Thermomagnetic Oscillators with FeRh

The thermomagnetic properties of iron-rhodium (FeRh) make it a suitable candidate for neuromorphic computing. FeRh undergoes a phase transition between 300 K and 400 K from an antiferromagnetic state with high resistivity at low temperature to a ferromagnetic state with low resistivity at high temperature. This unique characteristic enables the potential for FeRh to serve as a resistivity auto-oscillator in the development of energy-efficient computational systems that mimic the brain’s computing power. With a constant applied current through FeRh in its antiferromagnetic phase, FeRh heats up due to its high resistivity transitioning to its ferromagnetic phase, where the low resistivity causes it to cool down again to its antiferromagnetic phase. This may lead to auto-oscillations between the two phases and resistivity states. Our objective is to reduce the transition region to enhance energy efficiency by requiring a smaller current. To investigate FeRh’s potential for such an oscillator, thin films with varying Fe and fixed Rh sputtering powers were grown using sputter deposition from a pure Fe and pure Rh target onto a pre-annealed MgO substrate followed by post-annealing. Film thickness and deposition rates were determined using X-ray reflectivity (XRR). The magnetic response of the films was analyzed using a Magnetic Property Measurement System (MPMS), while resistivity measurements were performed using a Physical Property Measurement System (PPMS) and a vector magnet equipped with a cryostat, allowing a wide temperature range down to 3 K and up to 400 K. Composition dependence was explored by adjusting individual gun power. Films grown with Fe gun power ranging from 24–30 W exhibited the phase transition in both magnetometry and resistivity measurements. Films with 27 W Fe and 10 W Rh powers displayed the most promising phase transition, characterized by its small transition temperature range. Ongoing efforts aim to shift the transition to lower temperatures, with the ultimate goal of achieving thermomagnetic auto-oscillations of FeRh under a constant applied current.

This work was supported by NSF through the Illinois MRSEC (DMR-1720633) and DOE as part of the Q-MEEN-C EFRC (DESC0019273).