Exploring the impact of the inverse Faraday effect on all-optical helicity-dependent magnetization switching

TL;DR

The paper investigates the mechanisms of all-optical helicity-dependent switching (AO-HDS) in FePt granular media, focusing on the cooperative roles of magnetic circular dichroism (MCD) and the inverse Faraday effect (IFE) and how photon energy shapes switching efficiency. It combines ultrafast circularly polarized pulses ($<200$ fs) in the near-infrared range ($800$–$1500$ nm) with Kerr microscopy to map switched regions, supported by ab initio calculations of optical constants and a multi-scale switching model linking absorbed fluence to switching probability. Key findings show that MCD creates a spin-channel temperature separation near the Curie temperature $T_{ m C}$, enabling switching, while IFE contributes an induced magnetization $\,\Delta M^{\mathrm{ind}}\,=\,K_{\mathrm{IFE}} \cdot \frac{I}{c}$ that enhances switching at lower deposited fluences; increasing fluence raises thermal fluctuations, diminishing switching efficiency despite larger $\,\Delta M^{\mathrm{ind}}$. Unlike multilayer systems where domain-wall motion governs reversal, switching here is per-grain and governed by the balance of absorption-driven heating and the IFE contribution, with an optimal absorbed-fluence window that maximizes the effect of IFE before thermal disorder dominates. Altogether, the work provides design insights for energy-efficient AO-HDS in nanoscale grains, informing approaches to high-density optical magnetic storage.

Abstract

All-optical helicity-dependent magnetization switching (AO-HDS) is the quickest data recording technique using only ultrashort laser pulses. FePt grains provide an ideal platform for examining the interaction of effects conducting magnetization switching. We identify the magnetic circular dichroism (MCD) and the inverse Faraday effect (IFE) as the primary switching forces. Ultrafast photon absorption rapidly elevates electron temperatures, quenching magnetization. The MCD's helicity-dependent absorption ensures distinct electron temperatures, holding a finite switching probability by generating different spin noise rates in each spin channel. The IFE induces a magnetic moment, enhancing this probability. We present ultrashort laser pulse (<200 fs) AO-HDS experiments in the near-infrared spectral range from 800 nm to 1500 nm, demonstrating a correlation between switching efficiency and absorbed energy density. Elevating electron temperatures to the Curie point enables the IFE to induce a magnetic moment for deterministic switching in the quenched magnetization state. Unlike in films or multilayers, where domain wall motion and domain growth govern the switching process, increasing the MCD in nanometer-sized grains does not enhance switching efficiency. Electrons around the Curie temperature typically reach increased switching rates for higher induced magnetization generated by the IFE. The MCD sets the necessary switching condition, separating electron temperatures. The IFE generates a magnetic moment, directing spins toward the desired orientation and improving switching efficiency. Every laser pulse initiates a new switching probability for each grain, increasing the role of direction indication by the IFE. Stronger absorption assures higher induced magnetization at low switching fluences.

Figures (9)

• Figure 1: a) Schematic of the sample structure used in the experiments. The magnetic grains, shown in white and grey, populate the MgO seed-layer unordered. The arrows visualize the grains strong out-of-plane magnetization anisotropy. b) SEM picture of the sample surface with an average grains size $\sim10\,nm$. c) Experimental setup schematic. linearly polarized Laser pulses from a Ti:Sapphire laser are circularly polarized by a quarter wave plate and subsequently focused by a lens to a spot size of $40\,µm$, illuminating a sample mounted to a transitional stage.
• Figure 2: a) Kerr-microscopy image (16-bit gray scale values) of a demagnetized sample with lines switched by $\lambda_{\mathrm{c}} = 1300\, nm$ pulses at fluence $F = 8\,mJ\per\square cm$. The red arrow indicates the sample transition direction for the extracted profiles b) by integrating the area marked by the red rectangular box for signal-to-noise ratio reduction.
• Figure 3: Extracted line profiles switched using $\sigma^{-}$ helicity for all wavelengths $\lambda_{\mathrm{c}} = 800\,nm - 1500\,nm$ at a deposited average fluence $F=9.3\pm0.6\,mJ\per\square{cm}$ scaled on the left y-axis. The grey-shaded Gaussian function in the background represents the radial fluence distribution scaled on the right y-axis. The black dashed line marks the center of the written lines. The blue and red dashed lines define the line widths for switching with $\lambda_{\mathrm{c}} = 800\,nm$ and $\lambda_{\mathrm{c}} = 1450\,nm$ pulses respectively. The black dotted circle points out peaks originating from surface distortions.
• Figure 4: a) Switched magnetization $\Delta M/M_{\mathrm{S}}$ (lines are a guide to the eye) and b) switched area $A_{\mathrm{S}}$ (lines are linear fits) versus deposited laser fluence for the investigated spectrum. a) $\Delta M/M_{\mathrm{S}}$ reaches a peak at optimum fluence, then the switching efficiency decreases for increasing fluences. b) $A_{\mathrm{S}}$ increases with the deposited laser fluence. The gray lines (light gray exterior outlines the error) trace the spectral threshold fluence dependence increasing with the wavelength.
• Figure 5: a) Ab-initio calculated complex refractive index $n+ik$ with the real and imaginary part, respectively, calculated for both helicities and the resulting penetration depth $\Lambda_{\mathrm{opt}}$. b) Reflectivity $R$ and transmission $T$, and the resulting absorption $\alpha_{\mathrm{abs}}$ for a $10\,\mathrm{nm}$ thick FePt layer.