Effect of site dilution on the negative-coupling branch of the Ashkin–Teller critical line

Determine whether site dilution alters the clean Ashkin–Teller critical behavior for negative four-spin coupling $J_4<0$, where the clean model has correlation-length exponent $ u>1$, by obtaining more precise estimates of the polarization exponent $[?]$ and correlation-length exponent $ u$.

Background

For the site-diluted square lattice with occupation probability p=0.75p=0.75, the authors find clear evidence that uncorrelated disorder changes the critical behavior on the J4>0J_4>0 branch, where the clean Ashkin–Teller model has u[?][?]1u[?][?]1. In contrast, for J4<0J_4<0, the clean model has u>1u>1, so both the Harris and HBV criteria predict irrelevance of disorder. The measured [?] and [?] are broadly compatible with clean behavior, but the measured polarization exponent differs from its clean Coulomb-gas prediction by slightly more than twice the estimated uncertainty. The paper therefore leaves unresolved whether site dilution genuinely changes the critical behavior in this regime.

References

More precise estimates of \eta_p and \nu are therefore required to determine whether the clean critical behavior is altered by site dilution.

— Numerical Study of Stability of Clean Critical Points across Aperiodic, Topological, and Uncorrelated Disorder  (2609.17166 - Fan et al., 15 Sep 2026) in Section 3.3, Critical behavior on the diluted lattice