Undoped GaAs grown by the vertical gradient freeze method: growth and properties
ABSTRACT Undoped GaAs crystals were grown by the vertical gradient freeze technique in a multizone furnace. By optimization of the growth equipment and the thermodynamic conditions, crystals were obtained with a low density of microdefects and dislocations (etch pits density about 10 3 cm -2) as well as semi-insulating behaviour. Residual impurities and defect levels were investigated by optical (local vibrational-mode and IR absorption) and Hall effect measurements.
- [show abstract] [hide abstract]
ABSTRACT: CrO3-HF aqueous solutions have been applied to reveal and analyse defects in a large number of bulk and epitaxial GaAs crystals. This etching system has been used both in the dark (so-called DS method) and with light (so-called DSL method). The most important feature of the DS(L) system which distinguishes it advantageously from orthodox etchants such as molten KOH, Schell and RC solutions, is its ability to reveal a wide variety of defects, both crystallographic and chemical in nature. By a number of examples, proof is given that different defects which have been recognized by other methods in GaAs crystals previously, i.e. dislocations, stacking faults, twins, complex defects (streamers), precipitates, growth striations and recombinative impurity atmospheres, can be revealed and analyzed by a properly selected DS(L) procedure. In contrast with the orthodox etchants which form etch pits on defect sites, CrO3-HF aqueous solutions reveal almost all defects in the form of ridges and hillocks. The shape of these etch figures depends not only on the crystallography and geometry of the defects, but also on the nature of the related impurity atmospheres. An important feature of the present system is its sensitivity to light. For n-type and SI GaAs, both etch rates and defect sensitivity are strongly increased under illumination. The mechanisms of defect-revealing are also different: in the dark the shape of etch hillocks is mainly determined by lattice deformation around the defects, whereas under illumination the distribution of impurities plays a more important role. Using these concepts, it has been possible to make an inventory of different types of defects and to discriminate between different types of dislocations. With respect to the latter group of defects, the most important feature of the DS(L) system is its ability to distinguish (i) grown-in dislocations, (ii) glide dislocations, (iii) grown-in dislocations which glided after formation of impurity atmospheres and (iv) dislocations with a complex distribution of recombinative defects (for instance EL2). It is shown that with the DS/DSL method the slip planes of glide dislocations can also be determined. The DS(L) method is critically compared and/or calibrated with (i) other etching techniques used for defect studies in GaAs, such as AB, molten KOH, Schell and RC etchants, (ii) other characterization techniques, such as TEM and photoluminescence and (iii) numerous data from literature.Journal of Crystal Growth 11/1986; 78(2):191–217. · 1.55 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: We have evaluated the uniformity in [EL2], dislocation (or etch‐pit) density (EPD), resistivity, mobility, and carrier concentration for 3‐in., semi‐insulating GaAs wafers grown by the vertical‐gradient‐freeze (VGF) technique. Although slight W or U patterns were observed in [EL2] and EPD along the 〈110〉 directions, for the first time, nevertheless the overall uniformity was excellent, and comparable to that in the best In‐doped and whole‐boule‐annealed ingots grown by the liquid‐encapsulated Czochralski (LEC) technique. Based on results from implant‐activation studies on LEC wafers, it is estimated that the measured nonuniformities in EPD and [EL2] for the VGF wafers would contribute only about 1% to implant‐activation‐efficiency nonuniformities in Si‐implanted wafers designed for field‐effect transistor applications.Journal of Applied Physics 08/1989; · 2.21 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: We have developed a new growth process for GaAs that combines advantages found in several methods currently in commercial use, while at the same time minimizing many of the problems inherent in these presently used processes. The new technique, a form of vertical Bridgman (VB) growth, is capable of producing either doped (semiconducting) or undoped semi-insulating GaAs with very low dislocation density.Journal of Materials Research 01/1990; 5(7):1468-1474. · 1.71 Impact Factor
Materials Science and Engineering B28 (1994) 87-90
Undoped GaAs grown by the vertical gradient freeze method:
growth and properties
E. Buhrig, C. Frank, G. G~irtner, K. Hein, V. Klemm, G. Kiihnel, U. Voland
Technische Universitiit Bergakademie Freiberg, D-09596 Freiberg, Germany
Undoped GaAs crystals were grown by the vertical gradient freeze technique in a multizone furnace. By optimization of the growth
equipment and the thermodynamic conditions, crystals were obtained with a low density of microdefects and dislocations (etch pits
density about 10 3 cm-2) as well as semi-insulating behaviour. Residual impurities and defect levels were investigated by optical (local
vibrational-mode and IR absorption) and Hall effect measurements.
Keywords: Gallium arsenide; Crystallization; Electrical measurements; IR spectroscopy
The vertical gradient freeze (VGF) technique can be
regarded as an alternative to the vapour-pressure-
controlled Czochralski method . The particular
advantage of the VGF method in comparison with the
liquid-encapsulated Czochralski (LEC) method is the
very low dislocation density [2, 3]. With regard to the
horizontal Bridgman method , there are advantages
in the circular cross-section of the grown crystals and a
higher homogeneity [3,5]. Further features are a
growth process without any mechanical movements
and the compact design of the growth equipment.
There are also disadvantages, e.g. it is not possible to
observe the melt directly, and the growth rate is rela-
tively small. Furthermore, there is no way to influence
the convection of the melt.
The present work reports the process fundamentals
of the crystal growth of undoped VGF GaAs with low
dislocation density and the characterization of the
2. Crystal growth
The crystals were grown in a quartz ampoule which
was placed in a 15-zone furnace using an As source
with a defined temperature to stabilize the vapour
pressure. Pyrolytic boron nitride (pBN) was used as the
crucible material. The thermal regime was determined
with a computer-controlled process. First, the furnace
0921-5107/94/$7.00 © 1994
- Elsevier Science S.A. All rights reserved
design was optimized; a ratio of the diameter D to the
length l of nearly unity for each furnace zone turned
out to be a good choice. The overheating temperature
A T O of the melt corresponding to the relation
(d T/dx is the local temperature gradient) is the basis for
achieving a constant movement of the melt-point
isotherm [6, 7]. Suitable parameters are A T O = 12 K,
d T/dx = 1.8 K cm- t and 1 = 6.6 cm. Furthermore, it
was necessary to choose conditions and ensure a
growth rate low enough to transport away the melting
heat without marked deformation of the melting-point
isotherm and to avoid constitutional undercooling.
According to the relation
dT dT dx
dt dx dt (2)
(d T/dt is the temporal cooling rate during the gradient
freeze process and dx/dt the growth rate) a growth rate
dx/dt ~ 0.3 cm h-1 and a cooling rate close to 0.6 K
h- 1 are obtained .
The most essential point for the choice of the
chemical conditions is the As vapour pressure. Corre-
sponding to our investigations , an As source tem-
perature of 621 _+ 1 °C is necessary to produce high
quality GaAs. This corresponds to an As vapour pres-
sure PAs of about I bar. In this case, the crystal growth
E. Buhrig et al. / Mawrials Science and Lngineering B28 (I994) ,~7-~(~
is ensured with a small overpressure of As. So GaAs
crystals are formed with an As excess. This is necessary
to produce an EL2 concentration in the desired range.
It must be considered that all materials contained in
the growth equipment contribute to the total vapour
pressure. The conditions for such a growth system with
a pyrolytic BN crucible (source of B and N), quartz
ampoule (source of Si and O), graphite inserts (source
of C) and potential contamination by water (source of
H and O) are shown in Fig. 1.
Annealing experiments were carried out in equip-
ment similar to the growth furnace. Cubic samples with
dimensions of a =lcm cut from the ingot were
annealed in evacuated sealed ampoules at 1100 °C for
a time of 10 h followed by cooling at 100 K h- ~. The
As pressure in the ampoule was controlled by the
addition of As.
Fig. 2. Cellular structure of dislocations after selective photo
etching (interference contrast image; scale bar. 50(1 t~n~ I.
3. Crystal properties
3.1. Structural characterization
To investigate the structural properties of the crys-
tals in terms of their growth conditions, etching
techniques and analytical transmission electron mic-
roscopy (TEM) were used.
The dislocations are arranged in the well-known
cellular structure as demonstrated in an interference
contrast image after photo-etching in a CrO3-HF
solution  (Fig. 2 and Fig. 3). The mean dislocation
density (EPD) determined after KOH etching  is
clearly smaller than for LEC GaAs and shows values in
the range (0.5-4)x 103 cm -2, and the mean cell size is
about 1-2 mm. TEM investigations show decorated
dislocations. These decorations are arsenic tetrahedra
as demonstrated in Fig. 4, identified by the dark-field
~ 1,0 /
T = 1240 °C
GaAs +As +SiO 2 +BN +C +H20
Fig. 1. Dependence of vapour pressure on the components
(calculated for TAx = 617 °C).
Fig. 3. Decorated dislocations after selective photo-etching
(interference contrast; scale bar, 50 ~m).
/i!i!iiii~ ii ~ /~i ¸¸I
Fig. 4. TEM image of a stretched dislocation decorated by
arsenic tetrahedron (scale bar, 100 nm).
E. Buhrig et al. / Materials Science and Engineering B28 (1994) 87-90
Results of Hall effect and IR absorption investigations on undoped vertical-gradient-freeze-grown crystals
Hall effect IR
( X 1016 cm- 3)
1 X 107-8 X 107
3 x 107-2 X 108
1 x 101°-5x 1015 4000-6500
7 X 106-5 X 107
aNo changes compared with the as-grown state.
imaging technique and energy-dispersive X-ray analy-
sis. The dislocations are stretched between precipitates
owing to low thermal stresses during the crystal growth
and cooling. In the GaAs matrix, arsenic tetrahedra
were not detected at greater distances from the dis-
By annealing, the density of decorations on the
dislocations is reduced.
3.2. Electrical and optical properties
By Hall effect and conductivity measurements the
concentration n and mobility/~ of the charge carriers
were investigated. Depending on the growth and
annealing conditions the carrier concentration covers a
wide range between n = 107 and 1015 cm -3 and is n-
type in all cases. The residual impurities silicon (as Sica
donor) and carbon (as CAs acceptor) were measured by
local vibrational mode (LVM) absorption. The main
donor EL2 was detected by optical absorption at 1 ~m.
The results are listed in Table 1. On the basis of these
results the following can be stated.
(1) The different parameters measured on many
crystals show good reproducibility and the carrier
mobility/~ reaches values up to 6500 cm 2 V- 1 s- 1.
(2) The VGF process is characterized by a rela-
tively high Si concentration (about 1015 cm -3) owing to
the quartz ampoule and by a low carbon content
((1-5) × 1014 cm -3) (see also ). Therefore it is more
difficult in comparison with the LEC technique to grow
undoped GaAs with semi-insulating behavior.
(3) Apart from the impurity and defect levels
mentioned above (Si, C and EL2), additional acceptor
levels and donor levels shallower than EL2 must be
taken into account. The existence of such donor levels
was confirmed by temperature-dependent Hall effect
measurements (Fig. 5). In agreement with deep-level
transient spectroscopy on low resistivity samples, the
donor levels were identified as EL6 defects or mem-
bers of the EL6 group .
The EL2 concentration shows an average value of
(0.85_0.15)× 1016 cm -3, which is typical for VGF
10 7 t
E.~ = 0.16-0.24 eV
= 0.4. eV
= 0.76 eV
-T - :30OK
n~= 1 017.., 10_~Scm-'
i I '
1/T in K -1
Fig. 5. Temperature dependence of the carrier concentration on
different VGF samples.
÷ ÷ ÷ ~8-
0.00 r 0.00
0.00 20.00 40.00
Fig. 6. Radial distribution of carrier concentration n and
mobility kt of a semi-insulating VGF wafer.
E. Buhrig et al. / Materials Science and Engineering B28 (1994) 87-90
GaAs material [3, 5]. The electrical uniformity of a
semi-insulating wafer is shown in Fig. 6.
By annealing experiments, low resistivity samples
can be transformed to semi-insulating behavior (see
Table 1) similarly to what was reported by other
researchers . The related mechanism is not yet
clear; possibly new acceptors are formed. The mobility
values vary over a wide range. The low values of the
measured mobility in the annealed semi-insulating
samples are probably due to the effect of mixed con-
ductivity of electrons and holes.
We successfully grew undoped GaAs single crystals
at a very low temperature gradient using the VGF
method. The dislocation density was reduced to an
EPD of about 500 cm -2. TEM investigations showed
that the dislocations are stretched between precipitates
owing to low thermal stress. The decorations are
arsenic tetrahedra. Because of the high Si content
originating from the quartz ampoule it is difficult to
grow semi-insulating GaAs.
wafers show a good uniformity of electrical properties.
This work was supported by the Bundesministerium
f/ir Forschung und Technologie and Deutsche Agentur
fiir Raumfahrtangelegenheiten GmbH.
 T. Kawase et al., Proc. Conf. on Semi-Insulating IlI-V
Materials, Ixtapa, 1992, IOP, Bristol, 1993, p. 85.
 J.E. Clemens and J.H. Conway, Proc. Conf. on Semi-
Insulating lll-V Materials, Malm6, 1988, 1988, p. 205.
 R.E. Kremer, D. Franeomano, G.H. Beckhart and K.M.
Burke, J. Mater. Res., 5 (1990) 1468.
I4] K. Fujii, M. Hirata, H. Fujita and S. Takeda, in K. Sumeno
(ed.), Defect Control in Semiconductors, North-Holland,
Amsterdam, 1990, p. 667.
 D.C. Look, D.C. Walters, M.G. Mier, I.S. Sewell, I.S.
Sizelove, A. Akselrad and I.E. Clemans, J. Appl. Phys., 66
 K. Hein, E. Buhrig and H. G6hler, Cryst. Res. Technol., 27
 K. Hein, E. Buhrig and T. Richter, Cryst. Res. Technol., 27
 K. Hein, E. Buhrig and C. Frank, Metall., 47 (1993) 924.
 S. Hegewaid, K. Hein, C. Frank, E. Buhrig and M. John,
Cryst. Res. Technol., 29 (1994) 549.
 J.L. Weyher and J. van de Ven, J. Cryst. Growth, 78 (1986)
 J.G. Grabmaier and C.B. Watson, Phys. Status Solidi, 32
 G. Kiihnel and W. Siegel, Semicond. Sci. Technol., 6 (1991)
 B.E. Freidenreich, K.D. Bronsema, Z. Korzeniowski, B.
Nykiel, D. Francomano, J. Moore and R.E. Kremer, Proc.
Conf. on Semi-Insulating III-V Materials, Ixtapa, 1992,
1993, p. 105.