Recommended by

Indexed by

Pure isotropic proton NMR spectra in solids using deep learning

Manuel Cordova1*, Pinelopi Moutzouri1, Bruno Simões de Almeida1, Daria Torodii1, Lyndon Emsley1*

1 Institut des Sciences et Ingénierie Chimiques, École Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland

* Corresponding authors emails: manuel.cordova@epfl.ch, lyndon.emsley@epfl.ch
DOI10.24435/materialscloud:a7-59 [version v1]

Publication date: Dec 22, 2022

How to cite this record

Manuel Cordova, Pinelopi Moutzouri, Bruno Simões de Almeida, Daria Torodii, Lyndon Emsley, Pure isotropic proton NMR spectra in solids using deep learning, Materials Cloud Archive 2022.180 (2022), doi: 10.24435/materialscloud:a7-59.


The resolution of proton solid-state NMR spectra is usually limited by broadening arising from dipolar interactions between spins. Magic-angle spinning alleviates this broadening by inducing coherent averaging. However, even the highest spinning rates experimentally accessible today are not able to completely remove dipolar interactions. Here, we introduce a deep learning approach to determine pure isotropic proton spectra from a two-dimensional set of magic-angle spinning spectra acquired at different spinning rates. Applying the model to 8 organic solids yields high-resolution 1H solid-state NMR spectra with isotropic linewidths in the 50-400 Hz range.

Materials Cloud sections using this data

No Explore or Discover sections associated with this archive record.


File name Size Description
40.4 MiB Python code used to train and use the model and pre-trained model
13.0 MiB Datasets of variable-rate MAS 1H NMR experiments on which the model is applied
263.8 MiB NMR experiments used to determine the assignment of the proton spectra
182.2 KiB Assignment probabilities of the carbon chemical shifts of molnupiravir
47.7 KiB Quantum ESPRESSO DFT chemical shift computations to assign chemical shifts of MDMA hydrochloride


Files and data are licensed under the terms of the following license: Creative Commons Attribution Share Alike 4.0 International.
Metadata, except for email addresses, are licensed under the Creative Commons Attribution Share-Alike 4.0 International license.


MARVEL/DD1 machine learning NMR resolution

Version history:

2022.180 (version v1) [This version] Dec 22, 2022 DOI10.24435/materialscloud:a7-59