Separation and Purication Technology 123 (2014) 200213

Contents lists available at ScienceDirect

Separation and Purication Technology

journal homepage: www.elsevier.com/locate/seppur

Biogas desulfurization by adsorption on thermally treated

sewage-sludge
F.J. Gutirrez Ortiz , P.G. Aguilera, P. Ollero
Departamento de Ingeniera Qumica y Ambiental, Universidad de Sevilla, Camino de los Descubrimientos, s/n, 41092 Sevilla, Spain

a r t i c l e

i n f o

Article history:
Received 25 May 2013
Received in revised form 21 November 2013
Accepted 29 December 2013
Available online 3 January 2014
Keywords:
Biogas
Hydrogen sulde
Desulfurization
Breakthrough
Adsorption
Sewage sludge

a b s t r a c t
Biogas is a renewable source for power production, but the H2S present must be removed because it is
very corrosive and may damage the combustion engines. The adsorption using activated carbon is one
of the most used desulfurization methods. The operational life of the activated carbon could be extended
if the H2S concentration was reduced prior entering the activated carbon bed by using other cheaper
adsorbent. Sewage sludge is a possible inexpensive precursor to obtain adsorbents, and thus it would
be valorized. An experimental study was performed using three types of sludge from three Spanish locations, which were activated to increase their adsorption capacity. Two thermal treatments were tested
using nitrogen (pyrolysis) and air (calcination), as well as three heating temperatures. The adsorption
dynamics of the prepared adsorbents were investigated in a xed-bed column, determining the breakthrough curves and adsorption capacity of adsorbents. Besides, both their surface properties and their
chemical properties were analyzed to get more insight about the adsorbent behavior. In addition, the
effect of the oxygen content, relative humidity and the chemical impregnation, using different procedures, were also studied. As a relevant result, the adsorbent obtained by calcination at 700 C of one of
the three kinds of sludge showed a capacity twice of that of a commercial activated carbon without
impregnation, although somewhat lower than that of a commercial activated carbon impregnated with
a NaOH solution. The results showed that the use of this kind of precursors is very attractive to achieve
adsorbents with a relative high adsorption capacity valuable to apply them in an economically feasible
pretreatment.
2014 Elsevier B.V. All rights reserved.

1. Introduction
The organic matter that is present in a landll collecting municipal solid waste is degraded in an anaerobic process, which generates a gas mixture called biogas or landll gas. Biogas is obtainable
in a relatively economical way from anaerobic digestion [1]. The
main constituents of biogas are methane and carbon dioxide, but
it also contains nitrogen and oxygen, due to air incursion into the
gas collection system; oxygen concentration is continuously monitored and held low for safety reasons. Due to the high caloric value of methane and its high concentration, biogas is currently
considered a renewable energy source.
When biogas is used as a fuel for electricity generation, several
trace compounds may damage the combustion engines, requiring
expensive repairs and causing service interruptions. Among all of
the specic contaminants to biogas utilization, hydrogen sulde
(H2S) is the most relevant one, because it is highly corrosive during
the combustion process. Because hydrogen sulde concentration is
of up to 10002000 ppm, H2S removal is necessary to avoid
Corresponding author. Tel.: +34 95 448 72 68/60.
E-mail address: frajagutor@etsi.us.es (F.J. Gutirrez Ortiz).
1383-5866/$ - see front matter 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.seppur.2013.12.025

operational problems, such as corrosion in pipes, turbines or other

units [2], as well as environmental problems such as global warming or acid rain [3].
Biogas desulfurization is normally performed by wet scrubbing,
biological methods, adsorption, or selective catalytic oxidation [4].
The activated carbon adsorption process is normally penalized
with high operational costs related to the price of the adsorbent.
Among these methods, adsorption by dry treated sludge, such as
that coming from the municipal wastewater treatment, may be
benecial as a pretreatment of the biogas, before entering another
system, such as an adsorber using activated carbon. In this way, the
activated carbon would deal with biogas with a lower H2S concentration, and its useful operational life would be extended. Thereby,
sewage sludge, which is an inevitable byproduct of wastewater
treatment, could be valorized.
Generally, three sludge-stabilization methods are typically
used: thermal, chemical and biological [5]. Surplus sludge produced during the biological treatment of wastewater requires
costly disposal procedures. The most frequent methods of disposal
are landll, commercial composting and incineration. With
increasing environmental and legislative constraints, increasing

201

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213
Table 1
Immediate and elemental analyses of activated carbons and sewage sludge.
Parameter (dry basis) wt%

Activated carbon CAT

Activated carbon CAA

Sludge LL

Sludge LF

Sludge LG

Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash

88.12
0.76
0.79
1.37
0.35
Balance

73.32
0.81
0.48
8.10
0.30
Balance

38.49
5.42
6.54
19.42
1.32
Balance

39.12
5.48
5.86
19.77
1.26
Balance

36.90
5.07
4.72
16.38
1.54
Balance

Ash (wet basis)

Volatiles (wet basis)
Fixed carbon (wet basis)
Moisture (wet basis)

(8.25)
(2.48)
(85.09)
(4.18)

(14.67)
(11.84)
(59.81)
(13.69)

(27.11)
(57.42)
(9.57)
(5.90)

(26.79)
(58.56)
(8.63)
(6.02)

(32.96)
(54.62)
(5.53)
(6.89)

sludge production and lessening disposal options, new recycling

alternatives have to be found [6].
The adsorption dynamics of the prepared adsorbents was
investigated in a xed-bed column, using a simulated biogas and
determining the adsorption capacity of the adsorbents. Three
carbon-based adsorbents from dry sewage sludge (low-cost precursors) were tested after suffering a thermal treatment to increase
the adsorption capacity. The research work was focused on studying the effects of the thermal treatment, heating temperature, and
physicochemical properties of the precursors on the removal of
H2S, in the absence of oxygen and humidity. Additionally, the inuences of the relative humidity and the oxygen present in the biogas
entering the adsorber were tested, as well as the effect of chemical
impregnation.
One of the main novelty aspects of the present study is the use
of air as agent to thermally treat the precursors, instead of a pyrolysis (performed under N2 atmosphere), which has been the usual
thermal procedure in the previous studies encountered in the literature. This has a great importance, if the process is to be implanted
in a commercial scale, where using air instead of nitrogen would
reduce the operating costs. Another signicant point of this paper
is the use of a simulated biogas, and not a mixture of H2S and air
normally used in most studies found in the literature, because
CO2 may compete against H2S in adsorption depending on the porous structure of the adsorbent and the alkali constituents, since
both of them are acid gases.

2. Experimental section
2.1. Materials
Two commercial activated carbons, labeled as CAT and CAA,
were used as references to compare the results using the precursor-sludge. CAT is a fresh activated carbon, without impregnation,
while CAA is activated carbon impregnated with a NaOH solution.
Both of them were milled and sieved between 1.41 and 2.83 mm,
as the rest of the materials tested. Likewise, three types of air-dried
sewage-sludge were crushed to pass between 1.41 and 2.83 mm
sieve. They are referred to as LG, LL, and LF, respectively, based
on their Spanish provenance. Table 1 shows the elemental and
immediate analyses of the two activated carbons and the three
types of sewage-sludge, as received, i.e., before undergoing any
treatment.
Experiments were carried out by a certied mixture of CH4
(60 vol.%), H2S (2000 ppmv) and CO2 (balance).

2.2. Methods
Precursors were treated in order to improve their adsorption
capacity by thermal and chemical procedures. Then, they were

tested by an experimental unit to obtain their H2S breakthrough

capacity and characterized by different analytical techniques.
2.2.1. Thermal treatment of sludge
Physical activation involves calcination or pyrolysis of a precursor. The thermal treatment was performed in a tubular furnace,
illustrated in Fig. 1. The pyrolysis (heat treatment carried out in
an inert atmosphere) was performed with 100 g of precursor using
a nitrogen ow-rate of 1 L/min from room temperature up to
500 C. Beyond this temperature, no more nitrogen was used and
an air ow-rate of 0.125 L/min passed through the tubular furnace
until reaching the nal temperature (500 C, 700 C or 900 C).
During this latter step, calcination takes place and the metals present in the sample are oxidized. The code of this method was PA. As
a second method, the calcination was performed entering air from
room temperature up to the nal temperature. It was coded using
the letter A.
The heating rate was always 5 C/min and the holding time was
always 30 min, which was selected by a prior screening test. After
that time, samples were withdrawn from the furnace.
The samples were coded as follows: rst, the acronym of the
precursor (LL, LF and LG), then the maximum temperature reached,
and nally the treatment (PA or A).
2.2.2. Experimental adsorption unit
A lab-scale facility was designed and assembled to carry out the
adsorption tests. It consists of three main parts: gas feeding,
adsorption system and H2S analyzer. Fig. 2 depicts a scheme of
the experimental unit.
Experiments were carried out at controlled room temperature
(20 2 C), using a simulated biogas ow-rate of 1.1 L/min. The
xed bed tower was a glass tube (30-mm ID and 430 mm height),
with a perforated plate acting as gas distributor and adsorbent support. The adsorbent bed was 100 mm height. H2S was measured by
a continuous gas analyzer, based on ultraviolet radiation,
with three possible ranges (0200 ppm, 01200 ppm and
02500 ppm). A humidication system based on three bubblers
serially assembled was used for tests with humidication.
To investigate the effect of humidity on H2S removal efciency,
specic H2S removal tests were performed by two ways (at room
temperature): (1) by pre-humidifying the adsorbent using a saturated air ow-rate of 1.1 L/min for one hour (in this case, the PH
symbol is added at the end of the label), or (2) directly, by using
a saturated biogas (100% relative humidity at 20 C), adding H at
the end of the code.
Additionally, the effect of the oxygen on the H2S removal was
studied by adding oxygen (5 vol.%) to the dry biogas stream before
entering the xed bed.
Adsorption/removal capacities, x/M (mg H2S/g material), were
calculated by integrating the corresponding breakthrough curves
and by applying Eq. (1) [7]:

202

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

(1) Dry N2 bottle, (2) Pressure reducer, (3) Air feed, (4) Gas rotameters, (5) Tubular furnace, (6) Copper
spiral, (7) Ball valve, (8) Safety valve, (9) Bubblers, (10) Cooling water feed, (11) Drain
Fig. 1. Facility scheme for thermally treating the precursors in a controlled atmosphere.

ID Fan

To the Atmosphere
Rotameter
FI-2

Adsorption
tower
E-1

E-3
S.A1

Conditioning
system

Al-3
H2S Analyzer

U-tube
I-1

Pressure reducer
V-1

FI-1
Rotameter
E-2
Condensate
container

Biogas bottle
Composition
CH4 (60 vol%)
H2S (2000 ppm)
CO2 (balance)
Fig. 2. Scheme of the laboratory plant for performing the adsorption tests.

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213



Z ts
x
Q  MW

 c0  t s 
ctdt
M
w  VM
0

where Q is the total inlet ow rate (m3/s), w is the weight of sludgebased material introduced into the column (g), MW is the molecular
weight of H2S (34 g/mol), VM is the molar volume (22.4 L/mol), c0 is
the inlet gas H2S concentration (ppmv), c(t) is the gas outlet concentration (ppmv), and ts is the bed saturation/exhaustion time (s). The
test was stopped at the breakthrough concentration of 200 ppm;
thus, ts is the time corresponding to this concentration.
Experimental tests were validated by performing, at least, three
replicates, verifying thus the reproducibility and the feasibility of
the tests.
2.2.3. Preparation of impregnated precursors
With the aim of improving the activated carbons performance,
these may be impregnated with caustic materials such as NaOH
or KOH. Unlike activated carbons, sewage-sludge (either thermally
treated or not treated) cannot be immersed in stirred caustic solutions because they disaggregate and partially dissolve. Then,
impregnation procedures were performed by other two ways: (1)
milling the NaOH or KOH lentils and mixing with the sludge-based
material, and (2) atomizing a concentrated aqueous solution of
NaOH or KOH on the precursor. In this latter case, due to the addition of water, the precursors were partially dissolved forming an
unwieldy paste that led to a poor result after calcination. The
agents were added using an agent-to-precursor mass ratio ranged
from 0.10 to 0.40.
2.2.4. Characterization of sludge
Surface properties and chemical properties of the precursors
were determined to get more insight about the adsorbent behavior.
The surface properties of adsorbents that affect the adsorption
capacity are mainly surface area, pore volume, and pore size distribution [8].
The BET surface area as well as the micropores and mesopores
areas and volumes were obtained by means of the Micromeritics
ASAP 2420 System, which based on physisorption determination
that uses nitrogen as adsorptive. The IUPAC nomenclature was followed: micropores correspond to pore size lower than 2 nm and
mesopores ranges from 2 to 50 nm. The BET surface area and the
micropores volume were computed using the Harkins and Jura
Thickness equation, and the mesopores and macropores volume
were calculated using the t-method by BarrerJoynerHalenda. Before each test, samples were degasied at 150 C for ten hours.
Regarding the chemical analyses, direct analyses of major elements (C, H, N, S, and O) of the dry sludge and thermally treated
precursors (both before and after carrying out the H2S adsorption
test) were performed using LECO Instruments (CHNS-932). Trace
metals were determined by means of inductively coupled plasma
atomic (optical) emission spectroscopy (Horiba Jobin Yvon Ultima
2 High resolution ICPOES spectrometer).
In order to obtain structural information on an atomic scale of
both crystalline and non-crystalline (amorphous) materials, X-ray
Diffraction (XRD) analyses were carried out by a Bruker D8 Advance A25, equipped with an X-ray tube of 40 kV and 30 mA,
D2h of 370, step scanning with a step size of 0.015 (2h) and a
scan step time of 0.1 s.
X-ray Photoelectron Spectroscopy (XPS) was used (Leybold
Heraeus LHS-10/20) to discriminate on the different oxidation
states.
Scanning Electronic Microscopy (SEM) observations were carried out with a Philips XL30 microscope with secondary and backscattered electron imaging, equipped with an integrated Energy
Dispersive X-ray Spectrometer to analyze the elements and their
contents.

203

3. Results and discussion

Hydrogen sulde can be physically adsorbed, due to the porous
structure of the adsorbent, and chemically adsorbed, because of
the presence of some metals, acting as reagents or catalysts, and
the surface alkalinity.
Table 2 shows the content of potential catalytic or reagent metals in the three types of virgin dry sludge as well as for some of
them after thermal treatment. The possible catalytic activity of
some of the metals and the reactivity of some others towards
hydrogen sulde is discussed below. The results indicate high contents of calcium and iron in the adsorbents, which should have a
positive effect on the hydrogen sulde oxidation. Other alkali metals, such as sodium and potassium, are in a too low concentration
to have a signicant effect on the H2S removal. Apart from iron, alkali metals and alkaline earth metals, copper and zinc oxides also
have a catalytic activity for H2S removal [9,10].
Table 3 shows the weight loss of precursors in the tubular furnace as well as the adsorption capacity in the xed-bed experimentally obtained, in the absence of humidity and oxygen.
Pyrolysis or calcination result in changes in the sludge chemistry
with a signicant weight loss between 200 and 500 C due to volatization of organic matter, but probably also by dehydroxylation
of hydroxides and decomposition of inorganic salts as pointed previously by other researchers [11]. More than 50% of the dry weight
was lost during the thermal treatment.

3.1. Breakthrough curves: adsorption capacity

In order to establish a suitable comparison framework for the
precursors performance, two activated carbons were selected.
Fig. 3 shows that the adsorption capacity of the impregnated
activated carbon (CAA) is quite higher than that of the fresh one
(CAT). However, the former signicantly increases the cost of the
adsorbent (by threefour times). The adsorption capacity of CAT
and CAA was 4.42 and 12.75 mg/g, respectively.
Fig. 4 depicts the H2S breakthrough curves for the three adsorbents, using the PA thermal treatment. The sludge LF has a lower
adsorption capacity than that of the fresh activated carbon (CAT)
for all of the temperatures tested. The sludge LL presents higher
adsorption capacities than LF, although with small differences in
all of the temperature range. Finally, the sludge LG provides the
best results from 500 to 700 C, but a worse adsorption capacity
at 900 C. For some of the pyrolyzed precursors heated to 500 C
(LG 500 PA), it is possible to achieve an adsorption capacity higher
than that obtained with the fresh activated carbon (CAT). This is an
advantage from the energy standpoint in an industrial application;
nevertheless, not all of the treated precursors at 500 C make it
possible to overcome the fresh activated carbon adsorption capacity. The low capacity, less than that expected for the low temperature pyrolyzed materials, is due to their unfavorable volume of the
pores, as explained in the next section; furthermore, another reason could be the deactivation of catalytic centers in those pores
by carbon dioxide [12]. Up to 700 C treatment, the adsorption
capacity increases as the temperature rises. The adsorbent obtained by calcination (treatment A) at 700 C of the sludge LG
has a capacity twice of that of CAT but it is lower that impregnated
activated carbon (CAA). In general, the adsorption capacity of the
sludge-derived materials increases when increasing the calcination
temperature. At high temperature, solid-state reactions may occur
between the components of the sludge, resulting in active catalytic
phases in the form of oxides or carbonates, which matches well
with the literature [13]. However, at 900 C, the adsorption capacity decreased, probably due to sintering and pore closing.

204

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

Table 2
Trace metal composition of precursors and adsorbents obtained.
Precursors

Si (mg/g)

Al (mg/g)

Fe (mg/g)

Ca (mg/g)

Mg (mg/g)

Na (mg/g)

K (mg/g)

P (mg/g)

S (mg/g)

Cu (mg/g)

Zn (mg/g)

LL
LG
LF
LG500A
LG500PA
LG700A
LG700PA
LG900A
LG900PA
LL500PA
LL700A
LL700PA
LL900PA

91.77
126.11
163.48
112.47
100.32
140.21
266.66
109.53
103.80
111.38
112.15
102.40
147.23

17.67
19.15
17.35
39.79
38.39
44.89
49.93
50.30
48.17
39.74
42.72
43.07
48.93

30.43
31.14
16.19
49.57
46.80
64.61
66.62
67.46
66.59
24.38
28.43
28.18
29.42

36.32
43.43
35.21
57.78
54.04
105.56
108.51
83.76
88.34
40.51
41.67
44.49
24.34

11.07
9.21
10.92
18.52
17.78
22.44
18.22
23.46
21.68
11.30
11.96
12.84
14.23

1.63
1.61
1.53
4.76
4.55
4.88
4.05
5.48
5.60
1.77
2.56
3.08
3.05

3.21
3.69
3.26
10.06
9.72
12.04
10.73
13.27
12.58
9.07
9.79
10.13
11.66

18.11
16.48
18.12
33.43
32.32
41.68
39.96
42.02
40.17
43.51
46.63
47.53
55.33

9.05
11.25
8.64
9.09
7.45
12.13
13.08
14.36
12.66
4.11
3.88
4.36
4.64

0.36
0.18
0.44
0.25
0.24
0.33
0.33
0.34
0.32
0.44
0.49
0.49
0.57

0.56
0.38
0.52
1.08
1.07
1.12
1.09
1.02
0.61
2.05
2.23
2.28
1.80

PA denotes a thermal treatment in which a pyrolysis is carried out from room temperature up to 500 C, and then a calcination takes place from 500 C until reaching the
nal temperature (500, 700 or 900 C).
A denotes that the calcination was performed entering air from room temperature up to the nal temperature.

Table 3
Weight loss of adsorbents after being thermally treated and adsorption capacity.
Precursor

Weight loss (%)

Adsorption capacity
(mg H2S/g ads.) limit: 200 ppm

LF 500 PA
LF 700 PA
LF 900 PA
LL 500 PA
LL 700 PA
LL 900 PA
LF 500 A
LF 700 A
LF 900 A
LL 500 A
LL 700 A
LL 900 A
LG 500 PA
LG 700 PA
LG 900 PA
LG 500 A
LG 700 A
LG 900 A

53.79
58.46
62.16
52.59
57.40
61.95
53.21
58.29
61.88
52.73
57.49
61.31
51.80
56.78
61.60
51.00
56.85
61.05

2.23
2.12
4.37
3.42
4.71
3.80
1.41
2.19
3.68
3.37
4.93
3.72
5.15
8.63
2.89
3.65
6.64
2.67

Note 1: The precursor amount entered the electrical furnace was 100 g.
Note 2: The adsorption capacity of CAT and CAA was 4.42 and 12.75 mg/g, respectively.

Fig. 3. Breakthrough curves for the activated carbons used as reference framework.

Fig. 5 shows a comparison between the performance of the two

tested thermal treatments (PA and A). The adsorption capacities of
the treatment PA are somewhat higher than those are in the

treatment A for the precursor LG, while there are no signicant differences between both treatments for the precursor LL. Precursors
LF and LL have a similar behavior, although the former has lower

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

205

(a)

(b)

(c)

Fig. 4. Breakthrough curves for the precursors thermally treated at three temperatures in the absence of humidity and oxygen: precursor (a) LF, (b) LL, and (c) LG.

adsorption capacities. Likewise, both LG and LL present the highest

adsorption capacity at 700 C. This nding is very interesting, because for a full-scale plant, it would be more cost-effective to perform a treatment in which nitrogen was not necessary.
In order to compare with other materials or results from the literature, the presence of CO2, as a major compound along with CH4,

determines the nal adsorption capacity of both activated carbons

and the precursors used in our study, as above mentioned. Thus,
Yuan and Bandosz [12] used a simulated digester gas mixture
(40% of CO2, 60% CH4 at 25 C) containing 0.1% (1000 ppm) of
H2S, and they obtained adsorption capacities between 2 and
20 mg/g in all of their experiments, with mixtures of pyrolized

206

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

Adsorption capacity in mg H2S/g ads. Limit: 200ppm

Fig. 5. Comparison of adsorption capacities of precursors for the two thermal treatments.

Table 4
BET surface area and pore structure for some samples of the treated precursor LG.

BET surface area (m /g)

t-Plot external surface area (m2/g)
t-Plot micropore area (m2/g)
Pore volume (cm3/g)
Micropore volume (cm3/g)
Meso- and macropore volume (cm3/g)
Adsorption capacity (mg H2S/g ads)

LG 500 A

LG 500 PA

LG 700 A

LG 700 PA

LG 900 A

LG 900 PA

49.246
14.209
35.037
0.037
1.464E02
2.241E02
3.65

72.822
21.845
50.977
0.049
2.100E02
2.800E02
5.15

82.150
28.617
53.533
0.060
2.186E02
3.772E02
6.64

89.937
29.031
60.906
0.062
2.500E02
3.700E02
8.63

100.353
61.165
39.188
0.113
2.005E02
9.293E02
2.67

118.096
69.626
48.470
0.101
2.100E02
8.000E02
2.78

sewage sludge and metal sludge. Likewise, other researchers [8]

achieved adsorption capacity values of the same magnitude order
than in the present study, using a mixture of He and H2S, with
commercial and modied activated carbons. However, if a mixture
of H2S and humid air is used, the adsorption capacities are multiplied by 23 with gasied raw sludge [7].
3.2. Surface properties of the adsorbents
Tables 4 and 5 show the surface area measured by the BET technique and pore structure for some of the tested precursors.
Structural parameters change as the calcination temperature increases. Thus, it can be observed that the surface area increases as
the treatment temperature rises. Likewise, the treatment PA makes
it possible to enlarge the surface area more than the treatment A,
except for the precursor LL 700 PA that has a lower surface area
than the precursor LL 700 A. Hence, contrary to one would expect,
the overall BET surface area is not a good parameter to identify the
best adsorbent, because if it were so, the treatment temperature of
900 C should be chosen, but the highest adsorption capacity was
obtained at 700 C.
Up to 700 C, the increase in the values of specic surface area is
due to both the contribution of micro-, meso- and macropore volume, in the pore structure. The micropore volume achieves a maximum value, due to the decomposition of an inorganic matter and
release of gases, while a decrease in the volume of micropores is
observed at higher temperatures. However, an increase in mesopores and macropores volume and area is observed when increasing
the pyrolysis or calcination temperature, and, at 900 C, the mesoand macropore volume contribute in a high degree to the BET
surface area. On the other hand, the area due to the micropores
diminishes when heating from 700 to 900 C, and this result may
be related to the chemistry of the calcination process and due to

the sintering process as previously mentioned. However, taking

into account that the total pore volume generally increases as
the heating temperature rises, the micropore structure is likely
responsible for retaining the H2S molecules. Consequently, the
BET surface area of interest in that associated to the micropores,
and not the total value due to all of the pore structure. Therefore,
it is more important the t-Plot micropore area than the t-Plot
external surface area in the computation of BET surface area.
The growth of micropores and mesopores is observed with
increasing the temperature of the thermal treatment. All of the
sludge-derived samples have a signicant progress of mesoporosity because of the presence of inorganic matter. Probably, after H2S
adsorption, almost all of the micropores disappear. Unfortunately,
the analyzer used for the characterization does not tolerate sulfur
in the sampled gas, and this hypothesis could not be veried.
3.3. Further characterization of the adsorbents
X-ray diffraction (XRD) peaks clearly attributed to sulfur crystalline were not obtained. Fig. 6 depicts the XRD diffraction patterns for the precursor LG and the treatment PA, at the three
heating temperatures tested (500 C, 700 C and 900 C), including
an analysis of the crystalline phases. Calcite (CaCO3) in green,
quartz (SiO2) in red, and dolomite (CaMg(CO3)2) in blue, and phylosilicate, such as muscovite in pink, were identied at 500 C. At
700 C, a new phosphate phase was found containing Ca, Mg and
Na in blue (instead of dolomite), identied as merillite. Finally, at
900 C, it was identied anorthite (CaAl2SiO8) in orange, iron oxide
(Fe5.34O8) in crimson and calcium sulde phosphate (Ca10(PO4)6S).
As in previous studies [4,14], most of the deposited sulfur is in an
amorphous state: 45, 47 and 39 wt.% when heating to 500, 700 and
900 C, respectively. In addition, for the precursor LL (not shown),
calcite and dolomite were not detected at any temperature. At 700

207

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213
Table 5
BET surface area and pore structure for some samples of the treated precursors LL and LF.

LL 700 PA

LL 900 PA

LF 900 PA

13.363
5.066
8.297
0.015
0.003
0.011
3.42

94.384
16.857
77.527
0.051
0.031
0.020
4.93

79.870
15.089
64.779
0.044
0.026
0.018
4.71

122.040
56.200
65.840
0.093
0.027
0.066
4.19

126.090
67.900
58.190
0.094
0.024
0.069
4.12

(b)

3000

(a)

2000

Counts
1000

300

600

Counts

LL 700 A

900 1200 1500 1800 2100 2400

3000

BET surface area (m /g)

t-Plot external surface area (m2/g)
t-Plot micropore area (m2/g)
Pore volume (cm3/g)
Micropore volume (cm3/g)
Meso- and macropore volume (cm3/g)
Adsorption capacity (mg H2S/g ads)

LL 500 PA

10

20

30

40

50

10

60

20

2Theta (Coupled TwoTheta/Theta) WL=1,54060

30

40

50

60

3000
2000
1000

Counts

900 1200 1500 1800 2100 2400

300

600

Counts

3000

2Theta (Coupled TwoTheta/Theta) WL=1,54060

10

20

30

40

50

60

10

20

Counts

30

40

50

60

2Theta (Coupled TwoTheta/Theta) WL=1,54060

(c)

300

600

900 1200 1500 1800 2100 2400 2700

2Theta (Coupled TwoTheta/Theta) WL=1,54060

10

20

30

40

50

60

Counts

300

600

900 1200 1500 1800 2100 2400 2700

2Theta (Coupled TwoTheta/Theta) WL=1,54060

10

20

30

40

50

60

2Theta (Coupled TwoTheta/Theta) WL=1,54060

Fig. 6. XRD of adsorbents obtained from the precursor LG.

and 900 C a crystalline phase of phosphate with calcium and aluminum was measured (Ca9Al(PO4)7).
Fig. 7 shows the SEM images at a magnication so as to get
micrographs with a scale of 2000, 500, 50 and 20 lm for precursors

coded as LG 700 PAG and LL 700 PAG after the adsorption of hydrogen sulde. The nal G denotes that the sample corresponds to a
spent adsorbent. The images magnied at 2000 lm depict a more
irregular shape of LG precursor particles, with many grooves, while

208

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

Fig. 7. SEM micrographs of the adsorbents for LG and LL at 700 C and treatment PA, once used for H2S adsorption.

the LL precursor particles are spherical, more compact. At 500 lm,

it can be observed that the porosity (meso and macropore) of LG
700 PAG sample is higher than that of LL 700 PAG, as above-mentioned (Tables 4 and 5). In addition, the LG 700 PAG sample presents numerous laments, probably from algae and diatoms (see
Fig. 7 magnied at 20 lm), which provide calcium and phosphorous to the sample. At 20 lm, it can be observed a prismatic deposit of sulfur in the precursor LL. However, it is very interesting to
note that, in general, the sulfur distribution on the particle surface
was very concentrated for precursor LL, probably where active centers oxidize H2S to elemental sulfur, while for precursor LG the sulfur was located over the entire surface since the presence of active
reagents is uniform throughout the pellet. This was observed by
SEMEDS, which made it possible to analyze the contents of the
elements (Fig. 8). For precursor LL, Fig. 8a and b depict EDS analyses of a location near the sulfur deposit and another on the own deposit. On the sulfur deposit, very high concentrations of calcium
and magnesium were measured as well as a high sulfur content
(10.08 wt.%). The catalytic effect of calcium and magnesium oxides
dispersed within the carbon matrix makes it possible that

hydrogen sulde dissociates on the oxide active centers where it

is oxidized to elemental sulfur [15]. Near the sulfur deposit, the
sulfur concentration was 2.54 wt.%. In other samples analyzed, sulfur was not detected. However, for the precursor LG, sulfur was
mostly measured everywhere in the sample, with a sulfur concentration ranged from 2 to 4 wt.% (Fig. 8c and d). Additionally, the
iron concentration on the precursor LG was about 8 wt.%. This
may be very important, as explained in the next subsection.
In addition to the elemental sulfur detected by EDS, iron sulfate
was identied by XPS analysis, especially for precursor LG, which
contains a higher iron amount. Fig. 9 illustrates the general spectrum corresponding to the XPS analyses for LG 700 PAG and LG
500 PAG. First, it was detected carbon, nitrogen, oxygen, phosphorus, magnesium, calcium, silicon, aluminum, iron and sulfur. Sulfur
peak that appears in each spectrum is S (2p), because it is the most
intense photoelectron peak, while the peak S (2s) was not appreciated. On each general spectrum of Fig. 9, there is a record of the
sulfur spectrum in order to identify the oxidation states based on
the energy peak shift relative to the elemental sulfur peak, once
identied the other compounds in the general spectrum. For the

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

209

Fig. 8. EDS analysis of the adsorbents for LG and LL at 700 C and treatment PA, once used for H2S adsorption.

LG 700PAG and LG 500 PAG, the rst peak was observed at a binding energy of 169.7 eV and 169.6 eV, respectively. This veries the
sulfate existence, which could be iron sulfate as iron was detected
in the general spectrum. The second peak occurs at a binding energy of 164.2 eV, which corresponds to elemental sulfur. This peak
was not obtained for the LG 500 PAG sample.
3.4. Possible reactions involved in the adsorption process
If iron oxide is present, the following reactions may take place
[16]:

Fe3+ complex ions would shift reaction (4) to the right side [17].
Furthermore, the phosphoric acid is a well-known chelating agent,
with a complexing action. Taking into account the relative high proportion of phosphorous in the tested samples, phosphoric acid may
be probably formed during the adsorption process.
Additionally, other metals such as copper and zinc can act as
catalyst in a second stage, once sulfate is formed, based on the following reactions:

Me2 H2 S 2SO2
4 ! MeSs 2HSO4

Fe2 O3 3H2 S ! Fe2 S3 3H2 O

MeSs 2Fe3 ! Me2 2Fe2 S0

Fe2 S3 3=2O2 ! Fe2 O3 3S0

2Fe2 1=2O2 2HSO4 ! 2Fe3 2SO2

4 H2 O

In any case, it should be noted that the iron acts as a catalyst to

boost the nal hydrogen sulde removal converting it into elemental sulfur, which is stored in micropores. However, the absence of
oxygen hinders the regeneration of Fe3+. On the evidence of Fe3+
presence, a record of the iron spectrum was not performed. However, it is well known and widely accepted that exposure to air will
oxidize most of the surface iron atoms to Fe (III). Since the last step
in the thermal treatment of precursors was performed using air, it
is surely the presence of Fe3+.
On the other hand, possible redox-reactions involved in the process are the following:

2Fe3 H2 S 2Fe2 S0 2H

2Fe2 1=2O2 H2 O 2Fe3 2OH

One basic point is the low solubility of the intermediate metallic

suldes in acid media, which is a characteristic of Cu and Zn [18].
However, reactions (4)(8) are limited by the presence of a liquid phase (water) on the adsorbent surface. In the absence of
added water to the biogas or the precursor, water may be adsorbed
from the environment, once disposed in the bed, or coming from
the reaction (2). In any case, it does not seem very probable that
these redox-reactions can take place in a large extent. Therefore,
the adsorption capacity of the precursor is quite limited in this way.
In the above reactions, the absence of oxygen avoids the re-oxidation of Fe2+ to Fe3+ (reactions 5 and 8) so the enhancement of the
adsorption capacity lowers because the catalytic activity of some
metals is limited (reaction 3). As the hydrogen sulde diffuses into
the pellet, the active reagents begin to be depleted, so mainly the
physical adsorption takes place but at a lower rate [19].

210

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

40k

169.6

LG 500 PA
Intensity CPS

1350

35k
Mg(1s)

30k

1250

1200

C(AES)

Intensity CPS

1300

25k
Ca(AES)

SULFUR
SPECTRUM

1150

O(AES)

180

178

176

174

172

170

168

166

164

P(2s)
S(2p)

162

Binding Energy eV (-5.47)

O(1s)

20k

Si(2s)

N(AES)

Mg(AES)

Fe(AES)

P(2p)
Al(2s)

N(1s)

15k

Si(2p)

Fe(2p)
Ca(2s)

Ca(2p)

Mg(2s)
Al(2p)

10k

C(1s)

5k
O(2s)

0k
200

400

600

800

1000

1200

1400

Kinetic Energy eV
Mg(1s)

60k

169.7

1300

SULFUR
SPECTRUM

1250

Intensity CPS

LG 700 PA
50k

164.2

1200

1150

1100

Intensity CPS

40k
1050
180

178

176

174

172

170

168

166

164

162

P(2s)
S(2p)

Binding Energy eV (-4.13)

C(AES)

30k

Si(2s)

Ca(AES)

P(2p)

O(AES)
Fe(AES)

N(AES)

Mg(AES)

O(1s)

Al(2s)
Si(2p)

N(1s)

20k

C(1s)

Fe(2p)
Ca(2s)

Mg(2s)

Ca(2p)

Al(2p)

10k
O(2s)

0
200

400

600

800

1000

1200

1400

Kinetic Energy eV
Fig. 9. XPS analysis (general spectrum with embedded sulfur spectrum) of the precursor LG heated at 500 and 700 C, following the treatment PA.

Finally, the presence of alkali metals or alkaline earth metals

(Na, K, Ca and Mg) makes it possible to increase the H2S removal,
because the metal hydroxide reacts with the hydrogen sulde to
obtain hydrogen-sulde ions, which could be oxidized to elemental sulfur if oxygen was present, as described below.
Taking into account that the adsorbent bed continuously stores
hydrogen sulde from zero to saturation, the basicity of the adsorbent will change from moderately high values, corresponding to a
surface pH higher than 7.0, to low values. Therefore, probably, the
rst chemical adsorption that will probably take place is that
involving a process where elemental sulfur can be formed. Then,
when pH falls, the H2S removal leads to sulfuric acid, which may
plug the pores, but also provide an acid medium to enhance the
catalytic effect of other metals like Cu and Zn. Indeed, if sulfuric
acid is mainly formed, it may participate in a redox reaction to produce elemental sulfur.
3.5. Inuence of the relative humidity
Adsorption of hydrogen sulde in the presence of moisture was
also studied by obtaining the corresponding breakthrough curves.
Several tests were carried out, but the results did not bring out a

clear improvement. Fig. 10 shows the breakthrough curves obtained using saturated biogas. Results obtained by pre-humidifying
the adsorbent bed were similar, so they are not shown. No clear effect of the experimental conditions was noticed and the adsorption
capacities measured using dry and saturated biogas were almost
the same. A slightly better performance of the precursor in dry conditions may suggest that water contributes to deactivation of surface-active centers in the H2S retention, counteracting a clearer
positive effect. Probably, it is because the inorganic phase of the
sewage-sludge contains alkaline earth oxides that are nonreactive
with water, as pointed out by other researchers [20]. In addition,
water in the gas mixture hinders the adsorption capacity because
it plugs the pore structure. In any case, this nding agrees with a
result encountered in a previous study [12]; some other researchers found a positive effect (the breakthrough time increases as the
humidity increases), although in studies carried out using a mixture of air and H2S instead of a simulated or real biogas [9]. Finally,
the water present in the biogas is adsorbed on the adsorbent particles, providing a liquid phase where CO2 may be dissolved, forming carbonates and contributing to the formation of sulfurous acid,
which deactivates the basic catalytic sites, resulting in a lower
capacity.

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

211

(a)

(b)

(c)

Fig. 10. Breakthrough curves for the precursors thermally treated at three temperatures in the absence of oxygen: precursor (a) LF, (b) LL, and (c) LG.

The use of air may have an important effect due to the oxygen
content, as next described.
3.6. Inuence of the oxygen presence
To test the effect of the oxygen on the H2S removal, oxygen was
added to the dry biogas stream before entering the xed bed. For

safety reasons, the oxygen content was limited to 5 vol.%, due to

the methane presence. Compared to the results obtained in the absence of oxygen, the adsorption capacity of precursor LG is doubled
at 500 C and at 700 C, using the treatment PA, and tripled at
900 C, in the presence of oxygen. The activated carbons CAA and
CAT were also tested under these conditions and their adsorption
capacities increased from 12.8 to 77.4 mg/g, and from 4.4 to

212

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213

7.3 mg/g, respectively. This signicant increase in the adsorption

capacity was also detected for impregnated active carbons [21].
The adsorption capacity of the LG precursor has a moderately
adsorption capacity higher than that of CAT activated carbon. However, the CAA adsorption capacity is much higher than the best
adsorbent obtained (17.3 mg/g for LG 700 PA), in the presence of
oxygen. The oxygen may oxidize the hydrogen sulde to elemental
sulfur that would be physically adsorbed in micropores zone, xing
thus the sulfur to the adsorbent.
The primary dehydroxylation of inorganic species is expected to
occur at ca. 400 C [11], resulting in an increase in the basicity of
precursors. CAA activated carbon was impregnated with a solution
of NaOH, and the results showed the H2S breakthrough capacity increases six times. However, neither CAT activated carbon nor the
three kind of sewage sludge were impregnated with a basic solution, and the increase in the adsorption capacity was clearly lower
(somewhat less than double). This may indicate that the amount of
OH present on the surface is a factor that favors the adsorption
capacity. By increasing the basicity of the precursor, hydroxide ions
cause an increase in the HS- ion concentration. These ions can be
further oxidized to elemental sulfur, until the basic environment
is no longer maintained.
Therefore, possible reactions involved when oxygen is present
may be the following:

H2 S OH ! H2 O HS

HS 1=2O2 ! S0 OH

10

3.7. Inuence of a chemical impregnation

Chemical activation uses as agents mainly alkali and alkaline
earth metals, such as potassium or sodium hydroxide or carbonate
as well as zinc, aluminum or magnesium chlorides. It is published
that impregnated activated carbons remove gaseous pollutants via
an irreversible reaction between the additives and pollutants. A basic media is one of the required conditions for an efcient desulfurization process [22]. As it can be observed in Table 2, thermal
treatment increases the alkali and alkaline earth metals content
and, hence, the basicity of the precursors, leading to a higher
adsorption capacity. Most of the previous studies showed that
impregnated activated carbons have a very positive effect on the
desulfurization performance, increasing the adsorption capacity
of the adsorbent considerably [21,23]. However, when carrying
out different procedures of chemical activation and testing the
so-activated pretreated sludge, a signicant improvement was
not appreciated, i.e., the effect of the impregnation was not clearly
detected, probably to an inefcient impregnation procedure. Effectively, two different procedures of impregnation were applied to
the virgin precursors with and without performing thermal treatment. The adsorption capacity of adsorbents after atomizing a basic solution was lower than that obtained without impregnation.
The best performance was obtained by milling and mixing NaOH
with the precursor and after heating to 700 C, using the treatment
PA and an agent-to-precursor mass ratio of 0.25. Anyway, the
improvement was also small compared to the case in which no
impregnation was used.
4. Conclusions
Carbon-based adsorbents from thermally treated sewage sludge
can be used along with activated carbon to boost the removal efciency of the H2S present in a biogas, providing a longer operational life for the activated carbon and making the purication
process more cost-effective. Experimentally, different thermal

treatments, heating temperature and adsorption conditions were

tested, and the adsorption dynamics of the prepared adsorbents
were obtained in a xed-bed column, using three precursors from
sewage sludge (coded as LL, LF, LG). The physical and chemical
properties were widely studied to acquire a better knowledge of
the adsorbent behavior. The BET surface area of interest is that
associated to the micropores, and not the total value due to all of
the pore structure. Indeed, the development of porosity and formation of new mineral-like catalytic components depend on the thermal treatment for a given precursor. By this means, it was
identied the best precursor, which was that obtained from sewage-sludge LG pyrolyzed up to 500 C and then calcined up to
700 C (PA thermal treatment). The adsorption capacities of the
three kinds of sewage-sludge thermally treated were similar and
even higher than that of the non-impregnated commercial activated carbon. Precursors LL and LG showed similar capacities and
the best one was the precursor LG, probably due to surface characteristics, especially the micropore volume and its corresponding
surface area, as well as the content in Fe, Ca and Mg, which may
boost the chemical xation of H2S as elemental sulfur.
Additionally, a pyrolysis treatment is not necessary, contrary to
that mainly pointed out in the literature, since using air as an agent
in the thermal treatment allows high adsorption capacities near
those obtained by pyrolysis. Likewise, the presence of oxygen
(but limited for safety) in the biogas entering the adsorber will improve the process performance. The relative humidity did not enhance the adsorption efciency, although more experiments
should be carried out, using a biogas with oxygen. The presence
of alkali metals and iron seems to be a positive effect. Finally, a better impregnation procedure of the adsorbents should be investigated, since the two procedures tested did not increase the
adsorption capacity, contrary to that expected.
Acknowledgments
Authors thank to the Research, Technology and Investigation
Centre of the University of Seville (CITIUS) the aid given in the
analyses performed by the Services of Microanalysis, Microscopy,
X-Ray, XPS/ESCA, and Functional Characterization.
References
[1] P. Cosoli, M. Ferrone, S. Pricl, M. Fermeglia, Hydrogen sulphide removal from
biogas by zeolite adsorption. Part I. GCMC molecular simulations, Chem. Eng. J.
145 (2008) 8692.
[2] G. Garca, E. Cascarosa, J. Abrego, A. Gonzalo, J.L. Snchez, Use of different
residues for high temperature desulphurisation of gasication gas, Chem. Eng.
J. 174 (2011) 644651.
[3] R. Zhang, R.C. Brown, A. Suby, K. Cummer, Catalytic destruction of tar in
biomass derived producer gas, Energy Convers. Manage. 45 (2004) 9951014.
[4] Y. Xiao, S. Wang, D. Wu, Q. Yuan, Catalytic oxidation of hydrogen sulde over
unmodied and impregnated activated carbon, Sep. Purif. Technol. 59 (2008)
326332.
[5] A. Mndez, G. Gasc, M.M.A. Freitas, G. Siebiele, T. Stuczynski, J.L. Figueiredo,
Preparation of carbon-based adsorbents from pyrolysis and air activation of
sewage sludges, Chem. Eng. J. 108 (2005) 169177.
[6] M.J. Martin, E. Serra, A. Ros, M.D. Balaguer, M. Rigola, Carbonaceous adsorbents
from sewage sludge and their application in a combined activated sludgepowdered activated carbon (AS-PAC) treatment, Carbon 42 (2004) 13891394.
[7] A. Ros, M.A. Lillo-Rdenas, E. Fuente, M.A. Montes-Morn, M.J. Martn, A.
Linares-Solano, Dried sludges and sludge-based chars for H2S removal at low
temperature: inuence of sewage sludge characteristics, Environ. Sci. Technol.
40 (2006) 302309.
[8] R. Sitthikhankaew, S. Predapitakkun, R. Kiattikomol, S. Pumhiran, S.
Assabumrungrat, N. Laosiripojana, Comparative study of hydrogen sulde
adsorption by using alkaline impregnated activated carbons for hot fuel gas
purication, Energy Procedia 9 (2011) 1524.
[9] T.J. Bandosz, Carbonaceous materials as desulfurization media, in: J.M.
Loureiro, M.T. Kartel (Eds.), Combined and Hybrid Adsorbents, Springer,
2006, pp. 145164.
[10] H. Katoh, I. Kuniyoshi, M. Hirai, M. Shoda, Studies of the oxidation mechanism
of sulphur-containing gases on wet activated carbon bre, Appl. Catal. B
Environ. 6 (1995) 255262.

F.J. Gutirrez Ortiz et al. / Separation and Purication Technology 123 (2014) 200213
[11] A. Bagreev, T.J. Bandosz, D.C. Locke, Pore structure and surface chemistry of
adsorbents obtained by pyrolysis of sewage sludge-derived fertilizer, Carbon
39 (13) (2001) 19711979.
[12] W. Yuan, T.J. Bandosz, Removal of hydrogen sulde from biogas on sludgederived adsorbents, Fuel 86 (2007) 27362746.
[13] T.J. Bandosz, K. Block, Effect of pyrolysis temperature and time on catalytic
performance of sewage sludge/industrial sludge-based composite adsorbents,
Appl. Catal. B Environ. 67 (2006) 7785.
[14] O.C. Cariaso, P.L. Walker, Oxidation of hydrogen sulde over microporous
carbons, Carbon 13 (1975) 233239.
[15] A. Bagreev, T.J. Bandosz, On the mechanism of hydrogen sulphide adsorption/
oxidation on catalytic carbons, Ind. Eng. Chem. Res. 44 (2005) 530538.
[16] A. Kohl, R. Neilsen, Gas Purication, fth ed., Gulf Professional Publ., Houston,
1997.
[17] M.S. Horikawa, F. Rossi, M.L. Gimenes, C.M.M. Costa, M.G.C. Silva, Chemical
absorption of H2S for biogas purication, Braz. J. Chem. Eng. 21 (2004) 415
422.
[18] H. Maat, J.A. Hogendoorn, G.F. Versteeg, The removal of hydrogen sulde from
gas streams using an aqueous metal sulfate absorbent. Part I The absorption

[19]
[20]

[21]

[22]

[23]

213

of hydrogen sulde in metal sulfate solutions, Sep. Purif. Technol. 43 (2005)

183197.
S.G. Chatterjee, C. Tien, Adsorption with chemical reaction in a single
modied-carbon pellet, Chem. Eng. Sci. 44 (1989) 22832294.
M. Seredych, T.J. Bandosz, Desulfurization of digester gas on catalytic
carbonaceous adsorbents: complexity of interactions between the surface
and components of the gaseous mixture, Ind. Chem. Eng. Res. 45 (2006) 3658
3665.
D.Y. Choi, J.W. Lee, S.C. Jang, B.S. Ahn, D.K. Choi, Adsorption dynamics of
hydrogen sulde in impregnated activated carbon bed, Adsorption 14 (2008)
533538.
A. Bagreev, F. Adib, T.J. Bandosz, PH of the activated carbon surface as an
indication for its suitability for removal of hydrogen sulphide from wet air
streams, Carbon 39 (2001) 19051987.
Y. Xiao, S. Wang, D. Wu, Q. Yuan, Experimental and simulation study of
hydrogen sulde adsorption on impregnated activated carbon under anaerobic
conditions, J. Hazard. Mater. 153 (2007) 11931200.