Synthesis of Crystalline Carbon Nitride Thin Films by Electrolysis of Methanol–Urea Solution

Abstract

Polycrystalline carbon nitride films were deposited on Si (400) substrates by electrolysis of methanol–urea solution under
high voltage, at atmospheric pressure and at temperature below 350 K. Fourier transform infrared spectroscopy (FTIR)
measurements suggested the existence of both single and double carbon–nitrogen bonds in the film. X-ray diffraction (XRD)
spectrum showed various peaks for different d values which could be assigned to different crystalline carbon nitride phases.
Film morphology was studied by scanning electron microscopy (SEM) which indicated the existence of grains with average
grain size of f2.5 Am.

Full text

Synthesis of crystalline carbon nitride thin films by
electrolysis of methanol–urea solution
S. Kundoo, A.N. Banerjee, P. Saha, K.K. Chattopadhyay*
Department of Physics, Jadavpur University, Calcutta-700 032, India
Received 17 August 2002; accepted 5 September 2002
Abstract
Polycrystalline carbon nitride films were deposited on Si (400) substrates by electrolysis of methanol–urea solution under
high voltage, at atmospheric pressure and at temperature below 350 K. Fourier transform infrared spectroscopy (FTIR)
measurements suggested the existence of both single and double carbon–nitrogen bonds in the film. X-ray diffraction (XRD)
spectrum showed various peaks for different dvalues which could be assigned to different crystalline carbon nitride phases.
Film morphology was studied by scanning electron microscopy (SEM) which indicated the existence of grains with average
grain size of f2.5 Am.
D2002 Elsevier Science B.V. All rights reserved.
Keywords: Carbon nitride thin films
1. Introduction
In a theoretical study, Cohen [1] and Liu and Cohen
[2] predicted that the hphase of a new and hypothetical
material C
3
N
4
, which is derived from h-Si
3
N
4
after
substituting silicon by carbon atoms, should show a
hardness comparable to or even greater than that of
diamond, based on empirical calculations of bulk
moduli of the covalent solids. This prediction aroused
tremendous interest among scientists in the fields of
condensed matter physics and materials science to
synthesize this material. Apart from predicted extreme
hardness, carbon nitride films may also prove to be
useful as electron emitters, variable band gap semi-
conductors and transparent hard coatings. Much effort
has been devoted to synthesize crystalline carbon
nitride films. These include reactive sputtering of a
graphite target in an argon/nitrogen plasma [3], r.f.
plasma-assisted hot filament chemical vapour deposi-
tion using methane–nitrogen gas mixture [4] or mix-
ture of methane, ammonia and hydrogen [5],ion
implantation [6], laser ablation of graphite in an atomic
nitrogen atmosphere [7], etc. Recently, we had used
thermal plasma method for the production of C
3
N
4
using CVD of graphite powder and nitrogen [8]. The
resulting film was amorphous, but contained large
amount of nitrogen, even greater than stoichiometry.
In fact, most of these above efforts led to amorphous
CN
x
embedded with carbon nitride crystallites. It
turned out to be difficult to prepare samples with the
right stoichiometric composition of crystalline C
3
N
4
.
To date, samples containing sufficient amounts of
crystallized C
3
N
4
and with bulk modulus comparable
to the predicted value have not been obtained.
0167-577X/02/$ - see front matter D2002 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0167-577X(02)01172-2
* Corresponding author.
E-mail address: kkc@juphys.ernet.in (K.K. Chattopadhyay).
www.elsevier.com/locate/matlet
Materials Letters 57 (2003) 2193 – 2197

All the above-mentioned methods to obtain carbon
nitride films involved vapour deposition techniques.
These methods generally require complex equipment,
high vacuum and high temperature which are unfav-
orable for the synthesis of metastable compounds such
as C
3
N
4
because they need to be quenched to avoid
undesirable atomic arrangements [9]. Also, the control
of experimental conditions is difficult. Compared to
the vapour deposition techniques, the electrodeposi-
tion from liquid phase has a number of advantages
such as the simplicity of the apparatus, and this
process can be performed at atmospheric pressure
and at nearly room temperature. This process is also
attractive because of its low production cost and easy
scalability. Only in recent years, very few attempts
have been made to synthesize carbon nitride films by
electrodeposition from various organic liquids under
high voltages [9,10].Amongthem,Fuetal.[9]
reported the synthesis of amorphous carbon matrix
containing some mixed polycrystalline phases of a-
C
3
N
4
and h-C
3
N
4
. There is no other published report
on the deposition of crystalline carbon nitride films by
the electrodeposition process. In this letter, we report
the use of a simple electrochemical deposition method
to synthesize the crystalline carbon nitride films from
methanol–urea solution on Si (400) substrates.
2. Experimental
The apparatus used in our experiment was a simple
electrolytic cell system. The schematic diagram of the
system is shown in Fig. 1. The Si substrate with a
resistivity of 20 Vcm was mounted on the negative
electrode. The counter-electrode was a graphite plate.
Substrate size was 1.5 1.0 cm
2
. The distance
between the electrodes was nearly 7 mm. Before
deposition, the substrate was treated by 20% HF
solution, then cleaned by distilled water and finally
by ultrasonic cleaner. Urea dissolved in methanol
solution was used as the electrolyte. The solution
was prepared by mixing 1.5 g of urea (CO(NH
2
)
2
)
(>99.9%) per 1 l of analytically pure methanol
(99.5%). As the dielectric constant (e) of methanol
is higher (32.70) [11] and the growth rates of depos-
ited films increase with increasing evalues, methanol
was used as the source solution, instead of using any
other solution containing ethyl groups. A DC power
supply, which can be varied from 0 to 3 kV, was used
to apply high voltage to the cathode. During the
experiment, electrolysis voltage was kept constant at
1 kV, the corresponding current was varying from 200
to 180 mA, and the temperature of the solution was
nearly 350 K. The deposition time was 30 min.
The deposited films were characterized by Fourier
transform infrared spectroscopy (FTIR), X-ray dif-
fraction (XRD) and scanning electron microscopy
(SEM) studies.
3. Results and discussion
During the deposition, voltage was kept constant,
but the current density changed. When applied poten-
tial was increased to 1 kV, the current density increased
linearly with applied potential. After deposition being
started, substrate current density decreased with time
and the growth rate saturated gradually with time. This
decrease in substrate current density may be due to the
increasing resistance of the deposited film on the con-
ducting substrate. The deposition mechanism may be
inferred as follows.
Under high voltage, the molecules of CH
3
OH and
CO(NH
2
)
2
were polarized, two groups became partly
discharged and moved in the electric field. The alkyl
group (–CH
3
) from methanol and ( – NH
2
) radical
from urea both with a part of positive charge moved
towards the cathode and finally carbon nitride film
deposited on the substrate.
We carried out IR spectroscopic studies to analyze
the structure and bonding state of the deposited
sample. The FTIR absorbance spectrum of a typical
Fig. 1. Schematic diagram of the apparatus used for the electro-
deposition of carbon nitride film on Si substrate.
S. Kundoo et al. / Materials Letters 57 (2003) 2193–21972194

carbon nitride film deposited on Si substrate is shown
in Fig. 2. The spectrum was recorded in an FTIR
spectrometer (Nicolet Magna-750) from 400 to 4000
cm
1
by subtracting the absorption due to the Si
substrate. The spectrum shows different vibrational
modes of various bonding. The band at 2923 cm
1
corresponds to the asymmetric –CH
2
stretch for a
typical hydrocarbon. It can also be observed that there
are asymmetric –CH
3
stretch mode at 2957 cm
1
and
symmetric stretch mode at 2857 cm
1
. All these
vibrational frequencies are typical of tetrahedrally
bonded carbon (sp
3
bonding) in hydrocarbons. The
absorption at 3200–3500 cm
1
suggests the exis-
tence of NH
x
(x= 1,2) bonds. The strong absorption at
1400–1700 cm
1
can be attributed to the presence of
CjN bonds. The absorption bands around 1100–
1300 cm
1
can be assigned with C – N stretching
vibration. The appearance of these peaks confirms
the presence of carbon –nitrogen bond formation in
the lattice. There is no evidence for the existence of
CZN bonds.
XRD pattern of the film is shown in Fig. 3. The
spectrum was recorded in a diffractometer using the
Cu Ka
1
radiation operating at 20 kV with a normal
h–2hscanning. The spectrum showed various sharp
diffraction peaks which indicated that the film con-
tained crystalline structure. Six diffraction peaks were
found in the spectrum at 2hvalues of 14.9j, 22.24j,
31.62j, 35.37j, 45.37jand 56.25j. The identification
of the peaks is given in Table 1. Since no other carbon
phase can be fitted with these lines, these are sup-
posed to be due to C–N phases. All peaks belong to
two sets of crystal constants. One set is the predicted
h-C
3
N
4
phase with lattice parameters a=6.44 A
˚,
c= 2.45 A
˚, and the other is a new phase with lattice
constants almost twice those of h-C
3
N
4
(a* = 12.5 A
˚,
c* = 4.71 A
˚), as suggested by Gu et al. [6] for their
carbon nitride thin films produced by nitrogen ion
Fig. 2. FTIR spectrum of a typical carbon nitride film deposited on
Si substrate.
Fig. 3. XRD spectrum of the as deposited film.
S. Kundoo et al. / Materials Letters 57 (2003) 2193–2197 2195

implantation of carbon films. The structure of the
latter phase is not yet clear. It may be a kind of
CN compound. Interplaner spacings (d)ofh-C
3
N
4
and the other compound are also listed in the table for
comparison. dspacings were calculated based on the
assumption of a hexagonal structure using the follow-
ing formula:
1
d2¼4
3
ðh2þhk þk2Þ
a2þl
c

2
:
Our results are nearly similar to those obtained by
Gu et al. [6]. From the table, it can be seen that the
measured values are in agreement with the calculated
dvalues, but with some deviations. The largest
deviation of dvalue for the (100) peak from the
theoretical value is nearly 6%, but for other peaks,
the deviations are < 2%. It can be mentioned here that
for the (100) peak, Feng et al. [12 ] reported deviation
>9.5% of the dvalue for their ionised cluster beam-
deposited crystalline CN films. Since most of the
diffraction peaks were found to match with the theo-
retically predicted h-C
3
N
4
data, we can conclude with
confidence that polycrystalline h-C
3
N
4
film was
formed.
The information on strain and the particle size
was obtained from the full-widths-at-half-maximum
(FWHM) of the diffraction peaks. The FWHM’s
(h’s) can be expressed as a linear combination of
the contributions from the strain (e) and particle size
(L) through the following relation [13]:
bcosh
k¼1
Lþesinh
k:
Fig. 4 represents the plot of bcosh/kvs. sinh/k.
Slope of the graph gives the amount of strain which
comes out to be 2.44 10
3
and the intercept on
bcosh/kaxis gives the particle size as f42 nm.
Using the predicted value of bulk modulus [14] of h-
C
3
N
4
and from the above-mentioned value of strain,
the stress generated in the films was calculated. It
comes out to be 1.1 GPa. This high value of
generated stress may be responsible for the lattice
distortion and hence shift in XRD peak positions.
Film surface morphology was studied by scanning
electron microscopy. Fig. 5 shows the SEM micro-
graph of the deposited carbon nitride film. The figure
shows the existence of grains in the film, as com-
Table 1
Theoretically predicted and experimentally measured lattice spac-
ings (d) and Miller indices of the corresponding planes giving XRD
peaks
Experimental Theoretical
h-C
3
N
4
New phase
d(A
˚)d(A
˚) (hkl) d(A
˚) (hkl)
5.94 5.577 (100)
3.99 4.09 (210)
2.82 2.789 (200)
2.53 2.49 (320)
1.99 1.949 (111)
1.63 1.610 (220)
Fig. 4. bcosh/kvs. sinh/kplot.
Fig. 5. SEM micrograph of the film deposited on Si substrate.
S. Kundoo et al. / Materials Letters 57 (2003) 2193–21972196

monly observed in a polycrystalline film. Study of
the micrograph reveals the average grain size of the
order of 2.5 Am. The particle size measured from
SEM micrograph is quite larger than the value
obtained from XRD data. The disagreement of par-
ticle size measurement by the two methods is quite
common in other materials. However, the possibility
that the large grains seen in SEM picture might be
composed of smaller crystallites is also not ruled out.
Hence, it can be concluded that FTIR, XRD and
SEM results provided evidence for the existence of
polycrystalline h-C
3
N
4
thin film deposited on Si
(400) substrate.
4. Conclusions
Polycrystalline carbon nitride thin films have been
successfully deposited on Si (400) substrates by the
electrolysis of organic liquid. As the growth rate of the
film deposited from the liquid containing methyl
group is higher than those from the liquids containing
ethyl group, methanol–urea mixture was used for the
electrodeposition process. FTIR spectrum of the film
confirms the presence of both CUN and CjN bonds.
Most of the XRD peaks match well with the theoret-
ical h-C
3
N
4
data. Also, scanning electron micrograph
provides support for the existence of grains in the as
deposited carbon nitride thin film.
Acknowledgements
Two of us, S.K. and A.N.B., wish to thank Council
of Scientific and Industrial Research (CSIR), Govt. of
India, for awarding them Junior Research fellowships
during the execution of this work.
References
[1] M.L. Cohen, Phys. Rev. B 32 (1995) 988.
[2] A.Y. Liu, M.L. Cohen, Science 245 (1989) 841.
[3] J. Wei, P. Hing, J. Appl. Phys. 91 (2002) 2812.
[4] Y. Chen, L. Guo, E.G. Wang, J. Cryst. Growth 179 (1997)
515.
[5] H.K. Woo, Y. Zhang, S.T. Lee, C.S. Lee, Y.W. Lam, K.W.
Wong, Diamond Relat. Mater. 6 (1997) 635.
[6] Y.S. Gu, L.Q. Pan, X.R. Chang, Z.Z. Tian, J. Mater. Sci. Lett.
15 (1996) 1355.
[7] C. Niu, Y.Z. Lu, C.M. Lieber, Science 261 (1993) 334.
[8] S. Matsumoto, K.K. Chattopadhyay, M. Mieno, T. Ando, J.
Mater. Res. 13 (1998) 180.
[9] Q. Fu, J.T. Jiu, K. Cai, H. Wang, C.B. Cao, H.S. Zhu, Phys.
Rev. B 59 (1999) 1693.
[10] Q.-A. Xu, Z. Ning, M. Shen, Y. Xin, B. Zhao, J. Zhu, Surf.
Coat. Technol. 122 (1999) 188.
[11] Q. Fu, J.T. Jiu, C.B. Cao, H. Wang, H.-S. Zhu, Surf. Coat.
Technol. 124 (2000) 196.
[12] J. Feng, C. Long, Y. Zheng, F. Zhang, Y. Fan, J. Cryst. Growth
147 (1995) 333.
[13] S.B. Quadri, E.F. Skelton, A.D. Dinsmore, J. Yang, H.F. Gray,
B.R. Ratna, Phys. Rev. B 60 (1999) 9191.
[14] A.Y. Liu, M.L. Cohen, Phys. Rev. B 41 (1990) 10727.
S. Kundoo et al. / Materials Letters 57 (2003) 2193–2197 2197