α/G figure of merit for infrared photodetector materials Open Access

Various materials have been used in the history of infrared (IR) detector development. Figure 1 shows the approximate dates of their introduction in the active areas of infrared detectors. After World War II, modern IR detector technology began, supported by the discovery of the transistor in 1947 by W. Shockley, J. Bardeen, and W. Brattain.¹ Almost 80 years of technological progress led to the development of high-performance IR detectors. Infrared systems technology supported by new knowledge of semiconductor materials, advances in integrated circuit photolithography, and the Cold War impetus for military readiness has resulted in unprecedented advances in infrared imaging in a short period of time. On the other hand, in the last 30 years, a new family of compounds, commonly referred to as low-dimensional solids (LDS), has emerged to compete with the existing/standard commercial photodetectors. These include two-dimensional (2D) materials and nanowires (NWs)/quantum dots (QDs), perovskites, and organic materials—see the lower part of Fig. 1. Because these new groups of materials exhibit extraordinary photoelectric properties, they have been considered potentially promising for a new generation of high-performance infrared detectors. Type-II superlattice interband quantum cascade photodetectors (T2SL IB QCIPs) have been omitted from consideration in the paper. Although the performance of QCIPs is similar to that of HgCdTe photodiodes at room temperature, their future relevance is questionable due to their technological challenges and high production costs.² 

The parameters of an IR system are designed according to the application specification, which mainly depends on the environmental conditions in which the system will operate. This requires the IR system to meet a number of criteria, with detector selection being the most important. It turns out that different design solutions are used depending on the application requirements. The most important criteria for detector selection are spectral range, sensitivity, and operating temperature. Currently, the most popular infrared detector technologies are HgCdTe, InSb/III–Vs, Si:As BIB, and microbolometers, as is shown in Fig. 2.

For detectors operating in the near-infrared (NIR) range (1.0–1.7 μm), the most important material is InGaAs lattice-matched to InP. HgCdTe ternary alloys are mainly used in photodiodes covering the wavelength range from 2 to 20 μm. Type-II superlattices (T2SLs—both InAs/GaSb and InAs/InAsSb) have become alternatives to HgCdTe. Silicon blocked-impurity band (BIB) detectors (doped with Sb, As, and Ga) with a cutoff wavelength sensitivity limit in the range up to 30 μm operate at a temperature of 10 K. Ge:Ga photoconducting detectors are the best photon detectors operating in low-background conditions in the wavelength range from 40 to 120 μm at a very low temperature, about 2 K.

The introduction of microbolometer arrays into the global market in the 1980s was a milestone in the development of room-temperature IR cameras. However, microbolometers are thermal detectors with limited response times (usually in the millisecond range) and do not fulfill the conditions for multispectral detection. In this respect, photon detectors have a clear advantage, which on the other hand require cooling, usually cryogenic cooling. These requirements significantly limit the widespread use of photon infrared technology. For this reason, for several decades now, intensive efforts have been made to increase the operating temperature of this class of detectors to reduce the size (S) of IR systems, their weight (W) and power (P) consumption (abbreviated as SWaP). In the literature, the name of this class of photodetectors is adopted as HOT (high operating temperature) detectors. One of these research directions is also the search for a new class of photodetectors, including the aforementioned LDS photodetectors. These compact devices can be integrated into a wide range of applications, from wearable sensors for personal health monitoring to unmanned aerial vehicles for aerial inspections and surveillance.

Although many published articles indicate an optimistic future application of the new generation of IR photodetectors, this paper tempers this optimism. In this study, we used the α/G ratio as a criterion for comparing the performance of both standard and new-generation infrared photodetectors. Most of the considerations are limited to room temperature photodetectors because of the fact that the performance of the new generation of detectors is mostly specified for operation at this temperature. The evaluation of the performance of IR detectors based on the fundamental physical properties of materials confirms that it will still be difficult to achieve better potential performance of the new generation of photodetectors in the future compared to those presented by HgCdTe photodiodes.

As mentioned earlier, various material systems are used in the manufacture of IR photodetectors. In general, in standard detectors, the electron transitions in device structures are divided into interband and intraband transitions depending on whether the energy levels lie in the same energy band (conduction band or valence band). Interband transitions are characterized by radiative, Auger, and SRH g-r processes, where generally carrier lifetimes are longer with better device performance. Intraband transitions, on the other hand, are accompanied by phonon contribution, resulting in shorter carrier lifetimes and poorer performance.

Figure 3 shows the tunable bandgaps of different materials through line combinations with direct bandgaps (solid lines) and indirect bandgaps (dashed lines). Binary materials with slight lattice mismatches usually provide more efficient alloying when combined into ternary materials, i.e., GaAs and AlAs, GaAs with InAs, InAs with GaSb, and HgTe with CdTe. In the case of alloyed compounds, the cutoff wavelength limit can be smoothly adjusted from the very long-wavelength infrared (VLWIR) to the visible region by adjusting the appropriate stoichiometry of the compound.

Table I collects other relevant parameters of semiconductor materials at room temperature, which is important from the point of view of the photodetector design.

This section explains the basic differences in the photoelectric properties of the materials used in the new generation of photodetectors compared to the standard ones. Particular attention was paid to the influence of basic material properties (especially absorption coefficients, carrier lifetime, and carrier mobility) on the performance of photodetectors. The absorption coefficient was estimated for the threshold energy of 1.2 × E_g, where E_g is the semiconductor bandgap. Much of the information presented here is taken from previously published articles.^5–10 

The discovery of graphene (Gr) and derived two-dimensional (2D) materials has led to the development of next-generation optoelectronic devices, creating a new platform for various photonic device applications,^11,12 including photodetectors.^13,14

Currently, 2D compounds are promising materials for high-performance photodetectors due to their unusual properties:

• the ease of tuning the electron states of materials, with typical thickness less than 10 nm, by external fields (e.g., ferroelectric field, gate-induced electrostatic field, and localized photogating field);

• the wide variation of the energy bandgaps from 0 eV (graphene) to 6 eV (hexagonal boron nitride, h-BN) allows the production of photodetectors over a wide spectral range from the ultraviolet to the far infrared (even terahertz);

• in addition, the energy gap can be tuned by changing the number of monolayers, the stress, or the alloy stoichiometry; increasing the thickness of the material reduces the energy gap;

• in the case of some transition metal dichalcogenides (TMDs) (such as MoS₂, MoSe₂, WS₂, and WSe₂), monolayers are characterized by a direct energy gap, while bulk materials are indirect bandgap semiconductors.

Table II shows the electronic properties of the most important 2D materials used in the manufacture of optoelectronic devices.

Black phosphorus (bP) is characterized by an orthorhombic structure with two special directions: armchair and zigzag along the x and y axes, respectively. The highly anisotropic arrangement of phosphorus atoms in bP results in anisotropic optoelectronic properties. It turns out that the effective mass of carriers in the zigzag direction is about 10 times larger than in the armchair direction.

The strong in-plane anisotropy also affects the carrier mobility—for holes along the light effective mass direction as high as 1000 and about 500 cm²/V s along the heavy effective mass direction. The conductivity of bP is significant in samples with thicknesses from 2 to 5 μm. It is worth noting that bP always maintains a direct bandgap with thickness change. It should be noted that the energy gap takes on values between 0.2 and 2.0 eV.

A key issue in the properties of 2D materials is their stability under ambient conditions. Exfoliated bP flakes are highly hygroscopic, tending to take moisture from the air. Low stability and chemical degradation limit the use of bP in optoelectronic devices.

Layered TMDs are atomically materials with the chemical formula MX₂, where M is a transition metal atom (e.g., W, Mo, and Re) and X is a chalcogen atom (e.g., S, Se, or Te). One layer of M atoms is located between two layers of X atoms. There are three polytypes of 2D TMDs: trigonal—1T, hexagonal—2H, and rhombohedral—3R. They are characterized by different electronic properties—from metallic to semiconducting and even superconducting. In contrast, for graphene, for which optoelectronic properties are determined by s- and p-hybridization, the optoelectronic properties of TMDs are determined by the number of d electrons, i.e., the filling of d orbitals of transition metals and their coordination environment. The number of d electrons in a transition metal varies from 0 to 6 for group 4 to group 10 TMDs, respectively.

Bulk 2D TMD materials have an indirect energy gap typically around 1 eV. Monolayer TMDs, on the other hand, are semiconductors characterized by higher energy gaps. As the material thickness decreases from bulk to monolayer, the TMD band structure changes from a smaller indirect energy gap to a larger direct energy gap. This suggests that TMDs can be sensors of electromagnetic radiation in a wide spectral range by tuning the energy gap with the number of monolayers due to the quantum size effect. In addition, due to the layered crystal structure, the optical and electronic properties of these materials can be strongly modified by lattice deformations/strains. With respect to graphene, alternative 2D materials [such as molybdenum disulfide (MoS₂), tungsten disulfide (WS₂), and molybdenum diselenide (MoSe₂)] show higher absorption and cover a very wide part of the radiation spectrum from infrared to visible. The layer-number dependent energy gaps of noble TMDs such as PtSe₂ and PdSe₂ are characterized by a gradual transition from semiconductor (monolayer) to semimetal (bulk). These experimental observations are in agreement with theoretical simulations based on the density functional theory.

It turns out that the absorption range of 2D TMDs can be extended to the mid-IR spectral region by introducing defect- or edge-states inside the bandgap, which is facilitated by the relatively high edge-to-surface ratio. The absorption coefficient of 2D compounds is typically of the order of 10⁴–10⁶ cm⁻¹ (see Fig. 4), which implies that more than 95% of the sunlight is absorbed by the sub-micrometer-thick TMD layers. This high optical absorption is explained by dipole transitions between localized d states and excitonic coupling of such transitions.

As the number of monolayers increases, the carrier mobility in TMDs increases. In general, however, their mobility is low (usually less than 250 cm²/V s) and this phenomenon is difficult to overcome. For this reason, 2D materials do not show a clear advantage over conventional 3D bulk materials such as III–V compounds and HgCdTe ternary alloys. Similar to graphene, the carrier mobility of TMDs is limited by ripples, phonon scattering, impurity scattering, and interface scattering. Figure 5 shows experimentally measured room temperature carrier mobilities of typical group-6 TMD compounds compared to different noble TMD layers (PtSe₂, PtS₂, and PdSe₂) and bP on back-gate SiO₂ substrates. The carrier density in 2D materials depends on the doping levels and recombination centers and its typical value is 10¹² cm⁻². The range of mobility changes for different materials is highlighted from Refs. 5–10 and for organic materials from Ref. 16.

Carrier lifetime has been considered to be a very critical parameter in semiconducting materials since it affects the detector's response speed and sensitivity. In general, for 2D materials, the lifetime of carriers decreases as the power of incident radiation increases. For this reason, reliable lifetime measurements are conducted under low-power conditions of incident radiation. Figure 6 shows their typical values.^7–10 As we can see, the lifetime of carriers in TMDs does not exceed values of the order of nanoseconds, while in graphene it takes picosecond values.¹⁵ The range of carrier lifetime changes is highlighted from Refs. 6–10 and for organic materials and perovskites from Refs. 16 and 17, respectively. With the current state of technology, published parameter values (mobilities and carrier lifetimes) for quantum dots, organic materials, perovskites, and LDS solids differ by several orders of magnitude—as shown in Figs. 5 and 6.

Figure 6 also shows the carrier lifetime as a function of the cutoff wavelength for selected standard materials at 300 K. HgCdTe is characterized by the most favorable recombination mechanisms that determine the long carrier lifetime, which allows the photodetectors to operate in HOT conditions. The carrier lifetime of lightly doped HgCdTe (∼10¹³ cm⁻³), conditioned by the Shockley–Read–Hall (SRH) process, allows us to achieve ∼10 ms (MWIR) and ∼0.5 ms (LWIR), respectively. Similar values can be assumed for detectors operating at 300 K.⁴ It should be emphasized, however, that the high intrinsic concentrations ∼6 × 10¹⁵ cm⁻³ (MWIR) and 5 × 10¹⁶ cm⁻³ (LWIR) at 300 K limit the carrier lifetimes through the Auger processes—Auger 1 (n-type) and Auger 7 (p-type). The lifetimes of Auger 1 and Auger 7 carriers in HgCdTe are consistent with the dependence τ_Ai7 ≈ 6τ_Ai1.⁴ 

As we well know, traps have a decisive influence on the carrier lifetime determined by the SRH mechanism. III–V materials are characterized by a higher SRH trap density than II–VI ones. For this reason, the lifetimes τ_SRH of the latter are higher than for III–V compounds. II–VI compounds exhibit stronger ionic bonds than their III–V counterparts, which makes the electron wave function around the lattice sites more compact and less prone to the formation of band states due to crystal imperfections. The carrier lifetime of InAs/GaSb T2SLs (conditioned by SRH) is several tens of nanoseconds (for both MWIR and LWIR materials) and is attributed to the occurrence of Ga, while the “Ga-free” InAs/InAsSb T2SLs exhibit a longer lifetime, up to several microseconds in the MWIR.

Among the diverse types of perovskite materials, metal halide perovskites (MHPs) are emerging as a rising star in the field of optoelectronics. This subgroup of the perovskite family is characterized by a nearly cubic structure with a crystal unit cell given by the ABX₃ formula, where cation A can stabilize octahedra [BX₆]⁴⁻, formed with cation B and halide X (Cl⁻, Br⁻, and I⁻). The rapid progress in perovskite device technology is due to the relatively simple manufacturing process, low production cost, and high efficiency. Considerable progress is being observed in the construction of efficient light emission devices, especially those operating in the visible and near-infrared regions.

Perovskites are divided into three types based on their dimensions: three-dimensional (3D), two-dimensional (2D), and zero-dimensional (0D) perovskites, also known as quantum dots. The energy gap of perovskites depends on the selection of A-site cations with different ionic radii in the octahedral space. This causes lattice deformations (contraction or expansion) due to the tilt of the inorganic octahedral space and changes the bond length and B–X bond angle. Unlike conventional semiconductor materials, in which lattice defects create trap states located in the energy gap [between the edges of the conduction band (CB) and valence band (VB)], orbitals in perovskites are located inside or near the edges of the VB and CB bands, which makes them highly tolerant to defects.¹⁸ In addition, these defects do not act as trap states, so their impact on the electronic and optical properties of the devices is negligible. This feature is particularly advantageous in flexible devices susceptible to mechanical deformations. A critical problem associated with lead perovskite with organic cations is the thermal and structural instability¹⁹ that requires further research for a quick solution.

The perovskite absorber combines distinct features of the standard photovoltaics technologies to include: low weight and flexibility (GaAs, CdTe, and CIGS inorganic thin films), high power conversion energy (Si crystalline photovoltaic cells), scalable fabrication in low-temperature solutions, and wavelength tuning (by using organic thin films, dye-sensitized and quantum dot-based cells).²⁰  Figure 7 summarizes the most important optoelectronic properties (absorption coefficient and electronic properties) of perovskite solar cells in comparison with other thin-film photovoltaic cells.

Table III summarizes the unique properties of perovskite materials.²¹ The high efficiency of perovskite photovoltaic devices is attributed to the electron mobility (200 cm²/V s) together with the diffusion length (>1 μm). Their threshold absorption coefficients are at the level of 10⁵ cm⁻¹, which is determined by the s–p antibonding coupling. Another important property of perovskites is the low exciton binding energy <10 meV, which allows the excited carriers to migrate as free carriers.²² 

Single crystal perovskite materials are characterized by a greater advantage of crystalline states and exhibit better optical and electrical properties than polycrystalline films and microcrystals. This is due to structural advantages such as free grain boundaries, an ordered long-range crystal structure and high orientation resulting in greater stability.

For lead-iodide perovskites presently used for solar cells, such as MAPbI₃ and FAPbI₃, both hole and electron mobilities were found to be primarily restricted to the level of ∼200 cm²/V s being lower than for GaAs. Despite this, the metal halide perovskites exhibit high charge extraction, and the net carrier diffusion length also depends on the recombination lifetimes (typically below 1 μs—see Fig. 6) being higher than standard GaAs.²³ 

The development of QD photodetectors began in the 1990s. Since then, we have been observing an evolution of their development from devices based on self-assembled epitaxial dots, to a new generation of devices based on colloidal quantum dots (CQDs). The development of the latter has been observed in the last decade,²⁴ in which the active region consists of three-dimensional semiconductor nanoparticles.

IR quantum dots can be grown from different groups of semiconductors: group IV (Si, Ge, GeSn), IV–VI (PbS, PbSe, PbTe), III–V (InAs, InSb), II–VI (HgTe, HgSe), I–VI (Ag₂S, Ag₂Se), and triple I–III–VI (CuInS₂, CuInSe₂) and their alloys.²⁵ So far, most of the research has focused on quantum dots based on PbS and Hg-based CQDs. However, due to the toxicity of devices containing mercury quantum dots, alternative materials are sought. One of them is metal halide perovskite nanocrystals.²⁶  Table IV summarizes the advantages and disadvantages of CQD devices.

CQDs are usually synthesized chemically from the liquid phase using inexpensive reagents. Many CQD growth procedures have been adapted from the experience of VIS display companies to scale up the synthesis of dots with lower unit costs. The most commonly used technique for IR CQD production is the synthesis in the colloidal solution phase. The main problem of this process is the control of particle nucleation at various stages of growth in a solution containing both metal and anion precursors.

Due to hopping-like transport in CQDs, the carrier mobility, typically below 1 cm²/V s (Fig. 5), is much lower in comparison with that observed for standard IR materials. Carrier lifetimes are also low—below 0.1 μs (Fig. 6).

Initially, the main research efforts were focused on CQD photodetectors operating in the NIR and MWIR ranges. Despite the simple device architecture, their performance was limited by the dark current and 1/f noise. A major technological challenge was the control of the doping concentration. Current research is directed toward CQD photovoltaic devices with the aim of achieving lower dark current densities and lower 1/f noise.

The experimentally measured current responsivity of CQD photodiodes is in the range of 100 mA/W, corresponding to an external quantum efficiency (EQE) of ∼10%–50%. It is higher in the shorter wavelength range (SWIR) (see Fig. 8). As can be seen, the EQEs are much higher than those achieved for epitaxial (self-assembled) QD photodetectors—typically reported ∼2%. For comparison purposes, Fig. 8 also shows typically higher QEs for commercially available detectors based on InGaAs, InSb, HgCdTe, and T2SLs.

Organic semiconductors (OSCs) have been of great interest for more than 40 years, exhibiting the potential to transform significant technologies, such as photovoltaic energy, transparent screens/displays, efficient and reasonably priced white lighting, or flexible and robust electronics. However, most efforts have been directed on lighting/display development. The unique properties also make them suitable for photodetectors with spectral responsivity ranging from ultraviolet (UV) to NIR with panchromatic or selective tuning of specific wavelengths.

Compared to inorganic semiconductors (ISCs), OSCs offer several distinct advantages:

• in general, they are low-cost materials, and their manufacturing methods allow for low-temperature growth, low-cost surface scaling (compatible with high-performance roll-to-roll processing);

• majority of inorganic materials require high quality substrates, while in contrast due to high lattice mismatch tolerance and deformation-induced defect states, organic devices are fabricated on the plastic films, metal foils, or glass;

• organic optoelectronic materials possess the ability to adjust their optical bandgap through chemical modifications of their molecular structure without the need for filters, thereby simplifying the spectral selectivity of detectors; and

• the structural characteristics of organic molecules inherently provide advantages over inorganic materials in the development of large-area, flexible, wearable photodetectors making them well-suited for diverse and evolving future needs.

One of the fundamental disadvantages is the low carriers’ mobility related to the weak intermolecular interactions reducing performance compared to inorganic devices. In addition, majority of organic materials are found not to be very stable being susceptible to degradation by water vapor and oxygen exposure requiring special housing to reach the satisfactory device durability. Another problem is related to the organic materials purity—much lower than inorganic materials, with the consequent creation of electronic defects that reduce device performance.

OSCs may be categorized as single small molecules, oligomers (few monomer units), and polymers (many monomer units). The OSCs are characterized by the relatively low energy (≈10 kcal/mol) van der Waals (vdW) intermolecular bonds compared to, e.g., Si–Si (78 kcal/mol) covalent bonds. This means that the propensity to form ordered structures is mild; however, from the OSCs fabrication point of view, this is advantageous.

The physical properties of OSCs differ significantly from their ISC counterparts. In the former, the strong electron-phonon coupling along with the lack of long-range ordering results in the localization of charge carriers on a single molecule or few adjacent molecules. The consequence of this is relatively inefficient transport of charge carriers by tunneling among molecules (hopping). The carrier mobility is low, normally within 10⁻⁴–10⁻¹ cm²/V s range, and is much lower than those in perovskites (see Fig. 5). The mobilities in the range of tens of cm²/V s were reported in the latest generation of high-performance polymers (only for high carrier densities ∼10¹⁹ cm⁻³).²⁷ Furthermore, the molecules of OSC depend on vdW force interactions, and the delocalization of their charge carriers takes place within a confined range, leading to discrete and relatively narrow energy bands. On the other hand, however, such carrier localization makes these OSCs more tolerant to impurities than ISC counterparts, making the deposition techniques simpler and reducing processing costs. The strong exciton/charge localization and low dielectric constant of OSC contribute to high exciton binding energies (above 100 meV) compared to perovskite materials (below 20 meV). Carrier lifetimes in OSCs are in the range of 0.1 to 1 microseconds and are longer than in perovskites (see Fig. 6).

The OSCs are characterized by a high absorption coefficient. Typical α assumes 10⁴–10⁵ cm⁻¹, leading to absorption lengths (∼1/α) ∼100 nm. Figure 9 shows the absorption coefficients for selected photovoltaic materials. In contrast to ISCs, OSCs exhibit a high absorption coefficient only in a reduced spectral region (typical width of about ∼0.5 eV) which is presented in Fig. 9(a) for non-fullerene acceptors FBR and IDTBR, the donor polymer PTB7-Th,^29,30 and their blends. The absorption spectrum is asymmetric—the rising edge (in the lower energy region) is steeper than the absorption decay (in the higher energy region). The continuous lines [Fig. 9(b)] represent the exponential fit of the absorption edge to the Urbach energy. This characterization method is used to quantify the energetic disorder at the band edges of a semiconductor.

Typically, the absorption coefficient strongly influences the current responsivity, which determines the photodetection window. As shown in Fig. 10, GaAs and InGaAs photodiodes have the highest quantum efficiency close to 100%, while PbS CQD photodiodes have the lowest—about 30%.

Based on the fundamental photoelectric properties of various materials collected in Sec. III, we will estimate $α τ$ values for these materials for three important infrared spectral ranges: NIR (λ_c = 1.6 μm), MWIR (λ_c = 5 μm), and LWIR (λ_c = 10 μm).

Table V presents the estimated $α τ$ figure of merit for sets of material systems considered in Sec. III. These estimated values should be regarded as indicative due to the large discrepancies in published parameters, especially for LDS materials. Only the material data for HgCdTe and T2SLs are dependable. The absorption coefficients for HgCdTe and T2SLs were estimated for the threshold energy of 1.2 × E_g, where E_g is the semiconductor bandgap (see Fig. 4).

As can be seen from the table, the $α τ$ values for the standard NIR material systems (InGaAs, HgCdTe) are higher than those for LDS ones. This is mainly due to the much longer carrier lifetime in standard materials. Specifically, the carrier lifetime in HgCdTe with low doping concentration (below 5 × 10¹³ cm⁻³) is much longer (in the millisecond range) compared to their level in superlattice materials with doping 5 × 10¹⁴ cm⁻³ (in the microsecond range). In the case of InAs/GaSb SLs, the carrier lifetime is typically several tens of nanoseconds.

Different configurations of TMD photodetectors with performance comparable to HgCdTe photodiodes at room temperature are published in the available literature.^6,7,9 The most important conclusions regarding the absorption coefficients of TMDs and HgCdTe are as follows:

• the absorption coefficients of HgCdTe (E_g ≈ 0.1–0.3 eV) are below 10⁴ cm⁻¹ and for TMDs (E_g > 1 eV) above 10⁵ cm⁻¹ and

• for a hypothetical TMD with an energy gap of 0.1–0.2 eV, the threshold absorption coefficient can be expected to be below 10⁵ cm⁻¹.

However, the carrier lifetime in TMDs is below 10⁻⁸ s. Based on these data, it is reasonable to believe that the $α τ$ value is significantly higher for HgCdTe compared to TMDs.

It can be seen from the above discussion that, taking the α/G ratio as the figure of merit for materials used in photodetectors operating over a wide range of the infrared spectrum, standard materials such as InGaAs, HgCdTe, and T2SLs appear to be superior to TMDs and other LDS photodetectors. However, the differences in $α τ$ values for LDS NIR materials are negligible.

Figure 11 shows the experimentally measured highest spectral detectivities for photodetectors operating at 300 K in the wavelength range of 0.1–20 μm. Two basic performance limits are also marked: the signal fluctuation limit (SFL) and the background fluctuation limit known as BLIP (background-limited infrared photodetector). In general, the performance of experimentally measured LDS photodetectors matches well with standard photodetectors. However, some of them are overestimated, as indicated in magenta. The reasons for these overestimations have already been described in several articles.^31–34 The most important of these are as follows:

• incorrect noise assessment,

• incorrect determination of the radiation power density and active region of the device,

• noise and responsivity as opposed to bandwidth, and

• non-linearity in light response and electrical conductivity.

The first publication by Lawson et al. in 1959³⁵ triggered an unprecedented development of HgCdTe over the past six decades. Since 2007, the Rule-07 metric has been widely used in the IR community as a basis for predicting HgCdTe photodiode performance.³⁶ It is also a reference benchmark for alternative technologies such as III–V barrier detectors,^37,38 T2SL devices,^38,39 quantum dot photodetectors,^40,41 and two-dimensional (2D) material photodetectors.⁴² 

From the point of view of the operation of HOT LWIR HgCdTe photodetectors (including those operating at room temperature), the optimum detector configuration is a reverse-biased P–i–N photodiode.^5,43–46 For sufficiently low doping and long SRH carrier lifetimes, the dark current of the P–i–N HgCdTe photodiode is suppressed, and its performance is limited by the background radiation and estimated by Law 19.^45,46 The benchmark introduced by Rule 19 forces a new perspective on the ultimate performance of HgCdTe photodiodes operating particularly in the LWIR spectral range. As shown in Fig. 11, the potential performance of HOT HgCdTe photodiodes operating at longer wavelengths (above 4 μm) indicates that it is possible to achieve an order of magnitude higher detectivity (above 10¹⁰ Jones) compared to the value predicted by Rule 07³⁶ and Rule 22.⁴⁷ This detectivity is limited by the background at 2π FOV. In this context, it will be difficult for 2D photodetectors to compete with HgCdTe photodiodes. In calculations of the “Depletion” curve (Law 19) shown in Fig. 11, it was assumed a low doping concentration in the i-region (10¹³ cm⁻³) and τ_SRH = 1 ms.⁴⁶ The last value was approximated according to Kinch's monograph.⁴ This monograph presents the general theory of SRH recombination and estimates of SRH lifetimes indirectly deduced from the I–V characteristics of HgCdTe photodiodes (Table 5.2 on page 106).

We are currently witnessing tremendous technological progress in the development of a new generation of infrared photodetectors in which the active detector region is a low-dimensional solid. Among these, we can distinguish: two-dimensional (2D) materials, perovskites, nanowires/quantum dots, and organic materials. To date, a large number of articles (including review papers) have been published, many of which envisage the replacement of the currently dominant HgCdTe ternary alloy by new-generation materials.

In the present work, based on the α/G criterion, we have estimated the suitability of different material groups as active areas of IR photodetectors. The estimates carried out do not indicate better fundamental properties of the new generation of materials compared to the HgCdTe alloy system. In this context, it is rather difficult to compete LDS photodetectors with HgCdTe photodiodes.

The unique electronic and optical properties of 2D materials show them to be a promising alternative to standard infrared detectors. Even the highest/record detectivity for 300 K in the LWIR range has been reported in the literature (but, as it turned out, overestimated³²). Various techniques have been used to improve the sensitivity of 2D photodetectors, including electron trap layers and the photogate effect with a graphene fast transfer channel in FET detectors, but it has been found that the insufficiently high carrier mobilities and response times limit practical applications (a trade-off between sensitivity and response time is necessary). Practical applications of 2D materials require excellent synthesis and processing. To reach their predicted potential, these photodetectors still have a long way to go.

Perovskite and organic materials are generally used for visible photodetectors. Their long wavelength cutoff reaches the NIR region (≈2 μm) but with performance inferior to InGaAs photodiodes.

Generally, the CQD performance is lower than standard photodiodes except PbS CQD-based devices exhibiting detectivity comparable with InGaAs. Their big advantage is the low production cost of the FPAs.

The potential properties of HOT HgCdTe photodiodes operating in the long-wavelength infrared spectral range (above 4 μm) indicate that BLIP operating conditions can be achieved at room temperature. To lend credence to the above predictions, it is worth noting that fully depleted 640 × 512 P-i-N HgCdTe FPA photodiodes (in the MWIR and LWIR operating up to 250 and 160 K) have already been reported.^44,45,48 The theoretical estimations carried out in Ref. 46 indicate that further improvements in the quality of P-i-N HgCdTe depleted photodiodes will increase their operating temperature to 300 K.

This paper was supported by the Polish National Science Centre within Project No. UMO-2021/41/B/ST7/01532.

The authors have no conflicts to disclose.

A. Rogalski: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Supervision (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal).

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