Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals
Abstract Diffusion in alloys is an important class of atomic processes. However, atomistic simulations of diffusion in chemically complex solids are confronted with the timescale problem: the accessible simulation time is usually far shorter than that of experimental interest. In this work, long‐tim...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-02-01
|
| Series: | Advanced Science |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/advs.202304122 |
| _version_ | 1827361443863855104 |
|---|---|
| author | Hao Tang Boning Li Yixuan Song Mengren Liu Haowei Xu Guoqing Wang Heejung Chung Ju Li |
| author_facet | Hao Tang Boning Li Yixuan Song Mengren Liu Haowei Xu Guoqing Wang Heejung Chung Ju Li |
| author_sort | Hao Tang |
| collection | DOAJ |
| description | Abstract Diffusion in alloys is an important class of atomic processes. However, atomistic simulations of diffusion in chemically complex solids are confronted with the timescale problem: the accessible simulation time is usually far shorter than that of experimental interest. In this work, long‐timescale simulation methods are developed using reinforcement learning (RL) that extends simulation capability to match the duration of experimental interest. Two special limits, RL transition kinetics simulator (TKS) and RL low‐energy states sampler (LSS), are implemented and explained in detail, while the meaning of general RL are also discussed. As a testbed, hydrogen diffusivity is computed using RL TKS in pure metals and a medium entropy alloy, CrCoNi, and compared with experiments. The algorithm can produce counter‐intuitive hydrogen‐vacancy cooperative motion. We also demonstrate that RL LSS can accelerate the sampling of low‐energy configurations compared to the Metropolis–Hastings algorithm, using hydrogen migration to copper (111) surface as an example. |
| first_indexed | 2024-03-08T07:04:17Z |
| format | Article |
| id | doaj.art-02e1f30309cf455bb4ea9abd1882640e |
| institution | Directory Open Access Journal |
| issn | 2198-3844 |
| language | English |
| last_indexed | 2024-03-08T07:04:17Z |
| publishDate | 2024-02-01 |
| publisher | Wiley |
| record_format | Article |
| series | Advanced Science |
| spelling | doaj.art-02e1f30309cf455bb4ea9abd1882640e2024-02-03T05:02:44ZengWileyAdvanced Science2198-38442024-02-01115n/an/a10.1002/advs.202304122Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in MetalsHao Tang0Boning Li1Yixuan Song2Mengren Liu3Haowei Xu4Guoqing Wang5Heejung Chung6Ju Li7Department of Materials Science and Engineering Massachusetts Institute of Technology Cambridge MA 02139 USAResearch Laboratory of Electronics Massachusetts Institute of Technology Cambridge MA 02139 USADepartment of Materials Science and Engineering Massachusetts Institute of Technology Cambridge MA 02139 USADepartment of Materials Science and Engineering Massachusetts Institute of Technology Cambridge MA 02139 USADepartment of Nuclear Science and Engineering Massachusetts Institute of Technology Cambridge MA 02139 USAResearch Laboratory of Electronics Massachusetts Institute of Technology Cambridge MA 02139 USADepartment of Materials Science and Engineering Massachusetts Institute of Technology Cambridge MA 02139 USADepartment of Materials Science and Engineering Massachusetts Institute of Technology Cambridge MA 02139 USAAbstract Diffusion in alloys is an important class of atomic processes. However, atomistic simulations of diffusion in chemically complex solids are confronted with the timescale problem: the accessible simulation time is usually far shorter than that of experimental interest. In this work, long‐timescale simulation methods are developed using reinforcement learning (RL) that extends simulation capability to match the duration of experimental interest. Two special limits, RL transition kinetics simulator (TKS) and RL low‐energy states sampler (LSS), are implemented and explained in detail, while the meaning of general RL are also discussed. As a testbed, hydrogen diffusivity is computed using RL TKS in pure metals and a medium entropy alloy, CrCoNi, and compared with experiments. The algorithm can produce counter‐intuitive hydrogen‐vacancy cooperative motion. We also demonstrate that RL LSS can accelerate the sampling of low‐energy configurations compared to the Metropolis–Hastings algorithm, using hydrogen migration to copper (111) surface as an example.https://doi.org/10.1002/advs.202304122hydrogen diffusionlong‐timescale simulationsreinforcement learning |
| spellingShingle | Hao Tang Boning Li Yixuan Song Mengren Liu Haowei Xu Guoqing Wang Heejung Chung Ju Li Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals Advanced Science hydrogen diffusion long‐timescale simulations reinforcement learning |
| title | Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals |
| title_full | Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals |
| title_fullStr | Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals |
| title_full_unstemmed | Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals |
| title_short | Reinforcement Learning‐Guided Long‐Timescale Simulation of Hydrogen Transport in Metals |
| title_sort | reinforcement learning guided long timescale simulation of hydrogen transport in metals |
| topic | hydrogen diffusion long‐timescale simulations reinforcement learning |
| url | https://doi.org/10.1002/advs.202304122 |
| work_keys_str_mv | AT haotang reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT boningli reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT yixuansong reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT mengrenliu reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT haoweixu reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT guoqingwang reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT heejungchung reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals AT juli reinforcementlearningguidedlongtimescalesimulationofhydrogentransportinmetals |