[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-127013-en":3,"doc-seo-127013-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},127013,2336474466412,"Ezra","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Transferability and Accuracy of Ionic Liquid Simulations with Equivariant Machine Learning Interatomic Potentials","Ionic liquids (ILs) are promising electrolytes used in applications ranging from energy storage to solvent-like roles, yet their simulation and design often rely on expensive atomistic methods. This work evaluates whether machine learning interatomic potentials can become transformative for ILs by improving data efficiency and reliability. A compositionally transferable MLIP is demonstrated using limited training on mixtures, and accuracy is assessed for a newly synthesized IL through combined experimental characterization and comparison to DFT using ~200 DFT frames.","arXiv :2403 .01980v3 [physics .chem-ph] 15 Jul 2024  \nTransferability and Accuracy of Ionic Liquid Simulations with Equivariant Machine Learning Interatomic Potentials  \nZachary A. H. Goodwin, ∗ ,† Malia B. Wenny,‡ Julia H. Yang,†,¶ Andrea Cepellotti,† Jingxuan Ding,† Kyle Bystrom,† Blake R. Duschatko,† Anders Johansson,† Lixin Sun,† Simon Batzner,† Albert Musaelian,† Jarad A. Mason,‡ Boris Kozinsky,∗ ,†, §  \nand Nicola Molinari ∗ ,†, §  \n†John A. Paulson School of Engineering and Applied Sciences, Harvard University,  \nCambridge, MA 02138, USA  \n‡Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA  \n02138, USA  \n¶Harvard University Center for the Environment, 26 Oxford St., Cambridge, MA 02138  \n§Robert Bosch LLC, Research and Technology Center, Cambridge, MA 02142, USA  \nE-mail: [zgoodwin@seas.harvard.edu](zgoodwin@seas.harvard.edu) ; [bkoz@seas.harvard.edu](bkoz@seas.harvard.edu) ; [nmolinari@seas.harvard.edu](nmolinari@seas.harvard.edu)  \nAbstract  \nIonic liquids (ILs) are an exciting class of electrolytes finding applications in many areas from energy storage to solvents, where they have been touted as “designer solvents”as they can be mixed to precisely tailor the physiochemical properties. As using machine learning interatomic potentials (MLIPs) to simulate ILs is still relatively unexplored, several questions need to be answered to see if MLIPs can be transformative for ILs. Since ILs are often not pure, but are either mixed together or contain additives, we first demonstrate that a MLIP can be trained to be compositionally transferable, i.e. , the MLIP can be applied to mixtures of ions not directly trained on, whilst only being trained on a few mixtures of the same ions. We also investigate the accuracy of MLIPs for a novel IL, which we experimentally synthesize and characterize. Our MLIP trained on ∼200 DFT frames is in reasonable agreement with our experiments and DFT.  \nIonic liquids (ILs) are a unique and highly promising class of electrolytes that contain no solvent in their neat form. 1–3 This lack of solvent, such as organic carbonates or water, makes ILs advantageous for numerous applications from supercapacitors/batteries, 4–6 to gas storage, 7,8 to themselves acting as “green solvents” for chemical reactions. 1,3 These applications benefit from the use of ILs because ILs have extremely low vapor pressure, are non-flammable and can withstand large voltages without decomposing, 1–4 which are a set of properties that water and organic carbonates do not possess. The unique properties of ILs come, in part, from the large, highly asymmetric and ionic nature of the molecular species which comprise the IL. 1–4 Moreover, ILs can be mixed to tune desired physiochemical properties, creating a class of “designer solvents” with a huge chemical space.9  \nAs ILs are often comprised of large, complicated, molecular ions, this has meant atomistic simulations are necessary for quantitative predictions of these concentrated electrolytes. 4,6,10–12 Typically, classical molecular dynamics (MD) has been used to simulate ILs and calculate physiochemical properties, 4,6,12,13 largely because 1,000+ atoms can routinely be simulated for 1+ ns, which are the length and time scales typically required for ILs. The accuracy of the predictions from classical force fields is often under question, however, 14 which motivates some to use ab initio MD (AIMD) to simulate ILs.6 However, running density functional theory (DFT) is prohibitively expensive in comparison, which limits the simulations to short time scales (∼ 100 ps) and system sizes ( ∼ 100s of atoms) .6  \nMachine learning interatomic potentials (MLIPs) have promised a way to bridge this efficiency-accuracy gap in atomistic simulations. 6,15–19 In the context of ILs, there have been several works which develop MLIPs for specific electrolytes, such as the work of MontesCampos et al. , 20 Dajnowics et al. 14 and Ling et al.21 for room temperature ILs, and t","cbCaib7TGNOlDEDD","https://ap.wps.com/l/cbCaib7TGNOlDEDD","pdf",8801230,1,60,"English","en",105,"# Abstract\n## Transferable MLIP training for IL mixtures\n## Accuracy assessment for a novel IL\n## Comparison with DFT and experimental characterization","[{\"question\":\"What is the main goal of this work on ionic liquid simulations?\",\"answer\":\"To determine how to train compositionally transferable machine learning interatomic potentials (MLIPs) for ionic liquids and to evaluate their accuracy on a novel ionic liquid with limited training data.\"},{\"question\":\"How is compositional transferability achieved for MLIPs in this study?\",\"answer\":\"The study trains an equivariant MLIP to generalize to mixture compositions not directly included in training, showing transferability when training compositions are carefully selected.\"},{\"question\":\"How is the MLIP accuracy validated for the new ionic liquid?\",\"answer\":\"A novel IL is experimentally synthesized and characterized, and the MLIP predictions are compared against experiments and DFT using approximately ~200 DFT frames.\"}]","Transferability and Accuracy of Ionic Liquid Simulations with Equivariant Machine Learning Interatomic Potentials | PDF",1785936323,151,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"transferability-and-accuracy-of-ionic-liquid-simulations-with-equivariant-machine-learning-interatomic-potentials","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/transferability-and-accuracy-of-ionic-liquid-simulations-with-equivariant-machine-learning-interatomic-potentials/127013/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the main goal of this work on ionic liquid simulations?","Question",{"text":75,"@type":76},"To determine how to train compositionally transferable machine learning interatomic potentials (MLIPs) for ionic liquids and to evaluate their accuracy on a novel ionic liquid with limited training data.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is compositional transferability achieved for MLIPs in this study?",{"text":80,"@type":76},"The study trains an equivariant MLIP to generalize to mixture compositions not directly included in training, showing transferability when training compositions are carefully selected.",{"name":82,"@type":73,"acceptedAnswer":83},"How is the MLIP accuracy validated for the new ionic liquid?",{"text":84,"@type":76},"A novel IL is experimentally synthesized and characterized, and the MLIP predictions are compared against experiments and DFT using approximately ~200 DFT frames.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,109,114,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":21,"slug":108},5,"Comic","comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]