High-performance infrared photodetectors are used in a variety of fields, including for process diagnostics in industrial processes, thermal characterization, and astronomical research for civilian applications. During medical crises, such as the coronavirus disease 2019 pandemic, these detectors were utilized for the assessment of human body temperature and for identifying potential viral infections. In recent years, increasing global military tensions have given rise to a growing demand for array-based infrared detectors that can monitor the operational status of military equipment and detect the movements of personnel in real time. These applications are based on commercial compound semiconductor materials such as indium antimonide (InSb) and mercury cadmium telluride (MCT) [1].

InSb, a binary compound semiconductor, has a fixed band gap, which limits its detection range to the mid-wavelength infrared region corresponding to the band gap. As a result, it is incapable of multiwavelength detection, a critical property required for next-generation infrared detectors; in addition, its detection accuracy is limited when operating at temperatures above 77 K. In contrast, MCT is predominantly used for long-wavelength infrared detection. However, growing MCT wafers with a large wafer size and uniform composition is very costly. Attempts to address these limitations have extensively focused on GaAs-based quantum well infrared photodetectors, which, despite their potential, are also restricted by their ability to operate only at low temperatures and by their low quantum efficiency, which has limited their practical applications.

The InAs/GaSb type II superlattice (T2SL) enables the implementation of semiconductors with new electronic properties and physical characteristics through band gap engineering [2]. The InAs/GaSb T2SL structure is regarded as a compound semiconductor that could replace commercially used materials such as InSb and MCT [3]. Consequently, this structure has been significantly researched over the past two decades. The lattice constants of InAs and GaSb, which are used to form the T2SL structure, correspond very closely to each other, with values of 6.0585 and 6.0955 Å, respectively, at approximately 300 K. These closely corresponding lattice constants allow for the stable growth of InAs/GaSb T2SL devices with a thickness of approximately 3 μm.

The energy band structure of the InAs/GaSb T2SL has a broken type II band alignment, where the conduction band of InAs is positioned about 0.15 eV below the valence band of GaSb. Because of this broken band alignment, T2SL has unique characteristics, including a large electron effective mass (~0.04 m₀) compared to MCT, which detects in the same wavelength range. As a result, T2SL-based devices are predicted to exhibit a weaker tunneling current compared to MCT devices with the same cutoff wavelength. Additionally, due to the broken band structure of the T2SL, holes are localized in the GaSb region, whereas electrons are confined to the InAs region. This spatial separation of electrons and holes helps to lower the Auger recombination rate, which, consequently, has the potential to significantly increase the operating temperature of the device. Moreover, by controlling the thickness of the InAs and GaSb layers, band gap engineering can be employed to enable detection in the mid- or long-wavelength infrared ranges. Recent advancements in III-V compound semiconductor growth technologies have enabled the growth of InAs/GaSb T2SL structures up to approximately 3 inches in diameter. This development offers a more feasible approach for the future production of uniform materials for focal plane arrays (FPAs) [3].

In the early stages of InAs/GaSb T2SL development, device structures primarily employed a PIN configuration. The disadvantage of these devices is that the presence of the PN junction generates a dark current caused by tunneling through the Shockley-Read-Hall centers and diffusion phenomena. As a result, the operation of infrared detectors based on this device structure was limited to very low temperatures. To overcome this drawback, Maimon and Wicks [4] proposed the nBn structure, which involves the growth of a material with a large band gap (AlAsSb) during the fabrication of n-type InAs semiconductors. This design enables the dark current of the device to be partially suppressed. In contrast, an nBn device with the InAs/GaSb T2SL structure offers the advantage of inherently eliminating both the dark and diffusion currents associated with generation-recombination processes owing to the absence of the depletion region that is present in PIN structures [5]. Furthermore, the movement of the majority carriers (electrons) is restricted due to the barrier layer, while the minority carriers (holes) are unaffected by the barrier and can move freely between the electrodes. Consequently, the motion of the minority carriers could be freely controlled by adjusting the polarity of the applied voltage. Kim et al. [6] improved the flow of minority carriers (holes) by using a barrier layer, Al_xGa_1−xSb, with low aluminum content (x = 0.2). Additionally, to further enhance the electron barrier, they set the thickness of the barrier layer to 100 nm in their nBn-structured InAs/GaSb T2SL device. This device was developed as a FPA detector for the mid-infrared region, and its performance was characterized by a noise equivalent temperature difference of 23 mK [6].

Ting et al. [7] proposed a novel T2SL structure referred to as the complementary barrier infrared detector. In this device structure, an electron-blocking unipolar barrier is implemented on one side of the absorber layer, whereas a hole-blocking unipolar barrier is implemented on the reverse side. Specifically, the T2SL using InAs/AlSb served as a hole barrier, while the T2SL with InAs and GaSb enabled the formation of an electron barrier (eB). Upon exposure to infrared radiation, the holes generated in the absorber layer are directed through a tunnel junction in the eB region, which is connected to the bottom contact layer, InAsSb. The proposed T2SL structure exhibited a cutoff wavelength of 9.9 μm. The device demonstrated responsivity of 1.5 A/W and a dark current density of 0.99 × 10⁻⁵ A/cm² at 77 K.

The past decade has witnessed active research on high-performance interband cascade (IC) photodetectors based on the InAs/GaSb T2SL structure as an absorber material [8,9]. This structure consists of three important regions: for interband optical excitation, interband relaxation, and interband tunneling. This design incorporates band gap engineering of III-V compound semiconductors, as well as a multistage design comprising layers of GaSb, InAs, and AlSb with nanoscale thickness, which enables the realization of high-performance infrared detectors. Lei et al. [10] reported on the performance of a 6-stage interband IC detector based on an InAs/GaSb T2SL operating at 300 K. This infrared diode demonstrated a cutoff wavelength of 4.3 μm. The detectivity (D*) was calculated to be 1.5 × 10⁹ cm·Hz^1/2/W and responsivity showed 0.2 A/W. Additionally, an interband detector presented by Tian and Krishna [11] featured a 5-stage structure with the absorber region composed of InAs (7 ML) / GaSb (8 ML). The measured R_oA of this device was 1.25 × 10⁷ Ω·cm² at 120 K, and the Johnson-limited detectivity was found to be 9.73 × 10¹¹cm·Hz^1/2/W at 200 K [11].

A 5-stage IC photodiode design with an electron barrier was adopted in this study [9]. The sample was grown on a 2-inch (001) GaSb substrate using a Veeco Gen-10 solid-source molecular beam epitaxy system. Details of this growth process are reported elsewhere [11,12]. The IC T2SL photodiode consists of an absorber layer made of T2SL, an electron barrier layer, an electron relaxation layer, and electrode layers. The designed InAs (7 ML)/GaSb (8 ML) T2SL structure generates electron-hole pairs in response to incident infrared light. The generated electrons are injected into the electron relaxation region, which is composed of an InAs/AlSb multiple quantum well (MQW). Adjustment of the thicknesses of the InAs and AlSb layers enabled energy ladders with several stages to be created in the conduction band. In this energy ladder, intersubband transitions due to longitudinal optical phonon interactions occur, thereby allowing the electrons to tunnel into the valence band of the next stage [9,11]. Additionally, the InAs/AlSb MQW acts as a barrier for holes. The eB is composed of a GaSb/AlSb MQW.

The grown photodiode wafer was fabricated into single pixel diodes using traditional semiconductor processing techniques. Figure 1 shows a schematic diagram of the photodetector structure. The first step in the fabrication process involved the use of the spin-coating technique to uniformly deposit a positive photoresist, which was developed using a designed mask combined with ultraviolet exposure. After this lithographic process, a mesa structure of the photoresist was retained. Inductively coupled plasma etching with BCl3 gas was employed to etch down to the mid-level of the bottom electrode layer. Next, silicon dioxide was deposited using plasma-enhanced chemical vapor deposition, with the final film thickness being approximately 200 nm. In the next step, the areas designated for passivation, specifically the mesa sidewalls, were preserved, while the remaining areas were removed by dry etching. Finally, a Ti/Pt/Au stack was sequentially deposited on the top and bottom of the photodiode using electron beam deposition. The individual devices were then fixed onto a chip carrier using indium bonding to ensure effective heat transfer as the device measurements were conducted under cooling conditions. The electrical and optical properties of the fabricated devices were evaluated through wire bonding using Au. Figure 2 presents an optical microscope image of the device, after completion of the photolithography, mesa etching, passivation, and metal deposition.

Figure 1. Schematic diagram of the photodetector structure.

Figure 2. Optical microscope image of the device, after the completion of photolithography, mesa etching, passivation, and Ti/Pt/Au metal deposition.

Figure 3 illustrates the apparatus that was used to record the spectrum of the IC T2SL photodetector utilizing a Fourier-transform infrared (FTIR) spectrometer. The measurement system consisted of the FTIR spectrometer, a current amplifier, a cryostat, a temperature controller for the device, a detector interface module, and commercial software (Omnic) for controlling the observed infrared spectrum. The fabricated device was mounted on the cooling finger of the cryostat, which uses liquid helium for cooling. The electrode layer of the device was connected to the external circuit using Au wire bonding. The temperature of the device was controlled via a heater installed within the cryostat. The heater was linked to an external power supply to ensure that the measured temperature was automatically maintained at the desired value. The spectrum was acquired at various temperatures and biases. The data were analyzed using the Omnic software.

Figure 3. Spectral-response measurement system consisting of a FTIR spectrometer for spectrum acquisition of infrared devices, a cryostat for device cooling, and a high-performance current amplifier.

Figure 4 shows the FTIR spectra of the device that were acquired at temperatures ranging from 77 to 300 K. The cutoff wavelengths of the apparatus were approximately 4.0 μm at 77 K and 4.75 μm at 300 K. The cutoff wavelength was redshifted as the temperature of the device increased, indicating that the band gap was becoming narrower. At 77 K, the peak intensity at 3.5 μm was approximately 25 % lower compared to that at 2.75 μm. In contrast, at 300 K, the intensity at 3.5 μm decreased by about 10 % relative to the peak intensity at 2.75 μm. This difference is likely due to a change in the overlap of the wavefunctions of the InAs/GaSb T2SL, which increased at 300 K compared to 77 K, possibly due to temperature-induced strain effects. The absorption peak near 3.4 μm was observed at all temperatures and is believed to be caused by infrared absorption from atmospheric gases between the device and the light source in the FTIR setup. The applied bias was not observed to significantly affect the spectral distribution.

Figure 4. FTIR spectra of the device at temperatures ranging from 77 to 300 K. The cutoff wavelength determined by FTIR was approximately 4.0 μm at 77 K.

The dark current of the photodiode was measured by connecting a semiconductor parameter analyzer to the device mounted on the cryostat. During the dark current measurement, the device was covered with a special cylindrical cover to exclude infrared radiation originating from the ambient temperature of the experimental laboratory background. The temperature was maintained at a constant value throughout the measurement process. Figure 5 presents the results of the dark current density measurements, which were taken at temperature intervals of approximately 25 K, ranging from 50 to 300 K, excluding 77 K. The applied bias was varied from −2.0 to 1.0 V. The dark current densities measured under a bias of 0.1 V were 1.7 × 10⁻⁷, 8.1 × 10⁻⁷, and 1.7 × 10⁻⁵ A/cm² at 77, 150, and 200 K, respectively. Figure 6 presents the dark current density as a function of temperature for bias voltages of −50, 100, and 250 mV. The dark current density of the diode exhibits diffusion-limited behavior in the temperature range of approximately 200 to 300 K, contrary to temperatures below 100 K, at which the dark current density is almost saturated under low bias (−50 mV). In the low-temperature range, the variation in the dark current is substantial in response to changes in the bias. This phenomenon is presumed to be influenced by factors such as trap-assisted tunneling process and surface leakage current due to inadequate passivation.

Figure 5. Dark current density measurements. The dark current was measured using a semiconductor parameter analyzer while varying the temperature from 50 to 300 K.

Figure 6. Dark current density as a function of temperature for applied bias voltages of −50, 100, and 250 mV.

Figure 7 shows the differential resistance-area product (R_dA) of the photodiode as a function of temperature under a bias of −0.1 V. This R_dA is proportional to the carrier lifetime and inversely proportional to the square of the intrinsic carrier concentration. The calculated R_dA values are approximately 9,212 Ω·cm² at 200 K, 196,078 Ω·cm² at 120 K, and 295,787 Ω·cm² at 77 K. Considering that the typical operating R_dA for FPAs is over 1,000 Ω·cm², the fabricated photodiode is judged to possess suitable electrical characteristics for use in FPA wafers.

Figure 7. Calculated differential R_dA as a function of temperature under a bias of −0.1 V.

Figure 8 shows the calculated value of D* at zero bias using the following equation [13]:

Figure 8. Calculated detectivity derived using a high-performance noise measurement system, a chopper, and a blackbody source to obtain the fundamental data for the D* calculation. During the measurements, the temperature was varied between 77 and 300 K while maintaining a zero bias.

where R is the responsivity, Δf is the noise equivalent electrical bandwidth, N is the noise, and A_d is the detector area. The noise spectrum from the photodetector was measured using a low-noise preamplifier (Keithley 428) and an SR770 network analyzer. The D* values calculated in this study were found to be 5.4 × 10¹⁰ cm·Hz^1/2/W at 150 K and 6.2 × 10¹⁰ cm·Hz^1/2/W at 77 K under zero bias. As shown in Fig. 8, the D* value remained constant from 77 to 175 K, but decreased sharply below 200 K as the temperature of the device increased. This decrease in D* is attributed to the strengthening noise current, which lowers R_dA, thereby causing the observed decline in D*.

The optical and electrical characteristics of an IC photodetector were investigated using an InAs/GaSb T2SL as the absorber region for mid-wavelength infrared detection. The cutoff wavelength of the fabricated photodiode was measured to be approximately 4.0 μm at 77 K. The dark current density measured under a bias of −0.1 V was 1.7 × 10⁻⁷ A/cm² at 77 K. In the high-temperature range from around 200 to 300 K, the dark current density exhibited diffusion-limited behavior. In contrast, in the low-temperature region below 100 K, the dark current density approached saturation upon application of a low bias (−50 mV). This is presumed to be caused by surface tunneling, trapassisted recombination, or insufficient passivation at the etched device surface. The calculated R_dA was approximately 295,787 Ω·cm² at 77 K, demonstrating that the fabricated photodiode is suitable for use in FPAs. The calculated D* value for a single device under zero bias was 6.2 × 10¹⁰ cm·Hz^1/2/W at 77 K. Acknowledgments This work was supported by a grant from the National Research Foundation of Korea (NRF) funded by the Korean government (MSIT) (No. 2022R1F1A1072506). The author acknowledges the valuable advice provided by S. Krishna and Z. Tian regarding the device structure, and notes that Z. Tian grew the wafer. Conflicts of Interest The authors declare no conflicts of interest.