A difficulty appears when introducing the temperature for a chiral-invariant color potential V(r) between quarks. The naive gap equation at finite temperature, obtained from a minimum principle and a Bogoliubov approximation applied to the free energy, has the following problems: (1) in the limit T\ensuremath{\rightarrow}0, it is inconsistent with the gap equation formulated directly at T=0; (2) it is not invariant under the transformation V(r)\ensuremath{\rightarrow}V(r)+const, as one would expect if the Hilbert space is restricted to color singlets, as required by confinement. These difficulties are solved if one requires the thermal excitations to be global color singlets, or, equivalently, if one restricts to the color singlets the trace in the calculation of the free energy. We obtain the corresponding gap equation in the infinite-volume thermodynamic limit. This equation has now a T\ensuremath{\rightarrow}0 limit that possesses the same ground-state solution as the zero-temperature gap equation. Moreover, in the case of a confining potential the same ground-state solution remains when we switch on the temperature, so that chiral invariance is not restored at any value of the temperature. Because translational invariance is assumed, particle-hole thermal excitations are constructed as color-singlet pairs of plane waves that, due to the confining interaction, possess infinite energy. The particle-hole pairs cannot be excited with a finite cost of energy, preventing chiral-symmetry restoration at any temperature.
Read full abstract