Enhanced GaN decomposition in H2 near atmospheric pressures

Article abstract of DOI:10.1063/1.122354

GaN decomposition is studied at metallorganic vapor phase epitaxy pressures (i.e., 10-700 Torr) in flowing H₂. For temperatures ranging from 850 to 1050°C, the GaN decomposition rate is accelerated when the H₂ pressure is increased a

Full text of DOI:10.1063/1.122354

Enhanced GaN decomposition in H 2 near atmospheric pressures
D. D. Koleske, A. E. Wickenden, R. L. Henry, M. E. Twigg, J. C. Culbertson, and R. J. Gorman
Citation: Applied Physics Letters 73, 2018 (1998); doi: 10.1063/1.122354
View online: http://dx.doi.org/10.1063/1.122354
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/73/14?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Structural and optical characterization of GaN nanowires
J. Appl. Phys. 113, 204303 (2013); 10.1063/1.4808097
In situ GaN decomposition analysis by quadrupole mass spectrometry and reflection high-energy electron
diffraction
J. Appl. Phys. 104, 033541 (2008); 10.1063/1.2968442
GaN quantum dots doped with Tb
Appl. Phys. Lett. 88, 053102 (2006); 10.1063/1.2168504
Erratum: “Enhanced GaN decomposition in H 2 near atmospheric pressures” [Appl. Phys. Lett. 73, 2018 (1998)]
Appl. Phys. Lett. 75, 1646 (1999); 10.1063/1.124830
Characteristics of inductively coupled Cl 2 /BCl 3 plasmas during GaN etching
J. Vac. Sci. Technol. A 17, 2214 (1999); 10.1116/1.581749
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:
128.248.155.225 On: Sun, 23 Nov 2014 20:30:09

APPLIED PHYSICS LETTERS
VOLUME 73, NUMBER 14
5 OCTOBER 1998
Enhanced GaN decomposition in H2 near atmospheric pressures
D. D. Koleske,a) A. E. Wickenden, R. L. Henry, M. E. Twigg, J. C. Culbertson,
and R. J. Gorman
Laboratory for Advanced Material Synthesis, Code 6861, Electronics Science and Technology Division,
Naval Research Laboratory, Washington, DC 20375
͑Received 8 June 1998; accepted for publication 4 August 1998͒
GaN decomposition is studied at metallorganic vapor phase epitaxy pressures ͑i.e., 10–700 Torr͒ in
flowing H2. For temperatures ranging from 850 to 1050 °C, the GaN decomposition rate is
accelerated when the H2 pressure is increased above 100 Torr. The Ga desorption rate is found to
be independent of pressure, and therefore, does not account for the enhanced GaN decomposition
rate. Instead, the excess Ga from the decomposed GaN forms droplets on the surface which, for
identical annealing conditions, increase in size as the pressure is increased. Possible connections
between the enhanced GaN decomposition rate, the coarsening of the nucleation layer during the
ramp to high temperature, and increased GaN grain size at high temperature are discussed.
͓S0003-6951͑98͒03540-2͔
GaN has shown great promise for the production of blue
light-emitting diodes,1 lasers,2 and for high-power electronic
devices.3 Improvements in the material quality have played a
critical role in device performance breakthroughs.1,2 The
growth of high-quality GaN films on sapphire can be
achieved using metallorganic vapor phase epitaxy ͑MOVPE͒
at temperatures above 1000 °C and V/III ratios larger than
1000. For MOVPE growth, the temperature is typically
larger ͑i.e., 100–500 °C͒ than the threshold temperature for
GaN decomposition.4 The large V/III ratio is necessary to
offset the N desorption rate at the growth temperature used
for MOVPE growth,5 however, the magnitude of GaN de-
composition, which may occur during growth, is currently
not well understood.
In a previous paper we have suggested that decomposi-
tion of the GaN film during growth may enhance GaN order-
ing, by eliminating weakly incorporated Ga and N atoms.5
Growth conditions closer to equilibrium may be obtained
when the incorporation and decomposition rates of atoms
into the lattice are nearly equal with the incorporation rate
being slightly larger than the decomposition rate for a posi-
tive growth rate.5–7 In this letter, we focus on the GaN de-
composition kinetics in flowing H2. We show an enhance-
ment in the GaN decomposition rate for pressures Ͼ100 Torr
and describe how this enhanced decomposition may aid in
improving the quality of GaN growth at these higher pres-
sures.
For this study, GaN films were grown on a-plane sap-
phire at 76 Torr using a close-spaced showerhead reactor
design. After annealing the sapphire wafers in H2 at 1080 °C,
a thin ͑Ϸ200 Å͒ GaN nucleation layer was grown at 540 °C
followed by ramping to 1020–1030 °C for growth of 2–3
␮m of GaN. This growth process resulted in specular GaN
growth, with excellent across wafer thickness uniformity.
The temperature of the susceptor was calibrated by observing
the melting point of 0.005 in. diam Au wire and correlating it
to the setpoint of the heater measured with a thermocouple.
The temperature calibration in the reactor was found to be
reproducible to within 5 °C after eight months of use.
For the decomposition study, pieces of the GaN on sap-
phire were cleaved and weighed to within 0.1 mg using an
analytical balance. The pieces were reintroduced into the re-
actor and heated under varying conditions using a 6.0 SLM
flow of H2. Each piece was ramped at 25 °C per minute to
the annealing temperature. After annealing for a set time and
cooling, each piece was reweighed in air to determine the
mass loss. Repeated weighing of the same annealed pieces
over time resulted in reproducible weights to within 0.1 mg,
suggesting no net oxidation or water condensation of the
annealed sample. When the pressure is 40 Torr or greater,
liquid Ga droplets were observed. The liquid Ga droplets
were found to be very stable in air. With the liquid Ga cov-
ered surface sitting in air for over a week, the Ga droplets
could be coalesced into larger droplets by touching with a
tweezer, proving the Ga droplets are liquid and not signifi-
cantly oxidized. The Ga droplets were removed using dilute
HNO3 and rinsing with DI water. Each piece was weighed
again. The weight of the GaN decomposed was calculated by
subtracting the final weight from the initial weight. Likewise,
the weight of the liquid Ga droplets was calculated by sub-
tracting the final weight from the second weight. The weight
of desorbed Ga was calculated from the mass balance differ-
ence between the decomposed GaN and the liquid Ga. Fi-
nally, the piece was annealed at 1080 °C to remove the re-
maining GaN. The weights were converted to kinetic rates
͑atoms/cm2͒ by dividing by the sapphire area. The area of the
GaN thermal decomposition has been extensively stud-
ied in vacuum.4,8–12 From the previous studies, an activation
energy EA of 3.1 eV was measured for GaN decompo-
sition.9,10 This EA is only slightly larger than the EA of 2.8
eV for Ga desorption from liquid Ga,13 suggesting a possible
link between Ga desorption and GaN decomposition. These
studies showed that GaN decomposes into metallic Ga atoms
and N2 molecules, which both desorb from the surface.11 The
formation and desorption of GaN clusters has also been
observed.12
a͒
Electronic mail: koleske@estd.nrl.navy.mil
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:
2018
128.248.155.225 On: Sun, 23 Nov 2014 20:30:09

Appl. Phys. Lett., Vol. 73, No. 14, 5 October 1998
Koleske et al.
2019
FIG. 1. Ga droplets on the GaN surface after annealing at 992 °C for 10 min at pressures of ͑a͒ 40 Torr, ͑b͒ 76 Torr, ͑c͒ 150 Torr, and ͑d͒ 700 Torr. Note the
increase in size of the Ga droplets above 100 Torr. All images are the same scale with the bar on ͑d͒ indicating 100 ␮m.
irregularly shaped sapphire pieces was calculated using the
formula Aϭm/Td, where m is the weight of the sapphire, T
is the thickness of the sapphire wafer, which was typically
0.033 cm, and d is the sapphire density ͑3.98 g/cm3͒.
the GaN decomposition rate is also observed at temperatures
as low as 800 °C.
Currently, we do not understand the mechanism for the
enhanced GaN decomposition as the pressure is increased.
Clearly, from Fig. 2 the GaN decomposition rate coincides
with the increase of liquid Ga on the surface, however, it is
not clear if the Ga accumulation is the cause or a result of the
enhanced decomposition. Ga metal is known to dissociate H2
at high temperatures to form Ga hydrides,16 which may be
more mobile on the surface. This has been observed by Mor-
ishita and co-workers, who have shown that the Ga diffusion
length is increased when H2 or atomic H are used in the
molecular beam epitaxy growth of GaAs.17 In their interpre-
tation of the increased Ga diffusion length, Morishita et al.
speculated that the GaHx species are more weakly bound to
The buildup of Ga droplets on the GaN surface is shown
in Fig. 1 for a series of 10 min anneals at Tϭ992 °C at
pressures of ͑a͒ 40, ͑b͒ 76, ͑c͒ 150, and ͑d͒ 700 Torr. For
pressures less than 40 Torr no Ga droplets were observed,
which is consistent with previous vacuum experiments.8–12
As shown in Fig. 1, there is a noticeable increase in the size
of the largest Ga droplets as the pressure is increased from
76 to 150 Torr. The droplets are formed because the GaN
decomposition rate exceeds the Ga desorption rate from the
surface. The droplet size distribution suggests coalescence-
dominated growth of the Ga droplets, similar to Ga droplet
formation on GaAs when it is annealed above 600 °C.14
The surface Ga accumulation ͑droplet͒ rate, along with
the GaN decomposition and Ga desorption rates are plotted
in Fig. 2 as a function of pressure. Figure 2͑a͒ shows the
rates at 992 °C and Fig. 2͑b͒ shows the rates at 901 °C. It is
clear from Fig. 2͑a͒ that the GaN decomposition rate ͑filled
circles͒ increases as the pressure is increased. Furthermore,
the surface Ga accumulation rate ͑open squares͒ increases as
the pressure increases and appears to correlate with the in-
crease in the GaN decomposition rate. The Ga desorption
rate ͑filled diamonds͒ changes slightly as a function of pres-
sure, peaking near 100 Torr. We speculate that this small
increase is due to the liquid Ga surface area to volume ratio,
which is maximized from 76 to 150 Torr and decreases as
the size of the Ga droplets increase.
An Arrhenius plot of the Ga desorption rate kGa is shown
in Fig. 3. The data for Fig. 3 were measured at 40, 76, 150,
and 250 Torr. An exponential fit yields a preexponential of
(6.6Ϯ0.4)ϫ1029 cmϪ2 sϪ1 and an activation energy EA,Ga of
2.74Ϯ0.02 eV. This value for EA,Ga is in excellent agree-
ment with the value of 2.8 eV for Ga desorption from liquid
Ga,13 and 2.69 eV for Ga desorption from GaN in vacuum.15
It appears that the Ga desorption rate is independent of H2
pressure and depends solely on the liquid Ga surface area.
A more dramatic increase in the GaN decomposition rate
is observed at a temperature of 901 °C as shown in Fig. 2͑b͒.
The annealing times for the data at 40 and 76 Torr were 45
and 60 min, while for the other pressures the annealing time
was 10 min. As shown in Fig. 2͑b͒, there is almost an order
of magnitude increase in the GaN decomposition rate as the
pressure is increased from 76 to 150 Torr. As in Fig. 2͑a͒, the
increase in the GaN decomposition rate correlates well with
FIG. 2. Plot of the GaN decomposition rate ͑solid circles͒ at various pres-
sures at ͑a͒ 992 °C and ͑b͒ 901 °C. The Ga desorption rate ͑solid diamonds͒
and the rate of Ga accumulation ͑open squares͒ on the surface during the
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:
the increased Ga surface accumulation. A similar increase in
anneal are also plotted.
128.248.155.225 On: Sun, 23 Nov 2014 20:30:09

2020
Appl. Phys. Lett., Vol. 73, No. 14, 5 October 1998
Koleske et al.
shown in the atomic force microscope data of Ref. 20 for a
nucleation layer annealed at 100 Torr vs 140 Torr. In addi-
tion to increasing the nucleation layer coalescence, the en-
hanced GaN decomposition rate may aid in the ordering of
the epitaxial film by increasing the decomposition and incor-
poration rates,7 bringing the growth closer to equilibrium.6
Studies are underway to determine the pressure influence on
the coarsening of the nucleation layer before high-
temperature growth. This study illustrates a possible major
difference between reduced and atmospheric MOVPE GaN
growth.
The authors thank J. A. Freitas, Jr. and W. J. Moore for
characterization of the films. This work is supported by the
Office of Naval Research through the ONR Power Electronic
Building Block Program monitored by George Campisi.
FIG. 3. Arrhenius plot of the Ga desorption rate measured at four different
pressures versus the inverse temperature. All data are fit by kGaϭ(6.6
Ϯ0.4)ϫ1029 cmϪ2 sϪ1 exp(Ϫ2.74Ϯ0.02 eV/kBT).
1 S. Nakamura, M. Senoh, and T. Mukai, Appl. Phys. Lett. 64, 1687 ͑1994͒.
2 S. Nakamura, M. Senoh, S. Nagahama, N. Iwasa, T. Yamada, T. Mat-
sushita, H. Kiyoku, and Y. Sugimoto, Jpn. J. Appl. Phys., Part 2 35, L74
͑1996͒.
3 S. N. Mohammad, A. A. Salvador, and H. Morkoc¸, Proc. IEEE 83, 1306
͑1995͒; S. Strite and H. Morkoc, J. Vac. Sci. Technol. B 10, 1237 ͑1992͒;
R. F. Davis, Proc. IEEE 79, 702 ͑1991͒.
the surface, and therefore, can diffuse farther before
incorporation.17 The implication for the present study is that
any increase in the Ga diffusion length would more rapidly
uncover new areas of the GaN surface for N2 desorption,
which is 10–1000 times faster than the Ga desorption.5,15
The active dissociation of surface H2 may also occur on the
highly polar GaN surface, and this may be enhanced at high
pressures. Also, liquid Ga has recently been suggested to act
as a catalyst to initiate and enhance GaN decomposition.18
Further studies are underway to access the mechanism of the
enhanced GaN decomposition.
4 W. C. Johnson, J. B. Parsons, and M. C. Crew, J. Phys. Chem. 36, 2651
͑1932͒.
5 D. D. Koleske, A. E. Wickenden, R. L. Henry, W. J. DeSisto, and R. J.
Gorman, J. Appl. Phys. 84, 1998 ͑1998͒.
6 R. Heckingbottom, in Molecular Beam Epitaxy and Heterostructures, ed-
ited by L. L. Chang and K. Ploog ͑Martinus Nijhoff, Dordrecht, 1985͒, p.
71.
7 S. Yu. Karpov and M. A. Maiorov, Surf. Sci. 393, 108 ͑1997͒.
8 R. J. Sime and J. L. Margrave, J. Phys. Chem. 60, 810 ͑1956͒.
9 Z. A. Munir and A. W. Searcy, J. Chem. Phys. 42, 4223 ͑1965͒.
10 R. Groh, G. Gerey, L. Bartha, and J. I. Pankove, Phys. Status Solidi A 26,
353 ͑1974͒.
One major impact of this study on the growth of GaN in
our reactor is that the material quality is substantially im-
proved when GaN growth is conducted above 100 Torr.
When the GaN epitaxial layer is grown above 100 Torr, we
find a near doubling of the mobility (␮Ͼ500 cm2/V s for
11 R. C. Schoonmaker, A. Buhl, and J. Lemley, J. Phys. Chem. 69, 3455
͑1965͒.
12 A. S. Bolgar, S. P. Gordienko, E. A. Ryklis, and V. V. Fesenko, in Chem-
istry and Physics of the Nitrides, edited by G. V. Samsonov ͓in Russian͔
͑Naukova, Dumka, Kiev 1968͒, p. 151; also, see I. G. Pichugin and D. A.
Yas’kov, Izv. Akad. Nauk SSSR, Neorg. Mater. 6, 1973 ͑1970͒.
13 R. E. Honig and D. A. Kramer, RCA Rev. 30, 285 ͑1969͒.
14 M. Zinke-Allmang, L. C. Feldman, and W. van Saarloos, Phys. Rev. Lett.
86, 2358 ͑1992͒.
intentionally Si-doped films with nϭ2–3ϫ1017 cmϪ3
)
compared to growth at 76 Torr. Other groups using close-
spaced or high-speed rotating disk reactors have also re-
ported improved electric properties when their GaN epitaxial
growth is conducted above 100 Torr.19,20 In the films grown
above 100 Torr, the GaN grain size is increased to 2–5 ␮m
compared to grain sizes of Ͻ1 ␮m for growth at 76 Torr.
The reasons for the increased grain size are twofold and may
be related to the enhanced GaN decomposition that occurs at
higher pressures. It is known that during the temperature
ramp from the nucleation layer to epitaxial growth condi-
tions, the nucleation layer coalesces to form larger grains21
with increased roughness.22 If the ramp is conducted at high
pressures, the nucleation layer grain may coalesce to a larger
extent than at lower pressures. This increased coalescence is
15 O. Brandt, H. Yang, and K. H. Ploog, Phys. Rev. B 54, 4432 ͑1996͒.
16 H. Remy, Treatise on Inorganic Chemistry ͑Elsevier, New York, 1960͒, p.
18; W. R. S. Garton, Proc. Phys. Soc. London, Sect. A 64, 509 ͑1951͒.
17 Y. Morishita, Y. Nomura, S. Goto, and Y. Katayama, Appl. Phys. Lett. 67,
2500 ͑1995͒.
18 A. Pisch and R. Schmid-Fetzer, J. Cryst. Growth 187, 329 ͑1998͒.
19 B. T. McDermott, R. Pittman, E. R. Gertner, J. Krueger, C. Kisielowski,
Z. Lilienthal-Weber, and E. Weber, Talk D2.2 at Fall Material Research
Society Meeting, Boston, 1997.
20 J. Han, T.-B. Ng, R. M. Biefeld, M. H. Crawford, and D. M. Follstaedt,
Appl. Phys. Lett. 71, 3114 ͑1997͒.
21 A. E. Wickenden, D. K. Wickenden, and T. J. Kistenmacher, J. Appl.
Phys. 75, 5367 ͑1994͒.
22 J. C. Ramer, K. Zheng, C. F. Kranenberg, M. Banas, and S. D. Hersee,
Mater. Res. Soc. Symp. Proc. 395, 225 ͑1996͒.
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:
128.248.155.225 On: Sun, 23 Nov 2014 20:30:09