(3) the CO 2 +H 2 S mixture condition with 3.12 x 10-4 mol H 2 S and 0.
312
Effects of CO2 and H2S on
Corrosion of Martensitic Steels
in Brines at Low Temperature
Ruishu Feng,*,†,** Justin R. Beck,†,** Derek M. Hall,†
Aysel Buyuksagis,† Margaret Ziomek-Moroz,‡,*** and
Serguei N. Lvov,*,†,§,***
Department of Energy and Mineral Engineering, The EMS
Energy Institute, Department of Materials Science and
Engineering, The Pennsylvania State University.
U.S. Department of Energy, National Energy Technology
Laboratory
ARTICLE INFO
Article history:
Received Day Month Year
Accepted Day Month Year
Available Day Month Year
Keywords:
A. Sweet and sour corrosion
B. Corrosion rate
C. Corrosion mechanism
D. Martensitic carbon steels
E. Cold-climate corrosion
*Department of Energy and Mineral Engineering, The Pennsylvania State
University, University Park, PA 16802, USA
†The EMS Energy Institute, The Pennsylvania State University, University Park,
PA 16802, USA
‡U.S. Department of Energy, National Energy Technology Laboratory, 1450
Queen Avenue SW, Albany, OR 97321, USA
§Department of Materials Science and Engineering, The Pennsylvania State
University, University Park, PA 16802, USA
**First authors: Ruishu.Feng@NETL.DOE.GOV (R. Feng); jrb1005@psu.edu (J. R.
Beck).
***Corresponding authors: Tel.: +1 814 863 8377; Fax: +1 814 865 3248; Email:
lvov@psu.edu (S. N. Lvov). Tel.: +1 541 967 5943; Fax: +1 541 967 5880; Email:
Margaret.Ziomek-Moroz@NETL.DOE.GOV (M. Ziomek-Moroz).
ABSTRACT
Corrosion studies were conducted for martensitic carbon steels in
brine solutions at 4 °C and 10 MPa (1450 psi) to simulate the
subsurface environments that can be encountered in Arctic drilling.
Three conditions with a 5 %wt. NaCl brine were used: (1) the CO 2 only
condition with 0.312 mol of CO 2 per mol of H 2 O in brine, (2) the H 2 S
only condition with 3.12 x 10-4 mol of H 2 S per mol of H 2 O in brine, and
mol CO 2 per mol of H 2 O in brine. Two martensitic carbon steels were
selected for the investigation: a high strength low alloy carbon steel
commonly used for drill pipes, S-135 (G41000), and a newly designed
ultra-high strength low alloy carbon steel, UD-165 (G41300). The
experimental results showed that CO 2 :H 2 S ratios as high as 1000 can
result in predominant sour corrosion at 4 oC and 10 MPa.
Electrochemical and mass loss measurements showed a distinct drop
in corrosion rate (CR) when transitioning from sweet to sour corrosion,
from the order of 0.5 to 2 mm y-1 in the CO 2 only condition to the
order of 0.05 mm y-1 in the H 2 S-containing conditions. The linear
sweep voltammetry results indicated that the reaction mechanism was
also determined by H 2 S even when its concentration was lower than
CO 2 by three orders of magnitude due to specific adsorption of H 2 S.
For the Fe dissolution reaction, H 2 S(aq) probably followed the two-
electron mechanism, and CO 2 (aq) could have followed the two-
electron mechanism in the low polarization region and the one-
electron mechanism in the high polarization region. The cathodic Tafel
slopes at 4 oC were much larger in magnitude than the reported values
at 20 to 30 oC in literature, and the cathodic mechanisms need further
study at 4 oC and high pressure conditions. The modeling software was
used to simulate the investigated corrosion. A good agreement was
found between the experimental CR and the modeled CR for the CO 2
only and H 2 S only cases. However, the model predicted that CR in the
mixed case would match CR in the CO 2 only condition instead of the
H 2 S only condition. While the experimental corrosion behavior was
strongly dependent on the H 2 S content, the corrosion potential was
instead related to the CO 2 content and the brine pH. The inhibiting
effect of H 2 S on CR could be attributed to the formation of a protective
thin sulfide film. The corrosion products for all tests were found to be
quickly oxidized on exposure to air when the experimental system was
disassembled.
INTRODUCTION
The presence of CO 2 and H 2 S in condensed water phases has been of
considerable interest to scientists and engineers dealing with corrosion
in oil and gas operations, from production to processing.1–11 One of the
major effects of sweet corrosion is that dissolved CO 2 lowers the
solution pH and increases the rate of corrosion.12–16 Sour corrosion,
where H 2 S is present, has generally been found to reduce corrosion
rates when the H 2 S concentration is low.17–23 The proposed
mechanism for this drop is the formation of a relatively protective iron
sulfide film. However, while the reduced material loss may be
welcome, the tradeoff of H 2 S content is an increased susceptibility to
sulfide stress cracking (SSC). With drastically different failure modes
being possible, the corrosion behavior in mixed gas systems becomes
important for proper design of corrosion prevention and mitigation.
Currently, no widespread law or model exists for determining the
transition between sweet and sour corrosion. NACE SP0110 Appendix
A describes a rough rule for determining the sweet-sour transition by
the CO 2 :H 2 S ratio based on previous experimental results.24 A ratio
above 500 would be expected to produce predominantly sweet
corrosion, while a ratio below 20 would be expected to show
predominantly sour corrosion. A ratio between 20 and 500 may allow
for a mixed corrosion mechanism where both FeCO 3 and FeS could
coexist. Smith investigated the assumption of 500 on the upper end of
the CO 2 :H 2 S ratio.25 Thermodynamic calculations on the stability of
FeCO 3 and FeS showed that while a ratio of 500 could be justified, it
was found to have minor to moderate sensitivity to temperature and
solution ionic strength and extreme sensitivity to the thermodynamic
data quality. An example was that a 1 % change in the Gibbs free
energy value for FeCO 3 could change the ratio by over a factor of 10.
Another consideration beyond the corrosion products stability is
the kinetics for the corrosion process. Work has continued on
laboratory experiments investigating the kinetic effects of different
CO 2 :H 2 S ratios over a wide range of conditions. Experiments have
shown that CO 2 :H 2 S ratios from 500 to as high as 10,000 can still see a
decrease in corrosion rate by almost an order of magnitude relative to
the pure CO 2 condition.17–23 Results at higher H 2 S concentrations have
ranged from no significant effects to notable increases in corrosion
rates, though the corrosion rates in CO 2 -H 2 S mixtures were generally
lower than in pure CO 2 .18,19 While a decrease in material loss may be
favorable, the presence of the protective iron sulfide films can also
make the underlying steels more susceptible to pitting.15,26,27 The
transition to sour corrosion can also increase the risk of sulfide stress
cracking,4 and the recommendations from NACE MR0175 / ISO 15156
should be taken into account for selecting cracking-resistant
materials.28
The cold-climate corrosion oil country tubular goods is a great
concern for the integration and safety of oil and natural gas
exploration and production in low-temperature climates. Although the
temperatures are lower, the higher dissolved gas concentrations can
increase the risk of corrosion and failure, such as H 2 S and CO 2 in deep
wells.29,30 The cold-climate regions include more than the designated
Arctic and Antarctic areas. Perrigo suggested the regions defined by
the U.S. Army Cold Regions Research and Engineering Laboratory
(CRREL), where the approximate average temperature in the coldest
month is 0 oC (32 oF) or lower.30,31 Thus, the cold-climate corrosion is
prevalent beyond Arctic and Antarctic areas.
The purpose of the present work was to provide data and
discussion on corrosion of a high strength low alloy carbon steel, grade
S-135 (G41000), and an ultra-high strength low alloy carbon steel,
grade UD-165 (G41300) in sweet and sour environments at lower
temperatures related to offshore oil and gas operations in cold
climates. The tests investigated corrosion in brines at equilibrium with
CO2, H2S, and a mixed phase at a CO2:H2S ratio of 1000 at 4 oC and 10
MPa total pressure using electrochemical techniques, mass loss
measurements, software modeling, and surface analyses.
EXPERIMENTAL
Test Conditions
Representative of low temperature offshore drilling
conditions, a temperature of 4 °C was selected with a total system
pressure of 10MPa. The test solution contained 5 %wt. NaCl (0.9 mol
kg-1). The sweet/sour/mixed conditions were prepared by holding the
test solution in equilibrium with one of three non-aqueous phases:
pure CO2, a H2S/Ar mixture, and a CO2/H2S/Ar mixture. It should be
noted that 10 MPa was above the CO2 critical pressure (7.38 MPa) but
4 °C was below the CO2 critical temperature (31 °C).32 Care should be
taken to describe the system at high pressures when referring to such
substances as “gas” or “vapor” phases, and describing in terms of
molar fraction may reflect their compositions in the system more
accurately.
Using the molar fraction, the conditions and test procedure
were designed to allow consistent dosing of substances while
maintaining concentrations representative of deep-well
environments.10 Deep wells are predicted to encounter total pressures
up to and even above 100 MPa, and the “gas” phase may contain over
10 % CO2 and 0.01 % H2S. The following assumption was employed for
this study. The test systems were initially calculated to represent
exposure to 10 MPa CO2 and/or 0.01 MPa H2S (either gas or condensed
phase). A 1:1 volume ratio of aqueous:non-aqueous phase was
considered as coming to equilibrium. At equilibrium, each component
existed in all the phases inside the system. The total amount of each
component in the whole system was calculated, and the system could
be defined through the mol X:mol H2O ratios, where X represents the
non-water species such as NaCl, CO2, H2S, and Ar.
The initial calculation results had to be adapted to account for
physical limitations in the system when maintaining the proper mol
ratios. The final ratios used for laboratory testing were 0.312 mol
CO2:mol H2O and 3.12 x 10-4 mol H2S:mol H2O (maintaining CO2:H2S
=1000:1 for the mixed case). The Ar:H2O ratio was not controlled, but
rather argon was used to dilute H2S and balance the pressure between
tests. Phase equilibria calculations were performed using commercial
thermodynamic modeling software to determine the final amounts of
each species needed to achieve the internal autoclave volume (476 mL
after assembly) at 4 °C and a total pressure of 10 MPa. The system
compositions (in mol) for the three test environments are given in
TABLE 1.
Investigated Materials
Two martensitic drilling steels were tested: a high strength low
alloy carbon steel commonly used for drill pipes, grade S-135 (G41000),
and a newly designed ultra-high strength low alloy carbon steel, grade
UD-165 (G41300). Their chemical compositions are given in TABLE 2.
The primary differences in composition were lower chromium and
higher manganese, nickel, copper, and vanadium in UD-165 than in S-
135. The level of sulfur was also higher for UD-165. This resulted in UD-
165 with higher yield strength and ultimate tensile strength.33
These two steels were used as working electrodes with
dimensions of 10x10x5 mm for the electrochemical tests. One 10x10
mm face served as the electrode surface exposed to the test solutions,
and a 304 stainless steel (UNS S30400) wire was attached to the
opposite face as the electrical lead. The lead was coated with
shrinkable polytetrafluoroethylene (PTFE) tubing. Two to three steel
electrodes were placed into a PTFE holder and all sides other than the
10x10 mm electrode surface were sealed to the PTFE using an epoxy.
Mass loss samples were prepared using 25x20x2 mm plates. A 3
mm ID hole was drilled near the top of the plate so that the samples
could be suspended from a PTFE-coated wire. The electrode surface
and mass loss sample were both ground with 600 grit SiC paper and
then cleaned for 30 to 60 minutes in acetone inside an ultrasonic bath.
System Setup
The electrochemical measurements were performed using a
three-electrode system. The working electrodes consisted of two or
three steel samples prepared as described above. The counter
electrodes were 20x10x1 mm platinum plate or mesh. The electrode
leads were coated with PTFE. The counter electrode leads were
inserted through the PTFE assembly edge to place each counter
electrode parallel to a working electrode in the system. The schematic
diagram of the autoclave system is shown in FIGURE 1. Two mass loss
samples were suspended by PTFE-coated wire from the autoclave lid,
with care taken to ensure that they would remain in the aqueous
phase without contacting any of the inner surfaces.
 Custom double-junction silver-silver chloride electrodes were
prepared for the high pressure systems. The filling solution was 5 %wt.
NaCl(aq) to reduce the liquid junction potential between the reference
electrode and test solutions. The inner junction of the reference
electrode consisted of the Ag/AgCl electrode submerged in solution-
saturated zirconia sand inside an alumina tube. The top of the tube
was sealed with high temperature epoxy, and the bottom was
supported using a porous zirconia-magnesia cement. The second
junction was prepared by sealing a shrinkable PTFE tube around the
outside of the ceramic tube. Inside the extended PTFE tube bottom
was filled again with solution-saturated zirconia sand, and either
another cement frit or a perforated PTFE plug was used to support the
second junction. Two reference electrodes were used for each test in
case one failed before the end of the exposure, and fresh electrodes
were prepared for each assembled autoclave. The reference electrode
potential at 4 °C was calculated to be 0.248 V vs the standard hydrogen
electrode (SHE) using the data from OLI Analyzer. The results were
confirmed through comparison with calculation performed using
thermodynamic properties and activity coefficients from literature.34–36
All measurements were performed in a 600 mL autoclave vessel.
A PTFE liner was used to isolate the test solution from the autoclave
vessel. Temperature was controlled using a circulating chiller and a
copper coil cooling loop wrapped around the autoclave vessel. All test
systems were stirred using an impeller rotating at around 150
revolution per minute (RPM). A computer-controlled Gamry Reference
600 potentiostat was used to perform the electrochemical
measurements.
Experimental Procedure
To achieve 10 MPa total pressure and the system composition
shown in TABLE 1, the experimental procedure was as follows. First,
the autoclave was assembled dry with all electrodes and mass loss
samples. Only one kind of steel (either S-135 or UD-165) was present in
each assembled autoclave. The desired amount of NaCl(s) was placed
at the vessel bottom before it was sealed. Then, the sealed vessel was
then purged of oxygen by repeated and slow pressurization and
depressurization cycles up to 2 MPa (290 psi). Care had to be taken to
prevent gas bubbles from forming at the reference electrode junctions.
This was believed to have caused failure in some of the reference
electrodes by increasing internal resistance. During this process, the
cooling was started to bring the system down to 4 °C. This usually
involved letting the system chill overnight to ensure it was at a stable
temperature for gas injection. The temperature stability was typically
within ±1 °C, though on hot days the variability could be as high as ±3
°C.
After the autoclave was deaerated and cooled, the system
pressure was brought down to just above atmospheric pressure to
maintain positive pressurization, and H2S was added (if needed) using
a 10 %vol. H2S in Ar blend. The desired pressure was calculated using
the autoclave practical total volume of 476 mL after assembly and the
ideal gas law at 4 °C. In the H2S tests, the H2S addition was followed by
further addition of Ar to help balance the final pressure. The deaerated
deionized water was injected into the autoclave vessel using a high
performance liquid chromatography (HPLC) pump. The deionized
water was deaerated with Ar for at least one hour. The HPLC pump
was set at a constant flow rate of 10 mL min-1 and run until the
desired amount of water was transferred to the system. Then the
5%wt. NaCl solution formed inside the vessel as the aqueous phase.
For the H 2 S tests, water injection was the final step and was used to
bring the system to its final pressure. It should be noted that while the
target pressure was 10 MPa, the actual pressure was in the range of 6
to 8 MPa due to small errors in water addition and in determining the
autoclave practical internal volume after assembly. These errors were
magnified due to the small non-aqueous volume relative to the
incompressible aqueous phase.
For the CO 2 and CO 2 +H 2 S tests, CO 2 was added as the final step
after the water injection to bring the system up to 10 MPa. CO 2 was
added using a supercritical CO 2 pump which was a modified design of
the HPLC pump. Additional CO 2 had to be added over the first 12 hours
to compensate pressure drops due to the dissolution of CO 2 into the
aqueous solution.
After all components were added the autoclave was sealed for
the exposure period. The electrochemical measurements included
linear polarization resistance (LPR), electrochemical impedance
spectroscopy (EIS), cyclic voltammetry (CV), and linear sweep
voltammetry (LSV). LPR and EIS measurements were performed
regularly for at least 60 hours until the system approached a steady
state. LPR was measured at 0.1 mV s-1 over a range of ±15 mV around
the corrosion potential, E corr . EIS was performed with an amplitude of
10 mV over a frequency range from 200 kHz down to 5 mHz. After 60
hours, CV was run at sweep rates of 100, 50, and 10 mV s-1. LSV was
conducted at a scan rate of 1 mV s-1, sweeping up and down from E corr .
One of the working electrodes was used for surface analysis, so it was
not polarized with CV and LSV. Total exposure time ranged from 90 to
120 hours and was recorded for mass loss calculations. For data
analysis, the average and the standard deviation were obtained from
the measurements of the two or three working electrodes in each
condition at steady state.
After the electrochemical tests were completed the autoclave
systems were depressurized and disassembled. All electrodes and mass
loss samples were rinsed with distilled water, dried with argon, and
sealed in sample bags under argon. The working electrode used for
surface analysis was examined with scanning electron microscopy
(SEM) and energy dispersive X-ray spectroscopy (EDS). Mass loss
samples were cleaned and analyzed according to ASTM G1-03 using
Solution C.3.5 (HCl and hexamethylene tetramine) as described in
Appendix A1.37 Freshly polished, unexposed samples for each steel
were also cleaned to serve as calibrations and account for material loss
due to the cleaning procedure.
RESULTS AND DISCUSSION
Solution Speciation
The speciation in the three conditions were calculated using
commercial software.38 The calculated pH and the carbonate and sulfide
species in the aqueous phase are given in TABLE 3. The solution pH was
predicted to be around 3.0 for the two conditions containing CO 2 and
4.5 for the H 2 S only condition. From the CO 2 only condition to CO 2 +H 2 S
condition, pH and carbonate speciation were barely changed with the
addition of relatively small amount of H 2 S (total H 2 S:CO 2 =0.001). In
both conditions, CO 2 (aq) was the dominant active species and the
H 2 S(aq) concentration, if any, was three orders of magnitude lower. In
the H2S only condition, the sulfide species concentrations were
relatively higher than in CO 2 +H 2 S condition without the dissolution of
H 2 S in the CO 2 -rich phase. H 2 S(aq) was the dominant sulfide species in
the H 2 S-containing conditions.
Polarization Resistance and Corrosion Rate
The polarization resistances, R pol , of S-135 and UD-165 from LPR
and EIS in each condition at steady state are depicted in FIGURE 2.
Both steels showed an obvious correlation between R pol and the
presence of H 2 S. R pol increased by at least one order of magnitude in
the presence of H2S, indicating the inhibiting effect of small amount of
H 2 S on corrosion. This matched the behavior that has been reported
for low concentrations of H 2 S.18,19 For each steel, R pol in the H 2 S only
condition was about half of that in the CO 2 +H 2 S condition, which
suggested that a higher concentration of H 2 S(aq) could also accelerate
the corrosion considering that the H 2 S(aq) concentration was higher by
a factor of 4 in the H 2 S only condition as shown in TABLE 3.
The corrosion rates (CR) of S-135 and UD-165 were calculated
according to the Stern-Geary Equation using Rpol.2,39 The Tafel slopes
were calculated assuming the symmetry coefficient, β, was 0.5 and two
electrons were transferred in the Fe dissolution reaction. The CR from
LPR, EIS, and mass loss samples are given in TABLE 4. Inversely with
R pol , the corrosion rate dropped by at least one order of magnitude in
the presence of H 2 S. A small increase in corrosion rate was observed
when H 2 S(aq) concentration increased in the aqueous phase from
CO 2 +H 2 S condition to the H 2 S only condition; however, the CR
increase was negligible when compared with the decrease from the
CO 2 only condition.
It can be seen that a CO 2 :H 2 S ratio as high as 1000 was still
enough to transition from sweet to sour corrosion in the low
temperature and high pressure environment tested here. While a
decrease in material loss is generally welcome in corroding systems,
the tradeoff for the protection of iron sulfide can be an increased risk
for cracking. For this reason, laboratory corrosion measurements for
sour corrosion should be coupled with mechanical testing where
possible to determine if the decrease in corrosion rate due to the
transition from sweet to sour corrosion coincides with increased
susceptibility to sulfide stress cracking.
Very good agreement was seen between the LPR and EIS results in
all cases. Good agreement was seen between the corrosion rates from
the electrochemical and mass loss measurements in the CO 2 only
condition. However, the difference was greater when H 2 S was present,
where corrosion rates from mass loss samples were consistently larger
than those from the electrochemical tests. The difference was
attributed to relatively larger errors in the mass loss measurements
when the corrosion rates were lower. It was found that the material
lost from each exposure to the cleaning solution was of a similar order
to the mass lost from corrosion in H 2 S-containing conditions, which
likely resulted in an exaggerated corrosion rate.
In all three conditions, UD-165 gave a higher overall corrosion rate
than S-135 in the electrochemical tests, even after accounting for the
standard deviation between samples. The difference was roughly a
factor of two in the presence of H 2 S, but was as much as four times
larger in the CO 2 only system. The trend was seen in the mass loss data
for the CO 2 -containing systems but was not present in the H 2 S only
system. This may again have been due to measurement errors in the
mass loss samples.
Corrosion Modeling Results
The experimental results were compared with those calculated
for each condition using the commercial modeling software,40,41 and
the modeled CR results are given in TABLE 4. The CR calculations were
performed assuming a rotating disk of generic carbon steel (G10100) in
the test environments listed in TABLE 1 at 4 oC and 10 MPa with the
exception of 7 MPa in the H 2 S only system. Hydrodynamic effects were
approximated as a 10 mm OD rotating disk electrode at 150 RPM,
though hydrodynamic effects (rotating rate) were found to have only a
minor impact on the calculations at the test conditions.
The modeled CR in the CO2 only condition agreed with the
experimental CR for UD-165. For S-135, although it was four times
higher than the experimental CR, the modeled CR was still within a
reasonable range. Good agreement was seen in the H 2 S only condition
for both UD-165 and S-135 within a factor of two. Overall, the
modelling for these two conditions was considered to be quite good
when considering that a generic carbon steel was assumed in place of
the specific steels tested.
Substantial disagreement was seen in the CO 2 +H 2 S mixture. The
experimental CR results showed that the mixed case behaved very
close to the H 2 S only condition, indicating sour corrosion was
dominant. However, the modeled CR results were very close to the
prediction for CO 2 corrosion, resulting in overly high predicted
corrosion rates. Increasing the amount of H 2 S in the model, thereby
decreasing the CO 2 :H 2 S ratio, only increased the predicted corrosion
rate, showing that the issue was not a simple “tipping point” between
sweet and sour corrosion. Decreasing the amount of CO 2 in the model
(while keeping H 2 S constant) decreased the predicted corrosion rate
but did not reach the low rates in the H2S only case. The rotectiveness
of sulfide layers need to be better accounted for in the model.
TABLE 5 shows comparison between the experimentally
measured E corr and those calculated using the commercial modeling
software. It can be seen that, unlike the corrosion rate, the measured
E corr was strongly correlated with the presence of CO 2 rather than H 2 S.
This suggests that Ecorr was more a function of pH, which was
predicted to be 3.0 for the systems containing CO2 and 4.5 for the
system without CO 2 . This would affect the equilibrium potential for the
cathodic reactions (hydrogen evolution reactions), which could in turn
shift E corr to a more positive value when the pH is lower. It could also
affect the equilibrium potential for the anodic reaction by changing the
solubility of corrosion products in the aqueous phase, according to the
Nernst Equation. The modeled E corr values for all conditions were in
good agreement with the experimental data. Importantly, the model
correctly predicted the change in E corr when CO 2 was not present in
the system.
EIS Results
The representative Nyquist plots and Bode phase angle plots in
each condition after 60 hours are shown in FIGURE 3. Only one time
constant was observed in all three environments, and the Nyquist plots
were depressed semicircles. A dramatic increase in impedance was
observed in the presence of H2S, which was consistent with the LPR
results. Additionally, it can be seen that the transition to sour corrosion
doubled the maximum phase angle to close to 90 o, suggesting that the
double layer capacitances in H 2 S-containing conditions showed more
ideal capacitive behavior than those in the CO 2 only system. There was
no significant or consistent shift in the frequency at which the
maximum peak was observed, suggesting that the overall capacitance
of the corrosion product film did not change greatly with the change in
corrosion rate.
The EIS results confirmed the activation-controlled corrosion
behavior, which was the assumption used to calculate the corrosion
rates.36,39 The data was fitted using the Randles circuit without
diffusion shown in FIGURE 4. R sol represents the solution resistance,
and R pol represents the polarization resistance. A constant phase
element (CPE) was used to account for non-ideality of the double layer
capacitance, C dl . The R pol values from EIS fitting results are shown in
FIGURE 2 and were used to calculate the CR in TABLE 4.
Cyclic Voltammetry Results
 The representative CV curves for the two tested steels are shown
in FIGURE 5. All curves were corrected to take into account the
solution resistance determined from EIS. Three distinct trends were
observed. First and foremost, no obvious limiting currents were
observed at anodic regions and the sweep up and down plots were
almost overlaid, indicating that the anodic reactions were not limited
by mass transport. Secondly, a consistent peak was seen in the anodic
sweep for both steels in the H 2 S and CO 2 +H 2 S systems that was absent
in the CO 2 only system. It suggested that this peak thus corresponded
to a sulfide-related process. This peak was within about 100 mV from
the measured E corr in each case, indicating that the process was likely
to be present as part of the corrosion process. The third trend was that
the anodic slopes for the H 2 S and CO 2 +H 2 S curves were notably
steeper than for the CO 2 curve, suggesting that the presence of H 2 S
affected the corrosion mechanism in addition to reducing the available
reactive area. Trends for the cathodic side of the sweeps were
inconsistent and no similar conclusions were drawn. Sweeps at 100 mV
s-1 showed nearly identical behavior to those shown in FIGURE 5 at 50
mV s-1, only with larger current densities. In sweeps at 10 mV s-1 the
magnitude of the H 2 S-related peak was much smaller than at 50 mV s-1
and in some cases was difficult to distinguish. This correlation
indicated a capacitive feature of the process to which this H 2 S-related
peak corresponded.
Linear Sweep Voltammetry Results
The representative LSV plots for both steels in the three different
gas conditions are shown in FIGURE 6. All plots were corrected for R sol
determined from EIS. With the polarization of at least 200 mV at both
anodic and cathodic regions, no obvious limiting current densities were
observed in any curves, indicating that both steels were under active
corrosion in the acidic solutions containing CO 2 and/or H 2 S. This result
agreed with the reported effects of CO 2 and H 2 S in literature.15,21,42–44
In each condition, the average Tafel slope was obtained from the two
or three working electrodes, which are given in TABLE 6 along with the
standard deviation. For the CO 2 only case, two quasi-linear regions
were observed in the anodic region, the low polarization region and
the high polarization region. While the low polarization region could be
interfered by the cathodic reaction, the high polarization region could
be influenced by mass transport or a chemical step. Therefore, the
anodic Tafel slopes in both regions are listed. The CO 2 +H 2 S condition
and the H 2 S only condition had similar anodic (b a ) and cathodic (b c )
Tafel slopes, which were different from those in the CO 2 only
system.This indicated that the reaction mechanism was determined by
H 2 S even when its concentration was lower than CO2 by three orders
of magnitude. The cathodic Tafel slope for UD-165 in the CO 2 +H 2 S
condition was between the CO 2 only condition and H2S only condition;
however, it still confirmed that the presence of H 2 S decreased the
cathodic Tafel slope. In FIGURE 6, for UD-165 in CO2+H2S condition,
the slope at the further polarized anodic region was close to the anodic
Tafel slope in the CO 2 only condition, which indicated that CO 2
corrosion could have taken over when H 2 S was consumed at the
surface. This region was not taken for the anodic slope calculation.
The anodic reaction during corrosion was the dissolution of Fe.
For H 2 S corrosion, Shoesmith et al.45 proposed a chemisorption-
oxidation mechanism between Fe and H 2 S to form mackinawite in
saturated H 2 S solutions at 21 oC, and Ma et al.46 proposed a similar
mechanism with two consecutive electron transfers. These two
mechanisms are both analogous to a Fe dissolution reaction
mechanism proposed by Bockris et al.47 involving an intermediate with
OH-(aq). Iofa et al.48 and Zhang et al.8 described a mechanism involving
a chemisorbed hydrogen sulfide layer and two electrons transferred
simultaneously. For CO 2 corrosion, the anodic mechanism has been
frequently assumed to follow the Bockris mechanism, which could well
explain some experimental Tafel slopes from 0.03 to 0.04 V dec.-
1.9,44,49,50 However, Nesic et al.51 reported that b values transitioned
a
from 0.03 to 0.12 V dec-1 when pH increased from 4 to 5 at 22 oC,
indicating a mechanism transition. Moreover, Videm and Koren stated
that the anodic Tafel slope was 0.03 V dec-1 in the low potential region
and 0.12 V dec-1 in the higher potentials at 25 oC.14 Assuming the
symmetry coefficient to be 0.5, the b a value of 0.12 V dec-1 could be
correlated with a mechanism involving one electron transferred in the
rate determining step (RDS). It has been proposed by the authors
previously that at 85 oC the HS-(aq) reaction involved two electrons
(2e) transferring simultaneously as the RDS, and the HCO 3 -(aq)
reaction followed the one-electron mechanism (1e), and the OH-(aq)
reaction followed the Bockris mechanism.2 To check if these three
anodic mechanisms were also applicable in the conditions tested here
at 4 oC, the symmetry coefficients (β) were calculated from the
measured Tafel slopes following Equations (4), (8) and (9), and the
results are shown in TABLE 7.
Due to its physical meaning, β should be between 0 and 1, and β is
usually assumed as 0.5 for kinetics calculations. Thus, the Bockris
mechanism was not possible because its corresponding β values were
mostly negative. The corresponding β values were between 0 and 1 for
the two-electron mechanism and the one-electron mechanism, so both
mechanisms were possible in the three gas conditions at 4 oC. β is
usually considered as 0.5, and the mechanisms were postulated from
this perspective. The corresponding β values for the two-electron
mechanism were consistently around 0.5 in the CO 2 +H 2 S condition
and the H2S only condition. Hence, the presence of H 2 S could have
resulted in the two-electron mechanism. According to TABLE 3,
H 2 S(aq) was the most dominant reactive species in the H 2 S only case.
Therefore, H 2 S(aq) probably followed the two-electron mechanism,
which are written in Reactions (1) to (3). This mechanism was also
proposed and used by Iofa et al.48 and Zhang et al.8 The specific
adsorption of H 2 S changed the surface coverage in the CO 2 +H 2 S case
and shifted the mechanism from that in the CO2 only case.48 The two-
electron mechanism of H 2 S(aq) agreed with the anodic Tafel slopes
reported in literature for H 2 S-containing solutions, 0.04 to 0.05 V dec-1
at 20 to 30 oC, where β was around 0.6.43,52,53
The mechanism in the CO 2 only case was not as conclusive.
However, considering the β values for both S-135 and UD-165 in TABLE
7, it was suggested that CO 2 (aq) could have followed the two-electron
mechanism in the low polarization region and the one-electron
mechanism in the high polarization region. The one-electron
mechanism is written in Reactions (5) to (7). Although the proposed
CO 2 mechanisms did not include the Bockris mechanism which is
frequently assumed at room temperature, the one-electron
mechanism was in good agreement with the Tafel slopes reported at
pH>5 by Nesic et al.51 and in high polarization regions by Videm and
Koren.14 The mechanisms for H 2 S(aq) and CO 2 (aq) at 4 oC were
generally consistent with the proposed mechanisms at 85 oC by the
authors.2
The two-electron mechanism for H 2 S(aq):
Fe(s) + H 2 S(aq) + H 2 O(l) ↔ FeHS-(ad) + H 3 O+(aq)
(1)
FeHS-(ad) → FeHS+(ad) + 2e- (RDS) (2)
FeHS+(ad) + H 3 O+(aq) ↔ Fe2+(aq) + H 2 S(aq) + H 2 O(l)
(3)
 2.303𝑅𝑅𝑅𝑅
𝑏𝑏𝑎𝑎 = (4)
2(1−𝛽𝛽)𝐹𝐹
where R is the gas constant, T is the absolute temperature, and F is the
Faraday constant.
The one-electron mechanism for CO 2 (aq):
Fe(s) + CO 2 (aq) + 2H 2 O(l) → FeHCO 3 (ad) + H 3 O+(aq) + e-
(RDS) (5)
FeHCO 3 (ad) ↔ FeHCO 3 +(ad) + e-
(6)
FeHCO 3 +(ad) + H 3 O+(aq) ↔ Fe2+(aq) + CO 2 (aq) + 2H 2 O(l)
(7)
2.303𝑅𝑅𝑅𝑅
𝑏𝑏𝑎𝑎 = (8)
(1−𝛽𝛽)𝐹𝐹
Anodic Tafel slope in the Bockris mechanism:46
2.303𝑅𝑅𝑅𝑅
𝑏𝑏𝑎𝑎 = (9)
(1+𝛽𝛽)𝐹𝐹
The cathodic reaction during corrosion in a deaerated solution is
the hydrogen evolution reaction (HER). In an acidic solution where
H+(aq) is adequate, the reduction reaction of H+(aq) is dominant as the
cathodic reaction. The Volmer-Heyrovsky (V-H) mechanism and the
Volmer-Tafel (V-T) mechanism are the most commonly accepted
hydrogen evolution reaction mechanisms, and their Tafel slopes can be
calculated following Equations (10) and (11), respectively.2,54 The
dissolved CO 2 and H 2 S could form weak acids in the solution and
participate in the HER in two main ways. One way is to provide H+(aq)
through dissociation. The other way is the direct reduction of
H 2 CO 3 (aq) and H 2 S(aq).45,55 The Tafel slopes for HER with H+(aq),
H 2 CO 3 (aq), and H 2 S(aq) were frequently reported to be around -0.120
V dec-1 in literature at 20 to 30 oC, which is often used in modeling
calculation,43,44,48,52,53 and this value agreed with the V-H mechanism.
When the availability of the above hydrogen sources is limited, H 2 O(l),
in an abundant amount, could also be reduced to H 2 (g). The addition
of CO 2 (aq) and H 2 S(aq) could increase the cathodic current density,
and could also lead to a limiting current at high polarization due to
mass transport or a slow chemical step such as the hydration of CO 2
and the chemisorption of H 2 S.9,44,48 At even higher cathodic
polarization, the limiting current may transition to a Tafel slope again
showing the reduction of H 2 O(l), similarly to the results in Stern’s study
on the effect of pH on cathodic polarization.56
The corresponding β values for V-H and V-T mechanisms were
calculated from the measured cathodic Tafel slopes, and the results
are listed in TABLE 7. The cathodic Tafel slope of V-T was not affected
by β according to Equation (11), so no β value could be calculated for
V-T. Assuming β to be 0.5, the b c value for V-T was calculated to be
0.0275 V dec-1, which had a much smaller magnitude than the
measured Tafel slopes; therefore, the V-T mechanism was unlikely for
the tested conditions. A two-electron cathodic mechanism was also
included in TABLE 7 for comparison, and its Tafel slope was calculated
following Equation (12). Although the corresponding β values for the V-
H mechanism and the two-electron mechanism were within 0 to 1 as
shown in TABLE 7, they were very close to 0, especially for the two-
electron mechanism. Thus, the cathodic mechanism was more likely to
be the V-H mechanism than the V-T mechanism and the two-electron
mechanism, although the very low β values suggested some other
mechanisms may better explain the behavior. Considering the cathodic
Tafel slopes were taken in the linear regions 100 to 200 mV from E corr ,
the large Tafel slopes may have been complicated with mass transport
or a slow chemical step. The chemisorption of H 2 S could have caused
the smaller magnitudes of b c values in the H 2 S-containing conditions
than the CO2 only condition.
Cathodic Tafel slope in the V-H mechanism:54
2.303𝑅𝑅𝑅𝑅
𝑏𝑏𝑐𝑐 = − (10)
𝛽𝛽𝛽𝛽
Cathodic Tafel slope in the V-T mechanism:54
2.303𝑅𝑅𝑅𝑅
𝑏𝑏𝑐𝑐 = − (11)
2𝐹𝐹
Cathodic Tafel slope in the two-electron mechanism:
2.303𝑅𝑅𝑅𝑅
𝑏𝑏𝑐𝑐 = − (12)
2𝛽𝛽𝛽𝛽
Pourbaix Diagrams and Surface Analysis
Pourbaix diagrams were generated using the commercial
modeling software to predict the thermodynamically predominant
corrosion products in each condition, which are depicted in FIGUREs 7
to 9, respectively. At pH 3.0 for the CO 2 only and CO 2 +H 2 S conditions,
soluble Fe2+(aq) was predicted to be the thermodynamically favorable
product, which supported the assumption of active corrosion. At pH
4.5 for the H 2 S only condition, Fe2+(aq) and FeHS+(aq) were predicted
to be stable. The addition of H 2 S was predicted to favor the formation
of FeS(s) at relatively higher pH in FIGUREs 8 and 9. Considering the
inhibiting effect of H 2 S on corrosion rate, the protective FeS(s) might
have formed at the steel surface where the pH and Fe2+(aq)
concentration might be higher than in the bulk solution.21
In all three conditions, the corrosion scales on the sample surfaces
changed from black to orange quickly during rinsing with distilled
water and drying with argon after taking the samples out of the
autoclave. A large amount of oxygen was detected using SEM-EDS on
the surfaces for both S-135 and UD-165. More details about surface
analysis can be found in Reference 57. The discoloration and the
significant portion of oxygen on the sample surfaces may have been
due to oxidation of the existing corrosion products to hematite after
the system was disassembled. This showed that the corrosion product
stability was quite conditional, and the increase in temperature and
oxygen content could result in a rapid breakdown of the corrosion film.
CONCLUSIONS
Experimental results for two drill steels, S-135 and UD-165,
showed a distinct drop in corrosion rate when transitioning from sweet
to sour corrosion in 5%wt. NaCl solutions at 4 °C and 10 MPa total
pressure. It was found that a CO 2 :H 2 S ratio as high as 1000 was still
enough for sour corrosion to dominate in the conditions tested. The
H 2 S only and CO 2 +H 2 S conditions gave corrosion rates on the order of
0.05 mm y-1, whereas the corrosion rates in the CO 2 only condition
were on the order of 0.5 to 2 mm y-1.

The reaction mechanism was determined by the presence of H 2 S
even when its concentration was lower than CO 2 by three orders of
magnitude due to specific adsorption of H 2 S. The speciation modeling
indicated that CO 2 (aq) or H 2 S(aq) was the most dominant reactive
species in the tested conditions. For the anodic reaction / Fe
dissolution reaction, H 2 S(aq) probably followed the two-electron
mechanism. CO 2 (aq) could have followed the two-electron mechanism
in the low polarization region and the one-electron mechanism in the
high polarization region. The cathodic Tafel slopes at 4 oC were much
larger in magnitude than the reported values at 20 to 30 oC in
literature, and the cathodic mechanisms need further study at 4 oC and
high pressure conditions.
Comparisons were made between the experimental CR results
and those from a commercial software model. Good agreement was
seen in the CO 2 only and H 2 S only conditions. However, the model
predicted that corrosion in the mixed CO 2 +H 2 S system would match
that in the CO 2 only condition, whereas experimental CR results found
that it matched that in the H 2 S only condition. In all cases the model
was found to provide a good prediction for E corr . Corrosion potentials
in the CO 2 only and CO 2 +H 2 S systems were almost identical, while a
roughly 100 mV decrease was seen in the H 2 S only system. The
experimental results suggested that while the corrosion behavior was
strongly dependent on the H 2 S content, the corrosion potential was
instead related to the CO 2 content and pH.
In the Pourbaix diagrams, soluble species such as Fe2+(aq) and
FeHS+(aq) were predicted to be the thermodynamically predominant
products. The presence of H 2 S would favor the formation of FeS(s) at
the steel surface where the pH and Fe2+(aq) concentration may be
higher than in the bulk solution. The formation of a thin sulfide film
could be protective and decrease the CR. The corrosion products for all
tests were found to be quickly oxidized on exposure to air when the
system was disassembled.
DISCLAIMER
This report was prepared as an account of work sponsored by an
agency of the United States Government. Neither the United States
Government nor any agency thereof, nor any of their employees,
makes any warranty, express or implied, or assumes any legal liability
or responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents
that its use would not infringe privately owned rights. Reference herein
to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute
or imply its endorsement, recommendation, or favoring by the United
States Government or any agency thereof. The views and opinions of
authors expressed herein do not necessarily state or reflect those of
the United States Government or any agency thereof.
ACKNOWLEDGEMENTS
This work was completed as part of National Energy Technology
Laboratory (NETL) research for the Department of Energy's
Complementary Research Program under Section 999 of the Energy
Policy Act of 2005. The authors would like to acknowledge OLI systems
Inc. for providing access to OLI Analyzer modeling software and,
especially, Dr. Andre Anderko for his helpful suggestions on modeling.
The authors wish to acknowledge Roy Long (NETL Strategic Center for
Natural Gas and Oil), Dr. Kelly Rose and Dr. Jeffrey Hawk (NETL
Research and Innovation Center) for programmatic direction and
technical support.
REFERENCES
1. B.J. Berkowitz, and H.H. Horowitz, J. Electrochem. Soc.
129(1982): p. 468–474.
2. R. Feng, J. Beck, M. Ziomek-Moroz, and S.N. Lvov,
Electrochim.Acta 212 (2016): p. 998–1009.
3. E.C. Greco, and W.B. Wright, Corrosion 18 (1962): p. 119t–
124t.
4. Z.Y. Liu, X.Z. Wang, R.K. Liu, C.W. Du, and X.G. Li, J. Mater. Eng.
Perform. 23 (2014): p. 1279–1287.
5. F. Pessu, R. Barker, and A. Neville, “Early Stages of Pitting
Corrosion of UNS K03014 Carbon Steel in Sour Corrosion
Environments: The Influence of CO 2 , H 2 S and Temperature,”
CORROSION/2015, paper No. 5583 (Houston, TX: NACE, 2015).
6. G. Fierro, G.M. Ingo, and F. Mancia, Corrosion 45 (1989): p.
814–823.
7. S.N. Smith, and M.W. Joosten, “Corrosion of Carbon Steel By
H2S in CO 2 Containing Oilfield Environments,”
CORROSION/2006, paper No. 06115 (Houston, TX: NACE,
2006)..
8. G.A. Zhang, Y. Zeng, X.P. Guo, F. Jiang, D.Y. Shi, and Z.Y. Chen,
Corros. Sci. 65 (2012): p. 37–47.
9. . Y. Zheng, J. Ning, B. Brown, and S. Nesic, Corrosion 71
(2015): p. 316–325.
10. M. Ziomek-Moroz, J. Mater. Eng. Perform. 21 (2011): pp.
1061–1069.
11. W. Sun, S. Nešić, and S. Papavinasam, “Kinetics of Iron Sulfide
and Mixed Iron Sulfide/Carbonate Scale Precipitation in
CO 2 /H 2 S,” CRROSION/2006, paper No. 06644 (Houston, TX:
NACE, 2006).
12. J. Han, J. Zhang, and J.W. Carey, Int. J. Greenh. Gas Control 5
(2011): pp. 1680–1683.
13. A. Dugstad, “Fundamental Aspects of CO2 Metal Loss
Corrosion, Part I: Mechanism,” CORROSION/2015, paper No.
5826 (Houston, TX: NACE, 2015)..
14. K. Videm, and A.M. Koren, Corrosion 49 (1993): pp. 746–754.
15. F. Pessu, R. Barker, and A. Neville, Corrosion 71 (2015): pp.
1452–1466.
16. Wang, H., and S. Nešić, “CO 2 Corrosion Model Verification
Using Field Data,” CORROSION/2006, paper No. 06567
(Houston, TX: NACE, 2006).
17. E. Abelev, T.A. Ramanarayanan, and S.L. Bernasek, J.
Electrochem. Soc. 156 (2009): pp. C331–C339.
18. K.J Lee and S. Nesic, The Effect of Trace Amount of H 2 S on
CO 2 Corrosion Investigated by Using the EIS Technique,”
CORROSION/2005, paper No. 05630 (Houston, TX: NACE,
2005).
19. K. Videm and J. Kvarekvål, Corrosion 51 (1995): pp. 260–269.
20. E. Abelev, J. Sellberg, T.A. Ramanarayanan, and S.L. Bernasek,
J. Mater. Sci. 44 (2009): pp. 6167–6181.
21. Y.-S Choi, S. Nesic, and S. Ling, Electrochim. Acta 56 (2011): pp.
1752–1760.
22. J. Kvarekval, R. Nyborg, and M. Seiersten, “Corrosion Product
Films on Carbon Steel in Semi-Sour CO 2 /H 2 S Environments,”
CORROSION/2002, paper No. 02296 (Houston, TX: NACE,
2002).
23. M. Singer, B. Brown, A. Camacho, and S. Nesic, . “Combined
Effect of CO 2 , H 2 S and Acetic Acid on Bottom of the Line
Corrosion,” CORROSION/2007, paper No. 07661 (Houston, TX:
NACE, 2007).
24. NACE Standard SP0110-2010, “Wet Gas Internal Corrosion 51. S. Nesic, N. Thevenot, J.-L. Crolet, and D.M. Drazic,
Direct Assessment Methodology for Pipelines” (Houston, TX: “Electrochemical Properties of Iron Dissolution in the
NACE, 2010). Presence of CO2-Basics Revisted,” CORROSION/1996, paper
25. S.N.Smith, “Discussion of the History and Relevance of the No. 3 (Houston, TX: NACE, 1996).
CO2/H2S Ratio,” CORROSION/2011, paper No.11065 52. D.R. Morris, L.P. Sampaleanu, and D.N. Veysey, J. Electrochem.
(Houston, TX: NACE, 2011). Soc. 127 (1980): p. 1228–1235.
26. C. Ren, D. Liu, Z. Bai, and T. Li, Mater. Chem. Phys. 93 (2005): 53. X.L. Cheng, H.Y. Ma, J.P. Zhang, X. Chen, S.H. Chen, and H.Q.
pp. 305–309. Yang, Corrosion 54 (1998): p. 369–376.
27. J.Kvarekval, R. Nyborg, and H. Choi, “Formation of Multilayer 54. D. Landolt, Corrosion and Surface Chemistry of Metals, 1st ed.
Iron Sulfide Films During High Temperature CO2/H2S (EPFL Press, 2007).
Corrosion of Carbon Steel,” CORROSION/2003, paper No. 55. G.I. Ogundele and W.E. White, Corrosion 42 (1986): pp. 71–78.
03339 (Houston, TX: NACE, 2003). 56. M. Stern, J. Electrochem. Soc. 102 (1955): pp. 609–612.
28. ANSI/NACE MR0175/ISO 15156-1, “Petroleum and Natural 57. . R. Feng, J. Beck, D. Hall, A. Buyuksagis, S.N. Lvov, and M.
Gas Industries — Materials for Use in H2S-Containing Ziomek-Moroz, “Effects of CO2 and H2S on Corrosion of
Environments in Oil and Gas Production - Part 1: General Martensitic Steels in NaCl at Low Temperature,”
Principles for Selection of Cracking-Resistant Materials” CORROSION/2016, paper No. 7659, (Houston, TX: NACE,
(2009). 2016).
29. A. Bishop, C. Bremner, A. Laake, C. Strobbia, P. Parno, and G.
Utskot, Oilf. Rev. 22 (2011): pp. 36–49.
30. L.D. Perrigo, “An Overview of Corrosion in Cold Climates,”
CORROSION/2001, paper No. 01306 (Houston, TX: NACE,
2001).
31. P. Budzik, “Arctic Oil and Natural Gas Potential” (U.S. Energy
Information Administration, Office of Integrated Analysis and
Forecasting, Oil and Gas Division, 2009).
32. D.R. Lide, “CRC Handbook of Chemistry and Physics,” 84th ed.
(CRC Press, 2003), p. 6–53.
33. K.A. Rozman, “Corrosion and Fatigue of Oil Well Drilling
Steels,” Oregon State University, 2011.
34. R.A. Robie, B.S. Hemingway, and J.R. Fisher, U.S. Geol. Surv.
Bull. 1452 (1978).
35. Zaytsev, I.D., and G.G. Aseyev, Properties of Aqueous Solutions
of Electrolytes (CRC Press, 1992).
36. S.N. Lvov, Introduction to Electrochemical Science and
Engineering, 1st ed. (CRC Press, 2015).
37. ASTM Standard G1-03, “Standard Practice for Preparing
Cleaning and Evaluating Corrosion Test Specimens”, Annual
Book of ASTM Standards, vol. 03. 02 (West Conshohocken, PA:
ASTM International, 2003).
38. P. Wang, A. Anderko, and R.D. Young, Fluid Phase Equilib. 203
(2002): p. 141–176.
39. M. Stern, Corrosion 14 (1958): p. 60–64.
40. A. Anderko, P. McKenzie, and R.D. Young, Corrosion 57 (2001):
p. 202–211.
41. A. Anderko, and R.D. Young, “Simulation of CO 2 /H 2 S
Corrosion Using Thermodynamic and Electrochemical
Models,” CORROSION/1999, paper No. 99031 (Houston, TX:
NACE, 1999).
42. J. Tang, Y. Shao, J. Guo, T. Zhang, G. Meng, and F. Wang,
Corros. Sci. 52 (2010): p. 2050–2058.
43. Y. Zheng, B. Brown, and S. Nešic, Corrosion 70 (2014): p. 351–
365.
44. S. Nesic, J. Postlethwaite, and S. Olsen, Corrosion 52 (1996): p.
280–294..
45. D.W. Shoesmith, P. Taylor, M.G. Bailey, and D.G. Owen, J.
Electrochem. Soc. 127 (1980): p. 1007–1015..
46. H. Ma, X. Cheng, G. Li, S. Chen, Z. Quan, S. Zhao, and L. Niu,
Corros. Sci. 42 (2000): p. 1669–1683.
47. J.O. Bockris, D. Drazic, and A.R. Despic, Electrochim. Acta 4
(1961): p. 325–361.
48. Z.A. Iofa, V. V. Batrakov, and Cho-Ngok-Ba, Electrochim. Acta 9
(1964): p. 1645–1653.
49. M. Nordsveen, S. Nesic, R. Nyborg, and A. Stangeland,
Corrosion 59 (2003): p. 443–456.
50 G.A. Zhang, and Y.F. Cheng, Corros. Sci. 51 (2009): p. 87–94
 1

14 2
3
4
15
5 Non-
16 aqueous
6

7
8
9
10 Aqueous
11 12
13

FIGURE 1. Schematic diagram of the autoclave system. 1- Connection to pressure transducer; 2- Copper coil cooling loop; 3-
Autoclave stainless steel vessel; 4- PTFE liner; 5- Polymer-coated thermocouple; 6- Dip tubing; 7- Epoxy-coated stirrer; 8- PTFE-
coated working electrode lead; 9- Two Ag/AgCl reference electrodes; 10- Two mass loss samples; 11- PTFE assembly; 12- Two or
three steel working electrodes; 13- Two or three Pt counter electrodes; 14- Supercritical CO 2 pump; 15- H 2 S/Ar cylinder; 16-
CO 2 cylinder.

S-135
12000

10000

8000
Rpol / Ω cm2

6000

4000

2000

0
CO2 CO2+H2S H2S
 UD-165
5000

4000

Rpol / Ω cm2
3000

2000

1000

0
CO2 CO2+H2S H2S

LPR EIS

FIGURE 2. R pol from LPR and EIS for S-135 (top) and UD-165 (bottom) after 60 hours in the CO 2 only, CO 2 +H 2 S, and H 2 S only
cases. Conditions: 4 oC, 10 MPa, 0.9 mol kg-1 NaCl solutions, 150 RPM.

S-135 UD-165
8000 300
4000
150
200 100
6000 100
3000
50
-Zimag / Ω cm2

-Zimag / Ω cm2

0 0
4000 0 100 200 300 2000 0 50 100 150

2000 1000

0 0
0 2000 4000 6000 8000 0 1000 2000 3000 4000
Zreal / Ω cm2 Zreal / Ω cm2
10 10

-10 -10
Phase Angle / °
Phase Angle / °

-30 -30

-50 -50

-70 -70

-90 -90
0.001 0.1 10 1000 100000 0.001 0.1 10 1000 100000
Frequency / Hz Frequency / Hz
FIGURE 3. Nyquist plots and Bode phase angle plots for S-135 (left) and UD-165 (right) after 60 hours. Conditions: 4 oC, 10 MPa,
0.9 mol kg-1 NaCl solutions, 150 RPM. Symbols: x = CO 2 , o = CO 2 +H 2 S, + = H 2 S.
 Rsol Cdl

Rpol

Figure 4. Equivalent circuit used to fit EIS data.

S-135
0.04

Current density / A cm-2

0.03

0.02

0.01

-0.01

-0.02
-1 -0.75 -0.5 -0.25 0 0.25
Potential / V vs SHE

UD-165
0.04
Current density / A cm-2

0.03

0.02

0.01

-0.01

-0.02
-1 -0.75 -0.5 -0.25 0 0.25
Potential / V vs SHE

FIGURE 5. Cyclic voltammetry curves for S-135 (top) and UD-165 (bottom) at 50 mV s-1. Conditions: 4 oC, 10 MPa, 0.9 mol kg-1
NaCl solutions, 150 RPM. black = CO 2 , grey = CO 2 +H 2 S, dashed = H 2 S.
 S-135
0.1

Current density / A cm-2

0.01

0.001

0.0001

0.00001
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1
Potential / V vs SHE
UD-165
0.1
Current density / A cm-2

0.01

0.001

0.0001

0.00001
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1
Potential / V, vs SHE
FIGURE 6. LSV plots for S-135 (top) and UD-165 (bottom) at a scan rate of 1 mV s-1. Conditions: 4 oC, 10 MPa, 0.9 mol kg-1 NaCl
solutions, 150 RPM. black = CO 2 , grey = CO 2 +H 2 S, dashed = H 2 S.

FIGURE 7. Pourbaix diagram of Fe-H 2 O-CO 2 system. Conditions: 4 oC, 10 MPa, 0.9 mol kg-1 NaCl, 14.49 mol H 2 O, 4.52 mol CO 2 .
FIGURE 8. Pourbaix diagram of Fe-H 2 O-CO 2 -H 2 S system. Conditions: 4 oC, 10 MPa, 0.9 mol kg-1 NaCl, 14.40 mol H 2 O, 4.49 mol
CO 2 , 4.49×10-3 mol H 2 S, 0.0404 mol Ar.

FIGURE 9. Pourbaix diagram of Fe-H 2 O-H 2 S system. Conditions: 4 oC, 7 MPa, 0.9 mol kg-1 NaCl, 20.20 mol H 2 O, 6.30×10-3 mol
H 2 S, 0.5597 mol Ar.
 TABLE 1
Test conditions for corrosion tests (mol) at 4 oC
H2S CO 2 H2O NaCl Ar
CO 2 0 4.52 14.49 0.235 0
CO 2 +H 2 S 4.49×10-3 4.49 14.40 0.233 0.0404
H2S 6.30×10-3 0 20.20 0.327 0.5597

TABLE 2
Chemical compositions (%mass) of S-135 and UD-165
Steel Fe Cr Mn Mo Al Ni Cu Nb S
S-135 Bal 1.39 0.78 0.68 0.03 0.02 0.02 <0.01 0.002
UD-165 Bal 0.79 0.88 0.67 0.03 0.81 0.19 0.02 0.007
Steel P B C N V Si Sn Ti
S-135 0.006 <0.005 0.26 0.005 <0.01 0.3 <0.01 <0.01
UD-165 <0.005 <0.005 0.27 0.007 0.07 0.26 <0.01 <0.01

TABLE 3
The calculated pH and carbonate and sulfide species in the aqueous phase (mol/kg H 2 O)
CO 2 CO 2 +H 2 S H2S
pH 3.0 3.0 4.5
CO 2 (aq) 1.43 1.44 0
HCO 3 -(aq) 9.29×10-4 9.28×10-4 0
CO 3 2-(aq) 1.48×10-10 1.47×10-10 0
H 2 S(aq) 0 3.29×10-3 1.41×10-2
HS-(aq) 0 2.12×10-7 3.39×10-5
S2-(aq) 0 1.68×10-19 7.75×10-16

TABLE 4
Corrosion rate (mm y ) results for S-135 and UD-165
-1

Steel Gas LPR EIS Mass Loss Modeled

CO 2 0.418±0.074 0.429±0.061 0.590±0.019 2.19
S-135 CO 2 +H 2 S 0.0189±0.0070 0.0187±0.0080 0.0599±0.0147 2.36
H2S 0.0346±0.0042 0.0350±0.0050 0.0798±0.0143 0.0419
CO 2 1.82±0.16 1.49±0.27 1.77±0.32 2.19
UD-165 CO 2 +H 2 S 0.0334±0.0042 0.0328±0.0054 0.0877±0.0043 2.36
H2S 0.0510±0.0036 0.0515±0.0041 0.0678±0.0240 0.0419

TABLE 5
Experimental and modeled corrosion potentials (V vs SHE)
Gas S-135 UD-165 Modeled
CO 2 -0.353±0.001 -0.332±0.001 -0.374
CO 2 +H 2 S -0.337±0.004 -0.337±0.004 -0.376
H2S -0.444±0.002 -0.460±0.002 -0.463
 TABLE 6
The Tafel slopes (V dec ) at the three different gas conditions
-1

S-135 UD-165
Gas ba bc ba bc
(I) 0.054±0.007 (I) 0.099±0.022
CO 2 -0.755±0.000 -0.608±0.077
(II) 0.178±0.024 (II) 0.251±0.057
CO 2 +H 2 S 0.073±0.001 -0.395±0.064 0.063 -0.481
H2S 0.076±0.021 -0.312±0.098 0.092±0.001 -0.251±0.041
Note: In the CO 2 only case, (I) represents the low polarization region and (II) represents the high polarization region.

TABLE 7
Symmetry coefficients β calculated from the measured Tafel slopes
S-135 UD-165
Anodic reaction 2e 1e Bockris 2e 1e Bockris
(I) 0.49 -0.02 0.02 0.72 0.45 -0.45
CO 2
(II) 0.85 0.69 -0.69 0.89 0.78 -0.78
CO 2 +H 2 S 0.63 0.25 -0.25 0.57 0.13 -0.13
H2S 0.64 0.28 -0.28 0.70 0.40 -0.40
Cathodic reaction V-H V-T 2e V-H V-T 2e
CO 2 0.07 - 0.04 0.09 - 0.05
CO 2 +H 2 S 0.14 - 0.07 0.11 - 0.06
H2S 0.18 - 0.09 0.22 - 0.11
Note: In the CO 2 only case, (I) represents the low polarization region and (II) represents the high polarization region.