Diamondlike carbon deposition on silicon using radio-frequency inductive plasma of Ar and C2H2 gas mixture in plasma immersion ion deposition

Article abstract of DOI:10.1063/1.122469

Diamondlike carbon (DLC) was deposited on silicon using a plasma immersion ion deposition (PIID) method. Inductive radio-frequency plasma sources were used to generate Ar and C₂H₂ plasmas at low gas pressures ranging from 0.04 to 0.93 Pa. The film stress and hardness were sharply dependent upon bias voltage at an operating pressure of 0.04 Pa. A maximum hardness of 30 GPa and compressive stress of 9 GPa was observed at a pulsed bias of -150 V bias (carbon energy of 80 eV). The mechanical properties of DLC films are correlated with UV Raman peak positions which infer sp³-bonded carbon contents.

Full text of DOI:10.1063/1.122469

Diamondlike carbon deposition on silicon using radio-frequency inductive plasma of Ar
and C 2 H 2 gas mixture in plasma immersion ion deposition
D. H. Lee, X. M. He, K. C. Walter, M. Nastasi, J. R. Tesmer, M. Tuszewski, and D. R. Tallant
Citation: Applied Physics Letters 73, 2423 (1998); doi: 10.1063/1.122469
View online: http://dx.doi.org/10.1063/1.122469
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/73/17?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Influence of the filling gas on plasma focus assisted diamondlike carbon coating at room temperature
J. Appl. Phys. 101, 063307 (2007); 10.1063/1.2713086
Microbridge testing of plasma-enhanced chemical-vapor deposited silicon oxide films on silicon wafers
J. Appl. Phys. 97, 104909 (2005); 10.1063/1.1898449
Electronic and structural properties of doped amorphous and nanocrystalline silicon deposited at low substrate
temperatures by radio-frequency plasma-enhanced chemical vapor deposition
J. Vac. Sci. Technol. A 21, 1048 (2003); 10.1116/1.1586275
Preparation of electrically conductive diamond-like carbon films using i- C 4 H 10 / N 2 supermagnetron plasma
J. Appl. Phys. 89, 2737 (2001); 10.1063/1.1344586
Evaluation of the ion bombardment energy for growing diamondlike carbon films in an electron cyclotron
resonance plasma enhanced chemical vapor deposition
J. Vac. Sci. Technol. A 16, 2625 (1998); 10.1116/1.581392
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
155.97.178.73 On: Tue, 25 Nov 2014 13:21:03

APPLIED PHYSICS LETTERS
VOLUME 73, NUMBER 17
26 OCTOBER 1998
Diamondlike carbon deposition on silicon using radio-frequency
inductive plasma of Ar and C2H2 gas mixture in plasma immersion
ion deposition
D. H. Lee,a) X. M. He, K. C. Walter, M. Nastasi, J. R. Tesmer, and M. Tuszewski
Los Alamos National Laboratory, Los Alamos, New Mexico 87545
D. R. Tallant
Sandia National Laboratory, Albuquerque, New Mexico 87185
͑Received 11 May 1998; accepted for publication 24 August 1998͒
Diamondlike carbon ͑DLC͒ was deposited on silicon using a plasma immersion ion deposition
͑PIID͒ method. Inductive radio-frequency plasma sources were used to generate Ar and C2H2
plasmas at low gas pressures ranging from 0.04 to 0.93 Pa. The film stress and hardness were
sharply dependent upon bias voltage at an operating pressure of 0.04 Pa. A maximum hardness of
30 GPa and compressive stress of 9 GPa was observed at a pulsed bias of Ϫ150 V bias ͑carbon
energy of 80 eV͒. The mechanical properties of DLC films are correlated with UV Raman peak
positions which infer sp3-bonded carbon contents. © 1998 American Institute of Physics.
͓S0003-6951͑98͒00443-4͔
Diamondlike carbon ͑DLC͒ or tetrahedral amorphous
carbon ͑ta-C͒ has been studied for many years using various
deposition methods, most of which employ energetic ion
species. Examples of techniques are mass-selected ion-beam
deposition,1 filtered vacuum arc deposition,2 plasma beam
deposition,3 and plasma source ion implantation ͑PSII͒.4
DLC is an attractive material because of its unique properties
such as high hardness, low friction coefficient, and chemical
inertness.
It was reported earlier that compressive stress is associ-
ated with a high fraction of sp3 C–C bonds.2 DLC coatings
with higher compressive stress ͑Ͼ6 GPa͒ generally exhibit
higher hardness and higher film density ͑2–3 g/cc͒. The high
stress in a DLC coating, however, may limit the thickness of
an adherent film to less than 100 nm on a silicon substrate.
Strong adhesion to the substrate is normally required to de-
posit thick DLC films with high compressive stress; there-
fore, energetic ion deposition methods are frequently em-
ployed to make DLC films.
maintains many advantages of PSII,6,7 but also allows depo-
sition of hard DLC coatings. Many possible applications of
the PSII process have been reported elsewhere.7–9
The PIID chamber volume was ϳ1.8 m3 and the pres-
sure before deposition was below 1.3ϫ10Ϫ4 Pa. The sample
stage was cooled by refrigerated water at a temperature be-
low 20 °C. The silicon surface was sputter cleaned with Ar
ions before deposition. An argon plasma was generated using
two inductively coupled 0.46 MHz rf power sources to ob-
tain high ion density. The plasma sources were located in
mirror-symmetric positions with respect to the sample stage
in the PIID chamber to get a uniform ion distribution.10 Each
rf source was operated at 200 W. The Ar-ion density in the
plasma was about 1010/cm2 at a chamber pressure 0.04 Pa.
The plasma potential in the inductive argon plasma was 23 V
at 0.04 Pa, as measured by a Hiden analytical cylindrical
Langmuir probe. The pulsed-bias voltage was Ϫ600 V dur-
ing the Ar sputter-cleaning process and the incident dose was
ϳ5ϫ1016/cm2.
DLC film deposition on Si was investigated recently us-
ing a pulsed-glow discharge carbon plasma from acetylene
gas in plasma immersion ion processing.5 A high negative
pulsed-bias voltage ͑Ͼ1 keV͒ generated hydrocarbon ions
͑assumed to be mostly C2H2ϩ) at pressures above 1.3 Pa and
accelerated them onto a substrate material to deposit DLC
films. The resulting DLC films exhibited low compressive
stress ͑less than 0.5 GPa͒, functional hardness ͑ϳ17 GPa͒,
and strong adhesion. The low stress enabled deposition of
DLC films thicker than a micron.
The experimental results of DLC deposition onto silicon
substrates ͓either ͑100͒ or ͑111͒ orientations͔ in a plasma
immersion ion deposition ͑PIID͒ chamber using radio-
frequency ͑rf͒ inductive plasma sources will be discussed in
this letter. The low-energy PSII method used in this work is
referred to as the PIID process. The PIID process not only
After the Ar sputter cleaning, the negative pulsed-bias
voltage was first reduced from 600 V to the desired voltage,
and then acetylene (C2H2) gas was introduced into the Ar
plasma. This procedure prevents carbons from depositing
onto the Si surface and forming a graphitic film. This gra-
phitic film may provide a weak interface between the Si sub-
strate and the subsequently deposited DLC film. The pulsed-
bias voltage was varied from Ϫ70 to Ϫ300 V with a pulse
width of 20 ␮s and a pulse frequency of 10 kHz. The oper-
ating pressure was varied from 0.04 to 0.93 Pa. The deposi-
tion rate was dependent upon various conditions. It was 1.4
nm/min at a pressure of 0.04 Pa and a 2:1 argon-to-acetylene
gas ratio. The thicknesses of the DLC films were between
100–230 nm, as measured by a surface profilometer and an
in situ laser thin-film interferometer.
Hardness was measured by a Nanoindentor II ͑from
Nanoinstruments Inc.͒, utilizing the continuous stiffness op-
tion. The hardness values obtained at a 50 nm depth from the
a͒
Corresponding author. Electronic mail: dhlee@lanl.gov
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/98/73(17)/2423/3/$15.00 2423 © 1998 American Institute of Physics
155.97.178.73 On: Tue, 25 Nov 2014 13:21:03

2424
Appl. Phys. Lett., Vol. 73, No. 17, 26 October 1998
Lee et al.
FIG. 2. The maximal stress and hardness were obtained at the lowest gas
pressure of 0.04 Pa. The argon-to-acetylene ratio was 5:1 and the pulsed-
bias voltage was Ϫ200 V. Compressive stress values ranged from 4 to 7.5
GPa, corresponding to nanoindentation hardness values of 24–28 GPa.
Nanoindentation hardness values were obtained at a depth of 50 nm from
the surface and the film thickness ranged from 100 to 230 nm.
FIG. 1. RBS and ERD spectra of DLC film on silicon. The channeled RBS
spectrum ͑solid line͒ shows a small amount of oxygen. The inset plot is an
ERD spectrum used for hydrogen content measaurement. The scattering
angles were 167° and 29° for RBS and ERD measurements, respectively.
Figure 2 shows how DLC film stress and hardness
change with gas pressure. The maximal compressive stress
and hardness were observed at the lowest pressure of 0.04
Pa. Higher gas pressure during processing results in more ion
scattering and broadening of the ion energy distribution
͑IED͒. Ions in the low-energy portion of the IED make gra-
phitic carbon links and result in poor surface adhesion.
Pulsed-glow discharge plasma generation is possible only at
high pressure and high bias voltage, which are not optimal
conditions for tetrahedral DLC deposition. However, the in-
ductive rf plasma sources can generate high-density plasmas
at low pressure and allow use of low bias voltage in the PIID
chamber.10 Capacitively coupled rf sources also generate
plasmas at low pressure. However, the ion density is an order
of magnitude lower than that for the inductive source.10
The compressive stress was 9 GPa at a pulsed-bias volt-
age of Ϫ150 V ͑Fig. 3͒. A carbon ion energy of around 100
eV in various energetic deposition methods is known to pro-
duce maximal values in properties, such as density, sp3 frac-
tion, stress, and hardness.3 In this result, the negative pulsed
bias of 150 V to achieve maximal hardness and stress values
corresponds to a carbon ͑either ionic or atomic͒ energy of 80
eV. The carbon energy was estimated by considering the
surface impact of C2Hϩ2 and taking into account the plasma
potential of 23 V in argon plasma. The plasma potential in-
creased the actual accelerating voltage from Ϫ150 to Ϫ173
V.
sample surface were compared systematically for all DLC
samples. The curvature of a Si beam was measured using a
laser reflection method before and after deposition and the
stress was calculated using Stoney’s equation.11 The error in
the stress measurement was estimated to be within 15% of
the average stress value. The DLC film composition was
measured by Rutherford backscattering spectrometry ͑RBS͒
4
using 2 MeV Heϩ. The hydrogen content was obtained by
elastic recoil detection ͑ERD͒ measurement with a 75°
beam-incidence angle to the surface normal. Channeling
alignment in the Si substrate was used to enhance the signal
of low Z-number impurities ͑such as oxygen͒ in the DLC
films.
The addition of Ar gas into the process chamber has
many advantages. It assists in creating an inductive plasma
for hydrocarbon gas molecules. Also, Ar ions sputter clean
the surfaces of the inductive plasma sources and minimize
the deposition of a conductive carbon coating. Effective op-
erations of inductive sources rely on the generated rf electric
and magnetic fields penetrating through the chamber volume.
A conductive carbon coating on the source surface absorbs
the rf fields and prevents inductive plasma generation.10 En-
ergetic Ar ions also provide ion-assisted mixing and energy
transfer to surface neutral carbon atoms.
Figure 1 shows a RBS spectrum, acquired using a nor-
mal incidence beam, of a DLC film deposited at a pressure of
0.04 Pa, pulsed-bias voltage of Ϫ200 V, and an argon-to-
acetylene gas ratio of 2:1. The channeled RBS spectrum in-
dicates a negligible amount of oxygen contamination. The
inset plot is the ERD spectrum showing a uniform hydrogen
distribution in the DLC film. The argon content of the DLC
film was 3–4 at. % and the hydrogen content was approxi-
mately 25–30 at. %, depending on deposition parameters.
The hydrogen content increased slightly with reduced nega-
tive pulsed-bias voltage. The film density, estimated from
RBS measurements and RUMP simulation results and film
thickness, ranged from 2.5 to 3.1 g/cc. There was no notice-
able difference in film properties for different Ar-to-C2H2
Surface hardness values reach up to 30 GPa at Ϫ150 V
pulsed-bias voltage. It should be noted that the actual film
hardness might be higher than the measured value due to the
effect of the softer Si substrate. For films synthesized at 0.53
Pa, Fig. 3 shows that a plateau on maximal stress and hard-
ness exists over a wide bias range.
An advantage of using a low-pressure process is good
film adhesion. DLC films easily delaminated from a Si sub-
strate deposited at 0.53 Pa when the thickness was over 200
nm. In contrast, DLC grown at 0.04 Pa has stronger adhesion
to silicon and it was possible to achieve thicker than 200 nm
DLC films with a 9 GPa compressive stress. The hardness of
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
gas mixture ratios ranging from 2:1 to 5:1.
the coating deposited with a pulsed-bias voltage of Ϫ70 V
155.97.178.73 On: Tue, 25 Nov 2014 13:21:03

Appl. Phys. Lett., Vol. 73, No. 17, 26 October 1998
Lee et al.
2425
FIG. 4. UV ͑262 nm laser light͒ Raman spectra of DLC films and amor-
phous carbon. The negative pulsed bias was 150 V for the DLC films. Solid
lines are fits using two Gaussian functions. Peak shifts are indicated by
dashed lines for the most intense peaks at 1550–1590 cmϪ1 and the shoulder
peaks at 1360–1400 cmϪ1. The films are identified by their deposition pres-
sures and by ‘‘a-C,’’ for an amorphous carbon film deposited by pulsed
laser deposition under conditions that favor a high sp2 content.
FIG. 3. Stress and hardness change with negative pulsed-bias voltage. The
optimal pulsed bias of Ϫ150 V corresponds to carbon energy of 80 eV. At
higher gas pressure, the film adhesion was poor and the stress and hardness
were less than the DLC films deposited at lower gas pressures.
silicon allowed deposition of Ͼ200 nm thick DLC film with
9 GPa of compressive stress. These mechanical properties
were correlated with higher sp3 content inferred from the
UV Raman spectra of the DLC films.
and gas pressure of 0.53 Pa is very small compared to the
one at 0.04 Pa. Poor adhesion to the substrate may cause this
apparent low hardness value at high pressure due to the
delamination of the film during nanoindentation.
One of the authors ͑D.H.L.͒ would like to thank Kent
Scarborough and Jerome Reynolds for their technical assis-
tance, and Dr. John Baglin for his valuable discussion and
comments. This work was performed at Los Alamos Na-
tional Laboratory under Contract No. W-7405-ENG-36 and
supported by the Department of Energy’s Office of Basic
Energy Sciences.
An indication of sp3 ͑fourfold-coordinated͒ carbon
bonding in the DLC films was obtained using ultraviolet
͑UV͒-excited Raman spectroscopy.12 The samples were irra-
diated with a 262 nm UV laser ͑25 mW͒ focused into a
diffuse line on the sample to prevent laser-induced damage.
The spectral resolution was 25 cmϪ1. Figure 4 shows UV
Raman spectra of DLC films deposited by PIID. Also in-
cluded is the spectrum of an amorphous carbon film depos-
ited by pulsed laser ablation under conditions that favor for-
mation of sp2 ͑threefold-coordinated͒ carbon. The solid lines
are fits using two Gaussian functions, and the dashed lines
indicate shifts in the peak position between the spectra. The
peak shifts from 1579 cmϪ1 ͑Pϭ0.53 Pa͒ to 1590 cmϪ1 ͑P
ϭ0.04 Pa͒ implies a higher content of sp3-carbon bonds.13
With UV excitation, and in the context of the growth condi-
tions and the physical properties of these films, the appear-
ance of a relatively low-frequency shoulder ͑ϳ1360 cmϪ1͒ in
the Pϭ0.04 Pa films suggests an enhanced fraction of
sp3-carbon bonding, rather than an sp2-bonding structure.
The higher-frequency shoulders ͑1380–1400 cmϪ1͒ in the
other two films indicate the presence of extended
sp2-bonded structures.
1 J. W. Rabalais and S. Kasi, Science 239, 623 ͑1988͒.
2 D. R. McKenzie, D. Muller, and B. A. Pailthorpe, Phys. Rev. Lett. 67, 773
͑1991͒.
3 M. Weiler, S. Sattel, K. Jung, H. Ehrhardt, V. S. Veerasamy, and J. Rob-
ertson, Appl. Phys. Lett. 64, 2797 ͑1994͒; M. Weiler, S. Sattel, T. Giessen,
K. Jung, H. Ehrhardt, V. S. Veerasamy, and J. Robertson, Phys. Rev. B
53, 1594 ͑1996͒.
4 M. Nastasi, A. A. Elmoursi, R. J. Faehl, A. H. Hamdi, I. Henins, G. W.
Malaczynski, J. V. Mantese, C. Munson, X. Qui, W. A. Reass, D. J. Rej,
J. T. Scheuer, C. E. Speck, K. C. Walter, and B. P. Wood, Mater. Res.
Soc. Symp. Proc. 396, 455 ͑1996͒.
5 K. C. Walter, D. H. Lee, N. Baker, and M. Nastasi ͑unpublished͒.
6 J. R. Conrad, U.S. Patent No. 4,764,394 ͑1988͒.
7 D. J. Rej, Handbook of Thin Film Process Technology ͑Institute of Phys-
ics, Philadelphia, PA, 1996͒, Chap. E2.3.
8 K. C. Walter, J. T. Scheuer, P. C. McIntyre, P. Kodali, N. Yu, and M.
Nastasi, Surf. Coat. Technol. 85, 1 ͑1996͒.
9 J. V. Mantese, I. G. Brown, N. W. Cheung, and G. A. Collins, MRS Bull.
21, 52 ͑1996͒.
10 M. Tuszewski, I. Henins, M. Nastasi, W. K. Scarborough, K. C. Walter,
and D. H. Lee, IEEE Trans. Plasma Sci. ͑to be published͒.
11 G. G. Stoney, Proc. R. Soc. London, Ser. A 82, 172 ͑1909͒.
12 K. W. R. Gilkes, H. S. Sands, D. N. Batchelder, J. Robertson, and W. I.
Milne, Appl. Phys. Lett. 70, 1980 ͑1997͒.
In summary, this is a report of DLC thin-film deposition
on silicon using the PIID method in conjunction with an
inductive rf plasma of Ar and C2H2 gases. The maximum
hardness was 30 GPa and the compressive stress was 9 GPa
at a pulsed-bias voltage of Ϫ150 V ͑carbon energy of 80 eV͒
and a chamber pressure of 0.04 Pa. The strong adhesion to
13 V. I. Merkulov, J. S. Lannin, C. H. Munro, S. A. Asher, V. S. Veerasamy,
and W. I. Milne, Phys. Rev. Lett. 78, 4869 ͑1997͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
155.97.178.73 On: Tue, 25 Nov 2014 13:21:03