A

8TsptitsE©

K

1

tciwasiw[o

X

(

0d

Available online at www.sciencedirect.com

Electrochimica Acta 53 (2008) 3690–3700

Corrosion behavior of SM 80SS tube steel in stimulant solutioncontaining H2S and CO2

Z.F. Yin a,b,∗, W.Z. Zhao a, Z.Q. Bai b, Y.R. Feng b, W.J. Zhou c

a School of Materials Science and Engineering, Xi’an Jiaotong University, Xi’an 710049, PR Chinab The Key Laboratory for Mechanical and Environmental Behavior of Tubular Goods, Tubular Goods

Research Center, CNPC, Dianzier Road, Xi’an Shannxi 710065, PR Chinac School of Petroleum Engineering, Xi’an Shiyou University, Xi’an 710065, PR China

Received 9 August 2007; received in revised form 21 November 2007; accepted 2 December 2007Available online 23 December 2007

bstract

Scanning electron microscopy, X-ray diffraction and electrochemical measurement technique were applied to investigate the corrosion of SM0SS tube steel in stimulant solution with carbon dioxide (CO2) and hydrogen sulfide (H2S) at variable conditions of PCO2/PH2S and temperature.he results suggest that there exists a synergism of sweet corrosion and sour corrosion on the steel surface, corrosion attack increases in the initialtage and then decrease with the increase of PCO2 or PH2S; serious corrosion occurs in the PCO2/PH2S ranged from 31 to 520. In addition, the fittedarabola function equation Y = 0.47873 + 0.04014X – (3.23788E−5)X2 is established, and the most serious corrosion is 600 for PCO2/PH2S. Underhe moderate contents of PCO2 and PH2S, the corrosion scale consists of FeS0.9 and FeCO3; for relatively high PH2S, additive product FeS comes

◦

nto being at high temperature such as T = 150 C, product FeO(OH) is found in the corrosion scale. The H2S corrosion has a significant effect onhe whole reaction process and iron sulfide is superior to precipitating on the steel surface compared with iron carbonate. In addition, the surfacecales of iron sulfide almost act as a diffusion barrier and inhibit the corrosion by a coverage effect strongly depending on H2S concentration byIS measurement.2007 Elsevier Ltd. All rights reserved.

asurem

mpCibraiHt

eywords: SM 80SS tube steel; CO2/H2S corrosion; Corrosion scales; EIS me

. Introduction

To our knowledge, the complex work conditions such as highemperature, high pressure, multiphase flow, pH value and iononcentrations, make the corrosion of tube goods very seriousn oil and gas well. Since 1940s, the corrosion of tube goodsas concerned in order to understand and control it [1]. In the

bsence of protective films, an increase in CO2 partial pres-ure will result in an increase of corrosion rate, because withncreased CO2 partial pressure, the direct reduction of H2CO3

ill be accelerated due to an increase of H2CO3 concentration

2]. However, when other conditions are favorable for formationf protective iron carbonate films, increased CO2 partial pressure

∗ Corresponding author at: School of Materials Science and Engineering,i’an Jiaotong University, Xi’an 710049, PR China. Tel.: +86 29 8872 6201.

E-mail addresses: yinzhifu919@sohu.com, yinzhifu919@stu.xjtu.edu.cnZ.F. Yin).

ppitDiirc

013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.oi:10.1016/j.electacta.2007.12.039

ent

ay help to facilitate the film formation. At a given high enoughH, an increase in CO2 partial pressure results in an increase ofO3

2− concentration and a higher supersaturation, thus speed-ng up precipitation and film formation [3]. It has been illustratedoth experimentally [4,5] and computationally [6] that corrosionate changes significantly with respect to pH. However, pH hasn indirect effect on the formation of protective films (such asron carbonate) that influences remarkably the corrosion rate.igh pH leads to a decreased solubility of iron carbonate and

hus results in an increased precipitation rate, faster formation ofrotective films and hence reduction of the corrosion [3]. Tem-erature accelerates all the processes involved in CO2 corrosionncluding transport of species, chemical reactions in the bulk ofhe solutions and electrochemical reactions at the metal surface.epending on whether the solubility of protective films (such as

ron carbonate or other salts) is exceeded, temperature can eitherncrease or decrease the corrosion rate [3,7]. In the case of cor-osion where protective films do not form (typically at low pH),orrosion rate increases with increasing temperature. However,

ica A

aeod

obraT

i

Htaosnats(mCptcaoctiiFsircs

tTHfeacMaaecebtTw

piamdspNsru

btnrgrctp

2

2

ctd3wtgthol, weighted using a precision of 0.1 mg, and finally storedin desiccators for use.

Concentrations of various ions of the stimulant solution werelisted in Table 2. The solution was deoxygenated by pure nitro-

Z.F. Yin et al. / Electrochim

t a higher pH where solubility of protective films is likely to bexceeded, increased temperature would accelerate the kineticsf precipitation and facilitate protective films formation, thusecreasing the corrosion rate.

The internal CO2 corrosion of mild steel in the presencef hydrogen sulfide (H2S) represents a significant problem foroth oil refineries and natural gas treatment facilities. In theecent years the problem has become more important as thevailable reserves of oil possess a considerable amount of H2S.he presence of H2S, additional chemical reactions occurring

n the bulk of the solution include H2SKH2S←→H+ + HS− and

S−KHS−←→H+ + S2−. Although H2S gas is about three times

han CO2 gas, the acid created by the dissociation of H2S isbout three times weaker than carbonic acid. Hence, the effectf H2S gas on decreasing the solution pH is approximately theame as CO2 gas. Unlike dissolved CO2, dissolved H2S doesot need to undergo the slow hydration step in order to becomen acid. Depending on various environmental factors, differentypes of iron sulfide can be formed. Mackinawite is a tetragonalulfur-deficient iron sulfide with a composition of either FeS1− x

1 < x < 0.07) or Fe1 + xS (0.057 < x < 0.064) [8]. Mackinawite is aain corrosion product on the surface of carbon steel in saturatedO2/H2S solution. There are two ways forming mackinawite:recipitation of aqueous Fe2+ and S2− and direct chemical reac-ion of dissolved H2S with the metallic iron. Cubic FeS has aubic stoichiometric crystal structure and is only encountereds a corrosion product [8]. However, cubic FeS is considered asnly a metastable species. It is found that troilite appears as aorrosion product on carbon steel surface in aqueous H2S at lowemperatures, due to high local iron concentrations at the corrod-ng surface [9]. Pyrrhotite is more stable than mackinawite and its an iron-deficient iron sulfide with a composition ranging frome7S8 to stoichiometric troilite, FeS. Pytite has a cubic crystaltructure [8,10], and it is the most stable iron sulfide. Greigites a thio-spinel of iron [8], Fe3S4, which is thermodynamicallyelative to troilite and pyrite. In addition, greigite will form aorrosion product only if oxygen or sulfur is introduced into theolutions.

In 1965, Saridisco and Pitts presented two papers onhe corrosion of iron in an H2S–CO2–H2O system [11,12].hey concluded that during liquid phase corrosion of iron by2S–CO2–H2O, overall reaction is controlled partially by inter-

ace reaction and partially by passage (diffusion) of ions andlectrons across film. At low H2S concentrations, reaction mech-nism approaches complete diffusion control and at high H2Soncentrations it approaches complete interface control. Then,urata et al. evaluated the corrosion rate as a function of H2S

nd CO2 partial pressure and temperature [13]. They showedclear illustration of the basis for confusion in the early lit-

rature that cited FeS film as both increasing and decreasingorrosivity when H2S was added to CO2. Lichti et al. reviewsxperience with geothermal wells in New Zealand that produce

rines containing CO2 and H2S [14]. They found the forma-ion of corrosion product of mackinawite, troilite and pyrrhotite.hey concluded that sulfide films reduced corrosion rates, evenhen present in minute amounts. Afterwards Smith and Pacheco

cta 53 (2008) 3690–3700 3691

resent the results of lab tests that were run to evaluate the min-mum H2S levels required to form mackinawite [15]. The paperlso presents a simplified equation for the calculation of theinimum H2S required for mackinawite formation and for the

etermination of the critical H2S/CO2 ratio. This H2S effecteemed to vanish at higher H2S concentrations and higher tem-eratures (>80 ◦C) when a protective film forms [15,16]. In aorwegian study, it was showed that this effect of H2S could be

ignificant only in the low pH range (pH < 5) [17]. However, theeasons behind the H2S effect on CO2 corrosion are not entirelynderstood.

Although the interaction of H2S with low-carbon steels haveeen published by various authors [18,19], the understanding ofhe effect of H2S on CO2 corrosion is still limited because theature of the interaction with carbon steel is complicated. Soesearch on the interaction between H2S and CO2 has causedreat concern. The present work aims at investigating the cor-osion behavior of SM 80SS tube steel in stimulant solutionontaining H2S and CO2, it is helpful for predicting the poten-ial danger caused by corrosion and take measures to controlremature damage.

. Experimental

.1. Weight loss tests

The corrosion behavior was evaluated using an auto-lave (34.44 MPa) made by Cortest Company at highemperature and high pressure. Fig. 1 shows the schematiciagram of the autoclave. Rectangular test specimens, sized0 mm× 15 mm× 4 mm, were made from SM 80SS tube steelhose chemical composition was shown in Table 1. Before

he test, the surface of the test specimens were polished withrit silicon carbide papers progressively up to 800 grades,hen degreased with acetone and rinsed with absolute alco-

Fig. 1. Schematic illumination of corrosion apparatus.

3692 Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700

Table 1Chemical composition of SM 80SS tube steel (wt.%)

C 0.25Mn 0.41Si 0.22P 0.006S 0.006Cr 1.01Mo 0.32Ni 0.030V 0.006Ti 0.024

Table 2Concentration of various ions of stimulant solution (mg/L)

Cl− 39,425SO4

2− 15HCO3

− 195Ca2+ 12,215MK

g(tpwoSh0(t

2

tmr

2

acsMsPu

Table 4The average corrosion rates of SM 80SS tube steel under various CO2 partialpressures

PCO2 (vol.%) CR (mm/a)

5 1.6610 3.8713

(ueot

3

3

3

testttFtUi3Tsafiihfa

3

va

TT

BV

g2+ 1450+ + Na+ 8797

en for 14 h. Then, the system was given required partial pressurebut the total pressure is 30 MPa) and temperature. Table 3 showshe conditions of experiment. After 72 h, the corrosion sam-les are removed from the autoclave and rinsed with deionizedater. They were divided into two groups: the samples in groupne were descaled (the solution: 1 L HCl (ρ = 1.19 g/L), 20 gb2O3 and 50 g SnCl2), rinsed with water and absolute alco-ol, dried in nature state and weighted again with a precise of.01 mg. CO2 corrosion rate was represented by corrosion depthmm)/corrosion time (per year), i.e. mm/a. The samples in groupwo were not descaled, which would be used for surface analysis.

.2. Surface characteristics

Scanning electron microscopy (SEM) was utilized to inves-igate the surface morphology before and after corrosion. The

icrostructure of the corrosion sample was analyzed using X-ay diffractometer with filtered Cu K� radiation.

.3. EIS measurement

After the high temperature and high-pressure corrosion test,ll test samples were kept in analytic grade alcohol beforeorrosion. The test brine solution, with 3.5 wt.% NaCl, wasaturated with CO2. An EG&G Princeton Research (PAR)

odel 5210 lock-in amplifier and PAR Model 273A potential-tat/galvanostat controlled by a personal computer were used.AR Model 398 software and PAR Model 352 software weresed to calculate electrochemical impendence spectroscopy

0P

tt

able 3he variational ranges for test conditions

PCO2 (vol.%) PH2S(vol.%)

asal content 20.0 0.16ariational range 5–35 0.00016–3.2

5 4.135 6.67

EIS). Three-electrode electrochemical cell system with the sat-rated calmel reference electrode (SCE) and graphite auxiliarylectrode were used. The EIS measurements were carried out atpen-circuit potential using an alternating current voltage ampli-ude of 5 mV. The frequency varied from 5 mHz to 100 kHz.

. Results and discussion

.1. Weight loss tests

.1.1. The influence of CO2 partial pressureTable 4 shows the average corrosion rates (CR) of SM 80SS

ube steel under various CO2 partial pressures. In the system,xcept for the variable CO2 partial pressure, the PH2S is con-tant 0.16 vol.% (1600 ppm), and then the N2 was pressed up tohe total pressure (30 MPa). The result from Table 4 indicateshat the average corrosion rates are very great, compared withhe criterion of average corrosion rates (NACE RP-0775-91).rom the criterion, when the average corrosion rate was bigger

han 0.125 mm/a, the tube steels are subject to serious corrosion.nder the conditions, it can be seen that the CRs of the samples

ncrease with increasing PCO2 and the CR under the PCO2 of5 vol.% is about four times to that when the PCO2 is 5 vol.%.he reason is probably that with increased CO2 partial pres-ure, the direct reduction of H2CO3 will be accelerated due ton increase of H2CO3 concentration in the absence of protectivelms. So an increase in CO2 partial pressure will result in an

ncrease of corrosion rate. We can conclude that the average CRas an increasing trend from the limited test points. All resultsrom the tests are in accordance with the outcomes by Nesic etl. [2], Carlos et al. [20] and Sridhar et al. [21].

.1.2. The influence of H2S partial pressureTable 5 shows the average CR of SM 80SS tube steel under

arious H2S partial pressures. In the system, the variable PH2Sre as follows: 0.00016 vol.% (1.6 ppm), 0.016 vol.% (160 ppm),

.04 vol.% (400 ppm) and 3.2 vol.% (32,000 ppm), while theCO2 keeps constant 20 vol.%, and then the N2 was pumped up

o the total pressure (30 MPa). The result from Table 5 indicateshat the average corrosion rates are relatively lower than those

T (◦C) CCl− (g/L) C(Ca2++Mg2+) (g/L)

90 50 20 (15Ca2+ + 5 Mg2+)40–150

Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700 3693

Table 5The average corrosion rates of SM 80SS tube steel under various H2S partialpressures

PH2S(vol.%) CR (mm/a)

0.00016 0.020.016 0.080.04 0.363.2 0.12

Table 6The average corrosion rates of SM 80SS tube steel at various temperatures

T (◦C) CR (mm/a)

40 2.9660 3.87

11

uttP

dCoigiriwioHspw

3

twtThc1

FS

tpIft[(colHitc

3

taP

c5

P

TC

TC

C

S

00 6.1350 4.14

nder various PCO2 . Under the conditions, it could be found thathe CR of the samples increased with the PH2S increased, andhen the CR was up to a maximum value (0.36 mm/a) at theH2S of 0.04 vol.% for only the limited data points. After theatum point of 0.04 vol.% PH2S, with the PH2S increasing theR declined. When test conditions are favorable for formationf different types of protective iron sulfide iron carbonate films,ncreased PH2S may help to facilitate the film formation. At aiven high enough pH, an increase in CO2 partial pressure resultsn an increase of CO3

2− concentration and a higher supersatu-ation, thus speeding up precipitation and film formation. It isn agreement with the results of Nesic and Lee [3]. Therefore,e perhaps obtain the similar result for H2S partial pressure,

.e. increasing the H2S concentration can result in the increasef S2− concentration and accelerate the formation of sulfides.owever, because the data points of PH2S (vol.%) were uncon-

idered from 0.04 to 3.2, we could not predict the real criticaloint (maximum value) and the reasons for the phenomenon,hich would be carried out in the future work.

.1.3. The influence of temperatureThe average CR of SM 80SS tube steel at various tempera-

ures was shown in Table 6. In the system, the PH2S and PCO2

ere 0.16 vol.% (1600 ppm) and 20 vol.%, respectively, and thenhe N2 was pumped up to the total pressure (30 MPa) as above.

he results demonstrate that the average corrosion rates are veryigh for the four temperature points. Under the conditions, itould be found that the maximum CR of the samples occurred at00 ◦C, while over 100 ◦C, the CR decreased gradually. We have

prob

able 7orrosion characteristics of SM 80SS tube steel at high temperature and high pressur

PCO2 /PH2S

31 63 94 219 125000

(◦C) 90 90 90 90 90R (mm/a) 1.66 3.87 4.13 6.67 0.21

orrosiveness S S S S M

: serious corrosion; M: medium corrosion; L: light corrosion.

ig. 2. The relation between corrosiveness and PCO2 /PH2S for the representativeM 80SS tube steel using polynomial fitting method.

o point out that we should consider more temperature pointsredicting the maximum corrosion rate. In the paper reported,keda et al. research results suggested that the corrosion scalesormed on the surface of samples were loose and poor protec-ion at relatively high temperature for the matrix of the steel22]. Depending on whether the solubility of protective filmssuch as iron carbonate or other salts) is exceeded, temperaturean either increase or decrease the corrosion rate. In the casef corrosion where protective films do not form (typically atow pH), corrosion rate increases with increasing temperature.owever, when the solubility of protective films is exceeded,

ncreased temperature will accelerate the kinetics of precipita-ion and facilitate protective films formation, thus decreasing theorrosion rate.

.1.4. PCO2/PH2S quantitative analysisTable 7 shows corrosion characteristics of the representa-

ive SM 80SS tube steel at high temperature and high pressureccording to NACE RP-0775-91 criterion. It can be found thatCO2/PH2S was high or low, corrosion degree decreased, while,orrosion was very serious when PCO2/PH2S ranges from 31 to20.

Fig. 2 shows the relation between corrosiveness andCO2/PH2S for the representative SM 80SS tube steel using

olynomial fitting method. It is subjected to serious corrosionegion above dash dot lines of the CR (0.254 mm/a). Based on thebservation from Fig. 2 there existed a parabola function relationetween corrosiveness and PCO2/PH2S. The polynomial fitting

e

1250 520 6 125 125 125 125

90 90 90 40 60 100 1500.08 12.36 0.12 2.96 3.87 6.13 4.14

L S M S S S S

3694 Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700

F ons oM

e

Y

wti

3

3

somthptsscp

sscta9tsswttP

hecsc

FM

ig. 3. SEM images of corroded surface of SM 80SS steel under the conditiagnification about 200×.

quation could be expressed as follows:

= 0.47873+ 0.04014X− (3.23788E–5)X2 (1)

here X is the PCO2/PH2S and Y is the corrosion rate. As a result,he most serious corrosion occurs when the ratio of PCO2/PH2Ss 600.

.2. Analysis of corrosion morphology

.2.1. SEM images analysisIt can be seen from Fig. 3 that the corrosion scales on the

urface are compact and acerate porous characters at the PCO2

f 5 vol.%. With the increasing of PCO2 , the corrosion scalesorphologies have made great variation. When PCO2 was up

o 35%, crystal grains are bulky, irregular and dense, but theyave some gaps between the crystal grains. The interspaces areossible to be passages that the corrosion solution diffuses fromhe surface into the inner layer and even into the surface of the

ubstrate, which may results in some pitting corrosions on theurface of the substrate. In general, the fine and compact scalean impede corrosive ions through the scale, which has a goodrotection for the matrix. In contrast, the relatively big and loose

stP

ig. 4. SEM images of corroded surface of SM 80SS steel under the conditions of PCO

agnification about 200×.

f PH2S = 0.16 vol.% at 90 ◦C: (a) PCO2 = 5 vol.% and (b) PCO2 = 35 vol.%.

cale has an ion-passing property, leading to an obvious corro-ion character. The analytic results are in accordance with theorrosion rates as described above. In addition, our work onhe cross-section of samples suggests that the scale thickness isbout 52 �m at 5 vol.% PCO2 . It has a scale thickness of about9 �m at 35 vol.% PCO2 equaled to two times approximatelyhe one at 5 vol.% PCO2 . Moreover, uniform corrosion and somecattered acerate pits occurred on the surface of SM 80SS tubeteel. Masamura et al. [23] and Fierro et al. [24] presented thathen PCO2/PH2S > 200, CO2 played a leading role in this sys-

em, but, there would form a layer of compact FeS scale onhe tube steel surface reducing the corrosion rate. While, whenCO2/PH2S < 200, the FeS scale would firstly formed, whichobbled the protective FeCO3 scale coming into being. How-ver, the latter condition probably decreased or promoted theorrosion rate. Besides, generally three regimes comprised ofour regime, mixed regime and sweet regime in CO2/H2S systeman be classified based on the PCO2/PH2S of 500 [25].

Similarly, it could be found from Fig. 4(a) that the corro-ion surface was not very smooth, with the good coalescenceo the matrix. Compared with Fig. 4(a), with the increasing ofH2S, the surface scale became more compact and flat. Based

2 = 20.0 vol.% at 90 ◦C: (a) PH2S = 0.00016 vol.% and (b) PH2S = 3.2 vol.%.

Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700 3695

F ns ofM

otwrtcffa[idtcc

boicagaatcSw

3

sttmsTe

rcaa

•(x+ 1)Fe + (x+ 0.9)H2S→xFeS + FeS0.9+ (x+ 0.9)H2↑

ig. 5. SEM images of corroded surface of SM 80SS steel under the conditioagnification about 200×.

n the cross-section observed, the result shows that the scalehickness is about 51 �m at the PH2S of 0.00016 vol.%, thenith the increasing of PH2S, the scale thickness decreases. The

eason is that amounts of FeCO3 product scale is formed inhe former condition while a few forming iron sulfides in latterondition. However, the latter corrosion rate is higher than theormer according to the result above. But, the reason is not clearor this phenomenon. Murata et al. evaluated the corrosion rates a function of H2S and CO2 partial pressure and temperature13]. They showed a clear illustration of the basis for confusionn the early literature that cited FeS film as both increasing andecreasing corrosivity when H2S was added to CO2. The data ofhe scale thickness were in agreement with the condition of theritical PCO2/PH2S accepted commonly. As above, only generalorrosion occurred on the SM 80SS tube steel surface.

Fig. 5 shows that as temperature increased, the surface grainsecame bigger and looser than those at lower temperature. More-ver, the formation scales are relatively regular and minutelynterstitial. The interspaces are possible to be passages that theorrosion solution diffuses from the surface into the inner layernd even into the surface of the substrate. The SEM morpholo-ies were not in accordance with the other researches, probablyt the temperature of 150 ◦C the scale could not impede theggressive ions attack tube steel effectively, while acceleratedhe corrosion rate. It was suggested that there probably existed aritical temperature higher than 150 ◦C, which could protect theM 80SS tube steel effectively. The reasons for the phenomenonill be confirmed using other test methods.

.2.2. XRD analysisXRD analysis shown in Fig. 6(a) indicates that the corro-

ion scales mainly consisted of FeS0.9, FeS and FeCO3 whenhe PCO2 is 5 vol.%. Obviously, it could be seen from Fig. 6(b)hat the corrosion scales comprised FeS0.9 and FeCO3. The

ajority of strong peaks come from FeS0.9 substance andome weak peaks of FeS and FeCO3 occur in the spectra.he result indicates that the H2S corrosion has a significantffect on the whole reaction process and iron sulfide is supe-

F0

PCO2 = 20.0 vol.% and PH2S = 0.16 vol.%: (a) T = 40 ◦C and (b) T = 150 ◦C.

ior to precipitating on the steel surface compared with ironarbonate. Under the conditions with H2S, the reaction mech-nisms of SM 80SS altered with PCO2 , which could be writtens:

At low PCO2 :

ig. 6. XRD spectra for the surface scales under the conditions of PH2S =.16 vol.% at 90 ◦C: (a) PCO2 = 5 vol.%; (b) PCO2 = 35 vol.%.

3696 Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700

F2

•

u

mttt3Fcc8

•

•

F2

m4FaS

F

Cstcih

3

3

c

ig. 7. XRD spectra for the surface scales under the conditions of PCO2 =0.0 vol.% at 90 ◦C: (a) PH2S = 0.00016 vol.% and (b) PH2S = 3.2 vol.%.

At relatively high PCO2 :

Fe + 0.9H2S → FeS0.9+ 0.9H2↑

However, the mechanism for FeCO3 formation remainsnchangeable with the increasing of PCO2 .

By comparing Fig. 7(a) and (b), the corrosion scale wasainly composed of FeCO3 and a little quantity of FeS0.9 when

he PH2S was 0.00016 vol.% (1.6 ppm). As the PH2S increased,he contents of FeCO3 reduced and ferrous sulfide graduallyransformed to FeS. Especially when PH2S increased up to.2 vol.% (32,000 ppm), the corrosion scale only consisted ofeS. It can be concluded that H2S corrosion has overpoweringompeting advantage compared with CO2 corrosion. Under theonditions with addition of H2S, the reaction mechanism of SM0SS varied with PH2S, which could be presented as:

At low PH2S:

Fe + CO2+H2O → FeCO3+H2↑

(x+ 1)Fe + (x+ 0.9)H2S → xFeS + FeS0.9+ (x+ 1)H2↑

At relatively high PH2S:

Fe + H2S → FeS + H2↑

aTfiq

ig. 8. XRD spectra for the surface scale under the conditions of PCO2 =0.0 vol.% and PH2S = 0.16 vol.%: (a) T = 40 ◦C and (b) T = 150 ◦C.

It could be shown in Fig. 8(a) that the corrosion scale wasainly composed of FeS0.9 and FeCO3 at the temperature of

0 ◦C and 5 vol.% PCO2 . Obviously, it could be seen fromig. 8(b) that the corrosion scales consisted of FeO(OH), FeCO3nd FeS0.9. In the presence of H2S, the reaction mechanism ofM 80SS changed with temperature.

At the high temperature of 150 ◦C, it could be presented as:

e + 2H2O – 3e → FeO(OH) + 3H+

Generally, with the increasing of temperature the solubility ofO2 in water decreases which restrains the CO2 corrosion on the

urface of tube steel. To a great extent, CO2 corrosion accelerateshe H2S corrosion and there exists a synergism between CO2orrosion and H2S corrosion. Moreover, when the temperatures very high, the corrosion scales formed on the steel surfacesave a good protection.

.3. EIS analysis of corrosion scale

.3.1. High PCO2 conditionFig. 9 shows the Nyquist plot and the equivalent electric

ircuit of SM 80SS tube steel covered with corrosion scales

t PCO2 = 35.0 vol.% and PH2S = 0.16 vol.% (1600 ppm),= 90 ◦C in 3.5% NaCl solution. It can be seen from thetted result that there is an obvious capacitive arc at high fre-uency, which could be considered as capacitance of double

Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700 3697

el co

ebific

teirbrr

Z

tiarw

3

t23oaflur

TI

RCRCRZ

cictdbaelasaac

to

H

F

F

F

F

Fig. 9. The Nyquist plot and the equivalent electric circuit of SM 80SS ste

lectrode layer between the corrosion scale and electrode; War-urg impedance appeared clearly at low frequency, generally,t is a characteristic of diffusion process. In addition, it can beound that the equivalent capacitance first formed by ions pass-ng corrosion thick scale on sample surface, compared with thinorrosion scale.

For the equivalent electric circuit, where Cdl is the capaci-ance of double electrode layer between the corrosion scale andlectrode, C is the capacitance for the ions passing scale, and Zws the impedance corresponding to ions diffusion through cor-osion scale, while, Rs, Rtc and Rtw are the solution resistanceetween the working and reference electrodes, charge transferesistance of activation controlled reaction and charge transferesistance of diffusion controlled reaction, respectively.

By calculating the Warburg impedance (Zw) shown in Table 8,w = 0.5854 � cm2 s−0.5, suggests that Zw value is small and

he scale has weak protection because of the serious corrosion. Its in accordance with the results obtained from the morphologynd corrosion rate, using the mass loss test and SEM method,espectively. The EIS result in present work is also in agreementith that reported by de Moraes et al. [26].

.3.2. High PH2S conditionFig. 10 demonstrates the effect of very high H2S concentra-

ion on the characteristics of the impedance plots at PCO2 =0.0 vol.% and PH2S = 1.6 vol.% (16,000 ppm), T = 90 ◦C in.5% NaCl solution. It can be seen from the fitted result that therenly is obvious capacitive arc at high frequency, which indicatesdouble-layer capacitance. The capacitive arc is not uncommon

or iron dissolution in acidic media and it is suggested in theiterature that the heterogeneous surface roughness and the non-niform distribution of current density on the surface may beelated to it [27,28].

able 8mpedance fitting parameters obtained from EIS and equivalent electric circuit

s (� cm2) 4.067(F cm2) 1.313

tc (� cm2) 4.807

dl (F cm2) 0.07278

tw (� cm2) 61.56

w (� cm2 s−0.5) 0.5854

3

er(

TI

RCRCR

vered with corrosion scales at PCO2 = 35.0 vol.% in 3.5% NaCl solution.

The parameters have same meaning in the equivalent electricircuit as shown in Table 9. It can be presented that the polar-zation resistance calculated is 172 � cm2 suggesting the mildorrosion. The result suggests that the mechanism is still chargeransfer controlled in the presence of H2S, but the sulfide scaleetected on the electrode surface has inhibited the corrosion ratey a coverage effect. For the anodic reaction mechanism of CO2,n inductive loop arises in the low-frequency region when thelectrode surface in the absence of corrosion scales in the pastiteratures. The addition of H2S makes the inductive loop dis-ppear in Nyquist plot, probably relating to the adsorption ofulfides. When H2S dissolved in water it partially dissociatess H+ and HS− ions, because HS− has greater chemisorptionbility than HCO3

− and OH− ions [29], and great degree ofoverage induces disappearance of inductive loop.

In order to well understand the anodic dissolution process inhis work, the mechanisms of Fe mainly controlled by adsorptionf the HS− ion as follows:

S− +Fe = Fe(HS−)ads

e(HS−)ads= Fe(SH+)ads+ 2e

e(SH+)ads→ FeS1− x+ xSH− + (1− x)H+

e(SH+)ads+H3O+→ Fe2+ +H2S + H2O

e(SH+)ads+HS−→ FeS + H2S

.3.3. High temperature condition

Fig. 11 illustrates the Nyquist plot and the equivalent

lectric circuit of SM 80SS tube steel covered with cor-osion scales at PCO2 = 20.0 vol.% and PH2S = 0.16 vol.%1600 ppm), T = 90 ◦C in 3.5% NaCl solution. It can be seen from

able 9mpedance fitting parameters obtained from EIS and equivalent electric circuit

s (� cm2) 0.2099(F cm2) 0.02818

c (� cm2) 151.9

dl (F cm2) 0.021227

t (� cm2) 19.95

3698 Z.F. Yin et al. / Electrochimica Acta 53 (2008) 3690–3700

Fig. 10. The Nyquist plot and the equivalent electric circuit of SM 80SS covered with corrosion scales at PH2S = 1.6 vol.% in 3.5% NaCl solution.

S ste

ttaveil

8asibbdtms

TI

RCRCR

srctmtitc

dt

Fig. 11. The Nyquist plot and the equivalent electric circuit of SM 80S

he fitted result that the Nyquist plot maintains similar charac-eristics and exhibits only charge transfer controlled behavior. Itlso indicated that the reaction process was controlled by acti-ation effect. Impedance fitting parameters obtained from thequivalent electric circuit are shown in Table 10. From the polar-zation resistance (Rp) computed below, there probably existedittle influence of diffusion control.

On the other hand, the polarization resistance computed is6 � cm2, while 172 � cm2 as described at PH2S = 1.6 vol.%bove. Therefore, the polarization resistance from EIS mea-urements also indicates the decrease in corrosion rate withncreasing H2S concentration. From the above evidence, it cane concluded that the surface scales almost act as a diffusionarrier and inhibit the corrosion by a coverage effect strongly

epending on H2S concentration under the experimental condi-ions. The trend, by and large, is similar to that obtained by the

ethods used for the morphologies and corrosion rates of theurface scales as described above.

able 10mpedance fitting parameters obtained from EIS and equivalent electric circuit

s (� cm2) 0.4143

dl (F cm2) 0.05208

t (� cm2) 60.74(F cm2) 0.01427

c (� cm2) 24.68

I

T

Y

ws

Y

Ys

el covered with corrosion scales at T = 150 ◦C in 3.5% NaCl solution.

In nature, when the steel surface is covered with corrosioncale, the corrosion rate is controlled by ions diffusion rate in cor-osion scale. Diffusion impedance occurred in spectrum is justorresponding to the characteristics of ions diffusion throughhe scale. In the spectra of N80 steel, the additional capacitive in

iddle frequency indicate that charge transfer controlled elec-rochemical process paralleled to diffusion process also exist. Its well known that pits could be considered as active area wherehe corrosion process is controlled by charge transfer with higherorrosion rate.

It could be assumed that Ip is pitting current density, Id theiffusion current density, and θ is the coverage rate of pits. Thenotal corrosion current could be written as:

= Ipθ + Id(1− θ) (2)

he total admittance could be described as:

= Ypθ + Yd(1− θ) (3)

here Yd is the admittance caused by ions diffusion in corrosioncale, namely the Warburg impedance. It could be presented as:

√ 1

d = Y0(j�)1/2 coth(B jω), B = √

D(4)

0 is the Warburg coefficient, and l is the thickness of diffu-ion layer, here it is the thickness of corrosion scale, D is the

ica A

ia

Y

m

a

b

wr(r

θ

(

b

o

b

m

miai

4

riP

30mtoCiP

s

sptascicp(cladtcIds

A

dFCa

R

[

[[[

[

[[[

Z.F. Yin et al. / Electrochim

on diffusion coefficient in corrosion scale. In Eq. (3), Yp is thedmittance of pitting. It could be written as [30]:

p = 1

Rpt+ mb

a+ j�(5)

1

Rpt=

(∂(Ipθ)

∂E

)ss

(6)

=(

∂(Ipθ)

∂θ

)ss= Ip > 0 (7)

=(

∂θ′

∂θ

)ss

(8)

=(

∂θ′

∂E

)ss

(9)

here the subscript ‘ss’ denotes steady-state. Rpt is the transferesistance of the pitting. At stability condition, the value of a =∂θ′/∂θ)ss must be negative. The θ′ is the pits coverage variableate with time. It could be expressed as:

′ = dθ

dt= K[Id(1− θ)− Ipθ] (10)

K is the coverage-electricity conversion coefficient. For a =∂θ′/∂θ)ss = −K(Id + Ip) < 0, concluding, K > 0.

From Eq. (10), b could be rewritten as:

=(

∂θ′

∂E

)ss= −K

[∂[Id(1− θ)]

∂E− ∂(Ipθ)

∂E

]

Since Id is the diffusion current density, it is not the functionf voltage. Then

= − K

Rpt< 0

b < 0

According to the theory of Cao and Zhang [30], whenb < 0, the characteristics of capacitive impedance can occur

n impedance spectrum. In other words, the total impedance,s presented by Eq. (3), is composed of finite length diffusionmpedance parallel connecting with capacitive impedance.

. Conclusion

There exists a synergism of sweet corrosion and sour cor-osion on the steel surface, corrosion attack increases in thenitial stage and then decrease with the increase of PCO2 orH2S; serious corrosion occurs in the PCO2/PH2S ranged from1 to 520. In addition, the fitted parabola function equation Y =.47873+ 0.04014X− (3.23788E–5)X2 is established, and theost serious corrosion is 600 for PCO2/PH2S. At the tempera-

ure of 40, 60 and 100 ◦C, the effect mechanism and compositionf the corrosion scale were mainly controlled by synergism of

O2 and H2S; while, temperature is one very important influenc-

ng factor over 150 ◦C besides CO2 and H2S. For relatively highH2S, additive product FeS comes into being; at high temperatureuch as T = 150 ◦C, product FeO(OH) is found in the corrosion

[

[

cta 53 (2008) 3690–3700 3699

cale. From SEM images, general corrosion and some scattereditting corrosion on the surface of SM 80SS tube steel. Usinghe EIS measurement, at the conditions of PCO2 = 35.0 vol.%nd PH2S = 0.16 vol.% (1600 ppm), T = 90 ◦C in 3.5% NaClolution, i.e., there was relatively high content of CO2 gas, itan be seen from the fitted result that there is obvious capac-tive arc at high frequency and Warburg impedance appearslearly at low frequency, but the corrosion scales have poorrotection. While, at PCO2 = 20.0 vol.% and PH2S = 1.6 vol.%16,000 ppm), T = 90 ◦C in 3.5% NaCl solution, i.e. there is highontent of H2S, the polarization resistance is high meaning theow corrosion rate at 150 ◦C, the corrosion scale formation mech-nism is mainly controlled by activation, accompanying withiffusion control. The H2S corrosion has a significant effect onhe whole reaction process and iron sulfide is superior to pre-ipitating on the steel surface compared with iron carbonate.n addition, the surface scales of iron sulfide almost act as aiffusion barrier and inhibit the corrosion by a coverage effecttrongly depending on H2S concentration

cknowledgements

Financial support for the work by National Science Foun-ation of China (No. 50231020), and Key Laboratory Openingund of Corrosion and Protection of Tabular Goods Researchenter of China National Petroleum Corporation is gratefullycknowledged.

eferences

[1] A. Ikeda, M. Ueda, S. Mukai, Corrosion ’83, NACE, Paper No. 45, 1983.[2] S. Nesic, J. Postlethwaite, S. Olsen, Corrosion ’95, NACE, Paper No. 131,

1995.[3] S. Nesic, K.J. Lee, Corrosion ’02, NACE, Paper No. 131, 2002.[4] L.G.S. Gray, B.G. Anderson, M.J. Danysh, P.R. Tremaine, Corrosion ’89,

NACE, Paper No. 464, 1989.[5] L.G.S. Gray, B.G. Anderson, M.J. Danysh, P.R. Tremaine, Corrosion ’90,

NACE, Paper No. 40, 1990.[6] S. Nesic, M. Nordsveen, R. Nyborg, A. Stangeland, Corrosion ’01, NACE,

Paper No. 40. 2001.[7] E.W.J. van Hunnik, B.F.M. Pots, E.L.J.A. Hendriksen, Corrosion ’96,

NACE, Paper No. 6. 1996.[8] J.S. Smith, J.D.A. Miller, Br. Corros. J. 10 (3) (1975) 136.[9] D.W. Shoesmith, P. Taylor, M.G. Bailey, D. Owen, J. Electrochem. Soc.

127 (1980) 1007.10] L.A. Taylor, L.W. Finger, Carnegie Inst. Wash. Year Book, 69, 1971, p.

318.11] J.B. Sardisco, R.E. Pitts, Corrosion 21 (8) (1965) 245.12] J.B. Sardisco, R.E. Pitts, Corrosion 21 (11) (1965) 350.13] T. Murata, R. Matsuhashi, T. Taniguchi, K. Yamamoto, Offshore Technol-

ogy Conference, Paper 3507, 1979.14] K.A. Lichti, S. Soylemezoglu, K.D. Cunliffe, Proceedings of the New

Zealand Geothermal Workshop ’81, 1981, p. 103.15] P. Marcus, E. Protopopoff, J. Electrochem. Soc. 137 (9) (1990) 2709.16] A. Ikeda, M. Ueda, S. Mukai, Adv. Corros. 2 (1985).17] A. Valdes, R. Case, Z.M. Ramire, Z.A. Rui, Corrosion ’98, NACE Interna-

tional, Paper No. 22, 1998.18] B. Kaasa, T. Ostvold, Corrosion ’98, NACE International, Paper No. 62,

1998.19] J.L. Crolet, M. Pourbaix, A. Pourbaix, Corrosion ’91, NACE International,

Paper No. 22, 1991.

3 ica A

[

[[

[

[

[[

[27] E. McCafferty, Corros. Sci. 39 (1997) 243.

700 Z.F. Yin et al. / Electrochim

20] A. Carlos, T. Palacious, Y. Hernandez, Corrosion ’97, NACE, Paper No.20, 1997.

21] S. Srinivasan, R.D. Kane, Corrosion ’96, NACE, Paper No. 11, 1996.22] A. Ikeda, M. Ueda, S. Mukai, Advances in CO2 corrosion, NACE, vol. 1,

1984, p. 39.23] K. Masamura, S. Hashizume, J. Sakai, Corrosion ’87, NACE, 43 (6) (1987)

359.24] G. Fierro, G.M. Ingo, F. Mancla, Corrosion ’89, NACE, 45 (10) (1989)

814.

[

[[

cta 53 (2008) 3690–3700

25] B.F.M. Pots, R.C. John, Corrosion ’02, NACE, Paper No. 02235,2002.26] F. de Moraes, J.R. Shadley et al., Corrosion ’2000, NACE, Paper No. 30,

2000.

28] D.D. MacDonald, M.C.H. Mckubre, et al., Modern Aspects of Eletrochem-istry, vol. 14, Plenum Press, New York, 1982.

29] J.K. Heuer, J.F. Stubbins, Corrosion 54 (7) (1998) 566.30] C.N. Cao, J.Q. Zhang, Electrochim. Acta 35 (1990) 831.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具