You are on page 1of 46

Atlas of Microstructures

AF234/7226/MSch
Prepared by: Ellen Berghof-Hasselbcher 1) Peter Gawenda 1) Monika Schorr 1) Michael Schtze 1) John J. Homan 2)

1) Karl-Winnacker-Institut der DECHEMA e.V. Theodor-Heuss-Allee 25 60486 Frankfurt am Main Germany 2) Air Products and Chemicals, Inc. 7201 Hamilton Boulevard Allentown, PA 18195-1501 USA

The results, conclusions and recommendations given in this report refer to the specimens and data submitted as well as to the system conditions mentioned. Therefore, the results cannot be applied to other conditions. DECHEMA e.V. accepts no liability. It is only permitted to reproduce this report unabridgedly. Publication of the report or of single experimental results are subject to approval by DECHEMA e.V.

Atlas of Microstructures

ii

Atlas of Microstructures

All rights reserved, including translations ISBN 978-1-57698-046-0 No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the publisher. This document was prepared under the sponsorship of the Materials Technology Institute Inc. (MTI) and is approved for release. All data and information contained in this document are believed to be reliable; however, no warranty of any kind, express or implied, is made with respect to the data, analyses, or author of this document; and the use of any part of the document is at the user's sole risk. MTI, the author, or any person acting on its behalf, assume no liability and expressly disclaim liability, including without limitation liability for negligence, resulting from the use or publication of the information contained in this document to warrant that such use or publication will be free from privately owned rights. Published by Materials Technology Institute www.mti-global.org

iii

Table of Contents
Introduction . .......................................................................................................................... xx Description of Test Specimens................................................................................................. xx Creep rupture test specimens........................................................................................................ xx
Alloy 35/45 Exposure Conditions. ..................................................... Table 1 Chemistry of Alloy 35/45 Creep Rupture Stress Samples................... Table 2 Alloy HPMA Exposure Conditions................................................... Table 3 Chemistry of Alloy HPMA Creep Rupture Stress Samples. ................ Table 4 Alloy HP Modied Exposure Conditions.......................................... Table 5 Chemistry of Alloy HP Modied Creep Rupture Stress Samples. ....... Table 6

Service-exposed specimens............................................................................................................ xx
- History of Service Exposed Samples - Exposure Conditions of .................................................................... Table 7 - Chemistry of Service Exposed Samples............................................... Table 8

Sample Preparation and Microstructural Characterization....................................................... xx Metallographic Preparation...................................................................................................... xx Investigation Techniques.............................................................................................................. xx


- Dierential Interference Contrast Technique (LM-DIC) - Scanning Electron Microscopy (SEM-BSE) - Electron Probe Microanalysis (EPMA) - Interference Layer Metallography (LM-ZnSe) - Image Analysis by the False Color Technique (LM-FC) - Development of the Microstructure by Etching (LM-etched)
iv

Results of the Microstructural Analyses.................................................................................... xx Alloy 35/45................................................................................................................................ xx


- Results of image analysis of Alloy 35/45............................................. Table 9 - As Cast. ............................................................................................. Fig. 1-1a-j - Creep Rupture Samples..................................................................... Fig. 1-2a to 1-2j ..................................................................... Fig. 1-3a to 1-3r ..................................................................... Fig. 1-4a to 1-4r ..................................................................... Fig. 1-5a to 1-5r ..................................................................... Fig. 1-6a to 1-6z ..................................................................... Fig. 1-7a to 1-7r ..................................................................... Fig. 1-8a to 1-8r ..................................................................... Fig. 1-9a to 1-9z ..................................................................... Fig. 1-10a to 1-10r ..................................................................... Fig. 1-11a to 1-11r ..................................................................... Fig. 1-12a to 1-12z ..................................................................... Fig. 1-13a to 1-13v - Conclusions. ........................................................................................................................ xx

Alloy HPMA.............................................................................................................................. xx
- Results of image analysis of Alloy HPMA..............................Table 10 - As Cast. ............................................................................................. Fig: 2-1a to 2-1j - Creep Rupture Samples..................................................................... Fig. 2-2a to 2-2r ..................................................................... Fig. 2-3a to 2-3r ..................................................................... Fig. 2-4a to 2-4r


Creep Rupture Samples (continued)....................................................... Fig. 2-5a to 2-5r ..................................................................... Fig. 2-6a to 2-6r ..................................................................... Fig. 2-7a to 2-7z ..................................................................... Fig. 2-8a to 2-8r ..................................................................... Fig. 2-9a to 2-9r ..................................................................... Fig. 2-10a to 2-10r ..................................................................... Fig. 2-11a to 2-11r ..................................................................... Fig. 2-12a to 2-12r ..................................................................... Fig. 2-13a to 2-13r ..................................................................... Fig. 2-14a to 2-14r ..................................................................... Fig. 2-15a to 2-15r ..................................................................... Fig. 2-16a to 2-16r ..................................................................... Fig. 2-17a to 2-17s - Conclusions. ........................................................................................................................ xx

Alloy HP Modied...................................................................................................................... xx
- Results of image analysis of Alloy HP Modied.................................. Table 11 - As Cast . ............................................................................................ Fig. 3-1a to 3-1j - Creep Rupture Samples .................................................................... Fig. 3-2a to 3-2r ..................................................................... Fig. 3-2a to 3-2r ..................................................................... Fig. 3-3a to 3-3r ..................................................................... Fig. 3-4a to 3-4r ..................................................................... Fig. 3-5a to 3-5r ..................................................................... Fig. 3-6a to 3-6r ..................................................................... Fig. 3-7a to 3-7r
vi

- Creep Rupture Samples (continued)....................................................... Fig. 3-8a to 3-8r ..................................................................... Fig. 3-9a to 3-9r ..................................................................... Fig. 3-10a to 3-10r ..................................................................... Fig. 3-11a to 3-11r ..................................................................... Fig. 3-12a to 3-12r ..................................................................... Fig. 3-13a to 3-13r - Conclusions. ........................................................................................................................ xx

Service Exposed Samples.............................................................................................................. xx


- Results of image analysis of Service Exposed Samples ........................ Table 12 - Alloy HPMA.................................................................................... Fig. 2-SE-Fa to 2-SE-Fr ..................................................................... Fig. 2-SE-Ga to 2-SE-Gr - Alloy HP Modied ........................................................................... Fig. 3-SE-Ba to 3-SE-Br ..................................................................... Fig. 3-SE-Aa to 3-SE-Ar ..................................................................... Fig. 3-SE-Ca to 3-SE-Cr ..................................................................... Fig. 3-SE-Da to 3-SE-Dr ..................................................................... Fig. 3-SE-Ha to 3-SE-Hr ..................................................................... Fig. 3-SE-Ia to 3-SE-Ir ..................................................................... Fig. 3-SE-Ja to 3-SE-Jr ..................................................................... Fig. 3-SE-Ea to 3-SE-Ez - Conclusions. ........................................................................................................................ xx

Summary................................................................................................................................. xx Composition of the Precipitates. ................................................................................................... xx


Average Compositions of the Precipitates in atomic-%. ......................... Table 13
vii

Precipitation Kinetics Diagrams............................................................................................... xx


- Alloy 35/45 . .................................................................................... Fig. 1-PKD-a to 1-PKD-d ..................................................................... Fig. 1-PKD-e - Alloy HPMA ................................................................................... Fig. 2-PKD-a to 2-PKD-c ..................................................................... Fig. 2-PKD-d - Alloy HP Modied . .................................................................. Fig. 3-PKD-a to 3-PKD-d ..................................................................... Fig. 3-PKD-e

viii

Introduction
Up to and through the 1940s, furnace tubes were typically fabricated from wrought chromium steels and/or austenitic stainless steel alloys. Having rather low carbon concentrations, these alloys had poor creep strength and limited service lives. To increase tube life, greater carbon concentrations were required to promote precipitation hardening upon elevated temperature exposure. The increased carbon concentrations, however, generally resulted in excessive work hardening during conventional processing of wrought materials. The use of centrifugal casting, pouring molten metal in a horizontal spinning mold, allowed founders to develop high carbon alloys as the molten metal solidied into near nal shape without the need for subsequent metal working operations. Thus, renement of centrifugal casting processes was the gateway to further alloy development and more aggressive furnace operations. The rst widely used centrifugal cast alloy for steam-methane reformer tube applications was HK40 in the 1950s timeframe. HK40 is essentially the cast equivalent to wrought 310 stainless steel nominally containing 25 wt% chromium, 20 wt% nickel, with iron as the balance. However, HK40 nominally contains 0.40 wt% carbon while wrought 310 stainless steel contains only 0.08 wt% carbon. The increased carbon content and precipitation of primary carbides resulted in HK40 having greatly improved high temperature strength as compared to wrought 310 stainless steel. In the 1960s, the cast HP alloys (nominally 25 wt% chromium, 35 wt% nickel, 0.50 wt% carbon, with iron as the balance) were developed to provide greater creep strength as compared to HK40. The HK and HP alloys rely on precipitation of M23C6 and/or M7C3 carbides (where M is primarily chromium) for elevated temperature creep strength. The precipitated chromium carbides in the HK and HP alloys tended to coalesce as exposure temperatures approached 1800oF (982oC). The carbide coalescence decreased the creep strength and, therefore, limited the strength of these alloys at elevated temperatures. Microstructural changes that occur in the HK and HP alloys with

extended aging time and temperature have been well documented by Battelle Columbus Laboratories1. The Battelle data has proven to be a valuable resource in estimating reformer tube exposure conditions associated with reformer tube failure analyses and remaining life assessments. User demand for higher temperature/stronger alloys fueled continued alloy development resulting in the introduction of the HP-modied alloy in the 1970s. The HP-modied alloy had the same nominal chemistry of the HP alloy along with the addition of typically 1 wt% niobium. The niobium addition results in precipitation of M23C6, M7C3, and MC type primary carbides upon solidication. In the M23C6 and M7C3 carbides, niobium substitutes for some of the chromium with the complex niobium-chromium carbides being more stable at elevated temperatures as compared to chromium carbides. In the HP-modied alloy, niobium is the primary carbide forming element in the MC type carbides. In the 1980s, the demand for more severe design conditions and/or design lives in excess of 100,000 hours led to the introduction of the HP-micro-alloyed material. HP-micro-alloyed (or commonly designated as HPMA) material was based on the HP-modied chemistry with micro additions of alloying elements. In general, micro-alloying refers to intentional alloying additions at a concentration of 0.10 wt% or less. Titanium is the most common micro-alloying addition with other micro-alloying additions including zirconium, tantalum, or tungsten. Rare earth elements such as lanthanum, cerium, and/or yttrium may also be added. The micro-alloying additions provide a ne dispersion of MC type carbides that are stable at temperatures well in excess of 2,000oF (1093oC). In the 1990s, the 35Cr/45Ni alloy family grew in popularity. The 35Cr/45Ni alloy has similar creep strength to the HPMA alloy but with notably improved carburization resistance. The excellent carburization resistance made the 35Cr/45Ni alloy well suited for ethylene pyrolysis furnace tubes. As outlined above, alloy development has continued in centrifugally cast heat resistant alloys allowing users to design for and operate at more

Atlas of Microstructures

severe conditions. Unlike the HK40 and HP50 alloys, there is virtually no published data on the microstructural changes that occur in the HP-modied, HPMA, or 35Cr/45Ni alloys upon long term aging. In an eort to address industry need, the Materials Technology Institute (MTI) sponsored the Atlas of Microstructures project. The specimens analyzed in the Atlas of Microstructures were foundry stress rupture specimens generously donated by MetalTek International, Duraloy Technologies, Kubota Metal Corp., and Manoir Industries. In addition, service exposed samples, having longer aging times than the foundry stress rupture specimens, were donated by operating companies including Air Products & Chemicals, Syncrude Canada, Eastman Chemical, and MetalTek International. The thorough, detailed microstructural analyses were completed at the Karl-Winnacker Institut der DECHEMA in Frankfurt, Germany. The Atlas of Microstructures documents microstructural changes with increased aging time and temperature, identication and chemical composition of precipitated phases, as well as diagrams characterizing the kinetics of phase transformation for the HP-modied, HPMA, and 35Cr/45Ni alloy classes. This MTI Atlas of Microstructures bridges the gap from the Battelle project from the early 1970s to the most important alloys used in the petrochemical industry today.

Description of the Test Specimens


Creep rupture test specimens The specimens that had been taken from creep rupture tests can be summarized with regards to their exposure conditions and their chemistry in Tables 1-6. For each of the three materials investigated an as-delivered (as-cast) specimen as reference sample was available. For alloy 35Cr/45Ni and alloy HP Modied twelve creep specimens had been supplied for the investigations. For alloy HPMA sixteen creep specimens existed. The test temperatures varied between 1675 and 2100F (913 and 1149C), the maximum exposure times reached 12,289 h. The stress range was between 1.0 and 8.5 ksi (6.9 and 58.8 MPa). The scatter of the chemical compositions was in the allowed range for these materials.

continued on page 8

Atlas of Microstructures

Table 1: Alloy 35/45 Exposure Conditions Sample_ID Dechema Sample MTI A234_003 35/45-1 (c) A234_004 35/45-2 A234_005 35/45-3 A234_006 35/45-4 A234_007 35/45-5 A234_008 35/45-6 A234_009 35/45-7 A234_010 35/45-8 A234_011 35/45-9 A234_012 35/45-10 A234_013 35/45-11 A234_014 35/45-12 A234_015 35/45-13

Test-Temp F as cast 1800 1800 1922 1900 1900 1900 2012 2000 2012 2057 2050 2100

Test-Temp C as cast 982 982 1050 1038 1038 1038 1100 1093 1100 1125 1121 1149

Stress Ksi 3.30 2.90 3.40 2.09 1.60 1.54 2.00 1.80 1.15 1.60 1.00 1.50

Stress MPa 22.8 20.0 23.4 14.4 11.0 10.6 13.8 12.4 7.9 11.1 6.9 10.3

Expo. Time hrs 1,509 7,217 367 2,023 6,331 10,247 335 1,297 3,545 277 2,606 248

Table 2: Chemistry of Alloy 35Cr/45Ni Stress Rupture Samples Sample_ID Sample Dechema MTI Cr Ni Fe A234_003 35/45-1 (c) 35.85 44.59 bal A234_004 35/45-2 35.12 43.69 bal A234_005 35/45-3 34.84 44.43 bal A234_006 35/45-4 32.30 43.40 bal A234_007 35/45-5 32.39 44.77 bal A234_008 35/45-6 34.34 45.42 bal A234_009 35/45-7 34.41 45.37 bal A234_010 35/45-8 32.30 43.40 bal A234_011 35/45-9 32.65 44.61 bal A234_012 35/45-10 34.84 44.43 bal A234_013 35/45-11 32.30 43.40 bal A234_014 35/45-12 32.59 44.34 bal A234_015 35/45-13 34.84 44.43 bal

Mn 0.60 1.32 1.46 1.02 1.40 1.04 1.31 1.02 1.37 1.46 1.02 1.43 1.46

Composition (wt%) Si C 1.55 0.38 1.83 0.46 1.86 0.43 1.52 0.41 1.85 0.42 1.30 0.37 1.78 0.42 1.52 0.41 1.71 0.42 1.86 0.43 1.52 0.41 1.82 0.42 1.86 0.43

Nb 0.91 0.78 0.83 1.21 1.25 1.22 1.20 1.21 1.14 0.83 1.21 1.25 0.83

N 0.080 0.060 0.040 0.065 NR NR NR 0.065 NR 0.040 0.065 NR 0.040

Other Ti, Zr & W Ti, Zr & W Ti, Zr & W Ti Ti Ti Ti Ti, Zr & W Ti Ti, Zr & W

Atlas of Microstructures

Table 3: Alloy HPMA Exposure Conditions Sample_ID Dechema Sample MTI A234_018 HPMA-1 (c) A234_019 HPMA-2 A234_020 HPMA-3 A234_021 HPMA-4 A234_022 HPMA-5 A234_023 HPMA-6 A234_024 HPMA-7 A234_025 HPMA-8 A234_026 HPMA-9 A234_027 HPMA-10 A234_028 HPMA-11 A234_029 HPMA-12 A234_030 HPMA-13 A234_031 HPMA-14 A234_032 HPMA-15 A234_033 HPMA-16 A234_034 HPMA-17

Test-Temp. F as cast 1700 1700 1700 1750 1800 1800 1800 1832 1922 1950 1922 1922 2012 2012 2012 2012

Test-Temp. C as cast 927 927 927 954 982 982 982 1000 1050 1066 1050 1050 1100 1100 1100 1100

Stress Ksi 8.00 6.00 5.30 4.45 5.97 3.51 3.90 3.00 3.48 2.80 2.32 2.08 2.39 1.80 1.45 1.35

Stress MPa 55.2 41.3 36.6 34.1 41.2 24.2 26.9 20.7 24.0 19.1 16.0 14.3 16.5 12.4 10.0 9.3

Expo. Time hrs 183 2,338 6,072 8,359 177 2,436 6,478 12,289 137 2,558 4,864 11,778 105 2,714 5,580 8,990

Atlas of Microstructures

Table 4: Chemistry of Alloy HPMA Stress Rupture Samples Sample_ID Sample Dechema MTI Cr Ni Fe A234_018 HPMA-1 (c) 25.01 33.79 bal A234_019 HPMA-2 24.92 34,54 bal A234_020 HPMA-3 24.29 32.51 bal A234_021 HPMA-4 25.01 33.79 bal A234_022 HPMA-5 25.01 33.79 bal A234_023 HPMA-6 25.85 35.25 bal A234_024 HPMA-7 24.38 35.13 bal A234_025 HPMA-8 24.49 33.53 bal A234_026 HPMA-9 24.49 33.53 bal A234_027 HPMA-10 24.36 33.18 bal A234_028 HPMA-11 25.11 36.13 bal A234_029 HPMA-12 25.01 33.79 bal A234_030 HPMA-13 24.49 33.53 bal A234_031 HPMA-14 24.36 33.18 bal A234_032 HPMA-15 24.49 33.53 bal A234_033 HPMA-16 24.49 33.53 bal A234_034 HPMA-17 25.01 33.79 bal

Mn 0.65 0.78 1.31 0.65 0.65 0.89 0.79 0.56 0.56 1.37 0.91 0.65 0.56 1.37 0.56 0.56 0.65

Composition (wt%) Si C 1.08 0.45 0.80 0.50 1.37 0.42 1.08 0.45 1.08 0.45 1.12 0.47 0.85 0.46 1.14 0.44 1.14 0.44 1.73 0.46 1.33 0.48 1.08 0.45 1.14 0.44 1.73 0.46 1.14 0.44 1.14 0.44 1.08 0.45

Nb 0.65 0.49 0.40 0.65 0.65 0.95 1.08 0.51 0.51 0.83 0.65 0.51 0.83 0.51 0.51 0.65

N 0.06 NR NR 0.06 0.06 NR NR 0.05 0.05 NR NR 0.06 0.05 NR 0.05 0.05 0.06

Other Ti, Zr & W Ti W Ti, Zr & W Ti, Zr & W Ti & W Ti, Zr & W Ti, Zr & W Ti & Zr Ti & W Ti, Zr & W Ti, Zr & W Ti & Zr Ti, Zr & W Ti, Zr & W Ti, Zr & W

Atlas of Microstructures

Table 5: Alloy HP Modied Exposure Conditions Sample_ID Dechema Sample MTI Test-Temp. F A234_036 HPNb-1 (b) as cast A234_038 HPNb-2 1675 A234_039 HPNb-3 1700 A234_040 HPNb-4 1750 A234_041 HPNb-5 1778 A234_042 HPNb-6 1778 A234_043 HPNb-7 1800 A234_044 HPNb-8 1800 A234_045 HPNb-9 1800 A234_046 HPNb-10 1800 A234_047 HPNb-11 1850 A234_048 HPNb-12 1900 A234_049 HPNb-13 1900

Test-Temp. C as cast 913 927 955 970 970 982 982 982 982 1010 1038 1038

Stress Ksi 6.50 8.50 5.10 5.15 3.77 5.40 5.42 5.52 2.80 3.60 2.30 1.90

Stress MPa 44.6 58.8 35.2 25.5 26.0 37.4 37.4 37.4 19.3 24.8 15.8 13.1

Expo. Time hrs 659 59 794 286 1,185 191 4,467 7,833 10,637 707 2,555 5,373

Atlas of Microstructures

Table 6: Chemistry of Alloy HP Modied Stress Rupture Samples Sample_ID Sample Dechema MTI Cr Ni Fe A234_036 HPNb-1 (b) 25.30 34.93 bal A234_038 HPNb-2 25.34 34.88 bal A234_039 HPNb-3 25.26 33.56 bal A234_040 HPNb-4 25.34 34.88 bal A234_041 HPNb-5 25.11 32.85 bal A234_042 HPNb-6 25.11 32.85 bal A234_043 HPNb-7 24.88 33.33 bal A234_044 HPNb-8 24.86 33.79 bal A234_045 HPNb-9 24.86 33.30 bal A234_046 HPNb-10 25.26 33.56 bal A234_047 HPNb-11 25.62 34.75 bal A234_048 HPNb-12 25.26 33.56 bal A234_049 HPNb-13 25.26 33.56 bal

Mn 1.29 0.74 0.72 0.74 1.34 1.34 0.84 0.90 0.83 0.72 0.94 0.72 0.72

Composition (wt%) Si C 1.40 0.44 1.67 0.43 1.23 0.42 1.67 0.43 1.79 0.45 1.79 0.45 1.29 0.41 1.37 0.41 1.35 0.42 1.23 0.42 1.71 0.48 1.23 0.42 1.23 0.42

Nb 1.06 0.75 0.74 0.75 0.83 0.83 1.18 1.22 1.20 0.74 0.82 0.74 0.74

N NR 0.06 0.06 0.06 NR NR NR NR NR 0.06 0.06 0.06 0.06

Other W & Zr W & Zr W & Zr

W W W W & Zr W W & Zr W & Zr

Atlas of Microstructures

Metallographic Preparation
From the specimens available sections were taken for metallographic preparation with the surfaces to be investigated oriented parallel to the longitudinal direction of the tensile creep rupture specimens or perpendicular to the longitudinal axis of the tube specimens. Cutting was performed by using a precision sectioning machine with direct water cooling of the specimen. Some of the specimens had been delivered in the embedded state from which smaller sections were taken by saw cutting which were embedded again. Hot embedding in a conductive epoxy resin was used with a diameter of the moulds of 25 mm and after this the specimens were ground on SiC papers of the grades 180, 220, 320, 500, 1000 down to grit 2400 (Struers Standard 43-GB-1984, DIN 69176, Part 1,2,4) with a pressure of 70-80 N and water as a lubricant. Fine polishing was performed with a two step diamond polish of 3 m and 1 m followed by the nalizing polish with SiO2 suspension of size 0.02 m. The area on the specimen surface to be investigated was marked by four Vickers hardness imprints (HV1) with a distance of 500 m. Usually these areas were marked in the center of the specimen (in some cases, however, in addition regions close to the surface were investigated after marking).

Scanning Electron Microscopy (SEM-BSE)


After the LM-DIC investigations the same spot was investigated by the scanning electron microscope using the back-scattered electron imaging technique (SEM-BSE). In this case the dierences in contrast of the dierent precipitates resulting from dierences in the density can be used in order to distinguish between the phases.

Electron Probe Microanalysis (EPMA)


As a next step the phases documented by the other two techniques were analyzed quantitatively for selected specimens with regard to their chemical composition. In each case several spot measurements were taken with a beam diameter of 2 m. The results of these measurements were averaged. From the ratio between the dierent metals in the precipitates conclusions were drawn on the respective carbide type by taking carbon and nitrogen respectively from the dierence of the sum of the determined metal fractions and 100%.

Interference Layer Metallography (LM-ZnSe)


After the EPMA investigations the specimens were coated by a PVD process in an evaporation equipment (Edwards) at a vacuum of about 10-3 mbar3-7. As a coating material zinc selenide (ZnSe) was used which allowed the comparison of these marked areas in the colored state with the images from the other techniques. The evaporation source consists of a little vessel made of tantalum with ZnSe grains lled into this vessel. By resistance heating the vessel is heated to a point where the ZnSe starts to evaporate. The evaporation rate is controlled by controlling the heating current together with observation of changes of the color of the section surface. When reaching the desired color (violet as a macroscopic color) of the specimen surface the heating current is switched o immediately in order to achieve a reproducible thickness of the interference layer. Since the reection characteristics change periodically with the layer thickness and manual control of the heating current is not easy this coating technique requires signicant experience in order to achieve

Investigation Techniques
Dierential Interference Contrast Technique (LM-DIC)
For documentation of the microstructure in the light microscope at high magnication (500x and 1000x) the dierential interference contrast (LM-DIC) technique2 was used by which the dierent phases due to the dierent hardness reveal certain topographical structures which facilitates distinction of these phases.

Atlas of Microstructures

a suitable interference layer. In some cases the layer had to be removed again from the section surface several times and a new coating had to be applied in order to come to satisfactory results. Due to the small dimensions of the dierent phases there may be slight deviations between the dierent photographs taken by the dierent techniques.

The matrix was not binarized and was used as a background in true colors behind the false color image.

Development of the Microstructure by Etching (LM-etched)


Finally the interference layers were polished o and sections were etched for 30 seconds at 50C in etchant V2A (composition: 100 ml H2O, 100 ml HCl 1.19, 10 ml HNO3 1.40, 0.3 ml Dr. Vogels pickle. Dr. Vogels pickle is a mixture of organic solvents with Thiourea. It consists of 1-Methoxy- 2-propanol (40-50%), Thiourea (3-5%) and Nonylphenol-ethoxylate (5-7%). The interesting areas of the specimen were photographed at 200x and 500x magnication. By etching the matrix is partially removed so that edges are formed at the transition from matrix to precipitate. Due to the local reection situation of the light which hits the surface perpendicularly the edges of these phase boundaries appear dark in the photograph so that all phases have a dark seam and the grain boundaries become visible. (For more discussion on the etching technique, see the Addendum).

Image Analysis by the False Color Technique (LM-FC)


By the use of the automated image analysis system Leica QWin in combination with an automated laboratory microscope Leica DMLA the dierent fractions of the precipitations which had been characterized before by the other techniques were measured. For these measurements specimens were used which had been contrasted by the interference layer technique beforehand. An automated program routine was developed by which for each specimen several focused images were taken at representative spots in the specimen center and partially also at the specimen edges at a magnication of 1000x. The images were stored and in a second run the lower and upper threshold values for the RGB (red, green, blue)-colors of the dierent phases were determined and recorded. Since, during the coating process the ZnSe layer can vary from specimen to specimen, this procedure had to be performed for all specimens at least once. For specimens where the microstructure in the center and in the outer region was dierent this procedure had to be applied separately for the two regions. The program routine allows manual interaction with the measurement procedure, i.e. artifacts like creep pores or cracks can be eliminated in order not to inuence the measurement results. Based on the information in the binary memory, each of the phases was represented by a dened (false) color. For the dierent types of carbides and other phases the following false colors were selected: M7C3 - purple M6C - red M23C6 - yellow M2(C,N) - green M(C,N), MN - blue, cyan G-phase - magenta

Atlas of Microstructures

Table 9: Results of image analysis of Alloy 35Cr45Ni Creep Rupture Stress Samples Sample_ID Sample Temp. Temp F Stress Stress Expo. Time core C F MPa Ksi hrs M(C,N) M7C3 M23C6 M6C A234_003 35/45-1 (c) as cast as cast 0.8 3.9 A234_004 35/45-2 982 1800 22.8 3.30 1,509 0.5 7.9 A234_005 35/45-3 982 1800 20.0 2.90 7,217 0.3 7.8 A234_006 35/45-4 1050 1922 23.4 3.40 367 0.4 6.9 A234_007 35/45-5 1038 1900 14.4 2.090 2,023 < 0.05 8.3 < 0.03 A234_008 35/45-6 1038 1900 11.0 1.60 6,331 < 0.05 10.5 7.5 A234_009 35/45-7 1038 1900 10.6 1.54 10,247 0.1 15.1 10.7 A234_010 35/45-8 1100 2012 13.8 2.00 335 1.1 7.6 A234_011 35/45-9 1093 2000 12.4 1.80 1,297 0.4 10.8 0.3 A234_012 35/45-10 1100 2012 7.9 1.15 3,545 0.4 12.0 3.8 A234_013 35/45-11 1125 2057 11.1 1.60 277 1.1 7.2 A234_014 35/45-12 1121 2050 .06.9 1.00 2,606 0.2 12.7 9.2 A234_015 35/45-13 1149 2100 10.3 1.50 248 0.6 8.6 <0.3 Annotations: + few, ++ many, +++ very many

area fraction in % M2(C,N) G-Phase M(C,N) 1.5 1.8 0.7 3.0 0.8 0.3 0.3 < 0.02 < 0.05 1.1 0.2 0.6 0.2 0.6 outer zone / edge M7C3 M23C6 M6C 5.9 6.6 7.4 8.8 + + 7.8 10.2 8.9 7 9.9 6.3 12.5 4.2 13.1 +++ +++ 6.8 10.8 8.7 4.5 +++ 7.6 M2(C,N)

1.2

+ + + ++ ++ + ++ ++ + ++ ++

10

Atlas of Microstructures

Results of the Microstructural Analyses


Alloy 35Cr45Ni
The results from the image-analytical determination of the fractions of the precipitates in the microstructure are summarized in Table 9 (previous page) for all specimens investigated. As-Cast In the as-cast condition the microstructure consists of an austenitic matrix with a eutectic mixture of carbides of the types M7C3 and M(C,N) which is situated between the dendrites and along the grain boundaries respectively. The fraction of M7C3 amounts to about 4%, while that of M(C,N) lies at about 0.8%. The M7C3 carbides are rich in chromium, while the M(C,N) carbides are rich in niobium and/or titanium, Figures 1-1a to 1-1j. In the DIC representation the M7C3 carbides appear light gray and the M(C,N) have a reddish appearance and are slightly elevated above the matrix due to their higher hardness, Figures 1-1a and 1-1b. In the BSE image the Nb-rich M(C,N) particles appear signicantly brighter than the Cr-rich M7C3 particles and the Ti-rich M(C,N) which are almost black, Figures 1-1e and 1-f. After coating the specimen surface with ZnSe the MC particles containing high amounts of Nb appear in blue color, Figures 1-1g and 1-h, while the Ti-rich M(C,N) which could be detected in the DIC image by their strong intrinsic orange color were colored in light blue to white. The austenitic matrix and the M7C3 particles both have a reddish color, Figures 1-1g and 1-1h; however, the carbides can be distinguished quite clearly from the matrix as they are signicantly darker. After etching the carbides, which have a dark seam, can be clearly distinguished from the matrix, Figure 1-1i. In this gure it becomes evident that these have been precipitated in regions of interdendritic residual melts or at grain boundaries, respectively. In the false color image the M7C3 particles were

colored purple and the M(C,N) were colored blue, Figure 1-1j.

Creep rupture samples


1800F/982C 1800F/1,509 h/3.3 ksi (982C/1509 h/22.8 MPa) During exposure a transformation of the M7C3 particles had taken place. The microstructure reveals some single creep pores, Figure 1-2a, and a dominance of M23C6 carbides (7.9%) which had also formed very ne secondary precipitates, Figures 1-2a to 1-2d. Furthermore a small amount (1.5%) of G-phase can be observed which is frequently in direct contact with the M23C6 carbides and shows a gray color with a slightly reddish tone in the DIC image. It cannot be excluded that among those particles which had been identied as G-phase and which have a dot-like appearance (i.e. very small particles) could also be M6C carbides in some spots. Due to their small dimensions it had not been possible to identify these precipitates exactly either by their color contrast or by EPMA analysis. The G-phase (in the present analysis Ni16Nb6Si7) (according to the ASM Handbook of Metals also Ni18Ti8Si168) is supposed to have a negative inuence on creep lifetime9. Due to its high amount of Nb this phase appears bright in the BSE image, Figures 1-2e and 1-2f. After coating with ZnSe the color becomes dark blue, while the Ti and Nbrich M(C,N) particles (0.5%) which had appeared as orange precipitates in the DIC image become light blue to white, Figures 1-2c, 1-2d, 1- 2g and 1-2h. After etching, the contrast between these precipitates increases slightly, Figure 1-2i. In the false color image the following colors were chosen for the dierent phases: M23C6 yellow, M(C,N) and MN, respectively blue and cyan, respectively, G-phase magenta, Figure 1-2j. In the outer metal sub-surface zone of the specimen in addition the carbide M6C (6.3%) was found which has, however, not been documented photographically in the present case. 1800F/7,217 h/2.9 ksi (982C/7217 h/20.0 MPa) After 7,217 h the number of creep pores had increased in the micro-

Atlas of Microstructures

11

structure, Figure 1-3a, which is also true for the secondary precipitates of M23C6, Figures 1-3a to 1-3j. However, the volume fraction of the primary and secondary M23C6 carbides remained constant with 7.8%. The fraction of the G-phase has slightly increased with 1.8%. One striking feature is the bright, high Nb-containing dot-like inclusions in the G-phase which are residues of dissolved M(C,N), Figures 1-3c to 13h. The fraction of M(C,N) has decreased from 0.5 to 0.3% due to increased formation of G-phase. In the metal sub-surface zone M6(C,N) can replace the G-phase with a volume fraction of 12.5% and is most frequently directly in contact with the M23C6 carbide, Figures 1-3k to 13r. The composition of this phase can be taken from the table in Figure 1-3. In the DIC image this phase appears gray, while in the BSE image it can scarcely be distinguished from the matrix, Figures 1-3m and 1-3n. With increasing amounts of Ni this phase becomes brighter in the BSE image. In the ZnSe-coated condition the phase appears violet, Figures 1-3o and 1-3p, and in the false color image it had been assigned a red color, Figure 1-3r. The phase is resistant to V2A etchant, Figure 1-3q. The amount of M23C6 is slightly lower in the sub-surface area with 6.6% compared to the core area with 7.8%. In the outermost zone due to uptake of nitrogen the phase M2(C,N) appears which is slightly elevated in the metallographic section in the DIC image due to its higher hardness (not documented here). 1900F/1038C 1922F/367 h/3.4 ksi (1050C/367 h/23.4 MPa) Pores and cracks from the creep process can be clearly seen in the microstructure, Figure 1-4a. In the core of the specimen the phases M23C6, G-phase and M6C could be identied by EPMA. In the light microscopic investigations and BSE image furthermore the phase M(C,N) could be found, Figures 1-4a to 1-4j. Since the G-phase and M6C could not easily be distinguished from each other in the contrasted condition both phases were taken together in the image analysis procedure. The fractions given for these two phases in the table of Figure 1-4

are based on an estimation. The primary and secondary M23C6 particles amount to 6.9%, the G-phase to about 0.7%, the M6C carbides to 12% and the M(C,N) particles to 0.4%. Due to the high amount of Nb the M6C carbides in the BSE image appear with a similar level of gray as the G-phase, Figures 1-4e and 1-4f, while M(C,N) has the brightest appearance. In the sub-surface zone the amount of M23C6 increases slightly (7.4%), that of the M6C carbides reveals a signicant increase (4.2%), while the G-phase no longer exists in these positions, Figures 14k to 1-4r. Again, directly under the outer part of the metal sub-surface zone M2(C,N) can be detected. 1900F/2,023 h/2.09 ksi (1038C/2023 h/14.4 MPa) With increasing exposure time the amount of pores and cracks from creep has signicantly increased, Figure 1-5a. The fraction of primary and secondary M23C6 (8.3%) and of the G-phase (3.0%) increase in the core of the specimen, while almost no M6C (< 0.03%) and M(C,N) (< 0.05%) can be found, Figures 1-5a to 1-5j. In the sub-surface zone the amount of M23C6 (8.8%) and of M6C (13.1%) has signicantly increased. G-phase and M(C,N) are scarcely found there with a fraction of less than 0.02%, Figures 1-5k to 1-5r. Again, M(C,N) can only be found in the outermost zone of the sub-surface area. 1900F/6,331 h/1.6 ksi (1038C/6331 h/11.0 MPa) After exposure of more than 6,000 h a signicant coarsening and increase of the volume fraction of M23C6 (10.5%) and of M6C (7.5%) had taken place in the core and in the sub-surface area. The size of the pores has increased, Figures 1-6a and 1-6b. The secondary carbides have been dissolved or have become larger and blocky, Figures 1-6a to 1-6r. In the outermost zone of the sub-surface again M2(C,N) is detected which is slightly elevated in the LM-DIC image and appears somewhat darker than M23C6 and brighter than M6C, Figure 1-6d. In the SEM-BSE image this phase appears very dark so that it can be clearly distinguished

12

Atlas of Microstructures

from M6C, Figures 1-6u and 1-6v. After contrasting the phases by ZnSe M2(C,N) has a violet color similar to M6C with, however, a slightly brighter tone, Figure 1-6y. It was, however, not possible to distinguish between the two phases in the image analytical process due to an overlap in the color distributions in the red and blue range. In the false color image the M2(C,N) was colored green, Figure 1-6z. 1900F/10,247 h/1.54 ksi (1038C/10247 h/10.6 MPa) After more than 10,000 h the amount of M23C6 has increased to 15% and that of M6C to 10.7%. The size of the carbides has increased even more, Figures 1-7a to 1-7j, and a larger number of M23C6 carbides is found than after 6,331 h. A large number of pores from creep are found which have agglomerated increasingly to cracks oriented perpendicular to the stress direction. In the sub-surface zone additionally M2(C,N) is found, Figures 1-7k to 1-7r. 2000F/1093C 2012F/335 h/2.0 ksi (1100C/335 h/13.8 MPa) After 335 h of exposure at 2012F (1100C) the microstructure in the core reveals about 7.6% M23C6 (primary and secondary carbides) and about 1.1% M(C,N), Figures 1-8h to 1-8j. Again a large number of creep pores and cracks are observed. In the sub-surface zone furthermore about 6.8% of M6C carbides are present which frequently contain residues of the Nb-rich M(C,N) in their center, Figures 1-8k to 1-8r. 2000F/1,297 h/1.8 ksi (1093C/1297 h/12.4 MPa) With increasing exposure time under creep conditions the pores have grown signicantly, Figure 1- 9a. The M23C6 carbides have also grown in their size with the volume fraction in the core area of the specimen increasing up to 10.8%, Figures 1-9a and 1-9b. Besides a small fraction (0.3%) of M6C the microstructure also reveals the presence of the G-phase and of M(C,N). Due to an overlap in the colors the fractions could only be determined as a sum on the ZnSe-coated metallographic section, Figures 1-9g and 1-9h. It is estimated that about 1.2% are G-

phase, while 0.4% are M(C,N). The total fraction of these two phases was determined to be 1.6%. In the false color image the phases were partially re-drawn manually, Figure 1-9j. In the sub-surface zone again higher amounts of M6C (10.8%) are found beside M23C6 (10.3%), while the G-phase is missing, Figures 1-9k to 1-9r. The phase M2(C,N) is only found in the outermost area again, Figures 1-9s to 1-9z. 2012F/3,545 h/1.15 ksi (1100C/3545 h/7.9 MPa) In the core area of the specimen the fraction of the signicantly coarsened M23C6 carbides has increased up to 12% and the fraction of M6C amounts to 3.8%, Figures 1-10a to 1-10j. Furthermore some very ne secondary precipitates of M23C6 are also found. The G-phase is no longer present. The fraction of M(C,N) amounts to 0.4%. In the subsurface zone of the specimens about equal fractions of M23C6 and M6C are observed (8.8%). Furthermore the microstructure contains about 0.6% M(C,N) and M2(C,N), Figures 1-10k to 1-10r. 2050F/1121C 2057F/277 h/1.60 ksi (1125C/277 h/11.1 MPa) After 277 h of exposure at the further increased temperature of 2057F (1125C) several rows of creep pores oriented perpendicular to the tensile direction are observed, Figure 1-11a. The microstructure in the core of the specimen contains 7.2% of carbides of the type M23C6 (primary and secondary carbides) and 1.1% of the type M(C,N), Figures 1-11c to 1-11j. Furthermore in the sub-surface zone, Figures 1-11k to 1-11r, 4.5% of M6C are observed while the fraction of M(C,N) has signicantly decreased (0.2%). M2(C,N) is observed only locally in the outermost sub-surface area. 2050F/2,606 h/1.00 ksi (1121C/2606 h/6.9 MPa) Compared to the shorter exposure time coarsening and an increase in the number of M23C6 carbide to a fraction of 12.7% has taken place

Atlas of Microstructures

13

in the core area of the specimen with a simultaneous decrease of the ne secondary carbides, Figures 1-12a to 1-12j. The formation of a signicant fraction of carbides of the type M6C (9.2%) is associated with a dissolution of the carbides of the type M(C,N) whose fraction amounts only to 0.2%. In the metal sub-surface zone the carbide of the type M2(C,N) is found in a signicant amount, Figures 1-12k to 1-12z. 2100F/248 h/1.50 ksi (1149C/248 h/10.3 MPa) After brief exposure at the highest temperature of 2100F (1149C) a fraction of 8.6% of M23C6 primary and secondary carbides as well as of about 0.6% of M(C,N) is observed after 248 h with a small fraction of M6C (< 0.3%), Figures 1-13a to 1-13j. In the sub-surface zone the amount of M6C carbides is signicantly higher (7.6%) with also the amount of M23C6 having increased (9.9%). Furthermore in the outermost sub-surface zone which had been strongly oxidized and depleted in chromium M2(C,N) is also observed, Figures 1-13s to 1-13v.

Conclusions
The results of these investigations can be summarized as follows: The centrifugally cast material 35Cr/45Ni in the as-cast condition consists of an austenitic matrix with eutectic areas of chromium-rich carbides of the type M7C3 and of Nb or Ti-rich particles of the type M(C,N) which have precipitated between the grains or along the grain boundaries respectively. Already after 200-300 h of exposure in the temperature range of 1800-2100F (982 - 1149C) a complete conversion of the carbides of the type M7C3 into chromium-rich carbides of the type M23C6 takes place with the latter type of carbides forming secondary precipitates in addition. With increasing exposure time the number of these secondary carbides increases and also the fraction of the primary M23C6 carbides increases, accompanied by signicant growth. At the same time additional M6C and/or Gphase are formed. Formation of the latter phases initiates at the phase boundary between M23C6 and matrix. M6C is rich in Cr, Ni and Nb, while the G-phase (Ni16Nb6Si7) contains

less Cr, but is rich in Ni, Si and Nb. Frequently the G-phase shows a seam of small round M23C6 secondary carbides. M6(C,N) shows a stronger tendency of formation in the metal sub-surface zone compared to the core area which may be explained by depletion processes due to outward diusion of oxide-forming elements (e.g. Cr, Si) or by the uptake of nitrogen. The appearance of M6C seems to be stimulated at higher temperatures, i.e. it appears after shorter times. Furthermore the fraction increases with increasing exposure time. The G-phase is only found in the temperature range of 1800-2000F (982 - 1149C). In the lower part of this temperature range (1800F (982C)) this phase is still observed after 7,217 h, while at 2000F (1093 C) it has disappeared after 6,331 h. At 1900F (1038C) it can be found after 1,300 h, while after 3,545 h it no longer exists and seems to be replaced by M6C. At 2012F (1100C) this phase is not observed even after 335 h. In this case the lower Si content (1.52 mass-%) may have played a role9. An increased amount of silicon (> 1.6% mass-%) promotes the formation of G-phase by facilitating the conversion of NbC into G-phase. In the G-phase as well as in the M6C particles enrichment of Nb is observed which directs towards the dissolution of former MC10-15. The dissolution of MC with increasing exposure time is conrmed by a decrease of their volume fraction. In the cast state the fraction amounts to about 0.8%, while after short exposure times in which no G-phase has formed yet it amounts to 0.6-1.1%. When G-phase formation or after longer times G-phase and M6C formation start the fraction goes down to less than 0.02-0.4%. After several thousand hours above 1900F (1038C) the secondary carbides (M23C6) dissolve and form blocky, elongated particles. Besides these coarse carbides of the type M23C6 where the particles sometimes also have grown together only coarse carbides of the type M6C and in the sub-surface zone of the type M2(C,N) are present. The G-phase is no longer found.

14

Atlas of Microstructures

Table 10: Results of image analysis of Alloy HPMA Creep Rupture Stress Samples
Sample_ ID Sample Temp. C as cast 927 927 927 954 982 982 982 1000 1050 1066 1050 1050 1100 1100 1100 1100 Temp F F as cast 1700 1700 1700 1750 1800 1800 1800 1832 1922 1950 1922 1922 2012 2012 2012 2012 Stress MPa 55.2 41.3 36.6 30.7 41.2 24.2 26.9 20.7 24.0 19.1 16.0 14.3 16.5 12.4 10.0 9.3 Stress Ksi 8.00 6.00 5.30 4.45 5.97 3.51 3.90 3.00 3.48 2.80 2.32 2.08 2.39 1.80 1.45 1.35 Expo. Time hrs 183 2,338 6,072 8,359 177 2,436 6,478 12,289 137 2,558 4,864 11,778 105 2,714 5,580 8,990 area fraction in % M(C,N) 0.9 0.6 <0.3 0.8 0.8 0.6 1.2 0.5 0.8 0.7 1.0 1.1 0.8 0.8 0.8 1.0 1.1 M7C3 3.4 core M23C6 M6C 7.0 8.3 8.0 8.7 7.0 7.9 8.1 9.0 9.5 8.8 9.7 13.1 7.7 9.6 11.1 10.9 0.8 0.7 M2(C,N) G-Phase M(C,N) + + 0.8 + 0.4 0.4 0.5 + 0.4 0.9 + 0.8 + + + + outer zone / edge M7C3 M23C6 M6C M2(C,N) + ++ +++ + +++ + ++ <0.3 +++ 5.3 4.9 6.4 4.7 6.1 1.0 5.6 7.1 +++ 0.0 +++ +++ ++++ ++++ 9.8 5.1 15.2 +

A234_018 A234_019 A234_020 A234_021 A234_022 A234_023 A234_024 A234_025 A234_026 A234_027 A234_028 A234_029 A234_030 A234_031 A234_032 A234_033 A234_034

HPMA-1 (c) HPMA-2 HPMA-3 HPMA-4 HPMA-5 HPMA-6 HPMA-7 HPMA-8 HPMA-9 HPMA-10 HPMA-11 HPMA-12 HPMA-13 HPMA-14 HPMA-15 HPMA-16 HPMA-17

0.5 1.3 <0.2 1.7

Annotations: + few, ++ many, +++ very many, ++++ abundant

Atlas of Microstructures

15

Alloy HPMA
The results of the image analytical determination of the volume fractions of all specimens of this material are summarized in Table 10 (previous page). As-Cast Figures 2-1a to 2-1j show the as-cast condition after application of the dierent imaging techniques. In the austenitic matrix two dierent types of precipitates can be distinguished which are located on the grain boundaries or in areas of the residual melt. As the EPMA analyses reveal these precipitates are Nb or Ti-rich carbides MC or carbon nitrides of the type M(C,N) and Cr-rich carbides of the type M7C3 which can be distinguished even in the DIC image due to their shape, color and size, Figures 2-1a to 2-1d. M(C,N) precipitates have dierent modications with varying contents of Nb, Ti and Cr. In the BSE image 2-1b the heavier elements Zr and Nb appear brighter than the lighter ones. Therefore the Ti-rich core of M(C,N) appears darker. A better contrast is achieved by ZnSe coating, Figures 2-1g and 2-1h, after which the M7C3 appear in purple, the Ti-rich cores of the M(C,N) particles in light blue to white and the highly Nb-containing species in blue. Truecolor imaging allows a reliable identication of the dierent phases and a quantication by the binary image, Figure 2-1j. The fractions of M7C3 amount to 3.4% and of M(C,N) to 0.9%.

phase containing Cr, Ni, Si and Nb this creep specimen exhibits the phase M6C which also contains some Fe. This phase appears as light gray in the DIC image and is partially encapsulated by M23C6, Figures 2-2c and 2-2d. Frequently in the M6C particles, which can hardly be distinguished in the BSE image, Figures 2-2e and 2-2f, from the surrounding matrix bright Zrrich MC areas are also present. Coating of the specimen with ZnSe facilitates the detection of the M6C phase (which can easily be overlooked in the BSE image) by changing the color to violet. This is illustrated by the comparison of Figure 2-2f and Figure 2-2h. In the false color image, Figure 2-2j, the M23C6 (7.0%) are colored yellow, the M6C (0.8%) are colored red and the M(C,N) (0.6%) are colored blue. 1700F/2,338 h/6.00 ksi (927C/2338 h/41.3 MPa) Even after more than 2,000 h of exposure the microstructure has not changed signicantly, Figures 2-3a to 2-3r. In the core about 8.3% M23C6, 0.7% M6C and less than 0.3% MC are observed. The decrease in the amount of MC is probably due to the fact that according to chemical analysis this specimen contains less Nb and no Ti, as well as a lower amount of C, so that in the end fewer M(C,N) particles were formed. The sub-surface zone has a similar appearance of the microstructure. 1700F/6,072 h/5.30 ksi (927C/6072 h/36.6 MPa) Even after 6,000 h of exposure no signicant changes in the microstructure can be found. The fraction of M23C6 in the core area amounts to 8.0%, while the amount of M(C,N) has increased to 0.8% due to the chemical composition of the specimen, Figures 2-4a to 2-4f. In the sub-surface zone of the specimen some M6C (< 0.3%) had precipitated, Figures 2-4k to 2-4r. 1750F/8,359 h/4.45 ksi (954C/8359 h/30.7 MPa) After more than 8,000 hours at a slightly increased temperature the amount of M23C6 has increased slightly from 8.0% to 8.7% which is probably due to an increased precipitation of secondary carbides, Figures 2-5a to 2-5r. The amount of M(C,N) of 0.8% has remained constant

Creep Rupture Samples


1700F/927C 1700F/183 h/8.00 ksi (927C/183 h/55.2 MPa) Already after 183 h at a temperature of 1700F (927C) and a load of 8.0 ksi the M7C3 has been converted into the more stable M23C6. Furthermore M23C6 secondary carbides are observed in the grain interior, Figures 2-2b to 2-2d. The EPMA analyses reveal that the conversion of M7C3 into M23C6 is mainly due to an uptake of Ni from the matrix. The Ni content of M23C6 amounts to about 4-5 mass-%. As a further

16

Atlas of Microstructures

which is consistent with the identical compositions and phases formed in the specimens numbers HP-MA-4 and HP-MA-5. 1800F/982C 1800F/177 h/5.97 ksi (982C/177 h/41.2 MPa) In the austenitic matrix of the specimen again about 7.0% M23C6 (primary and secondary carbides) as well as 0.6% M (C,N) have been found, Figures 2-6a to 2-6r. The outermost sub-surface zone additionally contains about 5% M6C and M2(C,N) respectively (not documented here). The preferred formation of M6C in the sub-surface area is presumably due to the relatively high amount of Nb in the specimen and the outward diusion of Cr and Si which are needed for the formation of the oxide scale. Compared to M23C6 the carbide M6C is less rich in Cr and Fe, but contains more Ni, Si and Nb. The formation of M2(C,N) which is formed in the outermost sub-surface zone is probably due to outward diusion of Cr and nitrogen uptake from the environment. The matrix in the metal sub-surface zone was strongly depleted in Cr around these M2(C,N) particles. 1800F/2,436 h/3.51 ksi (982C/2436 h/24.2 MPa) After 2,436 h the amount of secondary carbides evidently increases slightly which is reected by an increase of M23C6 to 7.9%, Figures 2-7a to 2-7j. The high amount of M(C,N) of 1.2% is due to the composition of the material, cf. Table 4. Again in the metal sub-surface zone the formation of M6C and M2(C,N) is observed, Figures 2-7k to 2-7z. 1800F/6,478 h/3.90 ksi (982C/6478 h/26.9 MPa) After more than 6,000 h besides M23C6 (8.1%) and M(C,N) (0.5%) smaller amounts of M6C (0.5%) are also found in the core area, Figures 2-8a to 2-8j. In the metal sub-surface zone the amount of M6C is twice as high as in the core, Figures 2-8k to 2-8r. 1832F/12,289 h/3.0 ksi (1000C/12289 h/20.7 MPa)

After more than 12,000 h a coarsening of the primary carbides and a coagulation of secondary carbides have taken place, Figures 2-9a to 2-9r. The fraction of M23C6 amounts to 9.0%, M6C is observed neither in the core nor in the metal subsurface zone. The fraction of M(C,N) amounts to 0.8% and is clearly higher than in the case of specimen HP-MA-8 where in the core M6C had also been found. 1920F/1050C 1922F/137 h/3.48 ksi (1050C/137 h/24.0 MPa) At this further elevated temperature a signicant increase in the number of creep pores and cracks can be observed, Figure 2-10a. The M23C6 primary carbides seem to have coarsened and the secondary carbides show a needle-like and blocky appearance, Figures 2-10c to 2-10r. Their fraction amounts to 9.5% in the core area with that of M(C,N) amounting to 0.7%. Furthermore the core area contains 1.3% M6C. The metal sub-surface zone shows the following fractions: M23C6 5.6%, M23C6 0.4% and M6C and M2(C,N) (as a sum) 9.8%, Figures 2-10k to 2-10r. 1950F/2,558 h/2.80 ksi (1066C/2558 h/19.1 MPa) After more than 2,500 h the amount of creep damage has further increased, Figure 2-11a. In the core area the beginning of M6C formation (< 0.2%) has been found. The fraction of M23C6 amounts to about 8.8% and that of M(C,N) to about 1.0%, Figures 1-11a to 1-11j. The microstructure of the sub-surface zone contains the following phases: M(C,N) 0.9%, M23C6 7%, and M6C and M2(C,N) as a sum 5%, Figures 2-11a to 2-11r. 1922F/4,864 h/2.32 ksi (1050C/4864 h/16.0 MPa) After almost 5,000 h of exposure the amount of M23C6 carbides has increased up to 9.7%. The fraction of M(C,N) amounts to about 1%. M6C and M2(C,N) are neither observed in the core nor in the metal sub-surface zone, Figures 2-12a to 1-12r.

Atlas of Microstructures

17

1922F/11,778 h/2.08 ksi (1050C/11778 h/14.3 MPa) After almost 12,000 h a signicant coarsening and an increase in the amount of M23C6 carbides has taken place in the core area of the specimen with about 13.1%. M2(C,N) has spread from the metal subsurface zone into the core area, Figures 2-13c to 2-13j. Due to its higher hardness M2(C,N) can easily be distinguished from the other carbides in the DIC image. Even in the BSE image it shows a good contrast, Figures 2-13e and 2-13f. EPMA investigations reveal that this carbonitride is almost free of Ni and has the composition (Cr, Fe)2 (C,N), cf. table in Figures 2-13a and 2-13b. Values for the ratio between carbon and nitrogen cannot be given here since C and N had not been analyzed. Although the particles of M2(C,N) are rather coarse, which should facilitate their identication, it is not easy to distinguish them from M6C in the ZnSe-coated section, Figures 2-13e to 2-13r. The false color image, Figure 2-13j, was constructed by image manipulation in which the color green was assigned to M2(C,N). Besides about 0.8% M(C,N) about 1.7% M6C plus M2(C,N) were measured in the core area, Figures 2-13a to 2-13j. In the metal sub-surface zone no M23C6 was found. Besides about 15% M2(C,N) about 0.8% M(C,N) were present, Figures 2-13k to 2-13r. 2012F/1100C 2012F/105 h/2.39 ksi (1100C/105 h/16.5 MPa) Compared to specimen HP-MA-10 (1922F/137 h) (xxx) the amount of M23C6 carbides is lower with the precipitates being coarser, Figure 2-14a to 2-14r. Almost no dierences between the core area and the metal sub-surface zone are observed with 7.7% M23C6 beside 0.8% M(C,N). 2012F/2,714 h/1.80 ksi (1100C/2714 h/12.4 MPa) After 2,714 h the fraction of M23C6 carbides has increased due to an increase of secondary precipitates up to 9.6%. The fraction of M(C,N) of 0.8% remained about constant. The microstructure in the core and in the metal sub-surface zone is identical, Figures 2-15a to 2-15r.

2012F/5,580 h/1.45 ksi (1100C/5580 h/10.0 MPa) With increasing exposure time the amount of M23C6 increases further (11.1%). The number of the secondary carbides decreases since the very ne particles combine with the larger ones to form coarse particles and some of them seem to agglomerate with the primary carbides, Figures 2-16a to 2-16r. The fraction of M(C,N) lies around 1%. Occasionally high Nb-containing needles are found which most probably are nitrides, Figure 2-16b. The thickness of these needles is less than 1 m and their length reaches up to 100 m. In the DIC image they appear gray and after ZnSe coating violet. 2012F/8,990 h/1.35 ksi (1100C/8990 h/9.3 MPa) After almost 9,000 h of exposure time no visible changes in the microstructure are observed compared to the former specimen, Figure 2-17a to 2-17r. Even the fractions of M23C6 (10.9%) and of M(C,N) with 1.1% are comparable. Again highly Nb-containing precipitates are found, Figures 2-17b and 2-17s.

Conclusions
The results of the investigations can be summarized as follows: The centrifugally cast material HP-MA (microalloy) in its cast state consists of an austenitic matrix with a eutectic structure at interdendritic sites or along grain boundaries respectively, formed by Crrich carbides of the type M7C3 and Nb or Ti-rich carbonitrides or carbides of the type M(C,N). Already after 100-200 h of exposure in the temperature range of 1700-2012F (927 - 1100C) a complete conversion of the carbides of the type M7C3 into Cr-rich carbides of the type M23C6 takes place. The M23C6 carbides form secondary precipitates. With increasing exposure time the number of secondary carbides increases and later on also the number of primary M23C6 carbides rises accompanied by a strong growth in size. The fraction increases from about 7% to a maximum of 11-13%. At the same time some M6C is precipitated in the core

18

Atlas of Microstructures

area, while this is more pronounced in the metal sub-surface zone. At 1700F (927C) the M6C precipitates with an amount of less than 1% are dissolved after times longer than 2,500 h in the core of the specimen and exist in the metal sub-surface area up to about 6,000 h where, however, only an amount of less than 0.3% can still be detected. After more than 8,000 h no more M6C can be found. At the higher temperature of 1800F (982C) M6C can be found only in the metal sub-surface zone at times up to 2,500 h. After 6,500 h traces (< 0.5%) of M6C are observed in the core area, while the fraction in the metal sub-surface zone decreases signicantly. Due to the uptake of nitrogen M2(C,N) is also formed whose fraction was measured as part of the sum with M6C. After 12,000 h at 1832F M6C is neither found in the core nor in the metal subsurface zone. At temperatures above 1900F (1038C) small amounts of M6C (< 0.2-1.3%) are observed in the core after exposure times of 2,500 h which becomes dissolved after longer times. The formation of M6C and M2(C,N) occurs only in the metal sub-surface zone to a signicant extent with their fraction decreasing markedly after 2,500 h due to outward diusion processes. However, after almost 5,000 h M6C and M2(C,N) are observed neither in the core nor in the metal subsurface zone. After almost 12,000 h the extension of M2(C,N) is marked and reaches from the sub-surface zone into the core of the specimen. In the core only single precipitates of M6C are observed besides M23C6 and M(C,N), while in the outer part of the specimen only M2(C,N) and M(C,N) are present. At the maximum temperature of 2012F (1100C) the increase of the fraction of M23C6 with time is very marked. M6C and M2(C,N) are not present in the core and are observed only after short exposure time (100 h) and sporadically in the metal sub-surface zone. After longer exposure times in the core area some highly Nb-containing needle-shaped phases are formed. In the metal sub-surface zone an increasing depletion of oxide-forming elements takes place so that nally M6C and M2(C,N) are dissolved completely. In none of the specimens investigated was the G-phase detected.

Alloy HP Modied
The results of the image analytical determination of the volume fractions of the precipitates are summarized in Table 11. As-Cast Figures 3-1a to 3-1j show the as-cast condition after application of the dierent imaging techniques. The microstructure is comparable to that of the as-cast condition of the material HPMA. The precipitates along the grain boundaries or in the interdendritic residual melt areas are the Cr-rich carbides of the type M7C3 (3.9%), which appear bright gray in the DIC image, and Nb-rich carbides of the type M(C,N) with 0.9% which appear slightly red in the DIC image, Figures 3-1c and 3-1d. In the ZnSe-coated state M7C3 appears purple and M(C,N) blue, Figures 3-1g and 3-1h.

Creep Rupture Samples


1675F/913C 1675F/659 h/6.50 ksi (913C/659 h/44.6 MPa) At the lowest temperature of 1675F (913C) after 659 h in the core primary carbides of the type M23C6 are found together with elongated secondary carbides of the same type (together 8%). Furthermore some carbides of the type M6C (0.5%) and its secondary precipitates are observed together with about 1.8% G-phase and 0.3% M(C,N), Figures 3-2a to 3-2j. The M(C,N) precipitates besides a high amount of Nb contain also Zr, see table in Figures 3-2a and 3-2b. Presumably the formation of G-phase was favored by the high amount of Si of 1.67 mass-%, see Table 6. In the DIC image it was not easy to distinguish between the G-phase and the carbides of the type M6C since both of them appear gray, Figures 3-2c and 3-2d. In the BSE image the G-phase appears bright due to its high Nb content, while M6C does not dier much from the matrix, Figures 3-2e and 3-2f. The M(C,N) particles appear bright in the BSE image. These particles can, however, be distin-

Atlas of Microstructures

19

Table 11: Results of image analysis of Alloy HP Modied Creep Rupture Stress Samples
Sample_ID Sample Temp. C as cast 913 927 955 970 970 982 982 982 982 1010 1038 1038 Temp F F as cast 1675 1700 1750 1778 1778 1800 1800 1800 1800 1850 1900 1900 Stress MPa 44.6 58.8 35.2 35.5 26.0 37.4 37.4 37.4 19.3 24.8 15.8 13.1 Stress Ksi 6.5 8.5 5.1 5.2 3.8 5.42 5.42 5.42 2.8 3.6 2.3 1.9 Expo. Time hrs 659 59 794 286 1,185 191 4,467 7,833 10,637 707 2,555 5,373 area fraction in % M(C,N) 0.9 0.3 1.0 0.4 0.5 0.4 3.0 2.2 1.7 0.4 0.7 0.6 0.6 M7C3 3.9 core M23C6 M6C 8.0 8.0 7.5 6.7 7.1 7.6 6.4 7.1 11.1 7.8 8.4 9.5 0.5 0.7 2.4 1.1 0.7 M2(C,N) G-Phase M(C,N) + 1.8 (+) + (+) (+) + (+) ++ ++ ++ (+) + + + M7C3 ++ outer zone / edge M23C6 M6C M2(C,N) G-Phase +++ +++ +++ +++ +++ +++ +++ +++ ++++ +++ +++ ++++ + + ++ ++ + (+) (+) (+) ++ ++ ++ ++ +

A234_036 A234_038 A234_039 A234_040 A234_041 A234_042 A234_043 A234_044 A234_045 A234_046 A234_047 A234_048 A234_049

HPNb-1 (b) HPNb-2 HPNb-3 HPNb-4 HPNb-5 HPNb-6 HPNb-7 HPNb-8 HPNb-9 HPNb-10 HPNb-11 HPNb-12 HPNb-13

(+) (+) + + + +

3.9 0.9 2.5 1.0

Annotations: (+) very few, + few, ++ many, +++ very many, ++++ abundant

20

Atlas of Microstructures

guished from the G-phase due to their blocky shape. After ZnSe coating the M(C,N) particles appear in light blue to white, the G-phase in blue, the M6C particles in violet and the M23C6 carbides in magenta and are thus easy to distinguish. The G-phase particles and the M6C particles had to be measured together in the image analytical investigations since it was not easy to distinguish between the two of them and to measure their fractions separately. The fractions of the two phases were estimated manually. The false color image in Figure 3-2j needed some manual adjustment and the phases were colored as follows: M23C6 yellow, Gphase magenta, M(C,N) blue and cyan, M6C red. Etching with V2A etchant improves the visibility of the precipitates in the matrix, Figure 3-2i. The microstructure in the sub-surface zone is not dierent from that in the core, Figures 3-2o to 3-2r. 1700-1750F/913-955C 1700F/59 h/8.5 ksi (927C/59 h/58.8 MPa) After this short exposure time at 1700F (927C) the microstructure is full of secondary precipitates. The M23C6 carbides (8.0%) show typical eutectic appearance, Figures 3-3a to 3-3r. Besides a small amount of M6C (0.7%), some M(C,N) are also observed with about 1% in the core area as well as in the metal sub-surface zone. 1750F/794 h/5.10 ksi (955C/794 h/35.2 MPa) After this increased exposure time at 1750F (954C) the amount of M6C in the core area has slightly increased compared to specimen HPNb-3, while the amount of M(C,N) has clearly decreased, Figures 3-4a to 3-4j. Possibly some G-phase is present in the M6C carbides which, however, could not be evidenced by EPMA. The bluish coloring in the violet M6C seemed to indicate their presence, Figure 3-4a. Since the chemical composition of this specimen is identical to specimen HP-Nb2, the appearance of the G-phase would have been expected here. In the image analysis this phase was measured together with M6C or M(C,N) so that their fractions may appear slightly too high. In the sub-surface zone

of the specimen residues of dissolving M(C,N) are frequently detectible due to their blocky shape and are embedded into M6C, Figures 3-4k to 3-4r. Bluish colorings which would indicate the presence of G-phase were not detected here. 1778F/970C 1778F/286 h/5.15 ksi (970C/286 h/35.5 MPa) After this relatively short exposure time at 1778F (970C) the core area contains 6.7% M23C6, 1.1% M6C and 0.5% M(C,N), Figures 3-5a to 3-5j. Despite the relatively high amount of Si of 1.79 mass-% the G-phase could not be identied reliably. The metal sub-surface zone does not show any dierences compared to the core area, Figures 3-5k to 3-5r. 1778F/1,185 h/3.77 ksi (970C/1185 h/26.0 MPa) After this extended exposure time of almost 1,200 h no dierences with regard to the volume fractions of the phases identied by EPMA are observed compared to specimen HP-Nb-5, Figures 3-6a to 3-6r. The G-phase is not detected despite the high Si amount of 1.79 mass-%. 1800F/982C 1800F/191 h/5.42 ksi (982C/191 h/37.4 MPa) The specimens HP-Nb-7 to HP-Nb-9 possess a high amount of Nb (1.2 mass-%) which has a signicant eect on the microstructure. Besides M23C6 a large number of M(C,N) are precipitated in very ne structures and a eutectic appearance, Figures 3-7a to 3-7r. Their volume fraction amounts to 3% with an amount of 7.6% of M23C6. In the outermost zone of the specimen some M6C and M2N are encountered. 1800F/4,467 h/5.42 ksi (982C/4467 h/37.4 MPa) After an exposure time of 4,467 h the microstructure has not changed compared to specimen HPNb- 7, Figures 3-8a to 3-8r. The volume fractions of the carbides determined in the image analytical investigations

Atlas of Microstructures

21

are about equal to the former specimen. 1800F/7,833 h/5.42 ksi (982C/7833 h/19.3 MPa) Even after almost 8,000 h no changes in the microstructure can be found compared to the short exposure times, Figures 3-9a to 3-9r. Again the volume fractions of the carbides are unchanged. 1800F/10,637 h/2.80 ksi (982C/10637 h/19.3 MPa) Compared to specimens HP-Nb-7 to HP-Nb-9 this specimen HP-Nb-10 contains only 0.74 mass-% Nb, which is reected by the appearance of the M(C,N) precipitates. The microstructure does not show the typical eutectic characteristics, Figures 3-10a to 3-10r. The long exposure time of more than 10,000 h has led to a coarsening of the primary carbides of M23C6. The fraction of M(C,N) plus M23C6 amounts to 11%. In addition about 4% M6C and 0.4% M(C,N) were formed. In the metal sub-surface zone the amount of M6C is signicantly higher and due to the uptake of nitrogen some M2(C,N) is observed, Figures 3-10k to 3-10r. 1850-1900F/1010-1038C 1850F/707 h/3.60 ksi (1010C/707 h/24.8 MPa) This specimen contains a high amount of Si with 1.7 mass-%. The G-phase, however, could not be detected here. The following precipitates were observed: M23C6 7.8%, M2(C,N) 0.7% and M6C 0.9%. M(C,N) or their undissolved residues are frequently observed in the core of M6C particles, Figures 3-11a to 3-11j. In the outermost sub-surface zone additionally some M2(C,N) is present, Figures 3-11k to 3-11r. 1900F/2,555 h/2.30 ksi (1038C/2555 h/15.8 MPa) After 2,555 h at slightly increased temperature growth of the M23C6 carbides had taken place. The number of the secondary carbides has decreased. The volume fraction of M23C6 has increased to 8.4% and that of M6C to 2.5%, while the fraction of M(C,N) has remained almost constant with 0.6%, Figures 3-12a to 3-12r. In the outermost zone some M2(C,N) is also present, which is colored green in the false color

image in Figure 3-12r. 1900F/5,373 h/1.90 ksi (1038C/5373 h/13.1 MPa) After doubling the exposure time at a test temperature of 1900F (1038C) the M6C carbides have become partially dissolved, while again increasingly M23C6 secondary carbides were precipitated, Figures 3-13a to 3-13r. Their volume fraction of M23C6 increases up to 9.5%, while that of M6C decreases to 1%. The amount of M(C,N) with 0.6% remains stable.

Conclusions
The results of these investigations can be summarized as follows: The centrifugally cast material HP-Nb (modied alloy) in its cast version consists of an austenitic matrix with eutectic structures of the Cr-rich carbides of the type M7C3 and of Nb and Zr carbides of the type M(C,N). These structures are found along grain boundaries and in interdendritic areas. Already after an exposure time of 59-700 h and a temperature range of 1700-1900F ( 927 -1038C) a complete conversion of the carbides of the type M7C3 into Cr-rich carbides of the type M23C6 takes place with the latter also forming secondary precipitates. In addition in specimens with Nb contents between 0.74 and 0.83 mass-% M6C carbides are precipitated. The amount of M6C seems to increase in the lower temperature range up to 1750F (954C) with increasing exposure time. At the higher temperatures above 1778F (970C) rather a dissolution of M6C seems to take place with increasing exposure time. An exception is found for the specimen which had been exposed for 10,637 h at 1800F (982C) which had shown the highest amount of M6C (3.9%). Higher amounts of Nb (1.2 mass-%) seem to promote the precipitation of nely dispersed NbC as eutectic structures evidently stabilizing the microstructure and suppressing the formation of M6C. After longer exposure times at temperatures of 1800F (982C) the amount of M23C6 increases after times higher than 8,000 h and

22

Atlas of Microstructures

at 1900F (1038C) after times higher than 2,500 h from about 8% to maximum 11%. The G-phase was observed only for one specimen of this material which was subjected to the lowest test temperature of 1675F (913C) for 659 h. At higher temperatures for specimens with comparable or even higher Si content the G-phase could not be detected with condence.

Atlas of Microstructures

23

Service-Exposed Samples
Below the history of the Service-Exposed Tubes is described in detail. Exposure conditions and chemical composition are summarized in Tables 7-8. Table 7: Exposure Conditions of Service-Exposed Samples Tube Skin Tube Skin Hoop Hoop Expo. Sample_ID Sample ID Temp. Temp. Stress Stress Time Donated Dechema MTI Alloy F C ksi MPa hrs by A234_050 SE-A HPNb 1630 888 2.12 14.60 108,306 MetalTek A234_051 SE-B HPNb 1607 875 2.12 14.60 108,306 MetalTek A234_052 SE-C HPNb 1650 899 0.72 4.96 75,000 Eastman A234_053 SE-D HPNb ~1650 899 1.16 8.00 148,920 Eastman A234_054 SE-E HPNb ~2350 1288 1.30 9.96 35,000 APCI A234_055 SE-F HPMA ~1800 982 1.79 12.30 65,664 APCI A234_056 SE-G HPMA 1673 912 2.04 14.00 94,900 APCI A234_057 SE-H HPNb 1661 905 1.98 13.60 131,000 Syncrude A234_058 SE-I HPNb 1661 905 1.98 13.60 143,000 Syncrude A234_059 SE-J HPNb 1661 905 1.98 13.60 143,000 Syncrude Table 8: Chemistry of Service Exposed Samples Sample_ID Sample ID Dechema MTI Cr Ni A234_050 SE-A 24.5 34.1 A234_051 SE-B 24.5 34.0 A234_052 SE-C 24.9 31.7 A234_053 SE-D 24.5 34.5 A234_054 SE-E 25.3 32.6 A234_055 SE-F 24.9 32.8 A234_056 SE-G 24.7 34.0 A234_057 SE-H 23.9 31.8 A234_058 SE-I 24.4 33.3 A234_059 SE-J 24.3 33.4

Fe bal bal bal bal bal bal bal bal bal bal

Mn 0.40 0.34 0.68 0.83 1.20 1.00 0.70 0.50 0.60 0.57

Composition (wt%) Si C 0.90 0.45 0.90 0.39 1.10 0.42 1.90 0.40 2.50 0.40 1.30 0.42 0.90 0.57 1.10 0.42 1.40 0.42 1.40 0.39

Nb 0.80 0.75 0.74 1.40 1.20 0.98 0.50 1.20 1.20 1.20

N 0.047 0.060 0.074 0.062 0.074 0.038 0.104 0.128 0.090 0.102

Other Ti,Zr & W<0.1 Ti,Zr & W<0.1 Ti,Zr & W<0.1 Ti,Zr & W<0.1 Ti,Zr & W<0.1 Ti & W>0.1, Zr<0.1 Ti,Zr & W<0.1 Ti,Zr & W<0.1 Ti,Zr & W<0.1 Ti,Zr & W<0.1

24

Atlas of Microstructures

Samples SE-A & SE-B: These samples were removed from the same furnace after approximately 108,306 hours of service. The available history indicated that the furnace had 53 cold start-ups and 70 hot restarts over the service life. Tube samples SE-A & SE-B were removed from dierent tube segments and elevations within the furnace resulting in a slight dierence in estimated tube skin temperatures (see Table 7). Post service analysis found that samples SE-A and SE-B both had maximum diametral growths of less than 1%. Ambient temperature tensile testing of the aged material found the following: SE-A 66.9 Ksi / 462 MPa 37.6 Ksi / 259 MPa 5.9% SE-B 65.8 Ksi / 454 MPa 40.1 Ksi / 277 MPa 3.0%

Sample SE-D was extracted from the failed tube at the same elevation as the through-wall rupture. However, the sample was removed 180 degrees around the circumference from the rupture in an eort to obtain a microstructure representative of long-term ageing. Sample SE-E: The HP modied reformer tube was in service for approximately four years prior to rupture (see Table 8). The ruptured area of the tube had a maximum diametral growth of only 2.3%. Metallographic analysis revealed numerous, tight interdendritic cracks running nearly through wall with minimal creep voids. Higher magnication examination revealed a lamellar-like phase adjacent to the interdendritic cracks. The tight intergranular cracks and lamellar phase are characteristic of incipient melting. Incipient melting is liquation of low melting point constituents along the dendrite boundaries. Incipient melting of cast HP alloys generally occurs at approximately 2350F (1288C) or 100F (56C) below the bulk melting temperature of approximately 2450F (1343C).

UTS Yield Strength Elongation

Sample SE-C: This sample was removed within 6 inches (15.2 cm) from the outlet end of an HP-modied tube. The tube had been in service approximately 75,000 hours with no operational anomalies. Sample was included to observe the microstructural changes at relatively mild operating conditions. Sample SE-D: This tube failed approximately four months after a scheduled plant outage. The failure was likely the result of debris inadvertently left in the tube during the outage. The debris caused channeling of the feed gas resulting in a local hot spot and subsequent rupture. Immediately adjacent to the rupture, the secondary carbides were completely dissolved with notable coarsening of the primary carbides. The microstructural features suggested that the metal temperature at the ruptured area of the tube exceeded 2012F (1100C).

Atlas of Microstructures

25

Chemical analysis found deciencies in micro-alloying and/or carbon concentrations in many of the tube segments that experienced large diametral growths. Remaining life assessment via the Omega creep strain methodology was performed nding the creep strain rates as least 60 times greater than the average for HP micro-alloyed material in the Materials Property Council (MPC) database. The alloying deciencies resulted in the material being highly strain tolerant, likely capable of 15% diametral growth prior to rupture, but with poor creep strength. The remaining life testing indicated that the material had essentially no remaining life at the normal operating conditions. Sample SE-G: The HP micro-alloyed tube was removed from the furnace for destructive remaining life assessment. The tube had been in service approximately 95,000 hours with a design life of 100,000 hours (typical operating conditions listed in Table 8). The tube had a maximum diametral growth of 5%. Metallographic analysis found the onset of creep void alignment but no creep ssure formation suggesting considerable remaining life. Remaining life assessment was performed via the Omega creep strain method. Despite the actual service life reaching the original design life, the measured strain rates and calculated Omega indicated the tubes have an estimated remaining life of nearly 17 years at the normal operating conditions. The estimated tube life decreased approximately 44% with each 25 F (14C) increase in tube metal temperature. Sample SE-H: This HP-modied sample was extracted from a tube removed from a furnace after approximately 131,000 hours of operation (typical operating conditions in Table 8). The sample was removed from the red side of the tube approximately 16 inches (40.6 cm) above the oor of the top-red, downow reformer. Eddy current inspection of the tube immediately prior to removal indicated that the tube had a 30% through-wall indication and 2.0-

Micrograph of ne, interdendritic cracks and lamellar-like phase. (oxalic acid electrolytic, 500X)

The plant had a history of operational problems that tended to crush the reformer tube catalyst and promote channeling of the feed gas. The tube rupture resulted from a short-term temperature excursion. This sample was included to show the microstructural features as a result of incipient melting. Sample SE-F: During an outage after 7.5 years of service, diametral growths in excess of 10% were measured at an isolated area in the furnace. Metallographic analysis revealed only aligned creep voids with no creep ssures despite the large diametral growth. During normal operation, the tube metal temperature was in the 1650-1690F (899-921C) range. The microstructural features (i.e. carbide coalescence) and review of historical operating data indicated that the tube metal temperatures likely exceeded 1800F (982C) at an isolated area of the furnace for approximately 6 months leading up to the outage.

26

Atlas of Microstructures

2.4% diametral growth. Radiography could not conrm the indication. Subsequent metallography revealed small, linear indications, especially at the tube ID, where creep voids began to coalesce and align. Creep voids were present throughout the cross section but with little to no creep void alignment. Remaining life based on metallography was likely considerable, perhaps another 50,000 hours at the current design conditions. Caliper and strapping measurements found up to 5.6% diametral growth over the nominal diameter. Upon removal from service, tube remaining life assessment was performed using the Omega creep strain methodology. Even though the tube had surpassed the design life by nearly 50%, the calculated Omega value and measured creep strain rate values indicated that the tube still had 20,000 to 60,000 hours of remaining life at the current operating conditions. The obtainment of life well in excess of the original design life was likely due to the operating temperature being approximately 45F (25C) lower than the 1710F (932C) design temperature. Samples SE-I & SE-J: Samples SE-I & SE-J were obtained from a tube removed from the same reformer as sample SE-H (above). However, samples SE-I & SE-J had increased service time of approximately 143,000 hours. Sample SE-I was extracted approximately 24 inches (61 cm) above the reformer oor while sample SE-J was extracted from the same tube but 84 inches (213.4 cm) above the reformer oor. The samples were included in an eort to observe any microstructural dierences resulting from the slight elevation changes within the furnace. Strapping and caliper measurements indicated that sample SE-I had a diametral growth of 2% while sample SE-H had 3.5% diametral growth. Tube metal temperature measurements taken at these two elevations indicated a dierence of 9 F (5C), consistent with the diameter change. Omega creep strain remaining life assessment has not yet been completed in these samples.

Atlas of Microstructures

27

Service-Exposed Samples
The results of the image analytical determination of the volume fraction of the precipitates for the service-exposed tubes of the centrifugally cast materials are summarized in Table 12. Table 12: Results of image analysis of Service Exposed Samples
Sample_ID Sample Material Tube Skin Temp. F 1673 1800 1607 1630 1650 1650 1661 1661 1661 2350 C 912 982 875 888 899 899 905 905 905 1288 Hoop Stress Ksi 2.04 1.79 2.12 2.12 0.72 1.16 1.98 1.98 1.98 1.30 MPa 14.0 14.0 12.30 14.60 4.96 8.00 13.60 13.60 13.60 9.96 Expo. Time hrs 94,900 65,664 108,306 108,306 75,000 148,920 131,000 143,000 143,000 35,000 area fraction in % M(C,N) 0.3 <0.2 0.6 1.6 0.1 M23C6 12.9 6.2 7.0 9.2 8.0 5.5 7.4 6.5 7.9 4.0 core M6C 1.6 1.5 0.8 2.2 2.0 2.2 2.5 M2(C,N) G-Phase + 2.0 1.2 0.4 1.0 0.5 1.5 1.6 + M(C,N) + + (+) 0.2 outer zone / edge M23C6 M6C + M2(C,N) G-Phase +++ ++ + +++ + +++ + + +++ ++ (+) +++ ++ ++ 4.6 6.2 +++ ++ + 7.9 6.2 7.8 6.5 + (+)

A234_056 A234_055 A234_051 A234_050 A234_052 A234_053 A234_057 A234_058 A234_059 A234_054

SE-G SE-F SE-B SE-A SE-C SE-D SE-H SE-I SE-J SE-E

HPMA HPMA HPNb HPNb HPNb HPNb HPNb HPNb HPNb HPNb

1.3

Annotations: (+) very few, + few, ++ many, +++ very many

28

Atlas of Microstructures

Alloy HPMA 1673F/94,900 h/2.04 ksi (912C/94900 h/140 MPa) This specimen contains a very large number of M23C6 primary and secondary carbides, Figures 2- SE-Ga to 2-SE-Gr. Their volume fraction amounts to 12.9%. Furthermore in the center of the tube wall as well as in the metal sub-surface zone carbides and nitrides of the type M6C and M2(C,N) are present. The sum of these carbides amounts to about 1.6%. The M(C,N) have evidently been almost completely dissolved or were oxidized and appear white in the cross-section which had been coated with ZnSe. Their fraction amounts to 9.3%. The G-phase could not be identied in this specimen. Besides the precipitates a large number of creep pores could be observed. In the BSE image the M2(C,N) appear bright due to the high Nb content, see Table in Figures 2-SE-Ga and 2-SE-Gb. The high amount of N (0.1 mass-%) could explain the formation of M2N, which contains nearly 50 atomic-% Nb. As Nb is bound to N it is not available for the formation of G-Phase. High N-contents seem to prevent G-Phase formation. 1800F/65,664 h/1.79 ksi (982C/65664 h/12.3 MPa) After operation for 7.5 years at 1800F (982C) the microstructure of the tube contains a high level of creep pores and shows M23C6 primary and secondary carbides all over the entire cross-section with a fraction of about 6.2%. In addition less than 2.0% G-phase due to the high Si content of 1.4 mass-% and the low N content of 0.04 mass-% as well as residues of M(C,N) with about less than 0.2% are observed, Figures 2-SE-Fa to 2-SE-Fr. Alloy HP Modied 1607F/108,306 h/2.12 ksi (875C/108306 h/14.6 MPa) After an operation time of more than twelve years at 1607F (875C) only a few creep pores are observed. The carbides are arranged in a distinctive eutectic structure, Figure 3-SE-Ba to 3-SE-Br. Besides about 7.0% M23C6 primary and secondary carbides about 1.2% G-phase and

Atlas of Microstructures

29

1.5% M6C are present in the microstructure. M6C and G-phase are located in contact with M23C6 carbides or are found in areas of secondary M23C6 carbides. Possibly a small fraction of the G-phase particles are rather M(C,N) particles which can, however, not be distinguished due to their small dimensions. In the outer sub-surface zone some M2(C,N) is present. 1630F/108,306 h/2.12 ksi (888C/108306 h/14.6 MPa) The microstructure of this part of the tube which had been exposed to a slightly higher temperature is signicantly dierent from that of the section exposed to 1607F. (875C) The primary and secondary carbides have changed their shape and have coarsened and the number of secondary carbides has decreased due to agglomeration, Figures 3SE-Aa to 3-SE-Ar. The section contains besides 9.2% M23C6 also about 0.4% G-phase, 0.8% M6C and 0.6% M(C,N) in the core. In the metal sub-surface zone some M2(C,N) is observed, however no G-phase could be identied, Figures 3-SE-Ak to 3-SE-Ar. The dierent microstructures of both samples, A and B, cannot be explained by dierences in composition as their compositions are almost identical, Table 8. 1650F/74,000 h/0.72 ksi (899C/74000 h/4.96 MPa) The microstructure of this specimen which was exposed for 8.5 years reveals a similar appearance to that of the specimen exposed for 108,306 h at 1630F. (888C) Both show except of a higher Si content nearly identical compositions, Table 8.The structure is characterized by a eutectic-type of precipitate arrangement in areas of residual melt and along grain boundaries, Figures 3-SE-Ca to 3-SE-Cr. Besides about 8% M23C6 primary and secondary carbides, 1% G-phase and 2.2% M6C were measured. The M6C particles are sometimes found in combination with M23C6carbides in the dendrites of the matrix. 1650F/148920 h/1.16 ksi (899C/148920 h/8.0 MPa) After seventeen years of exposure the microstructure contains a large number of creep cracks resulting from creep pores which had grown

together, Figures 3-SE-Da and 3-SE-Db. The M23C6 primary carbides have coarsened and secondary carbides have not been found. In the core area about 5.5% M23C6 carbides are present besides 1.6% M(C,N), Figures 3-SE-Da to 3-SE-Dj. The composition of this sample shows higher Si and Nb contents. In the sub-surface zone 6.2% M6C are found together with 4.6% M23C6 and the M(C,N) particles had been almost completely dissolved in M6C (0.2%), Figures 3-SE-Dk to 3-SE-Dr. 1661F/131,000 h/1.98 ksi (905C/131000 h/13.6 MPa) After about fteen years of exposure at 1661F (905C) this specimen contains about 7.4% M23C6 primary and secondary carbides of a coarse structure. Furthermore in the core area as well as in the metal sub-surface zone about 2.0% M6C, 0.5% M2(C,N) and 0.1% M(C,N) are present, Figures 3-SE-Ha to 3-SE-Hr. The G-phase was not observed although the composition of the sample shows high Si and Nb contents. Possibly the high N content leads to bonding of Nb by the formation of M2N. In the metal sub-surface zone increasingly Nb and Cr-rich M2(C,N) are present. 1661F/143,000 h/1.98 ksi (905C/143000 h/13.6 MPa) After slightly longer exposure of sixteen years under comparable conditions this specimen contains 6.5% M23C6 primary and secondary carbides, 2.2% M6C and 1.5% G-phase in the core, Figures 3- SE-Ia- to 3-SE-Ij. M(C,N) could not be detected. The composition of specimen I shows high Si and Nb contents. Although the N content is high, no M2N was formed. In the sub-surface zone no G-phase was present. The fraction of M23C6 primary and secondary carbides amounts to 7.9% and that of M6C to 6.2%, Figures 3-SE-Ik to 3-SE-Ir. 1661F/143,000 h/1.98 ksi (905C/143000 h/13.6 MPa) This specimen originates from the same tube as the two foregoing specimens (H and I), however from a higher position. No dierences in the microstructure were found compared to the specimen exposed for the same duration and same temperature shown before, Figures 3-

30

Atlas of Microstructures

SE-Ja to 3-SE-Jr. Moreover the volume fractions of the precipitates, as determined by image analysis, are almost the same as for the foregoing specimen. The chemical composition shows nearly identical contents, as well, Table 9. 2350F/35,000 h/1.3 ksi (1288C/35000 h/8.96 MPa) The microstructure of this specimen which had been exposed to the highest temperature over a duration of four years contains a high number of interdendritically oriented cracks which reach almost through the entire wall thickness. The precipitates are about 4% M23C6, mostly of the primary type, and 1.3% M(C,N) which was rich in Nb and also contained some Ti, Figures 3-SE-Ea to 3-SE-Ej. Locally clear indications of beginning melting was observed. In these spots a ne lamellar eutectic M23C6 structure was present which partially contained some M6C, Figures 3-SE-Ek to 3-SE-Er. In the outer sub-surface zone around the M23C6 particles and within the dendritic matrix some local depletion features were observed which resulted in local over-etching of the microstructure, Figures 3-SE-Es to 3-SE-Ez.

no M(C,N) can be observed. One tube peaks out due to its high amount of M(C,N), while the amount of M23C6 is low. In this tube no G-phase had been formed which may be due to its chemical composition. Indeed a high content of Nb besides a high Si content was found which however does not explain why the G-phase had not formed here. Another tube contains Nb-rich M2(C,N) instead of G-phase. Again a high N content was found by chemical analysis. It can be suspected that the formation of Nb-rich M2N is enhanced by high N contents. Again it can be suspected that M2(C,N) is a precursor stage of the formation of G-phase. In other words M2(C,N) is possibly dissolved after longer exposure times and G-phase is formed. Evidently at 2350F (1288C) a critical temperature has been reached at which the material HP Modied shows molten areas in the microstructure due to the formation of lowmelting phases which can reduce the creep strengths locally to a dramatic extent.

Summary
Composition of the Precipitates
In Table 13 the average compositions in atomic-% of the dierent types of carbides and phases are listed for the dierent alloys. The values of all elements except for C, (N), which were taken from the dierence of the sum of the determined metal fractions and 100%, rest upon several averaged spot measurements for all samples of each alloy by EPMA. Comparison of the results shows a good agreement within the accuracy of the measurements for each type of phase. The existence of M2(C,N) of dierent compositions in alloy HPMA und HPNb is remarkable. Both alloys show probably Cr-Nitrides and Cr- and Nbrich M2(C,N).

Conclusions
Among the two service-exposed tubes of the alloy HPMA only the one exposed at 1800F (982C) for 7.5 years contained, besides M23C6 and residues of undissolved M(C,N), some G-phase in an amount less than 2%. The tube which had been exposed for eleven years at the lower temperature of 1673F (912C) showed M6C and M2(C,N) instead of G-phase. The chemistry of this specimen shows a very high N content which obviously leads to the formation of Nb-rich M2N. G-Phase cannot be formed because not enough free Nb is present. Possibly the highly Nb-containing M2(C,N) represents a precursor stage of the formation of the G-phase. The HP Modied service-exposed tubes with the exception of the two tubes with exposure time higher than 75,000 h contain between 0.4 and 1.6% G-phase besides 7-9% M23C6 and 0.82.5% M6C. In microstructures containing more than about 1% G-phase

Precipitation Kinetics Diagrams


The development of the contents (area-%) of the dierent types of carbides and phases as a function of exposure time on a logarithmic

Atlas of Microstructures

31

Table 13: Composition of the precipitates in atomic-% (Mean values of several measurements) Type Alloy Cr Ni Fe Si Nb 35/45 63.7 1.1 2.4 0.0 0.0 M7C3 HPMA 58.9 1.1 7.0 0.0 0.2 HPNb 61.9 1.0 5.9 0.0 0.2 35/45 4.0 2.1 1.0 0.0 49.0 M(C,N) 35/45 3.1 1.3 0.5 0.1 8.6 HPMA 3.6 2.1 1.6 0.6 48.4 HPNb 2.8 1.2 1.4 0.1 49.9 35/45 68.9 3.4 3.3 0.0 0.1 M23C6 HPMA 64.0 3.4 8.9 0.0 0.1 HPNb 66.5 3.4 8.7 0.0 0.1 35/45 34.0 27.1 1.9 13.0 8.3 M6C HPMA 38.0 23.9 3,4 12.0 4.6 HPNb 31.3 25.8 3.6 13.1 10.2 35/45 5.1 49.4 2.9 22.0 18.9 G-Phase HPMA 2.7 47.3 5.3 22.9 20.0 HPNb 4.0 47.7 5.8 21.8 19.6 35/45 64.8 0.7 0.8 0.0 0.3 HPMA 62.6 0.3 1.5 0.0 0.3 M2(C,N) HPMA 32.0 1.8 3.0 0.0 29.8 HPNb 63.0 0.5 1.9 0.0 0.3 HPNb 32.2 1.3 2.5 0.0 31.5 Annotations: n. a. - not analysed

Ti 0.2 0.0 n.a. 3.5 30.5 2.1 0.1-1.1 0.0 0.0 0.0 0.2 0.0 0.0 0.5 0.4 0.0 n.a. n.a. 1.4 n.a. 0.1

Zr 0.0 0.0 0.0 0.0 9.4 2.1 1.6 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0.0 0.1 n.a. n.a. 0.0 0.0 0.0

C, (N) 32.6 32.8 30.6 40.1 46.8 39.5 42.9 23.9 23.6 21.1 15.5 18.0 15.9 0.2 0.9 1.2 33.4 35.0 32.0 34.4 32.4

32

Atlas of Microstructures

scale for dierent temperature ranges are represented in the diagrams in Figures 1-PKD-a to 1-PKD-d, 2-PKD-a to 2-PKD-c and 3-PKD-a to 3-PKD-d. In the plots, which reveal trends for the increase or decrease of the contents of the dierent precipitates, error bars for the standard deviation are drawn. As these trends are rather complex and the diagrams are not easy to understand in a further step for each temperature range the upper and lower limits for each type of precipitate were plotted in the form of scatter bands. The results are given in Figures 1-PKD-e, 2PKD-d and 3-PKD-e, where the scatter bands are hatched in dierent colors for dierent temperature ranges. These plots show the range of fractions of MC, M23C6, M6C and Gphase expected for aged alloys of the types 35/45, HPMA and HP Modied for dierent temperature ranges based on the results of the image analysis measurements.

and vice versa, cf. Figures 1-PKD-c and 1-PKD-d. The scatter bands of the area fractions of the precipitates for the dierent temperature ranges are shown in Fig. 1-PKD-e. From these scatter bands the user can easily extract the expected area fraction of a certain precipitate (MC, M23C6, M6C and G-phase) for the specied temperature ranges. It must be pointed out, that the fraction of MC strongly depends on the alloy composition and on the formation of M6C or G-phase, respectively. It is noticeable that the MC-content is lower or MC does not exist at all, in samples in which M6C or the G-phase are present and the that M6Ccontent is lower or M6C does not exist at all, in samples, in which the G-Phase is present.

Alloy HPMA
The growth of the M23C6-fraction with increasing exposure time for three temperature ranges is similar to the development of this type of carbides in alloy 35Cr45Ni, Figure 2-PKD-a. The scatter of the amount of M(C,N) is shown in Figure 2-PKD-b. In the actual case its fraction predominantly depends on the composition of the sample, because here only in one case G-Phase is present. Figure 2-PKD-c shows the change in the fraction of M6C for the dierent temperature ranges. No clear tendency can be derived. In comparison to alloy 35Cr45Ni less M6C (max. 1.7%) is present. Only one of the samples was containing the Gphase. The scatter bands of the area fractions of the precipitates for the dierent temperature ranges are shown in Fig. 2-PKD-d. As only one of the SE-samples was containing the G-phase, no scatter bands for the G-phase were drawn. Attention should be given to the dierent scale of the y-axis for the M6C-area fractions in comparison to alloy 35Cr45Ni (Fig.1-PKD-e).

Alloy 35Cr45Ni
The development of the M23C6-Carbides is shown in Fig. 1-PKD-a. The content of M23C6 increases the faster the higher the temperature. The content of M(C,N) strongly depends on the composition of the alloy, especially on the alloy content of the M(C,N) forming elements Nb and Ti. The plot of a tendency does not make much sense here, Figure 1PKD-b. For the formation of carbides of the M6C-type a similar trend is observed, Figure 1-PKD-c. The area fraction is up to 10.7% after exposure times of 5,000-10,000 hours at higher temperatures (1900-2050F, 1038-1121C). Higher temperatures and exposure times >1,000 hours seem to promote the formation of M6C, which is rich in Cr, Si and Nb. The G-phase is observed in the temperature range of about 1800-2000F (982C-1093C) after exposure times of 500 <T<10,000 hours, Figure 1-PKD-d. For the formation of the G-phase high Nb fractions are necessary. Due to diusion processes the NbC dissolves, which is shown by a decreasing content of the M(C,N). In addition a large fraction of Si and Ni is needed for the formation of the G-Phase. The supply of Ni and Si results from the matrix. It is remarkable that the M6C-content is lower or M6C does not exist at all, in samples, in which the G-Phase is present

Atlas of Microstructures

33

Alloy HP Modied
For this alloy the content of M23C6-Carbides is more stable during exposure time. No clear dependence on the temperature range is shown in Fig. 3-PKD-a and 3-PKD-e compared to 35Cr45Ni (Fig. 1-PKD-a and 1-PKD-e) and HPMA (2-PKD-a and 2-PKD-d). The maximum fraction of M23C6 is 11%. One sample shows a very low fraction because of dissolution processes due to a temperature near the melting point (red value (SE-E) in Fig. 3-PKD-a and the beginning of melting, marked with an arrow in Fig. 3-PKD-e). The distribution of the M(C,N) fractions shows two levels depending on the respective content of Nb. High Nb-contents 1.18 mass-% seem to result in high M(C,N) fractions (1.0-3.0 area-%). Low Nb contents (0.74-0.83 mass-%) seem to result in lower M(C,N) fractions (1.0 area -%). In Fig. 3-PKD-e two green hatched scatter bands representing M(C,N) are shown for the temperature range of 1787-1800F (970-982C). The formation of M6C initially shows an increase in its fraction (max. 4%) up to an exposure time of about 10,000 hours. After long exposure times >50,000 hours this type of carbides partly seems to dissolve, while simultaneously in certain cases G-Phase (max. 2%) is formed (Fig. 3-PKD-c and 3-PKD-e). G-phase was only observed in the temperature range of 875-955C for exposure times between 750- 150,000 hours (Fig. 3-PKD-d and blue hatched scatter band in Fig. 3-PKD-e). In this temperature range in addition to M6C the G-phase exists. For higher temperature ranges G-Phase does not exist at all whereas M6C is sometimes present (green and red scatter bands in Fig. 3-PKD-e). The Project Team that reviewed this document was surprised at the rapid conversion of MC to M23C6 carbides and at the absence of G phase in some samples. Following receipt of the manuscript Dr. Schtze oered some additional insights. This discussion is in the Addendum.

The nancial support by MTI and the very pleasant collaboration with Galen Hodge, MTI are gratefully acknowledged.

Literature
L. Dillinger and D. Roach, Metallurgy and Metallography of Reformer Alloys, Battelle Columbus Laboratories, May 1, 1975 2 Metallograe, 14. Edition, Editors H. Schumann and H. Oettel, 2005 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim, ISBN 3-527-30679-X 3 Atlas der Interferenzschichten-Metallographie, Editors H.- E. Buehler and H. P. Hougardy, 1979 Deutsche Gesellschaft fr Materialkunde e.V, Frankfurt am Main, ISBN 3-88355-015-9 4 I. Aydin, H. - E. Buehler, Contrasting the Carbides M23C6, M7C3 and M6C in High Temperature Resistant Cast and Wrought Alloys, Praktische Metallograe 17, 1980, no. 9, 441 446 5 I. Aydin, H. - E. Buehler and A. Rahmel, Precipitation in the heat resistant nickel-base cast alloys G-NiCr 28 W and G-NiCr 50 Nb, Arch. Eisenhuettenwesen 54, 1983, no. 11, 461 466 6 I. Aydin, H. - E. Buehler and A. Rahmel, Determination of optical constants of carbides and carbonitrides in refractory steels, Microstructural Science 9, 1981, 55 63 7 I. Aydin, H-E. Buehler, A. Rahmel, Observations concerning nitridation of refractory austenitic materials in air and combustion gases, Werkstoe und Korrosion 31, 1980, 675 682 8 ASM Handbook, Volume 9, Metallography and Microstructures, 1998, ASM International, Materials Parc; OH 44073 9 Dr. D. Flahaut, D. T. Pool, 35Cr45Ni Micro-alloy Friend or Foe, presented at: ERTC Petrochemical Conference, Vienna, 2004 10 John J. Homan, G.E. Gapinski, Properties and Microstructures of Outlet Manifold Components, Ammonia Technical Manual, 2002 11 R. C. Ecob, R. C. Lobb and V. L. Kohler, The formation of G-phase
1

Acknowledgement
34

Atlas of Microstructures

Addendum
1. Etchant Cookbook Dr. Vogels pickle: Vogelsche Sparbeize (German name) is a commercial pickle, named after its inventor Dr. Otto Vogel (Patents: DRP. 309264, 346530 und 350771; Brit. P. 158768, 188713;.) Dr. Vogels Sparbeize is a chemical solution whose exact recipe is not known for the public. It partially acts as an inhibitor. Known contents are: 20% H2SO4, 1% H3PO4, 40-50% 1methoxy-2-propanol, 5-7% nonylphenol-ethoxylate, 3-5% thio-uric acid. It can be purchased e.g. from the following supplier: Buehler GmbH In der Steele 2 D-40599 Dsseldorf http://www.buehler-met.de http://www.buehler.com Purchase order number : 170014 quantity: 0,25l price: 26,00 Safety data sheet: www.buehler-met.de/pdf/msds/170014.pdf Most probably there will also be an American Buehler subsidiary with the address given on the above .com web page. V2A etchant This is a traditional etchant dating back to the middle of the last century which had been developed for the etching of austenitic stainless steels. We took the composition from the chapter etchants of Ni and Ni-alloys: recipe m10 of the following German book. Metallographisches, keramographisches und plastographisches tzen 6. berarbeitete Auage [Metallographic, keramographic and plastographic etching. A

Cookbook.] Unter Mitarbeit von: Carle, Veronika. 1994. 6. Auage. XII , 320 Seiten, 23 Abbildungen, 21x15cm (Materialkundlich-Technische Reihe, Band 1) Publisher: Gebr. Borntraeger, Berlin ISBN 3-443-23014-8 brosch., EUR 26.00 Similar etchants may be described also in American textbooks for metallography. As a cooking guide you can use: Composition: take 100 ml distilled H2O add successively 100 ml HCl 1.19 M, 10 ml HNO3 1.40 M, 0.3 ml of an inhibitor e.g. Dr. Vogels Sparbeize stir up the mixture and heat up to 50C Etching: move the sample in the hot etchant for 25-30 sec by using a pair of tweezers 2. Rapid conversion of M7C3 to M23C6 We know from experience, that the conversion of M7C3 to M23C6 is rapid in other comparable alloys, too (see Table 2 and Fig. 12, Fig. 38 and Fig. 66). If you consider the Time-Temperature-PreciptationDiagramm of the cast alloy 1.4857 whose composition is similar to HP-MA and HPNb, you see that for a temperature of 900C the transformation of M7C3 to M23C6 is already completed after 20 hours of exposure, Fig. 38. For higher exposure temperatures the conversion is even faster. So it is not surprising that we didnt nd M7C3 and M23C6 in coexistence, because the exposure times and temperatures are outside the the band of conversion. The reaction of the conversion is 2M7C3 + 9M M23C6.

Atlas of Microstructures

35

Source: I. Aydin, Ausscheidungsverhalten einiger hitzebestndiger, chromhaltiger Gulegierungen auf Eisen und Nickelbasis, Dissertation, RWTH Aachen, 1982

36

Atlas of Microstructures

3. Precipitation of the G-Phase The G-phase is an intermetallic silicide of the type Ni16Nb6Si7. Many of the intermetallic phases do not show a broad range of compositions, some are even socalled line-phases. Presumably stoichiometry does not allow for major variations concerning the composition also in the case of the G-phase. Intermetallics precipitate if specic equilibrium conditions are fullled. Indeed the G- phase was found most frequently in the 35/45 alloy in a temperature range of 982C-1150C. Possibly the higher Ni-content in comparison to the HP-MA and HP-Nb alloys beside Nb and Si favours the G-phase formation. In the HP-MA alloy it was only found after exposure at 982C / 65664hours. This service exposed sample contains 1.3 wt-% Si and 0.98 wt-% Nb measured after exposure. For the HP-Nb alloy the G-phase is only present in the temperature range of 875-955C. The Si content of these samples varies between 0.9-1.7 wt-% and the Nb content between 0.74-1.2 wt-%. The service exposed HP-Nb samples show G-phase in a temperature range of 875-905C for long time exposure of 75000-143000 hours. The wall thickness of these samples is higher in comparison to the creep rupture samples. Probably the diusion of carbide and G-phase forming elements is much slower at this lower temperature and the diusion path of Si to the surface of the sample is longer in samples of larger dimensions. Si may be present for a longer time for the formation of Gphase than in thinner samples where it may be consumed by oxidation at the surface. For the G-phase formation the following informations are known from the literature:  Ageing of Nb-containing castings results in the replacement of Cr7C3 carbides by M23C6, secondary precipitates of M23C6 and NbC, with the replacement of some of the pre-existing NbC by the G-Phase  The initial volume fraction of NbC and its rate of transformation to G-Phase were found to increase with the Si content

 Silicon The higher the Si content, the higher is the upper limit temperature of formation for G-phase, and the faster is the kinetics of formation  Niobium The higher the content of Nb the higher is the content of Gphase. Only at very high contents of Nb (more than 1.9 wt-%) the G-phase may partially be formed directly from NbC  Titanium Ti addition transforms NbC into NbTiC which is more stable and less easy to destabilise to form G-phase. Ti therefore inhibits G-phase formation. Maximum eciency is 0.1% wt-% Ti  The following chemical composition (wt-%) is preferred in order to limit Gphase formation:

Min Max

C 0.40 0.45

Si 1.3 1.6

Ni 42 44

Nb 0.7 0.9

Source: Dr. D. Flahaut, D.T. Pool, 35Cr45Ni Micro-alloy Friend or Foe, presented at: ERTC Petrochemical Conference, Vienna, 2004

Atlas of Microstructures

37

38

Atlas of Microstructures

You might also like