Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials
Abstract Mechanical behavior of 2D materials such as MoS2 can be tuned by the ancient art of kirigami. Experiments and atomistic simulations show that 2D materials can be stretched more than 50% by strategic insertion of cuts. However, designing kirigami structures with desired mechanical properties...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Publishing Group
2021-07-01
|
Series: | npj Computational Materials |
Online Access: | https://doi.org/10.1038/s41524-021-00572-y |
id |
doaj-677801e86c9c4479ada8ce35b66038ce |
---|---|
record_format |
Article |
spelling |
doaj-677801e86c9c4479ada8ce35b66038ce2021-07-11T11:18:12ZengNature Publishing Groupnpj Computational Materials2057-39602021-07-01711810.1038/s41524-021-00572-yAutonomous reinforcement learning agent for stretchable kirigami design of 2D materialsPankaj Rajak0Beibei Wang1Ken-ichi Nomura2Ye Luo3Aiichiro Nakano4Rajiv Kalia5Priya Vashishta6Argonne Leadership Computing Facility, Argonne National LaboratoryCollaboratory for Advanced Computing and Simulations, University of Southern CaliforniaCollaboratory for Advanced Computing and Simulations, University of Southern CaliforniaArgonne Leadership Computing Facility, Argonne National LaboratoryCollaboratory for Advanced Computing and Simulations, University of Southern CaliforniaCollaboratory for Advanced Computing and Simulations, University of Southern CaliforniaCollaboratory for Advanced Computing and Simulations, University of Southern CaliforniaAbstract Mechanical behavior of 2D materials such as MoS2 can be tuned by the ancient art of kirigami. Experiments and atomistic simulations show that 2D materials can be stretched more than 50% by strategic insertion of cuts. However, designing kirigami structures with desired mechanical properties is highly sensitive to the pattern and location of kirigami cuts. We use reinforcement learning (RL) to generate a wide range of highly stretchable MoS2 kirigami structures. The RL agent is trained by a small fraction (1.45%) of molecular dynamics simulation data, randomly sampled from a search space of over 4 million candidates for MoS2 kirigami structures with 6 cuts. After training, the RL agent not only proposes 6-cut kirigami structures that have stretchability above 45%, but also gains mechanistic insight to propose highly stretchable (above 40%) kirigami structures consisting of 8 and 10 cuts from a search space of billion candidates as zero-shot predictions.https://doi.org/10.1038/s41524-021-00572-y |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Pankaj Rajak Beibei Wang Ken-ichi Nomura Ye Luo Aiichiro Nakano Rajiv Kalia Priya Vashishta |
spellingShingle |
Pankaj Rajak Beibei Wang Ken-ichi Nomura Ye Luo Aiichiro Nakano Rajiv Kalia Priya Vashishta Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials npj Computational Materials |
author_facet |
Pankaj Rajak Beibei Wang Ken-ichi Nomura Ye Luo Aiichiro Nakano Rajiv Kalia Priya Vashishta |
author_sort |
Pankaj Rajak |
title |
Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials |
title_short |
Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials |
title_full |
Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials |
title_fullStr |
Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials |
title_full_unstemmed |
Autonomous reinforcement learning agent for stretchable kirigami design of 2D materials |
title_sort |
autonomous reinforcement learning agent for stretchable kirigami design of 2d materials |
publisher |
Nature Publishing Group |
series |
npj Computational Materials |
issn |
2057-3960 |
publishDate |
2021-07-01 |
description |
Abstract Mechanical behavior of 2D materials such as MoS2 can be tuned by the ancient art of kirigami. Experiments and atomistic simulations show that 2D materials can be stretched more than 50% by strategic insertion of cuts. However, designing kirigami structures with desired mechanical properties is highly sensitive to the pattern and location of kirigami cuts. We use reinforcement learning (RL) to generate a wide range of highly stretchable MoS2 kirigami structures. The RL agent is trained by a small fraction (1.45%) of molecular dynamics simulation data, randomly sampled from a search space of over 4 million candidates for MoS2 kirigami structures with 6 cuts. After training, the RL agent not only proposes 6-cut kirigami structures that have stretchability above 45%, but also gains mechanistic insight to propose highly stretchable (above 40%) kirigami structures consisting of 8 and 10 cuts from a search space of billion candidates as zero-shot predictions. |
url |
https://doi.org/10.1038/s41524-021-00572-y |
work_keys_str_mv |
AT pankajrajak autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials AT beibeiwang autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials AT kenichinomura autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials AT yeluo autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials AT aiichironakano autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials AT rajivkalia autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials AT priyavashishta autonomousreinforcementlearningagentforstretchablekirigamidesignof2dmaterials |
_version_ |
1721309121238532096 |