Available online at www.sciencedirect.

com

International Journal of Refractory Metals & Hard Materials 26 (2008) 6876 www.elsevier.com/locate/IJRMHM

Fracture of WCNi cemented carbides with dierent shape of WC crystals q

A.V. Shatov
a

a,b,*

, S.S. Ponomarev a, S.A. Firstov

Frantsevich Institute for Problems of Materials Science, 3 Krzhizhanovsky St., 03680 Kiev-142, Ukraine b Mercer Management Consulting, Inc., 125 Village Boulevard, Suite 270, Princeton, NJ 08540, USA Received 25 February 2007; accepted 18 March 2007

Abstract The dependences of fracture toughness and crack propagation on the shape equiaxiality of WC crystals are studied on WCNi cemented carbides with small addition of TiC. The fracture toughness K1C linearly correlates to mean linear path in binder phase regardless of the shape and contiguity of WC crystals as well as despite of the fracture path change in studied WCNi cemented carbides. It is experimentally shown that the fracture toughness K1C does not correlate versus contiguity of carbide crystals on WCNi cemented carbides when shape of WC crystals is changed. The relative amount of trans-crystalline fracture through the carbide crystals and the total area of fracture through the carbide phase increase, whereas, the relative amount of inter-crystalline fracture along carbidecarbide boundaries decrease on WCNi cemented carbides with atter WC crystals and lower values of contiguity of WC crystals. 2007 Elsevier Ltd. All rights reserved.
Keywords: Cemented carbides; Fracture toughness; Crack path; Shape of carbide crystals

1. Introduction The fracture of cemented carbides is believed to be a two-step process. The brittle inter- and trans-crystalline fracture of carbide phase is followed by the ductile rupture of binder phase and carbidebinder interfaces [1 12]. The area fraction of crack path through the carbide phase accounts for at least 5080% of the fracture surface area of cemented carbides and increases with the amount of carbide phase V C in the alloy [57,1317]. The increasV ing value of the mean linear intersect d of WC crystals causes the increase of the ratio of the surface area of trans-crystalline fracture through the carbide crystals (C) versus the inter-crystalline fracture of the carbide phase
This paper was partially presented at the 16th Plansee Seminar, Reutte, Austria, May 2005. * Corresponding author. Address: 8 Parkside Drive, Jamesburg, NJ 08831, USA. Tel./fax: +1 732 605 0416. E-mail address: alex_shatov@hotmail.com (A.V. Shatov). 0263-4368/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijrmhm.2007.03.002
q

(C/C) and at the same time the increase of the ratio of the surface area of the fracture through the binder phase (B) versus the fracture along the carbidebinder interfaces (B/C) regardless of the volume fraction of the carbide phase V C in the alloy [57,1317]. However, it is believed V that the rupture of the carbide phase makes a minor contribution to the fracture toughness of the cemented carbides due to relatively low value of the critical strain energy release rate of carbides (for WC G1C % 50 J/m2) compared to that of the cemented carbides (200500 J/m2) [4,5,7,1820]. The major contribution is made by the plastic deformation and rupture of the binder phase ligaments and carbidebinder interfaces. Despite the theoretical estimations of fracture energies, two experimental correlations of fracture toughness versus the mean linear path in binder phase k and versus the carbide crystals contiguity G are observed simultaneously (Fig. 1) [1,17]. The fracture toughness K1C increases linearly with k (Fig. 1a) but decreases with the contiguity G (Fig. 1b).

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876

69

Fig. 1. The correlation of the fracture toughness K1C with the mean linear path in binder phase k (a) and with the carbide crystals contiguity G (b) of WC Co cemented carbides with dierent mean diameters of WC crystals (DWC) and volume fraction of the binder phase between 5 and 37 vol% [1]; and with the mean linear intersects of WC crystals from 2.1 to 3.6 lm and volume fraction of the binder phase between 13 and 36 vol% [17].

The mean linear path in binder phase k is related to the contiguity G through the well-known stereological equation: k 1VC V C d V V 1 G 1

where V C is the volume fraction of carbide phase, and d is V the mean linear intersect of the carbide crystals. It is shown by numerous experiments on cemented carbides with unchanged shape of WC crystals that the contiguity G mainly depends on the volume fraction of the carbide phase V C V [2123]. Obviously, the change of the volume fraction of the carbide phase V C should be considered as a main even V though not direct reason for the observed correlations of K1C. To experimentally separate the eects of these stereological parameters on the value of K1C, k and G need to be changed regardless of the volume fraction of the carbide phase and the size of carbide crystals. In addition, no results were reported in regard to the dependence of the crack propagation on the contiguity and the shape of WC crystals in cemented carbides, whereas, the volume fraction of carbide phase V C is kept V constant in the alloy. Partially, this is due to the fact that the contiguity of carbide crystals is believed to be mainly determined by the volume fraction V C of the carbide phase V for conventional cemented carbides [2123]. On the other hand, not many results were reported about changing the shape of the carbide crystals in cemented carbides. As it was recently reported, the contiguity of WC carbide crystals can change even for a xed volume fraction of carbide phase when the shape of the WC crystals is changed [24]. The change of the shape to a atter triangular prism is produced by the addition of very small amount of TiC into WCNi cemented carbides [25]. The atter shape of WC crystals corresponds to a lower value of the shape equiaxiality PWC shape factor that was introduced and

measured in [25]. As it was found, the contiguity of carbide crystals G decreases when the shape equiaxiality of WC crystals PWC decreases [24]. The current study concentrates on the eect of the shape of WC crystals on the correlations of fracture toughness versus the mean linear path in binder phase k and versus the carbide crystals contiguity G on WCNi cemented carbides. In addition, the eects of the shape equiaxiality and the contiguity of WC crystals on the crack propagation in WCNi cemented carbides are studied by the qualitative comparison of the SEM micrographs and Auger electron spectroscopy of fracture surfaces of the alloys. 2. Experimental details A set of WCNi cemented carbides with binder contents of 8, 14, 22 wt% and small additions of TiC was produced by liquid phase sintering in vacuum for 1 h [25]. Quantitative electron-probe X-ray wavelength dispersive microanalysis (Superprobe-733, JEOL) with relatively large diameter of electron beam of about 50 lm was used to measure the chemical composition of the alloys. The measured values of Ti content in the alloys were between 0.04 and 0.4 wt%. The concentration of Ti in Ni binder phase was then calculated under assumption that Ti does not dissolve in WC. The specimens were sectioned, polished, and etched by the standard techniques for cemented carbides. Scanning electron microscopy (JSM T-20, JEOL) was used to produce micrographs. Standard quantitative metallographic parameters were measured on enlarged micrographs. That includes phase volume contents; mean linear intersects of carbide and binder phases; and the contiguity of WC crystals. In addition, the shape equiaxiality of WC crystals PWC was measured [25]. The critical stress intensity factor a.k.a. the fracture toughness K1C was measured by three-point bending of

70

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876

Table 1 Microstructure parameters and fracture toughness of WCNi cemented carbides with small TiC addition [24,25] Ni content (wt%) 8 8 14 14 14 14 14 22 22 Sintering temperature (C) 1500 1450 1450 1500 1400 1450 1500 1400 1400 Ti concentration in binder phase (wt%) %3.7 0.91 0.72 0.72 0.22 0.22 0.22 1.18 0.48 Mean linear intersect of WC crystals, d (lm) 0.55 0.03 0.76 0.03 0.83 0.03 0.91 0.03 0.89 0.03 0.93 0.03 0.91 0.03 1.01 0.03 0.94 0.03 Shape equiaxiality of WC crystals PWC 0.37 0.45 0.48 0.47 0.61 0.63 0.64 0.43 0.59 Contiguity of WC crystals G 0.65 0.58 0.24 0.25 0.34 0.32 0.30 0.11 0.18 Mean linear path in binder phase, k (lm) 0.25 0.01 0.32 0.01 0.35 0.02 0.38 0.02 0.41 0.02 0.41 0.02 0.39 0.02 0.56 0.02 0.58 0.02 Fracture toughness, K1C (MPa m1/2) 11.9 0.3 14.1 0.4 14.7 0.4 14.9 0.4 15.8 0.4 15.9 0.5 15.2 0.6 19.5 0.4 19.6 0.7 Porosity

%5% None None None None None None None None

810 ground specimens. Notches were introduced by spark erosion. Specimen cross section was (5 0.25) (5 0.25) mm2, while the longitudinal size was at least 35 mm, and knife supports were (20 0.5) mm apart. The fracture surface of the WCNi cemented carbides was studied by scanning electron microscopy (JSM T-20, JEOL) after the measurements of K1C. Separate specimens of the WCNi cemented carbides were fractured in a high vacuum chamber of Auger-electron spectrometer (JAMP10S, JEOL) at about (510) 107 Pa. The Auger-electron specters were taken from the relatively big areas of the fracture surfaces right after the fracture and additionally after 30 min of Ar+ ion beam etching in the high vacuum. 3. The eect of the shape of WC crystals on the fracture toughness of cemented carbides 3.1. Experimental results on microstructure and fracture toughness The results of the metallographic microstructure and fracture toughness measurements are presented in Table 1. All the sintered alloys except WC-8 wt% Ni alloy with the highest titanium concentration in binder phase of about 3.7 wt% were non-porous two-phase WCNi cemented carbides with the mean linear intersect of WC crystals between 0.75 and 1 lm, regardless of their shape. The WC-8 wt% Ni alloy with 3.7 wt% Ti concentration in Ni binder phase sintered even at 1500 C still exhibited high porosity of about 5%, while the mean linear intersect of WC crystals was only 0.55 lm that is less than on all the other sintered alloys. The porosity appears to be caused by the presence of the third phase enriched in Ti and detected by the X-ray microanalysis of the alloy. The presence of the third phase seems to decrease the amount of the liquid binder phase that retarded the shrinkage and slowed down the structure development of the alloy. Therefore, the experimental data of this alloy is presented for illustration purpose only. The microstructures of WC-14 wt% Ni with dierent Ti content are shown in Fig. 2. WC crystals have a well faceted shape. As reported in Ref. [25], Ti segregation on

Fig. 2. Microstructure of WC-14 wt% Ni cemented carbides with dierent shape equiaxiality of WC crystals: (a) PWC = 0.63 and (b) PWC = 0.48.

the carbidebinder interface causes the shape of WC crystals to become atter triangular prisms. The atter the prism, the lower the value of shape equiaxiality PWC of WC crystals [25]. It was previously observed that the shape equiaxiality PWC decreases monotonously with Ti concentration in the Ni binder phase regardless of Ni content in these WCNi cemented carbides [25]. The contiguity decreases with the decrease in the shape equiaxiality PWC of WC crystals for both non-porous WCNi alloys [24].

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876

71

Fig. 3. The dependence of the fracture toughness K1C on the shape equiaxiality of WC crystals PWC (a) and Ti concentration in binder phase (b) for WC Ni cemented carbides with dierent binder content. Presence of the pores decreases the fracture toughness of the alloys.

3.2. Dependence of K1C versus the shape equiaxiality PWC and Ti concentration in binder The measured dependence of the fracture toughness K1C on the shape equiaxiality of WC crystals PWC and Ti concentration in binder phase for WCNi cemented carbides with dierent binder content are shown in Fig. 3. The fracture toughness K1C does not show much of the change on the alloys with high content of the binder phase WC-22 wt% Ni. K1C slightly decreases with the decrease in shape equiaxiality of WC crystals and with the increase in Ti concentration in binder phase on non-porous WC-14 wt% Ni alloys. Pores sharply decrease the value of the fracture toughness on WC-8 wt% Ni. In general, both the change in chemical composition of the WCNi alloys and associated change of the shape of WC crystals can aect the fracture toughness. The addition

of TiC can change the properties of the binder phase, carbidebinder interfaces, and possibly the properties of the carbidecarbide boundaries. The shape equiaxiality of WC crystals aects the contiguity G [24] and consequently causes the redistribution of the binder phase and a decrease in the mean linear path in binder phase k (Table 1). In any case, if the shape and chemistry can directly aect the fracture toughness of the alloy, the deviation should occur on the correlations of K1C versus mean linear path in binder phase k and the contiguity of carbide crystals that are usually observed on cemented carbides (Fig. 1). 3.3. Correlations of K1C versus mean linear path in binder phase and contiguity of WC crystals The measured dependencies of the fracture toughness K1C on the mean linear path in binder phase k and the

Fig. 4. The correlation of the fracture toughness K1C versus the mean linear path in binder phase k (a) and versus the contiguity G of WC crystals (b) on WCNi cemented carbides with dierent shape equiaxiality of WC crystals PWC and dierent binder content.

72

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876

carbide crystals contiguity G are shown in Fig. 4. The fracture toughness K1C of the WCNi cemented carbides linearly increases with the mean linear path in binder phase k regardless of the shape change of WC crystals, Ti concentration in the binder phase, and the volume fraction of the carbide phase (Fig. 4a). On the contrary, the usually observed monotonous dependence of K1C on the contiguity G of WC crystals (Fig. 1b) breaks into two opposite dependencies. On one hand, the fracture toughness K1C increases with the contiguity of WC crystals caused by the growth in PWC and k when the volume fraction of the carbide phase is xed on both non-porous WC-14 wt% Ni and WC-22 wt% Ni cemented carbides (Fig. 4b). On the other hand, K1C decreases with the volume fraction of the carbide phase and related increase in the contiguity G when the shape of WC crystals is not changed (Fig. 4b). The narrow correlation of K1C with the contiguity G shown in Fig. 1b widens up into a range of values for each volume fraction of carbide phase when the shape of WC crystals is changed (Fig. 4b). In other words, the fracture toughness K1C does not directly correlate to the carbide crystals contiguity G on WCNi cemented carbides with dierent shape of WC crystals. 3.4. Summary of the correlations of K1C Thus, it can be concluded that the shape of WC crystals aects the value of fracture toughness K1C indirectly through the change in the contiguity of WC crystals and the associated change of the mean linear path in binder phase k according to stereological relationship (1). The resulted value of fracture toughness linearly correlates to k and is not related directly to the shape and/or the contiguity of WC crystals. The persistence of the correlation of fracture toughness versus k indicates that the small amount of Ti dissolved in the binder phase does not aect the fracture properties of the binder and interfaces in the alloys. 4. Crack propagation change due to the shape of WC crystals 4.1. SEM study of crack propagation change due to the shape of WC crystals in cemented carbides The micrographs of the fracture surfaces of the WCNi cemented carbides with dierent shape of WC crystals are presented in Fig. 5. A qualitative comparison of the SEM micrographs of the fracture surfaces reveals a change in the fracture path. The trans-crystalline fracture through the carbide crystals (C) occurs very rarely, whereas, the inter-crystalline fracture along the carbidecarbide interfaces (C/C) identied by a smooth carbide surface with no visible plastic deformation or dimples is widely present on the fracture surface of the alloys with relatively high value of the shape equiaxiality PWC (Fig. 5a). On the contrary, the trans-crystalline fracture through the carbide phase (C) identied by the river-like marking

Fig. 5. Fracture surface of WC-14 wt% Ni cemented carbides with dierent shape equiaxiality of WC crystals: (a) PWC = 0.63; (b) PWC = 0.48. Fracture modes: C carbide trans-crystalline; C/C carbide inter-crystalline; B/C fracture along the carbidebinder interfaces; B fracture through the binder phase.

pattern is observed quite often on the fracture surface of the alloys with the lower value of the shape equiaxiality of WC crystals PWC corresponding to a atter triangular prism shape of WC crystals (Fig. 5b). The inter-crystalline fracture along the carbidecarbide interfaces (C/C) is still present on these alloys (Fig. 5b), although not as much as on the alloys with the higher shape equiaxiality PWC and contiguity G of WC crystals (Fig. 5a). Even though both the fracture along the carbidebinder interfaces (B/C) characterized by closely spaced shallow dimples over carbide crystal surface and the fracture mode through the binder phase (B) characterized by deep dimples are widely present on these alloys, no visibly signicant change in their amount can be found by the rough qualitative analysis of the fracture surfaces (Fig. 5). 4.2. Auger-electron spectroscopy of the fracture surface 4.2.1. Ni to W ratio The results of Auger-electron spectroscopy of the fracture surfaces are presented in Table 2. The study of the fracture surface using Auger-electron spectroscopy

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876 Table 2 Ratio of Auger-electron dierential peak intensities on the fracture surface of WCNi cemented carbides with small TiC addition Ni content (wt%) Ti concentration in binder phase (wt%) Shape equiaxiality of WC crystals (PWC) 0.63 0.48 0.59 0.43 Ni to W ratio (INi/IW) Fresh fracture surface 5.8 4.7 8.3 6.2 After 30 min of Ar ion etching 2.0 1.6 5.0 3.0
+

73

Ti to W ratio (ITi/IW) Fresh fracture surface 0.14 0.13 0.23 0.21 After 30 min. of Ar+ ion etching 0.07 0.02 0.14 0.13

14 14 22 22

0.22 0.72 0.48 1.18

complements the results of the fracture surface analysis. The average ratio of nickel to tungsten Auger-electron differential peak intensities INi/IW measured on the fracture surface versus the shape equiaxiality PWC of WC crystals for both WC-14 wt% Ni and WC-22 wt% Ni cemented carbides is shown in Fig. 6a. The value of INi/IW is higher for the alloys with higher Ni binder content. The decrease of the relative intensity INi/IW after thorough etching of the fracture surfaces with Ar+ ions for 30 min indicates the ignicant presence of thin layers of Ni over WC crystals on the fresh fracture surface. These layers of Ni are obviously a result of the plastic deformation and fracture through the binder (B) and along the carbidebinder interfaces (B/C). The value of INi/IW decreases when PWC decreases on both WC-14 wt% Ni and WC-22 wt% Ni cemented carbides (Fig. 6a). The tendency of INi/IW to decrease with the decrease of PWC remains even after Ar+ ion beam etching. The unchanged tendency of INi/IW decrease with the decrease of PWC as on the fresh fracture surface, so as on the fracture surface after removing the thin layers of Ni, can be considered as an evidence of the eect of the shape equiaxiality PWC and the contiguity of WC crystals on the crack path through the carbide phase for both binder contents in the alloys regardless of the binder phase fracture.

To interpret the decrease in value of INi/IW when PWC decreases, let us consider the contribution of dierent fracture modes to the relative intensities of INi and IW. Disregarding the contribution of the atomic layers of the binder metal left on the carbide side of carbidebinder interfaces and present on the carbidecarbide interfaces as well as the contribution of tungsten dissolved in the binder phase, the Auger-electron dierential peak intensity of an element can be estimated as being proportional to the relative fracture surface area AiA of the phases that contain the element [1317]. Where i stand for one of the fracture modes, namely B, B/C, C/C, or C as described above. In this case, the ratio INi/IW is proportional to I Ni AB 0:5 AA / C A C=C : I W AA AA 0:5 AB=C A
B=C B=C

Here the carbidebinder fracture mode AA is assumed to contribute equally to the intensity of Ni in the binder on one side of the fracture surface and W in carbide phase on the other side of the fracture surface after the separation of the carbide and binder phases along the carbidebinder interface. According to formula (2), to decrease the ratio INi/IW for a given binder content in the alloy, either the area of

Fig. 6. The ratio of Ni to W (a) and Ti to W (b) Auger-electron dierential peak intensities versus the shape equiaxiality of WC crystals PWC from the fresh fracture surfaces prepared in high vacuum (circle dots) and after Ar+ ion beam etching for 30 min (square dots) on WC-14 wt% Ni and WC-22 wt% Ni cemented carbides.

74

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876

the fracture though the binder phase AB should decrease or A C=C the total area through the carbide phase (AC AA ) should A increase. The mean linear path in binder phase k decreases with the decrease of PWC on WC-14 wt% Ni cemented carbides (Table 1). The narrowing binder spaces can cause the area of the fracture through the binder phase AB to A decrease and perhaps increase the fracture area along the B=C carbidebinder interfaces AA . However, k only slightly changes on WC-22 wt% Ni alloys (Table 1) and cannot be associated with even bigger relative change in the ratio INi/IW with PWC on this alloy (Fig. 6a). Therefore, the decrease of the mean linear path in binder phase k cannot be considered as a main or only reason of the decrease in INi/IW with PWC. Hence, to decrease the value of INi/IW with PWC the total area of the fracture through the carbide C=C phase (AC AA ) should increase. A 4.2.2. Ti to W ratio The average ratio of titanium to tungsten Auger-electron dierential peak intensities ITi/IW measured on the fracture surface versus the shape equiaxiality PWC of WC crystals for both WC-14 wt% Ni and WC-22 wt% Ni cemented carbides is shown in Fig. 6b. It would be natural to expect that the relative amount of Ti on the fracture surface should be higher in the alloys with higher Ti content, i.e., in the alloy characterized by the lower value of the shape equiaxiality of PWC. Nevertheless, the relative amount of Ti decreases (Fig. 6b). The value of ITi/IW is higher for the alloys with higher Ni binder content. In addition, the value of ITi/IW is decreased by the Ar+ ion beam etching in roughly similar proportion as INi/IW on both alloys (Fig. 6a and b). This similarity to the Ni binder distribution indicates that titanium is dissolved in the binder phase and segregated to the carbidebinder interface as it was reported earlier [25]. The only striking dierence is the abrupt decrease of ITi/IW at lower binder content and lower value of PWC after the Ar+ ion beam etching. It seems that thin layers of carbidebinder interfaces (B/C) enriched in titanium segregated from the binder phase and removed by the Ar+ ion beam had a bigger relative area on the fresh fracture surface of the alloys with the lower value of PWC and the lower binder content in comparison to the higher binder contents. As it was reported earlier, higher content of Ti in the alloy and higher concentration of Ti in the binder phase is the reason for the lower value of shape equiaxiality PWC of WC crystals (see Table 1 and Ref. [25]). The decrease of PWC increases the relative amount of the trans-crystalline fracture through carbide crystals (Fig. 5) that is Ti-free because Ti does not dissolve in WC. Therefore, a lower amount of Ti is observed on the fracture surface of cemented carbides with lower shape equiaxiality PWC of WC crystals and higher content of Ti in the alloy and in the binder. However, the increase of Ti concentration in Ni binder phase might be responsible for the relatively slower decrease of ITi/IW value (Fig. 6b) in comparison to the decrease of INi/IW (Fig. 6a) on the fresh fracture surface.

4.3. Model for the crack propagation with the change of shape of WC crystals The observed change in crack propagation can be explained by the change in shape and the contiguity of WC crystals. The contiguity is higher when PWC is high (see Table 1 and Ref. [24]). Therefore, the likelihood of meeting the carbidecarbide contacts ahead of the crack is relatively high and the distance to the nearest carbide carbide contact is relatively low (Fig. 7a). The advancement of the crack through the carbidecarbide contacts results in high amount of the carbidecarbide (C/C) inter-crystalline fracture (Fig. 5a). When WC crystals change their shape to a atter triangular prism with lower value of PWC, the decreased value of the contiguity [24] decreases the amount and proximity and therefore, the likelihood of meeting the carbidecarbide contacts ahead of the crack. The atter shape also increases the relative size of the basal facets of WC crystals and consequently increases the likelihood for the crack tip to point to the basal facet of the crystal and increases both the average distance to the closest carbidecarbide contact and the average length of the path around the carbide crystal. In this case, the crack has either to increase the path through the binder phase and/or carbidebinder interface around the carbide crystal to reach the nearest carbidecarbide contact or to break through the WC crystal ahead of the

Fig. 7. Crack path in WCNi cemented carbides with dierent shape equiaxiality of WC crystals: (a) high PWC; (b) low PWC. The solid line represents the likely path of the fracture, the broken line the less likely path.

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876

75

crack (Fig. 7b). The increase of the path through the binder phase is energetically disadvantageous. Therefore, the crack breaks through the carbide crystals (Fig. 7b) and the amount of carbide trans-crystalline fracture (C) increases with the decrease of the shape equiaxiality PWC of WC crystals (Fig. 5b). In other words, the decrease of the shape equiaxiality PWC increases the eective size of WC crystals perpendicular to the crack direction that can be associated with the increase of the ratio of the carbide trans-crystalline fracture (C) to the carbidecarbide (C/C) inter-crystalline fracture in the alloys similar to what is observed on the conventional cemented carbides [57,1317]. On the other hand, the increase of the titanium-free trans-crystalline fracture (C) through the WC crystals contributes to the reduction of the relative concentration of titanium on the fracture surface (Fig. 6b). At the same time, the amount of the carbidecarbide (C/C) inter-crystalline fracture decreases on WCNi cemented carbides with the atter shape of WC crystals because of the lower value of the contiguity of carbide crystals (Fig. 5b). Apparently, the increase in the amount of the carbide trans-crystalline fracture AC supersedes the reducA C=C tion in the carbidecarbide inter-crystalline fracture AA to increase the total area of the fracture through the carC=C bide phase (AC AA ) when the shape equiaxiality PWC A of WC crystals decrease (Fig. 6a and formula (2)). 5. Conclusions It is experimentally shown that the fracture toughness K1C only correlates to the mean linear path in binder phase k on studied cemented carbides WCNi with varying shape of WC crystals. All the other parameters of the microstructure and the shape of WC crystals can only aect the fracture toughness indirectly through the change of k according to the stereological relationship (1). The usually observed monotonous correlation of K1C versus contiguity G of WC crystals breaks into two opposite dependencies when the shape of WC crystals changes. On one hand, K1C increases with the contiguity of WC crystals caused by the growth in shape equiaxiality PWC when the volume fraction of the carbide phase is xed on both non-porous WC-14 wt% Ni and WC-22 wt% Ni cemented carbides. On the other hand, K1C decreases with the volume fraction of the carbide phase and related increase in the contiguity G when the shape of WC crystals is not changed. The persistence of fracture toughness correlation versus k also indicates that the addition of TiC does not aect the properties of the binder phase, carbidebinder interfaces, and the carbidecarbide boundaries enough to change the fracture properties of the binder phase and the fracture toughness of the whole alloy. Small amount of pores and the third phase present on the WC-8 wt% Ni with excessive amount of TiC do not aect the fracture toughness directly, but obviously decrease the amount of the binder phase and the mean linear path in binder phase k in the alloy.

The decrease in the shape equiaxiality PWC of WC crystals causes a visible re-distribution of the fracture modes of WCNi cemented carbides. It is found that the amount of trans-crystalline fracture through the carbide phase (C) increases when the shape of WC crystals becomes atter triangular prism with lower value of shape equiaxiality PWC. The atter shape increases both the likelihood for the crack tip to meet the bigger basal facets and the average length of the path around the carbide crystals making it more advantageous to break through the carbide crystals than to go around it through the binder, carbidebinder, and the reduced amount of carbidecarbide boundaries. At the same time, the decrease of the shape equiaxiality PWC causes the contiguity of WC crystals to decrease. The lower value of the contiguity of carbide crystals apparently contributes to the decrease in the amount of carbide carbide inter-crystalline fracture (C/C) on the fracture surface of the alloys. Auger-electron spectroscopy results suggest that the total relative area of the fracture through the carbide phase C=C (AC AA ) increases with the decrease of PWC. It seems A that the increase of the amount of the carbide trans-crystalline fracture (C) caused by the change of WC crystals shape supersedes the reduction in the carbidecarbide inter-crystalline fracture (C/C) caused by the decrease of the contiguity of WC crystals in the alloys. The persistence of fracture toughness correlation versus k also indicates that the observed redistribution between the fracture modes of the carbide trans-crystalline fracture (C) and the carbidecarbide inter-crystalline fracture (C/C) does not aect the value of the fracture toughness. Acknowledgements The authors thank Dr. VV Sverdel and Dr. NA Yurchuk of the Frantsevich Institute for Problems of Materials Science for assistance in producing the specimens of cemented carbides, as well as Dr. OV Bakoun and Mrs. LB Dmitruck of the same Institute for help in measuring the mechanical properties. References
[1] Chermant JL, Osterstock F. Fracture toughness and fracture of WC Co composites. J Mater Sci 1976:193951. [2] Hong J, Gurland J. Direct observation of surface crack initiation and crack growth in the scanning electron microscope. Metallography 1981;14:22536. [3] Fischmeister HF. Development and present status of the science and technology of hard materials. In: Viswanadham RK, Rowclie D, Gurland J, editors. Proceedings of the rst international conference on the science of hard materials. New York: Plenum Press; 1983. p. 146. [4] Hong J, Gurland J. A study of the fracture process of WCCo alloys. In: Viswanadham RK, Rowclie D, Gurland J, editors. Proceedings of the rst International conference on the science of hard materials. New York: Plenum Press; 1983. p. 64969. [5] Sigl LS, Exner HE, Fischmeister HF. Characterization of fracture processes and fracture relevant parameters in WCCo hard metals.

76

A.V. Shatov et al. / International Journal of Refractory Metals & Hard Materials 26 (2008) 6876 In: Almond EA, Brookes CA, Warren R, editors. Proceedings of the second international conference on the science of hard materials. Bristol and Boston: Adam Hilger; 1986. p. 63144. Sigl LS, Exner HE. Experimental study of the mechanics of fracture in WCCo alloys. Metall Trans A 1987;A18:1299308. Sigl LS, Fischmeister HF. On the fracture toughness of cemented carbides. Acta Metall 1988;36:88797. Godse R, Gurland J. Applicability of the critical strain fracture criterion to WCCo hard metals. Mater Sci Eng 1988;A106:3316. Spiegler R, Fischmeister HF. Prediction of crack paths in WCCo alloys. Acta Metall Mater 1992;40:165361. Pezzotti G, Huebner H, Suenobu H, Sbaizeroc O, Nishida T. Analysis of near-tip crack bridging in WC/Co cermet. J Eur Cer Soc 1999;19:11923. Erling G, Kursawe S, Luyckx S, Sockel HG. Stable and unstable fracture surface features in WCCo. J Mater Sci Lett 2000;19:4378. Exner HE, Sigl L, Fripan M, Pompe O. Fractography of critical and subcritical cracks in hard materials. Int J Refract Met Hard Mat 2001;19:32934. Mason A, Kenny P. The structure and fractography of cemented tungsten carbide. Metallurgia 1970;82:20512. Lea C, Roebuck B. Fracture topology of WCCo hardmetals. Met Sci 1981;15:2626. Viswanadham RK, Sun TS, Drake EF, Peck JA. Quantitative fractography of WCCo cermets by Auger spectroscopy. J Mater Sci 1981;16:102938. Almond EA. Deformation characteristics and mechanical properties of hardmetals. In: Viswanadham RK, Rowclie D, Gurland J, editors. Proceedings of the rst international conference on the science of hard materials. New York: Plenum Press; 1983. p. 51761. Slesar M, Dusza J, Parilak L. Micromechanics of fracture in WCCo hardmetals. In: Almond EA, Brookes CA, Warren R, editors. Proceedings of the second international conference on the science of hard materials. Bristol and Boston: Adam Hilger; 1986. p. 65768. Lindau L. On the fracture toughness of WCCo cemented carbides. In: Taplin DMR, editor. Proceedings of the fourth international conference on fracture. New York: Pergamon Press; 1977. p. 21521. Warren R. Measurements of the fracture properties of brittle solids by Hertzian indentation. Acta Metall 1978;26:175969. Warren R, Matzke H. Indentation testing of a broad range of cemented carbides. In: Viswanadham RK, Rowclie D, Gurland J, editors. Proceedings of the rst international conference on the science of hard materials. New York: Plenum Press; 1983. p. 56382. Exner HE, Fischmeister HF. Structure of sintered tungsten carbide cobalt alloys. Arch Eisenhuttenw 1966;37:41726. Warren R, Waldron MB. Microstructural development during the liquid-phase sintering of cemented carbides. Powder Metall 1972;15: 166201. Lee HC, Gurland J. Hardness and deformation of cemented tungsten carbide. Mater Sci Eng 1978;33:12533. Shatov AV, Ponomarev SS, Firstov SA, Warren R. The contiguity of carbide crystals of dierent shapes in cemented carbides. Int J Refract Met Hard Mater 2006;24:6174. Shatov AV, Firstov SA, Shatova IV. The shape of WC crystals in cemented carbides. Mater Sci Eng 1998;A242:714.

[17]

[6] [7] [8] [9] [10]

[18]

[19] [20]

[11] [12]

[21] [22]

[13] [14] [15]

[23] [24]

[16]

[25]