Spin-Lattice Relaxation Time of Solid-State Spin Qubits from Molecular Dynamics Simulations

Solid-state spin qubits offer a transformative platform for quantum science and are having a profound impact on quantum information technologies. Their practical application is limited by their decoherence time, which is influenced by several processes, including spin-lattice relaxation induced by spin-phonon interactions. Recent advances in density matrix methods have allowed for the calculation of spin-lattice relaxation times at low temperature; however these approaches rely on assumptions about the number of phonons involved in the relaxation process, and they disregard anharmonic contributions. We circumvent some of these drawbacks by developing a molecular-dynamics approach based on the Kubo theory of magnetic absorption. Our method combines neural-network interatomic potentials with machine-learning models of zero-field splitting tensors. The computational efficiency of the data-driven models enables the simulation of nanosecond-long trajectories and the corresponding zero-field splitting time series with ab initio accuracy. Here, we apply the methodology to solid-state color centers across a wide temperature range and compare our results with experiments.