preparation and neutronic studies of tungsten carbide composite

PREPARATION AND NEUTRONIC STUDIES OF TUNGSTEN CARBIDE COMPOSITE T. DASH,a B. B. NAYAK,a* M. ABHANGI,b R. MAKWANA,b S. VALA,b S. JAKHAR,b C. V. S. RAO,b and T. K. BASUb a

Institute of Minerals and Materials Technology, Bhubaneswar-751013, Odisha, India Institute for Plasma Research, Bhat, Gandhinagar-382428, India

b

Received March 20, 2013 Accepted for Publication July 2, 2013 http://dx.doi.org/10.13182/FST13-663

Because of their desirable structural properties, WC, WCzB4C, and WC z TiC are possible materials for use in plasma-facing components of fusion reactors like tokamaks. In this work, seven different compositions of WC-W2C composites have been prepared (30 to 50 at. % C) by an arc plasma melting technique followed by furnace cooling. Efforts have been made to produce a composite that is very hard and tough and that has a high neutron absorbing capacity by adding B4C and TiC (5 to 15 wt% each) to the starting WC powder. Microstructures of the composites were studied by field emission scanning electron microscopy and transmission electron microscopy. Multiphasic structures of the composites exhibited an absence of pores. The WC z TiC and WC z B4C composites showed improvements in microhardness over pure WC. Typical samples of WC-W2C, WC z B4C, and

WC z TiC have been characterized by X-ray diffraction, X-ray photoelectron spectroscopy, and Brunauer-EmmettTeller techniques for analysis and correlation of material properties. When irradiated with 14-MeV D-T neutrons, it was observed that the pure WC melt-cast product exhibited a linear neutron absorption coefficient of 0.172 cm{1. The absorption coefficient was found to be a maximum (0.255 cm{1) for 5 wt% B4C added to WC as against Type 316LN stainless steel, which showed a value of 0.078 cm{1.

I. INTRODUCTION

properties, such as a high melting point (*2750uC), high hardness (16 to 22 GPa), high Young’s modulus (550 GPa), high fracture toughness (28 MPa?m1/2 when bonded with Co), high compressive strength (5 GPa at 20uC), high thermal conductivity [84 W/(m?K)], and high resistance to oxidation and corrosion.5,6 Some of these properties, like high Young’s modulus, high compression strength, and high hardness, may be useful to reduce mechanical degradation of the material due to displacement of atoms caused by neutron radiation. In industrial practice, fused tungsten carbide, a composite of WC and W2C, is prepared by melting WC and then allowing the melt to cool down to room temperature. The phase diagram of the W-C system shows a pseudoeutectic point at 2530uC and exhibits eutectoid decomposition below 2380uC to produce

Investigations into interactions of fast neutrons with potential structural materials of nuclear reactors are very important for a variety of specific purposes in the development and application of fission and fusion reactor techniques, for example, in the design of fusion power plants and fusion experimental devices such as ITER (Refs. 1 and 2). Nowadays, WC is considered a candidate material, in lieu of W, because the presence of C (a reasonably good moderator material for high-energy neutrons) would improve the neutron shielding performance of W (Refs. 3 and 4). WC is endowed with several superior material *E-mail: [email protected] FUSION SCIENCE AND TECHNOLOGY

VOL. 65

KEYWORDS: tungsten carbide, arc plasma, neutron absorption coefficient Note: Some figures in this paper may be in color only in the electronic version.

MAR./APR. 2014

241

Dash et al.

PREPARATION AND NEUTRONIC STUDIES OF TUNGSTEN CARBIDE COMPOSITE

device that takes not more than 10 to 15 min to attain a high temperature of the above order. The paper focuses on a neutron absorption study of some typical samples of WC-W2C and WC z B4C and WC z TiC composites using fast neutrons (14-MeV energy). Characterizations of typical samples were carried out using X-ray diffraction (XRD), transmission electron microscopy (TEM), field emission scanning electron microscopy (FESEM), X-ray photoelectron spectroscopy (XPS), Brunauer-Emmett-Teller (BET) analysis, and a microhardness tester.

II. EXPERIMENTAL

Fig. 1. Phase diagram of W-C system.5,7

a mixture of WC and W2C (Fig. 1) (Refs. 5 and 7). While some studies in the literature report higher hardness values for WC, some mention higher hardness values for W2C (Refs. 8 and 9). The carbide of boron (B4C) is a hard compound that is commonly used in nuclear applications as a neutron radiation absorbent.2,10 Titanium carbide (TiC) is also a hard compound exhibiting higher hardness than WC, and it is added to Co-bonded WC tool bits (for cutting metal and stone) to improve resistance to wear, corrosion, and oxidation.11 It is expected that if TiC and B4C are added in the WC-W2C matrix, a very hard, wear resistant, and high neutron absorbing composite material will result that would improve neutron shielding performance in toroidal and poloidal field coils and improve the thickness of the blank shield assembly. In the present work, an effort has been made to prepare seven different compositions of WC-W2C composites (30 to 50 at. % C) and six different compositions of WC z B4C and WC z TiC (5, 10, and 15 wt% of B4C or TiC) composites with very low porosity using an arc plasma melting method due to several advantages in comparison to induction and graphite furnace melting methods.5,12,13 While an induction furnace and a graphite furnace are relatively more expensive than an arc plasma furnace and take several hours to reach a high temperature (*2750uC, the melting point of WC), an arc plasma furnace, in contrast, is a very fast melting 242

Pellets made of WC z W, WC z B4C, and WC z TiC were prepared using the respective powders and employing polyvinyl alcohol as the binder. Uniaxial compaction of the pellets was carried out by applying pressures of 140.44 MPa for W z WC and 156.06 MPa for WC z B4C and WC z TiC for 30- to 60-s duration. The W and WC starting powders were procured from M/s Rapicut Carbides Ltd., India, and the B4C and TiC powders were procured from M/s Himedia Laboratories Pvt. Ltd. The specifications of the powders are as follows. W powder: grain size, 4 to 5 mm; WC powder: grain size, 4 to 7 mm; C weight percentage in WC: 6.11 to 6.16 (bound C), 0.05 max (free C); B4C powder: grain size, 68 mm; TiC powder: fine size (v10 mm). In situ melt-castings of air-dried (4 to 5 h) pellets were carried out in a 30-kW (direct-current) extended arc– type thermal plasma reactor using graphite crucible, and Ar was employed as the plasmagen gas. Ar (flow of 1.5 ,/min) was used to avoid oxidation in the melt-cast composites during the melting and cooling processes. XRD patterns of the typical composites (studied in powder form) were recorded by employing a PANalytical X’Pert Pro diffractometer (CuKa, l 5 0.15406 nm). Microstructures of melt-cast samples were observed by FESEM (ZEISS-SUPRA 55, backscattered mode) and TEM (TECNAI G2, 200 kV, FEI, The Netherlands). Surface area and volume of the pores in the composites were determined using the BET method (ASAP 2020) using the N2 adsorption-desorption technique. Vickers microhardness was evaluated by a LECO microhardness tester equipped with a diamond pyramid indenter at a load of 0.5 kg (for WC-W2C composites) and 0.1 kg (for WC z B4C and WC z TiC composites) with a dwell time of 13 s. For binding energy studies, XPS (S/N: 10001, Prevac, Poland) spectra were taken with AlKa (hn 5 1486.6 eV) radiation and a hemispherical energy analyzer. An instrument base pressure of 6|10{10 mbars was maintained during data acquisition. Neutron absorption studies of the typical composites were carried out using the 14-MeV sealed D-T neutron generator available at the Institute for Plasma Research,

FUSION SCIENCE AND TECHNOLOGY

VOL. 65

MAR./APR. 2014

Dash et al.


Gandhinagar (IPR), emitting 2|109 n/s. The neutron flux was measured using a neutron activation technique. The neutron source was first standardized by measuring the neutron flux (Cu was the neutron activation foil) at the four sides of the neutron generator tube (Fig. 2a). It was observed that approximately the same amount of neutron flux (flux * 1.38|107 n/cm2?s) was emitted at the surface on all four sides. After standardization, neutron irradiation of samples was carried out. The meltcast composite samples were placed at the top surface of the neutron generator tube, one Fe foil (activation foil) was placed over it, and another Fe foil was placed at one side of the source tube in such a way that its distance from the tube surface is equal to the thickness of the meltcast sample (Fig. 2b). The time of irradiation of the samples was 600 to 660 s. After irradiation, neutron activations were counted in the foils employing a wellcalibrated high-purity germanium detector and a multichannel analyzer to measure neutron-induced gamma activity. The foil-to-detector distance was maintained at 2 cm.

III. RESULTS AND DISCUSSION

It is well known that B4C is a good neutron absorber due to its high neutron absorption coefficient. Sawan3 reports the largest shielding improvement results from WC balls used in a vacuum vessel (packing fraction of 62 vol %) and finds a reduction in the magnet heating and insulator dose by a factor of 3 to 5. Neutron fluence and Cu damage are reduced by 30% in the case of WC (Ref. 3). The performance of WC is better than a material like Type 316LN stainless steel. Tungsten metal is also considered for various plasma-facing components because of its high melting point, high thermal conductivity, high resistance to sputtering and erosion, and good fast neutron attenuation characteristics.2,14 As a result, W is a primary candidate material for divertor components in fusion reactors (tokamaks). Keeping all these issues in sight and following a W-C phase diagram, as shown in Fig. 1, samples of seven different compositions were prepared in this investigation for WC-W2C composites

Fig. 2. Schematic view of the arrangements made for neutron irradiation: (a) for standardization of neutron generator and (b) for arc plasma melt-cast sample studies. FUSION SCIENCE AND TECHNOLOGY

VOL. 65

(50 to 70 at. % W / 50 to 30 at. % C) by varying the weight percentage of W in the WC matrix as follows: W, 0 to 55.5 wt%; WC, 100 to 44.5 wt%; and C, 6.14 to 2.72 wt%. The atomic percentages of C in the composites were as follows: WC-1 (50 at. % C); WC-2 (48.6 at. % C); WC-3 (45.4 at. % C); WC-4 (40 at. % C); WC-5 (38 at. % C); WC-6 (33.4 at. % C); and WC-7 (30 at. % C). The B4C and TiC additions were made using the arc plasma melting technique to prepare six different WC z B4C and WC z TiC composites (5, 10, and 15 wt% of B4C or TiC) comprising each type with three different samples. The XRD patterns (Fig. 3) of WC-W2C composites show that WC and W2C are the major phases that grow by our plasma method. Carbon was observed to grow as the minor phase in both the graphite phase [C(G)] and the diamond phase [C(D)], except for the WC-6 (33.4 at. % C) sample, where C(G) was the major phase. The crystal lattices of WC [space groupP6m2 (187)], W2C [space group-P3lm (162)], C(G) [2H polytype, space group-P63/mmc (194)] are hexagonal and that of C(D) [space group-Fd3m (227)] is cubic. From the observed XRD results, it is confirmed that the in situ melt-cast solid samples are WC-W2C composites. The FESEM microstructure of the WC-1 (50 at. % C) sample presented in Fig. 4a (at relatively less magnification) shows only two phases: white and gray, while its TEM picture (Fig. 4b), depicted at higher magnification, shows a polycrystalline microstructure of three types of phases comprising white, semi white/gray, and dark grains, depending upon the carbon content. The microstructures in Figs. 4a through 4d, by observations made both at lower and higher magnifications, clearly demonstrate that our arc plasma melt-cast samples are free from pores and surface defects. Such porosity-free samples grow by the melt-cast technique when the density of the liquid is very high, as happened in the present case (density of WC: 15.63 g/cm3). The XPS spectra of the WC-5 (38 at. % C) sample were recorded for the W4f, C1s, and O1s regions, and they are presented in Figs. 5a, 5b, and 5c, respectively. Deconvolution of the W4f region (Fig. 5a) displays two peaks at 31.6 and 33.7 eV, which are assigned to W4f7/2 and W4f5/2, respectively; they correspond to both hexagonal phases of WC and W2C (Ref. 15). The complex spectrum of C1s is deconvoluted into four distinct peaks, as shown in Fig. 5b. The peaks at 282.6 and 283.7 eV are attributed to carbidic carbon bonded to the W phases of WC and W2C, respectively.16 The other two peaks at 284.9 and 285.9 eV may be related to graphite carbon (sp2 bonded C) and diamond-like carbon (sp3 bonded C) (Ref. 16). The narrow spectrum of O1s (Fig. 5c) at 531.5 eV is related to OH{ (Ref. 17) due to the exposure of the sample in the air atmosphere. The XPS results match well with our XRD findings. Thus, XPS spectra again confirm that our samples are composed of WC and

MAR./APR. 2014

243

Dash et al.


Fig. 3. XRD patterns of arc plasma melt-cast tungsten carbide composite samples at different W-to-C ratios.

Fig. 4. FESEM and TEM microstructures of typical melt-cast samples for (a) FESEM of WC-1 (50 at. % C) sample, (b) TEM microstructure of WC-1 (50 at. % C) sample, (c) TEM micrograph of WC-8 (WC z 5 wt% TiC) sample, and (d) TEM micrograph of WC-10 (WC z 5 wt% B4C) sample. 244


VOL. 65

MAR./APR. 2014

Dash et al.


Fig. 5. XPS spectra observed for arc plasma melt-cast WC-5 (38 at. % C) sample: (a) spectra in the W4f region, (b) spectra in the C1s region, and (c) spectra in the O1s region.

W2C. Both XRD and XPS results of WC-5 (38 at. % C) allow us to conclude that the sample does not contain any oxide phase due its preparation conducted in an inert atmosphere (Ar). FUSION SCIENCE AND TECHNOLOGY

VOL. 65

BET surface area, Langmuir surface area, and pore volume of the WC-W2C (for all seven samples) composites were found mostly in the ranges 0.105 to 0.373 m2/g, 0.147 to 0.561 m2/g, and 0.001 to 0.004 cm3/ g, respectively. BET surface area, Langmuir surface area, and pore volume of the WC z TiC and WC z B4C composites were in the ranges 0.012 to 0.712 m2/g, 0.014 to 1.033 m2/g, and 0.005 to 0.013 cm3/g, respectively. These results are supported by the findings in the microstructures (Figs. 4a through d), and we may conclude that our samples are free from microporosity. The pure WC (WC-1: 50 at. % C) and eutectic composition WC-5 (38 at. % C) samples show microhardnesses in the ranges 1670 to 1708 VHN0.5 and 1818 to 1890 VHN0.5, respectively. When 5 wt% B4C and TiC each is added to pure WC, the microhardness in WC-10 (WC z5 wt% B4C) and WC-8 (WC z 5 wt% TiC) samples is seen to increase by 33.85% (1850 to 2582 VHN0.1) and 18.67% (1730 to 2100 VHN0.1), respectively. The increase in hardne occurs due to significantly higher hardness of B4C and TiC over WC. Neutronic studies of typical arc plasma melt-cast composite samples are presented in Table I. Neutron flux was measured both experimentally and also using an MCNP simulation code with and without our sample at 1- to 1.5-cm distance from the 14-MeV neutron generator. From Table I, it is observed that our experimentally determined results very closely match the MCNP simulation code results. Less neutron flux is observed when our sample is loaded at the S2 foil; i.e., the neutron flux emitted from the S2 foil (Fe) with the sample is less than that of the S1 foil (Fe) without the sample. Among the samples evaluated, the linear neutron absorption coefficient is found to be a maximum for WC-10 (WC z 5 wt% B4C), i.e., 0.255 cm{1, and a minimum for the Type 316LN stainless steel sample, i.e., 0.078 cm{1. The absence of any porosity in the melt-cast composite matrix and high neutron absorption cross section of boron (10B) together have contributed toward exhibition of higher neutron absorption coefficient values in the WC zB4C composites.10,18 Pure WC (WC-1: 50 at. % C) and the eutectic composition sample WC-5 (38 at. % C) show linear neutron absorption coefficients of 0.172 and 0.175 cm{1, respectively. It is noticeable from Table I that addition of TiC in WC reduces the linear absorption coefficient value (to 0.159 cm{1) of the composite. This may be attributed to (a) a poor scattering cross section of Ti for the high-energy neutrons (14 MeV) and (b) a poor capture reaction and absorption cross section of Ti for neutrons at lower energy (v1 MeV). Carbon is a reasonably good neutron moderator with a mean logarithmic reduction of neutron energy per collision (j) value of 0.1589, which is common in all composite samples studied here.

MAR./APR. 2014

245

Dash et al.


TABLE I Neutron Absorption Results of Typical Arc Plasma Melt-Cast Composite Samples Sample Neutron Activation and Absorption Property Details

WC-1 (50 at. % C)

WC-5 (38 at. % C)

WC-8 (WC z 5 wt% TiC)

WC-10 (WC z 5 wt% B4C)

Thickness of the sample (cm) Neutron flux activation by S1 (Fe) foil (experimentally determined) (n/cm2?s) Neutron flux activation by S2 (Fe) foil (experimentally determined) (n/cm2?s) Neutron flux activation by S1(Fe) foil (calculated by MCNP code) (n/cm2?s) Neutron flux activation by S2 (Fe) foil (calculated by MCNP code) (n/cm2?s) Neutron absorption (experimentally determined) (%) Linear neutron absorption coefficient (experimentally determined) (cm{1)

1 3.64|106

1 2.56|106

1.5 2.03|106

1 2.36|106

3.07|106

2.15|106

1.60|106

1.83|106

3.06|106

3.14|106

2.73|106

3.13|106

2.77|106

2.77|106

2.33|106

2.83|106

15.66

16.02

0.172

0.175

IV. SUMMARY

22.46

0.159

0.255

REFERENCES

An arc plasma melt-cast method was used to prepare WC-W2C and WC z TiC and WC z B4C composites. Characterization studies carried out by FESEM, TEM, and BET measurements show the nonporous and surfacedefect-free nature of the cast ingots. Growth of WC and W2C in composite form is concluded from XRD and XPS studies. With an addition of 5 wt% of B4C and TiC each in WC, the microhardness is found to increase by 33.85% and 18.67%, respectively. Linear absorption coefficients of the neutrons are found to be 0.255 cm{1 for the WC z 5 wt% B4C sample as against 0.078 cm{1 for Type 316LN stainless steel—a significant achievement, which is attributed to the high neutron absorption coefficient of 10B and the absence of pores coupled with reduced defects in the material.

1. C. THOMSER et al., ‘‘Plasma Facing Materials for the JET ITER-Like Wall,’’ Fusion Sci. Technol., 62, 1 (2012). 2. J. MATEˇJI’CˇEK et al., ‘‘Thermal Spray Coating for Fusion Applications—Review,’’ J. Therm. Spray Technol., 16, 1, 64 (2007); http://dx.doi.org/10.1007/s11666-006-9007-2. 3. M. SAWAN, ‘‘Radiation Shielding Requirements for Magnets in Fusion Reactors,’’ Proc. 4th Japanese-Mediterranean Workshop Applied Electromagnetic Engineering for Magnetic, Superconducting, Multifunctional, and Nanomaterials (JAPMED’4), Cairo, Egypt, September 17–20, 2005. 4. B. G. HONG et al., ‘‘Conceptual Design Study of a Superconducting Spherical Tokamak Reactor with a SelfConsistent System Analysis Code,’’ Nucl. Fusion, 51, 113013-1 (2011); http://dx.doi.org/10.1088/0029-5515/51/11/113013. 5. T. DASH et al., ‘‘Preparation of WC-W2C Composites by Arc Plasma Melting and Their Characterisations,’’ Ceram. Int., 39, 3279 (2013); http://dx.doi.org/10.1016/j.ceramint.2012. 10.016.

ACKNOWLEDGMENTS The authors T. Dash and B. B. Nayak are thankful to the Board of Research in Fusion Science and Technology, IPR, for its support for carrying out the reported research in WC-W2C and WC z B4C and WC z TiC composites. They are also thankful to the Director, Council of Scientific and Industrial Research, Institute of Minerals and Materials Technology (CSIR-IMMT), Bhubaneswar, for permission to carry out the work in the institute. 246

21.18

6. Z. WU et al., ‘‘Silica-Templated Synthesis of Ordered Mesoporous Tungsten Carbide / Graphitic Carbon Composites with Nanocrystalline Walls and High Surface Areas Via a Temperature-Programmed Carburization Route,’’ Small, 23, 2738 (2009); http://dx.doi.org/10.1002/smll.200900523. 7. E. RUDY, ‘‘The Phase Diagram of W-B-C,’’ AFML-TR69-117, Part V, p. 6, Wright-Patterson Air Force Base (1970).


VOL. 65

MAR./APR. 2014

Dash et al.


8. S. W. H. YIH et al., Tungsten Sources, Metallurgy and Applications, p. 387, Plenum Press, New York (1981). 9. A. MUKHOPADHYAY et al., ‘‘Recent Developments on WC-Based Buck Composites,’’ J. Mater. Sci., 46, 571 (2011); http://dx.doi.org/10.1007/s10853-010-5046-7. 10. T. MARUYAMA et al., ‘‘Effect of Fast Neutron Irradiation on the Properties of Boron Carbide Pellet,’’ J. Nucl. Sci. Technol., 34, 10, 1006 (1997); http://dx.doi.org/10.1080/ 18811248.1997.9733777.

14. H. RENNHOFER et al., ‘‘Neutron Flux from a 14-MeV Neutron Generator with Tungsten Filter for Research in NDA Methods for Nuclear Safeguards and Security,’’ AIP Conf. Proc., 1194, 36 (2009); http://dx.doi.org/10.1063/1.3275663. 15. R. OSPINA, ‘‘Influence of the Annealing Temperature on a Crystal Phase of W/WC Bilayers Grown by Pulsed Arc Discharge,’’ Vacuum, 81, 373 (2006); http://dx.doi.org/10. 1016/j.vacuum.2006.06.004.

11. H.-C. KIM et al., ‘‘Sintering Behavior and Mechanical Properties of Binderless WC-TiC Produced by Pulsed Current Activated Sintering,’’ J. Ceramic. Proc. Res., 8, 2, 91 (2007).

16. H. M. AJMERA, ‘‘Deposition of WNxCy for Diffusion Barrier Application Using the Imido Guanidinato Complex W(NiPr)Cl3[iPrNC(NMe2)NiPr],’’ J. Vac. Sci. Technol. B, 26, 5, 1800 (2008); http://dx.doi.org/10.1116/1.2981082.

12. B. B. NAYAK et al., in Proc. 20th Int. Conf. Composites/ Nano Engineering, Beijing, China, July 22–28, 2012.

17. M. FLORI et al., ‘‘A Study of Steel Surface Contamination by XPS Electron Spectroscopy,’’ J. Eng., V, 165 (2007).

13. A. SAHU et al., ‘‘DC Extended Arc Plasma Nitriding of Stainless and High Carbon Steel,’’ J. Mater. Sci., 35, 71 (2000); http://dx.doi.org/10.1023/A:1004736431235.

18. C. SUBRAMANIAN et al., ‘‘Development of Boron-Based Materials for Nuclear Applications,’’ BARC Newsletter, 313, 14 (2010).


VOL. 65

MAR./APR. 2014

247