Infrared Physics & Technology 79 (2016) 165–170

Contents lists available at ScienceDirect

Infrared Physics & Technology

journal homepage: www.elsevier.com/locate/infrared

Regular article

Roles of blocking layer and anode bias in processes of impurity-band

transition and transport for GaAs-based blocked-impurity-band
detectors
Xiaodong Wang a,⇑, Bingbing Wang a, Xiaoyao Chen b, Yulu Chen a, Liwei Hou a, Wei Xie a, Ming Pan a
a
No. 50 Research Institute of China Electronics Technology Group Corporation, 200331 Shanghai, China
b
Laboratory of Advanced Material, Fudan University, 200438 Shanghai, China

h i g h l i g h t s

Numerical simulations of GaAs-based BIB detectors have been performed.

Physical mechanisms of carrier transition and carrier transport have been studied.

Effective electric field for the detector is only located in the absorbing layer.

Blocking layer not only can suppress dark current but also attenuate responsivity.

Optimal anode bias exists to achieve the maximum responsivity.

a r t i c l e i n f o a b s t r a c t

Article history: Recently, GaAs-based BIB detector has attracted a lot of attention in the area of THz photovoltaic detec-
Received 8 August 2016 tion due to potential application values in security check and drug inspection. However, the physical
Revised 19 October 2016 mechanisms involving in carrier transition and transport are still unclear due to the poor material quality
Accepted 20 October 2016
and immature processing technique. In this paper, the dark current and THz response characteristics have
Available online 20 October 2016
thus been numerically studied for GaAs-based blocked-impurity-band (BIB) detectors. The key parame-
ters and physical models are constructed by simultaneously considering carrier freeze-out and impurity-
Keywords:
band broadening effects. Roles of blocking layer and anode bias in processes of impurity-band transition
Blocked-impurity-band (BIB)
Blocking layer
and transport are intensively investigated, and the results can be well explained by numerical models. It
Absorbing layer is demonstrated that the effective electric field for the detector is only located in the absorbing layer, and
Anode bias can determine to a large extent the magnitude of the dark current and THz response. While the blocking
Dark current layer not only can suppress dark current but also can attenuate responsivity due to its electric-field mod-
Responsivity ulation effect.
Electric field Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction and newly forming stars lies in the THz spectrum; (2) the molecu-
lar vibrational and rotational modes of most gaseous nebula corre-
The atmosphere can strongly absorb terahertz (THz) signal, spond to THz frequencies; (3) the signals from the furthest galaxies
making the realization of land-based THz detection difficult [1]. are highly doppler red-shifted, and strongest in the THz region [3].
However, space-based detection can overcome this disadvantage, In order to meet the extremely strict requirements for the
and achieve the extremely high signal-to-noise ratio (SNR) [2]. space-based detection including high resolution, large field of
Recently, the concept of space-based THz detection has attracted view, and high frame rate, THz detectors must possess qualifica-
a lot of attention in the field of atmosphere monitoring and astro- tions including high sensitivity, large array, and high speed [4,5].
nomical observation due to the following reasons: (1) The peak of Noise equivalent power (NEP) and response time of the state-of-
blackbody emission from planets, asteroids, cosmic dust clouds, the-art blocked-impurity-band (BIB) detectors are on the order of
1019 WHz1/2 and 108 s, respectively [6]. Moreover, array for-
mat of the BIB detectors can be as large as 2048  2048. Due to
⇑ Corresponding author. these outstanding performances in the frequency range from 0.9
E-mail addresses: [email protected] (X. Wang), [email protected]
(X. Chen).

http://dx.doi.org/10.1016/j.infrared.2016.10.008
1350-4495/Ó 2016 Elsevier B.V. All rights reserved.
166 X. Wang et al. / Infrared Physics & Technology 79 (2016) 165–170

to 20 THz, BIB detectors are universally acknowledged as the most The steady-state 2-D numerical simulations are performed for
suitable detectors for space-based THz applications [7]. the front-illuminated GaAs-based BIB detector, whose structural
BIB detectors can be fabricated based on different semiconduc- and physical models are constructed based on Sentaurus Device
tor material system, i.e., silicon (Si), germanium (Ge), and gallium [16], a commercial package by Synopsys. The fundamental equa-
arsenide (GaAs). Among them, Si-based BIB detectors with the cut- tions include Poisson’s equation, continuity equations, and
off frequency of 6 THz are most mature, and have been widely current-density equations [17]. Carrier recombination and genera-
investigated by technological researchers for practical applications tion rates are incorporated into the continuity equations by carrier
[8,9]. The energies of shallow impurity bound states formed in Ge recombination-generation terms [18]. Carrier-recombination
are much lower than those in Si, giving rise to a lower cutoff fre- terms consist of Shockley-Read-Hall, Radiative, and Auger recom-
quency around 1.4 THz for Ge-based BIB detectors [10–12]. GaAs- bination [19]. Carrier-generation term is primarily the photo-
based BIB detectors can further extend cutoff frequency to excitation generation. Besides, barrier tunneling effect, and veloc-
0.7 THz, and thus is perfectly suitable for the security application, ity saturation effect are simultaneously considered for electrons
especially for security check and drug inspection. Therefore, since and holes [20].
the first prototype device was successfully fabricated in UC Berke- The functional structure of GaAs-based BIB detector is impurity
ley [13], GaAs-based BIB detector has become a hot topic in the band from the energy-band point of view. Free carriers can be fro-
THz-detection research. However, because of restrictions from zen out in the impurity band at low temperature and excited to the
poor material quality and immature processing technique, the conduction band at higher temperature. In order to take this effect
studies on GaAs-based BIB detectors are still on the initial stage. into consideration, it is therefore necessary to introduce incom-
Especially, the physical mechanisms involving in carrier transition plete ionization model, which has been given in Ref. [7]. Carrier
and transport are still unclear. mobility is a key parameter, which can be influenced by tempera-
Physics-based numerical simulation provides an efficient and ture and electric field. Therefore, temperature- and field-
economical way for complementing experiment. Exploring the dependent mobility models must be considered in the simulation.
carrier-transport fundamentals, which are responsible for the Among them, temperature-dependent mobility model can be
opto-electrical behaviors of GaAs-based BIB detectors, allows for expressed as:
pushing ahead the further optimization of device performance.
The aim of this work includes: (1) developing physical models suit-
llow ¼ lRT ðT=T 0 Þf ð1Þ
able for predictive simulation of GaAs-based BIB detectors; (2) illu-
minating the roles of blocking layer and anode bias in processes of
impurity-band transition and transport; (3) revealing the underly- where llow is the low field mobility, lRT is the room-temperature
ing physics of carrier kinetics for dark current and THz response mobility, T is the device temperature, and T0 = 300 K. Additionally,
characteristics. Field-dependent mobility model can be expressed as:

llow
lðEÞ ¼ 1=b
ð2Þ
2. Device description and simulation models ½1 þ ðllow E=msat Þb 

Fig. 1 shows the schematic cross-section of front-illuminated

where E is the electric-field intensity, and vsat is the saturation
GaAs-based BIB detector, whose active layers consists of a 40-
velocity. llow obtained from Eq. (1) must be substituted into Eq.
lm-thick GaAs absorbing layer with Te concentration of
(2) to get the field-dependent mobility l(E).
5  1015 cm3, a nominally intrinsic blocking layer with Te concen-
Impurities including donor and acceptor can form energy levels
tration of 1  1013 cm3. These two layers are formed sequentially
within the bandgap, and can exert significant influences on the
on a high-conductivity GaAs substrate by chemical vapor deposi-
physical properties of semiconductors [21–26]. As the doing con-
tion (CVD). A 70-nm-thick highly Te-doped contact layer with peak
centration increases, the impurity level starts to overlay, and even-
concentration of 3  1019 cm3 is formed upon the blocking layer
tually become the impurity band [27–33]. As a result, the
by the ion implantation. The photosensitive mesa is formed by
broadening in the absorption coefficient spectrum and hopping
etching technology [14]. The anode is formed upon the contact
conductivity [34–38] can occur. In order to simulate the
layer, and the cathode is formed upon the substrate. Surface states
impurity-band effect, the fitted absorption coefficient a is coupled
induced by etching can serve as recombination centers [15], which
into the raytracing model, which has also be given in Ref [39].
can increase dark current and reduce responsivity. In order to sup-
Parameters for key models are listed in Table 1.
press the surface states as much as possible, the surface of the
device is passivated by Si3N4.
Table 1
Key parameters used in the simulation.

THz Wave Parameters (GaAs) Units Electron Hole

Band gap Eg eV 1.519 1.519
Dielectric constant er 1 11.7 11.7
A Refractive index n 1 3.25 3.25
Thermal conductivity j W/cmK 27.8 27.8
Blocking layer Heat capacity c J/(Kcm3) 1.63 1.63
Density of states N cm3 6.3  1014 1.2  1016
Effective mass (relative) m 1 0.067 0.481
Absorbing layer
Room-temperature mobility lR cm2/Vs 8500 400
C C Temperature exponent of mobility f 1 1 2.1
Saturation velocity vsat cm/s 1.125  107 1.125  107
Minimum saturation velocity vsat,min cm/s 5  105 5  105
GaAs high-conductivity substrate SRH lifetime sSRH sec 1  109 1  109
Radiative coefficient RRad cm3/s 2  1010 2  1010
Auger coefficient RAuger cm6/s 1  1030 1  1030
Fig. 1. Schematic cross-section of front-illuminated GaAs-based BIB detector.
 X. Wang et al. / Infrared Physics & Technology 79 (2016) 165–170 167

3. Results and discussion 0.83V 1.05V

350 0.86V 1.10V
0.90V 1.15V
Fig. 2 shows the dark current versus the anode bias for different

Electric field (V/cm)

300 100
0.95V 1.20V
thicknesses of the blocking layer (hBlo) at temperature (T) of 3.8 K. 1.00V

Electric field (V/cm)

A breakdown characteristics of dark current can occur when the 250
anode bias (VAnode) exceed a certain value, which is commonly 50
known as the threshold voltage (VTh). As observed, the dark current 200
is on the order of approximately 1017 A when VAnode < VTh, and
150
starts rising sharply with increasing bias when VAnode = VTh. Even- 0
tually, the dark current tends to be a saturated value with the mag- 100 10 20 30 40
Position ( µm)
nitude of 10 lA. In order to illustrate this phenomenon, the vertical
electric-field distributions for different VAnode are depicted in Fig. 3. 50
As shown in Fig. 3, the electric-field intensity in the blocking layer
0
is much larger than that in the absorbing layer. However, the
0 10 20 30 40
carrier-generation effect in the blocking layer is extremely weak
Position ( µm)
due to the low impurity concentration. On the contrary, a large
number of electrons in the absorbing layer can be generated due Fig. 3. Vertical electric field as a function of position at hBlo = 8 lm for different
to partial ionization of neutral donors. Therefore, the effective elec- VAnode, the origin of abscissa represents the interface between the anode and the
tric field for the detector is only located in the absorbing layer, and blocking layer. Inset shows the enlarged electric-field distributions in the absorbing
can determine to a large extent the magnitude of the dark current. layer.

As VAnode increases, the depletion region extends rapidly from the

top to the bottom of the absorbing layer. This can explain why
350 1µ m 7µ m
the dark current rise sharply when VAnode = VTh. On the other hand, 320 2µ m 8µ m
after the complete depletion of the absorbing layer, the width of 3µ m

Eelctric field (V/cm)

300 9µ m
depletion region (WD) will remain constant. It is known from Eq. 240 4µ m 10µ m
(3) that the generation-recombination current (Igr) as an important Electric field (V/cm) 250 5µ m 11µ m
6µ m 12µ m
contributor of the dark current is proportional to WD. As a result, 160
the dark current gradually tends to be saturated. 200

  80
qni AW D qV
Igr ¼ exp 1 ð3Þ 150
seff 2KT
0
100 0 2 4 6 8 10 12 14
where q is the unit electronic charge; ni is the intrinsic carrier con- Position ( µ m)
centration; A is the detector area; teff is the effective carrier lifetime; 50
K is the Boltzmann constant; T is the device temperature; V is the
voltage applied to the depletion region. 0
Fig. 2 also presents the effect of hBlo on the dark current charac- 0 10 20 30 40 50
teristics. It is found that a thicker blocking layer requires a larger Position ( µ m)
VAnode for dark current breakdown. This is because electric-field
Fig. 4. Electric-field intensity distributions at VAnode = 0.95 V for different thick-
intensity in blocker/absorber interface determines directly the
nesses of the blocking layer. Inset shows the enlarged electric-field distributions in
carrier-collection probability, and the blocking layer can consume the blocking layer.
most of electric field, making electric-field intensity of blocker/
absorber interface reduced. In order to verify our assumption,
electric-field intensity distributions at VAnode = 0.95 V for different
thicknesses of the blocking layer are presented in Fig. 4. It is absorbing layer is decreased monotonically with increasing hBlo,
demonstrate that the electric field in the blocking layer is increased indicating that the electric-field intensity of blocker/absorber
monotonically with increasing hBlo, while the electric field in the interface is reduce with the increased hBlo. If we define VTh as the
horizontal intercept of the maximum slope to the curve of dark
current versus anode bias, the dependence of VTh on hBlo is plotted
in the inset of Fig. 2. Our results reveal that VTh is not a linear func-
-6
10 1.0 tion of hBlo, instead, the slope of a plot of VTh versus hBlo increase
-8
gradually with increasing hBlo. It can thus be inferred that the
10
0.9 blocking layer can effectively suppress dark current, and a thicker
VTh (V)
Dark current (A)

10
-10 blocking layer can give rise to a stronger suppression capability.
0.8 1µm As shown in Fig. 1, GaAs-based BIB device adopts front-
10
-12
4µm illuminated scheme to realize THz detection. Namely, THz radia-
6µm tion illuminates on the device from the topside. Due to the absence
-14 0.7
10 8µm of the impurity-induced absorption in the blocking layer, THz radi-
0 2 4 6 8 10 12 14
-16 hBlo ( µ m) 10µm ation can transmit directly through the blocking layer and be
10 12µm absorbed gradually by the absorbing layer. THz absorption occur-
-18 13.5µm ring in the absorbing layer can be attributed to the impurity band
10
formed between the conduction band and the valence band.
Specifically, the shallow donor (Te) doped into the host material
0.0 0.2 0.4 0.6 0.8 1.0 1.2
(GaAs) by in-situ technique can form impurity level, which has a
Anode bias (V)
bonding energy of 5.7 meV (corresponding to wavelength of
Fig. 2. Dark current versus anode bias with different hBlo. Inset show the VTh as a 217 lm), and a 1 s to 2p excited state transition occurring at
function of hBlo. 4.3 meV (corresponding to wavelength of 286 lm). As doping
168 X. Wang et al. / Infrared Physics & Technology 79 (2016) 165–170

concentration increases, the impurity level can broaden and and thus independent of x; (2) n is a linear decreasing function of
become impurity band, resulting in a reduced energy gap between x, which means that n0 (x) is also independent of x. Therefore, it can
the bottom of the conduction band and the top of the impurity be concluded that JTotal is a monotonic increasing function of the
band. Therefore, GaAs-based BIB device can extend response wave- integration of E(x) over the entire region of absorbing layer
R 
length further into the low-frequency THz regime. In order to EðxÞdx . Turning back to Fig. 4, it is clear that an increase of hBlo
abs R
prove the extension of response wavelength, the optical wave- can result in an accelerating decrease in abs EðxÞdx. This can explain
length (k) and power density (P) of incident wave are chosen as why the decreasing tendency of responsivity becomes more and
328 lm and 4  104 W/cm2, respectively. Responsivity as a key more pronounced with the increased hBlo.
parameter to characterize device performance can be obtained by: Anode bias is another key parameter to influence the responsiv-
IL  I D ity of GaAs-based BIB detector. Fig. 6 shows the responsivity versus
R¼ ð4Þ VAnode for different hBlo at k = 328 lm. As observed, if hBlo is fixed,
Popt
responsivity increases first with the increased VAnode, after attain-
where IL and ID are the light current and the dark current, respec- ing a maximum value, responsivity starts to decrease, and eventu-
tively. IL–ID is the photo-generated current, and Popt is the optical ally tends to be a constant. A peak thus exists in the curve of
power. responsivity versus VAnode. Moreover, both abscissa and ordinate
Fig. 5 presents the responsivity as a function of hBlo for different of the peak is a strong function of hBlo, as shown in Fig. 7. An
VAnode at k = 328 lm. It can be found that if VAnode is fixed, the increase of hBlo gives rise to a linear decrease in the peak responsiv-
increase of hBlo can lead to a monotonic decrease of responsivity, ity (ordinate) and a linear increase in the peak bias (abscissa). In
and the decreasing tendency of responsivity becomes more and order to clarify the reason that both peak and steady-state points
more pronounced with the increased hBlo. Referring to Fig. 4, the exist in the curve of responsivity versus VAnode, electron-velocity
reason can be explained as follows: THz radiation absorbed by and recombination-rate profiles at hBlo = 8 lm for different VAnode
the absorbing layer can excite electrons from the impurity band are presented in Figs. 8 and 9, respectively. For simplicity, in the
into the conduction band, leaving an equal number of ionized following discussion we just take the situation when hBlo = 8 lm
donors in the impurity band. Under the applied electric field, two as an example. It is found from Fig. 6 that as for hBlo = 8 lm, the
different components of photocurrent can thus be generated. They peak bias and the steady-state bias are around 1.3 V and 1.7 V,
are electron current (Jele) and hopping current (Jhop), respectively. respectively.
Electron current is originated from drift process of excited elec- The velocity saturation model used in the simulation has the
trons in conduction band. Hopping current is originated from the following form:
hopping process of bound electrons in the impurity band, which llow
is a bound electron can escape from its donor by absorbing a pho- lðEÞ ¼   b 1=b ð6Þ
non, and move under the applied electric field, and then re-bind to 1 þ lmlow
sat
E

another donor by emitting a phonon. The total photocurrent (JTotal)

can be obtained by the following formula: where llow denotes the low-field mobility, and can be expressed as:
Z
J Total ¼ J ele þ J hop ¼ ½qEðnÞlele ðnÞ þ qEðnÞlhop ðnÞdn llow ¼ lR ðT=T 0 Þf ð7Þ
Z
where lR and f have been given in Table 1. msat is given by:
¼ qn0 ðxÞ½lele ðxÞ þ lhop ðxÞEðxÞdx ð5Þ

T
msat ¼ Amsat  Bmsat for msat > msat;min ð8Þ
where q is the unit electronic charge, n is the concentration of T0
excited electrons or ionized donors, E is the electric-field intensity,
otherwise, msat = msat,min. The exponent b is temperature dependent
lele is the mobility for electron drift in the conduction, lhop is the
according to:
mobility for electron hopping in the impurity band, and x is the posi-

bexp
tion along the vertical direction. If we assume when VAnode < 1 V, T
electric-field intensity in the absorbing layer is not high enough to b ¼ b0 ð9Þ
T0
trigger carrier-velocity saturation and the incident THz wave can
be linearly absorbed by the absorbing layer, following results can where T denotes the device temperature and T0 = 300 K.
be derived. (1) lele and lhop are constant in the absorbing layer
6
6
5
5
Responsivity (A/W)

1µm 3µm 8µm

4 2µm 4µm 9µm
Responsivity (A/W)

4 6.0 5µm 10µm

6µm 11µm
7µm 12µm
Responsivity (A/W)

3
13µm
3 5.7 13.5µm
0.776V
2
0.8V 5.4
2
0.84V
0.872V 1
1 5.1
0.92V 0.9 1.2 1.5 1.8
1V 0 Anode bias (V)
0 0.8 1.0 1.2 1.4 1.6 1.8 2.0
2 4 6 8 10 12 14 Anode bias (V)
hBlo (µm)
Fig. 6. Responsivity versus VAnode for different hBlo at k = 328 lm. Inset shows the
Fig. 5. Responsivity as a function of hBlo for different VAnode at k = 328 lm. enlarged view around the peak values.
 X. Wang et al. / Infrared Physics & Technology 79 (2016) 165–170 169

6.2 in the blocking layer is saturated. While, electric field in the

2.2 absorbing layer is not large enough to cause electron-velocity sat-
6.0 uration, instead, a larger electric field can give rise to a higher elec-
Peak responsivity (A/W)

2.0 tron velocity. According to Fig. 3, the electric-field intensity

5.8

Peak bias (A/W)

1.8 decreases linearly with shifting position from the blocker/absorber
5.6 Peak responsivity interface to the absorber/substrate interface, and the increased
5.4 peak bias 1.6 VAnode can enhance the electric field throughout the absorbing
5.2 layer. For this reason, electron velocity near the blocker/absorber
1.4
interface is larger than that near the absorber/substrate interface,
5.0 and the increased VAnode can also lift the electron velocity through-
1.2
4.8 out the absorbing layer. It can be concluded from Fig. 8 that if
1.0 VAnode > 1 V, electron velocity in the absorbing layer is approxi-
4.6
2 4 6 8 10 12 14 mately uniform and approaches the saturated value
hBlo (µm) (1.125  107 cm/s), which can also validate the assumption made
in the discussion of responsivity vs. hBlo characteristics.
Fig. 7. Peak responsivity (left axis) and peak bias (right axis) as a function of hBlo at On the other hand, carrier recombination which is the process
k = 328 lm. that an excited electron is recaptured by an ionized donor can also
contribute to THz response. Net carrier-recombination rate (U) is
given by:
7
10
U ¼RG ð10Þ
Electron velocity (cm/s)

where R is recombination rate, and G is generation rate. From Fig. 9,

two important pieces of information can be derived: (1) the sign of
0.83V
10
6 U is negative because G is larger than R, indicating the carrier-
0.84V
generation process predominates over the carrier-recombination
0.85V
process, and thus the absolute value of U is actually net carrier-
0.86V
generation rate; (2) the net carrier-generation rate decreases with
0.87V
increasing VAnode when VAnode > 1.3 V.
0.88V 0.92V
0.89V
Synthesis of the discussion of the preceding two paragraphs can
0.95V
5
0.90V
well explain why both peak and steady-state points exist in the
10 1.00V
curve of responsivity versus VAnode. Specifically, THz-response pro-
0 10 20 30 40
cess of GaAs-based BIB detector can be divided into two phases: (1)
Position ( µ m)
free electron and ionized donor are generated by THz absorption;
Fig. 8. Electron velocity profiles at hBlo = 8 lm for different VAnode. (2) electron drift and electron hopping occur under the applied
electric field before they are collected by the electrodes. When
VAnode < 1 V, THz-response process is dominated by the phase (2),
and responsivity increases with the increased VAnode due to the
-9.0 increased electric field. When VAnode > 1 V, THz-response process
Net recombination rate (cm s )
-3 -1

is dominated by phase (1) instead, and net carrier-generation rate

increases with increasing VAnode when VAnode < 1.3 V. The peak bias
-9.5 is around 1.3 V because net carrier-generation rate starts to decline
when VAnode = 1.3 V, and the steady-state bias is around 1.7 V
because net carrier-generation rate tends to be stable when
-10.0 VAnode = 1.7 V.
1.6V
1.7V
1.3V 1.8V
-10.5 4. Conclusion
1.4V 1.9V
1.5V 2.0V
In this paper, numerical simulation of dark current and THz
10 20 30 40 response characteristics for GaAs-based BIB detectors have been
Position ( µ m) reported. The physical mechanisms involving in carrier transition
and transport have been intensively studied to reveal the roles of
Fig. 9. Net recombination rate profiles at hBlo = 8 lm for different VAnode.
blocking layer and anode bias in processes of impurity-band tran-
sition and transport. Our results reveal that the effective electric
field for the detector is only located in the absorbing layer, and
In Eq. (6), E is the electric-field intensity. According to can determine to a large extent the magnitude of the dark current
Eqs. (6)–(9), the increase of E can lead to the decrease of l(E), and THz response. It is found that the blocking layer not only can
and eventually the saturation of electron velocity. As shown in suppress dark current but also can attenuate responsivity, which
Fig. 8, electron velocity in the blocking layer is uniform, and has is attributed to electric-field modulation by the blocking layer.
attained the saturated value (1.125  107 cm/s) regardless of the Additionally, the optimal anode bias exists to achieve the maxi-
magnitude of VAnode. However, the electron velocity in the absorb- mum responsivity, which is because of the competing effects
ing layer exhibits inhomogeneous distribution. The reason can be between the carrier transport and carrier generation. These inter-
explained as follows: electric field in the blocking layer is much esting results may provide supporting information for pushing
larger than that in the absorbing layer, so that electron velocity ahead the further optimization of GaAs-based BIB detectors.
170 X. Wang et al. / Infrared Physics & Technology 79 (2016) 165–170

Acknowledgement [18] X.D. Wang, W.D. Hu, X.S. Chen, W. Lu, The study of self-heating and hot-
electron effects for AlGaN/GaN double-channel HEMTs, IEEE Trans. Electron
Devices 59 (2012) 1393.
This work was supported by the National Natural Science Foun- [19] W. Hu, Z. Ye, L. Liao, H. Chen, L. Chen, R. Ding, L. He, X. Chen, W. Lu, A
dation of China (Grant No. 61404120). 128  128 long-wavelength/mid-wavelength two-color HgCdTe infrared focal
plane array detector with ultra-low spectral crosstalk, Opt. Lett. 39 (2014)
5130.
[20] W. Qiu, W. Hu, Laser beam induced current microscopy and photocurrent
References mapping for junction characterization of infrared photodetectors, Sci. China-
Phys. Mech. Astron. 58 (2015) 027001.
[1] W. Li, Z. Huang, J. Wang, M. Li, J. Gou, Y. Jiang, Thermal crosstalk simulation [21] V.V. Bannikov, A.L. Ivanovskii, FLAPW-GGA calculations of the influence of Mn,
and measurement of linear terahertz detector arrays, Infrared Phys. Technol. Fe, and Co impurities on the electronic and magnetic properties of layered
73 (2015) 73. oxychalcogenides LaCuSO and LACuSeO, Phys. Solid State 54 (2012) 1117.
[2] X.D. Wang, B.B. Wang, L.W. Hou, W. Xie, X.Y. Chen, M. Pan, Design [22] M. Amato, R. Rurali, S. Ossicini, Doping of SiGe coreshell nanowires, J. Comput.
consideration of GaAs-based blocked-impurity-band detector with the Electron. 11 (2012) 272.
absorbing layer formed by ion implantation, Opt. Quantum Electron. 47 [23] B.L. Zhou, B.H. Zhou, X.W. Chen, G.H. Zhou, Electronic transport for impurity-
(2015) 1347. doped armchair-edge graphene nanoribbons, Eur. Phys. J. B 85 (2012) 121.
[3] B.L. Cardozo, GaAs blocked-impurity-band detectors for far-infrared [24] R. Chandiramouli, First-principles insights on tuning band structure and
astronomy Doctoral thesis, University of California, Berkeley, (2004). transport property of GaN nanotube, Struct. Chem. 26 (2015) 375.
[4] A. Rogalski, Semiconductor detectors and focal plane arrays for far-infrared [25] L.S. Braginsky, M.V. Entin, Intervalley scattering by charged impurities in
imaging, Opto-Electron. Rev. 21 (2011) 346. graphene, JETP Lett. 101 (2015) 325.
[5] A. Rogalski, F. Sizov, Terahertz detectors and focal plane arrays, Opto-Electron. [26] S.H. Do, J.W. Hwang, K.Y. Choi, H. Zhou, H. Nojiri, Nonmagnetic impurity
Rev. 19 (2013) 406. effects in the decorated shastry-sutherland compound Cd(Cu1-xZnx)2(BO3)2
[6] B.A. Wilson, Workshop proceedings: sensor systems for space astrophysics in (0 6 x 6 0.2), J. Korean Phys. Soc. 63 (2013) 1028.
the 21st century, Ser. III Integr. Technol. Plan. 2 (1991) 37. [27] B.B. Wang, X.D. Wang, X.Y. Chen, L.W. Hou, W. Xie, M. Pan, Acquisition of
[7] X.D. Wang, B.B. Wang, L.W. Hou, W. Xie, X.Y. Chen, M. Pan, Analysis of dark optimal operating temperature for epitaxial Si: P blocked-impurity-band
current and spectral response mechanism for Si-based block-impurity-band detector based on temperature-dependent characteristics investigation, Opt.
detectors operating at terahertz regime, Opt. Quantum Electron. 48 (2016) Quantum Electron. 48 (2016) 126.
100. [28] N.V. Agrinskaya, V.I. Kozub, D.S. Poloskin, Suppression of the virtual Anderson
[8] D. Recht, J.T. Sullivan, R. Reedy, T. Bounassisi, M.J. Aziz, Controlling dopant transition in the impurity band of doped quantum well structures, JETP Lett.
profiles in hyperdoped silicon by modifying dopant evaporation rates during 94 (2011) 116.
pulsed laser melting, Appl. Phys. Lett. 100 (2012) 112112. [29] P. Pereyra, D. Weiss, Exchange energy and impurity band effects in the I–V
[9] P. Rauter, T. Fromherz, S. Winnerl, M. Zier, A. Kolitsch, M. Helm, G. Bauer, characteristics of (Ga, Mn)As/GaAs spin injectors, Phys. Rev. B 90 (2014)
Terahertz Si: B blocked-impurity-band detectors defined by nonepitaxial 245310.
methods, Appl. Phys. Lett. 93 (2008) 261104. [30] N.A. Bekin, V.N. Shastin, Stimulated emission on impurity-band optical
[10] T. Wada, Y. Arai, S. Baba, M. Hanaoka, Y. Hattori, H. Ikeda, H. Kaneda, C. Kochi, transitions in semiconductors, Quantum Electron. 45 (2015) 105.
A. Miyachi, K. Nagase, H. Nakaya, M. Ohno, S. Oyabu, T. Suzuki, S. Ukai, K. [31] T. Mizuno, Y. Nagamine, Y. Suzuki, Y. Nakahara, Y. Nagata, T. Aoki, T.
Watanabe, K. Yamamoto, Development for germanium blocked impurity band Sameshima, Impurity doping effects on impurity band structure modulation
far-infrared image sensors with fully-depleted silicon-on-insulator CMOS in two-dimensional n+ and p+ Si layers for future CMOS devices, Jpn. J. Appl.
readout integrated circuit, J. Low Temp. Phys. 184 (2016) 217. Phys. 54 (2015) 04DC05.
[11] M. Hanaoka, H. Kaneda, S. Oyabu, M. Yamagishi, Y. Hattori, S. Ukai, K. Shichi, T. [32] E.Y. Perlin, R.S. Levitskii, A.V. Ivanov, K.A. Eliseev, Nonlinear optical absorption
Wada, T. Suzuki, K. Watanabe, K. Nagase, S. Baba, C. Kochi, Development of in crystals with deep impurities: II. Band-impurity phototransitions involving
blocked-impurity-band type Ge detectors fabricated with the surface- free electrons, Opt. Spectrosc. 119 (2015) 216.
activated wafer bonding method for far-infrared astronomy, J. Low Temp. [33] A. Parisini, M. Gorni, A. Nath, L. Belsito, M.V. Rao, R. Nipoti, Remarks on the
Phys. 184 (2016) 225. room temperature impurity band conduction in heavily Al+ implanted 4H-SiC,
[12] H. Kaneda, T. Wada, S. Oyabu, R. Kano, Y. Kiriyama, Y. Hattori, T. Suzuki, K. Ide, J. Appl. Phys. 118 (2015) 035101.
M. Kato, K. Watanabe, Electrical and photoconductive properties at 1.8 K of [34] S.K. Deshpande, S.N. Achary, R. Mani, J. Gopalakrishnan, A.K. Tyagi, Low-
Germanium p+-i junction device fabricated by surface-activated wafer temperature polaronic relaxations with variable range hopping conductivity in
bonding, Jpn. J. Appl. Phys. 50 (2011) 066503. Fe TiMO6 (M = Ta, Nb, Sb), Phys. Rev. B 84 (2011) 064301.
[13] L.A. Reichertz, B.L. Cardozo, J.W. Beeman, D.I. Larsen, S. Tschanz, G. Jakob, R. [35] Z. Ovadyahu, Microwave-enhanced hopping conductivity: a non-Ohmic effect,
Katterloher, N.M. Haegel, E.E. Haller, First results on GaAs blocked impurity Phys. Rev. B 84 (2011) 165209.
band (BIB) structures for far-infrared detector arras, Proc. SPIE 5883 (2005) [36] M.A. Ormont, I.P. Zvyagin, Hopping conductivity of disordered semiconductors
58830Q. in the intermediate frequency range, J. Nanoelectron. Optoelectron. 9 (2014)
[14] J. Zhu, S. Liu, H. Jiang, C. Zhang, X. Chen, Improved deep-etched multilayer 766.
grating reconstruction by considering etching anisotropy and abnormal errors [37] H. Fu, K.V. Reich, B.I. Shklovskii, Hopping conductivity and insulator-metal
in optical scatterometry, Opt. Lett. 40 (2015) 471. transition in films of touching semiconductor nanocrystals, Phys. Rev. B 93
[15] V.A. Belyakov, V.A. Burdov, K.V. Sidorenko, Effects of surface defects on (2016) 125430.
radiative interband recombination in silicon highly doped with hydrogen-like [38] A. Artemenko, C. Elissalde, U.C. Chung, C. Estournes, S. Mornet, I. Bykov, M.
impurities, J. Surf. Invest.: X-Ray, Synchrotron Neutron Tech. 4 (2010) 987. Magione, Linking hopping conductivity to giant dielectric permittivity in
[16] Synopsys: Sentaurus Device User Guide, Synopsys Inc., USA, 2008. oxides, Appl. Phys. Lett. 97 (2010) 132901.
[17] W.D. Hu, X.S. Chen, Z.H. Ye, A.L. Feng, F. Yin, B. Zhang, L. Liao, W. Lu, [39] X.D. Wang, W.D. Hu, M. Pan, L.W. Hou, W. Xie, J.T. Xu, X.Y. Li, X.S. Chen, W. Lu,
Dependence of ion-implant-induced LBIC novel characteristics on excitation Study of gain and photoresponse characteristics for back-illuminated separate
intensity for long-wavelength HgCdTe-based photovoltaic infrared detector absorption and multiplication GaN avalanche photodiodes, J. Appl. Phys. 115
pixel arrays, IEEE J. Sel. Top. Quantum Electon. 19 (2013) 4100107. (2014) 013103.