In this chapter, the authors consider the formal description of the response and of the correlations displayed by the nuclear many-particle system at finite temperature. They work out the theoretical tools which allow to describe mean field properties, of both single-particle and collective motion, as well as their interweaving in highly excited systems. The authors show that the analysis of the response of the system to a weak external probe, and the treatment of correlations proceed in much the same way as they did at zero temperature. There is, however, a main difference to be kept in mind. At finite temperature, one no longer finds a unique state which describes the system, and a statistical description is called for. To study a quantum mechanical system, it is generally convenient to use the grand canonical ensemble, for which the energy of the system and the number of particles are not required to have definite values but only particular average values.
In this chapter, the authors consider the formal description of the response and of the correlations displayed by the nuclear many-particle system at finite temperature. They work out the theoretical tools which allow to describe mean field properties, of both single-particle and collective motion, as well as their interweaving in highly excited systems. The authors show that the analysis of the response of the system to a weak external probe, and the treatment of correlations proceed in much the same way as they did at zero temperature. There is, however, a main difference to be kept in mind. At finite temperature, one no longer finds a unique state which describes the system, and a statistical description is called for. To study a quantum mechanical system, it is generally convenient to use the grand canonical ensemble, for which the energy of the system and the number of particles are not required to have definite values but only particular average values.
This chapter aims to present the essence of the general theory in the particularly simple and convenient case of systems at finite temperature which may be treated in the grand canonical ensemble. It describes how thermal averages of time-ordered products of Heisenberg operators in imaginary time emerge naturally from path integrals and how to evaluate them in perturbation theory. The chapter considers at the outset what level of mathematical rigor one can actually expect and how in practice one may select and assess an approximation scheme. In view of the general lack of mathematical control on the convergence properties of the original series and the obvious ambiguity associated with regrouping divergent series, any such resummations must be understood ultimately on physical grounds. The ultimate objective of the quantum theory of many-particle systems is to understand experimentally observable properties of a diverse range of physical systems.
In the first three chapters we have considered problems related to the physical behaviour of one or a few particles (or some other physical objects which can be effectively described by one or a few (quasi)particles as, e.g., the random walk model in polymer physics considered in chapter 1). Though quantum field theory describes systems with an arbitrary number of degrees of freedom (an arbitrary number of (quasi)particles), in chapter 3 we actually applied it to systems with a restricted and small number of particles (e.g., for the description of the scattering process for a few particles). Another way to express this fact is to say that in chapter 3 we have considered field theories at zero temperature. However, the majority of realistic systems contain many identical (indistinguishable) particles such as atoms, electrons, photons etc. An attempt to describe these systems in terms of the individual trajectories of all particles is absolutely hopeless. Instead, we are interested in the collective behaviour of systems and describe them in terms of partition functions, mean values, correlation functions, etc. The methods of derivation, analysis and calculation of such collective characteristics constitute the subject of classical and quantum statistical mechanics.
When studying a many-body system, we rarely try to compute the N -particle wavefunction Ψ( r 1 , … , r N ). This calculation is usually hardly feasible, and we do not need all the information which is contained in the wavefunction. A reduced description in terms of a few variables is often sufficient if it allows one to explain the response of the system to an experimental probe. Correlation functions offer such a description. In Sec. 3.1, we show that the linear response of the system to an external field is given by a correlation function. In Sec. 3.2, we study the basic properties of correlation (or response) functions and show that they contain information about the excitations of the system. Furthermore, they are simply related to the imaginary-time Green functions introduced in Sec. 1.2.3 and can therefore be calculated perturbatively using the methods discussed in Chapter 1. In Secs. 3.3 and 3.4, we study the density–density and electromagnetic response functions. From the properties of the latter, we deduce a number of general features of electron systems (metals and superconductors). The one-particle Green function is just one type of correlation function and is studied in Sec. 3.5. Finally, in Sec. 3.6, we consider systems with a spontaneous broken symmetry and derive two important theorems due to Goldstone, and Mermin and Wagner.
Preparation, I have often said, is rightly two-thirds of any venture. Amelia Earhart Summary Correlation functions provide a direct way to characterize and analyze many-body systems, both theoretically and experimentally. In this chapter we review the properties of one- and two-body correlation functions in quantum systems, with emphasis on several key quantities: static density correlations that determine the energy and thermodynamic potentials, dynamic correlation functions such as response functions that describe excitations of the system, and Green's functions that are basic tools in the theory of interacting many-body systems. Correlation functions are central quantities in the description of interacting many-body systems, both in the theoretical formulation and in the analysis of experiments. In contrast to single numbers like the total energy, correlation functions reveal far more information about the electrons, how they arrange themselves, and the spectra of their excitations. In contrast to the many-body wavefunctions that contain all the information on the system, correlation functions extract the information most directly relevant to experimentally measurable properties. Dynamic current–current correlation functions are sufficient to determine the electrical and optical properties: one-body Green's functions describe the spectra of excitations when one electron is added to or removed from the system, static and dynamic correlations are measured using scattering techniques, and so forth. In this chapter we present the basic definitions and properties of correlation functions and Green's functions that are the basis for much of the developments in the following chapters. In general, a correlation function quantifies the correlation between two or more quantities at different points in space r , time t , or spin σ . Very often the correlation function can be specified as a function of the Fourier-transformed variables, momentum (wavevector) k , and frequency ω . It is useful to distinguish between a dynamic correlation function which describes the correlation between events at different times and a static or equal-time correlation function, by which we mean that of a property measured or computed with “snapshots” of the system. Also, the different correlation functions can be classified by the number of particles and/or fields involved.
A number of powerful methods exist in statistical mechanics to describe the interweaving of the variety of elementary modes of excitation taking place in a quantal many-body system. Among them, one can mention the Static Path approximation (cf., e.g., Feynmann and Hibbs ( 1965 )), the Monte-Carlo Shell Model Method (Koonin et al. ( 1997 )), and the thermal Green-function formalism (cf., e.g., Mahan ( 1981 )). The first two approaches have not yet been systematically used to describe nuclear motion at finite temperature. We shall briefly discuss them in the present chapter. The Matsubara techniques, which have been applied to the study of a large variety of phenomena, are discussed in Ch. 9 .
This chapter introduces the reader to the proper language in which to express with precision results of both theory and experiment concerning the response of a many-particle system to macroscopic and microscopic external probes. It shows why the Landau theory is rigorous in the long wavelength limit: as we pointed out earlier, the Landau theory is equivalent to neglecting multipair excitations in the calculation of the response function. The chapter suggests that the Landau theory succeeds because in the limit of small q and ω, the contribution of multipair excitations is usually negligible. By studying the structure of the linear response function, one may describe explicitly the transition from the collisionless high-frequency regime (where the system response is essentially microscopic), to the low-frequency hydrodynamic regime (where the response is governed by macroscopic average laws). The chapter considers a simple physical example of probe-liquid interaction, the inelastic scattering of an external particle beam by a neutral quantum liquid.