Electrodeposition of diamond-like carbon films in organic solvents using a thin wire anode

Article abstract of DOI:10.1016/S0009-2614(01)01040-5

Diamond-like carbon films (DLC) have been electrodeposited on silicon substrates in organic solvents using a thin tungsten wire anode under high potential. During the deposition, if the applied potential is sufficient enough, the spark could be observed even by the naked eye. It has been manifested that this method could produce an extreme environment which is favorable to form much more sp³ bonded carbon atoms compared with conventional electrolysis.

Full text of DOI:10.1016/S0009-2614(01)01040-5

2 November 2001
Chemical Physics Letters 348 (2001) 7±10
www.elsevier.com/locate/cplett
Electrodeposition of diamond-like carbon ®lms in organic
solvents using a thin wire anode
1
Hao Wang , Masahiro Yoshimura
*
Materials and Structures Laboratory, Tokyo Institute of Technology, Nagatsuta, Midoriku, Yokohama 226-8503, Japan
Received 8 August 2001; in ®nal form 31 August 2001
Abstract
Diamond-like carbon ®lms (DLC) have been electrodeposited on silicon substrates in organic solvents using a thin
tungsten wire anode under high potential. During the deposition, if the applied potential is sucient enough, the spark
could be observed even by the naked eye. It has been manifested that this method could produce an extreme envi-
ronment which is favorable to form much more sp3 bonded carbon atoms compared with conventional electroly-
sis. Ó 2001 Elsevier Science B.V. All rights reserved.
1. Introduction
examined to fabricate advanced materials includ-
ing carbon ®lms [1]. Gogotsi and Yoshimura [2]
reported carbon ®lm formation from aqueous so-
lution under hydrothermal condition. Recently
there is much interest in DLC and other carbon-
related ®lms deposited from liquid phase by elec-
trochemical methods since the work of Namba [3].
Wang et al. [4] reported that DLC ®lms could be
grown by electrolysis of methanol solution. Fu et
al. [5] reported that carbon nitride ®lms could be
synthesized by electrolysis of nitrogen-containing
organic solution. The electrochemical process has
some advantages compared with a gas-phase pro-
cess from an industrial point of view, including
low cost, possibility of coating on substrates with
complex shapes, control of ®lm thickness by elec-
trolysis charge, and composition by electrode po-
tential. Moreover, the electrochemical process is a
`soft solution process' [1], because the processing is
carried out under environmental friendly condi-
tions with relatively low energy (at low tempera-
tures, i.e., <200 °C). Meanwhile, there is
Diamond-like carbon (DLC) ®lms have been
the focus of considerable research e€orts in the
past three decades because of the similarity of their
properties to those of diamond. High energetic
processes like chemical vapor deposition (CVD)
and physical vapor deposition (PVD) have been
widely carried out to synthesize DLC ®lms.
However, they would cost environmentally and
economically due to the using of sophisticated
precursors and equipments, in addition to high
energetic processes. In this regard, solution pro-
cessing using aqueous and/or non-aqueous which
would be more environmentally friendly has been
* Corresponding author. Fax: +81-45-924-5323.
E-mail addresses: haowang@bjpu.edu.cn (H. Wang), yo-
shimu1@rlem.titech.ac.jp (M. Yoshimura).
1
Present address: Department of Material Science and
Technology, Beijing Polytechnic University, Beijing 100022,
China. Fax: +86-10-67-39-18-76.
0009-2614/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 9 - 2 6 1 4 ( 0 1 ) 0 1 0 4 0 - 5

8
H. Wang, M. Yoshimura / Chemical Physics Letters 348 .2001) 7±10
experiment evidence that most materials which can
be deposited from vapor phase can also be de-
posited in liquid phase using electroplating tech-
niques and vice versa [6].
copper rod. Before the experiment, one end of the
W wire was inserted into a conventional electrol-
ysis cell, with the electrolyte of 1 M KOH solution
and the ac voltage of 20 V, the wire was etched to
become sharper and sharper. Finally, the diameter
of the end part of the W wire can come up to be
less than 50 lm. The W wire was then bent to form
a structure that a pin (the sharp part of the W
wire) was vertically directed to a Si substrate plate.
The distance between the pin and the plate was set
to about 2 mm. The part of W wire which was
vertical to the solution was sealed by the glass
tube. Nitrogen gas was led into the solution to
exhaust oxygen to avoid the probable explosion
once when the spark happens. A magnetic stirrer
was utilized to promote the di€usion of solution.
The potential applied to the substrate can be
changed from 0 to 3000 V, the current limit is 100
mA. In this Letter, the typical samples were made
under the potential of 2000 V.
However, in the Namba [3] and Wang et al. [4]
experiments, both anode and cathode were plates
of graphite and Si substrate, no electrical glow
discharge had been reported in these papers, in-
dicating that the applied electrical energy was not
concentrated enough to breakdown the electrolyte.
In this Letter, a novel technique has been carried
out by using a thin metal wire as the anode.
During the experiment, when a certain high po-
tential was applied, spark might happen. It mani-
fested that, in this case, the energy distribution was
much more concentrated. Therefore, it could
produce an extreme environment which is much
more favorable for forming sp3 bonded C±C
bonds.
The deposits were characterized by scanning
electron microscopy (SEM), and Raman spec-
troscopy. The Raman spectra were excited using
the 244 nm of an intracavity frequency-doubled
2. Experimental
The schematic diagram of the deposition system
of this investigation is shown in Fig. 1. Analyti-
cally pure ethanol solvent was used as electrolyte.
A p-type Si h111i substrate with the resistivity of
10±20 X cm was mounted on a copper rod cath-
ode. The anode was a thin tungsten wire (the di-
ameter is about 0.5 mm) which was connected to a
‡
Ar laser.
3. Results and discussion
In the previous work by Namba [3] and Wang
et al. [4], both anode and cathode were plates of
graphite and Si, no spark had been observed
during the deposition process, indicating that the
applied electrical energy was not concentrated
enough to breakdown the electrolyte. Whilst in
this investigation, when a certain high potential
was applied, spark will happen. It manifested that,
in the case of a thin wire anode to the plate cath-
ode, the current density was much more concen-
trated. The relationships of the current and
temperature with deposition time of typical sample
(the applied potential was 2000 V) are shown in
Fig. 2. From Fig. 2 we can see that both the cur-
rent and temperature increase rapidly with time
almost linearly at the initial deposition time. At
this time, both the anode and cathode were cov-
ered with bubbles of gases. After about 100 s, the
current reached a maximum value, then decreased
Fig. 1. Schematic diagram of the deposition system.

H. Wang, M. Yoshimura / Chemical Physics Letters 348 .2001) 7±10
9
Fig. 2. The relationships of the current and temperature with
deposition time of typical sample (the applied potential was
2000 V).
gradually. However, the temperature continued to
increase, and ®nally reached a maximum value of
about 75 °C. At the time of about 1200 s, the
current began to increase, turbulence developed at
the anode and on further deposition time the spark
happened. We could even observe the spark with
the naked eye. The sequence of events leading to
spark conforms the proposal that, Joule e€ect
heating is the primary step in the breakdown of
normal electrolysis. To compare the structural
di€erence between the ®lms obtained by sparking
(the applied potential was 2000 V) and non-
sparking (1200 V), we have measured each ®lm
using SEM and Raman spectroscopy.
Figs. 3A and B show the SEM images of the
typical ®lms obtained by non-sparking and
sparking, respectively. We can see that the mor-
phology of the non-sparking ®lm is somewhat
corn-like, not compact. While the sparking ®lm is
composed of small, compact grains.
Raman spectroscopy is the most popular non-
destructive technique for the characterization of
carbon-related materials. Raman spectra of ®lm A
and B which were shown in Fig. 3 are demon-
strated in Fig. 4. Fig. 4A contains two broad
bands located at 1590 and 1380 cmÀ1, identi®ed
with the so-called Raman G and D bands, re-
spectively [7,8]. The sharp peak at 1550 cmÀ1 is
attributed to ambient O2 [9]. The G band corre-
Fig. 3. SEM images of typical (A) non-sparking ®lm, and (B)
sparking ®lm.
sponds to graphite-like layers of sp2 microdomains
in the ®lm, while the D band is attributed to bond
angle disorder in the graphite-like microdomains
induced by the linking with sp3 carbon atoms as
well as the ®nite crystalline sizes of sp2 microdo-
mains in the ®lm. The position, band-width of G
and D bands and the ratio of intensities of the D to
G bands (ID/IG) have been determined to be a
measure of the degree of structural disorder [7,8].
Besides the broad band at 1380 and 1590
cmÀ1, Fig. 4B shows another two distinctive
bands at about 1293 and 1452 cmÀ1. As was

10
H. Wang, M. Yoshimura / Chemical Physics Letters 348 .2001) 7±10
much more sp3 bonded carbon atoms compared
with conventional electrolysis.
4. Conclusion
(B)
(A)
In summary, DLC have been electrodeposited
on silicon substrates in ethanol solvent using a thin
tungsten wire anode under high electrical poten-
tial. During the deposition, if the applied potential
is sucient enough, the spark could be observed.
It has been manifested by Raman spectroscopy
that this method could produce an extreme envi-
ronment which is favorable to form much more
sp3 bonded carbon atoms compared with conven-
tional electrolysis.
Fig. 4. Raman spectra of (A) part A, (B) part B which are
shown in Fig. 3A.
demonstrated by Dillon et al. [7] and Beeman et
al. [8], a gradual transition from sp2 graphitic
spectra to sp3 diamond-like spectra were respon-
sible for the down shift of the D band to as low
as ꢀ1283 and ꢀ1265 cmÀ1, respectively. In the
investigation of Kumar et al. [10], they attributed
the lower peaks at 1300 cmÀ1 was due to sp3 C±C
vibration modes which does not have prefect
cubic symmetry. The band centered at about
1452 cmÀ1 had been reported in several literatures
[11,12]. This band was usually attributed to a
tetrahedrally bonded diamond precursor. Thus
we believe that the sample B which was prepared
under the sparking condition could be considered
to have appreciably improved diamond-like
character compared to that of non-sparking made
sample A.
It is proposed that, when electrolysis of a sol-
vent is carried out using a thin wire anode, a local
Joule heating might happen in the solvents, then
under suitable conditions of energy dissipation, the
glow discharge might happen if the applied voltage
is suciently high. In this case, a very shallow
primary reaction zone containing high concentra-
tions of radicals originating from the solvent can
be formed. Therefore, it can produce an extreme
environment which is closely analogous to those
brought about by plasma or ionizing radiation.
This is to say, some non-equilibrium reactions may
happen, or some metastable products may be ob-
tained. This environment is favorable to produce
Acknowledgements
Financial supports from Research Institute of
Solvothermal Technology, Kagawa Science and
Technology Forum, Takamatsu, Japan, are ac-
knowledged. The authors are thankful to Dr. T.
Fujiware and Dr. T. Watanabe for their help in
experiments, and Profs. M. Kakihana and Y.
Gogotsi for their discussions.
References
[1] M. Yoshimura, J. Mater. Res. 13 (1998) 796.
[2] Y.G. Gogotsi, M. Yoshimura, Nature 367 (1994) 628.
[3] Y. Namba, J. Vac. Sci. Technol. A 10 (1992) 3368.
[4] H. Wang, M. Shen, Z. Ning, C. Ye, C. Cao, H. Dang, H.
Zhu, Appl. Phys. Lett. 69 (1996) 1074.
[5] Q. Fu, J. Jiu, K. Cai, H. Wang, C. Cao, H. Zhu, Phys. Rev.
B 59 (1999) 1693.
[6] L.L. Maissel, R. Glang, in: Handbook of Thin Film
Technology, McGraw-Hill, New York, 1970, p. 5 (Chapter
5).
[7] R.O. Dillon, J.A. Woollam, V. Katkanant, Phys. Rev. B 29
(1984) 3482.
[8] D. Beeman, J. Silverman, R. Lynds, M.R. Andersn, Phys.
Rev. B 30 (1984) 870.
[9] K.W.R. Gilkes, H.S. Sands, D.N. Batchelder, J. Robert-
son, W.I. Milne, Appl. Phys. Lett. 70 (1997) 1980.
[10] S. Kumar, M. Malhotra, Diamond Relat. Mater. 7 (1998)
1043.
[11] P.K. Bachmann, D.U. Wiechert, Diamond Relat. Mater. 1
(1992) 422.
[12] B. Marcus, L. Fayette, M. Mermoux, L. Abello, G.
Lucazeau, J. Appl. Phys. 76 (1994) 3463.