LB_-31258
DE92 041165
Crystalline Growth of Wurtzite GAN on (111) GaAs
J. Ross,* M, Rubin,** and T.K. Gustafson*
*Department of FAec_cal Engineering
University of California
and
**Energy and Environment Division
Lawreace Berkeley Laboratory
University of California
Berkeley, California 94720
December 1991
This work was supported in part by the National Science Foundation under Contract No. 1-442427-21482, a
RockweU International Fellowship, and the U.S. Department of Energy under Contract No. DE-AC03-
76SF00098_ ,_, ,,_,_
i.?
, .
CRYSTALLINE GROWTH OF WURTZITE GAN ON (111) GaAs
J. ROSS*, M. RUB.IN** AND T.K. GUSTAFSON*
*University of CaJffomia, Department of Electrical Engineering, Berkeley, CA 94720
" **Lawrence Berkeley Laboratory, Berkeley, CA 94720
p
ABSTRACT
GaUium Nitride films were grown on (11l) Gallium Arsenide substrates using reactiv_
ff magnetron sputtering. Despite a 20% lattice __isrnatch and different crystal structure, wurtzite
GaN films grew epitaxially in basal orientation on (111) GaAs substa'ates. Heteroepitaxy was
observed for growth temperatures between 550-6000C. X-ray diffraction patterns revealed
(0002) GaN peak with a full-width-half-maximum (FWHM) as narrow as 0.17 °. Possible
surface recorL_'uctions to explain the epitaxial growth are presented.
INTRODUCTION
Gallium Nitride is a direct wide bandgap semiconductor (3.4 eV) having potential
applications for stimulated emi_ion in the blue, violet, and ultra-violet spectral range.
Development of GaN devices have been limited by problems in obtahfing p-type conduction and
convenient substrates for epitaxial growth. Recent reports of p-type conduction [1,2] iu Mg
doped sample_ show promise in this area. As seen in Table I, there are no readily available
substrates that are lattice matched to GaN in both lattice parameter and thermal expansion
[
coefficient. Sapphire has been frequently used despite a 16% lattice mismatch. Attempts on
i
other substrates include: silicon [3,4], gallium arsenide I3,5,6,71, gallium phosphide [3], and
silicon carbide [8]. Aluminum nitride has been used effectively as a thin-film buffer layer on
sapphire for improved GaN growth [9]. Table I summarizes the material data for GaN and
various substrates. The lattice parameters for the cubic crystals are given as the effective
spacing in the (I I1) plane eorrespoz_dingto "a" in the wurtzite system for easier comparison.
Table I. Lattice parameters and thermal expansion coefficients of various substrates.
......Material ..............................._c-o_ AaGaNtAasub T_ai exp._---
(A) (%) coef. (xl0"6'/K)
GaN a= 3.189 ..... 5.6
, c= 5o182 .... 7.7
A1N '_111 +2.5 5.3
c= 4.980 ..... 4.2
• ct-SiC (6H) a= 3.08 +3..4 4.2
c= 15.11 .... 4.8
A1203 a= 2.74 +16.1 "-'-"_."3"--'-'---
c= 12.991 ..... 8.5
GaAs (111) a= 3.99_- _0.2___._ - - 4, 8 ........
Despite the large mLsmatch,, GaAs is desirable due to its wide acceptance in the
electrooptic industry. Few researchers have used GaAs as a substrate for GaN growth possibly
because early comparisons showed sapphire to produce smoother and more oriented GaN films
[5]. Recently, the electrical properties of cubic GaN in a S-I-S structure on (100) GaAs grown
by modified molecular beam epitaxy has been reported [I0]. Concurrent work in Iapan,
¢
involving the growth of GaN on (11 I) GaAs by MBE has 'aLsoshown wurtzite GaN on (1I 1)
GaAs [6], however, our material appears to have narrower (0002) GaN x-ray diffraction peaks.
ql
GaN is typically grown by chemical vapor deposition (CVD) or modified molecular beam
epitaxy (MMBE). Sputter deposition has produced highly resistive GaN films in the past [1I],.
and although it is perhaps not suitable for tl_e growth of minority caz_ier ele':tronic devices,
sputter deposition can be a viable method for the study of GaN growth kinetics and rudimentax3,
GaN devices.
In this paper, we report the growth of highly oriented wurtzite GaN films on the (111)
face of GaAs. Little s_'."ainis observed, but crystalline growth is a sensitive function of
temperature. The growth conditions and possible explanations for the heteroepitaxy are
detailed. This is the first report of crystalline wm_tziteGaN on (I 11) GaAs by sputter deposition
techniques.
EXPERIEMNTAL PROCEDURE
The GaN films were. deposited usia-_gan US Gun-II 2 inch modular source. The target
was pure gallium (9.999999%) held in a stainless steel cup. The GaAs substrates wer_
degreased and etched before deposition in a <111> directional etch of 1 H20:5 H2SO 4. 1.
H20 2 at 65°C for 2 minutes. The substrates were then rinsed with alcohol and dried before
entering the chamber. The chamber was evacuated to less than ._O"7Torr, and then backfilled
with a mixture of N2 and Ar gas to 25 rnTorr. The substrates were heated to 500-700°C as
measured by a thermocouple clamped on the surface of the heating block. As soon _.s the
discharge was ignited, the Ga target liquified and slowly formed a nitrided crust. ENI sources
were used to deliver 110 watts of rf (_3.56 MHz) power to the 2 inch t_r_,c_. '_e growth rate
was measured by a quartz-crystal oscillator calibrated with a stylus prox.,,Jmeter. Growth rates
ranged from 1-5 A/s. After deposition, the substrates were cooled to 200°C irt 100% nitrogen
atmosphere at 30 mTorr. The choice of deposition parameters for ,_.pitaxial g_owth w_ partly
determined from our earlier work on sapphire substrates [12,131. Higher partial pressures of
nitrogen (25 mTorr) were needed to crystallize GaN on (111)GaAs compared tc, the sapphire
substrates. However, similar nitrogen flow rates (200 sccm) and N2:Ar ratios (7.3) were used.
RESULTS AND DISCUSSION
The GaN film's crystal orientation was analyzed using a Siemens x-ray diffractometer
(CuKcz, _.=0. 154 nra). For temperatures below 550 °C only mixed phases of GaN were
ootatned (Figure la). In this temperature range, the fi_ms were characterized by rough surfaces
with numerous defects as seen in Figure lb. For growth temperatures between 550..600 °C
=
2
I w t I......... i
(a) (1 1I) GaAs
5z5°c
25 mTorr
(I)
t (222) GaAs
GaN
(0002)
II(IoTo)/.(IOTI)
lO ZO 30 40 50 60 70 ,
2e
(b) :' ._ "': ' "'?" ' "%;"" " "' '"" _P Or "' '" " "
L _ . -,.,,,.._",a,'."" , ,,_'."" ".'--' : ",:
.":'._' " . * q¢" • 'w ;_ ". _,_p,]., D ., . ..
_ -,,"- ,.,W_,.,,_.,,.-.,".'L , , ,,,-X ..iv,, . -, -,
i_ "I'.;';,.',.,',;.,r,_b,_'._,__ ,_- . ; ., *,..;, _, ,. . r,:;..
l". ::.... "_",,._ :._ ,,.-_;.... . "' " .. • .- _ ,_.... _' _ _'; '
I_.' _':.JqL_']l_'_" ' .,,- , --' *° , ' -'t, nr_ " • ''_ _ "
t, ,:.;.,m_,/.>..._ ..,_-,-..;"-_,:r,s_"",:_. _". : :._._:
15,,-._..,,,,,...l_', .. _. • ]L _li'' • ,_ '-,.-,. I0 /.Lm
L_.". _ __.'.- -.. ti,'_ "_ _ ,'_ • .llii3_ ' ' " ," ' , 4,:-_llr,-.' .1
•"4_"""-- " ..al, "-,,:_-.,_.• -. • ', _:.'y., .. • . ":. ';._:.;,_,":.'I
,. - .: ., • . ... pr ;...,o_ . • ,: _ ; .;,,-.. _::
• -,,. . ........ ,_ -..,:l
. .
_._."_-,_. , " '.';, .-_" . _ Idi .... ".. " _.''
•-_ ._ _ • • ";w_ll[__. .. . _- -..' ,'9_, ".'¥_'; i
_'_'."-_ -'-.. "',.i - , ,ii m_ _ . , ..,, . ,.w',;. ,.,I
"_L _-" _.... _ : ' _"" "_ " ,. _" ' , .,li
.. . :. ,.
,_. .........,. , .,. • .,,,,, .,_,, _,,,
.... .. _. - • • . , _, . ., _ "_ .. _.._
.._._ :,:' ,'."r._:¢,. ' "" ": "'"" i;' "_'.......
• . • . . . .,'-- _j._.._.:.... .... _, ._ ,,:.. _,, . . .,. ._,; _ .,- ...._.:_,,,.
Figure l(a) X-:raydiffraction pattern of GaN film grown at 525°C; and (b) the surface of the
film characterized with many defects.
highly oriented basal plane GaN was achieved with much smoother surface morphology.
Figures 2a and 2b show the x-ray pattern and the surface morphology of a 200 nm film grown
at 580 °C. The I:%VHMof the (0002) peak is 0.17 °. The peak is located at 20=34.607 ° which
corresponds to a d-spacing of 2.589 A. The nteasured (0002) plane spacing is therefore 5.178
, A, which agrees well with the theoretical value of 5.182A. Despite the large lattice mismatch,
the (0002) planes do not appear appreciably strahaed. We did not observe any critical thickJaess
. phenomen. For films grown under similar conditions, no variation in peak location or width
was measured for film thicknesses ranging from 0.05 -2 p.m.
For substmte temperatures greater than 620'C the x-ray diffraction peaks vanished and
many films delaminated from the GaAs substrates. We found the delamination could be
3
(a)
. CaAs (1 i 1) 58,0 *C .
_, 25 mTorr
ul,-4
m CaN (0002)
. o
X --- _'-- FWHM = 0'.17 o
10 _m
Figure 2(a) X-ray diffracttion pat_m of GaN film grown at 580°C; and(b) the smoother
surface morphology of this f'flm.
mininfized and in some cases eliminated if the films were cooled down slowly from the growth
temperature (~100°/hr). Also, no appreciable GaN x-ray peaks were observed at pressures
below 20 retort. To ensure the x-ray diffraction peak at 20=34.607 ° is indeed the (0002)
wurtzite pha_ of GaN and not strained cubic GaN growing in the (111) direction, reflection
electron diffraction (RED) was employed. With the beam incident on the (1120) plane, only
spots in ver:ical rows were visible as predicted by theory. Carrierconcentrations for these GaN
films were ali n-type and greater than 2 x 1018 cm"3.
We believe the reason highly oriented wurtzite GaN was obtained on (I 11) GaAs can be
explained by examining the 2-D interface. Figure 3a shows the wurtzite and zinc-blende cry_al
structures with the <0002> and <l li> directions aligned. The only difference between the
wurtzite and zinc-blende structures in these directions is the stacking order of the layers. The
(111) GaAs face has the identical bond termination as the GaN basal plane substrate would.
4