Engineering Failure Analysis 79 (2017) 642–655

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

Failure cases related to material issues in wet-process phosphoric MARK

acid plant
K. Chandraa,⁎, Amrita Mahantia, Vivekanand Kaina, B. Sunil Kumara, Niraj Kumarb,
Vicky Gautamb
a
Materials Processing and Corrosion Engineering Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India
b
Heavy Water Board, Mumbai 400 094, India

AR TI CLE I NF O AB S T R A CT

Keywords: The present paper describes three failure cases of metallic components handling wet-process
Failure analysis phosphoric acid at ambient temperatures in a phosphate fertilizer plant. All the three cases of
Cracks failure were related not directly to the corrosive environment rather to wrong selection or
Intergranular corrosion inferior quality of materials. In the first case, pipeline of stainless steel 316L failed due to inferior
Microstructures
quality of material used in the elbow region. The elbow material was not a low carbon grade
Pitting corrosion
stainless steel and was also in heavily sensitized condition which led to intergranular corrosion
and intergranular cracking. The other two cases were related to failures of pump sleeves made up
of cast alloys equivalent to stainless steel 316 and Hastelloy C-276 respectively. SS 316 showed
through-wall pitting and cracking while Hastelloy C-276 had undergone extensive corrosion
along the interdendritic boundaries. Both the materials contained high carbon content which led
to heavy precipitation of carbides (Cr-rich carbides in SS 316 and Mo-rich carbides in C-276)
along inter-dendritic boundaries during solidification of the casting reducing their corrosion
resistance. Recommendations to avoid such failures are also suggested.

1. Introduction

Phosphoric acid is used in various industries such as chemical, fertilizer, mineral leaching, water purification, petroleum refining,
food, and metal production. One of its important uses is in fertilizer industry for manufacturing of phosphatic fertilizers. More than
95% of world production of phosphoric acid is obtained by wet-process phosphoric acid (WPA) [1] that involves leaching phosphate
rocks rich in calcium phosphate minerals such as apatite, Ca10(PO4)6(F,OH)2 with sulphuric acid (H2SO4) [2]. The composition and
purity of phosphoric acid produced by these processes is determined by the chemical composition of the raw mineral, i.e., phosphate
rock and the acid used. Phosphate rock contain many impurities: fluoride (F−), chloride (Cl−), iron oxide (Fe2O3), aluminium oxide
(Al2O3), pyrites (ferrous sulfide, FeS), and fossilized organic matter. These impurities impair the quality of phosphoric acid, and
enhance corrosion and deterioration of materials of construction used for equipment during production of phosphoric acid [2–4].
Otherwise, pure phosphoric acid is much less corrosive than other acids such as sulphuric and hydrochloric acids. It has no effective
oxidizing power and is classified as a non-oxidizing acid [3]. However, WPA is treated as an oxidizing acid due to the presence of
oxidizing impurities, e.g., Fe3 + and Al3 + ions.
As reported in literature, the sulphate and fluoride concentration of WPAs are in the range of 0.8–5% and 0.1–1.3%, respectively
[4]. The chloride concentration is generally lower than 500 ppm, but concentrations up to 1500 ppm have also been reported [4]. The

⁎
Corresponding author.
E-mail address: kchandra@barc.gov.in (K. Chandra).

http://dx.doi.org/10.1016/j.engfailanal.2017.05.039
Received 12 April 2017; Received in revised form 11 May 2017; Accepted 15 May 2017
Available online 16 May 2017
1350-6307/ © 2017 Elsevier Ltd. All rights reserved.
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Table 1
Composition of WPA used in the fertilizer plant.

Parameter Concentration

P2O5 (%w/w) 26.6

SO42 − (g/l) 46.1
F− (g/l) 27.8
Cl− (g/l) 1.1
NO3− (g/l) 0.6
SiO2 (g/l) 16.5
Al (g/l) 1.4
Fe (g/l) 2.6
Mg (g/l) 1.5
Ca (g/l) 0.6
Na (g/l) 0.6

impurities particularly chloride and fluoride in WPA are severely corrosive to the alloys. Several corrosion failures of metallic
components handling WPA have been reported in literature [4–6]. Severe corrosive conditions exist in the reaction tanks (erosion
corrosion of agitators), evaporators and during transport and storage of acid. Stainless steel (SS) particularly type 316 is reported to
have suffered severe corrosion during transportation of WPA [4]. The corrosion attack is reported to have occurred as general
corrosion or as pitting attacks concentrated mainly to tank bottoms, especially under gypsum deposits and on surfaces exposed to the
vapour phase at the top of the tanks.
In the present study, three failure cases of metallic components handling WPA in a phosphate fertilizer plant are analysed. All the
three cases of failure were related not directly to the corrosive environment rather to wrong selection or inferior quality of materials.

2. Case study # 1: failure analysis of SS 316L piping

2.1. Background

A pipeline made of SS 316L handling WPA in the fertilizer plant failed at the elbow region after five years of operation. The pipe
was 4.2 mm thick with an outer diameter (OD) of 51 mm. The vertical pipeline was handling liquid at an ambient temperature (40 °C
maximum) with a flow of 12 m3/h (maximum) at an operating pressure of 5 kg/cm2 (maximum). The failure was observed as hairline
cracks in the 90° elbow which was installed on this vertical pipeline. Table 1 shows the typical composition of WPA that flowed inside
the pipeline. The other two failed components reported below were also exposed to the same composition of the WPA. Table 2
presents the chemical composition of phosphate rock used as the raw material in the fertilizer plant.

2.2. Visual examination

The hairline crack is visible on the OD side of the failed elbow in Fig. 1(a). The weld between the elbow and the straight portion of
the pipe is also visible and the crack is present close to the weld. The elbow was cut into two halves to observe the inner diameter (ID)
surface of the pipe and is shown in Fig. 1(b). Welds between the elbow and the straight section of the pipes are visible near both the
ends. The bend region (falling within the two welds in Fig. 1(b)) appears grey in colour in contrast to the bright metallic appearance
of the straight section of the elbow. This indicates that the bend region had undergone heavy corrosion during service, while the
straight section of the pipe was relatively unaffected. Regions up to 3 mm from both the welds in the elbow also showed metallic
bright colour and appeared to be unaffected from corrosion (Fig. 1(b)). At the bottom of the image in Fig. 1(b), it is seen that the
corroded layers are removed as flakes leading to reduction in the pipe thickness. A black region is seen on top right corner of the
elbow which corresponds to the failed region observed on the OD surface in Fig. 1(a).

2.3. Surface examination

The ID and OD surfaces of the SS 316L elbow in the as-corroded condition were examined at higher magnifications using a stereo
microscope. Fine multiple cracks are seen on the OD surface at about 4 mm from the weld (Fig. 2(a)). All the cracks are inclined along
the circumferential direction of the elbow. Fig. 2(b) shows the ID surface of the elbow in the as-corroded condition as observed by a
scanning electron microscope (SEM). Heavy intergranular corrosion (IGC), grain droppings and several intergranular (IG) cracks are
evident on the corroded surface (Fig. 2(b)).

Table 2
Composition (%w/w) of major constituents of phosphate rock used in the fertilizer plant.

P2O5 CaO Al2O3 Fe2O3 MgO SiO2 F Cl Moisture

30.69 48.34 0.86 0.57 0.36 3.0 4.5 0.02 3.74

643
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 1. Photographs of the failed SS 316 elbow (a) OD surface showing fine cracks (indicated by an arrow) and (b) ID surface showing grey colour in the bend region
indicative of corrosion. The removal of the corroded layers (at the bottom of the image) as thin flakes is also visible.

2.4. Microstructural examination

A sample was cut near the failed region of the elbow. The longitudinal section of the cut sample was mounted and
metallographically polished to a finish of 1 μm. Fig. 3 presents the optical micrographs in the as-polished condition. Heavy IGC is
observed, more so, near the ID side which is in contact with the process fluid resulting in thinning of the pipe from the ID side. IG
cracks are also observed that appear to propagate in the direction of pipe thickness. In fact, IGC has propagated to the full thickness of
the elbow as indicated by presence of IG cracking and attack even near the OD side of the elbow (Fig. 3(a)). Due to preferential attack
along the grain boundaries, individual grains are removed leading to thinning of the material (Fig. 3(b)).
The microstructure of SS 316L elbow was examined after etching in two separate regions: one near the weld which was unaffected
from corrosion and the other away from the weld (referred as base metal) which had undergone IGC. Etching was done as per ASTM
A262 practice A [7] that involves electrolytic etching in 10% oxalic acid solution for 90 s at a current density of 1 A/cm2 and the

644
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 2. SS 316 elbow showing (a) fine cracks (indicated by arrows) near the weld on the OD surface in the stereo microscopic image and (b) IGC, IG cracking and grain
droppings on the ID surface in the SEM micrograph.

microstructures are shown in Fig. 4. The base metal of the elbow consists of ‘ditch’ microstructure (Fig. 4(a)) which means that the
material was heavily sensitized in the as-received condition. The microstructure of the elbow near the weld is shown in Fig. 4(b). It
consists of dual structure (lightly sensitized) which contrasts with the ditch structure observed in the base metal. The microstructure
of the straight section of the pipe was also examined after etching as per ASTM A262 practice A [7] and the microstructure is shown
in Fig. 4(c). In contrast to the heavy ditch microstructure of the elbow, the straight section of the pipe had the ‘step’ microstructure.
This indicates that, between the straight section and the elbow of SS 316L pipe, either the material composition or the final heat
treatment is different. To confirm, if the material compositions of these two sections are different, chemical composition was
measured using optical emission spectroscopy and the results are presented in Table 3. Based on the chemical compositions, it
becomes obvious that the alloy grades between the elbow and the straight section of the pipeline are different. The elbow is SS type
316 and is not a low carbon grade alloy (the carbon content of 0.055 wt% is much higher than the maximum value of 0.03 wt%
specified for SS 316L [8]). The straight section of the pipeline conforms to SS 316L with a low carbon content of 0.011 wt%.

3. Case study # 2: failure analysis of SS 316 sleeve of a pump impeller

3.1. Background

Failure of SS 316 sleeve of a pump impeller was noticed which was used for pumping WPA fluid at ambient temperatures
(maximum 40 °C). The function of the impeller sleeve is to protect the shaft from erosion, corrosion, wear and encountering the
pumping fluid. The failed pump sleeve which is a cylindrical hollow metal tube is shown in Fig. 5. The thickness of the tube sleeve is
3.5 mm but at one end, the tube is tapered and the thickness reduces from 3.5 mm to 1.7 mm. Visual examination indicated that the
pump sleeve failed due to pitting. Some of the pits had propagated to the full thickness of the tube in regions where the tube thickness
was lower. Pits were formed on the OD surface (mostly concentrated near one end of the sleeve) and on the ID surface just adjacent to
one end of the sleeve, as evident in Fig. 5.

3.2. Material

The chemical composition of the sleeve material is presented in Table 3. The material roughly conforms to type 316 SS with

645
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 3. Optical microscopic images of the SS 316 elbow on the longitudinal section in the as-polished condition showing heavy IGC, IG cracking and thinning of the
material at (a) lower and (b) higher magnifications.

deviations in composition of some elements (Ni, Si, C, S and P) with respect to the standard alloy composition. Ni is lower while other
elements (Si, C, S and P) are higher than the standard alloy composition of SS 316 [8]. The concentrations of C and P are 0.15 and
0.5 wt% which are much higher than the maximum concentrations of 0.08 and 0.045 wt% respectively specified for SS 316. It was
suspected that the SS 316 pump sleeve was in cast form, therefore, its ferrite content was measured using a ferrite meter for
confirmation. The ferrite content of the sleeve was in the range of 4.5–8.3% indicating the material to be a cast stainless steel.

3.3. Surface examination

The stereo microscopic images in Fig. 6(a) and (b) show the pits and corrosion attack on the OD and the ID surfaces of the sleeve
respectively. The pits on the OD surface are small and mostly spherical in shape. The pits on the ID surface are more wide and
irregular in shape. A few cracks were also observed which were seen to be initiated through the base of the pits. One such pit on the
ID surface through which cracks have initiated and propagated in the longitudinal direction is shown in the SEM image (Fig. 7(a)).
SEM micrograph in Fig. 7(b) corresponds to the region near the crack tip confirming that the cracks have propagated along the
dendritic boundaries.

3.4. Microstructural examination

A small sample was cut from the end of the pump sleeve where pits were concentrated. The cross-section of the end was mounted
and polished to a finish of 1 μm. Fig. 8(a) corresponds to the optical image in the as-polished condition showing a pit initiated on the
OD surface and propagating along the thickness of the tube sleeve. Corrosion attack is also visible near the surface and around the pit
along the inter-dendritic boundaries. Fig. 8(b) shows the general microstructure of the SS 316 pump sleeve on the longitudinal section
after etching as per ASTM A262 practice A [7]. The dendritic microstructure consists of austenite matrix with ferrite along the

646
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 4. Optical micrographs after etching as per ASTM A262 Pr. A. The microstructure of SS 316 elbow showing (a) ‘Ditch’ structure of the base metal and (b) ‘Dual’
structure close to the weld. (c) The straight section of SS 316L pipeline showing ‘step’ microstructure.

647
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Table 3
Chemical compositions of materials of the failed components.

Component Pipeline Pipeline Pump sleeve Pump sleeve

(Elbow section) (Straight section)

Material SS 316 SS 316 L SS 316 Hastelloy C-276

Cr 16.75 17.81 17.5 17.1
Mo 2.45 2.7 2.2 15.8
Ni 11.45 12.2 8.5 Bal.
Si 0.59 0.36 1.2 1.0
Mn 1.64 1.82 1.1 0.38
C 0.055 0.011 0.15 0.094
S 0.011 0.018 0.05 0.022
P 0.017 0.029 0.5 0.06
Fe Bal. Bal. Bal. 5.2
Others – – – W: 3.8, V: 0.23, Co: 0.18

dendritic boundaries. Chromium carbides are also formed within the ferrite phase and at the ferrite/austenite boundaries.

4. Case study # 3: corrosion of Hastelloy C-276 sleeve of a pump impeller

4.1. Background

During routine maintenance, a high corrosion degradation was noticed in a localized region of a Hastelloy C-276 pump sleeve.
The corroded pump sleeve which is a cylindrical hollow metal tube, is shown in Fig. 9. The thickness of the tube sleeve is 3.5 mm but
at one end, the tube is tapered and the thickness reduces to 1.8 mm. Visual examination indicated that a length of about 12 mm on
the OD side of the pump sleeve was severely corroded along the entire circumference. The corroded region lies at about 30 mm from
one end of the sleeve. The chemical composition of the Hastelloy C-276 pump sleeve is shown in Table 3 which is a Ni-based Ni-Cr-Mo
alloys.

4.2. Surface examination

Fig. 10 shows the boundary between the corroded and the non-corroded regions on the OD surface of the Hastelloy sleeve as
observed by the stereo microscope. Deep grooving and a loss in thickness are observed in the corroded region.

4.3. Microstructural examination

Fig. 11(a) shows the microstructure on the longitudinal section of the Hastelloy C-276 sleeve after etching. Etching was done by
swabbing the polished sample in a solution of 5 ml H2SO4 + 3 ml HNO3 + 90 ml HCl for about 30 s. The dendritic microstructure
indicated that the material is in cast form. The as-cast structure consists of austenite matrix with continuous network of large size and
fraction of M6C-type carbides (where, M = Mo, W, Si) at the dendritic boundaries [9,10]. The optical micrograph in Fig. 11(b)
corresponds to the longitudinal section of corroded C-276 in the as-polished condition. Severe attack is seen along the dendritic
boundaries and inter-dendritic arm spacing starting from the OD side (Fig. 11(b)). The white coloured primary, secondary and

Fig. 5. Photograph of the failed SS 316L sleeve of a pump impeller.

648
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 6. Stereo microscopic images of SS 316 pump sleeve near one end showing (a) small spherical pits on the OD surface (indicated by arrows) and (b) wider pits on
the ID surface.

tertiary dendritic arms are without any corrosion attack while the spaces between them are severely corroded. The interdendritic
corrosion problem in cast C-276 is caused by the presence of intermetallic phases and carbides. One of the intermetallic phases is μ-
phase which is enriched with Mo compared to the bulk composition [11]. The presence of Mo-rich phase/carbide in the
microstructure of C-276 was confirmed after SEM and energy dispersive spectroscopy (EDS) examination. SEM micrograph in
Fig. 12(a) presents the microstructure of Hastelloy C-276 showing formation of carbides or intermetallic phases along the dendritic
boundaries. Also, marked are the phases on which point EDS was carried out. Fig. 12(b)–(d) show the corresponding EDS spectra and
chemical composition of the phases. Since, carbon is not accurately measured with the EDS, it is not included in calculating the final
composition. All these phases are rich in Mo, W and Si in comparison to the matrix which could be related to M6C-type carbides. Mo
concentration as high as 42 wt% is measured (Fig. 12(b)) in the carbides.

5. Discussion

In the first case study, SS type 316 pipeline has shown high IGC and failure only in the elbow region. The elbow was joined with
the straight section of the pipe by welding. The straight section of SS 316L pipeline was not affected by corrosion in contrast to heavy
IGC observed in the elbow. The chemical analysis (Table 3) showed that the material of the failed elbow was of different grade than
the rest of the pipe. The failed elbow was SS type 316 with high carbon content of 0.055 wt%. In addition, the material was heavily
sensitized having ‘ditch’ microstructure whereby the grain boundaries are surrounded by chromium carbide and chromium depletion.
Austenitic stainless steels with such a microstructure are prone to IGC in corrosive environments [12] such as in the WPA. The cracks
may also initiate at the corroded grain boundaries at the grain boundaries in the presence of residual (due to welding or bending of
the pipe) or operating stresses (Figs. 2 and 3), since, these become weak. The remaining section of the pipe was type SS 316L with low
carbon content of 0.011 wt% and had a ‘step’ microstructure and, therefore, was unaffected from corrosion. Regions up to 3 mm of
the elbow from the welds was also unaffected from corrosion and showed a bright metallic surface appearance (Fig. 1(b)). This is

649
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 7. SEM micrographs corresponding to the ID surface of SS 316 pump sleeve showing (a) cracks initiating through the base of a pit and (b) crack propagation path
along the inter-dendritic boundaries.

because in these regions, chromium carbides and/or chromium depletion at the grain boundaries were partially dissolved due to the
heat of welding and therefore, showed ‘dual’ microstructure (Fig. 4(b)). It is therefore concluded that the failure of the pipeline
handling WPA was related to a higher carbon content and the sensitized microstructure of the SS 316 elbow.
The other two failure cases are of sleeves used in the pump for handling WPA at ambient temperatures. In the first case (case study
# 2), SS 316 sleeve has shown through wall pitting and cracking. The microstructural examination confirmed that the sleeve material
was in cast form (Fig. 8(b)). The microstructure consisted of austenite matrix with ferrite along the dendritic boundaries (Fig. 8(b)).
In general, cast stainless steel has a lower corrosion resistance compared to their equivalent wrought product due to localized
contamination, microsegregation, or lack of homogeneity. For example, mould quality may cause superficial compositional changes
and carbon pick-up from mould release agents that can affect corrosion resistance of the cast materials. In the present case, cast SS
316 contained very high carbon and phosphorous contents of 0.15 and 0.5 wt% respectively (Table 3). High carbon leads to heavy
precipitation of chromium carbides (and subsequent chromium depletion) in the microstructure which is precipitated within the
ferrite phase and at the ferrite/austenite boundaries during solidification of the casting (Fig. 8(b)). Additionally, a higher
phosphorous concentration is reported to cause reduced IGC resistance of austenitic stainless steels [12]. Therefore, such a
composition and microstructure of austenitic SS is prone to corrosion in an aggressive environment hence, the material showed severe
pitting and preferential corrosion along the interdendritic boundaries in WPA containing halide ions (Fig. 8(a)). A few of the pits have
also shown crack initiation through their base, since, these pits act as stress concentration sites. This could be possible as inside the
pits, the chloride and fluoride content is known to increase many folds compared to that present in the bulk solution [10,12]. In
addition, pits also act as stress concentration sites [10,12]. Therefore, stress corrosion cracks initiate from bottom of the pits and are
not observed to initiate on surfaces exposed to the bulk solution.
Because of inherent limitation of corrosion resistance of cast materials, wrought form of austenitic stainless steel type 316L should
be selected for the sleeve material for extended service life. Also, these materials should be qualified as per ASTM, A262 for resistance
to IGC.
In case study # 3, C-276 Hastelloy sleeve that showed high corrosion degradation in contact with WPA was also in cast form. The
microstructure consisted of austenite matrix with large and continuous carbide precipitates along the interdendritic boundaries. In
cast Hastelloy, formation of a variety of primary carbides (MC, M23C6 and M6C types) and intermetallic phases, e.g., sigma, μ-phase

650
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 8. Optical micrographs of SS 316 pump sleeve in the as-polished condition showing (a) a wide pit and corrosion attack around the pit on the cross-section and (b)
the general microstructure after etching on the longitudinal section, matrix is austenite and the dark regions are ferrite and carbides.

Fig. 9. Photograph of the corroded (indicated by an arrow) Hastelloy C-276 sleeve of a pump impeller handling WPA.

[9,10] together with segregation of Mo along the dendritic boundaries occur [13] during solidification. In the case of C-276 alloy,
most carbides are reported to be M6C-type, since, the M6C carbide usually replaces the M23C6 carbide in alloys containing over
approximately 6 wt% Mo plus its atomic equivalent in W, i.e., weight per cent Mo plus one half the weight per cent W [9]. As
analysed by SEM and EDS in the present study (Fig. 12), the carbides precipitated in C-276 are rich in Mo, W and Si and appear to be

651
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 10. Stereo microscopic image of the Hastelloy C-276 pump sleeve showing the boundary between the corroded and non-corroded regions.

related to M6C-type carbides.

As mentioned above, castings have typically high carbon contents. In the present case, cast C-276 contained 0.094 wt% carbon
which is on the higher side and due to slow cooling from the molten state after pouring into ceramic moulds, significant grain

Fig. 11. Optical micrographs of the Hastelloy C-276 pump sleeve on the longitudinal section (a) after etching showing as-cast microstructure consisting of continuous
network of carbides at the dendritic boundaries and (b) in as-polished condition near the OD side showing corrosion attack along the inter-dendritic boundaries.

652
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

Fig. 12. SEM micrograph of the microstructure of Hastelloy C-276 showing formation of carbides/intermetallic phases along the dendritic boundaries. The phases on
which point EDS was carried out are marked. (b) EDS spectrum 1 (c) EDS spectrum 2 and (d) EDS spectrum 3.

653
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

boundary precipitation of carbides takes place. Both M6C-type carbides and Mo-rich μ-phase (which can also form during cooling),
depletes the adjacent austenite in molybdenum. Molybdenum in solid solution in Ni-base alloys imparts corrosion resistance in non-
oxidizing chloride-bearing environments [10], and its depletion near the grain boundaries would render the material susceptible to
intergranular attack in WPA containing fluorides and chlorides. In aggressive reducing environments, even the matrix-intermetallic
interphase is attacked [14].
To increase the corrosion resistance of Hastelloy castings, it is necessary to dissolve carbides and intermetallic phases such as μ-
phase by solution annealing. M6C carbides can be dissolved at temperatures close to 1100 °C [13]. For the dissolution of μ-phase or
for the removal of Mo segregation, a higher annealing temperature (close to 1220 °C) is required [10,13,14], since the diffusion rate
of Mo is low.
All the three components showing failure or high corrosion degradation during service were exposed to WPA fluid at ambient
temperatures (40 °C maximum). As presented in Table 1, WPA contained 26.6% (w/w) P2O5 which is equivalent to 37% (w/w)
phosphoric acid concentration and as such is only mildly corrosive to stainless steels and Ni-based alloys at ambient temperatures
[3,15,16]. However, WPA also contained a variety of impurities especially fluoride, chloride and sulphate of high concentrations that
increases the aggressiveness of WPA. These impurities in WPA are known to cause high corrosion degradations [2–4,15–19] and
failures [4–6] of metallic components made of stainless steels. However, higher corrosion rates have mostly been reported at
relatively higher operating temperatures (> 60 °C) [3,15–17,20,21]. In contrast, certain soluble metal ion impurities in WPA are
beneficial and increase the corrosion resistance of the alloys. The metal ions such as Si4 +, Mg2 + and Al3 + are effective in containing
the aggressiveness of fluorides by forming strong complexes thereby reducing the amounts of free fluorides [2–4,19]. These metallic
impurities are in appreciable amounts (Table 1) in the WPA used in the plant. Despite a mildly corrosive environment and lower
operating temperatures (40 °C maximum), the three components showed failure and high corrosion degradation which are related to
material issues. In the first case, it is wrong selection of elbow material of SS 316 with high carbon and having a sensitized
microstructure. In the second case, the cast SS 316 sleeve is of inferior quality containing very high carbon and phosphorous content.
In the third case, the C-276 casting too has high carbon content with the microstructure containing high content of M6C-type
carbides, thereby, reducing the corrosion resistance. Therefore, chromium or molybdenum depletion, resulting either from
chromium-rich carbide or from molybdenum-rich intermetallic precipitates, cause severe corrosion in the form of intergranular
corrosion/interdendritic corrosion in WPA.

6. Conclusions and recommendations

a) SS 316L pipeline failed at the bend region because of inferior quality of the material used in the elbow. The straight section of the
pipe was low-carbon grade and had a ‘step’ microstructure and therefore worked well in contact with the WPA. However, SS 316
elbow had a high carbon content and was in sensitized condition (‘ditch’ microstructure) which led to intergranular corrosion and
IG cracking. It is recommended that the entire pipeline/elbow should be made of an L-grade type 316 SS and the material must be
in solution-annealed condition and qualified as per ASTM A262 [7] for resistance to IGC.
b) Both SS 316 and Hastelloy C-276 sleeves of the pump impeller handling WPA showed high corrosion rates and failure because
both the materials were in cast form and had very high carbon contents. Both these as-cast materials had shown heavy presence of
carbides along inter-dendritic boundaries and lack of homogeneity in the microstructure. For extended service life of the pump
sleeves, it is recommended to use wrought form of these materials which must be qualified as per ASTM A262 for austenitic
stainless steels and ASTM G28 [22] for Ni-base alloys for resistance to IGC.

Acknowledgements

We are thankful to Heavy Water Board for giving us the opportunity to carry out the failure analysis by providing the failed
components.

References

[1] C. Escriva-Cerdan, E. Blasco-Tamarit, D.M. Garcia-Garcia, J. Garcia-Anton, A. Guenbour, Temperature effect on the austenitic stainless steel UNS N08031 used in
the wet method phosphoric acid production, Chem. Eng. Trans. 32 (2013) 1717–1722.
[2] M. Schorr, B. Valdez, The phosphoric acid industry: equipment, materials, and corrosion, Corros. Rev. 34 (2016) 85–102.
[3] ASM Handbook, Corrosion: Environments and Materials, first ed., vol. 13C, ASM International, Materials Park, OH, 2006, pp. 736–744.
[4] T. Sydberger, S. Nordin, Corrosivity of wet process phosphoric acid, Corrosion 34 (1978) 16–22.
[5] M. Schorr, B. Valdez, R. Zlatev, N. Santillan, Agitator corrosion in wet phosphoric acid production, Mater. Perform. (Feb. 2007) 50–53.
[6] M. Schorr, B. Valdez, R. Zlatev, M. Stoytcheva, Erosion-corrosion in phosphoric acid production, Mater. Perform. 56 (Jan. 2010) 56–59.
[7] ASTM A262-15, Standard Practices for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels, ASTM International, West Conshohocken,
PA, 2015.
[8] ASTM A240/A240M-16a, Standard Specification for Chromium and Chromium-nickel Stainless Steel Plate, Sheet, and Strip for Pressure Vessels and for General
Applications, ASTM International, West Conshohocken, PA, 2016.
[9] G.P. Sabol, R. Stickler, Microstructure of nickel-based superalloys, Phys. Status Solidi 11 (1969) 11–52.
[10] E.E. Stansbury, R.A. Buchanan, Fundamentals of Electrochemical Corrosion, ASM International, Materials Park, OH, 2000, pp. 271–406.
[11] P. Crook, Corrosion resistant nickel alloys part 2, Adv. Mater. Process. (July 2007) 31–33.
[12] A.J. Sedriks, Corrosion of Stainless Steels, second ed., The Corrosion Monograph Series, A John Wiley & Sons, NY, 1996, pp. 231–236.
[13] Q. Zhang, G. Zhu, B. Hu, Effect of diffusion annealing on composition uniformity of Hastelloy C-276 ingot, Procedia Eng. 27 (2012) 1016–1023.
[14] H.M. Shalaby, Failure of Hastelloy C-276 pump impeller in hydrochloric acid, Eng. Fail. Anal. 15 (2008) 543–546.
[15] A. Alon, J. Yahalom, M. Schorr, Influence of halides in phosphoric acid on the corrosion behavior of stainless steels, Corrosion 31 (1975) 315–319.

654
K. Chandra et al. Engineering Failure Analysis 79 (2017) 642–655

[16] R.W. Kichner, An evaluation of nickel-rich alloys in wet process phosphoric acid, Werkst. Korros. 22 (1971) 581–592.
[17] R. Sanchez-Tovar, M.T. Montanes, J. Garcia-Anton, A. Guenbour, Influence of temperature and hydrodynamic conditions on the corrosion behavior of AISI 316L
stainless steel in pure and polluted H3PO4: application of the response surface methodology, Mater. Chem. Phys. 133 (2012) 289–298.
[18] S. Ghosh, T. Ramgopal, Effect of chloride and phosphoric acid on the corrosion of alloy C-276, UNS N08028, and UNS N08367, Corrosion 61 (2005) 609–619.
[19] P. Barraco, G. Chiaro, A. Barbera, Influence of impurities on stainless steels corrosion in wet process phosphoric acid, Fertil. Res. 6 (1990) 281–289.
[20] H. Li, Z. Jiang, H. Feng, S. Zhang, P. Han, W. Zhang, G. Li, G. Fan, Effect of temperature on the corrosion behavior of super austenitic stainless steel S32654 in
polluted phosphoric acid, Int. J. Electrochem. Sci. 10 (2015) 4832–4848.
[21] M. Ibanez-Ferrandiz, E. Blasco-Tamarit, D.M. Garcia-Garcia, J. Garcia-Anton, A. Guenbour, S. Bakour, A. Benckokroun, Effect of temperature on the corrosion
resistance of stainless steels in polluted phosphoric acid, ECS Trans. 25 (37) (2010) 49–61.
[22] ASTM G-2802, Standard Test Methods for Detecting Susceptibility to Intergranular Corrosion in Wrought, Nickel-rich, Chromium-bearing Alloys, ASTM
International, West Conshohocken, PA, 2015 (2015).

655