Wear 256 (2004) 433–455

On the tribological behaviour of mechanical seal face

materials in dry line contact
Part II. Bulk ceramics, diamond and diamond-like carbon films
G.A. Jones∗
Aeronautical and Mechanical Engineering Department, Newton Building, University of Salford, Salford M5 4WT, UK
Received 4 September 2002; received in revised form 2 July 2003; accepted 2 July 2003

Abstract
The evolving tribological demands of mechanical face sealing applications such as those associated with high-duty fluid seals and gas
sealing technology are driving the need for the development of new and novel seal face material solutions. Under conditions where the
combination of high pressure and high surface speeds create high levels of PV the seal designer will specify two hard faces, typically a
combination of reaction bonded silicon carbide, alpha sintered silicon carbide or tungsten carbide. Work has been carried out here aimed at
evaluating and increasing the understanding of the tribological nature of these hard face combinations and evaluating the potential of poly-
crystalline diamond (PCD) and diamond-like carbon (DLC) films. Experimentation has shown that under dry moderate contact conditions
mild deformation-controlled wear takes place with a polishing of the contact surfaces. Progressively increasing the contact pressure was
shown to eventually produce a transition to severe microcrack-controlled wear. Coating one of the contact surfaces with a PCD or DLC film
modified the contact dynamics producing low friction coefficients and in the case of the PCD film shifted the wear transition to a much higher
contact severity. A linear elastic contact model is developed that provides a framework for rationalising the results of this work and enables
the key features of mechanical face seal design to be related to the critical tribological characteristics of the seal face material systems.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Polycrystalline diamond film; Diamond-like carbon film; Ceramics

1. Introduction 2. alpha sintered silicon carbide/reaction bonded silicon car-

bide;
In Part 1 the tribological behaviour of mechanical carbon 3. in special circumstances: reaction bonded silicon
in contact with alpha sintered and reaction bonded silicon carbide + graphite composite/alpha sintered silicon car-
carbide was examined, here in Part 2 the tribological be- bide.
haviour of alternative material combinations are considered.
These material combinations, however, are sensitive to
In applications where the sealed pressures and the spe-
the operating conditions, particularly to the lack of lubricant
cific face loads are high and when the sealed liquid con-
flow and dry running, which can cause premature failure of
tains, or can precipitate abrasive solids, the seal designer will
the seal. Nevertheless, the evolving tribological demands of
consider the use of two hard faces. The standard seal face
mechanical face sealing applications such as those associ-
materials are silicon carbide and mechanical carbon. Un-
ated with high-duty fluid seals and gas sealing technology,
fortunately the mechanical carbon materials have failings in
are leading to the increasing use of two hard faces. There
terms of their low resistance to mechanical deformation and
is now a clear and significant need for the development of
third body abrasive wear. In applications were these char-
new and novel seal face material solutions.
acteristics are limiting, the seal designer will specify one of
Work has been carried out here to evaluate the tribologi-
the following combinations:
cal character of the hard face combinations mentioned and
1. tungsten carbide/reaction bonded silicon carbide; evaluate the potential of PCD and DLC films. PCD and
DLC films have the properties and tribo-structural charac-
∗ Present address: Tribo-Tech Consultancy, Gledco Engineered Materials
ter to present significant potential in mechanical face seal
applications.
Ltd, Bankfield Terrace, Leeds, LS4 2JR, UK. Tel.: +44-113-2751144;
fax: +44-113-2304724. In this work standard a-Ti:C2 H2 DLC films deposited
E-mail address: gordon@gajones.fsbusiness.co.uk (G.A. Jones). on to reaction bonded silicon carbide using the closed field

0043-1648/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0043-1648(03)00540-4
434 G.A. Jones / Wear 256 (2004) 433–455

Nomenclature method adopted enabled the examination of the PV capa-

bilities and the associated tribo-characteristics of the mating
a bearing area half space; line contact combinations.
c radius of a semi-circular surface crack
E elastic modulus
E effective elastic modulus 2. Experimental
k dimensionless parameter
KI stress intensity factor; mode 1 2.1. Experiment 1: Hard ceramic combinations
KII stress intensity factor; mode 2
KIc stress intensity factor due to An Amsler friction and wear tester operating in the pad
contact pressure on cylinder configuration under dry running conditions was
KIeff effective critical stress intensity factor used (Fig. 1). Details of the composition and properties of
KIr stress intensity factor induced by the the materials used and details of the combinations exam-
residual stress distribution ined are presented in the first two tables in Appendix A.
KIC
f critical stress intensity factor in the The material combinations were evaluated in terms of PV
film; mode 1 profile.
KIC
s critical stress intensity factor in the The experimental procedure was that the cylinder was ro-
substrate; mode 1 tated at a constant surface velocity of 0.5 m/s. The load was
KIIC
I critical stress intensity factor at the increased in increments up to the point of contact instability.
interface; mode 2 The surface temperature of the interface was monitored us-
l length of bearing area ing a thin wire thermocouple bonded to the stationary pad,
P normal load, contact pressure 0.25 mm from the contacting interface. The humidity in the
R cylindrical radius general ambient environment was monitored and maintained
T temperature rise at around 55–60 RH. The mating surfaces of the cylinder
V surface velocity and the pad were prepared to give the following parameters:

Greek letters Ra (␮m) Rq (␮m) Sm (␮m)

α, β, ξ Dundur’s parameters
µ friction coefficient Cylinder 0.5–0.7 0.7–1.0 >100
ν Poisson’s ratio Pad <0.08 0.05–0.1 >250
σf fracture stress
σr residual stress The outer surface of the cylinder was precision machined
σt maximum Hertzian surface tensile stress (surface ground) so that it presented a line contact with the
mating pad, see Fig. 1. This surface presentation and the tex-
ture parameters were selected in order to idealise the contact
unbalanced magnetron sputtering (CFUBMS) combining conditions and permit subsequent analysis by modelling the
plasma assisted CVD and PCD films deposited on to the contact to that of a plane on a regular undulating surface.
same substrate type using a combination of the plasma jet In mechanical seals it is common practice for both the seal
and the microwave plasma assisted chemical vapour deposi- faces to be lapped. This presents two very smooth surfaces
tion (MPACVD) have been examined. Details of these pro- with the likelihood of elastic contact between contacting sur-
cesses and deposition techniques are described elsewhere face asperities. The ceramic materials and coating systems
[1]. These coating/substrate systems were evaluated in con- used in this work have very low values of E/H (0.02–0.10)
tact with alpha sintered silicon carbide and various mechan- and for very fine surface finishes there is the potential for the
ical carbon materials, of varying degrees of graphitic order. contact to remain elastic up to very high contact pressures.
The experimental rationale adopted simulates the tribol- The rationale for leaving the outer surface of the cylinder
ogy of the sealing interface without the difficulty of complex in these experiments in the precision machined condition
design interactions of a mechanical face seal arrangement. is to facilitate the transition from elastic contact to plastic
In actual operation seal faces are not parallel, they deviate contact at lower contact pressures. It was decided that the
in terms of their macro-geometry in both the radial and tan- essential objectives of this work could be met by adopting
gential directions. The tangential deviation produces a cir- this approach.
cumferential waviness contact profile at the sealing inter- At least two test runs per material combination were car-
face and the contact at the peaks of the waves can, to a first ried out. Each material combination provided a PV profile,
approximation, be likened to a line contact geometry. Ac- which included a PV limit above which stable conditions
cepting this analogy, a pad on cylinder geometry was used. could not be sustained. The PV was calculated by relating
This presented line contact and permitted the tribo-dynamics the contact area to the normal load supported and the sur-
of the contact materials to be evaluated. The experimental face velocity of the rotating cylinder. The contact area, for a
 G.A. Jones / Wear 256 (2004) 433–455 435

Fig. 1. (a) Amsler friction and wear testing machine. (b) Test specimen arrangement, showing pad and cylinder geometry and the jig for holding the test
pads such that line contact is presented.

cylinder on a plane, was calculated from the Hertzian equa- 2.2. Experiment 2: Combinations involving PCD
tion for semi-width, a, such that the contact area = 2al, and DLC films
where l is the length of the cylinder, and the contact pres-
sure was derived by dividing the normal load by the calcu- Again a PV profiling experiment using the Amsler fric-
lated contact area. Any warping of the contact area resulting tion and wear tester in the pad on cylinder configuration
from tangential tractions was neglected. Also information under dry running conditions was used. Reaction bonded
was obtained on the wear process, friction coefficient, and silicon carbide pads to the specification in the first table
frictional heat generated. of Appendix A were used as substrates. Rotating cylinders
436 G.A. Jones / Wear 256 (2004) 433–455

were manufactured from the required mating material, again efficient increased with increasing load. Fig. 4 shows, for
to the relevant specification given in the table. these three hard face combinations, the temperature gener-
The procedure detailed in experiment 1 was repeated and ated, to an accuracy of ±0.1 ◦ C, on the stationary pad as a
the material system combinations examined are given in function of the PV value. It can be seen that the general rule
second table of Appendix A. The outer surface of the cylin- is that the frictional temperature increases with increasing
der was precision machined so that it presented a line con- PV. The reaction bonded SiC + graphite/alpha sintered SiC
tact with the mating pad. The PCD film on the pad surface combination produced less frictional heat, for a given load
was left in the ‘as-deposited’ condition. Examination of the and PV, when compared to that produced by the other two
available polishing techniques for PCD diamond films were hard face combinations. This shows the value that could be
shown to be cost prohibitive, time-consuming and poten- gained from the addition of a residual amount of graphite to
tially ineffective. For this work it was decided that the essen- one of the hard face materials.
tial objectives could be met by utilising the ‘as-deposited’
surface finish. The starting surface texture parameters of the 3.1.2. The wear process
pads and cylinders are detailed in Appendix A (fourth table). The nature of the PV profiling experiment meant that the
Two test runs per material combination were carried out. wear rate was very difficult to measure and quantifiable data
Each test run provided information on PV performance. was not collected. However, and more importantly for this
Each material combination provided a PV profile includ- work, the experimental observations of the wear processes
ing a PV limit above which stable conditions could not be are detailed below.
sustained. The PV was again calculated as in experiment The wear process observed for these hard ceramic com-
1. Also, information was obtained on friction coefficients, binations was typical of that associated with the wear of
wear rates, and frictional heat generated. ceramic materials. Relatively moderate contact conditions
produced very mild wear with the production of very finely
divided wear debris at a relatively low rate which became
3. Results agglomerated into patchy surface layers, a smooth polishing
of the contact surfaces and a steady frictional trace. Increas-
For the purposes of clarity the critical data collected from ing the PV, up to some critical value, caused a transition
both experiments 1 and 2 are presented in tabular form in to a more severe wear regime. Under these conditions the
Appendix A. ceramic materials showed a higher wear rate dominated by
brittle fracture, a coarser contact surface topography with lit-
3.1. Experiment 1: Hard ceramic combinations tle evidence of surface layer development and a fluctuating
frictional trace. Examination revealed that sliding between
3.1.1. General observations the two contacting surfaces at the point of wear regime tran-
The photomicrographs in Fig. 2 show the general mi- sition generated, from their surfaces, material in the form of
crostructural morphology of the subject materials. To control relatively large particles (>1 ␮m). These particles then be-
the material quality, samples of material from the batches came comminuted within the contact zone by repeated de-
from which they were taken, were examined metallograph- formation and fracture, and further contributed to the wear
ically to confirm quality and microstructural character. The process by introducing a third body abrasive component.
surface conditions of the mating materials before and af- In reaction bonded SiC + graphite/alpha sintered SiC
ter each test run are shown in Appendix A. In all tests, the combination the graphite in the former changed the wear
wear patterns exhibited on both of the mating faces had sim- generation process significantly. Under moderate contact
ilar features. Wear debris retained within the contact zone is conditions, the graphite formed a protective contact layer
comminuted by repeated plastic deformation and fracture, and at low loads this considerably influenced the frictional
and contributes to the wear process by introducing a third behaviour. The friction coefficient at loads of 49.05 N was
body abrasive component. Fig. 3 shows the friction coeffi- of the order of 0.1 compared to 0.58 and 0.65 for the tung-
cient, measured to an accuracy of ±0.01, as a function of sten carbide/reaction bonded SiC and the reaction bonded
increasing load for the combinations examined. The same SiC/alpha sintered SiC combinations, respectively. This ef-
trend as that identified for mechanical carbon materials [1], fect, however, diminished very quickly as the normal load
of reducing friction coefficient with increasing load, can be was increased above 58.86 N and the wear and frictional
seen for the tungsten carbide/reaction bonded SiC and the behaviour then started to resemble that of SiC on SiC.
reaction bonded SiC/alpha sintered SiC combinations. For
reaction bonded SiC + graphite/alpha sintered SiC combi- 3.2. Experiment 2: Combinations involving PCD
nation this trend was not fully observed. From the low PV and DLC films
value at the start of the test, increasing the PV produced a
steady reduction in friction coefficient. This continued until 3.2.1. Control of film material and microstructural quality
a minimum value was reached. Increasing the load beyond The control of film material and microstructural qual-
this point produced a reverse in trend and the friction co- ity was difficult. The supplier of the DLC films assessed
 G.A. Jones / Wear 256 (2004) 433–455 437

Fig. 2. Photomicrographs of the general microstructure of (a) reaction bonded silicon carbide, (b) alpha sintered silicon carbide, (c) graphite loaded
reaction bonded silicon carbide and (d) Co bonded tungsten carbide.
438 G.A. Jones / Wear 256 (2004) 433–455

Fig. 2. (Continued ).

film adhesion along with coating hardness as a matter any of the films evaluated that would bias the results
of quality control. These films were quoted to have crit- of these experiments. The quality of the substrates and
ical loads of >55 N measured on a Teer st200 scratch counterface materials was evaluated and controlled in
tester and hardnesses in the range 2100/2500 Hv. The the same manner as described in experiment 1. Sam-
supplier of the PCD films did not have the facility to of- ples of material from the batches from which they were
fer this kind of information. The films, however, were taken, were examined metallographically to confirm the
microscopically examined for consistency. The photomi- quality and microstructural character. Again the pho-
crographs in Fig. 5 show their general appearance. There tomicrographs in Fig. 2 show the general microstructural
were no obvious macroscopic or microscopic defects in morphology.

2.000

1.800

1.600
FRICTION COEFFICIENT

1.400

1.200

1.000
2
0.800

0.600
3
0.400 1
0.200

0.000
0 20 40 60 80 100 120 140 160 180
NORMAL LOAD (N)

Fig. 3. The relationship between the friction coefficient and the applied normal load for (1) reaction bonded SiC in contact with reaction bonded
SiC + graphite, (2) reaction bonded SiC in contact with alpha sintered SiC and (3) reaction bonded SiC in contact with tungsten carbide + 6% Co.
 G.A. Jones / Wear 256 (2004) 433–455 439

100

90
1

80

70

60
∆T ( C)
o

50
2
3
40

30

20

10

0
0 20 40 60 80 100 120 140
PV VALUE (MPa.m/s)

Fig. 4. The relationship between the frictional heat generated T (±0.1 ◦ C) and the PV value for (1) reaction bonded SiC in contact with reaction bonded
SiC + graphite, (2) reaction bonded SiC in contact with alpha sintered SiC and (3) reaction bonded SiC in contact with tungsten carbide + 6% Co.

3.2.2. PCD and DLC films in contact with alpha sintered the rise is 6.5× greater for the DLC/silicon carbide com-
silicon carbide bination than it is for the PCD/silicon carbide combination
and the final temperature at test termination is also greater
3.2.2.1. General observations. The surface condition of by 50%.
the mating materials before and after each test run is de-
scribed in Appendix A. The final condition of the contact 3.2.2.2. The wear process. In both the cases examined
surfaces is typical of film delamination. Wear debris retained here the point of contact instability was indicated by signif-
within the contact zone is comminuted by repeated plastic icant local disintegration or delamination of the deposited
deformation and fracture, and this contributes to the wear diamond or diamond-like carbon film. This disintegra-
and delamination process by introducing a third body abra- tion/delamination was accompanied by a significant in-
sive component. Fig. 6 shows the friction coefficient (±0.01) crease in frictional torque, frictional noise and frictional
as a function of increasing PV for the combinations exam- heat. The condition of the surface of the cylinder after the
ined. The same trend, as identified in [1] for mechanical film had started to disintegrate was very disturbed, having
carbon contacts and in experiment 1 for contact between two a coarse surface texture. In the case of the PCD/silicon
hard surfaces, of reducing friction coefficient with increasing carbide combination, there was evidence of relatively large
PV can be seen for the PCD/silicon carbide combination. For tracts of material having been removed by ploughing, tear-
the DLC/silicon carbide combination this trend was not fully ing or cutting. The edges of the cylinder in this case, which
observed. Increasing the PV from the low value set at the constitute the edges of the contact area, had also become
start of the test produced a rapid reduction in friction coeffi- heavily chipped. A closer examination of the surface of the
cient, falling to a minimum value at a PV of 46.52 MPa m/s. PCD film and its silicon carbide counterface just prior to
Increasing the PV beyond this point produced a reverse in test termination revealed that the film had already begun
this trend and the friction coefficient increased rapidly up to to fail. On both surfaces the presence of Hertzian ring and
a maximum value at a PV of 73.56 MPa m/s. At this max- cone cracks were observed within the wear track.
imum value the friction coefficient appears to stabilise and Again the nature of the PV profiling experiment meant that
the values recorded are similar to those recorded for SiC on the wear rate was very difficult to measure and quantifiable
SiC. This difference in PV capability is further borne out in data was not collected. The experimental observations of the
Fig. 7, which illustrates the frictional temperature generated wear and film failure processes, however, are detailed.
(±0.1 ◦ C) on the stationary pad as a function of the PV Microscopic examination of worn surfaces from both
value developed. It can be seen that in both instances the combinations shows that the deposited PCD or DLC film
friction temperature increases with increasing PV. However, wears via two predominant modes;
440 G.A. Jones / Wear 256 (2004) 433–455

Fig. 5. Photomicrographs of the general contact surface structure of the (a) DLC and (b) PCD films.

• delamination before they are worn through, due to inad- sive component. As the load was progressively increased,
equate adhesion; some critical point was reached, determined by the mechan-
• gradual wear (slow or fast) without delamination. ical, chemical and tribological characteristics of the coat-
ing substrate system, and the wear mechanisms produced
The former is self-explanatory and generally signifies gross film delamination. The significance of this process is
gross film failure. In the latter, the process appears to be shown when comparison is made between the performance
analogous to micro-polishing, which takes place under sta- of the PCD and DLC films. The DLC film was unable to
ble sliding conditions. In these experiments the latter wear sustain stable operation when the normal load extended be-
mechanism dominated in the early stages, when the con- yond approximately 19.62 N. The film failed at this point
tact conditions were not so severe. Under these moderate due, mainly, to gross film delamination. In contrast the PCD
conditions wear debris, formed from microfragmentation at film offers almost a 25-fold improvement in terms of PV ca-
discrete locations within the contact zone, was further com- pability. The photomicrographs in Fig. 8 show the polishing
minuted by repeated deformation and fracture and was ei- of the tips of the crystals in the film and the delamination
ther expelled as loose debris or filled the interstices within that takes place under severe contact conditions.
the surface topography of the wear track. This comminu- A closer examination of the wear tracks on the PCD/
tion and discharge of the wear debris contributed to the silicon carbide combination revealed the presence of par-
wear/polishing process by introducing a fine third body abra- tial Hertzian ring and cone cracks on both surfaces. The
 G.A. Jones / Wear 256 (2004) 433–455 441

1.6

1
1.4

1.2
FRICTION COEFFICIENT

0.8

0.6

0.4

0.2
2

0
0 50 100 150 200 250 300
PV VALUE (MPa.m/s)

Fig. 6. The relationship between the friction coefficient and PV value for (1) alpha sintered SiC in contact with reaction bonded SiC + DLC film and
(2) alpha sintered SiC in contact with reaction bonded SiC + PCD film.

40

35

30

25

2
∆T ( C)
o

20

15

10

0
0 50 100 150 200 250 300
PV VALUE (MPa.m/s)

Fig. 7. The relationship between the frictional heat generated T (±0.1 ◦ C) and the PV value for (1) alpha sintered SiC in contact with reaction bonded
SiC + DLC film and (2) alpha sintered SiC in contact with reaction bonded SiC + PCD film.
442 G.A. Jones / Wear 256 (2004) 433–455

Fig. 8. Photomicrographs (a)–(e), showing the progression of the polishing of the crystal tips of the PCD film into surface defects and consequent
propagation of Hertzian cone cracks and (f) final film delamination.
G.A. Jones / Wear 256 (2004) 433–455 443

Fig. 8. (Continued ).
444 G.A. Jones / Wear 256 (2004) 433–455

Fig. 9. Photomicrographs (a)–(d) showing the presence of Hertzian cone cracks on the PCD film. Parts (c) and (d) show Hertzian cracks at the edge of
film delamination.
 G.A. Jones / Wear 256 (2004) 433–455 445

Fig. 10. Photomicrographs of the carbon graphite contact film failure, viz. blistering of the C–G film on the ((a) and (b)) DLC and (c) PCD film.
446 G.A. Jones / Wear 256 (2004) 433–455

photomicrographs in Fig. 9 clearly show these half ring type materials. The PV profiling identified a classic transition
fractures. Notwithstanding good adhesion and high substrate between mild and severe wear. The transition was sudden
yield strength, tangential sliding under high friction condi- and can be interpreted as a change in wear mechanism from
tions can facilitate and/or accelerate the generation of a dy- deformation-controlled wear to microcrack-controlled wear.
namic series of Hertzian indentation-induced ring and cone At some critical load, the localised stresses within the con-
cracks. It appears from these results that this is the dominant tact zone were sufficiently high to cause microcracking and
mechanism determining film capability under the test con- relatively large particle pull out. The wear debris produced
ditions employed. The alpha sintered silicon carbide coun- caused further stress concentrations and surface roughen-
terface, up to the point of film failure was subjected to a ing which, in turn, helped to maintain the high stresses and
mild wearing regime with the production of very finely di- the high rate of wear. At low and moderate loads and with
vided wear debris at a relatively low rate which, as in ex- the surfaces relatively smooth, the wear process was one of
periment 1, became agglomerated into patchy surface layers gradual attrition of the surface asperities rather than com-
and a smooth polishing of the silicon carbide contact sur- plete detachment of large particles.
face. The eventual failure and removal of the film from the The critical load that propagates the transition to
contact zone presented contact between the counterface and microcrack-controlled wear can be further interpreted as
the substrate. The contact stresses then developed a transi- the load that generates surface tensile stresses sufficient to
tion to a more severe wearing regime. Under these condi- induce surface microcracks. The purpose of depositing a
tions the silicon carbide counterface showed a higher wear PCD or DLC film on to one of the hard ceramic surfaces is
rate dominated by brittle fracture and with features as per to provide a third phase between two hard contacting faces
experiment 1. that extends the mild wearing regime. In this work the PCD
films achieved this objective. The DLC films, however,
3.2.3. PCD and DLC films in contact with mechanical were not so successful. Figs. 11 and 12 and fifth table in
carbon Appendix A together provide a comparison of the tribo-
The surface conditions of the mating materials before and logical (PV) performance of the combinations examined
after each test run are shown in Appendix A. The wear pat- in experiment 1, those incorporating a PCD or DLC film
terns exhibited on both of the mating faces had similar fea- examined in experiment 2 and the seal face combination
tures for all PCD/mechanical carbon and DLC/mechanical examined in [1]. The PCD film on reaction bonded silicon
carbon contacts. In all cases, both the bearing surfaces ex- carbide in contact with alpha sintered silicon carbide offered
hibited evidence of carbon graphite contact film formation. a significant improvement in performance over that given
The stability of operation was shown to be a function of by the hard ceramic combinations. It also offered favourable
the balance between contact film generation and disintegra- results in comparison with the seal face combination of
tion, as described in [1]. The contact film failed in a seizure SiC/mechanical carbon. It appears from these results that
type of mechanism, and the photomicrographs in Fig. 10 il- the PCD film produces low friction and low traction stresses
lustrate this. Microscopic examination of the PCD and the for a given contact severity. The Hertzian contact pressures
DLC surfaces showed evidence that the deposited diamond necessary to raise the surface tensile stresses to a critical
or diamond-like film had delaminated in large areas. There value, as identified by the onset of film delamination or
was also evidence of a more gradual wear process taking the transition to a microcrack dominated wear regime on
place in which material has been removed from the PCD or the counterface, are significantly higher than those experi-
DLC film without interfacial delamination. enced with plain hard ceramic contacts and for the contact
The profiles of friction coefficient as function of PV were incorporating the DLC film.
similar to those reported in [1] for contacts between the reac- To further rationalise the results and observations of this
tion bonded silicon carbide and mechanical carbon. Fig. 11 work the following contact model is proposed. Hamilton
and fifth table in Appendix A detail this situation. Fig. 12 [2] and Hamilton and Goodman [3] considered the stresses
shows the temperature generated (±0.1 ◦ C) on the stationary associated with a sphere sliding over a semi-infinite half
pad as a function of the applied normal load. These results space and the greatest tensile stress σ t was shown to be at
show clearly that there is nothing significant to be gained the trailing edge of the contact spot. In the following model
by running DLC or PCD films against a mechanical carbon. their stress field analysis for a Hertzian sliding contact is
The factors governing the tribo-performance are determined applied with a modification to account for line contact.
and dominated by the mechanical carbon material and the
substrate material. 4.1. Contact model

The analysis by Hamilton had the following assumptions:

4. Discussion (1) there is a uniform friction coefficient over the contact
surface and (2) the contacting bodies behave like purely
The friction and wear of the hard ceramic combinations elastic and elastic/brittle solids. These assumptions apply to
examined here showed the characteristics typical of ceramic the hard combinations examined here. The stress σ x in the
 G.A. Jones / Wear 256 (2004) 433–455 447

2.000

1.800

1.600
FRICTION COEFFICIENT

1.400

1.200

1.000

0.800 3

0.600 2

0.400

0.200 1
0.000
0 50 100 150 200 250 300
(a) PV VALUES (MPa.m/s)

1.600

1.400

1.200
FRICTION COEFFICIENT

1.000

0.800

0.600 3

0.400

2
0.200 1

0.000
0 100 200 300 400 500 600
(b) NORMAL LOAD (N)

Fig. 11. (a) The relationship between the friction coefficient and PV value for (1) alpha sintered SiC in contact with reaction bonded SiC + PCD film,
(2) reaction bonded SiC in contact with tungsten carbide 6% Co and (3) reaction bonded SiC in contact with alpha sintered SiC. (b) The relationship
between the friction coefficient and the applied normal load permitting comparison in terms of the normal load required to produce contact stresses
necessary for transition to severe friction and wear: (1) alpha sintered SiC in contact with reaction bonded SiC + PCD film; (2) mechanical carbon grade
GEM 601 in contact with alpha sintered SiC [1]; (3) reaction bonded SiC in contact with tungsten carbide + 6% Co.

plane parallel to the direction of sliding, i.e. at right angles stress −2µP0 at the leading edge of the contact area (x =
to the axis of the cylinder, is given, after [4], by −a) and a maximum tensile stress 2µP0 at the trailing edge
 x (x = a). Thus,

 −2µP0 , |x| ≤ a

 a
 4µP
σx =  2 1/2  (1) σt = (2)

 x x πal

 −2µP 0 ± − 1 , |x| > a
a a2
and from Hertz:
and P0 = 2P/πal the peak surface contact stress.  
4PR 1/2
The tangential traction acting on the moving plane is neg- a= (3)
ative and the direct stress reaches a maximum compressive πE l
448 G.A. Jones / Wear 256 (2004) 433–455

70.0

60.0 3
4
50.0

2
40.0
∆T ( C)
o

30.0

20.0 1

10.0

0.0
0 100 200 300 400 500 600
NORMAL LOAD (N)

Fig. 12. The relationship between the frictional heat generated T (±0.1 ◦ C) and the applied normal load for (1) alpha sintered SiC in contact with
reaction bonded SiC + PCD film, (2) alpha sintered SiC in contact with reaction bonded SiC + DLC film, (3) mechanical carbon grade GEM 601 in
contact with alpha sintered SiC [1] and (4) reaction bonded SiC in contact with tungsten carbide + 6% Co.

 1/2
4PE a rationalisation of this work. Fig. 13 shows the contact
σt = µ (4)
πRl condition rationalised to a two-dimensional problem with
the coated half space subjected to a distribution of trac-
When σ t exceeds some critical stress σ c a transition from tion q(x) = µp(x), acting in the direction of sliding. From
mild to severe wear is produced. Now applying the rules of Eq. (5) the transition from mild deformation-controlled
linear fracture mechanics, the conditions for transition can wear to severe microcrack-controlled wear occurs when
be estimated from a semi-circular surface flaw loaded by an KI ≥ KIC . In the case of DLC and PCD films contact failure
external uniform tensile force and the stress intensity factor is very similar to this and in keeping with the analysis above
KI can be derived from and Hamilton [2] the following assumptions are now made:
KI = 0.713σt (πc)1/2 (5) (1) The friction coefficient is uniform over the contact sur-
face.
where c is the radius of the semi-circular surface crack.
(2) The film behaves as an elastic/brittle layer.
The transition to severe wear through intergranular micro-
cracking will proceed if the components of σ t increase the
stress intensity factor KI to the critical value KIC , the fracture Cylinder
toughness of the material. From Eqs. (2) and (5) it is clear R
that σ t is controlled by the operating conditions including Ec, νc
po
the environment within the contact zone, the composition of qo
X1
the ceramic materials, their elastic properties, and the geom-
2a c
etry of the contact arrangement. Of the components making Surface Film
up σ t the friction coefficient µ is of greatest significance. As h
c c Ef, νf
µ decreases due to changes in tribo-chemical reactions so
the contact pressure to elevate KI to the critical value KIC , Es,νs
and cause the transition to severe wear, will increase. This
X2 Substrate
is the primary objective of depositing a PCD film or DLC
film on to one of the contacting faces.
In applying this model to thin film/hard counterface com-
binations some subtle modifications are required to account Fig. 13. A coated half space, with flaws [13]. For simplicity the contact
for the influence of the thin film on one of the contacting condition is rationalised to this two-dimensional problem. The coated
half space is under sliding contact load with a cylindrical counterface of
surfaces. In the available literature there are a number of
radius R. The cylindrical counterface is elastic and the friction coefficient
excellent studies detailing the contact condition of surface is assumed to act between the sliding cylinder and the surface of the film.
coating systems [5–15]. The reader is directed to Oliveira The half space is subjected to a distribution of traction q(x) = µp(x)
and Bower [13] for a complete analysis. The following is acting in the direction of sliding.
 G.A. Jones / Wear 256 (2004) 433–455 449

(3) The silicon carbide cylinder behaves as a purely elastic 4.2. Use of the model
solid.
(4) The film, substrate and interface contain microcracks. The model is, clearly, an effective tool for describing and
(5) There is residual stress within the film, originating from clarifying the tribo-performance of the hard ceramic com-
the mismatch in elastic and thermal properties. binations examined in this work in dry line contact. The
model can, also, be related to mechanical seals were there
Failure of the thin film/substrate system under contact
are a number of adjacent contact zones within the sealing
loading will occur when one of the following fracture criteria
interface which support the axial closing force. The seal de-
is met:
signer, from this model, can evaluate the impact on seal face
• KI ≥ KIC
f for cohesive failure in the film; performance of the design criteria that affect the axial clos-
• KI ≥ KIC
s for cohesive failure in the substrate; ing force and the physical, surface and tribological proper-
• KII ≥ KIIC
I for failure at the interface and film delamina- ties of the seal face materials. In Eq. (4) there is the prospect
tion. of evaluating the influence of seal face shape in the form of
waviness amplitude and face width. Eq. (2) calculates the
In the case here, the toughness of the interface is quanti- maximum tensile stress σ t generated at the rear of the con-
fied by using stress intensity mode II (KIIC I ). The influence
tact zone, at the point of wear regime transition, to be around
of cracks at the interface is very complex and a combined in- 400 MPa for both the tungsten carbide/reaction bonded SiC
terfacial stress intensity factor could be proposed, i.e. Ki = and reaction bonded SiC/alpha sintered SiC material com-
(KI 2 + KI 2 )1/2 . However, as suggested by Oliveira and binations. This value is similar to the values quoted in the
IC IIC
Bower [13], O’Dowd et al. [14] and Wang and Suo [15], un- supplier’s literature for the bulk strengths of both reaction
der contact loading KIIC I would be the most dominant mech-
bonded and alpha sintered silicon carbide. It is therefore rea-
anism, i.e. KIICI K I . The mode II fracture toughness of
IC sonable to assume that in the discrete contact zones within
an interface, however, can be expected to be high and in the sealing interface the wear rate transition associated with
most practical circumstances, only small values of KII are σ t being of the order of 400 MPa will be realised. Eqs. (5)
induced at the tips of interface cracks by contact loading. and (6) rationalise these arguments in terms of linear frac-
Therefore, it is postulated here that the most likely cause ture mechanics. For reaction bonded SiC + graphite/alpha
of film failure and delamination is through cracks initiated sintered SiC combination the recorded σ t at transition is re-
at the contact surface, which subsequently propagate down duced by a 25% and is approximately 300 MPa. This reduc-
to the interface. It is, also, proposed that failure in the sub- tion arises because the pockets left in the surface structure
strate is the least likely of the criteria listed. Thus, an ef- by the removal of the graphite domains constitute surface
fective critical stress intensity factor KIeff can be derived by defects that act to develop KI ≥ KIC at lower values of σ t .
modification of Eq. (5) and is given by Also this transition is in part a measure of the effect of the
graphite contact film formed from the graphite domains.
KIeff = KIc + KIr = 0.713σt (πc)1/2 + k(α, β, ξ)σr (πc)1/2 The friction coefficients generated by this combination at
(6) low loads were significantly less than those recorded for the
other hard ceramic combinations. This produced in the early
where KIr is the stress intensity factors induced by the resid- stages of contact very low σ t but once the graphite contact
ual stress distribution, KIc the stress intensity factor due to film was removed from the contact zone the friction coeffi-
contact pressure, (α, β, ξ) the parameters describing elas- cient increased, σ t increased and in consequence KI ≥ KIC .
tic and thermal mismatch, respectively (α, β are Dundur’s The structure and derivation of the model allows an eval-
parameters for mismatch between elastic properties), and k uation of critical flaw size. Accepting that the σ t at the point
the dimensionless parameter, independent of α and β. of transition is representative of the fracture strength of the
Surface fracture (microcrack-controlled wear) of the film material, the radius of a surface flaw necessary for the prop-
occurs when the combined stress intensity factors due to agation of a ring crack can be evaluated. From the model the
residual stresses and the contact stress reach the fracture critical flaw diameter for both reaction bonded silicon car-
toughness of the film, i.e. KIeff ≥ KICf . If the residual stress bide and alpha sintered silicon carbide at transition is calcu-
component is, for the purposes of analysis, considered con- lated to be around 35 ␮m. Fig. 14 plots, from Griffiths, the
stant for the film/substrate systems examined the failure of fracture strength as a function of defect size for both silicon
the system (transition to microcrack-controlled wear) can be carbide types. From this it can be seen that the model pre-
said to occur when the contact stresses develop σ t such that dicts an acceptable and appropriate flaw size, thus the model
KIc tends KIeff ≥ KICf . can be used to evaluate the effect inherent flaw size on the
The model for the failure of the thin film/hard counter- tribo-performance produced.
face combination (transition to microcrack-controlled wear With a small amount of information, i.e. friction coef-
of one of the contacting surfaces) can now be said to occur ficient, fracture strength, fracture toughness and with an
when this criteria KIeff ≥ KICf is met or when K c ≥ K for
I IC understanding of the tribo-chemistry of the contact, this
the counterface is met. model enables an assessment of the tribo-performance of
450 G.A. Jones / Wear 256 (2004) 433–455

1600.00

Z = 1.50
Y = 2.00
1400.00 E = 360 GPa
2
γ = 25 J/M

1200.00
FRACTURE STRENGTH, σ f, (MPa)

1000.00

800.00

600.00

400.00

200.00

0.00
0 50 100 150 200 250
DEFECT SIZE, C, (µM)

Fig. 14. The influence of a planar elliptical crack on the fracture strength of silicon carbide.

hard ceramic combinations under dry running conditions to load of only 19.62 N the contact stresses developed σ t such
be made. The hard ceramic combinations examined in this that KIeff ≥ KICf .

work all tend to form oxide films on their sliding surfaces. The contact stresses at the point of failure for the PCD and
These films can constitute a third phase of low shear strength DLC films and also for the point of wear transition for the
that modifies the friction condition at the interface result- hard ceramic combinations are shown in Appendix A (sixth
ing in a reduced friction coefficient. Contact conditions that table). It can be seen that for both PCD and DLC films the
suppress these tribo-mechanisms will result in increases in failure stress σ t was approximately the same ∼90–106 MPa,
σ t , which will activate critical defects, develop KI ≥ KIC though the PV value at failure was significantly greater for
and lead to a transition to severe wear at lower severities of the PCD film. Also, the table, and of course the model,
contact. directly illustrates the potential of PCD films in dry tribo-
In applying the model to thin film/hard ceramic combina- logical applications where two hard contact faces are re-
tions the influence of the contact stresses superimposed onto quired. Introducing a PCD film between two silicon carbide
the residual stresses, derived from elastic and thermal mis- contacting faces significantly extends the PV capability by
match between the film and the substrate, can be evaluated. modifying the contact stresses generated such that the condi-
The model clearly shows the relationship between contact tion KIeff ≥ KICf or K c ≥ K is met at much higher contact
I IC
stresses and residual stresses and their combined impact of severities. The low friction coefficient developed by the PCD
on the tribological performance. The model in this context film is clearly very influential. Jahanmir et al. [16] postulated
presents a framework for evaluating the potential of thin that the low friction coefficient of diamond was related to the
films and thin film/substrate systems in various tribological formation of a thin film of graphite at the real area of con-
applications. In experiment 2, the PCD film showed that it tact. Miyoshi et al. [17] showed that, in vacuum the friction
could sustain stable operation up to PV values similar to coefficient of CVD diamond films increases from 0.02/0.04
those of mechanical carbon in contact with silicon carbide. to 1.5/1.8. They postulate that the vacuum removes some
This was primarily because of the very low friction coeffi- contaminant surface film from the contact area resulting in
cient developed. The DLC film, however, did not perform stronger interfacial adhesion in the contact zone giving rise
quite so well. This was because even at low loads the fric- to higher friction coefficients. Bowden and Hanwell [18]
tion coefficient was relatively high and at a normal applied and Dugger et al. [19] also attributed the increase in friction
 G.A. Jones / Wear 256 (2004) 433–455 451

under vacuum to the removal of adsorbed contaminants. sought in order to raise confidence and identify design direc-
Bowden and Young [20] presented data detailing the for- tions. The contact model developed here aims to contribute
mation of graphite on the contact surfaces of XTL dia- to this understanding. It demonstrates to both the seal de-
mond/XTL diamond contact in vacuum and at 1100 ◦ C. Ta- signer and the materials scientist the mechanisms and com-
bor [21] related the tribological behaviour of XTL diamond ponents governing the tribo-performance of these systems.
to surface chemistry-dependent adhesion. He suggested the Importantly the model enables the key features of mechan-
possibility of broken surface bonds. Other contemporary ical face seal design to be related to the critical tribological
works have reported similar effects making a connection characteristics of the seal face material systems. This is an
between dangling bonds and high friction. The common important connection for two reasons:
hypothesis appears to be that, notwithstanding the impor-
1. The future development of mechanical seal face material
tance of surface roughness and facet specific frictional
technology hinges on a greater understanding and subse-
anisotropy [22], the low frictional character of diamond and
quent development of the tribo-mechanisms that promote
diamond films is determined by a surface chemistry effect.
low wear and low frictional forces within the sealing in-
The model developed here lends further support to this
terface.
hypothesis but also relates it to the surface contact stresses
2. The relatively slow deposition rate and the relatively
and the consequent tribological performance. The notion of
small area of coverage of particularly the PCD type films
graphitisation of the contact surface or the formation of a
will require a significant research and development re-
hydrocarbon layer, onto which is adsorbed other adsorbates
source. There is an imperative, therefore, that a full and
offers a plausible explanation for the performance given
detailed understanding of the material systems, their rel-
here by the PCD film/hard ceramic combination. The model
evance and performance potential be developed.
however has provided a tribological context and framework
in which to evaluate and assess this explanation.

4.3. Relevance of the contact model to mechanical face 5. Conclusions

seal design
The evolving tribological demands of mechanical face
For dry running applications the seal face designer, seal applications and the requirement for the development of
whenever possible, will specify a mechanical carbon/silicon new seal face material solutions was the motivation behind
carbide combination. However, the ever-increasing techni- this work. The work aimed to evaluate and increase the un-
cal and commercial demands placed on seal performance derstanding of the tribological nature of the current standard
are a strong driving force for the development of new hard face combinations and evaluate the potential of PCD
and novel sealing solutions. The results from this work and DLC films. The results of the experimentation carried
and the model developed provide critical data regards the out, the subsequent analysis and the contact model devel-
tribo-performance of hard face combinations under dry oped have met these aims and have provided a developmen-
running conditions. The model here provides the seal de- tal framework for the material scientist and the mechanical
signer with both a diagnostic tool and developmental tool. face seal designer. It has been shown that the tribological
The model enables an evaluation of the likely impact of performance of the current standard hard ceramic combina-
the contact condition and environmental aspects on the tions under mechanical contact is related to the ability of
tribo-dynamics of the seal face arrangement. It provides the combination to sustain a mild wear rate regime up to
the format for seal face performance mapping. It provides high PV values. The work has shown that a transition from
a framework for the development of seal face materials and mild wear to severe wear takes place at some critical contact
seal face design. There is also use for this model in lubricated pressure. The potential of PCD and DLC films in modify-
mechanical face seals, in which hard face combinations are ing the contact dynamics such that this transition is shifted
used. In a liquid or gas lubricated mechanical seal there is to a higher PV value has been shown. The contact model
always a degree of mechanical contact either at stop/starts developed has provided a route for relating the features of
and during general operation due to the close dimensional mechanical face seal design to the critical tribological char-
tolerances associated with thin liquid films or thin gas films. acteristics of the seal face material systems. This work has
During these periods of partial lubrication the contacting shown that engineered tribological wear resistant thin films,
faces need to be able to cope with the tribo-stresses pre- particularly the PCD type films, could provide the sealing
sented. The model can be used to predict the likely outcome. industry with the next step forward in mechanical face seal
Currently all the major mechanical face seal manufactur- materials technology.
ers have a significant interest in the potential of diamond The experiments carried out and the contact model devel-
and diamond-like carbon films. Discussions with a number oped here are a first step and further refinement is currently
of mechanical seal manufacturers by this author has identi- ongoing. PCD films have relatively low fracture strength
fied that PCD and DLC films are finding tentative use but and low fracture toughness. This work, however, has il-
improved understanding of the material systems are being lustrated the importance of these properties along with the
452 G.A. Jones / Wear 256 (2004) 433–455

tribo-chemistry of the contact conditions to tribological performance obtained but a great deal of additional research
performance. Putting this in a mechanical face seal material is still required.
engineering context, it can be recognised that the fracture
strength and the fracture toughness of PCD type films are
predominantly dependent on crystallite size and morphol- Acknowledgements
ogy and the chemistry of the grain boundary phase. In
mechanical face seal materials there is the requirement for The author wishes to express his appreciation to Gledco
the more wear resistant pure sp3 structure. However, there Engineered Materials Ltd., US Graphite Inc. and the Univer-
is also the need for a highly polished surface, which makes sity of Salford for their support of this research. The advice
the sp2 + sp3 composite structure desirable. Importantly this and excellent technical assistance of Professor R.D. Arnell
work has begun to expose the relationship between these and Mr. G. France, University of Salford, are gratefully ac-
structural models (and their derivatives) and the tribological knowledged.

Appendix A. Critical data collected from both experiments 1 and 2

Details of the composition and properties of the materials used in this work

Materials Description Density Hardness Elastic Bend Thermal CTE × 10−6

(kg/m3 ) (Hv/∗ shore A) modulus strength conductivity (◦ C−1 )
(GPa) (MPa) (W/m K)
Alpha sintered SiC SiC: >98.5%, free Si: 0% 3100 3500 400 450 80 4.8
Reaction bonded SiC SiC: remainder, free Si: 10–12% 3010 2500–3500 390 300 73 4.8
Reaction bonded SiC + graphite SiC: remainder, free Si: 2850 – 260 180 70 –
10–12%, free graphite: 30%
Tungsten carbide + Co binder WC: 94%, Co: 6–8% 13500 1500–1600 550 1750 65 5.1
Mechanical carbon grade 001 Carbon + resin 1850 95∗ 23 79 10.37 5.8
Mechanical carbon grade 002 Carbon + graphite + resin 1830 105∗ 24 85 13.83 4.1
Mechanical carbon grade 003 Carbon + graphite + resin 1850 85∗ 22 75 13.83 4.9
a-Ti:C2 H2 DLC film Deposited by closed field unbalanced magnetron sputter ion plating + plasma assisted CVD to a thickness
of 3.5–4.5 ␮m with a critical load >55 N and hardness 2100–2500 Hv (Teer Coatings Limited)
PCD film Deposited by microwave plasma assisted CVD. Nucleated by agitation in a slurry of diamond powder.
Standard gas composition used to deposit crystalline film of final thickness of 5–6 ␮m (GEC Marconi
Materials Technology)

Material systems and combinations examined in this work

Pad Cylinder

Experiment 1a
Combination 1 Reaction bonded SiC Tungsten carbide + Co binder
Combination 2 Reaction bonded SiC Alpha sintered SiC
Combination 3 Reaction bonded SiC + 30% graphite Alpha sintered SiC
Substrate Deposited film Cylinder

Experiment 2a
Combination 4 Reaction bonded SiC PCD Alpha sintered SiC
Combination 5 Reaction bonded SiC DLC Alpha sintered SiC
Combination 6 Reaction bonded SiC PCD Grade 001
Combination 7 Reaction bonded SiC DLC Grade 001
Combination 8 Reaction bonded SiC DLC Grade 002
Combination 9 Reaction bonded SiC DLC Grade 003
a
At least two full test runs per material combination were carried out and for analytical purposes part test runs were
conducted.
 G.A. Jones / Wear 256 (2004) 433–455 453

Details of the surface condition of the contacting surfaces of the hard ceramic combinations examined in experiment 1, before
and after each test run

Material combination Stationary pad (1) Rotating cylinder (2)

Ra (␮m) Rq (␮m) Sm (␮m) Ra (␮m) Rq (␮m) Sm (␮m)

(1) Reaction bonded SiC/(2) alpha sintered SiC Before 0.06 0.08 336 0.60 0.8 101
After 4.00 7.8 178 0.60 1.0 70
Before 0.06 0.08 362 0.63 0.8 117
After 4.20 6.4 178 0.60 1.1 77
(1) Reaction bonded SiC/(2) tungsten carbide Before 0.06 1.0 232 0.73 0.92 54
After 1.20 1.60 178 1.28 1.82 170
Before 0.08 0.07 303 0.59 1.1 121
After 3.17 5.6 132 0.60 1.4 74
(1) Reaction bonded SiC + graphite/(2) Before 0.34 0.46 120 0.68 0.82 111
alpha sintered SiC
After 3.56 5.86 171 1.31 2.01 96
Before 0.39 0.38 121 0.71 0.84 76
After 1.32 1.78 185 4.06 7.9 152

Details of the surface condition of material combinations incorporating a PCD or DLC film in contact with alpha sintered
SiC

Material combination Stationary pad (1) Rotating cylinder (2)

Ra (␮m) Rq (␮m) Sm (␮m) Ra (␮m) Rq (␮m) Sm (␮m)

(1) Reaction bonded SiC + PCD film/(2) Before 0.25 0.40 113 0.60 0.83 98
alpha sintered SiC After 2.18 PCD film breakdown 0.82 0.98 137
Before 0.28 0.49 115 0.73 0.92 54
After 2.79 5.80 129 2.92 4.20 57
(1) Reaction bonded SiC + DLC film/(2) Before 0.06 0.08 550 0.71 1.02 53
alpha sintered SiC After 4.11 DLC film delaminated 1.96 1.82 165
Before 0.06 0.07 800 0.60 0.96 101
After 3.64 DLC film delaminated 0.86 1.21 91
(1) Reaction bonded SiC + PCD film/(2) Before 0.21 0.50 38 1.37 1.83 58
mechanical carbon grade 001 After 2.98 PCD film delaminated 1.69 – –
Before 0.28 0.49 35 1.27 1.92 54
After 2.89 PCD film delaminated 2.92 4.20 157
(1) Reaction bonded SiC + DLC film/(2) Before 0.16 0.24 38 1.71 1.02 53
mechanical carbon grade 001 After 4.11 DLC film delaminated 1.96 1.87 65
Before 0.16 0.26 47 1.60 0.96 101
After 3.64 DLC film delaminated 3.86 4.21 91
(1) Reaction bonded SiC + DLC film/(2) Before 0.20 0.30 30 1.45 – –
mechanical carbon grade 002 After DLC film delaminated 2.01 – –
Before 0.18 0.32 38 1.36 – –
After DLC film delaminated 2.34 – –
(1) Reaction bonded SiC + DLC film/(2) Before 0.16 0.24 60 1.47 1.53 56
mechanical carbon grade 003 After DLC film delaminated 1.85 – –
Before 0.17 0.25 42 1.63 – –
After DLC film delaminated 2.34 – –
454 G.A. Jones / Wear 256 (2004) 433–455

Comparison of the tribo-performance indicators for the critical combinations examined in this work
Material combination Maximum PV limit Friction coefficient T (◦ C)
normal load (MPa m/s) Maximum Minimum Mean
(N)
Experiment 1
Reaction bonded SiC/alpha sintered SiC 147.15 103.04 1.750 0.650 0.897 61
156.96 109.91 1.55 0.712 0.916 74
Reaction bonded SiC/tungsten carbide 147.15 128.94 2.00 0.622 0.951 54
156.96 137.54 1.75 0.643 0.887 62
Reaction bonded SiC + 30% graphite/alpha sintered SiC 156.96 95.063 0.656 0.078 0.255 90
156.96 95.063 0.617 0.101 0.263 110
Experiment 2
Reaction bonded SiC + PCD/alpha sintered SiC 490.50 262.09 0.375 0.070 0.169 16
343.35 219.10 0.250 0.071 0.130 18
Reaction bonded SiC + DLC/alpha sintered SiC 68.67 87.04 1.50 0.275 0.857 –
49.05 73.56 1.167 0.375 0.839 32
Reaction bonded SiC + DLC/mechanical 529.74 62.77 0.405 0.146 0.209 45
carbon grade 001 529.74 66.16 0.405 0.132 0.200 45
Reaction bonded SiC + DLC/mechanical 490.50 55.26 0.533 0.173 0.284 70
carbon grade 002 – – – – – –
Reaction bonded SiC + DLC/mechanical 588.60 71.41 0.405 0.169 0.245 60
carbon grade 003 588.60 71.41 0.241 0.169 0.354 58
Ref. [1]
Reaction bonded SiC/Mechanical carbon grade 001 441.45 59.36 0.421 0.158 0.268 66
441.45 59.36 0.421 0.184 0.274 63

Details of the contact stresses at the point of contact failure for all material combinations examined (for alpha sintered
SiC/RBSiC + DLC the values in parentheses represent the point where contact characteristics changed)

Material combination Maximum Hertzian Maximum Hertzian Maximum surface

contact pressure, p0 tangential stress, q0 tensile stress, σ t
(MPa) (MPa) (MPa)
Experiment 1
Reaction bonded SiC/alpha sintered SiC 269.39 197.55 395.10
269.39 188.57 377.15
Reaction bonded SiC/tungsten carbide 337.11 209.76 419.51
365.27 239.72 530.26
Reaction bonded SiC + 30% 248.54 144.37 288.77
graphite/alpha sintered SiC
240.77 144.63 297.88
Experiment 2
Alpha sintered SiC
Reaction bonded SiC + PCD 685.23 47.97 95.93
530.30 44.23 88.46
Reaction bonded + DLC 227.56 (121.63) 276.32 (48.65) 552.63 (106.43)
173.09 (121.63) 171.08 (45.61) 346.18 (91.22)
Reaction bonded + DLC
Mechanical carbon grade 001 164.10 22.16 45.54
164.10 25.85 51.70
Mechanical carbon grade 002 144.47 25.17 56.08
– – –
Mechanical carbon grade 003 170.44 24.16 51.78
170.44 25.89 48.33
Reaction bonded + PCD
Mechanical carbon grade 001 165.51 26.07 52.14
165.51 22.35 44.70
 G.A. Jones / Wear 256 (2004) 433–455 455

References [12] J.W. Hutchinson, Z. Sou, Mixed mode cracking in layered materials,
Adv. Appl. Mech. 29 (1992) 63–191.
[1] G.A. Jones, The tribology of the mechanical sealing interface: an [13] S.A.G. Oliveira, A.F. Bower, An analysis of fracture and delamination
evaluation of the role and potential of surface engineering, Ph.D. in thin coatings subjected to contact loading, Wear 198 (1996) 15–32.
Thesis, University of Salford, 1999. [14] N.P. O’Dowd, M.G. Stout, C.F. Shih, Fracture toughness of
[2] G.M. Hamilton, Explicit equations for the stress beneath a sliding alumina–niobium interfaces. Experiments and analyses, Philos. Mag.
spherical contact, Proc. Inst. Mech. Eng. 197C (1983) 53–59. A 66 (6) (1992) 1037–1064.
[3] G.M. Hamilton, L.E. Goodman, The stress field created by a circular [15] J.S. Wang, Z. Suo, Experimental determination of interfacial tough-
sliding contact, J. Appl. Mech. 88 (1966) 371–376. ness using Brazil-nut-sandwich, Acta Metall. 38 (1990) 1279–1290.
[4] K.L. Johnson, Contact Mechanics, Cambridge University Press, [16] S. Jahanmir, D.E. Deckman, L.K. Ives, A. Fieldman, E. Farabaugh,
Cambridge, 1985. Tribological characteristics of synthesised diamond films on silicon
[5] M. Comninou, The interface crack with friction in the contact zone, carbide, Wear 133 (1989) 73–81.
J. Appl. Mech. 44 (1977) 287–290. [17] K. Miyoshi, R.L.C. Wu, A. Garscadden, P.N. Barnes, H.E. Jackson,
[6] M. Comninou, The interface crack, J. Appl. Mech. 44 (1977) 631– Friction and wear of plasma deposited diamond films, in: Proceedings
636. of the International Conference and Meeting on Coatings and Thin
[7] M. Comninou, D. Schmueser, The interface crack in a combined Films, NASA Technical Memorandum 105926, San Diego, CA, 1993.
tension–compression and shear field, J. Appl. Mech. 46 (1978) 345– [18] F.P. Bowden, A.E. Hanwell, The friction of clean crystal surfaces,
348. Proc. R. Soc. London, Ser. A 295 (1966) 233–239.
[8] J.R. Rice, Elastic fracture mechanics concepts for interfacial cracks, [19] T.M. Dugger, D.E. Peebles, L.E. Pope, in: Y.-W. Chung, A.M.
J. Appl. Mech. 55 (1988) 98–103. Homola, G.B. Street (Eds.), Surface Science Investigations in
[9] R. Rezakhanlou, J. Von Stebut, Damage mechanisms of hard coatings Tribology: Experimental Approaches, ACS Symposium Series No.
on hard substrates: a critical analysis of failure in scratch and wear 485, ACS Books, Washington, DC, 1992, pp. 72–102.
testing, in: D. Dowson, C.M. Taylor, M. Godet (Eds.), Mechanics [20] F.P. Bowden, J.E. Young, Friction of diamond, graphite and carbon
of Coatings, Elsevier, New York, 1990, pp. 183–192. and the influences of surface films, Proc. R. Soc. London, Ser. A
[10] J.A. Blume, M. Ortiz, Effects of interface decohesion and skidding 208 (1951) 444–451.
on bimaterial crack tip fields, Int. J. Fracture 42 (1990) 117– [21] D. Tabor, in: J.E. Field (Ed.), The Properties of Diamond, Academic
128. Press, London, 1979, Chapter 10, pp. 325–350.
[11] C.F. Shih, Cracks on bimaterial interfaces: elasticity and plasticity [22] J. Wilks, E.M. Wilks, in: J.E. Field (Ed.), The Properties of Diamond,
aspects, Mater. Sci. Eng. A 143 (1991) 77–90. Academic Press, London, 1979, Chapter 12, pp. 351–382.