Cadmium selenide(CdSe)is an inorganic semiconductor with unique optical and electronic properties that make it useful in various applications,including solar cells,light-emitting diodes,and biofluorescent tagging.In o...Cadmium selenide(CdSe)is an inorganic semiconductor with unique optical and electronic properties that make it useful in various applications,including solar cells,light-emitting diodes,and biofluorescent tagging.In order to synthesize high-quality crystals and subsequently integrate them into devices,it is crucial to understand the atomic scale crystallization mechanism of CdSe.Unfortunately,such studies are still absent in the literature.To overcome this limitation,we employed an enhanced sampling-accelerated active learning approach to construct a deep neural potential with ab initio accuracy for studying the crystallization of CdSe.Our brute-force molecular dynamics simulations revealed that a spherical-like nu-cleus formed spontaneously and stochastically,resulting in a stacking disordered structure where the competition between hexagonal wurtzite and cubic zinc blende polymorphs is temperature-dependent.We found that pure hexagonal crystal can only be obtained approximately above 1430 K,which is 35 K below its melting temperature.Furthermore,we observed that the solidification dynamics of Cd and Se atoms were distinct due to their different diffusion coefficients.The solidification process was initiated by lower mobile Se atoms forming tetrahedral frameworks,followed by Cd atoms occupying these tetra-hedral centers and settling down until the third-shell neighbor of Se atoms sited on their lattice posi-tions.Therefore,the medium-range ordering of Se atoms governs the crystallization process of CdSe.Our findings indicate that understanding the complex dynamical process is the key to comprehending the crystallization mechanism of compounds like CdSe,and can shed lights in the synthesis of high-quality crystals.展开更多
Training machine learning interatomic potentials that are both computationally and data-efficient is a key challenge for enabling their routine use in atomistic simulations.To this effect,we introduce franken,a scalab...Training machine learning interatomic potentials that are both computationally and data-efficient is a key challenge for enabling their routine use in atomistic simulations.To this effect,we introduce franken,a scalable and lightweight transfer learning framework that extracts atomic descriptors from pre-trained graph neural networks and transfers them to new systems using random Fourier features—an efficient and scalable approximation of kernel methods.It also provides a closed-form finetuning strategy for general-purpose potentials such as MACE-MP0,enabling fast and accurate adaptation to new systems or levels of quantum mechanical theory with minimal hyperparameter tuning.On a benchmark dataset of 27 transition metals,franken outperforms optimized kernelbased methods in both training time and accuracy,reducing model training from tens of hours to minutes on a single GPU.We further demonstrate the framework’s strong data-efficiency by training stable and accurate potentials for bulk water and the Pt(111)/water interface using just tens of training structures.Our open-source implementation(https://franken.readthedocs.io)offers a fast and practical solution for training potentials and deploying them for molecular dynamics simulations across diverse systems.展开更多
The phase change compound Ge_(2)Sb_(2)Te_(5)(GST225)is exploited in advanced non-volatile electronic memories and in neuromorphic devices which both rely on a fast and reversible transition between the crystalline and...The phase change compound Ge_(2)Sb_(2)Te_(5)(GST225)is exploited in advanced non-volatile electronic memories and in neuromorphic devices which both rely on a fast and reversible transition between the crystalline and amorphous phases induced by Joule heating.The crystallization kinetics of GST225 is a key functional feature for the operation of these devices.We report here on the development of a machine-learned interatomic potential for GST225 that allowed us to perform large scale molecular dynamics simulations(over 10,000 atoms for over 100 ns)to uncover the details of the crystallization kinetics in a wide range of temperatures of interest for the programming of the devices.The potential is obtained by fitting with a deep neural network(NN)scheme a large quantum-mechanical database generated within density functional theory.The availability of a highly efficient and yet highly accurate NN potential opens the possibility to simulate phase change materials at the length and time scales of the real devices.展开更多
The prediction of material properties based on density-functional theory has become routinely common,thanks,in part,to the steady increase in the number and robustness of available simulation packages.This plurality o...The prediction of material properties based on density-functional theory has become routinely common,thanks,in part,to the steady increase in the number and robustness of available simulation packages.This plurality of codes and methods is both a boon and a burden.While providing great opportunities for cross-verification,these packages adopt different methods,algorithms,and paradigms,making it challenging to choose,master,and efficiently use them.We demonstrate how developing common interfaces for workflows that automatically compute material properties greatly simplifies interoperability and cross-verification.We introduce design rules for reusable,code-agnostic,workflow interfaces to compute well-defined material properties,which we implement for eleven quantum engines and use to compute various material properties.Each implementation encodes carefully selected simulation parameters and workflow logic,making the implementer’s expertise of the quantum engine directly available to nonexperts.All workflows are made available as open-source and full reproducibility of the workflows is guaranteed through the use of the AiiDA infrastructure.展开更多
The rapid advancements in artificial intelligence(AI)are catalyzing transformative changes in atomic modeling,simulation,and design.AI-driven potential energy models havedemonstrated the capability to conduct large-sc...The rapid advancements in artificial intelligence(AI)are catalyzing transformative changes in atomic modeling,simulation,and design.AI-driven potential energy models havedemonstrated the capability to conduct large-scale,long-duration simulations with the accuracy of ab initio electronic structure methods.However,the model generation process remains a bottleneck for large-scale applications.We propose a shift towards a model-centric ecosystem,wherein a large atomic model(LAM),pretrained across multiple disciplines,can be efficiently fine-tuned and distilled for various downstream tasks,thereby establishing a new framework for molecular modeling.In this study,we introduce the DPA-2 architecture as a prototype for LAMs.Pre-trained on a diverse array of chemical and materials systemsusing a multi-task approach,DPA-2demonstrates superior generalization capabilities across multiple downstream tasks compared to the traditional single-task pre-training and fine-tuning methodologies.Our approach sets the stage for the development and broad application of LAMs in molecular and materials simulation research.展开更多
Simulating catalytic reactivity under operative conditions poses a significant challenge due to the dynamic nature of the catalysts and the high computational cost of electronic structure calculations.Machine learning...Simulating catalytic reactivity under operative conditions poses a significant challenge due to the dynamic nature of the catalysts and the high computational cost of electronic structure calculations.Machine learning potentials offer a promising avenue to simulate dynamics at a fraction of the cost,but they require datasets containing all relevant configurations,particularly reactive ones.Here,we present a scheme to construct reactive potentials in a data-efficient manner.This is achieved by combining enhanced sampling methods first with Gaussian processes to discover transition paths and then with graph neural networks to obtain a uniformly accurate description.The necessary configurations are extracted via a Data-Efficient Active Learning(DEAL)procedure based on local environment uncertainty.We validated our approach by studying several reactions related to the decomposition ofammonia on iron-cobalt alloy catalysts.Our schemeproved to be efficient,requiring only~1000 DFT calculations per reaction,and robust,sampling reactive configurations from the different accessible pathways.Using this potential,we calculated free energy profiles and characterized reaction mechanisms,showing the ability to provide microscopic insights into complex processes under dynamic conditions.展开更多
基金supported by the National Natural Science Foundation of China(No.92370118)the National Science Fund for Excellent Young Scientist Fund Program(Overseas)of China,the Science and Technology Activities Fund for Overseas Researchers of Shaanxi Province,China,and the Research Fund of the State Key Laboratory of Solidification Proceeding(NPU)of China(No.2020-QZ-03).
文摘Cadmium selenide(CdSe)is an inorganic semiconductor with unique optical and electronic properties that make it useful in various applications,including solar cells,light-emitting diodes,and biofluorescent tagging.In order to synthesize high-quality crystals and subsequently integrate them into devices,it is crucial to understand the atomic scale crystallization mechanism of CdSe.Unfortunately,such studies are still absent in the literature.To overcome this limitation,we employed an enhanced sampling-accelerated active learning approach to construct a deep neural potential with ab initio accuracy for studying the crystallization of CdSe.Our brute-force molecular dynamics simulations revealed that a spherical-like nu-cleus formed spontaneously and stochastically,resulting in a stacking disordered structure where the competition between hexagonal wurtzite and cubic zinc blende polymorphs is temperature-dependent.We found that pure hexagonal crystal can only be obtained approximately above 1430 K,which is 35 K below its melting temperature.Furthermore,we observed that the solidification dynamics of Cd and Se atoms were distinct due to their different diffusion coefficients.The solidification process was initiated by lower mobile Se atoms forming tetrahedral frameworks,followed by Cd atoms occupying these tetra-hedral centers and settling down until the third-shell neighbor of Se atoms sited on their lattice posi-tions.Therefore,the medium-range ordering of Se atoms governs the crystallization process of CdSe.Our findings indicate that understanding the complex dynamical process is the key to comprehending the crystallization mechanism of compounds like CdSe,and can shed lights in the synthesis of high-quality crystals.
基金support of the Data Science and Computation Facility at the Fondazione Istituto Italiano di Tecnologia and the CINECA award under the ISCRA initiativeThis work was partially funded by the European Union—NextGenerationEU initiative and the Italian National Recovery and Resilience Plan (PNRR) from the Ministry of University and Research (MUR), under Project PE0000013 CUP J53C22003010006 “Future Artificial Intelligence Research (FAIR)”+3 种基金L.R. acknowledges the financial support of the European Research Council (grant SLING 819789)the European Commission (Horizon Europe grant ELIAS 101120237)the Ministry of Education, University and Research (FARE grant ML4IP R205T7J2KPgrant BAC FAIR PE00000013 CUP J33C24000420007 funded by the EU—NGEU).
文摘Training machine learning interatomic potentials that are both computationally and data-efficient is a key challenge for enabling their routine use in atomistic simulations.To this effect,we introduce franken,a scalable and lightweight transfer learning framework that extracts atomic descriptors from pre-trained graph neural networks and transfers them to new systems using random Fourier features—an efficient and scalable approximation of kernel methods.It also provides a closed-form finetuning strategy for general-purpose potentials such as MACE-MP0,enabling fast and accurate adaptation to new systems or levels of quantum mechanical theory with minimal hyperparameter tuning.On a benchmark dataset of 27 transition metals,franken outperforms optimized kernelbased methods in both training time and accuracy,reducing model training from tens of hours to minutes on a single GPU.We further demonstrate the framework’s strong data-efficiency by training stable and accurate potentials for bulk water and the Pt(111)/water interface using just tens of training structures.Our open-source implementation(https://franken.readthedocs.io)offers a fast and practical solution for training potentials and deploying them for molecular dynamics simulations across diverse systems.
基金funding from European Union NextGenerationEU through the Italian Ministry of University and Research under PNRR M4C2I1.4 ICSC Centro Nazionale di Ricerca in High Performance Computing,Big Data and Quantum Computing(Grant No.CN00000013)。
文摘The phase change compound Ge_(2)Sb_(2)Te_(5)(GST225)is exploited in advanced non-volatile electronic memories and in neuromorphic devices which both rely on a fast and reversible transition between the crystalline and amorphous phases induced by Joule heating.The crystallization kinetics of GST225 is a key functional feature for the operation of these devices.We report here on the development of a machine-learned interatomic potential for GST225 that allowed us to perform large scale molecular dynamics simulations(over 10,000 atoms for over 100 ns)to uncover the details of the crystallization kinetics in a wide range of temperatures of interest for the programming of the devices.The potential is obtained by fitting with a deep neural network(NN)scheme a large quantum-mechanical database generated within density functional theory.The availability of a highly efficient and yet highly accurate NN potential opens the possibility to simulate phase change materials at the length and time scales of the real devices.
基金This work is supported by the MARVEL National Centre of Competence in Research(NCCR)funded by the Swiss National Science Foundation(grant agreement ID 51NF40-182892)by the European Union’s Horizon 2020 research and innovation program under Grant Agreement No.824143(European MaX Centre of Excellence“Materials design at the Exascale”)and Grant Agreement No.814487(INTERSECT project).We thank M.Giantomassi and J.-M.Beuken for their contributions in adding support for PseudoDojo tables to the aiida-pseudo(https://github.com/aiidateam/aiida-pseudo)plugin.We also thank X.Gonze,M.Giantomassi,M.Probert,C.Pickard,P.Hasnip,J.Hutter,M.Iannuzzi,D.Wortmann,S.Blügel,J.Hess,F.Neese,and P.Delugas for providing useful feedback on the various quantum engine implementations.S.P.acknowledges support from the European Unions Horizon 2020 Research and Innovation Programme,under the Marie Skłodowska-Curie Grant Agreement SELPH2D No.839217 and computer time provided by the PRACE-21 resources MareNostrum at BSC-CNS+6 种基金E.F.-L.acknowledges the support of the Norwegian Research Council(project number 262339)and computational resources provided by Sigma2P.Z.-P.thanks to the Faraday Institution CATMAT project(EP/S003053/1,FIRG016) for financial supportKE acknowledges the Swiss National Science Foundation(grant number 200020-182015)G.Pi.and K.E.acknowledge the swissuniversities“Materials Cloud”(project number 201-003).Work at ICMAB is supported by the Severo Ochoa Centers of Excellence Program(MICINN CEX2019-000917-S)by PGC2018-096955-B-C44(MCIU/AEI/FEDER,UE),and by GenCat 2017SGR1506B.Z.thanks to the Faraday Institution FutureCat project(EP/S003053/1,FIRG017) for financial supportJ.B.and V.T.acknowledge support by the Joint Lab Virtual Materials Design(JLVMD)of the Forschungszentrum Jülich.
文摘The prediction of material properties based on density-functional theory has become routinely common,thanks,in part,to the steady increase in the number and robustness of available simulation packages.This plurality of codes and methods is both a boon and a burden.While providing great opportunities for cross-verification,these packages adopt different methods,algorithms,and paradigms,making it challenging to choose,master,and efficiently use them.We demonstrate how developing common interfaces for workflows that automatically compute material properties greatly simplifies interoperability and cross-verification.We introduce design rules for reusable,code-agnostic,workflow interfaces to compute well-defined material properties,which we implement for eleven quantum engines and use to compute various material properties.Each implementation encodes carefully selected simulation parameters and workflow logic,making the implementer’s expertise of the quantum engine directly available to nonexperts.All workflows are made available as open-source and full reproducibility of the workflows is guaranteed through the use of the AiiDA infrastructure.
基金supported by the National Key R&D Program of China(grantno.2022YFA1004300)the National Natural Science Foundation of China(grant no.12122103)+11 种基金supported by the National Key Research and Development Project of China(grant no.2022YFA1004302)the National Natural Science Foundation of China(grants nos.92270001 and 12288101)supported by the National Institutes of Health(grant no.GM107485 to D.M.Y.)the National Science Foundation(grant no.2209718 to D.M.Y.)supported by the Natural Science Foundation of Zhejiang Province(grant no.2022XHSJJ006)supported by the National Natural Science Foundation of China(grants nos.22222303 and 22173032)supported by the National Key R&D Program of China(grants nos.2021YFA0718900 and 2022YFA1403000)supported by the National Natural Science Foundation of China(grants nos.12034009 and 91961204)supported by the National Science Fund for Distinguished Young Scholars(grant no.22225302)Laboratory of AI for Electrochemistry(AI4EC),and IKKEM(grants nos.RD2023100101 and RD2022070501)supported by the National Natural Science Foundation of China(grants nos.12122401,12074007,and 12135002)Lastly,the computational resource was supported by the Bohrium Cloud Platform at DP Technology and Tan Kah Kee Supercomputing Center(IKKEM).
文摘The rapid advancements in artificial intelligence(AI)are catalyzing transformative changes in atomic modeling,simulation,and design.AI-driven potential energy models havedemonstrated the capability to conduct large-scale,long-duration simulations with the accuracy of ab initio electronic structure methods.However,the model generation process remains a bottleneck for large-scale applications.We propose a shift towards a model-centric ecosystem,wherein a large atomic model(LAM),pretrained across multiple disciplines,can be efficiently fine-tuned and distilled for various downstream tasks,thereby establishing a new framework for molecular modeling.In this study,we introduce the DPA-2 architecture as a prototype for LAMs.Pre-trained on a diverse array of chemical and materials systemsusing a multi-task approach,DPA-2demonstrates superior generalization capabilities across multiple downstream tasks compared to the traditional single-task pre-training and fine-tuning methodologies.Our approach sets the stage for the development and broad application of LAMs in molecular and materials simulation research.
基金We acknowledge support from the Data Science and Computation Facility and its Support Team at Fondazione Istituto Italiano di Tecnologia,the CINECA award under the ISCRA initiative(IscraB28_AmmoFeCo)the Max Planck Computing and Data Facilitythe Federal Ministry of Education and Research,Germany(Bundesministerium für Bildung und Forschung,BMBF,Hydrogen flagship project:TransHyDE Forschungsverbund AmmoRef,FKZ 03HY203A)for funding.
文摘Simulating catalytic reactivity under operative conditions poses a significant challenge due to the dynamic nature of the catalysts and the high computational cost of electronic structure calculations.Machine learning potentials offer a promising avenue to simulate dynamics at a fraction of the cost,but they require datasets containing all relevant configurations,particularly reactive ones.Here,we present a scheme to construct reactive potentials in a data-efficient manner.This is achieved by combining enhanced sampling methods first with Gaussian processes to discover transition paths and then with graph neural networks to obtain a uniformly accurate description.The necessary configurations are extracted via a Data-Efficient Active Learning(DEAL)procedure based on local environment uncertainty.We validated our approach by studying several reactions related to the decomposition ofammonia on iron-cobalt alloy catalysts.Our schemeproved to be efficient,requiring only~1000 DFT calculations per reaction,and robust,sampling reactive configurations from the different accessible pathways.Using this potential,we calculated free energy profiles and characterized reaction mechanisms,showing the ability to provide microscopic insights into complex processes under dynamic conditions.
