• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Development of Radiation Resistant Materials for Advanced Nuclear Power Plant

Posted on: Wednesday, 7 June 2006, 06:00 CDT

By Little, E A

A unique feature of the environment of nuclear power reactors is the presence of high energy neutron radiation, which can lead to degradation processes in the materials of critical components. An understanding of the operative elevated temperature radiation damage mechanisms enables these effects to be minimised by appropriate alloy selection together with compositional optimisation. An overview of these aspects relevant to advanced fission and fusion reactor systems is presented. Several key examples are highlighted and include radiation embrittlement of low alloy pressure vessel steels for pressurised water reactors, coating designs for fuel in high temperature pebble bed reactors, and selection and design of low activation materials for proposed tokomak fusion reactor devices. In addition, void swelling, non-equilibrium solute segregation and radiation induced precipitation processes in alloys at high displacement doses and relevant to several reactor systems are covered. Likely trends in future nuclear power plant design and associated materials requirements are discussed.

Keywords: Nuclear reactors, Neutron irradiation, Reactor materials. Microstructure, Mechanical properties

Introduction

Advanced nuclear power systems are being designed with the potential to make significant contributions to future energy demands in an environmentally acceptable manner. The safe, reliable and economic operation of such nuclear plant will be critically dependent on good materials performance. In all reactor systems, neutron irradiation can result in degradation of materials properties of core components and pressure boundary structures, and this is a central issue, since, with the exception of fuel elements, many components are designed for full service life. This overview highlights several important examples of developments in high temperature materials for nuclear applications aimed at minimising radiation effects in critical components. The paper covers not only current generation systems but also focuses on advanced reactor concepts and/or new reactor designs which may represent their near term and long term replacements.

The systems and components covered in detail include: (i) light water cooled thermal reactors typified by the pressurised water reactor, and for which materials development for the pressure vessel and core internals has been, and continues to be, a major research activity; (ii) high temperature gas cooled reactors based principally on the pebble bed design and which rely on the integrity and high performance of novel all ceramic fuel; and (iii) fusion reactors, of which the tokomak design based on magnetic confinement of plasma is favoured, and where major international efforts in the evaluation of radiation resistant materials for the plasma containment vessel are in progress. Pressurised water reactor technology is given prominence, since advanced systems in this category represent likely replacements to existing operating nuclear plant. Fast reactors have operated successfully in the past, but development has discontinued in many countries. However, the data base on irradiated materials behaviour from this design serves to underpin the understanding of several metallurgical processes encountered in thermal and fusion reactor systems, and therefore selective coverage is given.

Nuclear reactor fundamentals

Components in or near the cores of thermal and fast reactors are exposed to fluxes of neutrons with energies ranging from several MeV down to ~0-025 eV and ~10 keV, respectively. The neutrons are produced from fissioning (splitting apart) of the nuclei of atoms of suitable isotopes of uranium (U) or plutonium (Pu) fuel. In most thermal reactors the fuel is principally uranium enriched up to ~5% in the fissile isotope uranium-235 (^sup 235^U) and in modern systems is usually in the form of the oxide (UO^sub 2^). The reactor core must also contain a moderator (i.e. a material containing low atomic mass elements such as H or C, e.g. water or graphite) to slow the fission neutrons down to thermal energies by efficient neutron- nucleus elastic collisions. This is required to sustain a chain reaction, since U-235 is most effectively fissioned by thermal neutrons.

Natural uranium contains only ~0-7% of U-235, the remaining ~99- 3% being chiefly the relatively non-fissionable* isotope uranium- 238 (^sup 238^U). However, the latter can absorb fast neutrons and is thereby transmuted into plutonium-239 (^sup 239^Pu), which is a fissile isotope and thereby a potential nuclear fuel. Pu-239 is created to a limited extent in thermal reactors, the spent fuel from which is otherwise mainly waste U-238 with a depleted U-235 content of <~0.2%. Thus the Pu-239 can be separated out by chemical 'reprocessing', enriched (e.g. to ~15-25% Pu-239) and then used as a fuel in a fast reactor where the fission reaction is now sustained by a fast neutron spectrum - maintained by the absence of a moderator and a compact core configuration. Furthermore, a 'blanket' of U-238 can be placed around the Pu-enriched core to absorb fast neutrons and thereby create or 'breed' more Pu-239. The fast breeder reactor can thus generate more fissile fuel than it uses. Plutonium is now also used to a limited extent as fuel in thermal reactors which can be designed to 'burn' mixed oxide (MOX) fuel consisting of (U, Pu)O2 with ~5-10% Pu-239.

Natural thorium-232 (^sup 232^Th) can also be used in a nuclear fuel cycle and its application is directly analogous to that of U- 238. While non-fissionable, as a result of thermal neutron capture, it is transmuted into uranium-233 (^sup 233^U), which is fissile. Thus Th-232 either as a blanket or a component of fuel can be used to breed U-233 in a thermal reactor. The latter can then be separated out by reprocessing, enriched and used as the fuel charge in either a thermal or fast reactor. The thorium fuel cycle has been fully validated in prototype test reactor systems, but is unlikely to be used in the short term due to the abundance and success of uranium based fuel.

In proposed fusion reactor systems, neutrons are produced as a result of the fusion (joining together) of the nuclei of atoms of light elements, typically isotopes of hydrogen. The most accessible fusion reaction for near term devices is between the isotopes deuterium (^sup 2^H, designated D) and tritium (^sup 3^H, designated T), which results in the production of neutrons of energy 14.1 MeV, and to which the fusion containment structure is exposed.

Radiation damage in metallic reactor materials

Neutrons produce radiation damage in metals and alloys by two principal mechanisms:1,2

1. Direct displacement of atoms: An energy transfer of only ~25 eV is required to displace an atom from its stable lattice position. A typical fast neutron-atom collision creates an energetic recoiling atom ( the primary knock-on atom or PKA) which - if of sufficient energy (i.e. above ~ 1 keV and typically tens of keV) - then goes on to displace many neighbouring atoms from their equilibrium lattice sites, generating vacancy and self-interstitial point defects in equal numbers. Many point defects are thus created from each single neutron collision event. The spatial distribution of each damage region is markedly heterogeneous and consists of (i) a dispersion of initially single point defects and (ii) a highly localised concentration of point defects known as a 'collision cascade' at the end of the PKA track, with dimensions of the order of 10 nm or less, and in the form of a central core of vacancies with the associated self-interstitials ejected to the periphery. For such a defect configuration there is a strong inherent tendency for the cascade core to collapse spontaneously; depending on the material, this results in the creation of either a vacancy dislocation loop, or conceivably a microvoid or stacking fault tetrahedron. The surrounding shell of interstitials also forms clusters or dislocation loops both ballistically and by subsequent rapid migration and coalescence. A large fraction of the point defect population is annihilated simply by mutual recombination, while a small proportion escape the cascade volume as freely migrating defects. This classical picture of radiation damage has been supported in recent years by molecular dynamics atomistic calculations which simulate cascade formation in a computer modelled crystal lattice (see Bacon and Osetsky3 and references cited). Further insight into the subsequent evolution of cascade damage on a longer timescale has been obtained from kinetic Monte Carlo calculations (see Domain et al.4 and references cited). In addition, in some cases the defect clusters generated are resolvable by transmission electron microscopy, with features (number density, spatial distribution, geometry and size range) which are in accord with the above concepts of cascade formation (see Jenkins and Kirk5 and references cited). Nevertheless, a number of aspects relating to collision cascades still require clarification, particularly with regard to core collapse, cascade structure for high PKA energies and cascade dynamics in already damaged materials at high neutron doses.6 In metallurgically complex s\ystems such as alloyed steels, it is clear that the ultimate defect configuration affecting materials behaviour under irradiation will depend on, in addition to neutron and PKA energies, parameters such as irradiation temperature, pre-existing point and line defect populations, and point defect-dislocation and point defect-solute interactions (see review by Little1).

2. Creation of impurity atoms: Neutrons are also absorbed in nuclear reactions resulting in changes in atomic nuclei and thereby creating new elements as impurity atoms. A large number of such transmutations is possible involving both thermal and high energy neutrons, but the most important metallurgically are the (n,α) and (n,p) reactions, which generate helium and hydrogen gas, respectively. Helium is particularly important in irradiated steels and mainly arises from transmutation reactions involving nickel, boron and iron as follows: (i) thermal neutrons and nickel are involved in the two step reaction: ^sup 58^Ni(n,γ)^sup 59^Ni [arrow right] ^sup 59^Ni(n,α)^sup 56^Fe; (ii) thermal neutrons and boron are involved in the reaction: ^sup 10^B(n,α)^sup 7^Li; and (iii) fast neutrons (> 6 MeV) are involved in threshold type (n,α) reactions principally with nickel, but with significant contributions from other major alloying elements[dagger]2. Other reactions are also possible in fusion reactor environments.2,7

Since atomic displacements, as described above, rather than the neutron dose are linked to the irradiation induced changes, the damage is most meaningfully characterised in terms of the average number of times that an individual atom is displaced from its lattice site. This leads to a damage/exposure unit called 'displacements per atom' (dpa) which is generally accepted as a correlation parameter for mechanical property and microstructural changes in reactor materials irradiated in different neutron spectra. The dpa unit, calculated according to the widely accepted NRT model8 includes both the neutron energy spectrum and the total number of neutrons per unit area passing through the material, and has been incorporated into an international standard.9 Early correlations of reactor data were based on the neutron dose unit 'n cm^sup -2^ (>1 MeV)' for thermal reactors or 'n cm^sup -2^ (>0.1 MeV)' for fast reactors, where a threshold energy of 1.0 or 0.1 MeV defines a 'fast neutron' in the relevant neutron spectrum. Approximate conversion factors between the two units are: 1 10^sup 22^ n cm^sup -2^ (>0.1 MeV) [congruent with] 4.9 dpa for fast reactor core components, and 1 10^sup 19^ n cm^sup -2^ (>1 MeV) [congruent with] 0.015 dpa for thermal reactors (pressure vessels). In this publication, either dpa or n cm^sup -2^ units, as cited in the original publications, are quoted. It should be noted that the dpa unit on the NRT model still has deficiencies insofar as it does not take account of damage annihilation during cascade formation; thus within a cascade the fraction of surviving defects is strongly dependent on PKA energy and can be of the order of 20-30% of the NRT value, and also depends on irradiation temperature. A review of reactor dosimetry, covering further aspects including codes and calculational procedures, is given by Little and coworkers.10

Neutron damage in reactor structural materials can be evaluated by testing or examination of samples machined (in radiation shielded facilities) from spent or failed components, or by using standard or minature test pieces placed in rigs and irradiated in or close to the cores of materials test reactors. Fast neutrons travel large distances through most structural materials (tens of cm) and within the reactor core this results in uniform radiation damage throughout the component or specimen; materials peripheral to the core experience reduced neutron fluxes with energy/damage attenuation through their thickness and this is also the case within thick walled reactor pressure vessels. An alternative procedure is to irradiate samples with high energy ion beams in accelerators or cyclotrons, with the objective of simulating neutron radiation damage processes. In high energy heavy ion irradiations the incident particles essentially behave as PKAs, and for example, 30-40 MeV Cr, Ni or Fe ions have been used. Such ions penetrate only a few micrometres into the surface with the damage deposited in a narrow layer, but using appropriate specimen preparation techniques, this permits electron microscope characterisation of microstructural changes. Light ions of medium energy are more deeply penetrating, with the major part of the range at constant damage level; for example 3 MeV protons have a range of ~40 urn in steel and have been used to irradiate very thin mechanical property test pieces. Data interpretation must take account of the radiation damage configuration, since protons of this energy tend to undergo glancing angle collisions with lattice atoms, producing a series of isolated displacements along the trajectory. An advantage with simulation irradiations is that they produce displacement damage at highly accelerated dose rates such that many years of reactor operation can, in principle, be simulated in several hours. This permits rapid screening of many candidate alloys and/or heat treatments as well as investigations of variables such as dose, dose rate and irradiation temperature. However, since these variables interact in a complex manner in the evolution of the damage structure, ion beam simulation studies need to be evaluated with care to establish validity of the data; generally, benchmarking against reactor data and associated correlation procedures are mandatory. Most data trends covered in this presentation are based on reactor irradiations, except where indicated.

A range of microstructural and mechanical property changes occur in reactor materials as a consequence of the radiation damage processes highlighted above. These include radiation hardening and embrittlement, high temperature helium embrittlement and irradiation creep. Other important phenomena are cavity and gas bubble formation, which can result in macroscopic growth and dimensional instabilities. Furthermore the enhanced point defect population and its interaction with solutes can lead to radiation enhanced diffusion and solute segregation effects, respectively, which are responsible for a number of other important processes; these include radiation induced precipitation and precipitate dissolution, irradiation assisted stress corrosion cracking, and grain boundary weakening effects. Such a wide range of topics is beyond the scope of this paper, and therefore only selective relevant coverage is given.

In this review, a broad and generic perspective is taken, and the approach is to identify the key materials problems involving radiation for each of the principal reactor systems under consideration, and then to trace the rationale leading to optimised materials selection for radiation resistance. Ongoing developments of advanced plant are then discussed.

Materials behaviour under fast reactor irradiation

Fast breeder reactor (FBR) development was actively pursued worldwide up to about ten years ago, after which many facilities ceased operation, particularly in the UK and USA. For this reason, fast reactor design and technology are not considered in detail in this paper. It should be noted, however, that advanced FBR designs are currently being evaluated for the long term. These form part of the US-led Generation IV International Forum (GIF) in which ten countries, including the UK, are collaborating on the design of six innovative future reactor systems of various types for possible construction 20 years ahead or beyond and which include FBR designs using helium gas and liquid metals (Na and Pb) as coolants." The reasons for the renewed long term interest in FBRs, apart from electricity production, are:

(i) they can be configured to 'burn' unwanted long lived actinides in the spent fuel from nuclear power stations (i.e. transmuting them into shorter lived radionuclides) and thereby alleviating the nuclear waste burden

1 Transmission electron micrograph of a typical void distribution in type 316 austenitic steel after neutron irradiation to 60 dpa at 500C

(ii) their higher temperature operation may be of use in thermochemical hydrogen production for a perceived future 'hydrogen economy'.

A consequence of this earlier FBR research is that a large and potentially useful data base of irradiated materials behaviour exists. In particular, specific areas provide a remarkable insight into irradiation problems now encountered in reactor systems of current interest, such as light water reactors and fusion devices, and the FBR data base has contributed to the development of radiation resistant alloys for these designs. For this reason, three key irradiation processes first observed under fast reactor irradiation and now of direct relevance to other reactor systems, namely void swelling, radiation induced segregation, and radiation induced precipitation, are considered in detail.

Void swelling

The high heat fluxes encountered in the compact FBR core require very efficient heat transfer, and sodium as a liquid metal coolant has been successfully employed. In this system, core materials operate over the temperature range ~400-700C and experience a significant radiation effects problem known as void swelling, which can result in dimensional changes and distortion of core components (see review by Garner12). The effect is well established in many metals and alloys irradiated to high neutron doses (displacement doses >1-10 dpa) at elevated temperatures in the range 0.3-0.5T^sub m^ (where T^sub m^ is the absolute melting temperature). Austenitic stainless steels, for example, can exhibit volume changes of tens of per cent at high doses. The swelling is associated with the form\ation and growth of cavities (voids) throughout the microstructure, as originally revealed by Cawthorne and Fulton13 using transmission electron microscopy. As an example, Fig. 1 illustrates a typical void distribution in a type 316 austenitic steel irradiated at ~500C to ~60 dpa. In most materials, void swelling is characterised by three regimes at a given dose rate and temperature: an incubation period; a transient regime; and finally a steady state regime with monotonically increasing levels of swelling with increasing dose.

Theoretical treatments based on chemical reaction rate theory can explain the underlying physical processes and account quantitatively for the observed microstructures.14-17 Vacancies and self- interstitials - created in equal numbers by irradiation - are lost either by mutual recombination or by absorption into sinks such as dislocations. Surviving self-interstitials aggregate rapidly into dislocation loops, which expand, coalesce and finally form a dislocation network. Surviving vacancies cluster in association with gas atoms (generally helium produced by transmutation, but also surface active gases already present) to form embryonic cavities. The dislocations present in the system act as biased sinks for the preferential absorption of self-interstitials as a consequence of the differing strain fields associated with these point defects compared with vacancies. There is thus a net excess vacancy flux into neutral sinks such as the void embryos. When the latter contain a critical number of gas atoms (or, equivalently, reach a critical radius), biased driven void growth takes place, leading to breakaway or steady state swelling.

An alternative theoretical model capable of explaining aspects of void swelling has been put forward that gives a central role to collision cascades, and involves the concept of a 'production bias'.18-20 It is suggested that events immediately after cascade formation, which initially leads to localised partitioning of vacancies and self-interstitials, results in an ultimate imbalance of available vacancy versus self-interstitial point defect concentrations. It is suggested that the self-interstitials, which cluster at the cascade periphery, are partially eliminated by glide or sweeping up by dislocations, and then vacancy loops created at cascade centres evaporate, to provide an internal source of vacancies for void growth. It should be noted that recent analyses by Okita and Wolfer21 imply that the original rate theory can, to a good degree, incorporate these processes, provided certain diffusivity conditions are fulfilled.

Swelling resistant alloys

Considerable effort has been devoted to determining the factors controlling void swelling, in order to identify swelling-resistant materials.12,22,23 A schematic comparison of the swelling response of several categories of austenitic and ferritic alloys as a function of dose is given in Fig. 2, based on fast reactor irradiations over the temperature range 400-550C.23 Simple austenitic stainless steels of type 316 (17Cr-13Ni-2-5Mo, wt-%) or type 304 (18Cr-SNi) specification were originally reference core component alloys in prototype and demonstration FBR plant in the UK, USA and France. The early data from these reactors indicated that these austenitic steels exhibited significant void swelling problems, but that in the cold worked condition the swelling rates were significantly reduced. Austenitic stainless steels, however, were still favoured for fuel pin cladding and other core component applications since they possessed the required strength characteristics up to 650C, and therefore intensive development programmes were implemented to improve their swelling resistance, principally by compositional tailoring. These research programmes established the importance of certain solut additions in conferring swelling resistance - in particular Ti, Si and P.24 The result is that an optimised austenitic steel based on 15Cr-15Ni-1Ti with Si (0.7-0.9 wt-%) and P (0.05 wt-%) additions was developed for fuel pin cladding. Even higher levels of P have been shown to further enhance swelling resistance.25 Other alloy options were also available for cladding, and mention should be made of UK programmes which identified Nimonic PE16 nickel base alloy (~42Ni-17Cr-3Mo-1Ti- 1Al) as a low swelling, high strength candidate (Brown and co- workers26,27 and references cited). Commercial ferritic-martensitic steels based on the 9-12%Cr composition exhibit the highest swelling resistance, and this low swelling response appears to be a generic property of ferritic alloys as a class,23,28,29 as indicated in Fig. 2. Such alloys therefore appeared ideal for FBR applications, but their reduced strength above ~525C restricted their use to certain low stressed components such as subassembly wrappers (ducts), used to support clusters of fuel pins. To circumvent this limitation, programmes were implemented to explore the potential of fully ferritic oxide dispersion strengthened (ODS) variants, which possess good strength properties up to 650C. The alloys selected in UK programmes contain ceramic yttria dispersoids mechanically alloyed within a ~13%Cr matrix, and creep properties,30,31 recrystallisation,32,33 and irradiation behaviour34,35 were investigated. A key aspect of the research is that several of the above options, with certain modifications, now form the starting point for selection of candidate radiation resistant structural alloys in proposed fusion reactor designs, as described later.

2 Schematic comparison of Irradiation induced void swelling in austenitic and ferritic alloys as a function of displacement dose

Swelling suppression mechanisms

Considering first the austenitic steels, explanations for the role of solutes on swelling resistance rely on the concept of the critical size of the void embryo, as embodied in rate theory.36 Cavities containing less than the critical number of helium atoms grow slowly depending on the helium atom arrival rate while below the critical size, but those containing more than the critical number of gas atoms will grow at a rate governed by the influx of excess vacancies. It follows that if helium gas atoms can be dispersed widely such that the gas accumulation rate in any given embryo is very slow thereby delaying the acquisition of the critical number of gas atoms in that cavity for bias driven growth - then swelling can be reduced. This essentially prolongs the incubation period. Fine precipitates have been found to act as efficient traps for helium atoms, and in the various alloys described previously, distributions of TiC carbides, FeTiP phosphides and γ'-(Ni^sub 3^(Si,Ti)) formed during irradiation, act in this role.12,24 Clearly initial precipitate distributions arising from design heat treatment schedules also play an important part.

Turning next to the generic high swelling resistance of ferritic steels, there are significant mechanistic differences compared with austenitic alloys, and a variety of processes has been suggested (see reviews by Little23,29,37,38 and references cited). Interstitial solutes (C and N) with tetragonal strain fields in bcc iron interact strongly with both vacancy and self-interstitial point defects (trapping) and dislocations (Cottrell atmosphere formation) and furthermore can act in these roles conjointly with substitutional solutes in alloy steels. These processes enhance point defect recombination and reduce dislocation bias, respectively, to limit void nucleation and growth. In addition, the lath boundary structures in 10-12%Cr martensitic steels have been shown to act as significant point defect sinks with similar effects. Finally, the rather unique behaviour of interstitial dislocation loops in bcc iron is postulated to suppress swelling, based on the observation of loops of predominantly a<100> Burgers vector after fast reactor irradiation.23,39,40 Loops of either a<100> or a/ 2<111> geometry can form by shear from a common a/2<110> faulted nucleus by the Eyre-Bullough mechanism.41 Swelling suppression can originate as follows. The a<100> loops are not energetically favoured by this shear process, but act as stronger biased sinks relative to the a/2<111) type. The latter are forced to act as net vacancy sinks, and continually shrink after formation, thereby depressing the vacancy supersaturation and inhibiting void nucleation. Figure 3 illustrates dislocation loops of the two geometries, and the absence of voids, in fast reactor irradiated 12%Cr steel in support of this mechanism.23 Recently molecular dynamics (MD) computer simulations have predicted that a<100> loops can form by junction interaction of two geometrically favoured a/ 2<111> loops, in a variation of the original Eyre-Bullough mechanism.42 The role of dislocation loop geometry on radiation damage evolution in α-iron and ferritic steels is thus still being actively researched. In terms of experimental fast reactor data, the overall body of evidence suggests that in ferriticmartensitic steels, incubation doses approach or exceed ~100 dpa, with subsequent low void growth rate. However, an important point to make in terms of potential applications of these steels to fusion reactors is that their low swelling response needs to be reaffirmed after 14 MeV neutron irradiation to high doses, in case the operative swelling resistance mechanisms break down under conditions of high helium gas generation.43

Radiation induced segregation

Interactions between solute atoms and point defects can result in the coupled transport of solute atoms by point defect fluxes, giving rise to the non-equilibrium process of radiation induced segregation (RIS). The process is fundamentally different from thermal segregation insofar as it is driven by high fluxes of freely migrating radiation induced vacancies and self-interstitials generated at high displacement dose rates, theconcentrations of which can reach orders of magnitude greater than thermal equilibrium values. This type of segregation is now recognised as important in determining microstructural evolution in many alloy systems during elevated temperature neutron irradiation, and is a dominant irradiation phenomenon under fast reactor conditions; it is also relevant to irradiation assisted corrosion and other processes in PWR and BWR core internals - see later sections. Several reviews of RIS are available in the literature.44-46

3 Transmission electron micrograph showing dislocation loops near lath boundary (lb) in a 12%Cr martensitic steel after neutron irradiation to 46 dpa at 465 C. Large loops A and B are a<100> type with characteristic rectilinear geometry; small loops C are a/ 2<111> type with characteristic double arc images. The <100> directions marked are projections in the (111) plane

The direction of flow of the solute - i.e. either towards or away from typical point defect sinks, which can be interfaces, grain boundaries, cavities, etc. - will depend on the magnitude of the solute-point defect binding energy. In general, undersize solutes (e.g. Si or P in α-iron) bind strongly to self-interstitials in a mixed dumbbell configuration leading to marked enrichment at the sink. In contrast, oversize solutes (e.g. Cr, Mo in α-iron) exhibit weak binding to vacancies; this gives rise to solute depletion at the sink and corresponding enrichment in the matrix since the preferential exchange of solute atoms with vacancies moving towards the sink results in a flow of solute in the opposite direction. These simple generalisations do not explain all data trends.29 For example, Cr has been observed to locally enrich at interfaces but deplete in the adjacent matrix in 1012%Cr steels,47,48 which can be explained in terms of superimposed but competing thermal and radiation induced segregation; alternatively, such reversed direction of flow of oversize solutes can be due to cosegregation effects involving interstitial elements and/or linked flow with undersize species.48 Likewise, the RIS of Ni which generally enriches at sinks initially appears anomalous since it is marginally oversized in bcc iron; however, data suggests a high Ni- self-interstitial binding energy of ~1.0 eV.

These considerations apply essentially to dilute alloys (<1 at.- % solute) where the concept of an isolated bound defect-solute pair is valid. In concentrated solid solutions, alternative formulations such as the inverse Kirkendall effect49,50 are shown to predict parallel behavioural trends. Another aspect of the phenomenon is that as a function of temperature, segregation peaks at intermediate rather than high temperatures; reduced point defect mobility and recombination inhibits the process at low temperatures, while at high temperatures vacancy supersaturations diminish as their concentrations approach thermal equilibrium values. Theoretical treatments for RIS are available based either on rate theory51,52 or simplified analytical approaches.46,53,54 It should be emphasised that the accuracy of all RIS calculations depends on precise knowledge of binding and migration energies of solute-defect complexes, and in many cases this is not available.55

Experimental data exists demonstrating RIS in a wide range of austenitic steels and nickel base alloys12,56 and more recently in ferritic-martensitic steels with 9-13%Cr.29 The now routine availability of high spatial resolution field emission gun scanning transmission electron microscopy (FEGSTEM) has permitted accurate determinations of RIS solute concentration profiles at interfaces. As an example, Fig. 4 illustrates typical profiles for Cr, Ni, Si ( and Fe) on either side of a martensite lath boundary in a 12%Cr martensitic steel following irradiation at 465C to 46 dpa.47 The depletion of Cr and enrichment of Si, and particularly Ni, at the boundary are clearly evident.

Effects of radiation on phase stability

Elevated temperature irradiation to displacement doses typical of fast reactor conditions has significant effects on precipitate evolution in many alloy systems. The effects subdivide into three general categories:

(i) modification of phases present before irradiation, i.e. acceleration and/or retardation of dissolution processes and/or compositional changes to the array of thermally induced precipitate types created during initial preservice heat treatment, and/or equivalent effects during elevated temperature reactor service compared with purely unirradiated thermal aging

(ii) formation of new non-equilibrium (i.e. irradiation induced) precipitate types and which are not expected after thermal treatment for the temperatures, times and bulk compositions studied

(iii) bulk phase instability resulting in transformation from γ[arrow right]α to produce localised regions of matrix ferrite, or &947; giving localised regions of matrix austenite, in austenitic and ferritic-martensitic steels, respectively.

These effects may be understood in terms of basic processes leading to redistribution of alloying elements during irradiation, such as displacement cascade mixing, radiation enhanced diffusion involving both vacancy and self-interstitial point defect fluxes, and radiation induced solute segregation. In particular, the occurrence of new phases can be simply explained in terms of RIS to sinks such as grain boundaries, martensite lath boundaries and existing precipitate/matrix interfaces as described in the previous section, resulting in precipitation when local solubility limits are exceeded. Localised transformations to ferrite or austenite can likewise be rationalised in terms of build-up or depletion of chromium and nickel by RIS. Effects relating to radiation modification of phase diagrams, quantified mainly for binary solid solutions,57,58 may also be applicable.

4 Typical concentration profiles for Cr, Ni, Si and Fe on either side of a lath boundary in a 12%Cr martensitic steel after neutron irradiation to 46 dpa at 465C

There is a wealth of experimental data on precipitate evolution in fast reactor irradiated austenitic steels and nickel base alloys.12,56,59 as well as a range of ferriticmartensitic steels.12,29 Typical irradiation induced phases commonly observed in type 316 austenitic steel and other 300 series variants include nickel and silicon rich precipitates such as γ'- (Ni^sub 3^Si) and G phase. In 10-12%Cr martensitic steels Si and Ni rich diamond cubic phase (M^sub 6^X), bcc intermetallic χ phase, α' (bcc Cr rich ferrite) and both (Cr,Fe)^sub 3^P and (Cr,Fe)P type phosphides have been identified.60,61

Clearly the formation of these precipitated phases and the associated depletion of solutes from the matrix, together with the formation of irradiation defects (dislocation loops, voids and a dislocation network) will induce significant changes in mechanical properties, for example, embrittlement, hardening and/or strength loss, all of which are interlinked in a complex fashion.12,62 In addition, in ferritic alloys these changes will also include upward shifts in DBTT.

The pressurised water reactor

Light water cooled and moderated thermal reactors (LWRs) account for over 80% of all currently operating nuclear plant, as a consequence of proven and reliable technology. The designs subdivide into two closely related types; the pressurised water reactor (PWR) and the boiling water reactor (BWR). The former is generally regarded as the principal system and is thus covered in detail in this publication. The BWR differs from the PWR mainly in that the primary water is held at a lower pressure and is allowed to boil with the steam flowing directly to a turbine; however, the design of the pressure vessel and core internals, and the radiation damage problems affecting these components, are closely comparable. The technology for both reactor types is backed up by a substantial infrastructure and regulatory base in a number of countries.

LWR technology has been under continuous development since its inception, resulting in different generations of reactor systems which are conveniently grouped chronologically within a broad classification framework (which includes reactor types other than LWRs), as follows:

1. Generation I includes the original Westinghouse PWR and General Electric BWR designs developed in the 1950-1960s and their derivatives, such as the first set of French and German PWRs.

2. Generation II refers to designs currently in operation, such as the later Westinghouse systems, the French N4 reactors, the UK Sizewell 'B' reactor and others worldwide.

3. Generation III includes some advanced BWRs recently commissioned, advanced passive (AP) PWR designs now available and/ or about to be constructed, such as AP600/1000 and EPR.

4. Generation IV - as already mentioned - deals with advanced reactors essentially at the concept stage and for potential use beyond ~2020; under the aegis of GIF six systems have been identified which include a high temperature/pressure LWR (the other five are a very high temperature gas cooled reactor, a molten salt fast reactor and three other FBR types). The principal features of several of the above commercial LWR designations are briefly covered later in this review.

Reactor pressure vessel

The essential features of a modern PWR are shown schematically in Fig. 5.63,64 In terms of plant safety, the reactor pressure vessel (RPV) represents the most critical pressure boundary component. It performs a vital safety function as a barrier to fission product release. In addition, the RPV serves several operating functions: it supports and guides control rods, supports vessel internals, provides reactor coolant around the reactor core, and directs the reactor coolant flow that facilitates transfer of heat generated in the core to the steam generator. Typical RPV dimensions (e.g. those of Sizew\ell 'B') are: overall height 13.55 m, inside diameter 4.39 m, total thickness (opposite the core) 215 mm.64

5 Schematic illustration of a modern pressurised water reactor showing pressure vessel and internal components

The RPV is generally fabricated from high toughness quenched and tempered low alloy MnMoNi bainitic/ ferritic steel, typically of either A533B Class 1 (plate) or A508 Class 3 (the forging equivalent) specification (~0.2C-1.4Mn-0.5Mo-0.7Ni, wt-% - full composition given later); it should be noted that early plant utilised A302B steel, a low nickel plate variant, and some of these reactors are still in operation. A508-3 forging grade is now the preferred specification for most modern plant, in conjunction with a monoblock ring forging design to eliminate axial welds in the central region receiving the highest neutron dose. The inside of the vessel wall is clad with austenitic stainless steel (~3 mm thickness) to inhibit aqueous corrosion. Type 308 and 309 stainless steels have been generally used and are deposited by welding using an automated submerged arc process, with final stress relief being applied after the cladding process. Single layer or multilayer wire welding processes have been used in the USA, while strip cladding methods have been favoured in Europe. All vessel welds are post- weld heat treated at ~610C followed by furnace cooling to reduce residual stresses and temper any martensite in the heat affected zone.63,65

Standard PWR designs are typically of 900-1300 MWe power capacity and operate with inlet and outlet coolant temperatures of 290C and 325C, respectively, usually at a pressure of 15.5 MPa; the peak end- of-life neutron dose on the RPV opposite the core midplane (the 'beltline' region) typically reaches 2-3 10^sup 19^ n cm^sup -2^ (>1 MeV) ([congruent with]0.045 dpa) after a projected lifetime of 30-40 years. Many modern plants have the potential for life extension to 60 years, following regulatory safety assessment of RPV embrittlement levels. In addition, advanced PWRs (APWRs) - as described later - are under development based on these well established designs, but with specification lifetimes of 60 years and beyond; the operating parameters are generally comparable, including the end-of-life neutron dose, which can be limited by increasing the gap between the reactor core and the RPV.

Safety assessment methodology

The above radiation environment can produce significant changes in the mechanical properties of RPV steels.63,65 Thus, the yield stress increases, and there is an accompanying decrease in ductility. However, of key importance for RPV safety is the occurrence of radiation embrittlement, as exhibited by the change in notch toughness parameters and measured conventionally in the Charpy V-notch (C^sub v^) impact test. C^sub v^ tests reveal that ferritic steels exhibit a ductile to brittle transition temperature (DBTT) in which the energy of fracture increases with increasing test temperature on passing through the transition, followed by a region of relatively constant high fracture energy - the 'upper shelf - characterised by ductile fracture behaviour. Below the DBTT the steel behaves in a brittle manner and fracture occurs at low energies - the 'lower shelf. Radiation embrittlement is manifested as an increase in DBTT (traditionally specified at the 41 J (30 ft- lb) energy level), together with a decrease in the upper shelf energy (USE). These changes correspond, in fracture mechanics terms, to decreases in fracture toughness, implying decreased resistance to crack initiation (in the transition region) or ductile crack growth (in the upper shelf region) from pre-existing defects or postulated flaws in the RPV.

Radiation embrittlement is easily determined qualitatively from C^sub v^ tests, and the assessment of in-service degradation of RPVs is traditionally based on the testing of C^sub v^ specimens of RPV materials placed in surveillance capsules close to the inner vessel wall and removed periodically during the reactor lifetime. The location of the capsules some distance inwards from the vessel wall results in an accelerated dose rate condition, giving a 'lead factor' in neutron dose on the surveillance samples compared with the actual vessel. The test data therefore provide a forward prediction of future ongoing radiation embrittlement of the RPV - (for an overview of reactor surveillance methodology, see Little66). It is also important to note that C^sub v^ impact properties are specified in design and regulatory codes such as the ASME Boiler and Pressure Vessel Code, Section III67 and the US Federal Code 10CFR50,68 and limits specified for irradiation induced DBTT shifts; e.g. the minimum USE value for unencumbered PWR operation is 68 J (50 ft-lb). Hence, the assessment of post-irradiation C^sub v^ impact properties is currently mandatory.

However, in order to establish the influence of irradiation on RPV integrity, data is required on the magnitude of the degradation in fracture toughness (K^sub lc^) with irradiation. These measurements are rarely made on irradiated steels due to practical problems such as space limitations in reactor surveillance capsules, and the requirement for large specimen sizes for valid K^sub lc^ determination, which cannot be met. Instead, a methodology has been established in which the impact properties are used in conjunction with linear elastic fracture mechanics (LEFM) concepts, as embodied in ASME Section III.67 This code defines a curve representing the lower bound of unirradiated dynamic, crack arrest and static fracture toughness values as a function of temperature for A533B type steels, known as the reference fracture toughness (K^sub IR^) curve, in terms of the operating temperature and a 'nil ductility reference temperature', RT^sub NDT^. RT^sub NDT^ is established for the unirradiated steel using a combination of dropweight and C^sub v^ impact tests, and typically has a value around -12C. The irradiated RT^sub NDT^ value is then simply taken as the unirradiated value plus the DBTT shift at the 41 J energy level as established from C^sub v^ tests. The irradiated RT^sub NDT^ is then used to shift upwards in temperature the K^sub IR^ curve, which is assumed not to change in shape. An LEFM safety analysis of the RPV then follows standard procedures in which the stress intensity factors calculated from the various loads on the vessel when added together must be less than the K^sub IR^ value for the steel, as adjusted to account for the radiation embrittlement. This analysis is used to define the allowable pressure-temperature limits during plant heat-up or cool-down in order to maintain adequate safety margins against brittle fracture.

Recently there has been a move to extend and revise the above methodology with the introduction of the master curve (MC) concept,69 now incorporated into an ASTM standard.7 The method, essentially based on the work of Wallin,71,72 assumes that the fracture toughness-temperature curve in the transition region possesses a universal invariant shape (the MC), and this can now be indexed on the temperature scale by a reference temperature (T^sub o^) corresponding to a reference toughness level, taken as 100 MPa m0.5. This negates the ASME requirement to index from a combination of dropweight and C^sub v^ impact tests. The entire curve can then be redefined for the irradiated condition by input of the increased T^sub o^ after irradiation. The shift in T^sub o^ is now based on small scale fracture toughness tests on irradiated samples. Estimating procedures and validity tests for T^sub o^ are available.73,74 A significant advance has also been the underpinning of the invariant shape of the MC by multiscale and microstructural modelling of the cleavage fracture process.75-77

Radiation embrittlement in RPV steels

The factors controlling radiation embrittlement in low alloy steels, including the A533B and A508 grades used for PWR RPV construction, are now broadly understood in terms of observed irradiation induced microstructural changes. These effects can in turn be linked to material composition, starting microstructure and irradiation conditions, and hence the magnitude of the irradiation induced degradation in impact properties and associated fracture toughness understood as a function of these parameters. The key starting point for research in this area was the early C^sub v^ impact data trends as a function of dose, as illustrated in Fig. 6, which highlighted the importance of copper impurity as directly linked to severe 290C radiation embrittlement in older generation A533B steels and weldments; for example, the data demonstrate that severe embrittlement with DBTT shifts exceeding 130C at doses of 2 10^sup 19^ n cm^sup -2^ (>1 MeV) are possible.63,78

The accepted model for the observed hardening and embrittlement phenomena in RPV steels is that these effects are a consequence of at least two irradiation induced microstructural components:

1. A dominant aspect of the embrittlement process is the key role of minor elements. As described above, copper present as an impurity in RPV steels is the principal element of concern. Copper has a very low equilibrium solubility in α-iron, with a value <0.02 wt% at 290C.79 Hence even an alloy with 0 l%Cu is highly supersaturated at the PWR operating temperature, but with a diffusivity close to the self-diffusivity of α-iron,79 the copper is essentially immobile in the absence of irradiation. However, microstructural studies, given in detail later, demonstrate that the relatively low levels (>0.1 wt-%) of copper in PWR RPV steels will precipitate during neutron irradiation at 290C. The precipitation process occurs essentially by enhanced diffusion mechanisms driven by the irradiation induced vacancy flux, leading to fine copper precipitates, typically ~2 nm in diameter, which are nucleated and retai\ned as a coherent bcc phase and act as potent dislocation obstacles to give a strong hardening response.63 It is further noted that lower Cu levels in the range 0.03-0.10 wt-% still contribute to residual levels of radiation embrittlement.80

6 Measured irradiation embrittlement of early A533B pressure vessel steels and weldments showing dependence on copper content

2. Clusters formed directly by coalescence of point defects generated during irradiation induce a second component of irradiation strengthening termed 'matrix damage'. The clusters are considered to be either dislocation loops or microvoids, formed from the remnants of cascade collapse and stabilised by interaction with interstitial elements, particularly free nitrogen in solution.81 Matrix damage develops continuously with irradiation, resulting in a hardening component which is proportional to the half-power of the neutron dose. A characteristic of this damage component is that it becomes more dominant with decreasing irradiation temperature. The detailed mechanisms of point defect cluster hardening are complex.82

A significant data base exists on several aspects of the above copper precipitation model for radiation embrittlement, which may be summarised as follows:

1. All copper is removed from solid solution in the form of bcc rather than classical fee copper precipitates under irradiation and there is little or no subsequent overaging and growth of the copper precipitates; this results in a monotonie increase in radiation hardening and DBTT shift with increasing dose, followed by saturation in this embrittlement component.

2. It is believed that the copper precipitates interact with dislocations to give strengthening by a modulus hardening reaction (but see later discussion); this enables the strengthening component to be modelled theoretically and predictions to be made.

3. High nickel variants such as A533B and A508 weldments exhibit enhanced radiation embrittlement; under these conditions experimental evidence confirms that the nickel is alloyed into the copper precipitates leading to a higher precipitate volume fraction. The latter can partly explain the enhanced radiation hardening, but it is noted that the modulus hardening reaction itself may also be increased for nickel alloyed precipitates. Some studies also indicate that the copper precipitates are alloyed with manganese which can similarly influence the volume fraction of hardening centres and hence account for a further compositional effect on radiation hardening (but see later discussion on MnNi rich phases).

4. The initial distribution of copper is also important for subsequent precipitation under irradiation. Specifically, thermal pre-precipitation of copper in the form of large fee precipitates by stress relief of weldments reduces the level of copper remaining in solid solution and hence the magnitude of the radiation embrittlement. Similarly it is noted that some copper is removed from solution in the form of stable copper sulphide (Cu^sub 1.8^S) during welding, thereby again decreasing the net availability of copper to participate in radiation-induced precipitation.

7 Schematic diagram of predicted dose dependence of irradiation embrittlement in pressure vessel steels: (a) effect of copper content and dose rate; (b) subdivision into copper precipitation and matrix hardening components

5. The parametric variation of the copper precipitation process is reasonably well established.83 The rate of copper precipitation as a function of dose increases with increasing copper content as does the ultimate saturation value of hardening and DBTT shift, as illustrated in Fig. 7a. The rate of copper precipitation and related DBTT shift as a function of dose increases with both decrease in dose rate and decrease in irradiation temperature; however, the saturation value of the DBTT shift is controlled only by the copper content. The total DBTT shift as a function of dose including the matrix component is then as shown in Fig. 7b.

The above data trends represent a distillation of results of many investigations utilising a range of advanced microstructural characterisation techniques, including transmission electron microscopy (TEM), atom probe field ion microscopy (APFIM), positron annihilation (PA) and small angle neutron scattering (SANS) (for an overview of these techniques as applied to irradiated RPV steels, see Little and Gage85). These studies have revealed that the details of the copper precipitation process are in fact quite complex. A significant understanding has been made by Jenkins and coworkers86- 90 using high resolution TEM to compare thermal aging with electron and neutron irradiation in model Fe-Cu alloys. In 550C thermally aged Fe-1.3 wt-%Cu model alloy, the structures of Cu rich precipitates follow a complicated bcc [arrow right] 9R [arrow right] 3R [arrow right] fcc sequence with increasing aging time. Precipitates initially nucleate with a bcc structure, coherent with the ferrite matrix. Above a critical size ( ~12 nm) they lose coherency and transform martensitically to a twinned 9R structure. 9R is a close packed structure with ABC/ BCA/CAB stacking sequence, equivalent to fcc with regular stacking faults on every third close packed plane. This is followed by transformation to a 3R structure, which is a distorted fee structure obtained by removal of the regular stacking faults on every third (009)^sub 9R^ close packed plane. Further precipitate coarsening then results in the final stable fee structure.

Precipitate evolution under neutron irradiation differs significantly from the above.91 In simple Fe-Cu model alloys precipitates do not appear to coarsen above sizes of ~4 nm,92 while in RPV steels, as previously stated, the limiting size is ~2 nm.93 The differences compared with thermal aging are explained in terms of the role of collision cascades. Thus cascade induced precipitate dissolution may inhibit coarsening;94 alternatively, the point defect rich cascade centres may act as nucleation sites for Cu clusters but retain them in a loose form. Such small precipitates generally remain bcc, but there is some data to indicate transformation to a 9R structure of 2-4 nm precipitates in 270C neutron irradiated Fe1.3 wt-%Cu alloy.92 In irradiated RPV steels the precipitates can evolve with considerable compositional complexity; for example, APFIM has revealed a high density of Cu, Ni, Mn, Si, P enriched particles in a severely embrittled test weld irradiated to 8 10^sup 19^ n cm^sup -2^ (> 1 MeV) at 288C and there was also evidence of P segregation to the surfaces of Cu rich precipitates.95

A recent important development in understanding precipitate evolution is the prediction of irradiation induced MnNi rich phases in RPV steels containing high (1-1.5 wt-%) Ni levels such as weldments; these can be copper rich, but more significantly can be copper catalysed even in low Cu (~0.05 wt-%) material (see Odette and Lucas96 and references cited). The predictions are based partly on thermodynamic calculations, which reveal a new low temperature phase field present at ~290C in the Mn-Ni rich corner of the Fe-Cu- Mn-Ni phase diagram, coupled with the experimental evidence for Mn and Ni incorporation into Cu precipitates under irradiation, as detailed previously. While such MnNi type precipitates are equilibrium phases, the implication is that nucleation is controlled by clustering of copper, since only the latter element is supersaturated. A consequence is that their formation should be delayed at low Cu levels, giving essentially a new source of radiation embrittlement which in low Cu RPV steels emerges only at higher doses. This phenomenon - termed Mate blooming phase formation' - clearly has important practical consequences for long lifetime applications such as plant life extension or APWR construction, and full experimental validation is clearly required.

Other precipitation processes under irradiation have been reported, notably intragranular phosphide formation; this effect appears to occur when phosphorus levels are high but copper levels are low and again is accompanied by a hardening response.97,98 It is also worth mentioning that alloy carbide formation has been observed at high doses in Russian type VVER reactor steels and linked to the presence of chromium and vanadium.65 In A508 type MnMoNi steels, this could imply the possibility of redistribution of existing carbides within the steels at high doses, and such effects and related mechanical property changes again may need to be considered and evaluated for extended reactor lifetime applications.

In addition to matrix damage and copper precipitates, which both result in obstacles to the free movement of dislocations leading to strengthening by 'dispersed barriers', consideration also needs to be given to grain boundary embrittlement phenomena. It is well established that interfacial segregation of phosphorus to grain boundaries occurs in low alloy ferritic steels during thermal aging in the range 350-550C and results in intergranular embrittlement and an upward shift in DBTT, but generally without inducing matrix hardening. This is the well known phenomenon of temper embrittlement. A number of new aspects of phosphorus segregation in pressure vessel steels, including correlations with intergranular area fraction, microscopic fracture stress and fracture toughness, have recently been presented by Knott and coworkers.99-101 In irradiated RPV steels a similar segregation effect can be postulated to occur, but at lower temperatures, and based on two mechanisms:

(i) irradiation assisted phosphorus segregation can occur at 290C in which phosphorus, as an undersized solute, couples with the self- interstitial point defect flux and migrates to interfaces, namely non-equilibrium segregation

(ii) phosphorus segregation may take place by the normal vacancy controlled mechanism, but a\ugmented by the irradiation induced vacancy flux acting over longer available timescales.

Fractographic and microchemical studies on coarse grained P doped A533B steel after accelerated test reactor irradiation at 290C provides some evidence for irradiation induced P segregation102 and intergranular fracture coupled with P segregation has been observed in irradiated mild steel welds.103 Otherwise data on such processes is limited, and moreover they have not been identified as problems in current generation PWR plant. Nevertheless, the potential for P segregation is likely to increase for situations where the vessel lifetime is increased to 60 years, as in the case for APWR systems and plant life extension scenarios; clearly also, quite low solute concentrations may be sufficient to induce measurable DBTT shifts on the significantly extended timescales involved.

Embrittlement correlations

Models have been developed for radiation embrittlement which include the copper precipitation and matrix strengthening components; the two widely accepted theoretical treatments independently developed by Fisher et al.104 and Odette and coworkers105'106 enable changes in yield stress and accompanying DBTT shifts to be predicted in terms of copper content, irradiation temperature, neutron dose and dose rate. The schematic trends indicated in Fig. 7 essentially derive from these models. The derivation of the matrix damage component differs in the two models and, for example, in the Fisher et al. treatment is obtained empirically from accelerated irradiation data in the literature.107 The copper hardening component is modelled by reference to the thermal aging behaviour of copper in α iron and ferritic steels as a function of time, temperature and copper in solid solution; it is calculated from the peak hardening, time to peak hardening and time dependence of hardening derived from empirical fits to the thermal aging data. The approach assumes that irradiation changes the rate of precipitation but not the kinetics, and thus the time to peak hardening is dependent on the vacancy supersaturation during irradiation, as calculated from rate theory. The magnitude of the hardening by copper precipitates is derived from the Russell-Brown modulus hardening theory,108 in which strengthening arises as a consequence of the lower elastic modulus of the Cu compared to the Fe matrix. The dislocation energy per unit length in Cu is therefore less than in Fe and energy is required to pull the dislocation free from the precipitate, giving the observed hardening.

The above considerations strictly apply to strengthening by stable fcc Cu precipitates. Atomistic (MD) modelling has been carried out by Bacon and coworkers109-112 to theoretically examine the case of passage of a dislocation through a bcc Cu precipitate. The calculations predict that a dislocation could trigger transformation of the precipitate towards the fee structure with a resultant decrease in energy, implying a strong obstacle effect and a type of 'transformation hardening'. Detailed calculations with edge dislocations indicate both a precipitate size and a deformation temperature effect, with small precipitates up to 2 nm diameter still undergoing shear without transformation and thus remaining bcc.112

Standard procedures have been published for estimating DBTT shifts in RPV steels for licensing purposes; these procedures are issued by the regulatory bodies in several countries and are broadly similar, as documented by Grard.113 The most well established are the US Nuclear Regulatory Commission RG 1-99 revision 2 guidelines114 which give DBTT shifts in terms of composition and neutron dose. Specifically in RG 1-99 rev. 2 the DBTT shift is predicted from the copper and nickel contents for either plate/ forging material or weld metal. These guidelines were developed for early US RPV steels with relatively high copper contents and tend to underpredict DBTT shifts in modern low copper steels. In low copper steels, as previously stated, an influence of phosphorus on DBTT shift has been noted; in these circumstances the use of the earlier RG 1.99 rev. 1 version which includes a phosphorus term, or the French FIS (Fragilisation par Irradiation Suprieure) formula115 may be more appropriate. It is also worth noting that RG 1-99 rev. 2 specifies that RT^sub NDT^ at the quarter-thickness position in the vessel wall at end-of-life should be less than 93C (200F) for RPV beltline materials of new plant.

Finally the correlation between irradiation induced DBTT shift and change in fracture toughness needs to be considered. As previously stated, the lower bound fracture toughness versus temperature curve for the RPV steel is defined in the ASME code by the K^sub IR^ curve when indexed to the experimentally determined unirradiated RT^sub NDT^ of the material; irradiation is then simply taken into account by increasing the value of RT^sub NDT^ by the 41 J DBTT shift as determined from C^sub v^ impact tests. Implicit in this methodology is the assumption that changes in fracture toughness are equal to shifts in DBTT. Early data indicated clearly that the irradiation induced changes in fracture toughness are not always equal to changes in C^sub v^ DBTT shift.116 Furthermore, recent comparisons deduced from analysis of a large published data base concluded that while fracture toughness shifts are comparable for weld metals (albeit with relatively large scatter bands), for base materials the shifts are some 16% greater than DBTT shifts,117 i.e. C^sub v^ data can be non-conservative. The recently introduced master curve method described earlier will - if fully adopted - circumvent this problem area.

Inner vessel wall cladding and RPV integrity

In the above discussion, emphasis has centred on the influence of radiation embrittlement of the ferritic RPV steel and its role in reactor integrity assessment. In recent years, attention has also focused on the behaviour of the type 308/309 stainless steel cladding applied to the inner surface of the RPV, both in terms of its role in inhibiting crack propagat

Source: Materials Science and Technology; MST

More News in this Category