The primary challenge for long-wavelength infrared (λ = 8–13 µm) detection has long been the mitigation of dark current while achieving a high conversion efficiency of optical to electrical signals. Often overlooked is the bandwidth of detector response, despite several existing and expected future long-wave infrared high bandwidth applications. Here, we demonstrate ultra-fast response times in long-wave infrared detectors leveraging ultra-thin absorbers. The time response of the detectors is characterized using mid-infrared femtosecond pulses generated by an optical parametric amplifier, as a function of the device temperature and operating bias, as well as excitation wavelength. An equivalent circuit model for the detectors is presented and compared to our experimental results with excellent agreement. We demonstrate detector impulse response times of <100ps and 3 dB bandwidths in the GHz frequency range (f3dB > 3.5 GHz). Spectral response measurements confirm that the detectors have a resonant cavity mode enhanced response in the long-wave infrared, peaking at 10.2 µm. The presented detectors offer a potential solution for a range of high-frequency applications in the long-wave infrared.

The design and development of mid-infrared (mid-IR, 2–30 µm), and particularly long-wave infrared (LWIR, 8–13 µm), photodetectors typically involve a delicate balancing act between detector responsivity, which measures the electrical output for a given photon input, and detector noise, which is usually directly related to the detector dark current.1 Optimizing both simultaneously is quite difficult, as detector designs that increase the external quantum efficiency (EQE) very often result in larger dark currents, and efforts to minimize dark currents often have deleterious effects on detector quantum efficiency. The metric most closely identified with this balancing act is the specific detectivity (D*), which is the ratio of area-normalized responsivity to the noise spectral density (which is correlated with the dark current). D* allows for the comparison of detector architectures and materials across the mid-IR wavelength range. While D* is generally considered the standard for measuring the mid-IR detector performance, there are other metrics of growing importance to mid-IR applications and device development. Perhaps chief among these is the detector bandwidth.

Ultra-fast detectors have long been a fundamental component of the telecom infrastructure, and such detectors can reach bandwidths in the hundreds of GHz, at wavelengths where the fundamental processes responsible for dark current are far less punitive than in the mid-IR.2–4 For many mid-IR detector applications, bandwidth is far from the primary concern. At these longer wavelengths, detector bandwidths in the tens of kHz are often sufficient for many IR imaging systems and single-element sensors used in spectroscopy systems. These bandwidths, while challenging for bolometric detectors, are easily achievable in traditional epitaxial semiconductor-based LWIR photodetector device architectures. However, a growing number of IR applications could benefit greatly from detector bandwidths well above this level. Mid-IR free space communication and ranging,5–12 which both benefit from the reduced scattering and absorption in the mid-wave infrared (MWIR 3–5 µm) and LWIR atmospheric transmission windows, are two such examples. Novel sensing techniques, such as dual-comb spectroscopy (DCS)13–16 or cavity ringdown spectroscopy (CRDS),17 also require high-speed detectors capable of operating with bandwidths encompassing the microwave beat frequency (for DSC) or time constants much faster than the cavity ringdown time (for CRDS). With sufficiently fast detectors, there could even be opportunities to directly measure rotational and vibrational molecular dynamics.18 

The fastest LWIR detectors are the class of detectors leveraging intersubband transitions in III–V semiconductor heterostructures, such as quantum well infrared photodetectors (QWIPs)19,20 or quantum cascade detectors (QCDs).21–24 High bandwidth operation has also been shown in interband cascade infrared photodetectors (ICIPs).25–28 These detectors are the present detectors of choice for dual-comb spectroscopy and other high-speed applications but have quantum efficiencies so low that even with the minimal dark current in these detectors, D* is still orders of magnitude below that of competing materials and architectures.29–31 The II–VI ternary alloy HgCdTe (MCT) has been the most widely used material for mid-IR detection for many decades and offers an alloy-tunable narrow bandgap material with the potential to mimic the high-speed designs of short wavelength detectors. However, MCT growth is challenging due to material nonuniformity,32 and concerns about constituent toxicity33 have added urgency to the long-standing search for a III–V replacement material system for MCT.

Emerging from this search are III–V superlattice materials, comprising alternating nano-scale layers of lattice-matched or strain-balanced alloys with type-II (broken or staggered) band offsets.34 In such type-II superlattices (T2SLs), the superlattice minibands, controlled by the choice of material and layer thickness, can provide an effective bandgap lower than that of either of the constituent alloys.35 Such materials have been shown to reduce Auger scattering rates36 and offer a path forward for efficient LWIR detection at room temperature, or at least thermo-electrically cooled operation.37 Traditional detector designs effective for high-speed detectors at shorter wavelengths, such as reverse-biased photodiodes, are highly problematic in narrow bandgap systems, a result of Shockley–Read–Hall (SRH) processes and interband tunneling across the reverse-biased narrow bandgap absorber. Moreover, surface leakage currents endemic to narrow bandgap materials become increasingly problematic at smaller pixel dimensions. For this reason, the bariode (barrier diode) architectures,38–40 in which light is absorbed in a T2SL absorber region and collected across a wide bandgap barrier, have gained favor over traditional photodiode or photoconductor designs. The most typical implementation of such a structure is the nBn architecture,38 where photogenerated holes in an n-type absorber (n) transit across a wide bandgap barrier layer (B) and are collected in an n-type contact layer (n). The barrier layer blocks the flow of majority carriers (electrons), leading to a smaller dark current. A schematic of the nBn structure used in this work is shown in Fig. 1(a). Alternative versions of the nBn, leveraging the same advantages, fall into the general class of bariode devices, where valence/conduction band barriers limit the transport of majority carriers and dramatically reduce the detectors’ dark current. Such bariode architectures provide reduced SRH recombination/generation because much of the applied bias is dropped over the wide bandgap barrier instead of the lightly doped narrow bandgap layers. In addition, surface leakage is also reduced, thanks to the barrier layer acting to block parasitic surface conduction.

FIG. 1.

(a) Ultra-thin detector schematics. (i) The effective band structure of the detector considering the band edge positions and energy gaps of the T2SL minibands in the contact, barrier, absorber, and n++ ground plane, as well as a schematic of the electron–hole pair excitation and collection. (ii) Schematic of the LWIR T2SL band structure showing the valence (blue) and conduction (red) band edges of both constituent materials, as well as the lowest energy valence (blue) and conduction (red) band minibands. (b) Schematic of the as-grown layer stack, layer materials, and thicknesses, for the ultra-thin detector. (c) Normalized T = 77 K photoluminescence (black) and room temperature reflectivity (red) from the as-grown material.

FIG. 1.

(a) Ultra-thin detector schematics. (i) The effective band structure of the detector considering the band edge positions and energy gaps of the T2SL minibands in the contact, barrier, absorber, and n++ ground plane, as well as a schematic of the electron–hole pair excitation and collection. (ii) Schematic of the LWIR T2SL band structure showing the valence (blue) and conduction (red) band edges of both constituent materials, as well as the lowest energy valence (blue) and conduction (red) band minibands. (b) Schematic of the as-grown layer stack, layer materials, and thicknesses, for the ultra-thin detector. (c) Normalized T = 77 K photoluminescence (black) and room temperature reflectivity (red) from the as-grown material.

Close modal

Despite the benefits of bariode architectures for the electrical performance of T2SL detectors, the reduced electron–hole wavefunction overlap in the T2SL absorber results in weaker absorption coefficients and lower vertical mobility than would be achieved in bulk semiconductors of the same bandgap.41–44 Typical LWIR detectors based on T2SL absorbers have thicknesses of approximately the free-space wavelength to maximize absorption. This not only increases the dark current (detector noise) but also the collection time for photoexcited charge carriers, which must now diffuse across the thicker absorber region in order to be collected across the barrier. Thus, reducing the thickness of a LWIR bariode detector should provide a significant improvement in device bandwidth as well as a reduction in the dark current. However, such a reduction typically would come at the price of a marked decrease in the detector EQE.

There have been several demonstrations of ultra-thin mid-IR detector architectures, where a reduction in the absorber volume offers a dramatic decrease in the dark current while maintaining EQE via resonant optical responses. In the MWIR, approaches include resonant-cavity-enhanced infrared detectors,45 metal surface-wave-enhanced detectors,46 and guided mode resonance detectors.47,48 In the LWIR, we recently demonstrated ultra-thin nBn detectors that leverage a hybrid leaky-cavity/plasmonic mode for resonant absorption in absorber layers with thicknesses well below the free-space wavelength (∼λo/30).37,49 Such detectors show a significant improvement in dark current, with dark currents below the Rule 07 heuristic50 long used as the measure of the state-of-the-art dark current in MCT detectors (though recent MCT results have led to the Law 19,51 resetting the bar for mid-IR detector dark currents). While the driving force for the ultra-thin detector efforts has been to demonstrate higher D* and higher operating temperatures, the ultra-thin detectors should also show dramatic improvements in detector bandwidth due to the significant reduction in collection time associated with the ultra-thin absorber region.

Here, we demonstrate ultra-fast response times in ultra-thin LWIR detector structures. We leverage the basic hybrid plasmonic detector architecture in Ref. 49 but integrate the material into a device geometry designed to maximize microwave signal readout performance. The impulse response of the detectors is characterized by measuring the electrical signal in response to femtosecond mid-IR optical pulses. Response times under 100 ps are measured from the ultra-thin detectors. A circuit model of the detector structure is developed, and the simulated detector response is compared with the experimental results. The detector response is characterized as a function of temperature, applied bias, and excitation wavelength. We extract detector bandwidths in the GHz frequencies and analyze the potential of these detectors for high-speed mid-IR applications.

The detector structure used in this work is grown by molecular beam epitaxy (MBE) in a Varian Gen-II system with effusion sources for gallium, indium, aluminum, and silicon and with valved cracker sources for arsenic and antimony. The structure, from the surface to the substrate, consists of a 55 nm n-doped contact layer, an 80 nm barrier layer, a 372 nm n-doped absorber layer, and a 750 nm highly n-doped (n++) ground plane. The contact and absorber layers use the same superlattice design (10 nm InAs/2.2 nm InAs0.49Sb0.51), targeting an effective bandgap in the LWIR. The heavily doped n++ ground plane uses a wider bandgap (MWIR) superlattice design (5 nm InAs/1.1 nm InAs0.49Sb0.51) and is heavily doped to achieve a plasma wavelength of λp = 6 μm. The barrier layer uses an AlAs0.1Sb0.9 ternary alloy, slightly p-doped for the valence band alignment. A schematic of the as-grown layer structure is shown in Fig. 1(b). The design is based on the ultra-thin detector platform developed in previous work,37,49 where the highly doped ground plane supports a tightly bound hybrid leaky-cavity/surface plasmon polariton (SPP) mode in the absorber region. The thicknesses of the top contact, barrier, and absorber layers are chosen so that a strong optical enhancement is achieved at the target wavelength of 10 µm while maintaining the ultra-thin device dimensions required for low dark current. This strongly confined optical mode allows for a strong enhancement of absorption in an ultra-thin absorber layer with a thickness only a fraction of the operating wavelength.

Previous work on these ultra-thin detector architectures focused on the demonstration of the achievable optical enhancement and dark current reduction and used large mesa (>400μm×400μm) devices with metallic patch antennas designed to couple light into the SPP modes at the interface between the heavily doped (n++) ground plane and the nBn detector stack.37,49 For the purpose of demonstrating high-speed operation, we redesign the device to have top and bottom contacts forming a coplanar waveguide for signal readout and circular mesa diameters ranging from 20 to 100 µm, as the high-speed time response of photodetectors is often closely associated with their geometry. The patch antennas are omitted in this study in order to simplify the fabrication process. Enhancement is still achieved via the leaky-cavity mode, but coupling into the SPP is not possible without the antenna array. This study is focused solely on the time response characteristics of this detector architecture, and the slightly weakened optical response should not have a significant effect on the temporal response of the devices under test.

The fabrication process for the devices investigated in this work is schematically shown in Fig. 2(a). First, the mesa patterns are defined using a positive photoresist that doubles as the etch mask. A citric and phosphoric acid wet etch is used to etch down 750 nm, into the highly n-doped (n++) layer. Without removing the photoresist etch mask, a 750 nm layer of SiO2 is deposited using plasma-enhanced chemical vapor deposition (PECVD) to backfill the etched region. The photoresist is then removed, exposing the top surface of the mesas. A second layer of 200 nm SiO2 is then deposited to planarize the dielectric layer. The purpose of the thick dielectric layer is to passivate the mesa sidewalls and to reduce the parasitic capacitance formed between the top metal contact and the n++ layer. The top and bottom contact windows are then opened using photolithography and reactive-ion etching of the corresponding SiO2 regions. Finally, the metal contacts (30 nm Ti/15 nm Pt/300 nm Au) are deposited via photolithography, electron-beam evaporation, and lift-off. Figure 2 shows a schematic of the fabrication flow and a representative micrograph of a fabricated device. The dimensions of the top and bottom contacts are chosen to target a 50 Ω characteristic impedance while remaining compatible with the pitch size of the ground–signal–ground (GSG) probe used to bias the device and collect the generated photocurrent.

FIG. 2.

(a) Ultra-thin detector fabrication process flow schematics. (i) As-grown layer stack. (ii) Mesa definition by UV photolithography and citric and phosphoric acid wet etch. (iii) SiO2 backfill and photoresist lift-off. (iv) Second, planarizing SiO2 deposition. (v) UV photolithography and reactive ion etch of SiO2 aperture. (vi) UV photolithography, metal deposition, and lift-off for top and bottom contacts. (b) Optical micrograph of the fabricated detector. (c) Temperature-dependent current density vs voltage of fabricated devices. At low temperatures, the noise in the measurement is caused by the low dark current approaching the instrument noise floor.

FIG. 2.

(a) Ultra-thin detector fabrication process flow schematics. (i) As-grown layer stack. (ii) Mesa definition by UV photolithography and citric and phosphoric acid wet etch. (iii) SiO2 backfill and photoresist lift-off. (iv) Second, planarizing SiO2 deposition. (v) UV photolithography and reactive ion etch of SiO2 aperture. (vi) UV photolithography, metal deposition, and lift-off for top and bottom contacts. (b) Optical micrograph of the fabricated detector. (c) Temperature-dependent current density vs voltage of fabricated devices. At low temperatures, the noise in the measurement is caused by the low dark current approaching the instrument noise floor.

Close modal

The sample under test is mounted on a closed-cycle cryogenic probe station (ARS, Inc.) with temperature control from 10 to 300 K. The devices are contacted with a single GSG probe, with the signal line contacting the detector’s top contact and the ground lines contacting the device’s bottom contact. The GSG probe provides the applied bias to the device and reads out the electrical signal response to incident light. The temperature-dependent measurements are taken from 77 K up to 150 K. The current–voltage (IV) characteristics of the devices are measured directly using a source meter. During I–V measurements, the probe station window is covered with a metal plate to reduce unwanted photocurrent from background radiation.

In the time response measurements, the device is connected to a bias tee on the combined AC+DC port. The operational bias is supplied by a separate low-noise source meter to the DC port of the bias tee, and the photocurrent signal is collected from the AC-only port of the bias tee using a digital sampling oscilloscope. The incident optical pulse for the time response measurements is generated by an optical parametric amplifier (OPA, Light Conversion ORPHEUS) pumped with a ytterbium solid-state 1030 nm laser (Light Conversion CARBIDE). The temporal width of the pump pulse is <200 fs and confirmed with a frequency-resolved optical gating (FROG) pulse measurement system. The beam is expanded and attenuated with a neutral density filter to reduce the incident pulse energy on the detector. At an operating wavelength of 10 µm, the OPA beam profile is measured to have a full width at half maximum (FWHM) of 4.7 mm, resulting in an estimated photon energy flux of ∼200 nJ/cm2. For the bias-dependent time response measurements, the OPA wavelength is fixed at 10 µm, and the device bias is swept from 0 to −0.5 V.

The spectral response of the detectors is tested in two ways. First, the wavelength tunability of the OPA allows for the measurement of the wavelength-dependent impulse response of the detectors. In these measurements, the operating bias is fixed at −0.3 V, while the OPA wavelength is swept from 5 to 16 µm. The spectral response is extracted as the peak amplitude of the measured wavelength-dependent time response signal, normalized to the beam intensity (in photons/unit area) incident upon the detector surface. However, the short pulse width of the mid-IR probe does result in significant spectral broadening, particularly at long wavelengths, so the spectral measurements should be considered as low resolution.

In addition, the spectral response can be measured using a Fourier-transform infrared (FTIR) spectrometer (Bruker v80V). In this experiment, the device under test uses the same detector material, fabrication process, and CPW geometry as the devices used to extract the detector time response. In order to maximize the absolute signal from the detector, the devices with the largest diameter (D = 100 µm) are mounted and wire-bonded to a printed circuit board (PCB). The PCB package is then mounted onto the cold finger of a cryostat and electrically connected to the cryostat feedthroughs. Light from the FTIR’s internal glo-bar, after passing through the FTIR interferometer, is focused onto the device. The cryostat feedthrough is connected to a low-noise current preamplifier that both applies the operational bias (set to −0.5 V) and converts the photocurrent signal from the detector into a voltage signal that is sent back to the FTIR. As the FTIR interferometer scans, the measured photocurrent provides an interferogram whose Fourier transform is the spectral response of the detector to the incident light. The obtained spectra are then normalized to the response from a reference detector with a flat spectral response. This last normalization step accounts for the spectral weighting of the source and FTIR optics and provides the accurate relative spectral response of the devices.

Figure 2(c) shows the temperature-dependent dark current density vs applied bias (J–V) of four detector devices with mesa diameters ranging from 20 to 100 µm. At any given temperature, the J–V plots of the four devices, with mesa diameters of 20, 40, 60, and 100 µm, all show very similar current densities, indicating a uniformity in material quality and device fabrication, as well as well-passivated sidewalls. The dark current density for all devices increases dramatically with temperature, characteristic of photodetectors with narrow bandgap absorbers. The J–V plots are also clearly asymmetric with applied bias polarity, as would be expected for minority carrier-dominated transport in nBn detectors with a top contact layer much thinner than the absorber layer. In addition, we note the monotonic increase in dark current density over the operational range of the detectors (∼−0.1 to −0.5 V), which we attribute to a tunneling current, likely a result of the long-wavelength cutoff of the narrow bandgap absorber T2SL (λcutoff ∼ 15 µm). Resonant LWIR detectors, using the same architecture as we employ here, are able to demonstrate a dramatic reduction in tunneling current (and thus near diffusion-limited operation) at similar operating biases with a minimal effect on device responsivity by carefully increasing the absorber bandgap.37 Overall, the dark current profiles and magnitudes are largely consistent with previous results in Ref. 49. The low dark current of the ultra-thin nBn detector enables detector operations at temperatures significantly above liquid-nitrogen temperature. In the context of the present work, the high operating temperature of these detectors allows for the temporal characterization of detector operation up to, but not necessarily limited to, T = 150 K.

Figure 3 shows the measured time response from representative devices excited by a λ0 = 10 µm femtosecond pulse. This measurement provides the combined system impulse response of the detector, contacts, probe, bias tee, and RF cable. Figure 3(a) shows the time response from detectors of varying mesa diameters at T = 77 K. The signal amplitude shows a strong dependence on device diameter, with the larger surface area devices showing much stronger signals. Although not shown here, the integral of the time response signal increases linearly with the device area for the 20–60 µm devices, as would be expected. For the 100 µm device, the integral of the response is smaller than expected given the device area, which is attributed to the larger device dimensions becoming comparable with the lateral diffusion length of minority charge carriers in the T2SL absorbers,52,53 resulting in reduced carrier collection. In Fig. 3(b), where the time response at T = 150 K is shown, integrating the time response becomes difficult due to the parasitic negative voltage in the time response, an effect that we discuss in Sec. III C. We also observe a clear dependence of the temporal characteristics of the device response on device size. As the mesa diameter reduces, the detector time response narrows, as measured by the full width at half maximum (FWHM) of the measured signal, indicating faster device operation. At T = 77 K [Fig. 3(a)], we measure FWHM τ = 516 ps for the D = 100 µm device and τ = 91 ps for the D = 20 µm device. A similar decrease in the impulse response FWHM with device diameter is seen at T = 150 K [Fig. 3(b)], with τ = 316 ps for the D = 100 µm device and τ = 87.5 ps for the D = 20 µm device. The qualitative relationship between the device lateral geometry and the temporal response is largely expected and typical of high-speed photodiode detectors across the optical spectrum, a result of the reduction in device capacitance, which leads to a lower device RC time constant. Quantitatively, the temporal response of the nBn detectors in this work can be distorted by high-frequency parasitic effects, which results in a more complicated dependence on device geometry as discussed later.

FIG. 3.

(a) T = 77 K and (b) T = 150 K impulse response of detectors with diameters D = 20, 40, 60, and 100 µm under an applied bias V = −0.5 V, pumped with λ0 = 10 µm. (c) T = 77 K and (d) T = 150 K impulse response of the smallest diameter detector (D = 20 µm) for applied biases of V = 0, −0.1, −0.25, and −0.5 V, pumped at λ0 = 10 µm. The insets of (c) and (d) show the bias-dependent peak amplitude of the temporal response for the D = 20 µm device.

FIG. 3.

(a) T = 77 K and (b) T = 150 K impulse response of detectors with diameters D = 20, 40, 60, and 100 µm under an applied bias V = −0.5 V, pumped with λ0 = 10 µm. (c) T = 77 K and (d) T = 150 K impulse response of the smallest diameter detector (D = 20 µm) for applied biases of V = 0, −0.1, −0.25, and −0.5 V, pumped at λ0 = 10 µm. The insets of (c) and (d) show the bias-dependent peak amplitude of the temporal response for the D = 20 µm device.

Close modal

Figures 3(c) and 3(d) show the temporal response of the D = 20 µm detector as a function of the applied bias at T = 77 K and T = 150 K, respectively. The insets show the amplitude of the time response signal for D = 20 µm at both temperatures, which increases as the reverse bias increases and plateaus before V = −0.5 V, showing the responsivity saturation expected in nBn detectors. The low turn-on and saturation voltages of the detectors indicate a good band alignment for carrier transport, particularly the valence band offset at the barrier, which can cause significant offsets in device turn-on if not properly aligned.

For a given device, increasing bias does not significantly alter the temporal decay characteristics. This bias-independent temporal response differs from conventional detector architectures, such as p–i–n photodiodes, where an increasing reverse bias increases the field dropped across the intrinsic region and results in a drastically faster carrier drift across the device junction. Such an effect is not observed in the nBn detectors in this work, across the operational biases tested. It is likely that while both drift and diffusion mechanisms are present, the drift component is limited by the field drop across the absorber region, which can be pinned by the tunneling current that starts to dominate the dark current at higher biases [as suggested by the J–V characteristics in Fig. 2(c)].

While the detector temporal response signal largely follows the monotonic decay expected in response to an optical impulse, we do see two interesting features that cannot be modeled with a simple, single decay time. First, we observe a feature at t ∼ 200 ps in the smallest device (D = 20 µm), which also appears as a weak shoulder in the larger diameter devices’ responses. This is attributed to unexpected impedance mismatches in the measurement system, causing signal reflections. Second, we note that at higher temperatures, the detector response again shows a fast decay, but the output signal drops below 0 V to become negative before recovering to equilibrium.

In order to better understand the time response of our detectors, we develop an equivalent circuit model of our detectors, including the intrinsic detector structures, the CPW contact, and the bias tee-oscilloscope measurement back-end. The diagram of the equivalent circuit model is shown in Fig. 4(a). The red dashed box denotes the nBn detector, which is represented by a pair of parallel RC impedances. This modeling of the nBn detector is based on (with the terminologies also adopted from) Ref. 54. Cdiff is the capacitance formed by the stored charge near the barrier interface, Rdiff is the resistance associated with the barrier–top contact layer junction, and Cj is the capacitance associated with the barrier and the potentially depleted absorber layer (the capacitance formed between space charge regions across the absorber region). In our model, an additional resistance Rj, associated with the resistance of the absorber layer, is added in parallel to the junction capacitance Cj. This resistance was assumed to be very large in Ref. 54. For the LWIR nBn detectors in this work, this resistance is temperature dependent and its effect can be non-negligible. Rj can be regarded as the dynamic resistance in the measured I–V characteristics of the devices and, thus, decreases with increasing temperature.

FIG. 4.

(a) Equivalent circuit model of the nBn detector and the external measurement circuit. (b) The simulated response fitting of the experimentally measured response of the D = 100 µm device at 77 and 150 K. Only Rj, the time constant τ, and the amplitude of the current source InBn are changed to represent the temperature change, which is sufficient to reproduce the observed negative time response signals.

FIG. 4.

(a) Equivalent circuit model of the nBn detector and the external measurement circuit. (b) The simulated response fitting of the experimentally measured response of the D = 100 µm device at 77 and 150 K. Only Rj, the time constant τ, and the amplitude of the current source InBn are changed to represent the temperature change, which is sufficient to reproduce the observed negative time response signals.

Close modal

In addition to the temporal response associated with the behavior of the detector’s epitaxial stack, the device contacts can also play a role in the readout of signals with high-frequency components. First, the contacts form a coplanar waveguide geometry that, although designed to be impedance-matched to 50 Ω, can still have a slight impedance mismatch. This, along with unexpected impedance mismatches between components and adapter interfaces further down the measurement path, is considered to be the reason for the t = 200 ps feature observed in the time response of the 20 µm diameter device, as well as in the larger diameter devices, though in larger devices, this feature is mostly obscured by the broader detector response (see Fig. 3). In addition, the top contact metal strip can both form a parasitic capacitance and have a mutual inductance with the underlying n++ ground plane. In Fig. 4(a), the blue dashed box includes the metal readout contacts, with Rcontact being the series resistance of the contacts and Lcontact being the parasitic inductance associated with the top contact strip. The transmission line segment represents the coplanar waveguide formed by the top and bottom contacts, and Ccontact is the parasitic capacitance formed by the top metal contact and the underlying n++ layer, separated by the backfilled SiO2 layer. Finally, the green dashed box includes the bias tee, formed by an inductor and a capacitor pair. Rload represents the oscilloscope that captures the AC-only signal. The cables are assumed to be impedance-matched to 50 Ω and lossless to simplify the simulations. Through time-domain circuit response simulations, of which the results are shown in Fig. 4(b), with the source set to be an exponentially decaying pulse current, it is determined that the parasitic components associated with the top metal contact (Lcontact and Ccontact) cause the response signal to swing to negative voltages at higher temperatures. At lower temperatures, this effect is effectively screened out by the increased junction resistance (Rj) of the detector.

While the time-domain signals provide an intuitive understanding of the detector response, it is the frequency response of the detector that largely determines the practical use of the detector for high-speed applications. The frequency response can be calculated from the Fourier transform of the system impulse response. In the data presented in Fig. 3, this impulse response includes the response from the entire detection system, including the detector, contacts, probe, bias tee, and cables, all of which we consider to be components of the detector packaging and are, therefore, included in the Fourier transform calculation. Figure 5 shows the computed Fourier spectra of the detector responses. Consistent with the time response signals, the frequency response extends to higher frequency ranges with smaller device diameters. Note that the 3-dB bandwidth is measured from the highest point of the frequency response spectrum, which, due to the existence of parasitic effects, can be at a non-zero frequency. Alternatively, a theoretical 3-dB bandwidth can be computed from the single exponential decay current source signal extracted from the circuit model, which excludes the parasitic effects. It is noted here that in the frequency response calculated from the measured time response, because of the potentially non-zero offset in the location of the 0-dB frequency, the extracted 3-dB bandwidth could be even slightly higher than the theoretical 3-dB bandwidth calculated from the simulated current source signal. We cautiously acknowledge that because detector packaging is inevitable in practical applications, the discussion of the frequency response is based on the calculation from the measured time response, which is a more realistic representation of the entire detector package’s performance.

FIG. 5.

[(a) and (b)] Fourier transform frequency spectra of the time response signals from the four devices with mesa sizes ranging from 100 to 20 µm, under an applied bias of −0.5 V, at (a) 77 K and (b) 150 K, respectively. [(c) and (d)] 3-dB bandwidth extracted from the respective frequency spectra of the four devices, plotted as a function of the applied bias, at (c) 77 K and (d) 150 K, respectively.

FIG. 5.

[(a) and (b)] Fourier transform frequency spectra of the time response signals from the four devices with mesa sizes ranging from 100 to 20 µm, under an applied bias of −0.5 V, at (a) 77 K and (b) 150 K, respectively. [(c) and (d)] 3-dB bandwidth extracted from the respective frequency spectra of the four devices, plotted as a function of the applied bias, at (c) 77 K and (d) 150 K, respectively.

Close modal

The bias dependence of the frequency response shows the primary difference between our nBn detectors and the p–i–n detectors traditionally used, particularly at shorter wavelengths, for high-speed applications. Typically, p–i–n detectors can achieve a significant increase in bandwidth as reverse bias increases55 due to enhanced carrier drift across the biased depletion region and thus reduced transit times. Because drift is not the primary carrier transport mechanism in nBn detectors (at typical operational voltages), the bias dependence of our frequency response can differ from that of p–i–n photodiodes.

At low temperatures, shown in Fig. 5(c), the frequency response shifts very slightly with the applied bias, with the extracted 3-dB bandwidth actually decreasing with increasing reverse bias. When examining the bias-dependent time response signals of Fig. 3(c), this decrease in bandwidth is found to be related to a rise in the tail of the detector time response as the reverse bias increases. At higher temperatures [Fig. 5(d)], the effect diminishes and the frequency response becomes effectively independent of bias. This voltage dependence of the bandwidth is attributed to the change in the diffusion capacitance (Cdiff) as a function of bias. Cdiff is related to the stored charge at the barrier–top contact layer interface. This junction formed by the barrier and the top contact layer is analogous to a forward-biased junction in a traditional N–P–N transistor56 when the entire device is reverse-biased. An increase in the total reverse bias across the device indicates an increase in the forward bias across the junction, which can lead to an increase in the charge accumulation and thus Cdiff,54,57 that can manifest as a decrease in the overall frequency response. At higher temperatures, because more carriers are thermally generated in the narrow bandgap LWIR T2SL layers, the contact and barriers do not bend significantly under an external bias, reducing the potential for the conditions required for significant and bias-dependent charge accumulation. Thus, the diffusion capacitance and the frequency response become effectively bias-independent, as the drift process is minimal in the LWIR nBn detectors under investigation here. The frequency response at higher temperatures also differs in shape from that at low temperatures. This is directly associated with the difference in the time response at these temperatures. As shown in Fig. 3(c), at higher temperatures, the time response swings to negative voltages, which is attributed to the more prominent effects of parasitic components at higher temperatures, as discussed in Sec. III C. When transformed into the frequency domain, the frequency response thus adopts a high-pass filter-like behavior on the low frequency side, as shown in Fig. 5(c).

The area dependence of the detector bandwidth is shown in Fig. 5. While an area dependence for the device bandwidth is expected, as the various device capacitances scale with the device area, the dependence observed in our devices does not follow a quadratic dependence with the inverse of the device diameter. Instead, the device bandwidth is nearly linear with the inverse of the diameter (the bandwidth–diameter product is largely constant). This kind of dependence on the device diameter suggests that the detector response is possibly limited by the transport of the carriers within the thin, lightly doped top contact layer, instead of the transport from the absorber layer across the barrier layer.

These results suggest that a faster bandwidth is achievable, before reaching the ultimate limit associated with the vertical carrier transport in our layer stack, at the expense of further reducing the device size, or, alternatively, designing contiguous metal contact patterns that extend across the entire device mesa. The latter could be achieved in a way that can enhance detector responsivity by coupling the incident light into confined modes in the absorber,58,59 while simultaneously improving the detector bandwidth. For the smallest diameter device tested, having a 20 µm diameter mesa, the extracted detector 3 dB bandwidth is 3.8 GHz at 150 K. This is one of the fastest values recorded for an interband absorption-based LWIR detector architecture, being competitive with many of the commercially available high-speed LWIR detectors and shrinking the gap for achievable bandwidth of LWIR photodetectors leveraging interband and intersubband transitions. The bandwidths of the detectors presented in this work are comparable to those of the directly modulated LWIR quantum cascade detectors implemented in Ref. 12. At the same time, the detectivity of the ultra-thin T2SL detectors49 can be as much as an order of magnitude higher than quantum-well or quantum cascade detectors20,23 operating at the same temperature, a significant advantage of the ultra-thin T2SL detector architecture for future high-speed applications.

Finally, Fig. 6 shows the spectral response of the detector, tested both with the OPA and with the FTIR. In the FTIR-obtained spectrum, a peak in the spectral response is seen at around 10.2 µm, consistent with the spectral location of the cavity enhancement effect, seen as the dip in the reflection spectrum in Fig. 1(c), as well as the spectral response of ultra-thin detectors fabricated from similar layer stacks.37,49 Although not shown here in the normalized spectra, the intensity of the spectral response is observed to drop as temperature increases, which is also seen in the time response measurements [Figs. 3(a) and 3(c) vs Figs. 3(b) and (d)]. This is a somewhat unexpected trend for nBn detectors and is not yet fully understood for the devices under test, although similar temperature trends have been observed in previous studies60–62 and have been attributed to factors such as Auger recombination61 or the use of the ternary AlAs0.1Sb0.9 barrier.63 While the decreased responsivity does not seem to have any impact on the time-domain characteristics of the detectors, it does ultimately limit the detector performance at the higher operating temperatures studied in this work. Further studies will look to identify the cause of the decrease in response and optimize growth/fabrication to mitigate this effect.

FIG. 6.

Normalized spectral response of the representative D = 100 µm device, measured both with the FTIR and with the OPA. The FTIR spectra are obtained at both 77 and 150 K and normalized to their respective LWIR response peaks. The OPA spectrum is the extracted peak intensity of the response signals as a function of the OPA wavelength, normalized to the beam profile and photon energy at each wavelength, at the highest operating temperature tested (150 K) under an applied bias of −0.3 V.

FIG. 6.

Normalized spectral response of the representative D = 100 µm device, measured both with the FTIR and with the OPA. The FTIR spectra are obtained at both 77 and 150 K and normalized to their respective LWIR response peaks. The OPA spectrum is the extracted peak intensity of the response signals as a function of the OPA wavelength, normalized to the beam profile and photon energy at each wavelength, at the highest operating temperature tested (150 K) under an applied bias of −0.3 V.

Close modal

In the OPA-obtained spectrum, the peak response is observed at a wavelength of ∼11 µm, with the response remaining strong well past the spectral peak, even at 16 µm. This relatively high response on the longer wavelength side of the detector responsivity peak is attributed to the dispersion of the incident OPA excitation pulses. As the LWIR pulses are generated via multiple stages of nonlinear processes, the time–bandwidth product of the light pulse is expected to degrade, which will lead to a broader spectral width, especially for longer wavelength pulses. Even using the time–bandwidth product of the OPA at shorter operating wavelengths and applying this to the λ = 16 µm pulses, we obtain a best-case broadening of Δλ ∼ 5 µm. Thus, the OPA wavelength-dependent response curve should be considered as a low-resolution, long-wavelength side broadened spectral response. It remains clear, however, that the ultra-fast time response measured using the OPA at 10 µm truly represents the high-speed capabilities of ultra-thin nBn detectors at the LWIR.

We demonstrate a high-speed LWIR photodetector based on an ultra-thin T2SL nBn architecture. The detector exhibits fast time response signals, and the Fourier analysis shows that the detector bandwidth is not dependent on the bias, likely a result of limited field drop across the absorber region in the narrow bandgap nBn detector. The nBn structure coupled with an underlying highly doped ground plane allows for electrical and optical enhancements over conventional thick detector structures, enabling the ultra-thin detectors with a peak response at 10.2 µm operating at elevated temperatures. The photodetector in this work with the smallest circular mesa of 20 µm diameter recorded a detector bandwidth of 3.8 GHz at 150 K. In this work, the intrinsic high bandwidth of the ultra-thin detector architecture is demonstrated. Future efforts could more comprehensively explore the trade-offs and/or improvements required to extend the operating temperature and bandwidth of these devices. For example, implementing a pBp structure could potentially increase the speed of the time response via the faster transport of minority electrons compared to minority holes.64,65 Instead of a ring contact, patterned top contacts could shorten the path of carrier collection and increase the optical coupling and resonant absorption. In addition, further reducing the absorber thickness can improve the transport time and dark current simultaneously, a parameter space that can be explored if the reduced quantum efficiency and responsivity can be compensated through careful tuning of the optical resonance. The strong potential for both electrical and optical engineering of the ultra-thin detector paves the way for high operating temperature, high-speed LWIR photodetectors for free-space communication, advanced spectroscopic techniques, and a host of other existing and emerging LWIR high-bandwidth applications.

The authors J.A., M.A., and S.D. acknowledge the support from the CLAWS Applied Research for the Advancement of Priorities program of the Office of the Secretary of Defense and AFOSR Lab Task under Grant No. 23RWCOR002. Y.W., A.M., L.N., and D.W. acknowledge the support from the Air Force Research Lab via the University of Dayton Research Institute (Award No. RSC20061). Y.W. and D.W. acknowledge the support from the National Science Foundation (Award No. ECCS 1926187). This work was partly done at the Texas Nanofabrication Facility supported by NSF under Grant No. NNCI-202522.

The authors have no conflicts to disclose.

Yinan Wang: Conceptualization (equal); Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (lead); Software (lead); Validation (lead); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Aaron Joseph Muhowski: Data curation (supporting); Resources (equal); Writing – review & editing (supporting). Leland J. Nordin: Conceptualization (supporting); Resources (equal); Writing – review & editing (supporting). Sukrith U. Dev: Funding acquisition (equal); Project administration (equal); Supervision (equal); Validation (equal); Writing – review & editing (supporting). Monica S. Allen: Funding acquisition (equal); Project administration (equal); Supervision (equal); Validation (equal); Writing – review & editing (supporting). Jeffery W. Allen: Investigation (equal); Project administration (equal); Supervision (equal); Validation (equal); Writing – review & editing (supporting). Daniel Wasserman: Conceptualization (equal); Project administration (equal); Supervision (lead); Validation (equal); Visualization (lead); Writing – original draft (equal); Writing – review & editing (lead).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
A.
Rogalski
,
Infrared Detectors
(
CRC Press
,
2000
).
2.
Y.-G.
Wey
,
K.
Giboney
,
J.
Bowers
,
M.
Rodwell
,
P.
Silvestre
,
P.
Thiagarajan
, and
G.
Robinson
, “
110-GHz GaInAs/InP double heterostructure p-i-n photodetectors
,”
J. Lightwave Technol.
13
,
1490
1499
(
1995
).
3.
K.
Kato
,
A.
Kozen
,
Y.
Muramoto
,
Y.
Itaya
,
T.
Nagatsuma
, and
M.
Yaita
, “
110-GHz, 50%-efficiency mushroom-mesa waveguide pin photodiode for a 1.55-μm wavelength
,”
IEEE Photonics Technol. Lett.
6
,
719
721
(
1994
).
4.
H.
Ito
,
T.
Furuta
,
S.
Kodama
, and
T.
Ishibashi
, “
InP/InGaAs uni-travelling-carrier photodiode with 310 GHz bandwidth
,”
Electron. Lett.
36
,
1809
(
2000
).
5.
R.
Martini
,
C.
Bethea
,
F.
Capasso
,
C.
Gmachl
,
R.
Paiella
,
E.
Whittaker
,
H.
Hwang
,
D.
Sivco
,
J.
Baillargeon
, and
A.
Cho
, “
Free-space optical transmission of multimedia satellite data streams using mid-infrared quantum cascade lasers
,”
Electron. Lett.
38
,
181
(
2002
).
6.
A.
Soibel
,
M.
Wright
,
W.
Farr
,
S.
Keo
,
C.
Hill
,
R.
Yang
, and
H.
Liu
, “
Free space optical communication utilizing mid-infrared interband cascade laser
,” in
Free-Space Laser Communication Technologies XXII
(
SPIE
,
2010
), Vol.
7587
, pp.
241
250
.
7.
Y.
Su
,
W.
Wang
,
X.
Hu
,
H.
Hu
,
X.
Huang
,
Y.
Wang
,
J.
Si
,
X.
Xie
,
B.
Han
,
H.
Feng
et al, “
10 Gbps DPSK transmission over free-space link in the mid-infrared
,”
Opt. Express
26
,
34515
34528
(
2018
).
8.
O.
Spitz
,
P.
Didier
,
L.
Durupt
,
D. A.
Díaz-Thomas
,
A. N.
Baranov
,
L.
Cerutti
, and
F.
Grillot
, “
Free-space communication with directly modulated mid-infrared quantum cascade devices
,”
IEEE J. Sel. Top. Quantum Electron.
28
,
1200109
(
2022
).
9.
H.
Dely
,
T.
Bonazzi
,
O.
Spitz
,
E.
Rodriguez
,
D.
Gacemi
,
Y.
Todorov
,
K.
Pantzas
,
G.
Beaudoin
,
I.
Sagnes
,
L.
Li
et al, “
10 Gbit s−1 free space data transmission at 9 μm wavelength with unipolar quantum optoelectronics
,”
Laser Photonics Rev.
16
,
2100414
(
2022
).
10.
P.
Didier
,
H.
Dely
,
T.
Bonazzi
,
O.
Spitz
,
E.
Awwad
,
É.
Rodriguez
,
A.
Vasanelli
,
C.
Sirtori
, and
F.
Grillot
, “
High-capacity free-space optical link in the midinfrared thermal atmospheric windows using unipolar quantum devices
,”
Adv. Photonics
4
,
056004
(
2022
).
11.
P.
Didier
,
H.
Knötig
,
O.
Spitz
,
L.
Cerutti
,
A.
Lardschneider
,
E.
Awwad
,
D.
Diaz-Thomas
,
A.
Baranov
,
R.
Weih
,
J.
Koeth
et al, “
Interband cascade technology for energy-efficient mid-infrared free-space communication
,”
Photonics Res.
11
,
582
590
(
2023
).
12.
M.
Joharifar
,
H.
Dely
,
X.
Pang
,
R.
Schatz
,
D.
Gacemi
,
T.
Salgals
,
A.
Udalcovs
,
Y.-T.
Sun
,
Y.
Fan
,
L.
Zhang
et al, “
High-speed 9.6-μm long-wave infrared free-space transmission with a directly-modulated QCL and a fully-passive QCD
,”
J. Lightwave Technol.
41
,
1087
1094
(
2023
).
13.
I.
Coddington
,
N.
Newbury
, and
W.
Swann
, “
Dual-comb spectroscopy
,”
Optica
3
,
414
426
(
2016
).
14.
G.
Villares
,
A.
Hugi
,
S.
Blaser
, and
J.
Faist
, “
Dual-comb spectroscopy based on quantum-cascade-laser frequency combs
,”
Nat. Commun.
5
,
5192
(
2014
).
15.
Y.
Wang
,
M. G.
Soskind
,
W.
Wang
, and
G.
Wysocki
, “
High-resolution multi-heterodyne spectroscopy based on Fabry–Perot quantum cascade lasers
,”
Appl. Phys. Lett.
104
,
031114
(
2014
).
16.
G.
Ycas
,
F. R.
Giorgetta
,
E.
Baumann
,
I.
Coddington
,
D.
Herman
,
S. A.
Diddams
, and
N. R.
Newbury
, “
High-coherence mid-infrared dual-comb spectroscopy spanning 2.6 to 5.2 μm
,”
Nat. Photonics
12
,
202
208
(
2018
).
17.
M. D.
Wheeler
,
S. M.
Newman
,
A. J.
Orr-Ewing
, and
M. N.
Ashfold
, “
Cavity ring-down spectroscopy
,”
J. Chem. Soc., Faraday Trans.
94
,
337
351
(
1998
).
18.
J. J.
Klaassen
,
S. L.
Coy
,
J. I.
Steinfeld
, and
B.
Abel
, “
Direct measurement of rotational and vibrational relaxation in methane overtone levels by time-resolved infrared double-resonance spectroscopy
,”
J. Chem. Phys.
101
,
10533
10547
(
1994
).
19.
B.
Levine
, “
Quantum-well infrared photodetectors
,”
J. Appl. Phys.
74
,
R1
R81
(
1993
).
20.
D.
Palaferri
,
Y.
Todorov
,
A.
Bigioli
,
A.
Mottaghizadeh
,
D.
Gacemi
,
A.
Calabrese
,
A.
Vasanelli
,
L.
Li
,
A. G.
Davies
,
E. H.
Linfield
et al, “
Room-temperature nine-μm-wavelength photodetectors and GHz-frequency heterodyne receivers
,”
Nature
556
,
85
88
(
2018
).
21.
D.
Hofstetter
,
M.
Graf
,
T.
Aellen
,
J.
Faist
,
L.
Hvozdara
, and
S.
Blaser
, “
23 GHz operation of a room temperature photovoltaic quantum cascade detector at 5.35 μm
,”
Appl. Phys. Lett.
89
,
061119
(
2006
).
22.
F. R.
Giorgetta
,
E.
Baumann
,
M.
Graf
,
Q.
Yang
,
C.
Manz
,
K.
Kohler
,
H. E.
Beere
,
D. A.
Ritchie
,
E.
Linfield
,
A. G.
Davies
et al, “
Quantum cascade detectors
,”
IEEE J. Quantum Electron.
45
,
1039
1052
(
2009
).
23.
Y.
Zhou
,
S.
Zhai
,
F.
Wang
,
J.
Liu
,
F.
Liu
,
S.
Liu
,
J.
Zhang
,
N.
Zhuo
,
L.
Wang
, and
Z.
Wang
, “
High-speed, room-temperature quantum cascade detectors at 4.3 μm
,”
AIP Adv.
6
,
035305
(
2016
).
24.
J.
Hillbrand
,
L.
Matthieu Krüger
,
S.
Dal Cin
,
H.
Knötig
,
J.
Heidrich
,
A.
Maxwell Andrews
,
G.
Strasser
,
U.
Keller
, and
B.
Schwarz
, “
High-speed quantum cascade detector characterized with a mid-infrared femtosecond oscillator
,”
Opt. Express
29
,
5774
5781
(
2021
).
25.
H.
Lotfi
,
L.
Li
,
L.
Lei
,
H.
Ye
,
S.
Shazzad Rassel
,
Y.
Jiang
,
R. Q.
Yang
,
T. D.
Mishima
,
M. B.
Santos
,
J. A.
Gupta
, and
M. B.
Johnson
, “
High-frequency operation of a mid-infrared interband cascade system at room temperature
,”
Appl. Phys. Lett.
108
,
201101
(
2016
).
26.
Y.
Chen
,
X.
Chai
,
Z.
Xie
,
Z.
Deng
,
N.
Zhang
,
Y.
Zhou
,
Z.
Xu
,
J.
Chen
, and
B.
Chen
, “
High-speed mid-infrared interband cascade photodetector based on InAs/GaAsSb type-II superlattice
,”
J. Lightwave Technol.
38
,
939
945
(
2020
).
27.
Z.
Xie
,
J.
Huang
,
X.
Chai
,
Z.
Deng
,
Y.
Chen
,
Q.
Lu
,
Z.
Xu
,
J.
Chen
,
Y.
Zhou
, and
B.
Chen
, “
High-speed mid-wave infrared interband cascade photodetector at room temperature
,”
Opt. Express
28
,
36915
36923
(
2020
).
28.
L. M.
Krüger
,
J.
Hillbrand
,
J.
Heidrich
,
M.
Beiser
,
R.
Weih
,
J.
Koeth
,
C. R.
Phillips
,
B.
Schwarz
,
G.
Strasser
, and
U.
Keller
, “
High-speed interband cascade infrared photodetectors: Photo-response saturation by a femtosecond oscillator
,”
Opt. Express
29
,
14087
14100
(
2021
).
29.
X.
Lu
,
J.
Vaillancourt
, and
M. J.
Meisner
, “
Temperature-dependent photoresponsivity and high-temperature (190 K) operation of a quantum dot infrared photodetector
,”
Appl. Phys. Lett.
91
,
051115
(
2007
).
30.
J. C.
Campbell
and
A.
Madhukar
, “
Quantum-dot infrared photodetectors
,”
Proc. IEEE
95
,
1815
1827
(
2007
).
31.
A. V.
Barve
,
T.
Rotter
,
Y.
Sharma
,
S. J.
Lee
,
S. K.
Noh
, and
S.
Krishna
, “
Systematic study of different transitions in high operating temperature quantum dots in a well photodetectors
,”
Appl. Phys. Lett.
97
,
061105
(
2010
).
32.
A.
Rogalski
,
J.
Antoszewski
, and
L.
Faraone
, “
Third-generation infrared photodetector arrays
,”
J. Appl. Phys.
105
,
091101
(
2009
).
33.
A.
Han
,
E.
Han
, and
E.
Han Hv
, “
Directive 2011/65/EU of the European parliament and of the council of 8 June 2011 on the restriction of the use of certain hazardous substances in electrical and electronic equipment
,” Harting: Espelkamp, Germany,
2011
.
34.
D. L.
Smith
and
C.
Mailhiot
, “
Proposal for strained type II superlattice infrared detectors
,”
J. Appl. Phys.
62
,
2545
2548
(
1987
).
35.
D. Z.
Ting
,
A.
Soibel
,
L.
Höglund
,
J.
Nguyen
,
C. J.
Hill
,
A.
Khoshakhlagh
, and
S. D.
Gunapala
, “
Type-II superlattice infrared detectors
,” in
Semiconductors and Semimetals
(
Elsevier
,
2011
), Vol.
84
, pp.
1
57
.
36.
C.
Grein
,
P.
Young
, and
H.
Ehrenreich
, “
Minority carrier lifetimes in ideal InGaSb/InAs superlattices
,”
Appl. Phys. Lett.
61
,
2905
2907
(
1992
).
37.
L.
Nordin
,
A. J.
Muhowski
, and
D.
Wasserman
, “
High operating temperature plasmonic infrared detectors
,”
Appl. Phys. Lett.
120
,
101103
(
2022
).
38.
S.
Maimon
and
G.
Wicks
, “
nBn detector, an infrared detector with reduced dark current and higher operating temperature
,”
Appl. Phys. Lett.
89
,
151109
(
2006
).
39.
D. Z.-Y.
Ting
,
C. J.
Hill
,
A.
Soibel
,
S. A.
Keo
,
J. M.
Mumolo
,
J.
Nguyen
, and
S. D.
Gunapala
, “
A high-performance long wavelength superlattice complementary barrier infrared detector
,”
Appl. Phys. Lett.
95
,
023508
(
2009
).
40.
A. D.
Hood
,
A. J.
Evans
,
A.
Ikhlassi
,
D. L.
Lee
, and
W. E.
Tennant
, “
LWIR strained-layer superlattice materials and devices at Teledyne imaging sensors
,”
J. Electron. Mater.
39
,
1001
1006
(
2010
).
41.
N.
Yoon
,
C. J.
Reyner
,
G.
Ariyawansa
,
J. M.
Duran
,
J. E.
Scheihing
,
J.
Mabon
, and
D.
Wasserman
, “
Modified electron beam induced current technique for in(Ga)As/InAsSb superlattice infrared detectors
,”
J. Appl. Phys.
122
,
074503
(
2017
).
42.
Z.
Taghipour
,
S.
Lee
,
S.
Myers
,
E.
Steenbergen
,
C.
Morath
,
V.
Cowan
,
S.
Mathews
,
G.
Balakrishnan
, and
S.
Krishna
, “
Temperature-dependent minority-carrier mobility in p-type InAs/GaSb type-II-superlattice photodetectors
,”
Phys. Rev. Appl.
11
,
024047
(
2019
).
43.
I.
Vurgaftman
,
G.
Belenky
,
Y.
Lin
,
D.
Donetsky
,
L.
Shterengas
,
G.
Kipshidze
,
W.
Sarney
, and
S.
Svensson
, “
Interband absorption strength in long-wave infrared type-II superlattices with small and large superlattice periods compared to bulk materials
,”
Appl. Phys. Lett.
108
,
222101
(
2016
).
44.
A.
Rogalski
,
P.
Martyniuk
, and
M.
Kopytko
, “
InAs/GaSb type-II superlattice infrared detectors: Future prospect
,”
Appl. Phys. Rev.
4
,
031304
(
2017
).
45.
V.
Letka
,
A.
Bainbridge
,
A. P.
Craig
,
F.
Al-Saymari
, and
A. R.
Marshall
, “
Resonant cavity-enhanced photodetector incorporating a type-II superlattice to extend MWIR sensitivity
,”
Opt. Express
27
,
23970
23980
(
2019
).
46.
J. A.
Nolde
,
M.
Kim
,
C. S.
Kim
,
E. M.
Jackson
,
C. T.
Ellis
,
J.
Abell
,
O. J.
Glembocki
,
C. L.
Canedy
,
J. G.
Tischler
,
I.
Vurgaftman
et al, “
Resonant quantum efficiency enhancement of midwave infrared nBn photodetectors using one-dimensional plasmonic gratings
,”
Appl. Phys. Lett.
106
,
261109
(
2015
).
47.
A.
Kamboj
,
L.
Nordin
,
P.
Petluru
,
A. J.
Muhowski
,
D. N.
Woolf
, and
D.
Wasserman
, “
All-epitaxial guided-mode resonance mid-wave infrared detectors
,”
Appl. Phys. Lett.
118
,
201102
(
2021
).
48.
A.
Kamboj
,
L.
Nordin
,
A. J.
Muhowski
,
D.
Woolf
, and
D.
Wasserman
, “
Room-temperature mid-wave infrared guided-mode resonance detectors
,”
IEEE Photonics Technol. Lett.
34
,
615
618
(
2022
).
49.
L.
Nordin
,
P.
Petluru
,
A.
Kamboj
,
A. J.
Muhowski
, and
D.
Wasserman
, “
Ultra-thin plasmonic detectors
,”
Optica
8
,
1545
1551
(
2021
).
50.
W.
Tennant
,
D.
Lee
,
M.
Zandian
,
E.
Piquette
, and
M.
Carmody
, “
MBE HgCdTe technology: A very general solution to IR detection, described by ‘rule 07,’ a very convenient heuristic
,”
J. Electron. Mater.
37
,
1406
1410
(
2008
).
51.
D.
Lee
,
P.
Dreiske
,
J.
Ellsworth
,
R.
Cottier
,
A.
Chen
,
S.
Tallaricao
,
A.
Yulius
,
M.
Carmody
,
E.
Piquette
,
M.
Zandian
et al, “
Law 19: The ultimate photodiode performance metric
,” in
Infrared Technology and Applications XLVI
(
SPIE
,
2020
), Vol.
11407
, pp.
93
105
.
52.
L. K.
Casias
,
C. P.
Morath
,
E. H.
Steenbergen
,
G. A.
Umana-Membreno
,
P. T.
Webster
,
J. V.
Logan
,
J. K.
Kim
,
G.
Balakrishnan
,
L.
Faraone
, and
S.
Krishna
, “
Vertical carrier transport in strain-balanced InAs/InAsSb type-II superlattice material
,”
Appl. Phys. Lett.
116
,
182109
(
2020
).
53.
V.
Arounassalame
,
M.
Bouschet
,
R.
Alchaar
,
R.
Ferreira
,
F.
Carosella
,
A.
Ramiandrasoa
,
J.
Perez
,
N.
Péré-Laperne
,
P.
Christol
, and
I.
Ribet-Mohamed
, “
Anisotropic transport investigation through different etching depths in InAs/InAsSb T2SL barrier midwave infrared detector
,”
Infrared Phys. Technol.
126
,
104315
(
2022
).
54.
D.
Chen
,
K.
Sun
,
Y.
Shen
,
A. H.
Jones
,
A. A.
Dadey
,
B.
Guo
,
J. A.
McArthur
,
S. R.
Bank
, and
J. C.
Campbell
, “
Frequency behavior of AlInAsSb nBn photodetectors and the development of an equivalent circuit model
,”
Opt. Express
30
,
25262
25276
(
2022
).
55.
W.
Read
, Jr.
, “
A proposed high-frequency, negative-resistance diode
,”
Bell Syst. Tech. J.
37
,
401
446
(
1958
).
56.
M.
Reine
,
J.
Schuster
,
B.
Pinkie
, and
E.
Bellotti
, “
Numerical simulation and analytical modeling of InAs nBn infrared detectors with p-type barriers
,”
J. Electron. Mater.
42
,
3015
3033
(
2013
).
57.
R. T.
Howe
and
C. G.
Sodini
,
Microelectronics: An Integrated Approach
(
Prentice Hall Upper
,
Saddle River, NJ
,
1997
).
58.
T. W.
Ebbesen
,
H. J.
Lezec
,
H.
Ghaemi
,
T.
Thio
, and
P. A.
Wolff
, “
Extraordinary optical transmission through sub-wavelength hole arrays
,”
Nature
391
,
667
669
(
1998
).
59.
S. J.
Lee
,
Z.
Ku
,
A.
Barve
,
J.
Montoya
,
W.-Y.
Jang
,
S.
Brueck
,
M.
Sundaram
,
A.
Reisinger
,
S.
Krishna
, and
S. K.
Noh
, “
A monolithically integrated plasmonic infrared quantum dot camera
,”
Nat. Commun.
2
,
286
(
2011
).
60.
H.
Kim
,
O.
Cellek
,
Z.-Y.
Lin
,
Z.-Y.
He
,
X.-H.
Zhao
,
S.
Liu
,
H.
Li
, and
Y.-H.
Zhang
, “
Long-wave infrared nBn photodetectors based on InAs/InAsSb type-II superlattices
,”
Appl. Phys. Lett.
101
,
161114
(
2012
).
61.
K.
Michalczewski
,
Ł.
Kubiszyn
,
P.
Martyniuk
,
C.
Wu
,
J.
Jureńczyk
,
K.
Grodecki
,
D.
Benyahia
,
A.
Rogalski
, and
J.
Piotrowski
, “
Demonstration of HOT LWIR T2SLs InAs/InAsSb photodetectors grown on GaAs substrate
,”
Infrared Phys. Technol.
95
,
222
226
(
2018
).
62.
E.
Delli
,
V.
Letka
,
P. D.
Hodgson
,
E.
Repiso
,
J. P.
Hayton
,
A. P.
Craig
,
Q.
Lu
,
R.
Beanland
,
A.
Krier
,
A. R.
Marshall
, and
P. J.
Carrington
, “
Mid-infrared InAs/InAsSb superlattice nBn photodetector monolithically integrated onto silicon
,”
ACS Photonics
6
,
538
544
(
2019
).
63.
S.
Myers
,
A.
Khoshakhlagh
,
J.
Mailfert
,
P.
Wanninkhof
,
E.
Plis
,
M.
Kutty
,
H.
Kim
,
N.
Gautam
,
B.
Klein
,
E.
Smith
et al, “
Performance of InAsSb-based infrared detectors with nBn design
,” in
Infrared Remote Sensing and Instrumentation XVIII
(
SPIE
,
2010
), Vol.
7808
, pp.
23
31
.
64.
D.
Zuo
,
P.
Qiao
,
D.
Wasserman
, and
S.
Lien Chuang
, “
Direct observation of minority carrier lifetime improvement in InAs/GaSb type-II superlattice photodiodes via interfacial layer control
,”
Appl. Phys. Lett.
102
,
141107
(
2013
).
65.
G.
Umana-Membreno
,
B.
Klein
,
H.
Kala
,
J.
Antoszewski
,
N.
Gautam
,
M.
Kutty
,
E.
Plis
,
S.
Krishna
, and
L.
Faraone
, “
Vertical minority carrier electron transport in p-type InAs/GaSb type-II superlattices
,”
Appl. Phys. Lett.
101
,
253515
(
2012
).