Geometries of defects in nanodiamonds optimized with the low-cost methods: How good are they for the electronic g-tensor calculations?

Abstract

The geometry optimization of 30 paramagnetic defects, including biomedically attractive nitrogen-, silicon-, germanium-, and nickel-related color centers, is performed after their incorporation into hydrogenated nanodiamond (ND) of C84H64 size. The main aim is to examine the effectiveness of the low-cost methods, namely, PBEh-3c, r2SCAN-3c, B97-3c, HF-3c, and GFN2-xTB, in reproducing the geometries of these defects basing on the similarity between the results of the subsequent electronic g-tensor calculations. It is revealed that the overall performance of PBEh-3c, r2SCAN-3c, and B97-3c is very alike and can be considered as good, however, none of these “3c” approaches is able to cope with all tested geometries. The results of HF-3c, on the other hand, are disappointing, as this method is outperformed by computationally much more lighter GFN2-xTB. Additional calculations carried out for dangling bonds introduced into hydroxylated and aminated NDs show that all low-cost methods perform reasonably well for this type of defect but the largest quantitative discrepancies once again are demonstrated by HF-3c. The obtained findings lay the foundations for the future studies of larger NDs with the purpose to figure out the magnetic properties dependence on the size of NDs or defect positions within NDs.