Hybrid plasma CVD of diamond-like carbon (DLC) at low temperatures

Article abstract of DOI:10.1023/A:1018529512387

Diamond-like carbon coatings have been deposited onto various substrates at 100-150°C using a hybrid plasma assisted chemical vapour deposition technique activated by radio frequency at 13.56 MHz. The coatings have been characterized using a number of techniques including scanning electron microscopy, Raman spectroscopy, thermoanalysis and pin-on-disc wear testing. Results show the films to be diamond like, with the addition of nitrogen (prior to deposition) promoting the formation of crystallites. In addition the condition and type of substrate have been found to have a strong influence on the structural characteristics of the deposited diamond-like films. SEM analysis of diamond-like carbon coatings deposited onto metal matrix composite materials such as Si-Al MMC is reported. The hybrid CVD technology enabled films to be deposited evenly onto the porous MMC structure. Commercially manufactured drills, coated with DLC and titanium nitride (TiN), have been compared to examine their cutting wear resistance characteristics.

Full text of DOI:10.1023/A:1018529512387

JOURNAL OF MATERIALS SCIENCE 32 (1997) 1931 — 1936
Hybrid plasma CVD of diamond-like carbon (DLC)
at low temperatures
D. TITHER, W . AHM ED
Departm ent of Chem istry, The Manchester Metropolitan University, J ohn Dalton Building,
Chester Street, Manchester, M15 5GD, UK
E. AHM ED
Departm ent of Physics, B.Z. University, Multan, Pakistan
Diam ond-like carbon coatings have been deposited onto various substrates at 100—150 °C
using a hybrid plasm a assisted chem ical vapour deposition technique activated by radio
frequency at 13.56 MHz. The coatings have been characterized using a num ber of techniques
including scanning electron m icroscopy, Ram an spectroscopy, therm oanalysis and
pin-on-disc wear testing. Results show the film s to be diam ond like, with the addition of
nitrogen (prior to deposition) prom oting the form ation of crystallites. In addition the
condition and type of substrate have been found to have a strong influence on the structural
characteristics of the deposited diam ond-like film s. SEM analysis of diam ond-like carbon
coatings deposited onto m etal m atrix com posite m aterials such as Si—Al MMC is reported.
The hybrid CVD technology enabled film s to be deposited evenly onto the porous MMC
structure. Com m ercially m anufactured drills, coated with DLC and titanium nitride (TiN),
have been com pared to exam ine their cutting wear resistance characteristics.
1. Introduction
electro-optic and semiconductor devices. The optical
properties of diamond are important in lenses, laser
windows and optical devices, etc. [8].
Plasma assisted chemical vapour deposition
(PACVD) has become an important technology for
the deposition of a variety of coatings for a number of
applications, in the fields of electronics, tooling, surgi-
cal implants and optics [1, 2]. The major advantages
of PACVD over other thin film deposition techniques
include low deposition temperature, high deposition
rate, good control over stoichiometry, cleanliness and
low particulate levels. Although scientific interest in
carbon coatings, particularly diamond, has been con-
siderable over the past twenty years [2—7], it is only
recently that the commercial promise of thin films of
diamond and related materials such as diamond-like
carbon (DLC) has become an industrial reality [8].
DLC is an intermediate amorphous carbon coating
having some diamond-like characteristics [3] and has
been used in applications including cutting tools, dies,
surgical implants, combustion engine components [9],
fibre optic cables, catalysis, heat sinks and micro-elec-
tronics [10]. Diamond is a material with a unique
combination of physical, chemical, electrical and op-
tical properties which make it potentially attractive
for a large number of important industrial applica-
tions [11]. For example, its extreme hardness and
wear resistance makes it ideal for coating tools, drills,
moving parts in mechanical machinery such as car
engines, etc. In electronics its transparency, insulating
properties, radiation hardness and excellent thermal
conductivity make it a good candidate for future ap-
plications [12] in circuit packaging, high power,
Thin film carbon coatings can be divided into three
main categories; polycrystalline diamond, diamond-
like carbon (DLC) and graphite [3]. The challenge in
technological development for industrial application
onto large substrates at low temperature are signifi-
cant for polycrystalline diamond and become con-
siderably less demanding for graphite. A number of
techniques [13, 14] have been used to deposit carbon
coatings including hot filament chemical vapour de-
position (CVD), ion plating, plasma assisted CVD, ion
beam deposition, mass selected ion beam deposition,
combustion synthesis, etc. However, it has been shown
[5] that plasma assisted chemical vapour deposition is
one of the most promising techniques for low temper-
ature deposition of DLC and polycrystalline diamond
for large-scale industrial use. For low temperature
deposition the precursor gases need to be activated
using microwave, laser, radio frequency (RF) or d.c.
plasma to achieve respectable growth rates. Without
this enhanced precursor activation, extreme substrate
temperatures are required and this severely limits the
number of potential applications. In order to improve
the adhesion, crystallinity and density of the films, the
deposition system used in this work was modified.
A venturi magnetic confinement ring and a filament
were incorporated to increase the plasma intensity.
The high intensity bombardment of the growing film
could affect the properties of the deposited film in
0022—2461 ( 1997 Chapman & Hall
1931

a beneficial way. For many CVD coatings, the sub-
strate surface structure can have a pronounced influ-
ence on the coating characteristics. In this publication
DLC coatings, produced using a hybrid low temper-
ature RF plasma assisted CVD system (BEP Grinding
Ltd.) [15—17], were investigated using different sub-
strates.
2. Experimental details
The hybrid plasma assisted CVD system i.e. a system
which incorporates aspects of physical and chemical
vapour deposition [10], used to deposit the coatings
consisted of a gas handling and preheated assembly
reactor vessel, with a RF plasma generator and
vacuum pumping system and is shown in Fig. 1. The
plasma reactor comprised a double skinned water
cooled non-magnetic circular stainless steel vessel (C).
The loaded system had an ultimate vacuum of
1.333;10\ꢀ Pa using a rotary pump roots blower
combination. The substrate holder (E) was located
above the multi-pumping ports and was attached to
the powered electrode (D). The ports were fitted with
ceramic restrictors for plasma confinement. Precursor
gases were introduced into a preheating chamber
prior to entry into the plasma region. An RF power
supply and matching network was used to generate
the plasma between the two electrodes. To produce
a high negative bias potential at the substrate holder
at low input power, a stainless steel cage and venturi
magnetic confinement ring (B) was fitted to the in-
terior of the vessel. The cage and capital venturi ring
were electrically isolated and at floating potential to
the plasma. The RF power supply and matching net-
work impedance was set for automatic match in hy-
drogen between 2.7;10\ꢁ and 67 Pa. The substrates
were pre-cleaned prior to entry into the vacuum sys-
tem where they were subjected to additional plasma
cleaning in hydrogen and argon plasmas prior to
deposition. Both plasma cleaning and deposition were
carried out at 8 Pa with 300 and 600 V d.c. negative
bias voltages at the substrate, respectively. The gas
following technique was used with a single gas deliv-
ery line. Hydrogen, argon and/or nitrogen followed by
butane were introduced into the reactor, thus facilitat-
ing a smooth process and eliminating gas change-over
problems. Reflected power was always kept to below
10 W. To enhance ionization of the plasma, an elec-
tron filament was (D) incorporated into the reaction
zone. The deposition conditions were optimized using
Taguchi methods.
Figure 1 Schematic of the hybrid plasma assisted CVD system.
thermoanalysis technique was employed. It utilizes
temperature controlled heating of the sample in an
oxygen atmosphere resulting in the formation of
oxides of carbon. The amount of carbon oxides are
continuously monitored as a function of temperature.
Accurate data on the high temperature stability of
DLC coating/substrate is essential in predicting indus-
trial performance. A LECO Corporation RC-412 in-
strument was used for this investigation.
Pin-on-disc wear testing of a DLC coating on a
zirconia substrate was also carried out. The DLC was
not doped with a secondary material such as silicon
and it gave adequate coating adhesion without the use
of secondary materials [6] and an interface coating
was not considered necessary. A BICERI universal
wear machine, in the pin-on-disc configuration, was
used to undertake the investigation [19]. The test
procedures conformed to those now accepted interna-
tionally [20, 21]. The zirconia substrate was polished
to an Ra value of 35 nm prior to deposition. The
as-deposited DLC coating assembly was aged in air
under natural conditions for one year before testing.
The following wear conditions where used: geometry,
pin-on-disc with vertical pin of 10 mm diameter AISI
52100 ball; 20 N load; 0.033 m s\ꢁ speed; 14408 s
duration and 21 °C test temperature.
The coatings produced were characterized using
a variety of analytical techniques. These techniques
included Raman spectroscopy (RS) of coatings on
glass substrates, scanning electron microscopy (SEM)
of coatings on Al/Si MMC and thermoanalysis (TA) of
coatings on aluminium and steel. Wear resistance
characteristics when cutting metal were assessed using
1/4 inch (6.35 mm) drills coated with DLC and TiN,
powered by a Bridgeport CNC machine. The drills
were originally manufactured from high speed steel.
For the determination and identification of various
forms of carbon present in the films deposited [18] the
3. Results and discussion
Raman spectra were recorded between 200 and
1800 cm\ꢁ in order to obtain information regarding
the degree of crystallinity of the coatings. Figs 2 and
3 show Raman spectra recorded for two samples of
DLC on glass substrates grown under similar depos-
ition conditions using butane as the source gas. In the
second sample, nitrogen was introduced into the sys-
tem prior to deposition, the gas sequence being argon
and butane for sample 1 and argon, nitrogen and
butane for sample 2. In both cases the spectra are very
1932

reciprocal of crystallite sizes (D/GJ1/C ). The D/G
ꢀ
ratio indicates that the crystallite size is approximately
10 nm [1, 22].
DLC films were deposited using the hybrid plasma
assisted CVD system onto aluminium and stainless
steel substrates. Fig. 4 shows a thermoanalysis spec-
trum for films deposited onto an aluminium substrate
and Fig. 5 shows a spectrum for films deposited on
a stainless steel substrate. The coating deposited on
a stainless steel substrate exhibits two peaks, a strong
peak at 520 °C and a very weak peak at about 790 °C
perhaps due to the oxidation of two types of species.
However the film deposited on an aluminium sub-
strate shows a single peak at 535 °C. These films were
deposited under identical conditions in the hybrid
PACVD system. The major peak for synthetic dia-
mond appears at between 650—750 °C and for plasma
deposited polycrystalline diamond film at 1072 °C
[18]. For amorphous carbon the major peak is usually
shown between 420—490 °C [18]. It appears from the
initial analysis that for both experimental coatings,
irrespective of the substrate, small crystallites of dia-
mond in a graphite matrix are present, with those on
steel containing a greater degree of diamond crystal-
linity. These results show that the characteristics of the
substrate surface can have a significant effect on the
structural properties of the carbon coatings. This is
well known since CVD is a surface sensitive process
and thus highly dependent on the surface structure
and composition for a variety of coatings [23]. The
quantification of carbon on both substrates showed
the average surface concentration to be in the order of
150 lg cm\ꢀ and 70 mg cm\ꢀ for aluminium and
stainless steel, respectively.
Figure 2 Raman spectrum of DLC without nitrogen in the process.
Figure 3 Raman spectrum of DLC with nitrogen introduction prior
to deposition.
Fig. 6a shows a scanning electron microscope
(SEM) micrograph of DLC deposited by hybrid
PACVD at low temperature onto aluminium/silicon
metal matrix composites (MMC). The figure reveals
a typical MMC substrate structure which is porous
with numerous voids and a matrix of material. For
applications such as the modernized two-stroke en-
gine, structures of this type of material make them
light and durable. However, deposition of a number of
coatings onto these materials is difficult using conven-
tional vacuum deposition methods such as PVD and
CVD, due to the porosity and inhomogeneous surface
of the material. SEM examination reveals these voids
to be variable in size but typically in 5—25 lm range.
The DLC coating structure deposited onto the MMCs
is very dense with a rough surface. There are several
factors which contribute to the film density and ad-
hesion to the substrate. An important consideration is
that the magnetic confinement and the substrate bias
which confine the electrons and ions in the region
above the substrate and positive ions are accelerated
towards the substrate causing dense films and much
improved adhesion. The beneficial effects of ion-
bombardment on the density and adhesion of coatings
have been discussed in ion-assisted PVD processes
[6]. In other ion-assisted processes, such as unbal-
anced magnetron sputtering and dynamic recoil mix-
ing, the ions bombard the growing film giving superior
adhesion [8]. An additional feature of the current
similar with two major peaks between 1000 and
1800 cm\ꢁ and minor peaks at the lower end of the
spectra. The spectrum for sample 1 (without nitrogen)
shows the major peaks appearing at 1365 and
1575 cm\ꢁ. However the addition of nitrogen results
in better resolution and a shift from 1365 to 1350 cm\ꢁ
and from 1575 to 1585 cm\ꢁ. The two major peaks are
referred to as G and D bands [1, 22]. In crystalline
graphite only the G band appears at 1575 cm\ꢁ with
a narrow bandwidth. For pure diamond a sharp single
band is expected at 1332 cm\ꢁ. The D band occurs at
around 1360 cm\ꢁ and shows the presence of finite
crystallite sizes from pyrolytic graphite, glassy carbon
and carbon powders and could be a consequence of
spꢂ bonded species. It is found that the D band shifts
to higher wave number with increasing disorder. In
these experiments, the addition of nitrogen caused the
D band to shift slightly to a lower wave number,
perhaps indicating a small improvement in the degree
of crystallinity. The reason for nitrogen causing im-
proved crystallinity is unclear, however, it may be
related to the exclusion of oxygen from the surface
prior to deposition. This observation has been re-
ported in the literature [1] and requires further invest-
igation. These results indicate, due to the shift of the
D band, that the films consist of small crystallites of
diamond characteristic of DLC material. It has been
shown [1] that the D/G ratio is proportional to the
1933

Figure 4 Thermoanalysis of DLC on aluminium.
Figure 6(a) SEM of DLC coating onto MMC substrate. (b) SEM of
coated and uncoated regions of MMC substrate.
Figure 5 Thermoanalysis of DLC on stainless steel.
biased coating/substrate and their high energy causes
sputtering of the growing films thus giving rise to the
rough surface observed. The effects of the plasma
parameters, substrate temperature and system pres-
sure need to be investigated further.
Table I outlines the wear resistance tests performed
on 1/4 inch HSS drills using a Bridgeport CNC
machine. The use of a CNC machine ensured that all
tests were carried out under identical and consistent
conditions. These drills were coated with DLC and
were compared to drills coated with TiN. Test 2 shows
that DLC was superior to TiN. In this test the drills
were used on thyroplast steel.
Pin-on-disc wear testing illustrated some interesting
behaviourial information including an initial low fric-
tion and wear before a steady state is achieved. These
phenomena are shown in Figs 7 and 8. The displace-
ment wear, which is typical of DLC coatings, is seldom
mentioned by expert tribologists as the mechanisms at
work in the contact wearing process are not fully
understood [24, 25]. Wear test results are given in
Table II.
Based on microstructural observations, the behav-
iour of the DLC coating—substrate system when tested
for wear with the pin-on-disc machine can be ex-
plained in the following way. In the early test period
there is little interaction between the pin and coating
because interface temperatures are low and this results
in low friction coefficients. Increasing temperatures at
the contact point result in rapidly rising friction and
process is the use of an electron filament which in-
creases the ionization efficiency of the plasma as the
electrons continuously bombard the space in the gas
phase near the surface. An essential consideration for
improving wear performance in coated components is
the surface characteristics such as the degree of the
surface roughness. Fig. 6b shows an SEM micrograph
of DLC coated and uncoated region of MMC substra-
te. The uncoated region appears to have a flaky sur-
face whereas the coated region has a rough appear-
ance. Variable flakes up to about 30 lm appear in the
uncoated MMC and this effect could be due to the
surface preparation techniques employed. At higher
magnification the surface differences are more appar-
ent with the coated surface being much rougher than
the uncoated surface. One would expect a conven-
tional CVD process to procedure a surface roughness
not too different from the substrate [22]. Therefore it
is probable that in the modified CVD process other
factors need to be taken into consideration to explain
the unexpected observation. A possible explanation
for the rougher coated surface may be attributed to
high energy ion bombardment of the growing film due
to the use of an electron filament, magnetic confine-
ment and substrate bias. As mentioned earlier, all
these factors will have an additive effect and will tend
to increase the plasma intensity and ion bombard-
ment. Energetic positive ions bombard the negatively
1934

TABLE I Cutting and wear resistance trials on 1/4 inch steel drills coated with DLC and TiN
Drill number
11
Initial inspection comments
Number of holes
76 DU
Final inspection comments
Carbon, dark grey on tyre steel
AMG 1.3
2250 r.p.m.;657 mm min\ꢁ.
Arcs on chisel.
Cutting edges worn most at corners.
Coating off lands and flutes
behind lands.
Control
19
TiN on tyre steel
200 SC
AMG 1.3
2250 r.p.m.;657 mm min\ꢁ.
Cutting edge corners slightly worn.
Still cutting.
Carbon, dark grey on thyroplast steel
137 DU
AMG 1.5
1555 r.p.m.;178 mm min\ꢁ.
Cutting edge corners worn.
Heels slightly worn.
Coating off lands and flutes
behind lands.
Control
16
TiN on thyroplast steel
115 BO
AMG 1.5
1555 r.p.m.;178 mm min\ꢁ.
Worn down 13 mm front point.
Carbon, dark grey on stainless steel
132 SC
#11 SC
#4 DU
AMG 2.2
650 r.p.m.;85 mm min\ꢁ.
715 r.p.m.;93 mm min\ꢁ.
786 r.p.m.;102 mm min\ꢁ.
Swarf adhesion at cutting edges.
Chisel corner worn.
Cutting edge and corners worn.
Coating off lands and flutes.
Control
TiN on stainless steel
132 SC
#11 SC
#4 DU
AMG 2.2
650 r.p.m.;85 mm min\ꢁ.
715 r.p.m.;93 mm min\ꢁ.
786 r.p.m.;102 mm min\ꢁ.
Corners, cutting edges and
chisel badly worn.
Coating off lands and flutes.
15
Carbon, dark grey on cast iron
125 SC
125 SC
AMG 3.1
1540 r.p.m.;203 mm min\ꢁ.
Cutting edge and chisel
slightly worn.
Coating off lands.
Control
TiN on cast iron
AMG 3.1
1540 r.p.m.;203 mm min\ꢁ.
Cutting edge slightly worn.
Chisel corner worn.
Figure 7 Friction behaviour of DLC.
Figure 8 Wear characteristics of DLC.
consequent wear. A steady state is achieved and a rela-
tively constant proportion of the wear track is covered
with transferred material from the pin or ball. The test
then turns into a steel on transferred steel test and the
coefficient of friction gradually reduces as the contact
surface is worn smoother and lubricated with the
1935

TABLE II Pin-on-disc wear test results of DLC coating on zirco-
nia substrate
6. K. ENKE, Goram Advanced Materials Conference, London,
May 1990 (Presentation) (Goram Advanced Materials,
London, 1990).
Wear test results
7. H. E. SPEAR and J. P. DISMUHES, ‘‘Synthetic Diamond,
Emerging CVD Science and Technology’’, The Electrochemi-
cal Society Series (John Wiley & Sons, New York, 1994).
8. K. V. RAVI, Dimond Technology (John Wiley and Son Inc.,
New York, 1994).
Overall wear
Ball ring loss
Ring mass gain
Final friction coefficient
Minimum friction coefficient
Maximum friction coefficient
115 lm
0.0011 g
0.0008 g
0.48
0.23
0.83
9. D. TITHER, W. AHMED, M. SARWAR and E. AHMED,
J. Mater. Sci. ¸ett. 14 (1995) 1062.
10. D. TITHER, A. MATHEWS, P. JOHN, J. WILSON, Surf.
Coatings ¹echnol. 38 (1989) 251.
11. J. E. FIELD (Ed.), Diamond and Hard Materials (Academic,
London, 1979).
12. D. R. DECKER, A. J. SCHORR, IEEE ¹rans. Electron Devi-
ces, ED 17 (1970) 739.
13. J. RUSSEL, Diamond Deposition, 4 (1993) 172.
14. R. MESSIER, in Proceedings ICMC, 1987 (ICMI, San Diego,
1987).
DLC. Oxidation processes at the contact point and in
close proximity may also contribute to the physical
characteristics of the friction and wear of DLC coat-
ings on zirconia substrates [6].
15. Plasma 8 Workstation: Application note (BEP Grinding Ltd.,
Manchester, UK, 1994).
16. Low temperature RF Plasma Enhanced system 100: Applica-
tion Note (BEP Grinding Ltd., Manchester, UK, 1994).
17. W. MORAWSKI and D. TITHER, in Proceedings Catal-
Warsaw, 1992 (Catal-Warsaw, Poland, 1992).
18. Z. CIZEK, P. BOREK, J. FIALA and B. BOGDAIN,
Mikrochim. Acta (¼ien) III (1990) 163.
19. R. PENGELLY, Application notes, Universal Wear Testing
Machine, (BICERI Ltd, Slough, UK, 1994).
20. M. G. GEE, Guidelines for unlubricated sliding wear tests,
Part 1 General approach (National Physical Laboratory,
Teddington, UK, 1993).
21. Idem, Guidelines for unlubricated sliding wear tests, part 2.
Procedures for pin-on-disc testing (National Physical Labo-
ratory, Teddington, UK, 1993).
22. D. TITHER, A. BEHBI, P. HOLIDAY, A. MATHEWS,
A. G. FITZGERALD, A. E. HENDERSON, B. E. STOREY,
T. J. DINES, P. JOHN and J. I. B. WILSON, Carbon 28
(1990) 641.
4. Conclusions
DLC films have been deposited using a hybrid plasma
assisted CVD with RF excitation of butane with no
independent heating of the substrate. The addition of
nitrogen to butane prior to the deposition of the film
can aid growth of crystallites during deposition prob-
ably due to the elimination of oxygen from the surface.
As with all surface sensitive deposition processes in-
cluding CVD and PVD, both scanning electron
micrographs and thermoanalysis showed that the sub-
strate, such as MMCs, will have a significant influence
on the film properties [26]. Increasing the plasma
intensity by using a venturi magnetic confinement
ring, electron filament and substrate bias have been
shown to have beneficial effects on the film adhesion,
density and microstructure [22]. The hybrid plasma
CVD method was successful in depositing coatings
onto the inhomogeneous surface of the MMC sub-
strate [26]. DLC was also found to have superior
cutting and wear resistance properties compared to
TiN when deposited on HSS drills.
23. M. SARWAR, W. AHMED and D. TITHER, in Proceedings
of the First International Conference on Quality and Reliabil-
ity, Vol. 2, edited by W. S. Lau, T. C. Lee, H. C. Tsang and
C. Y. Tang, 11—12 April 1995, Hong Kong, (The Hong Kong
Polytechnic University, Hong Kong, 1995) pp. 233—236.
24. F. M. KUSTAS, M. S. MISRA, R. WEI and P. J. WILBUR,
¹ribology ¹rans. 36 (1993) 1.113.
25. D. KLAFFKE, R. WASCHE and H. CZICHOS, ¼ear 153
(1992) 149.
26. D. TITHER, W. AHMED, M. SARWAR, T. KOLSAKER
and E. AHMED, Islamabad J. Sci. (1996) (in press).
References
1. T. DINES and D. TITHER, Carbon 29 (1991) 225.
2. D. TITHER, ¹echnovation 10 (1990) 39.
3. D. V. FEDOSEEV and B. V. DERJAGUIN, ‘‘Science’’ paper
published by Nauka, Moscow, 1984.
4. P. BACHMAN, Phys. ¼orld 10 (1991) 16.
5. D. HOOVER, SUI-YUAN LYNN and DIWAKER GARG,
Solid State ¹echn. 104 (1991) 24.
Received 9 May 1995
and accepted 17 July 1996
.
1936