concentrator are located in an area where direct access to
fresh water is limited and water (either desalinated or seawater) is pumped a significant distance
from the coast to higher elevations to supply the plant. The application and effectiveness of these
measures to decrease fresh water usage is case specific and varies widely based on factors such as
mine location, ore body mineralogy, and impact on flotation performance. Brantes & Olivares
[2008] list average concentrator fresh water usage as 0.79 m3/t ore processed with a range of 0.3 2.1
m3/t ore processed.
For processing of copper sulphide concentrates through smelting, water is required for cooling and
washing of smelter gases as well as in the acid plant. Similar to concentrators, water usage can vary
between operations. A report commissioned by the Chilean Copper Commission (Cochilco) in 2008
[Brantes & Olivares, 2008] lists the average fresh water consumption for a smelter as 3.6 m 3/t
concentrate processed. Using Codelcos Ventanas smelter as an example, a fresh water value of 3.1
m3/t concentrate can be calculated based on numbers reported by the company in their 2011
sustainability report. Another source [Norgate & Lovel, 2006] quotes smelter water consumption of
7.8 m3/t copper cathode for Bolidens Ronnskar operation. Referring to Ronnskars more recent
sustainability report from 2012, a value of 8.2 m3 fresh water/t copper cathode can be calculated.
Assuming 25% copper contained in concentrate, a range of 2.0 3.6 m3/t concentrate will be used as
a benchmark to contrast the water consumption required for Tecks CESL copper process. To

reduce the variability in the analysis to follow, focus will be placed on comparing the following two
methods for refining copper concentrate: pyrometallurgical processing through a smelter-refinery,
and hydrometallurgical processing by use of the CESL process. As the same fresh water
consumption is assumed for the production of concentrates prior to the refining stage, the fresh
water consumption in the mine and concentrator is not considered in the analysis.

CESL CASE
Teck has developed a hydrometallurgical process, called the CESL process, as an alternative to
conventional smelting and refining of copper concentrates for the production of LME Grade A
copper cathode [Bruce et al., 2011]. The CESL copper process is a medium temperature pressure
leach for sulphide concentrates in the presence of chloride catalytic ions (12 g/L). Based on the
presence of chloride in the process, materials of construction for a CESL refinery are specified for a
corrosive environment, which uniquely positions the process for the incorporation of chloride
bearing salt water into the flowsheet. This includes salt water or, potentially, hypersaline bore
water prevalent in some desert regions. This section outlines testwork that has been carried out to
investigate the impact of salt water on metal recovery and the overall metallurgical performance of
the process.

Bench testwork
Bench testwork was recently conducted on a copper sulphide concentrate containing chalcopyrite
and bornite. To determine the appropriate test conditions, specifically the concentration of
dissolved salts in the recycled process solution, a METSIM model was constructed. Figure 1
presents a block diagram of the flowsheet used for modeling that highlights main water inputs to
the process.
In Pressure Oxidation (PO), concentrate is reground and then fed to an autoclave along with
recycled acid feed liquor. Oxygen is injected into the autoclave at 150C and 1380 kPag to oxidize
and leach the minerals. Depending on the degree of acid production in the autoclave, any
remaining oxidized minerals are leached from the PO solids at atmospheric conditions. Acid
required for Atmospheric Leach (AL) is provided from recycled, acidic raffinate generated in
Solvent Extraction (SX). Following AL the residue is washed in Counter Current Decantation (CCD)
with process water to retain soluble metals and chloride. Pregnant leach solution (PLS), consisting
of overflow streams from the PO thickener, AL thickener, and first CCD thickener, is sent to SX for
copper recovery using standard SX/EW technology. Acidic raffinate is sent to an Evaporator circuit
where water is removed to control the plant water balance. Condensate produced in the Evaporator
circuit is used as process water in the plant creating a closed water balance with no liquid effluents.
Finally, a portion of the Evaporator product is processed through a low pH (1.8) Neutralization
circuit for excess sulphate removal.

Figure 1 CESL process block diagram

Fresh water consumption and supply for the base case model (no salt water incorporation) are
shown in TABLE 1. Fresh water inputs were considered for substitution, in part, with salt water.
Table 1 Water supply and consumption distribution for the base case model
Water supply

m3/t conc.

Distribution

Water use

m3/t conc.

Distribution

Fresh water

1.63

51%

Residue Wash

2.19

68%

Evaporator condensate

1.56

49%

Grinding

0.53

17%

Reagent Makeup

0.28

9%

Electrowinning

0.19

6%

Water addition to Electrowinning (EW) requires low levels of impurities in the interest of product
quality as well as to protect the cathode blanks from chloride corrosion, making saline water
unsuitable for this application. It is also assumed that concentrate repulping is carried out in fresh
water. Salt water substitution was examined for all other water requirements which make up 77%
of the total water consumption in the plant. Water required for sealing of pumps as well as gland
water to the autoclave agitators was not considered.
Typical salt water composition used for modelling purposes is presented in TABLE 2. Taken from
an online source [Lenntech, 2014] these values were used to determine the solution composition for
bench testwork.

Table 2 Salt water composition

Element

Chloride

Sodium

Sulfate

Magnesium

Calcium

(Cl )

(Na )

(SO )

(Mg )

(Ca2+)

18.7

10.4

2.5

1.3

0.4

Concentration (g/L)

42-

2+

Several scenarios were run with increasing salt water content from 4%-30% of the total water
requirement for the plant. Acid feed concentration targets were taken from each of these scenarios
and a series of bench-scale leach tests were conducted in a 2L, titanium cast Parr autoclave.
Operating conditions for the tests are presented in TABLE 3. Each pressure oxidation test was
followed by an Atmospheric Leach conducted at a pH of 1.5-1.8 for 2 hours to complete the copper
leaching.
Table 3 Bench test leach operating parameters
Condition

Unit

Range

Solids density

g/L

200-210

Pressure

kPa

1380

Temperature

150

Retention time

min

60

Acid feed [FA]

g/L

60

Acid feed [Cl-]

g/L

11-24

Acid feed [Na+]

g/L

0-6

Acid feed [Mg2+]

g/L

0-2.8

Bench metallurgical results

Outcomes of greatest interest in this series of tests were the potential impact of salt concentration on
metal extraction and sulfur oxidation, which have the greatest economic impact on the process.
Finally the precipitation of sodium, one of the main impurities present in salt water, is considered.

Metal extraction
Copper extraction in the CESL process for a copper sulfide concentrate containing chalcopyrite and
bornite has been demonstrated to exceed 97% [Bruce et al., 2011]. Figure 2 presents the copper
extraction achieved at varying levels of salt water substitution for fresh water. Copper extraction for
all tests averaged 97.3%. All results fell within 2 standard deviations of this mean value.

Figure 2 Copper extraction versus salt water substitution to the CESL process

Sulfur oxidation
Sulfur oxidation within the CESL process has been demonstrated to be limited to a fraction of the
sulfur present in the concentrate. This limits operating costs associated with oxygen consumption
as well as limestone for neutralization of acidic process streams. Higher sulfur oxidation increases
the amount of gypsum residue produced, resulting in increased water consumption for washing of
solids.
Sulfur oxidation is closely related to percent pyrite contained in a concentrate but can be impacted
by other factors such as chloride and sulfate concentrations. Figure 3 presents the sulfur oxidation
measured in the bench tests. A small increase (5% absolute) in sulfur oxidation was noted with
increasing salt concentration in the autoclave feed solution.

Figure 3 Sulfur oxidation versus salt water substitution to the CESL process

Sodium precipitation
Sodium is one of the impurities present in salt water that does not have a convenient method of
removal (i.e. neutralization) from the CESL flowsheet. A certain portion of the sodium fed to the
autoclave precipitates as natrojarosite by reaction (1).

+3

) + 12

) (

) +6

(1)

From the bench results the percentage of sodium precipitation decreases with salt concentration in
the acid feed. This trend continues until sodium precipitation reaches a value of 20%, at which
point the trend stabilizes. This is an important factor to consider for flowsheet design as
precipitation of sodium in the autoclave, along with minor solution entrainment in the final
residue, are the only mechanisms for sodium removal in the CESL flowsheet as it is a closed water
balance.

Water usage
As the amount of salt water added to the flowsheet increases, there is a direct reduction in fresh
water consumption. At 30% salt water substitution fresh water additions are only required for
Electrowinning make-up solution and concentrate repulping in Grinding. Figure 4 presents a
comparison of the fresh water consumption for smelting and that required for the CESL process.
Fresh water consumption for the CESL base is 18 55% lower than that of the smelting case. With
30% salt water substitution, the fresh water consumption for the CESL flowsheet is 66 80% lower
than that of the smelting case. On an intensity of use basis smelting uses 3x to 5x more fresh water
than a modified CESL hydrometallurgical process (30% salt water substitution for fresh water).

Figure 4 Fresh water consumption comparison of smelting to the CESL process

Economic impact
Economic impact of salt water incorporation to the CESL flowsheet was examined on the basis of:
(i) decreased HCl additions to maintain chloride concentration in the plant inventory, and (ii) the
decreased cost related to reduced requirements for fresh water. Capital cost was not considered on
the basis of the CESL process already requiring moderate (12 g/L) levels of chloride in the process
inventory. Titanium would be specified in most applications, though recent work has shown that
application of super duplex steels in certain low temperature applications may be possible [Riha,
Bruce & Schwartz, 2013]. TABLE 4 presents the operating cost savings resulting from salt water
addition for a project where use of desalination is required.

Table 4 Operating cost savings of salt water addition when desalination is required
Salt Water Addition

Unit

0%

4%

8%

12%

15%

23%

31%

HCl addition

kg/t conc.

0.76

Fresh water addition

m3/t conc.

1.63

1.58

1.44

1.33

1.26

0.81

0.67

HCl cost savings

/lb Cu

0.27

0.31

0.31

0.31

0.31

0.31

Fresh water savings

/lb Cu

0.06

0.16

0.25

0.32

0.61

0.74

Total savings

/lb Cu

0.33

0.47

0.56

0.62

0.92

1.05

Unit costs used for HCl and desalinated water in this analysis are: US$280/t HCl (@34%) [vendor
quote] and US$2.8/m3 of desalinated water [Soruco & Philippe, 2012]. Based on this data the cost
savings for salt water addition were 0.33 1.05 cents/lb Cu cathode.
An internal tradeoff study conducted by Teck determined the costs for delivery of water to a site
located at an elevation of 2000 m above sea level to be $1.4/m3 for sea water. One source [Soruco &
Philippe, 2012] gives the operating cost of delivery for desalinated water as $2.8/m3. Using these
values three cost scenarios were considered for salt water substitution: a project where abundant
fresh water was available ($0.2/m3 fresh water), one where fresh water availability was limited
($1.0/m3 fresh water), and one using desalinated water ($2.8/m3 desalinated water). Figure 5
presents CESL process annual savings associated with 30% salt water substitution for each water
constraint scenario, based on 100,000 tpy copper production, compared to the CESL base case
where no salt water is used.

Figure 5 Yearly savings using salt water in the CESL process

In the case where abundant fresh water is available the introduction of salt water results in yearly
savings of $276,000. This is mainly attributed to the decreased requirement for HCl addition to
maintain chloride concentrations in the plant. For the case where fresh water supply is limited the
use of sea water would result in yearly savings of $900,000. For the case where desalinated water is
being used savings of using salt water amount to $2.3 million annually. This data only considers the

operating cost of producing desalinated water and conveying it to the project site; capital cost was
not considered. All three cases present an opportunity for cost savings. However, in cases where
fresh water is unavailable or where restrictions prevent its use, the use of salt water provides a
significant economic benefit when compared to the cost of desalinated water.

Pilot plant trial

Previous pilot plant testwork has been conducted, on a different, although mineralogically similar,
copper sulphide concentrate than that used for the bench testwork, which examined salt water
addition to the CESL flowsheet. As part of a client project a two week pilot plant trial was
conducted. For the case tested, integration of the CESL refinery with an existing heap leach was
considered as an attractive option. Figure 6 presents a diagram of the flowsheet considered.

Figure 6 Heap leach integration flowsheet

Heap leach integration allowed for the elimination of several unit operations based on the ability of
the heap leach to utilize acid produced through sulfur oxidation, negating the requirement for a
neutralization circuit. Based on the presumed evaporation expected within the heap leach
operations, the Evaporator circuit was also eliminated from the flowsheet. Salt water addition was
used to provide a large portion of the autoclave feed liquor (32%) as well as a portion of the wash
solution for the leach residue (25%).
TABLE 5 presents leach circuit metallurgical results for the pilot trial where the plant was operated
with and without the use of salt water. Similar to the bench testwork presented, copper extraction
was not impacted by salt water addition to the flowsheet. A slight increase in sulfur oxidation was
observed.

Table 5 Leach circuit metallurgical results for pilot trial

Operating results

Units

Fresh water

Salt water

Copper extraction

97

97

Sulphur oxidation

33

34

In addition to the variables presented in TABLE 5, performance of the integrated flowsheet was also
consistent for both cases. This included residue settling characteristics, residue stability, and SX
performance. Residues produced while incorporating salt water were washed in a thickener train.
Thick underflow densities, in excess of 50% solids, were maintained throughout the salt water trial.
Standard USEPA Toxicity Characteristic Leaching Procedure tests (TCLP) were conducted on the
washed residues produced. Samples of residues produced without salt water addition were tested
as well as those produced with salt water added to the flowsheet. Both samples met the USEPA
standard for all elements and no significant increase in leachable metals was detected for the
sample produced using salt water. Finally, solvent extraction performed well with no additional
transfer of impurity elements to the electrolyte, allowing for the production of LME Grade A
copper cathode.

CONCLUSION
As compared to the fresh water consumption rate of 2.0 - 3.6 m3/t concentrate for pyrometallurgical
smelting and refining of copper, the CESL base case provides an opportunity to decrease water
consumption by 20 - 50%, at 1.63 m3/t concentrate. The use of salt water decreases fresh water
consumption to less than 1.0 m3/t conc., a decrease of 66-80% when compared to smelting.
Substitution of fresh water with salt water in the CESL process provides an economic benefit in all
cases but this benefit is most pronounced for projects where an alternative to fresh water use is
required, such as desalination. Another added benefit of using salt water is the ability to process
directly concentrates from milling and concentrator operations that have adopted salt water
processing, while also capitalizing on the learnings and operating practices of these mills.
Water supply for mining and metal refining is a topic that has seen much attention over the last
decade and this trend is likely to continue. To meet increasing global demand for copper while
ensuring adequate water availability for all stakeholders, the copper industry needs to focus on
decreasing the requirements for fresh water and increasing efficiency in its use in all aspects of the
mineral and metal production process.

REFERENCES
Brantes R., Olivares G. (2008). Good Practices and the Efficient Use of Water in the Mining Industry. Santiago,
Chile: COCHILCO Directorate of Studies.
International Council on Mining and Metals. (2012). Water management in mining: a selection of case studies.
London, UK: Author.
Norgate, T.E., Lovel, R.R. (2006). Sustainable Water Use in Minerals and Metal Production. Water in Mining
Conference 2006 Proceedings, Brisbane, Australia.

Codelco 2011 Sustainability Report. (2011). Retrieved March 30, 2014 from
http://www.codelco.com/flipbook/reporte_sustentabilidad/2011/sitio/en/environmentalmanagement.html
Sustainability Report Ronnskar 2012. (2012). Retrieved March 30, 2014 from
http://www.boliden.com/Documents/Press/Publications/Sustainability%20Reports/Sustainability_Ron
nskar_2012_EN.pdf
Lenntech. (2014). Major ion composition of seawater. Retrieved from http://www.lenntech.com/compositionseawater.htm.
Bruce, R., Kadereit, H., Mayhew, K., Mean, R., Nagy, A., Wagner, O. (2011). Unlocking value in copper arsenic
sulphide resources with the copper-arsenic CESL technology. Proceedings of the 6th International Seminar
on Copper Hydrometallurgy 2011, Via del Mar, Chile.
Riha, J., Bruce, R., Schwartz, D. (2013). Corrosion studies at Tecks hydrometallurgical facility (CESL):
Identification of an alloy substitute for titanium in certain applications. COM 2013 Proceedings,
Montreal, Canada.
Soruco, L., Philippe, R. (2012). Upcoming trends in water supply costs for copper mining in arid regions.
Proceedings of the 3rd International Congress on Water Management in the Mining Industry,
Santiago, Chile.

chapter 5
Desalination and seawater use

sart to remove zinc and copper from a silver

Merrill Crowe barren leach solution

David Sanguinetti and

Farzad Mohammadi
BioteQ Environmental
Technologies, Canada
scar Lpez
BioteQ Water, Chile

The SART process (sulfidization-acidification-recycle-thickening) was

developed more than 15 years ago and in recent years has been applied at a
number of facilities. The vast majority of these facilities, as well as the
literature on the SART process, have focused on the removal of copper from
gold processing circuits. BioteQ has been involved in the design and
operation of a number of these facilities. Building on this experience,
BioteQ has recently performed a number of laboratory studies on applying
SART technology to remove zinc, particularly in Merrill Crowe applications.
This paper describes the results of one such study, where copper as well as
zinc were removed from the BLS of a silver Merrill Crowe circuit. In
addition to testing conventional SART technology, BioteQ also tested
BioSART, whereby the source of sulfide is changed from chemical
(NaHS) to biologically generated H2S. The advantage of BioSART is that
the reagent cost is significantly reduced, thus improving the operating
economics of the process. The results of this test work are presented,
demonstrating the ability to successfully recover the copper and zinc, both
separately and in a mixed concentrate, as well as regenerating a high
percentage of the WAD cyanide. The differences between conventional
SART and BioSART are explored and the operating and capital cost
implications are discussed.

INTRODUCTION
The SART process has been applied at an industrial scale in cases where cyanide soluble copper is
interfering with the gold recovery process (Littlejohn, Kratochvil & Hall, 2013), but it has not yet
been applied to cases where zinc or a combination of copper and zinc are responsible for
consuming cyanide. Additionally, all current industrial applications of SART use NaHS as a source
of sulfide, although it has been suggested that in certain cases biogenically derived H2S may be
advantageous (Adams, Lawrance & Bratty, 2008). In order to further elucidate this question BioteQ
proposed to compare in a lab study the use of H2S to NaHS in the recovery of both copper and zinc
from a Merrill Crowe barren leach solution (BLS) to determine the advantages and disadvantages of
each reagent.

The SART process

The SART process was first developed in the 1990s by SGS Lakefield Group and Teck Corporation
in response to issues associated with cyanide soluble base metals. The technology is an industrial
hydrometallurgical process for recovering aqueous cyanide and base metals from solution in
separate recyclable or saleable streams. SART was first applied commercially at the Telfer mine in
Australia (Barter et al., 2001) where the base metal of concern was copper. SART treatment involves
Sulfidization and Acidification of the feed cyanide solution, followed by capture and Recycling of
cyanide, then neutralization of the effluent and Thickening to separate precipitated solids. Sulfuric
acid is introduced to the feed liquor to reduce the solution pH to 4-5, breaking weak acid
dissociable (WAD) base metal cyanide complexes and forming hydrogen cyanide. Sulfide is
introduced to the reactor as hydrogen sulfide gas (H2S) or sodium hydrosulfide (NaHS), causing
the base metals to precipitate as metal sulfides, such as copper sulfide (Cu2S) or zinc sulfide (ZnS).
Silver in solution also reports to the sulfide precipitate. Any hydrogen cyanide gas that does not
remain in solution is scrubbed and returned to the subsequent neutralization stage, while metal
sulfide solids are collected in a high-grade concentrate. The effluent solution is then neutralized to
pH 10 or greater with caustic (NaOH) or lime (Ca(OH) 2) before the contained cyanide is recycled to
the leaching stage as sodium cyanide and/or calcium cyanide. If the effluent is neutralized with
lime then additional equipment, including a clarifier and filter press, is required to thicken and
collect gypsum (CaSO4.2H2O) to allow recycle of clear liquor. The following chemistry summarizes
the reactions when zinc is the base metal of concern:
Sulfidization and Acidification
Na2Zn(CN)4(aq) + NaHS + H2SO4 ZnS(s) + 3HCN + Na2SO4(aq) + NaCN(aq)

(1)

Recycling and Thickening

3HCN + Na2SO4(aq) + 1.5Ca(OH)2 CaSO4.2H2O(s) + 2NaCN(aq) + 0.5Ca(CN)2(aq) + H2O

(2)

The net result of the SART process is that base metal cyanide complexes are broken down, cyanide
is recycled as free cyanide and base metals are recovered as saleable metal sulfide concentrates. The
primary reagents required in the SART process are a sulfide source, sulfuric acid and a
neutralization agent. Use of SART technology yields direct economic benefit to the plant by
reducing operating costs due to cyanide consumption, recovering cyanide leached copper/zinc as a

saleable by-product and by reducing demands on detoxification/clean-up processes applied to

tailings and effluent (Lawrence & Lopez, 2011). Specific to the Merrill Crowe process the reduced
leaching efficiency of the recycled BLS, due to high zinc cyanide concentration, can be avoided with
application of the SART process. When the zinc content of the recycled BLS is increased, more
sodium cyanide has to be added via the fresh leach solution in order to keep the activity of the
combined leach solution constant. This will eventually lead to solution fouling and result in
unnecessarily high cyanide concentration in the leach solution, necessitating extra zinc addition
during the precipitation stages (Kurama & Catalsarik, 2000).

The Merrill Crowe process

Since the 1800s zinc has been used to recover precious metals from leachate. At an industrial scale
the process was first applied to gold mining by Charles Merrill. Thomas Crowe improved the
process efficiency with the introduction of solution deaeration. In gold/silver mining, the Merrill
Crowe process is a method of separating these metals from the cyanide solution. The solution is
first cleared with filtration or counter current decantation. Oxygen is then removed from the now
clear solution through application of a vacuum. Once de-oxygenated, zinc dust is added to the
solution to precipitate gold/silver (Gexla, Fuerstenau & Marsden, 1997). This is an electrochemical
process in that elemental zinc is oxidized (dissolution) and silver is reduced (deposited on zinc
particles). The efficiency of the process is a function of the available surface area of the zinc, which
in this case is zinc dust, and in return the amount of zinc required is dictated by the amount of
silver/gold to be recovered. The overall reaction for silver recovery is shown below.
2NaAg(CN)2 + Zn 2Ag + Na2Zn(CN)4

(3)

Under most circumstances the quantity of zinc dissolved in the leach solution is not significant
enough to cause problems for the process. In certain cases where there is a large concentration of
cyanide soluble zinc in the same ore body as the silver that is being leached, the loss of cyanide to
the zinc can become a significant factor in the process economics. In such cases, the application of
SART to the BLS to recover zinc and regenerate cyanide can be considered.

Case study background

The case study involves an early stage high-grade silver/zinc project that contains two distinct
mineralization targets: a shallow silver zone and a high-grade zinc oxide zone situated directly
below and to the east of the silver zone. The shallow silver zone resource has open pit potential and
would employ a Merrill Crowe process to recover the silver oxide.
Extraction of the silver mineralization will be accomplished through agitated leach with elevated
levels of cyanide solution. This high cyanide concentration leads to higher mineral dissolution in
general and given zincs affinity for cyanide, it will preferentially compete for the cyanide in the
leach solution, thereby reducing the availability of cyanide for silver recovery. This high
concentration of zinc cyanide complex may negatively impact the economics of the project through
high cyanide consumption and costly cyanide tailings destruction.
The purpose of the SART plant is to reduce the impact of cyanide soluble zinc on silver recovery.
This is done by liberating the zinc metal from cyanide, precipitating zinc as a sulfide concentrate
and regenerating the zinc complexed cyanide as free cyanide. The SART process therefore

decreases the net cyanide consumption in the leach circuit and is expected to recover marketable
zinc concentrate as an incremental revenue source.
In order to demonstrate the effectiveness of SART for this application a quantity of BLS generated
from the processing of a representative sample of ore was tested. The composition of this BLS is
given in TABLE 1.
Table 1 BLS feed assay
Physical Tests
Hardness (as CaCO3)

Units
< 43

mg/L

400

mg/L

WAD

7,370

mg/L

Total

7,500

mg/L

Free

7,090

mg/L

Arsenic (As)

15.0

mg/L

Cobalt (Co)

0.231

mg/L

Copper (Cu)

38.2

mg/L

Iron (Fe)

52.1

mg/L

Silver (Ag)

0.307

mg/L

Sodium (Na)

6,040

mg/L

Zinc (Zn)

2,450

mg/L

Anions and Nutrients

Sulfate (SO4)
Cyanides

Dissolved Metals

TEST METHODOLOGY
A schematic of the test apparatus is shown in Figure 1. The tests were conducted in batch format.
The test apparatus comprises two sealed reactors (the contactor and the HCN scrubber). The
contactor (1,000 mL) has ports for:

pH probe

ORP probe

Sampling

Headspace connection to the scrubber

Manometer

The contactor also has inlets for acid, NaHS, H2S gas and flocculent/coagulant addition. The HCN
scrubber was connected to the headspace of the zinc contactor and allowed gas from the headspace
of the contactor to bubble through the caustic scrubbing solution. The contactor headspace was also
connected to a manometer that showed the system pressure. Although all vessels were sealed, the
entire setup was placed in a fume hood for safety.

Figure 1 SART setup schematic

The BLS was quantitatively transferred into the contactor. The sulfuric acid and NaHS solution
tanks were volumetric cylinders isolated from the atmosphere, and contained 250 mL and 100 mL
of sulfuric acid and NaHS stock solutions at known concentrations, respectively. Pumps P101 and
P106 were variable speed peristaltic pumps (0 to 5 mL/min). Samples were withdrawn from the
contactor sampling port using a 60 mL syringe.
In the case of precipitation with H2S gas, BioteQ used a laboratory scale BioSART process that
employs two sealed reactors: the gas generator where hydrogen sulfide gas was produced and the
contactor where the hydrogen sulfide was mixed with the feed to be treated. NaHS was pumped
from the NaHS tank to the gas generator where it was contacted with sulfuric acid. The reaction
between NaHS and sulfuric acid produced pure H2S gas which was used for metal sulfide
precipitation. The sulfide dosage rate was controlled by a variable speed peristaltic pump. The
headspace of the gas generator was pumped through the solution in the contactor to facilitate
reaction and the headspace of the contactor and gas generator were linked to allow passive
pressure equalization. The pressure of the system was monitored using a manometer. A schematic
of the gas generator is shown in Figure 2.

Figure 2 BioSART laboratory setup schematic

DISCUSSION AND RESULTS

Metal recovery
Due to the presence of copper as well as zinc in the solution, it was of interest to test the
precipitation of zinc and copper as separate products as well as confirming the recovery of a highgrade zinc product to generate an incremental revenue source. In order to achieve these objectives,
sulfide and acid dosages were the process parameters that were controlled.
The tests were conducted in one and two-stage procedures to determine if copper and zinc could be
effectively and separately removed from solution. In single stage tests the entire dosage of acid and
NaHS required for the whole experiment was added in one stage with the goal of removing copper
and zinc from solution simultaneously. Two-stage tests aimed to first recover zinc from solution
while minimizing copper deportment in the first stage and then maximize copper recovery while
minimizing zinc deportment in the second stage. This was achieved by using different pH and
sulfide dose set points for the first and second stages. TABLE 2 shows selected results that illustrate
the removal and separation of copper and zinc.
Table 2 Select results showing copper and zinc removal
Test #

Stage

Sulfide source

(X of Y)

Sulfide dose

Cu recovery

Zn recovery

g/L

mg

mg

1 of 1

NaHS

2.57

88

98

48

100

2,535

1 of 1

H2S

2.62

90

98

48

100

2,535

1 of 2

NaHS

3.66

127

93

2,340

2 of 2

NaHS

3.66

127

96

46

100

2,499

Test #1 and #3 in TABLE 2 show that when using NaHS or H2S, almost all copper and zinc can be
recovered from solution. Compared to Test #6, using NaHS, stages 1 & 2 results show that 93% of
the zinc can be recovered from solution while only 6% of the copper is recovered in stage 1. During

stage 2 only 7% of the zinc (which remained in solution from stage 1) was precipitated while 90% of
the total copper was recovered.

Conventional SART vs. BioSART

The recovery of copper and zinc as copper sulfide and zinc sulfide can be achieved using different
sulfide sources. However, depending on the solution composition and specific process
requirements the efficiency of each sulfide source may be different. Further testing was done to
determine which reagent (NaHS or H2S) worked better under otherwise similar conditions.
Both one-stage and two-stage tests were trialled to compare the relative effectiveness of NaHS
versus H2S gas for copper and zinc recovery. The first set of tests (Test #1 & 3) were one-stage tests
intended to simultaneously recover copper and zinc as a mixed metal concentrate. These tests were
conducted under similar conditions. The second set of tests (Tests #6 & 7) were two-stage tests
intended to selectively recover zinc followed by copper and were also conducted under similar
conditions. TABLE 3 presents the experimental conditions as well as the measured copper and zinc
recoveries.
Table 3 Comparison of NaHS and H2S as a sulfide source
Test #

Stage
(X of Y)

Sulfide
source

Acid dose

Sulfide dose

Cu recovery

Zn recovery

g/L

g/L

mg

mg

1 of 1

NaHS

16.28

117

2.57

88

98

48

100

2,535

1 of 1

H2S

14.12

102

2.62

90

98

48

100

2,535

1 of 2

NaHS

12.45

90

3.66

127

93

2,340

2 of 2

NaHS

16.52

119

3.66

127

96

46

100

2,499

1 of 2

H2S

7.42

53

3.66

127

50

1,254

2 of 2

H2S

12.21

88

3.66

127

80

38

97

2,437

In Tests #1 and #3 it can be seen that, for comparable sulfide doses, the copper and zinc recoveries
were similar. Furthermore, the recoveries were higher than those expected based on the sulfide
dose; given that at a 88% sulfide dose only 88% of the total copper and zinc would be expected to be
recovered by sulfide precipitation. Under the conditions of more than 100% stoichiometric acid
addition it is possible that some of the zinc or copper precipitated as zinc cyanide or copper cyanide
will contribute to higher than expected metal recoveries. Examining Tests #6 and #7, a marked
difference between the NaHS results and the H2S results was observed when NaHS was used
(Test #6) compared to H2S (Test #7), higher overall recoveries of both metals were achieved. This
may be attributed to the batch-wise operation in the laboratory scale testing, resulting in a portion
of H2S not reacting with the solution and accumulating in the reactor headspace. Both tests
demonstrate that zinc can be precipitated with a very low level of copper contamination.
A final consideration when evaluating the performance of NaHS versus H2S is that any excess
sulfide added as NaHS remains in the SART effluent and cannot be recovered. However, most of
the excess H2S added in a H2S based SART process can be recovered and recycled, as it will be in
the gas phase. This represents a potentially important economic difference between the two sulfide
sources whereas H2S would use fewer reagents in terms of acid dosage and excess sulfide

consumption. The purity of the zinc sulfide is more of an issue if higher copper concentrations are
expected in the future. Based on the BLS for this project, the mass of copper that stands to be
recovered is only 1.9% of the mass of zinc that stands to be recovered. This means that even if all
the copper and zinc were simultaneously precipitated it would still produce nearly pure zinc
sulfide; decreasing the purity by around 2%.

Silver deportment
The BLS was measured as containing 0.30 mg/L silver and 3,342 mg/L zinc. The detection limit for
dissolved silver is 0.2 mg/L so an accurate determination of silver deportment cannot be attained by
dissolved metal analysis. Silver deportment can, however, be determined by looking at relative
silver and zinc composition of the solids. If there was 100% silver deportment to the solids, then
based on the zinc recovery the solids should be 0.012% - 0.015% silver. The solids assay shows
0.014% - 0.016% silver. This clearly demonstrates that there is very nearly 100% silver recovery in
the solids as would be expected.

Cyanide regeneration
In addition to the recovery of zinc, the value of the SART process is the regeneration of the cyanide
that is bound in WAD cyanide complexes. The calculated regeneration of cyanide during the lab
procedure was in excess of 98%. However, BioteQs field experience suggests that actual full scale
results are likely to be less than this so a value of 90% is used for the purposes of economic analysis.

ECONOMIC ANALYSIS
Using the results from the testing described above, a preliminary economic analysis was carried
out. The costs to erect a conventional SART plant as well as a BioSART plant were estimated, in
both cases assuming a single circuit recovering the zinc and copper together. As can be clearly seen
in Table 4, there is a compelling economic argument for both cases. The extra investment required
for the BioSART technology will be recovered in less than three years vs. the conventional SART,
making it an attractive way to increase the value of the project.
Table 4 SART/BioSART preliminary evaluation summary
SART

BioSART

Plant Design Capacity (m /hour)

550

550

Number of Circuits

Single

Single

Budgetary Installed Cost Estimate (C$ 30%)

26,000,000

51,000,000

Estimated Operating Expense (C$/A)

38,500,000

27,950,000

Annual Cyanide Regenerated (tonnes as NaCN)

28,400

28,400

Annual Value of Cyanide Regenerated (C$)3,4

48,350,000

48,350,000

Annual Zinc Recovery (tonnes as Zn)

11,580

11,580

Annual Revenue of Zinc Recovery (C$)5

16,000,000

16,000,000

Annual Profit (C$)

25,850,000

36,400,000

Simple Payback (years)

1.0

1.4

Based only on primary reagents and power included in operating costs.

Based on cyanide regenerated from WAD cyanide zinc and copper complexes.
3The value of cyanide regenerated is based on 50% of the available cyanide liberated from the decomposition
of dissolved Zn(CN)42- in the SART/BioSART process. Note that in the case of the Merrill Crowe process
without SART, 2 moles of cyanide are lost from the system as insoluble Zn(CN) 2 for every 4 moles of cyanide
that is complexed with zinc.
4 Based on NaCN purchase price of C$ 3,400/DMT.
5 Based on a zinc price of C$ 1.98/kg and a NSR of 70%.
1
2

CONCLUSIONS
The results from this test work point to a substantially wider range in applicability for SART
technology than has been implemented at an industrial scale to date. In addition to copper, SART is
also effective when zinc is consuming free cyanide. Beyond this, if the leach solution composition is
such that there are elevated concentrations of both copper and zinc, with careful control it is
possible to separate the copper and zinc into separate sulfide concentrates in addition to recovering
the cyanide. When doing so, it can be expected that high recoveries of both zinc and copper will be
achieved, as well as at least 90% regeneration of the cyanide.
It has also been demonstrated that there is more than one effective source of sulfide for the SART
process. In addition to the conventional use of NaHS, biogenically sourced H 2S, as in a BioSART
process, is also effective. NaHS has a lower capital cost, while biogenically sourced H 2S has a
substantially lower operating cost.

REFERENCES
Adams, M., Lawrence, R. & Bratty, M. (2008) Biogenic Sulphide for Cyanide Recycle and Copper Recovery in
Gold-Copper Ore Processing. Minerals Engineering, vol. 21, no.6, pp. 509-517.
Barter, J., Lane, G., Mitchell, D., Kelson, R., Dunne, R., Trang, C. & Dreisinger, D. (2001) Cyanide Management
by SART, Cyanide: Social, Industrial and Economic Aspects, C. Young (ed), TMS, pp. 549-562.
Gexla, C., Fuerstenau, M. & Marsden, J. (1997) Study of Merrill Crowe Processing. Part I: Solubility of Zinc in
Alkaline Cyanide Solution, International Journal of Mineral Processing, vol. 49, pp. 171-183.
Kurama, H. & Catalsarik, T. (2000) Removal of Zinc Cyanide from a Leach Solution by an Anionic Ionexchange Resin, Desalination, vol. 129, pp. 1-6.
Lawrence, RW. & Lpez, O. (2011) Improving the Economics of Gold-Copper Ore Projects Using SART
Technology, Proceedings of the World Gold Conference 2011, Montreal, Canada, 3-4 October.
Littlejohn, P., Kratochvil, D. & Hall, A. (2013) SART for Complex Gold Ores, Proceedings of the World Gold
Conference 2013, Brisbane, Australia, 26-29 September.

Development of the Albion Process

plant to treat refractory concentrates
from the gpm Gold Project

Mike Hourn and

Paul Voigt
Xstrata Technology,
Australia
Duncan Turner
Core Process Engineering,
Australia

The GPM Gold Project is located in Armenia, and consists of an open cut
mine at Zod, near the Azerbaijan border, and a CIL processing plant at
Ararat near the Turkish border. Mining at Zod commenced in 1976, and
focused on near surface oxide ores, which overlay deeper refractory sulfides.
Historical mining has now almost depleted the oxide ores, and the sulfide
content of ore delivered to the processing plant at Ararat is increasing. Gold
and silver recoveries through the Ararat plant are declining.
GeoProMining, the owners of the project, are expanding the Ararat
facility to deal with the increasing sulfide content in the ore. GeoProMining
are refurbishing an existing concentrator to recover a sulfide concentrate
from the ore, and are constructing an Albion Process plant to oxidize the
refractory concentrate. Xstrata Technology (XT) is providing the Albion
Process plant as a technology package.
Run of mine ore will be ground and subjected to flotation to produce
approximately 100,000 tonnes per annum of an 18% w/w sulfide concentrate
and 900,000 tonnes per annum of flotation tailings. The Albion Process
plant will oxidize the sulfide concentrate, to liberate gold and silver for
recovery. The oxidized concentrate and flotation tailings will be
recombined and processed through the existing CIL plant to recover over
90% of the gold and 85% of the silver as dore from run of mine ore.
Construction of the Albion Process plant is nearing completion and
plant commissioning will begin in March/April, 2014.

The GPM Gold Project

The GPM Gold Project is owned by GeoProMining Gold LLC and is located in Armenia. The project
consists of an open cut mine at Zod, near the Azerbaijan border, and a processing plant at Ararat
near the Turkish border. The Ararat plant has a milling and flotation facility built during the soviet
era, with a capacity of 1 million tonnes annually, and a CIL plant built in 1997, with a capacity of 1.5
million tonnes annually. The gold bearing ore, mined at the Zod Mine, is transported to the process
plant at Ararat via a state owned rail link.
The Zod deposit originally consisted of weathered oxide ores overlying deeper sulphides.
Arsenopyrite and pyrite are the major sulphide minerals. Historical mining has depleted the oxide
ores, and the processing plant at Ararat currently treats significant quantities of sulphide ore with
increasing amounts of gold locked within refractory sulphides.
The mineral reserves for the Zod mine, at August 2011, were estimated to be 14.2 million tonnes at a
gold grade of 4.3 g/t. The mineral resources were estimated to contain 28 million tonnes of ore in
the indicated category, at a gold grade of 4.2 g/t and 16 million tonnes of ore in the inferred
category, at a gold grade of 4.2 g/t.
GeoProMining are expanding the Ararat plant by re-commissioning an existing flotation
concentrator to recover a sulphide concentrate from the ore, and constructing an Albion Process
(Hourn & Turner, 2012) plant to oxidize the refractory concentrates. The Albion Process plant will
convert the sulphides to oxides, breaking down the sulphide matrix and liberating gold and silver
for recovery. Tailings from the concentrator and the Albion Process plant will be combined and
transferred to an existing CIL plant to recover the gold and silver as bullion.
Refurbishment of the concentrator and construction of the Albion Process plant commenced in 2013,
with the plant scheduled for commissioning in early 2014. The Albion Process plant will have the
design capacity to process up to 110,000 tonnes per annum of refractory concentrate from the Zod
mine.

Deposit geology
The Zod deposit is located in the Vardenis District of Western Armenia within a setting of
volcanogenic and volcanogenic-carbonate sequences, with gabbro-peridotite intrusions that have
metamorphosed to serpentinite (Konstantinov & Grushin, 1970).
Gold mineralization is associated with carbonate alteration of ultramafic rocks and is commonly
hosted within hydrothermal alteration zones, represented by talc carbonate and quartz-carbonate
assemblages. The ore is moderately hard with a medium level abrasion index.
Gold occurs as native free gold, finely dispersed gold in arsenical sulphides, gold tellurides and
secondary native gold remaining after oxidation of sulphides and tellurides. Silver occurs in its
native form in quartz, chalcopyrite and pyrite, and as silver tellurides.
The deposit has an average sulphur grade of 1.4 % w/w, with an average gold and silver grade of
4.54 g/t and 4.65 g/t, respectively. The arsenic grade across the deposit is 0.3 %w/w. The majority of
the sulphides occur as relatively coarse mineral grains. The dominant gangue minerals are quartz,
talc and chlorite, with minor magnesite, dolomite and calcite.

Development testwork
Development testwork for the GPM Gold Project began in 2009 initially with batch testwork, and
culminated in a continuous flotation and Albion Process pilot plant run over the months May and
June 2010. Approximately 4,600 kg of sulphide ore samples were collected from across the Zod ore
body to support the testwork program. The samples were classified by ore type, spatial location
and sample type and blended into 163 composites. The composites were then grouped into the four
major orebodies identified in the primary sulphide resource orebodies 1, 4, 16 and 23.
Diagnostic leaching and ore characterization testwork (Rohner & Andreatidis, 2010) confirmed that
an average gold recovery of only 48 %w/w was possible from the blended Life of Mine ore adopting
conventional carbon in leach (CIL) leaching methods, and that the majority of the refractory gold
was present in arsenical minerals, such as arsenopyrite and arsenical pyrite. Laser ablation work
showed that the majority of the pyrite had levels of arsenic in the lattice, averaging 0.9 %w/w.
Comminution testwork focused on generating comminution modeling parameters to determine the
capacity of the existing crushing and grinding circuit at the Ararat plant. The ore displayed an
average Bond Crushing Index of 10 kWh/t, an abrasion index of 0.085 and an unconfined
compressive strength of 59 kN. The Bond Ball Mill work Index was 16.5 kWh/t and the Bond Rod
Mill work index was 15.8 kWh/t. Modeling work by SMMC (Morrell, 2010) confirmed the milling
circuit at the Ararat plant would be capable of processing between 0.9 and 1 Mt/a of ore from the
Zod deposit, with minor refurbishment.
Batch and locked cycle flotation campaigns were completed on the testwork samples and a
flowsheet consisting of a bulk roughing and single cleaning stage was developed and taken
forward into a continuous pilot run. The continuous pilot plant testwork proved a sulphide
recovery of 93 % could be achieved from the Zod ores, at a mass recovery of 9 10 %. Gold
recovery to the sulphide concentrate was 87 %, at a silver recovery of 91 %. The sulphide grade of
the concentrate was in the range 16 18 %.
The analysis of the blended pilot plant cleaner concentrate is presented in TABLE 1.
Table 1 Cleaner Concentrate XRD Data
Mineral

Chemical Formula

Unidentified
Arsenopyrite

(%w/w)
20.03%

FeAs S
CuFeS2

6.78%

(Fe,Mg)3Fe3AlSi
3O10(OH) 83 )2
CaMg(CO

1.58%
0.40%

Magnesite

PbS
MgCO3

Plagioclase (Albite)

NaAlSi 3O8

4.50%

Pyrite

FeS2

36.63%

Pyrrhotite

FeS
SiO2
Mg3Si4O10 (OH)2

5.60%

Chalcopyrite
Clinochlore
Dolomite
Galena

Quartz
Talc

2.43%
1.30%
2.80%

3.90%
17.58%

The flotation tailings contained 13 % of the gold, and CIL testwork indicated a gold recovery of 60 %
could be achieved from the flotation tailings at modest reagent demand. The CIL plant at Ararat has
a capacity well in excess of the 1 million tonnes per annum treatment rate for the project, and so cotreatment of both the oxidized flotation concentrate and the flotation tailings was incorporated in the
design.
IsaMill signature plot testwork on the pilot plant composite concentrate sample returned a specific
grinding energy of 59 kWh/t to grind the concentrate to the target 80 % passing size of 10 m.
Extensive testwork was then carried out to determine the best oxidative leach pH for the finely
ground concentrates. The testwork examined oxidative leaching under mildly acidic conditions
for selective oxidation of the arsenopyrite minerals, and leaching at a more neutral pH. All tests
were carried out under atmospheric pressure, with oxygen gas as the oxidant.
Leaching at near neutral pH was ultimately chosen for the oxidative leach. Leaching at near
neutral pH allowed lower cost materials of construction to be used in the leaching circuit, and
resulted in a final residue with more stable arsenic phases when tested in accordance with the
USEPA TCLP protocol. Cyanide and lime demands were lowest for the residue generated under
near neutral pH, and the gold and silver recoveries were higher.
The two major oxidative leach reactions observed under the near neutral oxidative leaching
conditions were:
Pyrite:

FeS2 + 15/4O2 + 9/2H2O + 2CaCO3 = FeO.OH + 2CaSO4.2H2O + 2CO2

Arsenopyrite: FeAsS + 7/2O2 + 2H2O + CaCO3 = FeAsO4 + CaSO4.2H2O + CO2

Confirmatory bench scale oxidative leaching testwork was then carried out on flotation
concentrates from the four main orebodies at the Zod deposit. Economic modeling work that
compared the capital and operating costs for the Albion Process plant at varying levels of
sulphide oxidation was carried out using the batch leach test results. The modeling work showed
that a sulphide oxidation of 70% returned the highest Net Present Value for the project, at a
hurdle rate of 10 %. The gold recovery at this level of oxidation was 93%.
Continuous pilot plant oxidative leaching testwork was then carried out on a blended concentrate.
The continuous pilot testwork confirmed a sulphide oxidation of 70 % was required to achieve an
average gold recovery from the blended feed of 93%. The average silver recovery was 80%. A
design oxidation target of 75% was taken forward into detailed design of the oxidative leach circuit.
The oxygen demand for the concentrate to achieve the design oxidation of 75 % was 336 kg/tonne,
and the limestone demand was 326 kg/tonne. Mass and heat balance modeling indicated an average
operating temperature in the oxidative leach circuit of 96 oC.

Process plant description

Plant general
The Ararat Process Plant experiences a hot, dry summers and cold winters, with an absolute
maximum temperature of +42C; and a minimum of -30C. The region is located is classified as
semi-desert with an average rainfall of 238mm. The maximum 10 days depth of snow mantle is
35cm, at a design snow load pressure of 70 kg/m. The area is seismically active, and the plant is

designed to survive a magnitude 7.2 earthquake. A flowsheet For the GPM Gold Project is shown in
Figure 1.
RECOVERED
WATER

ORE

PRIMARY
MILLING

MILLED
ORE

BULK FLOTATION

FLOTATION
TAILINGS

CIL
FEED

CYANIDE LEACH
(CIL)
PLANT

CIL
TAILINGS

TAILINGS
STORAGE

FLOTATION
CONCENTRATE
ALBION PROCESS PLANT
OXYGEN
ISAMILL
GRINDING PLANT

OXYGEN PLANT

OXIDISED
RESIDUE

CRUSHED
LIMESTONE

OXIDATIVE
LEACHING PLANT

LIMESTONE
PLANT

Figure 1 GPM Gold Project Overall Flowsheet

Comminution and flotation

Ore is mined at the Zod Mine by open cut methods and delivered to a run of mine stockpile for
blending ahead of crushing. Crushed ore is trucked to a stockpile at a rail siding and loaded into
rail cars for transport to the Ararat Plant. Ore is recovered at the plant by tippler, into ore storage
bins. Ore is recovered from the bins by apron feeder and conveyed to the comminution circuit.
The comminution circuit consists of two parallel milling lines. Each line has a primary 1600 kW
Autogenous Grinding mill, operating in closed circuit with spiral classifiers, followed by two
secondary 630 kW Ball Mills, each operating in closed circuit with cyclones. The Ball Mills are
configured in series. Cyclone overflow from the second Ball Mill is currently transferred directly to
the CIL plant, however will gravitate to the flotation circuit once refurbishment is completed.
The flotation circuit has not been operated for over 10 years. The existing flotation plant equipment
consists of two banks with five cells in each, and two banks with four cells in each. Each cell is fitted
with dual 45 kW agitators, with a cell volume of approximately 32 m 3.
Slurry will first pass through a Pre-flotation Conditioner cell which will be operated without
aeration for conditioning of the slurry with MIBC frother. Conditioned slurry will then gravitate to
Pre-flotation cells, which will be used for pre-flotation of talc and carbonaceous slimes. The preflotation concentrate will be transferred to the 15 m diameter Flotation Tailings Thickener.
Tailings slurry from the pre-flotation stage will gravitate to a Rougher Conditioner cell, which will
be used for conditioning of the slurry with copper sulphate and collector prior to rougher flotation.
Conditioned slurry will flow to the Rougher/Scavenger cells to produce a rougher concentrate for
cleaning. The tailings slurry from Rougher/Scavenger flotation will be transferred to the Flotation
Tailings Thickener, with the thickened slurry pumped to the CIL circuit.
Rougher concentrate will be transferred to the Cleaner Flotation Bank. The final Cleaner
Concentrate will be advanced to the a 10 m diameter Flotation Concentrate Thickener, with the
thickened concentrate pumped to the IsaMill circuit for fine grinding. Cleaner tails will join the
rougher/scavenger tails and advance to the CIL circuit.

Albion process
The Albion Process is a combination of ultrafine grinding and oxidative leaching at atmospheric
pressure. The Albion Process technology was developed in 1994 by GlencoreXstrata and is
patented worldwide. There are four Albion Process plants currently in operation.
The first stage of the Albion Process is fine grinding of the concentrate. Most sulphide minerals
cannot be efficiently leached under normal atmospheric pressure conditions. The process of
ultrafine grinding results in a high degree of strain being introduced into the sulphide mineral
lattice. As a result, the number of grain boundary fractures and lattice defects in the mineral
increases by several orders of magnitude, relative to un-ground minerals. The introduction of strain
lowers the activation energy for the oxidation of the sulphides, and enables leaching under
atmospheric conditions. The rate of leaching is also enhanced, due to the increased mineral surface
area.
Fine grinding also prevents passivation of the leaching mineral by products of the leach reaction.
Passivation occurs when leach products, such as iron oxides and/or elemental sulphur, precipitate
on the surface of the leaching mineral. These precipitates passivate the mineral by preventing the
access of oxidants to the mineral surface.
After the concentrate has been finely ground, the slurry is then leached in agitated vessels with
oxygen to oxidize the sulphide minerals. Leaching is carried out under atmospheric pressure, and
autothermally. Excess heat generated from the oxidation process is removed through
humidification of the vessel off gases.
The average throughput for the Albion Process Plant will be 94,007 t/a of cleaner concentrate, with a
design factor of 15 % applied to the average rate to achieve a design rate of 108,108 t/a. The average
gold and silver throughput will be 127,000 and 131,000 ounces per annum, respectively.
The feed rate to the IsaMill Fine Grinding circuit will be 12.1 t/h, with a design feed rate of 13.9 t/h
of concentrate and a final 80 % passing size of 10 m. The IsaMill specific energy demand will be
798 kW. The available drawn power for an M3000 IsaMill is 1,120 kW and this mill was chosen for
the ultrafine grinding circuit. The completed IsaMill circuit at Ararat is shown in Figure 2.
Finely ground slurry will be pumped to the oxidative leach circuit. The oxidative leach circuit
consists of nine 240 m3 Albion Leach Reactors, each with a live height of 9.4 meters and a diameter
of 5.4 meters. Each reactor is agitated by a 160 kW dual impeller agitator, with oxygen delivered by
a bank six HyperSparge oxygen injection lances in each reactor. The slurry pH will be maintained at
5.0 5.5 in each reactor by limestone slurry dosing.

Figure 2 IsaMill Grinding Circuit GPM Gold Project

The design rate of sulphide oxidation within the oxidative leach will be 1800 kg/h. Under the near
neutral pH conditions employed in the oxidative leach, all of the sulphide will be oxidized to
sulphate, with a design oxygen requirement of 3750 kg/h. The Albion Leach Reactors have all been
designed to achieve an oxygen transfer rate of 4700 kg/h. The design oxygen capture efficiency in
the leach train is 80 %.
The oxygen mass transfer rate for the oxidation of the sulphide minerals is defined by the following
equation (Shuler and Kargi, 2002):
Oxygen Transfer Rate = KLa (Csat C)

(1)

Where:
KL = liquid film mass transfer coefficient for oxygen into the slurry, in units of m.s-1
a = the specific gas surface area, in units of m2.m-3 = m-1
Csat = the solubility of oxygen in the slurry at saturation, in units of g.m-3
C = the steady state oxygen level in the slurry, in units of g.m-3
The KL and a terms are typically combined in the form of a mass transfer coefficient for the
system. The design KLa for the Albion Leach Reactors is 0.12 s-1. Oxygen gas has poor solubility in
water, and so mechanical devices such as agitators and spargers are required to assist the mass
transfer. In the Albion Leach Reactor, oxygen gas is sparged into the vessel using the HyperSparge
supersonic gas injection lances. The HyperSparge oxygen injection system achieves very high
oxygen mass transfer rates at the interface between the supersonic gas jet and the impinging slurry,
reducing the amount of power required from the agitation system.
The agitator drawn power required to achieve the design mass transfer coefficient was determined
using an empirical correlation of the form (Nielsen and Villadsen, 1994):
KLa = A * Us * (Pg/( pSL V))

(2)

Where
A = a constant specific to the ionic strength of the leach solution
Us = the gas superficial velocity in the reactor, in units of m.s-1
Pg = the agitator drawn power under gassed conditions, in units of Watts
pSL = the density of the slurry, in units of kg.m-3
V = the volume of the slurry, in units of m-3
, = dimensionless empirical constants
The A, and parameters used for sizing the agitator were determined based on over 900
laboratory and pilot mass transfer tests. This correlation has been used successfully in the scale up
of all operating Albion Process plants to date. A drawn power requirement of 120 kW per Albion
Leach Reactor was determined using the correlation.

The residence time for the oxidative leaching circuit was designed based on the specific rate
constant for pyrite leaching measured in the batch and continuous leaching testwork. Pyrite
oxidation under near neutral pH conditions is first order (Singer and Stumm, 1970), allowing a
simple scale up. The residence time scale up was based on the method of Henein and Beigler
(Henein & Beigler, 1988). A design residence time of 40 hours was calculated for the oxidative leach
circuit.
Each Albion Leach Reactor was fabricated from lean duplex alloy steel having a diameter of 5460
mm and a live height in the range 9100 8100 mm. The Albion Leach Reactors were supplied in
modular sections for rapid assembly on site. Each Reactor was constructed from 15 panels, each
with a height of approximately 2 m and an arc length of 5.9 m. These panels were all fabricated off
site and imported to the plant site in shipping containers. Baffles, slurry risers, leach tank lids,
agitator support platforms and off gas stacks were all provided as part of the modular XT
equipment supply. Assembly of the oxidative leach train was rapid, with all nine leach reactors and
two slurry storage tanks complete within 8 weeks. The completed oxidative leach train is shown in
Figure 3.
Overflow slurry from the oxidative leaching circuit will gravitate via a slurry sampler to a 10 m
diameter thickener and be thickened to 45 %w/w solids prior to transfer to the CIL circuit.
Thickener overflow will be returned to the leach circuit to compensate for evaporative losses.

Figure 3 Oxidative Leach circuit GPM Gold Project

Limestone for the oxidative leach will be milled to an 80 % passing size of 45 microns in a 100 kW
overflow ball mill operating in closed circuit with cyclones. Cyclone overflow will report to a 200
m3 distribution tank and be circulated through the oxidative leach train by a ring main. Individual
dosing lines will add limestone slurry to each Albion Leach Reactor. Two 60 t/d VPSA oxygen
plants will operate in parallel to provide oxygen to the Albion Process Plant. Oxygen will be
delivered from each plant at a maximum flowrate of 1,745 Nm 3/h, at a purity of 93 % v/v.
The thickened oxidative leach residue and thickened flotation tailings will report to a 100 m 3 mixing
tank and be blended prior to feed to the CIL plant. The CIL Plant will process 137.5 t/h of feed
comprised of oxidized residue and flotation tailings. All six existing CIL tanks will be utilized,

providing a total residence time in the CIL circuit of 41 hours. The CIL Plant is expected to consume
5.3 kg/t of sodium cyanide and 10 kg/t of lime. Carbon levels in the CIL Plant will be 10 15 kg/m3,
with a design carbon loading of 2,500 g/t. Carbon movements will total 7.5 t/d, and the existing dual
AARL elution circuits will be used for carbon processing.
CIL Plant tailing will gravitate to a peroxide based cyanide destruction plant prior to being pumped
to tailings. The tailings will be deposited within the existing tailings impoundment, approximately
6 km from the Ararat plant site.

Project status
The Albion Process plant was provided to the GPM Gold Project as a Lump Sum technology
package by XT. The package included all detailed design, mechanical equipment, electrical,
instrumentation and control equipment, structural steel, flooring, handrails, piping and valves. The
scope of supply includes the fine grinding plant, oxidative leaching and thickening plant and the
supporting limestone, oxygen, flocculent and caustic reagent plants.
Mechanical design was completed in December 2012, with the majority of mechanical equipment
and fabricated components delivered to site by May, 2013. Site civil works were completed in
March, 2013, and construction activities have been underway since April 2013.
At the time of writing, construction activities at the Albion Process plant were largely complete
with only minor piping and electrical cabling remaining. Plant commissioning will begin in
March/April 2014.

REFERENCES
Hourn, M., & Turner, D., (2011) Commercialization of the Albion Process, Proceedings - ALTA Gold
Conference, 2012.
Konstantinov, M., & Grushin, V. (1970) Geologic Position of the Zod-Agduzdag Gold-ore Node in
Transcaucasia. International Geology Review, 12(12), 1447-1453.
Rohner, R., Andreatidis, J. (2010) Bankable Feasibility Report, Geopromining Gold LLC, Zod Upgrade Project
Volume 3, Metallurgy And Processing.
Singer, P., & Stumm, W. (1970) Acidic Mine Drainage: The Rate Determining Step. Science, 167, P.1121-1123.
Nielsen J, & Villadsen J. (1994) Bioreaction Engineering Principles. New York: Plenum Press. P 295342.
Shuler, M., & Kargi, F. (2002) Bioprocess Engineering: Basic Concepts, 2nd Ed. Englewood Cliffs, Nj: PrenticeHall. P 171.
Henein H., & Biegler, L. (1988) Optimization Of Reactor Volumes For Topochemical Reactions, Trans. Inst.
Min. Metall., 97, Pp.C215-223.
Morrell, S. (2010) Estimates of the Throughput of the Ararat Comminution Circuit.

Cerro Negro Process design,

from the design to operation

Cristian Ramos
Oroplata and Goldcorp
Inc, Argentina
Steve Dixon
Goldcorp Inc. Tucson, usa

Cerro Negro Mine is one of the most important gold deposit findings in
Argentina in the last 20 years. Goldcorp Inc. acquired the gold-silver
property with the acquisition of Andean Resources in 2010. Cerro Negro
property is located approximately 60 kilometers southeast of Perito Moreno
in the Province of Santa Cruz, Argentina. The project consists of the
development and construction of five underground mines and one open pit
mine, ore processing facilities and miscellaneous infrastructure and
support facilities. Based on expected throughput of 4,000 t/day @ 92%
availability, annual gold production at the first five years of full production
is expected to average 16,300 kg (525,000 ounces). First gold production is
expected in mid 2014.
The deposits are low-sulfidation, epithermal gold and silver hosted
within quartz veins and associated stockworks with high abrasion and
bond indexes. As previously mentioned the mill facility has been designed
to process 4,000 t/day @ 91% availability of ore by a conventional primary,
secondary and tertiary crushing, a 6.1 m of diameter by 10.1 m of lenght, 7.2
MW ball mill, leaching tanks, counter current decantation, Merrill Crowe
for gold-silver recovery, refinery, cyanide oxidation by INCOTM process and
conventional tailings disposal. The process design has been done
considering the International Cyanide Management Code (ICMI) and
flexibility criteria to support ore changes and future expansions.
Currently the project is in construction phase and in order to achieve
the commercial production quickly the comminution circuit will start as a
primary crushing AG/SAG mill configuration. This allows an opportunity to
evaluate the potential to get lower energy and steel consumption. This will
also be a challenge because there is not much experience and literature
about low aspect AG/SAG mill performance at an industrial scale.
This paper describes the process design evolution, the current
challenges and the adopted strategy for process start up.

INTRODUCTION
Cerro Negro is a gold and silver project located in the Santa Cruz province of the Patagonia region
of Argentina. The property is owned by Oroplata S.A. a wholly owned subsidiary of Goldcorp Inc.
An Australian company called Andean Resources was the owner of the Cerro Negro project until
the year 2010. The Feasibility study developed at that time consisted on underground mining of
Eureka and Bajo Negro Veins and an open pit mining of a veins system called Vein Zone, with the
subsequent ore processing with a rate of 1 850 tonnes per day @ 91% availability to produce Metal
Dore as final product. The process developed consisted of the following stages:

Primary crushing.

SABC circuit.

Conventional leaching.

Counter current decantation.

Tailings filtration.

Cyanide oxidation by Hydrogen peroxide.

Tailing repulping and final disposal on tailing storage facility.

Merrill Crowe

Refinery.

During the 2009 and 2010 exploration seasons, the Marianas and San Marcos orebodies were further
defined and demonstrated the potential of vastly increasing the Cerro Negro Project ore reserves.
In December 2010, Goldcorp purchased Andean Resources with its subsidiary Oroplata. The
increase in project reserves, in terms of tonnage and ore grade, justified increasing the plant design
to 4 000 tonnes per day. Goldcorp commissioned M3 Enginnering & Technology to revise the
feasibility study based the Internationational Cyanide Code Management guidelines, increase the
treatment capacity as well as modify certain aspects of the plant flowsheet including tailings
filtration, cyanide oxidation, and the comminution circuit. The revised design has been based in
increase the flexibility of the operation relative to the variability of the different ores.

METHODOLOGY
Cerro Negro mineral deposits are low sulphidation epithermal with high gold and silver contents
non refractory and free milling. Mineralization occurs in quartz veins within the metals of interest
are located. Gold is usually found as native gold and electrum while silver is mainly found
associated with gold as electrum and sulfosalts.
During the feasibility study an important metallurgical study was made in order to determine the
grindability characteristics, liberation, extraction process apropiated and tails treatment. Those tests
were mainly performed by AMMTEC Laboratories (Australia).
The Cerro Negro ore is mainly composed of quartz results in a high abrasion index of Ai = 0.7 and a
high power consumption with an average Ball Bond Index of 20 kWh/tn. The maximum value
found was about 23 kWh/tn.
Furthermore, the mineralogical studies showed that the grain size of the metals of interest is about
10 to 20 microns and the metallurgical tests have revealed that the most suitable for extraction of
the metals of interest is a conventional leaching process by cyanide. The consumption of cyanide,

free cyanide concentration in solution and gold and silver recoveries are within conventional values
for this kind of process.
Finally with the project acquisition by Goldcorp, M3 Enginnering worked on the process design
based on the previous work by Andean and new testworks done by SGS Chile according to the
following general guidelines:

Treatment of an average 4 000 tonnes per day for 365 days per year, after allowance for
availability while treating all ore types tested in the comminution testwork program

Design availability of 91% with standby equipment in critical areas

Sufficient plant design flexibility for treatment of all ore types as per testwork completed at
design throughput. A wide range of material characteristics requires design criteria for the
most demanding situation, ensuring that design criteria are very conservative.

The plant operation will be likely be extended, beyond current reserves.

International Cyanide Code Management.

Weather conditions: The Patagonia regin of Argentina is characterized with low

temperatures (-15C), snow in the winter and wind velocities up to 33 m/s (120 km/h).

The selection of these parameters is discussed in detail below.

Throughput and Availability

An overall plant availability of 91% was selected. Benchmarking indicates that welloperated plants with abrasive ores similar to those tested have consistently achieved these
levels of overall plant availability.
The throughput selected is a function of the anticipated mining production schedule. A
sustainable plant throughput of 183 dry metric tonnes per hour (t/h) was selected based on
91% availability, resulting in an average of 4 000 t/d processed.

Processing Strategy
The process design is based on treating the different ore types tested individually at the
design throughput rates. Ore hardness parameters were maintained as the previous study,
selected based on the 75th percentilethat is, 75% of the ore to be processed is expected to
be similar in hardness or softer than the ore hardness parameters used for design. The
revised Mill design the 95% value found was selected for sizing.

Head Grade
The plant is designed to treat high Gold and Silver grade with a ratio Gold/silver around 10
to 40. The design grades were for Gold content about 26 gr/t and a Silver content about 250
gr/t.

The overall approach was to provide a flexible process plant flowsheet that could handle the
variability in the metallurgical performance of the orebodies at the nameplate capacity.

Process description
The unit operations selected for Cerro Negro are well known in precious metal recovery. The
flowsheet incorporates the following major process operations:

Primary crushing with the product directly feeding a secondary stage crushing with closed
tertiary fine crushing plant designed that full nameplate capacity is accomplished in 12

hours operational time. This results in sufficient maintenance time to handle abrasive ores.
The circuit also has been designed to bypass the secondary or terciary crushing in the
operation of the mill in SAG or AG.

Fine crushed material feeding to a stockpile with 4000 t of ore as live capacity and 16000
tonne of ore as dead capacity. Two Apron feeders are installed to reclaim the ore and feed
the mill. Each one has the nameplate capacity which allows feed independiently in case of
maintenance or fail of one of them.

Ball mill grinding. As previously mentioned the mill can be performed as AG/SAG or ball
mill by changing the liners and discharge device. The ore is ground using barren solution.

Gravity concentration: The gravity concentration circuit is operated with cyanide. This
results in the simultaneous recovery and leaching of free gold.

The cyclone overflow reports to the grinding thickener. The overflow from the lead
thickener of the metal recovery counter-current decantation reports to the grinding
thickener. The grinding thickener underflow reports to leaching. The grinding thickener
overflow reports to the clarifying.

The clarifying thickener was selected to give the plant the flexibility to process ores with
various amounts of clays. The majority of the ore contains low clay. But there are locations
that have higher clay content. This unit will decrease the operating cost of the pressure
clarifiers.

Leaching: The ground slurry is leached by cyanide in 5 series tanks, each one is designed in
order to be bypassed in case of maintenance or operational issues, if it is needed the design
allows the Oxygen injection instead of Air injection. An Oxygen plant would be installed in
that case. Aditional room has been considered in order to install additional two tanks. This
would increase the residence time by 24 hours at 4000 t/d.

Counter-current decantation for metal recovery. Three high rate thickeners are being
installed. Each one can be bypassed in case of maintenance or operational issues. This is
followed by the cyanide recovery thickeners.

Counter current decantation for cyanide recovery. Three high rate thickeners are being
installed. Each one can be bypassed in case of maintenance or operational issues.

Pregnant solution clarification is performed using four pressure filters to preduce a solution
with low suspended solids. This circuit has been designed to have 3 filters operating with 1
stand by.

Merrill Crowe (Zn precipitation): a conventional circuit has been designed, a higher grade
than the ROM grade has been considered in order to oversize the equipment and provide
robustness in the face of head grade changes.

Mercury Retort Ovens. Four electrical mercury retorts ovens are being installed in order to
support changes in the mercury content and assure their maximum recovery.

Fire refining. One reverberatory furnace is being installed, that technology allows produce
large volumes of Dore with one pouring. In steady state 3 pourings per week are expected.

Cyanide oxidation by INCO Process. Two tanks have been installed with additional room
to install one more. The circuit allows the operation as series or parallel. Each tank has a
bypass for maintenance or operational issues.

Fresh and reclaim water supply. Water is provided by underground mine dewatering and
water wells, so there is an exccess of water which allows operate at high capacity without
reclaim water from TSF if it is needed.

Reagent preparation and distribution.

The overall process plant flowsheet is shown in Figure 1 and the Plant general arrangement in
Figure 2.

Figure 1 Cerro Negro process flowsheet

Figure 2 Cerro Negro plant general arrangement

Crushing and conveying

ROM rock will be dumped from haul trucks or a front-end loader through a square-grid grizzly
into a dump hopper. The primary crusher is a jaw crusher design to operate with a closed side
setting of 125 mm and produce a product with 80% passing (P80) of 115 mm in open circuit with a
vibratory grizzly.
The secondary section is a cone crusher in open circuit and a diverter gate which would allow
bypassing of material for maintenance in the secondary crushing section. Crushed product at 80%
passing46 mm would pass to the tertiary crushing circuit.
Tertiary crushed product is in closed circuit, combining with secondary product returning back to
the tertiary screen. Two cone crushers with similar dimensions and power at the secondary crusher
and a short head profile will be installed. The final product from crushing circuit will be 80%
passing 13 mm.
The secondary screen undersize material combines with tertiary screen undersize on the stockpile
feed conveyor, and passes to the fine ore stockpile. The stockpile will be covered to protect from the
wind and will have a live capacity of 4 000 t or approximately 24 hours of mill feed.
The ore conveying will be totally covered to protect the operators from the weather conditions. Two
dust collectors will be installed to assure a healthy enviroment for the operators.
The circuit has been designed to operate 12 hours per day with a 75% of availability during that
period of time. The crushers and screens are oversized in order to provide robustness at the circuit.

Grinding and classification

The primary grinding circuit will be designed to process an average of 4 000 t/day at 91% utilization
on a 24 h/day, 365 days/year basis. The grinding circuit consists of a single stage ball mill operated
in closed circuit with hydrocyclones. Grinding will occur with cyanide present in the mill. Ore will
be ground to final product size about a P80 = 65 micron.
A portion of the grinding circuit circulating load will be treated in a single high capacity,
continuous centrifugal gravity concentrator with the concentrate and the tail going back to the
cyclone feed sump. The plant had been designed for the future installation of an intensive leach
reactor (ILR). But the use of cyanide in the grind results in the gravity concentrator being the ILR.
Further surveys at full plant scale will allow to determine the technology appropriated downstream
of the centrifugal concentrator.
The configuration 3 stage crushing-ball milling has been selected because it had lower capital cost
and operating cost compared to a SABC circuit.
Grinding will be performed in a 6.1 m diameter by 10.1m (9.54 m effective grinding length) long
ball mill with a 7.25 MW motor operated in closed circuit with hydrocyclones. The ball mill will
discharge over a trommel screen. The undersized material will fall through the trommel into the
primary cyclone feed pump box where it is combined with tails from the gravity concentrator. Ball
chips will be rejected out the end of the trommel into a tote. A portion of the hydrocyclone
underflow will flow by gravity to a gravity concentration circuit. The remainder of the underflow
will return to the ball mill. Hydrocyclone overflow (final grinding circuit product) will flow by
gravity to a single deck, horizontal vibrating trash screen for removal of tramp material. Trash
screen oversize will discharge into a super sack to be periodically removed for disposal. The
equipment has been sized at a Ball Bond index found of 23 kWh/tn . Current testwork shows an
average ball Bond index of 18 kWh/tn so the mill would have a higher capacity than the designed.
Based on the flexible flowsheet the comminuion circuit can be configured as Primary
crushing/AG/SAG mill or Conventional crushing/Ball milling so that means an opportunity has
which can allow reduce the operative cost and improve the construction schedule.

During the early months of start up the ball mill will be operated as a single stage low aspect ratio
AG/SAG mill with a grate discharge, and with only primary crushed feed. The mill has been
designed with a feed chute to accommodate the SAG operation and a shell drilled to install
AG/SAG or Ball mill liners. In addition, the mill will be equipped with a variable speed drive to
control the optimum speed to prevent overgrinding, high liner wear, and scats formation. This will
allow the mill to be operated with a low ball charge and still obtain the desired product grind size.
The mill may be changed from a grate discharge to a typical overflow discharge mill to
accommodate AG/SAG or Ball operation.
There are operations that use low aspect AG/SAG mill successfully. South African mines have
operated single stage ROM feed AG mills. The Cannington mine in Queensland operates a milling
circuit of AG/vertimill. There are few metallurgical research papers of low aspect AG or SAG
milling. Vendors indicated it is feasible to obtain the required grinding product with a low aspect
AG/SAG mill, with an improved power efficiency and lower steel consumption than the 3 crushing
stages-ball mill circuit. Metallurgical researching by bench and pilot scale testwork is ongoing to
study the ore grindability and circuit performance and obtain a Grinding Circuit model by JK
SimMet.

There is also extra room at the grinding area to install a Vertimill and a pebble crushing circuit for
plant future plant expansion.
The cyclone overflow reports to the grinding thickener. The overflow from the lead thickener of the
metal recovery counter-current decantation reports to the grinding thickener. The grinding
thickener underflow reports to leaching. The grinding thickener overflow reports to the clarifying
thickener.

Leaching
Leaching is conventional using cyanide. Leaching of precious metals by cyanide occurs in a series of
five agitated leach tanks with compresed air injection to provide enough residence time to assure
the maximum gold and silver recoveries at the grinding size product.
The circuit has the flexibility to bypass tanks in case of maintenance or operational issues. There is
extra room to allow the installation of 2 more tanks. Oxygen injection has been considered during
the tanks design so in case of need Oxygen generation plant can be installed in order to improve the
leaching kinetics.
The product from leaching goes to the CCD stage.

Counter Current Decantation (CCD)

Previously tailing filtration circuit was considered by Andean for cyanide recovery. Filtration
testwork was done in order to study the rheology and filtration characteristics for the Cerro Negro
ore. The results obtained shown a poor filtration capability for the slurry with a filtration rate lower
than 80 kg/m2 hr. That low rate means the installation of large filtration plant for the tails treatment
with the possibility to be a bottleneck for the process. That option was dismissed and a
conventional CCD circuit was chosen for cyanide recovery.
The CCD circuit can be considered in two sections. The first section of 3 stages serves for the
recovery of precious metals leached into solution via three stages counter current thickener
washing. The concentration of gold and silver in the solutions are less than 2% of the soluble values
at the end of 3 stages.
The second section is three stages of cyanide wash thickeners to reduce and recover soluble cyanide
before the oxidation step. Approximately the 70% of cyanide is recovered in this section of the CCD
circuit. The product of the cyanide wash thickeners is sent to the Cyanide oxidation system.
The two step approach allows operational flexibility to apply at the third CCD tank the lower
cyanide grade solution available using either solution from the barren solution tank or directly
from reclaim of the first cyanide wash CCD.
Barren solution from the Merrill Crowe circuit is used as wash solution in the third CCD thickener.
The wash solution will flow counter current to the solids flow, increasing in precious metal
concentration as it continues to the first CCD thickener and lower cyanide levels in the cyanide
recovery CCDs.
The pregnant solution from the first CCD thickener will report to the grinding thickener. The CCD
area is fully bunded to contain any spillage and is equipped with vertical spindle sump pumps to
facilitate clean up. Large volumes of spillage would overflow this bound area to a lined pond to
assure capture in the event of a catastrophic failure.
Extra room has been designed in order to install a seventh thickener if it is needed.

Solution clarification and Merrill Crowe

The solution from grinding thickener is clarified by a large diameter clarification thickener and four
(three operating and one standby) self-cleaning pressure leaf clarifier filters in order to remove
suspend solids prior to the Merrill Crowe process.
Merrill Crowe process is used to precipitate the precious metals by deaeration of the solution
followed by zinc adition. The zinc precipitate is filtered in four (three operating and one standby)
plate and frame filters.

Cyanide oxidation of tailings

Metallurgical testwork has been performed in order to study the most efficient technology to be
applied at the tailings treatment. The INCO process was selected because of high efficiency of
cyanide oxidation.
The underflow from the last cyanide recovery thickener is sent to the Cyanide oxidation stage. In
the cyanide oxidation tanks, weak acid dissociable (WAD) residual cyanide will be oxidized to the
relatively non-toxic form of cyanate by the INCOTM process using sodium metabisulfite and air,
with copper sulfate as a catalyst. Milk of lime is added to maintain a slurry pH in the range of 8.0 to
9.0. The more stable iron cyanides are removed from solution as an insoluble ferrocyanide
precipitate. The cyanide levels in the tailing slurry are thereby reduced to an environmentally
acceptable level, less than 50 ppm WAD CN, before being pumping to the lined tailings storage
facility (TSF).

Tailings pumping and solution recovery

Tailings are pumped by three stage centrifugal pumping systems to the tailings storage facility
(TSF). There is one operating and one standby pump systems. Excess water on the TSF is recovered
by a pump and returned to the process plant reclaim water circuit for use in the cyanide wash
circuit.

Gold refinery
Filter cake harvested from the precipitate filters is loaded by fork lift into a mercury retort. The
recovered mercury is collected and handled separately from the gold and silver production. Once
moisture and mercury are removed, the precipitate is mixed with smelting fluxes and charged to
the 2 000 kg capacity reverb furnace. The precipitate and flux are smelted and the dore is poured
into ingots.
Slag from the smelt is crushed and precious metals entrained in the slag recovered by a vibrating
screen. The screen undersize is returned to the milling circuit.
Fume extraction equipment have been provided to remove emissions from the furnace. Ventilators
will be provided to ensure sufficient ventilation of the gold room.

Major equipment description

The TABLE 1 shows a list and most important dimensions for Cerro Negro equipment.

Table 1 Major equipment list

Area

Equipment name

Description

Power (W)

Primary
Crushing

Primary Crusher

Jaw Crusher 0.91m x 1.2m opening Metso C125 Closed

setting = 125mm

150 000

Fines
Crushing

Secondary Crusher

Standard cone crusher XL400 Raptor Cone Crusher

Closed side setting from 0.030 to 0.050 m, capacity 400550 t/hr, eccentric speed 315-380 rpm.

300 000

Fines
Crushing

Terciary Crushers

2 Short head cone crusher XL400 Raptor Cone Crusher

Closed side setting from 0.010 to 0.030 m, capacity 165275 t/hr and eccentric speed 315-380 rpm.

300 000

Grinding

Mill

6.1 m diameter by 10.1 m EGL (estimated)

7 250 000

Grinding

Hydrocyclones

(12) gMAX15 cyclones

Grinding

Grinding
Thickener

25 m diameter, Delkor Model SR 130-4, 400 kN

maximum operating torque, with four hydraulically
driven planetary gearboxes and pinions. 0.6 m additional
sidewall height

15 000

Leaching

Leaching tank

5 tanks; 16.3 m diameter x 16.9 m; 3526 m3 total capacity,

3200 m3 operating capacity; open top, vertical baffles,
upcomer discharge, steel top hat for air dispersion, and
nozzles for future oxygen injection sparging system

112 500

CCD

CCD Thickeners

6 thickeneres; 25 m diameter, Delkor Model SR 130-4, 400

kN maximum operating torque, with four hydraulically
driven planetary gearboxes and pinions. 0.6 m additional
sidewall heigh

15 000

Merrill
Crowe

Clarification
thickener

1 thickener, 30 m diameter, Delkor Model GB230 H, 200

kNm maximum operating torque, with hydraulically
driven planetary gearboxesvand pinions. 0.6 m
additional sidewall height.

11 000

Merrill
Crowe

Clarifier filters

4 filters MET-POR, Model MCRO-2600-72-3-WCD, 530

kPa operating pressure, vessel 1.8 m diameter x 3.2 m
length, 60 AISI 316 leaves with polypropylene, total area
83.6 m2 per filter.

Merrill
Crowe

Deaeration tower

4 m diameter x 6 m straight wall ASME section VIII

certified (Pressure Vessel)

Refinery

Precipitate filters

4 filters. Total capacity, 2.8 m3, total capacity 242 m2,

plate size 1.2 m x 1.2 m, 101 plates

Refinery

Retort Oven

4 retort ovens Summit Valley model 40MR1E, 1.1 m3

batch capacity, 815 C max operating temperature.

Refinery

Reverbetory
furnace

Fuel oil fired tilting reverberatory furnace, w/

combustion air blower and controls, Cutler model 4000
US, 2 000 kg of capacity.

Cyanide
oxidation

Cyanide oxidation
tanks

2 tanks; 7 m diameter x 8 m high; 308 m3 total capacity,

283 m3 operating capacity; steel top hat for air dispersion
and nozzles for future oxygen injection sparging system

75 000

CONCLUSION
The start up for the Process Plant is expected for mid June 2 014, the expected production for the
first year is between 4 000 kg (130 000 ounces) for Gold and 56 000 kg (1 800 000 ounces) for Silver.
The current results from metallurgical testworks and simulations project attaining the nameplate
capacity before the end of the year with the AG/SAG configuration and good gold and silver
recoveries. If there are issues with milling capacity, the installation of the crushing circuit will be
completed.
The flexibility of the design allows be confident about to achieve the nameplate capacity and the
production targets commited. Naturally the start up stage trends to be complex and stressful
however with a flexible design a smooth ramp up is expected.

ACKNOWLEDGEMENTS
The authors kindly acknowledge to Cerro Negro and Goldcorp management for their support.

NOMENCLATURE
CCD
TSF

counter current decantation

tailing storage facility

WAD CN weak acid dissociable cyanide

REFERENCES
M3 Engineering & Construction corp. May (2011), Cerro Negro Feasibility study update.
AMMTEC Ltd., Balcatta, Western Australia, , October (2008), Metallurgical Testwork conducted upon Two (2)
Samples of Gold Ore from the Cerro Negro Gold Project for Andean Resources Limited., Report No.
A114812.Report No A11942
SGS Minerals Services, Santiago, Chile, July, (2010), Pruebas Metalrgicas de Lixiviacin Y Molienda Informe
Preliminar Prepared for Oroplata S.A. Project 4416.
SGS Minerals Services, Santiago, Chile, September (2010), Leaching and Grinding Metallurgical Tests for San
Marcos Ore Final Report Prepared for Oroplata S.A. Project 4416.
SGS Minerals Services, Santiago, Chile, June (2011), Filtration pilot tests with cyanide leaching slurry tails. Final
Report Prepared for Oroplata S.A. Project 4522.
Steve Morell, Fremantle, WA, March (2007), The Effect of Aspect Ratio on the Grinding Efficiency of Open and
Closed Circuit AG/SAG Mills Ninth Mill operators conference.

chapter 6
Efficient water use and
water recycling

Improved models for predicting

heap leaching hydrodynamics

Saman Ilankoon and

Stephen Neethling
Royal School of Mines,
Imperial College London,
United Kingdom

Heap leaching is one of the most important hydrometallurgical

techniques for the extraction of valuable metals from low grade ores,
despite the low recoveries when compared to conventional mineral
extraction methods such as froth flotation followed by smelting. There are
a number of reasons for the lower recoveries including both chemical and
hydrodynamic effects. This paper will cover the development of improved
models for predicting the hydrodynamics of the complex unsaturated
flow within heap leaching systems.
The authors have shown that the liquid held within the particles has a
very different behavior to that between the particles and thus, in order to
successfully model these systems the effect of inter- and intra-particle
porosity on heap flow behavior needs to be separated. The model form was
developed theoretically and verified using column leaching liquid flow
experiments. This model was initially validated for non-porous spherical
particles and narrowly sizes copper ore particles of about 5% porosity.
In this paper the inter-particle flow model is validated for complex
particle size distributions that are representative of those found
industrially. In addition, liquid flow experiments were performed in a
pseudo 2D packed bed to investigate the horizontal liquid distribution
from, for instance, dripper points. The aim of this is to assess the relative
importance of capillarity compared to other dispersive effects in both the
inter and intra-particle spaces.

INTRODUCTION
Heaps exhibit complex unsaturated flow behaviour through the semi-consolidated porous ore
particles of wide size distribution that make up the heap. The liquid flow behaviour in industrial
heap leaching is an important aspect of the overall performance as it is the major driver in the mass
transport of leaching reagent to the ore surface as well as dissolved metal species out of the heap.
Despite this importance to the overall recovery, there have only been a very limited number of
systematic studies into the hydrodynamics of heaps (Murr, Schlitt & Cathles, 1981; Yusuf, 1984; de
Andrade Lima, 2006).
A typical heap has particles that are mainly in the size range of millimetres to a few centimetres
(Bartlett, 1992). This is significant as the Bond number goes from significantly below to significantly
above 1 over this range of particle sizes and the flow is thus in a transition region between gravity
and capillary dominated flow (Ilankoon & Neethling, 2013). Moreover, the flow characteristics
depend on factors such as the construction procedures of the heap, the application of the leach
solution, the size of the heap and the size distribution and the porosity of the rock particles (Roman,
1977; Murr, Schlitt & Cathles, 1981; Yusuf, 1984).
The applied leaching solution occupies space both between the particles as well as within the pore
spaces in the individual ore particles. Thus, the porosity of the heap has two distinct length scales,
namely that of the channels between the particles (i.e. interstitial space) and within the particles (i.e.
intra-particle space). The interstitial space will typically have a length scale of the order of
millimetres, and the typical length scale of the intra-particle space will be tens of microns and
smaller. As described above, the Bond number will be order of one for the fluid flow between the
particles. However, the pores within the particles will have Bond numbers that are many orders of
magnitude less than one, indicating capillary dominated flow. This distinct separation of length
scales means that the holdup within the particles will not have the same effect on liquid flow as the
holdup between the particles. New heap flow models should thus take into account of this
separation of length scales. In addition, these systems exhibit liquid holdup hysteresis, with liquid
holdup depending not only the current flow, but also on the history of liquid addition (Ilankoon &
Neethling, 2013).
As well as the vertical motion of the liquid, the horizontal spread of the liquid also has important
implications for leaching performance. Leaching solution distribution on the ore heaps has been
developed based on the technologies and materials associated with agricultural irrigation. In the
case of the drip emitter solution application the drip emitter lines are placed such that each drip
emitter covers an area of about one square meter (Bartlett, 1992). An understanding of leaching
solution spread from this initial liquid addition point is important in determining the wetted region
associated with each of these drippers.

METHODOLOGY
One of the main objectives of this study is to experimentally investigate and model the effect of the
particle porosity on the overall heap flow behaviour. This is done by comparing a bed consisting of
non-porous glass beads with a system consisting of narrowly sized copper ore particles and the
same system having realistic ore particle size distributions. In addition, the effect of capillary and
dispersive effects on liquid distribution will be studied using a pseudo 2-D packed bed filled with
both glass beads and ore particles.

Liquid holdup measurements were performed using circular Perspex columns of 243 mm diameter
and height of 300 mm and 500 mm (referred to as the 1-D system) by suspending the column from a
high precision load cell. Two types of particles were used in this study: a model system consisting
of randomly packed mono-dispersed non-porous glass spheres of 2, 10, 14 and 18 mm and an ore
system (i.e. porous system) with nine sets of narrowly sized copper ore particles in the range of 4-45
mm (see Ilankoon & Neethling, 2013 for more details including figures of experimental systems). In
addition, sets of experiments were conducted using mixtures of different narrow sized particles in
order to determine the validity of the inter-particle flow model for more realistic, but well defined
particle size distributions. These distributions included four different bi-dispersed mixtures made
up of 50% by weight from each size fraction with the following combinations: 8-11.2 + 16-20 mm,
11.2-13.2 + 20-26.5 mm, 13.2-16 + 26.5-31.5 mm, 16-20 + 31.5-37.5 mm. In addition, 6 wider size
distributions based on the Gaudin Schumann (GS) distribution were investigated. The details are
summarised in Table 1 and the size distributions cover typical of those found in industrial heap
leaching.
Table 1 Making realistic size distributions using GS distribution
Size range (mm)

Gradient of GS plot

2-31.5

0.5

1.0

1.5

2-45

1.0

1.5

The top of the packed bed was wetted using a novel liquid distributor, which consists of 32
individual drip points with equal flow from each point (see Ilankoon & Neethling, 2012a for more
details). This was mounted over the packed bed as a separate unit so as not to adversely affect the
gravimetric liquid holdup measurements (Ilankoon & Neethling, 2012a). The superficial liquid flow
rates investigated were 0.0075, 0.015, 0.03, 0.06 and 0.12 mm/s, which corresponds to 27-432
kg/m2/h. The lower values are within the typical solution application rates found in industrial
leaching and column leaching studies (de Andrade Lima, 2006; Roman, Benner & Becker, 1974). A
flow rate range covering at least an order of magnitude is required to validate the model as it takes
a power law form and hence the range of studied flow rates extends well beyond those encountered
in heap leaching.
The horizontal liquid flow behaviour was investigated using a Perspex pseudo 2D column that 800
mm long and 600 mm high, but only 100 mm deep. Figure 1 shows the column suspended from
two load cells so that the average liquid content could be measured. As with the 1D column,
experiments have been conducted in this column using both glass beads and ore particles with both
narrow size distributions and more realistic size distributions.
Liquid was added to the top of the particles across the narrow depth of the column, but at a single
horizontal point. This was done to mimic the behaviour around a single dripper. Liquid flow rates
of 0.6, 1.9, 2.5, 4.2, 10.6 and 16.2 L/h were used, which correspond to addition rates ranging from 7.5
L/m2/h to 200 L/m2/h and thus covers the typical heap leaching addition range, as well as extending
well above it. The liquid distribution at the bottom of the column was measured via a series of 25
liquid flow channels spaced 30 mm apart, each with its own liquid collection port (see Figure 1
right for locations). In addition to the measurement of the average liquid content via the load cells
and flow out of the bottom using liquid flow channels, the liquid motion was filmed so that the
spread could be visually assessed.

Load cell

Liquid addition

Liquid
distributor

100 mm

600 mm

Figure 1 2-D column rig filled with non-porous glass beads and ore particles

INTER-PARTICLE FLOW BEHAVIOUR WITH THE 1-D SYSTEM

Steady state and residual liquid holdup measurements with the model system
The two liquid holdup measurements that describe the flow behaviour are the steady state liquid
holdup and the residual liquid holdup. Liquid was introduced into the 1-D model system at the
lowest flow rate (i.e. 0.0075 mm/s) to measure the steady state liquid holdup, which takes about an
hour to achieve in the glass bead system. Then the packed was allowed to drain until no more
liquid dripped from the column, which generally occurred after about 15 minutes for the glass
beads, with this remaining liquid content referred to as the residual liquid holdup. The liquid flow
was increased to the next rate (i.e. 0.015 mm/s) and the same procedure was repeated. This was
done up to the maximum flow rate studied and then back down to the original flow rate so that the
hysteresis could be quantified and modelled. This resulted in a pair of steady state and residual
holdups for each of these flow rates. The full details of this can be found in Ilankoon & Neethling
(2012a).

Inter-particle flow model

The authors have developed a theoretical relationship to describe the inter-particle flow behaviour
initially for flow through this glass bead system (Ilankoon & Neethling, 2012a, 2012b), though this
was subsequently modified and extended in order to be applicable to porous ore particles (Ilankoon
& Neethling, 2013).
It was found out that rivulet flow was the dominant flow feature through the particles at the low
superficial velocities investigated in this work (Ilankoon, 2012). Since the residual liquid holdup ( )
is proportional to number of rivulets per cross-sectional area (
), the two main model
parameters in the inter-particle flow model that are used to characterise the rivulet flow are the
relative flow rate and the relative holdup. The relative flow rate ( ) is proportional to the average

flow per rivulet (

) and the relative holdup ( ) is proportional to the liquid holdup per
rivulet (
), where
is the superficial liquid velocity and
is the steady state liquid
holdup:

(1)

(2)

The inter-particle flow model (Equation 3) gives the relationship between and , assuming that
the velocity of a rivulet is given by the Darcy-Weisbach law and that, while the size of the rivulet
changes with flow rate, the cross-sectional shape of the rivulet is reasonably constant (Ilankoon &
Neethling, 2012a).

where

(3)

is a dimensionless drag coefficient,

is the average cross-sectional area of the residual liquid
in a rivulet, is the density and is the viscosity of the liquid. The pre-factor, , depends on the
Bond number (i.e. particle size), shape of the liquid rivulet (i.e. contact angle), its tortuosity (which
will depend on the particle packing) and the rheological properties of the fluid. The model
describes a power law relationship with an exponent of two between the main model parameters
for the model system, which was indeed found experimentally (see Figure 2).

Figure 2 The relationship between and

for a glass bead system and the ore system. The slope of the
solid line represents the squared relationship between the two parameters

Modifications of liquid holdup measurement methods for the ore system

Once the model had been developed based on the model glass bead system, it needs to be validated
for real ore systems. Initial experiments were performed with dry ore particles, but if was found out
that the liquid holdup measurement procedure used for the glass beads did not work. This was
because liquid drainage occurs from both within the ore particles and around the ore particles,
while the previous model was only applicable for the liquid held between the particles. A new
procedure to separate the contribution of these two regions to the overall liquid behaviour was thus
developed. The procedure is described in detail in Ilankoon & Neethling (2013). This procedure
starts with the column being filled with saturated ore particles (this is achieved by immersing the
particles in water for 3 days and then screening to remove external water). Once the particles are
placed in the column liquid addition is started immediately. Since capillarity is strong within the
particles, they will remain near saturated as long as liquid is being added to the system. This allows
us to measure the external steady state liquid holdup (generally it takes about an hour to reach
steady state compared to many hours with starting with dry ore particles).

Figure 3 The relationship between and

for realistic ore particle size distributions. The slope of the
solid line represents the squared relationship between two parameters and the results of 10 mm glass beads
are also shown

After achieving steady state, the flow into the column is stopped. It was observed that a large
amount of liquid drainage occurs in first ten minutes followed by very slow drainage for many
hours compared to the non-porous system which stops dripping after about 15 minutes with most
of the liquid lost in first few minutes. It can be assumed that this slow drainage is associated with
the drainage of liquid from within the ore particles, especially as it was found to be independent of

the steady state saturation and initial liquid addition rate into the system (Ilankoon, 2012). As this
long term drainage rate is associated with liquid lost from the particles, the external residual liquid
holdup at a particular flow rate can be estimated by adding the liquid flowing out of the particles
back onto the measured liquid content, which is the difference between the total liquid content and
the liquid content of the initially saturated particles (Ilankoon & Neethling, 2013).
Based on the calculated external steady state and residual liquid contents within the ore system, a
similar analysis to that for the glass beads can be performed. Figure 2 shows that real ore results
also follow a power law relationship with an exponent of 2, indicating that the model form
developed is also applicable to real ore particles, though only if the effects of the internally and
externally held liquid are separated. The pre-factor in the relationship for the model system is
approximately a factor of five greater than that for the ore system. This is probably due to the much
lower tortuosity of the flow paths in a system of spherical particles compared to that of the flow
between the angular ore particles.
While the results in Figure 2 are all for narrowly sized particles, the experiments have been
repeated for both bi-dispersed and more realistic poly-dispersed size distributions for the ore
particles. Figure 3 shows these results compared to some of the previous results. It can be seen that,
while the size distribution has an effect on the pre-factor, the model form remains appropriate.

RESULTS OF 2-D COLUMN EXPERIMENTS

Liquid holdup hysteresis
Liquid holdup hysteresis was observed in the 1-D packed bed system, with the liquid holdup
depending not only on the current liquid addition rate, but also on the liquid addition history
(Ilankoon & Neethling, 2012a). It was demonstrated that the main cause of this hysteresis was a
change in the number of flow paths within the system as the liquid addition was varied. New flow
paths are created as the liquid addition is increased, but do not disappear as the liquid addition is
decreased (Ilankoon & Neethling, 2012a).
A similar study has now been conducted to investigate hysteresis when using a point liquid source
within the 2D column. In Figure 4 it can be seen that, as we would expect from our 1D column
results, there is strong evidence of hysteresis within both the real ore and glass bead systems. The
liquid holdup with the ore is also much higher than that within similar sized glass beads, which is
due to a combination of the porosity of the ore particles holding liquid and the higher tortuosity of
the flow paths through the angular particles (the wettability/contact angle for ore particles and
glass beads are not similar, with both having low contact angles/high wettability)

Breakthrough times and steady state

Table 2 compares the breakthrough times (time taken for first drip of liquid to emerge from the
bottom of the column) for the two systems with different particle sizes and liquid addition rates. It
was found out that the breakthrough times for the porous systems are much higher than that for
the glass beads. This is due to the higher steady state liquid holdup in the ore system, as well as the
fact that liquid gets sucked into the slightly porous ore particles, thus retarding the liquid motion
down the column.

Figure 4 Liquid holdup hysteresis behaviour for the 2-D system with both 18 mm glass beads (bottom) and 2026.5 mm ore particles (top) starting from a dry bed

Table 2 Comparison of breakthrough times for 2-D system with non-porous glass beads and ore
particles at 0.6 L/h liquid addition along the centre of the bed
Size range (mm)

Particle Type

Breakthrough time (s)

Glass Beads

504

10

Glass Beads

198

14

Glass Beads

120

18

Glass Beads

107

20-26.5

Ore

1020

31-37.5

Ore

446

20-26.5 (at 4.2 L/h)

Ore

103

Horizontal liquid distribution

By measuring the horizontal liquid distribution at the bottom of the 2-D column at steady state the
extent of spread over the 600 mm height from the point source at the top can be measured. While
the vertical flow is gravity driven, the horizontal flow is due to a combination of dispersive and
capillarity effects, with the relative magnitude of these effects depending upon the channel size
distribution of the packed bed. Figure 5 shows the percentage liquid flowing out of the bed that
passes through each of the outlet channels. Channel 0 is in the centre of the column horizontally
and is directly below the liquid addition point. Figure 5 shows the comparison between the glass

beads and the ore at two different flow rates. This data shows that the extent of the spread is quite
similar in the glass beads and ore system, with both exhibiting a slight increase in the width of the
spread with increasing flow rate. This increase in the spread is actually quite modest given that the
flow rate has been increased by a factor of more than 25. The spread of these actively flowing flow
paths is very narrow as the width of these peaks is between about 100 mm and 200 mm over a drop
height of 600 mm. The spread becomes wider as the particle size distribution becomes wider.
45

90
70
60

Liquid flow out (%)

Liquid flow out (%)

80

18 mm
glass beads
20-26.5 mm
ore

50
40
30
20

40
35
30
25

18 mm
glass beads
20-26.5 mm
ore

20
15
10
5

10

0
-6 -5 -4 -3 -2 -1 0

Liquid outlet number

-6 -5 -4 -3 -2 -1 0 1 2 3 4
Liquid outlet number

Figure 5 Liquid distribution behaviour of the 2-D system with 18 mm glass beads and 20-26.5 mm ore
particles starting from a dry bed. Left- liquid addition rate of 0.6 L/h, right- 16.2 L/h

Figure 6 Wetted area of the 2-D system at different time intervals with 20-26.5 mm ore particles starting from a
dry bed: Liquid addition rate was 4.2 L/h

While the external flow paths and liquid distribution out of the bottom of the column establishes
itself very quickly, in the ore system the particles away from these actively flowing regions only
wet very slowly (see Figure 6), with the entire bed taking over 2 weeks to wet (eg. at a flow rate of
4.2 L/h the wetted area after 2 weeks is about 61%). While the bed becomes wetter over this period
(i.e. wetted area increases), this is not associated with any significant change in the liquid flow
distribution out of the bottom of the column.

CONCLUSIONS
The flow experiments were used to formulate an inter-particle flow model and it was validated for
both non-porous glass beads and slightly porous ore particles, which include both narrow size
distributions of ore and realistic particle size distributions. Even though the overall flow behaviour
between the non-porous and porous systems is significantly different, the inter-particle flow
behaviour between the two systems is quite similar, which indicates the importance of separating
the liquid held within the ore particles and the liquid held around the ore particles. The reason is
due to the very different length scales of the channels involved. Thus, heap fluid flow modelling
based on a single holdup is problematic as the relationship between the holdup with in the particles
and portion of flow within the particles versus the same relationship for the holdup outside the
particles is dramatically different.
In addition, a set of 2-D liquid flow experiments were discussed. It was shown that the spread of
the actively flowing paths is comparatively narrow for both the glass bead and ore systems, though
in the ore system capillarity eventually wets the bed over a period of a few days, though this not
accompanied by any significant change in the actively flowing paths.

ACKNOWLEDGEMENTS
This study was performed in the Rio Tinto Centre for Advanced Mineral Recovery at Imperial
College London. The authors gratefully acknowledge Rio Tinto for their financial support for this
project.

REFERENCES
Bartlett, R. W. (1992) Solution mining: leaching and fluid recovery of materials, Gordan and Breach Science
Publishers.
de Andrade Lima, L.R.P. (2006) Liquid axial dispersion and holdup in column leaching, Minerals Engineering,
vol. 19, pp. 37-47.
Ilankoon, I.M.S.K. (2012) Hydrodynamics of unsaturated particle beds pertaining to heap leaching , PhD
thesis, Imperial College London, United Kingdom.
Ilankoon, I.M.S.K. & Neethling, S.J. (2012a) Hysteresis in unsaturated flow in packed beds and heaps, vol. 35,
pp. 1-8.
Ilankoon, I.M.S.K. & Neethling, S.J. (2012b) Effect of particle porosity on liquid drainage in heaps Proc.
Hydroprocess 2012: 4th International Seminar on Hydrometallurgical Processes, Santiago, Chile.
Ilankoon, I.M.S.K. & Neethling, S.J. (2013) The effect of particle porosity on liquid holdup in heap leaching,
vol. 45, pp. 73-80.

Murr, L.E., Schlitt, W.J. & Cathles, L.M. (1981) Experimental observations of solution flow in the leaching of
copper-bearing waste, W.J. & Hiskey, J.B. (eds), Proc. 2nd SME-SPE Int. Solution Mining Symp.,
Denver, CO, Schlitt, AIME, New York, NY, pp. 271-290.
Roman, R. J. (1977) Solution channeling in leach dumps, Transactions of the Society of the AIME, vol. 262, pp.
73-74.
Roman, R.J., Benner, B.R. & Becker, G.W. (1974) Diffusion model for heap leaching and its application to scaleup, Transactions of the Society of the AIME, vol. 256, pp. 247-252.
Yusuf, R. (1984) Liquid flow characteristics in heap and dump leaching, M.Sc. Thesis, University of New
South Wales, Australia.

Hydro-mechanical-chemical coupled modeling of

copper heap leaching with comsol-phreeqc

Orlando Silva,
Albert Nardi, Gabriela
Romn-Ross, Jorge
Molinero and David Arcos
Amphos 21
Consulting, Spain
Pablo Quesada
and Eduardo Ruiz
Amphos 21
Consulting, Peru
Jordi Guimer
Amphos 21
Consulting, Chile

Several coupled Hydro-Mechanical-Chemical (HMC) processes are involved

during the construction and operation of a heap leach pad. Due to the
complexity of the problem, common existing heap leaching models solve
the different phenomena in an uncoupled or partially coupled way. Also,
often modeling of leaching of copper oxide ores assumes that mineral
dissolution/precipitation is described by the diffusion-controlled shrinking
core model and the system is composed by a single copper-containing
mineral and a single mineral representing the gangue. Although these
simplifications help to solve the problem, they could result in poor
predictions of the composition of the pregnant solution and consequently,
an over or under estimation of the copper recovery and acid
consumption. Further, improving the design of heap leach pads requires
performing an integrated analysis of the HMC phenomena. Mechanics
stability and mineral recovery efficiency cannot be estimated accurately
without considering multiple coupling of the complex chemistry, soil
mechanics and hydrodynamics. Here we present a coupled HMC model
that accounts for the kinetics of several mineral species and the evolution
of a muti-lift heap. Numerical simulations are performed with COMSOLPHREEQC (iCP), an interface that allows solving coupled unsaturated flow,
soil mechanics and reactive transport using the Sequential Non Iterative
Approach (SNIA). The results show that changes in porosity and
permeability due to mineral dissolution/precipitation and irrigation time
affect significantly the hydromechanical behavior of the heap leach pad. A
strong advantage of the present model is that it can treat multiple chemical
species simultaneously, providing solutions that are best for interpreting
all the chemical interactions within a leaching system. Finally, the stability
of the pad and the ore recovery efficiency are predicted with more accuracy
because the model considers the interaction between variable saturated
flow and mechanical deformation during the construction and operation
of the heap.

INTRODUCTION
Heap leach pads constitute a common practice in the mining industry. In their design and
management it is not common to take into account criteria based on coupled hydro-mechanicalchemical processes, and these are of paramount importance for two main reasons: (i) they could
play an important role on the physical stability of the pad and; (ii) they are keys for the
metallurgical efficiency, especially in the mid-long term. As an example of the relevance of such a
complex coupled phenomena, it could be mentioned that geotechnical stability of the heap is
related to the distribution of liquid pressures which, at the same time, is related to the permeability
through the Darcys law. In its turn, permeability will be affected by porosity changes, which
depends on the consolidation state and the dissolution/precipitation of minerals.
Hydrometallurgical performance of the heap ore is strongly linked with the abovementioned
coupled processes.
There are several models in the literature related to geotechnical, hydrogeological and geochemical
disciplines, but it is not common that these models are coupled in a real engineering problem.
Usual approaches for heap leaching models solve the different phenomena in an uncoupled or just
partially coupled way due to their resolution complexity. Some of the existing approaches were
formulated to solve hydrodynamics and solute transport (Decker & Tyler, 1999), and soil mechanics
and hydrodynamics (Pacheco et al., 2011).
In the last decade some efforts for incorporating the chemistry in Hydro-Mechanical (HM)
problems have been done (Liu & Brady, 2004; Mata et al., 2005; Mohajeri et al., 2011), but most of
them only considered simple geochemical models over synthetic cases. Also some attempts to solve
the Thermal-Hydro-Mechanical-Chemical (THMC) equations in the context of the design of
radioactive waste facilities have been reported (Guimares, Gens & Olivella, 2002). There is a
clearly demand of numerical simulators that can help in the design, planning and predicting of
leaching operations at industrial scale (Cross et al., 2006; Bennett et al., 2012; McBride, Gebhart &
Cross, 2012; McBride et al., 2013). Recently, McBride, Cross & Gebhart (2012) developed a
mathematical model based on CFD technology that consider operational aspects of a leaching
process such as actively 3D leached and unleached heap areas. Nevertheless, the flow through the
heap is modeled in 1D, which can lead to substantial errors in the prediction of metal recovery as
the lateral flow is disregarded. In addition, mechanical phenomena were not considered.
Improving the design of heap leach pads requires performing an integrated analysis of the HMC
phenomena. The key aspects of the construction and operation such as mechanical stability and
mineral recovery efficiency cannot be estimated accurately without taking into account multiple
coupling of a complex chemistry, the soil mechanics and the hydrodynamics. To our knowledge,
there is no numerical simulation of a coupled HMC model that accounts for the kinetics of several
mineral species and the evolution of a muti-lift heap. In this paper, we present a model that
considers all the above mentioned coupled phenomena and a numerical tool for the simulation of
copper heap leaching.

METHODOLOGY
In this work we provide a simulation of a synthetic copper heap leach pad case to demonstrate the
relevance of the HMC coupling, as well as the capabilities of the used numerical tool. In brief, the
methodology followed in this study consists of: (i) development of a numerical tool for the
simulation of HMC problems; (ii) application of the numerical tool for simulating copper heap

leaching; (iii) demonstration of the potential of the tool for the design, planning and prediction of
industrial heap leaching operations.

Hydro-mechanical-chemical model
Several coupled HMC processes are involved during the construction and operation of a heap leach
pad. In this section we describe the governing equations assumed in the modeling of the heap
leaching process. First, the mass balance equation of liquid under isothermal conditions is
expressed as follow

l C m S l S t pl l q l l t vol f extw

(1)

where l is the liquid density, S l is the liquid saturation, Cm is the specific moisture capacity, pl is
the liquid pressure, S is the storage term, q l is the liquid flow velocity, vol is the volumetric
deformation and f extw is an external source term. The liquid flow velocity equation is given by the
Darcys law (Bear, 1988)

ql

kk r

pl l g

(2)

where k is the intrinsic permeability tensor, k r is the relative permeability, is the dynamic
viscosity of the liquid and g is the gravity tensor. The van Genuchtens retention function (van
Genuchten, 1980) was considered to account for the relationship between liquid saturation and
pressure. Also, the relative permeability was calculated according to the model of van Genuchten
(1980).
On the other hand, we used the so called non-conservative formulation for reactive transport of a
chemical species (Bea et al., 2009)

t c Lt c a f extw c*a l r eq r kin

(3)

where S l l , c is the concentration vector of species, c a is the concentration vector of aqueous

species, c *a is the vector of external concentrations of species, is porosity, r eq and r kin are the
vectors of equilibrium and kinetics reactions, respectively. In (3) we have introduced the linear
operator accounting for advection, dispersion, diffusion and non-chemical sink-source terms,
Lt l q l l D f extw , where D is the tensor of dispersion and diffusion
coefficients. The equation (3) consists of a system of Ns equations, where Ns is the number of
chemical species. However, it can be reduced to a system of size N c, the number of chemical
components, by eliminating the equilibrium reactions. Chemical components are defined as linear
combination of species whose mass is not affected by equilibrium reactions (Steefel & MacQuarrie,
1996; Saaltink, Ayora & Carrera, 1998)

t m Lt m a f extw m *a Mr kin

(4)

In (4) m Mc is the vector of components, which can be decomposed in subcomponents

containing the chemical species of the component present in the aqueous ( m a ), gaseous ( m g ),
mineral ( m m ) and sorbed ( m d ) phases;

M is a component matrix, which is defined as a matrix

that eliminates the equilibrium reactive term, i.e., Mreq 0 (Saaltink, Ayora & Carrera, 1998); and

m *a is the vector of external aqueous component concentration.

Finally, the soil mechanics of a heap leach pad was assumed to be governed by the equilibrium
equations (Zienkiewicz & Taylor, 1989)

1 s S l l g 0

(5)

where is the total stress tensor and s is the soil density. In a variable saturated porous medium
the effective stress, , controls the mechanical behavior. We adopted a simple approach, in which
p a I S l S lr 1 S lr p a pl I , where pa is the air pressure and S lr is the residual
liquid saturation. The Cam-Clay elasto-plastic constitutive law (Wood, 1990) and plane strain
conditions were assumed to close the problem.
The initial suction and composition of the ore material layers are represented by pl0 and m 0a ,
respectively. Initial mechanical conditions assumed are zero strains and stresses for each layer. No
flow conditions are assumed at the boundaries of the bedrock. Also, no flow conditions are
imposed at the lateral boundaries of the ore layers and the free surface of those layers beneath the
current layer under irrigation. Atmospheric pressure and mass outflow conditions are imposed at
the lateral boundaries of the drainage system. An inflow condition (qin) is considered at the surface
of every layer to reproduce the irrigation process. Horizontal displacement (ux) at the lateral
boundaries of the rock foundation is set at zero. Also zero vertical (uy) and horizontal displacements
are considered at the bottom boundary of the bedrock. No mechanical constrains and no loads are
considered at the remaining boundaries.

Numerical implementation in COMSOL-PHREEQC

The governing equations (1-5) have been solved with the numerical interface COMSOL-PHREEQC
(iCP) (Nardi et al., 2014). iCP is a Java interface that couples the general purpose finite element
framework COMSOL Multiphysics and the geochemical simulator PHREEQC (Parkhurst &
Appelo, 2013). In brief, the tool is based on implementing an operator splitting technique for
solving the hydro-mechanical problem (a coupled or uncoupled system of partial/ordinary
equations and algebraic differential equations) on COMSOL and the chemical system of algebraic
and differential equations on PHREEQC. A problem is set in iCP through three types of input files:
(i) a COMSOL file for solving the hydro-mechanical phenomena, (ii) a set of PHREEQC files
defining the chemical system, including initial and boundary conditions and (iii) a XML file that
establishes the link between COMSOL and PHREEQC, set the time discretization and specifies the
outputs. In addition, to deal with the large computational requirements of THMC models, iCP
automatically divides the model domain into subdomains where geochemical reactions are solved
in parallel by balancing the computational load over multiple threads. In particular, the simulation
of this work comprised of 14641 elements simulating a 500 days leach operation and employing a
transport time step of 86400 s, required 1 GB of storage and 1 day of CPU time on a 3.8 GHz
processor with 4 cores and 16 GB RAM. These numbers are in the range of other codes and
computational methods that only solve flow and reactive transport (e.g., McBride, Gebhart & Cross,

2012; McBride, Cross & Gebhart, 2012).

Application example: simulation of a multi-lift heap leach pad

The present HMC model was applied to simulate the construction and operation of a synthetic
copper heap leach pad consisting of five ore layers settled on a terrain with a slight slope (2%). Each
layer (stacking levels) is simulated sequentially, mimicking the sequence of construction and
operation of a real heap leach pad as schematized in Figure 1. In this sequential simulation
approach the results obtained during the calculation of the previous layers (un-1, pn-1, cn-1) are used to
define the initial (u0, p0, c0) and boundary conditions of the next layer.

step 2

step 1

step 3

u0, p0 , c0 are the initial conditions

for displacement, pressure and
chemical composition

step n
u0, p0, c0

u0, p0, c0
construction and
irrigation time 1

end step 1

u0, p0, c0
u0, p0, c0

u2, p2, c2

un-1, pn-1, cn-1

u1, p1, c1

u2, p2, c2

un-1, pn-1, cn-1

construction and
irrigation time 2

construction and
irrigation time 3

end step 2

un-1, pn-1 , cn-1 are conditions

from previous step

end step 3

time

end step n

Figure 1 Sequence of construction of a heap leach pad layer by layer. The layer n is represented by step n.
State variables of the problem are displacement (u), liquid pressure (p) and concentration (c)

We have considered that the construction of each layer is performed during 10 days, followed by an
irrigation period of 90 days at a constant irrigation rate of qin = 7.5 l h-1 m-2 applied on the top
surface of each layer. The initial liquid pressure in the ore layer ( pl0 ) is considered equal to -30 kPa.
The temperature of the system is assumed to be 25 C. The rock and ore densities were set at 3000
kg/m3 and 2600 kg/m3, respectively. An initial porosity of 30% was assumed both for the liner
system and ore bed, while the initial intrinsic permeabilities were set at 510-10 m2 and 410-11 m2,
respectively. The longitudinal dispersivities of the liner system and ore bed were assumed equal to
50 m and 10 m, respectively. The respective transversal dispersivities were set at 10 m and 2 m.
These values were chosen to assure numerical stability of the simulation but they may be
representative of industrial heap leaching, where the total heap height can be up to 200 m and the
area under leach can be in the order of 106 m2 (Schlesinger et al., 2011).
Table 1 Chemical reactions considered in the geochemical model
Mineral

Kinetic reactions

Chrysocolla

CuSiH4O5 + 2H+ = Cu+2 + H2O + H4SiO4

Chalcopyrite

CuFeS2 + 2H+ = Cu+2 + Fe+2 +2HS-

Pyrite

FeS2 + 2H+ 2e- = Fe+2 + 2HS-

Albite

NaAlSi3O8 + 2H2O = Na+ + Al(OH)4- + 3H4SiO4

K-Feldspar

KAlSi3O8 + 8H2O = K+ + Al(OH)4- +3H4SiO4

Mineral

Equilibrium reactions

Jarosite-K

KFe3(SO4)2(OH)6 + 6H+ = Fe+3 + 6H2O + K+ +2SO4-2

Jurbanite

AlOHSO4 + H+ = Al+3 + SO4-2 + H2O

Basaluminite

Al4(OH)10SO4 + 10H+ = 4Al+3 + SO4-2 + 10H2O

Fe(OH)3(a)

Fe(OH)3 + 3H+ = Fe+3 + 3H2O

Schwertmannite

Fe8O8(OH)4.5(SO4)1.75 +20.5H+ = 1.75SO4-2 + 8Fe+3 +12.5H2O

We assumed that the mineral is composed by Chrysocolla (20% w/w), Chalcopyrite (10% w/w),
Pyrite (8% w/w), Albite (30% w/w) and K-Feldspar (32% w/w), and a Cu grade of 0.48%. This
composition may be representative of a zone of increased fracturing with a low Pyrite/Chalcopyrite
ratio. The chemical evolution of the above minerals is assumed to be governed by kinetics reactions
(Dreier, 1999; Kimball et al., 2010; Williamson & Rimstidt, 1994; Chou & Wollast, 1985, Schweda,
1989) (see TABLE 1). Equilibrium precipitation-dissolution reactions of Jarosite-K, Jurbanite,
Basaluminite, Fe(OH)3(a) and Schwertmannite are also taken into account (TABLE 1). To simplify
the problem, bioleaching of Chalcopyrite was not considered in the model. The chemical
compositions of the initial pore water and irrigation solution are shown in Table 2.
Table 2 Chemical composition of the initial pore water and irrigation solution
Initial pore water, m 0a

Irrigation solution, m *a

pH

5.95

0.94

Eh, mV

738

Al, mol/l

2.5610

C, mol/l

3.4910-2

1.0410-5

Ca, mol/l

3.4910-2

2.7010-3

Cl, mol/l

1.3010-6

9.6310-2

K, mol/l

1.8910-2

6.3010-4

Fe, mol/l

2.5010-7

7.5010-6

Mg, mol/l

3.9910-8

1.3010-5

Na, mol/l

10-14

6.1010-4

S, mol/l

3.7910-2

4.5010-2

Si, mol/l

9.2510-17

3.2010-4

1161
-10

10-22

RESULTS AND DISCUSSION

Figure 2 shows the evolution of concentrations of selected dissolved species at the output of the
drainage system. The evolution of pH shows the arrival of the acidic solution to the liner system in
a time of around 60 days (Figure 2a). Kinetic dissolution of Chrysocolla and Chalcopyrite is
triggered by the infiltration of the acidic solution and, as a result, an increase of Cu concentrations
is observed in the outlet flow (Figure 2b). The chemical evolution of Si and Al concentrations

(Figure 2c and 2d) clearly shows that under this low pH conditions kinetic dissolution of
aluminosilicates is also favored. This process is highly dependent on the pH and, over a long period
of time, porosity can be significantly modified.
Figure 3 shows the temporal evolution of Chalcopyrite and Chrysocolla concentrations. The
concentration of Chalcopyrite practically does not change, while the Chrysocolla is almost
completely dissolved at the end of the leaching of each layer. This is due to the combination of three
factors: the slow dissolution rate of Chalcopyrite, the fast dissolution rate of Chrysocolla and the
initial amounts of both minerals, which are very similar.

Figure 2 Evolution of pH (a) and concentrations of Cu (b), Si (c) and Al (d) in the pregnant leach solution.
Irrigation and drainage mass flows during each stage are also included

Figure 3 Evolution of the Chalcopyrite (left) and Chrysocolla (right) concentrations in the leach pad

The copper recovery is presented in Figure 4a. Note that the copper recovery associated to the
leaching of each layer displays a common pattern that reflects the difference in the leaching rates of
Chrysocolla and Chalcopyrite. During the first days of leaching of a new layer, the copper is
recovered mainly from Chrysocolla. The fast dissolution rate of Chrysocolla is associated to a
higher slope of the recovery curve. Once Chrysocolla has been spent, the copper recovery continues
from Chalcopyrite to a lower rate (lower slope of the copper recovery curve). This pattern is more
evident during the leaching of the first layer, which provides a copper recovery of about 30%. The
copper recovered from the leaching of subsequent layers is about 8%. At the end of the operation
(500 days), approximately the 60% of the copper has been recovered. The recovered copper
increases due to the leaching of residual mineral from previous layers, depending on the thickness
of the loaded layers and the irrigation times.
The competition between Chrysocolla and Chalcopyrite is more evident by looking at the acid
sulfuric consumption, which is shown in Figure 4b. During a given irrigation period, the consumed
acid increases almost at a constant rate until reaching a plateau coinciding with the construction of
a new layer. After 25 days from the start of a new irrigation period, the acid consumption
experiments a small decrease, due to the complete dissolution of Chrysocolla. Note that this time
coincides with the time at which the copper recovery changes its rate (see Figure 4a). These changes
in copper recovery and acid consumption last about 5 days, after which the acid consumption
increases again at the initial rate of the current leaching stage. That is, the copper recovery and acid
consumption curves directly reflect the different dissolution rates of the mineral species considered
in the model, specially the Chrysocolla (fast dissolution rate) and the Chalcopyrite (slow dissolution
rate). Also, the plateau of acid consumption is clearly associated to the behavior of the irrigation
and drainage flow rates (see Figure 2). In addition, acid consumption during the leaching of a new
layer is progressively lower than the acid consumption of previous leaching stages. This is because
the mass of mineral of the new layer is lower and the impact of acid consumption from previous
leached layers is minor.

Figure 4 Predicted copper recovery (a) and acid consumption (b)

The hydrodynamics behavior of the system is conditioned by the time required for the irrigation
solution to saturate the heap as it flows to the drainage system. The evolution of the liquid
saturation is shown in Figure 5a. After a short period of time, a quasi-steady state is achieved for
every layer. In the present case it takes about 40 days since the layer construction. We have set an
observation point to understand the evolution of the liquid pressure (red spot on Figure 5a). As
shown in Figure 5b, two stages are found for every leaching step: the construction and the irrigation

periods. The effect of construction does not generate pore water pressure (as occurs in undrained
conditions) as the ore is mainly unsaturated and the construction has been simulated in a
progressive manner. The increase of liquid pressure at the beginning of each leaching step is caused
by irrigation. Nevertheless, as shown in Figure 5b this increase is progressively lower because the
irrigation decreases in the next leaching step. The irrigation time plays an important role in the
system behavior. This is due to the influence of irrigation on the liquid pressure and mass of
recovered mineral. Some amount of copper remains in the pad close to the slope limits, because
irrigation is not very effective on those zones. Ideally, those zones could be exploited as well.

Figure 5 Evolution of the liquid saturation in the leach pad. The red circle on the left (b) denotes the
observation point for pressure evolution shown in (a)

Changes in porosity and permeability due to mineral dissolution/precipitation and irrigation time
can affect significantly the hydromechanical behavior of the heap leach pad. Figure 6a shows the
evolution of the global porosity due to the combined hydromechanical and chemical effects. We
found that there are zones of the ore where the porosity increases due to the mineral dissolution
and other zones where the consolidation of the layers leads to a porosity decrease. In turn, the
mechanical effects are traduced in an increase in the vertical stress developed mainly on the rock
foundation as the layers are progressively loaded. This vertical stress increase also reaches the top
of the third layer, where the vertical displacement reaches a maximum.

Figure 6 Evolution of global porosity (a) and vertical stress in the leach pad (b)

CONCLUSIONS
We presented iCP, a numerical tool that is able to solve simultaneously unsaturated flow, soil
mechanics and reactive transport in porous media. This tool allows integrating in a single
framework hydromechanical and geochemical coupled phenomena and the processes involved in
the construction and operation of heap leach pads. Geochemical changes affecting porosity and
permeability are coupled explicitly in the hydraulic and geomechanical calculations. The model
calibration may be conducted in a two-stage optimization process based on (i) experimental data
from batch tests and kinetics test columns for reaction rate parameters optimization in PHREEQC
and (ii) field scale measurements and operational indicators (e.g., pressures, flow rates, copper
recovery, acid consumption) for inverse modeling of hydro-mechanical-solute transport parameters
in COMSOL.
The simulation results show that changes in porosity due to mineral dissolution/precipitation and
irrigation time affect significantly the hydromechanical behavior of the heap leach pad, and vice
versa. For the synthetic case of a multi-lift heap considered in the study, the simulated copper
recovery and acid consumption responds directly to the dissolution rates of the copper bearing
minerals considered in the model. A strong advantage of the present model is that it can treat
multiple chemical species simultaneously, providing solutions that are best for interpreting all the
chemical interactions within a leaching system. In addition, the stability of the pad and the ore
recovery efficiency can be predicted with more accuracy because the model considers the
interaction between variable saturated flow, mechanical deformation and chemical reactions during
the construction and irrigation of the different layers of the heap. Finally, the modeling
methodology presented in this work can help to improve the stability of the heap pad and optimize
the mineral recovery efficiency.

ACKNOWLEDGEMENTS
We want to acknowledge the partners of the iMaGe consortium (www.image-modelling.net) for
supporting the iCP development.

REFERENCES
Bea, S. A., Carrera, J., Ayora, C., Batlle, F. & Saaltink, M. W. (2009) CHEPROO: A Fortran 90 object-oriented
module to solve chemical processes in Earth Science models, Computers & Geosciences, vol. 35(6),
pp. 10981112.
Bear, J. (1988) Dynamics of fluids in porous media, Dover, 764 pp.
Bennett, C. R., McBride, D., Cross, M. & Gebhart, J. E. (2012) A comprehensive model for copper sulphide
heap leaching Part 1 Basic formulation and validation through column test simulation,
Hydrometallurgy, vol. 127128, pp. 150161.
Chou L. & Wollast, R. (1985) Steady-state kinetics and dissolution mechanisms of albite, American Journal of
Sciences, vol. 285, pp. 963-993.
Cross, M., Bennett, C., Croft, T. N., McBride, D. & Gebhart, J. E. (2006) Computational modeling of reactive
multi-phase flows in porous media: Applications to metals extraction and environmental
recovery processes, Minerals Engineering, vol. 19, pp. 10981108.

Decker, D. L., & Tyler, S. W. (1999) Hydrodynamics and solute transport in heap-leach mining. Closure,
remediation and management of precious metals heap-leach facilities. Univ. of Nevada, Reno, pp. 113.
Dreier, J. E. (1999) The chemistry of copper heap leaching,
http://jedreiergeo.com/copper/article1/Chemistry_of_Copper_Leaching.html.
Guimares, L., Gens, A., & Olivella, S. (2002) Modelling the geochemical behaviour of an unsaturated clay
subjected to heating and hydration, Proc. 3rd Int. Conf. on Unsaturated Soils, pp. 7176.
Kimball, B. E., Rimstidt, J. D., & Brantley, S. L. (2010) Chalcopyrite dissolution rate laws, Applied Geochemistry,
vol. 27(7), 972-983.
Liu, J. & Brady, B. H. (2004) Simulations of a coupled hydro-chemo-mechanical system in rocks. Geotechnical
& Geological Engineering, 22(1), pp. 121133.
Mata, C., Guimares, L. D. N., Ledesma, A., Gens, A., & Olivella, S. (2005) A hydro-geochemical analysis of
the saturation process with salt water of a bentonite crushed granite rock mixture in an
engineered nuclear barrier, Engineering Geology, vol. 81(3), pp. 227245.
McBride, D., Gebhart, J. E. & Cross, M. (2012) A comprehensive gold oxide heap leach model: Development
and validation, Hydrometallurgy, vol. 113114, pp. 98108.
McBride, D., Cross, M. & Gebhart, J. E. (2012) Heap leach modelling employing CFD technology: A process
heap model, Minerals Engineering, vol. 33, pp. 7279.
McBride, D., Croft, T. N., Cross, M., Bennett, C. & Gebhart, J. E. (2013) Optimization of a CFD Heap leach
model
and
sensitivity
analysis
of
process
operation,
Minerals
Engineering,
http://dx.doi.org/10.1016/j.mineng.2013.11.010.
Mohajeri, A., Muhlhaus, H., Gross, L., & Baumgartl, T. (2011) Coupled Mechanical-Hydrological-Chemical
Problems in Elasto-plastic Saturated Soils and Soft Rocks Using escript, Advances in Bifurcation
and Degradation in Geomaterial, Springer Netherlands, pp. 269275.
Nardi, A., Idiart, A., Trinchero, P., de Vries, L. M. & Molinero, J. (2014) Interface Comsol-PHREEQC (iCP), an
efficient numerical framework for the solution of coupled multiphysics and geochemistry,
Computers & Geosciences, doi:10.1016/j.cageo.2014.04.011.
Pacheco, P. G. M., Purizaga, M., Huertas, J., & Romanel, C. (2011) Flow analysis and dynamic slope stability in
a copper ore heap leach, 2011 Pan-Am CGS Geotechnical Conference, 7 pages.
Parkhurst, D. L. & Appelo, C. A. J. (2013) Description of input and examples for PHREEQC version 3A
computer program for speciation, batch-reaction, one-dimensional transport, and inverse
geochemical calculations, U.S. Geological Survey Techniques and Methods, book 6, chap. A43,
497 p., available only at http://pubs.usgs.gov/tm/06/a43/.
Saaltink, M. W., Ayora, C., & Carrera, J. (1998) A mathematical formulation for reactive transport that
eliminates mineral concentrations, Water Resources Research, 7(34), pp. 16491656.
Schweda, P. (1989) Kinetics of alkali feldspar dissolution at low temperatures, D. L. Miles, ed., Water-rock
interaction 6: Proceedings Sixth International Symposium on Water-Rock Interaction, Edmonton,
International Association of Geochemistry and Cosmochemistry and Alberta Research Council,
pp. 609612.
Schlesinger, M. E., King, M. J., Sole, K. C. & Davenport, W. G. (2011) Extractive Metallurgy of Copper, Elsevier,
Fifth Edition.
Steefel, C. I. & MacQuarrie, K. T. (1996) Approaches to modelling reactive transport in porous media, Reviews
in Mineralogy and geochemistry, 34(1), pp. 85129.

van Genuchten, M. Th. (1980) A closed-form equation for predicting the hydraulic conductivity of unsaturated
soils, Soil Science Society of America Journal, 44, pp. 892898.
Williamson, M. A. & Rimstidt, D. (1994) The kinetics of electrochemical rate-determining step of aqueous
pyrite oxidation, Geochimica et Cosmochimica Acta, vol. 58(24), 5443-5454.
Wood, D. M. (1990) Soil behaviour and critical state soil mechanics, Cambridge University Press.
Zienkiewicz, O. C. & Taylor, R. L. (1989) The finite element method, McGraw Hill, vol. 1.

New approaches to old problems can dramatically

improve the viability of operations

David Robinson, Iztok

Livk, Jie Wu, Phillip
Fawell, Keith Barnard,
Rueben Rajasingam and
John Farrow
Minerals Down Under
National Research Flagship,
csiro, Australia

Most mineral processing operations face decreasing margins between

production costs and revenue, so that the industry more than ever is in
need of new innovations, to enhance existing equipment and to provide the
basis for the next generation of processing.
Feed and flow variations, conflicting demands of differing operational
components and insufficient attention to detail in both the design and
operation phases are amongst many other factors that account for some of
our problems. New diagnostic tools and previously impossible
measurements are increasingly frequently helping us identify opportunities
to improve the design and/or operation of ubiquitous unit operations such
as reactors, thickeners, mixer settlers and EW cells. The development of a
much deeper understanding of process fundamentals and new physical
modeling techniques has enabled the development of highly sophisticated
mathematical models of full-scale unit operations which have been
validated by in situ measurements on a wide range of mineral processing
operations. These tools ensure there is no longer the need to replicate the
inefficiencies of the past.
This paper highlights a range of existing inefficiencies and illustrates
how a combination of innovation, diagnosis and modeling has drawn
attention to the opportunities. Working closely with plant management
and operation based technical staff, many of these opportunities have been
realized and subsequent (and often minor) modifications to existing
equipment have led to demonstrable performance improvements and
benefits often beyond expectations.
Examples and related case studies will be presented from areas such as:
Improved slurry mixing in stirred tanks; Controlled bubble size and
increased mass transfer during gas sparging; Reduced energy requirement
for high viscosity slurry pumping; Enhanced flocculation and solid / liquid
separation; Improved fluid flow characteristics in solvent extraction settlers
for reduced entrainment loss and potential to increase throughput; and
Reduced (toward elimination) acid mist in electrowinning.

INTRODUCTION
The two high level, industry wide and seemingly ever-increasing pressures on operating plants
today are declining (and variable) quality of feed materials and increasing costs. The pressure on
feed quality relates not only to hardness and grade but also mineralogy (liberation and association),
gangue chemistry, impurity levels and process consequences such as leach kinetics and recoveries,
impurity deportment, reagent consumption, clay content, solid / liquid separation, etc. The pressure
on costs includes the cost of capital for new projects and expansions, as well as the ongoing costs
for labour, electricity, water and reagents, equipment, spares and maintenance, waste handling, etc.
Within our research group we have focussed a diverse group of scientists, mathematicians and
engineers toward developing technologies that challenge these paradigms and give our industry
options to ensure their long term future. This paper will highlight selected process engineering
innovations built on our established base of intellectual property include considerable input from
our research partners in both academia and industry.
Challenging the status quo is particularly difficult in an industry where there is a good deal of risk
aversion and the approach is to be a quick follower rather than the leader in new technology
implementation. Often the off-the-shelf package for unit operations in a complex flowsheet is not
optimal for local needs and often is actually fundamentally limited or flawed. Only when fresh
research perspectives and approaches are brought to bear on these well established or traditional
unit operations will progress really be made. It is the intention of this paper to present several
innovative new industry transformative technologies that can change the paradigm.

INNOVATION: TRIED, TESTED, MODELLED AND IMPLEMENTED

The CSIRO approach builds upon fundamental understanding and innovative engineering and is
based on the use of both mathematical and physical modelling, industry testing and thorough
validation before the roll-out of any innovation. The following examples are selected to illustrate a
number of process and engineering innovations that have progressed from laboratory concepts
through development, modelling and various stages of industry testing and validation. Several of
these innovations have already well established track record of significant impact in the minerals
industry while others have huge potential, but have still to be validated or rolled out to end users.

Increasingly efficient slurry mixing

Traditional mixing systems range from relatively simple, long shaft turbine impellors, through
multiple impellors, baffled systems and airlift agitation systems such as Pachucas. Many operating
companies and even commodity sectors have their preferred options. However, traditional mixing
technologies are in many cases associated with high capital costs, excessive power consumption
and high operating costs. Such systems are typically sensitive to solids size, solids specific gravity
and loading, prone to problems relating to tip speed, and can be a major challenge to restart after a
power failure or shut down.
To address some of these challenges CSIRO, in partnership with Queensland Alumina Ltd (QAL),
developed and patented Swirl Flow Technology (SFT)1 to exploit the relatively low energy swirling
motion created with a novel short shafted swirl flow rotor that sits near the top of the vessel. The

Patent family for SFT

rotor creates a tornado like effect, as shown in Figure 1, lifting solids from the base of the tank to the
rotor, and in doing so imparts energy to the solids to ensure good mixing as they continue to swirl
around the outside of the tank en route to the base of the tank again. The short shaft and relatively
small rotor result in huge capital cost savings. Improved mixing is achieved with less power
leading to significant energy savings. A lower rotor tip speed dramatically reduces wear on the
agitator compared to conventional systems. Due to the creation of high slurry velocities at the tank
wall, the swirling motion self-cleans the surface of the reactor suppressing scale build up, which
helps reduce down time and maintenance requirements (Wu et al., 2012). The unique SFT flow
pattern ensures good slurry mixing and reduced scaling rates in large scale (e.g. 30 m tall, 10 m
diameter) conical and flat bottom tanks.
The capability to model scaling rates enables development and optimization of the swirl flow
design in the laboratory, before being rolled-out at a plant-scale. Figure 2 shows reduced scale
formation on tank walls achieved in laboratory potassium nitrate crystallization by replacing a
conventional agitator design (A310 with baffles) with SFT agitation.
The successful application of SFT at the QAL alumina refinery has resulted in significant scale
reduction. The de-scaling downtime of their tanks has also been reduced by at least ~50%, thus
increasing tank availability and throughput. The first full size installation at QAL was done in 1997
and the benefits have now been well demonstrated through over 17 years of reliable operation.
Following a strategy whereby existing agitation systems are only replaced when necessary a total of
21 SFT agitated ~2 ML tanks are currently operating at QAL (Stegink et al., 2012).

(a)

(b)

(c)

Figure 1 Swirl Flow Technology (SFT) for agitating slurries: (a) SFT plant installation, (b) SFT in laboratory,
and (c) illustration of the SFT flow pattern

(a)

(b)

Figure 2 Extent of scaling in potassium nitrate crystallization tanks with different mixing technologies: (a)
A310 impellor with baffles, and (b) SFT agitation

On some operations power outages are not uncommon and the SFT allows start-up from an
unplanned shutdown since the rotor freely rotates high up in the tank within clarified liquor to start
the swirl flow mixing without the risk of the impellor being bogged down by a bed of settled solids.
With significantly lower capital and operating costs, and now well demonstrated operating
benefits, SFT is attracting attention from many operating companies and is being rolled out across
many applications including those in the bauxite/alumina, gold, and base metals industries.

Bubble size control and increased mass transfer during gas sparging
The optimisation of bubble size and subsequent mass transfer of gasses into solutions or slurries
can be a critical and often rate limiting step of many processes. For example; oxygen transfer is
important in a number of commonly employed metallurgical processes including cyanidation,
oxidative leach processes including pressure and bio-oxidation, iron oxidation and rejection
circuits, the INCO cyanide destruction process and a number of waste water treatment processes.
In the processes, where the gas-liquid mass transfer is the rate controlling step, the capacity design
is typically based on a volumetric mass transfer coefficient (K La) and the gas solubility in the
solution concerned. However, in a number of cases the designed oxygen transfer capacity for
systems is not achieved and establishing the cause and how this can be improved or optimized
often remains a challenge. Underperformance of oxygen transfer systems, particularly for slurries,
is related to a reduced Oxygen Transfer Rate (OTR), which may be caused by uncontrolled physical
and chemical factors. The formation of large air bubbles due to poor agitation or a poor gas inlet
system design is, however, the most common reason.
To solve the problem of large air bubbles, CSIRO has, in collaboration with Chemical Plant and
Engineering Ltd (CPE), developed the GLNozzle technology, which is designed to intensify the
gas / liquid mass transfer for gas sparging reactors in the minerals processing industry (Figure 3),
by making small bubbles reproducibly at a low cost. The GLNozzle employs a high velocity jet
design to break up gas bubbles in the nozzle prior to the bubbles being introduced into the reactor
vessel. The nozzle is engineered to be able to operate in highly abrasive slurry flows in such a way
that even at a high velocity there is minimum wear on the nozzle or associated equipment.

(a)

(b)

Figure 3 Air bubbles size and flow pattern visualized in a CSIRO laboratory-scale tank:
(a) with a conventional design air injection unit, and (b) using the CSIRO developed GLNozzle

Experience to date with a variety of applications of the GLNozzle demonstrates it is consistently a

low cost item and relatively simple to install in most applications, usually with no need to upgrade
the agitation system, as illustrated in Figure 4. The first installation at Stawell Gold Mine, Victoria
in 2003 immediately transformed the process and resulted in a remarkable 25% saving of oxygen
consumption. CPE has since successfully installed GLNozzle systems in many other mineral and
waste water treatment applications.

Figure 4 Installation of CSIRO GLNozzle as applied for enhanced oxygen transfer at the Stawell gold mine

Reduced energy requirement for high viscosity slurry pumping

The advantages of pumping higher pulp densities are almost self evident and the cost, availability
and desire to return/recycle water becoming ever more important across the globe. The challenge
in making our slurries denser is the rapid increase in energy demand for pumping slurries as their
viscosity rises with increasing solids concentration. The solution involves reducing the friction
between the walls of the pipe and the slurry, but this is not simple in practice for industrial scale
applications. There have been many attempts to reduce the wall friction with the introduction of
water through one, two, or more nozzles but these attempts have led to small and short lived
benefits and cause significant dilution of the slurries (Atherton et al., 1994; Schlecht et al., 1997).
CSIRO patented concepts introduce water (or any other reagent) into a slurry pipeline exactly
where it is needed, between the walls of the pipe and the slurry (shown schematically in Figure 5).
Pilot results are very encouraging, and application of this low cost drag reduction technology (DRT)
can result in very significant energy savings, whilst creating opportunity to pump higher pulp
density slurries, save water and increase production capacity in existing plants.
Recent tests have shown the drag reduction effect to be more persistent than expected even past
pipe connections, valves, round bends and across T-sections. Industrial testing of this DRT to
enhance the discharge from a paste thickener is underway. Success in this application has the
potential to eliminate thickener blockages, reduce the feed thickener underflow dilution
requirements, and enable an increase in the overall production capacity in the subsequent leach
circuits.

Figure 5 Schematic of the CSIRO drag reduction technology (DRT) for lubrication of pipe walls during slurry
pumping, and a picture of an installed unit

Enhanced flocculation and solid / liquid separation

The separation of solids from liquids is a feature of every hydrometallurgical operation; essential to
the separation of the pregnant leach solution (PLS) from leach residue, the separation of precipitates
(impurities and products) from leach solutions and waste streams. However, it is the unit operation
most often taken for granted. Failure or even just poor performance in this ubiquitous unit
operation will at least lead to unnecessary expense and poor plant performance but could, if left
long enough actually cripple a mineral processing operation.
Full-scale flocculation of slurries within a feedpipe, launder or feedwell prior to gravity thickening
very rarely produces the performance predicted from small-scale test work. This is generally
reflected in lower settling rates / throughputs, higher overflow solids concentrations, lower
underflow densities or higher flocculant consumptions. Most older thickeners are vastly over-sized
and quite forgiving of such process inefficiencies. However, modern thickeners are smaller in
diameter and their performance is predicated on consistently achieving good flocculation and
settling.
There are a range of factors that contribute to sub-optimal flocculation in thickeners. CSIRO has
established techniques for flocculating mineral suspensions under the tightest possible control of
the applied shear and reaction time during flocculation, with the ability to monitor aggregate
properties in real time. Turbulent pipe flow flocculation with in-stream monitoring by focused
beam reflectance measurement (FBRM) has allowed the kinetics of aggregate growth and breakage
to be qualified under suspension shear, flocculant dosage and solids concentration conditions of
practical significance, from which many insights have been gained (Heath et al., 2006a; Owen et al.,
2008). Higher flocculant dosages favour larger aggregates, but as the images captured in real-time
suggest (Figure 6), their irregular, low-density, fractal like structures mean that larger aggregates
are also more readily ruptured.

(a)

(b)

(c)

Figure 6 Real-time imaging of calcite flocculation (60 kg m-3) in turbulent pipe flow at flocculant dosages of:
(a) 40, (b) 60 and (c) 80 g t-1 (Mettler Toledo PVM probe, each image 1115 m full width)

Flocculation kinetics have been captured in a mathematical model of the flocculation process which
can be customized for a particular feed and flocculant system to describe the response to a wide
range of conditions (Heath et al., 2006b). This population balance (PB) model has then been used
within the CSIRO computational fluid dynamics (CFD) model of feedwells. This model has been
used extensively for the optimisation of feedwells designs around the world (Nguyen et al., 2012)
with very substantial financial and environmental benefits for operations.
Studies have produced numerous important insights for the design and operation of thickeners:
Feedwell modelling has led to the development of the first detailed design rules that can be used
to size feedwells, establish the most appropriate range for inlet velocities and position flocculant
sparges for the best flocculation performance.

 The model has highlights the importance of solids dilution being achieved prior to flocculation
to realise any benefit to throughput. Solids dilution streams exist within many feedwell designs,
but are rarely well utilized to enhance flocculation, and are often limited in the dilution they can
provide.
While polymer flocculants offer the potential for greater performance than coagulants, the much
lower dosages they require means that their efficient mixing through a slurry is critical to
achieving good fines capture, and hence high flocculant dilution and staged/split dosing is often
beneficial. Additional complications arise from the highly viscous nature of the high molecular
weight polymer flocculant solutions, with the long chain polymers requiring ageing times in the
order of hours on make-up (Owen et al., 2007) and potentially suffering activity losses from
pumping over extended distances (Owen et al., 2009). The activity of anionic flocculants are also
affected by the quality of the water used for make-up and dilution (Witham et al., 2012).
Operations that treat feeds with the potential for significant variations in mineralogy
(particularly in relation to clays) often report concerns with problematic feeds that lead to
reduced settling rates, high overflow solids or even the complete failure of feedwell flocculation.
In fact, many of these feeds are actually readily flocculated but a lower particle size (or more
correctly, a higher proportion of fines solids below ~10 m) increases the required flocculant
dosage, greatly reduces the optimum solids concentration for flocculation and, inevitably, the
settling flux that can be attained.

Improved fluid flow characteristics in solvent extraction settlers

Efficient solvent extraction (SX) performance is crucial to cost effective separation of metals in an
increasingly broad range of mineral processing operations. There are a number of production
challenges faced by operations, one being costly reagent loss and another being limitations in
throughput attributable sub-optimal operation. A multiple-pronged research approach has been
adopted, including a combination of SX chemistry and kinetics, small scale physical modelling,
development of a variety of measurement tools and probes for mixer-settler characterisation, and
the development and ongoing refinement of CFD models (Barnard and Robinson, 2013).
Site-based investigations of liquid flow patterns in operating commercial settlers have been
achieved by development and application of ultrasonic velocity profiling (UVP) (Mohanarangam et
al., 2013). Mapping flow patterns before and after furniture modifications allows specific and
detailed effects associated with the change to be quantified. For example, at Freeport McMoRans
Morenci site (Morgan et al., 2011) operational changes undertaken as a result of such measurements
resulted in a 60 per cent reduction in organic entrainment.
The ability to readily measure flow patterns in opaque SX settlers has also been used to validate the
associated CFD model of mixer-settlers. This has opened up the opportunity to undertake relatively
cheap and quick CFD-based investigations to assess the merits of many and varied proposed design
modifications (Lane et al., 2012). It was found that the flow pattern of the existing design was suboptimal, with substantial short-circuiting and various regions of recirculation evident impacting
negatively on phase separation and operating costs. To improve this, additional CFD simulations
were performed to investigate modifications to the internal settler design could improve the flow
pattern. Changes to the picket fences, for example, the number of pickets, the gap width between
pickets, the positions of the fences and the number of fences. It was predicted that a revised design
would reduce early exiting of fluid and could improve operational performance. The flow pattern
for one such alternative design containing three picket fences is illustrated in Figure 7.

(a)

(b)

(c)

Figure 7 (a) velocity field and streamlines 100 mm below surface as measured; (b) CFD simulation of
corresponding velocity vectors and; (c) revised settler design with three picket fences

Reduction (toward elimination) of acid mist in electrowinning

Acid mist is a pervasive problem in electrowinning operations, and has an adverse effect on human
health, leading to tightening of regulations on exposure limits. Acid mist is formed from droplets
ejected into the tankhouse atmosphere by gas bubbles busting at the electrolyte surface (Lhuissier
and Villermaux, 2012). The gas is primarily oxygen evolved from the anode and while various
approaches to alleviate the problem have been adopted by companies, with varying degrees of
success (Davis and Hopkins, 1994; Alfantazi and Dreisinger, 2003). However these approaches
attempt to treat the symptom (acid mist) rather than the problem and many are expensive and
varyingly effective. A deeper knowledge of the underlying mechanisms is required to enable
development of optimal technologies for effective acid mist mitigation or elimination.
We have establish the capabilities and expertise required to address some of the knowledge gaps
and for the first time, developed an in situ technique to measure under industrial scale and
conditions, the characteristics of oxygen bubble formation within a cell and the resultant acid mist
droplet formation above the cell. Coupled with this capability is the ability to quantify the total acid
mist produced under a given operating condition. Central to this capability is a transparent cell that
facilitates the use of state-of-the-art laser and particle/droplet image analyser (PDIA) diagnostic
techniques to establish velocity profiles, oxygen bubble and acid mist droplet size (down to 1 m in
diameter, Figure 8).

Figure 8 Electrowinning pilot plant (top), PDIA set up (bottom left) and sample bubble and droplet images
produced by the PDIA technique (bottom right)

Figure 9 shows how different operating conditions can influence the acid mist droplet size
distribution, providing insights into optimisation of operating conditions to minimise acid mist.
Acid mist quantification was achieved with the construction of an acid mist capture device. The
device comprises of an extraction hood connected to a variable speed extraction fan unit. Within
the hood an acid mist collection filter adequately captured and retained the acid mist while
maintaining a practical pressure drop (Figure 10).

Figure 9 Acid mist droplet size distribution results under three operating conditions

Figure 10 Acid mist capture device and extraction hood above EW cell.

The next phase of this work will involve expanding the range of operating parameters and
understanding their impact on acid mist generation. With the broader aim of acid mist elimination,
a CFD model will be developed to predict bubble flow within a cell and how changes to design
operating conditions affect the distribution of bubbles within the cell and the resulting acid mist.

THE FUTURE
The above examples of highly (or potentially highly) impactful innovations can often be
implemented in the design of new operations or retrofitted to existing operations to ensure the
minerals industry makes progress down the cost curve. With respect to greenfield operations, their
progress and success will depend upon competition and thus cost of capital. Relatively low capital
innovations, such as those presented above, can improve the return on investment both directly
through improved productivity and indirectly with additional benefits such as reduced footprint,
lower plant maintenance and operating costs, and improved safety characteristics and
environmental profile through reduced energy demand and improved water recovery.
The lower the capital barrier the more likely success in both attracting investment and establishing
a competitive rate of return for new plants. Similarly, for brownfield expansions or simply
improvement in existing plants, moving down the operating cost curve is particularly critical at this
time, when many operations are struggling with runaway costs. Again, the ability to implement
innovations at relatively low capital cost and dramatically improve process performance and
economics is essential.

The track record of the CSIRO team stands for itself with the above examples of technological
innovations, which include technologies proven in industrial applications, as well as, those with a
huge potential to have further significant impact on our industry. With the appropriate level of
industry engagement it is our intention to ensure the unique capability of CSIRO to pull together
diverse research and technology teams to address the challenges of mineral processing industry
will continue to deliver industry changing innovations into the future.

REFERENCES
Alfantazi, A.M., Dreisinger, D.B. (2003) Foaming behaviour of surfactants for acid mist control in zinc
electrolysis processes, Hydrometallurgy, 69(13), 5772.
Atherton, T., Schmidt, L., Kelly, S., Crow, H., Anderson, T.M. (1994) Sludge pipeline lubrication system, Patent
No. US5361797-A, Nov 8, 1994.
Barnard, K. R., Robinson, D.J. (2013) Improving solvent extraction technology. In Copper 2013 conference vol. IVHydrometallurgy, Ed. G. Urgarte, Chilean Institute of Mining Engineers, Santiago, Chile, 163177.
Davis, J.A., Hopkins, W.R. (1994) Recent developments in electrometallurgical tankhouse environmental
control, CIM Bulletin, 87(981), 8694.
Heath, A.R., Bahri, P.A., Fawell, P.D., Farrow, J.B. (2006a) Polymer flocculation of calcite: Experimental results
from turbulent pipe flow. AIChE J., 52, 1284-1293.
Heath, A.R., Bahri, P.A., Fawell, P.D., Farrow, J.B. (2006b) Polymer flocculation of calcite: Population balance
model. AIChE J., 52, 1641-1653.
Lane, G., Mohanarangam, K., Yang, W. (2012) Assessment of the flow pattern in a solvent extraction settler,
Ninth International Conference on CFD in the Minerals and Process Industries, C.B. Solnordal, P.
Liovic, G.W. Delaney and P.J. Witt, Eds., CSIRO Australia.
Lhuissier, H., Villermaux, E. (2012) Bursting bubble aerosols, Journal of Fluid Mechanics, 696, 544.
Mohanarangam, K., Yang, W., Barnard, K.R., Kelly, N.J., Robinson, D.J. (2013) Flow mapping of full scale
solvent extraction settlers using Pulsed Doppler UVP technique. Chem. Eng. Sci, 104, 925-933.
Morgan, J., Cashwell, W., Doole, S., Steeples, J. (2011) Entrainment reduction at Freeport-McMoRan copper &
gold Morenci operations, Solvent Extr. Ion Exch., 29 (56), 854-867.
Nguyen, T.V., Farrow, J.B., Smith, J., Fawell, P.D. (2012) Design and development of a novel thickener feedwell
using computational fluid dynamics. The Journal of the Southern African Institute of Mining and
Metallurgy, 112, 939-948.
Owen, A.T., Fawell, P.D., Swift, J.D. (2007) The preparation and ageing of acrylamide/acrylate copolymer
flocculant solutions. Int. J. Miner. Process., 84, 3-14.
Owen, A.T., Fawell, P.D., Swift, J.D., Labbett, D.M., Benn, F.A., Farrow, J.B. (2008) Using turbulent pipe flow to
study the factors affecting polymer-bridging flocculation of mineral systems. Int. J. Miner. Process., 87,
90-99.
Owen, A.T., Nguyen, T.V., Fawell, P.D. (2009) The effect of flocculant solution transport and addition
conditions on feedwell performance in gravity thickeners. Int. J. Miner. Process., 93, 115-127.
Schlecht, K., Schuster, U., Rockstroha, A. (1997) Device for conveying thick substances containing a great deal
of solid material, Patent No. US5605174-A, Feb 25, 1997.
Stegink, D., Davis, T., Wu, J. (2012) Operating experience with improved precipitation agitation system,
Proceedings of the 9th International Alumina Quality Workshop, 18-22 March 2012, Perth, Australia.

Witham, M.I., Grabsch, A.F., Owen, A.T., Fawell, P.D. (2012) The effect of cations on the activity of anionic
polyacrylamide flocculant solutions. Int. J. Miner. Process., 114117, 51-62.
Wu, J., Lane, G., Livk, I., Nguyen, B., Graham, L., Stegink, D., Davis, T. (2012) Swirl flow agitation for scale
suppression, Journal of Minerals Processing, 112-113, 19-29.

Characterization and in situ

monitoring of large scale heap
leach fluid dynamics

Danny Santis-Valenzuela
GeoSystems Analysis, Chile
Michael Milczarek,
Tzung-mow Yao and
Jason Keller
GeoSystems Analysis,
Inc., usa

The efficiency of solution and air flow dynamics significantly affects metal
recovery in heap leach operations. Quantifying these processes, however,
presents extreme challenges. Depending on the desired monitoring
parameters, monitoring systems can be designed to assess phreatic level
surface, in situ solution content, oxygen content and temperature, and air
injection pressures and air velocities. Sensors can be installed during heap
construction or via coreholes or drive points to depths limited only by the
achievable maximum depth of the chosen drilling method. Sensor
installation can also be coordinated with a heap pad characterization
program to determine in situ ore and permeability related properties.
The selection and application of appropriate monitoring methods
depends on the type of leaching operation. Cyanide leach solutions are less
corrosive and heap temperatures typically approximate ambient conditions
which allows for more flexibility in sensor selection. In addition, air
injection and oxygen concentration monitoring is not typically important
such that monitoring instrumentation can focus on phreatic level and
solution content/distribution monitoring. Conversely, acid leach heap
operations benefit greatly from in situ temperature and oxygen monitoring,
but these technologies require robust sensors in elevated temperature and
corrosive environments.
As a case study, we present the methodology and monitoring results
from a large-scale copper heap leach pilot project. The pilot heap was
designed to collect solution from different sectors of the heap to evaluate
efficiency of solution movement. Instrumentation was installed after the
placement of ore material using drilling methods at various locations and at
different depth intervals to monitor temperature, gaseous oxygen, solution
content and capillary pressure. Monitoring data was collected in real time
and transmitted to a control room with telemetry. Data collected from the
monitoring program were used to validate initial field scale characterization
and laboratory and column test results and support industrial scale heap
leach pad design and operation.

INTRODUCTION
In-situ monitoring of solution and air flow within an active heap leach facility can provide useful
information about the efficiency of the leaching process and create significant opportunities to
improve metal recoveries. For example, solution content monitoring can determine when and
whether the ore has been adequately wetted, oxygen content and temperature monitoring can
determine the distribution of air and temperature within the ore profile and whether aeration is
sufficient/efficient. Laboratory and column scale measurements of physical and hydraulic
properties of ore can provide valuable information in aiding heap leach design and leaching
operations (Guzman, Scheffel & Flaherty, 2008; Milczarek et al., 2013); however, these
measurements may not always be directly transferable to leach pad scale conditions (Bouffard &
West-Sells, 2009). Consequently, the ability to operationally monitor cribs and heaps can be useful
to quantify the effect of varying crush conditions, ore types, irrigation rates and, in the case of
copper sulphide ores, aeration schemes on solution and aeration efficiency. Moreover, in-situ
temperature, solution content, capillary pressure, solution chemistry, oxygen content, and air
pressure and velocity measurements can be used to validate laboratory and field characterization
tests and allow for real time adjustment of heap leach operations to increase leaching efficiencies.
In this paper, we provide an example of a 500,000 ton copper sulfide leach pad that was
instrumented and monitored to investigate large-scale heap leach fluid dynamics under varying
irrigation and aeration schemes and describe a conceptual model of hydraulic flow in the heap
derived from the monitoring data. The 18 m high heap had an approximately 90 m x 90 m leaching
area. Prior to stacking, areation lines were placed within an underlying gravel drainage layer to
provide oxygen to the leach ore for enhancing bio-assisted leaching. Two air-line grids were
established that allowed for the air source to originate from the east or from the west of the pad.
Data collected from the 1.3 year monitoring program were used to validate initial field scale
characterization and laboratory and column test results and support industrial scale heap leach pad
design and operation.

METHODOLOGY
Instrumentation
A plan view of the heap as well as location of instrument boreholes and solution collection modules
relative to the leaching core is provided in Figure 1. The heap leach monitoring system consisted of
nine 30 m by 30 m solution collection modules constructed beneath the gravel drainage layer.
Within each module was a borehole instrumented with temperature, oxygen/air piezometers,
raffinate content, and capillary pressure sensors at 3 m intervals below ground surface (bgs)
(diamonds in Figure 1). These boreholes also included a phreatic surface sensor installed 0.5 m
above the gravel drainage layer and suction lysimeters (for solution sampling) installed
approximately 6 and 12 m bgs. An additional sixteen boreholes were instrumented with
temperature and oxygen/air piezometer sensors at 3 m intervals and raffinate content and capillary
pressure sensors at approximately 5, 11, and 16 m bgs (circles in Figure 1).
The instrument boreholes were spatially located to provide data radiating from the center of the
leaching core; 21 instrument boreholes were located within the leaching core and 4 instrument
boreholes were located outside of the core to assess lateral migration of solution and air.
Instruments were installed in boreholes drilled with a 15 cm inside diameter hollow stem auger. All
downhole instruments were attached to the outside of 2-inch diameter PVC pipe and the PVC was
lowered inside the core barrel. The annulus between the monitoring instruments and the borehole
wall was backfilled with:

10/20 mesh graded sand mixed 50:50 with ore surrounding the suction lysimeters
Leach ore surrounding the advanced tensiometers and raffinate content sensors
3/8 inch gravel surrounding the air piezometers, oxygen sensors, and temperature sensors.

Backfill materials were added directly to the annulus of the borehole. A tag line was used to
determine the correct backfill depth for each material. Bentonite chips were added above each
backfill layer using a tremie pipe to seal between the instrument arrays and prevent surface and
subsurface water, air, and/or heat from preferentially migrating down the borehole. The bentonite
chips were hydrated after placement. All sensors were wired to data loggers for automated data
collection; data loggers were connected to the central control room via telemetry.

Figure 1

Plan view of heap instrument well locations and solution collection modules

Post-test coring
Following completion of testing, sonic coring of 52 coreholes located within and around the
leaching core of the heap was performed to collect samples for physical and hydraulic property
testing. Geologic logging of the cores was also performed to estimate ore material texture and
degree of oxidation (color). Particle size distribution (PSD) tests were performed on each 1.5 m
interval.
Laboratory hydraulic analyses were performed on twelve composited cores selected to represent
the range of particle size, core bulk density, predominant color characteristic (oxidized or un-

oxidized ore), and module performance data. Consolidation permeability tests were conducted to
determine the relationship of saturated and unsaturated hydraulic conductivity (Ksat and Kunsat)
and air permeability to overburden pressures representing approximate sample depths within the
heap. Details of the hydraulic test methods are described in Milczarek et al. (2013). The results of
hydraulic property analyses were then correlated with particle size distribution data and sample
depth to develop a predicted spatial distribution of heap hydraulic parameters across the heap.

RESULTS AND DISCUSSION

Estimated solution budget and drainage
An accounting of solution entering the heap and the estimated overall heap solution content is
presented in Figure 2Error! No se encuentra el origen de la referencia.. The predicted volumetric
solution content assumes an average initial solution content of 0.12 cm3/cm3 (0.07 g/g) prior to the
start of irrigation on November 7, 2007 and that the difference between the volume of solution
irrigated and the volume of solution lost to evaporation and drainage equates to solution storage
that is either added or subtracted from the solution content. Predicted solution contents increased
with the onset of double grid aeration then decreased during the drainage phase of the study when
irrigation was stopped (Figure 2). Predicted solution content increased again with the restart of
irrigation and continued to steadily rise for the remainder of the project, to approximately 0.216
cm3/cm3 (21.6 percent volumetric), by the end of operation. The continued increase in storage
represents solution going into storage within the leaching core and also represents any solution that
moved laterally outside the core.
Figure 3 present the ratio between drainage and irrigation (drainage/irrigation) for module
groupings along the north-south direction. Doubling the east aeration rate resulted in a significant
drop in drainage/irrigation from eastern modules 1, 4, 7 (0.76 to 0.18), a smaller decrease in the
middle modules 2, 5, 8 (1.08 to 0.82) and a large increase in drainage/irrigation from western
modules 3, 6, 9 (1.01 to 1.60). The effect of aeration direction is also observed during the alternating
aeration period. Drainage/irrigation along the east side modules (1, 4, 7) decreased when the
aeration source was from the east and increased when the aeration source was from the west.
Similar behavior is observed with the west modules (3, 6, 9); drainage/irrigation from the west
modules decreased during aeration from the west and increased when aeration was from the east.
Note that drainage/irrigation from the middle modules (2, 5, 8) remained relatively stable during
alternating aeration and that during the no aeration period, drainage/irrigation trends from the
north-south module groups resembled the single east side aeration trends.
It is possible that damage to solution collection drainage pipes could account for some
intercommunication of module solution or damage to aeration pipes could have resulted in
irregular aeration distribution throughout the heap. Nonetheless, the correlation of aeration
direction and rates to module drainage rates is strong, and is believed to be primarily caused by
entrapped air within the leach ore and aeration back pressure within the drainage layer as
discussed further below.

Figure 2

Figure 3

Irrigation rate over time and estimated solution budget

Average normalized drainage for module groups during different aeration schemes

Capillary pressure, phreatic surface, and temperature

Figure 4 provide soil water pressure potential (capillary pressure) contours interpolated from
advanced tensiometer measured capillary pressures at 4 m bgs at various time periods representing
single and double grid aeration from the east side. Capillary pressure data was also collected at 10
m and 15.5 m bgs but is not shown here due to space constraints. At all times capillary pressure was
variable both laterally and with depth. Negative capillary pressures (< 0; orange and red) represent
desired unsaturated leach conditions and very negative pressures (i.e. < -30 cm) indicate dry zones
that may not be receiving sufficient solution. Conversely, positive capillary pressures (blue) may
represent areas of entrapped air that reduce the effective porosity and downward solution flow,
resulting in perched solution conditions (pseudo-saturation). Figure 4 shows evidence of solution

mounding at the 4 m bgs interval in Module 9 during both single and double aeration periods
(mounding was also observed at 15.5 m bgs). Capillary pressure gradients can cause lateral solution
movement following from high to low (more negative) capillary pressures. Of note, Module 9
showed below average drainage rates whereas adjacent, and upgradient, modules 5, 6 and 8
showed above average drainage rates, suggesting that solution was flowing laterally from Module
9 into these adjacent modules.
Capillary pressures generally increased as aeration went from single east to double east aeration
(Figure 4). Capillary pressures subsequently decreased as aeration went to alternating aeration and
then to no aeration. Specifically, the doubling of air injection rates resulted in increased capillary
pressures which is most likely due to decreased leach ore hydraulic conductivity as air and solution
competed for flow paths. As aeration decreased or was alternated from east-west, competition of air
flow with solution flow was reduced and capillary pressures also decreased.

Figure 4 Capillary pressure contours at 4 m bgs at different time periods

Phreatic level monitoring 1.5 m above the drainage system liner (0.5 m into the leach ore), indicated
the phreatic surface gradient corresponded to changes in aeration regimes (date not shown).
Phreatic level changes also corresponded to changes in drainage rates; in general, as phreatic levels
decreased, drainage rates increased and vice-a-versa. The most probable mechanism for increased
phreatic levels (and decreased drainage) is aeration back pressure within the underlying drainage
layer and/or air entrapment in the leach ore above the drainage layer. This effectively reduces the
leach ore hydraulic conductivity and solution pressure in the overlying leach ore must increase to
allow the solution to drain. When aeration rates/pressures are reduced, entrapped air and back
pressure dissipates and more pore space is available for solution movement. This results in
increased leach ore hydraulic conductivity which also increases the drainage rate and decreases the
phreatic level.
Temperatures increased during the early period of heap operation and generally stabilized as
double aeration from the east began (Figure 5). At the onset of increased irrigation (4.5 l/m2/hr) on
5/7/2008 temperatures began to decrease and continued to decrease until irrigation was stopped
on 7/23/2008. At this time temperatures began to increase and then generally stabilized during the
alternating aeration scheme. Temperatures then decreased after aeration was stopped. Within the
core of the heap, temperatures were within the optimum temperature range for bioleaching (30C

to 60C) for the majority of operation. The increase in heap temperatures with the cessation of
irrigation also coincided with an increase in oxygen (data not shown).

Figure 5 Leaching core temperature for Module 5

Post leach ore sampling and testing

Average PSD for all post-leach core samples and the average PSD for pre-leach ore material (after
crushing and prior to agglomeration) indicated that leaching and decrepitation of the ore
significantly increased the amount of material passing all mesh sizes, with material passing the
#100 mesh increasing on average from 11.4 to 21.1 percent for pre-leach ore to post-leach ore
(Figure 6). The post-leach ore was also more well-graded, whereby the distribution of particles is
more evenly distributed between sieve sizes. Well graded ore material is easier to compact (higher
bulk density) which reduces the leach ore permeability under in-situ heap pressures.
The average percent passing the #100 mesh versus module percolation as a percent of total heap
percolation showed a trend of decreasing percolation with increasing percent passing the #100
mesh (Figure 7). Excluding Module 9, which was an outlier, a reasonable relationship between
percent passing the #100 mesh and percolation is obtained (R2 = 0.66). The lower drainage from the
eastern modules is likely due to these modules being closer to aeration source for the majority of
the measurement record and the decreased ore permeability caused by the greater percent passing
the #100 mesh. As discussed above, Module 9 experienced solution mounding at the 4 m bgs and
15.5 m bgs depths, in addition, the 10 m bgs intermediate depth showed more negative pressures
(much drier). There was an observed increase in fines at 3.5 m and 15.5 bgs in Module 9 which
could be due to increased solution contact and ore decrepitation at those depths, whereas the
intermediate depths showed much less decrepitation most likely due to lateral movement of
solution away from Module 9 into the adjacent modules.
Corehole logging indicated variable ore oxidation with the presence of jarosite (yellow ore)
precipitates indicating oxidizing conditions and absence of precipitates (grey ore) indicating
reducing conditions. The average percent of total core that was yellow (oxidized ore) and percent
that was gray (reduced ore) is presented in Figure 8. Other precipitates (i.e. goethite) observed were
generally less than a few percent within each module and are not presented. The average percent of
oxidized ore exceeded the average percent of reduced ore in all modules, except for Module 9 and
in coreholes located outside the perimeter of the modules (Figure 8).

Figure 6 Average particle size distribution for all post-leach samples and crush head samples

Figure 7 Average percent passing the #100 mesh versus module percolation as a percent of total heap
percolation

The greatest oxidation levels were observed in the eastern and western modules (except for Module
9) and also Module 5. Modules, 2, 8 and 9 showed the lowest oxidation. Higher oxidation in the
eastern and western modules may be due to the proximity of these modules to the aeration source,
whereas Module 9 most likely received less solution (and acid reagent) due to mounding and
lateral flow from above 4 m bgs (see above). Lower oxidation levels in Modules 2 and 8 could be
due to these modules being farther from the aeration source, however, Module 5 showed high
oxidation levels which indicates there were potentially greater aeration losses from the north and
south leaching core boundaries next to modules 2 and 8. Monitoring along these boundaries
showed generally drier ore conditions. Finally, oxidation outside the leach core was also observed,
indicating lateral movement of solution had leached ore outside of the irrigated core.

The observed variable ore wetting and oxidation indicates zones of low leaching efficiency that are
controlled by the fluid dynamics of the solution and air flow system. Understanding the interaction
of solution and air flow within the leach ore allows for designing and operating the leach pads to
optimize leaching efficiency and copper recovery.

Figure 8 Module average percent of yellow (oxidized) and grey (reduced) ore

Corehole samples selected for hydraulic property testing indicated that ore permeability decreased
with increasing fines and bulk density values. Consolidation-permeability measured Ksat values
showed strong correlation (R2 = 0.76)_ to PSD indicator values (see Milczarek, 2013). This
correlation was used to predict in-situ Ksat values based on the measured PSD indicator values,
and measured bulk density values at each depth interval. As was observed with the measured
data, the predicted Ksat typically decreased with depth and Modules 1, 4 and 7 showed a lower
predicted Ksat than other modules (due to the greatest amount of fines/decrepitation)

CONCLUSIONS
The in-situ monitoring system allowed real-time monitoring of solution content, temperature,
gaseous oxygen content and capillary pressure at various depths and locations in the heap. This
allowed better understanding of the movement of solution and air in the heap leach facility and
ultimately the efficiency of the leaching process. The initial air permeability of the heap was
sufficient to develop high temperatures; however, in-situ gaseous oxygen contents were observed
to be variable and declined rapidly with a subsequent decline in in-situ temperatures. Loss of
aeration efficiency in the leaching core may have occurred due to reduced leach ore air permeability
which caused air to move laterally outside the core where ore conditions were drier and air
permeability greater.
Individual module drainage rates were highly variable and decreased in proximity to the side
where air was injected into the heap. Based on the drainage rate and phreatic level data, it is
believed that aeration back pressure within the drainage layer caused solution mounding within
the leach ore above the drainage rock. During aeration from the east, elevated phreatic levels at the
leach ore/drainage layer interface caused lateral solution flow to down-gradient modules and
reduced drainage from the eastern modules. This occurred most significantly during double grid

east aeration and to a lesser extent during single grid east aeration. Phreatic level changes at the
drainage layer were not large enough to result in solution moving from west to east however, the
movement of aeration to the west from the east reduced aeration back pressure in the eastern
modules and increased drainage from the eastern modules and decreased drainage from the
western modules. There was the potential for intercommunication of module solution resulting
from damaged drainage collection piping or irregular aeration line distribution, however, the
strong correlation of module drainage rates and the heap.
In addition, entrapped air at various locations higher in the leach ore may have resulted in
increased capillary pressure and reduced hydraulic conductivity that resulted in solution
movement as evidenced by low drainage rates from Module 9 and elevated drainage rates in
adjacent modules. Module 9 also showed poor leach recovery below 4 m bgs where the high
capillary pressure conditions were observed. Solution was also observed to move outside of the
leach core as evidenced by the capillary pressure data, high oxygen levels and presence of oxidized
leach ore in the outside coreholes.
In-situ monitoring indicated a solution and air flow system that is highly dynamic,
multidimensional, and significantly influenced by the aeration system. The observed variable ore
wetting and oxidation indicates zones of low leaching efficiency that are controlled by the fluid
dynamics of the solution and air flow system. Data collected from the in-situ monitoring system
allowed for a more complete understanding of the leach recovery dynamics which can be used to
design and operate leach pads to optimize leaching efficiency and copper recovery.

ACKNOWLEDGEMENTS
This work would not have been possible without the support and collaboration over many years
with BHP Billiton, Barrick Gold, Freeport McMoRan Copper and Gold, and Newmont Gold under
which our methods have been developed and tested.

REFERENCES
Bouffard, S.C., and P.G., West-Sells (2009) Hydrodynamic behavior of heap leach piles: Influence of testing scale and
material properties, Hydrometallurgy 98:136142.
Guzman, A., Scheffel, R., Flaherty, S (2008) The fundamental of physical characterization of ore for leach, In
Hydrometallurgy 2009 - Proceedings of the sixth international symposium.
Milczarek, M., M. Banerjee, T-m Yao, J. Keller (2013). Ore permeability methods of evaluation and applications to
heap leach optimization. In Proceedings of the International Heap Leach Conference 2013, Vancouver,
B.C. Canada.

Phenomenological modeling of the

intensive cyanidation of gold and silver
concentrates in continuous reactors

Moiss Miranda
Universidad Nacional San
Agustin de Arequipa, Peru

At the San Jos mine the gold and silver minerals concentrate by floating.
From the concentrates, 60% is commercialized and 40% is leached with
cyanide for the production of dore bars.
The improvement in the quality of the concentrates questioned the
current intensive conditions of the cyanidation process. In order to optimize
the hydrometallurgical processes, in the last year, after the design and
development of an experimental test program, research was aimed at
obtaining a mathematical model for the extraction of gold and silver that
consider the phenomenological interactions of concentrate treatment by
intensive cyanidation and in the continuous reactors.
This paper presents the main results from the stages of the research and
industrial implementation, increased by 2% extraction for gold and 1% for
silver, with decreases in consumption of 28% sodium cyanide and 44% of
hydrogen peroxide. These decreases are important for reducing operational
costs and by the timely deactivation of free cyanide, which stops being a
simultaneous reaction in the cyanidation process, to become as expected,
the main reaction in the cyanide detoxifications step of the processs
tailings before its final disposal.
These results include a number of secondary advantages achieved with
the available resources, avoiding for example, reducing the particle size and
the use of new oxidizing agents. Among these advantages there is pH
stability, lower dissolution of cyanogen (copper and lead), and decreased
retention time in reactors. The actual kinetic validated by the model allows
for parameters to be set based on the quality ranges concentrated without
sacrificing current volume levels or pregnant solution quality. These
improvements are an addition to the efforts of the mining unit to achieve
100% production in dore bars, cyaniding the total of the concentrates.

INTRODUCTION
The technology Gekko Continuous InLine Leach Reactor (C-ILR) of intensive cyanidation consists of a
system which uses high levels of cyanide and an oxidant to leach the gold. It is usually
complemented by electrowinning (EW) producing precipitates to dore smelting. The process of the
case study corresponds to this system. This is ideal for the treatment of larger concentrates
amounts of flotation or gravimetric of low grade. This system is designed for the leaching and
recovery of gold in parallel.
In San Jose are produce an average of 90 tonnes per day of flotation concentrates, usually with gold
and silver grades of 100 g/t and 8000 g/t, respectively. 40% of these concentrates (38 tonnes per day)
are treated by intensive cyanidation. These concentrates are obtained from a granulometry design
of 65 % - m200 in the flotation feed. According to mineralogical studies, the installation of the C-ILR
plant did not consider changes in the granulometry of the concentrates; therefore, the particle size
distribution is unchanged to avoid unnecessary costs for regrinding, and the increase of suspended
solids (TSS) harmful in the EW recovery.
The two problems identified in the development C-ILR process were variability in extraction of
gold and silver, and the additional costs by cyanidation inputs. The first arises from the lack of
knowledge of the physical and chemical phenomena in the process, particularly in this intensive
medium, the second problem arises as a collateral effect of the criteria application on a single
variable to solving the first problem, and by some operational paradigms. Both problems were
considered as cause and effect in optimizing the process.

METHODOLOGY
Ores of gold and silver were identified, scattered as solid solution or as inclusions in pyrite, and to
a lesser extent as electrum (natural gold-silver alloys). The intensive cyanidation of gold and silver
requires maximum contact and optimum presence of reagents inside the reactors. Under these
conditions the Elsner reaction is spontaneous:

4 Au 8CN O 2 2 H 2 O 4 Au (CN ) 2 4OH

(1)

In early 2012, the high consumption of hydrogen peroxide hinted at the possible oxidation of free
cyanide along with the leaching of gold and silver, which would be a competition between the two
spontaneous reactions. The investigations were aimed at finding the best extractions and lower
consumption of reagents that make this possible, considering the phenomena of the extraction
system. (McCabe, Smith & Harriott, 1991)
Table 1 Characterization of the concentrates fed to the C-ILR plant
Au (g/t)

Ag (g/t)

Hg (g/t)

Pb (%)

Zn (%)

Fe (%)

Cu (%)

As (%)

Ins.(%)

-m200 (%)

84.50

7249.49

3.00

1.14

1.16

33.35

0.15

0.05

96.20

65.15

In Figure 1 is shown the flow sheet of intensive cyanidation process of this case, where the
configuration of the four Gekko continuous reactors is observed.
Between the extraction steps (cyanidation) and recovery (electrowinning) is the circuit
countercurrent decantation (CCD), where additional extractions have been shown in some cases.

Gold and silver

C-ILR 1

C-ILR 3

C-ILR 2

C-ILR 4

concentrates

H2 O2

H2 O2

Cyanide
CCD 1
CCD 2
CCD 3
CCD 4

ELECTROWINNING

CCD 5

Barren solution

CYANIDE

DETOXIFICATION

Cyanidation

EW precipitate

tailings

Figure 1 Flow sheet of the process of intensive cyanidation C-ILR

Initially the laboratory research was conducted in a Gekko mini reactor, with a testing program
designed to determine and optimize the influent variables in the process.

Taguchi design for optimization of the intensive cyanidation at laboratory scale

For these tests, the [liquid : solid] mass fraction was two to one (Butt, 2000), the [silver : gold] ratio
in the concentrates was 95 to one, with a residence of 20 hours inside the reactor.
Table 2 Factors and levels for the experimental design of Taguchi
Factor

Description

Plant
2012

Experimental levels
Level 1

Level 2

A
B
C
D
E

Free cyanide concentration (g/l)

Hydrogen peroxide concentration (Kg/t)
Alkalinity (pH)
Reactor rotational speed (rpm)
Silver-equivalent (g/t)

42
45
12
5
9 400

25
25
11
5
9 000

40
50
12.5
10
15 000

The five factors that may be influential in the extraction of gold and silver by intensive cyanidation
of the concentrates are: the initial concentration of cyanide, the ratio of hydrogen peroxide oxidant,
the alkalinity of the system, the rotational speed of the Gekko reactor, and concentrate grade
expressed in terms of silver equivalent. An orthogonal array was used to analyze the effect of the
five factors at two levels each.
Table 3 L8 Orthogonal array for experimental conditions
A

Yexp

1
1
1
1
2
2
2
2

1
1
2
2
1
1
2
2

1
1
2
2
2
2
1
1

1
2
1
2
1
2
1
2

1
2
1
2
2
1
2
1

98
98
96
95
97
97
94
95

Table 4 Analysis of variance and estimated coefficients from the model

Source
Constant
A
B
C
D
E
Residual error
Total

Coefficients
of level 1
96.25
0.50
1.25
0.00
0.00
0.25
-

Degrees of
freedom
1
1
1
1
1
2
7

Sum of
squares (SS)
2.00
12.50
0.00
0.00
0.50
0.50
15.50

adjusted sum
of squares
2.00
12.50
0.00
0.00
0.50
0.50

Adjusted mean
squares
2.00
12.50
0.00
0.00
0.50
0.25

F
factor
8.00
50.00
0.00
0.00
2.00

P
factor
0.106
0.019
1.000
1.000
0.293

The oxidizing effect is greater than the effect of cyanide and the concentrates grade. The alkalinity
and the speed of the reactor between the two levels observed do not influence in the extraction.
Excess hydrogen peroxide oxidizes free cyanide during leaching. The existence of overcosts by
reagents is inferred.
The optimum extraction of silver equivalent is 98.25%, with a coefficient of determination R2 0.968
at level 1 of the factors A, B and E.
Tests Designed to evaluate the effect of the waste that comes from reagents used in the flotation
process, they concluded that these waste do not interfere with cyanide and oxygen diffusion
through the Nerst boundary layer. Therefore washing steps were discarded from the concentrates
before the L8 orthogonal array testing. (Miranda, internal San Jose reports 2012).

Calculating the optimum retention for the population of particles of different sizes
The concentrates and the cyanide oxygenated solution in an alkaline medium (pH = 11.5) are
passed through a C-ILR reactor by plug flow, the concentrations change throughout the reactor
during contact between phases by direct current flow.
The concentrates feeding is constituted by a mixture of different sized particles (for xi=0 to xi=xm).
Keeping in mind that all concentrates remain on the C-ILR continuous reactor during the same time
, and knowing the kinetics for any of the resistors controlling the process, we can calculate the
extraction of gold or silver EAu/Ag(xi) for any fraction xi particle size. Therefore, you can calculate the
average extraction of gold and silver E* from concentrates that leave the reactor adding the
contributions of all size fractions to the overall extraction of particle population. (Levenspiel, 1987)

1 E

*
Au / Ag

xm

1 E

xi 0

Au / Ag

( xi )

F (Fx )
i

0 EAu/Ag 1

(2)

The particle size of the concentrate is between (20-150) m. The null tortuosity and intensive
concentrations of cyanide and oxygen, rule out the possibility that the physical mechanisms of
diffusion could control the kinetics of the cyaniding process. The chemical reaction is the
controlling step. (Smith, 1991)
3

1 E Au / Ag xi 1
t xi Au / Ag

t(xi)

(3)

Where the parameter in question is the retention time of the concentrates in the C-ILR reactor, we
define t(xi) as the time required to achieve maximum extraction of gold and silver in particles of size
xi on the steady state, in other words when EAu/Ag approximates to the unit (to one).
Table 5 Sieve analysis valued of gold and silver concentrates
Tyler
mesh

Mesh
(m)

xi
(m)

Weight
(%)

% Retained
(+)

% Passing
(-)

100
140
200
270
325
400
-400

150
106
75
53
43
38

178.38
126.10
89.16
63.05
47.74
40.42
19.00

6.81
13.61
15.62
13.48
6.50
6.79
37.18

6.81
20.42
36.04
49.52
56.03
62.82

93.19
79.58
63.96
50.48
43.97
37.18

TOTAL

100.00

Grades (g/t)

Distribution (%)

Gold

Silver

Gold

Silver

27.79
36.21
67.88
107.77
112.78
112.69
103.71

3 032.78
3 649.17
6 046.76
8 374.38
8 708.67
9 332.60
8 711.10

2.21
5.76
12.40
17.00
8.58
8.95
45.10

2.86
6.88
13.09
15.65
7.85
8.78
44.89

85.50

7 215.80

100.00

100.00

Cyanidation tests were performed on same size particles (average size) in order to reach the steady
state of gold and silver extractions, from the concentrates of an average size xi to a xi rich solution.
Substituting equation 3 into equation 2 for each one of the particle sizes, we obtain approximates of
optimal retention for all sized particles that remain invariable at the end of the cyaniding process.
The tests were conducted from the results obtained from Taguchi design.

Table 6 Tests Parameters for the optimal retention time

Description
Free cyanide concentration
Hydrogen peroxide concentration
Alkalinity - pH
Reactor rotational speed
Silver-equivalent
Passing 200 mesh

Units
g/l
Kg/t
rpm
g/t
%

Value
30
25
11.5
5
12346
63.96

Proportions greater to 50% of gold and silver are distributed in average size particles less than or
equal to 40 m. The average particle size (60-90) m containing about 30% of the values of gold and
silver. As expected, the coarse fractions will delay the overall kinetic of cyanide in the given
conditions.
Table 7 Calculated extractions for gold and silver from particles classified by sized and cyanided
separately
Tyler
mesh

Mesh
(m)

xi
(m)

Weight
(%)

Time (hours)
EAu(xi)==>1
EAg(xi)==>1

1-EAu

1-Eag

100
140
200
270
325
400
-400

150
106
75
53
43
38

178.38
126.10
89.16
63.05
47.74
40.42
19.00

6.81
13.61
15.62
13.48
6.50
6.79
37.18

33.20
21.10
13.20
8.00
5.25
3.00
2.00

37.30
26.50
18.70
13.25
10.10
8.50
4.00

0.012 77
0.004 76
0.000 00
0
0
0
0

0.011 05
0.006 35
0.000 13
0
0
0
0

100.00

14.20

16.96

0.017 53

0.017 52

0.982 47

0.982 48

TOTAL

Calculated extractions

It is observed that the retention time necessary to achieve estimated extractions by the Taguchi
model (98.25%) are 14 and 17 hours for gold and silver, respectively. It is concluded that the kinetic
of extraction of silver determines the retention time of the overall process of intensive cyanidation.

Experimental profile of cyanidation for the population of concentrate particles

To adjust the calculated retention times, cyanidation test were conducted in a laboratory on the
concentrates, without sorting by particle size.
For these tests the same parameters as shown in TABLE 6 were considered. Kinetics cyanidation of
gold and silver were evaluated, same as major cyanogen elements present in the flow supply
(flotation concentrates).
It is observed that the extraction of gold and silver on the population of multi-sized particles are
very close to those calculated from individual cyanidation of average sizes. The kinetics is
presented numerically in TABLE 8, and graphically in Figure 2.

Table 8 Population Kinetics of intensive cyanidation

Time
(hours)

Gold
(%)

Silver
(%)

Remaining free
cyanide (%)

Copper
(%)

Lead
(%)

Cyanide
consumption (%)

15

98.00

97.28

85.78

23.37

8.00

14.22

16

98.20

97.81

85.33

23.60

12.00

14.67

17

98.20

98.33

84.79

24.00

17.00

15.21

18

98.20

98.33

84.73

24.66

23.00

15.27

Extraction of gold reached a steady state at 16 hours of processing, exceeding 1.8 h with respect to
the information obtained in the previous tests. The extraction of silver offers greater accuracy in
relation to the preliminary results.
Lead and copper present in the pregnant solution is equivalent to 0.23 g/l and 1.62 g/l, respectively.
These levels of impurities can be controlled easily during the EW recovery stage.
100

Percentage

80

Gold e xtra ction

Silve r e xtra ction
Re ma ining fre e cya nide
Coppe r e xtra ction
Le a d e xtra ction
Cya nide consumption

60

40

20

time (hours)

12

15

18

Figure 2 Population Kinetics of intensive cyanidation

Intensive cyanidation tests at industrial stage

So far, the results of intensive cyanidation tests for certain particle sizes xi allowed to integrate the
kinetic model for the particle population of multiple sizes. This model has been experimentally
confirmed, at the same time the leaching of cyanogen elements that could affect the quality of EW
precipitate was evaluated.

Final testing conducted at industrial level evidences the optimization of the intensive cyanidation.
The improved results of extraction are shown in Figure 3.

Figure 3 Percentage of extraction at industrial level before and after the improvements

Model for the extraction of silver equivalent

From the quality and granulometric characterization of the concentrates, we substitute equation 3 in
equation 2 in terms of fractions extracted from the silver equivalent.

*
Ag Equiv

3
3
xm
xm
F ( xi )
F ( xi )
0.991 1
0.011 1

xi 0 t xi Ag F
xi 0 t xi Au F

(4)

To:

80 ( Ag : Au ) 120
7000 Ag ( g / t ) 8500
90 x i m 180

( hours ) 17.5
The kinetic model of equation 4 relates to the influence of the reactants through the following
population model:
*

E Ag
Equiv 0.9625 0.005 (CN ) 0.0125 ( H 2 O 2 ) 0.0025 ( Ag Equiv .)

(5)

From the calculated retention ( =17 hours) the capacity of the reactors was optimized by an
adjustment of the feed rates. TABLE 9 shows these changes.

RESULTS AND DISCUSSION

Table 9 Optimization of operational parameters in the C-ILR plant
Parameter
Dry tons per day (t/day)
Specific gravity
Pulp density (g/l)
Percentage of solids (%)
Solution flow rate (m3/day)
Mass flow of pulp (m3/day)
Pulp flow (m3/day)
Retention time (hours)
Free cyanide concentration (g/m3)
Hydrogen peroxide (Kg/t)
Alkalinity - pH
Silver equivalent in concentrates (g/t)
Ratio [silver : gold] in the concentrates
Extraction of silver equivalent (%)

January 2012

140

July 2012

Currently
48.0
4.0
1400
35.0
78.0
126.0
90.0
17.2
30
25
11.5
12 000
85 a 1 - 100 a 1
98.35

January 2013

Tota l cost C-ILR (% e xce ss)

Silve r e quiva le nt (% e xtra ction)
Tonna ge (% ca pa bility)
EW pre cipita te (% production)

130

Percentage

Until August 2012

38.4
4.0
1400
35.0
62.4
100.8
72.0
25
42
45
12
9 400
50 a 1 - 200 a 1
96.80

120

110

100

90

100

11

12

Year (00)

13

Figure 4 Evolution of operational costs and performance of the industrial process C-ILR

In Figure 4 shows that the improvement opportunity arises from increasing the quality of the
concentrates, in the second quarter of 2012.
The free cyanide consumption after 17.5 hours of process reaches a rate of 15.10%, equivalent to an
effective consumption of 14 kilogram per tonne of dry concentrate.
The alkalinity of the process (pH = 11.50) represents a sodium hydroxide consumption 3 kg/t, for
dilution conditions of pulp concentrate (liquid : solid = 2).
The concentration of dissolved oxygen inside the reactors reaches values of 14 mg/l from a dosage
of hydrogen peroxide equivalent to 25 kg/t.

CONCLUSIONS
The design of the phenomenological model of intensive cyanidation process (C- ILR) has
allowed identifying additional costs by reagents consumption and the opportunity to
significantly improve the efficiency of the process, by increasing the capacity of the treatment
and metallurgical extractions.
The optimizing of the C-ILR process has reduced consumption of sodium cyanide in 28.57%,
hydrogen peroxide in 44.40 % and sodium hydroxide at 18.75%. This is reflected in the reduction
of 24.32% of total operating costs, from August 2012.
The Greater knowledge of the kinetics of intensive cyanidation has allowed redesigning the
operating parameters to increase the capability by 26% in tonnage and 29.85% in metal content,
with 17.5 hours of average retention time.
The improved efficiency of the process showed extraction increases of 2.10% in gold and 1.31%
in silver. Simultaneously decreasing lead and copper present in the pregnant solution at 60.11%
and 22.89%, respectively. This also means better performance in the recovery by electrowinning
and in the dore quality.
The lower current rate of cyanide adds value to commitments with safety, occupational health
and environmental compromises for the company, ensuring the stability of free cyanide with an
economic alkalinity (pH = 11). These measures also reduce the costs of detoxification of cyanide
before the final disposal of the tailings.

REFERENCES
Butt John B., REACTION KINETICS AND REACTOR DESIGN, (2000) Marcel Dekker, pp. 592-599.
Levenspiel Octave, CHEMICAL REACTION ENGINEERING, (1987) Editions Repla, pp. 417-429.
McCabe Warren L., Smith Julian C., Harriott Peter, UNIT OPERATIONS OF CHEMICAL ENGINEERING,
(1991) McGraw Hill., pp. 616-627.
Smith J. M., CHEMICAL ENGINEERING KINETICS, (1991) Editorial Continental, pp. 737-740.

chapter 7
Water accounting and
water footprint

Sump Box for retaining organic entrainments in

electro-winning cells

Rodrigo Villarroel,
Fernando Penna,
Cristin Villaseca and
Hans Tippmann
Tecnocomposites, Chile

The electro-winning process for copper and other base metals require
previous stages to enrich the electrolyte entering into the electrolytic cells
to deposit metal by electrolysis. The elements used to generate the
enrichment of the electrolyte are organic chemical reactives that promote
selective metal separation. One of these is the extractant, a high cost
reagent that process plants are willing to recover for regeneration and reuse
in the process, so there is a special care to control losses. The organic
reagent entrainments occur permanently by losses during the solvent
extraction stage and by failures in the control process. Non controlled
entrainments go to electro-winning process up to electrolytic cells where
they burn and deteriorate, becoming unusable and, eventually,
contaminating the organic stocks. In addition, it generates problems in the
metal deposit, especially in the liquid-air interphase, the cathode top
border, obtaining a lower quality cathode that frequently is washed with
pressurized hot water. In some cases the contaminated top border is cut
off, with an important loss of metal.
To solve the problem of the organic entrainments, the authors of this
paper recommend applying the LateralFlow technology to quickly remove
this floating element from the surface of the electrolyte of each electrowinning cell, carrying it to the overflow before it burns and deteriorates.
Nevertheless, many process plants use scavenger cells as retainers for these
floating elements, and prefer to avoid re-entering these into the electrolyte
cycle. For this purpose, the LateralFlow technology is complemented with
a special Sump Box, a device which temporarily retains within it the
floating elements, as organic entrainments evacuated from the surface of
the electrolyte in the cell by the action of the LateralFlow, preventing that
this organic material continues to the electrolyte cycle, from where an
operator removes it manually or using a suction pump, or it is derived to an
alternative overflow trap.

INTRODUCTION
The electro-winning process for the production of copper and other deposited base metals implies
the use of refining plants, or else, plants with electro-winning tank-houses used jointly with solvent
extraction plants, known as SX-EW (ElectroWinning - Solvent eXtraction).
The solvent extraction process plant consists in large pools or tanks where the leaching solution
(originating from the leaching heaps) comes into contact with a non-miscible organic reactant and
extracts the copper through a selective process, to subsequently transfer it to a concentrated
electrolyte flow, in a re-extraction or stripping phase, obtaining what is called a rich electrolyte.
During this stage, organic entrainment occurs in the aqueous phase of the rich electrolyte, which is
a normal situation that the industry has addressed by incorporating filters to prevent them from
entering the E-W tank-house.
On the other hand, emphasis should be given to the very detrimental effect of organic entrainment
entering the electro-winning cells, as it may even become a critical situation for the normal
operation of the tank-house, depending on the level of contamination.
During a plants normal operation, the organic entrainment that reaches the tank-house after
having gone through the filters, usually fluctuates between 30 and 75 parts per million (ppm)
(Taylor, Alan (1995)), and there are plants that manage to control entrainment at values equal to or
less than 1 ppm. It is difficult to collect conclusive information regarding the level of organic
entrainment in the rich electrolyte, so the range of values presented in this document are for
reference, and probably conservative.
This constant presence of organic causes the following problems (Dr. Vargas C., (2006)):

Contamination of the cathodes obtained through electro-winning

Contamination of plastic balls and other elements generally used inside the cell

Loss of efficiency in the current transmission due to dirtying of equipotential busbars

Depositing and adherence problems due to soiling of permanent cathodes when blanks are
being replaced in the cells

Agglomeration of other impurities

Organic accumulation, deterioration and burning due to exposure to higher temperatures

inside the cell

Creation of a highly flammable crust over equipment and electrodes, with the risk of
ignition

Re-processing of contaminated cathodes

Higher water consumption for cathode washing.

These problems create quite onerous financial costs to EW plants. The presence of great quantities
of organic entrainments that occasionally occur, has much more serious effects, not only
endangering production but also the integrity of the whole tank-house. This is a specific situation
and, as mentioned earlier, is caused by operational problems in SX where, because of the difficulty
to control tank levels, a significant proportion of the total organic stock is subject to entrainment.
This results in most of the cells having an important organic stock, which entails an extremely high
risk of ignition and fire that may extend to the entire tank-house. In practice, these specific
entrainment events actually happen, and the action normally taken is to cut off the power to avoid

ignition sources and, on the other hand, to wash all electrodes, cell by cell, after having manually
removed the organic. There are times when the whole inventory of deposited copper may be lost.
Sometimes, the design of SX-EW plants considers including cleaning cells, called scavenger cells
or sacrificial cells that receive, from SX, electrolyte portions that usually present greater quantities
of organic, in the understanding and under the implicit assumption that it is a problem and that the
solution is not to increase the number of filters.
In plants that have scavenger cells, in numbers that usually correspond to 1/5 of the total number of
cells, organic concentration increases to tipically ranges between 10 and 50 ppm, and, on the other
hand, organic concentration on other cells, frequently called commercial cells, decreases. The
name implies that copper obtained from scavenger cells will be lower in price due to their quality,
unless additional measures are taken, such as the permanent removal of organic from these cells,
the use of an agitator tub with hot water for better washing of plates, cutting off the top border, or
others, and this portion of the production may be set apart from the commercial standard achieved
by the rest of the plant.
Considering this situation, Tecnocomposites developed and patented a product denominated
LateralFlow (Penna, F., Villaseca, C. & Villarroel, R. (2011)) consisting in a couple of plenums
mounted on the cells, one on each side. These plenums have perforations in their upper part that
allow suctioning the acid mist confined under flexible covers. In the lower part of each plenum,
LateralFlow incorporates a gutter, which extends facing every electrolyte, and through which the
electrolyte flows toward the drainage of the cell. Thus, the electrolyte that is injected and
distributed in the cell is forced to reach the surface and move perpendicularly to the axis of the cell,
in between the electrodes, generating a strong surface flow until it falls into the lateral gutters,
where there is a longitudinal flood-gate with an adjustable height that allows regulating the filling
level of the cell and with that, the upper part of the cathode deposit.
In the end where the cells drainage is located, a Modified Sump Box (Penna, F. (2013)) is
installed, which is a device manufactured in FRP, and located in the space between the wall of the
cell and the first anode, occupying the width of the cell and all of the available depth, and has the
following functions:

To prevent the cells electrolyte from overflowing directly into the drainage.

To receive the discharge from both gutters.

To deliver the lean electrolyte into the drainage.

To retain the organic recovered in the cell.

Cells without LateralFlow usually have a deflector at the end of the cell with the drainage,
generally used to contain spheres used as barriers to prevent the emanation of acid mist. Thus, the
free surface of the electrolyte in the cell, which contains the organic, does not have any point from
where the floating elements may evacuate or overflow. And if the design of the cell had a screen
instead of a deflector, to retain the spheres and yet allow the fluid to go through, only the organic
accumulated in the vicinity of the drainage would be evacuated, as the electrodes act as a barrier
that make it difficult for the floating elements to move along the cell. On the other hand, this
organic accumulation and the higher temperature achieved in the electrodes, promotes the
evaporation of the lighter organic fractions, mainly the extractant. This situation is known as
deterioration of the organic, and in practice it is addressed by removing manually the
deteriorated organic, as the plants do not have the infrastructure to treat and recover this element.

Thus, a drop of organic that appears in the surface of the cell due to its lower density, should take a
maximum of 10 seconds in overflowing into one of the two gutters of the LateralFlow, depending
on the geometry and the flows of the cell. It should be noted that the organic that enters the cell has
a low viscosity, contrary to when it has been deteriorated and burnt, when it turns into a brown or
black crust. Thus, cathodes in cells with LateralFlow are not affected by the presence of organic.
This matter is extremely important, as it promoted the fact that in a subsequent development, the
way to retain the organic recovered from the cell was addressed, with the clear objective of
reutilizing it. In plants where the flow of lean electrolyte mixes with the rich electrolyte fed to
commercial cells, this is very valuable, as the standard solution of LateralFlow removes the
organic from the cell very quickly, but it does not eliminate it from the flow.
The need arises, then, to incorporate elements to the sump box that will allow separating and
retaining the organic, bearing in mind that there are severe space limitations and extremely low
residence times. The latter is very relevant, as the traditional separation processes (API Publication
421, (1990)) of non-miscible phases are grounded in Stokes Law (Lamb, H. (1994)), which identifies
the force with which a drop may ascend as a function of the difference of densities between the
phases. This implies that for the separation to be effective there must be speeds and spaces that,
considering other factors such as viscosity, (which depends on the temperature) allow the
separation to take place.
The following Modified Sump Box, in contrast with the standard sump box, has inner gutters with
deflectors that promote the separated organic to be carried to a low agitation area, where the
organic is accumulated and retained, awaiting its manual removal.
Aiming to solve this problem, several prototypes were developed using various geometries in order
to achieve two essential objectives:

Recovery: to enable the organic to separate and float, generating a continuous phase on the
inner surface of the Modified Sump Box.

Retention: to have the separated organic remaining on the surface of a confined space and
prevent it from being entrained by the electrolyte flow, throughout time, out of the
Modified Sump Box, promoting an increasing and continuous accumulation.

Additionally, the design considers finding a way whereby, either through manual or automatic
operation, said organic may be removed for its subsequent reutilization.
The objective of this paper is to present the results obtained in a real scale experiment regarding the
performance of the Modified Sump Box, under two scenarios: organic entrainment at different
electrolyte flows, and variations in organic concentrations at the same flow rate. All of the above
assayed on a specific Modified Sump Box model.

METHODOLOGY
The assays were carried out in the production installations of the company. A commercial-sized cell
was used its general dimensions being 6.2 meters long, 1.45 meters deep, and 1.4 meters wide,
where a standard electrolyte distribution system was installed, consisting of a perforated 75 mm
PVC piping ring, fed by a centrifugal pump and an 8 cubic meters recirculation tank with oleophilic
fabric to retain the organic before recirculation. A set of LateralFlow plenums was also installed
and a Modified Sump Box measuring 1290 mm wide, 1200 mm deep and 150 mm long, which was
used for the assays whose results are presented herein. No electrodes were installed because it was

deemed that they do not affect the results of the sump box, and also to be able to visualize the
phenomenon throughout the cell.

Figure 1 Modified Sump Box with low agitation area (A), with accumulated organic (blue-colored paraffin)

For reasons of infrastructure and security, assays were carried out with drinking water (density: 1
gr/ml), to simulate the electrolyte, at ambient temperature (20 C), without sulfuric acid and
without dissolved copper. For the organic phase, commercial kerosene or paraffin (density: 0.81
gr/ml) was used, equivalent to the solvent that is generally used in the industry.
For the assays, paraffin was added directly to the cell, at the opposite end of the drainage and the
sump box. This addition was done in a constant and regular trickle, achieved by adjusting a smalldiameter valve and measuring a volume equivalent to 10% of the total volume added per assay,
during a proportional period.
The trickle fell directly onto the surface of the liquid in the cell, at the center of the surfaces third
portion farthest away from the drainage.
To remove the paraffin (organic phase) that had been separated and retained in the Modified Sump
Box, the floating elements were then removed introducing containers in such a way that the
overflow fell into them. Subsequently, the water and paraffin mixture was poured to a decanting
cone, to wait until the total separation of phases was complete, with the draining of the aqueous
phase and the recovery of the organic phase.

Figure 2 Decanting cone during the process of separation of phases.

Recovery assays using constant addition of organic and variable flow of water:
Four assays were carried out in which, during a period of 2 hours, 1500 ml of paraffin were added
through a constant and homogeneous trickle. Four cell-feeding flows were assayed within the usual
range used by the industry, which varies between 10 and 25 m3/h.

Recovery assays using constant water flow and variable organic concentration:
Three assays were carried out in which, during a period of 2 hours, different quantities of organic
(paraffin) were added, with a 10.5 m3/h water flow to the cell.

RESULTS AND DISCUSSION

The organic phase recovery assays with constant organic addition and variable water flow to the
cell, provided the results shown in the following TABLE 1 and in graphic form in Figure 3.
Table 1 Results of assays with variable flow
Flow of Water

Paraffin added Time of trickle Concentration

(m /h)
21,6
17,7
14,1
10,5

(l)
1,5
1,5
1,5
1,5

(h)
2
2
2
2

(ppm)
34,7
42,4
53,2
71,4

Recovered

Recovery

Volume (l)
0,32
0,38
0,44
0,58

(%)
21%
25%
29%
39%

Recovery at Constant Addition

45%
40%

Recovery (%)

35%
30%
25%
20%
15%
10%
5%
0%
0

10

15

Cell Flow (m3/h)

20

25

Figure 3 Chart of results shown in Table 1

Results for assays carried out with constant flow (10.5 m3/h) and variable organic phase, are shown
in TABLE 2 and in graphic for in Figure 4.

Table 2 Recovery with constant flow

Flow of Water
3

(m /h)
10,5
10,5
10,5

Paraffin added Time of trickle Concentration

(l)
0,6
1,5
2

(h)
2
2
1

(ppm)
28,6
71,4
190,5

Recovered

Recovery

Volume (l)
0,21
0,58
0,57

(%)
35%
39%
29%

Recovery at Constant Flow

45%
40%

Recovery (%)

35%
30%
25%
20%
15%
10%
5%
0%
0

50

100

150

200

Concentration (PPM)
Figure 4 Chart of results shown in Table 2

The potential of one drop of paraffin to float on an aqueous media (Stokes Law, equation 1), is
directly related to the difference in densities; in the situation assayed, the densities were: 1 gr/ml in
the case of water and 0.82 gr/ml in the case of paraffin. For an industrial scenario, the density of the
organic phase remains at similar values, but the aqueous media increases its density from 1 gr/ml to
values close to 1.28 gr/ml (usual for electrolytes with sulfuric acid in concentrations of 180 gr/l),
thus increasing the gap between phase densities to approximately twice.
Equation 1: Stokes Law for speed of ascent at low speeds (API Publication 421, (1990))

Where:
Vs is the ascent speed of the lower density bubbles (speed limit)
g is the gravitational acceleration,
p is the density of the bubble
f is the density of the fluid
is the viscosity of the fluid
r is the equivalent radius of the bubble.

CONCLUSIONS
From a hydrodynamic perspective, recovery of the organic phase in the Modified Sump Box shows
interesting results, bearing in mind that there are serious space limitations for the application of a
design based on low speeds that will promote the coalescence and flotation of the organic phase
(API Publication 421, (1990)).
Density of the aqueous media corresponds to water, which is around 1 gr/ml, while the density of
the electrolyte is around 1.28 gr/ml, depending mainly on the concentration of sulfuric acid. This
allows to extrapolate better results and to expect that in an industrial application with electrolyte,
the performance will be notably higher, due to the fact that difference in densities is more than
double the one used in the assays presented here.
The inverse relation between the feeding flow to the cell and the recovery of organic is consistent
with the theory, because at greater flows, the speeds inside the Modified Sump Box increase and
the probability of achieving separation and flotation decreases.
The relation between organic concentration and recovery with constant flow shows an unexpected
break in the tendency, as it was to be expected that at a higher organic concentration, recovery
would increase.
In a standard industrial scenario, the higher temperature in the electrolyte (around 45 C) decreases
viscosity and promotes better separation and flotation (O. Benavente, M.C. Hernndez y F. Sagredo
(2013)).
On the other hand, and from the point of view of production, incorporating a device like the sump
box with organic retention, together with the LateralFlow, decreases the reincorporation of this
element to the return of the electrolyte flow, which, with the preliminary design presented in this
paper, fluctuates between values of 30% and 40%, which results in cleaner production and less
subsequent cleaning operations

NEXT STEPS
Having proved that the Modified Sump Box and its separation and retention system functions,
now we expect to implement new modifications to its inner design, always seeking to optimize the
process.
Evaluate the performance of this Modified Sump Box in real operating conditions using electrolyte
and organic of common use in the industry, to determine the real results, which will allow to
correlate future assays to new versions of the sump box.
Carry out an assessment of the quality of the recovered organic, in order to verify the level of
eventual deterioration it may have, and thus, plan the management of the recovered organic.

REFERENCES
1.- Penna, F., Villaseca, C. & Villarroel, R. (2011) LateralFlow : A concept to control acid mist, organic
removal, height of deposition and other benefits for electro-winning cells Tecnocomposites S.A., 6th
International Seminar on Copper Hydrometallurgy: HydroCopper 2011, 6 al 8 Julio 2011, Via del
Mar, Chile.
2.- Dr. Vargas C., (2006) Curso de capacitacin operacional de plantas de electro obtencin, Departamento
de Ingeniera Qumica, Universidad de Santiago de Chile.

3.- Penna, F. (2013) Solicitud de Patente Modelo de Utilidad N 201303184, Chile.

4.- Lamb, H. (1994). Hydrodynamics (6th edition ed.). Cambridge University Press.
5.- API Publication 421, (1990) Monographs on Refinery Environmental Control Management of Water
Discharges. Design and Operation of Oil-Water Separators.
6.- Taylor, Alan (1995). Copper SX/EW: Basic principles & Plant Design. ALTA press.
7.- O. Benavente, M.C. Hernndez y F. Sagredo (2013). Estimation and modeling of organic entrainment in
aqueous solution (O/A) in laboratory tests of copper solvent extraction (SX) plant.

Failure mechanisms of lead electrowinning anodes

Abbas Mirza, Matt Burr and

Timothy Ellis
rsr Anode Group and
rsr Technologies, usa
Larry Webb and
Ricardo Cabrejas
rsr Anode Group and
Quemetco Metals
Limited, Inc., usa
Eben Lombard
rsr Anode Group and Castle
Lead Works,
South Africa
Andrew Johnston and
Franois Leclercq
rsr Anode Group and
Le Plomb Francais, France

Lead-based alloys are the primary anodes used in electrowinning from

sulfate-based aqueous systems. Lead anode technology has evolved over the
years migrating from pure lead and lead-antimonial alloys to the present
day lead-calcium-tin alloys for copper electrowinning and lead-silver alloys
for zinc electrowinning. It has also migrated from cast to mechanically
deformed, rolled, microstructures in search of improved mechanical
properties and higher corrosion resistance because a mechanically
deformed microstructure rapidly produces a protective oxide coating.
Although great strides have been made in the development of new alloys
and production processes, the industry still has unresolved issues related
to untimely corrosion which limits anode life and may lead to higher
contaminant levels in the metal being produced. Lead anodes corrode
because of the difference in the chemical/electrochemical potential across
the microstructural features of an anode. Given a very high purity material
we find that the grain boundary areas corrode significantly faster than the
rest of the grain. The desire to minimize grain boundary corrosion by a
balance of alloying element selection and microstructural design allows
the grain boundary area to be engineered. However, operational issues can
lead to unexpected corrosion behaviors which we will discuss moving
forward. The comments in this work, though directed toward copper
electrowinning, can be extended to similar phenomena in other metals
electrowon from sulfate media (i.e. zinc, nickel, cobalt, manganese). The
lifecycle of electrowinning anodes is dependent upon electrowinning
tankhouse operating conditions and maintenance of the anodes including
cleaning and straightening. This paper will focus on the operational
aspects of maximizing the utilization of lead electrowinning anodes.

INTRODUCTION
Lead-based alloys are the primary anodes used in electrowinning from sulfate-based aqueous
systems (Prengaman, Ellis & Mirza, 2010) and anode technology has migrated from cast
microstructure to mechanically deformed, rolled, microstructures (Prengaman & Morgan, 1992,
2001; Prengaman, 1987, 2000). The fundamental efficacy of lead-based anodes is dependent on the
properties of PbO2 (Prengaman & McDonald, 1980). In fact, lead anodes are used predominantly in
sufate-based electrolyte systems because of the protective ability of PbO 2. In particular, the
protective, adherent -phase, which is the brown PbO2 layer closest to the anode surface, isolates
the lead from the corrosive electrolyte. It is composed of rhombic, large, closely packed crystals.
The -phase of PbO2 which is black in color is the interface between the protective -PbO2 and the
electrolyte. The mechanism of electrochemical oxidation of lead to lead dioxide in sulfuric acid is
discussed in detail by Pavlov and Dinev (Pavlov & Dinev, 1980).
The oxidation of a lead anode is therefore determined not by the reaction rate of lead with sulfuric
acid but by the mass transport of the ions participating in the reaction through the electrochemical
layer. The rate of oxidation decreases drastically with the increase in the thickness of the layer and
is prevented almost completely if the corrosion layer does not dissolve, flake-off or become porous
(Hyvarinen, 1972). As illustrated in Figure 1, it is necessary to maintain an anode potential above
1.77V to maintain the protective -PbO2. Unless the anode potential is consistently maintained
above this value, the protective, adherent -PbO2 converts to the loosely adherent -PbO2 that
flakes off and causes cathode contamination.
Electrode Potential
(Volts, SHE)

Metal

Interface
-PbO2 / PbO2

1.77

Electrolyte/Anode
Interface
-PbO2 / -PbO2

-PbO2
1.69

-PbO2 / t-PbO

PbSO4 / -PbO2

Pb
1.18

t-PbO

0.25

Pb(OH)2

0.00

PbO / PbSO4

-0.30

PbSO4

PbSO4

Figure 1 Schematic representation of lead corrosion films (adapted from Burbank 1956)

CORROSION OF LEAD ANODES

General corrosion
When a lead-based anode is exposed to a sulfate-based electrolyte, a thin film of PbSO4 is formed
instantly. As soon as power is switched on, oxygen is evolved at the anode. The thin film of lead

sulfate is then converted to -PbO2. Upon further oxygen evolution, oxygen diffuses through PbO2 to form -PbO2.
To produce the lowest lead content in the copper cathodes, lead anodes should be cleaned
periodically to remove the loosely adherent portion of -PbO2. The objective is not to remove all the
-PbO2 layers, but only the non-adhering layers with a pressure water wash. The typical surface of
a cleaned lead anode in use with the adherent -PbO2 is shown in Figure 2.

Figure 2 Hard, dense, adherent -PbO2 (brownish color)

Chloride corrosion
While 20-30 ppm Cl- in a copper electrowinning electrolyte is helpful for grain refining/leveling of
the cathode deposits, higher chloride concentrations (>30 ppm) increase corrosion of the Pb anodes
due to the formation of soluble PbCl2 and other Pb-chloride complexes. A steady state
concentration of 55-60 ppm chloride in the electrolyte is too high, and a concentration above 100
ppm Cl- in the electrolyte is highly deleterious for the anodes. Chloride corrosion is highest at or
above the solution line because the higher oxygen concentration at the surface enhances the
formation of Pb-chloride complexes. Organic burn, discussed later in this paper, exacerbates the
solution line chloride corrosion. In order to increase the functional life of lead anodes, it is
imperative to keep chlorides levels in the copper electrowinning electrolyte to 20 30 ppm.

MnO4- ion corrosion

Copper electrowinning solutions often contain some manganese. Problems related to manganese in
the electrolyte have been discussed by Miller (1995) and Hughes et al (1998). Anode cleaning is very
important to remove these layers to maintain low power consumption. Cifuentes et al (2005)
investigated the effect of various impurities in a copper electrowinning electrolyte on the rate of
corrosion of lead anodes and found that manganese enhanced the corrosion rate of lead anodes
during current interruptions and precipitate significantly on the anode surface.
Manganese reacts with stable PbO2 corrosion layer on the anode surface as illustrated in Figure 3
making the PbO2 layer unstable. As the PbSO4 is converted back to PbO2 but is now deposited as

larger, softer, lighter loose layers, these layers spall off easily causing contamination of the copper
cathode. It is important to control the manganese content of the electrolyte to < 40 ppm. Addition of
100-200 ppm cobalt to the electrolyte helps in minimizing the deleterious effects of manganese
corrosion. Other measures to control manganese in copper electrowinning electrolyte are discussed
in detail by Miller (2010).

Figure 3 Disruption of PbO2 by MnO2 and formation of large, soft, loosely adherent layers

Addition of cobalt
The main purpose of cobalt addition to the copper electrowinning electrolyte is to depolarize the
oxygen evolution reaction on a lead anode. This decreases the oxygen evolution overpotential, and
results in savings in power consumption. Cobalt also protects the lead anode from corrosion and
the solution, i.e. cathode copper from lead contamination. The oxide layer in the presence of cobalt
is very thin and dense. The optimum level of cobalt in the electrolyte should be determined
experimentally for a particular tankhouse. Once determined, care should be taken to ensure that the
cobalt content of the electrolyte is maintained at the specified level particularly in the presence of
manganese in solution.

Hanger bar corrosion

The main reasons for hanger bar corrosion are:
a.

Intimate contact of the electrolyte with the hanger bar: The electrolyte level is sometimes raised
to overflow the electrolyte in order to float off the entrained organics and mitigate organic burn
issues. Electrolyte can also seep up to the hanger bars through the mist control foam or plastic
balls by capillary action

b.

Reaction of precipitated copper sulfate with protective lead.

c.

Penetration of the electrolyte between the copper hanger bar and the cast-around.

Electrolyte-hanger bar contact

If electrolyte is allowed to come in direct contact with hanger bars as shown in Figure 4, copper will
react spontaneously with H2SO4 to form CuSO4.5H2O. To increase the functional life of lead anodes,

it is therefore important not to immerse the hanger bars in the electrolyte and not to allow the level
of plastic balls or the mist suppressant foam to be within 3-4 of the hanger bars.

Figure 4 Corrosion of hanger bar due to immersion or seepage

Corrosion by copper sulfate

In covered copper tankhouses, acid rains down constantly on the anodes. As shown in Figure 5,
copper reacts with sulfuric acid to form cupric sulfate (CuSO 4.5H2O) which reacts spontaneously
with lead to form cuprous sulfate (Cu2(SO4)2).

Figure 5 Corrosion of lead by CuSO4:5H2O

Electrolyte penetration
This type of corrosion is illustrated in Figure 6 and is most prevalent in covered copper tankhouses.
Corrosion penetrates into the area between the lead covering the bar and the copper bar creating a
significant amount of copper sulfate.
In some cases the plastic can contact the cathode as it is not only a lifting device, but also a spacer.
When the plastic is coated with acid and it contacts the cathode, a small current can be generated
between the anode and cathode along the surface of the plastic lifting device.

Figure 6 Corrosion of hanger bar by acid penetration or corrosion cells

Solution line corrosion: Organic burn

Although animal glue and thiourea are used as leveling agent and grain refiner, respectively, in
copper electrorefining, they are not typically used in copper electrowinning because of their
deleterious effect on the organic phase in solvent extraction (Stantke, 1999). Therefore, guar in
concentrations of 1-5 ppm is used as a leveling agent in copper electrowinning. The degraded
derivative of guar can accumulate in the electrolyte. More importantly, there can be carryover of
organics (S/X reagent and diluent) into the electrolyte from solvent extraction. A typical example of
organic entrainment in electrolyte and solution corrosion is given in Figure 7.

Figure 7 Entrained organics on electrolyte surface

Solution line corrosion is often mistaken as weld corrosion where the weld line in welded anodes
corresponds with the solution level. Solution line corrosion is observed on rolled/slotted and cast
anodes where a weld line is absent. Since no alloy currently exists that can protect the anode, the
only suitable action is to prevent organics from entering the tankhouse. In order to minimize
organic burn, the organics from the solvent extraction process must be removed from the electrolyte
before it reaches the tankhouse. Column flotation, sand filters or scavenger cells can be used for this
purpose (Kordosky, 2002). Electrowinning cells should be split into scavenger and production cells

to minimize organic carryover from the Solvent Extraction Purification Plant (Beukes & Badenhorst,
2009). Improved mixer-settler design enhances the distribution of the dispersion exiting the mixer
and minimizes organic entrainment. In the case of settlers it is essential to provide for sufficient
floatation time to float the organics. Other means employed on the tankhouses include scrubbing
and scavenger cells as well as surface vacuuming. None of these methods are ideal as anodes will
be exposed to organic attack, even if it is only the anodes in the scavenging section of the
tankhouse.

Uneven electrode spacing

The electrode spacing should be as consistent as possible. Figure 8 indicates that this is not always
the case in a tankhouse. The inter-electrode spacing in copper tankhouses is usually about 4 (10
cm). Below this spacing, the risk of shorts due to electrode alignment and dendritic growth of
nodules becomes unacceptable. If a dendrite touches the anode, the protective PbO2 gets
overheated, is softened and spalls off under conditions of intense turbulence caused by oxygen
evolution. This process is autocatalytic and blows holes through the anode. Therefore, to increase
the useful life of lead anode and to decrease power consumption, it is important to maintain
uniform anode-cathode spacing and monitor and eliminate the shorts as soon as they are formed.

Progressive evolution of anode design

Anode design is a crucial yet often overlooked or undervalued aspect of anode life optimization.
Anode design is an important consideration in cathode plating efficiency. The anode hanger bar
was moved closer to the solution line in an effort to minimize the size of the lead anode. When this
is not taken into consideration with the design of the anode hanger bar, various modes of anode
damage follows owing to both corrosion and mechanical mechanisms.

Figure 8 Uneven electrode spacing

Corrosion stems from the exposed copper bars proximity to the electrolyte level as well as to the
presence of covers for acid mist control. With effective design, protection is extended to protect
such vulnerable areas of the hanger bar. Figure 9 demonstrates the evolution of anode hanger bar
encapsulation design from left to right.

Figure 9 Progressive evolution in hanger bar design

Copper plating on lead anodes

Strangely enough, copper sometimes deposits of lead anodes. What could possibly contribute to
this? It has been observed that this happens in two cases: (a) when there are dendrites shorting the
anode to the cathode and (b) when there is a power-off situation and the anodes are sitting in the
tankhouse electrically connected to the cathodes.

Dendrites
The local area around the region where the dendrite touches the anode will become cathodic and
copper will therefore deposit on the anode in this area. More importantly however, the protective
-PbO2 in the region around the dendrite will be reduced to PbSO4 which will flake off under
conditions of intense oxygen evolution in the surrounding areas. This autocatalytic process will
continue unabated until the dendrite blows a hole through the anode.

Power outage
Copper also plates on the lead anode as shown in Figure 10 if there is a power outage and if the
anodes and cathodes are left electrically connected in the tankhouse electrolyte. It is also found to
be a very common problem in cast Pb-Sb anodes compared to rolled Pb-Ca-Sn anodes. This is
because cast anodes have segregation and casting defects such as porosities, inclusions which lead
to non-uniform electrochemistry. As a result there are areas in a cast anode which become relatively
anodic to other areas more easily. On the other hand, rolled microstructure is very uniform with no
casting defects which results in uniform electrochemical behavior across its surface.

Figure 10 Copper plating on lead anode (With permission to reproduce from Nikkelwerk, 1/26/2014)

SUMMARY
The different types of corrosion processes affecting lead anodes discussed above are by no means
exhaustive but represent a vast majority encountered in practice. In spite of this, lead alloys are the
preferred material for electrowinning anodes from acidic sulfate solution because they are insoluble
in aqueous sulfate electrolyte systems and have the ability to form a protective PbO2 layer. They are
corrosion resistant, economical and have an acceptable operating voltage. There are alternate anode
technologies out there but none can compete economically with lead anodes. Moreover, limited,
long-term industrial test work has been carried out with the alternate anodes to prove their
technical, practical and commercial viability.
Rolled Pb-Ca-Sn anodes are used widely in copper electrowinning. They are strong, have a
particularly high corrosion resistance, good form stability, and they have a low potential drop
between the hanger bar and the lead sheet. The electrochemical performance of lead anodes is
independent of whether they are made from primary or secondary lead. If the anodes are
maintained properly according to the broad guidelines discussed in this paper, we can expect an
optimum functional life.

REFERENCES
Beukes, N.T. and Badenhorst, J., Copper Electrowinning: Theoretical and Practical Design, Hydrometallurgy
Conference 2009, The Southern African Institute of Mining and Metallurgy (2009).
Burbank, J., Anodization of Lead in Sulfuric Acid, J. Electrochemical Society, Volume 103, Issue 2, pp. 87-91
(1956).
Cifuentes, L., Astete, E., Crisostomo, G., Simpson, J., Cifuentes, G. and Pilleux, M., Corrosion and Protection of
Lead Anodes in Acidic Copper Sulphate Solutions, Corrosion Engineering Science and Technology,
Volume 40, Number 4, pp. 321-327, December (2005).
Hughes, C., Barnard, K, and Cheng, C.Y, and Larcombe, K. The Role of Contaminants in Phase Separation During

Copper Solvent Extraction, Proceedings ALTA Copper Hydrometallurgy Forum 1998 Brisbane, Qld
(1998).
Hyvarinen, O., The Effect of Silver and Cobalt on the Oxygen Evolution at Lead Anodes, Helsinki University of
Technology, Laboratory of Process Metallurgy, Otaniemi, Finland (1972).
Kordosky, G.A., Copper Recovery using Leach/Solvent Extraction/Electrowinning Technology: Forty Years of
Innovation, 2.2 million Tonnes of Copper Annually, The Journal of the South African Institute of Mining
and Metallurgy, pp. 445-450, November-December (2002).
Miller, G., "The Problems of Manganese and its Effects on Copper SX-EW Operations", Copper 95 - Cobre 95, Vina
del Mar, Chile (1995).
Miller, G., Methods of Managing Manganese Effects on Copper SX-EW Operations, ALTA 2010, Ni-Co-Cu
Conference, Perth Australia, May (2010).
Prengaman, R. D., Ellis, T. and Mirza, A., 10 Years Experience with Rolled Lead-Calcium-Silver Anodes, LeadZinc 2010, Edited by: A. Siegmund, I. Centomo, C. Geenen, N. Piret, G. Richards, and R. Stephens, TMS
(The Minerals, Metals and Materials Society) (2010).
Prengaman, R.D. and Morgan, C.E., RSR Corporation, Dallas, TX, Electrowinning Anode and Method of
Manufacture, U.S. Patent 5, 172,850, December 22 (1992).
Prengaman, R.D. and Morgan C.E., RSR Technologies, Electrowinning Anodes which Rapidly Produce a
Protective Oxide Coating, U.S. Patent 6,224,723 B1, May 1 (2001).
Prengaman, R.D., New Insoluble Lead Anodes for Copper Electrowinning, The Electrorefining and Winning of
Copper, TMS 116th Annual Meeting, Edited by: J.E. Hoffman, R.G. Bautista, V.A. Ettel, V. Kudryk, R.J.
Wesely, TMS, Denver, Colorado, February 24-26 (1987).
Prengaman, R.D., RSR Technologies, Electrowinning Anode, U.S. Patent 6,131798, October 17 (2000).
Prengaman, R.D. and McDonald, H.B., RSR Corporation, Stable Lead Dioxide Anode and Method for
Production, U.S. Patent 4,236,978, December 2 (1980).
Pavlov, D. and Dinev, Z., Mechanism of the Electrochemical Oxidation of Lead to Lead Dioxide Electrode in
H2SO4 Solution, J. Electrochemical Society, Volume 127 No. 4, pp. 855-863, April (1980).
Stantke, P., Guar Concentration Measurement with the CollaMat System, Fourth International Conference,
Copper 99 Cobre 99, Volume III, Electrowinning Theory and Practice (1999).

Steel dust leaching with organic acids for zincs

recovery using a bioelectrolitic cell

Byron Quiroga and

Ernesto de la Torre
Escuela Politcnica
Nacional, Ecuador

Steel dusts, considered as toxic wastes, are made of metal oxides (ZnO, PbO,
Fe3O4), which can be valued by techniques like leaching with HCl o H2SO4,
that are not selective with iron. Leachates have a pH between 1-2, therefore,
it is necessary to add NaOH for the solution to reach a pH of 4-5 that let the
electrode position of Zn. Organic acids (citric, oxalic, tartaric and acetic)
permits the leaching of Zn at this pH with less dissolution of Fe. The results
of this research show recoveries with: Citric acid, Zn 68%, Fe 12%; with
oxalic acid, Zn 0,2%, Fe 75%; with tartaric acid, Zn 56%, Fe 12% and with
acetic acid, Zn 70%, Fe 5%, working with pulps with 10% of solids at a
concentration of 100 g/L of organic acids. A bioelectrolitic cell was used for
zincs recovery, which has two sections divided by a cation exchange
membrane. Inside one of such sections, the leached solution is placed with
Al cathode, for zincs recovery; in the other section, 1M sulfuric acid is
added with lead anode and a current density of 300 A/m2. The function of
the membrane is to permit the transfer of only hydronium ions from the
sulfuric acid to trigger the formation of the first carboxylic acids. For this
reason, the pH of the solution decreases. This research presents good
economic perspectives for industrial application.

INTRODUCTION
The steel dust is fine solid particles found in the gases of furnaces in the steel production. They are
composed of elements such as iron, zinc, lead, cadmium (Madas, 2001). The steel industry
produces approximately fifteen kilos of steel dust per tonne of manufactured steel, which means
that about 6 million tons of steel dust is generated annually in the world.Not all the types of dust
produced in steel mills have enough zinc and lead to make worthy a project of recycling and
enrichment of zinc (Asadi et al., 1997). The compounds of such dusts are highly toxic so inUSA and
the European Union, they are considered as toxic and dangerous waste (RTP) (Machado et al.,
2006).
The procedures currently used for the recovery of heavy metals present in the steel dust could
bepyrometallurgical (Waelz, Plasmadust, etc.), hydrometallurgical (Zincex, Ezinex, etc.), or the
combinations of both. Such processes have the following common features: high costs of
investment, metals prices suffer considerable deductions and energy costs are significant (Pecner
Bemstein 1998).
The leached solution is treated by electrolysis.Lead anodes and aluminiumcathodes are used for the
recovery of Zinc. Zinc is deposited at the cathode and oxygen is formed at the anode, where also
the regeneration of the used acid takes place. Such acid is recycled tothe leaching stage. The
recovered metal is cleared automatically or manually, then it is melted in electric furnaces and
alloys are made (Jha, Kumar Singh, 2001).
In the Zn Recovery with cloridric acid, a bi-electrolyticcell was used with the same cathode and
anode, but separated by a waterproof cation exchange membrane. In the cationic side, the
electrolyte was composed of a chlorine solution, while in the anodic side,the electrolyte was a 1 M
H2SO4 solution.The usage of this type of cell was necessary because organic acids are weak
electrolytes that form oxygenatedfoam around aluminum cathode in normal cells, which causes the
no accession of Zn in the aluminum plaque.Also, the lead anode begins to disintegrate and reacts
with the solution.Thecationexchange membrane avoids the contact of the leachate solution with the
anode and, therefore,the formation of foam around the cathode, which in turn allows the
electrodeposition of Zn (Ricaurte et al., 2013).
In general, leaches are made of inorganic acids that are hazardous for the environment, so it is of
great interest to investigate new leaching solutions. This research presents an alternative with
organic acids: citric, oxalic, acetic and tartaric acid, which can give technical, economic and
ecological advantages for the industry.

METHODOLOGY
Steel dust was obtained from the cleaning systems of gaseous streams ofsteel
recyclingfurnaces.Spectrophotometry of atomic absorption (AA), fluorescence x-ray (FRX) and
scanning electronic microscopy (SEM-EDS)were used for chemical characterization.
Aqueous solutions of acetic, citric, oxalic, tartaric acids at different concentrations (10, 100 and 150
g/L) and vinegar (5% w/w acetic acid) were used as leaching means. In addition, with the best
leaching agent, tests varingthe percentage of solids (1, 10 and 30%)were carried out for 24 hoursand
leaching kinetics were established (Espinoza, de la Torre& Guevara, 2011).Firstly, the reactor was
loaded with 1 g steel dust in 100 mL solution. Then, it will be filtered to obtain a concentrate. The
sludge will be washed with 50 ml of distilled water to get the washing solution.The content of Fe,

Zn, Cd, Ni, Cr and Pb of the solutions and tailings will be analyzed by spectrophotometry of atomic
absorption (AA).
To evaluate the recovery of Zn by the process of electrolysisbi-electrolytic is going to be used. It will
be composed of a lead anode, aluminum cathode and cation ion exchange membrane (NEOSEPTA
type CMX).This membrane is in the middle of the cell.
AH2SO4 solution 1 M is going to be placedat the one end, where the lead plaque is located; it is the
anode side. At the other end of the cell, theleached solution is going to be placed, together with an
aluminum plaque that acts as a cathode. The two plates should be as close as possible to the
membrane to facilitate the electrodeposition.Thecurrent intensity is 400 A/m2.
The deposited Zn is then pulverized and homogenized with NH4Cl (ratio of 75 in weight of metal
deposits) as flux load, to be melted for 1 h in a furnace at 450 C.The Zn obtained from electrolysis
and fusion will be analyzed by scanning electron microscopy (SEM-EDS) to determine its purity.

RESULTS AND DISCUSSION

The steel dust taken from thegas line of the recycled steel melting furnace presents the following
characteristics:
Table 1 Chemical characterization of steel dust by fluorescenceX-ray (XRF) and atomic absorption
(AA)
Element

Content (%)

Element

Content (%)

Element

Content (%)

Zn

29.95

Al

1.22

2.03

Fe

18.70

0.52

Ca

1.94

Si

2.48

Cu

0.25

Mn

1.69

Mg

9.35

Cr

0.14

Bi

0.06

6.18

0.11

Cd

0.03

Cl

5.16

Ti

0.08

Sb

0.03

Pb

2.65

Sn

0.07

Ni

0.03

(Espinoza, de la Torre & Guevara, 2011)

Tests were carried out with a 100 mL leaching solution,1% of solids,100 g/L acid concentration.Two
brands ofcommercial vinegar were used as leaching agent (Sabor and Helmozt).The resultsare
shown below in figure 1 and 2.
Results show that the recovery of Zn is high for 70% and 68% for acetic and citric acid respectively,
but the acetic presents a lower recovery of Fe and Pb. Also, it could be appreciated that the metal
with the highest percentage of recovery is the Cd, but its presence in the sample is minimal. Its
percentage is 0.03%.

Metals recovery [%]

100
80

Zn
Pb
Cr
Ni
Cd
Fe

60
40
20
0

Acetic

Citric

Oxalic

Tartaric

Figure 1 Results from leaching of steel dust 1% solids concentration of 100 g/L of acid

Also it is appreciated that oxalic acid has not affinity with Zn because its recovery is low, but
itproduces a great recovery of Cr: 90%. Tartaric acid has a recovery of Zn of 56%, but it is not so
good in comparison to other leaching agents.The recovery of Pb is high and it represents a problem
for electrolysis.
Recovery of Zn is related to the dissociation constant of different acids, pKa, acetic: 4.75; citric: 3,08;
tartaric: 3,03; oxalic: 1.27, (Maron & Prutton, 2002). A higher pKa Zn recovery is greater because it
has a greater ability to form complexes, in addition electronegativities of Zn is high in relation to
the other metals. However, the Cd presents high recovery percentages in the different solutions due
to the electronegativity of this is higher than the Zn, but in the study is negligible because it is
found in low concentrations in the sample. By other hand, it is observed that the acid citric
recovered a large amount of the metals analyzed in figure 1, this occurs because of dissociates three
times, gradually increasing its pKa and recovers more metal in each dissociation.

Metals recovery [%]

100
80

Zn
Pb
Cr
Ni
Cd
Fe

60
40
20
0

Acetic

Vinegar H

Vinegar S

Figure 2 Results from leaching of steel dust 1% solids concentration of 100 g/L of acetic acid and vinegar
(concentration 50 g/L of acid)

As shown in Figure 2, the two brands of vinegar (acetic acid 50 g/L) presented a recovery of Zn of
70%, similar to the recovery of the 100 g/L acetic acid solution; but if we take into account the
recovery of Fe and Pb, which are impurities for obtaining the Zn by electrodeposition, we conclude
that the best leaching agent is vinegar H with recoveries of Zn 70%, Fe 5% and Pb19%.
Figure 3 shows the recovery of metals at different concentrations of acetic acid, as it has high
recoveries of Zn and low recoveries of Fe and Pb.
According to the results, we can say that fordiluted pulps, the concentration of acid does not affect
significantly the Zn recovery because its recovery is between the 70 to 80%, but for the Fe and Pb it
is an important factor.

Metals recovery [%]

100
80
60

Zn

40

Fe
Pb

20
0

50

100

150

Acetic acid concentration [g/L]

Figure 3 Comparison of the results of Zn, Fe and Pb recovery at different concentrations of acetic acid

In the case of Fe, the percentage of recovery is proportional to the concentration. Talking aboutPb, it
is appreciated that a lower recovery of this metal is obtained at a concentration of 50 g/L of acid.
According to this, it can be seen that the leaching agent which has the greatest affinity for Zn is the
acetic acid.The concentrations used in the following leaching tests are: 50 g/L because, at this
concentration, the lowestpercentage of impurities is recovered and 150 g/L,which obtained the
largest amount of Zn in solution.
The figure 4 shows the results of Zn and Fe recovery at different concentrations of acetic acid (50
and 150 g/L) in different percentage of solids (1, 10, 30%).
Theobtained results shows that the concentration of 50 g/L of acetic acid compared to the 150 g/L
does not show significant variation in the Zn recovery, working at different percentages of solids.
If we appreciate the recovery of Fe at different concentrations, we discovered that the behavior is
similar to the one described above. However, we can say that at 30% solids, the recovery of Fe is
higher and Zn recovery is lower, which is not favourable forelectrodeposition.

Metals recovery [%]

100
80
Zn a 50 g/L
acetic acid
Zn a 150 g/L
acetic acid
Fe a 50 g/L
acetic acid
Fe a 150 g/L
acetic acid

60
40
20
0

10

20

30

Amount of solids [%]

Figure 4 Comparison of the results of Zn and Fe recovery at different concentrations of acetic acid and
percentage of solids

The kinetics of leaching of Zn and Fe, which were obtained working with 10 and 30% of solids at
different concentrations of acetic acid (50 and 150 g/L), during 10 hours of treatment, are presented
in figure 5 and 6.

Metals recovery [%]

100
80

Zn 50 g/L
acetic acid
Zn 150 g/L
acetic acid
Fe 50 g/L
acetic acid
Fe 150 g/L
acetic acid

60
40
20
0

10

Time [h]

Figure 5 Zn and Fe recovery with 10% solids to 50 and 150 g/L of acetic acid at constant concentration

In the testof 10% solids, you can see that the curve of Zn recovery for both concentrations is similar.
For this, it can be concluded that 50 g/l acetic acid is sufficient to form complexes with Zn and it is
not so selective with other metals. In contrast, even though Zn is recovered at the concentration of
150 g/l, it saturates the sample, forcing other metals to react and form other compounds.
The higher recoveries of Zn were reached during the first 4 hours of leaching, in each of the
solutions of different concentration. However, the higher recovery was obtained at the
concentration of 50 g/L of acetic acid, corresponding to 70% of Zn, with a low recovery of Fe: 5.5%,
which value significantly helps the process of electrolysis.
Figure 6 shows that for the concentrations of 50 and 150 g/L, Zn recovery is 35 and 45%,
respectively, working with a time of leaching of 10 hours. This occurs because there isa large

amount of sample in solution and acid starts to react with various metals, that it is not favorable for
the recovery of Zn.
Zn concentration when working with 30% solidsand4 hours of leaching is higher than the one
obtained with 10% solids (25 and 50 g/L Zn in solution) respectively. However, a very important
aspect is the recovery of Fe, which at concentrations of 50 and 150 g/L of acid was 32 and 40%
respectively, which represents a lot of impurities in the electrolysis process, and would require an
intermediate stage of purification.

Metals recovery [%]

100
80

Zn 50 g/L
acetic acid
Zn 150 g/L
acetic acid
Fe 50 g/L
acetic acid
Fe 150 g/L
acetic acid

60
40
20
0

10

Time [h]

Figure 6 Zn and Fe recovery with 30% solids to 50 and 150 g/L of acetic acid at constant concentration

It is concluded that the best conditions for leaching with organic acids to recover Zn in steel dust
are: 50 g/Lacetic acid (vinegar), 10% solids, and constant agitation at 500 rpm in 4 hours
operation.Also, if we work in theseconditions, itis not necessary purification steps because the
concentrations of metals like Fe, Cd, Pbare low.Moreover, the pH of the solution has a value of 4.7,
so it is can betrasfered directly to the electrolysis process.

Zn recovery [%]

100
80
60
40

Zn

20
0

20

40

60

80

Time [min]

Figure 7 Results of Zn recovery at different times in the process of electrolysis

The bi-electrolytic cell was installed, as explained in the methodology, and Zn electrodeposition
was performed, the results can be seen in figure 7.
There is shown that the percentage of Zn in the plaques increases approximately 17% in the first 40
minutes, then the percentage of recovery does not change, because the amount of Zn in the solution
decreases over time until reaching the saturation point.
For the performedtests, it is known that the cell operation time is 40 min. For longer time, the
recovery is minimal and operating parameters are changed (pH decreases and increases the
intensity of current). The amount of Zn recovered in the process of electrodeposition was 80% in
relation to the amount present in the solution.
The material removed from the cathode was analyzed using the technique of scanning
electronmicroscopy SEM - EDX (Tescom - Bruker), and showed that Zn is approximately a 75% in
the analyzedsample.
To melt thecathodicZn, the blades obtained from electrolysis must be powdered manually.
Ammonium chloride was added as flux load. It was added in a proportion of75%of the Zn thatwas
going to be melted. (Espinoza, de la Torre & Guevara, 2012).
The pulverized bladesand flux load, previously homogenized, were placed in a ceramic crucible.
The crucible was placed in an electric muffle at 500 C for 1 hour. Pure metallic Znobtained in the
fusion was 95%, compared to the amount obtained in the electrodeposition. The rest of the material
was part of the slag.
The Zn obtained from fusion was analyzed by scanning electron microscopy SEM - EDX (Tescom Bruker). Results are shown below in TABLE 2.
Table 2 Composition of Zn metallic obtained from fusion
Elements

Composition [%]

Zn

89,06

8,86

Cl

2,08

According to the results of the fusion, impurities were removed with the slag and zinc with a purity
about the 90% was obtained.

CONCLUSIONS
Acetic acid was the best leaching agent of organic acids because it achieves a recovery of 70% Zn,
5% Fe and 13% Pb in solution at a concentration of 50 g/L (vinegar), with a percentage of solids of
10%, in 4 hours of dynamic leaching.
A high concentration of zinc is accomplished in leaches with150 g/L acetic acid and 30% of solids,
but is not favorable to the process because 40% Fe and 27%Pb were also recovered, which
represents large amount of impurities in the electrolysis process.
Electrodeposition of Zn was possible in a bi-electrolyticcell, at an intensity of current of 400 Am2,
with a pH of 4.7. These conditions recovered 80% ofcathodic Zn in aluminum plaque in 40 minutes;
Zn was present in a 75% in the deposits.

On the fusion of the cathodic Zn, using ammonium chloride as load flux, in a 75%proportion to the
amount of sample, at 450 C for 1 h, Zn metallic with a purity of 90%was obtained.

REFERENCES
Asadi, B., Mowla, D., Shariat, M. &Fathi, J. (1997) Zinc recovery from blast furnace flue dust.Hydrometallurgy.
Vol. 1 (47), p. 113. Shiraz, Irn, viewed February, 2014, http://www.sciencedirect.
com/science/article/pii/S0304386X9700039X.
Espinoza, C., de la Torre, E. & Guevara, A. (2012) Recuperacin de zinc a partir de polvos de acera.
Tesispregrado, p. 69. Quito, Ecuador.
Ismael, M. &Carvalho, J. (2002) Iron recovery from sulphate leach liquors in zinc hydrometallurgy. Minerals
Engineering 16, pp. 31-39, Lisboa, Portugal, viewed February, 2014, http://www.sciencedirect.
com/science/article/pii/S0892687502003102.
Jha, M., Kumar, V. & Singh, R. (2001) Review of hydrometallurgical recovery of zinc from industrial wastes.
Resources conservation & recycling 33, pp.2-10, Jamshedpur, India, viewed February, 2014,
http://www.scienc edirect.com/science/article/pii/S0921344900000951.
Machado, J., Brehm, F., Mendes, C., dos Santos, A. &Faria, A. (2006) Characterization Study of Electric Arc
Furnace Dust Phases. Mat. Res, pp.41-45, Santiago, Chile, viewed February, 2014, http://www.
scielo.org.co/scielo.php?script=sci_nlinks&ref=000084&pid=S0012-7353201100050002 600006&lng=en.
Madias, J. (2001) Tendencias tecnolgicas: reciclado de polvos de horno elctrico, p.38, Buenos Aires,
Argentina, viewed January, 2014, http://www.construccionenacero.com/Boletin06 09/T endencias
%20Tecnol %F3gicas-Reciclado%20de%20Polvos%20de%20Horno%20El%E9ctrico.pdf.
Maron, S. & Prutton, C. (2002) Fundamentos de Fisicoqumica, pp. 337-248, Mexico D.F, Mexico.
Nalini, J. & Sharma, D. (2004) Biohydrometallurgy for Nonsulfidic Minerals-a reiew.Geomicrobiology Journal,
pp. 135-138, doi: 10.1080/01490450490275271.
Peckner, D. & Bernstein, I. (1998) Handbook of Stainless Steels. , viewed February, 2014, http://library.dmr.go.
th/library/TextBooks/3187.pdf.
Ricaurte, J., de la Torre, E., Espinoza, C. & Guevara, A. (2013) Estudio comparativo de la recuperacin de Zn de
polvos de acera por lixiviacin con H2SO4 y HCl, electrodeposicin electroltica y bi-electroltica.
Revista Politcnica 33, pp. 73-74, Quito, Ecuador.

Flowmeter validation for

metallurgical balance in Minera
Michilla, using radiotracers

Francisco Daz and

scar Jimnez
Trazado Nuclear e
Ingeniera, Chile
Genny Leinenweber
csiro Chile International
Centre of Excellence in
Mining and Mineral
Processing and Universidad
de Antofagasta, Chile
Abraham Backit,
Armando Cceres and
Cristian Mundaca
Minera Michilla, Chile

Minera Michilla is located in the II Region of Chile, where copper oxide ore
is processed, from the Lince deposit, third deposits and sulfide minerals
from the underground mine Estefana and mixed sulfide minerals from
deposits exploited by others. The metallurgical processes involved are
crushing, agglomeration, heap leach, solvent extraction and electrowinning.
For metallurgical balances it is essential to know the copper
concentration and the solutions flow, therefore, the information provided
by online flowmeters is necessary for reliable results.
Conventional flowmeters installed in pipes may suffer disruptions or
failures by corrosion, fouling, presence of elements in the water such as
acids, salts and suspended solids. Radiotracers use is advantageous to
determine fluid velocities which are transported by pipelines, and which are
detectable by means of the radiation emitted by applying the technique of
the transit time between sensors at known distances, and determining the
passage time of the radiotracer. Pumped flow velocities have been measured
from 11 pipe solutions transport, obtaining velocities between 0.58 m/s and
2.5 m/s., corresponding to flows between 7,710 and 15,107 L/min.
Contrasting the measured flow from the radiotracer technique with the
information provided by the flowmeters, the following differences were
obtained: smaller than 5% in 4 flows, between 5 and 10% in 5 flows and 11
and 18% in 2 flows. With these new measured flows from radiotracer
technique, the metallurgical balances have a better fitting than the
flowmeters information provided.

INTRODUCTION
Michilla S.A. minning company is located nearby the locality of the same name in the II region, 200
km. north from Antofagasta, Chile. A dirt road takes the mining labor from the coastal installations
and is located about 1000 mt. over the sea level. The climate in Michilla is characterized by the
warm temperatures regulated by the offshore winds. Due to this favorable situation it is possible to
operate the mine and the heap leaching without interruption during the year.
Michilla mine processes the copper oxidized ores from the Lince deposit and third-party sites, as
well as the sulphide ores coming from the underground mine Estefania and the mixed sulphide
minerals coming from exploited sites by third-parties trough the processes of crushing
agglomeration - heap leach-solvent extraction and electrowinning (www.michilla.cl)
The mass balances are of a great importance in a mining labor, to know the flow of the solutions in
the processes of obtaining copper allows improving and optimizing the production processes.
In this work a flow rate speed hydraulic evaluation was done in different pipes of solution
transporting in the facilities of the Michilla mine with the purpose of having hard data that support
and verify the information online of specific instrumentation dedicated to those purposes, in a way
of having a cross-checked information of the flowmeters and this way validate, contrast and
calibrate them with the purpose of having more reliable solution flows and metallurgical balances.
The studied solutions were: rich, ILS, refine and organic, the ones are marked with the tracer I-131
in solution, injecting it in a nearby to impulsion pump suction. The tracer pass measurements are
done between two points of the flow transporting pipe to determine trough nuclear detectors in
real time.
The method used in this work is denominated radioactive tracer (Charlton, J.S., 1986) and consists
in that the radiation is physically inserted in the flow to determine by the addition of a measured
amount of the tracer. The I-131 emits radiation which is detectable trough the walls allowing its
measured online, that radiation decreases in time, reason why radioisotopes are used with
intensities useful in relation to the test period.

METHODOLOGY
Flow rate speed on pipes
Flow rate measurement by the transit time method is based on measuring the transit time of labeled
fluid particles between two cross-sections of the pipe at a known distance. Labeling of the fluid
particles is achieved by injecting a tracer into the flow upstream, then, the transit time is
determined from the difference of the mean arrival times of the tracer at each of the detector
positions located at two different cross-sections of the pipe, were the detectors are installed
(International Standar ISO 2975/VII, 1977; Dincer, T., 1967) .

Pipes flow rate speed analyzed

The following facilities where the solutions are transported by pipes by positive pressure were
analyzed:

Rich solutions 1, 2 and 3

Refine solutions 1, 2 and 3
ILS solutions 1, 2 and 3

Organic solutions 1 and 2

Tracer selection
The selection of the Radioactive Tracer was made based on:

Physicochemical behavior, according to the material that has to be traced.

Halftime life of the radioisotope, compatible with the duration of the transport and the
experiences.
Type and energy of the radioisotope radiation.

Due to the excellent dissolution in water I-131 is used as radiotracer, with activities in the order of 5
mCi. The halftime life is appropriate, considering the transport time to the mine, the experiences
duration (8 days) and the gamma radiation energy of the I-131 (Atomic Energy International
Agency, 1990). All this consideration allows the detection through thick HDPE pipes.

Tracer quantity
To calculate the amount of tracer required it is necessary to consider (Thyn, J., 2000):
Longer experience. It is considered around 3 times the theoretical half time, which results of
the quotient between the volume used by the solution and the flow. In this case from some
seconds to minutes.
The average time that is required to detect in the checkpoints.
Flow in which the tracer is injected.
Geometry of measure. In all of the cases it is measured directly over the transport pipes of
solutions, reason why it is not necessary relevant shield or considerable distance at the tracer
pass.
With the previous considerations it is determined the amounts of 5 mCi of I-131 are enough
to obtain count rates appropriate in each of the experiences with I-131 as tracer and injection
in pump suction.

Tracer injection
So that the function of the output represents directly the flow rate speed through the determination
of the transit time among 2 or more detectors, it is convenient that the income of the tracer to the
system to be the closest to a pulse. Is like this that the tracer gets into like a pulse in the time and
space, close to pump suction that impulses solutions.
Tracer injection is essential in the work, to ensure representative and reliable results. When there
are no facilities to inject the sample, a hole of a 1 mm diameter is done in the pipe, it is put a septum
and with the help of a syringe the tracer is injected, the septum avoids the sample to come back. To
seal the hole a metal band is put embracing the pipe in a way to reassure that there are no spillover
like in figure 1.

Tracer detection
The detection of the radioactive tracer (Knoll, G.F., 1989), Figure 2, is done through nuclear
detectors (scintillation NaI (Tl), Altaix) with its associate electronic for data transmission to a
monitoring station. These systems allow the acquisition of data simultaneously up to 12 control
points with 10 milliseconds minimum range. The probes are located over the pipe under study, in
order to observe only the tracer radiation passage, the probes have a lead collimator.

Figure 1 Injection of the radioactive tracer

Figure 2 Detectors of the radioactive tracer passage

RESULTS
The data acquired through radiation intensity are corrected by radioactive decay and natural
radiation background. The mathematical expression (Atomic Energy International Agency, 1990)
that summarizes the calculation of radiation intensity is:

=(

Where

(1)

Acor = Radiation activity, in counts per seconds (cps), corrected by background and radioactive
decay.
Amed = Experimentally measured activity
BG

= Background radiation.

= Measurement time.

t1/2

= Tracer halftime life.

To get the flow rate speed it is necessary to obtain the experimental half residence time of each
Distribution Residence Time Distribution Curve, applying the moments method, as follows:

Where

( )

( )

texp

= Experimental half time.

= Time in which the sample is collected.

Acor(t) = Activity of the tracer in time t.

(2)

The method of determining flow rate speed with radioactive tracer is about measuring the pass of a
tracer by two detection points, at least at a known distance and thus determines the flow rate speed.
The cross section to the flow is a constant value, that way the flow can be determined. For this case
the distance between detectors is measured with a certified tape and contrasted with a laser
instrument with a 2 mm. error, but for more security and error of 1% is assigned. The speed is
obtained through the quotient of the distance between detectors and the difference of the half time
pass between these.
As an example, in Graphic 1 we see the measurements of transit time in the pumping station rich
solution 1.

Graphic 1 Transit time in rich solution 1

Average flow rate speeds in pumping solution refine 1

In the TABLE 1 it is shown as an example the detailed information of the difference of the transit
times, distance between detectors and the speeds in each one of the experiences done in the solution
Refino 1 sulphides. It is used the error propagation.

Table 1 Speeds and transit times in solution refine 1 sulphides

Distance between

Speed flow rate

Exp.

Time [s]

detectors [m]

[m/s]

13,43 0,07

17,37

1,29 0,01

13,53 0,06

17,37

1,28 0,01

13,54 0,06

17,37

1,28 0,01

13,60 0,05

17,37

1,28 0,01

13,57 0,05

17,37

1,28 0,01

Average flow rate in solution transporting lines

With the internal diameters of each of the pipes informed by Michilla and the flow rate speeds
measured, it is obtained the flow of solutions transported by each of the lines. Simultaneously,
while the flow rate speed measurements are done, a lecture of the flowmeters installed on line for
their contrasting is taken. In TABLE 2 it is shown the final summary of the obtained results.
Comparing the data obtained through the flowmeters online and those obtained by the radioactive
tracer technique, necessary corrections and calibrations of the instrumentation can be done. Also,
TABLE 2 shows the comparative results of the flow rate measurements with flowmeters on line.
Table 2 Summary results
Internal
Diameter
[m]

Average
Speed
[m/s]

Rich 1 Sulphides

0,327

1,53

7.710

8.145

-5,65

Rich 2 Oxides

0,397

1,59

11.785

10.52

10,74

Rich 3 Ripios

0,353

1,42

8.319

8.581

-3,14

Refino 1 Sulphides

0,384

1,282

8.899

8.123

8,72

Refino 2 Oxides

0,441

1,36

12.441

11.708

5,89

Refino 3 Ripios

0,397

1,24

9.191

9.392

-2,19

ILS 1 Sulphides

0,409

1,52

11.994

12.394

-3,34

ILS 2 Oxides

0,555

1,04

15.107

14.865

1,60

ILS 3 Ripios

0,555

0,582

8.454

7.917

6,35

Organic SX-1

0,290

2,30

9.115

9.802

-7,53

Organic SX-2

0,313

2,52

11.619

13.68

-17,74

Solutions

Flow rate
Flow rate in
Difference
measured with flowmeters between flows
tracers [lts/min]
[lts/min]
[%]

CONCLUSIONS
The injection of the tracer is clean and perfectly representative of a pulse in all the experiences.
The intervals of data acquisition are appropriate. With higher times of acquisition, there is a risk of
losing information due to the fast pass of the tracer, especially with high speeds and short distance
between the detection points.
Differences lower than 5% regarding the flow rate measured through radiotracers and the
information given by the flowmeters online are found in the transporting lines of solutions rich 3,
refine 3, ILS 1 and ILS 2.
Differences between the 5 and 10% are obtained in the solutions rich 1, rich 2, refine 1, refine 2 and
ILS 3. Difference in the order of the 8%, being higher the flow indicated in the flowmeter, is
obtained in the solution Organic SX-1.
The biggest difference is found in the solution Organic SX-2, with a value of the order of 18% being
higher the value indicated in the flowmeter.
The values of the measurements of flow speed trough radioactive tracers are hard data that, ensure
an internal diameter of the pipes also hard, give the effective flow of solution transporting that
serves to validate, contrast and calibrate the instrumentation of the flowmeters installed online.
It is recommended to realize flow rate speed measurements at least every 3 months, to ensure a
proper operation of the flowmeters and that way realize metallurgical and flow balances reliable
and validated trough the cross flow measurements.

ACKNOWLEDGMENT
To Michilla S.A. Mining Company for the information given to carry this work done.

REFERENCES
Atomic Energy International Agency (1990) Guidebook on Radioisotopes Tracers in Industry Vienna.
Charlton, J.S. (1986) Radioisotope Techniques for Problem Solving Industrial Process Plants Leonard Hill.
Dincer, T. (1967) Application of radiotracer method in streamflow measurement, Editorial: Agency
International Atomic Energy. Vienna, pg. 93. Isotopes in Hydrology.
International Standar ISO 2975/VII (1977) Measurement of water flow in closed conduits- Tracer methods-Part
VII: Transit time method using radioactive tracers. ISO 2975/VII-1977 .
Knoll, G.F. (1989) Radiation Detection and Measurement. New York : Willey 2 Edition.
Thyn, J. (2000) Analysis and Diagnostics of Industrial Processes by Radiotracers and Radioisotope Sealed
Sources.
www.michilla.cl

Extraction of sulfides from aqueous effluents

kof the oil industry

Joo Bosco De Araujo

Paulo, Ana Karla Costa De
Oliveira, Elisa Maria
Bittencourt
Dutra De Sousa and Kssia
Larissa Pinheiro De Lima
ufrn, Federal University of
Rio Grande do Norte, Brazil

Produced waters are complex mixtures and have many contaminants as

finely dispersed oil, heavy metals and gases such as H2S and CO2,
originated from the composition of oil and natural gas. The H2S must be
removed from the effluent of the oil industry, since it is a highly toxic gas,
potentially lethal and when mixed with water it forms a corrosive acid to
pipelines, flowlines and others equipments. In the current Brazilian
environmental legislation established by The National Environmental
Council (CONAMA), the concentration of sulfides in these waters for
disposal in the environment must be less than 0.3 mg/L. In this paper, we
use solvent extraction for the recovery of sulfide species. We use three
commercial amines of the group of alkyl amines as extractants which are
dissolved into aviation kerosene (jet fuel). Produced water samples are
actual samples of the oil industry and have an initial concentration of 0.660
mg/L H2S. The extraction tests are done manually and carried out in batch
by means of separation funnels and agitation between the organic and
aqueous phases. The parameters studied were: The volumetric ratio Amine/
jet fuel (held in 0.25 vv.) and the volumetric ratio organic /aqueous phases
(held in 13 vv). The three tested amines, DUOMEEN T (N-tallow-1,3diaminopropane), ARCUARD 2 c-75 (Dicocoalkyldimethylcloride),
DUOMEEN O (N-oleyl-1,3-diaminopropane), present a character
hydrophobic and tensioactive character. The efficiency of the extraction
process was evaluated in terms of concentrations (mg/L) of H2S and oil
before and after the contact between phases. After tests in triplicate for
each of the three amines studied, it was concluded that the improved
efficiency of extraction occurred with the amine DUOMEEN O which
showed 76% of sulfide removal, secondly appeared to amine ARCUAD 2
c-75, showing 59% removal efficiency and finally the amine DUOMMEN T
which presented 40% removal.

INTRODUCTION
The processes of oil exploration are associated with the generation of produced waters. The volume
of produced waters from oil fields depends on factors such as: geological structure of the
generating rock, practices and time of production of reservoir fluids (Bayati compensation,
Noorjahan Shayegan, 2011). These effluents are complex mixtures of organic and inorganic
materials, suspended and dissolved, can contain heavy metals (Moraes, Paulo Medeiros, 2011), and
bring with them characteristic elements of the reservoir rock where fluids were produced, in
addition to chemicals added to prevent corrosion of the tubes (Burns et al., 1999) and alkalinization
due to injection wells recovery (Zang et al., 2010).
Acid gases like CO2 and H2S are also present as components in produced water, came from the oil
and natural gas composition. In the conventional processes of acid gases removal in natural gas,
these are removed by means of an Acid Gas removal Unit (AGRU), using the process of absorption
with an aqueous solution of amina (Rufford at al .2012). The H2S must be removed from the fluids
of the oil industry, since it is a highly toxic gas, potentially lethal and, when mixed with water,
forms a corrosive acid to pipelines and flowlines and other equipments (Rufford at al, 2012). In this
context, it is possible that acid gas concentrations are higher than the allowable limits established by
environmental law; it means concentration of sulfides lower than 0.3 mg/L into disposition waters.
The environmental agencies establish a concentration of oil of 20 mg/L as upper limit for the
disposal of effluents from the productivity process of hydrocarbons, (Burns et al., 1999).
One of the most used processes for removal of acidic components is the absorption with
alkaneamines; in this process, the acids components react with a liquid alkaneamine which works
as solvent, when contacting the liquid and gas phaes; in a later step, the acidic components are
removed in a regenerator, usually at low pressures and high temperatures. Huttenhuis et al., 2006
developed studies of gas solubility in aqueous solutions of alkaneamines. The solubility data have
been assessed and correlated with an Equation of State (EOS), as originally proposed by Frst and
Renon (Frst Renon, 1993). Park, 2008 used as absorbent, natural secondary metabolites, obtaining a
removal efficiency of 62%. Efficiency has increased to 96%, using a complex absorbent mixed with
1.0% of a chemical amine product.
According to Kohl and Nielsen (1997), primary amines react directly with H2S, CO2 and COS.
Secondary amines react directly with H2S, CO2 and COS, the latter to a lesser degree. Tertiary
amines react directly with H2S, indirectly with CO2 and in minor extension with COS. The most
common example is the tertiary amine methyl dyethaneamine (Kohl and Nielsen, 1997). There are
also the sterically hindered amines, characterized by the presence of a bulky substituent group next
to the nitrogen atom. According to Kohl and Nielsen, 1997, primary and secondary amines,
monoethaneamine (MEA), dyethanolamine (DEA), respectively, are very reactive and therefore
exhibit high rates of removal of acid gases.
Sulfur components present in hydrocarbons are removed, in the presence of appropriate catalysts,
by forming hydrogen sulfide (Eq. 1). In the second step, the H 2S is removed from the gas stream by
means of a washing solvent (amines) according to Eq. 2, followed by the same conversion to
elemental sulphur.
R-SH + H2 R-H + H2S
2RNH2+H2 S (RNH3)2 S

(1)
(2)

Bara, 2012 relates that aqueous amine solutions are the most common chemical solvents for removal
of H2S, where the acid is neutralized with solutions of alkaline amines. The H 2S reacts quickly and
directly with the 1st, 2nd or 3rd amines to form HS-disulfide and a cation of protonated amine.
Then, reacts with other disulfide amine to form sulphide and other equivalent protonated amine.
H2S + NR3 HS- + HNR3+

HS- + NR3 S-- + HNR3+

(3)
(4)

In this work is proposed a process where the contaminants are separated from the effluent through
solvent extraction process in batch scheme, where the constituents dispersed into aqueous phase
(oil and H2S) migrate preferentially to the organic phase, a solution of amines dissolved into
kerosene of aviation (jet fuel), after solvent action. The unit operation of solvent extraction involves
initially the transfer of contaminants from aqueous to the solvent phase and subsequently the
separation between both organic and aqueous phases, (Hadjiev et al 1995). In addition to
minimising the energy expenditure, this study presents an innovative aspect, since the amines in
general are fairly soluble in water due its polar character. Since the H2S removal processes, in
general, in the literature, do not report amines to removal acid gases which could be soluble into
organic medium. In this context, this paper presents the results of a preliminary investigation of the
removal of H2S and oil, simultaneously, in a real sample of produced water, using three commercial
amines, varying the amount of this extractant, in relation to the jet fuel solvent. The amines used in
this work possess simultaneously hydrophobic character and surfactant properties (alkylamines):
DUOMEEN O(OLEYL-RNHCH2CH2CH2NH2), ARCUAD 2C-75 (R2N+(CH2)2Cl) and DUOMEEN T
(TALLOW-RNHCH2CH2CH2NH2). The contents of H2S and petroleum (mgL) were evaluated
before and after contact between the phases.
The work presents results of a fundamental study conducted in order to investigate the efficiency of
removal of sulfides, as well as fined dispersed oil, contained in real samples of produced water. For
the solvent extraction process it was used an organic solution of amines into QAV (Jet fuel). The
variable studied was the concentration of amine (mg/L) into QAV on a volumetric basis (v:v). The
results show the influence of this variable on the efficiency of extraction of oil and sulfides,
simultaneously.

METHODOLOGY
Investigated system
The produced water used was obtained from entrance of an Effluent Treatment Station (ETS) of
PETROBRAS-BRAZIL. Thus, it consists of an actual sample for treatment in the batch system for
H2S and fine oil removal. The organic phase was composed by aviation kerosene (JET FUEL),
provided by the precedent company. The three different amines (DUOMEEN T, ARQUAD 2C-75
and DUOMEEN O), were obtained from AKZO NOBEL Co., (USA) and were added to the organic
phase, in different concentrations. The physico-chemical properties of the investigated system are
presented in TABLE 1.

Table 1 Physical-chemical properties of the investigated system

Products

Density
(x 103 kg/m3)

Viscosity
(x 10-3 Pa.s)

pH

Surface tension
(x 10-3 N/m)

Produced water
Jet fuel (QAV)
DUOMEEN T
ARQUAD 2C-75
DUOMEEN O

1.150
0.775
0.840
0.880
0.841

1.079
1.220
21
120
19

8.01
6.60
7.00
7.00
8.00

72.45
24.95
30.81
14.87
27.41

Experimental
The batch system is shown in Figure 1. It consists of a magnetic shaker (Tecnal-Brazil) (1) for the
homogenization of the extractant with the solvent (jet fuel) kept inside a beaker (2); then the organic
is sent through a micropump pump type (Cole-Parmer, Brazil) (3), to the a separation funnel (4),
where organic and aqueous phases (contaminated with sulfides and fine oil) are contacted. At the
base of the separation funnel, the treated water was analyzed in a DR 2000 spectrophotometer (5)
set in 665 nm to measure the concentration of sulphides. The content of oil and greases (TOG) is
obtained using the technique of absorbance on the infra-red region with equipment provided by
Wilks Co., (not shown in the schema).

Figure 1 General schema for sulphides and oil removal on batch scale

Sample preparation and experimental procedure

A volume of 2 mL of 0.1 M of solution of sodium hydroxide (NaOH) was added to 5 L of sample of
produced water for stabilization of sulfides by 48 h. An aliquot of 2 mL of this prepared water was
submitted to Spectrophotometric reading to evaluate content of sulfides. Another aliquot is
submitted to the infra-red equipment. The amines were mixed with jet fuel, in predefined
proportions, while stirring, and subsequently sent to the separation funnel. In this step, the

contaminated water was also admitted to the funnel for vigorous contact between phases made
manually.
The contaminants are transferred from aqueous to organic phase by agitation in the separation
funnel. The removal of sulfides is associated to a reaction of complexation by amines witch form a
neutral complex, while the removal of fine drops of oil is attributed to their solubilization into jet
fuel. The contaminant oil is soluble into jet fuel due chemical affinity, since jet fuel is a byproduct of
the original hydrocarbon.
At the exit of funnel, the treated water should not contain sulfides, and even oil, which had been
complexed and soluble in the organic phase, respectively. Thus, the output samples were
sequentially treated with Spectrophotometric readings in DR 2000 equipment. Other analyses as
TOG (content of oils and greases), pH, density, conductivity, turbidity, concentration of nitrate
were performed on aliquots of input and output samples. Table 2 shows the operational conditions
for bench essays.
Table 2 Operational conditions for bench essays
DUOMEEN O
Amines
(commercial names)

ARCUARD 2C-75

Concentration into
jet fuel (v:v, %)

0.01; 0.05; 0.10; 0.25;

0.50

O/A
(Organic/Aqueous ratio)
(v:v)

1/3

DUOMEEN T

The removal efficiency for sulfides and oil, from the bench tests, was calculated from the readings
in DR 2000 spectrophotometer and TOG (content of oils and greases) obtained by infra-red
equipment of samples comparing the concentrations at the entrance and exit of the separation
funnel according to Equation 5.

Ce Cs
Efe
* 100
Ce

(5)

Where, Efe is the efficiency of separation (%); Ce is the concentration of contaminant at the entrance
and Cs the concentration of contaminant at the exit of the funnel, respectively.

RESULTS AND DISCUSSION

Table 3 presents the results of removal of sulfides, using all three amines, showing that commercial
amine DUOMEEN O promoted better removal, dissolved in 0.25 v:v into jet fuel, providing
approximately 76% removal of sulfides. The commercial ARQUAD amine 2 c-75, on the same ratio
v:v with jet fuel, showed approximately 59%. The commercial amine DUOMEEN T, at its best essay
showed lower efficiency (approximately 42%), when compared to other amines, even using a ratio
of 0.50 v:v into jet fuel. TABLE 3 also shows turbidity analysis and TOG on treated water.

Table 3 Results of removal of sulphides, TOG and Turbidity on treated water with three
commercial amines
DUOMEEN O
Am/QAV(%)

Removal (%)

TOG (mg/L)

Turbdty (NTU)

0.01

56.00

14.00

23.01

0.05

63.00

10.00

18.00

0.1

70.15

8.00

15.00

0.25

76.21

5.00

11.00

0.5

73.00

4.00

9.80

Am/QAV(%)

Removal (%)

TOG (mg/L)

Turbdty (NTU)

0.01

50.05

46.00

43.40

0.05

51.00

34.00

38.00

0.10

57.00

27.00

36.00

0.25

59.50

22.00

33.46

0.50

51.50

14.00

33.3

Am/QAV(%)

Removal (%)

TOG (mg/L)

Turbdty (NTU)

0.01

38.56

31.00

59.80

0.05

30.58

20.00

50.00

0.10

31.38

13.00

45.67

0.25

40.00

9.00

45.00

0.50

42.00

10.00

45.00

ARQUAD 2C-75

DUOMEEN T

According to TABLE 3, samples of treated water by DUOMEEN O presented, in 0.25 v:v

(amine:QAV), low values of TOG and turbidity, leading to removal of considerable amounts of oil
and sulfides.
Figure 2 shows the efficiency of removal of sulfides when using three commercial amines in bench
essays. It is observed that DUOMEEN O and ARCUAD 2C-75 amines showed best results with 0.25
v:v (amine/QAV), on a ratio of 1/3 (organic phase/produced water).

Efficiency of removal of
Sulfides (%)

DUOMEEN O
ARQUAD 2C75
DUOMEEN T
Ratio Amine/QAV (v:v, %)
Figure 2 Efficiency of removal of sulfides

TOG of treated water (mg/L)

Figure 3 presents the results of TOG for treated water samples. Analyzing the results in terms of
final TOG, it turns out that in 0.25 v:v these values are low (Duomeen O and Duomeen T), showing
that there is a good oil removal on this range, which is confirmed by results of turbidity in water
(Figure 4).

y = -8,001ln(x) + 9,425
R = 0,9927
y = -5,792ln(x) + 2,7193
R = 0,923

Ratio Amine/QAV (v:v, %)

DUOMEEN O
ARQUAD 2C-75
DUOMEEN T
y = -2,65ln(x) + 1,8493
R = 0,9935

Figure 3 Results of TOG for treated water with different concentrations of extractants

Figure 4 shows that the higher turbidity values for commercial amine DUOMEEN T are due to the
fact of its high solubility into water which is attributed to the number of unsaturation (higher than
the commercial DUOMEEN O amine). In this way, although the results could be favorable to final
TOG, in terms of turbidity we had an increased value in the same range, due possibly to the affinity
of this amine by water.

Turbidity of treated water

(NTU)

y = -3,885ln(x) + 39,783
R = 0,8613
y = -2,695ln(x) + 30,373
R = 0,9644

ARQUAD 2C-75
DUOMEEN O
DUOMEEN T

y = -3,533ln(x) + 6,8958
R = 0,9902
Ratio Amine/QAV (v:v, %)

Figure 4 Results of turbidity of treated water with different concentrations of extractants

CONCLUSION
Results of the studies showed that amine DUOMEEN O was the most efficient in the removal of
76% of sulfides, when using a proportion of 0.25% v:v, of amine into jet fuel. The amine Arcuad 2C75 showed an efficiency of 59%, in the same proportional ratio and amine DUOMEEN T, presented
42% for a ratio of 0.50. The results of final TOG and turbidity demonstrate the possibility of
removal of sulfides and oil at the same time when using the amine DUOMEEN O at 0.25% v:v. In
spite of DUOMEEN T to present similar physical properties to amine DUOMEEN O, removal
efficiencies were very distinct, once the values of TOG and turbidity were greater than expected. It
is known that the structure of amine DUOMEEN T has largest number of unsaturated bonds, which
could lead to a partial solubility in water.

REFERENCES
Bara, J. E.(2012) Potential for Hydrogen Sulfide Removal Using Ionic Liquid Solvents. Green Solvents II ,
properties and applications of ionic liquids.2012, pp 155-167
Bayati, F.; Shayegan J.; Noorjahan A. 2011. Treatament of Oil Field Produced Water by Dissolved Air
Precipitation/Solvent Sublation, Journal of Petroleum Science and Engineering 80-26-31.
Burns K. A.; Codi S.; Furnas M.T.; Heggies, D. Holdway, D.; King B. A. and MCallister F. A.(1999). Dispersiom
and Fate of Produced Formation Water Constituents in an Australian Nothwest Shelf Shallow Water
Ecosystem. Marine Pollution Bulletin 38:593-603.
Dogan, E.C.,Turker, M. Dagasan,L., Arslan, A. (2012) Simultaneous Sulfide and Nitrite Removal from
Industrial Wastewaters under denitrifying Conditions. Biotechnology and Bioprocess Engineering.
V.17, pp 661-668.
Frst, W., Renon, H., Representation of excess properties of electrolyte solutions using a new equation of state.
AIChE J. 39 (1993) 335.
Hadjev D., Aurelle Y. (1995). Phase inversion: a method for separation of fine liquid-liquid dispersions. The
Chemical Engineering Journal. 58 pp. 45-51.

Huttenhuis, P.J.G., Agrawal, N.J. , Hogendoorn J.A. , Versteeg G.F. Gas solubility of H2S and CO2 in aqueous
solutions of N-methyldiethanolamine. Journal of Petroleum Science and Engineering 55 (2007) 122
134.
Kohl, A., Nielsen, R. Gas Purification. 5th Edition, Gulf Publishing Company, Houston, 1997.
Moraes, N. A.; Paulo, J. B. A; Medeiros, G. S.;2011. Influence of Main Processes Variables on the treatment of
waste waters using a new technology (MSPI). Brazilian Journal of Petroleum and Gas. V.5, N.2.75-85
Rufford, T.E., Smart S., Watson, G.C.Y. , Graham, B.F. , Boxall, J. , Diniz da Costa , J.C., May, E.F. The removal
of CO2 and N2 from natural gas: A review of conventional and emerging process technologies.
Journal of Petroleum Science and Engineering 94-95 (2012) 123154.
Young G. parker; 2008. Removal study of hydrogen sulfide using natural second metabolites. Biotechnology
and bioprocess. 13:325-331.
Zhang, Y.; Gao B.; Lu, L.; Yue, Q. Wang Q.; Jia, Y. (2010). Tratment of produced water from polymer flooding
in oil production by the combined method of hydrolysis acidification-dynamic membrane bioreactor
coagulation process. Journal of Petroleum Science and Engineering 74 1419.

chapter 8
Mine water and
effluent treatment

Use of CuPRO MEX extractants in copper solvent

extraction plants

Mauro Torres
snf FloMin, South
America, Chile
Brant Sudderth
snf FloMin, Inc.
Flat Rock, usa
Pablo Tapia,
Cristin Salgado,
Alvaro Arancibia and
Hctor Bravo
Mantoverde Division,
Anglo American Chile

SNF FloMin, Inc. has introduced a complete line of extractants for various
metals recovery. Among them are the CuPRO MEX Brand of copper
extractants that are based on the well known 2-hydroxy-5
nonylsalicylaldoxime and 2-hydroxy-5-nonylacetophenoneoxime.
A complete metallurgical test program with CuPRO MEX extractants
was started at laboratory and pilot plant scale at three of Anglo American
Chiles Divisions. Testing methods and the chemical, physical and
metallurgical comparisons of the CuPRO MEX 3302 and Acorga M
5640 extractants at continuous pilot plant scale are presented (both
formulations are a modified aldoxime with TXIB modifier, equivalents in
performance and quality). Data mainly for the Mantoverde mine is
presented, and these results include: recovery, net transfer, Cu/Fe
selectivity, re-extraction data and phase disengagement time in each phase
continuity. The pilot plant results were completed at each Division in
side-by-side mini-pilot plants, calibrated for precise comparison of different
reagents and/or flowsheets. Results show that the chemical/metallurgical
and the physical performances of CuPRO MEX reagents are identical or
superior to the equivalent extractants formerly used at these mines.
Based on the results obtained in the metallurgical program it was
decided to conduct long term plant scale tests at the Mantos Blancos,
Mantoverde and El Soldado Divisions.
This paper presents the pilot plant and industrial plant results for SNF
FloMins CuPRO MEX reagents at Mantoverde Division.
CuPRO MEX is a Trademark of SNF FloMin, Inc. ACORGA is a
Registered Trademark of Cytec Industries, Inc.

INTRODUCTION
Copper solvent extraction reagents have been provided predominately by two suppliers for many
years BASF (formerly Cognis Corporation) [1] and Cytec [2]. Both suppliers have been present
under different corporate names, but the core people and the reagents and technology provided by
both have been relatively constant over many years. There have been several attempts by other
suppliers to enter this market, most notably Shell International Chemicals [3] from the UK in the
1970s and 1980s and by Allied-Signal [4] (Inspec) in the 1990s. Both of these businesses, including
their products and/or plants, were eventually purchased by Henkel (a predecessor to
Cognis/BASF). More recently, Star Orechem Industries Ltd. in India made an effort to enter this
important market. After some limited success, mainly in Africa, this company and its products and
plant were purchased by Cytec, leaving still only the two major suppliers to provide the global
supply of critical solvent extraction reagents for copper recovery.
These other suppliers failed largely due to their lack of long-term commitment to the mining sector
in their business portfolios. Now, SNF FloMin, Inc. has made the strategic decision to provide
copper solvent extraction reagents as a part of its extensive portfolio of mining related chemicals
and services. SNF FloMin, Inc. is the PERFORMANCE MINING REAGENTS Company with a large
global presence and a firm commitment to the long term growth of its various portfolios, and it is
now the second largest flotation chemical company in the world. The purpose of this paper is to
demonstrate the efficacy and commercial success of the latest solvent extraction products being
offered by SNF FloMin, Inc. and its various global subsidiaries.
To date, laboratory and pilot plant studies have been carried out at SNF FloMins
metallurgical/analytical laboratory in Chile and at Anglo American Chiles El Soldado, Mantoverde,
and Mantos Blancos mines. Plant scale industrial trials are being carried out at the El Soldado
Division since August, 2012; at the Mantos Blancos Division since April, 2013; and at the
Mantoverde Division since June, 2013.
This paper presents the testing methods and the chemical, physical and metallurgical comparisons
of SNF FloMins CuPRO MEX 3302 to the ACORGA M5640 extractant at continuous pilot plant
and industrial scale. Data mainly for Mantoverde mine is presented, and these results include:
recovery, net transfer, Cu/Fe selectivity, re-extraction data and phase disengagement time each
phase continuity. The pilot plant results were completed in side-by-side mini-pilot plants,
calibrated for precise comparison of different reagents or flowsheets.
Similar tests were performed at the El Soldado Division, however, the reagent used there was LIX
9790N, a product supplied by BASF. Results for CuPRO MEX 3302 were very superior to the LIX
Reagent, and it was displaced in the industrial plant in 2012 by CuPRO MEX 3302.
Some of the key results related to the above test programs are presented as follows:
Description of Reagents CPM 3302 and M5640
Pilot Plant Protocols
Pilot Plant Results
Industrial Plant Results
Conclusions from Tests and Trials Performed
LIX is a Registered Trademark of BASF.

The results of the chemical and physical performance at lab scale of CuPRO MEX 3302 and M
5640 were discussed elsewhere [5] and will not be repeated here. In this work it was found a lower
loading capacity of both plant organics, indicating clearly that the plant organics have lost the
original loading characteristics as found for a fresh reagent. This could be due in part to the fact that
the extractant was over-modified by the selective degradation of the oxime, keeping in mind that
the TXIB modifier used is very stable and builds in concentration with time.
This over modification is the main reason why both Mantoverde and Mantos Blancos Divisions
decided to start the additions of unmodified aldoxime CuPRO MEX 3406 to their plant organic.
Over time, this will dilute the TXIB modifier content of the plant organic and will improve physical
properties of the organic phase such as viscosity and density. The target was to achieve the
theoretical Aldoxime/TXIB content of the fresh CPM 3302/M 5640 reagent.

DESCRIPTION OF REAGENTS
CuPRO MEX 3302: is a mixture of 2-hydroxy-5-nonylsalicylaldoxime (~ 48 wt.%) and TXIB (2,2,4Trimethylpentanediol diisobutyrate, ~ 28-30 wt.%) in a low aromatic hydrocarbon solvent.
CuPRO MEX 3406: is a 2-hydroxy-5-nonylsalicylaldoxime (~ 48 wt.%) in a low aromatic
hydrocarbon solvent.
ACORGA M 5640 [5]: is a mixture of 2-hydroxy-5-nonylsalicylaldoxime (~ 48 wt.%) and TXIB
(2,2,4-Trimethylpentanediol diisobutyrate, ~ 25-30 wt.%) in a low to medium aromatic hydrocarbon
solvent.

PILOT PLANT PROTOCOLS

The pilot plant program of Mantoverde Division was carried out in two solvent extraction circuits
in series configurations, consisting of two stages of extraction, 2 stages of stripping and one stage of
washing (2ES x 2S x 1W). Continuous flowrates were 100 cc/min PLS flow, and the circuit is
illustrated in Figure 1 below:

Figure 1 Series configuration for the industrial plant and pilot plant. 2ES x 2S x 1W Circuit

Figure 2 Photo of the Mini-Pilot Plant with 2ES x 2S x 1W stages. Two parallel/identical circuits were operated
simultaneously

The characteristics of the aqueous solutions, PLS and spent electrolyte were:
Table 1 PLS composition for Mantoverde
PLS
Mantoverde

Cu2+
(g/L)
8.11

pH
2.1

FeT
(g/L)
11.3

Fe3+
(g/L)
2.9

Mn
(g/L)
3.2

SO42(g/L)
203

Al
(g/L)
17.7

Mg
(g/L)
16.6

Table 2 Spent electrolyte composition for Mantoverde

Spent Elect
Mantoverde

Cu2+
(g/L)
39.8

H2SO4
(g/L)
194

FeT
(g/L)
0.98

Fe3+
(g/L)
0.86

Mn
(mg/L)
23

Co
(mg/L)
210

ORP
(mv/SHE)
715

CUPRO MEX 3302 VS. ACORGA M5640 IN THE PILOT PLANT KEY RESULTS
This phase of testing was oriented to evaluate the chemical/metallurgical performance of the two
equivalent formulations of reagents, CPM 3302 and M5640 when tested against real plant solutions.
The test program included a base line test using plant organic of Mantoverde Plant. The pilot
plants were first operated in duplicate in order to validate that both pilot plants were performing
identically. Among the reagents tested in the pilot plants were:
1. 100 % plant organic M5640 (base line in duplicate).
2. 100 % fresh organic CPM 3302.
3. 100 % fresh organic M5640.

Mantoverde results: CuPRO MEX 3302 vs. Acorga M5640

The most important results from these tests are summarized in TABLES 3, 4, 5 and 6 below:
Table 3 Average results for the Extraction parameters at Mantoverde
Cu PLS

TRIAL

Extract.

O/A

g/L

Cu Raff

Cu Rec.

Cu Rec.

Copper Rec.

g/L

E1 %

E2 %

100% Plant Organic

8.1

1.14

19.7

0.48

50.4

88.1

94.1

100% Fresh M5640

8.3

1.16

19.8

0.46

53.9

87.9

94.4

100% Fresh CPM 3302

8.3

1.16

19.7

0.44

55.0

88.2

94.7

Table 4 Average results for the organic parameters at Mantoverde

Loaded Org.

TRIAL

Stripped Org.

Max Load

Net Transfer

Selectivity

Cu g/L

Fe mg/L

Cu g/L

Fe mg/L

g/L Cu/1% v/v

Cu/Fe

100% Plant Organic

9.3

6.5

2.6

0.1

83.8

0.341

1439

100% Fresh M5640

9.5

2.0

2.7

0.0

85.4

0.341

4833

100% Fresh CPM 3302

9.7

1.9

2.8

0.0

86.3

0.345

5085

Table 5 Average results for the stripping parameters at Mantoverde

Cu Spent

Cu Rich

Cu Cut

Elect, g/L

Elect, g/L

g/L

100% Plant Organic

40.3

57.4

17.1

100% Fresh M5640

40.9

59.1

100% Fresh CPM 3302

40.9

59.4

TRIAL

H2SO4 SE

Stripping

g/L

Efficiency, %

2.63

196

72.3

18.2

2.92

193

70.7

18.5

2.85

193

70.5

O/A

Table 6 Average results for phase disengagement time at Mantoverde

TRIAL

Phase Disengagement
Time in Extraction

Phase Disengagement
Time in Stripping
AC, sec
OC, sec

OC, sec

AC, sec

100% Plant Organic

131

129

98

97

100% Fresh M5640

86

83

75

81

100% Fresh CPM 3302

85

82

67

74

MANTOVERDE COMMERCIAL INDUSTRIAL PLANT KEY RESULTS

It is most important to compare the performance of the CuPRO MEX reagents to the former
reagents on the commercial plant level. Results of the Mantoverde chemical lab data for the last
three years are presented below to demonstrate the benefits of the change to CuPRO MEX
Reagents.
The characteristics of the aqueous solutions, PLS and spent electrolytes and the operational
parameters were:

Table 7 Average results for the operational parameters at Mantoverde

2011

2012

June-13/March-14
(SNF FloMin)

Cu2+ PLS, g/L

6.74

6.91

6.53

pH PLS

2.23

2.18

2.36

FeT PLS, g/L

10.4

10.5

10.3

O/A in Extraction

1.12

1.12

1.13

O/A in Stripping

2.63

2.61

2.65

Cu2+ Spent Electrolyte, g/L

39.1

39.4

40.2

H2SO4 Spent Electrolyte, g/L

181

188

182

Extractant, v/v %

19.7

20.1

19.7

22

25

24

Parameter

PLS Temperature, C

Key Performance Indicators are measured continuously at Mantoverde

Table 8 Average results for the operational KPIs at Mantoverde
Parameter

2011

2012

June-13/March-14
(SNF FloMin)

Extraction Efficiency, %

95.9

95.9

96.2

Net Transfer, g/L Cu/1 v/v %

0.28

0.29

0.28

Stripping Efficiency, %

61.4

62.9

61.1

Total Iron in Electrolyte, g/L

1.4

1.3

1.0

Manganese in Electrolyte, ppm

28

23

18

O/A Entrainment in Raffinate, ppm

75

70

47

Extractant Consumption, Kg/ton Cu

2.6

2.1

1.7

The following graph shows schematically the consumption of extractant in the last three years at
Mantoverde SX Plant.

Figure 3 Extractant Consumption at Mantoverde Plant

CONCLUSIONS
The following can be concluded from the laboratory and pilot plant tests comparing CuPRO MEX
3302 and Acorga M5640:
The extraction efficiencies for M5640 and CPM 3302 are both very similar if not identical.

Net transfer is also identical for both M5640 and CPM 3302.

Stripping efficiency for M5640 and CPM 3302 are both very similar.

The selectivities, Cu/Fe, for M5640 and CPM 3302 are both very good and are of the same
order of magnitude.

The phase separation times in both continuities, in extraction as well as in stripping, are
very similar for both CPM 3302 and M5640.

In summary, it can be concluded from the laboratory and pilot plant studies that the
chemical/metallurgical and the physical performances of CPM 3302 and M5640 are similar
if not identical.

Finally, it can be concluded that the reagent CPM 3302 is compatible with the aqueous
solutions from Mantoverde (Mantos Blancos and El Soldado as well), both the PLS and the
electrolytes, and it performed in excellent fashion both chemically and physically at all
three Divisions.

The above comparable results are expected since both reagents consist of the same organic
molecules and chemical formulations and are otherwise equivalent in every way.
Below are the conclusions from the results of the industrial plant program:

Clearly, CuPRO MEX 3302 is at least equivalent to the Acorga M5640 in all aspects.

The initial addition of CuPRO MEX 3406 non-modified aldoxime reagent proved to be an

The periodic addition of CuPRO MEX 3406 should be considered for continued optimum

efficient method to return the plant organic performance to its original state.
performance of the plant organic.

To date, the extractant consumption rate has improved since the addition of the CuPRO
MEX Reagents and the permanent teamwork between Mantoverde and SNF FloMin. The
results of the laboratory and pilot plant tests have now been confirmed on the industrial
scale.

ACKNOWLEDGEMENTS
The authors would like to thank the managements of their respective companies, Anglo American
Chile and SNF FloMin, Inc. for granting permission to publish this data. Also, thanks are extended
to the many engineers, chemists and technicians who worked so diligently to ensure the accuracy of
the data.

REFERENCES
Agers, D. W. and DeMent, E. R., The Evaluation of New LIX Reagents for the Extraction of Copper and
Suggestions for the Design of Commercial Mixer-Settler Plants, AIME Annual Meeting, San
Francisco, Calif. 1972 TMS Paper Selection A72 87.
BIRCH, C.P., The Evaluation of the New Copper Extractant P-1", Proceedings of the International Solvent
Extraction Conference, ISEC 74, Lyon, September, 1974.
Van der Zeeuw, A.J., International Symposium on Copper Extraction and Refining, Las Vegas, 1976, Extractive
Metallurgy of Copper, AIME, New York, 1976, 1039-1055.
M. Amores, A.G. Coedo, F.J. Alguacil, Extraction of copper from sulphate solutions by MOC 45: Application
to Cu separation from leachates of a copper flue dust, Hydrometallurgy, Volume 47, Issue 1,
November 1997, Pages 99112
M. Torres B., M. Torres M., C. Salgado, H. Bravo, J. Fernandez, B. Sudderth, CuPRO MEX Extractants as
Equivalent Alternatives for Copper Recovery Copper 2013-Cobre 2013, IIMCh, 2013, pp. 1-12.
Private Communication, Cognis Corporation, 1997.

Experimental study of entrainment of

phases in copper solvent extraction

Jonathan Castillo
Universidad de
Atacama, Chile
Patricio Navarro and
Cristian Vargas
Universidad de
Santiago de Chile

Solvent extraction (SX) is a technology that has been widely used in

extractive metallurgy. The SX has been positioned as a key technique for
purification and concentration of solutions containing copper ions. The
copper solvent extraction on an industrial scale employs a mixer-settler
reactor, in which in a first instance two immiscible liquid phases are mixed
by mechanical stirring, and subsequently they are recovered separately.
This operation has a number of difficulties as crud formation, degradation
of organic reagent and entrainment of phases. This last problem involves
the transportation of organic micro-droplets in aqueous phase (or vice
versa), which does not coalesce with the respective phase. This
phenomenon is common and causes problems such as contamination of
the rich electrolyte (reducing the quality of electrodeposited cathode), and
heap leach (to be irrigated with solutions containing organic droplets), etc.
In this paper the results to evaluate the entrainment of organic phases in
the aqueous at laboratory level are presented. The conditions tested are:
agitation speed, agitation time and extractant concentration. The
experiments were conducted in a reactor equipped with four baffles which
had a capacity of 1 L and an internal diameter of 10 cm. The organic phase
used consists of a commercial extractant and diluent (LIX 984N and
Shellsol-2046AR), the aqueous phase consists of a copper sulfate solution at
pH 2. To measure the entrainment of phases UV-VIS spectrophotometry
was used. The results indicated levels of entrainment of phases between
160 ppm and 312 ppm. An analysis of the variables studied shows that the
phenomenon is mainly influenced by the size of droplets and the speed of
the phase separation.

INTRODUCTION
This paper shows the experimental results to determine the levels of entrapments (or entrainment)
in copper solvent extraction. In this context, the organic entrapment in aqueous O/A) is defined as
organic phase droplets dispersed in an aqueous phase; otherwise it is called aqueous into organic
entrapment (A/O). The entrapment of phases occurs by imperfect phase separation resulting from
incorrect operation or fractions of very small droplets that are generated during mixing. A number
of factors can interfere with operational control of entrapment: phase inversion, sudden
temperature changes, sudden changes in the flow of the solutions, degradation of the organic
phase, and equipment design, among others (Morgan, Cashwell et al. 2011).
In the industrial solvent extraction (SX) plants, a routine task is the replacement of the organic
phase due to losses at different stages in the process, which implies an increase in costs. The loss of
the organic phase in an SX plant is mainly due to three causes: degradation of the organic phase,
crud forming, and entrapment of phases. (Liu, Qiu et al. 2006, Ning, Cao et al. 2009, Zagorodnyaya,
Abisheva et al. 2010). The most common organic losses in a solvent extraction plant are generally
produced in the last stage of extraction, this operational problem can be a significant complication
in the case of the bioleaching (Dopson, Sundkvist et al. 2006).
Solvent losses due to entrapment of phases occur in all commercial solvent extraction plants,
commonly 10-100 ppm of (O/A) entrainment. The main problems are loss of reactants and transfer
trapping impurities, especially chloride in the electrowinning (EW) (Szymanowski 1993, Miller,
Readett et al. 1997).
The impurities transferred from one stage to another may cause any of the following problems in
the LIX-SX-EW plants:
Corrosion of permanent cathodes
Increasing the anodic corrosion
Contaminated copper cathodes
Impregnation of the leach pads with the organic phase
Cost increases due to loss of organic
Evolution (oxidation) of detrimental species such as manganese, iron and chloride
Local pollution with chlorine gas (evolution of chloride ion)
Increased costs for implementation and auxiliary equipment such as filters.

METHODOLOGY
The experimental methodology considers a series of batch tests in mechanically agitated reactors.
For this a synthetic PLS solution was used, and the organic phase was prepared based on
commercial reagents.

Equipment
The batch tests were performed in a cylindrical glass reactor with a total capacity of 1.0 liter (flat
bottom) which had an internal diameter of 10 cm and 14 cm high, it was equipped with four
stainless steel baffles of 1.0 cm wide.
The reactor maintains the temperature through a cooling jacket. The mixing impeller includes a
disk with six blades on the base and two blades on the surface; the impeller is mounted in the
center of the mixer-settler. Figure 1 shows the scheme of the reactor.

Figure 1 Reactor scheme (a) and impeller (b)

Molecular absorption (UV-VIS) to quantify the aqueous organic entrapments was used. The sample
preparation used cylindrical funnels (25 mL.) and reciprocal shaker (400 rpm).

Reagents
The organic phase consists of Shellsol 2046AR as diluent and LIX 984N as extractant both are
industrial quality.
The aqueous phase contains copper sulfate pentahydrate (CuSO4*5H20), anhydrous sodium sulfate
(Na2SO4) and sulfuric acid (H2SO4). All reagents are p.a. quality. Distilled water to dissolve and
prepare the electrolyte was used.
Cyclohexane (p.a. quality) for phases trapping analysis was used.

Procedure
A series of liquid/liquid extraction batch tests were performed. The tests consist of mixing organic
and aqueous phase by mechanical agitation. When the phases were separated 10 mL of aqueous
were sampled, the sample was placed in a funnel with equal volume of cyclohexane shaking it for 5
minutes, and the cyclohexane sample was analyzed at 223 nm. The result was correlated with the
calibration curve to obtain the magnitude of the entrapment (ppm).
The speed of the phase separation was determined using a scientific camera (Sony model Flea 2.0,
monochrome CCD sensor 2/3 "and equipped with an optical Fujinon 12.5 mm). The average droplet
size was estimated from the mathematical expression published by Kumar and Hartland [11],
which is shown in Equation 1.

v0

0, 53 c g 3 1
12 c
1 1
0, 53 c
108 c2 1 4, 56 0 ,73

(1)

Where vo is the initial rate of the phase separation, is the mean droplet diameter, , , correspond
to the density and viscosity (the subscript c refers to the continuous phase), and is the fraction of
dispersed phase.

RESULTS AND DISCUSSION

The experiments were carried out at 20 C, aqueous continuity, phase ratio 1:1, copper
concentration of 6 g/L, and initial pH of 2. The following variables were studied: Stirring speed,
stirring time and extractant concentration in organic phase.

Stirring time
Three levels of the stirring time effect were studied:
experimental results are shown in Table 1.

3, 10 and 30 minutes (at 800 rpm), the

Table 1 Quantification of the entrapment of phases test at different stirring time

Condition Phase entrapment, ppm V0, mm/s
3 min
164.99
1.579
10 min
188.06
1.370
30 min
227.94
1.361

, mm
0.3415
0.3167
0.3089

The results of TABLE 1 show a clear tendency with the increasing stirring time, the entrapment of
phases increased and a maximum of 227.94 ppm were obtained. The results are attributable to the
increase in the fraction of small droplets that are generated due to over-agitation, thus the average
diameter of drops decreases from an average value of 0.3415 mm to 0.3089 mm.

Stirring speed
The study of the stirring speed on the entrapment of phases considers three levels of 400, 800 and
1000 rpm. The mixing time was set at 3 minutes. TABLE 2 shows the results to evaluate the speed of
stirring.
Table 2 Quantification of the entrapment of phases test at a variable velocity of stirring
Condition
400 rpm
800 rpm
1000 rpm

Phase entrapment, ppm

296.13
164.99
312.67

V0, mm/s
2.038
1.579
1.370

, mm
0.3890
0.3415
0.3178

The results of TABLE 2 show a parabolic trend with respect to the entrapment of phases. At 800
rpm a minimum value (optimum) of 165 ppm is observed. The explanation is based on the effect of
speed of phase separation and the generation of a fraction of small sizes. In the case of stirring the
mixture at 400 rpm the phase separation proceeds extremely rapidly (2.038 mm/s), the advancing
front is moving faster than a certain amount of droplets, which move slowly, and they are finally
trapped in the electrolyte. On the other hand, stirring at 1000 rpm generates a large fraction of very
small droplets (mean droplet diameter of 0.3178 mm) which move slowly, and become trapped in
the aqueous phase. The value increases to 312.37 ppm.

Extractant concentration
To evaluate the entrapment of phases function of the extractant percentage, three levels of
extractant concentrations (10, 20 and 30% v / v) were considered. The phases were mixed at 800 rpm
for 3 minutes; the results are summarized in TABLE 3.
Table 3 Quantification of the entrapment of phases at different extractant concentration
Condition Phase entrapment, ppm V0, mm/s
10% E
152.53
1.579
20% E
172.91
1.534
30% E
263.22
1.450

, mm
0.3516
0.3637
0.3770

The first analysis emerges from the results of TABLE 3. This result shows that increasing the
concentration of extractant in the organic increases the entrapment of phases; this is due to the
influence of the extracting agent in the physical properties. As the concentration increases, so does
the viscosity and density of the organic phase, therefore, the speed of the phase separation has the
opposite behavior; consequently, the phase separation speed has the opposite behavior.

CONCLUSION
Regarding the speed and stirring time, the result is due to the generation of small droplets and slow
movement with respect to the average velocity of the interface. If the speed of displacement of a
drop is smaller than the drop of the interfase, the entrapment of phases eventually occurs.
In the case of the influence of the concentration of the extractant, the results obtained were due to
the strong influence of extractant in the physical properties of the organic phase, causing an
increase in the viscosity and density of the droplets, which move slowly through the aqueous
phase.

ACKNOWLEDGEMENTS
The authors acknowledge financial support from the Universidad de Santiago de Chile and also
appreciate the support provided by the Universidad de Atacama.

REFERENCES
Dopson, M., J.-E. Sundkvist and E. Brje Lindstrm (2006). "Toxicity of metal extraction and flotation chemicals
to Sulfolobus metallicus and chalcopyrite bioleaching." Hydrometallurgy 81(34): 205-213.
Liu, X.-r., G.-z. Qiu and Y.-h. Hu (2006). "Degradation of Lix984N and its effect on interfacial emulsion."
Journal of Central South University of Technology 13(6): 668-672.
Miller, G. M., D. J. Readett and P. Hutchinson (1997). "Experience in operating the girilambone copper SX-EW
plant in changing chemical environments." Minerals Engineering 10(5): 467-481.
Morgan, J., W. Cashwell, S. Doole and J. Steeples (2011). "Entrainment Reduction at Freeport-McMoRan
Copper & Gold Morenci Operations." Solvent Extraction and Ion Exchange 29(5-6): 854-867.
Ning, P., H. Cao, C. Liu, Y. Li and Y. Zhang (2009). "Characterization and prevention of interfacial crud
produced during the extraction of vanadium and chromium by primary amine." Hydrometallurgy
97(12): 131-136.
Szymanowski, J. (1993). Hydroxyoximes and Copper Hydrometallurgy, CRC PressINC.
Zagorodnyaya, A. N., Z. S. Abisheva, S. E. Sadykanova, V. V. Bobrova and A. S. Sharipova (2010). "The
characterisation and origins of interphase substances (cruds) in the rhenium solvent extraction circuit of
a copper smelter." Hydrometallurgy 104(2): 308-312.

Outotecs hydrometallurgical process

for recovery of copper from
challenging concentrates

Kari Valkama, Marko

Lampi and Janne Karonen
Outotec, Finland

Interest in exploiting low-grade and low quality copper ore deposits is

increasing due to the growing copper demand and declining ore grades. In
addition, diminishing oxide copper resources at existing solvent extraction/
electrowinning plants encourage us to find a way to economically leach
copper sulfide minerals.
Outotec has been active in developing hydrometallurgical copper
processing technologies and has launched a novel chloride based technology
suitable for treatment of a wide variety of copper concentrates including
concentrates with primary copper sulfides. Standard, readily available and
proven equipment can be utilized in the process and construction materials
are well defined and tested on an industrial scale. The technology is based on
atmospheric reactor leaching, solvent extraction and electrowinning of
copper. The majority of required acid is regenerated in copper
electrowinning from where it is transferred to leaching through solvent
extraction. Sulfuric acid is used for acid make-up. Copper minerals are
leached in sodium chloride solution near the boiling point of the solution
with oxygen feed and iron is precipitated during leaching. By-product silver
and gold are recovered if present in the concentrate. Chloride solutions allow
a selective copper recovery from a bleed stream as copper hydroxychloride,
atacamite, in a simple and a cost effective process. The process was tested
continuously in a bench scale pilot and it demonstrated high extractions in
both leaching and metals recovery circuits.

INTRODUCTION
World primary copper production is currently 17 million tons annually and a major part is
produced utilizing sulfidic raw materials (Edelstein, 2013). Smelting and electrorefining are the
main technology and chalcopyrite is the dominant copper mineral. There are less available good
quality ore bodies in the future and impurities in feed materials are increasing which increases the
costs of smelting. About 20 % of the total copper production is based on oxidic resources where
copper is produced by heap leaching, solvent extraction and electrowinning (Schlesinger et al.,
2011). Good oxidic ores are also depleting and oxidic copper operations could be converted into
sulfide mineral processing in the future. However, primary sulfides are challenging for the
conventional copper heap leaching.
Outotec has been following up these trends closely and technology development has been directed
to meet the future challenges in copper production industry. Hydrometallurgical chloride based
processing has been identified to be a very potential option to directly treat copper bearing sulfide
feed materials, especially chalcopyrite.
This paper presents the key features of Outotecs latest innovation regarding the
hydrometallurgical copper processing. The main driver for the development has been the need for
hydrometallurgical refining method with following characteristics:
Process needs to be suitable to process wide variety of feed materials in terms of copper
grade, impurity concentrations and mineralogy.
Precious metals recovery (gold and silver) is an integral part of the process.
Processing technology has to be based on proven technical solutions, unit processes and
equipment.
Process needs to be robust, inherently safe and deliver excellent environmental performance.
Metallurgical performance should be high in terms of metal recovery and finished product
quality.
Technology should be suitable for retrofits e.g. for existing heap leach SX EW operation,
which are converting to sulfides processing.

Process description
Outotecs chloride based process for copper refining is suitable for varied copper concentrates. It
has been designed especially for primary sulfides such as chalcopyrite and bornite but it is also
suitable for more easily leachable secondary sulfides such as chalcocite, covellite and digenite.
Copper concentrate is leached in atmospheric conditions and chalcopyrite passivation is avoided by
chlorides in the leach solution. Dreisinger (2009) has reviewed proposed processes in literature for
chalcopyrite leaching and they generally utilize either pressure leaching with or without a small
amount of chlorides, ultrafine grinding or high chloride concentration. Total pressure oxidation has
high oxygen consumption and acid generation, autoclaves in pressure leaching have high
maintenance costs; fine grinding is detrimental to solid/liquid separation leading to high CAPEX
and finally high chloride concentration makes hydroxyoxime based conventional copper SX-EW
uneconomical. Outotecs new process does not require fine grinding, pressure leaching or high
chloride concentrations. The concentrate is leached at moderate chloride concentration and sulfuric
acid is used instead of hydrochloric acid.

Copper is leached from the concentrate by atmospheric reactor leaching, purified with copper
solvent extraction and finally recovered with a traditional sulfate based copper electrowinning as
shown in figure 1. Process also includes silver solvent extraction and recovery. Most of the raffinate
from solvent extractions is recycled back to leaching.

Figure 1 Process block diagram

Copper and silver are leached at pH 2.02.8 with cupric ions in chloride solution according to
equation (1) which is the same reaction as in Outotecs HydroCopper process (Lundstrm et al.,
2005). Moderate chloride ion concentration is also required to solubilize silver because with a low
chloride concentration silver would precipitate as silver chloride. Oxygen is used to oxidize
cuprous ions back to cupric ions (2) and the acid comes from solvent extraction raffinate and makeup sulfuric acid.

CuFeS 2 (s) + 3 Cu 2 + (aq) 4 Cu + (aq) + Fe 2 + (aq) + 2 S 0 (s)

(1)

4 Cu + (aq) + O 2 (g) + 4 H + (aq) 4 Cu 2 + (aq) + 2 H 2 O(aq)

(2)

Iron is precipitated during copper leaching and the acid generated in iron precipitation is also used
for cuprous oxidation. The exact mineralogy of the iron precipitate depends on the operation

parameters. A small part of the elemental sulfur will oxidize to sulfuric acid during leaching. Gold
is not leached in these conditions and it can be recovered from the tailing (Miettinen et al., 2013).
As long as chloride concentration is sufficiently high, it does not affect greatly on copper leaching.
Copper is purified with solvent extraction using hydroxyoxime based reagents. Both
aldoxime/ketoxime blends and modified aldoximes are suitable for copper recovery. The higher the
chloride concentration is, the more prevailing copper chloride species will be and this hinders
copper solvent extraction. The optimal range of chloride concentration in this process is between
60115 g/L as it maximizes copper leaching without hindering copper solvent extraction too much.
A moderate chloride concentration can be compensated with increased number of stages or higher
O/A phase ratio in solvent extraction. A washing stage is required for chloride removal from the
loaded organic phase. Copper is stripped with lean electrolyte from Cu EW.
Silver is extracted from copper solvent extraction raffinate by tri-isobutylphosphine sulfide
(Cyanex 471x). The extraction is based on solvation mechanism as shown in eq. (3) where L
represents tri-isobutylphosphine sulfide.

Ag + (aq) + Cl - (aq) + L(org) AgClL(org)

(3)

Acid in the raffinate from copper solvent extraction does not affect negatively on silver extraction
(Abe & Flett, 1992). However, increasing chloride concentration would have an adverse effect on
silver solvent extraction as it makes AgCl2, AgCl32 and AgCl43 more prevailing species. If
concentrate contains low amounts of silver, a part of the Cu SX raffinate may bypass the Ag SX, in
order to concentrate silver in the solution and reduce the size of the solvent extraction. The loaded
organic is scrubbed with acidic solution and the silver is stripped with a thiosulfate solution.
Raffinate from solvent extraction contains the acid generated in copper electrowinning and it will
be recycled back to leaching. A small bleed stream has to be taken from the raffinate in order to
control the build-up of impurities such as Zn, Mg, etc. The bleed stream contains also copper
because copper solvent extraction is not a suitable process for near total removal of copper from a
PLS that has a high copper concentration. Chlorides in raffinate allow a selective precipitation of
residual copper as atacamite, Cu2(OH)3Cl. It is subsequently leached with raffinate and sulfuric
acid. The solution is returned to leaching.
Finally impurity metals are precipitated from bleed solution coming from atacamite precipitation.
Process solutions do not contain arsenic or antimony as they are co-precipitated with iron in
leaching. Mercury is co-extracted with silver in SX and thus removed from the process.

Materials of construction for chloride milieu

Materials selection and equipment design are in a central role in every industrial project in order to
build a successful operation. This applies also on chloride hydrometallurgical technologies. Outotec
has been working with chloride technologies for decades and has accumulated a large amount of
information and knowledge about material selection and equipment engineering. This work
includes the operation of a demonstration plant for about two years in Outotec Research Center in
Pori, Finland. During this operation industrial scale equipment (reactors, agitators, tanks,
thickeners, filter, pumps, heat exchangers, instruments, etc.) were operated in real process
conditions. Based on this experience a material specification documentation has been created for
hydrometallurgical chloride processing technologies in general and also specifically for copper

processing. Suitable and cost effective materials for the equipment can be selected from various
polymers, plastics and composites. Metals like titanium are used in special cases such as heat
exchangers and instrumentation.

METHODOLOGY
The main steps of the process were demonstrated in a bench scale pilot at Outotec Research Center
in Pori, Finland. The experimental settings consisted of copper and silver leaching, copper solvent
extraction, copper electrowinning, silver solvent extraction and bleed treatment.

Materials
A copper concentrate was used as the feed material. The concentrate contained 29.6 % Cu, 30.6 %
Fe, 33.4 % S, 2.18 % Zn and 0.38 % Pb. Practically all copper was in chalcopyrite. Other main
minerals were pyrite, sphalerite, pyrrhotite and galena. The concentrate used in the pilot had
particle size of P80 71 m and P50 31 m.

Equipment
The leaching area consisted of a pulping reactor and 45 leaching reactors. The effective volume of
the pulping reactor was 1 L and its pH and temperature were not controlled. The effective volume
of the first leaching reactor was 3 L and the rest leaching reactors were 5 L each. The concentrate
was continuously fed to the pulping reactor with a vibrating feeder in one hour batches. Oxygen
was fed under a mixer in all the leaching reactors. Temperature was 95 oC and pH targets were 1.7
and later 2.2. pH was controlled in the first reactor with acidic raffinate from solvent extraction and
any excess raffinate was fed to the second leaching reactor. pH in other reactors was controlled with
concentrated sulfuric acid or calcium hydroxide slurry. Reactors also contained a reflux condenser
and water feed to compensate the evaporated amount. After leaching the slurry was filtered and
filtrate was fed to copper solvent extraction.
Glass mixer-settlers were used in solvent extraction with 0.21 L mixers and 1 L settlers. The settlers
had a band heater and the temperature in mixers varied between 3040 oC. pH controls were not
used in any stages. The number of stages in copper solvent extraction was 3-2-2 for extraction,
washing and stripping, respectively. The number of stages in silver solvent extraction was 2-1-2 for
extraction, washing and stripping, respectively.
Rich electrolyte from copper stripping stage was pumped through a heated line to a recirculation
tank and from there to copper electrowinning. A part from the lean electrolyte from electrowinning
was pumped back to stripping stage and the rest went to recirculation tank. Total solution volume
in electrowinning cell and tanks was 44 L. Cobalt was added to the electrolyte at 100 mg/L
concentration in the beginning and guar was added continuously.
Most of the Ag SX raffinate was recycled back to leaching but a part of it was collected for bleed
treatment run. Bleed treatment was ran for 8 h from time to time once there were enough solution
for it. It contained two 2 L reactors in copper removal and one 5 L reactor in impurity removal.
Calcium hydroxide was fed according to pH control and the temperature was 40 oC. The residence
time was 2.6 h in copper removal and 3.5 h in impurity removal.

Analysis
Metals from solution samples were analyzed with inductively coupled plasma - optical emission
spectrometry (ICP-OES). Chloride concentration was analyzed with either ion chromatography or
ion selective electrode. Solid samples were first dissolved completely and then the solutions were
analyzed with ICP-OES.

Calculation of results
The exact measurement of the mass of the solids during leaching is inaccurate so leaching yields
Y(Metal) have been calculated (4) with respect to iron which was almost completely precipitated.
c(Metal, solid) is concentration of metal in solids of a sample and c(Metal, solid)0 is a concentration
of metal in solids of the feed.

c(Fe, solid) 0 c(Metal, solid)

Y(Metal) 1
c(Fe, solid) c(Metal, solid) 0

100

(4)

Concentrations in organic phase in solvent extraction were mostly calculated from mass balances
according to eq. (5) where Qaq and Qorg are aqueous and organic volumetric flow rates, c(Cu,org)0,
c(Cu,aq)0 and c(Cu,aq) are copper concentrations in either organic or aqueous phase before and
after a stage. Only barren organic phase samples were stripped empty and a copper concentration
in organic phase was calculated from the batch stripping.

c(Cu, org)

Q aq
Q org

(c(Cu, aq) 0 c(Cu, aq)) c(Cu, org) 0

(5)

Extent of silver extraction was calculated according to eq. (6) where c(Ag,aq)0 is silver concentration
in the Ag SX feed and c(Ag,aq)1 is silver concentration in the raffinate.

E(Ag) (c(Ag, aq) 0 c(Ag, aq) 1 )/c(Ag, aq) 0 100

(6)

RESULTS AND DISCUSSION

Leaching
The residence times were 16 h with four reactors and 19 h with five reactors. Chloride concentration
in PLS was between 8095 g/L during the pilot. Redox potentials in leaching reactors were during
the whole run in the range where chalcopyrite is leached in a chloride solution. Redox potentials
(Ag/AgCl vs Pt) in the last leaching reactor were between 520560 mV. Copper yields are shown in
figure 2. Copper leaching yields were generally above 95 % after the introduction of the fifth
leaching reactor. Silver and zinc leaching yields in the leaching reactor five were between 9394 %
and >97 %, respectively. Zinc yield changes followed copper yield changes quite accurately during
the pilot. Some bypass flow of the fines of the concentrate was observed during the pilot which
decreased leaching yields. This is due to small scale of the equipments where for example oxygen
efficiency is much lower than in industrial scale reactors. This leads to a higher ratio of oxygen flow
per cross section of the reactor which pushes fines to reactor overflow. The effect of bypass flow is
clearly seen between 50200 h after which it was controlled better by adjusting oxygen flow and
mixer speeds.

Figure 2 Copper leaching yields from leaching reactors 4 and 5 according to eq. (4)

Initially the leaching pH was 1.7 and color of the leach residue was dirty yellow. After 200 h the
leaching pH was increased to 2.2 and the color of the leach residue changed to red. Figure 3 shows
the effect of leaching pH to iron and total sulfur concentrations in solution. X-ray diffraction (XRD)
measurements confirmed that iron was mainly jarosite during 0-200 h and after that hematite and
goethite were detected.

Figure 3 The effect of iron precipitation form on iron and total sulfur concentrations in solution in leaching
reactors 4 (LR4) and 5 (LR5)

Solvent extractions
Typical measurements of copper solvent extraction from the pilot are shown in figure 4. At least
10 g/L copper should be left in the raffinate because of cupric ions role in oxidation in leaching. No
problems with impurities were encountered during the run. Concentrations of chloride, iron and
zinc in rich electrolyte are shown in figure 5. A slight build-up of zinc was noticed. Electrolyte from
recirculation tank was also analyzed but the chloride concentrations were around 10 mg/L. Two
washing stages were used in the pilot because lab scale mixer-settlers do not have a separate pump
and a mixer. Thus, the mixers have to function also as pumps between the stages which leads to
extensive entrainment and a high impurity load to washing. Outotec VSF technology has
dispersion overflow pump (DOP) and Spirok mixer which are able to minimize entrainment
(Pekkala et al., 2007; Nyman et al., 2003; Laitala, Pekkala & Rodrguez, 2007). One washing stage
would be sufficient in actual process based on Outotecs experience on aqueous in organic
entrainment with high copper extractant concentrations and VSF technology in industrial scale.

Figure 4 McCabe-Thiele diagrams of copper extraction and stripping in the pilot

Figure 5 Copper concentration and impurity build-up in the rich electrolyte

The silver extraction in silver solvent extraction is shown in figure 6.

Figure 6 Extent of extraction of silver

Copper electrowinning
The electricity current was adjusted so that lean electrolyte would contain 35 g/L Cu and
180 g/L H2SO4. This meant that due to the fluctuations in the rest of the process, current density was
not always at design value. The current density was generally between 250300 A/m2 once the
extraction had been stabilized. Copper cathode quality fulfilled LME A grade specifications during
the pilot.

Copper removal from bleed

The main components in the precipitate in copper removal were atacamite, gypsum and zinc
hydroxide. Residual iron was also precipitated so that iron concentration in solution was <2 mg/L.
Copper and zinc concentrations in solution are shown in Figure 7. When the dilution due to
addition of calcium hydroxide slurry was taken into account, the initial copper and zinc
concentrations in Figure 7 were 9.8 g/L and 9.0 g/L, respectively. The zinc concentration in the
solution depends on the concentrate impurities and the relative amount of bleed stream with
respect to raffinate. Even with this relatively high zinc concentration only a portion of zinc
co-precipitated at acceptable copper removal levels.

Figure 7 Copper and zinc concentrations in atacamite precipitation

CONCLUSIONS
A novel hydrometallurgical copper processing technology for primary copper sulfides has been
developed and process viability has been demonstrated for a purely chalcopyritic copper
concentrate in a 16 day continuous bench scale pilot operation. The pilot included the main parts of
the process and LME A grade copper cathode was produced from a concentrate. The process
utilizes only proven unit operations and it is suitable also for poor copper concentrates with low
copper concentrations and high impurity concentrations. A significant chloride concentration has to
be taken into account in materials of construction but Outotec has a very good knowhow of
materials in a chloride process on industrial scale at demonstration plant at Pori, Finland.
Process demonstrated very good performance in the pilot:
Leaching extractions for copper and silver are high, >95 % and >93 %, respectively
Conversion from chloride solution to sulfate solution in copper solvent extraction is very
efficient with negligible chloride transfer to electrowinning circuit
Copper cathode is of high quality and exceeds LME-A cathode chemical purity specification

REFERENCES
Abe, Y., Flett, D.S. (1992) Solvent extraction of silver from chloride solutions by Cyanex 471X, in: T. Sekine, S.
Kusakabe (Eds.), Solvent Extraction 1990, Elsevier Science Publishers B.V., Amsterdam, The
Netherlands, pp. 11271132.
Dreisinger, D. (2009) The Hydrometallurgical Treatment of Copper Concentrates, Perumin 29 Convencin
Minera, Arequipa, Peru, 14-18 September.
Edelstein, D.L. (2013) Copper, Mineral Commodity Summaries 2013, U.S. Geological Survey, Reston, Virgina,
pp. 49.
Laitala, H., Pekkala, P., Rodrguez, C. (2007) Innovations in the Outokumpu Technology`s VSF Cu SX
Process to reduce the total Investment cost while Maintaining High SX Process Efficiency,
HydroCopper 2007, Via del Mar, Chile, 16-18 May.
Lundstrm, M., Aromaa, J., Forsn, O., Hyvrinen, O., Barker, M.H. (2005) Leaching of chalcopyrite in cupric
chloride solution, Hydrometallurgy, vol. 77, No. 12, pp. 8995.
Miettinen, V., Haapalainen, M., Ahtiainen, R., Karonen, J. (2013) Development of chloride gold process,
Proceedings Gold Conference ALTA 2013, Perth, Australia, 25 May - 1 June.
Nyman, B., Ekman, E., Kuusisto, R., Pekkala, P., (2003) The OutoCompact SX Approach to Copper Solvent
Extraction, JOM, vol. 55, No. 7, pp. 2730.
Pekkala, P., Laitala, H., Nyman, B., Ekman, E., Weatherseed, M. (2007) Performance highlights of a modern
VSF SX Plant, ALTA 2007 Copper Proceedings, Perth, Australia, 24-25 May.
Schlesinger, M.E., King, M.J., Sole, K.C., Davenport, W.G. (2011) Extractive metallurgy of Copper, 5th ed., Elsevier,
Oxford, United Kingdom, pp. 810.

Change in extractant formulation

in Minera Franke

Francisco Villegas,
Ral Manrquez
and Ral Funes
kghm International, s.c.m
Franke, Chile

S.C.M. Franke located south-east of Taltal, Antofagasta, Chile, is a plant

Rodrigo Zambra and

David Acevedo
Cytec Chile

(SX), due to the high viscosity in the PLS (above 8 cP), mainly because of

designed to produce 30,000 tons of copper cathods per year via the
conventional hydrometallurgical process. The first years of operation of
Franke Mining presented several problems in the area of solvent extraction
their high content of Al and Mg, which resulted in a decrease in the stage
efficiency, with a negative impact in the SX recovery. Furthermore, this
high aqueous viscosity resulted in a high consumption of extractant and
diluent (high levels of entrainments O/A). Additionally, the content of iron
in the PLS is one of the highest in the region, reaching 32 gpL.
As part of the action plan to process this viscous PLS, SCM Franke
decided to change the formulation of the extractant from a ketoxime/
aldoxime blend (ratio 75/25) to an aldoxime modified with ester
(ACORGAM5910), due to its faster extraction kinetics which increases the
stage efficiency and diminishes the extractant requirements and
consequently the consumption. Additionally, improvement on the Cu/Fe
selectivity of this new formulation was measured.
This work shows the plant results after 12 months of operation with this
new formulation, confirming the expected good results in terms of higher
stage efficiency, lower extractant consumption and higher Cu/Fe selectivity.

INTRODUCTION
The deposit of SCM Franke, a fully-owned subsidiary of KGHM Polska Mied S.A., has a
hydrometallurgical plant designed to produce a maximum of 30.000 tons Cu/year. It is located in
district of Altamira within the municipality of Taltal, at an approximate distance of 77 km to the
southeast from the city of Taltal and 52 km northeast El Salvador and at an approximate altitude of
1.750 a.s.l.
The solvent extraction plant was designed according to the Outotec technology (VSFX Plant) with
an interlaced series parallel circuit (See Figure 1). Initially, the plant started its operation with an
extractant based on a 75/25 ketoxime/aldoxime mix, basically due to the potential presence of
nitrate in PLS and its configuration include only one stripping stage.
After a couple of years of operations, the PLS feeding the solvent extraction plant began to quickly
increase its viscosity (7cP). Analysis indicated a high level of sulfate in the PLS, with a high
presence of aluminum and magnesium, which are the main elements that influence the viscosity. In
addition to the above, the level of Fe in PLS increased to 32 gpL, the highest concentration in the
region. The problems of high PLS viscosity immediately started to produce an increase in the levels
of O/A entrainments, thus increasing the consumption of extractant and solvent. In addition the
high viscosity produced well-known operational problems such as heap impregnation, plugging of
the irrigation system, etc.
Furthermore, the high viscosity results in insufficient power from the mechanic agitation of both
phases, thus deteriorating of the mixer and the solutions approach to equilibrium. Low mixer
efficiency negatively impacts the global extraction recovery in SX.
In 2012 the situation became even more critical as the PLS feeding the SX plant increased futher in
viscosity, reaching values as high as 15 cP. This raised the specific consumption to 5,20 kg
Extractant/Ton Cu and 43 L/Ton Cu resulting from the increase of losses by organic entrainments.
Based on the above, SCM Franke carried out a work plan in order to avoid and mitigate the effects
that high viscosity produced on the solvent extraction process. Among the different works
performed, Cytec was requested to evaluate an alternative formulation of extractant considering the
use of the Acorga products, in order to:

Reduce the specific rate of consumption of the extractant.

Increase the stage efficiency (SE).
Increase the Cu recovery.
Increase the Cu/Fe selectivity.

The results of the pilot metallurgic tests and the results in the plant of the new formulation of
extractant are shown below.

Figure 1 Solvent extraction plant Configuration, SCM Franke

BACKGROUND INFORMATION
A common problem in the design of SX plants, process philosophy and auxiliary equipment, is the
ability to accurately determine the equilibrium of impurities on the basis of the escalation of leach
tests. These PLS impurities come from other elements dissolved into the aqueous phase during the
copper leach.
As known, in the LX-SX-EW copper operations, copper is transferred to the EW stage via SX, but
other elements are accumulated in the leach circuit until they reach an equilibrium value. It is quite
common not to consider impurities such as aluminum and magnesium due to the difficulty to
estimate their values in nominal conditions of the leach process, and only consider the impurities
such as iron, chloride, manganese or nitrate.
The accumulation of aluminum and magnesium may significantly increase the viscosity of the PLS,
thus generating a negative physical behavior, increasing the O/A entrainments and consumption of
reagents. At the same time, there is a potential reduction in the transfer of copper due to lower
stage efficiency (SE). Figure 2 (H. Yez and A. Soto, 2009), shows the effect of the levels of
aluminum and magnesium on the viscosity of the PLS.
Impact of impurities content on PLS viscosity
Synthetic solutions: Cu 6 gpl, Fe, 10 gpl, Al 15 gpl, Mg 15 gpl
100
90
80
70
60
50
40
30
20
10
0

Cu

Cu/Fe
Cu/Fe/Al
Contribucin a viscosidad, %

Cu/Fe/ /Mg

Figure 2 Impact of Al and Mg on PLS viscosity

Cu/Fe/Al/Mg

The physical and chemical effects of viscosity on the solvent extraction process are explained below.

Physical effect
The physical effect of PLS viscosity on the solvent extraction process is, basically, its influence on
the phase disengagement time mainly in aqueous continuity (figure 3). This often results in a larger
dispersion band or higher entrainment due to the high aqueous viscosity hindering organic droplet
coalescence or settling (Figure 4).

PD, seg

Phase disengagement times

450
400
350
300
250
200
150
100
50
0
PLS 1 (7.81 cP)

PLS 2 (4.21 cP)

PLS 3 (2.80 cP)

Org cont.

PLS 4 (1.69 cP)

Aq. cont

Figure 3 Phase disengagement time in virtue of the PLS viscosity

EFFECT OF VISCOSITY IN PLS ON THE DISPERSION BAND

E-1 STAGE

Dispersion Band (mm)

20
18
16
14

y = 0,5866x 2 - 2,8781x + 4,4617

12
10

R2 = 1

8
6
4
2
0
0

Viscosity (Cp)

10

Figure 4 Dispersion band in virtue of the PLS viscosity

Chemical effect
The viscosity is directly related to the mixer efficiency, which is a measure of the approach
equilibrium or the efficiency in the mass transfer that is produced in the mixer in relation to the
equilibrium that should occur between the PLS and the extractant. Mixer efficiency will depend on
the characteristics of both the organic and aqueous phases, as well as from the energy applied per
unit of volume. When the level of viscosity increases, the mix between the two phases is affected,

making it necessary to revise the level of agitation in order to achieve the optimal point for
metallurgical transfer provided the coalescence conditions are not affected.
The efficiency of approach to equilibrium or stage efficiency is a critical parameter to be checked in
a solvent extraction operation in order to assure an efficient operation from both the technical and
economic point of view. Calculations of stage efficiency may be made using different methods;
however, one of the most widely used methods is the McCabe-Thiele diagram. Other methods are
the extraction kinetics, (concentration change rate) and the direct method analyzing from the mixer.
The influence of viscosity on stage efficiency for set mixing conditions is shown below (H. Yez y
A. Soto, 2009):
Table 1 Stage efficiency in virtue of PLS viscosity

The impact of stage efficiency on SX recovery is shown below:

Figure 5 Influence of stage efficiency on the metallurgic result

Other variables affecting the stage efficiency are the following:

Temperature
Residence time and agitation speed
Impurities, degradation products and surfactants
Mix ratio
Extractant formulation

Previous studies evaluating extractant formulations (M. Soderstrom, T. Bednarski, E. Villegas,

2009), indicated a better stage efficiency and, consequently, a higher global recovery in SX would be
expected for formulated compounds based on nonyl salicyl aldoxime oximes.

Table 2 Stage efficiencies for different oximes

Therefore, SCM Franke decided to consider the change of formulation of the extractant used in the
plant and use the extractant Acorga M5910, an ester modified nonyl-aldoxime extractant.

RESULTS AND ANALYSIS

Comparative tests of extractants
SCM Franke performed standard tests to compare the extractant used in the plant
(ketoxime/aldoxime mix 75/25) with the extractant proposed by Cytec AcorgaM5910, mainly in
relation to its design characteristics:

Series parallel configurations with one stripping stage

Higher stage efficiency, given that the extractant AcorgaM5910, wich is a formulation of
ester modified aldoxime (C9) with higher kinetics, being less affected by the high viscosity
of the PLS.
Lower requirement of maximum load for the same production of Cu, therefore, less
consumption of extractant.
High Cu/Fe selectivity.

Information provided to SMC Franke by Cytec indicated that the proposed extractant has already
shown its benefits in plants such as El Abra, Minera Catemu, Toquepala, etc.
These tests were performed to show the metallurgical benefits in case of a possible change of
extractant in the plant. Additionally, its physical performance was also analyzed.
The characterizations of the PLS used in this evaluation (real plant) are shown below:
Table 3 PLS SCM Franke Characterization
Density
Viscosity
Superficial T.
Cu
pH
Fe
Mn
NO3ClAl
Mg
ENH

g/cc
cP
d/cm
g/L
g/L
g/L
g/L
g/L
g/L
g/L
mV

1,32
10,16
71,50
3,21
1,45
25,12
6,66
2,24
5,80
18,30
27,90
378

The comparison made between both extractants showed a better metallurgical response of Acorga
M5910 over the extractant used in plant, basically due to its higher stage efficiency (2% higher),
which resulted in 2 percentage points higher recovery in SX, a 40% higher Cu/Fe selectivity and
phase separation times slightly higher for the M5910 in organic continuity but lower in an aqueous
continuity. It is important to mention that all the variables were considered on an equivalent basis
for the evaluation (maximum load, O/A ratio, spent electrolyte, etc.)

Phase disengagement time

150

Time (sec.)

C.O.

C.A.

100

50

Fresh plant extractant

Acorga M5910

Figure 6 Phase disengagement time for evaluated extractants

Figure 7 Results of Fe co-extraction Isotherms for evaluated extractants

Table 4 Metallurgic result, extractant evaluation

Formulation
Ketox/Aldox 70/30 Mix
Acorga M5910

/v
Extractant
11,2
11,2

EAE
Ext.
79
81

EAE
Str.
95
95

% Rec.
E1-E2
72,3
75,1

% Rec.
E3P
89,1
90,0

% Rec.
SX
79,7
81,7

Production
Ton/year
19.011
19.483

Due to the results obtained in the tests it was decided to change the formulation of extractant from
a Ketoxime/Aldoxime mix 75/25 to an ester-modified aldoxime (Acorga M5910).

Results in plant
The main variable influencing the metallurgical response of the SX process is the viscosity of the
PLS. It is important tom mention that as of 2012 works started to be carried out in the leach area in
order to keep the viscosity under control. To this end, changes were made in the processes: discard
of raffinate solutions, dilution of the process liquor (water addition), and changes in the acid
distribution between the agglomerate and leach and mix of minerals fed into the plant.
It can be seen that the viscosity levels are extremely high, considering that the average viscosity in
the industry reaches values of 3,2 cP. This high viscosity depends on the concentration of Al and
Mg in the leach solutions.

Figure 8 Viscosity of PLS feeding the SX Plant

The high PLS viscosity produce a high consumption of reagent, thus increasing the operational
costs of the plant. After, a change in the formulation of extractant was made via monthly make up
in order to require a lower maximum load of extractant for the operating conditions of the plant
(PLS, spent electrolyte and configuration of SX plant).
The change in the formulation of the extractant was made as of December, 2012. Below is shown
the behavior of the stage efficiency of the solvent extraction plant for the 3 extract stages of the plant
as the first variable to be optimized as a consequence of the formulation change.

Change of extractant

Figure 9 Behavior of stage efficiency in relation to the PLS viscosity

In the graphic above, we can visualize the increase of the stage efficiency by 4 percentage points in
the average for the 3 extraction stages. It is important to mention that the increasing effect of this
variable starts to be visualized after eight months from the beginning of the addition of the new
formulation, equivalent to an 60% change in the inventory. The graph also considers the period of
lower temperatures due to winter season.
A better stage efficiency helped to increase the global recovery in SX by near 8%. Below is shown
the change of inventory to the new extractant in relation to the global recovery in SX.

Figure 10 Evolution of extraction efficiency in relation to the M59120 fraction in Plant

Other variables affecting the global recovery of SX were adjusted to new conditions of Cu in the
PLS, a work activities carried out in conjunction by Operations and Metallurgy of SCM Franke and
Cytecs technical staff. Among the adjusted variables it is important to mention an increase in the
acid concentration in the spent electrolyte in order to improve the stripping performance of the new
extractant, an increase in the levels of mixers, and adjustment in the copper-maximum load balance
and the optimization of the agitation speed.
Another objective for the SX plant was to reduce the consumption of extractant relative to
production. The results were the following:

Figure 11 Annual extractant consumption rate

The lower consumption evidenced by the new formulation of the extractant is mainly due to the
lower requirement of extractant for a higher production of fine Cu in relation to the year 2012.
Finally, the Cu/Fe selectivity, despite being a secondary problem in relation to viscosity, is of vital
importance due to the high concentration of Fe in the PLS, reaching values of 32 gpL. Below are
shown the plant results of selectivity and Fe concentration of the PLS for the 2012-2013 periods.

Figure 12 SCM Franke Cu/Fe selectivity

As shown, the Cu/Fe selectivity increased by 41% with the new formulation of the extractant,
despite the high level of Fe in PLS in the year 2013, increasing approximately 20 gpL. This result
allowed maintaining a Fe concentration lower than 1 gpL in the electrolyte, reducing the bleed of
electrolyte and minimizing the costs by losses of reagents in EW.

CONCLUSIONS

The change in the formulation of extractant allowed the mitigation of the reduction of the stage
efficiency and recovery in SX for the period of high viscosity (>10 cP).

Whenever there is a 60% of presence of Acorga M5910 in the organic phase in plant, there is
an increase in the stage efficiency and SX recovery.

The extractant consumption rate decreased by 25%, mainly due to a higher production as the
result of the increase in recovery in SX and lower requirement of extractant.

The Cu/Fe selectivity increased by 40%, despite a significant increase of Fe in PLS.

Finally, we came to the conclusion that the objectives proposed by the metallurgic and
operational areas for the solvent extraction area were achieved.

REFERENCES
Yez H. Soto A. Study of the impact of viscosity in PLS over solvent extraction, Seminario Cytec (2009),
Antofagasta, Chile.
Soderstrom M., Bednarski T., Villegas E. Eficiencia de aproximacin al equilibrio en plantas de extraccin por
solventes de cobre, Hydrocopper (2011), Via del Mar, Chile.

Recovery of the copper content of a mine

drainage by solvent extraction using a liquid
membrane emulsion process

Fernando Valenzuela,
Claudio Araneda, Carlos
Basualto and Jaime
Sapag-Hagar
Universidad de Chile

In this work, the treatment of an acidic mine drainage to extract its copper
content by a surfactant liquid membrane process in a stirred tank was
made. Cu(II) was removed using LIX-860 N-IC as extractants dissolved in
aviation kerosene. The effects of various factors, such as pH and metal
content in the feed solution, carrier concentration and surfactant content
in the membrane, acidity of the stripping solution, and stirring speed to
prepare the primary and the double emulsion, were reviewed. It was found
that the stirred tank designed was very efficient, achieving intimate contact
between the three participating phases. Copper extraction was favored in a
feed solution at pH 2.14, 3% w/w LIX-860 N-IC, and 100 g/L H2SO4 solution
as a stripping agent, although it would be more adequate to use a higher
content of acid especially if the concentrations of metallic ions in the AMD
increase. A 2% w/w content of Span 80 was appropriate to stabilize the
emulsions. As reported, the oximic carrier extractant transported
selectively copper, achieving a substantial degree of enrichment from feed
to strip liquor acceptor solutions. In the batch extractor, the best extraction
behavior was obtained when stirring speeds of 1,000 min-1 was used to
prepare the primary emulsion, and 350 min-1 to form the double emulsion.
The experimental data was analyzed in terms of a metal transport
mechanism based on a facilitated transport model, which takes into
account the chemical reaction between the metal and the oximic extractant
and the diffusion processes in the aqueous phase close to the external
interface of the membrane.
The results indicate that with adequate control and adjustment of these
experimental conditions the extraction extent of metal through the
membrane as well as their degree of enrichment in the stripping liquor can
be favored.

INTRODUCTION
The chemistry of the oxidation of sulfides, including pyrites, and the generation of ferrous ions and
subsequently ferric ions, is very complex, and this complexity has considerably inhibited the design
of effective treatment options. Although several chemical processes contribute to the formation of
acidic drainage, pyrite oxidation is by far the greatest contributor. In a previous paper we reviewed
the main chemical equations that govern the chemistry of acid formation in mines (Valenzuela et al.,
2005-b), however the following basic equation can be proposed for this process:
2FeS2 (s) + 7O2 (g) + 2H2O (l) 2Fe2+ (aq) + 4SO42- (aq) + 4H+ (aq)

(1)

On the other hand, the metal ion concentration in acidic mine drainage (AMD) is commonly low,
mainly due to dilution by natural surface and groundwater. Over the last few years some
remediation programs have been implemented to develop and apply new technologies to prevent
and control acidic drainage, aimed at evaluating the acidity of the drainage and removing or
recovering many valuable metals dissolved in it.
This work deals with the application of surfactant or emulsion liquid membrane (ELM) technology
to remove and recover copper from a sample of an acidic mine drainage collected close to a copper
mine located in the central part of Chile. The ELM method is in the pilot-scale testing stage, but
practical research activities have expanded quickly in recent years. The use of liquid membranes in
metal uptake is becoming the most promising approach to separation processes; Kumbasar and
Tutkun, 2004; Chakraborty, 2005). We have previously studied the application of solid supported
and surfactant liquid membranes in extractive metallurgy as well as in the decontamination
treatment of liquid industrial residues, and it is possible to conclude that the liquid membrane
methodology is particularly appropriate for the treatment of dilute solutions, offering many
advantages over currently used techniques. In order to approach an emulsion liquid membrane
process as an interesting industrial-application process for metal recovery or removal, it has been
necessary to study the behaviour of the ELM, that is, the kinetics mechanism of metal transfer from
the feed solution that contains the metal ions to the metal-acceptor solution making the strip liquor.
Many models for facilitated transport have been presented up to date.
The contribution of this work can be centered on two aspects: a) the analysis and explanation of
experimental results achieved in the batch-stirred tank, similar to industrial extractors and feasible
to be used in practical applications by means of a metal facilitated transport model deduced in a
former study in which we used a stirred kinetics transfer cell without double emulsion formation.
b) Additionally, in this work we have effectively succeeded in applying the use of emulsion liquid
membrane methodology to an actual sample of an acidic mine drainage from a copper mine. An
efficient extraction of copper with a surfactant liquid membrane was carried out in a batch-type
stirred tank in which a complete mix of the three phases aqueous feed, organic solvent and strip
liquor occurs.

METHODOLOGY
Materials
LIX-860 N-IC (5-nonylsalicylaldoxime) used as carrier extractant for copper was supplied by BASFChile, Span-80 (sorbitan monooleate) was used as a nonionic commercial surfactant. The organic
diluent was industrial-grade aviation kerosene. Samples of an acidic mine drainage from a copper
mine of Chile were used as the metal-containing feed solution which constitutes the external
aqueous phase of the liquid membrane system. It had a pH that varied in the range of 1 to 5, a
variable sulfate concentration between 4 and 5 g/L, and a density of 1.06 at 20 C. The received
sample presented an initial content of suspended solids of 615 mg/L which were separated by
settling and filtration previously to the treatment by the liquid membrane. Its average chemical
composition was the following: Cu(II) 350 mg/L, Zn(II) 250 mg/L, Fe 280 mg/L, and minor
quantities of other metals like Ca, Mg, Al, As, which did not affect the copper ion uptake. Sulfuric
acid solutions of various concentration ranges were used in the internal aqueous phase as metalacceptor stripping agents.

Procedure
The primary emulsion was prepared by vigorous mechanical agitation of the stripping aqueous
phase and the organic phase, what produced an emulsion with small droplets between 0.5 and 100
m in diameter at 1000 min-1. This organic solution consisted of the extractant and the surfactant
dissolved in kerosene. Then the double emulsion system was prepared using a varied range of
agitation. The experimental equipment was a batch-type stirred tank immersed in a constant
temperature water bath at 30 C. The vessel had an inner volume of 1.0 liter. The inner diameter
and the depth of the tank were 6.95 and 21.0 cm, respectively. The primary emulsion (50 mL) and
the feed solution (250 mL) were smoothly stirred with an IKA-Ruhrwerke RW20 stirrer equipped
with a turbine impeller with eight flat blades connected to a speed regulator. In this way the double
emulsion is formed with the globules of the primary emulsion encapsulated in the feed droplets.
After separation from the emulsion, the metals solutions were removed for analysis. Metal content
was measured by atomic absorption spectrophotometry on a Perkin Elmer 3110 apparatus. An
Oakton-500 pH meter was used for measuring the acidity of the acidic mine drainage.
Reproducibility to form the primary and double emulsion under the same conditions was almost
complete and with acceptable precision.

RESULTS AND DISCUSSION

Recovery of copper from acidic mine drainage
In order to recover the copper, the solution was treated in a first stage with an emulsion liquid
membrane contactor in the stirred tank in contact with an organic solvent formed by solutions of
LIX-860 N-IC as carrier extractant and Span-80 as surfactant in kerosene. The following variables
were investigated to study the selectivity of the extractant in the extraction of copper, compared to
zinc and iron: pH and metal content of the AMD, extractant concentration in the organic phase used
to prepare the liquid membrane, and acid strength of the sulfuric stripping liquor. Figure 1 shows
the influence of mine drainage pH on copper extraction. It is a plot of the disappearance of copper
from the raffinate in time, expressed as the [Cu]/[Cu o] ratio, where [Cu] and [Cuo] are the metal

concentration in the raffinate and its initial content in the AMD solution, respectively.
In Figure 1 it is seen that the removal of copper from the aqueous feed solution is fast and almost
quantitative above pH 2.0. The salicylaldoxime carrier is a strong copper extractant that selectively
extracts this metal from acid solutions like the AMD sample used in this study. At pH 2.14, over
98% of Cu(II) was extracted, while all the zinc and iron remain in the raffinate. The higher the initial
pH for extraction, the faster the removal of copper from the feed solution. However, although at pH
above 4.0 there is a large extent of copper extraction, there is also an important increase of iron coextraction.
1.0

1,0

pH 4.80
pH 3.84
pH 3.04
pH 2.14
pH 1.11
pH 0.58

0.6
0.4
0.2

0,8

[Cu]/[Cu0] [-]

[Cu]/[Cu0] [-]

0.8

0.6% LIX 860 N-IC

1.0% LIX 860 N-IC
2.0% LIX 860 N-IC
3.0% LIX 860 N-IC
4.0% LIX 860 N-IC
6.0% LIX 860 N-IC

0,6
0,4
0,2
0,0

0.0
0

10

20

30

40

50

60

10

15

20

25

30

Time [min]

Time [min]

Figure 1 Influence of feed solution pH

on copper extraction. Feed: 250 mg/L Cu;
Stripping solution: 2.04 mol/L H2SO4;
Organic solution: 3% w/w LIX-860 N-IC

Figure 2 Effect of carrier content in

organic solution on Cu(II) uptake
Feed: 350 mg/L Cu, initial pH 4.0
Stripping solution: 2.04 mol/L H2SO4;
and 2% w/w Span-80 in kerosene

The effect of a higher concentration of extractant in the organic phase is represented in Figure 2. It is
clear the permeation of metals through the liquid membrane is clearly dependent on the
concentration of the carrier in the organic solution. An increase of extractant content in the liquid
membrane brings about an increase in metal transport due to the existence of a larger amount of
carrier molecules at the available interface to complex the cupric ions. However, the carrier
extractant content in the organic phase cannot be increased indiscriminately because the increased
viscosity of the organic phase would cause an important decrease in the diffusion of the metal
complex in the liquid membrane.
Copper would exist over the pH range studied as the cation species Cu 2+, which would undergo an
extraction equilibrium as follows:
Cu2+ (aq) + 2 HX (org) = CuX2 (org) + 2 H+ (aq)

(2)

where HX denotes the chelating extractant LIX-860 N-IC.

As expected, metal extraction decreases as copper content in the AMD solution increases, which is
consistent with the chemical reaction that governs the Cu-LIX-860 N-IC complex. However,
although higher Cu(II) concentration in the acidic drainage would require the use of a higher
content of extractant in the membrane, in these experiments with surfactant liquid membranes the
contact of the three participant phases is very fast, causing a very rapid overall process kinetics,
thereby reducing the amount of carrier required for the same extraction compared with
conventional solvent extraction (SX) in mixer-settler reactors.

One of the most important variables for obtaining efficient metal transport in an ELM process is the
stripping agent concentration in the internal aqueous phase. The carrier extractant used in this
study, the oxime, correspond to acid compounds that form complexes with metals that must be
broken down with stronger acid solutions during the back-extraction step. It was seen that when a
1.02 mol/L solution is used, most of the copper disappears from the acid feed solution and is
transferred to the inner strip liquor. An even better result was achieved with 2.04 mol/L and 2.65
mol/L H2SO4 solution in a shorter time. However, it is not recommended to increase unnecessarily
the sulfuric acid concentration of the back-extraction solution due to the risk of degradation of
extractant through the loss of water molecules from the salicylaldoxime caused by the polyprotonation of the hydroxyl groups of this carrier.
As has been mentioned in previous papers, it is necessary to use an appropriate stirring speed to
prepare the primary emulsion, and a different one to produce the double emulsion. Many
experiments were carried out to check the importance of these variables. In both cases it was
established that a minimum of 1000 min-1 is required to form a stable primary emulsion, together
with the use of 2% w/w of the surfactant in the organic phase. The primary emulsion globule size
formed is adequate, and they turned out to be stable during the experiments, with rupture of the
emulsion being possible once the strip liquor was loaded and enriched with metal. The higher the
surfactant content in the liquid membrane, the more complicated is the final emulsion breakdown.
Therefore, it was not necessary to increase the amount of Span-80 in the organic phase above 2%
w/w because that amount was enough to form and stabilize the emulsion.
The surfactant does not have any effect on the transport of metals through the membrane. Under
the experimental condition of this study, it was not measured significant losses of carrier extractant
and the surfactant compound. It is the occurrence of emulsion difficulties like swelling, emulsion
breakdown, or phase entrainment what produces organic phase loss. However, in this study they
were controlled below 1% at least in the first minutes of the process, turning the loss of carrier and
surfactant practically negligible.
With respect to the effect of stirring speed of the double emulsion, extraction of Cu(II) from the feed
solution was enhanced as stirring was increased. Several experiments were carried out changing
this agitation rate in a range between 100 and 750 min-1. A minimum of 350 min-1 was required to
overcome the resistance of the external aqueous phase, which would be the controlling one in the
process, especially in those cases in which a rather low metal content exists in the AMD feed
solution. Increasing the stirring speed of the double emulsion brought about the generation of
smaller emulsion droplets, thereby increasing the interface area for the carrier/metal reaction, and
increasing the reaction sites. On the other hand, the use of excessive and unnecessary higher
stirring speeds causes coalescence of the globules, turning the primary emulsion unstable and
favoring the undesired transport of strip liquor towards the external continuous aqueous phase. In
all the subsequent experiments a stirring speed of 350 min-1 was selected.
In all these experiments the oximic extractant showed excellent selectivity for copper extraction
with respect to other metals present in AMD solution. It was found that it is completely feasible to
get a highly enriched copper strip liquor able to be driven to a final recovery step in a copper
sulfate crystallization process, because the same primary emulsion volume containing the same
strip liquor is contacted successively with different fresh feed fractions. For instance, after few
continuous contacts a content of 1.40 g/L was measured in the internal aqueous phase, which means
four times its content in the initial aqueous feed phase. No zinc load was detected around pH 2.0,
and most of the iron and other metals remained in the raffinate solution. Coextraction of iron with
the oxime tends to increase only if the pH of the acidic mine drainage is raised above 4.0. This

results make it feasible to separate Cu(II) from the acid mine drainage by the ELM method
adjusting the pH of the AMD around 2.0.

Analysis of experimental results

The experimental data obtained in this study are analyzed by a metal transport mechanism based
on a facilitated transport model, which takes account the chemical reaction between the metal and
the oximic extractant and the diffusion processes in the aqueous phase close to the external
interface of membrane. In the model, the double emulsion droplet system can be considered as a
large liquid membrane droplet where is assumed that the metal-carrier complex diffuses through
membrane towards the inner aqueous phase, while the extractant migrates in an opposite direction.
The water droplet inside the emulsion globule moves freely containing an average metal
composition. Furthermore, as the content of sulfuric acid in the strip liquor is kept much higher
than in feed solution, the reverse reaction of Eq. 2 would occur instantaneously at the stripping
interface.
The transfer rate of Cu(II) ion per unit surface area of emulsion droplet, JCu , can be expressed as

J Cu

Vaq ex d Cu

Aem
dt

[mol/cm2.sec]

(3)

where Vaq-ex is the volume of external aqueous phase, Aem is the total surface area of emulsion
droplets, and d[Cu]/dt is the change of metal concentration, [Cu ], with time. Aem can be estimated
from the total volume of emulsion droplet Vem and the surface area and volume of each drop, Sdrop
and Vdrop , as follows

Aem Vem

S drop
Vdrop

[cm2]

(4)

Then, the flux of metal can be rewritten as

J cu

Vaq ex D d Cu

Vem 6
dt

(5)

where D denotes the average diameter of emulsion droplet. Eq.(6) correspond to an expression for
copper flow, which was derived by assuming that the extraction rate would be controlled by the
chemical reaction to form copper-carrier chelating complex and by the diffusion step of metal in the
external aqueous phase film close to the interface with the emulsion. The transport mechanism also
assumes that there is a negligible mass transfer resistance in membrane based on that the apparent
membrane thickness would be too small compared with the external aqueous phase film neighbor
to liquid membrane.

J Cu

D k K 2 K
HX
3
1
2

HX Cu 1 / 2
0

1/ 2

H

(6)

DHX and k3, represents the diffusion coefficient of carrier extractant and the forward reaction rate
constant to form the CuX+aq species, respectively, as defined previously. As well, K1 and K2
correspond to the partition and the dissociation constant of oximic extractant. Suffix o denotes
initial concentration of each species.
Introducing this expression in Eq.(5) and rearranging it, the following equation is obtained:
1/ 2

2
d Cu Vem 6 DHX k 3 K1 K 2

dt
Vaq ex D
4

HX 0 Cu 1 / 2

1/ 2
H

(7)

Eq.(7) can be solved by integration by considering the initial conditions that at t = 0, [Cu]= [Cuo], as
follows:

Cu

Cu0

d Cu Q
t

HX 0 Cu 1 / 2 dt

1/ 2

(8)

where Q is defined by all the constant terms of Eq.(7) - whose values have been formerly estimated
- and is given by
1/ 2

2
Vem 6 DHX k 3 K1 K 2

Vaq ex D
4

(9)

In the calculation, the value of the partition constant, K1, used was the same as in the waterhydrocarbon system, that is, 7.210-4 [-] (Stary, 1964). The value of DHX , 1.2310-5 [cm2/s] was
estimated by Wilke-Chang, and the values of k3, 5.21011 [cm3/mol.s], and K2 ,1.110-12 [mol/cm3], used
were identical to those used in the previous paper (Valenzuela et al., 2005-b).
In most of cases, once equilibrium is reached, [H+] becomes practically constant around a value of
110-5 mol/cm3, then Eq.(8) becomes

Cu

1
Cu 0

1/ 2

HX 0 t
1/ 2
1/ 2
2 H Cu 0

(10)

In Figure 3 all the experimentally observed data are plotted according to Eq.(10). In the figure is
clearly appreciated that the experimental results are fairly well explained by the Eq.(10), specially in
the first minutes of process where difficulties of emulsion breakdown and swelling are controlled.

1.50

Exp. Observed [-]

1.0

1/2

0.6

0.4

1- ([Cu] / [Cu0])

1-([Cu] / [Cu0])

1/2

[-]

0.8

0.2

0.0
0

50

100

150

200

250
+ 1/2

300

350

400

450

500

1/2

[HX0]*t /(2[H ] [Cu0] ) [min]

Figure 3 Relation between 1- ([Cu]/[Cuo])1/2 and

([HXo]t) / (2 [H+]1/2 [Cuo]1/2) according to Eq.(10)

1.35
1.20
1.05
0.90
0.75
0.60
0.45
0.30
0.15
0.00
0.00 0.15 0.30 0.45 0.60 0.75 0.90 1.05 1.20 1.35 1.50
1/2

1-([Cu] /[ Cu0])

Calculated [-]

Figure 4 Comparison between

experimental and calculated values of
1- ([Cu]/[Cuo])1/2

The value of Q obtained from the slope of Figure 3 was 0.021, quite similar to those calculated using
the estimated diameter of the emulsion droplets of 0.5 m and the values of different constants
used in our previous work or reported in the literature. Figure 4 compares the experimentally
observed values of 1- ([Cu]/[Cuo])1/2 and the calculated according to Eq. 10. Figure 4 shows a quite
good agreement between both kind of values, validating the model of mechanism proposed to
explain the extraction of copper from the AMD sample by using the ELM in the batch-stirred tank.
However, the values calculated from the model tend to deviate from those observed
experimentally, especially at longer process times or a very low pH was used in the feed solution,
and when higher extractant content was used to prepare the primary emulsion, probably because
under these conditions, beyond the occurrence of emulsion difficulties, it is necessary to consider
other theoretical assumptions to explain the permeation of metal across the membrane. For
instance, likely under these conditions the mass transfer resistance in the ELM would becomes
significant or the process only be controlled by the diffusion or the chemical reaction. This point is a
matter of study by now and should be dealt in a further report.

CONCLUSIONS
The treatment of an acidic mine drainage by an emulsion liquid membrane process in a stirred tank
were studied with the purpose of optimizing the practical extraction and recovery of its copper
content. LIX-860 N-IC was used as carrier and Span 80 as the surfactant. It was found that the
stirred tank designed was very efficient, achieving intimate contact between the three participating
phases. Copper extraction was favoured in a feed solution at pH 2.14, 3% w/w LIX-860 N-IC, and
1.02 mol/L H2SO4 solution as stripping agent, although it would be more adequate to use a higher
content of acid specially if the concentrations of metallic ions in the AMD increase. A 2% w/w
content of Span 80 was appropriate to stabilize the emulsions. As reported, the oximic carrier
extractant transported selectively copper, achieving a substantial degree of enrichment from feed to
strip liquor acceptor solutions. In the batch extractor, the best extraction behavior was obtained
when stirring speeds of 1000 min-1 was used to prepare the primary emulsion, and 350 min-1 to form
the double emulsion.

The experimental data were analyzed in terms of a metal transport mechanism based on a
facilitated transport model, which takes account the chemical reaction between the metal and the
oximic extractant and the diffusion processes in the aqueous phase close to the external interface of
membrane. The experimental metal extraction results were approximately well explained by the
model, particularly in the early stage of the process when emulsion breakdown and swelling
difficulties are controlled.

ACKNOWLEDGEMENTS
The financial support of this work under FONDECYT Project 1040567 is gratefully acknowledged.
The authors also wish to thank Basf Chile Ltd. for providing the LIX 860 N-IC extractant.

REFERENCES
Valenzuela, F., Cabrera, J., Basualto, C. and Sapag-Hagar, J., (2005-b), Kinetics of copper removal from acidic
mine drainage by a liquid emulsion membrane, Miner. Eng., 18: 1224-1232.
Kumbasar, R., Tutkun, O., (2004), Separation and concentration of gallium from acidic leach solutions
containing various metal ions by emulsion type of liquid membranes using TOPO as mobile carrier,
Hydrometallurgy,75: 111-121.
Chakraborty, M., Murthy, Z., Bhattacharya, C., Datta, S., (2005), Process intensification: Extraction of
Chromium(VI) by emulsion liquid membrane, Sep. Sci. Technol., 40 : 2353-2364.
Stary, J., 1964, The Solvent Extraction of Metal Chelates, Pergamon Press, Oxford.
Valenzuela, F., Auspont, J., Basualto, C., Tapia, C., and Sapag, J., (2005-a), Use of a surfactant liquid membrane
contactor for zinc uptake from an acid aqueous effluent, Chem. Eng. Res. Des., 83: 247-255.

Improvements in the operation of SX

plants of Anglo American Copper

Juan Nez, Jessica Cueto

and Jos Fernndez
Mantos Blancos Division,
Anglo American, Chile
Cristian Salgado
Mantoverde Division,
Anglo American, Chile

Anglo American has used the extractant M 5640 in all the Divisions, which
is an aldoxime modified with TXIB formulation, except in Los Bronces,
where an aldoxime modified with tridecanol (TDA) formulation is used. It is
a well-known fact that the aldoxime degrades while the modifier does not.
With time the aldoxime/modifier ratio decreases significantly depending
on the operating conditions of each plant. In order to evaluate the effect of

Marco Torres
El Soldado Division,
Anglo American, Chile

this increasing modifier content in the organic of the plant, especially in

Gabriel Zarate
Anglo American, Chile

aldoxime without modifier does not affect the chemical performance of the

the organic losses, a complete metallurgical test work program was started
at laboratory and pilot plant scale. Results showed that the addition of
organic phase, but there is a significant decrease in the viscosity and
density of the organic phase. Based on the results obtained in the
metallurgical program, it was decided to run plant scale tests at both SX
plants, Mantos Blancos and Mantoverde. In these tests, instead of adding
the usual formulation of aldoxime modified with TXIB to maintain the
organic inventory, CuPRO MEX 3406 which is an aldoxime without
modifier is going to be added until the original aldoxime/modifier ratio is
reached. The plant tests started in April 2013 at Mantos Blancos and in June
2013 at Mantoverde and will last 18 months. The results obtained at the
laboratory and pilot plant scale, together with the preliminary plant test
results, are discussed in this paper.

INTRODUCTION
Anglo American has used the extractant M 5640 in all the Divisions, which is an aldoxime modified
with TXIB formulation, except in Los Bronces where an aldoxime modified with tridecanol
formulation is used. It is a well-known fact that the extractant degrades while the modifier does not
and as a consequence the aldoxime/modifier ratio decreases significantly through time, depending
on the operating conditions of each plant.
In order to evaluate the effect of this increasing modifier content in the organic of the plant,
especially in the organic losses, a complete metallurgical test work program was started at
laboratory and pilot plant scale. For this work, SNF FloMin reagents were used and the first step
was the validation of the SNF FloMin extractants as compared to the Cytec extractants. The
reagents CuPRO MEX 3302, equivalent to M 5640, and CuPRO MEX 3406 which is an
unmodified aldoxime were used.
Tests were run in parallel using the plant organic solution, from Mantos Blancos (MB) and
Mantoverde (MV), mixed in different proportions with both fresh extractants and then with the
unmodified aldoxime.
Based on the results obtained in the metallurgical program, it was decided to run plant scale tests at
both SX plants, Mantos Blancos and Mantoverde. In these tests, instead of adding the current
modified aldoxime to maintain the organic inventory, aldoxime without modifier is going to be
added until the original aldoxime/modifier ratio is reached. The plant tests started on April and
June 2013 respectively and will last 18 months.
The results obtained at the laboratory and pilot plant scale, together with the preliminary plant test
results, are discussed below.

EXPERIMENTAL PROCEDURES
Laboratory tests
Standard quality control tests were performed on 10 v/v % samples of each reagent. These
procedures are described in the SNF FloMin, Inc. Standard Quality Control Test Procedures [SNF
FloMin, Inc., 2013], and are very similar to those QC test procedures developed by other SX reagent
suppliers and commonly accepted by the copper industry. These tests consisted of extraction and
stripping equilibrium points, phase separation times in both aqueous and organic continuous
modes, copper/iron selectivity relative to the extraction equilibrium point, and kinetics of both
copper extraction and stripping. Additionally, physical properties were measured including the
densities, viscosities, interfacial tensions and surface tensions of the reagents against normal
standard aqueous solutions.

Pilot plant tests

The pilot plant programs at both the Mantoverde and the Mantos Blancos Divisions were carried
out in two solvent extraction circuits in series configurations, consisting of two stages of extraction,
two stages of stripping and one stage of washing (2ES x 2S x 1W). A similar program was
conducted at the El Soldado Division. Continuous flowrates were 100 cc/min PLS flow. The
Mantoverde and Mantos Blancos circuits are illustrated in Figure 1.

Figure 1 Circuit Configuration at MV and MB

Commercial tests
The commercial tests were started on April 2013 in Mantos Blancos and on June 2013 in
Mantoverde by adding CuPRO MEX 3406, a non-modified aldoxime supplied by SNF Flomin, as
the organic make up. Both SX plants have the same configuration, as it is shown in Figure 1.

RESULTS
Lab tests
The results of the chemical and physical performance of CuPRO MEX 3302 and M 5640 were
discussed elsewhere [M. Torres B., et al, 2013]. Among other results, a lower loading capacity of
both plant organics was found, indicating clearly that the plant organics have lost the original
loading characteristics as found in the fresh reagent. This could be partly due to the fact that the
extractant was over-modified by the selective degradation of the aldoxime, keeping in mind that
the TXIB modifier used is very stable and builds in concentration with time.
This over modification is the main reason why in both plants, Mantoverde and Mantos Blancos, it
was decided to start the addition of unmodified aldoxime (CuPRO MEX 3406) to their plant
organic. Over time, this will dilute the TXIB content of the plant organic and will improve physical
properties of the organic phase such us viscosity and density. The target was to achieve the
theoretical Aldoxime/TXIB content of the fresh CPM 3302/M 5640 reagent.
It has to be mentioned that the PLS of both plants is quite different, as it is shown in TABLE 1. As a
result, the extractant concentration is also different: 19.7% in Mantoverde and 14.0 % for Mantos
Blancos. In both plants, the diluent is Escaid 110.

Table 1 PLS Assay

Item
Cu, gpl
pH

MV

MB

8.1

3.2

2.1

1.6

11.3

2.0

Fe+3, gpl

2.9

0.7

Mn, gpl

3.2

1.3

Al, gpl

17.7

6.7

Mg, gpl

16.6

12.5

Cl-, gpl

2.0

53.1

SO4--, gpl

203

94.0

Fetotal, gpl

Pilot plant tests

The chemical performance of adding unmodified aldoxime to the plant organic, in terms of copper
extraction, copper unit transfer and Cu/Fe selectivity, is similar when compared to adding the
standard extractant.
However, there is a significant reduction in physical parameters like density, viscosity and phase
separation times as it is shown in TABLE 2 for Mantoverde and in TABLE 3 for Mantos Blancos.
These reductions are expected to decrease organic entrainment and therefore organic consumption.
Table 2 Physical Parameter Results at Mantoverde
Test

Aldoxime/Modifier Ratio

PST Extraction (sec)

PST Stripping (sec)

AC

OC

AC

Density Viscosity
(g/cc)

(cp)

Plant organic (PO)

1.42

131

129

98

97

0.848

4.71

90%PO+10% Aldoxime

1.50

110

99

86

89

0.846

4.58

80%PO+20% Aldoxime

1.67

101

97

74

76

0.843

4.45

Note:

OC

PST is phase separation time; OC is organic continuous and AC is aqueous continuous

Table 3 Physical Parameter Results at Mantos Blancos

Test
Plant organic (PO)
PO+Aldoxime
Note:

Aldoxime/Modifier Ratio

PST Extr (sec)

PST Strip (sec)

Density Viscosity

OC

AC

OC

AC

(g/cc)

(cp)

1.25

110

60

87

50

0,826

3.42

1.67

94

52

78

47

0,823

3.20

PST is phase separation time; OC is organic continuous and AC is aqueous continuous

Based on these encouraging results, it was decided to go to commercial tests which were on-going
at the time of preparation of this paper.

Commercial tests
The evolution of the aldoxime/modifier ratio together with the modifier (TXIB) concentration, after
six months, is shown in TABLE 4 for Mantoverde.

Table 4 Organic Parameters at Mantoverde

Date

Aldoxime/Modifier Ratio

TXIB (gpl)

15-05-2013

1.42

65.8

01-07-2013

1.44

61.5

05-09-2013

1.49

60.5

28-10-2013

1.52

58.1

21-11-2013

1.56

55.7

02-12-2013

1.61

54.0

As expected, the strategy of adding unmodified aldoxime has allowed a decrease in the reagent
consumption by almost 20%, as shown in TABLE 5. It has also allowed a decrease in the impurities
concentration in the electrolyte by reducing the aqueous entrainment as shown in TABLE 6.
Table 5 Reagent Consumption at Mantoverde
Year

Extr Cons
(kg/t)

2011

2.57

2012

2.07

2013 (Jan-May)

1.90

2013 (Jun-Nov)

1.55

Table 6 Electrolyte Impurity Concentration at Mantoverde

Jun-Nov

Fe

Mn

Cl-

(gpl)

(ppm)

(ppm)

2011

1.6

32

27

2012

1.2

24

25

2013

1.1

20

22

While doing the plant test, it was decided to evaluate at lab scale the optimum aldoxime/modifier
ratio for Mantoverde operating conditions using the net copper transfer to EW as a parameter.
Results, shown in Figure 2, indicated that the optimum ratio is around 1.6. By early December 2013,
this ratio has been achieved and therefore it was decided to stop adding the unmodified aldoxime
CuPRO MEXTM 3406 and start adding the extractant CuPRO MEXTM 3302.

200

N
e
t 190
C
u
T
r
a 180
n
s
f
e
r
t
o

170

E
W

(
)

t 160
p
d

150

0,8

1,0

1,2

1,4
1,6
Aldoxime/Modifier Ratio

1,8

2,0

2,2

Figure 2 Net Copper Transfer to EW and Aldoxime/Modifier Ratio in Mantoverde

In the case of Mantos Blancos, the evolution of the aldoxime/modifier ratio together with the
modifier (TXIB) concentration, after six months, is shown in Table 7. The lower aldoxime/modifier
ratio and the higher concentration of TXIB in the last sample were produced by an increase in the
extractant concentration due to the modification of the SX plant configuration from a series to a
parallel circuit.
Likewise in Mantoverde, it was also decided to evaluate at lab scale the optimum
aldoxime/modifier ratio for Mantos Blancos operating conditions which resulted to be around 1.8 as
shown in Figure 3. By the end of October, this ratio has decreased to 1.69 due to the reasons
mentioned above. Therefore, it was decided to add unmodified aldoxime until reaching a ratio of
1.8, which at the time of writing this paper has not been achieved.
Table 7 Organic Parameters at Mantos Blancos
Date

Aldoxime/Modifier Ratio

TXIB (gpl)

07-03-2013

1.25

47.1

23-05-2013

1.43

44.3

08-07-2013

1.60

39.8

12-09-2013

1.87

33.6

28-10-2013

1.69

44.7

100,0

95,0
N
e
t
C
u
T
r
a
n
s
f
e
r
t
o
E
W

90,0

85,0

80,0

75,0

70,0

65,0

(
60,0

t
p
d

55,0

50,0

1,0

1,2

1,4

1,6
Aldoxime/Modifier Ratio

1,8

2,0

2,2

Figure 3 Net Copper Transfer to EW and Aldoxime/Modifier Ratio in Mantos Blancos

CONCLUSIONS
AngloAmerican has used the extractant M 5640 in all the Divisions, except in Los Bronces, which is
a blend of an aldoxime and a modifier. As the extractant degrades while the modifier does not, the
aldoxime/modifier ratio has decreased significantly depending on the operating conditions of each
plant.
The effect of this increasing modifier content in the organic of Mantoverde and Mantos Blancos was
evaluated through a complete metallurgical test work program at laboratory and pilot plant scale
carried out by SNF Flomin. Lab and pilot plant results showed that the addition of aldoxime
without modifier does not affect the chemical performance of the extractant, but there was a
significant decrease in the viscosity and density of the organic phase.
Therefore, plant scale tests were run at both SX plants, Mantos Blancos and Mantoverde, adding the
reagent CuPRO-MEXTM 3406 which is an aldoxime without modifier until the original
aldoxime/modifier ratio was reached.
After six months, at Mantoverde the reagent consumption has been decreased by 20% and the
concentration of iron, manganese and chloride in the electrolyte has been significantly reduced.
In the case of Mantos Blancos, similar results are expected although they have not reached yet the
original aldoxime/modifier ratio due to a change in the SX plant configuration.

ACKNOWLEDGEMENTS
The authors wish to thank the Technical Vice-Presidency of Anglo American Chile for permission
to publish this paper. They also want to thank Mauro Torres, Technical Manager at SNF Flomin

South America, for organizing all the test work and to Brantley Sudderth, Principal Metallurgical
Consultant SNF Flomin Inc., for the useful discussions held throughout the development of this
work.

REFERENCES
SNF FloMin, Inc., Standard Quality Control Test Procedures, 2013.
M. Torres B., M. Torres M., C. Salgado, H. Bravo, J. Fernandez, B. Sudderth, CuPRO-MEX Extractants as
Equivalent Alternatives for Copper Recovery Copper 2013-Cobre 2013, IIMCh, 2013, pp. 1-12.

Extraction of Cu(II) ions through a

process coupling an ionic liquid
and supercritical CO

Rossana Seplveda and

Jonathan Castillo
Universidad de
Atacama, Chile
Julio Romero
Universidad de Santiago
de Chile
Jos Snchez
Institut Europen des
Membranes, Universit de
Montpellier II, France

Conventional treatments for removal of metals ions from aqueous

solutions require large amounts of organic solvents, which involve several
problems due to its high level of toxicity for human life and the
environment. As a result, research efforts have been focused on finding
new eco-friendly alternatives.
Ionic liquids (ILs) are organic salts that are liquid at room temperature,
with melting points lower than 373K, the main advantage in the use of
these compounds is their unique physicochemical properties that allow
them to be considered green solvents. ILs show a practically negligible
vapour pressure, which minimizes solvent losses related to its volatility.
Thus, ILs could be coupled to other eco-friendly operations.
This research is focused on the optimization of working conditions in
the removal process of Cu(II) from aqueous solutions, which are complexed
with 1,1,1-trifluoro-2,4-pentanedione (TFA), initially contained in the IL
1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide. Metal
complexes formed in the ionic liquid phase were re-extracted by using
supercritical CO2.
In a first step, the process was studied to assess the feasibility of the
liquid/liquid extraction, an extraction efficiency of 77.6% was obtained when
the initial concentration of Cu(II) in the aqueous solution was 250mg/L,
initial concentration of TFA in IL was 310mol/m3, the volumetric IL/aqueous
phase ratio was 1:1, initial pH of aqueous solution was equal to 5.3.
To optimize the liquid/liquid extraction, a buffer solution was used to
maintain the pH at 5.3, and the best extractions conditions were identified
at: Five min of stirring time, volumetric IL/aqueous phase ratio 1:7 and an
initial TFA concentration in IL of 60 mol/m3, obtaining a removal efficiency
of 83.54%. Preliminary assays, shows that the best condition for thre
stripping of metal complexes formed in the IL phase with supercritical
carbon dioxide at 18 MPa and 40C was 1 h of IL/CO2 contact time.

INTRODUCTION
Conventional treatments used to remove metals ions from aqueous solutions require the addition of
large amounts of organic solvents, which can be highly polluting (Cox & Reinhardt, 2004; Cox,
2004). In this way, a vast number of researches deal the identification of alternative solvents, which
are environmentally friendly. Among those unconventional solvents, ionic liquids emerge as a
versatile option with very tunable properties because of the diversity of its chemical structure.
A large body of literature (Cox & Reinhardt, 2004; Cox, 2004; Keskin et al., 2007; Roth, 2009)
identifies the most common ionic liquids as green solvents in terms of different properties such as
its practically negligible vapour pressure.
The term Room Temperature Ionic Liquid (RTIL) describes organic salts, which are liquid at room
temperature, in contrast to the well-known high-temperature molten salts. RTILs have been
disclosed as organic salts with melting points below 373 K (Roth, 2009; Yang & Dionysiou, 2004).
The number of potential ILs can be greater than 1018 (Holbrey & Seddon, 1999) due to the large
number of combinations of anions, cations and substituents. This variety allows managing the
relationships between structure and properties in order to be tailored to meet characteristics
required for a specific application. Thus, RTILs can be drawn from a wide range of compounds,
either organic (Izk et al., 2008; Poole & Poole, 2010), or metal ions (Cocalia et al., 2006; Liao et al.,
2011). The low volatility of the ionic liquids is a key property that makes them good candidates for
use as green solvents. However, this characteristic also causes problems for the separation and
recovery of the product. In this way, the design of intensified processes combining ILs with other
green technologies such as supercritical fluids could overcome this disadvantage. On the other
hand due to their physical and chemical properties of ionic liquids could be in the future an
alternative to conventional organic solvents currently used in industry. A pure component is
considered a supercritical fluid if its temperature and its pressure are higher than the critical values
(Brunner, 1994). The most common compound used as a supercritical fluid is carbon dioxide.
Supercritical carbon dioxide (SC CO2) has been used to remove metal ions from aqueous solutions
by the formation of complex between Cu(II) and a complex agent, due to its highly solubility in
supercritical CO2, allowing high extraction efficiencies in the stripping process (Murphy & Erkey,
1997a; Murphy & Erkey, 1997b), in literature it can be found a series of studies focused on analysing
the properties of the complexes formed between copper(II) ion and extractants type -diketone
(Kidd, Sager & Watson, 1967; Sallam et al., 2002) or Cu(II)-benzoyl acetone complexes extracted by
SC CO2, (Liu, Wang & Li, 2001) other works report specific points on the extraction of metals, such
as the solubility of the complexes in supercritical carbon dioxide (Murphy & Erkey, 1997b ; Kidd,
Sager & Watson, 1967).
In this work, an imidazolium-based ionic liquids is used to extract Cu2+ from aqueous solutions,
using 1,1,1-trifluoro-2,4-pentanedione (TFA) as extractant agent. In this way, the extraction
performance and mechanisms could be established and subsequently the IL with the best extraction
capacity was used in regeneration steps with supercritical CO 2.

METHODOLOGY
Studied system
The system studied in this work considers in a first step the liquid-liquid extraction of Cu2+ from
aqueous solutions using an imidazolium-based ILs, which contained TFA as extractant agent. As a

second step, the removal of metal complexes formed in the IL phase was achieved by using
supercritical carbon dioxide as stripping phase. The main goal of this study is the identification of
the IL with highest potential to replace the organic solvent used in a typical extraction process.

Materials
The ionic liquid 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, [bmim][Tf 2N],
were supplied by Sigma-Aldrich with purities higher than 99% for synthesis grade.
The -diketone used as extractant was 1,1,1-trifluoro-2,4-pentanedione (TFA).
For the preparation of the aqueous solutions we used copper sulphate pentahydrate (CuSO 45H2O),
and sulphuric acid (H2SO4) and anhydrous sodium sulphate (Na2SO4). All these compounds were of
analytical grade. The solutions were prepared in distilled water.
The purity of CO2 used for the supercritical extraction tests was 99.9%.
For atomic absorption spectroscopy analysis of ionic liquid samples containing Cu-TFA complexes
were dissolved in analytical grade ethanol.

Liquid-liquid extraction experiments

For experiments we used Cu(II) ions aqueous solutions of a concentration between 250 and 500
ppm. Meanwhile, the concentration of TFA solutions in ILs used for extraction was of 310 mol m-3.
Experiments were carried out at 25 C ( 1 C), putting into contact under vigorous stirring equal
volumes of TFA/IL and copper aqueous solution (3mL + 3mL) during 45 min in order to reach the
equilibrium conditions. Moreover, NaSO4 was added to the aqueous solution in order to obtain 0.1
M of SO4-2 in solution and to improve the phase separation. After the extraction run, both phases
were separated by centrifugation, the final pH of the aqueous solution was measured with a pHmeter and the final concentration of Cu(II) in the aqueous phase and IL was analysed as explained a
above. In a second step, the IL phase loaded with Cu(II) and recovered from the extraction
experiments was introduced in a view cell to be contacted with supercritical carbon dioxide in
order to achieve the equilibrium conditions, regenerating the extraction capacity of the IL by means
of the removal of the formed complexes.
The extraction performance was characterized by means of the extraction percentage (%E) and the
extracted mass of copper for each extraction step (ME i). These parameters are defined by the
following equations:

Amount of extracted metal

%E=
100
Initial amount of the metal in the aqueous solution
n

ME i = Mass extracted in step i

(1)
(2)

i=1

Supercritical extraction of metal-complexes formed in IL

After the separation of phases contacted in liquid-liquid extraction, the IL phase loaded with Cu(II)
and recovered from the extraction experiments was introduced in 100 mL view cell to be contacted
with supercritical carbon dioxide at 18 MPa and 40 C during 4 hours in order to achieve
equilibrium conditions. Supercritical extraction tests to remove the copper complexes were done in
a device specially designed for this purpose where the view cell containing the IL phase was loaded
with CO2, which was pressurized with high pressure air-driven pump Autoclave Engineers G35.

The ionic liquid was removed from the cell after this processing and sent to analysis by AAS as an
ethanol-based sample. Additional experiments were done using the same volume of ionic liquid in
cycles of liquid-liquid extraction and supercritical regeneration of the IL phase in order to verify the
feasibility to reuse a constant volume of IL as solvent maintaining its extraction and regeneration
capacity.

RESULTS AND DISCUSSION

Effect of the concentration of TFA and copper on the extraction performance
The results obtained for extractions with [bmim][Tf2N] show a continuous increasing of the
extraction percentage when the concentration of TFA, increases from 0 to 310 mol m-3. Indeed,
[bmim][Tf2N] seems to be the solvent that generates the best extraction equilibrium condition and it
forms neutral complexes, which could be extracted from the IL phase by means the contact with
supercritical carbon dioxide.
Table 1 Effect of TFA concentration on the extraction percentage. Initial concentration of Cu(II) 250
mg L-1, phase rate 1:1 (3 mL of each solution), SO42- =0.1 M, initial pH=5.3 0.2
TFA concentration, mol/m3
0
Stoichiometric
100
310

Extraction percentage, %
82.51 5%
82.13 5%
78.45 5%
71.2 5%

An additional set of experiments was done in order to verify the effect of the initial concentration of
Cu(II) in the aqueous phase on the extraction performance. Figure 1 show the results of extraction
percentage and equilibrium concentration of copper, respectively, for two different values of initial
concentrations of Cu(II): 250 and 500 mg L-1.

Accumulative extraction capacity of the ILs

The copper load capacity of IL was assessed following a procedure similar to the used for liquidliquid extraction tests described previously, but with an accumulative extraction of Cu(II) ions in
the same volume of IL. For this purpose the a same IL was used consecutively for the extraction of
fresh aqueous solutions containing Cu(II) ions, the process was carried out four times. TABLE 2
reports these results in terms of the extraction percentage and concentration of copper in the IL for
the IL when the initial concentrations of Cu(II)(aq) and TFA(IL) were 250 mg L-1 and 310 mol m-3,
respectively. Every extraction step was done with the same volumetric aqueous/IL ratio (1:1).
The [bmim][Tf2N] presents a very interesting results after four successive extractions, this IL shows
the a low decrease in the extraction percentage indicating that the the gradient to reach the
equilibrium concentration between the aqueous and IL phase decreases but slowly. Moreover it is
the only one to be able to capture copper continuously even after the third extraction period.

-1

300
Cu(II) in the ionic liquid, mg L

Extraction percentage, %

100
90
80
70
60
50
40
30
20
10
0

250
500
-1
Cu(II) in the feed, mg L
TFA free

250
200
150
100
50
0

250

500
-1

Cu(II) in the feed, mg L

-3

TFA=310 mol m

TFA free

TFA=310 mol m

-3

Figure 1 Left: Extraction percentage of Cu(II); right: Equilibrium concentration, both as a function of the initial
concentrations of Cu(II)(aq) and TFA(IL). Phase rate 1:1 (3 mL of each solution), SO42- =0.1 M, initial pH=5.3 0.2

Table 2 Result for test with four loads. Initial concentration of Cu(II) and TFA, 250 mg L-1 and 310
mol m-3 respectively, phase rate 1:1 (3 mL of each solution), SO42- =0.1 M, initial pH=5.3 0.2
Charge number
1
2
3
4

Accumulative extraction, %
82.51 0.5
82.13 0.5
78.45 0.5
71.2 0.5

Mass concentration
217 0.5
432 0.5
619 0.5
749 0.5

Copper complex removal from IL with supercritical carbon dioxide and extraction
cycles
The second part of this study involves the removal of copper complexes formed in the IL phase
after the extraction using a stripping phase of dense CO2 at supercritical conditions. In this way, a
set of experiments was implemented to remove the copper complexes from samples of the IL
obtained after the liquid-liquid extraction tests. Specifically, samples of 3 mL of IL with an initial
concentration of TFA equal than 310 mol m-3 and already used for the liquid-liquid extraction of
copper (250 mg L-1) from aqueous solutions , were placed in a 100 mL high-pressure extraction cell
and contacted with carbon dioxide at 180 bar and 40 C during 40 min. Then, the cell was
depressurized and the IL phase was recovered and analysed by AAS as ethanol-based sample.
Table 2 reports the concentration of copper before and after recovery with supercritical carbon
dioxide for [bmim][Tf2N]. From these results, it is possible to verify the higher extraction capacity of
[bmim][Tf2N], since the concentration of copper measured in its sample was 212 mg L -1.
Nevertheless, this high receiving capacity of [bmim][Tf 2N] involves a lower capacity to remove the
copper complexes formed during the extraction by means of dense CO2. After a regeneration step
with supercritical CO2, the 37.26% of copper species in the IL phase were removed from
[bmim][Tf2N]. This result could be explained by a higher affinity between the metal complex and
[bmim][Tf2N].
Moreover, these last experiments were complemented with additional tests for the assessment of
the extraction capacity of the IL regenerated with SC CO2. Thus, the IL samples stripped with dense
CO2 in the view cell were recovered in order to conduct liquid-liquid extraction experiments with
IL sample: the frist one using the regenerated IL and the second one by adding TFA to obtain the

same original concentration (310 mol m-3). The addition of fresh TFA was done under the
assumption that all non-reacted TFA in the first extraction was removed from the IL during the
treatment with SC CO2. The comparative results of the copper extraction with and without TFA
(310 mol m-3) for first extraction step and using stripped ILs are reported in TABLE 3.
Table 3 Concentration values of copper in the IL before and after treatment with supercritical
carbon dioxide at 180 bar and 40 C, and removal percentage
Concentration
of
copper in the IL, mg/L

Concentration of copper in the IL

after stripping (SC CO2 ), mg/L

Removal percentage,
%

212.0

133.0

37.26

Table 4 Extraction percentage values obtained for the three ILs during the first and second L-L
extraction step. Between these extraction steps the IL phase was regenerated by contact with SC
CO2
Extraction % for the first L-L
extraction step (%) (fresh ILs)
TFA free
13.60

Extraction percentage in the second extraction step

with stripped ILs (%)

CTFA = 310 mol m-3

Without addition of fresh

TFA

Adding
TFA, CTFA= 310 mol m-3

87.60

61.94

82.23

The IL show a decreasing extraction percentage close to 26.5% when were regenerated with dense
CO2 and reused without addition of TFA.
After these experiments, cycles of liquid-liquid extraction and stripping with SC CO2 were done
using the same original volume of IL. Figure 2 shows the extraction cycles, which considered
sequences of L-L extraction coupled to supercritical regeneration steps of IL. L-L extractions were
done using fresh aqueous CuSO4 solutions with an initial concentration of Cu(II) of 250 mg L -1.
[bmim][Tf2N] was regenerated after every L-L extraction step by means of SC CO2 at 180 bar and 40
C. Fresh TFA was added to the recycled [bmim][Tf 2N] under the assumption that all non-reacted
TFA in the last extraction was removed from the IL during the treatment with SC CO2. Extraction
percentages reported in Figure 2 for extraction-regeneration cycles with the same original volume
of [bmim][Tf2N] remain practically constant after 6 cycles. Thus, this procedure could become an
interesting alternative for the extraction of different types of metal ions and hazardous species from
aqueous media.

Extraction percentage, %

100
90
80
70
60
50
40
30
20
10
0

1st

2nd

3rd

4th

5th

6th

Liquid-liquid extraction
Figure 2 Extraction percentage of Cu(II) for L-L extraction cycles using recycled [bmim][Tf2N] regenerated
with supercritical carbon dioxide, initial concentration of Cu(II) in the aqueous phase 250 mg L-1, initial
concentration of TFA in IL 310 mol m-3, phase volume ratio 1:1 (3 mL of each solution), SO42- = 0.1 M. Fresh
TFA was added to the IL after every regeneration step

CONCLUSIONS
In L-L extraction tests, [bmim][Tf2N] showed high extraction capacity in terms of accumulative
extraction percentage. Additional extraction tests allow verifying that the effect of the anionic
medium of the aqueous solution (sulphate or chloride) on the extraction percentage was not
significant.
Moreover, IL used in L-L extraction of Cu(II) ions were regenerated using supercritical carbon
dioxide, which stripped the copper complexes formed in the IL phase. In these experiments,
[bmim][Tf2N] is still the best solvent for this application, since it was able to recover its extraction
capacity after several extraction cycles. The optional addition of fresh -diketone (TFA) in the
regenerated IL showed that the SC CO2 removes the most of the non-reacted TFA content with the
metal complex. The addition of fresh TFA in the IL after every supercritical regeneration step
allows recovering the most of its original extraction capacity.
Future studies must be conducted in order to optimize the operating conditions and reduce the
concentration of extractant and the required volume of IL. These researches will be focused on the
chemical aspects related to extraction mechanism verified in this work.

ACKNOWLEDGMENTS
The authors acknowledge financial support from the Universidad de Santiago de Chile and also
appreciate the support provided by the Universidad de Atacama. Chilean researchers R. Seplveda,
J. Castillo and J. Romero were supported by Project ICM-P10-003-F, Interdisciplinary Center of
Ionic Liquids, CILIS, granted by Fondo de Innovacin para la Competitividad, del Ministerio de
Economa, Fomento y Turismo, Chile.

REFERENCES
Brunner, G. (1994) Gas extraction: an introduction to fundamentals of supercritical fluids and the application
to separation processes, Berln.
Cocalia, V.A., Holbrey, J.D., Gutowski, K.E., Bridges, N.J. & Rogers, R.D. (2006) Separations of metal ions using
ionic liquids: The challenges of multiple mechanisms. Tsinghua Science and Technology 11: 188-608.
Cox, M. & Reinhardt, H. (2004) The use of solvent extraction in the recovery of waste, in Solvent extraction
principles and practice, ed by Rydberg J, Cox, M, Musikas, C. and Choppin, G. Marcel Dekker, Inc.,
New York.
Cox. M. (2004) Solvent extraction in hydrometallurgy, in Solvent extraction principles and practice, ed by
Rydberg J, Cox, M., Musikas, C. and Choppin, G. Marcel Dekker, Inc., New York.
Holbrey, J.D. & Seddon, K.R. (1999) Ionic Liquid. Clean Products and Processes 1: 223-236.
Izk, P., Ruth, W., Frei, Z., Dyson, P. & Kragl, U. (2008) Selective removal of acetone and butan-1-ol from water
with supported ionic liquidpolydimethylsiloxane membrane by pervaporation. Chem Eng J 139: 318321.
Keskin, S., Kayrak-Talay, D., Akman, U. & Hortasu, O. (2007) A review of ionic liquids towards supercritical
fluid applications. J Supercrit Fluids 43: 150-180.
Kidd, M.R., Sager, R.S. & Watson, W.H. (1967) Properties of some copper(II) and zinc(II) N-oxide and bdiketone complexes. Inorg Chem 6: 946-951.
Liao, C-Y., Peng, C-Y., Wang, H-C., Kang, H-Y. & Wang, H.P. (2011) Extraction of arsenic from a soil in the
blackfoot disease endemic area with ionic liquids. Nucl Instrum Methods Phys Res, Sect A 652: 925927.
Liu, J., Wang, W. & Li, G. (2001) A new strategy for supercritical fluid extraction of copper ions. Talanta 53:
1149-1154.
Murphy, J.M. & Erkey, C. (1997a) Copper(II) removal from aqueous solutions by chelation in supercritical
carbon dioxide using fluorinated b-diketones. Ind Eng Chem Res 36: 5371-5376.
Murphy, J.M. & Erkey, C. (1997b) Thermodynamics of extraction of copper(II) from aqueous solutions by
chelation in supercritical carbon dioxide Environ Sci Technol 31: 1674-1679.
Poole, C.F. & Poole, S.K. (2010) Extraction of organic compounds with room temperature ionic liquids. J
Chromatogr 1217: 2268-2286.
Roth, M. (2009) Partitioning behaviour of organic compounds between ionic liquids and supercritical fluids. J
Chromatogr 1216: 1861-1880.
Sallam, S.A., Orabi, A.S., El-Shetary, B.A. & Lentz, A. (2002) Copper, nickel and cobalt complexes of Schiffbases derived from b-diketones. Transition Met Chem 27: 447-453.
Yang, Q. & Dionysiou, DD. (2004) Photolytic degradation of chlorinated phenols in room temperature ionic
liquids. J Photochem Photobiol A: Chem 165: 229-240.

Lanthanide metals extraction with

extractants attached to surface of
magnetite nanoparticles

Jos Gaete,
Lorena Molina,
Fernando Valenzuela
and Carlos Basualto
Universidad de Chile

The industrial consumption of rare earth elements (REE) has been growing
year after year while the production does not grow at the same speed. One
of the largest difficulties for production is the fact that REE are highly
dispersed at relatively low concentrations in mineral deposits and because
present very similar chemical characteristics and behaviors. On the other
hand, it is well known that one of the more versatile and cheaper tools
available for separation of these metallic ions when are contained in
aqueous phase is the solvent extraction (SX) technique. The family of
organophosphorous acid extractants CYANEX272, D2EHPA, and CYANEX
301 has been used to achieve the separation of the four REE present in
bastnaesite ore: Lanthanum, cerium, praseodymium, and neodymium.
However, it is well known that the solvent extraction technique
presents potential risks of toxicity, flammable hazards, and pollution to
the environment. Furthermore, SX would present process efficiency losses
as crud formation and/or solvent entrainment into the aqueous phase.
The use of the mentioned extractants attached onto magnetic
nanoparticles surface could offer a powerful tool to overcome these
shortcomings by means of an external magnet, once the extractionadsorption of the metals has been achieved.
To appropriately apply this methodology, the particles must present the
smallest possible size, ideally in the nanoparticles range, thus ensuring the
maximum interfacial area. Magnetic nanoparticles must present an
individual magnetic-domain to be considered as superparamagnetic, a
property which assures their suitable behavior for their required use as
magnetic adsorbent.
The results showed that the extractants attached to the surface of
magnetic nanoparticles exhibited different extraction degrees for each of
the studied metals. In all cases, the load capacities by the functionalized
magnetite nanoparticles with CYANEX 272, D2EHPA and CYANEX 301
extractants were obtained.

INTRODUCTION
Rare earth materials REE, such as metals and/or their compounds, are characteristic because a
variety of physical properties, which means possibilities for wider applications as functional
materials. The difficulty for preparing rare earth metals of high purity has been a barrier to their
use, especially in the high-technology field. Since the existence of trace impurities affecting
mechanical, electrical, chemical and other properties of the materials, the obtainment of high purity
materials is also important to clarify their intrinsic properties and to find new functional materials.
Wider use of REE will require both lower-cost methods of refining and purification and the
development of many new applications that have been held back by the purity issue (Alguacil F.,
1997).
The solvent extraction technique has been applied practically to the mutual separation of rare earth
metals. The family of the organic extractants from phosphoric acid called by the generic names of
phosphoric, phosphonic, phosphinic and dithiophosphinic acids it seems to be the most promising
in achieving the extraction of rare earth metal ions from aqueous solutions by means of solvent
extraction (Abdeltawab A., 2002; Weiwei W., 2006; Yin J., 2007; Li H., 2006; Morais C. A., 2004;
Kondo K., 2002). Among of many extractants, PC-88A (2-ethylhexyl phosphonic acid mono-2ethylhexyl ester) is one of the most common, possessing high separation efficiency and low
solubility in water.
Although solvent extraction has been used successfully, there are some drawbacks. One of them is
the lower ability for enrichment. Usually, the volumes ratio between the solvent phase and the
aqueous phase is huge and, owing to its own solubility in the feed solution, the organic solvent, the
extractant and the extracted species as organic complex may then also be lost into the aqueous
phase (Nishihama S., 2002). Moreover, the difficulty of phase separation in multiple extraction and
back-extraction stages is another disadvantage.
The adsorption of metal ions on synthetized adsorbent bed materials such as activated charcoal,
alumina, silica, alumino-silica gels, and a broad line of metal oxides have been extensively used, as
well as other materials from natural origin, among of them bentonite, chitosan, cellulose, and other
vegetable raw materials. These materials are very cheap and plentiful. However, the main
hinderance for its massive use is the reduced surface area which produces an inefficient process
because the material spending is too high with a reduced load of adsorbed species. A good
performance requires the smaller particles as possible.
It is known, the smaller particles the larger surfaces area are obtained. Thus, the nanoparticles
(NPs), with an increasing relative surface area, are the smallest attainable. Thus, chemically active
nanoparticles are often between 110 nm, because in this range the proportion of surface atoms is
especially large. For example, in an NP consisting of 13 atoms, 92% of the atoms are surface atoms
(whereas in an NP of 55 atoms, only 76% of those are surface atoms; 147 (63%); 309 (52%); 1415
(35%)). Similarly, in an NP with a diameter of 1 nm, around 50% of the atoms are surface atoms,
with the percentage of surface atoms exponentially decreasing with an increasing diameter (10 nm
diameter10% surface atoms). These particles often exhibit physical and chemical properties that
are between those of their smallest components (such as metal atoms or metal oxide units) and
those of bulk materials, where only a few percent of the atoms lie on the surface (Rao C. N. R., 2004;
Somorjai G. A., 2010; Wilcoxon J. P., 2006; MacKenzie J. D., 2007; Kwon S. G., 2008; Thomas J. M.,
2006).
The large relative surface areas of NPs are comparable to those of porous materials, but without
problems of mass transport, because the active sites are highly accessible. However, the use of NPs

generates an additional difficulty because the fine particulated material separation is costly and
expensive.
In recent years, functionalized magnetic nanoparticles (MNPs), i.e. magnetic nanoparticles with
specific molecules attached to their surface, have gained an increasing interest because of their
potential applications. The functionalization method assigns to magnetic nanoparticles specific
properties of interaction or reaction while the magnetic attributes allow for easy separation by
using an external magnet. Examples of their uses include cell separation (Ying J., 2007), magnetic
resonance imaging (Lee J., 2006), drug delivery systems (Tobias N., 2005), proteins separation (Gu
H., 2006), and cancer treatments through hyperthermia (Akira I., 2003).
Among of magnetic materials the magnetite is the most widely used. It is nontoxic, cheap and is a
strongly ferromagnetic compound whose Curie point is 860 K, at which the effective magnetic
moment is 3.910-23 J/T (4.2 B) (Kirk Othmer Encyclopedia of Chemical Technology, 1981), but
changes its behavior as a superparamagnetic material when individual domains are obtained in the
size order of nanoparticles (NPs).
In this work, the organophosphorous acid extractants D2EHPA, LIX 272, and CYANEX 301 were
attached to the surface of superparamagnetic magnetite nanoparticles (MNPs) and used for
extracting lanthanide metal ions from a feed solution. The syntheses and characterization of MNPs
and its extraction behavior for lanthanum, cerium, praseodymium, and neodymium were carried
out.

EXPERIMENTAL
Reagents
In the experiments, the following pure analytical chemicals were used: FeCl24H2O and FeCl36H2O,
hydrochloric acid fuming 37%, Emsure Merck; ammonia solution 25% GR Merck; Acetone GR,
Merck; n-Hexane absolute, Emparta Merck; and Oleic Acid Ph Eu Sigma-Aldrich.

Procedure
For the obtainment of nanoparticles, the precipitation method of iron salts was used. In a 500 mL
beaker 50 mL of 0.2 M of Fe3+ and 50 mL of a 0.1 M of Fe2+ solutions were mixed, maintaining a 2:1
ratio between them. The pH was maintained at 2, while the solutions were degasified and the
working conditions remained under inert nitrogen atmosphere. Subsequently, the temperature was
increased to 80C and the pH was reached up to 9 by the addition of NH4OH solution. This mixture
was allowed to stand for a period of 30 min, the obtained precipitate was decanted using a magnet
and washed with degassed distilled water. This process is repeated three times being the obtained
nanoparticles dried and stored.
To stabilize the obtained nanoparticles, 0.65 mL of oleic acid are added, and for the subsequent
functionalization, 50 mL of a 0.03 M of an organophosphorous acid extractant emulsion was added
and placed in ultrasonic bath for 1h. The emulsions were previously prepared by saponification
adding NaOH for 12 h over the respective solutions of D2EHPA, CYANEX 272, or CYANEX 301
extractant dissolved in kerosene.
The obtained functionalized magnetic nanoparticles as product were retained by means of a magnet
and washed with several portions of pH 4 water. Finally, this product remained in water at the
same pH value and kept for later use.

For the extraction experiments, 50 mL of a 0.25 mM of lanthanide metal ion, as feed solutions, were
contacted with 7 mL of the functionalized nanoparticles concentrate during 1 hour at a stirring
speed of 600 rpm. Then the loaded nanoparticles were separated by magnetic decantation and the
raffinate was kept to determine the amount of remaining metal not extracted. The determination of
lanthanide metal concentration was made by the Arsenazo photometric method in the visible range
at 660 nm wavelength.

RESULTS AND DISCUSSION

Particle size by DLS
Bibliographical sources have indicated that the best magnetic behavior for magnetic particles is
obtained when the smallest particles are synthesized, ideally in the range of nanoparticles
(Magnetic Materials. Fundamentals and Device Applications, 2003). Thereby, the particles that are
constituted by monodomains non-coercive are superparamagnetic. This is one of the main
objectives to be achieved in this study.
Once the magnetite nanoparticles are coated with the extractants and after used them in the
extraction step, they should respond optimally when an external magnetic field is applied by
separating them from the depleted aqueous phase (raffinate).
This work presents a study of magnetite precipitation which was performed at different pH
conditions being the results analyzed by Dynamic Light Scattering (DLS) technique, whereby the
particle size and its distribution are obtained.
It was determined the optimum conditions for the magnetite precipitation, being determined that a
final pH 9 is required, yielding an average nanoparticle size of 134 nm. The graph of size and size
distribution for the uncovered magnetite nanoparticles (MNPs) and the stabilized (covered) with
oleic acid are shown in Figure 1 (a) and (b), respectively. In the same way, those magnetite
nanoparticles functionalized with CYANEX 272, D2EHPA, and CYANEX 301 are shown in
Figure 2 (a), Figure 2 (b), and Figure 3, respectively.
(a)

(b)

Figure 1 Size and size distribution of magnetite nanoparticles (MNPs):

(a) uncovered and (b) stabilized with oleic acid

(a)

(b)

Figure 2 Size and size distribution of functionalized magnetite nanoparticles (MNPs) with:
(a) CYANEX 272 and (b) D2EHPA extractants

Figure 3 Size and size distribution of functionalized magnetite nanoparticles (MNPs) with
CYANEX 301 extractant

The results of the average size for the obtained different nanoparticles are presented in TABLE 1.
Table 1 Average size for uncovered, stabilized, and functionalized magnetite nanoparticles
(MNPs)
Magnetite nanoparticles (MNPs)
uncovered
stabilized with oleic acid
functionalized with CYANEX 272
functionalized with D2EHPA
functionalized with CYANEX 301

Size
134.0
22.0
10.9
10.5
10.0

nm
nm
nm
nm
nm

The results indicate that the largest diameter is obtained for the uncovered nanoparticles of
magnetite. This outcome seems to be contradictory, because the smaller size is expected for this
kind of particle, but given the aqueous dispersion medium, it is highly probable that the particles
are in high degree of agglomeration. Also, it should be expected that the functionalized particles
were larger than those stabilized. However, in this case it seems a contraction is produced when the
extractant is added to the stabilized nanoparticles with oleic acid.

Magnetism by VSM
Vibrating Sample Magnetometer (VSM) measurements are necessary for determining the magnetic
capacity of the different synthesized MNPs.
In the Figure 4 are shown the magnetic responses, in emu/g units, when an external magnetic field
(H) in Oe units is applied, for the uncovered and the covered with oleic acid nanoparticles of
magnetite. Besides, in the Figure 5 the same responses for the functionalized magnetic
nanoparticles are presented.

80

60

40

40

20

20

Ms (emu/g)

Ms (emu/g)

60

80

covered with OA MNPs

uncovered MNPs

0
20

CYANEX 272
CYANEX 301
D2EHPA

0
-20

40

-40

60

-60
-80

80
-10000-8000 -6000 -4000 -2000

2000 4000 6000 8000 10000

H (Oe)

Figure 4 Magnetic responses against an applied

external magnetic field, for the uncovered
and the
covered MNPs

-10000 -8000 -6000 -4000 -2000

2000 4000 6000 8000 10000

H (Oe)

Figure 5 Magnetic responses against an applied

external magnetic field, for the functionalized
MNPs with: CYANEX 272, D2EHPA, and
CYANEX 301

The obtained curves show that all nanoparticles are constituted by a single magnetic domain due to
no-significant coercivity was observed and therefore all of them are superparamagnetic. Moreover,
the higher the covering for the nanoparticles, the smaller the magnetic response against an external
magnetic field.
The saturation magnetization capacity is lowered from 72, 55, until 46-49 emu/g for uncovered NPs
stabilized with oleic acid to functionalized magnetite nanoparticles, respectively. The obtained
value for those functionalized it was enough for retaining the loaded MNPs when the separation
step is conducted.

Extraction behavior
The extraction behavior for the four lanthanide elements with the different functionalized
magnetite nanoparticles was studied.
The results of the achieved load capacities by the functionalized nanoparticles are presented in
Figure 6.

16
14

mgLn/gNP

12
10
8

Cyanex 272

D2EHPA

Cyanex 301

2
0
La3+

Ce3+

Pr3+

Nd3+

Extracted lanthanide metal ion

Figure 6 Load capacities by the functionalized magnetite nanoparticles (MNPs)
with CYANEX 272, D2EHPA, CYANEX 301

In the Figure 6, the obtained load capacities by the functionalized magnetite nanoparticles with the
extractant CYANEX 272 was practically double than those functionalized with D2EHPA and
CYANEX 301 extractants are observed. It is also noted that for a given functionalized nanoparticle
no significant differences on selectivity for a given lanthanide element are obtained.
The best result was obtained for the functionalized magnetite nanoparticles with CYANEX 272
extractant for lanthanum metal ion. In this case, a 14.4 mgLa/gMNP of load capacity was achieved.

CONCLUSIONS
It was possible to obtain magnetite nanoparticles functionalized with the three proposed
extractants: CYANEX 272, D2EHPA, and CYANEX 301, with very similar sizes around 10 nm,
approximately.
The saturation magnetization capacity is lowered from 72, 55, until 46-49 emu/g for those magnetite
nanoparticles, uncovered, stabilized with oleic acid, and functionalized, respectively. The value
obtained for functionalized magnetite nanoparticles resulted to be enough good for retaining the
load MNPs, when the separation step is performing.
All the functionalized magnetite nanoparticles were able to extract the four studied lanthanide
metal ions. However, the best extraction result was obtained for magnetite nanoparticles
functionalized with CYANEX 272 extractant, especially for lanthanum metal ion, achieving a
loading capacity of 14.4 mgLa/gMNP.
The obtained results of this study have shown that the extraction of lanthanide metal ions is
possible with these nanoparticles under a quick and easy way.

REFERENCES
Abdeltawab A. A., Nii S., Kawaizumi F., Takahashi K., (2002), Separation and Purification Technology, 26, 265
272.
Akira I., Kouji T., Kazuyoshi K., Masashige S., Hiroyuki H., Kazuhiko M., Toshiaki S., Takeshi K., (2003),
Cancer Science, 94, 3, 308313.
Alguacil F. J., Rodriguez F. (1997), Revista de Metalurgia (Madrid) 33, 3, 187-196.
Gu H., K. Xu, C. Xu, Xu B. (2006), Chemical Communications, 9, 941949.
Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Third Ed, New York, USA, 1981. Vol 13,
pag. 775.
Kondo K., Kamio E., (2002), Desalination, 144, 1-3, 249-254.
Kwon S. G., T. Hyeon, (2008), Accounts of Chemical Research, 41, 16961709.
Lee J., Jun Y., Yeon S., Shin J., (2006), Angewandte Chemie, International Edition 45, 48, 8160-8162.
Li H., Guo F., Zhang Z., D. Z. Wang Li., (2006), Journal of Alloys and Compounds, 408-412, 995-998.
MacKenzie J. D., Bescher E. P., (2007), Accounts of Chemical Research, 40, 810818.
Magnetic Materials. Fundamentals and Device Applications, N. Spaldin. Cambridge University Press, Cambridge
CB2 2RU, UK. 1st Edition 2003, p. 42
Morais C. A., Ciminelli V. S. T., (2004), Hydrometallurgy, 73, 3-4, 237-244.
Nishihama S., Sakaguchi N., Hirai T., Komasawa I., (2002) Hydrometallurgy, 64, 35-42.
Rao C. N. R., Muller A., Cheetham A. K., The Chemistry of Nanomaterials: Synthesis and Applications, Wiley-VCH,
Weinheim, 2004.
Somorjai G. A., Li Y., Introduction to Surface Chemistry and Catalysis, Wiley-VCH, Weinheim, 2010.
Thomas J. M., Raja R., (2006), Topics in Catalysis, 40, 317.
Tobias N., Bernhard S., Heinrich H., Margarete H., Brigitte V.R., (2005), Journal of Magnetism and Magnetic
Materials, 293, 483496.
Weiwei W., Xianghan W., Shulan M., Hongfei L., Deqian L., (2006), Journal of Rare Earths, 24, 685-689.
Wilcoxon J. P., Abrams B. L., (2006), Chemical Society Reviews, 35, 11621194.
Yin J., Hu B., He M., Jiang Z., (2007), Analytica Chimica Acta, 594, 61-68.
Ying J., Lee R.M., Williams P.S., Jeffrey J.C., Sherif S.F., Brian B., Maciej Z., (2007), Biotechnology &
Bioengineering, 96, 11391154.

chapter 9
Acid rock drainage: Prediction,
prevention and treatment

Synthetic biology in extraction,

separation, and remediation

Christopher DaCunha
and Patrick Nee
Universal Bio Mining,
llc Tucson, usa

Advances in genetic engineering and synthetic biology offer new

approaches to the challenges of the mining industry. Traditional chemistry
is limited to small molecules, but during the four billion years of life on
earth many large molecules, such as proteins and enzymes, have evolved to
interact with the metals and the chemical reactions that take place in the
world around us. Synthetic biology allows us to harness these capabilities
for industrial purposes. The chemistry of copper extraction is being
modified by changing the genetic material in mining organisms. This will
allow for the processing of uneconomic ore types. In a separate effort,
microorganisms are being developed to deliver substrates to the chemical
reactions in metal extraction. In remediation, systems based on
microorganisms have been designed and tested for water treatment.
Biopolymers can be combined with industrial size exclusion filtration
systems to remove trace levels of Pb, Cd, As, Cu, Cr and Zn. Engineered
microorganisms can express metal specific binding proteins, sequestering
metals on their cell surface and allowing for separation by filtration or
centrifugation, a technology which is applicable to both metal separation
processes and water remediation. In addition to discussing the technical
developments in the field, the US regulatory environment for the use of
engineered organisms will be discussed.

INTRODUCTION
New genetic technologies and rapidly decreasing prices for genetic sequencing and DNA synthesis
are transforming the life sciences. Prices for sequencing are dropping at a rate which matches that
of the semiconductor industrys history, as shown in Figure X. The ability to create, or print,
DNA of lengths as long at tens of thousands of base pairs (soon to increase to millions of base pairs)
has accelerated the development of new genetic constructs. The development of new engineering
approaches in which genetic parts are compatible with each other has resulted in a rapid decrease
in the cost of development of new technologies, once priced at a billion US$ or more. As a result,
applications of biotechnology to industrial processes are growing rapidly, some examples of which
include biofuel production and the synthesis of high-value chemicals.
The metals mining industry currently utilizes biological processes in extraction and remediation.
Sulphide ores of copper, gold, uranium, and nickel are processed industrially with bioleaching, and
five additional metals have been extracted with biological leaching in the laboratory. Additionally,
bioremediation is utilized in the treatment of selenium and other pollutants. Phytoremediation is
used to reclaim former mine sites, planting grasses and shrubs on tailings and heaps. There is every
reason to believe the tools of synthetic biology will be able to expand the use of biological solutions
in the mining industry, having significant impact on the efficiency and ecological impact of mining.
This paper will review the growth and prospects of synthetic biology, will discuss some sample
applications in mining, and will attempt to predict the future direction of synthetic biology and
mining.

Synthetic biology
Synthetic biology is evolving extremely rapidly and there has been some difficulty in developing a
widely accepted definition. It is often defined as the design and construction of biological devices
and systems for useful purposes.
Advances in foundational technologies are driving the speed of evolution of synthetic biology.
DNA sequencing and synthesis are becoming dramatically faster and cheaper each year, with speed
and prices decreasing at rates approaching those of integrated circuits. As a greater number of
organisms are fully sequenced and their genes are identified, libraries of natural biosynthesized
molecules and their functions expand. Software modelling of genetic modifications and in-silica
simulations of designs are even enabling the design of synthetic proteins and enzymes.
Together these advances are accelerating a design-test-modify process of development similar to
the engineering cycles used in the software development industry.
As a result, the cost of developing biological applications has dropped from the billions of dollars
spend to develop the first genetically modified products, to projects measured in millions of dollars.
Some projects are even being crowd-funded by potential customers and expectations are to
deliver their synthetic biology applications having spent significantly less than US$1 million.

Relevance for metals mining

Biotechnology in metals mining has potential in improving extraction, separation, and remediation
processes because of the history of living systems.
Ever since the first microorganisms developed on earth, living systems have been co-evolving with
the atomic elements, minerals, and chemical processes of metals mining.

One leading hypothesis is that the first living organisms developed in environments similar to
some metals mining processes. The first cell membrane-like structures are hypothesized to have
formed from iron and sulphur (Martin, Russell; (Dec. 2002)). Subsequently, the first life forms are
believed to have been chemolithoautotrophs, or organisms that obtain their energy from chemical
processes rather than from consuming organic substances formed by other living organisms. These
organisms are similar to the chemolithoautotroph extremophile organisms used industrially in
bioleaching of copper today.
After these earliest life forms, organisms were required to develop methods to utilize, regulate, and
protect themselves from metals and the chemical conditions they create. As a result, the growing
libraries of sequenced and identified genes used by synthetic biologists include many which have
various interactions with these compounds and elements.

An example of synthetic biology in mining applications

In addition to extraction technologies, the authors have developed a biological metal binding
technology which can be used for remediation and for metal separation.
The system is based on modular biological parts. This modularity is a characteristic of synthetic
biology systems.
The first modular component is a metalloprotein. Metalloproteins are proteins which bind or
incorporate metals. These proteins are numerous and continue to be identified. Some bind
specifically to certain metals, and others are more promiscuous, which have binding affinity for
larger classes of metals, such as di or trivalent species.. With advancing understanding of protein
function, it may be possible to modify one of these proteins to create a synthetic biosynthized
protein which improves on the characteristics for a specific application.
The second modular component of the system is an outer membrane anchor. These anchors are
proteins which embed themselves in the outer membrane of a microorganism. Some of the
currently used anchors are naturally occurring, and some have had their native functions truncated
and modified for various applications. Anchors can have secondary functions as well, such are
cleavage sites when exposed to particular conditions, such as pH.
Combining the metalloprotein with the cell membrane anchor results in a hybrid molecule, a fusion
protein, which, when synthesized by the organism, migrates to and incorporates itself into the cell
membrane, with the attached metalloprotein on the external surface of the organism.
A third modular genetic component of the system is a regulatory gene to control the expression of
the components above. Differing levels of expression result in lesser or greater amounts of the
metalloprotein and anchor being synthesized. By varying this regulatory gene, the system can be
optimized for the specific application.
These self-replicating organisms can then be introduced into a liquid bearing the target metals and
the metals will be bound to the surface of the organisms until saturated.
The organisms can be allowed to gravity settle and the supernatant, with the bound metals
removed, can be decanted. Alternatively, anchors and attached metals could be removed in bulk by
utilizing secondary function characteristics.
Alternatively, because of the relatively large size of microorganisms (1-10 micrometers) relative to
metal ions, size exclusion micro-filtration systems can be used to separate the saturated organisms
from the metal-cleansed filtrate. These filters operate at much lower pressure differentials and

therefore require much lower power inputs to operate relative to reverse osmosis and nanofiltration systems.

Regulatory environment
Genetically modified organisms (GMOs) have been widely distributed and are currently approved
for used in 59 countries and are cultivated locally in 29 countries.i The majority of these genetically
modified organisms are crops which enter the human food chain.
Regarding industrial applications such as the ones addressed in this paper, in the United States,
because these applications are not food, drug, nor pesticide, they are regulated by the US
Environmental Protection Agencys Toxic Substance Control Act. This act regulates the use of new
chemical compounds in US commerce, and it requires a 90-day review period for commercial
release and a 60-day review period for R&D release. If the application is not found to present an
unreasonable risk to health or the environment, or a substantial or significant exposure, then the
genetically modified organism can be released for use in commerce. ii
In the United States, 47 microorganisms have been approved for release, primarily organisms used
in agriculture, in spite of the fact there is a possibility that they enter the food chain.iii Industrially
important organisms have a clear regulatory pathway which has been utilized by industrial
chemical manufacturers many times.
Regulatory frameworks in other countries focus on genetically modified organisms in agriculture,
and there appears to be no regulatory structure in place for the use of GMOs in industrial settings
in some of the major mining countries, such as Chile. Where there is a regulatory body, some
surmise that approval for genetically modified organisms in the food chain is a good indicator of
the likelihood of approval for use of an industrial genetically modified organism.
International treaties such as the Cartagena Protocol address plant safety, pesticide safety, food and
allergen safety, and notification. For the types of industrial applications addressed in this paper,
these treaties impose specific labelling, pre-notification of shipping, and pre-notification of use of
GMOs.

CONCLUSIONS
This example above shows the modular design of biological systems allows for combinations of
natural, modified, or completely synthesized genes and proteins to address challenging problems
in novel ways.
Areas where synthetic biology is currently being developed include extraction of metal, selective
separation of metals and minerals in process streams, the remediation of toxins in effluent streams,
and even in dust suppression. These systems can result in improvements in economic efficiency by
increasing recovery or reducing costs, or may allow operations to continue by bringing operations
to within regulatory thresholds.
The rapidly decreasing costs of biotechnology and its increasing predictability are enabling a new
set of tools to be brought to bear on problems which remain intractable with more traditional
technologies, and metals mining is an industry as ripe as others to benefit from this technological
revolution.

REFERENCES
Martin, Russell; (Dec. 2002) On the origin of cells:.., The Royal Society
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1693102/pdf/12594918.pdf

US Department of State, http://www.state.gov/e/eb/tpp/abt/biotech/, last accessed April 2, 2014.

Microbial Products of Biotechnology; Final Regulation Under the Toxic Substances Control Act ,
http://www.epa.gov/fedrgstr/EPA-TOX/1997/April/Day-11/t8669.htm, last accessed 4/2/2014.
iii TSCA Biotechnology Notifications, FY 1998 to Present, http://www.epa.gov/oppt/biotech/pubs/submiss.htm,
last accessed 4/2/2014
i

ii

Remediation of acid mine drainage

by microorganism

Pamela Gutirrez and

Mara Taboada
Universidad de
Antofagasta, Chile
Pamela Chvez
Aguamarina
Biominera, Chile

Acid mine drainage (AMD) is an environmental problem that mainly occurs

once mining operations have closed. These liquids are generated by the
exposure of sulfide minerals to oxygen and water present in the environment.
The aim of this project is to develop a bioremediation treatment of drains
generated by the mine through the use of the bacteria Bacillus pasteurii. This
process helps to reduce the concentration of heavy metals in AMD and also
increases the pH of the acidic water, allowing for the water to be recovered
and returned to productive mining processes or to the environment.
The work methodology consists of obtaining a culture medium that
allows for goodmicrobial growth and enzymatic expression. We work with
three growth media at room temperature and with different nutritional
characteristics, analitycal grade as a reference and industrial grade to get
an idea of large-scale application, allowing us to obtain a culture up to
90% cheaper.
Using 5% v/v of the culture medium with the microorganism and
synthetic AMD, it was possible to obtain a reduction of the Fe content by
38.8%, Al 90.1%, Mn 5.3% and Mg 7.4% and increase the pH to 7.3. This is
caused by trapping the cations in the crystalline structure of calcium
carbonate and increasing the pH.

INTRODUCTION
Technological progress and the current standard of life demand large quantities of raw materials in
diverse ecosystems from the environment. The mining process fully impacts the environment; it
pollutes the air, modifies geographical areas and decreases superficial water resources such as
groundwater. Generally, at the initiation of the mining process the oxidized ore is removed, to then
go through the sulfide ore, which, when it is exposed to the environmental conditions of water and
oxygen, begins to generate acid mine drainage, also known as acid rock drainage, a process which
can be accelerated by lixivian bacterias and results in the generation of acid water with a high
heavy metal content (Akcil and Koldas 2006). There are diverse methods to treat such water which
are classified in abiotic and biological. Within the abiotics system there is a passive system such as
the one that uses limestone in an anoxic environment and the active ones that use lime and
aeration. In the biological process, there is an active system that uses anaerobic bioreactors with
sulfate-reducing bacterias and as the passive systems, there are wetlands or marshes, permeable
barriers and bioreactors, (Hallberg and Johnson 2005).
Recent research has evaluated the incorporation of the calcium carbonate crystalline structure in
certain elements such as strontium (Fujita et al. 2004) (Mitchell and Ferris 2005), arsenic (Achal et al.
2012), and copper (Achal, Pan, and Zhang 2011). This incorporation has been studied using an
alkalophilic microorganism, which, when it hydrolyses the urea (nitrogen source), induces the
limestone precipitation, one of the most stable crystalline forms of calcium carbonate.
The precipitation of calcium carbonate is a reaction that is influenced by the following factors:
calcium concentration, inorganic calcium concentration, pH and the availability of nucleation sites
(Hammes and Verstraete 2002). The reactions of these processes are described in equation (1) and
(2).
(

) +2

(1)
(2)

The microbian hydrolysis of urea is approximately 1014 times faster than a non catalytic reaction
(Tobler et al. 2011), with the exposed antecedents and over the base that AMD remediation comes
from the ability of some microorganism to generate alkalinity and immobilize the metals, thereby,
reversing the AMD generation reactions (Johnson and Hallberg 2005). We studied the metals
abatement in the acid mine drainage using Bacillus pasteurii, considering a culture medium of
analytical and industrial quality with a dosification of 5% v/v of culture medium for the acid
drainage.

MATERIALS AND METHODS

Microorganism and growth conditions
B.Pasteurii (ATCC 11859) was used throughout the study with three different mediums for growth,
which used urea and diverse protein sources, TABLE 1. The experiments consider analytical and
industrial-grade reagents and distilled water. The urea was added after a filtration process because
the sterilization using an autoclave could decompose it. The analytical grade medium was sterilized
at 121C for 15 minutes and the industrial reagents were exposed to UV for 15 minutes before being

dissolvedin sterile distilled water. The cultures were grown at room temperature, with induced
aeration without pH adjustment. The cultures were inoculated at 1% v/v.
Table 1 Characteristics of media cultures
Growth Culture

Protein Quantity

Other

MC 1

High

MC 2

Medium

NaCl

MC 3

Low

NaCl

Analytical procedures
Culture medium
The changes in the system were monitored at different time intervals, measuring pH andmicrobial
growth through optical density of 600 nm. The ammonium concentration was measured by the
nessler assay. Absorbance readings were calibrated using NH4Cl as standard.

Removal of metals
For the removal of metals, synthetic AMD is mixed with 5% v/v of culture medium with the highest
enzyme expression and biomass. The AMD composition is shown in Table 2. For calculating
removal, the dilution effect of the mixing was considered. Samples were filtered (0.22um). Metal
concentration was measured using inductively coupled plasma atomic emission spectroscopy (ICPAES), Optima 2000 DV (PerkinElmer, Inc.).

RESULTS AND DISCUSSION

Growth and pH profile
The growth profiles were studied up to 86 hours in the analytical and industrial media, except MC
2 and 3 (industrial quality) which were measured for 113 h until they reached the stationary fase.
Bacillus pasteurii grew in all media and expressed enzymatic activity. Analytical media grew faster
than industrial media; this is influenced by the impurity and nutrient diffusion, Figure 1. Cultures
started with different pH, at 18 hours they had a pH around 9.0 ; which is the optimum for growth,
Figure 2 (Hammes and Verstraete 2002). This behavior is important because the bacteria present
tolerance and capacity for adapting to different kinds of medium cultures.
Table 2 Characteristics of synthetic AMD, composition in mg/L
Fe

Al

Mg

Mn

pH

475

64

25

90

2,0

Ammonia production
All cultures showed ammonium, which is a consequence of enzymatic activity (the control culture
did not show ammonium). The analytical culture presented ammonium first, while the industrial
grade ammonium concentration equaled 31 hours later. This condition is important because it is
possible obtain the same concentration of ammonium, but about 90% cheaper compared to the
analytical grade reagents.

Figure 1 Patterns of growth in the different media

Figure 2 Patterns of pH in the different media

Figure 3 Ammonia release in media culture

Heavy metal removal

For precipitation, culture media MC 2 was considered, because it has a high enzymatic activity and
more biomass than MC 3. This is important, because the cell wall has a negative charge, making
microorganisms ideal crystal nucleation sites (Hammes et al. 2003). When mixing the AMD with
MC 2, instantaneous precipitation is observed. With this mechanism it is possible to obtain a
reduction in Fe content by 38.8%, Al 90.1%, Mn 5.3% and Mg 7.4% and increased pH up to 7.3. The
decrease of heavy metals in the liquid phase is produced by two effects. The first effect is
incorporating cation in the calcium carbonate crystalline structure, and the second is an increase in
pH.

CONCLUSIONS
The microorganisms have shown the ability to adapt and grow in an industrial grade medium.
Moreover, they present the capacity to decrease heavy metal concentration and increase pH in acid
mine drainage by means of microbial metabolism for precipitated CaCO 3 present in Bacillus
pasteurii.

ACKNOWLEDGMENTS
The financial assistance for this study was received from Conicyt, project N 7812110015.

REFERENCES
Achal, Varenyam, Xiangliang Pan, Qinglong Fu, and Daoyong Zhang. 2012. Biomineralization Based
Remediation of As(III) Contaminated Soil by Sporosarcina Ginsengisoli. Journal of hazardous materials
201-202:17884. Retrieved May 19, 2012 (http://www.ncbi.nlm.nih.gov/pubmed/22154871).

Achal, Varenyam, Xiangliang Pan, and Daoyong Zhang. 2011. Remediation of Copper-Contaminated Soil by
Kocuria Flava CR1, Based on Microbially Induced Calcite Precipitation. Ecological Engineering
37(10):16015. Retrieved May 16, 2012
(http://linkinghub.elsevier.com/retrieve/pii/S0925857411001868).
Akcil, Ata, and Soner Koldas. 2006. Acid Mine Drainage (AMD): Causes, Treatment and Case Studies.
Journal of Cleaner Production 14(12-13):113945. Retrieved March 16, 2012
(http://linkinghub.elsevier.com/retrieve/pii/S0959652605000600).
Fujita, Yoshiko et al. 2004. Strontium Incorporation into Calcite Generated by Bacterial Ureolysis. Geochimica
et Cosmochimica Acta 68(15):326170. Retrieved March 8, 2014
(http://linkinghub.elsevier.com/retrieve/pii/S0016703704000341).
Hallberg, Kevin B., and D. Barrie Johnson. 2005. Biological Manganese Removal from Acid Mine Drainage in
Constructed Wetlands and Prototype Bioreactors. The Science of the total environment 338(1-2):11524.
Retrieved May 19, 2012 (http://www.ncbi.nlm.nih.gov/pubmed/15680632).
Hammes, Frederik, Nico Boon, Johan De Villiers, Steven Douglas Siciliano, and Willy Verstraete. 2003. StrainSpecific Ureolytic Microbial Calcium Carbonate Precipitation Strain-Specific Ureolytic Microbial
Calcium Carbonate Precipitation. 69(8).
Hammes, Frederik, and Willy Verstraete. 2002. Key Roles of pH and Calcium Metabolism in Microbial
Carbonate Precipitation. (Morita 1980):37.
Johnson, D. Barrie, and Kevin B. Hallberg. 2005. Acid Mine Drainage Remediation Options: A Review. The
Science of the total environment 338(1-2):314. Retrieved March 13, 2012
(http://www.ncbi.nlm.nih.gov/pubmed/15680622).
Mitchell, Andrew C., and F. Grant Ferris. 2005. The Coprecipitation of Sr into Calcite Precipitates Induced by
Bacterial Ureolysis in Artificial Groundwater: Temperature and Kinetic Dependence. Geochimica et
Cosmochimica Acta 69(17):41994210. Retrieved March 8, 2014
(http://linkinghub.elsevier.com/retrieve/pii/S0016703705002486).
Tobler, Dominique J. et al. 2011. Comparison of Rates of Ureolysis between Sporosarcina Pasteurii and an
Indigenous Groundwater Community under Conditions Required to Precipitate Large Volumes of
Calcite. Geochimica et Cosmochimica Acta 75(11):32903301. Retrieved March 11, 2013
(http://linkinghub.elsevier.com/retrieve/pii/S0016703711001803).

chapter 10
Water geochemistry

Solvent extraction for separation

of metals from hev Li-ion batteries
nitrate leach liquor

Sravya Kosaraju,
Christian Ekberg and
Stefan Allard
Chalmers University of
Technology, Sweden

Hybrid electric vehicles (HEVs) are gaining popularity as one of the more
economic choices for transportation. The Li-ion batteries with varying
chemistries are one of the favored batteries for these vehicles. The growing
numbers of HEVs will soon result in a large number of end-of-life batteries
in next decade. Currently there is little infrastructure to handle or recycle
end-of-life Li-ion batteries. The economic incentive for recycling will also be
soon lost as new chemistries are replacing the conventional cobalt
chemistries with other cheaper metals. However, recycling when
understood from a cumulative reasoning such as resource re-use and
reducing environmental load caused due to virgin metal extraction which
involves mining, transportation, processing emissions and load because of
improper waste management of end of life batteries. The Li-ion battery has
metals such as Li, Fe, Co, Ni, and Mn, which are in the black mass i.e., and
are the electrochemically active fraction of the batteries. Additionally the
current collectors are made of Al and Cu. After dismantling, heattreatment, and mechanical separation, the black mass is separated from
the current collectors. This black mass was then leached off its metals
using 1 M HNO3. Solvent extraction was chosen to achieve separation and
recovery of metals from this leachate with high purity.
In the current article, a screening study for the separation of Li(I), Ni(II),
Mn(II),Co(II), Fe(III) from nitrate leachate using both acidic and solvating
extractants was made. Solvent extraction studies were done using an
AKUFVE unit. The separation studies that were made with acidic extractant
were made with respect to pH and extractant concentration and studies
with solvating extractant were made with respect to nitrate concentration
and extractant concentration.

INTRODUCTION
Hybrid electric vehicles (HEVs) are gaining popularity as one of the more economic choices for
transportation [German & Cutler, 2004]. The Li-ion-batteries with varying chemistries are one of the
favored batteries for these vehicles[Ritchie, 2001]. The growing numbers of HEVs will soon result in
a large number of end of life batteries in next decade [Andersson & Rde, 2001]. Currently there is
little infrastructure to handle or recycle, end of life Li-ion batteries. The economic incentive for
recycling will also be soon lost as the new chemistries are replacing the conventional cobalt
chemistries with other cheaper metals. However, recycling when understood from a cumulative
reasoning such as resource re-use and reducing environmental load caused due to virgin metal
extraction which involves- mining, transportation, processing emissions and load due to improper
waste management of end of life batteries [Dewulf et al].The Li-ion battery has metals such as Li, Fe,
Co, Ni, Mn- which are in the black-mass i.e., the electrochemically active fraction of the batteries.
Additionally the current collectors are made of Al and, Cu. After dismantling, heat-treatment, and
mechanical separation, the black mass is separated from the current collectors. This black mass was
then leached off its metals using 1 M HNO3. Solvent extraction was chosen to achieve separation
and recovery of metals from this leachate with high purity.
The extractants- DEHPA and TBP, were previously used to study the separation of indium from
LCD screen panels by Yang et.al [Yang et al, 2013]. Stas et.al have used it to study the separation of
uranium[Stas et al, 2002]. DEHPA was used to separate molybdenum and tungsten [MorS et al,
1999]. Separation of Li from Li- ion battery waste was previously done by leaching follwed by pH
adjustment to alkaline conditions and precipitation either by carbonate additions [Granata et al,
2012] or crystallization by evaporation, in which case removal of other metals becomes important
[Hamzaoui et al, 2003]. Cobalt was selectevely separated from Li previously using PC-88A
extractant[Zhang et al, 1998].

METHODOLOGY
Each battery has many individual Li-ion cells (cathode/separator-electrolyte/anode) stacked one
after the other. The batteries tested for the purpose of this article had carbon as anode and for this
reason was left out of the present discussion where, only metal separation investigations are of
interest. The batteries were dismantled in inert atmosphere and heat-treated to separate the metallic
substrates Cu and Al from the black mass. After heat treatment, mechanical dismantling and
crushing steps, the black mass was then leached with 1M HNO3 to dissolve the constituent metals. The
present studies are made on synthetic solutions prepared with the same relative ratios of the metalions found in the leachate.
In the current article screening study for the separation of Li(I), Ni(II), Mn(II),Co(II), Fe(III) from
nitrate leachate using both acidic and solvating extractants was made. Solvent extraction studies
were done using an AKUFVE unit. In the scope of the present work, an investigation of separation
of Li(I), Ni(II), Mn(II),Co(II) and Fe(III) from nitrate leachate using DEHPA (Bis(2-ethylhexyl)
hydrogen phosphate, acidic extractant) and TBP (Tributyl Phosphate, solvating extractant),
separately was made. The separation studies with DEHPA were made with respect to pH and
extractant concentration while studies with TBP were made with respect to nitrate concentration
and extractant concentration.

Reagents
Li-ion batteries for this study were provided by ETC Battery and Fuel cells Sweden AB. The
extractants- DEHPA (Bis(2-ethylhexyl) hydrogen phosphate) and TBP (Tributyl Phosphate, 97%)
and HNO3 (69%, puriss), were procured from Sigma Aldrich. Solvent 70 aliphatic kerosene
distilled at 190-250C was from Statoil. Nitrate salts of Li, Mn, Ni, Fe, Co and sodium perchlorate
were all from Merck, Proanalysi.

AKUFVE
The AKUFVE apparatus was used for solvent extraction studies. The apparatus was described
clearly elsewhere [Rydberg, 1969; Albinsson, 1988, Ekberg et al, 2006]. This apparatus gives good
control of a small scale extraction with regard to temperature, pH, centrifuging for separation of
phases and automatic sampling reducing the huge uncertainties that can come from human
handling and subsequent organic phase entrenchment upon the aqueous phase. The schematic of
the AKUFVE apparatus with its main parts labeled is shown in Figure 1.

Figure 1 A schematic of the AKUFVE unit [Ekberg et al, 2006]

Chemical system used and experimental procedure

Two different systems were investigated- both with 1 M HNO3 aqeos media, 10E-5 M metal ions,
1. Separation of metals with DEHPA at concentrations 0.1, 0.5 and 1 M in solvent 70 diluent, at
increasing pH.

2. Separation of metals with TBP at concentration 0.1, 0.5 and 1 M in solvent 70 and separation
with varying concentrations of nitrate.
For the pH increase 0.5-3M NaOH was used. For the effect of nitrate the ionic strength was kept
close to 2, so as to keep the electrolyte ionic strength is much higher than any of the effects of the
counter-ions from the metal salts. In the same study (effect of nitrate the aqueous solution was
made with 10-5 M metal concentrations in 1 M HClO4 and 1 M NaClO4 to keep the starting H+
concentration the same as the leachate The D-values of the metals was checked before adding any
nitrate, to make sure that the metals are not extracted due to complexes formed with perchlorate.
Nitrate additions were made using 10.3M NaNO3. The temperature was set constant at 21 1C,
with a mixing speed of 5 000-4 000 rpm, and centrifuge speed of 15 000-16 000 rpm. The distribution
ratio (D-values) and the consequent percentage extraction (% E values), and the separation factor
(SF values) are calculated by the formulae shown below. The organic to aqueous volume ratio ()
was maintained at 1, at all times.
Aqueous samples that were taken from the solvent extraction experiments were diluted using 1 M
HNO3 and measured using an inductively couples plasma emission spectrometer (ICP-OES). The
distribution ratios (DM) were calculated by

(2)

=
%

(1)

( +(

))

(3)

The concentration of metal ions in the organic phase

, for each sample was calculated
,
as the difference between the starting aqueous phase metal concentration and remnant metal
concentration in aqueous phase after extraction in to the organic phase, as shown below. However,
since these values were not experimentally determined, it could potentially add uncertainties to the
presented data, due to the possibility of of losses because of a third phase formation or even volume
losses due to improper experimentation.

(4)

RESULTS AND DISCUSSION

Effect of concentration of DEHPA on separation
Figures 2-4, show the separation of metals with respect to pH at 0.1, 0.5 and 1 M concentrations of
DEHPA in solvent-70. In the 0.1 M DEHPA system, the log(D) values of Li remain constant. The

Log(D) values of Fe have an increasing trend until pH 1.5 and then remains constant for the rest of
the pH range. Mn log(D) values have an increasing trend from pH 1 and remain constant from pH
3. The log(D) values of Ni and Co increase from pH 2.5.
In the 0.5 M DEHPA system, the log(D) values for Li still remain constant, however they are a bit
higher than the 0.1 M DEHPA system. The log(D) for Fe remain constant until pH 2.2 and then
increase until pH 3, after which they remain constant. The log(D) values for Mn increase from pH 1
to pH 3. The values of log(D), Mn are considerably higher in the 0. 5 M DEHPA system as compared
to 0.1 M DEHPA. The log(D) curves for Ni and Co have the same trend as in 0.1 M DEHPA system ,
but higher values.
In the 1 M DEHPA system, the log(D) values for Li, slightly increase after pH 3. For Fe, they
increase sharply after pH 2.9 and remain constant after. The curve for Mn has an increasing trend
from pH 1 and remains constant after pH 3. The trend for Co and Ni remain the same as the trend
in the 0.5 and 0.1 M DEHPA system.

Co

Fe

Li

Mn

Ni

Log(D)

5
3
1

-1
-3

pH

Figure 2 Extraction curves for cobalt, iron, lithium, manganese and nickel by 0.1 M DEHPA in Solvent-70;
aqueous phase: 1 M HNO3. The data is based on two experiments at 21 C

Co

Fe

Li

Mn

Ni

Log(D)

5
3
1

-1
-3

pH

Figure 3 Extraction curves for cobalt, iron, lithium, manganese and nickel by 0.5 M DEHPA in Solvent-70;
aqueous phase: 1 M HNO3. The data is based on two experiments at 21 C

Co

Fe

Li

Mn

Ni

Log(D)

5
3
1

-1
-3

pH

Figure 4 Extraction curves for cobalt, iron, lithium, manganese and nickel by 1 M DEHPA in Solvent-70;
aqueous phase: 1 M HNO3. The data is based on two experiments at 21 C

Effect of concentration of TBP on extraction of metal ions

Figure 5 shows the separation of metals at 0.1, 0.5 and 1 M concentrations of TBP. The log(D)-values
of all metal-ions at each concentration were similar except for Li. The log(D) value for Li increased
with TBP concentration.

Co

Fe

Li

Mn

Ni

Log(D)

-0,3
-0,6
-0,9
-1,2
-1,5
-1,8

-1,1

-0,6

-0,1

Log([TBP])

Figure 5 Effect of TBP concentration on separation of of cobalt, iron, lithium, manganese and nickel. Org
phase: 0.1 to 1 M TBP in Solvent-70; aqeous phase: 1 M HNO3. The data is based on two experiments at 21 C

Co

Fe

Li

Mn

Ni

Log(D)

1
0
-1
-2

-1,5

-1

-0,5

Log([nitrate])

0,5

Figure 6 Effect of nitrate concentration on separation of cobalt, iron, lithium, manganese and nickel. Organic
phase: 0.1 M TBP in Solvent-70; aqeous phase: 1 M HClO4 and 1 M NaClO4. The data is based on two
experiments at 21 C

Effect of concentration of nitrate on extraction of metal ions

Figure 6 shows the effect of nitrate concentration on the extraction of the metals in the 0.1 M TBP
system. The log(D) trend of all metals increases after the nitrate concentration was higher than 0.4
M, especially Li.

CONCLUSION
With the acidic extractant DEHPA, Fe always remained strongly extracted in organic phase. All
metals were better extracted and separated in the 1 M DEHPA, higher concentrations of extractant,
made extraction of the metal ions happen at more acidic pH range, compared to the lower
concentrations of the extractant.
Increase in TBP concentration only increased the extraction of Li. When the nitrate concentration is
increased above of 0.4M the extraction of all the metal ions improves, more significantly for Li.
Further investigations with stripping might give an insight as to the fact if separation of metal ions
involved is possible with different stripping solutions. It could also have a possibility to be used
along with an acidic extractant such DEHPA, for synergistic extraction.

ACKNOWLEDGEMENTS
The authors wish to extend their gratitude to the Chalmers Energy Initiative (CEI) for funding this
work.

NOMENCLATURE
Li-ion
,

SFA/B
%E
AKUFVE

ICP-OES
S/L
DEHPA
TBP
Solvent 70
Black-mass

Lithium-ion (battery)
Total metal concentration in organic phase
Total metal concentration in aqueous phase
Organic to aqueous phase volume ratio
Distribution ratio of metal M (D-value)
Separation factor of metal A and B defined to be >=1
Percentage extraction
Apparatus for continuous measurement of partition factor in solvent extraction
(Anordning fr Kontinuerlig Underskning av Frdelningsjmvikter vid Vtske
vtske Extraktion)
Inductively coupled plasma-optical emission spectroscopy
Solid to liquid ratio
Extractant, Bis(2-ethylhexyl) hydrogen phosphate
Tributyl Phosphate
Paraffinic base oil -70 neutral solvent
The electrochemically active substance combined with carbon powder coated on
the electrodes

REFERENCES
Albinsson. (1988) Development of the AKUFVE-LISOL technique. In Solvent extraction studies of
lanthanide acetylacetonates PhD-Thesis thesis. CTH, S-412 96, Gothenburg, Sweden
Andersson, Rde.(2001) Metal resource constraints for electric-vehicle batteries. Transportation
Research Part D: Transport and Environment 6: 297-324.
Dewulf, Van der Vorst, Denturck, Van Langenhove, Ghyoot, et al.(Recycling rechargeable lithium
ion batteries: Critical analysis of natural resource savings. Resources, Conservation and
Recycling 54: 229-34.
Ekberg, Persson, degaard-Jensen, Albinsson, Andersson.(2006) Redox control in solvent extraction
studies using a PEEK AKUFVE unit. Solvent Extraction and Ion Exchange 24: 1-7.
German, Cutler. (2004) Hybrid Electric Vehicles. In Encyclopedia of Energy, pp. 197-213. New York:
Elsevier.
Granata, Moscardini, Pagnanelli, Trabucco, Toro.(2012) Product recovery from Li-ion battery
wastes coming from an industrial pre-treatment plant: Lab scale tests and process
simulations. Journal of Power Sources 206: 393-401.
Hamzaoui, M'Nif, Hammi, Rokbani.(2003) Contribution to the lithium recovery from brine.
Desalination 158: 221-4.
Mors, Dez, Coca.(1999) Solvent extraction of molybdenum and tungsten by Alamine 336 and
DEHPA in a rotating disc contactor. Separation and Purification Technology 17: 173-9.
Ritchie.(2001) Recent developments and future prospects for lithium rechargeable batteries. Journal
of Power Sources 96: 1-4.
Rydberg.(1969) Solvent extraction studies by the AKUFVE method. Acta Chemica Scadinavica 23
Stas, Dahdouh, Shlewit, Khorfan.(2002) Statistical study of factors affecting the co-extraction of
uranium and iron in the second cycle of extraction with DEHPA/TOPO in kerosene.
Hydrometallurgy 65: 23-30.
Yang, Retegan, Ekberg.(2013) Indium recovery from discarded LCD panel glass by solvent
extraction. Hydrometallurgy 137: 68-77.
Zhang, Yokoyama, Itabashi, Suzuki, Inoue.(1998) Hydrometallurgical process for recovery of metal
values from spent lithium-ion secondary batteries. Hydrometallurgy 47: 259-71.

Solvent extraction for separation

of metals from hev Li-ion batteries
sulfate leach liquor

Sravya Kosaraju,
Christian Ekberg and
Stefan Allard
Chalmers University of
Technology, Sweden

Battery driven automobiles are quickly gaining popularity. One of the

popular types of batteries for these vehicles is the Li-ion with varying
chemistries. The growing usage of these batteries, at the end of their life
have little infrastructure to be handled or recycled. As the previous cobalt
chemistries are now being replaced by other cheaper chemistries, the
incentive and hence the motivation to recycle these batteries is presently
scant. The motivation for recycling Li-ion batteries can be strengthened
when seen from the resource and environmental perspective rather than
just an economic one, at least until the number of end of life batteries
reach a significant number which is bound to happen in the coming
years. Recyling helps reduce the environmental effects of virgin metal
extractions and tackle the problem of improper disposal of end-of-life
batteries. When and if recycling is done with the objective of recovering
pure streams of the involved metals, it will become an economically,
environmentally and resource-wise a rewarding process to the end-of-life
Li-ion batteries or any other product. Some of the metals in a Li-ion
battery other than Li and Co are, Fe, Ni, Mn, which are in the black-mass
i.e., the the electrochemically active metal containing fraction of the
batteries. In addition Al and, Cu are present as pure metal foils that serve
as current collectors and are attached to the black mass. Solvent
extraction was chosen to achieve separation and recovery of metals with
higher purity from these leachates. In the scope of the present work, a
screening study for the separation of Li(I), Ni(II), Mn(II),Co(II) from sulfate
leachate using both acidic and solvating extractants was made. Solvent
extraction was chosen to achieve separation and recovery of metals with
higher purity from these leachates. The separation studies that were
made with acidic extractant were made with respect to pH and extractant
concentration and studies with solvating extractant were made with
respect to sulfate concentration and extractant concentration.

INTRODUCTION
Recycling vehicle scale Li-ion batteries soon will be necessary due to the growing trend in battery
run automobile market from as early as 2015 [Al-Alawi & Bradley, 2013]. The battery powered
vehicles require batteries with properties such as high specific volumetric and gravimetric energy
densities. Recycling these batteries also presents challenges due to the same properties. However,
recycling Li-ion batteries and focusing on the recovery of pure metal streams will help tackle the
problem of improper disposal of end of life batteries, along with giving economic and
environmental benefits by recovery of metals such as Ni, Mn, Co, Li, Cu and Al, which are often
found in these batteries [Ritchie & Howard, 2006]. For this particular reason the technique solvent
extraction is chosen.
In the scope of the present work, an investigation of separation of Li(I), Ni(II), Mn(II),Co(II) and
Fe(III) from sulphate leachate using DEHPA (acidic extractant) and TBP (solvating extractant),
separately was made. This study is interesting due to the fact that the extractants only differ by an
OH functional group. Also Li-ion is the smallest and the most electropositive metal ion, while the
rest of them are d- block elements with very similar oxidation numbers and ionic raddi and a wide
range of coordination numbers between 4 to 8. Thus the complexation mechanisama nd
subsequently their separation and extraction can be complex. The separation studies with DEHPA
were made with respect to pH and extractant concentration while studies with TBP were made
with respect to sulphate concentration and extractant concentration.
DEHPA and TBP were previously used to study the separation of indium from LCD screen panels
by Yang et.al [Yang et al, 2013]. Stats et.al have used it to studythe separation of uranium[Stas et al,
2002]. DEHPA was used to separate molybdenum and tungsten [MorS et al, 1999].

METHODOLOGY
Each battery has many individual Li-ion cells (cathode/separator-electrolyte/anode) stacked one
after the other. The batteries tested for the purpose of this article had carbon as anode and for this
reason was left out of the present discussion where, only metal separation investigations are of
interest. The batteries were dismantled in inert atmosphere and heattreated to separate the metallic
substrates Cu and Al from the black-mass. After heat treatment, mechanical dismantling and
crushing steps, the black mass was then leached with 1M H2SO4 to dissolve the metals contained in it. 1
M H2SO4 solution to dissolve the constituent metals. The present studies are made on stock
solutions prepared with the same relative ratios of the metal-ions found in the leachate.

Reagents
Li-ion batteries for this study were provided by ETC Battery and Fuel cells Sweden AB. The
extractants- DEHPA (Bis(2-ethylhexyl) hydrogen phosphate) and TBP (Tributyl Phosphate, 97%)
and HNO3 (69%, puriss), were procured from Sigma Aldrich. Solvent 70 aliphatic kerosene
distilled at 190-250C was from Statoil. Sulphate salts of Li, Mn, Ni, Fe, Co and sodium perchlorate
were all from Merck, Proanalysi.

Figure 1 TBP (Tributyl Phosphate) [Billard, 2013]

Figure 2 DEHPA (Bis(2-ethylhexyl) hydrogen phosphate) [Billard, 2013]

AKUFVE
The AKUFVE apparatus was used for current separation studies. The apparatus is described clearly
elsewhere [Rydberg, 1969, Albinsson, 1988, Ekberg et al, 2006]. This apparatus gives good control of
a small scale extraction with regard to temperature, pH, centrifuging for separation of phases and
automatic sampling reducing the huge uncertainties that can come from human handling and
subsequent organic phase entrenchment upon the aqueous phase.

Chemical system used and experimental procedure

Two different systems were investigated,
1.

Separation of metals with DEHPA at concentrations 0.1, 0.5 and 1 M in solvent 70 diluent,
at increasing pH.

2.

Separation of metals with TBP at concentration 0.1, 0.5 and 1 M in solvent 70 and separation
with varying concentrations of sulphate.

For the pH increase 0.5-3M NaOH was used. For the effect of sulphate the ionic strength was kept
close to 2, so as to keep the electrolyte ionic strength is much higher than any of the effects of the
counter-ions from the metal salts. In the same study (effect of sulphate the aqueous solution was
made with 10-5 M metal concentrations in 1 M HClO4 and 1 M NaClO4 to keep the starting H+
concentration the same as the leachate The D-values of the metals was checked before adding any
sulphate, to make sure that the metals are not extracted due to complexes formed with perchlorate.
Sulphate additions were made using 1.37 M Na2SO4. The temperature was set constant at 21 1C,
with a mixing speed of 5 000-4 000 rpm, and centrifuge speed of 15 000-16 000 rpm. The distribution
ratio (D-values) and the consequent percentage extraction (% E values), and the separation factor
(SF values) are calculated by the formulae shown below. The organic to aqeous volume ratio ()
was maintained at 1, at all times.

Aqueous samples that were taken from the solvent extraction experiments were diluted using 1 M
HNO3 and measured using an inductively couples plasma emission spectrometer (ICP-OES). The
distribution ratios (DM) were calculated by

(2)

=
%

(1)

( +(

(3)

))

The concentration of metal ions in the organic phase

, for each sample was calculated
,
as the difference between the starting aqueous phase metal concentration and remnant metal
concentration in aqueous phase after extraction in to the organic phase, as shown below. However,
since these values were not experimentally determined, it could potentially add uncertainities to
the presented data, due to the possibility of of losses because of a third phase formation or even
volume losses due to improper experimentation.

(4)

The dilution to the starting aqeous solution due to base additions at every step was taken into
consideration in the calculations.

RESULTS AND DISCUSSION

Effect of concentration of DEHPA on separation
Figures 3-5, show the separation of metals with respect to pH at 0.1, 0.5 and 1 M DEHPA. This
cationic extractant shows good separation of all the metal-ions from each other at all concentrations
of the extractant studied except for iron and lithium. As for the percentage extraction, for the 0.1 M
DEHPA system, the %E50 of lithium and iron was in the acidic region(pH<1). The %E50 of
manganese was at pH 2.5, for cobalt at pH 3.8, and nickel at pH 4.7.
With increased concentration of the extractant DEHPA the %E 50 for all metal-ions comes in at a
more acidic range, compared to the 0.1 M system. For the 0.5 M system the %E 50 of lithium and iron
still was in the pH range <1, while managanese %E50 is at pH 1.6, cobalt at pH 3, and, nickel at pH
3.4.
For the 1 M system the %E50 of lithium and iron remain in the pH range <1, while managanese %E 50
was at pH 1.3, cobalt at pH 2.2, and, nickel at pH 2.7.

Log(D)

6
5
4
3
2
1
0
-1
-2
-3

Co

Fe

Li

Mn

Ni

pH

Log(D)

Figure 3 Extraction curves for cobalt, iron, lithium, manganese and nickel by 0.1 M DEHPA in Solvent-70;
aqeous phase: 1 M H2SO4. The data is based on two experiments at 21 C

6
5
4
3
2
1
0
-1
-2
-3

Co

Fe

Li

Mn

pH

Ni

Figure 4 Extraction curves for cobalt, iron, lithium, manganese and nickel by 0.5 M DEHPA in Solvent-70;
aqeous phase: 1 M H2SO4. The data is based on two experiments at 21 C

Log(D)

6
5
4
3
2
1
0
-1
-2
-3

Co

Fe

Li

Mn

Ni

pH

Figure 5 Extraction curves for cobalt, iron, lithium, manganese and nickel by 1 M DEHPA in Solvent-70;
aqeous phase: 1 M H2SO4. The data is based on two experiments at 21 C

Effect of concentration of TBP on extraction of metal ions

Figure 6 shows the separation of metals at varying concentrations of TBP. TBP seems to act more as
group extractant. The D-values of all metal-ions at each concentration were similar. The increase in
concentration of TBP has increased the D-values for all-metal ions.

Effect of concentration of sulphate on extraction of metal ions

Figure 7 shows the effect of sulphate on the extraction of the metals in the 0.1 M TBP system. The
%E of all metals increases considerably after the sulphate concentration was higher than 0.15 M.

Log(D)

Co

-0,3

Fe

Li

Mn

Co

-0,4
-0,5
-0,6
-0,7

-1,2

-1

-0,8

-0,6

Log([TBP])

-0,4

-0,2

Figure 6 Effect of TBP concentration on separation of of cobalt, iron, lithium, manganese and nickel. Org
phase: 0.1 to 1 M TBP in Solvent-70; aqeous phase: 1 M H2SO4. The data is based on two experiments at 21 C

Co

Log(D)

Fe

Li

Mn

Ni

0,5
0

-0,5
-1

-1,5

-2,5

-2

-1,5

-1

Log([sulphate])

-0,5

Figure 7 Effect of sulphate concentration on separation of cobalt, iron, lithium, manganese and nickel.
Organic phase: 0.1 M TBP in Solvent-70; aqeous phase: 1 M HClO4 and NaClO4. The data is based on two
experiments at 21 C

CONCLUSION
The acidic extractant DEHPA showed good separation of all metal ions except iron and lithium.
Higher concentrations made extraction of the metal ions studied at more acidic pH range. This
could potentially be used in battery recycling but for waste streams which do not have iron, or
perhaps as an intial extraction system where in iron and lithium would be later separated with a
different extractant.
TBP acted as group extractant. Further investigations with stripping might give an insight as to the
fact if separation is possible with different stripping solutions. It could also have a possibility to be
used along with an acidic extractant such DEHPA, for synergistic extraction.

ACKNOWLEDGEMENTS
The authors wish to extend their gratitude to the Chalmers Energy Initiative (CEI) for funding this
work.

NOMENCLATURE
Li-ion
,

SFA/B
%E
AKUFVE

ICP-OES
S/L
DEHPA
TBP
Solvent 70
Black-mass

Lithium-ion (battery)
Total metal concentration in organic phase
Total metal concentration in aqueous phase
Organic to aqueous phase volume ratio
Distribution ratio of metal M (D-value)
Separation factor of metal A and B defined to be >=1
Percentage extraction
Apparatus for continuous measurement of partition factor in solvent extraction
(Anordning fr Kontinuerlig Underskning av Frdelningsjmvikter vid Vtske
vtske Extraktion)
Inductively coupled plasma-optical emission spectroscopy
Solid to liquid ratio
Extractant, Bis(2-ethylhexyl) hydrogen phosphate
Tributyl Phosphate
Paraffinic base oil -70 neutral solvent
The electrochemically active substance combined with carbon powder coated on
the electrodes.

REFERENCES
Al-Alawi, Bradley.(2013) Review of hybrid, plug-in hybrid, and electric vehicle market modeling Studies.
Renewable and Sustainable Energy Reviews 21: 190-203.
Albinsson. (1988) Development of the AKUFVE-LISOL technique. In Solvent extraction studies of lanthanide
acetylacetonates PhD-Thesis thesis. CTH, S-412 96, Gothenburg, Sweden
Billard. (2013) In Handbook on the Physics and Chemistry of Rare Earths. Elsevier

Ekberg, Persson, degaard-Jensen, Albinsson, Andersson.(2006) Redox control in solvent extraction studies
using a PEEK AKUFVE unit. Solvent Extraction and Ion Exchange 24: 1-7.
Mors, Dez, Coca.(1999) Solvent extraction of molybdenum and tungsten by Alamine 336 and DEHPA in a
rotating disc contactor. Separation and Purification Technology 17: 173-9.
Ritchie, Howard.(2006) Recent developments and likely advances in lithium-ion batteries. Journal of Power
Sources 162: 809-12.
Rydberg.(1969) Solvent extraction studies by the AKUFVE method. Acta Chemica Scadinavica 23
Stas, Dahdouh, Shlewit, Khorfan.(2002) Statistical study of factors affecting the co-extraction of uranium and
iron in the second cycle of extraction with DEHPA/TOPO in kerosene. Hydrometallurgy 65: 23-30.
Yang, Retegan, Ekberg.(2013) Indium recovery from discarded LCD panel glass by solvent extraction.
Hydrometallurgy 137: 68-77.

Editors

Fernando Valenzuela L.
Professor of Unit Operations and Mining and Metallurgical Chemistry at
The Universidad de Chile. He is a chemist and has a Masters degree in
Chemical Engineering from Kyushu University, Japan. He is author of 80 ISI
scientific publications, chapters of books and international patents on
mining wastewaters processing. Today, he is a member of the PhD in
Chemistry Faculty meeting at his university and holds the position of
Academic Director of his faculty.

colophon

This book was designed by Estudio Vicencio. The text was

output in the font Arnhem Pro designed by Fred Smeijers
and Ant designed by Rodrigo Ramrez. The cover
was printed on Environment paper and
the inside pages on Bond paper.
The first edition of 450 copies was
printed and bound in May
2014 by QuadGraphics,
Santiago, Chile.

***