Floating zone partial re-melting of B4C infiltrated with molten Si

Author’s Accepted Manuscript Floating zone partial re-melting of B4C infiltrated with molten Si I. Solodkyi, I. Bogomol, P. Loboda, D. Batalu, A.M. Vlaicu, P. Badica www.elsevier.com/locate/ceri

PII: DOI: Reference:

S0272-8842(17)31647-4 http://dx.doi.org/10.1016/j.ceramint.2017.07.203 CERI15923

To appear in: Ceramics International Received date: 11 July 2017 Revised date: 27 July 2017 Accepted date: 28 July 2017 Cite this article as: I. Solodkyi, I. Bogomol, P. Loboda, D. Batalu, A.M. Vlaicu and P. Badica, Floating zone partial re-melting of B 4C infiltrated with molten Si, Ceramics International, http://dx.doi.org/10.1016/j.ceramint.2017.07.203 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Floating zone partial re-melting of B4C infiltrated with molten Si

I. Solodkyia, I. Bogomola, P. Lobodaa, D. Batalub, A.M. Vlaicuc, P. Badicac a

National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Peremogy Ave.

37, 03056Kyiv, Ukraine b

c

University Politehnica of Bucharest, Splaiul Independentei 313, 060042 Bucharest, Romania

National Institute of Materials Physics, Atomistilor 405 A, 077125 Magurele, Romania

*

Corresponding Author – Ievgen Solodkyi, Address: National Technical University of Ukraine “Igor

Sikorsky Kyiv Polytechnic Institute”, Peremogy av. 37, Kyiv 03056, Tel.: +380 66 119 18 99; fax: +380 44 204 82 15. e-mail: [email protected]

Abstract

Green compacts of B4C or B4C added with 1vol% graphite were infiltrated with molten Si and subsequently were subject of processing by floating zone partial re-melting (FZPR). In FZPR only the low temperature fusible component, in this case Si, is melted. A fully dense B4C-based ceramic is obtained. It contains free-Si, SiC and B4C. In the center of the FZPR ceramic without graphite addition, the amount of Si is decreased when compared to the infiltrated material. Some impurity elements such as Al, Fe, or Ti detected in the raw B4C powder arepreferentially gathered at the edges of the sample. In the sample added with graphite, formation of a high amount of SiC in the infiltrated material hinders Si shift from the center to the edges. The pulling rate and the particle size of the B4C raw powders are also

important. It is recognized that sintering of powders larger than 10-20 m is usually difficult: our approach is demonstrated to be suitable for processing of B4C powders with a very different particle size, from 10 to 250 m. The FZPR ceramic had a Vickers hardness of 9-38 GPa depending on location of the indentation imprint and on the sample. A tensile strength of 114 -188 MPa that is up to about 2-3 times higher than for the infiltrated material was recorded. Work indicates that the proposed processing approach offers extended control possibilities towards fabrication of new composite materials not available by traditional technologies.

Keywords: B4C-Si, infiltration, floating zone partial melting, mechanical properties, microstructure

1. Introduction

Boron carbide(B4C) is a light-weight and superhard ceramic material with a strong covalent character [1]. Its mechanical properties recommend B4C for different protection, structural and tools applications [2-5]. Boron carbide shows also good high temperature mechanical bending strength [6] and itis also of interest for thermoelectric applications [7]. Boron carbide has the drawback of being a fragile material with low fracture toughness and, considering its covalent nature, it is a difficult-to-sinter material. Pressure-assisted sintering techniques [6, 8-11] and the use of additives [1-3] were successful in obtaining high density samples when using B4C powders with particle size less than 10-20 m. Larger particles impedes sintering and high densities cannot be reached, but the price of such powders is lower and for some applications they can be convenient. Presented problems can be somehow solved if using infiltration method. Infiltration of B4C green compacts with molten Si are reported in refs. [1216]. Samples have shown high density and acceptablemechanical quality. The method is cheap and considered suitable for fabrication of large parts. On the other hand, control of the Si-melt infiltration is challenging. Some Si reacts to form SiC, but a large amount of free Si of 15 wt. % or more is present in the samples and drastically decreases mechanical properties. This situation motivates search of novel or hybrid routes to improve infiltrated ceramics. In this regard, floating zone method is one alternative

since it can realize impurities cleaning. The impurities concentrate in the melt, and are moved in the traditional methodto one end of the sample. Floating zone method was also successfully applied for the growth of B4C single crystals or of B4C-based eutectic (composite) crystals [17-19]. While forcleaning of metals and growth of crystals the entire zone is melted, our concept for infiltrated ceramics is to melt in the hot zone only the low fusible component, namely Si (melting temperature of Si is 1414 C and of B4C is 2763 C). Expectations are also that theSi-melt will act as a solvent for other impurities available in the raw powders. By this processing route, not only the amount of free Si can be controlled, but also ‘dirty’ and large-particle B4C raw powders can be used for ceramics fabrication.In many cases, industrial production of B4C raw powders is realized by metal - thermal routes, which means that the resulting powders are inherently contaminated with the additives (such as Al, Fe, Mg, C), if special cleaning measures are not enforced. Our approach of floating zone partial re-melting (FZPR) of infiltrated materials is expected to promote fabrication of novel composites with new or improved properties. The approach allows, by using an independent extra processing step (floating zone partial melting), to add plus value to a material without changes in its already established production (infiltration). This principle has significant advantages [20]. In this work we present our investigations for the assessment of a combined approach comprised of infiltration and FZPR. Silicon is infiltrated in B4C green compacts made of small (< 10 m) and large particles (50-250 m) and subsequently partially re-melted using floating zone method. We show that our hybrid technique allows significant removal of free-Si and of other impurities. Control is enabled using different pulling rates. Particle size also plays an important role. Addition of free graphite modifies phase formation towards SiC formation and this influences free-Si – related processes. High density samples of B4C-SiC-Si are obtained and this result is regardless of the B4C particle size of the raw powders. Vickers hardness shows values from 9 to 38 GPa depending on the position where the indentation imprint was made and on the sample specific features. Atensile strength of 114 -188 MPa was measured and these values are 2-3 times higher than for the infiltrated material.

2. Experimental

Two commercially available powders, denoted B4C-1 (purity 98 %, and particle size of 50250m) and B4C-2 (purity 97 % and particle size of 1-10 m) were purchased from JSC

ZaporozhAbrasive and ChimReactiv co. Ltd. (Ukraine), respectively. According to our energy dispersive spectroscopy (EDS) investigations, detectable impurity elements were Al, Ti, Cu and Fe. Some amount of free graphite was detected by X-ray diffraction (Fig. 1).Silicon powder(purity 99.0% and particle size 12-50 m) was produced by ChimReactiv co. Ltd. Green samples were prepared by mixing the powder B4C-1 (sample A-green, Table 1) or mixtures of (B4C-1 and B4C- 2) (samples B-green, Table 1) and of (B4C-1, B4C-2, and graphite) (sample Cgreen, Table 1) with a 2.5% water solution of polyvinyl alcohol (PVA). Graphite fiber with a diameter of 10-15 m was supplied by Uglecompositeco. Ltd.,Ukraine. To obtain the powder mixtures of B4C-1 and B4C-2 with/without addition of graphite (samples A-C – green, Table 1), ball milling was performed in a plastic container for 30 min. Further, samples were pressed at 20 MPa into cylinders with a diameter of 12 mm and height of 25-50 mm and dried at 100C. From dimensional and mass characteristics of the green cylinders, their porosity was determined considering the theoretical bulk density of boron carbide (2.51 g/cm3). It is 40% when using the powder B4C-1 (sample A-green, Table 1) and 34 % when using the mixture of B4C-1 and B4C-2 (samples B-green and C-green, Table 1). The result is as expected, considering that packing is poor (hence, higher porosity) for a powder composed of very large particles, as in the case of B4C-1 raw powder. Next, the green samples were infiltrated (samples notation IN) with silicon in a vacuum furnace (10 Pa). The infiltration was carried out by placing silicon compacts on the top of the porous green samples. Compacts of Si were prepared by pressing the Si powder at 20 MPa. Heating rate was of 10 C/min up to 1450 °C. Samples were maintained for 2 minat the maximum temperature. Infiltrated samples were subject to FZPR using a “Crystal 206” (USSR) furnace equipped with an induction-type heater (samples notation FZPR). We used He atmosphere at an excess pressure of 1 atm (~105 Pa). The pulling rate of thesamples through the heating zone was fixed at 5 or10 mm/min (Table 1). The density of the infiltrated samples was determined by Archimedes method (Table 1). We found that for our process the segregation of Si and other impurities occurs at the outer lateral part of the cylindrical samples. Gradients and Si flow produce 3 zones with different microstructures. Details are addressed in Section 3.1. Specimens after infiltration or FZPR were prepared for microscopy observations. They were cut to reveal the axial cross section and polishedby using sandpaper and diamond paste. For microstructural characterization, we used scanning electron microscopy (SEM). Microscopes were SELMI and Zeiss EVO50 equipped with an EDS detector.

X-ray diffraction (XRD) patterns were taken with a Rigaku Ultima IV(Japan) diffractometer (CuKα radiation). Rietveld simulation of the spectra and estimation of the volume fraction of the phases was performed with PDXL software. Vickers hardness (HV) was measured on the polished surface of the samples for 500 gf (4.9 N) loading, with a dwell time of 10 s, using a CV-400DTS Micro Hardness Tester. We applied the standard procedure according to ASTM C 1327-03. The tensile strength of the samples (10 mm in diameter, D, and 10 mm in height, W) was measured following ASTM D 3967–95a. The load, P, was applied with a rate of 200 mm/min. The tensile strength σt (MPa) is evaluated by using eq. (1):

(1).

The tensile strength value of a sample (Table 1) was averaged for three tests.

3. Results and Discussion

3.1 Phase composition and microstructure of the infiltrated samples

The microstructure and phase assembly of B4C samples infiltrated with Si was investigated in literature [12, 15, 16 and therein refs.]. As-prepared composites, also known as reaction-bonded boron carbide (RBBC) show a specific core-rim microstructure. Namely, B4C particles are surrounded by the rim B12(B,C,Si)3 phase. The other phases in the sample are cubic SiC and Si. The rim structure is formed by partial or full congruent dissolution of the original B4C particles in the molten silicon. A local equilibrium is established through the precipitation from the B-containing Si-rich liquid (Liq.) on the surface of the B4C grains and within a ‘stoichiometric saturation’ model of the rim phase B12(B,C,Si)3. SiC also precipitates from the Si melt and it has a bar-like morphology which transforms into a platelike one and finally broadens into a polygonal shape. Authors of ref. [12, 15, 16] also pointed out that the presence of free-C or the presence of organic additives in the B4C green compact (before Si

infiltration) plays an important role in formation, growth and morphology of SiC. The proposed phase diagram of the B-C-Si system at 1480 C in the Si-rich corner indicates 4 regions of equilibrium: (i) Liq + B12(B,C,Si)3, (ii) Liq + SiC + B12(B,C,Si)3, (iii) Liq + SiC, and (iv) Liq. For our infiltrated samples A-C, XRD patterns, phase composition, and SEM images are presented in Fig. 1, Table 1, and Fig. 2, respectively. Indeed our samples A and B contain phases B4C, B12(B,C,Si)3, SiC, and Si as expected for a typical RBBC, but there are also notable differences. While the total amount of boron carbide (B4C + B12(B,C,Si)3) is comparable (67 % for A-IN and 66 % B-IN), sample A-IN has a lower amount of B12(B,C,Si)3 (17 %) than for sample B-IN (45%). We note that variation of the amount of B12(B,C,Si)3 between the two samples is large (45/17=2.6 times) and at the same time for SiC is comparable or slightly lower (10 % SiC in A-IN and 7 % in B-IN, 10/7=1.4 times). Moreover, in the sample A-IN with a lower amount of B12(B,C,Si)3 (17 %), the percentage of SiC is higher (10 %) when compared with sample B-IN (45 % of B12(B,C,Si)3 and 7% of SiC). Another useful piece of information is that the amount of Si and SiC in the samples A-IN and B-IN is approximately constant (23+10=33 % in A-IN and 27+7=34% in B-IN). Presented situation suggests that although reactions and their mechanisms are similar in the two samples, they are influenced by reactivity of the boron carbide powder with Si. This depends on processing, additives (polyvinyl alcohol in our samples) and specific features of the raw B4C powders/mixtures (particle size and distribution, impurities such as different metals or free-C, morphology, see section 2 Experimental). Apparently, regions containing B4C-2 powder (sample B-IN), i.e. containing small B4C particles are more reactive (Fig. 2c), and thus a higher amount of B12(B,C,Si)3 can form in the sample. The result may also indicate that dissolution of B4C into Si melt is the rate limitative factor for the formation of B12(B,C,Si)3. In turn, this process influences precipitation of SiC. Other details such as wetting are also important in formation of the microstructure. In Fig. 2c some SiC needles or plates can be visualized in the region containing B4C small particles. Hayun et al [12] found that morphology of SiC is of needle/plates – like when carbon is not added in the sample and of a polygonal shape when carbon is added in the samples. Their samples had an average B4C particle size of 5mfor the pristine sample and of 100 m for the C-added sample. We shall remind that indicated particle sizes are like those for our raw powders B4C-2 (small particles) and B4C-1 (large particles). In our samples A-IN and B-IN,free-C was not added,but our samples in the green state contained PVA. Therefore, expectations would have been to obtain polygonal rather than needle-like SiC. Results suggest for our samples occurrence of a more complex situation than indicated by Hayun et al [12]. When carbon fibers are added into system as for sample C-IN, equilibrium shifts in the direction of preferential formation of SiC and Rietveld phase-composition-analysis supports this idea. The highest amount of SiC was determined for sample C-IN (Table 1). At the same time, in this sample it is difficult

to detect the phase B12(B,C,Si)3 by XRD. Competition between formation of B12(B,C,Si)3 and SiC is related to carbon fibers presence and of the melt state. A newcore-rim structure involving C-fibers as core and SiC/Si as shell forms (Fig. 2d): C-fiber is covered by a thin layer of Si followed by a thick one of SiC. In some fully converted C-fibers into SiC, in the center, some residual Si is available (Fig. 2d). SiC forms by a dissolution-precipitation mechanism involving the Si-rich melt. In the Si-C phase diagram, in the Si-rich region, a eutectic reaction occurs at 1404 C (L  Si+ SiC) [21].

3.2 Phase composition and microstructure of the re-melted (FZPR) samples

Samples A-IN, B-IN and C-IN were subject to FZPR. As-resulted samples were named A-FZPR, B-FZPR and C-FZPR. One B-IN sample was processed for a higher pulling rate (10 mm/min instead of 5 mm/min as for all the other samples). This sample is denoted B-FZPR(R) (Table 1). After FZPR processing, cylindrical samples were cut longitudinally, i.e. parallel to the direction of the melt-zone movement. A schematic drawing shows a FZPR sample, longitudinal cross section after cutting along the direction of the melt-zone movement (Fig. 3). We also indicated the regions I – III where the microstructure changes (Figs. 4-6) and the areas of XRD (Fig. 7) measurements denoted L longitudinal, T – transversal, and S – on the outer surface of the cylinder. One observes that XRD – L, T and S measurements reveal the phase assembly from regions I, I+II+III, and III, respectively. It is important to underline that although we performed Rietveld quantitative analysis (Table 1), data should be regarded as qualitative with a high degree of scattering. The reasons are related to phase/microstructure uniformity, thickness, and continuity of a certain region I-III. Despite these limitations some general trends can be observed. The main result is that in our FZPR process, segregation of Si and of other impurities may occur at the outer lateral part of the cylindrical samples. For all FZPR samples (Table 1, Fig. 7) the amount of free-Si measured by XRD on area (S) is higher than for (L) or (T) meaning that a higher amount is on the sample edge (region III or II + III) than in its center (region I). Remarkable is also that for samples A-FZPR, B-FZPR and B-FZPR(R) the amount of SiC measured on area (S) is larger than on area (L). This result is supported by SEM backscattering images from Fig. 4 (a, b– A-FZPR, c, d– B-FZPR, f – B-FZPR(R), and g, h– C-FZPR ). Namely, regions II and III for these samples are lighter than the central part (region I) suggesting the presence of a higher amount of heavy elements such as Si and metal impurities.

For samples A-FZPR, B-FZPR, B-FZPR(R) other observations of interest are: (i) Regions I – III show different microstructure and thicknesses depending on sample. This suggests that B4C raw powders specific features and pulling rate influences FZPR processes although their background and trends are similar. (ii) The borders between regions (Fig. 4 a, c, fand Fig. 5), especially between I and II, are clearly defined and the microstructure of the region I (center of the sample) is relatively uniform when compared with region II. (iii) Region II shows some segregation of metal impurities Al, Fe, Ti. The EDS elemental map of Al follows the map of Si (Fig. 5). Titanium has a similar tendency although less obvious as for Al. Si and Fe, Al, or Ti are known to form binary eutectics at temperatures (Si-Ti at 1330 C, Si-Al at 577 C, SiFe at 1207 C[22]) lower than the melting point of Si (1414 C). Although our data suggest that the behavior of Si and of the metal impurities is similar due to the binary eutectic relationships described by the phase diagrams, one has to take into consideration the presence of boron and carbon. This can be the case for Cu (Fig. 5) for which apparently segregation is stronger in the region I rather than in region II. Si and Cu form a eutectic at 802 C [22], and B and Cu at 1028 C [23]. (iv) For each sample of A- or B- type (Table 1) the difference between the IN and FZPR state (as measured by XRD on the (L) area) consists in a higher amount of SiC and a lower amount of free-Si for FZPR. It results that free-Si gathers at the edge of the FZPR samples as concluded before, but, at the same time, it also reacts with carbon from B4C to form SiC during FZPR processing. Sample C added with C-fibers requires a special discussion. We already noted in Section 3.1 that this sample have shown in the IN state a high amount of SiC, that is about two times or more than for samples A-IN and B-IN. After FZPR processing of sample C, although the (S) area is rich in free-Si as addressed in the previous paragraph, sample C-FZPR does not show on the (L) and (T) areas much different values for SiC amount, while the amount of free-Si is high and around 30 wt. % for both (T) and (L) measurements (Table 1, Fig. 7). The amount of SiC in the IN and FZPR (on (L) areas) states is more or less constant. This situation suggests that formation of relatively high amount of SiC in the IN state hinders FZPR processes related to free-Si observed for samples A and B. A closer look at the microstructure of this sample (Fig. 6) reveals that region II has a complex macrostructure and a compact layer containing Si, C and some impurity metals (Cu, Al) can be observed. Formation of such a layer can change the behavior and distribution of free-Si across the I-III regions. Nevertheless, closer to the edge, in the region II, apparently there is a higher amount of Ti and Fe as for samples A-FZPR, B-FZPR and B-FZPR(R). In fact, observations regarding behavior of Fe, Cu, Al made for samples A-FZPR, B-

FZPR and B-FZPR(R) are valid for sample C-FZPR. The only deviation is for Ti which shows a needlelike morphology in some cases. Again, this is not special because we have observed needles in the region II for the other A-FZPR, B-FZPR and B-FZPR(R) samples. An example is presented for sample B-FZPR in Fig. 4e. We conclude that addition of C-fibers as for sample C has important consequences in the behavior of the Si-rich melt and FZPR processes of microstructure formation. Additional general observations are: (v) Phase B12(B,C,Si)3 is detected with difficulty in the XRD patterns of the FZPR samples and it was not taken into account for Rietveld analysis. (vi) Free-Si shows some unusually high intensities for certain (hkl) peaks in the XRD patterns measured on the (L) area of the FZPR samples. This suggeststexture occurrence for Si. The texture is of (hkl) or (0kl) type, i.e. it involves the diagonal planes in the cubic structure of Si. This observation requires further confirmation.

3.3 Mechanical properties

Indentations were performed on different single phases (Fig. 8 a, b) or mixtures of phases (Fig. 8 c) from the FZPR composites. For the single phase the average Vickers hardness (HV) and fracture toughness (KIc) values for different samples were 9-11 GPa and 1.1-1.3 MPam1/2 (white phase, Fig. 8a), 15-25 and 2-3.5 MPam1/2 (gray phase, Fig. 8b), and 30-38 GPa and 2.5-3.8 MPam1/2 (black phase). These values correspond to those for Si, SiC, and B4C, respectively [24-30]. Intermediate values of HV and KIc are obtained when indentation is performed on multiphase regions (Fig. 8c). Regions denoted 1 and 2 in Fig. 8b and c show cracks formation with a wavy shape. The fracturing mechanism is mainly of intragranular type, but intergranular fracturing is also active. The cracks arrest, deflection and the intergranular mechanism reflected by pull out of the grains (Fig. 9) can be observed. The tensile strength of the IN state is enhanced for the FZPR state (Table 1). The differences for a certain sample between the two states can be up to more than 2 times as determined for B type samples. The lowest difference is for C sample. A high strength is related to the presence of a higher SiC amount in the central part (L) of the samples. The dependence is approximately linear (Fig. 10).

4. Conclusion

Samples of B4C were infiltrated with Si and further processed by FZPR. Samples are dense and can use raw powders of B4C with sizes from few microns to hundreds of microns. The phase assembly, distribution and uniformity depends on the raw powder specific features, processing parameters (e.g. pulling rate), additives (C-fibers in this work) and metal impurities (Al, Cu, Ti, Fe). In the FZPR process Si and metal impurities can be gathered at the edge of the samples. This effect and formation of SiC are important for mechanical properties. Depending on sample, tensile strength of the FZPR state is enhanced up to more than two times when compared with the infiltrated one. New composites, not limited to B4C-Si system, can be obtained by our hybrid (infiltration followed by floating zone partial melting) approach. This method opens new possibilities for processing at reasonably low temperatures of refractor materials. For understanding of the processes and of their control and optimization further research is needed.

Acknowledgement

Ukrainian team acknowledges projects 0117U004301, 0117U003354, 0116U003737, 0116U006569 supported by Ministry of Education and Science of Ukraine. Romanian team acknowledges MEC-UEFISCDI, projects PNIII-P3-127.3BM/2016 (Copbil RomaniaUkraine), NEWCOMPOSITE,POC E-28 REBMAT, and Core Program PN16-480 P2, P3, Romania. Authors thank Dr. M. Enculescu for help with SEM/EDS measurements.

References

[1] V. Domnich, S. Reynaud, R.A. Haber, M Chhowalla, Boron carbide: structure, properties and stability under stress, J. Am. Ceram. Soc. 94 (2011) 3605-3628. [2] A.K. Suri, C. Subramanian, J.K. Sonber, T.S.R. Murthy, Synthesis and consolidation of boron carbide: a review., Int. Mater. Rev.55 (2010) 4-40. [3] F. Thevenot, Boron carbide – a comprehensive review, J. Eur. Ceram. Soc., 6 (1990) 205–225. [4] S.M. Walley, Historical review on high strain rate and shock properties of ceramics relevant to their application in armor, Adv. Appl. Ceram. 09 (2010) 446–466. [5] T.J. Holmquist, G.R. Johnson, Response of boron carbide subjected to high-velocity impact, Int. J. Impact. Eng. 35 (2008) 742–752. [6] O. Vasylkiv, D. Demirskyi, P. Badica,T. Nishimura, A.I.Y. Tok, Y.Sakka, H. Borodianska, Room and high temperature flexural failure of spark plasma sintered boron carbide, Ceram. Int. 42 (2016) 7001–7013. [7] K.F. Cai, C.W. Nan, M. Schmuecker, E. Mueller, Microstructure of hot-presses B4C-TiB2 thermoelectric composites, J. Alloys Compd. 350 (2003) 313-318. [8] H. Lee, R.F. Speyer, W.S. Hackenberger, Sintering of boron carbide heat-treated with hydrogen, J. Am. Ceram. Soc.85 (2002) 2131–2133. [9] M.W. Chen, J.W. McCauley, J.C. LaSalvia, K.J. Hemker, Microstructural characterization of commercial hot-pressed boron carbide ceramics, J. Am. Ceram. Soc. 88(2005) 1935–1942. [10] N. Cho, Z. Bao, R.F. Speyer, Density and hardness optimized pressureless sintered and post-hot isostatic pressed B4C, J. Mater. Res. 20(2005) 2110–2116. [11] P. Badica, H. Borodianska, S. Xie, T. Zhao, D. Demirskyi, P.F. Li, A.I.Y. Tok, Y. Sakka, O. Vasylkiv, Toughness control of boron carbide obtained by spark plasma sintering in nitrogen atmosphere, Ceram. Int. 40 (2014) 3053–3061. [12]S.Hayun, N. Frage,M.P. Dariel, The morphology of ceramic phases in BxC–SiC–Si infiltrated composites, J. Solid State Chem. 179 (2006) 2875–2879. [13]S. Hayun, N. Frage, H. Dilman, V. Tourbabin, M.P. Dariel, Synthesis of dense B4C–SiC–TiB2 composites, in ed. E. Medvedovski, Ceramic Armor and Armor Systems II, American Ceramic Society, Baltimore, MD, USA, 2006, pp. 37–44.

[14]S. Hayun, D. Rittel, N. Frage, M.P. Dariel, Static and dynamic mechanical properties of infiltrated B4C–Si composites, Mat. Sci. Eng. A 487 (2008) 405–409. [15] S. Hayun, A. Weizmann, M. P. Dariel, N. Frage, Microstructural evolution during the infiltration of boron carbide with molten silicon, J. Eur. Ceram. Soc. 30 (2010) 1007–1014. [16] S. Hayun, V. Paris, M.P. Dariel, N. Frage, E. Zaretzky, The morphology of ceramic phases in BxC– SiC–Si infiltrated composites, J. Eur. Ceram. Soc., 29 (2009) 3395–3400. [17] I. Bogomol, T. Nishimura, O. Vasylkiv, Y. Sakka, P. Loboda, Micro- structure and hightemperature strength of B4C–TiB2 composite prepared by a crucible less zone melting method, J. Alloy. Compd. 485 (2009) 677–681. [18] I. Bogomol, P.Badica, Y.Q. Shen, T. Nishimura, P. Loboda, O. Vasylkiv, Room and high temperature tougheningin directionally solidified B4C–TiB2 eutectic composites by Si doping, J. Alloys Compd. 570 (2013) 94–99. [19] I. Bogomol,T. Nishimura, Yu. Nesterenko, O. Vasylkiv, Y. Sakka, P. Loboda. The bending strength temperature dependence of the directionally solidified eutectic LaB6-ZrB2 composite, J. Alloys Compd, 509 (2011) 6123-6129. [20] P. Badica, A. Crisan, G. Aldica, K. Endo, H. Borodianska, K. Togano, S. Awaji, K. Watanabe, Y. Sakka, O. Vasylkiv, ‘Beautiful’ unconventional synthesis and processing technologies of superconductors and some other materials, Sci Technol Adv Mater 12 (2011) 013001. [21] R.W. Olesinski, G.J. Abbaschlan, The Si-C (Cabon-Silicon) system, Bull. Alloy Phase Diag. 5 (1984) 486-489. [22] T.B. Massalski, H. Okamoto, P.R. Subramanian, L. Kacprzak, Binary Alloy Phase Diagrams, 2nd ed., (eds. T. B. Massalski, J.L. Murray, L.H. Bennet, H. Baker), ASM International, 1990. [23] W. W. Zhang, Y. Du, H. Xu, W. Xiong, Y. Kong, W. Sun, F. Pan, A. Tang, Thermodynamic assessment of the Cu-B system supported by key experiment and first-principles calculations, J. Phase Equilib. Diffus. 30 (2009) 480. [24] V. Domnich, Y. Aratyn, W.M. Kriven, Y. Gogotsi. Temperature dependence of silicon hardness: experimental evidence of phase transformations, Rev Adv. Mater. Sci. 17 (2008) 33-41. [25] R. Hull (Ed.) Properties of crystalline silicon. In spec Publishing House, 1999, London, 1019 p.

[26] Ian G. Crouch (Ed.) The Science of Armour Materials, Elsevier Wood head publishing, 2017, 715 p. [27] H. Wang, K.T. Ramesh. Dynamic strength and fragmentation of hot-presses silicon carbide under uniaxial compression, Acta Mater. 52 (2004) 355-367. [28] M. Cengiz, F.C. Sahin. Spark Plasma Sintering of Boron Carbide Ceramics Using Metallic Silicon in Square Cross Section, ActaPhysicaPolonica 127 (2015) 1370-1372. [29] Q. An, W.A. Goddard, Microalloying Boron Carbide with Silicon to Achieve Dramatically Improved Ductility, Journal Phys. Chem. Lett. 5 (2014) 4169-4174. [30] A. A. Shul’zhenko, G. S. Oleinik, D.A. Stratiichuk, N.N. Belyavina, V.Ya. Markiv. Formation of Polycrystalline Boron Carbide B4C with Elevated Fracture Toughness, Powder Metallurgy and Metal Ceramics 44(2005)75–83.

Table 1 Samples, starting composition, phase composition, FZPR pulling rate, tensile strength (IN denotes infiltrated samples and FZPR is for infiltrated and subsequently FZPRed samples; R denotes rapid pulling in the FZPR;for L, T and S notation see text).

Sample

Powder mixture composition

Density(g/cm3 )

Phase composition by XRD, (wt. %)

(vol %)

A-IN A-FZPR

B4C-1

B4C-2

Large

Small

particles

particle s

100

-

FZPR pulling rate (mm/min )

Tensile strength , MPa

βSiC

Si

3

B4 C

-

17

50

10

23

-

88 ±11

2.54

-

77 (L)

12 (L)

11 (L)

5

114 ±14

31. 7 (S)

18. 8 (S)

49. 5 (S)

B12(SiBC)

B-IN

60

40

-

45

21

7

27

-

58 ±7

B-FZPR

60

40

2.66

-

72 (L)

18 (L)

10 (L)

5

138 ±15

2.74

-

68 (L)

25 (L)

7 (L)

10

169 ±17

-

-

61

20

19

5

179 ±18

2.50

-

45. 2 (L)

20. 8 (L)

34 (L)

5

188 ±22

BFZPR(R )

C-IN

99 %(60%B4C -1 + 40%B4C-2) +

2.62

1%C C-FZPR

99 %(60%B4C-1 + 40%B4C-2) + 1%C

44. 7 (T)

25. 3 (T)

11. 1 (S)

19. 4 (S)

30 (T) 69. 5 (S)

Fig. 1 XRD patterns for: raw powders of boron carbide B4C-1 and B4C-2 and for infiltrated samples AC. Phases are: 1 - B4C (PDF 35-798), 2 - C (graphite) (PDF 56-159), 3 – Si (PDF 27-1402), 4 – SiC (PDF 29-1129), 5 - B12(B,C,Si)3[12, 15, 16].

Fig. 2 SEM images in back scattering (BS) mode on the polished cross section of infiltrated samples: (a) – A-IN, (b) – B-IN, (c) – B-IN detail, (d) – C-IN. Black grain are B4C, dark gray – B12(B,C,Si)3, SiC – medium gray, Si – light gray, metal impurities (Ti, Cu, Al, Fe) – white.

Fig. 3 Schematic drawing of a FZPR sample: longitudinal cross section. Areas (L, T, and S) of XRD measurement and regions I-III with a different microstructure are indicated.

Fig. 4 SEM images in backscattering mode (BS) on the axial polished cross section of the FZPR samples (see Table 1) A-FZPR: (a) - general view and (b) – detail of image (a) taken on central region (I); B-FZPR: (c) - general view and (d) – detail of the image (c) taken on the central region (I), (e) – detail showing needle-like grains in the region II containing heavy elements (metals); B-FZPR(R): (f) – general view taken in BS mode at the border between regions I and II; C-FZPR: (g) – general view and (h) – detail of image (g) taken on region I. Double arrow shows the direction of the sample movement. Lighter phases indicate the presence of heavy elements such as Si and metal impurities.

Fig. 5 SEM-BS image taken on the axial polished cross section of the sample B-FZPR, EDS maps for different detected elements and red-green-blue (RGB) image representing the overlapped images for all elements.

Fig. 6 SEM image taken on the axial polished cross section of the sample C-FZPR, EDS maps for different detected elements and red-green-blue (RGB) image representing the overlapped images for all elements.

Fig. 7 Selected XRD patterns taken on FZPR samples. 1 - B4C (PDF 35-798), 3 – Si (PDF 27-1402), 4 – SiC (PDF 29-1129).

Fig. 8 Optical microscopy images on polished surfaces showing indentation imprints on the (a) –white (sample A-FZPR, HV=9 GPa, KIC=1.2 MPa·m1/2) and (b)- gray (sample A-FZPR, HV=16 GPa, KIC=1.9 MPa·m1/2) phases and (c) on a mixed region (sample C-FZPR, HV=13 GPa, KIC=2.3 MPa·m1/2). Regions 1 and 2 show arrest of a crack and intergranular cracking at the interface between two phases.

Fig. 9 SEM image showing fractured surface of the sample C-FZPR. White thick arrows show wavy multiple cracks and thin red arrows show grains or cracks as a result of a intergranular fracturing mechanism (pull out).

Fig. 10 Tensile strength dependence on SiC amount in the IN and FZPR A-C samples. Continuous line is the least square fit of the experimental data.