Determination of the g-, hyperfine coupling- and zero-field splitting tensors in EPR and ENDOR using extended Matlab codes

• Published in: Journal of magnetic resonance (1997) 2021-04, Vol.325, p.106956-106956, Article 106956
• Main Authors: Lund, Anders, Callens, Freddy, Sagstuen, Einar
• Format: Article
• Language: English
• Subjects: Coupling tensors; Data analysis; ENDOR; EPR; MatLab open code; Single crystals
• ISSN: 1090-7807; 1096-0856; 1096-0856
• DOI: 10.1016/j.jmr.2021.106956

[Display omitted] •Methods to obtain the g, zerofield and hyperfine coupling tensors.•Compact and transparent MatLab®codes to replace obsolete software.•Correction of crystal misorientation by a simplified method.•Functions for analysis of single crystal data of free radicals and triplet states.•Manual and functions at https://old.liu.se/simarc/downloads?l=en. The analysis of single crystal electron magnetic resonance (EMR) data has traditionally been performed using software in programming languages that are difficult to update, are not easily available, or are obsolete. By using a modern script-language with tools for the analysis and graphical display of the data, three MatLab® codes were prepared to compute the g, zero-field splitting (zfs) and hyperfine coupling (hfc) tensors from roadmaps obtained by EPR or ENDOR measurements in three crystal planes. Schonland’s original method was used to compute the g- and hfc -tensors by a least-squares fit to the experimental data in each plane. The modifications required for the analysis of the zfs of radical pairs with S = 1 were accounted for. A non-linear fit was employed in a second code to obtain the hfc -tensor from EPR measurements, taking the nuclear Zeeman interaction of an I = ½ nucleus into account. A previously developed method to calculate the g- and hfc -tensors by a simultaneous linear fit to all data was used in the third code. The validity of the methods was examined by comparison with results obtained experimentally, and by roadmaps computed by exact diagonalization. The probable errors were estimated using functions for regression analysis available in MatLab. The software will be published at https://doi.org/10.17632/ps24sw95gz.1, Input and output examples presented in this work can also be downloaded from https://old.liu.se/simarc/downloads?l=en.