Paper No.

05476
CORROSION 2005
MODELING OF H2S CORROSION
BY COUPLING OF PHASE AND POLARIZATION BEHAVIOR

Hans Hoffmeister
Institute for Failure Analysis and Failure Prevention ISSV e.V.
c/o University of the Federal Armed Forces
D-22039 Hamburg
Germany

ABSTRACT

From experimental evaluation of oil and gas field materials a synergy between the formation of dark
sulfide layers and hydrogen supported sulfide stress cracking is frequently reported in literature. The
present work describes a first approach to a deterministic H2S-corrosion model for calculation of
precipitation of FeS2 together with respective changes of pH at the anodic sites of the assumed
corrosion system for pure Fe. The model is based on coupling the calculated anodic polarization
curves to the precipitated equilibrium masses of Fe304 and FeS2 which in a "closed loop" time
stepwise procedure are calculated from the solute concentrations. The results show that as a
consequence of FeS2 precipitation the local pH at the anodic site is reduced together with the changes
of total concentrations of HS-, Fe ++ and H ÷ in the assumed diffusion layer. Also, the respective
changes in corrosion currents and potentials are demonstrated. With increasing bulk pH and total
pressures the acidification times from the start of the process to a local pH of 2.5 (t pn2.5) increase
while the respective mean corrosion currents are reduced. At higher pH levels increasing H2S -
contents lead to significant reductions in acidification times as well as corrosion currents. This effect
is however smaller at lower total pressures. The calculated pH reductions follow a similar course of
experimentally measured pH-levels during precipitation of iron sulfides in a 0.5 bar H2S-5g/1
NazSO4-solution at galvanostatic loading with 0.8 mA/cm2 in a closed cell. As a summarizing result
an example of a pH-vol.-% H2S domain diagram for constant total pressures, pH-reduction times is
established and discussed with respect to effects of anodically accumulated hydrogen on cracking
processes.

Keywords: HaS corrosion modeling, sulfide layers, polarization curves, coupling principle
sulfide stress cracking
Copyright
2001 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 77084-4906. The material presented and the views expressed in this paper are
solely those of the author(s) and not necessarily endorsed by the Association. Printed in U.S.A.

Publication Right
Government work published by NACE International with permission of the author(s). Requests for permission to publish this manuscript in any form, in part
or in whole must be made in writing to NACE International, Publications Division, 1440 South Creek Drive, Houston, Texas 77084-4906. The material
presented and the views expressed in this paper are solely those of the author(s) and not necessarily endorsed by the Association. Printed in U.S.A.

1
 INTRODUCTION

The effect of H2S on operability and life time of oil and gas production equipment has been
focussed on in failure as well as laboratory and modeling investigations 1-5. Common findings are that
H2S influences general corrosion rates and, by providing hydrogen pick up may sensitize the
materials to stress or strain induced sulfide stress cracking, (SSC). While the majority of the
investigations refer to carbon manganese steels also SSC of alloyed steels has been reported.

Frequently, 2,4 n2s induced reduction of ductility was observed to occur simultaneously with the
formation of dark sulfide scales, during general corrosion or, in particular, localized pitting. So far, no
attempts have been made to explain this quantitatively, although, from 6 it can be drawn that for
example, FeS2 is in thermodynamic equilibrium with aqueous solutions of pH as low as 1.

In further H2S- corrosion studies it was stated that:

• various stoichiometric compositions of iron sulfides may be formed depending on pH, ~2s and
time of exposition 7-9,
• sulfides are formed above very low levels of PH2s, 10-12
• the formation or dissolution of the sulfides control the corrosion rates, 7-10
• corrosion rates may depend on conductivity of the sulfides, 8
• the ohmic resistance of the sulfide depends on pH, 8
• corrosion rates may increase or decrease with increasing partial pressure of H2S,by affecting the
cathodic and anodic processes, 7.

In 1999 a significant first contribution to simulation of combined C O 2 - a n d H2S corrosion by

A.Anderko and R.D.Young 7 applies thermodynamic and electrochemical models. The calculated
corrosion results were found to correlate sufficiently with experimental findings regarding the effects
of partial pressures of CO2, H2S, of total pressure and temperature. However, the above mentioned
properties of the sulfide layers were so far not yet integrated.

In 2004, R.Hausler s summarizes previous experimental investigations but mentions, that it would be
too early for establishment of a model showing the various aspects of H2S-corrosion.

The effect of carbonate scaling on CO2-corrosion has been modeled extensively by S.Nesic and
coworkers 12 by application of similar procedures as A.Anderko and R.D.Young 7. However, the
specific effects of the physical properties of the scaling phase including their mass could not be
implemented in order to show the respective influence on polarization behavior.

Such coupling between phase properties and polarization behavior has recently been applied by
H,Hoffmeister 13,14 in modeling crevice corrosion of chromium and nickel. As a basic procedure
the phase masses which precipitate from the solution compositions are calculated from the
constitutional phase diagram w a t e r - m e t a l - metal hydroxide-metal chloride. By their electrical
resistance, these masses are then coupled to the slopes of the anodic polarization curves inside the
crevice thus establishing a time stepwise calculation loop for both the electrochemical and the
chemical crevice reaction.

As a consequence of the calculated crevice corrosion process, it was demonstrated that the often
discussed acidification during crevice corrosion starts after the complete dissolution of the buffering
metal hydroxide in equilibrium with metal chloride, metal and water. The resulting local

2
pH levels would decrease significantly and thus may provide potentials for local hydrogen induced
cracking at the end of the crevice.

In the present work, the same principle as in the previous papers will be applied to H2S-corrosion
including the change in local pH.

MODEL PRESENTATION

The corrosion system according to Figure 1 visualizes the pitting corrosion situation mentioned
frequently together with hydrogen sulfide cracking. It consists of separate cathodic and anodic
concentric surfaces of pure iron, each one of l m m 2 surface area. The following simplifying
assumptions are made:

• at the start of the process, the anodic surface is active, i.e., no solid phase except Fe is present,
• the cathodic reaction is controlled by diffusion of i-I+ ions into the diffusion layer with a given
thickness and their reaction with electrons provided by oxidation of Fe at the anodic surface,
• at the anodic site Fe ++ ions dissolve into the diffusion layer but do not diffuse into the bulk
solution,
• Fe ++ ions may form Fe304 as a passivating phase,
• HzS from the gas phase dissolves into the bulk solution and dissociates providing HS- and I-I+
ions,
• H S -ions migrate to the anodic surface providing charge balance with the iron ions,
• Fe ++ and H S ions may form various types of iron sulfides,
• for the present work, the formation of FeS2 is considered, following observations at actual
failures in 1
• the solid nonmetallic phases increase the anodic polarization resistance according to their
precipitated mass, their specific resistance and their density-(coupling principle)-,
• all reactions except the cathodic hydrogen reduction are at equilibrium,
• no substances are exchanged between the cathodic and anodic sites.

Chemical Reactions and Determination of Phase Masses at the Anodic Site:

The following Nernst equations are drawn from is 6 and are applicable to 25°C *

At the anodic surface Fe reacts according to:

Fe = Fe ++ + 2 e (1)
EoF~ = - 0.44 + 0.0295 log Fe ++ (la)

and may be at equilibrium with Fe304 and/or FeS2 in the aqueous solution of the diffusion layer
according to
3 Fe ++ + 4 H20 = Fe304 + 8 H + + 2 e (2)
0.2364 pH - 0.0886 log Fe ++
EoFe304 ---- 0 . 9 8 0 -- (2a)

and Fe ++ + 2 H2S = FeS2 + 4 I-I+ + 2 e (3)

EoFeS2 = -0,066 - 0,118 pH - 0,029 log Fe ++ - 0,059Iog H2S (3a)

3
H2S from the gaseous phase determines the H S concentration in the aqueous phase according to:

(H/S)ga = (H2S)aq (4)

log(H2S / PH2S)= - 0,99 (4a)

(n2S)aq- (ns') + n + (5)

log(HS / H2S) = - 7 - pH (5a)

For determination of phase masses the F e 3 0 4 - and F e S 2 - solubilities within the ternary equilibriae
with metallic Fe are required in terms of mass-% which are drawn from equations la-3a and equation
4a.

Fe- Fe304 - H20:

By combination of equations l a and 2a the solubility of Fe304 together with that of metallic Fe in
sulfide free aqueous solutions at RT with activity coefficient 3'Fe++is:

"~Fe++* o~ Fe ++ / (% H +)2= 10 14.72 (6)

Fe- FeS2- H20:

Combining equation l a, 3a, 4a and 5a it follows:

% HS- • ' ~ + ,%Fe ++ / %H + = 10 2.19 (7)

Fe - FesO4- FeS2 - H20:

Not considering an effect on Fe304 solubility of H S - ions, the combination of equations 5 and 6
give:
% W • %HS = 10-12.5 (8)

representing the solubility conditions for a solution being saturated by Fe as well as by Fe304 and
FeS2.

Figure 2 visualizes the respective solubility surfaces of Fe304 and FeS2 within the quaternary
equilibrium with metallic Fe and H20. (The solubility surface of Fe is not shown). It is obvious that
FeS2 precipitates already at low concentrations of Fe ++ and H +. For example, at pH 7 and 0.01% Fe ++
the transition from Fe304 to FeS2 would take place in the presence of 0.000667 % H S corresponding
to 0.00195 bar partial pressure of H2S. Taking these Fe ++ - and pH levels as practical lower limit
values of formation waters the precipitation of Fe304 was so far not considered in the present work.

For determination of FeS2 masses mFes2 the total contents Fet++, Ht+ and HSt and their equilibrium
solute contents F e~++H
, ~+ and HS~ together with the chemical composition of FeS2 (46.5% Fe, 53.5%
S) provide the mass balance for example according to

mves2 = M ,(%H + - %Ht+) / %H + (9)

4
Calculation of Hs+ is carried out by combining equation (8) with equations 10a,b representing the tie
lines of the ternary equilibrium FeS2 -Fe- H20:

%HSs -= 53.5 -%H~+, (53.5 - %HSt-) / %Ht + (10a)

%Fe~+ + = 46.5
o _0
%Hs+ ,(46.5 _
%F~ ++) / %Ht + (10b)

It follows that:

%Hs++=
(((4976- 53.5,%Fet ++ - 46.5,%HSt)/%Ht +- 153.5) + (((4976- 53.5,%Fet ++- 46.5,%HSt)/%Ht +
- 153.5) 2- (9952,(2488-53.5*%Fet ++ -46.5,%HSt- + %Fet ÷ ,%HSt) / (%Ht+)2)) 0.5) / (2,(2488-
•53.5,%Fet ++- 46.5,%HSt- + %F~++,%HSt)/(%Ht+) 2) (11)

Thus, the solute hydrogen ion concentration %H~++ and pH can be calculated from the total
concentrations of the ions of iron, sulfide and hydrogen. These total compositions will be calculated
from the electrochemical process as described below.

Electrochemical Reactions:

At the anodic side the concentration changes of the assumed surface layer of d = 0.001 mm thickness
are determined by 1 second time step calculations of corrosion currents ic from Tafel equations and
total Fe ++ ion concentrations from Faraday's law. For considering activity coefficients of the Fe ++ -
ions, equation 12 is drawn from literature data given in 16 and 17.

YFe++= 1/4.7,1og %Fe++/(-2.07)+0.15 (12)

As outlined above, no diffusion to the bulk solution is assumed. However, H S - ions are assumed to
migrate into the surface layer for balance of the positive surplus charge of the Fe++-ions. Thus, the
mass mns. of the HSt-ions follows the mass rnFe++of the Fe ++ - ions immediately according to:

m us- = 1.89 • m Fe++ (13)

The anodic polarization curve (Tafel-slope) is determined by the exchange current ioFe, the actual
Nernst potentials EoFe of the Fe++/Fe reaction (equation l a) and the inverse anodic polarization
resistance b~ which depends on the mass and the physical properties of precipitated phase 13,14
according to:
ba = 1/(1/bal + mFes2* p / Ac 2 * "~FeS2 ) (14)

bal -- 0.017 is the initial anodic Tafel slope without precipitated phase, taken from 16
~¢FeS2 " density = 4.7 mg/mm 3 from 18
p" specific resistance of passive layer, assumed 7000 mV,mm/log mA following 19
& " anodic surface area = 1 mm 2
ioFe" exchange current = 0,00007 mA as taken from 16

A detailed derivation of equation 14 is given in 13. It does not consider the resistivity of the aqueous
solution.

5
At the cathodic site hydrogen ions diffuse into the surface layer and are reduced by electrons mainly
provided by the above described anodic reaction. The cathodic polarization curve is controlled by the
Nernst potential Eon of the H/H + - reaction, equation 15,15a:

2H + + 2 e = H2 (15)

EoH -- - 0.0591,pH - 0.0293,1ogPH2 (15a)

together with the exchange current ioH = 1.17E-07 mA 16 and the inverse cathodic polarization
resistance bk (Tafel-slope) which is calculated according to:

bk =( log ie- log ion)/(Eon - Ec) (16)

The corrosion potential E¢ in equation 16 is calculated according to equation 17 taken from 20

considering hydrogen ion diffusion into the diffusion layer of thickness d with
diffusion coefficient DH = 0.01 mm2/s from 12:

E~ = 59.1,1og(10-PH/PH2 -t- ie*d/PH2*F*DH+) (mV/SHE) (17)

Following the above outlined local corrosion process, the cathodic site was additionally subjected to a
small passivation current. The respective constant anodic Tafel slope was assumed to b,o = 0.0006 log
mA/mV, with the exchange current ioF~ and a constant Nemst potential EoF~ which is determined by
the bulk solution Fe ++ - ion contents. For consideration of the passive current, the initial corrosion
potential Eci at the start of the corrosion process was implemented into equation 18 which provides
the net corrosion currents i¢ for each time step of 1 second of the calculation:

½ = 10A(1og ((log ioFe * bl, + (bao*(E~i-EoF~)+logioF~)*ba

+ bk, ba* (Eei-Eove))/( ba +bk))) (18)
The mixed corrosion potentials are finally determined by:

Ee=((bk,EoH)-logic+logion)/bk (mV/SHE) (19)

The following calculations were carried out for various levels of pH, H2S contents of the gas phase
and various total pressures which correspond to the partial pressures of gaseous hydrogen. The bulk -
Fe ++ concentration was 0,01%, the temperature 25°C.

RESULTS AND DISCUSSION

Corrosion Mechanism

As an example, for pH = 7, P= 15 bar and PH2S = 0.15 bar the corrosion mechanism is demonstrated
in Figures 3-4. Figure 3 shows the time dependant concentration increase of Fe ++ and H S of the
anodic surface layer together with the resulting decrease of the total H ÷ contents by dilution of the
initial content. The total concentration changes during the applied time steps are visualized in Figure
4 in terms of the H+ - and HS- - ion contents relative to the projected quaternary phase diagram 1-120 -
HS- - Fe ++ - H ÷. Starting with the initial concentration A, the layer solution is saturated only with Fe
metal. The precipitation of FeS2 initiates when the concentration path crosses the saturation limit for

6
FeS2 at B. The increasing masses of FeS2 are now calculated by the mass balance (equation 9) within
the three phase domain H20 - FeS2- Fe. With the total composition of point C the line CD represents
the equilibrium tie line between the aqueous solution of composition D and the solid phase FeS2. The
ratio of the distance C-D to the distance D- FeS2 is equivalent to the mass of FeS2 relative to the total
mass of the surface layer of 1 mm 2 area and 0.001 mm thickness.

It is stringent that with increasing total HS- - also the I-i+ - concentrations in solution are increasing in
direction of the arrow in Fig.4 because the respective tie lines are shifting to higher H+-contents of
point D. Basically, this stems from the transport of not only sulfur but also of hydrogen to the surface
layer.

The resulting increase of FeS2- masses and the coherent decrease of local pH in the surface layer are
demonstrated in Figures 5 and 6. In the present example, the calculated local pH decreases only
slightly from 7 to about 6.4 within 1500 s. The final rapid pH decrease is enforced by the f'mal
enrichment of Fe ++- and HS-- ions close to the sulfur - and iron contents of the precipitating FeS2.

In practice those finally low pH-levels may not be acquired due to diffusion effects, which is
demonstrated from experimental galvanostatic test results with a partially closed off pH- and iron
working electrode in Figure 7. It was therefore decided, to let the simulated process end at pH 2.5.
However, the frequently mentioned acidification in occluded cells (pits) 4 is demonstrated by the
simulation as well as by the experiment.

In Figure 8 the corrosion currents of the selected example increase with time but level off after the
end of the selected process time. The parallel corrosion potential exhibits a similar behavior as a
result of the simulated corrosion process. It should be mentioned that with other combinations of pH,
PH2Sand total pressure P the initial current increase is shifted to very short times followed by a rapid
decrease, also of local pH, Figure 9. Such behavior has been experimentally observed by Hausler 8
who correlated the corrosion rates to iron sulfide film growth rates.

Effect ofpH, total pressure P and H2S on anodic acidification time

To demonstrate the effects of the above mentioned parameters on anodic acidification together with
FeS2 precipitation the characteristic "process end times" for arriving at pH 2.5, tpaz..5 together with
the mean corrosion currents imeanduring this process time were selected.

Figure 10 demonstrates the increase of acidification times with pH and total pressure. At the same
time, the mean corrosion currents during the acidification time are decreasing with pH and total
pressure. It is obvious, that at pH below 6 the pH- and pressure effects on acidification times are less
pronounced as compared to higher pH levels while a reverse behavior is observed for the mean
corrosion currents.

The effect of H2S - contents is visualized in Figure 11, with the upper pH levels of Figure 10. With
increasing H2S from 1 to 10 vol.% the acidification times for 15 bar total pressure are more reduced
the higher the pH. At 7.5 bar total pressure, this reduction by HES is smaller, while at 1.5 bar, the
effect was still present from the calculations, but could not be shown in the diagrams.

In Figure 11, also the mean corrosion currents are reduced by increasing H2S contents at higher pH-
and total pressure levels.

It has to be emphasized that at lower pH levels the effect of H2S is much less pronounced and thus
can also not be visualized.

7
The results demonstrate that the three investigated parameters pH, pressure and H2S provide
combined effects on acidification and corrosion currents. For example, pH influences the anodic
solubilities as well as cathodic polarization.

In the available literature, the effect of H2S contents is frequently characterized by an initial decrease
and subsequent increase in corrosion rates ~0. K.L.J.Lee and S. Nesic 2~ mention that in their EIS
investigations sulfide films reduce corrosion rates at low H2S-concentrations up to 15 ppm while at
concentrations up to 250 pp9m the corrosion rates were slightly increased. Also, S.Ramachandran,
M.B.Ward and K.A.Bartrip restate the f'mdings of J.B.Sardisco, W.B.Wright and E.Greco 22 that the
corrosion rates increase with increasing H2S.

A different approach to explanation of H2S corrosion has been published by P.Marcus, 23 based on
Langrnuir adsorption mechanisms of elemental sulfur to the metal surface. However the precipitation
process from the aqueous solution of the nonmetallic layer was not elucidated and no coupling of
phases to anodic polarization was performed.

In the present work the simulation of local H2S corrosion was still restricted to a limited set of
parameters. Further work is in progress with extended ranges of diffusion layer thicknesses,
temperatures and activity coefficients which, in particular may be affected by the additional presence
of chlorides.

Therefore, a consistent comparison to other experimental results seems so far difficult due to lack of
sufficient basic investigations.

As a summarizing presentation of the results, Figure 12 shows the limiting levels of pH and vol.-%
H2S for constant acidification times t pro.5 and total pressures of 15 and 7.5 bar. In the range of bulk
pH between 8 and 8.5, at both pressures and selected acidification times of 6400 s for 7.5 bar and
8900 s for 15 bar, increasing H/S contents require an increase in bulk pH for arriving at the same
acidification times. It may be assumed that the presented simulated mechanism is also reflected by
the pH and H/S limits for sulfide stress cracking as shown in the respective domain diagram 3. This
would explain the above mentioned frequent observations of simultaneous iron sulfide precipitation
and sulfide stress cracking starting from pitting corrosion as shown in Figure 13 and 14.

CONCLUSIONS

Based on the above rigorous assumptions for localized corrosion the following conclusions are
drawn:

The concept of precipitated layer mass controlled anodic polarization has been implemented for
modeling of local H2S corrosion.

As a consequence of anodic precipitation of FeS2 the remaining aqueous solution in the anodic
diffusion layer acidifies to theoretically low pH levels.

The corrosion currents and potentials are controlled by precipitation and properties of nonmetallic
layers.

The acidification times for pH 2.5 increase with increasing pH of the bulk solution and total
atmospheric pressure, while the mean corrosion currents are reduced at the same time.

8
 The effect of H2S is most significant at higher bulk pH levels and pressures and shows that
acidification times and mean corrosion currents are reduced by increasing contents of H2S.

The calculated anodic acidification rates during precipitation of iron sulfide have been
experimentally verified.

The calculations need fimher refinement with respect to transport mechanisms polarization and
properties of the nonmetallic layers.

ACKNOWLEDGEMENTS

The author would like to thank the Department for Education and Research of the Federal German
Government and pipe manufacturing company H.Butting GmbH, Germany for f'mancial support.

REFERENCES

1. S.N.Smith,R.S.Pakalapati, "Thirty Years of Downhole Corrosion Experience at Big Escambia

Creek:Corrosion Mechanisms and Inhibition" NACE CORROSION 2004,paper 04744
2. Th.Boellinghaus,H.Hoffmeister, Slow Strain Rate Testing of Low Carbon Martensitic
Stainless Steels" EFC Publ.No.23, pp.274-285
3. Th.Boellinghaus,H.Hoffmeister, M.Littich, "On-line Sulfide Stress Cracking Monitoring of
13% Cr Pipe Welds at Realistic Weld Restraint Conditions in the Instrttmented Restraint
Cracking (IRC) Test" "EFC Publ.No.23, pp.286-303
4. Y.Miyata, M.Kimura,T.Toyooka, Y.Nakano,F.Murase,"Corrosion Performance of Weldable
12 % Chromium Stainless Steel Seamless Pipes" EFC Publ.No.23, pp.231-241
5. G.Schmitt, R.Buschmann, K.Ma,C.Bosch,B.Sadlowsky,"Flow Initiated Hydrogen Uptake of
Steel in Sour Environment", NACE CORROSION 2004,paper 04470
6. A.K.Singh,A.Pourbaix,"E-pH-Diagrams for the System Fe-S-H20" CEBELCOR, Rapports
Techniques, Vol. 166, E240
7. A.Anderko,R.D.Young,"Simulation of COa/HaS Corrosion using Thermodynamic and
Electrochemical Models" NACE CORROSION1999,paper 31
8. R.Hausler,"Contfibution to the Understanding of HaS-Corrosion" NACE CORROSION
2004,paper 04732
9. S.Ramachandran,M.B.Ward,K.A.Bartrip,"Molecular Modeling of Corrosion of Iron in H2S-
Environments" NACE CORROSION 2002,paper 02240
10. S. Nesic, "Modeling of Internal Corrosion of Oil and Gas Pipelines Made from Carbon Steel"
Proc.NACE CORROSION 2004 Research Topical Symposium, pp. 147-191
11. B.Brown, S.R.Parakala, S.Nesic, "COa Corrosion in the Presence of Trace Amounts of HAS"
NACE CORROSION 2004,paper 04736
12. S.Nesic, M.Nordsveen,R.Nyborg,A.Stangeland,"A Mechanistic Model for CO/ Corrosion
with Protective Iron Carbonate Films" NACE CORROSION 2001,paper 01040
13. H.Hoffmeister,"Modeling Crevice Corrosion of Pure Chromium by Coupling of Phase and
Polarization Behavior", CORROSION 2004, Vol.60,no.4, pp.369-377
14. H.Hoffmeister,"Modeling of Crevice Corrosion of Pure Nickel by Coupling of Phase and
Polarization Behavior at Various pH, Chloride and Oxygen Levels" NACE CORROSION
2004,paper 04289
15. M.Pourbaix, "Atlas of Electrochemical Equilibria in Aqueous Solutions", NACE
Houston,Texas USA, CEBELCOR 1966
16. H.Kaesche,"Die Korrosion der Metalle", Springer Verlag 3rd ed. 1990, p.19
17. J.Kielland, J.Amer.chem.Soc.59,p.1675, 1937

9
18. Hiitte, (Stoffhiitte), ed.1967, p.152, Verlag W.Emst & Sohn, Berlin
19. N.Shridar, D.S.Dunn, J.Eleetroehemieal Soc., vol. 144(12) p.4243-4253 (1997)
20. H.Kaesche,"Die Korrosion der Metalle", Springer Verlag 3rd ed. 1990, p.78.
21. K.L.J.Lee, S.Nesie:"EIS Investigation of CO2/H2S Corrosion NACE CORROSION
2004,paper 04728
22. J.B.Sardisco,W.B.Wright,E.C.Greeo, CORROSION 19 (1963) p.354
23. P.Marcus:"Sulfur Assisted Corrosion Mechanisms and the Role of Alloyed Elements" in
"Corrosion Mechanisms in Theory and Practice",P.Marcus and J. Oudar eds. Marcus Dekker
1995, pp.239-263

FeSz layer thickness

H÷

HS
Fe *÷

acidified solution
bulk solution iron

FIGURE 1: Local corrosion system geometry

10
 %HS'=t 0 2,816(%.H+/o~Fe++).
'20

10 HS'%

y--
..,.

1E-10 -~. "

j
~

,",
" t"O

"r ~tfi
e,i "7 W
~ W W ~
1E'I 2

H*% 1E-14
x_
1E-.16
- "- Fe"'/,

tfi "" -'--

t"w4

FIGURE 2: Solubility limits for Fe304 and FeS2 in Fe-saturated solutions of the
system H20 - Fe - S

60 - H+(%) - 0. 000000012
PHbum=7 HS
pH2S=0.15 i)ar
50 0.00000001

40
S 0.000000008

Fe++
30 0.000000006
E
/ 0.000000004
20-
H+
10- 0.000000002

O _

0 500 1000 1500 2000

time (s)

FIGURE 3: Time dependent total composition change of surface layer

11
 0,01
%Fe= 0.01
PHbuJk = 7
0,009
PH2S= O.15 bar
P = 15bar /
0,008

0,007 I I
/
/
0,006 ,.t..--/ • v • ,/... /

"r" 0,005
/
0,004

0,003 Fe- H20

0,002

0,001
A
~ J
0 0,000000002 0,000000004 0,000000006 0,000000008 0,00000001
H+ (%)

FIGURE 4: Total concentration changes in diffusion layer with reference to projected constitutional
diagram HzO-Fe-FeS2

0.01
P~u,k=7
PH2S=0.15bar
0.008 P=i5 bar

-~ 0.006-
E
w
FeS2
m
E 0.004-

0.002

~J Fe304
_

0 500 1000 1500 2000

time (s)

FIGURE 5" Time dependent precipitation of FeS2 and Fe304

12
8.5 -

.5 ..........

6.5 -

5.5-
pHbuJk=7
PH2S=0.15bar
4.5 -
P=15 bar

3.5 -

2.5 -

0 500 1000 1500 2000

time (s)

FIGURE 6: Time dependem change of local pH during anodic acidification

pH

_i i f~_/,~ P L__
pn2o--u.a ua~

5g/I Na2SO4
0,8

-_..__.__.~~
I

0 500 1000 1500 2000

time (s)

FIGURE 7: Experimental pH reduction

13
 Ec (mY)
0.0045 -482.3
pHbu~=7
_.....,..__--.-
0.004 pH~S=0.15 bar
P=15 bar f
0.0035 f
f
/ - -482.35
..~
<
E
0.003

0.0025
/ Ec

-482.4
.e

o
0.002

0.0015
-
/
- -482.45
0.001

0.0005
-

-
/
O _ --482.5
0 500 1000 1500 2000
t i m e (s)

Figure 8: Time dependent change of current and potential

current (mA)
6,5 -
I
pHbul. = 6
6- pH2S = 0.015 bar -
P = 1.5 bar
5,5 -

_
f

i
o
-!= 4 , 5 -
Q.
_

pH
3,5 -

2,5 -

500 1000 1500 2000

time (s)

FIGURE 9: Time dependent currents and acidification at pH 6

14
 current i mean (mA)
10000 ~k " P(baO ]J ' ! . . . . .L
9000 . . . . .

8000
1"5X 1 15//

7000 '\ , ,,
~ 6000 . ~ ~
5000 .........................
, . xt / 7
2000

0
8.5
6 6 . pH bulk

FIGURE 10: Effect of pH butkand pressure on acidification time t pH2.5 and i mean

current i m e a n 3 n ~ ) o 0 1
9900
t pH2.5 1
3
9400 5
10 %HzS
8900

8400 0.0008
A 15
W
lur}
¢,,i
-r- 7900 P(bar)
D.

O 7400 -
,.mE
15
6900 - - 0.0006

6400 -

5 9 0 0 - i" /.b "

, ,

5400 - - 0.0004
8.4 8.42 8.44 8.46 8.48 8.5

pH bulk

Figure 11" Effect ofpH buU,, pressure and H2S on acidification time t pI-I2.5 and i mean

15
 8.5

P(bar)
t pH2.5 (S)
7.5
6400
8.45

,,1¢
,,,,.,

t'~

-r-
15
Q. 8900

8.4

i
I

8.35
10 i
HzS (vol.%)

FIGURE 12" pH bulkand H2S - limit for 7.5 and 15 bar at various acidification times

~ '~

~,~ . . . . . ~d

~ , ~- ~:

FIGURE 13" Local corrosion at SSRT cracking (500:1) of Duplex Stainless Steel in 0.01%H2S, 25%
NaC1 deaerated brine, strain rate 106/s sulfur concentration of
corrosion product 2.16 % (EDX).

16
FIGURE 14: Duplex Stainless Steel SSRT crack surface at conditions of figure 13.
Surface sulfur concentration 1.08% (EDX)

17