QUANTUM-WELL infrared photodetectors (QWIPs) are

Transcription

1 872 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 41, NO. 6, JUNE 2005 High-Responsivity High-Gain In 0:53Ga 0:47As InP Quantum-Well Infrared Photodetectors Grown Using Metal Organic Vapor Phase Epitaxy Amlan Majumdar, Amit Shah, Mahesh Gokhale, Susanta Sen, Sandip Ghosh, Brij Mohan Arora, and Daniel Tsui Abstract We report the growth and fabrication of bound-tobound In 0 53 Ga 0 47 As InP quantum-well infrared photodetectors using metal-organic vapor phase epitaxy. These detectors have a peak detection wavelength of 8.5 m. The peak responsivities are extremely large with pk = 69 A/W at bias voltage = 34 V and temperature = 10 K. These large responsivities arise from large detector gain that was found to be =82at = 3 8 V from dark current noise measurements at =77Kand = 18 4 at = 3 4 V from photoresponse data at = 10 K. The background-limited temperature with F 1 2 optics is BLIP = 65 Kfor0 3 4 V. The highest value of peak detectivity is = cm Hz/W at = 2 9 V and = 77 K, while for BLIP = 65 K, the background-limited detectivity, which exhibits negligible bias dependence, is BLIP = cm Hz/W. Index Terms Infrared detectors, quantum-well (QW) devices. I. INTRODUCTION QUANTUM-WELL infrared photodetectors (QWIPs) are a low-cost and high-uniformity alternative to HgCdTe detectors for infrared imaging applications [1], [2]. Due to the mature nature of GaAs technology, AlGaAs GaAs QWIPs have been the topic of active research for the past decade. Although lattice-matched InGaAs(P) InP technology is extensively used in optical communication systems, only a few reports exist on the use of InGaAs(P) InP multiple QWs (MQWs) for QWIP fabrication [3] [12]. n-type InGaAs(P) InP QWIPs have considerably higher responsivity and photoconductive gain than n-type AlGaAs GaAs QWIPs with similar peak wavelengths. This gain, and hence, responsivity enhancement is due to the higher L valley energy separation in InP-based material system, which leads to both higher mobility and significantly larger photoelectron lifetime [13]. In addition, InP-based QWIPs have potential for two-color detection in Manuscript received December 15, 2004; revised February 7, The work at Princeton University is supported by grants from the Army Research Office and the Air Force Research Laboratory. A. Majumdar was with the Department of Electrical Engineering, Princeton University, Princeton, NJ 08544, USA. He is now with Components Research, Intel Corporation, Hillsboro, OR USA ( amlan.majumdar@intel.com). A. Shah, M. Gokhale, S. Ghosh, B. M. Arora are with Tata Institute of Fundamental Research, Mumbai , India ( sangho10@tifr.res.in; brij@tifr.res.in). S. Sen is with Institute of Radio Physics and Electronics, University of Calcutta, Calcutta , India ( susanta.rpe@caluniv.ac.in). D. C. Tsui is with the Department of Electrical Engineering, Princeton University, Princeton, NJ USA ( tsui@princeton.edu). Digital Object Identifier /JQE the midwavelength infrared (MWIR) and long-wavelength infrared (LWIR) ranges using In Ga As Al In As and InGaAs(P) InP MQWs, respectively, that are both lattice matched to InP. Most InP-based QWIPs demonstrated thus far have been grown by molecular beam epitaxy (MBE) [3] [9]. Many groups have grown high-quality lattice-matched In Ga As InP MQWs by metal organic vapor phase epitaxy (MOVPE), and studied the electronic and optical properties of these MQWs. However, only a few groups have demonstrated MOVPE-grown InP-based QWIPs [10] [12]. These lattice-matched detectors include bound-to-continuum (B-C) and miniband-to-miniband QWIPs with peak detection wavelengths in the 5 9 m range. In this paper, we report the growth and fabrication of bound-to-bound (B-B) In Ga As InP QWIPs using MOVPE, which a low-cost alternative to MBE. The understanding of B-B InGaAs InP QWIPs gained from this work would enable the fabrication of bound-to-quasicontinuum (B-QC) InGaAs InP QWIPs, which are known to have the best detector performance [1], [2]. The upper state in B-QC QWIPs is typically few mevs below the barrier, which leads to higher oscillator strength and quantum efficiency than B-C QWIPs, and higher photoconductive gain than B-B QWIPs. The outline of this article is as follows. We describe the detector structure in Section II, present the results of detector characterization in Section III, and conclude with a summary of this work and possible future work in Section IV. II. DETECTOR STRUCTURE We designed and grew B-B n-type QWIPs on (100) Fe-doped semi-insulating InP substrate. The MOVPE growth was carried out at a chamber pressure of 100 torr with the substrate at temperature C. We used arsine and phosphine as group V sources, trimethylindium and trimethylgallium as group III sources, and silane as an n-type dopant source. The photoactive area consists of 30 periods of 68-Å-wide In Ga As QWs that are separated by 340-Å-thick InP barriers. The QWs are uniformly doped with 5 10 cm Si donors while the barriers are undoped. The MQWs are sandwiched between 0.5- m-thick top and 1- m-thick bottom n -InP contact layers that contain cm Si donors. The composition and thickness of the QW and barrier materials were verified by high-resolution x-ray diffraction (HR- XRD) analysis. The HR-XRD measurements were carried out using a high-resolution diffractometer with a four crystal Bartel /$ IEEE

2 MAJUMDAR et al.: HIGH-RESPONSIVITY HIGH-GAIN In Ga As InP QWIPs GROWN USING MOVPE 873 Fig. 1. Rocking curve of (004) reflection of the QWIP structure measured using HR-XRD. The top curve is the measured data while the bottom curve is the simulated one. The simulated curve has been shifted down for clarity. monochromator in the incident beam path. A typical rocking curve of (004) reflection is shown in Fig. 1. The presence of well developed satellites shows the high quality of the MQW structure. The satellite peaks are fairly symmetrically placed with respect to the central peak, where both the substrate peak and the zero-order superlattice peak seem to have merged [14]. Using standard dynamical diffraction software, we calculated the diffraction profile using QW composition In Ga As and width 68 Å, and InP barriers of thickness 340 Å. The simulated curve is shown in the lower trace of Fig. 1. We find that the positions of satellite peaks are fairly close to the experimental ones. Further refinement is, however, necessary for a better match of the intensities of the simulated and the measured curves. One feature of our detector warrants further comment. We have used n -InP layers as contact layers as opposed to n -layers of the QW material that are generally used [10]. This amounts to having a conventional QWIP with thinner barriers at the two ends of the MQW structure. This is illustrated in Fig. 2 for a six-period InGaAs InP MQW structure at zero bias and K. The conduction band profile was calculated using relevant band parameters. 1 It is clear from Fig. 2 that Fermi level equalization at zero bias leads to electron transfer from the n -contacts into the first and last QW in the MQW region. The depletion width in the n -contacts is 200 Å. A large electric field is present in the first and the last InP barrier that are undoped. Apart from the first and the last QW that also have large electric fields, all other QWs remain under almost flat-band condition as the electrons in these wells come from the Si donors that are present in those wells. As a result of the large electric field in the first and the last InP barrier, the effective barrier height is lowered because the tunneling distance through these barriers is greatly reduced. This barrier height lowering near the two n -contacts could lead to higher photoconductive gain and better performance of our detectors. We also calculated the energy levels of the QWs using the transfer matrix method [2]. The calculated energy levels for the 1 We used the following values in our energy level calculations: E InP = 1:424 ev and E InGaAs = 0:816 ev for energy bandgap, 1E = 230 mev for conduction band offset, and m InP =0:079 m and $m InGaAs = 0:043 m for electron effective mass, where m is the free electron mass. Fig. 2. Calculated conduction band diagram of a 6-period In Ga As InP MQW structure at temperature T = 10 K. The first two energy levels are E and E with wave functions 9 and 9. E is the Fermi level. case of the six-period MQW structure at K and zero bias are shown in Fig. 2 for the four QWs in the middle that are under flat-band conditions. The energy separation between the ground state and the first excited state in each QW is mev. is 23 mev below the InP barriers, and therefore, the transition is a B-B one. From the computed wave functions and, which are indicated in Fig. 2, we calculated oscillator strength. In order to calculate the peak wavelength, we take the depolarization shift into account because the oscillator strength of the B-B transition is large and the QW doping is moderately high [15]. The QW plasma energy is - h mev, and therefore, the expected value of the transition energy is mev, which corresponds to m. III. DETECTOR CHARACTERIZATION A. Intersubband Absorption We first characterized the optical properties of the MOVPE-grown wafer by measuring intersubband absorption at room temperature K with a Fourier transform infrared (FTIR) spectrometer. The absorption spectrum, measured using the single-pass Brewster angle incidence geometry, is shown in Fig. 3. We fitted a Lorentzian line shape in energy, to the data and obtained a peak absorption coefficient of cm, a peak energy of mev, and a spectral width, defined as the full-width at half-maxima, of mev. As temperature is lowered, the absorption spectra is expected to get narrower with increasing by 30% to cm at K [1]. This would result in a low temperature quantum efficiency of % for unpolarized light. Here, m is the length of the active region. The peak energy, which corresponds to a wavelength of m, is smaller than that calculated for low temperatures. This is consistent with the observation of a small red-shift of with increasing in III V quantum wells [16]. The red-shift of can be explained if the temperature dependence of material parameters and effects such as exchange-correlation are taken into account [17]. The spectral width of mev mev %

3 874 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 41, NO. 6, JUNE 2005 Fig. 3. Absorption coefficient measured at room temperature using the single-pass Brewster angle =72 geometry. The dashed curve is a fit ofa Lorentzian line shape in energy, which yields a peak at E = 143 mev with a width =43meV. Inset: schematic of the Brewster angle geometry. obtained from our measurements is slightly larger than mev mev % obtained by Gunapala et al. for their MBE-grown B-B QWIP [3]. This larger width could arise from nonhomogenous broadening due to variations in the QW width, which appear to be slightly larger for our MOVPE-grown structure than that in MBE-grown structures. We determined the oscillator strength from the measured integrated absorption strength. Using the area of the Lorentzian line shape that is fitted to the data in Fig. 3 for, we find, which is in excellent agreement with the calculated value of. The slightly larger value of the deduced oscillator strength could be due to uncertainties in the material parameters used in our calculations or due to slightly larger actual doping density in the grown wafer. B. Responsivity We processed 45 edge coupled QWIPs with m mesas using conventional lithographic techniques for detector characterization. Spectral responsivity of these detectors was measured in the K temperature range. As shown in the inset of Fig. 4(a), we apply bias voltage to the top contact while keeping the bottom one grounded and measure photocurrent using standard ac lock-in techniques. Spectral responsivity of a typical detector at K is shown in Fig. 4. The peak wavelength is and m under positive and negative bias, respectively. The corresponding peak energies, and mev, are in good agreement with the computed value of mev. increases with from mat Vto and mat and V. While this wavelength shift is small, it is larger than what we calculated for symmetric InGaAs InP QWs under applied bias and what has been observed for symmetric AlGaAs GaAs QWs under applied bias [18]. This observation indicates the presence of asymmetry in the QWs, which is presumably due to the nonequivalence of InGaAs InP and InP InGaAs interfaces [14]. The cutoff wavelength, defined at the half-maxima, is and m under positive and negative bias, respectively. The bandwidth,defined as the full width at half maxima, Fig. 4. Responsivity R of a 45 -edge coupled detector at a temperature T = 10 K and under (a) positive and (b) negative bias V. (a) V = 3:4; 3:3;3:0;2:5;2:0; and 1:5 V (top to bottom curve). Inset: scheme for applying bias V to and measuring photcurrent I of an edge-coupled detector. (b) V = 03:5;03:0;02:5;02:0; and 01:5 V (top to bottom curve). is slightly larger for positive bias voltages with m % under positive bias and m % under negative bias. One should note that the responsivity spectra shown in both Figs. 4(a) and (b) is asymmetric with respect to. The spectra have a longer tail on the shorter wavelength side that leads to a larger broadening at shorter wavelengths. Both the asymmetric broadening of the responsivity spectra and the presence of a high-energy tail are due to the finite lifetime of photoelectron in the excited bound-state [19]. Under an applied bias that leads to band bending in the barriers, the probability of the bound photoelectrons to tunnel out of the QW increases with energy, which results in a shorter tunneling time. The shorter tunneling time broadens the line width of the excited bound-state due to the uncertainty principle. The broadening is energy-dependent and larger at higher energies, which leads to the long high-energy tail in the responsivity spectra. The peak values of responsivity are and 5.6 A/W at and V, respectively. These extremely large values of are reasonably higher than those of the previous MBE-grown n-type In Ga As InP QWIPs with similar peak wavelengths [3], [7]. These values of are also about five times larger than the ones obtained in AlGaAs GaAs QWIPs with similar peak wavelengths and QW doping densities [1], [2]. We plot the electric field dependence of in the K temperature range in Fig. 5. increases almost exponentially with, which indicates that electric field

4 MAJUMDAR et al.: HIGH-RESPONSIVITY HIGH-GAIN In Ga As InP QWIPs GROWN USING MOVPE 875 Fig. 5. Electric field dependence of peak responsivity R of our 45 -edge coupled detector in the temperature range T =10 60 K(filled gray symbols). Data from [3] and [10] [12] are also shown for comparison. The three data sets from [12] correspond to QW doping densities of cm (4), cm ( ), and cm (}). TABLE I QWIP STRUCTURE PARAMETERS assists the tunneling of photoelectrons out of the QW [1]. This confirms that the detector is a B-B QWIP as we had designed. In comparison to the K responsivity spectra, the higher temperature spectra have slightly larger values of and substantially higher values of ( % higher at K). The small red-shift of the peak position with increasing temperature can be accounted for from the temperature dependence of material parameters [16], [17]. The lowering of responsivity with decreasing temperature could be attributed to carrier freezeout at lower temperatures [20]. We have also plotted the electric field dependence of from [3], [10], [11], and [12] in Fig. 5. The details of these structures and ours are summarized in Table I. It is clear from this figure that our B-B detector has comparable or higher at electric fields greater than 25 kv/cm under positive bias. C. Dark Current We measured detector dark current in the range K. The - characteristics in the K range are plotted in Fig. 6. In this figure, we have also plotted the background photocurrent that was generated by 300 K radiation incident on the detector through optics. The detector temperature was fixed at K for this measurement. All - Fig. 6. Bias V dependence of dark current I (solid lines) and background photocurrent I (dashed line). Temperature T is in the K range in steps from 5 K for the I curves (bottom to top) while T =10Kfor the I curve. Inset: activation energy E versus V that was extracted from the T dependence of I. curves as well as the - curve are asymmetric. For the same is larger under negative bias while is larger under positive bias. The background-limited temperature is defined as the temperature when [1], [2]. From Fig. 6, we determine that K under positive bias up to V. We cannot compare this value of with those obtained for other InGaAs InP QWIPs as the - curves and values were not reported. The of our detector is, however, comparable to that of AlGaAs GaAs B-B QWIPs and slightly lower than that of AlGaAs GaAs B-C QWIPs with similar values of cutoff wavelengths [1], [2]. We also extracted electron activation energy from the temperature dependence of. The deduced values of are shown in the inset of Fig. 6. decreases with increasing because the top of the energy barrier becomes thinner at larger bias and, hence, becomes more transparent to tunneling [21]. At low bias voltages, mev. For a QW with barrier height, the conduction band offset,, where is the ground-state energy from the bottom of the QW and is the Fermi energy measured from. Using mev, mev that was found from our calculations, and mev for the nominal QW doping density, we obtain mev, which is in excellent agreement with the deduced value of 145 mev. The magnitude of is noteworthy. It is higher than that for the previous MBE- and MOVPE-grown QWIPs for the same electric field [3], [10]. The larger could be due to unoptimized InGaAs InP interfaces or traps in the InP barriers. Optimizing the growth conditions, i.e., chamber pressure and substrate temperature, and the use of purer sources should significantly lower dark current, and hence, considerably improve. It is also interesting to note that our QWIPs have larger yet smaller for positive bias. This behavior was also observed by Gunapala et al. [3] for their MBE-grown QWIPs but the reverse trend, i.e., smaller and larger for positive bias, was observed by Andersson et al. [10] for their MOVPE-grown QWIPs. The observation of larger and smaller under either positive or negative bias polarity is thought to be due

5 876 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 41, NO. 6, JUNE 2005 Fig. 7. Electric field dependence of noise gain g obtained from dark current noise measurements at temperature T =77Kand photoconductive gain g obtained from the photoreponse data at T =10K. The g [10] are also shown for comparison. values from [3] and the inherent asymmetry of the QWs [14]. Usually for MOVPE growth, InGaAs layers grown on InP result in sharp interfaces while InP layers grown on InGaAs do not due to a remaining tail of As. Due to the graded nature of the InP-on-InGaAs interface, we would expect a smaller under positive bias, which is opposite of what we found. One possibility is the presence of InP in a transition layer at the InGaAs-on-InP interface that would lower the conduction band offset at that interface, and hence, lower the activation barrier. We used a 1-s pause between the growth of alternate layers of InGaAs and InP. Optimizing the pause sequence further could solve the above-mentioned problem of a InGaAsP transition region between alternate In- GaAs and InP layers, and hence, lead to symmetric dark current and activation characteristics. D. Noise Gain and Photoconductive Gain We measured dark current noise at K as a function of and evaluated noise gain using. We used a low-noise current amplifier and a spectrum analyzer for these measurements [22], [23]. The deduced values of, which are plotted in Fig. 7 as filled circles, are extremely large compared to AlGaAs GaAs detectors [1], [2]. We found at V and 82 at V. We also evaluated photoconductive gain from the relation, where % is the low temperature quantum efficiency for unpolarized light and is the peak responsivity shown in Fig. 5. The extracted values of are shown in Fig. 7 as filled squares. We find at V. Such large gains in B-B QWIPs result from avalanche multiplication due to impact ionization initiated by photoelectrons whose kinetic energy, which is gained from the electric field, exceeds the energy level difference [24]. These high-energy photoelectrons lose part of their kinetic energy to ground-state electrons in that are then promoted to the first excited state, from where they can tunnel out to create gain. We have also plotted the electric field dependence of from [3] and [10] for comparison. The photoconductive gain of our detector is very similar to that of [10] and significantly larger than that of [3]. This similarity with [10] and difference with [3] can be explained by the difference in the contact layers. Our Fig. 8. Bias dependence of peak detectivity D at temperature T =77K (squares) and background-limited detectivity D for T < T =65K (circles). detector and the one in [10] has InP contacts while the detector in [3] has InGaAs contacts. As explained earlier in Section II, the use of InP contacts leads to the lowering of the effective barrier height near the two n -contacts, which leads to higher photoconductive gain. E. Peak Detectivity and Background-Limited Detectivity Since K under positive bias, the detector would be under dark current limited conditions at K. A commonly used figure of merit for such conditions is peak detectivity [1]. It is given by, where cm is the detector area and Hz is the bandwidth. The calculated values of are shown in Fig. 8. We obtained cm Hz/W at K and V. Due to different bias dependences of and peaks at V with a value of cm Hz/W. This value of is 16 times smaller than that deduced by Gunapala et al. [3] for their In Ga As InP B-B detector, which had a cutoff wavelength m. One should note that, where is the cutoff energy [1], [2]. Taking into account the fact that the cutoff energy of our detector is smaller as m, the of our detector is only 3 times smaller than that of the detector reported by Gunapala et al. [3]. We believe that the lower of our detector is due to larger dark current, and therefore, can be significantly improved by optimizing the growth conditions. For K, our detector is under background-limited conditions, where the relevant figure of merit is the background-limited detectivity, which is given by [1]. Here is the background photocurrent noise given by. The calculated values of are plotted in Fig. 8. exhibits negligible bias dependence and has a peak value of cm Hz/W. IV. CONCLUSION In conclusion, we have grown and fabricated lattice-matched n-type In Ga As InP QWIPs using MOVPE. The detector utilizes a B-B intersubband transition and has a peak wavelength of m. The peak responsivities are extremely large

6 MAJUMDAR et al.: HIGH-RESPONSIVITY HIGH-GAIN In Ga As InP QWIPs GROWN USING MOVPE 877 and reach A/W at V and K. The large responsivities result from large gain that was found to be at V from dark current noise measurements at K and at V from photoresponse data at K. The background-limited temperature is 65 K for V and optics, which is comparable to that of AlGaAs GaAs QWIPs with similar cutoff wavelengths. The peak detectivity is cm Hz/W at K and V while for K, the background-limited detectivity, which exhibits negligible bias dependence, is cm Hz/W. The detector dark current is larger than expected and can be reduced by optimizing the MOVPE growth conditions. This should significantly improve, and of the detector. Based on the promising results presented here, one can attempt the MOVPE-growth of B-QC In Ga As InP QWIPs that should lead to better detector performance. One can also attempt the growth of broad-band QWIPs using simple superlattices [25] or binary superlattices [26] by incorporating lattice-matched InGaAs InGaAsP InP MQWs as well as the growth of two-color MWIR/LWIR QWIPs by using In Ga As Al In As and In Ga As InP MQW stacks [27] [29]. ACKNOWLEDGMENT The authors would like to thank A. Bhattacharya for helpful suggestions, and A. D. Deorukhkar and R. B. Driver for their help with device processing. REFERENCES [1] B. F. Levine, Quantum-well infrared photodetectors, J. Appl. Phys., vol. 74, pp. R1 R81, Oct [2] K. K. Choi, The Physics of Quantum Well Infrared Photodetectors. River Edge, NJ: World Scientific, [3] S. D. Gunapala, B. F. Levine, D. Ritter, R. Hamm, and M. B. Panish, InGaAs/InP long wavelength quantum-well infrared photodetectors, Appl. Phys. Lett., vol. 58, pp , May [4] S. D. Gunapala, B. F. Levine, D. Ritter, R. A. Hamm, and M. B. Panish, Lattice-matched InGaAsP/InP long-wavelength quantum-well infrared photodetectors, Appl. Phys. Lett., vol. 60, pp , Feb [5] S. D. Gunapala, B. F. Levine, D. Ritter, R. Hamm, and M. B. Panish, InGaAs/InP hole intersubband normal incidence quantum-well infrared photodetector, J. Appl. Phys., vol. 70, pp , Mar [6] D. K. Sengupta, S. L. Jackson, D. Ahmari, H. C. Kuo, J. I. Malin, S. Thomas, M. Feng, G. E. Stillman, Y. C. Chang, L. Li, and H. C. Liu, P-type InGaAs/InP quantum-well infrared photodetector with peak response at 4.55 m, Appl. Phys. Lett., vol. 69, pp , Nov [7] C. Jelen, S. Slivken, T. David, M. Razeghi, and G. J. Brown, Noise performance of InGaAs-InP quantum-well infrared photodetectors, IEEE J. Quantum Electron., vol. 34, no. 7, pp , Jul [8] Y. Gusakov, E. Finkman, G. Bahir, and D. Ritter, The effect of strain in InP/InGaAs quantum-well infrared photodetectors on operating wavelength, Appl. Phys. Lett., vol. 79, pp , Oct [9] J. Jiang, K. Mi, R. McClintock, M. Razeghi, G. J. Brown, and C. Jelen, Demonstration of focal plane array based on Al-free GaInAs-InP QWIP, IEEE Photon. Technol. Lett., vol. 15, no. 9, pp , Sep [10] J. Y. Andersson, L. Lundqvist, Z. F. Paska, K. Streubel, and J. Wallin, Long-wavelength quantum-well infrared detectors based on intersubband transitions in InGaAs/InP quantum wells, Proc. SPIE, vol. 1762, pp , [11] L. Pham, X. S. Jiang, and P. K. L. Yu, High-responsivity intersubband infrared photodetector using InGaAsP InP superlattice, IEEE Electron Dev. Lett., vol. 14, no. 2, pp , Feb [12] M. Erdtmann, A. Matlis, C. Jelen, M. Razeghi, and G. Brown, Highresponsivity GaInAs/InP quantum-well infrared photodetectors grown by low-pressure metalorganic chemical vapor deposition, Proc. SPIE, vol. 3948, pp , [13] O. O. Cellek and C. Besikci, Detailed investigation of electron transport, capture, and gain in Al Ga As/GaAs quantum-well infrared photodetectors, Semicond. Sci. Technol., vol. 19, pp , [14] J. M. Vandenberg, D. Gershoni, R. A. Hamm, M. B. Panish, and H. Temkin, Structural perfection of InGaAs/InP strained-layer superlattices grown by gas source molecular-beam epitaxy: A high-resolution x-ray diffraction study, J. Appl. Phys., vol. 68, pp , Oct [15] S. J. Allen, D. C. Tsui, and B. Vinter, On the absorption of infrared radiation by electrons in semiconductor inversion layers, Solid State Commun., vol. 20, pp , Oct [16] B. C. Covington, C. C. Lee, B. H. Hu, H. F. Taylor, and D. C. Streit, Infrared intersubband absorption in GaAs/AlAs multiple quantum wells, Appl. Phys. Lett., vol. 54, pp , May [17] M. O. Manasreh, F. Szmulowicz, D. W. Fischer, K. R. Evans, and C. E. Stutz, Intersubband infrared absorption in a GaAs/Al Ga As quantum well structure, Appl. Phys. Lett., vol. 57, pp , Oct [18] A. Harwit and J. S. Harris, Observation of Stark shifts in quantum well intersubband transitions, Appl. Phys. Lett., vol. 50, pp , Mar [19] B. F. Levine, C. G. Bethea, K. K. Choi, J. Walker, and R. J. Malik, Tunneling lifetime broadening of the quantum well intersubband photoconductivity spectrum, Appl. Phys. Lett., vol. 53, pp , July [20] K. K. Choi, C. J. Chen, and D. C. Tsui, Corrugated quantum-well infrared photodetectors for material characterization, J. Appl. Phys., vol. 88, pp , Aug [21] B. F. Levine, C. G. Bethea, G. Hasnain, V. O. Shen, E. Pelve, R. R. Abbott, and S. J. Hsieh, High sensitivity low dark current 10 m GaAs quantum-well infrared photodetectors, Appl. Phys. Lett., vol. 56, pp , Feb [22] B. Xing, H. C. Liu, P. H. Wilson, M. Buchanan, Z. R. Wasilewski, and J. G. Simmons, Noise and photoconductive gain in AlGaAs/GaAs quantum well intersubband infrared photodetectors, J. Appl. Phys., vol. 76, pp , Aug [23] D. C. Wang, G. Bosman, Y. H. Wang, and S. S. Li, Noise performance of bound-to-miniband transition III-V quantum-well infrared photodetectors, J. Appl. Phys., vol. 77, pp , Feb [24] B. F. Levine, K. K. Choi, C. G. Bethea, J. Walker, and R. J. Malik, Quantum well avalanche multiplication initiated by 10 m intersubband absorption and photoexcited tunneling, Appl. Phys. Lett., vol. 51, pp , Sep [25] 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 , Apr [26] A. R. Ellis, A. Majumdar, K. K. Choi, J. L. Reno, and D. C. Tsui, Binary superlattice quantum-well infrared photodetectors for long-wavelength broad-band detection, Appl. Phys. Lett., vol. 84, pp , Jun [27] A. Majumdar, K. K. Choi, J. L. Reno, L. P. Rokhinson, and D. C. Tsui, Two-color quantum-well infrared photodetector with voltage tunable peaks, Appl. Phys. Lett., vol. 80, pp , Feb [28], Electron transfer based voltage tunable two-color quantum-well infrared photodetectors, Infrared Phys. Technol., vol. 44, pp , [29] A. Majumdar, K. K. Choi, J. L. Reno, and D. C. Tsui, Voltage tunable two-color infrared detection using semiconductor superlattices, Appl. Phys. Lett., vol. 83, pp , Dec Amlan Majumdar received the B.S. degree in electrical engineering from the Indian Institute of Technology, Kanpur, India, in 1995 and the M.A. and Ph.D. degrees in electrical engineering from Princeton University, Princeton, NJ, in 1997 and 2002, respectively. He has been with Components Research, Intel Corporation, Hillsboro, OR, since 2004, working on novel transistors for logic applications. Dr. Majumdar is a Member of the American Physical Society.

7 878 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 41, NO. 6, JUNE 2005 Amit Shah was born in Mumbai, India, in May He received the M.Sc. degree in physics from Mumbai University, Mumbai, India in He joined the Tata Institute of Fundamental Research, Mumbai, India, in 1991 as a Scientific Assistant, where he is currently a Scientific Officer. He has been working on the processing and characterization of semiconductor materials and devices. Currently, he is working on optoelectronic devices of III V compound semiconductors. He is the coauthor of several journal publications. Mahesh Gokhale was born in Daund, India, in December He received the M.Sc. degree in physics from Mumbai University, Mumbai, India, in He joined the Tata Institute of Fundamental Research, Mumbai, India, in 1992 as a Scientific Assistant, where he is currently a Scientific Officer. He has been working on the synthesis and characterization of III V compound semiconductor heterostructures grown by metal organic vapor phase epitaxy. He is the coauthor of several journal publications. Susanta Sen was born in Calcutta, India, in He received B.Sc. degree (hons.) in physics, and the B.Tech., M.Tech., and Ph.D. degrees, all in radio physics and electronics from the University of Calcutta, Calcutta, India, in 1970, 1973, 1975, and 1983, respectively. He joined the Institute of Radio Physics and Electronics, University of Calcutta, India, as a faculty member in 1978, where he has been a Professor since Between 1986 and 1989, he was on leave at the AT&T Bell Laboratories, Murray Hill, NJ, as a Postdoctoral Member of Technical Staff, and worked on quantum electron devices and pioneered the area of resonant tunneling transistors and their circuit applications. He has published over 45 research papers, six book chapters, and participated in over 12 international conferences. He had also assumed key positions in the organization of many international conferences. His current research interests are in the areas of optoelectronic devices including OEICs, VLSI device modeling in the sub-10-nm geometry, and FPGA-based instrumentation. Prof. Sen received the Homi Bhabha Fellowship in He is a Fellow of the Institute of Engineers, India, and a Life Member of the Computer Society of India. Sandip Ghosh was born in Delhi, India, in He received the B.Sc. degree in physics from the University of Delhi, Delhi, India, the M.Sc. degree in physics from the Indian Institute of Technology, Delhi, India, and the Ph.D. degree in physics from the Tata Institute of Fundamental Research-Mumbai University, Mumbai, India, in 1990, 1992, and 1998, respectively. After postdoctoral research at the University of Surrey, Surrey, U.K., and at the Paul Drude Institute, Berlin, Germany, he joined the Tata Institute of Fundamental Research as a Research Fellow in His research involves optical spectroscopy of semiconductor heterostructures. He has coauthored 24 papers in refereed journals. Dr. Ghosh received the Young Scientist Medal of the Indian National Science Academy in Brij Mohan Arora received the B.Tech. degree in electronics and electrical communication from the Indian Institute of Technology, Kharagpur, India, in 1965, and the M.S. and Ph.D. degrees in electrical engineering from the University of Illinois, Urbana-Champaign, in 1968 and 1972, respectively. Since 1972, he has been with the Tata Institute of Fundamental Research, Mumbai, India, where is currently a Professor in the Department of Condensed Matter Physics and Materials Science. He has worked in the areas of plasma display panel, semiconductor materials and devices. Currently, he is active in optoelectronics based on low-dimension structures. Daniel Tsui is the Arthur Legrand Doty Professor of Electrical Engineering at Princeton University, Princeton, NJ. Dr. Tusi is a Member of the National Academy of Science and the National Academy of Engineering, and the recipient of the 1998 Nobel Prize in Physics.