Journal of Nuclear Materials 356 (2006) 27–49

www.elsevier.com/locate/jnucmat

On the effects of irradiation and helium on the yieldstress changes and hardening and non-hardening

embrittlement of �8Cr tempered martensitic steels:Compilation and analysis of existing data

Takuya Yamamoto a, G. Robert Odette a,*, Hirotatsu Kishimoto a,Jan-Willem Rensman b, Pifeng Miao a

a Department of Mechanical Engineering, University of California, Santa Barbara, CA 93106, USAb NRG, P.O. Box 25, 1755 ZG Petten, The Netherlands

Abstract

Data on irradiation hardening and embrittlement of 8–10Cr normalized and tempered martensitic steel (TMS) alloyshas been compiled from the literature, including results from neutron, spallation proton (SP) and He-ion (HI) irradiations.Limitations of this database are briefly described. Simple, phenomenological–empirical fitting models were used to assessthe dose (displacement-per-atom, dpa), irradiation temperature (Ti) and test temperature (Tt) dependence of yield stresschanges (Dry), as well as the corresponding dependence of sub-sized Charpy V-notch impact test transition temperatureshifts (DTc). The Dry are generally similar for SP and neutron irradiations, with very high and low helium to dpa ratios,respectively. Further, the Dry trends were found to be remarkably consistent with the Ti and dpa hardening-dependence oflow alloy steels irradiated at much lower doses. The similar Ti and (low) dose dependence of Dry and DTc, as well as ananalysis of paired DTc–Dry datasets, show that embrittlement is typically dominated by a hardening mechanism belowabout 400 �C. However, the corresponding hardening-Charpy shift coefficient, Cc = DTc/Dry � 0.38 ± 0.18 �C/MPa islower than that for the fracture toughness reference temperature, T0, with DT0/Dry � 0.58 ± 0.1 �C/MPa, indicating thatsub-sized Charpy tests provide non-conservative estimates of embrittlement. The Cc increases at Ti > 400 �C, and DTc > 0are sometimes observed in association with Dry 6 0, indicative of a non-hardening embrittlement (NHE) contribution.Analysis of limited data on embrittlement due to thermal aging supports this conclusion, and we hypothesize that theNHE regime may be shifted to lower temperatures by radiation enhanced diffusion. Possible effects of helium on embrit-tlement for Ti between 300 and 400 �C are also assessed based on observed trends in Cc. The available data is limited, scat-tered, and potentially confounded. However, collectively the database suggests that there is a minimal NHE due to heliumup to several hundred appm. However, a contribution of helium to NHE appears to emerge at higher helium concentra-tions, estimated to be more than 400–600 appm. This is accompanied by a transition from transgranular cleavage (TGC) to

0022-3115/$ - see front matter � 2006 Elsevier B.V. All rights reserved.

doi:10.1016/j.jnucmat.2006.05.041

* Corresponding author. Tel.: +1 805 893 3525; fax: +1 805 893 8651.E-mail address: odette@engineering.ucsb.edu (G.R. Odette).

28 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

intergranular fracture (IGF). IGF generally occurs only at high Dry. Synergistic combinations of large Dry and severeNHE, due to helium weakening of grain boundaries, could lead to very large transition temperature shifts in first walland blanket structures at fusion spectrum dose levels above 50–75 dpa and in SP irradiations at much lower doses.� 2006 Elsevier B.V. All rights reserved.

1. Introduction

An important objective for the fusion materialscommunity is to develop a high quality databaseon the effects of irradiation on the constitutive andfracture properties of 8–10Cr normalized andtempered martensitic steel (TMS) alloys. This paperfocuses on changes in the tensile yield stress (Dry)and irradiation embrittlement, characterized bytransition temperature shifts (DTc), measured insub-sized Charpy V-notch impact tests. A high qual-ity database is needed to develop predictive modelsof the changes in Dry and DTc (and other proper-ties) as a function of the combination of all signifi-cant metallurgical and irradiation variables. Themetallurgical variables include the start-of-life alloycomposition (wt%), microchemistry and microstruc-ture, including the effects of thermo-mechanicalprocessing treatment (TMT). The primary irradia-tion variables include irradiation temperature (Ti)and the neutron flux (/), energy spectrum [/(E)]and fluence (/t). The neutron irradiation variablesare best represented in terms of the total and ratesof production of damaging species, including dis-placements-per-atom (dpa), helium, hydrogen andsolid transmutation products (appm). Post-irradia-tion testing and data analysis variables are also sig-nificant. For example, Dry depends on the testtemperature (Tt). Ultimately a comprehensive andhigh quality database will be analyzed with physi-cally based, multiscale models [1–3]. Such modelswill sequentially relate (a) the primary variables tomicrostructural evolutions; (b) the effects of theseevolutions on fundamental structure-sensitive con-stitutive and local fracture properties; (c) and theconsequences of changes in these fundamental prop-erties to more complex engineering properties, likeDTc (and corresponding shifts in the fracture tough-ness reference temperature, DT0 [1–3]). In theinterim, simpler phenomenological–empirical mod-els and physically motivated correlations will beused to analyze the growing, but imperfectdatabase.

There have been many irradiation studies aimedat contributing to such a database. Among the

most notable are the International Energy Agency(IEA) round robin project on the Japanese F82Hsteel [4,5], as well as the large program in Europecurrently focusing on various heats of the Eurofersteel [6–8]. There are also other data based onirradiations of TMS with spallation protons (SP),which generate very high levels of helium andhydrogen [9–13]. Other pertinent data includesaccelerator based ion irradiations, includingwith high energy He-ions (HI) [14,15]. In thispaper, we summarize the preliminary results ofan ongoing effort to tabulate and analyze exist-ing Dry and DTc data on TMS alloys. Oneobjective is to gain a practical, working knowledgeof what is required in a mechanical propertydatabase to make it functionally useful and of‘high quality’; and to assess the state of the exist-ing data in this context. A second objective is tocarry out a simple preliminary analysis to gaininsight on the variables that control Dry andDTc, including high levels of transmutation prod-uct helium.

The effects of high levels of helium under condi-tions of simultaneous displacement damage produc-tion and irradiation embrittlement are two of themost important issues facing the development ofTMS for fusion applications. Indeed, it is largelyconcerns about helium that have motivated the needfor the International Fusion Materials IrradiationFacility. Because of the high sink density that actsas traps, martensitic steels are generally believed tobe relatively immune to helium effects [16–18]. How-ever, at high helium levels, a significant populationof bubbles forms, with number densities, sizes andspatial distributions that depend on the irradiationtemperature and the alloy microstructure. Theamount and distribution of helium on grain bound-aries is likely to be particularly significant [19]. Oneschool of thought has asserted that helium plays adominant role in embrittlement [20–28], in somecases [20–22] even apparently showing a linear cor-relation between helium concentration and DTc. Acontrary view attributes a dominant role to irradia-tion hardening induced embrittlement, where thehardening is primarily associated with displacement

1 Chemical composition, normalizing–tempering temperaturesand times, hot/cold work strains; specimen size, geometry,orientation, test temperature, strain rate; irradiation source,dpa dose and dose rate, irradiation temperature, He and H;baseline versus post-irradiation tensile (ry, ru, rf, eu, et, RA),Charpy (DBTT, USE, ryd) – all with uncertainties.

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 29

damage. Thus, at least up to some concentration,helium only plays a secondary role [1,15], mainlythrough a direct or indirect effect on hardening.The sensitivity of hardening to helium generallyappears to be modest below very high levels. How-ever, to date it has not been possible to obtain reli-able experimental assessment of helium effects onfast fracture in candidate materials at damage rates,doses and temperatures relevant to the first wall andblanket environment. Thus there has been no clearexperimental resolution of this issue. Within theframework described below, we assess the issue ofhelium effects on fast fracture based on existingdata.

Our analysis addresses several specific issues.

• The dpa, Ti, Tt (for Dry) and metallurgical vari-able dependence of Dry and DTc.

• The relation between Dry and DTc to characterizehardening and non-hardening induced embrittle-ment.

• The reliability of Charpy-based DTc as a surro-gate for fracture toughness reference temperature(T0) shifts (DT0).

• The possible role of irradiation induced orenhanced non-hardening embrittlement (NHE)mechanisms on DTc.

• The possible effect of helium on fast fracture bothin terms of the local fracture mode and embrittle-ment shifts DTc.

2. Compilation of the existing irradiation hardening

and embrittlement data

Detailed descriptions of the development of acomputer accessible database will be provided inthe future. Most of the data is for neutron irradia-tions, but we also include a significant body ofresults of spallation proton (SP) irradiations, whichgenerate from �60 to 150 appm He/dpa, as well aslimited data from high energy He-ion (HI) irradia-tion studies implanting up to 5000 appm He. Cur-rently, the database consists of 834 entries for Dry

and 595 entries for DTc. However, not all of thehardening and embrittlement data entries containsufficient ancillary information to be useful in ouranalysis.

We first established a set of variables and param-eters for each Dry and DTc entry that are needed in adatabase for it to be useful. Basically, it is necessaryto have comparable baseline and irradiated data

and to know the details1 of (a) the metallurgicalvariables, including alloy composition and TMTprocessing history; (b) the irradiation variables; (c)the descriptions of the test procedures and methodsused to reduce, analyze and parameterize the data.Of course, it is also important to characterize uncer-tainties, assumptions and any possible biases inthese variables, methods and parameters. Thereare also often complications in an experiment thatare important to know, such as variations in Ti.Based on these criteria, it is necessary to concludethat the existing data generally do not yet comprisea high quality and optimally useful database. Apartial list of limitations we found is as follows:

• In some cases there was incomplete and uncertaincharacterization of the alloy and poorly definedpairs of baseline and irradiated data sets.

• Most often there was no indication of uncertain-ties in key variables, that are known to often bepoorly defined, or even biased, particularly Ti.

• Often there is no reported assessment of possibleuncertainties or bias in the test measurementsand data reduction-parameterization procedures.This is a particular issue for the DTc that oftenwere deduced from a small and, indeed, typicallyinsufficient number of specimens.

Due to such limitations, the analysis presentedbelow provides only a systematic, first-iterationassessment of the relations of Dry and DTc to dpa,Ti, Tt and metallurgical variables. However, we alsonote that some recent datasets [8,29] on Dry are ofmuch higher quality, which will be apparent in theanalysis that follows. Since many different alloyshave been irradiated under very different irradiationconditions, it is necessary to reduce the number ofindependent variables to make the analysis tracta-ble. This is discussed in Section 3.

3. Preliminary data analysis

We first carry out an analysis involving a mini-mum number of independent variables. Thus it is

30 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

necessary first to define broad alloy categories andTi bins to make the analysis tractable. The lumpeddata are naturally very scattered, but statistical fitscan still be used to establish broad, common trendsin Dry and DTc as a function of dpa, Ti and Tt (forDry). Predicted minus measured residual plots,including those for other potential variables notincluded in the primary fits, can then be used toidentify inaccuracies and biases in the first-iterationmodel. The first-iteration model can also be used toadjust the data to a common set of dpa and/or Ti

conditions. We divided the alloys into four broadcategories: (a) F82H; (b) Eurofer; (c) other 9CrWsteels containing 9 ± 1.5% Cr and 1–2% W; and(d) 9CrMo steels containing 9 ± 1.5% Cr and1 ± 0.5% Mo. Except at the lowest and highest irra-diation temperatures, the Ti binning, describedbelow, generally ranged from about ±15 to±25 �C around the nominal value. These Ti binsare on the order of the corresponding irradiationtemperature uncertainties. In general, the SP irradi-ation data have larger uncertainties in Ti. The testtemperature Tt is also a very important variable.Most often the Dry data is reported for Tt � Ti orTt � room temperature (�23 �C), and sometimesboth. The Dry generally decreases with increasingTt. Thus, we analyzed available datasets to quantifythis effect. However, since there is much more avail-able data, the Dry results presented in this sectionare for nominally Tt � Ti, except for Ti < 220 �C,where most of the data is for Tt � 23 �C. In a fewcases where only Tt 5 Ti data are available, thedata for Tt � Ti ± 50 �C were also included.

3.1. Analysis of the Dry data [7–12,14,23,29–66]

A useful general fitting expression for Dry is

Dry ¼ Drys½1� expð�dpa=dpa0Þ�p þ Dr0: ð1Þ

Here Drys is a saturation hardening, dpa0 specifiesthe dose transient prior to saturation, and p is aneffective dispersed-barrier hardening exponent. Theconstant hardening increment, Dr0, allows for either(a) a very rapid, but saturating, low dose hardeningmechanism; or (b) a threshold dose for the initiationof hardening. The saturating form of Eq. (1) can bephysically related to the depletion of solutes, in thecase of a precipitation hardening mechanism, or anexcluded volume type effect in the case of the accu-mulation of displacement damage-type defects. Theinitial rate of hardening is characterized byDrys=dpap

0.

More complex forms and detailed physical treat-ments could be used, but they must be based oncomprehensive microstructural observations andmore detailed mechanistic understanding and mod-els than are within the scope of this initial analysis.In the case of simple dispersed barrier hardeningwith one type, size and strength of hardeningfeature, and with a number density that initiallyincreases linearly with dose, p � 1/2. This is theso-called Makin and Minter model [30,31]. How-ever, p may deviate from a value of 1/2 for a widevariety of reasons. These include combinations ofincreases in the size of a fixed number of features,size-dependent feature dislocation obstacle strength,distributions in the number, sizes and strengths offeatures and superposition of the various contribu-tions to the overall strengthening. However, in viewof the limitations of the present database, and thedesire to make cross-comparisons between alloyclasses and different Ti, we sought simplicity andconsistency in the first-iteration fitting form, ratherthan additional complexity. Thus, we fixed a valueof p = 1/2 and Dr0 = 0. Note that this fitting expres-sion has the advantage that low dose data can beused to obtain estimates of the combined k =Drys/

pdpa0 fit parameter, even in the absence of

high dose data. Thus broad trends in the Drys anddpa0 can then be used in such cases to derive aself-consistent model within the range of the data.

Fig. 1(a)–(g) shows plots of Dry versusp

dpa forTi bins of (a) 25–220 �C (nominal 120 �C); (b) 250–280 �C (nominal 265 �C); (c) nominal 300 �C; (d)300–360 �C (nominal 330 �C); (e) 365–390 �C (nom-inal 380 �C); (f) 400–440 �C (nominal 420 �C); (g)and >440 �C (nominal 500 �C) [4,7–12,14,23,29,32–66]. The black solid lines are Dry = kr

pdpa fits

for the low dose data (note, kr � k) and the blacksolid-double-dashed lines are the fits for the saturat-ing form given in Eq. (1) with p = 1/2 and Dr0 = 0.In the case of neutron irradiations, different filledsymbols are used to code the alloy class with circles,diamonds and squares for F82H, 9Cr–W and 9Cr–Mo alloys, respectively. The first-iteration fits arefor irradiations with neutrons and SP (open sym-bols) and for all the alloy classes in each of the nom-inal Ti bins.

Fig. 1(a) shows the results for Ti < 220 �C. Themajority of the data is for SP irradiations of 9Cr–1Mo steels, primarily from LANSCE irradiations(open symbols) with �150 appm He/dpa, while thepartially filled diamonds show the data from SINQirradiations with �60 to 83 appm He/dpa. The

0

100

200

300

400

500

600

700

0 1 2 3 4 5 6

F82H9Cr-W9Cr-MoF82H (SP)9Cr-W (SP)9Cr-Mo (SP)F82H (HI)9Cr-Mo (HI)

Δσy

(MPa

)

√dpa

25 -220 oC

SP: : 60-83 appm/dpaother -150 appm/dpa

HI:6300appm/dpa

0

100

200

300

400

500

600

700

0 1 2 3 4 5 6

F82H9Cr-W9Cr-MoF82H (SP)9Cr-Mo (SP)F82H (HI)9Cr-Mo (HI)

Δσy

(MPa

)

√dpa

HI:6300appm/dpa

SP:50-83appm/dpa

250 - 280 oC

0

100

200

300

400

500

600

700

0 1 2 3 4 5 6

F82H9Cr-W9Cr-Mo9Cr-Mo (SP)F82H (HI)9Cr-Mo (HI)

Δσy

(MPa

)

√dpa

300 - 360 oC

SP:83 appm/dpa

d

a b

c

0

100

200

300

400

500

600

700

0 1 2 3 4 5 6

F82H(300°C)

Eurofer97 (300°C)

Δσy

(MPa

)

√dpa

300 oCLow dose RPVsteel trend

Fig. 1. Thep

dpa dependences of Dry for the temperature bins: (a) 25–220 �C (nominal 120 �C); (b) 250–280 �C (nominal 265 �C); (c)300 �C; (d) 300–360 �C (nominal 330 �C); (e) 365–390 �C (nominal 380 �C); (f) 400–440 �C (nominal 420 �C); and (g) >440 �C (nominal500 �C).

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 31

overall scatter bands for the SP and neutron irradi-ations overlap. Farrell reported that the heavydashed line shown in Fig. 1(a), which falls nearthe top of the SP irradiation scatter band, providesa good fit for both the SP and neutron data for irra-diations in the high flux isotope reactor for dosesbetween �0.05 and 1 dpa [48]. However, most ofthe LANSCE Dry data, at the highest He/dpa, tendto fall at the top of the SP irradiation scatter band inFig. 1(a), and hardening continues to a higher dpain this case. Thus this may signal the onset of aneffect of helium on Dry at very high He/dpa. Butthere are also a few LANSCE data points lying atthe lower bound of the scatter band, and the overallmean trend of both the neutron and SP irradiationhardening data appears to be saturating at Drys �

390 MPa at a low dpa0 � 0.8 dpa, albeit with verylarge uncertainty limits. Thus the effect of heliumon hardening at low Ti remains somewhat ambigu-ous. The single HI irradiation Dry data point, with ahigh He/dpa ratio �6300 appm/dpa, falls substan-tially above the scatter bands for both the SP andneutron irradiations.

In the case of Ti � 265 �C, shown in Fig. 1(b),most of the data is for neutron irradiations, exceptof a few SP data (50–83 appm He/dpa) includingthe highest dose irradiations to �9 dpa. The latterwas used to fit the saturating form for Dry. Satura-tion nominally occurs at a much higher dose, withdpa0 � 6.4, than for the Ti = 120 �C irradiations.There does not appear to be a large and systematicdifference in the Dry for the various alloy classes.

-200

-100

0

100

200

300

400

500

0 2 4 6 8 10

F82H

9Cr-W

9Cr-Mo

Δσy

(MPa

)

√dpa

365 - 390 oC

-400

-300

-200

-100

0

100

200

300

0 2 4 6 8 10

F82H

9Cr-W

9Cr-Mo

9Cr-Mo (HI)

Δσy

(MPa

)

√dpa

>450 oCHI: 6300appm/dpa

-200

-100

0

100

200

300

400

500

0 2 4 6 8 10

F82H9Cr-W9Cr-MoF82H (HI)9Cr-Mo (HI)

Δσy

(MPa

)

√dpa

400 - 440 oC

HI: 6300appm/dpa

e f

g

Fig. 1 (continued)

32 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

The nominal saturation Drys � 460 MPa, is actuallyslightly higher at 265 �C, compared to the lowestTi � 120 �C. The hardening rate for HI irradiationsis again substantially higher than for neutrons.

The data shown in Fig. 1(c) is based on a recentset of neutron irradiation Dry data reported byRensman and Lucon [8,29] for a narrow range ofTi � 300 �C. The data for various Tt have beenadjusted to a common value of 300 �C. The 9Cr–W data in this case are for various heats of Eurofer.Although limited to doses less than 10 dpa, the300 �C results constitute a very high quality datasetand can be considered a lynchpin for other parts ofthe database and analysis. While the dose does notreach saturation levels, the data are sufficient to pro-vide reasonable estimates for both Drys � 490 MPaand dpa0 � 6.8 dpa. As shown below, the 300 �CDrys is higher than at any other Ti, but both the Drys

and dpa0 are reasonably consistent with extrapola-tions of the trends at higher and lower Ti. The heavydashed line shows an extrapolation of the typicalhardening predicted for low Cu (<0.07%) low alloy(C–Mn–Si–Mo–Ni) quenched and tempered bainiticRPV steels irradiated at �290 �C to a much lowermaximum dpa of <0.06 dpa, is in remarkably goodagreement with initial Dry trend for the TMS alloysin the dose range up to several dpa [67,68].

With the exception of two spallation proton irra-diation data points (�80 appm He/dpa), the Dry atTi = 330 �C shown in Fig. 1(d) are for neutronirradiations. Note, this dataset also includes someF82H and Eurofer data at Ti = 300 �C previouslyreported by Rensman. There is an apparent trendtowards saturation at a higher dpa0 � 7.7 dpa thanat both Ti = 265 �C and 300 �C. In general theDry data for the 9Cr–1Mo alloys fall above the

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 33

hardening in the F82H and 9Cr–W alloys, with theSP irradiation data lying near the lower bound ofthe scatter band for this steel class. In this case,the nominal saturation Drys � 420 MPa is lowerthan that for both the 265 and 300 �C irradiations.

As shown in Fig. 1(e) and (f), the data is muchsparser at Ti P 380 �C, but, in this case, extend toa significantly higher dpa for the irradiations whichwere carried out in fast reactors. The Drys =119 MPa at Ti = 380 �C and =57 MPa at 420 �Care much less than for the lower Ti. There is nolow dose data in the transition in these cases to eval-uate dpa0. Thus we used extrapolations of the trendin Drys/

pdpa0 from lower Ti to estimate dpa0 in this

case. The Dry for the 9Cr–1Mo alloys are againhigher than for the other steels (note, there is noF82H data at Ti = 380 �C). There is a tendencytowards modest softening (Dry < 0) at Ti > 440 �C.Much higher hardening is observed for the HI irra-diations at both Ti = 420 and >440 �C.

Clearly, the reduced variable datasets are highlyscattered. However, this is not surprising since, inaddition to lumping various alloys within a broadclass, there are a variety of sources of uncertaintyand potential bias in the data and a range of Ti

and Tt, within the various data groupings. However,in spite of the scatter, the fits can be used to estab-lish broad trends in the database. Cross-plots of thefit parameters as functions of Ti are shown in Fig. 2.Fig. 2(a) plots the initial hardening rate coefficientk = Drys/

pdpa0, the Dry at 2 dpa and the satura-

-100

0

100

200

300

400

500

600

100 200 300 400 500

kΔσy @ 2dpa

Δσys

k(

MPa

/√dp

a),

Δσys

,Δσ y@

2dpa

(MPa

)

T ( C)

a

o

Fig. 2. (a) The Ti dependence of the initial hardening rate, k = Dry

hardening, Drys, evaluated from the fits to the data shown in Fig. 1(a)–normalized at 300 �C are also shown as the dotted-dash lines; (b) Ttemperatures (open symbols) were estimated by combining a linear fit

tion hardening, Drys. The Ti-dependence of theDry parameters decreases with increasing dosebelow Ti � 300 �C, but remains strong at higher Ti

even for Drys. The dotted-dash lines show a temper-ature dependence predicted by a relation

DryðT Þ ¼ fTðT ÞDryðT rÞ: ð2ÞHere fT(T) is a function independently derived byJones and Williams for very low dose irradiationhardening of both C–Mn ferritic and RPV bainiticsteels [69]. The reference temperature, Tr, atfT(Tr) = 1 is taken as 300 �C. The agreement be-tween the predictions based on data fit for verylow dose ferritic and bainitic steel irradiations andthe trend for much higher dose Dry(Ti) data forTMS alloys is striking, and lends considerable con-fidence to the results of the analysis shown inFig. 2(a). Fig. 2(b) shows the saturation doseparameter, dpa0, increases systematically withincreasing Ti up to �330 �C, and then apparently de-creases at higher Ti. Due to the paucity and scatterin the data, however, the estimates of Drys (at lowerTi) and dpa0 (at higher Ti), are very uncertain, indi-cating the need for additional data. Nevertheless theaverage trends are systematic and reasonably self-consistent compared to interpolations at Ti bothabove and below 350 �C.

As noted above, Rensman has suggested that theeffect of test temperature can also be treated by adose-independent adjustment factor, g(Tt) =Dry(Tt)/Dry(25 �C) that he derived from his

b

0

2

4

6

8

10

12

100 200 300 400 500

dpa 0

T (oC)

s/p

dpa0 (MPa/p

dpa), hardening at 2 dpa, and the saturation(g). The corresponding fT functions from Jones and Williams [69]he Ti dependence of the dpa0. The values for the three higherof the k data points shown in (a) and Drys.

0

20

40

60

80

100

120

0 50 100 150 200 250 300 350

F82H9Cr-1WVTaEurofer97JLF1Rensman's [8]

g(T t

)(%

)

Tt (ºC)

Fig. 3. The g(Tt) (solid line) from Rensman and fit to the UCSBry(Tt)/Dry(25 �C) database (dashed line).

34 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

Ti � 300 �C database [8]. This g(Tt), is shown inFig. 3 along with the corresponding g(Tt) for setsof Dry(Tt) in our database for 250 6 Ti 6 300 �C.The data are restricted to Tt 6 Ti + 50 �C to avoidexcessive damage annealing. The results scatteraround the Rensman’s g(Tt) but yield a similarmean trend shown by the dashed least squares fitline. Note that Rensman also found that g(Tt)shows a larger decrease in Dry with increasing Tt

for lower Ti � 60 �C, which is probably due toannealing effects.

Fig. 2 constitutes a first-iteration model forDry(dpa, Ti) for Ti � Tt. Fig. 4 shows model predic-

-100

0

100

200

300

400

500

600

700

0 2 4 6 8 10

200oC250oC300oC

350oC400oC450oC

500oC

Δσy

(MPa

)

√dpa

Fig. 4. Generic model predictions of Dry(dpa) for Ti = Tt from200 to 500 �C.

tions of Dry(dpa) created for Ti = Tt from 200 to500 �C using linear regression of k(Ti) and interpola-tion of Drys. The Dry(dpa,Ti) model expressionbetween the actual and nominal Ti can be used toadjust the individual data to a common set of condi-tions. The results of adjusting the Dry for the irradi-ations between Ti � 250 and 360 �C to a commonTi = Tt = 300 �C are shown in Fig. 5. While thereis considerable scatter, primarily due to uncertaintiesin Ti, clear separation of the data by the alloy groupis observed. Eq. (1) can be fit to the individual alloysubsets of data. The dpa0/rys (dpa/MPa) are7.7 ± 2.7/431 ± 53, 5.4 ± 1.7/395 ± 68, 10.2 ± 4.6/590 ± 62 and 7.2 ± 0.9/510 ± 22, for F82H, 9CrW,9CrMo and Eurofer97, respectively. The standarddeviation between the measured and predictedvalues is �59, 55, 57 and 22 MPa for F82H, 9CrW,9CrMo and Eurofer97, respectively. The limited SPirradiation Dry for F82H follows the neutron irradi-ation trends while those for 9CrMo appear to fallbelow the fitted trend, but this may be due to the factthat those specimens were irradiated �50 �C highertemperature for the last 15% of the dose [10]. Furtherrefinement of the model, in part based on analysis ofthe residuals, will be carried out in the future.

In summary, irradiation hardening can bedescribed by a saturating function of the

pdpa that

is characterized by a saturation hardening, Drys,and a saturation dose parameter, dpa0. The Drys

peaks at �490 MPa at Ti � 300 �C, with a weakdependence at lower Ti at less than �300 �C. How-ever, Drys decreases rapidly at higher temperatures,

0

100

200

300

400

500

600

700

0 1 2 3 4 5 6

F82H9Cr-W9Cr-MoEuroferF82H(SP)9Cr-Mo(SP)

Δσy

(MPa

)

√dpa

Ti = Tt = 300 ºC

Fig. 5. The Dry–p

dpa data adjusted to common Ti = Tt =300 �C along with fits for the different alloy classes.

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 35

with softening observed at Ti > 440 �C. The satura-tion dose parameter, dpa0, increases up to �330 �C,but also appears to decrease at higher Ti. The lowdose hardening trends at �300 �C in the TMS alloysare remarkably similar to those observed in muchlower dose irradiations of ferritic (C–Mn–Si) andlow alloy (C–Mn–Si–Mo–Ni) bainitic RPV steels.The SP Dry are similar to or slightly less than forneutron irradiations, at least for Ti between 250and 360 �C. The Dry also decreases with increasingTt, in a way that can be approximately accountedfor with the adjustment factor shown in Fig. 3,and can be primarily ascribed to the correspondingchange in the shear modulus. Finally, there are sys-tematic differences in the Dry trends observed forthe various alloy classes that will be assessed inmore detail in the future. Note a number of thesetrends were also identified in a neural network anal-ysis of the database.

In closing we add one important caveat to thisanalysis. Specifically, the dpa range is limited to rel-atively low doses, especially for data at Ti 6 380 �C.Thus these results should be revised as additionaldata becomes available, and extreme caution shouldbe applied in extrapolating the results to higher dpa,where high dose softening has been reported atTi � 400 �C [65], as well as general softening forhigher Ti P 430 �C [32,39,41,43,58,70,71].

3.2. Framework to assess various embrittlement

mechanisms

In this section we describe the framework to bet-ter assess several different mechanisms of irradiationembrittlement, including irradiation hardening andvarious non-hardening embrittlement (NHE) pro-cesses. It is noted that this framework has been usedto design a series of experiments that are being con-ducted as part of the US–Japan collaborative irradi-ation program in the high flux isotope reactor(HFIR).

1. ‘Brittle’ fracture occurs in bcc alloys when theelevated normal stress (rn = Mry) in front of anotch or crack tip exceeds a critical local fracturestress for transgranular cleavage (r�c) or inter-granular (r�ig) fracture over a critical microstruc-tural volume of material (V*), or rn(V*) P r�c .Here, ry is the yield stress and M is a constraintfactor that varies between about 2 and 4 depend-ing on the notch/crack geometry and the alloystrain-hardening rate [1,72,73].

2. At low irradiation temperatures (< � 375–400 �C), and in the absence of high levels ofhelium embrittlement, (DTc) is due to irradiationhardening and can be correlated with changes inyield stress (Dry), as DTc = CcDry [1,15,72,73].For Charpy impact tests. the value of the harden-ing-embrittlement coefficient Cc depends on bothDry and the elastic cleavage transition tempera-ture and upper shelf energy of a particular unir-radiated alloy [72].

3. As has been shown above, at high irradiationtemperatures (> � 440 �C) there is typicallyeither little or no hardening, or even somedegree of softening, with Dry 6 0. However,positive DTc may occur in this temperatureregime as the result of reductions in r* due toirradiation enhanced thermal aging processes[1]. Such NHE processes include precipitationor coarsening of brittle phases, grain boundarysegregation of solutes, such as phosphorous,and instabilities in the tempered martensiticsubstructure [1,24,74,75]. In 8–10Cr martensiticsteels containing significant quantities of tung-sten, a primary non-hardening embrittlementmechanism is precipitation of brittle Lavesphases on prior austenitic grain boundaries(PAGs). Purely thermal NHE occurs at�500 �C and above; but it appears that radia-tion enhanced diffusion decreases the lower legof the time–temperature NHE C-curve to about400 �C or less [1]. This is a concern, since com-binations of Dry and NHE may give rise to verylarge DTc. Significant NHE is signaled by large,or alternatively, negative (when DT > 0 andDry < 0) values of Cc and often, a transitionfrom transgranular cleavage (TGC) to intergran-ular fracture (IGF).

4. The accumulation of helium and hydrogen onPAGs may also lead to NHE. We will focus onthe potential effect of large amounts of heliumon NHE, but in some cases this may be difficultto distinguished from a corresponding highconcentration of hydrogen. The distribution ofhelium on the grain boundaries is critical. Forexample, a coarse distribution of large bubbleswould be expected to have a modest effect onDTc, while a monolayer-type film of boundarywould likely be most damaging [1].

5. As noted above, NHE mechanisms are generallyassociated with a transition from a TGC to IGlocal fracture mode. While oversimplified, aconceptual model of the competing effects TGC

36 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

versus IGF is very useful. The model posits thatthe TGC continues to be the fracture path ofleast resistance as long as it has a lower criticalstress than that for IGF. Thus, gradual weaken-ing of the PAGs by helium (and/or other mecha-nisms) would not be reflected in IGF until r�igfalls below r�c , where DTc depends synergisticallyon both r�ig � r�c (<0) and Dry. Of course in real-ity the transition would not be abrupt and theconceptual model is oversimplified.

6. High levels of helium may also contribute anincrement of hardening beyond that due to dis-placement damage alone. However, as noted pre-viously, the incremental hardening appears to bemodest up to fairly high concentration[9,10,16,76], at least at lower irradiationtemperatures.

7. A number of experimental studies have usednickel doping to produce high concentrations ofhelium by two-stage 58Ni(nth,c)� > 59Ni(nth,a)reactions. It is well established that nickel addi-tions result in additional hardening at lower irra-diation temperatures [33,77–79]. This may be dueto enhancement of hardening from defect clustersas well as fine scale nickel enriched precipitates[80,81]. More generally, nickel additions changethe transformation temperatures and kinetics inquenched and tempered steels, hence, modifyingthe overall microstructure. Thus increases inDTc associated with nickel doping may, or maynot, be due to helium.

8. Doping with boron isotopes (natural boron is�0.2 10B and 0.8 11B) can also be used to producelarge quantities of helium from the 10B(n,a)6Lithermal neutron reaction. However, the solubil-ity of boron in steels is extremely low, and borondoping is also confounded by (a) the non-uni-form distributions, which tend to segregate boronat PAGs and other interfaces and/or to formboride precipitate phases; (b) boron’s generaleffects on the microstructure and properties ofquenched and tempered alloys; (c) boron’s rolein strengthening grain boundaries; (d) the pro-duction of equal amounts of transmutant lithiumand helium while eliminating boron; (e) veryrapid transmutation and burn-out of boron atrelatively low dpa; and, (f) at low doses, excessdpa due to recoils from the boron n, a reactions.

9. The effects of such confounding factors can bepartially mitigated by careful experimentaldesigns. For example, comparisons of DT inpaired alloys doped with either 58Ni or 60Ni

(alternatively 10B or 11B) can help isolate theindependent effect of helium.

Within this framework the next section evaluatesthe relation between DTc and dpa and between DTc

and Dry. The DTc/Dry relation is then used to assessthe contributions of NHE, including potentialeffects of high levels of He.

3.3. Preliminary analysis of the DTc data

We carried out a similar analysis of DTc as afunction of dpa and Ti. However, in this case, thealloy classes were analyzed separately, since therewere more significant differences between the vari-ous steels and a common Ti binning was not asuseful. The data, which in most cases were very scat-tered, could be reasonably fit using a simpler non-saturating function

DT c ¼ kc

pdpa: ð3Þ

The DTc versusp

dpa data and kc fits are shown inFig. 6(a)–(c) [4,7,20,22,26,27,50–58,82–96]. Note,these plots do not include a comprehensive databaseassembled by Schneider and co-workers [93,94] on avariety of �9Cr–W and �9Cr–Mo steels, includingsome alloys with Ni additions, as well as F82H. Thisdatabase will be discussed further below. Fig. 7shows a cross-plot of kc from the fits in Fig. 6 as afunction of Ti. The 9Cr–Mo and F82H alloys showthe lowest and highest kc, respectively, with the9Cr–W steels generally falling between, but closerto F82H. The Ti dependence of the kc is in reason-able agreement with the corresponding Ti-depen-dence of the initial irradiation hardening as shownby the various solid and dashed lines in Fig. 7 thatare Dry(2dpa)/

p2 multiplied by DTc/Dry conver-

sion factors of = 0.18, 0.30 and 0.38 �C/MPa forthe 9Cr–Mo, 9Cr–W and F82H, respectively. Nota-bly, these values of DTc/Dry are much less than theDT0/Dry � 0.58 ± 0.1 �C/MPa found for fracturetoughness reference temperature shifts [1]. In thecase of F82H, a significant DTc persists in theTi > 400 �C, when irradiation hardening is minimal,or when softening is observed. Note that high, andeven negative values, of DTc/Dry are also observedin the Schneider’s database for Ti = 450 �C.

Fig. 8 plots the available paired DTc and Dry

datasets, including results from the preliminaryassessment of Schneider’s database. In the lattercase, the Dry is based on measurements of changesin dynamic yield stress. Fig. 8(a) shows the results

-50

0

50

100

150

200

9Cr-1Mo

50-60 °C

266-300 °C

365-400 °C

420-450 °C

500-550 °C

ΔTc

(˚C

)

-50

0

50

100

150

200

9Cr-(1-2)W

60-70 °C240-260 °C280-300 °C350-365 °C393-400 °C450 °C

ΔTc

(˚C

)√dpa

-50

0

50

100

150

200

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

0 1 2 3 4 5 6 7

F82H

60-80 °C230-255 °C300-352 °C400 °C450-592 °C

ΔTc

(˚C

)

√dpa

√dpa

ba

c

Fig. 6. Thep

dpa dependence of DTc for (a) F82H; (b) 9Cr–W; and (c) 9Cr–Mo steels and corresponding fits at various Ti.

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 37

for Ti < 400 and Fig. 8(b) is for 400 6 Ti 6 450 �C.Where possible, the Dry are for Tt = 25 �C. How-ever, the Tt = 25 �C data is limited for static tensiletest measurements of Dry, and the open symbols arefor cases when only Tt � Ti are available. The linesshow the nominal and ± one standard deviationsfits for static tensile data only. Note the data withDry < 0 are not used in these fits. Once again, thedata are highly scattered. The fitted DTc/Dry are�0.38 ± 0.18 �C/MPa for Ti < 400 and 0.76 ±0.42 �C/MPa for 400 6 Ti 6 450 �C bin. Schneider’sdatabase also showed similar trends with DTc/Dry � 0.41 ± 0.17 �C/MPa for Ti < 400 and0.75 ± 1.07 �C/MPa for Ti P 400 �C bin. Fig. 9shows the DTc/Dry as function of Ti, where the sta-tic and dynamic (d) DTc/Dry/yd were averaged overthe various Charpy Ti-bins. Error bars show thestandard deviation in each Ti bin. Both static anddynamic DTc/Dry are �0.4 up to Ti � 360 �C,

increasing to >1 �C/MPa with very large scatter atTi = 450. These trends are consistent with thoseobserved for F82H and the 9Cr–W steels in Fig. 6,showing that DTc/

pdpa persists above 400 �C when

Dry is �0, or is even <0.These results lead to two very important

conclusions:

• The DTc/Dry irradiations at Ti < 400 �C are gen-erally less than the corresponding fracture tough-ness indexed DT0/Dry shifts. Thus the DTc fromsub-sized Charpy tests are probably non-conser-vative, and may seriously underestimate theembrittlement potential of TMS alloys. Assum-ing a nominal DT0/Dry � 0.6 ± 0.1 �C/MPa anda maximum Dry = 600 MPa leads to an esti-mated DT0 � 360 ± 60 �C. Later we will showthat the DTc/Dry, hence presumably DT0/Dry aswell, may be even larger for steels containing

-20

0

20

40

60

80

100

120

140

0 100 200 300 400 500 600

F82H9Cr-W9Cr-Mox 0.38x 0.3x 0.18

kc(

=ΔT c

/√dp

a)

Ti (ºC)

Fig. 7. The symbols are the kc(=DTc/p

dpa) versus Ti, foundfrom the fits to the data for the various alloy groups shown inFig. 6. The dashed lines represent the average Dry (2 dpa,Ti)derived from the data in Fig. 1 multiplied by an alloy classdependent Cc shown in the legend.

38 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

high levels of helium, which would likely result inunacceptable T0 values.

• The DTc/Dry are larger for irradiations atTi P 400 �C; and in some cases DTc > 0 forDry < 0, indicating a contribution of NHE.

3.4. Non-hardening embrittlement

The persistence of DTc to higher temperaturesand large or negative values of DTc/Dry signal a

0

100

200

300

400

500

600

700

-100 0 100 200 300 400 500 600

F82H9CrW9CrMoSchneider's[89]

ΔTc

(ºC

)

Δσy (MPa)

Ti< 400 °C

Tt= RT

ba

Fig. 8. DTc versus Dry (or equivalent) plots from paire

non-hardening embrittlement (NHE) contribution.There are several potential sources of NHE thatoccur primarily at higher temperatures under ther-mal aging conditions, including some that may alsobe assisted by irradiation:

• Precipitation (Laves) and coarsening (carbides)of brittle grain boundary phases that act as cleav-age ‘trigger-particle’ or brittle grain boundaryfracture paths.

• Grain boundary segregation of trace impurities,such as P, and/or depletion of beneficial elementslike C.

• Instabilities in the dislocation and lath-packetsubstructures leading to larger effective subgrainsizes.

• Development of damage, particularly on grainboundaries, in the form of microcracks, gas bub-bles and creep cavities. The potential for NHE byhigh concentrations of helium is discussed below.

• In principle, high concentrations of transmutanthydrogen may also be a potential source ofembrittlement in TMS. Hydrogen embrittlementis enhanced by high strength levels, hence, mayact synergistically with irradiation hardening.Hydrogen is associated with the reduced cohe-sion of grain boundaries, where it is segregatesto much higher concentrations than in the bulk,hence may also interact with other NHE mecha-nisms, including those due to helium (see below).However, there are no systematic studies of theeffects of hydrogen embrittlement in TMS, except

0

100

200

300

400

500

600

700

-100 0 100 200 300 400 500 600

F82H Tt =Ti

F82H Tt =RT

9CrW Tt =Ti

9CrW Tt =RT

9CrMo Tt =Ti

9CrMo Tt =RT

Schneider's[89]

ΔTc

(ºC

)

Δσy (MPa)

Ti ≥ 400 °C

d datasets from (a) Ti < 400 and (b) Ti P 400 �C.

0

0.5

1

1.5

2

2.5

3

0 100 200 300 400 500

Δσy/ΔΤ

Δσyd /ΔΤ

Cc

=ΔT

/Δσ y

Ti(oC)

Fig. 9. Static and dynamic (designated by the subscript d) DTc/Dry/yd versus Ti averaged over the various Ti-bins. The error barsshow the standard deviation in each Ti bin.

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 39

at low temperatures following charging [97].Post-irradiation thermal desorption studies,including some H-ion implantation experiments,suggest that most of the hydrogen is expectedto diffuse out at Ti P 250 �C [16,98].

NHE due to precipitation and solute segregationcan result from radiation-enhanced diffusion(RED) and radiation induced segregation (RIS).Further, excess fluxes of point defects enhancestructural instabilities, like dislocation recovery. Inparticular, it is expected that the diffusion con-trolled, lower leg of a time–temperature DTc C-curve is shifted to lower temperatures by RED asis illustrated in Fig. 10. The Ti corresponding to abulk RED coefficient (D*) under irradiation that is

log ta

Ta

log t

ΔTc

RED

Thermal aging

Fig. 10. Schematic illustration of the effect of radiation enhanceddiffusion (RED) on time–temperature DTc embrittlement C-curves.

the same as the thermal diffusion coefficient (Dth)at the higher aging temperature Ta can be estimatedfrom

T i ¼ T a 1� ½RT u=Qd� lnðrÞf g: ð4Þ

Here Qd is activation energy for thermal diffusionand r is the D*/Dth ratio at Ti. Assuming a reason-able value of r = 104 at a dose rate of3 · 10�7 dpa/s and Qd = 300 kJ/mole, Eq. (4) pre-dicts a Ti � 405 �C for Ta = 600 �C. The NHEkinetics would be further accelerated if irradiationenhances other rate limiting steps like precipitatenucleation or dissolution. Note, however, the effectsof RED on grain boundary dominated transportprocesses are likely to be less significant and havenot been modeled. Further, NHE associated withRED processes like precipitation are bounded byfactors such as solute depletion, or may transformto slower kinetics, like coarsening versus growth ofa brittle phase.

The limited data on thermal embrittlement ofTMS alloys [24,25,74,75,99] is summarized inFig. 11 as DTc versus ta plots for various alloysand Ta. Notably, significant DTc are observed foraging ta 6 10000 h in the Ta range from 500 to650 �C. As expected, DTc increases with ta and Ta.The apparently linear increase with ta and the rathermodest Ta-dependence are not understood. ThisNHE due to thermal aging is probably primarilydue to precipitation of brittle Laves (nominallyFe2W) phases on the grain boundaries. Thus, asexpected, the alloy JLS-2 with 3% W has a some-what larger DTc compared to F82H with 2% W.The legend in Fig. 11 also shows the equivalent tem-perature under RED conditions based on Eq. (4)and the parameters cited above.

Fig. 12 plots an embrittlement time (or inverserate) parameter ln(ta/DTc) versus Ta for the F82HTMS alloy. The dashed line shows the predictionof Eq. (4) and associated parameters for the effectof RED. The open diamonds show the correspond-ing embrittlement rate for the irradiated data shownin Fig. 7 at �400 and 510 �C for a dpa rate of3.5 · 10�7/s (�10 dpa/year). The NHE data inFig. 12 is in reasonable agreement with the pre-dicted effect of RED, but has a temperature-depen-dence that is opposite to that predicted by thesimple model. In part, this may be due to ignoringa contribution to DTc from irradiation hardeningat 400 �C and a possible softening effect at�500 �C. However, these results do indicate thatsignificant DTc associated with irradiation enhanced

-20

0

20

40

60

80

100

0 2500 5000 7500 10000 12500

F82H(2W) 500/348 °CF82H(2W) 550/377 °CF82H(2W) 600/405 °CF82H(2W) 650/433 °CJLF1(2W) 550/377 °CJLF1(2W) 650/433 °CJLS2(3W) 600/433 °CEurofer(1W) 500/348 °C

ΔTc(°

C)

t a (h)

Fig. 11. The DTc versus aging time ta and temperature, Ta. The second set of entries in the figure caption are estimates of the equivalenttemperatures accounting for RED, assuming a dose rate of 3 · 10�7 dpa/s and Qd = 300 kJ/mole (see text for details).

300

400

500

600

700

12 13 14 15

ThermalIrrad.

T a (º

C)

ln (ta/ΔTc)

Fig. 12. An embrittlement rate parameter, ln(ta/DTc), versus Ta

for F82H. The dashed line shows the prediction of Eq. (4) andassociated parameters estimating the effect of RED. The symbolsshow the corresponding embrittlement rate for the irradiateddata shown in Fig. 7 at �400 and 510 �C assuming a dpa rate of3.5 · 10�7/s (�10 dpa/year).

40 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

NHE processes can be expected at Ti > 400 �C.While such NHE may or may not itself be a majorproblem, it will likely interact synergistically withother hardening and NHE mechanisms, thus poten-tially producing a significant contribution to DTc.

NHE is often accompanied by a transition fromtransgranular cleavage fracture (TGC) to intergran-ular fracture (IGF) along weakened PAGs or otherboundaries. As discussed elsewhere [1], the transi-

tion to IGF can be conceptually linked to gradualdecrease of a critical grain boundary fracture stressr�ig to the point where it falls below the critical stress,r�c , for TGF. Developing an IGF DTc model is animportant future objective.

3.5. He effects on fast fracture and embrittlement

Since the effects of fusion relevant helium levelson hardening appear to be modest, the major issueis the potential role of helium in NHE. The frame-work outlined above has been incorporated inrecent experiments. For example, a single variableexperiment is being carried out to compare theDT0, Dry and Ck for TMS doped with 58Ni and60Ni isotopes as well as natural Ni; and anotherexperiment is planned that will use alloys dopedwith 10B and 11B isotopes. However, these principlesare generally not reflected in past studies. Indeed,most previous experiments did not even include ten-sile specimens to provide Dry data as a complementto the DTc measurements from sub-sized Charpytests; and even the simple expedient of microhard-ness measurements to estimate irradiation harden-ing was rarely exploited. Further, there have beenrelatively few fractographic studies to characterizethe local fracture mode, and any transitions fromTGC to IGF in irradiated alloys as a function ofhelium content.

Nevertheless, it is still possible to analyze existingdata to try to detect, and even quantitatively esti-mate, the potential helium NHE by examining the

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 41

relation between DTc and Dry and/or Cc = DTc/Dry.Large increases in Cc with increasing helium signal aNHE helium contribution and vice versa. Of course,even if Cc appears to increase with helium, a corre-sponding NHE is not assured due to the confound-ing factors. Further, uncertainties in Cc must also beconsidered. Even relatively optimistic estimates ofuncertainties of only ±10 �C in DTc and ± 20 MPain Dry mean uncertainties in Cc of � ± 10/DTc

(�C/MPa) for an actual Cc = 0.4 �C/MPa. Thus arelatively large scatter in Cc values should beentirely anticipated, particularly at small DTc, andit is always useful to examine DT versus Dry scatterplots as well as the trends in Cc. Of course there areother indicators of helium to NHE. The most signif-icant is the transition from TGC to IGF above somecritical level of helium [10,12,16].

There are pertinent data from spallation protonirradiations based on the so-called small punchtests (sp) that can be used to estimate both irradi-ation hardening and transition temperature shifts[12,13]. However, the small punch technique isnot a true fracture test per se. An effective yieldstress (ryp) and corresponding irradiation harden-ing (Dryp) are evaluated based on the nonlinearitiesin punch load–displacement curves. An effectivetransition temperature (Tsp) and correspondingirradiation induced shifts (DTsp) are determinedby plotting the integrated load–displacementenergy at fracture, measured at the point of a largeload drop, as a function of the test temperature.Since small punch tests inherently involve static,low constraint conditions, with small values of M

(or ratios of the maximum principal stress relativeto the ry), the brittle (TGC or IGF) to ductile tran-sition, if any, occurs at temperatures that are muchlower than for standard, and even sub-sized,Charpy impact tests; and the corresponding shiftsin DTsp, due to irradiation embrittlement, are alsomuch lower than DTc. An empirical adjustment ofDT ðspÞ

c � 2:5DT sp is used here to evaluate DT ðspÞc

and Cc [12,13,15,100,101]. Finally, while highlyunusual in steels, brittle fracture may also occurin tensile tests with lower bound values of M � 1.As noted above this implies a very low value ofthe critical brittle fracture stress. However, evenin a tensile test, M is somewhat greater than 1 inthe necked region beyond the uniform strain limit.The maximum M increases with necking, and canbe related to the reduction in area at fracture andthe corresponding tri-axial stress state in thenecked region.

Fig. 13(a) and (b) summarizes the very limitedavailable Charpy-tensile test DTc versus Dry datawith helium variations in nominally similar alloyswith and without nickel or boron doping for irradi-ations at �300 and 400 �C, respectively [26,52,55,56,82,87,91,102]. Helium content shown in paren-thesis is common for both Charpy and tensile spec-imens except for some cases where data for acommon condition are not available; in these cases,both values of (Charpy/tensile) specimens areshown. The dashed lines indicate the scatter in thegeneral database for DTc versus Dry. Two data setsare for a 9CrMo alloy both undoped and dopedwith 2% Ni irradiated at 300 and 400 �C. The nickeldoping results in increases in helium from �14 to234 and 33 to 369 appm in these two cases, respec-tively. Both the doped and undoped alloys fall onvirtually the same DTc versus Dry line withCc � 0.40 �C/MPa at 300 �C and Cc � 0.82 �C/MPa at 400 �C. Thus while the nickel additionincreases the Dry in both these cases, there is noindependent NHE effect of either nickel, or helium,indicated by this data set. The corresponding DTc

for a 9Cr2W alloy doped with 2% nickel irradiatedat �300 �C with �115 appm He falls slightly abovethe DTc versus Dry line for the correspondingundoped steel with �5 appm helium, with Cc of0.44 and 0.32 �C/MPa, respectively. However, bothare well within the overall DTc versus Dry and Cc

scatter band. A generally similar result is foundfor natural B doped F82H and JLF1 steels irradi-ated at 300 �C, with �23–40 appm He, comparedto undoped alloys with �3 appm He. These resultsmay suggest a slightly stronger effect of boron com-pared to nickel doping, with values of �0.53 and0.75 (high helium) versus 0.43 and 0.58 (lowhelium), respectively.

The DTc versus Dry for the 12Cr1Mo Ni-dopedand undoped alloys in Fig. 13(a) and (b) also showno effect of helium and the data all fall within thenormal scatter bands [26,91,102]. However, a recentreview paper included 300 �C data on F82H dopedwith various isotopes of B showed a systematicincrease of Cc with increasing helium up to340 appm [103]. Thus these results may also suggesta stronger effect of boron compared to nickel dop-ing, perhaps due to B segregation to PAGs. Overall,however, the limited nickel and boron doping datashow little effect of helium on Cc up to concentra-tions �350 appm.

Dai and co-workers have reported a nearly linearcorrelation between the helium level and DTsp for

0

50

100

150

200

250

300

350

400

0 100 200 300 400 500 600

JLF1 HFRJLF1+B HFRF82H HFIR/HFRF82H+B HFR9CrMo HFIR9CrMo+2%Ni HFIR12CrMo HFIR12CrMo+2%Ni HFIR

ΔT(º

C)

Δσy (MPa)

(40) (3)(14)

(234)

(23)

(3)

Tirr.

=300 ºC

(71)

(234)

0

50

100

150

200

250

300

350

400

0 100 200 300 400 500 600

9Cr2W9Cr2W+2Ni9CrMo9CrMo+2Ni12CrMo12CrMo+2Ni

ΔT(º

C)

Δσy (MPa)

(369)

(33)

(5)

(115)

Tirr.

=400 ºC

(372)

(109)

HFIR (unless otherwise noted)

ΔTc

(°C

)

ΔTc

(°C

)

0

50

100

150

200

250

300

350

400

0 100 200 300 400 500 600

9Cr1Mo SPI@90-275 º CJLM CYRIC@115±35 ºCF82H SPI@80-265ºC

ΔT(º

C)

Δσy (MPa)

(588)

(125)

(140@80 ºC)

(285@150 ºC)

(375@180 ºC)

(680@265 ºC)

(165@90 ºC)

(310@160ºC)

(770@275 ºC)

ΔTc

(°C

)

d

e

Fig. 13. (a) and (b) The DTc versus Dry for data with helium variations due to B or Ni-doping in otherwise in nominally similar steelsirradiated at (a) 300 �C; and (b) 400 �C. (c) SP irradiation small punch test DTsp versus Dryp data with corresponding helium variations. (d)The fracture surface and (e) the bottom surface of punch-tested specimen with 770 appm He at 9.4 dpa showing a dominance of IGF[taken from Ref. [12]].

42 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

spallation proton irradiations of 9Cr1Mo T91 andF82H steels at nominal temperatures between �80and 275 �C from �2.5 to 9.4 dpa producing �140to 770 appm He [12,13]. Note the temperatures,He levels and dpa generally increase in tandem.Fig. 13(c) plots the small punch DT ðspÞ

c versus Dryp

data. The helium levels and nominal irradiationtemperatures are shown in the parenthesis near eachdata point. At the intermediate helium levels, thedata fall in the lower region of the overall DTc ver-sus Dry scatter band for this irradiation temperaturerange shown by the dashed lines. Jia and Dai alsonoted the fact that these data fall in the general scat-ter band of DT versus dpa for neutron irradiations[13]. However, there is a distinct break in the

DT ðspÞc versus Dryp trend for the data points at

680 appm He in F82H and 770 appm He in T91.Although the corresponding Csp values are neitherparticularly high nor anomalous, the onset of signif-icant amounts of IGF are observed in these cases,suggesting the emergence of a NHE contributionat very high helium levels. Note, IGF is readilyobserved both on the fracture surfaces and as anetwork of grain boundary cracks on the bottomsurface of the small punch disc as shown in themicrographs taken from Ref. [12] (Fig. 13(d) and(e)).

Of course, the SP-irradiation results based onsmall punch tests do not represent a single variableexperiment, and interpretations of this data should

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 43

be viewed with considerable caution. For example,high levels of hydrogen and other transmutationproducts may also play a role in the larger Cc

observed at the highest spallation proton dose, con-founding unambiguous conclusions about theeffects of helium acting alone. Fig. 13(c) also showsDT ðspÞ

c versus Drym data for 36 MeV a-ion irradia-tions to 100 and 600 appm He at 115 ± 35 �C basedon assessments of DTsp and hardening evaluatedusing Vickers microhardness tests (Drym) [15]. Inthis case, the DT ðspÞ

c versus Drym data fell near thetop of the general scatter-band, but there is essen-tially no effect of higher helium on the DT ðspÞ

c versusDrym relation.

Another useful source of paired DTc–Dryd datahas been developed by Schneider and co-workerson a variety of steels that ‘happen to have’ differentnatural boron contents [21,22,91,92]. The Schneiderdatabase includes mixed spectrum neutron irradia-tions from 0.2 to 2.4 dpa at temperatures between250 and 450 �C that produce up to �120 appmhelium. Based on empirical plots of DTc versushelium for subsets of this database, various authors[20,23] have also suggested a dominant role ofhelium on embrittlement. Indeed, fits to the overalldatabase for Ti < 400 �C do result in a weak corre-lation that would indicate a strong effect of helium,that would, if true, result in an enormousDTc � 1080 �C at 1000 appm He. Note, the 450 �Cdata indicates that DTc decreases with increasinghelium, but is excluded from this analysis, sincethe Cc are very scattered and, as expected, are onaverage systematically larger than those observedat lower irradiation temperatures, indicating aNHE mechanism that is not related to helium. Thisis consistent with other observations of increased Cc

at higher irradiation temperatures, but is not perti-nent to the assessment of NHE due to helium. Evenrestricting the analysis to 400 �C or less, the DTc ver-sus helium data is very scattered and the statisticalsignificance of the apparent correlation is low, withr2 � 0.21.

However, a much more significant shortcomingof this sort of analysis is that it neglects the largerange of Dryd in the irradiated steels. This masksthe combined effects of both differences in dose(dpa) as well as the wide range of metallurgical vari-ables representing the various alloys in the Schnei-der database. Both Dry and Cc = DTc/Dry (here,the dynamic yield stress increase) hence the corre-sponding DTc, depend on these variables, in a waythat is largely independent of either the alloy boron

or helium content. Note, Cc itself also increases withDry and depends on the alloy microstructure andunirradiated Charpy properties [5]. At lower irradi-ation temperatures, Dry is the most significant fac-tor in controlling DTc, thus it must be accountedfor before attempting to evaluate any potentialeffects of any other variables.

Notably, three of the alloys with the highestboron contents in the Schneider database alsocontained significant quantities of nickel (0.66–0.92% Ni). At 300 �C, the Dryd in these nickel-bear-ing alloys is higher by an overall average factor of�1.6 ± 0.25 relative to a set of nominally similarnickel free alloys. Indeed, there is a systematic over-all correlation between Dry and the alloy nickel con-tent. The nickel sensitivity of Dry decreases withincreasing Ti, dropping significantly between 350and 400 �C. Fig. 14(a) shows that the DTc versusDry generally fall in the same scatter band as the lar-ger overall database for �8Cr TMS, as shown bythe dashed lines. Fig. 14(b) plots the correspondingCc = DTc/Dry as a function of helium content. Leastsquares fits show an apparent effect of higher heliumincreasing DTc only for the data for irradiations at250 �C; and in that case, the effect of helium is pri-marily manifested in the range below 50 appm He,so is not very credible. The fits for the largest setof data for irradiations at 300 �C show only a veryweak apparent effect of helium, while at higher tem-peratures there is a negative association betweenhelium and Cc. A best fit to all the Cc data in theSchneider database for irradiations at Ti 6 400 �Cshown by the solid line, indicates a very weak,statistically insignificant, negative effect of helium.These results are not surprising, since all but onedata point is below 100 appm helium. The scatterin Cc = 0.41 ± 0.2 �C/MPa, shown by the horizon-tal dashed lines, is very consistent with the corre-sponding values for the larger overall TMS alloydatabase.

The band bounded by the short dashed lines onFig. 14(c) represent the general scatter plot for theCc found in the previous analysis, indicating no,or a weak, effect of helium up to �350 appm.Fig. 14(c) also shows the relation between Csp andhelium for the SP irradiation test data from Daiand Kasada as heavy dashed lines [12,15] normal-ized to a typical value of 0.4 �C/MPa at lowerhelium levels. The normalized Cc for the smallpunch test data exceed 1 �C/MPa at the highesthelium level, consistent with a significant NHE ofhelium and the observation of IGF.

0

50

100

150

200

250

300

350

400

0 100 200 300 400 500 600

250 ºC300 ºC350 ºC400 ºC

Δσyd

(MPa)

0.00

0.25

0.50

0.75

1.00

1.25

C(°

C/M

Pa)

0 25 50 75 100 125appm He

Schneider Database

1SD ±C range

LS FIT

400 °C

350 °C

300 °C

250 °C

0.0

0.5

1.0

1.5

Cc

(°C

/MPa

)

1 10 100 1000

appm He

T91(SPI)

F82H(SPI)

nomalized to Ccfor <350appm He

ΔTc

(°C)

Fig. 14. (a) DT versus Dryd from Schneider’s database compared to the overall scatter band for TMS irradiations at 6400; (b) Cc = DTc/Dryd for Schneider’s database as a function of helium content with a least squares fit line; (c) Cc values from Schneider database along withpaired lower-higher helium Cc values from Fig. 8(a) and (b) shown as heavy solid lines.

44 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

Another data set published by Henry and co-workers [10], at least indirectly pertinent to the issueof NHE, are the results of tensile tests at 25 and250 �C following SP irradiations of a 9Cr1Mo(EM10) steel in various thermo-mechanical treat-ment conditions at �260 �C to �9.8 dpa and750 appm He. The alloy conditions included as-tem-pered (AT), as-quenched (AQ) and tempered andcold-worked (CW). A strong effect of the alloythermo-mechanical treatment and test temperaturewas observed in this dataset. In three cases the spec-imens failed after considerable deformation atreductions in area (RA) from about 44% to 48%that were somewhat less than the correspondingRA of 58–68% in the unirradiated condition. The

RA ductility varied with test temperature andthermo-mechanical treatment condition andincreased with decreasing ry (given in the parenthe-sis) in the order AT at 250 �C (960 MPa); AT at25 �C (1150 MPa) and CW at 25 �C (1220 MPa).The fracture surfaces were a mixture of ductilemicrovoid coalescence and cleavage, with someIGF and transverse cracking. Assuming there isno strain hardening, the average fracture stress inthe necked region is on the order of �2000 MPa,consistent with typical values of r�c for cleavage.Indeed, the principal tri-axial stress would be evenhigher in the necked region, and this may compen-sate for some level of strain softening. The AQ alloytested at 25 �C had the largest ry of 1320 MPa and

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 45

lowest ductility with a RA of only 2% and �100%IGF. Given the small amount of necking prior tofracture it is not unreasonable to assume thatr�ig ¼ ry � 1300 MPa, which is much lower than typ-ical values of r�c . This low value may at least in part bedue to the high level of helium. This conclusion is sup-ported by the observation that the fracture mode was�100% ductile microvoid coalescence after neutronirradiations of the AQ alloy to a slightly higherstrength ry � 1330 MPa. A comparison of the frac-ture surfaces for the neutron and spallation protonirradiations is shown in Fig. 15 [10,19].

Jung and co-workers [14] have reported theresults of tensile tests at 25 and 250 �C on EM10and T91 after high-energy a-implantation at250 �C to 0.8 dpa and 5000 appm He. These irradi-ations result in high ry = 1079 to 1217 MPa accom-panied by very low ductility with the reduction inarea (RA) �0% along with �100% IGF. The ductil-ity was much higher and the ry much lower after a-implantations at 550 �C. The RA was also muchlower for T91 tensile tests at 550 �C compared to25 �C.

While no firm conclusions can be drawn from thetensile data, it is extremely significant that �100%IGF is observed at normal stress levels in the rangeof �1100 MPa (5000 appm He) to 1300 MPa(750 appm He). Further the results suggest that inaddition to helium itself, NHE and IGF also dependon both the irradiation and test temperatures. Ofcourse the presence of cracks or notches and/or highloading rates would make materials that are brittleeven in a static tensile test, even more so in normalstatic fracture tests as well as under dynamic loadingconditions.

Fig. 15. Tensile fracture surfaces of an AQ 9CrMo steel after (a) neutron�250 �C, showing ductile and IGF, respectively [from Refs. [10,19]].

Based on the observation of an apparent heliumenhanced Cc = DTc/Dry > 0.4 �C/MPa it is reason-able to assume that the corresponding toughnessreference temperature shift would also increase withhigh helium to C0 = DT0/Dry > 0.6 �C/MPa. Incombination with the model for Dry(dpa,Ti) asshown in Fig. 4, this hypothesis can be used to con-struct a simple model for DTc and DT0 as

DT c=0 � Cc=0ðHeÞDryðdpa;T iÞ: ð5Þ

Based on the normalized Csp(He) trends inFig. 14(c), we crudely estimate the Cc/k(He) as

Cc=0ðHeÞ ¼ ð0:4 or 0:6Þþ 0:0007ðHe� 500Þ �C=MPa

5006He6 1500; ð6aÞCc=0ðHeÞ ¼ 1:1 or 1:3 �C=MPa He> 1500: ð6bÞ

Here we have taken the absolute rate of increase ofCsp with helium above the threshold; using the nor-malized value would increase the helium coefficientby about a factor of �2. Fig. 16 shows the resultingDT0 and DTc predictions for Ti = 300 �C for typicalfusion reactor (Fig. 16(a), He/dpa = 10 appm/dpa)and SP irradiation (Fig. 16(b), He/dpa = 100appm/dpa) environments. Fig. 16 also shows theadditional effect of dynamic ðe0dÞ versus static ðe0sÞloading rate on the total shift. The T0 shift due to dy-namic loading, DTd, is given by a strain rate compen-sated temperature Td which is higher than theequivalent static fracture temperature using theexpression [104,105]

DT d ¼ T s ½1þ 0:035 lnðe0d=e0sÞ� � 1� �

: ð7Þ

The predicted shifts are large from hardening aloneand become enormous at high He levels. Assuminga starting Tc/0 of �100 �C, the irradiated Tc/0/d

and (b) spallation proton irradiations to the similar levels of ry at

0

200

400

600

800

1000

1200

0 10 20 30 40 50

ΔTo(ε's)

ΔTo(ε'd)

ΔTc

dpa

He/dpa = 100 (appm/dpa)

ε'd/ε's≈3 x 104

0

200

400

600

800

1000

1200

0 20 40 60 80 100

ΔTo(ε's)

ΔTo(ε'd)

ΔTc

ΔTo,Δ

Tc

(°C

)

ΔTo,Δ

Tc

(°C

)

dpa

He/dpa = 10 (appm/dpa)

ε'd/ε's ≈3 x 104

Fig. 16. Composite Cc/0/d(He) and Dry(dpa) model predictions of DT0, DTc and DTd (dynamic) for Ti = 300 �C for typical (a) fusionreactor (He/dpa = 10 appm/dpa); and (b) SP irradiation (He/dpa = 100 appm/dpa) environments.

46 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

quickly reach values in excess of the irradiation tem-perature (�300 �C). Notably, T0/d values of 400 �Cwould result in unacceptable median toughness val-ues of <40 MPa

pm, as well as lower shelf bounding

toughness. However, it is important to emphasizethat these plots cannot be viewed as ‘predictions’,but rather they are meant to emphasize the potentialimportance of the synergistic high helium-highhardening effects.

4. Discussion, summary and conclusions

The work reported here on developing and ana-lyzing a comprehensive database on irradiationeffects on the mechanical properties of �8Cr TMSalloys should be viewed as a first step, work inprogress in a long-term effort. However, the preli-minary evaluation of the Ti and dpa-dependenceof irradiation hardening, Dry, and embrittlement,DTc, has provided considerable insight that will helpto guide future research. First, it is clear that theexisting database is not sufficient. Future experi-ments must strive to avoid uncontrolled variablesand confounding variable combinations, as well asto provide sufficient detail to be compatible withinclusion in a high quality database. Such detailsshould include explicit estimates of uncertainties inkey parameters, such as the irradiation temperature,Ti.

In spite of the limitations in the database, how-ever, our analysis provided a first-iteration semi-empirical model of the dpa, Ti and Tt dependence

of Dry and DTc. This model will be refined in thefuture, but demonstrates that very large Dry

approaching 500–600 MPa can be anticipated athigh doses at Ti around 300 �C, especially at lowerTt. There appears to be little systematic differencebetween hardening produced by spallation proton(SP) and neutron irradiations with very high andlow He/dpa ratios, respectively. Our analysis alsoshows that neutron irradiation embrittlement isdominated by hardening below Ti � 400 �C forhelium levels up to �400 appm. However, the hard-ening-shift coefficient, Cc = DTc/Dry, for sub-sizedCharpy tests, is lower than for fracture toughnessreference temperature shifts, DT0. Thus DTc is anon-conservative measure of embrittlement. Clearly,future irradiation experiments must focus on estab-lishing a comprehensive DT0 database as well as forother fracture toughness properties, like ductiletearing resistance.

Our analysis also suggests that a non-hardeningembrittlement (NHE) contribution to DTc occursat Ti > 400 �C. The NHE can be rationalized onthe basis of decreases in thermal aging temperaturesdue to radiation enhanced diffusion (RED), typi-cally occurring at temperatures >500 �C. Whileprobably not a major issue per se, such NHE mayact synergistically with hardening and other NHEmechanisms, such as those due to high concentra-tion of and helium and possibly hydrogen.

We have collected and analyzed a variety of datapertinent to the issue of the effect of helium on fastfracture by evaluating the relation between DTc and

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 47

Dry. The hardening-shift relation was assessed bothin terms of DTc versus Dry scatter plots and trendsin Cc = D Tc/Dry. Large values of Cc, as well astransition from TGC to IGF, at high helium con-centrations were used as signatures of possibleNHE due to helium. While the data is scattered,and to some degree confounded by uncontrolledvariables, the results suggest that up to concentra-tions several hundred appm, helium has little effecton Cc or DTc. However, at higher concentrationsgenerated in SP and HI irradiations, the data is con-sistent with the hypothesis that accumulation ofhelium sufficiently weakens grain boundaries tothe point where a NHE effect emerges as signaledby both higher Cc and IGF. This hypothesis is alsosupported by the results of tensile tests on TMS inboth high strength irradiated state and implantedwith very high concentrations of helium. Notably,previous fractography studies to characterize IGFhave been very limited and should be a major focusof future work.

Acknowledgements

The authors explicitly acknowledge the extensiveresearch that produced the data analyzed in this pa-per and thank the researchers cited in the referencesfor their major contributions to the development ofmaterials for fusion reactors. Our many fruitfulcollaborations with many US colleagues in the Ad-vanced (Fusion) Materials program, Philippe Spatigof EPFL-CRPP and Yong Dai of PSI are also grate-fully acknowledged. The work described here wassupported by the US Department of Energy, Officeof Fusion Science (DE-FG03-94ER-5475).

References

[1] G.R. Odette, T. Yamamoto, H.J. Rathbun, M.Y. He, M.L.Hribernik, J.W. Rensman, J. Nucl. Mater. 323 (2003) 313.

[2] G.R. Odette, B.D. Wirth, D.J. Bacon, N.M. Ghoniem,MRS Bull. (2001) 176.

[3] S. Jumel, C. Domain, J. Ruste, J.-C. Van Duysen, C.Becquart, A. Legris, P. Pareige, A. Barbu, E. Van Walle, R.Chaouadi, M. Hou, G.R. Odette, R.E. Stoller, B.D. Wirth,J. Test. Eval. 30-1 (2002) 37.

[4] K. Shiba, A. Hishinuma, A. Toyama, K. Masamura,Properties of the reduced activation ferritic steel F82HIEA heat, JAERI-Tech 97-038 (1997) (in Japanese).

[5] K. Shiba, M. Suzuki, A. Hishinuma, J. Nucl. Mater. 233–237 (1996) 309.

[6] H.E. Hofmans, Tensile and impact properties of Eurofer97plate and Bar, NRG Petten 200023/00.38153/P.

[7] J. Rensman, H.E. Hofmans, E.W. Schuring, J. van Hoepen,J.B.M. Bakker, R. den Boef, F.P. van den Broek, E.D.L.van Essen, J. Nucl. Mater. 307–311 (2002) 250.

[8] J.-W. Rensman, NRG irradiation testing: report on 300 �Cand 60 �C irradiated RAFM steels, NRG Petten 20023/05.68497/P.

[9] K. Farrell, T.S. Byun, J. Nucl. Mater. 296 (2001) 129.[10] J. Henry, X. Averty, Y. Dai, P. Lamagnere, J.P. Pizzanelli,

J.J. Espinas, P. Wident, J. Nucl. Mater. 318 (2003) 215.[11] S.A. Maloy, M.R. James, G. Willcutt, W.F. Sommer, M.

Sokolov, L.L. Snead, M.L. Hamilton, F. Garner, J. Nucl.Mater. 296 (2001) 119.

[12] Y. Dai, X.J. Jia, K. Farrell, J. Nucl. Mater. 318 (2003) 192.[13] X. Jia, Y. Dai, J. Nucl. Mater. 323 (2003) 360.[14] P. Jung, J. Henry, J. Chen, J.-C. Brachet, J. Nucl. Mater.

318 (2003) 241.[15] R. Kasada, T. Morimura, A. Hasegawa, A. Kimura,

J. Nucl. Mater. 299 (2001) 83.[16] P. Jung, C. Liu, J. Chen, J. Nucl. Mater. 296 (2001) 165.[17] A. Kimura, R. Kasada, K. Morishita, R. Sugano, A.

Hasegawa, K. Abe, T. Yamamoto, H. Matsui, N. Yoshida,B.D. Wirth, T.D. Rubia, J. Nucl. Mater. 307–311 (2002)521.

[18] A. Hasegawa, H. Shiraishi, H. Matsui, K. Abe, J. Nucl.Mater. 212–215 (1994) 720.

[19] J. Henry, M.-H. Mathon, P. Jung, J. Nucl. Mater. 318(2003) 249.

[20] M. Rieth, B. Dafferner, H.-D. Rohrig, J. Nucl. Mater. 258–263 (1998) 1147.

[21] H.-C. Schneider, B. Dafferner, J. Aktaa, J. Nucl. Mater. 295(2001) 16.

[22] H.-C. Schneider, B. Dafferner, J. Aktaa, J. Nucl. Mater. 321(2003) 135.

[23] R. Lindau, A. Moslang, D. Preininger, M. Rieth, H.D.Rohrig, J. Nucl. Mater. 271&272 (1999) 450.

[24] B. Van der Schaaf, D.S. Gelles, S. Jitsukawa, A. Kimura,R.L. Klueh, A. Moslang, G.R. Odette, J. Nucl. Mater. 283–287 (2000) 52.

[25] B. Van der Schaaf, A.A. Tavasoli, Hardening and tough-ness from radiation and resistance characterization ofF82H and screening steels, NRG Petten 20027/99.25387/P.

[26] R.L. Klueh, D.J. Alexander, J. Nucl. Mater. 187 (1992) 60.[27] R.L. Klueh, D.J. Alexander, J. Nucl. Mater. 218 (1995) 151.[28] R.L. Klueh, D.S. Gelles, S. Jitsukawa, A. Kimura, G.R.

Odette, B. van der Schaaf, M. Victoria, J. Nucl. Mater.307–311 (2002) 455.

[29] E. Lucon, Mechanical properties of the European referenceRAFM steel (Eurofer97) before and after irradiation at300 �C (0.3–2 dpa), SCKCEN Report BLG-962, November2003, Mol, Belgium.

[30] M.J. Makin, A.D. Whapham, F.J. Minter, Philos. Mag. 7(74) (1962) 285.

[31] M.J. Makin, F.J. Minter, Acta Metall. Mater. 8 (10) (1960)691.

[32] A. Kimura, T. Morimura, M. Narui, H. Matsui, J. Nucl.Mater. 233–237 (1996) 319.

[33] R. Kasada, A. Kimura, H. Matsui, M. Narui, J. Nucl.Mater. 258–263 (1998) 1199.

[34] P. Spatig, R. Schaublin, S. Gyger, M. Victoria, J. Nucl.Mater. 258–263 (1998) 1345.

[35] R. Kasada, A. Kimura, H. Matsui, A. Hasegawa, M.Narui, J. Nucl. Mater. 271–272 (1999) 360.

48 T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49

[36] Y. Dai, S.A. Maloy, G.S. Bauer, W.F. Sommer, J. Nucl.Mater. 283–287 (2000) 513.

[37] E.V. van Osch, J.V. van Hoepen, J. Boskeljon, J. Rensman,Tensile properties of 2.5 dpa 300 �C neutron irradiated RAFM plate, powder HIP and EB and TIG weld, NRG Petten20023/99.26704/P.

[38] R.L. Klueh, J. Alexander, ASTM STP 1325 (1999) 911.[39] E.I. Materna-Morris, H.-C. Schneider, B. Dafferner, R.

Rolli, O. Romer, Irradiation influences on mechanicalproperties and structures of martensitic low-activationalloys, Presented in ICFRM-10.

[40] K. Shiba, personal communication.[41] H. Kurishita, H. Kayano, M. Narui, A. Kimura, M.L.

Hamilton, D.S. Gelles, J. Nucl. Mater. 212–215 (1994) 730.[42] F. Abe, T. Noda, H. Araki, M. Narui, H. Kayano, J. Nucl.

Mater. 191–194 (1992) 845.[43] R.L. Klueh, P.J. Maziasz, M. Vitek, J. Nucl. Mater. 141–

143 (1986) 960.[44] A. Kohyama, H. Matsui, K. Hamada, H. Simidzu, J. Nucl.

Mater. 155–157 (1988) 896.[45] R.L. Klueh, J. Nucl. Mater. 179–181 (1991) 728.[46] R.L. Klueh, J.M. Vitek, M.L. Grossbeck, ASTM STP 782

(1982) 648.[47] J. Russell Hawthorne, J.R. Reeds, J.A. Sprague, ASTM

STP 870 (1985) 580.[48] K. Farrell, T.S. Byun, J. Nucl. Mater. 318 (2003) 274.[49] V.K. Shamardin, V.N. Golovanov, T.M. Bulanova, A.V.

Povstyanko, A.E. Fedoseev, Z.E. Ostrovsky, Yu.D. Gon-charenko, J. Nucl. Mater. 307–311 (2003) 229.

[50] V.V. Rybin, I.P. Kursevich, A.N. Lapin, J. Nucl. Mater.258–263 (1998) 1324.

[51] A. Alamo, M. Horsten, X. Averty, E.I. Materna-Morris,M. Rieth, J.C. Brachet, J. Nucl. Mater. 283–287 (2000)353.

[52] R.L. Klueh, M. Sokolov, K. Shiba, Y. Miwa, J.P. Robert-son, J. Nucl. Mater. 283–287 (2000) 478.

[53] R.L. Klueh, J.J. Kai, D.J. Alexander, J. Nucl. Mater. 225(1995) 175.

[54] E.I. Materna-Morris, M. Rieth, K. Ehrlich, Effects ofradiation on materials, in: 19th International Symposium,ASTM STP 1366 (2001) 597.

[55] E.V. van Osch, M.G. Horsten, G.E. Lucas, G.R. Odette,Effects of radiation on materials, in: 19th InternationalSymposium, ASTM STP 1366 (2001) 612.

[56] J. Rensman, J. van Hoepen, J.B.M. Bakker, R. den Boef,F.P. van den Broek, E.D.L. van Essen, J. Nucl. Mater. 307–311 (2002) 245.

[57] J.L. Seran, A. Alamo, A. Maillard, H. Touron, J.C.Brachet, P. Dubuission, O. Rabouille, J. Nucl. Mater.212–215 (1994) 588.

[58] M.A. Sokolov, R.L. Klueh, G.R. Odette, K. Shiba, H.Tanigawa, ASTM STP 1447 (2004) 408.

[59] A. Alamo, V. Lambard, X. Averty, M.H. Mathon, J. Nucl.Mater. 329–333 (2004) 333.

[60] E. Lucon, R. Chaouadi, M. Decreton, J. Nucl. Mater. 329–333 (2004) 1078.

[61] E. Wakai, T. Taguchi, T. Yamamoto, F. Takada, J. Nucl.Mater. 329–333 (2004) 1133.

[62] C. Liu, H. Klein, P. Jung, J. Nucl. Mater. 335 (2004) 77.[63] N. Baluc, R. Schaublin, P. Spatig, M. Victoria, ASTM STP

1447 (2004) 341.[64] A. Kimura, Fusion Sci. Technol. 44 (2) (2003) 480.

[65] Y. Kohno, A. Kohyama, T. Hirose, M.L. Hamilton, M.Narui, J. Nucl. Mater. 271&272 (1999) 145.

[66] R.L. Klueh, J. Nucl. Mater. 132 (1989) 13.[67] E.D. Eason, J.E. Wright, G.R. Odette, NUREG/CR-6551,

US NRC, Washington DC, 1998.[68] G.R. Odette, T. Yamamoto, D. Klingensmith, H. Kishim-

oto, UCSB-NRC-03/2, 2003.[69] R.B. Jones, T.J. Williams, ASTM STP 1270 (1996) 569.[70] R.L. Klueh, P.J. Maziasz, J. Nucl. Mater. 187 (1992) 43.[71] J.L. Seran, V. Levy, P. Dubisson, D. Gilbon, A. Maillard,

A. Fissolo, H. Touron, R. Cauvin, A. Chalony, E. LeBoulbin, ASTM STP1125 (1992) 1209.

[72] G.R. Odette, P.M. Lombrozo, R.A. Wullaert, ASTM STP870 (1985) 841.

[73] G.R. Odette, M.Y. He, J. Nucl. Mater. 283–287 (2000) 120.[74] L. Schafer, J. Nucl. Mater. 283–287 (2000) 707.[75] H. Sakasegawa, T. Hirose, A. Kohyama, Y. Katoh, T.

Harada, K. Asakura, T. Kumagai, J. Nucl. Mater. 307–311(2002) 490.

[76] L.K. Mansur, J. Nucl. Mater. 318 (2003) 14.[77] G.R. Odette, G.E. Lucas, Radiat. Eff. Def. 144 (1998) 189.[78] G.R. Odette, G.E. Lucas, D. Klingensmith, B.D. Wirth, D.

Gragg, NUREG/CR-6778, 2002.[79] G.R. Odette, T. Yamamoto, D. Klingensmith, Philos. Mag.

85 (2005) 779.[80] D.S. Gelles, G.L. Hankin, M.L. Hamilton, J. Nucl. Mater.

251 (1997) 188.[81] N. Hashimoto, R.L. Klueh, J. Nucl. Mater. 305 (2002) 153.[82] R.L. Klueh, D.J. Alexander, J. Nucl. Mater. 233–237 (1996)

336.[83] I.V. Gorynin, V.V. Rybin, I.P. Kursevich, A.N. Lapin, E.V.

Nesterova, E.Yu. Klepikov, J. Nucl. Mater. 283–287 (2000)465.

[84] N. Baluc, R. Schaublin, C. Bailat, F. Paschoud, M.Victoria, J. Nucl. Mater. 283–287 (2000) 731.

[85] R.L. Klueh, D.J. Alexander, M. Rieth, J. Nucl. Mater. 273(1999) 146.

[86] E.V. van Osch, J.B.M. Bakker, R. den Boef, J. Rensman,NRG Petten 20023/99.26974/P.

[87] R.L. Klueh, D.J. Alexander, J. Nucl. Mater. 258–263 (1998)1269.

[88] T. Lechtenberg, J. Nucl. Mater. 133&134 (1985) 149.[89] R.L. Klueh, J.M. Vitek, W.R. Corwin, D.J. Alexander,

J. Nucl. Mater. 155–157 (1988) 973.[90] H. Kayano, A. Kimura, M. Narui, Y. Sasaki, Y. Suzuki, S.

Ohta, J. Nucl. Mater. 155–157 (1988) 978.[91] R.L. Klueh, D.J. Alexander, J. Nucl. Mater. 179–181 (1991)

733.[92] W.L. Hu, D.S. Gelles, ASTM STP 956 (1987) 83.[93] H.-C. Schneider, B. Dafferner, H. Ries, O. Romer, Irradi-

ation program MANITU: results of impact tests with theirradiated materials (2.4 dpa), FZKA6605, Forschungszen-trum Karlsruhe, 2001.

[94] H.-C. Schneider, B. Dafferner, H. Ries, S. Lautensack, O.Romer, Irradiation programme HFR phase 1b: results ofimpact tests with the irradiated materials (2.4 dpa), FZKA6976, Forschungszentrum Karlsruhe, 2004.

[95] R.L. Klueh, J.M. Vitek, J. Nucl. Mater. 132 (1985) 27.[96] E. Lucon, M. Decreton, E. van Walle, Fusion Eng. Des. 69

(2003) 373.[97] A. Kimura, H. Kayano, M. Narui, J. Nucl. Mater. (1991)

737.

T. Yamamoto et al. / Journal of Nuclear Materials 356 (2006) 27–49 49

[98] C. Liu, H. Klein, P. Jung, J. Nucl. Mater. 335 (2004) 77.[99] P. Fernandez, A.M. Lancha, J. Lapena, M. Hernandez-

Mayoral, Fusion Eng. Des. 58–59 (2001) 787.[100] Y. Ruan, P. Spaetig, M. Victoria, J. Nucl. Mater. 307–311

(2003) 236.[101] J. Kameda, X. Mao, J. Mater. Sci. 27 (1992) 983.[102] R.L. Klueh, J.M. Vitek, J. Nucl. Mater. 150 (1987) 272.

[103] S. Jitsukawa, A. Kimura, A. Kohyama, R.L. Klueh, A.A.Tavassoli, B. van der Schaaf, G.R. Odette, J.W. Rensman,M. Victoria, C. Petersen, J. Nucl. Mater. 329–333 (2004) 39.

[104] P. Spatig, G.R. Odette, E.G. Donahue, G.E. Lucas,J. Nucl. Mater. 283–287 (2000) 721.

[105] E.G. Donahue, G.R. Odette, G.E. Lucas, J. Nucl. Mater.283–287 (2000) 637.