1: \section{Introduction}
2: \label{1sec}
3: %
4: Quantum ergodicity describes the equidistribution of eigenfunctions
5: in the semiclassical limit ($\hbar\to 0$), in the sense that almost all
6: expectation values of quantum observables tend to a classical mean value of
7: the corresponding classical observable.
8: In particular, the Wigner functions associated with almost all eigenfunctions
9: of the Hamiltonian become equidistributed on the energy shell in phase space.
10: A sufficient condition for quantum ergodicity to hold
11: is ergodicity of the corresponding classical dynamics, as
12: was first stated by Shnirelman \cite{Shn74}. It is remarkable that quantum
13: ergodicity is one of the few results in quantum chaos for which there exist
14: mathematical proofs; the first complete ones were given by
15: Zelditch \cite{Zel87} and Colin de Verdi\`ere \cite{CdV85}. These
16: are results for scalar Hamiltonians, i.e. for the Laplacian on compact
17: manifolds or, more generally, for Schr\"odinger operators in $\rz^d$
18: \cite{HelMarRob87}. The case of quantum billiards was considered in
19: \cite{GerLei93,ZelZwo96}.
20:
21: If one wants to describe particles with internal degrees of freedom, such as
22: spin, one has to deal with matrix valued Hamiltonians. Since spin is a
23: purely quantum mechanical property, it is a priori not obvious what should
24: serve as the corresponding classical system. In \cite{BolGla00} quantum
25: ergodicity for Pauli Hamiltonians with spin 1/2 was proven under the condition
26: that a combination of the classical translational dynamics and a quantum
27: spin dynamics driven by the translational motion is ergodic. This kind of
28: mixed classical/quantum description in terms of a skew product dynamics
29: arises naturally in the semiclassical
30: analysis of Hamiltonians with spin \cite{BolKep98,BolKep99a,BolKep99b}.
31:
32: Although spin is a genuinely quantum mechanical property, there exists
33: an intuitive classical analogue given by the so-called vector model which
34: provides a picture of spin as a vector of fixed length whose dynamics
35: is similar to that of angular momentum. The mathematical description of this
36: model goes back to Thomas \cite{Tho27}, who gave an equation of motion for a
37: classical particle with a magnetic moment precessing in external fields.
38: In a semiclassical analysis of spinning particles various objects can be
39: related to this classical spin precession \cite{RubKel63,BolKep99a,Spo00}.
40: In the vector model the equations of motion are of first order in time,
41: and for pure spin precession can be formulated in a Hamiltonian framework.
42: Thus, after suitable normalisation of the spin vector, the unit sphere
43: $\Stwo$ serves as the phase space of classical spin precession.
44:
45: %In the vector model the phase space of a classical spin is the two-sphere,
46: %and pure spin precession can be described as a Hamiltonian flow on $\Stwo$.
47:
48: Based on ideas of Stratonovich \cite{Str57}, Gracia-Bond{\'\i}a and V\'arilly
49: \cite{GraVar88,VarGra89} developed a Wigner-Weyl calculus for arbitrary
50: spin $j$, in which quantum mechanical observables, i.e. hermitian
51: $(2j+1)\times(2j+1)$ matrices, are represented in terms of functions on
52: the phase space $\Stwo$. We use this description to rephrase the problem of
53: quantum ergodicity for non-relativistic particles with spin
54: and to prove for arbitrary spin the respective statement
55: under the condition that a different skew product of translational and spin
56: dynamics, where now the spin part is given by classical spin precession,
57: is ergodic.
58:
59: The paper is organised as follows. In section \ref{2sec} we briefly review
60: properties of the Weyl correspondence for systems with only either
61: translational or spin degrees of freedom. In section \ref{3sec} we then
62: discuss how in the semiclassical limit of systems with both kinds of
63: degrees of freedom the motion is governed by the skew product dynamics.
64: We also state a corresponding Egorov theorem. Section \ref{4sec} is
65: devoted to the proof of quantum ergodicity for Pauli Hamiltonians with
66: arbitrary spin. Some particular representations of the Wigner-Weyl transform
67: for spinors and a relation between ergodic properties of the two types of skew
68: products are discussed in two appendices.
69:
70:
71:
72:
73:
74:
75:
76: