Evaluation of Feasibility of Tungsten/Oxide DispersionStrengthened Steel Bonding with Vanadium Insert

Hiroyuki Noto1,+1, Akihiko Kimura2,+2, Hiroaki Kurishita3, Satoru Matsuo3 and Shuhei Nogami4

1Graduate School of Energy Science, Kyoto University, Uji 611-0011, Japan2Institute of Advanced energy, Kyoto University, Uji 611-0011, Japan3International Research Center for Nuclear Materials Science, Institute for Materials Research (IMR),Tohoku University, Ibaraki 311-1313, Japan4Graduate School of Engineering Department of Quantum Science and Energy Engineering,Tohoku University, Sendai 980-8579, Japan

A diffusion bonding (DB) technique to reduce thermal expansion coefficient mismatch between tungsten (W) and oxide dispersionstrengthened ferritic steel (ODS-FS) was developed by applying a vanadium (V) alloy as an insert material. In order to suppress · phaseprecipitation at the interface, DB of ODS-FS and V­4Cr­4Ti was carried out by introducing a Ti insert as a diffusion barrier between V­4Cr­4Tiand ODS-FS, and examined feasibility of W/V/Ti/ODS-FS joint for application to fusion reactor components by comparing the three-pointbending strength and microstructure between the joints with and without a Ti diffusion barrier layer. It is shown that the fracture strength of thejoint without a Ti insert was decreased by 25% after aging at 700°C for 100 h, but that with a Ti insert shows no change after the aging treatmentup to 1000 h. The result indicates that the introduction of a Ti insert leads to the prevention of the formation of · phase during aging and resultantcontrol of the degradation of the bonding strength. [doi:10.2320/matertrans.MG201213]

(Received October 24, 2012; Accepted February 7, 2013; Published March 25, 2013)

Keywords: oxide dispersion strengthened steel, diffusion bonding, · phase, barrier layer, bonding strength

1. Introduction

Tungsten (W) is considered as a leading candidate for theplasma facing material (PFM) or components (PFC) of ITER(International Thermonuclear Experimental Reactor) becauseof its very high melting point, superior resistance tospattering, etc. However, W is well known to be a brittlematerial, especially in the recrystallized state anticipated afterhigh heat loads during operation.1­3) Recently, innovativelytoughened W­1.1mass%TiC alloy in the recrystallized statewas developed in Tohoku University.4)

Oxide dispersion strengthened (ODS) steel is also wellknown as one of the advanced structural materials for thefusion system with high thermal efficiency, since ODSsteel has superior mechanical properties at high temper-atures to the previous reduced activation ferritic steels.5)

Among the ODS steels, ODS ferritic steels (FS) have ahigh potential for application to high-temperature structuralcomponents because of no ¡/£ transformation up to theliquid phase temperature. High performance 15%Cr­4%Al-ODS-FS was developed in Kyoto University, of whichthe oxide particles were reduced in size and increased innumber density by simultaneous additions of small amountof Ti and Zr.6) It is considered at this point that theapplication of W­1.1%TiC and ODS-FS to the divertersystem will provide the best combination for high thermalefficiency. Therefore, the technology development of thebonding between W­1.1%TiC and ODS-FS is cruciallyrequired.

A serious problem in bonding of ODS-FS and W materialsincluding W­1.1%TiC is a large difference in the coefficientof thermal expansion (CTE) between them,7,8) which causes

large thermal stresses at the bonding area and may inducecracking of W or cracks along joint boundary after bonding.In order to reduce the thermal stresses at the joint, aninsertion of insert material between W and ODS-FS can beeffective through thermal stress reduction.9) Therefore, theoptimization of new insert material is required.

In this study, a vanadium (V) alloy was selected as aninsert material for the joint because it exhibits the followingadequate properties that are needed for use as the insertmaterial between W and ODS-FS: (1) the CTE of V alloy isbetween W and ODS-FS, V alloy has (2) a higher thermalconductivity than ODS-FS, (3) a reduced activation charac-teristics, (4) a high solubility in W, and (5) nano-structured Valloy has radiation tolerance.10)

On the other hand, the application of V alloy as theinsert material between W and ODS-FS may lead to theprecipitation of Fe/V ·-phase that is known to be brittle, andtherefore the ·-phase formation is a critical issue for use of Valloy as an insert material of joints of ferritic steels. In thisstudy, diffusion bonding of V­4%Cr­4%Ti (V­4Cr­4Ti) andODS-FS was carried out at 1250°C, which is high enough toavert ·-phase formation during bonding. However, long timeoperation at around 700°C that is in the range of anticipatedtemperatures in fusion systems may cause embrittlement by·-phase formation.

One of the methods to suppress the formation of ·-phaseis another insertion of material to form multi inserted insertmaterial. Ti is a low activation element and possesses theCTE similar to V and a high solubility in V. Also, thediffusivity of V in Ti is rather small. It is expected that thelow diffusivity of V in Ti can retard the migration of V intoFe as a result of an insertion of Ti between them.

The objective of this study is to investigate the effect of aninsertion of Ti insert material on the formation of ·-phaseat the interface of V alloy and ODS-FS and the resultant

+1Graduate Student, Kyoto University+2Corresponding author, E-mail: kimura@iae.kyoto-u.ac.jp

Materials Transactions, Vol. 54, No. 4 (2013) pp. 451 to 455Special Issue on Materials-System Integration for Fusion DEMO Blanket©2013 The Japan Institute of Metals and Materials

changes in the bonding strength and microstructuresassociated with DB at 1250°C as well as aging at 700°C.

2. Experimental Procedures

The compositions of ODS-FS, V­4Cr­4Ti, Ti foil andamorphous foil used in this study are listed in Table 1. Aplate of ODS-FS, having superior swelling resistance andhigh temperature strength, was fabricated by mechanicalalloying, hot extrusion and hot rolling.5) V­4Cr­4Ti (NIFS-Heat2), having very low oxygen and nitrogen contentsand superior fracture toughness,11) was supplied from theNational Institute for Fusion Science (NIFS). The Ti andamorphous foils were commercially available, of which thethickness was 0.020 and 0.025mm, respectively. The Ti foilwas used as a barrier layer against inter-diffusion between Feand V: The amorphous foil contains melting point depressantelements (B, Si) for Fe, which may accelerate the inter-diffusion between Ti and ODS-FS.

Each of the ODS-FS and V­4Cr­4Ti plates was cut into acubic which measured 10mm © 10mm © 10mm. The bond-ing surfaces were mechanically polished with waterproofabrasive papers of #500, 800, 1200, 2400 and 4000 andbuff-polished with diamond paste with particle size of1.0 µm. For bonding experiments, a pair of cubic specimenswith an amorphous foil without a Ti barrier layer or with thebarrier layer, as shown in Fig. 1, was rapped with a Ta foil(thickness: 0.025mm) to protect the specimens from oxygenor nitrogen pick-up during bonding. The bonding wasperformed with a hot press machine at 1250°C and 10MPafor 1 h in a vacuum of 8.0 © 10¹4 Pa.

Each joint was cut into rectangular bars of 2mm ©3mm © 25mm for mechanical bending tests and 5mm ©10mm © 25mm (with notch) for XRD analyses of the joinedinterface of which the notch was introduced to reveal thejoined interface between ODS-FS and V­4Cr­4Ti as afracture surface.

Aging treatments were performed at 700°C for 50, 100,200, 1000 h for the joints rapped with a Ta foil and a Zr foilfor protection from oxygen or nitrogen contamination duringaging heat treatment.

The bonding strengths of the joints and aged specimenswere evaluated by three-point bending tests at room temper-ature with 13.2mm in span at a displacement rate of1 © 10¹3mm/s. The microstructures before and after theaging treatment were investigated by XRD and EPMAanalyses. The hardness distribution of the joint crossing theinterface was measured by nano-indentation method.

3. Results and Discussion

Before the practical fabrication of W/ODS-FS joint, afinite element method (FEM) analysis was carried out toevaluate internal stress (elastic stress) distribution generatedin a W/ODS-FS joint bonded at 1250°C followed by coolingto room temperature (27°C). Figure 2 shows the internalstress distribution on the cross sectional area, which isdescribed as ① in the figure of 1/4 of the whole joint.Comparing the Von-Misses stress in the W/ODS-FS jointwithout a V­4Cr­4Ti insert (a) to that with the insert (b),the internal stress built in (a) reaches approximately 2GPain maximum, whereas that in (b) reduces to approximately1.5GPa. Thus, the V alloy is an effective insert material forrelaxing the elastic stress caused by the difference in thethermal expansion coefficient between W and ODS-FS.Actually, the W­1.1%TiC/ODS-FS joint with a V-alloyinsert material was successfully produced, and no crack wasformed in the joint. However, it is expected that ·-phase willcause embrittlement during service at elevated temperatures.

Figure 3 shows the three-point bending stress­straincurves at room temperature for V­4Cr­4Ti/ODS-FS withand without a Ti barrier layer. The fracture strength, which isthe maximum fiber stress, ·, is estimated by the followingequation:

· ¼ 3PL=2Bt2 ð1ÞHere, P is the applied load, L is the span, B and t are thespecimen width and thickness, respectively. The fracturestrength before aging was 260MPa, which was maintained atthe same strength level until 50 h aging, regardless of a Tibarrier layer. After 100 h aging, however, the strengths wereaffected by the insertion of a Ti barrier layer. A reduction of

Table 1 Chemical compositions of materials used in this work (mass%).

Fe Cr W Ti Y O V Si B N C

ODS-FS Bal. 15 2 0.2 0.25 <0.05 <0.1 <0.01 <0.03

V­4Cr­4Ti 4.02 4.13 <0.02 Bal. <0.03 <0.02 <0.01

Ti foil Bal.

Amorphous foil Bal. 5 3

Fig. 1 Illustration of a set of joining materials including barrier insert usedfor preparing a diffusion-bonded V­4Cr­4Ti/ODS-FS joint.

H. Noto, A. Kimura, H. Kurishita, S. Matsuo and S. Nogami452

fracture strength by about 25% was recognized for thespecimens without a Ti barrier layer, but no furtherdegradation occurs after 1000 h aging. The degradation islikely due to embrittlement by ·-phase formation, as shownlater. On the other hand, for specimens with a Ti barrier layerno degradation was induced by aging. In addition, thefracture strengths of the joins with a Ti barrier layer showedan appreciable increase instead of any decrease even after1000 h of aging. The stress­strain behavior, however,indicated that the fracture mode of the joints was brittlebecause fracture occurred in the elastic deformation regionwithout showing any plasticity.

In order to clarify the possible causes of the above differentfracture strength, the joints were cut perpendicularly to thejoined interface of ODS-FS/V­4Cr­4Ti, and the revealedsurfaces (cross sectional areas) were examined.

Figures 4 to 7 show the SEM micrographs of the crosssectional area and the hardness profiles measured by nano-indentation methods (a), together with elemental mapping byEPMA (b). In the case of the joint without Ti insert beforeaging (Fig. 4), an inter-diffusion layer of approximately75 µm in thickness was observed in the ODS-FS adjacentto the interface of ODS-FS/V­4Cr­4Ti, and the hardnessremarkably increased in the layer. The EPMA results shown

in Fig. 4(b) indicate that the inter-diffusion layer contains V(see also Fig. 8). For the joint without Ti insert after theaging for 50 h (Fig. 5), the inter-diffusion layer broadensto approximately 95 µm in thickness: the growth of thediffusion layer produced by the aging for 50 h is not verysignificant.

In the case of the joint after the aging for 100 h withouta Ti insert (Fig. 6), an inter-diffusion layer similar to thoseshown in Figs. 4 and 5 is observed. The nano-indentation testresult also shows that the inter-diffusion layer is hardenedand may act as a crack initiation site.

In the case of the joint with a Ti insert after the aging for50 h (Fig. 7), on the other hand, the inter-diffusion layerindicative of the Fe/V ·-phase was not observed. This meansthat the Ti barrier insert has a beneficial effect of suppressingthe formation of the Fe/V ·-phase. Instead, another inter-diffusion layer, which was much thinner than the layer of theFe/V system with a higher hardness than each of ODS-FSand V­4Cr­4Ti, was observed. The formation of the thin andhard layer was confirmed for the joint before aging and noappreciable growth of the layer was observed after the agingfor 50 h.

Fig. 3 Effect of Ti diffusion barrier insert on three-point bend fracturestrength at room temperature for V­4Cr­4Ti/ODS-FS joints aged at700°C up to 1000 h.

(a) (b)

Fig. 4 (a) SEM image of the cross sectional area of V­4Cr­4Ti/ODS-FSjoint without Ti diffusion barrier insert before aging, showing hardnessprofile measured by nano-indentation method (inter-diffusion layer of75µm in thickness) and (b) elemental mapping by EPMA.

[MPa]

W

W

ODS - FS

V - 4Cr - 4Ti (1mm)

ODS - FS

X

Z

X

Z

(a)(b)

Fig. 2 Distribution of Von Misses stress (MPa) at room temperature in W/

ODS-FS joint bonded at 1250°C followed by cooling to room temper-ature, (a) without V­4Cr­4Ti insert material and (b) with V­4Cr­4Tiinsert material (thickness: 1mm).

(a) (b)

Fig. 5 (a) SEM image of cross sectional area of V­4Cr­4Ti/ODS-FS jointwithout Ti insert after aging at 700°C for 50, showing hardness profilemeasured by nano-indentation method (inter-diffusion layer of 95µm inthickness) and (b) elemental mapping by EPMA.

Evaluation of Feasibility of Tungsten/Oxide Dispersion Strengthened Steel Bonding with Vanadium Insert 453

In order to identify the structure of the thin and hard layer,XRD analyses were made for the fracture surface that wasoriginally a joined interface and revealed by fracturing atthe pre-introduced notch. Figure 8 is a result for 1000 h-agingwith and without the Ti insert. In the case where the barrierinsert exists, peaks of Fe2Ti are observed, and thereforethe thin and hard inter-diffusion layer in Fig. 8 is almostidentified to be Fe2Ti compounds.

The experimental results on bonding strength and micro-structural observation of the joints with and without the Tibarrier insert suggest that the role of the thin Fe2Ti layer isdifferent from that of the thick Fe/V ·-phase: It is noted thatthe thin Fe2Ti layer does not seem to have the detrimentaleffect of decreasing bonding strength by aging treatments,and thus we can say that the introduction of a Ti insert as abarrier is effective in fabricating joints of ODS-FS/V­4Cr­4Ti without the formation of ·-phase. In addition, the fracturestrength of ODS-FS/V­4Cr­4Ti joints with the barrier insertincreased with increasing aging time at 700°C. This resultcan be regarded as a benefit of the relaxation of CTEmismatch between V­4Cr­4Ti and ODS-FS, which occurred

at 700°C and increased with aging time. More investigationsare needed to clarify the different behavior between the thinFe2Ti layer and thicker Fe/V ·-phase layer.

An estimation of diffusion distance of constituting alloyelements gives us an insight into alloy mixing behavior

(a) (b)

Fig. 7 (a) SEM images of cross sectional area of V­4Cr­4Ti/ODS-FS joint with Ti insert after aging at 700°C for 50 h, showing hardnessprofile measured by nano-indentation method and (b) elemental mapping by EPMA.

Fig. 8 XRD patterns of fracture surfaces of V­4Cr­4Ti/ODS-FS joint withand without Ti insert after aging at 700°C for 1000h.

(a) (b)

Fig. 6 (a) SEM image of cross sectional area of V­4Cr­4Ti/ODS-FS joint without Ti insert after aging at 700°C for 100 h, showinghardness profile measured by nano-indentation method and (b) elemental mapping by EPMA.

H. Noto, A. Kimura, H. Kurishita, S. Matsuo and S. Nogami454

during bonding and aging at 700°C for 1000 h. The diffusiondistance (x) was roughly estimated by the equation; x =(2Dt)1/2, where D is diffusion coefficient of element inmatrix metal, t is the diffusion period. Table 2 summarizesthe diffusion distances of each element in metals for 1000 h.The estimated diffusion distance of Fe in Ti is close to 1 µm,thus likely that Fe2Ti compounds are promptly formed, butthat of V in Ti is 0.054 µm, resultantly Ti insert acts as adiffusion barrier layer for V. In the case of the joint withouta Ti insert, the inter-diffusion rate between Fe and V isexpected to be very low and the formation of Fe­Vcompound is considered to be suppressed even after1000 h-aging. This indicates that the Ti foil works as aneffective barrier insert. In particular, solid state diffusionbonding (SSDB) used in this study, unlike welding, neverresults in the formation of welded zone whose width, forexample, in electric beam welding is reported to be as largeas 1.0mm,12) and the amount of the ·-phase formed duringSSDB would be small. It is, therefore, considered that evenwhen the hardened narrow layer exists, the amount of Fe2Ticompound is small enough and the decrease in the fracturestrength by the aging for 1000 h is suppressed to be only 25%reduction.

4. Conclusions

Vanadium (V) has desirable characteristics required for aninsert material between W and ODS-FS that have significantdifference in the coefficient of thermal expansion. However,the use of a V insert may result in ·-phase formation dueto reaction with ODS-FS in the bonding zone. In this study,we prepared diffusion bonded joints of V­4Cr­4Ti and ODS-FS with and without a Ti diffusion barrier insert placedbetween them to prevent ·-phase formation. We examinedthe effects of the insertion of a Ti barrier insert on thebonding strength by three point bending tests before andafter aging at 700°C up to 1000 h. The main results are asfollows:

(1) The diffusion bonded ODS-FS/V­4Cr­4Ti joint with-out Ti barrier insert exhibited a decrease in fracture strengthby 25% by aging at 700°C for 100 h, compared with the jointbefore aging, although no further decrease was observed afteraging for 1,000 h. It is reasonable to say that the degradationof the bonding strength caused by the aging at 700°C is notvery significant.

(2) The diffusion bonded ODS-FS/V­4Cr­4Ti specimenswith Ti diffusion barrier insert exhibited no decrease in the

fracture strength up to 1000 h, and instead the fracturestrength increased with increasing aging time.

(3) Observations in the cross section of the joints withoutTi barrier indicate that Fe/V inter-diffusion phases wereformed and the increase in the hardness was noticeable in thediffused region. The width of the Fe/V inter-diffusion layerincreased with increasing aging time.

(4) The degradation of the joints without Ti barrier layerwas caused by a ·-phase in the Fe/V system, while theinsertion of a Ti insert as a diffusion barrier resulted in nodegradation of the joint strength. This is explained in terms oflower diffusivity of V in Ti and the Ti plays a role of barrierfor inter-diffusion of V and Fe. The increase in fracturestrength with increasing aging period can be regarded as abenefit of the relaxation of CTE mismatch between V­4Cr­4Ti and ODS-FS.

(5) In this study, we have suppressed ·-phase formationbetween V­Cr­4Ti and ODS-FS by the insertion of a Tiinsert, and resultantly the degradation of the bonding strengthhas been suppressed. On the other hand, the joints withouta Ti insert did not exhibit further significant decrease infracture strength after aging at 700°C up to 1000 h. It canbe stated that V alloys are applicable to bonding of W andODS-FS as an insert material where the relaxation of CTEmismatch is a primary concern.

Acknowledgments

The authors would like to express their gratitude toInternational Research Center for Nuclear Materials Science,IMR, Tohoku University for the help with the presentexperiments through the joint-use research. This study ispartly supported by LHD collaborative research of NationalInstitute of Fusion Science and US-Japan collaborativeresearch on Fusion Materials, TITAN.

REFERENCES

1) I. Smid, M. Akiba, G. Vieider and L. Plöchl: J. Nucl. Mater. 258­263(1998) 160­172.

2) W. W. Basuki and J. Aktaa: J. Nucl. Mater. 429 (2012) 335­340.3) H. Kim, J. El-Awady, V. Gupta, N. Ghoniem and S. Sharafat: J. Nucl.

Mater. 386­388 (2009) 863­865.4) H. Kurishita, S. Matsuo, H. Arakawa, T. Sakamoto, S. Kobayashi, K.

Nakai, T. Takida, M. Kato, M. Kawai and N. Yoshida: J. Nucl. Mater.398 (2010) 87­92.

5) A. Kimura et al.: J. Nucl. Mater. 417 (2011) 176­179.6) A. Kimura et al.: Proc. 10th ICAPP, (Atomic Energy Society of Japan,

2009) pp. 9220-1­9220-8.7) N. Oono, S. Noh, N. Iwata, T. Nagasaka, R. Kasada and A. Kimura:

J. Nucl. Mater. 417 (2011) 253­256.8) T. G. Wang, J. L. Chen, Y. Chen and Y. C. Wu: J. Nucl. Mater. 363­365

(2007) 1294­1298.9) H. E. Pattee: Bull. Weld. Res. Council 178 (1972) 1­42.10) T. Sakamoto, H. Kurishita, T. Furuno, T. Nakagawa, S. Kobayashi, K.

Nakai, S. Matsuo, H. Arakawa, A. Nishimura and T. Muroga: Mater.Sci. Eng. A 528 (2011) 7843­7850.

11) T. Muroga, T. Nagasaka, K. Abe, V. M. Chernov, H. Matsui, D. L.Smith, Z.-Y. Xu and S. J. Zinkle: J. Nucl. Mater. 307­311 (2002) 547­554.

12) S. Nogami, J. Miyazaki, A. Hasegawa, T. Nagasaka and T. Muroga:Proc. 15th Int. Conf. on Fusion Reactor Materials (J. Nucl. Mater).

Table 2 Diffusion distance of elements that were contained in the bondingmetals (700°C, 1000h).

Diffusion atom Metal Diffusion distance (µm)

Fe Ti 925

Ti Fe 16.78

V Ti 0.054

Fe V 0.174

V Fe 12.98

Evaluation of Feasibility of Tungsten/Oxide Dispersion Strengthened Steel Bonding with Vanadium Insert 455