Investigation of microstructure and mechanical properties at low and high temperatures of WC–6 wt% Co

Investigation of microstructure and mechanical properties at low and high temperatures of WC–6 wt% Co

Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181 Contents lists available at ScienceDirect Int. Journal of Refractory Metals a...

4MB Sizes 0 Downloads 26 Views

Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Contents lists available at ScienceDirect

Int. Journal of Refractory Metals and Hard Materials journal homepage: www.elsevier.com/locate/IJRMHM

Investigation of microstructure and mechanical properties at low and high temperatures of WC–6 wt% Co Satyanarayana V. Emani a, Ana Flavia C. Ramos dos Santos a, Leon L. Shaw a,⁎, Zheng Chen b a b

Department of Mechanical, Materials and Aerospace Engineering, Illinois Institute of Technology, Chicago, IL, USA Henan GrandMetals Corporation, Mengzhou, Henan, China

a r t i c l e

i n f o

Article history: Received 31 December 2015 Received in revised form 22 April 2016 Accepted 25 April 2016 Available online 03 May 2016 Keywords: WC-Co Grain size Hardness Toughness High temperature properties

a b s t r a c t The correlations of the microstructure, room temperature hardness and toughness, and high temperature compressive stress and strain of a series of WC–6 wt% Co materials with grain sizes at sub-micrometer ranges (b 0.30 μm) have been investigated. It is found that the average size and size distribution of WC grains play a critical role in both room temperature and high temperature (600 and 800 °C) properties. Furthermore, the maximum compressive stresses at 600 and 800 °C increase with increasing the room temperature hardness and decreasing grain sizes. However, the fracture strain in compression at 600 and 800 °C cannot be predicted from the room temperature toughness. This absence of a relationship is attributed to the facts that the fracture strain of compressive tests at high temperatures is dictated by significant plastic flow of the Co phase, whereas the toughness at room temperature is controlled by crack propagation with little plasticity. In general, small grain sizes (at ~0.20 μm) with or without a bi-modal grain size distribution are beneficial to the maximum compressive stress and fracture strain at 600 and 800 °C. In contrast, grain sizes at ~0.27 μm with a bi-modal grain size distribution are good for room temperature toughness, but offer inferior high temperature compressive stresses and fracture stains. Thus, the design of the microstructure at a fixed Co composition depends on the intended applications. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction HWC ¼ 1382 þ 23:1 dWC Since the first successful application of WC-Co as drawing dies in the 1920s [1], WC-Co has become the workhorse for hardmetal component applications today with 98% of hardmetal components made of cemented WC-Co [2]. Because of its important roles in manufacturing industry, the room temperature hardness and toughness of WC-Co have been investigated extensively [3–18]. One of the widely used hardness-microstructure relationships is proposed by Lee and Gurland [7], as shown in Eq. (1) which is established using 26 different WC-Co materials with WC particle sizes at about 1 μm or larger. HWC‐Co ¼ HWC VWC C þ HCo ð1–VWC CÞ

ð1Þ

where HWC-Co is the hardness of WC-Co, HWC and HCo are the hardnesses of WC and Co phases, respectively, VWC is the volume fraction of WC, and C is the contiguity of WC [7]. Further, HWC and HCo are in turn a function of the size of their respective phase (i.e., the WC particle size, dWC, and the mean free path of the Co phase, dCo), as shown in the following empirical equations [7]

⁎ Corresponding author. E-mail address: [email protected] (L.L. Shaw).

http://dx.doi.org/10.1016/j.ijrmhm.2016.04.009 0263-4368/© 2016 Elsevier Ltd. All rights reserved.

HCo ¼ 304 þ 12:7 dCo

−1=2

−1=2





kg=mm2

kg=mm2





ð2Þ ð3Þ

In service as cutting tools, cemented WC-Co could experience high temperatures ranging from 600 to 1000 °C depending on cutting speed [19]. As such, high temperature properties, particularly hardness, have also been studied widely [19–27]. However, it is well known that high temperature properties are more difficult to obtain than room temperature properties because of the requirement of more complicated experiment instrumentation and multiple competing deformation mechanisms at high temperature [21,22]. In spite of these challenges, many studies have revealed that high temperature hardness is proportional to room temperature hardness and both of them follow the Hall-Petch relationship shown in Eqs. (1), (2) and (3) [23,26,27]. Furthermore, deformation of WC-Co changes from a brittle manner to a ductile flow at around 700 °C [20,22–27]. By combining mechanical testing, mechanical spectroscopy, scanning electron microscopy (SEM) and transmission electron microscopy (TEM), Mari, et al. [21] have concluded that reduction in the bending strength at around 600 °C is mainly due to the softening of the Co matrix, while further reduction at around 900 °C is predominately derived from WC softening and additional

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

173

reduction above 950 °C is related to grain boundary relaxation. These conclusions are confirmed by other researchers later such as the work by Buss and Mari [20]. In this study the state-of-the-art WC-Co materials with grain sizes at sub-micrometer ranges (b0.30 μm) have been investigated for their relationship between room temperature properties (i.e., hardness and toughness) and high temperature compressive properties (compressive strength and fracture strain) and the dependence of these properties on the grain size and size distribution. Such a study is of scientific interest because nearly all of the previous high temperature studies and their conclusions are based on WC-Co materials with grain sizes near 1 μm or larger [19–23,26,27]. Furthermore, such a study is technologically important as all the state-of-the-art WC-Co materials in this investigation are acquired from companies in multiple countries including Australia, China, Germany, and Japan. As such, the insight derived from this study can be utilized by both the scientific community and the industry. 2. Experimental procedure Fig. 1. Image of an indent at 750× magnification indicating the diagonals.

2.1. Measurement of room temperature properties As summarized in Table 1, six (6) WC-Co samples from different commercial sources were investigated in this study. The as-purchased WC-Co rods measuring 3 mm in diameter were cut and polished using diamond slurry to a final polish of 0.25 μm. Hardness tests were performed on the polished samples using a diamond indenter at 5 kgf load for 15 s. Low load indentation (0.5 kgf) was also performed to see if there is any difference in the hardness. There was 1% to 3% difference in hardness values since cracking occurred in high loading cases and relaxed the indent, resulting in slightly lower hardness values. Although there was difference in hardness, the trend in hardness was the same for both loading cases. Therefore, for the purposes of this paper only the hardness at 5 kgf load was reported. Field-emission SEM (FESEM) images of each indent were taken at both 750× and 1 500× magnifications to determine indent sizes and crack lengths. Representative images of the indent, a crack emanating from the indent and some symbols used are presented in Figs. 1 and 2, respectively. The lengths of the diagonal and cracks were measured on the images using Image J (image analysis software). The Vickers hardness, Hv (GPa), and fracture toughness, Kc (MPa·m1/2), were calculated using Eqs. (4) and (5) [28], respectively.

E ¼ ðVCo  ECo Þ þ ðVWC  EWC Þ VCo ¼

  136 , 2 sin F 2 Hv ¼

ð4Þ d2

 K c ¼ 0:0193 ðHv aÞ

Note that Eq. (5) was first proposed by Niihara, et al. [28,29] based on the toughness data of WC-Co, many of which were measured using the standardized fracture toughness tests, such as chevron-notched beam and single-edge pre-cracked beam methods. Calibration using these standardized fracture toughness tests is important because a recent study [30] has unambiguously indicated that there is no single indentation toughness equation that can produce accurate fracture toughness values for all brittle materials. Instead, the indentation toughness equation can only be used for the material that has been used for the calibration. Eq. (5) was later confirmed by Shetty, et al. [31] using WC-Co again, and has been widely used by the WC-Co community after Niihara and Shetty [28,29,31–34]. In this study, the indentation toughness was evaluated with ten valid indentations on each specimen. The elastic modulus of each WC-Co rod was estimated using the rule of mixtures, Eq. (6), whereas the volume fraction of Co was calculated from Eq. (7) [18].

E Hv

2=5 

1



WCo =ρCo WCo =ρCo þ WWC =ρWC

ð6Þ ð7Þ

where Vx and Wx stand for the volume and weight fraction of the x phase respectively, ρx is the theoretical density, Ex is the theoretical elastic modulus, and VWC = 1 − VCo. The theoretical density and elastic

ð5Þ

L1=2

where F (with a unit of N) is the indentation force, d (m) is the average diagonal length (d1 + d2)/2, a (m) is the length of half indentation diagonal (i.e., d/2), E (GPa) is the elastic modulus, and L (m) is the average crack length emanating from the four corners of the indent, i.e., (l1 + l2 + l3 + l4) / 4. Table 1 The average of EDS area analyses of WC-Co samples (wt%) along with the calculated volume fractions and estimated elastic moduli. Sample ID

C (%)

V (%)

Cr (%)

Co (%)

W (%)

Vol fraction Co

Vol fraction WC

E (GPa)

S1 S2 S3 S4 S5 S6

7.44 7.77 7.66 7.70 7.77 7.43

0.40 0.39 0.20 0.34 0.25 0.23

0.66 0.57 0.33 0.39 0.41 0.41

5.79 5.61 5.91 5.69 6.23 6.50

85.72 85.68 85.90 85.87 85.36 85.45

0.097 0.094 0.099 0.096 0.104 0.109

0.903 0.906 0.901 0.904 0.896 0.891

659.5 660.9 658.6 660.3 655.6 653.8

Fig. 2. Image of a crack at 1 500× magnification indicating one of the crack lengths.

174

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

modulus of WC were taken to be 15.63 g/cm3 and 707 GPa, whereas the corresponding values of Co were 8.9 g/cm3 and 209 GPa, respectively [18]. Ten indents were taken per sample for measuring the hardness. 2.2. High temperature compression tests WC-Co rods of 3 mm in diameter and 5 mm in length were compressed at two different temperatures, 600 and 800 °C, using a “Gleeble-3800” testing system. The samples were heated to 600 or 800 °C and compressed at a constant displacement rate of 0.03 mm/min (i.e., an engineering strain rate of 6 × 10−3 min−1). The temperature of the specimens was kept constant within 2 °C of the target temperature. Fig. 3(a) shows the sample setup in the “Gleeble-3800” testing system with a specimen heated to 800 °C. We can observe that the specimen temperature is visibly uniform for the entire specimen except two ends. The temperature measurement and control were done by a K-type thermocouple that was directly in contact with the specimen surface at the central location at all times. Since the specimens were small and were not able to accommodate the strain gauge, the engineering strain was calculated based on the cross-head displacement and the initial specimen length. Fig. 3(b) shows the image of a sample with 10% compressive strain recorded by the “Gleeble-3800” testing system, while the actual measurement after testing indicated 9.6% strain. Therefore, the engineering strain recorded by the “Gleeble-3800” testing system was a very good approximation of the real engineering strain of the sample because all of the samples tested had ~10% strain or larger in this study and thus had small errors. The target compressive strain for all samples was 20%, but most of the samples failed before that. The engineering compressive stress was estimated using the initial cross-section area of the specimen. This was certainly different from the real engineering compressive stress because of the well-known barrel shape of compressive specimens. However, as can be seen from Fig. 3(b), the barrel shape of the specimen after 10% compressive strain is not severe and thus the estimation of the engineering stress using the initial cross-section area was reasonably good. Furthermore, all samples were estimated in the same way, making comparisons among them with no bias. The samples were all compressed under vacuum to avoid any oxidation. At least 3 specimens of each sample were tested for each temperature using the “Gleeble-3800” testing system. 2.3. Materials characterizations All of the WC-Co samples acquired in this study had a nominal composition of 6 wt% Co. This was confirmed by the composition analysis obtained from the energy dispersive spectroscopy (EDS) (Table 1).

Fig. 3. (a) Sample setup and the specimen heated to 800 °C. The thermocouple was directly mounted on the specimen. (b) A sample after 10% compression. Note that the sample is slightly barreled. Each division in (b) is 1 mm.

The grain sizes of all samples were analyzed using SEM images and Image J software. All samples were polished and etched strongly. For etching, the samples were immersed in a H2O2:HNO3 (9:1) solution at 65 °C for 10 min [35]. The etchant dissolved the cobalt phase aggressively and also mildly etched the WC phase. This helped in differentially etching the grain boundaries of WC grains, providing contrast between grains that would otherwise be counted as one. Grain size was analyzed using 3 images at high magnification (30,000×) and sampling all the grains in the frame. Image J was used to identify each grain and the area of each individual grain was measured. Fig. 4 represents the grain size measurement method. X-ray diffraction (XRD) was also performed to analyze differences in the phase composition and crystal structure of samples. Densities of all the samples were measured using the Archimedes principle using Eq. (8) [36]. ρ}WC−Co ¼

Weight in air Weight in air−Weight in water

ð8Þ

To obtain the theoretical density, ρ'WC−Co, of each WC-Co sample, we used Eq. (9) with the aid of the Co weight percentage derived from the EDS analysis (Table 1). The relative density of each sample is the ratio of its Archimedes density to the theoretical density. 0

ρWC−Co ¼ ρWC V WC þ ρCo V Co ¼

ρWC ρCo ρWC W Co þ ρCo W WC

ð9Þ

3. Results and discussion 3.1. Microstructure and composition The analysis of the phase composition obtained from XRD (Fig. 5) shows that there is no significant difference in either the crystal structure or the phase composition of the samples. All the samples contain only two observable primary phases, WC and Co. Non-stoichiometric phases such as W2C or η phases that are detrimental to the properties are not detected in the X-ray diffraction patterns. The phases related to the trace elements (b1 wt%) such as Cr and V observed in the EDS analysis (to be discussed later) are not found in the X-ray diffraction patterns presented in Fig. 5 because the amount of these phases, if present (not in the solid solution), fall below the detection limit of the XRD equipment.

Fig. 4. An example of grain size measurement using image J software. Note: individual grains are identified.

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Fig. 5. X-ray diffraction patterns of Samples S1 through S6 along with the standard WC and Co patterns. The observable Co peak is marked, while all other peaks correspond to WC.

The FESEM images of all the samples after etching are presented in Fig. 6 and the corresponding EDS analysis is summarized in Table 1. The EDS analysis shows that the compositions of all the samples are similar to each other with the Co weight percentage at ~6%, consistent with the nominal composition provided by vendors. Interestingly, all the samples contain a small amount of Cr and V. The EDS spot analysis reveals that most of Cr and V are present in the Co binder phase or at the WC-Co interface. The resolution of the EDS equipped in SEM (about 1 μm) does not allow further identification of the precise location at which Cr and V are present. It is known that the Co phase has very high solubility for vanadium carbide (~6 wt%) and chromium carbide (~12 wt%) [1]. Prior studies [37–39] have revealed that V and Cr additives can prevent grain growth. Furthermore, it has been suggested that vanadium carbide dissolves in the Co phase and segregates at the interface between the WC and Co phases, thereby preventing the grain growth [40,41]. In contrast, Cr could be found as carbides within the Co matrix forming Cr-rich Co7C3 type precipitates [39,42] or could go into solid solution and form other intermetallic precipitates [43]. However, other research suggests that Cr is present as precipitates along the WC-Co interface [44]. The significance of these additives cannot be overstated as small weight percentages alter the microstructural and mechanical properties significantly by altering the grain size and morphology [45]. Typically, these V and Cr carbides are used in conjunction to avoid non-uniform growth of WC and formation of binder agglomeration [43]. The alloys studied in this work have comparable compositions to the ones in the open literature, typically b0.6 wt%, because higher percentages of these phases are detrimental to oxidation resistance and fracture toughness [39,42]. From the microstructure shown in Fig. 6 it can be inferred that the samples have a significant variation in the grain size, morphology and distribution. It is interesting to note that six sets of the samples can be divided into three groups based on the grain size, morphology and distribution. Samples S1 and S2 belong to the first group which contains some large elongated WC grains along with many small grains. The second group includes samples S3 and S4 which have larger grain sizes. Furthermore, most of the grains in samples S3 and S4 are equiaxed. Samples S5 and S6 are in the third group which contains equiaxed grains with bi-modal size distribution. The average grain diameters of all of these samples are determined to be submicrometers ranging from 0.15 to 0.3 μm, as summarized in Fig. 7. The first group has the smallest average WC grain size (~ 0.17 μm), while the second group has the largest average WC grain size (~0.27 μm) and the third group is in the middle (~0.21 μm). The grain size distribution of these samples is presented in Fig. 8(a) where the number of grains is plotted as a function of grain size. The same data of the grain size distribution in

175

Fig. 8(a) is also presented in an “area distribution” fashion in Fig. 8(b) where the cross-section area of grains is plotted as a function of grain size. Note that the cross-section area of grains in each size range is obtained by summarizing the areas from all the grains within that specific grain size range. Note that the number distribution of grain size (Fig. 8a) is in good agreement with the trend of the average grain size shown in Fig. 7 since the average grain size in Fig. 7 is the number-averaged grain size. However, it is noted that the area distribution of grain size (Fig. 8b) does provide some new information which is not available from Fig. 7. That is, samples S3 and S6 can be considered to have a bimodal grain size distribution because they have a high area fraction (N10%) of grains with sizes larger than 0.65 μm, whereas the other samples do not (Figs. 8b and 9). As will be shown below, mechanical properties of WC-Co are affected by the average grain size as well as the grain size distribution. As shown in Fig. 6, samples S1 and S2 contain some high aspect ratio grains. To examine the effect of the aspect ratio of grains, the data in Fig. 8 is also plotted in Fig. 10 to include the morphology effect on the area distribution of grains. To produce Fig. 10 (i.e., the area distribution with the morphology effect), we have multiplied the area of each grain with its aspect ratio and summarized these multiplied areas within that specific grain size range. The aspect ratio of each grain is determined using Image J software which approximates the area of each grain using an ellipse. The ratio of the long axis to the short one of the ellipse defines the aspect ratio. It is noted that Fig. 10 is very similar to Fig. 8(b). Thus, the effect of the aspect ratio, if any, in the present study is expected to be small. All samples at the as-polished condition contain no observable porosity under SEM and optical microscopy, indicating that all samples are fully dense. This is consistent with the Archimedes density measurements which reveal that all samples have densities of 99% or higher (Table 2). Based on these data, it is determined that all samples are fully dense, and thus different mechanical properties to be discussed later are not due to sample densities since all samples are dense WCCo materials. 3.2. Room temperature properties The mean hardness and indentation toughness at room temperature along with the standard deviation for all samples are summarized in Fig. 11. In addition, the individual crack lengths emanating from the four corners of each indent for all samples are listed in Table 3. By comparing of Fig. 11 with Fig. 7, one can reach three general trends. First, small grain size leads to high hardness (e.g., the first group has the smallest grain size and the highest hardness, whereas the second group has the largest grain size and the lowest hardness). Second, large grain size results in high toughness, as demonstrated by the second group samples (S3 and S4). Third, high hardness materials have low toughness. For example, samples S1 and S2 have high hardness with low toughness, while samples S3 and S4 have low hardness with high toughness. These general trends are in excellent agreement with the results in the open literature. For example, many studies [3,4, 7–12,18,43,45,46] have demonstrated that hardness increases with decreasing grain sizes. Many studies [3,5,12,13,15,16,18,40,47] have also revealed that there is an inverse relationship between hardness and toughness, i.e., increasing hardness typically results in decreased toughness. In spite of the presence of the three general rules mentioned above, it is noted that there are exceptions to these general rules. For example, samples S5 and S6 have a similar average grain size, but S5 is much harder than S6 while S6 is tougher than S5. Samples S3 and S4 have similar average grain sizes and thus similar hardnesses, but S3 is much tougher than S4. These exceptions indicate that although the average grain size plays a critical role in determining hardness and toughness, but is not the only factor. These exceptions are not a surprise since a

176

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Fig. 6. FEG-SEM images of etched samples at 30,000× magnification of Samples S1 through S6 as indicated.

thorough study on 65 different WC-Co materials conducted by Schubert, et al. [47] has shown that for a given hardness value there could be a range of toughness values. They find that the specific combination of hardness and toughness is affected by the Co content, the concentration of grain growth inhibitors (such Cr3C2 and VC), the gross carbon content, and sintering conditions [47]. Given that the Co, C, V and Cr concentrations are very similar among the six WC-Co materials investigated in this study (Table 1), we have attributed the observed exceptions to the effect of grain size distribution, as explained below. By examining Figs. 8 and 9 carefully, it can be seen that S3 contains the highest area fraction of grains with sizes N0.65 μm among all the samples, and thus has the highest toughness. This is consistent with the general rule that the larger the grain size, the higher the toughness [12]. However, the high toughness in this case is achieved by bi-modal grain size distribution and thus obtained with little compromise in hardness when compared with sample S4. Similarly, S5 and S6 have similar average grain sizes, but S6 contains a higher area fraction of grains with sizes N0.65 μm than S5. As a result, S6 has a higher toughness

than S5 while the opposite is true for hardness. These exceptions indicate that bi-modal grain size distribution can alter hardness and toughness even at a constant average grain size. Furthermore, if properly designed, bi-modal grain size distribution can lead to increase in toughness with little reduction in hardness. 3.3. High temperature properties The typical high temperature compressive stress-strain curves of all samples at 600 °C and 800 °C are presented in Fig. 12(a) and (b), respectively. It is clear that all samples fracture in a brittle manner at 600 °C, whereas the stress-strain curves become ductile flow at 800 °C, indicating that the brittle to ductile transition occurs between 600 °C and 800 °C for all samples. This phenomenon is in good agreement with other researchers working on WC-Co with grain sizes near 1 μm or larger [20,22–27]. The previous studies [20,22] using mechanical spectroscopy have revealed that the brittle to ductile transition is related to softening of the Co phase.

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Fig. 7. Comparison in the average grain diameter for samples S1 through S6.

Comparisons between the room temperature hardness and the maximum compressive stresses at 600 and 800 °C are presented in Fig. 13(a) and (b), respectively. A general trend can be observed from these figures, i.e., WC-Co with high room temperature hardness also

Fig. 9. Grain size distribution indicating the effect of bi-modal distribution. Graphs represent the percentage of the area of the grains in the range indicated.

have high maximum compressive stresses at both 600 and 800 °C although one exception (sample S6 at 600 °C) is present. By comparing Fig. 13 with Fig. 7, one can also conclude that the smaller the average grain size, the higher the maximum compressive stresses at both 600 and 800 °C. In other words, the maximum compressive stresses at both 600 and 800 °C are strongly dependent of the average grain size. These phenomena are in good accordance with the previous studies using grain sizes near 1 μm or larger [23,26,27], showing a strong dependence of the high temperature hardness and strength on the room temperature hardness. To find out whether there is any quantitative relationship between the room temperature hardness and the maximum compressive stress, the maximum compressive stresses at 600 and 800 °C are plotted as a function of the room temperature hardness (Fig. 14). It is interesting to note that there is an empirical relationship between the room temperature hardness, Hv (GPa), and the maximum compressive stress at 600 °C, σmax (MPa), as shown in Eq. (10). σ max ¼ −302:7 Hv þ 1303:4

Fig. 8. Grain size distributions for samples S1 through S6: (a) the number of grains and (b) the cross-section area of grains as a function of grain size. The cross-section area of grains in each size range is obtained by summarizing the areas from all the grains within that specific grain size range. Note that the maximum value in the Y-axis for bar charts (a) and (b) is 15% and 10% respectively as indicated and the original of each bar chart (not shown) is coincided with the maximum value of the next bar chart.

177

ð10Þ

Fig. 10. Area distributions with the morphology effect as a function of grain size for samples S1 through S6. The area distribution with the morphology effect in each size range is obtained by multiplying the area of each grain with its aspect ratio and summarizing these multiplied areas within that specific grain size range. Note that the maximum value in the Y-axis for each bar chart is 10% as indicated and the original of each bar chart (not shown) is coincided with the maximum value of the next bar chart.

178

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Table 2 Densities of various samples.

Table 3 Crack lengths emanating from the four corners of each indent for all samples.

Sample ID

Archimedes density (g/cm3)

Relative density (%)

S1 S2 S3 S4 S5 S6

14.8 14.8 15.0 14.9 14.8 14.8

99.5 98.9 100.3 99.8 99.0 99.1

Sample S6 is the only exception to this empirical relation. We hypothesize that this exception is due to its bi-modal grain size distribution which provides a higher toughness than S5, as discussed before. A higher toughness allows the sample to deform more before fracture, thereby leading to a higher maximum compressive stress. As mentioned before, sample S3 also contains bi-modal grain size distribution, but does not display a higher maximum compressive stress than that predicted from the empirical relation. This may be due to its larger average grain size which results in a lower hardness (Fig. 11). Nevertheless, its maximum compressive stress is slightly above the line of the empirical equation (Fig. 14), implying some effects from its bi-modal grain size distribution. At 800 °C the maximum compressive stresses of all samples have decreased from the corresponding values at 600 °C (Fig. 14). This is due to softening of the Co phase because the deformation behavior at 800 °C is predominantly controlled by the flow properties of the Co phase [20,21]. It appears that S1 and S2 soften more than other samples, as indicated by their maximum compressive stresses below the trend displayed by other samples. This phenomenon may be related to grain boundary sliding when the average grain size is very small (~0.17 μm for S1 and S2). Although a higher room temperature hardness can result in higher maximum compressive stresses at both 600 and 800 °C, we are surprised to note that the fracture strain in compression cannot be predicted from the room temperature toughness. Fig. 15 displays the fracture strains of all samples at 600 and 800 °C. By comparing Fig. 15 with Fig. 11, one can conclude immediately that a higher room temperature toughness does not result in a higher fracture strain. This is particularly true at 800 °C. For example, S3 has the highest toughness at room temperature, but its fracture strain is smaller than S2, S5 and S6 at 800 °C. We postulate that the absence of the correlation between the room temperature toughness and high temperature fracture strain is due to the change in the fracture mechanisms. At room temperature the toughness is controlled by crack growth resistance which is in turn dictated by fracture pathways. It is known that at room temperature there

Fig. 11. Comparison of hardness and fracture toughness of samples S1 through S6.

Indent ID

Crack ID

Sample ID and crack length (mm) S1

S2

S3

S4

S5

S6

1

L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4

31.7 33.7 29.1 31.7 30.5 34.2 30.9 32.7 31.3 33.6 33.6 33.2 29.9 34.3 29.5 37.3 31.5 36.3 32.3 33.6 31.5 36.5 31.9 38.2 31.3 35.0 31.5 35.5 32.4 37.6 29.5 36.3 29.7 35.9 28.5 37.4 28.6 32.5 30.1 36.3

29.6 34.7 29.5 35.3 26.4 32.7 26.9 33.5 29.1 35.4 27.5 34.0 28.9 35.8 30.2 37.0 27.6 31.3 28.6 32.4 22.5 29.6 25.1 26.9 28.9 27.7 29.3 28.7 30.4 33.3 26.6 35.1 26.7 33.3 26.7 31.4 30.9 31.2 23.2 31.0

20.4 21.9 18.1 24.3 20.8 22.4 15.9 22.2 19.1 20.0 17.7 19.3 14.8 15.5 15.1 23.4 18.8 18.7 19.5 27.6 16.9 18.8 12.8 22.6 15.5 19.7 17.0 24.6 17.1 23.6 16.4 24.0 20.0 11.5 19.7 23.3 19.2 20.7 15.9 16.0

23.2 25.2 21.7 23.9 24.0 25.8 25.1 29.6 23.1 25.2 22.0 29.8 22.7 27.6 24.4 32.1 22.9 28.5 24.6 30.4 24.0 27.8 24.8 27.3 25.4 25.4 23.9 26.3 25.0 27.6 23.9 29.2 26.0 25.6 26.0 26.0 26.0 25.6 26.0 26.0

31.0 33.6 28.9 37.5 32.7 36.1 31.1 36.9 31.9 37.8 33.3 34.3 29.3 37.2 35.3 38.0 30.2 35.6 29.9 37.7 31.9 36.6 32.6 38.2 30.9 35.9 33.4 37.1 31.8 35.4 30.2 37.8 32.4 35.1 32.2 37.0 35.5 33.9 32.8 32.2

26.0 30.2 28.2 35.1 25.6 31.8 29.1 31.4 25.1 33.4 27.4 32.0 26.5 33.4 27.3 33.9 25.6 33.7 27.0 34.8 26.7 30.5 29.2 33.6 27.4 30.8 28.3 34.3 27.3 29.5 26.6 33.0 27.0 32.6 29.0 33.4 28.6 32.2 33.4 30.2

2

3

4

5

6

7

8

9

10

are four major fracture pathways in WC-Co, i.e., cracks propagate (i) along WC/WC boundaries, (ii) with the transgranular fracture of WC crystals, (iii) near the WC/Co interface, and (iv) through the Co phase [12]. Normally, the transgranular fracture in WC only accounts for ~ 10% of the total crack path, while the other three fracture paths, which are strongly affected by crack deflection, constitute ~90% of the total crack path [12]. As a result of these fracture characteristics, the room temperature toughness increases with increasing WC grain sizes because large WC grain sizes lead to larger and longer crack deflection and thus more tortuous crack paths with higher energy consumption. At high temperatures the fracture strain of compressive tests is not dictated by crack propagation with little plastic deformation; instead, it is controlled by softening and significant plastic deformation of the Co phase. As displayed in Fig. 12(b), plastic flow starts at very low stresses (b 1000 MPa) when compression is conducted at 800 °C, unambiguously exhibiting significant plastic deformation of the Co phase. However, at 600 °C a large portion of deformation is elastic with some plastic deformation at the high stress region (N~3500 MPa), suggesting that brittle to ductile transition occurs around 600 °C. Although the precise fracture mechanism at high temperature remains to be investigated in the future, we hypothesize that small WC grain sizes are beneficial to the fracture strain at 800 °C (Fig. 15) because small WC grain sizes give rise to a small mean free path of the Co phase at a constant Co concentration, which in turn means that the Co phase deforms under high constraint from the neighboring WC grains and thus can reach high flow stresses [48]. High flow stresses at one location can force the Co phase

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Fig. 12. Typical stress-strain curves of samples S1 through S6: (a) at 600 °C and (b) at 800 °C.

at other locations to participate in plastic deformation. As a result of the flow participation of the Co phase at all locations, the sample exhibits a large fracture strain. In contrast, when the WC grain sizes are large, the mean free path of the Co phase is also large for a given Co concentration. Thus, the flow stress of the Co phase at any location will be low because of less constraint from the neighboring WC grains [48]. The low flow stress of the Co phase at one location means low capability to force the Co phase at other locations to significantly participate in plastic flow. As a result, plastic flow can concentrate at weak location(s), leading to a low fracture strain. This proposed mechanism appears to be in good agreement with the data shown in Fig. 15, i.e., S2, S5 and S6 have smaller grain sizes than S3 and S4 and thus possess higher fracture strains than S3 and S4 at 800 °C. S1 appears to be the only exception to this proposed mechanism. However, at this stage it is not clear what has resulted in this exception. It is noted that the fracture strain at 600 °C is much less predictable. For example, S3 has the highest room temperature toughness (Fig. 11) and also the highest fracture strain at 600 °C (Fig. 15), suggesting that the high toughness at room temperature still has some influence at this temperature. However, S5 has a low room temperature toughness, but a high fracture strain at 600 °C, indicating no influence of the room temperature toughness. The complication at 600 °C is likely due to multiple competing fracture mechanisms at this temperature because brittle to ductile transition occurs at around 600 °C, as discussed before.

179

Fig. 13. Comparisons of the room temperature hardness with the maximum compressive stress at different temperatures for samples S1 through S6: (a) at 600 °C and (b) at 800 °C.

Taking room temperature and high temperature properties together, it can be said that there is no single WC-Co material that can serve all purposes. For cutting tools with operating temperature at ~ 800 °C WC-Co with the average grain sizes at ~0.20 μm is a good choice because such WC-Co offers high maximum compressive stress and high fracture

Fig. 14. The maximum compressive stress as a function of the room temperature hardness for all the samples at two different temperatures.

180

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181

Fig. 15. Comparison of maximum fracture strain at 600 °C and °800C for samples S1 through S6.

strain simultaneously. For impact applications at room temperature WC-Co with large average grain sizes (~ 0.3 μm) would be desirable because of its high toughness. For wear applications with little impact at room temperature WC-Co with small average grain sizes should be employed because of its high hardness. For 600 °C applications the situation can be more complicated. Nevertheless, the general trend is that WC-Co with small grain sizes at ~ 0.20 μm is good because these WC-Co materials offer high maximum compressive stresses (Fig. 13) with reasonable fracture strains (around 9% or higher as shown in Fig. 15). Finally, it should be emphasized that the trends identified in this study are based on the limited number of measurements (3 to 5 hightemperature tests per WC-Co specimen) and derived from samples with grain sizes in the range of 0.15 to 0.4 μm. Nevertheless, the trends identified are consistent with many prior studies on WC-Co with grain sizes near 1 μm or larger, particularly in the softening temperature of WC-Co and the dependence of its high temperature hardness and strength on the room temperature hardness [20,22–27]. However, the trends observed for the fracture strain are more complicated than the fracture strength. In particular, the fracture strain at 600 °C does not show a clear dependence on the grain size and the room-temperature hardness or toughness. Thus, more experiments are needed in the future to define whether there is any trend in the fracture strain at 600 °C with high confidence.

4. Summary and conclusions WC–6 wt% Co materials with grain sizes at submicrometer ranges (b0.30 μm) from various companies in multiple countries have been investigated for their relationships between room temperature properties (i.e., hardness and toughness) and high temperature compressive properties (compressive strength and fracture strain) and the dependence of these properties on the grain size and size distribution. This study has led to the following conclusions. 1. Small grain size leads to high hardness, whereas large grain size results in high toughness at room temperature. 2. A bi-modal grain size distribution at a constant average grain size (number-averaged) can lead to improvement in toughness with little compromise in hardness at room temperature. 3. The maximum compressive stresses at 600 and 800 °C increase with increasing the room temperature hardness. An empirical relationship between the room temperature hardness and the maximum compressive stress at 600 °C has been identified, as shown in Eq. (10).

4. Since the room temperature hardness increases with decreasing grain size, WC-Co with small grain sizes have higher maximum compressive stresses at 600 and 800 °C than the counterparts with larger grain sizes. 5. A bi-modal grain size distribution at a constant average grain size can enhance the maximum compressive stresses at 600 °C, particularly when the average grain size is small (b~0.21 μm). Such a phenomenon is likely due to a good combination of high hardness (due to small grain sizes) and high toughness (due to the bi-modal grain size distribution). 6. The fracture strain in compression at 600 and 800 °C cannot be predicted from the room temperature toughness. At 800 °C the fracture strain of compressive tests is dictated by significant plastic flow of the Co phase, whereas the toughness at room temperature is controlled by crack propagation with little plasticity. Because of the different deformation and fracture mechanisms the fracture strain at 800 °C is not correlated with the room temperature toughness. 7. Small grain sizes are beneficial to the fracture strain at 800 °C. This trend is likely related to the flow participation of the Co phase at all locations, which becomes possible when the mean free path of the Co phase is small and thus the Co phase deforms under high constraint from the neighboring WC grains. High constrained deformation leads to high flow stresses which in turn force the Co phase at other locations to participate in plastic deformation and thus a high fracture strain of the sample. 8. The transition from brittle to ductile fracture occurs at ~600 °C. As a result, multiple competing fracture mechanisms take place at this temperature, thereby no obvious relationship between the room temperature toughness and the fracture strain at 600 °C. Nevertheless, high toughness at room temperature may still result in high fracture strain at 600 °C because brittle fracture is one of the competing mechanisms at 600 °C. 9. This study demonstrates that the design of microstructure even for a fixed composition (like WC–6 wt% Co) depends on the intended application. For example, if the intended application is at 800 °C, an average grain size at ~ 0.2 μm is a good choice because such a WCCo offers high maximum compressive stress and high fracture strain simultaneously. For 600 °C applications an average grain size at ~ 0.2 μm with a bi-modal grain size distribution could be a good choice because these WC-Co materials can offer a good balance in the maximum compressive stress at 600 °C with reasonable fracture strain as well as a high room temperature toughness. Acknowledgements The authors S.E., A.R and L.S. acknowledge the financial support from GrandMetals Co., Ltd. GM2013032001, Mengzhou, Henan, China. A.R. is also thankful to the financial support of the Brazil Scientific Mobility Program (BSMP) sponsored by CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico). References [1] H.E. Exner, Physical and chemical nature of cemented carbides, Int. Met. Rev. 24 (1979) 149–173, http://dx.doi.org/10.1179/imtr.1979.24.1.149. [2] M.M. Schwartz, Engineering Applications of Ceramic Materials: A Source Book, Materials Park, 1985. [3] Z.Z. Fang, X. Wang, T. Ryu, K.S. Hwang, H.Y. Sohn, Synthesis, sintering, and mechanical properties of nanocrystalline cemented tungsten carbide – a review, Int. J. Refract. Met. Hard Mater. 27 (2009) 288–299, http://dx.doi.org/10.1016/j.ijrmhm. 2008.07.011. [4] I. Konyashin, B. Ries, F. Lachmann, Near-nano WC-Co hardmetals: will they substitute conventional coarse-grained mining grades? Int. J. Refract. Met. Hard Mater. 28 (2010) 489–497, http://dx.doi.org/10.1016/j.ijrmhm.2010.02.001. [5] K. Jia, T.E. Fischer, Abrasion resistance of nanostructured and conventional cemented carbides, Wear 200 (1996) 206–214, http://dx.doi.org/10.1016/S00431648(96)07277-8. [6] K. Jia, T.E. Fischer, Sliding wear of conventional and nanostructured cemented carbides, Wear 203–204 (1997) 310–318, http://dx.doi.org/10.1016/S0043-1648(96)07423-6.

S.V. Emani et al. / Int. Journal of Refractory Metals and Hard Materials 58 (2016) 172–181 [7] H.C. Lee, J. Gurland, Hardness and deformation of cemented tungsten carbide, Mater. Sci. Eng. 33 (1978) 125–133, http://dx.doi.org/10.1016/0025-5416(78)90163-5. [8] Z.H. Xu, J. Ågren, A modified hardness model for WC-Co cemented carbides, Mater. Sci. Eng. A 386 (2004) 262–268, http://dx.doi.org/10.1016/j.msea.2004.07.061. [9] M. Laugier, A microstructural model for hardness in WC–Co composites, Acta Metall. 33 (1985) 2093–2099, http://dx.doi.org/10.1016/0001-6160(85)90132-4. [10] L. Makhele-Lekala, S. Luyckx, F.R.N. Nabarro, Semi-empirical relationship between the hardness, grain size and mean free path of WC-Co, Int. J. Refract. Met. Hard Mater. 19 (2001) 245–249, http://dx.doi.org/10.1016/S0263-4368(01)00022-1. [11] B. Roebuck, E.A. Almond, Deformation and fracture processes and the physical metallurgy of WC–Co hardmetals, Int. Mater. Rev. 33 (1988) 90–112. [12] L.L. Shaw, H. Luo, Y. Zhong, WC-18 wt% Co with simultaneous improvements in hardness and toughness derived from nanocrystalline powder, Mater. Sci. Eng. A 537 (2012) 39–48, http://dx.doi.org/10.1016/j.msea.2012.01.005. [13] D. Sivaprahasam, S.B. Chandrasekar, R. Sundaresan, Microstructure and mechanical properties of nanocrystalline WC–12Co consolidated by spark plasma sintering, Int. J. Refract. Met. Hard Mater. 25 (2007) 144–152, http://dx.doi.org/10.1016/j.ijrmhm. 2006.03.008. [14] A. Michalski, D. Siemiaszko, Nanocrystalline cemented carbides sintered by the pulse plasma method, Int. J. Refract. Met. Hard Mater. 25 (2007) 153–158, http://dx.doi.org/10.1016/j.ijrmhm.2006.03.007. [15] S. Mi, T.H. Courtney, Synthesis of WC and WC-Co cermets by mechanical alloying and subsequent hot isostatic pressing, Scr. Mater. 38 (1997) 171–176. [16] S.I. Cha, S.H. Hong, B.K. Kim, Spark plasma sintering behavior of nanocrystalline WC– 10Co cemented carbide powders, Mater. Sci. Eng. A 351 (2003) 31–38, http://dx.doi. org/10.1016/S0921-5093(02)00605-6. [17] J.L. Chermant, F. Osterstock, Fracture toughness and fracture of WC-Co composites, J. Mater. Sci. 11 (1976) 1939–1951. [18] S.V. Emani, C. Wang, L.L. Shaw, Z. Chen, On the hardness of submicrometer-sized WC–Co materials, Mater. Sci. Eng. A 628 (2015) 98–103. [19] W. Acchar, U.U. Gomes, W.A. Kaysser, J. Goring, Strength degradation of a tungsten carbide-cobalt composite at elevated temperatures, Mater. Charact. 43 (1999) 27–32, http://dx.doi.org/10.1016/S1044-5803(98)00056-4. [20] K. Buss, D. Mari, High temperature deformation mechanisms in cemented carbides and cermets studied by mechanical spectroscopy, Mater. Sci. Eng. A 370 (2004) 163–167, http://dx.doi.org/10.1016/j.msea.2002.12.004. [21] D. Mari, S. Bolognini, G. Feusier, T. Viatte, W. Benoit, Experimental strategy to study the mechanical behaviour of hardmetals for cutting tools, Int. J. Refract. Met. Hard Mater. 17 (1999) 209–225, http://dx.doi.org/10.1016/S0263-4368(98)00078-X. [22] I.C. Lee, T. Sakuma, High-temperature tensile ductility in WC-Co cemented carbides, Metall. Mater. Trans. A 28 (1997) 1843–1847, http://dx.doi.org/10.1007/s11661997-0114-6. [23] Y.V. Milman, S. Luyckx, V.A. Goncharuck, J.T. Northrop, Results from bending tests on submicron and micron WC-Co grades at elevated temperatures, Int. J. Refract. Met. Hard Mater. 20 (2002) 71–79, http://dx.doi.org/10.1016/S0263-4368(01)00072-5. [24] F. Ueda, H. Doi, F. Fujiwara, H. Masatomi, Y. Oosawa, Tensile creep of WC-10% Co and WC-10% TaC-10% Co at elevated temperatures, Trans. Jpn. Inst. Metals 16 (1975) 591–600. [25] M.T. Laugier, Elevated temperature properties of WC-Co cemented carbides, Mater. Sci. Eng. A 105 (1988) 363–367, http://dx.doi.org/10.1016/0025-5416(88)90718-5. [26] Y.V. Milman, S. Luyckx, I. Northrop, Influence of temperature, grain size and cobalt content on the hardness of WC-Co alloys, Int. J. Refract. Met. Hard Mater. 17 (1999) 39–44, http://dx.doi.org/10.1016/S0263-4368(98)00038-9. [27] Y.V. Milman, S. Chugunova, V. Goncharuck, S. Luyckx, I.T. Northrop, Low and high temperature hardness of WC-6 wt% Co alloys, Int. J. Refract. Met. Hard Mater. 15 (1997) 97–101, http://dx.doi.org/10.1016/S0263-4368(97)81231-0. [28] K. Niihara, A fracture mechanics analysis of indentation-induced Palmqvist crack in ceramics, J. Mater. Sci. Lett. 2 (1983) 221–223, http://dx.doi.org/10.1007/BF007 25625.

181

[29] K. Niihara, R. Morena, D.P.H. Hasselman, Evaluation of KIc of brittle solids by the indentation method with low crack-to-indent ratios, J. Mater. Sci. Lett. 1 1982 January (1) 13–16. doi:http://dx.doi.org/10.1007/BF00724706. [30] G.D. Quinn, R.C. Bradt, On the Vickers indentation fracture toughness test, J. Am. Ceram. Soc. 90 (2007) 673–680, http://dx.doi.org/10.1111/j.1551-2916.2006. 01482.x. [31] D.K. Shetty, I.G. Wright, P.N. Mincer, A.H. Clauer, Indentation fracture of WC-Co cermets, J. Mater. Sci. 20 (1985) 1873–1882, http://dx.doi.org/10.1007/BF00555296. [32] F.A. Costa, F.F.P. De Medeiros, A.G.P. Silva, U.U. Gomes, M. Filgueira, C.P. De Souza, Structure and hardness of a hard metal alloy prepared with a WC powder synthesized at low temperature, 485 (2008) 638–642, http://dx.doi.org/10.1016/j.msea. 2007.08.058. [33] B. Reichel, K. Wagner, D.S. Janisch, W. Lengauer, Alloyed W–(Co, Ni, Fe)–C phases for reaction sintering of hardmetals, Int. J. Refract. Met. Hard Mater. 28 (2010) 638–645, http://dx.doi.org/10.1016/j.ijrmhm.2010.06.003. [34] V. Bounhoure, S. Lay, M. Loubradou, J.M. Missiaen, Special WC/Co orientation relationships at basal facets of WC grains in WC-Co alloys, J. Mater. Sci. 43 (2008) 892–899, http://dx.doi.org/10.1007/s10853-007-2181-x. [35] S.-W. Jung, J. Kim, S.-J.L. Kang, Etching for microstructural observation of cemented submicrometer-sized carbides, J. Am. Ceram. Soc. 84 (2001) 899–901, http://dx.doi. org/10.1111/j.1151-2916.2001.tb00765.x. [36] J. Wang, L.L. Shaw, Morphology-enhanced low-temperature sintering of nanocrystalline hydroxyapatite, Adv. Mater. 19 (2007) 2364–2369. [37] L. Sun, C. Jia, R. Cao, C. Lin, Effects of Cr3C2 additions on the densification, grain growth and properties of ultrafine WC-11Co composites by spark plasma sintering, Int. J. Refract. Met. Hard Mater. 26 (2008) 357–361, http://dx.doi.org/10.1016/j. ijrmhm.2007.08.009. [38] J. Weidow, H.O. Andrén, Grain and phase boundary segregation in WC-Co with small V, Cr or Mn additions, Acta Mater. 58 (2010) 3888–3894, http://dx.doi.org/ 10.1016/j.actamat.2010.03.038. [39] J. Weidow, S. Norgren, H.O. Andrén, Effect of V, Cr and Mn additions on the microstructure of WC-Co, Int. J. Refract. Met. Hard Mater. 27 (2009) 817–822, http://dx.doi.org/10.1016/j.ijrmhm.2009.02.002. [40] T. Yamamoto, Y. Ikuhara, T. Sakuma, ADVANCED high resolution transmission electron microscopy study in VC-doped WC–Co compound, Sci. Technol. Adv. Mater. 1 (2000) 97–104. [41] S. Lay, S. Hamar-Thibault, A. Lackner, Location of VC in VC, Cr3C2 codoped WC-Co cermets by HREM and EELS, Int. J. Refract. Met. Hard Mater. 20 (2002) 61–69, http://dx.doi.org/10.1016/S0263-4368(01)00071-3. [42] J. Zackrisson, B. Jansson, G.S. Uphadyaya, H.-O. Andr'en, WC-Co based cemented carbides with large Cr3C2 additions, Int. J. Refract. Met. Hard Mater. 16 (1998) 417–422, http://dx.doi.org/10.1016/S0263-4368(98)00048-1. [43] A. Upadhyaya, D. Sarathy, G. Wagner, Advances in alloy design aspects of cemented carbides, Mater. Des. 22 (2001) 511–517, http://dx.doi.org/10.1016/S0261-3069 (01)00003-6. [44] A. Delanoë, M. Bacia, E. Pauty, S. Lay, C.H. Allibert, Cr-rich layer at the WC/Co interface in Cr-doped WC-Co cermets: segregation or metastable carbide? J. Cryst. Growth 270 (2004) 219–227, http://dx.doi.org/10.1016/j.jcrysgro.2004.05.101. [45] S. Luyckx, M. Zunaid Alli, Comparison between V8C7 and Cr3C2 as grain refiners for WC-Co, Mater. Des. 22 (2001) 507–510, http://dx.doi.org/10.1016/S0261-3069 (01)00004-8. [46] E. Breval, J.P. Cheng, D.K. Agrawal, P. Gigl, M. Dennis, R. Roy, et al., Comparison between microwave and conventional sintering of WC/Co composites, Mater. Sci. Eng. A 391 (2005) 285–295, http://dx.doi.org/10.1016/j.msea.2004.08.085. [47] W.D. Schubert, H. Neumeister, G. Kinger, B. Lux, Hardness to toughness relationship of fine-grained WC-Co hardmetals, Int. J. Refract. Met. Hard Mater. 16 (1998) 133–142. [48] L. Xiao, R. Abbaschian, On the flow behavior of constrained ductile phases, Metall. Trans. A 24 (1993) 403–415.