Fractional anomalous determinants and the chiral phase transition

Abstract

At high temperature instantons form a dilute gas, so in QCD-like theories the breaking of the anomalous $U(1)_A$ symmetry is given by integral powers of the anomalous determinant, $\sim (\det \Phi)^{Q}$, where $\Phi\sim \overline{q}_L q_R$ is bilinear in the quark fields, and with untwisted boundary conditions, the topological charge, $Q$, is an integer. A syncretic model is constructed, which is manifestly"beyond Landau". In the chiral limit, at temperatures above the chiral phase transition, $T>T_\chi$, only integral powers of the anomalous determinant appear. Below $T_\chi$, following't Hooft et al. I assume that the topological charge $Q$ is fractional, as an integer times $1/N_c$, where $N_c$ is the number of colors. I suggest that consequently, fractional powers of the anomalous determinant appear in the chiral effective Lagrangian. For $N_f$ degenerate flavors, this generalizes the Witten-Veneziano term, valid for small $N_f/N_c$, to arbitrary $N_f/N_c$. In this model the chiral phase transition is generically of second order. The two exceptions are for one flavor, where it is probably crossover, and three flavors, where it could well be weakly first order. This can be tested in lattice QCD with $2+1$ flavors by comparing the (known) temperature dependence of the difference of the $\pi^a$ and $a_0^a$ propagators, to the chiral condensate of the strange quark, between $T_\chi$ and $\sim 2 \, T_\chi$. Analogous measurements are possible for one to four degenerate flavors about $T_\chi$. Lastly, I propose an operator for baryon number in the symmetric phase.