[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121490-en":3,"doc-seo-121490-105":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":20,"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},121490,1374391974585,"Genevieve","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Machine Learning the Energetics of Electrified Solid-Liquid Interfaces - Response-Augmented Approach","A response-augmented machine-learning framework is developed to model the energetics of electrified metal surfaces. Local descriptors are used to learn the work function as the first-order energy change induced by bias charges, with Born effective charges stabilizing the learning. The resulting extension of machine-learning interatomic potential architectures incorporates finite-bias effects up to second order. Application to OH on Cu(100) explains the experimentally observed pH dependence of the preferred adsorption site via non-Nernstian charge-induced site switching.","Editors' Suggestion  \nMachine Learning the Energetics of Electrified Solid-Liquid Interfaces  \nNicolas Bergmann, 1 Nicphore Bonnet,2 Nicola Marzari,2 Karsten Reuter, 1 and Nicolas G. Hörmann1,*  \n1Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin, Germany 2Theory and Simulation of Materials (THEOS), Ecole Polytechnique Fe´de´rale de Lausanne (EPFL), Lausanne 1015, Switzerland  (Received 13 May 2025; revised 9 August 2025; accepted 26 August 2025; published 29 September 2025)  \nWe present a response-augmented machine-learning (ML) approach to the energetics of electrified metal surfaces. We leverage local descriptors to learn the work function as the first-order energy change to introduced bias charges and stabilize this learning through Born effective charges. This permits the efficient extension of ML interatomic potential architectures to include finite bias effects up to second order. Application to OH at Cu(100) rationalizes the experimentally observed pH dependence of the preferred adsorption site in terms of a non-Nernstian charge-induced site switching.  \nDOI: 10. 1103/lm64-m3bn  \nPredictive-quality first-principles modeling and simulation has become indispensable for studying electrochemical interfaces. By accessing the detailed atomic structure and prevailing interactions, it provides deep mechanistic insight and generates ideas for the design of improved electrocatalysts [1–4] . Compared to its analog usage at solid-gas interfaces and thermal catalysis, first-principles modeling of electrified solid-liquid interfaces nevertheless still lags behind [5], largely because of difficulties to describe the extended double layer (DL) that builds up at biased conditions. Even within common approximations like implicit solvation [6–9], the description of complex and potential-dependent (near-surface) dynamics remains a challenging computational burden.  \nIn this respect, application ofmachine-learning interatomic potentials (MLIPs) [10–12] as fast surrogates to the firstprinciples calculations is highly appealing. MLIPs access longer time and larger length scales, a prerequisite to derive reliable macroscopic insights from the atomistic interfacial structures and reactions[13–16]. Notwithstanding, thenecessity to accurately capture the local fields that build up at the electrified interface prevents the straightforward usage of established short-range MLIP architectures [17]. To this end, an important strand of ongoing developments centers on incorporating long-range electrostatic interactions [18–25] . While this has led to a successful modeling of the atomistic structure of the interface at the potential of zero charge [26], there are still challenges at applied electrode potential  \n*Contact author: [hoermann@fhi.mpg.de](hoermann@fhi.mpg.de)  \nPublished by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Open access publication funded by the Max Planck Society.  \nconditions. For instance, underlying charge equilibration schemes can suffer from unphysical, partially metallic behavior of the electrolyte, leading to overpolarization in the presence of electric fields [27], while the need to solve for electrostatic potentials and atomic charges self-consistently generally increases the computational costs, thereby reducing the very efficiency gain sought with MLIPs. Explicit training of short-range MLIPs for different charge states in turn requires larger amounts of training data and is typically not size extensive [19,28] .  \nAs such, a complementary strand of developments appears promising: models incorporating the effects of electric fields by machine-learning relevant response effects. With previous works addressing isolated molecules [29–33], bulk solids [34–36], or liquid water [37–40], we here t","cbCaiqjwiQJt8ACN","https://ap.wps.com/l/cbCaiqjwiQJt8ACN","pdf",1942100,1,7,"English","en",105,"# Introduction\n## Predictive modeling needs for electrified electrochemical interfaces\n## Limitations of first-principles and implicit solvation\n# Machine-learning interatomic potentials with finite bias effects\n## Challenges with local electrostatic fields\n## Long-range electrostatics and charge equilibration issues\n# Response Analysis in z-Orientation (RAZOR)\n## Learning energy and force derivatives with respect to bias charge\n## Implicit solvent referencing of extended double layer\n# Demonstration on OH/Cu(100)\n## Potential-dependent adsorption-site switching\n## Connection to non-Nernstian behavior","[{\"question\":\"What is the core idea of the response-augmented machine-learning approach?\",\"answer\":\"The method learns how energetics change under introduced bias charges by using local descriptors to model the work function as a first-order energy change, with stabilization via Born effective charges.\"},{\"question\":\"How far can the presented ML interatomic potential extension capture finite-bias effects?\",\"answer\":\"It efficiently extends common MLIP architectures to include finite bias effects up to second order.\"},{\"question\":\"How does the approach explain the pH dependence of the preferred adsorption site for OH on Cu(100)?\",\"answer\":\"Simulations rationalize the experimentally observed site preference as a non-Nernstian charge-induced switching effect that depends on the interfacial potential.\"}]","Machine Learning the Energetics of Electrified Solid-Liquid Interfaces - Response-Augmented Approach | PDF",1785735907,18,{"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},"machine-learning-the-energetics-of-electrified-solid-liquid-interfaces-response-augmented-approach","",{"@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/machine-learning-the-energetics-of-electrified-solid-liquid-interfaces-response-augmented-approach/121490/",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","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the core idea of the response-augmented machine-learning approach?","Question",{"text":75,"@type":76},"The method learns how energetics change under introduced bias charges by using local descriptors to model the work function as a first-order energy change, with stabilization via Born effective charges.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How far can the presented ML interatomic potential extension capture finite-bias effects?",{"text":80,"@type":76},"It efficiently extends common MLIP architectures to include finite bias effects up to second order.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the approach explain the pH dependence of the preferred adsorption site for OH on Cu(100)?",{"text":84,"@type":76},"Simulations rationalize the experimentally observed site preference as a non-Nernstian charge-induced switching effect that depends on the interfacial potential.","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},[93,97,101,105,110,115,119,122,127,130,134],{"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":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},"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"]