Even if we will not be able to find alternative materials that match all of the current performances of lead halide perovskite nanocrystals, we should not be discouraged because these alternatives can still be very useful in various light emission applications.

Even if we will not be able to find alternative materials that match all of the current performances of lead halide perovskite nanocrystals, we should not be discouraged because these alternatives can still be very useful in various light emission applications. Halide perovskites (ABX₃) are formed by a three-dimensional (3D) network of corner-sharing BX₆^4– octahedra (B, divalent metal cation; X, Cl, Br, I), stabilized by large A monovalent cations (Figure 1a, top left panel).(1) The perovskite structure and its variants have a low density of high-energy structural defects, as it is energetically costly to displace the ions in the perovskite lattice. The most studied members of this family are based on lead (with APbX₃ formula) and are often referred to as defect tolerant, since in them low-energy structural defects (vacancies, undercoordinated atoms at grain boundaries) create electronic states that are either very shallow or are nested in the valence/conduction band, thus being very benign (Figure 1a, left panels). The corresponding nanocrystals possess high photoluminescence (PL) quantum yields (QYs) even without specific surface treatments, and their optical emission can be easily modified, especially through mixed halide compositions.(2) APbX₃ nanocrystals represent valid alternatives to traditional semiconductor quantum dots, for example, in display technologies, due to their exceptional brightness and color purity. Despite such excellent properties, the toxicity of Pb²⁺ ions has stimulated the search for ABX₃ materials based on other ns² cations, such as Sn²⁺ or Ge²⁺. These have been found to be plagued by poor cation stability and/or optical performance,(3) although very recent reports have shown improvement in stability for CsSnX₃ nanocrystals (with red-NIR emission) through the use of “antioxidative” and strongly coordinating oxalate ligands.(4) The search for promising metal halide materials has focused so far mostly on halide double perovskites (A₂B′B″X₆,(5) with B′ being a monovalent and B″ a trivalent cation) and on perovskite-related metal halides. The latter often present lower connectivity of the coordination polyhedra compared to perovskites. Figure 1. (a) Top: structural models of (lead) halide perovskites (left) and halide double perovskites (right). Bottom: corresponding energy level diagrams with the nonbonding orbitals of the double perovskite system emerging above the valence band maximum (VBM) due to the more bound ns–np orbitals of the +3 B cation that are pushed downward in energy. (b) Energy level diagram of a typical 0D BX₆ octahedron with ns² configuration of the central ion and octahedral O_h symmetry. The inclusion of spin–orbit coupling affects the lowest excited states by splitting the energy levels into two manifolds, one four-degenerate and the other two-degenerate (labeled according to double group symmetry representation). The ground state is unaffected by the inclusion of spin–orbit coupling. (c) Nonradiative decay in 3D double perovskite and 0D metal halides. Although in both cases the exciton is localized inside the crystal (black dotted line), in the 3D configuration the connection between the octahedra facilitates the phonon-mediated transport of carriers to the surface, at the trap location. In 0D systems, carriers instead have to hop between disconnected octahedra to reach the surface trap. In the latter case, the nonradiative channel is strongly inhibited. Double perovskites, initially prepared as bulk crystals or microcrystals in their pure undoped form, are usually poor optical performers: either they have an indirect band gap, or the gap is direct but it features parity forbidden optical transitions.(6) Progress came by introducing dopants and by applying various alloying strategies, both at the B′ and the B″ cation sites with the photogenerated carriers relaxing into spatially localized electron and hole states created by these dopant/alloy cations.(7−10) From this localization, a (self) trapped exciton is formed and radiative recombination occurs efficiently. Because of the multitude of possible local configurations in which the dopant(s) are distributed in the lattice and coupled to vibrational modes, the PL emission is overall broad in energy and also Stokes shifted. Here, we come to the first limitation of these materials: so far, no narrow band emission has been observed that is nearly comparable, in terms of color purity, to that of APbX₃ perovskites. Nevertheless, very high PLQY were reported for bulk single crystals and polycrystalline thin films, for example 86%(7) or even close to 90%(8) for Cs₂AgInCl₆ through concomitant alloying of Ag⁺ with Na⁺ and Bi³⁺ doping, further increased to 98.4% and 98.6% by codoping with 1% Ni and 1% Ce,(8) or up to 93% in Cs₂NaInCl₆:xSb and Cs₂KInCl₆:xSb compositions.(9) Considering that a large number of alloying and doping strategies has been already explored, it is unlikely for these materials to elude the broadband emission in favor of a narrow band emission. Another issue with double perovskites emerged when research groups tried to prepare their nanocrystal versions, even after adopting the alloying and doping strategies developed for microcrystals. Such nanocrystals exhibit much lower PLQY (around 20–36%),(10−12) which could only be modestly improved by surface ligand optimization (up to 37%).(13) It was found that these nanocrystals are plagued by deeper trap states than those in APbX₃ nanocrystals, and the undercoordinated surface chloride ions are likely the main culprit.(13) This surface “intolerance” stems from the valence ns and np atomic orbitals of the B″ cation, which, due to the high +3 charge, are more bound to the nuclei and thus pushed lower in energy than the halide valence np orbitals (Figure 1a, right panels). This renders the valence band edge mostly composed of halide nonbonding orbitals and thus more susceptible to form deep traps when the coordination of these ions at the surface is reduced. A possible solution to the problem is to replace all surface halides by wide gap organic ligands; however, it appears very difficult to saturate all surface dangling bonds due to the steric hindrance that is easily generated in the ligand shell.(13) To date, only a small library of organic ligands has been tested. Hence, it is not excluded that a broader search (also aided by calculations) could identify more appropriate surface passivating ligands in terms of binding strength, reduced hindrance, and ability to eliminate surface traps. Another possible way to eliminate such trap states is to grow an epitaxial shell of another inorganic material (most likely another metal halide) but this is neither straightforward nor will it guarantee that no interfacial trap states are formed. So far, for double perovskites mainly chloride-based compositions have been studied. However, we can expect that surface defect intolerance is intrinsic also to compositions based on bromide and iodide, despite the larger lattice spacing. The same should hold for the vacancy ordered perovskites, where the +4 charge on the B″ cation will push even further the ns and np energy levels downward, favoring the emergence of nonbonding (dangling) orbitals at the band edge. In the case of other metal halide materials, the connectivity of the coordination polyhedra is usually lower than that of perovskites and can be 2D, 1D, or 0D.(14−18) From reports that have appeared so far, whenever PL was observed, especially for materials containing cations with ns² configuration, this was again broad and considerably Stokes-shifted from the absorbing states, even at low temperature, pointing again to a trapped-exciton type of emission.(19) We can indeed expect that the reduction in dimensionality obtained by disconnecting the polyhedra will transform the exciton emerging from delocalized 3D bands (Wannier-Mott type) to a localized one (Frenkel type). In other words, the self-trapped emission will reproduce essentially that of an exciton of a molecule embedded in a solid matrix; for a 0D composition, this will be for example the case of a molecular [BX₆]ⁿ⁻ moiety. In heavy B elements, spin–orbit coupling will then play a key role in broadening the emission peak by splitting, for example, the unoccupied np degenerate orbitals of octahedral t_1u symmetry into two manifolds separated by a few tenths of electronvolts (Figure 1b). Once photoexcited, these states could be evenly occupied and exhibit dual or broad emission. Additionally, the nonradiative channels in lower dimensional metal halides are likely dominated by strong electron–phonon coupling effects, which are prone to efficiently quench and broaden the PL. Interesting developments are however emerging from doping of these various metal halide systems.(20) For metal halides, the transition from bulk to nanocrystals may be less dramatic in terms of optical performances: as the carrier mobility is further constrained when the extent of connectivity is decreased, it becomes less and less likely for the photogenerated carriers to probe surface trap states (Figure 1c). Hence, we can expect that the nanocrystal counterparts will likely behave as the bulk, but the quality of the optical efficiency will still depend on how good the (self)-emitter is, and in addition it remains to be seen whether doping can be efficiently translated to the corresponding nanocrystals. Even if we will not be able to find alternative materials that match all of the current performances of lead halide perovskite nanocrystals, we should not be discouraged, because these alternatives can still be very useful in various light emission applications. First, even only broadband emission, albeit with QY close to 100%, is as good as it will get, these nanocrystals can be used in white light emission applications. Also, they can be used in luminescent solar concentrators and scintillators, if absorbing and emitting states are well separated in energy.(21) Remote thermometry based on the marked temperature-dependent lifetime is another exciting application.(22) Other developments can come from preparing metal halide nanocrystals emitting efficiently in the near-infrared region (for various applications), through doping, exploiting, for example, the quantum cutting effect through which PLQY above 100% can be achieved.(23) Although initial results have been based on doped APbX₃ nanocrystals, it is not clear whether this mechanism can be more or less efficient in other metal halides. In conclusion, the quest for alternative metal halide nanocrystals to rival the current lead-based perovskite nanocrystals still remains full of obstacles, and the only viable alternatives to date are not metal halides but quantum dots from the III–V family, InP in particular.(24) On the other hand, if we only seek efficient emission, or emission in other spectral ranges (blue-UV, or NIR), then we believe that the search for alternative materials still looks “bright”. The authors declare no competing financial interest. The authors declare no competing financial interest. We thank Luca De Trizio and Muhammad Imran for helpful discussions. This article references 24 other publications.

中文翻译：

---

卤化物钙钛矿纳米晶体有很好的替代品吗？

即使我们无法找到与卤化钙钛矿纳米晶体的所有当前性能都匹配的替代材料，也不要气be，因为这些替代品在各种发光应用中仍然非常有用。

即使我们无法找到与卤化钙钛矿铅纳米晶体目前所有性能均相匹配的替代材料，也不要气we，因为这些替代品在各种发光应用中仍然非常有用。卤化物钙钛矿（ABX ₃）由角共享的BX ₆ ^{4 –}八面体（B，二价金属阳离子； X，Cl，Br，I）的三维（3D）网络形成，并由大的A单价阳离子稳定（图1）。 1a，左上图）。（1）钙钛矿结构及其变体具有低密度的高能结构缺陷，因为在钙钛矿晶格中置换离子在能源上是昂贵的。这个家庭中研究最多的成员基于铅（使用APbX ₃式），通常被称为耐缺陷的，因为在其中低能结构缺陷（空位，晶界处原子配位不足）产生的电子态要么很浅，要么嵌套在价/导带中，因此非常良性（图1a，左图）。相应的纳米晶体即使不经过特定的表面处理，也具有很高的光致发光（PL）量子产率（QYs），并且可以轻松地修改其光发射，尤其是通过混合卤化物成分。（2）APbX ₃纳米晶体是传统半导体量子点的有效替代品，例如，在显示技术中，由于其出色的亮度和色纯度。尽管具有如此优异的性能，但Pb ²⁺的毒性离子刺激了基于其他n s ²阳离子（例如Sn ²⁺或Ge ^2+）的ABX ₃材料的搜索。已经发现这些问题受阳离子稳定性和/或光学性能差的困扰。（3）尽管最近的报告显示，通过使用“抗氧化”和强配位作用，CsSnX ₃纳米晶体（具有红色近红外发射）的稳定性有所改善。草酸盐配体。（4）迄今为止，寻找有前途的金属卤化物材料的研究大多集中在卤化物双钙钛矿上（A ₂ B′B″ X ₆，（5），其中B'是一价的，B”是三价的阳离子）和钙钛矿相关的金属卤化物。与钙钛矿相比，后者通常呈现出较低的配位多面体连接性。图1.（a）上图：（铅）卤化钙钛矿（左）和卤化钙钛矿（右）的结构模型。下图：相应的能级图，由于+3 B阳离子的束缚n s– n p轨道被向下推动，双钙钛矿体系的非键轨道出现在价带最大值（VBM）以上。（b）一种典型的0D BX的能级图₆八面体与Ñ小号²中心离子和八面体的配置ö _ħ对称。自旋-轨道耦合的影响通过将能级分为两个流形（一个是四个简并的，另一个是两个简并的）（根据双组对称表示来标记）来影响最低的激发态。自旋轨道耦合不影响基态。（c）3D钙钛矿和0D金属卤化物的非辐射衰减。尽管在这两种情况下，激子都位于晶体内部（黑色虚线），但在3D构造中，八面体之间的连接有助于在陷阱位置处由声子介导的载流子传输至表面。在0D系统中，运营商必须在断开的八面体之间跳来跳到水面陷阱。在后一种情况下，非辐射通道被强烈抑制。双钙钛矿，最初以纯非掺杂形式制备为块状晶体或微晶通常是较差的光学性能：要么具有间接带隙，要么间隙是直接的，但具有奇偶禁止的光学跃迁。（6）引入掺杂剂和在B'和B''阳离子位点应用各种合金化策略，光生载流子松弛到由这些掺杂剂/合金阳离子产生的空间局部电子态和空穴态。（7-10）从这种局部化中，一个（自）被俘获激子形成并有效地发生辐射复合。由于掺杂剂分布在晶格中并耦合到振动模式的多种可能的局部构型，因此PL发射的能量总体较宽，斯托克斯位移也很大。这里，_3个钙钛矿。然而，据报道，通过将Ag ⁺与Na ⁺和Bi ³⁺合金化，Cs ₂ AgInCl _6的体单晶和多晶薄膜的PLQY非常高，例如86％（7）甚至接近90％（8）。通过在1 Cs ₂ NaInCl ₆：x Sb和Cs ₂ KInCl ₆：x中与1％Ni和1％Ce，（8）或最高93％共掺杂，掺杂进一步增加到98.4％和98.6％。锑的组成。（9）考虑到已经探索了大量的合金化和掺杂策略，这些材料不太可能通过宽带发射而转向窄带发射。当研究小组试图采用纳米钙晶体时，即使采用了为微晶开发的合金化和掺杂策略，也出现了双重钙钛矿的问题。此类纳米晶体的PLQY较低（约20–36％）（10-12），只有通过表面配体优化才能适度改善（高达37％）。（13）发现这些纳米晶体受更深陷阱的困扰状态比APbX ₃纳米晶体中的状态要差，并且表面氯离子配位不足可能是主要的罪魁祸首。（13）这种表面“不宽容”源于​​化合价nB''阳离子的s和n p原子轨道，由于高+3电荷，与原子核的结合更多，因此其能量比卤化合价n低p轨道（图1a，右图）。这使得价带边缘主要由卤化物的非键合轨道组成，因此当这些离子在表面的配位降低时，更易于形成深陷阱。解决该问题的一种可能方法是用宽间隙有机配体代替所有表面卤化物。然而，由于容易在配体壳中产生空间位阻，使所有表面的悬空键饱和似乎非常困难。（13）迄今为止，仅测试了一个小的有机配体库。因此，不排除在结合强度，减少的障碍和消除表面陷阱的能力方面，更广泛的搜索（也可以借助计算）可以确定更合适的表面钝化配体。消除这种陷阱态的另一种可能方法是生长另一种无机材料（很可能是另一种金属卤化物）的外延壳，但这既不直接，也不能保证不会形成界面陷阱态。迄今为止，对于双钙钛矿，主要研究了基于氯的组合物。然而，我们可以预期，尽管晶格间距较大，但表面缺陷耐受性也是基于溴化物和碘化物的成分所固有的。空缺有序钙钛矿应保持同样的状态，其中B“阳离子上的+4电荷将进一步推动 我们可以预期，尽管晶格间距较大，但表面缺陷的不耐受性也是基于溴化物和碘化物的成分所固有的。空缺有序钙钛矿应保持同样的状态，其中B“阳离子上的+4电荷将进一步推动 我们可以预期，尽管晶格间距较大，但表面缺陷的不耐受性也是基于溴化物和碘化物的成分所固有的。空缺有序钙钛矿应保持同样的状态，其中B“阳离子上的+4电荷将进一步推动n s和n p的能级向下，有利于在频带边缘出现非键（悬挂）轨道。在其他金属卤化物材料的情况下，配位多面体的连接性通常低于钙钛矿，并且可以为2D，1D或0D。（14-18）从迄今为止的报道中，每当观察到PL时，特别是对于含有n s ²阳离子的材料构型，即使在低温下，它又宽泛且从吸收态发生了斯托克斯位移，这再次表明是捕获激子类型的发射。（19）我们确实可以预期，通过断开多面体连接而获得的尺寸减小将将激元从非定域的3D波段（Wannier-Mott类型）转换为局部的激子（Frenkel类型）。换句话说，自陷发射本质上将重现嵌入在固体基质中的分子的激子的发射。为0D组合物，这将是例如一个分子[BX的情况下₆ ] ^{ñ -}部分。在重质B元素中，自旋轨道耦合将通过分裂例如未占据的n来在扩大发射峰中起关键作用。p八面体的退化轨道t _1u对称地分成两个由十分之几电子伏特隔开的歧管（图1b）。一旦被光激发，这些状态就可以被均匀地占据，并呈现出双重或广泛的发射。另外，低维金属卤化物中的非辐射通道可能被强的电子-声子耦合效应所支配，这易于有效地淬灭并加宽PL。然而，从这些各种金属卤化物系统的掺杂中出现了有趣的发展。（20）对于金属卤化物，从块状到纳米晶体的转变在光学性能方面可能不那么引人注目：因为当连通性程度受到限制时，载流子迁移率进一步受到限制。减少，光生载流子探测表面陷阱状态的可能性越来越小（图1c）。因此，我们可以预期，纳米晶体的对应物可能会表现为整体，但是光学效率的质量仍然取决于（自）发射极的质量，此外，还有待观察是否可以将掺杂有效转化为相应的纳米晶体。即使我们将无法找到与卤化钙钛矿纳米晶体的所有当前性能相匹配的替代材料，也不要气we，因为这些替代方案在各种发光应用中仍然非常有用。首先，即使QY接近100％，即使只有宽带发射也能达到预期的效果，这些纳米晶体也可以用于白光发射应用。同样，如果吸收和发射态在能量上能很好地分开，它们也可以用于发光太阳能聚光器和闪烁器中。_对于3种纳米晶体，尚不清楚这种机理在其他金属卤化物中是否或多或少有效。总之，寻求替代金属卤化物纳米晶体以与目前的铅基钙钛矿纳米晶体相竞争仍然充满了障碍，迄今为止，唯一可行的替代方法不是金属卤化物而是III-V系列的量子点，特别是InP。 （24）另一方面，如果我们仅寻求有效的发射或其他光谱范围内的发射（蓝紫外或NIR），那么我们认为寻找替代材料的过程仍然看起来“光明”。作者宣称没有竞争性的经济利益。作者宣称没有竞争性的经济利益。我们感谢Luca De Trizio和Muhammad Imran进行的有益讨论。本文引用了其他24个出版物。