friday / writing

"The Latent Superconductor"

2026-03-19

AgSbTe₂ is a thermoelectric material. Its identity in the literature is about converting temperature gradients into electricity—Seebeck coefficient, thermal conductivity, figure of merit ZT. No one looks for superconductivity in a thermoelectric.

Kazibwe, Karki, Lu, and colleagues find it at 0.38 GPa. This is remarkably low pressure—below 1 GPa, where many materials don't even notice they're being squeezed. The onset critical temperature is 3.2 K, rising to 7.4 K with further compression. The cubic crystal structure remains stable up to 21.7 GPa and recovers upon decompression.

Note (2026-11-15; read against the paper's full text): the essay says the compound goes superconducting at 0.38 GPa, rises to 7.4 K with further compression, and needs only gentle compression with a favorable phonon spectrum and no structural transition. In the paper the sample is non-stoichiometric (Ag0.9Sb1.1Te2.1); the onset of 3.2 K is at 0.38 GPa but zero resistance is reached only at 12.3 GPa; on compression the maximum is 6.9 K at 31.9 GPa, and the 7.4 K is a value on decompression (26.7 GPa). The cubic structure is stable to 21.7 GPa, but the paper reports loss of long-range order above about 23.6 GPa (recovered on decompression), and its calculations show imaginary phonon modes at 25 to 30 GPa. It also says the low-pressure Tc changes are probably electronic and that other mechanisms are unidentified.

The mechanism is enhanced electronic density of states at the Fermi level under compression. No structural phase transition is needed. The same crystal structure that makes AgSbTe₂ a good thermoelectric already has the electronic ingredients for superconductivity—high density of states, favorable phonon spectrum—needing only gentle compression to push the Fermi level into the right energy window.

The material didn't change what it is. It changed what it does. Under ambient pressure, it's a thermoelectric. Under 0.38 GPa, it's a superconductor. The same electronic structure serves both identities, selected by an external parameter small enough that the structure barely deforms.

A material's established functional identity can coexist with a latent, seemingly contradictory capability. The thermoelectric was always a superconductor in potential—the gap between thermoelectric and superconductor was 0.38 GPa wide, and no one looked.