Plasma Chemistry and Plasma Processing. Vol. 11, No.

3, 1991

Diamond S y n t h e s i s by D C T h e r m a l P l a s m a CVD
at 1 a t m

Z. P. Lu, ~ L. Stachowicz, ~ P. Kong, ~'2 J. Heberlein, t

and E. Pfender ~
Reeeit, ed Not,ember 20, 1990; revised January _.
~ 1991

Diamond co,stals and films have been succes.sJhllv synthesized by DC thermal

plasma jet CVD at a pressure of 1 arm. A not,el triple torch plasma reactor has
been used to generate a convergent plasma volume to entrain the participating gases.
Three coalescing plasma jets produced by this reactor direct the dissociated and
ionized gaseous species onto (100) silicon wc~iersubstrates where the diamond grows.
In a typical lO-min run, depending on the method ~l'substrate preparation, either
microco,stals with sizes up to 8 txm or continuous,films with thicknesses t~f I-2 txm
have been obtained. X-ray diflt'action, scanning electron microscop.v, and Raman
spectroscop.v ha t,e been used.l~r the characterization t!f the co'stals and t~l'the films.

KEY WORDS: Diamond; plasma synthesis; atmospheric pressure; triple-torch

plasma reactor.

I. I N T R O D U C T I O N
There have been increasing research efforts to produce diamond under
thermodynamic metastable conditions/i-4, Numerous processes have been
reported over the past decade including hot filament CVD, '5~ microwave
plasma-assisted CVD, ~"~ RF plasma-assisted CVD, ~7' arc discharge-assisted
CVD, ~'~ hollow-cathode CVD, ~ and oxygen-acetylene flame CVDJ ~°~
Among the various plasma-assisted chemical vapor deposition processes,
the majority of the research has been conducted under reduced pressures,
i.e., 1-100 tore. Attempts have been made to synthesize diamond at 1 atm '~
and, recently, there have been reports on the thermal plasma synthesis of
diamond films, c7'~-~4~ For example, Matsumoto e t al. '7~ briefly reported

Department of Mechanical Engineering, University of Minnesota, Minneapolis, Minnesota

55455.
: Now with the Idaho National Engineering Laboratory, Idaho Falls, Idaho 83415.
387
0272-4324 ~)1 090tl-I)3S7$Ot~ 511 {I I IC~l Plenum Publishing (orporatlon
388 Lu, Stachowicz, Kong, Heberlein, and Pfender

diamond deposition with an inductively coupled atmospheric-pressure

thermal plasma and Cappelli e t al.' H~ published a more detailed paper on
the subject recently.
Thermal plasmas are by definition in a state of equilibrium or, more
precisely, in local thermodynamic equilibrium (LTE). The high number
densities of species associated with LTE plasmas give rise to high collision
frequencies and, as a result, the heavy particles (atoms and ions) have
essentially the same temperature as the electrons. If such a plasma is rapidly
quenched, as for example at the surface of a cooled substrate, chemical
reactions will be "frozen," i.e., strong deviation from chemical equilibrium
will occur. As a consequence, the hydrogen ions, atoms, '~5-~7' and other
radicals, such as CH3, '~s'~'' which are believed to be responsible for the
successful synthesis of diamond, will have a much higher concentration
than if they would have in a corresponding equilibrium state.

2. E X P E R I M E N T S

Figure la shows the experimental apparatus. The heart of the system

consists of three identical DC plasma arc torches (only two of them are
shown in the schematic). The plasma torches are mounted on the top flange
of the reactor chamber in such a way that the plasma jets of the torches
coalesce, forming a converging plasma volume. A water-cooled gas feeding
probe is located above the temperature valley formed by the three plasma
jets. Details of the equipment can be found in a previous publication. '-~'"
These jets impinge on (100) silicon substrates which are glued onto a
pressurized water-cooled copper substrate holder with a ceramic-base high-
temperature adhesive. In some of the experiments, substrates are scratched
with a 1 # m grit diamond paste for several minutes and cleaned with acetone
in an ultrasonic bath for about 30 rain.
The reactor chamber is evacuated to 0.01 torr and flushed with argon.
The plasma torches are started in a pure argon environment at 1.0atm.
Hydrogen is then gradually added into the plasma torches. Additional
hydrogen and methane are fed through the center feeding probe. The
temperature valley between the plasma jet helps to direct the participating
gases onto substrates. The substrate temperature was roughly monitored by
a small strip of copper glued to the side of the silicon substrates. The
substrates are raised until the copper strip starts melting, indicating a
temperature of 1083°C, which is the melting point of copper. This is, of
course, not a temperature measurement but rather a rough indication.
Usually three silicon wafers are attached to the substrate holder for each
run. The arrangement is shown schematically in Fig. lb, and typical experi-
mental conditions are summarized in Table I.
Diamond Synthesis by Thermal P l a s m a CVD 389

CH4 + H2
I DC Power Supply X 3

~" Injector
Plasma Gas
Ar + H2

Plasma Torch X 3

Viewing Por~

i"
• [-
Coalesced Plasma Jets

Substrate Holder

~SiliconWafer
"'~Exhaust

a O

~ / / ' ~ Substrate Holder

"Copper Strip

Fig. I. (a) S c h e m a t i c of the triple torch p l a s m a reactor; (b) a r r a n g e m e n t of the substrates.

390 Lu, Stachowicz,Kong, Heberlein,and Pfender

Table I. Typical Experimental Conditions

Plasma torches: current 300-350 A, voltage 40-45 V, Ar flow rate 10-20 liters/min,
H~ flow rate 1-2 liters/rain
Process pressure: 1 atm
Gas feeding probe: H 2 flow rate 10 liters/rain, CH4 flow rate 3.5, 5.0, 6.5% of the total
H 2 flow rate
Substrate: (100) silicon scratched with 1 /~m diamond paste, about 1 cm 2 each
Deposition time: 10-15 rain

3. RESULTS AND D I S C U S S I O N
Before termination of the experiment, the methane gas flow is turned
off and the substrate stage is lowered away from the plasma zone to avoid
any deposition of unwanted carbon phases. The films appear grayish and
cover only part of the substrates (marked a in Fig. lb). Diamond films are
produced on an annular ring having a width of 3-6 mm on all three 1-cm 2
silicon wafers (Fig. lb). The produced films have been analyzed by X-ray
diffraction, SEM, and Raman spectroscopy.
Since the X-ray diffractometer has a wider beam than the film width,
two scans have been made. One is for the deposited film and the other for
a plain silicon wafer. The spectrum of the silicon wafer is shown in the
inset to Fig. 2, which shows the result for a test with a C H 4 concentration
of 5%.
Raman spectroscopy analysis has been performed to confirm diamond
formation (Fig. 3). An argon laser (514.5 nm) with a 1 ~m beam size is used
for the excitation. Representative spectra for the 5% methane runs are shown

Silicon Waferbefore Coating

7 ~ " T ' r , , r f
5O 5O ?0 80 g0 100 1t0. I~ t30

.~i-.~Si St ~ , ~ J D SL D

50. 60. 70. 80. 90. 100. 110. 120. 130. 140.

Fig. 2. X-ray diffraction pattern of the CVD diamond.

Diamond Synthesis by Thermal Plasma CVD 391

1332.5 c m - 1

e~ a
5OO0 b

0 • ,
lsoo 1~o 14oo l~o s~oo 1too lo~
Wave Number (cm-l)

Fig. 3. Raman spectrum of the CVD diamond.

in Fig. 3. The two spectra correspond to different locations as indicated in

Fig. lb. In both places, the sharp peak at 1332.5 cm ' confirms the formation
of diamond. Although spectrum a shows a slightly sharper peak than
spectrum b, they both show little formation of graphite.
SEM micrographs are shown in Figs. 4 and 5. Figures 4a, b show some
of the isolated microcrystals grown on a wafer which has been treated only
briefly with diamond paste or on the area b shown in Fig. lb. They show
different morphological features of the diamond single crystal polyhedrons.

• ; ,%'-,,
z
"

0026 2eKV x2,see levm Mo~9 , ." ." : : : ." ." ," .': ' ...~- r, ": :~

Fig. 4. (a, b) Diamond microcrystals; (c, d) partially developed diamond lilms.

392 Lu, Stachowicz, Kong, Heberlein, and Pfender

" . , !g,,
,>., . v,g
,. ~,,Q
®; :_

Fig. 5. (a, b) High nucleation density'along the scratch on the Si wafer; (c) continuousdiamond
lilm; (d) cross section of the continuous diamond film.

Figures 4c, d show film formation. Figures 5a and b show an interesting

phenomenon. The wafer used for that run was, in addition, slightly damaged
by scratching the surface with a sharp-tipped tweezer at several locations
during the sample preparation process. It can be seen that along the scratch
the nucleation sites for diamond particles are much higher than on the
remaining areas. Figure 5c shows a dense continuous film with well-faceted
surfaces. Figure 5d shows the cross section of a typical film with a thickness
of approximately 1.5 p.m. This translates to a deposition rate of 9 p.m/h.
The micrographs correspond to the run with 5% CH4. Similar morphologies
have been observed for the 3.5 and 6.5% CH4 runs.
Diamond is a thermodynamically metastable phase at low and moderate
pressures. Therefore, thermodynamic equilibrium calculations will not
reveal this phase. But such calculations will show which species are formed
for a carbon-hydrogen system under thermodynamic equilibrium condi-
tions. The results apply to regions remote from the fringes of the plasma
jets and from the cold substrate where thermodynamic equilibrium might
prevail. Although such equilibrium calculations have been performed by
several investigators, 's'-~'-24~little results have been reported for atmospheric
pressure conditions. A computer calculation for a thermodynamic equili-
brium carbon-hydrogen system (5% CH4 and 95% H:) at 1 atm is shown
in Fig. 6. The only solid phase presented in this graph is graphite which
Diamond Synthesis by Thermal Plasma CVD 393

103 H
H2
102

e-, 101
o

10°
e-,

E 10
o
CH
E 10
=
.-
i.
e~
10 .3
o.

10 .4

10-s

10-6
l0 -~
;00 1500 2500 3500 4500 5500

Temperature (K)
Fig. 6. Equilibrium composition ofaC-H system; 95% H, and 5%CH4 at latin.

forms in the temperature range from 1100 to 3000 K. The major species in
the system are H , , C H 4 , C (solid), C2H, H, and C (gas). Methane starts to
dissociate at approximately 1000 K and only 1% methane is left at 2000 K.
Hydrogen atoms start to appear at appreciable amounts at 2000 K, and at
4000 K almost all the hydrogen molecules are dissociated. Since the tem-
perature in our plasma exceeds 4000 K, there should be abundant atomic
hydrogen in the boundary layer in front of the substrate.

4. C O N C L U S I O N S
1. Diamond microcrystals and films have been successfully synthesized
in a novel DC plasma reactor at 1 atm.
2. Both the crystals and the films show good crystallinity although the
CH4 concentration used in the investigation is much higher than the con-
centrations used for a microwave plasma CVD process (about 1%).
3. The linear growth rate of the film is approximately 9 ~ m / h .

ACKNOWLEDGMENT
The authors wish to acknowledge Dr. James Butler at the Naval
Research Laboratory for the Raman analysis of some of the samples. This
work has been supported by DOE grant FG02-85ER-13433 and by
Crystalline Materials, Inc.
394 Lu, Stachowicz, Kong, Heberlein, and Pfender

REFERENCES

1. W. A. Yarbrough and R. Messier, Science 247, 688 (1990).

2. K. E. Spear, J. Am. Ceram. Soc. 72, 171 (1989).
3. J. C. Angus, F. A. Buck, M. Sunkara, T. F. Groth, C. C. Hayman, and R. Gat, M R S Ball.,
October, p. 38 (1989).
4. R. C. DeVries, Annu. Rev. Mater. Sci. 17, 167 (1987).
5. S. Matsumoto, Y. Sato, M. Tsutsumi, and N. Setaka, J. Mater. Sci. 17, 3106 (1982).
6. A. R. Badzian, T. Badzian, R. Roy, R. Messier, and K. E. Spear, Mater Res. Bull. 23, 531
(1988).
7. S. Matsumoto, M. Hino, and T. Kobayashi, AppL Phys. Lett. 51,737 (1987).
8. K. Suzuki, A. Sawabe, and T. lnuzuka, Jpn. J. AppL Phys. 29, 153 (1990).
9. B. Singh, O. R. Mesker, A. W. Levine, and Y. Arie, Appl. Phys. Lett. 52, 1658 (1988).
10. Y. Tzeng, C. Cutshaw, R. Phillips, T. Srivinyunon, A. Ibrahim, and B. H. Loo, Appl. Phys.
Lett. 56, 134 (1990).
11. N. Ohtake and M. Yoshikawa, J. Electrochem. Soc. 137, 717 (1990).
12. K. Kurihara, K. Sasaki, M. Kawarada, and N. Koshino, Appl. Phys. Lett. 52, 437 (1988).
13. S. Matsumoto, "'Deposition of Diamond from Thermal Plasma," Diamond and Diamond-
Like Materials Synthesis, M R S Extended Abstracts, 1988, G. H. Johnson, A. R. Badzian,
and M. W. Gels, eds., (Materials Research Society, Pittsburgh, Pennsylvania, 1988), p. 119.
14. M. A. Cappelli, T. G. Owano, and C. H. Kruger, J. Mater. Res. 5, 2326 (1990).
15. T. R. Anthony, Vacuum 41, 1356 (1990).
16. F. G. Celii and J. E. Butler, Appl. Phys. Lett. 54, 1031 (1989).
17. J. A. Mucha, D. L. Flamm, and D. E. lbbotson, J. Appl. Phys. 65, 3448 (1989).
18. H. Sugai, H. Kojima, A. Ishida, and H. Toyoda, AppL Phys. Lett. 56, 2616 (1990).
19. S. J. Harris and A. M. Weiner, ./. AppL Phys. 67, 6520 (1990).
20. Z. P. Lu and E. Pfender, "'Synthesis of AIN Powder in a Triple Torch Plasma System,"
Proc. of the 9th Int. Syrup. on Plasma Chem., Pugnochiuso, Italy, 1989, Vol. 2, R. d'Agostino,
ed., (University of Bari, Italy, 1989), p. 675.
21. Y. Mitsuda, T. Yoshida, and K. Akashi, "'The Growth of Diamond in MW Plasma under
Low Pressure," Proc. o/'the 8th Int. Symp. on Plasma Chem., Tokyo, Japan, 1987, Vol. 4,
K. Akashi and A. Kinbara, eds., (LOC, Tokyo, 1987), p. 2469.
22. P. K. Bachmann and H. Lydtin, "Diamond Forming Discharge," Mater. Res. Soc. Syrup.
Proc., Vol. 165, Characteri:ation of Plasma-Enhanced CVD Processes, Boston,
Massachusetts, 1989, G. Lucovsky, D. E. lbbotson, and D. W. Hess, eds., (Materials
Research Society, Pittsburgh, Pennsylvania, 1989), p. 181.
23. B. V. Spitsyn, 3". Crvst. Growth 99, 1162 (1990).
24. W. Piekarczyk, R. Messier, R. Roy, and C. Engdahl, d. Cryst. Growth 106, 279 (1990).