Quantum statistical mechanics 1st Edition William C. Schieve pdf download
Quantum statistical mechanics 1st Edition William C. Schieve pdf download
https://ebookgate.com/product/quantum-statistical-mechanics-1st-
edition-william-c-schieve/
https://ebookgate.com/product/heisenberg-s-quantum-mechanics-1st-
edition-mohsen-razavy/
https://ebookgate.com/product/non-hermitian-quantum-
mechanics-1st-edition-nimrod-moiseyev/
https://ebookgate.com/product/quantum-wave-mechanics-third-
edition-larry-reed/
https://ebookgate.com/product/group-theory-and-quantum-mechanics-
michael-tinkham/
Introduction to Quantum Mechanics David J. Griffiths
https://ebookgate.com/product/introduction-to-quantum-mechanics-
david-j-griffiths/
https://ebookgate.com/product/do-we-really-understand-quantum-
mechanics-laloe/
https://ebookgate.com/product/quantum-information-theory-and-the-
foundations-of-quantum-mechanics-1st-edition-christopher-g-
timpson/
https://ebookgate.com/product/quantum-mechanics-concepts-and-
applications-1st-edition-nouredine-zettili/
https://ebookgate.com/product/abc-s-of-quantum-mechanics-1st-
edition-v-rydnik/
This page intentionally left blank
QUANTUM STATISTICAL MECHANICS
www.cambridge.org
Information on this title: www.cambridge.org/9780521841467
© W. Schieve and L. Horwitz 2009
Preface page xi
1 Foundations of quantum statistical mechanics 1
1.1 The density operator and probability 1
1.2 The Gleason theorem and consequences 6
1.3 Calculation of averages of observables 9
Appendix 1A: Gleason theorem 12
References 18
2 Elementary examples 19
2.1 Introduction 19
2.2 Harmonic oscillator 19
2.3 Spin one-half and two-level atoms 27
Appendix 2A: the Fokker–Planck equation 34
References 35
3 Quantum statistical master equation 37
3.1 Reduced observables 37
3.2 The Pauli equation 39
3.3 The weak coupling master equation for open systems 42
3.4 Pauli equation: time scaling 46
3.5 Reservoir states: rigorous results and models 53
3.6 The completely positive evolution 54
Appendix 3A: Chapman–Kolmogorov master equation 57
References 59
4 Quantum kinetic equations 61
4.1 Introduction 61
4.2 Reduced density matrices and the B.B.G.Y.K. hierarchy 61
4.3 Derivation of the quantum Boltzmann equation 63
4.4 Phase space quantum Boltzmann equation 66
4.5 Memory of initial correlations 76
v
vi Contents
This book had its origin in a graduate course in statistical mechanics given by
Professor W. C. Schieve in the Ilya Prigogine Center for Statistical Mechanics at
the University of Texas in Austin.
The emphasis is quantum non-equilibrium statistical mechanics, which makes
the content rather unique and advanced in comparison to other texts. This was
motivated by work taking place at the Austin Center, particularly the interaction
with Radu Balescu of the Free University of Brussels (where Professor Schieve
spent a good deal of time on various occasions). Two Ph.D. candidate theses at
Austin, those of Kenneth Hawker and John Middleton, are basic to Chapters 3
and 4, where the master equations and quantum kinetic equations are discussed.
The theme there is the dominant and fundamental one of quantum irreversibil-
ity. The particular emphasis throughout this book is that of open systems, i.e.
quantum systems in interaction with reservoirs and not isolated. A particularly
influential work is the book of Professor A. McLennan of Lehigh University,
under whose influence Professor Schieve first learned non-equilibrium statistical
mechanics.
An account of relatively recent developments, based on the addition in the
Schrödinger equation of stochastic fluctuations of the wave function, is given in
Chapter 13. These methods have been developed to account for the collapse of the
wave function in the process of measurement, but they are deeply connected as
well with models for irreversible evolution.
The first six chapters of the present work set forth the theme of our book, par-
ticularly extending the entropy principle that was first introduced by Boltzmann,
classically. These, with equilibrium quantum applications (Chapters 7, 8, 9 and
possibly also Chapters 14 and 15), represent a one-semester advanced course on
the subject.
xi
xii Preface
functions. All of this requires a more advanced mathematical approach than the
earlier discussions in this book. However, it is necessary that a well-grounded
student of quantum mechanics know these things, as well as acquire the mathe-
matical tools, and therefore it is very appropriate here in a discussion of quantum
statistical mechanics.
Chapter 18 is in many ways an extension of Chapter 17. It is an outline of what
has been called extended statistical mechanics. Ilya Prigogine and his colleagues
in Brussels and Austin, in the past few years, have attempted to formulate many-
body dynamics which is intrinsically irreversible. In the classical case this may
be termed the complex Liouville eigenvalue method. As an example, the Pauli
equation is derived again by these nonperturbative methods. This is not an open-
system dynamics but rather, like the previous Chapter 17 discussion, one of closed
isolated dynamics. This effort is not finished, and the interested student may look
upon this as an introductory challenge.
The final chapter of this book is in many ways a diversion, a topic for personal
pleasure. The remarkable objects of our universe known as black holes apparently
exist in abundance. These super macroscopic objects obey a simple equilibrium
thermodynamics, as first pointed out by Bekenstein and Hawking. Remarkably,
the area of a black hole has a similarity to thermodynamic entropy. More remark-
able, the S-matrix quantum field theoretic calculation of Hawking showed that the
baryon emission of a black hole follows a Planck formula. Hawking introduced a
superscattering operator which is analogous to the extended dynamical theory of
Chapter 18.
To complete these comments, we would like to thank Florence Schieve for sup-
port and encouragement over these last years of effort on this work. She not only
gave passive help but also typed into the computer several drafts of the book as well
as communicating with the coauthor and the editorial staff of the publisher. The
second coauthor wishes also to thank his wife Ruth for her patience, understanding,
and support during the writing of some difficult chapters.
We also acknowledge the help of Annie Harding of the Center here in Austin.
Three colleagues at the University of Texas—Tomio Petrosky, George Sudarshan
and Arno Bohm—also made valuable technical comments. WCS also thanks the
graduate students who, over many years of graduate classes, made enlightened
comments on early manuscripts.
We recognize the singular role of Ilya Prigogine in creating an environment in
Brussels and Austin in which the study of non-equilibrium statistical mechanics
was our primary goal and enthusiasm.
Finally, WCS thanks the Alexander von Humboldt Foundation for making pos-
sible extended visits to the Max Planck Institute of Quantum Optics in Garching
and later in Ulm. LPH thanks the Center for Statistical Mechanics and Complex
xiv Preface
Systems at the University of Texas at Austin for making possible many visits over
the years that formed the basis for his collaboration with Professor Schieve, and the
Institute for Advanced Study at Princeton, particularly Professor Stephen L. Adler,
for hospitality during a series of visits in which, among other things, he learned of
the theory of stochastic evolution, and which brought him into proximity with the
University of Texas at Austin.
1
Foundations of quantum statistical mechanics
1
2 Foundations of quantum statistical mechanics
σ z , the state is represented by a vector, and the measurement of the spin in the z-
direction can yield probability one. For a general choice of γ ± , there is no vector
that can represent the state. In the first case the state is called pure, and it can
be represented by a projection into a one-dimensional subspace (in the previous
example, Pσ z = |σ z σ z |). This is equivalent to specifying the vector, up to a
phase, corresponding to the one-dimensional subspace. In the second case, it is
called mixed and does not correspond to a vector in the Hilbert space.
It is clear from the discussion of these examples that the a priori probabilities γ ±
are essentially classical, reflecting the composition of the beam that was prepared
in the macroscopic laboratory.
Although a density operator ρ of the type that we have defined in this exam-
ple appears to be a somewhat artificial construction, it is actually a fundamental
structure in quantum statistical mechanics (Dirac, 1958). It enables one to study a
complex system in the framework of an ensemble and in fact occurs on the most
fundamental level of the axioms of the quantum theory.
It was shown by Birkhoff and von Neumann (1936) that both quantum mechan-
ics and classical mechanics can be formulated as the description of a set of
questions for which the answer, as a result of experiment, is “yes” or “no.” Such a
set, which includes the empty set φ (questions that are absurd, e.g. the statement
that the system does not exist) and the trivial set I (the set of all sets, e.g. the state-
ment that the system exists), and is closed with respect to intersections and unions,
is called a lattice. A lattice that satisfies the distributive law
a ∩ (b ∪ c) = (a ∩ b) ∪ (a ∩ c) ,
where ∪ represents the union and ∩ the intersection, is called Boolean. These oper-
ations have the physical meaning of “or” (the symbol ∪), in which one or the other
of the propositions is true, and “and” (the symbol ∩), for which both must be true
for the answer of the compound measurement to be “yes.” An example of such a
lattice may be constructed in terms of two-dimensional closed regions on a piece
of paper. This is discussed again in the appendix to this chapter.
Both classical and quantum theories may be associated with lattices in terms,
respectively, of the occupancy of cells in phase space or states in the subspaces of
the Hilbert space. The questions a correspond, in the first case, to the phase space
cells (with answer corresponding to occupancy) and in the second to the projec-
tion operators Pα associated with a subspace Mα , with the answer corresponding
to the values ±1 which a projection operator can have. These values correspond
to evaluating the projection operator on vectors which lie within or outside the
subspace.
Birkhoff and von Neumann asserted that the fundamental difference between
classical and quantum mechanics is that the lattices corresponding to classical
4 Foundations of quantum statistical mechanics
mechanics are Boolean, and those corresponding to quantum mechanics are not.
The non-Boolean structure of the quantum lattice is associated with the lack of
commutativity of the projection operators associated with different subspaces:
a ∩ (b ∪ c) = (a ∩ b) ∪ (a ∩ c) . (1.1)
that is, this set of propositions is not Boolean. The construction is interesting in
that it illustrates the special structure of the topology of Hilbert spaces as well as
the notion of the non-Boolean lattice.
We start by constructing the union of the manifolds Mx and M y by their joint
linear span. Taking the standard definition of the Pauli matrices,
0 1 0 −i 1 0
σx = , σy = , σz = ,
1 0 i 0 0 −1
the projection operators into the subspaces with positive eigenvalues are
1 1 1 1
Px = (1 + σ x ) =
2 2 1 1
1 1 1 −i
Py = 1+σy =
2 2 i 1
1 1 1 0
Pz = (1 + σ z ) = .
2 2 0 0
The corresponding eigenvectors are given by projecting a generic vector v into the
respective subspaces. For
v
v= 1 ,
v2
1.1 The density operator and probability 5
The union of the subspaces Mx and M y is the closed linear span of vectors in both
subspaces. By taking the combination vx + iv y , it is easy to see that the vector vz
(and hence the subspace Mz ) is contained in Mx ∪ M y . To construct the distributed
operation
(Mz ∩ Mx ) ∪ Mz ∩ M y ,
we must use the construction for which the projection operator corresponding to
the intersection of two noncompatible subspaces is generated by an alternating
succession of projections into the two subspaces (Jauch, 1968). The products Pz Px
and Pz Py are, it so happens, idempotents up to coefficients less than one, i.e.
1 1 1
Pz Px =
2 0 0
1 1 1
(Pz Px ) =
2
4 0 0
6 Foundations of quantum statistical mechanics
and
1 1 −i
Pz Py =
2 0 0
2 1 1 −i
Pz Py = ,
4 0 0
n
which implies that both (Pz Px )n and Pz Py go to zero as n → ∞. Therefore,
Mz ∩ Mx = Mz ∩ M y = 0.
Clearly,
Mz ∩ Mx ∪ M y = (Mz ∩ Mx ) ∪ Mz ∩ M y .
Although Pz Px and Pz Py are not zero (the two corresponding vectors are not
orthogonal), the closed subspace that is common is empty. One can think of this
geometrically in terms of two lines that have some projection on the other, but
the intersection of the two lines is just a point of zero measure. Physically, this
implies that we cannot have a definite statement of the joint values of σ z and σ x or
σ y . The noncommutativity of the associated projections is essential; if they were
commutative, the product of projections would be a projection, and the products
would not converge to zero. It is clear from this example that compatible subspaces
would satisfy Boolean distributivity.
We shall later discuss the Wigner function, which appears to provide joint distri-
butions over noncommutative variables such as q and p; however, these functions
are not probabilities, since, although they are the coefficients of what might be
called the Weyl basis for the operator algebra of the quantum theory which appear
in expectation values, they are not positive (Wigner, 1936).
w : P → [0, 1]
1.2 The Gleason theorem and consequences 7
satisfies
The first follows from the fact that the sum over all disjoint a of w (a) is the total
probability measure on the set of all questions (and the sum over all disjoint Pa is
8 Foundations of quantum statistical mechanics
the unit operator). The second follows from the first; all eigenvalues of ρ are real
and positive with values less than or equal to unity. With these properties, one can
prove that the spectrum of ρ must be completely discrete.
Mackey (1963) has given a converse theorem. If the function w (a) can reach the
value unity on a one-dimensional subspace of the Hilbert space, the corresponding
density operator is just a projection into this one-dimensional subspace and can
be put into correspondence (up to a phase) with the vector of the Hilbert space
generating this one-dimensional subspace. Such a state is called pure. A state which
cannot reach the value of unity on any one-dimensional subspace is called mixed.
The proof is very simple. Let P0 be the projection onto a one-dimensional sub-
space generated by the vector φ 0 , and let us use the representation, taking into
account the discrete spectrum of ρ,
ρ= γ i ψ i ψ i . (1.6)
i
Here we use the Dirac ket ψ i to signify an element of the Hilbert space. Then if
Trρ P0 = 1, it follows that
Trρ (1 − P0 ) = 0,
or
γ i ψ i (1 − P0 ) ψ i = Tr γ i (1 − P0 ) ψ i
2
Tr = 0,
i i
for all i. Substituting into Eq. (1.6), we see that in this case we must have
ρ= γ i |λi |2 φ 0 φ 0 .
i
Furthermore, if the ψ i and φ 0 are normalized, |λi |2 = 1. Then, by Eq. (1.5) and
Eq. (1.6) (for the ψ i orthogonal), one sees that the sum of the γ i is unity; hence
ρ = φ 0 φ 0 ,
which is the projection operator into the subspace generated by φ 0 . This theorem
therefore identifies the pure states with vectors of the Hilbert space, and it is for this
reason that one often calls the vectors of the Hilbert space “states.” Every vector in
the Hilbert space corresponds to a pure state.
1.3 Calculation of averages of observables 9
equal to the classical probabilities of the occurrence of each quantum state in the
ensemble, i.e.
A = γ i ψ i A ψ i .
i
where
ρ= γ i ψ , ψ i . (1.10)
i
A = ak Tr (ρ Pk ) , (1.11)
k
where
Tr (ρ Pk ) = γ i ψ i Pk ψ i (1.12)
i
2
= γ i ψ i | φk
i
is the probability of finding the system in the subspace associated with Pk . This
probability is composed of two types of expectation: the quantum probability to
find the Pk in each state ψ i , and the classical probability for the occurrence of the
state ψ i (determined by the relative number of subsystems in that state).
The results that we have given can easily be extended to the most general case
of an observable with both discrete and continuous spectra without change in the
formal structure, although as we shall see later, there are special technical aspects
that arise in the continuous case (for example, in scattering theory). To see this,
we use the spectral representation theory of von Neumann. It was shown by von
Neumann (1955) that every self-adjoint operator A, corresponding to a physical
observable, has a spectral representation of the form
A= a d E (a) , (1.13)
where a takes on a continuous set of values (the real line), and the self-adjoint set
of operators E (a) is called a “spectral family.” It satisfies the property
with E (−∞) = 0 and E (∞) = I . It easily follows from these properties that
d E (a) , if a = b;
d E (a) d E (b) = (1.15)
0, otherwise
where a and b now refer to names given to infinitesimal intervals along the line
(i.e. for a small, d E (a) = E (a + a) − E (a)). The integral Eq. (1.13) is con-
sidered to be of Stieltjes–Lebesgue type, in the sense that if the weight function
ψ |d E (a)| ψ = d E (a) |ψ2 has a jump discontinuity at some point a0 , the
integral is evaluated as the difference between the values of E (a) |ψ2 above
and below the point a0 . If, in particular, d E (a) |ψ2 is zero in the neighbor-
hood of the point a0 (except at the point itself), so that the jump is isolated, one
obtains a contribution to any expectation value of A just from the point a = a0 (in
this neighborhood). The coefficient, since E (a)2 = E (a), is ψ| E (a0 + ε) −
E (a0 − ε) |ψ, where ε is infinitesimal. The operator E (a0 + ε) − E (a0 − ε)
may then be identified with one of the discrete projection operators appearing
in Eq. (1.7). Hence, the representation Eq. (1.11) includes both discrete and
continuous spectra. In Eq. (1.8) one then uses
2
ψ i A ψ i = ad E (a) ψ i ,
can describe the evolution of a density matrix for a pure state into a density matrix
corresponding to a mixed state. (For this more general evolution, Tr(ρ ρ̇) does not
vanish.)
Although, as we have previously emphasized, the density operator might appear
to be a somewhat artificial construction, combining both classical and quantum
probability notions to achieve an overall expectation value, it actually arises on
the most fundamental level of the quantum theory. Methods for the construc-
tion and study of this operator and its time evolution are the essential goal of
the techniques of statistical mechanics; the theory is constructed on this basic
foundation.
Good general references to the topics of this chapter are the books of Tolman
(1938), Dirac (1958), Farquahar (1964), Landau and Lifshitz (1970), Balescu
(1975), Dvurecenskij (1993), and Huang (1987). Extensive pertinent references
are given at the ends of later chapters.
represented. The closed subspaces of a Hilbert space, with their associated projec-
tion operators, form a set subject to the operations of intersection and union, and
contain the empty set and the set of all subsets, i.e. a structure called a lattice, iso-
morphic to the lattice of propositions (Birkhoff and von Neumann, 1936; Birkhoff,
1961; Piron, 1976), as mentioned earlier. For an irreducible proposition system,
in which there is only one minimal proposition (no superselection rules), every
self-adjoint operator corresponds to an observable. Let P (H ) be such a Hilbert
realization.
We now state the Gleason theorem (Gleason, 1957) (see Piron, 1976, for the
general case of a family of Hilbert spaces, for which there is a nontrivial set of
minimal propositions):
Theorem: Given a propositional system L = P (H ), where H is a Hilbert space
(of dimension ≥ 3) over the reals, complex numbers or quaternions, there exists
a unique function w ( p, b) defined on the atoms p (corresponding to the one-
dimensional subspaces of H ) and the propositions b of L which satisfies (as in
Eq. (1.2) and Eq. (1.3))
(1) 0 ≤ w ( p, b) ≤ 1 (1A.1)
(2) p ⊂ b ⇔ w ( p, b) = 1
(3) b ⊥ c ⇒ w ( p, b) + w ( p, c) = w ( p, b ∪ c) .
We begin the proof by noting that there is a vector f p in H , associated with the
atom p, satisfying
2
f p | f p = f p = 1.
Each proposition b in P (H ) can be represented by a projection operator Q into a
linear closed subspace of H . Then
w ( p, b) = f p |Q| f p
satisfies the conditions of the theorem.
Our principal task is then to show uniqueness. If there were another function
w ( p, b) satisfying these conditions, it would have to have a different value on
some pair p, b. For such functions, there would be another proposition q (an atom)
for which, in this case, w ( p, q) has a different value. However, if the function were
unique, the value would necessarily be the same. Such a q can be constructed as
follows. Note that
p ∪ b ∩ b ∪ p ∩ b = b
and that, since p and p are orthogonal,
p ∪ b ∩ b ⊥ p ∩ b.
14 Foundations of quantum statistical mechanics
However, w p, p ∩ b = 0, so
q = p ∪ b ∩ b (1A.2)
for an atom. The other function would, by construction, have a different value
for w ( p, q). We choose the two vectors f p and f q in such a way that f p | f q
is real. We may then consider just three vectors associated with the atoms p, q,
i.e. f p , f q and a vector (real) orthogonal to these. The restriction of w ( p, b) to
the three-dimensional real Hilbert subspace generated by f p , f q and a third vector
orthogonal to these still satisfies the conditions of the theorem. To complete the
the uniqueness of w in the case of the real three-
proof, it is then sufficient to prove
3
dimensional Hilbert space R . This construction, therefore, has the minimum
dimension necessary to carry out a proof of uniqueness.
To carry out the proof, let us assume p in w ( p, b) to be fixed. The lattice of
subspaces of R 3 is then the points and lines of the projective plane realized as the
intersection of R 3 with the tangent plane at p to the unit sphere. In the same way
as the complex plane is mapped onto the unit sphere including the point at infinity,
we are considering the plane as a (projective) representation of the sphere of unit
vectors in R 3 . (It may be helpful for the reader to draw his own diagrams for the
construction described here.)
We seek a unique function w(q), where we drop reference to p, now fixed,
defined at the points q of the plane which has the value 1 at p and 0 at the point(s)
at infinity.
If q lies on some arbitrary line L in the plane, then w (q) takes on a maximal
value at a point q0 where the line pq0 is perpendicular to the line L. This follows
from the fact that if q is a point on L, and q is its orthogonal complement on
L , q ∪ q on the line is just q0 . Hence, by (3) of Eq. (1A.1),
w (q) + w q = w (q0 )
or w (q0 ) ≥ w (q) .
We now note that w (q) decreases along the line L . To see this, consider a point
at q and a line L q perpendicular to pq. Move along this line to q1 ; we know by the
foregoing argument that
w (q) ≥ w (q1 ) .
Now erect a line at q1 perpendicular to pq1 and move to a point on this new line, r.
Clearly,
w (q1 ) ≥ w (r ) .
Now put another line at this point r , and connect it back to L q at the point q2 . Since
w (r ) ≥ w (q2 )
Appendix 1A: Gleason theorem 15
for 0 ≤ x ≤ 1
2
(i.e. λ > 1), and for a second relation,
1 − f (x) = f (1 − x) . (1A.6)
Lemma 2: If w (q) is continuous, then its value depends only on the angle between
the rays p and q.
The remaining two lemmas (lemmas 3 and 4) prove continuity.
To prove this lemma, let q and r be two points on the projective plane situated at
the same distance from p. To prove that w (q) = w (r ), we start by proving that for
any q0 ∈ qp sufficiently close to q, the signs of w (q0 ) − w (r ) and λ − λ0 , where λ
and λ0 are the distances pq and pq0 respectively, are the same. If λ > λ0 , we can
join q0 to r by a sequence q0 , q1 , q2 , ... of sequentially perpendicular steps, since
at each step λ1 ≥ λ0 , λ2 ≥ λ1 , ... up to r , which reaches λ, by construction (note
that we started with λ0 < λ). Then
since the lengths increase at every step. But we can take q0 arbitrarily close to q.
The same set of inequalities can be established in the other direction, starting with
a point r0 on pr , and hence w (q) = w (r ); i.e. the value of w (q) depends only on
the distance between p and q (the angle).
Lemma 3: If w (q) is continuous at some point q0 , then it is continuous at every
point.
We first show that if w (q) is continuous at q0 , it is continuous at each point q1
orthogonal to q0 . Then q0 and q1 lie symmetrically on both sides of the point of
a line from p perpendicular to q0 q1 . Denote an ε neighborhood of q0 by U , and
take a point q on the line q0 q1 in U ; further, consider the point q on the line q0 q1
orthogonal to q . As we have done before, we use the relations
w (q) + w q = w (q0 ) + w (q1 )
w (r0 ) + w (r1 ) = w r + w (q0 ) ,
where r0 , r1 and r are defined in a similar way on a line passing at some angle
through q, for which q and r are orthogonal and r0 ∈ U and r1 are orthogonal. It
follows from these relations that
|w (r1 ) − w (q1 )| = w (q0 ) − w (r0 ) + w r − w q
= w (q0 ) − w (r0 ) + w r − w (q0 ) + w (q0 ) − w q
≤ |w (q0 ) − w (r0 )| + w r − w (q0 ) + w (q0 ) − w q
≤ 3ε,
where we have used the bounding inequalities between the relation between the
w (q)’s and the distances. Our construction, furthermore, requires r , q ∈ Uq0 .
The subset r0 r1 ∈ U then forms an ε neighborhood of q1 and is therefore
18 Foundations of quantum statistical mechanics
References
Balescu, R. (1975). Equilibrium and Non-equilibrium Statistical Mechanics (New York,
John Wiley), revised 1999 as Matter out of Equilibrium (London, Imperial College
Press).
Birkhoff, G. (1961). Lattice Theory (Providence, American Mathematical Society).
Birkhoff, G. and von Neumann, J. (1936). Ann. Math. 37, 823.
Dirac, P. A. M. (1958). Quantum Mechanics, 4th edn. (London, Oxford University Press).
Dvurecenskij, A. (1993). Gleason’s Theorem and Its Applications (Dordrecht, Kluwer).
Farquahar, I. E. (1964). Ergodic Theory in Statistical Mechanics (London, Interscience).
Gleason, A. M. (1957). J. Math. Mech. 6, 885.
Huang, K. (1987). Statistical Mechanics, 2nd edn. (New York, John Wiley).
Jauch, J. M. (1968). Foundations of Quantum Mechanics (Reading, Addison-Wesley).
Landau, L. D. and Lifshitz, E. M. (1970). Statistical Mechanics (Reading,
Addison-Wesley).
Mackey, G. W. (1963). Mathematical Foundations of Quantum Mechanics (Reading,
Benjamin).
Piron, C. (1976). Foundations of Quantum Physics (Reading, Benjamin).
Tolman, R. C. (1938). The Principles of Statistical Mechanics (London, Oxford).
von Neumann, J. (1955). Mathematical Foundations of Quantum Mechanics (Princeton,
Princeton University Press).
Wigner, E. (1936). Phys. Rev. 40, 749.
2
Elementary examples
2.1 Introduction
Now we will turn to some elementary and familiar examples of quantum mechanics
to remind us of matters which will be used in the subsequent discussions. The focus
will be the harmonic oscillator and also the two-level atom and spin 12 systems
(Dirac, 1958; Louisell, 1973; Cohen-Tannoudji et al., 1977; Jordan, 1986; Liboff,
1998).
19
20 Elementary examples
To reduce this further, let us introduce the well-known creation (a † ) and annihila-
tion (a) operators. (Both are non-Hermitian.)
1
â = √ ωq̂ + i p̂ (2.10)
2ω
1
â † = √ ωq̂ − i p̂ (2.11)
2ω
From the commutation law, Eq. (2.3), we obtain
â, â † = 1. (2.12)
In this representation,
1
Ĥ = h̄ω â â +
†
. (2.14)
2
These relations are true in the Heisenberg as well as the Schrödinger picture.
Now, for the harmonic oscillator,
−iωt
U (t, 0) = exp −iωâ † ât exp .
2
Ĥ |E = E |E
N̂ = â † â = N̂ † .
aN − Na = a (2.16)
a N − Na = a .
† † †
22 Elementary examples
With these raising and lowering operators, we may construct a complete set of
states (Dirac, 1958). For normalized states we have
N̂ |n = n |n n integer and positive (2.17)
< n | n > = δ nn
√
â † |n = n + 1 |n + 1
√
â |n = n |n − 1
a |0 = 0
â †n |0
|n = √
n!
and completeness
∞
|n n| = I.
n=0
The energy is
1
En = ω n + .
2
In the number states, the harmonic oscillator von Neumann equation is
i ρ̇ nn = (E n − E n ) ρ nn
= ω n − n ρ nn .
The solution is simply
ρ nn (t) = exp − iω n − n t ρ nn (0). (2.18)
The diagonal and off-diagonal elements are uncoupled. Diagonal elements are
constant, and the off-diagonal elements oscillate, and
ρ nn (t) = ρ nn (0) = 1. (2.19)
n n
∞ ∞
√
a |α = cn (α) n |n − 1 = αcn (α) |n (2.23)
n=1 n=0
and shift indices n → n + 1. Take the scalar product with |m. We obtain the
recursion relation
√
cn+1 (α) n + 1 = αcn (α) . (2.24)
This gives
αn
cn (α) = √ c0 .
n!
Thus,
∞
αn
|α = c0 √ |n .
n=0
n!
It is easy to show
2
α 2n exp − α2
| n | α |2 = ,
n!
a Poisson distribution. From this n = α ∗ α, and
12
(n − n)2 1 1
= = .
n |α| 1
n 2
24 Elementary examples
so
− |α|2
|α = exp exp α â † exp −α ∗ â |0 , (2.25)
2
taking α to be complex. The completeness relation is
∞
d 2 α |α α| = 1 = |n n| , (2.26)
0
Now we introduce the first example met here of a phase space distribution func-
tion, P(αα ∗ , t), of Glauber (1963) and Sudarshan (1963). Here the “phase space”
is α, α ∗ . Now
d 2 α P αα ∗ , t = 1. (2.31)
The general solution is an arbitrary function f (α (t) , α ∗ (t)). If the initial value is
Gaussian in α, i.e.
P α, α ∗ , 0 = N exp − |α − α 0 |2 ,
then
P α, α ∗ , t = N exp − |α (t) − α 0 |2 .
For
P αα ∗, t = δ 2 (α (t) − α 0 ) ,
the coherent state propagates in time as exp iωt. This was first seen by Schrödinger
(1926).
Let us consider an extension of the harmonic oscillator by including a damp-
ing term. A particularly simple example is the phase damped oscillator with the
interaction
V = a † a + † a † a (2.37)
(Walls and Milburn, 1985; Gardiner, 1991). The von Neumann equation may be
written
1 2 2
ρ̇ = −iω a † a, ρ + K N̄ + 1 2a † aρa † a − a † a ρ − ρ a † a . (2.38)
2
This is the Lindblad form and is discussed in detail in Chapters 4, 5 and 6. Here
N̄ = exp 1ω −1 , and K is a damping constant. In the number representation,
( kT )
1
n |ρ̇| m = −iω (n − m) − K 2 N̄ + 1 (n − m) n |ρ| m .
2
2
The diagonal and off-diagonal elements n |ρ| m are still uncoupled. The solution
is immediate:
2 t
n |ρ (t)| m = exp (−iω (n − m) t) exp − (2 N̄ + 1)K (n − m) n |ρ (0)| m .
2
The off-diagonal elements decay as (n − m)2 K 2 N̄ + 1 to the constant diagonal
initial state n |ρ (0)| m. More will be said of this in the discussion of decoherence
in Chapter 12.
2.3 Spin one-half and two-level atoms 27
To obtain the equation for P (α), we use the operator correspondence discussed
in the appendix:
aρ → α P αα ∗ (2.39)
∂
a†ρ → α∗ − P αα ∗
∂α
∂
ρa → α − ∗ P αα ∗
∂α
∗
∗
ρa → α P αα
†
These are angular momentum commutation laws for half integer l. Now
σ i2 = 1, so
σ i σ j = iσ k . (2.44)
28 Elementary examples
They are not themselves Hermitian. Now we find the commutation laws,
σ ± , σ 1 − = ±σ 3 (2.46)
σ ± , σ 2 − = iσ 3
σ ± , σ 3 − = ∓σ 2
σ +, σ − − = σ 3,
and
σ 21. = σ 22 = σ 23 σ2 = 3 (2.48)
σ 2+ = σ 2− = 0.
The α state is spin positive (m s = +1) along the “3” direction, and β spin down
(m s = −1). Generally,
In this representation,
0 1 0 −i 1 0
σ1 = , σ2 = , σ3 = , (2.50)
1 0 i 0 0 −1
2.3 Spin one-half and two-level atoms 29
ri = Trρσ i .
The above operators have been written in the Schrödinger picture. If ρ 2 = ρ, it is
a pure state. If
1
0
ρ= 2 ,
0 12
then ρ 2 = ρ2 , and in this case, we have a mixture. We find si = 0. The spin is
unpolarized, since all directions are equivalent. A pure polarization state is
cos2 θ2 sin θ2 cos θ2 exp (−iθ )
ρ (θ , φ) = .
sin θ2 cos θ2 exp (iθ ) sin2 θ2
Here s = 12 μ, μ being a classical vector whose polar angles are θ, φ. Remember
that the mixture state is not a unique state |ψ .
The unperturbed spin Hamiltonian is
h̄ω
H= σ z, (2.55)
2
Another Random Scribd Document
with Unrelated Content
CHAPTER LXXII. THE REPLY OF THE GOPAS.
[228] The deity of war and son of Siva. Derived from Kirtika the
personified Pleiades: according to the legend having been
fostered and brought up by the nymphs so called. He was
so great an adept in the art of war that he was appointed
the commander of the celestial army in the war between
the gods and demons.
[230] The word in the text is Udapana—from Uda water, and the
root pa to drink. It may also mean a well. Here it means a
place where water is drunk. Near a well, as is still seen in
many place, there is a spacious pavement where people
may sit at ease and drink.
[231] Any real vessel made of wood, stone and in the shape of a
boat and used for holding or pouring out water, as a
bathing tub, a bathing vessel, a bucket or watering pot &c.
CHAPTER LXXIII. INDRA SENDS DOWN
PUNISHMENT.
[232] Literally the word means one who protects kine from go,
kine and the root pa, to protect.
[234] The Danava chief who was defeated by Rudra or Siva in the
war between the gods and demons.
[237] i.e. the waters will not overflood their banks as in the rainy
season.
[241] For a detailed account of the birth of all these sons see
chapter CXXIII of Adi Parva of Mahabharata.
[242] Kunti, while a maid, obtained a boon from the Rishi Durvasa
that whomever she would wish to have as her consort, he
would at once come to her. To make an experiment she
invoked the sun and Karna is the issue of her union with
him. The account of his birth is described in one hundred
and eleventh chapter of Adi Parva in the Mahabharata.
[243] Once on a time Pandu went out into a forest for hunting.
He struck a Rishi's son who was coupling with his mate in
the form of a deer. He imprecated a curse on Pandu saying
"As you have killed me in the form of a deer when I was full
of desire, so you, O foolish man, will certainly meet with the
fate that has fallen me. When you will go to your dear one,
full of desire as was the case with me, you will at that time,
certainly go to the land of the dead. Your wife will also
follow you." See slokas 30, 31, in chapter CXVIII in
Sambhava Parva of Adi Parva in the Mahabharata.
CHAPTER LXXV. RASA DANCE.
ebookgate.com