Material Performance Modulation And Material Design Based On Organic-inorganic Hybridization

Organic-inorganic hybrid materials offer outstanding optoelectronic properties,simple synthetic processes,and high degrees of tunability,rendering them pivotal in the field of optoelectronics.Whether it is the organic or inorganic components,the diversity within organic-inorganic hybrid materials provides a rich array of structural choices.Due to the immense tunability of the components in organic-inorganic hybrid materials,the efficient investigation of novel materials necessitates a comprehensive understanding of their formation mechanisms and a deep exploration of the structure-property relationships within these materials,thus facilitating the development of new organic-inorganic hybrid materials.Simultaneously,materials research methods based on high-throughput computing and machine learning can greatly assist in exploring the vast phase space of organic-inorganic hybrid materials.With the development of modern technologies such as big data,machine learning,artificial intelligence,and high-performance computing,a data-driven approach to studying material properties and structure-property relationships is becoming an inevitable trend in future materials science research.In this context,researchers need infrastructure that supports materials science research methods based on high-throughput computing and machine learning,enabling the efficient application of this emerging paradigm to deepen our understanding of structure-property relationships in materials and accelerate the development and design of new materials.Addressing the aforementioned challenges,this paper undertakes the following tasks:1.As the primary contributing author,the development of infrastructure supporting material science research methods based on high-throughput computing and machine learning is realized through the creation of the Artificial Intelligence-assisted,data-driven material intelligent design software package,JAMIP.This development encompasses modules related to high-throughput computing,including but not limited to a structure manipulation module,task monitoring and correction module,and a database interaction module,all of which were developed within JAMIP.The core framework of JAMIP includes data generation,data collection management and storage,as well as machine learning/data mining functionality centered around high-throughput computing.Within the high-throughput computing-related modules,I personally developed features and modules supporting data generation,including:（1）the structure manipulation module,which assists in generating candidate material structures for high-throughput computations;（2）the task monitoring and error correction module,responsible for overseeing and managing tasks during high-throughput computations and handling exceptional situations based on predefined error sets;（3）the database interaction module,which provides a user-friendly Python environment,obviating the need for users to learn complex structured database query languages,thereby reducing the barrier to database utilization.These contributions lay the foundation for high-throughput data generation patterns in fourth-paradigm material science research.Subsequently,leveraging the structural manipulation and high-throughput computation workflow functionalities within the high-throughput computing-related modules of JAMIP,we conducted structural optimization,self-consistent field calculations,and optical property calculations for a total of 696 halide perovskite materials.Analysis of the data revealed a weak coupling between electronic properties and atomic filling factors;bandgap values exhibited a general monotonic dependence on the electronegativity difference between B-site and X-site atoms in halide perovskite materials;and no conspicuous trend in optical transition intensities was observed for different candidate structures concerning bandgap variations.2.In response to the demand for designing novel organic-inorganic hybrid materials,a machine learning model was constructed based on limited training data to predict the formation energies and stability of organic-inorganic hybrid materials.The model achieved high prediction accuracy with a Mean Absolute Error（MAE）of 0.063 e V/atom and an R-squared（R²）value of 0.94,demonstrating its robustness and extensibility.Subsequently,feature engineering based on the trained model identified strong correlations between the formation energy of organic-inorganic hybrid materials and the periodic average values of constituent elements and bandgap averages.Finally,based on the essential features output by the model,a descriptor for predicting the formation energy of organic-inorganic hybrid materials was proposed,exhibiting a Pearson correlation coefficient（r）of 0.824,indicating a high level of predictive accuracy.This research provides both a methodology and software foundation for the design of new organic-inorganic hybrid materials.3.Employing principles of surface charge transfer doping and thermal stability analysis,several organic molecule dopants with potential applications in n-type diamond surface charge transfer doping were predicted,while the influencing factors affecting the effectiveness of n-type diamond surface charge transfer doping were identified.Methylviologen and benzylviologen were found to be the most suitable organic molecule dopants for oxygen-or fluorine-terminated diamond surfaces,achieving charge surface densities of 2.60×10¹³ cm^-2 and 9.20×10¹² cm^-2,respectively.The critical factors influencing doping effects were identified as the energy level difference between the highest occupied molecular orbital of organic molecules and the maximum valence band of the diamond substrate,as well as the interaction between organic molecules and the substrate surface.Moreover,an increase in the thickness and density of organic molecules enhanced the effectiveness of charge transfer doping,although marginal effects were observed.This research provides theoretical guidance for the exploration of surface charge transfer doping for n-type diamond through organic molecules.4.Theoretical investigations were conducted on high-performance triethylene glycol dimethyl ether triacrylate（TET）crosslinked flexible perovskite thin films prepared by experimental collaborators,aiming to understand the physical mechanisms behind the enhanced stability and optoelectronic properties resulting from the introduction of TET organic molecules into FAPb I₃ films.Utilizing a range of characterization techniques,including Fourier-transform infrared spectroscopy,X-ray diffraction,time-resolved photoluminescence,and ultraviolet photoelectron spectroscopy,we confirmed the crosslinking of TET molecules,the interaction between C=O groups in the organic component and Pb²⁺ions on the perovskite surface,and the enhancement of perovskite optoelectronic properties due to TET crosslinking.First-principles calculations revealed that TET doping exhibited a higher binding energy（72.0 k J/mole）compared to commercial PMMA,with the source of this binding energy originating from dipole-ion interactions between TET and the perovskite.Furthermore,TET doping led to a higher charge density on Pb²⁺ions（3.04>3.02）and a shorter Pb²⁺-O distance（2.57?<2.62?）compared to PMMA,resulting in stronger dipole-ion interactions.Analysis of the electronic structure before and after TET doping demonstrated an increase in the bandgap from 1.347 e V to 1.424 e V,confirming the passivation effect of TET on intrinsic defects in the perovskite.This research provides strong support for the application of polymer crosslinking in flexible perovskite solar cells.