Available online at www.sciencedirect.

com

Engineering Failure Analysis 15 (2008) 723–735

www.elsevier.com/locate/engfailanal

Mechanical strength and microstructural observations

for remaining life assessment of service exposed
24Ni–24Cr–1.5Nb cast austenitic steel reformer tubes
Jaganathan Swaminathan *, Prabha Prasad, Manoj Kumar Gunjan,
Krishna Gugloth, Prabir Kumar Roy, Raghubir Singh, Rabindranath Ghosh
National Metallurgical Laboratory, Materials Science and Technology Division, Burmamines, Jamshedpur, Jharkhand 831 007, India

Received 8 March 2007; received in revised form 16 June 2007; accepted 18 June 2007
Available online 30 June 2007

Abstract

Reformer furnace heater tubes made of cast austenitic alloy steels serve in the temperature range of 1073–1173 K and
pressure around 3–4 MPa. Though the design life is about 100,000 h (11.4 years), premature failures within 3–8 years
are more common, mainly caused by premature creep induced by overheating. The cast tube material undergoes microstruc-
tural changes with service exposure and as a consequence, degradation in strength. Microstructural changes, tensile strength
and accelerated stress rupture strength of 96,000, 105,120 and 131,400 h service exposed tubes of 24%Ni24%Cr1.5%Nb alloy
(equivalent to IN-519) served in a reformer furnace were analysed. Presence of coarsened carbides at the interdendritic
boundaries, absence of significant amount of secondary carbides in the intra-dendritic regions together with sigma (r) phase
at both inter and intra-dendritic regions and isolated creep cavities indicated severity of service induced ageing and damage
process accumulation. However, the stress rupture life obtained through Larson–Miller parametric method was good for
further life extension. Attempts to correlate the microstructural changes to remaining creep life assessment (RLA) proved
to be difficult as errors were mainly arising from the fact that temperature experienced by the individual tubes in a furnace
differs and for a single tube through out its vertical length it varies. Non-destructive ultrasonic scan method employed too
could not estimate the creep damage level accurately and the results vary with destructive test results. Lack of data on both
microstructural changes and mechanical strength of the service exposed tube materials makes RLA more difficult.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Remaining life assessment; Heater tubes; Creep; Austenitic alloys

1. Introduction

Primary reformer furnace unit of petrochemical plant produce hydrogen rich gases for the synthesis of
ammonia, methanol and other organic chemicals. The primary reformer furnace consists of catalyst filled

*
Corresponding author. Tel.: +91 657 2271709; fax: +91 657 2270527.
E-mail address: jsn@nmlindia.org (J. Swaminathan).

1350-6307/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2007.06.009
724 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735

heater tubes made of cast austenitic steel fired from outside. In plants using steam reforming reaction, naphtha
or natural gas and steam is passed inside the tubes and the ensuing endothermic reaction produces hydrogen
rich gases. The heater tubes experiences a temperature of around 450 °C at the top and around 800–900 °C in
the bottom zone. The heater tubes were designed for 100,000 h (11.4 years) of service but often fail prema-
turely within 3–8 years [1–13]. In many cases, the failure seems to be the result of overheating leading to pre-
mature creep [2–13]. The temperature on the outer side and inner side of the tube will be different due to the
endothermic reaction and for a single tube the temperature will not be uniform over its entire length and at the
same height inside the furnace, each tube may not experience the same temperature. Localised overheating
may occur due to catalyst collapse inside the tube or from burner misalignment. As such, for a given tube,
service induced microstructure changes may vary through out its length. Attempts to correlate remaining life
with microstructural degradation were found to be difficult as the microstructural changes were observed to be
non-monotonic [14]. Several of the non-destructive techniques too found lacking in evaluating the level of
creep damage and there is a difference in conclusion with regard to tube replacement based on NDT and
destructive stress rupture test results on the same tube piece [5–7,15]. Unlike that of thermal power plant mate-
rials, data on service exposed reformer tube materials are not much which further add to the difficulty in deci-
sion making for remaining life assessment (RLA). Our laboratory had received many heater tubes from
petrochemical industries for failure analysis and RLA [2–7]. The present paper discuses the results on mechan-
ical strength properties: hardness, tensile, accelerated stress rupture strength and microstructural observations
on service exposed 24Ni–24Cr–1.5Nb alloy steel tubes used in primary reformer furnace of a fertilizer plant.

2. Technical background

The primary reformer furnace is a Halder–Topsoe designed single row reformer having an induced draught
box heater and vertically installed catalyst tubes placed in two radiant chambers designated as AB and CD
(each with 144 nos. of tubes) and loaded with 35.6 m3 of catalyst. The heater tube is centrifugally cast and
each vertical tube in the furnace is made of three tube pieces welded together to form a height of
11,370 mm. The O.D. of each tube is 152 mm and the I.D. is 117.6 mm. Each radiant chamber is provided
with 288 numbers of forced draught burners. Naphtha and steam mixture preheated to 778–823 K is passed
downwards through the heater tubes. The operating skin temperature of the tube is 1153–1163 K and the out-
let process gas temperature is 1033 K and the flue gas temperature is 1293–1303 K. Operating pressure is
34 kg/cm2. The tubes were in service for 131,400 h (15 years). After 96,000 h of its service five tubes of section
CD were reported to be removed due to sagging and one tube of section AB was removed due to overheating
caused by catalyst collapse.
Portion of tubes removed after service exposure for 96,000, 105,120 and 131,400 h were tested for remaining
creep life assessment (RLA) at our laboratory. The tubes removed after 105,120 and 131,400 h service expo-
sure were cut off based on ultrasonic scan results during annual turnaround. The tubes section graded as ‘C’
(needing replacement in 2 years) was cut off for RLA. The heater tube material is 24Cr24Ni1.5Nb steel known
by its trade names: KHR24C, IN-519 and H24W. The specified chemical composition (in wt.%) and the values
obtained using Worldwide Analytical Systems AG make alloy analyser model PMI – Master working on the
principle of emission spectrophotometry are presented in Table 1. The chemical composition of the reformer
tube material under investigation conforms to the chemical composition specification for the grade.

Table 1
Chemical composition (in wt.%) of the reformer tube alloy
Element Specified Analysed
C 0.25–0.35 0.30
Si 1.5 (max.) 0.80
Mn 2.0 (max.) 0.73
Cr 23–26 25
Ni 23–26 23.6
Nb 1–2 1.5
Fe bal. bal.
 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735 725

3. Visual examination

The visual examination of 96,000, 105,120 and 131,400 h service exposed tubes revealed normal appearance
on both the inner and outer surfaces. There is no indication of carbon deposition on the inner surfaces of the
tube. The oxidation on the outer surface is not significant enough to take a note. Even though ultrasonic scan-
ning indicated (Level C) creep damage, no O.D. expansion was observed.

4. Metallography

Detailed optical and scanning electron microscopic studies were carried out only on metallographic spec-
imens cut from the 131,400 h service exposed tube. Metallographic samples were polished using a series of SiC
papers and finished with diamond polishing. Repeated polishing-etching cycle was carried out to eliminate
polishing artefacts. Samples were etched with Murakami’s etchant (50 g of K3Fe(CN)6 (potassium ferricya-
nide) + 50 g KOH (potassium hydroxide) + 100 ml distilled water at 348 K) to selectively reveal the sigma
(r) phase and with Glycergia etchant (30 ml HNO3 (nitric acid) + 10 ml HCl (hydrochloric acid) + 30 ml
Glycerol) to reveal the general structure with carbides.Meiji make optical microscope with Clemex image ana-
lyser software was used for optical microscopy and JEOL – SEM with Thermotron EDAX system was used
for electron microscopy.
Optical micrographs of 96,000 and 131,400 h service exposed tube samples etched with Glycergia etchant
are presented in Fig. 1. The outer wall region of both the tube samples showed columnar dendritic grains and
the inner wall region showed more equiaxed dendritic grains structure. The semi-continuous network of pri-
mary carbides at interdentritic boundaries shows some degree of coarsening and presence of secondary car-
bides in the intra-dendritic regions were less and appeared as bulky due to coarsening. Comparatively, in
the 96,000 h exposed sample, the degree of coarsening of the primary carbides is less as evident from the

Fig. 1. Optical micrographs of transverse section of service exposed sample etched with Glycergia (a) and (b) 131,400 h exposed (c) and (d)
96,000 h exposed. OW – near outer wall, IW – near inner wall.
726 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735

significant presence of lamellar eutectic of primary carbides, further, secondary carbides were almost absent in
the inner wall region and a few were present near the outer wall region. For the 131,400 h exposed sample, no
microcracks were observed but a few isolated creep voids were observed which needed SEM observation to
confirm.
Optical micrographs of the 131,400 h service exposed tube sample etched with Murakami’s etchant are pre-
sented in Fig. 2. Presence of fine, blocky as well as and coarsened needle like r phase precipitates was well
revealed at inter and intra-dendritic regions. The r phase content at the intra-dendritic region was less com-
pared to the interdendritic boundaries.
SEM observations on the 131,400 h service exposed tube samples etched with Glycergia etchant are pre-
sented in Fig. 3. The primary carbides at the interdendritic boundaries were shown to have coarsened from
their initial lamellar form and they were present as a semi-continuous network. The secondary carbides at
the intra-dendritic region were less in amount and whatever present appeared as blocky confirming the coars-
ened nature. The precipitate phases at inter and intra-dendritic grain boundaries were analysed through energy
dispersive – X-ray analysis (EDX) and are presented in Figs. 4 and 5, respectively. The spot analysis results
with their limitations in accuracy can give a fair indication of the type of precipitate phase present. The darker
shade phase was found to be rich in Cr and Fe and the light grey phase was rich in niobium. Amount of silicon
in the light grey phase is less to consider it as a Ni–Nb–Si phase. Cr-rich and Nb-rich phases co-exist at the
interdendritic grain boundaries, where as at the intra-dendritic regions they were present as isolated precipi-
tates. At the interdendritic boundaries, the light grey phase is present both as fine and coarser form. The dar-
ker phase is present only as coarser and formed a semi-continuous network. In the intra-dendritic region the
light grey phase is finer than the darker phase. This may be due to the higher growth rate of Cr-rich carbides
than the Nb-rich carbides. Several fine microcracks were observed (Fig. 6) in the Cr-rich carbide precipitate
phase at the interdendritic boundaries. Such type of cracking was also observed in the Cr-rich carbide of ser-
vice exposed HP grade alloys [12]. No microcracks were observed at the Nb-rich light grey phase. Isolated
microvoids were seen at the interdendritic carbide matrix interface. Presence of voids indicates that creep dam-
age was setting in the tube sample. Sigma phases and the Ni–Nb–Si phase were reported to promote embrit-
tlement in reformer alloys [12,16,17]. Observed tensile ductility for the service exposed samples indicate that
the embrittlement is not severe in the operating temperature range but significant at higher temperatures. In
the 131,400 h service exposed tube sample, Ni–Nb–Si rich phase was not observed and no reports were avail-
able on observation of Ni–Nb–Si rich phase in the service exposed 24Cr–24Ni–1.5Nb alloy.
SEM observation on the 131,400 h service exposed tube samples etched with Murakami’s etchant to reveal
the r phase precipitates is presented Fig. 7. The precipitate phases from which EDX spectra were taken were
marked by number and the SEM-EDX results for these precipitates were presented in Fig. 8. The precipitates
were rich in Fe, Cr and Ni with little Nb. Both fine and coarse r phase precipitates were observed at inter and
intra-dendritic regions. Appearance of Ca is due to impurity pick up while preparing the specimens.

Fig. 2. Optical micrographs of transverse section of 131,400 h service exposed sample etched with Murakami’s etchant. OW – near outer
wall, IW – near inner wall.
 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735 727

Fig. 3. Secondary electron micrographs of 131,400 h service exposed tube sample etched with Glycergia etchant. Transverse section – near
outer wall. Arrow indicates creep voids. Location from which EDX spectra obtained were indicated by arrow text.

5. Mechanical strength properties

5.1. Hardness

Hardness was measured on the tube cross section using a Vickers hardness tester and applying a load of
20 kg. The harness values obtained are as given in Table 2. Compared with the standard specification, it
appears that the hardness did not get degraded significantly up on service exposure. The observed difference
in hardness among the tubes investigated could be the result of difference in temperature experienced by the
individual tubes inside the furnace chamber.

5.2. Tensile strength properties

Standard tensile specimens with dimensions – M12 thread and 5 mm gauge diameter were machined out
from the tube samples and tested at room temperature and at high temperatures using INSTRON 8562
(100 kN capacity) servo electric material testing machine. The tensile properties: 0.2% proof strength (0.2%
PS), ultimate tensile strength (UTS) and elongation % obtained are given in Figs. 9–11 respectively. Compar-
ing the 0.2% PS (Fig. 9), the samples service exposed for 105,120 and 131,400 h showed a higher degradation
than the 96,000 h exposed sample. Marked reduction in 0.2% PS from standard specification was observed for
the 96,000 h exposed sample only at temperatures above 1123 K. Among the 105,120 and 131,400 h exposed
samples, the 0.2% PS was slightly higher for the later even as the exposure time is higher. Comparing the UTS
(Fig. 10), while all the service exposed samples showed strength reduction, the longest service exposed sample
(131,400 h) showed slightly higher value than the 105,120 h exposed one. The tensile ductility measured
(Fig. 11) in terms of elongation % also showed a similar variation in the trend. The tensile ductility for the
service exposed samples at room temperature was less than that of the specified minimum value. At higher
test temperatures, ductility had improved and the 96,000 h service exposed tube sample had shown higher
728 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735

Fig. 4. SEM-EDX results of the precipitate phases shown in Fig. 3b.

ductility than the standard values of the manufacturer. The ductility of the 105,120 h service exposed sample at
the higher test temperatures were closer to the specified minimum value compared to 131,400 h sample which
showed a reduction at temperatures above 1173 K. Improvement in tensile ductility for the service exposed
reformer tube alloys was observed in many of our investigations. Generally one expects a progressive strength
reduction with increase of service exposure time. The observed variations in tensile strength behaviour in the
present case are possible if the operating temperature experienced by the individual tubes vary from one tube
to another. Localised variations in temperature inside the furnace may arise due to burner problems or by heat
flow disturbances. The observed facts clearly indicated the need to exercise care while selecting the samples
and the possible difficulty that may arise out in evaluating the service exposed material for RLA.
 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735 729

Fig. 5. SEM-EDX results of the precipitate phases shown in Fig. 3c.

5.3. Accelerated stress rupture strength

Stress rupture specimens were cut from the bottom side of the supplied tube. 1/200 threaded specimens were
machined out and tested in ATS make single specimen lever arm creep testing machine with 15:1 lever ratio.
The test temperature was maintained with in ±2 °C of the set point. Test temperature and load was so chosen
to give a rupture time of 1000 h. Some tests which exceeded 1000 h were discontinued.
In the absence of standard rupture life data for the 24Cr24Ni1.5Nb alloy, a reputed manufacturer’s data
(minimum limit of the scatter band) was taken as the standard life data. Remaining creep rupture life estima-
tion was made using Larson–Miller parameter (LMP) by substituting the stress rupture time obtained at each
test temperature in the LMP equation, P = T(13.6 + log tr) where, P is the LMP parameter, tr is the stress
730 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735

Fig. 6. Secondary electron micrograph of 131,400 h service exposed tube sample showing cracks (circled) in the Cr-rich interdendritic
carbide phase. Etchant: Murakami’s.

Fig. 7. Secondary electron micrograph of 131,400 h service exposed tube sample etched with Murakami’s etchant. Numbers denote the
precipitates which were analysed using SEM-EDX.

rupture time in hours and T is the temperature in Kelvin. The LMP plot for the service exposed tube samples
was shown in Fig. 12. Comparing the data in the LMP plot, sample exposed for 105,120 h was better than the
sample exposed for 96,000 h even as the service exposure is higher by 9120 h. This highlights the possibility of
tube to tube variation in temperature experienced and its effect on life prediction. Such variation may also
arise if the test specimens were cut from the tube sample side which is not directly facing the burner. Hence
a check was done for the 105,120 h service exposed tube sample by cutting samples at 90° intervals and car-
rying out tensile test at 1173 K and stress rupture test at 1173 K/36 MPa. While the tensile results did not
show any difference in strength, stress rupture time showed a difference alone as noted in Table 3. The data
point for the south side (burner facing side) was also presented in the LMP plot (Fig. 12). The observed results
clearly indicated the effect of sampling and need for care while estimating the remaining creep life for life
extension.
The stress rupture tests on the 131,400 h service exposed tube were done using the specimens cut from the
burner facing side. The data points for this tube sample fall on the standard data line up to a temperature of
1173 K beyond which the deviation from the standard data was increasing indicating increased rupture
strength degradation at higher temperatures. This indicates the need to adhere to the permitted operating tem-
perature range. Any over temperature excursion will deteriorate the creep strength drastically for the service
exposed alloys.
Even though LMP plot for 105,120 h exposed tube sample showed good rupture life in the operating tem-
perature range, life extension was recommended for a period of three years (approx. 26,000 h) only and
 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735 731

Fig. 8. SEM-EDX results for the precipitates shown by marking in Fig. 7a: (a) precipitate 1; (b) precipitate 2 and (c) precipitate 3.
732 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735

Table 2
Vickers hardness (Hv20) test results for the service exposed reformer tubes
Exposure time (h) Inner side Mid-section Outer side
96,000 182 184 191
105,120 230 225 232
131,400 232 210 228
Specified (minimum) 180 180 180

Fig. 9. Plot of 0.2% proof stress variation with test temperature for the different service exposed tube samples.

Fig. 10. Plot of ultimate tensile stress variation with test temperature for the different service exposed tube samples.

advised to do recheck after that period. This decision was taken after noting the higher rupture life observed
even though the exposure time is higher and because of the highlighted possibility of tube to tube variation in
temperature experienced and within a tube sample, variations with respect to specimen locations. Based on the
LMP plot of 131,400 h service exposed sample which showed good rupture strength in the operating temper-
ature range, life extension to another 3 years was recommended. A note of caution was also added to the
observed increased degradation at temperatures above 1173 K.
 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735 733

Fig. 11. Plot of elongation percentage variation with test temperature for the different service exposed tube samples.

Fig. 12. Larson–Miller parametric plot for the service exposed tube samples super imposed on the standard manufacturer data.

Table 3
Accelerated stress rupture tests at 1173 K/36 MPa on samples cut at 90° intervals from 105,120 h service exposed 24Cr–24Ni–1.5Nb alloy
tube
Location in the tube Rupture time
North side 1036a
East side 659
South side (burner side) 400
West side 766
a
Test discontinued.

6. Discussion

In the present case of service exposed tube samples, service induced ageing had led to the coarsening of inter
and intra-dendritic carbides. The initial lamellar carbide form had coarsened but the coarsened carbide
network is not continuous indicating that the tubes might not have seen overheating to a significant level.
734 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735

Secondary carbides are few and coarsened. While iron rich r phase is present, Ni–Nb–Si rich phase is not
detected. Tawancy et al. [8] observed Ni–Nb–Si rich Laves phase in IN-519 tube materials exposed to
1223 K. Shinoda et al. [16] observed M23C6 and NbC for the as cast HK40 steel modified with various amount
of Nb additions. Formation NbC, M23C6 and Fe–Cr–Ni–Si rich sigma phase was observed only at Nb/C atom
ratio 0.54 (Nb–1.95 wt.% and C–0.47 wt.%) For Nb/C ratios 0.42, 0.5 and 0.88 the Fe–Cr–Ni–Si sigma phase
was not detected either by X-ray, optical or by EPMA analyses. The 24Cr–35Ni–1.5Nb2Si (HP grade) and Ti
microalloyed HP grade alloys were reported to show Ni–Nb–Si in the service exposed conditions [2,12,17].
The Ni–Nb–Si phase was explained as forming through transformation of NbC at service temperatures above
1123 K. The absence of Ni–Nb–Si rich phase in the present 24Cr–24Ni–1.5Nb alloy may be due to the lower Si
(0.8%) and Ni (23.6%) levels.
Isolated creep void formation at the carbide/matrix interface indicated the creep damage initiation but the
rupture strength is still good for life extension. IN-519 and HP-microalloyed grade tube samples with iso-
lated creep voids after service exposure for were found to have good stress rupture life [2,6]. Less continuous
nature of the moderately coarsened carbide network will make the coalescence of creep voids to form micro-
cracking, a difficult process. Observed decrease in tensile and stress rupture ductility at higher test temper-
atures above 1173 K clearly demonstrated the embrittled nature of the alloys after service exposure to
131,400 h.
For the 30,000, 44,000, 72,000 and 95,000 h exposed IN-519 alloy tube, Lebanowski [14] observed that the
microstructural degradation was not progressively increasing with operating time. An important point to be
noted in this observation is that the test samples are not from the same tube and not from the same level from
the inlet. Due to burner problems and localised flow conditions inside the furnace chamber and inside the indi-
vidual tube, each tube may experience different temperatures even at the same level of height in the furnace.
Hence, it is necessary for the plant to have the temperature of individual tubes be recorded at 2–3 locations in
the hottest zone. Samples for RLA should be cut from this region for destructive testing. Many times it was
observed [15] that at the normal operating conditions creep damage develop at the inner wall regions for the
advanced class of reformer alloys like HP-microalloyed. This makes ultrasonic method of looking at creep
damage in the mid wall region unreliable. Ultrasonic scan method to decide for tube replacement needed a
careful look is clearly demonstrated by the fact in the present case, the stress rupture properties are good
for life extension beyond that concluded from UT scan results. Measurement of tube diameteral expansion
is one good method to arrive at the selection of worst tubes for destructive testing but one should carefully
measure and record the diameter before the installation and at regular intervals. The level of expansion that
can be permitted depends on the grade of the alloy as each type of alloy shows different ductility which further
got modified due to service exposure. Attempts to correlate the microstructural changes in alloys used for
reformer tube application with ultrasonic characteristics were not known. Since the Cr–Mo alloys and some
stainless steels were shown to have good correlation in this regard, it is worth to attempt. Till then, the best is
the destructive method of accelerated stress rupture testing.

7. Summary

Cast 24Cr–24Ni–1.5Nb reformer tube alloy served for 96,000, 105,120 and 131,400 h were analysed for ten-
sile and stress rupture strength. Progressive degradation of strength was clearly observed mainly due to var-
iation in temperature experienced by the tube. Microstructural observation for the 131,400 h exposed sample
showed coarsened primary carbides, sigma phase. Secondary carbides were almost absent and isolated creep
voids at the carbide/matrix interface were present. But, stress rupture life indicated good amount of life for
further life extension. Observed destructive test results differed from ultrasonic scan method evaluation. Care-
ful tube sample selection and test specimen location is needed for reliable destructive test results.

Acknowledgment

The authors thank Director, National Metallurgical Laboratory, Jamshedpur, India for his kind permis-
sion to publish this paper.
 J. Swaminathan et al. / Engineering Failure Analysis 15 (2008) 723–735 735

References

[1] Shipley DG. Int J Pres Ves Piping 1983;14:21–34.

[2] Project report No. CIEP/2006/SSP 9302, National Metallurgical Laboratory, Jamshedpur 831007, India, 2006.
[3] Swaminathan J, Raghubir Singh, Das SK, Das G. In: Baldev Raj, Choudhury BK, Albert SK, editors. Pressure vessels and piping:
manufacturing and performance, in press.
[4] Ashok Kumar Ray, Sinha SK, Tiwari YN, Swaminathan J, Das G, Chaudhuri S, et al. Eng Fail Anal 2003;10:351.
[5] Project report No. CIEP/SSP/2001/0115, National Metallurgical Laboratory, Jamshedpur, India, 2001.
[6] Project report No. CIEP/SSP/2003/0141, National Metallurgical Laboratory, Jamshedpur, India 2003.
[7] Report No.CIEP/SSP/1999/077, National Metallurgical Laboratory, Jamshedpur, India, 2000.
[8] Tawancy HM, Ul Hamid A, Mohammed AI, Abbas NM. Failure analysis of catalytic steam reformer tubes. Anti-Corros Method
Mater 2005;52:337.
[9] Hamid AU, Tawancy HM, Mohammed AI, Abbas NM. Eng Fail Anal 2006;13:1005.
[10] Tawancy HM, Ul Hamid A, Mohammed AI, Abbas NM. Mater Des 2005.
[11] Shariat MH, Faraji AH, Ashraf Riahy A, Alipour MM. J Corr Sci Eng, 6 Paper H012, www.UMIST.org.
[12] Shu Shi. Doctoral Thesis, Ohio State University, 2006.
[13] Le May I, da Silveira, Vianna CH. Int J PresVes Piping 1996;66:233.
[14] Lebanowski J. Adv Mater Sci 2005;5:14.
[15] Swaminathan J, Raghubir Singh in Fourth Natl. Conf. on Creep, Fatigue and Creep–Fatigue Interaction. October 8–10, 2003, Indira
Gandhi Centre for Atomic Research, Kalpakkam, India.
[16] Haro RS, Lopez LD, Velasco TA, Viramontes BR. Mater Chem Phys 2000;66:90.
[17] Tawancy HM, Abbas NM. J Mater Sci 2005;40:2085.