|
|
| Type |
Symposium |
| Area |
Recent Trends in Computational Chemistry: Bigdata and Artificial Intelligence |
| Room No. |
Room 402 |
| Time |
FRI 16:00-: |
| Code |
PHYS3-5 |
| Subject |
Millions and Billions of Porous Materials |
| Authors |
JIHAN KIM
Chemical and Biomolecular Engineering, KAIST, United States |
| Abstract |
The talk will be divided into two parts. Firstly, I will describe an algorithm that we have developed to generate (in silico) novel hetero-interpenetrated porous materials from a set of existing materials. The method has yielded thermodynamically stable MOFs that can possess synergetic properties for gas adsorption/separations as well as bifunctional materials. Second, I will describe an algorithm that uses deep learning to generate new porous materials in a property space (as opposed to materials space). This method can be used to determine the performance limitations of various applications, which in effect can be used to determine whether or not a material can exceed certain targets (e.g. CO2 capture amount, methane storage amount) required for commercialization. |
| E-mail |
jihankim@kaist.ac.kr |
|
121st General Meeting of the KCS