Bulk ferromagnetism in nitronyl nitroxide crystals: a first principles bottom-up comparative study of four bulk nitronyl nitroxide ferromagnets (KAXHAS, YOMYII, LICMIT and YUJNEW)

Abstract

Searching for general trends that could help to design new bulk (3D) ferromagnets, we have carried out a first-principles bottom-up theoretical study of four bulk ferromagnets of the nitronyl nitroxide (NN) family of crystals: β-p-(nitro)phenyl-NN (KAXHAS), α-o-(hydroxy) phenyl-NN (YOMYII), α-2,5-(dihydroxy)phenyl-NN (LICMIT) and p-(methylthio) phenyl-NN (YUJNEW). This first-principles bottom-up theoretical study connects, in a rigorous form, the microscopic magnetic pair interactions (J _AB) with the macroscopic magnetic properties (e.g. magnetic susceptibility or heat capacity curves). The microscopic magnetic pair interactions are computed from the crystal structure using DFT methods. The network of non-negligible J _AB magnetic pair interactions between AB radicals provides the magnetic topology of the crystal.

 Using the room-temperature crystal structure (the only experimentally available) for KAXHAS, YOMYII and LICMIT, we found a 3D magnetic topology. However, although the computed magnetic susceptibility χ( T  ) curves reproduce, in all three cases, the ferromagnetic experimental χ( T  ) data, not all the J _AB pair interactions are ferromagnetic. For YUJNEW, there are three sets of crystallographic coordinates (room temperature, 114 and 10K) and its magnetic topology depends on the temperature: at room temperature, there is a 2D magnetic topology but, using the 114 and 10K crystal structures, the magnetic topology becomes 3D. The computed magnetic susceptibility curve only reproduces the experimental data when using the magnetic topology computed at 10K (at 114K, there is only a qualitative agreement). The dependence of the J _AB magnetic pair interactions on the geometry of the A–B radical–radical pair is also analysed. Our results show that small changes in the geometry induce changes in the sign of the interaction, a fact that can be only explained by considering that the nature of the interaction is a cooperative effect that depends on all atoms of the radical.

Acknowledgements

MD, FM and JJN acknowledge the Spanish ‘Ministerio de Ciencia y Tecnología’ (# BQU2002-04587-C02-02), and the Catalan ‘CIRIT’ (2001SGR-0044) for funding. The authors also thank CEPBA-IBM Research Institute, CEPBA and CESCA for allocation of CPU time on their computers. MD acknowledges the financial support of the ‘Ramón y Cajal’ program of the Spanish Ministry of ‘Educación y Ciencia’. This project was partly supported by EPSRC (UK) under grant GR/M86750 (ROPA).

Notes

†The J _AB pair interaction between S=1/2 radical molecules A and B is positive (negative) for ferromagnetic (antiferromagnetic) coupling when using any of the following three definitions of the Heisenberg Hamiltonian:

Here, we will use the first definition (equation (1)). An alternative formulation of the Heisenberg Hamiltonian (3), which can be called Hamiltonian (4), avoids the minus sign and, then, the ferromagnetic coupling is associated to a negative J _AB pair interaction. Notice that for Hamiltonians (1) and (2):

For Hamiltonian (3):

Finally, for Hamiltonian (4):

†The temperature dependence of the magnetic susceptibility was well reproduced by the ST model (J/k _B=+0.93K) with a positive Weiss temperature of θ=+0.46K, according to the following the Bleaney–Bowers expression corrected with an empirical mean field θ parameter: