Critical Intermediate Structure That Directs the Crystalline Texture and Surface Morphology of Organo-Lead Trihalide Perovskite

Abstract

We have identified an often observed yet unresolved intermediate structure in a popular processing with dimethylformamide solutions of lead chloride and methylammonium iodide for perovskite solar cells. With subsecond time-resolved grazing-incidence X-ray scattering and X-ray photoemission spectroscopy, supplemental with ab initio calculation, the resolved intermediate structure (CHNH)PbICl·CHNHI features two-dimensional (2D) perovskite bilayers of zigzagged lead-halide octahedra and sandwiched CHNHI layers. Such intermediate structure reveals a hidden correlation between the intermediate phase and the composition of the processing solution. Most importantly, the 2D perovskite lattice of the intermediate phase is largely crystallographically aligned with the [110] planes of the three-dimensional perovskite cubic phase; consequently, with sublimation of Cl ions from the organo-lead octahedral terminal corners in prolonged annealing, the zigzagged octahedral layers of the intermediate phase can merge with the intercalated methylammonium iodide layers for templated growth of perovskite crystals. Regulated by annealing temperature and the activation energies of the intermediate and perovskite, deduced from analysis of temperature-dependent structural kinetics, the intermediate phase is found to selectively mature first and then melt along the layering direction for epitaxial conversion into perovskite crystals. The unveiled epitaxial conversion under growth kinetics controls might be general for solution-processed and intermediate-templated perovskite formation.

Unveiling the Nanoparticle‐Seeded Catalytic Nucleation Kinetics of Perovskite Solar Cells by Time‐Resolved GIXS

“…In the early stage of the first 70 s of the thermal annealing process (Figure a), the intensity of the intermediate phase L 1 (040) peak centered at q z = 8.8 nm −1 gradually increased and became saturated. As identified in our previous work, the L 1 phase (of a monoclinic crystal structure with MA 2 PbI 2 Cl 2 ·MAI) comprises 2D bilayers of zigzagged octahedra of [PbI 4 Cl 2 ] 4− coordination. A schematic of the crystalline features of the L 1 phase is also provided in Figure a, where iodine ions occupy all the sharing corners and the chlorine ions occupy the terminal corners; the MAI layers are intercalated between the [PbI 4 Cl 2 ] 4− zigzagged octahedra layers (also refer to the Video S1 file in the Supporting Information).…”

“…A schematic of the crystalline features of the L 1 phase is also provided in Figure a, where iodine ions occupy all the sharing corners and the chlorine ions occupy the terminal corners; the MAI layers are intercalated between the [PbI 4 Cl 2 ] 4− zigzagged octahedra layers (also refer to the Video S1 file in the Supporting Information). Such an intermediate phase of L 1 is mainly associated with enhanced crystal growth kinetics (compared to the direct growth route from precursor to perovskite phase), where the formation of 3D perovskite crystals can be facilitated by interconnecting the 2D‐structured octahedra of L 1 intermediate phases . The result can be seen from the progressive conversion of the L 1 phase into the perovskite phase in the later stage of 200–650 s, as shown in Figure b, where a gradual diminishing of the (040) reflection at q z = 8.8 nm −1 (red arrow), accompanying a gradually enhanced growth of the (100) reflection of the high‐temperature cubic phase of perovskite with the scattering intensity at q = 10.1 nm −1 (blue arrow), is clearly observed .…”

“…To further quantitatively elucidate the perovskite formation mechanism assisted by PbS nanocrystals, we analyzed the evolution of the integrated peak intensity of the representative perovskite (100) reflection using Avrami analysis, α( t ) = 1 − exp{−[ k ( t − t 0 )] n }, which is often used to describe phase formation kinetics at a constant temperature. The relation can be reduced to a form of Sharp–Hancock presentation with ln{−ln[1 − α( t )]} = n ln( t − t 0 ) + n ln( k ), where the Avrami exponent n and the rate constant k can be determined from the slope and the intercept, respectively, with a properly chosen value of the induction time t 0 , as illustrated in Figure .…”

“…The relation can be reduced to a form of Sharp–Hancock presentation with ln{−ln[1 − α( t )]} = n ln( t − t 0 ) + n ln( k ), where the Avrami exponent n and the rate constant k can be determined from the slope and the intercept, respectively, with a properly chosen value of the induction time t 0 , as illustrated in Figure . Typically, n = 2 corresponds to accelerated nucleation, while n = 1 is sporadic nucleation, assuming no crystal size growth . We used the (100) p integral intensity of the cubic perovskite phase as the representative α in the Avrami analysis.…”

“…Figure S5 and Table S2, Supporting Information). Furthermore, the corresponding activation energy E a of the perovskite formation through L 1 ‐to‐perovskite transformation can be determined from the Arrhenius relation k ( T ) ∝ exp[− E a /( RT )], where k ( T ) is the temperature‐dependent Avrami rate constant, T the absolute temperature, and R the gas constant . By fitting the slope −( E a / R ) in Figure d, we obtained the activation energies E a for the perovskite formation via the L 1 ‐to‐perovskite transformation route of each film, which are E a = 145 ± 38 (pristine), 112 ± 12 (with 0.75 wt% PbS NCs), and 47 ± 5 (with 1.0 wt% PbS NCs) kJ mol −1 , respectively.…”