Syncreticity with fractional 't Hooft instantons in the chiral phase transition
by
The order of the chiral phase transition in the plane of temperature, $T$, and chemical potential, $\mu$, continues to puzzle. I conjecture that for massless quarks, the chiral phase transition at $(T_\chi,\mu_\chi)$ is syncretic: terms which break the axial $U(1)_A$ symmetry are fractional powers of 't Hooft determinants in the broken phase (for $N_c$ colors, an integer times $1/N_c$) but involve only integral powers of 't Hooft determinants in the chirally symmetric phase; terms symmetric under $U(1)_A$ remain of the same form for all $T$ and $\mu$. For three flavors, $N_f =3$, the chiral transition is predicted to be weakly first order, and could well be very weakly. For $N_f \geq 1$, the chiral transition is generically "beyond Landau", as for $T$ and imaginary $\mu$. This can be tested, now, for lattice QCD with $2+1$ flavors. In the plane of $T$-$\mu$ plane the phase diagram consists of two semi-circles: first for the restoration of chiral symmetry, and later for "deconfinement".