ION BEAM STUDIES OF STRAINS/DEFECTS IN
SEMICONDUCTOR MULTILAYERS
Anand P. Pathak1,3,∗ , S.V.S. Nageswara Rao1, D.K. Avasthi2,
Azher M. Siddiqui2, S.K. Srivastava2, F. Eichhorn3,
R. Groetzschel3, N. Schell3 and A. Turos4
1
School of Physics, University of Hyderabad, Central University (PO), Hyderabad 500 046, India.
2
Nuclear Science Centre, Post Box No. 10502, Aruna Asaf Ali Marg, New Delhi 110 067, India
3
Institute of Ion Beam Physics and Materials Research, FZR Rossendorf, 01314 Dresdon, Germany
4
Institute of Electronic Materials Technology, Ul. Wolczynska 133, 01-919, Warszawa, Poland
High Resolution X-Ray Diffraction (HRXRD) studies have been performed to study the effects
of Swift Heavy Ion (SHI) irradiation on In0.53Ga0.47As/InP lattice matched superlattice. Sample under the
study have been irradiated using 130 MeV Ag ions to two different fluences (5 x 1012 ions/cm2 & 5 x 1013
ions/cm2). Pre and post irradiation annealing (RTA) studies have also been performed. A finite tensile
lattice strain has been introduced due to the intermixing caused by ion irradiation and/or annealing
processes. The superlattice period is found to increase due to ion irradiation and annealing processes. The
superlattice structure used in this work was grown by Metal Organic Chemical Vapor-phase epitaxial
Deposition (MOCVD). Ion channeling measurements were earlier performed on the low dose sample.
Channeling and HRXRD measurements show good crystalline/interface quality of pristine and processed
samples. HRXRD also indicates the existence of the sharp boundaries across the superlattice interfaces. A
decreasing modulation of the intensities of satellite peaks is caused by a gradual diffusion of individual
layer interface. Such effects are more prominent for the irradiated and annealed sample.
Such structures made of III-V compound
semiconductors have more applications in the
optoelectronics because of their direct bandgap
nature. The multilayers with a small lattice
mismatch leading to a tensile or compressive
strain in the alternating layers are called strained
layer superlattice. With the advent of epitaxial
growth techniques, it is now possible to grow
crystals with mono layer precision. Atoms
deposited on a substrate take positions
corresponding to the potential minima of the
lattice sites. Hence in the strained-layer epitaxi,
despite the difference in substrate and deposit
lattice parameters, deposit atoms are constrained
to the substrate interatomic spacing in the plane
of the interface. Corresponding change occurs in
the perpendicular lattice parameter due to the
Poisson effect.
The strain in the epilayer due to such
tetragonal distortion improves the device
performance [4-6] and is a parameter for
tailoring the band structure and other relevant
quantities. The thickness and composition of the
epilayer decides the bandgap of SLS, which can
INTRODUCTION
Ion beam mixing experiments were
performed to introduce strain in an initially
lattice matched (In0.53Ga0.47As / InP) multilayer
system. HRXRD has been employed to study
these mixing effects. This experiment is a part of
our series of experiments [1-3] to study ion
beam based methods to measure and engineer
the bandgap/strain of semiconductor multilayers
for the integration of optoelectronic circuits.
Semiconductor
superlattices
have
potential device applications [4-7] for high
performance detectors, high speed and high
frequency digital and analog circuits because
they offer precise control over the states and
motions of charge carriers. Band structure of
these materials can be tailored to get required
electronic and optoelectronic properties. These
are basically multilayers with different bandgaps
on either side of each interface, and therefore
called as heterostructures. The band structure of
these materials depends on the band structure of
each layer and the band offsets at each interface.
∗
Corresponding author. Fax: +91-40-23010181 / 23010227 / 23010120, email: appsp@uohyd.ernet.in
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
593
be tuned by the ion irradiation [7]. Spatial
bandgap tuning of these materials is necessary to
integrate the opto-electronic circuits because
different devices require different bandgap. To
tune such important parameters of these
structures, post growth techniques like ion beam
mixing are found to be superior to the
manipulation during the growth process [7].
Swift Heavy Ion (SHI) beam mixing is more
suitable because of the advantage that the
intermixing can be confined to a narrow region
at the interface as against the lateral straggling
effects in low energy ion implantation. Here we
study the SHI irradiation and /or annealing
effects on the lattice matched In.53Ga.47As/InP
multilayer system. Next section gives the
experimental details while the following two
sections give discussion and conclusions of the
results obtained.
reflection <006> (InP) at ROBL Material
Research Station, Grenoble. The wavelength of
radiation was determined to be ~ 0.15nm from
the angular position of Bragg lines.
RESULTS AND DISCUSSION
In our earlier work [1] we studied the
low fluence sample ( i.e P523 & P523I) using
RBS/Channeling measurements. RBS spectra
showed ion beam induced inter diffusion of In,
Ga, As & P across the interfaces and RBS/C
showed that the irradiation can induce finite
tensile strain. Good reduction in RBS/C spectra
of irradiated sample indicates that the sample
structure was not spoiled by irradiation.
However we could not determine the exact value
of strains in that experiment. Here these values
are measured very accurately using HRXRD.
Effects of annealing (RTA) have also been
studied here.
TABLE 1: Measured strain and thickness values
Sample
P523 U
P523 URTA
P523 I
P523 IRTA
523 I2
523 I2RTA
FIGURE 1: HRXRD spectra of low dose samples
zoomed around the substrate peak, maintaining the
order given in label box; inset: complete interference
pattern of P523U.
s (nm)
(Nominal 30nm)
28.90
29.08
29.01
28.93
29.69
28.59
(∆d/d)⊥
-1.10 x 10-4
-1.44 x 10-4
-3.46 x 10-4
-3.59 x 10-4
-6.92 x 10-4
Fig.1 shows the HRXRD spectra of low
dose samples. These spectra are displaced
(Shifted upwards) on the intensity axis for
clarity. The inset shows the complete
interference pattern of P523U. Similar spectra of
high dose samples are shown in fig. 2. Very high
number of satellite orders in the HRXRD spectra
shown in these figures, indicate that the
interfaces are very sharp and the boundaries are
almost rectangular in shape. Also indicated fact
is that these interfaces remain considerably
sharp even after irradiation. The width of the
substrate peak varies with the material treatment.
The narrowest peak (More Lorentzian like) is
observed for the as grown sample and the
broadest peak (More Gaussian like) is observed
for the high fluence sample. This width is
reduced by RTA procedure but did not reach the
as grown value. This is caused mainly due to
the implantation damage in the substrate region
(about 13µ deep). The reduction in the intensity
of satellite peaks shows the interdiffusion of
elements across the interfaces. Superlattice
period and strain values are obtained using the
simple HRXRD formulas [2,3] based on Bragg’s
law. Measured strain values and other crystal
parameters are given in table 1. The dependence
of strain on the ion fluence (after RTA) is shown
EXPERIMENTAL
Ten periods of [In.53Ga.47As(15nm)
/InP(15nm)] were grown on an InP substrate
using MOCVD facility at Warsaw, Poland [1].
This sample is then cut in to three parts with one
part kept unirradiated (P523U). The other parts
were irradiated by 130 MeV Ag ions delivered
from 16MV Pelletron accelerator at Nuclear
Science Centre, New Delhi with two different
fluences 5 X 1012 ions/cm2 (P523I) and 5 X
1013 ions/cm2 (P523I2). Then each of these three
samples was again cut in to two parts and one
part was annealed (Rapid Thermal Annealing
(RTA)) at 450oC for 90 Sec. in N2 atmosphere
using the RTA facility at Dresden, Germany.
Annealed samples are referred with a tag “RTA”
at the end of the sample name. Including the
pristine sample in every set, we classify all these
samples into two sets namely low dose samples
(P523U, URTA, I & IRTA) and high dose
samples (P523U, I2 & I2RTA). All these
samples were characterized by HRXRD near the
594
TABLE2: Structure details of the superlattice found from HRXRD (Composition)
I n te n s it y
(a r b . u .)
Sample Name &
Structure
InxGa1-xAsyP1-y
InP
InxGa1-xAsyP1-y
InxGa1-xAs
InxGa1-xAsyP1-y
InP
1 0
7
1 0
6
1 0
5
1 0
4
1 0
3
1 0
2
1 0
1
Structure
Details
x 10
Layers
Substrate
P523U
X
Y
0.248
0.349
0.258
0.180
0.513
0.407
0.631
-
In P ( 0 0 6 ) P 5 2 3 U
1 0 0
- 1 5 0 0 0- 1 0 0 0 0 - 5 0 0 0
R e la t iv e
0
in c id e n c e
5 0 0 0
a n g le
θ
1 0 0 0 0 1 5 0 0 0
( a rc s e c )
FIGURE 2(a)
FIGURE 2(b)
P523I2
x
y
0.321
0.442
0.407
0.124
0.548
0.233
0.499
-
P523I2RTA
x
y
0.668
0.471
0.668
0.348
0.516
0.524
0.544
-
Simple irradiation changes the strain but the
interface quality could be improved after
annealing as it can be observed by the
asymmetry in the substrate peak. However
annealing alone doesn't change the strain value
significantly. Hence a tensile strain is induced in
an initially lattice matched system without loss
of the sample quality.
For a detailed understanding we have
analyzed the HRXRD data of high dose samples
using a computer code RADS Mercury. The
thicknesses and chemical content of all the
layers are taken as free parameters of the fitting
routine. Furthermore, a diffuse scattering profile
(due to uncorrelated lattice defects) is added to
the pure dynamic x-ray diffraction. Fig.2 shows
the HRXRD spectra of these high dose samples
along with proper fits. This analysis suggests
that the diffusion process will form a thin
InxGa1-xAsyP1-y on the top of every layer. This
effect is observed even in the pristine sample,
however the effect is more in irradiated and/or
annealed samples. Table 2 shows structure of the
sample which is obtained by fittings and table 3
shows the thickness of each layer. The thickness
(superlattice period) values given in table 3 are
little different from the values that are given in
table 1. It is because the superlattice was
considered as a homogeneous structure with
perfect nominal structure while calculating these
values from the interference pattern shown in
fig.1. These tables suggest the fact that the
diffusion is more effective in irradiated and
annealed sample (P523I2RTA).
FIGURE 2(c) FIGURE 2: HRXRD data with proper
fitting curves of high dose samples; a) P523U, b)
P523I2 and c) P523I2RTA.
in figure 3. As a general trend [2] the period of
the pristine sample is found to be less than that
of irradiated samples. The mixing effects are
more prominent for ion beam processed (high
fluence) and annealed sample as expected.
FIGURE 3: Fluence dependence on strain (after RTA)
595
TABLE3: Structure details of the superlattice found from HRXRD (Thickness)
Sample
treatment
P523U
P523I2
P523I2RTA
Multilayer
period(λ)
(Ao)
289.8
289.9
289.1
dInP
(Ao)
dInGaAs
(Ao)
151.5
152.7
137.8
131.3
121.6
116.0
dInGaAsP
above InP
(Ao)
6.8
8.3
7.9
dInGaAsP
below InP
(Ao)
0.2
7.3
27.4
Sum thickness of
diffusion layers / λ
0.024
0.054
0.122
REFERENCES
CONCLUSION
1.
It is shown that the irradiation can induce a tensile
strain in an initially lattice matched system with out
loss of crystalline/interface quality of the samples.
Broadening in the substrate peak is due to the
implantation damage in the substrate region,
which could be annealed out up to an extent by
RTA. Otherwise the basic structure of the
lattice is invariant under irradiation and
annealing process. Thickness, chemical
composition and strain values of each layer are
presented in tabular forms. Combining these
results with our earlier work [1,2] we conclude
this paper with more general conclusions. The
common trend in all the samples indicates the
gradual diffusion of In from surface and the
migration of Ga or As like atoms to the surface
regions due to the SHI irradiation and/or
annealing processes. The compressive strain is
found to decrease in the initially compressive
strained samples and tensile strain is induced
in an initially lattice matched system. General
trend indicates the increase in the superlattice
period after the irradiation.
2.
3.
4.
5.
6.
ACKNOWLEDGEMENTS
SVSNR thanks CSIR for SRF. APP & DKA
thank DRDO for supporting this work through
a research project. APP thanks Dr. B.
Schmidt for RTA work at FZR Rossendorf &
Dr R. Groetzschel for hospitality.
7.
596
Pathak A.P, Nageswara Rao S.V.S,
Siddiqui A.M, Lakshmi G.B.V.S,
Srivastava S.K, Ghosh S, Bhattacharya D,
Avasthi D.K, Goswami D.K., Satyam
P.V, Dev B.N and Turos A, Nucl. Insrt.
Meth. B, 193, 319 (2002).
Ion irradiation effects and ion beam
studies of semiconductor multilayers,
Nageswara Rao S.V.S, Pathak A.P,
Avasthi D.K, Muralidharan R, Tiwari U,
Srinivasan T, Metha S.K, Jain R.K,
Eichhorn F, Groetzschel R and Schell N,
submitted to World Scientific, presented in
international conference PSI 2002 at Puri,
India, March,( 2002).
Siddiqui A.M, Nageswara Rao S.V.S,
Pathak A.P, Kulkarni V. N,.Kesav Murthy
R, Eric Williams, Daryush Ila, Claudiu
Muntele, Chandrasekaran K.S and Arora
B. M., J. Appl. Phys., 90, 2824 (2001).
Charabonneu S, Poole P.J, Pive P.G,
Buchanan M, Goldberg R.D and Mitchell
I.V, Nucl. Inst. and Meth. (B) 106, 457
(1995).
Wan J.W, Thompson D.A and Simmons
J.G, Nucl. Instr. Meth. B, 106, 461 (1995).
HofsaB V, Kuhn J, Kaden C, Harle V,
Bolay H, Scholz F, Schweizer H, Hillmer
H, Losch R and Schlapp W, Nucl. Instr.
Meth. B, 106, 471 (1995).
Xia W, Hsu S.N, Han C.C, Papert S.A,
Zhu B, Cozzolino C, Yu P.K.L and Lau
S.S, Nucl. Instrum. Meth. B, 59/60, 491
(1991).