CE29 - Chimie : analyse, théorie, modélisation

Materials with Targeted Responsive Behavior – MataReB

The last 15 years have seen the emergence of entire new classes of crystalline nanoporous materials, which exhibit large or anomalous responses to external physical or chemical stimulation. These modifications of framework structure and pore dimensions also involve, in turn, a modification of other physical and chemical properties, making such materials multifunctional (or “smart materials”). One of the outstanding challenges in this field is the systematic synthesis of materials with controlled functionality and porosity. We propose here to remedy this by the development of novel computational chemistry methods that can predict the response of structures under various physical or chemical stimuli. These methods will then be applied on known materials to generate a training dataset for a machine learning procedure, thus allowing to provide rapid screening of large databases of hypothetical materials.

We have provided a systematic characterization of the mechanical properties of 600 000 hypothetical zeolitic structures, highlighting generic trends between mechanical properties and energetic stability. We showed that machine learning is a promising alternative approach, with better accuracy that the current screening approaches.

The computational chemistry toolbox developed in our group is able to predict the behavior of soft porous crystals, and guide the rational synthesis of novel materials with targeted physical properties for applications. This is a key result for prediction of nonlinear mechanical properties of materials at a large scale, traditionally considered a difficult problem.

We are starting very recently to use this data as a training dataset in order to predict the mechanical properties of an even larger number of materials, by training a machine learning algorithm. First results on this database are encouraging, although still relatively limited by the size of the training dataset. Our projects for the next year include increasing the amount of data available for training, in order to make more accurate predictions. We will, as well, use the data produced in order to train a new generation of classical force fields for aluminosilicate framework materials.

1. R. Gaillac, P. Pullumbi, T. D. Bennett and F.-X. Coudert, Chem. Mater., 2020, 32 (18), 8004–8011
2. S. Krause, J. D. Evans, V. Bon, I. Senkovska, S. Ehrling, P. Iacomi, D. M. Többens, D. Wallacher, M. S. Weiss, B. Zheng, P. G. Yot, G. Maurin, P. L. Llewellyn, F.-X. Coudert and S. Kaskel, Chem. Sci., 2020, 11 (35), 9468–9479
3. S. Chibani and F.-X. Coudert, APL Mater., 2020, 8 (8), 080701
4. F.-X. Coudert, Acc. Chem. Res., 2020, 53 (7), 1342–1350
5. L. R. Redfern, M. Ducamp, M. C. Wasson, L. Robison, F. A. Son, F.-X. Coudert and O. K. Farha, Chem. Mater., 2020, 32 (13), 5864–5871

Submission summary

The last 15 years have seen the emergence of entire new classes of crystalline nanoporous materials, which exhibit large or anomalous responses to external physical or chemical stimulation. These modifications of framework structure and pore dimensions also involve, in turn, a modification of other physical and chemical properties, making such materials multifunctional (or “smart materials”). One of the outstanding challenges in this field is the systematic synthesis of materials with controlled functionality and porosity. We propose here to remedy this by the development of novel computational chemistry methods that can predict the response of structures under various physical or chemical stimuli. These methods will then be applied on known materials to generate a training dataset for a machine learning procedure, thus allowing to provide rapid screening of large databases of hypothetical materials.

Project coordinator

Monsieur Coudert François-Xavier (Institut de Recherche de Chimie Paris)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

IRCP Institut de Recherche de Chimie Paris

Help of the ANR 218,621 euros
Beginning and duration of the scientific project: March 2019 - 42 Months

Useful links

Explorez notre base de projets financés

 

 

ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter