1 Introduction

WC–Co cemented carbides or hardmetals are well-established forefront materials in engineering and tooling applications, such as metal cutting, mining, rock drilling, metal forming, structural components, and wear parts (Prakash 2014; Konyashin and Ries 2022). Main reason behind it is the extensive range of microstructural assemblages that they may exhibit (García et al. 2019), as a direct consequence of their composite nature. It allows to manufacture multiple combinations of interdispersed ceramic (tungsten carbide particles) and metal (Co-rich solid solution acting as binder) phases (Jiménez-Piqué et al. 2017), with variable metal/ceramic ratio and different physical dimensions. It allows to tailor hardmetal properties to satisfy critical design requirements. In this regard, hardness and toughness has always been primary selecting parameters, particularly for wear-related or rupture-limited applications respectively. However, different from the former where evaluation methodology and physical meaning are clearly defined (see, e.g., Lee and Gurland 1978; Exner 1979; Roebuck et al. 1996; Shatov et al. 2014a; Chychko et al. 2022), assessment of the latter and understanding of its implication in the fracture behavior of hardmetals have usually being more complex to follow or rationalize (see, e.g., Kenny 1971; Sigl and Fischmeister 1988; Roebuck and Almond 1988; James et al. 1990; Roebuck et al. 1996; Torres et al. 2001a, Shatov et al. 2014b; Tarragó et al. 2018; Chychko et al. 2022). Nevertheless, similar to other hard and brittle materials, it is now well-established that mechanical failure in cemented carbides is mainly associated with the existence of flaws—either preexisting as a consequence of the processing procedure or induced during service operation—and their ensuing propagation. Moreover, previous statement is true concerning not only sudden fracture but also fatigue, where it has been shown that subcritical crack extension is the dominant regime that determines failure under cyclic loading in these materials (Fry and Garret 1988; Schleinkofer et al. 1996; Llanes et al. 2002; Tarragó et al. 2015a).

Following the above ideas, it is clear that investigation and analysis of the resistance to crack propagation under cyclic loads, within a fatigue mechanics framework, is key for microstructural design optimization of cemented carbides. Within this context, a first and mandatory requirement is the introduction of artificial cracks. Although there exists a wide range of techniques for such purpose, sharp indentation of unnotched samples and cyclic compression precracking of notched ones are the approaches commonly used for hard and brittle materials (see, e.g., Lawn and 1975; Suresh 1990). The latter allows to attain through-thickness long cracks and has been validated as a successful testing method for fracture toughness and fatigue crack growth (FCG) of cemented carbides in a large number of studies (Godse et al. 1988; James et al. 1990; Torres et al. 2001a; Torres et al. 2001b; Llanes et al. 2002; Tarragó et al. 2015a; Órtiz-Membrado et al. 2022). Meanwhile, the former is linked to the introduction of short controlled fissures of similar but generally longer length than natural flaws. For hardmetals, its implementation has been rather limited to determine indentation (Palmqvist) fracture toughness of the material through direct measurement of length of the cracks induced out of the corners of the residual imprint left after a Vickers indentation (see, e.g., Palmqvist 1957; Exner 1969; Exner et al. 1978; Shetty et al. 1985a, b; Sheikh et al. 2015; Chychko et al. 2022). This fact is somehow surprising, particularly when considering that use of sharp indentation for assessment of crack extension resistance (i.e. toughness) was indeed foreshadowed by the pioneering work conducted by Palmqvist (1957) in hardmetals. Since then, it has become the reference approach for evaluating and documenting fracture toughness of these materials (Chychko et al. 2022), particularly for hard and brittle grades as the one to be considered in this investigation. On the contrary, it has been rarely used as a technique for studying other mechanical or degradation phenomena (see, e.g., Almond et al. 1976). It is then the main objective of this investigation to use small indentation cracks to evaluate FCG behavior of hardmetals, following similar approaches as those implemented for similar purpose in structural ceramics (see, e.g., Hoshide et al. 1988; Liu and Chen 1991; Zhan et al. 1998; Gilbert et al. 2000). In doing so, multiple “controlled” surface flaws are introduced by sharp indentation in each tested sample. Main goal behind it is to document and rationalize the fatigue behavior for these small cracks by combining surface observation of stably extending cracks and fractographic inspection of the one that is ultimately linked to failure. For comparison purposes, FCG study is conducted in notched specimens where long cracks have been introduced, as it would allow to validate similarity between small and long fissures regarding both FCG mechanics, once indentation residual stresses are taken into consideration, and operative cyclic degradation mechanisms.

2 Material and experimental procedure

The cemented carbide used in this investigation was an experimental WC–Co hardmetal grade with nominal composition of 9.6%wtCo (including a minor addition of Cr) and mean carbide size of 0.7  \(\mu {\text{m}}\). Microstructural assemblage of the material was studied and documented by means of field-emission electron microscopy—FESEM (Carl Zeiss Merlin FESEM, Oberkochen, Germany). In doing so, the required surface condition was produced by conventional mechanical abrasion. It consisted of initial grinding using diamond-containing disks, followed by a three-step polishing sequence (employing diamond suspensions of 6, 3 and 1 µm) up to mirror-like condition, with a final colloidal silica stage. Figure 1 shows a representative image of the microstructural assemblage of the material studied. Hardness and fracture toughness for this material are 1620 HV30 and 10.5 MPa√m, as determined by indenting a polished surface with an applied load of 30 kgf (294 N) and testing to fracture a single-edge notched and precracked beam respectively. Detailed information on microstructural and mechanical characterization for the material here studied may be found in recent work published by Órtiz-Membrado and coworkers (2022).

Fig. 1
figure 1

FESEM micrograph showing microstructural assemblage of the hardmetal grade investigated in this study. White- and black- contrast phases correspond to WC particles and metallic binder respectively

Full size image

Indentation fracture behavior of cemented carbides is extremely sensitive to surface preparation (see, e.g., Exner 1969, 1979). Thus, surface finish of samples to be indented was similar to the one aimed for microstructural assemblage inspection, i.e. a high-quality smoothly polished test specimen surface. Then, controlled small cracks were introduced in 45 × 5 × 4 mm beams, by means of indentations performed using a Vickers tip. In doing so, a 10 kgf load was applied in the surface of the polished specimens. Three non-interacting controlled flaws were produced, 1 mm apart from each other, within the center span (four-point bending) of the prospective tensile surface of each tested specimen. It was done aiming to maximize fatigue data generated by each tested sample as well as yielding unbroken dummy imprints for post-mortem inspection. Special care was taken in order to obtain indentation cracks well aligned with the specimen edges, so that one pair of indentation cracks was always perpendicular to the specimen axis. A fully articulating four-point bend test fixture, with inner and outer spans of 20 and 40 mm respectively, was used for the mechanical tests. Fatigue testing was conducted at a maximum stress level of 512 MPa, corresponding to 70% of mean value (732 MPa) of fracture strength measured for three similarly indented samples. All the tests were conducted in a universal servohydraulic testing machine, using a sinusoidal wave with a testing frequency of 5 Hz and a load ratio of 0.1. Two specimens—each containing three indentations—were tested, yielding fatigue lives up to 40,000 cycles. FCG behavior of the induced fissures was evaluated by interrupting the tests after given number of cycles and optically examining crack extension. An optical image of one radial crack system, taken just before failure took place in order to highligh the significant stable crack extension induced under cyclic loading, is given in Fig. 2. Moreover, Fig. 3 illustrates the sequential steps undertaken in this study.

Fig. 2
figure 2

Optical image of one radial crack system (including indentation imprint), after being fatigue tested during 38,000 cycles, just before final failure (Nf = 38,524 cycles) took place

Full size image
Fig. 3
figure 3

Schematic image of the experimental methodology followed in this study to investigate short crack growth under cyclic loading of cemented carbides

Full size image

Attempting to validate similitude nature of small indentation fissures and long cracks, a comparative FCG investigation involving through-thickness cracked samples was conducted. It was done using 45 × 10 × 5 mm single edge precracked notch beams (SEPNB), with a notch length/specimen width ratio of 0.3. A precrack was introduced through application of cyclic compressive loads under reverse cyclic bending, and details may be found elsewhere (Torres et al. 2014; Órtiz-Membrado et al. 2022). As for the indented specimens, lateral sides of these samples were polished to follow crack extension using optical microscopy (see, e.g. Figure 4). FCG testing was carried out in two samples, using same conditions as before, and aiming to collect data within the stable crack growth regime.

Fig. 4
figure 4

FESEM micrograph showing lateral side of a notched and precracked sample. Fissure was first nucleated under cyclic compressive stresses (reverse bending); and then further propagated under tensile (bending) fatigue, until precracking-induced residual stresses were relieved

Full size image

For both types of specimens, i.e. containing either small or long fissures, detailed inspection of crack–microstructure interaction as well as the corresponding fractographic analysis on broken specimens were carried out by means of FESEM. Regarding the former, a combination of focused ion beam (FIB) milling and FESEM (Carl Zeiss Neon 40 FIB/FESEM unit, Oberkochen, Germany) was additionally employed, in order to document mechanisms operative during stable cyclic crack extension in indented specimens (Tarragó et al. 2015b). In doing so, a thin protective platinum layer was first deposited on small areas of interest, corresponding to regions close to tips of arrested cracks (from surviving dummy indentations), along surface crack paths. U-shaped trenches, perpendicular to the crack path and to the specimen surface, were produced by FIB. Subsequently, series of crack–microstructure interaction images were obtained by periodic removal of the material by selectively bombarding the material with Ga + ions, within the U-shaped crater parallel to the cross-sectional surface. For that purpose, FIB was selected to operate at a voltage of 30 kV and at a current intensity of 10 pA; while FESEM gun operated at 2 kV.

3 Results and discussion

3.1 FCG mechanics of small indentation and long through-thickness cracks

It is well-established that hard and brittle cemented carbides develop radial (Palmqvist) cracks when being sharp indented (see, e.g., Shetty 1985a). However, to the best knowledge of authors, it is not documented whether such geometry still applies or rather experiences a transition towards a fully developed radial/median (half-penny) system, after stable crack growth under cyclic loading. This information is required for quantification of indentation-induced residual stresses. Hence, aiming to document it, a serial sectioning procedure (see, e.g., Shetty et al. 1985b; Cook et al. 1994) was first implemented to assess profiles of cracks linked to dummy indentations that survived fatigue testing. Experimental findings are reported in Fig. 5. They show images taken from optical microscopy inspection after sequential polishing involving multiple steps. Information about the depth of the material removed, calculated from the relative change in the area of the remnant sharp imprint, is also included in such figure. It is evident that, before catastrophic failure took place, extended cracks were not connected to each other beneath the residual imprint; and thus, cracking system was still of Palmqvist type for surviving indentations. Accordingly, fatigue testing data will be reported as mean behavior exhibited by two individual cracks, with semi-elliptical configuration (length 2c and depth a), linked to each indentation imprint under consideration.

Fig. 5
figure 5

Sequential polishing of one surviving indentation (10 kgf) and corresponding crack system, including fissures that experienced stable growth under cyclic loading. Thickness of layer of material removed during sequential tomography is indicated in each image

Full size image

Graphs of crack length (c), directly measured at the surface, as a function of number of cycles (N) for fissures associated with each of the three indentations performed in the two tested specimens are shown in Fig. 6. In all the cases, same trends are followed and scatter of the experimental findings, considering the stochastic nature of fatigue, is rather low, similar to those previously documented for other brittle materials (Hoshide et al. 1988; Liu and Chen 1991; Zhan et al. 1998; Gilbert et al. 2000). Meanwhile, crack growth rates (da/dN) versus the maximum applied stress intensity factor (Kmax) are plotted in Fig. 7. In doing so, Kmax was calculated as the externally applied stress intensity factor (Kapp) evaluated from linear-elastic solutions (Newman and Raju 1979) for three-dimensional semielliptical surface cracks in bending in terms of crack depth, a; crack length, c; elliptical parameter angle, ϕ; shape factor, Q; specimen width, W; specimen thickness, B; and remote (outer surface) bending stress, σb:

$${K}_{app}={{{H}_{c}\sigma }_{b}\sqrt{\left(\frac{\Pi \pi c}{Q}\right)}} F(a/c,a/W,a/B,\phi )$$
(1)

where Hc is the bending multiplier and F is a boundary correction factor. The aspect ratio, c/a, was first estimated for the empirical equation recommended by ASTM E740 (1983) to correlate the shape of a semi-elliptical surface crack growing under bending, and later validated by direct measurement of crack geometry in FESEM micrographs of fractured surfaces (e.g. see, Fig. 11a below). Furthermore, it should be highlighted that FCG data is plotted as a function of Kmax, instead of against the more conventional stress intensity factor range (ΔK). Reason for it is the fact that FCG rates for cemented carbides are well-established to exhibit an extremely high dependence on the former fracture mechanics parameter (Fry and Garret 1988; Torres et al. 2001b; Torres et al. 2014; Tarragó et al. 2015a), this being particularly true for grades with relatively low binder content and fine carbide size (Llanes et al. 2002). Such a behavior is associated with relative predominance of static over cyclic modes of fracture during stable FCG, as a consequence of relatively low effective ductility of the constrained metallic binder. This is in agreement with FCG behavior reported for other brittle-like materials, such as structural ceramics or intermetallics, where extrinsic toughening mechanisms (i.e. developing behind the crack tip) are operative (Ritchie 1999).

Fig. 6
figure 6

Experimental data of surface crack length (c) as a function of number of cycles (N) for small radial fissures, emanating out of the three individual sharp indentation imprints introduced in the two tested samples. Solid and dashed lines indicate indentations corresponding to each specimen investigated, whereas indentatations are randomly numbered. Testing conditions: fatigue under load ratio of 0.1 and maximum applied remote (outer surface) bending stress of 512 MPa

Full size image
Fig. 7
figure 7

FCG rates as a function of applied Kapp.max, under load ratio of 0.1, for both small indentation and long through-thickness cracks

Full size image

In general, experimental data for short indentation-induced fissures show a clear gathering in two well-defined ranges of crack growth rates and applied stress intensity factor. However, analysis of individual data points allows to describe their kinetics behavior as rather undefined, as they are found to describe zig-zag trends where decreasing, constant and increasing rates as crack extends are mixed. This somehow anomalous-like behavior—from a fatigue mechanics analysis viewpoint—is indeed in total agreement with findings reported for short cracks propagating under cyclic loading in metallic, ceramic and composite materials. In this regard, differences in crack-tip shielding between long and small cracks is the basic reason usually invoked for rationalizing them (Suresh and Ritchie 1984; Ritchie and Dauskardt 1991), although specific crack-microstructure interactions behind it are quite diverse. Hence, depending on the material under consideration, operative toughening mechanisms, small-scale plasticity and/or residual stresses, among others, must be recalled for evaluation of the effective stress intensity at the crack tip, and this issue will be analyzed later.

FCG experimental data for long through-thickness cracks are also included in Fig. 7. Here, Kmax was determined using the expression given by Munz and Fett (1999) for the applied stress intensity factor (Kapp) in a notched and precracked specimen under pure bending:

$${K}_{app}={{\sigma }_{b}}Y\sqrt{a}$$
(2)
$$Y= \frac{1.1215\sqrt{\pi }}{{\beta }^{3/2}}\times \left[\frac{5}{8} - \frac{5}{12}\alpha + \frac{1}{8}{\alpha }^{2}{\beta }^{6}+ \frac{3}{8}{\text{exp}}\left(-\frac{6.1342\alpha }{\beta }\right)\right]$$
(3)

where σb is the applied remote (outer surface) bending stress, a is the length of the through-thickness crack, Y is a geometry factor that depends on the configuration of the sample with the notch and the manner in which the loads are applied, α = a/W and β = (1-α). This fatigue testing was conducted as a comparative study for evaluating similitude, regarding fatigue mechanics and mechanisms between small and long fissures. From the viewpoint of the former, although subcritical growth is observed at Kmax values much lower than fracture toughness for both crack types, it requires significantly higher applied Kmax values for long fissures than for small ones. Furthermore, different from the undefined and rather erratic FCG behavior of small flaws, long ones exhibit a well defined large-power dependence of FCG rates on Kmax. Here, it should be underlined that, based on previous experience, FCG data for long cracks were always recorded from fissures which had been first propagated (under tensile fatigue) until an accelerated growth behavior with further extension was discerned. It permitted to ensure that they were free of any residual stresses previously introduced during cyclic compression precracking (Godse et al. 1988; James et al. 1990; Torres et al. 2001a; Órtiz-Membrado et al. 2022). This is clearly not the case for small indentation flaws, as will be discussed below.

The application of a load by means of an indenter produces a displacement of matter at the contact of the specimen surface with the indenter and leaves residual stresses when the load is removed. These are responsible for the extension of cracks generated by sharp tips during unloading. On the basis of this response, monitoring cracks created by hardness indentation has become a quite popular, simple and convenient method for assessment of fracture toughness in brittle-like materials—from single crystal oxides to cemented carbides—since it was first advocated by Evans and Charles in the mid-70 s (Evans and Charles 1976). Thus, when submitted to an external stress, either mechanical or thermal, the radial indentation cracks propagates first in a stable way under the combined action of the residual contact stresses and the applied stress. Hence, to obtain the total stress intensity factor (Ktot) acting on an indented specimen, the residual stress intensity factor (Kres) associated with the indentation should be taken into account in addition to the nominal applied stress intensity factor (Kapp):

$${K}_{tot}={{K}_{app}}+ {K}_{res}$$
(4)

where Kapp is determined according to Eq. (1), whereas Kres is taken from the indentation fracture mechanics analysis developed by Shetty and co-workers (1985), on the basis of models proposed for semicircular Palmqvist surface cracks by Niihara and co-workers (Niihara et al. 1982; Niihara 1983). It is based on the assumption of a wedge-loaded fissure for a radial crack system in WC–Co hardmetals, and is given by:

$${K}_{res}=\beta \cdot \sqrt{HV\cdot \frac{P}{8c}}$$
(5)

where HV is the hardness (N/mm2), P is the applied load (N), 2c is the average of the length of the semicircular Palmqvist cracks at the imprint corners (mm), and β is a nondimensional constant, dependent on Young moduls and hardness of the material under consideration. It is calibrated by comparing the value of fracture toughness measured using this equation with those obtained by other well-established and reliable fracture mechanics methods. In doing so, Shetty and coworkers found 0.0889 as a best-fit β value, as far as fracture toughness of WC–Co grades was lower than about 15 MPa√m. In a more recent study addressing testing method and microstructural effects (Sheikh et al. 2015) such β value was validated for satisfactory correlation among fracture toughness values (up to 14 MPa√m) measured using three-point bending on Chevron notched specimens (“reference” baseline), Herzian indentation method and Palmqvist indentation test. Hence, the use of a value of 0.0899 for β in Eq. (5) is appropriated as hardmetal grade here studied exhibits a fracture toughness of 10.2 MPa√m.

Crack growth curves replotted as da/dN versus Ktot,max are shown in Fig. 8. It should be noticed that exclusive difference between Figs. 7 and 8 refers to the effective stress intensity factor acting as driving force for crack extension under cyclic loading for small indentation flaws. It is clear that accounting of the residual stresses induced by indentation results in overlapping of the FCG data for both crack types. This is in agreement with the results reported for other advanced ceramics, such as alumina, silicon nitride, silicon carbide and zirconia ceramics (see, e.g., Hoshide et al. 1988; Liu and Chen 1991; Zhan et al. 1998; Ritchie 1999; Gilbert et al. 2000), and validates the use of these small controlled indentation flaws for studying FCG behavior of hardmetals, as far as the residual field left after indentation unloading linked to a radial crack system, is considered in the fatigue mechanics analysis.

Fig. 8
figure 8

FCG rates as a function of Ktot,max (i.e. accounting for Kres in the case of small indentation flaws), under load ratio of 0.1, for both crack types studied

Full size image

These findings and their corresponding analysis allow to discuss on the possibility of choosing the use of either indentation-induced small fissures or long through-thickness cracks for evaluation of fracture toughness and FCG behavior of cemented carbides, depending on the specific microstructural assemblage of the grade under consideration. Hence, regarding cracks emanating from the corners of residual imprints, it is well-established that they are extremely small, or are even absent, when hardmetal grades with high binder content and/or coarse carbides are sharp indented. Under these conditions, one of the assumptions made by Shetty and co-workers in its wedge-crack model—total accommodation of the hardness impression volume by the radial expansion of the plastic zone—is not satisfied (Shetty et al. 1985a, b). As a consequence, Kres equation would either yield overestimated values or would not even be possible to be assessed (in the case that cracks are not nucleated); and thus, use of indentation technique becomes unfeasible for grades whose hardness value is below 1300 HV30. Meanwhile, concerning long cracks, precracking procedures based on cyclic compression of notched specimens are found to be quite difficult for hardmetal grades with low binder content and/or fine carbides, i.e. exhibiting higher levels of hardness and brittleness (e.g. Torres et al. 2001; Tarragó et al. 2015c). Furthermore, slopes implicit within crack growth kinetics under cyclic tensile loading are quite steep; and thus, gathering of FCG experimental data requires extremely meticuoulus, careful and time-consuming efforts (Llanes et al. 2002). In these cases, considering that Palmqvis indentation is a well-established practice for reliable determination of toughness on hardmetal grades whose hardness value is above 1300 HV, possible extension of its use under cyclic loading for documenting FCG behavior in those materials is a quite positive outcome of this study.

3.2 FCG mechanisms of small indentation and long through-thickness cracks

Similitude found between FCG behavior of small and long cracks was further investigated on the basis of the fatigue mechanisms involved. Hence, an extensive and detailed study of crack-microstructure interaction and fractographic features was conducted. Regarding the former, it included direct FESEM observation of stably grown cracks both at the surface as well as through FIB serial sectioning in the case of indented samples. Figure 9 shows representative images from the referred top-surface perspective. It is clear that, independent of crack type, fissures predominately run along the metallic binder phase through step-like paths that resemble small-scale striation features. They result from the similar submicrometric scales of cyclic plastic region ahead of the crack tip and effective thickness of the binder constrained between the hard carbides. Under these conditions, microscopic failure modes characterized by localized shear and zig-zag crack paths are expected (Tarragó et al. 2015c). This fatigue micromechanism is completely different from the one expected for invoking the reinforcement role played by the metal ductile ligaments for rationalizing the high toughness exhibited by cemented carbides under monotonic loading. Under these conditions, binder bridges at the crack fail by nucleation, growth and coalescence of microcavities, usually with starting locations close to either carbide corners or carbide–binder interfaces, where high triaxiality stress and strain conditions are fulfilled (Tarragó et al. 2015b). Such distinct scenarios are evidenced in Fig. 10, where FIB-milled cross-section images for indentations cracks stably grown under different loading conditions are compared. Here, stably extended cracks under monotonic loading have been attained from dummy indentations performed in specimens preliminary tested for assessment of fracture strength used as reference to define maximum applied stress in subsequent fatigue testings (see Sect. 2).

Fig. 9
figure 9

FESEM micrographs showing typical crack-microstructure interaction during stable FCG: a small indentation flaw; and b long through-thickness crack.

Full size image
Fig. 10
figure 10

FESEM micrographs of FIB milled cross sections (low magnification U-shaped trenches and high magnification detailed scenario) showing stable crack growth of small indentation fissures under monotonic (a and c) and cyclic (b and d) loading

Full size image

Concerning fracture surfaces for specimens containing small indentation flaws and long through-thickness cracks, they are found to be different at the mesoscale, as a direct consequence of crack geometry and residual stresses; but quite alike in terms of crack growth regimes and affiliated micromechanisms. As it is shown in Figs. 11a and b, independent of crack type, low-magnification fractographic inspection allows tagging four regions. The initial one (0) refers to the original fissures introduced by either indentation (including here the residual imprint) or cyclic compression precracking (including here the macroscopic notch partly observed in the image). The second one is the first stage driven by cyclic loading and involves stable crack growth. It is then followed by a transition region into the final fast fracture stage associated with unstable crack extension. Similitude between I, II and III regions in Figs. 11a and b is sustained by the specific fatigue and fracture micromechanisms discerned at high magnification for both crack types (Fig. 12a–f). Sequential visualization of these images, permits to understand the transition from stable crack extension in region I, characterized by surface markings of crystallographic nature within the binder (Torres et al. 2014; Tarragó et al. 2015a, 2015b; Llanes 2019; Órtiz-Membrado et al. 2022), into unstable one in region III, linked to sharply resolved dimples in the metallic phase, as a result of nucleation, growth and coalescence of microvoids. As expected, both failure modes co-exist in region II; thus, here referred as the transition stage, exhibiting it a fractographic scenario that would be expected to emerge in case fatigue testing would be conducted under higher load ratios for the material under consideration (Torres et al. 2014).

Fig. 11
figure 11

Fatigue fracture surfaces of a an indented sample; and b a notched and precracked specimen. Different regions are tagged, linked to distinct crack growth phenomena

Full size image
Fig. 12
figure 12

FESEM micrographs showing fractographic features (and involved micromechanisms) associated with stable (I zones), transition (II zones) and unstable (III zones) extension of small indentation flaws (a, c and e) and long cracks (b, d and f) under cyclic loading. Squares with solid and dashed lines indicate micromechanisms associated with stable and unstable crack propagation, respectively

Full size image

4 Conclusions

The fatigue crack growth behavior of a submicron-grained WC–Co cemented carbide has been investigated by using two different types of artificial cracks. It has been conducted by combining extensive mechanical testing, optical and scanning electron microscopy (including focused ion beam milling) characterization, and fracture mechanics analysis. From the results reported and their corresponding discussion, the following conclusios may be drawn:

  1. (1)

    The use of controlled small indentations flaws shows to be a valid and successful approach for studying and describing the fatigue behavior of WC–Co cemented carbides. This is sustained by the similitude found between fatigue mechanics exhibited by these artificial defects when compared to the ones determined and found when using long through-thickness cracks. In this regard, consideration of the residual stresses induced during indentation unloading are key to rationalize experimental findings gathered in tests using the different crack types investigated.

  2. (2)

    Similitude between small indentation flaws and long cracks is sustained by the alike crack-microstructure interaction and fractographic scenarios discerned as a result of cyclic loading. For both artificially introduced fissures, a transition from stable into unstable crack extension is clearly evidenced at both meso- and micrometric length scales, linked to fatigue-driven suppression of the commonly observed toughening mechanism for hardmetals, i.e. reinforcement due to ductile ligament bridging at the crack wake. As a result, stable crack extension under cyclic loading is evidenced at maximum stress intensity factors significantly lower than the fracture toughness measured under monotonic loading.

  3. (3)

    Considering that indentation fracture toughness is commonly invoked as key design parameter for selection and quality control of cemented carbides, previous conclusion has an immediate practical implication, as similar approach could be extended to investigate the FCG behavior of hardmetals. This is particularly important in the case of hard and brittle grades (hardness above 1300 HV30 and toughness below 14 MPa√m), where precracking under cyclic compression and FCG data gathering in notched specimens are quite challenging tasks. Meanwhile, it should be advised that implementation of such testing methodology to softer (hardness below 1300 HV) and tougher (toughness above 14 MPa√m) cemented carbides would not be successful. In these cases, the capability of sharp indentation for inducing well-developed radial crack systems is quite limited and, as a consequence, stress intensity factor linked to residual stresses would either yield overestimated values (extremelly small cracks) or would not even be possible to be assessed (in the absence of any crack emanating from the corners of residual imprint).