Dresden 2026 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MA: Fachverband Magnetismus

MA 42: Molecular Magnetism and Magnetic Particles / Clusters II

MA 42.7: Vortrag

Donnerstag, 12. März 2026, 11:15–11:30, HSZ/0004

Fully Quantum-Mechanical Simulations of Spin Crossover Behaviour in an Iron(II) Complex — •Ethan Crawford, Solveig Felton, and David Wilkins — Queen's University Belfast, Belfast, Northern Ireland

Spin-crossover (SCO) complexes change spin state when exposed to external stimuli such as light, temperature or pressure, allowing them to be used as sensors or magnetic memory devices. Predicting with precision when a crossover occurs would bring the technology closer to fruition, but it remains a significant challenge. Single-point energy calculations using Density Functional Theory (DFT) have been employed with limited accuracy, because they neglect the vibrational and other entropic contributions present at finite temperatures. Furthermore, a given calculation setup is not necessarily transferable to another complex. We trained a committee neural network potential (c-NNP) using data generated via Ab Initio Molecular Dynamics for an Iron (II) SCO complex. Two c-NNPs were trained, to model both the highest and lowest spin states respectively, encompassing all intramolecular interactions. The models were systematically improved using the query by committee approach. These systems were extended from a single complex to a bulk system combining the c-NNPs with a classical forcefield to account for the intermolecular interactions. The spin state for each complex is individually assigned to simulate experimental transitions. During the transition window, neither model can accurately predict the behaviour of the complex, but combining the three forcefields should replicate experimental behaviour.

Keywords: Spin Crossover; Ab Initio Molecular Dynamics; Density Functional Theory