4.13.3.2.1 Materials of construction

Nuclear safety-related concrete structures are com­posed of several constituents that, in concert, perform multiple functions (e. g., load-carrying capacity, radi­ation shielding, and leak tightness). Primarily, these constituents can include the following material sys­tems: concrete, conventional steel reinforcement, pre­stressing steel, and steel liner plate. The quality of these materials is established through regulations, qualification tests, and certification, followed by check­ing throughout construction. More detailed informa­tion on materials of construction than provided later is available elsewhere.35,58-60

Concrete is a composite material consisting of a binder (cement paste) and a filler of fine or fine and coarse aggregate particles that combine to form a synthetic conglomerate. Cement is a mixture of compounds made by grinding crushed limestone, clay, sand, and iron ore together to form a homoge­neous powder that is then heated at very high tem­peratures ranging from 1400 to 1600 °C to form a clinker.59 After the clinker cools, it is ground and mixed with a small amount of gypsum to regulate setting and facilitate placement. This produces the general-purpose Portland cement, which is mixed with water to produce cement paste that binds the aggregate particles together. (Current generation cements have higher tricalcium silicate (C3S) contents and are ground finer than previous cements. Current cements attain most of their compressive strength within a 28-day period, whereas the previous cements continued to gain strength after 28 days.6,61)

Portland cements are composed primarily of four chemical compounds: (C3S), dicalcium silicate (C2S), tricalcium aluminate (C3A), and tetracalcium aluminoferrite (C4AF). The type of Portland cement produced (e. g., general purpose, moderate sulfate resistance and heat of hydration, high early strength, low heat of hydration, and sulfate resistant) depends on the relative amounts of the four basic chemical compounds and fineness (high early strength). The calcium silicate hydrates (C-S-H) constitute about 75% of the mass. The C-S-H gel structure is made up of three types of groups that contribute to bonds across surfaces or in the interlayer of partly crystal­lized tobermorite material: calcium ions, siloxanes, and water molecules. Bonding of the water within the layers (gel water) with other groups via hydrogen bonds determines the strength, stiffness, and creep properties of the cement paste.

There are also a number of alternative or sup­plementary cementing agents that have been used in conjunction with Portland cement, and these are pulverized fly ash, ground granulated blast furnace slag (GGBFS), and silica fume. Fly ash is collected from the exhaust flow of furnaces burning finely ground coal and reacts with calcium hydroxide in the presence of water to form cement compounds consisting of calcium silicate hydrate. GGBFS is a by-product of the iron-making process and is formed by taking the hot slag, rapidly chilling or quenching it, and grinding into a powder. When mixed with water in the presence of an alkaline environment provided by the Portland cement, GGBFS hydrates to form cementing compounds consisting of calcium silicate hydrate. Silica fume is the condensed vapor by-product of the ferrosilicon smelting process. Silica fume reacts with calcium hydroxide in the presence of water to form cementing compounds consisting of calcium silicate hydrate. High alumina cement, con­sisting mainly of calcium aluminates, has been utilized as a cementitious material because of its rapid set and rapid strength gain characteristics and resistance to acidic environments, sea water, and sulfates. However, owing to certain conditions of temperature and humidity, the cement converts over time to a different hydrate having reduced volume (i. e., increased poros­ity and reduced strength), it is recommended that calcium aluminate cements not be used for structural applications (particularly in wet or humid conditions above 27 °C).62

Selection of the proper water content of concrete is critical, since too much water reduces the concrete strength and insufficient water makes the concrete un­workable. Hardening of concrete occurs as a result of hydration, which is a chemical reaction in which the major compounds in the cement form chemical bonds with water molecules and become hydrates. The hard­ened cement paste consists mainly of calcium silicate hydrates, calcium hydroxide, and lower proportions of calcium sulfoaluminate hydrate either as ettringite or monosulfate. About 20% of the hardened cement paste volume is calcium hydroxide. The pore solution is normally a saturated solution of calcium hydroxide within which high concentrations of potassium and sodium hydroxides are present. Proper curing of the concrete during this stage is essential, as it affects the concrete’s durability, strength, water-tightness, abrasion resistance, volume stability, and resistance to freezing and thawing.

Since cement is the most expensive ingredient in concrete, it is desirable to utilize the minimum amount necessary to produce the desired properties and characteristics. Aggregate typically occupies 60-75% of the volume of concrete and therefore its characteristics strongly influence the chemical, physical, and thermal properties of concrete, its mix proportions, and economy. (The balance of the concrete mix generally consists of 10-15% cement, 15-20% water, and air (5-8% if entrained).) Aggre­gates thus are important with respect to the concrete durability. The aggregates come in various shapes, sizes, and material types ranging from fine sand par­ticles to large coarse rocks. Selection of the aggregate material is determined in part by the desired char­acteristics of the concrete. Aggregate materials are available ranging from ultra-lightweight (e. g., ver — miculite and perite) to lightweight (e. g., expanded clay shale or slate-crushed brick), normal weight (e. g., crushed limestone or river gravel), and heavy­weight (e. g., steel or iron shot). Sometimes chemical or mineral admixtures are added during the mixing process to enhance durability (air entrainment), improve workability (enhanced placement and com­paction), modify hardening and setting characteris­tics, aid in curing, reduce heat evolution, or provide other property improvements.63

The concrete typically used in nuclear safety — related structures consists of Type II Portland cement,59 fine aggregates (e. g., sand), water, various minerals, or chemical admixtures for improving properties or performance of the concrete and either normal-weight or heavy-weight coarse aggregate.

American Society of Testing and Materials (ASTM) C 150,64 Type II Portland cement, typically has been used because of its improved sulfate resistance and reduced heat of hydration relative to the general — purpose Type I Portland cement. Both the water and fine and coarse aggregates are normally acquired from local sources and subjected to material charac­terization testing prior to use. Coarse aggregate can consist of gravel, crushed gravel, or crushed stone. Chemical (e. g., air-entraining or water-reducing) or mineral (e. g., fly ash or ground granulated blast fur­nace slag) admixtures have been utilized in many of the mixes to impart improved characteristics or per­formance. For those concrete structures in NPPs that provide primary (biological) radiation shielding, heavy-weight or dense aggregate materials, such as barites, limonites, magnetites, and ilmenites, may have been used to reduce the section thickness and meet attenuation requirements.

The constituents are proportioned and mixed to develop Portland cement concrete that has specific properties. Depending on the characteristics of the specific structure, the concrete mix may be adjusted to provide increased strength, higher durability, or better workability for placement. The hardened concrete typically provides the compressive load­carrying capacity for the structure. Specified con­crete unconfined compressive strengths typically have ranged from 13 to 55 MPa, with 35 MPa being a typical value achieved at 28 days age.

Concrete tensile strength is about one-tenth to one-fifth of its compressive strength, so concrete cannot be relied upon to withstand very high tensile stresses. This limitation is overcome by embedding steel reinforcement in the concrete so that the concrete and steel work in concert. In addition to resisting tensile loads, the bonded steel reinforce­ment is used to control the extent and width of cracks, especially where it is desirable to reduce member cross-sections. Steel reinforcement is also used in compression members to safeguard against the effects of unanticipated bending moments that could crack or even fail the member. The effectiveness of rein­forced concrete as a structural material depends on the interfacial bond between the steel and concrete so that it acts as a composite material, the passivating effect of the highly alkaline concrete environment to inhibit steel corrosion (see next section), and the similar coefficients of thermal expansion of the con­crete and steel. Most of the mild, or conventional, reinforcing steels used in NPPs to provide primary tensile and shear load resistance/transfer consist of plain carbon steel bar stock with deformations (lugs or protrusions) on the surface. These bars typically conform to ASTM A61565 or A 70666 specifications. The minimum yield strength for the steel reinforce­ment ranges from 280 to 520 MPa, with the 420 MPa strength material being most common.

Post-tensioning is a method of reinforcing (or strengthening) concrete with high-strength tendons to resist tensile loadings and to apply compressive forces to the concrete to provide increased resistance to concrete cracking. A number of NPP concrete containment structures utilize post-tensioned steel tendons that are designed to have (1) consistently high strength and strain at failure, (2) serviceability throughout their lifetime, (3) reliable and safe pre­stressing procedures, and (4) ability to be retensioned and replaced (nongrouted systems). The tendons are installed within preplaced ducts in the containment structure and post-tensioned from one or both ends after the concrete has achieved sufficient strength. After tensioning, the tendons are anchored by button-heads, wedges, or nuts. Corrosion protection is provided by filling the ducts with wax or corrosion- inhibiting grease (unbonded) or portland cement grout (bonded). (Although bonded post-tensioning tendons are less vulnerable to local damage than ungrouted tendons, ungrouted tendons have been primarily used in the United States because the grouted tendon systems cannot be visually inspected, mechanically tested, or retensioned in the event of a larger than anticipated loss of prestressing force.) Supplemental conventional reinforcing is also used to minimize shrinkage or temperature effects and to provide local load-carrying capacity or load transfer. Three major categories of post-tensioning system exist depending on the type of material utilized to fabricate the tendons: wire, strand, or bar that con­form to ASTM specifications A 421,67 A 416,68 and A 7 2 2,69 respectively. Minimum tensile strengths range from 1620 to 1725 MPa for the A 421 material and 1725 to 1860 MPa for the A 416 material. The A 722 material has a minimum tensile strength of 1035 MPa. Typical NPP tendon systems group sufficient numbers of wires, strands, or bars to have minimum ultimate strengths ranging from 2000 to 10 000 kN. The trend has been to increase the strength of the tendons to reduce the total number (e. g., in the early 1970s, the typical tendon had a capacity of 3000 kN and since then has progressed to capacities of 10 300 and 15 300 kN).19 With the exception of Robinsion 2 (bar tendons) and Three Mile Island 2 (strand tendons), plants that have post-tensioned containments utilize unbonded tendons so that the tendons can be inspected and replaced (if necessary). Bellefonte and Ginna each has grouted tendons (rock anchors) to which tendons are attached.

Leak tightness of reinforced and post-tensioned concrete containment vessels is provided by a steel liner plate. A typical liner is composed of steel plate stock <13 mm thick, joined by welding, and anchored to the concrete by studs (Nelson studs or similar conforming to ASTM A 10870), structural steel shapes, or other steel products. PWR containments and the drywell portions of BWR containments are typically lined with carbon steel (ASTM A 3671 or A 51672). The liners of LWR fuel pool structures typically consist of stainless steel (ASTM A 27673 or A 30474). The liners of wetwells also have used car­bon steel materials such as ASTM A 285,75 A 516, and A 53 7.76 Certain LWR facilities also have used carbon steel clad with stainless steel weld metal for liner members. Although the liner’s primary function is to provide a leak-tight barrier, it acts as part of the formwork during concrete placement and may be used in the support of internal piping/equipment. The liner is not considered to contribute to the strength of the structure.

After the cascade formation and relaxation, which last for a few picoseconds, the microstructure evolves over a longer time. The evolution of the microstruc­ture is driven by the bulk diffusion of point defects. For a better understanding of the microstructure evolution under irradiation, the elementary proper­ties of point defects, such as formation energy and migration energy, have first to be examined.

4.01.1.2.1 Vacancy formation and migration energies

Concerning the vacancy, all the atomic positions are identical in the lattice and so there is only one vacancy description leading to a unique value for the vacancy formation energy. Due to the rather low a—p phase transformation temperature, the measurement of vacancy formation and migration energy in the Zr hexagonal close-packed (hcp) phase is difficult. The temperature that can be reached is not high enough to obtain an accurately measurable concentration and mobility of vacancies.18 Nevertheless, various experi­mental techniques (Table 1), such as positron annihila­tion spectroscopy or diffusion of radioactive isotopes, have been used in order to measure the vacancy formation and migration energies or the self-diffusion

coefficient.18-26 The values obtained by the various authors are given in Table 1. It is pointed out by Hood18 that there is great discrepancy among the vari­ous results. It is particularly shown that at high tem­perature, the self-diffusion activation energy is rather low compared to the usual self-diffusion activation energy in other metals.18 However, as the temperature decreases, the self-diffusion activation energy increases strongly. According to Hood,18 this phenomenon can be explained assuming that at high temperature the vacancy mobility is enhanced by some impurity such as an ultrafast species like iron. At lower tempera­ture, the iron atoms are believed to form small pre­cipitates, explaining that at low temperatures the measured self-diffusion energy is coherent with usual intrinsic self-diffusion of hcp crystals. It is also shown that the self-diffusion anisotropy remains low for normal-purity zirconium, with a slightly higher mobil­ity in the basal plane than along the (c) axis.22,26,27 For high-purity zirconium, with a very low iron con­tent, the anisotropy is reversed, with a higher mobility along the (c) axis than in the basal plane.27

The vacancy formation and migration energies have also been computed either by MD methods, where the mean displacement distance versus time allows obtaining the diffusion coefficient, or by static computation of the energy barrier corresponding to the transition between two positions of the vacancy using either empirical interatomic potential7,28-34 or the most recent ab initio tools.35-38 Since the different sites surrounding the vacancy are not similar, due to the non-ideal c/a ratio, the migration energies are expected to depend on the crystallographic direction, that is, the migration energies in the basal plane Em and along the (c) direction E? are different. The results are given in Table 2.

The atomistic calculations are in agreement with the positron annihilation spectroscopy measurement but are in disagreement with the direct measure­ments of self-diffusion in hcp zirconium.2 As dis­cussed by Hood,18 and recently modeled by several authors,39, 0 this phenomenon is attributed to the enhanced diffusion due to coupling with the ultrafast diffusion of iron.

4.01.1.2.2 SIA formation and migration energies

In the case of SIAs, the insertion of an additional atom in the crystal lattice leads to a great distortion of the lattice. Therefore, only a limited number of configurations are possible. The geometrical descrip­tion of all the interstitial configuration sites has been

proposed for titanium byJohnson and Beeler41 and is generally adopted by the scientific community for other hcp structures (Figure 2).

• T is the simplest tetrahedral site, and O is the octahedral one, with, respectively, 4 and 6 coordi­nation numbers.

• BT and BO are similar sites projected to the basal plane with three nearest neighbors, but with dif­ferent numbers of second neighbors.

• BC is the crowdion extended defect located in the middle of a segment linking two basal atoms.

• C is the interstitial atom located between two adjacent atoms of two adjacent basal planes in the (2023) direction. This direction is not a close- packed direction, and allows easier insertion of the SIA.

• S is the split dumbbell position in the (c) direction.

The only way to have access to the SIA formation energy is from atomistic computations taking into account the different configurations of the SIA given previously. In their early work on titanium, Johnson and Beeler41 found that the most stable SIA configuration was the basal-octahedral site (BO). Several other sites were also found to be metastable, like asymmetric variants of the T and C sites. As reviewed by Willaime,35 the relative stabilities of the various SIA configurations were observed to depend strongly on the interatomic potential used (Table 3).

The mobility of SIAs can be estimated experimen­tally using electron irradiation at very low tempera­tures (4.2 K), followed by a heat treatment. During the recovery, the electrical resistivity is measured. The main recovery process was found around 100-120 K and analysis of the kinetics gives the SIA migration energy of Em ~ 0.26 eV.4

Atomistic computations have also brought results (Table 3) concerning the SIA migration energy. Sev­eral authors7,28-31,33-37 have found that the mobility of SIAs is anisotropic, with low migration activation energy for the basal plane mobility (E/J ~ 0.06 eV) and a higher migration activation energy in the (c) direction (E? ~ 0.15 eV). In the temperature range of interest for the power reactors (T~ 600 K), the diffusion coefficients obtained are the following: D1! = 8 x 10~9m2s_1 (in the basal plane) and

D? = 10~9 mm2s-1 (along the (c) direction). These authors have also shown that the anisotropy depends on the temperature. Computing the effective diffu­sion rate of SIAs in all directions, taking into account the multiplicity of the jump configurations for each type of migration, Woo and co-workers34,42 have obtained the anisotropy for self-interstitial diffusion as a function of temperature. It is shown that the SIA mobility is higher in the basal plane than along the (c) axis and that the anisotropy decreases when the temperature increases.

For many years it was assumed that void swelling would not be an issue for the 304 and 316 stainless components comprising the internals of power — producing light water-cooled reactors. Such a conclu­sion was easily accepted for boiling water reactors since steels used in the shroud assembly are separated from the core by a substantial water gap and therefore experience less than 5 dpa over a 40-year lifetime. For pressurized water reactors, however, the steel is much closer to the core and some regions can reach 80-100 dpa over 40 years. Swelling was still not thought to be a problem because swelling was per­ceived to inhabit a temperature range that did not extend down to the 280-290 °C inlet temperatures of PWRs, and based on most fast reactor irradiations, swelling was thought to vanish below 345 °C, the max­imum water temperature in PWRs. It was also thought that the lower dpa rates characteristic of PWRs would reduce vacancy supersaturations and would therefore inhibit void nucleation.

. — $ *

■ v

L.

Figure 57 Voids observed in Tihange baffle-former bolt made with cold-worked 316 stainless steel after irradiation at ~345°C to 12 dpa. Reproduced from Edwards, D. J.; Simonen, E. P.; Garner, F. A.; Greenwood, L. R.; Oliver,

B. A.; Bruemmer, S. M. J. Nucl. Mater. 2003, 317, 32-45.

Unfortunately, gamma and nuclear heating of thick plates can raise the internal temperatures in some areas of the baffle-former plates to temperatures above 400 °C, known to be prime territory for void swelling. Also, as seen in the previous section, void nucleation does not dominate the swelling response to decreasing dpa rate. The shortening of the tran­sient regime at lower dpa rates raised the strong possibility that void swelling might indeed occur in PWR internals. Theoretical considerations based on void nucleation also suggested that the temperature regime of swelling might move to lower temperatures with decreasing dpa rates. Therefore an effort was made to find stainless steels irradiated at lower dpa rates and/or lower temperatures.

The first clear example of void swelling in PWRs was found in a cold-worked 316 baffle bolt removed from the Tihange PWR reactor located in Belgium.39 The bolt was removed in response to an ultrasonic indication of cracking under the bolt head.

Although the bolt shown in Figure 57 was con­structed from cold-worked 316 austenitic stainless steel known to be more resistant to the onset of swelling than the annealed AISI 304 plate in which it was embedded, well-faceted voids of easily resolv­able size were clearly observed in three sections removed along the bolt axis. The doses in the bolt were relatively low and the calculated temperatures were also relatively low compared to typical fast reactor observations, but the swelling exceeded expectations based on fast reactor experience. As cold-worked 316 is known to always swell less than
annealed 304 at the same temperature and dpa rate the worrisome inference is that the 304 plate sur­rounding the bolt might be swelling at higher levels.

Significantly, hydrogen was also found to be stored in the microstructure at unexpectedly high levels that increased as void swelling increased along the bolt length.

Subsequently, voids were observed in other AISI 316 bolts from this same reactor by other research­ers148 often at even lower doses and temperatures, producing lesser but measurable amounts of swelling. An example is shown in Figure 58, but it should be noted that there appear to be two populations of cavities, a few that are recognizable as voids and an exceptionally high population of nanometer-sized cavities that are only visible using a large level of defocusing, similar to the behavior shown earlier in Figure 17.

Voids have been sometimes but not always observed in bolts of various steels removed from US PWRs.149,150 These studies were conducted before the need for defocusing was recognized, however. Small cavities that could be either voids or bubbles have also been observed in thin-walled flux thimble tubes removed from various PWRs.76,151-153 Neus- troev and coworkers also found voids in a thimble tube removed from a VVER operating in the Ukraine, noting that voids were observed at unexpectedly low temperatures and dpa levels.154

The potential for void swelling at PWR-relevant dpa rates and temperatures is best demonstrated in more comprehensive studies conducted in four USSR sodium-cooled fast reactors located in Russia and Kazakhstan. Whereas the inlet temperature of most Western or Asian fast reactors was of the order of 365-375 °C, the Soviet BOR-60 and BN-350 fast reactors had inlet temperatures of the order of 270-280°C. Components from regions below the core or in the reflector region have been extracted for study at dpa rates and temperatures that were

comparable to those of PWRs.155-161

A summary paper containing an overview of these studies shows that in all studies conducted on compo­nents removed from low flux positions in Soviet fast reactors, certain recurrent trends were observed.155 First, whenever the dpa rate was significantly lower at any investigated temperature, swelling was observed at surprisingly very low dpa levels. An excellent exam­ple is shown in Figure 59 where significant void swelling was observed at only 0.64 dpa at 350 °C.156 Second, whenever a comparison could be made within one reactor at a given temperature, the transient dura­tion decreased with lower dpa rate.157-159 Most impor­tantly, whenever temperatures approaching 280 °C could be reached, swelling was observed not only at these low temperatures but also at surprisingly low dpa levels.160,161 Other examples are shown in Figures 60 and 61.

Microstructural development of ferritic steels during service is driven by the interaction of neutrons and metal atoms. Collisions between the incident neutron and constituent atoms result in momentum transfer to the lattice atoms. If the transfer is above ^40 eV, atoms can be permanently displaced from their lattice sites resulting in vacancy-interstitial pairs. Indeed, lattice atoms with tens of kiloelectron volts may be created and a branching, tree-like distribution of displaced atoms formed, termed a displacement cas­cade. Vacancy and interstitial clustering may occur within the cascade. Vacancies and interstitials escap­ing from the cascade give rise to concentrations of vacancy and interstitial point defects throughout the material. The fate of the point defects formed in the irradiation depends most sensitively on irradia­tion dose rate and temperature, and also material factors such as composition.

This chapter focuses on the microstructural development of RPV steels. These steels experience a relatively low dose rate (and thus lifetime dose) at a relatively low temperature. The microstructure developed under these conditions is very different to that developed at high doses and high tempera­tures in the operating regime typical of fusion or fast reactors. The critical features are that at the temper­ature range of most operating pressure vessels both vacancy and interstitial point defects are mobile. Freely migrating defects (vacancies or interstitials) and mobile interstitial clusters escaping from the initial damage event may interact with point defect sinks, such as preexisting dislocations, recombine with each other, either directly or at solute traps, or cluster to form vacancy or interstitial clusters. The end result of the interactions described above is a microstructure comprising of a high density of small clusters in the matrix. These clusters may be point defect solute complexes, and, depending on steel composition, solute clusters formed from radiation-enhanced precipitation. At the dose, dose rate range, and temperature of interest to operating reactors, the clusters formed are usually thermally stable. Lastly, segregation of solutes to grain bound­aries or other sinks may occur.

There have been several studies on the effect of neu­tron irradiation on the strength of various types of SiC forms including reaction-bonded, sintered, pressure­less sintered, and CVD SiC materials.1’11’13’14’58’60-65

1

Figure 13 Irradiation temperature dependence of irradiated elastic modulus of CVD SiC, at ambient temperature, normalized to unirradiated values. The error bars are showing standard deviations for all the neutron data points and ranges of data scatter for the ion data points. Reproduced from Snead, L. L.; Nozawa, T.; Katoh, Y.; Byun, T-S.; Kondo, S.; Petti, D. A. J. Nucl. Mater. 2007, 371, 329-377.

СЛ

=J

"O

1.00

СЛ

o>

c

=J

о

>-

0.90

із

ф

DC

0.80 0.70

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Volumetric swelling (%)

Figure 14 Irradiation-induced change of elastic modulus versus swelling of CVD SiC. An estimation of the influence of lattice relaxation on elastic modulus is calculated using Tersoff potential. Reproduced from Snead, L. L.; Nozawa, T.; Katoh, Y.; Byun, T-S.; Kondo, S.; Petti, D. A. J. Nucl. Mater. 2007, 371, 329-377.

Si for reaction-bonded SiC, which typically segregate to grain boundaries during sintering, tends to have a significant influence on strength under neutron irradiation. For the case of sintered SiC with boron
compounds as sintering additives, the reaction of 10B(n, a)7Li causes the accumulation of helium bub­bles at and near the grain boundary phases under neutron irradiation.60-63 In contrast, unmatched

1.5
1 1 1 …………………………….. I 1 1 …………………………….. I г Hot pressed and sintered SiC forms
1.0
10	100

Figure 17 Fluence dependence of irradiated flexural strength of hot-pressed and sintered SiC normalized to unirradiated strength. Irradiation is variable but in the saturable swelling regime. Reproduced from Newsome, G. A.; Snead, L. L.;

Hinoki, T.; Katoh, Y.; Peters, D. J. Nucl. Mater. 2007, 371, 76-89.

2.0

Dose (dpa)

Figure 18 Flexural strength of CVD SiC at ambient temperature as a function of irradiation dose. Reproduced from Newsome, G. A.; Snead, L. L.; Hinoki, T.; Katoh, Y.; Peters, D. J. Nucl. Mater. 2007, 371, 76-89.

swelling between Si and SiC for reaction-bonded SiC causes disruption at the grain boundary, severely reducing the strength.11,19,60-64 Meanwhile, the high — purity materials such as CVD SiC exhibit superior irradiation resistance.

The irradiation effect on flexural strength of Rohm and Haas CVD SiC in a fluence range of 0.15-30 dpa is summarized in Figure 18. In comparing Figure 18 with Figure 17, it is clear that CVD SiC retains stability in strength to a much higher dose than the sintered and reaction-bonded forms of SiC. It is to be noted that in Figure 18, the data of Dienst does indicate a significant as-irradiated degradation in strength around 15 dpa. However, such degradation is not seen for the ^30 dpa irradiation of Snead. It is speculated that the degradation in the Dienst data may
have been due to statistical limitations of the study and/or due to issues with sample handling postirradi­ation. This issue is discussed in the Dienst reference.65 A compilation of strength data as a function of irradia­tion temperature is given in Figure 18, indicating no apparent correlation for the dose and temperature ranges studied. However, as with the fracture tough­ness data, irradiation-induced strengthening seems to be significant at 573-1273 K. The large scatter in flex­ural strength of brittle ceramics is inevitable, as the fracture strength is determined by the effective frac­ture toughness, morphology, and characteristics of the flaw that caused the fracture. Irradiation possibly modifies both the flaw characteristics and the fracture toughness through potential surface modification, relaxation of the machining-induced local stress, modifications of elastic properties, and fracture energy.

A typical means of describing the failure of cera­mics is through the use of Weibull statistics, which is a departure from the analysis of data that is assumed to follow a normal Gaussian distribution. In the two — parameter Weibull formalism, sometimes referred to as a weakest-link treatment, the failure probability F is described as

F(x) = 1 — e-(x"/x0)

where m is the Weibull modulus and xo is the distribution size parameter. A change in the Weibull statistics, indicating a higher scatter in as-irradiated flexural strength has been observed by previous authors, although the point could not be made convincingly because of limitations in the number of tests observed. In the earliest work known to the authors, Sheldon66 noted a 14% decrease in crushing strength of highly irradiated CVD SiC shells with an increase of the coefficient of variation from 8% to 14%. Price63 went on to a 4-point bend test using relatively thin (~0.6 mm) strips of CVD SiC deposited onto a graphite substrate. In his work, the flexural strength following a ^9.4 x 1025nm-2 (E > 0.1 MeV) irradiation was unchanged within the statistical scatter, but the scatter itself increased from about 10 to 30% of the mean flexural strength as described assuming a normal distribution. Unfortu­nately, there were not sufficient samples in Price’s work to infer Weibull parameters. In more recent work by Dienst,65 the Weibull modulus was reported to decrease from about 10 for irradiation of ~1 x 1026 nm — (E > 0.1 MeV). However, it is worth noting that the Dienst work used a very limited sample population (about 10 bars.)

Statistically meaningful data sets on the effect of flexural strength of CVD SiC have been re­ported by Newsome and coworkers14 and Katoh and coworkers.58,67 Figure 19 shows a compilation Weibull plot of the flexural strength of unirradiated and irra­diated Rohm and Haas CVD SiC taken from the two separate irradiation experiments carried out by New­some and cowokers14 and Katoh and coworkers.58,67 The sample population was in excess of 30 for each case. In Figure 19(a), the data was arranged by irradia­tion temperature, including data for unirradiated samples and 1.5—4.6 x 1026nm-2 (E > 0.1 MeV) dose range. It is likely that the Weibull modulus decreased by irradiation, appearing to be dependent on irradia­tion temperature. This is not easily visualized through inspection of Figure 19(a) unless one notes that there are significantly more low stress fractures populating the 573 K population. The scale parameters of flexural strength of unirradiated materials and materials irra­diated at 573, 773, and 1073 Kwere 450, 618, 578, and 592 MPa, respectively. The Weibull modulus of the flexural strength of unirradiated materials and materi­als irradiated at 573, 773, and 1073 Kwere 9.6, 6.2, 5.5, and 8.7, respectively, with significant uncertainty.

The work of Katoh, on identical material irra­diated at the same temperature as in the Newsome work, is at a slightly higher irradiation dose than the data of Newsome. As seen in Figure 19(b), the effect on the Weibull modulus undergoes a trend similar to that of Newsome, although the modulus for the 773 K and 1073 K irradiation of Katoh remained almost unchanged. Given the data discussed on the effect of irradiation on the Weibull modulus and scale parameter of CVD SiC bend bars, it is clear that the material is somewhat strengthened and that the Weibull modulus may undergo irradiation-induced change, with the greatest decrease occurring for the lowest temperature irradiation.

The fracture strength and failure statistics of tubular SiC ‘TRISO surrogates’ have been deter­mined by the internal pressurization test and the results are plotted in Figure 20. Thin-walled tubular SiC specimens of 1.22 mm outer diameter, 0.1 mm wall thickness, and 5.8 mm length were produced by the fluidized-bed technique alongside TRISO fuels.68 The specimens were irradiated in the HFIR to 1.9 and 4.2 x 1025nm-2 (E > 0.1 MeV) at 1293 and 1553 K. In the internal pressurization test, tensile hoop stress was induced in the wall of the tubular specimens by compressively loading a polyurethane insert.68,69

In Figure 20, Weibull plots of the flexural strength and internal pressurization fracture strength

of unirradiated and irradiated samples are presented. As with the Newsome and Katoh data, the sample population is large enough to be considered statistically meaningful. From this data, the mean fracture stress of tubular specimens is seen to increase to 337 MPa (from 297 MPa) and the Weibull modulus slightly decreased to 3.9 (from 6.9) after irradiation to 1.9 x 1025nm~2 (E> 0.1 MeV) dpa at 1293 K. The mean fracture stresses and Weibull moduli at 4.2 x 1025nm~2 (E> 0.1 MeV) were similar to those at 1.9 dpa. The results for 4.2 dpa irradiation indicate

that by increasing the irradiation temperature from 1293 to 1553 K, no discernible change in fracture stress distribution occurred. The horizontal shift indicates a simple toughening or an increase in frac­ture toughness alone. While the data for these surro­gate TRISO samples, irradiated through internal compression, are somewhat limited, the findings indi­cate that the trend in strength and statistics of failure are consistent with those found for the bend bars. Therefore, the general findings of the bend bar irra­diation on strength and Weibull modulus appear

3 2 1

„ 0

I

Iz

— -2 -3 -4 -5

appropriate for application to TRISO fuel model­ing. Specifically, a slight increase in the mean strength is expected (although it may be less signifi­cant at higher temperatures), and the statistical spread of the fracture data as described by the Weibull modulus may broaden. Unfortunately, a precise description of how the Weibull modulus trends with irradiation dose and temperature is not yet possible, although within the dose range and temperature covered by the data in Figures 19 and 20, a modest reduction is possible.

The most technologically important manifestation of SIC is hydrogen-induced cracking (also known as ‘cold cracking’) caused by the numerous sources of hydrogen in both fabrication and service, the relative ease of hydrogen entry into metals, its high diffusiv- ity, and its ability to weaken metallic bonds or form brittle second phases.36-39 Hydrogen cracking in steels and hydride-type cracking of zirconium alloys

30 Cr alloy cb—І a «3.58
O <P (j)G)
Partially coherent grain boundary Cr23C6
Grain boundary
(Cr, Fe)23C6
(a)	0.1um
Tension between carbides
Cracking between carbides
1um
Acc. V Spot Magn Det WD 20.0 kV 3.5 10000x SE 12.7
2 mm
Small cracks coalesce
‘Ductility Dip’ Crack
100 um
(c)
0.1%
-0.1%
О
-0.6%
3.575	3.580
(d)
Figure 10 Illustration of the mechanism of ductility dip cracking in Ni-Cr alloys. (a) Partially coherent (Cr, Fe)23C6 carbides form in reheated weld metal, which have misfit with the grain boundary. (b) Precipitation generates local grain boundary tensile stresses between carbides which (c) results in ductility dip cracking when sufficiently large global stresses are present during welding or are applied during tensile testing. (d) The increased misfit between the carbide and the matrix with increasing chromium concentration helps explain the susceptibility of alloy 690/EN52 and the resistance of A600/EN82 to DDC.

Figure 12 Correlation of the subsolidus cracking susceptibility of selected superalloys with the misfit and kinetics of second-phase precipitation (g or y"). Adapted from Young, G. A., et al. Welding J. Res. Suppl. 2008, 31S-43S; Prager, M.; Shira, C. S. Welding Res. Council Bull. 1968, 128; calculations done with JMatPro, Version 4.1.

have been treated recently in the literature and more extensive reviews are found elsewhere.40’41

However, it should be highlighted that low — strength austenitic alloys are resistant but not immune to hydrogen-induced cracking. Figure 13 shows a hydrogen-induced crack in a Ni-20Cr-3Mn — 2.5Nb-1Fe weld metal (EN82) that was produced by the combination of poor welding practice and the use of hydrogen-bearing shield gas. The 95%Ar-5%H2 shield gas helps minimize surface oxides and interpass grinding but results in ~12wt ppm hydrogen dis­solved in the filler metal. ‘Refuse welding’ or remelt­ing beads in an attempt to improve the tie-in and contour increases the plastic strain in the joint and can trigger cold cracking.42,43

The heat generated in the graphite (or energy depo­sition) is required for the calculation of the graphite temperatures, and in the case of CO2-cooled systems, it is required for the calculations of radiolytic weight loss. Both of these requirements are important in graphite stress analysis calculations.

In the case ofCO2-cooled systems it is assumed that the graphite radiolytic oxidation rate is proportional to the heat generated in the graphite. However, it is ionizing irradiation that causes the dissociation of the CO2. The energy deposition is produced by the inter­action of graphite atoms with three types of particles:

• Neutron interactions with graphite atoms (~40%).

• Fission g-rays (~60%).

• Secondary g-rays caused by absorption by materi­als outside the moderator (e. g., steel fuel pins in AGRs) and by inelastic scattering of carbon atoms (~1% in a Magnox reactor and ~10% in an AGR).

The main source of gammas and neutrons arises from the fuel, mainly from prompt fission, but there are some from delayed fission.

The ratios given above are for a central position in the core and for initial fuel loading. The ratio may change with position in the core and with graphite
weight loss. Furthermore, in graphite material test programs, the ratio between neutron and g-heating is likely to be significantly different, because of the dif­ferent materials used to construct the various reactor cores. It is therefore important that this ratio is known and the implication of a change in this ratio on material property changes, that is, the implication of the ratio between fast neutron damage versus radiolytic weight loss on graphite property changes, is understood.

The gamma and neutron spectrum varies with distance from the fuel and will vary with graphite density (i. e., will change with weight loss) and fuel design. A reactor is run at constant power, and there­fore, as weight loss increases, the spectrum (gamma and neutron) will change and become harsher (higher neutron and g-flux).

In the graphite, charged electrons are produced because of the following:

1. Compton scattering interaction of gamma with electrons within the carbon atoms.

2. Pair product in electrostatic field associated with carbon atoms.

3. Photoelectric absorption.

Compton scattering predominates, but electrons and charged carbon ions are also produced because of the displacement of carbon atoms in the moderator, and in principle this could be calculated.

Energy deposition is the energy released from the first collisions of primary gamma and neutrons.25 Energy deposition is calculated in watts per gram (W/g) and the spatial distribution can be calcu­lated using reactor physics codes such as McBend
(http://www. sercoassurance. com/answers/), WIMS, and WGAM. However, a crude estimation of energy deposition can be made by assuming that ^5% of the reactor power is generated in the graphite. This heat can then be proportioned to the rest of the core using interpolation and form factors, and estimates of the distribution within a moderator brick.

In conclusion, energy deposition is required to calculate graphite temperatures and radiolytic oxida­tion rates. Energy deposition can be estimated but is most accurately calculated using reactor physics codes. However, care must be taken because the ratio between neutron heating and g-heating, or more appropriately a direct measure of the ionizing irradi­ation, is important.

4.11.6.1 The Use of Titanium for Installed Sample Holders

During the construction of the Magnox and AGR reactors, graphite specimens were placed into ‘installed sample holders,’ the intention being that these samples could be removed at a later date to give information on the condition of the graphite core. To enhance the radiolytic weight loss of the graphite in the installed sample holders, titanium was used. Although this only slightly increased the g-heating, it did increase the number of electrons produced, because of an increase in pair production and Compton scattering caused by the higher atomic number or ‘Z-value’ of titanium compared to graphite (22 and 6, respectively).

Two main models for the mechanism of irradiation creep have been put forward but neither has any microstructural observations to support them. The first suggestion is that a model by Roberts and Cottrell97 for a-uranium may be appropriate. This model proposes that the graphite crystallite struc­tures will yield and shear because of the generation

8 10 12 14 16 18 20

Fluence (GWdte 1)

Figure 66 The effect of stress and irradiation on the interlayer spacing of graphite. Modified from Francis, E. L. Progress Report for the JNPC-Materials Working Party: Graphite Physics Study Group; UKAEA, TRG-M-2854 (AB 7/17604); 1965.

the need for present plants to extrapolate beyond current data and to predict the behavior of new graphite grades operating for longer lifetimes at higher temperatures than before means there is still a substantial amount of work for the graphite special­ist. Future understanding and validation of property/ microstructural change relationships that enable the prediction and interpolation of existing databases and the development of new graphite grades is now pos­sible using new characterization, modeling, and com­putation techniques. These allow the investigation of mechanisms and graphite behavior that were previ­ously impossible or impractical to conduct. Areas of particular interest are obtaining a better understand­ing of the mechanism of dimensional change and irradiation creep, and the development of a validated graphite failure

Several authors96,112,113,119-121,123-125 have studied

the deformation mechanisms using TEM by taking thin foils out of the specimens after testing. They have observed that, as for many other irradiated metals, after testing, numerous cleared bands free of irradiation defects are present in the material (Figure 15). These cleared bands are the consequence of the dislocation channeling mechanism reviewed in detail by Hirsch,110 Wechsler,126 and Luft.127 According

to several authors,128-130 the irradiation-induced loops, which are obstacles to dislocation glide, can be overcome by dislocations when a sufficient stress is applied, the loops being subsequently annihilated or dragged by dislocations following different pos­sible mechanisms.108-110,131,132 This process of

removal of irradiation loops by moving dislocations produces a cleared zone free of defects inside the grain. These obstacle-free channels or swaths will therefore constitute preferred areas for further dis­location gliding, leading to plastic strain localiza­tion at the grain scale with regions of very high local plastic strain surrounded by regions of almost zero plastic strain. According to Williams et at}1 and Adamson et at., 9 the local plastic strain could reach up to 100% inside these bands. Some dis­agreement on the activated slip systems seems to remain in the case of zirconium alloys. Indeed, some authors have observed channels along the prismatic planes101,119 for tests performed at 250 and 327 °C on a Zircaloy-2 containing 1500 ppm oxygen, whereas

more recently other authors113,124,125 have observed

channels along the basal plane as well as along the prismatic plane depending on the loading conditions. This discrepancy could probably be explained by the differences in the texture or test temperature used by the different authors. Nevertheless, it is now clearly proved113,124 that for materials with texture characteris­tic of RXA tubing or rolled sheets, with (c) axes ori­ented in the (r, в) plane with an angle between 20° to 45° to the radial (r) direction, and for internal pressure tests or transverse tensile test performed at 350 °C, only basal channels are observed for low plastic strain level. Therefore, most of the plastic strain is believed to occur by basal slip inside the channels. However, it is shown that, for an axial tensile test, basal slip is not active because of its very poor orientation and only prismatic and maybe pyramidal channels can be observed.

The fact that the basal slip becomes the easy glide slip system at 350 °C after irradiation constitutes a major change in the deformation mechanisms since, before irradiation, for the same test temperature it is the prismatic slip system that is the easy glide slip system. This change in the deformation mechanisms can be explained by the difference in the interaction between the irradiation-induced loops and the dis­locations gliding either in the basal plane or in the prismatic plane, as pointed out previously. Indeed, the junction created between a dislocation gliding in the basal plane and a loop is always glissile, whereas it is sessile when the dislocation is gliding in the pris­matic plane. Therefore, when the dislocation glides in the basal plane and encounters a loop, the loop can be dragged along the slip plane, leading to a progres­sive clearing of the basal channel.

Since the loops are cleared by gliding dislocations inside the channels, it is usually assumed133 that within the channels a strain softening occurs. This phenomenon is believed to be the cause of the decrease of the strain-hardening rate with irradiation and thus to the early localization of the deformation at the specimen scale, explaining the dramatic decrease of the uniform elongation after irradia — tion.96,133 According to several authors,119,127 the strong texture of the rolled sheets or tubing leads to an even stronger localization of the plastic strain. Indeed, due to the texture, the (c) axis of the hcp grains is along the (r в) plane in the case of a tube. Since for internal pressure test or transverse tensile tests the channels are along the basal plane, the basal channels can easily propagate from grain to grain, as has been shown by Onimus et a/.113,124 When the entire section of the specimen is crossed by disloca­tion channels, a strong necking is observed on the specimen. As was pointed out by Franklin eta/.,134 the RXA alloys are more susceptible to the plastic insta­bility since the dislocation tangles that remain in SRA alloys are believed to inhibit the easy glide and the plastic flow localization.

As discussed by Onimus and Bechade,135 the polycrystalline nature of the material is also believed to play an important role in the overall macroscopic response of irradiated zirconium alloys after irradiation. Indeed, the intergranular stresses that develop because of strain incompatibilities between grains can balance the local microscopic softening occurring in the dislocation channels up to the UTS.

Based on various mechanical data such as Knoop hardness test136 or plane strain and plane stress tensile tests, several authors93, 2 have shown that the irradia­tion decreases the plastic anisotropy of the RXA zirconium alloys. Concerning the SRA zirconium alloys, the mechanical behavior is already more iso­tropic before irradiation than RXA zirconium alloys137 and the relative decrease of the anisotropy is therefore lower.122 According to these authors,122,136 this decrease of the anisotropy of RXA zirconium alloys is due to the fact that the basal slip is more activated after irradiation than before irradiation.

This section begins with the optimization of chemis­try and initial microstructure to develop swelling and

creep-resistant ferritic steels. The microstructural instability during service exposure is briefly pre­sented. The superior swelling performance of ferritic steels is understood based on mechanisms of void swelling suppression. Following this, the irradiation — induced/-enhanced segregation/precipitation causing irradiation hardening is discussed. The irradiation creep and embrittlement, their mechanisms and meth­ods to combat the problems are highlighted. The R&D efforts of today to reduce the severity of embrittle­ment in ferritic steels, using modeling methods, are outlined. Finally, typical problems in the weldments

of ferritic steels, when used for out of core applica­tions, are presented, emphasizing the advantage of modeling in predicting the materials’ behavior.

4.03.4.1 Influence of Composition and Microstructure on Properties of Ferritic Steels

Rapid strides have been made the world over, in the design and development of advanced creep-resistant ferritic or ferritic-martensitic steels. The low alloy steels can be used as either 100% ferrite-martensite
or a mixture of both. It is possible to choose the required structure by the appropriate choice of either the chemistry or the heat treatment. For example, a completely ferrite matrix, yielding high toughness, can be obtained in steels with chromium content higher than 12%, with carbon reduced to less than

0. 03%. The same steel can be used to provide higher strength by choosing the 100% martensite structure, if carbon content is increased to about ~0.1%. The 9Cr steels have always been used in the 100% mar­tensite state. Extensive studies have been carried out on phase stabilities of these steels, with changes in chemistry and heat treatment.

The creep resistance of the plain Cr-Mo steels has, further, been increased by the addition of carbide stabilizers like Ti or V or Nb, leading to the modified variety of 9-12Cr-Mo steels. These

elements led24 to copious, uniform precipitation of Monte Carlo (MC) type of monocarbides, which are very fine and semicoherent. Such precipitates are very efficient in pinning the mobile dislocations, lead­ing to improved creep behavior at higher tem­peratures. These carbides are stable at temperatures higher than even 1273 K and hence, do not cause deterioration of long-term mechanical properties during service exposure.

The development of high creep-rupture strength 9-12% steels with various combinations of N, Mo, W, V, Nb, Co, Cu, and Ta is based on optimizing the constitution (Table 2.) and 8-ferrite content, increasing the stability of the martensite, dislocation structure and maximizing the solid solution and pre­cipitation hardening. The concentration of each ele­ment in ferritic steels has been optimized based on an in-depth understanding of the influence of the specific element on the behavior of the steel. The extensive studies related to optimization of chemistry are summarized in Table 3. Based on the strong scientific insights, large number of commercial steels have been developed (Table 4) in the later half of the last century.

Most of this family of ferritic-martensitic steels is used in the normalized and tempered condition or fully annealed condition to achieve the desirable phase. The type of structure that is deliberately favored in a given steel depends on the end application.

The microstructure of the steels in normalized and tempered conditions consists24 (Figure 5) of (a) martensite laths containing dislocations with a Burgers vector 1/2a0<111> with a density of approximately 1 x 1014m~2 (b) coarse M23C6 particles located at

prior austenite and ferrite grain boundaries with finer precipitates within the laths and at martensite lath and subgrain boundaries. M2X precipitates rich in Cr are isomorphous with (CrMoWV)2CN.

The initial microstructure of the normalized and tempered steels described above does not remain stable during service in a nuclear reactor. Pro­longed exposure at high temperature causes changes in the initial microstructure, which has been studied extensively. The M2X precipitates in the normal­ized and tempered stabilized 9Cr-1Mo steels are gradually replaced (Figure 6) by MX, intermetallic, and Laves phases during prolonged aging at high temperature.

The high temperature and the irradiation over prolonged time of exposure introduce microstructural instabilities. These instabilities are caused mainly by the point defects caused by irradiation and complex coupling of these defects with atoms in the host lattice, their diffusion or segregation and finally the precipitation. There is a recovery of the defect structure since the irradiation-induced vacancies alter the dislocation dynamics. There are three types of processes with respect to evolution of secondary phases: irradiation-induced precipitation, irradiation — enhanced transformations, and the irradiation modi­fied phases. It is seen that the evolution of these phases depends on the composition and structure of the steel and the irradiation parameters like the temperature, dose rate, and the dose. Evolution of irradiation — induced phases and their influence on hardening and embrittlement is discussed later.