242 IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 23, NO. 4, FEBRUARY 15, 2011
Noise Characterization of Midwave Infrared
InAs/GaSb Superlattice pin Photodiode
Katarzyna Jaworowicz, I. Ribet-Mohamed, C. Cervera, J. B. Rodriguez, and P. Christol
Abstract—We report on noise characterization of a midwave
infrared (MWIR) InAs/GaSb superlattice (SL) single detector.
The SL structure was made of eight InAs monolayers (MLs) and
eight GaSb MLs, with a total thickness of 2 m (440 SL periods).
This structure exhibits a cut-off wavelength of 4.8 m at 77 K. Ex-
tracted from current–voltage characteristics, zero-bias resistance
area product above cm at 80 K was measured.
Noise measurements were also performed under dark conditions.
The measurements reveal the absence of noise above 30 Hz.
Moreover, the detector under test remains Schottky noise-limited
up to a bias voltage of 200 mV typically, which confirms the
quality of the MWIR SL pin photodiode.
Index Terms—Dark current, InAs/GaSb superlattice, infrared
photodiodes, noise measurement.
I. INTRODUCTION
T
HE 3–5 m mid-wavelength infrared range (MWIR) is an
interesting spectral domain for high performance infrared
(IR) imaging technology for medical applications (diagnosis as-
sistance), industrial applications (process control) and military
applications (night vision). Type-II InAs/GaSb superlattice (SL)
photodiode has recently obtained tremendous results and can be
now considered as an interesting alternative technology to the
well-established InSb and HgCdTe MWIR photodetectors [1].
However, to enhance the device performances, a better knowl-
edge of electrical properties [2], carrier-lifetime [3], gain [4]
and, more importantly, noise measurements of SL photodiode
[5]–[7] are still necessary. Indeed, noise measurements allow
one to calculate the real detectivity, whereas most of the de-
tectivity values reported in the literature use a theoretical ex-
pression for the noise contribution which assumes a Johnson
noise-limited detector. The analysis of the physical origin of the
noise (Johnson noise, Schottky noise or noise) also allows
one to estimate the potentiality of InAs/GaSb SL detector tech-
nology. Paradoxically, only few papers deal with the study of
noise in SL detectors: [6] and [7] focus on longwave infrared
(LWIR) focal plane arrays (FPA) and single elements, respec-
tively, while [5] reports measurements on MWIR SL detectors
(but only at nearly zero bias voltage).
Manuscript received September 15, 2010; revised November 08, 2010; ac-
cepted November 13, 2010. Date of publication November 18, 2010; date of
current version January 28, 2011. This work was supported in part by the French
DGA.
K. Jaworowicz and I. Ribet-Mohamed are with the ONERA, 91761 Palaiseau,
France (e-mail: Katarzyna.Jaworowicz@onera.fr).
C. Cervera, J. B. Rodriguez, and P. Christol are with the Institut d’Electron-
ique du Sud (IES), Université Montpellier 2, UMR CNRS 5214, 34095 Mont-
pellier, France (e-mail: christol@ies.univ-montp2.fr).
Color versions of one or more of the figures in this letter are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/LPT.2010.2093877
In this letter we report on dark current and noise measure-
ments performed on p-i-n MWIR InAs/GaSb SL single detector.
II. EXPERIMENTS
The detector structure was grown by molecular beam epitaxy
(MBE) on p-doped GaSb substrate. It consists of a 200 nm-thick
Be-doped ( cm ) GaSb buffer layer, an undoped
InAs/GaSb SL active region surrounded by 10 periods of n and
p-doped SL with similar period composition, and a 20 nm-thick
Te-doped ( cm ) InAs cap layer. The unintention-
ally doped SL comprises 440 periods, corresponding to a total
thickness of 2 m; each period being made of 8 InAs mono-
layers (MLs) and 8 GaSb MLs. This structure exhibits photolu-
minescence emission in the MWIR, with a PL peak position at
4.8 m at 80 K [8]. Further details on MBE growth and struc-
tural characterizations have been published elsewhere [9].
Circular mesa diodes were realized using standard pho-
tolithography techniques with a mask set containing diodes
with several diameters ranging from 40 to 465 m. AuGeNi
and AuZn contact layers were deposited respectively on top
of the mesa (n-type InAs cap layer) and on the back of the
GaSb p-type substrate. A citric acid based solution was used
for mesa etching and immediately after, a photoresist AZ-1518
was spun onto the sample and polymerized in order to protect
the surface from ambient air [10]. Finally, the samples were
indium-bonded and packaged in submounts.
To perform dark current and dark noise measurements, the
SL detectors were placed on the cold finger of a continuous flow
SMC 7845.1 cryostat operated with helium or nitrogen. A BT
400 controller was used to reach the desired temperature with a
resolution of 0.01 K.
For dark current measurements, a Keithley 6430 subfem-
toamperemeter was used to both apply the bias voltage and read
the current delivered by the single detector. For dark current
noise measurements, we used a Keithley 428 transimpedance
amplifier and a spectrum analyzer (ONO SOKKI). In both
cases, the experimental setup was controlled by a computer
equipped with a GPIB (General Purpose Interface Bus) inter-
face and dedicated Python software.
III. RESULTS
Fig. 1 plots several dark current density curves as a function
of the bias voltage, for different operating temperatures (from
20 K up to 250 K). The sample under test is a cm
area pixel. A dark current density of about 1.1 A cm
at mV is measured at 80 K. At 20 K, the dark current is
so low that we reach the limitations of our experimental setup
(a dark current density of A cm corresponds to a dark
current of 40 fA). From dark current density curves, the
1041-1135/$26.00 © 2010 IEEE