SPARK PLASMA SINTERING OF NANOCRYSTALLINE BINDERLESS WC HARD METALS

Transcription

1 SPARK PLASMA SINTERING OF NANOCRYSTALLINE BINDERLESS WC HARD METALS Milan DOPITA, Chidambaram Rajagopalan SRIRAM, David CHMELIK, Anton SALOMON, Hans Jürgen SEIFERT Institute of Materials Science, Technical University of Freiberg, Gustav-Zeuner-Strasse 5, Freiberg, Germany, Abstract Nanocrystalline tungsten carbide powder, having a crystallite size around 37 nm and BET specific surface area 4.2 m 2 /g was sintered using the spark plasma sintering (SPS) technique. Sintered samples were investigated using different analytical methods providing detailed information on the microstructure and mechanical properties of materials. Density and porosity of specimens were determined using the Archimedes principle and optical and scanning electron micrographs. The X-ray diffraction (XRD) and high resolution transmission electron microscopy (TEM/HRTEM) investigations provided the information on the crystal real structure, crystallite size and lattice defects. The electron back-scatter diffraction (EBSD) measurements yielded the details about the grain size, frequency and distributions of grain boundaries. Finally the essential mechanical properties of sintered samples were obtained from the hardness and fracture toughness measurements. The influences of individual sintering conditions; sintering temperature and sintering time especially, on the microstructure and mechanical properties of sintered specimens were derived. Fully compact samples (relative density ~ 100%) having the Vickers hardness HV10 around 30 GPa and fracture toughness K Ic approximately 7.3 MPa m 1/2 were sintered from temperatures of and holding times 180 seconds. Specimens sintered at lower temperatures and shorter times showed lower density, which resulted in a significant drop in the sample hardness. Keywords: Tungsten carbide, nanocrystalline, binderless, SPS, hardness 1. INTRODUCTION The WC-Co hard metals are, because of their suitable mechanical properties as high hardness and high fracture toughness simultaneously, the most commonly used materials for the tooling applications, nowadays. Modern trends in the materials engineering leading to the improvement of the materials properties, decrease of the production costs and environmental protection tend to the effort to extend the serving life-time together with the significant improvement in the materials properties. In the hard metal industry this leads to the trend of preparing the materials with sub-micron or nanocrystalline microstructure as it has been shown that nanocrystalline hard metals have significantly improved mechanical properties in comparison to their coarse-grained counterparts [1, 2]. The conventional procedures as a classical liquid phase sintering however reached their limits in the preparation of the nanocrystalline materials, especially because of relatively long sintering times in combination with high temperatures which are necessary for the sintering of compact fully dense materials. The problems of long sintering times are significantly reduced using a modern sintering method spark plasma sintering (SPS). The presence of ductile metallic binder phase, most frequently cobalt in the WC-Co hard metals, has several important consequences on the mechanical properties of material as well as on the densification process. First, the presence of the ductile binder leads to the pronounced drop in the hardness of material and moreover it has in consequence considerable decay of corrosion and oxidation properties, particularly at

2 higher temperatures [3]. On the other hand the binder phase significantly helps during the densification while sintering close to the melting point of Co, which is extensively lower than melting point of the WC (2785 C). Thence it follows that it is nearly impossible to sinter binderless WC using classical sintering procedures. For successful specimen densification a simultaneous temperature and pressure application together with the WC grains surface activation is necessary. Both of those two conditions are successfully satisfied using the spark plasma sintering. 2. EXPERIMENTAL 2.1 Powder properties The commercially available tungsten carbide powder (H.C. Starck GmbH, Goslar, Germany) having BET specific surface area of 4.2 m 2 /g (corresponding to the mean grain size of d BET = 91 nm) and mean crystallite size determined from the X-ray diffraction measurements of approximately 37 nm was used in this work. The powder total and free carbon contents due to the vendor specifications were: C total = 6.08 ± 0.03 %, C free = 0.08 %, respectively. The oxygen contamination measured using the LECO TC-436 analyzer was 0.33 mass %. 2.2 Sintering and sintering conditions The samples sintering were done by the spark plasma sintering (SPS) technique. The SPS equipment FCT HP D 25/1 [4] (FCT Systeme GmbH, Raunstein, Germany) was used for sintering. WC powder was loaded into the graphite die having the inner diameter of 20 mm fabricated from high strength graphite. A graphite foil was used to avoid the contact of the powder with the die surface. To minimize the radiation heating losses and to avoid the creation of possible radial temperature gradients the whole graphite sample environment was surrounded with a carbon felt. The specimens were prepressed to 25 MPa and heated up by a pulsed electric current with the heating rate of 150 C/min up to 1600 C at which temperature the pressure of 80 MPa was applied. Consequently the temperature was increased with heating rate of 150 C/min up to the sintering temperature. After sintering the specimens were cooled down with the cooling rate of 150 C/min to the room temperature with continuous releasing of pressure. The samples series with sintering temperatures, and and holding times of 1, 3 and 10 minutes at sintering temperature was prepared. The whole sintering procedure was done in vacuum. 2.3 Specimens investigation Sintered cylindrical samples having the dimensions of 20 5 mm (diameter and height, respectively) were cut on two semi-cylinders. Section plane was carefully polished to remove strained surface of material created during cutting and consequently subjected to the detailed investigations. The polishing procedure followed the next steps; first, the specimens were mechanically grinded using silicon carbide grinding plates with decreasing roughness and subsequently polished with 15, 6, 3 and 1 µm diamond paste. Finally the samples were etched for 90 seconds using the OP-S suspension (Struers A/S, Denmark). The density and porosity of sintered specimens were determined using the Archimedes principle and the image analyses of optical and scanning electron microscope micrographs. The XRD measurements were performed using the URD6 diffractometer (Seifert, FPM, Freiberg, Germany). The goniometer working in conventional focusing Bragg-Brentano geometry was equipped with the graphite monochromator in the diffracted beam and filtered CuKα radiation (λ = nm) was used for the

3 measurements. Measured diffraction patterns collected in the angular range 2θ = , with a step of 2θ = 0.04 and time 10 seconds/step were refined using the full powder pattern refinement (rietveld method) to obtain the contents of present phases as well as the detailed information on the real structure of phases. For the rietveld refinement computer program Maud [5] was used. The morphology and microstructure were investigated using high-resolution scanning electron microscope (SEM) LEO 1530 (Carl Zeiss A.G., Jena, Germany) equipped with a field emission cathode, with a detector of back-scatter electrons (SEM/BSE) and Nordlys II EBSD detector (HKL Oxford Instruments, Bucks, UK). Hardness of sintered specimens was measured using the M4U-025 Vickers hardness tester (EMCO Test GmbH, Kuchl, Austria). An indenter load of 10 kg and indentation time 10 s were applied. Fracture toughness was determined from the measurements of the Vickers indentation crack lengths applying the theory proposed by Shetty et al. [6]. 3. RESULTS 3.1 Sintering characteristics, densification and density In Fig. 1 the shrinkage rates, corrected for the temperature expansion of graphite sample environment and piston, of samples sintered for 10 minutes at three different temperatures are shown. It is obvious from the plot that the powder grains surface activation and consequently the sintering process starts at approximately 1200 C (black solid arrow), whereas the sharp peaks at 1600 C correspond to the maximum pressure application (black open arrow). The steep drop down of the shrinkage rates at the end of sintering for samples sintered at and clearly demonstrates that the powder compaction was not totally finished in these two cases, while the smooth decay of shrinkage rate, nearly to zero, for sample sintered at indicates the end of the compaction procedure. Shrinkage rate [mm/min] C 1800 C 1900 C Temperature [ C] 100% powder compaction, no shrinkage rate change Fig. 1. Shrinkage rates of samples sintered at, and as a function of temperature. The beginning of the powder grains surface activation at approximately 1200 C is marked with black solid arrow, the maximum pressure application at 1600 C is marked with an open arrow. The red arrow points to the shrinkage rate smoothly decreasing to zero in the specimen sintered at, which means the end of sintering procedure and nearly 100% sample density. In Fig. 2a densities determined using the Archimedes principle and the relative densities as a function of sintering time for specimens sintered at, and are shown. In samples sintered at a pronounced increase of the density between the sample sintered for 1 and 3 minutes can be seen in plot, however the relative density of sample sintered for 10 minutes is still quite low around 98.5 %. In specimens sintered at the density linearly increases with sintering time, while in samples sintered at the density saturates already in the sample sintered for three minutes. The maximal relative density of samples sintered at and approaches nearly 100%.

4 In Fig. 2b a scanning electron micrograph acquired in the back scatter electron mode (SEM/BSE) of typical polished specimen morphology is shown. Individual grains (different gray scale corresponds to the different crystallographic orientation of grains) are clearly visible in the micrograph. The porosity usually determined from the optical or scanning electron microscopy is negligible as no clear pores can be found in the micrograph. Therefore, the common porosity determination based upon comparison of optical or SEM micrographs with standard pictures can not be applied to our samples and the porosity can only be roughly estimated from the deviation of the relative density from 100% (a) 100 (b) Density [g/cm 3 ] Fig. 2. Plot of the densities determined using the Archimedes principle and relative densities of specimens sintered at, and as a function of the sintering time (a). SEM/BSE micrograph of the specimen sintered at for 10 minutes (b). 3.2 Phase composition and crystallite size evolution The initial powder contained 97 vol.% of WC and 3 vol.% of W 2 C. During the sintering process the volume fraction of minor W 2 C phase increased to approximately 5 vol.%. As it can be seen in Fig. 3a this process is independent on the sintering temperature and sintering time. We suppose that the phase transform of some WC into the carbon poorer W 2 C phase is a consequence of the powder oxygen contamination. During the sintering process the oxygen molecules adsorbed on the WC grains surface react with WC transforming some carbon atoms into carbon monoxide and creating carbon poorer W 2 C phase Relative density [%] Mag = 100 kx 500 nm Volume fractions [%] (a) ; WC, W 2 C ; WC, W 2 C ; WC, W 2 C 0 Crystallite size [nm] (b) WC crystallite size Fig. 3. Plot of the phase composition (vol.%) as a function of sintering time, open symbols correspond to WC and solid symbols to W 2 C (a). Evolution of the WC crystallites size as a function of sintering temperature for specimens sintered at, and, respectively. Dotted red line corresponds to the crystallite size of WC in initial powder (b).

5 The size of coherently diffracting domains (crystallite size) of the WC phase determined using the XRD is nearly independent on the sintering time, while it strongly depends on the sintering temperature as can be clearly seen in Fig. 3b. The crystallite size increases from about 37 nm (in the initial powder) to approximately 110 nm, 130 nm and 150 nm in the samples sintered at, and, respectivelly. 3.3 Hardness and fracture toughness The measured Vickers hardness values HV10 slightly increases with increasing sintering time in the samples sintered at. Analogous increase in hardness can be observed in the specimens sintered at, however between one and three minutes of sintering, only. Hardness values of samples sintered for three and ten minutes are within the bounds of erros similar. Finaly the hardness of samples sintered at is independent on the sintering time and reaches the value of 30 GPa. For details on the hardness temperature and sintering time evolution see Fig. 4a. In Fig. 4b plots of the fracture toughness K Ic versus sintering time for samples sintered at, and are shown. The fracture toughness smoothly increases with increasing sintering time in all specimens under study. The increase in the fracture toughness is most pronounced between one and three minutes of sintering time, while it approaches its maxima around 7.3 MPa m 1/2. HV10 [GPa] (a) 0 Fracture toughness [MPam 1/2 ] (b) 6.0 Fig. 4. Plot of the measured Vickers hardnes HV10 as a function of sintering temperature for specimens under study (a). Plot of the fracture toughness K Ic determined from the lengths of the indentations cracks using the Shettys model (b). 4. DISCUSSION The samples relative density strongly depends on the sintering temperature and time. Specimens sintered at posses a lowest relative density, which corresponds to a highest porosity of samples. Even in specimens sintered for 10 minutes the relative density is 98.5%, only. With increasing sintering temperature the dependence of the samples relative density on sintering time decreases and in samples sintered at the density saturates already after three minutes of sintering. Mechanical properties of sintered specimens correlate strongly with relative densities (compare Figs. 2a and 4) - the higher the density the better are the values of hardness and fracture toughness.

6 The phase composition appears to be independent on the sintering conditions and does neither correlate with other microstructural properties nor mechanical parameters. With increasing sintering temperature the crystallite size of dominant WC phase increases significantly while the dependence of the crystallite size on the sintering time is inappreciable. In all samples under study the crystallite size of sintered samples is below 200 nm, which is a value accepted in the hard metal community as a boundary between nano and ultra-fine grained materials. Due to the clear difficulties concerning the binderless WC sintering, only few works were done so far on the sintering of nanocrystalline or fine-grained binderless WC [7-10]. However, it is rather hard task to compare published results with ours, since other authors used different starting powders (powders having different crystallite size) and various sintering equipments which resulted in diverse definition of sintering conditions (for example sintering temperature). Moreover, the mechanical properties of sintered specimens were measured using different indentation loads which make the direct comparison of results unreliable. Kim and coworkers [7] referred to the binderless WC having the relative density 97.6%, grain size 400 nm, hardness of 24.8 GPa and fracture toughness 6.6 MPa m 1/2, while Huang et al. [9] sintered samples with relative density 99.9%, grain size 280 nm, hardness of 27.4 GPa and fracture toughness 4.4 MPa m 1/2. In comparison to those papers our results published in this work indicate slightly higher hardness (30 GPa) and fracture toughness (7.3 MPa m 1/2 ) as well. This is probably caused by fact that our sintered specimens have smaller grain size (below 200 nm) than samples referred in [7] and [9]. 5. CONCLUSIONS Binderless WC hard metals were sintered from nanocrystalline powder. The influences of sintering temperature and sintering time, on the microstructure and mechanical properties of sintered specimens were derived. Correlation between individual determined microstructural and mechanical properties were discussed. The fully compact (relative density ~ 100%) samples having the Vickers hardness HV10 around 30 GPa and fracture toughness K Ic approximately 7.3 MPa m 1/2 were sintered from temperatures of and holding times 180 seconds. Specimens sintered at lower temperatures or shorter sintering times showed lower density, which resulted in a significant drop in the sample hardness. The hardness and fracture toughness values of fully compact sintered samples were higher than results published in literature so far. ACKNOWLEDGEMENT The authors would like to kindly acknowledge the financial support through the Dr. Erich-Krüger-Stiftung, this is subproject #8 from Freiberg High-Pressure Research Centre FHP, website: LITERATURE [1] Gille, G. et al., Submicron and ultrafine grained hardmetals for microdrills and metal cutting inserts, Int. J. of Refr. Metals & Hard Mater., 20, 1, (2002), [2] Fang, Z.Z. et al., Synthesis, sintering, and mechanical properties of nanocrystalline cemented tungsten carbide - A review Int. J. of Refr. Metals & Hard Mater., 27, 2, (2009), [3] Suzuki H., Cemented carbide and sintered hard materials, Maruzen Publishing Company, Tokyo, Japan (1986), [4] FCT Systeme GmbH, webpages: [5] Lutteroti L., MAUD, IUCR Commission for Powder Diffraction, CPD Newsletter (IUCr), 21, (1999). [6] Shetty D, et al., J Mater Sci 20, (1985), [7] Kim, H-C. et al., Consolidation and properties of binderless sub-micron tungsten carbide by field-activated sintering, Int. J. of Refr. Metals & Hard Mater. 22, (2004), [8] Girardini L., et al., Metal. Powder Report, 63, 4, (2008), [9] Huang, S.G., et al., Binderless WC and WC-VC materials obtained by pulsed electric current sintering, Int. J. of Refr. Metals & Hard Mater., 26, (2008), [10] Zhang, J., et al., Binder-free WC bulk synthesized by spark plasma sintering, J. of Alloys and Comp. 479, (2009),