CORROSION SCIENCE SECTION

Combined Effect of Carbon Dioxide, Hydrogen

Sulfide, and Acetic Acid on Bottom-of-the-Line
Corrosion

M. Singer,‡,* B. Brown,* A. Camacho,** and S. Nes̆ić*

aBstract introduction
This research work presents a study of the combined influence Carbon dioxide (CO2) corrosion in the presence of
of the partial pressure of hydrogen sulfide (H2S) and the con- acetic acid (CH3COOH) has been studied extensively
centration of free acetic acid (CH3COOH) on the general and by many different investigators.1-6 Consequently, the
localized carbon dioxide (CO2) corrosion at the bottom of the
corrosion mechanisms for this system are now well
line. Experiments were carried out for up to 21 days in three
defined and have been incorporated in prediction
4-in (101.6-mm) internal diameter flow loops at 70ºC with
models.7-8 The influence of acetic acid on the siderite
2 bars of CO2. The flow regime was stratified for all of the
experiments. It was found that trace amounts of H2S (from (FeCO3) film characteristics and formation is one of
0.004 bar to 0.13 bar) greatly retards the CO2 corrosion with the last areas where a widely accepted theory is not
general corrosion rates usually 10 to 100 times lower than agreed upon completely.9-12 The objective of this paper
their pure CO2 equivalent. However, the most protective con- was to try to improve our understanding of the influ-
ditions were observed at the lowest partial pressure of H2S, ence of the partial pressure of hydrogen sulfide (H2S)
while corrosion increased when more H2S was added. The and the presence of acetic acid on the CO2-influenced
presence of a mackinawite film on the coupon surface seems corrosion rate of carbon steel and the characteristics
to be the origin of this protectiveness. When 1,000 ppm of free of the corrosion product film formed.
acetic acid was added to the system, the general corrosion
The different chemical and electrochemical reac-
rate doubled compared with a baseline CO2 environment and
tions involved in CO2 corrosion in the presence of
by between one and two orders of magnitude compared to
baseline H2S/CO2 mixtures.
acetic acid are described below:

KEY WORDS: acetic acid, bottom-of-the-line corrosion, CO2/ Water dissociation:

H2S corrosion, mackinawite
H2O( l ) ↔ H(+aaqq ) + OH(–aq
OH aq )
(1)

Carbon dioxide dissolution:

CO2( g ) ↔ CO2( aaqq ) (2)
Submitted for publication July 14, 2009; in revised form, July 7,
2010.
‡
Corresponding author. E-mail: singer@[Link]. Carbon dioxide hydration (slowest step):
* Ohio University, Institute for Corrosion and Multiphase Technol- CO2( aq ) + H2O( l ) ↔ H2CO
CO3( aaqq ) (3)
ogy, 342 West State St., Athens, OH 45701.
** Ohio University, Institute for Corrosion and Multiphase Tech-
nology, 342 West State St., Athens, OH 45701. Present address: Carbonic acid dissolution:
Lloyd’s Register Capstone, Inc., 1505 Hwy. 6 South, Suite 250,
Houston, TX 77077. CO3( aq ) ↔ H(+aq ) + HCO3– ( aq )
H2CO (4)

ISSN 0010-9312 (print), 1938-159X (online)

CORROSION—Vol. 67, No. 1 11/000007/$5.00+$0.50/0 © 2011, NACE International 015004-1
CORROSION SCIENCE SECTION

Bicarbonate ion dissociation: Hydrogen sulfide dissociation:

O3( aq ) ↔
HCO H(+aq ) + CO23(–aq ) (5) H2S( aq ) ←
1 K
→ H(+aq ) + H
HS(–aq ) (14)

Acetic acid (HAc) dissociation: Bisulfide dissociation:

COOH( aq ) ↔ H(+aq ) + C
CH3COOH CH COO(–aq )
H 3C (6) HS(–aq ) ←
2 K
→ H(+aq ) + S(2aq– ) (15)

Proton reduction: Hydrogen sulfide reduction:

2H(+aq ) + 2e – → H2( g ) (7) 2e – → H2( g ) + 2HS
2H2S( aq ) + 2e 2HS(–aaqq ) (16)

Carbonic acid reduction: Bisulfide reduction:

2H2CO3( aq ) + 2e – → 2HCO
O3– ( aq ) + H2( g ) (8) 2HS(–aq ) + 2e
2e – → H2( g ) + 2S
2S(–aaqq ) (17)

Undissociated acetic acid reduction: Iron oxidation:

2e – → 2CH3COO(–aq ) + H2( g )
2CH3COOH( aq ) + 2e (9) +
Fe( s ) → Fe(2aaq –
q ) + 2e (18)

Iron oxidation:
Iron sulfide formation:
+
Fe( s ) → Fe(2aaq
q ) + 2e
–
(10) by precipitation:
+ K
Fe(2+ –
aq ) + HS( aq ) ←
←

→
sp,mck
mck
 FeS( s ) + H(+aaqq )
→ FeS (19)
Iron carbonate precipitation:
K sp,FeCO
Fe( aq ) + CO23(–aq ) ←
←

3
→ FeCO3( s ) (11) by direct (solid-state) reaction:
Fe( s ) + H2S → FeS
FeS( s ) + H2( g ) (20)
The choice of the expression of the equilibrium
constant governing the iron carbonate precipita- The chemical reactions proposed above are the
tion/dissolution is also of importance because many most widely accepted13 reactions describing sulfide
empirical equations exist. Recent work proposed by chemistry in aqueous solution. Other authors18-20 have
Sun13 suggests using the following equation derived proposed different pathways for iron sulfide (FeS) for-
from the work of Greenberg and Tomson14 and Silva, mation and dissolution. An example is the so-called
et al.:15 “direct” reaction:

2.1963 Fe( s ) + H2S → FeS

FeS( s ) + H2 (21)
g K sp,FeCO = ––59
log 59.3498 – 0.041377 × Tk – +
3 Tk (12)
24.5724 × log( Tk ) + 2.518 × I0.5 – 0.657 × I
724 It is assumed that the iron dissolution does not occur,
but rather a fast oxidation of solid iron transforms
where Ksp,FeCO3 is the solubility constant for iron car- it directly into solid iron sulfide attached to the steel
bonate (mol2/L2), Tk is temperature (K), and I is ionic surface. The mechanism of this reaction is still under
strength (mol/L). investigation, including the role of various species in
In addition, as more and more field conditions the formation of the different types of iron sulfide.
involve the presence of large quantities of H2S, the The uncertainty related to the expressions for the
prediction of sour corrosion appears today as one of equilibrium constants involved in H2S aqueous chem-
the most pressing matters in the oil and gas indus- istry is much more acute than with CO2 aqueous
try.16 The understanding of H2S corrosion mecha- chemistry and it is therefore necessary to state pre-
nisms lags clearly behind, even if a lot of effort has cisely what has been used in this work. Sun13 postu-
already been made in this direction.17 Although H2S lated the following equations were the most reliable:
gas is about three times more soluble than CO2 gas, from Suleimenov and Krupp:21
the acid created by dissociation of H2S is about three 
–  634.27 + 0.2709 TK – 0.11132 ×10 –3 TK2 –
16,719
– 261.9

. log TK 
 TK 
times weaker than carbonic acid (H2CO3). Hence, the KH = 10 (22)
2S
effect of H2S gas on decreasing the solution pH is
approximately the same as for CO2 gas. The different
from Suleimenov and Seward:22
chemical and electrochemical reactions involved in 20,565.7315
361 TK –1.6722 ×10 –4 TK2 –
782.43 + 0.361T –142.741
. ln TK
the H2S corrosion are described below and are added K1 = 10
TK
(23)
to the list above:
from Benning, et al.:23
Hydrogen sulfide dissolution: 2,848.779
.
KH S – 6.347
. + log( K1 )
TK
H2S( g ) ←
→ H2S( aq )
2
(13) K sp,mck = 10 (24)

015004-2 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

where KH2S is the solubility constant of hydrogen sul-

fide (mol/L/bar), K1 is the dissociation constant of
H2S (mol/L), Ksp,mck is the solubility limit of macki-
nawite (FeS) (mol/L), and Tk is temperature (K).
Sun13 also stated that there was not yet a reliable
expression for the second dissociation constant, K2,
and the use of this constant to calculate the solubility
limit should be avoided altogether. Consequently, to
avoid the HS– dissociation equilibrium constant, the
FeS solubility is calculated in the present paper using
the following equation:

Fe 2 +  × HS – 
K sp,mck =    
(25)
H +  FIGURe 1. Corrosion regimes in CO2/H2S corrosion as defined by
 
Pots, et al.29
Some valuable experimental work has been con-
ducted recently to determine the effect of small
amounts of H2S on the CO2 corrosion of carbon
steel.24-27 It was shown that small amounts of H2S
lead to a rapid and significant reduction of the CO2
corrosion rate. The reduction of the corrosion rate is
usually associated with the formation of a corrosion
product film on the metal surface, even if the bulk
conditions for supersaturation of FeCO3 or FeS are
not met. The analysis of the film usually shows the
presence of a very thin mackinawite film. It has been
reported that the formation of mackinawite on mild
steel is a very fast, direct surface reaction leading to a
solid adherent mackinawite layer.18-20,27-28 Depending
on various environmental factors, different thermody-
namically stable types of FeS can be formed. In some
cases FeS film can be nonprotective and result in
localized attack. For example, the formed layer can FIGURe 2. Corrosion product formation as a function of temperature
develop internal stresses that can lead to film frac- and H2S.30
ture, causing a potential localized attack problem.
Generally, three regimes in CO2/H2S systems have
been classified based on the concentration of H2S as The corrosion product map related to the
shown in Figure 1.29 Nevertheless, the mixed CO2/H2S formation of these three types of FeS film is shown
zone has been reported to begin at a much smaller in Figure 2.
ratio than the one that is displayed in the graph in The influence of organic acids on the relative
Figure 1.26 The chemistry of iron sulfide film forma- protectiveness of iron sulfide films adds another
tion is very complex and the film characteristics unknown to the problem since there is, to our knowl-
and morphology can change with test conditions edge, only one published paper on that subject.31
and time, which can lead to very different corrosivity The objective of this paper was to try to improve
scenarios. Smith and Pacheco30 reported that there our understanding of the influence of the partial pres-
are three main forms of iron sulfide that are com- sure of H2S and the presence of acetic acid on the
monly found in the field, which are mackinawite, CO2-influenced corrosion rate of carbon steel and the
pyrrhotite, and pyrite. characteristics of the corrosion product film formed.
—Mackinawite is a metastable form of FeS that
forms in the presence of small amounts of H2S. ExpErimEntal procEdurEs
—Pyrrhotite is believed to be more thermodynam-
ically stable than mackinawite because the pyr- Test Matrix
rhotite formation kinetics are much slower than Table 1 presents the experimental conditions
those of mackinawite. of each test. Only two parameters (free acetic acid
—The formation of pyrite is associated with high concentration and H2S partial pressure) are varied
H2S partial pressure and is believed to require around a set of baseline conditions (Test 1). The influ-
elemental sulfur. ence of these two parameters are studied separately

CORROSION—Vol. 67, No. 1 015004-3

CORROSION SCIENCE SECTION

Table 1
Experimental Conditions
Acetic Acid Series H2S Series Acetic Acid/H2S Series
Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9
Fe2+ Fe2+ Fe2+ Fe2+ Fe2+ Fe2+ Fe2+ Fe2+ Fe2+
 Duration pH (ppm) pH (ppm) pH (ppm) pH (ppm) pH (ppm) pH (ppm) pH (ppm) pH (ppm) pH (ppm)
At start NA(A)
NA 3.4 NA NA NA 4.2 7.9 4.3 9 4.4 6.5 4.8 40 4.4 56.3 3.9 76
After 2 days 4.6 0.4 NA NA 3.9 70 4.4 19 4.4 NA 4 25 NA NA 4.2 145 4.3 94
After 7 days 4.9 8.4 NA 10 3.6 40 4.4 19 4.5 NA 4.1 23 4.7 NA 4.5 110 4.3 NA
After 14 days 4.6 11 4 24 3.7 36 4.6 NA 4.4 18 4.3 25 4.7 70 4.5 150 4.1 170
After 21 days 4.8 11 4 17 3.7 32 4.7 18 4.5 20 4.3 26 4.7 35 4.6 170 4.3 140
(A)
Not available.

Table 2
Saturation Level with Regard to FeCO3 and FeS
   Test No. Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9
FeCO3 saturation(A) 0.12 0.005 0.002 0.13 0.06 0.03 0.24 0.7 0.13
FeS saturation (A) 0 0 0 1.6 12 11 3 9 62
(A)
Calculations are made using solubility constants obtained by Sun.13

Table 3
Acetic Acid Concentration
Calculated Free Acetic Acid Concentration
in the Liquid Phase (ppm)
Measured Total Acetate Species Based on the Amount Based on the Amount
([free HAc] + [Ac–]) in the of Acetate Species of Acetate Species
Test Liquid Phase with Ion Measured with Ion Introduced in the
No. Chromatograph (ppm) Chromatography Experimental Loop
Test 2 57 Between 50 and 55 Between 87 and 96
Test 3 675 Between 605 and 664 Between 895 and 944
Test 7 NA NA Between 46 and 57
Test 8 1,052 Between 656 and 846 Between 630 and 810
Test 9 1,120 Between 861 and 1,002 Between 772 and 895

(Tests 2 through 6) and then combined in Tests 7, 8, higher liquid flow rate for lubrication purposes. The
and 9. A set of measurements performed throughout values of liquid flow rates were, consequently, slightly
the tests (pH measurement, iron, and acetic acid con- different between the sour and sweet tests. The pri-
centrations) is presented in Tables 2 and 3. mary concern was to ensure that the flow regime was
The nine experiments conducted can be divided stratified with a superficial liquid velocity determined
into three groups investigating different aspects of the to be below 0.05 m/s.
corrosion process in a CO2 environment:
—influence of the concentration of free acetic acid Experimental loop
—influence of the partial pressure of H2S
—combined effect of the concentration of free ace- Three different large-scale flow loops were used
tic acid and the partial pressure of H2S in this study. Experiments were carried out in mul-
Apart from the acetic acid concentration and the tiphase, stratified flow with water and a mixture
partial pressure of H2S, all the other experimental of CO2/N2/H2S. The flow loops, made of Type 316
parameters were kept at a fixed value (system tem- (UNS S31600)(1) stainless steel and Alloy C276 (UNS
perature: 70°C, partial pressure of CO2: 2 bars, total N10276) (for the H2S experiments) all have very simi-
pressure: 3 bars, gas velocity: 5 m/s). The sweet and lar characteristics and can be divided in three main
sour tests were performed in two similar but indepen- component parts: the tank, the pump, and the loop.
dent flow loops. The design of the “sour” loop requires —The tank (2,000 L) is used for the liquid phase
conditioning and heating. It is filled with deion-
ized water. Acetic acid is added to reach the
(1) UNS numbers are listed in Metals and Alloys in the Unified Num-
bering System, published by the Society of Automotive Engineers requirements of the tests. A set of immersion
(SAE International) and cosponsored by ASTM International. heaters control the temperature.

015004-4 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

—Positive displacement progressive cavity pumps

or a gas blower are used to move the liquid or
the gas phase.
—The 4-in (101.6-mm) diameter flow loop is 30 m
long and horizontally leveled. The test sections,
where the measurements are taken, are located
at least 8 m downstream from the exit of the
tank. The test sections (Figure 3) are 1.5-m-long
pipe spool pieces. Each has up to four probe
ports (two at the top, two at the bottom). In this
paper, only the bottom-of-the-line results are
taken into account. Samples of condensed liq-
uid and in situ pH measurements were taken at
the test section.
More complete presentations of these loops have
been published previously in the open literature,26,32-33
and the reader is directed to consult these publica-
tions for more detailed information.
FIGURE 3. Test section of the H2S loop. Only the bottom ports were
used in this study.
Experimental Procedures
The tank was filled first with 1 m3 of deionized Table 4
water. Carbon dioxide (and nitrogen in some cases) Chemical Analysis of the Carbon Steels
was injected in the loop at a specific pressure. The liq- Used in the Experiements
uid phase then was heated up to the specific temper-
X65 Composition API 5L X65
ature by two electrical resistance heaters. The pump Element (%) (%)
was started and the gas/liquid mixture flowed around
C 0.13  <0.26
the loop in a stratified flow regime. Deoxygenation
Mn 1.16  <1.40
was performed by depressurizing several times until P 0.009 <0.03
the concentration of oxygen had been reduced suffi- S 0.009 <0.03
ciently (<50 ppb) so as to not affect the overall corro-
sion process.34 Oxygen concentration was measured
using colorimetric test kits. Once the deoxygenation Scale Formation
was completed, acetic acid and/or H2S concentra- The pH of the liquid phase at the bottom of the
tions were adjusted (see procedures below) to the line was kept at a value around 4 to 4.5 in each test
required levels. The corrosion probes then were intro- and the solution was always undersaturated with
duced under pressure at the test section, which was regard to iron carbonate precipitation. Therefore,
equipped with two probe ports and the experiment FeCO3 was not expected to form on the metal surface
began. Typically, two weight-loss coupons were intro- in any of the tests performed. On the other hand, the
duced at the beginning of the test and replaced after solution was always supersaturated with regard to
2 days of exposure. Then, 7 days later, one coupon FeS (in the tests with H2S). The saturation levels for
was replaced once again and the experiment contin- each type of scale are displayed in Table 4 with super-
ued for an additional 14 days. This way, there were saturation achieved when values were more than 1.
indications of the corrosion rate after 2, 7, 14, and
21 days of testing. A data-acquisition device was used Acetic Acid Concentration
to measure the total pressure and the gas/liquid tem- The acetic acid (HAc) concentration was adjusted
perature continuously. by adding a calculated amount of pure HAc in the
tank. The acetic acid solution was deoxygenated first
Liquid Phase Specification before being introduced into the tank using a high-
The liquid phase was made up exclusively from pressure vessel connected to the tank. Several liquid
deionized water. No salt was added. However, dis- samples then were taken and analyzed by ion chro-
solved ferrous iron (Fe2+) buildup occurred through- matography to verify the concentration of the total
out the test as a result of the corrosion process on the acetate species (free HAc + Ac–) present in the liquid
weight-loss coupons. Samples were taken regularly phase. The concentration distribution of free HAc and
from the liquid phase; Table 2 presents the evolu- acetate (Ac–) was calculated from the measured value
tion of Fe2+ concentration (obtained using a spectro- of pH. A differentiation was made between the free or
photometer) and pH during the whole duration of the undissociated acetic acid concentration (free HAc) and
tests. the total acetic acid concentration, which included

CORROSION—Vol. 67, No. 1 015004-5

CORROSION SCIENCE SECTION

of H2S was introduced in pure gas form at the begin-

ning of the test and checked regularly using a piston
pump and low-range standard detection tubes. The
length of color change in the detection tube reagent
was measured using calipers to increase the accuracy
of the value. Repeatability of this method was found
to be ±5%. The trace amounts of H2S introduced in
the loop were consumed fairly rapidly by the corro-
sion process and the H2S partial pressure had to be
adjusted almost every day to be kept within 20% of
the targeted value.

Corrosion Rate Measurement

The weight-loss coupons were not inserted into
the corrosion environment until the system had
reached steady state (stable temperature, pressure,
and flow velocities). The corrosion rates were mea-
FIGURe 4. Weight-loss coupons with PTFE coating at the back and sured with weight-loss coupons made of API X65 car-
the side (external diameter = 3.17 cm). bon steel. Samples consisting of circular coupons
(0.76 cm internal diameter, 3.17 cm external diameter,
0.5 cm thickness) with an exposed area of 7.44 cm2
all acetate-containing species (free HAc and acetate were polished using isopropanol (C3H8O) as coolant on
Ac–). To keep the concentration of free acetic acid con- silicon carbide (SiC) papers, up to 600 grit. After this
stant during the test, the pH of the liquid phase was preparation, they were covered with liquid polytetra-
adjusted if necessary by adding more acetic acid. fluoroethylene (PTFE) on the outer edges and under-
Table 3 presents the calculated free acetic con- side (Figure 4). Following 4 h to 6 h of curing at
centration at the bottom of the line for each test. Ion ambient conditions, the samples were held at 200°C
chromatograph measurements providing total ace- in an oven for 30 min according to manufacturer’s
tate species concentration ([HAc]free + [Ac–]) are shown recommendations. The uncovered steel surface was
where available. A range of concentration is provided then repolished with 600-grit SiC paper wetted with
each time to take into account the change in pH. The isopropanol, cleaned, dried, and weighed. A picture of
calculations of free undissociated acetic acid concen- a coupon after preparation is shown in Figure 4. The
tration were performed using the equilibrium con- coupons then were flush-mounted on the internal
stant of acetic acid dissociation proposed by Kharaka, pipe wall of the loop by using a specially designed
et al.,35 and the following equation (the total acetate probe holder. Therefore, only one face of the coupon
species concentration ([HAc]free + [Ac–]) is known: was in direct contact with the corrosive environment.
The exposure time was between 2 days and 21 days in

[HAc ]free =
([HAc]
free + [ Ac )
Ac – ] × [H+ ]
(26)
all experiments. Upon removal from the loop, the cou-
pon surface was flushed with isopropanol, to dehy-
+
HAc + [H ]
K HAc drate it, and images of the surface were taken. The
weight of the coupon after each test was registered.
It is important to note that the calculated concen- The ASTM G136 standard procedure was followed to
tration of free acetic acid was close to the required remove the corrosion products and determine the cor-
concentration, but there was, in most cases, a 20% rosion rate by weight loss. One coupon is generally
to 30% discrepancy. This discrepancy is most likely used for weight loss, and the other is preserved for
from the technical difficulties often met in large-scale corrosion product evaluation by scanning electron
loop tests to keep a high degree of accuracy in the microscopy (SEM) and energy-dispersive x-ray spec-
measurements but also includes errors in the mea- troscopy (EDS).
surement process. For clarity purposes, the concen-
tration of free acetic acid will be displayed as 100 ppm Materials Characterization
or 1,000 ppm (depending on the test conditions) Weight-loss coupons were machined from API
throughout this paper. X65 carbon steel prepared from the same piece of field
pipe line (33-cm outside diameter pipe section, 3.8 in
Gas Phase Composition [96.52 mm] thickness). The chemical analyses of this
In all the experiments, the gas phase was made X65 steel is shown in Table 5. Figure 5 shows the
of a mixture of CO2 and N2 (2 bars of CO2 and 0.7 bars microstructure of longitudinal and transversal cuts of
of N2, 0.3 bars of water vapor) for a total pressure of the X65 carbon steel. In this case, only the face of the
3 bars. For the H2S environment, the required amount coupon that would be in contact with the fluid was

015004-6 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

Table 5
Hardness (HRB) Results
X65 Longitudinal Cut X65 Transversal Cut
1 81.3 60.3
2 94.4 68.7
3 98.7 63.3
4 87.9 78.0
5 95.4 59.1
6 89.3 51.1
7 88.7 66.5
8 92.9 75.0
9 93.3 58.5
10 85.1 67.7
Average 90.7 64.8

Approx. tensile strength 90,000 psi for 90.7 HRB 56,000 psi for 65.7 HRB
Tensile requirements 77,000 psi (min.) 77,000 psi (min.)
Yield strength 65,000 psi (min.) 65,000 psi (min.)

(a) (b)
FIGURE 5. Microstructure of the X65 carbon steel: (a) longitudinal cut and (b) transversal cut.

Table 6
Test Matrix
Investigating Acetic Acid H2S Acetic Acid/H2S
Experiment No. 1 2 3 4 5 6 7 8 9
Free HAc tank (ppm) 0 100 1,000 0 0 0 100 1,000 1,000
pH2S (bar) 0 0 0 0.004 0.07 0.13 0.004 0.004 0.13
Common parameters—steel type: API X65; liquid phase composition: DI water; test duration: 3 weeks; absolute pressure: 3 bars; pCO2: 2 bars;
gas temperature: 70°C; gas velocity: 5 m/s; superficial liquid velocity: <0.05 m/s.

evaluated. Cuts were made using cooling fluids and culated and compared with the values designated for
the proper metallographic diamond blade to avoid the metal in the standards. The X65 showed a differ-
modifications of the microstructure. The microstruc- ence in hardness with direction. This change in hard-
ture of the X65 was finer in the longitudinal direction, ness is consistent with the change in microstructure
probably as a result of processing. Figure 5 shows a previously described.
microstructure typical of a microalloyed thermome-
chanical controlled processing (TMCP) pipeline steel. RESULTS AND DISCUSSION
Iron carbide was distributed in spheroidized form
instead of a lamellar arrangement. Hardness mea- Corrosion Rate Results
surements are recorded in Table 6. By converting Corrosion rate results are displayed in a series of
these values,37 approximate tensile strength was cal- graphs, Figures 6 through 12. Error bars representing

CORROSION—Vol. 67, No. 1 015004-7

CORROSION SCIENCE SECTION

FIGURE 6. Influence of the free HAc concentration. Evolution of the FIGURE 7. Influence of partial pressure of H2S. Evolution of the
general corrosion rate over time. (Absolute pressure [PT]: 3 bars, general corrosion rate with partial pressure of H2S. (PT: 3 bars, pCO2:
partial pressure of CO2 [pCO2]: 2 bars, pH2S: 0 bar, gas temperature 2 bars, free HAc: 0 ppm, Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s).
[Tg]: 70°C, gas velocity [Vg]: 5 m/s, superficial liquid velocity [Vl] <
0.05 m/s).

maximum and minimum values and the number of

coupons (number of repeated measurements) are dis-
played when applicable on each graph.
Influence of the Free Acetic Acid Concentration
— The observed influence of the free acetic acid con-
centration on the average corrosion rate was not sur-
prising. The acetic acid acted as a provider of protons
and at the same time added a new cathodic reac-
tion via the direct reduction of undissociated acetic
acid. Therefore, in the absence of protective corrosion
product film (as is the case here), the average corro-
sion rate increased when free acetic was present. As
is shown in Figure 6, the corrosion rate doubled if
1,000 ppm of free acetic acid was introduced in solu-
tion, reaching a maximum well above 10 mm/y. In all
cases, the corrosion rate might have varied slightly
FIGURE 8. Influence of the partial pressure of H2S. Evolution of the with time but was expected to remain rather high,
general corrosion rate over time. (PT: 3 bars, pCO2: 2 bars, HAc: since no protective film could form in the severe con-
0 ppm, Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s). ditions tested (pH always below 4.5).
Influence of the Partial Pressure of Hydrogen Sul-
fide — Results related to the influence of the partial
pressure of H2S are shown in Figures 7 and 8, which
represent the same data but are displayed in different
ways for purposes of clarity. The first H2S partial pres-
sure tested was 0.004 bar at a CO2/H2S ratio of 500
(at the limit between the reported sweet and mixed
regimes). However, there was obviously an overwhelm-
ing effect of H2S since the corrosion rates were glob-
ally reduced to 0.08 mm/y (almost a factor of 100). As
more H2S was added up to 0.07 bar and 0.13 bar (cor-
responding to CO2/H2S ratio of, respectively, 29 and
15), the tendency was reversed with the average cor-
rosion rate being only 10 to 20 times lower as com-
pared to the pure CO2 environment. With 0.13 bar of
H2S, the corrosion regime was assumed to be clearly
FIGURE 9. Combination of the partial pressure of H2S and the in the so-called “sour regime,” i.e., dominated by the
concentration of free HAc. Evolution of the general corrosion rate presence of H2S. It is assumed here that the main
over time. (PT: 3 bars, pCO2: 2 bars, pH2S: 0.004 bar, Tg: 70°C, Vg:
effect of high H2S content is the significant rate of the
5 m/s, Vl < 0.05 m/s).
additional cathodic reaction (H2S reduction) and that

015004-8 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

the FeS film formed during these tests has similar

protectiveness as in the previous tests with lower par-
tial pressure of H2S. Moreover, there seemed to be a
slight decreasing trend in time in the average corro-
sion in the presence of H2S, which could correspond
to the gradual formation of a corrosion product film
layer. Some FeS films (especially mackinawite) are
believed to form almost immediately with small quan-
tities of H2S and to create a diffusion barrier, which
prevents corrosive species (H+, H2CO3) from reaching
the surface.30
Combined Effect of the Acetic Acid and Hydrogen
Sulfide — Results are shown in a series of graphs, Fig-
ures 9 through 12. While the presence of 100 ppm of
acetic acid had little influence on the corrosion rate, FIGURE 10. Combination of the partial pressure of H2S and the
the effect was much stronger with 1,000 ppm of free concentration of free HAc. Evolution of the general corrosion rate
over time. (PT: 3 bars, pCO2: 2 bars, pH2S: 0.13 bar, Tg: 70°C, Vg:
acetic. The influence of this concentration of free ace-
5 m/s, Vl < 0.05 m/s).
tic acid on the general corrosion rate was even greater
compared to the pure CO2 environment, as described
above. The 1,000 ppm of HAc increased the aver-
age uniform corrosion rate by 50 times at 0.004 bar
H2S (Figure 9) and 8 times at 0.13 bar H2S (Figure 10)
with an average corrosion rate after 21 days of expo-
sure reaching 4 mm/y and 8 mm/y, respectively. The
beneficial effect of the protective sulfide scale shown
above was all but completely canceled by the pres-
ence of HAc, and the magnitude of the uniform corro-
sion rate was not acceptable anymore (Figure 11). In
the presence of 1,000 ppm of free acetic acid, the par-
tial pressure of H2S did not seem to have a very strong
influence with a corrosion rate being globally half its
sweet conditions’ equivalent (Figure 12).

Surface Analysis
The corrosion product layer was studied system- FIGURE 11. Combination of the partial pressure of H2S and the
atically for each test using SEM and EDS. However, concentration of free HAc. Evolution of the general corrosion rate
over time. (PT: 3 bars, pCO2: 2 bars, free HAc: 1,000 ppm, Tg: 70°C,
the complete characterization of corrosion product
Vg: 5 m/s, Vl < 0.05 m/s).
films (and especially sulfide films) requires x-ray dif-
fraction (XRD) analysis, which, unfortunately, was not
performed in this study because of limited availability
of the equipment and time constraint. Therefore, even
if the visual observations obtained by SEM give some
useful indications about the nature of the corrosion
product film, some caution should be taken when
interpreting the findings.
Influence of the Free Acetic Acid Concentration
— Figure 13 presents typical pictures of the corro-
sion product film that forms on the surface in a sweet
environment when the conditions for FeCO3 precipi-
tation are not met (low pH). The scale appeared to be
uniform and smooth when the coupon was removed
from the test loop, but cracks appeared rapidly dur-
ing the drying process (when isopropanol was sprayed
on the coupon surface immediately after the test to
dehydrate the steel surface). EDS and SEM analyses FIGURE 12. Combination of the partial pressure of H2S and the
of the film were consistent with the notion of cement- concentration of free HAc. Evolution of the general corrosion rate
with the partial pressure of H2S. (PT: 3 bars, pCO2: 2 bars, free HAc:
ite (Fe3C) being the undissolved component of the
1,000 ppm, Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s).
steel left behind by the corrosion process, although

CORROSION—Vol. 67, No. 1 015004-9

CORROSION SCIENCE SECTION

(a) (b)

(c) (d)

(e)
FIGURE 13. Test 1—Pure CO2 environment: (a) WL coupon after 14 days of exposure, (b) corrosion product layer X100, (c)
corrosion layer X100 back scatter, (d) corrosion product layer X200, and (e) EDS anaylsis of the corrosion layer. (PT : 3 bars,
pCO2: 2 bars, pH2S: 0 bar, free HAc: 1,000 ppm, Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s).

the presence of small amounts of FeCO3 cannot be weight-loss coupons; they were flushed with isopro-
ruled out at this stage. Fe3C forms a nonprotective, panol immediately after their removal from the loop.
very porous film, and has been reported in some case There was no indication of localized corrosion and the
to enhance the corrosion process by galvanic coupling corrosion process was strictly uniform. There was no
(Fe3C is electronically conductive).38-39 The numer- specific influence of the nature of the corrosion prod-
ous cracks observed in the SEM pictures of Figure uct layer once acetic acid was added. However, since
13 appeared during the dehydration process of the the corrosion rate increased with the addition of ace-

015004-10 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

tic acid, the weight of the layer increased as well (Fig-

ure 14).
Influence of the Partial Pressure of Hydrogen Sul-
fide — Once traces of H2S were added to the system,
the corrosion product layer seemed fairly differ-
ent. With a partial pressure of H2S of 0.004 bar, the
weight-loss coupons were covered with a very thin,
visually amorphous layer (Figure 15). Even if differ-
ent film structures could be observed on the sur-
face, EDS analysis shows a similar composition. Once
again, it is worth stressing that EDS is insufficient to
discriminate between different types of iron sulfides,
most of which have a similar elemental composition.
As the partial pressure of H2S increased, the corro-
sion became more severe and the corrosion product FIGURE 14. Influence of the acetic acid on the weight of the corrosion
layer became thicker. The film was poorly adherent product layer. The full lines represent tests performed without
to the surface and large parts actually fell off when free acetic acid. The dotted lines represent tests performed with
the coupons were processed after the tests. Even if 1,000 ppm of free acetic acid (PT: 3 bars, pCO2: 2 bars, Tg: 70°C,
EDS analysis gives a similar FeS film composition, the Vg: 5 m/s, Vl < 0.05 m/s).
structure of the film looks different when compared to
the amorphous FeS coinciding with hexagonal-shaped
crystals (Figure 16). Without XRD analysis, it is dif- of the presence of pyrrhotite at the higher H2S partial
ficult to characterize the film composition, but pyr- pressure. Pitting corrosion is observed in Figure 17(b)
rhotite or cubic FeS could form in these conditions and Figure 18(b) when 1,000 ppm of free acetic acid
and could match the hexagonal shape of the crystals is present in the solution. The surface profile analysis
observed.30 The weight of the film increased by a fac- presented in Figure 19 shows relatively low magnitude
tor of 100, when comparing 0.004 bar and 0.13 bar pitting corrosion, adding between 1 mm/y and 2 mm/y
of H2S corrosion coupons, while the corrosion rate to the already high average corrosion rate (4 mm/y for
itself only increased by a factor of 10. In the presence test 8 and 8 mm/y for test 9).
of 0.13 bar of H2S, the liquid phase is strongly super-
saturated with FeS. Precipitation of FeS and the direct CONCLUSIONS
surface reaction to form mackinawite are expected
to occur simultaneously. Internal stresses are prob- ❖  In the presence of 2 bars of CO2, the average cor-
ably caused by the growth of the corrosion product rosion at the bottom of the line (nonfilm-forming con-
film underneath the already existing layer (instead ditions) is approximately doubled when 1,000 ppm of
of on top of the existing layer seen in precipitation). undissociated acetic acid is present. The film present
Therefore, the growth of the scale associated with the on the metal surface (most likely formed of a mixture
direct surface reaction should be at the origin of the of Fe3C and FeCO3) is unprotective and no localized
scale breakdown observed with SEM. Large parts of corrosion could be observed.
the coupons where the scale peeled off are much more ❖  The presence of trace amounts of H2S (0.004 bar) in
corroded than the rest. This seems initially counter- the CO2 environment sharply decreases the corrosion
intuitive because conventional wisdom suggests that rate by two orders of magnitude. A FeS film, protective
locations with more mackinawite scale should be bet- at low H2S partial pressure, covers the surface.
ter protected against corrosion; however, once inter- ❖  As the partial pressure of H2S is increased to
preted in the light of the argument presented above, 0.13 bar, the tendency is reversed and the general
one can accept that locations with more corrosion corrosion rate is increased by an order of magnitude,
have seen more mackinawite formation. although it never reached the high rates observed in
Combined Effect of Acetic Acid and Hydrogen Sul- a pure CO2 environment. At higher H2S content, the
fide — Figures 17 and 18 present the SEM pictures of scale seems to break easily as a result of internal
the corrosion scale from 1,000 ppm of free acetic acid stresses and the steel not being evenly corroded.
and 0.004 bar or 0.13 bar of H2S. The FeS layer looks ❖  The introduction of 1,000 ppm of free acetic acid
very similar in both cases: a uniform film, entirely in the H2S/CO2 mixture negated the positive effects
cracked, thick, and non-adherent. The weight of the of partial pressure of hydrogen sulfide (pH2S) and led
film is, in both cases, around 6 times what it was in to average corrosion rates comparable to their pure
the absence of any acetic acid (Figure 14). The compo- sweet condition equivalents (same order of magni-
sition of the scale is expected to be mainly macki- tude). Some pitting and localized corrosion could be
nawite, but some crystals could be seen at 0.13 bars observed on the metal surface in the presence of ace-
of H2S (Figure 18[f]). Once again, it could be evidence tic acid.

CORROSION—Vol. 67, No. 1 015004-11

CORROSION SCIENCE SECTION

(a) (b)

(c) (d)

(e) (f)
FIGURE 15. Test 4—CO2 environment with traces of H2S–CO2/H2S: 500: (a) WL coupon after 21 days of exposure, (b)
WL coupon after removal of the layer, (c) corrosion product layer X100, (d) corrosion layer X50 back scatter, (e) corrosion
product layer X2,000, and (f) corrosion product layer X2,000. (PT: 3 bars, pCO2: 2 bars, pH2S: 0.004 bar, free HAc: 0 ppm,
Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s).

REFERENCES 4. Y. Garsany, D. Pletcher, B. Hedges, “The Role of Acetate in CO2

Corrosion of Carbon Steel: Has the Chemistry Been Forgotten?,”
CORROSION/2002, paper no. 02273, (Houston, TX: NACE,
1. K. George, S. Wang, S. Nes̆ić, C. de Waard, “Modeling of CO2 Cor- 2002).
rosion of Mild Steel at High Pressures of CO2 and in the Presence 5. Y. Sun, K. George, S. Nes̆ić, “The Effect of Cl– and Acetic Acid on
of Acetic Acid,” CORROSION/2004, paper no. 04623 (Houston, Localized CO2 Corrosion in Wet Gas Flow,” CORROSION/2003,
TX: NACE International, 2004). paper no. 03327 (Houston, TX: NACE, 2003).
2. C. De Waard, U. Lotz, “Prediction of CO2 Corrosion of Carbon 6. K. George, S. Nes̆ić, C. de Ward, “Electrochemical Investigation
Steel,” CORROSION/1983, paper no. 69 (Houston, TX: NACE, and Modeling of Carbon Dioxide Corrosion on Carbon Steel in the
1993). Presence of Acetic Acid,” CORROSION/2004, paper no. 04379
3. B. Hedges, L. McVeigh, “The Role of Acetate in CO2 Corrosion: (Houston, TX: NACE, 2004).
The Double Whammy,” CORROSION/1999, paper no. 21 (Hous- 7. S. Nes̆ić, J. Postlethwaite, S. Olsen, Corrosion 52, 4 (1996): p.
ton TX: NACE, 1999). 280, doi:10.5006/1.3293640.

015004-12 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

(a) (b)

(c) (d)

(e) (f)

FIGURE 16. Test 6—CO2 environment with H2S–CO2/H2S: 15: (a) WL coupon after 21 days of exposure, (b) WL coupon
after removal of the layer, (c) corrosion product layer X50, (d) corrosion layer X500 back scatter, (e) corrosion product layer
X2,000, and (f) corrosion product layer X10,000. (PT: 3 bars, pCO2: 2 bars, pH2S: 0.13 bar, free HAc: 0 ppm, Tg: 70°C, Vg:
5 m/s, Vl < 0.05 m/s).

8. M. Nordsveen, S. Nes̆ić, R. Nyborg, A. Stangeland, Corrosion 59, 5 13. W. Sun, “Kinetics of Iron Carbonate and Iron Sulfide Scale For-
(2003): p. 443, doi:10.5006/1.3277576. mation in CO2/H2S Corrosion” (Ph.D. diss., Ohio University,
9. J.L. Crolet, N. Thevenot, A. Dugstad, “Role of Free Acetic Acid on Chemical and Biomolecular Engineering Department, 2006).
the CO2 Corrosion of Steel,” CORROISON/1999, paper no. 24 14. J. Greenberg, M. Tomson, Appl. Geochem. 7 (1992): p. 185-190.
(Houston, TX: NACE, 1999). 15. C.A.R. Silva, X. Liu, F.J. Millero, J. Solution Chem. 31 (2002): p.
10. A. Dugstad, “The Importance of FeCO3 Supersaturation on the 97-108.
CO2 Corrosion of Carbon Steel,” CORROSION/1992, paper no. 14 16. M. Bonis, M. Girgis, K. Goerz, R. MacDonald, “Weight-Loss Cor-
(Houston, TX: NACE, 1992). rosion with H2S: Using Past Operations for Designing Future Fa-
11. S. Nes̆ić, K.L. Lee, Corrosion 59, 7 (2003): p. 616, doi:10.5006/ cilities,” CORROSION/2006, paper no. 6122 (Houston, TX:
1.3277592. NACE, 2006).
12. O. Nafday, S. Nes̆ić, “Iron Carbonate Film Scale Formation and 17. S. Smith, M. Joosten, “Corrosion of Carbon Steel by H2S in CO2-
CO2 Corrosion in the Presence of Acetic Acid,” CORROSION/ Containing Environments,” CORROSION/2006, paper no. 06115
2005, paper no. 05295 (Houston, TX: NACE, 2005). (Houston, TX: NACE, 2006).

CORROSION—Vol. 67, No. 1 015004-13

CORROSION SCIENCE SECTION

(a) (b)

(c) (d)

(e)
FIGURE 17. Test 7—CO2 environment with traces of H2S and acetic acid—CO2/H2S: 500: (a) WL coupon after 21 days of
exposure, (b) WL coupon after removal of the layer, (c) corrosion product layer X50, (d) corrosion layer X200, and (e) EDS
analysis. (PT: 3 bars, pCO2: 2 bars, pH2S: 0.004 bar, free HAc: 1,000 ppm, Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s).

18. W. Sun, S. Nes̆ić, D. Young, R. Woollam, Ind. Eng. Chem. Res. 47 23. L.G. Benning, R.T. Wilkin, H.L. Barnes, Chem. Geol. 167 (2000):
(2008): p. 1738-1742. p. 25-51.
19. D. Rickard, G.W. Luther, Chem. Rev. 107 (2007): p. 514-562. 24. A. Valdes, R. Case, M. Ramirez, A. Ruiz, “The Effect of Small
20. J. Vera, S. Kapusta, N. Hackerman, J. Electrochem. Soc. (1986): Amounts of H2S on CO2 Corrosion of Carbon Steel,” CORROSION/
p. 133-461. 1998, paper no. 22 (Houston, TX: NACE, 1998).
21. O.M. Suleimenov, R.E. Krupp, Geochim. Cosmochim. Acta 58 25. J. Kvarekval, “The Influence of Small Amounts of H2S on CO2
(1994): p. 2433-2444. Corrosion of Iron and Carbon Steel,” Eurocorr 97 (London, U.K.:
22. O.M. Suleimenov, T.M. Seward, Geochim. Cosmochim. Acta 61 IOM Communications for European Federation of Corrosion,
(1997): p. 5187-5198. 1999).

015004-14 CORROSION—JANUARY 2011

 CORROSION SCIENCE SECTION

(a) (b)

(c) (d)

(e) (f)
FIGURE 18. Test 9—CO2 environment with H2S and acetic acid—CO2/H2S: 15: (a) WL coupon after 21 days of exposure,
(b) WL coupon after removal of the layer, (c) corrosion product layer X50, (d) corrosion layer X5,000, (e) corrosion product
layer X500, and (f) corrosion product layer X500. (PT: 3 bars, pCO2: 2 bars, pH2S: 0.13 bar, free HAc: 1,000 ppm, Tg: 70°C,
Vg: 5 m/s, Vl < 0.05 m/s).

26. B. Brown, K.L. Lee, S. Nes̆ić, “Corrosion in Multiphase Flow Con- Management,” CORROSION/2002, paper no. 02235 (Houston,
taining Small Amounts of H2S,” CORROSION/2003, paper no. TX: NACE, 2002).
03341 (Houston, TX: NACE, 2003). 30. S. Smith, J.L. Pacheco, “Prediction of Corrosion in Slightly Sour
27. B. Brown, S. Reddy Parakala, S. Nes̆ić, “CO2 Corrosion in Pres- Environments,” CORROSION/2002, paper no. 02241 (Houston,
ence of Trace Amounts of H2S,” CORROSION/2003, paper no. TX: NACE, 2002).
04736 (Houston, TX: NACE, 2004). 31. A. Camacho, “CO2 Top-of-the-Line Corrosion in Presence of H2S”
28. W. Sun, S. Nes̆ić, Corrosion 65, 5 (2009): p. 291, doi:10.5006/ (Master’s thesis, Ohio University, Chemical and Biomolecular En-
1.3319134. gineering Department, 2006).
29. B.F.M. Pots, S.D. Kapusta, R.C. John, M.J.J. Simon Thomas, I.J. 32. M. Singer, S. Nes̆ić, Y. Gunaltun, “Top-of-the-Line Corrosion in
Rippon, T.S. Whitham, M. Girgis, “Improvements on de Waard Presence of Acetic Acid and Carbon Dioxide,” CORROSION/2004,
Milliams Corrosion Prediction and Applications to Corrosion paper no. 04377 (Houston, TX: NACE, 2004).

CORROSION—Vol. 67, No. 1 015004-15

CORROSION SCIENCE SECTION

(a) (b)
FIGURE 19. Tests 8 and 9—Localized corrosion with H2S and acetic acid—CO2/H2S: 15. Surface analysis performed on
weight-loss coupons after removal of the corrosion product layer: (a) Test 8—pH2S: 0.004 bar, free HAc: 1,000 ppm; WL
coupon after 21 days of exposure and (b) Test 9—pH2S: 0.13 bar, free HAc: 1,000 ppm; WL coupon after removal of the layer.
(PT: 3 bars, pCO2: 2 bars, free HAc: 1,000 ppm, Tg: 70°C, Vg: 5 m/s, Vl < 0.05 m/s).

33. C. Mendez, M. Singer, A. Camacho, S. Hernandez, S. Nes̆ić, “Ef- 36. ASTM G1-03, “Standard Practice for Preparing, Cleaning, and
fect of Acetic Acid, pH, and MEG on the CO2 Top-of-the-Line Cor- Evaluating Corrosion Test Specimens” (West Conshohocken, PA:
rosion,” CORROSION/2005, paper no. 05278 (Houston, TX: ASTM International, 2004), doi: 10.1520/G0001-03.
NACE, 2005). 37. B. Craig, Practical Oilfield Metallurgy and Corrosion, 2nd ed.
34. S. Wang, “Effect of Oxygen on CO2 Corrosion of Mild Steel” (Mas- Metcorr (Tulsa, OK: PennWell Books, 1993), p. 14.
ter’s thesis, Ohio University, Chemical and Biomolecular Engi- 38. J.L. Crolet, S. Olsen, W. Wilhelmsen, “Influence of a Layer of Un-
neering Department, 2009). dissolved Cementite on the Rate of the CO2 Corrosion of Carbon
35. Y.K. Kharaka, E.H. Perkins, W.D. Gunter, J.D. Debral, C.H. Steel,” CORROSION/1994, paper no. 4 (Houston TX, NACE,
Bamford, “Solmineq 88: A Computer Program for Geochemical 1994).
Modeling of Water Rock Interactions” (Menlo Park, CA: Alberta 39. J.L. Crolet, N. Thevenot, S. Nes̆ić, Corrosion 54, 3 (1998): p. 194,
Research Council, 1989). doi:10.5006/1.3284844.

015004-16 CORROSION—JANUARY 2011