PAL 2.0: a physics-driven bayesian optimization framework for material discovery.

The lack of efficient discovery tools for advanced functional materials remains a major bottleneck to enabling advances in the next-generation energy, health, and sustainability technologies. One main factor contributing to this inefficiency is the large combinatorial space of materials (with respect to material compositions and processing conditions) that is typically redolent of such materials-centric applications. Searches of this large combinatorial space are often influenced by expert knowledge and clustered close to material configurations that are known to perform well, thus ignoring potentially high-performing candidates in unanticipated regions of the composition-space or processing protocol. Moreover, experimental characterization or first principles quantum mechanical calculations of all possible material candidates can be prohibitively expensive, making exhaustive approaches to determine the best candidates infeasible. As a result, there remains a need for the development of computational algorithms that can efficiently search a large parameter space for a given material application. Here, we introduce PAL 2.0, a method that combines a physics-based surrogate model with Bayesian optimization. The key contributing factor of our proposed framework is the ability to create a physics-based hypothesis using XGBoost and Neural Networks. This hypothesis provides a physics-based "prior" (or initial beliefs) to a Gaussian process model, which is then used to perform a search of the material design space. In this paper, we demonstrate the usefulness of our approach on three material test cases: (1) discovery of metal halide perovskites with desired photovoltaic properties, (2) design of metal halide perovskite-solvent pairs that produce the best solution-processed films and (3) design of organic thermoelectric semiconductors. Our results indicate that the novel PAL 2.0 approach outperforms other state-of-the-art methods in its efficiency to search the material design space for the optimal candidate. We also demonstrate the physics-based surrogate models constructed in PAL 2.0 have lower prediction errors for material compositions not seen by the model. To the best of our knowledge, there is no competing algorithm capable of this useful combination for materials discovery, especially those for which data are scarce.

Blended Conjugated Host and Unconjugated Dopant Polymers Towards N-type All-Polymer Conductors and High-ZT Thermoelectrics.

N-Type thermoelectrics typically consist of small molecule dopant+polymer host. Only a few polymer dopant+polymer host systems have been reported, and these have lower thermoelectric parameters. N-type polymers with high crystallinity and order are generally used for high-conductivity ( σ ${\sigma }$ ) organic conductors. Few n-type polymers with only short-range lamellar stacking for high-conductivity materials have been reported. Here, we describe an n-type short-range lamellar-stacked all-polymer thermoelectric system with highest σ ${\sigma }$ of 78 S-1 , power factor (PF) of 163 µW m-1 K-2 , and maximum Figure of merit (ZT) of 0.53 at room temperature with a dopant/host ratio of 75 wt%. The minor effect of polymer dopant on the molecular arrangement of conjugated polymer PDPIN at high ratios, high doping capability, high Seebeck coefficient (S) absolute values relative to σ ${\sigma }$ , and atypical decreased thermal conductivity ( κ ${\kappa }$ ) with increased doping ratio contribute to the promising performance.

Untying the Cesium "Not": Cesium-Iodoplumbate Complexation in Perovskite Solution-Processing Inks Has Implications for Crystallization.

We illustrate the critical importance of the energetics of cation-solvent versus cation-iodoplumbate interactions in determining the stability of ABX3 perovskite precursors in a dimethylformamide (DMF) solvent medium. We have shown, through a complementary suite of nuclear magnetic resonance (NMR) and computational studies, that Cs+ exhibits significantly different solvent vs iodoplumbate interactions compared to organic A+-site cations such as CH3NH3+ (MA+). Two NMR studies were conducted: 133Cs NMR analysis shows that Cs+ and MA+ compete for coordination with PbI3- in DMF. 207Pb NMR studies of PbI2 with cationic iodides show that perovskite-forming Cs+ (and, somewhat, Rb+) do not comport with the 207Pb chemical shift trend found for Li+, Na+, and K+. Three independent computational approaches (density functional theory (DFT), ab initio Molecular Dynamics (AIMD), and a polarizable force field within Molecular Dynamics) yielded strikingly similar results: Cs+ interacts more strongly with the PbI3- iodoplumbate than does MA+ in a polar solvent environment like DMF. The stronger energy preference for PbI3- coordination of Cs+ vs MA+ in DMF demonstrates that Cs+ is not simply a postcrystallization cation "fit" for the perovskite A+-site. Instead, it may facilitate preorganization of the framework precursor that eventually transforms into the crystalline perovskite structure.

3,4,5-Trimethoxy Substitution on an N-DMBI Dopant with New N-Type Polymers: Polymer-Dopant Matching for Improved Conductivity-Seebeck Coefficient Relationship.

Achieving high electrical conductivity and thermoelectric power factor simultaneously for n-type organic thermoelectrics is still challenging. By constructing two new acceptor-acceptor n-type conjugated polymers with different backbones and introducing the 3,4,5-trimethoxyphenyl group to form the new n-type dopant 1,3-dimethyl-2-(3,4,5-trimethoxyphenyl)-2,3-dihydro-1H-benzo[d]imidazole (TP-DMBI), high electrical conductivity of 11 S cm-1 and power factor of 32 µW m-1 K-2 are achieved. Calculations using Density Functional Theory show that TP-DMBI presents a higher singly occupied molecular orbital (SOMO) energy level of -1.94 eV than that of the common dopant 4-(1, 3-dimethyl-2, 3-dihydro-1H-benzoimidazol-2-yl) phenyl) dimethylamine (N-DMBI) (-2.36 eV), which can result in a larger offset between the SOMO of dopant and lowest unoccupied molecular orbital (LUMO) of n-type polymers, though that effect may not be dominant in the present work. The doped polymer films exhibit higher Seebeck coefficient and power factor than films using N-DMBI at the same doping levels or similar electrical conductivity levels. Moreover, TP-DMBI doped polymer films offer much higher electron mobility of up to 0.53 cm2 V-1 s-1 than films with N-DMBI doping, demonstrating the potential of TP-DMBI, and 3,4,5-trialkoxy DMBIs more broadly, for high performance n-type organic thermoelectrics.