Thin Solid Films 370 (2000) 223±231
www.elsevier.com/locate/tsf
Supertough wear-resistant coatings with `chameleon' surface adaptation
A.A. Voevodin*, J.S. Zabinski
Materials and Manufacturing Directorate, Air Force Research Laboratory, AFRL/MLBT, 2941 P Street, Bldg. 654, Wright-Patterson Air Force Base, OH
45433-7750, USA
Received 24 September 1999; received in revised form 11 February 2000; accepted 29 February 2000
Abstract
The chameleon's ability to change skin color depending on environment to increase its chances of surviving served as an inspiration in the
development of self-adaptive supertough wear-resistant coatings. Surface chemistry, structure and mechanical properties of these thin (0.5
mm) coatings reversibly change with applied load and environment, providing the best wear protection. Coating designs developed in-house
are reviewed together with a critical analysis of design reports in the literature. `Chameleon' coatings were prepared using novel nanocomposite structures, consisting of crystalline carbides, diamond-like carbon (DLC), and transition metal dichalcogenides. Various mechanisms were activated to achieve surface self-adaptation and supertough characteristics. They included: transition of mechanical response from
hard and rigid to quasi plastic by grain boundary sliding at loads above the elastic limit; friction induced sp 3 ! sp 2 phase transition of the
DLC phase; re-crystallization and reorientation of the dichalcogenide phase; change of surface chemistry and structure from amorphous
carbon in humid air to hexagonal dichalcogenide in dry nitrogen and vacuum; and sealing the dichalcogenide phase to prevent oxidation.
These mechanisms were demonstrated using WC/DLC, TiC/DLC, and WC/DLC/WS2 coatings. The hardness of WC/DLC and TiC/DLC
composites was between 27±32 GPa and scratch toughness was 4±5 fold above that of nanocrystalline carbides. The WC/DLC/WS2
composites survived millions of sliding cycles in vacuum and air under 500±1000 MPa loading, and exhibited excellent friction recovery
in humid $ dry environmental cycling. Their friction coef®cients were about 0.1 in humid air, 0.03 in vacuum, and as low as 0.007 in dry
nitrogen. The proposed `chameleon' concept can dramatically increase wear-resistant coating applicability, durability, and reliability.
q 2000 Elsevier Science S.A. All rights reserved.
Keywords: Nanocomposite; Coatings; Tribology; Hardness
1. Introduction
Modern methods of vacuum deposition provide great
¯exibility for manipulating material chemistry and structure, leading to ®lms and coatings with unique properties
that are often unachievable in the bulk materials. In the ®eld
of wear-resistant coatings, one recent breakthrough is superhard materials, where hardness is dramatically increased by
using superlattice and amorphous/crystalline composite
structures [1±5]. These structures were designed on a
nanometer level to prevent operation of dislocation sources,
restrain existing dislocations in adjusted nanolayers (or
phases) with different elastic constants, and build-up material volume energy from grain boundary incoherence
strains. While superhard coatings are very important for
protection of cutting tools, most tribological applications
require tough and low friction coatings.
A material is generally considered tough if it withstands
* Corresponding author. Tel.: 11-937-255-9001; fax: 11-937-255-2176.
E-mail address:
[email protected] (A.A. Voevodin)
high levels of loading (tensile, compressive, shear, etc.) and
can dissipate strain energy without brittle fracture. A supertough coating must have high elastic modulus and high
hardness, as well as permit strain relaxation and crack termination. For materials used to prevent wear, there are additional requirements due to the nature of loading. A wearresistant coating must support high loads in sliding or rolling contact without failure by wear, cohesive fracture, or
loss of adhesion at the substrate interface. Most frequently, a
low friction coef®cient is required, which helps to reduce
friction losses and increase load support capability. The
later is clear from the fact, that typical coating failures
(deformations, cracks, delaminations, etc.) are caused by
the tangential stress which is proportional to the contact
load through the friction coef®cient. Finally, chemical and
thermal stability in tribochemical reactions with the counterpart and environment are required.
The development of tough wear-resistant coatings is quite
a challenging task, requiring considerations of substrate
materials, loading schemes, operational environment, tribochemical reactions, etc. Consequently, coatings must be
0040-6090/00/$ - see front matter q 2000 Elsevier Science S.A. All rights reserved.
PII: S 0040-609 0(00)00917-2
224
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
tailored to more or less de®ned applications, so that
currently there is no `universal' coating. Progress toward a
`universal' coating can be made by developing materials
which adapt their chemistry, mechanical characteristics,
and tribological properties, depending on the operating
environment, temperature, and the magnitude of loading.
During the last several years, a considerable effort in this
direction was made at AFRL/WPAFB for potential applications in tribological pairs of aerospace devices. The ®rst
advancements were solid lubricant composites made of
oxides and dichalcogenides (PbO/MoS2, ZnO/MoS2, ZnO/
WS2), which could operate in a broad temperature range [6±
8]. Advanced multilayer structures were designed to
combine these composites with buried diffusion barrier
layers and achieve surface self-adaptation during repeated
temperature cycling (J.S. Zabinski, S.D. Walck, J.E. Bultman, N.T. McDevitt, unpublished data). Recently, novel
wear-resistant materials were developed, which combine
nanocrystalline carbides (TiC, WC), dichalcogenides
(MoS2, WS2), and amorphous diamond-like carbon (DLC)
into nano-composite structures. The surface chemistry,
structure and mechanical behavior of these nanocomposite
materials could reversibly change in the tribological
contact, depending on applied loads and operational environment to maintain good friction and avoid wear. Such
adaptive behavior is well known in nature, as for example,
when a chameleon changes its color to match its environment and avoid predators. Development of `chameleon'
coatings with super-toughness, exceptional wear resistance,
environment stability, and low friction characteristics is
discussed in this paper together with a review of relevant
coating designs reported in the literature.
produced within the following nanocrystalline grains/amorphous matrix systems: TiN/a-Si3N4 [4,10,11], W2N/a-Si3N4
[5,12], VN/a-Si3N4 [5,12], TiN/c-BN [5,12], TiN/a(TiB2 1 TiB 1 B2 O3 ) [13], TiN/TiB2 [14], TiC/TiB2 [14],
TiN/Ni [15], ZrN/Cu [16], ZrN/Y [17], TiAlN/AlN [18],
CrN/Ni [19], Mo2C/a-(carbon 1 Mo2N) [20], TiC/DLC
[21,22], and WC/DLC [23,24]. Among carbon matrixes,
only hydrogen free DLC provided a composite with hardness of 30±40 GPa [21±24], which is approaching the 40±60
GPa hardness of ceramic-matrix composites [4,5], while
hydrogen terminated DLC matrixes could provide only
10±20 GPa [25,26]. In the references listed above, the
authors deliberately produced nanocomposite structures
with improved mechanical characteristics. In some earlier
works within multiphase systems similar structures were
possibly produced, but their nanocomposite nature was not
emphasized [27±29]. A detailed review of these and other
hard and super-hard coatings can be found in a most recent
work by VepcÏek [30] published when this manuscript was in
a revision.
The increase in hardness of composites in comparison to
that of single phase coatings was based on the suppression
of dislocation operation by using small 3±5-nm grains and
inducing grain incoherence strains when using 1-nm thin
matrix for grain separation [4,5]. The incoherence strain is
likely increased, when grain orientations are close enough to
provide interaction between matched but slightly misoriented atomic planes. In the absence of dislocation activity,
Grif®th's equation for crack opening was proposed as a
simple description of the composite strength, s [31]
2. Supertough composites with mechanical property selfadjustment
where E is elastic modulus, g s is surface energy of the grain/
matrix interface, and a is initial crack size, which was
accepted to be equal to the average diameter of the grains
[12]. The use of a full crack length instead of a half-length
calls for the factor four in the equation above. From this
equation, the composite strength can be increased by
2.1. Existing concepts of high-strength coatings
A concept for superhard (40±70 GPa) materials was
recently introduced, which is based on a combination of
nanocrystalline and amorphous phases in composite structures to suppress ductility and increase strength, using grain
boundary effects [4,5]. A variety of hard materials can be
used as the nanocrystalline phase, including nitrides,
carbides, borides, and oxides. The nanocrystalline grains
must be 3±10 nm in size and separated by 1±3 nm within
an amorphous matrix, which consists of other ceramics,
metals, carbon, etc. (Fig. 1). It has been suggested that the
nanocrystals should be oriented in a common direction
(have low angle boundaries) in order to provide interaction
across the amorphous matrix and maximize the desirable
super-hardness effect [9]. We shall give a brief analysis of
these designs before introducing the concept for supertough
`chameleon' nanocomposites.
At present, hard nanocomposite coatings have been
s
4Eg s
pa
1=2
1
Fig. 1. Design schematic of an amorphous/crystalline nanocomposite with
high-strength characteristics.
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
increasing elastic modulus and surface energy of the
combined phases, and by decreasing the crystalline grain
sizes. For superhard composites, in addition to the selection
of an appropriate material system, the elastic modulus arti®cially increases with decreasing grain sizes due to the
lattice incoherence strains and high volume of grain boundaries. In practice, grain boundary defects always exist, and a
3-nm grain size was found to be close to the minimum limit.
Below this limit, the strengthening effect disappears because
grain boundaries and grains become indistinguishable and
the stability of the nanocrystalline phase is greatly reduced
[1±5].
This composite design increases elastic modulus and
hardness, but does not necessarily yield high toughness
characteristics. First, dislocation mechanisms of deformation are prohibited and crack opening is the main mechanism to relax strains when stresses exceed the strength limit.
Second, Grif®th's equation does not take into account the
energy balance of a moving crack, which consists of the
energy required to break bonds and overcome friction
losses, potential energy released by crack opening, and
kinetic energy gained through crack motion [32]. From
crack energy considerations, a high amount of stored incoherence strain dictates a high rate of potential energy release
in the moving crack. In such conditions, a crack can achieve
the self-propagating (energetically self-supporting) stage
sooner, transferring into a macrocrack and causing brittle
fracture.
One way to improve composite toughness is to combine a
hard nanocrystalline phase with a soft metal matrix. This
approach has been widely explored in macrocomposites
made of ceramics and metals which are known as cermets
[33]. It was recently scaled down to the nanometer level in
thin ®lms made of hard nitrides and softer metal matrixes
[15±18]. When the dimensions of the metal matrix permit
operation of dislocations, the composite strength can be
described by the following form of the Grif®th±Orovan
model [31]
4E
g s 1 gp rtip 1=2
s
2
pa
3da
where g p is the work of plastic deformation, rtip is the curvature of the crack tip, and da is the interatomic distance.
According to this equation, tip blunting and the work of
plastic deformation considerably improve material strength,
counter-balancing lower elastic moduli of metals in comparison to that of ceramics. However, in the nanocomposite
design, dislocation operation may be prohibited because
the separation of grains by the metal matrix is very small.
For example, the critical dimension, D, for a Frank±Read
dislocation source is [31]
D Gbt21
3
where G is the modulus of rigidity approximated by the
expression for Young's modulus and Poisson's ratio,
225
G E
2 1 2n21 , b is the Burgers displacement, and t is
shear stress. For sliding wear contact with a 10 GPa contact
load and a 0.1 friction coef®cient, shear stress of about 1
GPa can be expected. At this level of stress, the critical size
of Frank±Read source operation is 7.4 nm for a Cu matrix
(E 110 GPa, v 0:34, b 0:181 nm) and 14.1 nm for a
Ni matrix (E 210 GPa, n 0:31, b 0:176 nm). The
contact load of 10 GPa represents a rather high loading
for real tribological contacts. Smaller loads would require
larger grain sizes for dislocation source operation. Since the
tendency toward plastic deformation increases with decreasing D, such soft materials as gold (Destimate 5:9 nm, at E
78 GPa, n 0:42, b 0:214) and silver (Destimate 5:5 nm
at E 74 GPa, n 0:37, b 0:204) could be better
choices for the metal matrix.
Matrix dimensions in hard nanocomposite coatings are
typically between 1 and 3 nm, which is well below the
critical size for dislocation source operation, even in very
soft metal matrixes. Therefore, the mechanical behavior of
such composites can be expected to be similar to that of
ceramic matrix composites. Con®rmation of this is found
in a recent report of a super-hardness effect for nanocomposites consisting of 35-nm ZrN grains within a thin Cu matrix
[16]. Thus, alternative mechanisms for strain release should
be sought in order to prevent brittle fracture and improve
toughness.
There is one unique feature of nanocomposites, which
prevents shattering of internally strained super-hard ceramic/ceramic and ceramic/metal composites into pieces.
Namely, they contain a high volume of grain boundaries
with a crystalline/amorphous transition across grain±matrix
interfaces, limiting initial crack sizes and helping to de¯ect
and terminate growing cracks. These mechanisms were used
as a primary explanation for the brittle resistance of novel
super-hard composites, whose ductility is quite low but yet
they can resist cracking during large elastic deformations
[30]. Alternatively to super-hard materials, grain boundary
diffusional atom ¯ow [34] and grain boundary sliding [35±
37] were suggested to improve ductility and provide superplasticity in single-phase nanocrystalline ceramics. The
most recent research indicates that high ductility can be
more easily achieved in multiphase structures [38] and
that grain boundary sliding is a primary mechanism of
superplasticity [39±42]. It was also found that equiaxial
grain shapes, high angle grain boundaries, low surface
energy, and the presence of an amorphous boundary phase
facilitate grain boundary sliding [33±35]. These ®ndings can
be expanded into the ®eld of hard wear-resistant coatings to
introduce ductility and prevent fracture under a high contact
load.
2.2. New concept development
The concept of `chameleon' wear-resistant coatings
mentioned in the introduction was explored on composites
consisting of hard nanocrystalline carbide and amorphous
226
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
diamond-like carbon (DLC) phases. To provide lubrication
in vacuum environments, they were also doped with a
dichalcogenide phase (discussed in Section 3). Hydrogenfree DLC was selected as a matrix material because it is
amorphous, has very high hardness (60±70 GPa), high elastic modulus (t600) GPa, low friction in humid ambient
environments, and low wear [43]. Toughness was expected
to be achieved by strain release via carbide nanograin sliding in the DLC matrix. Based on the literature analyses
discussed above, the following concepts were suggested
for a super-tough nanocomposite structure:
(i) combine crystalline carbide and amorphous DLC with
high elastic modulus to achieve correspondingly high
hardness;
(ii) maintain nanocrystalline grain sizes at the 10±20 nm
level to restrict initial crack size and create a large volume
of grain boundaries;
(iii) separate the nanograins within an amorphous matrix
thickness above 2 nm to prevent interaction of atomic
planes in the adjusted nanocrystalline grains and facilitate
grain boundary sliding, but less than 10 nm to restrict path
of a straight crack;
(iv) produce nanocrystalline grains with random orientation (high angle grain boundaries) to minimize grain incoherence strain and facilitate grain boundary sliding;
(v) release `over-limit' strain by grain boundary sliding at
loads exceeding the composite elastic strength, providing
hard or ductile behavior depending on the load;
(vi) terminate nanocracks by de¯ection at grain boundaries and by energy loss within the amorphous matrix.
Following this concept, nanocrystalline TiC and WC
grains have been embedded in a hydrogen-free DLC matrix
using a hybrid deposition technique, combining magnetron
sputtering and pulsed laser deposition [21,23]. Detailed
discussion of the technique can be found in Ref. [44]. The
following key features were most important for achieving
the objectives above:
(a) operation of two independent plasma sources (metal
from sputtering, carbon from laser ablation), which
provide a full range of composition control for each of
the elements;
(b) use of a high energy carbon plasma, facilitating crystalline carbide formation at near room temperatures;
(c) pulses of the ablated plasma, interrupting crystalline
growth and restricting grain sizes to the nanometer level.
These features enabled the precise control of, not only
chemistry, but also structure of both the amorphous DLC
phase and the nanocrystalline TiC or WC phases, which had
5±10-nm dimensions estimated by transmission electron
microscope (TEM) imaging (Fig. 2). The complete chemical and structural analyses of the composites can be found in
Refs. [23,45]. For mechanical tests, 0.5-mm thick coatings
were deposited onto stainless steel substrates. Thin Ti12xCx
and W12xCx functional gradient layers were used to improve
coating adhesion and relax interfacial stresses. Composition, structure, mechanical properties, and preparation of
such interlayers were described in Ref. [46].
The hardness of the nanocomposite coatings was about 32
GPa for TiC/DLC and 27 GPa for WC/DLC with correspondingly high elastic modulus, producing a hardness to
modulus ratio of about 0.1 [23,45]. Single-phase materials
with such high hardness would typically experience a brittle
fracture once the load exceeded their elastic limit. The behavior of the TiC/DLC and WC/DLC composites was in a
stark contrast to this expectation. Fig. 3 shows the wear
scars induced by a diamond tip of 0.2 mm radius loaded
Fig. 2. Transmission electron microscope images and corresponding electron diffraction patterns for TiC/DLC (a) and WC/DLC (b) nanocomposite coatings.
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
227
Fig. 3. Microphotographs of `plastic' deformation within TiC/DLC (a) and WC/DLC (b) nanocomposite coatings in scratch tests with 0.2 mm radius diamond
tip at 50 N load.
with 50 N and dragged across the coating surfaces. Under
such high contact load, both coatings exhibited surface
deformations, which visually appear to be `plastic'. This
was, however, not true plasticity since dislocation sources
were prohibited, but rather a result of what we believe is TiC
and WC grain boundary sliding in the DLC matrix.
A high-resolution secondary electron microscopy image
within the scratch reveals nanocrack openings at locations
where grain boundary sliding had occurred (Fig. 4). Such
nanocrack openings are similar to cavitations observed in
bulk nano- and micro- crystalline ceramics, which contributed to the deformation by boundary sliding [36,47]. The
interconnection of intergranular cavitation was identi®ed as
the main factor causing brittle failure and preventing superplasticity in microcrystalline metals and ceramics [48]. This
can be prevented when grain sizes and, hence, sliding cavitation are reduced to the nanometer level. Additionally, the
presence of an amorphous phase on the boundaries helps to
de¯ect and terminate cracks in addition to the enhancement
of boundary mobility [35,36]. Examples of de¯ection and
termination of nanocracks induced by grain sliding in a TiC/
DLC nanocomposite can be seen in Fig. 4. Grain boundary
sliding and crack termination mechanisms eliminated brittle
fracture and provided ductility within TiC/DLC and WC/
DLC composites observed on a macrosopic level in Fig. 3.
Thus, the TiC/DLC and WC/DLC coatings are hard and
non-plastic at contact loads below their elastic limit, but
change their behavior to the described `plastic' mode
above this limit. This avoids brittle fracture at high levels
of loading, and also creates the possibility to support high
contact loads by local coating compliance and load distribution onto larger areas. Such self-adjustment of the mechanical properties from hard to ductile resulted in dramatic
Fig. 4. High-resolution scanning electron microscope image of the surface
within `plastic' deformation zone of a TiC/DLC nanocomposite coating.
Arrows indicate places of possible grain boundary sliding, boundary cavitation, and nanocrack termination.
Fig. 5. Scratch toughness characteristics of the TiC/DLC and WC/DLC
coatings, as a function of the carbon content. The composition region
corresponding to the preparation of the nanocomposites in according to
the conceptual supertoughness design is shaded.
228
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
improvement of scratch toughness characteristics (Fig. 5).
Scratch toughness estimates were at least 3±4 times above
that of single-phase nanocrystalline TiC and WC coatings
deposited on the same substrates, with the same interlayers,
and the same thicknesses. This encouraging result justi®ed
the proposed super-tough coating concept.
3. Wear-resistant composites with chemical and
structural self-adjustment
Both TiC/DLC and WC/DLC supertough composite coatings were considered for controlling friction and wear in
aerospace applications. They exhibited very low wear
rates and friction coef®cients in ambient or humid environments [49,50], but had a potential problem for space lubrication. The problem was surface graphitization of DLC,
caused by friction induced sp 3 ! sp 2 interatomic bonding
transformation after several thousands sliding cycles [51].
Graphite requires an intercalation agent, such as water
molecules, for lubrication [52]. In the absence of graphite
intercalation, free p electrons of graphite participate in the
interplanar bonding of hexagonal sp 2 phase [53]. This,
together with strong covalent bonding between frictioninduced dangling bonds of sp 3 sites, leads to a very high
friction [51]. Hydrogenated DLC may provide a partial
solution due to hydrogen termination of the dangling carbon
bonds. However, hydrogenated DLC structures are much
softer, plus hydrogen may eventually be depleted [54].
This limits coating applications to shorter runs and smaller
contact loads as well as imposes a danger of graphitization
in non-favorable operating conditions, e.g. higher temperatures or oxidizing environments.
The most common solid lubricants for space applications
are transition metal dichalcogenides, from which MoS2 and
WS2 are well known to endure millions of sliding cycles in
high vacuum conditions, providing very low 0.02±0.04 friction coef®cients [55,56]. Their disadvantage is a low hardness and a high rate of oxidation in ambient conditions,
which leads to increased wear and deterioration of lubrication properties [57,58]. Also water adsorption between
plates of sandwiched S±Mo±S or S±W±S basal planes
increases the ®ction coef®cient [59,60]. Furthermore, these
materials should have fully crystalline hexagonal grains
oriented parallel to the friction surface for good lubrication
[60±62]. There were several reports in the literature, that
initially amorphous or nanocrystalline MoS2 ®lms were
found to experience crystallization and re-orientation into
hexagonal plates in sliding contacts [6,63±66]. This structural transformation can be used to achieve lubrication properties on the surface of nanocomposite coatings, where
dichalcogenide phases may be included in initially nanocrystalline or amorphous form.
A concept was developed, where poorly crystallized
dichalcogenide phases and nanocrystalline carbide phases
were encapsulated in an amorphous DLC matrix. The
concept pursued the following objectives:
(i) use DLC sp 3 ! sp 2 structural adjustment for ambient
environment lubrication;
(ii) use dichalcogenide re-crystallization and reorientation for high vacuum lubrication;
(iii) use friction to induce structural adjustments;
(iv) use friction and environmental action to adjust the
dominant surface chemistry from that of DLC in ambient
environments to that of dichalcogenide in vacuum.
The realization of these objectives was explored using the
W-C-S material system, within which nanocrystalline WC,
amorphous DLC, and poorly crystallized WS2 were
produced. The Ti±C±Mo±S system was found less favorable, due to a thermodynamically driven tendency to bind
sulfur into TiS2, which does not have good lubrication properties. Examples of TEM images of WC/DLC/WS2 nanocomposites are shown in Fig. 6, comparing structures of the
coatings with about 15 and 30 at.% S. At the higher sulfur
content, randomly oriented WS2 plates were identi®ed (Fig.
6b), while at lower sulfur content the WS2 phase was amorphous (Fig. 6a). Details on the chemistry, structure, and
tribology of WC/DLC/WS2 composites can be found in
Refs. [67,68]. Friction tests were performed in a pin-ondisk tribometer against 6-mm diameter 440C steel balls
under 100 g load and 0.16 m/s sliding velocities. Only the
main results are summarized here to discuss the proposed
`chameleon' concept.
Graphitization of DLC and crystallization of WS 2 in the
friction contact were con®rmed by micro-Raman spectroscopic analyses. Fig. 7 compares Raman spectra recorded
for friction surfaces of the as-deposited WC/DLC/WS2
coating, the same coating after 50 000 sliding cycles in
air with 50% relative humidity (RH), and after 50 000
sliding cycles in 10 27 Pa vacuum. Structural changes
from initially amorphous to either graphite-like carbon,
after tests in air, or to hexagonal WS2, after tests in
vacuum, were demonstrated. These structural changes
were accompanied by a corresponding variation in friction coef®cient. The friction coef®cient was about 0.15 for
sliding in air, compared to ,0.1 for single-phased DLC
[43,51]. For sliding in vacuum, the friction coef®cient was
on average about 0.03, compared to 0.02±0.04 for singlephase WS2, when its basal planes are oriented parallel to
the ®ction surface [60]. It was found that WS2-type lubrication in vacuum could be realized at sulfur contents as
low as 20 at.%. Coatings with this composition did not
posses crystalline WS2 grains based on the TEM studies.
The formation of hexagonal WS 2 oriented parallel to the
surface, was the result of the friction induced crystallization and re-orientation mentioned earlier.
Based on these observations and wear debris analyses in
Ref. [68], schematics of the tribological mechanisms are
proposed for the WC/DLC/WS2 composite sliding in air
and in vacuum. These mechanisms are presented in Fig. 8.
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
229
Fig. 6. Transmission electron microscope images and corresponding electron diffraction patterns for WC/DLC/WS2 nanocomposite coatings with 15 at.% (a)
and 30 at.% (b) of sulfur.
They were found to be repeatedly switchable, as the environment was cycled between humid air and dry nitrogen,
simulating ambient (humid) $ space (dry) cycling. The
following self-adaptive behavior was found in these tests.
For sliding in humid air, a graphite-like transfer ®lm is
formed by the sp 3 ! sp 2 transformation of the DLC matrix.
This provided lubrication and sealed the WS2 phase from
the oxidizing action of the environment (Fig. 8a). When the
environment was cycled to dry, the graphite-like transfer
®lm experienced unfavorable wear conditions due to the
depletion of intercalating water molecules. The graphitelike material and oxidation products (such as WO3) were
Fig. 7. Raman spectra recorded for friction surfaces of as deposited WC/
DLC/WS2 nanocomposite coating (a), the same coating after 50 000 sliding
cycles in humid air (b), and after 50 000 sliding cycles in vacuum (c).
quickly removed as debris, opening the sealed part of the
composite containing fresh WS2. Under the action of friction, WS2 was crystallized and re-oriented into hexagonal
plates, covering friction surfaces and providing lubrication
(Fig. 8b). In the next humid cycle, the WS2 transfer ®lm was
placed into unfavorable wear conditions, due to water
adsorption and oxidation, and was replaced by the
graphite-like transfer ®lm.
As a result of these mechanisms, a thin lubricating
layer was always present on top of a hard support coating
in both humid and dry environments. Such a con®guration
is ideal for reduction of friction and wear and permits
good coating endurance [69]. Over a million cycles of
operation were achieved with WC/DLC/WS2 coatings in
both 50% RH air and 10 27 Pa vacuum. Wear rate estimates after long duration runs in vacuum were at least as
low as 10 27 mm 3/N/m [62]. The process of surface adaptation to environmental changes was repeated several
times and is demonstrated in Fig. 9, which shows friction
coef®cient recovery in environmental cycling. One rather
unexpected result was an extremely low (0.007) friction
coef®cient recorded for the ®rst few dry nitrogen cycles.
Analyses of the literature on powder lubricants made of
MoS2/graphite and WS2/graphite indicated that this effect
could be a synergism between dichalcogenides and
carbon, discovered more than two decades ago [70±72]
but not yet completely understood.
The structural and chemical self-regulation of WC/DLC/
WS2 tribological surfaces was driven by the action of friction and environment. This adaptive behavior was selfdirected to extend coating endurance in adverse environments, providing an analogy to the `chameleon' protection
concept well known in nature.
230
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
Fig. 8. Mechanisms of a reversible self-adaptation in surface chemistry and structure of WC/DLC/WS2 nanocomposite coatings with the environment change
from humid air (a) to dry nitrogen or vacuum (b).
4. Conclusions
Recently, there has been considerable progress in the
development of superhard nanocomposite structures.
However, their designs do not provide mechanisms for
super-tough properties, since they are, in principle, directed
toward suppression of plastic deformation. While high
hardness is an important attribute of wear-resistant coatings, high toughness and low friction may have similar or
greater importance. In order to improve hard coating
toughness, some ductility should be introduced, which
ideally should be activated at loads above the elastic
strength limit. Thus, it is desirable to have a dual-property
material, which is ordinarily very hard, but switches its
behavior to ductile at extreme loading to prevent brittle
fracture. Since dislocation motion is prohibited in nano-
Fig. 9. Friction coef®cient recovery for a WC/DLC/WS2 nanocomposite
coating in dry $ humid environmental cycling.
composite structures, alternative mechanisms for ductility
should be sought. Grain boundary sliding is one of the most
promising due to the large volume of boundaries available
in nanocomposites.
In this work, TiC/DLC and WC/DLC nanocomposites
were used to show that grain boundary sliding and nanocrack termination mechanisms could provide self-regulation
of surface mechanical properties from hard to ductile. The
transition occurs when the load exceeds the composite elastic strength, preventing brittle fracture and distributing
contact load onto larger areas. As a result, composites
with about 30 GPa hardness `plastically' deformed and
exhibited extremely high scratch toughness.
The WC/DLC super-tough composite was modi®ed by
the introduction of the WS2 phase to provide friction selfadaptation, depending on the operating environment (humid
air, dry nitrogen, or vacuum). Here, the DLC phase was used
for lubrication in humid environments and the WS2 phase
was used for lubrication in dry nitrogen and vacuum. The
action of friction and environment was used to induce reversible structural transformations in initially amorphous DLC
and WS2 phases. This action was also used to self-regulate
between predominantly graphite-like lubrication in humid
air, and predominantly WS2 lubrication in vacuum or dry
nitrogen. The structure-chemistry self-adaptation of the friction surface resulted in low friction coef®cients and long
endurance in all environments and in dry $ humid environmental cycling.
Thus, self-regulation of the surface mechanical response,
structure, and chemistry, was achieved with changes of load
and environment. Such self-regulation is a result of the
proposed `chameleon' concept in the design of wear-resistant coatings, which can considerably expand the applicability and reliability of future hard coatings and tribological
materials.
A.A. Voevodin, J.S. Zabinski / Thin Solid Films 370 (2000) 223±231
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
M. Shin, L. Hultman, S.A. Barnett, J. Mater. Res. 7 (1992) 901.
W.D. Sproul, J. Vac. Sci. Technol. A 12 (1994) 1595.
B.M. Clemens, H. Kung, S.A. Barnett, MRS Bull. 2 (1999) 20.
S. Veprek, S. Reiprich, Thin Solid Films 268 (1995) 64.
S. Veprek, Thin Solid Films 317 (1998) 449.
J.S. Zabinski, M.S. Donley, V.J. Dyhouse, N.T. McDevitt, Thin Solid
Films 214 (1992) 156.
J.S. Zabinski, M.S. Donley, N.T. McDevitt, Wear 165 (1993) 103.
J.S. Zabinski, S.V. Prasad, N.T. McDevitt, Tribology for Aerospace
System, Proceedings of NATO Advisory Group of Aerospace
Research and Development (AGARD) Conference on Tribology for
Aerospace Systems, Sesimbra, Portugal, May 6±7, 1996, AGARD
NATO publication CP 589Canada Communication Group, Hull,
Quebec, 1996, p. 3/1.
J. Musil, Surf. Coat. Technol. 125 (2000) 322.
A.G. Dias, J.H. Van Breda, P. Moretto, J. Ordelman, Proc. 10th Euro.
Conf. on Chemical Vapour Deposition, Venice, Italy, September 10,
1995, J. de Physique, IV (5), 1995, p. C5.831.
J. Musil, H. Polakova, V. Cibulka, Czech. J. Phys. 49 (1999) 359.
S. Veprek, Surf. Coat. Technol. 97 (1997) 15.
C. Mitterer, P. Losbichler, M. Beschliesser, et al., Vacuum 50 (1998)
313.
C. Mitterer, P.H. Mayrhofer, M. Beschliesser, et al., Surf. Coat. Technol. 120±121 (1999) 405.
M. Irie, H. Ohara, A. Nakayama, N. Kitagawa, T. Nomura, Nucl.
Instrum. Methods Phys. Res. B 121 (1997) 133.
J. Musil, P. Zeman, H. Hruby, P.H. Mayrhofer, Surf. Coat. Technol.
120±121 (1999) 179.
J. Musil, H. Polakova, Surf. Coat. Technol. (in press).
J. Musil, H. Hruby, Thin Solid Films (2000) in press.
J. Musil, J. Vlcek, Mater. Chem. Phys. 54 (1998) 116.
M. Benda, J. Musil, Vacuum 55 (1999) 171.
A.A. Voevodin, S.V. Prasad, J.S. Zabinski, J. Appl. Phys. 82 (1997)
855.
M.P. Delplancke-Ogletree, O.R. Monteiro, J. Vac. Sci. Technol. A 15
(1997) 1943.
A.A. Voevodin, J.P. O'Neill, S.V. Prasad, J.S. Zabinski, J. Vac. Sci.
Technol. A 17 (1999) 986.
O.R. Monteiro, M.P. Delplancke-Ogletree, R.Y. Lo, R. Winand, I.
Brown, Surf. Coat. Technol. 94/95 (1997) 220.
H. Dimigen, C.-P. Klages, Surf. Coat. Technol. 49 (1991) 543.
K. Bewilogua, H. Dimigen, Surf. Coat. Technol. 61 (1993) 144.
J.E. Sundgren, B.-O. Johanson, S.-E. Karlsson, Thin Solid Films 105
(1983) 353.
H. Holleck, J. Vac. Sci. Technol. A 4 (1986) 2661.
O. Knotek, F. Lof¯er, G. Kramer, Surf. Coat. Technol. 59 (1993) 14.
S. Veprek, J. Vac. Sci. Technol. A 17 (1999) 2401.
G.E. Dieter, Mechanical Metallurgy, 2nd ed., McGraw-Hill, New
York, 1976.
M. Marder, J. Finberg, Phys. Today 49 (9) (1996) 24.
S.T. Mileiko, Metal and Ceramic Based Composites, Elsevier, New
York, 1997.
J. Karch, R. Birringer, H. Gleiter, Nature 330 (1987) 556.
O.D. Sherby, J. Wadswoth, Prog. Mater. Sci. 33 (1989) 169.
T.G. Nieh, J. Wadsworth, F. Wakai, Int. Mater. Rev. 36 (1991) 146.
T.G. Langdon, Mater. Sci. Eng. A 166 (1993) 67.
C.C. Koch, D.G. Morris, K. Lu, A. Inoue, MRS Bull. 24 (2) (1999) 54.
R.W. Siegel, G.E. Fougere, in: M.A. Otooni, R.W. Armstrong, N.J.
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
231
Grant, K. Ishizaki (Eds.), Grain Size and Mechanical Properties ±
Fundamentals and Applications, Boston, USA, November 27±
December 2, 1995, Materials Research Society Symposium Proceedings, 362, 1995, p. 219.
T.G. Langdon, Mater. Sci. Forum 189/190 (1995) 31.
O.D. Sherby, T.G. Nieh, J. Wadsworth, Mater. Sci. Forum 243 (1997)
11.
J. Schiotz, F.D.D. Tolla, K.W. Jacobson, Nature 391 (1998) 561.
A.A. Voevodin, M.S. Donley, J.S. Zabinski, J.E. Bultman, Surf. Coat.
Technol. 77 (1995) 534.
A.A. Voevodin, M.A. Capano, A.J. Safriet, M.S. Donley, J.S.
Zabinski, Appl. Phys. Lett. 69 (1996) 188.
A.A. Voevodin, J.S. Zabinski, J. Mater. Sci. 33 (1998) 319.
A.A. Voevodin, M.A. Capano, S.J.P. Laube, M.S. Donley, J.S.
Zabinski, Thin Solid Films 298 (1997) 107.
M. Yan, T.G. Langdon, Metal. Mater. Trans. A: Phys. Metal. Mater.
Sci. 27 (1996) 873.
D.J. Schissler, A.H. Chokshi, T.G. Nieh, J. Wadsworth, Acta Metal,
Mater. 39 (1991) 3227.
A.A. Voevodin, J.S. Zabinski, Diam. Rel. Mater. 7 (1998) 463.
A.A. Voevodin, J.P. O'Neill, J.S. Zabinski, Thin Solid Films 342
(1999) 194.
A.A. Voevodin, A.W. Phelps, M.S. Donley, J.S. Zabinski, Diam. Rel.
Mater. 5 (1996) 1264.
R.H. Savage, J. Appl. Phys. 19 (1948) 1.
M.N. Gardos, Proceedings of the First World Tribology Congress on
New Directions in Tribology, September 8±12, 1997, London,
Mechanical Engineering Publications, London, 1997, p. 229.
H. Zaidi, T.L. Huu, D. Palmer, Diam. Rel. Mater. 3 (1994) 787.
R.L. Fusaro, Lubr. Eng. 3 (1995) 182.
P.D. Fleischauer, Proceedings of the First World Tribology Congress
on New Directions in Tribology, September 8±12, 1997, London,
Mechanical Engineering Publications, London, 1997, p. 217.
W.O. Winer, Wear 10 (1967) 422.
S.V. Prasad, J.S. Zabinski, N.T. McDevitt, Tribol. Trans. (1995) 38.
C. Pritchard, J.W. Midgley, Wear 13 (1969) 39.
S. Prasad, J. Zabinski, Nature 387 (1997) 761.
P.D. Fleischauer, ASLE Trans. 27 (1984) 82.
S.V. Prasad, J.S. Zabinski, J. Mater. Sci. Lett. 12 (1993) 1413.
M.R. Hilton, R. Bauer, P.D. Fleischauer, Thin Solid Films 188 (1990)
219.
J. Moser, F. Levy, J. Mater. Res. 8 (1993) 206.
J.S. Zabinski, M.S. Donley, S.V. Prasad, N.T. McDevitt, J. Mater. Sci.
29 (1994) 4834.
D.N. Dunn, K.J. Wahl, I.L. Singer, in: N.R. Moody, W.W. Gerberich,
N. Burnham, S.P. Baker (Eds.), Fundamentals of Nanoindentation and
Nanotribology, San Francisco, USA, April 13±17, 1998, MRS Symp.
Proc., 552, 1998, p. 451.
A.A. Voevodin, J.P. O'Neill, J.S. Zabinski, Surf. Coat. Technol. 116±
119 (1999) 36.
A.A. Voevodin, J.P. O'Neill, J.S. Zabinski, Tribol. Lett. 6 (1999) 75.
K. Holmberg, H. Ronkainen, A. Matthews, Proc. 1st World Tribol.
Cong. on New Directions in Tribology, September 8±12, 1997,
Mechanical Engineering Publications, London, 1997, p. 251.
F.G. Fisher, A.D. Cron, R.G. Muber, Publication of National Lubrication Grease Institute 46, NLGI, 1982, p. 190.
A.M. Petlyuk, L.N. Sentyurikhina, O.V. Lazovskaya, T.P. Yukhno,
Trenie Iznos (J. Fric. Wear) 8 (1987) 740.
M.N. Gardos, Tribol. Trans. 31 (1988) 214.