WorksPhysics and Astronomy
Size dependent solid-solid crystallization of halide perovskites
Abstract
The efficiency and stability of halide perovskite-based solar cells and light-emitting diodes directly depend on the intricate dynamics of solid-solid crystallization[1-23]. In this study, we employ a multi-scale approach using random phase approximation, density functional theory, machine learning potentials, reduced charge force fields, and both enhanced sampling biased and brute-force unbiased molecular dynamics simulations to understand the solid-solid phase transitions in cesium lead iodide perovskite. Our simulations uncover that the direct phase transition from the non-perovskite to the perovskite involves the formation of stacked-faulted and low-dimensional intermediate structures. Through extensive large-scale all-atom simulations encompassing up to 650,000 atoms, we observe that solid-solid crystallization may require the formation of a sufficiently large critical nucleus to grow into a faceted perovskite crystal. Based on simulations, we determine that utilizing (100)-faceted seeded crystallization could offer a promising path for manufacturing high-performance and stable perovskite solar cells.
Introduction
Halide perovskites[24, 25, 26, 27, 28] are photo-active materials with tunable band gap (1.1 − 2.3 eV) to manufacture record efficiency solar cells (PSCs)[11, 12, 29, 30, 31, 32, 33]. Despite their superior optoelectronic properties for making low-priced efficient photovoltaics and light emitting diodes, their wide-spread industrialization is still suffering because of the poor reproducibility and stability using current manufacturing processes.
Where one of the key problem is the lack of suitable and reproducible synthesis recipes to produce defect free highly stable thin-films of halide perovskites. Therefore, it is of key importance to comprehend the details of their formation process which can assist to design better experimental methodologies. In particular, wide range of experiments have established that the solid-solid phase transitions dominate the crystallization process and control the micro-structure evolution of perovskite thin-films. In this work, we perform molecular dynamics simulations and try to understand the details of crystallization of the widely studied and thermally stable cesium lead iodide (CsPbI).
CsPbI has four experimentally known polymorphs[1, 34, 35, 36, 37, 38], often recognized as δ, γ, β and α-phases visualized in Figure 1. Among them, δ-phase is the most stable form, and a yellow-colored photo-inactive material consists of edge-sharing lead-iodide octahedra surrounded by cesium cations. On the other hand, γ, β and α-phases are black coloured thermally stable photo-active perovskites made of corner-sharing lead-iodide octahedra with different tilted angles, however meta-stable compared to its δ-form. The perovskite phases have suitable band-gap of ∼ 1.7 to ∼ 1.9eV for making tandem solar cells with the current leading silicon technology (efficiency ∼ 26%) and theoretically can reach efficiencies beyond 30%[29]. However, to construct a competitive low priced and stable > 30% efficient and stable perovskite-silicon or all perovskites tandem solar cells[31, 32, 39, 40, 41, 42, 43, 44, 45, 46] involving CsPbI: one of the main quests is to synthesize/crystallize defects-free corner-sharing phases of CsPbI and eliminate/prevent the formation of edge-sharing δ-phase.

Figure 1. Experimentally known polymoprhs of CsPbI3: Figure shows the different phases of CsPbI3 with Pb-I structures as octahedra with red color iodides on corners and Cs atoms as light-green spheres.
Absolutely gigantic and strenuous experimental efforts[14] have been employed to reach the present day synthesis recipes. A wide range of processes and uncountable additives have been tried to improve the crystallization process.
Now, it is common to utilize additives like dimethylammonium[6, 47, 48, 49], phenylethylammonium[50], bromide[51, 52, 53, 54] and chloride[11, 55] to make efficient and optimistically stable black-phase of cesium lead iodide. Nevertheless, existing manufacturing processes continue to struggle with issues of reproducibility and stability, prompting ongoing efforts to enhance them through extensive experiments aimed at achieving progressively higher efficiencies and industrial scale stable PSCs. Therefore, it is of key importance to understand and optimize the synthesis routes to ensure the stability and reliability of perovskite-based devices in practical applications. A comprehensive atomic-level picture of the nucleation and growth processes can help understand the current experimental recipes’ limitations and, in turn, assist to design improved, reproducible and even newer methodologies for making PSCs and LEDs.
Towards this goal, we perform molecular dynamics (MD) simulations of the crystallization of cesium lead iodide. A simulation study of the full-scale experiment is challenging and herculean task. Here, we focus on understanding the solid-solid phase transitions in perovskites which are experimentally known to controls the synthesis process of perovskite thin-films. We explore two key aspects that have not been extensively investigated in the context of phase transitions in halide perovskites.
Direct non-perovskite () phase to perovskite phase transition
In conventional experimental procedures, the thermodynamically stable -phase initially crystallizes and subsequently undergoes transformation into the black-phase through annealing at elevated temperatures. However, direct conversion processes have historically yielded less stable and less efficient perovskite solar cells. Therefore, first we start with the MD simulations of the direct conversion of edge-sharing -phase of CsPbI to the corner-sharing photo-active phase. Previously Grünwald, Limmer and co-workers[56, 57] have conducted notable co-existing simulations, revealing a liquid-like interface mediating the phase transition between edge-corner-sharing polymorphs of CsPbI. Another significant contribution by Pramchu et al.[58] employed variable-cell nudged-elastic band calculations to investigate reverse phase transitions from the corner-sharing to the edge-sharing phase of CsPbI, primarily aiming to gain insights into the degradation process of smaller systems using DFT calculations.
However, to understand the crystallization of perovskite black phases, it is necessary to reveal formation pathway of meta-stable perovskite phases from a direct heating the -phase as per experimental procedure. Here, we perform MD simulations of the direct phase transition, initiated solely from the -CsPbI. First-order phase transitions between crystalline phases are often associated with substantial energy barriers[59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75], potentially occurring on time scales ranging from seconds to minutes[76, 77, 78, 79, 80].
To address this, we employ biased enhanced sampling methods[61, 81, 82, 83, 84, 85, 86], specifically umbrella sampling[62, 81, 87] and OPES[84, 88]. These methods explore the free energy surface by applying a bias potential to the potential energy surface. The bias potential is crafted as a function of collective coordinates, commonly referred to as collective variables (CVs). The choice of suitable CVs is crucial[64, 89, 90, 91, 92] for addressing complex geometries in multi-species systems, such as the process under investigation. In this study, we introduce a locally ordered CV[93, 94] based on the local crystallographic information of perovskite ()[10, 95, 96, 97] for species = (Cs, Pb, and I), centered around positions of Pb atoms.
Where is defined as:
Where are the number of species within a distance cutoff . Taking inspiration from Ornstein–Zernike[98], Kirkwood[99] equations and others[100, 101]: the final local symmetric order centering Pb positions is calculated by multiplying these individual local contributions[94] from different species:
The final expression for the CV becomes the sum for all Pb species:
Using , we conduct biased simulations employing a reduced charge potential of CsPbI[57]. We successfully observe the crystallization of perovskite structure, presented in Figure 2 and supplementary movie SM1. Additional simulation details can be found in the Methods section. Our observations reveal that the solid-solid phase transition is a complex, multi-step process, generating various stack-faulted structures, anti-phase domains, and -edge dislocations[102, 103] during the transition from full edge-sharing to corner-sharing structures of CsPbI.

Figure 2. Solid-solid phase transition by enhanced sampling: Figure shows the simulated transformation of edge-sharing -phase to corner-sharing -phase of CsPbI. Pb-I structures are represented as yellow octahedra with orange color iodides on coners, and Cs atoms are depicted with blue spheres.
To further scrutinize these observations, we employ an alternative CV based on permutation invariant vector (PIV) distances[104, 105]. Initially, PIV distances are calculated for the starting configuration concerning the reference structures of the edge-sharing () and corner-sharing () phases of . Subsequently, we utilize these indices to define the functional form of the CV as follows:
Where is the PIV distance from the corner-sharing phase, is the PIV distance from edge-sharing phase, are the coordinates of the system, are the coordinates of the reference -phase phase and are the coordinates of the reference -phase.
Using this functional form in Equation (5), we perform biased simulations, and again successfully observe a similar solid-solid phase transition pathway in .
Motivating from these exciting observations, we further investigate direct solid-solid phase transitions with the machine learning potentials (MLPs). In this work, Nequip MLPs[106] are built from dataset of forces, energies and stresses acquired by performing DFT driven MD simulations of in the isothermal–isobaric ensemble.
However, the accuracy of DFT calculations may depend on the exact exchange-correlation functional (XC) adopted in the procedure. Therefore, we first conduct an elementary comparison of widely employed XC functionals against reference first-principles many-body electronic structure method of random phase approximations (RPA+HF)[107, 108, 109, 110, 111] which can account for dynamic electronic screening and long-range dispersion interactions. Figure 3A-C show the relative ground state energies of different perovskite polymorphs against the thermodynamically stable -phase of CsPbI. The best results are obtained for the PBE+D3(BJ)[112, 113, 114, 115], r2SCAN+rVV10[116], optB86b[117] functionals that include van der Waals interactions. We note that although acceptable relative energies are obtained with LDA[118, 119], however this might be due to error cancellations where exchange energy is generally underestimated, and the binding energy is overestimated. To choose the best, we perform an additional calculation of ground state relative energies between -phase and manually constructed polymorph of CsPbI based on the framework of Pb-I face-sharing octahedra[23]. This structure is prepared by replacing formamidnium cations with cesium into hexagonal phase of formamidnium lead iodide (FAPbI), illustrated in Figure 3D. Numerous experimental investigations[120, 121, 122] have consistently demonstrated that the introduction of cesium into hexagonal-FAPbI rapidly form the black perovskite phase, and prevent the formation of its hexagonal face-sharing counterpart.

Figure 3. Relative Energies of different polymorphs using RPA+HF and a range of DFT functionals
Consequently, one anticipates that a face-sharing structure doped with cesium would exhibit instability and higher energy levels compared to the -perovskite phase, a phenomenon corroborated by the ground state energies presented in Figure 3D. Taking into account the comprehensive set of results presented in Figure 3, we opt for the utilization of the r2SCAN+rVV10 functional, considering its suitable depiction of relative energies across various structural configurations. It is worth noting that Shuxia Tao and co-workers[123, 124] have also previously utilized SCAN+rVV10[125] to investigate point defects in various perovskites. Subsequently, we perform DFT based ab-initio MD (AIMD) simulations[126, 127] using the selected functional. Nequip MLPs are generated based on the dataset acquired from the AIMD simulations. The Methods section provide details of these simulations. Using MLPs, we again perform a series of biased simulations to further scrutinize the phase transformation dynamics. Remarkably, the observed pathway is similar to the transformation sequence previously portrayed in Figure 1. This noteworthy consistency between the outcomes obtained through distinct simulation methodologies, RCPs and MLPs underscores the robustness and reliability of our findings.
To further understand the structural transformations, next we perform a detailed analysis of simulated trajectories using smooth overlapping of atomic positions (SOAP)[128, 129, 130] and atomic cluster expansion (ACE)[131, 132, 133] descriptors.
SOAP and ACE, known for capturing detailed atomic variations, are complemented by principal component analysis (PCA) and uniform manifold approximation and projection (UMAP)[134] for dimensionality reduction. This combined approach enhances our ability to comprehend intricate molecular relationships and dynamics. The resulting Figure 4 represents our simulated trajectories, highlighting clusters of similar structural conformations and uncovering underlying patterns in the simulated trajectories. Beyond the discerned mixtures of corner-sharing and edge-sharing structures highlighted in Figure 4, our findings suggest the potential existence of distinct tilted perovskite configurations (see clusters that are not circled and highlighted in Figure 4) on the free energy surface, see Figure 5. It is noteworthy to acknowledge the prior simulations conducted by Wiktor, Erhart and co-workers, have employed Glazer notations to rationalize the presence of diverse locally tilted octahedra during tilt-based phased transformations in cesium lead bromide[135]. Whereas, we show that the wide range of tilted-perovskite phases could exist on the free energy surface during solid-solid crystallization and can be quantified using machine learning techniques. Our observations broadens the understanding of the diverse structural landscapes in halide perovskites and can be easily extended to wide range of perovskite materials, in particular in oxide perovskites where tilt-based phase transformation play a crucial role in determining their electronic properties for designing electronic devices.

Figure 4. Classifications of trajectories with SOAP and ACE
Figure 5. Qualitative free energy profile for direct solid-solid phase transition.
Nonetheless, to achieve a realistic material simulation of phase transition dynamics, it is important to extend beyond the collective atomic motions of approximately 1000s of atoms. It is crucial for capturing the detailed macroscopic mechanisms and dynamics of solid-solid crystallization on a larger scale ensuring a direct atomic visualization of experimental processes, in contrast to small-scale simulations. This transition sets the stage for the subsequent section, wherein we try to delve into the investigation of the macroscopic aspects of solid-solid phase transitions.
To comprehend the macroscopic dynamics of solid-solid phase transitions, we initiated biased simulations on relatively large-scale structures comprising up to 35,000 atoms. As illustrated in Figure 2, our observations reveal that the phase transition progresses through the sliding of octahedra. To emulate this on large scales, we adopt an initial configuration of the -phase, characterized by the relatively larger dimensions along the X and Z directions, however relatively smaller dimensions along the Y-direction, approximately measuring 250 × 24 × 250 Å, illustrated in Figure 6A. Once again, using enhanced sampling simulations with a multi-species structure factor as CV, we excitingly witness the emergence of corner-sharing perovskite structures, illustrated sequentially in Figure 6A-B-C-D. Upon analyzing the trajectory data and Figure 6B, we again observe that the morphological evolution occurs through the mixed formation of stack-faulted or edge-distorted configurations. This phenomenon is directly observable in Figure 6B-D, where significant amount of layers featuring mixed edge-sharing and corner-sharing arrangements are crystallized during the phase transition. Supplementary movie SM2 provides a detailed atomic-scale depiction of this phase transformation.

Figure 6. Solid-solid phase transition by enhanced sampling: Figure shows the simulated transformation of edge-sharing -phase (A) to corner-sharing perovskite-phase of CsPbI. Selectively Pb-I structures are represented as blue octahedra to guide the eye.
However, a crucial insight emerges from our large-scale simulations. As demonstrated in Figure 6D, the process gives rise to several defects, particularly observed in the neighbouring regions where mixed-corner-edge-sharing layers have crystallized. Despite the subsequent conversion of these mixed configurations to full corner-sharing perovskite structures during the biased simulations, the formation of defects is evident, and defects prevailed in the final configuration. These are easily recognizable by incomplete or broken octahedra (and empty spaces) in Figure 6D. In opposite to that, where parts of the -phase that undergo direct conversion to the perovskites instead of mixed corner-edge-sharing structures, exhibit fewer defects, and can be readily visualized in Figure 6A, B and Supplementary movie SM2. From this comprehensive analysis, we propose that one of the mechanism of underlying defect formation in halide perovskites may strongly involve the formation of long-range intermediate structures during the crystallization process.
Simulating structural transformations arising from biased simulations involving more than 35,000 atoms surpasses the capabilities of current code implementations and available computational resources. As a result, in the following section, we undertake brute force MD simulations to investigate various aspects of solid-solid crystallization in greater depth.
Large scale seeded crystallization
Earlier experiments[76, 79, 136, 137, 138] and simulations[72, 139, 140] investigating solid-solid phase transitions in materials other than perovskites, have noticed that following the formation of initial smaller nuclei in a foreign crystalline solid, the subsequent solid-solid crystallization process may obey classical nucleation pathways[62, 141, 142, 143, 144] on larger length scales. Seeking a fundamental macroscopic level understanding in perovskite, we perform unbiased brute force seeded simulations[145, 146, 147, 148, 149, 150, 151], a sophisticated approach that could allows us to explore and probe underlying mechanisms governing macroscopic solid-solid phase transitions in halide perovskites.
To start with, we prepare different structural configurations by incorporating spherical structures (seeds) of perovskite phase into extensive yellow-phase structures, details are provided in the Methods section. We select perovskite seeds with varying radii ranging from 3.0 to 7.0 nm which could offers a comprehensive exploration of nucleation and growth processes at different initial conditions. Starting from equilibrated configurations with different initial seeds we carry out brute force MD simulations using RCPs and monitor the time evolution of fraction of perovskite in these systems. To assess the quantity of perovskite within our simulations, two distinct metrics are employed. Firstly, the relative intensity of the corresponding (100) X-ray peak of perovskite-phase is utilized, allowing for tracking and comparison with experimental data. Additionally, a local-bond order parameter [93, 152, 153], is employed to precisely quantify the number of formula units corresponding to different phases within the simulation box. To compute the X-ray peaks for these extensive structures, we employ a simplified Debye form of structure factor, more simulation details are provided in the Methods section.
From simulations, we observe that as the fraction of perovskite decreases, the inserted seed becomes subcritical and completely transform into -phase. Conversely, as the fraction increases, the cluster transitions to a post-critical state. This approach allows us to ascertain whether the inserted cluster is critical, similar to the methodologies employed in earlier seeded simulations involving ice[144] and carbon-graphite-diamond transitions[72]. Figure 7(A) and 7(C) depict the initial configurations for seed sizes of 3.7 nm and 7.0 nm, respectively, serving as the starting points for our simulations. Subsequently, Figure 7(B) and 7(D) show the concluding configurations for the respective seeded configurations. First, the visual analysis of these figures directly reveals that the smaller perovskite seed experiences significant contraction, as illustrated in Figure 7(B), ultimately transforming into the -phase characterized by an increase in edge-sharing structure.
Figure 7. Seeded simulations: (A) Snapshot depicting the starting configuration with a seed size of 3.7 nm, and (B) illustrates its time evolved configuration. (C) Presents a snapshot of the initial structure with a seed size of 7.0 nm, while (D) showcases the evolved structure. To enhance clarity, middle slices from the 3D box’s center are shown here to guide the eye. The representation is focused on Pb-I octahedra for clear distinction between the two phases.
Conversely, the larger seed (7.0 nm) exhibits substantial growth, resulting in the formation of a very larger perovskite crystal structure. Figure 8 and Figure 9 illustrate the perovskite fraction over time for various simulated perovskite seeds.

Figure 8. Seeded simulations: Figure shows the time evolution of amount of perovskite phase calculated using .

Figure 9. Seeded simulations: Figure shows the time evolution of amount of perovskite phase calculated using structure factor.
Interestingly, a noteworthy observation arises at the boundary between the two phases, where the atoms exhibit liquid-like behavior. This underscores a dynamic moving interface between the coexisting phases, suggesting that solid-solid phase transition is facilitated by the propagation of a liquid-like front[57, 76, 144, 154]. The interplay between liquid-like behavior and crystallization at the phase boundary adds a layer of complexity to the understanding of solid-solid crystallization in perovskites.
To complement with visual observations, simulation movies SM3 and SM4 are provided in the SI for the dynamical insights into the temporal evolution of these systems. Importantly, it is crucial to note that these simulations are conducted at significantly lower temperatures of 600, 650 and 700K than the calculated melting temperature 800K of CsPbI, see the Methods section for details. This deliberate choice ensures the liquid-like front does not results from the melting of the crystal.
In the following sections, we delve into the mechanistic details of the solid-solid nucleation mechanism. Figure 10 offers an insightful perspective into the growth profile of a larger seed across different temperatures. The impact of temperature on growth rate is evident, with a pronounced acceleration at higher temperatures beyond 650K, similar to the earlier reported experiments[155]. However, this heightened growth rate is simultaneously accompanied by the emergence of defects within the crystal structure. Subsequently, we scrutinize these configurations and find the formation of different type of point defects, as portrayed in Figure 11. For clarity, we selectively depict connected Pb atoms, highlighting displaced or defected Pb atoms with cross marks on the evolving perovskite lattice.
Figure 10. Seeded simulations with 7 nm seed: Figure illustrates the time evolution of the normalized (100)-peak in the structure factor at different temperatures for a seed radius of 7.0 nm.

Figure 11. Seeded Simulations. Figure displays a configuration of the ultimately crystallized structure. For clarity, only a slice from the center of the seed is presented, featuring connected Pb atoms exclusively to guide the eye.
Furthermore, we conduct room-temperature relaxation of these final structures at first-principle level of calculations. Performing direct ab-initio simulations for systems of this magnitude is utterly unfeasible. We achieve this by employing MLPs through the neuroevolution potential (NEP) at r2SCAN+rVV10 level and using GPUs accelerate MD (GPUMD) code for molecular dynamics simulations of these large scale configurations. Detailed procedures are explained in the Methods section. These structurally relaxed configurations can serve as valuable input for the comparisons with transmission electron microscopy (TEM) experiments and spectroscopic experiments, thereby offering direct visualizations for tracking the phase transition, formation of defects and help experiments to obtain atomic-level details of the heterogeneities during crystallization process.
Moreover, a significant finding in our investigation is related to the dynamic evolution of the perovskite phase. We discern a distinct discrepancy in the evolution rates of different facets of the perovskite phase. Specifically, the (100)-facet displays accelerated growth compared to its counterparts, as evident in Figure 7 and Figure 11. These figures illustrates the rapid expansion of the (100)-facet boundary, originating from an initial spherical seed. However, when considering the boundary of the (110)-facet and the -phase, interestingly, we observe the formation of a coherent phase between them. Figure 11 provides direct evidence of this phenomenon, where the green color interface highlights the (110)-facet of the grown crystal, while the yellow color interface represents the (100)-facets.
To further investigate this growth process of respective (100)-perovskite-facet, we perform solid-solid co-existing simulations of utilizing NEP MLPs and GPUMD code. Detailed methodologies and simulation set-ups are provided in the Methods section. Brute force MD simulations are conducted at temperatures of 600K, 650K, and 700K, revealing a layer-by-layer transformation[23] of the -phase into the perovskite phase at different temperatures. This whole transformation is illustrated in Figure 12 and Supplementary Movie SM5. Further characterization of the layer-by-layer transformation is evaluated by calculating the change in the number of Pb species belonging to the perovskite phase using the previously defined lq4. Figure 13 shows the time evolution of the perovskite phase across eight repeated MD runs started with the variable momenta. The step-increase in the number of perovskite Pb atoms indicates that upon the onset of perovskite crystallization at the co-existing interface, a complete layer of the -phase transitions into the perovskite phase. Additionally, we computed the free energy profile of this process using the equation . The free energy profile of the co-existing simulations at T=700K is depicted in Figure 14 and eventually indicate a layer-by-layer nature of the transformation, as evidenced by the individual free energy minima separated by barriers on the order of .

Figure 12. Solid-solid co-existing simulations of : Figures A to E illustrate the temporal progression of structural changes during co-existing simulations. For clarity, only magenta-colored Pb-I octahedra are displayed alongside yellow cesium atoms. These visualizations were generated using VMD software[156].

Figure 13. Time evolution of local bond order parameter () for solid-solid co-existing simulations of

Figure 14. Free energy profile of the solid-solid co-existing simulations of
Now, we investigate scenarios involving sub-critical nuclei. In such instances, we observe a distinctive behavior in which the (110)-facet of the perovskite seed undergoes conversion to the -facet of the -phase at an accelerated rate compared to other facets, as highlighted in Figure 7 B. Furthermore, we also perform co-existing simulations of the -perovskite facet with the -phase at 700 K are aimed to understand the interface dynamics. Interestingly, seven time repeated simulations lasting up to 2 microseconds did not show any significant increase in the perovskite structure. This stability contrasts with the behavior observed for the perovskite facet, where co-existence with the -phase led to a complete conversion to perovskite structure by formation of a dynamical interface with a liquid-like character as shown in Figure 12C and Figure 12D. In the case of the facet, however, a distinct interfacial behavior emerged. Here, the facet formed a well-defined and stable boundary with the -phase, lacking the mobile characteristics of the boundary atoms observed with the facet. This difference in interfacial dynamics between the two facets suggests a potential influence of crystallographic orientation on the stability and transformation behavior of the perovskite phase.
At the end, we try to analyze and compute the associated free energy barrier with the seeded crystallization process. We achieve this by gathering multiple histogram data derived from X-ray peak intensity across multiple seeded simulated runs, and combining them into a single run.

Figure 15. Seeded simulations: Figure presents the qualitative free energy of crystallization across various seeded systems.
Utilizing the combined data, we calculate the free energy using the equation , where represents the probability distribution of the (100)-perovskite peak, as explained in prior discussions and detailed in the Methods section. Figure 15 reveals a qualitative free energy barrier, highlighting the necessity to overcome an energy barrier on the order of 10 for solid-solid nucleation at 700K. Excitingly, our collective observations pave the way for quantifying the rates of solid-solid nucleation process at different conditions utilizing a modified classical nucleation theory framework[79, 157]:
Here, represents the surface area, and denotes the volume of the perovskite phase. Additionally, signifies the number density of the product nuclei, denotes the chemical potential difference between the bulk phases, represents the surface tension, and signifies the positive misfit strain free energy per unit volume, which depends on the elastic modulus of the solids. Finally, refers to the energy of pre-existing defects per unit volume .
Discussion and Conclusion
In conclusion, our study offers new insights into the atomistic mechanism for the solid-solid phase transitions in perovskites. Numerous X-ray crystallography experiments investigating the synthesis process have already established the active and crucial involvement of edge-sharing crystalline intermediate structures in the overall crystallization of perovskite thin-films. From simulations of direct phase transition from edge-sharing non-perovskite to perovskite structure, we find the formation of intermediate mixed edge-corner sharing structures that commonly identified as stacking faults or dislocation in wide range of perovskites. This knowledge is complemented by earlier findings from transmission electron microscopy (TEM) experiments[158], which have identified analogous intermediate edge-sharing stack-faulted domains in halide perovskites. Most importantly, photoemission electron microscopy based experimental investigations[159] have further elucidated the presence of edge-dislocations and stack-faulted structures is harmful to the stability of halide perovskite under operational conditions. Recent experiments also revealed their possible photoplastic[160, 161, 162] nature of such structures, and signifying their capability to undergo movement under the influence of light during solar cell operations, therefore highlighting them as possible notorious degradation centers for PSCs.
From our results and analysis, one can conclude that the synthesis methodologies that involve the direct heating of the -phase to perovskite can give rise to the formation of these defected intermediate structure and potentially stable on free energy surface. Hence, a careful attention is required in the development of large-scale industrial processes for producing PSCs those involving the direct annealing of the -phase to perovskite. Secondly, our extensive investigation through large-scale seeded simulations, have unveiled crucial finding of the preferred occurrence of nucleation on a macroscopic level, characterized by a continuous transition marked by the formation of well-defined faceted perovskite crystals. Significantly, our simulations reveal the accelerated growth of the (100)-facet, which efficiently cut through the -phase.
We observed that (100)-oriented seeds promoting the formation of the (100) facet could strategically optimize the perovskite crystallization process. Earlier experiments[3, 7, 8] on demonstrated that the controlled solid-solid phase transitions have produced highly efficient and stable PSCs. Building on our current study and observations of layer-by-layer transformation via a highly mobile interface, we propose two industrial scale synthesis methodologies to crystallize defect-free, and highly crystalline thin-films for PSCs and perovskite LEDs. First, strategies of surface vapor treatment-based can be devised to produce highly crystalline thin-film PSCs by a top to bottom layer-by-layer transformation as demonstrated by earlier experiments on FAPI. Alternatively, molecular additives and chlorine can be employed, for instance, to initially create (100)-oriented domains of quasi-2D perovskite or CsPbICl serving as seeds for the uniform conversion of the -phase to the complete perovskite structure. Therefore, simulations can ultimately contribute to the production of stable perovskite electronics.
Moreover, we provide valuable insights into how a potentially stable boundary may form between the (110)-facet of the perovskite crystal and its corresponding -phase. Specifically, it may occur when the perovskite crystal reaches a critical size. Drawing inspiration from previous work by Haibo Zeng and co-workers[155], who engineered co-existing -CsPbI thin-films to obtain all-inorganic white LEDs[163], our study provides a platform to understand the impact of different process conditions for example effect of temperatures on free energy barriers for stabilizing the -interface.
Therefore, we open avenues for enhancing the efficiency and stability of PSCs and perovskite LEDs building from all-atom MD simulations.
Our study has been limited. Specifically, we have not fully explored the formation and significance of more complex structures, such as CsPbI[14], which could play a crucial role in the synthesis of perovskites. Therefore, looking ahead, our research road-map involves exploring the influence of non-stoichiometric complex structures, seed shapes and commonly used additives on phase transitions through precise finite temperature molecular dynamics studies of nucleation, utilizing large scale machine learning potentials derived from accurate quantum chemical calculations[164, 165, 166, 167, 168, 169, 170]. Accurate calculations of free energy barriers for the multi-step crystallization process will serve as a crucial parameter for calculating the rates of phase transitions to determine the actual kinetic stability of these meta-stable perovskites: one of the key unsolved problem for halide perovskites.
We have taken elementary steps in understanding the complex crystallization of multi-species materials and towards the design of atom by atom industrial process design, however importantly, the first-principles-inspired all-atom computational framework presented in this study offers promising opportunities for investigating complex structural transformations across a diverse range of multi-species materials. Examples include ferroelectric[171, 172, 173, 174, 175, 176, 177, 178], multiferroics[179], and superconducting[180, 181, 182, 183, 184, 185] materials with their potential to reduce global electricity consumption[186, 187, 188, 189].
Methods
MD simulations with RCPs
We utilized a reduced point charge[190, 191] inter-atomic potential for CsPbI following the descriptions provided in the references[57]. The production runs are carried out in the isothermal-isobaric ensembles, using the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) code (version 31 Mar 2017)[192]. Non-bonded interactions were restricted by a 1.0 nm cutoff, and electrostatic interactions were managed using the particle-particle-particle-mesh Ewald method. Temperature control was achieved using velocity rescaling thermostat[193] with a relaxation time of 0.1 ps, while pressure control utilized a Parrinello-Rahman barostat[194, 195, 196] with a relaxation time of 10 ps. OPES and umbrella sampling biased simulations were conducted using PLUMED 2.8[197, 198], interfaced with LAMMPS. To scrutinize finite-size effects concerning yellow to black phase transitions, we employed supercells of varying sizes, ranging from 32, 128, 512 to 1024 formula units (f.u.) of -CsPbI. In the context of seeded simulations, we started by generating three distinctive configurations featuring spherical perovskite seed sizes with radii of 3.7 nm, 5.0 nm, and 7.0 nm. These seeds were embedded within -phase structures, characterized by approximate dimensions of 16x16x16 nm (88540 atoms), 23x23x23 nm (272310 atoms), and 33x31x33 nm (675180 atoms), respectively. It is crucial to emphasize that the selection of the simulation box size, representing -phases, was purposefully set to be more than at least four times the radius of the perovskite seeds. This consideration ensures that the perovskite seeds do not interact with each other within the confines of periodic boundary conditions in MD simulations. Additionally, meticulous attention was given to maintaining suitable distances between pairs of atoms, aligning with the Pb–Pb, Pb-Cs, and Pb-I distances between the two phases. Subsequent to the observation of perovskite seed growth and dissolution in these three initial systems, we expanded our configurations. Additional simulations were conducted, introducing perovskite seeds with radii of 4.0 nm, 4.5 nm, 4.8 nm, 5.0 nm, 5.3 nm, 5.5 nm, and 6.0 nm into a simulation box with dimensions approximately 33x31x33 nm, encompassing around total 678770 atoms. These subsequent simulations were intentionally designed first to repeat earlier simulations, and to offer qualitative insights into the associated free energy barrier for the process. The structure factors were computed employing a methodology similar to previous studies[10], and are commonly depicted, as shown in Figure 16, for different polymorphs. To facilitate analysis of the crystallization process, we have normalized the intensities relative to the initial configurational intensities, thus initiating values from 1.0 in Figure 9 and Figure 10.

Figure 16. Debye structure factors for perovskite() and non-perovskite() phases of CsPbI.
Melting point
To determine the melting point of CsPbI using the reduced charge inter-atomic potential, we carried out co-existing simulations[87, 199] involving crystalline seeds composed of 6400 atoms in contact with a melt containing equal number of atoms. Subsequently, isothermal-isobaric simulations were executed over a temperature range spanning from 720 to 850 K. The progression of potential energy at these temperatures is illustrated in Figure 17, indicating that perovskite melts above 800 K and grows below 800 K. It is noteworthy that the experimental melting point of CsPbI is around 750 K[200].
Figure 17. The figure shows the time evolution of potential energy in co-existing simulations at different temperatures.
local q4 analysis
To quantify the number of formula units within the perovskite phase, whether in direct coexistence simulations or clusters in seeding simulations, we employ the rotationally invariant local-bond order parameter, denoted as [93, 152, 153, 197, 198]. This parameter is determined by the relative positions of a tagged particle and its neighbors within a defined distance. In our investigation, we restrict the calculation of to the Pb species within the systems, considering a cutoff distance of 8 Å. Figure 18 illustrates the histogram of values of a finite-temperature simulated trajectories for perovskite and non-perovskite phases. Notably, a threshold can be established at , enabling the calculation of the number of perovskite formula units in a given system. To further validate this approach, we conduct a simple test

Figure 18. Histogram illustrating the distribution of for a combined simulated trajectory of and phases.
on a trajectory created by combining an equal number of frames from simulated perovskite and non-perovskite phases. As depicted in Figure 19, effectively distinguishes and counts only the number of perovskite units. Similarly, the quantity of -phase can be assessed by considering .

Figure 19. Time evolution of local exceeding a threshold of 0.5 in a combined simulated trajectory featuring both and phases.
RPA and DFT calculations
In this study, all DFT and RPA calculations were executed using the VASP code[201, 202]. The projected Aagmented wave functions (PAW) with PBE potentials[203] were employed for DFT calculations, and LDA potentials were used for LDA functional. RPA+HF results were also obtained using PBE PAW potentials, given their comparability with those derived from PAW GW potentials[111]. We systematically compared various exchange-correlation (XC) functionals and dispersion methods, including LDA[118, 119], PBE[204], PBEsol[205], r2SCAN[206], PBE+D3(BJ)[112, 113, 114, 115], r2SCAN+rVV10[116], optB86b[117] and RPA+HF[107, 108, 109, 110, 111]. Each polymorph was represented using 20-atom supercells, and a cutoff energy of 600 eV was established.
Different k-point grids of , , , , and for DFT calculations and for RPA+HF calculations k-point grids of , , , , were used for , , , , and manually constructed face-sharing phase, respectively. AIMD[126] simulations using k-points were conducted to generate the training dataset, utilizing the r2SCAN+rVV10 functional. These simulations involved 40-atom supercells for each polymorph and were carried out within the isothermal-isobaric (NPT) ensemble, employing a 2fs time step. Pressure and temperatures were controlled using the Parrinello-Rahman barostat and Langevin thermostat[207], respectively.
Machine Learning Potentials
Nequip MLPs
The Nequip[106] MLPs are derived from AIMD simulations data based on r2SCAN+rVV10 DFT level of theory, as described in above sections. For the training process, we employed three interaction layers, implementing a radial cutoff of 8 Å and embedding interatomic distances in a basis comprising eight Bessel functions. Feature representations were selected up to rotation orders of . The loss function incorporated forces, energies, and stresses obtained from DFT calculations, and optimization was executed using Adam with a learning rate of 0.003 and a batch size of 10. The training procedure was conducted on a V100 GPUs.
NEP-GPUMD simulations
For large-scale simulations using MLPs, we utilized the neuroevolution potentials (NEPs)[208, 209] model in GPUMD code[210, 211], as extensively discussed recent publications[212, 213]. NEPs feature a neural network architecture where local atomic environments are delineated by radial and angular components. The radial component of the atomic environment descriptor is designed through linear combinations of Chebyshev basis functions, while the three-body angular part is similarly constructed from Legendre polynomials. In this study, we tried two different setups, first with a radial and angular cutoffs at 6 Å and 5 Å, respectively, and second with using radial and angular cutoffs cutoffs at 8 Å and 4 Å, respectively, both with radial and angular expansions at orders 12 and 8. The neural network architecture consisted of hidden layer comprising 50 neurons, and the training process involved 100,000 generations. The ultimate model for CsPbI was trained based on forces, energies, and stresses derived from rSCAN+rVV10 DFT calculations. In order to perform co-existing solid-solid simulations, we prepare configurations comprising different polymorphs with a minimal lattice mismatch between them. Specifically, for the case of (100)||(100), a supercell consisting of 2520 atoms (with dimensions of 88.046 x 37.734 x 37.734 Å) representing the -phase is combined with a 2560-atom supercell (with dimensions of 83.472 x 38.320 x 35.522 Å) representing the -phase. The final configuration is attained by averaging the lattice parameters, resulting in dimensions of 175.0 x 37.847 x 36.448 Å. For the case of (100)||(110), we first create the (110)-facet of the -phase utilizing ase utilities. Subsequently, it is combined with a supercell of the -phase, ensuring minimal mismatch between the two structures. All structures are provided in the supplementary materials.
Acknowledgements
This research is funded by Swiss National Science Foundation through Post.Doc mobility fellowship P500PN_206693. I am sincerely thankful to Professors Daan Frenkel, Michael Graetzel, Nitin Padture, Ali Alavi, Volker Heine, and Carlos Vega for helpful discussions, encouragements, and useful guidance. I thank Professors Erin Johnson, Michael Sprik, Gabor Csyani, and Alessandro Laio for helpful insightful discussions. I am greatly thankful to Dr. Felix Musil for discussion and help regarding SOAP kernels. I thank Dr. Michele Invernizzi for discussions on enhanced sampling simulations. I am greatly thankful to Dr. Jiahuan Zhang, Dr. Zaiwei Wang, Dr. Chen Li and Dr. Pepe Marquez for highly insightful discussions on experimental data of phase transitions between - phase transitions. I thank Ben Shi and Dr. Michele Simoncelli for carefully reading the manuscript and providing helpful comments. I am indebted and greatly thankful to Yusuf Hamied Department of Chemistry, Professor Angelos Michaelides for hosting and providing necessary computational resources through Rogue-GPU cluster, Nest-CPU cluster, UK Materials and Molecular Modelling Hub YOUNG CPU/GPU cluster, which is partially funded by EPSRC (EP/T022213/1, EP/W032260/1 and EP/P020194/1) and ARCHER2 UK National Supercomputing Service.
Data availability
All parts of this study are easily reproducible: all kinds of input/output files, starting configurations, notified output configurations, plumed input files, lammps input files, RPA/DFT input files and MD trajectories are available on open access server Zenodo: 10.5281/zenodo.10938457, Github: https://github.com/ahlawat-paramvir/plumed2, https://github.com/ahlawat-paramvir/CsPbI3_phase_transitions and also from author.
References
- [1]G. E. Eperon, G. M. Paternò, R. J. Sutton, A. Zampetti, A. A. Haghighirad, F. Cacialli, and H. J. Snaith, Inorganic caesium lead iodide perovskite solar cells, Journal of Materials Chemistry A 3, 19688 (2015).
- [3]T. Liu, Y. Zong, Y. Zhou, M. Yang, Z. Li, O. S. Game, K. Zhu, R. Zhu, Q. Gong, and N. P. Padture, High-Performance Formamidinium-Based Perovskite Solar Cells via Microstructure-Mediated δ-to-α Phase Transformation, Chemistry of Materials 29, 3246 (2017).DOI
- [6]Y. Wang, M. I. Dar, L. K. Ono, T. Zhang, M. Kan, Y. Li, L. Zhang, X. Wang, Y. Yang, X. Gao, Y. Qi, M. Grätzel, and Y. Zhao, Thermodynamically stabilized β-CsPbI3−δ-based perovskite solar cells with efficiencies >18%, Science 365, 591 (2019).
- [7]H. Lu, Y. Liu, P. Ahlawat, A. Mishra, W. R. Tress, F. T. Eickemeyer, Y. Yang, F. Fu, Z. Wang, C. E. Avalos, B. I. Carlsen, A. Agarwalla, X. Zhang, X. Li, Y. Zhan, S. M. Zakeeruddin, L. Emsley, U. Rothlisberger, L. Zheng, A. Hagfeldt, and M. Grätzel, Vapor-assisted deposition of highly efficient, stable black-phase FAPbI3 perovskite solar cells, Science 370, eabb8985 (2020).DOI
- [8]J.-W. Lee, S. Tan, T.-H. Han, R. Wang, L. Zhang, C. Park, M. Yoon, C. Choi, M. Xu, M. E. Liao, S.-J. Lee, S. Nuryyeva, C. Zhu, K. Huynh, M. S. Gooarsky, Y. Huang, X. Pan, and Y. Yang, Solid-phase heteroepitaxial growth of α-phase formamidinium perovskite, Nature Communications 11, 5514 (2020).
- [10]P. Ahlawat, A. Hinderhofer, E. A. Alharbi, H. Lu, A. Ummadisingu, H. Niu, M. Invernizzi, S. M. Zakeeruddin, M. I. Dar, F. Schreiber, A. Hagfeldt, M. Grätzel, U. Rothlisberger, and M. Parrinello, A combined molecular dynamics and experimental study of two-step process enabling low-temperature formation of phase-pure α-FAPbI3, Science Advances 7, eabe3326 (2021).
- [11]S. M. Yoon, H. Min, J. B. Kim, G. Kim, K. S. Lee, and S. I. Seok, Surface Engineering of Ambient-Air-Processed Cesium Lead Triiodide Layers for Efficient Solar Cells, Joule 5, 183 (2021).DOI
- [12]J. Park, J. Kim, H.-S. Yun, M. J. Paik, E. Noh, H. J. Mun, M. G. Kim, T. J. Shin, and S. I. Seok, Controlled growth of perovskite layers with volatile alkylammonium chlorides, Nature 616, 724 (2023).DOI
- [14]Z. Wang, J. Zhang, W. Guo, W. Xiang, and A. Hagfeldt, Formation and Stabilization of Inorganic Halide Perovskites for Photovoltaics, Matter 4, 528 (2021).DOI
- [23]P. Ahlawat, Crystallization of FAPbI3: Polytypes and stacking faults, The Journal of Chemical Physics 159, 151102 (2023).DOI
- [24]C. R. Kagan, D. B. Mitzi, and C. D. Dimitrakopoulos, Organic-Inorganic Hybrid Materials as Semiconducting Channels in Thin-Film Field-Effect Transistors, Science 286, 945 (1999).
- [25]A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka, Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells, Journal of the American Chemical Society 131, 6050 (2009).DOI
- [26]J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park, and N.-G. Park, 6.5% efficient perovskite quantum-dot-sensitized solar cell, Nanoscale 3, 4088 (2011).
- [27]M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami, and H. J. Snaith, Efficient Hybrid Solar Cells Based on Meso-Superstructured Organometal Halide Perovskites, Science 338, 643 (2012).DOI
- [28]H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Grätzel, and N.-G. Park, Lead Iodide Perovskite Sensitized All-Solid-State Submicron Thin Film Mesoscopic Solar Cell with Efficiency Exceeding 9%, Scientific Reports 2, 591 (2012).DOI
- [29]S. Bremner, C. Yi, I. Almansouri, A. Ho-Baillie, and M. Green, Optimum band gap combinations to make best use of new photovoltaic materials, Solar Energy 135, 750 (2016).DOI
- [30]J. Jeong, M. Kim, J. Seo, H. Lu, P. Anlahrat, A. Mishra, Y. Yang, M. A. Hope, F. T. Eickemeyer, M. Kim, Y. J. Yoon, I. W. Choi, B. P. Darwich, S. J. Choi, Y. Jo, J. H. Lee, B. Walker, S. M. Zakeeruddin, L. Emsley, U. Rothlisberger, A. Hagfeldt, D. S. Kim, M. Grätzel, and J. Y. Kim, Pseudo-halide anion engineering for α-FAPbI3 perovskite solar cells, Nature 592, 381 (2021).DOI
- [31]X. Chu, Q. Ye, Z. Wang, C. Zhang, F. Ma, Z. Qu, Y. Zhao, Z. Yin, H.-X. Deng, X. Zhang, and J. You, Surface in situ reconstruction of inorganic perovskite films enabling long carrier lifetimes and solar cells with 21% efficiency, Nature Energy 8, 372 (2023).
- [32]Interactive Best Research-Cell Efficiency Chart.
- [33]H. Chen, A. Maxwell, C. Li, S. Teale, B. Chen, T. Zhu, E. Ugur, G. Harrison, L. Grater, J. Wang, Z. Wang, L. Zeng, S. M. Park, L. Chen, P. Serles, R. A. Awni, B. Subedi, X. Zheng, C. Xiao, N. J. Podraza, T. Fileter, C. Liu, Y. Yang, J. M. Luther, S. De Wolf, M. G. Kanatzidis, Y. Yan, and E. H. Sargent, Regulating surface potential maximizes voltage in all-perovskite tandems, Nature 613, 676 (2023).
- [34]C. K. Møller, Crystal Structure and Photoconductivity of Cæsium Plumbohalides, Nature 182, 1436 (1958).DOI
- [35]C. C. Stoumpos, C. D. Malliakas, and M. G. Kanatzidis, Semiconducting Tin and Lead Iodide Perovskites with Organic Cations: Phase Transitions, High Mobilities, and Near-Infrared Photoluminescent Properties, Inorganic Chemistry 52, 9019 (2013).DOI
- [36]D. Trots and S. Myagkota, High-temperature structural evolution of caesium and rubidium triiodoplumbates, Journal of Physics and Chemistry of Solids 69, 2520 (2008).DOI
- [37]R. J. Sutton, M. R. Filip, A. A. Haghighirad, N. Sakai, B. Wenger, F. Giustino, and H. J. Snaith, Cubic or Orthorhombic? Revealing the Crystal Structure of Metastable Black-Phase CsPbI 3 by Theory and Experiment, ACS Energy Letters 3, 1787 (2018).
- [38]A. K. Jena, A. Kulkarni, Y. Sanehira, M. Ikegami, and T. Miyasaka, Stabilization of α-CsPbI 3 in Ambi- ent Room Temperature Conditions by Incorporating Eu into CsPbI3, Chemistry of Materials 30, 6668 (2018).
- [39]G. E. Eperon, T. Leijtens, K. A. Bush, R. Prasanna, T. Green, J. T.-W. Wang, D. P. McMeekin, G. Volonakis, R. L. Milot, R. May, A. Palmstrom, D. J. Slotcavage, R. A. Belisle, J. B. Patel, E. S. Parrott, R. J. Sutton, W. Ma, F. Moghadam, B. Conings, A. Babayigit, H.-G. Boyen, S. Bent, F. Giustino, L. M. Herz, M. B. Johnston, M. D. McGehee, and H. J. Snaith, Perovskite-perovskite tandem photovoltaics with optimized band gaps, Science 354, 861 (2016).
- [40]X. Y. Chin, D. Turkay, J. A. Steele, S. Tabean, S. Eswara, M. Mensi, P. Fiala, C. M. Wolff, A. Paracchino, K. Artuk, D. Jacobs, Q. Guesnay, F. Sahli, G. Andretta, M. Boccard, Q. Jeangros, and C. Ballif, Interface passivation for 31.25%-efficient perovskite/silicon tandem solar cells, Science 381, 59 (2023).DOI
- [41]M. Jošt, E. Köhnen, A. B. Morales-Vilches, B. Lipovšek, K. Jäger, B. Macco, A. Al-Ashouri, J. Krč, L. Korte, B. Rech, R. Schlatmann, M. Topič, B. Stannowski, and S. Albrecht, Textured interfaces in monolithic perovskite/silicon tandem solar cells: advanced light management for improved efficiency and energy yield, Energy & Environmental Science 11, 3511 (2018).
- [42]S. Mariotti, E. Köhnen, F. Scheler, K. Sveinbjörnsson, L. Zimmermann, M. Piot, F. Yang, B. Li, J. Warby, A. Musienko, D. Menzel, F. Lang, S. Keßler, I. Levine, D. Mantione, A. Al-Ashouri, M. S. Härtel, K. Xu, A. Cruz, J. Kurpiers, P. Wagner, H. Köbel, J. Li, A. Magomedov, D. Mercerreyes, E. Unger, A. Abate, M. Stolterfoht, B. Stannowski, R. Schlatmann, L. Korte, and S. Albrecht, Interface engineering for high-performance, triple-halide perovskite-silicon tandem solar cells, Science 381, 63 (2023).
- [43]J. Liu, M. De Bastiani, E. Aydin, G. T. Harrison, Y. Gao, R. R. Pradhan, M. K. Eswaran, M. Mandal, W. Yan, A. Seitkhan, M. Babics, A. S. Subbiah, E. Ugur, F. Xu, L. Xu, M. Wang, A. U. Rehman, A. Razzaq, J. Kang, R. Azmi, A. A. Said, F. H. Isikgor, T. G. Allen, D. Andrienko, U. Schwingen-schlögl, F. Laquai, and S. De Wolf, Efficient and stable perovskite-silicon tandem solar cells through contact displacement by MgFₓ, Science 377, 302 (2022).
- [44]Z. Wang, L. Zeng, T. Zhu, H. Chen, B. Chen, D. J. Kubicki, A. Balvanz, C. Li, A. Maxwell, E. Ugur, R. dos Reis, M. Cheng, G. Yang, B. Subedi, D. Luo, J. Hu, J. Wang, S. Teale, S. Mahesh, S. Wang, S. Hu, E. D. Jung, M. Wei, S. M. Park, L. Grater, E. Aydin, Z. Song, N. J. Podraza, Z.-H. Lu, J. Huang, V. P. Dravid, S. De Wolf, Y. Yan, M. Grätzel, M. G. Kanatzidis, and E. H. Sargent, Suppressed phase segregation for triple-junction perovskite solar cells, Nature 618, 74 (2023).DOI
- [45]S. G. Ji, I. J. Park, H. Chang, J. H. Park, G. P. Hong, B. K. Choi, J. H. Jang, Y. J. Choi, H. W. Lim, Y. J. Ahn, S. J. Park, K. T. Nam, T. Hyeon, J. Park, D. H. Kim, and J. Y. Kim, Stable pure-iodide wide-band-gap perovskites for efficient Si tandem cells via kinetically controlled phase evolution, Joule 6, 2390 (2022).DOI
- [46]S. S. Mali, J. V. Patil, J.-Y. Shao, Y.-W. Zhong, S. R. Rondiya, N. Y. Dzade, and C. K. Hong, Phase-heterojunction all-inorganic perovskite solar cells surpassing 21.5% efficiency, Nature Energy 8, 989 (2023).
- [47]W. Ke, I. Spanopoulos, C. C. Stoumpos, and M. G. Kanatzidis, Myths and reality of HPbI3 in halide perovskite solar cells, Nature Communications 9, 4785 (2018).DOI
- [48]H. Meng, Z. Shao, L. Wang, Z. Li, R. Liu, Y. Fan, G. Cui, and S. Pang, Chemical Composition and Phase Evolution in DMAI-Derived Inorganic Perovskite Solar Cells, ACS Energy Letters 5, 263 (2020).
- [49]C. F. J. Lau, Z. Wang, N. Sakai, J. Zheng, C. H. Liao, M. Green, S. Huang, H. J. Snaith, and A. Ho-Baillie, Fabrication of Efficient and Stable CsPbI3 Perovskite Solar Cells through Cation Exchange Process, Advanced Energy Materials 9, 1901685 (2019).
- [50]Y. Jiang, J. Yuan, Y. Ni, J. Yang, Y. Wang, T. Jiu, M. Yuan, and J. Chen, Reduced-Dimensional α-CsPbX3 Perovskites for Efficient and Stable Photovoltaics, Joule 2, 1356 (2018).DOI
- [51]R. E. Beal, D. J. Slotcavage, T. Leijtens, A. R. Bowring, R. A. Belisle, W. H. Nguyen, G. F. Burkhard, E. T. Hoke, and M. D. McGehee, Cesium Lead Halide Perovskites with Improved Stability for Tandem Solar Cells, The Journal of Physical Chemistry Letters 7, 746 (2016).
- [52]W. Li, M. U. Rothmann, A. Liu, Z. Wang, Y. Zhang, A. R. Pascoe, J. Lu, L. Jiang, Y. Chen, F. Huang, Y. Peng, Q. Bao, J. Etheridge, U. Bach, and Y. Cheng, Phase Segregation Enhanced Ion Movement in Efficient Inorganic CsPbIBr2 Solar Cells, Advanced Energy Materials 7, 1700946 (2017).
- [53]W. Zhou, Y. Zhao, X. Zhou, R. Fu, Q. Li, Y. Zhao, K. Liu, D. Yu, and Q. Zhao, Light-Independent Ionic Transport in Inorganic Perovskite and Ultrastable Cs-Based Perovskite Solar Cells, The Journal of Physical Chemistry Letters 8, 4122 (2017).
- [54]W. Xiang, Z. Wang, D. J. Kubicki, W. Tress, J. Luo, D. Prochowicz, S. Akin, L. Emsley, J. Zhou, G. Dietler, M. Grätzel, and A. Hagfeldt, Europium-Doped CsPbI2Br for Stable and Highly Efficient Inorganic Perovskite Solar Cells, Joule 3, 205 (2019).
- [55]K. Wang, Z. Jin, L. Liang, H. Bian, H. Wang, J. Feng, Q. Wang, and S. F. Liu, Chlorine doping for black γ-CsPbI3 solar cells with stabilized efficiency beyond 16%, Nano Energy 58, 175 (2019).
- [56]J. Lin, M. Lai, L. Dou, C. S. Kley, H. Chen, F. Peng, J. Sun, D. Lu, S. A. Hawks, C. Xie, F. Cui, A. P. Alivisatos, D. T. Limmer, and P. Yang, Thermochromic halide perovskite solar cells, Nature Materials 17, 261 (2018).
- [57]C. G. Bischak, M. Lai, Z. Fan, D. Lu, P. David, D. Dong, H. Chen, A. S. Etman, T. Lei, J. Sun, M. Grünwald, D. T. Limmer, P. Yang, and N. S. Ginsberg, Liquid-like Interfaces Mediate Structural Phase Transitions in Lead Halide Perovskites, Matter 3, 534 (2020).
- [58]S. Pramchu, A. P. Jaroentijitthichai, and Y. Laosirita-worn, Effects of bromine substitution for iodine on structural stability and phase transition of CsPbI3, Applied Surface Science 496, 143593 (2019).
- [59]Statistical Physics (Elsevier, 1980).
- [60]K. Binder, Theory of first-order phase transitions, Reports on Progress in Physics 50, 783 (1987).
- [61]D. Frenkel and B. Smit, Understanding molecular simulation: from algorithms to applications, 2nd ed., Computational science series No. 1 (Academic Press, San Diego, 2002).
- [62]S. Auer and D. Frenkel, Prediction of absolute crystal-nucleation rate in hard-sphere colloids, Nature 409, 1020 (2001).
- [63]C. X. Du, G. Van Anders, R. S. Newman, and S. C. Glotzer, Crystal shape-driven solid–solid transitions in colloids, Proceedings of the National Academy of Sciences 114, 10.1073/pnas.1621348114 (2017).
- [64]B. Peters, Reaction rate theory and rare events (Elsevier, Amsterdam ; Cambridge, MA, 2017) oCLC: ocn969963116.
- [65]P. Xiao, J.-G. Cheng, J.-S. Zhou, J. B. Goodenough, and G. Henkelman, Mechanism of the CaIrO 3 post-perovskite phase transition under pressure, Physical Review B 88, 144102 (2013).
- [66]G.-R. Qian, X. Dong, X.-F. Zhou, Y. Tian, A. R. Oganov, and H.-T. Wang, Variable cell nudged elastic band method for studying solid–solid structural phase transitions, Computer Physics Communications 184, 2111 (2013).
- [67]D. Frenkel, Entropy-driven phase transitions, Physica A: Statistical Mechanics and its Applications 263, 26 (1999).
- [68]W. Qi, Y. Peng, Y. Han, R. K. Bowles, and M. Dijkstra, Nonclassical Nucleation in a Solid-Solid Transition of Confined Hard Spheres, Physical Review Letters 115, 185701 (2015).
- [69]M. Grünwald, E. Rabani, and C. Dellago, Mechanisms of the Wurtzite to Rocksalt Transformation in CdSe Nanocrystals, Physical Review Letters 96, 255701 (2006).
- [70]M. Grünwald and C. Dellago, Nucleation and Growth in Structural Transformations of Nanocrystals, Nano Letters 9, 2099 (2009).
- [71]D. Zahn, Nucleation mechanism and kinetics of the perovskite to post-perovskite transition of MgSiO3 under extreme conditions, Chemical Physics Letters 573, 5 (2013).
- [72]R. Z. Khaliullin, H. Eshet, T. D. Kühne, J. Behler, and M. Parrinello, Nucleation mechanism for the direct graphite-to-diamond phase transition, Nature Materials 10, 693 (2011).
- [73]J. Anwar and D. Zahn, Polymorphic phase transitions: Macroscopic theory and molecular simulation, Advanced Drug Delivery Reviews 117, 47 (2017).
- [74]J. Rogal, E. Schneider, and M. E. Tuckerman, Neural-Network-Based Collective Variables for Enhanced Sampling of Phase Transformations, Physical Review Letters 123, 245701 (2019).
- [75]K. J. Caspersen and E. A. Carter, Finding transition states for crystalline solid–solid phase transformations, Proceedings of the National Academy of Sciences 102, 6738 (2005).
- [76]Y. Peng, P. Wang, Z. Wang, A. M. Alsayed, Z. Zhang, A. G. Yodh, and Y. Han, Two-step nucleation mechanism in solid–solid phase transitions, Nature Materials 14, 101 (2015).
- [77]E. Sanz and C. Valeriani, Mediated by a liquid, Nature Materials 14, 15 (2015).
- [78]S.-c. Zhu, X.-Z. Yan, J. Liu, A. R. Oganov, and Q. Zhu, A Revisited Mechanism of the Graphite-to-Diamond Transition at High Temperature, Matter 3, 864 (2020).
- [79]M. Li, Z. Yue, Y. Chen, H. Tong, H. Tanaka, and P. Tan, Revealing thermally-activated nucleation pathways of diffusionless solid-to-solid transition, Nature Communications 12, 4042 (2021).
- [80]D. Frenkel, R. J. Vos, C. G. De Kruif, and A. Vrij, Structure factors of polydisperse systems of hard spheres: A comparison of Monte Carlo simulations and Percus–Yevick theory, The Journal of Chemical Physics 84, 4625 (1986).
- [81]G. Torrie and J. Valleau, Nonphysical sampling distributions in Monte Carlo free-energy estimation: Umbrella sampling, Journal of Computational Physics 23, 187 (1977).
- [82]P. Virnau and M. Müller, Calculation of free energy through successive umbrella sampling, The Journal of Chemical Physics 120, 10925 (2004).
- [83]T. Huber, A. E. Torda, and W. F. Van Gunsteren, Local elevation: A method for improving the searching properties of molecular dynamics simulation, Journal of Computer-Aided Molecular Design 8, 695 (1994).
- [84]M. Invernizzi and M. Parrinello, Rethinking Metadynamics: From Bias Potentials to Probability Distributions, The Journal of Physical Chemistry Letters 11, 2731 (2020).
- [85]J. Debnath and M. Parrinello, Gaussian Mixture-Based Enhanced Sampling for Statics and Dynamics, The Journal of Physical Chemistry Letters 11, 5076 (2020).
- [86]P. Maragakis, A. Van Der Vaart, and M. Karplus, Gaussian-Mixture Umbrella Sampling, The Journal of Physical Chemistry B 113, 4664 (2009).
- [87]P.-R. Ten Wolde, M. J. Ruiz-Montero, and D. Frenkel, Simulation of homogeneous crystal nucleation close to coexistence, Faraday Discussions 104, 93 (1996).
- [88]M. Invernizzi, P. M. Piaggi, and M. Parrinello, Unified Approach to Enhanced Sampling, Physical Review X 10, 041034 (2020).
- [89]M. A. Rohrdanz, W. Zheng, M. Maggioni, and C. Clementi, Determination of reaction coordinates via locally scaled diffusion map, The Journal of Chemical Physics 134, 124116 (2011).
- [90]J. McCarty and M. Parrinello, A variational conformational dynamics approach to the selection of collective variables in metadynamics, The Journal of Chemical Physics 147, 204109 (2017).
- [91]M. M. Sultan and V. S. Pande, Automated design of collective variables using supervised machine learning, The Journal of Chemical Physics 149, 094106 (2018).
- [92]V. Rizzi, D. Mendels, E. Sicilia, and M. Parrinello, Blind Search for Complex Chemical Pathways Using Harmonic Linear Discriminant Analysis, Journal of Chemical Theory and Computation 15, 4507 (2019).
- [93]W. Lechner and C. Dellago, Accurate determination of crystal structures based on averaged local bond order parameters, The Journal of Chemical Physics 129, 114707 (2008).
- [94]P. Ahlawat, M. I. Dar, P. Piaggi, M. Grätzel, M. Parrinello, and U. Rothlisberger, Atomistic Mechanism of the Nucleation of Methylammonium Lead Iodide Perovskite from Solution, Chemistry of Materials 32, 529 (2020).
- [95]H. Niu, P. M. Piaggi, M. Invernizzi, and M. Parrinello, Molecular dynamics simulations of liquid silica crystallization, Proceedings of the National Academy of Sciences 115, 5348 (2018).
- [96]L. Bonati and M. Parrinello, Silicon Liquid Structure and Crystal Nucleation from Ab Initio Deep Metadynamics, Physical Review Letters 121, 265701 (2018).
- [97]P. Ahlawat, Molecular dynamics simulations of nucleation of hexagonal($\delta$) and cubic($\alpha$)-FAPbI$_3$ perovskites from solution (2023), arXiv:2306.14172 [cond-mat, physics:physics].arxiv.org/abs/2306.14172
- [98]L. S. Ornstein, Accidental deviations of density and opalescence at the critical point of a single substance, Proc. Akad. Sci. 17, 793 (1914).
- [99]J. G. Kirkwood, Statistical Mechanics of Fluid Mixtures, The Journal of Chemical Physics 3, 300 (1935).
- [100]J. Van Leeuwen, J. Groeneveld, and J. De Boer, New method for the calculation of the pair correlation function. I, Physica 25, 792 (1959).
- [101]E. E. Salpeter, On Mayer’s theory of cluster expansions, Annals of Physics 5, 183 (1958).
- [102]P. Hirel, P. Carrez, and P. Cordier, From glissile to sessile: Effect of temperature on <110> dislocations in perovskite materials, Scripta Materialia 120, 67 (2016).
- [103]K. Song, J. Liu, N. Lu, D. Qi, and W. Qin, Direct atomic-scale imaging of a screw dislocation core structure in inorganic halide perovskites, Physical Chemistry Chemical Physics 24, 6393 (2022).
- [104]G. A. Gallet and F. Pietrucci, Structural cluster analysis of chemical reactions in solution, The Journal of Chemical Physics 139, 074101 (2013).
- [105]S. Pipolo, M. Salanne, G. Ferlat, S. Klotz, A. Saitta, and F. Pietrucci, Navigating at Will on the Water Phase Diagram, Physical Review Letters 119, 245701 (2017).
- [106]S. Batzner, A. Musaelian, L. Sun, M. Geiger, J. P. Mailoa, M. Kornbluth, N. Molinari, T. E. Smidt, and B. Kozinsky, E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials, Nature Communications 13, 2453 (2022).
- [107]D. Bohm and D. Pines, A Collective Description of Electron Interactions. I. Magnetic Interactions, Physical Review 82, 625 (1951).
- [108]M. Gell-Mann and K. A. Brueckner, Correlation Energy of an Electron Gas at High Density, Physical Review 106, 364 (1957).
- [109]J. Harl, L. Schimka, and G. Kresse, Assessing the quality of the random phase approximation for lattice constants and atomization energies of solids, Physical Review B 81, 115126 (2010).
- [110]M. Kaltak, J. Klimeš, and G. Kresse, Cubic scaling algorithm for the random phase approximation: Self-interstitials and vacancies in Si, Physical Review B 90, 054115 (2014).
- [111]T. Braeckevelt, R. Goeminne, S. Vandenhaute, S. Borgmans, T. Verstraelen, J. A. Steele, M. B. J. Roefaers, J. Hofkens, S. M. J. Rogge, and V. Van Speybroeck, Accurately Determining the Phase Transition Temperature of CsPbI 3 via Random-Phase Approximation Calculations and Phase-Transferable Machine Learning Potentials, Chemistry of Materials 34, 8561 (2022).
- [112]A. D. Becke and E. R. Johnson, A density-functional model of the dispersion interaction, The Journal of Chemical Physics 123, 154101 (2005).
- [113]E. R. Johnson and A. D. Becke, A post-Hartree-Fock model of intermolecular interactions: Inclusion of higher-order corrections, The Journal of Chemical Physics 124, 174104 (2006).
- [114]S. Grimme, J. Antony, S. Ehrlich, and H. Krieg, A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements, H-Pu, The Journal of Chemical Physics 132, 154104 (2010).
- [115]S. Grimme, S. Ehrlich, and L. Goerigk, Effect of the damping function in dispersion corrected density functional theory, Journal of Computational Chemistry 32, 1456 (2011).
- [116]J. Ning, M. Kothakonda, J. W. Furness, A. D. Kaplan, S. Ehlert, J. G. Brandenburg, J. P. Perdew, and J. Sun, Workhorse minimally empirical dispersion-corrected density functional with tests for weakly bound systems: r 2 SCAN + rVV10, Physical Review B 106, 075422 (2022).
- [117]J. Klimeš, D. R. Bowler, and A. Michaelides, Van der Waals density functionals applied to solids, Physical Review B 83, 195131 (2011).
- [118]P. Hohenberg and W. Kohn, Inhomogeneous Electron Gas, Physical Review 136, B864 (1964).
- [119]W. Kohn and L. J. Sham, Self-Consistent Equations Including Exchange and Correlation Effects, Physical Review 140, A1133 (1965).
- [120]J. Lee, D. Kim, H. Kim, S. Seo, S. M. Cho, and N. Park, Formamidinium and Cesium Hybridization for Photo- and Moisture-Stable Perovskite Solar Cell, Advanced Energy Materials 5, 1501310 (2015).
- [121]C. Yi, J. Luo, S. Meloni, A. Boziki, N. Ashari-Astani, C. Grätzel, S. M. Zakeeruddin, U. Röthlisberger, and M. Grätzel, Entropic stabilization of mixed A-cation ABX 3 metal halide perovskites for high performance perovskite solar cells, Energy & Environmental Science 9, 656 (2016).
- [122]L. Merten, A. Hinderhofer, T. Baumeler, N. Arora, J. Hagenlocher, S. M. Zakeeruddin, M. I. Dar, M. Graetzel, and F. Schreiber, Quantifying Stabilized Phase Purity in Formamidinium-Based Multiple-Cation Hybrid Perovskites, Chemistry of Materials 33, 2769 (2021).
- [123]H. Xue, G. Brocks, and S. Tao, First-principles calculations of defects in metal halide perovskites: A performance comparison of density functionals, Physical Review Materials 5, 125408 (2021).
- [124]H. Xue, J. M. Vicent-Luna, S. Tao, and G. Brocks, Compound Defects in Halide Perovskites: A First-Principles Study of CsPbI 3, The Journal of Physical Chemistry C 127, 1189 (2023).
- [125]H. Peng, Z.-H. Yang, J. P. Perdew, and J. Sun, Versatile van der Waals Density Functional Based on a Meta-Generalized Gradient Approximation, Physical Review X 6, 041005 (2016).
- [126]M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias, and J. D. Joannopoulos, Iterative minimization techniques for ab initio total-energy calculations: molecular dynamics and conjugate gradients, Reviews of Modern Physics 64, 1045 (1992).
- [127]M. A. Carignano, S. A. Aravindh, I. S. Rogan, J. Even, and C. Katan, Critical Fluctuations and Anharmonicity in Lead Iodide Perovskites from Molecular Dynamics Supercell Simulations, The Journal of Physical Chemistry C 121, 20729 (2017).
- [128]A. P. Bartók, M. C. Payne, R. Kondor, and G. Csányi, Gaussian approximation potentials: The accuracy of quantum mechanics, without the electrons, Physical review letters 104, 136403 (2010).
- [129]F. Musil, S. D. Je, Y. Yang, J. E. Campbell, G. M. Day, and M. Ceriotti, Machine learning for the structure–energy–property landscapes of molecular crystals, Chemical Science 9, 1289 (2018).
- [130]F. Musil, M. Veit, A. Goscinski, G. Fraux, M. J. Willatt, M. Stricker, T. Junge, and M. Ceriotti, Efficient implementation of atom-density representations, The Journal of Chemical Physics 154, 114109 (2021).
- [131]R. Drautz, Atomic cluster expansion for accurate and transferable interatomic potentials, Physical Review B 99, 014104 (2019).
- [132]G. Dusson, M. Bachmayr, G. Csányi, R. Drautz, S. Etter, C. Van Der Oord, and C. Ortner, Atomic cluster expansion: Completeness, efficiency and stability, Journal of Computational Physics 454, 110946 (2022).
- [133]W. C. Witt, C. Van Der Oord, E. Gëlzinytė, T. Järvinen, A. Ross, J. P. Darby, C. H. Ho, W. J. Baldwin, M. Sachs, J. Kermode, N. Bernstein, G. Csányi, and C. Ortner, ACEpotentials.jl: A Julia implementation of the atomic cluster expansion, The Journal of Chemical Physics 159, 164101 (2023).
- [134]L. McInnes, J. Healy, N. Saul, and L. Großberger, UMAP: Uniform Manifold Approximation and Projection, Journal of Open Source Software 3, 861 (2018).
- [135]J. Wiktor, E. Fransson, D. Kubicki, and P. Erhart, Quantifying Dynamic Tilting in Halide Perovskites: Chemical Trends and Local Correlations, Chemistry of Materials 35, 6737 (2023).
- [136]C.-C. Chen, A. B. Herhold, C. S. Johnson, and A. P. Alivisatos, Size Dependence of Structural Metastability in Semiconductor Nanocrystals, Science 276, 398 (1997).
- [137]K. Jacobs, D. Zaziski, E. C. Scher, A. B. Herhold, and A. Paul Alivisatos, Activation Volumes for Solid-Solid Transformations in Nanocrystals, Science 293, 1803 (2001).
- [138]Y. Peng, W. Li, T. Still, A. G. Yodh, and Y. Han, In situ observation of coalescence of nuclei in colloidal crystal-crystal transitions, Nature Communications 14, 4905 (2023).
- [139]P. A. Santos-Florez, H. Yanxon, B. Kang, Y. Yao, and Q. Zhu, Size-Dependent Nucleation in Crystal Phase Transition from Machine Learning Metadynamics, Physical Review Letters 129, 185701 (2022).
- [140]P. Ball, Size matters, Nature Materials 21, 1341 (2022).
- [141]P. Bolhuis, M. Hagen, and D. Frenkel, Isostructural solid-solid transition in crystalline systems with short-ranged interaction, Physical Review E 50, 4880 (1994).
- [142]P. G. Debenedetti, Metastable liquids: concepts and principles, Physical chemistry (Princeton University Press, Princeton, N.J, 1996).
- [143]X.-M. Bai and M. Li, Test of classical nucleation theory via molecular-dynamics simulation, The Journal of Chemical Physics 122, 224510 (2005).
- [144]E. Sanz, C. Vega, J. R. Espinosa, R. Caballero-Bernal, J. L. F. Abascal, and C. Valeriani, Homogeneous Ice Nucleation at Moderate Supercooling from Molecular Simulation, Journal of the American Chemical Society 135, 15008 (2013).
- [145]A. Z. Panagiotopoulos, Direct determination of phase coexistence properties of fluids by Monte Carlo simulation in a new ensemble, Molecular Physics 61, 813 (1987).
- [146]X.-M. Bai and M. Li, Calculation of solid-liquid interfacial free energy: A classical nucleation theory based approach, The Journal of Chemical Physics 124, 124707 (2006).
- [147]D. Gonzalez-Salgado, A. Dopazo-Paz, P. Gomez-Alvarez, J. M. Miguez, and C. Vega, Solid-Solid and Solid-Fluid Equilibria of the Most Popular Models of Methanol Obtained by Computer Simulation, The Journal of Physical Chemistry B 115, 3522 (2011).
- [148]J. R. Espinosa, C. Vega, C. Valeriani, and E. Sanz, Seeding approach to crystal nucleation, The Journal of Chemical Physics 144, 034501 (2016).
- [149]J. R. Espinosa, E. Sanz, C. Valeriani, and C. Vega, Homogeneous ice nucleation evaluated for several water models, The Journal of Chemical Physics 141, 18C529 (2014).
- [150]J. R. Espinosa, E. Sanz, C. Valeriani, and C. Vega, On fluid-solid direct coexistence simulations: The pseudo-hard sphere model, The Journal of Chemical Physics 139, 144502 (2013).
- [151]K. Binder and P. Virnau, Phase transitions and phase coexistence: equilibrium systems versus externally driven or active systems - Some perspectives, Soft Materials 19, 267 (2021).
- [152]P. Rein Ten Wolde, M. J. Ruiz-Montero, and D. Frenkel, Numerical calculation of the rate of crystal nucleation in a Lennard-Jones system at moderate undercooling, The Journal of Chemical Physics 104, 9932 (1996).
- [153]J. S. Van Duijnveldt and D. Frenkel, Computer simulation study of free energy barriers in crystal nucleation, The Journal of Chemical Physics 96, 4655 (1992).
- [154]S. Pogatscher, D. Leutenegger, J. E. K. Schawe, P. J. Uggowitzer, and J. F. Löffler, Solid–solid phase transitions via melting in metals, Nature Communications 7, 11113 (2016).
- [155]J. Chen, J. Wang, X. Xu, J. Li, J. Song, S. Lan, S. Liu, B. Cai, B. Han, J. T. Precht, D. Ginger, and H. Zeng, Efficient and bright white light-emitting diodes based on single-layer heterophase halide perovskites, Nature Photonics 15, 238 (2021).
- [156]W. Humphrey, A. Dalke, and K. Schulten, VMD: Visual molecular dynamics, Journal of Molecular Graphics 14, 33 (1996).
- [157]Y. A. Chu, B. Moran, G. B. Olson, and A. C. E. Reid, A model for nonclassical nucleation of solid-solid structural phase transformations, Metallurgical and Materials Transactions A 31, 1321 (2000).
- [158]S. Chen, C. Wu, Q. Shang, Z. Liu, C. He, W. Zhou, J. Zhao, J. Zhang, J. Qi, Q. Zhang, X. Wang, J. Li, and P. Gao, Atomic structure and electrical/ionic activity of antiphase boundary in CH3NH3PbI3, Acta Materialia 234, 118010 (2022).
- [159]S. Kosar, A. J. Winchester, T. A. S. Doherty, S. Macpherson, C. E. Petoukhoff, K. Frohna, M. Anaya, N. S. Chan, J. Madéo, M. K. L. Man, S. D. Stranks, and K. M. Dani, Unraveling the varied nature and roles of defects in hybrid halide perovskites with time-resolved photoemission electron microscopy, Energy & Environmental Science 14, 6320 (2021).
- [160]M. Li, Y. Shen, K. Luo, Q. An, P. Gao, P. Xiao, and Y. Zou, Harnessing dislocation motion using an electric field, Nature Materials 10.1038/s41563-023-01572-7 (2023).
- [161]K. W. P. Orr, J. Diao, M. N. Lintangpradipto, D. J. Batey, A. N. Iqbal, S. Kahmann, K. Frohna, M. Dubajic, S. J. Zelewski, A. E. Dearle, T. A. Selby, P. Li, T. A. S. Doherty, S. Hofmann, O. M. Bakr, I. K. Robinson, and S. D. Stranks, Imaging Light-Induced Migration of Dislocations in Halide Perovskites with 3D Nanoscale Strain Mapping (2023), arXiv:2304.09554 [cond-mat, physics:physics].
- [162]C. Ducati, J. F. Orri, A. Iqbal, T. Yang, T. Doherty, T. Selby, Y.-H. Chiang, M. Anaya, C. Allen, and S. Stranks, Visualising nanoscale bias-induced degradation in halide perovskite solar absorbers, preprint (In Review, 2023).
- [163]T. A. Wani, J. Shamsi, X. Bai, N. Arora, and M. I. Dar, Advances in All-Inorganic Perovskite Nanocrystal-Based White Light Emitting Devices, ACS Omega 8, 17337 (2023).
- [164]G. H. Booth, A. J. W. Thom, and A. Alavi, Fermion Monte Carlo without fixed nodes: A game of life, death, and annihilation in Slater determinant space, The Journal of Chemical Physics 131, 054106 (2009).
- [165]H.-J. Werner and M. Schütz, An efficient local coupled cluster method for accurate thermochemistry of large systems, The Journal of Chemical Physics 135, 144116 (2011).
- [166]P. R. Nagy, G. Samu, and M. Kállay, Optimization of the Linear-Scaling Local Natural Orbital CCSD(T) Method: Improved Algorithm and Benchmark Applications, Journal of Chemical Theory and Computation 14, 4193 (2018).
- [167]T. Gruber, K. Liao, T. Tsatsoulis, F. Hummel, and A. Grüneis, Applying the Coupled-Cluster Ansatz to Solids and Surfaces in the Thermodynamic Limit, Physical Review X 8, 021043 (2018).
- [168]Z.-H. Cui, H. Zhai, X. Zhang, and G. K.-L. Chan, Systematic electronic structure in the cuprate parent state from quantum many-body simulations, Science 377, 1192 (2022).
- [169]N. A. Bogdanov, G. Li Manni, S. Sharma, O. Gunnarsson, and A. Alavi, Enhancement of superexchange due to synergetic breathing and hopping in corner-sharing cuprates, Nature Physics 18, 190 (2022).
- [170]M. Ochi, TC++: First-principles calculation code for solids using the transcorrelated method, Computer Physics Communications 287, 108687 (2023).
- [171]V. Garcia, M. Bibes, L. Bocher, S. Valencia, F. Kronast, A. Crassous, X. Moya, S. Enouz-Vedrenne, A. Gloter, D. Imhoff, C. Deranlot, N. D. Mathur, S. Fusil, K. Bouzehouane, and A. Barthélémy, Ferroelectric Control of Spin Polarization, Science 327, 1106 (2010).
- [172]W. Zhong, D. Vanderbilt, and K. M. Rabe, First-principles theory of ferroelectric phase transitions for perovskites: The case of BaTiO 3, Physical Review B 52, 6301 (1995).
- [173]H. Y. Wang, Y. Iwasa, M. Kawasaki, B. Keimer, N. Nagaosa, and Y. Tokura, Emergent phenomena at oxide interfaces, Nature Materials 11, 103 (2012).
- [174]S. Manipatruni, D. E. Nikonov, C.-C. Lin, T. A. Gosavi, H. Liu, B. Prasad, Y.-L. Huang, E. Bonturim, R. Ramesh, and I. A. Young, Scalable energy-efficient magnetoelectric spin–orbit logic, Nature 565, 35 (2019).
- [175]S. Manipatruni, D. E. Nikonov, C.-C. Lin, B. Prasad, Y.-L. Huang, A. R. Damodaran, Z. Chen, R. Ramesh, and I. A. Young, Voltage control of unidirectional anisotropy in ferromagnet-multiferroic system, Science Advances 4, eaat4229 (2018).
- [176]J. Mannhart and D. G. Schlom, Oxide Interfaces—An Opportunity for Electronics, Science 327, 1607 (2010).
- [177]Y. Zhang, E. Parsonnet, A. Fernandez, S. M. Griffin, H. Huyan, C.-K. Lin, T. Lei, J. Jin, E. S. Barnard, A. Raja, P. Behera, X. Pan, R. Ramesh, and P. Yang, Ferroelectricity in a semiconducting all-inorganic halide perovskite, Science Advances 8, eabj5881 (2022).
- [178]T. Kimura, T. Goto, H. Shintani, K. Ishizaka, T. Arima, and Y. Tokura, Magnetic control of ferroelectric polarization, Nature 426, 55 (2003).
- [179]J. Wang, J. B. Neaton, H. Zheng, V. Nagarajan, S. B. Ogale, B. Liu, D. Viehland, V. Vaithyanathan, D. G. Schlom, U. V. Waghmare, N. A. Spaldin, K. M. Rabe, M. Wuttig, and R. Ramesh, Epitaxial BiFeO 3 Multiferroic Thin Film Heterostructures, Science 299, 1719 (2003).
- [180]A. Ohtomo and H. Y. Hwang, A high-mobility electron gas at the LaAlO3/SrTiO3 heterointerface, Nature 427, 423 (2004).
- [181]N. Reyren, S. Thiel, A. D. Caviglia, L. F. Kourkoutis, G. Hammerl, C. Richter, C. W. Schneider, T. Kopp, A.-S. Rüetschi, D. Jaccard, M. Gabay, D. A. Muller, J.-M. Triscone, and J. Mannhart, Superconducting Interfaces Between Insulating Oxides, Science 317, 1196 (2007).
- [182]J. G. Bednorz and K. A. Müller, Perovskite-type oxides—The new approach to high-T c superconductivity, Reviews of Modern Physics 60, 585 (1988).
- [183]T. He, Q. Huang, A. P. Ramirez, Y. Wang, K. A. Regan, N. Rogado, M. A. Hayward, M. K. Haas, J. S. Slusky, K. Inumara, H. W. Zandbergen, N. P. Ong, and R. J. Cava, Superconductivity in the non-oxide perovskite MgCNi3, Nature 411, 54 (2001).
- [184]J. A. Bert, B. Kalisky, C. Bell, M. Kim, Y. Hikita, H. Y. Hwang, and K. A. Moler, Direct imaging of the coexistence of ferromagnetism and superconductivity at the LaAlO3/SrTiO3 interface, Nature Physics 7, 767 (2011).
- [185]K. Lee, B. Y. Wang, M. Osada, B. H. Goodge, T. C. Wang, Y. Lee, S. Harvey, W. J. Kim, Y. Yu, C. Murthy, S. Raghu, L. F. Kourkoutis, and H. Y. Hwang, Linear-in-temperature resistivity for optimally superconducting (Nd,Sr)NiO2, Nature 619, 288 (2023).
- [186]S. Datta and B. Das, Electronic analog of the electro-optic modulator, Applied Physics Letters 56, 665 (1990).
- [187]N. Jones, How to stop data centres from gobbling up the world’s electricity, Nature 561, 163 (2018).
- [188]A. Andrae and T. Edler, On Global Electricity Usage of Communication Technology: Trends to 2030, Challenges 6, 117 (2015).
- [189]R. Ramesh, Materials for a Sustainable Microelectronics Future: Electric Field Control of Magnetism with Multiferroics, Journal of the Indian Institute of Science 102, 489 (2022).
- [190]I. M. Zeron, J. L. F. Abascal, and C. Vega, A force field of Li+, Na+, K+, Mg2+, Ca2+, Cl-, and SO42- in aqueous solution based on the TIP4P/2005 water model and scaled charges for the ions, The Journal of Chemical Physics 151, 134504 (2019).
- [191]S. Blazquez, M. M. Conde, J. L. F. Abascal, and C. Vega, The Madrid-2019 force field for electrolytes in water using TIP4P/2005 and scaled charges: Extension to the ions F−, Br−, I−, Rb+, and Cs+, The Journal of Chemical Physics 156, 044505 (2022).
- [192]A. P. Thompson, H. M. Aktulga, R. Berger, D. S. Bolintineanu, W. M. Brown, P. S. Crozier, P. J. In ’T Veld, A. Kohlmeyer, S. G. Moore, T. D. Nguyen, R. Shan, M. J. Stevens, J. Tranchida, C. Trott, and S. J. Plimpton, LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales, Computer Physics Communications 271, 108171 (2022).
- [193]G. Bussi, D. Donadio, and M. Parrinello, Canonical sampling through velocity rescaling, The Journal of Chemical Physics 126, 014101 (2007).
- [194]G. J. Martyna, D. J. Tobias, and M. L. Klein, Constant pressure molecular dynamics algorithms, The Journal of Chemical Physics 101, 4177 (1994).
- [195]M. Parrinello and A. Rahman, Polymorphic transitions in single crystals: A new molecular dynamics method, Journal of Applied Physics 52, 7182 (1981).
- [196]W. Shinoda, M. Shiga, and M. Mikami, Rapid estimation of elastic constants by molecular dynamics simulation under constant stress, Physical Review B 69, 134103 (2004).
- [197]G. A. Tribello, M. Bononi, D. Branduardi, C. Camilloni, and G. Bussi, PLUMED 2: New feathers for an old bird, Computer Physics Communications 185, 604 (2014).
- [198]M. Bonomi, D. Branduardi, G. Bussi, C. Camilloni, D. Provasi, P. Raiteri, D. Donadio, F. Marinelli, F. Pietrucci, R. A. Broglia, and M. Parrinello, PLUMED: A portable plugin for free-energy calculations with molecular dynamics, Computer Physics Communications 180, 1961 (2009).
- [199]R. García Fernández, J. L. F. Abascal, and C. Vega, The melting point of ice Ih for common water models calculated from direct coexistence of the solid-liquid interface, The Journal of Chemical Physics 124, 144506 (2006).
- [200]D. B. Straus, S. Guo, and R. J. Cava, Kinetically Stable Single Crystals of Perovskite-Phase CsPbI3, Journal of the American Chemical Society 141, 11435 (2019).
- [201]G. Kresse and J. Furthmüller, Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set, Computational Materials Science 6, 15 (1996).
- [202]G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Physical Review B 54, 11169 (1996).
- [203]G. Kresse and D. Joubert, From ultrasoft pseudopotentials to the projector augmented-wave method, Physical Review B 59, 1758 (1999).
- [204]J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized Gradient Approximation Made Simple, Physical Review Letters 77, 3865 (1996).
- [205]G. I. Csonka, J. P. Perdew, A. Ruzsinszky, P. H. T. Philipsen, S. Lebègue, J. Paier, O. A. Vydrov, and J. G. Ángyán, Assessing the performance of recent density functionals for bulk solids, Physical Review B 79, 155107 (2009).
- [206]J. W. Furness, A. D. Kaplan, J. Ning, J. P. Perdew, and J. Sun, Accurate and Numerically Efficient r²SCAN Meta-Generalized Gradient Approximation, The Journal of Physical Chemistry Letters 11, 8208 (2020).
- [207]M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids (Oxford University Press, 2017).
- [208]Z. Fan, Z. Zeng, C. Zhang, Y. Wang, K. Song, H. Dong, Y. Chen, and T. Ala-Nissila, Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport, Physical Review B 104, 104309 (2021).
- [209]Z. Fan, Improving the accuracy of the neuroevolution machine learning potential for multi-component systems, Journal of Physics: Condensed Matter 34, 125902 (2022).
- [210]Z. Fan, Y. Wang, P. Ying, K. Song, J. Wang, Y. Wang, Z. Zeng, K. Xu, E. Lindgren, J. M. Rahm, A. J. Gabourie, J. Liu, H. Dong, J. Wu, Y. Chen, Z. Zhong, J. Sun, P. Erhart, Y. Su, and T. Ala-Nissila, GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations, The Journal of Chemical Physics 157, 114801 (2022).
- [211]Z. Fan, W. Chen, V. Vierimaa, and A. Harju, Efficient molecular dynamics simulations with many-body potentials on graphics processing units, Computer Physics Communications 218, 10 (2017).
- [212]E. Fransson, J. Wiktor, and P. Erhart, Phase Transitions in Inorganic Halide Perovskites from Machine-Learned Potentials, The Journal of Physical Chemistry C 127, 13773 (2023).
- [213]E. Fransson, J. M. Rahm, J. Wiktor, and P. Erhart, Revealing the Free Energy Landscape of Halide Perovskites: Metastability and Transition Characters in CsPbBr3 and MAPbI3, Chemistry of Materials 35, 8229 (2023).
