2018
DOI: 10.1021/jacs.8b08800
|View full text |Cite
|
Sign up to set email alerts
|

Machine-Learning Prediction of CO Adsorption in Thiolated, Ag-Alloyed Au Nanoclusters

Abstract: We propose a machine-learning model, based on the random-forest method, to predict CO adsorption in thiolate protected nanoclusters. Two phases of feature selection and training, based initially on the Au 25 nanocluster, are utilized in our model. One advantage to a machinelearning approach is that correlations in defined features disentangle relationships among the various structural parameters. For example, in Au 25 , we find that features based on the distribution of Ag atoms relative to the CO adsorption s… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

2
100
1
1

Year Published

2019
2019
2024
2024

Publication Types

Select...
9

Relationship

0
9

Authors

Journals

citations
Cited by 126 publications
(104 citation statements)
references
References 52 publications
2
100
1
1
Order By: Relevance
“…More specifically, the relationship between features and the target property should be strong. [ 9,22 ] Therefore, each material is uniquely described by seven element feature matrixes: 1) the Mendeleev number ( N Men ), 2) the number of unpaired electrons of metallic elements and the unoccupied orbits of non‐metallic elements ( N e ), 3) the electronegativity (χ), 4) the covalent radius ( R cov ), 5) the enthalpies of atomization (Δ H at ), 6) the first ionization energy ( IP 1 ), and 7) the polarizability ( P ). In each feature matrix, the main diagonal consists of the elemental properties of the atom itself, and the non‐diagonal elements correspond to the ratios of elemental property values between the atoms i and j, except for the feature matrix N e , in which the non‐diagonal elements are in the form of summation.…”
Section: Figurementioning
confidence: 99%
“…More specifically, the relationship between features and the target property should be strong. [ 9,22 ] Therefore, each material is uniquely described by seven element feature matrixes: 1) the Mendeleev number ( N Men ), 2) the number of unpaired electrons of metallic elements and the unoccupied orbits of non‐metallic elements ( N e ), 3) the electronegativity (χ), 4) the covalent radius ( R cov ), 5) the enthalpies of atomization (Δ H at ), 6) the first ionization energy ( IP 1 ), and 7) the polarizability ( P ). In each feature matrix, the main diagonal consists of the elemental properties of the atom itself, and the non‐diagonal elements correspond to the ratios of elemental property values between the atoms i and j, except for the feature matrix N e , in which the non‐diagonal elements are in the form of summation.…”
Section: Figurementioning
confidence: 99%
“…Our two machine learning models based on the neural network algorithm show RMSE of 0.0521 eV for CO adsorption energy and 0.0614 eV for HOCO formation energy on testing sets. To place our model accuracy in a more straightforward context, we compared our errors to a similar work in predicting CO adsorption energy in Thiolated Ag-alloyed Au nanoclusters 38 , which finds a much higher RMSE at ~0.17eV using over 2000 data points for training. Another work using machine learning for predicting adsorption energies of CH4 related species (CH3, CH2, CH, C, and H) on the Cu-based alloys 39 reported the best performance of RMSEs around 0.3 eV after an extra tree regression algorithm.…”
Section: Acs Paragon Plus Environmentmentioning
confidence: 99%
“…Nanoclusters, with the advantages of precise compositions and well-defined structures, provide an exciting opportunity to grasp the structure–property correlation at the atomic level 1,2. The quantum size effect of nanoclusters endows them with a plethora of properties, such as photo-luminescence (PL), catalysis, chirality, and magnetism, to name a few 1,2. The property manipulation at the atomic level has long been a hot topic, and has allowed a series of nanoclusters with controllable chemical–physical properties to be produced 1f–h,2g,i,k,l…”
Section: Introductionmentioning
confidence: 99%