Quantum moment hydrodynamics

and the entropy principle

(1) ∗ (2) †
P. Degond and C. Ringhofer

(1) MIP, UMR 5640 (CNRS-UPS-INSA)

Université Paul Sabatier
118, route de Narbonne, 31062 TOULOUSE cedex, FRANCE
email: degond@mip.ups-tlse.fr
(2) Department of Mathematics,
Arizona State University,
Tempe, Arizona 85287-1804, USA
email: ringhofer@asu.edu

Abstract
This paper presents how a non-commutative version of the entropy extremalization
principle allows to construct new quantum hydrodynamic models. Our starting point
is the moment method, which consists in integrating the quantum Liouville equation
with respect to momentum p against a given vector of monomials of p. Like in the
classical case, the so-obtained moment system is not closed. Inspired from Levermore’s
procedure in the classical case [26], we propose to close the moment system by a quantum
(Wigner) distribution function which minimizes the entropy subject to the constraint that
its moments are given. In contrast to the classical case, the quantum entropy is defined
globally (and not locally) as the trace of an operator. Therefore, the relation between the
moments and the Lagrange multipliers of the constrained entropy minimization problem
becomes non-local and the resulting moment system involves non-local operators (instead
of purely local ones in the classical case). In the present paper, we discuss some practical
aspects and consequences of this non-local feature.

Key words: Density matrix, quantum entropy, quantum moments, local quantum equilibria,
quantum BGK models, quantum hydrodynamics,
AMS Subject classification: 82C10, 82C70, 82D37, 81Q05, 81S05, 81S30, 81V70,
∗
supported the European network HYKE, funded by the EC as contract HPRN-CT-2002-00282. P. Degond
wishes to express his gratitude to F. Méhats for very valuable remarks and comments.
†
supported by NSF award Nrs. DMS0204543 and DECS0218008.

1
1 Introduction
The aim of this paper is to present a new approach to quantum hydrodynamics. More precisely,
starting from the quantum Liouville equation, we derive a whole hierarchy of moment models
including quantum hydrodynamical models as well as higher order moment models. The so-
obtained models will be referred to as ’Quantum Moment Hydrodynamics’.
The derivation of quantum hydrodynamic models from first principles has attracted consid-
erable attention in the recent past. The quest for such models is driven by the growing field of
nanotechnology applications. The derivation of reliable yet computationally affordable many-
particle quantum models determines the possibility of efficient industrial design and fabrication
of the next generation of devices. However, few attempts have been successful in this direction.
Indeed, away from the complete resolution of the Schrdinger equation (or even worse, of the
quantum Liouville equation), which is computationally expensive, and the use of continuum
models with ad-hoc phenomenological closure and limited reliability, few alternatives are avail-
able. The starting point for the derivation of quantum hydrodynamic models is the quantum
Boltzmann equation of the form

i~∂t ρ = [H, ρ] + Q(ρ), (1.1)

for the effective single particle density matrix ρ(x, y), where H is the Hamiltonian H =
~2 2
− 2m ∗ |∇x | + V (x), the symbol [., .] denotes the usual commutator and the operator Q models

particle collisions. ~ denotes the reduced Planck constant and m∗ is the particle (effective)
mass. To derive fluid like models for macroscopic quantities it is convenient to consider instead
the equivalent formulation via Wigner functions, which is of the form
1
∂t fw + divx ( pfw ) − θ[V ]fw = Qw (fw ), (1.2)
m∗
where the Wigner function fw (x, p, t) and the pseudo - differential operator θ are related to the
density matrix ρ and the potential V via the Wigner transform
Z
−d ~ ~ i ~ ~
fw (x, p, t) = (2π) ρ(x− η, x+ η)eiη·p dη, θ[V ] = [V (x+ ∇p )−V (x− ∇p )] (1.3)
Rd 2 2 ~ 2i 2i

where ρ(x, y) is the integral kernel of ρ and d is the space dimension (an integer multiple of 3
according to the number of degrees of freedom). Moment systems for the Wigner - Boltzmann
equation (1.2) are equations for a given set of moments
Z
mn (x, t) = κn (p)fw (x, p, t)dp , n = 0, .., N , (1.4)
Rd

which are obtained by building the corresponding moments of (1.2). The main problem in
closing this system, i.e. expressing the highest order moments in the resulting equations in
terms of the lower order moments, is that the collision operator Q (or Qw in (1.2)) is not
available in a sufficiently simple form, to be used in a Chapman - Enskog - like approach to
moment closures. Collision operators for scattering with phonons on the weak coupling limit

2
have been derived and analyzed in [2], [5], [12] and quantum versions of Fokker - Planck type
operators have been derived in [6] and [3].
First attempts towards the derivation p of quantum hydrodynamic models have used BKW
waves. Writing the wave-function as ψ = n(x, t) exp(iS(x, t)), where n is the probability of
presence and S is the phase and inserting it into the Schrödinger equation gives rise to a set
of two equations for n and n∇S which mimic the clasical density and momentum conservation
equations. Compared with the classical case, the momentum equation involves an additional
term, called the Bohm potential. This approach has been investigated in [16], [23], [20], [17].
This is however a pure-state model which does not incorporate mutli-particle effects and which
can be viewed as a zero-temperature model. Finite temperature effects have been taken into
account in [18] through the use of a non linear Schrödinger equation and gives rise to a mo-
mentum equation with a nonlinear enthalpy relation. The equation-of-state is local (i.e. the
pressure (or the enthalpy) depends locally on the density at the same point). In [22], it is
argued that the Bohm potential approach is not consistent with the entropy condition. The
approach presented in this paper brings a cure to this problem.
Other approaches [13], [11] make use of moment closures of the Wigner equation using semi-
classical asymptotics (i.e. a limit ~ → 0 where ~ is the Planck constant) for thermodynamical
equilibrium states as closures. They give rise to local equations-of-state as well, which coincide
with the previous theories up to constant factors. Related to this approach are moment closure
theories of the Wigner equation using small-field asymptotics for thermodynamical equilibria
[14], [15]. These give rise to nonlocal equations-of-state. However, the nonlocality enters only
through the potential.
The idea behind the work presented in this paper is to use a thermodynamic approach using
the minimal amount of information about the collision operator necessary to derive moment
equations. That is we assume knowledge about the corresponding moments of the collision
operator and the existence of an entropy which is dissipated by the collisions. Our approach
bears strong similarities with the theory of NESOM (for NonEquilibrium Statistical Operator
Mechanics) by Zubarev and coworkers [36], [28] (see also the review paper by Luzzi [27]).
However, we give a neater (and to some extent, more practical) mathematical framework which,
we believe, will be useful for further developments of the theory We shall elaborate more on
the relation between our theory and NESOM in section 4, remark 4.2.
Our approach consists in taking moments of the density matrix equation (or quantum
Liouville equation), and then, closing the resulting set of moment equations with an equilibrium
density matrix. This equilibrium is found as an extremum of the entropy functional subject to
the constraints that its moments coincide with those of the density matrix we are considering.
This approach is therefore similar to Levermore’s closure moment hierarchies [26] or to the
extended thermodynamics approach [29] for classical systems
To be more, specific, let a set of polynomials κ(p) = (κ0 (p), . . . , κN (p)) be given. To
any ρ, we associate a set of moments m[ρ] = (m0 (x), . . . , mN (x)) defined by duality as the
representation
P of the linear functionals λ → Tr{ρ Op(λ · κ)}, where λ = (λ0 (x), . . . , λN (x)),
λ · κ = i λi (x)κi (p) and Op means the Weyl quantization of a symbol (i.e. a function of
position x and momentum p) into an operator. This definition of moments by duality will
prove more convenient in connection with the entropy minimization principle. If the chosen

3
set of polynomials is equal to (1, p, |p|2 /2m∗ ), the associated moments correspond to the usual
hydrodynamic quantities, i.e. local density, momentum and energy (per particle). If a larger
set is chosen, the associated moments correspond to higher order hydrodynamic quantities
like the pressure tensor, heat flux vector, higher order heat flux tensor, etc. , according to
the terminology of extended thermodynamics [29]. For instance, the pressure tensor will be
associated with the monomials pi pj where pi and pj denote the components of the momentum
vector in the i-th and j-th direction respectively.
Now, we turn to the definition of an equilibrium density matrix which satisfies given moment
constraints. Such an equilibrium minimizes the entropy (defined as H(ρ) = Tr{h(ρ)} where h is
a suitable convex function, like e.g. the Boltzmann entropy h(ρ) = ρ(ln(ρ)−1)). In this work, we
show that this constrained minimization problem has the solution ρm = (h0 )−1 (Op(µ · κ)) where
µ = (µ0 (x), . . . , µN (x)) are the Lagrange multipliers of the moment constraints {m[ρ] = m
given}. (h0 )−1 is the inverse function of the derivative of h. Functions of operators are given a
meaning in the sense of functional calculus.
It is now possible to get back to the problem of deriving moment models from the quantum
Liouville equation. Taking the moments of the Liouville equation (which is now a precise
concept), we can close the chain of moment equations by using the equilibrium just defined.
According to the chosen set of polynomials and the number of moments involved, we obtain
a hierarchy of moments systems which we call Quantum Moment Hydrodynamics. We note
that such a closure is different from a single-state closure as it takes into account many-particle
interactions through the entropy minimization procedure. However, we should recover the
single-state closure by a zero-temperature asymptotics (for ~ being fixed) of our moment models.
The investigation of this point is left to future work. Also, the moment systems should obviously
constitute an approximation of the original quantum kinetic system (1.1). However, since the
expression of Q(ρ) is not known, assessing the accuracy of this approximation is extremely
delicate.
The goal of the present paper is to develop these concepts. The paper is organized as
follows: in section 2, we recall Levermore’s approach to moment closure hierarchies in the
classical case, however giving it a presentation which makes its extension to the quantum
case more natural. Then, in section 3, we extend Levermore’s approach to the quantum case,
developing our definition of moment. We elaborate more on the fluid entropy and compare
our theory with NESOM in section 4. Quantum moment hydrodynamics systems are derived
and studied in section 5. In particular, it is shown that quantum hydrodynamics for the usual
set of moments (density, momentum and energy) differs from classical hydrodynamics by a
possibly non-scalar pressure tensor and non-zero heat flux vector. These are related non locally
to the basic moments through the equilibrium density matrix. Higher order quantum moment
systems exhibit similar features. Finally, as noticed in this section, it is possible to make sense
to quantum BGK relaxation systems. A conclusion is drawn in section 6. Finally, technical
details are deferred to two appendices (sections 7 and 8).
It should be stressed that most of the mathematical properties stated in this paper are given
only formal proofs. Fully rigorous proofs will require a lot of mathematical developments which
are beyond the scope of the present paper. However, a practical usage of these new models
does not require that all the mathematical theory is settled. Indeed, this paper should rather

4
be viewed as a presentation of new models and as a programme definition for future work.
The results of the present work have been announced in [9].

2 Entropy minimization principles and moment closure

hierarchies in the classical case
In this section, we mainly review the approach proposed by Levermore [26] in the classical
case. However, we shall present the entropy minimization principle in a slightly different (but
completely equivalent) form which will make it more suitable to an extension to the quantum
case.
In classical kinetic theory, the basic object is the particle distribution function f (x, p, t)
where x ∈ Rd is the position, p ∈ Rd is the momentum, and t is the time. f is a probability
distribution and is therefore normalized according to
Z
f (x, p) dx dp = 1 . (2.1)
R2d

In all this work, we shall assume that the total number N of particles in the system is fixed. This
hypothesis is not essential in the classical case, but it will greatly simplify the presentation of the
quantum case. The distribution function f is a solution of the so-called Boltzmann equation:
∂f
+ {Ĥ, f } = Q(f ) , (2.2)
∂t

where Ĥ(x, p) is the particle Hamiltonian, {Ĥ, f } is the Poisson bracket

{Ĥ, f } = ∇p Ĥ · ∇x f − ∇x Ĥ · ∇p f ,

and Q(f ) is a collision operator, which models the interactions of the particles among themselves
or with the surrounding medium.
On the other hand, the physics of continua considers averaged quantities which only depend
on the position variable such as the mean density, momentum or energy. These quantities are
defined as moments of the distribution function f . More specifically, let κi (p), i = 0, . . . , N be
N + 1 independent functions of p and denote by κ(p) = (κi (p))i=0,...,N the vector of monomials.
The associated moments ki (f )(x) of f (x, p) are defined by:
Z
ki (f )(x) = f (x, p) κi (p) dp , ∀x .
Rd

The equations for k(f ) = (ki (f ))i=0,...,N are obtained by multiplying (2.2) by κi (p) and inte-
grating with respect to p:
Z Z
∂ki (f )
+ κi (p) {Ĥ, f } dp = κi (p) Q(f ) dp (2.3)
∂t Rd Rd

5
In most cases the integrals in (2.3) cannot be expressed in terms of the functions ki (f ) alone.
In order to reduce system (2.3) to a closed system for the functions ki , some assumptions
must be made on the distribution function f . According to statistical physics, the most likely
distribution function (upon a certain number of realizations) is a minimum of the entropy
functional among functions whose moments are given by the functions ki . Following [26], we
use this Ansatz and replace f in the integrals of (2.3) by this distribution function.
The minimization principle (or Gibbs minimization principle, see [4]) can be formulated as
follows. Let h be a smooth strictly convex function defined on [0, ∞) and define the entropy
functional H(f ) acting on functions f (x, p) associated with h by:
Z
H(f ) = h(f (x, p)) dx dp ,
R2d

Let m = (mi (x))i=0,...,N be N + 1 given functions of the position variable x. We are concerned
with the following minimization problem (Gibbs problem): find the solution f m of

H(f m ) = min {H(f ) | f satisfies ki (f )(x) = mi (x) , ∀x, ∀i} . (2.4)

The normalization condition (2.1) is not included in the constraints of (2.4). Rather, we assume
that the constant function is contained in κ(p), say κ0 (p) = 1. Then,
R m0 (x) is theR probability of
presence of a particle in the neighbourhood of x and is such that m0 (x) dx = f (x, p) dx dp.
We restrict the set of moments to those satisfying
Z
m0 (x) dx = 1 , (2.5)
Rd

so that the constraint (2.1) is satisfied, as soon as the constraints of (2.4) are satisfied.
In most physics textbooks, the tradition is to use the opposite sign for the entropy (i.e.
to suppose that it is a concave function) and to write the Gibbs principle as a constrained
maximization problem. It is the mathematicians’ usage however to use the opposite convention.
Of course, the two conventions are completely equivalent.
We note that this problem can be equivalently stated as
Z
m
H(f ) = min {H(f ) | f satisfies Kλ (f ) = m · λ dx , ∀λ(x) = (λi (x))i=0,...,N } , (2.6)
Rd

where, we have introduced

Z Z
Kλ (f ) = k(f ) · λ(x) dx = f (x, p) κ(p) · λ(x) dx dp , (2.7)
Rd R2d

and, for two vectors a and b of RN +1 , we write a · b =

P
i ai bi . The arbitrary functions
λ = (λi (x))i=0,...,N are the Lagrange multipliers of the constraints.
According to classical optimization theory, the constrained optimization problem (2.4) can
be equivalently formulated in terms of a saddle-point problem for the Lagrangian
 Z 
Lm (f, λ) = H(f ) − Kλ (f ) − m(x) · λ(x) dx , (2.8)
Rd

6
where the Lagrange multipliers λ = (λi (x))i=,...,N are functions of x. The saddle-point problem
is formulated as follows:
H(f m ) = min max Lm (f, λ) = max min Lm (f, λ) , (2.9)
f λ λ f

where now the minimum over f or the maximum over λ are unconstrained problems.
Let us now consider the unconstrained problem
Lm (fλ , λ) = min Lm (f, λ) . (2.10)
f

Then, the necessary condition for extremality leads to

fλ (x, p) = (h0 )−1 (λ(x) · κ(p)) (2.11)
where h0 is the derivative of h and (h0 )−1 is the inverse function of h0 (which exists since h0 is
strictly increasing). fλ is called the equilibrium distribution function and λ is called the system
of entropic variables.
Now, the solution f m of the constrained minimization problem (2.4) is an equilibrium dis-
tribution function fµ where µ = µm is a solution of the unconstrained maximization problem
H(f m ) = Lm (fµm , µm ) = max Lm (fλ , λ) . (2.12)
λ

The resolution of (2.12) leads to

f m = fµm , (2.13)
where µm is such that
ki (fµm ) = mi , ∀i = 1, . . . N , (2.14)
We note that this relation is equivalent to saying that
Z
Kλ (fµ ) =
m m(x) · λ(x) dx , ∀λ(x) , (2.15)
Rd

This last relation is in a form which will be easily extended to the quantum case. The entropic
variables µ and the moment (or conservative) variables m are dual (in the Legendre transform
sense) through the fluid entropy (see section 4 where this aspect is developed in the quantum
case).
Now, following [26], a closed set of moment equations can be derived from (2.3) by replacing
f in the integrals appearing in (2.3) by the solution of the Gibbs principle associated with the
constraint that the moments are ki (f ). This leads to a closed system of equations for the
moments mi = ki (f ) which is written as follows:
Z Z
∂mi m
+ κi (p) {Ĥ, f } dp = κi (p) Q(f m ) dp (2.16)
∂t Rd Rd

For future use, we note that system (2.16) can be written in weak form as:
Z Z Z
∂ m
m · λ dx + λ(x) · κ(p) {Ĥ, f } dp dx = λ(x) · κ(p) Q(f m ) dp dx , (2.17)
∂t Rd R2d R2d

7
for all λ(x). We shall extend this weak form of the hydrodynamic equations to the quantum
case. The key point is to interpret the integrals in (2.17) in an operator way.
We now comment on the form (2.4) of the Gibbs minimization principle. In[26], the Gibbs
principle is given a local form. For a function f (p), we denote by H̃ the ’local’ entropy functional
Z
H̃(f ) = h(f (p)) dp .
Rd

Then, for any given (N+1)-tuple of real numbers m̃ = (m̃i )i=0,...,N , the local Gibbs principle
reads: find f˜m̃ (p) which realizes

H̃(f˜m̃ ) = min {H̃(f ) | f (p) satisfies Ki (f ) = m̃i , ∀i} . (2.18)

Of course, for f (x, p), we have

Z
H(f ) = H̃(f (x, ·)) dx .
Rd

The solution of the ’local’ Gibbs minimization principle (2.18) is obviously

f˜m̃ = (h0 )−1 (µ̃m̃ · κ) , (2.19)

where µ̃m̃ is such that

Ki (f˜m̃ ) = m̃i , ∀i . (2.20)
This defines a local mapping m̃ ∈ RN +1 → µ̃m̃ ∈ RN . The solution of the global Gibbs principle
(2.13) is obviously related to that of the local one (2.19) by:

µm (x) = µ̃m(x) .

However, the global form (2.4) of Gibbs principle can be extended to the quantum case, while
the local form cannot. Quantum mechanics is a non-local theory and requires that a non-local
Gibbs principle be used.
Let us review some classical cases. Suppose that the Hamiltonian is given by

|p|2
Ĥ(x, p) = ε(p) + V (x, t) , ε(p) = ,
2m∗
where ε(p) is the kinetic energy, V (x, t), the potential energy and m∗ the particle mass. Suppose
that the vector κ of monomials is a d + 2-dimensional vector (d being the dimension of the
physical space) consisting of

κ(p) = (1, (pi )i=1,...,d , |p|2 /(2m∗ )) . (2.21)

The functions involved in κ(p) correspond to the physically conserved quantities (i.e. the
probability of presence, the d components of the momentum per particle and the energy per
particle). As often in the literature, we shall label the components of this vector according to

8
κ(p) = (κ0 (p), (κi (p))i=1,...,d , κd+1 (p)). In thermodynamics, the associated vector of Lagrange
multipliers is usually written
µc u 1
λ=( , , − ) := (λ0 , (λi )i=1,...,d , λd+1 ) ,
T T T
where µc ∈ R is the chemical potential, u ∈ Rd is the mean velocity and T ∈ R+ is the
temperature. The associated moment vector in turn is written

m = (n, q, W ) := (m0 , (mi )i=1,...,d , md+1 ) ,

where n ∈ R+ is the probability of presence, q ∈ Rd is the mean momentum per particle and
W ∈ R+ is the mean energy per particle. Multiplying n, q and W by the total number of
particles N gives the number density, the fluid momentum and the fluid energy.
Let us consider specifically the Boltzmann entropy:

h(f ) = kB f (ln f − 1) , (2.22)

where kB is the Boltzmann constant. In this case, the equilibrium distribution function fλ is
the classical Maxwellian:
d d
1 X |p|2 1 X |p|2
fλ = exp{ (λ0 + λi pi + λd+1 ∗ )} = exp{ (µc + ui p i − ∗
)} , (2.23)
kB i=1
2m k B T i=1
2m

Then, the mapping m → µm relates (n, q, W ) to (µc , u, T ) in the following way:

µc + m∗ |u|2 /2 1 d
n = h−d ∗
P (2πm kB T )
d/2
exp , q = m∗ nu , W = m∗ n|u|2 + nkB T .
kB T 2 2
With this transformation, the Maxwellian takes the more familiar form

n |p − m∗ u|2
fλ = exp{− }. (2.24)
(2πm∗ kB T )d/2 2m∗ kB T
Supposing
R that the collision operator is mass, momentum and energy conservative, i.e.
satisfies Q(f )κi (p) dp = 0 for all considered κi , eqs (2.16) can be simplified and gives rise to
the usual classical hydrodynamic equations:
∂n
+ ∇ · (nu) = 0 , (2.25)
∂t
∂nu
m∗ ( + ∇ · (nu ⊗ u)) + ∇(nkB T ) = −n∇V , (2.26)
∂t
∂W
+ ∇ · (W u) + ∇ · (nkB T u) = −n∇V · u . (2.27)
∂t
Therefore, the above considered moments give rise to the balance equations for the physically
conserved quantities (i.e. mass, momentum and energy).

9
 If other moments than those corresponding to the physically conserved quantities, such as
the heat flux (corresponding to κ(p) = pi |p|2 /2), the pressure tensor anisotropy (i.e. κ(p) =
pi pj ), or the higher order heat flux tensor (i.e. κ(p) = pi pj pk ) are considered, new sets of
hydrodynamic-like equations are obtained. This gives rise to the so-called ’moment closure
hierarchies’ of Levermore [26], which are also closely related with the theory of extended ther-
modynamics [29]. Then, the equilibrium distribution functions (2.11) depend on a larger di-
mensional vector of Lagrange multipliers λ than the Maxwellians (2.23), which therefore appear
as a special case of these equilibria. If one linearizes these equilibria about the Maxwellians
(considering that the higher order moments, after convenient normalization, are small), one
obtains another moment closure method first proposed by Grad [19].
However, the following considerations should be borne in mind. First, the N -tuple of mono-
mials κ(p) cannot be completely arbitrary. It has to satisfy a certain number of requirements
such as gallilean invariance, or the fact that the moments of fλ must be defined. These con-
straints are detailed in [26]. Second, the existence and uniqueness of solutions of the mini-
mization problems is by far not guaranteed. Moment realizability conditions, i.e. conditions
that guarantee that the constrained minimization problem has a solution, as well as uniqueness
conditions have been studied in [24], [25], [32], [1]. When moments
R corresponding to physically
non conserved quantities (i.e. such that the corresponding Q(f )κi (p) dp is non zero) are con-
sidered, a particular care must be taken in the closure relation for the collision operator part of
eq. (2.3). This must be done in order to capture the correct dynamics in situations which are
close perturbations of the usual gas dynamics equations. This point is detailed in [26]. Finally,
moment systems of the form (2.16) are by construction hyperbolic, which guarantees a certain
degree of well-posedness (at least in a linear sense). This point is developed in [26].
We are now going to investigate how these considerations can be extended to the quantum
case. For that purpose, we shall use the non local version of the Gibbs minimization principle.

3 Entropy minimization principles in the quantum case

We first state the entropy minimization principle in the quantum case. To be specific, we
consider a system whose objects can be described by wave-functions ψ(x) belonging to the
Hilbert space X = L2 (Rd ). In this context, the quantum equivalent of the distribution function
is the density matrix ρ which is a positive, trace-class Hermitian operator on X satisfying [31]:
Tr{ρ} = 1 . (3.1)
The normalization condition (3.1) is the quantum counterpart of (2.1). Then, the spectrum of
ρ consists of a decreasing sequence of positive eigenvalues (α` )`=1,...,∞ tending to 0 as ` → ∞
and the associated eigenvectors φ` constitute a Hilbert basis of X. Moreover, because of (3.1),
we have ∞
X
α` = 1 .
`=1
The elementary wave-function φ` represents a pure state, while the datum of a density matrix
ρ represents a mixed state. The eigenvalues α` represent the probability for the system to be

10
in the state φ` .
Any observable defined by a Hermitian operator A on X gives rise to an observation hAiρ
on the system modeled by ρ according to the formula

hAiρ = Tr{ρA} ,

where ρA denotes the operator multiplication of ρ and A. Examples of such observables are
the mean position defined by the position operator X which operates through the multipli-
cation by x, or the mean momentum defined by the momentum operator P = −i~∇x where
~ is the reduced Planck constant and i2 = −1. More generally, we shall consider operators
Op(a) obtained from any symbol a(x, p) of the position x and momentum p through the Weyl
quantization by Z  
1 x+y p(x−y)
Op(a)φ = d
a , p φ(y)ei ~ dp dy . (3.2)
(2π~) R2d 2
The Weyl quantization is such that any real valued symbol gives rise to a Hermitian operator
(under regularity conditions that we shall not detail see for instance [34]). The Weyl quantiza-
tion (3.2) is formally related to the Wigner transform (1.3) via the formula ρ = (2π~)d Op(fw ).
Thus, Weyl quantization and Wigner transform are (up to a factor (2π~)d ) inverse operations
one to each other and Z
Tr(ρOp(a)) = a(x, p)fw (x, p)dxdp (3.3)
R2d
holds. Wigner functions are therefore a convenient tool to express local moments in quantum
mechanics. Entropy principles, however, are better expressed in terms of density matrices
and operators. A class of symbol that we shall be particularly interested in are a(x, p) =
λ(x) · κ(p) where λ = (λi (x))i=0,...,N is an arbitrary N+1-tuple of real valued functions and
κ = (κi (p))i=0,...,N is the N+1-tuple of moment monomials.
We now choose a given N+1-tuple of real valued moment monomials κ, such that κ0 (p) = 1.
For any given operator ρ, we shall denote by Kλ (ρ) the expectation value of the operator
Op(λ(x) · κ(p)) (denoted Op(λ · κ) for short) i.e. for any N+1-tuple of real valued functions
λ = (λi (x))i=0,...,N :
Z
Kλ (ρ) = Tr{ρOp(λ · κ)} = λ(x) · κ(p)fw (x, p)dxdp . (3.4)
R2d

Kλ (ρ) is therefore the observation corresponding to a certain combination of the moment mono-
mials κi (p), weighted by x-dependent factors λi (x). The mapping λ → Kλ (ρ) is a linear func-
tional defined on the set of real valued functions λ(x) which, by duality, defines a N+1-tuple
of real valued functions m[ρ] := m(x) = (mi (x))i=0,...,N , which are functions of x, according to
Z
Kλ (ρ) = λ(x) · m(x) dx , ∀λ(x) real valued . (3.5)
Rd

These moments are nothing but the local moments of the Wigner distribution function fw in
the usual sense: Z
m[ρ](x) = κ(p)fw (x, p, t) ds dp .

11
Therefore, (3.5) is a way to express local moments of the Wigner distribution function in terms
of the density operator ρ. This formula is clearly the quantum equivalent of (2.7). Therefore,
in the quantum framework, there is no local relation between ρ and its moments m, but rather,
a functional one, through the duality (3.5).
For the sake of clarity, we give explicit expressions of the moments m[ρ] expressed in terms
of the density matrix ρ. The proof of the following two Lemmas is just an exercise in Fourier
transforms using the definition (1.3) of the Wigner function fw and the equivalence formula
(3.3), and is left to the reader. We start with the following expression of Op(λ · κ).

Lemma 3.1 Let β = (β1 , . . . , βd ) ∈ Nd be a multi-index (with N the set of natural integers)
and denote by pβ = pβ1 1 . . . pβd d and ∂/∂xβ = ∂/∂xβ1 1 . . . ∂/∂xβd d . Then, for any smooth real or
complex-valued function ν(x), we have the two following equivalent expressions of the operator
Op(pβ ν):
βd  β1   
β
X
|β|
Xβ1 βd 1 ∂ν ∂φ
Op(p ν)φ = (−i~) ... ... (3.6)
γ1 γd 2|γ| ∂xγ ∂xβ−γ
γ1 =0 γd =0
β1 βd    
β 1 X X β1 βd
Op(p ν) = |β| ... ... pγ νpβ−γ , (3.7)
2 γ1 γd
γ1 =0 γd =0


βi
where |β| = β1 + . . . + βd and is the binomial coefficient. We denote by ν the multipli-
γi
cation operator by the function ν(x). Because ν does not commute with pi , the orders of the
factors in the right-hand side of (3.7) matters.

The local moments are given as operators acting on the integral kernel of the operator ρ by:

Lemma 3.2 Let β ∈ Nd be a multi-index and let mβ [ρ] be the moment of ρ associated with
the monomial pβ i.e. satisfying
Z Z
β β
mβ (x) = p fw (x, p)dp, Tr{ρ Op(p ν(x))} = mβ (x)ν(x) dx ,
Rd Rd

for any test function ν(x). Let ρ(x, x0 ) be the integral kernel of the operator ρ in the position
representation, i.e. the operator ρ acts on any square integrable function φ(x), x ∈ Rd according
to Z
ρφ(x) = ρ(x, x0 )φ(x0 ) dx0 . (3.8)
Rd
Then, mβ is given by
 β1
|β| X βd      
i~ X β1 βd |γ| ∂ ∂ 
mβ (x) = ... ... (−1) ρ  . (3.9)
2 γ1 γd ∂x γ
∂x 0 (β−γ) 
(x,x)
γ1 =0 γd =0

12
 This lemma shows that, as soon as the integral kernel ρ(x, x0 ) is smooth enough, the moment
mβ is a function. Otherwise, mβ may be a singular distribution.
Now, we turn to the definition of entropy. By functional calculus, any continuous function h
(and by duality, any measure) defined on the interval R+ gives rise to an operator h(ρ) defined
by
X∞
h(ρ)φ = h(α` )(φ, φ` )X φ` ,
`=1
where (·, ·)X denotes the scalar product in X (supposed linear with respect to the left entry and
antilinear with respect to the right entry). Let h be a strictly convex function on R+ . Then,
we define the quantum entropy H of ρ with respect to h according to
H(ρ) = Tr{h(ρ)} , (3.10)
We now formulate the quantum entropy minimization
R principle: let m(x) be a N+1-tuple
of moment functions (defined on R ) such that m0 dx = 1. Find ρm such that
d
Z
m
H(ρ ) = min {H(ρ) | ρ satisfies Kλ (ρ) = λ(x) · m(x) dx ,
Rd
∀ real valued λ(x) = (λi (x))i=,...,N } . (3.11)
We see that this minimization problem is the quantum equivalent of problem (2.6). The con-
strained optimization problem (3.11) can be rephrased as a saddle-point problem for the La-
grangian  Z 
Lm (ρ, λ) = H(ρ) − Kλ (ρ) − m(x) · λ(x) dx , (3.12)
Rd
according to
H(ρm ) = min max Lm (ρ, λ) = max min Lm (ρ, λ) , (3.13)
ρ λ λ ρ

where now the minimum over ρ or the maximum over λ are unconstrained problems. Again,
we stress the fact that λ = (λi (x))i=0,...,N is a N+1-tuple of functions of x.
To pursue the analysis, we need the following two lemmas, the proofs of which are deferred
to the appendix in section 8. From now on, derivatives denoted with a δ will refer to Gâteaux
derivatives. The Gâteaux derivative δH/δρ of a function H(ρ) (if it exists) is a linear form
acting on increments δρ according to
 
δH 1
δρ = lim [H(ρ + tδρ) − H(ρ)] .
δρ t→0 t
A necessary condition for extremality of H is that its Gâteaux derivative vanishes (first order
Euler-Lagrange equation of the extremality problem).
Lemma 3.3 Let h be a strictly increasing continuously differentiable function defined on R+ .
Consider that H(ρ) is defined on the space of Trace-class positive self-adjoint operators ρ. Then
H is Gâteaux differentiable and its Gâteaux derivative δH/δρ is given by:
∞
δH X
δρ = h0 (α` )δρ`` = Tr{h0 (ρ)δρ} , (3.14)
δρ `=1

13
where α` are the eigenvalues of ρ, and δρ`` are the diagonal values of the perturbation operator
δρ in the basis of the eigenfunctions φ` of ρ.
Lemma 3.4 Let h be a strictly convex twice continously differentiable function on R+ . Then,
H is strictly convex and is twice Gâteaux-differentiable, with :
δ2H X h0 (α` ) − h0 (αr )
(δρ, δρ) = |δρ`r |2 , (3.15)
δρ2 `,r
α` − α r

where the quotient is understood to be h00 (α` ) when α` = αr . The perturbation operator δρ is
assumed Hermitian.
In fact, Nier [30] has proved infinite Frechet differentiability of h (provided h is a C ∞
function) and has given proofs of formulae (3.14) and (3.15). However, his proof uses the
theory of Hellfer and Sjöstrand [21] of almost analytic extensions in functional calculus. In the
appendix, for the reader’s convenience, we give elementary proofs of the weaker statements 3.3
and 3.4.
Let us now consider the unconstrained problem
Lm (ρλ , λ) = min Lm (ρ, λ) . (3.16)
ρ

Then, we have:
Lemma 3.5 The necessary condition for extremality for the unconstrained minimization prob-
lem (3.16) is
ρλ = (h0 )−1 (Op(λ · κ)) (3.17)
where (h0 )−1 is the inverse function of h0 .
The operator ρλ is called the equilibrium density operator associated with the N+1-tuple λ
of entropic variables. We stress the fact that, in this theory, λ is a N+1-tuple of functions of the
position variable x and not mere constants. Formula (3.17) must be understood in the sense of
functional calculus, which is possible via the spectral theorem, since the operator Op(λ · κ) is
Hermitian [31]. The proof of Lemma 3.5 is deferred to the appendix, section 7.
Now, the solution ρm of the constrained minimization problem (3.11) is an equilibrium
density operator ρλ where λ = µm is a solution of the unconstrained maximization problem
H(ρm ) = Lm (ρµm , µm ) = max Lm (ρλ , λ) . (3.18)
λ

We have:
Lemma 3.6 The solution ρm of the constrained Gibbs minimization problem (3.11) or equiv-
alently, of the unconstrained maximization problem (3.18) is given by
ρm = ρµ m , (3.19)
where µm is such that Z
Kλ (ρµm ) = m(x) · λ(x) dx , ∀λ(x) , (3.20)
Rd

14
 By duality, relation (3.20) expresses that the moments of ρµm are m (see (3.5)). It is the
quantum extension of (2.15). Again, the entropic variables µ (which are functions of x) and the
moment (or conservative) variables m (which are also functions of x) are dual (in the Legendre
transform sense) through the fluid entropy (which is now a functional on functions of x, see
section 4). For the proof see the appendix, section 7

4 The quantum fluid entropy

We now discuss the concept of fluid entropy in the framework of the present theory. Let
m = (mi (x))i=0,...,N be a given N-tuple of fluid moments (which are functions of the position
variable x). We define the fluid entropy S(m) by

S(m) = H(ρm ) . (4.1)

For S, we can prove:

Lemma 4.1 (i) S is strictly convex.

(ii) We have
δS
= µm , (4.2)
δm
where δS/δm is the Gâteaux derivative of S with respect to m.

Eq. (4.2) shows another aspect of the duality between µ and m, namely that they are dual
through the entropy. Of course, this relation extends a well-known relation in the classical case.
The proof can be found in the appendix, section 7.
To highlight the duality between the entropic variables µ and the moments (or conservative
variables) m, we show how relation (4.2) can be inverted by means of the Legendre dual of the
entropy. Define Z
Σ(µ) = S(m) − µ · m dx , (4.3)
Rd

where m is such that (δS/δm)(m) = µ (or in other words, such that µm = µ, i.e. m is the set
of moment of ρµ ). We have the following Lemma, the proof of which is given in the appendix,
section 7.

Lemma 4.2 (i) Σ is strictly concave.

(ii) We have
δΣ
= −m , (4.4)
δµ

The function Σ is sometimes called the Massieu-Planck function [4].

15
Remark 4.1 In equilibrium statistical mechanics (see e.g. [4]) the constraints are global i.e.
are written
Tr{ρκi } = mi , ∀i = 0, . . . , N ,
where (κi )i=0,...,N are the operators corresponding to the global observables (e.g. the total
energy: κ = Ĥ). and (mi )i=0,...,N are N + 1 real numbers. Among these constraints, the
first one, corresponding to κ0 = 1 and m0 = 1, has a special status. Indeed, it does not say
anything about the thermodynamical state of the system, but simply expresses that ρ is a
statistical operator, i.e. satisfies (3.1). The associated equilibrium is written:
N
X
ρµ = (h0 )−1 (µ0 + µi κi ) ,
i=1

where µi are now constants. If we further specialize to the Boltzmann entropy (2.22), we get:
N N
1 X 1 1 X
ρµ = exp{ (µ0 + µi κi )} = exp{ ( µi κi )} , (4.5)
kB i=1
Z kB i=1

where Z is the so-called partition function. Z can be viewed as a function of µ1 , . . . , µN if ρµ

is constrained to satisfy (3.1), indeed:
( N
)
1 X
Z{µ1 , . . . , µN } = Tr exp{ ( µi κi )} .
kB i=1

It can be shown that ln Z coincides with −Σ (up to a constant), i.e. we have:

δ ln Z
= mi , i = 1, . . . , N .
δµi
In our case however, since the µi ’s are no longer constant but instead functions of x, it is
impossible
P to factor out the contribution of µ0 outside the exponential. This is because µ0 (x)
and i µi (x)κi are operators which do not commute in general. So, in our framework, there
is no such object as a partition function. However, the Massieu function Σ is well-defined by
(4.3) and can be used to define the moments m as functionals of the Lagrange multipliers µ.

Remark 4.2 At this point, it is appropriate to compare our approach to the NESOM theory
(for NonEquilibrium Statistical Operator Mechanics), which has been pioneered by Zubarev et
al (see e.g [36]) and has been later expanded by Luzzi, Vasconcellos, and coworkers. The reader
can refer e.g. to [28] and [27] for reviews. This approach consists in extending formula (4.5)
into a formula for a local equilibrium according to
N Z
1 1 X
ρµ = exp{ ( µi (r, t)κi (r, t) dr)} , (4.6)
Z kB i=1

16
where κi (r, t) are local versions of the moment operators κ corresponding to the conserved
quantities. R
The expression µi (r, t)κi (r, t) dr inside the exponential bears strong similarities with our
introduction of the observable µ · κ and formula (4.6) is close to our formula (3.17) which, in
the case of the Boltzmann entropy (2.22), would reduce to
1
ρµ = exp{ (Op(µ · κ))} , (4.7)
kB
In the cited references, the local operators κi (r, t) are not further precised. If we assume that
they are localizations of the global moment operators κi (p) in the way outlined P R in the introduc-
tion, i.e. something like κi (r, t) = Op(κi (p)δ(x − r)), then the expression i µi (r, t)κi (r, t) dr
coincides with Op(µ · κ). However, unless this is identification is made, the two approaches do
not give the same results. Also, as pointed out in the previous remark, the factoring out of
partition function in (4.6) looks somehow suspicious. However, the two approaches have clearly
similar roots and the added value of our approach is to provide a neat mathematical framework
to the notion of locally conserved variables by means of the operators Op(µ · κ) and the dual
concept of moments, as developed in section 3. This framework may in turn prove powerful for
future extensions and applications of the theory.

5 Quantum moment hydrodynamics

5.1 General framework
The time evolution of the density matrix can de derived from the Schrdinger equation and is
given by the quantum Liouville equation for the density matrix (or Von-Neumann equation):
∂ρ
i~ = [Ĥ, ρ] + i~Q(ρ) , (5.1)
∂t
where Ĥ is the particle Hamiltonian, [Ĥ, ρ] = Ĥρ − ρĤ is the commutator of Ĥ and ρ and
Q(ρ) is an abstractly defined collision operator taking care of dissipation phenomena. To fix
the ideas, we can think of Ĥ as being the simple Hamiltonian on Rd

~2 |p|2
Ĥφ = − ∆φ + V φ = Op( + V )φ , (5.2)
2m∗ 2m∗
where V = V (x, t) is the potential.
About the collision operator Q, we specifically request the following:
(i) Q locally conserves the moments associated with κ(p), i.e.:
Z
Qw (fw )(x, p) κi (p) dp = 0 , i = 0, . . . , N , ∀x ∈ R3 ,

where Qw is the Wigner transform of Q,

17
 (ii) Q(ρ) dissipates the quantum entropy, i.e.

Tr{Q(ρ)h0 (ρ)} ≤ 0 .

An example of such an operator is the BGK operator of section 5.5 (we shall prove in [8] that
it is consistent with entropy dissipation i.e. that property (ii) holds). In [10], we also derive a
Boltzmann-like collision operator for binary quantum collisions precisely on the basis of these
two requirements.
From (5.1), we want to derive a system of conservation equations using a similar route as
that exposed in section 2 for the classical case. First, from (5.1) we derive a system of equations
for the moments, and then, use the equilibrium density matrix ρλ to express all the quantities
that cannot be directly expressed as moments. However, there is a significant difference from
the classical case, in that the moments are not defined directly but rather, from duality through
the linear forms Kλ (ρ). Therefore, the moment equations come naturally in duality form (or
in weak form, in the sense of Partial Differential Equations theory).
Define m[ρ] to be the moments of ρ, through the duality relation (3.5). To obtain the
moment equations, we compose (5.1) on the right by Op(λ · κ) and take the trace. We obtain:
Z   
∂ i
m[ρ(t)](x) · λ(x) dx = Tr ( − Ĥ, ρ + Q(ρ))Op(λ · κ) , ∀λ(x) . (5.3)
∂t Rd ~

Eq. (5.3) is the quantum equivalent of (2.3). However, the right-hand side cannot be expressed
in general in terms of the moments m[ρ(t)]. This is the closure problem.
We close eq. (5.3) by using the solution ρm = ρµm of the Gibbs minimization problem (3.11)
with moments m = m[ρ(t)]. We find
Z   
∂ i
m(x, t) · λ(x) dx = Tr ( − Ĥ, ρµm(t) + Q(ρµm(t) ))Op(λ · κ) , ∀λ(x) . (5.4)
∂t Rd ~

Eq. (5.4) is a quantum moment closure system for the set of moments m, which we refer to as
’Quantum Moment Hydrodynamics’. Rewriting (5.4) in terms of Wigner functions instead of
density matrices, one can now return to a strong formulation of the moment system. Equation
(5.4) is equivalent to the weak formulation of
Z Z
1 m m
∂t m + κ(p){∇x · ( ∗ pfw ) − θ[V ]fw }dp = κ(p)Qw (fwm )dp, (5.5)
Rd m Rd

where fwm (x, p, t) is the closure Wigner function corresponding to ρµm(t) , i.e.
Z
m −d −1 −d ~ ~
fw (x, p, t) = (2π~) Op (ρµm(t) ) = (2π) ρµm(t) (x − η, x + η)eiη·p dη (5.6)
Rd 2 2
holds.
We now investigate several examples and consequences of this methodolgy. In all these
examples, otherwise explicitely stated, we shall restrict to the case of the Hamiltonian (5.2).

18
5.2 Quantum hydrodynamics
Our first use of the present framework is the derivation of the quantum hydrodynamic equa-
tions, which are the quantum counterpart of the classical hydrodynamic system (2.25)-(2.27).
For this purpose, we again consider the d+2-tuple of moment functions (2.21) i.e. κ(p) =
(1, (pi )i=1,...,d , |p|2 /(2m∗ )). We first derive the moment equations from the Wigner function
formulation (5.5). In analogy to the classical case, change from the moment variables mi to
density n, ensemble velocity u and temperature T via the formulas
Z Z
m ∗
m0 = n = fw dp, mi = qi = m nui = pi fwm dp, i = 1, .., d
Rd Rd

n(m∗ |u|2 + dkB T ) |p|2 m

Z
md+1 =W = = f dp
2 Rd 2m∗ w
The moment system (5.5) is then of the form

∂t n + ∇x · (nu) = 0, (b) ∂t m∗ nu + ∇x · (m∗ uuT n + P ) + n∇x V =< p >coll , (5.7)

|p|2
(c) ∂t n(m∗ |u|2 + dkB T ) + ∇x · [n(m∗ |u|2 + dkB T )u + 2P u + 2qH ] + 2n∇x V · u =< >coll
m∗
with the pressure tensor P and the heat flux qH given by
Z Z
1 ∗ ∗ T m 1
P = ∗ (p − m u)(p − m u) fw dp, 2qH = |p − m∗ u|2 (p − m∗ u)fwm dp . (5.8)
m Rd (m∗ )2 Rd

The closure, i.e. P and qH , has to be computed using the closure Wigner function fwm which
in turn is given by the closure operator ρµm via (5.6). Note that the quantum hydrodynamic
equations coincide with the classical hydrodynamic system (2.25)-(2.27), except for the form
of the pressure tensor P and the heat flux qH . (For the Maxwellian closure (2.23), P = kB nT I
2
and qH = 0 holds.) The terms < p >coll and < |p| m∗
>coll on the right hand side of (5.7) denote
the corresponding moments of the collision operator evaluated at fwm and have to be computed
from the closure in terms of n, u, T in the same way as the pressure tensor and the heat flux.
They vanish identically in the case when collision operator Qw conserves mass momentum and
energy.
We now turn to the computation of fwm and ρµm . Let λ be a Lagrange multiplier, i.e. a
d+2-tuple of real valued functions of x. Using the interpretation of classical thermodynamics,
we can write λ = ( µT̄c , T̄ū , − T̄1 ) where µc (x), ū(x), and T̄ (x) will be called respectively local
chemical potential, velocity and temperature. Note that ū and T̄ will in general be different
from the ensemble velocity and temperature u and T in (5.7) which are defined directly from
the moments of the Wigner function, in contrast with the classical case for which u = ū and
T = T̄ holds. Then:
|p|2
 
1
λ(x) · κ(p) = µc (x) + ū(x) · p − . (5.9)
T̄ (x) 2m∗

19
A straightforward computation leads to the expression of Op(λ · κ):

Op(λ · κ)φ := Op[µc , ū, T̄ ]φ

~2  µ 1 ~2

1 1  ū  ū
c 1
= ∇ · ( ∇φ) − i~ ∇ · φ + · ∇φ + + ∆ φ . (5.10)
2m∗ T̄ 2 T̄ T̄ T̄ 4 2m∗ T̄

It is readily checked that Op[µc , ū, T̄ ] is a Hermitian operator.

Now, as soon as T̄ (x) ≥ 0 a.e. (which we are going to assume from now on), Op[µc , ū, T̄ ]
is an unbounded operator from below and is bounded from above. Thus, assuming that the
spectrum of Op[µc , ū, T̄ ] consists of a discrete sequence of eigenvalues (a` [µc , ū, T̄ ])`=1,...,∞ , we
have lim`→∞ a` = −∞. Let us denote by φ` the associated Hilbert basis of eigenfunctions. By
definition, the equilibrium operator ρλ = ρµc ,ū,T̄ = (h0 )−1 (Op[µc , ū, T̄ ]) is given, for any φ ∈ X
by:
∞
X
ρµc ,ū,T̄ φ = α` [µc , ū, T̄ ] (φ, φ` )X φ` , α` [µc , ū, T̄ ] = (h0 )−1 (a` [µc , ū, T̄ ]) . (5.11)
`=1

Like in section 2, let us denote Rby m = (n, q, W ) the vector of moments, with n(x) > 0 the
probability of presence (satisfying n(x) dx = 1), q(x) ∈ Rd the mean momentum per particle,
and W (x) the mean energy per particle (i.e. respectively the density, momentum and energy
divided by the total number of particles N ). The equilibrium operator ρn,q,W , subject to the
constraints (3.20) is a solution of the maximization problem (3.18), which, in the present case,
is written:
(∞ Z   )
X µc ū 1
max g(a` [µc , ū, T̄ ]) + (x)n(x) + (x) · q(x) − (x)W (x) dx , (5.12)
µc ,ū,T̄
`=1 Rd T̄ T̄ T̄

with g(s) = h ◦ (h0 )−1 (s) − s (h0 )−1 (s). We note that the first term in the curly bracket of (5.12)
is nothing but the Massieu-Planck potential (4.3) and that (5.12) itself is a reformulation of
relation (4.4) as a minimization problem.
Now, if we specialize to the Boltzmann entropy (2.22), we have (h0 )−1 (s) = es , g(s) = −es ,
and so ρµc ,ū,T̄ = exp{Op[µc , ū, T̄ ]} is given by (5.11) with

α` [µc , ū, T̄ ] = ea` [µc ,ū,T̄ ] . (5.13)

Since ρµc ,ū,T̄ is a statistical operator, it satisfies (3.1) and therefore, we must have
∞
X
ea` [µc ,ū,T̄ ] = 1 . (5.14)
`=1

In particular, this implies that all eigenvalues of Op[µc , ū, T̄ ] must be strictly negative or in
other words, that −Op[µc , ū, T̄ ] must be an elliptic operator. Furthermore, the eigenvalues
must tend sufficiently fast to −∞ for the series (5.14) to be convergent and have a sum equal
to 1. Therefore, we must restrict the set of trial functions of the maximization problem to

20
functions which guarantee such a property to the operator. Now, the optimization problem
(5.12) for the Boltzmann entropy (2.22) is written:
( ∞ Z   )
X µc ū 1
max − ea` [µc ,ū,T̄ ] + (x)n(x) + (x) · q(x) − (x)W (x) dx , (5.15)
µc ,ū,T̄
`=1 Rd T̄ T̄ T̄

We can guess that this problem has some chances to have a unique solution. For instance,
if we focus on the extremal value with respect to T̄ (with fixed ū/T̄ and µc /T̄ ), we see that
a` [µc , ū, T̄ ] is an increasing function of T̄ . Therefore, the first term in (5.15) is a decreasing
function of T̄ while the second term is an increasing one. Furthermore, as T̄ approaches 0,
a` [µc , ū, T̄ ] → −∞, exp a` → 0 and the first term tends to 0, while the second one tends to −∞
(we suppose that W > 0). Conversely, if T̄ → ∞, we have a` [µc , ū, T̄ ] → 0, exp a` → 1 and the
sum diverges, making the first term tend to −∞, while the second one tends to 0. Therefore,
the expression to be maximized in (5.15) tends to −∞ at the boundaries of the domain of
variation of T̄ . Of course, this argument is not rigorous, since T̄ is a function and some care
has to be taken in making sense to the fact that T̄ tends to 0 or to ∞. Also, the influence of
the other parameters ū/T̄ and µc /T̄ has to be studied. A rigorous investigation of this problem
is deferred to future work.
In classical mechanics, the system of hydrodynamics equations is hyperbolic, i.e. the matrix
of the derivatives of the flux functions with respect to the state variables is diagonalizable with
real eigenvalues. These eigenvalues are the speeds of propagation of the various types of waves.
Asking whether system (5.7) is hyperbolic in this sense would be meaningless. Indeed, since the
flux functions are not local functions of the state variables, it would be meaningless to compute
the matrix of derivatives in this way. Hyperbolicity provides a well-posedness theory, at least
locally in time [35], [33]. The well-posedness of quantum hydrodynamics systems is an open
problem so far. The fact that an entropy is decreasing in time as the following proposition
states, should be an important ingredient in such a theory.

Proposition 5.1 Let the collision operator Qw in (1.1) dissipate the entropy, i.e. let

Tr{Q(ρ)h0 (ρ)} ≤ 0, ∀ρ , (5.16)

hold. Then, any solution (n, u, T ) of system (5.7) satisfies the entropy dissipation relation:

∂
S(n, q, W ) ≤ 0 . (5.17)
∂t

Proof: To prove the entropy conservation relation, we go back to the notation m = (n, q, W )
and use µm = (µc /T, u/T, −1/T ) for the Lagrange multiplier of the constraint m. We write,
thanks to (4.2): Z
d δS ∂m ∂m
S(m) = = µm dx .
dt δm ∂t Rd ∂t

21
Now, using (5.4) and the cyclicity of the trace, we have
 
d i  m
 0
S(m) = Tr − Ĥ ρµm(t) , Op(µ · κ) + h (ρµm(t) )Q(ρµm(t) ) .
dt ~
But, by construction, the operators Op(µm · κ) and ρµm(t) commute. Therefore, the right-hand
side is less than zero and the result follows.

We note that entropy conservation is not restricted to the quantum hydrodynamic model
(5.7) but is valid for all quantum moment closure systems (5.4). In clasical hydrodynamics, it
is a well established fact that the entropy of smooth solutions is constant in time. However,
classical hydrodynamic models have discontinuous solutions (shock waves) the entropy of which
is strictly decreasing with time. It is very unlikely that quantum hydrodynamic models exhibit
shock waves solutions. The meaning of the model for discontinuous solutions would even be
very unclear. For instance, what sense should we give to the operator (5.10) if the coefficients
are discontinous functions ?
Another interesting question is how the closure (5.7)-(5.8) relates to the Bohmian single
state closure analyzed in [16] and [17]. The single state closure corresponds to the case of zero
temperature, when all particles become statistically completely independent, i.e. the Boltzmann
or Fermi-Dirac distribution reduces to a δ− function. Since, when using the Boltzmann entropy,
we close the moment system essentially by exp(Op(µκ)) the same limit can probably be carried
out by letting T → 0 in (5.7)(a)(b) for finite ~. Carrying out this limit is however not by no
means simple, since it involves computing the limiting solution of the minimization problem.

5.3 Quantum moment hydrodynamics

With this term, we refer to all quantum moment closure systems (5.4) constructed from a larger
basis of monomials than that of hydrodynamics. Therefore, we consider a finite subset B of Nd
(where N is the set of natural integers) and consider multi-indices β = (β1 , . . . , βd ) ∈ B. We
suppose that 0 ∈ B, ei ∈ B for all i ∈ {1, . . . , d} (where ei = (βj )j=1,...,d with βj = δij ). Let b be
the cardinal of B and consider the b-tuple of moment monomials κ(p) = (pβ )β∈B . Additionally,
we suppose that the space generated by κ(p) contains |p|2 . In this way, it contains the hydro-
dynamic monomials (1, pi , |p|2 ) and the associated quantum closure system (5.4) will contain
the quantum hydrodynamic system (5.7) as a subcase. The vector κ may contain homogeneous
polynomials which are not monomials, like |p|4 . However, for notational simplicity, we shall
restrict to κ(p) constructed from simple monomials pβ and leave this straightforward extension
to the reader.
We can now reproduce the procedure developed in the previous section. We just summarize
it now. First, let us denote by λ(x) = (λβ )β∈B a b-tuple of Lagrange multipliers. Then, the
operator Op(λ · κ) is given by:
X
Op(λ · κ) = Op(λβ · pβ ) , (5.18)
β∈B

where the expression of Op(λβ · pβ ) is given in Lemma 3.1.

22
 From the analysis conducted in section 5.2 (at least with the Boltzmann entropy (2.22)), the
maximization problem (3.18) has a non empty set of solutions only if there is a non empty set of
λ such that the operator −Op(λ·κ) is elliptic. In fact, we shall request that the set of such λ has
a non empty interior. This condition is analogous to condition (III) of Levermore’s approach
[26] and is likely to lead to identical constraints on the set of monomials to be considered. This
point will be investigated in more detail in future work. This requirement leads to a restriction
on the possible sets B.
We now suppose that this requirement is fulfilled and that the operator Op(λ · κ) has a
sequence of negative eigenvalues (a` [λ])`=1,...,∞ such that a` [λ] → −∞ as ` → ∞ and that
the associated eigenvectors φ` form a complete orthonormal basis. Let us now denote by
m(x) = (mβ )β∈B a given set of moments. The maximization problem (3.18) is now stated as
follows (from now on, we shall restrict to the case of the Boltzmann entropy (2.22)):
( ∞ Z X )
X
max − ea` [λ] + λβ (x)mβ (x) dx . (5.19)
λ Rd β∈B
`=1

Let us denote by µm the value of λ which solves this maximization problem (assuming that the
solution exists and is unique) and ρm = ρµm , the associated operator
∞
m
X
ρµ m φ = α` [µm ] (φ, φ` )X φ` , α` [µm ] = ea` [µ ] . (5.20)
`=1

We shall derive the quantum moment hydrodynamics system in a similar form as for the
quantum hydrodynamics system. The derivation of a form like (5.7) shall be done in a future
work. Let us consider the β-th component λβ of an arbitrary λ and denote by bβ`r [λβ ] the matrix
element of the operator Op(λβ pβ ) in the eigenbasis φ` of ρm . As in section 5.2, we denote by
Ĥ`r the matrix element of the Hamiltonian in the same basis, and by Q`r (ρ) the matrix element
of Q(ρ). We have:
nh i o X
Tr Ĥ, ρm Op(λβ pβ ) = Ĥ`r (αr − α` )bβr` [λβ ] . (5.21)
`,r

Finally, the system of quantum moment hydrodynamics equations can be written according
to:
Z
∂ iX X
mβ (x, t)λβ (x) dx = − Ĥ`r (αr − α` )bβr` [λβ ] + Q`r (ρ)bβr` [λβ ] , (5.22)
∂t Rd ~ `,r `,r
∀λβ (x) , ∀β ∈ B , (5.23)
This provides an evolution system for the quantities mβ , which is well adapted to a Galerkin
discretization. For this system also, provided that Q is entropy dissipative, (i.e. property (5.16)
is satisfied), the entropy dissipation inequality (5.17) applies and the entropy S(m) is decreasing
in time.
We point out that the solution of the maximization problem (5.20) does not always exist in
the classical case (see in particular [24], [25]). It is yet an open problem, and formidably more
difficult, to solve it in the quantum case.

23
5.4 A Galerkin discretization
The main difficulty in solving the quantum hydrodynamic model (5.7) is of course the evaluation
of the pressure tensor and the heat flux vector through formulae (5.8). Indeed, this evaluation
requires the solution of the optimization problem (5.12) or (5.15). We now roughly outline the
computational complexity (and feasibility) of this problem. The moment equations (5.3) render
themselves to a natural Galerkin discretization in space, which preserves the entropy principle.
We start by choosing a set of scalar basis functions λj (x), j = 1, .., J in space and approximate
the moment vector m = (m0 , .., mN ) in (5.3) by
J
X
mn (x, t) ≈ znj (t)λj (x), n = 0, .., N . (5.24)
j=1

Equation (5.3) is then replaced by

Z
i
λj (x) ∂t m(x, t)dx = Tr{(− [Ĥ, ρm ] + Q(ρm ))Op(λj κ)}, j = 1, .., J , (5.25)
Rd ~
where ρm is the solution of the constrained minimization problem
Z
m
Tr(h(ρ )) = min{Tr(h(ρ)) : Tr{ρOp(λj κ)} = λj mdx, j = 1, .., J} . (5.26)
Rd

Thus, we have replaced the minimization problem (3.11) by a problem with finitely many
constraints leading, consequently to only J(N + 1) Lagrange multiplyers. Next, we replace
the density matrix ρ in (5.26) by a finite expansion. We choose orthonormal basis functions
ψk (x), k = 1, .., K, and approximate the density matrix ρ by
K
X
ρm (x, y) ≈ m
Rkl rkl (x, y), rkl (x, y) = ψk (x)ψl (y)∗ ,
k,l=1
m
where Rkl is a Hermitian matrix and the star exponent denotes complex conjugation. In order
to arrive at a finite minimization problem, we also have to replace the operator Op(λj κ) by a
finite matrix, i.e.
XK Z Z
jn jn
Op(λj κ) ≈ Γkl rkl (x, y), Γkl = ψk (x)∗ Op(λj κn )(x, y) ψl (y)dxdy ,
k,l=1 Rd Rd

and replace the minimization problem (5.26) by

Z
m jn
Tr(h(R )) = min{Tr(h(R)) : Tr{RΓ } = λj mn dx, j = 1, .., J, n = 0, .., N } . (5.27)
Rd

The symbol Tr in (5.27) denotes now just the usual matrix trace and h is now to be understood
as the function of a matrix. The Galerkin equations (5.25) are then replaced by
Z K
X i
λj (x)∂t mn (x, t)dx = Γjn m m
kl Tr{(− [Ĥ, ρ ] + Q(ρ ))rkl }, j = 1, .., J, n = 0, .., N .
R d
kl=1
h
(5.28)

24
(5.28) is a system of ordinary differential equations for the expansion coefficients znj (t) after
inserting (5.24) for the components mn of the moment vector. The system (5.27)-(5.28) now
clearly satisfies the same entropy relation as given by Proposition 5.1 on a discrete level. This
can be seen by multiplying (5.28) by µjn and summation over jP and n, where µjn is the Lagrange
multiplier of the minimization problem (5.27), i.e. h0 (Rm ) = jn µjn Γjn holds. Two questions
arise:
• How to choose the basis functions ψk for the density matrices and the basis functions λj
for the moments ?

• How large should K 2 , the number of density matrix basis functions, be compared to J,
the number of moment basis functions?
There is considerable freedom in the answer to the first question. One possibility would be
to choose the ψk as the eigenfunctions of ρm itself, making the matrix Rm diagonal. This
makes the evaluation of h(R) trivial, but has the disadvantage that we have to compute the
matrices Γjn and the matrix corresponding to the Hamiltonian Ĥ anew for each time step. It is
probably preferable to choose a ’good’ basis {ψk } (say eigenfunctions of the Hamiltonian, which
are also the eigenfunctions of the equilibrium solution e−β Ĥ ), compute the Γjn once and for all,
and rather deal with the problem of computing the matrix function h(R) in the optimization
procedure. An Palternative would be to use the entropy variables µjn as primary variables, since
m 0 −1 jn
R = (h ) ( jn µjn Γ ) has to hold. This has the advantage of eliminating the optimization
procedure but makes the ODE system (5.28) implicit, since the term on the left hand side of
(5.28) is then given by
Z
0 0
X
λj mn dx = Tr{Γjn (h0 )−1 ( µj 0 n0 Γj n )} .
Rd j 0 n0

The easiest way to invert this equation might again be by using an optimization procedure (i.e.
a discrete version of (5.19). As to the second question: The number of degrees of freedom in
the minimization problem (5.27) is K 2 (the number of elements in a K × K hermitian matrix
if we count real and imaginary parts separately). Therefore, K 2 > J(N + 1) has to hold, in
order to have more variables than constraints. How much larger K 2 has to be will depend on
how well we wish to approximate the minimization problem. This will have to be determined
by numerical experiment.
We conclude this section with a remark about the usefulness of Wigner functions in this
context. While Wigner functions are convenient for expressing local moments they are not the
appropriate tool in this approach. The reason for this is the following. The reason why Wigner
functions were useful in expressing the moment equations is that the moments of [H, ρ] can be
expressed in terms of the moments of ρ in the Wigner picture, up to a few closure terms. This
property is lost here, due to taking a finite number of expansion terms for the density matrix.
If we define by fkl the Wigner transform of the basis element ρkl of the density matrix, we
obtain Z Z
jn
Γkl = Tr{Op(λj κn )ρkl } = λj κn fkl drdp
Rd Rd

25
and (5.28) becomes
Z Z Z K
X 1
λj (x)∂t mn (x, t)dx = Γjn
kl {−∇x ( pf m ) + θ[V ]f m + Qw (f m )}fkl dxdp, (5.29)
Rd Rd Rd kl=1 m∗

j = 1, .., J, n = 0, .., N .
Written out in more detail, the right hand side is of the form
Z K
X 1
λj (x0 )κn (p0 )fkl (x0 , p0 )fkl (x, p){−∇x ( pf m ) + θ[V ]f m + Qw (f m )}(x, p)dxdpdx0 dp0 .
R4d kl=1 m∗

If now
K
X
fkl (x0 , p0 )fkl (x, p) = δ(x0 − x)δ(p0 − p)
kl=1

were to hold, the right hand side of (5.29) would simplify in the same way as it did in Section
5.2. For a finite number K of expansion terms this can, however, not be assumed.

5.5 Quantum BGK model

Finally, we end this section about the quantum moment hierarchies by outlining the use which
can be made of the equilibrium density matrix ρm under the moment constraint m. Indeed, a
density matrix equation including relaxation to local equilibria can be written in the spirit of
the BGK model of rarefied gas dynamics. Such a model can be written:
∂ρ i
= − [Ĥ, ρ] − νc (ρ − ρm[ρ(t)] ) , (5.30)
∂t ~
where we haved denoted by m[ρ(t)] the moments associated with ρ at time t according to (3.5).
Usually, we shall take the hydrodynamic moments. We shall prove in [8] that such an operator
is consistent with entropy decay. The quantity νc is a collision frequency. By construction,
composing on the right by Op(λ · κ) where κ are the hydrodynamic monomials shows that
the relaxation term ρ − ρm[ρ(t)] does not appear in the evolution of the hydrodynamic moment
equations (5.3). However, the scaling νc → νc /ε, where ε  1 is a small parameter representing
the Knudsen number (i.e. the ratio of the collision scale to the macroscopic scale), allows
to justify (at least formally) the hydrodynamic closure (5.4). In most of the literature, only
relaxation towards global equilibria are taken into account through equilibria ρm associated with
global constraints (like in remark 4.1). The present approach allows to give a meaning to ’local
equilibrium relaxation’ in a quantum framework. In fact, the term ’local’ is slightly misleading
in this case. It refers to the fact that moment constraints are functions of x. However, the
relation between the moment constraints m and the equilibrium density matrix ρm is not a
local one, but a functional one.
A natural question is the existence of a series of perturbative models obtained through an
expansion of the solutions of (5.30) in powers of ε (after rescaling νc → νc /ε), in the spirit of

26
the Chapman-Enskog expansion of the Boltzmann equation of gas dynamics (see [26]). This
question will be investigated in future work.
Another question is related to the possible existence of ’diffusion-like’ limits of model (5.30)
(after a simulatneous rescaling νc → νc /ε and ∂ρ/∂t → ε∂ρ/∂t, with ε  1). This question is
investigated in [8].

6 Conclusion and future work

We have developed a systematic approach to construct quantum hydrodynamical models from
appropriate closures of the quantum Liouville equation. This approach is made possible by
an adequate definition of the local moment of a density operator. This concept is defined by
duality through the observations of the system on a suitable class of test functions. Then,
the concept of an equilibrium density matrix, a solution of the entropy minimization principle
subject to the constraints of given moments, is developed. Taking successive moments of the
quantum Liouville equation (in the above defined sense), it is possible to close the chain of
equations by this equilibrium operator, giving rise to a hierarchy of models called Quantum
Moment Hydrodynamic models. We have studied in more detail the quantum hydrodynamic
model corresponding to the usual hydrodynamical moments and have shown that it is formally
similar as the classical one apart from nonlocal expressions of the pressure tensor and heat-flux
vector.
Obviously, this paper leaves a lot of mathematical questions open: existence of solutions for
the entropy minimization problem, well-posedness of the quantum hydrodynamic equations,
derivation of efficient numerical solvers, well-posedness of quantum BGK models, existence of
perturbative series expansions of solutions in the spirit of the Chapman-Enskog expansion,
derivation of quantum diffusion models (like the quantum drift-diffusion), etc. Other pending
questions are concerned with the physical restriction of the model. One can think of taking
into account non constant particle number (thus requesting the use of Fock spaces), Fermi-
Dirac or Bose-Einstein statistics, spin and relativity effects together with the inclusion of the
magnetic field through the Dirac equation, etc. All these questions clearly open a large field of
investigations for the future.

27
References
[1] P. Andriès, P. Le Tallec, J. P. Perlat and B. Perthame, The Gaussian-BGK model of
Boltzmann equation with small Prandtl number, preprint.

[2] P. Argyres: Quantum kinetic equations for electrons in high electric and phonon fields,
Phys. Lett. A 171 North Holland ,1992.

[3] A. Arnold, J. Lopèz, P. Markowich, J. Soler: An analysis of quantum Fokker - Planck

models: A Wigner function approach, preprint ,2002.

[4] R. Balian, From microphysics to macrophysics, Springer, 1982.

[5] J. Barker, D. Ferry: Phys. Rev. Lett. 42 ,1997.

[6] A. Caldeira, A Leggett: A path integral approach to quantum Brownian motion Physica
A, 121:587-616 , 1983.

[7] P. Degond, Mathematical modelling of microelectronics semiconductor devices, Proceedings

of the Morningside Mathematical Center, Beijing, AMS/IP Studies in Advanced Mathe-
matics, AMS Society and International Press, 2000, pp. 77–109.

[8] P. Degond, F. Méhats, C. Ringhofer, On a Quantum Energy-Transport model, in prepara-

tion.

[9] P. Degond, C. Ringhofer, A note on quantum moment hydrodynamics and the entropy
principle, C. R. Acad. Sci. Paris Ser 1 335 (2002), pp. 967–972.

[10] P. Degond, C. Ringhofer, A note on binary quantum collision operators conserving mass
momentum and energy, submitted.

[11] D. Ferry, H. Grubin, Modelling of quantum transport in semiconductor devices, Solid State
Phys. 49 (1995), pp.283–448 .

[12] F. Fromlet,P. Markowich,C. Ringhofer: A Wignerfunction Approach to Phonon Scattering,

VLSI Design 9 pp.339-350 ,1999.

[13] C. Gardner, The quantum hydrodynamic model for semiconductor devices, SIAM J. Appl.
Math. 54 (1994), pp. 409–427.

[14] C. Gardner and C. Ringhofer, The smooth quantum potential for the hydrodynamic model,
Phys. Rev. E 53 (1996), pp.157–167.

[15] C. Gardner and C. Ringhofer, The Chapman-Enskog Expansion and the Quantum Hydro-
dynamic Model for Semiconductor Devices, VLSI Design 10 (2000), pp. 415–435.

[16] I. Gasser and A. Jüngel, The quantum hydrodynamic model for semiconductors in thermal
equilibrium, Z. Angew. Math. Phys. 48 (1997), pp. 45–59 .

28
[17] I. Gasser and P. A. Markowich, Quantum Hydrodynamics, Wigner Transforms and the
Classical Limit, Asympt. Analysis, Vol. 14, No. 2, pp. 97-116,1997.

[18] I. Gasser, P. Markowich and C. Ringhofer, Closure conditions for classical and quantum
moment hierarchies in the small temperature limit, Transp. Th. Stat. Phys. 25 (1996), pp.
409–423.

[19] H. Grad, On the kinetic theory of rarefied gases, Comm. Pure Appl. Math. 2 (1949), pp.
331–407.

[20] H. Grubin, J. Krekovski, Quantum moment balance equations and resonant tunneling struc-
tures, Solid-State Electron. bf 32 (1989), pp. 1071-1075.

[21] B. Hellfer, J. Sjöstrand, Equation de Harper, Lect. Notes in Physics 345 (1989), pp. 118–
197, Springer.

[22] S. Jin and X. T. Li, Multi-phase Computations of the Semiclassical Limit of the Schrodinger
Equation and Related Problems: Whitham vs. Wigner, preprint, University of Wisconsin
(http://www.math.wisc.edu/ jin/publications.html).

[23] A. Jüngel, P. Pietra, A discretization scheme for a quasi-hydrodynamic semiconductor

model, Math. Models Meth. Appl. Sci. 7 (1997), pp. 935–955.

[24] M. Junk, Domain of definition of Levermore’s five-moment system, J. Stat. Phys. 93

(1998), pp. 1143–1167.

[25] M. Junk, Maximum entropy for reduced moment problems, Math. Methods and Models in
the Applied Sciences 10 (2000), pp. 1001–1025.

[26] C. D. Levermore, Moment closure hierarchies for kinetic theories, J. Stat. Phys., 83 (1996),
pp. 1021–1065.

[27] R. Luzzi, untitled, electronic preprint archive, reference arXiv:cond-mat/9909160 v2 11 Sep

1999.

[28] V. G. Morozov and G. Röpke, Zubarev’s method of a nonequilibrium statistical operator

and some challenges in the theory of irreversible processes, Condensed Matter Physics 1
(1998), pp. 673–686.

[29] I. Muller and T. Ruggeri, Rational Extended Thermodynamics, Springer Tracts in Natural
Philosophy, volume 37, Second edition. 1998.

[30] F. Nier, A variational formulation of Schrödinger-Poisson systems in dimension d ≤ 3,

Comm. PDE 18 (1993), pp. 1125–1147.

[31] M. Reed and B. Simon, Methods of modern mathematical physics, vol 1, Functional
analysis, Academic Press, 1980.

29
[32] J. Schneider, Entropic approximation in kinetic theory, preprint, Analyse non linaire ap-
plique et modlisation, Universit de Toulon et du Var, BP 132, 83957 La Garde cedex,
France (2001).

[33] D. Serre, Systmes de lois de conservation, tome I et II Diderot, Paris, 1996.

[34] M. A. Shubin, Pseudodifferential operators and spectral theory, Springer, 1980.

[35] J. Smoller, Shock waves and reaction-diffusion equations, Grundlehren der mathematis-
chen wissenschaften 258, Springer, New-York (2nd edition), 1994.

[36] D. N. Zubarev, V. G. Morozov and G. Röpke, Statistical mechanics of nonequilibrium

processes. Vol 1, basic concepts, kinetic theory, Akademie Verlag, Berlin, 1996.

30
7 Appendix: Quantum entropy minimization principle:
proofs
In this appendix, we give the proofs of the statements related with the quantum entropy
minimization principle.
Proof of Lemma 3.5: The Euler-Lagrange equation for the minimization problem (3.16)
is      
δLm  δH  δKλ 
= − = 0. (7.1)
δρ (ρλ ,λ) δρ ρλ δρ ρλ
Thanks to the linearity of Kλ (ρ) with respect to ρ, we have, for any self-adjoint, trace class
operator δρ:  
δKλ  X
δρ = Tr{(δρ)Op(λ · κ)} = δρ`r ar` ,
δρ ρ `,r

where we denote by a`r the matrix element of the operator Op(λ · κ) in the eigenbasis of ρ.
Then, thanks to (3.14), the Euler-Lagrange equation (7.1) is written:
∞ ∞
!
X X
h0 (α` (ρ))δρ`` − δρ`r ar` = 0 ,
`=1 r=1

for all Hermitian, trace class operators δρ. Choosing δρ`0 r0 = δ`0 ` δr0 r + δ`0 r δr0 ` or δρ`0 r0 =
i(δ`0 ` δr0 r − δ`0 r δr0 ` ) where δ`0 r0 is the Kronecker symbol, with ` 6= r, we deduce that a`r must
be equal to zero, i.e. that ρ must be diagonal in the basis where the operator Op(λ · κ) is
diagonal. Such a basis always exist, since Op(λ · κ) is Hermitian (because λ(x) and κ(p) are
real functions) at least in the generalized sense, (i.e. in the sense of the spectral measure if
Op(λ · κ) has continuous spectrum, see [31]). Then, taking δρ`0 r0 = δ`0 ` δr0 ` , we deduce that

h0 (α` (ρ)) = a`` ,

i.e.
α` (ρ) = (h0 )−1 (a`` ) ,
which is precisely the definition of ρ being given by (3.17).

Proof of Lemma 3.6: The Euler-Lagrange equation of the maximization problem reads
(where we drop the superscript m):
     
δLm  δρλ  δLm 
+ = 0. (7.2)
δρ (ρµ ,µ) δλ µ δλ (ρµ ,µ)

But, by the Euler-Lagrange equation of the minimization problem (7.1), since ρµ realizes the
minimum, the first term is identically zero. Therefore, µ is characterized by
 
δLm 
= 0.
δλ µ

31
By linearity of Lm with respect to λ, we have:
   Z 
δLm 
δλ = − Tr{ρµ Op(δλ · κ)} − m(x) · δλ(x) dx = 0 ,
δλ µ Rd

for any N-tuple δλ = (δλi (x))i=1,...,N of arbitrary functions δλi (x), which is exactly relation
(3.20) (with δλ = λ). It is easy to show that this extremum is indeed a maximum. This
classical point is left to the reader.

Proof of Lemma 4.1: (i) We use that H is strictly convex (by virtue of Lemma 3.4.
Therefore, let m and m0 be two moment vectors and t ∈ [0, 1]. Then:
0
S(tm + (1 − t)m0 ) = H(ρtm+(1−t)m )
Z
= min {H(ρ) | Tr{ρOp(λ · κ)} = λ · (tm + (1 − t)m0 ) dx , ∀λ(x)} .
Rd

But we have
0 0
Tr{(tρm + (1 − t)ρm )Op(λ · κ)} = tTr{ρm Op(λ · κ)} + (1 − t)Tr{ρm Op(λ · κ)}
Z Z
= t λ · m dx + (1 − t) λ · m0 dx
d Rd
Z R
= λ · (tm + (1 − t)m0 ) dx
Rd

0
Therefore, tρm + (1 − t)ρm satisfies the constraints and consequently:
0 0
H(ρtm+(1−t)m ) ≤ H(tρm + (1 − t)ρm ) .

Now, using the strict convexity of H, we have:

0 0
H(ρtm+(1−t)m ) < tH(ρm ) + (1 − t)H(ρm ) ,

or,
S(tm + (1 − t)m0 ) < tS(m) + (1 − t)S(m0 ) ,
which proves the strict convexity of S, i.e. point (i).
(ii) We write:

δS δ
= (Lm (ρµm , µm ))
δm δm   m
δLm m δLm m δρλ m δLm m δµ
= (ρµm , µ ) + (ρµm , µ ) (µ ) + (ρµm , µ )
δm δρ δλ δλ δm
δLm
= (ρµm , µm ) ,
δm

32
where the last equality follows from the fact that µm is a solution of the Euler-Lagrange equation
(7.2). Therefore, for any perturbation δm = (δmi (x))i=1,...,N , we have
Z
δS δLm m
δm = (ρµm , µ )δm = µm (x) · δm(x) dx ,
δm δm Rd

which exactly means (4.2) by duality.

Proof of Lemma 4.2: (i) Σ can be defined as

Z
Σ(µ) = min {S(m) − µ · m dx} ,
m Rd

and a minimum of linear functions, which are concave functions, is concave.

(ii) We have: Z Z
δΣ δS δm δm
δµ = δµ − δµ · m dx − µ· δµ dx .
δµ δm δµ Rd Rd δµ
But, with (4.2), the first term of the right-hand side cancels the last one and we have
Z
δΣ
δµ = − δµ · m dx ,
δµ Rd

which proves (4.4).

8 Appendix: properties of the quantum entropy H

The proofs of lemmas 3.3 and 3.4 can be found e.g. in [30]. Here, we give an elementary proof
of these results.
Proof of Lemmas 3.3 and 3.4: Let δρ be a self-adjoint trace-class operator. By the definition
of the Gâteaux differentiability, we wish to investigate if the following limit exists:
 
δH 1
δρ = lim (Tr{h(ρ + tδρ)} − Tr{h(ρ))} .
δρ t→0 t

We can choose variations δρ the expressions of which are given by either of the following
formulae in the eigenbasis φ` of ρ :
θ
δρ`0 r0 = (δ`0 ` δr0 r + δ`0 r δr0 ` ) , (8.1)
2
iθ
δρ`0 r0 = (δ`0 ` δr0 r − δ`0 r δr0 ` ) , (8.2)
2
where δ`0 r0 is the Kronecker symbol and θ is a real number. Indeed, any variation δρ can be
decomposed into a (possibly infinite but convergent) sum of such variations. The elementary

33
δρ is still trace-class and self adjoint. It is then enough to investigate the influence of the
perturbation on the space spanned by (φ` , φr ). We investigate the following three cases:
(i) ` = r. Then, α` (ρ + tδρ) = α` (ρ) + tθ and therefore,
t2 θ2 00
h(α` (ρ + tδρ)) = h(α` (ρ)) + tθ h0 (α` (ρ)) + h (α` (ρ)) + o(t2 ) ,
2
as t → 0. Therefore, in this case,
δH δ2H
δρ = h0 (α` (ρ))δρ`` , 2
(δρ, δρ) = h00 (α` (ρ))|δρ`` |2 .
δρ δρ

(ii) ` 6= r and α` 6= αr . Then, in the basis (φ` , φr ), the matrix ρ + tδρ is written as follows
(respectively in the cases (8.1) and (8.2)):
   
α` θt/2 α` iθt/2
ρ + tδρ = or ρ + tδρ = .
θt/2 αr −iθt/2 αr
In these two cases, the eigenvalues of ρ + tδρ are the same. Suppose that α` > αr to fix the
ideas. Then, they are given by
1 p  t2 θ2
α` (t) = α` + αr + (α` − αr )2 + θ2 t2 = α` + + o(t2 ) ,
2 4(α` − αr )
1  p  t2 θ2
αr (t) = α` + αr − (α` − αr )2 + θ2 t2 = αr − + o(t2 ) ,
2 4(α` − αr )
as t → 0. Thus,
t2 θ2 h0 (α` ) − h0 (αr )
h(α` (t)) + h(αr (t)) = h(α` ) + h(αr ) + + o(t2 ) .
4 α` − αr
Therefore, in this case, we have:
δH δ2H h0 (α` ) − h0 (αr )
δρ = 0 , 2
(δρ, δρ) = (|δρ`r |2 + |δρr` |2 ) .
δρ δρ α` − αr

(iii) ` 6= r and α` = αr = α. Then,

α` (t) = α + |tθ|/2 , αr (t) = α − |tθ|/2 .
It follows that
t2 θ2 00
h(α` (t)) + h(αr (t)) = 2h(α) + h (α) + o(t2 ) .
4
Then, we deduce that:
δH δ2H
δρ = 0 , (δρ, δρ) = h00 (α) (|δρ`r |2 + |δρr` |2 ) .
δρ δρ2
Collecting the results in the above three cases, leads to formulae (3.14) and (3.15).

34