Deposition profile control of carbon films
on submicron wide trench substrate using H-assisted plasma CVD
Tatsuya Urakawa1, Takuya Nomura1, Hidehumi Matsuzaki1, Daisuke Yamashita1, Giichiro Uchida1,
Kazunori Koga1,4, Masaharu Shiratani1,4, Yuuichi Setsuhara2,4, Makoto Sekine3,4, and Masaru Hori3,4
1
Department of Electronics, Kyushu University, Fukuoka, Japan,
2
Joining and Welding Research Institute, Osaka University, Osaka, Japan,
3
Dept. of Electrical Engineering and Computer Science, Nagoya University, Nagoya, Japan,
4
JST,CREST, Tokyo, Japan
Abstract: We have demonstrated three kinds of deposition profiles of carbon films on
substrates with submicron wide trenches using H-assisted plasma CVD of Ar + H2 + C7H8.
The three deposition profiles are sub-conformal, conformal and anisotropic deposition, for
which carbon is deposited on top and bottom of trenches without being deposited on
sidewall of trenches. Experimental deposition profiles are determined by the balance
between deposition of carbon containing radicals and etching by H atoms. Irradiation of ions
hardens films and hence decreases the etching rate. When the etching rate surpasses the
deposition rate of carbon containing radicals, no deposition takes place there. Therefore, a
high H atom flux is the key to anisotropic deposition.
Keywords: carbon film, deposition profile, plasma CVD, trench
1. Introduction
Carbon has attracted much attention by the very
important advancement in its history such as the
development of the chemical vapor deposition
(CVD) method of diamond, diamond-like carbon
(DLC) [1, 2], and hydrogenated amorphous carbon
(a-C:H) [3, 4] as well as the discovery of C60 [5],
carbon nanotubes [6], and "freestanding" graphene
[7]. DLC films first reported in the 70s were
deposited by ion beam deposition [1, 2], and a-C:H
films introduced in the beginning of 80s were
deposited by rf plasma CVD [3]. Since then a-C:H
films deposited by rf plasma CVD have been studied
intensively by many researchers [4]. In addition, the
most widely used technique for DLC films
deposition is rf plasma CVD [8-12]. DLC is an
amorphous network solid, containing a high fraction
of carbon sp3 sites, but also sp2 sites and hydrogen
[12]. DLC films have attractive properties, such as
high mechanical hardness, wear resistance, optical
transparency and chemical inertness [8-12] and
hence have widespread applications as protective
coatings in several areas such as car parts, microelectromechanical systems (MEMS) and as magnetic
storage disks [12]. Deposition profile of the carbon
films in trenches is one of the concerns to realize
coatings on submicron wide trench substrates.
So far, we have succeeded in controlling
deposition profile of Cu films on trench, and have
realized sub-conformal, conformal and anisotropic
deposition, for which Cu is filled without being
deposited on sidewall of trenches, using a H-assisted
plasma CVD method [13-18]. We have applied the
method to carbon film deposition on trenched
substrates in order to control deposition profile [19].
In this paper, we report dependence of
deposition rate of carbon films in trenches on the
aspect ratio as a parameter of a discharge power of H
atom source PH. Based on the results we propose a
deposition kinetics model to identify the key to
anisotropic deposition.
2. Experimental
Experiments were performed using the Hassisted plasma CVD reactor, in which a
capacitively-coupled main discharge and an
inductive-coupled discharge for an H atom source
were sustained as shown in Fig. 1. This reactor
provided independent control of generation rates of
deposition radicals and H atoms. For the main
discharge, a mesh powered electrode of 85 mm in
diameter and a plane substrate electrode of 85 mm in
diameter were placed at a distance of 33 mm. The
discharge of H atom source was sustained with an rf
induction coil of 100 mm in diameter placed at 65
mm above the substrate electrode of the main
discharge. The excitation frequency is 13.56 MHz
and the supplied power PH = 0-500 W. The
excitation frequency of the main discharge was 28
MHz and the supplied power Pm was below 45W.
An rf bias voltage of 400 kHz was applied to the
substrate for controlling kinetic energy of ions
incident on it. The supplied power Pbias was 0-5 W.
Toluene (C7H8) was vaporized at 150oC, and
supplied with H2. The total flow rate of H2 and Ar
was 90 sccm. The total pressure was 13 Pa.
3. Results and Discussion
First, we have measured PH dependence of
electron density in the capacitively-coupled main
discharges using the plasma absorption probe.
Figure 2 shows the results. Electron density is nearly
constant irrespective of PH, indicating independent
control of generation rates of deposition radicals and
H atoms.
Next, we have studied effects of PH on deposition
rates at the bottom and sidewall of trenches. Figures
3 (a) and (b) show cross-section SEM images of
carbon films deposited on a trenched substrate. Subconformal deposition profile is obtained for PH = 0
10
1.5 10
R = (H2 / (H2+Ar))
R=33.3%
R=88.9%
1 10
e
-3
n (cm )
10
Thickness and deposition profile of carbon films
in trenches were obtained with a scanning electron
micro scope (SEM:JEOL, JSM-6320FZ). Electron
density, ne, in the capacitively-coupled discharges
was measured at 7 mm above the center of the
substrate electrode with a plasma absorption probe
using a network analyzer (Agilent Technologies,
E5071B) [20].
9
5 10
0
0
100
200
300
400
500
PH (W)
Figure 2. PH dependence of electron density in main
discharge.
Pm = 45 W, H2 + Ar 90 sccm, C7H8 2.5 sccm,
pressure 13 Pa, PH = 0-500 W.
Figure 1. Experimental setup.
Figure 3. Cross-section SEM images of carbon films
deposited on a substrate with trenches.
Substrate temperature 100 oC, Pm = 45 W, ion energy = 45
eV, H2 80 sccm, Ar 10 sccm, C7H8 2.5 sccm, pressure 13 Pa.
(a) PH = 0 W, and (b) PH = 500 W.
No deposition takes place at the sidewall of
trenches of all aspect ratio in Fig. 4 for the gas flow
ratio R = 33.3 % and PH = 500 W. In other words,
we have succeeded in deposition carbon films on
trenched substrates in an anisotropic way. The
deposition rates at the top, sidewall, and bottom are
nearly the same for the aspect ratio of 0.8, R =
33.3 % and PH = 0 W, namely we have realized
conformal deposition. Under other conditions except
these two, we have obtained sub-conformal
deposition.
Based on the results in Figs. 2-4, we propose a
deposition kinetics model as follows. Experimental
deposition profiles are determined by the balance
between deposition of carbon containing radicals and
etching by H atoms. When the etching rate surpasses
the deposition rate of carbon containing radicals, no
deposition takes place there. Irradiation of ions
induces structural modification at the film surface
[21]. The etching rate for the modified hard films is
significantly lower than that for the unmodified films.
Therefore etching rates at the top and bottom is lower
than that at sidewall, because ion fluxes on the top
and bottom are higher than that on the sidewall.
Moreover, incident deposition radical flux per
surface area at the sidewall and bottom is lower than
that at the top. Because of the lower incident
deposition radical flux per surface area and the
higher etching rate, the deposition rate at the
sidewall is the lowest. We can realize anisotropic
deposition rate at top (nm/min)
25
R=89% 500W
R=33% 500W
20
15
10
R=89% 0W
R=33% 0W
5
0
0
deposition rate at side (nm/min)
Figure 4 shows aspect ratio dependence of
deposition rate for PH = 0 W and 500 W, respectively.
The deposition rate at the top is constant irrespective
of the aspect ratio, whereas the deposition rates at
the sidewall and bottom tend to increase with
decreasing the aspect ratio. This is because the
incident deposition radical flux per surface area in a
trench increases with decreasing the aspect ratio. All
deposition rates at the top, sidewall and bottom
decreases with increasing PH from 0 W to 500 W.
The decrease in deposition rate for R=33% is larger
than that for 89%, indicating that the H atom flux for
R=33% is higher than that for 89%.
1
2
0
1
2
3
aspect ratio
25
R=89% 0W
R=33% 0W
20
R=89% 500W
R=33% 500W
15
10
5
0
0
1
2
0
1
2
3
aspect ratio
deposition rate at bottom (nm/min)
W in Fig. 3 (a), whereas anisotropic deposition
profile for PH = 500 W is obtained in Fig. 3 (b).
25
R=89% 0W
R=33% 0W
20
R=89% 500W
R=33% 500W
15
10
5
0
0
1
2
0
1
2
3
aspect ratio
Figure 4. Aspect ratio dependence of deposition rate at
top, bottom, side of trenches.
Substrate temperature 100 oC, Pm = 45 W, ion energy = 45
eV, H2 + Ar 90 sccm, R (H2 / (H2+Ar)) = 88.9% or 33.3%,
C7H8 2.5 sccm, pressure 13 Pa, PH = 0 W or 500 W.
deposition with increasing H atom flux to suppress
sidewall deposition. Therefore, a high H atom flux is
the key to anisotropic deposition.
4.Conclusions
We have studied the aspect ratio dependence of
deposition rate of carbon films on trenched
substrates as a parameter of PH. The following
conclusions are obtained in this study.
1) Electron density in the main discharge does not
depend on PH, indicating that the H-assisted plasma
CVD reactor provides independent control of
generation rates of deposition radicals and H atoms.
[5] H. W. Kroto, J. R. Heath, S. C. Obrien, R. F. Curl, and R. E.
Smalley, Nature, 318, 162 (1985).
[6] S. Iijima, Nature, 354, 56 (1991).
[7] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y.
Zhang, S. V. Dubonos, I. V. Grigorieva, A. A. Firsov, Science,
306, 666 (2004).
[8] J. Robertson, Mater. Sci. and Eng., R 37, 129-281 (2004).
[9] J. Robertson, Prog. Solid State Chem., 21, 199 (1991).
[10] J. Robertson, Surf. Coatings Technol., 50, 185 (1992).
[11] A.A. Voevodin and M.S. Donley, Surf. Coatings Technol.,
82, 199 (1996).
[12] J. Robertson, Phys. Stat. Sol. (a), 205, 2233 (2008).
[13] M. Shiratani, H. J. Jin, K. Takenaka, K. Koga, T. Kinoshita,
and Y. Watanabe, Sci. Technol. Adv, Mater., 2, 505 (2001).
2) We have demonstrated three kinds of deposition
profiles of carbon films on substrates with
submicron wide trenches using H-assisted plasma
CVD of Ar + H2 + C7H8. The three deposition
profiles are sub-conformal, conformal and
anisotropic deposition.
[14] K. Takenaka, M. Shiratani, M. Takeshita, M. Kita, K. Koga,
and Y. Watanabe, Pure Appl. Chem., 77, 391 (2005).
3) Experimental deposition profiles are determined by
the balance between deposition of carbon containing
radicals and etching by H atoms. Irradiation of ions
hardens films and hence decreases the etching rate.
When the etching rate surpasses the deposition rate
of carbon containing radicals, no deposition takes
place there. A high H atom flux is the key to
anisotropic deposition.
[17] J. Umetsu, K. Koga, K. Inoue, H. Matsuzaki, K. Takenaka,
and M. Shiratani, Surf. Coat. Technol., 202, 5659 (2008).
[15] K. Takenaka, M. Kita, T. Kinoshita, K. Koga, M. Shiratani,
and Y. Watanabe, J. Vac. Sci. Technol., A22, 1903 (2004).
[16] K. Takenaka, K. Koga, M. Shiratani, Y. Watanabe, and T.
Shingen, Thin Solid Films, 506-507, 197 (2006).
[18] J. Umetsu, K. Inoue, T. Nomura, H. Matsuzaki, K. Koga, M.
Shiratani, Y. Setsuhara, M. Sekine, and M. Hori,
Proceedings of 30th International Symposium on Dry
Process, (Tokyo, Japan), 35 (2008).
[19] T. Nomura, K. Koga, M. Shiratani, Y. Setsuhara, M. Sekine,
and M. Hori,. MRS Proceedings, 1222, DD05-16 (2009).
[20] H. Kokura, K. Nakamura, I. P. Ghanashev, and H. Sugai,
Jpn. J. Appl. Phys., 38, 5262 (1999).
Acknowledgements
This work was partly supported by JST, CREST
and MEXT.
References
[1] S. Aisenberg and R. Chabot, J. Appl. Phys., 42, 2953 (1971).
[2] S. Aisenberg and R.W. Chabot, J. Vac. Sci. Technol., 10, 104
(1973).
[3] K. Enke, H. Dimigen, and H. Hübsch, Appl. Phys. Lett., 36 (4),
291 (1980).
[4] K. Holmberg and A. Matthews, Coatings Tribology, Tribology
Series, 28, 440 (1994).
[21] W. Jacob, Thin Solid Films, 326, 1 (1998).