0018-8646/99/$5.00  (C)  1999 IBM
   
Plasma-deposited diamondlike carbon and related materials

A. Grill 


Diamondlike carbon (DLC) prepared by plasma-enhanced CVD is a metastable 
amorphous material characterized by attractive optical, electrical, chemical, 
and tribological properties. DLC films can be prepared at low temperatures from 
a large variety of precursors, and can be modified by the incorporation of 
different elements such as N, F, Si, or metals. The diversity of plasma 
conditions and precursors used for the deposition of DLC films provides 
flexibility to tailor their properties for specific needs and potential 
applications. This paper presents a review of the preparation of DLC and 
related materials, their properties, and their use in different applications. 

Introduction 

Diamondlike carbon (DLC) is an amorphous, in most cases hydrogenated, 
metastable material. The designation diamondlike has arisen because the 
material is characterized by properties that are to a certain extent similar to 
those of diamond: high hardness, high wear resistance and low friction 
coefficient, chemical inertness, high electrical resistance, and optical 
transparency in the visible and infrared. DLC films were first deposited 
byAisenberg and Chabot in 1971 by quenching a beam of C+ ions accelerated 
in a vacuum of 10**-6 Torr to a negatively biased substrate [1]. The 
attractive properties of the material have stimulated a large amount of 
research on their deposition and characterization, and on the development of 
relevant applications. 

Diamond and graphite are stable forms of carbon with well-defined 
crystallographic structures. Natural diamond is a crystalline material, and the 
diamond films fabricated by various CVD methods are composed of diamond 
microcrystallites up to tens of microns in size. Crystalline diamond is 
composed entirely of tetrahedrally coordinated sp3-bonded carbon. Diamond 
and diamond films are thus constituted of a well-defined material with fixed 
properties. In contrast, DLC films lack any long-range order and contain a 
mixture of sp3-, sp2-, and sometimes even sp1-coordinated carbon 
atoms in a disordered network. The ratio between the carbon atoms in the 
different coordinations depends on deposition conditions and, in hydrogenated 
DLC films, has been found to be a strong function of the hydrogen content of 
the films. While missing a long-range order, DLC may have medium-range 
ordering, as is discussed later. The properties of DLC therefore cover a wide 
range of values between those of diamond, graphite, and hydrocarbon polymers, 
as illustrated in Figure 1. 

"Pure" DLC composed of carbon and/or hydrogen can be classified as follows: 
hydrogenated amorphous carbon, a-C:H, and amorphous carbon, a-C. The first type 
contains from less than 10% [3] to 60% [4] hydrogen; incorporation of hydrogen 
in this type of DLC is important for obtaining diamondlike properties. The 
second type, amorphous carbon, contains less than 1% hydrogen. 

In the quest to change and improve the properties of DLC films, various 
materials derived from carbon films have been developed. Such materials are 
similar in structure to DLC but, in addition to carbon and/or hydrogen, they 
include nitrogen (NDLC or CNx films), silicon (SiDLC), fluorine (FDLC), and 
metal atoms (MeDLC). Most modifications have been made to DLC to reduce its 
(typically high) internal compressive stresses (N, Si, metal incorporation), or 
to reduce its surface energy for further lowering of its already low friction 
coefficients (F, Si-O incorporation) [5, 6]. In 1989 Liu and Cohen [7] made 
theoretical predictions of the potential existence of a crystalline material 
C3N4, similar to Si3N4. This has led to extensive studies of CNx films; 
however, in spite of various claims, such a crystalline structure has not 
yet been unequivocally identified. 

PECVD-deposited DLC films grow as continuous films, even at a thickness less 
than 50 nm; they are uniform over wide areas, and their roughness can be a 
fraction of an angstrom [5]. Because of their attractive properties, DLC films 
and their modifications have been technologically developed for a variety of 
applications, and new potential applications are being investigated. The 
following sections present a general review of the preparation techniques for 
DLC and related films and their structure, physical, and tribological 
properties, as well as an overview of the studies of such materials at the IBM 
Thomas J. Watson Research Center. A description of earlier studies at IBM can 
be found elsewhere [8]. 

Preparation techniques 

Since diamondlike carbon is a metastable material, DLC and related films must 
be deposited while their surfaces are continuously bombarded with energetic 
ions in order to obtain diamondlike properties. The metastable structure of DLC 
films most likely originates from the thermal and pressure spikes produced by 
the impinging energetic species on the growth surface [9]. Since the deposition 
by Aisenberg and Chabot, a variety of techniques based on beam and plasma 
techniques have evolved for the preparation of DLC. Such methods include single 
low-energy beams of carbon ions, dual ion beams of carbon and argon, ion 
plating, rf sputtering or ion-beam sputtering from carbon/graphite target, 
vacuum-arc discharges, or laser ablation. Details of the different deposition 
methods and of specific references can be found elsewhere [5]. While some of 
the mentioned methods are being extensively used in investigations of 
nonhydrogenated carbon (taC = tetrahedral carbon), the main techniques for 
depositing diamondlike carbon films are, however, based on plasma-enhanced 
chemical vapor deposition (PECVD) reactors and magnetron sputtering. 

PECVD deposition of DLC is generally performed in plasmas sustained by rf 
excitation [2, 10-13], but dc discharge systems [14-16], microwave plasmas 
[17], and ECR systems [18, 19] have also been used. High-density plasma (HDP) 
tools have been used recently for deposition of modified DLC films [20, 21]. 
PECVD deposition of DLC and related films is preferably performed in systems 
using parallel-plate reactors, as illustrated in Figure 2. Of the various 
possible geometries, parallel-plate reactors are preferred because they allow 
the deposition of uniform films over large areas and can be scaled up 
relatively easily for coating large areas. In the rf PECVD reactors, the 
substrates are placed on the rf-powered electrode (Figure 2), thus attaining a 
negative dc self bias, which is dependent on the precursor gases used, 
deposition parameters (e.g., power, pressure), and relative area of electrodes 
[22]. In dc PECVD the plasma is sustained by applying the negative dc voltage 
to the substrate electrode [16]. 

Any hydrocarbon can be used as a precursor for DLC deposition by the PECVD 
method, provided it has sufficient vapor pressure to be transported into the 
reactor. Details on the use of specific precursors can be found elsewhere [5]. 
Hydrogen or argon is sometimes added to the hydrocarbon precursor used in the 
PECVD reactors; however, the effect of such additions to the properties of DLC 
is not obvious [23, 24]. The growth rates of the films may decrease with 
dilution of the precursor with hydrogen, and in some cases the addition of 
hydrogen can have a detrimental effect on the film properties. In other cases 
it was found that under certain plasma conditions, the addition of argon to the 
hydrocarbon precursor can improve DLC properties [25, 26]. 

In addition to the large variety of precursors, DLC deposition conditions also 
spread over large ranges of parameters. DLC depositions have been carried out 
in rf PECVD systems at an rf power of 15-1000 W, a negative substrate bias of 
100-1500 V dc, a pressure of 6 mTorr-7.5 Torr, and a substrate temperature 
between room temperature and 250[degree]C. While power density is a more 
significant deposition parameter than rf power, it is not always possible to 
determine this value from published results, because reported depositions have 
been performed in reactors of different sizes and geometries. Nevertheless, it 
seems that power densities of the order of 0.1-0.5 W/cm**2 are generally 
used to obtain DLC films in PECVD systems. Deposition of DLC must be carried 
out at substrate temperatures below 325[degree]C to prevent graphitization 
[14]. 

In the PECVD processes, DLC grows mainly by condensation of the radicals formed 
in the plasma; however, while the concentration of the ions in the plasma is 
several orders of magnitude smaller than that of the radicals, ion bombardment 
is an important factor affecting the growth and properties of the films. With 
use of the right conditions, the films deposited on the biased electrode of an 
rf reactor are hard DLC. Under the same conditions, the films deposited on the 
grounded electrode are soft and polymerlike. The deposition rate on the biased 
electrode is also much higher than on the grounded electrode. This can be 
explained by assuming that the ion bombardment leads to increased fragmentation 
of the hydrocarbon species arriving at the surface to ever more 
hydrogen-deficient radical species, resulting in an increased sticking 
coefficient on the bombarded surface [27]. The near-surface excited region is 
quenched by the underlying cold substrate. Angus and Jansen [28] have proposed 
that the composition of a DLC film, and its average nearest-neighbor 
coordination, adjust to a minimum possible energy state through the 
incorporation of hydrogen. 

Deposition of DLC films by PECVD is usually performed in rf-sustained plasmas, 
although dc plasmas have also been used for PECVD of DLC films. It is generally 
believed that because of the insulating characteristic of DLC, it may not be 
possible to apply the dc voltage through the films in order to sustain the 
plasma and grow them. Therefore, dc deposition of DLC has been performed in 
PECVD systems with an additional screen in front of the sample electrode [14, 
15] or by using a triode configuration [29]. However, it has been shown that 
highly insulating DLC films can be deposited on electrically conductive 
substrates using a dc-powered diode-type parallel-plate reactor [16]. The 
deposition is possible at sufficiently high bias because, as shown later, the 
resistivity of the DLC films is not constant and drops several orders of 
magnitude at high electric fields, enabling bias transfer to their surface. DLC 
films several micrometers thick have been grown by dc PECVD at high rates using 
low power densities (~0.25 W/cm**2) [16]. 

Modified DLC films can be prepared by PECVD by admixing volatile precursors of 
the modifying elements to the main gas feed. Thus, NDLC (or CNx) films have 
been prepared from mixtures of hydrocarbons with nitrogen [30], SiDLC films 
have been prepared by adding diluted SiH4 to the hydrocarbon [31], and FDLC 
films have been deposited from pure fluorocarbons or fluorocarbons diluted with 
hydrogen [32]. Metal incorporation was achieved by sputtering a metal or metal 
carbide target in a plasma of argon with hydrocarbon mixture [33, 34]. 

Bias and temperature effects 

The substrate bias appears to be the dominant factor determining the properties 
of DLC films. Their hardness, density, and refractive index increase with 
increasing bias, while their hydrogen content and optical gap usually decrease 
with increasing bias [5, 16, 25]. It should, however, be noted that in most 
PECVD deposition cases the bias is not controlled directly as an independent 
process variable. In rf PECVD deposition systems, the substrate bias is 
frequently changed by changing the rf power (W) and/or the pressure in the 
reactor (p). According to Catherine [29], the self-bias of the substrate 
(V[SUB]B) in a specific reactor is related to these two parameters by the 
expression 

V[SUB]B [open infinity] (W/p)**1/2.

The substrate bias in PECVD systems can thus be varied by adjusting the bias 
directly, or by adjusting the rf power and/or the pressure in the reactor. 
Varying the substrate bias by changing the rf power affects not only the 
phenomena occurring at the surface of the growing film but also the 
fragmentation of the precursors in the plasma. Reports about the "effect of 
bias" on the properties of DLC films may therefore reflect the combined effects 
of two independent plasma parameters, W and p, and thus variations in the 
active source species produced in the plasma by changes in the degree of 
fragmentation of the precursor gas. In addition, for identical gases and 
identical values of pressure and rf power, the self-bias also depends on the 
reactor geometry, making it difficult to extrapolate results from one reactor 
to another. 

At the relatively high pressures used in PECVD of DLC films, the ions are 
scattered by collisions and reach the substrate with a distribution of energies 
lower than the applied bias. The average impact energy (E) of ions bombarding 
the growing film in the rf plasma therefore depends on the dc bias and on the 
pressure in the reactor, according to the relation [35] 

E [open infinity] V[SUB]B/p**1/2.

It has been claimed [2] that a substrate bias higher than 100 V dc is required 
to obtain hard DLC-type films. However, this lower bias threshold can depend on 
the pressure during deposition, because operation at high pressure reduces the 
average energy of species impinging on the growth surface. At sufficiently low 
pressures, the threshold value is relaxed, and hard, wear-resistant films can 
be obtained at negative bias smaller than 100 V dc [8, 13]. 

The minimum bias required to obtain diamondlike characteristics can also be 
affected by the precursor used to deposit the films. Thus, for films deposited 
by dc PECVD, a higher bias is required to obtain films with DLC characteristics 
from cyclohexane than is required for films deposited from acetylene [25]. 
However, for the same deposition system, enhancement of the ion bombardment of 
the films by addition of argon to the precursor enables the preparation of hard 
DLC films from cyclohexane at the lower bias. 

The substrate temperature during deposition is another parameter which can 
affect the properties and deposition rate of DLC films. The deposition rate of 
DLC increases with decreasing temperature and increasing ion bombardment. The 
decrease of the rate of deposition with increasing substrate temperature can be 
explained by assuming that growth takes place through the interaction between 
the surface and a physisorbed layer of hydrocarbon. This assumption is 
consistent with the low values of growth activation energies (0.05-0.5 eV) and 
with the increase of the deposition rate with increasing ion bombardment [29]. 

It was reported that the optical gap of DLC films decreases with increasing 
substrate temperature, with a rapid decrease taking place above 250[degree]C 
[14, 36]. While some properties of DLC films, such as their index of refraction 
[37] or wear resistance, are a weak function of the deposition temperature, 
substrate temperature can have a significant impact on their hydrogen content 
and thermal stability [38]. Deposition at temperatures above 325[degree]C has 
been found to result in a sharp decrease of their electrical resistivity and in 
significant film softening [14]. DLC deposition is therefore performed at 
substrate temperatures well below 300[degree]C. Low deposition temperature may 
lead in certain conditions to the formation of a polymerlike film, or a 
polymerlike fraction embedded in a hard DLC layer [13]. 

Structure and composition of DLC and related materials 

The properties of DLC films are determined by the bonding hybridization of the 
carbon atoms and the relative concentration of the different bonds, i.e., 
sp3:sp2:sp1, as well as by the hydrogen distribution among the 
different types of bonds. In hydrogenated DLC films, some of the carbon bonds 
are terminated by hydrogen. A fraction of the hydrogen can be incorporated as 
single atoms, as well as molecules in voids. Voids have been assumed to explain 
the hardness or densities of DLC films [39, 40]. While direct evidence of such 
voids has not been reported, it has been shown that depending on deposition 
conditions, more than 50% of the hydrogen incorporated in DLC films can indeed 
be unbounded hydrogen [16, 41]. 

The structure of hydrogenated DLC films can be described as a random network of 
covalently bonded carbon in the different coordinations. Robertson [42] claimed 
that in addition to the short-range order defined by carbon hybridization and 
hydrogen content, a substantial degree of medium-range order on the ~1-nm scale 
also exists in the films. The material can accordingly be described as a 
network of graphitic clusters of fused sixfold rings, linked into islands by 
sp3 bonds. The formation of the graphitic clusters relieves the strain in 
the DLC structure. According to Robertson's model, the existence of the optical 
gap in DLC films is the result of the correlation of hexagonal rings on a very 
fine scale and does not necessarily prove a preponderance of diamondlike 
sp3 bonding [42]. The width of the optical gap was claimed to depend on 
the degree of medium-range order and was found to vary inversely with the 
square root of the size of the clusters, being proportional to 6N-1/2, N being 
the number of aromatic rings in a cluster. 

Angus and Jansen have described the structure of hydrogenated DLC (a-C:H) by a 
random covalent fully constrained network (FCN) model [28]. The model was based 
on the assumption that a random covalent network is mechanically constrained 
when the number of constraints equals the number of mechanical degrees of 
freedom. The authors found that in a covalent network composed solely of 
sp2 and sp3 carbon sites, in which some clustering of aromatic rings 
occurs, the optimal ratio of the number of sites for each coordination, 
designated in [28] as sp3/sp2, is determined by the atomic fraction 
of hydrogen incorporated in the film, X[SUB]H, and is given by 

sp3/sp2=6X[SUB]H-1/8-13X[SUB]H.

The DLC structure can be described according to the FCN model as a 
three-dimensional array of rings, containing mostly six-membered rings but also 
five- and seven-membered rings. Such a cross-linked structure lacks long-range 
order, and the material appears amorphous [28]. According to Equation (3), the 
optimal coordinated carbon network can exist only for bound hydrogen 
concentrations 0.167 < X[SUB]H < 0.615. The FCN model thus defines the range of 
hydrogen concentrations for which it is possible to obtain amorphous 
hydrogenated carbon with diamondlike properties, but does not indicate the 
distribution of hydrogen among the different carbon sites. The model takes into 
account only the concentrations of hydrogen bound to carbon atoms. Since some 
hydrogen can be unbound, the total hydrogen content in DLC films may extend 
beyond the upper limit. 

o   Bonding characterization 

The structural characterization of DLC films is complicated by their amorphous 
nature. A variety of spectroscopic techniques, such as Fourier transform 
infrared absorption (FTIR), Raman, NMR, and electron energy loss spectroscopy 
(EELS), have been used to characterize the structure and composition of DLC 
films. While each of these techniques provides some qualitative information 
about their structure, the interpretation of the results is often indirect and 
sometimes even misleading. 

FTIR is often used to characterize hydrogenated DLC and related films. Typical 
for the FTIR spectra of hydrogenated diamondlike carbon is the wide absorption 
band centered at about 2900 cm**-1. This band is a superposition of 
absorption peaks of stretching vibrations from different spm CHn 
configurations, with n, m = 1-3 [5]. FDLC films are characterized by a large 
absorption band centered at about 1300 cm**-1, corresponding to the 
superposition of CFn peaks (n = 1-3), and a smaller peak centered at about 
1750 cm**-1, corresponding to superposition of C-CF2 and FC=CF2 
absorption peaks [32]. The incorporation of nitrogen in DLC results in a strong 
increase in the IR absorption band at 1700-1100 cm**-1, a broad absorption 
band at about 3380 cm**-1, and a narrower peak at 2210 cm**-1 [30]. 
Typical FTIR spectra are illustrated in Figure 3, and the interpretation of the 
absorption peaks can be found in the papers referenced above. 

Because each individual spm CHn configuration is characterized by a 
specific IR absorption peak, it is often assumed that one can use the 
deconvoluted FTIR peaks to analyze the relative hybridization ratio, 
sp3/sp2, of the carbon atoms and use the total intensity of the broad 
peak centered at 2900 cm**-1 to quantify the hydrogen content of the DLC 
films [12, 43, 44]. Such analyses have been performed assuming a constant 
oscillator strength and no occurrence of preferential hydrogen bonding to a 
particular type of carbon. However, the proportionality coefficients may differ 
by an order of magnitude for the -CH bond and the -CH2 or -CH3 bonds 
[45]. Therefore, FTIR cannot be used for determination of the carbon 
hybridization or hydrogen content of the films [41]. It may, however, be used 
for qualitative comparisons of films deposited in the same reactor, and can be 
used together with forward recoil elastic scattering (FRES) measurements of 
hydrogen content in the films to determine relative fractions of unbound 
hydrogen in similar films [41], or to evaluate the general types of bonding 
between different elements in DLC films or modified films [30, 32]. FTIR 
spectroscopy may also be used for qualitative identification of changes in a 
film undergoing postdeposition treatments, such as thermal annealing [11, 12, 
26, 37]. 

EELS is another technique used in determining the sp2 fraction in 
nonhydrogenated DLC, especially in the so-called tetrahedral amorphous carbon 
(taC), and then estimating the fraction of sp3 hybridized carbon (sp3 
fraction = 1 - sp2 fraction). The method has also been attempted on 
hydrogenated DLC [46], and it has been found that the films are sensitive to 
the electron dose exposure. By using low electron-beam doses to minimize 
irradiation damage, it was found that the sp3/sp2 ratio increases 
with increasing hydrogen content in the films. However, these findings are only 
qualitative for hydrogenated DLC films. 

High-resolution 13C NMR spectroscopy appears to be the most reliable 
method for determining the relative concentrations of sp2- and 
sp3-hybridized carbon, as well as the local atomic environment of carbon 
in each state in hydrogenated DLC [13, 47-49]. Taking the spectra both with and 
without the use of proton decoupling, one can identify the carbon atoms to 
which at least one hydrogen atom is bound [13]. Because the C-H spin 
interaction causes the carbon peak to be broadened into the baseline of the 
spectrum, only the signal from carbon atoms with no bound hydrogen is obtained 
when the proton signal is not decoupled. With the natural abundance of about 1% 
13C, about 100 mg of DLC is required in order to obtain a significant 13C NMR 
signal. Such studies have shown that, within detectable limits ([> or =]2%), 
no "pure diamond" (tetrahedral carbon-to-carbon bonding) was found in the DLC 
films, indicating that hydrogen is required to obtain and stabilize the sp3 
coordination. 

o   Hydrogen in DLC 

Incorporation of hydrogen in the carbon films deposited by PECVD or sputtering 
appears to be generally required in order to obtain "diamondlike" optical and 
electrical properties. The hydrogen content in DLC films and its depth profile 
with a resolution of 4 nm can be measured by the nuclear reaction 
H(15N[alpha], [gamma])C: 

15N + 1H --> 12C + 4He+[gamma].

The measurement is done by bombarding the films with 15N ions and 
counting the 4.33-MeV [gamma] rays produced by the reaction, which is resonant 
at 6.4 MeV [13]. A more accessible and generally used method for measurement of 
the total hydrogen content of DLC films is the forward recoil elastic 
scattering (FRES) or elastic recoil detection (ERD) technique [38, 50]. While 
FRES is simpler than the nuclear reaction method, it has far inferior depth 
resolution (about 50 nm vs. 4 nm for nuclear reaction). 

The total hydrogen content critically determines film structure at the atomic 
level (the ratio between sp3- and sp2-coordinated carbon atoms), and 
therefore the physical properties of the films. Hydrogen content is also key to 
obtaining a wide optical gap and high electrical resistivity, because it 
passivates the dangling bonds in the amorphous structure [51]. DLC films 
deposited from hydrocarbons must include hydrogen in concentrations from 17% to 
61% in order to obtain diamondlike properties [28]. The loss of hydrogen 
through annealing at high temperatures generally causes a collapse of the 
structure to a graphitelike phase dominated by sp2 bonds [37, 52]. As is 
discussed later, the hydrogen content of DLC films is also an important factor 
controlling their tribological properties [53]. It was suggested in 1981 that 
significant amounts of hydrogen in DLC films may not be bonded to carbon atoms, 
but may be trapped interstitially [40]. The existence of unbound hydrogen was 
proven by comparing FTIR and FRES data of different DLC films, and it was found 
that in some cases the fraction of unbound hydrogen in DLC can be up to about 
50% of the total hydrogen incorporated in DLC films [16, 41]. 

o   Optical properties 

DLC films are typically transparent in the infrared, with the exception of the 
CH absorbing bands mentioned earlier, are weakly absorbing in the visible 
spectrum, and are increasingly absorbing with decreasing wave length in the UV. 
Their hydrogen content is critical in controlling their optical properties, and 
removal of hydrogen from hydrogenated DLC films causes the loss of IR 
transparency [54, 55]. 

A wide range of optical gap values (E[SUB]opt), spanning the range from 0.38 to 
2.7 [2, 56-58], were reported for DLC films prepared under presumably similar 
conditions, indicating the dependence of the property on the deposition system. 
For otherwise similar deposition conditions, E[SUB]opt was found to decrease 
strongly for DLC films deposited above 250[degree]C [5, 14]. This behavior 
reflects the role of hydrogen in stabilizing the structure of DLC layers. Films 
deposited at lower temperatures (in the range 25-250[degree]C) contain 
significant concentrations of hydrogen; most of the sp3-coordinated carbon 
atoms, as well as a substantial fraction of the sp2-coordinated carbons, 
are bound to at least one hydrogen atom. When the release of hydrogen is 
induced in these materials, they revert to the configuration of lowest energy, 
i.e., graphite. 

The index of refraction (both the real part n and imaginary part k) and its 
spectroscopic variation have been found to be dependent on the preparation 
conditions and hydrogen content of the films. Its value at 632.8 nm has been 
found to increase from 1.7 to 2.25, with an increasing rf power/pressure ratio 
during deposition [59]. The dependence of the index of refraction on the 
wavelength is strongly connected to the preparation conditions and can show 
opposite behaviors [60, 61]. The index of refraction at 632.8 nm of DLC films 
deposited from acetylene by rf PECVD at a range of substrate temperatures, a 
bias of -80 V, and 0.1 W/cm**2 was found to be insensitive to the deposition 
temperature and to have a value of n = 1.9 +- 0.2 [37]. It is, however, 
dependent on the ion bombardment of the growing film and increases with 
increasing ion bombardment. Ion bombardment increases with increasing bias, 
decreasing pressure, and increasing Ar dilution, when Ar is added to the 
precursor [25], resulting in an increasing index of refraction, as illustrated 
in Figure 4. 

The index of refraction is also affected by the hydrogen concentration in the 
DLC films, as determined by the deposition condition mentioned above. It 
generally increases with decreasing hydrogen content, as shown elsewhere [25, 
38, 62]. It is, however, dependent on the concentration of bound hydrogen and 
not total hydrogen content in the film [16]. A higher index of refraction 
usually indicates DLC with stronger cross-linking, higher hardness, and better 
wear resistance [25]. 

o   Electrical properties 

Owing to the range of properties of the naturally occurring forms of carbon, 
i.e., graphite and diamond, the electrical properties of diamondlike carbon 
films can vary from that of a semimetal to that of a wide-bandgap insulator. 
Electronic transport itself is quite complex in DLC as a result of its 
disordered structure. Diamond is characterized by high electron mobilities and, 
since it is a wide-bandgap semiconductor, diamond-based devices such as diodes 
can continue to function properly at temperatures in excess of 500[degree]C; a 
wide variety of applications for such high-temperature electronics are being 
considered. In contrast, DLC has neither the wide energy gap nor the thermal 
stability to operate in such an environment. The electronic and optical 
properties of hydrogenated DLC behave to a certain extent in a fashion similar 
to those of hydrogenated amorphous silicon. 

The electrical properties of DLC films have been modeled assuming that the band 
structure consists of only a mobility gap and that carriers residing in gap 
states are localized [63]. This mobility gap produces semiconductor behavior, 
but the existence of a high density of such gap states significantly degrades 
the otherwise desirable properties of semiconducting materials. A more 
quantitative description of the band structure of diamondlike carbon layers can 
be found elsewhere [4, 63]. Hydrogenated DLC can be doped similarly to 
amorphous silicon; however, the doping is inefficient, since high levels of 
doping are required to produce significant changes in its electrical properties 
[63, 64]. 

Diamondlike carbon films are generally characterized by high electrical 
resistivities spanning a large range of values, from 10**2 to 10**16 
[Omega]-cm, depending on the deposition conditions [14, 36, 64]. Their 
electrical resistivity has been found to be strongly reduced by up to seven 
orders of magnitude by incorporation of metals [33, 65]. Nitrogen incorporation 
also reduces the electrical resistivity of the films. 

The values of electrical resistivities of the DLC films discussed above have 
generally been reported at low electric fields, and in most cases no indication 
has been given of the dependence of the resistivity on the electric field. 
Electrical resistivities in fields up to several MV/cm have been measured by 
Grill et al. for DLC and nitrogen-doped DLC films deposited from several 
precursors at a variety of conditions [64]. The electrical resistivities are 
shown in Figure 5 as a function of the electric field E. As can be seen, the 
resistivities vary strongly with E, indicating that the DLC films do not 
exhibit ohmic behavior. The resistivities of these DLC films change only 
slightly up to a field of 3 x 10**5 V/cm, but decrease sharply at higher 
fields. Furthermore, the resistivities are strongly dependent on deposition 
conditions, and the resistivities of dc PECVD films deposited at identical 
conditions are strongly dependent on the precursor used [see Figure 5(a)]. As 
can also be seen in Figure 5(b), the incorporation of nitrogen into the DLC 
films reduced their resistivity over the whole range of electric fields. 
Incorporation of 11% nitrogen (film N2) decreased the resistivity by about five 
orders of magnitude compared to the undoped film (CY1) [64]. 

Hydrogen, which stabilizes the sp3 bonds and determines the carbon 
hybridization ratio, is therefore required for obtaining a high electrical 
resistivity. However, electrical resistivity differences of several orders of 
magnitude have been found between films having small or no differences in total 
hydrogen content [64]. Because the electrical properties of DLC films depend on 
the carbon hybridization, which in turn is affected by C-H bonding, the 
electrical conductivity is determined by the bound and not by the total 
hydrogen concentration in the films and, for films deposited under similar 
conditions, the conductivity decreases with increasing concentration of bound 
hydrogen in the films [64]. 

Special consideration has recently been given to the dielectric constant (k) of 
DLC and FDLC films [66]. Low-k materials are needed for the "back-end-of-line" 
(BEOL) multilevel interconnect structures of ULSI circuits to improve their 
performance. It was found that by adjusting deposition conditions, it is 
possible to change the dielectric constant of DLC films in the range 2.7-3.8, 
as illustrated in Figure 6. The integration of such films in BEOL structures, 
however, imposes further requirements on the low-k material, such as low 
stresses and thermal stability at 400[degree]C. It has been found that for DLC 
films, the intrinsic stresses decrease with decreasing k values, while the 
thickness changes increase with decreasing k values, as illustrated in 
Figure 7. The increased changes reflect the reduced thermal stability of films 
with low k values. 

FDLC films with dielectric constants k < 2.8 (see Figure 8) have been prepared 
having intrinsic stresses below 200 MPa. The FDLC films seem to be stabilized 
after a first annealing at 400[degree]C in an inert ambient. Such films are 
therefore potential candidates as low-k insulators in BEOL interconnect 
structures; however, integration issues, especially those related to 
potentially reactive fluorine, have yet to be effectively addressed. 

o   Chemical resistance 

At room temperature, DLC films are chemically inert to practically any solvent 
and are not attacked by acids, alkalis, or organic solvents. The films are 
inert even to strong acid mixtures, such as the "acid etch" (HNO3:HF = 
7:2) and to exposure to alkali solutions at 85[degree]C for several hours. As a 
result of their chemical resistance, DLC films can be used as 
corrosion-resistant coatings. The films and their modifications can be removed 
from a substrate by exposure to atomic oxygen or fluorine species generated in 
a plasma, which react with the carbonaceous films to produce volatile COx 
and CFy species which are pumped out of the system. Reactive ion etching 
in oxygen- or fluorine-containing plasmas can be used to pattern DLC films. The 
rate of etching can be strongly modified by incorporating Si in the films, and 
SiDLC can be used as an etch stop during the etching of DLC or FDLC films [31]. 

o   Mechanical properties 

DLC films are characterized by high hardness, spanning the range from 10 to 30 
GPa [68], associated with high intrinsic compressive stresses in the range from 
0.5 to 7 GPa [9, 10, 13, 38, 39, 69]. The high internal stresses limit the 
thickness of films that can be used for various applications, often to less 
than 1 [muon]m. Adhesive forces must overcome the high stresses in order to 
prevent delamination of the film from the substrate. Some of the mechanical 
properties of DLC have been explained by Robertson [68] using the constraint 
counting model of Phillips and Thorpe [69, 70], taking into consideration the 
elasticity of the individual bonds and the connectivity and coordination of the 
network. 

The stresses in DLC films are generally directly correlated to the fraction of 
sp3 carbon [71]. For rf PECVD films, the intrinsic stresses therefore 
depend on the combination of deposition parameters which affect the sp3 C 
hybridization. Because of the interdependence of the plasma parameters, stress 
dependency reported with regard to one specific parameter is valid for only a 
specific plasma system. Stresses were found to decrease with increasing 
deposition pressure [10, 39] and to have a bias relation which depends on other 
deposition parameters [10, 70, 71]. 

Of the factors affecting the sp3 hybridization, the more important one is 
the average energy of the ions bombarding the film, which accounts for both 
substrate bias and pressure in the plasma and is given by Equation (2). Figure 
9, which presents a plot of the stresses in DLC films deposited by dc PECVD 
from several precursors as a function of the average ion energy, shows that the 
stresses in the films deposited from one precursor do increase monotonically 
with increasing average ion energy. However, for identical average ion 
energies, the stresses are strongly dependent on the gases used to deposit the 
films. The stresses in films deposited from acetylene by rf PECVD at similar 
power densities [38] are also shown in Figure 9. Although the rf films were 
deposited at conditions corresponding to much lower ion energies (E values), 
the stresses are much higher in these films than in the dc films. The results 
shown in Figure 9 thus indicate that the factors determining the internal 
stresses in DLC films are much more complex than just the direct effects of 
bias or even of average ion energy. 

The intrinsic stresses in DLC films are also correlated with changes in 
hydrogen content obtained by changing deposition parameters [39, 72]. Strong 
variations in stresses have been observed in DLC films in which the differences 
between the total hydrogen contents were too small to predict or explain the 
large differences in the stresses [16, 24]. In addition, it was observed that 
films containing smaller amounts of hydrogen had higher stresses than DLC films 
containing higher amounts of hydrogen [24]. As illustrated in Figure 10 for 
dc-deposited DLC films, the stresses increase with an increasing fraction of 
unbound hydrogen, without reflecting a direct precursor dependence. 

Stresses in DLC films have been decreased by doping the films with nitrogen 
[24, 30, 73] to values as low as 0.22 GPa [24]. Such NDLC films contain 
significant numbers of NH bonds, which increase with increasing concentrations 
of nitrogen [30]. The stress reduction in nitrogen-doped DLC films was 
explained [24, 73] by using the overconstraining model of DLC [28]. According 
to this model, replacement of CH with NH bonds in NDLC reduces the average 
coordination number and the overconstraining in the NDLC films, resulting in 
decreased internal stresses. However, other effects may contribute to the 
stress reduction in DLC as a result of nitrogen incorporation. 

o   Tribological properties 

Wear protection is needed for surfaces in kinetic contact, in situations which 
generally also require low friction between the moving surfaces. As a result of 
their high hardness, DLC films are utilized as wear-resistant protective 
coatings for metals and for optical or electronic components. The use of DLC is 
especially attractive in applications where the thickness of the protective 
film is limited to less than 50 nm, as for example in the case of magnetic 
recording media, where the trend toward higher-density data storage has led to 
the requirement of very low flying heights between a disk and a recording head. 
The coating must be able to protect the magnetic media against wear and 
corrosion but must also be thin enough not to impede the achievement of high 
recording density. 

Tribological (friction and wear) characterization is usually performed on a 
pin-on-disk apparatus [8]. The disk coated with the DLC film is rotated at a 
fixed speed, while a pin with a spherical tip is loaded against it with a 
chosen normal load. After a number of rotations, the wear of the disk is 
calculated from the cross section of the profile of the produced wear track. 
The friction force between the pin and the disk can generally also be measured 
in "pin-on-disk" tools, thus enabling the determination of the friction 
coefficient. Because detailed reviews of the tribology of DLC and related 
materials can be found elsewhere [6, 74], the following contains only a summary 
of the tribological properties of DLC films. 

A variety of factors contribute to friction and wear, yet friction is 
controlled mainly by the formation and breaking of chemical bonds at the 
interface between the sliding parts, while breaking of bonds within a part 
controls the wear [75]. A low friction coefficient and low wear are two 
attractive properties of diamond. However, the friction coefficient of diamond 
depends strongly on the environment and is mainly controlled by surface 
dangling bonds. Diamond has a low friction coefficient in humid air or dry 
nitrogen because of the presence of a contaminant layer of low shear strength 
on its surface. Clean diamond surfaces can interact strongly and result in high 
friction coefficients. The friction coefficient of diamond against diamond in 
vacuum can increase to values as high as 1.0 when an increase of the 
temperature causes the desorption of surface hydrogen, leaving interacting 
dangling bonds [75]. Saturation of dangling bonds by atomic hydrogen or other 
adsorbates reduces the friction coefficient to [muon] = 0.02. There is also 
evidence that the debris formed during sliding is contributing to the low 
friction in diamond [76]. Note that water [77-79] and atomic hydrogen [80] act 
as lubricants for graphite and improve its wear and lubricating behavior. 

In spite of the high chemical inertness of DLC films, their tribological 
behavior is controlled to a large extent by their surface chemistry. The latter 
is dependent on the method used for the preparation of the films, but can also 
be affected by the testing environment. The tribochemical behavior of DLC films 
(i.e., the effect of environment on their friction and wear) is well accepted 
for DLC film tribology [74]. 

As summarized in other reviews [6, 74], DLC films deposited by PECVD have 
several common tribological characteristics: 

1. DLC films display a behavior similar to that of diamond in UHV but not in a 
humid atmosphere. 

2. The friction coefficient of hydrogenated DLC is low in humid nitrogen or 
oxygen, extremely low in dry nitrogen or ultrahigh vacuum (UHV), and very high 
in dry oxygen. 

3. Loss of hydrogen through annealing at high temperature causes a marked 
increase in the friction coefficient in UHV, but not in a humid environment. 

4. Both friction and wear can be affected by a transfer layer which forms in 
most cases during friction. The transfer layer in the wear scars was observed 
directly with microlaser Raman spectroscopy, which showed that the layer has a 
disordered graphite structure that is different from that of the original DLC 
film [81]. Raman spectra taken from the wear track of the DLC film also 
revealed evidence of graphitization. 

Initial explanations of the tribological behavior of DLC materials were based 
on their mechanical properties and surface energy [34]. The films attain 
hardness values (H) typical of inorganic ceramics, but have Young's modulus 
values (E) significantly lower than those of ceramics. The corresponding H/E 
ratios are relatively high, corresponding to those of polymers. The surface 
energy (S) of DLC is also more like that of polymers than metals or ceramics. 
The combination of the high H/E with low S/H ratios (S/H values being smallest 
among the other type of materials) was used to explain the low adhesion of 
other materials to DLC films, their low friction coefficient, and low wear 
[34]. The excellent tribological performance of these materials was attributed 
to a combination of favorable properties which are ceramiclike on the one hand 
(high hardness) and polymerlike on the other (high elasticity, H/E, and low 
surface energy). 

However, it has become clear in recent years that in all environments the 
tribological behavior of DLC films is controlled by an interfacial transfer 
layer formed during friction. The transfer layer is formed by a 
friction-induced transformation of the top layer of a DLC film into a material 
of low shear strength. This transformation may be brought about by 
friction-induced annealing caused by thermal and strain effects generated 
during sliding [81]. The shear strength of the transfer layer and its adhesion 
to sliding surfaces can be affected by the environment and by contact load and 
sliding speed. The composition of the transfer layer can also be affected by 
the material of the sliding counterpart. The low friction and ultralow wear of 
DLC and counterparts can be explained by the low shear strength of the transfer 
layer [6], which can also be affected by the testing environment [82]. 

A recent compilation of friction coefficients of DLC films [83] shows that 
their friction coefficients span a range of [muon] = 0.007-0.4 in vacuum, below 
10**-4 Pa, while in ambient air at relative humidities of 20% < RH < 60% 
they span a range of [muon] = 0.05-1.00. Most typical, however, are the ranges 
[muon] = 0.007-0.02 in vacuum and [muon] = 0.1-0.4 in ambient air. The large 
spread in the values of the friction coefficient are caused by variations in 
the structure and composition of the films. The transfer layer described above 
has a lubricating effect, and its formation can be enhanced by hydrogen but may 
be restricted in the presence of water or oxygen. The increase of the friction 
coefficient of hydrogenated DLC films in humid air was attributed by Gardos to 
the increase in the van der Waals bond strength of hydrogen bonding to adsorbed 
water molecules (~5 kcal/mol) compared to the bonding of hydrocarbons (~2 
kcal/mol) [75]. 

Hydrogen passivates the dangling bonds in hydrogenated DLC films and permits 
only weak interactions between the DLC and the sliding partner. When hydrogen 
is lost from a hydrogenated DLC film by annealing, the formed dangling bonds 
cause strong interactions between the surfaces in contact, resulting in 
increased friction in UHV or dry nitrogen, similar to that reported for both 
diamond or graphite [80, 84]. It was shown that removal of hydrogen from DLC by 
annealing the films above 550[degree]C caused an increase in the friction 
coefficient to [muon] = 0.68 in UHV or dry nitrogen, indicating that the 
presence of hydrogen in these films is essential for obtaining very low 
friction coefficients [76]. It has recently also been shown that the 
achievement of a low friction coefficient in UHV requires a hydrogen content of 
more than 40% [53]. 

Certain applications may require modification to preserve wear resistance but 
change other characteristics (e.g., increase electrical conductivity or reduce 
surface energy). It has been shown that incorporation of F, O, N, and Si in DLC 
could modify water wetting angles of the modified films [85]. The incorporation 
of F or Si increased the contact angle of water, while the incorporation of O 
and N decreased the contact angle. The effect of F and Si on the surface energy 
of DLC films was attributed to the reduction primarily of the polar part of the 
surface energy, due to the loss of sp2 C hybridization and dangling bonds 
[85]. Yet DLC films containing F and Si can be prepared to be as wear-resistant 
as the pure hydrogenated DLC [32, 86], as illustrated in Figure 11 for FDLC 
films deposited by dc PECVD. Similar results have been obtained for FDLC films 
deposited by rf PECVD [86]. The incorporation of Si in DLC films has been 
reported to render their friction against steel and the wear of the steel 
counterpart insensitive to moisture [87]. Me-DLC films containing 15% Ta, W, 
Ti, or Nb, having wear resistance similar to that of DLC and friction 
coefficients <0.2, but higher conductivity (up to 0.005 [Omega]-cm), have been 
deposited by dc magnetron sputtering of the metals in acetylene [85]. Modified 
DLC films incorporating different elements such as N, Si, and metals can 
display tribological properties similar to those of pure DLC films, and these 
properties are affected more by preparation conditions and testing environments 
than by the incorporated element. 

The dependency of the tribological behavior of DLC materials on preparation 
parameters and tribotesting conditions is summarized in Figure 12. 

Applications 

The unique properties of DLC films and their modifications, together with the 
possibility of adjusting the properties by choosing appropriate deposition 
parameters, make them suitable for a variety of applications. The exploited 
properties include high wear resistance and low friction coefficient, chemical 
inertness, infrared transparency, high electrical resistivity, and, 
potentially, low dielectric constant. Owing to their IR transparency, DLC films 
can be used as antireflective and scratch-resistant wear-protective coatings 
for IR windows or lenses made of Ge, ZnS, or ZnSe, where they also act as an 
antireflective coating at the used wavelength of 8-13 [muon]m [88, 89]. 
Aluminum mirrors used in optical imaging deteriorate with time as a result of 
exposure to atmosphere. This deterioration can be prevented by coating the 
surface of the mirrors with DLC films [90]. Because of their absorbance in 
visible light, DLC coatings can be applied as scratch-resistant coatings on 
sunglasses. The low deposition temperatures of DLC films allow their use as a 
wear-protective layer on products made of plastic. Currently, DLC films are 
used for protection against abrasion of sunglass lenses made of polycarbonate 
[91]. 

The most widespread use of DLC films is in wear and corrosion protection of 
magnetic storage media. Nanosmooth DLC films are especially useful where the 
protective coating must be very thin (<50 nm), such as in magnetic hard drives. 
When the storage disk is started or stopped, sliding contact occurs between the 
disk and slider, and this can cause mechanical damage and wear unless a 
suitable protective coating is present between the surfaces in contact. The 
protective coating must be resistant to wear and corrosion, but also thin 
enough so as not to impede the achievement of high recording density; DLC 
coatings are ideal for such applications. DLC is used as a corrosion- and 
wear-protective coating for both magnetic disks and magnetic heads, and is also 
used for corrosion protection of metal films during the manufacturing of 
magnetic heads. Tapes for video recording or magnetic data storage, using 
ferromagnetic metal as a recording medium, as well as the metallic capstans in 
contact with the tapes, are also protected with DLC coatings to reduce wear and 
friction, thus extending the life of the tapes. 

Diamondlike carbon films appear to be biocompatible, and applications are being 
developed for their use in biological environments. One major reason for the 
failure of metallic implants in the body is their corrosion by body fluids. In 
vitro tests have shown that DLC coatings allow cells to grow without 
inflammatory response [92]. Because of their chemical inertness and 
impermeability to liquids, DLC coatings could protect the implants against 
corrosion and serve as diffusion barriers. DLC films are being considered for 
use as coatings of metallic as well as polymeric (e.g., polyurethane, 
polycarbonate, and polyethylene) biocomponents in order to improve their 
compatibility with body tissues [90, 93]. Their wear resistance is another 
useful property for bioapplications. Diamondlike carbon films, deposited on 
stainless steel and titanium alloys used for components of artificial heart 
valves, have been found to satisfy both mechanical and biological requirements 
and to be capable of improving the performance of these components [94]. The 
same properties may make DLC films useful as protective coatings for joint 
implants. 

As already indicated, DLC films and their modifications are currently being 
considered for use as low-dielectric insulators for the interconnect structures 
of ULSI chips. A better understanding of the means to control their thermal 
stability and other integration problems should improve their potential for 
this application. 

Summary 

Diamondlike carbon films can be prepared from a large variety of precursors 
using the PECVD technique. Film deposition is carried out at substrate 
temperatures between room temperature and 250[degree]C under ion bombardment of 
the substrate and of the growing film. Although the term DLC was adopted to 
describe these films and their modifications, the films consist essentially of 
amorphous networks of sp3- and sp2-bonded carbon, containing hydrogen 
in amounts between 16 and 61%. The films are characterized by infrared 
transparency, a significant optical gap, high electrical resistivity, low 
dielectric constant, high hardness and internal compressive stresses, low 
friction coefficients, and chemical inertness. 

The properties of the films are affected by preparation conditions, and, for a 
given preparation system, depend on the amount of hydrogen incorporated in the 
films. Some of that hydrogen can be incorporated as unbound hydrogen, and the 
relative fraction of bound or unbound hydrogen also affects the properties of 
the films. Although the type of hydrocarbon used as a precursor often appears 
to have only an insignificant effect on film properties, this is not always 
true. The properties of DLC can be tailored by adjusting deposition parameters 
such as substrate bias and temperature, power density, and precursor. 

The wear resistance of DLC films appears to be strongly dependent upon their 
deposition conditions. It can be maintained while modifying the material 
through incorporation of different elements. Both the wear resistance and the 
friction coefficients of DLC films and their modifications are strongly 
affected by the testing environment and are controlled by a transfer layer 
formed between the surfaces in kinetic contact. 

Because of their attractive properties and the ease of preparation of extremely 
smooth films at relatively low substrate temperatures, DLC films have found a 
large variety of practical applications, and it is expected that the range of 
applications will expand further. Although currently limited to applications at 
temperatures below film deposition temperatures, their applicability may 
eventually extend to higher temperatures. 

References 

1. S. Aisenberg and R. Chabot, J. Appl. Phys. 42, 2953 (1971). 

2. P. Koidl, C. Wild, R. Locher, and R. E. Sah, in Diamond and Diamondlike 
Films and Coatings, NATO-ASI Series B: Physics, R. E. Clausing, L. L. Horton, 
J. C. Angus, and P. Koidl, Eds., Plenum Publishing Co., New York, 1991, p. 243. 

3. P. V. Koeppe, V. J. Kapoor, M. J. Mirtich, B. A. Banks, and D. A. Gulino, J. 
Vac. Sci. Technol. A 3, 2327 (1985). 

4. J. Kaplan, F. Jansen, and M. Machonkin, Appl. Phys. Lett. 47, 750 (1985). 

5. A. Grill and B. Meyerson, in Synthetic Diamond: Emerging CVD Science and 
Technology, K. E. Spear and J. P. Dismukes, Eds., John Wiley & Sons, Inc., New 
York, 1994, p. 91. 

6. A. Grill, Surf. Coat. Technol. 94-95, 507 (1997). 

7. A. Y. Liu and M. L. Cohen, Science 245, 841 (1989). 

8. A. Grill, B. S. Meyerson, and V. V. Patel, IBM J. Res. Develop. 34, 849 
(1990). 

9. C. Weissmantel, in Thin Films from Free Atoms and Molecules, K. J. Klabunde, 
Ed., Academic Press, Inc., New York, 1985, p. 153. 

10. K. Enke, Thin Solid Films 80, 227 (1981). 

11. B. Dischler, A. Bubenzer, and P. Koidl, Solid State Commun. 48, 105 (1983). 

12. P. Couderc and Y. Catherine, Thin Solid Films 146, 93 (1987). 

13. A. Grill, B. Meyerson, V. Patel, J. A. Reimer, and M. A. Petrich, J. Appl. 
Phys. 61, 2874 (1987). 

14. B. Meyerson and F. Smith, J. Non-Cryst. Solids 35/36, 435 (1980). 

15. M. Ham and Karen A. Lou, J. Vac. Sci. Technol. A 8, 2143 (1989). 

16. A. Grill and V. Patel, Diamond Films & Technol. 1, 219 (1992). 

17. R. R. Nimmagadda, A. Joshi, and W. L. Hsu, J. Mater. Res. 5, 2445 (1990). 

18. K. Yano, H. Oyama, K. Shimizu, M. Yanokura, Y. Abe, H. Kokai, and Y. 
Sakamoto, in Proceedings of the 9th International Symposium on Plasma 
Chemistry, R. d'Agostino, Ed., IUPAC, Bari, Italy, 1989, p. 1444. 

19. W. Varhue and P. Pastel, J. Mater. Res. 5, 2441 (1990). 

20. K. Endo and T. Tatsumi, Appl. Phys. Lett. 68, 2864 (1996). 

21. S. Takeishi, H. Kudo, R. Shinohara, M. Hashino, S. Fukuyama, J. Yamaguchi, 
and M. Yamada, J. Electrochem. Soc. 144, 1797 (1997). 

22. A. Grill, Cold Plasma in Materials Fabrication; From Fundamentals to 
Applications, IEEE Press, New York, 1994. 

23. O. Matsumoto, H. Toshima, T. Katagiri, and Y. Kanzaki, in Proceedings of 
the 7th International Symposium on Plasma Chemistry, C. J. Timmermans, Ed., 
IUPAC, Eindhoven, Netherlands, 1985, p. 39. 

24. A. Grill and V. Patel, Diamond & Related Mater. 2, 1519 (1993). 

25. A. Grill and V. Patel, Diamond & Related Mater. 4, 62 (1994). 

26. A. Grill and V. Patel, in Applications of Diamond Films and Related 
Materials, Third International Conference, A. Feldman, Y. Tzeng, W. A. 
Yarbrough, M. Yoshikawa, and M. Murakawa, Eds., NIST Special Publication 885, 
Washington, DC, 1995, p. 771. 

27. G. J. Vandentop, M. Kawasaki, K. Kobayashi, and G. A. Somorjai, J. Vac. 
Sci. Technol. A 9, 1157 (1991). 

28. J. C. Angus and F. Jansen, J. Vac. Sci. Technol. A 6, 1778 (1988). 

29. Y. Catherine, in Diamond and Diamondlike Films and Coatings, NATO-ASI 
Series B: Physics, R. E. Clausing, L. L. Horton, J. C. Angus, and P. Koidl, 
Eds., Plenum Publishing Co., New York, 1991, p. 193. 

30. A. Grill and V. Patel, Diamond Films & Technol. 2, 25 (1992). 

31. K. L. Saenger, A. Grill, A. G. Schrott, V. Patel, S. A. Cohen, and D. C. 
Edelstein, in Advanced Metallization and Interconnect Systems for ULSI 
Applications in 1996, R. Haveman, J. Schmitz, H. Komiyama, and K. Tsubouchi, 
Eds., Materials Research Society, Pittsburgh, 1997, p. 499. 

32. A. Grill and V. Patel, Diamond Films & Technol. 6, 13 (1996). 

33. H. Dimigen, H. Hubsch, and R. Memming, Appl. Phys. Lett. 50, 1056 (1987). 

34. H. Dimigen and C. P. Klages, Surf. Coat. Technol. 49, 543 (1991). 

35. A. Bubenzer, B. Dischler, G. Brandt, and P. Koidl, J. Appl. Phys. 54, 4590 
(1983). 

36. J. C. Angus, P. Koidl, and S. Domitz, in Plasma Deposited Thin Films, J. 
Mort and F. Jansen, Eds., CRC Press Inc., Boca Raton, FL, 1986, p. 89. 

37. A. Grill, V. Patel, and B. S. Meyerson, J. Mater. Res. 5, 2531 (1990). 

38. A. Grill, V. Patel, and B. S. Meyerson, in Diamond and Diamondlike Films 
and Coatings, NATO-ASI Series B: Physics, R. E. Clausing, L. L. Horton, J. C. 
Angus, and P. Koidl, Eds., Plenum Publishing Co., New York, 1991, p. 417. 

39. J. W. Zou, K. Schmidt, K. Reichelt, and B. Dischler, J. Appl. Phys. 67, 487 
(1990). 

40. J. A. Reimer, R. W. Vaughan, J. C. Knights, and R. A. Lujan, J. Vac. Sci. 
Technol. 19, 53 (1981). 

41. A. Grill and V. Patel, Appl. Phys. Lett. 60, 2089 (1992). 

42. J. Robertson, Adv. Phys. 35, 317 (1986). 

43. B. Dischler, A. Bubenzer, and P. Koidl, Appl. Phys. Lett. 42, 636 (1983). 

44. M. Shimozuma, G. Tochitani, H. Ohno, and H. Tagashira, in Proceedings of 
the 9th International Symposium on Plasma Chemistry, R. d'Agostino, Ed., IUPAC, 
Bari, Italy, 1989, p. 1462. 

45. M. Tanaka, T. Iwata, F. Fujimoto, K. Komachi, K. Kobayashi, H. Yamashita, 
and M. Haba, Nucl. Instrum. Methods Phys. Res. B 45, 223 (1990). 

46. L. Ponsonnet, C. Donnet, J. M. Martin, A. Grill, and V. Patel, Thin 

Solid Films 319, 97 (1998). 

47. S. Kaplan and A. Dilks, Thin Solid Films 84, 419 (1981). 

48. F. Jansen, M. Machonkin, S. Kaplan, and S. Hark, J. Vac. Sci. Technol. A 3, 
605 (1985). 

49. R. H. Jarman and G. J. Ray, J. Chem. Soc. Chem. Commun. 107, 1153 (1985). 

50. L. A. Farrow, B. J. Wilkens, A. S. Gozdz, and D. L. Hart, Phys. Rev. B 41, 
10,132 (1990). 

51. H.-C. Tsai and D. B. Bogy, J. Vac. Sci. Technol. A 5, 3287 (1987). 

52. B. Dischler, in Amorphous Hydrogenated Carbon Films, Proceedings of the 
European Materials Research Society Symposium, P. Koidl and P. Oelhafen, Eds., 
Les Editions de Physique, Paris, 1987, p. 219. 

53. C. Donnet and A. Grill, Surf. Coat. Technol. 94-95, 456 (1997). 

54. J. Kaplan, F. Jansen, and M. Machonkin, Appl. Phys. Lett. 47, 750 (1985). 

55. M. P. Nadler, T. K. Donovan, and A. K. Green, Thin Solid Films 116, 241 
(1984). 

56. M. J. Mirtich, D. M. Swec, and J. C. Angus, Thin Solid Films 131, 245 
(1985). 

57. F. W. Smith, J. Appl. Phys. 55, 764 (1984). 

58. S. Craig and G. L. Harding, Thin Solid Films 97, 345 (1982). 

59. T. J. Moravec and J. C. Lee, J. Vac. Sci. Technol. 20, 338 (1982). 

60. S. Lin and S. Chen, J. Mater. Res. 2, 645 (1987). 

61. D. R. McKenzie, D. Muller, and B. A. Pailthorpe, Phys. Rev. Lett. 67, 773 
(1991). 

62. A. Grill, V. Patel, K. L. Saenger, C. Jahnes, S. A. Cohen, A. G. Schrott, 
D. C. Edelstein, and J. R. Paraszczak, Mater. Res. Soc. Symp. Proc. 443, 155 
(1997). 

63. B. Meyerson, "Characterization of Hydrogenated Amorphous Carbon Films: 
Electrical and Optical Properties," Ph.D. Thesis, City University of New York, 
NY, 1981. 

64. A. Grill, V. Patel, and S. Cohen, Diamond & Related Mater. 3, 281 (1994). 

65. C. P. Klages and R. Memming, Mater. Sci. Forum 52 & 53, 609 (1989). 

66. A. Grill, V. Patel, S. A. Cohen, D. C. Edelstein, J. R. Paraszczak, and C. 
Jahnes, in Advanced Metallization and Interconnect Systems for ULSI 
Applications in 1996, R. Haveman, J. Schmitz, H. Komiyama, and K. Tsubouchi, 
Eds., Materials Research Society, Pittsburgh, 1997, p. 417. 

67. A. Grill, V. Patel, and C. Jahnes, J. Electrochem. Soc. 145, 1649 (1998). 

68. J. Robertson, Surf. & Coatings Technol. 50, 185 (1992). 

69. D. Nir, Thin Solid Films 146, 27 (1987). 

70. M. F. Thorpe, J. Non-Cryst. Solids 57, 355 (1983). 

71. M. A. Tamor, W. C. Vassell, and K. R. Carduner, Appl. Phys. Lett. 58, 592 
(1991). 

72. J. W. Zou, K. Reichelt, K. Schmidt, and B. Dischler, J. Appl. Phys. 65, 
3914 (1989). 

73. D. F. Franceschini, C. A. Achete, and F. L. Freire, Appl. Phys. Lett. 60, 
3229 (1992). 

74. A. Grill, Wear 168, 143 (1993). 

75. M. N. Gardos, in Synthetic Diamond: Emerging CVD Science and Technology, K. 
E. Spear and J. P. Dismukes, Eds., John Wiley & Sons, Inc., New York, 1994, p. 
419. 

76. R. Memming, H. J. Tolle, and P. E. Wierenga, Thin Solid Films 143, 31 
(1986). 

77. R. H. Savage, J. Appl. Phys. 19, 1 (1948). 

78. D. H. Buckley, Friction, Wear, and Lubrication in Vacuum, NASA SP-277, 
Scientific and Technical Information Office, Washington, DC, 1971. 

79. J. K. Lancaster, Wear 34, 275 (1975). 

80. H. Zaidi, A. Mezin, M. Nivoit, and J. Lepage, Appl. Surf. Sci. 40, 103 
(1989). 

81. A. Erdemir, C. Bindal, J. Pagan, and P. Wilbur, Surf. Coat. Technol. 76/77, 
559 (1995). 

82. C. Donnet, M. Belin, J. C. Auge, J. M. Martin, A. Grill, and V. Patel, 
Surf. Coat. Technol. 68/69, 626 (1994). 

83. C. Donnet, Condensed Matter News 4, 9 (1995). 

84. L. P. Hayward, Surf. Coat. Technol. 49, 554 (1991). 

85. M. Grischke, K. Bewilogua, K. Trojan, and Heinz Dimigen, Surf. Coat. 
Technol. 74/75, 739 (1995). 

86. C. Donnet, J. Fontaine, A. Grill, V. Patel, C. Jahnes, and M. Belin, Surf. 
Coat. Technol. 94-95, 531 (1997). 

87. A. K. Gangopadhyay, P. A. Willermet, M. A. Tamor, and W. C. Vassell, 
Tribol. Internat. 30, 9 (1997). 

88. P. Koidl, A. Bubenzer, and B. Dischler, Symposium on Optical Surface 
Technology, Proc. SPIE 381, 186 (1983). 

89. A. Lettington, in Amorphous Hydrogenated Carbon Films, Proceedings of the 
European Materials Research Society Symposium, P. Koidl and P. Oelhafen, Eds., 
Les Editions de Physique, Paris, 1987, p. 359. 

90. A. H. Lettington, in Applications of Diamond Films and Related Materials, 
Materials Science Monographs, Y. Tzeng, M. Yoshikawa, M. Murakawa, and A. 
Feldman, Eds., Elsevier, New York, 1991, p. 703. 

91. F. M. Kimock and B. J. Knapp, Surf. Coat. Technol. 56, 273 (1993). 

92. A. C. Evans, J. Franks, and P. J. Revell, Surf. Coat. Technol. 47, 662 
(1991). 

93. C. J. McHargue, in Applications of Diamond Films and Related Materials, 
Materials Science Monographs, Y. Tzeng, M. Yoshikawa, M. Murakawa, and A. 
Feldman, Eds., Elsevier, New York, 1991, p. 113. 

Received December 2, 1997; accepted for publication March 16, 1998 

Alfred Grill

IBM Research Division, Thomas J. Watson Research Center, P.O. Box 218, Yorktown 
Heights, New York 10598 (grilla@us.ibm.com). Dr. Grill is a Research Staff 
Member and manages the Electronic Materials and Structures group in the 
Telecommunication Technologies Department at the Thomas J. Watson Research 
Center. He received B.S., M.Sc., and Ph.D. degrees in physics from the Hebrew 
University in Israel in 1964, 1966, and 1973, respectively. In 1975 he joined 
the Materials Engineering Department at the Ben-Gurion University in Israel, 
where, in 1986, he became a Full Professor. In 1989, he joined IBM at the 
Thomas J. Watson Research Center, where he is studying the PECVD growth of 
diamondlike carbon and related materials as well as low- and 
high-dielectric-constant materials. In 1992, he received an IBM Outstanding 
Innovation Award for his work on carbon overcoats for magnetoresistive heads. 
Dr. Grill has published extensively on the plasma-assisted deposition of films 
and surface treatment of materials; he is an author or coauthor of 25 patents. 
He is a member of the Materials Research Society and the Electrochemical 
Society, and a member of the editorial boards of the journals Diamond Films and 

Technology and Diamond and Related Materials.