Quasi-2D perovskites for efficient solar cells and LEDs

Perovskites are a large family of crystalline ceramics with 3D structures based on that of the natural mineral calcium titanium oxide. In general, they have a chemical formula of the form AMX₃, where X is an anion bonded with different-sized cations (A and M). In the case of organic–inorganic hybrid perovskites, the A-site is occupied by a small organic cation, the M-site is a transition metal (e.g., lead, tin, or germanium), and the X-site is a halogen (e.g., iodine, bromine, or chlorine ions). Such organic–inorganic hybrid perovskites are an emerging class of materials that have so far led to great advances in the performance of solution-processed optoelectronic devices.^{1, 2} On one hand, they enable inexpensive processing, and both optical and electronic tunability. The caveat, however, is that many as-formed lead halide perovskite thin films lack chemical and structural stability, and thus undergo rapid degradation in the presence of moisture or heat.³

For perovskite materials to make a comparable impact in light emission, it is necessary to overcome their slow radiative recombination kinetics (a result of the free-carrier nature of electrons and holes in perovskites at room temperature). In other words, the electrons and holes in perovskites are free carriers at room temperature instead of bound excitons, and the radiative emission rates of these materials are therefore slow. Particularly at moderate excitation densities, the desired radiative processes are overwhelmed by the undesired non-radiative recombination processes. In reality, therefore, electroluminescent devices (e.g., LEDs) based on perovskites have had only limited to modest efficiencies that fail to compete against the non-radiative processes.

To improve LED performance, perovskite films are typically engineered to be thin (about 20nm) so that they confine the injected charges and maximize the carrier density.⁴ However, pinholes are unavoidable in such thin films and the average external quantum efficiency (EQE) of perovskite LEDs has, until recently, remained moderate.^{5, 6} In recent work, a new strategy was developed to spatially confine the injected charges within organolead bromide (CH₃NH₃PbBr₃) nanograins, and thus increase the probability of dissociated excitons regaining their bound form. The reduced grain size thereby increased the radiative recombination rate. This resulted in an enhanced photoluminescence quantum yield (PLQY) and an impressive EQE of 8.53% in the visible region.⁷ Although this strategy was very successful for bromide perovskites, organolead iodide (CH₃NH₃PbI₃) has an exciton binding energy that is three times lower, as well as spatial confinement. It has therefore not yet been possible to produce the much-needed increase in PLQY for IR-emitting perovskites with this material.

Layered perovskites are intermediate to those with either a 2D or a 3D structure, and they can be systematically synthesized by introducing a large organic cation—phenylethylammonium (PEA)—at a carefully chosen stoichiometry. PEA has a large ionic radius, which is incompatible with a 3D perovskite structure. Perovskites that are based on this cation therefore tend to crystallize into a layered 2D structure. Consequently, it is possible to continuously tune the dimensionality of metal halide perovskite compounds by mixing stoichiometric quantities of lead iodide (PbI₂), methylammonium iodide (MAI), and PEA iodide to yield compounds with different numbers of layers (n) in the series PEA₂(CH₃NH₃)_n–1Pb_nI_3n+1. We have therefore investigated a novel platform of mixed organic, dimensionally tuned quasi-2D perovskite thin films that can be used to bridge the gap between 2D and 3D materials.⁸ We used a two-step spin coating, solvent engineering fabrication process⁹ to produce our high-quality, ultrasmooth perovskite layers, without pinholes and with different n values.

We find that members of our quasi-2D perovskite family combine the enhanced stability of 2D perovskites with the excellent optoelectronic parameters of 3D perovskites (including long-range photocarrier diffusion). Indeed, by using density function theory (DFT), we have shown that a low energy of formation, exacerbated in the presence of humidity, explains the propensity of perovskites to decompose back to their precursors. In addition, we find (also with DFT) that intercalation of PEA between perovskite layers introduces quantitatively appreciable van der Waals interactions. These drive an increased formation energy and should thus improve material stability. Our quasi-2D perovskite films therefore exhibit improved stability (see Figure 1), while retaining the high performance characteristics of conventional 3D perovskites.

Figure 1. Illustration of unit cells of organolead halide perovskite (C₈H₉NH₃)₂(CH₃NH₃)_n-1Pb_nI_3n+1(where C is carbon, H is hydrogen, N is nitrogen, Pb is lead, and I is iodine) with different average numbers of layers (n). These structures demonstrate the evolution of the perovskite dimensionality from 2D (i.e., n=1) to 3D (n = ∞). Device performance (expressed as power conversion efficiency, PCE) as a function of n is also shown. Increased performance is achieved with higher n values, but with corresponding decreases in stability. PEA: C₈H₉NH₃(phenylethylammonium). MAI: CH₃NH₃I(methylammonium iodide).

In our work, we explored whether increasing the average number of layers in our quasi-2D perovskites (i.e., that make up the average dimensionally tuned perovskites in a solid-state material) could be used as a way to enhance the PLQY of the materials. Specifically, we studied a mixed-material perovskite that was composed of different quantum-sized, tuned grains.¹⁰ Our results—see Figure 2(a)—show that there is a steeper and earlier rise (lower threshold intensity) in PLQY for our quasi-2D perovskites than for 2D and 3D perovskites. We also studied—see Figure 2(b) and (c)—the EQE and radiance of quasi-2D electroluminescent perovskite devices as a function of n. In the radiance results for the n=5 perovskite, we observe a clear turn-on of light emission at 3.8V. For the same device, we achieved a radiance of 80Wsr⁻¹m⁻² at a current of 95mAcm⁻², when it was driven at 7.4V. On average (based on more than 30 devices), we obtained an EQE of 8.4% at 64mAcm⁻² and 6.4V, which we determined on the basis of the Lambertian emission profiles.

Figure 2. (a) Evolution of photoluminescence quantum yield (PLQY), with increasing excitation intensity, (b) external quantum efficiency (EQE) as a function of current density, and (c) device radiance versus voltage for perovskites with different n values.

We also sought to gain a deeper insight into the dynamics of the photocarriers in our quasi-2D perovskites (compared with control samples). For this part of our study we used transient absorption and time-resolved photoluminescence spectroscopy to enable the characterization of carrier transport and recombination processes on ultrafast timescales. In this way, we obtained a view of the transfer and recombination dynamics of photogenerated charges. Our results indicate that the multiphased perovskite materials channel energy across an inhomogeneous energy landscape and thus concentrate carriers on smaller bandgap emitters. In other words, the mixed quantum-confinement (and thus mixed-bandgap) nature of the materials causes them to efficiently funnel photoexcitations to their lowest-bandgap components. This funneling occurs with greater rapidity than competing non-radiative processes and therefore explains the tremendous improvement in the PLQY of our perovskites. In addition, it ensures that their radiative recombination successfully surpasses trapping, and thus non-radiative recombination. We have therefore used our new materials to build near-IR LEDs that exhibit an EQE of 8.8% and a radiance of 80Wsr⁻¹m⁻². This represents the brightest and most-efficient solution-processed near-IR LEDs produced to date. Furthermore, we have used our materials to achieve the first, certified, hysteresis-free solar power conversion in a planar perovskite solar cell. With this cell, we obtain a certified power conversion efficiency of 15.3% and we observe a greatly improved performance longevity.

In summary, we have developed novel mixed organic, dimensionally tuned perovskite thin films that combine the good stability of 2D perovskites and the excellent optoelectronic properties of 3D perovskites. In particular, we find that by tuning the number of layers in our quasi-2D perovskites, we can vastly improve the PLQY of the materials. In addition, we can use these materials in LEDs and solar cells to achieve high efficiencies. In our ongoing work, we are exploring whether replacing the MAI cation with formamidinium (to further reduce the bandgap of the quasi-2D perovskites) may be a way to engineer solar devices with even higher performance and without hysteresis. We are also investigating new strategies to overcome the low efficiency of radiative recombination in electroluminescence devices that are based on CH₃NH₃PbI₃. Moreover, by tailoring the composition of the perovskites, it is also possible to apply our concept to visible emissions and white LEDs.