Infrared detectors are widely applied in communications，medical treatment，astronomical observation，infrared imaging，night vision，and guidance^［1-3］. With the development of society，the demands for infrared detectors have become diversified，which has continuously promoted the development of infrared detection technology. Infrared detectors have developed from the first generation to the third generation，and the detection mechanism is constantly improving and developing（as shown in Fig. 1）. At present，the new generation of infrared detection technology is developing in the direction of large area array，two-color and multi-color，digitization，high frame rate，hyperspectral and single-photon detection^［4］. The core competitiveness will focus on the concept of “SWaP”：smaller size，lower weight，and lower power consumption^［5-7］.

In order to suppress the influence of dark current，the infrared detector needs to work at or below the liquid nitrogen temperature，which greatly increases the weight and power consumption of the detectors and reduces the reliability^［8］. Therefore，increasing the operating temperature of infrared detectors is an important research direction. If the temperature required for normal operation of the detector is increased，not only can the purpose of “SWaP” be achieved，but also the startup speed of the device can be accelerated. It can also significantly increase the reliability of the detector. Considerable efforts have been made aiming to increase detector operating temperatures to realize high-operating temperature（HOT）detectors over the past decades^［9］. Among the new emerging strategies cascade infrared detectors^{［10，11］}，InAs/GaSb type-II superlattices（T2SLs）^［12］，and barrier structures^［13］ have considerable progress.

In 2006，Maimon and Wicks proposed the nBn structure device for the first time，and developed it on a short-wave infrared detector，using InAs material as the absorbing layer and AlAsSb as the barrier layer^［14］. Subsequently，the nBn structure was applied to InSb^［15-17］，InAsSb^［18-21］，T2SLs^［22-29］，and HgCdTe^［30-35］ detectors. Related technologies have been developed，and some products have been applied. The nBn detector is designed to block the flow of majority carriers through reasonable energy band structural design，while minority carriers can flow freely. Fig. 2（a）shows the schematic diagram of the energy band structure of nBn detector. It consists of an absorbing n-type layer，a large bandgap barrier layer，and a thin n-type contact layer. The thickness of the absorbing layer should be comparable to the absorption length of light in the device which is typically several microns，while the contact layer is for biasing the device. There is a special type of heterojunction between the barrier layer and the n-type layers：all the bandgap difference appears in the conduction band offset which blocks the flow of majority carrier electrons，and there is zero valence band offset which allows the free passage minority carrier holes^［23］. To suppress the dark current effectively，the barrier should be located near the minority carrier collector and away from the region of optical absorption^［36］. Placing the barrier in this location allows photogenerated holes to flow unimpeded to the contact while majority carrier dark currents，re-injected photo-current，and surface currents are all blocked. Besides，the same doping type in the barrier and active layers is the key to maintaining a low，diffusion-limited dark current^［37］.

Dark current is an important indicator that restricts the detection performance of IR detectors. Generally，the composition of dark current is very complex，which includes diffusion current，G-R current，tunneling current that includes direct tunneling current and trap-assisted tunneling current，etc. In addition，due to surface defects and non-ideal electrodes of the device，there are surface leakage currents^{［38，39］}. The magnitude of the G-R current is linearly related to the intrinsic carrier concentration，while the diffusion current is proportional to the square of the intrinsic carrier. As the operating temperature rises，the intrinsic carrier concentration of the material will also increase significantly，and the dark current of the device will transfer from the recombination limit to the diffusion limit.

Generally，the G-R current is derived from the thermal excitation and recombination of carriers that occur in the depletion region of the p-n junction. And the SRH current is the G-R current which caused by the defect centers in the depletion region. Impurities and defects in semiconductors will form certain energy levels in the forbidden band which can promote the excitation and recombination of carriers. The nBn detectors with a large bandgap barrier layer will increase the energy required for electrons to transition to the impurity and defect levels，thereby suppressing the generation of SRH current. Due to the restriction of the growth process，surface passivation process，sample shape，and other external factors，surface leakage will also be introduced，and it may even become the dominant mechanism of device dark current. The surface leakage current in nBn detectors is suppressed by its multi-carrier barrier^{［40，41］}. The spatial makeups of current components in nBn detectors are shown in Fig 2（b）. The diffusion current is formed by the diffusion and drift of carriers on both sides of the depletion region of the p-n junction under the action of an electric field. Auger recombination^［42］ is that when a carrier transitions from a high energy level to a low energy level or recombines through a bandgap，the released energy is transferred to another carrier and transitions to a higher energy level. Owing to Auger diffusion currents being generated in the same region as the photocurrents in nBn barrier detector，Auger currents share the same spatial makeup as the photocurrent^［43］. Fig. 2（c）is the current components of traditional p-n photodiode. The majority-carrier dark current，surface leakage currents，and SRH generation current can almost be eliminated because of the existence of the unipolar barrier layer.

The typical Arrhenius plot of the dark currents as a function of temperature in nBn detector and in a conventional diode is shown in Fig. 2（d）. Generally，the diffusion current varies as T³exp（-E_g0/kT），while the G-R current varies as T^3/2exp（-E_g0/2kT）which is dominant by the generation of electrons and holes by SRH traps in the depletion region.（E_g0 is the bandgap extrapolated to zero temperature，T is the temperature，and k is Boltzmann’s constant）. T_c is defined as the crossover temperature at which the diffusion and G-R currents are equal^［44］. Consequently，the dark current is then limited to the diffusion contribution. The slope of the lower portion of the Arrhenius plot for the standard photodiode is roughly half that of the upper portion. The properties of the nBn detector（blue line）show the extension of the high-temperature diffusion-limited region to temperatures below T_c. In the low-temperature region，nBn detector has a lower dark current than the conventional photodiode. Compared to the conventional photodiode，it should exhibit a higher signal-to-noise ratio at the same temperature and operate at a higher temperature with the same dark current. Moreover，the Teledyne Technologies published an empirical expression in 2007，called “Rule 07”，for estimation the performance of state-of-the-art HgCdTe photodiodes dark current versus normalized wavelength-temperature product（λ_cT）^［45］. Hereafter，Lee et al. suggested replacing Rule 07 by Law 19 corresponding exactly to the background limited curve for room temperature^［46］. The internal photodiode current may be several orders orf magnitude below Rule 07 versus given cut-off wavelength and operating temperature. Ideal nBn barrier detectors are expected to exhibit diffusion limited performance originating from Auger 1 generation matching Rule 07^［47］. To date，however，experimentally prepared barrier devices still show diffusion limited performance.

Compared to the conventional p-n photodiode，because of the near elimination of SRH and surface leakage currents，the nBn photodetector requires less cooling to operate optimally and has simpler processing requirements. Under the conditions of lattice matching and energy band matching，the nBn structure could be extended to different material systems which will be discussed in later chapters.

III-V materials can provide stronger chemical bonds，and the band edge is less dependent on the composition，hence it has higher chemical stability^［48］. Furthermore，considering the cost of detector materials，III-V materials have great advantages. The diameter of the antimonide material wafer reaches the order of 10 cm，so the detectors fabricated by III-V compounds with larger areas and higher uniformity^{［49，50］}. The Bandgap alignment of several III-V materials is shown in Fig. 3（a）. Among the III-V materials for fabricating barrier detectors，there is a material system consisting of the near lattice-matched InAs/GaSb/AlSb family of semiconductors—the 6.1 Å material family with room-temperature energy gaps ranging from 0.36 eV（InAs）to 1.61 eV（AlSb）^［51］. Fig. 3（b）shows the schematic diagram of the valence band offset as a function of lattice constant^［52］. This makes the 6.1 Å lattice-matched family the potential materials for designing the barrier detectors with the desirable unipolar band alignments by adjusting the elemental composition of the compound. Moreover，the bulk GaSb has a direct bandgap in the short-wave IR（SWIR），InAs in the MWIR，and InAs/Ga（In）Sb superlattices in the SWIR，MWIR，and LWIR^［53］. The 6.1 Å materials can be epitaxially grown on GaSb and GaAs substrates. In particular，4-inch diameter GaSb substrates became commercially available in 2009，offering improved economy of scale for fabrication of large format FPAs arrays^［48］. At present，the barrier detectors have been successfully applied to the III-V materials system and commercialized.

The InAsSb is an important candidate for fabricating MWIR detectors because of the high carrier mobility，relatively small dielectric constant，and self-diffusion coefficient at room temperature. Due to a nearly zero band valence offset with respect to AlAsSb in the valence band，InAsSb has emerged to play a dominant role in the designing of the nBn detectors. The InAsSb were grown on either GaAs（100）or GaSb（100）substrates and the n-type doping is usually reached by either Si or Te elements.

Klipstein and Weiss have described the detailed growth procedure and device’s characterization of InAs_1-xSb_x/AlAs_1-ySb_y nBn MWIR detector^{［20，44］}. In 2011，the Klipstein team achieved the InAsSb/AlAsSb nBn detector with a cutoff wavelength of ~4.1 μm and a dark current close to ~3×10^-7 A/cm² at 150 K^［54］，as shown in Fig. 4（a-b）. The device was fabricated nearly lattice-matched to the GaSb substrate，which exists a typical strain of 500 ppm or less. It discussed the cutoff wavelength increases with decreasing absorber layer doping because of the Moss-Burstein effect^［55］，which blocks interband optical transitions in the absorber layer with final states in the conduction band that lie below the Fermi energy. Also，the quantum efficiency（QE）of the detector is dependent on the thickness of the absorber layer. The FPA detectors with a 320×256 format and 30μm pitch were fabricated which have high operability and a NETD of 16 mK at 150 K，using f/3 optics and a 20 ms integration time^［44］. By the standard two-point nonuniformity correction procedure of SCD，excellent image quality has been achieved，as shown in Fig. 4（c）.

In 2012，D’Souza et al. have demonstrated the nBn detector in the InAsSb/AlAsSb materials system grown by molecular beam epitaxy（MBE）on GaSb and GaAs substrates^［56］. The device has a 5 μm cut-off operated at 200 K. The dark current density in the pyramidal structured diodes is reduced by a factor of 2-3，which is consistent with the volume reduction due to pyramid formation. Soibel described InAsSb/AlAsSb nBn detectors with cut-off wavelengths near 4.5 μm in 2014^［57］.

The nBn structure is considered theoretically by Martyniuk and Rogalski^{［58，59］}. The device structure and stimulated energy band diagram are shown in Fig. 4（d-e）. Fig. 4（f）shows the I-V characteristics as a function of temperature^［60］. The alloy composition x of the InAs_1-xSb_x absorber layer is 0.195 with a cutoff wavelength of ~4.9 μm at 150 K. And J_dark is 1.0×10^-3 A/cm² at 200 K and 3.0×10^-6 A/cm² at 150 K（Fig. 4（f））. The detectors are dominated by diffusion currents at -1 V bias at the QE peaks.

Subsequently，Klipstein et al. presented one of the first commercial nBn array detectors operated in the blue part of the MWIR window of the atmosphere（3.4~4.2 μm）and launched on the market by SCD in 2014 known as “Kinglet”（640×512，15 μm），which is the first III-V XBn detector developed by SCD to meet low SWaP applications^［50］. The working temperature of the focal plane array is 150 K，the power consumption is 3 W，and the weight is less than 300 g.

HOT Hercules（1280×1024，15 μm）is SCD’s second nBn MW HOT product. It was launched in 2014 and uses the InAsSb material system. The operating temperature of this million-pixels component can reach 150 K^{［61，62］}. In order to improve the performance of the detector，SCD has also developed the Blackbird series of MWIR digital detectors，the pixel pitch is reduced to 10 μm，with lower power consumption and higher photoelectric performance. The Blackbird series of nBn-InAsSb barrier detectors（1280×1024，10 μm）have the same photoelectric performance as planar junction InSb detectors，but the working temperature is up to 150 K^［63］.

InSb is the III-V semiconductor with the narrowest bandgap and the highest electron mobility. InSb IR detectors are intrinsic absorption in the MWIR（3~5 μm）spectral region，as well as extremely high quantum efficiency and responsivity. Therefore，InSb detectors can achieve extremely high thermal sensitivity and excellent image quality，which has become one of the most important MWIR detectors. However，the InSb device has to work at a low temperature of 77 K which greatly restricts its application.

Evirgen et al. first reported the experimental research progress of the InSb IR detector based on the nBn structure in 2014^［64］. The structure and energy band of the device are shown in Fig. 5（a-b）^{［64，65］}. The nBn structure was fabricated by MBE. While the Al-content of the barrier is around 20%，the values of dark current density are easily fitted by current density due to thermionic emission（J_TE）⁵，given by the following equation^［66］：J_TE=A*T²e^-Φ/kT，where A* is Richardson constant including the conduction band effective mass of the n-type contact layer and Φ is the energy difference between the top of the InAlSb barrier and the fermi energy in the contact layer（Fig. 5（c））^［64］. It shows that the total current of the device is mainly from the majority electrons of the n-doped InSb cap layer. In order to achieve a sufficiently high conduction band barrier to block electrons while maintaining a lower valence band barrier to allow the flow of photo-generated holes，the barrier layer of the device adopts a graded Al composition design. The dark current of InSb nBn detector reaches 1×10^-9 A/cm² under 77 K and -50 mV bias（Fig. 5（d）），which is 2×10^-5 A/cm² under 135 K and -50 mV bias（Fig. 5（e））^［65］.

As shown in Fig. 5（f），compared with the InSb PIN detector fabricated by MBE and standard planar process of SOFRADIR，the dark current of nBn device is significantly suppressed. The nBn detector can operate at a temperature of around 120 K，which is 20 K higher than the PIN device and 40 K higher than the standard planar process^［67］. Such a result demonstrates the potentiality of the application of the InSb nBn detector.

In addition，the DAT-CON company of Slovenia has developed MWIR FPA based on XBn-InSb detectors，including the CLRT series（640×512，15 μm）and CLRT HD series（1280×1024，10 μm）. The detective spectral range is 3.4~5.1 μm and NETD reaches 25 mK^［20］.

The type-II superlattice material has become the focus of the development of the third generation of infrared detectors for its inherent lower Auger G-R rate，large effective mass of electrons（small tunneling current between bands，small dark current），adjustable energy band，and good uniformity. From the perspective of the T2SLs structure，the electrons and holes are spatially separated and localized in self-consistent quantum wells formed on both sides of the heterointerface^［37］. The unique “broken bandgap” energy band structure enables T2SLs IR detectors to achieve higher device performance and operating temperature in MW/LW IR detection. In addition，the advantage of T2SLs for IR detectors is the property to tune the wavelength by adjusting one of the material components and fixing the other^［68］，which is especially crucial in the design of nBn detectors.

InAs/GaSb heterojunctions were found by Sakaki in 1977，which are generally considered to be a material that can replace the current mainstream HgCdTe with its unique characteristics^［69］. In the InAs/GaSb T2SLs system，the AlAsSb barrier layer is generally used to fabricate the nBn detector，owing to the InAs/GaSb T2SLs having an almost zero valence band offset relative to the AlAsSb barrier^［48］.

Rodriguez in the university of New Mexico applied the nBn structure to the T2SLs detector for the first time in 2007^［70］. In 2015，SCD presented the long-wave FPA “Pelican-D LW”（640×512，15 μm）using the pBp structure，which is the first LW FPA product of SCD^［71］. The absorber layer and electrode layer of the device are InAs/GaSb T2SLs，while the barrier layer is InAs/AlSb T2SLs. The alignment between mini-bands is shown in Fig. 6（a）. The operating temperature of the detector is 77 K，and the cut-off wavelength is 9.5 μm. The dark current is as low as 4.4×10^-5 A/cm² at 77 K.

Northwestern University reported a short-wave T2SLs photodetector based on the nBn structure in 2017^［72］. The device was fabricated on GaSb substrate based on InAs/AlSb/GaSb T2SLs，and AlAs_0.1Sb_0.9/GaSb H-structure SL design was used as the large bandgap electron barrier（Fig. 6（b））. The photodetector has a peak responsivity of 0.65 A/W at room temperature at 1.9 μm. At 300 K and 400 mV bias，the dark current density is 8×10^-3 A/cm² and the D* is 1.51×10¹⁰ cm∙Hz^1/2W^-1. At 150 K，the dark current density is 9.5×10¹⁰ A/cm²，the quantum efficiency is 50%，and the D* is 1.12×10¹³ cm∙Hz^1/2W^-1.

The T2SLs IR detectors have developed rapidly in the past 10 years. The focus of future research is on the device structure design and FPA technology. The barrier device structure is expected to further reduce the dark current of the device and enhance the performance.

In 1959，Lawson invented the direct bandgap ternary compound semiconductor material with continuously adjustable bandgap width—Hg_1-xCd_xTe^［1］. By changing the component x，the bandgap width meets the infrared detection of three atmospheric windows of 1~3 μm，3~5 μm，and 8~14 μm. Meanwhile，there is an absolutely small change of lattice constant with component x making it possible to grow high-quality heterostructures. The advantages of the high quantum efficiency of the HgCdTe material make it a key core IR device that is widely used in the optoelectronic system of military equipment^{［73，74］}. However，there are several drawbacks associated with HgCdTe photodetectors. Firstly，the weak ionic Hg-Te bonds and high Hg vapor pressure make the epitaxial growth is more challenging than typical III-V compounds. In addition，HgCdTe is easy to oxidize in the air for its unstable physical and chemical properties.

Considering the successful application of nBn barrier detectors for III-V materials，it is expected to be introduced into HgCdTe materials to overcome the processing technology limitations of typical p-n junctions. However，the implementation of nBn structure in the HgCdTe material is not ideal due to the existence of non-zero valence band offset at the absorber–barrier interface which enormously limits their performance^{［43，75］}. At low bias，the valence band barrier（∆E_v）inhibits the holes flow to the contact cap layer as a result the devices always exhibit lower sensitivity and detectivity. In order to align the valance band，a high operating bias—“turn-on” voltage，typically greater than the bandgap energy is required depending on the operating wavelength，which leads to strong band-to-band and trap-assisted tunneling due to high electric field within the depletion layer^［68］.

Itsuno et al^{［30，76，77］}. and Kopytko et al^［78］. introduced the nBn structure to HgCdTe for the first time and performed theoretical simulation calculations on the device. The band structure is shown in Fig. 7（a）. Compared with the p-n junction structure，the nBn structure presents tight dark current density，responsivity，and detectivity at a temperature from 50 K to 95 K and it shows that the HgCdTe unipolar barrier device based on the nBn structure is very promising for the MWIR and LWIR detection. Thereafter，the research group realized the fabrication of HgCdTe nBn devices. The material parameters are shown in Fig. 7（b）. The dark current is saturated with a reverse bias of approximately -0.8 V shown in Fig. 7（c）and a cut-off wavelength of approximately 5.7 μm is observed^［76］. Subsequently，the results of HgCdTe devices based on the nBn structure and optimization have been reported successively internationally.

In 2020，Voitsekhovskii et al. presented an MWIR HgCdTe nBn detector grown by MBE on GaAs substrate^［80］. To further suppress the surface current，there is passivation of the Al₂O₃ layer grown by plasma-enhanced atomic layer deposition（PE-ALD）. The dark current density of the detector is 3.1×10^-6 A/cm² at 184 K and has a bias of -1 V.

At present，although the application prospects of nBn HgCdTe photodetectors have been extensively explored，their research is mostly at the theoretical stage. The application of nBn architecture to HgCdTe presents a serious challenge due to the difficulty of achieving an ideal nBn band structure which has a zero valence band. The existence of a valence band offset seriously limits the device performance^［81］. Present strategy in the development of HgCdTe nBn detectors concentrates on decreasing or even removing the ∆E_v in the barrier layer，which will make it have lower operating bias，lower dark current，and operate at higher temperatures. There are several ways being proposed to eliminate valence band offset，such as composition optimization and doping modulation technique of barrier layer^［82-84］ and superlattice barrier layer^{［85，86］}. The appropriate p-type doping of the barrier layer can reduce ∆E_v and increase the ∆E_c. According to previous reports，by applying the barrier layer based on the Type III CdTe/HgTe superlattice structure replacing the single-layer barrier structure，the ideal energy band structure can be obtained by adjusting the thickness and doping of the CdTe and HgTe layers^［85］. To a certain extent，it is worth noting that barrier detectors are the most promising approach to achieve HgCdTe detectors operating at close to room temperature. Table 1 is the nBn infrared detector products that have been launched.

Considering the strict requirements of lattice and band matching of designing nBn barrier detectors with conventional materials is extremely challenging. Two-dimensional materials are innovated to fabricate nBn detectors for their self-passivated surfaces，tunable band structures，and avoidable lattice mismatch and interface defects. The ultra-thin and atomic-level flat ideal interface of 2D materials provides the possibility for designing high-performance heterojunction devices. Recently，Hu et al. have presented a nBn unipolar barrier photodetectors based on a tungsten disulfide/hexagonal boron nitride/palladium diselenide（WS₂/h-BN/PdSe₂）heterostructure in which WS₂ is used as a visible-wavelength n-type photon absorber，h-BN as the barrier and PdSe₂ as the contract layer^［94］. The structure of the device is shown in Fig. 8（a）. 2D materials are prepared by the mechanical peeling method，and through non-destructive fixed-point transfer，a heterojunction without lattice mismatch and thermal mismatch is obtained.

The band structure achieved from simulation analysis using Sentaurus-TCAD is illustrated in Fig. 8（b）with drain biases ranging from 0 V to -0.5 V. It can be seen from the simulation results that there is an obvious electronic barrier in the conduction band. From the output characteristic curves under 520 nm laser illumination depicted in Fig. 8（c），the nBn photodetector exhibits a dark current of the order of pA，a photocurrent of 20 μA. Under the room temperature at 520 nm，the detectivity of the device reaches up to 2.7×10¹² cm∙Hz^1/2W^-1.

The barrier photodetectors based on 2D material are currently in the initial research stage，and there are no other relevant literature reports. The above-mentioned research results will create a foundation for subsequent research. Therefore，the barrier photodetectors fabricated by 2D material could have good prospects in IR photoelectric detection.

The efforts in IR detection technology particularly aim to increase the operating temperature reaching the HOT demand^［7］. A relatively novel nBn structure based on a unipolar barrier has materials and thus enhances the operating temperature because it is not limited by G-R current and tunneling currents. Research based on various material systems mentioned above shows that the barrier structure has great potential.

Significant effort has been devoted to developing nBn barrier detectors based on the III-V materials. Up to now，nBn barrier detectors based on III-V materials（InAsSb，InAs/GaSb T2SLs，et al.）have been successfully industrialized，and the performance of the devices has also been greatly improved. Among these，III-V T2SLs materials，such as InAs/GaSb and InAs/InAsSb，are considered to be the most promising competitor to HgCdTe IR detectors. However，T2SLs IR detectors suffer from a major limitation related to the short SRH carrier lifetime（<100 ns experimentally）. Although the intensive research effort has been devoted over the last decade，little improvement has been acquired. Compare with the state-of-the-art HgCdTe detectors，its performance advantage is not significant mainly due to the lower quantum efficiency and shorter minority carrier lifetime limited by SRH GR mechanisms. However，it appears that T2SLs are especially potential in the design of barrier detectors because of the ability to tune the wavelength by adjusting one of the material components and fixing the other. What’s more，the high quality，high uniformity of fabricated FPAs of T2SLs material system is very suitable for today's application needs and in an early stage of development which reached rapid progress in recent years，it will have a great promise in the design of nBn barrier detectors for the future IR solutions.

In recent years，new low-dimensional materials are progressing rapidly in material growth，device fabrication，and characterization. At the same time，the production of 2D-materials heterojunctions can avoid the limitation of lattice matching of the traditional heterojunction interface，thereby realizing more optoelectronic devices with new functions，high sensitivity，and room temperature IR detection. Therefore，combing the traditional bulk materials with excellent photoelectric properties with low-dimensional materials，it is effective to optimize the heterojunction interface and avoid the complex fabrication process and large interface damage of traditional bulk materials.

The nBn barrier structure based on energy band engineering has continuously made breakthroughs in low dark current and high operating temperature. This new IR detector structure with strategically important technology is expected to lead to the further development of a new generation of infrared detection technology.