Review—Factors Inﬂuencing Sulfur Induced Corrosion on the Secondary Side in Pressurized Water Reactors (PWRs)

The work presented in this review paper is part of a project to investigate corrosion degradation of steam generator (SG) tubing and tube support materials by sulfur compounds at reduced or intermediate oxidation levels (designated here as S x ). The stability of these S x species in aqueous solutions is discussed, and the electrochemical and chemical interactions between S x and metallic surfaces and the subsequent impact of these interactions on metal corrosion are reviewed. Factors inﬂuencing S x -induced uniform corrosion, passivity breakdown and stress corrosion cracking (SCC) considered here include: pH, electrode potential, temperature, S x species, alloy compositions and other impurity ions like Cl − . pro- motes the crack propagation rate of metals because ﬁlm breakdown is easy in the presence of stress, as has been reported on sensitized Alloy 600. The samples showed an increased maximum current and longer repair time when compared to unstressed sample. 157 A recent work 158 found that both tensile stress and compressive stress can lead to an increased surface reactivity of the passive ﬁlm formed on Alloy 800 in S 2 O 32 − solution at room temperature, as observed by scanning electrochemical mi- croscopy (SECM). However, the underlying mechanism, especially the reason why the compressive stress leads to an increased surface reactivity of the passive ﬁlm, needs to be further explored.

Electricity has played an important role in the history of human civilization and greatly influences the quality of life in modern society. The ever increasing demand for electricity and environmental concerns warrant the need to seek safer, cleaner and more efficient methods to generate electricity, especially in the developing countries such as China and India where local economies are booming. However, the air pollution from coal-fired plants in China, for example, is deteriorating the environment and negatively affecting human health, thus validating the necessity to generate power in a more environmentally friendly manner. Nuclear power generation can provide a reliable supply of electricity, with low carbon emissions and relatively small amounts of waste that can be safely stored and eventually disposed of. Pressurized water reactors (PWRs) are used in the large majority of nuclear plants worldwide. Fig. 1 shows a pictorial explanation of a PWR which mainly includes two loops−primary coolant loop (also called primary side) and secondary coolant loop (also termed as secondary side). The primary side refers to the heat generation system to make steam in the secondary side (shown as the orange loop in Fig. 1), and the secondary side denotes the Turbo-generator system (shown as the blue loop in Fig. 1).
The steam generator (SG) shown in Fig. 1 is one of the core components in PWRs, it facilitates heat transfer from the primary to secondary side. The main metallic components in a SG includes SG tubing, tube support plate (TSP) and tubesheet (TST). Primary coolant flows through the inside of the SG tubes and secondary water and steam are heated outside the SG tubes. Several early PWRs used austenitic stainless steel (SS) as SG tubing materials, but later changed to Alloy 600MA ("MA" means mill annealed a ) due to the alkaline and Cl − -induced stress corrosion cracking (SCC) of SS, then changed to use Alloy 600TT ("TT" means thermally treated b ) and generally to Alloy 690TT and Alloy 800NG c (UNS N08800). [2][3][4] Carbon steel was initially used for the tube supports but was found to be prone to corrosion, 2 so was later changed to SS of various grades (Types 304, 316, 405, 409, 410). 2 All the SG tubing, TSP and TST materials are of interest in this review paper, and are listed in Table I. It is worth noting that from batch to batch, their chemical compositions may vary slightly.
As SGs operate with high temperature and high pressure water, corrosion is an issue that threatens safe operation and causes reactor outages. On the primary side, primary water SCC, although historically important, does not fall within the main scope of this review paper because corrosion induced by sulfur is not a problem in the primary side. On the secondary side, corrosion is most likely to occur in z E-mail: dahaixia@tju.edu.cn; jingli.luo@ualberta.ca a Ideally time and temperature for mill annealed is 1100 • C for 0.5 h. 2 b Ideally time and temperature for thermally treated is 715 • C for 10 h. 2 c "NG" means nuclear grade. the heat-transfer crevice (HTC) d that is formed, for instance, between SG tubing and TSP at the secondary side. 2,5,6 Inside the HTCs, some impurities from the bulk environment can concentrate and deposit, producing an often aggressive local environment and leading to sharp gradients in temperature, electrochemical potential, concentration of impurities, and fluid density. 2 The impurities in the HTC include S, Cu, Pb, Cl, iron oxide, silicate and organic species (as schematically shown in Fig. 2). This review paper mainly focuses on the sulfur species at reduced or intermediate oxidation levels (designated here as S x ), which has been well known to be deleterious to passivity. S x can cause corrosion degradation of metals. Much laboratory work has been conducted in order to clarify the electrochemical behavior and manage the corrosion degradation of SG materials in PWRs. S x -induced SCC (S x -SCC) is considered one of the submodes e of SCC in PWRs, as summarized by Staehle and Gorman. 4 Most S x -SCC occurring on SS and nickel-based Alloy 600 on the secondary side has been reported as intergranular stress corrosion cracking (IGSCC) due to a sensitization treatment. 4 Hence, TT treatment, which is applied to Alloys 600 and 690 to make them more resistant to primary water SCC (or low potential SCC, LPSCC), should be conducted at a suitable temperature range for avoiding sensitization of the alloy. Several articles reviewing S x -induced corrosion and S x -SCC have been published in the past years. [8][9][10][11][12] IGSCC of austenitic sensitized SS in S x environment at temperatures below 100 • C was explored, 8 the role of S 2 O 3 2− on steel corrosion at low temperature was reviewed, 9 and intergranular attack (IGA) and IGSCC of Alloys 600 and 690 in high-temperature acidic solutions containing SO 4 2− and Cl − were reviewed. 10 Recent advances using electrochemical methods coupled with surface analysis have resulted in an improved understanding of interactions of SG passive film with electrolytes at microscale. However, broad corrosion problems regarding the influence of solution pH and temperature on the S x species distribution and the subsequent impact on pitting corrosion, passivity degradation and SCC have not been systematically summarized so far.
This review paper brings together the information from the studies relevant to S x -induced corrosion in PWRs, conducted over approximately the last 40 years. The emphasis in the first part of this review is on S x species distribution, which is greatly affected by solution pH and temperature, and the electrochemical potential. The purpose d Heat-transfer crevice refers to the crevice formed between SG tube and the other materials in a SG, the heat from the primary side can transfer through SG to the crevice. 2 e Here, SCC submodes on the secondary side of steam generator tubing in PWRs include: low potential stress corrosion cracking (LPSCC), high-potential stress corrosion cracking (HPSCC), acidic stress corrosion cracking (AcSCC), alkaline stress corrosion cracking (AkSCC), lead stress corrosion cracking (PbSCC), low-valence stress corrosion cracking (S x -SCC), organic stress corrosion cracking (OgSCC), doped steam stress corrosion cracking (DSSCC), and low-temperature stress corrosion cracking (LTSCC). 2  of second part is to identify the dependence of S x -induced corrosion degradation on the primary variabilities of pH, electrode potential, temperature, S x species, alloy compositions, alloy structure and other impurity ions. The third part will discuss the S x -SCC of SG tubes and TSP materials.

The Source of S x Species in the Heat-Transfer Crevice
Sulfate (SO 4 2− ) at ppb or ppt levels is present on the secondary side of SG tubing. 2,13 Also, SO 4 2− can leak from the resin beads in the condensate polisher into the SG tubing. 2,13 The purposeful addition of hydrazine (N 2 H 4 ) (e.g. ∼100 ppb) 14 to maintain the corrosion potential of SG within a safe region can cause a reduction of SO 4 2− to other S x , which are detrimental to the passivity of SG alloys. S x species in deposits in heated crevices is of primary importance, since some S x species can promote SCC and pitting corrosion of SS and Alloy 600, 4  compounds that were found in the TTS sludge material. 15 This kind of attack in Bruce Unit 4 most likely occurred during startup situations where oxidizing conditions developed during the shutdown. The IGA identified in Alloy 600 SG tubes and the TSP at the Palisades Nuclear Power Plant g and Arkansas Nuclear One-1 in 1983 was also associated with S x at low temperatures. 16 Sulfur and chloride species were found in the pits of Alloy 600 tubes removed from Millstone 2 h , indicating that the sulfur species are likely involved in the pitting process. 16 Alloys 800 and 690 have not been reported to have the occurrence of S x -SCC and pitting corrosion under service conditions, but they have been extensively reported in laboratory tests. f a nuclear power station in Ontario, Canada. g a nuclear power plant in Michigan, USA. h a nuclear power plant in Connecticut, USA. In this section, we will discuss the source of S x species in the HTC. The HTC can be found in three common types: 2 (1) TSP/SG tubing crevice, (2) TST/SG tubing, crevice and (3) sludge//SG tubing crevice. S 2− species are present in the HTC due to the reduction of SO 4 2− by H 2 N 4 within the crevice during full power operation, and S 2− can be oxidized to S 2 O 3 2─ and polythionate under the oxidizing conditions that occur during shutdown. The reduction of SO 4 2− to lower valence S x (e.g. S 2− ), which has been confirmed experimentally by Sakai et al., 17 Sala et al., 18 and Allmon et al., 19 is thermodynamically feasible since the N 2 /N 2 H 4 half-cell equilibrium is very negative relative to the S 6+ /S 2half-cell equilibrium. 2 For instance, 0.9∼21.3% SO 4 2− was reduced to S x in the presence of N 2 H 4 and ferrous ions at 345 • C and 155 • C within one month. 15 Carbon steel 508 and 410 SS also significantly catalyze the reduction of SO 4 2− under SG operating conditions. The Alloy 800 SG tubing material is much more inert to reduction of SO 2− 4 than both materials mentioned above. 16 Unfortunately, the reaction kinetics of H 2 N 4 under secondary side conditions and the reaction rates between H 2 N 4 and SO 4 2− or oxygen have not been measured and should be investigated in future work. 7 These reactions within the HTC may be primarily heterogeneous because the crevice is actually a system involving at least steam and water, as schematically shown in Fig. 2.
The temperature of the outside circumference surface of the SG tube within HTC is close to the primary temperature T P , but lower than the temperature of the secondary water (T S ). Therefore, there is a superheat temperature of (T P -T S ). 2,7 The superheat in HTCs is the primary factor affecting S x concentration and the valence state of S x species. 2 The condition within the crevice is very complicated, because it is filled with many deposits other than S x , including Cu, Pb, Cl, iron oxide, silicate, organic species (as schematically shown in Fig. 2). Due to the sharp gradients in temperature, electrode potential, fluid density and concentration of impurities, the environment in the crevice is likely to be very aggressive. However, not all the reactions within the crevice are associated with degradation in-service, because the complex environment results in precipitation of stable and relatively nonthreatening species (many ionic species precipitate out due to loss of water polarity). This is one reason why degradation is not often observed in-service. Also, corrosion product, such as magnetite, in the HTC may play an important role in corrosion, but is seldom studied. The concentrations of the various deposits are significantly different, depending on their vapor pressures because of higher concentrating species having lower vapor pressures. 2 The pH and electrochemical potential with in the crevice can significantly affect the S x species distribution. However, the concentrations of S x species within the crevice are not well determined.

The Stability of Sulfur Species
Sulfur compounds.-Sulfur compounds (S x ) found in the HTCs of secondary side of PWR are expected to be present as various of species with oxidation states from −2 to +6. 13 Some species can adsorb (chemisorb) on the SG tubing surface, and may further be electroreduced or electro-oxidized. In fact, the environment in the HTCs is very complicated and it contains solid particles, steam, and water; of more concern is the complexity of the chemistry in the crevice itself and the wide variation of pH (∼4 to 9), cations (Na, Cu, Ca, Mg, Pb, etc.) and anions (e.g. S, Cl, etc.) possibly present. Fig. 3 schematically shows the diversity and formation conditions of chemical reactions of S x species and their interconversion under specific conditions. 20 Most of them are unstable or metastable except SO 4 2− , S and S 2− . The oxidation states of S can be fractional because S atoms can be present different oxidation states in a compound. Table II shows the structures and space-filling models of some common sulfur species. The complicated condition in the HTC may change the course of  surface reactions occurring on SG tubing, with potential consequences in pitting corrosion, uniform corrosion, passivity degradation, and SCC. 21 Therefore, the analysis of the stability of sulphur species is useful when identifying their roles on uniform corrosion and localized corrosion.
Sulfate (SO 4 2− ).-Sulfate (SO 4 2− ) at ppb or ppt levels can be detected in the secondary side feedwater. SO 4 2− is a stable anion because the S atom is surrounded symmetrically by four oxygen atoms, thus it can be reduced only in a strong reducing environment. 4 The reducing agent hydrazine (N 2 H 4 ) that is present on the secondary side of SG tubing can reduce SO 4 2− to S 2− . 4 This reaction, does not happen readily at room temperature and likely, is highly dependent on temperature although this is still unclear. 4 Experimental evidence has confirmed that Fe, Ni, Cr and their alloys can act as catalysts for SO 4 2− reduction reactions at high temperature. 11,15 In addition to PWR condition, SO 4 2− can be reduced to S 2− by sulfate-reducing bacteria that is extensively present in soil or petroleum. 20 Dithionate or metabisulfate (S 2 At high anodic electrode potential or under oxidising conditions, 11] or disproportionated into SO 4 2− and S 0 or SO 3 2− and S 2− : 27 Whether the corresponding acid of S 2 O 3 2− (H 2 S 2 O 3 ) is stable or not remains unknown to date. The structure of H 2 S 2 O 3 is speculative, as shown below in schematics 1 and 2 below, 28 because the molecule of H 2 S 2 O 3 has never been observed directly. [16] Elemental sulfur (S 0 ).-Elemental sulfur (S 0 ) is likely present in the HTC due to oxidation of S 2− . 13 S 0 is detrimental to passivity of SG tube and tube support materials. 13 Another source of S 0 is the electrochemical reduction of S 2 O 3 2− in corrosion pits, as reported by Newman's group. 29 They found that pitting corrosion products in SS and nickel-based alloys in S 2 O 3 2− -containing solutions contain S 0 and H 2 S. At room temperature, S 0 is most often found as cyclic S 8 . 30 S • can be further reduced to yield H 2 S in acidic conditions, 20 as shown in Equations 10 and 17: Hydrogen sulfide (H 2 S).-Hydrogen sulfide (H 2 S) is a special species as it is likely present in the HTC as a gaseous phase, and dissolved in water, in equilibrium with the partial pressure of H 2 S in the gas. H 2 S can be generated in a similar way as S 0 during pitting corrosion, as mentioned in Elemental sulfur (S 0 ) section. 29 It dissociates to HS − and S 2− in aqueous solution: 9 H 2 S → HS − + H + [18] Thermodynamics of corrosion in environments containing s x .-Corrosion thermodynamics provides the framework for bounding submodes of SCC and passivity degradation on the secondary side of SG tube materials. The corrosion thermodynamics can be shown in forms of Pourbaix diagrams, volt-equivalent diagrams (VEDs) and E−temperature (E−T) diagrams. Calculations of S−H 2 O and S−metal−H 2 O thermodynamics are not easy as S exists in a variety of oxidation states (including fractional states) from −2 to 6, and more importantly, S can adsorb on Fe, Ni and Cr with a broad pH range and play a role in corrosion acceleration 31 (the mechanism will be discussed in details in S 0 section). (

1) S−H 2 O
To investigate the electrochemical behaviors of alloys in S x −containing solutions, the valence of S x as functions of the electrode potential and the solution pH has been studied with an E−pH diagram of the S−H 2 O system. 27,[32][33][34] This type of E−pH diagram and distribution curves of aqueous species are of theoretical and practical importance for understanding the secondary side environment. Such diagrams involving H 2 S, HS − , S 2− , S 0 , HSO 4 − , S 2 O 3 2− , and SO 4 2− (some of them are metastable in solution) have been reported at 25 • C, 33 200 • C, 34 and 300 • C. 34,35 Some metastable ions were also included in the studies, because E−pH diagrams can predict the metastability of those species, including S 2 O 3 2− , SO 3 2− , and S 2 O 6 2− . 13,33 However, the E−pH diagram of S−H 2 O system gives no information on the corrosion of alloy elements in the SG tubing materials. There- gives a framework for bounding submodes of SCC and passivity degradation on the secondary side of SG tube materials. Theoretical calculations made by Marcus and Protopopoff 31,36-38 considered S adsorbed on Fe, Ni, Cr, or Cu in water at 25 and 300 • C, as shown in Fig. 4. They described the stability regions of metal, metal oxides, metal hydroxides and the formation of metal sulfide. It should be noted that Cr is very resistant in near-neutral pH S x -containing environments due to the formation of Cr 2 O 3 , but Fe and Ni are not so resistant because the formation of sulfide in certain potential range. As seen from Figs. 4e and 4f, the presence of sulfur has limited effect on Cr, at 25 • C and 300 • C, and no metal sulfides are formed. In addition, S • can adsorb on Fe, Ni, Cr surface in a wide potential range to catalyze anodic dissolution. It is a chemisorption process on metal surfaces due to an electronegative attraction, with partial charge transfer, resulting in a partial dipole (+) charge on the metal surface. 31,39,40 This leads to in decreased activation energy necessary needed for metal dissolution, thus accelerating anodic corrosion reactions. 31,39,40 E−Temperature (E−T) diagrams were recently developed by Nickchi and Alfantazi 41 who used them to evaluate the thermodynamic stability of metal (Fe, Ni, Cr)-H 2 O systems at temperatures up to 200 • C, and studied the corrosion behavior of high purity Fe, Ni, and Cr in NaSO 4 solutions. The advantages of these diagrams are their capabilities of presenting the phase stability as a function of temperature. However, only SO 4 2− was considered in the work, 41 and no other important S x species were taken into consideration.
Volt equivalent diagrams.-The E−pH diagram in Fig. 4 includes several S x species only, because it is too complex to convey useful information if all the S x species are included. Volt equivalent diagrams (VEDs) i developed by Macdonald and Sharifi-Asl, 42 were used to represent thermodynamic stability of S x in a convenient way. In VEDs, equilibrium information for any given S x species is represented as a i VET of a species is the equilibrium potential for the reduction reaction of the species with respect to elemental sulfur S 0 multiplied by the average sulfur oxidation state in the compound. 41 ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 207.241.231. 81 Downloaded on 2019-04-27 to IP  function of average sulfur oxidation state. Figs. 5a and 5b shows VEDs for the S-H 2 O system at pH = 0 and 10.5, respectively, at 25 • C (the activity of all S x species is 1), with positive slopes in acidic solutions and negative slopes in alkaline solutions. One can judge the stability of S x species by these rules: 42 if a species A, for instance, lies above a line joining any two or more other compounds (B and C), A will tend to be decomposed into B and C; If A lies below B and C, B and C will tend to react to form A.
VEDs are useful tools to evaluate the thermodynamic stability of almost all S x species in S−H 2 O system, but they do not give information in S− Metal−H 2 O systems.
Summary of S thermodynamics.-Thermodynamic calculations such as E−pH and E−T diagrams provide a holistic picture of the stability of metal (including their dissolved anions or cations) and S x species as functions of temperature, solution pH, and electrode potential. Species distributions and VEDs depict the thermodynamic stability of S x species. Combined use of these diagrams can present visual information about corrosion state and stability of species. The information that is not provided in E−pH and VED diagrams includes the following: (1) Thermodynamic calculations offer little information about the corrosion process and the underlying mechanism of corrosion. Besides, the reaction kinetics are also not provided, but the kinetics of SO 4 2− reduction is particularly important in enabling corrosion occurrence on SG, TSP and TTS materials.
(2) Most developed diagrams are representative of a pure metal, but are not valid for alloys that contain more than one alloy element. A recent important work has focused on E−pH diagrams for Ni−Cr−Fe alloys and Pb adsorption on Ni−Cr−Fe alloy surfaces in aqueous solutions in a temperature range of 25∼300 • C, 43 and such diagrams in S x -containing environment are needed. (3) During the corrosion process, some intermediate products are formed that are not considered in the thermodynamic calculations. (4) Thermodynamic calculations are performed based on the equilibrium state of metal and its corresponding cations in solution only, but the effects of other soluble ions on the equilibrium are neglected. When conducting metal corrosion experiments in high temperature and high pressure water to simulate the secondary side condition, the results are greatly affected by Fe 2+ , Ni 2+ , Cr 3+ , and S x concentrations, temperature, complexing agents and pH. These variables usually change over time as corrosion reactions progress, therefore, the thermodynamic stability range of ions and compounds also change with time, which should be taken into consideration in the experiment. In addition, localized corrosion such as pitting corrosion and SCC is often the primary concern and local chemistry conditions in pits and cracks are also constantly in flux. Recently, Santucci Jr. et al. 44 developed chemical stability diagrams (CSDs) to present the relative stabilities of ions, chemical compounds, and complexes of an element as a function of bulk solution chemistry (pH and metal ion concentration). With the help of CSDs, data interpretation and experimental design in simulated secondary side conditions can be improved. (5) In real situations, the corrosion may significantly be inhibited by solid deposition on the tubing surface. As shown in Fig. 2, these depositions may include iron oxide, silicate, and organic species, which are not considered in the thermodynamic calculations. (6) In E−pH diagrams, the average pH value in bulk solution, not the pH on the anodic and cathodic reaction sites, is used. However, in real corrosion systems, the pH in anodic and cathodic regions is likely different from the one in the bulk solution.

Factors That Influence S x −Induced Passivity Degradation
pH.-As SG tubing alloy and TSP materials are in passive state (a metal protected with an oxide film), the outermost surface of the passive film interacts with H + or OH − in the solution when it is exposed to an aqueous solution, resulting in dissolution of the passive layer. The formation and dissolution kinetics of passive film and other film properties are mainly governed by the solution pH and potential, which influences the level of protection the passive layer provides to the metal in direct and indirect ways. 45 The TSP material used in early times was carbon steel, which is also in passive state in the bulk secondary side water. 4 In this part, we mainly discuss the effect of solution pH on the degradation of passive metal/alloys. is present only in near neutral or alkaline solutions whereas H 2 S exists only in the solutions where the pH is lower than 7 (S 2 O 3 2− is a metastable species, as discussed in Thiosulfate (S 2 O 3 2− ) section). Solution pH, on the one hand, affects the S x species that is present in solution, and on the other hand, it also determines the H + and OH − concentrations. These factors ultimately determine the electrochemical behavior of the metal in a complex way.
pH and surface charge.-Solution pH determines the surface charge of a passive film, therefore, there exists a critical pH at which the surface of the adsorbent is neutral, i.e., the point of zero charge (PZC) (also called pH PZC ). [46][47][48] Because the ions with the same charge are repulsive, 45 solution pH significantly influences the passivity breakdown via changing the surface charge. When the pH is lower than the pH PZC , such as in acidic solutions, the surface is positively charged. The dissolution products of the passive film are Fe(H 2 O) 6 2+ , Cr 3 (OH) 4 5+ , and Cr(OH) 2 + , which accumulate in the interfacial layer, and Cl − and S 2 O 3 2− appear to reach the film surface due to anionselective When the passive film surface is negatively charged, such as in alkaline solution, film dissolution reactions are shown in Equations 23 and 24. The Cl − and S 2 O 3 2− attack on the passivated alloy surface is significantly limited. 49 pH effect on film dissolution.-The pH effect on film dissolution can be estimated from the E-pH diagrams as shown in Fig. 4. The Fe-Cr-Ni alloy may undergo dealloying in caustic conditions due to the formation of soluble CrO 4 2− , 50 as shown in Fig. 4f. For instance, Cr was depleted while Fe was enriched in the passive film formed on alloy 800 in alkaline solution at 300 • C. 51 This observation agrees with other results 52 that a Cr-depleted and Fe-enriched anodic film formed at 300 • C on an Alloy 690 exposed to an alkaline crevice chemistry. In the presence of S 2 O 3 2− , S 2 O 3 2− can weaken the selective dissolution of Fe, Ni and Cr from the film, 51,53 possibly due to the formation of a stable sulfate or sulfide species on the surface of passive film. Unfortunately, the underlying mechanism was not interpreted in depth due to a lack of experimental evidence.
The metal cation transport rates during film formation also have significant effects on film composition. 54,55 Faster diffusing species will pass through to the outer layer, while slower diffusing components, such as Cr, will be oxidized due to their stagnant movement and remain in the inner layer. The diffusion rates of these alloy ions in the oxide have been measured 54,55 and show a well defined order: Therefore, film compositions are mainly influenced by these two aspects: (1) solution pH, and (2) metal cation transport rate through the oxide film. With respect to the S x effect on film dissolution, especially at high temperature, the mechanism is poorly understood. Atomic modeling and calculations regarding the interaction of S x with the passive film are needed to assist clarification of the film dissolution mechanism at the atomic scale.
Electrode potential.-As shown in the E−pH diagrams (Fig. 4), the electrode potential can significantly affect the S x species distribution as well as the film composition. Under the fully deaerated secondary side conditions, the electrochemical potential of the cathode reaction is dominated by the H 2 O/H 2 equilibrium, depending on the pH and the pressure of H 2 . 3,13 In the HTC, the electrochemical potential of SG tube is influenced by the deposits including S x species, iron oxide, silicate and organic species. During PWR shutdown, sometimes O 2 is admitted to the secondary side and it can oxidize some S x species to higher valence species. 13 As a result, the electrochemical potential of SG tube is changed due to a change in environmental conditions. S x species distribution.-As indicated both in the thermodynamic predictions (Fig. 4) and in laboratory tests, 24,56 the electrode potential has a significant effect on the S x species: at high anodic potential, S x present as the oxidized species, such as SO 4 2− and HSO 4 2− ; at low potential, S x is reduced mainly to HS − and H 2 S. However, the kinetics of those reactions are poorly understood.
Film composition.-Film composition is significantly affected by the electrode potential, which can be estimated from the E−pH diagram in Fig. 4. The effect of electrode potential on film dissolution rate can be directly observed from the polarization curves. In most cases, the dissolution rate of a passive metal at the free corrosion potential is as low as nA·cm −2 . Driving the potential in the anodic direction, the film dissolution rate increases until the film breaks down. For SG tubing containing Cr (alloy 690 is an example), the polarization curves show two regions in simulated crevice chemistries containing S 2 O 3 2− in alkaline solutions: a passivity region at low potential and a transpassivity region at high potential; the latter is ascribed to the dissolution of Cr 2 O 3 in the passive film to CrO 4 2−50,57 (also termed as dealloying).

Alloy composition.-Iron/nickel/chromium.-SG tubing materials
should be self-passive in normal service conditions of the secondary side. The contents of Fe, Ni, and Cr in alloys significantly affect their localized corrosion resistance in Cl − + S 2 O 3 2− solutions. 58,59 Cr is considered to be beneficial for passivity in S x containing environments due to the formation of protective Cr 2 O 3 in near neutral conditions. 60 The work of Marcus et al. 61 claimed that excessive Fe and Ni contents in alloys will make the alloys more susceptible to accelerated anodic dissolution in the presence of reduced sulfur species, because the passivity can be impaired by the more thermodynamically favorable but less protective FeS, NiS, Ni 3 S 2 , 60,62 leading to a high passive current density. In Cl − -only solutions, Alloys 600, 690 and 800 are all sensitive to pitting corrosion, but Alloy 600 is relatively less susceptible to pitting than 690. 63 In the presence of S 2 O 3 2− , alloys with high Fe and Ni contents are more sensitive to crevice corrosion and pitting corrosion in chloride solutions. 58,59 One of the most important issues of concern is the critical step for pitting corrosion. A recent perspective paper proposed that the critical step is pit growth stability under aggressive conditions, 64 and Alloy 800 in 0.6 M Cl − +0.075 M S 2 O 3 2− solutions is an example. In this case, passive film is easy to breakdown, leading to a fast pitting growth rate. If the solutions are not harsh, e.g., Alloy 800 in 0.001 M Cl − +0.075 M S 2 O 3 2− solutions, pits are not easy to initiate, the passive film breakdown becomes the critical step for pitting corrosion.
Other alloying elements.-(1) Molybdenum Mo, as an alloying element, is present in 316 SS and Alloy 690 (see Table I). TSP materials like 316 SS with higher molybdenum (Mo) content are more resistant to pitting corrosion than the Mo-free 304 SS. 65,66 The role of Mo in solid solution in the pitting corrosion of metals was summarized in three points: [67][68][69][70] (1) Mo-containing alloys are more resistant to pitting corrosion.
This effect is likely provided by Mo 6+ locally enriched on the film surface. (2) Mo slows the kinetics of anodic dissolution after film breakdown and pitting initiation, as reported by Marcus and Olefjord. 69 (3) Mo located on the surface binds adsorbed S and removes it from the surface (by dissolution), thus improving the corrosion resistance when S x species are present. This binding effect was originally reported for nickel alloys. 70 (2) Nitrogen Nitrogen (N) is present in solid solution as well as in precipitates in Alloy 800 and 690 (see Table I). N plays dual roles in the corrosion resistance of alloys. Alloying N is beneficial to the tensile property and corrosion resistance of the metal, and it also promotes the formation of TiN inclusions on the metal (which is usually harmful). 71 The effect of solid solution N on increased corrosion resistance in acidic solutions is because the solid solution N can interact with H + to form NH 4 + at the interface of the metal/passive film. 71 [N] + 4H + + 3e − → NH 4 Equation 26 leads to higher pH. This mechanism may also be applied in localized corrosion, as the pH in the corrosion pits is acidic. However, in alkaline solution, the mechanism is not clear. Alloying N has a synergistic effect with Mo on increasing the pitting potential. 72 Solid solution N obviously plays a role in corrosion resistance for Alloys 800 690 600, but it is hard to correlate N with the differences in corrosion resistance among them, because corrosion resistance is jointly impacted by many other factors. As of this writing, the role of N on S x -induced corrosion has not been studied systematically.
TiN inclusions are present in alloys 690 and 800, mainly in square and rectangular shapes. 73 The size generally varies from submicron to several microns. Broken TiN and TiN clusters may facilitate the corrosion fatigue of Alloy 690. 74 However, their role in localized corrosion in simulated PWR conditions has not been investigated.
Temperature.-Increasing temperature leads to a change in S x species distribution. Increasing solution temperature degrades the protectiveness of the passive film and increases the corrosion rates in S x environment since corrosion is thermally activated. 75 Each 10 • C increase in temperature corresponds to about 50 kcal/mol decrease in activation energy for corrosion reaction, thus doubling the corrosion rate. 4 Increasing operating temperatures have been regarded as the reason of increased uniform corrosion rates for Alloy 600MA tubes. 4 At high temperature, the dissolution rate of the passive layer is higher than the dissolution rate at low temperature because the vacancy generation and transfer rate are enhanced. 75 The corrosion of SG tubing alloys has been postulated to occur at low temperature due to the presence of both O 2 and S x species during PWR shutdown or startup evolutions, as well as at high temperature during full power operation. 29,76 Early work mainly investigated the pitting corrosion of Alloys 600, 690, and 800 in solutions containing Cl − and S 2 O 3 2− in low temperature ranges (below 100 • C), 77-79 and recent work was conducted at elevated temperatures up to 315 • C. 51,60 Increasing temperature, the pitting potential of 304 SS and 316L SS decreased in Cl − -only or Cl − + S 2 O 3 2− solutions. 66,80 Valence of S x species.-The electrochemical reactions of Soxygen compounds have been reviewed by Hemmingsen. 20,81 Some S-oxygen compounds are presented in the order of oxidation state in Fig. 3. The impact of the valence of S x species on corrosion degradation is briefly reviewed in the following sections. S 0 .-S 0 has been reported to significantly increase the corrosion of carbon steel under both aerated and deaerated conditions. 82,83 Early TSP used carbon steel as the material. This type of materials is subjected to catalysis of cathodic and anodic reactions in acidic environments based on the following proposed reactions: 82,83 [28] which gives the overall corrosion reaction as: The additional cathodic reaction in acidic environment is: In neutral conditions, the additional cathodic reactions are expected to be: [32] In the presence of Cu 2+ in the HTC, the following cathodic reaction is also involved: 13 Cu 2+ + 2e − → Cu [33] Adsorbed S • catalyzes anodic dissolution during which there is no change in the oxidation state during the dissolution process. 84 This catalytic effect on the anodic dissolution has been reported on Fe, Ni, Fe-Cr, Fe-Ni, carbon steel, Fe-Ni-Cr alloys. 84 The mechanism has been merged together by Marcus et al.: [85][86][87] the adsorbed S • can disrupt oxide formation and block H 2 O adsorption and dissociation, therefore inhibit the oxide film formation. Adsorbed S 0 on metal surface results in a dipole (δ−) charge on the adsorbed S atom and a dipole (δ+) charge on metal surface, which reduces the activation energy necessary for active dissolution of metal.

S 2-.-S 2and S 2 O 3
2− have similar impacts on uniform corrosion and pitting corrosion. 88,89 However, S 2is only stable in alkaline solution. In acidic solution, S 2can interact with H + to form HS − and H 2 S, which are aggressive for SG and TTS material; in alkaline solutions, S 2is also stable and the corrosion products on SS are FeS, FeS 2 , NiS, Ni 2 S 3 , which are less protective. 90 FeS deposit was reported to be electrochemically reactive and led to a large interfacial capacitance. 90 6 2− , and ultimately to SO 4 2− , or be reduced to S 0 and S 2-(or to polysulfides), depending on the redox potential and the solution H + concentration. 24,56 During the electrochemical oxidation of S 2 O 3 2− in both linear potential and galvanic voltammograms, an oscillatory behavior of current/potential is often observed. 92 S 2 O 3 2− is known to enhance uniform corrosion as well as pitting corrosion for many alloys. For example, in the presence of S 2 O 3 2− , the corrosion rate of plain carbon steel was greater than 1 cm per year in simulated pipeline water composition. 93 In cases of pitting corrosion of 316 SS, S 2 O 3 2− can damage the passive film seriously in the presence of Cl -, which has been extensively reported. 94 A mechanism has been proposed: S 2 O 3 2− can stabilize metastable pits during passive film breakdown induced by Cl -. 95,96 The combined effect of Cland S 2 O 2− 3 on localized corrosion occurs when the Cl -/S 2 O 3 2− concentration ratio is high, ensuring that Cladsorption is dominant, and as a result, the initial film breakdown is induced by Cl -. 97 If the Cl -/S 2 O 3 2− concentration ratio is low, insufficient amounts of Clare available for breakdown of the passive film, therefore pitting corrosion is mitigated. Obviously, pitting does not occur in a solution containing only S 2 O 3 2− . These electrochemical behaviors mentioned above have been found on 304L and 316L. 98 The interactions of Cl − and S x with passive film at high and low Cl − /S 2 O 3 2− concentration ratios is discussed later in details in Interactions of Cl − and s x with passive film at high and low Cl − /S 2 O 3 2− concentration ratios section. The reason that S 2 O 3 2− alone or S 2 O 3 2− + small amount of Cl − is not very detrimental to passivity is still in debate and the following three reasons have been proposed: (1) an excess of S 2 O 3 2− can neutralize the acidic conditions within a pits, therefore retard pit growth. 88 (2) under the condition where Cl -/S 2 O 3 2− concentration ratio is low, an excessive amount of S 2 O 3 2− will retard the adsorption of Cl-, therefore passive film does not break down easily and no pitting corrosion is observed, as reported; 99

C58
Journal of The Electrochemical Society, 166 (2) C49-C64 (2019) (3) possibly due to its large anion size, S 2 O 3 2− is unable to occupy the vacancies in the passive layer. 100 S 2 O 3 2− itself can hardly enter the lattice of oxygen vacancy unless it is reduced to S 2-. 99 In absence of Cl -, S 2 O 3 2− may interact with the passive film and slightly alter the composition of the passive film, which has been confirmed by radioactive labeling and Auger electron and x-ray photoelectron spectroscopy. 101,102 It was found that S 2 O 3 2− accumulation on the 304SS surface is irreversible and occurs over a broad electrode potential range of -1.0 to -0.50 V Ag/AgCl . 101,102 This irreversible surface behavior is attributed to S 2 O 3 2− incorporation into the passive film. However, the authors concluded that S 2 O 3 2− accumulation is reversible on 316 SS because the Mo content in 316 SS is higher than that in 304 SS. 101,102 Possibly, the Mo cation creates a barrier that makes S 2 O 3 2− penetration into the passive film of 316 SS energetically more difficult than into the passive film of 304 SS. A recent study conducted by Faichuk et al. 103 also showed that S 2 O 3 2− accumulation on Alloy 600 (Mo-free alloy) was irreversible because S is detected in the passive film.
It has been reported that S 2 O 3 2− may "facilitate" hydrogen adsorption and permeation into metallic materials in acidic solutions, 104,105 but these effects are attributed mainly to the decomposition products of S 2 O 3 2− (H 2 SO 3 , HSO 3 -, and S 0 ), as S 2 O 3 2− is readily decomposed in acidic environments to give the above products, and because experimental results obtained in the same acidic environment with SO 3 2showed much less hydrogen permeation. Although S 2 O 3 2− can facilitate hydrogen permeation into the passive film, the presence of O 2 in the S 2 O 3 2− solution inhibits hydrogen permeation. 106 Hydrogen adsorption enhancement is lower in neutral solution where S 2 O 3 2− remains mainly undecomposed. Thus, the major effects of S 2 O 3 2− are caused by itself in neutral media and by its decomposition products in acidic media. It should be noted that the pitting potential of metal cannot be accurately measured if the oxidation potential of S x is lower than the pitting potential. Future work should focus on fully modeling the interaction of S 2 O 3 2− with a passive surface at the atomic scale in order to figure out whether S can enter into the passive film and determine the associated impact on corrosion.
2is present in the HTC and is measured during hideout return studies j . It is considered as a metastable ion and tends to decompose to yield S 2 O 3 2− or S 0 . S 4 O 6 2has been found to significantly enhance the uniform corrosion rate of Alloy 800 in acidic solution, as it increases cathodic and anodic reaction rates. 107 S 4 O 6 2has been widely reported to promote SCC initiation, localized corrosion, and uniform corrosion in lab tests. 108 S 4 O 6 2is also considered as one of the detrimental ions that cause passivity degradation and SCC in the secondary side of PWR, as failure cases have been given in part 2. 13 The impact of S 4 O 6 2on corrosion and its decomposition products should not be neglected when conducting experiments and interpreting the results in S 4 O 6 2containing solutions. 4 2-does not harm passivity of SG tube materials at low temperature 109,110 whereas it may be harmful at high temperature if the pH value is low, due to the electrochemical reduction of SO 4 . 2-111 SO 4 2may induce localized corrosion in the presence of S 2 O 3 2− or Cl -, as reported by Newman 112 who found that an Fe-19Cr-10Ni alloy underwent pitting corrosion in a potential range of −325 ∼ −100 mV SCE using scratch tests in SO 4 2-+ S 2 O 3 2− solution. Newman 112 claimed that the reason for rapid pitting corrosion was the reduction of j Hideout and hideout return are impurities deposited during boiling and dissolved back to solution during shutdown. During boiling, the solutions accumulate in flow-restricted regions on the secondary side of the SG, such as crevices formed at the intersections between the SG tubes and the tube sheet, the tube-support structure and deposits have accumulated on the tube surface and on the tube sheet. These processes lead to the formation of concentrated solutions (hideout) in these flow-restricted regions. During shutdown and cooldown, steam voids collapse, areas are rewetted, and impurities are released back into the bulk steam generator water -a phenomenon known as hideout return. Hideout return studies correspond to water composition analysis in the HTC conducted at the shutdown or cooldown time. S 2 O 3 2− to adsorbed S and S 2at the scratch site at room temperature (22 ±1 • C). In this situation, SO 4 2just acts as the support electrolyte in the presence of S 2 O 3 2− at room temperature. Polarization curves of Ni immersed in SO 4 2solution at 315 • C in acidic solutions containing Clshow that the anodic current density increased as the SO 4 2concentration increased from 0 M to 0.015 M. 41 These results indicate that SO 4 2is the aggressive anion in mixed Cland SO 4 2systems at elevated temperatures, due to the electrochemically reduction of SO 4 2-; the anodic current density is increased and the metal dissolution is activated in Ni-Fe-Cr alloys in the presence of SO 4 2-, even at relatively low concentrations of SO 4 2-. It is unlikely that SO 4 2itself is responsible for accelerating anodic dissolution in Ni alloys and is more likely that the reduced sulfur species produced at high temperatures plays a major role. Despite that some corrosion phenomena have been reported, the underlying passivity mechanism has not been well clarified. on the passive film degradation, 117 because H 2 S showed a diffusion effect, not a migration effect during localized corrosion. At the acidic pH values in and near a pit in the passive film, H 2 S is uncharged and will always be depleted inside the pit. S 2 O 3 2− is an anion, and although it is decomposed in acidic solutions, the decomposition is slow enough so that there is always pit enrichment of S 2 O 3 2− by electromigration.
H 2 S in aqueous solution significantly enhanced the uniform corrosion rate of many metallic materials including SG TSP (carbon steel and SS), and the corrosion product was FeS (sometimes polymorphous) and Fe 1−x S (hexagonal crystal). [118][119][120] FeS mainly forms at low temperature and low H 2 S partial pressure while Fe 1−x S is formed at high temperature and high H 2 S partial pressure. Solution pH and temperature have significant effects on the protection of FeS; 120 increases in pH and temperature lead to increased corrosion resistance of FeS film due to its compact structure. 120 In a review of corrosion products of steels in H 2 S environments, 30 it was pointed out that thermodynamics and detailed kinetics of the corrosion product formation and transformation are not well understood.
Experiments involving H 2 S are complicated and dangerous because (1) H 2 S is toxic and corrosive, (2) H 2 S can contribute to the currents measured using electrochemical techniques (similar to S 2 O 3 2− ), because of the oxidation and reduction of H 2 S ,121 and (3) reactive and porous ferrous sulfide films increase the interfacial capacitance and introduce diffusional effects. 122 Tsujikawa, et al. 123 and Kappes et al. 26 developed a solution containing S 2 O 2− 3 ions, which can be used to mimic the sour gas environments that are generally found in the petroleum industry.
Summary of S x effects on corrosion.-All S x species should be considered when investigating corrosion mechanisms of metals in S xcontaining environments in the lab. For instance, the initial added S 2 O 3 2− will be decomposed to many other species via electrochemical reductions or oxidations, complicating the determination of corrosion processes. The decomposition of S x species is dependent on the functions of electrode potential, solution pH, temperature, and oxygen concentration, and the compositions of the materials under investigation. 59 S x species can transform to higher or lower valences, complicating the determination of corrosion processes.
Interactions of Cl − and S x with passive film.-Cl − is considered to be a detrimental ion for many alloys, including Fe-Ni-Cr alloys used in PWRs. Cl − −induced pitting corrosion of Fe-Ni-Cr alloys has been extensively explored in the past decades, and the effect of S x on Cl − −induced corrosion has been well documented. 51,[124][125][126][127][128][129][130] This section will begin with the effects of Cl − on film degradation at high and low Cl − /S 2 O 3 2− concentration ratios.  97 claimed that if sufficient Cl − ions were available to breakdown the passive layer, the adsorption of S 2 O 3 2− within the pits and the electrochemical reduction of S 2 O 3 2− would stabilize the metastable pits and catalyze the pitting growth. A critical concentration ratio at which pit growth rate reached a maximum value was defined as the Cl − :S 2 O 3 2− ratio = 250:1 for type 304 SS at 20 • C in 1 M NaCl. 66 This critical ratio is highly dependent on materials and temperature. 140 If the Cl − /S 2 O 3 2− concentration ratio is low, the ions have no combined effect, and pits are not initiated. One reasonable reason is that there is not enough Cl − to break down the passive film, 99,141 as discussed in S 2 O 3 2− section.

Interactions of Cl − and S x with passive film at high and low
Impact of sulfate.-SO 4 2− alone cannot trigger localized corrosion of Alloys 600, 690 and 800 at temperatures below 100 • C; however, it can induce localized corrosion of Alloys 600, 690 and 800 in the presence of S 2 O 3 2− and Cl − . The presence of 8000 ppm SO 4 2− suppressed anodic dissolution reactions of alloy 600 in Cl − solution, perhaps through the formation of SO 4 2− salts. 142 The sulfate anion is much more likely to precipitate out with stable ionic species in reaction with cations at low temperature. Pitting corrosion occurred in solutions where SO 4 2− /S 2 O 3 2− concentration ratios ranged from 1.6 to 58. 134 Newman 79 claimed that S 2 O 3 2− -induced pitting of commercial AISI 304 SS occurred readily when {[SO 4 2− ]+[Cl − ]/[S 2 O 3 2− ]} was in the range of 10 to 30. Cl − was not necessary for pitting; the right proportion of S 2 O 3 2− and SO 4 2− ions could take part in an acidification process that leads to pitting. Ag 2 SO 4 may be used as a localized corrosion inhibitor for alloy 690 in NaCl solution at room temperature because Ag 2 SO 4 could interact with Cl − to form AgCl of low solubility (see Equation 34), altering the interface of the double layer structure and mitigating anodic dissolution. 143 Ag 2 SO 4 + 2NaCl → 2AgCl + Na 2 SO 4 [34] However, Ag 2 SO 4 has not been used in real PWRs at the temperature of 300 • C because SO 4 2− is considered detrimental to passivity at high temperature, due to the formation of reduced species. The addition of a relatively small amount of SO 4 2− to a 0.1 M Cl − solution results in a significant increase in current density in the anodic range of Ni and Ni alloys at 315 • C, 60 suggesting that SO 4 2− is detrimental to Ni-based alloys. 60 Cullen et al. 144 also reported a similar phenomenon: there was a combined effect of SO 4 2− and Cl − at high temperatures on nickel-based alloys, and the role of Cl − is to breakdown the passive film, and reduced sulfur species (from SO 4 2− reduction) adsorbed on these sites to promote pit growth. 2 contamination should be avoided during pressurized water reactor (PWR) startup and wet layup conditions, because O 2 significantly increases the corrosion potential, and changes the chemical conditions on the secondary side of the SG tubing. The possible ingress of O 2 during startup and shutdown should be prevented or greatly minimized. The role of O 2 on corrosion can be summarized as the following three aspects:

Oxygen (O 2 ) concentration.-O
(1) O 2 can oxidize the lower valence of S x , i.e. S 2− , to high valence species S 2 O 3 2− which is more aggressive to passivity. 13 (2) O 2 contributes to the cathodic reaction, therefore increases the cathodic current density, causing the increase in the anodic dissolution rate of carbon steel. 106 (3) O 2 can oxidize Fe 2+ and Cu to form Fe 3+ and Cu 2+ . These species are detrimental to passivity of SG tube and tube support materials. 3

S x Influences on Passive Film Properties
Film thickness.-Passive film thickness in SG tubing is mainly influenced by the exposure temperature, and increases as the temperature rises. 145 At room temperature, the passive film thickness is in the range of several nm, but the thickness can increase to hundreds of nm at 300 • C. 126 As the temperature in a PWR rises, anion vacancy generation rate increases at the metal/film interface, film dissolution increases at the film/solution interface, and the transport rates of oxygen vacancies and cation vacancies are enhanced. Anion and cation vacancies generation rate at the metal/film interface is faster than the film dissolution rate and as a result, film thickness increases. 126 Film thickness cannot be correlated with the corrosion resistance of the passive film because corrosion resistance is mainly determined by the composition, defect level, and structure of the passive film.
Film structure.-Film structure greatly influences the SCC susceptibility of a passive metal. Passive film with a crystalline structure is superior to the passive film with an amorphous structure. 146,147 In most cases, Ni and Fe hydroxides form the outer layer of the passive film whereas the inner layer contains oxides that protect the film from corrosion. A film enriched with crystalline Cr 2 O 3 is beneficial for passivity. 148 The heat treatment process of nickel-based alloys is important, because it highly influence the chromium depletion, precipitate evolution, and resistance to IGA. 149 It is unknown whether the presence of S x can change the structure of the passive layer on an alloy, and an atomic-scale mechanism is needed for clarification in future work. A recent paper reported the effect of Cl − on the structure of the passive film formed on a FeCr 15 Ni 15 single crystal, and confirmed that Cl − was accumulated at the metal/film interface, leading to a lattice expansion on the metal substrate, undulations at the film/substrate interface, and structural inhomogeneity on the film side. 150 What should be investigated in the future is whether S has the similar effect as Cl − on film structure.
Semiconductivity.-Most metallic materials used on the secondary side of steam generator tubing in PWRs are in a passive state, i.e., a passive layer is formed on them. In most cases, the passive layers are a p-type, n-type, or n-p-type semiconductor. Metal oxides of Fe and Cr are n-type whereas Ni is p-type. The n-type oxide can be reduced easily, which is detrimental to passivity, 151 Alloy 800 in Cl − solutions is an example. 99,141 Some results reported so far seem to be contradictory, as the nickel-based Alloy 600 shows either n-type or p-type semiconducting properties. 152 This is possibly due to the limitations of the techniques in determining the semiconductivity. As the semiconductivity is usually experimentally determined by Mott−Schottky technique, during the potential scan, the results are always influenced by the Faradic current generated by the dissolution of a passive film, casting a dubious light on the experimental evidence. 153 Semiconductivity influenced by S x is not well understood yet. Recent investigations on Alloy 800 and 690 show that the adsorbed S 0 on passive film surface can accept electrons to form S 2− by occupying the oxygen vacancies V •• O , as shown in Equation 35. 99,130,141 However, a direct evidence is still needed to confirm this.
Pit growth rate.-Pit growth in an alloy leads to rapid material failure. S 2 O 3 2− accelerates pit growth rate in Cl − solution when the Cl − :S 2 O 3 2− concentration ratio is high. As discussed in Interactions of Cl − and s x with passive film at high and low Cl − /S 2 O 3 2− concentration ratios section, there is a critical Cl − :S 2 O 3 2− ratio at which the pit growth rate reaches a maximum. Increasing or lowering this ratio led to a decreased pit growth rate or no pit initiation. For instance, in the case of of 316 SS and 304 SS in solutions containing S 2 O 3 2− (0.01 to 500 mM) and Cl − (1 to 1000 mM), a critical concentration ratio was around 10∼50. 65 Concentration ratios lower than 1 and higher than 1000 prevented pit formation. 65 A similar phenomenon was reported by Wang et al. who used image analyses to study pit growth on alloy 800. 154 Pits were identified from the dark regions from corrosion images. Pit growth rate in 2D can be characterized by counting the occupied dark areas on the whole alloy surface. However, this method offers little information on pit depth because of the limitation of the 2D corrosion images.

S x -Induced Stress Corrosion Cracking (S x -SCC)
S x −SCC is one important SCC submode owing to the concentrated S x species in the HTC, which has been measured during hideout return studies. 4 As discussed in Valence of s x species section, many S x species are detrimental to passivity, therefore it is not surprising that these species can promote SCC. For instance, SCC of SG tube materials at some CANDU stations have been attributed to a low-temperature attack by S 2 O 3 2− and S 4 O 6 2− . 13 SCC of sensitized Alloy 600 SG tubes was detected in Three-Mile Island PWR under wet−layup conditions in the mid-1990s. 4 Sensitized TSP materials such as 304 SS and 316 SS have shown great S x -SCC susceptibility. 4 Summary of SCC initiation work on metallic materials in S x -containing solutions is shown in Table III. In this part, we discuss the factors affecting S x -SCC and the possible remedies.
Stress.-The stress may stem from the material itself or extraneous interference such as corrosion product. The corrosion product generated from carbon steel corrosion results in a slow straining and denting of the SG tube, which accelerate SCC of the SG tube. 4 Stress induces film rupture, alters diffusion kinetics of atoms in the passive film, and oxidation kinetics on grain boundary. 155 Both of these can greatly enhance SCC susceptibility. It is known that stress promotes the crack propagation rate of metals because film breakdown is easy in the presence of stress, as has been reported on sensitized Alloy 600. 156 The repassivation of stressed Alloy 800 C-ring samples showed an increased maximum current and longer repair time when compared to unstressed sample. 157 A recent work 158 found that both tensile stress and compressive stress can lead to an increased surface reactivity of the passive film formed on Alloy 800 in S 2 O 3 2− solution at room temperature, as observed by scanning electrochemical microscopy (SECM). However, the underlying mechanism, especially the reason why the compressive stress leads to an increased surface reactivity of the passive film, needs to be further explored.
Electrode potential.-The effect of electrode potential on S x -SCC can be estimated in the E-pH diagrams as shown in Fig. 4. The potential range in which the metal is active is also the region where metal is prone to S x -SCC, while the relative minimum SCC susceptibility usually corresponds to good passivity of these alloys. 4 For instance, heat sensitized nickel-based Alloy 600 underwent active dissolution between -300 mV SCE and +400 mV SCE in low temperature S 2 O 3 2− solution, it also suffered from IGSCC susceptibility, 139 because of the Ni-rich and chromium-depleted phase at the grain boundaries (see Table III).
Solution pH.-As can be estimated from E-pH diagrams shown in Fig. 4, iron based Alloy 800 and nickel-based Alloy 600 are active in acidic solutions, therefore S x -SCC of Alloy 600 and 800 easily occurs in an acidic condition. 159,160,41 Alloy 690 with more Cr content is susceptible to SCC in alkaline solutions since Cr is in the form of CrO 4 2− , 50 as shown in Fig. 4f. There are different corrosion states shown in Figs. 4b, 4d, 4f: pitting and uniform corrosion occur below pH 3 (at 300 • C) while pitting and SCC occur between pH 3 and 5 (at 300 • C).
Temperature.-S x -SCC susceptibility increases with temperature, which is not surprising since the passive region shrinks as the tempera-ture elevates 4 (see and compare the E-pH diagrams at 25 • C and 300 • C shown in Fig. 4). This is because S x species distribution is changed at high temperature, therefore, the interaction of S x with passive surface is totally different from the one at low temperature. In the case of sensitized Alloy 600 in 0.1 M S 2 O 3 2− solution, the activation energy is as low as 3.5 kcal/mol. 4 At high temperature, even ingress of SO 4 2− can accelerate the metal dissolution rate, particularly in Ni and Ni-rich alloys that may not form sufficiently compact oxide films in aggressive acid solutions. Severe SCC, IGA and pitting corrosion susceptibility of alloy 800 in acid SO 4 2− solutions at 280∼320 • C have been reported by Persaud et al. 60,161,162 In laboratory experiments, traces of H 2 S were detected and S was detected at the crack tip although oxide state was not reported. 60,161,162 Species.-S x species with valences of −2, 0 and +6 are thermodynamically stable at low temperature, although other intermediate oxidation states of S x can form via either reduction of SO 4 2− by H 2 N 4 or oxidation of S 2− by O 2 introduced to the SG water during wet layup or startup. 13 S x -SCC mainly occurs in solutions containing S 2− , S 2 O 3 2− , or S 4 O 6 2− regardless of the low or high temperature, and S x -SCC has been reported on 304 SS, Alloys 600 and 800, see the summary in Table III. SO 4 2− does not accelerate S x -SCC at low temperature but can promote SCC at high temperature due to the electrochemical reduction of SO 4 2− . 143 The kinetics of all the reactions regarding S x under startup and layup conditions are still not well understood. 78 Additionally, whether magnetite and Cu 2+ have any synergistic effect with S x on SCC should also be investigated.
Alloy composition.-Alloys 600MA and 800 are susceptible to S x -SCC in near neutral conditions while Alloy 690 shows good S x -SCC resistance. 163,164 Almost all SG tube and SS materials are prone to S x -SCC at 350 • C in alkaline S 2 O 3 2− -containing solution. 4 The results indicate that S x -SCC occurs extensively on SG tubing materials as well as SS. In acidic solutions, where sulfur-species are most relevant, higher Cr content in Alloy 690 generally imparts more resistance to S x -SCC due to the better ability to form and sustain a passive oxide after rupture or impairment (or in competition with sulfur), as compared to Alloy 600. 164 Alloy structure.-Some treatment methods for SG and TSP materials will significantly affect their structure. Sensitization treatment of Alloy 600 and 304 SS leads to an increased S x -SCC susceptibility, 156,165 but mill annealed Alloy 600 sustains no S x -SCC susceptibility at low temperatures. 77,166,167 This is because sensitization treatment generates a Ni-rich and chromium-depleted phase at the grain boundaries of Alloy 600 and 304 SS. 156,165 Cold work of 316L SS led to an acceleration of SCC and a crack growth rate in oxygenated high temperature water, 168 but whether cold work has the same effect on SG tubing and TSP materials in S x environment is presently not known.

Prevention or amelioration of S x -SCC.-Strategies
that can prevent or ameliorate S x -SCC are briefly discussed in this section, these methods including surface treatment, water chemistry optimization and SG structure design.
(1) Surface treatment is frequently used to enhance the SCC resistance of metallic materials. Laser shock peening 169 and iterative thermomechanical processing k170 have been reported to improve the resistance of Alloy 600 to SCC in S 4 O 6 2− solution. (2) The S x concentrations in the secondary side can be lowered by exploring suitable strategies for water chemistry control, to minimize the formation of such species caused by the hydrazine reduction of SO 4 2− . (3) From an engineering perspective, it is better to avoid the formation of HTC. If the crevices between the SG tube and TSP cannot be eliminated, it should be designed so as to keep them as tight as possible in practice. Additionally, line contact arrangements, rather than surface to surface contact, should be used as the contact mode between the tube and tube support.

Future Work
(1) Exploring the passive film properties in S x -containing environments at the nano and atomic scales The passivity degradation mechanism at the nanoscale is not clear, so a multiscale approach combining atomic modeling with advanced electrochemical and surface analyses techniques is needed to fully understand alloy corrosion mechanisms. 171 The details of the film structure in an atomic scale, film thickness and semiconductivity influenced by S x should be investigated. The roles of alloying elements with low concentrations such as Mo, S, and Cu in pitting and SCC initiation are not fully understood at the atomic scale, and this can be further investigated by in-situ techniques (such as Raman spectroscopy) that could be valuable to fill the knowledge gaps such as the small quantities of sulfur segregation and oxidation state information of sulfur. The electron transfer reactions occurring on a passive film influenced by S x also need to be clarified. The repassivity kinetics of passive film at high temperature have been studied by in situ scratch electrode systems, 172 but attention should also be paid to the reliability and data reproducibility of this technique.
(2) Environment at heat transfer crevice occluded sites The specific condition in the HTC is very complicated because it is a two-phase system or steam blanketed. Unfortunately, the specific conditions are not well determined. The kinetics of the reduction of high valence S x to low valence S x is not known. There are various impurities in a HTC, the role of magnetite (Fe 3 O 4 ) and Cu 2+ on S x −induced corrosion is not clear. Future work should be focused on systematically investigating the corrosion degradation of SG, TTS, and TSP materials in the simulated crevice chemistries.
(3) S x -induced stress corrosion cracking (S x -SCC) mechanism at the atomic scale S x is involved in SCC initiation and propagation. S x is detected in the crack tip 162 on Alloy 800 at high temperature, but its valence is unknown. Our knowledge on how S x influences the composition and fracture toughness of passive films at the atomic scale only remains at a superficial level, and should be expanded through future work. In addition, the roles of other impurities (Cu 2+ ) on S x -SCC should also be studied systematically.
(4) Simulated corrosion degradation in bulk boiling conditions, the nucleation boiling surface, and the flow boiling condition Most corrosion experiments under simulated PWR conditions are conducted in a static autoclave, which is different from the environment on the secondary side of steam generator tubing. Future work should focus on the electrochemical behavior of SG tubing alloys in bulk boiling conditions or even at nucleation boiling surfaces and flow boiling conditions where the electrochemical reactions are significantly influenced by mass transfer. Bulk boiling condition refers to the situation where the electrolyte is heated to boiling (about 100 • C). Nucleation boiling surface simulates the SG service condition where the inner temperature of SG is high, e.g., 150 • C, and the outside of the SG is exposed to electrolyte, therefore the water at the SG tubing surface is boiled. The flow boiling condition simulates the real SG service condition which is similar to a nucleation boiling surface, i.e., the temperature inside SG tubing is as high as 300 • C. Flow boiling is a complex condition involving elevated pressure and temperature which generate vapor and results in the flowing vapor-liquid two phases. Recently, Chen et al. 173 established a novel experimental setup to simulate flow boiling condition and to study the solid-liquid mass transfer under flow boiling conditions. In such condition, the electrochemical process is governed by mass transfer, which has not been investigated previously.

Conclusions
This paper reviews the factors influencing S x -induced corrosion on the secondary side in PWRs. The following conclusions can be made: (1) The stability of S x species in aqueous solutions highly depends on pH, temperature and electrochemical potential. S x species with valences of +6, 0, −2 are stable in the relevant pH and temperature ranges. (2) pH, electrode potential, temperature, S x species, alloy compositions and other impurity ions such as Cl − can significantly influence the uniform corrosion and localized corrosion of SG tubing and TSP materials in S x environment. (3) S x -SCC can occur on Alloy 600, 800 and other SSs, but highly depends on solution pH, temperature and alloy compositions. (4) Although great progress has been made regarding S x -induced corrosion on the secondary side of PWRs, future work should focus on the corrosion mechanism at the atomic scale utilizing advanced surface analyses techniques and atomic modeling.