Nanostructured diamond-like carbon films characterization

Article abstract of DOI:10.1016/j.jallcom.2009.10.195

In this work, we have studied the influence of the substrate surface condition on the roughness and the structure of the nanostructured DLC films deposited by high-density plasma chemical vapor deposition. Four methods were used to modify the silicon wafers surface before starting the deposition processes of the nanostructured DLC films: micro-diamond powder dispersion, micro-graphite powder dispersion, and roughness generation by wet chemical etching and roughness generation by plasma etching. The reference wafer was only submitted to a chemical cleaning. It was possible to see that the final roughness and the sp³ hybridization degree (that is related with the structure and chemical composition) strongly depend on the substrate surface conditions. The surface roughness was observed by AFM and SEM and the hybridization degree of the DLC films was analyzed by Raman Spectroscopy. Thus, the effects of the substrate surface on the DLC film structure were confirmed. These phenomena can be explained by the fact that the locally higher surface energy and the sharp edges may induce local defects promoting the nanostructured characteristics in the DLC films.

Full text of DOI:10.1016/j.jallcom.2009.10.195

Journal of Alloys and Compounds 495 (2010) 620–624
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
Nanostructured diamond-like carbon films characterization
∗
A.P. Mousinho , R.D. Mansano, M.C. Salvadori
Laboratório de Sistemas Integráveis do Departamento de Engenharia de Sistemas Eletrônicos da Universidade de São Paulo, av. Prof. L. Gualberto,
158, trav. 3, São Paulo – SP, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 July 2008
Received in revised form 22 October 2009
Accepted 23 October 2009
Available online 4 November 2009
In this work, we have studied the influence of the substrate surface condition on the roughness and the
structure of the nanostructured DLC films deposited by high-density plasma chemical vapor deposition.
Four methods were used to modify the silicon wafers surface before starting the deposition processes
of the nanostructured DLC films: micro-diamond powder dispersion, micro-graphite powder dispersion,
and roughness generation by wet chemical etching and roughness generation by plasma etching. The
reference wafer was only submitted to a chemical cleaning. It was possible to see that the final roughness
Keywords:
3
and the sp hybridization degree (that is related with the structure and chemical composition) strongly
Plasma process
Nanostructured films
Thin films
DLC
Carbon films
depend on the substrate surface conditions. The surface roughness was observed by AFM and SEM and
the hybridization degree of the DLC films was analyzed by Raman Spectroscopy. Thus, the effects of the
substrate surface on the DLC film structure were confirmed. These phenomena can be explained by the
fact that the locally higher surface energy and the sharp edges may induce local defects promoting the
nanostructured characteristics in the DLC films.
©
2009 Elsevier B.V. All rights reserved.
1
. Introduction
nisms of DLC films have been proposed, but very difficult to show
an exact model for growing DLC films, because several parameters
can influence in the deposition mechanisms of DLC films (depo-
sition techniques, time, precursor, DC BIAS applied powers, etc.)
[8–16]. Using the HDPCVD for the deposition of nanostructured DLC
film, the deposition mechanism is not influenced by sub-plantation
or by preferential sputtering of the sp2 and sp3 sites, because in a
HDPCVD system there is no DC bias applied to the samples. In this
case, the main mechanism for obtaining nanostructured DLC films
occurs changing the surface energy. The surface energy is increased
when different substrate topographies and different kinds of micro-
and nano-powders on the surface substrate (diamond and graphite
powder) are used. In this work, nanostructured DLC films were
One of the many forms of carbon, diamond-like carbon (DLC)
or tetrahedral amorphous carbon (ta-C), consists mainly of sp3
bonded carbon atoms. However, the DLC does not have a unique
composition, but consists of a mixture of amorphous and crystalline
phases [1–3]. DLC films contain carbon atoms in a variety of dif-
ferent coordinations. There are tetragonal coordinated sp carbon
atoms present in pure diamond as well as trigonal sp coordination
3
2
1
found in graphite, and possibly some sp -coordinated atoms. DLC
films may contain microcrystalline diamond and graphite as well as
a disordered structure containing a mixture of configurations. The
physical properties of these films depend on the method of prepara-
tion. For a given method, the ratio of sp to sp -coordinated carbon
can be affected by the presence of hydrogen in the films [1]. The
beneficial properties of DLC stem from the continuous rigid random
3
2
3
2
obtained with low surface roughness and high sp /sp hybridiza-
tion ratio [17,18]. The different substrate surface preparations have
strong influence on the characteristics of the nanostructured DLC
films, as is shown in this work.
3
networks of sp carbon atoms, and the properties can essentially
3
2
be tailored by controlling the sp /sp ratio [4–7]. The deposition
mechanisms of DLC can be described in both macroscopic and
microscopic terms. The macroscopic view attributes the forma-
2
.
Experimental
The nanostructured DLC films were produced in an HDPCVD system that uses a
2
tion of DLC to a phase transition between sp (a-C) and denser,
new concept of plasma generation. In this case, a planar coil coupled to an RF system
for plasma generation, and an electrostatic shield for plasma densification are used,
in this mode high-density plasmas are obtained. For the ion acceleration another
RF system in planar configuration is used as seen in Fig. 1. The substrate tempera-
ture was not controlled, but measured by a type K thermocouple, which indicated
3
metastable sp bonded DLC. The microscopic view describes the
ion processes giving rise to sp3 bonding. Many deposition mecha-
◦
that the temperature never exceeded 90 C. For the deposition of the films, pure
methane plasma processes were used. The substrates used to deposit the films were
three-inch diameter silicon wafers, 380 ␮m thick and orientation (1 0 0). They were
submitted to a Piranha clean, followed by a diluted HF dip, before the modification
of the surface. The surface modification is important for the increase of the surface
∗ Corresponding author. Tel.: +55 11 30915660; fax: +55 11 30915664.
E-mail addresses: mousinho@lsi.usp.br (A.P. Mousinho), mansano@lsi.usp.br
R.D. Mansano), mcsalvadori@if.usp.br (M.C. Salvadori).
(
0
925-8388/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2009.10.195

A.P. Mousinho et al. / Journal of Alloys and Compounds 495 (2010) 620–624
621
Fig. 2. Rms Roughness as a function of the process parameters.
Fig. 1. Scheme of the high-density plasma chemical vapor deposition system. (1)
Planar coil with an electrode shield, (2) pyrex window, (3) electrode, (4) secondary
distributor of gases and (5) main distributor of gases.
For the nanostructured DLC films deposited by HDPCVD, just as
for the deposition of others kinds of carbon films, the growth mech-
anisms for obtaining these films are not totally known. In this work,
we used different surface topographies for obtaining different
kinds of nanostructured DLC films. In this situation the roughness
depends on the surface topography. Films deposited on the ref-
erence samples and on the samples with diamond and graphite
powder, showed less roughness than the DLC films deposited on
the etched surface (plasma etching or wet chemical etching). For
the DLC films deposited on the samples with powders, the rough-
ness increases with the deposition time, and for the films deposited
on the etched samples, the roughness decreases with the deposi-
tion time. The roughness for the plasma etched samples before the
depositions were approximately 150 nm and for the wet chemical
etched samples were approximately 55 nm. In Fig. 3, we compared
atomic force microscope (AFM) images from nanostructured DLC
films deposited by the HDPCVD system. And in Fig. 4, we com-
pared SEM micrographics from these films. These images are from
samples with a deposition time of 4 h. The surface morphologies
of DLC films are very different from each other, and all samples
have different grain size. The different grain sizes are dependent
on the surface topography before the film deposition. Larger grain
size can produce more roughness. Different surface topographies
lead to a different DLC film growth. For example, for the DLC films
deposited on rough surfaces (plasma etching), at first, the atoms are
accumulated in the holes that are existent on the surface substrate,
but they cannot fill all the holes on the surface, in this situation
the roughness is larger for smaller deposition time (for the deposi-
tion time of 1 h, the roughness was 120 nm). When the final time is
increased, the atoms fill the holes and are accumulated around the
islands (formed by the filled holes) and the roughness decreases.
This effect promotes the reduction of the surface energy. This effect
is related to the DLC films growth rates. A smaller surface energy on
the substrate is related to a smoothing effect in the films deposition,
and in this condition, the roughness is smaller, because the atoms
have more easiness to form the DLC films. If the surface energy
is high, the roughness increases, because the atoms need the high
energy to be leagued and to form the film. In this situation, the final
roughness of the DLC films is strongly related to the surface energy
[19,20].
energy of the silicon wafers. Four methods were used for the surface modification of
the samples: diamond powder (<1 ␮m) dispersion (deposition by spinner, 1000 rpm,
3
0 s); graphite powder (<5 ␮m) dispersion (deposition by spinner, 1000 rpm, 30 s);
roughness generated by plasma etching (RIE system, SF6 plasma, 6.67 Pa, 100 W,
2
min); roughness generated by wet chemical etching (one part of fluoridric acid and
◦
nine parts of nitric acid or TMAH (2.5%wt @ 80 C) solution and the reference sub-
strate was only submitted to a chemical cleaning). For each method, the DLC films
were deposited with four different deposition times: 1, 2, 3 and 4 h. After prepa-
ration of the samples, the DLC films were deposited. The parameters of the DLC
films depositions were: 2.00 Pa, 250 W (coil power, RF, 13.56 MHz, remote plasma),
4
0-sccm methane. In these conditions, the deposition rate was 47 ± 1 nm/min. The
deposition rate is independent of the surface topography, and is directly related to
the process conditions. For evaluating the deposition rate, we measured the DLC
thickness, using a Dektak 3030 perfilometry from Sloan-Vecco. For the analyses of
the nanostructured DLC films, we used an AFM microscope model SPM 9500J3 (Shi-
madzu). Each sample was analyzed in the AFM in the tapping mode. For each sample,
we have made 10 analyses: 5 analyses in different areas of 15 ␮m × 15 ␮m and 5 ana-
lyzes in different areas of 5 ␮m × 5 ␮m. The scanning electronic microscope used for
analyzing the DLC films was a FEI NOVAnano400 microscope. We analyzed the sur-
face topography of the samples and the lateral profile after the final deposition. We
analyzed the nanostructured DLC films with a vis–UV Raman Spectroscopy system.
The spectra were collected using a Renishaw micro-Raman 2000 spectrometer on a
4
0× objective with a photo multiplicator. Unpolarized Raman spectra were acquired
−1
at ꢀ = 514.5 nm, the spectral resolution was about 4–6 cm and the power on the
sample was kept well below 1 mW. For calculating the intensities and the areas of
the peaks in the Raman spectra, we have made the deconvolution of these spectra
(
with Gaussian fit), using the Microcal Origin program, with Peak Fitting Mode. The
results obtained are shown in this work.
3
. Results and discussion
High-density plasma is quite for the growth of good quality
nanostructured DLC films. But the properties of these films are
largely dependent on the deposition conditions. The roughness of
the nanostructured DLC films is important, in this situation, because
it is affected by surface topography. Fig. 2 shows the Rms rough-
ness of the nanostructured DLC films deposited by the HDPCVD
system as a function of the deposition time. The Rms roughness
was analyzed in two different sizes in each sample (5 ␮m × 5 ␮m
and 15 ␮m × 15 ␮m) and we made five analyses per size. The mea-
surements were made in this way to check if the DLC roughness
does not change when the analyzed area changes, checking the
data reproducibility. With the AFM analyses we checked if rough-
ness does not change with the analyzed area for the nanostructured
DLC films, and the standard deviation was less than 1%. Observing
Fig. 2, we could see that roughness depends strongly on the surface
topography of the samples and on the final deposition time.
For longer deposition time, the atoms cover every island and
the spaces between one island and another, forming a continuous
film (for the deposition time of 4 h, the roughness was 40 nm). The
same effect occurs with the films deposited on the sample etched
with chemical wet. For the DLC films deposited on the surface
with diamond and graphite powder, the roughness increases with

6
22
A.P. Mousinho et al. / Journal of Alloys and Compounds 495 (2010) 620–624
Fig. 3. AFM images of nanostructured DLC films, with final time deposition of 4 h: (1) with diamond powder, (2) with graphite powder, (3) with plasma etching, (4) with
wet chemical etching and (5) reference sample.
related to the formation of sp bonds due to the presence of C–C sp3
bonds related to the tetrahedral structures in the DLC films. For the
reference samples the high density used for depositing the films,
promotes high energy during the growth films that increases sp3
bonds in these films.
3
the deposition time. For these processes, the initial surface energy
increases with the deposition time, and the roughness increases
too. These effects occur because of the influence of the surface
topography substrate before the nanostructured DLC films depo-
sition [21,22].
We have noticed that the DLC films deposited in reference sam-
ple by the HDPCVD system, showed the D and G bands well-defined
in the Raman spectra and the central position of these bands is
Due to the presence of other kinds of structures formed in
the DLC films, the modification of the surface topography sub-
strates before the deposition process. We observed that the D and
−1
−1
around 1330 and 1540 cm , respectively. For these samples, when
G bands appeared at around 1270 and 1500 cm , respectively.
the deposition time increases, the increasing of the ID/IG ratio
For the films deposited by the HDPCVD system on surfaces with
diamond and graphite powder, and etched by chemical wet and
plasma, the Raman spectra showed the first-order G band (around
occurs. The increasing of the ID/I ratio is related to the increas-
G
3
ing of the sp bonds in the DLC films. This effect can be confirmed
with the presence of a small peak around 950–1000 cm , this is
−
1
−1 −1
1500–1600 cm ) and D band (around 1230–1340 cm ), and some
Fig. 4. SEM micrographs of nanostructured DLC films, with final time deposition of 4 h (with 40,000×): (a) and (b) with diamond powder; (c) with graphite powder, (d)
reference, (e) with plasma etching and (f) with wet etching.

A.P. Mousinho et al. / Journal of Alloys and Compounds 495 (2010) 620–624
623
Fig. 5. ID/IG ratio as a function of the process parameters.
Fig. 6. Typical Raman spectrum for nanostructured DLC films deposited with pure
methane plasma by HDPCVD system.
second-order peaks (combination bands at approximately 1000,
−1
1
750 and 1860–925 cm ). The second-order bands are related
DLC films deposited on substrates with diamond powder show
complex structures in the film composition; in this case the films
have high disorder, which can be observed with the increasing of
the D band. The use of diamond powder promotes the tetrahe-
dral carbon bonds, with the high content of sp3 bonds that can
to the presence of the nanostructures in these films (tetrahedral
structure formation with high sp3 bonds concentration) and the
formation of carbon nanotubes during the growth deposition [23].
The intensity for all bands in the Raman spectra depends on the
surface topography substrate and on the deposition time.
−
1
be confirmed with the second-order band around 1000 cm . The
final structure in nanostructured DLC films deposited with diamond
powder is composed of amorphous carbon with graphite clusters,
First, we analyzed the D and G bands of the DLC films deposited
by the HDPCVD system. Fig. 5 shows the ID/IG ratio as a function of
the deposition time. We noticed that ID/IG ratio is strongly depen-
dent on the deposition time and of the surface topography used. For
the films deposited on the reference sample, diamond powder and
wet chemical etching, the ID/IG ratio increases when the deposition
time is increased. For the films deposited on a surface with graphite
powder and plasma etching, the ID/IG ratio, decreases when the
final time deposition is increased. These effects are related to the
second-order bands due to the presence of nanostructures in the
films and the surface energy generated by the different surface
topography used. For the films deposited on the surface etched with
plasma, the initial surface shows high roughness (approximately
3
tetrahedral carbon (with high sp content) and carbon nanotubes
−
1
(
around 1800–1970 cm region).
With the DLC films obtained on surface etched by chemical wet,
the ID/IG ratio increases with the deposition time (see Fig. 6), and the
disorder increases too. In these samples, the formation of tetrahe-
dral carbon structures in the films have occurred, but the carbon
nanotubes formation has not. The intensity of the second-order
3
band related to sp bonds, increase with the deposition time. These
effects are related to the flat surface generated when the samples
are submitted to a wet chemical etching.
The DLC films deposited on surfaces with graphite powder
showed minor ID/IG ratio, due to the long chains formation of
graphitic structures with some tetrahedral carbon bonds clusters in
the graphitic matrix, but these effects are decreased when the depo-
sition time increases. In DLC films deposited on surfaces etched
1
50 nm). In this situation, the initial surface has structures with
high aspect ratio (characteristic of the plasma etching processes).
The structure formed due the plasma etching are thin and the
distance between one and the other is small. In this case, the
surface energy is bigger, due to the presence of the privileged
points of deposition and this promotes the sp2 bonds formation,
that increase with the deposition time, generating DLC films with
with plasma the I /I ratio decreases with the deposition time.
D
G
The second-order band related to the tetrahedral bonds are formed
but with low intensity. The different nanostructures formation for
these films decrease when the deposition time increases and are
related to the initial high roughness of the substrates.
graphite-like characteristics. With the use of graphite powder, we
noticed a similar effect, but in this situation, the increasing of sp2
bonds in the films, were promoted by the presence of graphite
powder on the substrate surface. One of the characteristics of the
graphite grains is the formation of bi-dimensional structures that
4. Conclusions
2
privileged sp bonds. This effect is stronger when the deposition
time is increased, promoting the ID/IG ratio decreasing.
In this work, it was studied the influence of the surface substrate
topography in the formation of nanostructured DLC films deposited
by the HDPCVD system. We have observed that the structure forma-
tionofthese films dependsstronglyonthe substratecharacteristics.
For different surface topography used in this work (diamond and
graphite powder, plasma etching and wet chemical etching), the
surface energy is changed, which influences the roughness and
the structure formation of these films (graphitic, tetrahedral, car-
bon nanotubes, etc). For example, in the DLC films deposited
using diamond powder, tetrahedral carbon formation has occurred
(with high sp3 content) in amorphous carbon matrix, and car-
bon nanotubes. The formation of these structures depends on the
deposition time. On the other hand, this effect has not occurred
in the DLC films deposited on a surface with graphite powder,
The presence of diamond powder on a substrate surface pro-
3
motes the formation of tri-dimensional bonds (sp ). Diamond
powder is commonly used as seed for growing diamond films. For
the DLC films deposited by the HDPCVD system, the powder pro-
motes the increasing of the ID/IG ratio that is related to the sp3
content. This effect increases when the deposition time is increased.
For the films deposited on a surface etched with wet chemical, the
ID/IG increased with the increasing of deposition time. The rough-
ness generated with the wet chemical etching is smaller (compared
with the plasma chemical etching), and in this case, the surface of
the substrate before the deposition films is flatter. This effect pro-
motes to the decreasing of the surface energy, contributing to the
formation of clusters of sp3 bonds.

6
24
A.P. Mousinho et al. / Journal of Alloys and Compounds 495 (2010) 620–624
where the DLC showed only bi-dimensional graphitic structure for-
mations. This way, it was possible to confirm the effects of the
surface energy in the results of the DLC films structure formation.
[4] C. Martino, F. Demichelis, A. Tagliaferro, Diamond and Related Materials 4
1995) 1210–1215.
5] Z.Y. Chen, Y. Yu, J.P. Zhao, X. Wang, T.S. Shi, Journal of Applied Physics 83 (1998)
281–1285.
(
[
1
[
[
6] R. Raul, Journal of Physics D: Applied Physics 41 (2008).
7] O.S. Panwar, M.A. Khan, S. Kumar, Indian Journal of Pure & Applied Physics 47
Acknowledgements
(
2009) 141–148.
[
8] A.H. Lettington, Carbon 36 (1998) 555–560.
The authors would like to thank Dr. Marcos Massi and Mr. Sil-
vio A. de Souza for supporting the AFM analyses, and Mr. Nelson
Ordonez, Mr. Alexandre Camponucci for technical input and sup-
port, Mr. Dave Devito for supporting the English corrections, and
FAPESP, CAPES and CNPq for financial support.
[9] J. Robertson, Diamond and Related Materials 3 (1994) 361–368.
10] J. Robertson, Diamond and Related Materials 14 (2005) 942–948.
11] J. Robertson, Diamond and Related Materials 2 (1993) 984–989.
12] E. Neyts, Diamond and Related Materials 13 (2004) 1873–1881.
[13] J. Xu, Diamond and Related Materials 16 (2006) 45–49.
[14] B. Andre, Diamond and Related Materials 1 (1992) 307–311.
15] Y. Lifshitz, Diamond and Related Materials 4 (1995) 318–323.
16] A. Keudell, Diamond and Related Materials 11 (2002) 969–975.
[17] M.C. Salvadori, Diamond and Related Materials 10 (2001) 2190–2194.
18] J. Robertson, Materials Science and Engineering Reports 37 (2002) 129–281.
[
[
[
[
[
References
[
[
[
1] Q. Wei, J. Narayan, International Materials Reviews 45 (2000) 133–164.
2] X. Liu, D.M. Phototiadis, J.A. Bucaro, J.F. Vignola, B.H. Houston, Material Science
and Engineering A (2002).
3] A.C. Ferrari, J. Robertson, Philosophical Transactions of the Royal
Society A: Mathematical, Physical and Engineering Sciences 362 (2000)
[19] S. Meier, Surface & Coating Technology 202 (2007) 1267–1270.
[20] L.C. Cune, Nanomaterials and Biostructures 2 (2007) 315–319.
[21] J. Xu, Diamond and Related Materials (2006).
[22] C. Casiraghi, Diamond and Related Materials 14 (2005) 913–916.
[23] S.D.M. Brown, Physical Review B 61 (2000).
[
2
477–2512.