We discuss the imaginary chemical potential approach to the study of QCD at nonzero temperature and density, present results for the four flavor model in the different phases and show that this method is ideally suited for a comparison between lattice data and phenomenological models.
QCD and a complex µ BResults from simulations with an imaginary chemical potential can be analytically continued to a real chemical potential, thus circumventing the sign problem [1], [2] [3]. In practice, the analytical continuation is carried out along one line in the complex µ plane: first along the imaginary axes, and then along the real one. It is then meaningful to map this path in the complex µ 2 plane: because of the symmetry property Z(µ) = Z(−µ) this can be achieved without losing generality. In the complex µ 2 plane the partition function is real for real values of the external parameter µ 2 , complex otherwise: the situation resembles that of ordinary statistical models in an external field. Hence, the analyticity of the physical observables [4] as well as that of the critical line [2] follows naturally.The phase diagram in the temperature, (real) µ 2 plane is sketched in Fig.1, where we omit the superconducting and the color flavor locked phase, which (unfortunately) play no rôle in our discussion. The region accessible to numerical simulations is the one with µ 2 ≤ 0: at a variance with other approaches to finite density QCD which only use information at µ = 0 [5] [6] [7] the imaginary chemical potential method exploits the entire halfspace. And we will also argue that there are physical questions which can be addressed without analytic continuation. * Electronic address: delia@ge.infn.it † Electronic address: lombardo@lnf.infn.it It encodes reality for real µ 2 , contains the physical scale T c , is dimensionally consistent, gives T (µ = 0) = T c , T (µ = 0) < T c . We refer to Section IV of Ref.[3] for our results on the critical line in the four flavor model, and their discussion in terms of model calculations. Here it suffices to remind ourselves that the second order approx-