Stress corrosion cracking and localized corrosion of downhole tubular steels in petroleum fluid containing CO2/Cl−

There are a number of different techniques capable of being used to measure corrosion within equipment. The most simple, the use of metal coupons, usually causes the process to be shut down, is manpower intensive, and has a time delay in getting the required corrosion information. Electrical Resistance (ER) techniques are often used but their response is very sensitive to temperature and they cannot differentiate between general and localized corrosion. Electrochemical techniques, such as linear polarization resistance (LPR), electrochemical noise (EN), electrochemical impedance spectroscopy (EIS), harmonic distortion analysis (HDA), and electrochemical frequency modulation (EFM), have the capability of solving most of those drawbacks. Electrochemical probes can be mounted permanently in most equipment, give regular measurements of the intensity of corrosion, and some can detect localized corrosion. Of all of the electrochemical techniques, EN has the most potential for being used successfully to measure general and localized corrosion rates of equipment. The EN technique was studied in the late 1970s and early 80s as a means of detecting localized (stochastic) corrosion phenomena, such as occurs with pitting, crevice and cavitation attack. EN measurements are based on fluctuations in electrochemical potential and corrosion current that occur during corrosion. Electrochemical potential is related to the driving force (thermodynamics) of the reaction, while corrosion current is related to the rate of reaction (kinetics) of the reaction. The idea is that random electrochemical events on the surface of a corroding metal will generate noise in the overall potential and current signals. Each type of corrosion (for example general corrosion, pitting corrosion, crevice corrosion, and stress corrosion cracking) will have a characteristic “fingerprint” or “signature” in the signal noise. This “fingerprint” can be used to predict the type and severity of corrosion that is occurring. By comparison, conventional electrochemical techniques such as LPR, EIS, HDA and EFM rely on a steady-state analogy for the determination of general corrosion rates. Early studies were carried out using potential EN measurements, using time domain, statistical and frequency domain analyses to characterise the electrochemical response of systems undergoing localised corrosion. Current EN measurements followed quickly using zero resistance ammetry to study the current noise between two identical electrodes. For general corrosion processes, EN has been demonstrated independently by several workers to provide information similar to LPR. Noise technology has been used to study systems undergoing very low to very high rates of corrosion, for example, coatings performance, passive systems undergoing pit initiation/propagation, condensing systems, systems undergoing stress corrosion cracking, and general corrosion through to the very high corrosion rates experienced during chemical cleaning processes. This review will describe: state of the art methods and probes used to measure EN, data acquisition requirements, theory to analyze the signal and to relate the signal to corrosion rates and types, the results of EN field trials, and laboratory results in environments similar to gaspipelines.

This review considers the problems of general corrosion and stress corrosion cracking for Fe-Cr-Ni alloys in caustic environments.Environments considered are primarily NaOH-H 2 0 over a broad range of temperatures.Information is presented in areas of thermodynamics, electrochemical kinetics, corrosion rates, SCC phenomena, structures of passive films, and inhibitors.-Materialsconsidered are iron, chromium, and nickel, as well as binary and ternary alloys. ii SUMMARY AND CONCLUSIONSThis review considers the corrosion behavior of Fe-Cr-Ni alloys in caustic environments, which are mainly aqueous solutions containing alkaline metal hydroxides (Na,K,Li).There is no specific information herein on anhydrous fused salts nor on sodium-based solutions.This review was prepared in support of the liquid metal cooled, fast breeder reactor program from the point of view that any accidentally spilled or leaked sodium would oxidize or hydrolyze to produce NaOH-HaO-H 2 0 solutions.These solutions, in a general way, are very aggressive toward this alloy system and particularly toward iron base alloys.A series of soluble species is formed at higher pH's which have identities like HFeOg and FeOf.The corrosion of these alloys in caustic environments takes the form of either (a) general corrosion or (b) stress corrosion cracking, depending subtly on specific features of the environment.The review considers first the pure materials, iron, nickel, and chromium.This is followed by the binary and then the ternary alloys.Each section considers stress corrosion, general corrosion, and electrochemistry of the respective material system.Structures of protective oxide films are also discussed.In addition to considering the general technological and mechanistic behavior, methods for prevention or diminution of both general corrosion and SCC are described.With respect to stress corrosion cracking there is a paucity of data compared to chloride-induced SCC.Important general observations from this review are summarized as follows:1. Stress corrosion cracking becomes most severe for iron base alloys.This is contrary to the trend for chloride SCC where the very iron-rich alloys are immune.2. In general the effects of temperature, alloy, stress, and caustic concentration are critically interrelated.In iron SCC will' occur at temperatures as low as 70°C 0 The detailed nature of this interrelationship is not completely clear but these variables produce results in expected directions.Susceptibility to SCC is decreased as the total Ni and Cr content is raised, as temperature is reduced, as stress is lowered, and as the [0H~] is lowered.3. The onset of caustic SCC is shown to be critically related to the electrochemical potential.The mean potential at which caustic iii SCC occurs for iron is generally independent of the temperature and caustic concentration.This range of potentials is associated with the anodic peak of the potential-time curve.A secondary range of cracking occurs near the transpassive transition.These patterns are reasonable in terms of patterns observed for other alloys.h.The caustic SCC, in general, appears to be dominated by electrochemical parameters.The initiation and propagation of stress corrosion cracks appear to be related critically to transient anodic processes; a reducible species (0 2 , H 2 0) is required for cracking to be sustained.5. The information on general corrosion rates and electrochemical parameters will be useful in estimating rates of general corrosion of Fe-Cr-Ni alloys in caustic solutions.iv IV CORROSION OF NICKEL 6l A. Corrosion under Stress 6l B. General Corrosion 6l 1. Corrosion in Sodium Hydroxide 6l 2. Polarization Behavior 62 V CORROSION

During operation of oil- and gasfields the new developed weldable low carbon martensitic 13% Chromium steels may be subjected to local corrosion by chloride containing wet phases. Crevice corrosion at defective welds, coatings or flanges may contribute to unexpected failures and shutdowns at temperatures and potentials below that of local pitting corrosion. In addition to established standardized test methods, crevice corrosion susceptibility can be quantified by measured corrosion starting times and corrosion currents. In the present investigation, both parameters are monitored by a „Remote Crevice Assembly“ (RCA) providing quantitative consistence between local corrosion rates and measured net corrosion currents at a cathode/anode area ratio of 45000/1. Two materials with different Nickel and Molybdenum contents are tested at free corrosion conditions in CO2 purged formation water and nitrogen purged artificial seawater. Increasing potentials at test start are providing reduced crevice corrosion incubation times and increased corrosion currents, depending on fluid and type of steel. Weld simulating quenching from 1000°C of the anode provided stronger crevice corrosion only at the steel with higher Ni and Mo alloy content, which, otherwise was less crevice corrosion sensitive in the as delivered state as compared to the lower alloyed steel. Increasing temperatures of the cathode and the anode resulted in strong increases of corrosion currents. However, at temperatures above 40°C, pitting corrosion started at the cathode which resulted in reduction of crevice corrosion net currents. As a conclusion, the susceptibility to crevice corrosion of 13%Cr martensitic steels should be considered, also in possibly unintended pipeline operations and can be quickly established by short RCA tests.

The present study was designed to evaluate the long-term corrosion behavior of 3Cr and 13Cr steels in a single-phase flow environment containing aqueous carbon dioxide (CO2), hydrogen sulfide (H2S), and dissolved chlorides (Cl-). Long-duration tests, each lasting 40 or 100 days, were conducted using a flow loop equipped with a Thin Channel Flow Cell. The experimental setup aimed to replicate field conditions with varying H2S concentrations over time and included surface conditioning in NaCl brine before exposure to the corrosive solution. The results indicated that 3Cr and 13Cr steels exhibited different uniform and localized corrosion behaviors. The 3Cr steel exhibited average corrosion rates ranging from 0.03 to 0.12 mm/y, depending on specific test conditions and H2S concentrations. A beneficial effect of preconditioning was noted, potentially related to the formation of corrosion product layers with higher amounts of chromium rich oxides. However, localized corrosion was found in 3Cr steel when exposed to an increased H2S concentration of 100 mbar, even if its surface was preconditioned and the H2S concentration was increased gradually throughout the test duration. The 13Cr steel demonstrated a more stable corrosion rate, maintaining an average rate of approximately 0.01 mm/y, even with increased H2S content. However, minor localized corrosion was still detected at low H2S concentrations.

A range of Fe-Cr-Ni alloys are used in nuclear power plants due to their high mechanical strength and corrosion resistance. As a major component of these alloys, Ni contributes to the formation of austenite microstructures. In many nuclear reactor applications the alloys are exposed to ionizing radiation (g-rays) that can continuously decompose water to a range of highly redox active species (such as •OH, H2O2, O2). The change in water chemistry induced by radiation can affect both the general and localized corrosion behaviour of these alloys. In particular, its influence on the susceptibility of an alloy to localized corrosion such as crevice and stress corrosion cracking (SCC) are important for assessment of reactor component aging. As well, corrosion products that dissolve into the reactor coolant can be transported to the reactor core where they can be neutron activated. This can create a radiological hazard for reactor maintenance personnel. Thus, quantifying the rate of general corrosion and the corrosion product transport in a reactor are important safety, operation and maintenance issues. It has been well established that the type of oxide that forms on a surface is a critical factor in determining both general and localized corrosion behaviour. Both alloy composition and water chemistry affect the type of oxide that forms. Systematic studies that examine the effect of alloy composition on corrosion are rare and those available have been conducted mostly from a metallurgical perspective. There are even fewer studies that examine the combined effects of alloy composition and corrosion environment. A fundamental understanding of the mechanism by which the Fe to Ni ratio in a steel alloy influences the corrosion rate in different environments will provide valuable information that can be used for alloy selection and usage. We have investigated the effect of the Fe to Ni ratio of alloy on oxide formation and growth kinetics. For this study, customized steel alloys with different Ni contents and constant Cr content (18 ± 1 %) were prepared. The Ni content ranged from 15 to 25 wt.%, within the range between AISI 316L and Alloy 800 (Fig. 1). The corrosion kinetics was investigated using a combination of coupon corrosion tests and electrochemical measurements. These measurements were augmented by oxide surface and depth analyses using several spectroscopic and imaging techniques, and by post-test solution analyses for dissolved metal loss. We found that the oxide formed at pH 10.6 and 80 °C has a graded layer structure, consisting of a mixed spinel oxide (FeCr2O4/Fe3O4/NiFe2O4) as an inner layer and Ni(OH)2 as an outer layer. The increase in Ni content from 15 to 25 wt.% does not affect the inner layer but increases the thickness of the Ni(OH)2 layer. The increase in Ni content also decreases the total amount of dissolved metal loss. Gamma-irradiation of the corroding system increases the corrosion potential on these alloys. Correspondingly, gamma-radiation accelerates the formation of passive oxide layers and decreases the overall rate of metal dissolution. The observed kinetics of oxide growth and metal dissolution can be explained by the competition kinetics of oxide growth and dissolution for oxidized metals. Figure 1

During localized (crevice and pitting) corrosion, a local cell is established between an anode within a crevice or pit and a cathode on the surrounding passive surface. Data are presented to show that concentrated acidic chloride solutions, simulating corrosion product hydrolysis within a crevice or pit, produce potentials which are active (negative) to the normal surface passive potential. This behaviour explains the previously observed active drift of corrosion potential after initiation of crevice or pitting attack in dilute chloride solutions. The active state in concentrated chloride solutions was quite noble (positive) compared to the active state in more dilute solutions. Thus, there is no need to invoke ohmic resistance effects to account for the active state within a crevice or pit. Experiments were devised in which the local anode within a crevice was physically separated from the nearby passive-surface cathode. When the two were coupled together electrically, the cathode surfaces were polarized nearly to the unpolarized local anode potential, with only a few millivolts anodic polarization at the anode within the crevice. The rate of localized corrosion appears from the data to be limited by the rate of dissolved-oxygen reduction on the cathode surfaces. Thus, localized corrosion in dilute chloride solutions will be increased by (a) raising the temperature, (b) adding an oxidizer such as Fe 3+ ions, or (c) substituting external anodic polarization for dissolved oxidizers. The overall potential, E corr acquired by a specimen undergoing pitting or crevice corrosion is demonstrated to be near the protection potential, E p below which pitting corrosion cannot propagate. Any potential active to E corr and E p results in cathodic polarization and suppression of the anode reaction in a crevice or pit. Since both E corr and E p vary with the extent of previous localized attack, E p is not a unique property of the alloy as has been sometimes suggested and is of limited value in classifying alloy resistance to localized corrosion.

In recent years there has been an increased interest in drilling deeper geothermal wells to obtain more energy output per well with the corresponding higher temperature and pressure and increased corrosiveness of the geothermal environment. To explore the potential of the high alloy austenitic stainless steel UNS S31254 in future deep geothermal wells corrosion testing was done in simulated geothermal environment at 180°C and 350°C with a pressure of 10 bar. The simulated environment was composed of steam with H2S, HCl and CO2 gases, with a pH of 3 upon condensation. The testing was done in a flow through reactor for 1 and 3 week exposures. The stainless steel UNS S31254 performed well at 180°C with negligible corrosion rates both for the 1 and 3 week tests and no localized corrosion damage detected. After the testing at 350°C localized corrosion and substantial amount of NaCl crystals were observed on the surface of the samples. Microstructural and chemical composition analysis revealed large cracks in the cross-section of the sample most likely due to chloride induced stress corrosion cracking. The measured corrosion rate for the 1 and 3 week test was 0.024 mm/year and 0.24 mm/year respectively.

Localized corrosion describes dissolution processes concentrated at specific areas on the surfaces of metals. In some types of localized corrosion, enhanced dissolution rates arise from partial or complete destruction of the protection normally afforded by the passive oxide film covering the metal surface. Oxide breakdown can be due to mechanical rupture (stress corrosion cracking), the chemical action of aggressive anions such as chloride (pitting corrosion), the impaction of solid particles on the surface (erosion corrosion), or the concentration of corrosion products within small solution-filled gaps (crevice corrosion). Other localized corrosion processes are initiated at metal compositional inhomogeneities such as grain boundaries in alloys (intergranular corrosion), or interfaces between dissimilar metals (galvanic corrosion). The economic impact of all forms of localized corrosion is severe. For example, pitting and stress corrosion cracking together account for about one fourth of equipment failures in the chemical process industries. Metal dissolution rates during localized corrosion are high enough so that large concentration or potential gradients are typically found near the dissolving metal surface. Characterization of these gradients is a necessary precursor for understanding the mechanisms controlling the corrosion rate. Thus, experimental research on localized corrosion has always been closely coupled to quantitative analysis of mass transport processes by mathematical modeling. In this chapter, three examples are presented which illustrate the range of models applied to localized corrosion processes, reflecting the particular interests of the authors. Section II, written by Hebert, is a review of recent work on the modeling of pitting corrosion. The remainder of the chapter communicates results of recent work by Tribollet on galvanic corrosion (Sect. III) and on the simulation of the impedance in crevice-type geometries.

A 3-week corrosion testing of UNS N06625 was conducted in supercritical fluid at 350°C and 10 bars. The aim of corrosion testing was to simulate high temperature geothermal environment i.e. IDDP-1 conditions where previous on-site corrosion testing of UNS N06625 and other alloys had been conducted. The simulated environment had lower concentration of H2 concentration of H2S and CO2 S and CO2 in the steam comparing IDDP-1 environment. In addition, no silica scaling was precipitated on the samples nor HF was used in the simulated experiment. The corrosion rate was determined with weight loss comparison and the corrosion forms were analyzed with SEM, XEDS and light microscope. The result of the simulated experiment shows that some localized corrosion is occurring and the corrosion rate of UNS N06625 in simulated environment is similar to the corrosion rate observed in IDDP-1.

The four-point bending method and weight loss method were used to study respectively the stress corrosion cracking behavior and weightlessness corrosion situation of the baosteel BG80S steel in the simulated field environments. The experimental results shown that BG80S won't occur to stress corrosion cracking when the maximum loading stress is 85% Rt0.5 ; the corrosion rates increase with the rising of temperature which is from 40°C to 80°C under the dynamic and static conditions of the simulated environments; the dynamic corrosion rates are between 1.5558 and 1.7523mm/a and the corrosion rates are 0.4827～1.4078mm/a under the static conditions, both of which belong to a serious corrosion category; the form of corrosion is uniform corrosion under the dynamic conditions; because the corrosion products exist micro defects under the static conditions of 80°C, the experimental samples have had the localized corrosion.

Nickel based alloys such as Alloy 600 and 690 have been used as the steam generator (SG) tubing materials in a pressurized water reactor (PWR) due to their high corrosion resistance. However, many types of corrosion have occurred in highly caustic environments containing some oxidizing impurities, especially in SG sludge piles, because the highly caustic conditions can be developed in the heated crevices of PWR SG’s (Jacko, 1990). Among those impurities, Pb is well known to assist in stress corrosion cracking (SCC) of the SG tubing in the caustic environments (Sakai et al., 1990). Many authors reported on the cracking modes of SCC in Pb-contaminated solutions and the role of Pb on the passive films formed on nickel-based alloys to explain the mechanism of SCC (Hwang et al., 1997; Hwang et al., 1999; Kim et al., 2007; Kim et al., 2008). Recently, there was an approach to investigate the mechanism by distinguishing between the initiation and propagation stages of Pb-assisted SCC using an electrochemical noise (EN) technique (Kim & Kim, 2009). The EN is defined as a fluctuation of the electrochemical potential or current which is observed experimentally to be associated with localized corrosion processes (Cottis et al., 2001; Stewart et al., 1992). From the analyses of the EN parameters such as the frequency of events, the average charge of events, the noise resistance (Al-Mazeedi & Cottis, 2004; Cottis, 2001; Sanchez-Amyay et al., 2005), and the mean free time-to-failure (Kim & Kim, 2009; Na & Pyun, 2007, Na et al., 20007a, Na et al., 2007b), various types of localized corrosion are distinguishable from each other. Therefore, the EN monitoring technique has become a useful tool for characterizing such localized corrosions as pitting corrosion, crevice corrosion and SCC. This work is aimed to analyze the EN generated during Pb-assisted SCC of Alloy 600 at high temperature. The EN was measured from C-ring specimens in the highly caustic solution containing oxidizing impurities in two different ways: in a potentiostatic controlled current noise (PCCN) mode, the electrochemical current noise (ECN) was measured from the stressed C-ring specimen by applying an anodic potential. In an uncorrelated three electrode current and potential noise (UCPN) mode, the electrochemical potential noise (EPN) and the ECN were measured simultaneously from the stressed C-ring specimen. Changes in an amplitude and time interval of the ECN and EPN, and variations in power spectral density of the ECN and EPN were analyzed in terms of the initiation and propagation of Pb-assisted SCC of nickel-based alloys in the highly caustic solutions at high temperature.

The nature and rates of the chemical and electrochemical reactions that occur within the occluded regions of a given alloy are controlled by the local electrochemical potential and the local solution composition. The very small physical dimensions of these regions lead to challenges in both measurement and modeling. When performed in a coordinated and complementary way, measurements and modeling provide insights into the controlling processes of a range of localized corrosion phenomena, including crevice corrosion, pitting, intergranular corrosion, and stress corrosion cracking. Examples of attempts to overcome the measurement challenges are described for a range of corrosion scenarios, including identification of the critical ionic species in stainless steel crevice corrosion and in the corrosion of aircraft lap joints, operando measurement of chemistry and potential simultaneously within stress corrosion cracks, and monitoring of water layer thickness in salt spray testing. Examples of work addressing the challenges in modeling localized corrosion including intergranular corrosion of AA5XXX alloys, scaling laws in crevice corrosion, the extent to which the Laplace Equation can be used and applied to geometrically complex galvanic structures, and an approach to modeling localized corrosion for extraordinarily long service times. Finally, suggestions regarding future avenues of research are provided.

Corrosion immersion tests of 3%Cr steel and API X65 pipeline steel were carried out in the high-pressure high-temperature autoclave, in order to simulate the typical corrosion environment of oil and gas transported pipeline. The effects of temperature and CO2 partial pressure on corrosion behaviors were investigated, and the purpose of this paper was to compare the corrosion resistance of 3%Cr steel with API X65 pipeline steels in CO2 corrosion environment. The results show that, the corrosion of both 3%Cr steel and API X65 pipeline steel were slight in mild corrosion environment, so we could choose API X65 steel in these environment. With temperature or CO2 partial pressure increasing, the corrosion resistance of 3%Cr steel was superior to API X65 pipeline steel. The analysis of composition and morphology revealed that the corrosion product film of 3%Cr steel was composed of FeCO3, Cr(OH)3 and Cr2O3. An amorphous scale without grain boundaries and pores was formed on the surface of 3%Cr steel due to the Cr element enrichment, which offered considerable resistance to corrosive medium, therefore the corrosion rate was decreased and local corrosion was inhibited.

One point of technological importance in the use of stainless steels for natural water environments such as sea and fresh waters lies with their liability to stress-corrosion cracking. The stress-corrosion crack initiation in these environments is almost always via localized corrosion such as pitting or crevice corrosion. Although the critical initiation potentials of pitting and crevice corrosion have been clearly defined and the experimental method of determination been well standardized, neither definition nor standard method has been established for stress-corrosion cracking. In this paper, the critical initiation conditions have been discussed for the kind of stress-corrosion crack that originates from corrosion crevice with intergranular stress-corrosion cracking occurring in the sensitized Type 304 stainless steel/neutral chloride solution environment system as an example. Following conclusions have been drawn: (1) The Stress-corrosion repassivation potential, ER,SCC, can be determined in the cyclic polarization tests using a specimen provided with an artificial crevice and applying a static load on it, (2) The ER,SCC thus determined was about 100 mV lower than ER,CREV, the corrosion-crevice repassivation potential determined for the same specimen but with no load applied, (3) Effects of applied stress, degree of sensitization, test temperature, and [NaCl] concentration on ER,SCC were documented, and (4) The ER,SCC agreed with the critical initiation potential for stress corrosion cracking, VC,SCC, determined in the potentiostatic holding test.

Stress corrosion cracking resistance and crevice corrosion resistance are important characteristics for alloys used in sour environments. When crevice corrosion occurs, crevice corrosion would enhance the initiation of stress corrosion cracking. In this study, effect of the presence of crevice in 4 point bent beam specimen are discussed. Type 316L, duplex stainless steel;CR22 (22%Cr-5%Ni-3%Mo), five high Ni alloys {Alloy A (20%Cr-32%Ni- 4.5%Mo), NIC42 (22%Cr-42%Ni-3%Mo), Alloy B (21%Cr-25%Ni-3%Mo), Alloy C (19%Cr-26%Ni-4.5%Mo) and Alloy D (20%Cr-24%Ni-6%Mo)} and ferritic stainless steel;Alloy E (27%Cr-4%Mo) were used. Tests were conducted in autoclaves simulating sour environments (60~120°C, 0.3~20% NaCl, PH2S =0~1 atm.). In case of ferritic stainless steel and alloys containing more than 32%Ni, stress corrosion cracking in creviced specimens was not found under conditions tested, although crevice corrosion was found. In case of Type 316L, duplex stainless steel and alloys containing 25%Ni, stress corrosion cracking took place in creviced specimens at the portion where crevice corrosion was occured under some conditions. These results indicate that ferritic stainless steel and alloys containing more than 32%Ni have high stress corrosion cracking resistance in conditions tested.