Long wavelength pseudomorphic InGaPAsSb type-I and type-II active layers grown on
GaAs
S. R. Johnson, P. Dowd, W. Braun, U. Koelle, C. M. Ryu, M. Beaudoin, C.-Z. Guo, and Y.-H. Zhang
Citation: Journal of Vacuum Science & Technology B 18, 1545 (2000); doi: 10.1116/1.591423
View online: http://dx.doi.org/10.1116/1.591423
View Table of Contents: http://scitation.aip.org/content/avs/journal/jvstb/18/3?ver=pdfcov
Published by the AVS: Science & Technology of Materials, Interfaces, and Processing
Articles you may be interested in
Highly tensile-strained, type-II, Ga 1 − x In x As / GaSb quantum wells
Appl. Phys. Lett. 96, 062109 (2010); 10.1063/1.3303821
GaAs-substrate-based long-wave active materials with type-II band alignments
J. Vac. Sci. Technol. B 19, 1501 (2001); 10.1116/1.1386380
Strained InGaAs/GaPAsSb heterostructures grown on GaAs (001) for optoelectronic applications in the
1100–1550 nm range
J. Appl. Phys. 88, 3004 (2000); 10.1063/1.1287233
Observation of direct (type-I) transitions in type-II InGaAs/AlAsSb heterostructures lattice matched to InP grown
by molecular beam epitaxy
J. Vac. Sci. Technol. B 18, 1586 (2000); 10.1116/1.591432
Molecular beam epitaxial growth of InGaAsN:Sb/GaAs quantum wells for long-wavelength semiconductor lasers
Appl. Phys. Lett. 75, 178 (1999); 10.1063/1.124311

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 209.147.144.21 On: Tue, 10 Feb 2015 00:50:22

Long wavelength pseudomorphic InGaPAsSb type-I and type-II active
layers grown on GaAs
S. R. Johnson,a) P. Dowd, W. Braun, U. Koelle, C. M. Ryu, M. Beaudoin, C.-Z. Guo,
and Y.-H. Zhang
Center for Solid State Electronics Research, Department of Electrical Engineering,
Arizona State University, Tempe, Arizona 85287-6206

共Received 10 October 1999; accepted 14 January 2000兲
GaPAsSb type-I quantum wells and symmetric InGaAs/GaPAsSb/InGaAs type-II quantum wells are
proposed as long wavelength active layers for GaAs based optoelectronic devices. Room
temperature photoluminescence from 1.2 to 1.5 ␮m is observed from the type-II quantum wells,
while low temperature photoluminescence at 1.25 ␮m is observed from type-I quantum wells. A
reflection high-energy electron diffraction peak that is extended parallel to the surface, while its
width is perpendicular to the surface is narrow, suggests that spontaneous lateral composition
modulation occurs in the GaPAsSb layers. © 2000 American Vacuum Society.
关S0734-211X共00兲02703-7兴

I. INTRODUCTION
Vertical-cavity surface-emitting lasers 共VCSELs兲 operating at 1.3 and 1.55 ␮m are desirable for a wide range of
applications, such as optical communication and optical interconnects. The development of optical interconnects is important for many applications such as image processing in
synthetic aperture radar, real-time tasking of multisensor
platforms, dynamic target recognition, virtual reality systems, rapid searches of extremely large dynamic databases,
and local area networks. Long wavelength VCSELs that operate at high data rates are anticipated to be key components
in the development of optical interconnects, gigabit Ethernet,
fiber to the home, and optical local area networks.
Hybrid integration of VCSELs and photodetectors with Si
circuitry is a promising low-cost technology for future highspeed free-space optical interconnects. Because of the ever
decreasing operating voltages in Si integrated circuits 共ICs兲,
longer wavelength 共i.e., lower operating voltage兲 VCSELs
will soon be required for integrated optical interconnects. In
a few years Si IC voltages will be below the threshold voltages of the 980 and 850 nm lasers that are now commonly
used in short distance optical communication. Therefore, the
development of longer wavelength 共smaller band gap兲 materials for short distant optical links is critical. The logical
choices for these longer wavelengths are 1.3 and 1.55 ␮m as
these wavelengths are already extensively used in longer
haul optical communications.
There has been considerable interest in developing long
wavelength VCSELs using a number of different approaches
based on InP and GaAs substrates. Fabrication of VCSELs
using the InP/InGaAsP system has several drawbacks when
compared to the GaAs based system. For example, both refractive index contrast and thermal conductivity are superior
in the GaAs based system.1 These two attributes are important for the fabrication of suitable distributed Bragg reflector
共DBR兲 mirrors for VCSELs. A complex approach that overa兲

Electronic mail: shane.johnson@asu.edu

1545

J. Vac. Sci. Technol. B 18„3…, MayÕJun 2000

comes the limitation of InP based DBRs is wafer fusion,
where InP based active regions are bonded to GaAs based
DBRs.2
It can be difficult to find materials with band gaps suitable
for 1.3 ␮m emission that can be grown on GaAs. In one
approach, quantum dot 共QD兲 structures have been developed
that produce photoluminescence 共PL兲 at up to 1.3 ␮m.3 More
recently, 1.31 ␮m edge-emitting QD lasers have been
demonstrated,4 where long 共2.1 mm兲 cavity and high reflectivity facet coating were required to achieve lasing. Room
temperature operation of QD-based VCSELs has also been
achieved,5 but the lasing wavelength was only 1.15 ␮m PL at
wavelengths of up to 1.3 ␮m has also been observed in
GaAsSb/InGaAs bilayer type-II quantum well 共QW兲
samples.6 To date, the longest emission wavelength in GaAsbased VCSEL structures is with a single step growth 1.18
␮m, using a single InGaNAs QW.7 Though it is difficult to
incorporate significant levels of N in InGaAs, a 1.3 ␮m
edge-emitting laser with an InGaNAs QW 共containing 1%
N兲 has been realized.8
To overcome the limitations of InP based DBRs and wafer bonding, and because GaAs is in general a more mature
technology, we are developing novel long-wavelength type-I
and type-II InGaPAsSb QW systems for the GaAs system.
These structures are grown coherently strained on GaAs by
solid source molecular beam epitaxy 共MBE兲. We propose
type-I QWs 共0.00⭐x⭐0.04,
Inx Ga1⫺x Py Asz Sb1⫺y⫺z
0.3⭐y⭐0.5, 0.0⭐z⭐0.2兲 for wavelengths up to 1.3 ␮m and
symmetric
type-II
Inx Ga1⫺x As/GaPy Asz Sb1⫺y⫺z /
Inx Ga1⫺x As QWs 共0.3⭐x⭐0.4, 0.1⭐y⭐0.5, 0.0⭐z⭐0.5兲 for
wavelengths up to 1.55 ␮m. Preliminary results indicate that
1.3 and 1.55 ␮m 共0.95 to 0.80 eV兲 wavelengths can be
reached without exceeding 3% compressive strain levels.
The samples presented here were grown using a VG
V80H solid source MBE system equipped with As, P, and Sb
group-V valved crackers, Ga, In, and Al group-III effusion
cells, Si and GaTe cells for n-type doping, and a Be cell for
p-type doping. This MBE system is outfitted with a 15 keV

0734-211XÕ2000Õ18„3…Õ1545Õ4Õ$17.00

©2000 American Vacuum Society

1545

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 209.147.144.21 On: Tue, 10 Feb 2015 00:50:22

1546

Johnson et al.: Long wavelength pseudomorpic InGaPAsSb

FIG. 1. ⌫ point band gap energy vs strain for coherently strained GaPAsSb
on GaAs. Negative strain values indicate compressive strain. The right-hand
shaded region is direct band gap material while the left-hand shaded region
is indirect band gap material.

reflection high-energy electron diffraction 共RHEED兲 system
with a charge coupled device camera and software for realtime acquisition of RHEED data. This RHEED system has
the ability to phase-lock 共or synchronize兲 the data acquisition
to substrate rotation,9 which allows for several real-time observation modes during substrate rotation. The control of the
group-V element ratio can be problematic when incorporating three group-V materials in a single layer. RHEED is one
of the key tools used to sort out the conditions required for
the optimum growth of these structures.
II. TYPE-I QUANTUM WELLS
Though not obvious, because GaP is an indirect band gap
material with a band gap greater than 2 eV, emission wavelengths up to 1.3 ␮m are achievable using pseudomorphic
共In兲GaPAsSb type-I QWs on GaAs. This is mainly due to the
large band gap bowing parameter of GaPSb. Furthermore,
the small size of the P atom helps to reduce the strain in the
material. However, it is not possible to use GaPSb by itself
because it is an indirect band gap material for the strain
levels required for a two-dimensional growth mode in MBE,
i.e., less than about 3% strain. The novelty of this approach
is that the addition of a small amount of As or InAs to
GaPSb results in a suitable direct band gap active material
for wavelengths up to 1.3 ␮m. The attributes of this system
that make it suitable for GaAs-based lasers are the large
bowing band gap parameter of GaPSb and the large range of
mixed P, As, and Sb compositions where 共In兲GaPAsSb is a
direct band gap material.
The ⌫ point band gap energy versus strain for coherently
strained GaPAsSb on GaAs is shown in Fig. 1. 共Note that in
Fig. 1 the ternaries GaPAs, GaPSb, and GaAsSb ternaries
border the GaPAsSb quaternary region.兲 In the region labeled ‘‘direct band gap’’, the lowest energy band to band
transition is between the ⌫ point of the conduction band and
the ⌫ point of the valence band. These direct band gap com-

1546

FIG. 2. Low temperature photoluminescence spectrum from a strain compensated GaPAsSb type-I quantum well grown on GaAs. A band diagram of
the type-I quantum well is shown in the inset. The 8 nm GaPAs spacers are
1.5% tensilely strained while the 8 nm GaPAsSb QW is 3% compressively
strained.

positions are suitable as active materials for lasers. In the
region labeled ‘‘indirect band gap’’, the lowest energy band
to band transition is an indirect transition that occurs between either the X or the L valley of the conduction band and
the ⌫ point of the valence band. These compositions are not
suitable as active materials for lasers. In the indirect band
gap region, the lowest energy point of the conduction band
occurs at the L valley for compositions nearer GaSb and at
the X valley for compositions nearer GaP.
Most of the bowing parameters used in the calculations
for Fig. 1 were obtained from the literature. However, since
no data is available for the L valley bowing parameter of
GaPSb, the X valley value was used in its place. Measurements of the GaPSb material system indicate that the band
gap bowing parameters for this material are very large, 3.8
eV at the ⌫ point and 2.7 eV for the X valley.10 The bowing
parameters for GaPAs and GaAsSb are much smaller than
those for GaPSb and were obtained from Adachi, Refs. 11
and 12, respectively. Finally Vegard’s Law is assumed in the
strain calculations, i.e., strain is a linear function of mole
fraction.
A schematic band diagram of a strain compensated 1.3
␮m GaPAsSb type-I QW is shown in the inset of Fig. 2. The
8 nm GaPAs spacers are 1.5% tensilely strained and the 8
nm GaP0.37As0.08Sb0.55 QW is 3% compressively strained.
The use of strain compensation layers permits the growth of
multiple unrelaxed QWs that are necessary for the fabrication of VCSELs. Figure 2 shows preliminary low temperature 共4 K兲 PL results for the strain compensated type-I QW
共nominally shown in the inset兲. The excitation density was
250 W/cm2 and the peak of the spectrum occurred at 1.25
␮m. Though room temperature PL was not observed from
this sample an emission wavelength of about 1.3 ␮m is expected at room temperature when the temperature dependence of the band gap is considered. This result indicates that
the GaP0.37As0.08Sb0.55 is a good candidate for a 1.3 ␮m
type-I active material.

J. Vac. Sci. Technol. B, Vol. 18, No. 3, MayÕJun 2000

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 209.147.144.21 On: Tue, 10 Feb 2015 00:50:22

1547

Johnson et al.: Long wavelength pseudomorpic InGaPAsSb

FIG. 3. Room temperature photoluminescence spectra from InGaAs/
GaPAsSb/InGaAs type-II quantum wells grown on GaAs. Varying the composition of the GaPAsSb layer changes the emission wavelength. Each layer
is nominally 3% compressively strained. A band diagram and the electron
and hole wave function of the type-II quantum well are shown in the inset.

III. TYPE-II QUANTUM WELLS
A type-II QW is formed using two semiconductor materials that have a staggered band alignment, where the conduction band and valence band edges of one semiconductor
are both shifted in the same direction relative to the band
edges of the other semiconductor. In the classical picture, the
resulting optical transition in such a structure is perceived as
being spatially indirect because the electrons 共holes兲 are
thought to reside in lowest 共highest兲 part of the well. However, in the more accurate quantum picture where the electron and hole wave functions are considered, the transition is
direct in the sense that it is the spatial overlap of the wave
functions that is important. With this system a smaller effective band gap 共optical transition兲 is obtained at the cost of
wave function overlap. The idea to use a type-II heterojunction for infrared emission was first proposed by Kroemer and
Griffiths.13
In the case of symmetric type-II QWs, a smaller effective
band gap is achieved at the cost of a broader electron wave
function and a narrower hole wave function 共see Fig. 3 inset兲. While for asymmetric type-II QWs the electron and hole
wave functions are spatially offset. In both cases, this results
in a reduction of the electron-hole wave function overlap,
which reduces the oscillator strength when compared to
type-I QWs. However, theoretical calculations show that in
spite of the reduced wave function overlap, the gain levels
required for lasers are attainable in symmetric type-II
QWs.14 For an In0.4Ga0.6As/GaP0.15As0.40Sb0.45 /In0.4Ga0.6As
共2.3 nm–2.3 nm–2.3 nm兲 QW structure embedded in GaAs,
the calculated transition wavelength is close to 1.3 ␮m and
the electron–hole wave function overlap is better than 65%.
Such a large electron–hole wave function overlap is the
foundation of the symmetric type-II QW design. A peak gain
of 104 cm⫺1 is predicted for a carrier density of
1019 cm⫺3, 14 which is not unreasonable for the operation of
VCSELs.

1547

FIG. 4. Reflection high-energy electron diffraction images taken through the
growth cycle of a symmetric InGaAs/GaPAsSb/InGaAs type-II quantum
well: 共a兲 GaAs surface prior to first InGaAs layer, 共b兲 end of the first InGaAs
layer, 共c兲 central GaPAsSb layer, 共d兲 start of second InGaAs layer, 共e兲 end of
second InGaAs layer and 共f兲 start of GaAs overgrowth.

A schematic of the band-edge alignment and the electron
and hole wave functions of the proposed type-II QWs is
shown in the inset of Fig. 3. The band offsets have been
calculated using the model-solid theory15 and the effects of
strain have been included. Though not shown, GaPAs strain
compensation layers can also be used to extend this system
to multiple QWs.
Two symmetric type-II QWs samples 共A and B兲 with different compositions for the center GaPAsSb layer were
grown at 505 °C on semi-insulating GaAs. Each individual
InGaAs and GaPAsSb layer thickness and strain are estimated to be 3 nm and 3%, respectively. Depending on the
composition of the central GaPAsSb layer, the peak of the
emission spectrum can be shifted from 1.2 to 1.5 ␮m 共see
Fig. 3兲. Increasing the Sb to P flux ratio shifts the emission
spectrum to longer wavelengths. Theoretical modeling shows
that these extended wavelengths are attributed to a symmetric type-II QW.
Figure 3 shows the PL spectra for the two samples studied. For sample A, the peak of the PL spectrum is at 1.3 ␮m
with a full-width at half-maximum 共FWHM兲 of 110 meV.
For sample B, the peak wavelength is close to 1.5 ␮m with a
FWHM of 130 meV. These broad linewidths 共in the type-II
structures兲 are believed to be due to band filling,16 fluctuations in the well width and composition,17 and InGaAs well
asymmetry due to indium segregation in the growth direction.
The inhomogeneous contribution to the broadening of the
linewidth in Fig. 3 is due to a Stranski–Krastanov transition
that occurs very close to the completion of growth of each
InGaAs layer 关see RHEED images Figs. 4共a兲–4共b兲 and 4共d兲–
4共e兲兴. High-resolution secondary ion mass spectrometry measurements indicate that the indium is segregating, and is
therefore present in the GaPAsSb layer, while P and Sb incorporate in their intended position. In addition to the ratios
of the three group-V elements, the properties of the
GaPAsSb layer strongly depend on the indium segregation.
The first InGaAs layer forms a rough surface on which the
GaPAsSb layer deposits preferentially smoothing the growth

JVST B - Microelectronics and Nanometer Structures

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 209.147.144.21 On: Tue, 10 Feb 2015 00:50:22

1548

Johnson et al.: Long wavelength pseudomorpic InGaPAsSb

surface 关see Fig. 4共c兲兴 before the second InGaAs layer is
deposited.
RHEED results indicate that the surface is disordered
when the group-V fluxes are switched back to solely arsenic
after the growth of the GaPAsSb layer 关see Figs. 4共c兲 and
4共d兲兴. Surface order can be inferred from the intensity of the
RHEED patterns, where a well ordered surface results in
sharp, intense diffraction peaks with a low background 关see
Fig. 4共c兲兴 and a disordered surface results in a low intensity
diffraction peak 关see Fig. 4共d兲兴. Furthermore during
GaPAsSb deposition, the diffraction peak is extended parallel to the surface, while its width perpendicular to the surface
is narrow 关see Fig. 4共c兲兴.
This unusual shape can be explained by a lateral composition modulation in the growing layer. The different groupIII and group-V elements involved are each from different
periods of the periodic table, resulting in strong scattering
contrast. Both a lateral phase separation of the GaPAsSb
with homogeneous layer thickness due to strain modulation
by the InGaAs template or deposition of homogeneous
GaPAsSb between the InGaAs mounds with exposed InGaAs could lead to such a modulation. A combination of
these two growth modes is also possible. Such a modulation
with a periodicity around 40 nm would lead to a diffraction
spot broadening of about 1% of the Brillouin zone. This
agrees with the observed spot size.
The layer structure of the type-II QW samples was examined
using
cross-sectional
transmission
electron
microscopy.18 Lattice imaging showed no visible sign of
structural defects. However, the electron micrographs of
sample B showed in plane undulating contrast within the
QW structure, confirming the thickness and composition
variations inferred from the RHEED measurements.
Though these material systems sound very promising
there are several growth issues that must be solved to ensure
that high quality optoelectronic materials can be grown.
Apart from problems like strain relaxation and interface segregation, the role of possible spontaneous lateral composition
modulation warrants further study.
IV. CONCLUSIONS
Long wavelength GaPAsSb type-I and InGaAs/GaPAsSb
type-II active materials for GaAs based optoelectronic de-

1548

vices are proposed. Low temperature photoluminescence
emission at 1.25 ␮m is observed from the type-I GaPAsSb
quantum well. Room temperature photoluminescence at 1.2–
1.5 ␮m is observed from the type-II quantum wells; increasing the Sb to P ratio in the GaPAsSb layer extends the emission wavelength. During GaPAsSb deposition, the reflection
high-energy electron diffraction peak is extended parallel to
the surface, while its width perpendicular to the surface is
narrow, indicating that spontaneous lateral composition
modulation occurs.

J. Piprek and S. J. B. Yoo, Electron. Lett. 30, 866 共1994兲.
J. J. Dudley, D. I. Babic, L. Yang, B. I. Miller, R. J. Ram, T. Reynolds, E.
L. Hu, and J. E. Bowers, Proc. LEOS 93, IEEE Lasers and Electro-Optics
Society 1993 Annual Meeting, p. 560.
3
R. P. Mirin, J. P. Ibbetson, K. Nishi, A. C. Gossard, and J. E. Bowers,
Appl. Phys. Lett. 67, 3795 共1995兲.
4
D. L. Huffaker, G. Park, Z. Zou, O. B. Shchekin, and D. G. Deppe, Appl.
Phys. Lett. 73, 2564 共1998兲.
5
M. Yamada, T. Anan, K. Tokutome, K. Nishi, A. Gomyo, and S. Sugou,
Proc. LEOS 98, IEEE Lasers and Electro-Optics Society 1998 Annual
Meeting, p. 149.
6
M. Peter, K. Winkler, M. Maier, N. Herres, J. Wagner, D. Fekete, K. H.
Bachem, and D. Richards, Appl. Phys. Lett. 67, 2639 共1995兲.
7
M. C. Larson, M. Kondow, T. Kitatani, K. Nakahara, K. Tamura, H.
Inoue, and K. Uomi, IEEE Photonics Technol. Lett. 10, 188 共1988兲.
8
K. Nakahara, M. Kondow, T. Kitatani, M. C. Larson, and K. Uomi, IEEE
Photonics Technol. Lett. 10, 487 共1998兲.
9
W. Braun, H. Möller, S. R. Johnson, and H.-Y. Zhang, J. Vac. Sci. Technol. B 17, 474 共1999兲.
10
M. J. Jou, Y. T. Cheng, H. R. Jen, and G. B. Stringfellow, J. Cryst.
Growth 93, 62 共1988兲.
11
S. Adachi, Physical Propeties of III–V Semiconductor Compounds
共Wiley, New York, 1992兲.
12
S. Adachi, J. Appl. Phys. 61, 4869 共1987兲.
13
H. Kroemer and G. Griffiths, IEEE Electron Device Lett. EDL-4, 20
共1983兲.
14
C. Z. Guo, S. L. Chen, P. Dowd, and Y.-H. Zhang, Appl. Phys. Lett.
共submitted兲.
15
C. G. Van de Walle, Phys. Rev. B 39, 1871 共1989兲.
16
Y.-H. Zhang, R. H. Miles, and D. H. Chow, IEEE J. Sel. Top. Quantum
Electron. 1, 749 共1995兲.
17
A. Chin and W. J. Chen, Appl. Phys. Lett. 69, 443 共1996兲.
18
P. Dowd, W. Braun, D. J. Smith, C. M. Ryu, C.-Z. Guo, S. L. Chen, U.
Koelle, S. R. Johnson, and Y.-H. Zhang, Appl. Phys. Lett. 75, 1267
共1999兲.
1
2

J. Vac. Sci. Technol. B, Vol. 18, No. 3, MayÕJun 2000

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 209.147.144.21 On: Tue, 10 Feb 2015 00:50:22