of the superlattice. By analyzing the spectral responses with dif-ferent biases across the energy filter, the photoelectron distribu-tion and relaxadistribu-tion in the second miniband of the superlattice can be resolved. In this paper, a model based on the rate equation is proposed to extract the related parameters of the photoelectron transport. The photoelectrons suffer from both intraminiband and interminiband relaxation. The extracted interminiband relaxation time is 11 ps, while the intraminiband one is in the subpicosecond range. In addition, this structure can also be utilized as an infrared photodetector. Our analysis shows that the fast intraminiband re-laxation is the dominant mechanism of the maximum achievable responsivity of the detector and the theoretical model not only pro-vides the insight of the photoelectron transport in the superlattice miniband but also is useful for designing a superlattice infrared photodetector.
Index Terms—Infrared photodetector, intersubband, quantum-well infrared photodetector, superlattice.
I. INTRODUCTION
T
HE BENEFITS of using intersubband infrared photode-tectors for large imaging system have driven studies on electron transport in the popular adopted quantum-well (QW) structures for optimization of performances. One of the methods to analyze the electron transport is using optically stimulated hot-electron spectroscopy (OSHES) [1]. In this method, an en-ergy filter is designed to separate photoelectrons with different energies. Therefore, the energy distribution of the photoelec-trons can be resolved.In addition to QWs, superlattices are another promising struc-ture for infrared detection and has drawn much attention. Su-perlattice with a graded barrier was fabricated for photovoltaic detection in the wavelength ranges of 3.6–6.3 and 8–10.5 m in 1988 [2] and 1990 [3], respectively. In 1991 [4], superlat-tice alone was applied in the detection for 5–10- m wavelength range. Superlattice with a blocking layer for low bias operation was demonstrated in 1992 [5]. The earlier works of the
super-Manuscript received July 10, 2002; revised October 21,2002. This project is supported by National Science Council of Taiwan under Contract NSC 89-2215-E-002-058.
C.-C. Chen, H. C. Chen, M.-C. Hsu, W.-H. Hsieh, and C.-H. Kuan are with the Department of Electrical Engineering and, Graduate Institute of Electronics En-gineering, National Taiwan University, Taipei 10617, Taiwan, R.O.C. (e-mail: kuan@cc.ee.ntu.edu.tw).
S.-Y. Wang is with the Institute of Astronomy and Astrophysics Academia Sinica, Taipei 10617, Taiwan, R.O.C.
C.-P. Lee is with the Department of Electronic Engineering, National Chiao Tung University, Hsinchu 30050, Taiwan, R.O.C.
Digital Object Identifier 10.1109/JQE.2002.807175
lower operating voltage, more suitable for low background and low operating temperature conditions [6] than the conventional QWIP. However, experimental and theoretical studies about the transport and relaxation of the photoelectrons within the mini-band of the superlattice do not attract much attention.
In order to study the photoelectron transport and relaxation within the superlattice miniband, a structure that is composed of a 14-period GaAs–Al Ga As superlattice and a high-pass energy filter (similar to the of OSHES method) is studied in this paper. In addition, this structure is also applicable as an infrared photodetector for 10 m [7]. Because of the energy filter, the photoresponse is tunable by external bias to enable multiwavelength operation and is therefore useful in versatile applications, including target discrimination and temperature sensing. Meanwhile, according to the measured voltage-dependent photoresponses, a model based on the rate equations that involves with the absorption, transport and relaxation of the electrons in the superlattice is proposed.
This paper is organized as follows. Section I is a brief in-troduction. The structure and measured photoresponses are pre-sented in Sections II and III, respectively. In Section IV, a model is suggested to analyze the photoelectron transport from the ob-served voltage-dependent photoresponses. In the analysis, it is found that the photoelectrons in the second miniband suffer not only interminiband but also intraminiband relaxation. The pho-toelectrons relax back into the first miniband through intermini-band relaxation, while the intraminiintermini-band one causes them to ac-cumulate in the bottom states of the second miniband. The latter process is much faster than the former and poses a limitation on the short-wavelength responsivity. Section V summarizes this paper.
II. SAMPLESTRUCTURE
The schematic band diagram of the structure for analyzing the photoelectron transport in the superlattice miniband is shown in Fig. 1. The structure, grown by molecular beam epitaxy on a semi-insulating GaAs substrate, contains a 500-nm heavily doped GaAs bottom contact layer, a 150-nm Al Ga As high-pass energy filter, a 14-period GaAs–Al Ga As su-perlattice, and a 400-nm heavily doped top contact layer. Con-tact layers are both doped with 1 10 cm of Si. Each pe-riod of the superlattice consists of 6.5 nm GaAs well doped
with 5 10 cm of Si and 3.5-nm undoped Al Ga As
barrier. The shadowed region in Fig. 1 shows the two mini-bands formed in the superlattice structure. Each miniband
Fig. 1. Schematic band diagram of the superlattice infrared photodetector. The superlattice is composed of 6.5-nm GaAs well and 3.5-nm Al Ga As barrier. The 150 nm Al Ga As current blocking layer and the superlattice barriers are both undoped and the GaAs well region of the superlattice is doped with 52 10 cm of Si. From the conduction band edge of the GaAs, two minibands are formed in the superlattice region that ranges from 47 to 60 meV and 174 to 242 meV, respectively.
Fig. 2. Spectral responsivity from 0 to00.6 V at various biases at 25 K. The solid curves are the measured data and the dashed curves are the simulated ones. The responsivity at 9.7m is suppressed at zero bias, while it increases with applied bias and dominates the spectral responsivity at00.6 V.
tains 14 resonant-tunneling states. The conduction band offset of Al Ga As relative to GaAs is taken as 810x meV [8]. Rel-ative to the conduction band edge of GaAs, the first miniband ranges from 47 to 60 meV and the second miniband ranges from 174 to 242 meV.
The barrier height (186 meV) of the energy filter is designed to be higher than the bottom state of the second miniband in the superlattice. At around zero bias, photoelectrons with energy higher than the energy filter can pass through it, while those with energy lower are blocked. At high biases, due to the assistance of the strong electric field on the blocking barrier, part of the blocked photoelectrons can tunnel through the current blocking barrier and contribute to the photocurrent. As a result, the energy filter can be used to distinguish the energy of the photoelectron and resolve the photoelectron distribution versus energy.
The sample was fabricated into a 200 400 m mesa by standard photolithography, chemical wet etching, evaporation and lift-off processes. The top and bottom contacts were made by evaporating 100-nm Ni–Ge–Au and 200-nm Au. After evap-oration and lift-off, thermal annealing at 390 C was performed
Fig. 3. Calculated and measured responsivity versus bias at 6.7 and 9.7m. The solid (open) squares and triangles represent the measured (calculated) responsivity at 6.7 and 9.7m, respectively. The open circles represent the responsivity at 6.7m calculated by counting only the photoelectrons with energy higher than the energy filter.
to make ohmic contacts with the contact layers in the semi-conductor. Finally, a 45 facet on the substrate was polished to allow the detector to probe the TM polarized infrared radiation.
III. MEASUREDSPECTRALRESPONSIVITY
The spectral photoresponses were measured by a lock-in am-plifier at various temperatures and biases. A glowbar radiator was used as the light source and the light wavelength was se-lected with a monochromator. The peak responsivity was cali-brated by a blackbody radiation source heated at 500 C. The observed spectral photoresponses show little temperature de-pendence and the representative one at 25 K is shown in Fig. 2. It is obvious that at zero bias, the detected wavelength is pri-marily at 6.7 m. While negative bias increases in magnitude, another peak at 9.7 m rises and dominates the spectral re-sponsivity at high bias region. Therefore, the spectral photore-sponse is tunable by the external bias between 6.7 and 9.7 m, which allows the structure to be applicable in versatile applica-tions including temperature sensing and target discrimination. This experimental result agrees with the aforementioned design principle.
The measured voltage dependent responsivity at 6.7 and 9.7 m are shown as the solid squares and triangles in Fig. 3, respectively. The nonzero 6.7- m photoresponse at zero bias indicates the photocurrent is caused by the emission of the pho-toelectrons with energy higher than the energy filter. It is also noted that the 6.7- m photoresponse does not saturate with the bias magnitude, but increases with it instead. This is attributed to the tunneling of the relaxed photoelectrons accumulating in the bottom state of the second miniband through the energy filter by the electric field.
These voltage dependences of the photoresponse show clues for the photoelectron transport in the second miniband of the superlattice. In the next section, a model for the photoelectron transport is suggested to explain the measured voltage-depen-dent photoresponse.
Fig. 4. Schematic drawing of the photoelectron relaxation processes in the second miniband.
IV. MODEL FOR THERESPONSIVITY
As is schematically drawn in Fig. 4, photons with energy excite electrons from the first miniband into an excited state with energy in the second miniband. The number of excited pho-toelectrons that move toward or backward the energy filter are equal in probability. During transport, two primary scattering processes for the photoelectrons are expected to be phonon and impurity scattering. According to the conservation of energy and crystal momentum, the electron scattering with a phonon emitted in a bulk semiconductor is allowed only in the forward direction [9]. Since the structure under investigation is operated at low temperatures ( 55 K), phonon emission is the domi-nant process. Therefore, the phononrelated scattering of photo-electrons in the superlattice structure is expected to be in the forward direction (scattering angle 90 ). The electron scat-tering by impurities in a bulk semiconductor is inversely pro-portional to the forth power of the sine of the scattering angle in the case of negligible screening [10]. Therefore, the scattering of photoelectrons in the superlattice miniband by impurities is also expected to be in the forward direction. The forward-scat-tering property of the photoelectrons simplifies our analysis by decoupling the rate equations related with the respective num-bers of photoelectrons moving toward, and backward from, the energy filter.
Those photoelectrons moving backward from the energy filter can be supplied by the electrons from the top contact instead of the bottom one because of the obstruction by the energy filter to the electrons from the bottom contact. Hence, the photoelectrons moving backward from the energy filter cause merely internal electron circulation and do not contribute to the measurable photocurrent. As a result, only the flux of photoelectrons moving toward the energy filter is responsible for the measured photocurrent. In the following analysis, only the number of photoelectrons moving toward the energy filter is considered.
During the transport in the second miniband, the photoelec-trons may sequentially relax their energy and finally accumu-late in the bottom state of the second miniband. This relaxation mechanism is called the intraminiband relaxation and results
trons gain their kinetic energy through the electric field on the energy filter and also have the scattering probability which in-creases with their kinetic energy.
In an ideal case, the photoelectrons with energy may con-tribute to the carrier flux , which can be calculated from the electron wavefunction. Without taking the scattering in the en-ergy filter into consideration, the transmission probability of the photoelectrons through the blocking layer is easily esti-mated with WKB approximation. In reality, both the intramini-band and interminiintramini-band relaxations in the superlattice have to be considered, while the scattering effect in the energy filter is actually an intraband relaxation and has similar nature with the intraminiband relaxation of photoelectrons in the superlat-tice miniband. Therefore, in our model, this effect is included in the intraminiband relaxation in the superlattice. In summary, as indicated in Fig. 4, the whole mechanism of the voltage-de-pendent spectral responsivity consists of the light absorption in the superlattice region, the photoelectron relaxation during the transport in the second miniband, and the transmission through the energy filter. The scattering in the energy filter is included in the intraminiband relaxation for simplicity.
The aforesaid mechanisms will be discussed one by one in Section IV-A–IV-C, and then a complete equation for the voltage-tunable responsivity will be presented in Section IV-D.
A. Absorption Coefficient
Instead of considering the miniband in the 14-period super-lattice as a continuous one, we calculate the measured quan-tity by counting the contributions from the 14 discrete reso-nant-tunneling states in the miniband. However, it is observed from Figs. 2 and 5 that neither the measured absorption spec-trum nor the spectral responsivity shows the fine structure orig-inated from the 14 discrete transitions. It is attributed that each discrete transition is broadened with an energy comparable to the one separating the neighboring states. In the following anal-ysis, Lorentzian broadening is assumed in the modeling of the absorption spectrum. In addition, we also assume the spectral responsivity is broadened in the same fashion as the absorption spectrum for simplicity.
The transfer matrix method [11] was used to calculate the energies and corresponding wavefunctions of the resonant-tun-neling states that form the two minibands in the superlattice region. According to the calculated oscillation strength, the electrons in the resonant-tunneling state in the first miniband of the superlattice can only be excited to the corre-sponding resonant-tunneling state in the second miniband
Fig. 5. Measured and calculated absorption spectrum for the TM-polarized light of the superlattice structure.
by absorbing infrared light. The overall absorption coefficient counting each individual transition from the first to the second miniband is given by [12], [13]
(1)
where the last multiplication factor in the summation is at-tributed to Lorentzian broadening and is the function defined in the summation and represents the absorption from
the transition. From the curve fitting of
the absorption spectrum, the superlattice parameters can be extracted. The substantial well and barrier widths are, in fact, 6.1 and 4.4 nm, respectively, and the first and second miniband ranges are 53–62 meV and 188–254 meV. These parameters will be utilized in the subsequent analysis of the voltage-de-pendent spectral responsivity. Fig. 5 shows the measured absorption spectrum, along with the fitting results.
B. Carrier Flux and Transmission Probability
One photoelectron represented by wavefunction can gen-erate carrier flux given by [14]
(2) Fig. 6(a) shows the carrier flux of the resonant-tunneling states in the second miniband multiplied by the area of the mesa. The carrier flux shows a maximum near the middle of the second miniband corresponding to the excitation wavelength of 7.6 m
Fig. 6. (a) Carrier flux and normalized photoelectron number induced by 6.7-m light at 0 V. The carrier flux is maximum at around the middle of the second miniband and is minimum in the bottom state of the second miniband. Most of the photoelectrons accumulate in the bottom state of the second miniband due to the fast intraminiband relaxation. (b) Simulation parameters of intraminiband and interminiband relaxation time constants along with carrier escaping time at several biases.
and the maximum value is 35 (5.8) times larger than the bottom (top) of the second miniband. The flux peak at 7.6 m does not appear in the experimental data of the responsivity. It is at-tributed to the relaxation of photoelectrons, which will be dis-cussed in Section IV-E.
The carrier flux is partially reflected by the energy filter and partially passes through it. The associated transmission proba-bility through the blocking layer can be calculated by using the WKB approximation and depends on the carrier energy and the electric field applied on the blocking layer. It can be written as in (3), shown at the bottom of the next page, where is the photoelectron energy, is the electric field applied on the en-ergy filter, and and are the length and barrier height of the energy filter, respectively.
C. Photoelectron Number Generated by Incident Light
The number distribution for the photoelectrons toward the energy filter is actually time-dependent and denoted by . The associated steady-state value is
de-noted by , as defined before. It is due to the
photoelectrons generated by the transition
with incident power and those relaxed from higher energy states than . For the incident power, the photoelectron
anisms such as phonon scattering and impurity scattering. The relaxation of photoelectrons between states only in the second miniband is attributed to the intraminiband relaxation, while re-laxation from the second to the first miniband is the intermini-band relaxation as schematically shown in Fig. 4. To find the photoelectron number in the second miniband generated by in-cident light, the rate equation of each resonant-tunneling state in the second miniband is analyzed. For simplicity, in the following analysis, the momentum perpendicular to the growth direction is neglected. Consider the photoelectron number as a result of
the transition. For the incident optical power
with photon energy , the rate equation at energy in the second miniband is given by
(5)
where is the photoelectron generation rate,
and are the intraminiband and interminiband re-laxation time, respectively, and is the mesa area of the de-tector. It should be noted that the scattering in the energy filter, which is expected to increase with the applied bias, is included in the parameter , since the intraband relaxation in the en-ergy filter has similar nature with the intraminiband relaxation. Therefore, the intraminiband relaxation is actually bias-de-pendent and expected to decrease with bias. The last term in (5) represents the total loss due to the carrier flux . It should
be noted that no matter whether or , the carrier
flux of the photoelectron always represents a loss for ,
be-to the transition rate in (4). For each state with
en-ergy lower than , the generation rate comes
from the intraminiband relaxation from the states with energy between and . In addition to the intraminiband relaxation time, the function is introduced for the intraminiband relaxation as the probability of carriers relaxed from energy state into a lower energy state in the second miniband. Therefore, the generation rate can be written with as in (7), shown at the bottom of the page.Therefore, the steady-state
light-induced carrier number can be found by
solving (6) and (7).
D. Voltage-Dependent Spectral Responsivity
The spectral responsivity is defined as the photocurrent di-vided by the incident power. Because of the intraminiband re-laxation, the photocurrent may be due to the photoelectrons with energy . Therefore, the responsivity can be written as
(8) Because almost all of the external bias is dropped on the en-ergy filter, the electric field on the energy filter may be written as , where is the length of the energy filter. In order to compare the relaxation rate with the escaping carrier flux, the voltage-dependent escaping time constant is defined as
(9) for for for (3) (7)
which represents the time for the photoelectrons to escape the entire superlattice and tunnel through the blocking layer.
It is noted in (8) is obtained by considering only
the transition. The overall voltage-dependent
spectral responsivity can be expressed as
(10)
where is the number of resonant-tunneling states in each miniband. Here, we are going to make several further assump-tions in the calculation of (10). First, since for a high-energy state in the second miniband, there are more states with lower energy into which the photoelectrons can be relaxed, the in-traminiband relaxation rate from the state is therefore as-sumed to be proportional to the number of the states with energy lower than . That is, the intraminiband relaxation time
con-stant is assumed to be and is voltage
dependent to include the scattering effect in the blocking layer. Second, the probability is supposed to be uniformly dis-tributed for the state with energy that is lower than , i.e., . Third, is assumed to be indepen-dent of for simplicity. In this model, the carrier flux and tun-neling probability are calculated values and the interminiband and intraminiband relaxation times are fitting parameters. In ad-dition, the barrier height of the energy filter is also treated as a fitting parameter, since it is grown inaccurately. The calculated results of (10) for the voltage-dependent spectral responsivity are shown by the dashed curves in Fig. 2. The extracted bar-rier height is 208 meV from the conduction band-edge of GaAs. Shown in Fig. 6(b) are the voltage-dependent intraminiband re-laxation time along with the interminiband one and the escaping time under several bias voltages. Fig. 6(a) shows a represen-tative normalized carrier distribution versus resonant-tunneling state energy in the second miniband at zero bias induced by the illumination of the short-wavelength (6.7 m) light. The agree-ment between the experiagree-mental and simulated spectral respon-sivity offers some insight into the photoelectron transport and relaxation mechanisms, which will be discussed next.
E. Discussion
As expected, the photoelectron number distribution in Fig. 6(a) shows a peak at the state , into which the electrons are excited by photons. However, due to the intraminiband re-laxation, almost 97.7% of the total photoelectrons are relaxed and accumulated in the bottom state of the second miniband. The extracted interminiband relaxation time from the fitting is 11 ps for our sample. This value is consistent with the intersub-band relaxation time in the quantum-well infrared photodetec-tors which is in the range from several picoseconds to tens of picoseconds [14]–[18].
The escaping time for the photoelectrons under several bias voltages is shown in Fig. 6(b). It is noted from (3) that under zero bias, the transmission probability approaches zero if the photoelectron energy is lower than the energy filter, while the transmission probability equals 1 if the photoelectron energy is higher than the energy filter. This causes the steep increment of escaping time below the barrier height 208 meV of the energy
filter under zero bias. The fitting barrier height corresponding to the aluminum fraction is actually higher than the designed value of 186 meV ( ). The conduction band offset for Al Ga As is taken as 810 meV . In the growth of the sample, the Al sources for our superlattice barrier and energy filter were different. The superlattice barriers are grown accurately but the energy filter seems not. It is our belief that the fitting parameter is more accurate than the designed one since the applied bias must be as high as 0.6 V to extract the accu-mulated carriers in the bottom state of the second miniband, as shown in Fig. 3.
The extracted intraminiband relaxation time is much shorter than the interminiband one and varies from several tenths of pi-coseconds to several femtoseconds for the bias between 0 and 0.6 V, as shown in Fig. 6(b). The intraminiband relaxation rate increases with the second miniband energy since, for a high-en-ergy state in the second miniband, there are more states with energy lower than it, into which the photoelectrons can be re-laxed into. Besides, the aggravation of intraminiband relaxation rate with applied voltage is attributed to the scattering in the energy filter. The fast intraminiband relaxation compared with the interminiband one is consistent with the fact that the spec-tral responsivity does not show a maximum at the wavelength 7.6 m, corresponding to the maximum of carrier flux which is 35 times lager than that in the bottom of the second mini-band. It is because only a small fraction of the photoelectrons can stay in the high-energy states of the second miniband while most of them are accumulated in the bottom state of the second miniband by sequential intraminiband relaxation. The detailed physical mechanisms about the relaxation time would be com-plicated and will be studied in the future.
Fig. 3 shows the measured and calculated responsivity versus bias voltage at 6.7 and 9.7 m. The data for the responsivity at 6.7 m reveals some interesting results due to the relaxation of the photoelectrons. Here we discuss the responsivity at 6.7 m in more detail. The open squares and circles shown in Fig. 3 are the calculated data corresponding to the responsivity caused by all the excited photoelectrons in the second miniband and by the photoelectrons with energy higher than the energy filter, re-spectively. Due to the small tunneling probability at low bias, most of the photoelectrons accumulated in the bottom state of the second miniband can not contribute significantly to the re-sponsivity. As can be observed from the proximity of the open squares and circles at low biases in Fig. 3, the responsivity is pri-mary dominated by the emission photocurrent, which is caused by the photoelectrons with energy higher than the energy filter. Although the photoelectrons with energy higher than the energy filter are much smaller in number than those accumulated in the bottom state of the second miniband, due to the larger car-rier flux than the smallest one of the bottom state as shown in Fig. 6(a), they can generate 8 mA/W of the responsivity at zero bias.
Even though more photoelectrons accumulated in the bottom state of the second miniband can be collected under high bias, they only generate 32 mA/W at 0.6 V, which is four time larger than the responsivity at zero bias. It is because of the smallest carrier flux of the bottom state in the second miniband. This in-dicates that the responsivity can be enhanced by reducing the
dependent responsivity, including the nonzero responsivity at zero bias at short wavelength, the increment of short-wavelength responsivity with the magnitude of bias, and the responsivity at long wavelengths, which is mainly tunneling photocurrent.
From the discussion above, it was found that the much faster intraminiband relaxation is the dominant factor in limiting the maximum achievable responsivity of the superlattice infrared photodetector. However, the large ratio of the interminiband time constant over the intraminiband one suggests that the su-perlattice structure can be utilized in the design of quantum-cas-cade lasers to achieve population inversion [20], [21]. Neverthe-less, the tunable spectral responsivity exhibits the potentiality in versatile applications, including temperature sensing and target discrimination.
V. SUMMARY
We have designed a superlattice with an energy filter to in-vestigate the photoelectron transport in the superlattice mini-band. In addition, this structure can be applicable as an infrared photodetector, which shows the advantages of lower oprating voltage (0 0.7 V), wider and tunable spectral responsivity, and flexible miniband engineering. A model based on the rate equation is suggested for the measured voltage-dependent be-havior of the photoresponse. From this model, it is found that the intraminiband relaxation process (subpicosecond) is much faster than both the interminiband relaxation (11 ps) and the carrier escaping time (with a minimum of 0.75 ps) and is the dominant mechanism for the responsivity. This indicates that the responsivity can be increased by the reduction of fast in-traminiband relaxation. Therefore, this model not only provides an insight into the transport of the photoelectrons in the super-lattice with an energy filter, but also a tool to further improve the detector performance.
REFERENCES
[1] K. K. Choi, The Physics of Quantum Well Infrared Photodetectors, Sin-gapore: World Scientific, 1997, ch. 8.
[2] A. Kastalsky, T. Duffield, S. J. Allen, and J. Harbison, “Photovoltaic detection of infrared light in a GaAs/AlGaAs superlattice,” Appl. Phys. Lett., vol. 52, pp. 1320–1322, Apr. 1988.
[3] O. Byungsung, J.-W. Choe, M. H. Francombe, K. M. S. V. Bandara, and D. D. Coon, “Long-wavelength infrared detection in a Kastalsky-type superlattice structure,” Appl. Phys. Lett., vol. 57, pp. 503–505, July 1990.
[4] S. D. Gunapala, B. F. Levine, and N. Chand, “Bound to continuum su-perlattice miniband long wavelength GaAs=Al Ga As photoconduc-tors,” J. Appl. Phys., vol. 70, pp. 305–308, July 1991.
[5] K. M. S. V. Bandara, J.-W. Choe, and M. H. Francombe, “GaAs/AlGaAs superlattice miniband detector with 14.5m peak response,” Appl. Phys. Lett., vol. 60, pp. 3022–3024, June 1992.
[11] S. L. Chang, Physics of Optoelectronic Devices. New York: Wiley, 1995, ch. 4.
[12] M. Helm, W. Hilber, and T. Fromherz, “Infrared absorption in superlat-tices: A probe of the miniband dispersion and the structure of the impu-rity band,” Phys. Rev. B, vol. 48, pp. 1601–1606, July 1993.
[13] S. Nojima, “Intraband optical absorption in semiconductor superlat-tices,” Phys. Rev. B, vol. 41, pp. 10 214–10 217, May 1990.
[14] K. L. Wang and P.-F. Yuh, “Theory and applications of band-aligned superlattices,” IEEE J. Quantum Electron., vol. 25, pp. 12–19, Jan. 1989. [15] J. A. Brum and G. Bastard, “Resonant carrier capture by semiconductor
quantum wells,” Phys. Rev. B, vol. 33, pp. 1420–1423, Jan. 1986. [16] B. F. Levine, A. Zussman, S. D. Gunapala, M. T. Asom, J. M. Kuo, and
W. S. Hobson, “Photoexcited escape probability, optical gain and noise in quantum well infrared photodetectors,” J. Appl. Phys., vol. 72, pp. 4429–4443, Nov. 1992.
[17] K. K. Choi, C. Y. Lee, M. Z. Tidrow, and W. H. Chang, “Performance assessment of quantum well infrared photodetectors,” Appl. Phys. Lett., vol. 65, pp. 1703–1705, Sept. 1994.
[18] E. Rosencher, F. Luc, P. Bois, J. Nagle, and Y. Cordier, “Capture time versus barrier thickness in quantum-well structures measured by infrared photoconductive gain,” Appl. Phys. Lett., vol. 63, pp. 3312–3314, Dec. 1993.
[19] E. Rosencher, B. Vinter, F. Luc, L. Thibaudeau, P. Bois, and J. Nagle, “Emission and capture of electrons in multiquantum-well structures,” IEEE J. Quantum Electron., vol. 30, pp. 2875–2888, 1994.
[20] F. Capasso, A. Tredicucci, C. Gmachl, D. L. Sivco, A. L. Hutchinson, A. Y. Cho, and G. Scamarcio, “High-performance superlattice quantum cascade lasers,” IEEE J. Select. Topics Quantum Electron., vol. 5, pp. 792–807, May 1999.
[21] A. Tredicucci, F. Capasso, C. Gmachl, D. L. Sivco, A. L. Hutchinson, and A. Y. Cho, “High-performance quantum cascade lasers with electric-field-free undoped superlattice,” IEEE Photon. Technol. Lett., vol. 12, pp. 260–262, Mar. 2000.
Chun-Chi Chen was born in Taiwan, R.O.C. He received the Ph.D. degree in
electrical engineering from National Taiwan University, Taipei, Taiwan, R.O.C., in 2002. His studies focused on the design and analysis of multicolor superlattice infrared photodetectors.
He is currently with the Research and Development Department of Taiwan Crystal Corporation, Taiwan, R.O.C., involved in developing high-precision oscillators.
H. C. Chen was born in Tainan, Taiwan, in 1977. He received the B.S. and
M.S. degrees in electrical engineering from National Taiwan University, Taiwan, R.O.C., in 1999 and 2001, respectively.
Since 2001, he has been with Media Tek Inc., Taiwan, R.O.C. His research in-terests include infrared photo-detectors, coding theory, optical-storage systems, and VLSI architecture.
Mao-Chieh Hsu was born in Taipei, Taiwan, R.O.C. He received the M.S.
degree in electrical engineering in physics from National Taiwan University, Taipei, Taiwan, in 2000. His Ph.D. studies focus on the design and characteri-zation of infrared devices using the intersubband transitions in quantum well or superlattice structures.
He currently serves in the Research and Development Department, Copax Photonic Corporation, Taiwan, R.O.C., responsible for developing high-speed fiber communication receivers.
Wen-Hsing Hsieh was born in Taipei, Taiwan, R.O.C. He received the M.S.
degree in physics from National ChungHsing University, Taichung, Taiwan, R.O.C., in 1997, and is currently working toward the Ph.D. degree in electrical engineering at National Taiwan University, Taipei, Taiwan, R.O.C.
His current research interests include the high-frequency magnetotransport properties of a two-dimensional electron system and spin obital effects in quantum dots or quantum-point contacts.
Chieh-Hsiung Kuan (M’95) was born in Taipei, Taiwan, R.O.C., in 1962. He
received the B.S. degree in electrical engineering from National Taiwan Uni-versity, Taipei, Taiwan, R.O.C., in 1985, and the M.S.A. and Ph.D. degrees in electrical engineering from Princeton University, Princeton, NJ, in 1990 and 1994, respectively. His Ph.D. work involved dark current and noise characteris-tics of the infrared hot-electron transistors, in cooperation with the U.S. Army Laboratory, Fort Monmouth, NJ.
In 1994, he joined the Department of Electrical Engineering, National Taiwan University, as an Associate Professor. His current research interests include in-frared photodiodes for room-temperature operation, quantum-well inin-frared pho-todetectors and lasers, and topics on how to measure and suppress the noise in detectors.
Dr. Kuan is a member of Phi-Tau-Phi.
Shiang-Yu Wang was born in Taiwan, R.O.C., in 1972. He received the B.S.
degree in physics from National Taiwan University, Taipei, Taiwan, R.O.C., in 1994 and the Ph.D. degree in electronic engineering from National Chiao Tung University, HsinChu, Taiwan, R.O.C., in 1999. His dissertation focused on quantum-well and quantum-dot infrared detectors.
He is an Assistant Research Fellow with the Institute of Astronomy and Astro-physics, Academia Sinica, Taiwan, R.O.C. His research interests include molec-ular-beam-epitaxy techniques, growth and characterization of quantum-well and self-organized quantum-dot infrared detectors, and infrared instrument develop-ment for large telescopes.
Dr. Wang is a member of Phi Tau Phi.
Chien-Ping Lee (M’80–SM’94–F’00) received the B.S. degree in physics from
National Taiwan University, Taipei, Taiwan, R.O.C., in 1971 and the Ph.D. de-gree in Applied Physics from the California Institute of Technology, Pasadena, in 1978.
He was with Bell Laboratories and later Rockwell International until 1987. While at Rockwell, he was a Department Manager, responsible for developing high-speed semiconductor devices. He became a Professor at National Chiao-Tung University, Hsinchu, Taiwan, R.O.C., in 1987. He was also appointed Director of the Semiconductor Research Center and, later, the first Director of the National Nano Device Laboratory. Currently, he is the Director of the Nano Science and Technology Center, National Chiao Tung University. He is well recognized in the field of semiconductor research. He is an expert in compound semiconductor devices and was the pioneer of the development of optoelectronic integrated circuits (OEIC), high electron mobility transistors (HEMTs), and ion-implanted MESFETs. His current interests include semiconductor nano structures, quantum devices, spintronics, and heterjunction bipolar transistors. He has graduated 20 Ph.D. students and more than 40 Master’s degree students.
Dr. Lee was the Founding Chair of the IEEE LEOS Taipei Chapter and has also served in the IEEE EDS Taipei Chapter. He has organized and served in sev-eral international conferences. He was awarded the Engineer of the Year Award from Rockwell in 1982, the Best Teacher Award from the Ministry of Education in 1993, the Outstanding Engineering Professor Award from the Chinese Insti-tute of Engineers in 2000, the Outstanding Research Award from the National Science Council in 1993, 1995, and 1997, the Outstanding Scholar Award from the Foundation for the Advancement of Outstanding Scholarship in 2000, and the Academic Achievement Award from the Ministry of Education in 2001.
