The influence of growth conditions on carrier recombination mechanisms in 1.3μm
GaAsSb/GaAs quantum well lasers
N. Hossain, K. Hild, S. R. Jin, S.-Q. Yu, S. R. Johnson, D. Ding, Y.-H. Zhang, and S. J. Sweeney
Citation: Applied Physics Letters 102, 041106 (2013); doi: 10.1063/1.4789859
View online: http://dx.doi.org/10.1063/1.4789859
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/102/4?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Carrier recombination in 1.3 μ m Ga As Sb ∕ Ga As quantum well lasers
Appl. Phys. Lett. 89, 173509 (2006); 10.1063/1.2369649
Unusual increase of the Auger recombination current in 1.3 μm GaInNAs quantum-well lasers under high
pressure
Appl. Phys. Lett. 82, 2335 (2003); 10.1063/1.1566468
Growth and optical properties of GaAsSb quantum wells for 1.3 μm VCSELs
J. Vac. Sci. Technol. B 19, 1948 (2001); 10.1116/1.1406153
Temperature dependence of continuous wave threshold current for 2.3–2.6 μm InGaAsSb/AlGaAsSb separate
confinement heterostructure quantum well semiconductor diode lasers
Appl. Phys. Lett. 74, 2990 (1999); 10.1063/1.123989
Analysis of temperature dependence of the threshold current in 2.3–2.6 μm InGaAsSb/AlGaAsSb quantum-well
lasers
Appl. Phys. Lett. 74, 2743 (1999); 10.1063/1.124000

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
209.147.144.22 On: Tue, 03 Feb 2015 17:54:48

APPLIED PHYSICS LETTERS 102, 041106 (2013)

The influence of growth conditions on carrier recombination mechanisms
in 1.3 lm GaAsSb/GaAs quantum well lasers
N. Hossain,1 K. Hild,1 S. R. Jin,1 S.-Q. Yu,2,3 S. R. Johnson,2 D. Ding,2 Y.-H. Zhang,2
and S. J. Sweeney1,a)

1
Advanced Technology Institute and Department of Physics, University of Surrey, Guildford, Surrey GU2 7XH,
United Kingdom
2
School of Electrical, Computer, and Energy Engineering, Arizona State University, Tempe,
Arizona 85287, USA
3
Department of Electrical Engineering, University of Arkansas, Fayetteville, Arkansas 72701, USA

(Received 13 November 2012; accepted 15 January 2013; published online 28 January 2013)
We investigate the temperature and pressure dependence of the threshold current density of
edge-emitting GaAsSb/GaAs quantum well (QW) lasers with different device characteristics.
Thermally activated carrier leakage via defects is found to be very sensitive to the growth
conditions of GaAsSb QWs. An optimization of the growth conditions reduces the nonradiative
recombination mechanisms from 93% to 76% at room temperature. This improvement in carrier
recombination mechanisms leads to a large improvement in the threshold current density from 533
Acm2/QW to 138 Acm2/QW and the characteristic temperature, T0 (T1), from 51 6 5 K
C 2013 American Institute of
(104 6 16 K) to 62 6 2 K (138 6 7 K) near room temperature. V
Physics. [http://dx.doi.org/10.1063/1.4789859]
Semiconductor lasers emitting at 1.3 lm are of considerable importance to overcome the bandwidth limitation of optical fiber communication systems (OFCS) over relatively
short distances, for example, between cities and in metropolitan areas, fiber to the home and local area networks. The
conventional InGaAsP/InP quantum well (QW) material system used to make such lasers suffers from poor temperature
characteristics resulting in the need to incorporate sophisticated temperature control electronics into the package, leading to a large increase in cost.1 Moreover, due to the lack of
lattice matching and high refractive index contrast materials
to form all-epitaxial distributed Bragg reflectors (DBRs), it
is also very difficult to fabricate monolithic vertical cavity
surface emitting lasers (VCSELs) in the InP material system.2 The GaAs-based 1.3 lm GaInNAs QWs3,4 and 1.3 lm
InAs/GaAs quantum dots5,6 have also been investigated
extensively. The performance of such lasers is still far from
ideal. GaAsSb/GaAs lasers may offer a solution in the search
for an uncooled, thermally stable and cheaper semiconductor
lasers for 1.3 lm OFCS.7,8 GaAs permits the growth of near
lattice-matched GaAs/AlGaAs DBRs, which have superior
optical and thermal properties when compared to other III-V
DBRs.9,10 Furthermore, the fabrication of GaAs based
1.3 lm VCSELs can take full advantage of the industrial
standard 850 nm VCSEL fabrication technology, which is
attractive from a manufacturing point of view.11–13 However,
the optimization of the growth conditions to grow highquality GaAs1xSbx/GaAs (x ¼ 0.3 necessary for 1.3 lm
emission) QW active material is one of the major challenges
to make this material system commercially viable.14–16 The
aim of this letter is to compare device characteristics of
GaAs1xSbx/GaAs QW lasers to aid in the design and opti-

a)

Author to whom correspondence should be addressed. Electronic address:
s.sweeney@surrey.ac.uk.

0003-6951/2013/102(4)/041106/4/$30.00

mization of GaAsSb/GaAs-based edge-emitting lasers and
VCSELs.
The devices in this study were grown using solid source
molecular beam epitaxy (MBE) growth technique under
similar growth conditions. We note that VCSELs are more
typically grown by the metal organic chemical vapor deposition (MOCVD) growth technique. However, the growth
temperature is an important parameter and the optimal temperature for mixed group-V’s depends on the composition
and the composition for a given As/Sb flux ratio depends on
growth temperature for a given set of fluxes, making the
optimization process difficult. Another important factor is
the level of impurities such as oxygen on the growing surface, which also affects the optimal growth temperature.
MBE systems with lower background impurities (compared
to MOCVD) in general grow higher quality material and
with more control compared to MOCVD grown devices.9,17
For this reason, MBE has been more widely investigated to
grow Sb-based materials. The grown structures consist of
a triple GaAs0.9P0.1/GaAs/GaAs0.7Sb0.3/GaAs/GaAs0.9P0.1
(9 nm/5 nm/7 nm/5 nm/9 nm) strain compensated QW active
region grown at 495  C (device A) and 490  C (device B).
The GaAs0.9P0.1 strain compensating layers allows growing
the maximum number of highly strained QWs and reduces
strain driven in-plane Sb segregation.8 The active region in
each device is sandwiched between two 20 nm Al0.25Ga0.75As
layers, two 150 nm graded-index AlGaAs layers, one 2 lm
n-type Al0.65Ga0.35As cladding layer followed by 500 nm
GaAs buffer layer at the bottom, and one 2 lm p-type
Al0.65Ga0.35As cladding layer followed by a 100 nm GaAs
cap layer at the top. The broad-area edge-emitting lasers were
processed by defining 50 and 100 lm wide ridges. Ti/Pt/Au
p-contact stripes ranging from 2 to 32 lm and AuGe/Ni/Au
n-metal contacts were deposited. Further details of the growth
and processing of these devices can be found in Ref. 8. The
devices were measured as-cleaved.

102, 041106-1

C 2013 American Institute of Physics
V

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
209.147.144.22 On: Tue, 03 Feb 2015 17:54:48

041106-2

Hossain et al.

Appl. Phys. Lett. 102, 041106 (2013)

Temperature dependent measurements over the range of
60–300 K were performed by using a standard closed-cycle
cryostat set-up. The emission wavelengths (at room temperature (RT)) of devices A and B are found to be very similar at
1.27 lm and 1.26 lm, respectively. However, the measured lowest threshold current density (Jth) of device B is
138 Acm2/QW at RT, which is significantly lower than
that of device A (533 Acm2/QW). Fig. 1 shows the T0
(derived from 1=T0 ¼ dlnJth =dT) for device A to be 77 6 3 K
at 200 K compared with T0 ¼ 230 6 5 K for device B. T0
drops to 51 6 5 K (device A) and 62 6 2 K (device B) at RT.
T0 (at RT) in these devices is approximately the same as that
of conventional InGaAsP/InP devices (typically 50–60 K
near RT15). Similar to T0, we find that T1 (derived from
1=T1 ¼ dlngd =dT) for device A (device B) is 508 6 11 K
(636 6 4 K) at 200 K, which drops to 104 6 16 K (138 6 7 K)
at RT. Indeed, when compared with 1.2 lm InGaAs/GaAs
devices, T1 is significantly lower in these devices.18 The
lower T1 is an indication of thermally activated recombination processes,19 which may occur in these lasers at higher
temperatures. However, both T0 and T1 for device B are
higher than that for device A over the temperature range
studied, which is consistent with the reduced Jth of device B
than that of device A. A non-ideal internal quantum efficiency (gi) may also contribute to the low T1 values in these
devices. The inverse of differential quantum efficiency (gd)
against the cavity length (Lcav) was measured from which gi
and internal optical loss (ai) of device B were determined
(using the relation mentioned in Ref. 20) to be 76 6 1% and
12 6 0.3 cm1, respectively. We could not compare the values of these parameters of device B with device A due to the
unavailability of multiple cavity length devices. We note
that the gi is lower and ai is higher in device B (which is
showing improved performance in this work) than that
of other GaAs-based materials (for which internal quantum
efficiency and internal optical loss are 83%-85%21 and
10 cm1,22 respectively). The low gi and high ai in device B
is consistent with the large threshold current density in these
devices compared to other GaAs-based lasers, which operate
at similar wavelengths for which Jth/QW  100 Acm2.23
Achieving an ideal QW laser operation in GaAs1xSbx/GaAs

material system therefore requires an understanding of the
important carrier recombination mechanisms.
Fig. 2 shows the normalized (at T ¼ 60 K) temperature
dependence of Jth and its radiative component (Jrad), which is
extracted from the pinning level of the measured integrated
spontaneous emission at laser threshold (further details of
spontaneous emission measurement technique can be found
in Ref. 24). Here, Jrad is determined by assuming that (a) Jrad
is proportional to the integrated spontaneous emission rate at
laser threshold and (b) that non-radiative recombination is
negligible at the lowest temperature. Jrad therefore provides a
measure of the maximum radiative component of Jth as a
function of temperature. From the measured Jth and Jrad, for
device B, we estimate that the relative ratio of the radiative
and non-radiative currents to Jth are 24% and 76%,
respectively, at RT, compared with 7% and 93% for device A. Thus, a non-radiative recombination process dominates Jth at RT in both structures. It can also be seen that in
addition to having a lower absolute Jth, device B has a higher
Jrad/Jth than in device A. A photoluminescence study of
GaAs0.7Sb0.3/GaAs QW lasers (similar to devices A and B)
suggests that antimony segregation causes lower photoluminescence intensity and a larger full-width at half maximum
and hence a higher Jth.8 Antimony segregation introduces
defects at the GaAsSb active region—GaAs spacer interface
forming a thermally activated loss mechanism as more energetic carriers increasingly recombine via the defect states. It
follows that the lower growth temperature of device B
reduces Sb segregation and the associated defect density and
consequently reduces the thermally activated non-radiative
processes, thereby reducing Jth for device B. It is important
to maintain the improved growth conditions for repeated
growths. The active material mole fraction must be continually and accurately calibrated to maintain the desired mole
fraction and the substrate to substrate thermal couple temperature offset must be calibrated every run so that the temperature is controlled within a few degrees Celsius run to run.
This can be done using pyrometry or band edge thermometry. We note that while the growth temperature of GaAsSb
QWs offers a possible explanation of the reduced Jth of device B, this in itself is insufficient. Others factors, such as

FIG. 1. ln[Ith] and ln[gd] as a function of temperature.

FIG. 2. Normalized (at T ¼ 60 K) temperature dependence of Jth (full squares
and circles). Temperature dependence of Jrad (open squares and circles).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
209.147.144.22 On: Tue, 03 Feb 2015 17:54:48

041106-3

Hossain et al.

Appl. Phys. Lett. 102, 041106 (2013)

device processing may also has affected the performance of
these devices, which remains the subject of further investigation. We also note that the presence of defects in these devices suggests that the analysis in Fig. 2 may overestimate Jrad
at low temperature due to there being another non-radiative
path active at low temperature. Considering this nonradiative path (at 60 K) in our analysis, the non-radiative
contribution at RT in device A and device B will be higher
than 76% and 93%, respectively. However, due to the
higher defect density in device A, non-radiative contribution
at RT in device A compared to device B will be even higher
than the estimated values in Fig. 2. Hence, the possible overestimation of Jrad in Fig. 2 does not affect our current analysis and conclusions.
To further probe the recombination mechanisms, high
pressure techniques were utilized. The application of hydrostatic pressure mainly affects the conduction band (CB)
causing an increase in the direct band gap of III–V semiconductors and is therefore an ideal method to investigate the
important band gap dependent non-radiative processes (like
Auger, carrier leakage, and defect-related recombination).
With increasing pressure, the Auger recombination process
reduces, carrier leakage mechanisms increase or remain stable, and defect related recombination remains constant. Further details of these dependencies can be found in Ref. 25.
Fig. 3 shows the measured pressure dependence of threshold
current (normalized at 0 kbar) at RT. Also shown is the ideal
expected variation of Jrad / E2g ,26 where Eg is the bandgap
(taken from Eg ¼ hc=k, where k is the measured lasing
wavelength). It can be clearly seen that the threshold current
increases with pressure much faster than Jrad for both devices A and B, which confirms that the lasers are not operating
in a radiatively dominated regime. Jth for devices A and B
increase by 27% and 44%, respectively, up to 7 kbar at
RT suggesting the presence of carrier leakage in these devices as observed earlier in similar devices.1,15,16 The lower
rate of increase in Jth with pressure in device A compared to
device B may be due to the higher growth temperature of device A, which increases Sb segregation and thus the concentration of defects. The fact that Jth for device A has a

stronger temperature dependence but weaker pressure dependence indicates the presence of defects which are shallow
and therefore not pressure dependent.
Fig. 4 shows the measured pressure dependence of threshold current (normalized at 0 kbar) at 120 K. It can be seen that
the threshold current of device B increases with increasing
pressure at the similar rate to Jrad, indicating that radiative
recombination dominates at this low temperature. A slight
deviation of threshold current from Jrad / E2g may be due to a
couple of factors. A weak type-I conduction band alignment
with a conduction band offset, DEc ¼ 19.0 6 19.1 meV, has
been reported at the GaAs0.7Sb0.3/GaAs interface of these
materials,27 as shown in Fig. 5. Pressure coefficients for the
band gap for both devices A and B are measured to be
7.7 meV/kbar. On the other hand, the well known pressure
co-efficient of GaAs is 10.7 meV/kbar.1 Hence, with increasing pressure, DEc increases at a rate of 3 meV/kbar if we
assume that hydrostatic pressure predominantly affects the
conduction band28,29 and forms a stronger type-I band alignment at the GaAs0.7Sb0.3/GaAs interface (inset of Fig. 5),
which increases the electron-hole overlap. As a result, lower
carrier injection is required to achieve the required gain. The

FIG. 3. Measured pressure dependence of threshold current and ideal Jrad at
RT.

FIG. 5. Band alignment at GaAs0.7Sb0.3/GaAs interface. With increasing pressure, DEc increases and forms a type-I band alignment at the GaAs0.7Sb0.3/
GaAs interface (inset).

FIG. 4. Measured pressure dependence of threshold current and ideal Jrad at
120 K.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
209.147.144.22 On: Tue, 03 Feb 2015 17:54:48

041106-4

Hossain et al.

reduction of threshold carrier density (nth) with pressure will
decrease Jrad, which would cause Jrad deviate from the simple
Jrad / E2g model as described in Ref. 30. Furthermore,
the presence of Auger recombination may also give rise to a
lower pressure dependence of threshold current, as has been
observed previously in Ref. 1.
In summary, it is found that the threshold current of
GaAs1xSbx/GaAs QW lasers is dominated by non-radiative
recombination mechanisms which are responsible for the
poor temperature sensitivity of the devices resulting in low T0
and T1 values at RT. Pressure dependence measurements suggest that carrier leakage mechanism dominates the threshold
current density at room temperature in these devices. Thermally activated carrier leakage via defects is found to be very
sensitive to the growth conditions of QWs. Optimization of
growth conditions leads to an improvement in the radiative
recombination process from 7% to 24%, results a lower
Jth from 533 Acm2/QW to 138 Acm2/QW and a higher T0
from 51 6 5 K to 62 6 2 K at RT. Optimization of 1.3 lm
GaAs1xSbx/GaAs QW lasers will rely upon reducing the
thermally activated carrier leakage via the defect states and
by careful optimization of growth conditions.
The authors gratefully acknowledge the EPSRC (UK),
grants GR/T21516/1 and EP/H005587/1, the Royal Academy
of Engineering, and the Kwan Trust Fund in supporting this
work.
1

K. Hild, S. J. Sweeney, S. Wright, D. A. Lock, S. R. Jin, I. P. Marko, S. R.
Johnson, S. A. Chaparro, S.-Q. Yu, and Y.-H. Zhang, Appl. Phys. Lett. 89,
173509 (2006).
2
W. Braun, P. Dowd, C. Z. Guo, S. L. Chen, C. M. Ryu, U. Koelle, S. R.
Johnson, Y. H. Zhang, J. W. Tomm, T. Elsasser, and D. J. Smith, J. Appl.
Phys. 88, 3004 (2000).
3
S. R. Jin, S. J. Sweeney, S. Tomić, A. R. Adams, and H. Riechert, Appl.
Phys. Lett. 82, 2335 (2003).
4
A. D. Andreev and E. P. O’Reilly, Appl. Phys. Lett. 84, 1826 (2004).
5
Y. Qiu, P. Gogna, S. Forouhar, A. Stintz, and L. F. Lester, Appl. Phys.
Lett. 79, 3570 (2001).
6
I. P. Marko, A. R. Adams, S. J. Sweeney, D. J. Mowbray, M. S. Skolnick,
H. Y. Liu, and K. M. Groom, IEEE J. Sel. Top. Quantum Electron. 11,
1041 (2005).

Appl. Phys. Lett. 102, 041106 (2013)
7

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 (1999).
S.-Q. Yu, D. Ding, J.-B. Wang, N. Samal, X. Jin, Y. Cao, S. R. Johnson,
and Y.-H. Zhang, J. Vac. Sci. Technol. B 25, 1658 (2007).
9
M. S. Noh, R. D. Dupuis, D. P. Bour, G. Walter, and N. Holonyak, Appl.
Phys. Lett. 83, 2530 (2003).
10
J. J. Wierer, D. A. Kellogg, and N. Holonyak, Appl. Phys. Lett. 74, 926
(1999).
11
M. Dinu, J. E. Cunningham, F. Quochi, and J. Shah, J. Appl. Phys. 94,
1506 (2003).
12
M. J. Cherng, G. G. Stringfellow, and R. M. Cohen, Appl. Phys. Lett. 44,
677 (1984).
13
Y.-K. Su, C.-T. Wan, R. W. Chuang, C.-Y. Huang, W. C. Chen, Y. S.
Wang, and H.-C. Yu, J. Cryst. Growth 310, 4850 (2008).
14
C.-A. Chang, R. Ludeke, L. L. Chang, and L. Esaki, Appl. Phys. Lett. 31,
759 (1977).
15
N. Hossain, S. R. Jin, S. J. Sweeney, S.-Q. Yu, S. R. Johnson, D. Ding,
and Y. H. Zhang, Proc. SPIE 7616, 761608 (2010).
16
N. Hossain, K. Hild, S. J. Sweeney, S.-Q. Yu, S. R. Johnson, D. Ding, and
Y. H. Zhang, in Proceedings of IEEE Photonics Global Conference
Singapore (IEEE, 2010), pp. 1–3.
17
S.-W. Ryu, and P. D. Dapkus, Electron. Lett. 38, 564 (2002).
18
N. Tansu, Y.-L. Chang, T. Takeuchi, D. P. Bour, S. W. Corzine, M. R. T.
Tan, and L. J. Mawst, IEEE J. Quantum Electron. 38, 640 (2002).
19
N. Hossain, S. R. Jin, S. Liebich, M. Zimprich, K. Volz, B. Kunert, W.
Stolz, and S. J. Sweeney, Appl. Phys. Lett. 101, 011107 (2012).
20
S. J. Sweeney and A. R. Adams, “Optoelectronic devices and materials,”
in Springer Handbook of Electronic and Photonic Materials (Springer,
2006).
21
X. D. Wang, S. M. Wang, Y. Q. Wei, M. Sadeghi, and A. Larsson, Electron. Lett. 40, 1338 (2004).
22
Y. Q. Ning, Y. F. Sung, Z. H. Jin, and L. J. Wang, Proc. SPIE 5644, 614
(2005).
23
G. Adolfsson, S. M. Wang, M. Sadeghi, and A. Larsson, Electron. Lett.
43, 454 (2007).
24
S. J. Sweeney, A. F. Philips, A. R. Adams, E. P. O’Reilly, and P. J. A.
Thijs, IEEE Photonics Technol. Lett. 10, 1076 (1998).
25
K. Hild, S. J. Sweeney, I. P. Marko, S. R. Jin, S. R. Johnson, S. A.
Chaparro, S.-Q. Yu, and Y.-H. Zhang, Phys. Status Solidi B 244, 197
(2007).
26
A. R. Adams, M. Silver, and J. Allam, Semicond. Semimetals 55, 301
(1998).
27
S. R. Johnson, C. Z. Guo, S. Chaparro, Yu. G. Sadofyev, J. Wang, Y. Cao,
N. Samal, J. Xu, S. Q. Yu, D. Ding, and Y.-H. Zhang, J. Cryst. Growth
251, 521 (2003).
28
C. G. Van de Walle and M. M. Martin, Phys. Rev. B 35, 8154 (1987).
29
N. Hossain, I. P. Marko, S. R. Jin, K. Hild, S. J. Sweeney, R. B. Lewis, D.
A. Beaton, and T. Tiedje, Appl. Phys. Lett. 100, 051105 (2012).
30
N. F. Masse, A. R. Adams, and S. J. Sweeney, Appl. Phys. Lett. 90,
161113 (2007).
8

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
209.147.144.22 On: Tue, 03 Feb 2015 17:54:48