Abstract Real-time finite temperature Green's functions are discussed on the basis of the definition of new dynamical pictures. Causality appears explicitly. Feynman diagrams are formally identical to the ones of zero temperature. The vanishing of disconnected diagrams follows naturally.
摘要 本文从解析性的讨论出发,应用朗道关系改进Дзялошинский的图解法,建立了双时因果格林函数的费曼图解法,它具有正确的零温权限,并可直接用来计算各种非平衡过程的动力学性质。 Abstract Starting from the discussion of analytic properties and using the Landau relation to improve the diagrammatic technique of Dzyaloshinskii, the Feynman diagrammatic technique for double time dependent causal Green's function is established. There is a correct limit of zero temperature and may be used to calculate the various dynamical properties of nonequlibrium processes. 作者及机构信息 章思俊 1. 中国科学院 Authors and contacts CHANG SHIH-CHUN 1. 中国科学院 参考文献 [1] 施引文献
Abstract Analytic properties of the causal Green's function in statistical mechanics are investigated. It is shown that the causal function can be obtained as a certain limit of an analytic function. The dual field introduced by Umezawa and Takahashi in their field‐theoretical formulation of statistical mechanics (thermo field dynamics), plays an important role to make the spectral functions real. In a few examples, it is shown explicitly that the Feynman diagram technique is applicable in the thermo‐field dynamics.
We study the dynamical equation of the time-ordered Green's function at finite temperature. We show that the time-ordered Green's function obeys a conventional Dyson equation only at equilibrium and in the limit of zero temperature. In all other cases, i.e. finite temperature at equilibrium or non-equilibrium, the time-ordered Green's function obeys instead a modified Dyson equation. The derivation of this result is obtained from the general formalism of the non-equilibrium Green's functions on the Keldysh time-loop contour. At equilibrium, our result is fully consistent with the Matsubara temperature Green's function formalism and also justifies rigorously the correction terms introduced in an ad hoc way with Hedin and Lundqvist. Our results show that one should use the appropriate dynamical equation for the time-ordered Green's function when working beyond the equilibrium zero-temperature limit.