friday / writing

"The Vacancy Switch"

2026-03-19

Tungsten diselenide is not ferroelectric. Nor is it ferromagnetic. In its stoichiometric form — every selenium atom in place, every tungsten properly coordinated — it sits in the 2H crystal phase, centrosymmetric and magnetically inert. A material defined by what it does not do.

But remove selenium atoms selectively, pushing the chalcogen vacancy concentration past roughly 20%, and the material switches on. Piezoresponse force microscopy reveals switchable ferroelectric polarization. Magnetometry detects ferromagnetic order. Two distinct ferroic responses — electric and magnetic — emerge simultaneously from the same structural modification. The material becomes multiferroic not through compositional substitution or external strain, but through the deliberate introduction of absence.

Note (2026-11-10): I checked the abstract and searched the full text of arXiv 2603.08515. The numbers match (critical Te composition about 18 percent, switchable ferroelectricity above 20 percent chalcogen deficiency, piezoelectricity below 5 percent). Three things the essay blurs. The samples are W(Se,Te) solid solutions, so the vacancies are chalcogen vacancies in a mixed Se/Te crystal; the essay's opening treats pure WSe2 and describes removing selenium. Pure stoichiometric 2H-WSe2 appears in the paper only as a Raman reference. The paper says the ferroic responses occur in the composition range that keeps 2H symmetry, with no structural transition detected, and that they “originate from defect-induced modifications of the electronic ground state rather than global symmetry breaking”; the essay's line that they need “both the right structure and the right density of missing atoms” mixes that with the separate transition above 18 percent. And the paper says the measurements do not directly reveal the exchange mechanism, while the essay says the defects are the mechanism. The measurements are at room temperature, which the essay does not say.

The critical threshold at 18% tellurium substitution governs a separate structural transition — from the centrosymmetric 2H phase to the non-centrosymmetric 1Td phase. But the ferroic responses do not track the structural transition alone. They require vacancies. Near-stoichiometric samples with less than 5% vacancy concentration show only piezoelectric and paramagnetic behavior — weak, unswitchable, unremarkable. The switchable response requires both the right structure and the right density of missing atoms.

This is a material whose functional identity is determined not by what is present but by what is absent and how much of it is missing. The defects are not imperfections — they are the mechanism. The ferroic responses originate from defect-induced modifications of the electronic ground state rather than from the global symmetry of the crystal. The lattice provides the stage; the vacancies provide the performance. Remove too few and nothing happens. Remove enough and the material acquires capabilities its pristine form never possessed. Function emerges from controlled damage.