Octupolar-Driven Light-Induced Hysteresis in Antiferromagnetic FePS₃

Studying light–matter interactions in low-dimensional quantum materials is an important field in modern condensed matter physics and promise new ways to functionalities that cannot be achieved by conventional electronics. For instance, van der Waals (vdW) magnets stand out for their unique ability to retain magnetic order down to the atomic limit, despite the constraints of reduced dimensionality and where their layered structure, held together by weak interlayer forces, permits mechanical exfoliation to few-layer or even monolayer thickness, which unlock new regimes where spin, charge, and lattice degrees of freedom become highly tunable. This tunability is not limited to traditional means such as chemical doping, electrostatic gating, or strain; light itself can act as a non-contact, ultrafast, and reversible control parameter. By coupling directly to the electronic orbitals, optical fields can perturb a system far from equilibrium, open unconventional excitation channels, and potentially stabilize exotic states that do not exist in the equilibrium ground state. However, the prospect of using light to directly manipulate electronic polarization in magnetic semiconductors remains largely unexplored. In ferroelectrics and some multiferroics, polarization control is typically achieved through electric or magnetic fields, while in semiconducting magnets, polarization effects are often subtle, masked by spin ordering, or constrained by crystal symmetry. Moreover, achieving a steady-state optical modification of polarization—rather than a transient one observable only in ultrafast pump–probe experiments—poses its own challenges. Compounding the difficulty, polarization changes in such materials can be driven by higher-order multipolar contributions, not just simple dipoles, making their detection and interpretation far from straightforward. Now FePS₃ which is an antiferromagnetic semiconductor in the family of transition metal thiophosphates can provide a compelling testbed for addressing these challenges. Below its Néel temperature (~117 K), it has Ising-type spin ordering without a structural phase transition, meaning any observed in-plane anisotropy in its optical response can be attributed to electronic and magnetic effects alone. This property makes FePS₃ a rare platform for studying light-induced polarization in the absence of lattice distortions. Still, despite its suitability, no prior work had convincingly demonstrated that light could induce hysteresis in its electronic polarization—a signature of memory effects and a hallmark of multiferroic-like behavior.

To this account, new research paper published in Advanced Material and conducted by Dr. Kyung Ik Sim, Dr. Byung Cheol Park, Taesoo Kim, Dr. Byeong Wook Cho, Dr. Jae Hoon Kim, Professor Eun-Mi Choi, and led by Professor Young Hee Lee from the Sungkyunkwan University, the researchers developed a new method to induce and control hysteresis in the electronic polarization of the antiferromagnetic semiconductor FePS₃ using light alone. They selectively triggered either pure dipolar polarization or a combination of dipolar and octupolar components by tuning the excitation energy, the latter breaking mirror symmetry and enabling a robust, path-dependent response. This approach revealed an entirely optical route to create memory-like behavior in a layered magnetic semiconductor, without the need for external electric or magnetic fields.  The research team began by preparing pristine FePS₃ flakes through mechanical exfoliation, transferring them onto quartz substrates under inert conditions to prevent oxidation. They verified each flake’s thickness using atomic force microscopy, ensuring sufficient bulk to probe intrinsic properties while maintaining high optical quality. They then placed the samples inside a low-temperature confocal optical spectroscopy setup, capable of delivering broadband light between 1.4 and 2.2 eV with precise polarization control. The authors directly monitored directional differences in transmission and absorbance from 10 K up to well above the Néel temperature by aligning the light’s electric field along either the crystallographic a-axis (zigzag) or b-axis (armchair).

Above 117 K, the team found the spectra for both polarizations were indistinguishable, confirming the in-plane isotropy of the paramagnetic state. Cooling below the magnetic ordering temperature revealed a clear divergence: absorption for light polarized along the a-axis no longer matched that for the b-axis. This linear dichroism emerged without any detectable phonon shifts in Raman measurements, ruling out structural distortions and pointing to a purely electronic origin tied to spin ordering. Using detailed fits to the transmission data, the researchers extracted the real and imaginary parts of the dielectric constant. They observed that below TN, absorption decreased near 2.0 eV for E ∥ b while increasing near 1.6 eV for E ∥ a, a pattern consistent with light-driven charge transfer between orthogonal crystal directions in the spin-ordered state. The authors examined the dielectric polarization itself and found at 1.6 eV excitation, the induced polarization aligned strongly with the a-axis, while at 2.0 eV it switched to the b-axis. These energies correspond to distinct intersite d–d transitions between Fe²⁺ ions, differing in spin configuration and energetic cost. Angular scans of the dielectric response revealed a surprising feature: at 1.6 eV, the polarization pattern was mirror-symmetric, producing identical traces for forward and reverse angular sweeps. But at 2.0 eV, the symmetry broke down—forward and backward paths diverged, forming a reproducible hysteresis loop.

Fourier analysis showed why. The 1.6 eV response was purely dipolar, while the 2.0 eV case required an additional octupolar component. This higher-order term, absent at lower energies, broke the mirror symmetry of the magnetically ordered lattice and enabled the hysteresis. The effect’s magnitude tracked the temperature dependence of the antiferromagnetic order parameter, vanishing above TN. It was robust across samples of different thicknesses and stable over months which confirmed it as an intrinsic property.

In conclusion, the new study by Professor Young Hee Lee and colleagues demonstrated for the first time that light can act as direct and selective control parameter for electronic polarization in a two-dimensional antiferromagnetic semiconductor. They also showed a robust hysteresis in the polarization state of FePS₃ under specific optical excitation. Indeed, the findings of Sungkyunkwan University scientists reframes our understanding of multiferroicity in layered magnets, showing that purely optical routes can break symmetry in a spin-ordered background strongly enough to generate memory-like behavior. We believe one of the most impactful aspects is the identification of an octupolar electronic polarization component as the primary driver of the hysteresis. Higher-order multipoles have long been treated as secondary or elusive in solid-state systems, often overshadowed by dominant dipolar effects. Here, the octupolar term is not only detectable but essential, dictating how the polarization responds to light and why it retains a path-dependent signature. This recognition broadens the conceptual toolkit for designing materials in which multiple polarization channels—dipolar, octupolar, and potentially beyond—can be engineered and selectively addressed. Moreover, the implications extend to device physics. If polarization states in a magnet can be optically written, read, and switched without contact, the possibilities for ultrafast, low-power, and non-volatile data storage become tangible. Because the effect arises in a semiconducting host, it could be integrated with photonic or optoelectronic circuitry more readily than metallic or purely insulating multiferroics. Furthermore, the strong energy selectivity observed—hysteresis appearing only near 2.0 eV— which suggests that devices could be spectrally encoded, with different wavelengths targeting distinct polarization channels for multiplexed functionality.