Direct identification of diamond growth precursor using almost pure CH4 or C2H2 near growth surface

Article abstract of DOI:10.1063/1.118352

Diamond growth was studied by injecting thermally decomposed Cl atoms into CH₄/H₂ or C₂H₂TH₂. Owing to the extremely short residence time (25 μs) and low gas temperature (< 1000 °C) in the decomposition system, the gas reaction is insignificant. Therefore, the carbon species near the substrate surface can be nearly identical to the input carbon source. With 0.3% CH₄ being the input carbon source, CH₄ remained the dominant carbon species near the surface (only 2.5% C₂H₂ was formed), and an almost continuous diamond film was deposited after 2 h growth. Raman spectra confirmed the formation of diamond. With 0.15% C₂H₂ being the input carbon source, C₂H₂ remained the dominant carbon species near the surface (10% CH₄ was formed), but only a few very small particles were deposited. Therefore, we conclude that CH₃ radicals seem the only diamond growth precursor under the Cl-rich conditions, whereas C₂H₂ is not efficient to grow diamond.

Full text of DOI:10.1063/1.118352

Direct identification of diamond growth precursor using almost pure CH 4 or C 2 H 2
near growth surface
Jih-Jen Wu and Franklin Chau-Nan Hong
Citation: Applied Physics Letters 70, 185 (1997); doi: 10.1063/1.118352
View online: http://dx.doi.org/10.1063/1.118352
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/70/2?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Step dynamics in the homoepitaxial growth of 6H-SiC by chemical vapor deposition on 1° offcut substrate using
dichlorosilane as Si precursor
J. Appl. Phys. 113, 184904 (2013); 10.1063/1.4803881
High Pressure Spectroscopic Study of the Action of NEthylamines and Water on the 2Nitropropane/Nitric System
AIP Conf. Proc. 706, 1265 (2004); 10.1063/1.1780468
Growth of diamond films on Ti–6Al–4V substrates and determination of residual stresses using Raman
spectroscopy
J. Vac. Sci. Technol. A 18, 2486 (2000); 10.1116/1.1285992
Study of hydrogenated diamond-like carbon films using x-ray reflectivity
J. Appl. Phys. 86, 289 (1999); 10.1063/1.370792
Activation energy for diamond growth from the carbon–hydrogen gas system at low substrate temperatures
Appl. Phys. Lett. 70, 173 (1997); 10.1063/1.118348
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.120.242.61 On: Wed, 26 Nov 2014 11:32:42

Direct identification of diamond growth precursor using almost pure CH4
or C2H2 near growth surface
Jih-Jen Wu and Franklin Chau-Nan Honga)
Department of Chemical Engineering, National Cheng-Kung University Tainan, Taiwan 701,
Republic of China
͑Received 24 June 1996; accepted for publication 5 November 1996͒
Diamond growth was studied by injecting thermally decomposed Cl atoms into CH4 /H2 or
C2H2 /H2. Owing to the extremely short residence time ͑25 ␮s͒ and low gas temperature ͑Ͻ1000 °C͒
in the decomposition system, the gas reaction is insignificant. Therefore, the carbon species near the
substrate surface can be nearly identical to the input carbon source. With 0.3% CH4 being the input
carbon source, CH4 remained the dominant carbon species near the surface ͑only 2.5% C2H2 was
formed͒, and an almost continuous diamond film was deposited after 2 h growth. Raman spectra
confirmed the formation of diamond. With 0.15% C2H2 being the input carbon source, C2H2
remained the dominant carbon species near the surface ͑10% CH4 was formed͒, but only a few very
small particles were deposited. Therefore, we conclude that CH3 radicals seem the only diamond
growth precursor under the Cl-rich conditions, whereas C2H2 is not efficient to grow diamond.
© 1997 American Institute of Physics. ͓S0003-6951͑97͒00302-1͔
Diamond film deposition has been of great interest due
to its enormous industrial potentials; yet, the diamond
growth mechanism is difficult to elucidate due to the com-
plex chemistry involved in the deposition system. The im-
portance of atomic hydrogen in diamond chemical-vapor
deposition ͑CVD͒ is now well recognized, but the diamond
growth precursor is still not clarified. Most researchers tend
to believe that CH3 is the major diamond growth
precursor;1–5 however, others have also proposed that C2H2
can grow diamond efficiently.6,7
CH4 or C2H2 dominant carbon species near the surface can
be obtained.
A schematic diagram of the hot-tube system employed in
this study is shown in Fig. 1. Pure Cl2 is dissociated into Cl
atoms in the innermost graphite tube ͑3 mm o.d., 1 mm wall,
and 25 mm length͒, the temperature of which can be mea-
sured through a view hole in the outer graphite tube. The
Cl-atom jet from the innermost tube is then mixed with the
Under typical diamond CVD conditions, hydrocarbon
radical reactions initiated by H atoms proceed so rapidly that
hydrocarbon compositions near the substrate surface are
nearly independent of the identity of the input carbon
source.8–10 Therefore, by minimizing or avoiding the hydro-
carbon reactions with H atoms several groups have at-
tempted to keep the carbon species near the growth surface
identical to the input carbon source in order to identify the
diamond growth precursor.3–6
In this letter we also report the identification of the dia-
mond growth precursor by using the CH4 or C2H2 dominant
carbon species near the growth surface created by injecting
thermally decomposed Cl atoms into CH4 /H2 or C2H2 /H2
feed, respectively. Due to the fact that the Cl—Cl bond ͑58
kcal/mol͒ is much weaker than the H—H bond ͑104 kcal/
mol͒, Cl2 can be readily dissociated at a much lower tem-
perature than H2. Besides, Cl atoms react rapidly with H2
forming H atoms through the fast exchange reaction,
ClϩH2→HϩHCl. H atoms formed can facilitate the dia-
mond growth. In addition, Cl atoms can initiate the free radi-
cal reactions and abstract hydrogen from the surface C–H
faster than H atoms, and thus enhance the diamond growth
rate.11 Pan et al.12 have successfully synthesized diamond
films by injecting thermally decomposed Cl atoms into a
CH4 /H2 mixture. Accidentally, we find that in a similar sys-
tem the gas phase reaction is relatively insignificant so the
a͒
Author to whom correspondence should be addressed.
Electronic mail: hong@mail.ncku.edu.tw
FIG. 1. The schematic diagram of the hot-tube system.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
Appl. Phys. Lett. 70 (2), 13 January 1997 0003-6951/97/70(2)/185/3/$10.00 © 1997 American Institute of Physics 185
129.120.242.61 On: Wed, 26 Nov 2014 11:32:42

TABLE I. The composition of gaseous carbon species ͑C2H2 and CH4͒ near
the substrate surface measured by mass spectrometry under the conditions:
͑a͒ in a hot-filament CVD reactor using 1.5% CH4 /H2, at a filament tem-
perature of 2050 °C, reactor pressure of 15 Torr, and total flow rate of 100
sccm; ͑b͒ in the hot-tube system using 0.3% CH4 and 8% C2 in the total flow
͑H2 in balance͒, at a reactor pressure of 5 Torr and total flow rate of 400
sccm; ͑c͒ the same as ͑b͒ except using 0.15% C2H2 as carbon source.
Condition
CH4 ͑%͒
C2H2 ͑%͒
͑a͒
͑b͒
͑c͒
45.0
97.5
9.5
55.0
2.5
90.5
CH4 and H2 stream from the outermost tube and ejected
though a hole of 5 mm diameter to impinge on the substrate.
The distance between the exit hole and substrate is 5 mm.
The diamond growth characteristics employing CH4 and
C2H2 as carbon sources were compared under the following
conditions: carbon mole fraction of 0.3% ͑i.e., 0.15%
C2H2͒, Cl2 mole fraction of 8%, the innermost graphite tube
temperature of 1500 °C, substrate temperature of 600 °C, re-
actor pressure of 5 Torr, and total flow rate of 400 sccm.
Silicon substrates were pretreated for 30 min in an ultrasonic
bath of acetone which contained a 15 ␮m diamond powder
suspension. A 0.5 mm i.d. quartz tube was located on the
substrate surface to evacuate the gas sample through a 6.4
mm o.d. stainless tube ͑1 m long͒ into a mass spectrometer
for analysis.
Mass spectrometry analysis was performed to determine
the carbon species compositions near the growth surface.
Fast radical reactions and recombination inside the sampling
tube prevent the radical species from being detected. Both
C2H2 and CH4 are the major carbon species detected, while
other hydrocarbon species are negligible. In order to estab-
lish its reliability, mass spectrometry analysis was first con-
ducted in a hot-filament CVD reactor using 1.5% CH4 /H2
feed at a reactor pressure of 15 Torr. As listed in the row ͑a͒
of Table I, the mole ratio of CH4 to C2H2 near the surface is
close to 1. The conversion of CH4 into C2H2 is around 60%.
This result agrees very well with that analyzed by molecular-
beam mass spectrometry ͑MBMS͒ reported by McMaster
et al.10 ͑also with hot-filament CVD at a pressure of 20
Torr͒. The MBMS technique is capable of in situ detection of
the stable and radical gas species near the surface by avoid-
ing fast radical reactions during sampling. As a result, the
agreement indicates that the CH4 to C2H2 ratio analyzed by
our mass spectrometry is close to the actual value near the
substrate surface despite the reactions occurring inside the
sampling tube. In the hot-tube system, the CH4 and C2H2
compositions near the surface are listed in the rows ͑b͒ and
͑c͒ of Table I. With CH4 being the input carbon source,
CH4 remains the dominant carbon species near the substrate
surface and C2H2 formed is only 2.5% of the total carbon
species. There is only a 5% conversion of CH4 into C2H2.
Similarly, with C2H2 being the input carbon source, C2H2
remains the dominant carbon species and only 9.5% CH4 is
formed. Also, the conversion of C2H2 into CH4 is only 5%.
Evidently, contrary to the hot-filament CVD results, the
CH4–C2H2 interconversion reaction in our hot-tube system is
insignificant.
FIG. 2. Scanning electron microscopy photographs of diamond depositions
in the hot-tube system using ͑a͒ 0.3% CH4, ͑b͒ 0.15% C2H2, and 8% Cl2, at
a reactor pressure of 5 Torr and total flow rate of 400 sccm.
The minor gas phase reaction from the exit to surface
must be due to a very short residence time. In our hot-tube
system, the gas stream flowing out of the outer graphite tube
exit has a velocity estimated to be 2ϫ104 cm/s, which is
comparable to that in a plasma torch system.13 The Peclet
number is defined as the ratio of convective mass transfer
FIG. 3. The Raman spectrum of the diamond deposit in Fig. 2͑a͒. The peak
position is at 1332cmϪ1
.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
186 Appl. Phys. Lett., Vol. 70, No. 2, 13 January 1997 J.-J. Wu and F. C.-N. Hong
129.120.242.61 On: Wed, 26 Nov 2014 11:32:42

rate to diffusive mass transfer rate and is given by uL/D,
where u is a characteristic flow velocity, L a characteristic
length scale, and D the diffusion coefficient. L is 5 mm and
D is estimated to be 1300 cm2/s ͑at 1000 °C, 5 Torr͒. The
Peclet number is calculated to be on the order of 10, indicat-
ing that convection, instead of diffusion ͑as in a hot-filament
CVD͒, is the dominant mechanism of mass transport. The
residence time in the hot-tube system is 25 ␮s, which is
much smaller than the residence time of 2ϫ103␮s for the gas
to diffuse from the filament to substrate in a typical hot-
filament CVD at 20 Torr.14 Furthermore, the outer graphite
tube temperature was around 1000 °C, much lower than the
filament temperature in a hot-filament CVD, indicating a
much lower average gas temperature in the hot-tube system.
Owing to the lower gas temperature and much shorter resi-
dence time, the extent of gas phase reaction in the hot-tube
system is far less than that in a conventional hot-filament
CVD, resulting in the gaseous carbon species near the sub-
strate surface possessing the characteristic structure of the
input carbon source.
conditions. Martin and Hill3 have shown that the efficiency
of diamond growth from the CH3 radical is about ten times
higher than that for C2H2; however, their film produced from
C2H2 contains a much smaller fraction of diamond phase
than that from CH4.15 Besides, we do not observe the dia-
mond growth from C2H2 on diamond crystallite seeds as re-
ported by Martin.6 More results will be published soon. Our
results tend to be consistent with those by Lee et al.5
In conclusion, due to the small residence time and low
gas temperature in the hot-tube system, the CH4–C2H2 inter-
conversion is insignificant so almost pure CH4 or C2H2 near
the substrate surface can be obtained. Diamond can be syn-
thesized relatively fast with the CH4 dominant carbon spe-
cies near the growth surface; however, diamond growth is
negligible with the C2H2 dominant carbon species near the
growth surface. Under the Cl-rich conditions, the CH3 radi-
cal seems the only diamond growth precursor whereas
C2H2 is inefficient to grow diamond.
The financial support of this work, by the National Sci-
ence Council of the Republic of China under Contract No.
NSC-82-0405-E-006-479, is gratefully acknowledged.
In order to identify the diamond growth precursor, the
conditions under which the carbon species was almost pure
CH4 or C2H2 ͓rows ͑b͒ and ͑c͒, respectively, in Table I͔ were
employed to grow diamond. With 0.3% CH4 being the input
carbon source, an almost continuous diamond film was de-
posited after 2 h growth, as shown in Fig. 2͑a͒. The Raman
spectrum of the deposit in Fig. 2͑a͒ is shown in Fig. 3, and
the 1332 cmϪ1 peak confirms the formation of diamond. On
the other hand, with 0.15% C2H2 being the input carbon
source, only a few very small particles were deposited, as
demonstrated from the very small bright spots in Fig. 2͑b͒.
The above results are very reproducible. The deposition rate
with almost pure C2H2 is negligible compared to that for
almost pure CH4. Since CH4 is less likely to be the diamond
growth precursor, the CH3 radical formed by
CH4ϩH→CH3ϩH2 or CH4ϩCl→CH3ϩHCl should be the
diamond growth precursor as proposed by Harris.1 There-
fore, we conclude that the CH3 radical seems the only dia-
mond growth precursor whereas C2H2 is not under the above
1 S. J. Harris, Appl. Phys. Lett. 56, 2298 ͑1990͒.
2 C. J. Chu, M. P. D’Evelyn, R. H. Hauge, and J. L. Margrave, J. Mater.
Res. 5, 2405 ͑1990͒.
3 L. R. Martin and M. W. Hill, J. Mater. Sci. Lett. 9, 621 ͑1990͒.
4 C. E. Johnson, W. A. Weimer, and F. M. Cerio, J. Mater. Res. 7, 1427
͑1992͒.
5 S. S. Lee, D. W. Minsek, D. J. Vestyck, and P. Chen, Science 263, 1596
͑1994͒.
6 L. R. Martin, J. Mater. Sci. Lett. 12, 246 ͑1993͒.
7 M. Frenklach and K. E. Spear, J. Mater. Res. 3, 133 ͑1988͒.
8 D. G. Goodwin and G. G. Gavillet, J. Appl. Phys. 68, 6393 ͑1990͒.
9 K. L. Menningen, C. J. Erickson, M. A. Childs, L. W. Anderson, and J. E.
Lawler, J. Mater. Res. 10, 1108 ͑1995͒.
10 M. C. McMaster, W. L. Hsu, M. E. Coltrin, and D. S. Dandy, J. Appl.
Phys. 76, 7567 ͑1996͒.
11 J.-J. Wu and F. C.-N. Hong ͑unpublished͒.
12 C. Pan, C. J. Chu, J. L. Margrave, and R. H. Hauge, J. Electrochem. Soc.
141, 3246 ͑1994͒.
13 D. S. Dandy and M. E. Coltrin, J. Mater. Res. 10, 1993 ͑1995͒.
14 D. S. Dandy and M. E. Coltrin, J. Appl. Phys. 76, 3102 ͑1994͒.
15 S. J. Harris and L. R. Martin, J. Mater. Res. 5, 2313 ͑1990͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
Appl. Phys. Lett., Vol. 70, No. 2, 13 January 1997 J.-J. Wu and F. C.-N. Hong 187
129.120.242.61 On: Wed, 26 Nov 2014 11:32:42