School of Physics, CRANN Institute and AMBER, Trinity College, Dublin 2, Ireland
Title of the talk
Spin-Phonon Relaxation: theory, ab initio simulations and machine-learning modelling.
Abstract
The interaction between spin and lattice’s vibrations, namely the spin-lattice interaction, is one of the main limitations to the spin lifetime in insulating materials and the control of such interaction is of key im- portance for the development of technologies based on spin, such as quantum sensing and spintronics. Although spin-lattice relaxation has a central role in the physics of magnetism and magnetic resonance, its theoretical formulation is mostly based on a phenomenological approach and a quantitative under- standing of its microscopic origin is often lacking.
In this contribution I will show the progresses in developing a computational approach able to tackle the challenge of predicting spin-lattice (T1) relaxation from first principles. The formalism is based on a spin Hamiltonian description of magnetism and exploits the theory of open quantum systems in order to describe the dissipative effect of phonons on the spin degrees of freedom. This formalism is then mapped onto electronic structure theory, such as Density Functional Theory and Complete Active Space SCF, in order to determine all the parameters of the model in a full first-principles fashion[1]. Moreover, I will show how machine learning can be used to interpolate electronic structure results[2,3] and speed up the calculation of spin-phonon coupling terms and lattice force constants by orders of magnitude, making it possible to tackle both first and second order perturbative effects[4,5]. The method is applied to a series of molecular crystals of Kramers ions, including prototypical molecular qubits, such as S = 1/2 V(IV) complexes[1,2], and single-ion magnets, such as S = 3/2 Co(II)[5] and J = 15/2 Dy(III) compounds[6]. Ab initio spin dynamics is able to explain the dependence of spin-lattice relaxation time with respect to temperature for all the investigated compounds, making it possible to individuate the origin of slow relaxation for all the most relevant classes of magnetic molecules. Finally, thanks to the fully-ab initio nature of the method, it is also possible to disentangle all the spin and molecular interactions leading to spin-phonon coupling, making it possible to discuss future directions for this field and possible strategies for the enhancement of spin lifetime in solid-state magnetic molecules.
References
[1] A. Lunghi, S. Sanvito, Science Advances, 2019, 5, eaax7163
[2] A. Lunghi, S. Sanvito, Science Advances, 2019, 5, eaaw2210
[3] A. Lunghi, S. Sanvito, The Journal of Physical Chemistry C, 2020, 124, 5802-5806
[4] A. Lunghi, S. Sanvito, The Journal of Physical Chemistry Letters, 2020, 11, 6273–6278
[5] A. Lunghi, S. Sanvito, The Journal of Chemical Physics, 2020, 153, 174113
[6] M. Briganti, et al., The Journal of the American Chemical Society, 2021, 143, 13633–13645
