Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells

The photoelectric conversion efficiency (PCE) of perovskites remains beneath the Shockley-Queisser limit, despite its significant potential for solar cell applications. The present focus is on investigating potential multicomponent perovskite candidates, particularly on the application of machine le...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Zhaosheng Zhang (4603021) (author)
مؤلفون آخرون: Sijia Liu (1751842) (author), Qing Xiong (136442) (author), Yanbo Liu (180774) (author)
منشور في: 2025
الموضوعات:
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
_version_ 1852023664627679232
author Zhaosheng Zhang (4603021)
author2 Sijia Liu (1751842)
Qing Xiong (136442)
Yanbo Liu (180774)
author2_role author
author
author
author_facet Zhaosheng Zhang (4603021)
Sijia Liu (1751842)
Qing Xiong (136442)
Yanbo Liu (180774)
author_role author
dc.creator.none.fl_str_mv Zhaosheng Zhang (4603021)
Sijia Liu (1751842)
Qing Xiong (136442)
Yanbo Liu (180774)
dc.date.none.fl_str_mv 2025-01-13T06:58:10Z
dc.identifier.none.fl_str_mv 10.1021/acs.jpclett.4c03580.s002
dc.relation.none.fl_str_mv https://figshare.com/articles/dataset/Strategic_Integration_of_Machine_Learning_in_the_Design_of_Excellent_Hybrid_Perovskite_Solar_Cells/28193341
dc.rights.none.fl_str_mv CC BY-NC 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Cell Biology
Cancer
Science Policy
Space Science
Biological Sciences not elsewhere classified
Chemical Sciences not elsewhere classified
Information Systems not elsewhere classified
vgg16 surpassed xception
solar cell applications
selecting solar materials
random forests lagged
photoelectric conversion efficiency
including decision trees
graph neural networks
exceptional photoelectric properties
convolutional neural networks
centered symmetry functions
body tensor representation
least favorable performance
perovskites remains beneath
ewald sum matrix
8 </ sub
3 </ sub
2 </ sup
2 </ sub
decision tree models
cnn models utilizing
throughput screening method
substitutional defects compared
machine learning establishes
exhibited superior performance
efficiently identify high
performance perovskites
throughput screening
machine learning
r </
sine matrix
>< sup
gcsconv exhibited
efficientnetv2b0 exhibited
gnn models
strategic integration
slight advantage
significant potential
robust foundation
results indicated
queisser limit
present focus
ideal combination
gnns ).
gcsconv outperformed
data set
considerable improvement
cnns ),
band gaps
adjacency matrices
acsf ),
dc.title.none.fl_str_mv Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
dc.type.none.fl_str_mv Dataset
info:eu-repo/semantics/publishedVersion
dataset
description The photoelectric conversion efficiency (PCE) of perovskites remains beneath the Shockley-Queisser limit, despite its significant potential for solar cell applications. The present focus is on investigating potential multicomponent perovskite candidates, particularly on the application of machine learning to expedite band gap screening. To efficiently identify high-performance perovskites, we utilized a data set of 1346 hybrid organic–inorganic perovskites and employed 11 machine learning models, including decision trees, convolutional neural networks (CNNs), and graph neural networks (GNNs). Four descriptors were utilized for high-throughput screening: sine matrix, Ewald sum matrix, atom-centered symmetry functions (ACSF), and many-body tensor representation (MBTR). The results indicated that LightGBM and CatBoost somewhat surpassed XGBoost in decision tree models, but random forests lagged. Among the CNN models utilizing the same four descriptors, CustomCNN and VGG16 surpassed Xception, while EfficientNetV2B0 exhibited the least favorable performance. When the sine matrix and Ewald sum matrix served as adjacency matrices in GNN models, GCSConv exhibited a considerable improvement over GATConv and a slight advantage over GCNConv. Significantly, GCSConv outperformed other models when utilized with the Ewald sum matrix. The ideal combination of descriptors and algorithms identified was MBTR + CustomCNN, with an <i>R</i><sup>2</sup> of 0.94. Subsequently, three perovskites exhibiting appropriate Heyd–Scuseria–Ernzerhof (HSE06) band gaps were identified to define the defects. Among them, CH<sub>3</sub>C(NH<sub>2</sub>)<sub>2</sub>SnI<sub>3</sub> exhibited superior performance in both vacancy and substitutional defects compared to C<sub>3</sub>H<sub>8</sub>NSnI<sub>3</sub> and (CH<sub>3</sub>)<sub>2</sub>NH<sub>2</sub>SnI<sub>3</sub>. This high-throughput screening method with machine learning establishes a robust foundation for selecting solar materials with exceptional photoelectric properties.
eu_rights_str_mv openAccess
id Manara_7ba34bd9bd85d2ee78a7b824098a18d4
identifier_str_mv 10.1021/acs.jpclett.4c03580.s002
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/28193341
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY-NC 4.0
spelling Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar CellsZhaosheng Zhang (4603021)Sijia Liu (1751842)Qing Xiong (136442)Yanbo Liu (180774)Cell BiologyCancerScience PolicySpace ScienceBiological Sciences not elsewhere classifiedChemical Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedvgg16 surpassed xceptionsolar cell applicationsselecting solar materialsrandom forests laggedphotoelectric conversion efficiencyincluding decision treesgraph neural networksexceptional photoelectric propertiesconvolutional neural networkscentered symmetry functionsbody tensor representationleast favorable performanceperovskites remains beneathewald sum matrix8 </ sub3 </ sub2 </ sup2 </ subdecision tree modelscnn models utilizingthroughput screening methodsubstitutional defects comparedmachine learning establishesexhibited superior performanceefficiently identify highperformance perovskitesthroughput screeningmachine learningr </sine matrix>< supgcsconv exhibitedefficientnetv2b0 exhibitedgnn modelsstrategic integrationslight advantagesignificant potentialrobust foundationresults indicatedqueisser limitpresent focusideal combinationgnns ).gcsconv outperformeddata setconsiderable improvementcnns ),band gapsadjacency matricesacsf ),The photoelectric conversion efficiency (PCE) of perovskites remains beneath the Shockley-Queisser limit, despite its significant potential for solar cell applications. The present focus is on investigating potential multicomponent perovskite candidates, particularly on the application of machine learning to expedite band gap screening. To efficiently identify high-performance perovskites, we utilized a data set of 1346 hybrid organic–inorganic perovskites and employed 11 machine learning models, including decision trees, convolutional neural networks (CNNs), and graph neural networks (GNNs). Four descriptors were utilized for high-throughput screening: sine matrix, Ewald sum matrix, atom-centered symmetry functions (ACSF), and many-body tensor representation (MBTR). The results indicated that LightGBM and CatBoost somewhat surpassed XGBoost in decision tree models, but random forests lagged. Among the CNN models utilizing the same four descriptors, CustomCNN and VGG16 surpassed Xception, while EfficientNetV2B0 exhibited the least favorable performance. When the sine matrix and Ewald sum matrix served as adjacency matrices in GNN models, GCSConv exhibited a considerable improvement over GATConv and a slight advantage over GCNConv. Significantly, GCSConv outperformed other models when utilized with the Ewald sum matrix. The ideal combination of descriptors and algorithms identified was MBTR + CustomCNN, with an <i>R</i><sup>2</sup> of 0.94. Subsequently, three perovskites exhibiting appropriate Heyd–Scuseria–Ernzerhof (HSE06) band gaps were identified to define the defects. Among them, CH<sub>3</sub>C(NH<sub>2</sub>)<sub>2</sub>SnI<sub>3</sub> exhibited superior performance in both vacancy and substitutional defects compared to C<sub>3</sub>H<sub>8</sub>NSnI<sub>3</sub> and (CH<sub>3</sub>)<sub>2</sub>NH<sub>2</sub>SnI<sub>3</sub>. This high-throughput screening method with machine learning establishes a robust foundation for selecting solar materials with exceptional photoelectric properties.2025-01-13T06:58:10ZDatasetinfo:eu-repo/semantics/publishedVersiondataset10.1021/acs.jpclett.4c03580.s002https://figshare.com/articles/dataset/Strategic_Integration_of_Machine_Learning_in_the_Design_of_Excellent_Hybrid_Perovskite_Solar_Cells/28193341CC BY-NC 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/281933412025-01-13T06:58:10Z
spellingShingle Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
Zhaosheng Zhang (4603021)
Cell Biology
Cancer
Science Policy
Space Science
Biological Sciences not elsewhere classified
Chemical Sciences not elsewhere classified
Information Systems not elsewhere classified
vgg16 surpassed xception
solar cell applications
selecting solar materials
random forests lagged
photoelectric conversion efficiency
including decision trees
graph neural networks
exceptional photoelectric properties
convolutional neural networks
centered symmetry functions
body tensor representation
least favorable performance
perovskites remains beneath
ewald sum matrix
8 </ sub
3 </ sub
2 </ sup
2 </ sub
decision tree models
cnn models utilizing
throughput screening method
substitutional defects compared
machine learning establishes
exhibited superior performance
efficiently identify high
performance perovskites
throughput screening
machine learning
r </
sine matrix
>< sup
gcsconv exhibited
efficientnetv2b0 exhibited
gnn models
strategic integration
slight advantage
significant potential
robust foundation
results indicated
queisser limit
present focus
ideal combination
gnns ).
gcsconv outperformed
data set
considerable improvement
cnns ),
band gaps
adjacency matrices
acsf ),
status_str publishedVersion
title Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
title_full Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
title_fullStr Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
title_full_unstemmed Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
title_short Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
title_sort Strategic Integration of Machine Learning in the Design of Excellent Hybrid Perovskite Solar Cells
topic Cell Biology
Cancer
Science Policy
Space Science
Biological Sciences not elsewhere classified
Chemical Sciences not elsewhere classified
Information Systems not elsewhere classified
vgg16 surpassed xception
solar cell applications
selecting solar materials
random forests lagged
photoelectric conversion efficiency
including decision trees
graph neural networks
exceptional photoelectric properties
convolutional neural networks
centered symmetry functions
body tensor representation
least favorable performance
perovskites remains beneath
ewald sum matrix
8 </ sub
3 </ sub
2 </ sup
2 </ sub
decision tree models
cnn models utilizing
throughput screening method
substitutional defects compared
machine learning establishes
exhibited superior performance
efficiently identify high
performance perovskites
throughput screening
machine learning
r </
sine matrix
>< sup
gcsconv exhibited
efficientnetv2b0 exhibited
gnn models
strategic integration
slight advantage
significant potential
robust foundation
results indicated
queisser limit
present focus
ideal combination
gnns ).
gcsconv outperformed
data set
considerable improvement
cnns ),
band gaps
adjacency matrices
acsf ),