Paper No.

11260 2011

Prediction of Uniform CO2 Corrosion of Mild Steel under Inert Solid Deposits

Jin Huang, Bruce Brown, Yoon-Seok Choi and Srdjan Nešić

Institute for Corrosion and Multiphase Technology

Department Of Chemical & Biomolecular Engineering
Ohio University
342 West State Street
Athens, OH 45701
USA

ABSTRACT

The effect of an inert solid deposit on uniform CO2 corrosion of mild steel is modeled based on a
mechanistic electrochemical CO2 corrosion model. Laboratory testing has shown that the dominant
factors introduced by the inert solids deposit are related to surface coverage, where both anodic and
cathodic reaction rates are decreased because of less active surface area being exposed. The inert
solid deposits also create a mass transfer barrier for corrosive species which limits the rate of the
cathodic reactions. An existing mechanistic electrochemical model was modified to account for these
effects and was capable of capturing the main features of uniform CO2 corrosion of mild steel under
inert solid deposits.

Key words: underdeposit, CO2 corrosion, mild steel, model, mass transfer

INTRODUCTION

Severe crevice and pitting corrosion problems can be found under solid deposits in oil and gas
pipelines[1]. Localized corrosion can occur under these deposits as they can provide an environment
which is chemically and physically different than the areas which are uncovered. Such heterogeneities
may lead to formation of galvanic corrosion cells, affect inhibitor performance or harbor bacterial growth
leading to MIC,[2],[3]. Underdeposit corrosion is more prevalent at the bottom of horizontal lines and
where flow rates are lowest. However, there are very few studies to be found in the open literature
related to the mechanisms of underdeposit CO2 corrosion[4]. Most of the available literature refers to
the effect of deposits on corrosion inhibitor performance.[5] -[8] Since deposits have been reported as an
important factor which may lead to severe CO2 corrosion, it’s very important to understand first the
mechanisms of uniform corrosion under solid deposits, before focusing on the effect they have on
corrosion inhibitor performance.

©2011 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole, must be in writing to NACE
International, Publications Division, 1440 South Creek Drive, Houston, Texas 77084. The material presented and the views expressed in this paper are
solely those of the author(s) and are not necessarily endorsed by the Association.
1
Real-life scenarios for under deposit CO2 corrosion found in oil/gas pipelines are very complex. An in
situ deposit is likely to be neither pure nor inert. Rather it has complex composition and even some
reactivity. Typical deposits consist of combinations of inorganic solids such as sand, scale and
corrosion products, and organic matter such as wax and inhibitor residues. In addition, oxygen (O2),
acetic acid (CH3CO2H), hydrogen sulfide (H2S) and bacteria were found in some deposits.

In the present phase of the project on underdeposit corrosion, the initial step was made to measure and
define the effect of pure and inert deposits on uniform CO2 corrosion of mild steel. Then an effort to
model the observed behavior was made using a simple mechanistic CO2 corrosion model such as the
one proposed by Nešić, et al. [4]

EXPERIMENTAL PROCEDURE

Experiment set up

Experiments were conducted at atmospheric pressure in a three-electrode glass cell, Figure 1. The cell
was filled with 2 liters of 1 wt% NaCl solution. CO2 was continuously bubbled through the cell. API# 5L-
X65 mild steel was used as the working electrode (WE) for electrochemical measurements. Platinum
wire was used as a counter electrode (CE) and a KCl saturated silver-silver chloride (Ag/AgCl)
reference electrode (RE) was connected to the cell externally via a Luggin capillary. A glass pH
electrode was immersed in the electrolyte to monitor the pH during the experiment. Hydrochloric acid
(HCl) or sodium bicarbonate (NaHCO3) was added to adjust the pH at the beginning of the test to
desired value, which didn’t change much throughout the duration of the test. The temperature was
maintained within ± 1°C using a hot plate and a thermocouple with feedback control.

The corrosion process was studied using electrochemical techniques including linear polarization
resistance (LPR), potentiodynamic sweeps (PDS) and electrochemical impedance spectroscopy (EIS).
Electrochemical measurements were made using a Gamry† potentiostat under computer control.

Figure 1: Three-electrode glass cell apparatus.

†
Trade Name
# American Petroleum Institute (API), 1220 L St. NW, Washington, DC 20005.

2
Material

Working electrode
X-65 mild steel, a typical pipeline material was used for all corrosion tests. The chemical composition of
the steel is given in Table 1. The exposed area of the WE was 8.00 cm2.

Table 1
Chemical composition of X65 mild steel (mass % balance is Fe) [9]
C Mn Si P S Cr Cu Ni Mo Al
X65 0.065 1.54 0.25 0.013 0.001 0.05 0.04 0.04 0.007 0.041

Solid deposit
Sand and SiO2 powder were used as solid deposits in this work. Both of the deposits are the same
base inorganic material, i.e. silica (SiO2) sand. The main difference between them is their grain size
and shape as can be seen in the SEM images shown in Figure 2. The properties of these deposits are
listed in Table 2. The density and porosity of each type of deposit material was experimentally
determined as previously reported.[10]

Sand SiO2 powder

Figure 2: SEM images of deposit materials.

Table 2
Deposit properties

Grain size Bulk density* Porosity#

SiO2 powder 44 µm 0.75 g/cm3 75%

sand 240 µm 2.5 g/cm3 39%

* The mass of many particles of the material divided by the total volume they occupy.
#
The percentage of the volume of voids in a material composed of particles to the total volume
the particles occupy.

Each deposit material was rinsed with DI water stored in a CO2 purged solution similar to the test
solution and transferred by pipette onto the tested sample shown in Figure 3.

3
 Figure 3: Corrosion sample and sample holder.

Procedure

In each experiment, CO2 was bubbled through the electrolyte for a minimum of 60 minutes to saturate
and deaerate the test solution.

The surface of the WE was sequentially polished using 200-, 400- and 600-grit silicon carbide paper,
washed and rinsed with isopropyl alcohol before each experiment, mounted in the specimen holder and
immersed into the electrolyte. The free corrosion potential measurement began immediately after the
immersion and normally stabilized within ± 1 mV after 5 - 10 minutes.

Baseline “bare” steel corrosion tests were conducted at the same conditions as the under deposit
corrosion tests. For each under deposit corrosion test, the test sample was immersed in the electrolyte
for one hour without a deposit, where the free corrosion rate and solution resistance were measured by
LPR and EIS respectively. The deposit was then introduced onto the WE steel surface and the test was
further carried out for 24 hours.

The WE was polarized ±5 mV from open circuit potential (Eoc) during the LPR measurements at a scan
rate of 0.125 mV/s to obtain the polarization resistance (Rp) which was then used to calculate the
general corrosion rate of mild steel. For the EIS measurement, a sinusoidal potential signal ±5 mV
peak-to-peak around Eoc was applied to the WE with scanning frequencies from 5 kHz to 1 mHz. This
scan enables the identification of both the solution resistance (RS) at the highest frequency range and a
“fingerprint” of the various reactions involved in the corrosion process at the lower frequency ranges.
The potentiodynamic sweeps were conducted at the end of each experiment by sweeping the potential
at a scan rate of 0.125 mV/s in the cathodic (more negative) direction from Eoc to Eoc-0.4 V, and then
disconnecting and pausing until a stable Eoc was attained, then the potential of the WE was swept in the
anodic (more positive) direction from Eoc to Eoc+0.2 V using the same scan rate. Experimental
conditions are summarized in Table 3. After each test, the WE surface morphology and composition
were analyzed by SEM and EDX.

4
 Table 3
Experimental conditions
Parameter Conditions

Test material X65 mild steel

Test solution DI water + 1wt% NaCl

Temperature 25°C - 80°C

CO2 partial pressure 0.54 bar - 0.96 bar

Solution pH 4~6

Deposit Silica sand, SiO2 powder

Deposit thickness 2 ~ 10 mm

Test duration 24 hour

Sweep rate 0.125 mV/s

Polarization resistance From -5 mV to 5 mV vs. Eoc

EIS ±5 mV vs. Eoc from 5 kHz to 1 mHz

Potentiodynamic sweep From -400 mV to +200 mV vs. Eoc

ELECTROCHEMICAL MODEL

The experimental observations led to the development of a mechanistic model for uniform CO2
corrosion under inert solid deposits, first requiring consideration of the cathodic and anodic reactions
involved.

Cathodic reactions

Initially CO2 is hydrated to form carbonic acid in water:

CO2 + H 2O ⇔ H 2CO3

These species can then undergo partial dissociation:

H 2CO3 ⇔ H + + HCO3−

HCO3− ⇔ H + + CO32 −

In CO2 corrosion at a pH range of 4 to 6, the presence of CO2 leads to a much higher corrosion rate
[
than would be found in a solution of a strong acid at the same pH. 9] This is because the presence of
dissolved CO2 acts as a “reservoir” of carbonic acid and can help with replacement of consumed

5
hydrogen ions. Water reduction is also considered as one of the potential cathodic reactions in the
present model. The three cathodic reactions considered in the model are:

H + + e− → H
H 2CO3 + e − → H + HCO3−

H 2O + e − → H + OH −

H+ reduction
Generated current from hydrogen reduction ( H + + e − → H ) is calculated by the following equation[9]:

1 1 1
= + d
i( H + ) iα ( H + ) ilim (H + )
(1)

where charge transfer current density ( iα ( H + ) ) is:

η
−
iα ( H + ) = ε ⋅ i0( H + ) × 10 bc

(2)
i0( H + ) ΔH ( H + ) 1 1
ref
= exp(− ( − ))
i0( H + ) R T Tref (3)

RT
bc = 2.303
0.5F (4)
η = E − Erev
(5)
Tref = 25°C. R is gas constant R = 8.315 J/mol/K. F is Faraday constant, F = 96485 C/equiv.

Diffusion limiting current density is usually defined as:

d
ilim (H + )
= km ,eff F [ H + ]b
(6)
where km ,eff is the mass-transfer coefficient and is calculated to take into account the flow effect on
diffusion of species through the hydrodynamic boundary layer[9]. However, in this work, all tests were
conducted in stagnant condition with the diffusion of species impeded primarily by the presence of inert
solid deposit [10]. In the first approximation, the permeability ( κ ) of the surface deposit layer for
transport of species depends on the superficial porosity ε and the shape and connectivity between the
pores, expressed via the tortuosityψ :

κ = ψε (7)

6
It was found that the superficial porosity can be approximated with volumetric porosity [9] and the
tortuosity can be related to porosity, so in the in the present work we can replace the mass transfer
coefficient with:

DH + ⋅ ε 1.3
km,eff =
L (8)

where L is deposit thickness.

[
Other parameters in the above equations are: 12]

Symmetry factor (at 25oC):

α c = 0.5

Exchange current density:

∂ log i0( H + )
= −0.5
∂pH (9)
Reversible potential:
2.303RT 2.303RT
Erev ( H + ) = − pH − log pH 2
F 2F (10)

H2CO3 reduction
The current density contributed by H2CO3 reduction is derived in a similar way as for H+ reduction:

1 1 1
= + r
i( H2CO3 ) iα ( H2CO3 ) i lim( H 2CO3 )
(11)
η
−
iα ( H 2CO3 ) = ε ⋅ i0( H 2CO3 ) × 10 bc
(12)

r
At stagnant conditions, the chemical reaction limiting current density ilim( H 2 CO3 ) is related to diffusion

coefficient[9] and is found to be related to deposit thickness and effective diffusion coefficient as:
−1 0.5
H 2CO3 ) = F ⋅ CCO2 ⋅ ( K hyd k hyd Deff L )
r f
ilim(
(13)

where the equilibrium constant for the CO2 hydration reaction is K hyd = 2.58 × 10−3 [9]. khyd
f
is the
forward reaction rate for the CO2 hydration reaction. It is a function of temperature, and is calculated by
using the following equation[11]:

11.715
169.2 −53.0log T −
f
khyd = 10 T
(14)

7
The effect of the deposit is expressed in the effective diffusion coefficient as:

Deff = D ⋅ ε 1.3
(15)
and
CCO2 ,b = k CO
d
× pCO2
2
(16)

Where k CO
d
is Henry’s constant and is a function of temperature:[12]
2

d
kCO2
= 0.0454(1.6616 − 5.736 ×10−2 t + 1.031×10−3 t 2
−9.68 ×10−6 t 3 + 4.471×10−8 t 4 − 7.912 ×10−11 t 5 ) (17)

H2O reduction
Since water is considered to be present in “unlimited” quantities at the metal surface, it is then assumed
that water reduction is limited by a charge-transfer process and the Tafel equation applies:

η
−
iH 2O = ε ⋅ i0, H 2O ×10 bc
(18)

It was determined experimentally[10] that the exchange current density for water reduction is
i0, H 2O ≈ 3 ×10−5 A/m2 .

Anodic reaction

Iron dissolution in water is the only reaction considered for the anodic contribution:

Fe → Fe 2 + + 2e −

From experimental observation[12] iron dissolution is under charge transfer control:

iFe = ε ⋅ i0, Fe ×10 ba

(19)

where,
i0, Fe ΔH ( Fe ) 1 1
= exp(− ( − ))
i0, Fe ref
R T Tref (20)

Tref = 20D C , ΔH ( Fe ) = 40KJ/mol i0, Fe ref = 1 A/m 2 and ba = 40 mV[12].

RESULTS

8
 [
Detailed experimental results have been presented and discussed in a previous publication 10]. It was
concluded there that the effect of inert solid deposit on CO2 corrosion of mild steel was a combination
of a surface coverage effect and a mass transfer effect. The surface coverage effect is due to the direct
contact of sand with the metal surface reducing the amount of metal surface area available for
corrosion, thus reducing the magnitude of both anodic and cathodic reactions. Available metal surface
area was directly proportional to deposit porosity. The mass transfer effect was related to the depth of
the deposit and the porosity and tortuosity of the path for diffusion of species to the steel surface.
These results will be compared below with the predictions made by the model described above.

MODEL VERSUS EXPERIMENTS

The underdeposit CO2 corrosion was first tested for a bare steel corrosion condition. Figure 4 is the
comparison of a potentiodynamic sweep predicted from the present model with experimental data. It
can be seen that the anodic and cathodic portions of the potentiodynamic sweep capture the corrosion
processes very well and the prediction is in good agreement with the experiment.

-0.2
-0.3 Bare steel CO2 corrosion
E vs Saturated Ag/AgCl / V

-0.4 pH 5.0, 25°C

-0.5
-0.6
-0.7
Exp. sweep
-0.8
-0.9 Pred. sweep

-1
-1.1
H2CO3 reduction
-1.2
0.01 0.1 1 10
i / (A/m2)

Figure 4: CO2 corrosion of bare X65 steel at pH 5, 25°C, 1 wt% NaCl solution, pCO = 0.96 bar.
2

The effect of deposit porosity

In Figure 5, anodic and cathodic potentiodynamic sweeps conducted on X65 mild steel covered by a 2
mm deposit of either SiO2 powder or sand are contrasted with those conducted on bare steel at bulk
solution pH 5 and 25°C. Experimental results are in good agreement with the predictions of individual
reactions generated with the present model. Both anodic and cathodic reactions are diminished with the
decrease in deposit porosity. The relatively small decrease in the anodic current stems from the
decrease in the available metal surface (by coverage). The decrease in the cathodic current is
predominantly because of the increased difficulty in the diffusion of corrosive species involved in the
cathodic reactions.

9
 ‐0.4

‐0.5

‐0.6
E vs. Ag/AgCl / V

‐0.7

‐0.8

‐0.9
Exp. Bare steel Pred. Bare steel
‐1 Exp. SiO2 (porosity 75%) Pred. SiO2 (porosity 75%)
Exp. Sand (porosity 39%) Pred. Sand (porosity 39%)
‐1.1
0.01 0.1 1 10
i / A/m2
Figure 5: Effect of deposit porosity on CO2 corrosion of X65 steel at pH 5, 25 °C, 1 wt% NaCl solution,
pCO2 = 0.96 bar, 2 mm deposit.

The effect of deposit thickness

Anodic and cathodic potentiodynamic sweeps conducted for an X65 mild steel surface covered with
different sand depths are compared with those conducted on bare steel at bulk solution pH 5 and 25°C.
With the change in the deposit thickness, the anodic current did not change substantially as can be
seen in Figure 6. This is because increasing the deposit thickness only increases the mass transfer
resistance of corrosive species for cathodic reactions. The measured potentiodynamic sweeps are in
good agreements with the prediction. The thicker the deposit is the more tortuous the path of diffusion
for species to go through, which results in a smaller cathodic current.

10
 ‐0.4

‐0.5

‐0.6
E vs. Ag/AgCl / V

‐0.7

‐0.8

‐0.9
Bare steel Pred. Bare steel
‐1 Exp. 2mm Sand Pred. 2mm Sand
Exp. 5mm Sand Pred. 5mm Sand
‐1.1
Exp. 10mm Sand Pred. 10mm Sand
‐1.2
0.01 0.1 1 10
i / A/m2
Figure 6: Deposit thickness effect at pH 5, 25 °C, 1 wt% NaCl solution, pCO = 0.96 bar, sand porosity =
2

39%.

Temperature effect

Anodic and cathodic potentiodynamic sweeps on X65 mild steel covered with a 10 mm sand deposit
were conducted at 25°C and 80°C (Figure 7). The corrosion rate of mild steel under this type of deposit
did not increase with increasing temperature as would be observed in bare steel corrosion. However,
some iron carbonate crystals were observed on the metal surface under the deposit which was not
seen in bare steel corrosion.

11
 ‐0.4

‐0.5

‐0.6
E vs. Ag/AgCl / V

‐0.7

‐0.8
Exp.data at 25°C
‐0.9 Pred.data at 25°C
Exp.data at 80°C
‐1 Pred.data at 80°C

‐1.1
0.01 0.1 1 10
i / A/m2
Figure 7: Temperature effect at pH 5, 1 wt% NaCl solution, 25 °C and 80 °C, 10 mm sand deposit (porosity
= 39%).

Corrosion rate prediction

The general corrosion rates predicted by the present model were compared with experimental
measurements using LPR. Figure 8 compares the corrosion rate for different porosities, and Figure 9 is
for different deposit thickness. The B value used in calculating corrosion rate from LPR data was 13
mV. The prediction of corrosion rate at 25°C is in good agreement with the experimental results.
However, the one at 80°C differs from the one measured from LPR, see Figure 10. The difference
could be due to the use of an inappropriate B value in the LPR calculation at 80°C and due to formation
of iron carbonate which was not accounted for in the model.

12
 1
0.9
0.8
Corrosion rate / mm/y

LPR
0.7
Prediction
0.6
0.5
0.4
0.3
0.2
0.1
0
bare steel SiO2 (75%) Sand (39%)

Figure 8: Comparison of predicted corrosion rate with experiment data. Experiments are conducted at pH
5, 25 °C, 1 wt% NaCl solution, pCO2 = 0.96 bar, 2 mm deposit (porosities 75% & 39%).

1
0.9
0.8
Corrosion rate / mm/y

0.7 LPR
0.6 Prediction
0.5
0.4
0.3
0.2
0.1
0
0 2 5 10
Deposit thickness L / mm
Figure 9: Comparison of predicted corrosion rate with experiment data. Experiments are conducted at pH
5, 25 °C, 1 wt% NaCl solution, pCO = 0.96 bar, sand deposit porosity 39%.
2

13
 1
0.9
0.8
Corrosion rate / mm/y

0.7 LPR
0.6 Prediction
0.5
0.4
0.3
0.2
0.1
0
25 80
Temperature T / °C
Figure 10
Figure 10: Comparison of predicted corrosion rate with experiment data at different temperature. pH 5, 25
°C, 1 wt% NaCl solution, pCO2 = 0.96 bar, 10 mm sand deposit (porosity 39%).

CONCLUSIONS

Based on the results discussed in this work, the following conclusions were made:
1. The general corrosion rate of mild steel decreased by a factor of 3 to 5 after sand deposit was
introduced, both at 25°C and 80°C;
2. Both anodic and cathodic current are decreased because of less active surface area availability due
to deposit coverage.
3. The inert solid deposit creates a mass transfer barrier for corrosive species and limits cathodic
reactions.
4. The extension of a simple mechanistic model can be used capture this behavior and predict the
effect of a solid deposit on CO2 corrosion of mild steel reasonably well.

ACKNOWLEDGEMENTS

The authors would like to thank the following companies for financial support: Baker Petrolite, BG
Group, BP, Champion Technologies, Chevron, Clariant Oil Services, ConocoPhillips, Encana, ENI
S.p.A., ExxonMobil, WGIM, Total, NALCO Energy Services, Occidental Petroleum Co., Petrobras,
PETRONAS, PTT, Saudi Aramco, Inpex Corporation. Advice and comments from Professor Brian
Kinsella is also greatly appreciated.

14
 NOMENCLATURE

ba, bc anodic and cathodic Tafel slope, mV/decade

Cco2 bulk concentration of dissolved carbon dioxide, mol/m3
D diffusion coefficient, m2/s
Erev reversible potential, V
F Faradaic constant, =96485 coul/equiv.
[H+]b [H+]s bulk and surface concentration of H+ (mol/m3)
i current density, A/m2
ia activation component of the total current density, A/m2
i0 exchange current density, A/m2
idlim diffusion limiting current density, A/m2
Irlim chemical reaction limiting current density, A/m2
kdCO2 Henry’s constant, mol/m3/bar
km mass transfer coefficient, m/s
L deposit thickness, m
PCO2 partial pressure of CO2 gas, bar
R universal gas constant, 8.314J/mol/K
t temperature, °C
T Kevin temperature, K
α a , α c, apparent transfer coefficients
µ viscosity, kg/m/s
κ permeability
ψ tortuosity
ρ Porosity
η = E – Erev, overpotential, V

REFERENCES

[1] H. D. Reus, E. L. J. A. Hendriksen, and M. Wilms, “Test methodologies and field verification of
corrosion inhibitors to address under deposit corrosion in oil and gas production systems”,
CORROSION/05, paper no. 288 (Houston, TX: NACE, 2005).
[2] D. I. Horsup, J. C. Clark, B. P. Binks, P. D. I. Fletcher, J. T. Hicks, “I put it in, but where does it go?-
the fate of corrosion inhibitors in multiphase systems”, CORROSION/07, paper no. 617 (Houston,
TX: 2007).
[3] D. H. Pope, B. Cookingham, R. Day, G. Pogemiller, “Mitigation strategies for microbiologically
influenced corrosion in gas industry facilities”, CORROSION/89, paper no. 192 (Houston, TX:
NACE, 1989)
[4] M. Nordsveen, S. Nesic, R. Nyborg, and A. Stangeland, “A mechanistic model for carbon dioxide
corrosion of mild steel in the presence of protective iron carbonate films – part 1: theory and
verification”, Corrosion 59, 5 (2003): pp. 442 - 456A.
[5] A. Pedersen, K. Bilkova, E. Gulbrandsen, and J. Kvarekval, “CO2 corrosion inhibitor performance in
the presence of solids: test method development”, CORROSION/08, paper no. 632 (Houston, TX:
NACE 2008).
[6] G. Hinds, S. Zhou, Al. Turnbull and P. Cooling, “A multi-electrode approach to evaluating inhibition
of underdeposit corrosion in CO2 environments”, CORROSION/09, paper no. 445 (Houston, TX:
NACE, 2009).

15
[7] K. Lepkova and R. Gubner, “Development of standard test method for investigation of under-
deposit corrosion in carbon dioxide environment and its application in oil and gas industry”,
CORROSION/10, paper no. 331 (Houston, TX: NACE, 2010).
[8] M. Achour, J. Kolts, P. Humble, R. Hudgins, “Experimental evaluation of corrosion Inhibitor
performance in presence of iron sulfide in CO2/H2S environment”, CORROSION/08, paper no. 344
(Houston, TX: NACE, 2008).
[9] S. Nesic, J. Postlethwaite, and S. Olsen, “An electrochemical model for prediction of corrosion of
mild steel in aqueous carbon dioxide solutions”, Corrosion Science 52, 4 (1996): pp. 280-294
[10] J. Huang, B. Brown, X. Jiang, B. Kinsella and S. Nesic, “Internal CO2 corrosion of mild steel
pipelines under inert solid deposits”, CORROSION/10, paper no. 379 (Houston, TX: NACE, 2010)
[11] D. Stachewski, Chemie-Ing.-Techn., Vol. 41(1969), pp. 1.111
[12] J. O’M. Bockrts, D. Draztc, A. R. Desptc, “The electrode kinetics of the deposition and
dissolution of iron”, Electrochimica Acta 4, 4 (1961) : p 325 - 361

16