Ccoreson Science 157 2019) 392-405, journal homepa Content lists available at ScienceDisect Corrosion Science www.elsevier.com/locate/corsci The formation of FeCOs and Fe30, on carbon steel and their protective capabilities against CO, corrosion at elevated temperature and pressure ‘Yong Hua’ Richard Barker”, Anne Neville” sont, , Shusheng Xu’, Yun Wang®, Wassim Taleb’, Jianbo Sun”, Lei Zhang’, “rasan of inert Suraes, Schoo of Mecham geo erty of Lt, Led 182, Unt Kien Soh of echanat and Berne Igri Chine Unt of Parlin. ni, Shanda. Cina Carson an rato Genre, Unt Sec an ety Bei 50 Kuyuan oa, Bei, Cina ARTICLE INFO ABSTRACT Rowers “This study inesigates he comosion performance of NES carbon sie elevated temperate (up C0 250°C) Fede and CO, partial pressures (up 28.5 br pCO). A detailed appraisal of haw the eoconon rots can protect Foo {gulnst general and lease corrosion Is peseted. The morphology and chemical composition of corrosion zcomaon products were determined using various microscopic and spectencople techniques, with locale corrosion Fes ein determine y sree proilomety An neree in temperate oe edtion in CO, pil presi favour the formation of protective mngnetite lays Tes theamoynamically more sable and more protective ‘than ron carbonate i these conditions 1. Introduction IHCOjgy + 26° Hig) + 2005) eo HO) + 2° Hi + 20H) ea Carbon dioxide (CO,) corrosion of carbon steel pipelines is 2 ‘common occurrence in the oil and gas industry. The CO corrosion mechanism has recently received renewed interest resulting in an im proved understanding ofthe anodie and cathode reaction pathways at low temperatures (< 90°C) [1-3]. Despite the CO, corrosion me- ‘chanism below 90°C being widely investigated, the electrochemical reactions and formation of corrosion products beyond 90°C have re- ‘ceived significantly loss attention. ‘When carbon steel is exposed to CO, environments, the main anodic reaction Which oceuts is the dissoltion of ion Peay > Fey + 26 o ‘The corresponding cathodic reactions below 90°C have been con: firmed to be hydrogen evolution from a combination of H", Hs00s, HCO3” and H,0 reduetion (noting that HCO, and HCOs- contribute to the cathode reaction via a ‘buffering effect’ whereby they dissociate at the steel surface, producing the H* fons which subsequently undergo hydrogen-evolution}: By + 26° > Hay) ea) 2HyCOyg) + 2 - Hye) + HCO; ev, "Corresponding author. ‘mal dares: ¥ uo Meds uk CY. Hu), ups: /do.og/10.1016/}.corsc.2019.06.016 ‘The formation of corrosion products is normally a key process inthe (C0. corrosion mechanism and thee presence ean significantly alter the rate and mechanism of corrosion. The corrosion products capable of forming on earbon stel vary depending upon the exposed environment and operating conditions, affording different levels of protection to the steel substrate depending upon their formation kinetics, chemistry and ‘morphology [4-8]. Falure to account for the presence of such products can reslt in overpresiction of the corrosion severity and a high degree of conservatism being integrated into CO, carosion prediction models. Research indicates that the most common corrosion product to form on carbon stel is ron carbonate (FeCOs) at temperatures up to 90 Cin (COz environments [9-12]. When carbon ste! is exposed to CO-sat rated solutions, Fe’ ions can accumulate within the process uid. Once the proc of the activites of Fe?” and CO,?~ ions teach and exceed a particular saturation limit, FeCOs ean precipitate via Eq. (2) Fe + 003 PCO) @ A FeC0 layer can reduce the corrosion rate by over one order of magnicude in some instances (4,11,13]. The formation of such a layer can block aetive sites on the steel surface and restrict the transport of electrochemically-active species [8,10,11,13,14]. Authors have also Received 30 September 2018; Received in revised form 2 June 2019 Accepted 17 Jue 2019 ‘Anilale online 19 June 2019, (0010:938K/ © 2019 Elsevier L.A rights reserved‘Table 1 Elemental compasiton of X65 steel (Ww. reported that FeCO, films can adopt a double-layer structure under specif conltions, consisting of an outer erysallie layer and an inner amorphous or nano-rystalline film. In this instanee, the protection ‘fforded to the substrate has been suggested to arise predominantly from the inner layer (10,15). In addition tothe formation of FeCOs, a number of addtional cor rosion products have been reported at temperatures below 90°C, some ‘of which include iron hydroxide (Fe(OH),), ehukanovite (e-{OH) C03) and magnetite (Fe50,) [15-19]. Toxeards higher pH and higher temperature, researchers tend to report the dominance of either FeO, Fes(OH_C0s oF FeyOs [15,16,18-20]. From literature, i¢ is evident that transitions to different corrosion products occur as temperature is increased, owing to a change in the thermodynamie stability of different compounds. At low temperatare (6070), the steel surface is typically free feom corrosion products, ‘except in specific circumstances where solution pH is high or the so lution is highly supersaturated with respeet to corrosion produets at the steel-eleetrolyte interface for prolonged periods [9,19,21, ‘As the operating temperature exceeds 60°C, numerous authors have Feported corrosion rates declining with increasing temperature, with lssolution rate typically peaking in the range 60-120°C, depending upon the exaet operating eonditions [4,11,16,21,22]. Crtieal factors influencing the sensitivity of the locaton ofthe corrosion peak include (CO> partial pressure [11,16,22], pH [23,2], water chemistry and flow ‘characteristics [25,26]. Typically, only FeCO, can be observed on the surface of carbon steel immersed in & NaCl solution saturated with COe ‘at temperatures of around 80°C, However, as temperatures rises be ‘yond this temperature, authors have reported both changes in the ap pearance ofthe FeCOs layer (in tems of thickness and grain sizes), as ‘well as transitions in the surface species formed, with Fe:Oq being ‘commonly reported. As an example, Shannon [27] evaluated the for: mation of corrosion products in geothermal brines. Ata pH of 4.8, no ‘corrosion products were observed at S0C, whilst FeCO, was recorded ‘at 150°C. Farther increase of temperature to 250°C resulted in a cor. rosion product comprising of 7086 FesO and 30% FeCO,, Increasing pH 075 resulted in Fe30s forming at both 150 and 250°C, suggesting that Increased pH and temperature favour the formation of FesO,. i Hi Gamat Swe 157 (2019) 92-405 In another study, Yin et al, (20) reported a change in corrosion product structure in their experiments, with finer PeCOs erystals and thinner layers being associated with superior eorresion protection as temperature was inereased from 50 to 180 C for carbon steel exposed 0 CO; environments. Again, above 100°C, the onset of trace amounts of FesO4 formation was reported in conjunction with FeCOx, This ob: servation was supported by the findings of Han eta. (17) who utilised srazing incidence XRD and TEM to study the protective layers formed fon carbon stel at pH § and 80°C in a COpsaturated NaCl brine sol tion. Their results showed that FeCOs was able to form under such conditions, however, the high local pH within the film created f+ ‘ourable environment for the formation of FeO, at the boundaries between FeCOs and the steel substrate. ‘Tanupabringsen et sl. [18] constricted Pourbalx diagrams for an Fe—COpH,0 system at elevated temperatures using thermodynamic theory. The diagrams were correlated with the observed corrosion products on carbon steel up to 250°C through consideration of surface DH within each experiment. For the 4 day’ experiments performed be tnveen 80 and 150°C, FeCO, and Fe,(OH),CO, were observed as the dominant corrosion products, with corrosion rate peaking at 120°C, whereas FeO, was detected at 200 and 290°C, Variation of exposure time showed that Fex{OH)2COs ultimately transformed into FeCOs, consistent with the caleulated relative thermodynamie stability ofeach compound, ipplementary experiments were performed by Tanupabringsen etal. [24] in another study over 20h to evaluate the susceptibility of eatbon steel to corrosion at temperatures of 80°C, 120°C, 150°C and 200°C. In these experiments the pH and dissolved CO, was controlled ¢ 0.020. to determine th isolated effect of temperature. Their fin: lings indicated that the corrosion rates reduced with increasing tem- perature and that the corresion product was exclusively FeCO3 below 150°C, but comprised of FeCO, and FeO, when the temperature reached 150°C, Despite the recent research into the CO, corrosion mechanism at levated temperatutes, there sil emains some questions regaeding the preferential formation of the surface products and their controlling factors. In addition, the formation of Fe304 and the extent to which it protects carbon steel from localised attack has not been thoroughly investigated ‘This current study is conducted over a wide temperature range of 90°C to 250°C in an effort to determine the role of simultaneous in- creases in temperature, CO2 partial pressure and pH on the corrosion produets and how these relate to general and localised corresion be haviour, Furthermore, studies are performed to isolate the effets of temperature and CO, partial pressure on FeCO; and FeO, formation. Fig. 1. Schematie of autolave setup.Gamat Swe 157 (2019) 92-405 rabie2 Totti fo const fer ans xp aus enperue i opti we i iat a a = coon 2 = = ee _ _ : : some : : . : rabies Tx for ons tet ne vrs CO al ese ronan ne Teme '© Mae, pam Radka nt A pene, Tol aw on ho ste : See _ _ = = ssa : [ : a ae a si bl pressurising to a specific O02 partial pressure at 25°C before heating to des ee ———C== i F tron ths suay are proved nae 2nd Table Te inal se ges ous § lution pH at elevated temperatures and COy partial pressures were 2 4 pete wing utente sotwre 25) and ae le proved Ty po on i ‘Table 2 and Table 3. le } Upon completion of each ts the samples were removed rm the eas 0% F autoclave, cleaned with distilled water and dried thoroughly. In ac am ‘a cavance wih AST Standard G18 (9), tne corn eds on = me mee the surace were removed wth Care's slat, alowed by Hsing Fig. 2 Average general corrosion rates of 6 carton ste over 48 of posure to CO,-satrated 4.91% NaCl solution as funtion of temperature Exact test conditions at each temperature are oie in Table 2 In the frst series ofthe experiments, the starting pressure and pH in the autoclave are allowed 10 evolve naturally With temperature (achieved by shutting in the system at room temperature and pressure before heating). In the second series of cess, the pH, CO» partial pressure and temperature are all controlled to isolate the individual ‘effets of CO2 partial pressure and temperature, and draw conclusions regarding the role of ech parameter 2. Experimental procedure 2.1, Material and methods ‘The chemical composition of arbon steel (API SL X65) used in this study is provided in Table 1, The mass loss samples were machined into ‘ses of 25 mm diameter and thickness of 5 mm. The stel sample su faces were wetground up to 600 grit SiC paper, followed by rinsing ‘with dstiled water, ethanol and drying with compressed air. Samples ‘were weighed immediately using an electronic balance (within an ac- ‘curacy of 0.01 mg). Two samples (27 cm) were exposed to a 660 ml solution within @ one litre eapacity autoclave in each experiment. Fig, 1 shows a schematic representation of the autoclave setup. The brine used in all experiments was de-aerated by continuously purging ‘with CO2 for a minimum of 12h, Two samples were placed on Polyether ether ketone (PEEK) sample holder within the autoclave Before transferring the test solution into the autoclave, all the lines were Fished out using high pressure CO», The prepared CO; saturated brine vas then carefully delivered into the autoclave, followed by With distilled water and ethanol before drying with compressed ar. “The corrosion rates were calculated by using Ea. (4) 70am aT @ where V¢ is the corrosion rate of the sample in mim/y, Am is the mass loss in grams, pis the density of the sample in gem, A i the exposed cea in em? and Tis the immersion time in hours 2.2, Surface analysis AIL SEM mages were acquired using a Carl Zeiss EVO MAIS sean hing electron mieroscope (SEM). A 20KeV accelerating voltage and ‘8mm working distance were used for all images. ‘The exytaline nature of corrosion products was determined using X-ray diffraction (XRD) (DB advance, Bruker) and obtained by em- ploying Cu Ka radiation over a scattering angle range of 2Theta; = 10 {70° The step size was 0.083 per second and approximately 49min were required per san, Raman spectroscopy vwas used to identify potential amorphous corrosion products formed on the surface which XRD would be inable to detect. An Ar ion laser with low intensity power (ess than 196) was used and approximately 40 min was required for each scan point. 23, Focused ion beam scanning electron microscopy and transmission tlectron microscopy ‘The corrosion product formed at the material interface in specific experiments was sectioned using 2 focused ion beam scanning electron microscope (FIB-SEM) and analysed further using transmission electron microscopy (TEM). More details relating tothe techniques applied and the microscope operating parameters ean be found in a previous pub: lieation. (5),tect ©) Gamat Swe 157 (2019) 92-405 Bist ct eee. Prete ~ yh des Car seed temperature @ Feces Fig. 3. SEM images ofthe corroslon produts formed on X65 carton steel surfaces after 4Bh exposure toa COrsaturted 4.9% NaCl brine at temperatures of (8) 10°C, (b) 150°C c) 200 Cand (d) 250°C. Exact test conditions at each temperature are outlined in Table 2 Intensity abt) Bebes toon ate, mm eae eet Postion Theta] (Copper) ig 4. XRD patterns ofthe corrsion pets formed on NBS scl surfaces after 4h exposure to 9 COvsturated 4:9¥t.% NaCl brine at temperatures ranging from 90 to 250°, Exact test conditions at each temperature ae out lined in Table 2 124, Non-contact interferometry ‘To quantify the localised attack on the carbon steel samples, an 1NPrasx 3D Surface Metrology System was used. The corrosion products formed on the surface were First removed by using Clarke's solution. Localised corrosion was the quantified through the implementation of ASTM Standard G46.94, [30] (eer carson ae ese Fig. 5. Average general corason rates of N65 carbon see! over 48 of ex posure to CO, saturates 4.9 wt. NaCl soliton at 200°C aa anction of ©O, Dota presse. Exact test conditions at each temperate are tine’ in ‘Table 2. Results 2.1, General corrosion behaviour of K6S with increasing temperature Fg. 2 presents the corrosion rates and corrosion produet mass of| X65 carbon stel exposed to CO, saturated 4:9 w.% NaCl solution for 48h at elevated temperatures. The results show that X65 has highestGamat Swe 157 (2019) 92-405 Peetu Fig. 6. SER images of the corrosion products formed on X65 carbon sel suraces after 48h exposure to a COysturatd 4.91% NaCl bene at 200°C and CO; pal presses of) 2:7 bar, (b) 6 ar (€) 154 bean () 28.5. act est conditions at each temperate are outlined in Sa ae woo] ES a Se -_ wd hace sl ansnomenerennrtiene Psiten -7That (Cope) Fig. 7. XRD patterns of the consion products fem on X65 easbon ste! sutfaces after 48D exposure to a COgsaturted 49yt% NaCl Brine at 200°C ‘and diferent CO, patil pressures. attest conditions at each temperaire are outlined in corrosion rate of 3.19 mm/year at the Jowest temperature of 90°C ‘General corosion rates of 057, 0.36 and 0.38 mm/Year are recorded for temperatures of 150°C, 200 C and 250°C, respectively over 48h of ‘exposure. The general corrosion rate decreases with increasing tem perature, with nearly an order of magnitude reduction being recorded by inereasing temperature from 90 to 200°C. Interestingly, the coro: sion product mass reduces in conjunction with the corrosion rate, signifying thatthe lowest corrosion rate is associated with the lowest corrosion product mass. The results suggest that with increasing tem perature, protective film formation is accelerated. The analysis of cor rosion product morphology and composition are described in the fol lowing sections 3.2. Analysis of corrosion product morphology and compostion ‘The reduction in X65 general corrosion rate in conjunction with increased temperature (Fig, 2) suggests that protective film formation plays critical role in the COs corrosion mechanism. Fig. 3 shows SEM mages of the soo! surface after 48h of exposure to each (est tem: perature evaluated in Fig. 2. After 48h, a dense and compact corrosion produet layer appears on the ste! surface at temperatures of 90°C and 150°C (Fig. 3(a) and (b)) which cover the entire steel surface. As the temperature is increased to 200°C, gaps between the crystals beeome evident and their coverage reduces. AC 250°C, crystals are randomly distributed on the steel surface and become fewer in number. Perhaps lnruitivey, st may be expected from analysis of the SEM images thatthe corrosion product formed at 250°C would be the least protective. However, the reality Is on the contrary, with the substrate having the lowest coverage of erystals possessing the lowest general corrosion rate lover the duration of the experiment. XRD was employed to determine the crystal structure ofthe various corrosion products formed on each steel surface and the corresponding Alitfraetion pattems are shown in Fig. 4. The patterns show that the dominant crystalline phase detected on all steel surfaces is PeOOs,Corrosion rate (mm/year) Conrasion rate mayear so’ "350° 200°C 250°C a7ben 1 General corrosion ate, mm/year Locales coresion ate, mmiyear @ ‘which relates to the large, cubic crystals clearly visible at temperatures ‘90°C, 150°Cand 200°C. Traces f FesO coexisting within Une FeCOs ‘corrosion produet were also detected at @ temperature of 250°C. It i possible that FesOs exists on the steel surface at 200°C, however, its ‘amount is below the detection limit of XRD. The lower corrosion rates at higher temperature, and the gradual transition from FeCOs to Fey0s suggests that the FojO, layer offers superior corrosion protection ‘compated 10 FeCOs. In addition, the Inetease in temperature up t0 250°C resulted in FeO, becoming the more stable eorrosion product ‘compared to FeCO, under these conditions. 38. The influence of CO, paral pressure on the general corrsion rates of 1X65 at 200°C ‘The effec of CO partial pressure onthe corrosion behaviour of X65, carbon steel ata fixed temperature of 200°C was evaluated, with the substrate mass loss and corrosion produet mass results provided in Fig. 5. Fig. 5 shows chat the general corrosion rate ineteased as CO- partial pressure increased. The lowest corrosion rate of 0.17 mm/year was recorded at 200°C and 2.7 bar of COs patil pressure. A general cor rosion rate of 0.41 mmy/year was recorded when CO. partial pressure Increased to 28.5bar a the same temperature of 200°C. This effect can atleast partly be attebuted to the seduction in pH achieved within ‘creasing CO. partial pressure, which increases the corrosivity of the test, solution by increasing the H concentration, whieh Is able co support the cathode hydrogen evolution rexction. ‘34, Analysis of corrosion product morphology and ther compositions SEM images for X65 exposed to diferent CO. partial pressures at @ fixed temperature of 200°C and immersion time of 48h are showin in Fig. 6. Traces of crystalline corrosion products (confirmed as FeCOs from XRD pattems in Fig. 7) an be observed at 2.7 bar CO partial pressure, as shown in Fig. 6(2), in conjunction with Foj0y, also de termined from XRD analysis in Fig. 7 Increasing CO. partial pressure to {6 dar (Pig, 6(b)) resulted in a greater quantity of FeOOs forming on the ste! surface. Gaps betwen the precipitated crystals are still evident a the CO2 partial pressure is increased to 15.4 ar (Fig. 6(€)- However, ‘a dense and compact FeCO, layer forms on the steel surface at a CO3 Partial pressure of 28.Sbar (Fig. 6(@)). The images reveal that in ‘creasing CO partial pressute as the opposite effect on the formation of FeCO3 compared to increasing temperature ie. reduced COy partial pressure and increased temperature both supress the formation of C03 and favour the formation of FeO Star Gamat Swe 157 (2019) 92-405 Fig. 6 Average general and localised comesion rates of X65 carbon ses over 48h exponie to 42 COrsnturated 492% NACI soltion a8 2 function of temperature with CO, pressure ranges between 13. and 15.4 bar al (b) C02 partial pressure at 3 fixed temperstre of DOC Exact test conditions a each temper ture ate oti’ in Tables 2 and, Ababa 28S bar ‘General cotesion ate, mye floclsedearosion ate, mniyear ) ‘The XRD patterns for the samples exposed to various OO. partial pressure fora fixed temperature of 200°C (Fig. 7) indicated that the «xytalline phases were FeCOs and Fes0« at 2.7 and 6.5 bar COs partial pressure, Larger crystals which are clearly visible on the X65 steel at 15.4 and 28.5 bar of CO, partial pressure were determined to be FeCOs, according to the XRD analysis in Fig. 7, [twas not clear whether FesOx coexisted within the corrosion products at these two higher pressures. 35, Loclied corrosion assessment One particular concern in such severe high temperature and high pressure environments is the accurrence of localised pitting corosion. Fig. 8 provides a comparison betwen the general and localised cor rosion behaviour of earbon steel across all test conditions outlined in “Tables 2 and 3. Fig. 9 shows the morphologies of the localised/piting tak at various temperatures and COs partial pressures. ‘The localised corrosion rates shown in Fig. 8 producea similar trend with increasing temperature and pressute tothe general cortosion rates presented in the same figure. From Fig. 8a), the localised corrosion rate was found to reduce with increasing temperature from 9.2 mm/year to 0.9 mmvyear as temperature was increased from 90 to 250°C. Fur- thermore, itis interesting to note thatthe sizes and depths of the To calised/pitted regions decreased as temperature was increased from 90°C to 250°C, as shown in Fig, fa) and 91D) Fig. 8(b) indicates that as CO, partial pressure was increased at a constant temperature of 200°C, the extent of localised corrosion in: creased with the general corrosion behaviour, with localised material loss rates increasing from 0.74 to 3.7 mm/year as COs partial pressure was increased from 2.7 to 28.5 bat Large regions of localised attack can be observed after removal ofthe corrosion products on the surface as shown in Fig. {e) and (). In all the tests performed, the localised corrosion rates are ap: proximately two or thee times higher than the general corrosion rates lover this particular period of time, Iti important to stess that these measurements provide an indication as to the extent of pitting/localised attack over 48h only and should not be extrapolated to long term be haviour without further analysis. Nonetheless, they provide a compar {son of the severity of each envionment over the initial stages of lo calised attack within 48h, 3.6 Isolating the roe of temperanure (ne aspect which is not fully clear from the previous tests isthe role of solely temperature on the formation of FesO« and how this influences the ratio of FeCOy/Pes04 on the steel surface. To provide someYt, at Gamat Swe 157 (2019) 92-405 retort ip rare 250°C and 15:4 bay fee ea errs Fig. 9, Examples of proflometry and SEM images ofthe localse/pitng attack denied afer removal of coroson prac on X65 carbon steel surfaces at various temperatures and 0; partial pressures of (a) 13.3har a 90°C, (b) 15.4 a 250°C, (2) 15.4 bar a 200°C and (2) 28.5 ar at 200°C.Yt, at Gamat Swe 157 (2019) 92-405 Sead ate @ Fig. 9. contin)Yt, at era Gamat Swe 157 (2019) 92-405 Fig. 10, SEM images of the corosion products formed on XS carbon steel surfces afer AB exposure toa COpsaturted 49449 NACI Brine at pH 4.9 and temperature of (a) 90°C and (b) 250°C Fig. 11, SEM images ofthe corrosion products form on XS carbon steel surfaces ater 48h exposure toa CO, saturated 49 wt. NaC brine at pH SA and 200°C ‘with CO, partial pressures of a) 27 bar and (b) 285. Indication as to the individual effet of temperatute, the role is con: sidered here through an additional experiment at 250°C with the pH maintained at 4.92 and a similar partial pressure to the test conducted at 00°C ealier (13.3 ban). ‘The result from the additional test in comparison to the previous ‘experiment at 90s shown in Fig, 10. Increasing temperate from 90 to 250°C whist maintaining pH at “4.9 results in @ reduction in €or. rosion rate from 3.2 to 0.9 mmnyear. However, this not as substantial a reduction as when pH Is inereased from 4.9 to 6.3 t 250°C, where the corrosion rate was reduced fo 0. mm/year. The SEM images in Fig. 1 show that solely increasing temperature also reduces the amount of FeCO3 on the steel surface and promotes FesO« (validated by XRD), in ‘agreement with the findings of Tanupabrungsun etal. (18) through the development of thelr Pourbaix diagrams. Furthermore, compating 0) with Fig. 32), it ean also be seen that increasing pA at ‘constant temperature of 250°C favours FeO. formation and supressed the precipitation of FeCOs, 37, Isolating the role of CO2 paral pressure similar analysis to that with temperature can be made to show the role of solely COs partial pressure changes on the formation of FeCO, and Fes0s, Additional experiments at 200°C with the pH maintained at 5.4 and COy partial pressures of 2.7 and 28.5 bar are performed in this section, The general corrosion rate behaviour of X65 are shown in jn conjunction with SEM images of each steel surface. Inter estingly, the signlfeant inetesse in COs partial pressure has little effect fn the general corrosion rate, increasing it from 0.8 to 0.4 mm/year. However, comparison of the SEM images in ig. 11 shows that there isa change inthe proportion of PeCO onthe steel surface asa result of the partial pressure increase i. ineresing partial pressure increases the ‘quantity of FeCOs on the steel surface. ‘38, Fuster analysis of FeCOs and FesO4 formation on XOS steel surface at 200°C ‘To further understand the chemisty ofthe corrosion product layer developed on the steel surface at 200°C (the experiment which pro- duced the lowest general corrosion rate), a combination of Raman spectroscopy and FIB-SEM/TEM was employed, The use of Raman spectroscopy enabled the determination of the corrasion procict film locally on the X65 steal surface. The spectra provided in Fig, 12 relate to two scans conducted at the surface; one on a large cubic crystal as shouin in Region A and one within Region B. The spectrum for Region A shouts «wo main peaks located at 290 and 1086 em", confirming the crystals. as FeCO,, The two main peaks within the spectrum for Region B located at $43 and 668 em™? confirms the presence of Fes0s between the FeCO, crystals. It is this layer of PesO, which appears to be largelyGamat Swe 157 (2019) 92-405 000 21000 Region A~ ret soo reco, at 1086 «0000 al 2 a ae Feso.at 668 3 3 m0 Tate 2 000 Ebrccosst200 i. 0000 ae 3000 ° 2000 100 oo u00 100 ao 1100 3600 Raman shife (em) pee Ceres Te :S) Fig. 12. Raman spectra of the corrosion produets formed on X65 surface at 200°C for Region A i scan of an FeCO, crystal, whist Reyloa Bis a scan of a fine erst responsible forthe increased general and localised corrosion resistance ‘of X65 under these conditions, given its dominance as surface cov erage when compared tothe FeCO, crystals, 29, FIR-SEM and TEM analysis ‘To complement the Raman spectroscopy analysis, FIB-SEM/TEM, was used to further characterise the eorrosion product formed between eC crystal and substrate. The sample exposed to 200°C and 2.7 bar ‘of CO; partial pressure for 48 h was selected and two locations covered ‘with different corrosion products were selected, as showin in Fig 13 Figs. 13(a) and (c) shows SEM images of the two selected areas of the sample surface. The FIB samples prepared from each of these two rogions were attached to Gu TEM grids (Fig. 13(b) and ()). The FI. ‘SEM image in Fig. 15(b) shows that the interface bowen the FeCO3 ‘rystal and the steel substrate Is separated by a thin interfacial layer (approximately 200 am thick). Based on the analysis of the eross-sec tion from Fig. 1344), the thin layer Is present across the entice steel surface as well as below the FeCO, crystals. Therefore, the low general ‘corrosion observed under these conditions can be attributed this thin layer (identified through Raman analysis as Fes03), Further analysis ofthe interlayer between the FeCOs erystal and the substrate depicted in Fig. 13(b) was performed using selected area aman shift (em) Cee) S 4handa COpsaturted 4.9% NaCl solution a 27 bar CO; patil presure; line FeO, layer electron diffretion. Fig. 14(a) provides a higher magnification TEM Image ofthe sample shown in Fig. 19(b) Fig. 14(b), (€ and (€ indicate the local electron diffraction patterns from different arens across the sample. Based on the analysis ofthe difiaction patterns, the interlayer is shown to comprise crystalline and nano-polyerystalline FesOs Fig. 15(@) provided a higher magnification TEM image of corrosion produet imaged in Fig. 13(). Indexing ofthe diffraction patterns pro {luced from the thin corrosion layer observed on the sample surface (ig. 15(6) produced ¢-spacing values characteristic of FeO, The EDX line scan analysis in Fig. 15(€) corroborates with the difraction pat ‘ems collected from the TEM in that only Fe and O can be identified from the corrosion product layer and have an atomie ratio with one nother which beats a strong resemblance to FeO, 4, Discussion 44.1. Propose film formation mechanism Referring to the literature, ic well known that in @ C02 environ: ment, the solution pit at the stee-electrlyte interface can be sig nificantly higher than that of the bulk solution (811. This sas a result of the flux of Fe? from the steel surface and the consumption of H* as a result of the eathodie hydrogen evolution reaction atthe steel surfaceoon See 157 C19) 92-405, Thin separating layer Fig. 13, SEM images and FID prepared cos etions of wo regis on the surface of XS cron tel exposed to a CO, saturated 4.9¥L.% NaCl solution af 200°C for 48h with a CO. paral pressure of 27 bar (171, According to measurements by Han et al. [21] for @ bulk solution 1 pit 4, the surface pil was recorded as being approximately 6, How ‘ever, asthe bulk pH Increased 6.6, the dscrepaney between the bulk ‘and surface pH significantly reduced. A surface pHi of “6.9 was mea sured under these conditions (determined through the application of imesh capped pH probe), Inthe experiment performed within this study at 200°C and COs partial presure of 2.7 ba, the bulk solution pH was calculated at “6.4 Although the surface pH will be higher than this, based on the mea: surements performed by Han et al. (21), the anticipated surface pH wil be higher than this. Furthermore, the significant suppression af corzo- sion rate as a result ofthe Fey04/FeCOs fllm minimises the surface fx ‘of Fe2* from the steel surface, as well as the consumption of H*. This means that their effect on the local pH is also minimised once the protective layer becomes established, reducing the discrepancy be: ‘neeen bulk and surface pt In View of the TEM images and the SEM images at 200°C, itis suggested that the higher local pH at the stel-clectrolyte interface results i the formation of the thin Fes0. layer via Reaction (5) and (6) (noting that these reactions produce H* ions, potentially acidifying the Interface during formation) CO; + Hig) > FeO) + COQ) + 2A) +20 6 AF 4 AO > FO + By #2 . ‘The reaction Is followed by a decline in interfacial pH (due to the reduction in corrosion rate and also potentially the production of H* as 8 result of FeO, formation). In terms of other experiments performed in this study, as the op- erating conditions change, the theemodynamie stability and the kineties of FesOq and FeCO; evolves with the operating conditions. Consequently, the corrosion produet which dominates the surface clearly depends on their relative thermodynamic stability in the en: vironment, but alsa the kinetic of formation of each corosion product (ushich is governed by temperature and CO2 partial pressure), 4.2, Isolating the roe of temperacure Fig, 10 ilustrates the effect of temperature on the formation of corrosion products more clearly between 90 and 250°C for a constant (COz partial pressure of 10bar at 25°C. At a temperature of 90°C, the stable state of corrosion product is FeCOs, As temperature is inereased te 250°C, FejO, becomes thermodynamically more stable relative to FeC0s, Based on Fig. 8 and Fig. 9, a reduction in general and localised corrosion rates is recorded as the temperature increases which can be related to the protective eapabilityof the corrosion produets (FeCO3 orPeery ee ey © 14. (@) TEM image of coresion produ in area electron difaction patterns corresponding to Regions A Band C, respectively. Fes04) formed on the steel surface. Numerous authors report that eCO) is able to block aetive sites on the steel surface as well as acting ‘8s a diffusion barrier to electrochemically active spectes. [413,14] However, the lowest corrosion rate was observed at 250°C when only trace amounts of FeCO, were apparent on the steel surface and Fej0, was present in detectable levels sing XRD. The rests suggest that the formation of Fe,0y appears to be more effective at reducing the sus ceptibility of the surface to general (and localised corrosion for that matter) as opposed to the crystalline FeCOs layer. 4.3. boating the role of CO» partial pressure Similar observations in terms of the reduced general and localised ‘corrosion rates as well asthe corrosion product coverage on the surface ‘were observed a8 COp partial pressute decreased from 28.5 ar to 27 bar, Fig, 1 illustrates the effect of CO. partial pressure on the formation of corrosion products between 2.7 and 28.5 bar fra constant temperature of 200°C. The increase in CO, partial pressure appears Co make the formation of FeCO more favourable relative to FeO on the surface. The observations suggest that the inereased eoverage of FeCOs was attributed to the increased CO> partial pressure and that the Gamat Swe 157 (2019) 92-405 eee Diffraction pattern for Region C Paonia @ 1), denying regions where sect aren elcton dation patterns were collected (D/C selected formation of FesOx appears to be more effective at reducing the sus- cepubiity of the surface to general and localised corrosion. 5. Conclusions ‘The corrosion behaviour of X65 carbon steel in CO, saturated brine was studied at temperatutes from 90°C to 250°C and COs partial pressures between 2.7 bar and 28.5 ba in experiments of 48h duration. From this study, the following conclusions can be made: AA very dense and compact FeCOs crystalline layer was detected at 90°C and the corosion products were mainly comprised of FeCO3 and FejOq at 250°C. Localised corrosion was also shown to reduce with Increasing temperature, ‘The lower C2 pressures favoured the formation of FeO, while higher CO. pressures favoured the transition to the FeCOs phase. ‘The superior corrosion resistance at lower OO. pressures and high temperatures was attributed to the thin FeO. layer Increasing temperature and pH promoted Fe,0, formation, while the increased precipitation of FeCOs is observed as a result of increasing (00s partial pressure.Yt, at ey 070 60 050 040 030 ono 00 ‘Atomic fraction so Gamat Swe 157 (2019) 92-405 ca Regio Postion fm) © Fig. 15. (a) TEM image ofcozosion product in Fig. 10(4), Kenting the region where the selected area electron diaction pattern and the EDX lie ean were colecteds(h) selected area electeondifatin patter corresponding to repion A; (€) EDX line sean result trough the substrate and cresion product aver References UU) Ame, NCCE. Bandera RA More, OR. Matos New isis on he ae of Go nthe meshanem of aron see coon, Care S120 (2017) 299-50, 121 T Almas, ML Bonet, e Mora, 0. Matos Deusen on" oy cerson a mild ste: ee f CO, 0 rn dain ean” BA tyra, Brow Nese, Comoe Se 129 2017) 14615) Carson Sees, aonb isk pai ‘A Kanyanan, Brow, 8 Nes, Bcroehemisy of COs corson of mils: Stet of 60, on Uisolton etn, Cts. S129 (2017) 146-15, ‘Sun and 3 Nese, Bess eve kino eon carbonate sale precipitation 1a CO? crrxon.caneson 2006, paper, 20060535. Fare M Cll Brown, 8 Nate He Catal Evlutoa of dselaton recs a teint of xb tron 0, enone sued by HS, onos. Sc. 82 (2) QO10) 500-517, Yeu R Backer CT, M Ward A Neve Rlatng Won conse mpg 1 Fur 98201 183-103. Yue, Raker, A Neve et of mperaute a heal water conten for eer and lc cmon of X55 noms i te apo of prs {ya J: Greet. Gas Cool 31 2014 43-6 [We Pang, Cis, KW Gao, Prmaton mechan and protective propery ti cenon pct sean 370s unde prec CDs ences, (Garo Seto (015) 404-420. ‘A Dugsat, The inpoance of FeC0S supersaturation on the Coz fad Seas CORROSION paper, 19820) Goo, YX Pag. Zag Qi, W. Ch, My, Meehan popaties of (oy cote po ser tn ato face rt Care SCL So uo) ean) 296-290. ‘Yc. Png X- Png, PQ, Xi, KM Gao, custo of he 0. coon Ineo enero prt pesre and super coat, Ces 59 (2012) 18697 Bingham, Ko, Layeoc J Burwel . Kapp, Kington, . Winns sta "anehraron Xray station dof eae tomato rng CO. coon 0 {atbon steals and magnesium choi solos, Coes. 8 36 C2012) @ "1 3 tr "1 6 » no, an nay as) a0 a5) a6 on os) a en eu a es ea es) o-04, ‘Hus Barer, A Neve, Compas of cores bao for X48 cabon ‘esl inspereicl COyattursted water and warr-iaed/ ans oipe ‘fica COs Superet Fae 97 (2015) 224-297 Sinn Wig We hang, 2H. ang 7 To AX Ly Characters eemation mechani of eosion season cron X65 ste in COs ‘rome Acta Metal Sin 45 (1) (2009) 84 {Palco J Shey, Chrcterstsof orson sales om sels ia a Ost fated Natl eine, Caren 4 (2) (191) 122-127 ‘tha MU, Asa, C2 behave of cro a trian ee (tone coverage) ence in Dy Comon 1 08) 2851 ‘Han, D Yura I ali, A Tet 8. Neb, Chersty and stuctre of the sie fit on ml tel CD oso eaionment In Eg Che es. $8 (3) e05 e2e-aa0a "Tanapabrargn D. Young, Brown, S. Nei Consrcton and Verein of Porta Digs oe OD crrson of Mb Stee! Va wp 0 250 In De mare, 2 Jang. Pj Eom, An a tu syncroen raiatin fang inldece Xa) deacon suo ton lc cri etree oe 152 (10) (205) R39. 1952 Vin ¥ Fog. Zo, 7 al, Ln fc of temper CDs cmon of Earbor tel Src and Interface Ass At nt J devote Dv. on AL “eet. of surtiens traces ad hn is 4 (6) (200) 517-52 ‘Tara MC Ste, Hemmingsen fr of Fe, superar and ‘atid expo on te 0, coma te xo ts, Coon 69 (6) 20) santa ML opal. appa and R Zang, Modeling te Dison Bets Trough te tro Carbonate Layer in the Caton Dade Cnrnen Crh Stl, CORROSION 8 NACE ner: San log, CANCE Fess, Rr, Nevibe, Theale of pon loan ease behavior {81X65 carton sta In 00st bps, Corson 71 (12) 2015) 1482-146, ‘.Tandabvungsi,B Brow ao Nes, let of pH on COZcroson of mld Sta at etd operates: CORKOSION 2013, pepe, 201888, Thee GS ney, Campton et dad i owYt, a (26) 5 Nei S. Wag J Ca¥. na, negate > Conon multiphase How Mel In 0 lnematona!Sympetam 00 Oli Cnet, Scetyo eta Ensncrs 2004. fa iS, af hemi cones nt it ao (28) Blutisele 7.11 commer stare package om Bxpo Group Items fox move atin se pie Gran Smee 157 2019) 392-405, (29) AsmM, supine 61-48, standard Practice or Preparing and Bvaluatag ors Test Specs, ASTM nteroatonal West Cansaocen, PA, 2003, (0) AS, Sa 454, Sad ae er amin rin of Ping (a1) Sim BA me You Nes Mesh cp pe des or is mesures tn cel ering etal suc, AL Hecocem 40 (3) (Goto 68400.