Integrated Material System Modeling of Fusion Blanket

A. Sagara1,+1, R. Nygren2, M. Miyamoto3, D. Nishijima4, R. Doerner4, S. Fukada5,Y. Oya6, T. Oda7, Y. Watanabe8,+2, K. Morishita8, F. Gao9 and T. Norimatsu10

1National Institute for Fusion Science, Toki 509-5292, Japan2Sandia National Laboratories, Albuquerque, NM 87123-1129, USA3Shimane University, Matsue 690-8504, Japan4University of Californian, San Diego, CA 92093-0417, USA5Kyushu University, Fukuoka 812-8581, Japan6Shizuoka University, Shizuoka 422-8529, Japan7Department of Materials Science and Engineering, University of Tennessee, Knoxville, USA8Institute of Advanced Energy, Kyoto University, Kyoto 611-0011, Japan9Pacific Northwest National Laboratory, Richland, WA 99352, USA10Institute of Laser Engineering, Osaka University, Suita 5650871, Japan

The blanket of fusion reactors is a multifunctional system that breeds tritium, harvests heat from the burning plasma, and protects the othercomponents and the environment. The common task in the US-J TITAN (Tritium, Irradiation and Thermofluid for America and Nippon) projectidentifies cross-linked considerations in the blanket system modeling on the basis of the material research in each task. In this paper, we reviewthe main outputs of this task: elucidation of the effect of helium on deuterium retention in the tungsten wall, analysis of tritium transfer in a Pb­Li liquid breeder, experimental and computational studies on the effects of radiation damage on hydrogen trapping, and modeling of the tritiumbarrier in a double-tube heat exchanger. [doi:10.2320/matertrans.MG201210]

(Received October 19, 2012; Accepted February 6, 2013; Published March 15, 2013)

Keywords: blanket, tungsten, tritium, irradiation

1. Introduction

Fusion systems of magnetically confined fusion energy(MFE) or inertial fusion energy (IFE) are enclosed within ablanket system that breeds tritium, converts the energy offusion output into heat for power conversion, transfers thisheat with high efficiency, and provides nuclear shielding tothe in-vessel components and the environment. Thus, theblanket system must be optimized not as a single functionbut as an integrated multifunctional system incorporating theboundary conditions of burning core plasmas, plant require-ments for power conversion, and tritium handling (involvingsafety issues and environmental protection). The authors haveworked on common-task integrated system modeling underthe framework of the Japan­US Joint Research ProjectTITAN (Tritium, Irradiation and Thermofluid for Americaand Nippon).1) In this project, cross linking between taskswas regarded as a multiphysics modeling problem for theoptimal design of a fusion blanket, as shown in Table 1.

For each task, tritium transfer, thermofluid mechanisms,and irradiation synergy are investigated under typical reactorconditions: plasma-wall interactions with heat and particles,nuclear and chemical reactions in the blanket material, MHDeffects in high magnetic field, heat and mass transfer in eachsubsystem of the heat exchanger and tritium separationsystem, and the intermittent effects of intense transient heatloads. The time constants for tritium and material behaviorsare separated and evaluated from these tasks, and thecharacteristics of the system components are verified by anexperiment.

The common task creates a self-consistent model byintegrating the separate systems of tritium, heat andthermofluid. At the same time, via feedback of importantconsiderations to each task, the validity of our experimentalresults is assessed in the interfaces between research goalsin each task, in order to promote the development of anintegrated system modeling methodology.

In this review, we focus on material research in Tasks 1­1;1­2; and 2­1, 2­2, 2­3 (Table 1). The common critical issuein MFE and IFE in terms of fuel self-sufficiency andenvironmental safety is tritium mass transfer through theheat exchangers. Therefore, a typical heat-exchanger system,designed to identify appropriate boundary conditions formaterial researches, is also discussed.

2. Helium (He) Effect on Deuterium (D) Retention inTungsten

A critical consideration in fusion reactor design is thetritium recycling properties of the first wall of the fusionblanket. The effects of helium on the retention of hydrogenisotopes in tungsten have been extensively reported, butremain contentious. While some studies have reportedthat retention is enhanced by helium implantation, othershave reported the opposite.2­10) These discrepancies may beattributed to differences in experimental conditions. Theresults of several previous studies are shown in Figs. 1 and 2,and numerical data are summarized in Table 2. Results for Dretention by energetic ion implantation (open symbols inFigs. 1 and 2) are consistent across studies, with D retentionsaturating at around 3 © 1020D+m¹2, even for He+ fluenceas large as 1022He+m¹2. Given that D retention in theabsence of He+ is ³1 © 1020D+m¹2, we observe that He+

+1Corresponding author, E-mail: sagara.akio@LHD.nifs.ac.jp+2Present address: Japan Atomic Energy Agency, Aomori 039-3212, Japan

Materials Transactions, Vol. 54, No. 4 (2013) pp. 477 to 483Special Issue on Materials-System Integration for Fusion DEMO Blanket©2013 The Japan Institute of Metals and Materials OVERVIEW

1020 1022 1024 10261018

1019

1020

1021

D r

eten

tion

, RD /

D m

-2

He fluence, ΦHe

/ He+ m-20

Y.Oya [2]S.Nagata [3]H.Iwakiri [4]Y. Sakoi [5]H.T.Lee [6]B.I.Khripunov [7]H.T.Lee [8]M.Miyamoto [9]D.Nishijima [10]Y. Sakoi [5]

Open: Ion irradiationFixed: Plasma exposure

(He 700 K)

(He 1600 K)

Fig. 1 D retention in ion-implanted or plasma-exposed tungsten, as afunction of He+ fluence.

1018 1020 1022 1024 10261018

1019

1020

1021

D r

eten

tion

, RD /

D m

-2

D fluence, ΦD / D+ m-2

Open: Ion irradiationFixed: Plasma exposure

Y.Oya [2]S.Nagata [3]H.Iwakiri [4]Y. Sakoi [5]H.T.Lee [6]B.I.Khripunov [7]H.T.Lee [8]M.Miyamoto [9]D.Nishijima [10]Y. Sakoi [5]

(He 700 K)

(He 1600 K)

Fig. 2 D retention in ion-implanted or plasma-exposed tungsten, as afunction of D+ fluence.

Table 2 Summary of conditions of various He+ and D+ implantation and plasma exposure experiments.

Ion energy:HeD

Ion flux:!He/He+m¹2 s¹1

!D/D+m¹2 s¹1

Ion fluence:)He/He+m¹2

)D/D+m¹2

Implantationtemperature

Y. Oya3.0 keV3.0 keV

(0.2­1.8) © 1018

1.0 © 1018(0.2­1.8) © 1022

1.0 © 1022R.T.

S. Nagata10 keV1keV

®

®

4.2 © 1018­6.0 © 1021

3.0 © 1021R.T.

H. Iwakiri8 keV4keV

®

®

(0.1­2.0) © 1021

1019­1022R.T.

Y. Sakoi3 keV3keV

5.0 © 1018

1.0 © 10181.0 © 1021

1.0 © 1021R.T.

H. T. Lee500 eV500 eV

1018­1019

1.0 © 10191021­1023

(0.5­2.0) © 1023300K

B. I. Khripunov4MeV250 eV

®

®

(1.0­3.0) © 1022

(0.27­1.3) © 1026373K

H. T. Lee500 eV500 eV

1018­1019

1.0 © 10193.0 © 1021

2.0 © 1023300K

M. Miyamoto120 eV120 eV

5.0 © 1020

1.0 © 10222.5 © 1024

5.0 © 1025573K

D. Nishijima20­25 eV80 eV

(0.25­4.8) © 1022

4.0 © 1021(1.8­9.0) © 1021

3.0 © 1025700­1600K (He)550K (D)

Y. Sakoi10 eV3keV

1.0 © 1021

1.0 © 10183.0 © 1022

1.0 © 1021573K (He)R.T. (D)

Table 1 Task structure and research items in the TITAN project. The common task integrates system modeling by connecting each taskunder the many boundary conditions existing in the fusion blankets.

Task Subtask Facilities Research items

Task 1Tritium andmass transferblanket

1­1 Tritium and mass transfer in first wallSTAR/TPEPISCES

Tritium retention and transfer behavior and mass transfer in first wall

1­2 Tritium behavior in blanket systems STAR Tritium behavior through elementary systems of liquid blankets

1­3 Flow control and thermofluid modeling MTOR Flow control and thermofluid modeling under strong magnetic fields

Task 2Irradiationsynergism

2­1 Irradiation-tritium synergism HFIR STARIrradiation effects on tritium retention and transfer behavior in firstwall and structural materials

2­2 Joining and coating integrity HFIRSynergy effects of simultaneous production of tritium and helium onjoining and coating integrity

2­3 Dynamic deformation HFIREffects of irradiation and simultaneous production of tritium andhelium on dynamic deformation of structural materials

Common TaskSystem integrationmodeling

MFE/IFE system integration modelingIntegration modeling for mass transfer and thermofluid through firstwall, blanket and recovery systems of MFE/IFE

A. Sagara et al.478

implantation enhances D retention by two- to three-fold.Above a fluence of 1 © 1019He+m¹2, He exerts no furtherenhancing effect on D retention. While He bubble formationand radiation damages create potential D trapping sites,higher density and larger He bubbles may enhance desorptionof D by suppressing its diffusion toward the bulk tungsten.The lack of annealing effects in ion implantation experimentswould have reduced the ion flux, thereby contributing to Dretention in the implant region. In different plasma exposureexperiments, D retention is quite scattered even at the sameD+ fluence, as shown in Fig. 1. A strong mediator of Dretention in tungsten is the formation of He bubbles andoccupation of trap sites by He atoms. As previouslymentioned,10) tungsten exposed to He plasma at a hightemperature (1600K) exhibits reduced He retention capacity,while additional exposure to D plasma exposure induces Dtrapping. However, exposure to lower-temperature He plasmaenhances the formation of He bubbles and He retention,leading to reduced D retention.

Under Task 1­1, the effects of He on D retention oftungsten were clarified in D + He mixed plasma exposureexperiments, conducted in the Plasma Interaction withSurface and Components Experimental Simulator (PISCES)facility at the University of California, San Diego.9,11)

Complementary ion irradiation experiments were performedat Shimane University.5) Transmission electron microscopy(TEM) observations of prethinned W samples exposed tomixed D + He plasma in PISCES revealed the formation ofhigh-density nanosized He bubbles in the near surface region,accompanied by suppression of blister formation andsignificantly reduced D retention.9) The volume fraction ofbubbles, estimated from TEM cross-sectional observationsand ellipsometric measurements, exceeds the percolationthreshold, beyond which bubbles interconnect at the nearsurface region.11) The percolating bubble clusters provide adiffusion path to the surface for D atoms exposed to theplasma, allowing them to escape. D retention is similarlyreduced under the sequential irradiation of helium anddeuterium ions. Preirradiation of 3 keV He+ at high fluence(above 1.0 © 1023He+/m2) results in a drastic reduction ofD retention. On the other hand, at lower He fluences (up to1.0 © 1022He+/m2), retention increases with increasing Hefluence.5) This suggests that defects induced by He irradiationmay trap D atoms at lower He fluences, where He bubblesare not yet interconnected.

To verify the abovementioned experimental observations,we consider that multiscale modeling, combined with varioussimulation techniques (e.g., molecular dynamics simulationand finite element method), is required. Multiscale modelingcan capture a wide range of spatial and temporal phenomenaoccurring in real-time fusion reactors. For example, energyis transferred from implanted D to target W atoms in afemtosecond timeframe, while He bubbles are expected toform and migrate over a much longer period, given that theimplantation depth of D in W and the diffusion distance of Din W are on the order of several millimeters.

3. Tritium Absorption and Diffusion in Pb­Li

Task 1­2 of the TITAN project is the quantitative analysis

of tritium absorption and diffusion in a promising blanketcooling material (Pb­Li).12­17) First, the solubility of T in Pb­Li must be determined. Previous data of H isotope solubilityin Pb­Li eutectic alloys are highly scattered. Therefore, inour study, solubility measurements were carried out by twodifferent methods: (1) a constant volume method at the IdahoNational Laboratory (INL) and (2) a transient permeationmethod at Kyushu University. The former experimentdetermined T solubility in Pb­Li, while the latter determinedthe solubility and diffusivity of H and D in Pb­Li. The resultsare presented in a separate Task 1­2 paper of the specialissue: “Clarification of tritium behavior in Pb­Li blanketsystem”.18)

The second task is system design integration, with theapplication of a Pb­Li breeding material to the fusion reactorsystems. Pb­Li is proposed as a test blanket module (TBM)in ITER, although other candidates, such as He-cooledlithium lead (HCLL) and dual coolant lithium lead (DCLL),have been proposed by the European Union and UnitedStates. In Japan, a water-cooled ceramic (Li2TiO3) breederblanket is used as the ITER­TBM. Blanket designs using Pb­Li as coolant and breeder of magnetic fusion have not beenconsidered previously. On the other hand, a wet-wall Pb­Lidesign is proposed for the Koyo-fast commercial fusionreactor. A schematic of this design, with a list of designparameters, is shown in Fig. 3. Pb­Li flows into the top ofthe reactor chamber to protect the metal chamber fromdamage by heavy neutron irradiation. Besides its role as atritium breeder, Pb­Li receives heat from particles createdby D-T fusion and the n-6Li reaction within the blanket.The energy conversion and tritium recovery systems in theexternal Pb­Li flow for IFE are essentially same as those forMFE, except that the flowing Pb­Li directly comes in contactwith the high-temperature plasma on the wet wall of thevacuum chamber. Tritium concentration in the Pb­Li flow iscontrolled by the tritium generation rate in the chamber andby the Pb­Li flow rate. The Pb­Li temperature is maintainedbetween 300 and 500°C. Majority of the tritium is recoveredby a Pb­Li­He counter-current flowing system, but a smallportion escapes from the outlet of the recovery system intothe heat exchanger. Tritium balance and permeation ratethrough heat-exchanger tubes were initially estimated as255Ci/day. We consider that this comparatively high tritiumpermeation rate can be mitigated by coating the heatexchanger tubes with ceramic. A 1000-fold reduction inpermeation implies a tritium leakage rate below 1Ci/day,which complies with the safety standards. Experimentalverification of permeation reduction is presented in the Task1­2 section of the above mentioned paper.

4. Effect of Radiation on Tritium Retention in TungstenExposed to Ion and Neutron Irradiations

The effect of radiation on tritium accumulation in tungsten(a promising candidate for the first wall of fusion blankets)has attracted much attention since radiation damage is likelyto increase tritium retention. Most previous research hasfocused on ion irradiation rather than neutron irradiation;however, the formation of radiation damage structuresdepends on the irradiation conditions. Hence, accurate

Integrated Material System Modeling of Fusion Blanket 479

prediction of tritium retention in fusion reactors requirescareful analysis and interpretation of these experimentalresults, which has not necessarily been accomplished to date.In Task 2­1, the effects of neutron and ion irradiation wereinvestigated by computer simulations19) as well as byirradiation experiments with neutrons and ions.20­22)

We developed a simulation code based on the MonteCarlo technique for modeling the accumulation and releasebehaviors of hydrogen isotopes interacting with vacancies intungsten.19) The code utilizes the results of first-principlecalculations as well as data obtained in previous experiments.It can integrate systems containing vacancies from externalsources, for example, from TRIM code. We then evaluatedthe behavior of hydrogen loaded by plasma exposure or ionirradiation. The results are presented as depth profiles intungsten and thermal desorption spectra. From the simulationresults, hydrogen introduced by plasma exposure was foundto be localized in a region of high defect density createdby ion irradiation of tungsten. In contrast, in neutron-irradiated tungsten, hydrogen was spread over a wide region(Fig. 4). Thermal desorption behavior of introduced hydro-gen also differed between ion-irradiated and neutron-irradiated samples, consistent with the experiment.20) Thesedifferences are attributed to a difference in the vacancydistribution between ion-irradiated (localized distribution)and neutron-irradiation (uniform distribution) systems. Infuture work, the accuracy of simulation results will beimproved by (1) balancing long-time irradiation processeswith a rapid diffusion process, (2) preventing unrealisticaccumulation of hydrogen, and (3) modeling the releaseof hydrogen forcibly loaded into a region already containinghigh hydrogen density.

5. Multiscale Modeling of Microstructural Change inSiC during Irradiation

The main component of SiC/SiC composites, used asblanket structural materials for nuclear fusion reactors, iscubic silicon carbide (¢-SiC). Fusion reactor materials aresubject to various point defects such as vacancies (V), self-

interstitial atoms (SIAs), and those induced by helium andhydrogen atoms displaced by high-energy incident neutronsfrom the fusion core plasma. These processes occur onpicosecond-order timeframes and across nanometer-orderdistances. The resulting defects thermally migrate and formdefect clusters (SIA clusters, V clusters and cavities) oversubmicrosecond times and across submicrometer distances.The microstructural changes caused by defect clustersdegrade the performance of the material; therefore, suchchanges should be accurately predicted and controlled.Material response to irradiation is inherently a multiscalephenomenon, as described above, and should be modeled bymultiple complementation of experimental and computationaltechniques over appropriate time and distance scales.23,24)

To understand the microstructural changes in ¢-SiC duringirradiation, a variety of transmission electron microscopyexperiments have been conducted on material test fission

Fig. 4 Monte Carlo simulation results on distribution of deuteriumimplanted into ion-irradiated or neutron irradiated tungsten at 473K.Deuterium implantation depth is around 5 nm, corresponding to a plasmaexposure of 200 eV deuterium. The label “15­10 nm” signifies that thevacancy distribution is centrally localized at 15 nm with FWHM 10nm(assuming normal distribution), and “uniform” signifies a uniformdistribution of defects. The localized and uniform distributions correspondto ion and neutron irradiation, respectively. The peak trap concentration isset to 4% for all conditions.19)

Fig. 3 Li16Pb84 loop and steam Rankine cycle in a Laser fusion reactor.

A. Sagara et al.480

reactors and ion accelerators. While these experiments haveelucidated the species and sizes of defect clusters undervarious irradiation conditions,25,26) the formation mechanismof defect clusters remains poorly understood. In the presentstudy, the kinetics of defect cluster formation in ¢-SiCwere numerically evaluated through a multiscale modelingapproach, focusing on the nucleation and growth processesof SIA clusters. We note that SiC is a binary compoundcomprising silicon and carbon atoms; therefore, defectclusters of various chemical compositions could be formed,depending on the irradiation conditions. Modeling defectcluster formation in a compound material requires specialcare, as discussed below.

First, the formation, binding and migration energetics of¢-SiC defects were investigated by classical moleculardynamics (MD) and molecular statics (MS), combined withthe method of Gao­Weber empirical interatomic potential,27)

which employs not only experimental observations but alsoab-initio calculations. In the MD and MS calculations, theformation energy of the most energetically favorable SIAclusters was derived as a function of size and chemicalcomposition ratio (Si/C) of the clusters.28,29) Knowledgeof the formation energy is crucial for evaluating bindingenergy between defects since the binding energy correspondsto the thermal stability of the defects. Migration energiesof isolated silicon and carbon interstitials (SIAs) were alsoderived.30)

On the basis of the defect energetics obtained from theatomistic calculations, the nucleation and growth processes ofSIA clusters were then investigated by the kinetic MonteCarlo (KMC) method. Here the KMC model constructedallows for statistical fluctuations in the inflow/outflow ofisolated SIAs into/from an SIA cluster. In the KMCsimulations, the ratios of the diffusion fluxes between siliconand carbon interstitials (DI

SiCISi: DI

CCIC) were set as 1 : 1,

1 : 10 and 1 : 100, while the total diffusion flux (DISiCI

Si +DI

CCIC) was kept constant at 1.0 © 1019m2/s. The formation

kinetics of SIA clusters in ¢-SiC during irradiationwere roughly classifiable into two temperature-dependentclasses.31) At relatively high temperatures, the thermalstability of an SIA cluster is crucial, and the chemical

composition of the cluster is almost stoichiometric (i.e.,Si/C = 1), as shown in Fig. 5(a). In contrast, at relativelylow temperatures, where cluster thermal stability is nolonger crucial, the composition of SIA clusters can deviatemarkedly from stoichiometric (Fig. 5(b)). Such informationis very important for prediction of the microstructuralchanges in compound materials during irradiation. As such,it will form the basis of a model incorporating reaction-rate-theory analysis and phase-field method, for evaluatingmicrostructural changes over a prolonged time and lengthscales.

6. Leakage Control of Tritium through Heat Cycles ofFusion Reactor

Tritium leakage by diffusion through heat exchangersoccurs in both MCF and IFE. While retaining the systemefficiency, the tritium leakage must be reduced by a factorof 1/1010 to reach that of a current fission plant. To reducetritium permeation from the primary liquid metal or sodiumloop into the secondary water loop, a heat exchangerincorporating small-diameter tubes containing an oxidizerwas proposed.32) An inert gas containing a small amount ofoxidizer flows through the tubes, oxidizing tritium inter-cepted from the primary liquid metal coolant. The tritiatedwater was conveyed to a tritium recovery system, minimizingleakage into the secondary water loop. Evaluation of thisdesign indicated that the tritium leakage through the heatexchanger was reduced by 1/105, with an acceptable increasein the size of the heat exchanger. This scheme is compatiblewith the coating technique, using compounds such as Er2O3

and ZrO2, which may further reduce permeation by a factorof 1/104.18)

The permeation rate and chemical form of tritium afterpermeation are known to depend on surface conditions andtemperature. Figure 6 shows a simplified model of tritiumpermeation through the wall of a heat exchanger. A virtualgap with zero thickness between the wall and the liquid LiPbis assumed. The tritium concentration in metals and thetritium partial pressure in the gap are assumed to be relatedby Sieverts’ law. Following the permeation of wet surfaces,

0

10

20

30

40

501:10

1:1

Num

ber

of C

-int

erst

itia

ls in

an

SIA

-clu

ster

Number of Si-interstitials in an SIA-cluster

1:100

0 10 20 30 40 500 10 20 30 40 500

10

20

30

40

50

1:1

1:10

Num

ber

of C

-int

erst

itia

ls in

an

SIA

-clu

ster

Number of Si-interstitials in an SIA-cluster

stoichiometric(Si/C=1)

Each path closely follows the diagonal line.

973K(a)

C-rich

673K(b)

DISi CISi : DIC CIC = 1:100

SiC

stoichiometric(Si/C=1)

Fig. 5 KMC simulations of nucleation and growth path of SIA clusters in ¢-SiC at (a) 973K and (b) 673K.

Integrated Material System Modeling of Fusion Blanket 481

almost all of the tritium atoms form HTO through isotopeexchange reactions. However, some tritium forms T2 whenthe surface is dry and the temperature is high. Henry’s law isassumed to relate the tritium concentration in pressurizedwater to the T2 in the vapor phase.

Figure 7 shows cross sections of the steam tube proposedin this study. The thick walls of the tube are fitted withsmall-diameter tubes filled with an oxidizer and carrier gas.Tritium permeating from the outer LiPb will be oxidized onthe surfaces of the small tubes and conveyed to the tritiumrecovery system, thereby reducing the amount of tritiumreaching the inner surface of the steam pipe.

In the following calculations, the partial pressures oftritium in the outer and inner virtual gaps are assumed as1 Pa and zero, respectively. The permeation rate is 8.3 ©1012 atomsm/m2Pa1/2 s at 745K. The solubility and thermalconductivity are 7 © 1021 atoms/m3 Pa1/2 and 20W/mK,respectively. The apparent thermal conductivity and perme-ation of tritium from outer to inner surfaces are calculatedusing a commercially available finite element code (ANSYS).

Figure 8 shows the normalized tritium permeation througha heat exchanger maintained at a constant heat flow. Theexpected reduction of tritium flow is approximately 1/105.

Applying this system to the heat cycle of a 1GWe plant,the tritium concentration in the second water loop is predictedas 400MBq/cm3 after one year of full power operation.Therefore, for safety reasons, a tritium recovery system must

be designed for the second loop. Almost all of the injectedtritium can be recovered and reused as fuel after isotopeseparation.

7. Summary

The common task in the US-J TITAN project identifiescross-linked considerations in the blanket system modelingon the basis of the material research in each task. The resultsare summarized as follows:(1) The helium effects on hydrogen isotope retention in

tungsten were investigated by reviewing the results ofprevious studies and by conducting D + He mixedplasma exposure experiments in PISCES. These resultsreveal that D retention is reduced by high-density Hebubbles forming in the near surface region.

(2) Research into permeability, diffusivity and solubility ofhydrogen isotopes in a Pb­Li alloy as a tritium breederis summarized. The blanket design must achieve a lowtritium leak to the outside and high tritium recoveryfrom a breeder loop.

(3) A simulation code to model the accumulation andrelease of hydrogen isotopes interacting with vacanciesin tungsten was developed using a Monte Carlotechnique. Simulation results revealed that the behaviorof hydrogen isotopes depends strongly on the modeof irradiation (i.e., whether the specimen is ion- orneutron-irradiated).

(4) Multiscale modeling of microstructural changes in abinary compound material during irradiation haselucidated the nucleation and growth processes ofdefect clusters.

(5) Tritium permeation through a heat exchanger equippedwith double tubes filled with an oxidizer can be reducedto 1/104 that of bare stainless tubes, without degradingthe heat exchange rate.

Acknowledgment

This work is supported by the Japan­US cooperationprogram TITAN sponsored by Japan-MEXT and US-DOE.

10-5

0.0001

0.001

0.01

0.1

1

Base pipe 3mm x 12tubes 3mm x 24tubes 3mm x 30tubes

Normalized permeation per unit lengthNormalized permeation per system

Number of oxidizer tubes

No

rmal

ized

per

mea

tio

n p

er u

nit

len

gth

Fig. 8 Normalized tritium permeation through a heat exchanger equippedwith different configurations of double pipes. The heat flow through theexchanger is constant.

LiPb

SUS316Water

Ar+O2(a)

LiPb+T

Water

Ar+O2

Ar+O2+T2O(b)

Fig. 7 Double tube design for the heat exchanger. Cross section (a)perpendicular to the water flow (a) and (b) parallel to the flow.

Vapor

Virtual gas phasewith zero volume

LiPb

water

SUS316

LiPb

water

SUS316

T2T22THTO

Vapor

T2 2T HTO

Sievert law

Henrylaw

Fig. 6 Model of tritium permeation through a wall of a heat exchanger.

A. Sagara et al.482

REFERENCES

1) T. Muroga et al.: Fusion Sci. Technol. 60 (2011) 321­328.2) Y. Oya, M. Kobayashi, R. Kurata, W. Wang, N. Ashikawa, A. Sagara,

N. Yoshida, Y. Hatano and K. Okuno: J. Nucl. Mater. 415 (2011) S701­S704.

3) S. Nagata and K. Takahiro: J. Nucl. Mater. 290­293 (2001) 135­139.4) H. Iwakiri, K. Morishita and N. Yoshida: J. Nucl. Mater. 307­311

(2002) 135­138.5) Y. Sakoi, M. Miyamoto, K. Ono and M. Sakamoto: J. Nucl. Mater.,

in press, http://dx.doi.org/10.1016/j.jnucmat.2012.10.003.6) H. T. Lee, A. A. Haasz, J. W. Davis, R. G. Macaulay-Newcombe, D. G.

Whyte and G. M. Wright: J. Nucl. Mater. 363­365 (2007) 898­903.7) B. I. Khripunov et al.: J. Nucl. Mater. 415 (2011) S649­S652.8) H. T. Lee, A. A. Haasz, J. W. Davis and R. G. Macaulay-Newcombe:

J. Nucl. Mater. 360 (2007) 196­207.9) M. Miyamoto, D. Nishijima, Y. Ueda, R. P. Doerner, H. Kurishita, M. J.

Baldwin, S. Morito, K. Ono and J. Hanna: Nucl. Fusion 49 (2009)065035.

10) D. Nishijima, T. Sugimoto, H. Iwakiri, M. Y. Ye, N. Ohno, N. Yoshidaand S. Takamura: J. Nucl. Mater. 337­339 (2005) 927­931.

11) M. Miyamoto, D. Nishijima, M. J. Baldwin, R. P. Doerner, Y. Ueda, K.Yasunaga, N. Yoshida and K. Ono: J. Nucl. Mater. 415 (2011) S657­S660.

12) Y. Edao, S. Fukada, S. Yamaguchi and H. Nakamura: Fusion Eng. Des.85 (2010) 53­57.

13) Y. Edao, H. Okitsu, H. Noguchi and S. Fukada: Fus. Sci. Technol. 60(2011) 1163­1166.

14) Y. Edao, H. Noguchi and S. Fukada: J. Nucl. Mater. 417 (2011) 723­726.

15) S. Fukada and Y. Edao: J. Nucl. Mater. 417 (2011) 727­730.16) D. Masuyama, T. Oda, S. Fukada and S. Tanaka: Chem. Phys. Lett. 483

(2009) 214­218.17) Y. Edao, S. Fukada, H. Noguchi, Y. Maeda and K. Katayama: Fus. Sci.

Technol. 56 (2009) 831­835.18) S. Fukada, T. Terai, S. Konishi, K. Katayama, T. Chikada, Y. Edao, T.

Muroga, M. Shimada, B. Merrill and D. K. Sze: Mater. Trans. 54(2013) 425­429.

19) T. Oda, M. Shimada, K. Zhang, P. Calderoni, Y. Oya, M. Sokolov, R.Kolasinski, J. P. Sharpe and Y. Hatano: Fus. Sci. Technol. 60 (2011)1455­1458.

20) M. Shimada, Y. Hatano, P. Calderoni, T. Oda, Y. Oya, M. Sokolov, K.Zhang, G. Cao, R. Kolasinski and J. P. Sharpe: J. Nucl. Mater. 415(2011) S667­S671.

21) Y. Oya, M. Shimada, M. Kobayashi, T. Oda, M. Hara, H. Watanabe, Y.Hatano, P. Calderoni and K. Okuno: Phys. Scr. T145 (2011) 014050.

22) Y. Shimada, G. Cao, Y. Hatano, T. Oda, Y. Oya, M. Hara and P.Calderoni: Phys. Scr. T145 (2011) 014051.

23) S. Sharafat and K. Morishita: J. At. Energy Soc. Jpn. 50 (2008) 724.24) K. Morishita and S. Sharafat: J. At. Energy Soc. Jpn. 50 (2008) 803.25) Y. Katoh, N. Hashimoto, S. Kondo, L. L. Snead and A. Kohyama:

J. Nucl. Mater. 351 (2006) 228­240.26) T. Yano and T. Iseki: Philos. Mag. A 62 (1990) 421­430.27) F. Gao and W. J. Weber: Nucl. Instrum. Meth. B 191 (2002) 504­

508.28) Y. Watanabe, K. Morishita and A. Kohyama: J. Nucl. Mater. 417 (2011)

1119­1122.29) K. Morishita, Y. Watanabe, A. Kohyama, H. L. Heinisch and F. Gao:

J. Nucl. Mater. 386­388 (2009) 30­32.30) Y. Watanabe, K. Morishita, A. Kohyama, H. L. Heinisch and F. Gao:

Nucl. Instrum. Meth. B 267 (2009) 3223­3226.31) Y. Watanabe, K. Morishita and Y. Yamamoto: Nucl. Instrum. Meth. B

269 (2011) 1698­1701.32) T. Norimatsu, H. Saika, H. Homma, M. Nakai, S. Fukada, A. Sagara

and H. Azechi: Fus. Sci. Technol. 60 (2011) 893­896.

Integrated Material System Modeling of Fusion Blanket 483