Journal of the Korean Physical Society, Vol. 39, No. 4, October 2001, pp.

617621

High Density Plasma Etching of GaN Films in Cl2 /Ar Discharges

with a Low-Frequency-Excited DC Bias
Y. H. Im, C. S. Choi and Y. B. Hahn
School of Chemical Engineering and Technology, Semiconductor Physics Research Center,
Chonbuk National University, Chonju 561-756
(Received 10 April 2001)
Gallium nitride films (undoped, n-type and p-type GaN) were etched in inductively coupled
Cl2 /Ar plasmas. To control the ion energy, we applied a dc bias voltage excited by low frequency
of 100 kHz to the susceptor electrode. Etch rates were strongly affected by the inductively coupled
plasma (ICP) source power, the rf chuck power, the reactor pressure and the etch gas concentration.
Fast etch rates of 5300 9300
A/min, depending on materials, were obtained at 700 W ICP, 150
W rf, 40 mTorr and 25 % Cl2 . The surface roughness was relatively independent of the chuck
power up to 200 W and showed quite smooth morphology (rms roughness of 1.1 1.3 nm). The
stoichiometry was maintained at the etched surfaces of the undoped and the p-type GaN films, but
the n-type GaN showed some depletion of nitrogen from the surface.

I. INTRODUCTION

CH4 //H2 /Ar and CH4 //H2 /N2 chemistries; they obtained etch rates of 2500
A/min; the rate was approximately two times faster in CH4 //H2 /Ar [18]. Hahn et
al. first reported ICl- and IBr-based plasma chemistries
for ICP etching of GaN, InN and AlN [19,20]. They obtained a maximum selectivities of 30 for InN/AlN and
14 for InN/GaN. They also reported the effect of additive noble gases (He, Ar, Xe) on Cl2 -based ICP etching
of GaN, InN, and AlN [21,22], and they concluded that
efficient breaking of the III-nitrogen bond is crucial for
higher etch rates. It is worthwhile to note that all the
investigators mentioned above used 2 MHz for the chuck
power. Much less work has been done on the ICP etching
of doped GaN films in Cl2 /Ar discharges, especially in
terms of a low frequency dc-bias chuck power to control
the ion energy.
In this paper, we report a parametric study of ICP
etching of undoped, n-type and p-type GaN films with a
low frequency, 100 KHz, rf power applied to the sample
chuck. The effects of etch gas concentration, ICP source
power, rf chuck power and reactor pressure on etch rates
were investigated. Practically fast etch rates were obtained at moderate ICP conditions.

The main applications of the III-nitrides are photonic

devices such as laser diodes and light-emitting diodes
(LEDs) [112]. Currently, all of the LEDs and a majority
of the lasers are ridge wave-guide structures in which the
mesas are formed by dry etching [13]. Hence, most of the
previous etching studies have been focused on obtaining
the relatively large etch depths (2 4 m) typical of
ridge or facet heights, where the effect of the final surface
morphology on the field is less important. Obtaining
a fast etch rate requires an ion-enhanced reactive dry
etch technique. The most significant advancement in dry
etching of the III-nitrides has been the utilization of highdensity plasma sources.
In recent years, the majority of high-density plasma
etchings have been performed using inductively coupled
plasma (ICP) etch systems because of their superior uniformity and control as well as their lower cost of ownership [1416]. One of the main advantages of ICP systems
is independen control of the bulk plasma density and the
ion energy. For control of the ion energy, which enhances
physical sputtering and, thus, anisotropic etching, 2 MHz
or 13.56 MHz is generally used for the radio frequency
(rf) chuck power.
Shul et al. first reported the ICP etching of GaN
in Cl2 /H2 /Ar ICP-generated plasmas with etch rates as
high as 6900
A/min [15,17]. Vartuli et al. were the
first to report ICP etching of GaN, InN, and AlN with

II. EXPERIMENTAL
The GaN films were grown at 1100 C on Al2 O3 substrates by using metal organic chemical vapor deposition
(MOCVD). The total layer thickness was 1.5 m, and the
carrier concentrations were 1.0 1017 /cm3 for n-GaN (Si
doped) and 7.9 1016 /cm3 for p-GaN (Mg doped).

E-mail: ybhahn@moak.chonbuk.ac.kr.
-617-

-618-

Journal of the Korean Physical Society, Vol. 39, No. 4, October 2001

Fig. 1. Effect of chlorine concentration on the etch rates

of undoped, n- and p-type GaN films (700 W ICP, 150 W rf
and 40 mTorr).

Fig. 3. Effect of ICP source power on the etch rates of

undoped, n- and p-type GaN films (150 W rf, 40 mTorr and
25 % Cl2 ).

Etching was performed in a planar-type Vacuum Science ICP system (VSICP-1250A), in which the ICP
source operates at 13.56 MHz. A detailed description
of the etch system is available elsewhere [23]. The temperature of the back-side cooled chuck was held at 25

C. The ion energy was controlled by using the applied

rf chuck power having a low frequency of 100 kHz. The
Cl2 /Ar mixture with total gas loads of 20 standard cubic
centimeters per minute (sccm) was injected into the reactor through electronic mass flow controllers (MFCs). For
the etch rate experiments, the samples were masked with
Apiezon wax; etch depths were obtained from stylus profilometry measurements of the etched samples after removal of the wax. Surface morphology and near-surface
chemistry were examined using atomic force microscopy
(AFM) operating in the tapping mode with a Si tip and
Auger electron spectroscopy (AES), respectively.

III. RESULTS AND DISCUSSION

Fig. 2. Effect of reactor pressure on the etch rates of

undoped, n- and p-type GaN films (700 W ICP, 150 W rf
and 25 % Cl2 ).

Figure 1 shows the effect of the Cl2 concentration of

the feed gas on the etch rate and the dc bias at 700 W
ICP, 150 W rf and 40 mTorr. The etch rates of undoped, n- and p-GaN films decreased with Cl2 concentration
beyond 50 %, while the dc bias increased overall with
increasing etch gas concentration. The increase in the
dc bias can be explained by the fact that compared to
pure Ar discharges, additional collisional energy losses
are present with increasing Cl2 content, resulting in less
production of ions. A lower ion flux, thus, induces lower
etch rates at higher Cl2 concentrations.
The effect of the reactor pressure on the etch rate and
the dc bias was examined by varying the pressure from 10
to 50 mTorr at 700 W source power and 150 W rf chuck
power. The flow rate of the etch gas was held constant at
5 sccm Cl2 and 15 sccm Ar. Figure 2 shows that the etch
rates of n-GaN increase substantially with pressure, but
all materials show overall a similar trend. The increase
in the etch rate with pressure might be attributed to an
increase in the concentration of neutrals with increasing
pressure and to the longer mean free paths of the ions due
to the low frequency of 100 kHz. Figure 2 also shows a
substantial increase in the dc bias voltage with increasing
reactor pressure. The increase in the dc bias voltage (or
decrease in the ion density) with pressure is probably due
to higher recombination rates of ions and electrons in the
plasma and, thus, to a lower number of ions arriving at
the bottom electrode.
The effects of the ICP source power on the etch rates
of undoped, n- and p-type gallium nitrides and on the
dc-bias voltage are shown in Fig. 3. During these experiments the reactor pressure, the rf chuck power, and
the gas feed rate were held constant at 40 mTorr, 150
W, and 5 sccm Cl2 /15 sccm Ar, respectively. The etch
rates increased with the ICP source power while the dc

High Density Plasma Etching of GaN Films in Cl2 /Ar Discharges Y. H. Im et al.

-619-

Fig. 4. Effect of rf chuck power on the etch rates of undoped, n- and p-type GaN films (700 W ICP, 40 mTorr and
25 % Cl2 ).

bias decreased. The highest etch rates were obtained at

a moderately high ICP power of 700 W: 9300
A/min
for n-GaN, 5300
A/min for p-GaN, and 7100
A/min
for undoped GaN. The decrease in the dc biase with the
ICP source power was mainly due to the increased ion
density at higher ICP source powers.
Figure 4 shows the effects of rf chuck power on the
etch rates and the dc bias. The etch rates and the dc
bias voltages increased with the rf chuck power or ionbombarding energy, typical results for an ICP etching
system [1621,23]. The increase in the etch rate with
the chuck power can be explained by enhanced sputter
desorption of etch products as the ion energy increases.
The surface morphologies of etched n-type and undoped GaN films were examined using AFM. Figure 5
shows the effect of the reactor pressure on the surface
morphology of the n-GaN etched at 25 % Cl2 (5 sccm
Cl2 /15 sccm Ar), 700 W ICP and 150 W rf chuck pow-

Fig. 5. AFM images of the surface of n-GaN films etched

varying pressure (700 W ICP, 150 W rf and 25 % Cl2 ).

Fig. 6. AFM images of the surface of undoped-GaN (top)

and p-GaN (bottom) films etched varying pressure (150 W
rf, 40 mTorr and 25 % Cl2 ).

ers. The surface morphology in terms of the root mean

square (rms) roughness was between 0.71 and 1.54 nm,
depending on the reactor pressure. P-GaN also showed
similar results. The effect of the ICP source power on the
surface morphology of etched undoped- and p-GaN films
was also investigated, and the results are shown in Fig.
6, which shows little dependence on the source power,
except at higher powers (>700 W). The increased roughness at higher ICP powers is believed to be mainly due to
an increase in the incident ion flux toward the substrate
surface. Although not illustrated, AFM images showed
that the surface morphology of undoped GaN was relatively independent of the chuck power up to 200 W,
resulting in a fairly smooth morphology (rms roughness
of 1.1 1.3 nm). Figure 7 presents the SEM images of
the etched profiles of undoped GaN at 50 and 200 W rf
powers, and shows quite smooth surfaces.
Figure 8 shows the AES surface scan from unetched
control and p- and n-type GaN films after etching with 5
sccm Cl2 /15 sccm Ar discharges at 700 W ICP, 150 W rf
and 40 mTorr. There is oxygen present from the native
oxide that grew on the samples during the transfer from
the reactor to the AES chamber. Carbon is also present
from the same exposure to ambient air. The surface scan
results are very similar to those for the unetched control
sample, indicating that stoichiometry at the surface was
maintained and, thus, equirate removal of etch products
was obtained. However, n-GaN showed some depletion
of nitrogen from the surface, which may actually reduce
the contact resistance when an n-type ohmic contact is
deposited. It is also interesting to see that the oxygen
peak intensity was substantially decreased after the ICP
etching, indicating removal of oxides form the surface.

-620-

Journal of the Korean Physical Society, Vol. 39, No. 4, October 2001

Fig. 7. SEM images of the etch profiles of undoped GaN

films etched at 700 W ICP, 40 mTorr and 25 % Cl2 under
different rf chuck powers: 50 W (top) and 200 W (bottom).

IV. SUMMARY AND CONCLUSIONS

High density plasma etching of undoped, n- and ptype GaN films with Cl2 /Ar discharges in a planar type
reactor was carried out with low-frequency (100 kHz)
dc biases applied to the sample chuck. The etch rates
of the GaN films decreased with increasing Cl2 concentration beyond 50 % while the dc bias increased overall
with increasing Cl2 content. The etch rate also increased
substantially with increasing pressure, mainly because
of increased neutral concentration and longer mean free
paths of ions at low frequency. The highest etch rates
were obtained at a moderately high ICP power of 700 W:
9300
A/min for n-GaN, 5300
A/min for p-GaN and
7100
A/min for undoped GaN. Although all the GaN
films showed some dependence on etch conditions, overall they showed similar etch characteristics. The etched
surface was quite smooth in terms of rms roughness. The
stoichiometry at the etched surface was maintained, indicating equirate removal of etch products, and n-GaN
showed some depletion of nitrogen from the surface.
ACKNOWLEDGMENTS

Fig. 8. AES surface scans from the unetched control and

the p- and n-GaN films etched at 700 W ICP, 150 W rf , 40
mTorr and 25 % Cl2 .

This work was supported by Grant No. 2000-2-30700005-3 from the Basic Research Program of the Korea
Science and Engineering Foundation (KOSEF).
REFERENCES
[1] R. J. Shul, G. B. McClellan, S. A. Casalnuova, D. J.
Rieger, S. J. Pearton, C. Constantine, C. Barrat, R. F.
Karlicek, Jr., C. Tran and M. Schurman, Appl. Phys.
Lett. 69, 1119 (1996).
[2] R. J. Shul, GaN and Related Materials, edited by S. J.
Pearton (Gordon and Breach, N.Y., 1997).
[3] H. P. Gillis, D. A. Choutov and K. P. Martin, JOM 48,
50 (1996).
[4] I. Adesida, A. Mahajan, E. Andideh, M. A. Khan, D.
T. Olsen and J. N. Kuznia, Appl. Phys. Lett. 63, 2777
(1993).
[5] M. E. Lin, Z. F. Fan, Z. Ma, L. H. Allen and H. Morkoc,
Appl. Phys. Lett. 64, 887 (1994).
[6] H. Lee, D. B. Oberman and J. S. Harris, Jr., Appl. Phys.
Lett. 67, 1754 (1995).
[7] W. Pletschen, R. Niegurch and K. H. Bachem, in Proc.

High Density Plasma Etching of GaN Films in Cl2 /Ar Discharges Y. H. Im et al.

[8]
[9]
[10]

[11]

[12]

[13]
[14]

[15]

Symp. Wide Bandgap Semiconductors and Devices (Electrochemcal Society, Pennington, N. J., 1995), Vol. 95-21,
p. 241.
S. J. Pearton, C. R. Abernathy and F. Ren, Appl. Phys.
Lett. 64, 2294 (1994).
L. Zhang, J. Ramer, J. Brown, K. Zhang, L. F. Lester
and S. D. Hersee, Appl. Phys. Lett. 68, 367 (1996).
R. J. Shul, R. D. Briggs, S. J. Pearton, C. B. Vartuli, C.
R. Abernathy, J. W. Lee, C. Constantine and C. Barratt,
Mat. Res. Soc. Symp. Proc. 449, 969 (1997).
H. Cho, C. B. Vartuli, S. M. Donovan, C. R. Abernathy,
S. J. Pearton, R. J. Shul and C. Constantine, J. Vac. Sci.
Technol. A 16, 1631 (1998).
H. Cho, C. B. Vartuli, S. M. Donovan, K. D. Mackenzie, C. R. Abernathy, S. J. Pearton, R. J. Shul and C.
Constantine, J. Electron. Mat. 27, 166 (1998).
S. Nakamura, GaN and Related Materials, edited by S.
J. Pearton (Gordon and Breach, N.Y.,1997).
S. A. Smith, C. A. Wolden, M. D. Bremser, A. D. Hanser,
R. F. Davis and W. V. Lampert, Appl. Phys. Lett. 71,
3631 (1997).
R. J. Shul, R. D. Briggs, S. J. Pearton, C. B. Vartuli, C.
R. Abernathy, J. W. Lee, C. Constantine and C. Barratt,

-621-

Mat. Res. Soc. Symp. Proc. 449, 969 (1997).

[16] Y. B. Hahn and S. J. Pearton, Korean J. Chem. Eng.
17, 304 (2000).
[17] R. J. Shul, G. B. McClellan, S. J. Pearton, C. R. Abernathy, C. Constantine and C. Barratt, Electron. Lett.
32, 1408 (1996).
[18] C. B. Vartuli, S. J. Pearton, J. W. Lee, J. D. Mackenzie, C. R. Abernathy, R. J. Shul, C. Constantine and C.
Barratt, J. Electrochem. Soc. 144, 2844 (1997).
[19] Y. B. Hahn, D. C. Hays, H. Cho, K. B. Jung, C. R.
Abernathy, S. M. Donovan, S. J. Pearton, J. Han and R.
J. Shul, Mat. Sci. Eng. B 60, 95 (1999).
[20] D. C. Hays, A. Cho, K. B. Jung, Y. B. Hahn, C. R.
Abernathy and S. J. Pearton, Appl. Surf. Sci. 147, 134
(1999).
[21] Y. B. Hahn, D. C. Hays, S. M. Donovan, C. R. Abernathy, J. Han, R. J. Shul, H. Cho, K. B. Jung and S. J.
Peaton, J. Vac. Sci. Technol. A 17, 768 (1999).
[22] H. Cho, Y. B. Hahn, D. C. Hays, K. B. Jung, S. M. Donovan, C. R. Abernathy, S. J. Pearton and R. J. Shul, MRS
Internet J. Nitride Semicond. Res. 4S1, G6.56 (1999).
[23] B.-C. Cho, Y.-H. Im, J.-S. Park, T. Jeong and Y. B.
Hahn, J. Korean Phys. Soc. 37, 23 (2000).