Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=yaac20

Advances in Applied CeramicsStructural, Functional and Bioceramics

ISSN: 1743-6753 (Print) 1743-6761 (Online) Journal homepage: https://www.tandfonline.com/loi/yaac20

The influence of microstructure on the mechanicalproperties of polycrystalline diamond: a literaturereview

Thomas A. Scott

To cite this article: Thomas A. Scott (2018) The influence of microstructure on the mechanicalproperties of polycrystalline diamond: a literature review, Advances in Applied Ceramics, 117:3,161-176, DOI: 10.1080/17436753.2017.1389462

To link to this article: https://doi.org/10.1080/17436753.2017.1389462

Published online: 02 Nov 2017.

Submit your article to this journal

Article views: 875

View Crossmark data

Citing articles: 1 View citing articles

The influence of microstructure on the mechanical properties of polycrystallinediamond: a literature review*Thomas A. Scotta,b

aDepartment of Materials, University of Oxford, Parks Road, Oxford, UK; bCentre for Doctoral Training in Diamond Science and Technology,Warwick, UK

ABSTRACTPolycrystalline diamond (PCD) is an extremely high-performance cutting tool material used inthe machining of rock, high-strength, non-ferrous metal alloys and carbon-fibre-reinforcedcomposites. It is favoured for its exceptional hardness and wear resistance which results in atleast an order of magnitude improvement in performance over previous technologies inalmost all metrics. However, PCD suffers from unpredictable brittle fracture and degradationat high temperature during service which limits its capabilities in cutting applications. Theliterature on the link between its microstructure and its mechanical properties, includingstrength, toughness and flaw size distribution as measured by pseudo-static tests, isinvestigated. The conclusions of the seminal paper on this topic are re-examined in the lightof modern ceramics research and an alternative explanation is put forth for the strength–grain size relationship published in this paper. All known literature values for strength andtoughness vs. grain size and binder content are collated showing no overall trend in strengthwith binder content but moderate trends in all other combinations. The common claim ofweak grain boundaries is brought into question in the light of the lack of any evidence ofthis fracture mode being evident in pseudo-static tests. The industrial literature on weartesting and failure modes of PCD in service and service-like tests is examined to bridge thegap between pseudo-static and dynamic, application-based experiments. Six main failuremodes are recorded and summarised with intergranular fracture being the mostconspicuously absent from the pseudo-static tests. It is suggested that the temperaturegenerated by friction in dynamic tests causes the weakening of grain boundaries, resulting ina transition from transgranular to intergranular fracture and a call for further research in thisarea is made.

ARTICLE HISTORYReceived 22 May 2017Accepted 4 October 2017

KEYWORDSPolycrystalline diamond; PCD;microstructure; mechanicalproperties; fracturetoughness; strength; flaw sizedistribution; thermaldegradation

Introduction

Brief history of diamond synthesis

Almost as soon as diamond was discovered to be anallotrope of carbon in 1797 by Smithson Tennant [1],efforts were made to produce it synthetically. Manyclaimed success but a distinct lack of evidence andreproducibility suggests that these early reports werefallacious. Simultaneously the most famous and tragicof which was that of Henri Moissan. He attempted touse the solubility of carbon in iron combined with arapid quench from white heat to induce the pressurerequired to transform graphite into diamond [2].Although he claimed success, testimony from hiswidow stated that one of his assistants had introducednatural diamond flakes into his samples ‘to please theold man’ and cut short his tediously long experiments[3]. Later cathodoluminescence work confirmed thatthe diamonds were indeed natural in origin [4].

It was not until 1955 [5] (with further details givenin 1959 [6]) that the first credible report of diamondbeing synthesised was published by researchers at

General Electric. They discovered that a molten metalcould be used to lower the temperatures and pressuresneeded to synthesise diamond making it feasible giventhe technology of the time but did not understand themechanism of growth. The role of the molten metalwas strongly contested until a study in 1981 concludedthat it must play a role as both a solvent for graphiticcarbon and a catalyst for the re-precipitation of dia-mondiferous carbon [7], lowering the activation energybarrier of its formation.

Polycrystalline diamond

HistoryThe first instance of a sintered, polycrystalline dia-mond (PCD) compact was published in 1970, 12years after it was predicted that a synthetic form ofnatural PCD, carbonado, would be a highly desirablematerial to be able to make [8] by the same author.The following year, it was found that using cobalt asa binder could reduce the temperature and pressurerequired for successful sintering by up to 24 and

© 2017 Institute of Materials, Minerals and Mining. Published by Taylor & Francis on behalf of the Institute.

CONTACT Thomas A. Scott thomas.scott@materials.ox.ac.uk Department of Materials, Parks Road, Oxford OX1 3PH, UK*This paper was the winning entry of the 2017 Davidge Award.

ADVANCES IN APPLIED CERAMICS, 2018VOL. 117, NO. 3, 161–176https://doi.org/10.1080/17436753.2017.1389462

27%, respectively [9]. This technological innovationallowed the mass production of PCD to be realisedand still accounts for the way in which most PCD ismade today.

In 1973, General Electric patented the use of a car-bide-supported PCD drill bit for use in oil and gas dril-ling [10]. From then onwards, PCD cutters madeinroads into various material removal industries suchas the machining of wood-based products [11], non-ferrous metals [12] and most recently, composites[13]. In 2010, PCD cutters accounted for 65% of thefootage drilled by the oil and gas industry [14], pro-jected to rise to over 80% by 2016 [15]. Such marketdomination has been due in large part to the cost sav-ings made possible by PCD’s remarkable properties.PCD cutters provide up to two orders of magnitudeimprovement in tool life over other industry stan-dard-cutting materials such as cemented tungsten car-bide depending on the specific application [16–18].This allows far less time and money to be expendedon replacing worn tools; something that is particularlycostly in drilling applications, leading to vastly reducedoperating costs and hence increased profits. Outstand-ing hardness and abrasion resistance are the two mainmaterials properties of PCD that have allowed suchhuge performance increases compared to othermaterials. But, as ever, with high hardness comes lowtoughness. Although PCD is far less brittle than singlecrystal diamond, its resistance to impact and shockloading are two of the main issues that must be over-come in improving PCD as a cutting material.

Synthesis of PCDPCD synthesis begins with diamond powder; this canconsist of a single carefully controlled grain size, awide distribution of grain sizes or multiple singlesizes mixed together. The powder is compacted onthe top of a cemented tungsten carbide substratewhich may have a textured top surface to mitigatethe residual stresses that develop at the diamond–sub-strate boundary. This is encapsulated by a metal canis-ter and inserted into a capsule that in turn goes into ahigh pressure press. There are two common designs ofpress: belt and cubic. Cubic presses apply pressurealong each of the three Cartesian axes, whereas beltpresses apply pressure from above and below and con-straint circumferentially. Presses typically apply 1400–1800◦C and 4–10 GPa of pressure. The capsule isdesigned so that a constituent of it becomes moltenat high temperature, distributing the applied pressurehydrostatically.

The four major processes that occur during the sin-tering of PCD are outlined in Figure 1. The resultantproduct is a layer of two-phase ceramic predominantlycomprised of micron-sized diamond grains sinteredtogether with a high degree of diamond–diamondbonding atop a thick tungsten carbide/cobalt substrate.

The diamond grains retain roughly the same mor-phology post-sintering and the cobalt binder occupiesthe spaces between the grains. The micro-structure isa complex combination of the original diamond grains,newly precipitated diamond [19], metal binder andvarious intermetallic phases [22]. In addition, thereare regions of plastically deformed material [20] andquite substantial microscopic and macroscopic residualstresses [23].

Leaching of PCDAlthough cobalt is very useful in acting as a catalyst fordiamond synthesis, it unfortunately also contributes tothe thermal degradation of PCD under the high-temp-erature conditions of drilling [24]. A patent was filed in1978 describing a method of removal of the cobalt bin-der phase to increase PCD’s thermal stability, calledleaching [25]. Leaching involves heating PCD in a mix-ture of concentrated acid for long periods of time todissolve the cobalt. While this does increase the ther-mal stability, the mechanical properties are degraded[26] and it is a time-consuming and dangerous process.Electrolysis has also been used to leach PCD [27] withgood results. Few studies have been done on the effectleaching has on the mechanical properties of diamondand these will be discussed in later sections.

Traditional mechanical testing of PCD

Techniques, such as notched and unnotched beam flex-ure, ball-on-three ball flexure and diametral com-pression, are discussed in this section. The term‘traditional’ is taken to mean mechanical deformationof macroscopic samples. Almost all of the work on dia-mond’s mechanical properties cite the work of Lammer[28] from 1988. A recent review paper entitled ‘Themechanical and strength properties of diamond’ byField quotes it as if it is the state of the art [29].Although Lammer’s work made significant improve-ments over the early work carried out by Gigl [30]and Roberts [31], there are several points to beaddressed regarding the conclusions of the paper.

The seminal paper on the mechanical propertiesof PCDThe first work to examine the mechanical properties ofPCD in great depth was written by Lammer in 1988.This paper presented a big step forward in the under-standing of this research area and still influenceswork undertaken today. However, some of its con-clusions warrant a re-examination in the light of mod-ern understanding of ceramic materials and given itsprominence in the field, the entirety of the present sec-tion has been dedicated to doing just that.

Figure 2 illustrates the relationship between trans-verse rupture stress and the inverse of the square rootof nominal grain size deduced by Lammer for various

162 T. A. SCOTT

PCD grades. It shows two distinct regions in behaviour.The bimodal nature of the curve is attributed to anexplanation by Rice [32] based on work on other cer-amics showing that the inherent critical flaw size c isindependent of the grain size d. The argument con-tinues that for fine-grained materials, where theinherent flaws may span numerous grains, the strengthis controlled by the polycrystalline fracture toughness.For coarse grains, where the critical flaw is within asingle grain, it is controlled by something akin to asingle-crystal fracture toughness. The crossover pointbetween these two regimes is expected to be wheregrain size, d, is equal to flaw size, c. From Table 1,c/d = 1 occurs between 12–30 μm which agrees withthe approximate crossover point between ‘region 1’and ‘region 2’ in Figure 2 of 30 μm.

However, the grain size used in this analysis is theaverage diameter of particles in the original powderprior to sintering. During the high-pressure sinteringprocess, particles are crushed due to large stresses atpoint contacts [21], widening the particle size

Figure 1. Schematic representation of the four major stages in the sintering process of PCD [19–21].

Figure 2. Plot of transverse rupture strength against theinverse of the square root of grain size from Lammer [28].Two distinct regions are shown in the curve transitioning ata value that represents a grain size of approximately 30 μm.

ADVANCES IN APPLIED CERAMICS 163

distribution andpushing it to lower diameters. This pro-cess is much more severe for large particles as the num-ber of contact points is fewer causing higher localstresses. As particle size decreases, the comminutionlimit is eventually reachedmeaning little or no crushingoccurs. Therefore, the assertion that the crossover pointoccurs where c=d is likely untrue as ‘nominal grain size’largely, but not linearly, overestimates the true grainsize.

Additionally, two distinct fracture behaviourswould be expected to be seen. The ‘single crystal frac-ture toughness’ for large grain sizes and the ‘polycrys-talline fracture toughness’ for small-grain sizes. It isalso implied that large-grained materials should exhi-bit transgranular fracture while small-grainedmaterials show intergranular fracture. As can beseen from Figure 3, no such clear distinction is evidentand although fracture in large-grained (≥30 μm)material is described as being transgranular, there isno mention of the fracture mode of fine-grainedmaterials and no corroborating SEM images. Theexplanation given for the observed trend is derivedfrom Rice’s work stating that ‘thermal and elastic ani-sotropy between diamond and binder phases couldresult in high interfacial stresses that allow micro-cracks to propagate along the diamond-cobalt bound-aries’. While high-residual stresses have beenobserved due to the thermal expansion mismatch of

diamond and cobalt [33], fracture occurs predomi-nantly transgranularly in PCD [34] so micro-cracksalong the grain boundaries, if they do exist, have mini-mal effect on the fracture behaviour. He then goes onto say that other factors such as ‘impurities in thematrix material, the amount and size distribution ofthe secondary phase, and the degree of plastic defor-mation and fragmentation of diamond grains duringsynthesis’ could affect the fracture toughness. Someof these postulations are almost certainly true, butfurther work is required to show which hold merit.

As it has been established that the flaw size is notindependent of the grain size, Lammer’s data can bereinterpreted. It is commonly found in ceramics thatfor large-grained materials, the individual grains actas critical flaws and so the fracture behaviour followsIrwin’s modification of the Griffith equation

sc = KIc

Y���pc

√ (1)

‘Region 1’ in Figure 2 is a straight line that passesapproximately through the origin on a graph ofstrength sc against inverse, root grain size d−1/2,suggesting that this typical behaviour is indeed fol-lowed. Also given the roughly constant value of c/dfor grain sizes above 30 μm in Table 1 it can safely beassumed that the flaw size is proportional to thegrain size in this region.

By comparing Equation (1) to the gradient of ‘region1’, the implied fracture toughness of the material can becalculated from Equation (2) by substituting Y = 2/pand c = d/2, assuming that the flaws are semi-ellipticaland surface breaking

K impIc = 1.12 ·m

���2p

√(2)

where m is the gradient of ‘region 1’.The result of this calculation gives K imp

Ic = 6.67MPa

���m

√. Figure 4 shows this result along with the

measured fracture toughness of the materials in ‘region1’ (materials C-G, values taken from Figure 3). Lam-mer’s measurements follow the work of Devin et al.[35], using laser cut notches in the centre of diametralcompression samples. If the notch created by this tech-nique has a much greater radius that the inherent flawsize of the material, it is considered blunt and will over-estimate the toughness [36]. A treatment by Fett andMunz [37] calculates the true fracture toughness

Table 1. Critical flaw size calculated from strength and toughness for each grain size material also showing the ratio between thetwo.

A B C D E F G

ccrit,mm 6 16 18 91 19 68 105d,mm 2 12 30 125 30 95 150c/d 3.00 1.33 0.60 0.73 0.63 0.72 0.70

From Lammer [28].

Figure 3. Plot of fracture toughness against nominal grain sizemeasured by diametral compression of specimens with lasercut notches in them from Lammer [28]. Given that the drawntrend appears to be on the same order as the scatter in thedata, its validity is drawn into question.

164 T. A. SCOTT

from the apparent value by assuming a small, incipient,sharp crack protrudes from the blunt notch using thefollowing equation:

Ktrue = Kapp tanh(2Y������da/r

√) (3)

where Y is the geometrical shape factor, da is the inci-pient crack length and ρ is the notch-root radius. Giventhat Equation (3) reduces the apparent fracture tough-ness by a greater amount for smaller incipient cracksizes (and hence smaller grain sizes), the calculatedvalue of K imp

Ic in Figure 4 and the measured toughnessvalues are in reasonable agreement and a single-valuedKIc in ‘region 1’ is plausible. Unfortunately, the notch-root radius was not given and so the actual adjustmentcannot be performed.

If ‘region 2’ merely had a different, single-valuedpolycrystalline fracture toughness, it would still beexpected that the data would form a straight linethrough the origin, only with a different gradient.This is evidently not the case. Instead, a much weakerdependence on grain size is seen, suggesting thatanother factor is controlling the strength.

StrengthData from other authors measuring the strength ofPCD are tabulated in Table 2. Belnap and Griffo [41]compared the strength of PCD monoliths and PCDcomposites with additional tungsten carbide/cobaltadmixed. They found that decreasing the diamondcontent of the composite resulted in a decrease in theflexural strength. The two different grain sizes theyused resulted in similar flexural strengths within thebounds of the error. Along with showing that there isa negative strain rate and positive temperature depen-dence on strength, Petrovic et al. [38] found that a lar-ger grain size resulted in a lower strength material. Thiswas ascribed to larger flaws found in the coarser-grained material; they noted that the typical flaw sizefound in the surface of the specimen was proportionalto the grain size. Despite reporting the strength of onegrade of PCD, Morrell et al. [40] do not provide anyinformation on which grade this was. Finally, McNa-mara et al. [39] found that a small decrease in thecobalt content of the PCD resulted in a small decreasein the strength.

The data in Table 2 are also shown graphically forgrain size and binder content in Figures 5 and 6.From Figure 5, a reasonably clear trend can be seenthat decreasing the grain size increases the strength.This agrees with the general trend shown by Lammerbut two distinct regions are not obvious in these data.The correlation between flaw size and grain sizereported by Petrovic et al. disagrees with Lammer’ssuggestion that flaw size is independent of grain size;however, further results are needed to verify thetrend and an investigation into the flaw size distri-bution for various grades of PCD should be carried out.

In Figure 6, a less obvious trend is seen betweenstrength and binder content, although there may be ageneral positive correlation, the large scatter in thedata adds significant uncertainty. As it is difficult tovary the binder content independently of grain size,understanding the role of the binder phase on themechanical properties from literature data isproblematic.

Figure 4. Re-plot of toughness values for materials C-G fromLammer [28]. The green line shows the ‘implied’ toughness cal-culated for these materials from Equation (2).

Table 2. Strength data for various grades of PCD collated from numerous authors.Author Technique Strength (MPa) Weibull modulus Grain size (μm) Binder content (wt-%)

Petrovic et al. [38] 3PB 750+ 75 – 6 23650+ 75 – 30 7

McNamara et al. [39] - 1075+ 82 13.4 – 8.51034+ 129 8.3 – 7.5

Morrell et al. [40] B3B 1453+ 29 14.5 – –Belnap and Griffo [41] 3PB 1815+ 100 16.0 3.5 23.5

1714+ 95 21.9 5 20Lammer [28] DC 1550+ 185 3.4 2 13

1256+ 175 4.3 12 111180+ 175 4.4 30 11444+ 150 3.4 125 12

Marro et al. [42] B3B 1610+ 230 – 4 101120+ 160 – 10 7.51000+ 200 – 20 6

The following terminology is used for the technique column: 3PB – 3 point bending, B3B – ball-on-three-ball, DC – diametral compression. The error shown inthe strength column is the standard deviation except for Lammer where it is the range in the measured results.

ADVANCES IN APPLIED CERAMICS 165

Fracture toughnessMiess and Rai [43] investigated the fracture toughnessof a wider range of grain sizes using the same methodas Lammer. Their results show a plateau in fracturetoughness rather than the drop-off found by Lammer.Along with citing Rice’s work, they suggest that asgrain boundaries in ceramics are typically weakerthan the grains, increasing the area of grain boundaryby decreasing the average grain size increases theamount of weak material and hence lowers the fracturetoughness. This argument suffers from the same flaw asthat of Lammer – no fracture paths are shown in theirwork and later work has shown that fracture in PCDoccurs predominantly transgranularly.

Double-torsion tests were carried out on variousgrades of PCD by Lin et al. [44] to measure the fracturetoughness and investigate sub-critical crack growth.They measured unusually high values of fracturetoughness, in the region of 13MPa

���m

√. A review on

the technique [45] notes that the use of a guiding

groove, as was done in their work, is now recognisedto be undesirable as the effects of the stress concen-tration it causes are poorly understood [46]. It is alsonoted that crack deflection is a sign of a poorly alignedor balanced loading system. It is unsurprising that Linet al. observed crack deflection as with diamond’sexceptionally high stiffness, even the smallest misalign-ment will result in large-stress gradients. If the align-ment of the specimen was slightly off, the stressintensity at the crack tip could be much lower thanexpected, resulting in the unusually high value of frac-ture toughness observed. However, despite its potentialdisadvantages, the results of this work are particularlyinteresting as a natural crack is grown in the samplerather than the artificial notches used in other tests.

Lin et al. ’s work shows several micrographs oftransgranular fracture, Figures 7 and 8, furthersuggesting that PCD does not preferentially fracture

Figure 8. Higher magnification backscattered electron SEMimage of a crack running through a larger diamond grain.The authors describe this as indicative of crack deflection; how-ever, it appears the degree of deflection is relatively small.From Lin et al. [44].

Figure 5. Graphical representation of the strength vs. grain sizedata from Table 2. Studies were omitted if they did not havesufficient data on grain size. A general positive trend can beseen.

Figure 7. Backscattered electron SEM image of a mostlystraight crack path showing transgranular fracture. Its micro-structure is described as ‘coarse diamond grains (20–50 μm)embedded in a fine-grained diamond matrix, containing littlecobalt’. From Lin et al. [44].

Figure 6. Graphical representation of the strength vs. bindercontent data from Table 2. Studies were omitted if they didnot have sufficient data on binder content. The scatter in thedata is large and no obvious overall trend can be seen.

166 T. A. SCOTT

along the grain boundaries. The authors state thatcrack deflection appears to be a significant tougheningmechanism in PCD, however, by examining the micro-graphs provided it can be seen that this is a minimaleffect. It is also explained that in three of the samplestested, the crack advanced in discrete jumps with astress relaxation occurring after each jump. Theauthors suggest that this may be due to the presenceof the cobalt phase, presumably the plasticity therein,but due to only sporadic description of the micro-structures of the samples tested, it is impossible toform any kind of correlation.

While others have measured the fracture toughnessof PCD using single-edged v-notched beam (SEVNB)bending [40, 47], the most recent and in-depth studyof this type is by McNamara et al. [48]. However, theanalysis used in this paper, the theory of critical dis-tances, was originally developed for use with metalcomponents [49]. The ‘critical distance’ rc is initiallytaken to be the nominal grain size – a parameterwhich has already been seen to be inappropriatewhen discussing sintered PCD micro-structures.Upon realising the difficulty in defining it meaning-fully, the authors surmise that it is merely a fitting par-ameter. Figure 9 shows that a small change in rc resultsin a large change in the calculated fracture toughnesswhile still producing an acceptable fit to the exper-imental data. The authors note that this is not as cata-strophic an issue for fine-grained micro-structures as itis for coarse grained as the degree of particle crushing isnot as great, simplifying the choice of rc.

McNamara et al. also studied the effect that ‘residualmetal removal depth’ (leaching depth) had on the frac-ture toughness of PCD. Figure 10 shows that forcoarse-grained material, the toughness remains

constant until a certain depth and that fine-grainedmaterial’s toughness begins to degrade linearly assoon as leaching begins. It is unclear exactly why thistrend exists, but the authors suggest that fine-grainedPCD may be more sensitive to leaching as it has ahigher volume fraction of cobalt and hence there willbe more pores on the surface to act as flaws post-leaching.

The collated data on fracture toughness are shownin Table 3. Figure 11 shows that when aggregated,there is no discernible trend in fracture toughnessagainst grain size. However, within each individualstudy, a similar trend is apparent – that increasingthe grain size increases the fracture toughness butwith increasingly diminishing effect. The main con-clusion that can be drawn from this is that the changesin fracture toughness as the grain size varies are vastlyoutweighed by another effect that varies betweenstudies. One possible issue is that three different testingmethodologies are used within this data set. Evenwithin a single-testing method such as SEVNB bend-ing, different studies can be rendered incomparable ifthe notch roots are of different radii due to its strongeffect on the apparent toughness [50]. It is highly likelythat a similar effect will be present between differenttesting methodologies.

Figure 12, on the other hand appears to tell a com-pletely different story. Of the four plots of collateddata from the literature, fracture toughness againstbinder content appears to be the strongest. Thereseems to be a strong negative relationship betweenthe toughness and the amount of binder in the PCDwhich is surprising given that others have suggested[34] that the binder phase contributes a tougheningeffect. However, the interaction between cracks in

Figure 9. Fitting of various values of critical distance rc toexperimental data by McNamara et al. [48]. It can be seenthat varying rc from 1 to 30 μm changes the calculated valueof fracture toughness KIc dramatically while maintaining anacceptable fit to the experimental data.

Figure 10. Behaviour of the fracture toughness of variousgrades of PCD as the leaching depth (RMRD) is increased.The number in each label refers to the original diamond par-ticle size. Materials A, B and C contained 8.5, 7.5 and 6.5% bin-der phase, respectively. It is unclear whether this is volume orweight percentage. From McNamara et al. [48].

ADVANCES IN APPLIED CERAMICS 167

PCD, the binder phase and the overall toughness ofthe material are likely much more complicated thanthis graph appears to show. For example, the bindercontent does not vary independently of grain sizeand the methodology effect may confound the resultsfurther. One must be wary of drawing any concreteconclusions from the trend shown in this graph.Although there appears to be a low amount of scatter,the results of each study are tightly grouped in bindercontent space rather than each showing a similartrend. A more carefully controlled investigation intothe effect the binder has on fracture toughness istherefore required.

Flaw size distributionA piece of information conspicuous by its absence fromthe literature on PCD is the flaw size distribution.Despite being briefly discussed by Lammer and Petro-vic et al., no substantive effort to understand its influ-ence has been published. As indicated in previoussections, it is likely that this information would helpto explain the trends seen in the strength of PCD.One such technique that can probe the surface flawsize distribution in diamond is Hertzian indentation[51].

The first example of Hertzian indentation of dia-mond was by Howes and Tolansky [52] who were

Table 3. Fracture toughness data for various grades of PCD collated from numerous authors.Author Technique Fracture toughness (MPa

��m

√) Grain size (μm) Binder content (wt-%)

Petrovic et al. [38] SDC 3.4+ 0.34 2 304.4+ 0.44 4 –2.3+ 0.23 6 274.5+ 0.45 16 –5.0+ 0.50 28 186.0+ 0.60 29 185.0+ 0.50 30 185.8+ 0.58 45 –5.6+ 0.56 46 –6.4+ 0.64 76 –6.0+ 0.60 120 –

McNamara et al. [39] 3PB 9.55 10 8.59.53 10 7.58.52 10 6.5

Lin et al. [44] DT 14.28+ 0.20 30 –13.13+ 0.57 10 –13.43+ 0.66 20 –

13.03 – –13.12 – –

Morrell et al. [40] 3PB 8.8+ 0.75 – –Lammer [28] SDC 6.86+ 0.43 2 13

8.81+ 0.46 12 118.89+ 0.37 30 117.49+ 0.78 125 12

Devin et al. [35] SDC 5 – –6 – –

The following terminology is used for the technique column: SDC - slotted diametral compression, 3PB – 3 point bending, DT – double torsion. The errorshown in the strength column is the standard deviation in the results.

Figure 11. Collated data on the variation of fracture toughnesswith grain size. Although no overall trend can be seen, withineach study, a similar trend of initial increase followed by a pla-teau appears to exist.

Figure 12. Collated data on the variation of fracture toughnesswith binder content for various studies. Barring a single anom-alous point, a reasonably strong negative correlation appearsto exist. However, a common trend is not apparent across allof the data sets, resulting in data points grouped by study.

168 T. A. SCOTT

attempting to artificially induce the small, polygonalcrack figures that Tolansky and Halperin [53] hadseen on the octahedral faces of many natural diamonds.Lawn and Komatsu [54] then published a paperexplaining that the mechanism behind the deformationthat occurred when one of these so-called ‘pressurecracks’ was formed was completely elastic in nature.Field and Freeman [55] later used Hertzian indentationto measure the fracture toughness of single-crystal dia-mond. By plotting the square of the applied load vs. thecube of the cone-crack-base radius, something that isconsiderably easier to measure in transparentmaterials, the gradient can be used to calculate the frac-ture energy [56] and hence the fracture toughness. Thiswas calculated as 3.4 MPa

���m

√.

Despite this early work into single crystal and natu-ral diamonds, it was not until 2015 that the techniquewas applied to PCD. Marro et al. [34] performedHertzian indentation on various grades of PCDunder both monotonic and cyclic loading conditions.Although the technique of acoustic emission isreasonably well documented in the literature on Hert-zian indentation, it was not used to detect fracture inthis study. The authors instead loaded samples ofPCD with a spherical indenter up to predeterminedloads and then examined them in an optical micro-scope to determine whether or not fracture had

occurred. Although the data they collected was dis-tinctly qualitative in nature, they did find that afine-grained micro-structure was more resistant tothe initiation of a ring crack as compared to coarse-grained specimens. The authors present micrographsthat show local toughening by means of a reducedcrack-opening displacement as the crack passesthrough a binder pool (Figure 13). Initially, thisseems to disagree with the trend shown in Figure 12as an increased binder percentage appears to resultin a less tough material which could be taken to indi-cate that the binder phase is more prone to fracturethan the diamond. However, it would appear thatinteractions are more complicated than that simplisticview. The material shown in Figure 13 comprises 10wt-% binder which from Figure 12 suggests it shouldbe a highly tough material. In light of this, it seemspossible that the binder phase can still provide atoughening effect despite an increased percentage ofit not resulting in an overall tougher material.

Figure 14 shows SEM images of crack paths in acoarser-grained micro-structure. The crack bifurcatesas it passes through a large-diamond grain. From theinformation given and the background literature, it isquite unclear why this might happen.

The second paper on this topic published by Marroet al. deals with the damage tolerance of PCD materials

Figure 13. Backscattered electron images of a transgranular crack path in finer-grained PCD. The arrows show where the author hasdeemed the binder phase to be showing local toughening by means of a reduced crack-opening displacement. From Marro et al.[34].

Figure 14. Backscattered electron image of a transgranular crack path through a coarse-grained PCD material. The crack exhibitsbifurcation as it passes through particularly large grains. From Marro et al. [34].

ADVANCES IN APPLIED CERAMICS 169

after indentation by measuring the residual strength bymeans of ball-on-three-ball flexure [42]. This is a veryrelevant topic for PCD research as one can imagineHertzian-like fractures being caused by loaded contactwith rock asperities which could then lead to cata-strophic failure of the cutter. The authors found thatall materials retained their strength with indentationsof up to 400 N, likely due to no ring crack or damageof any form being evident at this load. At 1000 N, thefine-grained material experiences a drop in strengthof around 65% to a strength below that of the coarse-grained material (Figure 15).

One possible explanation for this is the initial flawsize distribution. If the flaws do scale proportionallywith the grain size as suggested by Petrovic et al. [38]then the lower initial strength of the coarse-grainedmaterial can be explained by larger initial flaws. AfterHertzian indentation, the ring crack introduced couldbe much larger than the flaws in the fine-grainedmaterial but of a similar size to those already presentin the coarse-grained PCD. This would then explainwhy the strength of the fine-grained material dropsdramatically but that of the coarse-grained remainsroughly constant – essentially the fine-grained materialis much stronger to start with and therefore is moresensitive to damage on the scale of that caused by Hert-zian indentation. However, without a measure of theinitial flaw size distribution, it is impossible to tell ifthis explanation is correct.

Wear testing

The vast majority of the applications of PCD involvewear between two bodies; drilling of rock, machiningof metal, wood and composites to name the most

common. In light of this, an understanding of howPCD behaves when it is subjected to wear is highlyimportant if we are to improve its wear resistance.

Drilling of rockHibbs and Lee conducted a study into the wearmechanisms that operate when PCD is used in arock removal process [57]. They used very coarse-grained PCD, approximately 100-mm grain size, tocut a helical groove in a mortar rock cylinder con-taining hard particles of quartz. Figure 16 showsthe two distinct wear morphologies that they foundon the surface of the PCD after testing. The authorsput forward an interesting hypothesis explaining thetwo differing morphologies. During sintering, the dia-mond grains undergo a large amount of plastic defor-mation due to the effect of high pressure combinedwith high temperature. It is suggested that this plasticdeformation prevents easy cleavage of grains due tothe disruption it causes to the lattice resulting inthe chipping of small pieces of diamond from grainsas seen in Figure 16(a). However, some grains maynot be plastically deformed to the same degree bythe chance arrangement of grains around them caus-ing a bridging effect and preventing them fromexperiencing the same high stresses. These grainsthen retain their easy cleavage planes and can frac-ture transgranularly as seen in Figure 16(b). Asthere have been no studies on the effect of plasticallydeforming diamond on its fracture behaviour to theauthor’s knowledge, this remains an interestinghypothesis to test.

Another significant conclusion from their work iswith regard to the wear resistance of the grain bound-aries. The authors found that grains contained heavyscratch damage in the grinding direction. Cobaltpools had been preferentially removed from the surfacelayer but grain boundaries neither preferentially frac-tured even when they were very close to the active cut-ting edge (Figure 17(a)) nor were preferentiallyabraded even when they were parallel to the cuttingdirection (Figure 17(b)). This gives further evidencethat the grain boundaries do not represent a signifi-cantly weaker link as part of the PCD micro-structureeven if the cobalt pools along them are removed, con-trary to what many authors have suggested. When alarge area was found to have broken away, a particu-larly large grain or evidence of poor intergranularbonding was often found showing that inhomogene-ities in the sintering of PCD may act as critical flawsfor fracture.

Lin et al. [58] identified three mechanisms of wearin a detailed study of cutters damaged in both labora-tory rock-cutting tests and field operations. ‘Smoothwear’ occurred where diamond grains appear to havebeen polished to a flat surface with no significant frac-ture occurring. This type of wear was predominantly

Figure 15. Graph showing the residual strength of variousgrades of PCD after varying Hertzian indentation loads wereapplied. The number in each label refers to the starting dia-mond particle size and the terminating letter denotes themodality: S – single mode, B – bimodal. The strength ofPCD20S and 10B remains roughly constant regardless of theHertzian load applied but PCD04S experiences a large dropin strength above 1000 N. From Marro et al. [42].

170 T. A. SCOTT

seen in cutting through strong, abrasive, homogeneousrock where the drill head had a large number of bitsdistributed over its surface. Grooves oriented in thedirection of cutting were also observed similar tothose shown in other work [59, 60]. The rock beingcut contained hard quartz particles but given the con-siderably higher hardness of diamond even to a veryhigh temperature [61, 62], these grooves could nothave been caused in a typical abrasive manner. Theauthors suggested that a combination of thermal andmechanical load on the diamond must have allowedeither oxidation or graphitisation to occur causingthese grooves to form.

The second wear morphology was labelled ‘micro-chipping’; although it may sound similar to one ofthe morphologies shown in Hibbs and Lee’s work, itcan be seen by looking at Figure 18 that it is in factquite different. This wear mechanism is characterisedby small flakes of material chipping off where theplane of fracture is roughly parallel to the directionof cutting. The mechanism only appeared when cuttingrock with a higher uniaxial compressive strength butlow abrasiveness and after varying the cutting deptha number of times. The authors therefore concludedthat this morphology was caused by some thermal ormechanical fatigue mechanism, citing work by Dunnand Lee who showed that the fracture stress of PCD

cutters was significantly lowered by cyclic loading[63]. They noted that a higher negative rake angle(associated with higher forces in the direction of thecut) resulted in worse damage by microchipping. Themechanism proposed is that the cracks develop fromtensile failure of the top face of the cutter. With ahigher negative rake angle (i.e. the axis of symmetryof the cutter moving further away from parallel withthe direction of cut) there is less material able to sup-port the cutting edge and so larger tensile stresses arepresent on the cracks that develop here leading toincreased microchipping.

The final morphology seen was that of gross fractur-ing. This involved large chunks of the cutter beingremoved instantaneously and represents the most pro-blematic wear/fracture mechanism for drilling oper-ations. Gross fractures were found to initiate on thecurved surface of the PCD cutter rather than on thetop face. Given that this type of fracture often occurredslightly after a change in rock formation during drilling[64], the authors suggested that it was as a result ofmechanical overloading. To test this hypothesis, theyloaded cutters against a steel platen in a similar geome-try to that of rock cutting. They found that this loadingcondition caused fractures very similar to the ‘grossfractures’ seen in operationally used cutters and thatby using a harder steel platen (causing an increased

Figure 16. Wear morphologies seen after rock drilling by Hibbs and Lee [57]. Panel (a) shows sub-grain chipping and panel (b)shows transgranular fracture of multiple grains.

Figure 17. Resistance of grain boundaries to fracture and abrasion after rock drilling shown by Hibbs and Lee [57]. Panel (a) showsthat a grain boundary has not preferentially fractured despite being close to the cutting edge and panel (b) shows that there is nopreferential abrasion of a grain boundary parallel to the cutting direction.

ADVANCES IN APPLIED CERAMICS 171

stress at the cutter’s edge), gross fracturing occurred atlower loads.

Yahiaoui et al. [65] performed a study comparingthe wear resistance of fine (6+ 1mm) and coarse(17+ 4mm)-grained PCD sintered using two HPHTregimes. A standard regime (HPHT1) and one with ahigher pressure but shorter dwell time. In the introduc-tion, the author states that it is ‘well known by the man-ufacturers [that] PDC cutters with a fine graindistribution are more resistant to abrasion than cutterswith coarse grains’ citing a review paper on the currentstate of PDC bit technology [15]. However, the resultsof the study did not agree as it was found that thecoarse-grained PCD was more wear resistant thanfine grained. The authors did not look for a mechanis-tic difference in the wear behaviour and suggested thatother microstructural parameters might be moreimportant to the wear behaviour than simply averagegrain size.

The influence of binder removal on the wear behav-iour of PCD was examined by Liu et al. [27]. The cobaltwas removed from PCD to various depths by electroly-sis and the wear ratio (ratio between amount of workpiece removed and amount of cutter worn away) eval-uated for each removal depth over a fixed sliding dis-tance. It was found that the wear ratio of the cutterincreased monotonically up to a removal depth ofaround 200mm showing that the removal of the binderphase increases the wear resistance of the cutter. Thiswas also seen in a visual comparison with binder-removed cutters having a far smaller wear scar thanthose with residual metal still present. Deep grooveswere noted on the wear scars of the binder-removedcutters. Although the authors did not suggest a mech-anism by which this could have happened, they suggestthat it may constitute a self-sharpening behaviour asthese grooves reduce the contact area between the cut-ter and the work piece, reducing heat and friction at thecontact and hence improving the cutting efficiency.XRD measurements showed that there was a completeabsence of graphite on the wear scars of the binder-removed material. This potentially corroborates Linet al.’s suggestion that graphitisation is a significantsource of wear and shows that catalysis by the binderphase is the pathway by which this happens. Another

possible explanation is that the removal of the binderreduces the friction of the contact [66] and hence inhi-bits graphitisation by a virtue of a lower temperatureinterface.

In a study of microstructural characterisation of thethermal degradation of PCD after milling of granite[67], Westraadt et al. found that volume expansiondue to graphitisation was a major contributor to failureof the cutter. This finding contradicts much of theindustrial wisdom that it is the thermal expansion ofthe cobalt that causes thermally induced cracking.They found through comparison with static annealingexperiments that the temperatures generated at the cut-ting interface must have exceeded 800◦C which is suf-ficient to cause the cobalt-catalysed conversion ofdiamond to graphite. Another interesting finding oftheir work is that much of the cracking they observedwas intergranular. This suggests that there may be adifferent fracture mechanism activated when PCD isused for drilling applications compared to pseudo-sta-tic mechanical tests at room temperature. If cobaltalong the grain boundaries is causing conversion ofdiamond to graphite at high temperature, resulting init expanding, it seems logical that the material wouldfracture along these grain boundaries rather than trans-granularly as is seen at low temperature.

Machining of titaniumA growth area for PCD is in the machining of high-strength titanium alloys that are particularly difficultto machine with traditional carbide and high-strengthsteel tooling. Turning tests of five different diamondgrain sizes on a billet of titanium found that the toollife increased as grain size increased from 1.3 to14mm [68]. Although, the largest grain size of 39mmperformed the worst by far, experiencing grain-pulloutand ‘sub-grain transgranular fracture’.

A study was conducted to measure the effect of cut-ting speed on the wear behaviour of PCD when used toturn Ti–6Al–4V titanium alloy [69]. It was found thatthe predominant wear mechanism depended on thecutting speed used. At the lower-cutting speed, thewear mechanism was described as attritional with arough, discontinuous wear surface and a completelack of adhered material to the tool. At the higher-

Figure 18. Micrographs of the wear morphology labelled ‘microchipping’ by Lin et al. [58].

172 T. A. SCOTT

cutting speed, a smooth wear surface was foundcharacteristic of diffusive/dissolutional wear. The pres-ence of adhered material on the cutter would seem tosupport this. Finite element modelling simulationswere used to calculate the expected temperatures atthe cutting interface, the results of which predicted atemperature of 897◦C for the lower-cutting speed and991◦C for the higher. The authors point out that atemperature difference of only 100◦C is not enoughto explain a difference in diffusivity so large that itwould cause the completely different wear mechanismsseen between the two cutting speeds. It was thereforesuggested that a temperature-induced phase changein the chip material caused the diffusivity of toolelements into the titanium to increase by orders ofmagnitude. This hypothesis appears to be backed upby the alloy’s phase diagram (taking pressure intoaccount). A secondary argument is that the HCP αphase of titanium contains only six facile slip systems,whereas the BCC β phase has 12. This means that at thelower-cutting speed where the chip material is pre-dicted to be of the α phase, the chip will not plasticallydeform as easily, increasing the stress on the cuttingedge and causing it to fracture more frequently.

Machining of wood compositesBai et al. [70] performed tests involving the machiningof laminate wood composites using four different PCDgrain sizes (2, 10, 25 and 75mm). It should be notedthat these are the average sizes of the original diamondgrains not a measurement of the sintered grain size.Micro-cracks were observed in the 10-mm grain sizematerial (Figure 19). It was suggested that it is the

formation of these microcracks that are responsiblefor the ‘rupture and disintegration of the PCD tool’.The authors also suggest that these microcracks canpropagate along grain boundaries. It may be that thisoccurs in larger-grained microstructures, being respon-sible for the often mentioned ‘grain pullout’ but it hasbeen shown on numerous occurrences that the grainboundaries do not act as a preferential path for failureso this statement seems unlikely without any evidenceprovided to support it. An interesting claim is madethat the ‘inherent cleavage cracks generated in HTHPsynthesis also contribute to the formation of cleavagewear’, however, as the paper cited against this claimis in Chinese [71], further investigation has proved dif-ficult. The final conclusion of this paper is that middle-sized diamond grains (10mm) have the highest wearresistance over both large and small grain sizes, furtherhighlighting the error in what is ‘known by allmanufacturers’.

Philbin and Gordon [11] analysed the wear mechan-isms that operated when cutting two different types ofwood composite with PCD. They found that when cut-ting a more homogeneous material, abrasive wear wasdominant, with voids where individual grains had beenpulled out being seen. When a material with a protec-tive, hard layer was cut, multi-grain chips wereobserved at the cutting edge instead.

A summary of all of the wear mechanisms describedherein is given in schematic form in Figure 20.

Summary

It is clear from the literature presented in this docu-ment that there are still many fundamental aspects ofthe relationship between microstructure and mechan-ical behaviour of PCD that are poorly understood.Some of the assumptions made to help understandthis behaviour are based either on questionable con-clusions or long held industrial wisdom, both ofwhich are unhelpful in furthering the scientific under-standing of the material.

A better explanation for the strength-grain sizebehaviour of PCD is suggested based on considerationof the flaw size distribution, but there is currently littleexisting literature to corroborate this. There appears tobe consensus between authors that increasing the grainsize of PCD increases its toughness, counter to theeffect seen in many traditional ceramics, however, noobvious reason has been discovered for thisphenomenon.

Many authors present micrographs of transgranularfracture occurring in PCD yet when explaining thefracture behaviour, persist in describing grain bound-aries as weak paths for fracture. It is clear that thesetwo arguments are incongruent and that a betterunderstanding of the fracture behaviour of PCD isrequired. In a number of the wear studies, different

Figure 19. Micro-cracks observed in 10-mm grain size PCDafter machining of a laminate composite by Bai et al. [70].

ADVANCES IN APPLIED CERAMICS 173

wear morphologies such as smooth wear, grain pulloutand microchipping are seen under different conditions.Intergranular fracture is almost never seen in pseudo-static testing but frequently appears in the results ofwear tests. It is suggested that the main driving factorbehind this phenomenon is the effect of temperatureon the material at the grain boundaries which is asyet relatively unstudied.

The binder phase clearly has an important effect onthe mechanical properties of PCD. Marro et al. showedthat it had a local toughening effect as cracks passedthrough it. However, the collated data in section ‘Frac-ture toughness’ suggests that increasing the overall

binder content of a PCD material reduces its tough-ness. Leaching appears to greatly improve the wearbehaviour of PCD while also moderately degradingits toughness which appears to be contradictory tothe previous statement. The interaction of cracks andthe binder phase is clearly complicated, but it is diffi-cult to vary the binder content of PCD independentlyof the grain size so a more direct way of analysingthis interaction is required.

Finally, it has been suggested that the plastic defor-mation of diamond grains that occurs during sinteringcould have an effect on the mechanical properties ofPCD. Some grains may remain undeformed if bridging

Figure 20. Schematic representation of all of the wear mechanisms of PCD described above. (a) Sub-grain microchipping due toplastic deformation blocking easy cleavage planes [57], (b) bridging of individual grains prevents plastic deformation allowingsingle grains to fracture along easy cleavage planes [57], (c) smooth wear caused by dissolution of carbon into the work pieceand/or chemical attack [58], (d) microchipping caused by tensile cracks opening up on the PCD table [58] (figure taken from refer-ence) (e) gross fracturing caused by excessive normal force on the cutter [58] (figure taken from reference), (f) pullout of an entiregrain caused by a weakening of the grain boundary [68].

174 T. A. SCOTT

by neighbouring grains occurs allowing these defor-mation-free grains to act as critical flaws. However, lit-tle research has been conducted investigating thiseffect.

By investigating these gaps in the literature, a muchmore in-depth and accurate understanding of how themicrostructure of PCD affects its mechanical proper-ties can be gained. This will allow continued develop-ment in the field of high performance cutting toolmaterials creating technological advancements in theaerospace, automotive, resource exploration and trans-portation industries of the same or greater magnitudeas when PCD first entered the market.

Acknowledgments

I would like to acknowledge my DPhil supervisors ProfRichard Todd and Dr Dave Armstrong for their support,encouragement and advice and the invaluable technicalexpertise and wisdom of the Element Six Global InnovationCentre team including but not limited to Serdar Özbayrak-tar, Jonee Zunega, Roger Nilen, Valentine Kanyanta, EdEardley and Rod Cruise.

Disclosure statement

No potential conflict of interest was reported by the author.

Funding

This work was supported by Engineering and PhysicalSciences Research Council [AR0743] in conjunction withthe Centre for Doctoral Training in Diamond Science andTechnology.

References

[1] Tennant S. On the nature of the diamond. Philos TransR Soc Lond. 1797;87:123–127.

[2] Moissan H. Nouvelles expériences sur la reproductiondu diamant [New experiments on diamond reproduc-tion]. C R Acad Sci Paris. 1894;118:320–326.

[3] Bridgman PW. Synthetic diamonds. Sci Am. 1955;193(5):42–46.

[4] Collins AT. The Hannay diamonds. Ind Diam Rev.1975;35:434–437.

[5] Bundy FP, Hall HT, Strong HM, et al. Man-made dia-monds. Nature. 1955;175(4471):642–643.

[6] Bovenkirk HP, Bundy FP, Hall HT, et al. Preparationof diamond. Nature. 1959;184(4693):1094–1098.

[7] Badzian AR, Klokocki A. On the catalytic growth ofsynthetic diamonds. J Cryst Growth. 1981;52:843–847.

[8] Hall HT. Sintered diamond: a synthetic carbonado.Science. 1970;169(3948):868–869.

[9] Katzman H, Libby WF. Sintered diamond compactswith a cobalt binder. Science. 1971;172(3988):1132–1134.

[10] Wentorf Jr. RH, Rocco WA. Diamond tools formachining. United States patent US 3745623 A. 1973.

[11] Philbin P, Gordon S. Characterisation of the wearbehaviour of polycrystalline diamond (PCD) tools

when machining wood-based composites. J MaterProcess Technol. 2005;162–163:665–672.

[12] Wang YG, Lui B, Song JY, et al. Study on the wearmechanism of pcd tools in high-speed milling of Al-Si alloy. Adv Mater Res. 2012;381:16–19.

[13] Kim D, Beal A, Kwon P. Effect of tool wear on holequality in drilling of carbon fiber reinforced plastic-titanium alloy stacks using tungsten carbide and poly-crystalline diamond tools. J Manufact Sci Eng.2015;138(3):03100603101031006–1 –031006–11.

[14] Bellin F, Dourfaye A, KingW, et al. The current state ofPDC bit technology: Part 1. World Oil. 2010;231(09):41–46.

[15] Bellin F, Dourfaye A, KingW, et al. The current state ofPDC bit technology: Part 3. World Oil. 2010;231(11):67–71.

[16] Jahanmir S, Ramulu M, Koshy P. Machining of cer-amics and composites. Taylor & Francis; 1999.New York: Manuf Eng Mater Process.

[17] Arsecularatne JA, Zhang LC, Montross C. Wear andtool life of tungsten carbide, PCBN and PCD cuttingtools. Int J Mach Tool Manu. 2006;46(5):482–491.

[18] Wang DH, Wern CW, Ramulu M, et al. Cutting edgewear of tungsten carbide tool in continuous and inter-rupted cutting of a polymer composite. Mater ManuProcess. 1995;10(3):493–508.

[19] Walmsley JC, Lang AR. Characteristics of diamondregrowth in a synthetic diamond compact. J MaterSci. 1988;23(5):1829–1834.

[20] Walmsley JC, Lang AR. Transmission electron micro-scopic observations of deformation and microtwinningin a synthetic diamond compact. J Mater Sci Lett.1983;2(12):785–788.

[21] Özbayraktar S. PCD Sintering Model. De BeersIndustrial Diamond Technical Note. 1997.

[22] Westraadt JE, Sigalas I, Neethling JH. Characterisationof thermally degraded polycrystalline diamond. Int JRefract Metals Hard Mater. 2015;48:286–292.

[23] McNamara D, Alveen P, Damm S, et al. A Ramanspectroscopy investigation into the influence of ther-mal treatments on the residual stress of polycrystallinediamond. Int J Refract Metals Hard Mater.2015;52:114–122.

[24] Mehan RL, Hibbs LE. Thermal degradation of sintereddiamond compacts. J Mater Sci. 1989;24:942–950.

[25] Bovenkerk H, Gigl P. Temperature resistant abrasivecompact and method for making same. United Statespatent US 4,224,380. 1980.

[26] McNamara D. The mechanical and fracture propertiesof polycrystalline diamond as a function of microstruc-ture [dissertation]. National University of Ireland;2015.

[27] Liu C, Kou Z, He D, et al. Effect of removing internalresidual metallic phases on wear resistance of polycrys-talline diamond compacts. Int J Refract Metals HardMater. 2012;31:187–191.

[28] Lammer A. Mechanical properties of polycrystallinediamonds. Mater Sci Technol. 1988;4(11):949–955.

[29] Field JE. The mechanical and strength properties ofdiamond. Rep Prog Phys. 2012;75(12):126505.

[30] Gigl PD. High-pressure science and technology, Vol. 1.New York: Plenum Press; 1979.

[31] Roberts DC. Syndite – its mechanical and physicalproperties. Ind Diam Rev. 1979;39:237–241.

[32] Rice RW. Fracture mechanics of ceramics, Vol. 1.New York: Plenum Press; 1974.

ADVANCES IN APPLIED CERAMICS 175

[33] Catledge SA, Vohra YK, Ladi R, et al. Micro-Ramanstress investigations and X-ray diffraction analysis ofpoly crystalline diamond (PCD) tools. Diam RelMater. 1996;5(10):1159–1165.

[34] Marro FG, Mestra A, Jiménez-Piqué E, et al.Damageinduced by monotonic and cyclic spherical indentationin polycrystalline diamond (PCD). Int J Refract MetalsHard Mater. 2015;49:292–301.

[35] Devin LN, Maistrenko AL, Simkin ES, et al. Resistanceto cracking of hard-metal composite materials. SovPowder Metall. 1982;21:419–423.

[36] Nui LS, Chehimi C, Pluvinage G. Stress field near alarge blunted tip v-notch and application of the con-cept of the critical notch stress intensity factor (nsif)to the fracture toughness of very brittle materials.Eng Fract Mech. 1994;49(3):325–335.

[37] Fett T,MunzD. Stress intensity factors andweight func-tions. Southampton, UK: Computational MechanicsPublications; 1997. (Advances in fracture series).

[38] Petrovic M, Ivankovic A, Murphy N. The mechanicalproperties of polycrystalline diamond as a function ofstrain rate and temperature. J Eur Ceram Soc.2012;32(12):3021–3027.

[39] McNamara D, Carolan D, Alveen P, et al. The influ-ence of microstructure on the fracture statistics ofpolycrystalline diamond and polycrystalline cubicboron nitride. Ceram Int. 2014;40(8):11543–11549.

[40] Morrell R, Danzer R, Supancic P, et al. Meso-scalemechanical testing methods for diamond compositematerials. Int J Refract Metals Hard Mater. 2010;28(4):508–515.

[41] Belnap D, Griffo A. Homogeneous and structuredPCD/WC-Co materials for drilling. Diam Rel Mater.2004;13(10):1914–1922.

[42] García-Marro F, Mestra A, Kanyanta V, et al. Contactdamage and residual strength in polycrystalline dia-mond (PCD). Diam Rel Mater. 2016;65:131–136.

[43] Miess D, Rai G. Fracture toughness and thermal resist-ance of polycrystalline diamond compacts. Mater SciEng A. 1996;209(1–2):270–276.

[44] Lin T-P, Cooper GA, Hood M. Measurement of thefracture toughness of polycrystalline diamond usingthe double-torsion test. J Mater Sci. 1994;29:4750–4756.

[45] Shyam A, Lara-Curzio E. The double-torsion testingtechnique for determination of fracture toughnessand slow crack growth behavior of materials: a review.J Mater Sci. 2006;41(13):4093–4104.

[46] Tait RB, Fry PR, Garrett GG. Review and evaluation ofthe double-torsion technique for fracture toughnessand fatigue testing of brittle materials. Exp Mech.1987;27(1):14–22.

[47] Petrovic M, Carolan D, Ivankovic A, et al. Role of rateand temperature on fracture and mechanical proper-ties of PCD. Key Eng Mater. 2011;452:153–156.

[48] McNamara D, Alveen P, Carolan D, et al. Fracturetoughness evaluation of polycrystalline diamond as afunction of microstructure. Eng Fract Mech.2015;143:1–16.

[49] Taylor D. Crack modelling: a technique for the fatiguedesign of components. Eng Fail Anal. 1996;3(2):129–136.

[50] Damani R, Gstrein R, Danzer R. Critical notch-rootradius effect in SENB-S fracture toughness testing. JEur Ceram Soc. 1996;16(7):695–702.

[51] Warren P. Determining the fracture toughness ofbrittle materials by Hertzian indentation. J EurCeram Soc. 1995;15(3):201–207.

[52] Howes VR, Tolansky S. Pressure crack-figures on dia-mond faces. I. The octahedral face. Proc R Soc A.1955;230(1182):287–293.

[53] Tolansky S, Halperin A. Oriented ring cracks on dia-mond. Proc Phys Soc B. 1954;67(6):473–476.

[54] Lawn BR, Komatsu H. The nature of deformationaround pressure cracks on diamond. Philos Mag.1966;14(February):689–699.

[55] Field JE, Freeman CJ. Strength and fracture propertiesof diamond. Philos Mag A. 1981;43(3):595–618.

[56] Roesler FC. Brittle fractures near equilibrium. ProcPhys Soc B. 1956;69(10):981–992.

[57] Hibbs LE, Lee M. Some aspects of the wear of polycrys-talline diamond tools in rock removal processes. Wear.1978;46(1):141–147.

[58] Lin TP, Hood M, Cooper GA, et al. Wear and failuremechanisms of polycrystalline diamond compact bits.Wear. 1992;156:133–150.

[59] Miller D, Ball A. Rock drilling with impregnated dia-mond microbits – an experimental study. Int J RockMech Min Sci. 1990;27(5):363–371.

[60] Miller D, Ball A. The wear of diamonds in impregnateddiamond bit drilling. Wear. 1991;141:311–320.

[61] Wentorf RH, DeVries RC, Bundy FP. Sintered super-hard materials. Science. 1980;208:873–880.

[62] Osburn HJ. Wear of rock cutting tools. Powder Metall.1969;12(24):471–502.

[63] Dunn KJ, Lee M. The fracture and fatigue of sintereddiamond compact. J Mater Sci. 1979;14:882–890.

[64] Brett FJ,Warren TM, Behr SM. Bit whirl – a new theoryof PDC bit failure. Soc Petrol Eng. 1990;5(4):275–281.

[65] Yahiaoui M, Paris JY, Delbé K, et al. Quality and wearbehavior of graded polycrystalline diamond compactcutters. Int J Refract Metals Hard Mater. 2016;56:87–95.

[66] Damm O, Tank K, Tomlinson P. Bearing assembly. EPPatent 0,617,207. 1998.

[67] Westraadt JE, Sigalas I, Neethling JH. Characterisationof thermally degraded polycrystalline diamond. Int JRefract Metals Hard Mater. 2015;48:286–292.

[68] Pretorius CJ, Soo SL, Aspinwall DK, et al. Tool wearbehaviour and workpiece surface integrity when turn-ing Ti-6Al-2Sn-4Zr-6Mo with polycrystalline diamondtooling. CIRP Ann Manuf Technol. 2015;64(1):109–112.

[69] Schrock DJ, Kang D, Bieler TR, et al. Phase dependenttool wear in turning Ti-6Al-4V using polycrystallinediamond and carbide inserts. J Manuf Sci Eng.2014;136(4):041018.

[70] Bai QS, Yao YX, Bex P, et al. Study on wear mechan-isms and grain effects of PCD tool in machining lami-nated flooring. Int J Refract Metals Hard Mater.2004;22(2–3):111–115.

[71] Li Y, Han XL, Wang XQ. Research on mechanicalproperties and test methods of polycrystalline dia-mond. Diam Abrasives Eng. 2001;5:27–30.

176 T. A. SCOTT