Large negative persistent photoconductivity in InAs ∕ AlSb quantum wells
Yu. G. Sadofyev, A. Ramamoorthy, J. P. Bird, S. R. Johnson, and Y.-H. Zhang
Citation: Applied Physics Letters 86, 192109 (2005); doi: 10.1063/1.1926407
View online: http://dx.doi.org/10.1063/1.1926407
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/86/19?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Electron spin relaxation time in (110) InGaAs/InAlAs quantum wells
J. Appl. Phys. 116, 023507 (2014); 10.1063/1.4887803
Excitation wavelength dependence of the anomalous circular photogalvanic effect in undoped InGaAs/AlGaAs
quantum wells
J. Appl. Phys. 115, 083509 (2014); 10.1063/1.4867039
Relaxation and dephasing of the intersubband transitions in n -type In As ∕ Al Sb multi quantum wells
Appl. Phys. Lett. 91, 071918 (2007); 10.1063/1.2769948
Study of Spin‐Orbit Interaction in AlxGa1−xAsySb1−y/InAs Quantum Well by Means of Persistent
Photoconductivity Effect on Weak Localization
AIP Conf. Proc. 893, 569 (2007); 10.1063/1.2730019
Positive and negative persistent photoconductivities in semimetallic Al x Ga 1−x Sb/InAs quantum wells
J. Appl. Phys. 86, 3152 (1999); 10.1063/1.371182

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:
209.147.144.21 On: Mon, 09 Feb 2015 17:26:56

APPLIED PHYSICS LETTERS 86, 192109 共2005兲

Large negative persistent photoconductivity in InAs/ AlSb quantum wells
Yu. G. Sadofyev, A. Ramamoorthy, J. P. Bird, S. R. Johnson,a兲 and Y.-H. Zhang
Department of Electrical Engineering and Center for Solid State Electronics Research, Arizona State
University, Tempe, Arizona 85287-5706

共Received 20 December 2004; accepted 28 March 2005; published online 5 May 2005兲
Negative persistent photoconductivity in backgated InAs/ AlSb quantum wells under red and blue
light emitting diode illumination is observed to depend on both illumination energy and intensity.
During these experiments the electron sheet density is varied by one order of magnitude from
5.6⫻ 1011 to 5.8⫻ 1010 cm−2. This behavior is attributed to the role of optically excited holes in the
AlSb barrier layers near the InAs quantum well. Furthermore, a long relaxation time in the electron
sheet density is observed and attributed to the slow movement of electrons from the InAs quantum
well to ionized deep levels in the AlSb barriers. © 2005 American Institute of Physics.
关DOI: 10.1063/1.1926407兴
InAs/ AlSb-based quantum structures are very attractive
for applications in spintronics and high-speed electronic devices. The advantages of this material system include the low
electron effective mass and large g* factor of InAs and the
strong electron confinement that results from the large energy difference 共1.35 eV兲 between the conduction bands of
AlSb and InAs. It is well established that, in nominally undoped InAs/ AlSb structures, a two-dimensional electron gas
共2DEG兲 is formed in the InAs well with a sheet density on
the of order 1012 cm−2. The proposed sources of these carriers include donorlike defects in the AlSb barriers, donors at
the InAs/ AlSb interfaces, and surface donors in the GaSb
cap layer.1 An interesting feature of the InAs/ AlSb system is
a bipolar behavior of its persistent photoconductivity.2 Positive persistent photoconductivity, namely, an increase in the
2DEG density, occurs for illumination by photons with energies lower than about 1.55 eV, and has been attributed to the
photoionization of deep donors.2 Negative persistent photoconductivity 共NPPC兲, on the other hand, occurs for illumination by photons with energies larger than about 1.55 eV. The
NPPC effect has been attributed to photoexcited holes that
drift toward, and recombine with electrons in the InAs quantum well 共QW兲 关2,3兴 and the capture of photoexcited electrons by deep donors in the AlSb barriers that are located
slightly above the first quantized subband in the InAs QW.4
In similar InAs/ AlGaSb structures, which have a larger
type-II valance band offset than InAs/ AlSb, the NPPC effect
has been attributed to the capture of photoexcited electrons
from the InAs well by ionized deep donors in the AlGaSb
barriers 5 and compositional fluctuations at the AlGaSb/ InAs
interface.6
This article reports the study of the NPPC effect in backgated InAs/ AlSb QWs where the electron density in the QW
is decreased by a factor of 2.6 using red light emitting diode
共LED兲 illumination and a factor of 9.7 共from 5.6⫻ 1011 to
5.8⫻ 1010 cm−2兲 using blue light emitting diode 共LED兲 illumination. Previous work reports a factor of two using a red
LED 共Ref. 7兲 and factors of three3 and five7 using a green
LED which is attributed to the role of photogenerated holes
in the AlSb barriers.
a兲

Author to whom correspondence should be addressed; electronic mail:
shane.johnson@asu.edu

Undoped InAs/ AlSb QWs structures were grown on
共100兲 n+ GaAs substrates by molecular-beam epitaxy. In order to accommodate the lattice-constant differences between
GaAs, AlSb, and InAs, and to increase the electrical resistance between the InAs QW and the GaAs substrate, a composite buffer layer was grown. This layer consisted of 500
nm of AlAs, grown on a 200 nm undoped GaAs buffer, followed by a 2 ␮m thick AlSb metamorphic buffer layer, and a
25-period 共2.5 nm/ 1.25 nm兲 AlSb/ GaSb superlattice. The
superlattice serves as a smoothing layer for subsequent
molecular-beam epitaxy growth and as an etch-stop layer
during Hall-bar preparation by wet chemical etching. The
active part of the structure consists of a 25 nm AlSb barrier,
a 15 nm InAs QW, a 40 nm AlSb barrier, and a 6 nm GaSb
cap layer. These layers were grown at a substrate temperature
of 480 °C, and the shutter sequence was chosen to provide
InSb-like bonds at both interfaces of the InAs QW.
After growth, Hall bars were defined by mesa etching in
a solution of H3PO4 : H2O2 : H2O 共1:1:10兲 at room temperature. Since the developers commonly used in photolithography attack GaSb and AlSb, 10 nm of amorphous Ge was
evaporated on top of the GaSb layer to serve as a protection
layer. At low temperatures, where our electricalcharacterization studies were performed, we have confirmed
that the Ge layer does not create any significant shunting of
the QW layer. The phosphoric acid-based solution removes
the Ge film, together with the other layers of the structure.
After mesa etching, Si3N4 was deposited at a temperature of
50 °C on the sidewalls and floor of the mesa by plasmaassisted chemical vapor deposition. Several non-alloyed
ohmic contacts were formed to the top surface of the InAs
QW using a two-step etch in phosphoric acid and ammonia
hydroxide based solutions to remove the upper layers of Ge,
GaSb, and AlSb, and then directly depositing Cr–Au on the
exposed InAs. Last, an In film was deposited on a backside
of the n+ GaAs wafer to serve as a backgate.
In our study of NPPC, red and blue LEDs were mounted
in close proximity to the Hall-bar sample and a predetermined constant current ILED was supplied to either the blue
or the red LED, for an interval of a about 30 s. The influence
of the illumination was monitored by measuring the change
in the voltage drop between pairs of potential probes, both in
the absence and in the presence of a small applied magnetic
field. After each dose of illumination, relaxation effects were

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:
0003-6951/2005/86共19兲/192109/3/$22.50
86, 192109-1
© 2005 American Institute of Physics
209.147.144.21 On: Mon, 09 Feb 2015 17:26:56

192109-2

Appl. Phys. Lett. 86, 192109 共2005兲

Sadofyev et al.

FIG. 1. Shubnikov–de Haas oscillations measured before 共lower curve兲 and
after illumination with red light 共next four curves兲 and blue light 共upper two
curves兲. For clarity, the curves are shifted vertically by the resistances
shown on the left.

FIG. 2. Linear increase in electron mobility with sheet density under different backgate voltages 共Vg = 0 , −0.5, and +1.5 V兲, before 共labeled initial兲 and
after illumination with red light 共labeled red LED兲 and blue light 共labeled
blue LED兲.

The resistance relaxation in our sample is shown in Fig.
3, where 30 s of illumination under the blue LED at a 1 ␮A
drive current, increases the resistance well above the initial
steady-state value, then after turning the LED off the resistance approaches a new higher resistance. A similar type of
behavior is observed using the red LED, although the effect
of red light on the sample resistance is not as dramatic. As is
consistent with the NPPC effect, this new steady state is a
metastable one, where after thermal cycling between room
temperature and 4.2 K the electron concentration and the
resistance of the sample return to their initial values. The
inset in Fig. 3 shows a long time-scale relaxation process that
takes place after illumination where the resistance R increases a further 4 k⍀, saturating at RSAT = 17 k⍀. This long
scale relaxation involves a single time constant of 7740 s and
is likely a result of the slow movement of electrons from the
QW to ionized deep levels in the AlSb barriers. Interface
states, or fluctuations of the potential at the InAs QW interfaces, can be ruled out since a random distribution of energies is typical for defect states at interfaces, and should not
give rise to a relaxation process with a single time constant.

observed in the resistance; therefore we waited for approximately 30 min, before beginning the magnetotransport measurements.
The influence of illumination from the red and blue
LEDs on the longitudinal magneto-resistance and the 2DEG
sheet density is shown in Fig. 1 using zero backgate voltage
共Vg = 0兲. The lower curve shows the Shubnikov–de Haas oscillations for the initial measurements performed at 4.2 K,
where the electron sheet density is 5.6⫻ 1011 cm−2 and the
electron mobility is 98 950 cm2 / Vs. 共At 300 K the electron
sheet density was 8.1⫻ 1011 cm−2 and the electron mobility
was 28 730 cm2 / Vs.兲 The next four curves show measurements performed after illumination from the red LED, where
the 2DEG density is reduced to 2.1⫻ 1011 cm−2 for the highest level of illumination used 共ILED = 100 ␮A兲. The upper two
curves show the measurements performed after illumination
from the blue LED, where a further decrease in the electron
density is realized. Although not shown for clarity, a 2DEG
density of 5.8⫻ 1010 cm−2 is reached for ILED = 50 ␮A. At
this density, the extreme quantum limit is reached, where
only the lowest spin-resolved Landau level is occupied at a
magnetic field as low as 3 T. The illumination effects are
cumulative, starting with the lowest to the highest red LED
drive current and then from the lowest to the highest blue
LED drive current. Upon warmup and recooling and repeating a blue LED measurement we find that the 2DEG density
is reproduced within 20%, indicating that the cumulative effect does not substantially impact the results.
The presence of the composite buffer layer in this structure permits the application of backgate voltages in the range
−7.5⬍ Vg ⬍ + 7.5 V, with a leakage current below ±10 nA.
After each red LED illumination step, the magnetoresistance
was measured for different backgate voltages 共Vg = 0 , −0.5,
and +1.5 V兲. In general, the influence of the backgate was
found to be similar to that observed prior to illumination,
with a negative 共positive兲 backgate voltage reducing 共increasing兲 the 2DEG sheet density. The low temperature
2DEG mobility decreases linearly with sheet density 共see
FIG. 3. Behavior of the NPPC effect before, during, and after illumination
Fig. 2兲; due to a decrease in electron screening of impurity
withtoblue
shows the deviation of the resistance from
its satura- to IP:
This article is copyrighted
as indicated in the article. Reuse of AIP content is subject
the light.
termsInset
at: http://scitation.aip.org/termsconditions.
Downloaded
3
scatters.
tion value 共RSAT = 17 k⍀兲.
209.147.144.21 On: Mon, 09 Feb 2015 17:26:56

192109-3

Appl. Phys. Lett. 86, 192109 共2005兲

Sadofyev et al.

FIG. 4. Electron sheet density vs LED drive current, before 共labeled initial兲
and after illumination with red light 共labeled red LED兲 and blue light 共labeled blue LED兲.

The electron sheet density as a function of LED drive
current and color is shown in Fig. 4. The red LED drive
current is shown on the lower x-axis scale and the blue LED
drive current 共scaled up by a factor of 50 000兲 is shown on
the upper x-axis scale. The initial density before illumination
is given by the solid circle labeled initial, and the densities
after illumination under red or blue light are given by the
solid circles labeled red or blue LED. The straight solid line
is a power law fit to the data; this fit illustrates a weak,
approximately −1 / 8, power dependence of the sheet density
on LED drive current. The 50 000 scaling factor required to
arrange the blue data in a line with the red data, indicates that
blue light is 共50 000兲−1/8.5 ⬇ 3.6 times more effective at reducing the 2DEG density than red light. When plotted in this
manner the dependence of electron density on illumination
intensity does not appear to saturate, however, due to the
weakness of the power law, it appears to saturate when plotted on a linear scale. The functional form of the effect of
light intensity on NPPC has not been reported, however
qualitatively, a weak dependence on illumination intensity
has been previously observed.2,7
For visible wavelengths, the absorption of light in the
AlSb barriers and InAs QW is dominated by band–to–band
absorption; therefore the photogeneration of electron-hole
pairs must be the prevailing process. The absorptance 共fraction of light entering layer that is absorbed兲 in the InAs QW
is about 20% 共␣ ⬃ 1.5⫻ 105 cm−1兲 for red 共⬃600 nm兲 light
and 60% 共␣ ⬃ 6 ⫻ 105 cm−1兲 for blue 共⬃450 nm兲 light.
While the absorptance in the upper 共40 nm thick兲 AlSb barrier is about 4% 共␣ ⬃ 1 ⫻ 104 cm−1兲 for red light 共indirect

band-gap transition兲 and about 90% 共␣ ⬃ 6 ⫻ 105 cm−1兲 for
blue light 共direct band-gap transition兲. As a result about 2%
of the blue light 共that enters the sample兲 is absorbed in the
InAs QW 共after passing through the Ge, GaSb, and AlSb
layers兲 compared to about 13% for the red light 共⬃7 times
less blue light is absorbed in the InAs QW兲. On the other
hand, about 40% of the blue light is absorbed in the upper
AlSb barrier compared to about 3% for red light 共⬃13 times
more blue light is absorbed in the upper AlSb barrier兲. Interestingly enough, about 1% of both the red and blue light is
absorbed in the lower 共25 nm thick兲 AlSb barrier; ⬃50% of
the red light reaches the lower barrier, where little of it is
absorbed, and ⬃1% of the blue light reaches the lower barrier, where most of it is absorbed.
In conclusion, illumination with red or blue light rapidly
drives the 2DEG population into a new steady state where
photogenerated holes near the InAs QW drift toward, and
recombine with electrons in the InAs QW. This process takes
place over a few seconds and quickly terminates, indicating
that the transport of holes to the InAs QW suddenly ceases;
likely due to a reduction in the built-in electric field as the
corresponding photogenerated electrons drift toward, and
neutralize, surface donors and to a lesser extent, donors in
the upper AlSb barrier. The ability of the holes to drift toward and recombine with electrons in the InAs QW weakly
increases with illumination intensity and strongly increases
with illumination energy; indicating that the electron-hole
pair generation rate 共which is proportional to intensity and
several times larger for blue light兲, and possibly even the
level excited above the band edge, play an important role in
hole transport and electron-hole recombination at the InAs
QW. The further rapid decrease in the 2DEG density observed once illumination is stopped is likely due to the recombination of QW electrons with some of the photogenerated hole population next to the QW. The final metastable
steady state is reached after a long time relaxation of electrons that slowly move from the InAs QW to ionized deep
levels in the AlSb barriers.
1

C. Nguyen, B. Brar, H. Kroemer, and J. H. English, Appl. Phys. Lett. 60,
1854 共1992兲.
2
Ch. Gauer, J. Scriba, A. Wixforth, J. P. Kotthaus, C. Nguyen, G. Tuttle, J.
H. English, and H. Kroemer, Semicond. Sci. Technol. 8, S137 共1993兲.
3
G. Tuttle, H. Kroemer, and J. H. English, J. Appl. Phys. 65, 5239 共1989兲.
4
F.-C. Wang, W. E. Zhang, C. H. Yang, M. J. Yang, and B. R. Bennet, Appl.
Phys. Lett. 69, 1417 共1996兲.
5
I. Lo, W. C. Mitchell, M. O. Manasreh, C. E. Stutz, and K. R. Evans,
Appl. Phys. Lett. 60, 751 共1992兲.
6
W. C. Wang, L. C. Tsai, J. C. Fan, Y. F. Chen, and I. Lo, J. Appl. Phys. 86,
3152 共1999兲.
7
P. F. Hopkins, A. J. Rimberg, R. M. Westervelt, G. Tuttle, and H. Kroemer,
Appl. Phys. Lett. 58, 1428 共1991兲.

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:
209.147.144.21 On: Mon, 09 Feb 2015 17:26:56