Crevice Corrosion of Stainless Steels 904L, 2205, 2507 in High Temperature Sulfuric Acid Solution Containing Chlorides

CORROSION SCIENCE SECTION
Crevice Corrosion of Stainless Steels 904L,

Downloaded from corrosionjournal.org by University of Michigan-Ann Arbor on 06/16/18. For personal use only.
2205, and 2507 in High-Temperature Sulfuric

Acid Solution Containing Chlorides:
Influence of Metal Cations
M. Lindgren,‡,* E. Huttunen-Saarivirta,** H. Peltola,* J. Romu,*** T. Sarikka,*** H. Hänninen,*** and P. Pohjanne**

CORROSION 2018.74:225-240.
ABSTRACT INTRODUCTION
Crevice corrosion influences the integrity of materials in many Hydrometallurgical treatment of ores has been pro-
industrial sectors, such as hydrometallurgy. The conditions in moted by declining ore grades and the emergence of
hydrometallurgical processes are severe, with high concen-
secondary raw material sources. For example, copper
trations of sulfuric acid leachant as well as high-chloride im-
ores that are too low-grade for concentration process or
purity, oxygen, and metal cation contents. In this study,
crevice corrosion behavior of highly-alloyed austenitic and contain such undesirable impurities that they cannot
duplex stainless steels is investigated in aerated high- be smelted may be refined by hydrometallurgical route.1
temperature sulfuric acid solution with 1 g/L chlorides in the In the hydrometallurgical treatment of mineral raw
presence of three types of metal cations: Fe3+, Cu2+, and Fe2+. materials, sulfuric acid is a common leaching agent
The experimental work contains the determination of the because of its low cost, environmental friendliness,
oxidizing capacity of the solutions, immersion crevice corro- and possibility for regeneration. The use of elevated
sion tests followed by characterization of the corrosion attack, temperatures significantly accelerates the leaching
and potentiodynamic polarization measurements. The results
kinetics and, therefore, typical process temperatures
show that the alloy with the highest PREN value, grade 2507,
approach the boiling point of the solution, being about
was not susceptible to crevice corrosion under the test con-
ditions. In the case of grades 904L and 2205 of relatively equal
100°C. From the corrosion point of view, chlorides are
PREN values, the extent and nature of crevice corrosion attack the main impurities in the leaching process and they
was different and dependent on the ratio of activities of chlorides may originate either from raw materials or from make-up
to sulfates and the type of metal cations. The results are water. Typical chloride concentrations are in the
presented and discussed in the light of IR drop mechanism. magnitude of a few hundred milligrams per liter. Oxygen
and ferric ions, Fe3+, the latter also originating from
KEY WORDS: cations, chloride, crevice corrosion, stainless steel,
the mineral raw materials, can act as oxidants to en-
sulfuric acid
hance the leaching. Because of this, the redox po-
tential of the solution is high, promoting localized, i.e.,
pitting and crevice, corrosion of passivating metals if
critical values are exceeded. Ferrous iron, Fe2+, is always
Submitted for publication: June 22, 2017. Revised and accepted:
September 28, 2017. Preprint available online: September 28, present and has an inhibiting effect on uniform cor-
2017, https://doi.org/10.5006/2565. rosion of the main construction materials in sulfuric acid
‡
Corresponding author. E-mail: [email protected].
* Outotec (Finland), Pori Research Center, P.O. Box 69, 28101 Pori,
environment.2 Cupric ions, Cu2+, if present, are ben-
Finland. eficial for the uniform corrosion resistance of passivating
** VTT Technical Research Centre of Finland Ltd., Kemistintie 3, 02044 alloys but may promote localized corrosion.
Espoo, Finland.
*** Aalto University School of Engineering, P.O. Box 14100, 00076 Aalto, The aggressive nature of the leaching environ-
Finland. ment is reflected in the materials selection. Depending
ISSN 0010-9312 (print), 1938-159X (online)
CORROSION—Vol. 74, No. 2 18/000037/$5.00+$0.50/0 © 2018, NACE International 225
on the exact solution chemistry, austenitic corrosion- raise to current (I increases), (ii) the potential range of
resistant grades, such as 904L (EN 1.4539, UNS the active peak increases (EA/P increases) as a result of
N08904(1)), and duplex ferritic-austenitic corrosion- an increase in temperature or a change in the elec-
resistant grades, such as 2205 or 2507 (EN 1.4462, trolyte composition (pH,14 chlorides), (iii) R increases
UNS S32205; and EN 1.4410, UNS S32750, respec- because of the formation of constriction by gas
tively), can be selected. However, crevice corrosion is the bubbles or corrosion products,15 or (iv) the corrosion
most difficult corrosion mode to take into account in potential of applied potential ESURF shifts in the
the choice of materials, as crevices may vary from the negative direction. Then, the crevice corrosion initiates
geometrical and structural features, such as crevices at some distance xA/P from the crevice mouth.
created by bolts, nuts, welds, and flange connections, to Considering only the IR > Δ∅* criterion, crevice
those created by an unpredictable formation of ad- corrosion is less likely when oxidants with noble elec-
herent deposits or accumulation of solids from the trode potential are used, as Δ∅* increases. This has
process. been demonstrated to hold true for iron in strongly
The current understanding of crevice corrosion of buffered solutions or in solutions close to the hy-
passivating metals is based on three models:3 (1) pas- drolysis equilibrium, i.e., in cases where the pH change
sive dissolution leading to a gradual acidification and is precluded.16 Abdulsalam and Pickering investi-
a general passivity breakdown,4 (2) so-called IR (ohmic) gated the effect of the oxidizing power on the electrode
drop mechanism,5 and (3) stabilization of metastable potential profile and in situ measured crevice corro-
pits by an occluded crevice geometry,6-7 or a combina- sion for nickel in 1 N H2SO4 under conditions of effective
tion of these. Typically, the generation of more ag- convective mixing of the crevice and bulk electro-
gressive environment in the pit or crevice involves lytes.9 In other words, in their experimental system, the
acidification of the environment and the ingress of composition of the crevice electrolyte did not change
negative chloride ions to counterbalance the positively and hence pure IR drop effect was measured. They
CORROSION 2018.74:225-240.
charged metal cations, and the crevice environment stated that both higher ESURF and increase in the
depletes with respect to oxygen. The special charac- amount of oxidant (oxygen) available in the crevice
teristics of the hydrometallurgical leaching solutions meant the lower susceptibility or rate of crevice cor-
are that the leaching solutions are very acidic to begin rosion. Abdulsalam, et al.,17 further demonstrated by
with (pH < 1) and the solutions contain negatively applying an external potential that the higher the
charged sulfate ions in addition to chlorides and applied potential, the lower the crevice corrosion rate. In
oxidizers, probably also influencing the crevice cor- contrast, a different view emerges from the studies in
rosion behavior. which the composition of the crevice electrolyte is
In general, the IR drop mechanism is widely allowed to change with time. Sridhar and Dunn in-
acknowledged to explain the crevice corrosion of passive creased the applied potential in steps and monitored
metals with the changes in the solution composition potential, current density, Cl−concentration, and pH
inside the crevice aggravating the chemical environ- of the bulk in the mouth and tip of the crevice.11 They
ment. According to the IR drop mechanism, the observed that the environment inside the crevice was
localized corrosion starts when: altered only when the external potential was maintained
above a certain value. When this critical value was
IR > Δ∅ (1) exceeded, there was an incubation period followed by a
rapid increase in the current density. This event was
where: succeeded by a decrease in pH and potential in the
Δ∅ = ESURF − EA=P (2) crevice. Similarly, in potentiostatic and open-circuit
tests of Alloy 825 (UNS N08825) in 1,000 ppm Cl− at
In these equations, I is the current flowing out of 95°C, higher applied potentials were conducive to
the crevice, R is the resistance of the crevice electrolyte, faster crevice corrosion initiation times.18 This means
and Δ∅* is the difference between the applied po- that, in contrast to pure IR experiments, when the
tential ESURF and the electrode potential of the active/ environment was allowed to change, the higher applied
passive transition EA/P. In some cases,8 e.g., pure potentials promoted the occurrence of crevice corro-
metallic nickel in sulfuric acid,9 the active peak origi- sion. This is consistent with the establishment of
nally exists in the polarization curve and immediate various critical potentials for the prediction of the
crevice corrosion ensues. In contrast, in less corrosive occurrence of crevice corrosion in stainless steels
environments10 or for more corrosion-resistant and nickel-based alloys.19-20
alloys,11-12 an incubation period is required for the Crevice corrosion resistance of austenitic
active peak to appear. This can occur when (i) the and duplex stainless steels can be influenced by
passive current increases13 or metastable pitting gives alloying.21-22 In duplex stainless steels, the two-
(1)
phase microstructure implies that the austenite and
UNS numbers are listed in Metals and Alloys in the Unified Num-
bering System, published by the Society of Automotive Engineers ferrite phases may be attacked differently based on
(SAE International) and cosponsored by ASTM International. their different chemical composition and thus
226 CORROSION—FEBRUARY 2018

TABLE 1
Test Materials, Their Nominal Chemical Compositions (in wt%), and PREN Values
Grade EN Microstructure C N Cr Ni Mo Others PREN
904L 1.4539 Austenitic 0.01 – 20 25 4.3 1.5Cu 34

2205 1.4462 Duplex 0.02 0.17 22 5.7 3.1 – 35
2507 1.4410 Duplex 0.02 0.27 25 7 4 – 43
TABLE 2
Target Compositions of the Solutions
Solution H2SO4 (g/L) Cl− (g/L) Fe3+ (g/L) Cu2+ (g/L) Fe2+ (g/L)
1 50 1 5 – –
2 50 1 – 5 –
3 50 1 – – 5
4 50 1 0.5 – –
polarization behavior within the crevice.23 In the zone Before the tests, the specimens were cleaned with eth-
of the most severe attack, the microstructural features anol, rinsed with deionized water, and dried.
are often no longer distinguishable, but in the areas of
lighter attack, either ferrite22-24 or austenite25 may be Methods
preferentially attacked even within different locations Immersion Crevice Corrosion Tests — Prior to the
CORROSION 2018.74:225-240.
of a single sample. attachment of the crevice formers, the specimens

In this study, crevice corrosion behavior of were weighed. The specimens were then equipped with
highly-alloyed austenitic and duplex stainless steels is flat or serrated polytetrafluoroethylene (PTFE) crevice
investigated in high-temperature sulfuric acid solu- formers with inner and outer diameters of 12.5 mm and
tion with 1 g/L chlorides, under the conditions simu- 27.6 mm, respectively, which were tightened onto the
lating leaching processes of ores. The focus is on the surfaces to a torque of 6 N·m using a moment wrench.
effect of the amount and nature of metal cations on the The serrated crevice formers contained 1 mm wide
degradation of the materials. Instead of using an and 1 mm thick grooves. The stainless steel specimens
external current source to apply the potential to the together with the crevice formers were attached to an
sample, oxidizing ions Fe3+ and Cu2+, and metal annular specimen holder made of titanium, and the ring
cations Fe2+ are used to see whether their chemical was placed on the bottom of the reactor. The speci-
nature has an influence besides their different elec- mens were insulated from the titanium bolts by PTFE.
trode potentials. The extent and characteristics of at- The reactors used in the experiments were heated and
tack as well as reasons for dissimilar behavior of the thermally insulated glass reactors of the liquid volume
three stainless steel grades are described and discussed of 5 L.
in this paper. Four experiments with equal sulfuric acid, 50 g/L
(0.5 M), and chloride concentrations, 1 g/L (0.028 M),
EXPERIMENTAL PROCEDURES but with different additions of oxidizing ions or metal
cations were performed. In the experiments, three types
Materials of cations: ferric iron (Fe3+), cupric ion (Cu2+), and
Studies were conducted for three stainless steel ferrous iron (Fe2+) at the initial concentration of 5 g/L
grades: AISI 904L, 2205, and 2507. The typical chem- were included, with one type of cation in each test.
ical compositions of the alloys and their pitting re- Additionally, the influence of the ion concentration,
sistance equivalent numbers, PREN = Cr + 3.3 Mo + 0.5 g/L vs. 5 g/L, of ferric iron (Fe3+) was examined in
16 N, are shown in Table 1. The surface finish of the the fourth experiment. In all cases, the solution tem-
test materials was 2E, i.e., annealed, shot blast, and perature was 90°C and the immersion time of the
pickled. The materials were obtained as sheets with samples was 4 weeks (672 h). The target parameters of
the thickness of 3 mm, from which specimens of the size the test solutions are given in Table 2. Only one
of 50 mm × 50 mm were laser cut. The edges of the specimen of each material was tested using the flat
specimens were ground and treated with a pickling crevice formers, but parallel tests were run with the
paste (Polinox P Rapid†) for 60 min in order to remove serrated crevice formers allowing the detected trends to
the chromium-depleted surface layer possibly induced be verified.
by laser cutting, and also to repassivate the edges. The test solutions were prepared from reagent
grade chemicals and deionized water. The addition of
†
Trade name. chlorides was made in the form of hydrochloric acid,
CORROSION—Vol. 74, No. 2 227

TABLE 3
Analyzed Values of the Test Solutions Before the Tests
H2SO4 Cl− Fe3+ Fe(tot) Cu2+ Redox at 23°C Redox at 90°C
Solution (g/L) (mg/L) (mg/L) (mg/L) (mg/L) (mVAg/AgCl) (mVAg/AgCl) pH at 23°C pH at 90°C
1 52.7 942 5,100 5,110 <1 507 676 0.31 0.66

2 52.6 940 – <1 5,220 294 440 0.27 0.49
3 52.0 970 370 5,250 – 338 443 0.35 0.44
4 52.1 1,170 550 574 – 409 675 0.32 0.57
and the additions of cations as sulfates: Fe2(SO4)3, The anodic polarization measurements for all
CuSO4, and FeSO4. The solutions were analyzed before three stainless steels were performed in the electrolyte
and after the tests in order to check the match with the that was the base in all four test solutions in im-
target values and to reveal the changes during the tests. mersion crevice corrosion experiments, i.e., aqueous
The redox potentials and the pH values were measured solution heated to 90°C and containing 50 g/L sul-
from the initial test solutions at room temperature, furic acid, H2SO4, and 1 g/L chlorides (Cl−, added as
after the solutions had reached the test temperature HCl). Similarly, measurements for all three grades
(90°C), and from the final solutions at the end of the were conducted in concentrated chloride solutions
tests. In the initial test solutions, the open-circuit containing a five-fold amount of chlorides compared
potentials (OCP) of grade 904L stainless steel were to the base solution, i.e., 50 g/L sulfuric acid and 5 g/L
additionally measured at 90°C. In both measurements, chlorides (added as HCl). Additionally, for grade 904L,
redox and OCP, an Ag/AgCl (saturated KCl) reference the influence of the oxidizing ions (adding 0.5 g/L and
electrode was used. The counter electrode in the OCP 10 g/L of Fe3+ in the base solution), chloride con-
CORROSION 2018.74:225-240.
measurements was platinum. The solutions were centration (adding 0, 1, 5, and 10 g/L of Cl− in the base
stirred by rotary magnets (8 mm × 40 mm) throughout the solution), and the addition of sulfate (SO2−
4 ; 45 g/L as
tests. Minor vaporization occurred from the solutions Na2SO4 into the solution containing 50g/L H2SO4 and
during the 4-week test period, although the reactors were 5 g/L Cl−) were investigated.
sealed with lids. This was compensated by deionized
water additions on a regular basis. RESULTS
Characterization — After the tests, the specimens
were photographed (Nikon D 700†), and inspected Solutions and Their Oxidation Capacity
both visually and under a stereo microscope (Zeiss The analyses of the solutions before the immer-
Stemi 2000-C†) in order to evaluate the occurrence of sion crevice corrosion tests are presented in Table 3. The
corrosion. Based on the stereomicroscopy inspections, table shows that the measured initial compositions
the most interesting specimens were selected for were consistent with the targeted ones. In Solutions 1
scanning electron microscopy (SEM) examinations. and 4 (5 g/L and 0.5 g/L Fe3+), the dissolved iron was
SEM studies were conducted for the specimens in almost totally in ferric form, as intended. In turn, in
plane and cross section using a Zeiss Merlin VP Com- Solution 3 (5 g/L Fe2+), a minor part of the dissolved
pact SEM† equipped with a Bruker X-flash† energy iron was in ferric (Fe3+) form, although the addition was
dispersive spectrometer (EDS) and a Bruker Quantax† made as ferrous sulfate (Fe2+). This occurred con-
electron backscattered diffractometer (EBSD). sistently when the solution preparation procedure was
Electrochemical Measurements — The three stain- repeated. The valence of copper ions in Solution 2 was
less steel grades were subjected to potentiodynamic not analyzed, but the addition was made as Cu2+ ions.
polarization measurements to reveal their overall cor- The equilibrium concentrations, activities, and
rosion behavior in the test solutions and in simulated pH of the solutions as a function of temperature were
crevice solutions. In order to have the specimens in an calculated with the HSC Software†,26 both for the
identical condition, the surfaces were ground to 600 target compositions (Table 2) and the measured com-
grit surface finish. The measurements were performed positions (Table 3). The calculations took into account
using an Avesta cell, each of the stainless steel grades the non-ideality of the solution by using the Pitzer’s
as the working electrode, platinum as the counter model to describe ion-ion interactions. Table 4
electrode, and an Ag/AgCl (saturated KCl) or a sat- summarizes the calculated activities for chlorides and
urated calomel electrode (SCE) as the reference elec- sulfates, their ratios, and calculated pH values for
trode. For the uniform presentation of the results, all solutions at 23°C and 90°C. Despite a relatively high
potential values in this paper are presented with respect sulfuric acid concentration, not many sulfate ions
to Ag/AgCl (saturated KCl). The measurements were (SO2−
4 ) existed in the equilibrium solutions and their
performed from −50 mVOCP to 960 mV vs. reference activities were low. This is because sulfuric acid does
electrode and back. The scan rate in the measure- not dissociate completely in aqueous media to H+ and
ments was 0.167 mV/s. SO2−
4 ions and an intermediate species, bisulfate

TABLE 4
Summary of the Calculated Activities and pH(A)
a(Cl−) Calculated a(SO2−
4 ) Calculated a(Cl−)/a(SO2−
4 ) Calculated a(Cl−)/a(SO2−
4 ) Calcu- Calculated Calculated
Based on Measured Based on Measured Based on Measured lated Based on Target pH, T = pH, T =
Solution Concentrations Concentrations Concentrations Concentrations 23°C 90°C
1 (5 g/L 0.011 0.0016 7.2 7.6 0.37 0.50

Fe3+)
2 (5 g/L 0.017 0.0017 9.8 10.4 0.32 0.43
Cu2+)
3 (5 g/L 0.017 0.0019 9.2 9.8 0.35 0.47

Fe2+)
4 (0.5 g/L 0.019 0.0012 15.5 13.9 0.31 0.39
Fe3+)
(A)
Chloride to sulfate activity ratio was calculated two ways: using measured and target concentrations of the solutions.
TABLE 5
Analyzed Values of the Test Solutions After the Tests
Solution H2SO4 (g/L) Cl− (mg/L) Fe3+ (mg/L) Fe(tot) (mg/L) Cu2+ (mg/L) pH at 90°C
1 59.7 1,130 1,306 8,450 16 0.42

2 54.6 1,010 – 66 5,490 0.44
3 51.5 955 1,320 5,630 – 0.43
4 52.1 1,150 270 985 – 0.41
CORROSION 2018.74:225-240.
(HSO−4 ), is formed. At moderate temperatures and (440 mV) was at the same level as in the Fe2+-containing
concentrations, bisulfate ion does not completely solution. The redox potential values systematically
dissociate further to H+ and SO2− +
4 ions, so H and HSO4
− increased by 33% to 65% during heating depending on
remain as the major species in the solution.27 The the solution composition. To sum up Tables 3 and 4,
calculated ratio of the activities of chlorides and sulfates at the test temperature of 90°C, the redox potentials
varied between the solutions. The highest and the were identical for the solutions containing ferric iron
lowest chloride activity to sulfate activity ratios, 15.5 (Solutions 1 and 4), but they differed with respect to the
and 7.2, were determined for the solutions with 0.5 g/L ratio of the activities of chlorides and sulfates, 7.2 and
and 5 g/L ferric iron, respectively. The corresponding 15.5 for the ferric iron concentrations of 5 g/L and
ratios for the copper- and ferrous iron-containing 0.5 g/L, respectively. In the case of the test solutions
solutions were intermediate to those and had relatively containing copper (Solution 2) and ferrous iron (Solu-
equal values, 9.8 and 9.2. The calculated pH values tion 3), both the redox potential and the ratio of the
showed similar trends with temperature to the experi- activities of chlorides and sulfates were equal.
mental ones. However, the lack of a clear correlation Composition of the test solutions after the tests is
between the calculated and the measured pH values in given in Table 5. Comparison of the final and the initial
different solutions reflects the fact that at this high solution compositions enables a few observations to
sulfuric acid concentration, pH is not a sensitive mea- be made. In the solutions containing Fe3+, Solution 1
sure of the solution acidity anymore. (5 g/L) and Solution 4 (0.5 g/L), some of the initial Fe3+
As shown in Table 3, the redox potential, i.e., the transformed into Fe2+, with the redox potential also
oxidation capacity of the solution, was the highest for being lowered as compared to the initial value. The
the solutions with ferric iron: Solution 1 (Fe3+ = 5 g/L) total Fe content was also higher than its initial value,
had the redox potentials of 510 mV and 680 mV at 23°C evidently as a result of the corrosion of the specimens.
and 90°C respectively, and Solution 4 (Fe3+ = 0.5 g/L) In Solution 3 (5 g/L Fe2+), some of the original Fe2+
had the redox potential values of 410 mV and 680 mV. transformed into Fe3+, accompanied by a rise in the
This is in line with the results of Yue, et al.,28 who oxidation capacity. In Solutions 1 (5 g/L Fe3+) and
stated that the redox potential of a Fe3+/Fe2+ couple in 2 (5 g/L Cu2+), a slight increase in all measured
the sulfuric acid solution is a function of the ratio of concentrations was observed. As the contents of all
ferric to ferrous iron and the absolute amounts have constituents were higher than their initial values, this
only a minor influence. The ferric iron-containing indicates a slight evaporation of water during the
solutions were followed in the oxidation capacity by experiment. The increase in the total dissolved iron
Solution 3 (Fe2+ = 5 g/L, 340 mV and 440 mV). The content in Solution 2 was a result of the corrosion
oxidation capacity was lowest for Solution 2 at 23°C of the specimens, similarly to the detected Cu2+ in
(Cu2+ = 5 g/L, 290 mV), but the value at 90°C Solution 1.

800 (a) 7
904L 2205 2507
700
6
600
Potential (mVAg/AgCl)
Mass Loss (g)

500
4
400
3
300
2
200
Solution 1: Fe3+ = 5 g/L OCP(ave) = 700 mV
2+
Solution 2: Cu = 5 g/L OCP(ave) = 410 mV 1
100 Solution 3: Fe2+ = 5 g/L OCP(ave) = 410 mV
3+
Solution 4: Fe = 0.5 g/L OCP(ave) = 580 mV
0 0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 Solution 1 Solution 2 Solution 3 Solution 4
Time (s) (5 g/L Fe3+) (5 g/L Cu2+) (5 g/L Fe2+) (0.5 g/L Fe3+)
Test Environment
FIGURE 1. Open-circuit potential records for the specimens of grade
904L in the test solutions during the first 2 h of the experiment. (b) 7
904L 2205 2507
6
5
The OCP records of the grade 904L electrode,
measured for 2 h from the onset of the experiments in Mass Loss (g) 4
each test solution at the beginning of the experiment,
are presented in Figure 1. The highest values were 3
CORROSION 2018.74:225-240.
obtained in Solutions 1 (5 g/L) and 4 (0.5 g/L) with

2
ferric iron, consistent with the redox potential mea-
surements. Distinct from the observation that the 1
redox potential does not depend on the amount of ferric
0
iron, the OCP of grade 904L was about 100 mV Solution 1 Solution 2 Solution 3 Solution 4
higher in the solution with 5 g/L Fe3+ than with 0.5 g/L (5 g/L Fe3+) (5 g/L Cu2+) (5 g/L Fe2+) (0.5 g/L Fe3+)
Fe3+. The lowest OCP values were detected in Solu- Test Environment
tion 2 (5 g/L Cu2+), again similarly to the redox po- FIGURE 2. Measured mass loss of the specimens of each test
tential, but the difference to Solution 3 (5 g/L Fe2+) material using the two types of crevice formers. (a) Crevice former
was minimal. with flat surfaces. (b) Crevice former with serrated surfaces.
Immersion Crevice Corrosion Tests and

Characterization of the Attack found, but pitting corrosion was detected in one
The measured mass losses are presented in passivated edge (Figures 4[a] and [b]), thus explaining
Figure 2 for flat and serrated crevice formers. Grade the high-mass loss in the absence of crevice corro-
904L experienced measurable mass loss in all test sion. The specimens of test material 2205 experienced
environments when tested with flat crevice formers, but crevice corrosion only in the ferric iron-containing
only in the two ferric iron-containing solutions when solutions (Solutions 1 and 4). Additionally, edge corro-
serrated crevice formers were used. The flat crevice sion was apparent in the solution with 5 g/L ferric
former generated more demanding crevice conditions. iron (Solution 1, Figure 4[c]). In the case of specimens of
For grade 2205, evident mass loss was seen only in the grade 2507, no signs of corrosion attack were ob-
two ferric iron-containing solutions (Solutions 1 and served. Table 6 provides a summary of the observed
4). For grade 2507, no mass loss was detected in any of forms of corrosion.
the test solutions. Based on the appearance of the Figure 3 enables some observations on the
sample surfaces observed with stereomicroscopy, no characteristics of the crevice corrosion attack when
uniform corrosion occurred. This is consistent with using the flat crevice formers, which produced more
the iso-corrosion curves and corrosion data for uniform severe crevice corrosion attack in the specimens than
corrosion given in the Outokumpu Stainless Corrosion the serrated ones. In grade 904L, the extent of attack
Handbook.29 was greatest near the edges of the crevice, but otherwise
Photographs of the specimens are shown in the attack was relatively uniformly spread under-
Figure 3. Crevice corrosion was evident in the speci- neath the crevice former. In grade 2205, the attack
mens of grade 904L in the ferric iron-containing within the covered areas was spread much more
solutions (Solutions 1 and 4) and in the copper- unevenly than in the case of grade 904L, with the
containing solution (Solution 2). In the solution with materials degradation occasionally following micro-
ferrous iron (Solution 3), no crevice corrosion was structural features of the material. These observations

TABLE 6
Summary of the Observed Forms of Corrosion in the Case of Both Crevice Former Types
Flat Crevice Former
Solution Metal Cation 904L 2205 2507
1 5 g/L Fe3+ Crevice corrosion Crevice corrosion & edge corrosion No corrosion
2 5 g/L Cu2+ Crevice corrosion No corrosion No corrosion
3 5 g/L Fe2+ Edge pitting No corrosion No corrosion
4 0.5 g/L Fe3+ Crevice corrosion Crevice corrosion No corrosion
Serrated Crevice Former

Solution Metal Cation 904L 2205 2507
3+
1 5 g/L Fe Crevice corrosion Crevice corrosion No corrosion
2 5 g/L Cu2+ No corrosion No corrosion No corrosion
3 5 g/L Fe2+ No corrosion No corrosion No corrosion
4 0.5 g/L Fe3+ Crevice corrosion Crevice corrosion No corrosion
are supported by the SEM investigations. In grade be 2.3 mm in grade 904L and 0.3 mm in grade 2205.
904L (Figure 5), the edge areas around the crevice Similarly, in Solution 4 (0.5 g/L Fe3+), the attack had
formers were most heavily attacked, with areas progressed to the depths of 1.5 mm and 0.25 mm. In
around the external rather than internal periphery of other words, crevice corrosion was more severe for
the circular crevice former showing the heaviest at- grade 904L than for grade 2205, although the mass
tack (Figures 5[a] and [b]). Cross-sectional examina- losses in the case of Solution 1 (5 g/L Fe3+) predict the
tions revealed that the attack was occasionally spread opposite.
CORROSION 2018.74:225-240.
laterally in the material (Figures 5[c] and [d]). In grade

2205 (Figures 6 and 7), degradation of the material Electrochemical Measurements
occurred more unevenly in the areas under the crevice The test solutions used in the electrochemical
former than in grade 904L, with only very little attack measurements are shown in Table 8. Cyclic polarization
in some areas, but evidently more material loss in some curves for the three stainless steel grades in the base
other locations. Detailed investigations systemati- solution containing 50 g/L H2SO4 and 1 g/L Cl− are
cally revealed selective dissolution (Figures 6[c] and [d], shown in Figure 9. Slight differences were observed
7[c] and [d]). Based on EBSD characterization, aus- between the test materials in the critical values for
tenite was primarily attacked, whereas the areas of the current density, ipp, and potential, Epp, to reach the
ferrite phase remained intact (Figure 8). EDS analy- passive state. The highest ipp and Epp values, 0.6 mA/cm2
ses taken from distinct areas of austenite and ferrite in and −170 mV, respectively, were detected for grade
both duplex grades 2205 and 2507 (Figure 8[a] and 904L. In turn, the lowest ipp, 0.01 mA/cm2, was
Table 7) disclosed compositional differences between obtained for grade 2507, while the lowest Epp, −230 mV,
the two phases. According to the analyses in Table 7, was obtained for grade 2205. Concerning the
austenite was relatively enriched in nickel and man- tendency for crevice corrosion, these results mean
ganese, whereas ferrite contained relatively more that the active peak may be most easily reached in
chromium, molybdenum, and silicon. When converted the case of grade 904L, followed by grade 2205. In the
into PREN values, that of austenite (31 in grade 2205, case of grade 2507, the active region barely existed.
38 in grade 2507) was lower than that of ferrite (37 in However, once the passive state was reached, the
grade 2205, 44 in grade 2507). The effect of nitrogen passive current density ipass was relatively equal in all
was excluded from these calculations as its determi- cases, ranging from 2 μA/cm2 to 5 μA/cm2. In the
nation with EDS is associated with uncertainty. If it is case of duplex steel grade 2205, a shoulder was detected
assumed that the ferrite phase contains the saturation in the beginning of the passive region around 0 mV.
amount of 0.05% N and the rest of N is assigned to the This can relate to the fact that the austenite and ferrite
austenite phase,30 the PREN value of austenite will phases show active dissolution at different poten-
increase only by two units with respect to that of tials.30-31 In contrast, for grade 2507, no such shoulder
ferrite, still demonstrating a substantial difference in in the passive region of the polarization curve could be
PREN values between the two phases. seen. For all three test materials, transpassive disso-
The measured mass loss did not systematically lution started at the same potential, at about
correlate with the extent of crevice corrosion in the 900 mVAg/AgCl. None of the measured curves included a
specimens of grades 904L and 2205 as a result of the hysteresis loop between the forward and reverse
fact that occasionally other corrosion mechanisms re- scans that indicates tendency for pitting and crevice
lated to the edges of the samples were active, too corrosion.
(Figure 4). For example, in Solution 1 (5 g/L Fe3+), the The influence of three key solution constituents,
depth of the crevice corrosion attack was measured to the oxidizing ion Fe3+ content, the chloride ion

(a) (b) (c)

(d) (e) (f)

CORROSION 2018.74:225-240.
(g) (h) (i)
(j) (k) (l)
FIGURE 3. Photographs of the specimens after the testing using flat crevice formers. ([a] through [c]) Solution 1. ([d] through
[f]) Solution 2. ([g] through [i]) Solution 3. ([j] through [l]) Solution 4. The images on the left ([a], [d], [g], and [j]) are of grade
904L, the images in the middle column ([b], [e], [h], and [k]) are of grade 2205, and those on the right ([c], [f], [i], and [l]) are of
grade 2507.
Cl− content, and the sulfate ion SO2− 4 content, on the between chlorides and sulfates (Table 8) and shifted the
ratio of the activities of chloride and sulfate ions is given corrosion potential upward (Figure 10[a]). The active/
in Table 8. Their influence on the polarization curves passive transition was not resolved in the polarization
is shown in Figure 10 using grade 904L as an example. curves with Fe3+ additions, probably because the
The addition of Fe3+ decreased the activity ratio transition was below the OCP and hidden by the

(a)
(b) (c)
200 µm EHT = 15.00 kV Signal A = SE2 Date: 15 Feb 2016 200 µm EHT = 15.00 kV Signal A = SE2 Date: 15 Feb 2016
WD = 14.5 mm Mag = 64× Aperture size = 30.00 µm WD = 16.9 mm Mag = 64× Aperture size = 30.00 µm
FIGURE 4. (a) A photograph and ([b] and [c]) SEM images showing the pitting corrosion in the passivated edges of the
specimens. ([a] and [b]) Grade 904L in Solution 3 (Fe2+ 5 g/L). (c) Grade 2205 in Solution 1 (Fe3+ 5 g/L). Photographs are
from the experiments using flat crevice formers.
CORROSION 2018.74:225-240.
TABLE 7
Average Composition of the Phases Present in Stainless Steel Grades 2205 and 2507 Determined by Five EDS Analyses Each,
and Calculated PREN Values
Grade Fe Cr Ni Mo Si Mn PREN
2205 Austenite 68.0 21.2 6.6 2.4 0.4 1.4 31

2205 Ferrite 66.1 24.0 4.3 4.0 0.5 1.2 37
2507 Austenite 63.6 24.1 8.2 3.1 0.3 0.8 38
2507 Ferrite 66.1 27.1 5.2 4.8 0.3 0.6 44
cathodic part of the curve. The corrosion potential further lowered to 200 mV. The addition of sulfate,
values obtained in the cyclic polarization tests agreed Figure 10(c), slightly decreased the ipp needed for pas-
well with the values obtained in the OCP measure- sivation to the same level as in the base solution
ments. The increase in chloride content, in contrast, (50 g/L H2SO4, 1 g/L Cl−), evidently because of the
raised the activity ratios between chlorides and sul- relatively equal ratio of activities between chlorides
fates and brought clear changes in the behavior of grade and sulfates, Table 8. Simultaneously, ipass was lowered
904L, particularly at chloride contents of 5 g/L and by more than one order of magnitude and the pitting
higher. The curves in the absence of chlorides and at the potential was increased by more than 400 mV, con-
chloride content of 1 g/L were almost equal, but sistent with the observations in the base solution
further increase in the chloride content caused a dra- (50 g/L H2SO4, 1 g/L Cl−).
matic increase in ipp and ipass, and pitting corrosion Keeping in mind the dramatic changes in the
occurred instead of transpassive dissolution. At the behavior of grade 904L at the chloride concentration of
chloride content of 1 g/L, ipp was 0.6 mA/cm2, but 5 g/L and the fact that the crevice solutions typically
increased first to 2.4 mA/cm2 and then further to enrich in chlorides, polarization curves for all three test
5.3 mA/cm2 with the increasing chloride content. The materials were determined at this chloride content
corresponding increase in ipass was higher than in ipp, (Figure 11). Here, dissimilarities in the behavior were
with a rise by more than one order of magnitude, evident between the alloys. Grade 2507 was least
while increasing the chloride content from 1 g/L to 5 g/L, influenced by the high chloride content of the solution,
and even slightly more when reaching the chloride with the forward curve being almost identical to that
content of 10 g/L. With the chloride contents of 5 g/L detected at the lower chloride content (Figure 9) despite
and 10 g/L, grade 904L exhibited pitting corrosion at a more distinct active-passive transition, and the
high potentials instead of transpassive dissolution. The formation of a hysteresis loop between the forward and
pitting potential in the 5 g/L chloride solution was reverse scans at the high potentials. The measured
380 mV and in the 10 g/L chloride solution it was hysteresis loop was, however, closed in the beginning of

(a) (b)
1 mm EHT = 15.00 kV Signal A = SE2 Date: 15 Jan 2016 100 µm EHT = 15.00 kV Signal A = SE2 Date: 15 Jan 2016
(c) (d)
CORROSION 2018.74:225-240.
300 µm EHT = 15.00 kV Signal A = SE2 Date: 22 Mar 2016 300 µm EHT = 10.00 kV Signal A = SE2 Date: 22 Mar 2016
WD = 31.6 mm Aperture size = 30.00 µm Mag = 38× WD = 28.2 mm Aperture size = 30.00 µm Mag = 38×
FIGURE 5. SEM images showing the corrosion attack in grade 904L in Solution 4 with 0.5 g/L Fe3+. ([a] and [b]) Top view.
([c] and [d]) Cross-sectional view. Photographs are from the experiments using flat crevice formers.
the reverse scan but opened at slightly lower poten-

DISCUSSION
tials. This type of behavior is not a manifestation of
pitting corrosion but is rather a result of the occur-
Effect of Sulfate Ions on Localized Corrosion
rence of crevice corrosion in the sample. Repassivation
It is well established that pitting corrosion of
was detected for grade 2507 below the potential of stainless steels occurs when the corrosion potential
about 500 mVAg/AgCl. During the forward scan, the exceeds a critical value Epit in the passive range of the
higher chloride content brought a shoulder to the polarization curve. Epit decreases with the logarithm of
polarization curve at about −80 mV which was not seen the chloride concentration or activity.33 In addition,
with 1 g/L chloride but has been observed for super pH and temperature exert some influence on the critical
duplex stainless steels, e.g., in 3 M HCl.30 pitting corrosion temperature. Sulfate ions inhibit
Grades 904L and 2205 behaved in a relatively pitting corrosion because of the competitive absorption
consistent manner, and the passive currents were the of chloride and sulfate ions,34 and anions with the
same for both grades. Both grades also showed a clear highest charge will be preferentially accumulated at the
active peak. In the case of grade 2205, the active peak metal/solution interface.4 Similarly, the presence of
showed a bimodal shape suggesting that the aus- sulfates has been shown to inhibit crevice corrosion. For
tenite and ferrite phases dissolve at different potentials. example, Brossia and Kelly35 demonstrated the effect
In 4 M H2SO4 + 1 M HCl, active dissolution of the for austenitic 304 stainless steel (UNS S30400) in
ferrite phase was reported to occur at Ecorr + 50 mV and neutral chloride solutions. They stated that the in-
that of the austenite phase at somewhat higher po- hibition of crevice corrosion was primarily a result of the
tential of Ecorr + 150 mV.32 Based on the similarity of the competitive migration and/or absorption between
−
environments this suggests that the austenite phase SO2−
4 and Cl . A similar conclusion is made in Abd El
dissolves at a slightly higher potential than the ferrite Meguid and Abd El Latif,36 explaining the delay in
phase, enabling a smaller IR drop to cause the dis- achieving the critical concentration of chlorides in the
solution of austenite in a crevice. crevice. Pardo, et al.,37 observed that the critical

(a) (b)
1 mm EHT = 15.00 kV Signal A = SE2 Date: 13 Jan 2016 20 µm EHT = 15.00 kV Signal A = SE2 Date: 13 Jan 2016
(c) (d)
CORROSION 2018.74:225-240.
100 µm EHT = 15.00 kV Signal A = SE2 Date: 13 Jan 2016 20 µm EHT = 15.00 kV Signal A = SE2 Date: 13 Jan 2016
FIGURE 6. Top-view SEM images showing the corrosion attack in grade 2205 in Solution 4 with 0.5 g/L Fe3+. Photographs
are from the experiments using flat crevice formers.
pitting and crevice corrosion temperatures of super solutions. This low value suggests that crevice and/or
duplex and super austenitic stainless steel grades pitting corrosion should have occurred in all the
increased when the test solution was acidified with current test environments, which was not the case for
sulfuric acid. grade 2205. For corrosion-resistant high-nickel Alloy
When this study’s immersion test results are 825, a repassivation potential of 50±100 mVAg/AgCl was
compared to the results from immersion tests conducted obtained in 1 g/L Cl− at 95°C,18 which is lower than
for welded specimens in a relatively similar environ- those seen in these tests for any of the steel grades.
ment, they are well in line. Ekman2 immersed creviced When the test solutions are compared to each
stainless steel samples in 10% H2SO4 with 0.7 g/L Cl−, other, it can be noted that they are fairly similar with
5 g/L Fe3+, and 4 g/L Cu2+ at 90°C and observed that respect to the ratios between the activities of chlorides
grade 904L showed pitting corrosion in the heat- and sulfates. Then, the primary factor that governs the
affected zone, but grades 2205 and 2507 were free from occurrence of crevice corrosion is the OCP of the
corrosion. In the solution with 1% H2SO4, 5 g/L Fe3+, material. Considering the IR drop mechanism for
and 4 g/L Cu2+ with a chloride concentration of 1 g/L Cl−, crevice corrosion, Equation (2) states that the higher
and at a slightly higher temperature, 98°C, grades the surface potential, the larger the IR drop that is
904L and 2205 displayed uniform, pitting, and crevice required to initiate crevice corrosion. However, when
corrosion, but grade 2507 showed only minor uniform the polarization curves are investigated, it can be seen,
corrosion and pitting corrosion in the weld. e.g., in Figure 10(b), that the higher surface potential
Comparing the results to investigations per- translates into a higher current which increases the IR
formed at higher temperatures and relatively small drop. The effect of increasing IR surpasses the effect of
chloride concentrations, the beneficial effect of sul- increasing Δ∅* and consequently, the critical condition
fates can be noted. The experimental and modeling is more readily attained in solutions with higher
results of Anderko, et al.,38 give for grade 2205 a oxidizing power, such as in Fe3+-containing solutions,
repassivation potential of 30 mVAg/AgCl at 95°C with compared to the solutions with a lower oxidizing
similar chloride activities as in the current test power, such as Cu2+-containing solutions.

(a) (b)
1 mm EHT = 15.00 kV Signal A = SE2 Date: 14 Mar 2016 1 mm EHT = 15.00 kV Signal A = SE2 Date: 14 Mar 2016
(c) (d)
CORROSION 2018.74:225-240.
20 µm EHT = 15.00 kV Signal A = SE2 Date: 14 Mar 2016 20 µm EHT = 15.00 kV Signal A = SE2 Date: 14 Mar 2016
WD = 33.5 mm Mag = 700× Aperture size = 120.0 µm WD = 33.5 mm Mag = 1.50 K × Aperture size = 120.0 µm
FIGURE 7. Cross-sectional SEM images showing the corrosion attack in grade 2205 in Solution 4 (0.5 g/L Fe3+).
Photographs are from the experiments using flat crevice formers.
(a) (b)
30 µm 30 µm
Mag: 2,000× HV: 20 kV WD: 19.5 mm Mag: 2,000× HV: 20 kV WD: 19.5 mm
FIGURE 8. (a) Cross-sectional SEM image and (b) the corresponding EBSD band contrast map for grade 2205 in Solution 4
(0.5 g/L Fe3+). In (b), the detected face-centered cubic (fcc) phase, i.e., austenite, is shown in red and the body-centered
cubic (bcc) phase, i.e., ferrite, in blue. The images are from the tests performed using serrated crevice formers, but the
phenomenon and the extent of corrosion attack are similar in the specimens with flat crevice formers.
Occurrence of Crevice Corrosion and well established from microelectrode experiments that
Implications of the IR Drop Mechanism corrosion damage spreads after initiation both toward
The maximum loss of thickness in grade 904L the mouth of the crevice but also inward.12 In duplex
was detected at the distance of approximately 0.5 mm to steel grades 2205 and 2507, the loss of thickness was
1.5 mm inward from the crevice mouth (Figure 3). It is homogeneous under the crevice former. The activation

TABLE 8
Calculated Values for the Activities of Chlorides and Sulfates as well as Their Ratios in the Solutions Used for the
Electrochemical Measurements
Solution H2SO4 (g/L) Cl− g/L SO2−
4 (g/L) Fe3+ (g/L) a(Cl−) a(SO2−
4 ) a(Cl−)/a(SO2−
4 )
Base solution 50 1 0.018 0.0012 15.1

Addition of ferric iron 50 1 0.5 0.0066 0.0013 5.3
50 1 10 0.0000025 0.0021 0.0012
Influence of Cl− content 50 0 0 0.0013 0
50 5 0.090 0.0010 87.9
50 10 0.18 0.00087 206.5

Addition of sulfate 50 5 45 0.087 0.0044 19.8
1,200 an aggressive crevice solution. The mobility of chlo-

904L
ride ion is roughly twice that of the sulfate ion
1,000 2205
2507
(DCl− = 2.03 × 10−9 m2 =s andDSO2− 4
= 1.07 × 10−9 m2 =s
800 at room temperature, DCl− = 6.57 × 10−9 m2 =s and
DSO2− = 3.73 × 10−9 m2 =s at 96°C),40 thus it may be

600 4
expected that chloride is the ion moving preferentially to
400 the crevice area. In the current study, this is not as
straightforward as that because the solution is initially
200
highly acidic (pH 0.5) and contains a high amount of
negatively charged sulfate ions in the bulk solution and
CORROSION 2018.74:225-240.
0
also inside the crevice. Actually, the given explanation
–200
holds primarily for the acidification model, whereas for
–400 materials exhibiting active-passive behavior, the IR
1×10–7 1×10–6 1×10–5 1×10–4 1×10–3 1×10–2 1×10–1 1×100 1×101
drop model explains the crevice attack even in the
Current Density (mA/cm2)
complete absence of acidification and chloride accu-
FIGURE 9. Cyclic polarization curves for the test materials. Solution: mulation (yet these are not completely excluded). The IR
50 g/L H2SO4, 1 g/L Cl−, 90°C. Cl− was added as HCl. drop model is therefore well suited for the description
of the studied three stainless steel grades in the ex-
amined leaching environment, particularly as the pH
values are initially so low that minor changes in the H+
by the IR drop mechanism explains the differences in concentration do not cause noteworthy changes
the crevice morphology, given that grade 904L had the in pH.
lowest PREN (with a possibility of pitting within the The application of the IR drop model has some
crevice) and that the duplex grades underwent prefer- implications in the interpretation of the polarization
ential dissolution of the phase with a lower PREN. data, as the active-passive transformation and the
When examining the polarization curves measured in influence of various ions (oxidizing species, chlorides,
the bulk solution (Figure 10), it can be seen that the sulfates) on it are in a key role. As shown in Figure 10,
required IR drop for crevice corrosion to occur in such a the addition of Fe3+ shifts the OCP radically upward,
solution is very high, of the order of 800 mV in the hiding the clear active-passive transition from the
case of 5 g/L ferric iron (Solution 1) and almost 700 mV polarization curve of grade 904L. The increase in the
in the case of 0.5 g/L ferric iron (Solution 4). The chloride content raises the chloride activity and
required IR drop is by far higher than what has been increases the active current densities. The addition of
typically measured, e.g., by microelectrode arrays.11 sulfate (45 g/L) restores the ratio of activities between
Consequently, it can be expected that the materials are chlorides and sulfates to the same level as in the base
spontaneously passive in the bulk solution, but the solution (50 g/L H2SO4, 1 g/L Cl−) and brings the
potential drop associated with the critical crevice current density values within the active area to the same
dimensions (the critical depth and the width of the level as in the base solution. However, because of the
crevice) brings them into the active region. Indeed, as fact that the ionic species take part in the reactions
stated by Engelhardt and Macdonald,39 the rate of through their activities which are determined by the
anodic dissolution of the metal determines the depth of equilibrium chemistry of the solution, the influence of
attack inside the corroding crevices. chloride additions on the ratio of activities between
It is commonly accepted that the acidification of chlorides and sulfates, and on the corrosion behavior of
the solution and the migration of negatively charged the test materials is much stronger than that of
chloride anions to balance the effect of the positively sulfates. Therefore, when negatively charged species are
charged metal cations are responsible for the creation of attracted into the crevice, the corrosion effects of

1,200
(a) 1,200 No Fe3+ 904L
1,000 0.5 g/L Fe3+ 1,000 2205

10 g/L Fe3+ 2507
800
800
600 600
400 400
200 200
0 0
–200 –200
–400 –400
1×10–6 1×10–5 1×10–4 1×10–3 1×10–2 1×10–1 1×100 1×101 1×10–5 1×10–4 1×10–3 1×10–2 1×10–1 1×100 1×101 1×102
2
Current Density (mA/cm ) Current Density (mA/cm2)
(b) 1,200 FIGURE 11. Cyclic anodic polarization curves for the three stainless
904L 5% H2SO4 904L 5% H2SO4 + 1 g/L Cl–
904L 5% H2SO4 + 5 g/L Cl– 904L 5% H2SO4 + 10 g/L Cl
– steel grades in the solution containing 50 g/L H2SO4 and 5 g/L
1,000
Cl−, 90°C.
800
600
400
increases the corrosion potential of the material to
various extents. Second, the ions interact with the
200
other constituents of the occluded cell solution. The
CORROSION 2018.74:225-240.
0 addition of ferric or ferrous ions changes the con-

centration of ions naturally released from stainless
–200
steels during corrosion. In the case of the duplex
–400 stainless steels, the presence of copper brings also a new
0.0001 0.001 0.01 0.1 1 10
ion into the bulk and occluded cell solutions as they
are not alloyed with copper. Conversely, grade 904L is
(c) 1,200 904L 5% H2SO4 + 5 g/L Cl– + 45 g/L SO4
2– alloyed with copper. Under neutral conditions, copper
1,000 904L 5% H2SO4 + 5 g/L Cl– has not been detected to enrich into the crevice corroded
surface.41 Under highly acidic conditions the en-
800
richment can, however, be expected as copper is alloyed
600 to improve uniform corrosion resistance in sulfuric
400
acid media. The concentrations that can enter through
dissolution into the occluded cell solution can still be
200 minor compared to the concentration in the bulk so-
0 lution, 5 g/L. Other constituents of the occluded cell
solution can be thought to be nickel and chromium
–200
when allowing for congruent dissolution.
–400 Considering oxidation capacity, it is well known
0.001 0.01 0.1 1 10
that a higher potential conveys to more extensive local
damage. In this respect it is expected that the solu-
FIGURE 10. Anodic polarization curves for grade 904L in the solution tions with the ferric ions generate the most severe
containing 50 g/L H2SO4, 90°C. (a) Cl− content 1 g/L. Without Fe3+, crevice corrosion damage, as indeed was observed for
with 0.5 g/L Fe3+, and 10 g/L Fe3+. (b) As a function of Cl− content. (c) grades 904L and 2205. However, the redox potentials
With and without SO2− 4 addition. and the ratio of the activities of chlorides and sulfates
in the two remaining solutions, the one with copper ions
(Solution 2) and the other one with ferrous ions
chlorides dominate those of sulfates. Considering the (Solution 3), were similar, but crevice corrosion was
observed only in grade 904L in the copper-containing
greater mobility of the chloride ions compared to the
solution (Solution 2). This suggests that the ion-ion
sulfate ions, such a situation is highly probable.
interactions in the occluded cell solution generate
more aggressive conditions in the presence of copper
Effect of the Metal Cations ions than in the presence of ferrous ions. One possible
The presence of the oxidizing ions, i.e., ferric or explanation for the differences in the occurrence of
cupric ions, may influence the localized corrosion of crevice corrosion in Fe2+- and Cu2+-containing solu-
stainless steels in two ways. First, their presence tions, when the ratios of the activities of chlorides and

sulfates and the OCP values are similar, may relate to In the case of duplex steel grade 2205, the
the difference in the complexing tendency of Fe2+ and presence of the two phases with dissimilar resistance to
Cu2+. In aqueous solutions, metal ions make com- corrosion, evaluated on the basis of the PREN values,
plexes with water molecules. Chloride ions can substi- explains the uniform lateral spreading of the corrosion
tute one or several water molecules depending on the attack within the crevice. The austenite phase has a
reaction constants; thus, chloride may also act as a much lower PREN value than that of the overall bulk
ligand for metal ion coordinates. For Fe2+, only one material, and active dissolution of austenite occurs at
water ligand can be substituted by chloride, giving a slightly higher potentials than that of ferrite, i.e., a
formation constant of log K1 = 0.36.42 For copper, the smaller IR drop is required to dissolve the austenite
formation constant for the first water substitution by phase, and the attack proceeds along the evenly dis-
chloride is log K1 = 0.1 and the second cumulative tributed austenite areas. Although the experimental
formation constant log K2 = −0.6. This indicates that the PREN values rely on the EDS spot analyses where part of
complex of Fe2+ and Cl− is more stable than that the collected information may come from the sur-
between Cu2+ and Cl−. Then, one may expect that in the roundings, the PREN values provide enough relative
Fe2+-containing solution there are fewer free chloride information to enable the comparison, even though
ions that can interact with the passive film of the they are not absolute PREN values. Similarly for the
stainless steel, and this reduces the crevice corrosion duplex steel grade 2507, austenite was the phase with
tendency, which was seen experimentally. the lower PREN value and it was thus more prone to
localized corrosion than the ferrite phase.
Comparison Between the Stainless Steel Grades
The PREN value predicts the resistance of a CONCLUSIONS
stainless steel to localized corrosion attack. The
obtained results proved that PREN of 43 in the case of The investigation of crevice corrosion behavior of
CORROSION 2018.74:225-240.
grade 2507 was high enough to avoid crevice corrosion three stainless steel grades: 904L, 2205, and 2507, in a
in the aggressive environment containing 50 g/L leaching environment based on H2SO4 containing
H2SO4 and 1 g/L Cl− despite the presence of oxidizing chlorides and oxidizing ions enables the following con-
ions. Although the oxidizing ions were demonstrated clusions to be made:
to ennoble the corrosion potential of the materials sig- v The extent and nature of crevice corrosion attack
nificantly, the passive region detected for grade 2507 varies between the steel grades of nominally identical
was so wide that the passive-active transition was PREN values (austenitic 904L and duplex 2205 have
clearly prevented. Even in the concentrated chloride PREN values 34 and 35, respectively) and between the
solution, the passivity was preserved well above the solutions with dissimilar chemistries (ratio of activi-
redox potential of the Fe3+-containing test solutions. ties of chlorides and sulfates, type of oxidizing ions). The
In the case of grades 904L and 2205, the PREN duplex steel grade 2507 (PREN 43) is not susceptible
values were almost equal, but there were differences in to crevice corrosion under any of the studied conditions.
the degradation mechanism resulting from micro- v Grade 904L undergoes the most severe crevice
structural dissimilarities. With respect to the electro- corrosion among the studied alloys, with the corrosion
chemical response, the two alloys behaved relatively attack being evident in the solutions containing ferric
consistently in the solution containing 50 g/L H2SO4 and iron (5 g/L and 0.5 g/L Fe3+) and cupric ions (5 g/L Cu2+),
1 g/L Cl− and in the concentrated chloride solution. together with chlorides (1 g/L Cl−).
Minor differences in the parameters enabling the active- v In the duplex stainless steel 2205, crevice corro-
passive transition and the initiation of pitting corrosion is evident in the solutions containing ferric iron
sion were detected between the alloys and these may (5 g/L and 0.5 g/L Fe3+). Here, the ratio of activities of
have some role in the observations between the four chlorides and sulfates is particularly important with
test solutions. However, what is important with regard to respect to the extent of corrosion attack, with a lower
the crevice corrosion susceptibility is the significant ratio leading to greater material loss. The austenite and
dependence of the behavior of these two materials on the ferrite phases undergo dissolution at different rates,
chloride concentration, as demonstrated in Figures 9 austenite phase being preferentially attacked in the
and 11. For both materials, the passive region shrinks studied leaching environments.
and the passive current ipass increases with increasing v The influence of enrichment of chlorides and sul-
chloride content. In the case of austenitic alloy 904L, fates in the crevice on the behavior of the test materials
the most severe damage was detected within some is demonstrated through anodic polarization mea-
distance from the crevice mouth, which is consistent surements in concentrated chloride solutions. Although
with the current understanding of the damage evo- the passive region is significantly influenced by the
lution.43-44 Indeed, the external periphery of the crevice concentration of these ions, the active peak shifts
former was surrounded by a larger cathodic area than to higher (increase in chloride content) or lower
the inner periphery, explaining the location of the (increase in sulfate content) current densities upon
damage near the crevice mouth. altering their amount and ratio of activities.

ACKNOWLEDGMENTS 18. D.S. Dunn, G.A. Cragnolino, N. Sridhar, Corrosion 56, 1 (2000):
p. 90-104.
19. F. Bocher, R. Huang, J.R. Scully, Corrosion 66, 5 (2010):
The work was performed within the Digital, Internet, p. 055002-1 to 055002-15.
Materials & Engineering Co-Creation platform (DIMECC 20. M.A. Rodriguez, R.M. Carranza, R.B. Rebak, Corrosion 66, 1 (2010):
p. 015007-1 to 015007-14.
Oy) and its Breakthrough Steels and Applications (BSA) 21. U. Steinsmo, T. Rogne, J.M. Drugli, P.O. Gartland, Corrosion 53, 1
program. We gratefully acknowledge the financial sup- (1997): p. 26-32.
port from the Finnish Funding Agency for Technology 22. D. Han, Y. Jiang, B. Deng, L. Zhang, J. Gao, H. Tan, J. Li, Corrosion
67, 2 (2011): p. 025004-1 to 025004-7.
and Innovation (Tekes) and participating companies. 23. D. Han, Y. Jiang, C. Shi, Z. Ji, J. Li, Corros. Sci. 53 (2011): p. 3796-
VTT Technical Research Centre of Finland Ltd is also 3804.
24. Y.Z. Yang, Y.M. Jiang, J. Li, Corros. Sci. 76 (2013): p. 163-169.
acknowledged for funding. At VTT, the technical staff,
25. J.N. Al-Khamis, H.W. Pickering, J. Electrochem. Soc. 148, 8 (2001):
particularly Tuomo Kinnunen, is thanked for conduct- p. B314-B321.
ing the electrochemical measurements. 26. HSC Software, version 8.1, www.outotec.com/hsc.
27. H. Sippola, “Thermodynamic Modelling of Aqueous Sulfuric Acid”
(Ph.D. thesis, Aalto University, Finland, 2015).
28. G. Yue, L. Zhao, O.G. Olvera, E. Asselin, Hydrometallurgy 147-148
REFERENCES (2014): p. 196-209.
29. Outokumpu Stainless Corrosion Handbook, 10th ed. (Helsinki,
1. H.R. Watling, Hydrometallurgy 140 (2013): p. 163-180. Finland: Outokumpu, 2009).
2. S. Ekman, E. Torsner, “Special Stainless Steel Grades for the 30. R.A. Perren, T.A. Suter, P.J. Uggowitzer, L. Weber, R. Magdowski,
Hydrometallurgical Industry,” Proceedings of Copper 2010 (Ham- H. Böhmi, M.O. Speidel, Corros. Sci. 43 (2001): p. 707-726.
burg, Germany: GDMB, 2010), p. 1857-1870. 31. E. Symniotis, Corrosion 46, 1 (1990): p. 2-12.
3. Z. Szklarska-Smialowska, Pitting and Crevice Corrosion (Houston, 32. M. Feminia, J. Pan, C. Leygraf, P. Luukkonen, Corros. Sci. 43
TX: NACE International, 2005). (2001): p. 1939-1952.
4. J.R. Galvele, J. Electrochem. Soc. 123, 4 (1976): p. 464-474. 33. H.P. Leckie, H.H. Uhlig, J. Electrochem. Soc. 113 (1966):
5. H.W. Pickering, J. Electrochem. Soc. 150, 5 (2003): p. K1-K3. p. 1261-1267.
6. B.A. Kehler, J.R. Scully, Corrosion 61, 7 (2005): p. 665-684. 34. S.-I. Pyun, S.-M. Moon, J. Solid State Electrochem. 3 (1999):
7. N.J. Laycock, J. Stewart, R.C. Newman, Corros. Sci. 39, 10-11 p. 331-336.
CORROSION 2018.74:225-240.
(1997): p. 1791-1809. 35. C.S. Brossia, R.G. Kelly, Corrosion 54, 2 (1998): p. 145-154.
8. M.I. Abdulsalam, Corros. Sci. 47 (2005): p. 1336-1351. 36. E.A. Abd El Meguid, A.A. Abd El Latif, Corros. Sci. 46 (2004):
9. M.I. Abdulsalam, H.W. Pickering, Corros. Sci. 41 (1999): p. 351- p. 2431-2444.
372. 37. A. Pardo, E. Otero, M.C. Merino, M.D. Lopéz, M.V. Utrilla, F.
10. A.M. Al-Zahrani, H.W. Pickering, Electrochim. Acta 50 (2005): Moreno, Corrosion 56, 4 (2000): p. 411-418.
p. 3420-3435. 38. A. Anderko, N. Sridhar, M.A. Jakab, G. Tormonen, Corros. Sci. 50
11. N. Sridhar, D.S. Dunn, Corrosion 50, 11 (1994): p. 857-872. (2008): p. 3628-3647.
12. F. Bocher, F. Presuel-Moreno, J.R. Scully, J. Electrochem. Soc. 155, 39. G.E. Engelhardt, D.D. Macdonald, Electrochim. Acta 65 (2012):
5 (2008): p. C256-C268. p. 266-274.
13. K. Cho, H.W. Pickering, J. Electrochem. Soc. 138, 10 (1991): 40. S.-K. Lee, “The Coupled Environment Models for Localized
p. L56-L58. Corrosions: Crevice Corrosion and Stress Corrosion Cracking”
14. T. Kaji, T. Sekai, I. Muto, Y. Sugawara, N. Hara, J. Electrochem. Soc. (Ph.D. diss., Pennsylvania State University, 2013).
159, 7 (2012): p. C289-C297. 41. S. Zakipour, C. Leygraf, Corrosion 37, 1 (1981): p. 21-28.
15. H.W. Pickering, Corrosion 42, 3 (1986): p. 125-140. 42. J.A. Dean, Lange’s Handbook of Chemistry, 14th ed. (New York, NY:
16. K. Cho, M.I. Abdulsalam, H.W. Pickering, J. Electrochem. Soc. 145, McGraw-Hill Inc., 1992), p. 8.83-8.97.
6 (1998): p. 1862-1869. 43. G.F. Kennell, R.W. Evitts, Electrochim. Acta 54 (2009): p. 4696-
17. M.I. Abdulsalam, H. Alghamdi, S.T. Arab, Y.T. Al-Janabi, “Sus- 4703.
ceptibility of Superaustenitic Stainless Steel Alloys to Crevice 44. W. Sun, L. Wang, T. Wu, G. Liu, Corros. Sci. 78 (2014): p. 233-243.
Corrosion in Chloride Solutions,” Proceedings of EuroCorr 2011
(Stockholm, Sweden: European Corrosion Congress, 2011).

Crevice Corrosion of Stainless Steels 904L, 2205, 2507 in High Temperature Sulfuric Acid Solution Containing Chlorides

Uploaded by

Copyright:

Available Formats

Crevice Corrosion of Stainless Steels 904L, 2205, 2507 in High Temperature Sulfuric Acid Solution Containing Chlorides

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Crevice Corrosion of Stainless Steels 904L, 2205, 2507 in High Temperature Sulfuric Acid Solution Containing Chlorides

Uploaded by

Copyright:

Available Formats

CORROSION SCIENCE SECTION

Crevice Corrosion of Stainless Steels 904L,

2205, and 2507 in High-Temperature Sulfuric

M. Lindgren,‡,* E. Huttunen-Saarivirta,** H. Peltola,* J. Romu,*** T. Sarikka,*** H. Hänninen,*** and P. Pohjanne**

226 CORROSION—FEBRUARY 2018

904L 1.4539 Austenitic 0.01 – 20 25 4.3 1.5Cu 34

of a single sample. attachment of the crevice formers, the specimens

CORROSION—Vol. 74, No. 2 227

1 52.7 942 5,100 5,110 <1 507 676 0.31 0.66

228 CORROSION—FEBRUARY 2018

1 (5 g/L 0.011 0.0016 7.2 7.6 0.37 0.50

3 (5 g/L 0.017 0.0019 9.2 9.8 0.35 0.47

1 59.7 1,130 1,306 8,450 16 0.42

CORROSION—Vol. 74, No. 2 229

Mass Loss (g)

obtained in Solutions 1 (5 g/L) and 4 (0.5 g/L) with

Immersion Crevice Corrosion Tests and

230 CORROSION—FEBRUARY 2018

Serrated Crevice Former

laterally in the material (Figures 5[c] and [d]). In grade

CORROSION—Vol. 74, No. 2 231

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

(j) (k) (l)

232 CORROSION—FEBRUARY 2018

2205 Austenite 68.0 21.2 6.6 2.4 0.4 1.4 31

CORROSION—Vol. 74, No. 2 233

the reverse scan but opened at slightly lower poten-

234 CORROSION—FEBRUARY 2018

CORROSION—Vol. 74, No. 2 235

236 CORROSION—FEBRUARY 2018

Base solution 50 1 0.018 0.0012 15.1

50 10 0.18 0.00087 206.5

1,200 an aggressive crevice solution. The mobility of chlo-

DSO2− = 3.73 × 10−9 m2 =s at 96°C),40 thus it may be

CORROSION—Vol. 74, No. 2 237

1,000 0.5 g/L Fe3+ 1,000 2205

0 addition of ferric or ferrous ions changes the con-

238 CORROSION—FEBRUARY 2018

CORROSION—Vol. 74, No. 2 239

240 CORROSION—FEBRUARY 2018

You might also like

M. Lindgren,‡,* E. Huttunen-Saarivirta,** H. Peltola,* J. Romu,* T. Sarikka,* H. Hänninen,* and P. Pohjanne