1. Introduction

In general, the term “perovskite” is used to describe any material with the same structure as CaTiO₃. Pure perovskites present a general formula of AMX₃, where A and M are cations and X is an anion that binds to both cations. While M is coordinated to six X anions, A is coordinated to twelve X anions (Fig. 1). Therefore, they consist of anionic M-X semiconducting frameworks and charge-compensating cations [1].

 

Fig. 1 Schematic representation of the dimensionality (D) of the inorganic framework of metal halide perovskites.

Download Full Size | PDF

In the case of lead halide perovskites, M is Pb and X is a halogen (Cl, Br, I, or a combination of them). This family of perovskites can be further divided into two groups of perovskites according to the nature of the cations: all inorganic lead-halide perovskites and organic-inorganic lead halide perovskites (or hybrid lead halide perovskites). Lead halide perovskites contain an anionic lead-halogen semiconducting framework and charge-compensating ammonium salts (hybrid lead halide perovskites) or inorganic cations (all inorganic lead-halide perovskites). In principle, the dimensionality (D) of the inorganic framework can vary from three to zero. Small-sized organic or inorganic cations can fit into the PbX₆ octahedra of the 3D framework. If the cation is too large, the 3D perovskite structure would be inappropriate and the dimensionality of the inorganic framework would change to 2D, 1D, or 0D (Fig. 1). In the 0-D hybrid perovskite structure, the PbX₆ octahedra are isolated, while in the 1D the PbX₆ octahedra are connected in a chain, and in the 2D the PbX₆ octahedra are connected in layered sheets at the corners [2–4] The stoichiometry of the perovskite changes with the dimension of the inorganic framework. Thus, the APbX₃, A₂PbX₄, A’₂APbX₅, and A₂PbX₆ stoichiometry possesses a 3D, 2D, 1D, and 0D framework, respectively.

Regarding the dimensionality in the material, genuine nanomaterials have three dimensions on the nanoscale [5]. In principle, perovskite nanoparticles (NPs) can be prepared with different shapes and with inorganic framework dimensionalities varying from 0D to 3D. In this report we focus on the preparation and unique optical properties of organic-inorganic lead halide perovskite and all-inorganic cesium lead perovskite nanoparticles with a 3D framework and therefore with the APbX₃ general formula. Compared with the preparation of other types of NPs, such as semiconducting CdSe and metal NPs, the synthesis of halide perovskite NPs has been a very recent target and their successful preparation is partially arising by making use of the knowledge acquired from decades of synthesis of the other types of nanoparticles.

2. Ammonium-lead halide perovskite nanosystems

Nanoparticulate APbX₃ material can be prepared by either a template method, by making use of nanoporous films (template method), or by using organic ligands as the agents that control the growth and eventually allow the NPs to disperse in an organic medium (non-template method).

2.1 The template method

Nanoparticulate lead halide perovskite MAPbX₃ (MA = CH₃NH₃⁺) material was first prepared by using TiO₂, Al₂O₃, and ZrO₂ mesoporous films combined with the perovskite precursors (MAX and lead halide). The deposition of the MAPbX₃ perovskites can be performed via a one-step or a two-step coating procedure. The one-step spin-coating technique was used in the pioneer work of Miyasaka et al. [6] for the preparation of MAPbX₃ (A = CH₃NH₃⁺, X = Br^- and I^-) nanoparticles in mesoporous material, using a solution of MAX and lead halide (PbX₂) dissolved in a polar aprotic solvent, DMF or γ-butyrolactone. The porosity of the TiO₂ film (thickness 8-12 µm), prepared with nanocrystalline TiO₂, allows the formation of 2-3 nm-sized MAPbBr₃ and MAPbI₃ NPs with a cubic and tetragonal perovskite structure, respectively, on the TiO₂ NP surface. Studies on the application of the MAPbI₃/TiO₂ material as visible-light sensitizers in photovoltaics showed that this material (with an extended spectral response up to 800 nm) produces a high photoelectric conversion in dye-sensitized liquid-junction solar cells [6].

Later, Park et al. reported the formation of MAPbI₃ and ETAPbI₃ (ETA = CH₃CH₂NH₃⁺) NPs on TiO₂ film [7–9]. Thus, an equimolar ratio of MAI and PbI₂ in γ-butyrolactone was spread over the TiO₂ film (20 nm-sized TiO₂ NPs) and the solution was allowed to penetrate into the film pores before the spin-coating step. Then the film was dried at temperatures from 40 °C to 160 °C for 30 minutes. The color of the perovskite on the TiO₂ film varied from yellow to black (absorbance in the 400-850 nm range) when the concentration of the precursor solution increased from 10% to 40%. The MAPbI₃ perovskite (2.5 nm average diameter) crystallized in a tetragonal perovskite structure, distributed homogenously within the pores and on the surface of the TiO₂ NPs (Fig. 2(a)-2(b). On the other hand, ETAPbI₃ NPs crystallized in the orthorhombic phase, they were distributed sparsely on the TiO₂ surface, and their average size was 1.8 nm [8].

 

Fig. 2 a-b) TEM images of MAPbI₃ nanoparticles deposited on TiO₂ (Scale bar 20 nm and 2 nm) Adapted from [7], Copyright 2011 The Royal Society of Chemistry. c) Cross-sectional SEM image of a photovoltaic device with MAPbI₃ nanocristal on TiO₂, prepared by the two-step sequencial method. Reprinted from [18], Copyright 2013, Macmillan Publishers Limited.

Download Full Size | PDF

The MAPbI₃ nanoparticulate material presents an absorption coefficient of 1.5 x 10⁴ cm⁻¹ at 550 nm [7], similar or even higher than that of the typical photovoltaic materials used as sensitizers in solar cells; this makes them a good material for absorbing light in photovoltaic applications. Grätzel et al. have shown that MAPbI₃ NPs prepared by spin-coating from the precursor solution on anatase nanosheet TiO₂ film could act not only as a light absorber but also as a hole transporter in MAPbI₃/TiO₂ heterojunction solar cells [10].

Snaith et al. [11] reported the formation of mixed lead halide MAPbI₂Cl perovskite nanocrystals on Al₂O₃ film by spin-coating from a DMF solution of the precursors (MAI and PbCl₂) on an Al₂O₃ film, followed by annealing at 100 °C for 45 minutes in air. The MAPbI₂Cl crystallized in a tetragonal perovskite structure (diffraction peaks at 14.20°, 28.58°, and 43.27°). The MAPbI₂Cl/Al₂O₃ material showed a broad absorption spectrum (400-850 nm), was less moisture-sensitive during the preparation process than the MAPbI₃ nanocrystals, and performed better in solar cell devices than those based on TiO₂ mesoporous films.

The green luminescence of the MAPbBr₃ perovskite NPs was shown by Miyasaka et al. [12] using a thin Al₂O₃ film as the mesoporous media for the preparation of the NPs. The mesoporous Al₂O₃ film (1 µm thick) was prepared with Al₂O₃ nanopowder (40-50 nm) and the MAPbBr₃ material by spin-coating from a precursor solution (MABr plus PbBr₂ in DMF) at different concentrations (1-10 wt%). High resolution transmission electron microscopy (HRTEM) images showed the formation of 5 nm-sized MAPbBr₃ NPs on the spherical Al₂O₃ surface. The material exhibited an intense photoluminescence (PL) peak at 523 nm when they were prepared by using 1 wt% of the precursor solution. The weak PL observed when using higher concentrations was attributed to the formation of the bulk MAPbBr₃ material that did not fit into the Al₂O₃ porous film.

Mixed MAPbBr_3-x Cl_x perovskites on Al₂O₃ mesoporous film (1.5 µm thickness), prepared with 50 nm-sized Al₂O₃ NPs, also showed high PL with a blue-shifted emission from 530 nm to 484 nm when the chloride ratio increased [13]. The X-ray diffraction (XRD) pattern of the film showed the diffraction peaks of cubic MAPbBr₃ perovskite (lattice constant 5.91 Å) that shift to greater angles when the Cl ratio increases. The morphology of the lead halide perovskite on Al₂O₃ film was not described. The PL of MAPbBr_3-x Cl_x was higher than that of the MAPbBr₃ prepared under the same conditions.

The deposition of perovskites by one-step coating permits the precursor solution to infiltrate into mesoporous oxide films. However, the concentration of the precursor solution has to be low enough and the solubility of the precursor salt has to be high enough for a complete penetration of the solution into the pores. Factors such as solution concentration, solvent type, spin-coating speed, and drying temperature can affect the infiltration and, as a consequence, the perovskite morphology and efficiency in its photovoltaic application [14–16].

In the two-step sequential method, PbX₂ is first spin-coated on a mesoporous film followed by dipping into a MAX solution thus giving rise to the MAPbX₃ perovskite [17, 18]. Alternatively, PbI₂ can be added to nanoporous TiO₂ (anatase, 500 nm thick) film, thus forming 22 nm-sized PbI₂ nanoparticles, and the material can then be transformed into the MAPbI₃/TiO₂ film by exposing it to a solution of MAX [18]. The dipping of the TiO₂/PbI₂ nanocomposite film in a solution of MAI in 2-propanol (10mg/mL) for 20 seconds produced MAPbI₃ nanocrystals with a tetragonal perovskite structure. The confinement of PbI₂ in the nanopores of the TiO₂ film facilitated the conversion while forcing the perovskite to adopt a morphology confined to the nanoscale.

Cuboid-like crystals and nanowires of MAPbX₃ were prepared by the two-step coating method. Thus, MAPbI₃ cuboid crystals (an 800 nm-sized polycrystalline material) were obtained by spin-coating on PbI₂/TiO₂ film (nanocrystalline TiO₂ film) from isopropanol solutions of MAI [19]. However, the addition of a small amount of DMF to the MAI solution gave rise to MAPbI₃ nanowires (average diameter and length of 100 nm and 1 µm, respectively). The effect of DMF in the wire formation was attributed to locally dissolved PbI₂ acting as a preferential site to the one-dimension structure growth. This effect was not observed when using other solvents, such as dimethyl sulfoxide and γ-butyrolactone. The MAPbI₃ particulate material showed broad absorption spectra (400-750 nm) and a PL peak near 775 nm and 765 nm for MAPbI₃ cuboids and nanowires, respectively. The MAPbI₃ nanowires performed well in solar cells.

2.2 The non-template method

Pérez-Prieto et al. [20] reported the synthesis of 6-nm sized MAPbBr₃ colloidal perovskite NPs by using octylammonium bromide (OABr), and the non-coordinating solvents, oleic acid (OLA) and 1-octadecene (ODE), to confine the three-dimensional inorganic framework (Fig. 3). While the methyl ammonium cations were embedded in the voids of a set of corner-sharing PbX₆ octahedra, the longer alkyl chain cations only fitted the periphery of the octahedra set with their chains dangling outside it. Thus, OA ions act as the capping ligands of the NP, limiting the growth of the array that extends in three dimensions.

 

Fig. 3 a) Image and schematic representation of colloidal MAPbBr₃ perovskite NPs under UV lamp (at 365 nm) synthesized by the method reported by Pérez-Prieto et al. Adapted from Ref [20], Copyright 2014 American Chemical Society. b-c) Comparison between the absorption (b) and emission spectra (c) of MAPbBr₃ NPs (20% quantum yield) synthesized by the non-template method (green dark) and of that with enhanced quantum yield (83%) prepared by the same method by changing the molar ratio between the perovskite precursors (green).

Download Full Size | PDF

The synthesis of the NPs takes place under mild conditions; thus, a solution of OLA in ODE was stirred and heated at 80 °C and then OABr, MABr, and lead bromide dissolved in DMF were sequentially added. The molar ratio between the total ammonium salt, MABr plus OABr, and PbBr₂ was 1:1. Then, acetone was added to induce the precipitation of the NPs, and the unreactive material was separated by centrifugation. The XRD spectrum of the MAPbBr₃ NPs showed they crystallized with a cubic phase (space group Pm3m), whereas HRTEM images showed the formation of spherical NPs with an average size of 6 nm. The energy dispersive spectroscopy (EDS) analysis confirmed the 72/23 Br/Pb stoichiometry. The NPs exhibited an absorption peak at 525 nm and a PL peak at 527 nm, blue shifted compared with that of MAPbBr₃ bulk. The colloidal nanoparticles showed a PL quantum yield (PLQY) of 20% in toluene and were dispersible in a variety of aprotic organic solvents.

More luminescent MAPbBr₃ colloidal perovskite NPs were reported by the same research group by fine-tuning the molar ratio between MABr, PbBr₂, OABr, and ODE [21]. The synthesis of the NPs was performed using a larger molar ratio between the total ammonium (OABr plus MABr) and PbBr₂ salts (a 4:1 molar ratio instead of the 1:1 molar ratio used in the previous report), while the same amount of OLA and ODE was maintained. The increase of ammonium salts enabled the formation of 7 nm-sized MAPbBr₃ with the PL peak at 526 nm and a PLQY of 67%. Further PL enhancement was obtained when the synthesis was performed under the same conditions but in the absence of OLA. Thus, 5.5 nm-sized NPs with the PL peak at 520 nm and 83% PLQY were obtained showing that the solvent had an important effect in the NP perovskite synthesis. The NPs showed a crystalline cubic perovskite structure (space group = Pm3m).

The organic and inorganic composition of the NPs was studied by TGA and ¹H-NMR. The TGA showed the weight loss of organic compounds (OABr plus MABr) before reaching 360 °C and lead bromide before reaching 650 °C. The one-time purified NPs were transformed back into their components by adding deuterated DMSO and the ¹H-NMR analysis corroborated the presence of MABr, OABr, and ODE and permitted their relative ratio quantification. The combination of both TGA and NMR analyses revealed that the NPs were obtained with a high chemical yield and that the NPs were capped with OABr and ODE, which cooperated to confer NPs with colloidal stability.

Recently, the performance of MAPbBr₃ NPs prepared with OABr by using the non-template method was compared with that of those prepared with octadecylammonium bromide (ODABr). The long alkyl chain ammonium salts were heated in ODE at 120 °C and then the DMF solution of MABr and PbBr₂ were added [22]. The XRD analysis of both NPs showed diffraction peaks that fit MAPbBr₃ cubic phase structure together with diffraction peaks of OABr (14°) or ODABr (10.91°, 14.30°, 19.00 ° and 28.7°). The HRTEM images showed that the OA-capped and ODA-capped NPs exhibited an average size of 3.4 nm and 6.9 nm, respectively. A broader particle distribution was reported for the NPs synthesized with ODABr which was attributed to the poor solubility of ODABr in ODE. The NPs synthesized with OABr showed a PL peak at 525 nm (absorption at 513 nm), 20 nm blue-shifted compared with the PL of MAPbBr₃ bulk (at 535 nm), and PLQY of 20% (compared with fluorescein). The NPs synthesized with ODABr showed a PL peak at 529 nm (absorption at 527 nm) and PLQY of 10%. Compared with the PL peak of the MAPbBr₃ bulk material, that of the OA-capped MAPbBr₃ NPs was more blue-shifted than that of the ODA-capped NPs. This is consistent with the larger size of the latter NPs and thus also consistent with less quantum confinement.

With a view to applying APbBr₃ NPs in light-emitting electrochemical cells (LECs), MAPbBr₃ and FAPbBr₃ (FA = formamidinium) NPs, measuring less than 10 nm in diameter, were synthesized following the non-template method reported by Pérez-Prieto et al. with the difference that THF was used for the precipitation and dispersion of the colloidal NPs for compatibility with the electrolyte solution used in their application in LECs [23]. The NPs showed a cubic structure (space group Pm3m) with lattice constants of 5.939(2) and 6.000(2) Å for MAPbBr₃ and FAPbBr₃, respectively. The OA-capped NPs exhibited an absorption peak and PL peaks at 527 nm and 542 nm (MAPbBr₃), and at 532 nm and 545 nm (FAPbBr₃), whereas the PLQY were 15% and 5% for MAPbBr₃ and FAPbBr₃, respectively. The ¹H-NMR analysis (in deuterated DMSO) of the NPs showed the presence of MABr or FABr and OABr.

In addition, MAPbBr₃ nanoplates have been synthesized by adapting the non-template method [24, 25]. Thus, a mixture of ODE and octylamine (OA) or oleylamine (OAm), instead of OABr, was heated at 80 °C and then a DMF solution of MABr and PbBr₂ was added; the NPs were precipitated with acetone and re-dispersed in toluene. In the case of using OA, crystalline MAPbBr₃ square nanoplates (with a cubic phase) with an average length of 70 nm and thickness of 15 nm, together with 5 nm-sized MAPbBr₃ NPs embedded in the nanoplates, were obtained. Smaller nanoplates (5-20 nm) were synthesized using OAm, and the nanoplate size decreased when increasing the OAm concentration.

Interestingly, the 70 nm-sized MAPbBr₃ nanoplates were used as the starting material for the synthesis of MAPbBr_3−x Cl_x and MAPbBr_3−x I_x nanoplates using a reversible halide-exchange method in solution. The halide-exchange strategy consisted in mixing MAPbBr₃ nanoplates dissolved in toluene with a solution of MACl or MAI in isopropyl alcohol at room temperature. After the exchange reaction, the MAPbBr_3−x Cl_x or MAPbBr_3−x I_x products were separated from the reaction mixture by centrifugation, washed with isopropyl alcohol, and dispersed in toluene. The change in the perovskite composition was visible due to the change in the color of the solution (from white to black). The composition of mixed halide nanoplates was studied by XRD and correlated with XPs and EDS analyses. The diffraction peaks shifted with the increase of the iodide ratio from those of MAPbBr₃ (cubic phase, a = 5.90 Å) to those of MAPbI₃ (tetragonal phase, a = 8.80 Å, c = 12.685 Å) for MAPbBr_3−x I_x. Similar observations were found when comparing MAPbBr_3−x Cl_x with MAPbCl₃ (cubic phase, a = 5.675 Å). The HRTEM images of mixed lead halide perovskites showed the formation of nanoplates with a similar size as that of the initial MAPbBr₃ nanoplates. The absorption and PL emission spectra of the colloidal nanoplates and film (10 µm prepared on silicon substrates) showed a band gap tuning from 400 nm to 800 nm for MAPbCl₃ and MAPbI₃, respectively. The rate of the Br-exchange reaction depended on the nature and concentration of the ammonium halide. The substitution of bromide by iodide was faster than that by Cl; this was attributed to differences in the solubility of the product that governs the thermodynamics of the exchange reaction.

Recently, the precipitation method was studied for the synthesis of MAPbX₃ nanocrystals by using the mixture of perovskite precursors MAX and the lead halide together with a long ammonium halide dissolved in a polar solvent (good solvent) like DMF, which was injected into non-polar solvent (bad solvent) under stirring, thus promoting the precipitation of nano- and micro-sized perovskite particles. Thus, MAPbBr₃ and MAPbI₃ nanocrystals with different shapes were synthesized by injecting a precursor solution of MAX (X = Br, I), OAX, and lead halide PbX₂ dissolved in a polar solvent (DMF, γ-butyrolactone, or acetonitrile) into toluene while being stirred, causing thus the precipitation of perovskite nanocrystals, which were separated by centrifugation and washed with toluene [26]. The TEM images showed that long and thin MAPbI₃ nanowires (1500 nm x 34 nm) were obtained by using a fast addition of the precursor solution (in acetonitrile) into toluene, and rods (810 nm x 54 nm) were observed by using a slow addition rate (0.1 mL/min). In the case of MAPbBr₃ nanocrystals, the size and shape depended on the OABr concentration. The TEM images showed that nanowires (500 nm x 47 nm) were obtained when using a large amount of OABr, while plate-like nanocrystals (150 nm x 30 nm) were obtained when decreasing the concentration of MABr and OABr. MAPbI₃ wires and rods exhibited an emission peak (on film) at 756 nm and 760 nm, respectively, and PLQY of 1.7% and 1.4%, whereas MAPbBr₃ wires and plates showed an emission peak at 530 nm and 520 nm, respectively, and PL QY of 13% and 0.43%. The PL peak were blue-shifted compared with that of the bulk MAPbI₃ and MAPbBr_3, with sizes in the 0.3-2 µm and 0.2-0.8 µm range, prepared in absence of OAX, thus the PL peaks were at 540 nm and 762 nm, respectively. The crystallization process was controlled by the change in the solubility and by the concentration of the long chain ammonium, which controlled the size and optical properties of the nanocrystals [26–29].

High luminescent colloidal MAPbX₃ (X = Cl, Br and I) perovskite QDs were synthesized at room temperature by the re-precipitation method using n-octylamine and OLA as co-ligand of perovskite QDs by Dong et al. [30]. In the synthesis of MAPbBr₃, the precursor solution was prepared dissolving MABr, PbBr₂, OA, and OLA in DMF. The precursor solution was added into toluene under stirring and the precipitates were discarded by centrifugation, Fig. 4. The HRTEM images showed the formation of MABr₃ QDs with an average diameter of 3 nm. Amines with different alkyl chains (dodecylamine, hexadecylamine and hexylamine) and acids with a longer alkyl chain (octanoic acid and butyric acid) were used for the synthesis of MAPbBr₃ QDs. However, the authors did not describe their effect on the size and optical properties of the QDs. The MAPbBr₃ QDs synthesized with n-octylamine showed an absorption and emission peak at 505 nm and 515 nm, respectively, and 50-70% PLQY. The XRD analysis showed the formation of a crystalline MAPbBr₃ cubic phase with broader peaks than that observed with micro-sized MAPbBr₃ bulk material prepared without OA. The EDS analysis showed that the Br/Pb ratio was close to 3.5, thus indicating a Br-rich surface. The presence of n-octylamine as capping ligand was confirmed by XPS, were N 1s spectrum of QDs showed a binding energy at 399.0 and 401.6 eV attributed to n-octylamine and methyl amine, respectively.

 

Fig. 4 a) Schematic illustration of the re-precipitation method used for the synthesis of MAPbBr₃ QDs. b) Images of colloidal solutions (under ambient light and UV light (excitation at 365 nm) and PL spectru of mixed lead halide QDs synthesized by re-precipitation method. Adapted from [30], Copyright 2015 American Chemical Society. c) Images of colloidal solutions under UV lamp (excitation at 365 nm) and PL spectra of MAPbBr₃ QDs synthesized at different temperatures by modification of the re-precipitation method. Adapted from [31]. Copyright 2015, The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

Download Full Size | PDF

This re-precipitation method was also used for the synthesis of mixed halide perovskite MAPbX₃ QDs with different Br/I and Br/Cl composition. MAPbBr_3-xCl_x colloidal QDs showed an progressive blue shifted PL from 515 nm to 407 nm with the increase of x; while MAPbBr_3-xI_x red-shifted 515 nm to 734 nm (Fig. 4(b)). The precipitation process was controlled by the change in the solubility in the solvent mixture while the long chain alkyl acids avoided the aggregation providing stability in non-polar solvent.

Rogach et al. [31] reported a variation of the re-precipitation method described above for the synthesis of MAPbBr₃ QDs, consisting in the variation of the temperature of the bad solvent in the 0-60 °C range. The precursor solution was prepared dissolving MABr, PbBr₂, OAm, and OLA with ultrasonic agitation in DMF. Toluene was used as bad solvent and was cooled or heated before the injection of the precursor solution under stirring. The insoluble bulks precipitated after the injection were separated from the solution by centrifugation and the supernatant contained the MAPbBr₃ QDs. QDs of 1.8 nm, 2.8 nm, and 3.6 nm average diameters were obtained by using 0, 30, and 60 °C, thus showing that the QD size can be controlled with the temperature. The XRD analysis of the QDs synthesized at 30 °C was registered after freeze-drying the QD sample to remove the solvent, instead of performing the simple direct evaporation, thus avoiding the formation of bulk by the presence of traces of DMF solvent in the sample. The XRD of QD fitted with the cubic-MAPbBr₃ perovskite structure, and showed broader peaks compared with the XRD of discarded precipitates (bulk perovskite). The PL and the PLQY was tuned between 470 and 520 nm and between 74 and 93%, respectively, by changing the temperature in the 0-60 °C range. The increase in the PLQY was attributed to the surface capping with longer alkyl chain ligand and to the high crystallinity of the nanoparticle by the precipitation temperature. Although the re-precipitation methodology is a fast and easy strategy for the synthesis of a high luminescent MAPbX₃ QDs, the yield of the synthesis is very limited due to the formation of large nanoparticles together with the QDs.

Finally, MAPbI₃ porous nanostructures have recently been by using MAPbI₃(TEG)₂,, which was prepared by mixing TEG (triethylene glycol) and lead iodide at 60 °C and then MAI [32]. The MAPbI₃(TEG)₂ compound crystallized with an orthorhombic unit cell with the PbI₆ octahedra forming one-dimensional chain by face-sharing and two TEG coordinated to one counterion (CH₃NH₃)⁺, and it exhibited absorption in the 400-450 nm range. Interestingly, this compound was transformed into MAPbI₃ by injection of the precursor dissolved in TGE into dichloromethane, in which the TEG is dissolved while the MAPbI₃ was insoluble and crystallized forming a three dimensional structure (tetragonal phase), Fig. 5(a), with a broad visible absorption range (400-800 nm). The HRTEM images showed the formation of MAPbI₃ NPs with shape and size ill-defined when they were precipitated into dichloromethane. Therefore a capping agent, dodecylammonium iodide (DAI), was added in order to improve the colloidal stability of the MAPbI₃ NPs. TEM and SEM images revealed the formation of 40-400 nm-sized cuboid NPs with nanopores of 20-30 nm when using a low DAI concentration (Fig. 5(b)). Smaller MAPbI₃ NPs with a narrower size distribution were obtained with increasing the DAI concentration (Fig. 5(c)). The formation of nanoporous MAPbI₃ nanocrystal was described by the author as a crystal to-crystal transition (orthorhombic to tetragonal) of precursor into MAPbI₃ crystals that caused the spinoidal demixing of TEG, while DAI stabilized the interfaces, thus acting as the capping agent and controlling the growth and shape of the nanocrystals.

 

Fig. 5 a) Image of the crystal structure of (TEG)₂MAPbI₃ (1), TEG = triethylene glycol depicting the [PbI₆] octahedral (yellow faces) of the unit cell (blue lines), and the crystal-to-crystal transition into MAPbI₃ (2) by the loss of TEG after injection in dichloromethane. b-c) SEM micrograph of the porous MAPbI₃ single crystals (scale bar: 400 nm and 2 μm). Adapted from [32]. Copyright 2015, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Download Full Size | PDF

3. Cesium lead halide perovskite nanosystems

In 1958 Møller reported the synthesis, crystallography, and photoconductivity of cesium lead halides (CsPbX₃; X = Cl, Br, or I) perovskites [33]. This first determination of CsPbX₃ structure was limited to the accuracy of the experimental technique available at that time. In 2008 Trots, D. M. et al. [34] determined the structure of CsPbI₃ perovskite within temperature ranges of 298–687 K and 298–714 K by single-crystal diffraction. These perovskites crystallize in orthorhombic, tetragonal, and cubic polymorphs of these perovskite lattice with the cubic phase being the high-temperature state for all compounds.

Mitzi et al. studied the interesting optoelectronic properties of organic–inorganic perovskite materials about twenty years ago [35]. Lately, solution-processed organic–inorganic hybrid perovskites, such as MAPbI₃, have become one of the most notable materials because they exhibit nearly 20 % solar-cell efficiency [6]. The preparation of organic-inorganic APbX₃ nanoparticulate material by using the porous of mesoporous titania and alumina material and its potential in photovoltaics prompted the preparation of colloidal APbX₃ nanoparticles by a non-template strategy, specifically by using a good organic capping agent, such as long alkyl chain ammonium salts [20] (see previous sections).

The high luminescence efficiency of colloidal organic-inorganic perovskite nanoparticles, in particular those with APbX₃ formula, has recently motivated the preparation of all-inorganic CsPbX₃ (X = Cl, Br, I, and mixed Cl/Br and Br/I) nanoparticles by using the corresponding precursors, organic surfactants, and solvents and heating the solution at high temperature as in the case of CdSe quantum dots (QDs). There are still few papers reporting on their preparation, chemical and photochemical stability, and performance of CsPbX₃ nanoparticles, but they have already motivated a great interest on this type of nanomaterials.

Kovalenko et al. prepared cesium lead halides (CsPbX₃, X = Cl, Br, and I) nanoparticles, which integrate high stability and high PL quantum yield [36]. CsPbX₃ NPs with cubic phase are prepared at high temperature; this is a very metastable state for CsPbX₃ since the bulk material converts into cubic polymorph only above 315°C. Bulk CsPbI₃ crystallizes in an orthorhombic phase (a yellow phase) at room temperature and this phase is non-photoluminescent. Theoretical calculations are consistent with the bulk cubic CsPbI₃ phase to have 17 kJ/mol higher internal energy than the orthorhombic polymorph, whereas such difference is of 7 kJ/mol for CsPbBr₃. Solution synthesis of CsPbI₃ at 305 °C yielded cubic-phase 100–200 nm particles with weak, short-lived emission at 714 nm, and this evidenced the importance of size reduction for stabilizing the cubic phase. Cubic, 4–15 nm-sized CsPbI₃ NPs recrystallized into the yellow phase only upon prolonged storage (months), whereas all other compositions of CsPbX₃ NPs appeared to be fully stable in a cubic phase.

3.1 Synthesis of colloidal CsPbX₃ perovskites

The synthesis of colloidal perovskites is carried out by using precursors (PbX₂ and Cs-carboxylate) and surfactants/solvents (usually OAm, OLA, and ODE; trioctylphosphine (TOPO) is additionally added in the case of CsPbCl₃ to solubilize PbCl₂). Cubic CsPbX₃ perovskite NPs are synthesized through the fast, hot-injection methodology, similarly to the preparation of CdSe quantum dots (QDs) but performing the injection at lower temperatures (< 200 °C). For example, CsPbX₃ NPs can be synthesized through hot-injection of a cesium carboxylate (oleate or stearate) to a PbBr₂ solution containing the surfactants/solvent, 1-5 s later followed by fast cooling of the reaction mixture. The temperature during the synthetic process is a critical factor for the nanocrystal growth; it must be high enough to allow the rearrangement and annealing of atoms during the synthesis process, while being low enough to promote the crystal growth. The NPs are usually precipitated at room temperature and centrifuged, though centrifugation at 0 °C or addition of solvents such as tert-butanol or acetone to the crude reaction can be helpful for a complete precipitation of the smaller NPs. The wet pellet is re-dispersed in an organic solvent (toluene, octane, and hexane) for the NP characterization. The surfactants do not only allow the solubilization of the PbX₂ salt but they can also control the NP growth, passivate the NP surface, and the organic-capped NPs can be stored as colloid for several months [36–41]. CsPbX₃ NPs have been prepared with sizes ranging from 3 nm to 14 nm and the reaction temperature plays a key role in the tuning of the NP size; thus, the NP size increases (decreases) with the reaction temperature. In addition, mixed-halide CsPb(Cl/Br)₃ and CsPb(Br/I)₃ NPs can be prepared by combining appropriate ratios of PbX₂ salts. As in the case of the organic-inorganic lead halide perovskites, CsPb(Cl/I)₃ perovskites cannot be obtained due to the large difference in ionic radii.

CsPbX₃ nanowires have been prepared in air-free conditions by injecting cesium oleate to a mixture containing the lead halide, OLA, and OAm in ODE at 150–250 °C. After 5-10 min the reaction mixture is cooled by an ice-water bath and the aggregated nanowires can be separated by centrifugation and washed with hexane. A detailed analysis of the reaction evolution of CsPbBr₃ with the reaction time evidenced sequential morphological changes. The formation of nanocubes with sizes ranging from 3 to 7 nm at the initial stage is followed by the formation of nanowires with diameters around 9 nm and square-shaped nanosheets, while the amount of nanocubes decreases. Then (about 40–60 min later), the nanosheets dissolved and nanocubes with diameters below 12 nm and lengths up to 5 μm are dominant together with 200 nm-sized crystals. Finally, prolonged heating leads to the disappearance of the nanowires to lead mainly to large crystals. The growth of CsPbI₃ nanowires requires elevated temperatures (> 180 °C) and takes place with faster kinetics; consequently, the reaction is less controllable and the size distribution of the nanowires is wider. Furthermore, the preparation of CsPbCl₃ nanowires is less efficient. A control experiment changing the reaction solvent from ODE to OAm showed that OAm slowed the kinetics but increased the efficiency of the nanowire preparation, thus suggesting a surfactant-directed one-dimensional growth.

Interestingly, postsynthetic chemical transformations can be used to change the NP composition [37, 38]. Thus, Kovalenko et al. reported on the fast partial or complete anion-exchange at low temperature in CsPbX₃ (X = Cl, Br, I) perovskite nanoparticles. The photoluminescence of the material can be tuned over the entire visible spectral region (410–700 nm) while maintaining high quantum yields of 20–80% and narrow emission line widths of 10–40 nm (from blue to red) by adjusting the halide ratios in the colloidal solution. Furthermore, fast inter-nanocrystal anion-exchange leads to uniform CsPb(Cl/Br)₃ or CsPb(Br/I)3 compositions simply by mixing CsPbCl₃, CsPbBr₃, and CsPbI₃ NPs in appropriate ratios, Fig. 6 [37, 38].

 

Fig. 6 Schematic representation of the different routes and precursors for the anion exchange reactions on CsPbX₃ (X = Cl, Br, I) NCs . Reprinted from [38], Copyright 2015, American Chemical Society.

Download Full Size | PDF

Moreover, several attempts of preparing crack-free, NP-only films of CsPbBr₃ perovskites by using one-time purified CsPbBr NPs have failed, but it has recently been reported a smart design of the experimental setup to enable their fabrication [40]. Specifically, it consists in the addition of a CsPbBr₃ NP solution into a tube containing an antisolvent (1:1 volume ratio) and a glass substrate, followed by centrifugation of the sample at 8000 rpm for 30 min (Fig. 7(a)). The CsPbBr₃ NPs are physically adsorbed on the substrate due to the centripetal acceleration and the resulting film-coated substrate can be removed from the tube and dried to remove residual supernatant. The supernatants are substantially eliminated from the film but the NPs remain capped with the ligands; i.e., the NPs remain stabilized by the ligand retained during the deposition process. The PL QY of the film (18%) was three times lower than that of the colloidal CsPbBr₃ NPs prepared in the absence of the substrate, but the optical features of the films such as the PL peak (at ca 511 nm) did not change over several weeks under ambient conditions. Remarkably, although the XRD spectrum of CsPbBr₃ NPs prepared as powders presented the characteristic of random orientation, the CsPbBr₃ NPs assembled in the film exhibited preferential orientation of (100) and (200) planes coplanar with the substrate planes, Fig. 7(b). This single-step casting of perovskite NP films also enables the preparation of CsPbI₃ and CsPbCl₃ and mixed-halide CsPb(I/Br)₃ NP films.

 

Fig. 7 (a) Schematic illustration of the centrifugal casting process, which enables simultaneous purification and film fabrication using crude solutions of as-synthesized CsPbX₃ nanocrystals. (c) Comparison between the XRD pattern of the powder and that of the centrifugally cast film of CsPbBr₃ nanocrystals. Adapted from [40], Copyright 2015, American Chemical Society.

Download Full Size | PDF

4. Comparison between lead halide perovskites and CdSe nanoparticles

The effects of quantum confinement as a function of the particle size is one of the most widely studied phenomena. The exciton energy increases as the particle size is reduced below or near the bulk Bohr exciton radius giving rise to highly size-tunable unique optical properties [42].

Traditional CdSe QDs exhibit intense PL only when their size is comparable to the Bohr excitonic diameter (5.6 nm) [43], where the strong quantum confinement of the charge carriers enhances the transition probability. In CdSe QD materials, tunable optical properties are exhibited by particles smaller than 10 nm. However, this requirement of quantum confinement can result in spectral broadening from the size-distribution and a high density of trap states because of the large surface to volume ratio of the nanoparticle. Smaller sized QDs exhibit higher optical gap than the larger ones and as a consequence, the chromaticity and PL QY changes with the QD concentration due to both self-absorption and Förster resonance energy transfer (FRET) processes. Thus, PL red-shifts occur in close-packed films either because higher energy light emitted by a smaller QD is re-absorbed (self-absorption) by a larger QD with a smaller optical gap, and/or the excited smaller QD non-radiatively (FRET) transfers its energy to a larger QD. Such processes dominate at high concentrations due to the close proximity of the QDs. The optical gap of the CdSe QDs is strongly dependent on their size and, as a consequence, the preparation of several batches with PL peak within an error of ± 5 nm is challenging.

CdSe QDs require a strong confinement of charge carriers to exhibit high PL QY and large inhomogeneity in the optical gaps of the QDs within an ensemble arises even after achieving a narrow size distribution. On the other hand, ensemble CdSe QDs with sizes considerable larger than the excitonic diameter show homogeneity in optical gap, but they do not exhibit intense luminescence.

The exciton Bohr diameter estimated for CsPbX₃ perovskites is up to 12 nm (5 nm for CsPbCl₃; 7 nm for CsPbBr₃, and 12 nm for CsPbI₃) [36]. Kovalenko et al. reported this year on the colloidal synthesis of monodisperse, 4–15 nm-sized CsPbX₃ nanoparticles with cubic shape and cubic perovskite crystal structure. They exhibit not only compositional bandgap engineering but also size tunability of their bandgap energies through the entire visible spectral region of 410–700 nm. The CsPbX₃ NPs present narrow emission line widths (12–42 nm), high quantum yields (50–90%), and short radiative lifetimes (1–29 ns).

Recently, Nag et al. [41] have reported on the low size-dependence of the optical properties of highly luminescent CsPbBr₃ NPs. They prepared colloidal 11 nm-sized CsPbBr₃ NPs with cubic morphology and orthorhombic crystal structure (similar to bulk CsPbBr₃). The absorption edge and emission maxima blue shifted slightly from the 11 to 8 nm cubes. The weak quantum confinement of the charge carriers led to batch to batch reproducibility, with the optical gap lying within ± 1 nm for different synthesis and with similar QYs. The emission of these NPs preserved after months even when stored under ambient conditions and the PL from colloidal CsPbBr₃ NPs remained stable under prolonged exposure to UV light.

Therefore, the weak confinement of the charge carrier in CsPbBr₃ NPs is sufficient enough to show high transition probability for luminescence. Nearly identical effective mass of electron (me = 0.15 electron mass) and hole (mh = 0.14 electron mass) leads to an equal extent of confinement for both charge carriers, which in turn is expected to increase the transition probability for PL by increasing the overlap between electron and hole wavefunctions. The fact that PL is not much influenced by the size-distribution of the 11 nm CsPbBr₃ NPs explains other features of these NPs, such as narrow FWHM of ensemble, batch-to-batch reproducibility, and negligible influence of self-absorption and FRET on emission energy. Therefore, a high emission combined with a non-dependence of the optical properties on the NPs size can be advantageous. This weak confinement on the other hand will inhibit significant size-dependent tuning of emission color, but this can be overcome by controlling the composition of CsPbX₃, where X can be a combination of Cl, Br, and I.

The bright emission of organic-inorganic lead halide perovskites and that of the all-inorganic CsPbX₃ perovskites evidences the considerable reduction of midgap trap states, which severely reduces the PL of core chalcogenide nanoparticles. This is remarkably in the case of organic-inorganic lead halide perovskites, since they are prepared in solution at extraordinarily mild temperatures and these conditions are generally considered to cause a high density of structural defects and mid-gap trap states.

An exception in CsPbX₃ perovskites is that of the perovskite nanowires synthesized by Yang et al. [44]. Whereas CsPbBr₃ nanowires exhibited a narrow PL spectrum which corresponded to excitonic emission with a small degree of quantum confinement, the CsPbI₃ PL spectrum presented two peaks, a thin peak at 446 nm and a broad peak at 523 nm. The former has been ascribed to the excitonic emission similar to that of CsPbBr₃, but the broad, low-energy peak of CsPbI₃ has been attributed to the formation of self-trapped excitons. Studies on the PL temperature dependence of the nanowire showed that CsPbI₃ nanowires exhibited only the broad mission at low temperature while after heating at >100 K the excitonic emission peak appeared and growed monotonically and red-shifted with the temperature. Similar studies with the CsPbBr₃ nanowires revealed a small blue-shift of their excitonic emission with increasing temperature. This different behavior has been attributed to the balance between lattice expansion/contraction and electron–phonon coupling; whereas the lattice term can be dominant in the bromide perovskite, electron–phonon coupling dominates band gap behavior in CsPbI₃ nanowires.

An additional disadvantage of CdSe NPs for light-emitting device applications is the decrease in the optical gap with increasing temperature that can change the chromaticity of an LED with operational temperature. Interestingly, Nag et al. [41] showed that 11 nm-sized CsPbBr₃ NPs did not exhibit any change in PL peak position in the 25-100 °C temperature range. This behavior has been attributed to the electronic structure of CsPbBr₃, where the optical gap predominantly arises from the 6s to 6p transition of Pb²⁺, with less influence from the bromide ion (similarly for the case of the organo-lead perovskites) [45]. Therefore, thermal expansion of the lattice, which would decrease the cation–anion interaction, does not decrease the optical gap.

A practical advantage of CsPbX₃ NPs is the facile access to the blue–green spectral region of 410–530 nm via one-pot synthesis [36]. In comparison, common metal chalcogenide colloidal QDs such as those of CdSe need to be extremely small (≤5 nm) to emit in the blue–green, as-synthesized they exhibit rather low PL QYs (≤5%) due to mid-gap trap states. In addition, CdSe QDs are comparatively less chemically and photochemically stable and require coating with an epitaxial layer of a more chemically robust, wider-gap semiconductor, such as CdS.

Single photon emission has been discovered in a very limited number of artificially engineered fluorescent materials (e.g., CdSe QDs) [46]. This capacity has recently been demonstrated in single CsPbBr₃ and CsPbI₃ NPs [47]. At room temperature, the PL intensity of single CsPbBr₃ NPs switched between the “on” and “off” periods intermittently (blinking), which has been attributed to formation of charged excitons and the associated non-radiative Auger recombination. Additionally, the Auger process favored almost the complete single photon emissions from single CsPbBr₃ NPs since it greatly suppressesed the radiative recombination of multiexcitons. There are several advantages compared with CdSe NPs: CsPbBr₃ perovskite NPs exhibit a two orders of magnitude increase of the absorption cross section and the PL lifetime of single NPs is greatly shortened at both room and cryogenic temperatures. Moreover, the dark-exciton emission frequently found in metal chalcogenide such as CdSe NPs is negligible in CsPbBr₃ perovskite NPs.

Nag et al. [41] investigated the behavior of individual 11nm-sized CsPbBr₃ NPs cast on a silica substrate. Figure 8 shows the PL spectrum of a representative single-NC, the average spectra of 90 single-NPs, and the PL spectrum from an ensemble in solution. The width of both the single-NP PL spectrum as well as of the sub-ensemble of 90 individual NPs are only slightly smaller than that of the ensemble. These authors found that the individual CsPbBr₃ NPs underwent unambiguous temporal fluctuations in their PL emission under continuous illumination (Fig. 8). The vast majority (ca. 90 %) of the individual NPs (several hundred studied) displayed substantial blinking suppression, that is, remained mostly emissive (on-time >85 %), and the nature of blinking was not severely affected over a wide range of excitation powers. In addition, they did not observe the switching to a completely non-emissive state but most of the NPs exhibited frequent flickering, i.e., the emission intensity fluctuated between “bright” and “dim” levels with very short (<60 ms) dim-time bursts. This implies that the time durations for complete absence of emission does not exceed a few tens of milliseconds. Such flickering behavior contrasts with that of conventional CdSe-based core–shell QDs, where clear fluctuations in emission intensity occur between a bright (on) and a dark (off) level, and on-/off-time durations as well as the net on/off ratio for individual NPs depend on the density of photo-generated carriers.

 

Fig. 8 a) PL image of single CsPbBr₃ nanocrystals at excitation power of 9 W cm⁻² and 60 ms exposure time; intensity line profiles show extremely high signal to background b) Normalized PL emission spectra of an ensemble of nanocrystals in solution (green dashed dot line) compared to that of a representative single-nanocrystal (red solid line) and that of 90 single nanocrystal (blue dashed line); Arrows depict spectral line-widths. c) Scatter plot of spectral FWHM and spectrally integrated intensity for 90 single NCs; Dashed lines represent average values of integrated intensities (vertical line) and FWHM (horizontal line) for 90 nanocrystals. d) Characteristic temporal fluctuation of PL intensity for three typical single nanocrystals marked (1-3) in (a), along with proportion of nanocrystals (for several hundred single nanocrystals) which exhibit corresponding intensity/blinking behaviors; Dashed lines mark approximately 50 % intensity compared to the “bright” level. Reprinted from [41], Copyright 2015, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Download Full Size | PDF

Recently, the dynamics of carrier trapping and recombination within high CsPbBr₃ NPs with high PLQY (ca. 79%) have been studied by ultrafast transient absorption spectroscopy, showing that the high PL can be attributed to negligible electron or hole trapping pathways in CsPbBr₃ QDs [48]. Thus, ca. 94% of lowest excitonic states decayed with a single-exponential time constant of 4.5 ± 0.2 ns. Interestingly, the excitons in CsPbBr₃ QDs was efficiently dissociated in the presence of electron or hole acceptors (namely, benzoquinone and phenothiazine as the electron and hole acceptor, respectively). These studies showed that these perovskites are promising in efficient light-harvesting and light-emitting devices.

An intriguingly phenomenon is that found in CsPbX₃ NPs [41] in which all the single NPs were almost non-emissive and remained in a dark state. However, upon continuous illumination, radiative recombination was initiated (over seconds) and near saturation brightness was achieved within few hundreds of milliseconds. Furthermore, after initial illumination for 1 min the NP PL disappeared when kept in the dark for 1 min, and emission restarted after a delay upon exposure to light. This evidenced photoinduced activation of the NPs and slow-timescale (tens of seconds) deactivation processes in the absence of radiation. It has been suggested that such temporally delayed initiation of radiative emission for these perovskite NPs is due to saturation of non-radiative traps within individual NPs, similar to that recently observed for localized emission centers within MAPbI₃ microcrystals [49].

A drawback of the organic-inorganic hybrids for their application in light-emitting devices can be their thermal stability [50]. Comparatively, all-inorganic halide perovskites exhibit more stability, e.g., the Cs component of CsPbBr₃ is not lost by sublimation after heating up to 250 °C in air. Hodes et al. have shown that CsPbBr₃ perovskite material worked equally well as the organic-inorganic analogue in generating the high open circuit voltages. An important disadvantage of all-inorganic halide perovskites is the relatively high cost of cesium compared to the organic ammonium salts [51].

CsPbX₃ NPs are largely free from mid-gap trap states, similar to their MAPbX₃ analogues [52]. Both molecular solutions of MAPbX₃ and colloidal solutions of CsPbX₃ NPs share the common feature of facile solution deposition on substrates. Further, CsPbX₃ NPs are readily miscible with other optoelectronic materials (polymers, fullerenes and other nanomaterials) and future surface-capping ligands for further adjustments of the electronic and optical properties, and solubility in various media.

5. Conclusion

Organic-inorganic (hybrid) and all-inorganic lead halide perovskites, in particular APbX₃ where A is an organic cation (methylammonium or formamidinium) or cesium cation and X = Cl, Br, I, respectively, are of great interest in photovoltaic devices and as luminescent materials for light-emitting devices. They exhibit advantges as photoluminescence systems in both as ensemble and as a single nanoparticle compared to the traditional quantum dots, such as those based on CdSe. The lead halide perovskites can be prepared as nanoparticulate material by using the pores of mesoporous films and also as colloidal nanoparticles and all-NP films;, all of them exhibiting enhanced optical properties with respect to the bulk material. There are several methods for their preparation, but there is still a demand for their obtention with a high chemical yield. The preparation of the all-inorganic lead halide perovskites requires high temperatures, whereas the organic-inorganic lead halide perovskites are prepared under mild conditions. In spite of that, both perform well as luminescent systems. Experimental and theoretical studies on this class of materials are ongoing to understand better their unique photophysical behavior.