Characterization of cathodically deposited carbonaceous films on a silicon substrate

Article abstract of DOI:10.1023/A:1018638314997

Carbon-based deposits were electrochemically formed on silicon substrates in ethanol at room temperature. This work was based on the work reported by Namba, who described the electrochemical deposition of diamond from organic solutions. The deposits were analyzed using a scanning electron microscope, energy dispersive spectroscopy (EDS), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy and electrochemical impedance spectroscopy. Scanning electron micrographs showed some crystalline deposits on the silicon. EDS was unable to identify carbon in the film, but did reveal impurities such as sodium, potassium, calcium and zinc. It was later established that the impurities most likely came from impurities in the graphite used for a counter electrode. XPS showed the presence of carbon species, and subsequently Raman spectroscopy was used to classify further the carbon deposits. Raman spectroscopy showed the presence of amorphous carbon in some films, but no diamond peak was observed for any of the films. EIS revealed that the impedance of the deposited films was nearly identical to that of the uncoated silicon, and did not resemble the impedance of diamond. Thus, in this work, carbon-based films were formed electrochemically, but these films were not diamond.

Full text of DOI:10.1023/A:1018638314997

JOURNAL OF MATERIALS SCIENCE 32 (1997) 3123—3128
Characterization of cathodically deposited
carbonaceous films on a silicon substrate
S . E. KW IATEK*, V. DES AI
Departm ent of Mechanical and Aerospace Engineering, University of Central Florida,
Orlando, FL 32816, USA
P. J . M O RAN
* Mechanical Engineering Departm ent, United States Naval Academ y, Annapolis, MD 21402,
USA
P . M . N ATIS HAN
Naval Research Laboratory, Washington, DC 20375, USA
Carbon-based deposits were electrochem ically form ed on silicon substrates in ethanol at
room tem perature. This work was based on the work reported by Nam ba, who described the
electrochem ical deposition of diam ond from organic solutions. The deposits were analysed
using a scanning electron m icroscope, energy dispersive spectroscopy (EDS), X-ray
photoelectron spectroscopy (XPS), Ram an spectroscopy and electrochem ical im pedance
spectroscopy. Scanning electron m icrographs showed som e crystalline deposits on the
silicon. EDS was unable to identify carbon in the film , but did reveal im purities such as
sodium , potassium , calcium and zinc. It was later established that the im purities m ost likely
cam e from im purities in the graphite used for a counter electrode. XPS showed the presence
of carbon species, and subsequently Ram an spectroscopy was used to classify further the
carbon deposits. Ram an spectroscopy showed the presence of am orphous carbon in som e
film s, but no diam ond peak was observed for any of the film s. EIS revealed that the
im pedance of the deposited film s was nearly identical to that of the uncoated silicon, and did
not resem ble the im pedance of diam ond. Thus, in this work, carbon—based film s were
form ed electrochem ically, but these film s were not diam ond.
1. Introduction
electrical and mechanical properties, has also been
investigated. In this study, temperature was increased
in an effort to improve deposition rates. Alternate
methods of depositing diamond films in an organic
solution were also examined.
In recent years, several attempts have been made to
electrochemically deposit diamond or diamond-like
carbon films. Most notably, in 1992, Namba [1] re-
ported the deposition of diamond-phase carbon films
on a silicon substrate at low temperatures using an
organic solution of ethyl alcohol. Electrodeposition of
carbon films in an ethylene—glycol solution, and car-
bon fibres in a solution of carbon tetrachloride in an
organic solution, have been reported [2, 3]. Because
the electrodeposition takes place near room temper-
ature, it greatly increases the number of substrates
that can be used. Chemical vapour deposition (CVD),
which is the most common method of diamond film
deposition, must be done at temperatures greater than
800°C [4]. Therefore, only substrates with high
melting temperatures can be used. Depositing the
films at low temperatures also minimizes the thermal
stresses that can develop in both the films and the
substrate as they cool.
2. Experimental procedure
A schematic diagram of the experimental apparatus is
shown in Fig. 1. A flat specimen holder was used to
create the electrochemical cell. This holder places
a flat, circular specimen in physical contact with the
electrolyte and in electrical contact (by means of
a metal rod) with the power supply. A boron-doped
P-type silicon material with 1 0 0 orientation was used
as the substrate. The specimen was circular with an
area of 0.493 mmꢀ and a thickness ranging from
475—575lm. A graphite rod, a platinum wire and
a palladium wire were all used at various times as
counter electrodes for individual experiments.The
electrolyte was approximately 850 ml denatured ethyl
alcohol for the first four experiments and a 100% pure
ethyl alcohol for the remaining experiments. The dis-
tance between the anode and cathode was approxim-
ately 3.8 cm. Voltages varied from 1250—1850 V and
This work examined the nature of carbon films
formed using Namba’s experimental procedures [1],
and further investigated the feasibility of electrochemi-
cally depositing diamond films. Further characteriza-
tion of the deposited films, including the evaluation of
0022—2461 ( 1997 Chapman & Hall
3123

ments were made using a monochromatic. AlKa X-ray
source (X-ray energy was 1446.6 eV). The spot size was
600 lm, the pass energy was 50 eV, and the base
pressure was 8;10\ꢁ torr (1 torr"133.322 Pa) or
lower.
Deposits were also examined using a scanning elec-
tron microscope (SEM) and X-ray microanalysis. Se-
lected samples were analysed by Raman spectroscopy
performed in air and at room temperature. The sam-
ples were illuminated by directing radiation from an
Ar> ion laser emitting 40 mW at 514.6 nm, through
a microscope and on to the sample surface. The inelas-
tically scattered light was dispersed in a Spex Triple-
mate triple-pass monochromator where its intensity
was measured with a liquid-nitrogen cooled CCD
array. Electrochemical impedance spectroscopy (EIS)
was also used to investigate the electrical properties of
the deposited films on two samples.
Figure 1 Schematic diagram of experimental apparatus to deposit
carbon films cathodically on a silicon substrate.
currents ranged from 0.5—0.9 mA. Temperature was
varied from 20—52 °C, depending on the particular
run. The sample was driven cathodically in all tests.
Table I shows the experimental details of each experi-
ment.
3. Results and discussion
Scanning electron microscopy detected deposits on
every sample, although the appearance of the deposit
varied. The most common morphologies are shown in
Figs 2 and 3. The deposits either appeared granular or
showed a crystalline pattern. The thickness of the
deposits varied over the entire surface of the sample.
Energy dispersive spectroscopy was also used to
analyse all samples. Energy dispersive spectroscopy
(EDS) detected unexpected elements such as calcium,
zinc, potassium and sodium in several deposits. It was
also noted that during the second test, the ethanol
changed colour to a pale yellow and the current in-
creased during the deposition process. It is thought
that the high driving voltage caused the ethanol to
break down, and impurities in the ethanol were depos-
ited on the sample. Therefore, a 200 proof, 100% pure
ethanol was used in all remaining tests to prevent any
impurities from being deposited.
Specimens were prepared by rinsing with tap water,
distilled water and the denatured ethanol. After rins-
ing all pieces separately, the substrate was placed in
the specimen holder, rinsed with distilled water and
ethanol, and immediately placed in the electrolyte.
Deposition times varied in length from 20—140.5 h.
After the samples were removed from the solution
they were allowed to air dry before they were placed in
the SEM. The time between removal from solution
and SEM examination ranged from 5 min to 1 day.
The surfaces of selected samples were characterized
using X-ray photoelectron spectroscopy. The escape
depth for photoelectrons is of the order of nanometres
so that X-ray photoelectron spectroscopy (XPS) was
used to analyse the near-surface region. XPS measure-
TABLE I Parameters of the experiments
Sample
Counter
electrode
Ethanol
Temperature
Voltage
(V)
Current
(mA)
Time
(h)
( ° F)
( ° C)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Graphite
Graphite
Graphite
Pd
Denatured
Denatured
Denatured
Denatured
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
200 proof
68
68
68
20
20
20
20
49
44
44
47
20
20
20
20
20
20
52
20
20
20
20
1250
1750
1850
610
0.5—0.355
0.5—1.26
0.5—0.292
0.5—0.281
0.29—0.19
69
138.5
140.5
193
65
49.5
27
114
72
94
95
50
115.5
49
68
Pd
120
112
112
116
68
68
68
68
68
68
126
68
68
68
2000
2000
1000
1220
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
Graphite
Graphite
Pt
0.15—0.175
0.085—0.065
0.091—0.178
0.115—0.098
0.080—0.033
0.120—0.074
0.096—0.076
0.076—0.034
0.084—0.055
0.081—0.072
0.080—0.114
0.076—0.035
0.075—0.060
0.115—0.060
Graphite
Graphite
Graphite
Graphite
Graphite
Graphite
Graphite
New Graphite
Graphite
Graphite
Graphite
20
117.5
137.5
24
68
50.5
3124

Figure 4 EDS spectrum of silicon sample 12 that was generated
using a graphite counter electrode that had been used for 11 prior
runs. No impurities were noted.
Figure 2 Scanning electron micrograph of sample 12. The sample
was run at 1000 V for 50 h in 200 proof ethanol at room temper-
ature. The graphite counter electrode had been used in the 11
previous runs. The micrograph shows both granular and crystalline
deposits.
Figure 5 EDS spectrum of silicon sample 16 that was generated
with a previously unused graphite counter electrode. Impurities
calcium and potassium were noted.
electrode might be the source of the impurities. To test
this hypothesis, a new graphite rod was used as
a counter electrode. The EDS spectra showed that the
amount of impurities increased dramatically com-
pared to the previous samples. Fig. 4 (sample 12)
shows the spectrum of a sample done after eleven runs
had been performed using the same graphite counter
electrode. Fig. 5 (sample 16) shows the spectrum of
a sample that was generated using a previously unused
graphite counter electrode. Notice the considerable
difference in the amount and presence of the deposited
impurities. Thus, it appears that the graphite counter
electrode, when driven at high voltages, may have
been the source of some of the impurities.
Figure 3 Scanning electron micrograph of sample 16. Sample was
run at 1000 V for 117.5 h in 200 proof ethanol at room temperature.
This test utilized a new graphite counter electrode. The micrograph
shows a granular deposit only.
XPS analysis was used to examine selected deposits
to identify carbon species and to confirm the presence
of impurities as determined using EDS. Fig. 6 shows
an XPS spectrum obtained from a carbon disc which
was used as a reference material. The binding energy
of the large peak, 284.3 eV, is in reasonable agreement
Deposition of the impurities continued, even with
the pure ethanol, but it was observed that the amount
decreased as more tests were run. Because all samples
had been run using the same graphite counter elec-
trode, it was speculated that the graphite counter
3125

Figure 6 XPS of the carbon disc used as a reference to detect carbon
in the deposits.
Figure 8 XPS of sample 16 showing the large peaks indicative of
carbon and oxygen, as well as the peaks indicating the impurities.
sodium, zinc, calcium, magnesium and potassium. Accelerating
voltage of 100 eV, step"1.2 eV, AlK radiation, scan time"33
a
\
min 22 s, 20 scans.
—
—
Figure 7 XPS of sample 8 showing the presence of C—C, C—O, C O
bonding.
with literature values for C—C bonds in graphite [5, 6].
The two peaks to the higher binding-energy side of the
main peak have binding energies of 285.8 and 287.2
eV. The peak at 285.8 eV is indicative of C"O [7]
bonding while the peak at 287.2 eV suggests C"O
bonds [7]. These latter two peaks are most likely the
result of residuals from cleaning with acetone and
ethanol and/or impurities from the environment.
Fig. 7 show an XPS spectrum for sample 8. The
spectrum shows the presence of C—C bonds at 284.7
eV, C—O bonds at 286.3 eV [7], and C"O [7] or
possibly O—C"O [8] bonds at 288.5 eV. The C—O
peak and the C"O and/or O—C"O peaks account for
45% of the carbon spectrum. The peaks at 292.9 and
295.7 eV are the 2p —2p doublet for potassium.
Thus, the XPS analysis using a narrow scan in the
region of the C 1s peak, has shown the presence of
C—C, C—O, and C"O and/or O—C"O in the deposits
as well as potassium. It should be noted that XPS
cannot be used unambiguously to identify diamond
because diamond peaks have been reported to have
binding energy values ranging from 283.5—286.2 eV
depending on the surface treatment [9—11]. Therefore,
as discussed below, Raman spectroscopy was used to
determine if the carbon deposits had any diamond
character.
Figure 9 Raman spectrum of sample 16, showing the peaks at 1350
and 1588.1cm\ꢂ. These peaks are indicative of amorphous carbon.
greater than those of the deposited impurities, illustra-
ting that the relative amounts of carbon and oxygen
are higher than that of the impurities.
Raman spectroscopy was used to analyse several
samples to determine if the carbon deposits were, in
fact, diamond or diamond-like carbon. Raman is
a technique that can be used to differentiate easily
between diamond and other forms of carbon [12, 13].
Amorphous carbon peaks were identified on some
samples, but the distinctive diamond peak at
1332 cm\ꢂ was not observed on any of the samples.
Second-order silicon peaks, due to the silicon substra-
te were also generated on a few samples. Fig. 9 shows
the Raman spectrum from a deposit that is indicative
of amorphous carbon (1350 and 1588.1 cm\ꢂ) [14]. It
was also noted the samples that showed amorphous
carbon peaks were the samples that had the largest
amounts of impurities deposited on them. It would
appear that the presence of impurities may be acting
as a catalyst to the formation of amorphous carbon.
Lastly, EIS was carried out on two samples to
investigate the electrical properties of the deposited
film. Using Bode amplitude and phase plots, the impe-
dance of the coated silicon was compared with the
ꢃꢄꢀ
ꢂꢄꢀ
XPS was also used to confirm the presence of the
impurities detected on the samples by EDS. Fig. 8
shows a broad range scan of sample 16 and shows the
presence of sodium, magnesium, potassium, calcium,
nickel and zinc. These findings are consistent with the
EDS data generated earlier. It is also notable that the
intensity of the carbon and oxygen peaks are much
3126

4. Conclusions
1. XPS showed the presence of C—C, C—O and
C"O, bonds, as well as several impurities (sodium,
magnesium, potassium, calcium, nickel, zinc). Because
XPS is not a precise enough method to distinguish
between diamond and graphite, Raman spectroscopy
was used to classify the carbon deposits further.
2. Raman spectroscopy demonstrated the presence
of amorphous carbon in the films. The two broad
peaks in the range 1570—1510 cm\ꢂ and 1390—1350
cm\ꢂ are indicative of the presence of amorphous
carbon. Although the films are not diamond, some
amorphous carbon was deposited.
Figure 10 Bode amplitude plot of (᭢) bare and (£) coated silicon.
3. The capacitance of the deposited film, as deter-
mined by EIS, is closer to that of the silicon substrate
than to the capacitance of diamond. This is illustrated
in the Bode amplitude plot, which clearly shows that
at low frequencies, the capacitance of the film is very
dissimilar to the capacitance of diamond. Based on
this data, it seems that the film is very porous, which
would result in the electrical properties of the film
being comparable to that of the substrate .
4. A limited number of studies showed that the
graphite rod, when used as a counter electrode and
driven at high anodic voltages, released impurities
into the solution. These impurities were then depos-
ited on to the substrate. Thus, care must be taken any
time graphite is used as a counter electrode at high
anodic voltages. Further testing will be done to sub-
stantiate this claim.
Figure 11 Bode phase plot of (᭢) bare and (£) coated silicon.
5. Raman spectroscopy and EDS revealed that the
films with the largest amounts of impurities were those
identified by Raman spectroscopy as containing
amorphous carbon. Deposits with low amounts of
impurities only show a second-order silicon peak on
the Raman spectra. It is possible that the deposited
impurities are acting as a catalyst to the formation of
amorphous carbon.
6. Earlier reports [1] of diamond film deposition
by electrochemical means have not been substan-
tiated. Although the same experimental procedures
were followed, this work showed no evidence of
diamond. The results of Raman spectroscopy
showed no diamond peak, no diamond crystals were
observed under SEM, and the electrochemical impe-
dance spectroscopy reported a capacitance of the
deposited film that was very different from that of
real diamond.
Figure 12 Bode amplitude plot of (£) bare and (᭢) coated silicon,
as compared to (᭿) real diamond (linear fit).
bare silicon. A Bode amplitude plot was then con-
structed with the capacitance of a diamond film and
the impedance values of the bare and coated silicon.
Figs 10 and 11 show the similarities in EIS data
between the coated and bare silicon. The film does not
seem to affect the electrical properties of the substrate,
indicating that the film is so porous that it allows
wetting of the silicon to take place. Fig. 12 compares
the capacitance of the bare and coated silicon.
The straight line having a slope of !1 represents
the capacitance of a diamond film. At low frequencies,
there is a large difference in the impedance of the
bare and coated silicon and the diamond film. The
difference becomes smaller as the frequency
increaces, but not enough to match that of a diamond
film.
Acknowledgement
We acknowledge the USNA-NRL Cooperative Re-
search Program for its assistance. In addition, we
thank Pehr Pehrsson and Terri McCormick, Naval
Research Laboratory, for help with the Raman spec-
troscopy, Dr. John Murray for assistance with the EIS
study, and John Hein for sample preparation.
References
1. Y. NAMBA, J. »ac. Sci ¹echnol. A 10 (1992) 3368.
2. T. SUZUKI, Y.MANITA, T. YAMAZAKI, S. WADA, and
T. NOMA, J. Mater, Sci. 30 (1995) 2067.
3127

3. M. J. SHANE, B. K. CHINH, and M. J. SAILOR, ‘‘Extended
Abstracts,’’ Electrochemical Society, Fall Meeting, Chicago,
IL, 8-13, October 1995.
4. A. H. DEUTCHMANN, R. J. PARTYKA, Adv. Mater. Proc.
June (1989) 29.
5. J. F. MOULDER, W. F. STICKLE, P. E. SOBOL and
K. D. BOMBEN, ‘‘Handbook of X-ray Photoelectron
Spectroscopy’’ (Perkin—Elmer, Eden Prairie, MN, 1992)
p. 225.
6. C. D. WAGNER, in ‘‘Practical Surface Analysis’’, Vol. 1, 2nd
Edn, edited by D. Briggs and M. P. Seah (Wiley New York,
NY, 1990) Appendix 5.
10. P. E. PEHRSSON, in ‘‘Proceedings of the Third International
Symposium on Diamond Materials’’, edited by J. P. Dismukes,
K. V. Ravi, B. Lux and N. Setaka, (The Electrochemical
Society, Pennington, NJ, 1993) p. 668.
11. R. GRAUPER, J. RISTEIN, and L. LEY, Surf. Sci. 320, (1994)
201.
12. D. E. NEWBURY, D. C. JOY, P. ECHLIN, C. E. FIORI and
J. I. GOLDSTEIN, ‘‘Advanced Scanning Electron Micro-
scopy and X-Ray Microanalysis’’ (Plenum Press, New York
1986) p. 289.
13. D. S. KNIGHT, and W. B. WHITE, J. Mater. Res. 4 (2) (1989)
385.
7. D. BRIGGS, ibid., Ch. 9.
8. G. P. LOPEZ, D. G. CASTNER and B. D. RATNER, Surf.
Interface Anal. 17 (1991) 267.
14. M. YOSHIKAWA, G. KATAGIRI, H. ISHIDA, A. ISHI-
TANI and T. AKAMATSU, Appl. Phys. ¸ett. 52 (1988) 1639.
9. R. G. CAVELL, S. P. KOWALCZYK, L. LEY, R. A. POL-
LAK, B. MILLS, D. A SHIRLEY and W. PERRY, Phys. Rev.
B 7 (1973) 5313.
Received 23 August
and accepted 30 October 1996
.
3128