INSTITUTE OF PHYSICS PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS

J. Phys. D: Appl. Phys. 39 (2006) R311–R327 doi:10.1088/0022-3727/39/18/R01

TOPICAL REVIEW

Tribology of diamond-like carbon films:recent progress and future prospectsAli Erdemir1 and Christophe Donnet2

1Energy Technology Division, Argonne National Laboratory, Argonne, IL 60439, USA2University Jean Monnet and University Institute of France, Laboratoire Traitement du Signal etInstrumentation, UMR 5516, Saint-Etienne, France

Received 16 February 2006, in final form 23 March 2006Published 1 September 2006Online at stacks.iop.org/JPhysD/39/R311

AbstractDuring the past two decades, diamond-like carbon (DLC) films have attracted anoverwhelming interest from both industry and the research community. These filmsoffer a wide range of exceptional physical, mechanical, biomedical and tribologicalproperties that make them scientifically very fascinating and commercially essential fornumerous industrial applications. Mechanically, certain DLC films are extremely hard(as hard as 90 GPa) and resilient, while tribologically they provide some of the lowestknown friction and wear coefficients. Their optical and electrical properties are alsoextraordinary and can be tailored to meet the specific requirements of a givenapplication. Because of their excellent chemical inertness, these films are resistant tocorrosive and/or oxidative attacks in acidic and saline media. The combination of sucha wide range of outstanding properties in one material is rather uncommon, so DLC canbe very useful in meeting the multifunctional application needs of advancedmechanical systems. In fact, these films are now used in numerous industrialapplications, including razor blades, magnetic hard discs, critical engine parts,mechanical face seals, scratch-resistant glasses, invasive and implantable medicaldevices and microelectromechanical systems. DLC films are primarily made of carbonatoms that are extracted or derived from carbon-containing sources, such as solidcarbon targets and liquid and gaseous forms of hydrocarbons and fullerenes.Depending on the type of carbon source being used during the film deposition, the typeof bonds (i.e. sp1, sp2, sp3) that hold carbon atoms together in DLC may vary a greatdeal and can affect their mechanical, electrical, optical and tribological properties.Recent systematic studies of DLC films have confirmed that the presence or absence ofcertain elemental species, such as hydrogen, nitrogen, sulfur, silicon, tungsten, titaniumand fluorine, in their microstructure can also play significant roles in their properties.The main goal of this review paper is to highlight the most recent developments in thesynthesis, characterization and application of DLC films. We will also discuss theprogress made in understanding the fundamental mechanisms that control their veryunique friction and wear behaviours. Novel design concepts and the principles ofsuperlubricity in DLC films are also presented.

(Some figures in this article are in colour only in the electronic version)

1. Introduction

Carbon is one of the most abundant elements in our planet.It is the sixth most common element and exists in 94% ofall known substances. The vast field of organic chemistry

is mainly based on carbon. Undoubtedly, it is one of themost important building blocks for many chemicals, drugsand nutritional products on which our well-being and modernlife style depend. Carbon is also the essential ingredient ofnumerous key engineering materials possessing exceptional

0022-3727/06/180311+17$30.00 © 2006 IOP Publishing Ltd Printed in the UK R311

Topical Review

0

20

40

60

80

100

Ha

rdn

ess

(G

Pa)

0.1 0.3 0.5 0.7

COF

(t)a-C

(t)a-C:H

Hard coatings

Superhard nc-coatingsDiamond

0

20

40

60

80

100

Ha

rdn

ess

(G

Pa)

0.1 0.3 0. .7

COF

(t)a-C

(t)a-C:H

Hard coatings

Superhard nc-coatingsDiamond

(t)a-C

(t)a-C:H

Hard coatings

Superhard nc-coatingsDiamond

Figure 1. A schematic representation of hardness and coefficientsof friction (COF) of carbon-based and other hard coatings.

properties. While some of these are very soft (graphite,polymers, plastics, etc), others are extremely hard and resilient(such as natural and synthetic diamonds and various carbides).Most of the recently discovered nanomaterials (fullerenes,nanotubes, nano-onions, nanofibres, etc) are also carbon-based, and they are currently being used in the fabrication of amyriad of nano-to-micro-scale devices. Carbon is also used inthe synthesis of numerous hard coatings including diamond,diamond-like carbon (DLC), carbon nitride, transition metalcarbides and boron carbide. Because of their superhardness,superhigh thermal conductivity and superlow friction, someof these coatings have attracted tremendous interest in recentyears from both the industrial and scientific communities, andtheir friction and wear properties have, in particular, been thesubject of numerous scientific studies [1–10].

Other carbon forms (such as graphite, graphite fluoride,carbon–carbon composites and glassy carbons) are alsovaluable as low-friction engineering materials, and they areoften used as solid lubricants by the industry to combatfriction and wear under conditions where the uses of liquidlubricants are neither possible nor desirable [1, 5–8]. Recentlydeveloped carbon-based bulk micro- and nano-compositesand their coatings represent a new class of smart materialsproviding impressive friction, wear and corrosion propertieseven at high ambient temperatures [1, 11].

Among the many properties of an engineering material,hardness and stiffness may play the most important roles inits ability to carry the load and hence its wear resistance.In general, materials with high hardness and stiffness havehigh wear resistance. Diamond represents a prime example,with its highest known hardness and extreme wear resistance.However, there is no universal correlation between hardnessand friction coefficients of different materials and/or coatings[12]. Nonetheless, figure 1 classifies various coatings withrespect to their hardness and friction characteristics to point outthe fact that most carbon films are capable of providing not onlyhigh hardness but also low friction. In particular, DLC filmsappear to provide the broadest range of hardness and frictionvalues, while some of the recently developed nanocompositecoatings are able to provide superhardness but lack lubricityor low friction [12].

In brief, the field of carbon-based materials and coatingshas enjoyed strong and growing interest from all kinds ofscientific and commercial disciplines. In particular, diamondand DLC coatings have attracted the most attention in recentyears, mainly because they offer a wide range of exceptional

properties for a wide range of demanding applications.Accordingly, in this review, we attempt to highlight some ofthe most important developments in the field of DLC filmsin general and their tribology in particular. The present stateof the art in scientific research and industrial practices thatinvolve DLC films is also surveyed. A relatively short reviewpaper such as this cannot address all the important facets ofthese films. Hence, we will focus our attention mainly on themost important developments of the last decade or so. Wewill also summarize the recent research on their friction andwear mechanisms. Several excellent review papers and bookchapters already exist on the earlier developments concerningDLC and other carbon-based films [1, 2, 4, 13, 14]

2. Historical perspective

2.1. Inception and early studies

Historically, the earliest research on DLC films can be tracedback to 1953. Even though Eisenberg and Chabot are oftenconsidered as the true pioneers of this technology (mainlybecause of their comprehensive studies in the early 1970s [15],Schmellenmeier had produced such carbon films back in1953 and hence he was probably the earliest pioneer of DLCtechnology [16]. From the start of their studies, these scientistshave immediately realized that these carbon films were ratherunique and possessed some unusual mechanical and electricalproperties. For one thing, they were mechanically veryhard and resistant to scratching; they also possessed highdielectric constants, high index of refraction and excellentoptical transparency. Furthermore, these early DLC films werechemically inert and hard to remove or etch out from coatedsurfaces by dipping into strong acidic solutions.

During the mid-1970s, Holland et al and a few otherresearchers also developed an interest in DLC films. Theseresearchers were able to derive DLC from a number ofgaseous hydrocarbon sources by applying an r.f. bias to thesubstrate materials and thus creating a plasma [17,18]. In thisrespect, their deposition process was somewhat different fromthe one that Eisenberg and Chabot had used. Weissmentaland his co-workers were the very first group of scientistswho performed extensive electron microscopy and electron-energy loss spectroscopy work on DLC films to elucidatetheir structural and chemical nature [19]. Some of the earlierresearchers had thought that these films were perhaps made ofcrystalline diamond, but the microscopic work by Weissmentalet al proved otherwise; these films were made of amorphouscarbon.

Despite their several attractive properties, DLC filmsdid not draw much attention throughout the 1970s and evenuntil the mid-1980s. This neglect may, in part, have beendue to the fact that around the same time, the creation ofcrystalline diamond films using low-pressure chemical vapourdeposition (CVD) had been announced, and most researcherswere working on the hot topic of diamond but not on somethinglike it [20, 21]. However, inherent difficulties plagued thelarge-scale production of diamond as thin films or free-standinglarge crystals that everybody was dreaming of producingat reasonable costs. Nevertheless, these dedicated researchactivities resulted in the development of high-quality diamondfilms that are now used in key industrial applications [22, 23].

R312

Topical Review

During the 1980s, only a few systematic studies weredirected towards the production and in-depth structuraland/or chemical characterization of DLC films. Most ofthe mechanical and tribological characterization works werecarried out during the 1990s. Some of the very first kinds ofDLCs produced by Eisenberg and Chabot [15] were tested fortheir durability as wear-resistant coatings on eyeglasses andcutting edges of steel blades. Limited test results confirmedthat these films were capable of substantially improving thedurability of both the eyeglasses and the steel blades [24].

In a classic review paper from 1985, Arnoldussen andRossi from IBM announced potential applications of DLCfilms as protective overcoats for magnetic recording media[25]. Such a possibility had already been explored to someextent by King in 1981, and some interesting results werereported by him at that time [26]. Specifically, in hisexperimental studies, the newly developed DLC-coated discmedia exhibited much superior wear resistance comparedwith the other types of overcoats that were being used byindustry. Based on his studies, King concluded that theseDLC-based recording media would have been a significantstep forward in advancing magnetic recording technology, andobviously he was right. During the late 1980s, numerousother studies specifically focused on the development andtribological characterization of DLC-type carbon overcoats formagnetic recording media (see, for example, [26, 27]). Fromtime to time, several excellent review papers have appeared onthe uses of DLC as an overcoat for hard disc drives. For morereading on this subject [28–30] are particularly interesting.In this paper, we will not cover this subject any further,mainly because several excellent review papers address boththe scientific and application-oriented issues in various carbonovercoats for magnetic storage media. The current trend is tofurther reduce the film thickness and the gap between the headand disc, with an ultimate goal of achieving contact recordingthat can potentially revolutionize the field.

2.2. Systematic studies

Most of the systematic studies on sliding friction and wearbehaviour of DLC films were carried out during the 1990s.Only a few noteworthy studies appeared during the 1980s, andthey were mostly carried out by Enke and his coworkers [31].In their studies, the friction and wear coefficients of DLC filmswere indeed confirmed to be low, but they were also found tobe sensitive to the test environments. Compared with graphiteand diamond, some of the very first DLC films used in theirstudies exhibited relatively poor tribological performance inhumid test environments but impressive performance in inertor dry test media. This finding was in contrast to the frictionalbehaviour of both graphite and diamond, which happen toprovide low friction and wear in humid test environments,but relatively high friction and wear in inert gases or vacuum[8, 32, 33].

During the 1980s, only a few other tribological studiesspecifically focused on the friction and wear of DLC films[34,35]. During the 1990s, the interest in the tribology of DLCfilms gained momentum. As a result, several key publicationsemerged in a number of archival journals. Papers published byGrill et al [36], Miyoshi and his co-workers [37], Ronkainen

0

50

100

150

200

250

Nu

mb

er o

f p

ub

licat

ion

s

71-8

019

9019

9219

9419

9619

9820

0020

0220

04

Year

DLC

DLC for tribology

Figure 2. Number of publications per year on DLC coatings and ontheir friction and wear.

et al [38], Erdemir et al [39] and Donnet et al [40] are primeexamples of such systematic studies on the tribology of DLCfilms. All and all, these early studies laid the foundationfor more in-depth studies and reinforced the notion that DLCfilms are indeed unique and thus warrant further study for abetter understanding of their friction and wear mechanisms.Figure 2 summarizes the number of scientific papers per yearthat have been published on DLC films since the early 1980s.Again, it is interesting to see that over the years, scientificinterest in these films has grown tremendously. They arenow considered as one of the most important tribologicalcoatings in numerous application fields. Excellent reviewsof work done on DLC films over the past several years areprovided in [1, 2, 4, 13, 14, 35, 41, 42]. In particular, a recentcomprehensive review by Robertson is an excellent source forfurther information on various types of DLC films and theirproperties [42].

2.3. Practical applications

Since their discovery back in 1971, DLC films havecome a long way to become one of the most valuableengineering materials for a number of industrial applications,including microelectronics, manufacturing, transportation andbiomedical fields. Until the mid-to-late 1990s, very fewapplications took advantage of the unique properties of DLCfilms. In fact, except for the magnetic storage media, DLCfilms were not used in large volumes by industry. A fewcompanies tried DLC films on eyeglasses and laser barcodescanners to improve their resistance to abrasive wear and/orscratching. Over the years, several new versions of DLC filmshave been developed, and with the introduction of industrial-scale, more-robust coating systems, the production of high-quality DLC films has become rather easy and inexpensive.In the late 1990s, DLC films were used extensively in razorblades and fuel injector systems of diesel engines. Overthe past few years, researchers have made great strides incontrolling film chemistry and hence properties. Current DLCfilms are highly optimized and hence have the capacity to meetthe increasingly more stringent application requirements ofnumerous mechanical systems [1, 10, 42–44].

At present, high-quality DLC films are readily availablefrom commercial sources. Some of these DLC coatings are

R313

Topical Review

Activated reactive evaporation Cathodic arc deposition

Bias sputter depositionIon-beamassisted

deposition

Dual ion-beam sputtering

Ion plating

Figure 3. Schematic representation of various deposition processes that can be used to deposit DLC films.

extremely hard and resilient, while others are relatively soft butcan provide some of the lowest friction and wear coefficients.Films that contain unique crystalline nanostructures and/ornano-phases are also available and have the ability to meet theincreasingly more stringent application conditions of advancedmechanical devices. These multifunctional nanocompositeDLC films are now routinely produced by both CVD andphysical vapour deposition (PVD). They are currently beingused or tested for numerous applications, ranging from razorblades to microelectromechanical systems, from engine partsto articulated hip and knee joints, from bearings to machinetools and dies [1, 4, 43, 44].

3. Synthesis and classification of DLC films

3.1. Synthesis of DLC films

Currently, several kinds of PVD and CVD methods canbe used to deposit DLC films. The range of depositiontemperatures is wide, from sub-zero to 400 ◦C. Dependingon the type of deposition method being used, the range ofdeposition pressure, bias voltage, etc may also be variedover broad ranges [42]. Such a high degree of flexibility indeposition parameters is not always feasible with other coatingtypes, including crystalline diamond or transition metal nitrideand carbide films, all of which may require narrow rangesof deposition parameters for an ideal microstructure and/orchemical stoichiometry. Another feature that makes DLC filmsunique is that they can be deposited on all kinds of substrate

materials. However, the ability of DLC films to establish strongbonding or adhesion can vary widely with the chemical natureof these substrate materials. For most tribological uses, DLCfilms must attain strong bonding to their substrates; otherwise,they can prematurely fracture and delaminate from the surfaceunder the influence of high normal and/or shear forces thatdevelop during sliding contacts [45].

Strong interfacial bonding or adhesion can be attainedeasily between DLC and carbide- and silicide-formingsubstrates (such as Si, Ti, W and Cr). The adhesion of DLCcoatings to other metallic and ceramic substrates may not beas strong but can be improved by the deposition of an initialbond layer on these substrates prior to DLC deposition. Thesebond layers are typically selected from those elements that areknown to be strong carbide- or silicide-formers such as Si, Ti,Cr, W and Nb. These elements can chemically react with theatoms of the substrate materials and thus insure strong bonding.The deposition of these interface layers is ideally done inthe same deposition chamber and before the start of DLCdeposition. Such a practice minimizes the introduction of pointdefects and chemical impurities between the superimposedlayers and allows high precision control of the entire depositionmethodology. Figure 3 shows schematics for some of theplasma-based processes that can be used in the deposition ofDLC films.

In the gas discharge plasmas of the PVD and CVDprocesses mentioned above, usually a hydrocarbon gas (suchas methane or acetylene) is used as the precursor for carbon.Films derived from such hydrocarbon gases contain not only

R314

Topical Review

HydrogenMolecule

HydrogenAtom

BondedHydrogen

Figure 4. Molecular dynamic simulation of atomic structure of ahydrogenated DLC film.

carbon but also considerable amounts of hydrogen in theirmicrostructures, and they are often referred to as hydrogenatedDLC films. Using the same deposition processes, one canalso deposit super-hydrogenated DLC films by establishinghigher than normal hydrogen-to-carbon ratios. For example,introduction of more hydrogen gas into discharge plasmasduring film growth can lead to the formation of highlyhydrogenated DLC films that contain more than 40 at. %hydrogen [3]. Compared with hydrogen-free DLCs, thesefilms are relatively soft but exhibit some of the lowestfriction and wear coefficients, as will be discussed later.Researchers have not yet experimentally determined theatomic-scale structures of these films, but researchers haveattempted to simulate them by a variety of computationaltools. Figure 4 is a molecular dynamic simulation of theatomic structure of a typical hydrogenated DLC film havinga network of carbon atoms with three- and four-fold atomiccoordination and different forms of hydrogen within theirstructure.

Solid carbon materials (such as graphite, glassy carbonor carbon–carbon composites) can also be used as carbonsources in the deposition of DLC films. These carbon sourcesare mostly used in cathodic arc-PVD, laser ablation (or pulselaser deposition, PLD), ion-beam assisted deposition andmagnetron sputtering processes [19, 46–49]. The DLC filmsproduced by conventional sputter deposition may contain largeamounts of sp2-bonded carbon atoms, and hence they tendto be much softer than those DLCs deposited by arc-PVDand PLD methods. However, in recent years, closed-fieldunbalanced magnetron sputtering, filtered-cathodic arc andPLD methods have become increasingly popular, especiallyfor the production of super-dense, ultra-thin (a few nanometresthick) DLC films, mainly for use in magnetic hard discapplications. These processes enable larger amounts ofsp3-bonding and hence superhardness, which are necessaryfor superior tribological performance and longer durabilityeven at very small thicknesses [48–50]. In the past, nano-to-micrometre size particles generated during deposition ofsuch films used to be a major problem, but with the use

of advanced magnetic filtering systems, these problems havebeen overcome, and the latest films are dense, uniformand almost particle free. Because of their percentage ofhigh sp3-bonded carbon atoms (up to 80%), DLC filmsproduced by arc-PVD and PLD methods are often referredto as tetrahedral amorphous carbon [51]. They possesssuper-high hardness and stiffness, but like other DLC films,they are thermally insulating and could be made opticallysemitransparent.

The development of femtosecond PLD processes offersnew possibilities for the deposition of more advanced DLCfilms. The use of femtosecond pulses (typically 150 fs)with very high power densities (up to 1013 W cm−2) canablate carbon atoms with much higher velocities and hencekinetic energies (up to 1 keV), compared with the conventionalnanosecond laser ablation processes [52]. The high kineticenergy facilitates the implantation of some of the impingingcarbon atoms into the substrate materials, thus ensuringstrong adhesion between substrate and DLC. These coatingsconsist of large proportions (up to 70 at. %) of sp3-bondedcarbon atoms, exhibit rather low internal stresses and havefewer micro-particles or droplets on their surfaces, comparedwith the DLC films deposited by conventional PLD andcathodic arc-PVD methods [53]. DLC films deposited byfemtosecond PLD (containing 70% of sp3 carbon) haveexhibited friction coefficients of about 0.1 and wear rates of1.6 × 10−8 mm3 N−1 m−1 when tested in ambient air under acontact pressure of 0.5 GPa [53, 54].

One of the most promising features of the femtosecondPLD is associated with its unique ability to depositnanostructured and doped or nano-alloyed DLC films.Recently, this technique has been used successfully to depositmetal-doped DLC films, with a very precise control of themetallic concentration within the carbonaceous matrix [55].This technique can also be used in the deposition of DLCfilms consisting of metallic clusters or nano-phases in the10–200 nm range [56]. Furthermore, by alternating ablationbetween carbon and metal targets, one can also produce nano-layered DLC films that are made of alternating layers ofmetallic and carbon phases [57]. Such coating architecturesmay combine a range of attractive physical properties such ashigh thermal and/or electrical conductivity, low friction andwear and supertoughness.

Hydrogenated DLC films are generally soft and do notpossess very high internal stresses [45]. Hence they canbe deposited rather thickly without creating any problemsregarding adhesion or cracking. However, hydrogen-free DLCfilms are very hard and often exposed to high levels of internalstresses. Their thickness is generally limited to 1 µm or so.Doping with certain elements allows deposition of thickercoatings with lesser degree of internal stresses, but the hardnessof the film may also decrease [42]. As mentioned above, withthe use of the femtosecond PLD method, one can deposit verythick hydrogen-free DLC films without a high degree of stressbuildup [53]. Whatever the technique, the fraction of sp3

bonding is maximized for ion-dominated processes with ionenergies ≈100 eV. The specific properties of these DLC filmsare mainly due to the high kinetic energies of the impingingparticles, while film growth is governed by a subplantationprocess instead of conventional condensation, as in amorphous

R315

Topical Review

Figure 5. Nano-wear resistance of various DLC films produced bydifferent processes. The effect of nitrogen (N) content on wear isalso shown for DLC films produced by filtered cathodic arc process(courtesy of H Hyoda, Fujitsu Co.).

(a) CH films [42,51]. Consequently, the substrate temperatureduring film deposition is a critical parameter. Above atransition temperature of 250 ◦C, a noticeable decrease in thehigh sp3 content and density is observed. This transitiontemperature decreases with increasing ion energy during filmgrowth.

Because the film structure is a strong function of theenergy of the impinging species, this ability depends onseveral plasma parameters, mainly the bias voltage and thegas pressure. The bias voltage is primarily controlled bythe RF or pulse-dc power sources, but the desirable rangeof gas pressures in the deposition chamber is obtained by aseries of gas flow metres. The amounts of sp3-bonding andhydrogen in DLC films are affected by the average impactenergy of the impinging atoms. More precisely, at the lowestimpact energies, the gaseous precursor is not sufficientlydecomposed, and as a result, a polymer-like carbon film witha predominance of =CH2 groups is generally obtained. Atintermediate impact energies, the hydrogen content is reducedand sp3 type bonding is favoured, thus leading to the so-called‘diamond-like’ qualities. However, if the impact energiesbecome too high, a graphite-like carbon network is obtained,mainly because of an increase in the disordered sp2-likebonding.

3.2. Classification of DLC films

Most DLC films are structurally amorphous and, as discussedabove, they can be synthesized by plasma-based PVD and CVDmethods. Depending on the deposition methods and carbonsources used, the structural chemistry of the resultant filmsmay also differ substantially, and such differences may, in turn,lead to large variations in their properties. Figure 5 comparesthe nano-wear characteristics of various DLC films producedby different deposition methods. As is clear, depending onthe deposition method used, the wear resistance of DLC filmsdiffers substantially.

Structurally, DLC films are made of sp2- and sp3-bondedcarbon atoms. Trace amounts of sp1 bonding are also feasibleunder certain deposition conditions, but the bulk of the bondsare predominantly sp2 and sp3 [42]. The relative amount

Figure 6. Ternary phase diagram for various DLC films withrespect to their sp2, sp3 and hydrogen contents.

Table 1. Structure, composition and properties of two forms ofDLC.

Composition and properties ta-C a-CH

Hydrogen content (at.%) < 5 20–60sp3(%) 5–90 20–65Density (g cm−3) 1.9–3.0 0.9–2.2Thermal stability (◦C) < 600 <400Optical gap (eV) 0.4–1.5 0.8–4.0Electrical Resistivity (� cm−1) 102–1016

Index of refraction 1.8–2.4Compressive stress (GPa) 0.5–5Hardness (GPa) <80 <60Young modulus (GPa) <900 <300

of sp2 versus sp3-bonded carbon atoms varies a great dealfrom one DLC film to another. Films with a high proportionof sp2-bonded carbon atoms tend to be relatively soft andbehave more like graphite during tribological tests, while filmswith more sp3-bonded carbons are more like diamond, andhence they are superhard and provide impressive tribologicalproperties [50]. If the films are derived from a hydrocarbonsource (such as acetylene or methane) then large amountsof hydrogen may also be present within their structures.Figure 6 represents the various DLCs and other carbon filmsin a ternary phase diagram. This diagram was proposed byFerrari and Robertson, who have performed perhaps the mostcomprehensive structural and chemical studies on these filmsusing spectroscopic techniques [51].

In the ternary diagram shown in figure 6, the regionsof various DLC films are clearly identified, and based onthe fraction of sp3 bonds and hydrogen content, the filmsare classified into several kinds, ranging from hydrogenatedamorphous carbons (or a-C : H) to tetrahedral amorphouscarbon (or ta-C [51]. Table 1 summarizes the basic propertiesof these carbon films with respect to their sp2, sp3 andH contents. Such a wide range of film structures andcompositions and the diversity of methods available for theproduction of DLC films are not possible with other typesof hard coatings. In addition to the ones shown in thephase diagram of Ferrari and Robertson, there exist severalmore DLCs consisting of different kinds of alloying elements,discrete compound phases in a nanocomposite and superlatticeor nano-layered coating architectures, as will be discussedlater.

R316

Topical Review

4. Friction and wear of DLC films

The unique tribological behaviour of DLC films may varya great deal from one type to another. Test conditions andenvironments can also play a major role in their friction andwear. In particular, the friction values reported for various DLCfilms span the range 0.001–0.7, which probably represents thewidest range of friction among all other materials or coatings[1]. As far as wear performance is concerned, certain DLCfilms are very soft and easily scratchable, while others areextremely hard and resistant to wear (the normalized wearrates of such films are as low as 10−11 mm3 N−1 m−1). Sucha large disparity in friction and wear properties of DLC filmsappears to stem from a complex combination of intrinsic orfilm-specific factors and extrinsic or test-condition-specificfactors. Intrinsically, the friction and wear behaviours of thesefilms are strongly affected by their chemical and structuralnature. Extrinsically, the tribotest conditions (includingmaterial parameters, such as the nature of the substrate andcounterface materials, contact pressure, nature of motion,speed, ambient temperature during sliding test and the natureand/or chemistry of the test environment) may play significantroles. Therefore, the following sections will provide a detaileddiscussion of such factors.

4.1. Major causes of friction in DLC films

As with most other materials and coatings, the factors thatinfluence the frictional behaviour of DLC films are manyand may vary from one type of DLC to another. Asmentioned earlier, it is possible to divide these factors intotwo broad categories: intrinsic factors and extrinsic factors.Intrinsically, the degree of sp2 versus sp3 bonding as wellas the relative amounts of hydrogen and/or other alloyingelements in the structure or on the sliding surfaces of DLCfilms can have a strong effect on their friction and wearbehaviours. Extrinsically, the frictional behaviour of thesefilms can be affected by the extent of chemical, physical andmechanical interactions between the rubbing surfaces of DLCfilms and their surroundings. The physical roughness of thesliding surfaces of DLC films can also have a strong influenceon friction and wear. The contribution from each type ofinteraction to overall friction can vary a great deal and maydepend strongly on the specific test conditions or parameters(load, speed, type of motion, distance, etc) that are beingemployed during the sliding tests. The chemical nature of testenvironments, ambient temperature and the type of counterfacematerial that is being rubbed against the DLC films can alsoplay major roles. The presence or absence of a transfer film onthe sliding surfaces of counterface materials and the physicaland/or chemical nature of such films can also influence friction.

4.1.1. Physical and mechanical interactions. Physicallyrough surfaces can certainly cause high friction and severewear losses in most sliding contacts. Specifically, if thesliding surfaces are very rough, a high level of mechanicalinterlocking can take place between surface asperities and leadto high frictional losses (especially during the run-in or initialstages of sliding tests). A prime example is the inherentlyrough surface finish of microcrystalline diamond films, which

0

0.05Fri

ctio

n C

oeff

icie

nt

0.1

0.15

0.2

0 100 200 300 400 500 600

Time (s)

Ra=20 nm

Ra=0.7 nm

Figure 7. Effect of surface roughness on initial and steady-statefriction coefficients of highly hydrogenated DLC films.

can cause high friction and severe wear losses during slidingcontacts. In fact, recent systematic studies have demonstratedthe existence of an almost linear correlation between surfaceroughness and friction and wear coefficients of sliding diamondsurfaces [1, 58]. In general, it was found that the higher thesurface roughness, the greater the friction and wear losses.

Most DLC films are structurally amorphous and hencethey can closely mimic the original surface roughness of thesubstrate materials. Roughness increase due to preferredgrowth orientations is not possible with amorphous DLC filmsbut is possible with crystalline diamond films. When depositedon highly polished surfaces (such as Si wafers), the surfacefinish of DLC films is extremely smooth; however, if depositedon a lapped or rough ground surface, their surface finish is veryrough. A rough surface finish may also result from the typeof deposition process that is used. For example, the filmsdeposited by conventional PLD and arc-PVD methods mayresult in numerous nano/micro-particles and/or droplets beingejected from the solid carbon sources and then deposited onthe surface of the substrate materials along with the actualDLC films. The presence of such microparticles on slidingsurfaces can cause high friction and wear losses, especiallyduring the initial run-in periods of sliding tests. In general,regardless of the type, the smoother the DLC films, the lowertheir friction and wear coefficients (provided that other testsparameters are kept the same). Figure 7 demonstrates that ifthe DLC films are deposited on an originally rough or lappedsurface, the measured friction coefficients are much higher thanfilms produced on a highly polished surface.

4.1.2. Adhesive interactions. Apart from the adverse effectsof surface roughness on friction and wear, the extent ofchemical and/or adhesive interactions between sliding DLCsurfaces may also strongly influence their friction. In thesefilms, the adhesive interactions can primarily result fromseveral types of bonding. Among others, covalent bondinteractions between unoccupied or dangling σ -bonds ofsliding carbon film interfaces can account for a significantsource of adhesion. Covalent bonding is the strongest type incarbon-based materials, and if it is not taken care of then very

R317

Topical Review

0.2

0.3

0.4

0.

Fri

ctio

n C

oe

ffic

ien

t

5

0.6

0.7

0.8

0.9

0 2000 4000 6000 8000 1 104

1.2 104

Time (s)

Dry Nitrogen

Humid airin

Test Conditions: Load: 10 NSpeed:0.5 m/s

Temperature: 22-23oC

Moist Air In

Figure 8. Effect of moist air on the frictional behaviour of a ta-Cfilm. Note that its friction is very high in dry nitrogen but dropsrapidly as soon as humid air is let into the test chamber.

strong adhesion and hence friction may result between suchsliding surfaces. For example, some papers report frictioncoefficients of more than 1 in ultra-high vacuum (UHV) forboth the bulk and thin-film versions of diamond [59,60]. Suchbonding may also occur between sliding diamond surfaces atvery high temperatures [32, 61]. Apparently, under UHV andhigh-temperature sliding conditions, the surface adsorbatesthat normally pacify those σ -bonds are either absent ormechanically and thermally removed from the sliding surfaces.The covalent bonds that become free and active then interactwith each other and give rise to the very high frictioncoefficients that were reported earlier [59–61]. The same istrue for hydrogen-free DLC films. In particular, tetrahedralamorphous carbon (or ta-C) type DLC films possess a highdegree of sp3-bonded carbon in their structure. As shown infigure 8, the friction coefficient of such films in dry nitrogen(with near zero relative humidity) is rather high, i.e. about 0.7.However, when moisture is introduced into the test chamber,their friction coefficient drops sharply to values less than 0.3.The high friction in dry nitrogen is largely attributed to the highlevel of covalent σ -bond interactions between sliding contactsurfaces, while the lower friction in moist air is mostly due tothe passivation of these σ -bonds by oxygen and water, whichreduce the extent of adhesive interactions between these slidingsurfaces.

The van der Waals forces, π–π∗ interaction, capillaryforces and electrostatic attractions may also be present betweensliding DLC surfaces and can certainly increase adhesion and,hence, cause friction. Ionic, metallic and magnetic bondinteractions do not exist in carbon films. The weakest bonding,the van der Waals force, exists between most surfaces broughtinto very close proximity of each other. Little lateral force isneeded to overcome this bond. The lamellar sheets of MoS2

and a few other solid lubricants are held together by theseforces, and that is why such solids are easy to shear and henceself-lubricate.

Among others, π–π∗ interaction between sliding DLCsurfaces may account for significant adhesive bonding. Thistype of interaction is relevant to crystalline graphite, and hence

it plays a major role in the friction and wear behaviour. Theseverity of π–π∗ interaction is reduced by the presence ofwater and a few other molecules in test chambers [8]. Indry test environments, the friction coefficient of graphite ismore than 0.3; while in moist air, it is only around 0.15. DLCfilms with numerous sp2-bonded carbon atoms and withoutany hydrogen (specified by a-C in the ternary diagram infigure 6) behave similarly to graphite in tribological testing.Hard ta-C or hydrogen-free DLC films generally exhibit lowerfriction (i.e. 0.1–0.2) in humid air. In ta-C films, relativelyhigh friction can cause a local shear-induced graphitizationat nano-to-micro-scales, as described below in section 4.1.5.Consequently, it is not surprising that the friction coefficientof ta-C films decreased with increasing humidity [62] becausewater molecules are known to intercalate between graphitelayers and facilitate their slip over each other. In contrast,the friction of a-CH films generally increases with humidity[63–65]. Figure 9 summarizes some of the bond types andtheir energies with respect to the frictional behaviour of DLCfilms.

4.1.3. Tribochemical interactions. The sliding contactsurfaces of most carbon films are chemically very stable and,hence, are generally inert towards outside species under staticconditions. They do not normally enter into major chemicalinteractions with liquids and/or solid materials when broughtinto direct contact. However, under the influence of dynamicsliding contacts, these surfaces may interact with counterfacesand with the gaseous molecules (such as water molecules,oxygen and hydrocarbons) in their surroundings. Dependingon the extent and nature of such interactions, steep fluctuationsmay be observed in DLC frictional behaviour [39, 40, 62–65].This finding clearly suggests that a gas–surface interactionis feasible and perhaps the main cause of the fluctuations infriction. Some of the gaseous species are highly polar (suchas water molecules), and they can physically interact withthe surface carbon atoms of DLC films to form a layer ofphysisorbed species. The rate of formation of such layersis thought to be very high; even if these layers are removedmechanically during sliding, they can replenish or re-formquickly and change the frictional behaviour of these films[63, 65].

Systematic studies by Heimberg et al [63] and Dickrellet al [65] have confirmed that the velocity dependence of thefriction coefficients of a-C : H films was indeed due to stronggas–surface interactions. From a series of tests in controlledenvironments, these authors concluded that the longer theexposure time between subsequent sliding passes, the higherthe friction of DLC films [63, 65]. These initial findings werefurther corroborated by recent model experiments in whicha clear dependence on exposure time as opposed to velocitywas found. Recent efforts to model the transient-to-steady-state frictional behaviour of a-C : H films (with respect toLangmuir’s fractional surface adsorption and removal rates)compared favourably with the friction data generated earlierby Heimberg et al [63].

Among the gaseous species that may exist in testchambers, oxygen and water molecules were shown to havethe strongest effects on the friction and wear behaviour ofDLC films. For example, in vacuum and inert gases, friction

R318

Topical Review

Figure 9. Friction mechanisms of various DLC films at molecular level.

Figure 10. Effect of humid air and humid nitrogen on the frictionbehaviour of a a-C : H film. Note that humid air is more detrimentalto friction than humid nitrogen.

coefficients of less than 0.01 are feasible with certain DLCfilms (especially highly hydrogenated DLC films) [3, 40, 66,67]; however, when oxygen and/or moisture are introducedinto the test chambers, their friction coefficients may increasesubstantially, as shown in figure 10. In highly moist air,condensed water molecules can give rise to capillary forcesthat increase friction.

If sliding takes place in a liquid (such as water, oil orfuel), the frictional behaviour of DLC films may be primarilydominated by the physical (such as viscosity) and chemicalnature (such as polarity, chemical affinity and molecularsize) of such liquids. In certain types of DLC films,tribochemical interactions between sliding surfaces and theadditives in oils have led to the formation of low-frictiontribofilms (such as MoSx and WSx) [68, 69]. Recently,researchers have demonstrated superlow friction and wear onhydrogen-free DLC films with the addition of glycerol mono-oleate (GMO) to poly-alpha olefin base oils [70]. Detailed

surface analytical studies have revealed that such impressivetribological behaviour was primarily due to the formation ofan OH-terminated carbon surface causing very little σ -bondinteractions.

The tribochemistry of sliding interfaces that involve aDLC coating is complex and has been the focus of manyfundamental studies in recent years. Despite being highly inert,DLC films interact or react with the chemical species in theirsurroundings as well as with materials that are being rubbedagainst them. Recently, a new concept based on tribocharging,tribomicroplasma generation and triboemission phenomenahas been proposed by Nakayama [71]. For the tribosystemsthat consist of a diamond and/or a DLC film, the intensityof tribomicroplasmas generated at sliding interfaces increaseslinearly with the electrical resistivity of these films. The highlyinsulating DLC films produce the greatest triboplasmas andhence trigger strong tribochemical interactions.

The incorporation of certain alloying elements (such asSi, F, B, P, N and various metals) into DLC films may inducea significant effect on the tribochemical mechanisms of thesefilms [72]. The reactivity of such atomic species with oxygenor water vapour during sliding (such as Si, which forms siliconhydroxides during the friction process) may induce higherinertness of the carbonaceous network. As a consequence,the coefficient of friction can be made less sensitive to therelative humidity of the surrounding environment through theincorporation of silicon, as discussed in [73–75]. DLC filmscontaining small amounts of Si, F, B and S exhibit frictioncoefficients almost independent of the relative humidity, withina range 5–85% [1, 76].

4.1.4. Thermal interactions. Upon exposure to elevatedtemperatures, DLC films may undergo gradual transformationfrom a highly disordered or amorphous state to an increasinglyordered or graphitic state [77,78]. This occurs mainly because

R319

Topical Review

0

0.1

0.2

Fric

tion

Coe

ffici

ent

0.3

0.4

0.5

0 1000 2000 3000 4000 5000

400 C

200 C

23 C

Time (s)

Figure 11. Friction behaviour of ta-C at different temperatures.

these films are thermodynamically unstable, and when externalthermal energy is provided, the carbon and/or other atoms maybegin to re-arrange themselves and assume thermodynamicallymore stable bonding configurations, such as graphite. Ifthe DLC film is hydrogenated, some of the hydrogen atoms(especially the ones that are not bonded to carbons) may beginto diffuse out and leave a relatively porous structure behind.With such changes in chemical and structural morphology, thefriction and wear behaviour of DLC films also change. Inthe case of hydrogenated DLC films, the end product maybe a highly porous and graphitic thin layer that is very softand hence wears out quickly. Hydrogen-free or ta-C typefilms may have much higher endurance limits to elevatedtemperatures. The structural integrity or density of these filmsdoes not change much until reaching very high temperatures(i.e. 500 ◦C), at which point they may also gradually transformto graphite. The hydrogenated DLC films (depending on thedegree of hydrogenation or hydrogen content) may provide lowfriction up to about 300 ◦C during short-duration (i.e. ≈1 h)sliding tests, but during repeat or extended tests, they tend towear at a much faster pace than at room temperature [77].Hydrogen-free or ta-C type films may last longer; however,their friction coefficients tend to increase with increasing testtemperatures (presumably due to the thermal desorption ofwater and other adsorbed species from their sliding surfaces).Figure 11 compares the frictional behaviour of the same ta-Cfilm at room temperatures, 200 and 400 ◦C.

4.1.5. Third-body interactions. The frictional behaviour ofDLC films can be influenced by the presence or absence ofthird bodies or transfer layers on their sliding surfaces [64, 79–84]. These layers mostly form on the surfaces of uncoatedcounterface balls and pins sliding against the DLC-coated flatsor discs. The type or composition of counterface material mayalso play some role in the kinetics and/or thermodynamics oftransfer film formation. Normally, those counterface materialsthat are known to be very strong carbide formers (Ti, Fe, W, Si,etc) tend to generate such layers much faster and with a muchhigher degree of coverage and strong bonding. Conversely,non-carbide formers (such as Cu) may not form a strongly-bonded transfer layer on their sliding surfaces, and the frictioncoefficients of the DLC films sliding against such non-carbide

Figure 12. Typical transfer layer that was formed on a steel ballduring a sliding test in dry nitrogen (magnification: 50X).

formers are generally high. Figure 12 shows a transfer layerthat was formed on a steel ball during a sliding test in drynitrogen.

The third bodies or debris particles are generally producedby wearing of sliding contact surfaces. These debris particlesare trapped at the sliding contact interfaces and undergo severephysical grinding action as well as chemical reaction with theuncoated ball surface and other species in their surroundings.These debris particles are often smeared on one or both sidesof the sliding pairs as a thin film. In the case of most DLCfilms, the debris particles and transfer layers have a disorderedmicrostructure, and their Raman spectroscopy and electrondiffraction patterns resemble those of a disordered graphite[82–86]. Hauert [87] has shown that the chemistry involvedin the formation of a transfer layer during a tribologicalexperiment depends critically on experimental conditions suchas the contact force and the sliding velocity: at the highestload and sliding velocity, a continuous transfer film is formedon sliding surfaces of steel balls, with a composition similarto the original DLC film. On the contrary, at the lowestloads and sliding velocities, fiber-like debris particles fromdiscontinuous wear are generated, and these particles have aRaman signature that is very close to that of graphite. Thesetwo opposite experimental configurations have been tested atthe same relative humidity (60%) but under different contactpressures. Since the tests were run in humid air, it is possiblethat the debris particles may have contained considerableamounts of oxygen, but this was difficult to verify with Raman.

Lower humidity tends to increase the rate of graphitizationof third-body particles mainly because of the reduced coolingeffect of water molecules on flash heating. The rate ofgraphitization can be reduced by lowering ambient temperatureor by further increasing the humidity level in test chambers.Again, this observation can largely be attributed to thesuppression of the temperature rise at contact spots where flashheating can take place.

When tests are run in vacuum, the partial pressure ofwater vapour (pH2O) appears to dominate the mechanismsby which frictional interactions occur. For example, duringrecent experimental studies with a hydrogenated DLC filmin ultrahigh vacuum, a transition from the ultralow frictionregime (10−2 range) to moderate friction regime (10−1 range)

R320

Topical Review

occurred when the water vapour pressure increased from0.1 hPa (RH = 0.4% at 23 ◦C) to 1 hPa (RH = 4% at23 ◦C) [67, 88]. Similar observations were made on a highlyhydrogenated DLC film during tests under a wide range ofwater vapour pressures by other vapour pressures by otherresearchers [89, 90]. In all these tribological tests, DLC filmswere applied on the ball sides, and thus the sliding tookplace between two DLC surfaces. Under conditions whereDLC was present on only one of the sliding surfaces, suchas the disc surface, the formation of a carbon-rich transferfilm on the ball side was paramount for achieving ultralowfriction [66]. However, when tests were performed underincreasing partial pressures of water vapour, the friction tendedto increase significantly. At higher humidity levels, thetransfer layers formed on ball surfaces were much thinner (asdetected by Auger electron spectroscopy). Consequently, thecounterface and the tribotesting parameters, together with thetest environment, play a crucial role in the kinetics of theformation and composition of the transfer films and, thus,strongly influence the friction and wear behaviour of DLCfilms. Recent systematic studies by Scharf and Singer furtherconfirmed that the formation and periodic loss of transfer filmsfrom sliding sapphire ball surfaces are the main causes offluctuations in the friction behaviour of DLC coatings [79].

5. Recent advances and future prospects

5.1. Novel compositions

Because of their increasing popularity and diverse potentialapplications, DLC films have gone through numerousmodifications in their microstructure and chemistry duringthe past decade. Using advanced deposition techniques,researchers have developed more exotic DLC films that aretruly nano-structured, -composite or –alloyed to provide muchimproved physical, mechanical and tribological properties. Inparticular, doping DLC with certain metals, metalloids andgaseous species (such as Ti, B, S, Si, Cr, F, W and N) hasbecome popular for various applications. Systematic studieshave confirmed that compared with their predecessors, thesenano-alloyed or doped-DLC films are capable of providingsuperior mechanical, tribological, thermal and electricalproperties [30,52,53,72]. Adding Si, Ti and W into DLC filmswas proved to provide better friction and wear properties underlubricated sliding conditions and high resistance to scuffingunder severe contact pressures [68,91,92]. For example, withthe addition of W, up to 50% reduction in friction is reported byPodgornik et al [68]. Mechanistically, the superior tribologicalbehaviour of W-containing DLC films under lubricated slidingconditions may have been due to the formation of WSx typetribofilms on their sliding surfaces. Figure 13 compares thefrictional properties of a standard and a B-doped DLC film inopen air. As is clear, the presence of dopants in DLC films candefinitely improve their friction and wear performance.

Among all other alloying elements, nitrogen occupies aspecial place in the field of DLC films. Nitrogenated DLC(also referred to as ‘carbon nitride films’) provides significantlyhigher hardness and superior tribological performance whenused in magnetic hard disc applications [30, 93–96]. Asuperlow friction version of carbon nitride has recently been

0

0.05

0.1

0.15

Fri

ctio

n C

oe

ffic

ien

t 0.2

0.25

0.3

0 500 1000 1500 2000 2500 3000 3500

Standard DLCBoron-doped DLC

Time (s)

Figure 13. Friction behaviour of a standard and a boron-doped DLCfilm in humid air.

pioneered by Kato and his co-workers [5, 97]. When tested indry nitrogen or nitrogen gas flow against Si3N4 balls, such filmsreached friction coefficients less than 0.01. These films had asignificant positive impact on the tribological performance ofSiC ceramics when tested in water [98].

Interest in doping carbon films with nitrogen started backin the 1980s, but this idea gained increased momentum duringthe 1990s when Cohen and his co-workers predicted that thesynthesis of a crystalline form of carbon nitrides (in particular,β-C3N4) is feasible and, if realized, such a material couldprovide hardness values much greater than that of naturaldiamond [99]. Despite intense research by many scientistsin past years, the production of such crystalline C3N4 phaseshas not yet been achieved or independently confirmed, butthe amorphous carbon nitride films have since flourished andbeen used in industrial applications. Apparently, these filmspossess the impressive mechanical, tribological and corrosionproperties that are desirable in magnetic hard disc applications;hence, most of the hard disc overcoats are currently made ofcarbon nitrides [5, 30, 91, 93, 94, 96].

5.2. Novel architectures

Apart from the monolithic or single-phase DLC films men-tioned above, production of multilayered and nanostructuredDLC films has also become popular in recent years. Such filmstypically consist of an amorphous carbon matrix impregnatedwith nano-scale crystalline phases. In the case of multilayerfilms, alternating layers of amorphous carbon and crystallineor amorphous layers of other metals and/or compounds are pro-duced at thicknesses ranging from a few nanometres to tens ofnanometres. Such films can easily be deposited on appropriatesubstrates by preferably using hybrid deposition systems con-sisting of not only sputtering but also cathodic arc-PVD, ion-beam deposition or femtosecond pulsed laser ablation. Thesenanocomposite and multilayered DLC films are tough and re-sistant to microcrack initiation and growth during mechanicalor tribological uses [100].

R321

Topical Review

Some of the very popular multi-layered films engineeredin recent years consist of alternating layers of carbon and W orcarbon and Cr [100, 101]. In particular, DLC films consistingof C/W multilayers are currently being used by industry toovercome friction, wear and scuffing problems in criticalengine components [102, 103]. Mechanically, most of themultilayered films are relatively soft but can provide excellenttoughness and high resistance to scuffing and adhesive wearin rolling and sliding contacts, especially under lubricatedsliding conditions. In particular, the DLC films based onC/W (either nanocomposite or multilayered) work extremelywell under lubricated sliding conditions, and mechanistically,such superior performance is attributed to the formation ofWSx-rich boundary films on sliding contact surfaces [68].They are currently used as protective films on gears, pistonrings, bearings, fuel injectors and several other demandingengine applications [102–106]. In recent years, such filmshave been the subject of more intense scientific studies andattracted much broader interest from several industrial endusers [107–110]. The nanocomposite DLC films are primarilymade of nanoscale carbide phases that are evenly dispersedor embedded in the amorphous DLC matrix [111, 112]. Thesize and concentration of these nanophases can be tailored toachieve a wide range of mechanical, electrical and tribologicalproperties [101, 111, 113].

As is clear from the foregoing, the DLC films haveevolved rapidly over the past two decades. The structure andcomposition of these films (H content, fraction of H bonded toC, C hybridizations, nature of bonds, alloying elements, etc)are strongly influenced by the average impact energy of theatomic and/or ionic species that impinge on the growing filmsurface and by the deposition temperature. The latest DLCsare both structurally and chemically unique and can meet theincreasing demands of advanced mechanical systems. With theintroduction of hybrid deposition processes in recent years, thelatest films with a nanocomposite or multilayered architecturenow offer much improved properties and better performanceunder dry and lubricated sliding conditions.

5.3. Patterned DLC films

In an attempt to further improve the performance and durabilityof DLC films under lubricated sliding conditions, researchershave lately been creating special textures on their slidingsurfaces. In particular, high-precision patterns created onDLC surfaces by excimer lasers have been shown to improvethe tribological properties of these films under boundary-lubricated sliding conditions [114]. Furthermore, with the useof femtosecond lasers, even greater improvements in frictionand wear have been achieved. For example, with the useof ultrashort laser pulses (in the 100 fs range, with powerdensity as high as 1013 W cm−2), researchers were able toablate various materials without much collateral damage (i.e.no splashing of molten metal and negligible heat-affectedzone). This behaviour is more typical of surfaces preparedby conventional lasers [115]. Lifetime increases up to a factorof 10 have been reported for the femtosecond-laser-patternedTiN [116] and TiCN [117] films. Improved durability andfrictional performance of the patterned surfaces have mainlybeen attributed to the fact that shallow dimples created on

sliding surfaces act as reservoirs for lubricants and improve thehydrodynamic efficiency of such surfaces. These dimples canalso trap abrasive wear particles that may have been generatedduring sliding contacts and thus reduce the risk of third-bodywear.

Recently, laser patterning of wear-resistant DLC films hasbeen done by two methods: applying a coating over already-patterned substrates (indirect processing) or by direct laserpatterning of an as-deposited DLC film [118]. Dimple depths(>10 µm) that yield positive tribological improvements aremuch greater than the DLC film thickness (<5 µm). Dumitruet al report that debris particles were conveniently trappedwithin the dimples created by the indirect laser processing, thuspreventing third-body wear and hence the breakdown of thewhole tribological system. According to Voevodin et al [119],the three-dimensional design considerations have considerablyimproved the tribological characteristics of hard coatings bypermitting solid lubricant replenishment inside the slidingcontacts. Such patterning may have additional positive effectsby carrying away the heat from sliding interfaces. In [119], afunctionally-gradient Ti–TiC–TiC/DLC coating with an upperlayer of tough nanocrystalline/amorphous composite was usedfor load support, crack prevention and stress equalization. Thiscoating was processed by laser irradiation to form groovedtracks along wear paths, which were then filled with MoS2

to provide a solid lubricant reservoir in the lateral dimensionof the coating. The three-dimensional coating was testedin long-duration sliding tests at fixed and variable humidity.The coating exhibited environmental adaptation, with frictioncoefficients of 0.15 in humid air and 0.02 in dry nitrogen. Thewear life was increased by at least one order of magnitudewhen compared with that for a hard coating with a top MoS2

layer but without three-dimensional laser patterns.

5.4. Novel carbon films and structures

In another major development, DLC films have beensuccessfully deposited on flexible polymers and rubber-likeflexible substrates [120, 121]. These new coatings possesseda compositionally graded microstructure and thus had agradient of mechanical properties. They are most convenientlysynthesized by plasma-enhanced CVD processes in whichbias voltage and gas discharge or plasma composition arecontrolled. Havert et al and Zhang et al have shown thatthe wear rate of such DLC coatings can be minimized furtherby controlling the sequence of the alternating layers and byvarying the thickness of each layer [122, 123]. In addition tothese coatings that hold promise for applications on flexibleor polymeric substrates, new forms of carbon films withstructure and properties close to those of DLC are nowbeing deposited, and they have shown excellent mechanicaland tribological properties. These films are produced bymagnetron sputtering by establishing a unique combination ofbias voltage and substrate temperature during film deposition.The films produced under such condition have a fullerene-likemicrostructure, as confirmed by high-resolution transmissionelectron microscopy [124]. Because of their unique structures,these films are very hard and resilient and have exceptionalability to recover elastically. Studies by Neidhardt et alconfirmed a solid lubrication capability for such films under a

R322

Topical Review

wide range of sliding conditions and increased the prospects forsome applications for these fullerene-like carbon films [125].

In another interesting development researchers havesynthesized a carbon film that is derived from carbide-basedmaterials and coatings. Specifically, the metallic part inthese carbides is selectively removed by high-temperaturechlorination, and the carbon atoms left behind are rearranged toform the carbon films. Being directly derived from the carbidesubstrates, they are called carbide-derived carbon (CDC) films.These films are truly nanostructured and consist of bothamorphous and crystalline forms of carbon (i.e. nanocrystallinegraphite, diamond, fullerene-like structures, such as carbonnano-tubes, -onions and –horns). Recent tribological studieshave confirmed that CDC films have impressive friction andwear properties, especially after a post-process hydrogentreatment [126]. These films are not produced in a plasmadeposition system and do not involve energetic bombardment;however, they possess similar characteristics to those of DLCin terms of their tribological behaviour. The friction and wearmechanisms observed in these films are very similar to thoseof the DLC films. Other examples of novel carbonaceousmaterials with interesting tribological properties include nano-onions and fluorinated single- and multi-wall nanotubes.Recent exploratory research has shown that all these carbonforms are able to lower friction and extend the wear life ofsliding interfaces even under extreme sliding conditions [127].

6. Superlubricity in DLC films

As discussed in section 4, the sliding friction and wearcoefficients of DLC films are generally low, but depending onthe kinds of extrinsic and intrinsic factors, these coefficientsmay vary widely. To achieve superlow friction (µ < 0.01) inDLC films, one has to reduce or eliminate the main causes offriction. One of the requirements is to make sure that the slidingsurfaces of DLC films are atomically smooth (like cleavedmica surfaces). Such a smoothness, combined with molecularflexibility, is necessary for eliminating mechanical interlockingand/or asperity-asperity interactions during sliding. Secondlyand most important, these surfaces must have the highestdegree of chemical passivity or inertness so that they donot enter into any type of adhesive bonding or chemicalinteractions with counterface materials.

Because of their amorphous structures, DLC films canbe made extremely smooth. When deposited on atomicallysmooth or highly polished substrates (such as Si wafers orcleaved sapphire or mica surfaces), DLC films can attainthe kinds of smoothness that are needed for eliminatingthe deleterious effects of surface roughness on friction. Atthicknesses as low as 2–5 nm, they provide very uniformcoverage and mimic the original surface roughness of theunderlying substrates. If necessary, the sliding surfaces ofDLC films can be polished after the deposition.

As for achieving a high degree of chemical inertnesson sliding DLC surfaces, researchers have pursued twocomplementary or closely-related approaches. Specifically,they have either used a hydrogen-rich gas discharge plasmaduring film deposition [3,40,62,66,128–131] or they haveintroduced hydrogen gas into the test chamber duringtribological testing [132–134]. The main purpose of both

0

0.005

0.01

0.015

0.02

0

0.002

0.004

0.006

0.008

0.01Friction Coefficient Wear Rates (X10-7 mm3/N.m)

Source Gas Chemistry

0.3 7

H/C=1

H/C=10

Figure 14. Friction and wear coefficients of DLC films derivedfrom various source gases in a PECVD system (test conditions:load, 10 N; speed, 0.5 m s−1; temperature, 22–23 ◦C; testenvironment, dry N2).

approaches was to eliminate those dangling σ -bonds byreaction with hydrogen. Indeed dangling bonds can causevery strong covalent bond interactions and, hence, highfriction during sliding. Because of its small size in theatomic and protonic forms, large amounts of hydrogen (upto 50 at. %) can be incorporated into DLC films. For theproduction of hydrogen-rich DLC films, one can use a PVD orCVD deposition system such as plasma-enhanced chemicalvapour deposition (PECVD), ion-beam-assisted deposition,magnetron sputtering, cathodic arc-PVD and PLD. In the caseof PECVD, H2 gas is blended with such hydrocarbons asCH4 and C2H2 up to 90 vol % and released into the depositionchamber [3, 128–131]. In the cases of arc-PVD, PLD or directsputtering, one has to blend hydrogen with argon or other gasesused in the generation of discharge plasmas. In the case ofPECVD, a gas composition of 75% H2 and 25% CH4 (whichcorresponds to a hydrogen-to-carbon ratio of 10 in the plasma)is more than enough to produce a highly hydrogenated DLCfilm (containing more than 40 at. % hydrogen).

Such a high degree of hydrogenation of DLC films appearsto play a pivotal role in friction and wear especially when testsare run in inert or vacuum test environments. Figure 14 showsthe relationship between friction and wear coefficients of DLCfilms derived from various source gases having H/C ratios of1–10. In general, the higher the H/C ratio in the gas dischargeplasma, the lower the friction coefficients. When tested ina clean and dry test environment, a hydrogen-free DLC filmtypically provides friction coefficients of 0.6–0.7, while aDLC film derived from pure methane gas exhibits frictioncoefficients of 0.015–0.02 [3, 128–131]. The H/C ratio is 4for pure methane source gas, while it is zero for the hydrogen-free DLC produced in a cathodic-arc PVD system. As shownin figure 14, films derived from C2H2 (whose H/C ratio is 1)provide friction coefficients of 0.3–0.4 under the same slidingconditions, whereas the friction coefficient for ethylene with anH/C = 2 is between that of methane and acetylene. The lowestfriction is achieved in the plasma that contains 10 hydrogensfor each carbon atom. In short, those films grown in hydrogen-poor plasmas exhibit high and unsteady friction, while the filmsgrown in hydrogen-rich plasmas provide very steady and lowfriction.

R323

Topical Review

c

Figure 15. Illustration of sliding contact interface of hydrogen terminated a-C : H surfaces.

Figure 14 also compares the wear performance of variousDLC films grown in gas discharge plasmas with various H/Cratios. Similarly to the friction results, the wear of DLC filmsappears to relate strongly to the H/C ratio of gas dischargeplasmas. In general, the higher the H/C ratio, the lower thewear coefficient (suggesting that hydrogen plays an importantrole not only in friction but also in the wear behaviour of DLCfilms).

Based on the friction test results from hydrogen-richand/or -poor DLC films, the following mechanism is proposedto explain the role of hydrogen in the frictional behaviourof DLC films. Because hydrogen has a strong chemicalaffinity towards carbon, it bonds strongly to some of thecarbon atoms and thus effectively passivates their unoccupiedor free σ -bonds. Once passivated, such carbon atoms becomechemically inert and cause very little adhesive interactionsduring sliding. Such reduction in adhesive interactions acrossthe sliding interface results in low friction. This explanation forthe low-friction behaviour of bulk diamond and thin diamondfilms is widely accepted [135, 136]. C–H bonding is covalentand extremely strong (stronger than single C–C bonds) andhence very difficult to remove from the surface unless slidingtests are performed in high vacuum or at high temperatures[137]. Under such circumstances, it is very-well known thatsliding diamond surfaces begin to exhibit high friction.

When extra hydrogen is used during DLC deposition,several important events take place and determine the structuralchemistry of the DLC films being produced on a substratesurface. First, this condition can lead to increased hydrogenconcentration within the bulk, as well as on the surface. Mostof these hydrogen atoms are paired with σ -bonds, but someunbonded free hydrogen may also exist at interstitials. Infact, computer simulation of highly hydrogenated DLC filmssuggests that, in addition to bonded hydrogen, considerableamounts of unbonded or free hydrogens may exist inatomic and molecular forms (see figure 4). High hydrogenconcentration within the DLC films and on the surfaceshould effectively diminish or even eliminate the possibility ofunoccupied σ -bonds remaining and participating in any strongadhesive interactions during sliding. Free hydrogen within thefilms may serve as a reservoir and can replenish or replace

those hydrogen atoms that may have been lost due to thermalheating and/or mechanical action during sliding.

Secondly, hydrogen is highly effective in etching outor removing sp2-bonded or graphitic carbon forms duringdeposition. The removal of such graphitic carbons preventsthe formation of planar graphitic clusters that can give rise toπ–π* interactions. When DLC films are prepared in highlyhydrogenated gas discharge plasmas, strong C-H bondingrather than C=C double bonding should be favoured. Asexplained in section 4, residual π -bonding that can result fromC=C double bonds in DLC can give rise to friction. Finally,some of the carbon atoms (at least those on the surface) couldbe double-hydrogenated: that is, two hydrogen atoms bondedto each carbon atom on the surface. This bonding can occur onthe unreconstructed (100) surfaces of diamond structures underspecial or supercritical conditions that may have been createdby energetic hydrogen bombardment in a highly hydrogenatedgas discharge plasma. The double-hydrogenated carbon atomswill increase the hydrogen density of these surfaces and thusprovide better shielding or passivation and hence superlowfriction. Such a friction model is presented in figure 15 forpartially dihydrated sliding DLC surfaces [62].

As mentioned earlier, other forces such as vander Waals and capillary forces, as well as electrostaticattraction/repulsion, may cause adhesion and hence frictionat the sliding interfaces of DLC films. When friction tests arerun in a clean, dry nitrogen environment, the extent of capillaryforces due to moisture precipitation on the sliding surfacesshould be minimal or essentially absent. As for the van derWaals forces, they will be present at the sliding interfaces, buttheir relative contributions to overall frictional force should beinsignificant mainly because of the very high contact loadsused in sliding tests. As for electrostatic attraction, sincethe DLC films are in general dielectric, their sliding surfacescan certainly accumulate static electrical charges. Then, themain question is whether these charges will cause attraction orrepulsion. When the free electrons of hydrogen atoms pair withthe dangling σ -bonds of carbon atoms, the electrical chargedensity is permanently shifted to the other side of the nucleusof the hydrogen atom and away from the surface. Such ashift in charge density allows the positively charged hydrogenproton in its nucleus to be closer to the surface than the

R324

Topical Review

electron, which is used up by the σ -bond of the surface carbonatoms. Therefore, the creation of such a dipole configurationat the sliding interface should give rise to repulsion rather thanattraction between the hydrogen-terminated sliding surfacesof the DLC films [62]. In support of the proposed mechanismdescribed above, Dag and Ciraci have recently demonstratedthe existence of strong repulsive forces between H terminateddiamond (001) surfaces, especially when the distance betweensuch surfaces is below 2.5 Å [138]. In short, there is no doubtthat hydrogen plays a critical role in the frictional behaviourof most carbon films in general and DLC films in particular.

Superlow friction in DLC films can also be achievedby OH termination of surface carbon atoms. In a series ofrecent studies, Kano et al achieved ultralow friction (0.03)on hydrogen-free DLC or ta-C films under lubricated slidingconditions [70]. Specifically, they blended a poly-alpha olefinbase oil with glycerol mono-oleate and used it as a lubricantfor sliding ta-C surfaces. Surface analytical studies confirmedthe presence of a layer of OH on the surface. Based on thesefindings, they proposed that alcohol function groups of GMOand the surface carbon atoms of ta-C were mechanically andtribochemically activated to result in a strongly bonded OHlayer on ta-C surfaces. Just like hydrogen termination, OHtermination of the dangling σ -bonds appears to have resultedin such ultralow friction.

Kato and his coworkers also achieved ultralow friction(less than 0.01) by blowing nitrogen gas into the slidinginterfaces of Si3N4 balls and ion-beam-deposited CNx coatings[97]. Based on numerous other studies in dry nitrogen, vacuumand elevated temperatures, they concluded that nitrogen–surface interactions (bolstered by mechanical action andtribochemical reaction) may have been the reason for theultralow friction behaviour of these CNx films. Overall, theseand other dedicated studies suggest that the control of surfacechemistry or chemical interactions at sliding DLC interfacesis extremely important for the friction and wear behaviourof these films. Specifically, by controlling or effectivelyeliminating the intrinsic and extrinsic sources of friction inDLC films, one should be able to achieve ultra- and super-low friction coefficients under both dry and lubricated slidingconditions.

7. Summary and future direction

Unlike most other coatings, DLC is unique and may beformulated in many ways. Some are hydrogen-free, othersare highly hydrogenated, while some others are doped witha range of alloying elements. A range of nano-compositeand –structured DLC films is also available and represents thelatest trend in the field. The degree of sp2 and sp3 bonding inthese films can be controlled to formulate DLC films with morediamond- or graphite-like qualities. There now exist severaldeposition processes that can produce high-quality DLC filmsat reasonable costs. From a tribological point of view, theintrinsic and extrinsic factors can play significant roles inthe friction and wear behaviour of DLC films. Intrinsically,the extent of sp2 versus sp3 bonding and the amount ofother elemental species (H, N, F, etc) can make a significantdifference in their friction and wear. Extrinsically, the testconditions and environments, as well the type of counterface

materials being used, may play major roles. Previous researchhas shown that the sliding friction and wear properties of DLCfilms can be optimized by controlling the chemistry of gasdischarge plasmas during film deposition. Ultra- and super-low friction is feasible with highly hydrogenated DLC filmswhen tests are performed in inert gases or vacuum. Whensliding surfaces of hydrogen-free or CNx type DLC films are re-conditioned by tribochemical reactions, they can also provideextremely low friction and wear coefficients. Mechanistically,the ultra- and/or super-low friction behaviour of these carbonfilms is related to the formation of a highly passive surface thathas little or no chemical or physical interactions with the slidingcounterfaces. Because of their impressive friction and wearproperties and low deposition costs, DLC films are presentlyused to combat friction and wear in a wide range of engineeringapplications.

Acknowledgments

This work is supported by the US Department of Energy,Office of Energy Efficiency and Renewable Energy, FreedomCar and Vehicle Technologies Program, under ContractW-31-109-Eng-38. The authors thank their students andcollaborators who participated in the preparation, testing andcharacterization of the DLC coatings discussed in this chapter.

© US Government

References

[1] Erdemir A and Donnet C 2000 Modern Tribology Handbooked B Bhushan (Boca Raton, FL: CRC Press) pp 871–908

[2] Erdemir A 2002 J. Eng. Tribol. 216 387[3] Erdemir A, Eryilmaz O L and Fenske G 2000 J. Vac. Sci.

Technol. A 18 1987[4] Grill A 1997 Surf. Coat. Technol. 94 507[5] Kato K, Umehara N and Adachi K 2003 Wear 254 1062[6] Fusaro R L and Sliney H E 1970 ASLE Trans. 13 56[7] Rabinowicz E and Imai M 1964 Wear 7 298[8] Savage R H 1948 J. Appl. Phys. 19 1[9] Burton R A and Burton R G 1989 Proc. 35th Meeting IEEE

Holm Conf. on Electrical Contacts (Chicago, IL) (NewYork: IEEE) pp 31–4

[10] Erdemir A and Donnet C 2006 Wear: Materials, Mechanismsand Practice ed G W Stachowiak (New York: Wiley)pp 191–209

[11] Erdemir A, Eryilmaz O L, Urgen Mm Kazmanli K, Mehta Nand Prorok B Nanomaterials Handbook ed Y Gogotsi(Boca Raton, FL: CRC Press) p 685

[12] Ruff A W 2000 Modern Tribology Handbook ed B Bhushan(Boca Raton, FL: CRC Press) pp 523–61

[13] Grill A 1993 Wear 168 143[14] Tsai H C and Bogy D B 1987 J. Vac. Sci. Technol. 5 3287[15] Eisenberg S and Chabot R 1971 J. Appl. Phys. 42 2953[16] Schmellenmeier H 1954 Exp. Technik Physik 1 49[17] Holland L and Ojha S M 1976 Thin Solid Films 38 L17[18] Spencer E G, Schmidt P H, Joy D C and Sansalone F J 1976

Appl. Phys. Lett. 29 118[19] Weissmantel C, Bewilogua K, Schurer C, Breuer K and

Zscheile H 1979 Thin Solid Films 61 L1[20] Deryagin B V and Fedoseev D B 1975 Sci. Am. 233 102[21] Angus J 1994 Synthetic Diamond: Emerging CVD Science

and Technology ed K E Spear and J P Dismukes (NewYork: Wiley) p 21

[22] Spitsyn B V, Bouilov L L and Derjaguin B V 1988 Prog.Cryst. Growth Charact. 17 79

R325

Topical Review

[23] Messier R, Badzian A R, Badzian T, Spear K E, Bachmann Pand Roy R 1987 Thin Solid Films 153 1

[24] Eisenberg S 1984 J. Vac. Sci. Technol. 2 369[25] Arnoldussen T C and Rossi E M 1985 Ann. Rev. Mater. Sci.

15 379[26] King F K 1981 IEEE Trans. Magn. 17 1376[27] Khan M R et al 1988 IEEE Trans. Magn. 24 2647[28] Talke F E 1995 Wear 190 232[29] Johnson K E, Mate C M, Merz J A, White R L and Wu A W

1996 IBM J. Res. Dev. 40 511[30] Cutiongco E C, Li D Chung Y W and Bhatia C S 1996 J.

Tribol. Trans. 118 543[31] Enke K, Dimigen H, and Hubsch H 1980 Appl. Phys. Lett. 36

291[32] Bowden F P and Young J E 1951 Proc. R. Soc. Lond. A 208

444[33] Rowe G W 1960 Wear 3 274[34] Imamura A, Tsukamoto T, Shibuki K and Takatsu S 1988

Surf. Coat. Technol. 36 161[35] Andersson L P 1981 Thin Solid Films 86 193[36] Grill A, Patel V and Meyerson B 1991 Surf. Coat. Technol.

49 530[37] Miyoshi K, Wu R L C and Garscadden A 1992 Diamond.

Relat. Mater. 1 639[38] Ronkainen H, Koskinen J, Anttila A, Holmberg K and

Hirvonen J-P, 1992 Surf. Coat. Technol. 55 428[39] Erdemir A, Switala M, Wei R and Wilbur P 1991 Surf. Coat.

Technol. 50 17[40] Donnet C, Belin M, Auge J C, Martin J M, Grill A and Patel

V 1994 Surf. Coat. Technol. 68–69 626[41] Robertson J 1986 Adv. Phys. 35 317[42] Robertson J 2002 Mater. Sci. Eng. 37 129[43] Lettington A H 1998 Carbon 36 555[44] Hauert R 2003 Diamond. Relat. Mater. 12 583[45] Moon M W, Jensen H M, Hutchinson, J W, Oh K H and

Evans A G 2002 J. Mech. Phys. Solids 50 2355[46] Bubenzer A, Dischler B, Brandt G and Koidl P 1983 J. Vac.

Sci. Technol. A 1 305[47] Voevodin A A, Donley M S and Zabinski J S 1997 Surf. Coat.

Technol. 92 42[48] Ronkainen H, Koskinen J, Anttila A, Holmberg K and

Hirvinen J-P 1992 Diamond. Relat. Mater. 1 639[49] Horsfall R H 1998 Proc. 41st Annual Technical Conf. of

Society of Vacuum Coaters (Boston, MA) p 60[50] Sullivan J P, Friedmann T A and Hjort K 2001 MRS Bull. 26

309[51] Ferrari A C and Robertson J 2000 Phys. Rev. B 6 114095[52] Qian F, Craciun V, Singh R K, Dutta S D and Pronko P P

1999 J. Appl. Phys. 86 2281[53] Garrelie F, Loir A S, Donnet C, Rogemond F, Le Harzic R,

Belin M, Audouard E and Laporte P 2003 Surf. Coat.Technol. 163–164 306

[54] Loir A S, Garrelie F, Donnet C, Subtil J L, Belin M,Forest B, Rogemond F and Laporte P 2005 Appl. Surf. Sci.247 225

[55] Benchikh N, Garrelie F, Donent C, Bouchet-Fabre B, WolskiK, Rogemond F, Loir A S and Subtil J L 2005 Thin SolidFilms 482 287

[56] Perriere J, Millon E, Seiler W, Boulmer-Leborgne C, CraciunV, Albert O, Loulergue J C and Etchepare J 2002 J. Appl.Phys. 91 690

[57] Benchikh N, Garrelie F, Donnet C, Wolski K, Fillit R Y,Rogemond F, Subtil J L, Rouzaud J N and Laval J Y 2006Surf. Coat. Technol. 200 6272

[58] Gupta B K, Malshe A, Bhushan B and Subramaniam V V1994 J. Tribol. 116 445

[59] Chandrasekar S and Bhushan B 1992 Wear 153 79[60] Miyoshi K, Wu R L C, Garscadden A, Barnes P N and

Jackson H E 1993 J. Appl. Phys. 74 4446[61] Gardos M N and Soriano B L 1990 J. Mater. Res. 5

2599[62] Erdemir A 2001 Surf. Coat. Technol. 146 292

[63] Heimberg J A, Wahl K J, Singer I L and Erdemir A 2001Appl. Phys. Lett. 78 2449

[64] Erdemir A, Bindal C, Fenske G R and Wilbur P 1996 Tribol.Trans. 39 735–44

[65] Dickrell P L, Sawyer W G, Heimberg J A, Singer I L, Wahl KJ and Erdemir A 2005 J. Tribol. 127 82

[66] Donnet C, Le Mogne T, Ponsonnet L, Belin M, Grill A,Patel V and Jahnes C 1998 Tribol. Lett. 4 259

[67] Andersson J, Erck R A and Erdemir A 2003 Surf. Coat.Technol. 163 535

[68] Podgornik B, Hren D and Vizintin J 2005 Thin Solid Films476 92–100

[69] de Barros-Bouchet M I, Martin J M, Le-Mogne T andVacher B 2005 Tribol. Int. 38 257

[70] Kano M et al 2005 Tribol. Lett. 18 245[71] Nakayama K 2004 Surf. Coat. Technol. 188 599[72] Donnet C 1998 Surf. Coat. Technol. 100–101 180[73] Oguri K and Arai T 1990 J. Mater. Res. 5 2567[74] Gangopadhyay A K, Willermet P A, Tamor M A and

Vassel W C 1997 Tribol. Int. 30 9[75] Wu W J, Pai T-M and Hon M H 1998 Diamond. Relat. Mater.

7 1478[76] Gilmore R and Hauert R 2001 Thin Solid Films 398–399 199[77] Erdemir A and Fenske G R 1996 Tribol. Trans.

39 787[78] Wu W J and Hon M H 1999 Surf. Coat. Technol. 111 134[79] Scharf T W and Singer I L 2002 Tribol. Trans. 45 363[80] Erdemir A, Nichols F A, Pan X, Wei R and Wilbur P 1993

Diamond. Relat. Mater. 3 119[81] Donnet C, Belin M, Martin J M, Grill A and Patel V 1994

Surf. Coat. Technol. 68–69 626[82] Kim D S, Fischer T E and Gallois B 1991 Surf. Coat.

Technol. 49 537[83] Erdemir A, Bindal C, Pagan J and Wilbur P 1995 Surf. Coat.

Technol. 76–77 559[84] Erdemir A, Bindal C, Fenske G R, Zuiker C and Wilbur P,

1996 Surf. Coat. Technol. 86–87 692[85] Liu Y, Erdemir A and Meletis E I 1997 Surf. Coat. Technol.

94–95 463[86] Sanchez-Lopez J C, Erdemir A, Donnet C and Rojas T C

2003 Surf. Coat. Technol. 163 444[87] Hauert R 2004 Tribol. Inter. 37 991[88] Andersson J, Erck R A and Erdemir A 2003 Wear 254 1070[89] Gao F, Erdemir A and Tysoe W T 2005 Tribol. Lett. 20 221[90] Kim H I, Lince J R, Eryilmaz O L and Erdemir A 2006

Tribol. Lett. 21 53[91] Ban M, Ryoji M, Fujii S and Fujioka J 2002 Wear 253 331[92] Vercammen K, Van Acker K, Vanhulsel A, Barriga J, Arnsek

A, Kalin M and Meneve J 2004 Tribol. Int. 37 983[93] Wang E G 1997 Prog. Mater. Sci. 41 241[94] Khurshudov A G and Kato K 1996 Surf. Coat. Technol.

86–87 664[95] Fernandez A, Fernandez-Ramos C and Sanchez-Lopez J C,

2003 Surf. Coat. Technol. 163–164 527[96] Qi J, Chan C Y, Bello I, Lee C S, Lee S T. Luo J B. and Wen

S Z 2001 Surf. Coat. Technol. 145 38[97] Kato K. Koide H. and Umehara N. 2000 Wear 238 40[98] Zhou F, Kato K and Adachi K 2005 Tribol. Lett. 18 153[99] Cohen M L 1995 Mater. Sci. Eng. A 209 1

[100] Strondl C, van der Kolk G J, Hurkmans T, Fleischer W, TrinhT, Carvalho N M and de Hosson J T M 2001 Surf. Coat.Technol. 142 707

[101] Hovsepian P E, Kok Y N, Ehiasarian A P, Erdemir A Wen JG and Petrov I 2004 Thin Solid Films 447 7

[102] Gahlin R, Larsson M and Hedenqvist P 2001 Wear 249 302[103] Johnston S V and Hainsworth S V 2005 Surf. Eng. 21 67[104] Hershberger J, Ozturk O, Ajayi O O, Woodford J B, Erdemir

A, Erck R A and Fenske G R 2004 Surf. Coat. Technol.179 237

[105] Kalin M and Vizintin J 2005 Wear 259 1270[106] Alzoubi M F, Ajayi O O Woodford J B, Erdemir A and

Fenske G R 2002 Lubr. Eng. 58 21

R326

Topical Review

[107] Feng B, Cao D M, Meng W J, Rehn L E, Baldo P and Doll G2001 Thin Solid Films 298 210

[108] Lian G D, Dickey E C, Ueno M and Sunkara M K 2002Diamond. Relat. Mater. 11 1890

[109] Meng J W, Tittsworth R C and Rehn L E 2000 Thin SolidFilms 377 222

[110] Zehnder T and Patscheider J 2000 Surf. Coat. Technol. 133138

[111] Voevodin A A and Zabinski J S 1998 Diamond. Relat. Mater.7 463

[112] Meng W J and Gillispie B A 1998 J. Appl. Phys.84 4314

[113] Cao D M, Feng B, Meng W J, Rehn L E, Baldo P M andKhonsari M M 2001 Appl. Phys. Lett. 79 329

[114] Agreev V P, Glushko T N, Dorfman V F, Kuzmichev A V andPypkin B N 1991 Proc. SPIE 1503 453

[115] Le Harzic R, Huot N, Audouard E, Jonin C, Laporte P,Valette S, Fraczkiewicz A and Fortunier R 2002 Appl.Phys. Lett. 80 3886

[116] Kononenko T M, Garnov S V, Pimenov S M, Konov V I,Romano V and Borsos B 2000 Appl. Phys. A 71 627

[117] Dumitru G, Romano V, Weber H P, Haefke H and Gerbig Y2001 Proc. WLT Laser (Munich) p 351

[118] Dumitru G, Romano V, Weber H P, Pimenov S M,Kononenko T M, Hermann J, Bruneau S, Gerbig Y andShupegin M 2003 Diamond. Relat. Mater. 12 1034

[119] Voevodin A A, Bultman J and Zabinski J S 1998 Surf. Coat.Technol. 107 12

[120] Cuong N K, Tahara M and Yamauchi N 2003 Surf. Coat.Technol. 174–175 1024

[121] Aoki Y and Ohtake N 2004 Tribol Int. 37 941[122] Hauert R, Patscheider J, Knoblauch L and Diserens M 1999

Adv. Mater. 11 175

[123] Zhang W, Tanaka A, Xu B S and Koga Y 2005 Diamond.Relat. Mater. 14 1361

[124] Czigany Z, Neidhardt J, Brunell I F and Hultman L 2003Ultramicroscopy 94 163

[125] Neidhardt J, Hultman L, Broitman E, Scharf T W andSinger I L 2004 Diamond. Relat. Mater. 13 1882

[126] Erdemir A, Kovalchenko A, Mcnallan M J, Welz S, Lee A,Gogotsi Y and Carroll B 2004 Surf. Coat. Technol.188–189 588

[127] Vander Wal R L, Miyoshi K, Street K W, Tomasek A J,Peng H, Liu Y, Margrave J L and Khabashesku V N 2005Wear 259 738

[128] Erdemir A 2002 Mater. Res. Soc. Symp. Proc. 697 391[129] Erdemir A, Erylmaz O L, Nilufer I B and Fenske G R 2000

Surf. Coat. Technol. 133–134 448.[130] Erdemir A, Nilufer I B, Eryilmaz O L, Beschliesser M and

Fenske G R 1999 Surf. Coat. Technol. 121 589[131] Erdemir A, Eryilmaz O L, Nilufer I B and Fenske G R 2000

Diamond. Relat. Mater. 9 632[132] Donnet C and Grill A 1997, Surf. Coat. Technol. 94–5 456[133] Donnet C, Fontaine J, Grill A and Le Mogne T 2000 Tribol.

Lett.9 137

[134] Fontaine J, Belin M, Le Mogne T and Grill A 2004 Tribol.Int. 37 869

[135] Bowden F P and Young J E 1951 Proc. R. Soc. Lond. A 208444

[136] Gardos M N 1994 Tribology and wear behaviour of diamondSynthetic Diamond ed K E Spear and J P DismukesElectrochemical Society Series (New York: Wiley)pp 419–504

[137] Su C and Lin J C 1998 Surf. Sci. 406 149[138] Dag S and Ciraci S 2004 Phys. Rev. B 70 241401

R327