Wear 298–299 (2013) 32–41

Contents lists available at SciVerse ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

A study on PDC drill bits quality

M. Yahiaoui a,n, L. Gerbaud b, J.-Y. Paris a, J. Denape a, A. Dourfaye c
a
Laboratoire Génie de Production, Université de Toulouse, Ecole Nationale d’Ingénieurs de Tarbes, 47 avenue d’Azereix 65016 Tarbes, France
b
Centre de Géosciences, Mines Paritech, France
c
Varel Europe, France

a r t i c l e i n f o a b s t r a c t

Article history: The quality of innovating PDC (Polycrystalline Diamond Compact) bits materials needs to be
Received 12 July 2012 determined with accuracy by measuring cutting efficiency and wear rate, both related to the overall
Received in revised form mechanical properties. An original approach is developed to encompass cutting efficiency and wear
21 November 2012
contribution to the overall sample quality. Therefore, a lathe-type test device was used to abrade
Accepted 5 December 2012
specific samples from various manufacturers. Post-experiment analyzes are based on models establish-
Available online 2 January 2013
ing coupled relationships between cutting and friction stresses related to the drag bits excavation
Keywords: mechanism. These models are implemented in order to evaluate cutting efficiency and to estimate wear
PDC cutters of the diamond insert. Phase analysis by XRD and finite element simulations were performed to explain
Wear rate
the role of physicochemical parameters on the calculated quality factor values. Four main properties of
Quality factor
PDC material were studied to explain quality results obtained in this study: cobalt content in samples
Cobalt carbide
Residual stresses that characterizes hardness/fracture toughness compromise, undesired phase as tungsten carbide
Leaching weakening diamond structure, diamond grains sizes and residual stresses distribution affecting
abrasion resistance.
& 2012 Elsevier B.V. All rights reserved.

1. Introduction commonly used in several works to describe PDC/rock behavior [3].

Excavation performance depends on cutting efficiency which is
The main tools employed in the drilling industry are roller cone initially determined by the sample depth of cut. During friction,
and drag bits. Roller cone bits work by impact excavation and are cutting efficiency evolves and the change is closely linked to wear
currently used in hard rock formations because of a convenient wear flat formation on PDC cutters. Because a long bit life could be related
resistance. Drag bits rather operate by shear mode in softer rock to to a poor cutting performance and vice versa, this paper proposes an
medium hard formations. Nevertheless, they suffer from thermal objective quality criterion to clearly classify PDC cutters. Drilling
abrasive wear and impact damage while drilling interbedded forma- mechanisms and material analyzes are taken into consideration to
tions. As excavation rate is directly related to the overall cost, the interpret the grading of the testing bits.
drag bits using PDC (Polycrystalline Diamond Compact) cutters are
really attractive compared to roller cone bits. In fact, PDC bits could
drill twice faster and longer than roller bits even in hard forma- 2. PDC samples
tions [1]. Petroleum and hydrothermal investigations in deep geolo-
gical formations lead to manufacturing new bits materials able to Six cutters coming from various manufacturers (referred from A
drill at higher temperature, in more abrasive and harder geological to F) were selected to represent a large range of properties. Cutters
fields. Such innovating materials, sintering processes and design, are made of a tungsten carbide cylinder surmounted by a diamond
recently developed to improve drill bits hardness and fracture table (Fig. 1a). Material parts have a diameter of 13 mm: the
toughness, also require new strategies in quality assessment. Drag tungsten carbide cylinder has a height of 8 mm and the diamond
bits are mostly damaged by abrasion [2] and thus quality can be layer is around 2 mm thick. Diamond tables have a chamfer of
defined by two main parameters: materials wear rate and excavation 451  0.4 mm except for sample C where it is 451  0.7 mm. These
performance. Wear rate calculus by Archard’s model has been cutters were sintered by HPHT (i.e. High Temperature and High
Pressure) at a temperature over 1400 1C under a pressure close to
5.5 GPa (Fig. 1b) [4].
n
Corresponding author. Tel.: þ33 5624 42700; fax: þ 33 5624 42708.
Tungsten carbide prismatic grains in a binder cobalt phase form
E-mail addresses: [email protected], the substrate part (Fig. 2a). The mean grain size of tungsten carbide is
[email protected] (M. Yahiaoui). around 2 mm with minimum and maximum values observable under

0043-1648/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.wear.2012.12.026
 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41 33

Nomenclature FT total drag force, N

G grinding ratio
a back rake angle, deg I sum of maximum peak of present phases
e intrinsic specific energy, J m3 ICoCx CoCx XRD maximum peak intensity
Z cutting efficiency Idiamond diamond XRD maximum peak intensity
m friction coefficient IWC WC XRD maximum peak intensity
rx cobalt mass content at distance x k wear rate, m3 N1 m1
z cutting coefficient L excavation distance, m
Ac cross-sectional area of cut, m2 LT total excavation distance, m
Af wear flat area, m2 Q quality factor
D infiltration coefficient, m2 s1 R2 coefficient of determination
E specific energy, J m3 t time of infiltration, s
E0 initial specific energy, J m3 u cutting capacity, m
Em cutting dissipated energy, J VC cutter worn volume, m3
Fc cutting force component, N VR cut rock volume, m3
Ff friction force component, N Wm cutter mechanical work, J
FN total normal force, N xi infiltration transition position, m
F 0N initial normal force, N

a micrometer and over 10 mm (Table 1). Jeol JSM-7000F field distribution was measured on the samples (Fig. 3) using energy
emission scanning electron microscope observations revealed aggre- dispersive X-ray spectrometry (EDX) analyzes with Bruker XFlash
gates of micrometric diamond grains also surrounded by cobalt 4010 detector. To perform semi-quantitative measurements, the
(Fig. 2b). Samples A, E and F have been exposed to a chemical post- detector was calibrated with copper located close to samples
treatment called ‘‘leaching process’’ [5]. This treatment removes before each observation campaign. For these measurements, the
interstitial cobalt grain boundaries on the diamond layer beyond six samples were longitudinally cut by electroerosion, polished
several tens of micrometers (Fig. 2c). and metalized with palladium. The SEM was adjusted at 15 kV
The cobalt phase in the diamond part is due to the infiltration with a working distance of 15 mm. The electron beam intensity
of cobalt from the tungsten carbide substrate during sintering. was set around 100 counts per second to enable a high speed
Commonly, cobalt proportion can represent 6–18 wt.% in tung- analysis. The cobalt mass content distribution was evaluated with
sten carbide substrate and 2–8 wt.% in the diamond part. a step of 500 mm along a line on sections.
The cobalt distribution in samples follows a law that can be expressed EDX characterizations showed that all samples have similar
as a solution [6] of differential equations from Fick’s laws (Eq. (1)). cobalt content (r0 ) around 3 wt.% in the diamond material
  whereas cobalt content of tungsten carbide (r10 ) part can vary
1
erfc pffiffiffiffiffiffiffiffiffiðxxi Þ from 8 to 17 wt.% (Table 2). The square root of D  t permits to
rðxÞ ¼ ðr0 r10 Þ 2 Dt  þ r10 ð1Þ evaluate dispersion of the inflection i.e. metal ability to spread
1
erfc  pffiffiffiffiffiffiffiffiffixi from tungsten carbide to diamond. D depends on diamond/WC
2 Dt grains size and on sintering temperature. At sintering tempera-
In this equation, rðxÞ represents axial cobalt mass content from ture, molten cobalt moves by capillarity through voids between
diamond face (where rðxÞ ¼ r0 ) to the bottom of the tungsten diamond grains. Larger voids are directly associated with larger
carbide part (where rðxÞ ¼ r10 ). D is the infiltration coefficient, t is grain size which favors displacement of cobalt [7]. Moreover,
the time of infiltration and xi expresses infiltration transition metal infiltration in diamond structures increases with tempera-
position between PDC and WC-Co materials. The cobalt ture as viscosity of molten cobalt decreases.
SEM observations only permit the measurement of diamond
aggregates (see Table 1). The aggregates size does not represent
the diamond grain size distribution in the sample and cannot be
directly related to the mechanical behavior of PDC materials. pffiffiffiffiffiffiffiffiffi
Considering that t is almost equal for the six samples, D  t
parameter permits to qualitatively evaluate diamond grain sizes
pffiffiffiffiffiffiffiffiffi
rather than of aggregate ones. Here, B and C displays D  t values
two times higher than those of samples A, D, E and F. Theses
results may be due to higher diamond grain sizes in samples B
and C than in the others.
HPHT sintering
over 1400 °C and 5.5 GPa Leaching process:
depletion of cobalt
Diamond powder
Chamfer
3. Experimental study
2 to 8%wt Co

A vertical lathe-type device was used to simulate drilling

WC-Co Co Interface
6 to 18%wt Co design conditions. Cutters brazed on sample holders were adjusted
downward on the lathe shaft. Ring-stone counter-faces were
made of a manufactured mortar rock (1 m in external diameter,
Manufacturing Process
0.5 m in internal diameter and 0.6 m thick with a density of
Fig. 1. PDC cutter: (a) photography of a cutter; (b) manufacturing process of a 2210 kg m3 ). This mortar ensures homogeneous chemical
cutter. composition (silica content of 80 wt.%) and mechanical properties
34 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41tpdel

Polymer
resin
Leached

Chamfer

Non-leached

Fig. 2. SEM images at 15 kV: (a) sample B tungsten carbide-cobalt part by secondary electron analysis; (b) sample B diamond part by backscattered electron analysis;
(c) illustration by secondary electron and colorized energy dispersive X-ray spectrometry (EDX) mapping of cobalt leaching zone for sample F. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

Table 1
Geometry and microstructural properties of PDC samples A to F measured by SEM image analysis.

Cutters A B C D E F

Section view 8.3  6.2 mm

Side diamond thickness (mm) 2.3 70.05 1.6 7 0.05 2.17 0.05 2.4 70.05 2.0 70.05 2.1
Diamond aggregate size (mm) 13.3 73.9 15.1 7 5.8 9.87 4.8 8.1 75.1 11.6 7 4.6 11.07 5.0
min o 4:5 min o 7:2 min o 3:1 min o 2:3 min o 3:2 min o 3:4
max4 21:8 max 429:9 max4 23:8 max4 20:5 max 421:4 max4 22:4
WC-Co grain size (mm) 2.5 71.1 1.9 7 1.7 2.07 1.3 1.4 71.1 1.7 7 1.0 2.27 1.1
min o 0:8 min o 0:3 min o 0:5 min o 0:4 min o 0:5 min o 0:3
max4 5:7 max 49:8 max4 8:3 max4 9:1 max 46:2 max4 7:0
Leaching depth (mm) 707 4 – – –  100a 3257 30

a
Not measured, manufacturer data.

(compressive strength of 21.6 MPa and young modulus of mortar ring works for a total length of 25 500 m. At the end of
48.1 GPa). Experiments were carried out according to real drilling each sequence, the height of material lost was measured to
conditions (Fig. 4a): normal load ranged from 3000 to 5000 N, calculate cutting active area Ac and cutter worn volume V c .
back rake angle at 151, penetration depth of 2 mm and mean
cutting speed of 1:8 m s1 . Tests were conducted in atmospheric
environment and no lubricant was added into the contact. 4. Wear rate analysis
Experiments were performed following four sequences on the
mortar ring and each sequence represents three radial round-trips PDC drill bits are made of tens of cutters (e.g. VTD616 tools
(i.e. an excavation length about 510 m). As a consequence, one from Varel International have 48 face cutters). Operating
 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41 35

parameters as weight and torque on bit, defined for a constant while cutter B obtains the highest rate value over
penetration speed, could be considered as independent of the 16  108 mm3 N1 m1 (Fig. 5b).
number of bits [8]. This condition is assumed by a correct cutters
repartition on the bit, which ensures a homogenized wear on
every PDC [9]. Therefore, wear behavior and drilling performance
information based on single cutter experiments are relevant and 5. Cutting capacity
can be extrapolated to understand the whole tool behavior.
Afterward, Fairhurst and Lacabanne [10] assume that cutting Detournay and Defourny [11] established a linear relationship
action and sliding friction can also be considered as independent between torque on cutter (i.e. F T ) and weight on cutter (i.e. F N ).
(Eq. (2)) in the drilling procedure (Fig. 4). Thus, normal force F N This equation involves three constant parameters e, m and z which
applied onto a cutter can be expressed as the sum of friction are respectively the intrinsic specific energy, the friction coefficient
normal force F fN and cutting normal force F cN . Moreover, an and the cutting coefficient (Eq. (5)). The friction coefficient
approximation of F fN can be expressed as the difference between represents the ratio between the transversal and normal friction
normal force applied to a cutter and initial value of the normal forces. The cutting coefficient is calculated by the ratio of the
force, noted as F 0N normal cutting component on the transversal one. The intrinsic
specific energy is defined by the ratio between the normal cutting
F N ¼ F fN þ F cN  F fN þF 0N ð2Þ
force and the cutting active area Ac . This parameter has the
The same approach is assumed for the transverse force F T , so dimension of a stress and can be related to the strength of the
that (Eq. (3)) rock toward the cutting action.

F T ¼ F fT þ F cT  F fT þ F 0T ð3Þ F T ¼ mF N þAc ð1mzÞe ð5Þ

Because this study is clearly a case of abrasive friction between Moreover, specific energy E represents the dissipated energy
cutter and hard rock, Archard’s model is an interesting choice. Em needed to cut a unitary volume of rock V R . The energy Em
This model has been extensively involved in tribological studies equals the transverse force F T multiplied by the cutter travel
because of its simple linear relationship (Eq. (4)) between wear distance L. The lateral displacement of the tool front face implies
volume V c and the product of the normal force F fN by the sliding that dug volume V R equals the product of active area Ac by
distance L (Fig. 5a) distance L. As a consequence, F T and Ac measurements permit
V c ¼ k  F fN L  k  ðF N F 0N ÞL ð4Þ

The coefficient of proportionality k is usually called wear rate FN

and could be expressed as a function of rock or cutter hardness,
but only proportionality is considered here, and its meaning is not FT
identified. Back rake angle :α
Wear rate calculations show that cutter A gives the best wear
behavior with a rate value lower than 1  108 mm3 N1 m1 , Sample holder

WC-Co excavation
C
PD

Cutting active area :Ac

Af

c
FN f
c
F
f FN
F
c
FT f
Ac FT
Af

Fig. 3. EDX longitudinal measurement of cobalt content from diamond to Fig. 4. Schematics of cutter/rock mechanical interactions: (a) overall cutter
tungsten carbide material for sample B. disposition; (b) cutting components; (c) friction components.

Table 2
Cobalt mean content and migration parameters obtained by EDX profiles analyzes for samples A to F (measurements of cobalt content exclude leached fields).

Cutters A B C D E F

Co wt.% in Diamond 2.9 70.3 2.97 0.8 2.77 0.4 3.5 7 0.4 3.4 7 0.2 3.2 70.2
Co wt.% in WC-Co 8.2 70.2 16.7 70.2 9.87 0.2 11.2 7 0.2 9.6 7 0.1 7.7 70.1
pffiffiffiffiffiffiffiffiffi
D  t (mm) 0.2 70.2 0.87 0.1 0.57 0.1 0.1 7 0.1 0.3 70.1 0.17 0.1
xi (mm) 3.0 70.1 3.37 0.2 2.77 0.1 2.4 7 0.1 2.6 7 0.1 2.8 70.1
36 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41tpdel

0.5 1.2
2
R =98.4 % 1.1
Wear volume,Vc (mm3)

0.4

Cutting efficiency,
1.0
0.3
0.9

0.2 0.8

0.7
0.1

0.6
0.0
0 5 10 15x106 0 2 4 6 8 10 12x103
f
Fn · L (N.m) Distance (m)

22

150
Wear rate,k (10-9mm3·N·m-1)

20
Cutting capacity, u (m)

18
100
16

14
50

12

0 10
A B C D E F A B C D E F
PDC sample PDC sample
Fig. 5. Wear rate evaluation: (a) Archard’s model applied to sample F; (b) wear Fig. 6. Cutting efficiency analysis: (a) Cutting efficiency vs. distance model for
rates results. sample F; (b) cutting capacity results.

specific energy calculations (Eq. (6)) law as a function of excavating distance seems adequate to
empirically evaluate the relative cutting efficiency behavior for
Em FT FN
E¼ ¼ ¼m þ ð1mzÞe ð6Þ each series of PDC samples. A constant coefficient u, expressed in
VR Ac Ac
meters, is introduced here and is identified as ‘‘cutting capacity’’
The ratio between e and E characterizes drilling process (Eq. (8))
efficiency Z, which represents the cutting part in the overall  
L
mechanical action. Eventually, a pure cutting process means that Z ¼ exp  ð8Þ
u
m ¼ 0 then E ¼ e  E0 where E0 is the initial value of E (Eq. (7))
The lowest value of cutting capacity u of 12 km was measured
Ee with B and the highest value around 20 km was obtained with A
Z¼ ¼ 0 ð7Þ
E E and F cutters (Fig. 6b).
For all cutters, experiments revealed that cutting efficiency Z
plotted as a function of distance L follows a decreasing and
nonlinear curve (Fig. 6a). When excavation test starts, Z is closed 6. Quality model
to 1 (new cutter fully efficient) and as expected, Z tends to 0 when
the distance L becomes greater (mathematically infinite). Obviously, separate analyzes of wear endurance and cutting
The experimental cutting efficiency Z can reach values higher efficiency do not clearly discriminate the relative quality of
than 1 at the beginning of the test, but these values are due to cutters. The grade assessment of PDC samples depends on its
fluctuations in rock homogeneity and transitory periods occurring tribological behavior (i.e. friction and wear), cutting efficiency and
before cutting process stabilization. Accordingly, an exponential rock cutting resistance (i.e. intrinsic specific energy). A low wear
 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41 37

rate and a high cutting efficiency of the sample ensure a high According to quality factor results (Fig. 7), cutter A is clearly
quality cutter. In order to compare cutter performances, a quality the best one with a Q factor about 169  104 , which is confirmed
criterion must involve both wear rate and cutting efficiency. by its low wear rate and its high cutting efficiency. In contrast, the
For that purpose, we first consider the grinding ratio G, which is cutter B registers the worst Q factor of only 4  104 .
the ratio between rock wear volume V R and cutter wear volume Four parameters influence cutter behavior: the amount of
V c . The grinding ratio is commonly used to estimate cutters cobalt in the diamond table, the diamond grain size distribution,
resistance by abrasive wear. By combining previous equations, residual stresses and defaults (inclusions, cracks or porosities).
this ratio can be directly related to the parameters e, m, k and Z Cobalt ductile phase is directly associated to wear kinetic
(Eq. (9)) and the higher the cobalt proportion, the easier the abrasion of
a drilling tool. In other words, master cobalt distribution in
VR Ac L mF cT mF cT
G¼ ¼ f
¼ ¼ c diamond permits to handle tools wear resistance. In this study,
Vc k  FNL f
ekF T ekðF T F T Þ cobalt content differences in diamond parts are not significant
!1  1 and no rank can be made in relation with cobalt changes in
m FT m 1
¼ 1 ¼ 1 ð9Þ samples. In addition, leached samples can not be discriminated
ek F 0T ek Z
from non-leached samples by the quality factor defined here.
Cutters D, E and F are made by the same manufacturer and
By introducing the experimental cutting capacity parameter u,
the leached depth is the only parameter differentiating them.
the grinding ratio can be expressed as follows (Eq. (10)):
Samples E and F are leached on less than 400 mm depth which
2 31
corresponds to 20% of diamond layer thickness. Leaching depths
m6 6 1 7 represent respectively 14% and 40% of the final worn diamond
G¼   17
ek 4 1 5 thicknesses of samples E and F. This could explain the low
exp   L
u influence of the leaching process on long excavation distances.
   1 As seen above, cutters B and C may have a higher diamond
m L
¼ exp 1 grain size than others samples. These two cutters have the highest
ek u
wear rates and lowest quality factors. As expressed by Bellin et al.
" #
þ 1  n 1 [12], PDC cutters with a fine grain distribution are more resistant
m L X L
¼ þ ð10Þ to abrasion than cutter with coarse grains. Nevertheless, it is an
ek u n ¼ 2 u
accepted fact that a coarse grain distribution permits a better
From here, it is interesting to define the quality factor Q, which impact resistance.
is a dimensionless value integrating all variations of cutting
efficiencies Z over travel distance LT (Eq. (11)). This distance
should be high enough to measure significant wear evolution on
samples. LT was set at 12 580 m for the six samples. 7. Phase analysis and quality factor
m u
Q¼ ð11Þ The samples were submitted to XRD (X-ray diffraction)
eLT k
measurements to study chemical differences between each diamond
These results display the interest in confronting m, e, LT , k and u tables (Fig. 8). Diffractograms were recorded on a X’PERT Philips MRD
parameters and evaluating the quality factor (Table 3). First, this diffractometer with CuK a radiation source beam at 40 kV and 50 mA.
definition of quality factor takes into account coefficients m and e. XRD measurements are also defined by 2y Bragg angles from 101 to
The friction coefficient m is directly associated to the tribological 1601 with a step size of 0.021.
system formed between the cutter, the rock and the debris. Cobalt can exhibit three possible crystalline structures: hexa-
The intrinsic specific energy e is more linked to the rock mechanical gonal close packed a-Co phase, face-centered cubic b-Co phase
properties as compressive strength. These two coefficients and cubic also e-Co phase [13]. In addition to diamond, two
highlight the contribution of PDC and rock properties in cutter phases associated to cobalt element were identified on diffracto-
quality. Then, the ration u on k proves the compromise between grams: cobalt carbide (CoCx ) [14] and traces of cobalt (e-Co)
cutting efficiency Z and wear rate k in the quality formula. During (Fig. 9).
experiments, rock inhomogeneities and contact variations can As displayed on diffractogram of WC-Co part, the substrate
influence wear kinetic which affects the performance assessment contains only e-Co phase. During sintering, cobalt elements
of a cutter. Therefore, factor quality Q normalizes wear behavior migrated from WC-Co substrate towards PDC. This process
with regard to dissipated energy during the cutting process. explains the detection of e-Co phase in diamond table.
Eventually, Q is inversely proportional to LT which is fixed for The cobalt carbide phase is formed during sintering of PDC
all tested samples. This comment only expresses that for higher material. Thus, cobalt cannot be seen as a catalyzer of diamond
excavation distances, cutters need higher cutting capacity and formation but only as a precursor because it forms a new product
wear resistance to be qualified as good. That is why the quality with carbon. Actually, Akhaishi et al. [15] described the cobalt
factor depends on the targeted excavation distance. action during sintering as follows:

Table 3
Summary of mechanical results for samples A to F at LT ¼ 12 580 m.

PDC samples m e (106 J m3 ) k (109 mm3 N1 m1 ) u (km) Q (104 )

A 0.261 70.007 24.9 70.8 9.67 0.6 19.6 70.6 169 75.4
B 0.196 70.004 30.47 0.9 164.7 78.6 12.07 0.4 47 0.1
C 0.218 70.005 30.07 0.8 62.07 3.7 13.4 70.4 12 70.4
D 0.234 70.005 25.5 70.7 28.57 1.2 16.3 70.6 41 70.4
E 0.2307 0.011 28.4 71.5 16.67 1.1 14.2 70.5 55 72.0
F 0.2107 0.011 29.1 71.2 25.77 1.2 19.8 70.7 44 70.2
38 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41tpdel

 At first graphite, formed on diamond grain surfaces at high

temperature, dissolves into cobalt liquid phase above eutectic
150 until saturation.
 From saturation, diamond precipitates and dissolution can
Quality factor, Q (104)

start again until saturation of pure diamond into cobalt.

 Therefore, diamond grain size grows, densification increases
by Oswald ripening and cobalt carbide appears in grain
100
boundaries.

Like diamond, cobalt carbide is a metastable phase at room

temperature. For example, Ishida and Nishizawa [16] note that,
50 over a pressure of 4.5 GPa, Co3 C becomes a stable phase which
corresponds to a lower pressure than requires synthetic diamond
sintering conditions.
XRD peak intensity (Idiamond , ICoCx or Iwc ) depends on the
0 concentration of the identified crystallized phase. Also, diffracted
X-rays by diamond grains are modified by the amount of absorbent
A B C D E F cobalt phases surrounding them. Cobalt element has a higher atomic
number than carbon element and likewise absorption coefficient.
PDC sample Thus, X-rays penetrate deeper in leached samples than non-leached
Fig. 7. Quality factors results with LT ¼ 12 580 m. ones. This explains the higher intensities measured for leached
samples A, E and F. In addition, an interesting way to display relative
proportions of phases in the PDC material consists to calculate the
10
4 ratio between maximum peak intensity of the studied phase and the
8
Intensity (counts)

6 sum of maximum peak of all present phases (I) in the sample

4 (Table 4). This calculation only leads to relative results because the
A D phase attenuation coefficients are not taken into account [17].
2
As expected, the results obtained here show that the leaching process
3
10 eliminates the cobalt carbide phase. This reinforces the fact that, in
PDC table, the matrix surrounding diamond grains is essentially made
C of cobalt carbide phase.
F
Tungsten carbide phase was detected in the diamond table on
only samples B, C and D. Manufacturers well-know that WC grain
pollution, accidentally introduced during pressing and sintering,
E B can weaken diamond grain cohesion [18]. Sample B has clearly
the greatest proportion of WC and this can explain its low Q value
relatively to other samples.
Intensity (%)

100 Diamond WC CoCx

50

0
8. Residual stresses and quality factor
40 80 120 160 40 80 120 160
2·Theta 2·Theta Three-dimensional finite element analyzes were carried out to
evaluate post-sintering residual stresses. The diamond/carbide
Fig. 8. Diffractograms of PDC samples A, B, C, D, E and F classified from low Q
(sample B) to high Q (sample A) and calculated stick pattern of diamond, tungsten interface designs of samples B, C and D (i.e. also E and F) are
carbide and cobalt carbide. considered. All samples are supposed to be sintered at a tempera-
ture of 1380 1C under an isostatic pressure of 5.5 GPa [19]. Then, a
drop in the temperature until 1000 1C is taken into account and
pressure is maintained to pass the solidus of the carbide part.
100 Eventually, a last cooling is assumed until 20 1C at atmospheric
3 PDC part
40x10 WC-Co part
pressure. Physical properties used for numerical simulations are
identical (Table 5). To perform numerical simulations, the software
80
Phase Card intensity

35 Abaqus was used considering a steady state calculation and a

Diamond WC tetragonal mesh of 0.5 mm size.
30 Radial, axial and shear stresses are calculated at the end of a
(counts)

CoCx ε-Co 60
(%)

simulation. Negative values of stresses are related to compressive

25 state and positive ones to traction. Residual stresses are expressed
40
relatively to the maximum traction values in radial, axial or shear
20 directions obtained with a flat interface design. This approach
20
15
Table 4
Calculus of relative proportion of cobalt carbide and tungsten carbide in PDC
10 0 samples A to F.

45 50 55 60 65 70 75 PDC samples A B C D E F
2 Theta
ICoCx =I 3.5 70.9 7.9 70.7 8.1 7 0.6 11.4 7 0.9 0.9 7 0.9 0.7 70.4
Fig. 9. Cobalt allotropic phases: sample D diamond and WC-Co parts IWC =I – 16.4 70.9 6.8 7 0.7 8.3 7 1.1 – –
diffractograms.
 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41 39

Table 5
Physical properties of PDC and WC-Co [20].

Material Density Thermal conductivity Specific heat Thermal expansion coefficient Young’s modulus Poisson
(kg m3 ) (W m1 K1 ) (J kg
1
K1 ) (106 K1 ) (GPa) ratio

PDC 3510 543 790 2.5 890 0.07

WC-Co 15 000 100 230 5.2 579 0.22

Table 6
Residual stress distributions for flat interface design, samples B, C and D (i.e. also E and F).

Radial stress distribution Axial stress distribution Shear stress distribution

Flat

avoids numerical calibrations and considers only changes in Simulation results indicate that the flat interface between
stress distributions and amplitudes comparatively to the simple diamond and WC-Co shows the highest gradient of residual
flat interface (Table 6). stresses. At the interface, the radial stresses are compressive in
The high traction residual stresses favor cracks propagation the diamond and are symmetrically in traction in the substrate at
in grain boundaries and lead to diamond grain decohesion. respectively 100% and 100% by definition. In addition, the radial
The change in thermal expansion coefficient between WC-Co stresses reach more than 40% of the maximum traction value in
and diamond induces an overall cap-like radial stress distribution localized zones on the diamond front face. The axial stresses
in compression in the PDC table and in traction in the substrate. exceed 60% on the perimeter and in the middle of the PDC table.
This distribution was well illustrated by cap type cracks in the Maximum axial values are compressive in the edge of the carbide
diamond part in several microsections performed by electroero- part and tensile in the heart of this part. Most of the shear stresses
sion (Fig. 10a). The radial stresses produce compressive field lines is concentrated at the interface and reaches 60% (or symmetri-
in the PDC material (Fig. 10b) and the addition of traction axial cally  60%) near the cylindrical edge of the sample. These
stresses can cause cap-like crack propagation guided by these considerations justify the work of manufacturers in trying to
field lines. avoid traction field at the tip of a cutter by performing more
40 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41tpdel

PDC

WC-Co

Radial compressive stresses

Axial traction stresses

Shear traction stresses
Shear compressive stresses

Fig. 10. Cap type crack formation: (a) cap type cracks observed on a microsection realized by electroerosion; (b) illustration of radial, shear and axial stresses actions in the
PDC part.

complex interface designs. Indeed, all the studied designs modify Four factors influencing the quality of cutters have been
the distribution of residual stresses and reduce traction fields highlighted in this study:
compared to the flat interface.
The sample C reveals a diamond table with the highest  The diamond grain size distribution is the most sensitive
compressive state (around 40%) in the radial direction. parameter toward cutters abrasive resistance. As already
The substrate part counterbalances with high traction values expressed in previous studies and confirmed here, the finest
(higher than 60%) near the interface. As the other manufacturers’ diamond grains are the more abrasive resistant cutters.
designs, axial stresses are lightly in traction in the middle of the  The cobalt element is a precursor of diamond formation and
PDC part ( o 20%) and in compression until (up to  100%) at the not a catalyzer, because cobalt carbide remains after PDC
edge of the sample. High traction shear stresses were calculated sintering. The increase of cobalt content in the diamond part
( 4 60%) at the interface near the edge in the diamond. increases samples wear rates. However, depleting cobalt on
The samples B and D register a distribution of radial stresses few hundred micrometers has not a significant effect on
less contrasted than sample C. Traction does not exceed 60% and quality factor when long distance tests are performed.
compression is lower than  60%. Sample D has low radial traction  Low quality cutters contain tungsten carbide particles in
fields (o 20%) at the tip of the diamond part. The samples B and D diamond part.
have also high axial stress concentration zones (4 60%) well  Traction residual stresses in diamond promote cracks propa-
localized at the interface angularities or waves in the WC-Co part. gation and induce lower abrasion resistance by weakening
It is also interesting to note that sample B distributes traction shear grain boundary.
stress in the two materials with regards to undulating design.
Eventually, considering abrasion resistance, the sample C has
the most efficient design that permits a high compressive state at Acknowledgment
the sample tip compared to the other samples. This sample is
followed by the cutter B in terms of compressive values in This work was performed under the program ANR-09-MAPR-0009
diamond. However, as seen above, these cutters have the worst of Agence Nationale de la Recherche. We thank Armines Geosciences
quality factor results. Hence, residual stresses have to be considered laboratory for performing the wear experiments and the Varel Europe
with respect to cutters performances but diamond grain sizes have Company for providing cutters used in this study.
more significant impact on the relative quality of the six cutters. The
edge of sample A is based on the same interface as sample D but it
has a more complex design in the center. As sample B, the complex References
design along the interface permits to distribute residual stresses
in the three directions. This enhancement of interface design could [1] J.L. Wise, D.W. Raymond, C.H. Cooley, K. Bertagnolli, Effects of Design and
Processing Parameters on Performance of PDC Drag Cutters for Hard-Rock
have contributed to the high quality factor estimated for the Drilling, Technical Report, Sandia National Laboratories, 2002.
sample A. [2] A. Ersoy, M. Waller, Wear characteristics of PDC pin and hybrid core bits in
rock drilling, Wear 188 (1995) 150–165.
[3] A.K. Wojtanowicz, E. Kuru, Mathematical modeling of PDC bit drilling process
based on a single-cutter mechanics, Journal of Energy Resources Technology
115 (4) (1993) 247–256.
9. Conclusion [4] H.P. Bovenkerk, F.P. Bundy, H.T. Hall, H.M. Strong, R.H. Wentorf, Preparation
of diamond, Nature 184 (1959) 1094–1098.
[5] N.D. Griffin, P.R. Hughes, High volume density polycrystalline diamond with
A comparison among various PDC cutters submitted to wear tests working surfaces depleted of catalyzing material, United States patent US
requires an overlap of information. The PDC cutters assessment tends 6,562,462 B2 2003.
[6] J. Crank, The Mathematics of Diffusion, Oxford University Press, 1956.
to balance ability to withstand abrasive wear and to remain efficient
[7] S.-M. Hong, M. Akaishi, H. Handa, T. Osawa, S. Yahmaoka, Behaviour of cobalt
as long as possible. For this purpose, a quality factor was proposed as infiltration and abnormal grain growth during sintering of diamond on cobalt
the ratio of the cutting capacity on wear rate multiply by coefficients substrate, Journal of Materials Science 23 (1988) 3821–3826.
associates to the cutter/rock contact. To determine the cutting [8] D.A. Glowka, Development of a Method for Predicting the Performance and
Wear of PDC Drill Bits, Technical Report, Sandia National Laboratories, 1987.
capacity, an exponential law was fitted to the cutting efficiency [9] L. Gerbaud, S. Menand, H. Sellami, PDC bits: all comes from the cutter rock
decreasing with the excavation distance. interaction, in: IADC/SPE Drilling Conference, Miami, États-Unis, 2006, p. 1.
 M. Yahiaoui et al. / Wear 298–299 (2013) 32–41 41

[10] C. Fairhurst, W. Lacabanne, Some principles and developments in hard rock [16] K. Ishida, T. Nishizawa, The C–Co (carbon–cobalt) system, Journal of Phase
drilling techniques, in: Sixth Annual Drilling and Blasting Symposium, 1956, Equilibria 12 (1991) 417–424.
pp. 12–25. [17] R.P. Goehner, M.C. Nichols, X-ray powder diffraction-quantitative analysis,
[11] E. Detournay, P. Defourny, A phenomenological model for the drilling action in: ASM Handbook, ASM International, 1986, pp. 693–695.
of drag bits, International Journal of Rock Mechanics & Mining Sciences and [18] Z. Qi, The manufacture of PDC for cutting tools, Science and Technology of
Geomechanics 29 (1992) 13–23. New Diamond (1990) 415–416.
[12] F. Bellin, A. Dourfaye, W. King, M. Thigpen, The current state of PDC bit [19] T.N. Butcher, R.M. Horton, S.R. Jurewicz, S.E. Scott, R.H. Smith, Polycrystalline
technology, World Oil (2010) 41–46. diamond cutters having modified residual stresses, United States patent US
[13] V.A. de la Pena O’Shea, I.P.R. de Moreira, A. Roldan, F. Illas, Electronic and 6,220,375 B1, 2001.
magnetic structure of bulk cobalt: the alpha, beta, and epsilon-phases from [20] F. Chen, G. Xu, C. Ma, G. Xu, Thermal residual stress of polycrystalline
density functional theory calculations, The Journal of Chemical Physics 133 (2) diamond compacts, Transactions of Nonferrous Metals Society of China
(2010) 024701.
(2010) 227–232.
[14] A. Badzian, A. Klokocki, JCPDS-ICDD XRD card CoCx 00-044-0962, 1981.
[15] M. Akhaishi, H. Kanda, Y. Sato, N. Setaka, T. Ohsawa, O. Fukunaga, Sintering
behaviour of the diamond–cobalt system at high temperature and pressure,
Journal of Materials Science 17 (1982) 193–198.