RESEARCH ARTICLE | SEPTEMBER 01 1974

Structure of vapor‐deposited GaxIn1−xAs crystals 

Haruo Nagai

J. Appl. Phys. 45, 3789–3794 (1974)

https://doi.org/10.1063/1.1663861


View Export
Online Citation

12 June 2024 11:46:50

Structure of vapor-deposited Gax ln 1 _ x As crystals
Haruo Nagai
Musashino Electrical Communication Laboratory, Nippon Telegraph and Telephone Public Corporation,
Musashino-shi, Tokyo, Japan
(Received 2 January 1974; in final form 10 June 1974)

Deformation of the crystal lattice of layers of Gaxlnt_xAs (0.5 < x < 1.0) grown onto GaAs
substrates by a chemical vapor deposition process was investigated. For this study, Ga and In were
transported as chlorides with HCl and AsH) was the source of As. It was observed that the lattice
constant along the direction of the film thickness was larger than that along the surface. This fact
indicates that the crystal lattice of GaxIn t _ xAs is compressed along the substrate surface as a result
of lattice mismatch. The influence of both the composition of the epitaxial layers and the crystal
plane of the substrate surface on the degree of this deformation was investigated. It was also found
that the epitaxial orientation grew inclined to the substrate crystal when the crystal surface of the
substrate was pOO j or {II OJ.

I. INTRODUCTION carrier concentrations in the range 1016 _10 17 cm- 3 • The

effect of the doping of the epitaxial layers was not
Ga)n1_xAs crystals are interesting materials for two
tested. Since GaAs and GaxIn1_~s have cubic crystal
main reasons. First, the value of energy gap covers the
lattices (zinc blende); crystal planes of the same index
range 0.9-3 Mm which is of practical interest for in-
both in the substrate and in the epitaxial layer must be
frared-emitting diodes 1 and detectors.2 Second, theyof-
parallel if an ideal epitaxial relation is expected [Fig.
fer the prospect of having interesting properties for
1 (a)J. But they are not parallel if the crystal lattice of
transport phenomena. 3
the epitaxial layer is deformed [Fig. 1 (b)] or if the
Preparation of Ga"Inl_xAs bulk crystals by the Bridg- epitaxial layer grows inclined to the substrate crystal
man or the Czochralski methods is very difficult be- plane [Fig. l(c)]. These two phenomena may occur si-
cause of the change of the composftion during the crys- multaneously [Fig. l(d)]. These relations are tested by
stalUzation. Therefore, the most convenient methods to x-ray diffraction. The angle between the two crystal
prepare the Ga)n1_x As are thought to be vapor or liquid planes is known by measurement of the angle between the

12 June 2024 11:46:50

epitaxial growth using GaAs crystals as substrates. Ac- two corresponding diffraction peaks in changing the in-
tually, some properties of Ga)n1_ xAs epitaxial crystals cident directions of x rays. 7 Its principle is illustrated
obtained by these methods have been reported. 4 ,5 We in Fig. 2. Two diffraction peaks of the substrate and the
present here detailed x-ray studies of the effect of the epitaxial layer were observed by rotating the sample
heteroepitaxy on the properties of the epitaxial layers. using the parallel x-ray beam 1 mm in length and 150
Mm in width. An x-ray Lang camera was used for this
In heteroepitaxial systems, it is known that there is
purpose. When the crystal surface of the substrate was
strain because of the mismatch between lattice constants
{100}, the angles between two corresponding {400},
and because of the differences in the thermal expansion
coefficients. As a result, bending of the structure, de- {5H}, or {422} planes in the epitaxial layer and the sub-
formation of the crystal lattice, and generation of the strate were measured. When the Substrate surface was
{lll}, planes of {333} and {422} were investigated.
misfit dislocations can occur. In this report, we present
The degree of deformation of the crystal lattice can
some observations on deformation of the Ga)n1_xAs
crystal lattice deposited on a GaAs substrate from the be calculated from these data.
vapor phase (0.5 < x < 1.0). In addition to deformation,
inclination of the epitaxial crystal was found when the
substrate surface was {100} or {HO}. This phenomenon
is also discussed.

II. EXPERIMENT
GaxIn1-xAs crystals were grown on GaAs substrates in
the temperature range 700-750°C. The apparatus used
for the growth was the same as that used for the vapor
growth of GaAs by a Ga-HCI-AsHs-H2 system, where
Ga-In alloys were used instead of Ga. The examined
crystalline orientation of the GaAs substrate were {100}
and {lll}. The cutting accuracy of the substrate surface
was made within 1°. The substrate surface was treated
by two methods. One method is a chemical etching by a
4 H2S04 : 1 H20: 1 Il20 etching solution at 80-90 °C, and
the other method is a mechanochemical polish. The
thickness of the substrates was in the range 2-5 mm,
and the epitaxial layer was in the range 5-10 Mm. In
this case the bending of the system 4 ,6 can be neglected. FIG. 1. Relation between the epitaxial and the substrate crys-
Generally, deposited crystals were undoped n type with tal lattice.

3789 Journal of Applied Physics, Vol. 45, No.9, September 1974 Copyright © 1974 American Institute of Physics 3789
3790 Haruo Nagai: Structure of vapor-deposited Gax In1_xAs 3790

(1) (2) two planesare parallel x (GaxInl_xAs)

~~2
090 0·80 070 0·60 050
(1) 0,=9,- 92 30
_ ~ Bragg

~=~==~r:J..:i~=d1 an~:
(2) O2=91- 92
two planes are inclined
Bragg angle 9, (1) 0;=(9,-92) + ~

(2) O2=(9, - 9z> - rJ..

..
::. Oe-Os
.~ as
.~
FIG. 2. Principle of the measurement of the angle between
the two crystal plane.
1·0 2·0 3·0 4·0
ideal misfit factor (%)
III. RESULTS
FIG. 4. Relief of the lattice mismatch by the deformation vs
A. Deformation of the crystal lattice mismatch factor.

The lattice constant perpendicular to the film is larger

than that along the surface. In Fig. 3, the influence of strate surface is {100}. These two planes make an angle
the composition of the epitaxial layers on the degree of up to 250" in the direction of the (110). Therefore, in
deformation is shown. The deformation is expressed by these {100} samples, inclination and deformation of the
the ratio of the lattice constant perpendicular to the sur- crystal lattice are observed simultaneously. When the
face to that along the surface. It was ascertained exper- substrate surface was treated by a chemical etching
imentally that the deformation of the crystal lattice in most samples have the epitaxial layers as shown in Fig.
the substrate did not occur. Figure 4 shows the relation

12 June 2024 11:46:50

between the mismatch factor and the relief of the lattice
mismatch by the deformation of the epitaxial layer. This a) b)
figure was calculated from the result of Fig. 3. The lat-
tice deformation was checked varying the thickness of an ~gi
epitaxial layer (Gaa.93Ino.o7As) thickness 9 J1.m) by chem-
ical etching. In this case, it was concluded that the de-
gree of deformation was constant along the thickness
direction within the accuracy of measurement.

B. Inclination of the crystal lattice

(i) The {100} planes of the epitaxial layer and the sub-
strate are not parallel when the crystal plane of the sub-
{S1H diffraction
J
...___.... {S11ldiffraction
(011)
epitaxial
1.0030 epitaxial layer thickness 5-10 pm
. layer(100)
c) d)
• (100)
A (111)A
1.0025
CII
01.0020
"-
CII
U '·0015

1.0010
--...
{S11} diffraction
com
substrate (100)
0·9 0·8 0·7 0·6 0·5 0·4 FIG. 5. Reflection x-ray topographs of a {IOO} sample. Epi-
X (Ga x ln _ As) taxial layer is GaO.88InO.12As; thickness, 9 j,I m. Substrate
1x thickness, 5 mm. The discrimination between the (110) and
(IIO) axes was done by the method of etching in the
FIG. 3. Effect of the composition of the epitaxial layers on Cr03-HF-H20 system (Ref. 8). In this sample, the angle be-
the deformation of the epitaxial crystal lattice. The composi- tween the two {100} planes of the epitaxial layer and the sub-
tion was determined from the lattice constant. strate is 205".

J. Appl. Phys., Vol. 45, No.9, September 1974

3791 Haruo Nagai: Structure of vapor-deposited Gax In 1 -x As 3791

TABLE 1. Inclination, deformation, and lattice constants TABLE II. Inclination of the epitaxial layers on different sub-
of the epitaxial layers on some different substrate surfaces. strate surfaces. [The composition of the epitaxial layer was
[The composition of the epitaxial layer was estimated to be estimated to be Gao.SsInO.12As from the lattice constant on sur-
Gao.SsInO.12As from the lattice constant on substrate surface (1). face (1). The thickness of the substrate was 2 mm, and that
The thickness of the substrate was about 4mm, and that of the of the epitaxial layer was 7 It m].
epitaxial layer was 7 It m on surface (1) J. 80 a
Substrate surface
Substrate surface 80 a d 100• t
b
d100•s !fullw. (sec)
(sec) (A) (A) d 100•s
along (110) , 55
(1) {I00}±1' 20 5.7132 5.6995 1.0024 along (110), 0
(2) 4.5 0 inclined along 540 5.7076 5.6962 1.0020 (2) 1. 0 inclined along {n 0)
0
along (11 0), 0
(110) from the surface from the surface of (1) along (110), 0
of (1)
(3) 1. 0 0 inclined along (110) along (110), 0
(3) 8. 8 0 inclined along 1100 5.7036 5.6962 1. 0013 from the surface of (1) along (110), 195
(110) from the surface
of (1) "Numbers following the directions are angles between the {IOO}
planes of the epitaxial and substrate crystaL
aAngles between the {100} planes of the epitaxial and substrate
crystal along the direction (11 0).
bd 100 •t and d 100 •s denote the lattice constant of the boo} along
the direction of the thickness and the direction of the surface, In the region x> 0.8, the degree of deformation in-
respectively. The calculation of the lattice constants was creases with the indium content, but in the region
performed assuming that the d 100 of GaAs that was used as the x < O. 8, it decreases with the indium content. When the
substrate was 5.6532 A. Measurement was made at 26 'C. values of lattice constant difference are small enough,
the mismatch is thought to be relieved only by the
5. The epitaxial layer is divided into two regions which elastic deformation of the crystal lattice of the epitaxial
have different orientatio~ of the {lOO} planes. The layer. When the mismatch reaches the limiting values,
boundary lies along the (110). Two regions divided by
one boundary are symmetric with respect to the inclina-
tion of the {IOO} plane. When the substrate surface was
treated by a mechanochemical polish, boundaries like
those of Fig. 5 were not seen, but the {lOO} planes of

12 June 2024 11:46:50

the substrate and the epitaxial layer were not parallel,
and the direction of the inclination was (110). The same
experiment, which used the accurately cut substrates
(the cutting accuracy was made within 0.1°), showed the
same results.
(ii) When the substrate surface is {110} , the {1l9}
planes of the epitaxial layer and that of the substrate
are not parallel. The direction of the inclination is
(100).

(iii) When the substrate surface is {lll} A, the {l1l}

planes of the epitaxial layer and that of the substrate are
parallel within the accuracy of measurement.
(iv) A single substrate surface of {100} or{lll} was
divided into several places which were angles lapped
in various directions and at different angles. Gaxln1_xAs
layers were grown on these substrates and examined.
The results are shown in Tables I-III. The inclination
of the epitaxial layer increased with the inclination of the
substrate surface from {100} and {Ill} (see Tables I and
III).
....
EE I
IV. DISCUSSIONS
A. Deformation of the crystal lattice
The lattice constant of InAs (6.058 A) is larger than d _

that of GaAs (5.653 A), and it is reported that Vegard's (0"')

law can be applied to GaxIn1_xAs. 9 Figure 3 shows that
the crystal lattice of GaxIn1_xAs is smaller when parallel (400) d iffroction
to the substrate surface and larger when parallel to the
thickness direction because of the lattice mismatch. The FIG. 6. Reflection x-ray topograph of a crosshatched pattern
on a hoo} substrate. Epitaxial layer is GaO.9sInO.02As; thick-
effect of the thermal expansion coefficients can be ne- ness, llitm. Substrate thickness is 2 mm. (a) and (b) are
glected because the difference of the coefficients of GaAs inclined 5" along (110) and (110), respectively, from the {lOO}
and InAs is small. surface.

J. Appl. Phys., Vol. 45, No.9, September 1974

3792 Haruo Nagai: Structu re of vapor-deposited Gax In 1 -x As 3792

TABLE III. Inclination of the epitaxial layers on some different boundary

a)~~
substrate surfaces. [The composition of the epitaxial layer
was estimated to be GaO.90InO.1oAs from the lattice constant on ___ C)
surface (1). The thickness of the substrate was about 3 mm,
and that of the epitaxial layer was 8 P. m on surface (1)1. ...... -""

Substrate surface 60
(sec) a
\-0,.,
~
(1) {Inho. 2° o '" ~(Olll-'"
(2) 3.0° inclined along (211) 187
from the surface of (1)
(3) 5.0° inclined along (211) 238
from the surface of (1)

aAngles between the {Ill} planes of the epitaxial and substrate

crystal along the direction of (211).

FIG. 8. Growth direction of the epitaxial layers on various

the misfit dislocations are formed. The lattice mismatch substrate surfaces.
is relieved by both the dislocations and lattice deforma-
tion in this region. The region x> 0.8 is thought to be
this case. As shown in Fig. 4, 80-90% of the lattice the substrate surface treated by a mechanochemical
constant difference is relieved by misfit dislocations and polish is flat. Generation of the boundary in the epitaxial
10-20 % is relieved by the lattice deformation in this layer seems to be related to the configuration of the
composition range. The generation of the misfit disloca- substrate surface. Figure 7 shows the stacking-fault
tions in these GaxIn1_xAs-on-GaAs structures have been shadow which is thought to be evidence of the lateral
observed 1o by Enstrom et al. Also in this work, a cross- growth. 11 The direction of the shadow shows the growth
hatched pattern which is thought to be caused by the mis- direction. On the sample of Fig. 5, the directions of
fit dislocations was observed by x-ray topography and an the shadows are different in different regions. They are
example is shown in Fig. 6. In the region x < 0.8, de- in opposite directions in the two regions inclined sym-
fects that took the form of short-line imperfections ori- metrically with each other, as shown in Fig. 5(d). On

12 June 2024 11:46:50

ented along (110) and (flO) directions on the {100} sur- the sample that has no boundary, the direction of the
face were remarkable. They are thought to be stacking shadows is uniform approximately along the (110)
faults with line structure and increase with indium con- direction.
tent. For example, they are about 1-10 )lm in length The (110) direction is thought to be one of the easy-
and their density is about 106 cm- 2 on the surface of the growth directions in this mixed crystal system. There-
Gao.49InO.51As epitaxial layer (thickness, 7 )lm). Gener- fore, when the substrate surface is near the {lOO} and
ation of the stacking faults may be related to the large
value of mismatch of lattice constants in the region. It
might be possible that, when the mismatch of lattice a) (100) b)
constants becomes large, these defects are also gener- ( 100)
ated in order to relieve the mismatch.

B. Inclination of the crystal lattice

The causes of the phenomena described in (i-iv) are
not well known yet. The real orientation of the substrate
surface is thought to be important. Although the accura- "'----'"' (0 11)
cy of the cutting of the substrate is within 1 the sub-
0
,

strate surface is rounded because of the chemical etch-

c) d)
ing. The surface of the sample shown in Fig. 5 is spher-
ical and has a radius of curvature of about 350 mm. -100,..,m
Etching at 80-90°C by a 4H 2S04 : 1 H2 0 2 : 1 H2 0 etching
solution makes this curvature in 60". On the other hand,

FIG. 9. (a) and (b) Generation of the boundary line on the con-
cave substrate surface. (c) Epitaxial layer is GaO.97InO.osAs;
thickness, 16 p. m. The radius of curvature of the substrate
surface is 2 mm. (d) Epitaxial layer is GaO.96InO.04As; thick-
ness, 24 p. m. The radius of curvature of the substrate sur-
FIG. 7. Stacking-fault shadows on a {IOO} sample. face is 3mm.

J. Appl. Phys., Vol. 45, No.9, September 1974

3793 Haruo Nagai: Structure of vapor-deposited Gax In1_xAs 3793

(100) ::-... The effect of misorientation of a {100} substrate was

a) .. -------~8, studied by determination of the relation of the inclination

._.l~~:._ . =~~=~ . .). ~o

epltaxlal ) _., _••••
flOO
along (110) between the substrate crystal and the epi-
taxiallayer. This is shown in Fig. 10. The angle be-
substra te(100)/ tween the epitaxial layer and the substrate {lOO} planes
passes through zero as the inclination of the substrate
surface from {lOO} increases [Figs. 10(a)-10(c)]. The
b) ...------=-. degree of inclination of the substrate surface due to the
chemical etching is within the case of Fig. 10(a). At a
8=0 certain angle of the substrate surface, the epitaxial lay-
o
er grows parallel to the substrate crystal [Fig. lO(b)].
In the composition range of this experiment, 82, near
lOis obtained for 80 = 0 (see Table II). 80 and 82 are ex-
c) ....--- pected to be the functions of the composition of the epi-
taxial layer, but systematic research has not been done
__ ._._._._._._. _~9~(.·_· because the quantitative reproducibility of the data is not
good. This problem of the reproducibility should be re-
lated to the difference in the substrate surface. Though
.>/. direction of the stacking-fault
all the substrate surfaces were treated in the same
shadows
manner, it was difficult to say whether they were in the
same condition. But it was recognized qualitatively that
FIG. 10. Relation between the inclination of the substrate sur-
face and of the epitaxial layer. 6 1 <6 2 <6 3 , The inclination an- the values of 80 became small in the cases of Figs. 10(a)
gle of the substrate surface due to chemical etching is in the and 10(c) when the lattice mismatch became small.
range of 6 1,
A growth model that can explain the inclination of the
{100} plane of the epitaxial layer in the case of Fig.
10(a) is shown in Fig. 11. In this model, the surface of
has a curvature because of chemical etching, the lateral the substrate is assumed to be composed of {lOO} steps.
growth direction of Ga"In1_~s is shown in Figs. 8 (a) and If the epitaxial crystal grows by the lateral advance of
8(b). The epitaxial crystal inclines with respect to the

12 June 2024 11:46:50

easy-growth direction, and the epitaxial layer is divided
growth direction
into two regions. A boundary lies in the direction of • lO : lattice mismatch
(110) through a point on the real {100}. If this is correct (om L···-··
we can prevent the generation of the boundary on the
{lOO} surface by polishing the substrate surface flat
[Fig. 8(c)] or by inclining the substrate surface by up to
1 from the {100} plane along the (110) direction before
0

chemical etching [Fig. 8(d)]. In both cases, if this argu-

ment is correct the growth direction becomes unique on
the whole surface, and no boundary is generated. Actu-
ally, no boundary was seen in these substrates
V -I - .........

experimentally.
substrate(100)j I I :mean value of step
The boundary on the convex surface can be observed +length
only by the x-ray topograph. No boundary exists on the epitaxial e/.. --,;-T----·n
surface of the epitaxial layer. But when the substrate layer(100)--,
surface is concave, the boundary appears on the \ ---4
epitaxial layer surface and can be observed by optical .-
f--
microscope. The substrate surface was prepared to
have a grooved configuration as shown in Fig. 9(a). The
two directions of the steps advance downwards to meet
with each other [Fig. 9(b)]' and a clear boundary line is
formed at the center, as shown in Fig. 9 (c). When the
epitaxial layer is grown on the spherical convex surface,
it is known that the facets are formed on the low-index
planes. 12 In this Ga"In1_~s-on-GaAs system, {lOO} fac- tan e/.. =A~
ets were found on the convex spherical substrate sur-
face. But when the substrate surface was a concave
FIG. 11. Model of the {100} interface between GaAs and
sphere, formation of the facets was not observed, and G~Inl-xAs. To simplify the condition, the lattice mismatch
the generation of the boundary line as shown in Fig. 9(d) in the horizontal direction is neglected in this figure. Actually,
was characteristic. This boundary line was found to when the step length 1 is large enough compared to the lattice
consist of an array of grooves by the optical interfer- constant of the epitaxial crystal, this condition will be satis-
ence method, and this seems similar to the grooves fied virtually by the effect of misfit dislocations and lattice
deformation. This model can explain the fact that the inclina-
which appears at the grain boundary when heated for a tion becomes small with the decrease of the lattice mismatch
long time. 13 and the inclination of the substrate surface from {lOO}.

J. Appl. Phys., Vol. 45, No.9, September 1974

 3794 Haruo Nagai: Structure of vapor-deposited Gax Inl-X As 3794

steps, mismatch of the lattice constant may cause in- tration of a certain impurity may cause the difference
clination of the epitaxial layer as shown in this figure. of lattice constant, and it is known that the impurity con-
The mean value of the step length and the lattice mis- centration depends on the orientation16 of the substrate
match are the factors that decide the inclination. The surface even though the epitaxial layers are formed at
values of the inclination of {100} planes estimated from the same condition. Also, the change of the composition
this model lie near the values (15-250") obtained from can be a cause of lattice constant change in these terna-
the experiment. The reason why this phenomenon occurs ry compounds Therefore, the real cause of the change
0

only in the (110) and not in (110) is not explained by this of the lattice constant shown in Table I is not know yet.
model. This anisotropy between the (110) and (flO) di- Research is in progress concerning this point.
rections may be related to that observed in the bending 6
and misfit dislocations 14 of this heteroepitaxial system. V. CONCLUSIONS
When the inclination of the substrate surface from {100}
(0 In the Ga)n1_xAs-on-GaAs heteroepitaxial system,
becomes large [Fig. 10(c)], the mean length of the step
the crystal lattice of the epitaxial layer is deformed be-
becomes small. If this value is smaller than the dis-
cause of the lattice mismatch.
tance between the misfit dislocations which is expected
from the misfit factor, the lattice mismatch will not be (Ii) Besides the lattice deformation, the inclined
relieved sufficiently on each step. In this case, the sub- growth of the epitaxial layer was observed. (iii) The
grain boundarylike structure 15 is thought to be stable at crystal orientation of the substrate surface strongly af-
the interface. This will result in the more downward fects the phenomena mentioned above.
slope of the epitaxial layer as shown in Fig. 10(c). The
inclination of the substrate surface along (flO) or other ACKNOWLEDGMENTS
directions does not affect this phenomenon when it is
The author wishes to thank Dr. Y. Mizushima for
small (within about 1°), but the effect appears at a large
helpful discussions and suggestions. The author also
inclination angle. The same relation as in Fig. 10(c) is
wishes to thank T. Hattanda for his useful discussions.
observed in this case.
As has already been mentioned, tilting of the epitaxial
layer on {111} substrate was not initially observed. lC. J. Nuese and R. E. Enstrom, IEEE Trans. Electron
When the inclination of the substrate surface from {111} Devices ED-19, 1068 (1972).
is small (within about 1°), the epitaxial layer grows par- 2B. F. Williams, Appl. Phys. Lett. 14, 273 (1969).

12 June 2024 11:46:50

3C. Hilsum, Proceedings of European Semiconductor Device
allel to the substrate crystal. When the inclination of the Research Conference, Munich, 1971, pp, 77-85 (unpublished).
substrate surface from {lll} becomes large, the epi- 4R. W. Conrad, P. L. Hoyt, and D. D .. Martin, J. Electro-
taxial layer grows inclined (see Table III), but the in- chern. Soc. 114, 164 (1967).
clination direction of the epitaxial layer is opposite to 5G.A. Antypas, J. Electrochem. Soc. 117, 139:~ (1970).
6H. Nagai, J. Appl. Phys. 43, 2454 (1972).
the case of {100} substrate. The stacking-fault shadows
7T. Hattanda and A. Takeda, Jpn. J. Appl. Phys. 12, 1104
were not distinct on the {llO} and {lll} samples. There- (1973) .
fore, the discussion of the growth mechanism is limited By. Tarui, Y. Komiya, and Y. Harada, J. Electrochem.
thus far on the {100} surface. Soc. 118, (1971).
9E . F. Hockings, 1. Kudman, T.E. Seidel, C.M. Schmeltz,
Also, it was found that the degree of the deformation andE.F. Stegmeier, J. Appl. Phys. 37,2879(1966).
and the lattice constant of the epitaxial crystal were l~.E. Enstrom, D. Richman, M.S. Abrahams, J_R. Appert,
different on these inclined substrate surfaces [in the D. G. Fisher, A. H. Sommer, and B. F. Williams, Proceed-
case of Fig. 1O(c)] compared to the epitaxial layer on ings of the Third International Symposium on GaAs and
Related Compounds, Aachen, 1970, pp. 30-40 (unpublished).
{100} and {lll} on the same substrate. The effect of the
1!H. Sunami, T. Terasaki, N. Miyamoto, and J. Nishizawa,
temperature and gas flow were checked and found to be J. Appl. Phys. 40,4670 (1969).
neglected. Table I shows this fact in the case of {100} 12H. A. Allen, J. Electrochem. Soc. 117, 1417 (1970).
substrate. From the subgrain boundarylike model pro- 13R . A. Swalin, Thermodynamics of Solids (Wiley, New York,
posed above, the decrease of the degree of deformation 1962), pp. 206-209.
14M. S. Abrahams, J. Blanc, and C. J. Buiocchi, Appl. Phys.
on the inclined surface is reasonable. But the decrease
Lett. 21, 185 (1972).
of the lattice constant cannot be explained. There are 15J. Friedel, Dislocations (Pergamon, London, 1964),
some possibilities that can explain the difference of lat- pp. 279-281.
tice constant. For example, a difference of the concen- 16J.V. Dilorenzo, J. Cryst. Growth 17, 189(1972).

J_ AppL Phys., VoL 45, No.9, September 1974