Stress corrosion cracking (SCC) of austenitic stainless steels in high temperature light water reactor (LWR) environments

Stress corrosion cracking (SCC) of austenitic stainless steels in high temperature light water reactor (LWR) environments

9 Stress corrosion cracking (SCC) of austenitic stainless steels in high temperature light water reactor (LWR) environments P. L. A n d r e s e n, G...

9MB Sizes 0 Downloads 43 Views

9

Stress corrosion cracking (SCC) of austenitic stainless steels in high temperature light water reactor (LWR) environments

P. L. A n d r e s e n, GE Global Research Center, USA

Abstract: This chapter focuses on stress corrosion cracking (SCC) of austenitic stainless steels in high temperature light water reactor (LWR) environments, and emphasizes common grades of wrought austenitic stainless steel over cast, ferritic or martensitic stainless steels. Corrosion fatigue and environmental effects on fracture are related forms of degradation, and are also addressed. This chapter provides insight into the SCC dependencies and the underlying processes and mechanisms. There are 20–30 key, interdependent variables, and this complexity makes a purely empirical approach intractable. Research has transformed this complexity to one in which the origin of the effects and their interaction is reasonably well understood and quantified. Key words: stress corrosion cracking, crack growth rate, high temperature water, light water reactors, stainless steel, nickel base alloys, sensitization, stress intensity factor, corrosion potential, irradiation effects, prediction.

9.1

Introduction

Many incidents of stress corrosion cracking (SCC) in austenitic stainless steels have occurred in boiling and pressurized water reactors (BWRs and PWRs), and laboratory studies of this phenomena extend over 50 years [1–28]. The corrosion and SCC behavior of structural materials in high temperature water have many unique elements, and extrapolation of data and intuition based on low temperature (<100 °C) response is of limited use, especially when considering low temperature, aggressive forms of SCC such as in halide cracking, including in concentrated MgCl2 environments at 100–154 °C. By contrast, SCC in stainless steels in relevant high temperature LWR environments represents a lower SCC susceptibility or growth rate that ideally satisfies the need for long life in nuclear power plant (NPP) applications. The pressure boundary design codes (e.g., ASME Section III [29]) address only cyclic loading, and reflect an evolution of their historical emphasis on fatigue and fracture; SCC is simply addressed by noting that the design must ensure it will not occur. Thus, the concepts of SCC immunity and thresholds (conditions below which SCC does not occur) are attractive, although careful 236 © Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

237

studies show that immunity rarely exists [10–12]. High resistance to initiation can be achieved by careful attention to design issues such as low stress; no crevices or sharp corners; proper selection of materials, heat treatment and fabrication/welding/grinding; good control of environment including during transient operation; minimizing grinding or other sources of surface cold work; polishing/treating surfaces to minimize surface roughness and create compressive stresses, etc. However, in large welded structures ‘initiation circumventing’ phenomena (broadly, anything that results in much faster initiation than expected from smooth surface data) have contributed to extensive SCC in operating plants. While continued efforts to improve initiation are important, complete reliance on a thin 1–50 mm skin in welded structural components is unwise, and ‘inherent resistance’ to SCC crack advance remains a key consideration. A crack growth rate approach is supported by the vastly better experimental techniques for quantifying crack growth than crack initiation. SCC is often schematically shown as a confluence of stress, material and environment factors (Fig. 9.1), with an implication that there is a very small central region of susceptibility. It would be wise to view such diagrams as representing an iso-susceptibility boundary, with the central overlap representing high susceptibility or growth rate, and increasingly larger circles overlap representing boundaries of lower susceptibility or growth rate. Figure 9.1 also shows the underlying crack tip processes that more directly control SCC growth, as well as how these processes are affected by radiation. Improved crack detection and monitoring techniques, coupled with improved testing techniques, permit SCC growth rate response to be observed into regimes that were previously hidden ‘below ground level’, and had given the impression of immunity and masked the inter-relationships of related phenomena. The modern view is that all variables follow a well-behaved continuum, with growth rates often dropping by, e.g., 100¥ or 1000¥ – but not to zero – for changes in corrosion potential or other parameters. Examples include corrosion potential (SCC occurs in deaerated water), degree of sensitization or neutron fluence level (SCC occurs in unsensitized stainless steel), water purity (SCC occurs in ultra high purity water), crack tip stress intensity factor, K (SCC has been readily observed in at ~5 MPa÷m), etc. Thus it is much more realistic and appropriate to consider a continuum in SCC response as a function of material, environment and stressing parameters, with the need to define boundaries of adequately low susceptibility or growth rate, rather than viewing very low susceptibility as immunity. The objective of this chapter is to provide a broad framework for understanding and interpreting SCC (especially the crack growth rate response) of austenitic stainless steels; to define linkage and commonality of SCC in structural materials; to define the dependencies (e.g., corrosion potential, water purity, stress intensity factor (K), neutron fluence, sensitization, etc.);

© Woodhead Publishing Limited, 2010

238

Understanding and mitigating ageing in nuclear power plants Crack chemistry, mass transport

Stress

Environment

Oxide rupture rate at crack tip Microstructure

Passivation rate at crack tip (a) Solution renewal rate to crack tip Stress

Df anionic   transport

Oxide rupture rate at crack tip

Environment Microstructure

g-field –

Crack tip f (A) , pH

Hardening Relaxation

Passivation rate at N-fluence crack tip G.B. denudation Segregation (b)

9.1 SCC is often shown schematically as an overlap of stress, environment and microstructure, although the small central area should be viewed as a region of high SCC susceptibility. The underlying factors include mass transport, dynamic strain that damages the protective oxide, and the repassivation process. This structure provides a basis for anticipating the effects of irradiation. The complexity of SCC is reflected in the large number of influential variables and the associated requirement that all 20 to 40 in a given system be adequately controlled.

to propose controlling processes and SCC mechanisms; to describe SCC mitigation approaches; and to summarize the state of ability to predict SCC. This understanding facilitates appropriate mitigation measures to be taken for NPP ageing and plant-life management (AM and PLiM, respectively) to achieve safe, reliable and long-term performance/operation in the system, structure or component (SSC) involved.

9.2

Historical problems and structures affected

At the start of commercial nuclear power over 50 years ago, it was reasonable to select stainless steels for use in high purity, unaggressive water chemistries

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

239

at moderate stresses, and this decision was supported by field experience in other industries and accelerated testing. However, in retrospect, such accelerated tests did not have sufficient sensitivity to ensure high SCC resistance over long operating times. Earlier and more incidents of SCC developed in boiling water reactors (BWRs) because of the oxidizing conditions that exist in the coolant, although the incidence of SCC in pressurized water reactors (PWRs), with very low dissolved oxygen, has also increased in the last 25 years. The common grades of austenitic stainless steel are types 304, 316, 321 and 347 stainless steel (Table 9.1). These are generically called 18-8 stainless steels because they have approximately 18% Cr and 8% Ni, along with 0.04–0.07% C, ~1% Mn and ~0.5% Si. The low carbon specification (e.g., 304L and 316L) is <0.03% C, and 0.015–0.020% C is typical. Type 316 stainless steel includes 2–3% Mo, while Type 321 incorporates ~0.5% Ti and Type 347 ~0.5% Nb. C, N and Ni stabilize the austenite phase, and reducing their concentration or adding Cr and Mo promotes formation of the ferrite phase. Stainless steel weld metals, such as Type 308/308L, are designed to form 3–12% ferrite, which is necessary to avoid hot cracking during weld solidification. Ferritic and martensitic stainless steels typically have < 1% Ni. Common nickel alloys (which are all austenitic) include Alloy 600 (~15.5% Cr, ~8% Fe), Alloy 690 (~30% Cr, ~9% Fe) and Alloy 800 (~30% Ni, ~20% Cr, bal Fe). Nickel alloy weld metals, such as Alloy 182 (~15% Cr) and Alloy 82 (~20% Cr), are used both to weld nickel alloys as well as for most dissimilar metals welds. Alloys 152 and 52 are the higher Cr (~30%) weld metals. In BWRs, water boils on the fuel cladding surfaces, and the steam, after going through the steam separators and dryer, directly drives the turbine. The steam is condensed and then the water is demineralized and returned as feedwater to the pressure vessel. Water is circulated upward through the Table 9.1 Typical composition of common grades of austenitic stainless steel, (wt%) AISI Grade UNS

Fe

Cr

Ni

Mn*

Si*

C*

Other

304 304L 316 316L 321 347

Bal Bal Bal Bal Bal Bal

18–20 18–20 16–18 16–18 17–19 17–19

8–12 8–12 10–14 10–14 9–12 9–13

~1.2 ~1.2 ~1.2 ~1.2 ~1.2 ~1.2

~0.5 ~0.5 ~0.5 ~0.5 ~0.5 ~0.5

~0.045 ~0.020 ~0.045 ~0.020 ~0.045 ~0.045

– – 2–3 Mo 2–3 Mo ~0.5 Ti* ~0.5 Nb*

S30400 S30403 S31600 S31603 S32100 S34700

* Mn is 2% maximum; Si is 1% maximum C is 0.08% maximum in non-L grades, typically 0.045% in modern heats, 0.065% in older heats; L-grades are typically 0.015 – 0.020% Ti in 321 is 5¥(C+N) minimum, 0.7% maximum Nb in 347 is 10 ¥ C minmum, 1.0% maximum

© Woodhead Publishing Limited, 2010

240

Understanding and mitigating ageing in nuclear power plants

core and steam separators, and down through the annulus between the core shroud and the pressure vessel. An average water molecule recirculates 7–10 times before leaving the reactor as steam. The accumulation of impurities in the reactor water is controlled by the reactor water clean-up system. Radiation–primarily by neutrons–produces radiolysis of water that simplistically generates hydrogen (H2) and hydrogen peroxide (H2O2). Most of the H2 partitions to the steam phase, leaving behind a net oxidizing environment in the reactor water. Nuclear core reactivity is controlled by moving control rods, containing neutron absorbers such as B or Hf, that penetrate through the bottom head of the reactor pressure vessel. Normal water chemistry (NWC) is considered to have 100–200 parts per billion by weight (ppb, or parts per 109) dissolved O2 and ~10 ppb dissolved H2, although the conditions in the core are different due to radiolysis. All US and some international BWRs inject H2 (hydrogen water chemistry or HWC) to mitigate SCC by reducing the O2 and H2O2 – the H2 level in the reactor water can vary from ~40 to ~250 ppb. Most US BWRs employ NobleChem™, which creates an electrocatalytic surface on all wetted components. This requires much less H2 injection for SCC mitigation, typically resulting in reactor water with 35–40 ppb H 2. This also avoids an increase in radioactive N16, which shifts from soluble (e.g., NO3–) to volatile (e.g., NO or NH3) at higher levels of H2, and increases the radiation level in the piping and turbine (called turbine shine). N16 forms at very low concentrations from transmutation of O16 in the core, and has a half-life of 7.1 seconds. PWRs circulate pressurized primary coolant water through the core, where it is heated from ~290 °C to ~323 °C. It then flows through the inside of the steam generator tubing, with boiling occurring on the outside in the secondary water, which produces steam that drives the turbine. The primary water (that flows through the core) contains boric acid (H3BO3) and lithium hydroxide (LiOH), with ~2 bar of dissolved H2 (~3 parts per million (ppm) H2 or ~35 cc H2 per kg H2O) to minimize radiolysis. The reactivity during the fuel cycle is controlled by adjusting/tapering the boron (B) level from ~1500 ppm to < 50 ppm at the end of cycle. Li (as LiOH) is adjusted to maintain an approximately constant pH at temperature (pH300C ~7.1), and so is typically adjusted/tapered down from ~3 ppm to ~0.3 ppm over the fuel cycle (the target pH300C has varied by plant and over time). Control rods that penetrate from the upper head are used primarily as a safety system and for shutdown. Over-pressure in the primary system is maintained using a pressurizer – a smaller pressure vessel connected to the primary system in which water is electrically heated to ~343 °C, where the vapor pressure is about 151.7 bar, about 34.5 bar above the vapor pressure of 323 °C water. The secondary system uses high purity, low O2 feedwater with O2 scavengers such as hydrazine or morpholine. Trace level impurities (ppb levels) that

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

241

accumulate during steam generator operation are ‘blown down’ by releasing water from the bottom of the steam generator. In BWRs, stainless steels are extensively used for piping, pressure vessel cladding and structures inside the pressure vessel, including the core shroud (which separates the up-flow through the core from the downward flow in the annulus), the core plate (which supports the bottom of the fuel), the top guide (which aligns the top of the fuel bundles), the shroud dome (the dome is above the core, and supports the steam separators), the steam separators, the steam dryer (above the separators), etc. The earliest incidents of SCC in BWRs occurred in stainless steel fuel cladding [30–33] (before zirconium alloys were used). Radiation produces atom displacement damage that causes grain boundary segregation (including Cr depletion) and radiation hardening, which significantly increase susceptibility to SCC (as discussed later). SCC then developed in creviced, cold-worked and/or furnace sensitized components, and these factors were eliminated from designs. Subsequently, SCC occurred in small, then increasingly larger diameter, stainless steels pipes that were weld sensitized; these were primarily type 304 stainless steel, which typically had carbon levels > 0.06%. Most problems occurred in piping external to the pressure vessel (e.g., recirculation and reactor water clean-up piping), and considerable efforts were devoted to understanding and mitigating SCC in these piping systems. Most piping was replaced by a low carbon grade which incorporated nitrogen to retain strength. This was designated as ‘LN’ (low carbon, added nitrogen) or ‘NG’ (nuclear grade) stainless steel. Subsequently, SCC developed in many core internals, initially in highly stressed components (such as absorber tubes that contain the B4C neutron absorber material, which swells over time), core spray piping welds, and core shroud welds (Fig. 9.2). Related structures (e.g., the top guide) and materials (e.g., Alloy 182 nickel-base weld metals) also exhibited cracking. The control blades and absorber tubes are replaceable, but SCC can permit dissolution of the B4C and potentially affect control rod insertion. Various approaches have been taken to manage SCC in core internals, including hydrogen injection (H2 water chemistry), electrocatalysis (NobleChem™) [34–37], mechanical restraints (tie rods) that limit motion of the shroud in the unlikely event of through-wall circumferential cracking, etc. In PWRs, the incidents of SCC were slower to appear, but have increased steadily over time and include stainless steel fuel cladding, baffle former bolts, pressurizer heater sleeves, canopy seals in the control rod drives, steam generator safe ends, etc. Some debate continues regarding whether SCC occurs primarily under unusual circumstances, such as upset water chemistry (e.g., O2 in make-up water), crevice/boiling situations (like pressurizer heat sleeves), severely ground (surface cold-worked) piping welds, very high O2 conditions such as in unvented control rod drives (where air is pressurized

© Woodhead Publishing Limited, 2010

242

Understanding and mitigating ageing in nuclear power plants

Internals holddown bracket

Steam dryer and brackets Core spray piping and spargers

Shroud head bolts Feedwater nozzle weld butters

Core spray safe end to nozzle weld butters

Top guide rim welds and beam

Shroud head and separators

Core shroud weld Jet pump riser brace

SRM dry tubes and in-core housings

Recirc. inlet and outlet nozzle to safe end weld butters

Core plate rim welds Recirc. inlet safe ends

Access hole cover and shroud support shelf

CRD stub tube to housing welds (SS)

Shroud-to-shroud support weld Recirc. piping welds

Jet pump riser elbow Alloy 182/ 600

304, 304L, or 347SS

9.2 Schematic of a BWR showing materials and areas where cracking has occurred.

and trapped during start-up), etc. However, with a large number of incidents and the widespread observation of SCC of cold-worked stainless steel in laboratory tests in PWR primary water, it is likely that SCC of stainless steels is a generic issue whose incidence is likely to continue to rise somewhat with operating time. The presence of ‘cold work’ from shrinkage strains in the weld heat affected zone is a universal issue, and is addressed later. Examples of intergranular SCC in BWR and PWR components are shown in Fig. 9.3. The emphasis in this chapter is on SCC in stainless steels, but SCC has also been observed in many nickel-base alloys and weld metals, including in BWRs (Alloy 600 shroud head bolts, Alloy X-750 jet pump beams, and various Alloy 182 weldments) and PWRs (Alloy 600 steam generator tubing, Alloy 600 upper head and lower head penetrations, Alloy 182 and 82 weld metals in various locations), etc., so some comments will be directed at showing the areas of common dependencies and mechanisms in stainless steels and nickel alloys.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

243

1.3 cm Crack

Iascc pwr Baffle Bolt

2 1

3

9.3 Examples of intergranular SCC in stainless steel: BWR pipe welds (left), BWR control rod sheath (middle) and PWR baffle former bolts (right).

9.3

Stress corrosion cracking (SCC) dependencies – introduction

Increasingly sophisticated measurement techniques and patient observation have eroded the historical concepts of immunity and thresholds to SCC in high temperature water [11–13, 19, 21], and have revealed a continuum in SCC response among the relevant materials, environments and stresses. High resolution crack depth measurements monitor SCC growth rates into regimes that were previously hidden ‘below ground level’, which had previously given the impression of immunity and uniqueness. The modern view is that all variables follow a well-behaved continuum, with growth rates often changing dramatically for changes in corrosion potential (Figs 9.4 and 9.5), water purity (Figs 9.4–9.7), stress intensity factor (Fig. 9.8) and other parameters. If a low growth rate is achieved, another variable may become more aggressive and cause the growth rate to increase to a readily observable level; demonstration of the underlying interconnection requires high resolution measurements. The interdependency is widespread, with the effect of a given variable generally dependent on the state of all other variables (e.g., Figs 9.4–9.8). It is important to understand that corrosion potential represents a mixed electrochemical potential, and does not refer to the ‘potential or rate of corrosion’. As oxidants (e.g., O2 or H2O2) are

© Woodhead Publishing Limited, 2010

244

Understanding and mitigating ageing in nuclear power plants

added to water, the corrosion potential of iron- and nickel-base materials increases, although the corrosion rate often does not. When not specified in more detail, corrosion potential refers to the mixed electrochemical potential on the relevant surface, usually an iron- or nickel-base structural material. The absence of SCC growth rate thresholds (i.e., where SCC ceases) is demonstrated for sensitization, corrosion potential and water purity in Figs 9.4–9.7, which show data (and predictions) for sensitized and unsensitized stainless steels and nickel alloys in moderate to high purity 288 °C water. Figures 9.4 and 9.5 include data from the SKI/EPRI round robin [38] (the smaller symbols at about +150 to +200 mVshe, or mV referenced to the standard hydrogen reference electrode), and also include data (larger symbols in Fig. 9.4b) for carefully controlled changes in potential and at low potential. Similar data exist for the non-threshold response vs. temperature, buffered 304 STAINLESS STEEL 1

25 mm CT Specimen 10

–6

Furnace sensitized; 15 C/cm 288 °C water; 0.1-0.3 mS/cm Constant load; 25 Ksi√in

5

2

10 8 6

Crack propagation rate, mm/s

42.5 min/h

10–7

11

14

14.2 min/h Theoretical curves

a aa 9

mS/cm 0.1

10–8 Hydrogen 12 water chemistry b b 10–9

–600

0.3 0.2

3 7 2

4

Normal water chemistry (ex-core)

–400 –200 0 +200 Corrosion potential, mVshe (a)

+400

9.4 SCC growth rate vs. corrosion potential for stainless steels tested in 288 °C high purity water containing 2000 ppb O2 and 95–3000 ppb H 2.

© Woodhead Publishing Limited, 2010

1.E-05

Crack growth rate, mm/s

1.E-06

Sensitized 304 Stainless Steel 30 MPa√m, 288 °C water 0.06–0.4 mS/cm, 0–25 ppb SO4 filled triangle = constant load open squares = “gentle” cyclic

¨200 ppb O2 ¨500 ppb O2 ¨2000 ppb O2

Stress corrosion cracking of austenitic stainless steels

Screened Round Robin data – highest quality data – corrected corr. potential – growth rates corrected   to 30 MPa√m

42.5 28.3

1.E-07

14.2 min/h GE PLEDGE Predictions 30 MPa√m

1.E-08

0.5

0.25

2000 ppb O2 Ann. 304SS 200 ppb O2 0.1

0.06 mS/cm

0.06 mS/cm Industry mean 30 MPa√m 1.E-09 –0.6 –0.5 –0.4 –0.3 –0.2 –0.1 0.0 0.1 Corrosion potential, Vshe (b)

9.4 Continued

© Woodhead Publishing Limited, 2010

0.2

0.3

0.4

245

Understanding and mitigating ageing in nuclear power plants 1.E-05 Sensitized 304 Stainless Steel 30 MPa√m, 288 °C water 0.06–0.4 mS/cm, 0–25 ppb SO4 SKI Round Robin Data filled triangle = constant load open squares = “gentle” cyclic

¨200 ppb O2 ¨500 ppb O2 ¨2000 ppb O2

246

Crack growth rate, mm/s

1.E-06 CW A600

316L (A14128, square) 304L (Grand Gulf, circle) non-sensitized SS 50%RA 140 C(black) 10%RA 140C (grey) 1.E-07

42.5 28.3 14.2 min/h

CW A600 GE PLEDGE Predictions 30 MPa√m Sens SS

0.5

2000 ppb O2 Ann. 304SS 200 ppb O2

0.25

1.E-08 0.1

0.06 mS/cm

GE Pledge Predictions for Unsensitized Stainless Steel (upper curve for 20% CW) 1.E-09 –0.6 –0.5 –0.4

–0.3 –0.2 –0.1 0.0 0.1 Corrosion potential, Vshe

0.2

0.3

0.4

(c)

9.4 Continued

water chemistry, presence vs. absence of grain boundary carbides, cold work, etc., as discussed later. The behavioral characteristics that reflect similarities and continua in SCC response include: ∑

BWR and PWR primary water chemistry, including pure water and the buffered B/Li chemistry in PWR primary water ∑ type or grade of stainless steel ∑ stainless steels vs. nickel alloys ∑ corrosion potential, controlled primarily by oxidants (dissolved O2 and H2O2), reductants (H2) and pH, and the effects of impurities such as chloride and sulphate ∑ temperature

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

247

∑ intergranular morphology, even in unsensitized materials ∑ grain boundary Cr depletion vs. grain boundary carbides ∑ cold work from bulk deformation, surface cold work, weld residual strain, etc. ∑ neutron irradiation ∑ stress intensity factor (K) ∑ general corrosion rate ∑ grain boundary silicon ∑ effects of environment on fracture ∑ crack initiation.

1.E-05 Sensitized 304 Stainless Steel 30 MPa√m, 288 °C water 0.06–0.4 mS/cm, 0–25 ppb SO4 SKI Round Robin Data filled triangle = constant load open squares = “gentle” cyclic

4 dpa 304 SS

Crack growth rate, mm/s

1.E-06 316L (A14128, square) 304L (Grand Gulf, circle) non-sensitized SS 50%RA 140 C(black) 10%RA 140C (grey) 1.E-07

CW A600 GE PLEDGE Predictions 0.5 mS/cm 30 MPa√m 0.25 Sens SS 0.1

1.E-08

CW A600

2000 ppb O2 Ann. 304SS 200 ppb O2

0.06

GE Pledge Predictions for Unsensitized Stainless Steel (upper curve for 20% CW) 1.E-09 –0.6 –0.5 –0.4 –0.3 –0.2 –0.1 0.0 0.1 Corrosion potential, Vshe (a)

0.2

0.3 0.4

9.5 SCC growth rate vs. corrosion potential in 288 °C high purity water for stainless steels, irradiated stainless steel and various nickel-base alloys.

© Woodhead Publishing Limited, 2010

Understanding and mitigating ageing in nuclear power plants

1.E-06

Sensitized 304 Stainless Steel 30 MPa√m, 288 °C water 0.06–0.4 mS/cm, 0–25 ppb SO4 SKI Round Robin Data filled triangle = constant load open squares = ‘gentle’ cyclic

¥ 750 20% CW HTH AH

316L (A14128, square) 304L (Grand Gulf, circle) non-sensitized SS 50%RA 140 C(black) 10%RA 140C (grey)

1.E-07

¨200 ppb O2 ¨500 ppb O2 ¨2000 ppb O2

1.E-05

Crack growth rate, mm/s

248

718 CW A600 42.5 28.3 14.2 min/h

CW A600

20% CW ¥ 750 AH

GE PLEDGE Predictions 30 MPa√m Sens SS

0.5

2000 ppb O2 Ann. 304SS 200 ppb O2

0.25 1.E-08 0.1

¥750 718 HTH

0.06 mS/cm

GE Pledge Predictions for Unsensitized Stainless Steel (upper curve for 20% CW) 1.E-09 –0.6 –0.5 –0.4

–0.3 –0.2 –0.1 0.0 0.1 Corrosion potential, Vshe (b)

9.5 Continued

© Woodhead Publishing Limited, 2010

0.2

0.3

0.4

249

¨2000 ppb O2

Alloy 182, Alloy 600 & St. Steel 30 MPa√m, 288 °C Water Sens 304 SS Round Robin (open) Circles = constant load Triangles = ‘gentle’ cyclic SO4 & Cl data in pink

¨500 ppb O2

1.E-05

¨200 ppb O2

Stress corrosion cracking of austenitic stainless steels

Includes: • All alloys tested • All heat treatments • No cold work data • Medium screening

1.E-06

Crack growth rate, mm/s

42.5 28.3 1.E-07

14.2 min/h Expected 16¥ peak in CGR for alloy 182 vs. H2 in pure water

1.E-08

CGR data in pure water in Ar or low H2

1.E-09 –0.6 –0.5 –0.4

0.25 0.1 0.06 mS/cm

–0.3 –0.2 –0.1 0.0 0.1 Corrosion potential, Vshe (c)

SS Pledge Prediction * Normal YS

0.2

0.3

0.4

9.5 Continued

9.3.1 Need for SCC insights and modelling based on data and fundamental understanding The number of factors (and sub-factors) and their complexity (interdependency) make a purely empirical approach to understanding and quantifying SCC intractable. Without interdependencies, evaluating five levels for each of 20 variables would require 5 ¥ 20 = 100 tests. But with interdependency, the number leaps to 520 = 1014 experiments. Of course, all experiments would need to be definitive and reproducible, which are notorious problems in SCC experiments. Another approach is to emphasize fundamental understanding and modeling of SCC. The challenge is that there are hundreds of processes that might be

© Woodhead Publishing Limited, 2010

Understanding and mitigating ageing in nuclear power plants

Crack growth rate, mm/s

10–6

10–7

316L Stainless Steel 25 mm CT Specimen Constant load 288 °C water Test conditions: 0 C/cm2 EPR ª 27.5 MPa√m 200 ppb O2

10–8

Predicted curves from Pledge Code for typical range in ECP 10–9

100 10–1 Solution conductivity, mS/cm (a)

101

10–7

Crack growth rate, mm/s

250

10–8 Theoretical prediction for typical range in corrosion potential 10–9 304 Stainless steel Water 288 °C; 200 ppb O2 ~25 Ksi√in; 15 C/cm2 (27.5 MPa√m) 10–10

10–1 100 Solution conductivity, mS/cm (b)

101

9.6 Predicted and observed SCC growth rates of stainless steel in 288 °C BWR water as a function of water purity, e.g., from additions of chloride or sulphate.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels Sens. 304 Stainless Steel Deaerated, 289 °C water fc ª – 0.6 Vshe SSRT 10–6/s

10–2

in/h

Crack growth rate, mm/s

10–4

251

10–5 10–3

Prediction 1 ¥ Œ· app

TG + Ductile IG + Ductile

10–6

10–2

102

10–1

100 H2SO4, ppm (a)

10–4

101

102

Sensitized Type 304 Stainless Steel Factor of improvement for change in conductivity from 0.5 to 0.1 mS/cm

Factor of improvement

K = 5 MPa√m

101

K = 25 MPa√m

SSRT = 1 ¥ 10–6/s

100

–0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 Corrosion potential, Vshe (b)

0.2

0.3

9.7 (a) Elevated growth rate can occur in deaerated water if sufficient H2SO4 is added. (b) The factor of improvement for a change in water purity depends on corrosion potential and stressing.

controlling SCC, and while formulations exist to describe those processes, often the constants cannot be accurately estimated in high temperature water. Historically, electrochemists have tried to describe SCC as a predominantly

© Woodhead Publishing Limited, 2010

Understanding and mitigating ageing in nuclear power plants

10

–5

4

Stress intensity, ksi√in 6 8 10 20 30 40 60 80 Sens. 304 Stainless Steel 288 °C Water

Crack growth rate, mm/s

10–6

10–4

10–7

NRC disposition line

Theory 15 C/cm2­, –50 m Vshe 0.5 mS/cm Theory 15 C/cm2­, –50 m Vshe 0.2 mS/cm Theory 15 C/cm2­, –200 Æ – 500 m Vshe 0.2 mS/cm

10–8

10–9

10–10

10–3

4

10–5

10–6

Crack growth rate, in/h

252

10–7

6 8 10 20 30 40 60 80 Stress intensity, MPa√m

9.8 Predicted and observed SCC growth rate vs. stress intensity factor for various materials and water chemistry conditions in 288 °C water.

electrochemical phenomenon, but with little success. Mechanical engineers have focused on stress and stress intensity factor and cycling/vibration, also with limited success. Nuclear engineers have modeled radiation damage; materials scientists considered phases and nucleation and diffusion, etc. All of these factors are important, but only within a more complex context of the whole of SCC. The most effective approaches have evaluated existing data and dependencies, hypothesized the underlying processes (e.g., the effects of water chemistry or mechanics), performed critical experiments to validate the hypothesis, then modeled and quantified the critical parameters. From an engineering perspective, SCC is often, and simplistically, characterized as a confluence of stress, environment and metallurgy (Fig. 9.1). But this is only an engineering, macroscopic, external (to the crack) perspective. Crack advance must instead be understood from the perspective of the crack tip

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

253

system, and then the role of more readily measurable, external, engineering parameters on the crack tip must be described. Historically, the occurrence of SCC in each material or environment has been considered unique, but as increasingly careful studies are performed, perhaps 80% of the dependencies and characteristics are common among iron and nickel-base structural materials. This section will emphasize the description of individual effects, as itemized above – primarily for stainless steels, but with some discussion of the dependencies and mechanisms shared with nickel alloys. The next section will provide a mechanistic approach for interpreting these parameters, including their interdependencies, from the perspective of the crack tip system.

9.3.2 Distinctions between BWRs and PWRs The primary differences between BWR and PWR primary conditions are associated with: coordinated changes vs. time in the B and Li level in PWR primary water that increases the pH at temperature in pure water from 5.65 (BWR) to ª7.2 (PWR); the H2 fugacity (from ª40 to 3000 ppb H2); and temperature (274/288 °C vs. 288/323/343 °C for the PWR core inlet, core outlet and pressurizer). Of these differentiating factors, temperature is the most important in stainless steels, whereas for nickel alloys, both temperature and H2 are important. Irrespective of whether BWRs operate at high (electrochemical) corrosion potential or low corrosion potential on the exposed surfaces (that is, with or without H2 addition and/or NobleChem™), the crack tip itself is always deaerated and at a low corrosion potential. The details of secondary (steam generator) chemistry in PWRs (a few early BWRs also had steam generators) is beyond the scope of this chapter, and the comments below should be viewed as a broad, introductory perspective. Once-through and recirculating steam generator designs are both in use, and there are typically three or four steam generators per PWR. The boiling process leaves behind non-volatile species, and steam generators do not have a clean-up system (such as the reactor water clean-up system in BWRs) but rather depend on periodic ‘blow-down’ to purge the system of impurities. Improved control of feedwater chemistry has been key to controlling corrosion; this includes control of impurities, dissolved oxygen and other oxidants, and additives to control pH and scavenge oxidants. The control of pH was once done using phosphate, but essentially all commercial steam generators now use all volatile treatment (AVT) comprising species like hydrazine, morpholine, EDTA, etc., most of which also scavenge oxidants. Lead (Pb) and sulfur (S) are major concerns, as is the corrosion product ‘sludge’ that builds up on the tube sheet and can induce stresses and strains on the tubing as well as an occluded (crevice) chemistry. The heat exchanger tubes are the primary SCC problem in steam generators, and most Alloy 600 tubing

© Woodhead Publishing Limited, 2010

254

Understanding and mitigating ageing in nuclear power plants

has been changed from mill annealed (most susceptible) or thermally treated Alloy 600 to Alloy 690. Thermal treatments for both Alloy 600 and Alloy 690 are typically ~705 °C for ~12 hours, which produced grain boundary carbides without Cr depletion. Canadian (CANDU) and German (PWR) plants generally use Alloy 800, which at 30Ni-20Cr has a composition between a stainless steel and Alloy 600, and has given very good service. (CANDU reactors are heavy water moderated, but the secondary side is similar to that in LWRs.)

9.3.3 BWR and PWR primary water chemistry The presence of the B and Li chemistry in PWR primary water sometimes leads to the assumption that the SCC growth rate behavior is fundamentally different from that in BWR pure water. Certainly the presence of oxidants and high (electrochemical) corrosion potential in BWRs produces an elevated growth rate compared to deaerated water (Figs 9.4 and 9.5). However, the crack tip, where the SCC process takes place, is deaerated in all cases. Also, the vast majority of BWRs now operate at low corrosion potential using surface catalysis and a small amount of H2 [34–37], as discussed in the Section 9.8. The net H2 at the surface is typically 35–40 ppb H2, and the corrosion potential of structural materials is close to the potential in deaerated water, which is predominantly controlled by the H2/H2O reaction. Thus, a major distinction in chemistry between BWR and PWR primary water comes down to the role of B and Li in PWRs, whose concentrations are typically ~1500 ppm B as H3BO3 and 3 ppm Li as LiOH at the start of the fuel cycle, tapering to < 50 ppm B (sometimes <10 ppm) and ~0.3 ppm Li at the end of cycle. B and Li levels are coordinated to maintain an approximately constant pH300C of ~7.0–7.3. Laboratory studies involving dozens of duringthe-test changes in B and Li concentrations [39–43] showed no effect on crack growth rate in stainless steels or nickel alloys, even when changing from deaerated pure water (Fig. 9.9). This is consistent with expectations based on the Pourbaix diagram (Fig. 9.10), because shifts in pH produce a change in corrosion potential (controlled by the H2/H2O reaction) that is exactly parallel to the metal-metal oxide reactions for Ni, Fe and Cr (of these, only Ni/NiO is close to the H2/H2O reaction). In the mid-pH range, the solubility of metal cations (e.g., Ni2+) or anions (e.g., HNiO2–) is limited; at more extreme pHs, a significant effect of pH might be expected. In the presence of oxidants, the relatively concentrated B and Li chemistry plays a large role, producing growth rates that are perhaps 10¥ higher than in pure water (Fig. 9.11). Thus, oxidants should be avoided in PWR primary water. Indeed, the common practice of adding of H2O2 during PWR shutdown (typically at <130 °C) to react with the dissolved H2 should be evaluated to determine the extent to which cracking may be accelerated.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

To 1100 ppm B, 2 ppm Li @2675h

–8

pH325C constant at ~7.25

c283 – 0.5TCT of A600 CRDM, 325C 27.5 MPa√m, 30 cc/kg H2, Varying B/Li

11.19

Pt potential

11.14 1200

1700

2200 Test time, h (a)

Ct potential 2700

0

–0.2 –0.4 –0.6 –0.8 –1

3200

SCC#3 – c315 – Alloy 600, CRDM Tube, 93510

11.23

0.2

0.4

Outlet Conductivity ∏ 0.01

11.225

0.2

11.205

To 600 ppm B, 2.2 ppm Li @2325h

Crack length, mm

11.21

To Constant K @2011h

11.22 11.215

11.2

11.195

Conductivity, mS/cm or Potential, Vshe

3.5 ¥ 10 mm/s

0.4

4.3 ¥ 10 mm/s

0 6.6 ¥ 10–9 mm/s

–0.2 –0.4

c315 – 0.5TCT of A600 CRDM, 325C 17.5 MPa√m, 18 cc/kg H2, Pure Water

–9

–0.8

11.19

Pt potential

11.185 2000

2200

–0.6

2400 2600

2800 3000 3200 3400 Test time, h (b)

Ct potential

–1 3600 3800 4000

9.9 Crack length vs. time for Alloy 600 tested in 325 °C water containing H2 under constant K conditions. Various changes in B, Li and pHT, including additions to pure water, produced no discernible change in crack growth rate.

© Woodhead Publishing Limited, 2010

Conductivity, mS/cm or Potential, Vshe

11.24

0.6

To 60 ppm B, 0.3 ppm Li @3315h

11.29

0.8

Conductivity ¥ 0.01 To 3200 ppm B, 7 ppm Li @1880h

Crack length, mm

11.34

To 1100 ppm B, 2 ppm Li

11.39

SCC#2 – c283 – Alloy 600, CRDM Tube, 93510

To Constant K @1201h

11.44

255

Understanding and mitigating ageing in nuclear power plants 2.50

(Ni++)

(Ni(OH)3–)

2.00 (Ni(OH)2aq)

(NiOH+) 1.50

Potential (Volts VS. SHE)

(b)

1.00

NiO2 Ni

0.50

++

H

2 /H 2O

(a)

0.00

Fe 2O

3

–0.50

Ni3O4

Fe

≠H2

3O 4

Fe

NiO

–1.00

≠pH

Ni –1.50

–2.00

Ni(OH)3–

0

4

pH

8

12

16

(a)

Dissolved H2, cc/kg

256

75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 260

Film reduction and/or no film formation Film formation Experimentally measured Ni/NiO transition

Experimentally measured Ni/NiO transition

Ni regime NiO regime

280

300 320 340 Temperature (°C) (b)

360

380

9.10 (a) Ni–H2O Pourbaix diagram at 300 °C. (b) Right Ni/NiO phase boundary as a function of H2 fugacity and temperature [44].

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

28.5 28.4 28.3

2.5 ¥ 10–6 mm/s

1 ¥ 10–8 mm/s

0.1 0 –0.1 –0.2 –0.3 –0.4 –0.5 –0.6

28.2 6 ¥ 10–6 mm/s

CT Corrosion Potential

–0.7

Potential, Vshe or Conductivity, mS/cm

Crack length, mm

28.6

33 MPa√m, R = 0.7 0.01 Hz + 900 s hold

pH288C ~ 6.79 To 95 ppb H2 @2518h

28.7

c85 – 1T CT of Sens. SS – AJ9139 95 ppb H2, 1200 ppm B, 2.2 ppm Li, 288C

To 200 ppb O2 @2426h

28.8

257

–0.8 28.1 2375 2400 2425 2450 2475 2500 2525 2550 2575 2600 2625 2650 Time, h

9.11 SCC crack length vs. time for sensitized stainless steel. Adding 200 ppb O2 at 2279 h causes a marked increase in growth rate in water with 1200 ppm B and 2.2 ppm Li as LiOH at 288 °C.

9.4

Stress corrosion cracking (SCC) dependencies – materials and water chemistry

9.4.1 Type or grade of stainless steel A brief introduction to the common grades of stainless steels was given in Section 9.2. A lot of emphasis has been made on differences among grades of stainless steel, but the primary differences relate to whether sensitization and grain boundary Cr depletion occurs, which is described in more detail in the next section. Among common grades of austenitic stainless steel, there is little distinction in their SCC growth response if they are in the same metallurgical condition. That is, if there is a similar Cr depletion profile, cracks in all types of stainless steels will grow at similar rates. If unsensitized stainless steels are cold worked to the same yield strength, they all grow cracks at nearly identical rates [13–21]. This can be seen in prior and future figures (e.g., Figs 9.4, 9.8 and 9.20). Austenitic stainless steels or nickel alloys with low carbon and/or additions of Nb or Ti (also Mo) have less tendency to sensitize, and thus L-grade (low carbon) or NG-grade (low carbon, high nitrogen) or Nb/Ti-stabilized stainless steels can show less SCC susceptibility than standard Type 304 stainless steel. However, sensitization is still possible in all of these grades. They have similar work hardening characteristics, so cold working produces

© Woodhead Publishing Limited, 2010

258

Understanding and mitigating ageing in nuclear power plants

a similar increase in tensile yield strength. Stainless steels and Alloy 600 of a similar condition (level of sensitization or cold work) all exhibit about the same SCC growth rates in BWR water – there is very little difference as a function of Mo, Ti or Nb [13–21]. The fact that stainless steels and nickel alloys (e.g., Alloy 600 and Alloy 182/82 weld metal) exhibit similar crack growth rates and response vs. corrosion potential and water purity (Figs 9.4 and 9.5) suggests that a similar crack advance mechanism is operative.

9.4.2 Stainless steels vs. nickel alloys Subsequent sections will identify some unique elements of SCC in austenitic stainless steel vs. nickel alloys. In broad terms, their SCC growth rate dependencies are 80% similar. Figure 9.12 shows that the crack growth rate in 288 °C water is close to identical in stainless steels and nickel alloys, and the electrochemical response (e.g., corrosion potential and corrosion rate vs. dissolved O2) is the same. Since the metal ion solubility is also comparable, it is reasonable to expect that the crack chemistry will be essentially identical. It is not surprising that the effect on SCC of aqueous impurities, such as Cl– and SO4=, is also similar for these two classes of material. A primary difference is related to the higher temperature dependence of SCC in nickel alloys, which typically has an activation energy of ~135 kJ/ mol [41, 45–46] vs. ~80–100 kJ/mol for stainless steels [15, 16, 19, 47], which may be associated with the higher diffusivity and creep in the nickel alloys. A second difference is that the corrosion potential in deaerated water is close to the Ni/NiO phase boundary (Fig. 9.10a), and far from the Fe/ Fe3O4 phase boundary; it is effectively impossible to add sufficient H2 to drop the corrosion potential near the Fe/Fe3O4 phase boundary. For nickel alloys, being near the phase stability for nickel might be expected to be important. Indeed, the growth rate of nickel alloys peaks near the Ni/NiO phase boundary, with a peak height of 2.5–3¥ for Alloy 600 and 8–20¥ for Alloy 182/82 weld metal [41–43]. While few comparative crack initiation tests have been performed on stainless steels and nickel alloys, field experience indicates that crack initiation occurs more readily in nickel alloys in PWR primary water, e.g., at 320–340 °C than at 274 °C (BWRs) or than in stainless steel. This is probably because of the higher creep rate in nickel alloys at elevated temperatures or in stainless steels. While thermal creep at these temperatures would not be a significant problem at constant load, it becomes more important as crack advances and the strain field is redistributed and creates dynamic strain.

9.4.3 Corrosion potential and water chemistry There are two distinct mechanisms for the effects of corrosion potential on SCC growth rate, one related to the oxidants in the bulk water that © Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

12.15 1000

0.2 Outlet Conductivity 1.8 ¥ 10 mm/s 1.9 ¥ 10–8 mm/s

2.7 ¥ 10–7 mm/s 1200

0.1

CT Potential 1400

1600

1800 Time, h (a)

–7

2000

–0.3 –0.4 –0.5

2200

2400

0.2

CT Potential

0.1

23.5 4700

2.1 ¥ 10 mm/s

2.0 ¥ 10 mm/s

3.2 ¥ 10–8 mm/s

–7

To 2000 ppb O2 @5107h

To 95 ppb H2 @4816h

Outlet Conductivity To const 31.1 MPa√m To 200 ppb @3449h O2 @4174h

Crack length, mm

23.7

23.55

–0.6

0.3 Pt Potential

23.6

–0.1 –0.2

23.75

23.65

0

–7

0 –0.1 –0.2 –0.3

Annealed + 20%CW Alloy 600 30 MPa√m, 2000 ppb O2, Pure Water

–0.4 –0.5 –0.6

4800

4900

5000 Time, h (b)

5100

5200

Potential, Vshe or Conductivity, mS/cm

12.25

0.3

To 2000 ppb O2 @2108h

12.35

To 6% H2 in Ar @1245h

12.55 To static load @367h Pt/Rh coated @960h

Crack length, mm

12.65

12.45

0.4

SSC of c126 – 316L SS 20% Cold Work 27.5 MPa√m, 2000 ppb O2, Pure Water

Potential, Vshe or Conductivity, mS/cm

12.75

259

–0.7 5300

9.12 Crack length vs. time for an unsensitized, 20% cold-worked 316L stainless steel (a) and Alloy 600 (b) showing similar growth rates and effects of corrosion potential.

create a gradient in corrosion potential and altered crack chemistry, and one related to the presence of reductants (especially H2) at the crack tip. Oxidants increase the corrosion potential on the external surface and create an electrochemical (differential aeration) cell vs. the region inside the crack (where O2 is consumed, and therefore the corrosion potential is low). This

© Woodhead Publishing Limited, 2010

260

Understanding and mitigating ageing in nuclear power plants

electrochemical cell produces a shift in the crack chemistry. In deaerated water, changes in dissolved H2 or shifts in pH (from B/Li, ammonia, etc.) also shift the corrosion potential, but these species are not consumed in the crack, and thus the corrosion potential is essentially identical throughout the crack, and no unique crack chemistry forms. Altered crack chemistry can also form from temperature gradients (especially boiling). Oxidants markedly elevate the corrosion potential on the free surface above that in argon (Ar) or nitrogen (N2) deaerated water, e.g., to between lines (a) (H2/H2O) and (b) (H2O/O2) in Fig. 9.10a. The elevation in corrosion potential occurs only on the external surface and near the crack mouth; it does not persist far into the crack because oxidants are rapidly consumed once any convective flow decays within the crack. Oxidants have an indirect effect on SCC; the potential gradient that forms near the crack mouth (Fig. 9.13) produces a change in crack chemistry that affects the entire crack. Figure 9.13 shows that electromigration, which concentrates anions in the crack, is balanced by back-diffusion of species out of the crack (if convection is present, it generally overwhelms the contributions of electromigration or diffusion). The dynamic equilibrium between electromigration and diffusion occurs a short distance into most cracks, and after the dynamic equilibrium in concentration is initially formed, the species move deeper into the crack by ordinary diffusion. The gradient in corrosion potential is determined by both the external corrosion potential and the corrosion potential within the crack. The corrosion Cl–

Time evolution after increase in ECP of –0.5 to 0.1 Vshe

fC = +0.1

O 2 , fC fC = –0.5 Vshe

Df

e–

e– Cl–, SO42–, OH–

1 Ni Æ Ni2+ Æ H+ 2 e– Microcell 1

Ni Æ Ni2+ + 2e– + H2O Æ NiO + 2H+

Zn2+ Æ e 3 Microcell



4 O2 + 2H2O + 4e– = 4OH–

3 H2 Æ 2H+ + 2e–

2 2H+ + 2e– Æ H2 JA = –DADCA – zmCAFDf + CAV flux = diffusion + f-driven + convection

9.13 Schematic of crack chemistry transport processes in high temperature water with O2.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

261

potential in the crack is controlled by the pH and the concentration of H2 in the crack, but it remains along the H2/H2O line that controls the potential in deaerated water (Fig. 9.10a). The change in pH in the crack is generally more extreme than the ~1.5 unit change in high temperature pH from the addition of B/Li to deaerated water. If most impurities are species like Cl– and SO4=, acidification occurs. This produces a somewhat higher corrosion potential in the (deaerated) crack because of the effect of pH on line (a) in Fig. 9.10a, which reduces the potential gradient somewhat. The potential gradient does not always cause acidification of the crack. Acidification requires the presence of anions other than OH– (such as SO4= or Cl–) to balance the charge of the H+ cation. In high purity water containing oxidants, the pH in the crack shifts in the alkaline direction (higher pH), with the increase in OH– concentration balanced by an increase in metal ion solubility. An alkaline shift also occurs if species such as sodium hydroxide (NaOH) or potassium hydroxide (KOH) are the predominant impurities. As the crack tip moves in the alkaline direction (higher in pH), the corrosion potential in the deaerated crack drops because of the pH change (line (a) in Fig. 9.10a), and the potential gradient increases. However, the balancing cations (e.g., Na+) also tend to be rejected from the crack by the potential gradient. The effect of oxidants on crack growth is quite similar for sensitized and unsensitized, cold-worked stainless steel (Fig. 9.4), and for Alloy 600 and Alloy 182 weld metal (and other nickel alloys and irradiated stainless steel, Fig. 9.5). Detailed examples of the crack length vs. time response for changes in corrosion potential are shown in Fig. 9.12 for stainless steel and Alloy 600 – many other similar examples are shown in References [10–23]. The oxidant concentration itself is not the controlling factor. Electromigration is controlled by the potential gradient, not by the oxidant concentration (or gradient) per se. The relationship between dissolved O2 and corrosion potential is not linear, but rather follows a complex relationship (Fig. 9.14). Above ~2–20 ppb O2 (in tests with well-controlled chemistry), the corrosion potential follows a roughly logarithmic relationship, as predicted by the Nernst equation. At lower oxidant concentrations, the potential falls rapidly to that associated with deaerated water. The reason for the rapid drop is that O2 becomes mass transport limited, with O2 consumed on the surface (by corrosion or reaction with H2) faster than it can diffuse though the nearsurface, stagnant boundary layer. The O2 concentration at which the potential drops depends on the O2 reaction rate (including corrosion rate and reaction with H2) and the convection (flow) rate. As the oxidant level and corrosion potential decrease, the growth rate also decreases but does not cease. Another mechanism of corrosion potential effect on crack growth is associated with the presence of H2 in deaerated water (note that this mechanism can also be operative if the bulk water has stoichiometrically more H2 than

© Woodhead Publishing Limited, 2010

262

Understanding and mitigating ageing in nuclear power plants

400

200

Electrode potential, mVshe

g g g g

0

–200

g g

–400

g

g

e

eee

e

gg

d

• • • •

–600

–800

100

Laboratory and in-reactor data for 304/316 stainless steel 288 °C water Unirradiated condition Conductivity range .1 – .5 mS/cm Variety of flow rates Both bright and oxidized surfaces

101 102 Dissolved oxygen content, ppb

103

104

9.14 Effect of dissolved oxygen on corrosion potential in 288 °C water. Note that modern measurements made with improved water purity and higher refresh rates tend to represent the data where the large drop in corrosion potential occurs at the lowest O2 concentration. Also, improved, more accurate reference electrodes show the corrosion potential to be perhaps 50 mV higher than the upper line.

O2). The presence of H2 always decreases corrosion potential; H2 decreases the corrosion potential in aerated or deaerated water, although by a smaller magnitude than the increase from oxidants. For a tenfold increase in H2 in deaerated water, the corrosion potential of Pt, stainless steel and nickel alloys decreases by 55 mV at 281 °C and 60 mV at 332 °C (according to the Nernst equation, which depends on temperature in Kelvin). Unlike oxidants, H2 is not consumed in the crack – some H2 might be created by corrosion reactions, or some might be lost by transport through the metal, but these are small factors. Generally, it is reasonable to assume that the H2 level at the crack tip is the same as the bulk H2 concentration (or the H2 that remains after reaction with oxidants in the bulk water). An increase in H2 causes the H2/H2O line to shift vertically downward on a Pourbaix diagram (Fig. 9.10a), and for stainless steels and nickel alloys the corrosion potential in deaerated water is exactly thermodynamic and identical to the response on catalysts like platinum (Pt) [10–23, 39–43]. Among Fe, Cr and Ni, only Ni/NiO has a metal-metal oxide equilibrium near the H2/H2O line. Thus, it is reasonable to expect that the crack growth

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

263

rate response of nickel alloys might be affected, but stainless steels not. References [10, 13–19, 21] give examples of the lack of effect of H2 on the SCC growth rate of stainless steel, although it should be noted that effects of H2 are observed in crack initiation under corrosion fatigue initiation [48, 49], corrosion fatigue crack growth [50] and slow strain rate testing [51, 52], perhaps because the corrosion rate is higher at low corrosion potential [10, 14]. For nickel alloys, a peak in crack growth rate is observed at the Ni/NiO phase boundary (Fig. 9.10a), which occurs at different H2 levels as a function of temperature (Fig. 9.10b). The growth rate response is symmetrical and the growth rate similar on the NiO and Ni-metal sides of the phase boundary, making it unlikely that there is a change in crack growth mechanism. Morton et al. [44, 53–56] found that the peak height is higher for Alloy 82 weld metal and Alloy X-750 than for Alloy 600 (Fig. 9.15a), but that the peak height and shape were unchanged over the 260–360 °C range they evaluated. Andresen et al. [39–43] have confirmed the effect on Alloy 600 and Alloy 182 weld metal (Fig. 9.15b), but have seen a stronger effect of H2 (higher peak) of 15–20¥. Bruemmer et al. [57] also observed a higher peak for Alloy 182. This is likely due to their use of constant K (no cycling) vs. Morton’s use of periodic partial unloading throughout the test, because cycling biases low growth rate data upward more than high growth rate data. Figure 9.5c shows the effect of H2 in the context of oxidants. At high corrosion potential the presence of oxidants causes a relatively high crack growth rate. As the oxidant level decreases, the growth rate drops. If H 2 is added to deaerated water, the corrosion potential begins to shift down. For example, 100% O2 bubbled in pure water at standard temperature and pressure (STP) gives ~42 ppm dissolved O2, but even 0.1 ppm O2 can increase the corrosion potential by >500 mV compared to deaerated water. 100% H2 gives 1.58 ppm dissolved H2, and shifting by 100¥ (e.g., from 10 ppb to 1000 ppb H2) only causes a 114 mV decrease in potential at ~300 °C. As H2 is changed and the Ni/NiO phase boundary crossed, there is a peak in the crack growth rate of nickel alloys, as shown in Fig. 9.15. While it has not been verified in deaerated pure water, the data of Morton et al. showing that the peak height is unchanged from 260–360 °C. The observation described above shows that the crack growth rate is unaffected by B/Li additions to deaerated water [39, 40] and provides a strong basis for anticipating that H2 affects SCC growth of nickel alloys in pure water. Since most BWR structural materials are exposed to 274 °C water, the Ni/ NiO phase boundary and associated peak in crack growth rate occurs at ~200 ppb H2 (Fig. 9.10b), which is the regime in which ‘medium HWC’ BWRs operate.

© Woodhead Publishing Limited, 2010

264

Understanding and mitigating ageing in nuclear power plants SCC growth rate/minimum SCC growth rate

3.0 K = 60 ksi√in, 338 °C K = 25 ksi√in, 338 °C K = 60 ksi√in, 316 °C

2.5

2.0

1.5

1.0

0.5

0.0 –100

–50

0 EcPNi/NiO – EcP (mV) (a)

50

100

Effect of H2 on Crack Growth Rate of 182 Weld Metal – c380, 325C 6.E–08

Observed Predicted

Crack growth rate, mm/s

5.E–08

Predictions scaled to CGR at 10.4 cc/kg H2. The decrease in CGR was larger than predicted based on an 16X peak vs. H2.

4.E–08 3.E–08 2.E–08 1.E–08

0.E+00 4.16

10.4

10.4 10.4 Dissolved H2 in test, cc/kg (b)

26

80

9.15 Crack growth rate of Alloy 600 (a) [44] and Alloy 182 (b) [42, 43] in high temperature water as a function of potential/H2. The peak growth rate occurs very close to the Ni/NiO phase boundary, with a characteristic height and width associated with the material and condition.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

265

9.4.4 Temperature The effect of temperature is somewhat complex when considered in the range of ~25–350 °C. Focusing on operating temperatures (274–288 °C for BWRs and 288–343 °C for PWRs) and low corrosion potential conditions, stainless steels and nickel alloys show a significant temperature dependency, although the effect on crack growth is stronger on nickel alloys (at least Alloy 600 and its weld metals) at ~ 135 kJ/mole [41, 45, 46] vs. ~80–100 kJ/ mole for stainless steel [15, 16, 19, 47]. For 135 kJ/mole, the difference in growth rate is ~13¥ from 288 to 343 °C, while for 90 kJ/mole, the difference is 5.6¥. The lower activation energy for stainless steel translates to a lower incidence of cracking of these components in PWRs compared to Alloy 600 and its weld metals, but the limited incidence of cracking in stainless steel PWR components is certain to increase with time. For an activation energy of 135 kJ/mol, a temperature change from 274 and 343 °C increases the growth rate by ~28¥, and explains why good SCC mitigation is achieved by reducing corrosion potential of materials in BWRs (most structural materials operate at 274 °C), while problems have developed in the low corrosion potential PWR primary water, especially in nickel alloys. Temperature also affects the dissolved H2 concentration in the coolant, where the peak in growth rate occurs in nickel alloys (at the Ni/NiO phase boundary) and the dissolved H2 associated with the phase boundary varies with temperature (Fig. 9.10b). While Attanasio and Morton [44] reported that the peak in crack growth rate may not always occur exactly at the Ni/NiO phase boundary, the data are too few and too scattered to convincingly support a deviation from the Ni/NiO phase boundary, especially since the deviation was sometimes slightly positive and sometimes slightly negative. The effect of temperature over wider ranges (to lower temperatures) is more complex. When the environment is more aggressive (resulting in higher growth rates), there is a tendency for the increase in growth rate vs. temperature to be monotonic (Fig. 9.16), but if the water chemistry is unaggressive (e.g., less oxidizing), a peak in growth rate is generally observed at ~175 °C (Fig. 9.16). An interest in this wider temperature regime is usually related to plant start-up and shutdown, but these involve many other changes in addition to temperature. In BWRs, the water often has higher impurity levels and is more oxidizing due to prior exposure to air, from radiolysis and/or because H2 is not injected during shutdown or start-up. There are also differences in pressure loading, vibration, and differential thermal strains, e.g., in the ferrite containing stainless steel weld metal vs. the base metal (all austenite) due to the different coefficients of thermal expansion. In PWRs, H2O2 is generally added during cool down to consume most of the H2 and minimize transport of radioactive crud; during heat-up, the water is more oxidizing due to exposure to air during shutdown.

© Woodhead Publishing Limited, 2010

SCC of high strength, cold worked austenitic steels in pure water, oxygen 0.1–300 ppm, K = 50 MNm–3/2 10–5

X50MnCr185, Rp0.2 = 1000 MPa X6MnCrN1818, Rp0.2 = 1500 MPa

–7

10–8

10–9

300 250 200

°C

25 10–3

Q = 10 kcal/mole

Air Sat’d 0.224 mS/cm HCl 0.077 mS/cm HCl

10–6 10–4

10–7

10–11 0.0015

(a)

100

Open = temp. ≠ Close = temp. Ø

10–10

10–12 1.6

150

in/h

Da , (m/s) Dt Stress corrosion crack growth rate,

© Woodhead Publishing Limited, 2010

10

Effect of Temperature Constant K = 33 MPa√m + R = 0.5, 0.01 Hz every 1000s 200 ppb O2, 0.27 mS/cm H2SO4

0

2.0 2.4 2.8 3.2 3.6 Reciprocal temperature, 1/T (1/K)·103

4.0

Sens Sens Sens Sens

304 St.St. AJ9139, C49 Alloy 600 NX8608, C46 304 St.St. 2P4932, C52 304 St.St. 71635, C53

0.002 0.0025 0.003 Inverse temperature, K–1

10–5

0.0035

(b)

9.16 Effect of temperature on stainless steel in aggressive (a) and more typical (purer) water chemistry (b). Even in less aggressive environments, an increase in growth rate is observed as temperature is increased.

Understanding and mitigating ageing in nuclear power plants

10–6

20

Crack growth rate, mm/s

10–5

Temperature, T, (°C) 200 160 100 60

266

288

Stress corrosion cracking of austenitic stainless steels

267

9.4.5 Intergranular crack morphology, even in unsensitized materials While an intergranular (IG) crack morphology is not surprising for sensitized (Cr-depleted grain boundaries) stainless steels and nickel alloys, it might not be expected in unsensitized materials. Figure 9.17 shows the IG crack morphology in unsensitized, cold-worked Type 316L stainless steel. There is no evidence of a role of grain boundary segregants (e.g., S, P, B, N and C), since similar (or faster) IG growth rates are often observed in high purity alloys, and no significant effect of heavily impurity doped alloys was observed [58]. The preference for cracks to follow the grain boundary can be strongly promoted by grain boundary chemistry (e.g., Cr depletion or Si enrichment), but is also promoted by preferential deformation under loading in the grain boundary, which is a common explanation for intergranular SCC in stainless steels with no measurable chemical difference in the grain boundary (especially Cr depletion). One measure of the preference for cracking in the grain boundary is the tendency for the crack morphology to shift from IG to

9.17 Intergranular SCC morphology in unsensitized, 20% cold-worked Type 316L stainless steel tested in 288 °C BWR water.

© Woodhead Publishing Limited, 2010

268

Understanding and mitigating ageing in nuclear power plants

transgranular (TG) with an increase in DK and frequency, the applied strain rate, or even Kmax in tests without cycling. In heavily sensitized stainless steel in aggressive (oxidizing and impure) water, IG cracking can be sustained to relatively high DK or frequency, whereas very resistant materials in unaggressive environments can shift to TG morphology even at very low DK or frequency. Indeed, materials that are very resistant to SCC may exhibit TG cracking without cycling. No fundamental difference should be attributed to a TG vs. an IG crack path – an IG morphology reflects a chemical or strain rate preference between the grain boundary vs. slip planes or twin boundaries. The tendency to dismiss TG observations originates primarily from the early use of highly accelerated tests, e.g., employing slow strain rate or cyclic loading.

9.4.6 Grain boundary cr depletion vs. grain boundary carbides Sensitization resulting from grain boundary chromium (Cr) depletion occurs in stainless steels and nickel alloys because various types of Cr carbides form in the temperature range of ~500–750 °C. Chromium carbide formation ties up significant amounts of Cr in and near the grain boundaries, thus rendering the matrix in the boundary poor in Cr, relative to the bulk composition, and susceptible to corrosion and SCC. The most common Cr carbide in stainless steel is Cr23C6, although other forms exist. Carbon (C) atoms diffuse much faster than Cr atoms, and the growth of the carbide produces a Cr depletion profile near the grain boundary. Since grain boundary diffusion is faster than matrix diffusion, the entire grain boundary is depleted in Cr, even though the carbides are discontinuous. Depending on the carbide formed, it becomes stable below about 1000 °C, but the nucleation rate is fastest at a peak temperature between about 650 and 750 °C (depending on composition of the steel), and is more sluggish at higher or lower temperatures. Below about 550 °C, nucleation is sluggish, although in cold-worked stainless steels, carbides can form at temperatures as low as 400 °C in a few hours. Grain boundaries are the preferred sites for nucleation, although in cold-worked materials extensive intragranular nucleation occurs. Cr borides can also form and cause sensitization, usually in materials with low C or elevated B. Time-temperature-sensitization (TTS) diagrams are used to show the time to nucleation vs. temperature. Once the carbide nucleates, it can grow over a wide temperature range. Two key factors are the Cr activity at the carbide interface, and the Cr diffusion kinetics. Above 900–1000 °C, the Cr activity at the carbide interface is higher than the Cr activity in the matrix, and the carbide dissolves. At 700–800 °C, the Cr activity at the carbide boundary is somewhat below the matrix level, and the level of Cr depletion in the grain boundary is moderate (e.g.,

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

269

12–15% Cr). The Cr depletion profile is broad because of the more rapid diffusion. At lower temperatures (e.g., 500 °C), the Cr activity at the carbide interface is very low, so the grain boundary Cr concentration is lower (e.g., 5–8% Cr) and the Cr profile is narrow. At LWR operating temperatures of ~ 300 °C, carbides continue to grow, but very slowly – this is called low temperature sensitization. For many stainless steels, a short high temperature (e.g., 1050 °C for 30 min) heat treatment dissolves the Cr carbides; this is called a solution anneal. When all of the free carbon is consumed, carbides stop growing but the Cr continues to diffuse, eventually eliminating the Cr depletion profile – a process called healing. This is difficult to achieve in stainless steels because of its lower Cr diffusivity, but is used commercially in nickel alloys (e.g., in thermally treated steam generator tubing) to create grain boundary carbides without Cr depletion. Sensitization is often measured using standard corrosion or electrochemical tests, such as the EPR test (electrochemical potentiokinetic repassivation). The EPR test reflects the area (width) along the grain boundary where the Cr concentration is below ~15% Cr–it does not measure the minimum Cr level in the grain boundary, which controls SCC susceptibility [59, 60]. Thus, EPR data can be very misleading, because very high readings can result from higher temperature sensitization heat treatments that have very wide but relatively shallow Cr depletion, while very low readings can result from low temperature treatments that give narrow but deep Cr depletion profiles. Sensitization is mitigated by controlling the alloy composition and restricting exposure to temperatures where it occurs most readily. Low C (L-grade) alloys, such as types 304L and 316L stainless steels, are common, and the presence of ~2.5% Mo in 316(L) further inhibits carbide nucleation. Current melting practice typically results in a 0.04–0.05% C steel, down from the 0.07–0.08% C range, which was common 40 years ago. The specification for L-grade steels is <0.03% C, but in practice a C content of 0.015–0.020% is common. The addition of titanium (Ti) or niobium (Nb) (in types 321 and 347 stainless steel, respectively) helps to reduce or eliminate sensitization by reacting with the free C in TiC or NbC carbides, which are more stable and form at higher temperatures than Cr carbides, and limit the free carbon available. These MC carbides are primarily intragranular and, because they form at higher temperature, do not result in grain boundary Cr depletion. However, during welding, the area adjacent to the weld fusion line heats above ~ 1200 °C, and MC carbides can partially dissolve. Because cooling to 800 °C in welds is rapid compared to solidification in an ingot, the MC carbides do not have a chance to tie up the C, and the cooling rate below 800 °C can permit the formation of grain boundary Cr carbides and Cr depletion. IG corrosion can occur in the heat affected zone and is known as knife line attack. IG SCC will obviously also be accelerated.

© Woodhead Publishing Limited, 2010

270

Understanding and mitigating ageing in nuclear power plants

Historically, SCC in nickel alloys has been considered to be fundamentally different in PWRs than in BWRs because grain boundary carbides were advantageous in PWRs (e.g., in steam generator tubing and other Alloy 600 components) but detrimental in BWRs (e.g., furnace and weld sensitization in stainless steels and Alloy 600). However, controlled studies on stainless steels with grain boundary carbides, but without Cr depletion, show that the carbides were beneficial in BWR chemistries (Fig. 9.18). Cr depletion is the detrimental factor, primarily under oxidizing (external) water chemistry conditions where the crack tip chemistry/pH is aggressive; sensitization is a much smaller factor in PWR or BWR environments where the materials are at low corrosion potential.

9.5

Stress corrosion cracking (SCC) dependencies – cold work, stress intensity factor and irradiation

9.5.1 Cold work from bulk deformation, surface cold work, weld residual strain, etc. Components that had bulk cold work cracked relatively early in BWR service, and subsequent designs restricted cold work to less than a few 1

0.4 Comparison of 20% CW SSs 27.5 MPs√m, 2000 ppb O2, Pure water

0.3

Crack length, mm

0.8

0.2

0.7

Alloy 600 2 ¥ 10–7 mm/s

Last 60% of test at 200 ppb O2

0.6

0

347 SS 2.3 ¥ 10–7 mm/s

0.5 0.4

–0.1 304L SS 1.5 ¥ 10–7 mm/s

316L SS 0.3 1.6 ¥ 10–7 mm/s

–0.2 –0.3

304 GB Carbides 2 ¥ 10–8 mm/s

0.2

–0.4 –0.5

0.1 0

0.1

0

200

400

600 800 Test time, h

1000

1200

–0.6 1400

9.18 Crack length vs. time in 288 °C water for 20% cold-worked stainless steels (and Alloy 600) with and without grain boundary carbides. Grain boundary carbides were formed at 621 °C for 24 h, then Cr depletion was eliminated by an equilibration heat treatment at 950 °C for 5 h, followed by water quenching.

© Woodhead Publishing Limited, 2010

Conductivity, mS/cm or potential, Vshe

0.9

Stress corrosion cracking of austenitic stainless steels

271

percent. However, surface cold work from machining, grinding and related processes is more difficult to control or quantify, and has been responsible for accelerated initiation in many LWR components. Importantly, shrinkage from weld solidification produces residual strains as well as residual stresses near the weld. The development of electron backscattered diffraction (EBSD) permitted measurement of residual strains on a very fine scale [61–64]; EBSD was extensively calibrated using both tensile and double-cone compression specimens of many materials. It was discovered that peak weld residual strains of 20–30% (equivalent room temperature tensile strain) are often present near welds. These peak values occur at the root of the weld (e.g., at the inside diameter of welded pipes) and decrease toward zero within a few millimeters from the fusion line (Fig. 9.19). Lower residual strains were observed near the final (outside) welding passes, and they also tapered towards zero away from the fusion line. The importance of the weld residual strains on SCC was directly measured on weld heat affected zone aligned specimens (e.g., Fig. 9.20), which exhibited high crack growth rates consistent with bulk cold-worked specimens. The historical inattention to weld residual strain in sensitized stainless steel piping

20

Various BWR stainless steel weld HAZs

Strain, %

15

10

5

0

0

5

10 15 20 25 30 Distance from weld fusion line, mm

35

9.19 Equivalent room temperature tensile strain in the heat affected zone adjacent to the weld fusion line determined from electron backscattered diffraction. The peak strain always occurs near the fusion line and at the root of the weld, with peak values of 20–30% strain being common both in stainless steel and Alloy 600 HAZs.

© Woodhead Publishing Limited, 2010

272

Understanding and mitigating ageing in nuclear power plants 0.4

11.9

0.35 0.3

11.7 3.2 ¥ 10–7 mm/s

Corrosion potential of Pt

0.25

To 200 ppb O2 @565h

11.6 11.5 11.4 11.3 11.2

Corrosion potential of CT

0.2 0.15 0.1

Outlet conductivity 0.05

0

100

200

300

400 Time, h

500

600

700

Conductivity, mS/cm or potential, Vshe

11.8

Crack length, mm

2.2 ¥ 10–7 mm/s

0.5T CT of HAZ Aligned 348 Weld 288C, 2 ppm O2, 27.5 MPa√m

0 800

9.20 SCC response for a Nb-stabilized, high N-bar stainless steel specimen aligned along the weld HAZ. This and similar observations confirm the detrimental role of shrinkage strains adjacent to weld HAZs.

is not surprising because the peak in sensitization in Type 304/316 stainless steel does not occur near the weld fusion line, but typically 5–15 mm away where the thermal profiles favor nucleation and growth of grain boundary carbides. At the 5–15 mm distance, the residual strain decreases to quite low levels, and ignoring its contribution represents a small oversight. As SCC has developed in unsensitized stainless steel weld heat affected zones, the role of residual strain has become very important, and indeed most cracks occur close to the weld fusion line. The effect of bulk cold work on SCC has been studied in many materials (Fig. 9.21) [7–22], and is best characterized in terms of a basic effect of yield strength on SCC growth rate. Factors that increase the yield strength (e.g., cold work, precipitation hardening and irradiation) tend to increase SCC growth rates, all other factors being equal. The crack growth rates are accelerated at high corrosion potential, but remain relatively high in deaerated or in PWR water (Figs 9.4, 9.5 and 9.12). Indeed, Fig. 9.4 shows that lower crack growth rates can be achieved at low potential on sensitized materials than on cold-worked materials. While the incidence of SCC in stainless steel welds has not been as high in PWRs as in BWRs, the growth rates observed in the laboratory at low corrosion potential suggest that it will become more evident over time. High yield strength materials (which can develop from high cold work,

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels 1.e–07

Crack growth rate, mm/s

Unsensitized 304, 304L & 316L SS & A600 288 °C High purity water, 95 or 1580 ppb H2 CT Tests at 27.5 – 30 MPs√m  = High martensite SS Predicted  = Alloy 600 Response

273

Very high martensite

2 Sensitized Points for Comparison

1.e–08

Very low or no martensite

Annealed   Cold Worked at –55 °C or +140 °C 1.e–09

0

1.E–06

200

300 400 500 Yield strength, MPa (a)

600

Crack growth rate, mm/s

Unsensitized 304, 304L & 316L SS & A600 288 °C High purity water, 2000 ppb O2 CT Tests at 27.5 – 30 MPa√m  = High martensite SS  = Alloy 600 Predicted

700

800

Very high martensite

Response

2 Sensitized Points for Comparison

1.E–07

1.E–08

100

Very low or no martensite

Annealed   Cold Worked 0

100

200

300 400 500 Yield strength, MPa (b)

600

700

800

9.21 Effect of yield strength and martensite content on the SCC growth rate on stainless steel and Alloy 600 in 288 °C, high purity water (ª0.06 mS/cm outlet) containing 95 or 1580 ppb H2 (a) or 2 ppm O2 (b).

neutron irradiation, etc.) can exhibit both high growth rates and a more limited effect of corrosion potential at high stress intensity factor or under ‘gentle’ cyclic loading conditions (Fig. 9.22), which typically enhance the SCC growth rate only very slightly. In some high yield strength materials, rapid crack advance is observed during the reloading portion of the waveform, with rates

© Woodhead Publishing Limited, 2010

Understanding and mitigating ageing in nuclear power plants 0.3 Outlet conductivity

13.75

0.2

13.6

13.55 13.5

0 –0.1

2.2 ¥ 10–7 mm/s

–0.2

To 1.58 ppm H2 @3396h

Crack length, mm

13.65

To R = 0.7, 0.001 Hz + 85,400s hold @1808h To 2000 ppb O2 @3037h

0.1 13.7

–0.3 –0.4 CT potential

c157 – 0.5T CT of 316LSS, 50%CW 32 – 34.5 MPa√m, Pure water

13.45

PT potential

K slowly rises from 32 to 34.5 MPa√m in this graph 13.4 3100

16.8

3150 3200

3250

3300 3350 3400 Time, h (a)

3450

–0.8 3500 3550 3600

Because c157 is growing faster than c156, K slowly rises from 45 to 49 MPa√m in this graph

16.55 16.5 16.45 5500

Average rate: 5 ¥ 10–7 mm/s

5560

5580 5600 5620 Time, h (b)

–0.2 –0.3 –0.4

CT potential

–0.5 –0.6 –0.7

PT potential 5520 5540

0.1

–0.1

To 100% H2 @5559h

16.6

340C @4818h 100% N2 @5419h R = 0.7, 0.001 Hz+ 400s hold @5176h

16.65

During rapid growth: ~ 1 ¥ 10–4 mm/s over 500 s load rise time

0.2

0

c157 – 0.5T CT of 316LSS, 50%CW 45–49 MPa√m, Pure water

To To To 85

Crack length, mm

16.7

–0.6 –0.7

Outlet conductivity 16.75

–0.5

Conductivity, mS/cm or potential, Vshe

13.8

Conductivity, mS/cm or potential, Vshe

274

5640

–0.8 5660 5680 5700

9.22 Crack length vs. time for a 0.5TCT specimen of unsensitized 316L stainless steel ‘cool’ worked to 50% showing the effect of gentle unloading cycles on environmental crack advance on stainless steel whose yield strength is elevated by cold work.

of 10–4 mm/s and higher. Whether this ‘gentle’ cyclic phenomenon at a load ratio, R = 0.7 (R is Kmin/Kmax) would be observed under higher frequency, lower amplitude conditions still needs to be evaluated. High growth rates are

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

275

also observed as the stress intensity factor increases at constant K. Figure 9.23 shows the very high growth rates and, while there is some reduction as the corrosion potential is decreased, the rate rapidly increases as stress intensity factor increases.

3.4 ¥ 10–7 mm/s

4 ¥ 10–7 mm/s Pt potential

1100

1200

20

18.5 18 17.5 17 16.5

16 3800

–6

3820

5 ¥ 10–6 mm/s

3840

1700

2 ¥ 10 mm/s

Outlet conductivity

4 ¥ 10 mm/s

1600

–0.2 –0.3 –0.4

0.3 –4

3 ¥ 10–5 mm/s

3900

–0.1

0.4

1.5 ¥ 10–6 mm/s

3860 3880 Time, h (b)

0

–0.6 1800

c181 – 0.5T CT of 316L 50%CW 2000 ppb O2, Pure water

61 MPa√m

Crack length, mm

19

1500

CT potential

To constant load & 95 ppb H2 @3851h

19.5

53 MPa√m To 2 ppm O2 @2389h NobleChem @1837h To 0.001 Hz + 9000s hold @3037h

Pt potential

1300 1400 Test time, h (a)

3920

0.2 Specimen failed @3927h Actual K at failure = 124 MPa√m

1000

0.1

–0.5

CT potential 12 900

0.2

0.1 0 –0.1 –0.2 –0.3 –0.4 –0.5

–0.6 3940

9.23 Crack length vs. time for 0.5TCT specimens of unsensitized types 304L and 316L stainless steel ‘cool’ worked to 50% showing the effect of high yield strength and high stress intensity on environmental crack advance and fracture toughness.

© Woodhead Publishing Limited, 2010

Conductivity, mS/cm or potential, Vshe

13

1.1 ¥ 10–6 mm/s

Outlet conductivity

0.3

Conductivity, mS/cm or potential, Vshe

14

Stop cycling @1255h K = 45 MPa√m

15

To 6% H2 in Ar @829h To R = 0.7, 0.001 Hz + 9000s hold @890h

Crack length, mm

16

5.2 ¥ 10–6 mm/s

48 MPa√m

17

End of test @1758h K = 64 MPa√m

0.4 c182 – 0.5T CT of 304L ~50%CW 2000 ppb O2, Pure water

276

Understanding and mitigating ageing in nuclear power plants

At lower temperatures (<150 °C), martensite in stainless steel can have a significant effect on cracking. In high temperature water, the effect of martensite was evaluated by rolling or forging at –55 ºC where very high levels of deformation-induced martensite can form. However, Fig. 9.21 shows that there is no consequential difference in crack growth rate at a given yield strength in materials with high or low (or no) martensite, including in Alloy 600, where martensite does not form. To evaluate a possible role of hydrogen on SCC of stainless steels, H2 permeation was measured [7–22, 63–68]. Hydrogen exists in the coolant as a dissolved gas (H2), but dissociates to H0, adsorbs on the surface, then permeates the metal in atomic form. Hydrogen permeation is controlled by the coolant H2 fugacity, which is non-zero even in water containing no H2 (Fig. 9.24). Once hydrogen had permeated into the 4.6 mm ID tube, reducing the H2 fugacity in the water readily produced dissociation of H2 and permeation back into the water. Since the hydrogen permeation rate is high compared to the H2 generation rate from corrosion, radiolytic proton injection, or transmutation, it is very unlikely that a high H 2 fugacity (i.e., well above the coolant H2 fugacity) is generated in metals exposed to hot water. Measurements of high hydrogen concentration in metals [25, 69] are a reflection of hydrogen storage locations, e.g., in heavily cold-worked materials (as atoms), or at radiation-induced voids (as atoms on the surface and H2 in the void) – not of a high hydrogen fugacity that can have extraordinary effects on sub-critical crack advance. No difference in the growth rate of stainless steels was observed when changing from N2 deaerated water to various levels of H2 [13, 15–19]. Similarly, the effect of electrocatalytic species (e.g., Pt) on the surface has no accelerating effect on the crack growth rate in deaerated water [34–37]. Under dynamic strain conditions, such as corrosion fatigue, some detrimental effect of lower corrosion potential from dissolved H2 is observed on crack initiation [48, 49] and crack growth rates of stainless steels [50] as noted earlier.

9.5.2 Irradiation It is not the intent of this chapter to discuss the details of irradiation assisted SCC, which are presented in references [7, 24–28]. Irradiation assisted is an apt term – the primary effects of irradiation on SCC (Figs 9.1 and 9.5a) are associated with radiation hardening, radiation induced segregation (especially Cr depletion and Si enrichment), radiation creep relaxation and radiolysis (insofar as it elevates the corrosion potential of the metal). Figure 9.5a highlights the effect of radiation hardening – the large-triangle data point at low corrosion potential shows that the growth rate is similar to cold-worked stainless steel of similar yield strength. The large-triangle

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels 10000

0.2 Outlet conductivity To 1580 ppb H2 @2459h

Closed end tube of 304LSS 95 ppb H2, Pure water

8000

4000 3000 2000

0

85 microns/h

–0.1 –0.2 Repeat data 1000h later –0.3 –0.4 CT potential

30 microns/h

Pt potential

–0.6 –0.7

1000 0 2350

2370

2390

2410

2430

2450 2470 Time, h (a)

2490 2510

–0.8 2530 2550

0.3

10000 Outlet conductivity

9000

0.2

Pt potential

8000

0.1

CT potential

7000 H2 pressure, microns

–0.5

315 °C water –90 microns/h

6000 5000

0 –0.1

304LSS closed end tube 2000 ppb O2, Pure water

–0.2 –0.3

4000 3000

–0.4

H2 Pressure ¥ 10

2000

–0.5 –0.6

1000 0 4090

4140

4190

Time, h (b)

4240

4290

Conductivity, mS/cm or potential, Vshe

5000

To 95 ppb H2 @1332h

H2 pressure, microns

7000

0.1 Conductivity, mS/cm or potential, Vshe

9000

6000

277

–0.7 4340

9.24 Hydrogen permeation vs. time and coolant H2 fugacity in unsensitized Type 304L stainless steel.

data at high potential shows the effect of both hardening and segregation, in that the growth rate is higher than for the unirradiated data that is either cold worked or sensitized. The Cr depletion profile from irradiation is a few nanometers (nm) wide, ~100¥ narrower than exists from thermal sensitization. But a wide range of Cr depletion widths were evaluated using complex heat

© Woodhead Publishing Limited, 2010

278

Understanding and mitigating ageing in nuclear power plants

treatments and it was shown that SCC depends on the minimum Cr level, not the width [59, 60] of the Cr profile near grain boundaries. Radiation creep relaxation is often beneficial because components under constant displacement loading experience load relaxation vs. time, so that if cracks do not nucleate early in life (nor are reloaded later in life), the probability of SCC can decrease with neutron fluence because the load drops below a third of its original value within a few displacements per atom (dpa). Above ~ 320 °C and above a few dpa, radiation swelling can occur. While the PWR core outlet temperature is only ~323 °C, gamma heating in stainless steel components such as baffle plates and baffle former bolts can produce an elevation in their internal temperature by about 30 °C. Because core baffle plates are typically fabricated from annealed stainless steel, while the baffle former bolts are typically fabricated from ~15% cold-worked stainless steel, and because swelling is delayed in cold-worked stainless steel, differential swelling can occur preferentially in the plates. Early in life (within ~5 dpa) the bolts undergo radiation-induced relaxation by > 80%, but subsequently differential swelling in the plates can reload the bolts and promote SCC – the balance between reloading and relaxation is complicated.

9.5.3 Stress intensity factor An understanding of the effects of the (crack tip) stress intensity factor (K) is key both to rationalize the historical development of cracks and to disposition their future growth. K is proportional to stress (s) times the square root of crack length (a) (K = B s÷a, where B is a factor that accounts for crack and component geometry). Quantifying the effects of K on SCC growth rates has had a long and troubled history. Simple transgranular (TG) fatigue precracking followed by static (esp. bolt) loading is a very poor approach for quantifying the effects of K, as well as most other parameters. Decreasing K in large steps is also likely to produce crack arrest in cases where it would not otherwise occur. Modern techniques [20, 21, 38, 70] employ transitioning to IG SCC, and re-transitioning at each K level or employing a –dK/da technique where the K decreases only as crack growth occurs. Using such techniques (Figs 9.6, 9.25 and 9.26), a consistent K dependency can be measured, with no evidence of crack arrest (KISCC), at least down to ~5 MPa÷m. A similar K dependency has been observed for irradiated stainless steel [66], and for nickel alloys in BWR water [7, 21, 65]. Effects of K can easily be mis-measured and misinterpreted. In constant displacement specimens there is a strong tendency toward low or no growth rates at low K (i.e., much lower than would be observed with good techniques), while reasonable growth rates are often measured at high K. The net effect is to increase the K dependency because the lower K crack growth rate data are biased low. Somewhat conversely, specimens with no side grooves tend to

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

279

711 mm SCH80 Pipe-ranganath analysis Buchalet/Bamford Method Circumferential crack 55

Upper limit

Mean

33

+ 30

Residual stress Ksi

22

142

11

Lower limit

0

40

Stress intensity, MPa√m

35

5

10

Uper limit

Lower limit

+8.1 1.423 Crack depth

Pressure stress 41 MPa deadweight stress   7 MPa Thermal stress 41 MPa

15 20 25 Crack depth, mm (a)

30

35

15

304SS, 2-sided weld, EPRo = 5 C/cm2 0.3, 0.2, 0.1 mS/cm, +0.175 Vshe Symmetrical sres profile 3 ¥ 1019 n/cm2-y

30 25 20 15

sres with +69 MPa above nominal (shroud.mj3)

0.3 mS/cm

10

0.2 0.1

5 dK/da for 0.3 mS/cm

10 5 0 0

dK/da, MPa√m per mm

Sress intensity, MPa√m

44

0 10

20 Crack depth, mm (b)

30

9.25 Residual stress through wall (a, inset diagram) and resultant stress intensity factor vs. through-wall crack depth (a) for a typical large diameter pipe weld, with all welding passes made from the outside diameter. (b) Stress intensity factor vs. crack depth for a BWR core shroud that is welded by alternate passes from the inside and the outside diameter. © Woodhead Publishing Limited, 2010

Understanding and mitigating ageing in nuclear power plants

58.1

Stress intensity factor

58 6020

11.88 11.83

To constant K @816h At 19.8 MPa√m @2146h

Crack length, mm

11.93

11.78 11.73

40 35 30

6025

25

6030

6035 6040 Test time, h (a)

6045

6050

20 6055

20

12.03 11.98

45

1.5 ¥ 10–7 mm/s

58.05

50

5.7 ¥ 10–8 mm/s

1168 11.63 2200

2700

Stress intensity factor 3.4 ¥ 10–8 mm/s

2.5 ¥ 10–8 mm/s

18 16 14 12 10 8 6

c209 – 0.5TCT 316L 20% CW 20 MPa√m, 2000 ppm O2, Pure water 3200

3700 Test time, h (b)

4200

4700

4

Stress intensity factor, MPa√m

58.15

1.0 ¥ 10–5 mm/s

End Varying-K at @6044h

58.2

To varying-K at @6029h 86.6 MPa√m per mm

58.25

55 3.5 ¥ 10–7 mm/s

Unload @3006h

Crack length, mm

58.3

To 27.5 MPa√m, R = 0.5, 0.001 Hz @5897h

58.4 58.35

60

c317 – 2TCT of 316L +20% RA at 140C 27.5 MPa√m, 2 ppm O2, 100 ppb SO4

Stress intensity factor, MPa√m

58.45

Re-initiated varying-K drop at 21.7 MPa√m per mm @2452h

280

2 0 5200

9.26 Effect of +dK/da and –dK/da at values relevant to plant components. Growth rates up to 1000¥ faster than constant K conditions have been observed under +dK/da conditions. By contrast, under representative –dK/da conditions, the crack growth rate is sustained and the observed rates are similar to those obtained at constant K.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

281

exhibit extensive crack branching above moderate K, and this decreases the local K at each branched crack tip. This biases the higher K data downward, and yields a lower K dependency. In general, the agglomeration of many sets of data that involve different heats and test techniques tends to dilute any trend, including K, giving a shallower K dependency. The issue of crack branching can be important in the analysis of plant components. As the K increases, extensive branching may occur, and this can lead to a significant deviation from a fixed power-law dependency to a plateau-like behavior (fixed growth rate vs. K). However, if there are reasons why the crack plane is constrained – e.g., the narrow plane where the weld residual strain is high, adjacent to the fusion line – then plateau behavior may not exist. Another important consideration is the change in K as the crack depth increases. In plant components, K changes almost completely because of crack advance since, e.g., changes in pressurization are small and represent a small fraction of the total stress. Unfortunately, most laboratory tests have been performed at constant K or constant load, and make no effort to simulate the actual profile in K vs. crack depth. Tests to evaluate +dK/da have shown elevated growth rates [71, 72] because there is a positive feedback mechanism that causes an increase in growth rate (da/dt), which in turn causes an increase in K vs. time (dK/dt = da/dt · dK/da). Growth rates up to 1000¥ higher than constant K data have been observed under realistic +dK/da conditions. Under –dK/da conditions, the possibility that cracks might arrest was an attractive prospect, but –dK/da data have shown sustained cracking in ~25 cases, some involving a more than twofold reduction in K. Stainless steels and nickel alloys show the same response under dK/da conditions.

9.6

Stress corrosion cracking (SCC) dependencies – miscellaneous

9.6.1 General corrosion rate The general corrosion rates of stainless steels and nickel alloys (i.e., Alloy 600 and Alloys 182/82 weld metals) are very similar (Fig. 9.27). Corrosion rate can be important for many reasons, including: ∑

the loss of a thin polished or compressive layer or a passivation film that was designed to mitigate SCC initiation; ∑ release of corrosion product into the water; ∑ altering the rate of consumption of oxidants as they diffuse into the crack; ∑ shifting the corrosion potential when the oxidant concentration is low (a higher corrosion rate will consume more O2 and cause the corrosion potential to drop, e.g., at 10 ppb vs. 1 ppb, Fig. 9.14); and

© Woodhead Publishing Limited, 2010

282

Understanding and mitigating ageing in nuclear power plants

1000

Electrode potential, mV(she)

800

304 SS polarization measurement in deaerated, high purity water (< 2 ppb O2) ast 288 °C

600

After preoxidation in 200 ppb O2 for 3 weeks

400

After preoxidation in 150 ppb H2 for 3 weeks

200 After preoxidation in 1 ppm H2O2 for 3 weeks

0

After preoxidation in deaerated water for 3 weeks

–200 –400 –600 0.01

1000

Electrode potential, mV(she)

800

0.1

1 10 Current density, mA/cm2 (a)

100

1000

Alloy 600 polarization measurement in deaerated, high purity water (< 2 ppb O2) at 288 °C and 150 cc/min

600 400 After preoxidation in deaerated water for 3 weeks

200 After preoxidation in 200 ppb O2 for 3 weeks

0

After preoxidation in 150 ppb H2 for 3 weeks

–200 –400 –600 0.01

0.1

1 10 Current density, mA/cm2 (b)

100

1000

9.27 Corrosion rate vs. prior exposure for stainless steel (a) and Alloy 600 (b) in 288 °C pure water. Exposure in a given dissolved gas chemistry was for several weeks, then a rapid change was made to deaerated water to make the polarization measurements.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels



283

enhancing the environmental effect, especially on surfaces, e.g., for corrosion fatigue.

While seemingly contradictory, the corrosion rate at ‘high corrosion potential’ is lower than at ‘low corrosion potential’ because the passive oxide films that form at low potential are not as protective. The data shown in Fig. 9.27 are short-term data obtained by exposing the material in a specific environment for ~2 weeks, then changing to Ar-deaerated water so that O2 and H2 do not interfere with electrochemical measurements of the metal corrosion reaction. Long-term corrosion tests show lower corrosion rates because the rate appears to decay parabolically with time. The similarity in general corrosion behavior means that the corrosion potential response vs. O2 and H2 is very similar among these alloys, as is the short crack and long crack chemistry development. The higher corrosion rate at low potential may explain why the corrosion fatigue crack initiation life of stainless steels and nickel alloys is reduced at low vs. high corrosion potential [48, 49].

9.6.2 Grain boundary silicon At sufficiently high concentrations in the grain boundary, Si can enhance SCC because it oxidizes to SiO2 at all relevant potentials and is soluble in high temperature water. The primary concern is for irradiated materials, where Si can segregate to levels above 20 at%, and can cause high crack growth rates, and no effect of corrosion potential or stress intensity factor on crack growth rate. This has been observed in stainless steels (Fig. 9.28) and nickel alloys with elevated bulk Si levels [73].

9.6.3 Effects of environment on fracture Environmental effects on fracture is an emerging area that can be subdivided into J-R tearing resistance and fracture toughness (e.g., KIC-Env) [74, 75]. Both stainless steels and nickel alloys can exhibit environmental effects on fracture in light water reactor environments, and very large reductions have been observed at lower temperatures in Alloy 82 weld metal (Fig. 9.29). Exposure to high temperature water allows H2 gas molecules to dissociate and permeate as hydrogen atoms throughout the material, although the effects of hydrogen are generally more pronounced after cooling to <130 °C. Pre-saturation of the metal with hydrogen by exposure to high temperature water can also create the opportunity for rapid failure, as distinct from slower tearing fracture, where (without pre-exposure) there is a rate-limiting step associated with hydrogen dissociation and permeation into the metal. To date, several laboratories have observed sudden fracture while performing SCC testing, with KIC values as low as ~75 MPa÷m (Fig. 9.30).

© Woodhead Publishing Limited, 2010

Understanding and mitigating ageing in nuclear power plants

16.1 15.7 Pt potential

Outlet conductivity

12.5

200

250

300

350 400 450 Test time, h (a)

500

550

600

c270 – 0.5TCT of A182 +3%Si + 26%RA 27.5 MPa√m, 2 ppm O2, Pure water

11.6

3.4 ¥ 10–7 mm/s

Pt potential

11.5

CT potential

11.4 700

750

–0.3 –0.4 –0.5

–0.6 650

0.5 0.4

0.1

800

850

0

To 2000 ppb O2 @997h

11.7

–0.2

0.2 Outlet conductivity

11.8

–0.1

0.3

To 95 ppb H2 @633h

Crack length, mm

11.9

To constant K @277h

12.1 12

0

2.2 ¥ 10–6 mm/s

4 ¥ 10–6 mm/s

12.1 150

0.1

900 950 1000 Test time, h (b)

–0.1 –0.2 –0.3 –0.4 –0.5 1050 1100 1150

Conductivity, mS/cm or potential, Vshe

12.9

To 6% H2 in Ar @312h

13.3

To constant K @226h

13.7

To 0.001 Hz + 9000s hold @167h

Crack length, mm

14.1

1.0 ¥ 10–6 mm/s

Conductivity, mS/cm or potential, Vshe

0.2

CT potential

14.9 14.5

0.3

Varying-K drop to 15 MPa√m @630h

15.3

0.4

c183 – 0.5T CT of 5Si-SS 15%CW 29.7 MPa√m, 2000 ppb O2, Pure water

Being Varying-K decrease @492h

284

–0.6 1200

9.28 Crack length vs. time for a 0.5TCT specimen of an unsensitized model ‘stainless steel’ containing 5% Si ‘cool’ worked to 22% reduction in area (a) and a wrought Alloy 182 with 3% Si with 26% RA (b).

9.6.4 Crack initiation This chapter has emphasized SCC growth, partly because it can be quantified more accurately than crack initiation, and partly because the study and use of initiation data poses a variety of conceptual and pragmatic concerns. These

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels T= 340

Air (A2a,b) 29 (150 cc/kg)

54 °C Water 54 °C Water

15

68

150 cc/kg

338 °C Water

Welded in 100%Ar

285

Alloy 82H (T-S) Weld A2a,b

245

327

Air (C4a) 2 (150 cc/kg)

54 °C Water

54 °C Water 50 13 93 °C Water

3 (150 cc/kg)

Welded in 100%Ar

Weld C4a

121 °C Water 150 24

454

Air (C4c) 54 °C Water

4 (150 cc/kg)

54 °C Water

50 36

54 °C Water

150 cc/kg

93 °C Water 150 23 149 °C Water

150 cc/kg

338 °C Water

150 cc/kg

144 Welded in 100%Ar

Weld C4c

230 310

26

Air (C2) 54 °C Water

4 (150 cc/kg)

54 °C Water

4 (50 cc/kg) 0

Welded in Ar-5%H2

500 JIC, kJ/m2

Weld C2 1000

9.29 J-R data of Brown and Mills on various Ni alloy 82H weld metal showing a large reduction in fracture resistance in water compared to air [76].

include challenges associated with the absence of continuous monitoring (even of failure); use of sufficient replicates and surface area so that the results are statistically significant; and uncertainty in defining the average, distribution or extremes of surface condition in plant components that is key to assuring the applicability and relevance of the data. Conceptually, there is no agreement on the meaning of crack initiation (e.g., a depth of 1 mm, of 50 mm, of 1 grain size, when conveniently detectable by ultrasonic inspection, when long crack growth rate response is achieved, etc.). Most examinations of cracked plant components show significant surface cold work, and indeed

© Woodhead Publishing Limited, 2010

286

Understanding and mitigating ageing in nuclear power plants

most surfaces show heavy damage when examined in detail (Fig. 9.31a), and how well understood or reproducible that the surfaces are has a strong bearing on the initiation behavior.

17.6

17 16.8 16.6 16.4

c192 – 0.5T CT of 304L 70%CW ~41 MPa√m, 2000 ppb O2, Pure water 1.6 ¥ 10–5 mm/s

Outlet conductivity

2 ¥ 10–6 mm/s

16.2 6715

0.25 End of Test @6749h

17.2

0.3

CT potential Start Varying-K at 43.3 MPa/m per mm @6687h

Crack length, mm

17.4

0.35

Actual KIC = 75.5 MPa√m

Pt potential

6720

6725

6730 6735 Test time, h (a)

6740

0.2 0.15 0.1 0.05

Conductivity, mS/cm or potential, Vshe

17.8

0 6750

6745

18.3

0.4 Pt potential CT potential

14.3

K rising because test is controlled by specimen c275

13.3 1.7 ¥ 10 mm/s

12.3 3900

This tandem specimen grew rapidly over 2 week holiday and failed @5914h

15.3

0.2

c276 – 0.5TCT A182 Weld Metal, 288 °C 27.5 MPa√m, 2 ppm O2, Pure water

–7

To Constant K @4515h

16.3

To 30 ppb SO4 @2961h To R = 0.8, 0.001 Hz + 85,400s hold @3900h

Crack length, mm

17.3

2.4 ¥ 10–7 mm/s

0

–0.2

–0.4 KIC ~ 90 MPa√m

Conductivity, mS/cm or potential, Vshe

Outlet conductivity

–0.6 4400

4900 Test time, h (b)

5400

5900

9.30 Examples of sudden failure of (a) cold worked stainless steel and (b) Alloy 182 weld metal tested in 288 °C water at increasing K until failure occurred. The load and crack depth at failure are very well defined, and the resulting KIC (not necessarily obtained valid conditions) is relatively low [75]. The fracture surfaces are illustrated in (c) and (d) and show the demarcation between scc and fast fracture. © Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

287

(c)

(d)

9.30 Continued

Crack initiation can be tracked to relatively small dimensions, both on smooth specimens and specimens with defects such as a CT specimen with a controlled radius, blunt notch. The transition from short crack growth to long crack growth (Fig. 9.31b) generally occurs in the 20–50 mm range of

© Woodhead Publishing Limited, 2010

288

Understanding and mitigating ageing in nuclear power plants

40.0 Transverse cross-section

Plastic strain (%)

30.0

20.0

10.0

0.0 0

Crack length from notch, microns

400

50

100 150 Distance from surface (microns) (a)

350

Sensitized 304 SS 25 mm CT C37 (1y) Kmax = 33 MPa√m, R = 0.7, 0.01 Hz 200 ppb 02, 0.5 mS/cm HCI

300

Cracking from blunt notch, wet machined with a 220 grit wheel

200

250

Blunt notch CT specimen 250 200

200 ppb O2 0.5 mS/cm HCl

150 100

6% H2 in Ar 0.5 mS/cm HCl

50

50 0 80

0 130 100

120

140

160

180 200 Time, h (b)

220

180 240

260

280

9.31 (a) Example of surface cold work present on many surfaces when section obliquely polished and evaluated by EBSD. (b) Transition from short crack to long crack growth in a carefully wet ground blunt notch CT specimen of sensitized stainless steel.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

289

crack depth, with the growth rate being slower when the crack is shorter – apparently related to nucleation of isolated cracks and their coalescence to a single crack. Repeated interruption of slow strain rate specimens revealed a strong effect of surface preparation, temperature (Fig. 9.32a) and impurities (Fig. 9.32b). Many crack initiation tests use very aggressive loading, temperature, surface preparation and/or water chemistry conditions, sometimes so aggressive that

7

Pure water

5 4 55 mS/cm H2SO4 All

3

A40 Na2CO3

A20 H2SO4

A25,A22 HCl+H2SO4 3.3+0.53 ¥ 10–7/S A26, H2SO4

1

4

5

6 pH288°C (a)

7

Weld + 400 °C/10d 40 Weld + 500 °C/24h 600 °C/24h 600 °C/24h + shot peen Open symbols-% strain Closed symbols-% life

40

Initiation, % strain

NaCl, A39 NaHSO4 A27

Neutral

2

A30 Na2SO4

30

8

100

80

60 20 40 10

0 100

20

150

200 Temperature, °C (b)

250

0 300

9.32 Strain to crack initiation for sensitized stainless steel tested by slow strain rate at 3.3 ¥ 10–7 s–1 in 288 °C water and repeatedly interrupted to detect initiation.

© Woodhead Publishing Limited, 2010

Initiation, % of life

Crack initiation %e

6

0

A37 NaOH

10 mS/cm Impurity 304SS Weld + 500 °C/24h Cert 3.3 ¥ 10–7/S 200 ppb Oxygen A13

290

Understanding and mitigating ageing in nuclear power plants

they call into question the relevance of the data. For example, in BWR tests it is common to use graphite wool as a crevice agent, but the wool releases large amounts of impurities. U-bend and crevice bent beam specimens usually are interrupted for SCC evaluation, which makes the time to initiation ambiguous. Test designs that allow for controlled, active load and continuous monitoring are more attractive, such as the ‘Keno’ test which can test up to 150 specimens in one autoclave [76]. Failure of any specimen is revealed by a numbered indicator ball. Good statistical confidence can be obtained (Fig. 9.33). The effect of many variables on crack initiation and growth are similar, e.g., temperature, corrosion potential, water purity, irradiation, etc. This suggests that a similar mechanism may be responsible for the early stages of the formation of a mechanically distinct geometry that will tend to grow in preference to its surroundings and the later stages when a long crack exists. The sole reliance on a ‘paper thin’ layer whose characteristics and SCC behavior is difficult to quantify with confidence and relevance is not wise. While every effort should be made to control component design and surface characteristics to delay initiation, the concept of inherent resistance in the bulk material to crack advance is a safer focus.

324 MPa 324 MPa 47 Ksi 47 Ksi creviced

99

%

95 90 80 70 60 50 40 30

193 MPa 28 Ksi creviced

193 MPa 28 Ksi

Stress, 20X

20 Stress w/ crevice, 10X

10

Crevice @193 MPa, 6X

Crevice @324 MPa, 3X

5 3 2

Sensitized 304 SS 288 °C, 80 wppm O2, 0.87 mS/cm H2SO4

1

10

100 Time to failure, h

1000

9.33 Effect of stress and crevicing on the crack ‘initiation’ of sensitized stainless steel in actively loaded smooth specimens whose failure is detected on-line [77].

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

9.7

291

Mechanism of stress corrosion cracking (SCC)

Despite the many common characteristics across environment, temperature and material (e.g., Figs 9.1–9.6), SCC tends to be viewed narrowly in terms of specific observations, with the inherent implication that different mechanisms and dependencies are involved. However, the crack tip is deaerated and at low potential in LWR environments, and a well-behaved continuum exists for changes in corrosion potential, water purity, temperature, stress intensity factor, material and condition, cold work/yield strength, etc. (Figs 9.1–9.6). Thus, it is reasonable to propose that a common crack advance mechanism applies to all of these materials, with some special factors that must be accounted for in specific materials or environments.

9.7.1 Underlying crack advance mechanism and primary sub-processes When the protective oxide film is removed, engineering alloys are highly reactive in aqueous environments, not unlike sodium in water (whose oxide/ hydroxide, NaOH, is highly soluble and therefore non-protective). Thus, most models of SCC advance (or, more generically, environmentally assisted cracking, EAC, which encompasses the spectrum of corrosion fatigue ´ slow strain rate ´ constant load loading) in hot water involve strain induced disruption of the protective oxide film at the crack tip. Following such disruption, rapid metal oxidation and subsequent repassivation proceeds, a process that can be quantitatively linked to crack advance (Fig. 9.34) [7–9]. Even in hydrogen embrittlement or film cleavage models of crack advance, oxide rupture plays an essential role, e.g., in enhancing the kinetics of (atomic) H0 formation and avoiding the restricted kinetics of H0 transport through oxide films. Hydrogen embrittlement is sometimes invoked as a primary mechanism of crack advance in high temperature water, but there is a significant array of contrary evidence [13–21]. Hydrogen in the metal has been shown to be proportional to the H2 fugacity in the coolant, and hydrogen permeation into and back out of closed (permeation) tubes is also controlled by the square root of the coolant H2 fugacity. All iron and nickel-base structural materials show a similar effect of corrosion potential (e.g., Figs 9.1–9.5), and their growth rates rise dramatically as the oxidant level rises (usually with the H2 level decreasing), whether the crack tip pH moves acidic or basic. Efforts to achieve even small amounts of crack growth under static load in 300 °C gaseous H2 requires high pressure, and generally only achieves a modest decrease in fracture toughness (vs. growth at a constant and reasonable K). For temperatures below 150 °C, the role of hydrogen embrittlement can become much more pronounced, and crack growth rates in water and gaseous

© Woodhead Publishing Limited, 2010

292

Understanding and mitigating ageing in nuclear power plants e◊

VS J H2O

VT

Ê ˆ Penetration distance Á = M · Qf ˜ Ë nrF ¯

Oxidation charge density, Qf

J M+ dc

Qf

Average crack tip penetration

Qf

Q Q VT = M · f = M · f · e nrF t f nrF e f Crack side penetration VS = M · ip nrF Oxide nucleation and growth

Bare-surface dissolution

Time

tf Oxide rupture e tf = f e

9.34 Schematic of the corrosion (oxidation charge density) vs. time for metal that is unstrained (lower curves, ‘crack side penetration’) or strained to produce damage to the protective oxide (upper curves), with subsequent repassivation. While shown for dissolution the model equally applies to chemical oxidation in which the anodes and cathodes are separated.

H2 can be similar. Of course, hydrogen might have a role in enhancing dislocation mobility at 300 °C, but dislocations are more mobile at 300 °C, and hydrogen is ubiquitous and permeates in atomic form readily in metals exposed to high temperature water. In high temperature water environments, the slip–film rupture–oxidation (S/FR/O) model of crack advance (Fig. 9.34) is the most widely accepted model, and it has been extensively developed by Ford and Andresen for light water reactor environments [7–9]. It ascribes fundamental importance to dynamic strain that occurs at the crack tip (not, e.g., to stress intensity factor) and to the kinetics of film repair, which are related to local material chemistry and crack tip solution chemistry. The nature of passivity in hot water is different than at lower temperatures, and the oxides are relatively thick (0.1–1 mm), even on highly corrosion resistant materials such as Alloy

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

293

690 (30% Cr) and Alloy 22 (22% Cr, 13% Mo). Thus, while the crack tip solution chemistry and local (e.g., grain boundary) material composition are very important and predictable, no theoretical basis has been developed to predict the resulting repassivation response. Thus, investigators [7–9, 77] have relied on measurements of repassivation response. Since the repassivation response generally follows a power law decay of fixed slope ‘n’, the S/FR/O model can be reduced to a very simple formulation that derives from the periodicity of strain (oxide ‘rupture’) and the repassivation response: –n



it = i0 ÈÍ t ˘˙ Ît 0 ˚



i0 t 0n Vt = M e n z r F (1 – n ) e fn ct



V t = A (ect )n

Methods of direct, in-situ measurements of crack tip strain rate have not been devised, especially since the deformation behavior is highly inhomogeneous, in terms of specific slip planes within one grain, specific grain and grain boundary response, and vs. time. Thus, correlations to engineering/macroscopic stressors, such as stress intensity factor (K), stress intensity factor amplitude (DK), frequency, and applied strain rate (for slow strain rate tests) have been developed. Continuum mechanics formulations, such as proposed by Gao and Hwang [78], have also been used (e.g., by Shoji [77] and Young et al. [79]), but they rely on many assumptions and approximations, and have not been carefully validated against SCC observations, e.g., as a function of K, yield strength, rising/falling K profiles, growth rate (i.e., varying water chemistry), etc. While relying on measurements of dissolution currents during repassivation to provide conceptual and quantitative data in many EAC systems, the S/ FR/O model does not conceptually require separated anodic dissolution and cathodic reactions, since the process of film rupture and reformation is equally applicable to gaseous environments such as steam. Indeed, the continuum in EAC from high temperature water (e.g., 200–360 °C) to steam (e.g., >400 °C) can be conceptually explained by the shift from a dissolution dominated crack advance process at lower temperatures (where there is a ‘catalytic’ benefit associated with separated anodic and cathodic reactions) to oxidation dominated crack advance at higher temperatures (where direct chemical oxidation can occur readily). The sub-mechanisms associated with the effects of water chemistry – including dissolved O2 and H2, corrosion potential, aqueous impurities, etc. – were discussed in Sections 9.3.3 and 9.4.3. Anodic and cathodic

© Woodhead Publishing Limited, 2010

294

Understanding and mitigating ageing in nuclear power plants

reactions near the crack mouth create the potential gradient and associated crack chemistry. Since these reactions are remote and uncoupled from the crack tip reactions associated with crack advance, this sub-mechanism can be isolated and solved independently. The sub-mechanisms related to grain boundary compositional inhomogeneities, e.g., from sensitization or radiation segregation, can similarly be isolated and addressed independently of the overall crack advance mechanism. The sub-mechanism related to repassivation is addressed by measurement – no useful models exist to describe film formation, the extent of corrosion during repassivation, etc., as a function of crack tip material and crack tip water chemistry. Simulating the crack tip conditions and measuring the repassivation kinetics is not trivial, but has been done. The sub-mechanism related to crack tip deformation kinetics is the most difficult, as measurements of deformation rate at growing crack tips are very difficult, and continuum mechanics formulations and finite element models are inadequate for providing a robust description of the dislocation kinetics at a crack tip. However, they do provide guidance regarding the effect of temperature, stress, stress intensity factor, stress intensity factor amplitude, etc. For example, under inert conditions, fatigue crack growth occurs by reversed slip, and the amount of reversed slip per increment in crack advance can be accurately estimated. Reversed slip can readily be associated with a crack tip strain rate, which permits good estimation of the crack tip deformation under cyclic conditions. Similarly, under slow strain rate test conditions, there is a sound basis for estimating a crack tip strain from the applied strain rate and the number of cracks that form. The biggest challenge is for constant K conditions, where sustained crack tip deformation is maintained by virtue of sustained crack advance, which requires that the strain field at the crack tip be redistributed. Nonetheless, there are empirical formulations that have been developed and provide reasonably accurate SCC prediction [7–9] of the spectrum of loading conditions (cyclic, slow strain rate, constant) and the inter-related effects of loading and material and water chemistry effects (Fig. 9.35).

9.8

Stress corrosion cracking (SCC) mitigation

Some of the mitigation approaches should be clear from the preceding sections; mitigation can be viewed in terms of moving or shrinking the size of one or more of the circles in Fig. 9.1. Factors that minimize SCC initiation include low-stress designs, absence of sharp corners or crevices, control of surface cold work during fabrication and welding, and improved surface characteristics. Many of these techniques are only applicable to new components, but processes have been developed that reduce one of more of the stressors, namely stress, cold work, surface roughness and surface

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels 10–3

Crack propagation rate, cm/s

10–4 10–5

295

Theoretical and observed crack propagation rate vs. strain rate sensitized 304 stainless steel epr 15 C/cm2 288 °C Constant load --- SSRT --- Corrosion fatigue

10–6 10–7

Theory 8 ppm O2 0.5 mS/cm

10–8

Theory (SSC) deaerated 0.1 mS/cm

10–9 10–10 10–10

Theory (fatigue) deaerated 0.1 mS/cm

10–9

10–8

10–7 10–6 10–5 10–4 Crack tip strain rate, 1/s

10–3

10–2

10–1

9.35 SCC growth rate of sensitized stainless steel vs. crack tip strain rate spanning constant K, slow strain rate and corrosion fatigue loading. Because the curves diverge, the change in growth rate (factor of improvement) associated with the indicated water chemistry change varies with loading condition [7].

degradation. While surface stresses can be redistributed using various types of shot, laser and water-jet peening processes, decreasing the surface cold work, surface roughness and surface degradation requires removal of some of the surface. Perhaps the most promising technique addresses all of these factors [80]. It is more common to consider techniques that mitigate SCC growth rates, because it is essentially impossible to prove that components that have been in service for years or decades have not already developed small cracks, or to ensure that new components will not crack. Because extensive cracking of stainless steel was observed first in BWRs, the most fully developed techniques for SCC mitigation apply to that system. Billions of US dollars were spent replacing piping systems with low-carbon and/or elevated nitrogen stainless steels that were resistant to weld sensitization. But replacement of pressure vessel internals is far more costly, and the most efficient mitigation techniques involve optimized water chemistry, since they provide systemwide benefits. Hydrogen water chemistry (HWC) [81] was the first water chemistry mitigation technique developed. It involved injecting H2 into the feedwater, and was effective because the annulus of most BWRs provided an optimal range of gamma radiation for recombination of oxidants and H2. This significantly

© Woodhead Publishing Limited, 2010

296

Understanding and mitigating ageing in nuclear power plants

reduced the corrosion potential in the external piping and in the lower plenum (below the core) – its effectiveness in the core region, where there is a high neutron flux, is less clear. Adequate mitigation generally requires a ‘moderate’ H2 injection rate that makes the bulk water reduce enough to promote reduction of NO3 to volatile forms (NO2, NO and NH3). Because there is always some transmutation of O16 to N16, moderate HWC causes a 5–7-fold increase in the radiation levels in the steam lines and at the turbine. Since the half-life of N16 is only 7.2 seconds, it quickly decays if it remains soluble as NO3 or NO2. Because H2 partitions to the steam phase, the rate and cost of the continuously injected H2 is relatively high. These drawbacks led to the development of electrocatalytic surface technologies, most notably NobleChem™ and OnLine NobleChem™. By introducing very low levels of platinum ion (usually in the form of Na2Pt(OH)6) into the reactor water – either at ~130 °C for NobleChem™ or during fullpower operation for OnLine NobleChem™) – all wetted surfaces become sufficiently catalytic that oxidants (e.g., O2 and H2O2) and H2 (a reducant) will fully react on the surface. This creates a very low corrosion potential even when the bulk water retains significant levels of oxidant – provided there is a stoichiometric excess of H2 vs. oxidants (Fig. 9.36). Since water is the product, only a small amount of H2 has to be injected to achieve the 2:1 molar ratio associated with its reaction with O2 (2H2 + O2 Æ 2H2O), and essentially no increase in radiation level occurs. Other techniques that have been considered or applied include stress mitigation (e.g., with last-pass heat sink welding, or mechanical stress improvement process) and full annealing of welds to eliminate weld residual stresses and strains. An important factor in evaluating mitigation methods is recognizing that the factor of improvement is dependent on the conditions of test or component. Figure 9.36b shows that the benefit associated with water chemistry changes depends on loading and corrosion potential, and Fig. 9.35 shows that the divergence of the lines translates to a different benefit depending on the loading conditions of the specimen or component.

9.9

Prediction of stress corrosion cracking (SCC) and irradiation assisted stress corrosion cracking (IASCC)

There are a number of narrow applicability, empirical models of crack advance, e.g., for SCC of stainless steels in BWR water, SCC of irradiated SS in BWRs, SCC of Alloy 600 in PWR primary water, SCC of Alloy 182 weld metal in PWR primary water, etc. However, the limits of such empirical approaches were recognized 30 years ago [7–9], because with 20–30 primary variables, with the effect of most variables being interdependent on the others, the matrix of experiments required to identify the dependencies exceeds

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels 400

nmca-18

300 1.0 ppm O2, 288 °C, 150 cc/min.

200

nmca-18 (ust) nmca-19 nmca-19 (ust)

100 ECP, mV (she)

297

nmca-18: 60 ppb Pt+20 ppb Rh, 200 ppb O2, 48 hours, 50 °C nmca-19: 60 ppb Pt+20 ppb Rh, 200 ppb O2, 48 hours, 88 °C

0 –100 –200 –300 –400 –500 –600

0

1

2 3 4 Molar ratio of H2/O2 in water (a)

5

6

O2

0 Bulk water with O2 and H2 Metal with catalyzed surface

2H2 + O2 Æ H2O on surface Stagnant water layer (b)

9.36 (a) Corrosion potential of an electrocatalytic surface vs. H2 to O2 molar ratio. The corrosion potential drops significantly once stoichiometric excess H2 conditions are achieved at the surface. (b) Schematic of an electrocatalytic surface showing rapid reaction of O2 and H2 at the surface, mass transport through the stagnant boundary layer, and the bulk levels of O2 and H2. If H2 is in excess, all of the O2 arriving at the surface is reacted, and the corrosion potential drops significantly.

1014. So the conceptual model described in Section 9.7 was created with engineering inputs, such as stress intensity factor, neutron fluence, corrosion potential (on the external surface), bulk water chemistry, grain boundary Cr depletion, etc. While many common measurements are imperfect – for example, solution conductivity does not account for specific anion effects – better inputs can be used if available.

© Woodhead Publishing Limited, 2010

298

Understanding and mitigating ageing in nuclear power plants

The model evaluated critical processes to properly model the individual effects – such as corrosion potential and water purity, which combine to affect the crack tip chemistry – and to understand the repassivation response in deaerated water of varying pH and anion concentrations. The overall concept is to develop, evaluate and quantify a model based on the crack tip system, then compare it to controlled laboratory data and then to well-defined plant data. The model provided excellent agreement with a diversity of laboratory data, e.g., for the effects of corrosion potential, water purity, stress intensity factor, cold work, sensitization, fluence, etc., as can be seen by the predicted lines in Figs 9.4 – 9.8 and 9.21. Insofar as well-documented field cracking data existed, reasonably accurate predictions can be made for the time when cracks appear (e.g., in recirculation piping, Fig. 9.37) as a function of distinguishing characteristics of plant operation – in this case, the impurity concentrations that are reflected in the average plant coolant conductivity. Difference in BWR water purity dominated the behavior of other components, including the incidence of cracking in shroud head bolts (Fig. 9.38). The predicted response based on average conductivity matched the plant response in most cases, but there were three outliers, where the incidence of cracking was high at low average conductivity. Investigation revealed that these plants were unique in having very high coolant conductivity early in life, followed by low conductivity subsequently. The early exposure caused more extensive cracking, and these cracks were then at higher K where growth could occur more readily after the plant conductivity improved. The predicted curve in these cases was based on year-by-year conductivity data rather than an overall lifetime average. Cracking in core internal structures, such as the core shroud, were accurately predicted years in advance of their detection in-plant (Fig. 9.39). Cracking in some cases was associated with significant irradiation damage, and in other cases cracking occurred in areas of low neutron fluence. An important factor in welded, bolted or other constant displacement loaded structures is that radiation creep causes stress relaxation at the same time as it causes radiation segregation and hardening. Although no residual stress measurements vs. neutron fluence have been made to verify the quantitative role of radiation creep, there is a strong basis for its effect. SCC in highly irradiated structures, like stainless steel control rods, has also been predicted [7–9, 24, 25, 82], although a range of responses is expected since swelling of the neutron-absorbing B4C pellets plays a large role. Studies on irradiated stainless steel from BWR control rods in the laboratory were performed by Jacobs as described in references [24, 25]. Figure 9.40 shows the predicted vs. observed response for these tests in one range of stress. Other examples of predicted response in sensitized stainless

© Woodhead Publishing Limited, 2010

Fraction of wall thickness penetrated

Stress corrosion cracking of austenitic stainless steels

299

28≤ dia. schedule 80 304 piping theoretical vs observed igscc penetration Theoretical Residual predication stress Mean Upper limit

0.5 0.4 0.3

Browns Ferry–1 0.326 ms/cm

Hatch–2 0.401 mS/cm

N.d.t. resolution limit

0.2

Big Rock Point 0.15 ms/cm

vermont Yankee 0.216 ms/cm

0.1

20

40

60

80

100 120 140 160 On–line, months (a)

180

200 240

260

10≤–28≤ dia. sch 80 304 piping in BWR environment Observed data with, in parentheses, the no. of cracks quoted Theoretical relationships based on “pledge” code, assuming Ecorr = OmVshe

Conductivity (Geo. mean), mS/cm

0.5

0.4

(42) (2)

0.3

0.2

(1) Curve # sresid. 1 Mean 2 Upper

0.1





Degree of sensitizN 15C/cm2 15C/cm2

Upper 20C/cm2

3

20

(19)

(50)

(6)

(6) (3) (19) (19) (2) (2)

1 2

10

Mean condition

(50)

3

Worst condition

40 60 80 100 120 140 On-line time (months) for wall penetration, a/t = 0.25 (b)

160

9.37 Predicted and observed response of sensitized stainless steel piping as a function of plant water purity (solution conductivity).

steels, unsensitized stainless steel, irradiated stainless steel and other materials are discussed in References [7–9, 24, 25, 82].

9.10

Future trends

Advanced LWR designs – such as Advanced BWR (ABWR), Economic Simplified BWR (ESBWR), AP1000 (Advanced PWR), European Passive

© Woodhead Publishing Limited, 2010

300

Understanding and mitigating ageing in nuclear power plants

% components with IGSCC divided by on-line months

1

Creviced Alloy 600 Shroud Head Bolts

0.9

UT inspections at BWR plants with detection at 10% of wall

0.8

BWR's with high mS/cm excursions not reflected in plant average

0.7 0.6 0.5

Predicted from stress distribution Eq. EPR 13 C/cm2 fc 150 mVshe Uniform Stress a0 50 mm

0.4 0.3 0.2

Prediction for high conductivity Early in Plant Life

Prediction for Avg. Conductivity

0.1

0.5

0.1 0 0

0.2 0.3 0.4 Avg. plant conductivity, mS/cm

0.6

9.38 Predicted and observed effect of average plant water purity (solution conductivity) on incidence of SCC in Alloy 600 shroud head bolts in BWRs [7]. The three outlier plants had high coolant conductivity early in life, then good water purity (low conductivity) subsequently, so that the average conductivity did not accurately characterize their response. When predictions were made on a yearby-year basis, the agreement was good.

Reactor (EPR), and others – incorporate simplifications, improved materials, better inspectability, etc. Nonetheless, in April 2008 there were 439 nuclear power plants operating in 31 countries (104 in the United States, 59 in France, 55 in Japan, 31 in the Russian Federation and 20 in Korea, among others), and the need to maintain their safe and reliable operation must remain a priority. The development of improved materials, fabrication and joining techniques, control of microstructure and surface conditions, etc., also remains a priority. High nickel alloys such as Alloy 690 are unattractive in the core because nickel activates and complicates inspection and servicing during outages. Welding creates residual stresses and strains (and possible weld defects), and these have been the primary causative factor in SCC in LWRs. The importance of microstructure is highlighted in observations of high growth rates in cold-worked Alloy 690 [83–85]. Despite efforts to carefully fabricate components, most surfaces have significant surface damage, which can include both surface cold work and compositional changes from heat treatment. The desire to achieve component reliability over 40 or 60+ years of plant operation will require sustained, diligent efforts.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels Core Shroud Analysis #J30/31e

25

32 mm thick 304SS, 2-sided weld 0.15 mS/cm, EPR0 = 0 C/cm2 Symmetrical sres profile Stepped thru-wall flux 5 ¥ 1019 n/cm2-y at ID +0.20 Vshe

20

Crack depth, mm

301

sres with + 103 MPa above nominal

15 sres with +69 MPa above nominal

10

Indication #4: prior UT current average current maximum

5

0

0

Boat sample

100

200 Time, months

300

400

9.39 Predicted and observed SCC in stainless steel core shrouds in BWRs [7, 24, 25].

Fluence, n/cm2 (E > 1 MeV)

1022

3

25 ¥ 1021

Theoretical V Observed Relationships, Double Ligament Specimens 304 Stst, 32 ppm O2 Water, 288 °C

5

18 30 43 48 44 46 42 40 61 62 60 7

59

50-70 KSI

(345 – 483 MPa)

30-50 KSI

(207 – 345 MPa)

71

69

70

1020

(483) 70 1019 101

102

Theory

103

(207) 30

104

Applied tensile stress KSI (MPa)

105

Time to failure, h

9.40 Predicted and observed SCC response of irradiated stainless steel tested under constant load conditions in oxygen-saturated (~40 ppm) water at 288 °C.

© Woodhead Publishing Limited, 2010

106

302

9.11

Understanding and mitigating ageing in nuclear power plants

Sources of further information and advice

There are a number of publications and proceedings focused on SCC of stainless steels and other structural materials (i.e., nickel alloys, weld metals and pressure vessel steels) that are excellent introductory sources of information [1–5], which include publications that cover a broader spectrum of environmental cracking phenemona. The IAEA, NRC and EPRI are organizations that are sources of extensive information, and the latter two organizations have undertaken major efforts in proactive materials degradation management. The US Nuclear Regulator Commission website (www.nrc. gov) has extensive links to plant incidents and laboratory reports. Important conferences include the Env Deg and Fontainvraad conferences, along with nuclear sessions such as the annual NACE Corrosion conference. An annual gathering of technical experts involved in environmental cracking in LWRs – the International Cooperative Group on EAC – meets every year (contact the author). Some company, national laboratory and contractor reports are available from their websites (e.g., www.epri.com) or by direct contact, e.g., with Areva, GE, Westinghouse/Toshiba and Hitachi. Journals in which SCC data in high temperature water environments are published include Corrosion, J. of Nuclear Materials, Corrosion Science and others.

9.12

References

1. Proc. 1st–13th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE/ANS/TMS, 1983–2007. 2. Proc. 1st–6th Int. Symp. Fontevraud, Contribution of Materials Investigation to the Resolution of Problems Encountered in Pressurized Water Reactors, 1986–2006. 3. Proc., Chemistry and Electrochemistry of Corrosion and SCC: A Symposium Honoring the Contributions of R.W. Staehle, Ed. by R.H. Jones, TMS, Feb. 2001. 4. Proc. Parkins Symp. on Fundamental Aspects of Stress Corrosion Cracking, ed. by S.M. Bruemmer et al., AIME, 1992. 5. Stress Corrosion Cracking and Hydrogen Embrittlement of Iron-Base Alloys, Firminy, France, June 1973, Ed. by R.W. Staehle, J. Hochmann, R.D. McCright and J.E. Slatern, NACE, Houston, TX, 1977. 6. Fundamental Aspects of Stress Corrosion Cracking, Ed. by R.W. Staehle, A.J. Forty, and D. VanRooyen, Ohio State Univ., NACE, Houston, TX 1967. 7. F.P. Ford and P.L. Andresen, ‘Corrosion in Nuclear Systems: Environmentally Assisted Cracking in Light Water Reactors’, in Corrosion Mechanisms, Ed. by P. Marcus and J. Ouder, Marcel Dekker, pp. 501–546, 1994. 8. F.P. Ford and P.L. Andresen, Proc. Third International Symposium on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, Ed. by G.J. Theus and J.R. Weeks, The Metallurgical Society of AIME, 1988, p. 789. 9. P.L. Andresen and F.P. Ford, Mat. Sci. Eng., Vol. A1103, 1988, p. 167. 10. P.L. Andresen and G.S. Was, ‘SCC of Unirradiated Stainless Steels and Nickel Alloys in Hot Water’, 17th International Corrosion Congress, Las Vegas, NACE, Houston, TX, 2008.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

303

11. P.L. Andresen, ‘Perspective and Direction of Stress Corrosion Cracking in Hot Water’, Proc. Tenth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE, 2001. 12. P.L. Andresen, T.M. Angeliu and L.M. Young, ‘Immunity, Thresholds, and Other SCC Fiction’, Proc. Staehle Symp. on Chemistry and Electrochemistry of Corrosion and SCC, TMS, Feb. 2001. 13. P.L. Andresen, T.M. Angeliu, L.M. Young, W.R. Catlin and R.M. Horn, ‘Mechanisms and Kinetics of SCC in Stainless Steels’, Proc. Tenth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE, 2001. 14. P.L. Andresen, L.M. Young, W.R. Catlin and R.M. Horn, ‘Stress Corrosion Crack Growth Rate Behavior of Various Grades of Cold Worked Stainless Steel in High Temperature Water’, Corrosion/02, Paper 02511, NACE, 2002. 15. P.L. Andresen, P.E. Emigh and L.M. Young, ‘Mechanistic and Kinetic Role of Yield Strength/Cold Work/Martensite, H2, Temperature, and Composition on SCC of Stainless Steels’, Proc. Int. Symp. on Mechanisms of Material Degradation in Non-Destructive Evaluation in Light Water Reactors, Osaka, Japan, May 2002, published by Inst. of Nuclear Safety System, Japan, 2002. 16. P.L. Andresen, P.W. Emigh, M.M. Morra and R.M. Horn, ‘Effects of Yield Strength, Corrosion Potential, Stress Intensity Factor, Silicon and Grain Boundary Character on the SCC of Stainless Steels’, Proc. 11th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, ANS, 2003. 17. P.L. Andresen, P.W. Emigh and L.M. Young, ‘Mechanistic and Kinetic Role of Yield Strength/Cold Work/Martensite, H2, Temperature, and Composition on SCC of Stainless Steels’, Invited overview, Proc. of 10th Anniversary Symposium of Institute for Nuclear Systems Safety, Osaka, Japan, May 2002. 18. P.L. Andresen, ‘Factors Influencing SCC and IASCC of Stainless Steels in High Temperature Water’, PVP, Vol. 479, ASME, 2004. 19. P.L. Andresen and M.M. Morra, ‘IGSCC of Non-sensitized Stainless Steels in High Temperature Water’, J. of Nuclear Materials, Vol. 383, Issues 1–2, December 2008, pp. 97–111. 20. P.L. Andresen, ‘Perspective and Direction of Stress Corrosion Cracking in Hot Water’, Proc. Tenth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE, 2001. 21. P.L. Andresen and M.M. Morra, ‘SCC of Stainless Steels and Ni Alloys in High Temperature Water’, Corrosion, Vol. 64, 2008, pp. 15–29. 22. P.L. Andresen, ‘Critical Processes to Model in Predicting SCC Response in Hot Water’, Paper 05470, Corrosion/05, NACE, Houston, TX, 2005. 23. P.L. Andresen and L.M. Young, ‘Characterization of the Roles of Electrochemistry, Convection and Crack Chemistry in Stress Corrosion Cracking’, Proc. Seventh International Symposium on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE, 1995, pp. 579–596. 24. P.L. Andresen, F.P. Ford, S.M. Murphy and J.M. Perks, Proc. Fourth International Symposium on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, Ed. by D. Cubicciotti and G.J. Theus, NACE, 1990, pp. 1–83. 25. P.L. Andresen, ‘Irradiation Assisted Stress Corrosion Cracking’, in Stress Corrosion Cracking: Materials Performance and Evaluation, Ed. by R.H. Jones, ASM, Materials Park, OH, 1992, pp. 181–210. 26. S.M. Bruemmer, E.P. Simonen, P.M. Scott, P.L. Andresen, G.S. Was and J.L.

© Woodhead Publishing Limited, 2010

304

Understanding and mitigating ageing in nuclear power plants

Nelson, ‘Radiation Induced Material Changes and Susceptibility to Intergranular Failure of Light Water Reactor Core Internals’, J. Nucl. Mater., Vol. 274, 1999, pp 299–314. 27. G.S. Was and P.L. Andresen, ‘SCC Behavior of Alloys in Aggressive Nuclear Reactor Core Environments’, Corrosion, Vol. 63, No. 1, 2007, pp. 19–45. 28. G.S. Was and P.L. Andresen, ‘The Nature of SCC in Irradiated Stainless Steels and Nickel-base Alloys in LWR Environments’, 17th Int. Corrosion Congress, Las Vegas, NACE, Houston, TX, 2008. 29. ASME Boiler and Pressure Vessel Code, Sections III and XI, ASME, New York. 30. H. Hanninen and I. Aho-Mantila, ‘Environment-Sensitive Cracking of Reactor Internals’, Proc. Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, Traverse City, AIME, 1987, pp. 77–92. 31. R.L. Cowan and G.M Gordon, ‘Intergranular Stress Corrosion Cracking and Grain Boundary Composition of Fe-Ni-Cr Alloys’, Stress Corrosion Cracking and Hydrogen Embrittlement of Iron-Base Alloys, Firminy, France, June 1973, Ed., by R.W. Staehle, J. Hochmann, R.D. McCright and J.E. Slatern, NACE, Houston, TX, 1977, pp. 1063–1065. 32. J.S. Armijo, J.R. Low and U.E. Wolff, Nuclear Applications, Vol. 1, 1965, p. 462. 33. T.J. Pashos et al., ‘Failure Experience with Stainless Steel Clad Fuel Rods in VBWR’, Trans. Am. Nuclear Society, Vol. 7, No. 2, 1964, p. 416. 34. Y.J. Kim, L.W. Niedrach, M.E. Indig and P.L. Andresen, ‘Applications of Noble Metals in Coatings and Alloys for Light Water Reactors’, Journal of Metals, Vol. 44, No. 2, 1992, pp. 14–18. 35. P.L. Andresen, ‘Application of Noble Metal Technology for Mitigation of Stress Corrosion Cracking in BWRs’, Proc. Seventh International Symposium on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE, 1995, pp. 563–578. 36. P.L. Andresen, Y.J. Kim, T.P. Diaz and S. Hettiarachchi, ‘Mitigation of SCC by Online NobleChem’, Proc. 13th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, Canadian Nuclear Society, 2007. 37 P.L. Andresen, Y.J. Kim, T.P. Diaz and S. Hettiarachchi, ‘Online Catalytic Mitigation of SCC at Parts Per Trillion Level’, Paper 1683, Corrosion/08, NACE, Houston, TX, 2008. 38. P.L. Andresen, K. Gott and J.L. Nelson, ‘Stress Corrosion Cracking of Sensitized Type 304 Stainless Steel in 288C Water: A Five Laboratory Round Robin’, Proc. Ninth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1999. 39. P.L. Andresen, P.W. Emigh, M.M. Morra and J. Hickling, ‘Effects of PWR Primary Water Chemistry and Deaerated Water on SCC’, Proc. 12th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, TMS, Snowbird, August 2005. 40. P.L. Andresen and J. Hickling, ‘Effects of B/Li/pH on PWSCC Growth Rates in Ni-base Alloys’, EPRI Final Report 1015008 (MRP-217), August 2007. 41. P.L. Andresen, ‘Mitigation of PWSCC in Nickel-base Alloys by Optimizing H2 in Primary Water’, Report to EPRI, Report 1016603 (MRP-252), December 2008. 42. P.L. Andresen, J. Hickling, K.S. Ahluwalia and J.A. Wilson, ‘Effects of Hydrogen on SCC Growth Rate of Ni Alloys in High Temperature Water’, Corrosion, Vol. 64, No. 9, 2008, p. 707.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

305

43. P.L. Andresen, J. Hickling, K.S. Ahluwalia and J.A. Wilson, ‘Effect of Dissolved H2 in Primary Water on the SCC Growth Rate of Ni Alloys’, Proc. Int. Conf. on Water Chemistry of Nuclear Reactor Systems, Berlin, VGB, 2008. 44. S.A. Attanasio and D.S. Morton, ‘Measurement of the Ni/NiO Transition in Ni-Cr-Fe Alloys and Updated Data and Correlation to Quantify the Effect of Aqueous Hydrogen on Primary Water SCC’, Proc. 11th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems, ANS, 2003. 45. ‘Materials Reliability Program: Crack Growth Rates for Evaluating Primary Water Stress Corrosion Cracking (PWSCC) of Alloy 82, 182, and 132 Welds’, Report 1006696 (MRP-115), EPRI, Palo Alto, CA. 46. D. Morton, S. Attanasio, E. Richey, G. Young, ‘In Search of the True Temperature and Stress Intensity Factor Dependences for PWSCC’, 12th International Conference on Environmental Degradation of Materials in Nuclear Systems, 2005. 47. K. Arioka, T. Yamada, T. Terachi and G. Chiba, ‘Cold Work and Temperature Dependence of SCC Growth of Austenitic Stainless Steel in Hydrogenated and Oxygenated High Temperature Water’, Corrosion, Vol. 63, No. 12, 2007, pp. 1115–1123. 48. O.K. Chopra and D.J. Gavenda, ‘Effects of LWR Coolant Environments on Fatigue Lives of Austenitic Stainless Steels’, J. Pressure Vessel Technol., Vol. 120, 1998, pp. 116–121. 49. M. Higuchi and K. Iida, ‘Reduction in Low–Cycle Fatigue Life of Austenitic Stainless Steels in High–Temperature Water’, in Pressure Vessel and Piping Codes and Standards, PVP, Vol. 353, ASME, New York, 1997, pp. 79–86. 50. S. Ritter and H-P Seifert, ‘Corrosion Fatigue Crack Growth Behavior of Austenitic Stainless Steels under Simulated LWR Conditions’, Proc. 17th International Corrosion Congress, Las Vegas, NACE, Houston, TX, 2008. 51. K. Fukuya, K. Fujii, M. Nakano, N. Nakajima and M. Kodama, ‘Stress Corrosion Cracking on Cold-Worked 316 Stainless Steels Irradiated to High Fluence’, Proc. 10th Int. Symp. on Env. Deg. of Materials in Nuclear Power Systems – Water Reactors’, TMS, Snowbird, 2003. 52. K. Arioka, Y. Kanashima, T. Yamada and T. Terachi, ‘Influence of Boric Acid, Hydrogen Concentration and Grain Boundary Carbides on IGSCC Behaviors of SUS 316 Under PWR Primary Water’, Proc. 11th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, ANS, 2003. 53. S.A. Attanasio, D.S. Morton, M.A. Ando, N.F. Panayotou and C.D. Thompson, ‘Measurement of the Ni/NiO Phase Transition in High Temperature Hydrogenated Water Using the Contact Electrical Resistance (CER) Technique’, Proc. 10th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems, NACE, 2001. 54. D.S. Morton, S.A. Attanasio and G.A. Young, ‘Primary Water SCC Understanding and Characterization Through Fundamental Understanding in the Vicinity of the Ni/NiO Phase Transition’, Proc. 10th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE, 2001. 55. S. Attanasio, J. Mullen, J. Wuthrich, W. Wilkening, D. Morton, ‘SCC Growth Rates of Alloy 182 and 82 Welds’, NRC Conference on PWR Vessel Penetration Inspection, Cracking and Repair, Gaithersburg, MD, September 2003. 56. D. Morton, S. Attanasio, E. Richey, G. Young and R. Etien, ‘Updated Data and Correlation to Quantify the Effect of Aqueous Hydrogen and Low Temperature on the SCC Growth Rate of Nickel-base Alloys in Primary Water’, Proc. Alloy 600 Conference, Atlanta, June 2007, EPRI, Palo Alto, CA. © Woodhead Publishing Limited, 2010

306

Understanding and mitigating ageing in nuclear power plants

57. S.M. Bruemmer, J.S. Vetrano and M.B. Toloczko, ‘Microstructure and SCC Crack Growth of Nickel-Base Alloy 182 Weld Metal in Simulated PWR Primary Water’, Proc. 13th Int. Symp. on Env. Degradation of Materials in Nuclear Power Systems – Water Reactors, CNS, 2007. 58. C.L. Briant and P.L. Andresen, ‘Role of S, P and N Segregation on Intergranular Environmental Cracking of Iron and Nickel Base Alloys in High Temperature Water’, Proc. Third Int. Conf. Degradation of Materials in Nuclear Power Industry – Water Reactors, Traverse City, TMS-AIME, Warrendale, PA, 1989, pp. 371–382. 59. P.L. Andresen, ‘Effect of Noble Metal Coating and Alloying on the Stress Corrosion Crack Growth Rate of Stainless Steel in 288C Water’, Proc. Sixth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1994, pp. 245–253. 60. S.M. Bruemmer, B.W. Arey and L.A. Charlot, Proc. 6th Int Symp on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1994, pp. 277–285. 61. T.M. Angeliu, P.L. Andresen, J.A. Sutliff and R.M. Horn, ‘Intergranular Stress Corrosion Cracking of Unsensitized Stainless Steels in BWR Environments’, Proc. Ninth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1999, pp. 311–318. 62. T.M. Angeliu, P.L. Andresen, E. Hall, J.A. Sutliff, S. Sitzman, ‘Strain and Microstructure Characterization of Austenitic Stainless Steel Weld HAZs’, Corrosion/2000, Paper 00186, NACE, 2000. 63. P.L. Andresen, T.M. Angeliu, W.R. Catlin, L.M. Young and R.M. Horn, ‘Effect of Deformation on SCC of Unsensitized Stainless Steel’, Corrosion/2000, Paper 00203, NACE, 2000. 64. P.L. Andresen, T.M. Angeliu and L.M. Young, ‘Effect of Martensite and Hydrogen on SCC of Stainless Steels’, Corrosion/01, Paper #01228, NACE, 2001. 65. P.L. Andresen, L.M. Young, P.W. Emigh and R.M. Horn, ‘Stress Corrosion Crack Growth Rate Behavior of Ni Alloys 182 and 600 in High Temperature Water’, Corrosion/02, Paper 02510, NACE, 2002. 66. P.L. Andresen, ‘Similarity of Cold Work and Radiation Hardening in Enhancing Yield Strength and SCC Growth of Stainless Steel in Hot Water’, Corrosion/02, Paper 02509, NACE, 2002. 67. D.S. Morton, S.A. Attanasio, J.S. Fish, and M.K. Schurman, ‘Influence of Dissolved Hydrogen on Nickel Alloy SCC in High Temperature Water’, Corrosion/99, Paper 99447, NACE, 1999. 68. D.S. Morton, S.A. Attanasio, G.A. Young, P.L. Andresen and T.M. Angeliu, ‘The Influence of Dissolved Hydrogen on Nickel Alloy SCC: A Window to Fundamental Insight’, Corrosion 2001, Paper 01117, NACE, 2001. 69. A.J. Jacobs, ‘Hydrogen Buildup in Irradiated Type 304 Stainless Steel’, 13th Symp. Radiation Induced Changes in Microstructure, Ed. by F.A. Garner, N.H. Packan and A.S. Kumar, STP 956, Vol. II, ASTM, 1985, p. 239. 70. P.L. Andresen, ‘SCC Testing and Data Quality Considerations’, Ninth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1999. See also, P.L. Andresen, ‘Experimental Quality Guidelines for SCC Testing’, GE CRD, January 30, 1998. 71. P.L. Andresen and M.M. Morra, ‘Effects of Positive and Negative dK/da on SCC Growth Rates’, Proc. 12th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors’, TMS, Snowbird, August 2005.

© Woodhead Publishing Limited, 2010

Stress corrosion cracking of austenitic stainless steels

307

72. P.L. Andresen and M.M. Morra, ‘Effect of Rising and Falling K Profiles on SCC Growth Rates in High Temperature Water’, Journal of Pressure Vessel Technology, Vol. 129, No. 3, 2007, pp. 488–506. 73. P.L. Andresen and M.M. Morra, ‘Effects of Si on SCC of Irradiated and Unirradiated Stainless Steels and Nickel Alloys’, Proc. 12th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors’, TMS, Snowbird, August 2005. 74. P.L. Andresen, ‘Emerging Issues and Fundamental Processes in Environmental Cracking in Hot Water’, Corrosion, Vol. 64, No. 5, 2008, pp 439–464. 75. C.M. Brown and W.J. Mills, ‘Load Path Effects on the Fracture Toughness of Alloy 82H and 52 Welds in Low Temperature Water’, Proc. 12th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors’, TMS, Snowbird, August 2005. 76. P.L. Andresen, F.P. Ford, T.M. Angeliu and R.M. Horn, ‘Stress Corrosion Cracking Initiation in Austenitic Stainless Steel in High Temperature Water’, Proc. Ninth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1999. 77. T. Shoji, ‘Progress in the Mechanistic Understanding of BWR SCC and Its Implications to the Prediction of SCC Growth Behavior in Plants’, Proc. 11th Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, ANS, 2003. 78. Y.C. Gao and K.C. Hwang, ‘Elastic Plastic Fields in Steady Crack Growth in a Strain-Hardening Material’, 5th Int. Conf. on Fracture, 1981, pp. 669–682. 79. L.M. Young, P.L. Andresen and T.M. Angeliu, ‘Crack Tip Strain Rate: Estimates Based on Continuum Theory and Experimental Measurement’, Corrosion/2001, Paper 01131, NACE, 2001. 80. H. P. Offer, R.M. Horn, A.Q. Chan and M.M. Morra, ‘Assessment of the Mitigation of SCC by Surface Stress and Material Improvements’, 13th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems, Canadian Nuclear Society, 2007. 81. B.M. Gordon, R.L. Cowan, C.W. Jewett and A.E. Pickett, ‘Mitigation of Stress Corrosion Cracking through Suppression of Radiolytic Oxygen’, Proc. 1st Int. Symp. on Environmental Degradation of Materials in Nuclear Power System – Water Reactors, NACE, 1983, p. 893. 82. F.P. Ford, P.L. Andresen, T.M. Angeliu, H.D. Solomon, R.M. Horn, R.L. Cowan, ‘Prediction and Mitigation of Cracking in BWR Core Components’, Proc. Ninth Int. Symp. on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, AIME, 1999. 83. P.L. Andresen, M.M. Morra, J. Hickling, K.S. Ahluwalia and J.A. Wilson, ‘Effect of Deformation and Orientation on SCC of Alloy 690’, Corrosion/09, Paper 4840, NACE, Houston, TX, 2009. 84. D.J. Paraventi and W.C. Moshier, ‘Alloy 690 SCC Growth Rate Testing’, Workshop on Cold Work in Iron- and Nickel-Base Alloys, Ed. by R.W. Staehle and J. Gorman, June 2007, EPRI, Palo Alto. 85. D.J. Paraventi and W.C. Moshier, ‘Alloy 690 SCC Growth Rate Testing’, Proc. EPRI Alloy 690 Workshop, Atlanta, 31 October 2007.

© Woodhead Publishing Limited, 2010