Real-time Simulations of Spin Dynamics with Time-dependent Density Functional Theory.

Understanding the behaviour of electrons in materials subjected to rapidly changing magnetic fields is crucial for advancements in areas such as data storage and spintronics, a technology that exploits the intrinsic spin of the electron alongside its charge. Researchers are now refining computational methods to model these dynamic processes with increasing accuracy and efficiency. A team comprising Jacopo Simoni, Wuzhang Fang, and Andrew C. Grieder from the University of Wisconsin-Madison, alongside Xavier Andrade, Alfredo A. Correa, and Tadashi Ogitsu from Lawrence Livermore National Laboratory, and Yuan Ping from the University of Wisconsin-Madison, detail their work on extending time-dependent density functional theory (TDDFT), a quantum mechanical method used to calculate the evolution of many-electron systems, to incorporate non-collinear spin dynamics. Their article, “Spin non-Collinear Real-Time Time-Dependent Density-Functional Theory and Implementation in the Modern GPU-Accelerated INQ code”, presents an implementation within the INQ software package, designed to leverage the parallel processing capabilities of graphics processing units (GPUs) to simulate these complex interactions in both magnetic clusters and solid-state materials, potentially enabling detailed analysis of phenomena like magnon behaviour and ultrafast spin dynamics induced by laser excitation.

Accurate computational modelling of magnetism necessitates sophisticated techniques, and researchers have recently developed a non-collinear time-dependent density functional theory (TDDFT) implementation within the INQ software package to address limitations in simulating spin dynamics. TDDFT, a quantum mechanical method, calculates the time evolution of many-electron systems responding to external fields, and the non-collinear extension allows for modelling magnetic systems where the magnetic moments are not aligned in a single direction. This implementation leverages the parallel processing power of graphics processing units (GPUs) to efficiently model these complex systems, enabling detailed computational studies of magnetic phenomena that were previously inaccessible.

The implementation incorporates crucial magnetic effects, including exchange-correlation magnetic fields, which arise from the interactions between electrons, spin-orbit coupling, a relativistic effect linking an electron’s spin to its orbital motion, and the interaction between the electronic system and external magnetic fields. This comprehensive treatment allows researchers to investigate complex magnetic phenomena in real-time, opening avenues for exploring a wide range of materials and phenomena at the nanoscale and beyond. The code now enables the study of magnon dynamics – the quantum of spin waves – and ultrafast spin dynamics induced by laser excitation, providing a powerful tool for materials scientists and physicists.

Researchers validated the implementation by simulating spin dynamics in both magnetic clusters and solid-state materials following light excitation, confirming the code’s accuracy and reliability. Crucially, the code’s capabilities extend to predicting spectroscopic signatures, specifically magnetic circular dichroism and pump-probe Kerr rotation. Magnetic circular dichroism measures the difference in absorption of left- and right-circularly polarized light, while pump-probe Kerr rotation probes the magnetization dynamics by measuring changes in the polarization of reflected light. These predicted signatures offer a powerful tool for interpreting experimental results and validating theoretical models.

The code facilitates the modelling of real-time dynamics of magnons, quantized spin waves that carry energy and momentum within a magnetic material, and the study of ultrafast spin dynamics following excitation by laser light. Furthermore, the code allows for the investigation of responses to both linearly and circularly polarized illumination, offering insights into fundamental magnetic processes.

The implementation allows for the calculation of spectroscopic signatures, such as magnetic circular dichroism and pump-probe Kerr rotation, establishing a direct link between theoretical predictions and experimental observations. By comparing these simulated spectroscopic signatures with experimental data, researchers can validate their models and gain a deeper understanding of the underlying physics governing spin dynamics in various materials.

The methodology captures the evolution of magnons and ultrafast spin dynamics induced by both linearly and circularly polarised laser pulses, providing a comprehensive picture of magnetic behaviour on femtosecond timescales. This capability allows for detailed analysis of complex magnetic phenomena, paving the way for the development of novel spintronic devices.

The implemented non-collinear TDDFT framework facilitates the calculation of spectroscopic signatures, including magnetic circular dichroism and pump-probe Kerr rotation, establishing a direct link between theoretical predictions and experimental observations. These calculations enable the validation of the methodology and further understanding of the underlying physical mechanisms, thereby enhancing the reliability and accuracy of computational models.

The efficient GPU-based implementation ensures scalability, allowing for simulations of increasingly large and complex systems, expanding the scope of research and enabling investigations into more realistic materials. The INQ code, with its enhanced capabilities, provides a powerful platform for exploring the fundamental aspects of spin dynamics and developing novel spintronic devices, offering a versatile tool for materials scientists and physicists. Future work will focus on extending the methodology to investigate more complex materials and phenomena, such as interfacial magnetism and spin transport.