UC Santa Barbara
UC Santa Barbara Previously Published Works
Title
High-quality InAsyP1-y step-graded buffer by molecular-beam epitaxy
Permalink
https://escholarship.org/uc/item/6327m3jq
Journal
Applied Physics Letters, 82(19)
ISSN
0003-6951
Authors
Hudait, M K
Lin, Y
Wilt, D M
et al.
Publication Date
2003-05-01
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
High-quality InAs
y
P
1À
y
step-graded buffer by molecular-beam epitaxy
M. K. Hudait and Y. Lin
Department of Electrical Engineering, The Ohio State University, Columbus, Ohio 43210
D. M. Wilt
NASA Glenn Research Center, Cleveland, Ohio 44135
J. S. Speck
Department of Materials Science, University of California, Santa Barbara, California 93106
C. A. Tivarus, E. R. Heller, and J. P. Pelz
Department of Physics, The Ohio State University, Columbus, Ohio 43210
S. A. Ringel
a)
Department of Electrical Engineering, The Ohio State University, Columbus, Ohio 43210
共Received 11 November 2002; accepted 13 March 2003兲
Relaxed, high-quality, compositionally step-graded InAs
y
P
1⫺y
layers with an As composition of
y⫽ 0.4, corresponding to a lattice mismatch of ⬃1.3% were grown on InP substrates using
solid-source molecular-beam epitaxy. Each layer was found to be nearly fully relaxed observed by
triple axis x-ray diffraction, and plan-view transmission electron microscopy revealed an average
threading dislocations of 4⫻ 10
6
cm
⫺ 2
within the InAs
0.4
P
0.6
cap layer. Extremely ordered
crosshatch morphology was observed with very low surface roughness 共3.16 nm兲 compared to
cation-based In
0.7
Al
0.3
As/In
x
Al
1⫺ x
As/InP graded buffers 共10.53 nm兲 with similar mismatch and
span of lattice constants on InP. The results show that InAs
y
P
1⫺ y
graded buffers on InP are
promising candidates as virtual substrates for infrared and high-speed metamorphic III–V devices.
© 2003 American Institute of Physics. 关DOI: 10.1063/1.1572476兴
InAs
y
P
1⫺ y
alloys are of interest for both infrared opto-
electronic and high-speed electronic device applications due
to their wide range of band-gap energies from 0.36 eV to
1.35 eV and the large band offset energies possibly using
InAs
y
P
1⫺ y
/In
x
Ga
1⫺ x
As heterostructures grown on InP.
1–3
InAs
y
P
1⫺ y
alloys are also of interest for compositionally
graded buffer applications, where the span of lattice con-
stants between InP and InAs provides the opportunity for
generating ‘‘virtual substrates’’ on InP to support a wide va-
riety of lattice-mismatched devices based on In
x
Ga
1⫺ x
As,
In
x
Al
1⫺ x
As, and InAs
y
P
1⫺ y
. This is currently being ex-
plored for thermophotovoltaic 共TPV兲 devices based on
In
x
Ga
1⫺ x
As, where the band gaps required for optimal TPV
system conversion efficiencies in the range of 0.5–0.6 eV
necessitate In
x
Ga
1⫺ x
As compositions (x⫽ 0.69–0.81) that
generate a significant lattice mismatch with respect to the InP
substrate.
4–7
The use of an anion 共group-V兲-based alloy, such
as InAs
y
P
1⫺ y
for compositionally graded buffers, compared
with more common graded buffer alloy choices, such as
In
x
Ga
1⫺ x
As and In
x
Al
1⫺ x
As, which can also bridge the lat-
tice constant mismatch between active device layers and the
InP substrate, offers a potential advantage since control of
the growth rate 共indium flux兲 is decoupled from control of
the layer composition 共As:P flux ratio兲. The addition of the
group-V sublattice, as an independently controlled variable,
has the effect of widening the parameter space for the growth
of such graded buffers that is otherwise constrained for cat-
ion 共group-III兲-based graded buffers where growth rates and
compositions are both dictated by the group-III sources. This
is particularly advantageous for solid-source molecular-beam
epitaxy 共MBE兲 growth since optimizing the group-III fluence
with respect to both composition and growth rate is ex-
tremely time consuming and would require substantial
growth interruptions that may compromise interface quality.
In this letter, we report the growth of high-quality relaxed
InAs
y
P
1⫺ y
step-graded buffers by solid-source MBE that
show great promise for virtual substrates applications.
InAs
y
P
1⫺ y
compositionally step-graded 共four steps兲 lay-
ers with As mole fractions 共y兲 from 0.05 to 0.40 were grown
on 共100兲 semi-insulating InP substrates using solid-source
MBE equipped with valved cracker sources for arsenic and
phosphorus. Substrate oxide desorption was done at 510 °C
under a phosphorus overpressure of ⬃1⫻ 10
⫺ 5
Torr, which
was confirmed by the observation of a strong (2⫻ 4) reflec-
tion high-energy electron diffraction 共RHEED兲 pattern, in-
dicative of a clean 共100兲 InP surface. An initial 0.2
m thick
undoped InP buffer layer was then deposited to generate a
smooth surface at ⬃485 °C under a stabilized P
4
flux prior to
the growth of InAs
y
P
1⫺ y
step-graded buffers. After InP
growth, the P
4
flux was reduced to the required value for the
InAs
y
P
1⫺ y
growth and the As valve was opened before in-
troducing In into the growth chamber. The exposure time of
As on the InP surface was minimized in order to avoid the
formation of an InAsP interlayer due to As–P exchange on
the InP surface.
8,9
For all InAs
y
P
1⫺ y
layers, the growth rate
was 0.75 monolayers/s, as determined by RHEED intensity
oscillations at a constant substrate temperature of 485 °C
controlled by a pyrometer-based feedback control system.
The first three undoped step-graded layers were each grown
to a thickness of 0.4
m, followed by a 1.7
m thick n-type
共Si-doped兲 InAs
0.4
P
0.6
layer with n⬃3⫻ 10
16
cm
⫺ 3
for char-
a兲
Author to whom correspondence should be addressed; electronic mail:
ringel@ee.eng.ohio-state.edu
APPLIED PHYSICS LETTERS VOLUME 82, NUMBER 19 12 MAY 2003
32120003-6951/2003/82(19)/3212/3/$20.00 © 2003 American Institute of Physics
Downloaded 09 May 2003 to 128.111.74.212. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
acterization purposes. The total mismatch between the cap
layer and the InP substrate is ⬃1.3%.
Figure 1 shows a cross-sectional transmission electron
microscopy 共XTEM兲 image of a typical InAs
0.4
P
0.6
/
InAs
y
P
1⫺ y
/InP step-graded buffer structure. The composi-
tions shown in Fig. 1 were determined using triple axis x-ray
diffraction 共described below兲. The image of Fig. 1 shows a
high contrast at the graded buffer layer interfaces due to
misfit dislocations with no threading dislocations 共TDs兲 ob-
servable in the 1.7
m thick InAs
0.4
P
0.6
cap layer using
XTEM. Hence, to accurately quantify the dislocation density,
plan-view transmission electron microscopy 共TEM兲 mea-
surements were performed, the results of which are shown in
Fig. 2. By considering several fields of view, the average TD
density 共TDD兲 in the relaxed InAs
0.4
P
0.6
cap was found to be
4⫻ 10
6
cm
⫺ 2
. It should be noted that both etch pit density
共EPD兲 measurements using AgNO
3
:CrO
3
:HF:H
2
O 共A–B
etch兲 and electron-beam-induced current measurements on
InGaAs p-n junctions grown on the InAsP buffer were also
performed. These measurements revealed matching values of
⬃1⫻ 10
5
cm
⫺ 2
for EPD and dark spot density, respectively,
in substantial disagreement with the plan-view TEM results
and significantly underestimating the true TDD value. This
exemplifies the general difficulty in quantifying TDD values
in high-quality relaxed buffers.
The relaxation state of the InAs
y
P
1⫺ y
graded buffer
structure was evaluated using high-resolution triple axis
x-ray diffraction measurements. Figure 3 shows reciprocal
space maps 共RSMs兲 for the 共004兲 and 共224兲 reflections. From
the RSM in Fig. 3共a兲, the diffraction intensity maximum for
each layer in the buffer is almost centered on the substrate
reciprocal lattice point 共along the vertical line drawn here兲,
indicating minimum lattice tilt with respect to the substrate.
For the asymmetric 共224兲 reflection in Fig. 3共b兲, the intensity
contours corresponding to the step-graded buffer and the fi-
nal 1.7
m layer makes an angle of ⬃32° with respect to the
substrate reciprocal lattice intensity contours indicating that
the material is almost fully relaxed, since the angle between
共004兲 and 共224兲 is ⬃35°. To further quantify the relaxation of
each layer, the lattice parameters in the growth plane a
储
, and
in the growth direction a
⬜
, were determined. The relaxed
lattice constant, a
layer
and the relaxation, R of the layers were
evaluated using
10
FIG. 1. Cross-sectional TEM image of InAs
0.4
P
0.6
structure grown on 共100兲
InP substrate using InA
y
P
1⫺ y
step-graded buffers.
FIG. 2. Plan-view TEM micrograph of InAs
0.4
P
0.6
layer grown on step-
graded InAs
y
P
1⫺ y
buffers.
FIG. 3. 共a兲 Symmetric 共004兲 and 共b兲 asymmetric 共224兲 reciprocal space
maps of InAs
0.4
P
0.6
layer grown on InP substrate using a four-step
InAs
y
P
1⫺ y
layer, showing that all of the layers are almost fully relaxed. Q
x
coordinates are linked to the lattice parameter parallel to the layer plane a
储
by Q
x
⫽ 2/(
冑
2a
储
), Q
y
coordinates are linked to the lattice parameter per-
pendicular to the layer plane a
⬜
by Q
y
⫽ 2/a
⬜
. Q
x
and Q
y
are expressed in
terms of reciprocal lattice units. Here, is the x-ray wavelength.
3213Appl. Phys. Lett., Vol. 82, No. 19, 12 May 2003 Hudait
et al.
Downloaded 09 May 2003 to 128.111.74.212. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
a
layer
⫽
共
2C
12
a
储
⫹ C
11
a
⬜
兲
/
共
2C
12
⫹ C
11
兲
, 共1兲
R⫽
共
a
储
⫺ a
o
兲
/
共
a
layer
⫺ a
o
兲
. 共2兲
In these expressions, a
o
is the InP substrate lattice parameter,
and C
11
and C
12
are the elastic constants of each ternary
InAs
y
P
1⫺ y
layer obtained using Vegard’s law and the elastic
constants of InP and InAs.
2
From the RSM data in Fig. 3, the
percent relaxation was determined to be ⬎99%, 98%, 97%,
and 93% for the InAs
0.21
P
0.79
, InAs
0.32
P
0.68
, InAs
0.35
P
0.65
,
and InAs
0.4
P
0.6
layers in the buffer stack, respectively, noting
an experimental error of ⫾3%. This level of relaxation is
consistent with the thickness of each layer being well in ex-
cess of their respective critical thicknesses, and indicates ef-
ficient relaxation of the 1.3% total misfit strain.
The most distinctive feature of InAs
y
P
1⫺ y
graded buffer,
however, is the surface morphology. Figure 4共a兲 shows a
typical atomic force microscopy 共AFM兲 image of the relaxed
InAs
0.4
P
0.6
surface. The expected crosshatch morphology that
is characteristic of strain relaxation using compositionally
graded buffers
11–13
is clearly evident. However, compared to
graded buffers consisting of cation-based grades on InP such
as In
x
Al
1⫺ x
As 共Ref. 14兲 or In
x
Ga
1⫺ x
As,
7
grown to span the
identical range of lattice constant and strain on InP, the
InAs
y
P
1⫺ y
surface is far superior with respect to root-mean-
square 共rms兲 roughness, peak-to-valley height, and unifor-
mity. For comparison, Fig. 4共b兲 shows an AFM image of the
surface of an In
x
Al
1⫺ x
As graded buffer stack comprised of
In
0.7
Al
0.3
As/In
x
Al
1⫺ x
As/InP, which incorporates almost the
same total misfit 共1.2%兲 over the identical range of lattice
constants, was found to be greater than 80% relaxed with a
TDD less than 10
7
cm
⫺ 2
. A detailed investigation comparing
the structural properties of these buffers is beyond the scope
of this letter and is the subject of a separate publication.
14
Nevertheless, it is clear that the InAsP buffer results in much
more ordered crosshatch and a far less granular background
superimposed over the crosshatch. Analyzing the AFM data
reveals a rms roughness that is more than three times lower,
3.16 nm, for the InAsP graded buffer as opposed to more
than 10.53 nm for the graded InAlAs structure. The peak-to-
peak roughness difference is even more dramatic due to the
poor uniformity for the InAlAs structure. The vastly im
proved surface morphology for the InAsP graded buffer is
believed to be due to advantages of grading the mole fraction
of the group-V sublattice, which neither influence growth
rate nor require temperature changes for MBE growth, hence
providing an extra degree of freedom compared to grading
on the group-III sublattice. Detailed investigations on this
comparison and reporting of device performance as a func-
tion of graded buffer type are the subjects of forthcoming
publications.
In conclusion, relaxed high-quality compositionally step-
graded InAs
y
P
1⫺ y
layers with As compositions of y⫽ 0.4,
corresponding to a lattice mismatch of ⬃1.3% were grown
on InP substrates using solid-source MBE. Plan-view TEM
revealed an average TDD of 4⫻ 10
6
cm
⫺ 2
. An extremely
ordered crosshatch morphology was observed with very low
surface roughness compared to cation-based graded buffers
with a similar mismatch on InP. Hence, MBE-grown
InAs
y
P
1⫺ y
step-graded buffers hold great promise as a vir-
tual substrate technology for InP-based infrared devices.
This work is supported in part by a National Science
Foundation Focused Research Group 共FRG兲 grant 共No.
DMR-0076362兲.
1
M. Beaudoin, A. Bensaada, R. Leonelli, P. Desjardins, R. A. Masut, L.
Isnard, A. Chennouf, and G. L’Espe
´
rance, Phys. Rev. B 53, 1990 共1996兲.
2
I. Vurgaftman, J. R. Meyer, and L. R. Ram-Mohan, J. Appl. Phys. 89,
5815 共2001兲.
3
R. P. Schneider Jr. and B. W. Wessels, J. Electron. Mater. 20, 1117 共1991兲.
4
M. W. Wanlass, J. J. Carapella, A. Duda, K. Emery, L. Gedvilas, T. Mo-
riarty, S. Ward, J. D. Webb, X. Wu, and C. S. Murray, AIP Conf. Proc.
460, 132 共1999兲.
5
N. S. Fatemi, D. M. Wilt, R. W. Hoffman, Jr., M. A. Stan, V. G. Weizer, P.
P. Jenkins, O. S. Khan, C. S. Murry, D. Scheiman, and D. J. Brinker, AIP
Conf. Proc. 460, 121 共1999兲.
6
T. J. Coutts, Renewable and Sustainable Energy Rev. 3,77共1999兲.
7
M. W. Wanlass, J. S. Ward, K. A. Emery, M. M. Al-Jassim, K. M. Jones,
and T. J. Coutts, Sol. Energy Mater. Sol. Cells 41, 405 共1996兲.
8
G. Hollinger, D. Gallet, M. Gendry, C. Santinelli, and P. Viktorovitch, J.
Vac. Sci. Technol. B 8,832共1990兲.
9
J. M. Moiser, M. Bensoussan, and F. Honzay, Phys. Rev. B 34,2018
共1986兲.
10
M. Fatemi and R. E. Stahlbush, Appl. Phys. Lett. 58,825共1991兲.
11
E. A. Fitzgerald, Y.-H. Xie, M. L. Green, D. Brasen, A. R. Kortan, J.
Michel, Y.-J. Mii, and B. E. Weir, Appl. Phys. Lett. 59,811共1991兲.
12
E. A. Fitzgerald, Y.-H. Xie, D. Monroe, P. J. Silverman, J. M. Kuo, A. R.
Kortan, F. A. Thiel, and B. E. Weir, J. Vac. Sci. Technol. B 10, 1807
共1992兲.
13
J. S. Speck, M. A. Brewer, G. Beltz, A. E. Romanov, and W. Pompe, J.
Appl. Phys. 80, 3808 共1996兲.
14
M. K. Hudait, Y. Lin, C. A. Tivarus, E. R. Heller, J. P. Pelz, J. S. Speck,
and S. A. Ringel 共unpublished兲.
FIG. 4. AFM images from the surface of 共a兲 InAs
0.4
P
0.6
layer grown on InP
substrate using a four-step InAs
y
P
1⫺ y
layer and 共b兲 In
0.7
Al
0.3
As layer grown
on InP substrate using a five-step In
x
Al
1⫺ x
As layer. The scan area and the
rms roughness were 40
m⫻ 40
m 3.16 nm, and 10.53 nm, respectively.
Total mismatch is 1.2% and layers are more than 80% relaxed.
3214 Appl. Phys. Lett., Vol. 82, No. 19, 12 May 2003 Hudait
et al.
Downloaded 09 May 2003 to 128.111.74.212. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
