PbE (E = S, Se) Colloidal Quantum Dot-Layered 2D Material Hybrid Photodetectors

Hybrid lead chalcogenide (PbE) (E = S, Se) quantum dot (QD)-layered 2D systems are an emerging class of photodetectors with unique potential to expand the range of current technologies and easily integrate into current complementary metal-oxide-semiconductor (CMOS)-compatible architectures. Herein, we review recent advancements in hybrid PbE QD-layered 2D photodetectors and place them in the context of key findings from studies of charge transport in layered 2D materials and QD films that provide lessons to be applied to the hybrid system. Photodetectors utilizing a range of layered 2D materials including graphene and transition metal dichalcogenides sensitized with PbE QDs in various device architectures are presented. Figures of merit such as responsivity (R) and detectivity (D*) are reviewed for a multitude of devices in order to compare detector performance. Finally, a look to the future considers possible avenues for future device development, including potential new materials and device treatment/fabrication options.


Introduction and Background
Lead chalcogenide (PbE, E = S, Se) quantum dots (QDs) are desirable materials for implementation in photodetectors because of their potential for improving upon current technologies in the near-infrared (near-IR) and mid-IR spectral regions [1][2][3][4]. Improving capabilities in this region is of great consequence for applications in environmental monitoring [5][6][7], motion sensing [8], fiber-optic communications [9,10], X-ray detection [11], and biological imaging [12], where Pb-based QDs have been shown to be less destructive to cells than there Cd-based counterparts when coated with silica. PbE QDs have size-tunable bandgaps from 800 to 4000 nm (1.55-0.31 eV) [13,14], as shown in Figure 1A, and their solution processability allows for facile integration onto both rigid and flexible substrates through spin-coating, dip-coating, and ink-jet printing [4,15]. The ease with which QD solutions can be integrated, as sensitizers for the near-IR wavelength regions into current silicon based technologies, which are based on complementary metal-oxide-semiconductor (CMOS) architectures make them ideal candidates for further study [16]. Along with the profound effect size has on the properties of QDs, altering the surface chemistry via ligand exchange or other post-deposition treatments can also greatly alter the physical and electronic properties of QD films [17][18][19][20][21][22][23], affecting properties such as carrier type and the smoothness of the QD film. Quality lead sulfide (PbS) QD photodetectors were initially reported in 2005 by Sargent, et al. [24]; during the years since then many efforts have been made to understand, and improve QD devices [17,[25][26][27][28][29].
Successfully absorbing incident photons and collecting the photogenerated carriers from the QD film is paramount to fabricating QD-based photodetectors. PbE QDs are strong absorbers that efficiently However, these values pale in comparison to those of layered 2D materials such as graphene, which can have carrier mobilities of up to 15,000 cm 2 /Vs [69]. Layered 2D materials have been of interest for optoelectronic devices due to their promising characteristics and physical versatility, since the successful isolation, through micromechanical cleavage, of a single graphene layer in 2004 [70]. Layered materials have strong in-plane bonding but weak layer-to-layer bonding through van der Waals interactions, enabling cleavage or exfoliation into two-dimensional layers of single unit cell thickness [71,72]. Graphene is an especially interesting material for broadband photodetection purposes because it has zero band gap, meaning that it absorbs light over a wide range of photon energies, from the ultraviolet (UV) to far-infrared [73][74][75][76][77]. Additionally, graphene displays ultrafast carrier dynamics [78][79][80], tunable optical properties via electrostatic doping [81][82][83], and high carrier mobility [84][85][86][87]. The high carrier mobility enables ultrafast conversion of photons to electrical currents or voltages [88,89]. However, the lack of a bandgap in graphene makes it difficult to fabricate devices with low dark current and high on/off ratios because of the presence of free carriers. Techniques such as fabricating graphene nanoribbons via nanostructuring [90][91][92], nanopatterning to create a graphene nanomesh [93,94], and chemical functionalization [95][96][97][98] have been used to engineer a bandgap in graphene [99]; however, these processes often lead to diminished mobility and add unwanted steps and cost to the fabrication process.
Although graphene has been at the forefront of layered 2D materials research, transition metal dichalcogenides (TMDs) and other layered 2D materials such as phosphorene [a single 2D layer of black phosphorous (BP)] have also shown promise for potential applications in which a non-zero bandgap is advantageous [72,100]. Phosphorene has a unique 2D structure (see Figure 2B), which causes it to display anisotropic carrier mobility [101], meaning that mobility within the plane is dependent on the direction in which the carrier is traveling. Phosphorene is also highly unique in that it displays p-type characteristics [101], and a high work function [102] making it a valuable material for hole transport and hole injection [103]. Figure 2 shows the energy band structures of graphene, phosphorene, and molebdynum disulfide (MoS2), a TMD that is one of several that display a transition from indirect to direct bandgap as the sample is reduced from bulk to single monolayers Reproduced with permission from [13]. Copyright American Chemical Society, 2004. (B) Illustration of a lead sulfide (PbS) QD with various types of ligands (left), and the corresponding (by color) conduction and valence band energy levels each ligand produces (right). The positions of the valence and conduction bands of PbS QDs with different ligands are presented in the graph to the right. Beginning with Br -(brown color, lowest band positions) moving clockwise around the schematic of the QD until it reaches benzenethiol (red color, highest band positions). Reproduced with permission from [18]. Copyright American Chemical Society, 2014. (C) Transfer characteristics of a 5.9 nm PbSe QD film after solid-state exchange with sodium selenide (Na 2 Se) (black), which removes long oleate ligands and enriches the surface in Se, and subsequently upon lead chloride (PbCl 2 ) treatment for durations of 1 h (blue), 6 h (green) and 12 h (red) at 65 • C, which enriches the surface in metal. I D , drain current; V G , gate voltage; V DS , drain-source voltage. Reproduced with permission from [64]. Copyright American Chemical Society, 2014.
However, these values pale in comparison to those of layered 2D materials such as graphene, which can have carrier mobilities of up to 15,000 cm 2 /Vs [69]. Layered 2D materials have been of interest for optoelectronic devices due to their promising characteristics and physical versatility, since the successful isolation, through micromechanical cleavage, of a single graphene layer in 2004 [70]. Layered materials have strong in-plane bonding but weak layer-to-layer bonding through van der Waals interactions, enabling cleavage or exfoliation into two-dimensional layers of single unit cell thickness [71,72]. Graphene is an especially interesting material for broadband photodetection purposes because it has zero band gap, meaning that it absorbs light over a wide range of photon energies, from the ultraviolet (UV) to far-infrared [73][74][75][76][77]. Additionally, graphene displays ultrafast carrier dynamics [78][79][80], tunable optical properties via electrostatic doping [81][82][83], and high carrier mobility [84][85][86][87]. The high carrier mobility enables ultrafast conversion of photons to electrical currents or voltages [88,89]. However, the lack of a bandgap in graphene makes it difficult to fabricate devices with low dark current and high on/off ratios because of the presence of free carriers. Techniques such as fabricating graphene nanoribbons via nanostructuring [90][91][92], nanopatterning to create a graphene nanomesh [93,94], and chemical functionalization [95][96][97][98] have been used to engineer a bandgap in graphene [99]; however, these processes often lead to diminished mobility and add unwanted steps and cost to the fabrication process.
Although graphene has been at the forefront of layered 2D materials research, transition metal dichalcogenides (TMDs) and other layered 2D materials such as phosphorene [a single 2D layer of black phosphorous (BP)] have also shown promise for potential applications in which a non-zero bandgap is advantageous [72,100]. Phosphorene has a unique 2D structure (see Figure 2B), which causes it to display anisotropic carrier mobility [101], meaning that mobility within the plane is dependent on the direction in which the carrier is traveling. Phosphorene is also highly unique in that it displays p-type characteristics [101], and a high work function [102] making it a valuable material for hole transport and hole injection [103]. Figure 2 shows the energy band structures of graphene, phosphorene, and molebdynum disulfide (MoS 2 ), a TMD that is one of several that display a transition from indirect to direct bandgap as the sample is reduced from bulk to single monolayers [104,105], which is key for efficient photon absorption [71]. Photodetectors utilizing layered 2D materials have been widely reported [89,103,[106][107][108][109][110][111][112], as have studies on charge transport and electronic properties of these materials [113][114][115][116][117][118][119][120]; however, the relatively low absorption, attributable to the atomically thin profile [121], remains a serious challenge.  [104,105], which is key for efficient photon absorption [71]. Photodetectors utilizing layered 2D materials have been widely reported [89,103,[106][107][108][109][110][111][112], as have studies on charge transport and electronic properties of these materials [113][114][115][116][117][118][119][120]; however, the relatively low absorption, attributable to the atomically thin profile [121], remains a serious challenge. An attractive option is to combine QDs with layered 2D materials in hybrid devices that allow one to harness many of the desirable features of QDs, including strong, size tunable absorption, long exciton lifetimes and advanced phenomena such as carrier multiplication, while exploiting the excellent charge transport properties of 2D materials. Accordingly, in a typical hybrid QD-2D photodetector, the QD layer acts as the light-absorbing, charge-generating layer, while the 2D layer acts as the transport layer [122], with charge transfer between the layers being a key process for optimization. Such charge transfer affected by various parameters such as ligand length (i.e., QD-2D spacing) and relative band alignment [123], in a similar manner to the role these factors play in charge transfer within QD-only films as discussed above. In this paper, we will briefly review types of photodetectors and their figures of merit and then discuss recent advancements in hybrid QD-layered 2D photodetectors, with emphasis on graphene-QD hybrid and TMD-QD hybrid devices.

Photoconductors
A photoconductor is a two-contact optoelectronic device in which two ohmic source and drain electrodes are separated by a photoactive layer (i.e., a QD film; Figure 3A). Operated under an applied bias, photoconductors detect temporary changes in the carrier mobility, density, or both under incident illumination due to photogeneration of carriers in the photoactive layer [124]. Under a moderate field, the majority carriers (which can be either holes or electrons, depending on the material) have a higher mobility than the minority carriers. This results in the majority carriers having a shorter transit time to traverse the photoactive area to an electrode, while minority carriers remain left behind. If holes, for example, are the majority carrier, as they are swept out of the detector, charge neutrality is maintained by additional holes supplied from the other electrode. Therefore, effectively, holes can circulate the detector many times during the carrier lifetime, resulting in gain (a measure of how much the response departs from linear dependence) [4,125]. Higher gain leads to detectors An attractive option is to combine QDs with layered 2D materials in hybrid devices that allow one to harness many of the desirable features of QDs, including strong, size tunable absorption, long exciton lifetimes and advanced phenomena such as carrier multiplication, while exploiting the excellent charge transport properties of 2D materials. Accordingly, in a typical hybrid QD-2D photodetector, the QD layer acts as the light-absorbing, charge-generating layer, while the 2D layer acts as the transport layer [122], with charge transfer between the layers being a key process for optimization. Such charge transfer affected by various parameters such as ligand length (i.e., QD-2D spacing) and relative band alignment [123], in a similar manner to the role these factors play in charge transfer within QD-only films as discussed above. In this paper, we will briefly review types of photodetectors and their figures of merit and then discuss recent advancements in hybrid QD-layered 2D photodetectors, with emphasis on graphene-QD hybrid and TMD-QD hybrid devices.

Photoconductors
A photoconductor is a two-contact optoelectronic device in which two ohmic source and drain electrodes are separated by a photoactive layer (i.e., a QD film; Figure 3A). Operated under an applied bias, photoconductors detect temporary changes in the carrier mobility, density, or both under incident illumination due to photogeneration of carriers in the photoactive layer [124]. Under a moderate field, the majority carriers (which can be either holes or electrons, depending on the material) have a higher mobility than the minority carriers. This results in the majority carriers having a shorter transit time to traverse the photoactive area to an electrode, while minority carriers remain left behind. If holes, for example, are the majority carrier, as they are swept out of the detector, charge neutrality is maintained by additional holes supplied from the other electrode. Therefore, effectively, holes can circulate the detector many times during the carrier lifetime, resulting in gain (a measure of how much the response departs from linear dependence) [4,125]. Higher gain leads to detectors with high responsivities, but also higher noise levels which hinders high sensitivity applications [8]. Taking advantage of differential carrier mobility also means photoconductors typically have slower response times than a photodiode, due to the temporal response also being determined by the lifetime of the trapped carriers. Photoconductors are more common for IR applications such as thermal imaging [126] and motion detection [127].

Phototransistors
Phototransistors are essentially photoconductors bridged by a "gate" electrode, typically a metal or degenerately doped semiconductor separated by a dielectric spacer layer. The gate electrode provides the ability to modulate transport using an applied gate voltage; a schematic of a simple phototransistor can be seen in Figure 3B. Phototransistors function in a similar manner as field-effect-transistors (FETs) and are even referred to as optical FETs (OFETs) on occasion [27]. Applying a gate bias introduces charges into the conductive channel, which can be used to tune transport by, e.g., filling trap states [27], as in the case for devices based on p-type lead chalcogenide QD films, for which introduction of holes by a negative applied bias increases the conductivity. Phototransistors have become a preferred architecture of the hybrid QD-2D detectors because the level of control offered by the ability to modulate carrier dynamics and concentrations within the device allows for more versatility than a typical photoconductor.

Photodiodes
Photodiodes employ an internal electric field established within the absorbing layer of the device to enhance the efficiency with which photogenerated carriers are separated and subsequently collected. The internal electric field is created by the pairing of materials with majority carriers of opposite charge, which interdiffuse and undergo recombination near the junction to create a depletion region featuring a charge gradient and built-in field. The field favors very fast, unidirectional transport by drift for photogenerated electrons and holes (in opposite directions, respectively), largely preventing recombination. In contrast carriers generated outside the depletion region travel by much slower diffusion to reach either an electrode or the depletion region [125], allowing more time for recombination; accordingly, the size of the depletion region is an important factor in performance. An example of a p-n junction photodiode device structure is presented in Figure 3C. Photodiodes are typically operated under moderate reverse bias to create an even wider depletion region to minimize transit time as much as possible, but not too wide or transit-time effects will limit the response frequency [125]. Photodiodes are often categorized by the types of materials forming the junction (e.g., Schottky junction, p-n junction, p-i-n junction), but the principles of separating and collecting photogenerated charges using a built-in electric field are similar [4], and the relation between device thickness and carrier diffusion length (as determined by carrier lifetime and mobility) is critical to performance. For reference, among QD-based devices, PbE QD solids have exhibited record-high diffusion lengths of up to 230 nm [128]. Photodiodes have a gain of 1 unless operated in avalanche mode, under large reverse biases, where impact ionization and carrier multiplication can result in gains higher than 1. Photodiodes can also be operated under a zero bias condition; however, under reverse bias conditions, the device can have greater bandwidth and wider linear dynamic range. Applications requiring fast response times are heavily reliant on photodiodes [125], as current commercially available photodiodes can have response times as fast as 35 ps. Operation is similar to a photoconductor, but the device is fabricated on a substrate that allows for the possibility of applying a gate voltage [VG] to tune transport in the photoactive area. (C) Schematic of a photodiode device. A p-n junction enhances charge transport by creating an internal electric field. The green shaded region indicates the depletion region, which can be tuned by altering the bias voltage [VB]. Carriers generated in the depletion region are quickly separate and are collected, while carriers generated outside of the depletion region must avoid recombination while they diffuse to either depletion region or contact. All devices are displayed as having a current readout; this signal is typically converted into voltage by a load resistor for easy readout using an oscilloscope.

Figures of Merit
Figures of merit [124,125,129] that account for device variables, such as device active area, response time, and spectral region are key for evaluating and comparing detector performance. The first measure of how effective a photodetector can be is how efficiently incident photons are Operation is similar to a photoconductor, but the device is fabricated on a substrate that allows for the possibility of applying a gate voltage [V G ] to tune transport in the photoactive area. (C) Schematic of a photodiode device. A p-n junction enhances charge transport by creating an internal electric field. The green shaded region indicates the depletion region, which can be tuned by altering the bias voltage [V B ]. Carriers generated in the depletion region are quickly separate and are collected, while carriers generated outside of the depletion region must avoid recombination while they diffuse to either depletion region or contact. All devices are displayed as having a current readout; this signal is typically converted into voltage by a load resistor for easy readout using an oscilloscope. [124,125,129] that account for device variables, such as device active area, response time, and spectral region are key for evaluating and comparing detector performance. The first measure of how effective a photodetector can be is how efficiently incident photons are converted into electron-hole pairs and subsequently collected as photocurrent. The external quantum efficiency (EQE) is a measure of how well the device absorbs and converts incident photons into photocurrent, while the internal quantum efficiency (IQE) is a measure of how efficiently absorbed photons are converted into photocurrent. The incident photon flux (φ in ) can be calculated by dividing the incident power by the energy of the incident photon. Similarly, the absorbed photon flux (φ abs ) can be calculated by multiplying the φ in by the fraction of light that is absorbed [130].

Figures of merit
In the preceding equations, q is the elementary charge and I ph is the photocurrent, which can be calculated by I ph = I illum − I dark , where I illum is the photocurrent under illumination and I dark is the dark photocurrent. The fraction of light absorbed by a QD layer can be calculated with the Beer-Lambert law; for many films this equation can be written in terms of optical thickness (τ), which can be σ i n i , where σ i and n i are the attenuation coefficients and number densities of each attenuating species in the absorbing medium. The absorbance can then be calculated with the relationship A = τ/ ln10. Ellipsometry can also be used to determine refraction indexes and extinction coefficients of materials, which can then be used to create a model for photon absorption [33,131].
The responsivity [R(λ)] of a detector is a measure of the electrical signal output relative to the optical signal input, similar to the EQE. The units of R(λ) are amperes per watt [A/W] and it can be calculated by the following equation [124]: where λ is the wavelength of the incident photons, h is Planck's constant, c is the speed of light, ω is the modulation frequency, τ is the time constant, and G (λ) is the photoconductive gain. Photodiodes have a gain of 1 unless they are operated in avalanche mode, while for photoconductors and phototransistors the gain is equal to the majority carrier lifetime divided by the transit time [G(λ) = τ lifetime /τ transit ]. Transit time can be calculated from the carrier mobility and bias voltage by τ transit = L 2 /(µV bias ) [130]. Noise equivalent power [NEP(λ)] is a measure of the detector sensitivity, defined as the optical power at which the signal-to-noise ratio (SNR) is 1, giving the minimum power the detector can effectively detect per square root of bandwidth [124], and is obtained using: where I n is the total noise current, and NEP (λ) is in units of W/Hz 1/2 . Total noise current is the sum of all the noise sources, which includes low-frequency flicker noise (1/f ), thermal noise (I th ) and shot noise (I sh ). The root-mean-square value of thermal noise and shot noise current is given by I th = √ 4kTB/r and I sh = 2qBI dark , respectively, where k is Boltzmann's constant, T is absolute temperature, r is the resistive element source contributing to noise, B is the noise bandwidth, and I dark is the dark current [3]. The most useful figure of merit is the detectivity [D*(λ)], which accounts for different configurations and detector areas [3], allowing for comparison across all devices. The D*(λ) of a photodetector is expressed in units of Jones (cm Hz 1/2 /W), and is obtained using the following equation: where A is the detector active area (cm 2 ). D*(λ) is proportional to R(λ), and it is also an indirect function of applied bias, temperature, modulation frequency, and wavelength [3,124]. When reporting D*(λ) the value should be accompanied by the measurement conditions to ensure that presented values are both correct and can be properly used for comparison of multiple devices.
The linear dynamic range [DR (λ)] shows the range over which the photocurrent increases linearly with increasing incident optical power and provides for the range of optical power over which detector should be utilized. Ideally, R (λ) should remain constant with the increase of optical intensity. DR (λ) has units of decibels (dB) and is expressed as: where P(λ) max is the incident optical power when the photocurrent saturates [124].

Graphene-QD Hybrid Detectors
As described above, graphene has been an attractive material for photodetection applications since its discovery, due to high carrier mobilities and potential broad spectrum applicability [130], and QDs are of interest because of strong, tunable light absorbance. Sensitizing graphene-based photodetectors with PbE QDs can improve D* of a device by improving absorption in the short-wave IR and visible regions. Photon absorption in a layer of QDs creates electric charges that can then be transferred to the graphene for fast transport. In the case of phototransistors and photoconductors, the high carrier mobility combined with the long carrier lifetimes in the QD layer would potentially allow the carriers to circulate the device many times before recombining, leading to high gains.
In 2012, Konstantatos et al. [122] reported a hybrid graphene/PbS QD phototransistor with ultrahigh gain, sparking much interest in the system and inspiring many variations to be fabricated in the following years. In the seminal device, single and bilayer graphene flakes were prepared by mechanical exfoliation and deposited on Si/SiO 2 substrates, followed by gold contact deposition and, finally, PbS QD deposition via spin-coating. Ligand exchange from oleic acid to EDT was performed during the QD deposition phase through a layer-by-layer approach, with a resulting film thickness of ≈80 nm. A schematic of the device can be seen in Figure 4A. The photo-generated holes in the PbS QD layer are transferred to the graphene layer while the photo-generated electrons remain trapped in the PbS QD layer. Due to the high hole mobility in graphene, these carriers are allowed to circulate in the device, while charge conservation in the graphene channel is enabled by hole replenishment from the source electrode for every hole collected at the drain, resulting in a gain of 10 8 electrons per photon. The vast improvement of the gain led to an R on the order of 10 7 A/W and a D* of 7 × 10 13 Jones. A short gate voltage pulse can be used to purge the charge carriers from the QD layer as shown in Figure 4D, effectively resetting the device and increasing the operating speed, which is advantageous for imaging applications. 9 of 1 × 10 9 A/W with shorter capping ligands such as thioglycerol (TGL) (length 0.5 nm). Ideally, photocurrent as a function of incident photon energy of devices should closely follow the absorption features of the QDs used, as seen in Figure 4G, indicating increased sensitivity in the range where QDs are strongly absorbing. Laser shock imprinting has also been used as a method to improve contact between graphene and QDs in graphene/PbS QD/graphene sandwich structure [134], resulting in improved response time and current on/off ratio.   Later in the same year, Sun et al. [132] used chemical vapor deposition (CVD) rather than mechanical exfoliation to prepare the graphene transport layer. CVD allows for precise control over thickness, as well as the possibility of fabricating larger area devices than what is feasible with mechanically exfoliated graphene. Upon deposition of PbS QDs capped with pyridine ligands onto un-doped graphene, the measured Dirac point shifts from 0 to 50 V, indicating p-type doping of the graphene layer, as can be seen in Figure 4B. Along with this shift in the Dirac point, there is a sizable decrease in electron mobility from 1000 cm 2 /Vs to 440 cm 2 /Vs while the hole mobility remained unchanged at 1000 cm 2 /Vs. Responsivity of the detector was found to be affected by the thickness of the QD layer, with a saturation point occurring at ≈150 nm, as can be seen in Figure 4C. The devices using CVD-grown graphene showed R values of 1 × 10 7 A/W, similar to that of detectors using mechanically exfoliated graphene.
Ligand length has a significant effect on the efficiency of charge transfer from QDs to graphene [133], as can be seen in Figure 4E by the more significant shift in the Dirac point of the graphene/QD system when a shorter capping ligand is utilized, indicating increased coupling between graphene and QDs. Figure 4F shows the responsivity of the QD/graphene system can reach values on the order of 1 × 10 9 A/W with shorter capping ligands such as thioglycerol (TGL) (length 0.5 nm). Ideally, photocurrent as a function of incident photon energy of devices should closely follow the absorption features of the QDs used, as seen in Figure 4G, indicating increased sensitivity in the range where QDs are strongly absorbing. Laser shock imprinting has also been used as a method to improve contact between graphene and QDs in graphene/PbS QD/graphene sandwich structure [134], resulting in improved response time and current on/off ratio.
Multi-heterojunction phototransistors synthesized by spin-coating alternating graphene and PbSe QD layers show the importance of using graphene as the bottom layer of the device [135], with graphene bottom layer device displaying electron and hole mobilities of µ E = 147 cm 2 /Vs and µ H = 137 cm 2 /Vs, while QD bottom layer devices showed µ E = 14 cm 2 /Vs and µ H = 59 cm 2 /Vs. Intercalation of graphene layers within a PbS QD film also improves the charge carrier extraction of the device by counteracting the limitation of diffusion length in QD films. Placing graphene layers, separated by a distance less than the carrier diffusion length of the QD film, periodically through a QD/graphene film results in higher photocurrents than devices with only a bottom graphene layer [136].
Electrohydrodynamic nanoprinting of colloidal PbS QDs onto graphene FETs with varying quantum dot layer thicknesses is a potential method for realizing small footprint detectors with high spatial resolution [137]. The responsivity of the photodetectors increases with increasing layer thicknesses up to 130 nm. However, the noise current is found to be independent of the layer thickness. Additionally, responsivity and noise current are both linearly dependent on the applied drain voltage and drain current. As a result, the specific detectivity is independent of the drain voltage, and the detector can be operated at lower drain voltage thus reducing power consumption. D* values of at least 10 9 Jones are reported without degradation of the charge carrier mobilities in graphene from the electrohydrodynamic printing of QDs [137].
Response time of phototransistors with thicker QD films (>100 nm) is still regulated by the diffusion of carriers through the QD sensitizing layer. However, this can be overcome by a device architecture that combines a graphene-colloidal QD photodiode and a high-gain phototransistor; a schematic of the devices can be seen in Figure 5A. Transforming the electrically passive sensitizing layer to an active one through an applied electric field in the photodiode significantly enhances charge collection, due to carrier drift instead of solely relying on diffusion [138]. The QD photodiode consists of a top-contact (e.g., indium tin oxide (ITO)) acting as the cathode of the QD photodiode, whereas graphene acts as the hole acceptor contact and the charge transport channel for the phototransistor. Progressively increasing the bias voltage causes the depletion region to grow, enhancing efficiency of the charge collection ( Figure 5D). The hybrid device architecture results in a sub-millisecond temporal response, a gain-bandwidth product on the order of 10 8 , a linear dynamic range in excess of 110 dB, and very high sensitivity with experimentally measured D* of 1 × 10 13 Jones.
In addition to the properties described above, graphene's weak electrostatic screening effect, finite density of states and mechanical flexibility [139] makes it a versatile material for conducting electrodes. Due to its relative transparency at wavelengths greater than 1000 nm, compared to ITO [140], there has been a push for using graphene as the transparent conducting electrode in IR detection and imaging applications. Responsivity of graphene-based and ITO-based PbS QD photodiodes from wavelengths of 1100-1800 nm operated at different reverse bias conditions are shown in Figure 5C. The responsivity of graphene-based vs. ITO-based PbS photodiodes are 0.112 and 0.076 A/W at 1530 nm, respectively, and increase to 0.69 and 0.50 A/W at bias of −1 V. Photocurrent response of the devices under IR illumination (1530 nm) was found to increase linearly with light power. Ambipolar vertical phototransistors ( Figure 5B) [141] utilizing graphene as an electrode have been fabricated using both PbS [142] and PbSe [139] QDs. In a vertical phototransistor, the channel length is determined by the film thickness, which is much shorter than a typical lateral phototransistor channel, leading to faster response times. Vertical phototransistors utilizing graphene as source electrode with PbS QDs displayed temporal response times of 14 ms, which improves to 8 ms when graphene is mixed within the QD layer, as seen in Figure 5E,F, while PbSe QD detectors showed a response time of 7 ms.

Merging QDs with TMDs and Other Layered 2D Materials
While graphene-QD hybrid detectors display high gain, they suffer from high dark currents due to the semi-metallic nature of graphene. Transition metal dichalcogenides (TMDs) have relatively large bandgaps (1-2.5 eV) [72], which make them interesting candidates for applications that require high sensitivity. Using 2D TMDs rather than graphene presents a trade-off in carrier mobility and ultimately gains in exchange for lower dark conductivities. In 2015 Kufer et al. [143] published the first hybrid MoS2-PbS QD photodetector, in which micromechanically exfoliated MoS2 nanosheets  Figure (A,D) reproduced with permission from [138]. Copyright Springer Nature, 2016. Figure (B) reproduced with permission from [141]. Copyright AIP Publishing, 2016. Figure (C) reproduced with permission from [140]. Copyright AIP Publishing, 2011. Figure (E,F) reproduced with permission from [142]. Copyright American Chemical Society, 2017.

Merging QDs with TMDs and Other Layered 2D Materials
While graphene-QD hybrid detectors display high gain, they suffer from high dark currents due to the semi-metallic nature of graphene. Transition metal dichalcogenides (TMDs) have relatively large bandgaps (1-2.5 eV) [72], which make them interesting candidates for applications that require high sensitivity. Using 2D TMDs rather than graphene presents a trade-off in carrier mobility and ultimately gains in exchange for lower dark conductivities. In 2015 Kufer et al. [143] published the first hybrid MoS 2 -PbS QD photodetector, in which micromechanically exfoliated MoS 2 nanosheets were used as electrically controllable transport layers, resulting in responsivities on the order of 6 × 10 5 A/W. D* of bilayer and few-layer devices were found to be 2 × 10 11 and 5 × 10 11 Jones, respectively. At high negative back-gate bias the MoS 2 channel is depleted from free carriers in the dark state, giving the detector the potential to reach high sensitivity in the shot noise limit with D* shot-noise limit reaching up to 7 × 10 14 Jones at V G of −100 V ( Figure 6A). Application of a semiconducting titanium(IV) oxide (TiO 2 ) buffer layer at the interface of MoS 2 and PbS QDs, as presented in Figure 6B, preserves the gate modulation by suppressing the high density of localized sub-band-gap states that pin the Fermi level [144]. The maintained gate control over carrier density in the conduction channel allows for low noise operation similar to pristine MoS 2 devices, resulting in a D* of 5 × 10 12 Jones, an improvement of more than 1 order of magnitude compared to MoS 2 /PbS devices without a buffer layer [144]. Figure 6C shows the responsivity as a function of irradiance as well as the decay time of MoS 2 /PbS devices when a TiO 2 buffer layer is utilized.
Applying methods that were previously used in QD-only devices to provide a built-in p-n junction via energy level modification through ligand engineering can also have positive impacts on hybrid devices. The combination of tetrabutylammonium iodide (TBAI) and EDT is a well-known ligand combination used to create such a built-in potential in QD photovoltaics, resulting in more efficient charge carrier separation in the QD layer. Combining a layer of EDT-treated PbS QD with a layer of TBAI-treated PbS QD along with and MoS 2 transport layer in a vertical phototransistor, as shown in Figure 6D, resulted in fast response times (960 µs), and D* on the order of 10 11 Jones under applied gate voltage of −40 V [145].
The range of TMD materials used in hybrid PbE QD photodetectors has also been expanded to include tungsten disulfide (WS 2 ) [146] and tungsten diselenide (WSe 2 ) [147] which show higher carrier mobilities than MoS 2 . Phototransistors fabricated using CVD to fabricate a p-type WSe 2 monolayer on Si/SiO 2 substrate, coated with PbS QDs with TBAI ligands, displayed rise times of 7 ms and a decay time of 480 ms. The responsivity of hybrid WSe 2 -PbS device could be tuned by the applied gate voltage, but operating in depletion mode was not necessary, which overcomes a drawback for the MoS 2 -QD hybrid device. The highest D* of the device was found to be 7 × 10 13 Jones, with a responsivity of up to 2 × 10 5 A/W, as seen in Figure 6E. WS 2 /EDT-capped PbS QD devices were fabricated in a similar manner resulting in rise and decay times of 153 µs and 226 µs, respectively; however, the responsivity (14 A/W), and D* (3.9 × 10 8 Jones) of the devices were lower than the WSe 2 counterpart. Detectors utilizing both WS 2 and MoS 2 combined with larger (8.0 ± 1.7 nm diameter) PbS QDs, with an absorption peak near 1.8 µm show compelling results with responsivities of 1400 A/W, at 1.8 µm excitation operated at 1 V bias, and D* as high as 10 12 Jones at room temperature for the WS 2 based devices [148]. The devices employing larger dots showed better results using Zn 2 I and mercaptopropionic acid (MPA) ligands, rather than the traditional EDT ligand, as this ligand combination has been shown to produce higher mobilities in larger PbS QD films [149].

12
counterpart. Detectors utilizing both WS2 and MoS2 combined with larger (8.0 ± 1.7 nm diameter) PbS QDs, with an absorption peak near 1.8 µm show compelling results with responsivities of 1400 A/W, at 1.8 µm excitation operated at 1 V bias, and D* as high as 10 12 Jones at room temperature for the WS2 based devices [148]. The devices employing larger dots showed better results using Zn2I and mercaptopropionic acid (MPA) ligands, rather than the traditional EDT ligand, as this ligand combination has been shown to produce higher mobilities in larger PbS QD films [149]. Other layered 2D materials have also been used in hybrid 2D-PbE QD photodetectors. Tin disulfide (SnS 2 )/PbS QD hybrid photodetectors show distinct photoresponse towards photons of different wavelengths [150]. Mechanically exfoliated SnS 2 nanosheets (5 layers) sensitized with EDT-capped PbS QDs and gold electrodes, yielded a broadband, spectrally distinctive photodetector which displays positive photocurrent at wavelengths below 520 nm (the cutoff absorption wavelength of SnS 2 ) and negative photocurrent at wavelengths above 520 nm, as shown in Figure 7A,B. This spectral selectivity is accounted for by illumination-modulated barrier height between the gold electrode and the SnS 2 nanosheet. Upon NIR illumination, only PbS QDs are absorbing incident photons, and photogenerated electrons flow into SnS 2 nanosheets, reversing the p-type doping effect in the dark and shifting the Fermi level of SnS 2 nanosheets upwards. Consequently, the contact between SnS 2 nanosheets and Au electrodes is Schottky in nature, accounting for the observed negative photoconductivity. Under UV illumination, SnS 2 nanosheets absorb incident photons, resulting in carrier density increasing in the SnS 2 channel, which overrides the contribution from photogenerated carriers in the PbS QDs, leading to the observed positive photoconductivity [150].
The 2D black phosphorous (BP) nanosheets, also called "phosphorene", have also recently emerged as a potential candidate for photodetection devices due to high carrier mobilities and anisotropy. Phosphorene has typically suffered from formation of phosphoric acid on the surface under ambient conditions, which causes device performance to degrade. However recent work has shown that treatment with EDT recovers the desirable properties of the device even after degradation [151]. Since EDT is also a common ligand treatment for QD devices as well, Lee et al. decided to combine phosphorene and PbS QDs into a hybrid photodetector, resulting in responsivities of 5.36 × 10 8 A/W and a D* of 1.89 × 10 16 Jones [152]. A hybrid BP/PbS QD photodetector with a cascade-type energy band structure can be fabricated by using ligand chemistry to alter the energy bands of two different layers of PbS QDs. EDT and cetyltrimethylammonium bromide (CTAB) ligands are utilized to form an additional energy barrier at the interface of bilayer QDs. A high responsivity of 1.1 × 10 7 A/W, a D* of 1.75 × 10 15 Jones and a low noise equivalent power of 4.3 × 10 7 pW/Hz 1/2 are achieved at a bias of 1 V without gate voltage modulation, as shown in Figure 7C,D,E, respectively. The responsivity is an order of magnitude higher compared to the phosphorene/PbS photodetector that used only EDT ligands [153]. PbSe QDs have also been integrated with 2D Bi 2 O 2 Se, a 2D material with a relatively narrow bandgap around 0.8 eV, in devices showing impressive responsivities on the order of 10 3 A/W, when excited with 2000 nm excitation and operated under 100 V bias [154]. Table 1 Figure (A,B) reproduced with permission from [150]. Copyright Springer Nature, 2016. Figure (C-E) reproduced with permission from [153]. Copyright Royal Chemical Society, 2019.

Outlook
Hybrid PbE QD-layered 2D photodetectors display great promise for improving detector capabilities, especially in the spectral regions beyond the 1 µm bandgap of Si-based detectors, with much progress made in recent years. However, there are still many avenues for continued progress in improving the figures of merit. There exists a wide and expanding array of studies of PbE QDs aimed at: controlling synthesis to improve size distribution and stability [155][156][157][158]; QD size and composition effects on the optoelectronic properties [39,49,159]; in-solution and layer-by-layer ligand  Figure (A,B) reproduced with permission from [150]. Copyright Springer Nature, 2016. Figure (C-E) reproduced with permission from [153]. Copyright Royal Chemical Society, 2019. Table 1. Hybrid PbE QD-layered 2D photodetectors and their corresponding figures of merit. All detectors presented in this table are phototransistors, with one detector that from [138] utilizing a hybrid phototransistor/photodiode geometry. EDT = ethanedithiol, TBAI = tetrabutylammonium iodide, TGL = thioglycerol, DTG = 2, 3-dimercapto-1-propanol, MPA = mercaptopropionic acid, and CTAB = cetyltrimethylammonium bromide.

Outlook
Hybrid PbE QD-layered 2D photodetectors display great promise for improving detector capabilities, especially in the spectral regions beyond the 1 µm bandgap of Si-based detectors, with much progress made in recent years. However, there are still many avenues for continued progress in improving the figures of merit. There exists a wide and expanding array of studies of PbE QDs aimed at: controlling synthesis to improve size distribution and stability [155][156][157][158]; QD size and composition effects on the optoelectronic properties [39,49,159]; in-solution and layer-by-layer ligand exchange techniques [18,61]; and inter-and intra-layer charge transfer in QD films with other materials [28,45,160]. Many of these findings have yet to be applied to hybrid QD/2D material hybrid devices. Particularly promising directions would involve application of less-commonly used ligand/surface treatments [18,61], expanding the spectral range further into the IR by increasing QD size [13,148], and use of ALD to improve mobility [66,67]. Research and development of new layered 2D materials and techniques for improvement in synthesis of current materials is also an ongoing direction. Strides made in recent years show the great potential of PbE QD-layered 2D hybrid photodetectors on account of facile operation at room temperature, low cost, flexible substrate compatibility, and high figures of merit. Continued progress in the field is possible through continued research, implementation of new materials, surface treatments, and device engineering.
Author Contributions: Conceptualization, investigation, and writing of the original draft was done by T.N.; Supervision, reviewing, editing, and funding acquisition were performed by H.L. and J.P. All authors have read and agree to the published version of the manuscript.
Funding: This research was funded by the National Aeronautics and Space Administration (NASA) Cooperative Agreement NNX15AL51H.