Raman spectroscopy measures vibrational modes by scattering light off phonons. The scattered light's frequency shift identifies the mode; its intensity depends on symmetry selection rules. Conventional Raman treats left- and right-circularly polarized light identically — the modes don't care about the handedness of the probe.
In pyrite FeS₂, they do (arXiv:2603.21756). The circular intensity difference — left versus right polarization — reverses sign between neighboring {111} crystal faces. The effect appears only for the doubly degenerate Eₘ phonon mode and is absent in all other modes. The crystal distinguishes left from right, but only through one specific vibration on one specific face.
Note (2026-11-10): I checked the abstract of arXiv 2603.21756. It reports Raman optical activity in pyrite FeS2, a centrosymmetric and nonmagnetic crystal with T_h symmetry whose lowest admitted electric toroidal multipole is the octupole of rank three, with a circular intensity difference of about 85 percent of the mean intensity at 633 nm for the doubly degenerate Eg phonon only, reversing sign between neighboring {111} facets and strongly suppressed on {001}; first-principles calculations reproduce the Eg selectivity and the sign reversal with excitation energy. The essay calls pyrite non-centrosymmetric and the octupole an eighth-rank order; the abstract says centrosymmetric and rank three. It also says the calculations reproduce the sign reversal between facets, which is not what the abstract states.
The origin is not ordinary chirality. Pyrite's space group (Pa3̄) is non-centrosymmetric but not conventionally chiral. What it possesses is electric toroidal octupolar symmetry — an eighth-rank multipolar order that conventional probes (dipole, quadrupole) cannot detect. The Raman optical activity is the signature of this higher-rank symmetry, visible because the Eₘ mode couples to it while other modes do not.
First-principles calculations reproduce the sign reversal quantitatively, confirming that the effect is intrinsic to the crystal symmetry and not an artifact of surface conditions or measurement geometry.
The structural insight: the crystal has more symmetry structure than lower-rank probes can detect. Dipole measurements see nothing unusual. Quadrupole measurements see nothing unusual. The octupolar order is invisible to standard spectroscopy — not because it's weak but because the probe's rank is too low to couple to it. Raman optical activity, by comparing circular polarizations, accesses the rank needed. The information was always in the crystal. The measurement had to become sophisticated enough to read it.