All-Inorganic Metal Halide Perovskite Nanocrystals: Opportunities and Challenges.

The past decade has witnessed the growing interest in metal halide perovskites as driven by their promising applications in diverse fields. The low intrinsic stability of the early developed organic versions has however hampered their widespread applications. Very recently, all-inorganic perovskite nanocrystals have emerged as a new class of materials that hold great promise for the practical applications in solar cells, photodetectors, light-emitting diodes, and lasers, among others. In this Outlook, we first discuss the recent developments in the preparation, properties, and applications of all-inorganic metal halide perovskite nanocrystals, with a particular focus on CsPbX3, and then provide our view of current challenges and future directions in this emerging area. Our goal is to introduce the current status of this type of new materials to researchers from different areas and motivate them to explore all the potentials.

Introduction

Everything seems to be speeding up in the new century, especially in the field of scientific research. For example, researchers have spent more than 20 years to achieve a photoconversion efficiency (PCE) of over 20% for silicon-based (Si-based) photovoltaics.[1] In striking contrast, the PCE of solar cells based on metal halide perovskite (ABX3, where A = organic group; B = Pb, Sn, Sb, etc.; X = Cl, Br, or I) has been rapidly promoted from ∼3.8% to over 22% within just several years.[2−5] In addition to solar cells, various applications have been demonstrated based on metal halide perovskite materials, such as light-emitting diodes (LEDs),[6,7] photodetectors,[8] lasers,[9,10] flexible electronics,[11] and so on. Many scalable manufacturing protocols have also been developed in a relatively short period to meet the demand of practical applications, along with more detailed characterizations of their physiochemical properties by state-of-the-art instruments. As a result, thousands of papers on these emerging materials have been published over the past several years.

Despite the great success of organic metal halide perovskite (OHP) materials, several problems remain to be resolved, including particularly their long-term stability under ambient condition. It has been discovered that they are sensitive to many factors, such as moisture, oxygen, light, and heat.[12] Although many methods have been proposed to solve the problem, the low intrinsic stability of organic metal halide perovskite materials originated from the organic groups has not been improved significantly. It is, therefore, not a surprising idea to replace the organic group with inorganic ions to produce all-inorganic metal halide perovskite (IHP) materials.[13]

In 2015, the Kovalenko group reported the first successful preparation of CsPbX3 nanocrystals (NCs) using a hot-injection method.[14] Compared with the OHP materials, the CsPbX3 materials possess several advantages, including a higher melting point (>500 °C), higher thermal stability, and higher stability against photobleaching, which make them more suitable candidates for optoelectronic applications.[15,16] The photophysical properties of CsPbX3 NCs are also excellent: they show high photoluminescence quantum yields (PLQYs, up to 90% without any post-treatment), wide color gamut (up to 140%), and narrow photoluminescence emission line-widths that can be easily tuned over the entire visible spectral region by manipulating the halide composition. For the traditional II–VI or III–V semiconductor NCs, it has been well-established that the particle size strongly affects their emission peak, which often brings problems in the consistency of the optical property as the particle size may vary from batch to batch during practical synthesis of such NCs. In contrast, much higher consistency in the optical property can be obtained in CsPbX3 NCs because their emission peak position is mainly determined by the halide composition. With the advantages mentioned above, CsPbX3 NCs have been regarded as promising materials for next-generation optoelectronic applications, as made evident by the emergence of over a thousand relevant papers in the past three years.[17]

In this Outlook, we first briefly summarize the recent advances in the field of IHP NCs, followed by our thoughts on the future directions of this field.

Synthetic Strategies and Morphology Control

The Kovalenko group first reported the all-inorganic metal halide perovskite NCs in 2015,[14] which represents the launch of a new field in luminescent nanomaterials. As depicted in Figure a, a typical synthesis to colloidal CsPbX3 NCs involves simply the injection of Cs-oleate into a three-neck flask containing an octadecene solution of predissolved PbX2 at 140–200 °C. Several seconds later, monodisperse CsPbX3 nanocubes, as shown in the inset of Figure a, are obtained after a rapid cooling down of the reaction system using an ice bath. The CsPbX3 NCs show widely tunable emission wavelength (410–700 nm), high PLQY (up to 90%), narrow-emission line-widths (12–42 nm), and wide color gamut (up to 140% of NTSC standard). This method is similar to the synthesis of traditional II–VI and III–V semiconductor quantum dots (QDs), where high temperature and inert gas protection are usually required to improve the crystallinity of the product.

Figure 1

Schematic illustration of the methods for the synthesis of colloidal perovskite nanocrystals: (a) hot-injection method, (b) supersaturated recrystallization method, (c) ultrasonication-assisted method, (d) solvothermal method, and (e) microwave-assisted method. The images were modified with permission from refs (a) (14), (b), (18), (c) (19), (d) (21), and (e) (22).

Compared with traditional II–VI and III–V semiconductor NCs such as metal chalcogenides and pnictides, the chemical bonding of CsPbX3 NCs is more ionic. In principle, CsPbX3 NCs can be prepared in a less stringent environment. It was soon discovered by many groups that high-quality CsPbX3 NCs can be produced under moderate synthetic conditions, for example, at room temperature and ambient condition. Zeng and co-workers reported a ligand-assisted reprecipitation method, where Cs+, Pb2+, and X– ions were predissolved in a polar solvent and then injected into a nonpolar solvent, leading to rapid nucleation and growth processes (Figure b).[18] This method can produce gram-scale CsPbX3 NCs and, more importantly, avoid the need for inert gas protection and high temperature. This work was quickly followed by the development of many more facile, low-cost, and large-scale synthetic strategies to the preparation of high-quality CsPbX3 NCs. For instance, Polavarapu et al. demonstrated that highly luminescent CsPbX3 NCs could be prepared through an ultrasonication approach (Figure c).[19,20] Zhang et al. have developed solvothermal (Figure d)[21] and microwave-assisted (Figure e)[22] methods to synthesize CsPbX3 NCs efficiently under air. Some postsynthetic methods have also been developed to produce CsPbX3 NCs. In fact, there are several members in the cesium lead halide family, including CsPbX3, Cs4PbX6, and CsPb2X5. Although it is still under debate about which one is photoluminescent, an interesting transformation among the three phases has been noticed. Alivisatos et al. reported that CsPbX3 could be transformed to nonluminescent Cs4PbX6 NCs after being treated with an excessive amine.[23] Manna et al. pointed out that Cs4PbX6 NCs could be regarded as a lead-deficit structure, which can be converted to CsPbX3 by treating Cs4PbX6 with excessive PbX2.[24] Meanwhile, we recently considered Cs4PbX6 NCs as a CsX-rich structure.[25] Because of the high solubility of CsX in water, nonluminescent Cs4PbX6 NCs can be converted to highly luminescent CsPbX3 NCs by treating Cs4PbX6 NCs with water through a CsX-stripping process. CsPbBr3 nanocubes can also be converted to CsPb2Br5 nanosheets with the assistance of dodecyl dimethylammonium bromide (DDAB).[26]

It has been well-established that the morphologies of nanomaterials might affect their properties greatly. Much effort has therefore been devoted to preparing CsPbX3 NCs with different morphologies, including nanocubes,[14,27,28] nanowires,[29−33] nanospheres,[34,35] nanorods,[35,36] nanoplate,[37−40] two-dimensional nanosheets,[41−43] and so on. Cubelike CsPbX3 NCs are the most common morphology among reported works. Their formation may be attributed to two major factors. First, CsPbX3 is usually cubic or orthorhombic phase with near-cubic crystal lattice. On the other hand, the fast anion mobility of ionic CsPbX3 and fast reaction kinetics make it challenging to realize oriented growth. Low-dimensional CsPbX3 NCs tend to be prepared at a low reaction temperature. For example, precise control over the thickness of the platelike structure has been achieved at a lower reaction temperature. Alivisatos et al. reported that CsPbX3 nanoplates with a thickness ranging from one to several layers could be obtained when the reaction temperature was kept low (90–130 °C),[44] although the product was a mixture of nanoplates with various thicknesses. Recently, Zhang and co-workers found that simply heating the premixed precursors to the desired temperature (80–160 °C) could lead to the formation of CsPbBr3 nanoplates with a uniform thickness.[40] The thickness of nanoplates could be precisely tuned from single layer to several layers by controlling the reaction temperature, with a higher temperature resulting in thicker nanoplates. As the thickness of nanoplates is smaller than the Bohr radius (∼3.5 nm) of CsPbBr3, a blue emission could be observed. Meanwhile, the edge length could be varied by tuning the reaction time, with a longer reaction time leading to a larger edge length. Ligands were found to also play an essential role in the morphology control. Deng et al. reported a reprecipitation method to the synthesis of CsPbX3 nanocubes, nanospheres, nanorods, and nanoplatelets using ligands with different lengths of alkyl chains.[35]

Optical Properties

CsPbX3 NCs have been considered as one of the most promising materials in optoelectronic devices and other related fields, mainly because their optical properties are intrinsically different from those of other nanostructured luminescent materials. For example, without any surface treatment, CsPbX3 NCs can exhibit high PLQY and narrow-emission fwhm, while surface engineering is usually critical for traditional II–VI and III–V semiconductor quantum dots. More importantly, the PL emission of traditional metal chalcogenide and pnictide QDs is extremely sensitive to their particle size, leading to poor optical consistency for materials synthesized in different batches. In contrast, the PL peak position of CsPbX3 NCs is determined mainly by their halide composition. As reported by Kovalenko et al.,[14] the PL emission of CsPbX3 NCs can cover the full visible range (410–700 nm) when different halide precursors were used. This approach has proven to be a convenient and highly reproducible way to determine the optical properties in comparison with the traditional size-controlled method.

The PL peak position of CsPbX3 NCs can also be simply realized through post-chemical-transformation (Figure a–c).[45−47] Because of the high charge carrier mobility, the halide ions can be replaced partially or completely by each other. As demonstrated by the Kovalenko group, the anion-exchange process could be achieved in a fast and facile manner.[47] The bright PL emission could be tuned to cover the entire visible range (410–700 nm) by treating presynthesized CsPbX3 NCs with different amounts of other halide ions at room temperature. Interestingly, the morphology of the NCs could be well-maintained. The PLQYs were also maintained (20–80%), with narrow-emission line-widths (10–40 nm). In addition to the above-mentioned anion-exchange method, quantum-confinement-effect-induced PL spectra variation has also been observed when the dimension of NCs was smaller than their Bohr radius (∼2.5 nm for CsPbCl3, 3.5 nm for CsPbBr3, and 6.0 nm for CsPbI3).[14,37,48] Doping CsPbX3 NCs with divalent cations (such as Zn2+, Cd2+, Mn2+) represents another efficient and effective way to tune their optical properties.[49−52] For example, when Mn2+ ions were incorporated into wide-band-gap CsPbCl3 and CsPb(Cl/Br)3 NCs, the d–d transition of Mn2+ could cause strong yellow–orange emission (Figure e,f).[51]

Figure 2

(a) Schematic illustration of the anion-exchange process within the cubic CsPbCl3 NCs. (b) The XRD patterns indicate similar crystallite size before and after anion exchange. (c) Bright emission covering the entire visible spectral region can be realized with the anion-exchange approach. Images were modified with permission from ref (47). (d) Schematic illustration of partial replacement of Pb2+ with divalent cations. Images were modified with permission from ref (49). (e) Digital image and (f) PL spectra showing bright yellow emission of Mn2+-doped CsPbCl3 nanocrystals. Images were modified with permission from ref (51). (g, h) PLQY of CsPbBr3 can be improved to near 100% by postsynthetic surface treatment using thiocyanate. Images were modified with permission from ref (53).

In addition to the composition-dependent PL emission, another very attractive optical property of CsPbX3 NCs is the tolerance of ultrahigh density of defects (up to 1–2 atom %, typical as vacancies), which is much higher than conventional binary compound QDs, resulting in high PLQY without the need of any surface treatment. To improve the PLQY of traditional metal chalcogenide and pnictide QDs, one may use another semiconductor with a wider band gap as the shell material to passivate the surface. Because of the low stability of the ionic CsPbX3 NCs, it has been very difficult to achieve such surface passivation. Some progress has been achieved by modifying the synthetic strategies or post-treatment with other compounds. For example, Alivisatos et al. reported that thiocyanate treatment could dramatically promote the PLQY of CsPbBr3 NCs from 70% to nearly 100% (Figure g,h).[53] In this case, thiocyanate passivates the lead-rich surface, diminishing the shallow electron traps of CsPbBr3 NCs. Recently, Shen et al. prepared CsPbI3 NCs with up to 100% PLQY by using organo-lead trioctylphosphine-PbI2 as the precursor.[54]

Application

Thanks to their exciting photophysical properties, CsPbX3 NCs have been utilized in many photoelectric devices, including solar cells, light-emitting diodes (LEDs), photodetectors, and lasers. In particular, CsPbX3 NCs are promising candidates in the field of solar cells. The PCE of CsPbX3-based solar cells has grown rapidly over the past several years. In 2016, Luther et al. reported a solar cell based on phase-stable α-CsPbI3 QDs with a PCE of about 10.77%.[55] Later, a record-high PCE of 13.43% was achieved through an A-site cation halide salt (AX) treatment (Figure a–d).[56] Zhao et al. reported a solar cell of PCE of 11.8% with high phase stability at room temperature for months and at 100 °C for over 150 h.[57] In addition to the rising performance, the important advantages of CsPbX3-based solar cells also include higher thermal stability compared with OHP-based solar cells. As confirmed by Cahen et al., the performance of CsPbX3-based solar cells can be maintained for up to 2 weeks under constant illumination, which is much better than that of OHP-based solar cells.[58] Although much progress has been made, currently the PCE of IHP-based solar cells is still much lower than that of OHP-based solar cells. How to further improve the efficiency of IHP-based solar cells remains a quite challenging question.

Figure 3

(a) Schematic illustration and (b) SEM image of the cross-section of a solar cell device. (c) NREL-certified J–V characteristics from forward bias to reverse bias. (d) NREL-certified stabilized current at a constant voltage of 0.95 V. Images were modified with permission from ref (56). (e) Schematic illustration and (f) cross-sectional TEM image of an LED device. Images were modified with permission from ref (27). (g) Schematic illustration of the first CsPbI3 NCs photodetector. Images were modified with permission from ref (61).

Zeng et al. demonstrated the first application of CsPbX3 NCs in LEDs.[27] The structure of the device is illustrated in Figure e,f. The color of LEDs can be tuned from blue to orange by adjusting the content and the category of anions. However, the external quantum efficiency (EQE) of the first reported CsPbBr3 LED was only 0.12%. The low EQE of CsPbX3-based LEDs could be attributed to the excessive ligands that form an insulating layer. To solve this problem, Zeng et al. developed a new method to control the ligand density and balance the surface passivation and carrier injection. The EQE has been promoted to 6.27%.[59] More recently, Chiba et al. reported a novel washing process using an ester solvent to remove excess ligands and achieved the highest EQE (8.73%) of CsPbBr3-based LED devices.[60]

The photodetector is another exciting application of IHP NCs. The first reported IHP-based photodetector was made by Ramasamy et al.[61] As illustrated in Figure g, CsPbI3 NCs were used as the main component. A very good on/off photocurrent ratio of 105 was achieved by using the very simple device. This work opened a new door to the application of IHP NCs. After that, much effort has been devoted to improving the performance of IHP-NC-based photodetectors. For example, Zeng et al. developed a room-temperature healing method to treat CsPbBr3 film, resulting in significantly improved performance of photodetectors.[62] In addition to the cubelike structure, CsPbX3 NCs with different morphologies, such as nanorods[63] and nanosheets,[41,64] have also been used to improve the performance by utilizing their unique feature in charge carrier transfer. Overall, the reports on the IHP-based photodetectors are limited, and more efforts are still needed to improve the performance.

Because of their high absorption coefficient and low density of defects, IHP NCs have also been used to fabricate lasing devices.[65−67] For example, Sun et al. first reported the potential application of CsPbX3 QDs in lasers.[67] Xiong et al. prepared high-quality CsPbX3 nanoplates through a vapor-phase van der Waals epitaxy method,[65] and then used the well-defined product for lasing. They were able to realize multicolor and low-threshold lasing, and obtained so far one of the highest values of mode line-width (0.14–0.15 nm). Despite these achievements, the stability and the excitation mechanism are still quite challenging problems to be addressed.

Outlook

As summarized above, intensive efforts have been made in the development of all-inorganic metal halide perovskite nanomaterials since the first study in 2015.[14] Significant progress has been achieved in the controlled preparations of IHP NCs and optimizations of their properties over the past three years. Their promising applications in different areas have also been partially demonstrated. Despite these great successes, the research in this field is still in its early infancy. Several issues must be addressed before the widespread practical applications of IHP NCs become possible. In this section, we discuss our perspectives regarding challenges and future research in this area.

Stability Issue

Presumably, the most prominent issue with IHP NCs is still their structural stability against chemical (particularly moisture), thermal, and photodisturbances. For example, the colloidal stability of IHP NCs is often questionable because of the relatively weak binding of common ligands to the particle surface. Acids and amines with long alkyl chains, such as oleic acid and oleylamine, are the most widely used ligands for the synthesis of IHP NCs. These ligands can nevertheless easily detach themselves from the nanocrystal surface, leading to the aggregation and structural damage of IHP NCs. New ligand chemistry is therefore needed. Recently, Sun et al. attempted to address this issue by replacing traditional oleic acid/olelyamine with octylphosphonic acid (OPA) and found that the stability of CsPbX3 NCs could be significantly enhanced.[68] Thanks to the relatively strong interaction between OPA and lead ions, the resulting CsPbX3 NCs remained highly photoluminescent after eight purification cycles. It was found that OPA could provide much better protection even in the presence of fewer ligands on CsPbX3 NC surfaces (only ∼4.6% ligand left in the OPA-CsPbX3 system in comparison with ∼29.7% for OA/OLA-CsPbBr3). As a result, the EQE of the LED device has been promoted to 6.5% (Figure a). Wu and co-workers also found that the addition of trioctylphosphine oxide (TOPO) into the oleic acid/oleylamine system can significantly improve the stability of IHP NCs against antisolvent cleaning.[69]

Figure 4

(a) Schematic structure of OPA-capped CsPbBr3 NCs. (b) J–V–L curve of devices based on different ligands (blue line for OA/OLA-capped CsPbBr3 NCs, and red line for OPA-capped CsPbBr3 NCs). Images were modified with permission from ref (68). (c) Schematic structure of CsPbBr3 NCs capped by long-chain zwitterionic molecules. Images were modified with permission from ref (70).

Zwitterionic ligands that possess several anchoring groups can also provide effective protection for IHP NCs (Figure b).[70] Compared with conventional acids and amines, zwitterionic ligands show stronger adhesion to the surface of IHP NCs via special chelating effect. Ligand exchange on presynthesized IHP NCs is another effective way to passivate the surface of NCs. Bakr et al. reported that both the PLQY and stability of CsPbX3 NCs could be dramatically improved after ligand exchange with didodecyl dimethylammonium bromide and bidentate 2,2′-iminodibenzoic acid.[71,72] Recently, it has been found that anchoring CsPbBr3 NCs onto a substrate can improve their stability. By using a presynthesized aminated silica sphere as the substrate, Zeng et al. successfully grew CsPbBr3 NCs on the silica spheres.[73] Since all NCs are anchored on the surface of silica, photoinduced regrowth and deterioration of NCs were inhibited, leading to enhanced stability against moisture and light exposure.

In addition to the surface passivation with organic ligands, another way of improving the stability of IHP NCs is to encapsulate these vulnerable NCs into a stable and inert shell made of different materials, such as graphene, silica, amorphous alumina matrix, and polymer.[74−81] For example, after encapsulation within silica/alumina monolith, CsPbBr3 NCs showed much higher stability against light irradiation than naked ones (Figure a–c).[81] Although researchers have proposed some strategies to maintain the original properties of IHP NCs after the encapsulation, there are still several major questions.

Figure 5

(a) TEM images of CsPbBr3 NCs-silica/alumina monolith (SAM). Photostability of the CsPbBr3 QDs-SAM powder (b) under illumination with a 470 nm LED light and (c) sealed with PDMS on the LED chip (5 mA, 2.7 V). The images were modified with permission from ref (81). (d) TEM image of the obtained CsPbBr3/SiO2 Janus NCs. (e) HRTEM image of a single CsPbBr3/SiO2 NC. (f) HAADF-STEM image and elemental mapping images. (g) Photographs of (I) CsPbBr3/SiO2 NCs, (II) WT-CsPbBr3 NCs, and (III) HI-CsPbBr3 NCs thin film stored in humid air (40 °C and humidity of 75%). The images were modified with permission from ref (82).

First, how to modify IHP NCs with another material on a single-particle level is still a big challenge. In many applications, e.g., cell labeling, small particles are preferred. In the examples mentioned above, many IHP NCs were encapsulated together into the shell material, resulting in very large particles (several hundred nanometers to several micrometers in size). All of the reported IHP NCs are hydrophobic because their surfaces are covered with long-alkyl-chain ligands. Since most of the reported shell materials are hydrophilic, it is challenging to grow hydrophilic materials onto the hydrophobic surface directly. Recently, for the first time, we found that Janus CsPbX3/oxide (SiO2 and Ta2O5) nanoparticles can be obtained through an interfacial synthesis process,[82] in which presynthesized Cs4PbX6 NCs and oxide precursors were mixed in hexane and then treated with water. At the water/hexane interface, Cs4PbX6 NCs were converted to CsPbX3 NCs through a CsX-stripping process because of the high solubility of CsX in water.[25] Meanwhile, the hydrophobic ligands were removed during the stripping process, leaving a hydrophilic surface to the environment. Simultaneously, the hydrolysis of oxide precursor happened at the interface, leading to the formation of Janus NCs (Figure d–f). Thanks to the surface modification, the as-prepared Janus nanoparticles showed improved stability in air (Figure g). However, how to achieve a full coating of IHP NCs on a single-particle level remains a challenge to be addressed.

Second, how to coat IHP NCs with a stable and active shell is a problem. For traditional II–VI or III–V semiconductor NCs, encapsulating them into an active and more stable shell can not only improve the stability, but also tune the photophysical properties. One of the most widely used shell materials is ZnS.[83] However, most of the reported shell materials for IHP NCs are organic or amorphous materials. No full coating of IHP NCs into an active layer has been reported. This question can be partially attributed to the hydrophobic surface we have discussed above. On the other hand, the strong interaction between Pb and chalcogen anions (S2–, Se2–, or Te2–) prevents the coating of the metal chalcogenide. To achieve such kind of coating, developing some sandwich-like structures might be a plausible approach, in which an ultrathin inert shell could be coated onto the surface of IHP NCs first, followed by the coating of the metal chalcogenide layer (Figure ). By engineering the thickness and composition of the middle layer, the photophysical properties might also be tuned. Semiconductors, oxides, and polymers are promising shell materials to protect IHP NCs. With the protection of shell materials, the IHP NCs could be dispersed into water or other polar solvents. More importantly, the toxicity can be minimized by the core–shell configuration. These advantages may bring new opportunities to IHP NCs in some important applications, such as LCD displays, photocatalysis, and cell imaging.

Figure 6

Schematic illustration of the design of sandwich-like structures with possible shell materials and promising applications.

Toxicity of Lead

The toxicity of lead is another inescapable problem of IHP before industrialization. So far, some progress has been achieved in the preparation of environmentally friendly IHPs by replacing lead with low-toxicity or nontoxic metal, such as tin, bismuth, and antimony. For example, CsSnX3 has been considered as a promising alternative, which exhibits different properties due to the introduction of Sn.[84−86] Orthorhombic CsSnI3 NCs reported by Kanatzidis et al. not only show the properties of a p-type direct band gap but also possess high hole mobility, which come from the intrinsic defects associated with Sn vacancies.[87] However, the development and application of CsSnI3 NCs have been limited seriously by their poor stability as they are extremely sensitive to moisture, oxygen, and thermal treatment. What’s worse, its PLQY (<1%) is too low to meet the requirement for practical applications. Cs3Bi2X9 and Cs3Sb2X9 NCs have also been successfully synthesized, which exhibited much higher PLQY compared to CsSnX3 (Figure a–c). However, the PL spectra can cover only part of the visible range (350–560 nm).[88−91] Novel lead-free double-perovskite NCs, such as Cs2AgBiX6 and Cs2InAgX6 NCs,[92−94] have been successfully obtained, which provide new members for the development of perovskite materials (Figure d–f). In general, however, these lead-free IHP NCs have lower efficiency compared with lead-based ones. It is still urgently needed to develop new lead-free materials with high performance or find effective means to improve the performance of the current ones. The rapid development of computational chemistry might be able to offer hints to the rational design of more potential candidates. Zhao et al. exploited lead-free halide perovskite materials using first-principle calculations. Some double-perovskite structures, such as Cs2InSbCl6 and Cs2InBiCl6, have been regarded as promising materials for stable and efficient candidates for the application of solar cells.[95]

Figure 7

(a) Cs3Bi2Br9 unit cell, XRD patterns of Cs3Bi2Br9 NCs, and TEM image of Cs3Bi2Br9 NCs. (b) Photographs of as-obtained colloidal Cs3Bi2X9 and XRD patterns of NCs containing pure and mixed halides. (c) Steady-state absorption and PL spectra of NCs containing pure and mixed halides. The images were modified with permission from ref (88). (d) Structure of Cs2AgBiBr6. (e) TEM image of Cs2AgBiCl6 NCs. (f) TEM image of Cs2AgBiBr6 NCs. The images were modified with permission from ref (93).

Photocatalysis

The potential application of IHP NCs in photocatalysis is another very attractive topic because of their excellent photophysical properties, including tunable band gap, high absorption coefficient, broad absorption spectrum, high charge carrier mobility, and long charge diffusion lengths. Kuang and co-workers reported that CsPbBr3 NCs could be used to trigger the photocatalytic reduction of CO2 in ethyl acetate (Figure a,b),[96] and observed relatively good stability. More importantly, a 25.5% increase in the catalytic performance has been observed when CsPbBr3 NCs were combined with graphene oxide. The improved performance has been attributed to the enhanced electron extraction and transport. Recently, lead-free IHP NCs have also been used for photocatalytic reduction of CO2. By adopting a similar reaction system, Cs2AgBiBr6 NCs were used as the photocatalysts (Figure c,d).[92] Although the photocatalytic performance is still poor compared to other reported semiconductor catalysts, this work demonstrates the vast potential of IHP NCs in environmentally friendly photocatalysis.

Figure 8

(a) Schematic illustration of CO2 photoreduction over the CsPbBr3 QD/GO photocatalyst. (b) Catalytic performance of CsPbBr3 QD and CsPbBr3 QD/GO. The images were modified with permission from ref (95). (c) Catalytic performance of Cs2AgBiBr6 NCs in CO2 reduction. (d) Schematic illustration of the photoreduction of CO2 over Cs2AgBiBr6 NCs. The images were modified with permission from ref (92).

Despite the successful demonstration of IHP NCs in photocatalysis, there is still a long way to go toward practical applications. First, the recombination rate of photogenerated charge carriers in IHP NCs is high, leading to low utilization efficiency of electrons and poor performance. Second, the low stability of IHP NCs in frequently used polar media such as water and alcohol is still the biggest problem. For example, compared to nonaqueous ethyl acetate, water is a much better reaction medium for CO2 reduction or other photocatalysis reactions. However, none of the reported IHP NCs can be dispersed in water so far. Additionally, when carbon-containing nonpolar solvents were used as the reaction media, CO or CH4 might be generated from the solvents rather than from the photocatalytic reaction, causing big trouble in the productivity calculation. One unexplored but possible way to overcome these challenges is to modify such IHP NCs with wide-band-gap materials or metal NCs that can form hetero- or core–shell nanostructures. In this way, the stability might be improved, and the charge separation and the catalytic performance could be enhanced.