Advances in Nanoparticle-Based Fabrication Techniques for Infrared Detectors: A Comprehensive Review

Abstract

The field of infrared (IR) photodetection is undergoing rapid development through the emergence of solution-processable nanoparticle (NP)-based materials and fabrication strategies. This review critically examines recent advances in fabrication approaches for NP-based IR detectors, emphasizing the relationship between synthesis, surface engineering, deposition processes, and device architecture in determining detector performance. Representative material platforms are discussed, including colloidal quantum dots (CQDs) such as PbS and HgTe, which enable tunable operation from the near-infrared (NIR) and short-wave infrared (SWIR) to selected mid-wave (MWIR), long-wave (LWIR), and emerging very-long-wave infrared (VLWIR) regimes depending on material composition and operating conditions. Further platforms including plasmonic metal NPs, black phosphorus, and topological nanomaterials are evaluated for their unique mechanisms of optical enhancement and broadband response. Fabrication approaches including continuous-flow synthesis, ligand exchange, blade coating, inkjet printing, electrophoretic deposition, and other scalable solution-processing methods are analyzed with respect to their influence on film quality, charge transport, interface engineering, and integration compatibility. The review further compares major device architectures, including photoconductors, photodiodes, plasmonic absorbers, and phototransistors, using key performance metrics such as specific detectivity (D*), responsivity (R), response speed, and operating temperature, while emphasizing the importance of measurement conditions in cross-platform comparisons. Critical challenges including dark-current generation, 1/f noise, transport limitations associated with ligand chemistry, environmental instability of narrow-bandgap materials, manufacturability constraints, and toxicity considerations are also discussed. Emerging directions such as neuromorphic sensing, CMOS-compatible integration, and sustainable lead-free nanomaterials are highlighted. By linking nanoscale material design and fabrication processes to device-level performance, this review provides a framework for advancing NP-based IR technologies toward scalable and application-relevant sensing systems.

1. Introduction

1.1. Overview of Infrared Photodetectors Based on Nanomaterials

Infrared detectors transform IR photons into measurable electrical signals. Traditional technologies, such as those that use mercury cadmium telluride (MCT) or III–V semiconductor heterostructures, provide high performance but face challenges related to high costs, complex fabrication processes (often necessitating molecular beam epitaxy), limited spectral tunability, and the requirement for cryogenic cooling [1,2,3]. In addition to costs, the inherent material characteristics of these solid, bulk semiconductors restrict their compatibility with flexible electronics and silicon-based readout circuits, leading to a need for approaches on new materials, specifically polymers [4] or NP based systems [5]. As illustrated in Figure 1, NP-based IR detectors offer significant advantages over traditional technologies, particularly in terms of solution-processability, tunable absorption, and compatibility with flexible substrates.

Among the conventional designs, metal–semiconductor–metal (MSM) photodetectors have been widely studied due to their simple planar structure, low intrinsic capacitance, and appropriateness for high-speed applications [6,7]. In a typical MSM device, interdigitated metal electrodes are placed on the semiconductor surface, forming back-to-back Schottky contacts. While they offer advantages in terms of ease of fabrication and speed, traditional MSM detectors often suffer from high dark currents (I_dark) caused by Schottky barrier reduction and thermionic emission, which can limit their detectivity, especially in the IR range [8,9]. The I_dark in an MSM detector is described by Equation (1).

I_dark∝A^*T²exp(−qϕB/Κ_BT)

(1)

where A* is the Richardson constant, T is the temperature, and ϕB is the Schottky barrier height, q is the elementary charge and Κ_B is the Boltzmann constant.

The operating principle of IR photodetectors is based on converting incident IR photons into measurable electrical signals through photon absorption, charge generation, separation, transport, and collection at the electrodes [8,9,10]. Depending on the detector architecture, this process may occur through photovoltaic, photoconductive, photogating, hot-electron, or photothermal mechanisms. In photovoltaic devices such as p–n and p–i–n photodiodes, absorbed photons generate electron–hole pairs that are separated by an internal electric field, producing photocurrent with low noise and fast response [11]. In photoconductors and phototransistors, illumination modulates the conductivity of the active layer under an applied bias, often enabling internal gain but introducing longer carrier lifetimes and increased 1/f noise [12,13,14]. In plasmonic NP systems, localized surface plasmon resonance (LSPR) generates energetic carriers that can overcome Schottky barriers and contribute to photocurrent generation [15].

For NP-based detectors, these processes are strongly influenced by quantum confinement, surface chemistry, ligand coupling, and interparticle charge transport [16,17,18]. Unlike bulk semiconductors, where carrier motion occurs in continuous bands, NP-based solids consist of electronically coupled nanostructures in which charge transport proceeds through hopping or tunneling between neighboring particles. Consequently, detector performance depends not only on intrinsic material properties but also on synthesis conditions, surface passivation, and film assembly methods [19,20,21].

Engineering of NPs provides direct methods to adjust ϕB and decrease A* through manipulation of surface dipoles [1,10]. The utilization of NPs represents a considerable transformation, not only in materials but also in the re-evaluation of device architectures such as the MSM design. By employing quantum confinement in semiconductor NPs, for instance, colloidal quantum dots (CQDs), or localized surface plasmon resonance (LSPR) in metal NPs, the absorption spectrum can be precisely customized across the short-wave (SWIR), mid-wave (MWIR), and long-wave IR (LWIR) regions [11,12,13,14]. The bandgap energy (E_g) of a CQD is adjustable according to the Brus Equation (2).

E_g^(QD) = E_g^(bulk) + h²/8r_(QD)²(1/m_e^* + 1/m_h^*) − 1.786e²/ϵ_rϵ₀r_(QD)

(2)

where E_g^(QD) is the band gap energy of the NP. E_g^(bulk) is the band gap energy of the same material in bulk form. h is Planck’s constant, r_(QD) is the radius of the spherical quantum dot, m_e^* is the effective mass of the excited electron, and m_h^* is the effective mass of the electron hole. e is the elementary charge of an electron, ϵ_r is the relative permittivity (dielectric constant) of the semiconductor material, and ϵ₀ is the vacuum permittivity.

The photon detection process in NP films can be summarized in four sequential stages (Figure 2): (i) photon absorption and exciton generation within the NP, (ii) exciton dissociation into free carriers, (iii) carrier transport across the NP network, and (iv) charge extraction at the contacts. Each stage contributes to overall detector performance metrics including responsivity (R), specific detectivity (D*), external quantum efficiency (EQE), response speed, and noise-equivalent power (NEP) [22,23,24]. Carrier transport efficiency is particularly sensitive to ligand length and trap density because long insulating ligands improve colloidal stability but reduce carrier mobility, whereas aggressive ligand exchange may increase conductivity at the expense of environmental stability [20,21,22]. Therefore, fabrication and surface engineering become central design variables rather than secondary processing steps.

Figure 2 summarizes the general operating framework of NP-based IR photodetectors and highlights the multiscale relationship between material properties and detector performance. The detection process begins with IR photon absorption within the nanostructured active layer, followed by the generation of photocarriers or plasmon-induced hot carriers, depending on the material system. These carriers then migrate through the NP network via transport mechanisms governed by interparticle coupling, ligand chemistry, trap states, and film morphology before being collected at the electrodes to generate photocurrent. The figure further emphasizes that optimization of each stage directly influences key device metrics including responsivity (R), specific detectivity (D*), response speed, and noise characteristics.

The integration of nanocrystal films into MSM configurations is particularly attractive, as the solution-processable, low-temperature deposition of NPs aligns perfectly with the planar geometry of MSM electrodes, thereby facilitating straightforward fabrication on flexible or unconventional substrates [16,17]. Furthermore, the nanoscale architecture and tailored surface states of NP-based films can be utilized to modify charge injection and transport dynamics at the substrate–NP interface, offering innovative solutions to tackle conventional MSM issues such as I_dark [18,19]. For example, surface-ligand passivation with halides (I⁻, Br⁻) can reduce mid-gap trap-state densities from > 10¹⁷ cm⁻³ to < 10¹⁶ cm⁻³, effectively suppressing Shockley–Read–Hall recombination noise [20]. Thus, NPs can be synthesized in solution through scalable, cost-effective techniques, paving the way for versatile, large-area, and on-chip integrated photodetectors [21,22].

This review builds upon fundamentals in IR detector physics and material requirements [23,24] and will explore how advancements in NP synthesis and patterning of NPs in materials are revitalizing traditional device architectures. At the same time, the switch from traditional IR detectors to NP-based detectors will allow for reducing production costs.

Figure 3 provides a schematic overview linking NP materials, spectral selectivity strategies, fabrication methods, device architectures, and performance metrics, highlighting how nanoscale engineering controls both sensitivity and wavelength discrimination.

1.2. Spectral Selectivity in Nanoparticle-Based IR Photodetectors

Spectral selectivity, which is the ability of a photodetector to preferentially respond to a defined wavelength range while suppressing out-of-band signals, is an increasingly important characteristic for imaging, spectroscopy, environmental sensing, biomedical diagnostics, and multispectral detection systems [25,26,27]. Unlike conventional bulk semiconductor detectors, where selectivity is primarily determined by the intrinsic bandgap and external optical filters, NP-based IR photodetectors provide multiple routes to engineer wavelength discrimination directly within the active material and the device architecture. The direct way to wavelength selectivity is bandgap engineering through quantum confinement in semiconductor NPs and CQDs. By controlling NP size and composition, the absorption edge can be continuously shifted, allowing selective detection across near-infrared (NIR), short-wave infrared (SWIR), mid-wave infrared (MWIR), and long-wave infrared (LWIR) regions [15,28,29,30]. For example, reducing CQD dimensions increases quantum confinement and enlarges the effective bandgap, whereas larger particles extend absorption toward longer wavelengths. Additional selectivity can be achieved through optical cavities and resonant structures. Fabry–Perot cavities, distributed reflectors, plasmonic metasurfaces, and metamaterial absorbers enhance optical confinement at selected wavelengths and suppress undesired spectral contributions [31,32,33].

Such approaches are particularly attractive for compact multispectral systems because they reduce dependence on external optical filters. At the device level, heterostructure and multilayer engineering further improve selectivity by controlling carrier generation and extraction. Graded junctions, stacked NP absorbers, and spectrally selective transport layers enable wavelength-dependent carrier collection and broadband-to-narrowband response conversion [34,35,36]. More recently, patterned CQD arrays and printed heteropixel architectures have enabled filterless multispectral imaging and compact computational spectrometers [37]. Spectral selectivity can also emerge from anisotropic and polarization-sensitive materials, including black phosphorus and low-dimensional nanostructures, whose absorption depends on crystallographic orientation and incident polarization [38,39]. Combined with machine-learning-assisted spectral reconstruction and detector arrays, these approaches are extending NP-based IR detection from single-band sensing toward integrated spectroscopic and intelligent imaging platforms. Although enhanced selectivity generally improves functional capability, it often introduces trade-offs between bandwidth, absorption efficiency, fabrication complexity, and signal-to-noise ratio. Consequently, optimization of spectral selectivity should be considered together with responsivity (R), specific detectivity (D*), and manufacturability.

1.3. Literature Search and Selection Methodology

This review was prepared using a structured literature survey of peer-reviewed journal articles, conference proceedings, and selected review papers related to NP-based IR photodetectors. The literature search was primarily conducted using the Web of Science, Scopus, Google Scholar, IEEE Xplore, and ScienceDirect databases. Publications from approximately 2010 to early 2025 were considered, with emphasis on recent advances reported during the last five years. Keywords and combinations of terms used during the search included “nanoparticle infrared detector”, “colloidal quantum dot photodetector”, “HgTe CQD”, “PbS infrared photodiode”, “plasmonic infrared detector”, “solution-processed IR detector”, “SWIR photodetector”, “MWIR nanomaterials”, “surface passivation”, “IR detector fabrication”, “specific detectivity”, and “CMOS-compatible infrared imaging”.

The selection criteria prioritized studies experimentally reporting device performance evaluation, scalable fabrication strategies, or significant advances in material engineering, spectral tunability, noise reduction, and room-temperature operation. Additional emphasis was placed on publications presenting quantitative metrics such as specific detectivity (D*), responsivity (R), response time (τ), dark current (I_dark), and operational wavelength range under clearly stated measurement conditions. Review articles were included primarily to provide historical context and identify emerging research trends. The comparison tables summarize representative “state-of-the-art” results selected based on a combination of reported device performance, relevance to NP-based IR detection, fabrication maturity, and citation impact. Because measurement conditions vary substantially across detector architectures and spectral regions, the reported values should be interpreted as representative benchmarks rather than directly comparable rankings.

In this review, IR detection is subclassified into near-infrared (NIR, 0.75–1.4 µm), short-wave infrared (SWIR, 1.4–3.0 µm), mid-wave infrared (MWIR, 3–5 µm), long-wave infrared (LWIR, 8–12 µm), and very-long-wave infrared (VLWIR, > 12 µm) if not further specified.

2. Key NP Materials Systems: Synthesis, Properties, and Metrics

2.1. Colloidal Quantum Dots (CQDs): Beyond Size Tunability

Lead sulfide (PbS) and lead selenide (PbSe) CQDs are particularly significant, as their band gaps can be adjusted from the visible spectrum to the MWIR by altering their size [25,26,27]. Recent advancements in their synthesis, such as the continuous-flow microreactor technique, have facilitated production of CQDs with remarkable uniformity and reproducibility between batches, which is essential for commercial focal plane arrays (FPAs) [28]. Furthermore, the introduction of mercury telluride (HgTe) CQDs has broadened the operational range of CQDs into the VLWIR (> 8 µm) at room temperature, due to the narrow bulk bandgap of HgTe CQDs [29,30]. A significant challenge for HgTe CQDs is the stabilization of the elemental Hg⁰ state; however, advanced synthesis techniques utilizing HgCl₂ precursors in hexadecanethiol/oleylamine mixtures have enhanced oxidation resistance, achieving D* > 10⁹ Jones at 10 µm and 300 K [29,30,31]. The quantum efficiency and absorption coefficients of these materials make them particularly beneficial for high-performance photodetectors, as detailed in system-level assessments [32]. The surfaces of CQDs are passivated with organic ligands, which are crucial for maintaining stability and promoting electronic transport, although they also pose challenges for the development of conductive films [33]. The balance between steric stabilization and charge transport is measured by the inter-dot coupling energy, which can range from < 1 meV for long-chain oleic acid to > 50 meV for short inorganic ligands such as arsenite, AsS₃³⁻, directly affecting film mobility [34].

2.2. Plasmonic Metal NPs: Engineering the Local Field

Au, Ag, and Al NPs exhibit notable localized surface plasmon resonance (LSPR), which can be tuned into the IR spectrum by altering their shape, size, and the dielectric environment surrounding them [35,36]. For IR applications, anisotropic structures such as Au nanorods, bipyramids, and hollow nanoshells are specifically designed [37,38,39]. The LSPR wavelength λ_max for a nanorod is described by the relationship λ_max∝aspect ratio × ϵ_m^1/2, where ϵ_m represents the dielectric constant of the medium, enabling precise adjustments from 1 to 10 µm [37]. These NPs are primarily employed as sensitizers to enhance light absorption in adjacent semiconductor active layers or to create hot-carrier-based detectors [38,39]. In hot-electron detectors, the internal quantum efficiency (η_IQE) is frequently low (< 1%) due to swift electron–electron scattering. Nevertheless, the integration of a plasmonic nanostructure with an energy-selective electron acceptor, such as a TiO₂ or graphene layer, can improve η_IQE by selectively extracting carriers with energies exceeding the Schottky barrier [40]. Their integration can be optimized through modeling techniques aimed at increasing photodetector responsivity and reducing noise [41].

2.3. Emerging Low-Dimensional and Hybrid Nanostructures

NPs made from narrow-bandgap materials, including graphene quantum dots (GQDs) [42], black phosphorus nanosheets [43], and topological insulator NPs like Bi₂Se₃ [44], are currently being studied for their unique carrier dynamics and extensive IR response. 2D black phosphorus (BP) features a direct bandgap that varies with layer thickness, ranging from 0.3 eV in bulk form to approximately 2.0 eV in monolayer form, thus providing an exceptional platform for broadband detection. Nonetheless, its susceptibility to environmental factors is being mitigated through in situ covalent functionalization using aryl diazonium salts, which enhances its operational lifespan in air from a few hours to several months while maintaining a high hole mobility of over 100 cm²/V·s [45]. Additionally, hybrid systems such as CsPbI₃ perovskite nanocrystals integrated within a PbS CQD matrix have shown synergistic benefits: the perovskite effectively passivates surface traps in PbS, while the CQD matrix facilitates a conductive pathway, resulting in a fivefold increase in responsivity (R) compared to films composed of a single component [46].

Table 1. Synthesis strategies for nanomaterials used in IR detection.

Synthesis Strategy	Principle	Representative Materials	Key Advantages	Main Limitations	Typical Feature Size	Typical Applications	Refs.
Hot-Injection	rapid precursor injection at elevated temperature to separate nucleation and growth	PbS, PbSe, CsPbX3	excellent size control; high crystallinity	batch-to-batch variation; precursor sensitivity	2–10 nm	CQD photodiodes, SWIR detectors	[25,26,47]
Continuous-Flow/Microreactor	flow-based synthesis with controlled residence time and heat transfer	PbS	scalable; reproducible; precise thermal control	more complex equipment	3–8 nm	large-scale CQD production	[48]
Cation Exchange	partial/complete ion replacement in preformed nanocrystals	PbSe, CdSe, Cu2Se	enables composition tuning and heterostructures	surface defects may form	Tunable	bandgap engineering; heterojunction devices	[25,26,49,50]
Solvothermal/Hydrothermal	high-pressure reaction in sealed reactor	Bi2Se3, BP nanosheets	simple processing; ligand-free synthesis	broader size distribution	5–100 nm	broadband IR absorbers	[43,44]
Seed-Mediated Growth	controlled growth on pre-existing seeds	Au, Ag	shape and aspect-ratio control	complex growth kinetics	aspect ratio 2–10	plasmonic IR structures	[51]
Ligand-Assisted Precipitation	rapid crystallization using precursor–antisolvent interaction	Perovskite nanocrystals	low-temperature and scalable	limited monodispersity	3–15 nm	solution-processed photodetectors	[52]
Colloidal Synthesis via Metal Chalcogenide Precursors	controlled reaction of metal and chalcogen precursors	PbS, AgBiS2	improved stoichiometry and reduced defects	higher precursor cost	4–10 nm	high-quality CQD films	[53,54]
Epitaxial Nanostructure Growth (SK/VW/FM)	strain- or surface-driven crystal growth on lattice-engineered substrates	Ge/Si QDs, GeSiSn QDs, III–V QDs	excellent crystallinity; low defect density; precise interfaces; CMOS compatibility	high equipment cost; limited throughput; thermal constraints	self-assembled QDs (typically 5–50 nm)	Si-compatible IR photodiodes, QD detectors, integrated photonics	[55]

provides a comparative overview of the principal synthetic methodologies employed for producing NPs suitable for IR detector applications. The table categorizes synthesis approaches according to their reaction mechanisms, scalability, structural precision, and compatibility with IR-active materials. The hot-injection method remains the most widely adopted strategy for CQDs, offering exceptional size control and high crystallinity through rapid nucleation–growth separation at elevated temperatures [25,26,47]. However, its batch-based nature and sensitivity to precursor reactivity limit scalability. In contrast, continuous-flow or microreactor synthesis enables improved reproducibility and scalable production under precisely controlled laminar flow conditions, achieving narrower dispersity but requiring more complex instrumentation [48].

Cation-exchange techniques provide a powerful route for post-synthetic bandgap engineering and heterostructure formation via ion substitution within preformed nanocrystals [25,26,49,50]. This approach is particularly valuable for PbSe/CdSe and Cu₂Se/PbSe systems, although it may introduce surface defects that influence electronic transport. Solvothermal and hydrothermal methods, performed under high-pressure autoclave conditions, offer ligand-free and relatively simple synthesis routes for layered or narrow-bandgap materials such as Bi₂Se₃ and BP nanosheets, albeit with broader size distributions (5–100 nm) [48,49]. For plasmonic and anisotropic nanostructures, seed-mediated growth allows precise control over morphology and aspect ratio (2–10), which is critical for tuning localized surface plasmon resonance (LSPR) into the IR regime [51]. Ligand-assisted reprecipitation provides a low-temperature, scalable pathway for the synthesis of perovskite nanocrystals (3–15 nm), though it typically sacrifices monodispersity [52]. Finally, metal chalcogenide complex-based colloidal synthesis enables improved stoichiometric control and reduced defect density in materials such as PbS and AgBiS₂ at the expense of higher precursor costs [53,54].

Beyond colloidal synthesis approaches, epitaxial formation of nanostructures represents an important route for producing highly ordered IR-sensitive nanomaterials with superior crystallinity and interface quality. In contrast to solution-processed NP assembly, epitaxial growth forms nanostructures directly on crystalline substrates through controlled atom-by-atom deposition and strain-driven self-organization. Epitaxial nanostructure growth is a key method for making IR photodetectors when precise interfaces and semiconductor compatibility are needed. Unlike colloidal methods, it grows nanostructures directly on crystalline substrates using self-assembly modes (SK, VW, FM–see below), enabling highly uniform CQD and heterostructures such as GeSi, III–V, or GeSiSn. These structures offer low defect density, good carrier transport, low dark current, and reproducibility. However, the approach is complex and requires high thermal budgets and sophisticated equipment, limiting scalability. It remains ideal for high-performance, Si-compatible IR photodiodes and CQD arrays where uniformity and interface quality are critical [55].

Several epitaxial growth modes are relevant for IR photodetector production. The Frank–van der Merwe (FM) mode promotes layer-by-layer growth and is commonly used for smooth heterointerfaces. Volmer–Weber (VM)growth leads to island formation due to weak substrate interactions, while the Stranski–Krastanov (SK) mode combines initial wetting layers followed by spontaneous QD formation resulting from accumulated strain. The SK mechanism is particularly important for IR devices because it enables self-organized CQD with strong three-dimensional carrier confinement. Ge CQDs grown on silicon through SK epitaxy have shown extended absorption into the SWIR region while maintaining compatibility with complementary metal–oxide–semiconductor (CMOS) processing. More recently, GeSiSn CQD heterostructures have attracted attention because Sn incorporation reduces the bandgap and shifts optical absorption toward longer IR wavelengths. Compared with colloidal NP processing, epitaxial nanostructures generally provide lower defect densities, improved carrier mobility, and superior interface control, which reduce dark current and enhance specific detectivity (D*). However, these advantages come at the expense of increased fabrication complexity, higher thermal budgets, and reduced scalability relative to solution-processed methods.

Table 2. Nanomaterial systems for IR detection.

Material	Type	Tunable Range (µm)	Key Advantages	Key Challenges	Primary Synthesis Route	Typical Mobility (cm2/V·s)	Performance record (D*, Jones)	Refs.
PbS/PbX2 CQDs	semiconductor	0.8–4.0	good tunability, high absorption	surface traps, ligand resistivity	hot-injection	10−2–101	> 1013 at 1.55 µm, 300 K	[56]
HgTe CQDs	semiconductor	3–12	room-T VLWIR operation	oxidation, stability	modified hot-injection	10−3–10−1	2 × 1010 at 5 µm, 300 K	[29,30,31]
Au/Ag NPs	plasmonic metal	1–10	strong LSPR, ultrafast response	low R, cost	seed-mediated growth	N/A (Metal)	N/A (speed <100 fs)	[35,37,40]
Graphene QDs	2D	< 10	high mobility, flexibility	low absorption	top-down etching	~102	1011 at 1.5 µm, 300 K	[42,57]
BP nanosheets	2D semiconductor	1–8	anisotropy, high mobility	ambient instability	liquid exfoliation	102–103	1011 at 3.4 µm, 300 K	[43,45]
CsPbX3 nanocrystals	perovskite	< 1.8	high defect tolerance, fast response	ion migration, thermal stability	ligand-assisted reprecipitation, hot-injection	10−4–10−2	109 at 1.5 µm, 300 K	[47]
Bi2Se3 NPs	topological insulator	2–8	topological surface states	complex synthesis, impurities	coll. synthesis	N/A	109 at 4 µm, 300 K	[58]

provides a comparative overview of the principal NP-based material platforms currently explored for IR photodetection, highlighting the relationships among spectral tunability, carrier transport, synthesis strategies, and device-level performance. The comparison reveals that different NP systems occupy distinct performance niches, with no single material simultaneously optimizing spectral range, mobility, environmental stability, and manufacturability. PbS/PbX₂ (X = Cl, Br, I) CQDs remain the dominant platform for NIR and SWIR detection due to their exceptional bandgap tunability through quantum confinement, strong optical absorption, and compatibility with low-temperature solution processing [56]. Their reported specific detectivities (D*) exceeding 10¹³ Jones at 1.55 µm and room temperature position them among the highest-performing solution-processed IR materials. However, device performance is strongly limited by surface trap states and ligand-induced tunneling barriers, which increase 1/f noise and reduce carrier mobility. Consequently, ligand engineering and surface passivation remain central to further performance improvements.

HgTe CQDs extend detection into the mid-wave infrared (MWIR), long-wave infrared (LWIR), and very-long-wave infrared (VLWIR) regions (3–12 µm), enabling room-temperature operation in spectral regimes traditionally dominated by epitaxially grown HgCdTe detectors [29,30,31]. Despite their strategic importance for thermal imaging and spectroscopy, HgTe CQDs generally exhibit lower mobilities (10⁻³–10⁻¹ cm²/V·s) and higher dark currents than PbS-based systems because of their narrow bandgap and increased susceptibility to oxidation and trap-assisted tunneling. This illustrates a recurring trade-off in NP-based IR detectors: extending spectral coverage toward longer wavelengths often comes at the expense of transport stability and noise performance.

Plasmonic Au and Ag NPs operate through localized surface plasmon resonance (LSPR), fundamentally differing from semiconductor CQDs that rely on interband absorption [35,37,47]. Their primary advantage lies in ultrafast carrier dynamics (< 100 fs) and strong local electromagnetic field enhancement. However, plasmonic enhancement frequently improves responsivity more effectively than detectivity because thermal noise and inefficient hot-electron extraction limit the achievable signal-to-noise ratio. As a result, plasmonic architectures are more attractive for ultrafast photonic applications than for ultra-sensitive imaging systems. Among emerging two-dimensional (2D) materials, graphene quantum dots (GQDs) offer high carrier mobility (~10² cm²/V·s), broadband operation, and excellent mechanical flexibility, supporting IR detection up to 10 µm with D* values approaching 10¹¹ Jones [42,57]. Nevertheless, the intrinsically weak optical absorption of graphene-derived materials necessitates optical confinement or plasmonic enhancement strategies. BP nanosheets exhibit high mobilities (10²–10³ cm²/V·s) and unique in-plane anisotropy, enabling polarization-sensitive photodetection across the 1–8 µm range [43,45]. However, rapid ambient degradation and oxidation remain major obstacles to large-scale deployment.

CsPbX₃ perovskite nanocrystals (X = I, Br, Cl) showed excellent defect tolerance, high absorption coefficients, and rapid photoresponse in the visible-to-NIR region [47]. Yet, unlike PbS CQDs, their long-term operational reliability is hindered by ion migration and thermal instability. Similarly, topological insulator NPs such as Bi₂Se₃ provide broadband IR absorption enabled by protected surface states, although current device performance remains limited by synthesis complexity and impurity-related scattering [58].

In addition to colloidal and solution-processed NP systems, Si-compatible nanostructured photodetectors represent an important class of IR detection technologies due to their compatibility with mature CMOS manufacturing infrastructure. In particular, heterostructures based on self-organized Ge QDs grown on Si substrates and GeSiSn heterostructure photodiodes have attracted increasing attention for extending silicon photonics into the IR region. Unlike colloidal systems that rely primarily on solution-phase assembly, these architectures employ nanoscale heterostructure engineering through epitaxial growth and strain-driven self-organization. GeSi quantum dots are typically synthesized via the Stranski–Krastanov growth mode, producing nanoscale islands with quantum confinement effects and modified electronic band structures. The resulting heterostructures support interband and intraband optical transitions that enable photodetection from the NIR to the SWIR regime while maintaining compatibility with silicon platforms. Ge quantum-dot photodetectors also benefit from reduced lattice mismatch relative to III–V systems and enable monolithic integration with Si readout electronics and photonic circuits. Beyond Ge QDs, GeSiSn heterostructure photodiodes have emerged as a promising route for extending detection toward longer IR wavelengths through bandgap engineering. Incorporation of Sn reduces the bandgap and can promote direct-bandgap behavior at sufficient concentrations, enabling enhanced absorption and improved carrier collection in the SWIR and MWIR regions. GeSiSn structures additionally provide opportunities to improve spectral selectivity and reduce thermal leakage currents through engineered confinement. However, challenges remain in controlling strain relaxation, defect density, Sn segregation, and thermal stability during growth and processing. Although GeSi and GeSiSn platforms generally do not yet match the broad spectral tunability achievable with HgTe CQDs, their compatibility with wafer-scale fabrication and silicon photonics makes them attractive candidates for future integrated IR imaging and spectroscopic systems.

While the primary focus of this review remains NP-enabled fabrication approaches, these Si-compatible nanostructured platforms illustrate complementary pathways for integrating nanoscale IR absorbers with scalable electronic and photonic technologies.

3. Advances in Fabrication and Integration Techniques

3.1. Synthesis and Surface Engineering: From Lab to Fab

Modern hot-injection and cation-exchange methods enable precise CQD size control within 5% for narrow spectral response [59,60]. Continuous-flow synthesis enhances scalability, with segmented-flow droplet reactors producing PbS CQDs that showed an 85 meV photoluminescence linewidth at 1500 nm for sharper detector cut-offs [61]. Advances in ligand engineering, such as replacing long-chain oleic acid with shorter alternatives like halides or hybrid perovskites, improved inter-dot coupling and carrier mobility, enhancing photoconductive gain and responsivity [62,63]. Solid-state ligand exchange using gaseous formic acid (HCOOH) created a conductive Pb-formate matrix, achieving film mobility of 16 cm²/V·s, three orders of magnitude higher than conventional materials [64,65]. In situ spectroscopic techniques (FTIR, XPS) are commonly employed to measure ligand exchange efficiency and correlate it with trap-state density [66,67]. As illustrated in Figure 4, NP synthesis and surface engineering jointly determine the structural and electronic properties of the active layer by controlling particle size, surface defects, and interparticle coupling, which ultimately govern IR detector performance.

3.2. Deposition and Patterning Techniques: Enabling Large-Area Electronics

For deposition and patterning of NPs, a number of methods are available.

• Spin-coating and blade-coating: Remain the conventional techniques for laboratory-scale and proof-of-concept devices. Recent research highlights the importance of solvent engineering and meniscus-guided coating methods to produce more uniform and thicker films [68,69]. For IR absorbers that necessitate thick films (> 500 nm), blade-coating with solvent mixtures (such as octane:butyronitrile) that create a “stick-slip” motion can yield highly uniform, pinhole-free CQD layers, thereby decreasing pixel non-uniformity in arrays to below 5% [70].
• Inkjet printing and 3D aerosol jet printing: Enable digital and additive patterning of NP-based inks, facilitating the monolithic integration of multispectral pixel arrays onto flexible substrates or readout circuits. This marks a significant progress towards the creation of low-cost focal plane arrays (FPAs) [71,72,73]. The capability of aerosol jet printing to concentrate NP aerosol stream allows for the direct writing of CQD patterns with approximately 10 µm linewidth onto pre-patterned Complementary Metal-Oxide-Semiconductor (CMOS) readout integrated circuits (ROICs), which is a crucial advancement towards heterogeneous integration [74].
• Layer-by-Layer (LbL) assembly & electrophoretic deposition (EPD): Provide exceptional control over the thickness and composition of films. EPD, which employs an applied electric field to propel charged NPs to an electrode, is especially effective in forming dense, tightly packed films of large HgTe CQDs for LWIR detection, enhancing packing density by around 30% compared to spin-coating [29,30,75,76,77].
• Spray pyrolysis & electrospray deposition: These are innovative techniques for large-area, conformal coatings. Ultrasonic spray pyrolysis, in which a precursor mist is pyrolyzed on a heated substrate, can deposit high-quality, doped CQD films in a single step, paving the way for direct in situ doping during deposition [78].

Table 3. Deposition and patterning techniques.

Method	Principle	Scalability	Film Uniformity	Resolution	Throughput	Best For	Refs.
Spin-Coating	centrifugal force	lab-scale	high (thin)	wafer-scale	high	R&D, test structures	[76,77]
Blade/Bar-Coating	meniscus guidance	pilot-line	excellent (thick)	mm scale	very high	large-area flexible substrates	[78]
Inkjet Printing	drop-on-demand	high	moderate	20–50 µm	medium	multispectral pixel patterning	[71,72,73,74,79]
Aerosol Jet Printing	focused aerosol stream	high	good	~10 µm	medium	direct write on 3D/CMOS	[74,80]
Electrophoretic Deposition (EPD)	electric field-driven	medium	very high (dense)	electrode-defined	medium	dense LWIR films	[81]
Spray Pyrolysis	in situ reaction from the mist	high	moderate	cm scale	high	conformal, doped films	[77,82]

compares the principal thin-film deposition and patterning techniques used for fabricating NP-based IR photodetectors, evaluating each method in terms of working principle, scalability, film uniformity, resolution, and throughput. The table highlights the trade-offs between laboratory simplicity, industrial scalability, and patterning precision that ultimately determine their suitability for large-area or CMOS-integrated IR systems.

Table 3 also compares representative deposition and patterning strategies used for NP-based IR detectors; however, their importance extends beyond manufacturability because each method directly influences film morphology, interfacial quality, and ultimately detector performance. Processing conditions affect active-layer thickness, ligand removal efficiency, defect density, contact resistance, surface roughness, pixel uniformity, and compatibility with readout electronics, all of which govern responsivity, dark current, response speed, and specific detectivity (D*). Spin-coating remains the dominant method for laboratory-scale device fabrication because centrifugal spreading produces highly uniform thin films with precise thickness control and good wafer-scale reproducibility [75,76]. These characteristics support reduced surface roughness and stable junction formation, contributing to lower leakage current and improved D*. However, spin-coated films often require repeated deposition and post-treatment cycles for thicker absorbers, while material waste and limited scalability restrict translation to industrial manufacturing.

Blade (bar) coating enables meniscus-controlled film formation and is better suited for large-area and roll-to-roll processing [78]. Compared with spin-coating, blade-coated layers offer improved material utilization and greater control over thick-film deposition, which is beneficial for IR absorbers requiring extended optical path lengths. Nevertheless, film thickness gradients and solvent-drying effects must be controlled because local non-uniformity can increase trap density and generate pixel-to-pixel performance variation in detector arrays. Digital additive methods, including inkjet printing and aerosol jet printing, establish a stronger link between fabrication and system integration. Inkjet printing allows localized deposition and mask-free patterning with resolutions of approximately 20–50 µm, enabling direct fabrication of multispectral pixels and focal plane arrays (FPAs) [71,72,73,74,79]. Device performance is strongly influenced by droplet coalescence, coffee-ring effects, and solvent evaporation kinetics, which determine local thickness, ligand distribution, and electrical continuity. Aerosol jet printing further improves pattern resolution (~10 µm) and permits direct writing onto three-dimensional and CMOS-compatible substrates [74,80]. This capability reduces alignment complexity and supports heterogeneous integration with ROIC platforms, although optimization of surface roughness and contact resistance remains necessary for reproducible operation. Electrophoretic deposition (EPD) produces dense NP assemblies through electric-field-driven assembly and is particularly attractive for MWIR/LWIR devices that require thick absorbing layers [81]. The dense morphology can improve optical absorption and reduce void formation; however, excessive compaction may hinder ligand exchange and carrier extraction, increasing recombination losses and limiting response speed. Spray pyrolysis provides a scalable route for conformal deposition over large areas and supports doped or compositionally graded films [77,82]. Although throughput is high, thermal processing conditions strongly influence grain growth, interface quality, and defect formation. Maintaining uniform morphology is especially important for preserving low dark current and ensuring high manufacturing yield. Overall, fabrication should be viewed as a performance-defining parameter rather than merely a processing step. Future progress will depend on co-optimization of synthesis, deposition, post-processing, and device architecture to simultaneously achieve high D*, low noise, array uniformity, ROIC compatibility, and scalable manufacturing.

3.3. Device Architecture Innovations: Engineering the Charge Pathway

Different applications allow us to define different materials with different demands.

• Photoconductors: This represents the fundamental architecture. The gain (G) in a photoconductor is expressed as G = τ/τt, where τ denotes the carrier lifetime, and τt signifies the transit time. In films composed of NPs, τ is notably influenced by trap states. By incorporating a minor proportion of reduced graphene oxide (rGO) into a PbS CQD film, percolation pathways for holes can be established, which enhances τ and enables gains exceeding 10⁴, albeit with a reduction in response speed [83].
• Photodiodes: The incorporation of CQD films into p–n or p–i–n junctions leads to a significant decrease in I_dark. A key advancement is the implementation of graded heterojunctions, where the size of the CQD (bandgap) is progressively altered throughout the depletion region. This configuration generates a built-in quasi-electric field that facilitates carrier separation, thereby minimizing the reverse bias required for complete depletion and consequently reducing tunneling I_dark [84]. Heterojunctions that utilize materials such as graphene or ZnO are the subject of extensive research [85,86,87].
• Plasmon-enhanced and metamaterial detectors: The integration of shaped Au nanorods or nanoantennae with CQD channels leads to significant enhancements in the local field. Metasurfaces, which are periodic arrangements of resonant elements, can be engineered to achieve perfect absorption at a designated IR wavelength within an ultra-thin (< λ/10) CQD layer, significantly improving quantum efficiency without elevating noise levels [88].
• Phototransistors and gate-modulated devices: The use of an NP-based film as the channel material in a field-effect transistor (FET) configuration permits direct control of gain through the gate voltage. PbSe CQD FET photodetectors have shown ultrahigh gain (> 10⁵) and low noise-equivalent power (NEP) by functioning in the subthreshold regime, where the conductivity of the channel is exponentially responsive to changes in photo-induced gate potential [25,26,89,90].

Beyond single-pixel photodetectors, recent studies have demonstrated the scalability of CQD architectures toward array-based IR sensing systems and integrated spectrometers. A notable example is the development of HgTe CQD homojunction photodiode arrays integrated with Fabry–Perot resonant cavities for SWIR spectroscopy applications [91]. In these architectures, p-type, intrinsic (i), and n-type CQD layers form a p-i-n homojunction structure that improves carrier collection efficiency and suppresses interfacial carrier loss compared with conventional heterojunction designs. The incorporation of resonant cavity structures further enhances optical absorption and spectral selectivity while maintaining room-temperature operation. Using this strategy, 256-channel CQD detector arrays covering the 1.4–2.5 µm spectral range with spectral reconstruction capability approaching 7 nm resolution have been reported, highlighting the potential of NP-based detector arrays for compact and low-cost integrated IR spectrometers [91].

Figure 5 schematically shows the fabrication of a photodetector using various techniques.

Table 4. Device Architectures and Their Key Characteristics.

Architecture	Operating Mechanism	Key Metrics	Selectivity Mechanism	Advantages	Key Challenges/Noise Mechanism	Refs.
Photoconductor	conductivity modulation	R, gain, NEP	bandgap-controlled absorption	simple, high gain	high 1/f noise, low speed	[16,92]
Photodiode (p–i–n)	built-in field separation	D*, R, EQE, Idark	junction + optical cavity	low noise, fast, photovoltaic	complex fabrication, interface traps	[93,94]
Plasmonic Hot-Electron	hot carrier injection	τ, ηIQE	resonance wavelength	ultrafast, sub-bandgap	very low ηIQE, thermal noise	[95,96]
Hybrid Heterojunction	band alignment engineering	R, spectral range	band alignment	broadband, multifunctional	interface recombination	[46,97]
Phototransistor (FET)	gate-modulated conductivity	gain, NEP, specific detectivity	gate-tunable spectral response	tunable gain, very high sensitivity	gate hysteresis, stability	[98,99]
Metamaterial Absorber	resonant cavity enhancement	absorption %, Q-factor	resonant spectral filtering	near-unity absorption in an ultra-thin layer	narrowband, complex fabrication	[100,101]
Avalanche Photodiode (APD)	impact ionization and internal carrier multiplication under reverse bias	gain (M), D*, bandwidth, excess noise factor	primarily determined by absorber bandgap and optical filtering a	internal gain, high sensitivity, fast detection	Multiplication noise, high bias voltage, temperature sensitivity	[102]
Single-Photon Avalanche Diode (SPAD)/Silicon Photomultiplier (SiPM)	Geiger-mode avalanche triggering from single-photon events	photon detection efficiency (PDE), timing jitter, dark count rate	defined by absorber spectral response b	single-photon sensitivity, ultralow signal detection	afterpulsing, dark counts, dead time	[103]

^a Can be extended using hybrid NP absorbers, resonant cavities, or wavelength-selective heterostructures. ^b Enhanced through optical filters, cavity engineering, wavelength conversion layers, or NP-assisted spectral tuning.

provides a comparative overview of the principal device architectures employed in NP-based IR detectors, summarizing their operating mechanisms, key performance metrics, inherent advantages, and dominant noise or fabrication challenges. The table highlights how architectural design plays a decisive role in balancing responsivity (R), detectivity, speed, and system complexity. The photoconductor represents the simplest implementation, operating through photogenerated conductivity modulation under an applied bias. It typically delivers high R and photoconductive gain, making it attractive for proof-of-concept devices [16,92]. In both APD and SPAD architectures, spectral selectivity is generally governed by the absorber material and optical coupling design rather than the avalanche process itself; therefore, integration with wavelength-selective NP layers and resonant optical structures offers a route toward spectrally engineered high-gain IR detection.

However, performance is often limited by pronounced low-frequency 1/f noise and slow response times (τ) due to trap-mediated carrier dynamics in NPs. In contrast, the p–i–n photodiode relies on built-in electric field separation of photocarriers within a depletion region, enabling low I_dark, fast response, and photovoltaic operation without external gain. Key parameters include specific detectivity (D*), external quantum efficiency (EQE), and responsivity (R). Although this architecture offers superior noise performance compared to photoconductors, it requires more complex fabrication and careful control of interface trap states to suppress recombination and tunneling currents [93,94]. Plasmonic hot-electron detectors exploit localized surface plasmon resonance (LSPR) to generate energetic carriers that overcome a Schottky barrier [95,96]. These devices exhibit ultrafast response times and enable sub-bandgap detection, making them attractive for high-speed and narrowband applications. However, their η_IQE is typically low, and thermal noise can limit detectivity. The hybrid heterojunction architecture integrates NPs with complementary materials such as graphene, ZnO, or perovskites, to engineer band alignment and facilitate efficient charge separation [46,97,104]. This design supports broadband and multifunctional detection, including extended spectral coverage and enhanced responsivity. The primary challenge lies in minimizing interfacial recombination and controlling defect states at heterogeneous interfaces. Phototransistor (FET)-based detectors incorporate NP-based films as the channel material within a field-effect geometry, enabling gate-modulated conductivity [98,99]. These devices can achieve extremely high gain and specific detectivity, particularly when operated in the subthreshold regime. Nevertheless, issues such as gate hysteresis, bias instability, and long-term reliability remain significant concerns. Finally, metamaterial absorber architectures use resonant cavity or metasurface designs to achieve near-unity absorption within ultra-thin NP layers [100,101]. Performance is often characterized by absorption percentage and quality factor (Q-factor). While these structures dramatically enhance light–matter interaction without increasing film thickness, they are typically narrowband and involve larger fabrication complexity.

Although NP synthesis and deposition methods are commonly discussed as independent processing steps, their influence extends directly to detector-level performance. Parameters including nanocrystal size dispersion, ligand composition, film thickness, packing density, and post-deposition treatment determine charge transport, optical absorption, dark current, and device reproducibility. Consequently, fabrication choices strongly influence specific detectivity (D*), responsivity (R), response speed, and manufacturability. Synthetic routes affect defect formation and surface chemistry at the nanocrystal level. For example, hot-injection synthesis typically produces highly monodisperse CQDs, reducing energetic disorder and improving carrier transport, whereas scalable continuous-flow approaches improve batch reproducibility and support industrial implementation. Surface ligand engineering further governs interparticle spacing and electronic coupling; shorter inorganic or mixed ligands generally improve mobility but may reduce environmental stability. Deposition and patterning methods additionally determine active-layer morphology and integration compatibility. Uniform coating methods reduce thickness variation and improve pixel-to-pixel consistency, while direct-write techniques facilitate monolithic integration with readout integrated circuits (ROICs). Therefore, fabrication optimization should be considered a central determinant of detector performance rather than merely a processing step.

4. Performance Benchmarks, Noise Analysis, and System Integration

HgTe CQD photodiodes now achieve D* > 10¹⁰ Jones at room temperature in the 3–5 µm MWIR range, approaching cooled MCT detector benchmarks [29,30,31,105,106,107]. For NP photodiodes, D* = R√AΔf/i_n, where R is responsivity, A is detector area, Δf is bandwidth, and i_n is the total noise current [108]. PbS CQD photodiodes exceed 10¹³ Jones in the SWIR (1–2 µm) [109,110]. Plasmonic hot-electron detectors offer ultrafast response times < 1 ps despite lower responsivity [111]. NP integration with silicon photonics and CMOS ROICs advances practical applications [112,113,114]. Direct inkjet printing of PbS CQD arrays onto CMOS ROIC wafers has produced 320 × 256 SWIR imaging systems with NETD < 50 mK [115]. The specific detectivity D^∗ is a normalized sensitivity metric that accounts for detector area and measurement bandwidth, enabling comparison between devices of different geometries. In practice, reported D^∗ values may vary substantially depending on whether the noise current is experimentally measured or estimated from dark-current-limited shot noise assumptions. Consequently, direct comparison between different studies should be approached cautiously unless measurement conditions, bandwidth, modulation frequency, and noise evaluation methodologies are explicitly specified.

Table 5. Selected performance metrics of NP-based IR detectors.

Material/Structure	Spectral Range (µm)	Peak Wavelength (µm)	D* (Jones) a	R (A/W)	Bias Voltage (V)	Operation T (K)	Response Time	Architecture	Ref.
PbS CQD Photodiode	1–2	1.54	> 1011	0.2–1.2	−0.5	300	µs	p–n on Si heterojunction	[116]
HgTe CQD Photodiode	< 2.8	1.55	1011	0.24	0.6	300	µs	p–i–n on Si	[117]
Au/graphene/Au	< 1.33	1.33	2.4 × 1013	10−3	0	300	360 ns	plasmonic Schottky	[118]
Graphene–PbS CQD Hybrid	1.45–1.59	1.55	1010	107	5	300	Ms	photoconductor	[119]
CsPbI3/PbS	< 1.7	1.55	4.5 × 1013	11.12	0.53	300	Ms	bulk heterojunction	[120]
Modified HgTe CQD	8–18	18	6.6 × 108	0.3	3	80	µs	MSM-photoconductor	[121]
PbSe-MoS2	2–3	3	7.7 × 1010	137.6	1	300	Ms	broadband absorber	[122]
HgTe CQD Homojunc. array with Fabry–Perot Cavity	1.4–2.5	2.5	4.64 × 1011	0.54	1	300	µs	56-channel CQD spectrometer array	[123]

^a The reported D* values were derived under different specific experimental conditions and calculated using dark-current-limited noise assumptions rather than direct noise spectral measurements. Therefore, direct comparison with detectors characterized using measured noise spectra should be treated cautiously.

summarizes selected state-of-the-art performance metrics reported for representative NP-based IR photodetectors across different material systems, device architectures, and spectral regions. The comparison includes spectral coverage, specific detectivity (D*), responsivity (R), operating temperature, response time (τ), and structural configuration, offering a comprehensive benchmark of current technological capabilities. PbS CQD photodiodes operating in the SWIR (1–2 µm) showed detectivities exceeding 10¹¹ Jones at room temperature, with responsivities ranging from 0.2 to 1.2 A/W and microsecond-scale response times [116]. Their implementation in p–n heterojunctions on silicon highlights compatibility with CMOS platforms and strong potential for low-cost imaging arrays.

HgTe CQD photodiodes extend detection from visible into the SWIR (< 2.8 µm), while maintaining room-temperature operation. With D* values on the order of 10¹¹ Jones and microsecond response speeds [117,124]. Plasmonic Au/graphene/Au Schottky structures showed exceptionally high detectivity (2.4 × 10¹³ Jones) and ultrafast response of about 360 ns, although responsivity (R) remains relatively low (~10⁻³ A/W) [118]. These devices highlight the trade-off between speed and quantum efficiency inherent in hot-carrier-based architectures. Hybrid systems such as graphene–PbS CQD photoconductors achieve extraordinarily high R values (up to 10⁷ A/W) due to photogating effects and internal gain mechanisms [119]. However, their response times are typically in the millisecond regime, reflecting gain–speed trade-offs. D* values around 10¹⁰ Jones indicate strong sensitivity but with higher noise contributions compared to photovoltaic architectures. The CsPbI₃/PbS bulk heterojunction photodetector showed an exceptionally high reported specific detectivity approaching 4.5 × 10¹³ Jones at room temperature near 1.55 µm under low-bias operation [120]. However, this value should be interpreted cautiously because the detectivity was derived using dark-current-based noise estimation rather than direct experimental noise spectral measurements. In addition, the reported performance was obtained under optimized laboratory conditions and has not yet been broadly reproduced across independent studies. While the result highlights the strong potential of hybrid perovskite/CQD architectures for visible-to-SWIR photodetection, it currently represents an outlier rather than a universally established benchmark. The unusually high D* likely originates from the combined effects of efficient charge separation, suppressed recombination at the bulk heterojunction interface, and photoconductive gain mechanisms. Nevertheless, further studies employing standardized noise characterization protocols and statistically reproducible device datasets are required before direct comparison with mature HgTe or PbS CQD technologies can be made. For LWIR (8–18 µm) detection, HgTe CQD operating as photoconductors exhibit detectivity approaching 6.6 × 10⁸ Jones at 80 K with microsecond response times, marking significant progress toward LWIR operation [121]. Finally, nanostructured PbSe–MoS₂ perfect absorber architectures showed high responsivity (137.6 A/W) and detectivity (~7.7 × 10¹⁰ Jones) in the 2–3 µm range, leveraging resonant cavity enhancement to maximize absorption in ultra-thin active layers [122].

Cross-comparison of the reported performance metrics must be approached carefully because D*, R, and τ strongly depend on measurement wavelength, modulation frequency, detector area, applied bias, operating temperature, and noise extraction methodology. In several reports, detectivity values were estimated from shot-noise-limited I_dark assumptions rather than measured noise spectra, which can significantly overestimate practical detector performance. Furthermore, architectures exhibiting extremely high R values often rely on photogating or trap-assisted gain mechanisms that inherently increase τ and 1/f noise. Consequently, direct ranking across photodiodes, photoconductors, phototransistors, and plasmonic detectors may be misleading unless measurement conditions are normalized. This highlights the need for standardized benchmarking protocols in NP-based IR photodetector research.

Figure 6 illustrates a representative architecture of an NP-based IR photodetector and highlights the sequential processes that convert incident IR radiation into an electrical signal. The device consists of an active NP layer deposited between charge-selective transport layers and contacted by top and bottom electrodes integrated on a supporting substrate. Incoming IR photons are absorbed within the NP film, where size-dependent electronic structure and quantum confinement enable tunable spectral sensitivity across different IR regions. Following photon absorption, electron–hole pairs are generated and separated either by an externally applied electric field or by the built-in potential of a junction structure. Efficient charge transport depends strongly on interparticle coupling, ligand chemistry, carrier mobility, and the density of trap states within the NP solid. Surface passivation and controlled ligand exchange are therefore essential to suppress recombination and reduce dark current (I_dark). The extracted carriers are collected at the electrodes and converted into a measurable electrical output, typically characterized through responsivity (R), specific detectivity (D*), response time, and noise-equivalent power (NEP). The figure also emphasizes that device performance emerges from the coupling of multiple design levels, including NP material selection, surface engineering, film deposition quality, interface design, and device architecture, rather than from material properties alone. This multiscale interaction explains why fabrication strategy directly influences final detector sensitivity, selectivity, speed, and scalability.

Operating temperature is one of the most important variables governing IR detector performance because specific detectivity (D*), dark current (I_dark), response speed, and noise mechanisms are strongly T-dependent. Although room-temperature operation is frequently highlighted as an advantage of NP-based detectors, direct comparison across detector platforms requires careful normalization with respect to spectral region, measurement conditions, and cooling requirements. In general, reducing operating temperature suppresses thermally generated carriers and decreases dark current, leading to improved D*. However, lower temperatures may also modify carrier mobility, trap occupancy, recombination dynamics, and response kinetics. In NP solids, transport frequently occurs through thermally activated hopping and tunneling mechanisms; therefore, excessive cooling may not always yield proportional performance improvements. For SWIR devices (~0.9–2.5 µm), PbS and related CQD photodiodes often achieve high room-temperature D* values (> 10¹¹–10¹³ Jones) due to relatively low thermal carrier generation and favorable bandgap energies. In contrast, MWIR (3–5 µm) and LWIR (> 8 µm) operation requires narrower bandgap materials, making dark current increasingly dominant at room temperature and often necessitating cooling strategies. Comparison between NP-based detectors and conventional IR technologies such as HgCdTe (MCT) and III–V superlattice detectors should be interpreted carefully because performance depends strongly on spectral regime and operating temperature. For SWIR detection, room-temperature NP detectors increasingly approach practical performance levels required for imaging and spectroscopy while offering advantages in solution processability and CMOS compatibility. However, for MWIR and particularly LWIR applications, cooled MCT detectors continue to provide superior detectivity and lower noise under equivalent measurement conditions. For example, state-of-the-art cooled MCT focal plane arrays routinely achieve D* values exceeding 10¹¹–10¹² Jones in the MWIR and LWIR ranges under cryogenic operation, whereas room-temperature HgTe CQD systems typically report lower detectivity and narrower operating windows. Therefore, room-temperature NP detector operation should not be interpreted as a direct replacement of cooled MCT systems but rather as an alternative pathway for applications prioritizing cost, integration flexibility, compactness, and reduced system complexity.

5. Critical Challenges, Foundational Limits, and Emerging Frontiers

Despite the significant advancements outlined in earlier sections, the journey towards the broad commercialization of NP-based IR detectors is hindered by various interconnected scientific and engineering obstacles. A thorough examination of these challenges, based on recent scholarly work, is crucial for steering future investigations.

5.1. Material Stability Beyond Surface Passivation

The intrinsic instability of narrow-bandgap nanomaterials, especially HgTe CQDs and black phosphorus (BP), continues to be a major constraint. Although halide passivation greatly enhances the longevity of PbS CQDs [15], HgTe CQDs are prone to oxidation and decomposition through the reaction HgTe + O₂ → Hg⁰ + TeO₂, which compromises the crystal structure and optical characteristics [31]. Recent approaches have moved beyond static encapsulation. In situ development of conformal wider-bandgap shells (e.g., CdS on CdTe) via cation exchange forms quasi-type-II core/shell structures that protect the core while promoting charge extraction, prolonging operational lifespan by an order of magnitude [123]. However, these methods often introduce new interfacial defect states. In operando X-ray absorption spectroscopy is essential to link chemical degradation pathways with real-time device performance degradation [125]. Figure 7 presents a comparative overview of the principal performance metrics used to evaluate IR photodetectors and illustrates the multidimensional trade-offs that govern detector optimization. The figure highlights key parameters, including specific detectivity (D*), responsivity (R), response time (τ), noise-equivalent power (NEP), spectral selectivity, dark current (I_dark), and operating temperature, emphasizing that detector performance cannot be assessed through a single metric alone.

Specific detectivity (D*) represents the ability of a detector to distinguish weak optical signals from noise and is strongly influenced by dark current and noise sources. Responsivity (R) quantifies the electrical output generated per unit optical input and reflects the efficiency of photon absorption, carrier generation, and collection. Response time characterizes the speed of detector operation and determines suitability for dynamic imaging and high-frequency sensing applications. NEP defines the minimum detectable optical signal and is inversely related to sensitivity. The figure further emphasizes that these parameters are interdependent and often involve competing trade-offs. For example, increasing internal gain can enhance R but may simultaneously increase response time and noise, while extending operation toward longer IR wavelengths often introduces higher I_dark and reduced D*. Temperature also plays a critical role because lower operating temperatures suppress thermally generated carriers and improve D*, particularly in MWIR and LWIR detectors.

5.2. Dark Current (I_dark) and Noise: Probing the Microscopic Origins

The I_dark present in NP-based photodiodes consists of a combination of diffusion, generation-recombination (G–R), and trap-assisted tunneling (TAT) components. Although the p–i–n architecture mitigates diffusion current, TAT current, represented by I_TAT∝exp (−E_t/kT), where E_t denotes the trap energy level, frequently prevails in mid-IR devices [126]. The design known as the “bulk barrier” or “graded heterojunction” [127] is a complex architectural solution. By manipulating a spatially varying composition, it establishes a progressively increasing electron affinity that functions as an energy filter, effectively obstructing one type of carrier and reducing TAT. Recent simulations that integrate density functional theory (DFT) with nonequilibrium Green’s function (NEGF) formalism have refined these graded profiles, forecasting a decrease in I_dark by a factor of 100 in comparison to abrupt junctions [128]. Additionally, low-frequency 1/f noise, which is prevalent in disordered NP solids, has been directly associated with fluctuations in charge carrier numbers due to trapping and detrapping at interface states. Measurements of correlated noise and admittance spectroscopy indicate that ligand-exchange processes aimed at minimizing mid-gap states, for instance, utilizing metal chalcogenide complexes, are more effective than those that solely enhance mobility [129].

5.3. Charge Transport and Mobility

The mobility of carriers (μ) in a CQD solid is constrained by two primary factors: the tunneling barrier that exists between the dots, which is influenced by the inter-dot distance and the dielectric constant of the ligands, and the energetic disorder, which is determined by the distribution of sizes and the uniformity of surface traps. The quest for so-called mono-atomic ligands such as S²⁻, I⁻, or NH₂⁻ has proven to be revolutionary, as it has decreased the inter-dot spacing to less than 1 nm and facilitated band-like transport with mobilities surpassing 10 cm²/V·s for electrons [130]. Nevertheless, a new approach highlights the importance of mixed ligand shells [131]. In this work, a composite passivation layer that includes a short conductive ligand such as formate, HCOO⁻, to enhance transport, along with a longer, cross-linkable ligand such as mercaptopropionic acid, to provide mechanical stability, has resulted in films exhibiting high mobility (μ > 5 cm²/V·s) and remarkable durability against solvent processing [131]. To understand the charge percolation pathway in these hybrid solids, sophisticated techniques like conductive atomic force microscopy (C-AFM) can be used to map the variations in the nanoscale conductivity [132].

5.4. Scalability and Uniformity: From Wafers to Webs

Transitioning from lab-scale spin-coating to high-throughput manufacturing is complex. Blade-coating and slot-die coating provide excellent thickness control over large areas [133], but maintaining uniformity in nanocrystal size and ligand coverage during continuous deposition remains difficult. Convergent synthesis methods, which more rigorously distinguish nucleation and growth phases, produce CQDs with narrower size distributions (< 4%) that resist Ostwald ripening, enhancing pixel-to-pixel uniformity in coated arrays [134]. Electrohydrodynamic jet printing achieves sub-micron resolution and prints high-viscosity NP inks for direct patterning of multilayer device stacks without masks [135]. Combining these deposition techniques with in-line optical spectroscopy for real-time process monitoring represents the next advancement in process control [136].

5.5. Emerging Frontiers: New Materials and Novel Concepts

Lead-free III–V colloidal quantum dots (CQDs) have attracted significant attention for near-infrared (NIR) to short-wave infrared (SWIR) photodetection applications [137]. Some examples have been studied for this purpose. AgBiS₂ nanocrystals show broadband absorption but lower defect tolerance [138]. InAs CQDs offer SWIR tunability but suffer from high surface trap densities [139]. Cation-ordered double perovskites (e.g., Cs₂AgBiBr₆) provide stable, low-toxicity materials, though band gaps remain too wide for IR detection [140].

Neuromorphic computing: Memristive properties of certain NPs (e.g., perovskites under bias) enable devices integrating photodetection with synaptic plasticity, where photocurrent history influences conductance to facilitate image preprocessing (e.g., edge detection) at the sensor level [141,142].

Multispectral/polarimetric detection: Mixing NP inks enables multispectral pixels. Inkjet printing of different-sized CQDs (1.2, 1.5, and 2.0 µm) into single pixels has been achieved [143]. Anisotropic optical characteristics of aligned 2D nanosheets like BP permit filterless polarization detection [144,145].

Fundamental limits: As NP devices approach bulk semiconductor performance, questions emerge about ultimate detectivity limits from Heisenberg’s uncertainty principle and whether topological insulator NPs can utilize protected surface states for noiseless transport [146].

Table 6. Current challenges and mitigation strategies.

Challenge Category	Specific Issue	Current Mitigation Strategies	Performance Impact	Key Research Needs	Refs.
Material stability	oxidation, ligand desorption, and ion migration (perovskites)	inorganic shells, cross-linked ligands, and ALD encapsulation	extends operational lifetime from h to > 1000 h	in operando degradation studies; predictive stability models	[144,145]
Electronic noise	high Idark, 1/f noise, trap-assisted tunneling	graded heterojunctions, bulk barriers, trap-passivating ligands	reduces Idark by 2–3 orders of magnitude; lowers NEP	correlated noise-defect spectroscopy; DFT modeling of trap states	[146]
Charge transport	low mobility, energetic disorder	atomic/mixed ligand exchange, doping, energy-level alignment	increases μ (mobility) from 10−3 to > 10 cm2/V·s	nanoscale mapping of transport	[147]
Scalability	film non-uniformity, low throughput, and pattern resolution	slot-die/blade coating, electrohydrodynamic printing, convergent synthesis	enables uniform films > 100 cm2; < 5% pixel variation	in-line metrology; roll-to-roll process integration	[148]
Next-gen. materials	Pb/Hg toxicity, stability, and scalability challenges; limited transport performance in lead-free alternatives	Ag2Se/SiO2/Si heterojunction	D* > 1010 Jones for some Pb-free materials in SWIR	discovery of new narrow-gap, defect-tolerant compositions	[149]
Novel functionality	single-band spectral response, limited processing	multispectral inkjet printing, neuromorphic device design	enables hyperspectral and pre-processed imaging	co-design of materials, devices, and algorithms	[150]

summarizes the major scientific and technological barriers limiting the performance and commercialization of NP-based IR detectors, together with state-of-the-art mitigation strategies, their impact on device metrics, and critical future research directions. The table highlights that progress in this field requires coordinated advances in materials chemistry, device physics, and scalable integration technologies. Material stability remains one of the most pressing concerns. NP systems, particularly HgTe CQDs and halide perovskites, are vulnerable to oxidation, ligand desorption, and ion migration. Approaches such as inorganic shell growth, cross-linked or hybrid ligand systems, and atomic layer deposition (ALD) encapsulation have significantly improved environmental robustness, extending operational lifetimes from only a few hours to more than 1000 h under controlled conditions [144,145].

Deeper insight into degradation pathways via in operando characterization and predictive modeling is still required. Electronic noise (high I_dark, 1/f noise, trap-assisted tunneling) fundamentally constrains D* [141]. Graded heterojunctions, bulk blocking layers, and trap-passivating ligands have reduced I_dark by 2–3 orders of magnitude and lowered NEP. Future work should correlate microscopic defect states with macroscopic noise via advanced spectroscopy and DFT-based trap modeling. Charge transport limitations from low mobility and energetic disorder limit speed and sensitivity [147]. Mono-atomic/mixed ligand exchange, doping, and energy-level alignment have increased mobility from ~10⁻³ to > 10 cm²/V·s. However, spatial heterogeneity in transport requires nanoscale mapping (C-AFM, TEM-EELS). Scalability challenges include film non-uniformity, limited throughput, and insufficient pattern resolution for FPAs [148]. Slot-die coating, blade coating, Electrohydrodynamic printing, and convergent synthesis enable uniform films > 100 cm² with < 5% pixel variation. Industrial adaptation requires in-line metrology and roll-to-roll integration. Pb/Hg toxicity drives development of Pb-free alternatives; Ag₂Se/SiO₂/Si heterojunctions show D* > 10¹⁰ Jones in SWIR [149]. Novel functionality expansion via multispectral inkjet printing and neuromorphic NP-based devices enables hyperspectral imaging and in-sensor preprocessing [150]. Future advances require co-design, integrating materials development, device engineering, and computational algorithms.

Table 7. Future research directions and opportunities for NP-based IR detectors.

Research Focus Area	Strategic Goal	Emerging Approach/Technology	Expected Impact	Key Challenges	Refs.
Stable surface chemistry	extend operational lifetime	dynamic ligand exchange; self-healing surface layers	> 10× stability in air	maintaining conductivity after passivation	[151]
Noise suppression	lower NEP and improve D*	defect-engineered CQDs, noise-filtered readout circuits	D* > 1013 Jones at 300 K	interface charge trapping	[152]
Quantum transport and tunneling control	band-like transport	mixed ligand strategies	μ > 10 cm2/V·s	structural disorder	[153]
Scalable manufacturing	transition to roll-to-roll	in-line metrology; robotic ink formulation	low-cost, large-area IR sensors	film homogeneity	[154]
Lead-free alternatives	environmental safety	AgBiS2, InAs, perovskite derivatives	sustainable, RoHS-compliant detectors a	low mobility, narrow range	[155,156]
Neuromorphic/intelligent sensing	combine detection + processing	perovskite memristive pixels, optoelectronic synapses	Real-time edge computing	scalability, reproducibility	[104,157]
Multispectral/polarimetric systems	enhance imaging functionality	printed heteropixels, anisotropic 2D NPs	filterless multispectral, polarimetric imaging	crosstalk, spectral selectivity	[158]

^a RoHS = Restriction of Hazardous Substances.

presents a strategic roadmap outlining emerging research priorities aimed at advancing NP-based IR detectors beyond current performance and integration limits. The table connects specific research focus areas with their technological goals, enabling approaches, anticipated performance impact, and remaining technical challenges. A central priority is stable surface chemistry, where the goal is to extend operational lifetime under ambient conditions [151]. Emerging strategies such as dynamic ligand exchange and self-healing surface passivation layers aim to provide more than a tenfold improvement in air stability. However, preserving high electrical conductivity while implementing robust passivation remains a major challenge, as stronger surface binding often reduces carrier mobility.

In the area of noise suppression, efforts focus on reducing the noise-equivalent power (NEP) and increasing specific detectivity (D*). Defect-engineered CQDs combined with advanced, noise-filtered readout circuits are projected to enable room-temperature detectivities exceeding 10¹³ Jones [152]. Yet, interface charge trapping and defect-induced fluctuations continue to limit low-frequency noise performance. Quantum transport and tunneling control represent another frontier, targeting band-like carrier transport in NPs. Mixed ligand strategies techniques seek to reduce energetic disorder and increase carrier mobility beyond 10 cm²/V·s [153]. Achieving this requires overcoming structural disorder and maintaining long-range electronic coupling across the NP assemblies. Scalable manufacturing is essential for commercialization. Transitioning from batch processing to roll-to-roll production with in-line metrology and robotic ink formulation promises low-cost, large-area IR sensor fabrication [154]. However, film homogeneity and reproducibility across industrial scales remain primary technical barriers.

Figure 8 depicts a roadmap of the evolution of NP-based IR detection technologies from early proof-of-concept demonstrations to future intelligent sensing platforms. The timeline highlights the transition from fundamental material exploration toward system-level integration and application-oriented development within about 20 years.

The first stage (early exploration in the years 2000–2010) focused primarily on demonstrating the feasibility of NP photoconductors for IR detection, establishing the foundational concepts of solution-processable photodetectors. The second stage (material advances, 2010–2016) emphasized improvements in NP synthesis, composition control, and surface stabilization, particularly for colloidal quantum dots (CQDs) such as PbS and HgTe, enabling enhanced optical absorption, spectral tunability, and environmental stability. The third phase (device engineering, 2016–2020) marked the emergence of advanced architectures, including heterojunction photodiodes, PIN structures [124], and phototransistors designed to improve carrier extraction, reduce dark current, and enhance specific detectivity (D*). Subsequently, the array & imaging stage (2020 to present) reflects the shift toward scalable device integration, including focal plane arrays (FPAs), multispectral imaging systems, and compact spectrometer platforms based on NP detector arrays. Finally, the roadmap projects future directions toward Intelligent & Integrated Systems, where NP-based IR detectors are expected to combine with artificial intelligence, heterogeneous electronics integration, neuromorphic sensing concepts, and multifunctional imaging platforms. Overall, the figure shows how progress across materials, fabrication, device design, and system integration collectively drives the transition of NP-based IR detectors from laboratory-scale research to practical sensing technologies.

Driven by environmental regulations such as the Restriction of Hazardous Substances (RoHS) [155,156], lead-free alternatives such as AgBiS₂, Ag₂Se, InAs, and perovskite derivatives are being explored. Although these materials demonstrated competitive detectivity (D* > 10¹⁰ Jones in some SWIR devices), they face challenges including lower carrier mobility, higher defect densities, and limited spectral tunability compared to Pb/Hg-based CQDs. Achieving high performance while maintaining RoHS compliance remains a key research direction [156,157].

Emerging paradigms such as neuromorphic and intelligent sensing aim to integrate detection and signal processing. Perovskite memristive pixels and optoelectronic synapses offer real-time edge computing at the sensor level [104,157]. Challenges remain in scalability, reproducibility, and system-level integration. Multispectral and polarimetric systems using printed heteropixel arrays and anisotropic 2D NPs enable filterless multispectral and polarization-sensitive detection [158], though minimizing pixel crosstalk and maintaining spectral selectivity are ongoing priorities.

Figure 9 summarizes the future outlook and remaining challenges for NP-based IR detection technologies by presenting the balance between technological opportunities and the barriers that must be overcome for large-scale implementation. The left side highlights emerging opportunities, including the development of lead-free and environmentally friendly nanomaterials, scalable large-area and solution-based fabrication routes, integration with CMOS platforms and flexible electronics, and the incorporation of artificial intelligence (AI) for computational imaging and intelligent signal processing. These advances are expected to expand the application space of NP-based IR detectors into areas such as the Internet of Things (IoT), wearable sensing, space technologies, biomedical diagnostics, and compact spectroscopic systems. The right side illustrates the principal challenges limiting commercialization and widespread adoption. These include carrier trapping and associated low-frequency (1/f) noise, environmental instability and oxidation of sensitive nanomaterials, difficulties in achieving high device uniformity and reproducibility over large areas, limitations associated with forming low-resistance and stable electrical contacts, and the need for cost-effective manufacturing processes. Together, these factors emphasize that future progress will require coordinated optimization across materials design, fabrication processes, device engineering, and system-level integration to achieve high-performance, scalable, and reliable IR detection technologies.

5.6. Critical Analysis of Performance Trade-Offs in NP-Based IR Detectors

One of the central trade-offs in NP-based IR photodetectors is the inverse relationship between photoconductive gain and response speed. Photoconductors and phototransistors often achieve exceptionally high responsivity through photogating effects and long carrier lifetimes associated with trap-mediated charge separation. In graphene–CQD hybrids, for example, trapped carriers can modulate channel conductivity for extended durations, producing internal gain values exceeding 10⁶–10⁷. However, these same long-lived trap states slow carrier recombination dynamics and increase response times from the microsecond to millisecond regime. Consequently, devices optimized for ultrahigh gain often exhibit limited modulation bandwidth and poor suitability for high-frame-rate imaging applications. In contrast, photovoltaic architectures such as p–i–n photodiodes typically sacrifice gain to achieve lower noise and faster operation. This trade-off highlights the importance of balancing gain mechanisms with carrier extraction kinetics depending on the target application.

5.6.1. Gain–Bandwidth Trade-Off in Photoconductive Architectures

The relationship between surface chemistry and electronic noise remains one of the defining challenges in CQD-based IR detectors. Surface ligands are essential for colloidal stability and solution processability, yet long organic ligands introduce tunneling barriers that reduce carrier mobility and increase energetic disorder. Incomplete ligand exchange additionally creates surface trap states that promote carrier trapping/detrapping processes, which are widely recognized as the origin of 1/f noise in NP solids. Although shorter inorganic or hybrid ligands improve charge transport, aggressive ligand removal can simultaneously increase surface defect density and compromise environmental stability. As a result, optimization of ligand chemistry requires balancing conductivity, passivation quality, and long-term device reliability. Recent strategies involving mixed-ligand systems, atomic ligands, and inorganic shell passivation have demonstrated reduced dark current and suppressed low-frequency noise, but the microscopic correlation between defect states and noise spectra remains insufficiently understood.

5.6.2. Plasmonic Enhancement: Responsivity (R) Versus Detectivity (D*)

Plasmonic nanostructures significantly enhance local electromagnetic fields and optical absorption through localized surface plasmon resonance (LSPR). While this enhancement frequently increases responsivity (R), it does not necessarily translate into proportional improvements in specific detectivity (D*). In many plasmonic hot-electron devices, optical gain is accompanied by increased thermal noise, Schottky leakage current, or nonequilibrium carrier scattering losses. Consequently, devices exhibiting strong R enhancement may still possess moderate D* values if the associated noise current is also increased. Furthermore, plasmonic enhancement is often highly wavelength- and geometry-dependent, limiting broadband performance. Therefore, the practical value of plasmonic architectures depends not only on field enhancement magnitude but also on whether noise generation and carrier extraction efficiencies can be simultaneously controlled.

5.6.3. Limitations of Hot-Electron Internal Quantum Efficiency

Despite strong local-field enhancement in plasmonic nanostructures, the internal quantum efficiency (IQE) of hot-electron IR detectors generally remains low. This limitation originates from several competing ultrafast processes. Following plasmon excitation, most absorbed energy rapidly dissipates through electron–electron and electron–phonon scattering before energetic carriers can overcome the Schottky barrier. In addition, only a small fraction of generated hot carriers possess both sufficient energy and appropriate momentum for successful injection across the metal–semiconductor interface. Momentum mismatch, interfacial defects, and short carrier mean free paths further reduce collection efficiency. Consequently, although plasmonic devices achieve ultrafast temporal response, their conversion efficiency remains substantially lower than that of conventional semiconductor photodiodes.

5.6.4. Practical Challenges in Neuromorphic NP-Based IR Detectors

Neuromorphic NP-based photodetectors represent an emerging direction that combines sensing, memory, and signal preprocessing within a single device platform. Although proof-of-concept demonstrations have shown optoelectronic synaptic behavior and wavelength-selective memory effects, several practical limitations remain unresolved. Device endurance under repeated switching cycles is often limited by ion migration, trap-state evolution, or structural degradation. Retention times and conductance linearity also vary substantially between devices, complicating reliable analog signal processing. Furthermore, significant device-to-device variability arising from stochastic nanomaterial assembly and interface disorder currently limits large-scale reproducibility. As a result, neuromorphic NP-based detectors remain at an early exploratory stage, and substantial advances in materials uniformity, defect engineering, and fabrication control are necessary before practical deployment in edge-computing imaging systems becomes feasible.

Overall, these trade-offs show that optimization of NP-based IR detectors cannot rely on maximizing a single performance metric. Improvements in responsivity, gain, spectral range, or speed are frequently coupled to penalties in noise, stability, reproducibility, or manufacturability. Future progress therefore requires co-design strategies integrating materials chemistry, interface engineering, device physics, and system-level considerations rather than isolated optimization of individual parameters.

5.7. Environmental, Stability, and Manufacturability Considerations for NP-Based IR Detectors

Although PbS and HgTe CQDs remain the benchmark materials for high-performance NP-based IR photodetectors, their practical deployment raises important concerns beyond detector performance alone. Lead- and mercury-containing systems provide excellent spectral tunability, mature colloidal synthesis routes, and currently deliver the highest room-temperature specific detectivity (D*) across SWIR to LWIR regimes. However, these materials face increasing challenges related to environmental impact, long-term stability, and large-scale manufacturability. From an environmental and regulatory perspective, both Pb- and Hg-based CQDs are subject to restrictions associated with hazardous substances and electronic waste regulations, including RoHS-related considerations for commercial deployment. Encapsulation strategies and sealed-device architectures can reduce direct exposure risks, but they introduce additional fabrication complexity, increase cost, and may complicate recycling and end-of-life processing. Material stability is another major limitation. PbS CQDs commonly suffer from oxidation and ligand desorption, which generate surface trap states and increase dark current over time. HgTe CQDs are particularly sensitive to air and moisture exposure. Encapsulation approaches based on atomic layer deposition (ALD), inorganic shell growth, and cross-linked ligand networks improve operational lifetime but often introduce transport penalties through increased carrier barriers or reduced mobility. Manufacturability also remains challenging. High-performing CQD devices typically require tightly controlled ligand exchange, oxygen-free processing, and multilayer deposition protocols to maintain reproducibility across large areas. These requirements complicate wafer-scale integration and may reduce fabrication yield compared with mature epitaxial semiconductor technologies.

Lead-free alternatives including AgBiS₂, Ag₂Se, InAs, Si-compatible nanostructures, and perovskite-derived materials are being actively explored to improve sustainability and regulatory compliance. Nevertheless, these systems generally remain behind PbS and HgTe in detector performance. Their lower performance is attributed to a combination of (i) reduced carrier mobility due to less favorable band transport pathways, (ii) higher defect and trap densities that increase recombination and noise, (iii) more limited bandgap tunability restricting spectral coverage into MWIR/LWIR regions, and (iv) less mature surface passivation chemistries that hinder efficient charge extraction. As a result, while several lead-free systems already showed D* values exceeding 10¹⁰ Jones in SWIR operation, they have not yet consistently matched the combination of detectivity, stability, and spectral breadth achieved by optimized PbS and HgTe platforms.

Future progress will likely depend not only on discovering alternative compositions but also on developing defect-tolerant electronic structures, improved ligand chemistries, and scalable encapsulation methods that preserve both environmental compatibility and device performance.

6. Conclusions and Outlook

This thorough review has attempted to trace the transformative evolution of NP-based technologies, transitioning from laboratory curiosities to promising candidates in the realm of IR detection. We have elaborated on how colloidal quantum dots (CQDs), plasmonic metal NPs, and novel low-dimensional materials, through precise synthetic control, provide unmatched spectral tunability ranging from the short-wave to the very-long wavelength IR, an achievement that traditional bulk semiconductors cannot replicate. In addition to materials, the review has rigorously examined the collaborative advancement of fabrication techniques, encompassing high-precision continuous-flow synthesis, atomic-scale ligand engineering, and scalable deposition methods such as blade-coating and additive digital printing. These innovations collectively facilitate the design of advanced device architectures ranging from low-noise photodiodes and ultrafast plasmonic detectors to gate-tunable phototransistors that are expanding the limits of responsivity, detectivity, and operational speed.

Nevertheless, the journey towards widespread technological adoption and the market displacement of established technologies is not solely an engineering challenge; it signifies a fundamental frontier in materials science. The primary limitations, such as environmental instability of narrow-gap nanocrystals, the ongoing trade-off between charge transport and trap-state density, and the integration gap with metal–oxide–semiconductors (CMOS), are intricately linked to the nanoscale physics of surfaces, interfaces, and disorder [159]. Consequently, future advancements necessitate a paradigm shift from empirical optimization to predictive, physics-informed design. This involves utilizing first-principles calculations and machine learning to identify new, stable material compositions and surface passivating agents (ligands) [97,111], employing in situ and in operando spectroscopic tools to trace degradation pathways in real-time [96], and creating multiscale models that relate atomic-scale defects to macroscopic device noise [80,98]. The primary objective is to create NP solids not merely as random assemblies but as meticulously engineered “meta-materials” that possess specific charge transport and optoelectronic characteristics. The future perspective goes beyond mere incremental enhancements in performance. We foresee the integration of NP-based IR detectors with various technological advancements, such as heterogeneous integration with silicon photonics for on-chip spectroscopy [108,109], neuromorphic engineering aimed at developing sensors with built-in preprocessing capabilities for edge computing [92,112], and the creation of sustainable, heavy-metal-free active materials to tackle environmental and regulatory issues [110]. The unique solution-processability and flexibility of NPs make them particularly suitable for applications in wearable health monitoring, distributed environmental sensor networks, and lightweight, low-power imaging systems for autonomous vehicles and drones.

In conclusion, the field is at a pivotal moment. The upcoming years will not be characterized by the question of whether NP-based IR detectors can function but rather by the efficiency and reliability of their manufacturing and integration into larger systems. Future success will depend on ongoing, interdisciplinary collaboration among chemists, physicists, materials scientists, and engineers. By adopting a co-design approach that connects material properties, device architecture, and fabrication processes from the beginning, the aspiration for high-performance, low-cost, and versatile IR sensing, previously limited to high-end defense and scientific applications, can transform into a widespread reality, opening new avenues in safety, health, and our comprehension of the world.