Influence of de-tuning and non-radiative recombination on the temperature dependence
of 1.3 μm GaAsSb/GaAs vertical cavity surface emitting lasers
Konstanze Hild, Igor P. Marko, S. R. Johnson, S.-Q. Yu, Y.-H. Zhang, and Stephen J. Sweeney
Citation: Applied Physics Letters 99, 071110 (2011); doi: 10.1063/1.3625938
View online: http://dx.doi.org/10.1063/1.3625938
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/99/7?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Anomalous lasing of high-speed 850nm InGaAlAs oxide-confined vertical-cavity surface-emitting lasers with a
large negative gain-to-cavity wavelength detuning
Appl. Phys. Lett. 105, 061104 (2014); 10.1063/1.4892885
Detecting and tuning anisotropic mode splitting induced by birefringence in an InGaAs/GaAs/AlGaAs verticalcavity surface-emitting laser
J. Appl. Phys. 111, 043109 (2012); 10.1063/1.3682648
Vertical cavity surface emitting lasers with the active layer position detuned from standing wave antinode for
picosecond pulse generation by gain switching
J. Appl. Phys. 110, 123101 (2011); 10.1063/1.3668321
Single-mode 1.27 μ m InGaAs vertical cavity surface-emitting lasers with temperature-tolerant modulation
characteristics
Appl. Phys. Lett. 86, 211109 (2005); 10.1063/1.1935755
Ultrafast (GaIn)(NAs)/GaAs vertical-cavity surface-emitting laser for the 1.3 μ m wavelength regime
Appl. Phys. Lett. 74, 2271 (1999); 10.1063/1.123821

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:
209.147.144.24 On: Thu, 05 Feb 2015 17:54:29

APPLIED PHYSICS LETTERS 99, 071110 (2011)

Influence of de-tuning and non-radiative recombination on the temperature
dependence of 1.3 lm GaAsSb/GaAs vertical cavity surface emitting lasers
Konstanze Hild,1 Igor P. Marko,1 S. R. Johnson,2 S.-Q. Yu,2,3 Y.-H. Zhang,2
and Stephen J. Sweeney1,a)

1
Advanced Technology Institute and Department of Physics, University of Surrey, Guildford, Surrey GU2 7XH,
United Kingdom
2
Center for Photonics Innovation, Electrical, Computer and Energy Engineering Department, Arizona State
University, Tempe, Arizona 85207-8806, USA
3
Department of Electrical Engineering, University of Arkansas, 3217 Bell Engineering, Fayettville, Arkansas
72701, USA

(Received 4 April 2011; accepted 6 July 2011; published online 17 August 2011)
In this letter, we measure the pure spontaneous emission and lasing emission from a working
vertical cavity surface emitting laser (VCSEL) for a wide range of temperatures. From this
spontaneous emission, we gain insight into the temperature dependence of the radiative component
of the threshold current (and hence gain). Together with the temperature dependence of the
threshold current, the cavity mode to gain peak alignment and the temperature dependence of an
equivalent active region edge emitting laser, we show how non-radiative recombination coupled
C 2011 American
with gain-cavity de-tuning influences the thermal properties of these devices. V
Institute of Physics. [doi:10.1063/1.3625938]
Vertical cavity surface emitting lasers (VCSELs) were
originally proposed by Iga et al.1 in 1977 and demonstrated
in 1979.2 Today, VCSELs can be found in many applications
ranging from computer mice, short haul communications3
through to oxygen sensors.4 One particularly important
potential application of VCSELs is in metropolitan area networks. Due to the zero dispersion and an absorption minimum of silica fiber available close to 1.3 lm, there has been
significant effort devoted to producing VCSELs around this
wavelength. Conventional edge-emitting lasers in this wavelength range are traditionally InP-based. However, the relatively small refractive index contrast between InGaAsP/InP
alloys, compared to GaAs/AlGaAs, means that many multiples of layer pairs are required to produce the high reflectivity distributed Bragg reflectors (DBRs) as necessary for
VCSEL operation, which can lead to both higher losses,
higher operating voltages, and poorer thermal behavior.
GaAs/AlGaAs DBRs have been used to produce 850 nm
VCSELs for which InGaAs/GaAs quantum well (QW) active
regions are also a mature technology. The challenge, therefore, has been to produce GaAs-based active regions providing operation around 1.3 lm. InGaAs/GaAs becomes very
highly strained as the wavelength is pushed towards 1.3 lm,
hence alternative approaches such as InAs/GaAs quantum
dots5 and GaInNAs/GaAs based QWs have been the subject
of intensive studies. The properties of quantum dot lasers are
far from ideal since their threshold current increases quickly
with temperature. Explanations include non-radiative processes such as Auger recombination, gain saturation, or carrier excitation.6 Quantum dot VCSELs at 1.28 lm have all
recently been demonstrated.7 1.3 lm GaInNAs(Sb)-based
QW VCSELs have been produced at 1.53 lm under pulsed
excitation,8 however, the material quality of dilute nitrides
remains an issue.9 Another possible GaAs-based system
a)

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

0003-6951/2011/99(7)/071110/3/$30.00

which has received less attention utilises GaAsSb/GaAs
QWs.10 GaAsSb-based edge emitting lasers and VCSELs
close to 1.3 lm have both been produced with continuous
wave (cw) operation at room temperature.11,12
One of the most important characteristics of a laser diode
is its sensitivity to changes in the ambient temperature. Devices with low temperature sensitivity are desirable for commercial use by reducing their reliance on active temperature
control which is expensive and significantly adds to the overall
energy demands of the system. Due to their small size, selfheating exacerbates this issue in VCSELs. Long wavelength
VCSELs have a complex temperature behavior due to the different material dependent recombination processes and the
additional effect of the temperature dependent misalignment
of the cavity mode wavelength (CM) with respect to the gain
peak wavelength. In this paper, we quantify and separate these
two effects and determine the extent to which they couple and
govern the temperature dependent properties of GaAsSb/
GaAs-based VCSELs. A key aspect of this study has been to
measure the unamplified spontaneous emission (SE) from
VCSELs during operation. From this, we have been able to
observe carrier density pinning in the VCSELs above threshold and consequently used this to determine the intrinsic radiative component of the threshold current. Thus, we are able to
directly observe the interaction of the VCSEL cavity with the
material gain and its effect on the temperature dependence of
the radiative and non-radiative processes occurring.
In this study, we investigated devices processed from
wafers grown by solid source MBE. The QW region consists
of three, 7 nm thick GaAs0.64Sb0.36 QWs with 5 nm GaAs barriers with the addition of GaAsP layers for strain compensation
to ensure that the strain-thickness product is not exceeded. The
30 lower and 25 upper pairs of AlGaAs/GaAs DBRs, respectively, enclose a cavity length of 1.5 times the free space emission wavelength of 1.26 lm. The VCSELs were fabricated
into devices with an oxide aperture and an optical window of
9 lm in diameter. Further details can be found in Refs. 12 and

99, 071110-1

C 2011 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:
209.147.144.24 On: Thu, 05 Feb 2015 17:54:29

071110-2

Hild et al.

13. To minimize self-heating effects, devices were driven
using 200 ns long pulses at a duty cycle of 10 kHz.
We begin by considering the temperature behavior of
edge emitters with nominally the same active layer structure
to probe the temperature dependence of the carrier recombination processes in the absence of the resonant VCSEL cavity. In order to do this, we have measured the threshold
current and its radiative component (Jrad) in isolation. Jrad
was measured by collecting the spontaneous emission
through a transparent window milled into the substrate of the
device using a focused ion beam technique.14 Comparing the
temperature dependence of the pure radiative current and the
total threshold current density of 1.6 kA/cm2, we found that
more than 90% of Jth at room temperature in the edge-emitting lasers is due to non-radiative recombination, previously
shown to be due to Auger recombination and leakage.15
In the VCSELs, the presence of a strongly wavelength dependent DBR reflectivity has a significant influence on the
VCSEL device properties. In particular, the CM which determines the lasing wavelength of the VCSEL, red-shifts with
increasing temperature at a much lower rate than the gain peak
wavelength whose temperature dependence approximately follows the thermally induced band gap shift.16,17 From measurements of the laser wavelength of edge emitting lasers, this is
found to be 0.36 nm/K in agreement with calculations using
the Varshni relation.18 The change in the cavity mode is due to
the much smaller change in the refractive indices of the semiconductor materials with temperature which comprise the
Bragg reflectors, measured to be 0.07 nm/K. Hence, there is an
optimum operating temperature where the CM aligns with the
peak of the QW gain spectrum. In the absence of any other
temperature dependencies in the device, the threshold current
will be at a minimum close to this temperature since maximum
gain is available at the cavity mode wavelength. As stated earlier, in the VCSEL, the lasing wavelength is principally determined by the cavity mode, whereas in the edge emitting laser,
it is determined approximately by the gain peak wavelength.
Comparing the temperature dependence of the CM and the
gain peak wavelength provides an indication of the temperature at which the gain peak and cavity mode energies align.
Fig. 1 shows the temperature dependence of the peak
lasing wavelength of an edge-emitting laser11,15 and a
VCSEL. Here, we assume that the temperature dependence
of the lasing wavelength for the edge emitting laser (based
on the same active region as the VCSEL) is equivalent to the
temperature dependence of the gain peak for the VCSEL.
The temperature dependence of the VCSEL lasing wavelength also directly provides the temperature dependence of
the cavity mode (for normal incidence). These data suggest
that the gain-CM alignment occurs at a temperature of approximately 270(65) K. Confirmation of this comes from independent photomodulated reflectance spectroscopy results on
unprocessed VCSEL wafer material.19 In the comparison of
the temperature dependence of the edge electroluminescence
and the reflectivity (and thus the cavity mode), the alignment
is found to occur at approximately 280(610) K. Further evidence for this comes from measuring the derivative of the
reflectivity in temperature dependent electroreflectance measurements.20 With this method, an alignment of the cavity
mode and the QW electroluminescence have been found at

Appl. Phys. Lett. 99, 071110 (2011)

FIG. 1. (Color online) Temperature dependence of the lasing emission peak
energy for a GaAsSb edge emitting laser (circles) and VCSEL (squares).
The two lines intersect at 270 K, corresponding to gain cavity-mode
alignment.

260(610) K. These data, therefore, suggest that the alignment
occurs in the temperature range 260–280 K.
To investigate the temperature dependence of the
VCSEL properties further, we measure both the lasing emission collected from the top of the device and the unamplified
spontaneous emission as emitted from the side of the device.
By measuring this at laser threshold, we can determine the
temperature dependence of the threshold current and its radiative component. To measure the unamplified spontaneous
emission from the VCSEL, the VCSEL wafer is cleaved
close to the contact ring. The emitted light from the VCSEL
is collected using two multi-mode silica optical fibres. One
optical fibre collects the lasing emission from the top of the
device while the second optical fibre, securely held 500 lm
from the side of the device, simultaneously measures the
unamplified spontaneous emission. The emission is measured
as a function of both current and temperature (180-300 K)
using a closed-cycle cryostat. Typical above threshold spectra
measured from the top and side of the VCSEL are shown in
Fig. 2(a) for a temperature of 210 K. It is clear that the side
emission is much broader than the lasing emission consistent
with it being unamplified SE. We also note that at this temperature, the VCSEL is “de-tuned” with the unamplified SE
peak (gain peak) at a shorter wavelength than the VCSEL lasing wavelength (CM peak).
To find the radiative current (which is proportional to the
SE), we find the intensity of the integrated SE at threshold.
This intensity from the side together with the lasing emission
at 210 K as a function of current is shown in Fig. 2(b). At
threshold, pinning of the SE can be seen. This is as expected
since the carrier density is clamped above threshold due to the
lasing process. The intensity of the SE at threshold is measured for a range of temperatures. Since the integrated spontaneous emission is proportional to the radiative current, this,
therefore, represents the temperature dependence of Jrad. To
obtain Jrad in absolute units, it is assumed that, at low temperature, Jth ¼ Jrad (since non-radiative recombination is negligible at the lowest temperatures, as determined from edgeemitting laser studies) hence the integrated spontaneous emission is normalized to Jth at the lowest temperature, as shown

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:
209.147.144.24 On: Thu, 05 Feb 2015 17:54:29

071110-3

Hild et al.

FIG. 2. (Color online) (a) Lasing emission and smoothed spontaneous emission spectra (at 210 K). The gain to cavity-mode misalignment is apparent
from this graph. (b) Spontaneous emission (edge emission) together with the
lasing emission (surface emission) for a VCSEL also at 210 K. The pinning
of the spontaneous emission can be seen above threshold.

in Fig. 3. This provides a maximum value for Jrad (T). The
fact that Jrad closely follows Jth in the low temperature range
justifies this assumption.
At low temperature Jrad (open triangles in Fig. 3) is initially high. With increasing temperature, the gain red-shifts
faster than the cavity mode (as shown in Fig. 2(a)) and Jrad
decreases due to the increased gain at the cavity mode. This
effect dominates over the typical linear increase in Jrad with
temperature in a QW.14 At low temperature Jth (filled circles
in Fig. 3) follows Jrad with a minimum at 225 K since here
Jth  Jrad. The fact that Jth increases above 225 K, while Jrad
continues to decrease (due to improved gain-cavity alignment), indicates the presence of non-radiative recombination
at the higher temperatures (shaded area in Fig. 3). The source
of the non-radiative current is likely to be due to a combina-

FIG. 3. (Color online) Temperature dependence of Ith and Irad in the
VCSEL. The decrease in Jth and Jrad with increasing temperature in the low
temperature range is due to gain to cavity-mode tuning. Non-radiative
recombination causes the increase in Jth with temperature above 230 K
and can be seen as the increase in the shaded area.

Appl. Phys. Lett. 99, 071110 (2011)

tion of Auger recombination and leakage, as found from
previous pressure and temperature dependence studies on
edge-emitting lasers based upon the same active region.15
At 280 K based upon these data, we estimate that the
non-radiative recombination accounts for at least 60% of Jth
compared to the almost 90% measured for the edge emitter
at this temperature. In order to achieve a flat temperature dependence around room temperature in these devices, a design
where the gain and cavity-mode align at higher temperature
(>300 K) would be preferable, as the tuning-effect close to
room temperature could, at least partially offset the increase
in non-radiative current with increasing temperature and is
the subject of ongoing investigations.21
In conclusion, we have experimentally measured the
temperature dependence of the radiative current in a working
VCSEL. Together with the temperature dependence of the
threshold current of the VCSEL, this has allowed us to determine and quantify the influence of non-radiative recombination on VCSEL performance. These results clearly illustrate
the importance of accounting for non-radiative recombination as well as de-tuning when designing a VCSEL for temperature stable operation around room temperature.
The authors gratefully acknowledge EPSRC(UK) for
supporting this work under Grants GR/T21516/01 and EP/
H005587/1.
1

K. Iga, IEEE J. Sel. Top. Quantum Electron. 6(6), 1201 (2000).
H. Soda, K. Iga, C. Kitahara, and Y. Suematsu, Jpn. J. Appl. Phys. 18(12),
2329 (1979).
3
J. D. Ingham, R. V. Penty, and I. H. White, J. Lightwave Technol. 24(3),
1283 (2006).
4
H. P. Zappe, M. Hess, M. Moser, R. Hövel, K. Gulden, H.-P. Gauggel, and
F. Monti di Sopra, Appl. Opt. 39(15), 2475 (2000).
5
O. B. Shchekin and D. G. Deppe, IEEE Photon. Technol. Lett. 14, 1231 (2002).
6
I. P. Marko, A. R. Adams, S. J. Sweeney, D. J. Mowbray, M. S. Skolnik,
H. Y. Liu, and K. M. Groom, IEEE J. Sel. Top. Quantum Electron. 11(5),
1041 (2005).
7
D. W. Xu, S. F. Yoon, C. Z. Tong, L. J. Zhao, Y. Ding, and W. J. Fan,
IEEE Photon. Techol. Lett. 21(17), 1211 (2009).
8
J. S. Harris, T. O’Sullivan, T. Sarmiento, M. M. Lee, and S. Vo, Semicond.
Sci. Technol. 26, 014010 (2011).
9
S. Tomic, E. P. O’Reilly, R. Fehse, S. J. Sweeney, A. R. Adams, A. D.
Andreev, S. A. Choulis, T. J. C. Hosea, and H. Riechert, IEEE J. Sel. Top.
Quantum. Electron. 9, 1228 (2003).
10
T. Anan, K. Nishi, S. Sugou, M. Yamada, K. Toukutome, and A. Gomyo,
Electron. Lett. 34, 2127 (1998).
11
K. Hild, S. J. Sweeney, S. Wright, D. A. Lock, S. R. Jin, and I. P. Marko,
Appl. Phys. Lett. 89, 173509 (2006).
12
P. Dowd, S. R. Johnson, S. S. Feld, M. Adamcyk, S. A. Chaparro, J.
Joseph, K. Hilgers, M. P. Horning, K. Shiralagi, and Y.-H. Zhang, Electron. Lett. 39, 987 (2003).
13
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).
14
S. J. Sweeney, A. F. Philips, A. R. Adams, E. P. O’Reilly, and P. J. A.
Thijs, IEEE Photon. Technol. Lett. 10, 1076 (1998).
15
K. Hild, S. J. Sweeney, I. P. Marko, S. R. Jin, S. R. Johnson, S. A. Chaparro, S. Yu, and Y.-H. Zhang, Phys. Status Solidi B 244(1), 197 (2007).
16
S. Ghosh, S. Constant, T. J. C. Hosea, and T. E. Sale, J. Appl. Phys. 88,
1432 (2000).
17
G. Knowles, S. Tomić, S. Jin, R. Fehse, S. J. Sweeney, T.E. Sale, and A.
R. Adams, Phys. Status Solidi B 235(2), 480 (2003).
18
Y. P. Varshni, Physica (Utrecht) 34, 149 (1967).
19
G. Blume, Ph.D. thesis, University of Surrey, 2007.
20
G. Blume, T. J. C. Hosea, S. J. Sweeney, S. R. Johnson, J.-B. Wang, and
Y.-H. Zhang, IEE Proc.: Optoelectron. 152(2), 110 (2005).
21
N. Hossain, S. R. Jin, S. J. Sweeney, S.-Q. Yu, S. R. Johnson, D. Ding,
Y.-H. Zhang, Proc. SPIE 7616, 761608 (2010).
2

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:
209.147.144.24 On: Thu, 05 Feb 2015 17:54:29