1 ICMCTF San Diego April 2002, Paper B1-2-2, Surf. Coat. Technol. (in press) Superhard nc-TiN/a-BN and nc-TiN/a-TiB x /a-BN Coatings Prepared by Plasma CVD and PVD: A comparative Study of their Properties P. Karvankova a , M. G. J. Veprek-Heijman a , O. Zindulka b , A. Bergmaier c and S. Veprek a * a Institute for Chemistry of Inorganic Materials, Technical University Munich, Lichtenbergstr. 4, D-85747 Garching b. Munich, Germany b SHM Ltd., CZ-788 03 Novy Malin 266, Czech Republic c Physics Department E12, Technical University Munich, James-Franck-Str. D-85747 Garching Abstract We present a comparative study of the preparation and properties of superhard "Ti-B-N" coat- ings deposited by plasma CVD and by Vacuum Arc Evaporation (PVD) of Titanium combined with Plasma CVD of TiB 2 and BN. Using high frequency plasma CVD at a total pressure of sev- eral mbar with TiCl 4 , BCl 3 , N 2 and H 2 as reactants, superhard (H V ≈ 40 50 GPa) nanocompo- site coatings were successfully and reproducibly deposited and characterized in terms of me- chanical properties (indentation & SEM), phase composition (XPS and ERD) and nanostructure (XRD, SEM). Using reactive sputtering, several authors reported about the preparation of super- hard TiN/TiB 2 coatings with only a small fraction of BN. Efforts to increase the fraction of the BN phase resulted in soft films. In contrast, plasma CVD yields superhard nc-TiN/a-BN and nc- TiN/a-TiB 2 /a-BN coatings in a wide range of the fractions of BN and TiB 2 phases. This is attrib- uted to the high chemical activity of nitrogen under the conditions of plasma CVD. Industrial scale vacuum arc evaporation PVD in combination with plasma CVD is used for preparation of "Ti-B-N" coatings. In the system nc-TiN 1-x /a-TiB 2 with a minor fraction of the a-BN phase su- perhard coatings with a very low fraction of microparticles and resultant low surface roughness of R m ≈ 0.15 m were successfully prepared and those are intended for testing under a variety of cutting conditions. * Corresponding author: Tel.: ..49-89-2891 36 24; fax: ..49-89-2891 36 26 E-mail address: [email protected]1. Introduction Superhard (H ≥ 40 GPa) coatings are receiving increasing attention because of their po- tential applications for wear protection (e. g. on tools for cutting, forming and stamping). Only diamond and cubic boron nitride are intrinsically superhard. Superhardness can be achieved in a variety of ordinary hard coatings either by energetic ion bombardment during their deposition
18
Embed
Superhard nc-TiN/a-BN and nc-TiN/a-TiBx/a-BN coatings prepared by plasma CVD and PVD: a comparative study of their properties
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
ICMCTF San Diego April 2002, Paper B1-2-2, Surf. Coat. Technol. (in press)
Superhard nc-TiN/a-BN and nc-TiN/a-TiBx/a-BN Coatings
Prepared by Plasma CVD and PVD: A comparative Study of their Properties
P. Karvankovaa, M. G. J. Veprek-Heijmana, O. Zindulkab, A. Bergmaierc and S. Vepreka *aInstitute for Chemistry of Inorganic Materials, Technical University Munich,
(which results in a variety of effect including a high compressive stress) or by the formation of
an appropriate nanostructure, such as superlattices and nanocomposites (see [1 - 4] and refer-
ences therein). The hardness enhancement due to the energetic ion bombardment is easy to
achieve in a variety of coatings including the so called "nanocomposites" M(1)nN/M(2). Here
M(1) is a hard and stable transition metal nitride, such as ZrN, Cr2N, TiN, ... and M(2) is another
metal which does not form any thermodynamically stable nitride, such as Ni, Cu,... (for a review
on these "nanocomposites" see [5] and also [1]). However, the superhardness is lost when the
high stress and other beneficial effects induced by the energetic ion bombardment relax upon an-
nealing to 400 � 600 °C [1] [4] or upon a long term storage in air.
The superhard nanocomposites prepared according to the generic design principle [6] [7]
show a very high thermal stability up to 1000 � 1100 °C [8] [9]. This design principle is based on
the formation of stable nanostructure due to a strong, spinodal decomposition in a binary or ter-
nary system [10]. The most studied systems so far were nc-TiN/a-Si3N4, nc-W2N/a-Si3N4, nc-
VN/a-Si3N4, nc-(TiAl)N/a-Si3N4, and nc-TiN/a-Si3N4/a- & nc-TiSix prepared by plasma induced
CVD at a relatively large partial pressure of nitrogen of several mbar which assured the desired
phase segregation (see [7 - 10]).
The Ti-B-N system represents another class of superhard coatings which were prepared
by a variety of PVD methods, such as vacuum arc evaporation (e.g. [11] [12]) and sputtering (e.
g. [13 - 16] and further references in the review [7]). Because of limited space available here we
can mention only some highlights of the many papers on the Ti-B-N system which are relevant
to our present work. In the PVD techniques used, the partial pressure of nitrogen is orders of
magnitude lower than in our plasma CVD and, therefore, the segregation into TiN and BN is not
completed. As result, a variety of Ti, B and N containing phases are formed depending on the
exact conditions (N2 pressure, substrate temperature, ion bombardment,...). Using conventional
sputtering at a low pressure results either in the formation of a homogeneous phase TiB2Nx [17 -
19], or a TiNxBy phase consisting of TiN lattice in which the nitrogen sites are occupied by 33 %
of boron, 49 % of nitrogen and 18 % of vacancies [20], or amorphous Ti-B-N coatings [21] [22].
Other authors have found phase segregation with the formation of binary phases TiN and TiB2
[13 - 15] [23] [24] corresponding to the "region 4" of the equilibrium phase diagram [15] 1. At a
lower boron content, nanocrystalline TiN and quasi-amorphous TiB2 phase are formed whereas a
higher boron content leads to a structure consisting of quasi-amorphous TiN and nanocrystalline
TiB2 [23] [24]. A large number of "Ti-B-N" coatings with different composition and structure
1 The authors refer to the original work of Novotny et al. from 1961 [25] not considering the more recent literaturein which a variety of other phases was reported [26].
3
prepared by a variety of methods have been reported, such as a mixture of phases TiN, TiB2 and
BN or nc-TiN phase imbedded in a BN matrix [27 - 29], but in most cases, the coatings were
relatively soft of ≤ 11 GPa [30] [31]. Héau and Terrat reported a hardness of 80 GPa in Ti-B-N
coatings deposited by magnetron co-sputtering from Ti and TiB2 targets and having the compo-
sition 13 at.% B, 37 at.% Ti and 50 at.% N, but the hardness was unstable and decreased to about
46 GPa after several days [32]. The BN matrix can be hexagonal, amorphous or cubic [30 - 33],
or even a mixture of fcc TiN, orthorhombic TiB, c-BN and hexagonal Ti-B-N [34 - 37]. Hard-
ness of 40 - 70 GPa was reported in many papers [13 - 20] [24] [27] [33 - 38] [40 - 42]. In gen-
eral, the hardness as a function of composition did not follow the rule of mixture but displayed a
maximum at a given composition where the microstructure of the films was very uniform (no
columnar growth) and the crystallite size was a few nm. With increasing nitrogen content, when
the h-BN phase was formed, the hardness of the film strongly decreased. This decrease was at-
tributed to the formation of soft h-BN [15] [19] [22] [28] [38] [40] [43].
The purpose of our present paper is to further prove the generality of our concept for the
design of superhard nanocomposites also for this system. The thermal stability of the fully segre-
gated, binary TiN + BN system depends on the pressure of nitrogen. The decomposition TiN +
2BN → TiB2 + (3/2)N2 commences about 1350 °C at nitrogen pressure of 1 mbar and the de-
composition temperature increases to 1600 °C at P(N2) = 500 mbar. At a lower nitrogen pressure
the decomposition temperature decreases and a complex multiphase diagrams applies including a
variety of TixBy, TinNm and further mixed phases TixByNz [26]. This explains the complexity of
phases found in coatings deposited by the PVD techniques (see above) under a low and variable
nitrogen pressure or even without nitrogen. Under conditions of plasma CVD, when the nitrogen
pressure of several mbar is sufficiently high, the binary system of stoichiometric TiN and BN
should be stable. Therefore, only under these conditions the spontaneous formation of the nc-
TiN/a-BN nanocomposite should occur. By analogy with our earlier studied systems nc-TiN/a-
Si3N4, nc-W2N/a-Si3N4, nc-VN/a-Si3N4, and nc-TiN/a-Si3N4/a- & nc-TiSi2 we predicted that the
maximum hardness should be obtained at the percolation threshold when there is about one
monolayer thin continuous tissue of a-Si3N4 or a-BN between the nanocrystals of a stable transi-
tion metal nitride [6] [44]. In this paper we provide an evidence for this prediction and compare
the properties of the coatings deposited by plasma CVD with those deposited by vacuum arc
PVD.
2. Experimental
4
The nc-TiN/a-BN coatings were deposited by means of plasma CVD in a high frequency
(HF) discharge operating at 13.56 MHz and power 100 Watts. The apparatus was similar to that
described in our earlier papers [6] [44] with the following modifications: The reactor, made of
silica glass, was inserted into an electrical oven and kept at 565 °C. The substrate holder was
used as one electrode, the other one was a grounded nickel sheet around the silica tube. At a con-
stant total pressure of 3 mbar the gas flows were: N2 � 5 sccm, H2 � 50 sccm, TiCl4 � 1.7 sccm
and BCl3 was varied between 0 and 1.7 sccm. Under these conditions the deposition rate was
around 0.9 nm/s and the thickness of the coatings between 6 and 10 µm. Because the energy of
ions bombarding the growing film is under these conditions small, the biaxial compressive stress
was below about 1 GPa and, therefore, there was no hardness enhancement due to the stress in
the coatings.
The "Ti-B-N" (essentially nc-TiN/a-TiB2) coatings were deposited by vacuum arc evapo-
ration of Ti as described in [33] and introducing borazine (B3N3H6) but no nitrogen. For the two
series of samples studied here the deposition temperature was 600 and 700 °C, respectively and
the bias of the substrates � 100 V.
The crystallite size and crystalline phases were determined by means of X-Ray diffrac-
tion (XRD) using both Bragg-Brentano and glancing incidence geometry. The crystallite size
was determined by means of Warren-Averbach analysis [45] because for the largest crystallite
sizes the contribution of random strain to the Bragg peaks broadening is significant and, there-
fore, the simple Scherrer formula cannot be used. The morphology was investigated by means of
scanning electron microscopy (SEM) equipped with energy dispersive analysis of X-rays (EDX).
Because of the low sensitivity of EDX to low-Z elements and because the cut-off due to the be-
ryllium window at the energy corresponding to the B line, the elemental analysis was done by
means of elastic recoil detection (ERD) as described in [46] [55]. The annealing experiments
were done in 1 atm. of pure nitrogen (99.999) keeping the sample at the given temperature for
0.5 hour.
The hardness measurements were done by means of the automated load-depth sensing
technique as a function of the applied load and only the load independent hardness (typically
between a load of 30 to 150 mN and indentation depth ≥ 0.3 µm) was taken as reliable. The in-
dentometer Fischerscope 100 equipped with a microscope and possibility to program a series of
indentations at different chosen lateral positions on the coatings was used. The hardness values
obtained from the Fischerscope was verified by measuring the size of the remaining indentation
in SEM and calculating the hardness from the equation H = 0.927⋅L/AP where L is the applied
5
load and AP is the projected area of the indentation [47]. Further details of our hardness meas-
urements are given in [7] [48].
Leybold LH 10 surface analytic system was used for the measurement of the phase com-
position by means of X-ray photoelectron spectroscopy (XPS). The spectra were excited with the
AlKα source and recorded in ∆E = const. mode. Repetitive scans of selected spectral regions and
signal averaging were used in order to obtain a sufficient signal-to-noise ratio.
3. Results and discussion
3.1. Coatings deposited by plasma CVD
The only crystalline phase which was detected by means of XRD in all coatings dis-
cussed here was the fcc TiN showing a preferential (200) orientation. The lattice parameter was
by about ≤ 0.3 % slightly enhanced but no systematic variation with the boron content could be
found. Although the coatings had gold-like color the ERD showed a slight substoichiometry of
about 4 to 8 at. %.
In Fig. 1 we report results of two series of samples. Initially, the substrate was fixed to
the sample holder at one point only (open symbols) later by two (full symbols) which clearly
helped to improve the lateral homogeneity of the deposition and to lower the chlorine content at
high total flow rates of the chlorides. In the first series the chlorine content of the coatings in-
creased above the flow rate ratio of 0.6 to 2.5 � 4.5 at. % which deteriorated the properties. In the
second series the chlorine content was less than 1.5 at. % even at the highest chlorine flow rates
and below about 0.6 at. % for flow ratio ≤ 0.6. Although all recent coatings are deposited with
the two point fixing we include here also the older ones in order to underline the general depend-
ence of the hardness on the coverage of the surface of TiN nanocrystals by BN (we removed the
error bars for the sake of lucidity). From Fig. 1a we can see that with increasing BCl3/TiCl4 flow
ratio the hardness increases up to 50 GPa and, at even higher flow rates, it slightly decreases still
being in the range of about 40 GPa. The vertical line in Fig.1a separates the region where the
fraction of TiB2 phase measured by XPS is small and can be neglected as compared to BN (left
hand side) and that where the content of this phase increases and becomes comparable (right
hand side). The crystallite size (Fig. 1b) of 10 � 40 nm for flow ratio ≤ 0.6 is much larger than in
our other coatings (nc-TiN/a-Si3N4, nc-W2N/a-Si3N4, nc-VN/a-Si3N4, and nc-TiN/a-Si3N4/a- &
nc-TiSi2) and it decreases to 3 � 7 nm at higher BCl3 flow.
One can see from Fig. 1 that although the crystallite size and the boron content signifi-
cantly change between the BCl3/TiCl4 flow ratio of 0.2 and 1.1 the hardness does not show any
significant change in that range. The only clear change is the increase of the hardness from the
6
usual value of 23 � 25 GPa for pure TiN to 40 � 50 GPa with the addition of boron for the
BCl3/TiCl4 gas flow ratio between 0 and 0.15. This is further emphasized by Fig. 2a which
shows the hardness vs. crystallite size for coatings which were so far completely analyzed (i. e.
also by ERD and XPS, which is very much time consuming). On the other hand, if the specific
surface area is calculated from the measured crystallite size and the coverage of the surfaces of
the TiN nanocrystals is calculated from the boron content determined by ERD and XPS (see be-
low) it is seen in Fig. 2b that maximum hardness is achieved at about one monolayer coverage.
The specific surface area of the TiN nanocrystals was calculated assuming a regular
shape. This is based on our earlier investigations of the nc-TiN/a-Si3N4 [49] and nc-TiN/a-
Si3N4/a-TiSi2 [50] by means of high resolution transmission electron microscopy which reveals
fairly regular and uniform shape of the TiN nanocrystals. However, because the TiN nanocrys-
tals are randomly oriented, it is impossible to obtain image of the grain boundaries and of the ex-
act shape of the nanocrystals even when high resolution, at which one can image the TiN lattice
plains, is used (see Fig. 3 in [49]). In spite of these problems, the results shown in Fig. 2b agree
surprisingly well with those obtained earlier for the binary nc-TiN/a-Si3N4 (see Fig. 5 in [6] and
Fig. 4 in [10]), nc-W2N/a-Si3N4 (see Fig. 3 in [44]) and ternary nc-TiN/a-Si3N4/a-TiSi2 (see Fig.
3 in [10]). In spite of the limited accuracy of such calculations due to unavoidable errors of the
analyses and lack of knowledge of the exact shape of the TiN nanocrystals, this agreement is
very encouraging because it lends a strong support to our generic concept [6].
3.2. Coatings deposited by vacuum arc PVD
Also in these coatings the only crystalline phase detected by XRD was the fcc of TiN
with a preferential (220) orientation and crystallite size of 6 � 7 nm. The boron content was 22
and 24 at. % for samples from the two series studied here. In spite of such a relatively high boron
content no Bragg reflection corresponding to a TiBx phases was found. Because no nitrogen was
introduced during the deposition (i. e. all nitrogen originated from the addition of borazine) the
samples were substoichiometric nc-TiN1-x/a-TiB2 with a minor contribution of the BN phase (see
below) with about 23 - 26 at. % of N and they had metallic gray color which changed to gold
upon post-annealing in N2. The load independent hardness varied between about 45 and 55 GPa
(for an indentation depth of ≥ 0.3 µm).
Because the coatings were deposited at negative bias they had a biaxial compressive
stress of ≥ 5 GPa which relaxed upon annealing to a temperature of ≥ 800 °C. This was accom-
panied by a decrease of the measured hardness to about 30 GPa. The exact value of this decrease
is difficult to quantify because even in the pure nitrogen the residual oxygen impurity due to the
7
desorption from walls caused appreciable oxidation of the TiB2 phase, evaporation of BOx and
resultant loss of boron from the coating (see below). The nc-TiN/a-BN superhard nanocompo-
sites are more stable against oxidation than the nc-TiN/TiB2 as already found also for the nc-
TiN/a-Si3N4 [6].
3.3. XPS investigation
The XPS investigation of Ti containing coatings is subjected to an unavoidable problem
of the contamination with oxygen during the transport of the coatings from the deposition gear to
the XPS which cannot be fully removed by sputtering due to a final surface roughness. Moreo-
ver, the Ar+ sputter-cleaning results in a preferential removal of nitrogen. For example, when a
pure stoichiometric TiN, whose stoichiometry was verified by means of Rutherford backscatter-
ing spectroscopy (RBS), was sputter-cleaned by 3 keV Ar+ and, afterwards the stoichiometry
determined by means of XPS a composition of about TiN0.85 was found in our system [51]. Ac-
counting for these changes the overall composition determined by ERD and XPS agrees rea-
sonably well. However, because of these artefact we used ERD to calibrate our XPS spectra.
From Fig. 3 one can clearly see that in the nanocomposites prepared by means of plasma
CVD at a relatively high partial pressure of nitrogen of about 0.3 mbar as compared with the
PVD techniques BN is the dominant boron containing phase when the total boron content does
not exceed about 8 at.%. The TiB2 fraction can be neglected in these coatings and we can calcu-
late the coverage of the TiN nanocrystals by the amorphous BN. In the case of coatings with
higher boron content we have to take the fraction of a-BN and a-TiB2 into the calculation of the
coverage.
The assignment of the peaks to BN and TiB2 is in agreement with the data of other
authors [14] [52] [53]. The peak at a binding energy of about 186 was assigned to TiB or to bo-
ron dissolved in substoichiometric TiN1-x [54]. However, the following argument supports the
assignment of this peak to boron atoms at the TiN surface being bonded simultaneously to Ti and
N. This is supported by results shown in Fig. 4 which shows the fraction of the three phases vs.
the crystallite size. It is well known that less than about 10 � 20 nm small nanocrystals are free of
any defects because of a relatively high, destabilizing contribution of such defect to the total free
energy of such nanocrystal combined with a short diffusion length to the grain boundaries. In
Fig. 4 we cannot see any dependence of the fraction of the phases on the crystallite size although
the latter changes between about 7 and 35 nm. Therefore we attribute, at least preliminary, the
peak at Eb ≈ 186 eV to boron atoms located at the surfaces of the TiN nanocrystals.
8
The dominant boron phase in the PVD "Ti-B-N" coatings is TiB2 with a much smaller
fraction of BN and even much smaller TiB component (see Fig. 3b). This is in agreement with
the majority of results reported by other authors for Ti-B-N coatings prepared by PVD tech-
niques.
3.4. Thermal stability
The thermal stability of nc-TiN/a-BN nanocomposites deposited by plasma CVD on
stainless steel substrates is comparable to that of the nc-TiN/a-Si3N4 and nc-TiN/a-Si3N4/a-TiSi2
reported earlier [8] [9]. One example is shown in Fig. 5.
Because the crystallite size remains fairly constant upon annealing up to 900oC the de-
crease of the hardness upon annealing to ≥ 900 °C is due most probably to the diffusion of chro-
mium from the stainless steel substrate as observed by EDX. Relaxation of a compressive stress
is unlikely because the lattice parameter measured by XRD remained almost unchanged up to the
highest annealing temperature (not shown here for lack of space). In our earlier paper we have
found that the thermal stability of the previous superhard nanocomposites increased to ≥ 1100 °C
with crystallite size decreasing to about 3 nm (see Fig. 6 in [8]). The stability of the nc-TiN/a-BN
nanocomposites reported here seems to be similar for the range of crystallite sizes obtained so far
but more work is needed in order to prepare such nanocomposites with crystallite size below 5
nm.
The "Ti-B-N" coatings prepared by vacuum arc PVD and CVD from borazine show a
smaller oxidation resistance and also somewhat smaller thermal stability upon annealing in pure
nitrogen. As already mentioned, there is an indication that the somewhat larger decrease of the
hardness upon annealing to 800 °C may be due partially to the relaxation of compressive stress.
However, because the XPS showed clearly a lost of boron after the annealing of these coatings,
evidently due to the above mentioned oxidation by residual O-impurities in the oven and subli-
mation of BOx (not shown here for lack of space), we attribute the observed softening to the lost
of the TiB2 tissue. Thus, it is quite clear that the high hardness of 45 � 55 GPa in our Ti-B-N
coatings is due predominantly to the formation of nanostructure and not to enhancement caused
by energetic ion bombardment during the deposition as in the case of the ZrN/Ni, Cr2N/Ni [4]
and other so called "nanocomposites" [5] prepared by magnetron sputtering at a low pressure.
Moreover, Hammer et al. [13] have shown that in the Ti-B-N coatings deposited at room tem-
perature and having a relatively low hardness of about 20 GPa the superhardness of 40 GPa can
be achieved upon thermal post-treatment which results in a spontaneous formation of a nanos-
tructure. In many other papers there is also a contribution of the high energetic ion bombardment
9
during the deposition and the formation of compressive stress to the higher values of the reported
hardness and it is difficult to distinguish whether the nanostructure or the ion bombardment play
the decisive role (see discussion in [1]).
Conclusions
Superhard nanocomposites consisting of dominant phases nc-TiN/a-BN were prepared by
plasma CVD. Their hardness reached 45 � 50 GPa at a composition when the surface of the TiN
nanocrystals was covered by approximately one monolayer of a-BN. This lends further support
to our generic concept for the design of novel superhard nanocomposites in a variety of different
systems. "Ti-B-N" coatings deposited by PVD in combination with CVD of boron phases from
borazine contain as main boron phase TiB2. In agreement with many earlier papers they also
reach hardness of about 50 GPa but show, at least so far known, a lower oxidation stability. The
latter could be the reason for somewhat worse performance of these coatings in cutting tests un-
der conditions of hard, fast and dry machining. This is in agreement with the conclusions of [13].
Several open questions require further study. In particular it is necessary to elucidate the possi-
bilities of preparing perfectly stoichiometric superhard nc-TiN/a-BN with crystallite size less
than 5 nm and to study their oxidation resistance and thermal stability.
Acknowledgment
This work was supported by the European Commission under the project G1RD-CT99-
0222 "NACODRY" and by NATO SfP Project 972379 Protection Coatings.
10
References
[1] S. Veprek, P. Karvankova and H. Männling, MRS Symp. Proc. 697 (2002) in press.
[2] S. Veprek and A. S. Argon, Surf. Coat. Technol. 146-147 (2001) 175.
[3] S. Veprek and A. S. Argon, J. Vac. Sci. Technol. B 20 March/April (2002).
[4] P. Karvankova, H. Männling, S. Veprek, Surf. Coatings Technol. 146-147 (2001) 280.
[5] J. Musil, Surf. Coat. Technol. 125 (2000) 322.
[6] S. Veprek, S. Reiprich, Thin Solid Films 268 (1995) 64.
[7] S. Veprek, J. Vac. Sci. Technol. A 17 (1999) 2401.
[8] A. Niederhofer, P. Nesladek, H.-D. Männling, K. Moto, S. Veprek and M. Jilek, Surf. Coat-
ings Technol. 120-121 (1999) 173.
[9] H.-D. Männling, D. S. Patil, K. Moto, M. Jilek and S. Veprek, Surf. Coating Technol. 146-
147 (2001) 263.
[10] A. Niederhofer, T. Bolom, P. Nesladek, K.Moto, C. Eggs, D. S. Patil and S. Veprek, Surf.
Coat. Technol. 146-147 (2001) 183.
[11] R. A. Andrievski, I. A. Anisimova and V. P. Anisimov, Thin Solid Films 205 (1991) 171.
[12] R. A. Andrievski, J. Mater. Sci. 32 (1997) 4463.
[13] P. Hammer, A. Steiner, R. Villa, M. Baker, P. N. Gibson, J. Haupt and W. Gissler, Surf.
Coat. Technol. 68/69 (1994) 194.
[14] T. P. Mollart, M. Baker, J. Haupt, A. Steiner, P. Hammer and W. Gissler, Surf. Coat.
Technol 74-75 (1995) 491.
[15] W. Gissler, M. A. Baker, J. Haupt, P. N. Gibson,R. Gilmore and T. P. Mollart, Diamond
Film and Technol. Vol. 7, No.3 (1997) 165.
[16] P. Losbichler, C. Mitterer, P. N. Gibson, W. Gissler, F. Hofer and P. Warbichler, Surf.
Coat. Technol. 94-95 (1997) 297.
[17] M. S. Wong and Y. C. Lee, Surf. Coat. Technol. 120-121 (1999) 194.
[18] H. Deng, J. Chen, R. B. Inturi and J. A. Barnard, Surf. Coat. Technol. 76-77 (1995) 609.
[19] T. P. Mollart, J. Haupt, R. Gilmore and W. Gissler, Surf. Coat. Technol. 86-87 (1996) 231.
[20] C. Héau, R. Y. Fillit, F. Vaux and F. Pascaretti, Surf. Coat. Technol. 120-121 (1999) 200.
[21] R. Wiedemann, V. Weihnacht and H. Oettel, Surf. Coat. Technol. 116-119 (1999) 302.
[22] R. D. Ott, C. Ruby, F. Huang, M. L. Weaver and J. A. Barnard, Thin Solid Films 377-378
(2000) 602.
[23] P. H. Mayrhofer and C. Mitterer, Surf. Coat. Technol. 133-134 (2000) 131.
[24] C. Mitterer, P. H. Mayrhofer, M. Beschliesser, P. Losbichler, P. Warbichler, F. Hofer, P. N.
Gibson, W. Gissler, H. Hruby, J. Musil and J. Vlcek, Surf. Coat. Technol. 120-121 (1999) 405.
11
[25] H. Novotny, F. Benesovsky, C. Brukl and O. Schob, Mh. Chem. 92 (1961) 403.
[26] P. Rogl and J. C. Suster, eds., Phase Diagrams of Ternary Boron Nitride and Silicon Nitride
Systems, ASM International, Materials Park, Ohio 1992.
[27] L. Chaleix and J. Machet, Surf. Coat. Technol. 91 (1997) 74.
[28] J. L. He, S. Miyake, Y. Setsuhara, I. Shimizu, M. Suzuki, K. Numata and H. Saito, Wear
249 (2001) 498.
[29] M. A. Baker, T. P. Mollart, P. N. Gibson and W. Gissler, J. Vac. Sci. Technol. A 15(2)
(1997) 284.
[30] J. F. Pierson, F. Bertran, J. P. Bauer and J. Jolly, Surf. Coat. Technol. 142-144 (2001) 906.
[31] L. Maya, C. E. Vallet and J. N. Fiedor, J. Vac. Sci. Technol. A 15(4) (1997) 2007.
[32] C. Héau and J. P. Terrat, Surf. Coat. Technol. 108-109 (1998) 332.
[33] P. Holubář, M. Jílek and M. �íma, Surf. Coat. Technol. 120-121 (1999) 184.
[34] Q. Q. Yang, L. S. Wen, X. Z. Chen, Y. Z. Zheng and Y. Z. Zhuang, Vacuum Vol. 46
(1995) 181.
[35]Q. Q. Yang, L. H. Zhao and L. S. Wen, Surf. Coat. Technol. 78 (1996) 27.
[36] Y. Qiaoqin, Z. Lihua, X. Hanning, Z. Nanfang and H. Qizhong, Trans. Nonferrous Met.
Soc. China 7 (1997) 97.
[37] Q. Q. Yang, L. H. Zhao, H. Q. Du and L. S. Wen, Surf. Coat. Technol. 81 (1996) 287.
[38] C. Mitterer, M. Rauter and P. Rödhammer, Surf. Coat. Technol. 41 (1990) 351.
[39] C. Mitterer, P. Losbichler, F. Hofer, P. Warbichler, P. N. Gibson and W. Gissler, Vacuum
50 (1998) 313.
[40] C. Rebholz, H. Ziegele, A. Leyland and A. Matthews, J. Vac. Sci. Technol. A 16(5) (1998)
2851.
[41] C. Héau, N. Guillon, R. Y. Fillit and J. Machet, Surf. Coat. Technol. 97 (1997) 60.
[42] C. Pfohl and K.-T. Rie, Surf. Coat. Technol. 116-119 (1999) 911.
[43] C. Mitterer, P. Rödhammer, H. Störi and F. Jeglitsch, J. Vac. Sci. Technol. A 7(4) (1989)
2646.
[44] S. Veprek, M. Haussmann and S. Reiprich, J. Vac. Sci. Technol. A 14(1996)46.
[45] H. P. Klug and L. E. Alexander, X-Ray Diffraction Procedures, 2nd ed., J. Wiley, New York
1974.
[46] S. Veprek, A. Niederhofer, K. Moto, T. Bolom, H.-D. Männling, P. Nesladek, G. Dollinger
and A. Bergmaier, Surf. Coating Technol. 133 (2000) 152.
[47] G. Tabor, The Hardness of Metals (Clarendon Press, Oxford, 1951).
12
[48] S. Veprek, S. Mukherjee, P. Karvankova, H.-D. Männling, J. L. He, K. Moto and A. S. Ar-