InP Low‐Dimensional Nanomaterials for Electronic and Optoelectronic Device Applications: A Review

The suitable direct bandgap, high carrier mobility, biologically nontoxicity, and size‐dependent physical properties of low‐dimensional indium phosphide InP (0D, 1D, and 2D) have attracted great interest from scientists. The appealing optical and electronic properties make them promising for the fabrication of state‐of‐the‐art nanoscale electronic and optoelectronic devices including photodiode, photodetector, and solar cells. This Review focuses on the recent development of low‐dimensional InP materials. The synthesis methods and growth mechanisms of high quality low‐dimensional InP are comprehensively recapped. The multifunctional applications in electronic and optoelectronic devices of low‐dimensional InP are discussed, and typical strategies to resolve the main challenges limiting the performance of low‐dimensional InP‐based application are then reviewed. Finally, a brief perspective on the challenges and opportunities of low‐dimensional InP in synthesis, electronics, and optoelectronics applications is also provided.


Introduction
As a representative member of III-V compound semiconductors, indium phosphide (InP) demonstrates a direct bandgap (≈1.34 eV), high carrier mobility (4600 cm 2 V −1 s −1 ), high breakdown field (≈5 × 10 5 V cm −1 ), and high photoelectric conversion DOI: 10.1002/adsr.202200101efficiency (22.1%). [1]Nanostructured InP are widely used in high-speed electronics, satellite communications, critical semiconductor IC applications, high speed internet access, photodetectors (PDs), photovoltaics, and real time multimedia file transfer. [2]For example, InGaAs/InP single-photon detector systems have received special attention due to the low dark count rate, high photon count rate, and very accurate time resolution. [2]ne of the most exciting research areas in semiconductors over the past few decades is the low-dimensional nanostructures and devices.The "low-dimensional materials" include atomically thin 2D materials, [3] 1D nanostructures like nanowires (NWs), [4] nanotubes [5] and nanobelts, [6] 0D quantum dots (QDs), [7] fullerenes, [8] and small organic molecules. [9]It represents a class of materials with at least one reduced dimension which is far less than the incident light wavelength.For example, 2D materials are atomically thin crystalline solids with large lateral dimensions as compared to thickness.A wealth of unusual physical phenomena that occur when charge and heat transport are confined to a plane, such as charge density waves and high-temperature superconductivity, 2D excitons, commensurate-incommensurate transition, etc. [10] 2D materials also have a strong light-material interaction, which makes it possible for them to provide the basis for high-quality optoelectronic devices. [11]ompared with 2D materials, 2D motion constraint in 1D atomic crystal highly confines the mass/heat/carrier transport along the axial direction, [12] while the size reduction of 0D materials (QDs and fullerenes) in all three dimensions.The restricted electron motion leads to discrete atom-like electronic structure and size-dependent energy levels.This enables the design of nanomaterials with widely tunable light absorption, bright emission of pure colors, control over electronic transport, and a wide tuning of chemical and physical functions because of their large surface-to-volume ratio. [13]It is expected that low-dimensional InP can inherit the above advantages, which offers great potentials for the design and fabrication of functional electronic and optoelectronic devices.
Although there are a few reviews summarizing the properties of InP quantum dots or nanowires, comprehensive discussion about the device applications of low-dimensional InP materials (0D, 1D, and 2D), especially for functional electronics and opto-  [29] Copyright 2011, Wiley.
electronics is still lacking.In this Review, we start from the synthesis method of low-dimensional InP, and discuss the respective characteristics.After that, electronic and optoelectronic applications of low-dimensional InP are introduced, and the future development is also prospected.

Synthesis of InP QDs
Many approaches have been applied for the synthesis of QDs, including hot-injection, heat-up, seed-mediated approach, etc.The uniformity of particle size and low density of surface traps are highly required, which are the basis of highly efficient quantum yield (QY).

Hot-Injection
Hot-injection is a representative approach in terms of high uniformity and general applicability. [14]The rapid injection of the precursors induces a sudden supersaturation of monomers, which is followed by a short burst of nucleation. [15]The hotinjection technique includes two schemes (Figure 1a).Scheme I: rapid injection of one [B] precursor into a thermal reaction medium containing another [A] precursor; Scheme II: sequentially or simultaneously injecting the [A] and [B] precursors into the thermal medium.
Murray et al. invented the hot-injection method and applied it for the synthesis of CdS(Se/Te) QDs, [16] Peng's and Prasad's groups were inspired and synthesized high quality InP nanocrystals for the first time in 2002. [17]They injected P(SiMe 3 ) 3 precursors into octadecene (ODE) solution of indium acetate with fatty acids as the capping ligands (300 °C, Scheme I).The process was relatively green and cost-effective since the organophosphine/organophosphine oxide solvent was replaced by ODE.Researchers reported that InP QDs were synthesized in a solvent containing weak coordinating fatty-acid esters or fatty amines, which enables low-temperature reaction (<190 °C, Scheme I) and avoids a tedious degassing process. [18]The InP/ZnS core/shell nanocrystals can emit light ranging from 450 to 750 nm, with a high QY over 40%.After surface passivation with GaP shell and removing of traps, a considerably improved QY of 85% was achieved. [19]ompared with Scheme I, the progress of Scheme II is relatively slow.Using trioctylphosphine-cluster as precursors, Xu et al. synthesized size-tunable and nearly monodisperse InP QDs with absorption peak covering most part of the visible window (between ≈480 and 660 nm). [20]Such hot-injection approach with clusters precursors can also be applied for synthesizing CdSe and other II-VI QDs. [21]ue to the fast process, less toxicity, high reaction rate, and relatively high-quality of products, the hot-injection technique is the most frequently adopted and well-developed method for the preparation of colloidal nanocrystals.By further optimizing the reaction parameters, it is expected to guarantee the controllable and repeatable synthesis of InP QDs.

Heat-Up Method
Heat-up method needs to mix all precursors together in advance and then raise the temperature to reaction conditions, which requires the precise control of precursors, ligands, and temperature (low panel of Figure 1a). [22]Mićić et al. first reported the well-crystallized InP QDs in 1994 using the heat-up process. [23]nP QDs were obtained with a mean diameter of 26.1 ± 7.5 Å by heating chloroindium oxalate and P(SiMe 3 ) 3 in solution at 270 °C for several days.However, the uncontrolled decomposition kinetics make it difficult to prepare monodisperse nanocrystals with narrow size distribution. [24]In addition, the QY was as low as 1% because of their significant sensitivity to surface trap states, which are caused by phosphorous surface vacancies and dangling Reproduced with permission. [30]Copyright 2016, Wiley.b) Schematic of a cation exchange synthesis of nanocrystals.Reproduced with permission. [34]Copyright 2020, Wiley.
bonds. [15]With further efforts to eliminate the surface trap states by photochemical etching with HF, the QY can be considerably improved up to 20-40%. [25]Through the reaction of InCl 3 and P(Si(CH 3 ) 3 ) 3 in trioctylphosphine oxide at elevated temperatures, the size distribution of InP nanocrystals was narrowed down to 20 to 50 Å in diameter. [26]ore-shell InP/ZnS nanocrystals were synthesized using a convenient one-pot method, and demonstrated size-dependent emission in the range of 480-590 nm with QY of ≈68%. [27]The addition of zinc would deteriorate the absorption features of core InP QDs, resulting in broad emission line widths. [28]By using one-step method, the emission full width of Zn carboxylate covered In(Zn)P QDs at half maximum (FWHM) can be dramatically decreased to as low as 36 nm with QY up to 67% for the green emitting QDs.
The essential difference between hot-injection and heat-up method is the nanocrystal formation stage (Figure 1b).In the hotinjection method, the injection process brings about substantial nucleation.By contrast, in heat-up approach, nanocrystals do not form until the high energy barrier for homogeneous nucleation is supersaturated for burst nucleation. [29]However, the precursors in a heat-up synthesis typically experience a substantial variation of temperature, resulting in complicated reaction processes.Therefore, the experimental variables such as compositions of precursor and ligand, reaction temperature, and heating rate in a heat-up synthesis, need to be carefully optimized to minimize the size distribution, which is caused by the temporal overlap of nucleation and growth.

Seed-Mediated Approach
As for seed-mediated method, spontaneous nucleation is suppressed due to the reduction of energy barrier in heterogeneous nucleation.The growth process is based on external crystal nuclei, thereby achieving controllable and predictable growth of the nanostructures. [15,30]As shown in Figure 2a, the energy barrier for the heterogeneous nucleation is one half of that for the homogeneous nucleation.Through a seed-assisted hot-injection method, Jang and co-workers synthesized InP QDs with narrow size distribution by using polydisperse InP seeds with a mean diameter of 2.7 nm.The InP QDs with averaged diameter of 3.3 nm obtained a near-unity QY(100%), narrow FWHM (35 nm), and highly symmetric spectral emission profile. [31]common issue in preparing large spherical particles is the generation of faceted or rod-like nanostructures.It can be well explained by the varying of growth rates at different facets or the competition between particle growth and the capping action of stabilizers.Lee and co-workers overcame this problem by using oleic acid capping agent, and successfully prepared 1.9 nm InP QDs using the presynthesized oleic acid capped InP QD "seeds." [28,32]The prominent benefit of seeded growth method is the exceptional controllability of size, morphology, and uniformity of InP QDs.By simultaneously changing nanocrystal composition during the synthesis, core-shell structured InP QDs can be readily prepared.

Cation Exchange
Cation exchange is one of the methods of ion exchange.Ion exchange is a well-explored strategy for modifying the compositions and properties of crystalline materials. [33]The advantages of ion exchange including morphology retention, rapid reaction rates, and tunable thermodynamics, which make ion exchange method a particularly attractive tool to synthesize InP QDs and related heterostructures (Figure 2b). [34]Compared with hot-injection and heat-up methods, which are highly dependent on the morphology of precursors, ligation, and solvent, the cation exchange is a postsynthetic process controlled by thermodynamic and kinetic factors. [35]he ion exchange reaction in InP nanocrystals is driven by the difference in lattice energies, thus thermal condition is required to enhance ions diffusion. [33]Beberwyck and Alivisatos synthesized InP QDs by incubating preformed Cd 2 P 3 QDs and InCl 3 :tri-n-octylphosphine complex at 270 °C, and other III-V QDs such as GaAs, InAs, and GaP can also be obtained using the same strategy. [36]u 3−x P QDs is a representative template to prepare InP QDs.Manna and co-workers demonstrated that the exchange of Cu + and In 3+ ions starts from the peripheral corners Cu 3−x P nanocrystals and toward the center. [37]Nanoscale Kirkendall effect takes place in most cation-exchange processes when the solvation of host outweighs the guest desolvation, which leads to the accumulation of vacancies in the nanocrystals and resulting cracked morphology. [38]In addition, defect formation is considered detrimental for optoelectronic applications, leading to collapse of the anion framework. [35,39]Researchers found that high-temperature annealing can eliminate defects, which also implies that cation exchange at low temperatures is insufficient for crystallization effectively decoupled ionic interdiffusion and ripening processes. [38]The cation exchange method provides a facile reaction route for nanocrystals with well-controlled size and morphology, offering accessibility to hetero-nanostructures with nonequilibrium composition and crystal phase.However, for the synthesis of III-V QDs by cation exchange, mass defects still cannot be avoided, especially when low-valent cations were employed.

Other Methods
Other methods, for example, aqueous-phase synthesis and microwave-assisted synthesis are also proposed for the synthesis of InP QDs.Aqueous-phase synthesis becomes cornerstone of InP biological applications. [40]The common way of obtaining water-soluble nanocrystal is to exchange the surfactant on the surfaces for a hydrophilic. [41]However, this phase transfer generally causes a significant decrease in photoluminescence (PL) efficiency, which requires further processing to maintain the PL efficiency. [41]Microwave-assisted synthesis features rapid heating technique, enhanced material quality and size distributions.It does not need the changing of precursors, reactants or other postreaction during processing, but it is uneasy to realize coreshell nanostructures from a single step. [42]Microfluidic synthesis consists of microreactors with integrated heaters and fluidic chips, which provides nanoliter and sub-nanoliter volume scale chemical reactions.With the advantages of precise control, millisecond-level response time and suitability for a variety of materials, and it has attracted the attention of more and more researchers. [43]

Synthesis of InP Nanowires
The synthesis of InP NWs mainly follows two mechanisms: vapor-liquid-solid (VLS) [44] and solution-liquid-solid (SLS). [45]ince the dopant engineering, axial or radial heterojunction fabrication, growth process manipulation are difficult to control in SLS method. [46]Therefore, we will mainly discuss the examples of InP NWs prepared on basis of VLS mechanism.
VLS mechanism was first proposed by Wagner et al. to explain the growth of Si NWs by reducing silicon tetrachloride (SiCl 4 ) in the presence of metal nanoparticles. [44]The catalytic liquid alloy phase adsorbs the vapor-phase precursor to the supersaturation level, and subsequent crystal growth occur from the seeds at the liquid-solid interface [47,48] (Figure 3a).Typically, the metal seed is pushed up by the growth of the NWs, thus leading to the formation of nanowire with a metal ball at tip.Bakkers and co-workers synthesized InP NWs using metalorganic chemical vapor deposition (MOCVD) with trimethylindium and phosphine as molecular precursors at 420 °C. [49]By controlling the concentration of impurity Zn dopants, InP NWs with twinning superlattices in long-range order were obtained (Figure 3b).
For the laser-assisted catalytic growth (LCG), a pulsed laser is utilized to ablate targets for the generation of gas source.Morales and Lieber pioneered the utilization of an LCG to prepare InP NWs. [46]Using controllable laser power and tunable gold particle size, precise and controllable growth of InP NWs was achieved (Figure 3c,d). [50]In addition, the diversity of solid targets enables the synthesis of binary group III-V material NWs. [51]LCG is a simple, clean, and fast synthesis method of NWs due to the reduction of by-products, simple precursors, and absence of catalysts.However, the main disadvantage of LCG is the low productivity due to high input energy and the small laser-irradiating area for evaporating the target materials.
Both nanoimprint lithography (NIL) and selective area epitaxy (SAE) are available approaches for fabrication of nanowire arrays.NIL is an unconventional lithography technique for highthroughput patterning with high accuracy and low cost. [52]Wallentin et al. demonstrated the utilization of NIL for the synthesis of vertical InP NWs arrays (Figure 3e). [53]Growth of InP NWs driven by VLS mechanism was realized with gold particles as catalytic seeds, which are initially arranged in the array with the assistant of NIL.By carefully controlling the size and spacing of gold seeds, samples with different NWs diameters (130 to 190 nm) and array pitches (470 or 500 nm) can be obtained.SAE enables the fabrication of nanostructures with lithographically defined positioning, catalyst-free, high-uniformity, and silicon compatibility. [54]Using SAE method, Gao et al. reported the growth of stacking-fault-free and taper-free wurtzite (WZ) InP NWs array (Figure 3f).The diameter of NWs ranges from 80 to 600 nm, and the nanowires reveal a high quantum efficiency of ≈50%. [55]SAE grown InP NWs are usually tapering-free and have low impurity concentrations due to high growth temperature. [56]he morphology of NWs is highly related to the phase structure of InP crystal. [55,57,58]Fukui and co-workers revealed that high proportion of zinc blende (ZB) segments will result into the tapered morphology of the NWs. [55,57]As shown in Figure 3g, in the wurtzite (WZ) segment, { " 2110} facets are vertical to the (111)A substrate, while in the ZB segment, { " 111} facets are inclined 19.5°from ⟨111⟩ vertical direction to the InP (111)A substrate.Therefore, a higher proportion of ZB segments will result in a more tapered morphology of the NWs due to the preferred sidewall facets and radial growth preference.Motohisa and Lieber discussed the phase structure of InP crystals grown at different temperatures. [59]High-resolution TEM image and selectedarea electron diffraction patterns indicate that NWs obtained at 660 °C were mainly WZ phase.Statistic analysis shows that 92% of the atomic-layer stacking order follows WZ configuration.By contrast, NWs with ZB phase were obtained as the growth temperature decreased to 600 °C, and high density of stacking faults and twinning dislocations appear (Figure 3h).SAE mainly occurs with a slow growth rate, which badly blocks the formation of highaspect-ratio NWs. [56]

Synthesis of 2D InP
Since the isolation of graphene in 2004, studies on ultrathin 2D materials have grown rapidly in the fields of condensed matter physics, material science, chemistry, and nanotechnology. [60]ue to high specific surface area and confinement of electrons, InP crystals in 2D configuration are expected to exhibit intriguing physical properties.However, the strong covalent bonding in  [47] Copyright 2018, Wiley.b) Transmission electron microscopy (TEM) image of a wire containing segments of intrinsic InP and Zn doped segments (top) and higher magnification TEM image (bottom).Reproduced with permission. [49]Copyright 2008, Springer Nature.c,d) TEM and HRTEM images of InP nanowire grown by laser ablation method.Scale bars are 50 nm and 5 nm, respectively.Reproduced with permission. [50]Copyright 2001, American Chemical Society.e) Scanning electron microscope (SEM) images of NIL-defined InP nanowire arrays.Reproduced with permission. [53]Copyright 2013, AAAS.f) Selective-area epitaxy of InP NWs.Reproduced with permission. [55]Copyright 2014, American Chemical Society.g) HRTEM image of InP nanowires containing WZ and ZB segments (lift) and corresponding structural image (right).Reproduced with permission. [57]Copyright 2011, American Chemical Society.h) InP nanowires growth temperature 660 °C (lift) InP nanowires growth temperature 600 °C (right).Scale bar: 10 nm.Reproduced with permission. [59]Copyright 2007, American Chemical Society.
nonlayered InP crystals hinders the layer-by-layer exfoliation and impedes the anisotropic growth.
Javey and co-workers first demonstrated the templated liquidphase (TLP) growth of high-quality polycrystalline InP thin-films with large grain size up to 100 μm, [61] and they further developed a technique that enables direct "writing" of single-crystalline III-V semiconductors on amorphous substrates. [62]As shown in Fig- ure 4a, indium crystals are first transformed into the liquid phase during growth process, but remains mechanically confined by the SiO x template.Phosphorus passes through the SiO x cap, combines with liquid In, and finally precipitates in the form of InP crystals.To prevent further nucleation, a phosphorous depletion zone is highly required, which also contributes to the gradual nucleation of InP crystals. [61]The growth of InP in predefined 2D geometries can be finally achieved, and each individual pattern is a single crystal with a preferential texturing along the (1 0 n) direction with n ranging between 1 and 2 (Figure 4b,c).
Fu and co-workers designed AuIn 2 alloyed substrate to improve the interfacial interaction, and obtained ultrathin InP atomic crystal with a thickness of only 6.3 nm (Figure 4d-f). [63]he simulation results show that the adhesion energy of InP layer on alloy substrate is much greater than that on the InP or Indium substrate.Therefore, the formation of InP on the alloy substrate tends to proceed in the 2D layer-by-layer growth mode.
A remote epitaxy technique was recently proposed for the preparation of 2D III-V semiconductors, including InP, GaAs, and GaP.For example, the homoepitaxial growth of InP films was conducted on InP substrate by MOCVD at 450 °C with itrimethylindium and indium as precursors.Monolayer graphene is introduced as intermediate layer, which allows the rapid release and transfer of InP films due to the weak vdW force at heterinterface. [64]DFT calculations demonstrates that the epitaxy is possible within a 9 Å gap between substrate and epilayer because the charge density still exists.The epitaxial InP film can (Scale bar: 10 mm).Reproduced under terms of the CC-BY license. [62]Copyright 2016, Springer Nature.d) Schematic illustration of the growth process of III-V crystals.e) The AFM image of an ultrathin InP single crystal grown on AuIn 2 .f) Bright-field scanning TEM (BF-STEM) image of the InP crystal along the [111] zone axis.Reproduced under terms of the CC-BY license. [63]Copyright 2020, The authors, published by Springer Nature.g) Schematic illustration of exfoliation process.h) Photographs of single-crystalline InP (001) films exfoliated.i) High-resolution X-ray diffraction patterns of InP (001) films exfoliated.Reproduced with permission. [64]Copyright 2017, Springer Nature.
be easily exfoliated by a Ti/Ni stressor and successfully transferred onto arbitrary substrates with good crystal quality (Figure 4g-i).Although many efforts have been devoted to the preparation of 2D InP, methods to obtain high-quality atomically thin layers of InP are still urgently needed.The lacking of high-quality 2D InP also makes it difficult to probe the optical and optoelectronic properties.

Emerging Applications of Low-Dimensional InP
Versatility of applications for low-dimensional InP are explored due to the unique structure and physical properties.In this section, the applications of low-dimensional InP and heterostructure including electronic device, optoelectronic device, biosensing, and imaging are reviewed.

Electronic Devices
The electrical properties and transport characteristics of InP NWs have been extensively studied.The 1D geometry of NWs can offer great electrostatic control due to the reduced dimensions, and devices in gate-all-around configurations enables the lowering drive voltages and decreasing of power consumption. [1,65]nP NWs are usually polytypic, and the crystal structure plays a key role on the conductivity.For polymorphic InP NWs, the zincblende segments can act as traps for carriers, hindering the carrier transport and affects the electrical performance. [66]For example, field effect transistor (FET) made of polytypic InP NWs shows very low mobilities of only 100 cm 2 V −1 s −1 . [67]Many efforts have been devoted to controlling the NW crystal form, orientation, diameter, etc., with an aim to minimize the adverse effects defects on the electronic transport properties of NWs. [68]Ho et al. changed Au catalyst into Pd catalyst and successfully prepared nonpolar-oriented defect-free WZ InP NWs. [69]Benefiting from the substantial reduction in crystal defects, the FET-based on InP NWs reveals an excellent I ON /I OFF ratio of ≈2.4 × 10 5 at room temperature together with a record-high electron mobility 2, 000 cm 2 V −1 s −1 (Figure 5a,b).
A small amount of dopant atoms can change the conductivity of semiconductors, which is crucial for the electrical properties of InP NWs.Researchers study the electrical properties of different doping InP NWs, which are controlled by selective doping. [2,62]or example, Lieber et al. demonstrated NWs with different doping types, and studied the electrical properties of different p-p, n-n, p-n cross-junctions (Figure 5c,d). [2]The n-and p-type NW building blocks are also available for highly integrated device arrays and functional networks. [70]ompared with bare InP crystals, InP-based heterostructures can be well engineered with specific electronic band structures, which are expected to exhibit excellent photodetection capability.Thelander and co-workers reported an InP-GaAs axial NWs het-  [69] Copyright 2018, American Chemical Society.c) SEM image of crossed nanowire device.d) I-V curves of p-n junctions.Reproduced with permission. [2]Copyright 2001, Springer Nature.e) SEM image and f) corresponding transfer curves of InP/GaAs NWs.Reproduced with permission. [71]Copyright 2012, American Chemical Society.erostructure with an n-i-p doping profile, and the FET demonstrates a small subthreshold slope down to 50 mV dec −1 and high I ON /I OFF ratio of 10 7 (Figure 5e,f). [71]Low-temperature electrical measurements suggests that the trap-assisted tunneling effect associated with the narrow bandgap segment of InGaAsP dominated the electrical transport behaviors.In addition, heterojunctions related to InP NWs also reported for constructing electronic devices, such as InP/ZnS, [72] InAs/InP, [73] InP/InGaAs, [74] InGaAs-InP, [75] etc.

Optoelectronic Devices
Low-dimensional InP and hybrid structure with high carrier mobility, high absorption coefficient, and tunable bandgap property make them ideal materials for optoelectronics device.The optoelectronic applications of low-dimensional InP are reviewed, focusing on the photodetectors, light-emitting devices, and photovoltaic devices.

Photodetectors
Low-dimensional InP photodetector features high responsivity, fast response speed, and large polarization sensitivity from the visible to the long-wave infrared region. [2,76]These properties endow low-dimensional InP with great potential for application as photodetectors.
Electron-hole pairs are excited in semiconductors when photons with energies above the bandgap are absorbed.Electrical characteristics of photoactive material vary with the change of carrier density induced by external incident light and thus achieve the detection.InP NWs photodetectors can be fabricated in the form of single NWs (horizontally between the two electrodes, FET-like structure) or ensemble of NWs (regularly arranged array).Since the first demonstration of an InP NWs photodetector by Lieber and co-workers in 2001, [2] plenty of optical sensitive devices have been reported based on low-dimensional InP crystals.
In general, bare InP crystal without heterojunction can realize high photosensitivity, however, the defects will reduce the electron mobility and affect the recombination process of photoexcited electron-hole pairs, thus leading to the deterioration of optoelectronic performance. [77]77a] A lot of efforts have been mbade to reduce the impact of defects and orientations.For example, Ho et al. demonstrated pure WZ InP NWs with nonpolar growth orientations of ⟨2̅ 110⟩ and ⟨1̅ 100⟩ via Pd-catalyzed growth, which exhibit a high photoresponsivity of 10 4 A W −1 and fast rise and decay times of 0.89 and 0.82 s, respectively [69] (Figure 6a,b).Bond rotation is strictly prohibited when the In and P atoms configure in the nonpolar WZ ⟨2̅ 110⟩ and ⟨1̅ 010⟩ planes, resulting into a minimized concentration of stacking faults and other defects.
The optoelectronic properties of low-dimensional InP crystals can be effectively modulated by integrating with foreign materi-  [69] Reproduced with permission. [69]Copyright 2018, American Chemical Society.c) Schematic of InP/InAsP NW array devices.d) Photocurrent at 300 K for different applied biases.Reproduced with permission. [79]Copyright 2018, American Chemical Society.e) Schematic of the InP QD/BP hybrid photodetector.f) Responsivity and detectivity of the InP QD/BP photodetector at different optical power densities.Reproduced with permission. [80]Copyright 2019, American Chemical Society.
als.The heterostructure demonstrates specific electronic states, precise control of charge transport that are absent in singlecomponent system, thus contributing to the enhanced photosensitivity, fast response, high gain, wide detection range even selfpowered capability. [78]For example, InP/InAsP NWs formed by self-assembly achieved a gain of 12, where InP is used for photon absorption. [78]In addition, the infrared response from 3 to 20 μm can be achieved by axially growing InAsP low-bandgap quantum disks (QDisc) in InP NWs by MOCVD (Figure 6c,d). [79] hybrid photodetector was fabricated by spin-coating InP QD light-absorbing layer onto 2D black phosphorus (B-P) channel.The device demonstrates a high responsivity of 1 × 10 9 A W −1 and a detectivity of 4.5 × 10 16 Jones under 405 nm light illumination (Figure 6e,f). [80]he hybrid structure based on low-dimensional InP crystals can also be applied for constructing avalanche photodetectors. [81]or example, by integrating a single InAsP QDs into the avalanche multiplication region of an InP NWs photodiode, singleshot electrical readout can be implemented. [82]The hybrid device also reveals a high conversion rate of 95%, high gain of 2.3 × 10 4 , low dark current of ≈0.2 pA, and linear polarization selectivity (Figure 7a,b).This work represents a significant step towards achieving single-shot electrical readout and offers a new functionality for on-chip quantum information circuits.
The utilization of single InP NWs for terahertz (THz) detection in a nascent field has great potential for the realization of highly integrated THz systems.In 2016, Jagadish and co-workers demonstrated a single InP nanowire-based THz photoconductive detector, which exhibits excellent sensitivity, broadband perfor-mance (0.1 to 2.0 THz), high carrier mobility high rate (≈1260 cm 2 V −1 s −1 ), and low dark current (≈10 pA) (Figure 7c,d). [83]hrough the doping engineering of InP NWs, Jagadish and coworkers demonstrate high-quality terahertz detection with a 2.5fold improvement in signal-to-noise ratio compared to undoped InP NWs PD, which is comparable to conventional bulk terahertz PD. [85] Photogating effect is effective to change the electrical conductance of semiconductors, thus enabling the modulation of photoelectric gain and bandwidth of photodetector. [86]Zheng et al. reported a sidegated single InP NWs photodetector with ferroelectric polymer as photogating layer. [84]The device reveals a high photoconductive gain of 4.2 × 10 5 , responsivity of 2.8 × 10 5 A W −1 , and high detectivity of 9.1 × 10 15 Jones (Figure 7e,f).Nearsurface electrons will be repelled under the strong ferroelectric field applied on the NWs, thus causing a sharp upward bending of the energy level.The photoexcited electrons and holes are spatially isolated with long lifetimes for recombination, thus leading to the high gain and the picoampere-level currents. [86]

Light-Emitting Devices
The recombination of electrons with holes leads to the energy releasing in the formation of photons, where the color of the light is determined by the bandgap of the semiconductor.Although most of state-of-the-art quantum dots light-emitting devices (QLED) are based on Cd-based semiconductor, the toxicity element Cd brings hidden trouble for their consumer applications. [87]Cd- Absorption of one photon per pulse yields a signal (red) and the dark current (black).Reproduced with permission. [82]Copyright 2012, Springer Nature.c) InP NW detector with strip-line geometry.d) Decay of conductivity lifetime (top) and THz induced current (bottom).Reproduced under terms of the CC-BY license. [83]Copyright 2016, The authors, published by American Chemical Society.e) 3D schematic of ferroelectric side-gated InP NWs photodetector.f) Plots of photoresponsivity and detectivity.Reproduced with permission. [84]Copyright 2016, American Chemical Society.
free InP-based QLED is burgeoning as an ideal substitution in recent years. [88]The first InP@ZnSeS QLED was reported in 2011, [89] which delivers poor external quantum efficiency (EQE) of ≈0.008%.The quality of InP QDs plays important role for the performance of QLED.The delocalized electrons are easily trapped by surface ions or vacancies upon excitation, and the resulting defect states would cause the dissipation of excitation energy. [90]Jang and co-workers reported a high performance QLEDs based on InP/ZnSe/ZnS QDs by surface passivation.The device reveals a high EQE of 21.4%, together with a maximum brightness of 100 000 cd m −2 and an extremely long lifetime of one million hours (Figure 8a,b). [31]The InP-based QDs are expected to be optimized from the following four aspects: 1) controlling of the precursor injection rate for the enhancement of the uniformity of InP core, 2) In situ etching of oxide during shell passivation process to improve the optical properties of InP QDs, 3) thickness increasing of ZnSe interlayer to suppress the nonradiative Auger recombination and Förster resonance energy transfer, and 4) utilization of hexanoic acid to reduce the alkyl length for better charge injection.Besides, the inorganic salt ZnF 2 mildly reacts with carboxylic acid at a high temperature and in situ generates HF, avoiding the danger of adding HF directly.The continuous and mild production of HF makes the InP/ZnSe/ZnS shell more uniform than adding HF directly.Meanwhile, InP/ZnSe/ZnS -based QLED achieves the highest peak external quantum efficiency of 22.2%. [91]part from the quality of InP QDs, structure optimization of device such as designing inverted structure, the introduction of electron transport layers (ETLs) and hole injection layers are also applied to improve the performance of QLED devices. [88,92]For example, using ZnMgO nanocrystals as electron transport layers, threefold enhancement of emitting efficiency can be demonstrated in InP QLEDs. [93]Peng and co-workers also reported an InP QLEDs by introducing Zn 0.9 Mg 0.1 O nanocrystals as ETL. [94]he device delivers 12.2% peak EQE and maximum brightness >10 000 cd m −2 (Figure 8c,d) due to improvement of the charge balance and the mitigation of exciton quenching.To improve the charge balance in the InP-QLED, there is a need of new high mobility hole transport material with deep HOMO level.Dibenzothiophene and tertiary amine units (DBTA) have been employed as new HTL with high hole mobility and a deep HOMO level. [95]The fabricated red InP-QLED with DBTA HTL and optimized ZnMgO ETL showed an extremely high EQE of 21.8% (Figure 8e).

Photovoltaic Device
Photovoltaic (PV) device can convert the energy of light directly into electricity by the photovoltaic effect.InP has exceptional optoelectronic characteristics and ideal direct bandgap, thus allowing the maximization power conversion efficiency (PCE) under AM 1.5G spectrum. [96]In comparison to Si and GaAs, InP is considered as an ideal materials candidate for commercial PV device applications because of the stronger radiation resistance and longer minority carrier lifetime. [97].Reproduced with permission. [31]Copyright 2019, Springer Nature.c) Scheme of QLEDs using ZnMyO ETL.d) EQE and current efficiency versus luminance of the optimal device.Inset: corresponding CIE coordinates.Reproduced with permission. [94]Copyright 2019, American Chemical Society.e) InP-based QLED device with DBTA HTL.Reproduced with permission. [95]Copyright 2018, The authors, published by American Chemical Society.c) SEM image of InP NWs array solar of exceeding the ray optics limit.d) 1-sun J-V curve.Reproduced with permission. [53]Copyright 2013, AAAS.e) SEM image of a cleaved nanowire solar cell device with ITO hemispheres.f) J-V curve of top enhanced absorption.Reproduced with permission. [101]Copyright 2016, American Chemical Society.
Understanding intrinsic limiting factors of single NW devices can provide valuable information for the development of largearea devices.Through carrier doping, surface passivation, introduction of dielectric shells, and p-n junction, high PCE can be achieved in InP PV devices. [98]Compared with the device in horizontal structure, high absorption efficiency (>1%) can be realized in NW device with vertical configuration due to large absorption cross-sections. [99]With the assistance of a nanoprobe installed within an SEM, researcher can correlate the parameter of individ-ual NWs with arrayed solar cells.Through feedback from single nanowire measurements, Borgström and co-workers identified the relationship between performance-limiting parameters and growth parameters, which contributes to a remarkable enhancement of PCE for InP NWs array solar cells from less than 2% to 15.0% (Figure 9a,b). [100]ther optimization approaches, including resonant light trapping, surface cleaning, and top design are also proposed to improve the performance of InP array solar cells. [53,101]Benefiting  9c,d). [53]The solar cell based on InP NWs array with surface cleaning and top enhance absorption reveals an extremely high PCE of 17.8% (Figure 9e,f). [101]

Bioimaging and Biosensing
Hybrid materials based on InP QDs are appropriate for the development of bioimaging and biosensing. [102]Although Cd-based QDs are currently the most widely studied, researchers prefer to turn to InP QDs for biological applications due to minimal intrinsic toxicity. [103]The incorporation of biomolecules into InP QDs enables the construction of hybrid nanostructures, which can be applied for optical bioimaging and biosensing.Several attempts have been reported to achieve optical tissue window imaging, tumor detection, tumor therapy, and complete metabolism. [104]or example, mercaptoacetic acid-coated InP QD conjugated with folic acid for overexpressed tumor cells such as human oral epidermoid carcinoma cells (KB) (Figure 10a). [104]NIR region (≈750-900 nm) lies in the spectral region where blood and tissue absorb minimally but are still detectable by the instrument, making them ideal for in vivo imaging applications of QDs.Gao et al. demonstrted mercaptopropionic acid (MPA)-coated InAs/InP/ZnSe QDs that exhibit 800 nm emission wavelength (QD800-MPA) and diameter smaller than 10 nm (Figure 10b). [105]or intra vivo fluorescence imaging, high tumor uptake of with excellent contrast to the surrounding tissues, which is promis-ing for tumor detection in living micethe InP QDs (Figure 10c).Meanwhile, 800 nm emission wavelength and diameter of 10 nm ensures low tissue background and high tissue penetration and allows renal clearance, respectively.
Rapid diagnosis and targeted drug treatment have also been desired in recent years.For example, functionalized InP nanocomposites (IMAN) with 640 nm PL peak was synthesized by modifying InP QDs with a vascular endothelial growth factor receptor 2 monoclonal antibody and miR-92a inhibitor (Figure 10d). [104]Functions including targeted drug delivery, infrared imaging in vivo and induction apoptosis of human myelogenous leukemia cells were successfully implemented (Figure 10e).IMAN provides chemotherapy strategy that against cancer cells by its targeting function and utility in 3D tumor imaging.

Conclusions and Perspectives
This review provides a platform of short yet recapitulative coverage that focus on the synthesis methods of low-dimensional InP and its optoelectronic/optical applications, such as photodetectors, biological applications, solar cells, etc. Besides, the lowdimensional InP also provides great opportunities for fundamental physics/chemistry research and future applications.
Due to the 3D confinement, the quantum confinement effect of 0D InP nanocrystals is notable.For the InP QDs with diameter of 1 nm, a remarkably increased energy bandgap can be expected, as high as 4.6 eV. [106]InP QDs feature excellent sizedependent bandgap, large absorption coefficient, broad color tunability, low toxicity, and low-cost solution process. [107]Compared with 1D nanowires and 2D nanosheets, InP QDs feature with mass manufacturing capability.These properties render InP QDs promising candidates for a wide diversity of applications, especially in light emitting [108] and biomedical imaging. [109]In addition, InP QDs can also be used as the light-absorbing layer of photodetectors. [80]It should be noted that one persisting drawback for the synthesis of InP QDs is the low availability of suitable phosphorus precursors.In view of the large-scale production of InP QDs, development of stabilizers and solvents, which contributes to the sustainability of the syntheses, also needs to be considered.The preparation of InP QDs still needs rigorous experimental conditions due to the high reactivity of [P] toward [In] as well as impurities such as water and oxygen. [110]In view of applications, long-term stability under harsh conditions should be considered, which is highly necessary for developing consumer electronics such as solar cells and multicolor-LED-based displays.
InP nanowire solar cells demonstrate high absorption efficiency due to increased light absorption related to the unique 1D structure, allowing the substantial cost reductions. [53]The effective strain relaxation provided by the open side surfaces of NWs enables the combination of highly mismatched heterogeneous materials. [111]In addition, electrons and holes in 1D channel undergo impact ionization with high probability, thereby achieving large multiplication factors, which facilitates the construction of InP avalanche photodiodes. [82]Noted that the surface roughness and interface traps would degrade the subthreshold performance and result into the poor PL efficiency, [112] therefore NWs need to be passivated by a shell material with large bandgap. [4]Considering the high toxicity and environmental friendliness, heavy metals (Cd or Pb) utilized as seeds for NWs growth needs to be substituted by Au, and it poses great challenge to synthesis largescale NWs for consumer electronics such as solar cells and LEDbased displays. [13]Development of strategies for NWs assembly would help to construct more complex electronic or photonic systems that involve InP materials with multiple dimensionalities.However, achieving of high-quality interfacial contact faces great challenges due to cylindrical structure of 1D atomic crystals.
Although the synthesis of 0D and 1D InP crystals has been extensively studied, there is no mature and effective approach for the synthesis of 2D InP due to the strong covalent bonding between InP atoms.Therefore, development of synthesis methods for 2D InP is of high urgency.
The applications of low-dimensional InP materials for photonics and optoelectronics are exciting.For example, InP QDs with high emission efficiency can be applied to nonclassical quantum light sources such as single-photon and entangled-photon sources.More importantly, InP nanostructures with low dimensionalities can also be merged into complementary metal-oxidesemiconductor (CMOS) platform.There are two approaches to integrate InP into CMOS platform, either by epitaxially growing InP materials on silicon wafers to form monoliths, or by directly transferring InP nanostructures and devices for hybrid integration.The electronic-photonic codesign based on InP-Si hybrid platform will change the way of chip-to-chip and chip-to-system communications in the future.

Figure 1 .
Figure 1.Schematic of hot-injection and heat-up approach.a) Synthesis InP QDs by hot-injection (top) and heat-up (bottom) techniques.b) Calculation of hot-injection (top) and heat-up (bottom) nanocrystal formation process.Reproduced with permission.[29]Copyright 2011, Wiley.

Figure 2 .
Figure 2. Schematic illustration of seed-mediated approach and cation exchange.a) Two nucleation modes: homogeneous nucleation (top) in solution and heterogeneous nucleation (bottom) on the seed's surface (left) and plot of Gibbs free energy change for homogeneous (orange line) and heterogeneous (blue line) nucleation (right).Reproduced with permission.[30]Copyright 2016, Wiley.b) Schematic of a cation exchange synthesis of nanocrystals.Reproduced with permission.[34]Copyright 2020, Wiley.

Figure 3 .
Figure 3. Schematic diagram of VLS growth and characterization of InP NWs.a) Schematics for VLS growth mechanisms.Reproduced with permission.[47]Copyright 2018, Wiley.b) Transmission electron microscopy (TEM) image of a wire containing segments of intrinsic InP and Zn doped segments (top) and higher magnification TEM image (bottom).Reproduced with permission.[49]Copyright 2008, Springer Nature.c,d) TEM and HRTEM images of InP nanowire grown by laser ablation method.Scale bars are 50 nm and 5 nm, respectively.Reproduced with permission.[50]Copyright 2001, American Chemical Society.e) Scanning electron microscope (SEM) images of NIL-defined InP nanowire arrays.Reproduced with permission.[53]Copyright 2013, AAAS.f) Selective-area epitaxy of InP NWs.Reproduced with permission.[55]Copyright 2014, American Chemical Society.g) HRTEM image of InP nanowires containing WZ and ZB segments (lift) and corresponding structural image (right).Reproduced with permission.[57]Copyright 2011, American Chemical Society.h) InP nanowires growth temperature 660 °C (lift) InP nanowires growth temperature 600 °C (right).Scale bar: 10 nm.Reproduced with permission.[59]Copyright 2007, American Chemical Society.

Figure 4 .
Figure 4. Growth of 2D InP.a) Schematic for TLP crystal growth.b) SEM images of an array of InP circles and c) corresponding inverse pole figure maps.(Scalebar: 10 mm).Reproduced under terms of the CC-BY license.[62]Copyright 2016, Springer Nature.d) Schematic illustration of the growth process of III-V crystals.e) The AFM image of an ultrathin InP single crystal grown on AuIn 2 .f) Bright-field scanning TEM (BF-STEM) image of the InP crystal along the[111] zone axis.Reproduced under terms of the CC-BY license.[63]Copyright 2020, The authors, published by Springer Nature.g) Schematic illustration of exfoliation process.h) Photographs of single-crystalline InP (001) films exfoliated.i) High-resolution X-ray diffraction patterns of InP (001) films exfoliated.Reproduced with permission.[64]Copyright 2017, Springer Nature.

Figure 5 .
Figure 5. Electronic devices with 1D InP.a) HRTEM images of nonpolar-oriented defect-free WZ InP NW. b) FET device transfer curve based on defectfree InP.Reproduced with permission.[69]Copyright 2018, American Chemical Society.c) SEM image of crossed nanowire device.d) I-V curves of p-n junctions.Reproduced with permission.[2]  Copyright 2001, Springer Nature.e) SEM image and f) corresponding transfer curves of InP/GaAs NWs.Reproduced with permission.[71]Copyright 2012, American Chemical Society.

Figure 6 .
Figure 6.Optoelectronic devices with low-dimensional InP.a) Typical SEM image of Pd-InP NWs device (top) and corresponding device schematic (bottom).b) Transfer characteristics with/without light.[69]Reproduced with permission.[69]Copyright 2018, American Chemical Society.c) Schematic of InP/InAsP NW array devices.d) Photocurrent at 300 K for different applied biases.Reproduced with permission.[79]Copyright 2018, American Chemical Society.e) Schematic of the InP QD/BP hybrid photodetector.f) Responsivity and detectivity of the InP QD/BP photodetector at different optical power densities.Reproduced with permission.[80]Copyright 2019, American Chemical Society.

Figure 7 .
Figure 7. Three types of photodetectors based on InP NW. a) SEM image of InP nanowire p-n junctions containing InAsP QD and schematic.b)Absorption of one photon per pulse yields a signal (red) and the dark current (black).Reproduced with permission.[82]Copyright 2012, Springer Nature.c) InP NW detector with strip-line geometry.d) Decay of conductivity lifetime (top) and THz induced current (bottom).Reproduced under terms of the CC-BY license.[83]Copyright 2016, The authors, published by American Chemical Society.e) 3D schematic of ferroelectric side-gated InP NWs photodetector.f) Plots of photoresponsivity and detectivity.Reproduced with permission.[84]Copyright 2016, American Chemical Society.

Figure 9 .
Figure 9. Photovoltaic device based on InP NWs array.a) InP NWs array with the tungsten nanoprobe contacting a single nanowire.b) J-V characteristic of before (red) and after (blue) optimization.Reproduced under terms of the CC-BY license.[100]Copyright 2018, The authors, published by American Chemical Society.c) SEM image of InP NWs array solar of exceeding the ray optics limit.d) 1-sun J-V curve.Reproduced with permission.[53]Copyright 2013, AAAS.e) SEM image of a cleaved nanowire solar cell device with ITO hemispheres.f) J-V curve of top enhanced absorption.Reproduced with permission.[101]Copyright 2016, American Chemical Society.

Figure 10 .
Figure 10.InP QD targeted imaging and therapy.a) PL Images of QD-FA in KB cells.Reproduced with permission. [104d] Copyright 2005, American Chemical Society.b) UV/vis absorption and PL emission spectra of QD800-MPA.c) In vivo imaging of LS174T tumor-bearing mice after the tail-vein injection of QD800-MPA-HSA nanoparticles.Reproduced with permission.[105]Copyright 2010, Wiley.d) PL spectra of the IMAN.e) Schematic diagram of IMAN for targeted tumor imaging and treatment.Reproduced with permission.[104c]Copyright 2017, American Chemical Society.