Engineering Failure Analysis 118 (2020) 104918

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

An investigation on corrosion failure of a shell-and-tube heat

T
exchanger in a natural gas treating plant
H. Panahia, A. Eslamia, , M.A. Golozara, A. Ashrafi Lalehb
⁎

a
Department of Materials Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran
b
Pardis Petrochemical Complex, Pars Special Economic Energy Zone, Asaloye 196-75391, Iran

ARTICLE INFO ABSTRACT

Keywords: This study investigates failure of a shell-and-tube heat exchanger used in a natural gas treating
Amine plant. The shell was made of carbon steel grade A516, while its interior wall was cladded by
Cracking grade 304 stainless steel. Tubes and tube-sheet of the heat exchanger were made of grade 304L
Crevice corrosion austenitic stainless steel. After almost 6 years of service, tube failure was observed adjacent to the
Shell-and-tube heat exchanger
tube-sheet, which caused gas leakage, and consequently an increase in internal pressure of the
shell. Detailed examinations were performed on the failed tubes. Results showed that crevice
corrosion formed within the gap between tube external surface and tube sheet, followed by
subsequent cracking was the main cause of failure. Mitigation strategies were suggested for fu-
ture failure prevention.

1. Introduction

Heat exchangers are crucial equipment in petroleum refining [1,2]. They provide a noticeable reduction in energy consumption
since it exchanges heat from a hot stream to a cold stream. Just like all other industrial equipment, heat exchangers fail due to several
mechanisms [3]. Corrosion is the most known cause of failure in all types of heat exchangers, especially for shell-and-tube exchangers
[1,4–6]. Tube obstruction, leakage, and decline of heat transfer are results of corrosion in this specific type of exchanger. Crevice
corrosion is a localized corrosion known as a source for numerous industrial failures [7,8]. This type of corrosion happens when there
is a gap, wide enough for liquid media to penetrate inside the gap (crevice), also sufficiently tight to be motionless [9]. Previous
studies proved that most of the passive metals (e.g. austenitic stainless steels and aluminum alloys) failed by crevice corrosion due to
localized concentration cell formation inside the crevice [9,10]. Tube-sheet and baffles are susceptible zones for crevice corrosion in
shell-and-tube exchangers. These parts have holes which tubes inserted through them and have probable remained gaps after the
expansion process.
Various methods can be used for expanding tubes inside tube-sheet, which are based on expansion pressure and residual contact
pressure calculations [11]. Welding is an alternative for expansion of fixing tubes inside the tube-sheet; however, it has its own sets of
problems, particularly formation of residual stresses and subsequent cracking due to thickness difference between tubes and tube-
sheet [12]. Expansion process is strongly in concern, since if done inappropriately it may cause serious problems. In this regard,
failure of a new heat exchanger in dimethyl ether to propylene process was investigated which occurred during its first start-up
operation [4]. Results showed that Stress Corrosion Cracking (SCC) caused the failure. Residual stresses at seam weld had initiated
the SCC at the expansion joints. In addition, chloride contaminated pre-heating steam, facilitated the SCC. In addition to welding

⁎
Corresponding author.
E-mail address: [email protected] (A. Eslami).

https://doi.org/10.1016/j.engfailanal.2020.104918
Received 17 March 2020; Received in revised form 27 July 2020; Accepted 11 September 2020
Available online 14 September 2020
1350-6307/ © 2020 Elsevier Ltd. All rights reserved.
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

induced residual stresses, forming induced residual stress are detrimental mostly at exists of U-tube bundles. Theses stresses might
cause failure in either tube-sheet or tubes after the expansion process such that Xu et al. reported [13]. Miscalculations in amount of
required stress for expansion process might also be another cause of failure. In this regard, according to a case study carried out by Sui
et al. [14], inappropriate expansion with lower pressure than calculated amounts resulted in long narrow gaps, which were con-
sidered to be susceptible sites for SCC in presence of chloride ions. Also, in this regard, Hydraulic expansion of tube-to-tube-sheet
joints for shell-and-tube heat exchangers was investigated by Ma et al. [15]. Various expansion procedures, and pull-out and push-out
forces were evaluated. Based on their study, increase in expansion percentages was not beneficial and over-expansion could cause of
leakage at the joints. Expansion percentage was a function of tube size and material.
Amine solutions are not considered as inherent corrosive solutions [16–18]. In fact, these solutions contain various chemical
reactants reducing their pH, or can get contaminated by aggressive ions such as Chlorine, during the natural gas treatment cycles
[19,20]. These phenomena’s increase the corrosiveness of amine solutions for steels [21]. Corrosion in amine solutions is more severe
for carbon steels than stainless steels, considering the fact that they do not form an effective passive layer [22]. However, presence of
aggressive ions such as Chlorine in amine solutions as impurities dramatically can increases the localized corrosion of stainless steel
components used in natural gas treating plants [23].
Although there has been some researches investigating failure of heat exchangers, it should be mentioned that each equipment
has its own failure scenario due to its complicated service condition might. Therefore, the purpose of this study is to investigate the
premature failure of a heat only after 8 years of service, while it was expected to operate at least for 15 years.

2. Background

2.1. Process description

In order to absorb acidic gases (CO2 and H2S), Activated Methyl Diethanol Amine (aMDEA) solution and natural gas must enter in
an absorption tower. Amine solution saturated by acidic gas, known as rich amine solution must be heated after the absorption
process, and then sent to the stripper tower. In the stripper tower, hot rich amine solution releases absorbed acidic gases and becomes
lean amine during a solution recycling process [1]. Shell-and-tube heat exchanger is an equipment that increases the temperature of
amine solution before entering the stripper tower. Hot, pressurized natural gas enters in this exchanger through tubes and cold amine
solution enters inside the shell, so natural gas (so-called synthesis gas) increases solution temperature as it cools down. Table 1
summarizes working condition of the shell-and-tube heat exchanger.

2.2. Failure description

A shell-and-tube heat exchanger failed only after 8 years of service due to leakage from the tubes, while this equipment was
expected to operate for at least 15 years of service. Circumferential corrosion observed on the outer surface of tubes where placed
inside the tube-sheet. Leakage of natural gas from tubes increased the internal pressure of the shell and caused explosion. Materials
specification of the failed shell-and-tube heat exchanger is shown in Table 2.
Fig. 1 shows a schematic of the failed heat exchanger and a section of the tubes inside tube-sheet. Fig. 2 represent a sample of
observed gap between tubes and tube-sheet.

3. Experimental procedure

3.1. Visual examination

Fig. 3 shows images from tube samples of the tube-shell exchanger. As shown in this figure, a black scale has covers external
surface of the tubes. Also, wall thinning has occurred at the outside of the tubes accompanied by cracking at the location with
minimum thickness. As seen from Fig. 3 there are barrier like lines on the outer surface of all of the failed tubes as shown by arrows.
The interior surface of the tubes is clean, without any corrosion damages. The average profile of the outer diameter of the corroded
tubes is shown in Fig. 4.
Failed tubes were cleaned in acetone in an ultrasonic bath, then examined by an OPTIKA stereomicroscope Model SZM-2. As can
be seen from Fig. 5, the external surface of the tube is rough. In addition, some pits can be observed on the rough surface. Further
examination was carried out using a Scanning Electron Microscope (SEM), Model Philips X230.

Table 1
Shell-and-tube heat exchanger's working conditions.
Parameter Inside shell Inside tubes

Pressure (Bar) 0.99 32.34

Temperature (°C) Inlet 123 164
Outlet 123 136

2
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Table 2
Specifications of failed shell-and-tube heat exchanger.
Item Shell side Tube side

Shell material AISI 304L AISI 304L

Tube material AISI 304L AISI 304L
Tube-sheet AISI 304L AISI 304L
Operating fluid aMDEA Natural gas

Fig. 1. Schematic of the failed shell-and-tube heat exchanger.

Fig. 2. Camera image showing a gap formed between tube and tube-sheet of the shell and tube failed exchanger.

3.2. Microscopic examination

Fig. 6(a)–(c) show SEM images from the surface of failed tubes. As seen from Fig. 6(a), there is a clear boundary between severely
corroded and uncorroded areas. In addition, pitting corrosion is visible from image (b) and (c) at different magnifications on the
corroded area, in accordance with the stereomicroscope image shown in Fig. 5.

3.3. Chemical composition of the tube materials

Table 3 shows the chemical composition of the tube sample from heat exchanger, performed by a Hitachi optical emission
spectrometer, Model Foundry-Master Pro2. The chemical composition of the failed tube is in accordance with AISI 304L standard
[24].
Energy-dispersive X-ray spectroscopy (EDS) tests were performed on the surface of the failed samples to investigate micro-scale
chemical analysis. Fig. 7(a)–(c) shows the results of the analysis.

3
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Fig. 3. Samples of failed tubes (a) circumferential corrosion and cracking (b) corrosion and cracking on one side of the tube.

Fig. 4. Average profile of outer diameter of failed tubes.

Fig. 5. Stereomicroscope image from corroded zone on the outer surface of tubes, beside a crack.

4
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Fig. 6. Outer surface of tubes: (a) the boundary between sound and corroded zones (b) pitting corrosion (c) pitting corrosion at higher magnifi-
cation.

Table 3
Chemical analysis of amine exchanger sheet in comparison to AISI standard (wt.%).
Sample Content

C Mn Si Cr Ni S P Fe

Exchanger Tube 0.018 1.76 0.32 18.28 8.7 0.024 0.004 Bal.
AISI 304L Standard < 0.03 ≤2 < 0.75 ≥18 ≥8 < 0.03 < 0.045 Bal.

From analysis presented in Fig. 7(a), two chloride peaks are visible. As seen from Fig. 7(c), it can be concluded that chloride peaks
in Fig. 7(a) are related to pits. In fact, Chloride concentration is noticeable inside pits.

3.4. Microstructure of heat exchanger tubes

Fig. 8 shows the Optical Microscope image of the failed tubes at different magnification taken by an Olympus GX71microscope.
The microstructure mainly consists of small amounts of deformation-induced twins in austenite grains with ASTM grain number 12.

3.5. Surface scale analysis

Fig. 9 shows X-ray diffraction analysis from the surface scales formed on the surface of a failed tube. A Stoe STADI-P X-ray
diffractometer used in this regard. As can be seen from this figure, the main corrosion products are Fe2CO3 and Fe2(SO4)3 and FeS.
Since stainless steels essentially do not have severe corrosion in amine solutions [25], it could be stated that Iron (Fe) in surface scales
should be due to CO2 and H2S corrosion of carbon steel in other parts of the unit, mostly fabricated from A516 carbon steel.

5
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Fig. 7. EDS microanalysis from external surface of failed tube. (a) large scale including pits (b) small area of sound zone near pits (c) micro-analysis
inside a pit.

3.6. aMDEA solution analysis

A Metrohm Model 827 PHlab was used for measuring pH of samples taken from aMDEA inside the shell. The average pH of the
semi-lean amine solution inside the shell was 9.23, which is in the normal range based on the process information of the plant.
Ion Gas Chromatography (IGC) analysis of the amine solution done by Metrohm 930. Table 4 shows results of this test. According
to previous studies [25,26], formate anion content in the aMDEA should be under 500 ppm, while samples contain over 1500 ppm.
Formate anion is capable to produce formic acid, which is corrosive to stainless steels [16]. Succinate and Acetate are not considered
as formate since their limit is 1000 ppm [25,26].
Leakage of oxygen is a common problem in natural gas treating plants leading to oxidative degradation and consequently,
formation of corrosive agents, e.g. formate anions [25–27]. It is worthy to mention that caustic addition (for neutralization) could
also be a reason for the increase in fomate level because of higher kinetics of formate formation by the addition of some chemicals
(e.g. NaOH) [28].

6
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Fig. 7. (continued)

Fig. 8. Optical Microscope image from the microstructure of 304stainless steel used as amine exchanger tubes, (a) etched for grain size mea-
surement (b) etched for microstructure evaluation.

4. Overall discussion, failure scenario, and recommendations

The main reason for failure is presence of gaps between tubes and tube sheets. These gaps remained after tube expansion,
inappropriate for environments that contain corrosive agents. Such gaps provide suitable sites for crevice corrosion in passive metals,
e.g. stainless steels. Previous studies provided a mechanism for crevice corrosion of passive alloys especially stainless steels in
chloride-containing media [10]. When a creviced part of stainless steel placed in chloride and air containing neutral solution, the
passive film will thicken while consuming oxygen. This is a cathodic reaction as shown by Reaction (1):

1/2 O2 + 2H2O + 2e− →4OH− (1)

The result of this reaction is a concentration gradient for oxygen inside and outside of the crevice accompanied by anodic
dissolution inside the crevice by Reaction (2):

M → M++ + 2e− (2)

In the Second stage, since metal cations pass into the solution and hydrolyze (reaction (3)), the metal ion concentration in the
crevice increases. As these hydroxides precipitate, OH– ions will be removed from the solution and generating H+ ions, which reduces
the pH consequently, as shown in the following (Reaction (3)).

7
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Fig. 9. XRD patterns of the corrosion product on the outside of tubes.

Table 4
Analysis of Semi-Lean and Lean amine solution.
Basic degradation compounds Amount

HSS Anions (ppm)

Acetate 212
Formate 1468
Succinate 238
Glycolate <1
Propianate 106
Butyrate <1
Chloride 3
Nitrate 2
Nitrite <1
Oxalate <1
Sulfite 4
Sulfate 10
Phosphate 6
Thiocyanate 2
Thiosulfate 9

Cation content (mg/kg)

Cr <1
K 4
Na 79
Ni 13
Fe 6
Mg <1
Ca <1

Mn+(Cl−)2 + 2H2O → M(OH)2 + 2H+ + Cl− (3)

In order to maintain the neutrality of the crevice, cations move out and anions (e.g. chloride) come inside the crevice (due to an
autocatalytic reaction) and increase the aggressiveness. The third stage is the breakdown of the passive film when the solution inside
the crevice is sufficiently aggressive to demolish the passive layer. Finally, the fourth stage is the propagation of crevice corrosion.
Previous studies [29], proved that acidified gap or so-called crevice is a proper site for pitting formation, which observed in the
present case study.
Based on chemical analysis, the amine solution contains minor amounts of formate higher than standard range[25,26]. Amine
degradation is known as the origin of various failures in natural gas treating systems, including corrosion [30]. When oxygen enters
the acid gas containing aMDEA at high temperatures, the result is degradation and formate formation according to Reaction (4)

8
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

[26,28]. Formate could react with chromium oxide in the passive layer of stainless steel as shown by following reactions.

(4)
(5)
Although chemical analysis showed that amine solution contains only 3 ppm chloride, this could be detrimental as there are
crevices formed between the tube and tube-sheet in the equipment. It is believed that the synergy of mentioned corrosive agents (e.g.
heat stable acids, acidic gases and chloride) in the concentrated cells inside the crevice zones is the main reason for failure.
Some recommendations in order to minimize future failure are as following:

• Elimination of crevices formed in the intersection of tube-sheet and baffles with tubes by proper designation. In this regard, full
tube expansion is preferred. In this regard, Sui et al. [11] conducted a research including simulation on variables of expansion
process. It was found that there is a linear relationship between the minimum expansion pressure and the initial radial gap, and
the minimum expansion pressure increased linearly with tube wall thickness deviation.
• Tighter control of corrosive agents, such as Chloride, Formate and Oxygen, if possible.
• Using more resistive alloys to crevice corrosion, such as the AISI 316L or 316LN.

5. Conclusions

Main conclusions from this case study are:

• Crevice corrosion accompanied by subsequent cracking was the main reason of tube failure.
• Inappropriate production (especially expansion process) provided a suitable gap or crevice.
• Presence of Chloride, heat stable acids, H S and CO caused sophisticated cell inside the crevice.
2 2

• Proper designation and tube expansion, tight control of corrosive agents, and using more resistive alloys to crevice corrosion could
mitigate future failure occurrence.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgments

Special thanks to the staff of the Pardis Petrochemical complex, also Materials Engineering Department of Isfahan University of
Technology for their cooperation.

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.engfailanal.2020.104918.

References

[1] A.L. Kohl, R. Nielsen, Gas Purification, Gulf Professional Publishing, 1997.
[2] K. Raju, Fluid Mechanics, Heat Transfer, and Mass Transfer: Chemical Engineering Practice, John Wiley & Sons, 2011.
[3] K. Thulukkanam, Heat Exchanger Design Handbook, CRC Press, 2013.
[4] S. Xu, C. Wang, W. Wang, Failure analysis of stress corrosion cracking in heat exchanger tubes during start-up operation, Eng. Fail. Anal. 51 (2015) 1–8.
[5] L. Hakka, S.F. Bosen, H. Liu, The Role of Anion Contaminants on Corrosion in Refinery Amine Units, American Institute of Chemical Engineers, 1995.
[6] R.T. Mousavian, E. Hajjari, D. Ghasemi, M.K. Manesh, K. Ranjbar, Failure analysis of a shell and tube oil cooler, Eng. Fail. Anal. 18 (1) (2011) 202–211.
[7] S. Allahkaram, P. Zakersafaee, S. Haghgoo, Failure analysis of heat exchanger tubes of four gas coolers, Eng. Fail. Anal. 18 (3) (2011) 1108–1114.
[8] Z.G. Yang, Y. Gong, J.Z. Yuan, Failure analysis of leakage on titanium tubes within heat exchangers in a nuclear power plant. Part I: Electrochemical corrosion,
Mater. Corros. 63 (1) (2012) 7–17.
[9] E. Bardal, Corrosion and Protection, Springer Science & Business Media, 2007.
[10] J. Oldfield, W. Sutton, Crevice corrosion of stainless steels: i. a mathematical model, Br. Corros. J. 13 (1) (1978) 13–22.
[11] R. Sui, W. Wang, Y. Liu, Y. Qu, Effects of austenitic stainless steel tube and tubesheet hole dimensional deviations on the hydraulic expansion pressure, Int. J.
Press. Vessels Pip. 138 (2016) 1–7.
[12] S. Xu, W. Wang, Numerical investigation on weld residual stresses in tube to tube sheet joint of a heat exchanger, Int. J. Press. Vessels Pip. 101 (2013) 37–44.
[13] S. Xu, Y. Wei, D. Guo, L. Zhang, W. Wang, Numerical investigation of thermo-mechanical stress in U-tube including forming effect for the SCC failure analysis,
Eng. Fail. Anal. 77 (2017) 126–137.
[14] R. Sui, W. Wang, Y. Liu, D. Li, W. Li, Root cause analysis of stress corrosion at tube-to-tubesheet joints of a waste heat boiler, Eng. Fail. Anal. 45 (2014) 398–405.
[15] H. Ma, H. Yu, C. Qian, Z. Liu, J. Zhou, Experimental study of hydraulically expanded tube-to-tubesheet joints for shell-and-tube heat exchangers, Procedia Eng.
130 (2015) 263–274.
[16] R.N. Maddox, J.M. Campbell, Gas Conditioning and Processing: Gas Treating and Sulfur Recovery, Campbell Petroleum Services, 1998.
[17] S. Rennie, Corrosion and materials selection for amine service, Mater. Forum, Citeseer, 2006, pp. 126–130.
[18] J. Kittel, M. Bonis, G. Perdu, Mitigating corrosion in sweet gas units: a comparison between laboratory, 2014.
[19] A. Krzemień, A. Więckol-Ryk, A. Smoliński, A. Koteras, L. Więcław-Solny, Assessing the risk of corrosion in amine-based CO2 capture process, J. Loss Prev.

9
H. Panahi, et al. Engineering Failure Analysis 118 (2020) 104918

Process Ind. 43 (2016) 189–197.

[20] C. Gouedard, D. Picq, F. Launay, P.-L. Carrette, Amine degradation in CO2 capture. I. A review, Int. J. Greenhouse Gas Control 10 (2012) 244–270.
[21] Y.-S. Choi, D. Duan, S. Nesic, F. Vitse, S.A. Bedell, C. Worley, Effect of oxygen and heat stable salts on the corrosion of carbon steel in MDEA-based CO2 capture
process, Corrosion 66 (12) (2010) 125004-125004-10.
[22] E. McCafferty, Introduction to Corrosion Science, Springer Science & Business Media, 2010.
[23] P. Rooney, T. Bacon, M. DuPart, K. Willbanks, Effect of Chlorides on Solution Corrosivity of Methyldiethanolamine (Mdea) Solution, Corrosion97, NACE
International, 1997.
[24] F. Hardesty, Metals Handbook Ninth Edition, Volume 1. Properties and Selection: Irons and Steels: Bruce P. Bardes (Ed.), American Society for Metals, Ohio,
1978. xxi+ 793 pages, over 1300 illus., ISBN 0-87170-007-7 hardback US $56.00, Elsevier, 1980.
[25] A. Erfani, S. Boroojerdi, A. Dehghani, M. Yarandi, Investigation of carbon steel and stainless steel corrosion in a MEA based CO2 removal plant, Pet. Coal 57 (1)
(2015) 48–55.
[26] P.C. Rooney, M. DuPart, Corrosion in alkanolamine plants: causes and minimization, CORROSION 2000, NACE International, 2000.
[27] R. Mesgarian, Corrosion management in gas treating plants (GTP’s): comparison between corrosion rate of DEA and MDEA a case study in sour gas refinery, in:
International Conference on Industrial Engineering and Operations Management, Bali, Indonesia, 2014.
[28] G. Badea, D. Ionita, P. Cret, Corrosion and passivation of the 304 stainless steel in formic acid solutions, Mater. Corros. 65 (11) (2014) 1103–1110.
[29] E.C. Hornus, M.A. Rodríguez, R.M. Carranza, R.B. Rebak, Comparative study of the crevice corrosion resistance of UNS S30400 and UNS S31600 stainless steels
in the context of Galvele’s model, Corrosion 73 (1) (2016) 41–52.
[30] H. Panahi, A. Eslami, M. Golozar, Effect of heat stable acids on corrosion and stress corrosion cracking initiation of 304 and 316 stainless steels in activated
methyl diethanol amine (aMDEA) solution, Corros. Eng., Sci. Technol. (2019) 1–9.

10