Photosensors based on colloidal quantum dots

The environment is perceived by humans mainly by vision and hearing organs, with 70% of information being gained by vision. Despite the unique characteristics of the human eye as an optical photosensor[1], it can perceive only things in a rather narrow range of electromagnetic radiation wavelengths, from ~0.38 μm to ~0.72 μm (visible region), while leaving subjects in X-ray (10^–6 – 10^–2 μm), ultraviolet (0.1 – 0.4 μm) and infrared (0.8 – 100 μm) electromagnetic wavelengths beyond the visibility.

The attempts to fabricate devices that would detect and visualize ultraviolet and infrared radiation sources, which were actively started after the World War II, initiated a new research area called optoelectronics; the milestones of its development have been described in a number of publications[2-6]. Optoelectronic devices form the basis for night vision, thermal imaging, infrared direction finding, laser location, optical data transmission and other systems. Of particular interest are research and development related to the elemental base of modern optoelectronics, that is, photosensors sensitive to near-infrared (NIR, 0.8 – 1.0 μm), short-wave infrared (SWIR, 1.0 – 3.0 μm), mid-wave infrared (MWIR, 3 – 5 μm), long-wave infrared (LWIR, 8 – 14 μm), very long-wave infrared (VLWIR, 14 – 30 μm), and extremely-long wave infrared (30 – 100 μm) electromagnetic radiation.

There exist a variety of approaches to the design of photosensing devices. The design and manufacturing of photosensors for the wavelength range of 0.4 – 0.9 μm covering the visible and near-infrared light are largely related to image intensifiers, which use the external emission of electrons excited by absorption of photons focused by an IR lens on the photocathode of an electro-optical converter[6].

For spectral wavelengths λ > 0.9 μm, development of photosensors is mainly associated with the solid-state semiconductor photoelectronics based on the internal photoelectric effect. In photonic sensors, the observed signal is the measure of excitation by incident photons of the electronic system of a semiconducting material forming the photosensitive structure. There are a number of architectures for photonic elements that form the basis of virtually all types of photosensors used in both the infrared and ultraviolet and visible regions. Most often, single-area and line photosensors with no more than 200 elements are designed on the basis of photoresistors. There are also examples of photodetector arrays. For example, a latest-generation photodetector array is a two-dimensional GaAs/AlGaAs-based quantum well infrared photodetector (QWIP), which uses the photoresistive effect[4]. Energy barrier photodiodes can be based on both homo-(p-n) and heterojunctions of various types (P-n, N-p, P-p, n-N), Schottky barriers, p-i-n homo- and heterojunctions, avalanche gain, bariode heterosystems, injection diodes, etc. Most of modern 2D photodetector arrays are based on photodiodes. Depending on the required sensitivity range, operating temperature and operation speed, photodetectors are manufactured using various materials such as Ge and Si (in particular, those doped with Hg, Au, Zn, B, As, P, Al, In, S, Mg, and other ions), PbS, PbSe, PbTe, CdS, CdSe, InSb, InP, GaAs, InAs, InGaAs, CdHgTe, and so on[7-9]. In the far-IR region, apart from photodetectors, non-selective thermal detectors (microbolometers) are often used[4].

The modern trend of development of solid-state IR photosensors is associated with the design of 2D photodetector arrays. Currently, they are being developed virtually for all semiconductor compounds that have a sufficient photosensitivity in the ultraviolet and near-, mid- and far-IR regions and in the terahertz (THz) region. These camera arrays (N × N format in systems without image scanning) and TDI subarrays [M × N format, where М = 2, 4, 8, 12, for the time delay and integration (TDI) mode in optical mechanical scanning systems) are referred to as second-generation solid state photosensing devices[4].

A considerable progress has been made in the development of second-generation solid-state semiconductor photosensors. Imaging arrays comprising more than a few million elements with 1280 × 1024, 1920 × 1080, 2048 × 2048, and 4096 × 4096 formats (Figure 1) have now been fabricated for the main infrared spectral regions, 0.9 – 1.7 μm, 3 – 5 μm and 8 – 12 μm, corresponding to the transparency windows of the Earth’s atmosphere[10, 11]. Lately, special attention has been paid to the development of multiband devices sensitive in several spectral regions. The possibilities of increasing the operating temperature are being investigated. The ways for reducing size, weight and power consumption of photosensors [size-weight-and-power (SWaP) concept of development] are being explored.

Figure 1

Photosensitive element based on Cd_xHg_1–xTe photodiode (l_c = 2.5 µm) of 4096×4096 format for the James Webb space telescope; l_c is the cut-off wavelength[10].

Meanwhile, in recent years, certain drawbacks of the currently developed technologies of infrared and ultraviolet photosensors based on solid-state semiconductor electronic materials, structures and approaches have become more obvious, and they restrict some applications of the developed devices. First of all, this refers to complexity of integrating photosensor arrays with microelectronic readout integrated circuits (ROIC), which restricts the format and the minimum size of photosensitive elements; high cost of the starting semiconductor materials; the need to use cryogenic operating temperatures; and sophisticated and expensive manufacturing equipment, resulting in high cost of the final product. This stimulates the search for alternative approaches to the production of components for optoelectronics, new photosensitive materials and techniques for multi-element photosensors and photosensor arrays operating in ultraviolet and infrared spectral ranges.

The last decades witness the rapid increase in the number of studies directed towards the design of photosensors using low-dimensional materials and structures. Examples are structures based on graphene and graphene-like and related 2D nanomaterials, non-graphene 2D materials composed of Group IV, V, VI and IIIa elements, nano-sized layers of topological insulators like bismuth telluride, etc[12-15]. The design of photosensors based on type-II superlattices (T2SL) is being successfully developed. The T2SL structures are alternating nano-sized layers with different band-gap semiconductors (InAsSb/InSb, InGaAs/GaAsSb, InAs/GaSb, PbTe/PbS, PbTe/ SnTe)[16]. The properties of T2SL structures often eliminate some drawbacks inherent in solid-state semiconductor photosensors based on 3D single crystals and epitaxial layers, e.g., increase the sensor operating temperature owing to significantly lower dark currents.

The last two decades witnessed the appearance a new trend in the design of photosensors, particularly, the use of so-called colloidal quantum dots (CQDs), the spectral properties of which are specified by the average size of semiconductor nanoparticles, which is varied in the 2 – 50 nm range. Meanwhile, the width of the luminescence spectrum of a single nanoparticle at room temperature is only 20 – 30 nm. Thus, quite a small set of luminescent quantum dots based on A^IIB^VI, A^IIIB^V or A^IVB^VI type semiconductors easily cover the spectral range from 350 to 3500 nm owing to variation of their size[17, 18].

The size dependence of the absorption and luminescence spectrum properties of ultradisperse semiconductors was first demonstrated in 1981 by A.I.Ekimov and A.A.Onushchenko from the S. I.Vavilov State Optical Institute, in relation to CuCl nanoparticles grown in multicomponent silicate glasses during thermally initiated diffusion phase decomposition of a supersaturated solid solution, and theoretically substantiated by Al.L.Efros and A.L.Efros[19, 20]. Independently, L.E.Brus the same dependence for a CdS dispersion[21, 22]. Over the following several years, this effect was experimentally detected and confirmed for ultradisperse particles of a number of semiconductors such as CdS, CuBr, CdSe, PbS, ZnS, Zn₃P₂, and Cd₃P₂[23-28]. The term ‘quantum dot’ (QD) was first introduced in 1988 by Reed et al.[29], who studied the electron transport in quasi-zero-dimensional semiconductor nanocrystals and showed that an electron behaves in a nanocrystal as in a three-dimensional potential well. When the nanocrystal size becomes smaller than the exciton Bohr radius, the continuous spectrum of allowed charge carrier energies (characteristic of the conduction and valence bands of bulk crystals) is replaced by a discrete spectrum. The size of such nanocrystals consisting of 10³ – 10⁵ atoms and called quantum dots was found to range from 2 to 50 nm. A similar size quantization of the energy spectrum was also observed for nanocrystals formed as nanoparticle dispersions in organic or aqueous solutions. These nanocrystals are called colloidal quantum dots.

In 1993, Murray, Norris and Bawendi[30] proposed a facile and highly efficient chemical method for the synthesis of quantum dots, which was called high-temperature colloidal synthesis. This method not only opened up new opportunities for studying CQDs, but also clarified the real prospects for their practical applications.

As a recognition of the significance of this research area and priority of the pioneering studies that initiated a new field of photonics that actively develops now, Louis Bruce, Alexey Ekimov and Alexander Efros were awarded the R.W.Wood Prize in 2006, while in 2008, Louis Bruce was awarded the Fred Kavli Prize in nanoscience. In 2023, Moungi Bawendi, Louis Bruce and Alexey Ekimov won the Nobel Prize in Chemistry.

The CQD-based approach to the fabrication of infrared photodetector arrays, which has actively developed in recent years, markedly simplifies the technology, reduces the pixel pitch limitation for the photosensitive elements and considerably reduces the cost of photodetector arrays[31, 32]. The vigorous development of this area in the last few years resulted in the fabrication of the first commercial imagers and cameras based on them, which operate in the spectral range from 0.4 to 2.1 μm[33-36].

A number of reviews published over the past few years generally address the use of CQDs for a wide range of applications[37, 38]. In 2023, Mamiyev and Balayeva[39] considered various applications of PbS-based CQDs. The key achievements in the synthesis of PbS CQDs and the fabrication of thin films based on them are addressed by W.M.M.Lin et al[40].

In addition, there are a number of specialized reviews devoted to particular issues of photosensor applications of CQDs. For example, V.Pejovic et al.[41] analyzed the advances in the CQD image sensor technology as of June, 2022. The protosensor applications of PbSe and PbS CQDs are addressed by M.C.Gupta et al.[42] and H.Wu and Z.Ning[43].

Ponomarenko et al.[44] analyzed the architectures of photosensitive elements and described the advances in the CQD-based photoelectronics as of the beginning of 2021.

The present review gives the first comprehensive and detailed account of the architectures, fabrication methods and properties of photonic sensors based on colloidal quantum dots of Group II, IV and VI elements. The review describes the synthetic approaches, the influence of ligands and morphology of various CQDs used in photosensors. Functional diagrams are considered for photoresistor, photodiode and phototransistor elements with absorption layers based on HgTe, HgSe, PbS and PbSe colloidal quantum dots for operation in various spectral ranges, including hybrid elements using functional layers made of 2D materials. The parameters of the best optical devices based on colloidal quantum dot structures as of the beginning of 2024 are given. On the basis of publications of the last few years, we analyze the potential and development trends for the photonic sensors composed of colloidal quantum dots of compounds of Group II, IV and VI elements as one of the most promising new areas of quantum photosensorics.

The fabrication of second-generation two-dimensional photodetectors required the development of way for electrical connection (hybridization) of hundreds of thousands and millions of photosensitive elements made of semiconductor materials (Figure 2) with an equal number of microelectronic input devices arranged in a silicon readout integrated circuit (ROIC). This is exemplified in Figure 2, which shows the connection of an indium antimonide photodiode array to silicon ROIC. The development of the hybrid technology started in the second half of the 1970s. This required elaboration of a method for growing indium bumps with size and height variations of approximately fractions of a micrometre; development of special high-precision equipment for the thermocompression bonding of a group of bumps located on a silicon ROIC to a similar group located on an array of photosensitive elements[45]. The above hybridization operation is a key issue for all second-generation photodetector arrays. Using this operation, photonic detector arrays were fabricated, including those of a mega-pixel format, based on InSb (1280 × 1024, 1920 × 1536), InGaAs (1280 × 1024) and CdHgTe (2048 × 2048, 4096 × 4096) photodiodes and multielement AlGaN arrays (320 × 256, 640 × 512)[11, 12, 46-51]. The hybridization technique substantially influences the photosensor quality, since connection defects (bonding disruption between single bumps of the photodiode array and ROIC wafer due to their different heights, high resistance in the contact layer, etc.) give rise to non-functional photosensitive elements.

Figure 2

Hybridization of InSb photodiode array with Si ROIC wafer for signal readout and preliminary processing using indium bumps: (a) hybridized photosensitive elements and Si ROIC wafer; (b) two array elements connected by columnar indium bumps to the ROIC input devices (Т1 transistor sources).

The hybridization operations performed using the above-described indium bumps restrict most appreciably characteristics of photosensor arrays for near and short-wave spectral regions of 0.7 – 3.0 μm. The diameter of columnar indium bumps with minor variations of the diameter and height can hardly be made smaller than 3 – 5 μm. Meanwhile, the maximum pixel pitch for photosensitive elements for arrays operating in the 0.4 – 1.7 μm that provides the best spatial resolution for an electro-optical system is approximately 2 μm. For the mid-wave infrared range (MWIR) of 3 – 5 μm, the pixel pitch should be approximately 4 μm. In the design of CdHgTe-based avalanche photodiodes (APD) in the early 2000s, a hybridization method of APD array with Si ROIC without using columnar or spherical indium bumps was developed; this gave so-called high density vertically integrated photodiodes (HDVIPs)[52]. The architecture of an avalanche p-n-n⁺ photosensitive element based on the CdTe/ HgCdTe pair hybridized by the HDVIP process with silicon readout input is shown in Figure 3[52-54]. The diameter of the vertical via formed by ion etching in CdTe/HgCdTe is approximately 6 μm. However, this hybridization method did not provide the required pixel pitch for photosensitive elements either.

Figure 3

Hybridization of a p-n-n⁺ avalanche photodiode with the input to silicon readout circuit by the HDVIP process without using columnar (or spherical) indium bumps (in relation to the CdTe/ HgCdTe array). The Figure was created by the authors using published data[52-54].

Thus, hybridization of photodiode arrays and signal readout electronic circuits using columnar bumps or via holes (HDVIP method) is not only labour-consuming, requires the use of highly precise and expensive equipment, decreases the percentage yield of acceptable arrays and increases the cost of products, but also limits the pixel pitch of photosensitive elements for arrays not smaller than 3 – 5 μm, which has an adverse effect on the spatial resolution of optoelectronic devices and complexes using such arrays.

Unlike the thermal imaging devices operating in the mid (3 – 5 μm) and far (8 – 12 μm) spectral regions and recording the intrinsic radiation of items heated to 240 – 970 K, the devices operating in the 0.7 – 1.0 μm and 1.0 – 3.0 μm ranges record the intrinsic radiation of items heated to temperatures above 970 K and reflected radiation of, first of all, natural sources (Sun, Moon, stars, natural radiation of the night sky). The images obtained using special instruments in these spectral ranges by measuring the radiation reflected from an observed object have shadows and half-tones and thus allow highly accurate recognition of these objects or parts of objects that have no temperature contrast. This is illustrated by Figure 4, demonstrating that only visible (0.38 – 0.72 μm) and short-wave IR (1.0 – 3.0 μm) images make it possible to confidently recognize the facial features and, hence, identify a person[55]. Due to the inevitable differences in the reflectivity of different fabrics and dyes, the range of 1.0 – 3.0 μm makes it possible to distinguish between camouflage outfits of different countries and thus to make the friend-or-foe identification. This indicates the relevance of development of advanced photosensors providing high-quality images in the NIR and SWIR ranges of electromagnetic radiation. The major part of the visible and near-IR ranges is covered by night vision devices (NVD); the photosensors of these devices are represented by generation III and III+ vacuum electrical to optical converters with GaAs-based photocathodes with a negative electron affinity and cut-off wavelength of not more than 0.93 μm and quantum efficiency of approximately 30% (Figure 5, curve 6)[56]. The Sb – Rb – Cs or Sb-K – Cs bialkali photocathodes (e.g., K₂CsSb and Rb₂CsSb), photocathodes based on GaAsP ternary solid solution, and Na-K – Sb – Cs multi-alkali photocathodes [e.g., (Cs)Na₃KSb] (Figure 5, curves 3, 4, 5) have a higher quantum efficiency (30 – 55%), but their spectral sensitivity is limited to the visible and near-infrared spectral regions. The same is true for the super bialkali (SBA) and ultra bialkali (UBA) type photocathodes based on bialkali compounds manufactured by a specific technique resulting in an increase in the absolute sensitivity, but not in an extension of the spectral range[57]. The efforts of the last decades on the search for photocathodes with a cut-off wavelength extended to the short-wave IR (SWIR) region resulted in the development of new-generation photocathodes based on InGaAs and InGaAsP solid solutions and рhotocathodes based on InP/InGaAsP and InGaP/InGaAs heterojunctions with cut-off wavelengths at about 1.6 μm. The greatest progress was attained for photocathodes with InP/InGaAs heterojunction (Figure 5, curve 9). However, the wide use of such devices in night vision technology is hindered by their low quantum efficiency, not exceeding a few percent, and high cost.

Figure 4

Facial images of the same person in different spectral ranges: 0.4–0.7 µm (visible), 1.0–3.0 µm (SWIR), 3.0–5.0 µm (MWIR) and 8.0–14.0 µm (LWIR)[55].

Figure 5

Photosensitivity spectra of some photocathodes used in modern EOCs[56].

The key NVD parameters (range of vision, probability of detection and recognition, noise protection and noise resistance, power consumption, etc.) depend on the external conditions under which NVD is used. The key external conditions of NVD operation (in passive systems without artificial light source) include

— spectral composition and level of the natural night illumination (NNI);
— atmospheric parameters along the observation track;
— atmospheric haze;
— brightness coefficients of natural entities and observation targets.

The natural night illumination of the Earth’s surface at different times of the day, in various regions, at different weather conditions, etc. has been investigated for many years. Generally, the results of these studies can be summarized as follows[58-60].

The major source of NNI is moonlight, which provides spectral irradiance of 1.8 × 10^–3 W m^–2 μm^–1 at λ ≈ 0.7 μm near the Earth’s surface; for the wavelength λ ≈ 1.6 μm, this value decreases to approximately 1.2 × 10^–3 W m^–2 μm^–1 (Figure 6a)[58]. The spectral irradiance generated by stars in the visible region in the wavelength range of 1.4 – 1.8 μm is much lower. On a clear moonless night, the main natural sources of illumination are the airglow, starlight and zodiacal light. The main contribution to the ground illumination is made by airglow (Figure 6b)[58]. The night airglow is caused by various processes taking place in the upper atmospheric layers: recombination of ions formed upon рhotoionization during daytime under the influence of the solar radiation; reactions between oxygen, nitrogen and hydroxyl OH– accompanied by photon emission; luminescence caused by the passage of cosmic rays in the upper atmosphere; reactions involving lithium and sodium; and combination reactions between nitrogen and oxygen atoms accompanied by photon emission.

Figure 6

Spectral irradiance of natural night illumination: (a) with a full Moon; (b) airglow in a moonless night[58].

Comparison of these characteristics for the 0.4 – 0.9 μm and 1.4 – 1.8 μm spectral ranges shows that[61]

(a) the spectral irradiance from NNI regularly increases on going from the visible [from λ ≈ 0.38 μm (violet) to 0.78 μm (red)] to IR spectral region (λ ≥ 0.9 μm); indeed, in the absence of moonlight, the average spectral irradiance for λ = 0.6 – 0.8 μm is (1.5 – 3.0) × 10^–5 W m^–2 μm^–1, while in the 1.4 – 1.8 μm wavelength range, it increases to (6.0 – 7.0) × 10^–4 W m^–2 μm^–1;
(b) air transparency is much higher in the near-IR range than in the visible range; for example, for the meteorological visibility SM = 10 km, transmittance of a 1 km thick atmospheric layer is 0.72 at λ = 0.6 μm and 0.93 at the centre of the transparency window (1.4 – 1.8 μm);
(c) brightness of the atmospheric haze decreases more than 10-fold in the 1.4 – 1.8 μm wavelength range compared with thee visible region;
(d) NNI changes during the night from 10^–1 W m^–2 to 2.5 × 10^–5 W m^–2 (i.e., by almost four orders of magnitude) in the 0.4 – 0.9 μm spectral range, while in the 1.4 – 1.8 μm range, this change is from 1.6 × 10^–1 W m^–2 to (3 – 4) × 10^–3 W m^–2; (e) the contrast (observation target/background) is 1.4 – 1.5 times higher in the 1.4 – 1.8 μm wavelength range than in the 0.4 – 0.9 μm range.

A comparative estimation of the ranges of vision \( L_1 \) and \( L_2 \) for two NVDs with identical optical systems, but different spectral sensitivity ranges was reported by Gusarova et al[61]. The \( L_1/L_2 \) ratio was calculated by the formula:

where \( L_1 \) is the distance of action of NVD at which the detection of an object ensures recognition, \( K_1 \) is contrast, \( \tau_{a\lambda} \) is the spectral atmospheric transmittance between the target and the observer, \( S_{1\lambda} \) is the relative spectral sensitivity of a conventional detector operating in the 1.4 – 1.7 μm range, equal to \( S_{1\lambda}/S_{1\lambda max} \), \( E_{\lambda} \) is the spectral irradiance on the target surface; \( L_2 \), \( K_2 \), \( S_{2\lambda} \) are the same parameters for EOC with the sensitivity range of 0.4 – 0.9 μm. The estimated \( L_1/L_2 \) ratios indicate that for a low illumination level on a horizontal track, \(E_H = 1.4 \times 10^{–3} lux \) and meteorological visibility \(S_M = 2.5 km \), the \( L_1/L_2 \) ratio is 13, while for \(S_M = 20 km \), \( L_1/L_2 = 8 \) .

This estimate convincingly demonstrates the considerable advantage of photonic sensor arrays with cut-off wavelength λ_c of approximately 1.8 – 2.0 μm for night vision devices. Meanwhile, the manufacture of these devices all over the world is currently based on vacuum electrical to optical converters operating in the range of 0.4 – 0.9 μm.

The solid-state photonic sensors with cut-off wavelength λ_c ≈ 1.8 μm are now largely manufactured using semiconductor solid solutions In₀₅₃Ga₀₄₇As as epitaxial layers grown on InP substrates for which there exist a technique for the fabrication of photodiodes with a photosensitive element pitch of 12 – 15 μm and an assembly technique with signal readout and processing integrated circuits made of silicon by means of indium bumps[11, 12, 48, 49]. Due to the active optical absorption of visible light, the InP substrate restricts the short-wavelength cut-off to approximately λ ≈ 0.9 μm. In order to extend the photosensitivity range to the visible part of the spectrum, various methods are used for thinning or eliminating the InP substrate[62-64]. This may provide a photosensor quantum efficiency of approximately 60% at λ = 0.5 μm[64]. In recent years, considerable progress was also attained in the development of methods for reducing the pixel pitch for InGaAs-based photodiode arrays by switching from the indium bumps to copper-based planar microcontacts (cupper-to-copper microcontacts), the basic diagram for which is shown in Figure 7[48, 64-68]. However, this type of hybridization does not eliminate the above-noted drawbacks of photosensor arrays based on 3D materials, related to the necessity to use highly precise and expensive equipment and high labour consumption, resulting in high cost of products.

Figure 7

Hybridization of a 1280×1024 photodiode array based on InGaAs solid solution semiconductor with Si ROIC: (a) using indium bumps (17 µm pitch); (b) using planar Cu–Cu microcontacts (5 µm pitch); (c) TEM image of planar Cu–Cu microcontacts (3 µm pitch). The Figure was created by the authors using published data[48, 64, 66].

Quite a few recent publications are devoted to the design of photosensitive elements operating in the near and short-wave IR regions using CQD-based low-dimensional structures[69-74]. The use of CQD-based structures as the element base of photosensorics provides a number of characteristics that are of fundamental importance for the key technology of photodetector arrays. The most significant of these characteristics are the following:

— CQD-based arrays are completely free from In or Cu metal microcontacts, usually utilized for connecting the photodiode arrays to silicon multiplexers; CQD-based photosensitive layers can be formed by spin-coating, jet or aerosol printing at moderate temperatures directly on the surface of input devices of silicon ROIC (Figure 8)[75];
— currently, there are mega-pixel (2.1 MP) CQD PbS array formats amounting to 1920 × 1080 at a 15 μm pitch[76]; however, the pixel pitch and format are restricted only by those of the silicon multiplexor; therefore, the fabrication of supermegapixel devices is possible in the future;
— the two smallest pixel pitches known for SWIR sensors (1.62 and 1.82 μm) were implemented in ROIC used with CQD-based photodiodes in which there are no limitations related to the minimum size of indium bumps[77, 78]; — a 640 × 512 array based on HgTe CQDs with a pitch of 15 μm for the SWIR spectral range is markedly less expensive than the same array based on the epitaxial InGaAs heterostructure[32];
— wide range of spectral sensitivity Sl(l) of PbS nanocrystal; furthermore, the right edge of Sl(l) dependence can be controlled by varying the quantum dot size via variation of conditions of CQD synthesis[69, 79].

The first attempts of using PbS nanocrystals as quantum dots for recording electromagnetic radiation in the wavelength range of 0.975 – 1.3 μm were undertaken in 2004 and reported by McDonald et al[80].

The authors[80] described photoresistors in which PbS nanocrystals distributed throughout the MEH-PPV polymer served as the photosensitive medium. The photoconductivity spectra of the MEH-PPV : PbS CQD nanocomposites for variable PbS CQD size showed maxima at 0.98, 1.2 and 1.355 μm wavelengths. The photosensitive layer was deposited on an electrode coated by indium tin oxide (ITO). McDonald et al.[81] described the first photodiode with the ITO/PPV/PbS CQD : MEH-PPV/Mg architecture. In 2009, McDonald et al.[82] described the architecture of a photosensitive element obtained by deposition of 256 × 256 pixel array on Si ROIC from a solution. The photosensitive element consisted of photodiodes based on bulk heterojunctions produced by mixing PbS colloidal quantum dots with two organic compounds, poly(3-hexylthiophene-2,5-diyl) (P3HT) and phenyl-C61-butyric acid methyl ester (PCBM) (Figure 9)[82]. A better contact between the photosensitive layer and the ITO electrode was attained by using an interlayer based on the poly(3,4-ethylenedioxythiophene) (PEDOT) : poly(styrene sulfonic acid) (PSS) composite. Since then, lead sulfide colloidal quantum dots have been recognized as one of the most promising materials of IR photosensorics.

Figure 9

(a) Architecture of a photosensitive element based on the PbS CQD:P3HT:PCBM bulk heterojunctions; (b) structure of PCBM (phenyl-C61-butyric acid methyl ester)[82].

Structure of MEH-PPV

Structure of P3HT

Structure of PEDOT:PSS

Lately, two architectures have been most popular for the development and fabrication of photodetector arrays based on PbS colloidal quantum dots, particularly, one using the energy barrier of the PbS CQD/C₆₀ heterojunction (Figure 10, Figure 11)[75, 83-89] and one using the energy barrier across the p-PbS CQDs/n-TiO₂ contact (Figure 12)[90-93]. The former architecture was used to design 640 × 512, 1280 × 1024 and 1920 × 1080 photodiode arrays with an extended sensitivity range from 0.35 to 2.1 μm and photosensitive element pitch of 15 μm, which are commercially distributed[76, 83-89]. The latter architecture was used to design image sensors with 640 × 480, 640 × 512 and 768 × 512 formats and a reduced pixel pitch of 5 μm[90]. Using this architecture, photodetectors with very small pixel pitches of 2.1 and 1.82 μm were studied[34, 91-93]. A pixel pitch of 1.62 μm was attained for photosensitive elements based on PbS CQDs with microlenses and copper reflective coatings[78].

Figure 10

Photosensitive element with an energy barrier at the PbS CQD/fullerene C₆₀ contact: (a) architecture of a photosensitive element of 1 mm in diameter; (b) energy diagram; (c) current–voltage characteristic of the structure: in the dark (dark) and under illumination (light) from a source with a spectral composition and power of 100 mW×cm^–2 corresponding to solar radiation[75, 83-89].

Figure 8

Diagram of integration of photosensitive element array based on CQD barrier structure to Si ROIC with the top contact/С₆₀/CQD PbS/bottom contact architecture: (a) photodiode array hybridized with Si ROIC; (b) two С₆₀/CQD elements electrically connected to the input devices of ROIC (Т1 transistor sources). The energy barrier is formed at the С₆₀/PbS CQD contact. The Figure was created by the authors using published data[75, 88].

Figure 11

Architecture of a 640×512 photosensor array with a 15 µm pitch based on PbS CQDs.

The architecture of a single photosensitive element with a PbS CQD/C₆₀ heterojunction energy barrier is shown in Figure 10a. This is a sequence of nanolayers ITO (top contact)/MoO₃ (EBL)/PbS CQD/C₆₀/BCP (HBL)/bottom contact where MoO₃ serves as the electron blocking layer (EBL). In the CQD and fullerene C₆₀ layers, photons of electromagnetic radiation are absorbed, excitons appear and decay to give electron – hole pairs, which are separated at the boundary between PbS CQDs and C₆₀. The electrons are transported through the C₆₀ layer, pass through the hole blocking layer (HBL) and are collected by the aluminium bottom contact. 2,9-Dimethyl-4,7-diphenyl-1,10-phenanthroline [bathocuprine (BCP)] acts as HBL. Similarly, holes are transported through the PbS CQD layer, pass through the EBL layer and are collected at the indium tin oxide (ITO) (In₂O₃)_0.9(SnO₂)_0.1 top contact. The energy diagram of this structure and the current – voltage characteristic (CVC) (in the dark and under illumination from a radiation source with a spectral composition corresponding to solar radiation) are shown in

Figure 10

Photosensitive element with an energy barrier at the PbS CQD/fullerene C₆₀ contact: (a) architecture of a photosensitive element of 1 mm in diameter; (b) energy diagram; (c) current–voltage characteristic of the structure: in the dark (dark) and under illumination (light) from a source with a spectral composition and power of 100 mW×cm^–2 corresponding to solar radiation[75, 83-89].