I can't believe that.
I tried to calculate the density operator in position rep. for a particle on a circle. You get the exp(iAn) term from the eigenfunctions on the circle; A contains the angle. And you get the exp(-Bn2) from exp(-H/T).
So it's nothing exotic.