[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122906-en":3,"doc-seo-122906-105":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},122906,137441390410,"Hazel","https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984",8,"Research & Report","Modeling Chemical Processes in Explicit Solvents with Machine Learning Potentials","Solvent effects govern chemical reactivity by stabilizing intermediates and transition states, reshaping reaction-rate behavior, and altering product distributions, yet they remain difficult to model with adequate accuracy. A general strategy is presented for building reactive machine learning potentials for solution-phase processes, combining active learning with descriptor-based selection and automated workflows to obtain data-efficient training sets across relevant chemical and conformational space. The method is demonstrated on a Diels-Alder reaction in water and methanol, yielding MLP predictions consistent with experimental results and explaining solvent-dependent rate differences.","Modeling Chemical Processes in Explicit Solvents with Machine Learning Potentials  \nHanwen Zhang, Veronika Juraskova, and Fernanda Duarte ∗ Chemistry Research Laboratory, 12 Mansfield Road, Oxford, OX1 3TA  \nE-mail: [fernanda.duartegonzalez@chem.ox.ac.uk](fernanda.duartegonzalez@chem.ox.ac.uk)  \nAbstract  \nSolvent effects influence all stages of the chemical processes, modulating the stability of intermediates and transition states, as well as altering reaction rates and product ratios. However, accurately modelling these effects remains challenging. Here, we present a general strategy for generating reactive machine learning potentials (MLPs) to model chemical processes in solution. Our approach combines active learning with descriptor-based selectors and automation, enabling the construction of data-efficient training sets that span the relevant chemical and conformational space. We demonstrate the versatility of this strategy by applying it to investigate a Diels-Alder reaction in water and methanol. The generated MLPs exhibit excellent agreement with experimental data and provide insights into the differences in reaction rates observed between the two solvents. Our strategy offers an efficient approach to the routine modelling of chemical reactions in solution, opening up avenues for studying complex chemical processes in an efficient manner.  \n1  \n[https://doi.org/10.26434/chemrxiv-2023-ktscq](https://doi.org/10.26434/chemrxiv-2023-ktscq ORCID:)[ ORCID:](https://doi.org/10.26434/chemrxiv-2023-ktscq ORCID:) [https://orcid.org/0000-0002-6062-8209 Content](https://orcid.org/0000-0002-6062-8209 Content) not peer-reviewed by ChemRxiv. License: CC BY 4.0  \nIntroduction  \n(Bio)chemical and industrially relevant reactions occur predominantly in the liquid phase. 1 ,2 The influence of solvent on reaction rates was first documented by Berthelot and Pean de Saint-Gilles in 1862 3 and later formalised by Menschutkin in 1890 .4 Since then, experimental and computational techniques have been developed to elucidate and quantify the origin of solvent effects on reactivity and selectivity.5 From an atomistic point of view, solvent effects arise from the interactions between solute and solvent molecules, which, although generally weak, have a significant impact on the overall reaction dynamics.  \nIn computational chemistry, solvent effects are modelled using implicit or explicit solvent models. The former represents solvents as a polarizable continuum, offering computational simplicity and efficiency. However, it fails to capture the contributions arising from solutesolvent interactions, including entropy and pre-organisation effects. 6 On the other hand, explicit solvent models provide an atomistic representation of the solvent but at a much higher computational cost since they require the use of ab initio molecular dynamics (AIMD) . This cost becomes significant when attempting to compute free energies, where extensive sampling is required to obtain statistically meaningful ensembles. Hybrid approaches, such as quantum mechanics/molecular mechanics (QM/MM), can alleviate this computational cost by describing only the reactive part at a QM level while the environment is described classically. However, the inclusion of mobile solvent molecules into the QM part creates additional technical difficulties, such as discontinuities at the boundary between the QM and MM regions.7 ,8  \nIn recent years, machine learning-based potentials (MLPs) have emerged as powerful surrogates for existing modelling techniques, representing complex potential energy surfaces (PES) with an accuracy comparable to QM methods but at a significantly lower computational cost.9 , 10 Since the pioneering work of Behler and Parinello (BP) on neural network-based potentials (NNP), 9 several MLP approaches have been developed, including models directly derived from BP, such as DeepMD 11 and ANI, 12 message-passing NN models, 13 such as  \n2  \n[https://doi.org/10.26434/chemrxi","cbCais30ZMwcXCsl","https://ap.wps.com/l/cbCais30ZMwcXCsl","pdf",4733954,1,32,"English","en",105,"# Abstract\n# Introduction\n## Solvent effects in chemical processes\n## Implicit vs explicit solvent models\n## Machine learning potentials for potential energy surfaces\n## Challenges in training data for solution-phase reactions","[{\"question\":\"Why are solvent effects important in modelling chemical reactions?\",\"answer\":\"Solvent interactions influence intermediate and transition-state stability, which changes reaction rates and product ratios. These effects also stem from solute–solvent interactions that impact overall reaction dynamics.\"},{\"question\":\"What is the difference between implicit and explicit solvent models?\",\"answer\":\"Implicit models treat the solvent as a polarizable continuum, improving efficiency but missing detailed solute–solvent interaction contributions. Explicit models represent solvent atoms directly but are much more computationally expensive, especially for free-energy calculations requiring extensive sampling.\"},{\"question\":\"How does the proposed machine learning strategy address training-data challenges?\",\"answer\":\"It uses active learning with descriptor-based selectors and automation to build data-efficient training sets. This helps cover the relevant chemical and conformational space needed to capture both minima and transition-state regions in solution.\"}]","Modeling Chemical Processes in Explicit Solvents with Machine Learning Potentials | PDF",1785813594,81,{"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},"modeling-chemical-processes-in-explicit-solvents-with-machine-learning-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/modeling-chemical-processes-in-explicit-solvents-with-machine-learning-potentials/122906/",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","2026-08-04",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},"Why are solvent effects important in modelling chemical reactions?","Question",{"text":75,"@type":76},"Solvent interactions influence intermediate and transition-state stability, which changes reaction rates and product ratios. These effects also stem from solute–solvent interactions that impact overall reaction dynamics.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the difference between implicit and explicit solvent models?",{"text":80,"@type":76},"Implicit models treat the solvent as a polarizable continuum, improving efficiency but missing detailed solute–solvent interaction contributions. Explicit models represent solvent atoms directly but are much more computationally expensive, especially for free-energy calculations requiring extensive sampling.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the proposed machine learning strategy address training-data challenges?",{"text":84,"@type":76},"It uses active learning with descriptor-based selectors and automation to build data-efficient training sets. This helps cover the relevant chemical and conformational space needed to capture both minima and transition-state regions in solution.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]