We developed a new method of GaN growth using Radical-Enhanced Metalorganic Chemical Vapor Deposition (REMOCVD) technology by which Gallium Nitride (GaN) grows at low temperatures without ammonia gas. In this method, we investigated the effect of N2/H2 plasma on the GaN substrate surface cleaning prior to the growth of homoepitaxial GaN. In-situ reflection high-energy electron diffraction (RHEED) and atomic force microscope (AFM) were used to investigate the surface morphology of the cleaned GaN substrates. The interface between GaN substrate and homoepitaxially grown GaN by REMOCVD was evaluated by transmission electron microscope and the crystal quality was evaluated by X-ray diffraction. The in-situ N2/H2 plasma cleaning at 600 °C shows a smooth surface morphology with streak diffraction lines observed by RHEED. Since the homoepitaxial growth of GaN was performed at 800 °C, the cleaned GaN substrate temperature was ramped up from 600 °C to 800 °C with and without plasma exposure to compare the effect of plasma. Homoepitaxially grown GaN on GaN substrates whose temperature was ramped up with plasma exposure showed good crystal quality with no threading dislocations at the interface. It was found that N2/H2 plasma plays a significant role in the GaN surface cleaning for good quality crystal growth.

III-V nitride semiconductors draw a lot of attention because of its wide band gap and its potential application in blue-ultraviolet light emitters, and high electron mobility transistors (HEMT’s).1,2 Most of the commercially available gallium nitride (GaN) based devices are heteroepitaxially grown on foreign substrates such as sapphire, SiC or Si due to the limited availability of bulk GaN crystals.3,4 Recently, bulk GaN crystals are commercially available and devices like light-emitting diode (LED),5,6 Schottky diodes,7 HEMTs8 grown on bulk GaN substrates were demonstrated. In homoepitaxial growth, the substrate surface and the interface between the subsequently grown layers play a significant role in the structural and electronic properties of the devices. It is therefore a key factor to establish an effective and efficient method of GaN substrate surface cleaning, resulting in high quality homoepitaxial growth. To reduce the surface contamination, various cleaning methods such as ex-situ solvent cleaning9 together with in-situ cleaning like thermal annealing in ammonia10 or thermal degassing in ultra-high vacuum11 were studied in the past. Radio Frequency (RF) plasma treatment of GaN surfaces by N2,12,13 H2,14 BCl3/Cl2,15,16 N2/Ar217 were also studied at elevated temperatures.

RF N2-plasma treatment at room temperature and at low temperatures can reduce the surface oxide whereas the effective removal of surface contamination takes place at a much higher temperature (950 °C). Prolonged N2 plasma exposure however results in surface damage and loss of step definition.12,18,19 The combination of thermal degassing and RF N2 plasma were also studied for the removal of surface oxide. There was prolonged thermal degassing and subsequent plasma assisted cleaning starting at low temperatures, followed by a short period of plasma cleaning at high temperatures that seem to effect the removal of surface oxide.12 In many of the previous studies, the cleaning was performed at lower temperatures < 700 °C whereas the GaN growth takes place at higher temperatures > 1000 °C. A detailed understanding on the effect of the temperature ramp up towards the cleaned GaN substrate prior to the growth is missing and it is necessary to study its effect.

In this study, we have investigated the effect of N2/H2 RF plasma exposure during temperature ramp-up of the substrate prior to GaN growth by radical-enhanced metalorganic chemical vapor deposition (REMOCVD). Surface treatment of GaN by N2/H2 plasma was evaluated by field-emission scanning electron microscope (FE-SEM), atomic force microscope (AFM), and in-situ reflection high energy electron diffraction (RHEED). The interface between the GaN substrate and the grown GaN was evaluated by transmission electron microscope (TEM) and the crystal quality was evaluated by X-ray diffraction (XRD).

The schematic diagram of our newly developed REMOCVD system is shown elsewhere.20 The system consists of a vacuum chamber, in which a showerhead is equipped for a gas introduction. The showerhead also plays a role as a top electrode of capacitively coupled plasma (CCP), which contacts to a very high frequency (100 MHz) power source through a matching box. Below the showerhead distanced at 150 mm, a sample susceptor is equipped, which is used as the ground electrode. The exhaust lines are connected to a turbo molecular drag pump and the background pressure was around 10-4 Pa. An N2 and H2 mixture gas with the flow rate of 750 sccm for N2 and 250 sccm for H2 was introduced through the showerhead of the top electrode. The pressure was maintained to be 300 Pa by an automatic pressure controller equipped to the exhaust line. Trimethylgallium (TMG) was used as a precursor with a flow rate of 0.1 sccm, which was controlled by a needle valve. Ammonia species (NHx) and reactive nitrogen species (N-radicals), including excited N2 and N atoms, were generated in the plasma discharge region. The decomposition of TMG molecules occur thermally near the vicinity of the substrate, far from the shower head electrode. More details of the developed REMOCVD system and effect of plasma with TMG have been studied and reported already.20 

As the first step, the susceptor inside the chamber was cleaned at 900 °C, RF 400 W, for 20-30 minutes, to remove any contamination inside the chamber. Then, substrates were mounted on a carbon susceptor that was monitored by a thermocouple for controlling the substrate temperature. GaN growth was carried out on commercially available bulk GaN substrates (NGK Co. Ltd.) prepared by the flux method. Prior to the growth, a wet cleaning procedure was carried out. The substrate was ultrasonically cleaned in an acetone bath, and in an iso-propanol bath for 5 min and was then rinsed with deionized (DI) water. The substrate was then dipped in 5% hydrofluoric acid for 5 minutes and rinsed with deionized (DI) water. Finally, the substrate was dried using a N2 gun and loaded into the chamber for subsequent in-situ cleaning. Inside the chamber, N2/H2 plasma cleaning of the substrate was carried out by varying the (i) substrate temperature and (ii) RF plasma power to obtain a smooth surface morphology. After plasma cleaning, the substrate temperature was ramped up to 800 °C at which GaN was grown for 6 hrs with a N2/H2 gas mixture of 750/250 sccm, under the chamber pressure of 300 Pa and with the plasma power of 700 W.

To understand the effect of N2/H2 plasma on GaN substrate surface morphology, the substrate temperature and the RF power were varied and evaluated by FESEM and AFM. When the temperature was increased from 600 °C to 900 °C with RF power of 400 W for 5- 10 minutes, it was found that, the GaN surface was greatly damaged beyond 600 °C. The increase of RF power beyond 400 W at 600 °C was also found to damage the GaN surface. From this study, the plasma cleaning condition was performed at 600 °C with 400W for 5-10 mins under which the GaN surface showed smooth morphology as observed by SEM and strong streaky diffraction lines were observed by RHEED. After this optimized plasma cleaning, the substrate temperature was raised to 800 °C with and without plasma exposure and the surface property was evaluated by AFM and RHEED as shown in Fig. 1. Figures 1(a) and (c) show the AFM and RHEED images of a GaN substrate ramped up without plasma treatment (sample-A) and Fig. 1 (b) and (d) show the GaN substrate ramped up with the plasma power of 200 W (sample-B). For sample A, a very rough surface with the RMS value of 4.018 nm and the white spots represents Ga which was confirmed by the EDX. The RHEED shows a spotty pattern because of the rough surface. For sample-B, AFM shows a smooth morphology with the RMS value of 0.220 nm and RHEED patterns show strong streak lines which are the same as after plasma cleaning. Heating of the substrates without plasma exposure may activate the separation of Ga-N bonds resulting in N-out diffusion. While under nitrogen plasma exposure, N-out diffusion may be suppressed.

FIG. 1.

Growth temperature ramp-up from 600 °C to 800 °C without plasma and with plasma exposure. AFM shows (a) rms of 4.018 nm, (b) rms of 0.220 nm, RHEED shows (c) spotty pattern, and (d) streak lines respectively.

FIG. 1.

Growth temperature ramp-up from 600 °C to 800 °C without plasma and with plasma exposure. AFM shows (a) rms of 4.018 nm, (b) rms of 0.220 nm, RHEED shows (c) spotty pattern, and (d) streak lines respectively.

Close modal

GaN was then grown on sample-A and sample-B and their interface was investigated by cross-sectional TEM as shown in Fig. 2. The total thickness of the grown epilayer was 700 and 650 nm respectively for sample -A and sample -B. The growth rate of GaN was found to be 116 and 108 nm/hr for sample -A and sample-B, respectively. The bright field TEM revealed a rough interface with the generation of additional threading dislocations for GaN grown on sample-A as shown in Fig. 2(a). For GaN grown on sample-B, uniform growth was performed so that it was difficult to recognize the interface between the substrate and the grown GaN. No additional threading dislocation was observed for sample B.

FIG. 2.

Cross-sectional TEM images with bright field. The GaN grown on substrate (a) ramped up without plasma and (b) ramped up with plasma exposure. Yellow dotted lines are the interface between the grown GaN and the GaN substrate.

FIG. 2.

Cross-sectional TEM images with bright field. The GaN grown on substrate (a) ramped up without plasma and (b) ramped up with plasma exposure. Yellow dotted lines are the interface between the grown GaN and the GaN substrate.

Close modal

XRD measurements were performed for evaluating the structural crystal quality of the homoepitaxially grown GaN. Fig. 3 shows the omega scan of GaN (0002) plane. It shows a full width half maximum (FWHM) of GaN grown on sample-A (black line) and sample-B (red line). The FWHM rocking curve of the bulk GaN substrate was 100 arcsec and the FWHM of homoepitaxially grown GaN on sample-A was 210 arcsec and that of sample-B was 122 arcsec. The crystal quality was greatly improved for sample-B. Hence, it can be concluded that the application of plasma power during temperature ramp-up to the growth temperature seems to be due to the prevention of N-out diffusion. RF Plasma power treatment is therefore essential for the homoepitaxial growth of GaN.

FIG. 3.

XRC-FWHM of the GaN grown on substrate (a) ramped up without plasma, and (b) ramped up with plasma exposure.

FIG. 3.

XRC-FWHM of the GaN grown on substrate (a) ramped up without plasma, and (b) ramped up with plasma exposure.

Close modal

In conclusion, GaN thin films of smooth and flat surface morphology were obtained by our newly developed REMOCVD system. This new growth method has the advantage of being able to grow GaN at a low temperature of 800oC, compared with the conventional MOCVD method which requires a high temperature (>1000°C) growth. After plasma cleaning, the GaN substrate temperature ramp up without plasma exposure resulted in rough and spotty patterns, which may be due to N-out diffusion. However, GaN substrate temperature ramp-up with plasma exposure showed smooth surface morphology and RHEED streak patterns, which are suitable for the homoepitaxial GaN growth. The smooth interface was observed by TEM, and high quality GaN with the XRC-FWHM of 122 arcsec was obtained for plasma cleaned GaN substrate. N2/H2 plasma plays a significant role in GaN surface cleaning and temperature ramp-up prior to GaN growth and was found to be essential for the high quality homoepitaxial GaN.

This work was partly supported by the Knowledge Cluster Initiative of the Ministry of Education, Culture, Sports, Science and Technology. The authors thank NGK Co. Ltd. who supplied bulk GaN substrates for this study.

1.
M. A.
Khan
,
J. W.
Yang
,
W.
Knap
,
E.
Frayssinet
,
X.
Hu
,
G.
Simin
,
P.
Prystawko
,
M.
Leszczynski
,
I.
Grzegory
,
S.
Porowski
,
R.
Gaska
,
M. S.
Shur
,
B.
Beaumont
,
M.
Teisseire
, and
G.
Neu
,
Appl. Phys. Lett.
76
,
3807
(
2000
).
2.
S. J.
Pearton
,
J. C.
Zolper
,
R. J.
Shul
, and
F.
Ren
,
J. Appl. Phys.
86
,
1
(
1999
).
3.
D. F.
Storm
,
M. T.
Hardy
,
D. S.
Katzer
,
N.
Nepal
,
B. P.
Downey
,
D. J.
Meyer
,
T. O.
McConkie
,
L.
Zhou
, and
D. J.
Smith
,
J. Cryst. Growth
456
,
121
(
2016
).
4.
D. F.
Storm
,
T. O.
McConkie
,
M. T.
Hardy
,
D. S.
Katzer
,
N.
Nepal
,
D. J.
Meyer
, and
D. J.
Smith
,
J. Vac. Sci. Technol. B, Nanotechnol. Microelectron. Mater. Process. Meas. Phenom.
35
,
02B109
(
2017
).
5.
K.
Okamoto
,
H.
Ohta
,
D.
Nakagawa
,
M.
Sonobe
,
J.
Ichihara
, and
H.
Takasu
,
Japanese J. Appl. Physics, Part 2 Lett.
45
(
2006
).
6.
A.
Chakraborty
,
B. A.
Haskell
,
S.
Keller
,
J. S.
Speck
,
S. P.
Denbaars
,
S.
Nakamura
, and
U. K.
Mishra
,
Jpn. J. Appl. Phys. Part 2 Lett.
44
(
2005
).
7.
Y.
Wang
,
S.
Alur
,
Y.
Sharma
,
F.
Tong
,
R.
Thapa
,
P.
Gartland
,
T.
Issacs-Smith
,
C.
Ahyi
,
J.
Williams
,
M.
Park
,
M.
Johnson
,
T.
Paskova
,
E. A.
Preble
, and
K. R.
Evans
,
Semicond. Sci. Technol.
26
(
2011
).
8.
E.
Frayssinet
,
W.
Knap
,
P.
Lorenzini
,
N.
Grandjean
,
J.
Massies
,
C.
Skierbiszewski
,
T.
Suski
,
I.
Grzegory
,
S.
Porowski
,
G.
Simin
,
X.
Hu
,
M. A.
Khan
,
M. S.
Shur
,
R.
Gaska
, and
D.
Maude
,
Appl. Phys. Lett.
77
,
2551
(
2000
).
9.
L. L.
Smith
,
S. W.
King
,
R. J.
Nemanich
, and
R. F.
Davis
,
J. Electron. Mater.
25
,
805
(
1996
).
10.
S. W.
King
,
J. P.
Barnak
,
M. D.
Bremser
,
K. M.
Tracy
,
C.
Ronning
,
R. F.
Davis
, and
R. J.
Nemanich
,
J. Appl. Phys.
84
,
5248
(
1998
).
11.
V.
Bermudez
,
D.
Koleske
, and
A.
Wickenden
,
Appl. Surf. Sci.
126
,
69
(
1998
).
12.
T.
Hashizume
,
S.
Ootomo
,
S.
Oyama
,
M.
Konishi
, and
H.
Hasegawa
,
J. Vac. Sci. Technol. B Microelectron. Nanom. Struct.
19
,
1675
(
2001
).
13.
R.
Kawakami
,
M.
Niibe
,
Y.
Nakano
,
S.
Chen
,
H.
Kondo
,
K.
Ishikawa
,
R.
Kometani
,
K.
Takeda
,
M.
Sekine
, and
M.
Hori
,
Appl. Phys. Express
6
,
056201
(
2013
).
14.
T.
Inagaki
,
T.
Hashizume
, and
H.
Hasegawa
,
Appl. Surf. Sci.
216
,
519
(
2003
).
15.
H. W.
Choi
,
S. J.
Chua
,
A.
Raman
,
J. S.
Pan
, and
A. T. S.
Wee
,
Appl. Phys. Lett.
77
,
1795
(
2000
).
16.
L. K.
Li
,
L. S.
Tan
, and
E. F.
Chor
,
J. Cryst. Growth
268
,
499
(
2004
).
17.
R.
Kometani
,
K.
Ishikawa
,
K.
Takeda
,
H.
Kondo
,
M.
Sekine
, and
M.
Hori
,
Appl. Phys. Express
6
(
2013
).
18.
S.
Gangopadhyay
,
T.
Schmidt
,
C.
Kruse
,
S.
Figge
,
D.
Hommel
, and
J.
Falta
,
J. Vac. Sci. Technol. A Vacuum, Surfaces, Film.
32
,
051401
(
2014
).
19.
K. N.
Lee
,
S. M.
Donovan
,
B.
Gila
,
M.
Overberg
,
J. D.
Mackenzie
,
C. R.
Abernathy
, and
R. G.
Wilson
,
J. Electrochem. Soc.
147
,
3087
(
2000
).
20.
Y.
Lu
,
H.
Kondo
,
K.
Ishikawa
,
O.
Oda
,
K.
Takeda
,
M.
Sekine
,
H.
Amano
, and
M.
Hori
,
J. Cryst. Growth
391
,
97
(
2014
).