Machine Learning Screening for 2D Materials
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The commercialization of 2D materials is hindered by the difficulty in achieving mass production without compromising their excellent properties, and traditional experimental methods are inefficient and costly in discovering new materials with high elastic modulus and shear modulus.
Innovation Solution
A device and method utilizing deep learning, machine learning, and high-throughput calculation to generate, classify, and derive new 2D materials by predicting space groups, analyzing chemical similarity, and performing element substitution, leveraging generative adversarial networks, random forest models, and density functional theory for property analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional experimental methods are used to discover new 2D materials, then material properties can be verified through physical experiments, but the process is inefficient, costly, and time-consuming
Solution Approach 1:
The patent performs preliminary computational screening of material properties using density functional theory (DFT) calculations and machine learning models before physical experimentation. This preliminary action filters out unsuitable candidates, ensuring that only materials with predicted high elastic modulus and shear modulus proceed to experimental verification, thereby improving efficiency while maintaining reliability
Solution Approach 2:
The patent replaces traditional mechanical experimental testing with computational mechanics approaches including DFT calculations and machine learning-based property prediction. This substitution enables rapid assessment of material properties without physical experiments, dramatically improving discovery efficiency while maintaining sufficient accuracy for screening purposes
2Measurement precision
If traditional experimental methods are used to discover new 2D materials, then accurate material characterization can be achieved, but the cost and time requirements increase significantly
Solution Approach 1:
The patent performs preliminary computational characterization using established DFT methods and machine learning models to predict key material properties including elastic modulus, shear modulus, and structural stability. This preliminary characterization provides accurate enough data for screening purposes without the time investment required for full experimental characterization
Solution Approach 2:
The patent segments the material characterization process into two stages: (1) computational screening stage using DFT and machine learning for rapid property prediction, and (2) experimental verification stage for final confirmation. This segmentation allows most materials to be evaluated quickly computationally, reducing overall discovery time while maintaining measurement precision for final selections
3Reliability
If new 2D materials are synthesized using mechanical exfoliation and chemical vapor deposition, then material quality can be maintained, but mass production becomes difficult
Solution Approach 1:
The patent performs preliminary computational assessment of synthesis feasibility and structural stability for candidate materials before attempting production. By predicting which materials are most likely to be synthesizable and maintain quality, the approach guides selection toward materials that balance quality requirements with manufacturability considerations
Solution Approach 2:
The patent explores parameter changes in material composition through computational screening of different chemical formulas and structures. By identifying materials with optimal compositional parameters that balance structural stability, mechanical properties, and synthesis feasibility, the approach enables selection of materials more amenable to scalable production while maintaining quality
4Strength
If existing 2D materials like graphene and MoS are used, then excellent electrical and mechanical properties are achieved, but bonding with other materials remains difficult
Solution Approach 1:
The patent applies local quality by screening for materials with specific local chemical environments and surface properties that enhance bonding capability while maintaining overall structural integrity. By identifying materials with appropriate surface chemistry and local atomic configurations, the approach enables improved interfacial bonding without sacrificing the excellent mechanical properties of 2D materials
Solution Approach 2:
The patent explores composite material concepts by screening for 2D materials that can be combined with other materials through computational prediction of heterostructure formation. By identifying materials with complementary properties and favorable interface characteristics, the approach enables development of composite structures that leverage both the mechanical strength of 2D materials and the bonding versatility of partner materials
Data Source
AI summary
A technology capable of finding new 2D materials with high elastic modulus and shear modulus through using deep learning, machine learning, and high-throughput calculation methods are described. A device for modeling 2-dimensional (2D) material using machine learning includes a material classification unit receiving data on the virtual inorganic material chemical formulas from the virtual inorganic material generation unit and classifying a 2D material among the plurality of virtual inorganic material chemical formulas into a preliminary 2D material, a space group analysis unit receiving data on the preliminary 2D material from the material classification unit, predicting the space group of the preliminary 2D material, and selecting the preliminary 2D material having the same space group as an existing 2D material as a structurally similar 2D material.