Infrared sensors are indispensable in a wide range of applications, including night vision, thermal imaging, environmental monitoring, and healthcare diagnostics^{[1, 2]}. Unlike visible or ultraviolet (UV) light, infrared radiation excels in detecting temperature variations and penetrating obstructions like smoke, fog, or haze, ensuring precise observations in diverse environmental conditions^[3]. Moreover, these sensors uniquely identify both thermal signatures and intrinsic material properties, such as chemical composition, offering insights that other forms of light cannot provide. This dual functionality renders infrared sensors crucial for sectors such as meteorology, astronomy, and military operations, where traditional light sources are inadequate. Initially limited by slow response times and low sensitivity, early thermal infrared sensors like thermopiles and radiometric thermometers were restricted in their performance in practical use. Modern sensors, employing the photoelectric effect, have significantly evolved. These sensors utilize advanced materials and technologies such as high-quality narrow-bandgap semiconductors from the Ⅲ−Ⅴ and Ⅱ−Ⅵ groups, combined with silicon readout circuits, to demonstrate superior infrared imaging capability^{[4, 5]}. The advent of wearable consumer electronics has escalated demands on infrared sensors, particularly emphasizing reduced size, weight, price, power consumption, and enhanced overall performance (SWaP³)^[6]. Traditional sensors, with their rigid structures, often fail in dynamic environments due to susceptibility to damage, further exacerbating challenges related to their heavier weight, higher power consumption, and complex manufacturing processes.

The development of flexible electronic technologies is rapidly transforming the optoelectronic devices, demonstrating significant potential for innovative applications. These technologies are particularly promising in the realm of infrared sensors, which can employ flexible, lightweight, and stretchable substrates to maintain functionality under bending, stretching, and twisting. These sensors hold considerable potential for revolutionizing fields such as biomimetic image sensors, and wearable health monitoring medical diagnostics. Their inherent flexibility allows conformity to curved or irregular surfaces, thereby widening their utility in complex settings. The adoption of innovative substrates and advanced fabrication techniques reduces the weight and power consumption of these sensors, enhancing portability, energy efficiency, and user comfort, which is critical for prolonged use in portable electronics.

The shift from rigid to flexible sensors marks a significant evolution, involving a complete overhaul of material properties and device architectures. This transformation includes the development of materials sensitive to infrared light through the integration of new materials and fabrication techniques. In particular, flexible infrared sensors—whether photoelectric or thermal—present unique challenges compared to their UV−visible counterparts. Photoelectric infrared detectors, with their smaller bandgaps, are more prone to thermal noise, requiring enhanced material stability and noise suppression. Thermal infrared detectors, which detect infrared radiation through temperature changes, demand higher thermal conductivity and resilience to mechanical deformation to ensure consistent performance under bending or stretching. These distinct detection mechanisms highlight the need for careful thermal management and material engineering in infrared sensor design, which differs significantly from UV−visible sensors. Optimal design is crucial for the practical application of these sensors, where the selection and arrangement of flexible substrates and electrodes must be meticulously engineered. Although the introduction of flexible substrates has led to new interface challenges potentially affecting the photoresponse, strategic modifications in material composition and heterostructure design have been employed to optimize performance.

With the growing complexity of applications, single-function flexible infrared sensors solely for detecting light intensity no longer suffice. Today's applications often demand sensors that can handle multiple functionalities such as light detection, signal storage, signal amplification, logical operations, and multispectral detection, among others. However, spatial constraints also make multifunctional integration within a single-layer architecture challenging. Three-dimensional (3D) integration techniques, which involve vertically stacking various functional layers, significantly increase sensor pixel density and facilitate multifunctional integration. This approach meets the demands of complex applications and substantially boosts sensor performance and adaptability. Moreover, the evolution towards smart integration is evident as flexible infrared sensors begin to incorporate sensing, storage, and computational functions within a single pixel, creating a highly-integrated, all-in-one sensory system. Moreover, inspired by biological sensory system such as human brains, advanced sensors that utilize unconventional device physics such as non-linear photoresponsivity, short-to-long term persistence of photoconductivity, gate-tunable spike firing behavior, etc. has been developed to perform data processing or logic operation in optoelectronic sensors. These neuromorphic sensory hardware could simplify or ultimately implement artificial intelligence algorithms, such as deep learning networks, recurrent neural networks, and reinforcement learning models, designed to optimize decision-making processes based on adaptive learning, akin to biological brains. The integration enables the sensors to process data more efficiently, respond dynamically to environmental changes, and reduce latency, significantly boosting the system’s performance and reliability^[7].

This review elucidates the basic mechanisms and key parameters of flexible infrared photodetectors and thoroughly examines the configurations of major infrared photodetector arrays and their applications. We further explore the use of emerging materials including inorganic, organic, and hybrid semiconductors, as well as inorganic low dimensional semiconductors with tunable or narrow bandgaps in flexible sensors. Additionally, we detail advanced fabrication methods and material/devices design strategies that enhance sensor functionality. The discussion extends to material design and integration strategies, highlighting composition and interface engineering, novel flexible techniques and system-level integration technologies that significantly improve sensor performance and applicability. This review also covers the extended applications of these technologies, notably in the domains of wearables and health monitoring, as well as biomimetic vision. Finally, it delineates the strategic trajectories for the advancement of flexible infrared optoelectronic sensors, highlighting their progression towards sophisticated smart sensor technologies and the realization of ultra-high-performance sensor capabilities. By outlining the state of the art in integrated flexible infrared optoelectronics, we hope to provide guidance on the promising avenues to continue advancing this field.

Typically, Infrared radiation, which spans a broad spectrum from near-infrared to far-infrared (0.75−1000 µm), is categorized into several specific bands (Fig. 1(a))^[8−10]. Near-infrared is critical in optical communications and biomedical diagnostics for its minimal loss and deep penetration capabilities^[11]. Short-wave infrared (SWIR, 1−3 µm), capable of penetrating atmospheric water vapor, is vital in remote sensing and night vision applications^[12]. Mid-wave infrared (MWIR, 3−8 µm), known for its sensitivity to temperature variations, is crucial in military and environmental monitoring^[13]. Long-wave infrared (LWIR, 8−12 µm), used primarily in thermal imaging and gas detection, captures thermal radiation emitted by objects^[14]. Very long-wave infrared (VLWIR, 12−30 µm) is essential in astronomy and high-precision temperature measurements, while far-infrared (30−1000 µm) has its niche in specialized fields like astronomy and materials science due to its unique spectral information^{[15, 16]}.

Infrared photodetectors are primarily categorized into two types based on their detection mechanisms: photoelectric sensors and heat-effect-based sensors^{[17, 18]}. Heat-effect-based sensors, including photothermal and thermal radiation sensors, operate on principles derived from thermal effects. Photothermal sensors generate a thermal voltage through a temperature gradient that arises from differential light-induced heating. Thermal radiation sensors measure light power through changes in material resistance, which vary with temperature. Any object with a temperature above absolute zero emits electromagnetic radiation, which includes infrared radiation among others depending on the temperature. Planck's law of black-body radiation precisely describes the relationship between this radiation and the object’s temperature and its emission wavelength:

1M(λ,T)=2πhc2λ5[exp(hcλkT)−1]−1.

Here M(λ,T) is the spectral radiance, λ is the wavelength, T is the temperature, h is Planck's constant, k is Boltzmann's constant, and c is the speed of light. While these sensors are advantageous for broad spectral detection across MIR, FIR, and terahertz ranges, their relatively slow response times may not be suitable for applications requiring quick reactions.

The development of photoelectric infrared sensors has enhanced high-precision detection, excelling in rapid response for dynamic applications, low-light detection, and high-resolution spectral analysis. These detectors are categorized based on their detection mechanisms into photovoltaic and photoconductive types. Photovoltaic detectors operate under zero or low reverse bias, where photogenerated carriers are separated by the built-in electric field. In contrast, photoconductive detectors require an external electric field to drive the photogenerated carriers towards the electrodes for collection.

Further classification of photon-based infrared detectors includes interband transitions, subband transitions, miniband transitions, and impurity band transitions. Interband transition detectors, such as mercury cadmium telluride (HgCdTe) and indium gallium arsenide (InGaAs), interband transitions occur when electrons absorb energy and transition from the valence band to the conduction band^{[19, 20]}. The process is contingent upon the alignment of the material's bandgap with the energy of incident photons. Effective infrared detection is achieved when the bandgap of the material precisely matches the energy of the infrared light, enabling the sensor to detect specific wavelengths. The mechanism is particularly suited for a broad range of infrared wavelengths, especially those closely aligned with the material’s bandgap characteristics. The capability makes interband transition-based detectors versatile tools for applications requiring sensitivity across diverse infrared spectra. The efficiency of interband transitions in infrared detectors hinges on the absorption characteristics of the materials used, primarily determined by the material's bandgap (E_g). The fundamental relationship between a material’s bandgap and the detected wavelength (λ) can be expressed by the equation:

2λ=hcEg,

where h is Planck’s constant, c is the speed of light, and E_g is the bandgap energy. Materials with smaller bandgaps detect longer wavelengths. For instance, to detect wavelengths up to 0.75 µm effectively, a material’s bandgap should ideally not exceed 1.6 eV. Materials with bandgaps around this value typically exhibit reduced optical absorption at these shorter wavelengths, impacting detection efficiency. Adjusting the bandgap energy is the most obvious way to tailor the detection capabilities to specific wavelength ranges, thereby enhancing the sensitivity and selectivity of the detectors for various applications. Fig. 1(b) illustrates the response ranges of several commonly used infrared-sensitive materials, providing a visual reference for these detection capabilities.

Subband (intraband) transitions occur within specially designed band structures such as quantum wells or quantum dots, where electrons move between discrete energy levels or subbands. The mechanism is particularly responsive to narrow spectral ranges, making it ideal for precise spectral analysis or ultrafast optical detection. Additionally, miniband transitions facilitated by superlattice structures allow electrons to transition across broader infrared wavelengths. The capability enhances detector efficiency and is suitable for applications requiring coverage across extensive infrared spectral ranges. Furthermore, impurity band transitions involve electrons moving between impurity bands and the conduction or valence bands, created through intentional doping. The precise control enables detectors to respond to long-wave infrared radiation, predominantly used in astronomical observations and high-precision temperature measurements. These strategies afford a high degree of control over the electronic structure of materials, enhancing the adaptability of detectors to meet specific application needs. The versatility is crucial in the design and manufacturing of infrared detectors, enabling their critical roles across various technological and scientific applications.

Applying photon-based detection techniques to flexible infrared sensors introduces multiple challenges. First, conventional bulk infrared materials such as HgCdTe and InGaAs used in these sensors may undergo structural alterations in their lattice when bent. Such deformations can degrade the material's optoelectronic functionality, thereby affecting its responsiveness to infrared wavelengths. Furthermore, integrating precisely quantum confined or superlattice structures within flexible substrates is challenging, requiring revolutionized fabrication and substantially increased cost. Environmental variations pose another significant challenge, threatening the electrochemical stability and longevity of flexible devices. Effective interface engineering becomes crucial in optimizing the electronic and optical compatibility between different materials, ensuring efficient charge carrier injection and extraction. To overcome these obstacles, the development of new materials, enhancement of device designs, and innovations in manufacturing techniques and interface treatments are essential. These advancements will ensure that flexible infrared sensors achieve high performance and durability, meeting the growing demands of their application fields.

The performance of infrared photodetectors is primarily defined by key parameters that determine their efficiency and adaptability across various environments. Among these, detectivity (D^*), responsivity (R), and response time are the most critical for evaluating overall sensor performance.

Detectivity measures the sensor's ability to detect weak signals amidst background noise. It is typically expressed using the formula:

3D*=A⋅ΔfNEP,

where A is the area of the photodetector, Δf is the bandwidth, and NEP (noise equivalent power) is the minimum detectable power per square root bandwidth. For practical applications where the responsivity and the dark current I_dark are known, the detectivity can also be calculated using the simplified formula:

4D*=R⋅A2qIdark,

where q is the electronic charge. The formula is particularly useful in highlighting the device’s sensitivity by factoring in the intrinsic noise generated by the dark current.

Responsivity indicates the efficiency of the photodetector converting incident light into an electrical signal:

5R=IlightPin,

where I_light is the photocurrent generated under illumination, and P_in is the incident optical power. For voltage-type infrared sensors, responsivity indicates the efficiency with which the detector converts incident light into an electrical voltage:

6R=VoutPin,

where V_out is the output voltage generated under illumination. Measured in volts per watt (V/W), the parameter is crucial for assessing the sensitivity of voltage-based sensors, especially when the focus is on voltage response rather than current.

The response properties of infrared sensors, specifically the rise and fall times, are critical metrics that determine the sensor's ability to adapt to changes in infrared light intensity. Rise time is defined as the duration required for the sensor's output to increase from 10% to 90% of its peak following a sudden increase in light intensity. Conversely, fall time refers to the time it takes for the output to decrease from 90% back to 10% once the light source is removed. Short response times are indispensable in dynamic environments where rapid and precise detection of infrared signals is necessary. The capability is crucial across various applications, including thermal imaging, motion sensing, and real-time spectroscopy, where immediate response and recovery are required to perform effectively.

As for more practical applications, the development of high-operating temperature (HOT) detectors marks a significant technological advance^[21]. These detectors can operate effectively above liquid nitrogen temperature or near ambient temperatures, significantly reducing reliance on complex cooling equipment, thereby optimizing SWaP³, which is foundational to the advancement of flexible infrared sensors^{[22, 23]}. Currently, the only commercial product capable of room temperature operation in the far-infrared spectrum is based on the thermoelectric/bolometric effect. However, these infrared detectors are limited by their slower response speeds and disadvantages in multispectral imaging. To achieve better performance, including higher detection rates, improved response speeds, and broader spectral responses, infrared photon detectors require cryogenic cooling, which is a major barrier to the widespread application of infrared technology. However, operating at low temperatures increases the system’s complexity and reduces its portability. Therefore, developing flexible photodetectors capable of high-performance detection at room temperature is a primary research direction in infrared detectors. One key challenge is managing dark current effectively, a major factor affecting the performance of HOT detectors. Non-equilibrium detector designs, such as Auger suppression and optical immersion, are effective strategies for achieving high-temperature detectors^{[24, 25]}. For instance, the Auger recombination rate (R_Auger) can be approximated as:

7RAuger=C×n3,

where C is the Auger recombination coefficient, typically measured in cm⁶/s, a parameter specific to the material's electronic structure, and n represents the carrier concentration in cm⁻³.

In narrow-gap materials with higher carrier concentrations, Auger recombination significantly impacts the performance of high-sensitivity detectors. Recent strategies for achieving high-temperature infrared detectors include utilizing barrier structures like nBn, which help reduce generation−recombination leakage in the material, thus minimizing dark current production. Another innovative approach is through the concept of photon-trapping detectors, which help reduce dark current caused by carrier movement within the detector material by reducing the volume of detector material needed. These strategies not only enhance detector operational efficiency but also pave new paths for the practical application of high-temperature detectors.

In addition to suppressing Auger recombination, optical immersion techniques are employed to allow infrared optoelectronic devices to operate without cooling. Optical immersion involves placing the detector within a medium with a high refractive index, enhancing the concentration of incident light and effectively increasing the ratio of optical to electrical areas of the detector, thereby boosting photoelectric conversion efficiency and sensitivity. The technique minimizes the physical size of the detector needed, reduces costs, and refines the device architecture, thereby improving overall system performance.

Furthermore, new materials such as InAsSb and type-Ⅱ superlattices have demonstrated uncooled operational performance comparable to traditional thermal detectors with higher cutoff frequencies^{[26, 27]}. Primarily based on group Ⅲ−Ⅴ semiconductors systems and capable of epitaxial growth on high-quality GaAs substrates, these materials provide excellent crystal quality and device compatibility. The characteristic not only offers exceptional optoelectronic performance but also the potential for integration with existing silicon-based technologies, suggesting a significant role in future integrated circuits and optoelectronic systems. The integration capability provides the potential to develop more efficient and compact infrared detection systems, making these materials highly applicable in scientific and commercial applications. To meet the commercial demands for SWaP³, the integration of these materials with flexible substrates is necessary. In addition to conventional semiconductor materials, a new generation of narrow-bandgap low-dimensional materials, as well as organic or hybrid materials, are being considered for future uncooled infrared devices. This will be detailed further in the materials and fabrication strategies section. However, to date, no practical devices based on this technology have been developed, necessitating the combination of new materials and low-temperature processing techniques compatible with flexible engineering.

In the design of infrared image sensors, resolution and fill factor are crucial technical parameters that determine imaging performance. Resolution, measured in pixels per inch (PPI), indicates the sensor's ability to discern the smallest spatial details. High resolution is critical for applications requiring detailed imaging, such as precision medical imaging and remote sensing. Enhancing resolution involves reducing pixel size to increase the number of pixels per unit area, which can be mathematically expressed by the formula:

8PPI=W2+H2L,

where W is the number of horizontal pixels, H is the number of vertical pixels, and L is the diagonal length of the planar sensor. The fill factor is defined as the proportion of the photosensitive area relative to the total pixel area. Increasing the photosensitive area within each pixel can improve the fill factor, thus enhancing the utilization rate of the array sensor. The formula for calculating the fill factor is:

9Fillfator=SpixelSpixel+Sgap,

where S_pixel represents the effective sensing area of a pixel, which is the portion of the pixel capable of detecting light. S_gap refers to the area of the gap between two adjacent pixels. This gap area does not contribute to light detection and therefore counts as part of the non-sensitive region surrounding each pixel. Theoretically, enhancing resolution in the development of flexible infrared sensors, a coordinated effort to reduce pixel size and increase the fill factor is required. However, excessive miniaturization excessively can degrade light absorption and optical guide efficiency. It establishes crucial limits on minimizing pixel size and maximizing fill factor without compromising the overall performance of the sensor. The trade-off significantly restricts the potential for commercial application success. To navigate these challenges, researchers are actively exploring new semiconductor materials, micro-lens technologies, and enhancements to internal pixel structure design. These innovations aim to boost pixel density without sacrificing photoelectric conversion efficiency. Ultimately, such advancements ensure robust imaging performance and expand the application scope of the sensors, aligning technological development with commercial viability.

The operation of image sensor involves several key stages: photocurrent generation, pixel addressing, and optical signal mapping. The generation of photocurrent is foundational, relying on the structural configuration within each pixel to efficiently capture incident light and convert it into electrical signals. Additionally, pixel addressing, a vital step for imaging functionality, selects specific pixels within the array for signal readout. These arrays vary significantly in design and function. Image sensors are classified into three types based on the pixel architecture in the photodetector array: (1) photoconductor photodetector array, (2) photodiode photodetector array, and (3) phototransistor photodetector array^[28]. For instance, photoconductor arrays operate on the photoconductivity effect where incident light modifies the material's conductivity. Such arrays are straightforward to manufacture, making them cost-effective for large-scale production. However, as pixel density increases, challenges such as intricate wiring, cross-talking become more pronounced. The issue is depicted in Fig. 2(a), which illustrates a traditional interdigitated electrode setup struggling with spatial efficiency. Photodiode arrays, utilizing the photovoltaic effect, convert light directly into electrical signals. Their design typically features a vertical cross-bar configuration, as shown in Fig. 2(b), optimizing pixel density without compromising resolution—essential for high-definition imaging applications. Despite their benefits, these arrays face their own challenges, including crosstalk between pixels, which can diminish imaging quality if not adequately addressed through advanced structuring, such as the integration of blocking diodes or the incorporation of thin-film transistor (TFT) technology. Phototransistor arrays, distinct from photodiodes, incorporate transistor-based amplification to enhance both sensitivity and signal gain. The configuration provides more precise control over photo-generated carriers, thereby improving image quality as pixel count increase. Fig. 2(c) presents a phototransistor array layout with horizontally arranged electrodes and vertically stacked gate configurations, specifically designed to reduce crosstalk and ensure consistent resolution across the array.

Traditional photodetector arrays, including photoconductor, photodiode, and phototransistor arrays, form the backbone of optoelectrical imaging due to their reliability, straightforward manufacturing, and adaptability, which are highly valued in scientific research for customizable experiments. Despite these advantages, these arrays face significant challenges such as complex wiring, high noise susceptibility, and notable crosstalk, which can undermine their effectiveness in high-resolution or low-noise applications. Consequently, there is an urgent need to develop more advanced architectures in photodetector technology to enhance performance, reduce noise, and minimize crosstalk. This innovation drive aims to extend the capabilities of these sensors to meet and surpass the stringent requirements of contemporary imaging applications, pushing the boundaries of current technology to better accommodate demanding environments.

Building on the fundamental understanding of classic photodetector arrays and their limitations, it is crucial to examine advanced pixel architectures that address these challenges and enhance infrared sensing capabilities. Passive pixel sensor (PPS) and active pixel sensor (APS) represent significant advancements in photodetector technology, overcoming traditional issues such as crosstalk and complex wiring. These enhanced functionalities not only improve image quality but also reduce noise, making these systems ideal for high-stakes applications where precision is paramount. PPS, optimized for integration on flexible substrates, suits wearable technology applications, while APS, with its sophisticated in-pixel amplification, is well-suited for high-performance commercial imaging systems.

A PPS includes an optical sensor paired with a single switching element per pixel^[29]. In optical sensor arrays, pixels can be structured using one of three traditional architectures such as photoconductor, photodiode, and phototransistor arrays^[30]. The streamlined design enhances high-resolution imaging capabilities, making it well-suited for applications requiring lightweight, thin form factors, such as wearable technology and flexible electronics. The addition of a switching element per pixel permits independent activation, markedly reducing crosstalk and improving image clarity by isolating readout noise from adjacent pixels.

Building on the foundational design of the PPS, Zalar and colleagues introduced a reverse diode array to each pixel, comprising a photodiode and a rectifying diode arranged in a "back-to-back" configuration^[31]. The innovation leverages the photodiode array’s advantages while minimizing rectification loss under illumination, effectively reducing crosstalk. The resulting sensor, featuring a low-crosstalk design, achieves a resolution of 262 ppi, simplifies readout processes, reduces power consumption, and decreases wiring complexity. The approach offers advantages for applications requiring high pixel density, moderate sensitivity, and low-cost manufacturing. Further advancing the PPS architecture, Gasparini and his team developed a photodetector based on non-fullerene materials, marking a significant innovation in infrared imaging^[32]. Their detector uses a novel non-fullerene electron acceptor, rhodanine-benzothiadiazole-coupled indacenodithiophene, combined with poly(3-hexylthiophene). The combination broadens the spectral response to the near-infrared region, achieving an external quantum efficiency (EQE) of 69% and a responsivity of 0.42 A/W, unprecedented in traditional PPS architectures. Moreover, this design demonstrates excellent device stability and dynamic response, making it ideal for high-speed imaging and high-quality image capture. Gasparini's work not only enhances the performance of the PPS architecture but also ensures compatibility with conventional silicon-based active matrix backplanes, allowing easy integration into existing large-area imaging systems. These technological advancements open new possibilities for infrared imaging technology, particularly in applications like biomedical imaging, environmental monitoring, and security, significantly enhancing their potential for impactful applications.

However, despite these advantages, PPS configurations may still encounter wiring noise and crosstalk due to the minimal signal captured by each pixel, potentially leading to a lower signal-to-noise ratio (SNR). Metal−oxide−silicon (MOS) field-effect transistors and TFTs are typically employed as switching elements, critical for effective signal management within these sensors. While PPS systems offer benefits of simplicity and integration ease, their susceptibility to residual noise and crosstalk necessitates meticulous design to ensure optimized performance, particularly in applications that are less demanding or cost-sensitive.

In contrast, an APS integrates an amplifier circuit alongside the optical sensor and switching element within each pixel. This setup allows for the direct amplification of the sensor signal at the pixel level, significantly enhancing the SNR and reducing issues related to wiring noise and crosstalk. APS technology is widely used in commercial imaging applications, especially in complementary metal oxide semiconductor (CMOS) image sensors, where high image quality is achieved through complex circuitry integrated directly into each pixel. Typical APS configurations include a reset switch, a select switch, and an amplification circuit, providing capabilities that surpass the basic light detection functionality of PPS systems.

Expanding upon these technological foundations, Wang and colleagues developed an innovative organic photodetector incorporating an organic light-dependent resistor, an organic resistor, and an organic field-effect transistor, collectively forming an organic photosensitive voltage divider^[33]. This novel, integrated system can instantaneously generate increased voltage under light exposure, effectively amplifying the light signal without enlarging the photosensitive area. The design delivers high contrast between illuminated and dark states, robust current output, and maintains high pixel density. The breakthrough provides new opportunities to enhance the performance and expand the applications of organic photodetectors, particularly demonstrating vast potential in flexible displays and wearable devices.

Expanding on this groundwork, Kim and his team developed a flexible photodetector array utilizing an APS architecture, merging the advantages of organic and metal oxide materials^[34]. Each pixel within this system features dedicated amplification and readout circuits, enhancing the capacity for rapid and high-quality image capture. Integration of a current enhancement feature within each pixel markedly increases the photocurrent output while preserving device flexibility, allowing adaptation to irregular or curved surfaces. The approach is especially apt for high-end optoelectronic applications, including wearable technology and biomedical monitoring, representing a new trajectory in flexible electronics technology that exploits APS architecture to improve image quality and functionality in challenging environments.

These advancements in advanced architecture represent a significant leap forward, melding innovative design with functional applications to enhance the versatility of flexible infrared sensors. By seamlessly integrating complex circuitry such as the APS architecture, these developments not only push the boundaries of organic electronics but also set new benchmarks for the evolution of sensor technologies. Such innovations are crucial in applications demanding high resolution and sensitivity, where they effectively mitigate noise and signal degradation. APS is exceptionally advantageous for demanding applications that require high resolution and sensitivity, effectively addressing noise and signal degradation challenges inherent in simpler systems. The incorporation of complex circuitry within each pixel allows APS to deliver detailed and high-quality images, essential for sophisticated scientific research, high-end photography, and accurate surveillance systems.

Recent advancements in emerging materials have catalyzed significant breakthroughs in the field of flexible infrared detectors. These materials, compatible with flexible fabrication processes, exhibit superior photoelectric properties. They span various categories, including bulk inorganic semiconductors, organic and inorganic−organic hybrid semiconductors, and low-dimensional forms such as 2D, 1D, and 0D semiconductors. Table 1 systematically summarizes the key performance metrics for infrared photodetectors across these categories, highlighting their diverse capabilities and potential applications.

Bulk inorganic semiconductors refer to narrow bandgap 3D materials, specifically thin films or bulk inorganic materials with considerable thickness. The physical properties and performance of these materials are uniform throughout the bulk and are not significantly affected by quantum confinement effects. It includes traditional semiconductors such as Group Ⅳ (Si, Ge)^{[35, 36]}, Group Ⅴ (Te, Se_xTe_1−x)^[37−40], Ⅲ−Ⅴ(GaAs, InGaAs, InAlAs)^{[41, 42]}, and Ⅱ−Ⅵ (HgCdTe, CdS, CdSe, InSb) semiconductors, etc.^{[43, 44]}. As a result of the extensive library of traditional infrared photosensitive materials, a variety of narrow bandgap bulk inorganic semiconductors are compatible with flexible applications upon thinning or/and structural engineering.

Traditional thin film materials such as germanium, which are optimal for near-infrared applications due to their high carrier mobility, become less effective beyond 1.55 µm where their absorption coefficient significantly decreases. An et al. demonstrated a flexible titanium nitride/germanium tin (TiN/GeSn) photodetector that achieves enhanced performance through sub-bandgap absorption, extending its photodetection range^[45]. The fabrication process of the device is illustrated in Fig. 3(a), which includes a photograph of the developed sensor (Fig. 3(b)). Initially, the 90 nm top GeSn layer and the middle 300 nm SiO₂ insulation layer from a GeSn-on-insulator substrate are removed via wet etching. The GeSn film is then transferred onto a polyethylene terephthalate (PET) substrate. A 30 nm thick TiN layer is subsequently sputtered onto the GeSn, creating a TiN/GeSn heterostructure. Simulations indicate that within the 1400 to 2000 nm wavelength range, the TiN layer boosts the average absorption from 0.13 to 0.33. Furthermore, the Schottky barrier height of 0.49 eV between TiN and GeSn extends the photodetection wavelength up to 2530 nm, thereby enhancing light absorption across the detection range.

Materials such as indium arsenide (InAs) and indium antimonide (InSb) are more suitable for mid-infrared applications due to their smaller bandgaps, which are essential for maintaining stability and efficiency. The lattice mismatches with GaAs substrates are approximately 7.2% for InAs and 14.6% for InSb, resulting in high threading dislocation densities in the epitaxial layers (Fig. 3(c)). Despite the lattice mismatch challenges when grown on traditional substrates, which often lead to high dislocation densities, recent advancements in epitaxial growth techniques have improved their performance. Woo and colleagues have enhanced detection capabilities in flexible InAs thin-film mid-infrared photodetectors through high-yield wafer bonding and innovative heteroepitaxial techniques^{[46, 47]}. They introduced a step-graded In_xAl_1−xAs (0.5 < x < 1) buffer layer to effectively bridge the gap between materials with significant lattice mismatches. This approach, compared to traditional linear grading methods, substantially improves the peel-off interface morphology and cuts the threading dislocation density in half. The resulting flexible InAs photodetectors demonstrate excellent optical performance, with a peak room-temperature detectivity of 1.21 × 10⁹ Jones at 3.4 µm, and outstanding device reliability, as illustrated in the cross-sectional schematic of the device, which shows a dense photosensitive layer and a well-formed contact interface (Fig. 3(d)). These advancements position flexible InAs photodetectors as a promising option for next-generation infrared imaging sensors.

While bulk inorganic semiconductors have been the mainstay for commercial infrared detection, transitioning these materials to flexible applications poses significant challenges due to interface incompatibilities and the costs associated with complex manufacturing processes. Techniques like low-temperature deposition and transfer printing show promise for integrating these materials onto flexible substrates. These methods are likely to enhance the mechanical flexibility and photoelectric performance of the detectors, paving the way for the next generation of flexible electronic devices. In summary, while bulk inorganic semiconductors are currently the only commercialized materials for infrared detection, continuous material innovations and advancements in fabrication techniques are essential to expand their applications in flexible technologies.

Organic and hybrid semiconductors are capturing attention for their potential in flexible infrared detectors, prized for cost-effectiveness, ease of fabrication, and versatile physical properties^{[48, 49]}. These materials are typically produced via scalable solution-based processes that strategically balance affordability with effective performance^{[50, 51]}. Their inherent flexibility and tunable photoelectric properties make them ideal for emerging applications in wearable technology, biomonitoring, and environmental sensing. The functionality of organic semiconductors largely depends on the diversity and tunability of their molecular structures, which allow precise control over bandgaps and photoelectric performance through methods like electron-accepting unit integration or donor−acceptor (D−A) copolymerization^[52−54].

Similar to many organic photovoltaic (OPV) devices, the active layers of organic infrared photodetectors typically feature a bulk heterojunction (BHJ) structure^[55]. The structure consists of a finely mixed permeable network of electron donors and acceptors, providing a substantial interface area that aids effective exciton dissociation and forms a bi-continuous transport network, crucial for enhancing charge mobility within the device. Such a configuration improves the charge transport efficiency towards the electrodes and plays a vital role in the overall enhancement of the device's sensitivity and performance.

Building on the foundational concepts of bulk heterojunction structures, recent advancements, such as those made by Siegmund et al. have further pushed the boundaries of organic photodetection technology. They innovatively leveraged intermolecular charge transfer (CT) absorption enhanced by resonant optical microcavities for high-performance narrow-band near-infrared photodetection^[56]. This approach capitalizes on CT states between donors and acceptors, inducing additional optical transitions below the bandgap. By integrating a microcavity structure composed of two silver mirrors—one fully reflective and the other partially transmissive—that also serve as electrodes, the device compactly combines optical and electrical functionalities. The photoactive mixture of C₆₀ as the electron acceptor and zinc phthalocyanine (ZnPc) as the electron donor is positioned between two transparent transport layers, facilitating selective charge extraction upon illumination. Enhanced by the Fabry−Pérot interference effect within the optimized microcavity, this design markedly boosts the photocurrent through enhanced CT absorption. Further advancing the integration of organic and inorganic components, Xiang et al. developed a high-performance, heavy-metal-free flexible photodetector that is sensitive to λ = 1.5 µm photons, as depicted in Fig. 4(a). The device cleverly combines a diketopyrrolopyrrole-based polymer/PC₇₀BM BHJ with inorganic upconversion nanoparticles (UCNPs) of NaYF₄:15%Er³⁺, effectively broadening the response to infrared wavelengths^[57]. The UCNPs, synthesized via hydrothermal methods, are incorporated into the organic host matrix through a solution process, followed by spin-coating onto a substrate. The nanoparticles convert infrared photons into visible light, which the adjacent organic layer absorbs, generating electron−hole pairs and enhancing the photocurrent, illustrated in Fig. 4(b). The device exhibits robust mechanical properties, rapid operational speed, and a photoresponse of 0.44 mA/W under 1.5 µm illumination.

In the realm of infrared photodetection technology, traditional organic materials encounter performance bottlenecks, particularly in terms of efficiency and the range of detectable wavelengths. Addressing these challenges, Wei et al. pioneered the development of photodetectors based on narrow bandgap conjugated polymers, which exhibit superior photoresponse extending from the visible to infrared regions up to 2000 nm. Utilizing the Stille coupling method, they synthesized two novel polymers: poly(benzobisthiadiazolebithiophene-tellurophene) (PBTT) and poly(benzobisthiadiazolebithiophene-4,4'-dioctyloxy-[2,2'-bithiophene]) (PBTB), as showcased in Fig. 4(c)^[58]. These materials demonstrate exceptional broadband absorption, evidenced in the optical absorption spectra (Figs. 4(d) and 4(e)), highlighting their capability from the visible to the infrared spectrum. Notably, the smooth surfaces of the PBTT and PBTB films optimize charge transport and improve contact with electrodes, enhancing photocurrent efficiency and device stability. Specifically, PBTB films enhance π−π interactions, reducing grain boundary impediments and effectively boosting charge mobility, thereby enhancing overall device performance.

In recent years, organic−inorganic hybrid perovskites have emerged as a transformative force in optoelectronics, finding applications in devices ranging from solar cells and LEDs to lasers and photodetectors^[59−64]. These materials are characterized by their excellent photoelectric properties, ease of processing, and cost-effectiveness^[65−68]. Structured in an ABX₃ format, 'A' represents a monovalent cation such as methylammonium (MA⁺), formamidinium (FA⁺), or cesium (Cs⁺); 'B' is a divalent metal cation like lead (Pb²⁺), tin (Sn²⁺), or bismuth (Bi²⁺)^[69]; and 'X' is a halogen anion (Cl⁻, Br⁻, or I⁻). Non-toxic alternatives such as Sn²⁺ and Bi²⁺ based perovskites are becoming increasingly prominent due to their ability to absorb into the near-infrared spectrum, which meets stringent environmental and performance standards^[70].

As advancements in organic and hybrid semiconductors continue, the scope and efficacy of flexible infrared photodetectors are set to expand. Despite the advantages of these materials, such as cost-effectiveness, straightforward processing, and room-temperature operation, they are primarily confined to the NIR and SWIR spectrums and struggle with environmental stability and longevity. The compatibility issues with fine lithography techniques also hinder the development of high-resolution devices. Future research should therefore prioritize the development of materials that not only exhibit broader spectral responses but also align with advanced fabrication processes, ensuring their integration into a wider array of applications.

Advancing from organic and hybrid semiconductors, low-dimensional materials such as 2D, 1D and 0D materials are revolutionizing infrared photodetection^[71]. When the layer thickness is comparable to, or smaller than, the de Broglie wavelength of a thermalized electron, the quantized energy of an electron resident in the layer must be accommodated, in which case the energy−momentum relation for a bulk semiconductor material is no longer applicable. The de Broglie wavelength is expressed as λ_dB = h/p, where h is the Planck’s constant and p is the electron momentum (λ_dB = h50 nm for GaAs). The scenario of energy/momentum quantization happens when the size of a matter reduces from at least one dimension yields substantial advantages in particular for optoelectronics compared with its 3D bulk counter parts, leading to the revolution of quantum-confined or low dimensional electronics and photonics and Nobel prize winning graphene and quantum dots. These materials address the inherent limitations of organic semiconductors by offering expanded spectral responses and enhanced environmental stability. Their broad absorption capabilities extend well beyond the NIR and SWIR spectra, overcoming the spectral constraints of organic alternatives. Low-dimensional semiconductors also bring exceptional mechanical flexibility and high carrier mobility, which are critical for developing optoelectronic devices that maintain performance under environmental stress while operating at room temperature. Their quantum confinement and reduced dimensionality facilitate precise control over electronic properties, enhancing device performance and efficiency. The complex synthesis techniques required for these materials, such as vapor deposition and mechanical exfoliation, are compatible with advanced lithographic methods. This compatibility is essential for crafting high-resolution devices and extending their application across a broader range of advanced technologies. By overcoming the challenges associated with organic semiconductors, low-dimensional materials are set to significantly enhance the scope and effectiveness of flexible infrared photodetectors, promising to exceed the performance requirements of future applications and driving forward the evolution of semiconductor technologies.

Known for their exceptional mechanical flexibility and electrical properties, 2D infrared-sensitive materials are becoming pivotal in the development of flexible optoelectronic devices, especially in infrared photodetection^{[72, 73]}. The range of 2D materials currently responsive to infrared light includes graphene, transition metal dichalcogenides (TMDCs, such as MX₂ where M is a transition metal and X is a chalcogen element like S, Se, and Te)^[74], MX compounds (including GeSe and SnS)^[75], M₂X₃ compounds (such as Sb₂Se₃, Bi₂Se₃, Bi₂S₃)^[76], black phosphorus (BP)^[77], oxyselenides (like Bi₂O₂Se, Bi₂Se₂S)^[78], black arsenic phosphorus (As_xP_1−x)^[79], and 2D Te^[80], among other emerging materials.

Graphene, the first isolated 2D material, is renowned for its broad spectral response and high carrier mobility, absorbing light from ultraviolet to terahertz wavelengths^{[81, 82]}. Its excellent mechanical and electrical properties make it ideal for high-performance flexible infrared sensors. Moving beyond graphene, TMDCs offer advantages due to their variable bandgaps and structural versatility, making them suitable for visible to NIR applications. To extend their spectral response, innovative approaches have been employed, such as combining TMDCs with other 2D materials to enhance their properties. For instance, significant advancements have been demonstrated by integrating SnSe and MoTe₂ with graphene. Xu et al. enhanced SnSe's infrared detection capabilities to 10.6 µm using a sputter deposition technique that exploits both photoconductivity and photothermal effects. The method enables the detection of infrared light by altering the electrical resistance of the SnSe film in response to temperature changes caused by sub-bandgap photon absorption^[83].

However, despite the potential of TMDCs and other 2D materials, challenges remain in their practical application, particularly concerning large-scale production and the enhancement of photoelectric properties. Innovative approaches such as the non-destructive dispersion strategy developed by Velusamy et al. have shown promise. This strategy uses long-chain polymers to improve the flexibility and performance of TMDC films, significantly enhancing their photoelectric characteristics. Figs. 5(a) and 5(b) illustrate the schematic diagrams of the 2D TMDC nanosheet structures (M: Mo, W, Re, Nb, Zr, Ta; X: S, Se, Te), along with representative amine-terminated polymers and their broad absorption spectra^[84]. The process begins by mixing MoSe₂ powder with polystyrene amine (PS−NH₂) and subjecting it to ultrasonication to form a stable dispersion. The mixture is then centrifuged and vacuum dried to prepare the films, as shown in Fig. 5(c). A simple solution-mixing method allows for the adjustment of MoSe₂ and MoS₂ composite films, achieving selective detection across a wide range of light wavelengths, as depicted in Figs. 5(d) and 5(e). This method opens a new avenue for the efficient, scalable fabrication of TMD-based photonic devices.

The 2D materials such as graphene, black phosphorus, and TMDCs have shown potential in the field of infrared photodetectors. However, their practical application still faces numerous challenges. For instance, the relatively low light absorption capabilities of monolayer graphene and black phosphorus—approximately 2.3% and 3%, respectively—limit their efficiency in photoelectric conversion without further optimization. Moreover, the technology for producing large-area, high-quality 2D materials remains underdeveloped, affecting the uniformity and reliability of material performance^[85]. Currently, production methods like mechanical exfoliation or vapor deposition can produce high-quality samples but often do not provide the economic feasibility or repeatability needed for mass production. Additionally, integrating 2D materials into existing device architectures is crucial for their commercial application. Universal fabrication strategies, compatibility with traditional silicon-based technologies, and environmental stability are key issues that need addressing in ongoing research and development. Overall, while 2D materials hold broad application prospects as infrared photodetectors, achieving their commercial application necessitates in-depth research and technological innovation in material preparation, performance optimization, and device integration.

Quantum dots (QDs), zero-dimensional nanomaterials composed of semiconductor nanocrystals ranging from 1 to 10 nm, demonstrate enhanced photoelectric properties due to quantum size effects and confinement^{[9, 10]}. The properties make QDs ideal for infrared photodetection, particularly when using group Ⅱ−Ⅵ materials like cadmium selenide (CdSe) and cadmium telluride (CdTe), group Ⅳ−Ⅵ materials such as lead sulfide (PbS) and lead selenide (PbSe), as well as group Ⅲ−Ⅴ materials like indium arsenide (InAs) and among others^{[86, 87]}. Their tunable bandgaps and excitonic properties allow for precise control over photoelectric conversion properties, which can be optimized by adjusting the particle size to improve infrared detection performance.

The efficiency of quantum dots in converting light to electricity benefits significantly from their high surface-to-volume ratio, which enhances light absorption and photosensitivity within the infrared range. These materials are commonly synthesized through cost-effective, low-temperature methods such as hot injection and cation exchange, facilitating scalable production. For instance, a notable development by Liu and colleagues involves a flexible PbS colloidal quantum dots (CQD) photodiode array designed for multispectral image fusion (Fig. 6(a))^[88]. This array employs a p−i−n heterostructure that optimizes the extraction and transport of photogenerated carriers, depicted in Fig. 6(b), which effectively reduces carrier recombination and expands the device's spectral response from X-rays to NIR. Addressing manufacturability and stability, enhancements in quantum dot inks have been achieved by integrating polymers such as polyimide, which improves film passivation and stability, as demonstrated by the reduction in dark current and increased uniformity in devices by Liang and colleagues (Fig. 6(c))^[89]. Additionally, the films show enhanced infrared absorption and longevity, shown in Fig. 6(d). Further advancing the technology, Zhou et al. have introduced a solution-processed CdSe QD photodetector that incorporates silver nanoparticles to enhance infrared to visible light conversion, achieving high detectivity and rapid response times (Fig. 6(e)), with capabilities demonstrated through a shadow mask imaging strategy (Fig. 6(f))^[90].

Despite their outstanding photoelectric performance and adaptability to flexible substrates, which make them suitable for wearable electronics, quantum dot-based infrared detectors still face challenges such as limited spectral range and environmental stability issues, including heavy metal toxicity, and low carrier mobility. The low mobility in quantum dot-based detectors is primarily due to the quantum confinement effects, which restrict the movement of charge carriers within the small dimensions of the dots. Additionally, the hopping transport mechanism between dots is inherently less efficient than the band-like transport found in bulk materials. Future research should concentrate on developing narrower bandgap materials that are environmentally benign and stable, improving surface modification techniques to enhance carrier mobility, and advancing recycling methods to enhance the commercial viability of these technologies.

In the dynamic field of semiconductor technologies, quasi-1D semiconductors such as nanowires (NWs) and nanorods (NRs) are emerging as viable alternatives to their 2D and 0D counterparts^[91]. Unlike 2D materials, which are characterized by their expansive planar structures, and 0D materials, known for their point-like discrete energy states, 1D semiconductors offer unique benefits with their extremely high aspect ratios and small nano-scale diameters ranging from several to tens of micrometers. These characteristics endow 1D materials with an enhanced surface-to-volume ratio and superior photoelectric properties—key advantages that significantly boost the performance of flexible infrared photodetectors. Si and Ge NWs are particularly notable in this category due to their high carrier mobility and thermal stability. Si NWs, for example, are valued for their adjustable bandgap and high carrier mobility, attributes that are crucial for the development of highly sensitive near-infrared sensors. Despite their potential, the mechanical flexibility and cost-effectiveness of these 1D materials pose challenges, especially when scaling up for high-performance wearable technologies. Addressing these challenges, Wang et al. utilized an electrospinning technique to fabricate a flexible NIR photodetector featuring a 1D all-organic heterostructure, consisting of a polyacrylonitrile (PAN) and polyaniline (PANi) core-shell design (Fig. 7(a)). The construction significantly enhances charge separation and collection, paving the way for the development of cost-effective, high-efficiency flexible optoelectronic devices^[92].

In practical applications, where performance requirements are stringent, devices often rely on external amplification circuits to enhance light sensitivity and image recognition capabilities. However, these additions can increase power consumption and complicate integration and noise management. To combat this, Ran et al. introduced an infrared detection system employing gallium-doped indium oxide (In₂O₃) NWs configured in a top-gate field-effect transistor (FET) architecture (Fig. 7(b))^[93]. The setup not only improves gate efficiency but also reduces parasitic capacitance and optimizes integration with TFT circuits, achieving a remarkable light sensitivity of 7.6 × 10⁴ under 1342 nm illumination—a record for similar systems (Fig. 7(c)). To address the challenge of limited effective areas in individual nanowires or nanorods, Meng et al. developed a method to synthesize wafer-scale tellurium (Te) van der Waals nanomeshes at just 100 °C. These nanomeshes, which can grow on a variety of substrates—including rigid, flexible, and curved surfaces—exhibit high field-effect hole mobility and ultrafast light response characteristics, making them suitable for high-performance flexible infrared detectors (Fig. 7(d))^[94].

To address the issues of uniformity and insufficient effective area in 1D materials, Liu et al. introduced a hybrid active layer comprising Ge QDs decorated on reduced graphene oxide fragments^[95]. The integration of dual infrared-sensitive materials not only maintains a rapid response speed but also enhances the infrared spectral response. Similarly, Mukherjee and colleagues developed a hybrid 0D/2D structure by integrating 2D TMDCs materials, specifically MoS₂, with 0D PbS quantum dots^[96]. The synthesized hybrid nanostructures exhibit extended photo-generated carrier lifetimes, a critical feature that significantly improves the photoelectric conversion efficiency, thereby greatly enhancing the performance of photodetectors.

Given the unique attributes of 1D semiconductors and the emerging hybrid-dimensional structures, future research and development efforts should aim to refine synthesis and integration techniques. 1D materials, characterized by exceptional carrier transport and photoelectric conversion efficiencies, face challenges in uniformity and scalability in large-scale production. Hybrid-dimensional structures can effectively mitigate these issues, enhancing the functionality and applicability of these materials in advanced applications^[97]. Advanced fabrication methods such as vapor deposition and self-assembly are essential for improving production efficiency and device performance. Moreover, hybrid-dimensional semiconductors, combining different dimensional materials like quantum dots with nanowires or two-dimensional layers, show enhanced photo-response due to improved carrier lifetimes and interaction dynamics. These synergies could pave the way for next-generation flexible infrared detectors with superior performance across a broad spectrum of applications.

In the development of flexible electronics, particularly infrared photodetectors, molecular beam epitaxy (MBE) and similar technologies lay the groundwork by fabricating high-quality semiconductor materials^[98]. However, the integration of these materials onto flexible substrates predominantly hinges on the innovative application of post exfoliation and transfer techniques. These methods are essential for adapting traditionally rigid, high-quality materials to the dynamic, bendable surfaces required in wearable and portable technologies. The essence of post exfoliation and transfer techniques is their ability to separate high-quality epitaxial HgCdTe, GaAs, and InGaAs and other similar films, from their original rigid substrates and transfer them onto flexible ones. Importantly, these techniques circumvent the limitations associated with lattice mismatch and the high-temperature growth environments typically required for these materials, facilitating their integration into flexible devices.

For instance, the epitaxial lift−off (ELO) technique has been developed, enabling the integration of high-quality active layers onto flexible substrates. The process is vital for fabricating flexible and reliable optoelectronic devices. Woo and colleagues demonstrated the fabrication of a high-detection flexible InAs mid-infrared photodetector array using ELO^[47]. The device features a vertical p−i−n structure and undergoes processes including bonding metal deposition, mesa etching, wafer bonding, hetero-epitaxial lift−off, and top metal deposition (Figs. 8(a) and 8(b)). The structural integrity of the device is maintained even after repeated bending and recovery cycles, highlighting its robustness.

Further advancing this technology, Pan and colleagues have successfully grown and transferred HgCdTe (111) epitaxial thin films directly on a 2D transparent mica substrate through MBE and post exfoliation and transfer techniques^[99]. The method achieves a peak responsivity of approximately 110 V/W at 3500 nm and 80 K, and around 8 V/W at room temperature under a 25 V/cm bias. The weak van der Waals bonding between HgCdTe and the mica substrate facilitates an etch-free layer peeling/transfer process. The technique has proven effective for creating ultra-thin and flexible HgCdTe layers, potentially revolutionizing the fabrication of flexible infrared sensors. The innovative method employs the weak van der Waals forces between HgCdTe and mica to facilitate an etch-free layer transfer process. The approach not only streamlines manufacturing but also minimizes potential damage to the epitaxial layers. Demonstrating high responsivity at critical infrared wavelengths and temperatures, these films prove ideal for the development of flexible, high-performance infrared sensors.

Post exfoliation and transfer techniques have demonstrated considerable success in the large-scale transfer of van der Waals thin films, contributing significantly to the field of flexible electronics. The capability to fabricate extensive areas of high-quality films highlights the scalability and adaptability of these techniques for modern device applications. The recent work by Yang and colleagues exemplifies this success, showing the effective use of these techniques for growing 2.0 × 2.0 cm² T_d-MoTe₂ van der Waals thin films directly on flexible mica substrates^[100]. The weak van der Waals interactions between MoTe₂ and mica substrates have showcased the tremendous potential of this approach, facilitating an etch-free layer transfer that maintains the material's integrity and functionality. The flexible photodetector exhibits an ultra-broadband sensitivity ranging from ultraviolet to sub-millimeter waves (325 nm to 566.0 µm). Moreover, the sensor is capable of high-resolution terahertz imaging and can detect shielded objects. These achievements not only showcase the potential application of semimetal T_d-MoTe₂ in future high-performance broadband photodetectors but also pave the way for the development of flexible wearable optoelectronic devices using transferred large-area vdW crystals. Clearly, these inorganic infrared photosensitive materials, epitaxially grown using techniques such as MBE, are compatible with traditional photolithography and etching processes. The compatibility facilitates the straightforward patterning of array devices based on epitaxial films after they have been lifted off.

Despite the significant advancements in fabricating flexible infrared sensors through post-exfoliation and transfer techniques, several challenges persist. Key challenges in the fabrication of flexible infrared sensors include potential damage and contamination during the transfer process. These issues can severely compromise the integrity and performance of the thin films, reducing their quality and introducing complexities in the production process. Ongoing research and development are vital to refine these techniques and minimize both damage and contamination. Additionally, exploring alternative approaches, such as low-temperature vapor deposition, holds promise. Investigating these innovative deposition techniques is essential for fully realizing the potential of flexible infrared detectors and ensuring their successful commercialization across various applications.

Building upon the advancements in post exfoliation and transfer techniques, the field of flexible infrared sensors is increasingly turning towards innovative deposition methods that offer precise control over material properties at reduced temperatures. Low temperature vapor deposition, encompassing both physical vapor deposition (PVD) and chemical vapor deposition (CVD), represents a paradigm shift towards methods that allow for the deposition of thin films and advanced materials without the high thermal stress typically associated with traditional high-temperature processes^[101]. These deposition techniques, fundamental to the fabrication of flexible infrared sensors, are particularly beneficial because they maintain the structural integrity and functional properties of materials while being deposited directly on flexible substrates, making them ideal for constructing flexible infrared photodetectors.

PVD is a method that transitions materials from solid or liquid states to gases and deposits them onto substrates as thin films^[102]. With its lower operating temperatures, PVD is particularly advantageous for flexible substrates like polymers, which are susceptible to thermal damage under high-temperature processes. Evaporative deposition and sputtering, two common PVD techniques, are preferred in the production of flexible electronics for their ability to deposit materials without deforming or decomposing the substrate. PVD's straightforward and cost-effective approach facilitates large-scale production, reducing costs and enhancing manufacturing efficiency.

Demonstrating PVD's adaptability, Zhou et al. successfully developed stable and efficient infrared photoelectric sensors by depositing an Sb_0.405Te_0.595 film on a substrate pre-coated with a thin Sn layer, achieving broad spectral detection from the visible to MWIR range (405−4500 nm) (Fig. 9(a))^[103]. The photoactive layers significantly improved detection capabilities in the MWIR spectrum, reducing response times by approximately a thousandfold. The SEM images showcasing film morphologies over varied deposition durations (Fig. 9(b)) illustrate the ability to tailor phase composition and crystal morphology by adjusting growth durations. Advancing PVD's capabilities further, Wen et al. utilized physical vapor transport deposition to grow Sb₂Se₃ films epitaxially on flexible mica substrates (optical image in Fig. 9(c)), with X-ray diffraction confirming their superior crystallinity, evidenced by narrow rocking curve half-widths of 0.25° (Fig. 9(d)). These epitaxial films nearly doubled the photocurrent compared to traditional non-epitaxial Sb₂Se₃ film photodetectors (Fig. 9(e)), demonstrating the extended capabilities of PVD in enhancing photodetector performance^[44].

The versatility of PVD allows for the optimization of the deposited films' types, properties, and functionalities. Illustrating this adaptability, Wang et al. achieved a breakthrough by fabricating a mixed BHJ film through co-evaporation of DP-OMe and C₆₀ using thermal deposition. This innovation notably improved the performance of NIR photodetectors^[104]. The semi-transparent detectors excelled in EQE across both small (6.44 mm²) and large (256 mm²) areas, achieving EQEs of 34% to 36% and detectivities of 1.4 × 10¹³ and 1.1 × 10¹² Jones, respectively. The performance matched that of conventional silicon-based inorganic photodetectors, highlighting the effectiveness of organic materials in advanced photodetection applications. Furthermore, these organic photodetectors have been successfully employed in NIR imaging and biosensing and integrated with NIR organic light-emitting diodes, showcasing their potential in invisible light communication and underscoring the practical applications of PVD-fabricated devices.

While PVD demonstrates promising applications in the fabrication of advanced materials like topological semimetals, amorphous films, and hybrid materials, contributing to the development of fast-response, high-stability flexible photodetectors, it predominantly produces polycrystalline or amorphous structures^[105]. The limitation can be restrictive in applications requiring high-performance single-crystal films. Although tube furnaces can grow two-dimensional single crystals, scaling these methods to support large-area material production remains a challenge. Consequently, ongoing research aimed at improving and controlling the crystalline quality of PVD-deposited films is crucial, potentially broadening the technology's application in high-performance areas such as wearable devices and environmental sensing.

CVD surpasses the limitations of PVD in depositing 2D infrared-sensitive materials, especially in crafting large-area high-quality films^[106]. CVD introduces precursor gases onto a heated substrate, where they react to form solid films under controlled conditions, such as gas flow, reaction temperature, and process duration^[76]. This method is particularly effective for creating large-scale, high-quality materials that are challenging with PVD^[107]. For example, Choi et al. employed plasma-enhanced CVD (PECVD) to develop ultra-flexible 2D-MoS₂/Si heterojunction photodetectors at temperatures below 200 °C^[108]. The approach bypasses PVD's thermal constraints on flexible substrates and directly deposits MoS₂ onto flexible silicon, enhancing light absorption and interaction. The resulting heterostructure demonstrated exceptional photosensitivity under near-infrared illumination (λ = 850 nm), achieving a responsivity of 10.07 mA/W and a specific detectivity of 4.53 × 10¹⁰ Jones.

Expanding the range of materials and designs, CVD also allows for the deposition of 1D NWs. GaSb NWs, notable for their narrow direct bandgap of 0.726 eV and high hole mobility, are particularly effective for flexible NIR photodetectors due to their superior infrared response. Utilizing silver as a catalyst, Sa et al. achieved size-controlled synthesis of high-quality GaSb nanowires, which exhibit excellent electrical and mechanical flexibility and are compatible with CMOS processes, enhancing photodetector performance with a responsivity of 618 A/W and detectivity of 6.7 × 10¹⁰ Jones under 1550 nm laser illumination^[91].

Further advancing CVD techniques, Hoang et al. optimized the process by lowering the operating temperature to approximately 150 °C using metal−organic CVD (MOCVD) (Fig. 10(a))^[109]. The innovation facilitated the synthesis of high-quality, highly crystalline single-layer MoS₂ directly on ultra-thin polymer and glass substrates. The method eliminates the need for material transfer, preserving the quality of film and achieving large-area growth of single-layer MoS₂ on flexible parylene substrates (Fig. 10(b)). The transparency of both the substrate and MoS₂ film enables these phototransistors to capture light from both top and bottom, expanding their utility in flexible electronics. 2D MoS₂-based flexible phototransistors also exhibit robust switching characteristics across a spectral range of 405 to 904 nm. Overall, CVD's precise control over the deposition process on large-area substrates provides significant advantages in manufacturing flexible infrared photodetectors. The direct growth of high-quality films and nanowires on flexible substrates via CVD minimizes damage during the material transfer process.

In the realm of arrayed devices, the majority of inorganic infrared-sensitive materials fabricated using low-temperature vapor deposition strategies are patterned through processes compatible with traditional photolithography, lift−off, and etching techniques. The compatibility simplifies the fabrication of arrayed devices significantly. For materials that are not compatible with conventional patterning strategies, an alternative approach involves depositing an entire photosensitive layer directly, followed by a sacrificial protective layer. This layering facilitates the subsequent removal of excess active material using etching techniques. The final device patterning is achieved by removing the sacrificial layer, thus preparing the patterned materials and devices for integration.

Low-temperature vapor deposition techniques, such as PVD and CVD, are pivotal in the development of flexible and smart infrared sensors. These methods facilitate the deposition of inorganic, infrared-sensitive materials onto substrates that cannot endure high temperatures, preserving both their mechanical and structural integrity. Additionally, they ensure compatibility with established semiconductor processes like photolithography and etching, which improves integration into existing manufacturing workflows and enhances both scalability and efficiency. Despite these advantages, the techniques require costly equipment and must operate under stringent vacuum conditions, limiting the diversity of applicable materials and complicating the production process. This restricts their use in cost-sensitive markets. Consequently, exploring deposition technologies that are both more versatile and cost-effective, while accommodating a wider range of materials, is a crucial research direction. Such advancements could significantly benefit the development of the next generation of flexible electronic devices, where cost efficiency and material diversity are essential.

Low temperature solution processing techniques such as spin coating, doctor blading, roll-to-roll coating, dip coating, transfer printing, and inkjet printing present a cost-effective and scalable method for fabricating functional films on extensive flexible substrates^[50]. These methods employ straightforward solution deposition followed by annealing, a process that not only yields large area and high-quality films but also, owing to the low-temperature conditions, renders these techniques particularly effective for the fabrication of flexible infrared optoelectronic sensors. It is particularly adept at depositing a diverse thin film and array of infrared-sensitive functional materials including perovskites, organic semiconductors such as poly(3-hexylthiophene) (P3HT) and fullerene derivatives (PCBM), 2D materials, and QDs^{[110, 111]}. Ge and colleagues have leveraged solution processing to develop innovative flexible hybrid photodetectors that function efficiently from the visible to MIR spectra at ambient temperatures^[50]. In this process, a balanced mixture of P3HT and PCBM in 1,2-dichlorobenzene is spin-coated onto an indium tin oxide (ITO) substrate pre-treated with poly(styrenesulfonate) (PSS) and thermally annealed at 130 °C. It is followed by the transfer of a three-dimensional graphene (3DG) film and the evaporation of aluminum electrodes, culminating in a device that exhibits a remarkably high responsivity of 5.8 × 10⁵ A/W and detectivity of 3 × 10¹⁵ Jones, capable of detecting picowatt-level light. Further enhancements in fabrication efficiency and cost reduction are exemplified by the synthesis of heterostructures that combine PbS QDs and ZnO nanoparticles (NPs) through solution-based methods^[112]. The fabrication processes and the schematic of light detection are shown in Figs. 11(a) and 11(b), respectively. The heterostructures show enhanced and broadened visible to NIR absorption, a fact confirmed by comparative absorption spectra (Fig. 11(c)). I−V curves under various wavelength illuminations demonstrate high sensitivity and a large on/off ratio, attributable to the robust light-absorbing properties and NIR sensitization by PbS QDs (Fig. 11(d)). Additionally, these devices maintain excellent durability and stability, even after 500 bending cycles.

Traditional coating techniques such as spin coating and doctor blading are not directly capable of fabricating patterned material arrays^{[113, 114]}. However, by integrating surface-functionalization strategies, it is possible to prepare patterned arrays on substrates with varying surface energies. This effectiveness arises because materials processed via solution methods tend to nucleate and grow preferentially in areas with higher surface energy (hydrophilic regions), forming films, while areas with lower surface energy lack the conditions necessary for material growth. This results in the formation of high-density, patterned arrays of active layer materials^[115]. Expanding upon these developments, flexible printing technology, a branch of solution processing, employs methods such as inkjet and transfer printing to deposit functional materials directly onto various substrates. The strategy facilitates large-scale production at reduced costs and enables direct patterning without additional lithographic steps, substantially streamlining the fabrication process. A prominent application of this approach is detailed in Fig. 11(e), where Ag/Ge₂Sb₂Te₅/Ag heterostructures are fabricated using full-template printing, producing Ge₂Sb₂Te₅ microwires with a responsivity up to 6.7 × 10⁴ A/W under a 980 nm wavelength. These devices achieve the highest NIR photodetector resolution to date, reaching 1.0 × 10³ ppi. Assessments of devices with varying curvature radii under identical light intensity show no significant photocurrent variations, underscoring the robust potential of flexible printing technology in producing high-performance photonic devices (Fig. 11(f))^[114].

In conclusion, low temperature solution processing techniques present a highly efficient and economical approach for the production of functional films on large-area flexible substrates^[116]. The technology reduces costs and supports the development of innovative materials, enhancing prospects for flexible electronics. The advancements are poised to resolve technological challenges, fostering more efficient, cost-effective, and environmentally friendly solutions in flexible photodetector technology.

Composition engineering focuses on optimizing photoelectric performance and mechanical flexibility by modifying the composition and structure of materials^[117]. In flexible infrared photodetectors, varying combinations of materials substantially influence device performance. For instance, material bandgaps and carrier mobility can be adjusted through alloying and doping to enhance photodetection response. Alloying, a common approach in composition engineering, involves blending different semiconductor materials to modify their photoelectric properties. Within the scope of composition engineering, the introduction of an In_0.8Al_0.2As barrier layer in flexible p−i−n InAs thin-film photodetectors as demonstrated by Woo and colleagues optimizes device performance. The barrier layer significantly impedes electron flow, reducing leakage current by 283 times. Bend tests further confirm the high mechanical stability and reliability of the detector in flexible electronics^[47].

Doping is a prevalent method for adjusting material conductivity and enhancing light absorption characteristics. Xia and colleagues have developed two-dimensional high electron-donating central core acceptors, YZ and YZ1 (Fig. 12(a)), specifically designed for NIR organic photodetectors. These materials exhibit excellent spectral complementarity and energy level matching with the PCE-10 donor due to precise doping (Fig. 12(b))^[118]. Extending the YZ terminal conjugation system expands the absorption spectrum and minimizes trap states (Fig. 12(c)), thus enhancing charge transport. The modification not only broadens their light-harvesting capabilities but also significantly improves the efficiency of the devices. The energy level diagram and device structure of the p−i−n organic photodetector with the YZ1 active layer, as shown in Fig. 12(d), demonstrate a high responsivity of 0.27 A/W and a detectivity of 9.24 × 10¹³ Jones at 1000 nm, making it competitive with commercial silicon photodiodes. Moreover, perovskite materials have expanded their light absorption into the near-infrared region by doping with metal or organic cations, or by altering halide ion types. Experiments indicate that cation doping can control the carrier diffusion length in perovskites, a critical factor considering the absorbance coefficient and carrier diffusion length in device design (Fig. 12(e)). Simulation of charge distribution at various incident wavelengths illustrates this effect (Fig. 12(f)). Feng et al. have achieved a redshift in the EQE spectrum by adjusting the I and Br doping ratio in the spray precursor solution for quasi-2D phenethylammonium/methylammonium lead halide (PEA₂FA_n−1PbnX_3n+1) perovskites, as shown in Fig. 12(g)^[119]. Optimal spray speeds and solution concentrations allow for control over film thickness from tens of nanometers to several hundred micrometers, achieving FWHM less than 20 nm for narrow-band light response. Spray strategies have enabled in situ deposition on lensless hemispherical surfaces, particularly suited for applications requiring high sensitivity and precise wavelength recognition, such as advanced imaging systems, multi-wavelength sensors, and smart recognition technologies.

The appropriate combination of materials by enhancing device stability and reliability across diverse environments is crucial for commercial applications. Composition engineering, through subtle adjustments in material interactions and properties, not only advances the development of novel flexible photodetectors but also opens new possibilities for the design and realization of future high-performance flexible electronic devices.

Interface engineering concentrates on refining the interaction between materials, substrates, and electrodes to boost the performance and durability of flexible infrared photodetectors. Key strategies such as surface treatments and the incorporation of specialized interface layers significantly improve interface quality, decrease defects, and enhance carrier transport efficiency^[120]. Through methods like surface treatment and the deployment of interface layers, interface engineering effectively minimizes defects at the junctions, crucial for advancing carrier transport efficiency and thereby enhancing device functionality. Although current research on interface passivation primarily targets rigid substrates, the techniques developed are largely applicable to flexible sensors. Typically, the interfaces between infrared-sensitive materials and their electrodes or substrates exhibit numerous defect states. These defects capture photogenerated carriers, increasing carrier recombination and diminishing the photoelectric conversion efficiency.

In the realm of solution processing, especially when applying spin-coating techniques to hydrophobic flexible substrates, challenges such as uneven distribution, weak adhesion, augmented interface defects, and diminished environmental stability are common. The problems often result from poor compatibility at the material−substrate interface, which adversely affects device performance and durability. To reduce these challenges, techniques such as plasma treatment, UV irradiation, and chemical modification are used to optimize the substrate's hydrophilicity, improve adhesion, and reduce interface defects, thus enhancing the device's overall photoelectric efficiency and environmental resilience. Moreover, introducing a passivation layer at the interface can effectively diminish these defect states. For example, employing chemical treatments or depositing thin layers of organic or inorganic materials can seal off dangling bonds or other defect states at the interface, reducing carrier recombination and boosting the device's photoelectric performance. An illustration of the effect is seen in the response times of carbon nanotube (CNT) structures. On SiO₂ substrates, the response time is about 40 ms, which is considerably better compared to the 1200 ms on PMMA substrates^[121]. By transferring CNT/SiO₂ structures onto PMMA, teams have markedly shortened the response time of new flexible infrared detectors to 50 ms. The detectors excel at monitoring subtle thermal or far-infrared radiation and have proven capable of detecting minor movements like a human finger twitch, showcasing their expansive potential in practical scenarios.

Interface engineering, this field also encompasses strategies for interface modification and passivation, each aimed at amplifying device performance. Interface modification frequently involves chemically introducing functional molecules to the material's surface to increase interface transmission efficiency. Additionally, interface passivation entails adding layers that seal defects and hinder non-radiative carrier recombination, thereby improving photoelectric conversion efficiency and stability. For instance, Jang et al. utilized an alcohol-soluble tetrakis(phenylethynyl)benzene unit with sulfonate chains as a passivation layer to address oxygen vacancies in ZnO, effectively enhancing the efficiency of organic photodetectors. The innovation establishes a foundation for further research and underscores the critical role of passivation layers in advancing semiconductor performance^[122]. In additional studies, Liu and colleagues used focused ion beam technology to precisely tailor the WSe₂ layer, removing Se atoms to form Se vacancies and heterojunction structure. The introduction of a heterostructure of graphene and WSe₂ enhanced electron mobility and fortified the device's mechanical stability. Using PBDB-T as a passivation layer significantly curbed surface charge recombination caused by Se vacancies, markedly boosting the photoelectric conversion efficiency and stability, allowing these photodetectors to demonstrate robust light response and reliability in practical applications^[123].

In conclusion, by consistently enhancing interface characteristics and implementing effective passivation strategies, substantial progress can be achieved in the development and commercial deployment of flexible infrared photodetectors, especially in vital areas such as environmental monitoring and health diagnostics.

Optimization of heterojunction structures is pivotal in enhancing the interface structures between different materials to boost device performance. In photodetectors, heterojunctions play a key role as they efficiently separate and transport photogenerated carriers, thus minimizing carrier recombination and enhancing photoelectric conversion efficiency^[124]. Typically, this optimization involves fine-tuning the bandgap alignment between materials to achieve ideal carrier dynamics, as well as improving material contacts and carrier transport efficiency through refined interface engineering. Fig. 13(a) delineates the structure and operational principles of a flexible photodetector based on a dual heterojunction of SWCNT/graphene/MoS₂, while Figs. 13(b) and 13(c) showcase the optical image and the fabrication process of the device, respectively. The integration of gadolinium iron garnet (Gd₃Fe₅O₁₂) thin films at the dual heterojunction interface significantly reduces dark current, optimizing device efficiency^[125]. Within this configuration, the MoS₂ layer acts as the visible light-absorbing layer, graphene serves as a transparent electrode, and together with SWCNT, forms a van der Waals heterojunction that optimizes the separation and transport of photogenerated carriers. The SWCNT layer, functioning as a NIR absorbing layer, broadens the detector's response range and bolsters its performance across a wide spectrum. The setup improves the photodetector’s photoresponse and enhances its flexibility, making it adaptable to a variety of application environments.

Further strategies include constructing multidimensional heterostructures that combine materials of various dimensions. Integrating 0D, 1D, 2D, and 3D materials enhances photodetection across an extended spectral range and significantly improves the mechanical flexibility and system integration of the devices^{[126, 127]}. As an illustrative example, Mukherjee et al. have combined 2D MoS₂ nanosheets with 0D PbS QDs to achieve enhanced optical responses that extend into the SWIR region, while still maintaining a robust visible light response for multispectral photodetection^[96]. Expanding further, by dispersing 1D V₂O₅ NWs onto 2D MoS₂ and establishing metallic contacts on MoS₂, devices can absorb light ranging from ultraviolet to NIR regions^[128]. The technique, which strategically distributes 1D materials on 2D substrates, introduces a novel approach to broaden the absorption range of photodetectors, presenting vast potential in sectors like optoelectronics, sensors, and advanced photodetection.

The optimization of heterojunction structures focuses on enhancing the intrinsic performance of materials and improving device functionality through efficient interface design and the integration of multidimensional materials. The comprehensive approach significantly elevates the photoelectric conversion efficiency and broadens the application spectrum of the devices. Such strategies are integral to the progress of cutting-edge technologies, including flexible photodetectors, playing a crucial role in pushing the envelope in fields such as environmental monitoring and health diagnostics.

Device integration strategies are essential for achieving high performance and reliability in developing flexible infrared photodetectors. The strategies, which include flexible design techniques and integration technologies, ensure both the mechanical flexibility and photoelectric performance of the devices, expanding their applicability in real-world scenarios. The foundation of deploying flexible infrared photodetectors lies in the implementation of flexible design technologies. Optimizing device architecture and material selection, the technologies, including techniques for flexible substrate, electrodes design and integrated flexible structural design, are pivotal in realizing high-performance flexible infrared detectors.

As the cornerstone of flexible design technologies, flexible substrate design is paramount for maintaining device functionality under mechanical stress. Traditional rigid substrates, such as silicon, provide excellent electrical performance but do not offer the flexibility required for applications that necessitate bending and adaptability. Therefore, the shift towards flexible materials is imperative to overcome these limitations and enhance device adaptability and practicality. Notably, Park et al. have advanced significantly in this area by employing ultra-thin, 3 µm thick parylene films as substrates. The substrates demonstrate unprecedented operational stability even under severe mechanical deformations, including bending radii smaller than 3 µm and after enduring more than 1000 bending cycles^[129]. The ultra-thin substrates significantly reduce the device's volume and weight while substantially increasing its flexibility. As a result, photodetectors can easily conform to irregular surfaces or soft biological tissues, broadening their applications in emerging fields such as wearable devices and biomedical sensing.

As the demand for wearable devices and flexible electronics continues to grow, photodetection technologies that perform well under various mechanical stresses are increasingly necessary. Wang et al. introduced ultra-flexible organic photovoltaic devices based on solution-processed TPU substrates that can be wrapped around a hair with a diameter of 50 µm (Fig. 14(a)). Fig. 14(b) illustrates the structure of a stretchable flexible infrared detector that maintains a stable dark current even at 22% strain^[130]. Thanks to its p−i−n structure, the built-in electric field enables the OPD to operate in a self-powered mode, achieving a detection rate of 1.3 × 10¹² Jones in the near-infrared region, marking the highest performance among similar devices. Furthermore, the OPD has proven capable of conducting heart rate monitoring at strains up to 30% or undergoing 800 stretching cycles at 10% strain while still maintaining excellent stability.

Currently, a variety of polymeric materials are employed as flexible substrates in the development of infrared photodetectors, including polyimide (PI)^[43], polyethylene terephthalate (PET)^[116], polyethylene naphthalate (PEN)^[131], silicone, thermoplastic polyurethane (TPU)^[130], polyetherimide (PEI)^[42], polydimethylsiloxane (PDMS)^[132], parylene^[129], polypropylene (PP)^[133], and polyvinyl chloride (PVC)^[134]. Additionally, cost-effective options such as carbon cloth and paper are also widely used^{[135, 136]}. Beyond these conventional organic flexible substrates, certain ultra-thin inorganic materials, like mica and flexible silicon wafers, demonstrate excellent flexibility and bending capabilities, making them suitable for flexible infrared sensors. Metal foils, including aluminum and copper, are particularly valued for their exceptional conductivity and flexibility, facilitating the creation of high-performance flexible electronic devices without the need for additional metal electrodes, thus streamlining the manufacturing process^[128]. Each material offers distinct advantages: PI is preferred for its superb mechanical flexibility and thermal stability, maintaining performance at temperatures up to 400 °C. PET, while cost-effective and easy to process, offers transparency but has a lower thermal resistance, generally below 150 °C. PEN, providing greater mechanical strength and thermal resistance than PET, can withstand temperatures around 200 °C, making it ideal for more demanding flexible applications. Silicone stands out for its outstanding elasticity and biocompatibility, making it suitable for wearable devices and biomedical sensors across a broad temperature range. TPU is favored for its flexibility and abrasion resistance, ideal for portable and wearable electronics.

The strategic selection and design of these materials are critical in advancing flexible infrared photodetector technologies. By optimizing the choice and configuration of flexible electrodes and substrates, it is possible to significantly enhance the mechanical flexibility and photoelectric performance of these devices. Such advancements promote the broader application of flexible infrared photodetectors in various practical settings.

Traditional electrodes, such as gold and silver along with their derivatives like thin films, nanowires, and nanoparticle networks, are widely used due to their excellent conductivity and mechanical flexibility^{[137, 138]}. Despite their advantages, these materials often struggle with mechanical stability and performance degradation under extreme mechanical stress^[139]. Addressing the shortcomings of traditional electrodes, researchers have developed an innovative flexible photodetector featuring a sandwich structure that integrates the inorganic semiconductor Ag₂S with PI substrates. In this configuration, a ductile silver layer acts as both an electrode and a buffer, strengthening the bond between Ag₂S and PI. The design significantly mitigates mechanical stress within the Ag₂S film, enhancing its flexibility. Remarkably, after undergoing more than 10 000 bending cycles, the Ag₂S/Ag/PI film maintains its conductive performance, underscoring its excellent mechanical stability and electrical durability. Additionally, this Ag₂S/Ag-based flexible photodetector demonstrates stable responses across a wide spectral range from ultraviolet to near-infrared, even under extreme bending, thus proving its suitability for various environmental applications^[140].

Moreover, transparent electrodes, crucial for flexible and stretchable applications, must maintain excellent conductivity without impeding light transmission. Traditional options like indium tin oxide (ITO) and fluorine-doped tin oxide (FTO) offer high transparency but suffer from brittleness, limiting their use in applications demanding ultra-flexibility. Conductive polymers such as PEDOT provide a durable and cost-effective alternative, facilitating large-scale application on flexible substrates through solution processing techniques, which is ideal for mass production. For instance, research by Zhu et al. has led to the development of a novel flexible broadband photodetector using a solution-processed PEDOT film treated with formamidinium iodide (FAI) as a transparent electrode. This configuration achieves high transparency across the 300 to 7000 nm range and excellent conductivity. Incorporating a dual-layer film of perovskite and PbSe QDs as the photoactive layer allows the photodetector’s spectral response to extend into the infrared region. At room temperature, the device maintains impressive photo-response performance, rapid response times, and superior flexibility, illustrating the potential of transparent polymer electrodes in next-generation flexible photodetectors and paving new paths for electrode engineering.

In scenarios where electrode materials form heterojunctions with photosensitive materials, the choice of electrode is crucial. Ti₃C₂ MXene electrodes, for example, exhibit superior transparency and conductivity compared to traditional Au electrodes. The electrodes form a van der Waals heterostructure with Te microplates, markedly optimizing NIR photodetection efficiency. Under 1064 nm laser illumination, Ti₃C₂−Te photodetectors achieve a switch ratio seven orders of magnitude higher than Au−Te-based detectors, reaching 9.51 × 10⁷ and significantly boosting photo-response performance. With 63% transparency in the 500−1000 nm wavelength range, the electrodes show great promise for wearable optoelectronic devices^[141]. Similarly, Liu and colleagues have significantly improved infrared photodetector performance by incorporating graphene electrodes known for their exceptional conductivity and transparency. Integrating graphene with n-type zinc oxide creates an effective interface barrier that substantially reduces dark current, thus enhancing the signal-to-noise ratio and the on/off switch ratio. This development underscores the crucial role of graphene electrodes in boosting the efficiency and stability of infrared detectors, marking a pivotal advancement in photodetector technology^[95].

Cost considerations are paramount in advancing the commercial viability of infrared photodetectors. One innovative approach involves utilizing graphite electrodes deposited by pencil drawing, combined with liquid-exfoliated Bi₂Se₃ nanoplates as the photosensitive material. The design, depicted in Fig. 14(c), demonstrates high photocurrent, responsiveness, and durability under 1064 nm infrared illumination. Thanks to its flexible paper substrate, the detector maintains stability even when bent, offering a cost-effective solution for manufacturing flexible infrared detectors. Future research will continue to optimize the conductive performance and mechanical stability of flexible electrodes to reduce costs and meet commercialization demands^[135].

In the evolving field of flexible infrared sensors, integrating flexible substrates and electrode technologies has partially enhanced the wearability and flexibility of devices. However, these technologies alone often fall short in applications requiring adaptation to dynamic and complex shapes. Therefore, comprehensive flexible structural design strategies are pivotal for fully realizing device flexibility to effectively address complex application environments. This encompasses both flexible optoelectronic devices and stretchable optoelectronic devices.

For flexible optoelectronic devices, it's crucial to account for various mechanical stresses encountered during actual usage while ensuring the stability of electronic and optoelectronic functions. The bending stiffness (D) of a device, which significantly affects its flexibility, can be calculated using the equation^[142]:

10D=Et312(1−v2),

where E represents Young's modulus, t is the thickness of the device, and v is the material's Poisson's ratio. The formula illustrates that reducing the thickness of the device significantly decreases its bending stiffness, thereby enhancing its flexibility. The adjustment is vital for applications that require sensors to conform closely to varied surfaces without compromising functionality. This involves optimizing flexibility and durability by managing the device's bending stiffness and Young's modulus. Since the application scenarios for flexible infrared photodetectors may involve complex surfaces and dynamic environments, excessively thin films are prone to damage due to stress. Therefore, the final thickness of flexible optoelectronic devices must be carefully considered to facilitate their integration into flexible infrared detection systems.

In the design of flexible and smart infrared sensors, the selection of materials with low Young's moduli, such as flexible polymers, plays a crucial role. Specifically, the formula for Young's modulus is:

11E=σ(ϵ)ϵ=F/A0ΔL/L0.

It explains the mechanical dynamics where F is the applied force, A₀ is the initial cross-sectional area, and ΔL and L₀ are the change in length and the original length of the device, respectively. Using materials with low stiffness ensures that the sensors remain flexible and functional under mechanical stress, adapting seamlessly to the contours of the various surfaces without significant damage. In the realm of flexible infrared sensors, ensuring the right material properties is paramount. A Young's modulus that is too low can leave sensors overly soft, thereby undermining their mechanical stability and reducing their lifespan. It is vital, therefore, to establish a balance between flexibility and mechanical strength during the design process. This balance ensures that the sensors are not only effective in their function but also durable and reliable over time. By optimizing material properties and sensor design, developers can create devices that meet both performance and longevity standards crucial for advanced applications.

In situations where repeated mechanical movements cause flexible device failure, stretchable devices that adjust to dynamic changes significantly improve user experience^[143]. Various strategies have been developed to reduce stress caused by mechanical deformation in stretchable optoelectronic devices, categorized by the hardness of the materials used^{[144, 145]}. Strategies employing hard materials include island−bridge configurations and kirigami-inspired designs^[146], which utilize the high performance of ultrathin hard materials combined with geometric engineering optimization^[147−150].

Kim and colleagues have significantly enhanced the performance of stretchable devices by introducing island−bridge structures combined with single-crystalline silicon nanomaterials and non-coplanar mesh designs. These designs allow circuits to withstand up to 140% linear stretching while maintaining high electron mobility, mechanical reliability, and electrical performance, opening new possibilities for biomedical devices and wearable technologies^[151]. Lee et al. further developed this concept by employing island−bridge structures with elastomeric substrates featuring surface relief structures, enabling stable operation under various reversible deformation modes such as bending, twisting, stretching, or compressing. The approach not only ensures continuous functionality under extreme strain but also allows for greater flexibility in designing devices with high area coverage and excellent stretchability^[152]. Moreover, Tang and colleagues have designed reconfigurable materials through kirigami structures, achieving expansive area enhancement and mechanical property improvement. Their design, through hierarchical cutting and hinge optimization, significantly strengthens the material's tensile resistance and expandability, making it suitable for complex mechanical deformations in wearable electronic devices and offering new design strategies for stretchable electronic devices^[153].

The soft material strategy utilizes intrinsically stretchable materials, facilitating mechanical property adjustments suitable for developing stretchable optoelectronic devices. The pre-stretching technique involves pre-stretching the stretchable substrate before depositing conductive materials, enhancing the materials' stretch tolerance in actual applications. However, in intrinsically stretchable devices, mismatches in mechanical stiffness between materials, layers, and device units pose major challenges, with interface control determining device characteristics and stretchability levels. Gu et al. have successfully developed multi-layered gradient-assembled polyurethane (GAP) conductors incorporating gold nanoparticles. This structure not only enhances conductivity under extreme mechanical strains (up to 300%) but also supports the development of high-performance stretchable lithium-ion batteries. Furthermore, the hierarchical nano-composite structure of GAP accelerates the rapid manufacturing of advanced energy storage devices, demonstrating great potential in adjusting mechanical properties for biomedical optoelectronic device development. Through innovative multi-layered interface design, not only is the performance and durability of stretchable electronic devices enhanced, but new possibilities are also opened for the further development of wearable technologies and biomedical devices^[154].

At the macroscopic level, the focus in developing stretchable infrared sensors is on the comprehensive integration of device components and electrical connections. The design incorporates interconnect structures capable of withstanding high mechanical strains, optimizing the interface between individual devices and their substrates. This advanced multi-level interface design significantly enhances the performance and durability of stretchable electronic devices, paving the way for novel applications in wearable technology and biomedical devices. Recently, the adoption of low-melting-point liquid metals has become prevalent due to their exceptional stretchability. These metals effectively resolve the modulus mismatch between rigid components and flexible substrates through their inherent fluidity and malleability. Under pressure, they naturally deform to fill gaps, achieving seamless mechanical bonding. Their high conductivity ensures that electronic functions are maintained effectively, even when substrates are bent or stretched. This property is particularly crucial for wearable and bioelectronic devices that demand high flexibility and stable electrical connections. In a notable study, Zhuang et al. developed an innovative three-dimensional integrated electronic skin (P3D-eskin) that merges high-density inorganic electronic components with a stretchable organic fiber substrate, SBS. This electronic skin utilizes multi-layer EGaIn liquid metal circuits and hybrid liquid metal solder to create a soft, stretchable, and breathable structure that diverges significantly from traditional rigid circuit boards. It supports continuous system-level functions such as data acquisition, signal processing, intervention, and wireless communication. Additionally, its high breathability and moisture permeability effectively prevent skin inflammation during prolonged contact, maintaining stable electrical connections even under extreme strains (up to 1500%). In comparison to PDMS-based electronic skin, P3D-eskin shows a reduction in thickness and rigidity by approximately 54% and 60%, respectively, demonstrating enhanced system-level integration and functionality^[155].

The above design strategies from flexible and stretchable electronics are equally applicable to infrared photodetectors. Integrated flexible structural designs not only enhance the flexibility of these devices but also ensure performance stability and extended operational life through precise adjustments in material bending stiffness and Young's modulus. These universal design principles provide effective guidance for infrared photodetectors, enabling them to maintain functional integrity while adapting to the challenges of complex operational environments.

System-level integration technologies, encompassing 3D and multifunctional integration, significantly enhance both the diversity and performance of devices, setting trends for next-generation high-performance flexible electronics. The two complementary technologies optimize spatial layouts, increase the density of sensors, and expand their functional scope. The combined effect of the technologies boosts device performance and versatility, shaping the developmental trends of advanced flexible electronics.

3D integration transcends the constraints of traditional in-plane integration, which is limited by physical and technological challenges, by effectively utilizing vertical space. The approach is pivotal for the development of flexible infrared photodetectors, enabling the vertical stacking and interconnection of multiple electronic components^[156]. Fig. 15(a) illustrates this architecture, which not only improves the integration of photodetector units with signal amplification and processing units, essential for advanced functionalities but also addresses the challenges inherent in traditional 3D heterogeneous integration. Such integration often involves complex wafer fabrication and bonding processes that require intricate procedures like through-silicon-vias and solder ball bonding, which can complicate and limit chip integration due to their complexity^[157]. The research by Kang et al. presents a advanced approach to the 3D integration for 2D materials, initially intended for artificial intelligence applications, yet profoundly relevant to the development of flexible infrared sensors. Their utilization of 2D transition metal dichalcogenides (TMDCs) via bottom-up synthesis into a compact, multi-layered structure (as depicted in Fig. 15(b)) highlights a technique that significantly boosts data processing efficiency. This approach reduces processing times, lowers voltage loss, and shrinks device footprint, thus offering a valuable model for the design of next-generation flexible infrared sensors. It exemplifies how methodologies devised for one specific area can cross-fertilize and drive significant technological progress in other areas, thereby facilitating advanced three-dimensional integration within sensor technology^[158].

The utilization of 3D integration extends beyond enhancing pixel density, enabling innovative applications in photodetector design. Utilizing Ⅲ−Ⅴ semiconductors, Wang et al. formed multilayer structures on GaAs substrates using MOCVD technology. The process included AlAs sacrificial layers and quantum well (QW) functional layers. In manufacturing, etching the AlAs layer released stress, causing planar nanofilms to spontaneously curl into tubular structures, thus enhancing the photoresponse of QW infrared photodetector without external optical coupling. This innovative structure optimizes light coupling and photoelectric conversion efficiency, achieving omnidirectional detection and broad coupling efficiency over wide incident angles. By adjusting the number of coils in curled QW infrared photodetector, precise control over photocurrent and responsiveness is achieved, thus enhancing device performance. The example highlights the potential of 3D self-assembly technology in improving nanofilm functionality and photoelectric conversion efficiency, offering new pathways for the development of advanced photodetectors and related devices^[159].

The 3D integration technologies substantially increase the functional density and performance of flexible infrared photodetectors. The method not only achieves device miniaturization but also broadens the scope for deploying photodetectors in diverse applications, thereby driving the technological advancement of high-performance flexible electronic devices. Additionally, 3D integration establishes a robust basis for advancing multifunctional integration technologies.

Benefiting from the advancements in 3D integration, multifunctional integration technologies enable the incorporation of various sensors and functional modules within a single system. The integration significantly broadens the application spectrum of flexible infrared photodetectors, especially in sectors such as medical diagnostics, environmental monitoring, and smart manufacturing. For example, the simultaneous monitoring of multiple physiological or environmental parameters significantly improves device efficiency and intelligence. In medical applications, integrating temperature sensors, heart rate monitors, and blood oxygen saturation sensors into a single flexible device simplifies the architecture while enhancing the accuracy and timeliness of data collection.

Zhu et al. have pioneered a skin-like NIR-Ⅱ photodetector and array featuring a simple bilayer structure composed of highly conductive MXene (Ti₃C₂Tx) and photosensitive PbS QDs. The configuration effectively promotes the separation of photogenerated carriers at the bilayer interface and ensures efficient transport between spatially separated layers. The robust interfacial bonding between the MXene and PbS QDs significantly improves both the photoelectric performance and mechanical stability of the device, demonstrating a high responsivity (1000 mA/W), rapid light response, and excellent mechanical stability, with performance retention exceeding 95% after 500 bending cycles. The device's design not only caters to the fabrication of various practical high-performance flexible wireless photodetection systems but also supports comprehensive applications such as optical communication, NIR-Ⅱ imaging, and proximity sensing^[160].

Despite the promising prospects of multifunctional integration technologies, challenges persist in integrating multiple sensors on a flexible substrate without interference while maintaining high sensitivity and rapid response capabilities. Shen et al. developed a flexible 1T1R (one transistor-one resistor) photodetector based on a 2D Sb₂Se₃ film, covering a broad detection range from visible light to NIR. Employing low-temperature PECVD, the 2D Sb₂Se₃ film was directly deposited in the sensing area to achieve a flexible 1T1R structure. The design maintains high stability and functionality under various bending states, significantly enhancing the photoresponse compared to traditional 1R structures. This demonstrates the potential of multifunctional integration technologies in non-volatile memory and integrated computing devices, which often employ the 1T1R structure^[161].

Furthermore, traditional full-color photodetectors, which rely on complex color filters and interferometric optical components, face limitations that hinder their widespread application. To overcome these barriers, Kim et al. developed a novel 2D pixelized full-color photodetector that integrates different-sized colloidal QDs (e.g., PbS, CdSe, CdS) with amorphous indium gallium zinc oxide semiconductors at temperatures below 150 °C using a monolithic integration technique (Fig. 15(c)). The approach covers a wide spectral range from ultraviolet to infrared and introduces chelating chalcogenide metal ligands to reduce trap states and enhance charge carrier transport efficiency. The resulting full-color photodetector exhibits extremely high photoresponsivity across a broad wavelength range, capable of accurately identifying and differentiating light sources of various wavelengths, and supports phototransistor arrays on flexible platforms suitable for broad-spectral imaging sensors and human-centric bio-devices^[162].

As multifunctional integration technologies continue to evolve, future research should focus on optimizing interconnection technologies and managing interfaces to enhance signal decoupling and prevent interference. These improvements are crucial for developing wearable health monitors and advanced environmental sensing networks, and for expanding smart sensors into new application domains^[163].

Flexible infrared photodetectors have evolved from focusing solely on physical flexibility to meeting specific functional requirements. Innovations in materials, manufacturing, and design have enabled the expansion of these detectors into new applications such as wearable health monitoring and biomimetic curved imaging, while also suggesting potential for further technological advancements and broader applications.

Wearable devices and health monitoring present unique demands for flexible infrared photodetectors, which are ideally suited for real-time physiological parameter monitoring due to their lightweight, flexibility, and low power consumption. These detectors utilize non-invasive, high-sensitivity, and broad spectral response capabilities to monitor key physiological parameters such as heart rate, respiratory rate, and body temperature. When integrated into smartwatches and medical patches, these sensors deliver accurate physiological data across various activity states, ensuring conformity to human contours for a comfortable wearing experience^[164−166].

A cornerstone technology, photoplethysmography (PPG), employs optical methods to detect volumetric changes in blood flow within arteries over time, extracting critical parameters like heart rate, heart rate variability, and blood oxygen saturation. PPG signals are captured by illuminating the subcutaneous tissue with light at green, red, or NIR wavelengths and detecting the transmitted or reflected light with photodetectors. The sensors convert light absorption changes from blood volume fluctuations into electrical signals to produce PPG waveforms and extract physiological parameters. In the field of wearable health monitoring, Liang et al. have developed a new large-area flexible CQD photodiode that significantly enhances photodetector performance. Incorporating PI into the CQD ink enhanced the passivation, durability, and morphological quality of the ink, resulting in a photodiode with low dark current, high uniformity, stable performance, and detectivity surpassing 10¹³ Jones. The technology's implementation in wearable PPG signal measurement under ambient light conditions using lower power consumption and cost has demonstrated accurate heart rate detection under everyday lighting conditions, validating its practicality and efficiency for real-time health monitoring (Fig. 16(a))^[89].

The demand for remote measurement of vital signs such as heart rate and respiratory rate presents significant challenges in non-invasive health monitoring. The challenges necessitate large-field, easily integrable, and inconspicuous sensors, such as large-area thin-film photodiodes, which must distinguish faint light signals from background interference over long distances. Ollearo et al. have developed a solution-processed film photodiode based on a tandem perovskite-organic structure specifically designed for remote non-invasive health monitoring. The device's extremely high NIR responsivity and superior optical noise filtering capability allow the device to effectively distinguish faint light signals from background noise, particularly suited for long-distance monitoring of heart rate and respiratory rate. The active layer, produced by spin-coating and paired with a transparent ITO/poly[di-(4-phenyl)(2,4,6-trimethylphenyl)amine] front electrode and a reflective back electrode, demonstrates high stability (over 8 h) and exceptional quantum efficiency (over 200%), significantly surpassing traditional silicon-based sensors. The tandem photodiode configuration, similar to a solar cell but with lower dark currents and a broader dynamic range, maximizes patient comfort and ensures the accuracy of monitoring data in hospital beds and other scenarios through wireless methods (Fig. 16(b))^[167].

Liu and colleagues have expanded the scope of wearable health monitoring devices to include mid and far-infrared spectra through the development of a flexible photodetector utilizing the heat-effect. The device combines monolayer graphene with heterogeneous metal electrodes, achieving an ultra-broadband response that spans from ultraviolet to millimeter waves. Designed for flexibility, it conforms well to human skin, making it suitable for extended wear and continuous monitoring. Additionally, this photodetector simplifies the traditionally complex integration process required for terahertz and millimeter wave detectors, enhancing its practicality for wearable technologies. In field tests, the device successfully conformed to a simulated human wrist, demonstrating its capability to image shielded objects and various materials across a wide spectrum. The successful deployment of this technology in practical scenarios underscores its potential to transform health monitoring and personal medical devices, aligning with the ongoing trend of integrating advanced technology into wearable health solutions^{[168, 169]}.

The integration of cutting-edge technologies into wearable health solutions is transforming the field of health monitoring and personal medical devices. Flexible infrared photodetectors have evolved beyond basic blood oxygen saturation monitoring to become essential components in smart apparel and electronic skin. These detectors enhance the real-time tracking of vital signs such as heart rate and body temperature. Advances in material science and nanotechnology are improving the functionality of these detectors, broadening their use from routine health tracking to full-scale medical diagnostics. Such developments are poised to significantly enhance health insights and improve patient care.

Human eyes, a highly optimized optical systems, capture and process light signals through their spherical globes and curved retinas^{[170, 171]}. The naturally evolved structure enables high-resolution focus across a wide viewing angle while minimizing image distortion and enhancing edge clarity. Inspired by the natural visual systems, biomimetic curved imaging technology aims to replicate these capabilities, achieving more natural viewpoints and improved imaging performance, especially in applications requiring expansive fields of view and high image quality^[172]. Curved detectors, in contrast to traditional flat photodetectors, better mimic the curvature of the human eye, thus reducing optical distortions and enhancing image quality^[173].

Significant advancements in flexible infrared curved imaging sensors have been made, with pioneering work by Tang et al. demonstrating the use of flexible CQD photovoltaic detectors. The detectors, made from HgTe CQDs, possess excellent mechanical flexibility and rapid response capabilities, crucial for effective curved imaging. Integration of Fabry−Perot resonators within the detectors has further enhanced light absorption capabilities, achieving a peak detection rate of 7.5 × 10¹⁰ Jones at room temperature, and presenting new prospects for developing high-resolution infrared electronic eyes^[173]. Addressing performance limitations in biomimetic infrared sensors, Ran et al. have developed a revolutionary biomimetic infrared detection amplification system that significantly enhances infrared imaging performance, especially in terms of light sensitivity and image contrast (Fig. 16(c)). The approach uses NW field-effect transistors to amplify the output signals of infrared detectors, achieving a light sensitivity of 7.6 × 10⁴ at a wavelength of 1342 nm, substantially surpassing traditional systems. The incorporation of artificial neural networks (ANNs) has optimized the image processing capabilities of the IRDA system, improving image contrast and recognition efficiency. This enhanced sensitivity and contrast imaging capability show great potential for applications that mimic natural visual systems^[93]. Recent innovations have led to significant advancements in infrared curved sensor technology, exemplified by the work of Ding et al. who have developed a biomimetic hemispherical infrared imaging device. The device is inspired by the pit organs of snakes and utilizes high-density ionic thermoelectric polymer nanowire arrays that function similarly to nerve cells, responding significantly to temperature changes. It features a 625-pixel array embedded on a hemispherical substrate, achieving an ultrawide field of view of up to 135°, which exceeds the natural capabilities of pit organs. Remarkably, this sensor operates effectively at room temperature without the need for external power sources or cooling systems. This development not only demonstrates a major leap forward in infrared sensor technology but also establishes a strong foundation for the future enhancement of bioinspired infrared imaging devices^[172].

Furthermore, techniques such as ultrathin design, origami/kirigami design, island−bridge structure, fractal web structure, and in situ growth of nanowires are common strategies employed for achieving curved surfaces in biomimetic imaging systems^[174]. The ultrathin design approach significantly enhances device flexibility and adaptability, allowing direct adhesion to complex curved structures such as concave hemispheres (Fig. 17(a))^[150]. Origami and kirigami designs transform flat materials into multifunctional 3D structures without breaking, enhancing spatial adaptability (Fig. 17(b))^[175]. The island−bridge structure design, featuring serpentine metallic pathways, absorbs mechanical stress, preventing structural damage and maintaining high electrical and mechanical performance under stress (Fig. 17(c))^[176]. The fractal web structure, inspired by natural mesh structures like spider webs, achieves uniform stress distribution and exceptional mechanical stretchability (Fig. 17(d))^[177]. in situ growth of nanowires on hemispherical substrates maintains a monocrystalline structure and enables direct adaptation to curved surfaces, offering high-density and excellent optoelectronic performance (Fig. 17(e))^[178].

Biomimetic vision sensor based on flexible and curved photodetector arrays with low aberration, wide FOV, and potentially reduced size and weight, offering substantial advances in the field of robotics and machine vision. Future research will focus on further optimizing these technologies, enhancing signal decoupling, and preventing interference to address the complexity and stability challenges introduced by the added functional modules. These improvements will facilitate the development of wearable health monitors and advanced environmental sensing networks, expanding smart technologies into new application areas.

Flexible optoelectronic sensors are vital in dynamic environments, enduring various stresses to maintain function. High bending stresses can break the photosensitive layer, reducing light absorption. Damage to the transport layer and interfaces impairs carrier transport and charge collection, while even slight deformations may worsen phonon scattering and defect mechanisms, complicating carrier transport control and dark current suppression. Therefore, selecting and optimizing materials with excellent photoelectric properties and flexibility is crucial. The reliance on low-temperature processes for flexible substrates demands balancing photoelectric properties with defect management. Defects like irregular atomic configurations and high defect state densities severely affect properties such as carrier mobility and lifespan, challenging power reduction and control over photoconductive effects. Moreover, integration strategies face substantial hurdles in material compatibility, interface quality, and 3D stacking, requiring meticulous design of isolation layers and shielding to ensure system reliability. Additionally, multifunctional sensor integration tackles signal decoupling and noise management, using advanced processing to maintain accuracy and stability. Overall, developing high-performance flexible infrared sensors involves a systematic integration of material selection, fabrication, and integration strategies to effectively navigate these technological challenges.

Comprehensive optimization towards SWaP³ is crucial for the next generation optoelectronic sensors. While flexible infrared photodetectors have shown potential across various applications, performance constraints—especially critical metrics like dark current, signal-to-noise ratio, response speed, and their integration with flexible ROIC—remain to be addressed^[179]. These limitations currently hinder their deployment in vital sectors such as security surveillance, medical imaging, and environmental monitoring, making the push towards ultra-high performance a key research direction.

First, the overall performance of flexible infrared detectors based on narrow bandgap materials are to be improved. These materials enhance optical absorption capabilities but also pose challenges due to their lower bandgaps, resulting in higher dark currents. In dark environments, these materials facilitate thermal excitation of carriers from the valence to the conduction band, generating undesired currents. The dark current I_d can be described by the following equation：

12Id=A*T2e−Eg2kT,

where E_g represents the bandgap energy, T is the absolute temperature, k is the Boltzmann constant, and A is a constant dependent on material and device architecture. The equation quantitatively illustrates how dark current increases with decreasing bandgap energy and rising temperature, offering a theoretical basis for material selection and device architecture optimization to mitigate dark current.

To delve deeper into carrier dynamics, the thermionic emission equation further elucidates how carriers are thermally excited to the conduction band under narrow bandgap conditions:

13J=A*T2e−ΦBkT.

Here, Φ_B represents the barrier height. The equation is pivotal not only for enhancing the understanding of dark current mechanisms but also for dissecting the influence of material properties on carrier dynamics. Detailed analysis of these parameters allows for the optimization of infrared detector design.

Moreover, narrow bandgap materials are more susceptible to thermal excitation mechanisms, leading to higher dark currents. The increase in dark current is closely associated with non-radiative recombination processes, which are particularly significant in narrow bandgap semiconductors. Non-radiative recombination occurs when carriers recombine without photon emission, influenced primarily by internal defects such as phonons (lattice vibrations), thermionic effects, and other lattice structural defects (e.g., dislocations and impurities). These defects act as trap centers, capturing carriers and thus increasing the probability of non-radiative recombination, which in turn elevates dark current. Given the narrower band structure, electrons are more susceptible to phonon interactions as they can participate in scattering processes at lower energy states. Furthermore, the higher carrier generation rate in narrow bandgap materials makes them more sensitive to internal defects. The heat generated by non-radiative recombination raises device temperatures, thereby increasing the number of thermally excited carriers and further enhancing dark current. To mitigate the effects of poor carrier separation caused by non-radiative recombination, an increase in external electric field strength is necessary, although this also leads to an increase in dark current.

Cooling operations form a key optimization strategy for photonic-type infrared detectors based on narrow bandgap materials. By lowering operational temperatures, the temperature-dependent equation for dark current significantly reduces the number of thermally excited carriers, effectively reducing dark current. The measure not only decreases thermally excited carriers but also improves performance by reducing the non-radiative recombination rate described by SRH theory. Additionally, cooling helps alleviate non-radiative recombination caused by phonons and internal defects, further controlling dark current. Lower temperature operations also enhance the signal-to-noise ratio and response speed by increasing carrier mobility and reducing noise levels, thereby optimizing detector performance in low-light conditions and enhancing their potential for high-performance applications such as security surveillance, medical imaging, and environmental monitoring.

In addition to cooling strategies, the integration of multilayer heterostructure designs and high-purity semiconductor materials, combined with high-temperature annealing and purification processes, effectively controls carrier movement in dark conditions, reduces the incidence of defects and impurities, and minimizes non-radiative recombination and trap release, thereby comprehensively enhancing the performance of flexible infrared detectors. For instance, the p−i−n photodiode structure plays a critical role in controlling dark current through the strategic setting of low doping concentrations in the intrinsic (i) region. The design effectively broadens the depletion zone, significantly reducing dark current. In this configuration, the p-type and n-type semiconductor layers are typically crafted from wide bandgap materials, serving as barriers to prevent carrier diffusion into the absorptive layer, thereby not contributing to the dark current. When an appropriate reverse bias is applied, the low-doped i-region can be fully depleted, essentially devoid of free carriers, which substantially lowers dark current. Additionally, dark current in p−i−n structures displays a marked temperature dependency, increasing with temperature as intrinsic carrier concentration rises. The characteristic allows p−i−n structures to maintain low dark currents while delivering high-performance photo-detection, making them particularly suitable for precision optical detection and imaging systems that require high sensitivity and low power consumption. For a p−i−n photodiode, the dark current density can be defined by the following formula^[23]:

14Jdep=qnitdepτSRH.

In the expression, t_dep denotes the width of the depletion zone, τ_SRH represents the Shockley−Read−Hall lifetime. The dark current associated with this region shows a dependency on temperature, as indicated by the intrinsic carrier concentration n_i​.

The substantial breakthroughs achieved in optimizing the SWaP³ of flexible infrared sensors have not only enhanced sensor performance but also expanded their potential for high-performance applications, meeting the modern technological demand for efficient, low-power, and highly adaptable devices. Size and weight optimizations directly enhance the portability and applicability of the sensors, necessitating the development of new micro- and nano-manufacturing techniques to achieve miniaturization and lightweight design. This involves synthesizing and processing new materials, and precisely controlling and integrating nano-scale structures. Cost control is critical for the widespread adoption of flexible infrared sensor technology, requiring efforts at every stage from material selection and manufacturing processes to device integration, exploring cost-effective alternative materials, and simplifying production processes. Power consumption optimization is crucial for enhancing system efficiency and extending lifespan, making the development of intelligent power management and adaptive energy demand control technologies an effective strategy for reducing power usage. Despite these advancements, current technology and material development levels have not yet fully met the high standards for commercialization, making achieving a comprehensive commercial breakthrough a long-term goal. This requires researchers to conduct in-depth studies and interdisciplinary collaborations in fields such as materials science, nanotechnology, optoelectronics, and integrated circuit design.

Moreover, with the rapid increase in pixel density and the integration of multifunctional sensors within flexible infrared sensor arrays, there is a surge in data processing demands. Traditional processing methods, reliant on frequent data transmission, result in low energy efficiency due to the segregated operation of sensing, processing, and storage units in the von Neumann architecture. The architecture necessitates constant data retrieval and writing during processing, causing delays and high energy consumption. In this context, data must be transmitted between sensors and processors, especially in large-scale multisensory systems, making this approach inefficient and slow. The future direction involves smart sensor arrays capable of efficient information recognition, processing, and feedback. Recent breakthroughs in smart optoelectronic sensors spanning from ultraviolet to infrared have been significant, although much focus remains on rigid substrate-based visible light sensors. However, the efficiency and intelligence of flexible infrared sensors are compelling, showing converging development trends.

The smart sensors fall into two categories: in-memory computing photonic sensors and synaptic photonic sensors^[180]. The development of in-memory computing sensors represents a trend towards more integrated and intelligent sensor computation^{[181, 182]}. In these architectures, low-level sensory processing focuses on enhancing data quality through noise filtering and signal strengthening, typically achieved through simple logic or small-scale computing units integrated closely with the sensors. In contrast, high-level sensory processing handles more complex tasks like pattern recognition and machine learning, necessitating higher computational power and possibly more complex computing units adjacent to the sensors. These sensors process data internally, reducing reliance on central processing units. The integrated processing not only improves immediacy in data handling but also reduces power consumption and latency by minimizing long-distance data transmission, crucial for applications requiring real-time responses, such as autonomous driving and wearable technologies. In addition, this technology facilitates high integration of sensors and processing units through advanced manufacturing techniques, potentially enhancing both device performance and cost-efficiency. Employing 3D integration and multifunctional integration technologies, sensors and computing units can be closely integrated at the micro to nanoscale, significantly reducing device size and price while maintaining high performance^[183−185].

Another innovative type is the synaptic sensor, which merges optoelectronic technology with synaptic dynamics to mimic biological synapses^[186−189]. Particularly suited for flexible infrared sensing technologies due to their ultra-low power consumption and intelligent sensing capabilities, these sensors utilize the photoelectric properties of materials and conductance changes to emulate synaptic behaviors in response to light signals. Their design’s plasticity and adaptability, akin to biological systems, are achieved by modulating conductance to replicate synaptic strength variations. These sensors feed their electrical output into neural network models, like deep learning networks, dynamically adjusting synaptic weights to support adaptive learning and implement neuroplasticity mechanisms such as short-term plasticity and long-term potentiation or depression. Such features enable artificial intelligence to improve predictions and responses, enhancing the accuracy of health monitoring and environmental sensing.

For applications involving flexible infrared sensors, the low power characteristics of synaptic photonic sensors are especially crucial as they often operate in energy-constrained environments over extended periods. These sensors perform initial data processing at the hardware level without frequent data transmission or complex processing steps, significantly reducing energy consumption. By directly filtering and preprocessing information at the sensor level, they reduce reliance on traditional digital signal processing, further lowering overall system power consumption. For instance, synaptic infrared photonic detectors can rapidly analyze fluctuations in heart rate and breathing patterns, detecting anomalies quickly and providing advanced health management services.

Moreover, by integrating computing tasks directly within sensors, redundant data transmission is significantly reduced, optimizing energy management and processing speed, and addressing data handling and energy consumption issues in existing sensor networks. This approach not only enhances efficiency but also fosters overall system performance optimization, broadening the application prospects of flexible infrared sensors in medical, environmental monitoring, and wearable device sectors.

In conclusion, the advancement of flexible infrared optoelectronic sensors will depend on the fusion of multidisciplinary technologies, including microelectronics, nanotechnology, and material science, to achieve higher integration and low power consumption. Future research will need to address how to further reduce device size and cost while maintaining high performance, as well as how to effectively manage thermal performance and energy consumption challenges.

Flexible infrared sensors are pivotal for narrowing the boundaries in realizing next-generation technologies of wearable health monitoring and biomimetic vision. We conclude the latest research advancements in enhancing the performance and broadening the application horizons of these devices. We have examined critical design parameters and architectures that influence the functionality of infrared detector arrays, as well as a wide array of novel materials and sophisticated fabrication techniques designated for flexible applications. Additionally, we have dissected the roles of innovative material designs, the integration of 3D and multifunctional technologies, and the expanded utility of flexible infrared sensors. Looking ahead, the review forecasts the evolution of flexible infrared photodetectors toward ultra-high-performance and intelligently integrated systems, underscoring their transformative potential across various fields including robotics, autonomous driving, medication, military, etc.