BLOCH APPROXIMATION IN HOMOGENIZATION

AND APPLICATIONS

CARLOS CONCA

, RAFAEL ORIVE

, AND MUTHUSAMY VANNINATHAN

SIAM J. MATH. ANAL. c 2002 Society for Industrial and Applied Mathematics
Vol. 33, No. 5, pp. 11661198
Abstract. The classical problem of homogenization of elliptic operators with periodically os-
cillating coecients is revisited in this paper. As is well known, the homogenization process in a
classical framework is concerned with the study of asymptotic behavior of solutions u

of boundary
value problems associated with such operators when the period > 0 of the coecients is small. In
a previous work by C. Conca and M. Vanninathan [SIAM J. Appl. Math., 57 (1997), pp. 16391659],
a new proof of weak convergence as 0 towards the homogenized solution was furnished using
Bloch wave decomposition.
Following the same approach here, we go further and introduce what we call Bloch approximation,
which will provide energy norm approximation for the solution u

. We develop several of its main

features. As a simple application of this new object, we show that it contains both the rst and
second order correctors. Necessarily, the Bloch approximation will have to capture the oscillations
of the solution in a sharper way. The present approach sheds new light and oers an alternative for
viewing classical results.
Key words. homogenization, Bloch waves, correctors
AMS subject classications. 35B27, 35A25, 42C30
PII. S0036141001382200
1. Introduction. In this paper, the classical problem of homogenization of el-
liptic operators with periodically oscillating coecients is revisited. As is well known,
the homogenization process is concerned with the study of the behavior of solutions
u

of boundary value problems associated with such operators when the coecients
are periodic with small period > 0. For an excellent introduction to this subject,
the reader is referred to the book of A. Bensoussan, J.-L. Lions, and G. Papanico-
laou [5]. In a previous work by C. Conca and M. Vanninathan [11], a new proof
of weak convergence of u

towards the homogenized solution u

was furnished using

Bloch wave decomposition. Following the same approach, we go further and introduce
what we call Bloch approximation of the solution u

. As a simple application of this

new object, we treat the problem of correctors in homogenization. At this point, it
is worthwhile to remark that the homogenized solution u

is merely the weak limit

of solutions u

as 0. The idea behind introducing correctors is to look for terms

(called rst order correctors) which, when added to the homogenized solution, pro-
vide an approximation in the energy norm for all suciently small. Second order
correctors yield an error estimate in the energy norm of order O(). The main feature
of Bloch approximation is that it contains both the rst and second order corrector
terms. Another important feature is that it is easily computable in principle.

Received by the editors January 17, 2001; accepted for publication (in revised form) September
18, 2001; published electronically April 8, 2002. This work has been partially supported by FONDAP
through its Programme on Mathematical-Mechanics.
http://www.siam.org/journals/sima/33-5/38220.html

Departamento de Ingeniera Matematica, Facultad de Ciencias Fsicas y Matematicas, and

Centro de Modelamiento Matematico, UMR 2071 CNRS-UChile, Universidad de Chile, Casilla 170/3
- Correo 3, Santiago, Chile (cconca@dim.uchile.cl).

Departamento de Matematica Aplicada, Facultad de Ciencias Matematicas, Universidad Com-

plutense de Madrid, 28040 Madrid, Espa na (orive@sunma4.mat.ucm.es).

IISc-TIFR Mathematics Programme, TIFR Center, P.O. Box 1234, Bangalore, 560012, India
(vanni@math.tifrbng.res.in).
1166
BLOCH APPROXIMATION 1167
Historically, a classical way of obtaining such correctors is to work in the physical
space and use multiscale expansion of the solution, which was rst introduced in the
basic book just cited. As we will see, the method of Bloch waves sheds new light
and oers an alternative for viewing the classical results. This method naturally
leads us to work in the Fourier space, and thus in a framework dual to the one
used in L. Tartar [21], whose method is very general and justies, in particular, the
rst term obtained via the multiscale expansion. (There are also other methods
of justication based on the analysis in physical space; cf. G. Nguetseng [17] and
G. Allaire [1].) However, it is important to mention that the Bloch wave method
does not presuppose any multiscale structure of the solution; on the contrary, such
a structure of the solution will be a consequence of the present method. Although
correctors are generally not unique, our approach yields a posteriori the same ones as
those obtained in [5].
Bloch waves and their applications are not by any means new. It is a classical tool
which has been in use in solid state physics since the pioneering paper of F. Bloch [6].
However, the basic idea was introduced in the mathematical literature much earlier
by G. Floquet [13]. For later developments, let us cite the works of F. Odeh and
J. Keller [18] and M. Reed and B. Simon [19]. A deep analysis concerning the par-
tial regularity of the spectrum of Schr odingers equation with periodic potential was
carried out by C. Wilcox [22]. Our point of view regards periodic medium as a per-
turbation of homogeneous ones. In this context, the book by T. Kato [15] provides
excellent analysis when the parameter of perturbation is a scalar. We end this rather
incomplete list by citing G. Allaire and C. Conca [2], [3], and P. Gerard et al. [14].
We feel it is also appropriate to cite a recent work by M. Avellaneda, L. Berlyand,
and J.-F. Clouet [4], in which the BlochFloquet approach is used to provide new ho-
mogenization results and handles the boundary layer terms for frequency dependent
problems. To conclude, let us refer the reader to C. Conca [8] for a more complete
general survey on Bloch waves.
Before proceeding further, we mention a word about the notations adopted in
what follows. Unless mentioned explicitly, the usual summation convention with
respect to the repeated indices is understood. The constants appearing in various
estimates independent of are generically denoted by c, c
1
, c
2
, etc. Apart from the
usual norms in Sobolev spaces H
1
, H
2
, we will also use the following seminorms:
[v[
H
1
=
_
_
_
N

j=1
_
_
D
j
v
_
_
2
L
2
_
_
_
1
2
, [v[
H
2
=
_
_
_
N

j,k=1
_
_
D
2
j,k
v
_
_
2
L
2
_
_
_
1
2
.
Now let us introduce the problem to be studied in this work. We consider the
operator
A
def
=

y
k
_
a
k
(y)

y

_
, y R
N
, (1.1)
where the coecients satisfy
_

_
a
k
L

#
(Y ), where Y =]0, 2[
N
, i.e., each a
k
is a
Y -periodic bounded measurable function dened on R
N
, and
> 0 such that a
k
(y)
k

[[
2
R
N
, y Y a.e.,
a
k
= a
k
k, = 1, . . . , N.
(1.2)
1168 C. CONCA, R. ORIVE, AND M. VANNINATHAN
For each > 0, we consider also the operator A

, where
A

def
=

x
k
_
a

k
(x)

x

_
with a

k
(x) = a
k
_
x

_
, x R
N
. (1.3)
In homogenization theory, it is usual to refer to x and y, the slow and the fast variables,
respectively. They are related by y =
x

. Associated with A

, let us consider the

boundary value problem
A

= f in , u

H
1
0
(), (1.4)
which is posed in an arbitrary bounded domain in R
N
and where f is a given
element in L
2
(). It is classical that the above problem admits one and only one
solution.
From the classical work [5], it is known that one can associate to A

a homogenized
operator A

given by
A

def
=

x
k
_
q
k

_
. (1.5)
The homogenized coecients q
k
are constants and their denition is given below.
The solution u

of (1.4) converges weakly in H

1
0
() to the so-called homogenized
solution u

characterized by
A

= f in , u

H
1
0
(). (1.6)
In the present paper, we do not consider the eects of boundaries, postponing
them to a subsequent article [9]. In the case of R
N
, it is natural to replace the
operator A

by (A

+I). In that case, if w

satises
_
(A

+I)w

= g in R
N
,
w

in H
1
(R
N
)-weak,
(1.7)
where g is a given function in L
2
(R
N
), then it can be seen that (see Proposition 6.1
below)
w

in L
2
(R
N
)-strong. (1.8)
In view of the above result, there is no concentration of L
2
-energy at innity, and
therefore throughout this paper we will consider a sequence u

and a function f
L
2
(R
N
) satisfying
_

_
A

= f in R
N
,
u

in H
1
(R
N
)-weak,
u

in L
2
(R
N
)-strong.
(1.9)
The central issue in the analysis of the rst order correctors is to obtain functions
u

1
H
1
(R
N
), which can be easily constructed and have the following characteristic
property:
|u

1
|
H
1
(R
N
)
0 as 0. (1.10)
BLOCH APPROXIMATION 1169
By denition, second order correctors u

2
H
1
(R
N
) will enjoy the property
|u

1

2
u

2
|
H
1
(R
N
)
c. (1.11)
One of the purposes of this article is to carry out a more general construction than
the classical one for correctors, namely, Bloch approximation

, which contains all

the above correctors and justies the procedure. Apart from this,

contains a lot
of information about the periodic medium which will be amply demonstrated in this
paper.
1.1. Survey of the previous results. In the classical book [5] the authors
obtain an asymptotic expansion (with y =
x

) of the form
u

(x) = u

(x) +
_

k
(y)
u

x
k
(x) + u
1
(x)
_
+
2
_

k
(y)

2
u

x
k
x

(x) +

(y)
u
1
x

(x) + u
2
(x)
_
+ .
(1.12)
Here,
k
is the unique solution of the cell problem
_

_
A
k
=
a
k
y

in R
N
,

k
H
1
#
(Y ), M
Y
(
k
)
def
=
1
[Y [
_
Y

k
dy = 0.
(1.13)
The function
k
is characterized as the unique solution of
_

_
A
k
=a
k
+a
km

y
m

y
m
(a
mk

)M
Y
(a
k
)M
Y
_
a
km

y
m
_
in R
N
,

k
H
1
#
(Y ), M
Y
(
k
) = 0.
(1.14)
Further, u
1
(x), u
2
(x), . . . are independent of and satisfy equations of the type A

u
j
=
g
j
in R
N
, where, for instance, g
1
(x) = b
jk
D
3
jk
u

, where b
jk
are constants:
b
jk
= M
Y
_
a
jm

k
y
m
+a
k

j
_
j, k, = 1, . . . , N.
With these notations, the classical formula of the homogenized coecients is as fol-
lows:
q
k
= M
Y
_
a
k
+a
km

y
m
_
k, = 1, . . . , N.
(Another characterization of q
k
is given in Proposition 1.5 below.) Using the above
expansion, the rst order corrector term is obtained in [5]. More precisely, we have
the following.
Theorem 1.1. We assume that the coecients a
k
satisfy assumptions (1.2),
f L
2
(R
N
), and the solution
k
W
1,
(Y ), k = 1, . . . , N. Then the rst order
corrector is dened by
u

1
(x) =
k
_
x

_
u

x
k
(x),
1170 C. CONCA, R. ORIVE, AND M. VANNINATHAN
which means that
|u

1
|
H
1
(R
N
)
0 as 0.
In this paper, we obtain a more general result using a dierent approach intro-
duced in [11]. The basic tool of this new approach is Bloch waves associated with
A which we dene now. Let us consider the following spectral problem parameterized
by R
N
: nd = () R and = (y; ) (not identically zero) such that
_
A(; ) = ()(; ) in R
N
, (; ) is (; Y )-periodic, i.e.,
(y + 2m; ) = e
2im
(y; ) m Z
N
, y R
N
.
(1.15)
Next, we dene (y; ) = e
iy
(y; ), and (1.15) can be rewritten in terms of as
follows:
A() = in R
N
, is Y -periodic. (1.16)
Here, the operator A() is dened by
A()
def
=
_

y
k
+i
k
__
a
k
(y)
_

y

+i

__
, (1.17)
which can be rewritten as
A() = A+i
k
C
k
+
k

a
k
(y) (1.18)
with
C
k

def
= a
kj
(y)

y
j

y
j
(a
kj
(y)). (1.19)
It is clear from (1.15) that the (, Y ) periodicity condition is unaltered if we
replace by ( + q) with q Z
N
, and can therefore be conned to the dual
cell Y

= [
1
2
,
1
2
[
N
. It is well known (C. Conca, J. Planchard, and M. Van-
ninathan [10]) that for each Y

, the above spectral problem admits a discrete

sequence of eigenvalues with the following properties:
_
0
1
()
m
()
m 1,
m
() is a Lipschitz function of Y

.
Besides, the corresponding eigenfunctions denoted by
m
(; ) and
m
(; ) form or-
thonormal bases in the spaces of all L
2
loc
(R
N
)-functions which are (; Y )-periodic and
Y -periodic, respectively; these spaces are denoted by L
2
#
(; Y ) and L
2
#
(Y ). It is
worthwhile to remark that these eigenfunctions in fact belong to the spaces H
1
#
(; Y )
and H
1
#
(Y ), respectively, where
H
1
#
(; Y ) =
_
L
2
#
(; Y )

y
k
L
2
#
(; Y ) k = 1, . . . , N
_
,
H
1
#
(Y ) =
_
L
2
#
(Y )

y
k
L
2
#
(Y ) k = 1, . . . , N
_
.
BLOCH APPROXIMATION 1171
The functions
m
(; ) and
m
(; ) (referred to as Bloch waves) introduced above
enable us to describe the spectral resolution of A (an unbounded self-adjoint operator
in L
2
(R
N
)) in the orthogonal basis e
iy

m
(y; )[m 1, Y

. More precisely, we
have the following.
Theorem 1.2. Let g L
2
(R
N
). The mth Bloch coecient of g is dened as
follows:
(B
m
g)() =
_
R
N
g(y)e
iy

m
(y; )dy m 1, Y

.
Then the following inverse formula holds:
g(y) =
_
Y

m=1
(B
m
g)()e
iy

m
(y; )d.
Further, we have Parsevals identity:
_
R
N
[g(y)[
2
dy =
_
Y

m=1
[(B
m
g)()[
2
d.
Finally, for all g in the domain of A, we have
Ag(y) =
_
Y

m=1

m
()(B
m
g)()e
iy

m
(y; )d.
To obtain the spectral resolution of A

in an analogous manner, let us introduce

Bloch waves at the -scale:

m
() =
2

m
(),

m
(x; ) =
m
(y; ),

m
(x; ) =
m
(y; ),
where the variables (x, ) and (y, ) are related by y =
x

and = . Observe that

m
(x; ) is Y -periodic (in x) and
1
Y

-periodic with respect to . In the same

manner,

m
(; ) is (; Y )-periodic because of the relation

m
(x; ) = e
ix

m
(x; ).
Note that the dual cell at -scale is
1
Y

and hence we take to vary in

1
Y

in what
follows. With these notations, we have the following result analogous to Theorem 1.2.
Theorem 1.3. Let g L
2
(R
N
). The mth Bloch coecient of g at the -scale is
dened as follows:
(B

m
g)() =
_
R
N
g(x)e
ix

m
(x; )dx m 1,
1
Y

.
Then the following inverse formula and Parsevals identity hold:
g(x) =
_

1
Y

m=1
(B

m
g)()e
ix

m
(x; )d,
_
R
N
[g(x)[
2
dx =
_

1
Y

m=1
[(B

m
g)()[
2
d.
Finally, for all g in the domain of A

, we get
A

g(x) =
_

1
Y

m=1

m
()(B

m
g)()e
ix

m
(x; )d.
1172 C. CONCA, R. ORIVE, AND M. VANNINATHAN
Using the above theorem, the classical homogenization result was deduced in
[11]. Let us recall the main steps. The rst one consists of considering a sequence
u

H
1
(R
N
) satisfying (1.9). We can express the equation A

= f in R
N
in the
equivalent form

m
()(B

m
u

)() = (B

m
f)() m 1,
1
Y

. (1.20)
In the homogenization process, one can neglect all the relations for m 2. More
precisely, it is proved in [11] that the following result holds.
Proposition 1.4. Let
v

(x) =
_

1
Y

m=2
(B

m
u

)()e
ix

m
(x; )d. (1.21)
Then |v

|
L
2
(R
N
)
c.
Thus we can concentrate our attention only on the relation corresponding to the
rst Bloch wave:

1
()(B

1
u

)() = (B

1
f)()
1
Y

. (1.22)
The homogenized equation in the Fourier space
q
k

u() =

f() R
N
(1.23)
is obtained from (1.22) by passing to the limit as 0. Here, the symbol stands
for the classical Fourier transformation

f() =
1
(2)
N/2
_
R
N
f(x)e
ix
dx.
To this end, the following results were established and applied in [11].
Proposition 1.5. We assume that a
k
satises (1.2). Then there exists > 0
such that the rst eigenvalue
1
() is an analytic function on B

def
= [ [[ < , and
there is a choice of the rst eigenvector
1
(y; ) satisfying
_

1
(; ) H
1
#
(Y ) is analytic on B

1
(y; 0) = p
(0)
(= [Y [
1/2
=
1
(2)
N/2
).
Moreover, we have the relations

1
(0) = 0, D
k

1
(0) =

1

k
(0) = 0 k = 1, . . . , N,
1
2
D
2
k

1
(0) =
1
2

(0) = q
k
k, = 1, . . . , N,
and there exist constants c and c such that
c[[
2

1
() c[[
2
Y

, (1.24)
0 <
(N)
2

m
() m 2, Y

, (1.25)
where
(N)
2
is the second eigenvalue of the spectral problem for A in the cell Y with
Neumann boundary conditions on Y .
BLOCH APPROXIMATION 1173
Apart from the above result of regularity on the Bloch spectrum, we need to
prove that the rst Bloch transform is an approximation to the Fourier transform.
This result is naturally expected from the fact that

1
(x; ) (2)
N/2
, as 0,
R
N
.
Proposition 1.6. Let g

and g be in L
2
(R
N
). Then
(i) if g

g weakly in L
2
(R
N
x
), then

1
Y

1
g

g weakly in L
2
loc
(R
N

) pro-
vided there is a xed compact set K such that supp(g

) K ;
(ii) if g

g in L
2
(R
N
x
), then

1
Y

1
g

g in L
2
loc
(R
N

).
These results easily lead us to the following homogenization theorem in R
N
.
Theorem 1.7. We consider a sequence u

satisfying (1.9). Then

a

k
u

q
k
u

in L
2
(R
N
) k = 1, . . . , N.
In particular, u

satises A

= f in R
N
.
Once the homogenization result in R
N
is established, it is an easy matter to
deduce the corresponding result in a bounded domain by localization techniques
using a cut-o function T() (see [11]).
1.2. Presentation of new results: The Bloch approximation. Let us con-
sider the sequence u

satisfying hypotheses (1.9). The Bloch approximation of u

is
dened by the following formula:

(x)
def
=
_

1
Y

()e
ix

1
(x; )d, x R
N
. (1.26)
First of all, let us remark that this object is not dicult to be computed in principle.
Our goal throughout this paper is to study properties of this function and particularly
its relations with various correctors terms. It is worth noticing that

is dened only
in terms of the rst Bloch mode

1
. We will see in section 3 that higher Bloch modes

m
, m 2, do not contribute at all in the analysis of the correctors of rst and
second order in the energy norm. (It will be interesting to know whether these higher
order modes play a part in the analysis of correctors in stronger norms H
2
, . . . , etc.
For H
2
-estimates, we refer to our work in [12].) Thus we are motivated to introduce
the projection onto the rst Bloch mode: for all g L
2
(R
N
), we dene
P

1
g(x) =
_

1
Y

1
g()e
ix

1
(x; )d, x R
N
. (1.27)
We note by the item (ii) of Proposition 1.6 that the Fourier transform u is an
approximation of B

1
u

. Therefore, heuristically speaking, the Bloch approximation

is close to P

1
u

and hence to u

. With these notations, we will prove the following

theorem.
Theorem 1.8. Assume that the coecients a
k
satisfy (1.2). Let u

be the
sequence introduced in (1.9). Then if f L
2
(R
N
), we have
(u

) 0 in H
1
(R
N
). (1.28)
Furthermore, we have the estimate
[u

[
H
1
(R
N
)
c|f|
L
2
(R
N
)
. (1.29)
1174 C. CONCA, R. ORIVE, AND M. VANNINATHAN
It is worth remarking that even though error estimates of the type (1.29) are
sometimes found in the literature, they are usually obtained using the maximum
principle with more regularity hypotheses on a
k
and f. Here, we obtain these natural
estimates under optimal hypotheses.
Thanks to the above result, we are reduced to expanding

in terms of in order
to be able to compare it with the classical correctors for u

. To fulll this task, it is

clear from the denition of

that it is necessary to obtain asymptotic expansions

of the rst eigenvalue

1
() and the rst Bloch mode

1
(; ). (In addition, for our
purposes below, we need an asymptotic expansion of the rst Bloch transform B

1
g()
for which we refer the reader to section 5. These results strengthen earlier results,
particularly those of Proposition 1.6.) We now state results in this direction, and their
proofs will be taken up in the following sections along with other auxiliary results.
First, we introduce some test functions
k
,
km
,
kmn
dened by the following cell
problems (observe that the rst ones are nothing but the functions already introduced
in (1.14)):
_

_
A
k
= (a
k
q
k
)
1
2
_
C
k

+C

k
_
in R
N
,

k
H
1
#
(Y ), M
Y
(
k
) = 0.
(1.30)
_

_
A
km
=
1
3
_
(a
k
q
k
)
m
+ (a
m
q
m
)
k
+ (a
mk
q
mk
)

C
k

m
C

mk
C
m

k
_
in R
N
,

km
H
1
#
(Y ), M
Y
(
km
) = 0.
(1.31)
_

_
A
kmn
=
1
4!
D
4
kmn

1
(0)
1
4
_
C
n

km
+C
k

mn
+C

mnk
+C
m

nk
_
+
1
3!
_
(a
k
q
k
)
mn
+ (a
m
q
m
)
kn
+ (a
km
q
km
)
n
+ (a
mn
q
mn
)
k
+ (a
n
q
n
)
km
+ (a
kn
q
kn
)
m
_
in R
N
,

kmn
H
1
#
(Y ), M
Y
(
kmn
) = 0.
(1.32)
Proposition 1.9. All odd order derivatives of
1
at = 0 vanish, i.e.,
D

1
(0) = 0 Z
N
+
, [[ odd.
All even order derivatives of
1
at = 0 can be calculated in a systematic fash-
ion. For instance, the fourth order derivatives have the following expressions: for all
k, , m, n = 1, . . . , N
1
4!
D
4
kmn

1
(0) =
1
4
1
[Y [
_
Y
_
C
n

km
+C
k

mn
+C

mnk
+C
m

nk
_
dy
BLOCH APPROXIMATION 1175

1
3!
1
[Y [
_
Y
_
(a
k
q
k
)
mn
+ (a
m
q
m
)
nk
+ (a
mn
q
mn
)
k
+ (a
nk
q
nk
)
m
+ (a
km
q
km
)
n
+ (a
n
q
n
)
km
_
dy.
Various derivatives of
1
at = 0 can also be calculated in a systematic fashion.
Proposition 1.10. We have the following expressions:
D
k

1
(y; 0) = ip
(0)

k
(y),
1
2!
D
2
k

1
(y; 0) = p
(0)

k
(y)
(2)
k
p
(0)
,
1
3!
D
3
km

1
(y; 0) = ip
(0)

km
(y)
i
3
_

(2)
k

m
(y) +
(2)
m

k
(y) +
(2)
mk

(y)
_
p
(0)
,
1
4!
D
4
kmn

1
(y; 0) = p
(0)

kmn
(y)
1
3!
_

(2)
k

mn
(y) +
(2)
m

nk
(y) +
(2)
mn

k
+
(2)
nk

m
(y) +
(2)
km

n
(y) +
(2)
n

km
(y)
_
p
(0)
+
(4)
kmn
p
(0)
with

(2)
k
=
1
2!
1
[Y [
_
Y

k
dy,

(4)
kmn
=
1
[Y [
_
Y
1
4
_

mn

k
+
kmn

+
nk

m
+
kn

n
_
dy

1
[Y [
_
Y
1
6
_

kn
+
km

n
+
k

nm
_
dy
+
1
[Y [
1
2
_

(2)
k

(2)
mn
+
(2)
km

(2)
n
+
(2)
kn

(2)
m
_
.
We note that all odd order derivatives of
1
at = 0 are purely imaginary and
all even order derivatives are real.
Since
1
(; ) is proved to be analytic for [[ , we can expand it and thus give
rise to an asymptotic expansion of

which is as follows:

(x) = u

(x) +
k
_
x

_
u

x
k
(x)
2
_

k
_
x

_
+
(2)
k
_

2
u

x
k
x

(x) + . (1.33)
This can be rigorously proved. Our next result is a sample where we specify the
precise hypotheses needed to justify the above expansion up to three terms.
Theorem 1.11. Assume that the hypotheses of Theorem 1.8 hold.
(i) If u

H
1
(R
N
), then
|

|
L
2
(R
N
)
c|u

|
H
1
(R
N
)
.
(ii) If f L
2
(R
N
) and
k
W
1,
#
(Y ), where
k
is the solution of (1.13) and

k
(x) =
k
_
x

_
, then we have
_
_
_
_

k
u

x
k
_
_
_
_
H
1
(R
N
)
c|f|
L
2
(R
N
)
.
1176 C. CONCA, R. ORIVE, AND M. VANNINATHAN
(iii) If f H
1
(R
N
) and
k
,
k
W
1,
#
(Y ), where
k
is the solution of (1.30),

(2)
k
are constants dened in Proposition 1.10, and

k
(x) =
k
_
x

_
, then
_
_
_
_

k
u

x
k
+
2
_

k
+
(2)
k
_

2
u

x
k
x

_
_
_
_
H
1
(R
N
)
c
2
|f|
H
1
(R
N
)
.
It is important to note that these above expansions are of Taylor type owing to
the analyticity of
1
() and
1
(; ). This is the main dierence between this approach
and the classical one found in [5], where the expansion has a multiscale structure.
Concerning the hypotheses on the smoothness of functions
k
and
k
in state-
ments (ii) and (iii), it is worth mentioning from regularity theory of elliptic boundary
value problems that W
1,
-estimates are hard to come by. This is why they are usually
assumed in homogenization theory. However, several numerical studies with simple
bers show that these assumptions are valid. Thus they are reasonable hypotheses to
work with as far as certain applications are concerned.
The expansions of
1
(),

1
(; ), and B

1
g() obtained in Propositions 1.5 and 1.9
and Propositions 5.1, 5.2, and 5.3 below have further interesting consequences which
will be developed in a forthcoming paper. For the time being, we will be content
with a few remarks. Since higher order modes can be neglected, the rst eigenvalue

1
() along with the rst eigenvector
1
(; ) represent the periodic medium under
consideration. Their contributions occur somewhat separately without interaction at
the levels of homogenized equation and correctors. More precisely, the rst eigenvalue

1
() contributes at various levels through its derivatives at = 0. The rst eigen-
vector
1
(; ) and its rst derivatives contribute through the rst Bloch transform
B

1
g() and its expansion described in Propositions 5.2 and 5.3.
In the homogenized equation, for instance, we see the product of the second order
derivatives of
1
() at = 0 with the 0th order term of B

1
g(), namely, g(). We
see a similar structure in the correctors, too. There seem to be situations where both
interact and contribute jointly in a manner dierent from the above. One example
of such a situation is the study of the propagation of waves in a periodic medium.
It appears that the homogenized medium is not good enough to provide an approx-
imation to the propagation for large times because of the appearance of dispersion
eects shown numerically in F. Santosa and W. W. Symes [20]. We feel that this is an
appropriate place to highlight the improvements achieved in this work with respect
to [20]. Apart from the mathematical rigor, the main point is that the third order
derivatives of
1
() at = 0 are shown to be zero even in the multidimensional case.
(In fact all odd order derivatives vanish.) Moreover, our arguments are more general
compared with the one-dimensional case covered in [20]. This will have consequences
in the propagation of waves in periodic media. We plan to cover these aspects in a
future publication.
We conclude this introduction by saying how the rest of this paper is organized.
Section 2 is devoted to certain fundamental lemmas which are indispensable. As an
immediate application, we prove in section 3 that the higher order Bloch modes are
negligible. Taylor expansions for
1
and
1
are obtained in section 4 which proves
Propositions 1.9 and 1.10. Section 5 is devoted to the description of the asymptotic
behavior of the rst Bloch transform B

1
whose denition is given in Theorem 1.2.
Finally, in section 6, we present the proofs of the main results, namely, Theorems 1.8
and 1.11.
BLOCH APPROXIMATION 1177
2. Fundamental lemmas. In this section, we prove a series of results which
generalize Parsevals identity stated in Theorem 1.3. These estimates will be useful
later for the analysis of correctors. The following two lemmas are easily seen to be
generalizations of well-known classical results for .
Lemma 2.1. For all g H
1
(R
N
), we have
c
1
[g[
2
H
1
(R
N
)

1
Y

m=1

m
()[B

m
g()[
2
d c
2
[g[
2
H
1
(R
N
)
,
where c
1
and c
2
are constants independent of and g.
Proof. First of all, by uniform ellipticity of A

, we have

_
R
N
[g[
2
dx
_
R
N
A

g gdx
_
R
N
[g[
2
dx.
We can rewrite the middle term by applying the Plancherel identity:
_
R
N
g(x)h(x)dx =
_

1
Y

m=1
B

m
g()B

m
h()d g, h L
2
(R
N
). (2.1)
Indeed, using the spectral resolution of A

, we get
_
R
N
A

g gdx =
_

1
Y

m=1

m
()[B

m
g()[
2
d.
This completes the proof.
By using the duality between H
1
(R
N
) and H
1
(R
N
), we deduce the following.
Lemma 2.2. For all g H
1
(R
N
), there exist c
1
and c
2
which are independent
of and g, such that
c
1
|g|
2
H
1
(R
N
)

1
Y

m=1
1
1 +

m
()
[B

m
g()[
2
d c
2
|g|
2
H
1
(R
N
)
.
Proof. It is well known that (A

+ I): H
1
(R
N
) H
1
(R
N
) is an isomorphism.
For every g H
1
(R
N
) there exists a unique solution u H
1
(R
N
) of A

u +u = g in
R
N
. We can express the previous equation in the equivalent form
(

m
() + 1)B

m
u() = B

m
g() m 1,
1
Y

.
Therefore, an application of the CauchySchwarz inequality yields
g, v) =
_

1
Y

m=1
(

m
() + 1)B

m
uB

m
vd

_
_

1
Y

m=1
(

m
() + 1)[B

m
u[
2
d
_
1/2
_
_

1
Y

m=1
(

m
() + 1)[B

m
v[
2
d
_
1/2

_
_

1
Y

m=1
[B

m
g[
2
(

m
() + 1)
d
_
1/2
_
_

1
Y

m=1
(

m
() + 1)[B

m
v[
2
d
_
1/2
1178 C. CONCA, R. ORIVE, AND M. VANNINATHAN
for all v H
1
(R
N
), g H
1
(R
N
). Here, , ) denotes the H
1
(R
N
) and H
1
(R
N
)
duality pairing. By virtue of Lemma 2.1 and Parsevals identity, the second term in
the right-hand side is equivalent to the H
1
-norm of v. Thus we deduce the existence
of a constant c
1
such that
c
1
|g|
2
H
1
(R
N
)

1
Y

m=1
1
1 +

m
()
[B

m
g()[
2
d,
which is the lower estimate in Lemma 2.2. To prove the upper estimate is enough
to use the continuity of the solution u H
1
(R
N
) with respect to the right-hand side
g H
1
(R
N
).
In our next lemma, we consider g

= g

() a measurable function dened on

1
Y

, and another function = (y; ) measurable with respect to (y; ) and Y -

periodic in y. We then introduce
G

(x) =
_

1
Y

()e
ix

_
x

;
_
d, x R
N
. (2.2)
The following result estimates its L
2
(R
N
) and H
1
(R
N
) norms.
Lemma 2.3. We assume g

L
2
(
1
Y

) and L

(Y

; H
1
#
(Y )). Then we
have
|G

|
2
L
2
(R
N
)
=
_

1
Y

[g

()[
2
|(; )|
2
L
2
(Y )
d,
[G

[
2
H
1
(R
N
)
=
_

1
Y

[g

()[
2
|i(; ) +
1

y
(; )|
2
L
2
(Y )
N
d.
Proof. We expand (y; ) as a function of y in the orthonormal basis
m
(y; )

m=1
where is a parameter:
(y; ) =

m=1
a
m
()
m
(y; ).
Introducing this expression in (2.2), we get
G

(x) =
_

1
Y

()

m=1
a
m
()e
ix

m
(x; )d.
Applying Parsevals identity of Theorem 1.3, we get
|G

|
2
L
2
(R
N
)
=
_

1
Y

[g

()[
2

m=1
[a
m
()[
2
d.
This completes the proof of the rst part of the lemma if we use Parsevals identity
in L
2
(Y ):
|(; )|
2
L
2
(Y )
=

m=1
[a
m
()[
2
Y

. (2.3)
For the second part of the lemma, we formally dierentiate G

(x) with respect

to x. We obtain

x
G

(x) =
_

1
Y

()e
ix
_
i
_
x

;
_
+
1

_
x

;
__
d.
BLOCH APPROXIMATION 1179
We remark that the above integral is of the same type as the one analyzed in the rst
part. This completes the proof.
The next lemma presents H
1
-estimates on the Bloch modes.
Lemma 2.4. We suppose that the coecients a
k
satisfy (1.2). Then there exists
a constant c depending on |a
k
|
L

(Y )
such that
_
_
_

m
y
k
(; )
_
_
_
L
2
(Y )
c
1

m
()
1/2
Y

, m 1, k = 1, . . . , N. (2.4)
To prove this, let us introduce the bilinear forms associated with the operators
A() and A, respectively.
a(; , ) =
_
Y
a
k
(y)
_

+i

__

y
k
+i
k

_
dy,
a(, ) =
_
Y
a
k
(y)

y

y
k
dy.
The basic estimates on them are obtained in [10, p. 190]: There exist constants c, c
which are independent of Y

such that for all H

1
#
(Y ),
c (||
2
L
2
(Y )
N
+[[
2
||
2
L
2
(Y )
) a(; , ) c (||
2
L
2
(Y )
N
+[[
2
||
2
L
2
(Y )
), (2.5)
c ||
2
L
2
(Y )
N
a(, ) c ||
2
L
2
(Y )
N
. (2.6)
Proof of Lemma 2.4. For simplicity, we denote
m
(; ) by
m
(). We recall that
it satises
a(;
m
(), ) =
m
()(
m
(), ) H
1
#
(Y ). (2.7)
To deduce (2.4), it is enough to take =
m
() and use (2.5).
Our next result concerns the estimation of expressions which are inverse to (2.2).
We dene
J

g() =
_
R
N
g(x)e
ix

_
x

;
_
dx for
1
Y

, (2.8)
where g = g(x) is a measurable function dened on R
N
and = (y; ) is a measurable
function dened on Y Y

. We assume that is Y -periodic in y. The required

hypotheses on these functions will depend on the estimate deduced on J

g. This is
illustrated in the results that follow which are analogous to classical estimates on the
Fourier transform.
Lemma 2.5.
(i) If g L
2
(R
N
) and L

(Y

; L
2
#
(Y )), then we have
|J

g|
L
2
(
1
Y

)
|g|
L
2
(R
N
)
||
L

(Y

;L
2
#
(Y ))
.
(ii) If g H
1
(R
N
) and L

(Y

; H
1
#
(Y )), then we have
|(1 +[[
2
)
1/2
J

g()|
L
2
(
1
Y

)
c
_
|g|
L
2
(R
N
)
||
L

(Y

;L
2
(Y ))
+
1
|g|
L
2
(R
N
)
|
y
|
L

(Y

;L
2
(Y )
N
)
_
.
1180 C. CONCA, R. ORIVE, AND M. VANNINATHAN
Proof. The idea is to consider the product space L
2
(Y

; L
2
#
(Y )) and expand
(y; ) in two steps. First using the fact that

m
(; )

m=1
is an orthonormal basis
in L
2
#
(Y ), we get
(y; ) =

m=1
a
m
()

m
(y; ) y Y, Y

.
Next, for each m, we can expand a
m
() in the usual Fourier series:
a
m
() =

nZ
N
a
mn
e
2in
Y

.
The corresponding Parsevals identities are as follows:
|(; )|
2
L
2
(Y )
=

m
[a
m
()[
2
Y

,
_
Y

[a
m
()[
2
d =

nZ
N
[a
mn
[
2
m N.
Using this expansion, we can rewrite J

g as follows:
J

g() =

m=1

nZ
N
a
mn
e
2in
_
R
N
g(x)e
ix

m
_
x

;
_
dx,
which, according to the denition of B

m
g(), is equal to
J

g() =

m=1

nZ
N
a
mn
e
2in
B

m
g() =

m=1
a
m
()B

m
g().
By the CauchySchwarz inequality,
[J

g()[
2

m=1
[a
m
()[
2
__

m=1
[B

m
g()[
2
_
= |(; )|
2
L
2
(Y )
_

m=1
[B

m
g()[
2
_
||
2
L

(Y

;L
2
#
(Y ))
_

m=1
[B

m
g()[
2
_
.
The proof of (i) is complete if we integrate the above inequality with respect to

1
Y

and apply Theorem 1.3. For the proof of (ii), we multiply (2.8) by (i
k
)
and obtain
(i
k
)J

g() =
_
R
N
g(x)(i
k
)e
ix

_
x

;
_
dx,
which, by integration by parts, can be rewritten as
(i
k
)J

g() =
_
R
N
g
x
k
(x)e
ix

_
x

;
_
dx
1
_
R
N
g(x)e
ix

y
k
_
x

;
_
dx.
BLOCH APPROXIMATION 1181
It is now sucient to apply (i) to each of the terms on the right-hand side of the
above relation.
Next, we will need some properties of the classical discrete Fourier transform
in our asymptotic description of the rst Bloch transform. In particular, we are
interested in the relation between discrete and continuous Fourier transforms. To this
end, let us begin by introducing some necessary notations. Let Y

Z
N be the mesh
of R
N
generated by the cell Y . More precisely, Y

= x

+ Y where x

= 2 is
the origin of the cell Y

. We recall the denition of the discrete Fourier transform of

a function corresponding to this mesh: Let p > N be given. For g W
1,p
(R
N
) with
compact support we dene
F

g() =

Z
N
g(x

)e
ix

1
Y

. (2.9)
It is worthwhile to recall that W
1,p
(R
N
) is embedded in C
0
(R
N
) when p > N, and so
g(x

) is well dened.
Lemma 2.6. For g W
1,p
(R
N
) (p > N) with compact support K, we have
(i)
N
(

1
Y

g)()
1
(2)
N/2
g() for R
N
.
(ii) |
N
F

g|
L
2
(
1
Y

)
c[K[
p2
2p
|g|
L
p
(R
N
)
+ |g|
L
p
(R
N
)
N
, [K[ denotes the
measure of K.
(iii)
N

1
Y

g
1
(2)
N/2
g in L
2
(R
N
).
Proof. To prove (i), we multiply (2.9) by
N
to get

N
F

g() =
1
(2)
N

Z
N
g(x

)e
ix

[Y

[.
We regard the right-hand side of the above equality as a Riemann sum of the integral
1
(2)
N
_
R
N
g(x)e
ix
dx
which converges to it as 0.
To prove (ii), we observe that the right-hand side of (2.9) is nothing but the
Fourier series in the variable
1
Y

. Therefore, by Parsevals identity, we get

N
_

1
Y

[F

g()[
2
d =

Z
N
[g(x

)[
2
.
We multiply this relation by
N
and rewrite it as

2N
_

1
Y

[F

g()[
2
d =
1
(2)
N

Z
N
[g(x

)[
2
[Y

[. (2.10)
To estimate the right-hand side of the above equality, we integrate the inequality
[g(x

)[
2
2
_
[g(x)[
2
+[g(x) g(x

)[
2
_
, x Y

,
over Y

to obtain
[g(x

)[
2
[Y

[ 2
_
_
Y

[g(x)[
2
dx +
_
Y

[g(x) g(x

)[
2
dx
_
. (2.11)
1182 C. CONCA, R. ORIVE, AND M. VANNINATHAN
Since p > N, we can use the classical Morreys inequality (see Brezis [7, p. 167]) to
deduce
[g(x) g(x

)[ c
1
N
p
[g[
L
p
(Y

)
N
.
Using both the above estimate in (2.11) and the H older inequality and summing over
Z
N
, we complete the proof of (ii).
To prove the statement (iii) we rst use (i) and (ii) to deduce that

1
Y

()F

g()
1
(2)
N/2
g() in L
2
(R
N
)-weak.
Let us now expand
_
_
_
_

1
Y

g
1
(2)
N/2
g
_
_
_
_
2
L
2
(R
N
)
=
2N
|F

g|
2
L
2
(R
N
)

2
N
(2)
N/2
(

1
Y

g, g)
+
1
(2)
N
| g|
2
L
2
(R
N
)
.
Now relation (2.10) shows that

2N
|F

g|
2
L
2
(R
N
)

1
(2)
N
_
R
N
[g[
2
dx =
1
(2)
N
| g|
2
L
2
(R
N
)
.
Thanks to the above weak convergence, the second term converges to

2
(2)
N/2
1
(2)
N/2
| g|
2
L
2
(R
N
)
.
This simple computation establishes the strong convergence in L
2
(R
N
).
3. Higher Bloch modes are negligible. In this section, we consider a se-
quence of solutions u

of the equation with f H

1
(R
N
):
A

= f in R
N
, u

H
1
(R
N
). (3.1)
Let us recall that the above equation is equivalent to (1.20) in the Bloch space. In
what follows, we present a systematic method of obtaining estimates on the solution
in Sobolev spaces L
2
and H
1
. In particular, we show that the component of u

in the
higher Bloch modes does not play any role in the analysis of correctors of rst and
second order provided f is suciently smooth. Thus we consider v

dened in (1.21),
which is nothing but the projection of u

corresponding to all higher Bloch modes.

Estimates on v

derived in this section improve Proposition 1.4.

Proposition 3.1. We have the following estimates for f L
2
(R
N
):
(i) [v

[
H
1
(R
N
)
c|f|
L
2
(R
N
)
,
(ii) |v

|
L
2
(R
N
)
c|f|
H
1
(R
N
)
.
Proof. To show (i), we apply Lemma 2.1 with g = v

and use (1.20). We obtain

|v

|
2
L
2
(R
N
)
N
c
_

1
Y

m=2
1

m
()
[B

m
f()[
2
d
c sup
m2,
1
Y

m
()
|f|
2
L
2
(R
N
)
.
BLOCH APPROXIMATION 1183
Proof of (i) is complete since we have (cf. (1.25))
sup
m2,
1
Y

m
()

1

(N)
2

2
. (3.2)
For the proof of (ii), we apply Lemma 2.2 with g = f and (1.20). We have
|v

|
2
L
2
(R
N
)
=
_

1
Y

m=2
[B

m
u

()[
2
d
=
_

1
Y

m=2
1

m
()
2
[B

m
f()[
2
d.
Writing
1

m
()
2
[B

m
f()[
2
=
1 +

m
()

m
()
2
[B

m
f()[
2
1 +

m
()
and using (3.2), we deduce that
1

m
()
2
[B

m
f()[
2
c
2
[B

m
f()[
2
1 +

m
()
.
The proof is complete if we use Lemma 2.2.
While the above proposition shows that v

can be neglected at the level of the

rst order correctors (cf. (1.10)), the next result will demonstrate that v

can be
neglected at the level of correctors of rst and second order. These ner estimates
require naturally higher order regularity of f but not of the coecients a
k
(y). Let us
state the following proposition, whose proof is similar to the previous one and hence
will not be repeated.
Proposition 3.2. We have the following estimates for f H
1
(R
N
):
(i) [v

[
H
1
(R
N
)
c
2
|f|
H
1
(R
N
)
,
(ii) |v

|
L
2
(R
N
)
c
2
|f|
L
2
(R
N
)
.
Assuming a
k
are in W
1,
#
(Y ) and further assumptions, we can obtain H
2
-
estimates on the solution. This is dicult as it involves more subtleties (see [12]).
4. Taylor expansion of the rst Bloch eigenvalue and eigenvector. The
purpose of this section is to indicate a systematic method to compute derivatives
of the rst Bloch eigenvalue
1
() and the rst Bloch eigenvector
1
(; ) at = 0.
In particular, we will prove Propositions 1.9 and 1.10. Recall that
1
() and
1
(; )
depend analytically on in a small neighborhood B

of = 0. At the cost of reducing

this neighborhood, we claim that the branch
1
(; ) can be chosen so that the
following conditions are satised simultaneously:
B

1
(; ) H
1
#
(Y ) is analytic, (4.1)
|
1
(; )|
L
2
(Y )
= 1 B

, (4.2)
Im
_
Y

1
(y; )dy = 0 B

. (4.3)
1184 C. CONCA, R. ORIVE, AND M. VANNINATHAN
In what follows, we will see that the above conditions uniquely x the eigenvector

1
(; ). We remark that the condition (4.2) is classical, whereas the condition (4.3)
is somewhat unusual and can be achieved as indicated below. The idea consists of
multiplying
1
(; ) by a complex number (
1
() +i
2
()) where
1
() and
2
() are
real analytic with respect to and are chosen such that
Im
_
Y
(
1
() +i
2
())
1
(y; )dy = 0.
If we dene
d() = (d
1
(), d
2
())
def
=
_
Im
_
Y

1
(y; )dy, Re
_
Y

1
(y; )dy
_
,
then the above condition is equivalent to

1
()d
1
() +
2
()d
2
() = 0 B

.
Obviously, one such choice which is analytic is as follows:

1
() = d
2
(),
2
() = d
1
().
Of course, the above procedure has destroyed condition (4.2) (but not condition (4.1)).
However, it can be regained by dividing by [d()[. This is possible because d(0) ,= 0
by our choice of
1
(; 0) (see Proposition 1.5).
Thanks to our choice of the branch satisfying (4.1)(4.3), we will now draw some
consequences which will simplify the computations below. In fact, dierentiating (4.2)
with respect to , we successively get for all k, , m, n = 1, . . . , N
ReD
k

1
(; ),
1
(; )) = 0, (4.4)
ReD
2
k

1
(; ),
1
(; )) + ReD
k

1
(; ), D

1
(; )) = 0, (4.5)
_
ReD
3
km

1
(; ),
1
(; )) + ReD
2
k

1
(; ), D
m

1
(; ))
+ ReD
2
km

1
(; ), D

1
(; )) + ReD
k

1
(; ), D
2
m

1
(; )) = 0,
(4.6)
_

_
ReD
4
kmn

1
(; ),
1
(; )) + ReD
3
km

1
(; ), D
n

1
(; ))
+ ReD
3
kn

1
(; ), D
m

1
(; )) + ReD
2
k

1
(; ), D
2
mn

1
(; ))
+ ReD
3
kmn

1
(; ), D

1
(; )) + ReD
2
km

1
(; ), D
2
n

1
(; ))
+ ReD
2
kn

1
(; ), D
2
m

1
(; )) + ReD
k

1
(; ), D
3
mn

1
(; )) = 0,
(4.7)
where ; ) denotes the scalar product in L
2
#
(Y ). On the other hand, dierentiation
of (4.3) yields
Im
_
Y
D

1
(y; )dy = 0 Z
N
+
. (4.8)
From these sets of relations, it follows that
_
Y
D

1
(y; 0)dy = 0 Z
N
+
with [[ odd. (4.9)
BLOCH APPROXIMATION 1185
4.1. First order derivatives. If we dierentiate the eigenvalue equation
(A()
1
())
1
(; ) = 0 once with respect to
k
, we obtain
D
k
(A()
1
())
1
(; ) + (A()
1
())D
k

1
(; ) = 0. (4.10)
Taking the scalar product with
1
(; ), we get
[D
k
(A
1
)]
1
,
1
) = 0, (4.11)
where we have suppressed the dependence on for ease of writing. We will continue
with this convention in what follows provided there is no ambiguity. It follows from
(1.18) that
D
k
A(0) = iC
k
Y

, (4.12)
where the operator C
k
is dened in (1.19).
If we evaluate the relation (4.11) at = 0 and use the structure of C
k
, we
immediately get that
D
k

1
(0) = 0 k = 1, . . . , N. (4.13)
The next step is to compute the rst order derivatives of
1
at = 0. To this end,
we go back to (4.10) and use (4.13). We obtain
AD
k

1
(; 0) = D
k
A(0)
1
(; 0) = iC
k

1
(; 0).
Taking into account (4.9) and the above equation, we can solve for D
k

1
(y; 0) and
obtain
D
k

1
(y; 0) = i
1
(y; 0)
k
(y) = ip
(0)

k
(y), (4.14)
where, we recall,
k
satises (1.13) and the constant p
(0)
was xed in Proposition 1.5.
Thus, the rst order derivative is completely determined and
D
k

1
(y; 0) is purely imaginary. (4.15)
4.2. Second order derivatives. Our starting point is the relation (4.10), which
we dierentiate once with respect to . We obtain
[D
2
k
(A
1
)]
1
+ [D
k
(A
1
)]D

1
+ [D

(A
1
)]D
k

1
+ (A
1
)D
2
k

1
= 0.
(4.16)
Taking the scalar product with
1
, we get
[D
2
k
(A
1
)]
1
,
1
) +[D
k
(A
1
)]D

1
,
1
) +[D

(A
1
)]D
k

1
,
1
) = 0 (4.17)
for all B

. If we use the information obtained in section 4.1 on D

k

1
(0), D
k

1
(; 0),
D
k
A(0), and
D
2
k
A() = 2a
k
(y) k, = 1, . . . , N, Y

, (4.18)
1186 C. CONCA, R. ORIVE, AND M. VANNINATHAN
we obtain
1
2!
D
2
k

1
(0) =
1
[Y [
_
Y
a
k
(y)dy
1
2[Y [
_
Y
(C
k

(y) +C

k
(y))dy
=
1
2
(q
k
+q
k
) = q
k
k, = 1, . . . , N. (4.19)
As before, the next step is to compute D
2
k

1
(; 0). For this purpose, we go back to
(4.16) and rewrite it with = 0 as follows:
AD
2
k

1
(; 0) =
_
2(a
k
q
k
) +C
k

+C

k
_

1
(; 0).
By comparing the above equation with (1.30) and using the simplicity of the eigenvalue
under consideration, we see that D
2
k

1
(; 0) is of the form
1
2!
D
2
k

1
(y; 0) = p
(0)

k
(y)
(2)
k
p
(0)
for some constant
(2)
k
. Thanks to (4.5) and (4.8), we can infer that

(2)
k
and D
2
k

1
(; 0) are real. (4.20)
Moreover,
(2)
k
admits the expression given in Proposition 1.10.
4.3. Third order derivatives. From the calculations done so far, it is now
clear how to proceed further to calculate higher order derivatives. Therefore we will
be brief here. Dierentiating (4.16), we get
_

_
[D
3
km
(A
1
)]
1
+ [D
2
k
(A
1
)]D
m

1
+ [D
2
m
(A
1
)]D
k

1
+ [D
2
km
(A
1
)]D

1
+ [D
k
(A
1
)]D
2
m

1
+ [D

(A
1
)]D
2
km

1
+ [D
m
(A
1
)]D
2
k

1
+ (A
1
)D
3
km

1
= 0.
(4.21)
Taking the scalar product with
1
, we get
_

_
[D
3
km
(A
1
)]
1
,
1
) +[D
2
k
(A
1
)]D
m

1
,
1
) +[D
2
m
(A
1
)]D
k

1
,
1
)
+ [D
2
km
(A
1
)]D

1
,
1
) +[D
k
(A
1
)]D
2
m

1
,
1
)
+ [D

(A
1
)]D
2
km

1
,
1
) +[D
m
(A
1
)]D
2
k

1
,
1
) = 0.
(4.22)
To conclude that D
3
km

1
(0) = 0, it is enough to use the following information in the
above relation:
_
D
k
A is purely imaginary, D
2
k
A is real, D
3
km
A = 0,

1
(0), D
2
k

1
(0) are real, D
k

1
(0) is purely imaginary.
(4.23)
It is evident that the above argument is very general and so can be used to establish
that all odd order derivatives of
1
at = 0 vanish. This proves the rst part of
Proposition 1.9.
BLOCH APPROXIMATION 1187
To nd the third order derivatives of
1
at = 0, we realize that (4.21) denes
a periodic problem for D
3
km

1
(; 0) which can be compared with (1.31). Further, the
relation (4.9) says that its average vanishes. These observations are enough to get the
expression of D
3
km

1
(; 0) given in Proposition 1.10. We conclude by observing the
following important property:
D
3
km

1
(y; 0) is purely imaginary. (4.24)
4.4. Fourth order derivatives. To arrive at the expressions for the fourth
order derivatives of
1
and
1
at = 0 given in Propositions 1.9 and 1.10, we follow
the same arguments as in section 4.3.
5. Convergence of the rst Bloch transform to the Fourier transform.
This section is devoted to the proof of the next proposition which shows the sense in
which the Fourier transform is approximated by the rst Bloch transform.
Proposition 5.1.
(i) For every g L
2
(R
N
) with compact support, we have

1
Y

()B

1
g() g() in L

loc
(R
N

).
(ii) If g L
2
(R
N
), we have

1
Y

()B

1
g() g() in L
2
(R
N

).
This will be a consequence of a more general result. In order to state it, we need
to introduce some new notations. We associate with every function = (y; ) dened
on Y Y

which is Y -periodic in y the following function:

(0)
() =
1
[Y [
_
Y
(y; )e
iy
dy, Y

. (5.1)
With this notation, we have the following proposition.
Proposition 5.2. We suppose L

(Y

; L
2
#
(Y )). Then for all g W
1,p
(R
N
)
with compact support K and with p > N, we have

1
Y

()
_
J

g() (2)
N/2

(0)
() g()
_
0 in L
2
(R
N

), (5.2)
where, we recall, J

g was dened in (2.8).

The proof will be taken up later. Admitting it for the moment, we turn our
attention to the following proof.
Proof of Proposition 5.1. If g L
2
(R
N
) with compact support K, we have for all
R
N
[

1
Y

()B

1
g() g()[ [

1
Y

()(B

1
g() g())[ +[(

1
Y

() 1) g()[
c[K[|g|
L
2
(R
N
)
|
1
(; )
1
(; 0)|
L
2
(Y )
+[(

1
Y

() 1) g()[.
If [[ is bounded, then by using the fact that the map
1
(; ) L
2
#
(Y ) is
Lipschitz near = 0, we deduce
[[
1
(; )
1
(; 0)|
L
2
(Y )
c.
1188 C. CONCA, R. ORIVE, AND M. VANNINATHAN
This completes the proof of (i).
The proof of (ii) is more involved. First, according to Theorem 1.3, we have the
uniform estimate
_

1
Y

[B

1
g()[
2
d
_
R
N
[g(x)[
2
dx,
and so, by the usual density arguments, it is enough to prove (ii) with g D(R
N
).
We can now complete the proof using Proposition 5.2. Indeed, with =

1
, we see
that

(0)
() p
(0)
and B

1
g() = J

g() R
N
,
which implies, by Lebesgues dominated convergence theorem, that
(2)
N/2

1
Y

()
(0)
() g() g() in L
2
(R
N

).
Proof of Proposition 5.2. The key point is that the variation of (
x

; ) with
respect to x is faster than that of g. To exploit this, we consider the -mesh Y

Z
N
generated by the cell Y which was already introduced at the end of section 2. We
decompose
J

g() =

Z
N
_
Y

g(x)e
ix

_
x

;
_
dx (5.3)
=

Z
N
g(x

)
_
Y

e
ix

_
x

;
_
dx +r

1
(),
where
r

1
() =

Z
N
_
Y

(g(x) g(x

))e
ix

_
x

;
_
dx. (5.4)
The rst term on the right-hand side of (5.3) can be, by means of the change of
variables x = x

+y, transformed into

[Y [
N
F

g()
(0)
(),
where F

g is the discrete Fourier transform of g and

(0)
is dened in (5.1). Since
we know that

1
Y
()
N
F

g()
1
(2)
N/2
g() in L
2
(R
N
) (cf. Lemma 2.6), our
hypothesis on ensures that
_
_
_

1
Y

()
_
[Y [
N
F

g() (2)
N/2
g()
_

(0)
()
_
_
_
L
2
(R
N
)
0. (5.5)
Thus, to complete the proof, it is enough to show that
|r

1
|
L
2
(
1
Y

)

c(K)
(1
N
p
)
||
L

(Y

;L
2
#
(Y ))
|g|
L
p
(R
N
)
. (5.6)
To this end, we rewrite r

1
in a slightly dierent form, namely,
r

1
() =
_
R
N
g

1
(x)e
ix

_
x

;
_
dx, (5.7)
BLOCH APPROXIMATION 1189
where
g

1
(x) =

Z
N
(g(x) g(x

))
Y

(x). (5.8)
We already know how to estimate integrals of the type (5.7) in L
2
(R
N
) (see Lemma
2.5), and so we can deduce (5.6) provided we have the estimate
| g

1
|
L
2
(R
N
)

c(K)
(1
N
p
)
|g|
L
p
(R
N
)
. (5.9)
Thanks to our hypothesis, we can deduce a stronger estimate, namely,
| g

1
|
L
p
(R
N
)

c
(1
N
p
)
|g|
L
p
(R
N
)
, (5.10)
where c is a constant independent of K, the support of g. We note that (5.10) is a
simple consequence of Morreys estimate (see [7, p. 167]).
Finally, we note that (5.9) can be obtained from (5.10) with c(K) = c[K[
1
2
p
and
a simple application of the H older inequality.
The proof of Proposition 5.2 shows that the result can be strengthened by as-
suming suitable smoothness on g. Our next result is an example in this direction. It
introduces naturally the following quantities:

(k)
() =
1
[Y [
_
Y
(y; )y
k
e
iy
dy k = 1, . . . , N, Y

. (5.11)
Then we have the following corrector result for J

g.
Proposition 5.3. We suppose L

(Y

; L
2
#
(Y )). Then for all g W
2,p
(R
N
)
with compact support K and with p > N, we have

1
Y

()
_
J

g() (2)
N/2
_

(0)
() +i
k

(k)
()

g()
_
0 in L
2
(R
N

).
Proof. We follow the idea of the proof of Proposition 5.2. We decompose J

g()
as
J

g() =

Z
N
_
Y

g(x

) +g(x

) (x x

) e
ix

_
x

;
_
dx +r

2
(), (5.12)
where
r

2
() =

Z
N
_
Y

g(x) g(x

) g(x

) (x x

) e
ix

_
x

;
_
dx. (5.13)
We can estimate r

2
() as follows:
|r

2
|
L
2
(
1
Y

)

c(K)
(2
N
p
)

2
||
L

(Y

;L
2
#
(Y ))
[g[
W
2,p
(R
N
)
. (5.14)
This, in fact, will be a consequence of Lemma 2.5, because we can represent r

2
as
follows:
r

2
() =
_
R
N
g

2
(x)e
ix

_
x

;
_
dx (5.15)
1190 C. CONCA, R. ORIVE, AND M. VANNINATHAN
with
g

2
(x) =

Z
N
(g(x) g(x

) g(x

) (x x

))
Y

(x), (5.16)
which admits the following estimates:
| g

2
|
L
2
(R
N
)

c(K)
(2
N
p
)

2
[g[
W
2,p
(R
N
)
, (5.17)
| g

2
|
L
p
(R
N
)

c
(2
N
p
)

2
[g[
W
2,p
(R
N
)
. (5.18)
As before, (5.17) will be a consequence of (5.18) with c(K) = c[K[
1
2
p
.
To establish (5.18), what we need is a generalization of Morreys inequality for
W
2,p
functions, namely,
[g(x)g(x

)g(x

) (xx

)[
c
(2
N
p
)
[xx

[
2
N
p
[g[
W
2,p
(Y

)
xY

. (5.19)
Admitting the above estimate, it is an easy matter to prove (5.18). But the above
estimate is a consequence of Morreys inequality for the gradient g W
1,p
(R
N
) and
the following representation: for all x Y

,
g(x) g(x

) g(x

) (x x

) =
_
1
0
g((1 t)x

+tx) g(x

) (x x

)dt.
This completes the proof of the estimate (5.14) on r

2
. Thus, as expected, r

2
tends to
zero more rapidly. The same cannot be said for the rst term on the right-hand side
of (5.12). Indeed, it is equal to
[Y [
_

N
(F

g)()
(0)
() +
N+1
_
F

g
x
k
_
()
(k)
()
_
. (5.20)
According to Lemma 2.6, we have the following convergence (apart from (5.5)):

1
Y

_
[Y [
_

N
_
F

g
x
k
_
()
1
(2)
N/2
i
k
g()
_

(k)
()
_
0 in L
2
(R
N

). (5.21)
This clearly allows us to complete the proof.
6. Proof of the main convergence results. Applying the previously devel-
oped techniques and results, we are now in a position to prove the main convergence
results stated in section 1.2 of the introduction (namely, Theorems 1.8 and 1.11 and
the statement (1.8)). We begin by recalling briey the set-up. We take f L
2
(R
N
)
and consider a sequence u

satisfying (1.9), i.e.,

_

_
A

= f in R
N
,
u

in H
1
(R
N
)-weak,
u

in L
2
(R
N
)-strong.
(6.1)
BLOCH APPROXIMATION 1191
6.1. No concentration of energy at innity. Our hypothesis that u

in L
2
(R
N
)-strong may, at rst sight, look articial. But this in not the case. If
is bounded and smooth, then it is classical that the weak convergence in H
1
() will
automatically imply the strong convergence in L
2
(). This is not the case in R
N
. To
make comparisons, the correct operator to consider is (A

+ I) instead of A

in R
N
.
In that case, we have the following.
Proposition 6.1. Assume that w

satises
_
(A

+I)w

= g in R
N
,
w

in H
1
(R
N
)-weak,
(6.2)
where g is a given function in L
2
(R
N
). Then
w

in L
2
(R
N
)-strong.
Proof. First of all, following the idea of Proposition 3.1, we can neglect higher
Bloch modes of w

and w

. More precisely, we can show

_

1
Y

m=2
[B

m
w

()[
2
d c
4
,
_

1
Y

m=2
[B

m
w

()[
2
d c
2
.
Therefore, it remains to prove
_

1
Y

[B

1
w

() B

1
w

()[
2
d 0. (6.3)
Equation (6.2) gives the relation
(1 +

1
())B

1
w

() = B

1
g(),
1
Y

.
We use it to write

1
Y

()(B

1
w

() B

1
w

()) =

1
Y

()
B

1
g()
1 +

1
()
w

() (

1
Y

()B

1
w

() w

()).
According to Proposition 5.1, the last term tends to zero in L
2
(R
N
). It suces to
show

1
Y

()
B

1
g()
1 +

1
()
w

() 0 in L
2
(R
N

). (6.4)
Note that w

satises the homogenized equation A

+ w

= g in R
N
, which is
equivalent to
_
1
2
D
2
k

1
(0)
k

+ 1
_
w

() = g(), R
N
.
1192 C. CONCA, R. ORIVE, AND M. VANNINATHAN
So, (6.4) is reduced to

1
Y

()
B

1
g()
1 +

1
()

g()
1 +
1
2
D
2
k

1
(0)
k

0 in L
2
(R
N

). (6.5)
The above expression can be written in the form
a

+b

,
where
a

=
_
1 +
1
2
D
2
k

1
(0)
k

__

1
Y

()B

1
g() g()
_
,
b

=
_

1
()
1
2
D
2
k

1
(0)
k

_
g(),
c

= (1 +

1
())
_
1 +
1
2
D
2
k

1
(0)
k

_
.
Now we have the convergence
a

1
Y

()B

1
g() g()
1 +

1
()
0 in L
2
(R
N

)
because [1 +

1
()] 1 and by the virtue of Proposition 5.1.
The convergence of
b

is not immediate. To show this, we split the energy into

three parts, taking > 0 as a xed constant:
_
||
1
||
_
b

_
2
d +
_
||
1
||>
_
b

_
2
d +
_
||>
1
_
b

_
2
d.
In the rst two parts, we use the estimate

1
()
1
2
D
2
k

1
(0)
k

c[[
3
for [[
1
, (6.6)
which holds since
1
(0) = D
1
(0) = 0 (see Proposition 1.5). In the rst integral, we
have c

1 and [b

()[ c
3
[ g()[, and consequently it is less than
c
2
_
R
N
[ g()[
2
d
and hence converges to zero. In the second integral, we have
c

1
2

1
()D
2
k

1
(0)
k

c[[
4
c[[
3
since [[ > 0.
With regards to b

, we still have [b

()[ c[[
3
[ g()[, and so the second integral also
converges to zero.
In the third integral, we use the bounds
[b

[
_

1
() +
1
2
D
2
k

1
(0)
k

_
[ g()[,
c

1
() +
1
2
D
2
k

1
(0)
k

.
BLOCH APPROXIMATION 1193
Thus the third integral is estimated from above by
_
||>
1
[ g()[
2
d.
Obviously, this tends to zero as 0 since g L
2
(R
N
).
6.2. Corrector result in R
N
. This section is devoted to the proof of Theo-
rem 1.8 concerning the Bloch approximation

. The proof consists of several steps

which correspond to estimations of the required energy in dierent regions in the
Fourier space (in a neighborhood of the origin [[ and in its complement [[ > ).
Step 1. We decompose u

as follows:
u

= v

+P

1
u

,
where v

and P

1
u

are dened in (1.21) and (1.27), respectively. Thanks to Proposi-

tion 3.1, it is enough to prove
|P

1
u

|
L
2
(R
N
)
0, (6.7)
[P

1
u

[
H
1
(R
N
)
c|f|
L
2
(R
N
)
. (6.8)
Step 2. We estimate the energies in the region [[ >
1
. To this end, we
introduce the quantities

,
(x) =
_

1
Y

||>
1
u

()e
ix

1
(x; )d, (6.9)
P
,
1
u

(x) =
_

1
Y

||>
1
B

1
u

()e
ix

1
(x; )d. (6.10)
We will obtain the estimates
|
,
|
L
2
(R
N
)
c|f|
H
1
(R
N
)
, (6.11)
[
,
[
H
1
(R
N
)
c|f|
L
2
(R
N
)
, (6.12)
|P
,
1
u

|
L
2
(R
N
)
c|f|
H
1
(R
N
)
, (6.13)
[P
,
1
u

[
H
1
(R
N
)
c|f|
L
2
(R
N
)
. (6.14)
We start with (6.14). Using Lemma 2.3 with =
1
and inequalities (2.5), we get
[P
,
1
u

[
2
H
1
(R
N
)
c
_

1
Y

||>
1
[B

1
u

()[
2

1
()d.
Now (6.14) easily follows if we use (1.22) and (1.24). Next, we prove (6.12). Following
the above procedure, we get
[
,
[
2
H
1
(R
N
)
c
_

1
Y

||>
1
[ u

()[
2
[[
2
d. (6.15)
1194 C. CONCA, R. ORIVE, AND M. VANNINATHAN
If f L
2
(R
N
), then it is well known that u

H
2
(R
N
) and
_
R
N
[[
4
[ u

()[
2
d c
_
R
N
[

f()[
2
d. (6.16)
Combining (6.15) and (6.16), we easily get (6.12). We now show (6.11). By Parsevals
identity, we have
|
,
|
2
L
2
(R
N
)
=
_

1
Y

||>
1
[ u

()[
2
d c

2
_

1
Y

||>
1
[[
2
[

f()[
2
d,
since u

and f are related by the homogenized equation A

= f in R
N
. This clearly
implies
|
,
|
2
L
2
(R
N
)
c

2
_

1
Y

||>
1
(1 +[[
2
)
1
[

f()[
2
d = c

2
|f|
2
H
1
(R
N
)
.
The proof of (6.13) is completely analogous.
Step 3. Now, we consider the energies in [[
1
. To this end, let us dene

(x) =
_
||
1
(B

1
u

() u

())e
ix

1
(x; )d (6.17)
and show that
|

|
L
2
(R
N
)
0, (6.18)
[

[
H
1
(R
N
)
c|f|
L
2
(R
N
)
. (6.19)
To prove (6.18), we decompose the integrand as follows:
B

1
u

= B

1
(u

) + (B

1
u

).
By Parsevals equality, the rst term in the L
2
-norm is bounded above by |u

|
L
2
(R
N
)
which, by our hypothesis, converges to zero. That the second term con-
verges to zero in L
2
(R
N
) is proved in Proposition 5.1.
Next, we turn are attention to the proof of (6.19). By Lemma 2.1, we have
[

[
2
H
1
(R
N
)
c
_
||
1

1
()[B

1
u

() u

()[
2
d. (6.20)
To estimate the above integral, we write the integrand as
B

1
u

() u

() =

1
()
1
(B

1
f()

f()) +
_

1
()
1

_
1
2
D
2
k

1
(0)
k

_
1
_

f().
Thus we get, using (1.24), that
_

_
[

[
2
H
1
(R
N
)
c
_
||
1
[B

1
f()

f()[
2
[[
2
d
+ c
_
||
1

1
()

1
()
1

_
1
2
D
2
k

1
(0)
k

_
1

2
[

f()[
2
d.
(6.21)
BLOCH APPROXIMATION 1195
To estimate the rst term on the right-hand side of (6.21), we represent the integrand
as
B

1
f()

f()
[[
=
_
R
N
f(x)e
ix
(

1
(x; )

1
(x; 0))
[[
dx.
Applying Lemma 2.5 and using |
1
(; )
1
(; 0)|
L
2
(Y )
c[[ for [[ , we get
_
||
1
[B

1
f()

f()[
2
[[
2
d c
2
|f|
2
L
2
(R
N
)
.
The second term on the right-hand side of (6.21) can be rewritten, using the homog-
enized equation, as
_
||
1
[

1
()
1
2
D
2
k

1
(0)
k

[
2

1
()
[ u

()[
2
d.
Using (6.6) and (1.24), we see that the above integral is estimated from above by
c
2
_
||
1
[[
4
[ u

()[
2
d c
2
|f|
2
L
2
(R
N
)
.
This establishes (6.19) and hence the result.
6.3. Asymptotic expansion of the Bloch approximation. In this conclud-
ing section, we prove Theorem 1.11.
Proof of (i). We have the following decomposition:

(x) u

(x) = z

(x) +
,
(x) +u
,
(x), (6.22)
where
z

(x) =
_

1
Y

||
1
u

()e
ix
(

1
(x; )

1
(x; 0))d, (6.23)
u
,
(x) =
1
(2)
N/2
_
||>
1
u

()e
ix
d, (6.24)
and
,
is dened in (6.9).
The second term has already been estimated in the L
2
-norm (see (6.11)). The
same proof shows that the third term admits a bound
|u
,
|
L
2
(R
N
)
c|f|
H
1
(R
N
)
c|u

|
H
1
(R
N
)
. (6.25)
To estimate the rst term on the right-hand side of (6.22), we must proceed
dierently. In fact, it is essential to use Lemma 2.3. We see then that
|z

|
2
L
2
(R
N
)
=
_

1
Y

||
1
[ u

()[
2
|

1
(; )

1
(; 0)|
2
L
2
(Y )
d.
Using the Lipschitz continuity of the map
1
(; ) L
2
(Y ) for [[ , we see
that the above integral can be majorized, and we obtain
|z

|
2
L
2
(R
N
)
c
2
_
||
1
[ u

()[
2
[[
2
d c
2
[u

[
2
H
1
(R
N
)
. (6.26)
1196 C. CONCA, R. ORIVE, AND M. VANNINATHAN
This nishes the proof of (i). We note that we cannot, in general, assert that
[u

[
H
1
(R
N
)
c|f|
H
1
(R
N
)
as we are working on the entire space R
N
.
Proof of (ii). Because of (i), it suces to prove

k
u

x
k

H
1
(R
N
)
c|f|
L
2
(R
N
)
. (6.27)
To this end, we use once again the decomposition (6.22) for (

) in terms of z

,
, and u
,
. For
,
, we have the estimate (6.12). For u
,
, we can easily derive
the estimate
[u
,
[
2
H
1
(R
N
)
c
_
||>
1
[[
2
[ u

()[
2
d c

2
|f|
2
L
2
(R
N
)
. (6.28)
Thus, we are reduced to obtaining the estimate

k
u

x
k

H
1
(R
N
)
c|f|
L
2
(R
N
)
. (6.29)
To this end, we use the representation
u

x
k
(x) =
1
(2)
N/2
_
R
N
(i
k
) u

()e
ix
d,
and combine it with the representation (6.23) for z

. We get
z

(x)

k
(x)
u

x
k
(x) =
_
||
1
u

()e
ix
_

1
(x; )

1
(x; 0) ip
(0)

k
(x)
k
_
d

_
||>
1
ip
(0)

k
(x)
k
u

()e
ix
d. (6.30)
To estimate the rst term on the right-hand side of (6.30), we appeal to Lemma 2.3.
Further, we use
_
_
_
1
(; )
1
(; 0) ip
(0)

k
()
k
_
_
_
H
1
(Y )
c[[
2
for [[ . (6.31)
The estimate on the second term on the right-hand side of (6.30) is more straightfor-
ward. We nally get

k
u

x
k

2
H
1
(R
N
)
c
2
_
R
N
[[
4
[ u

()[
2
d.
This completes the proof of (6.29) and hence (ii).
Proof of (iii). Consider again the decomposition (6.22). Thanks to (6.9) and
(6.15), we have the estimates
|
,
|
L
2
(R
N
)
c
2
|f|
L
2
(R
N
)
, (6.32)
[
,
[
H
1
(R
N
)
c
2
[f[
H
1
(R
N
)
. (6.33)
BLOCH APPROXIMATION 1197
Similar techniques imply
|u
,
|
L
2
(R
N
)
c
2
|f|
L
2
(R
N
)
, (6.34)
[u
,
[
H
1
(R
N
)
c
2
[f[
H
1
(R
N
)
. (6.35)
On the other hand, it is clear from the representation (6.30) that
_
_
_z

k
u

x
k
_
_
_
L
2
(R
N
)
c
2
|f|
L
2
(R
N
)
. (6.36)
Thus, it is enough to obtain the estimate

k
u

x
k
+
2
(

k
+
(2)
k
)

2
u

x
k
x

H
1
(R
N
)
c
2
[f[
H
1
(R
N
)
. (6.37)
Thanks to (6.33) and (6.35), we are reduced to showing that

k
u

x
k
+
2
(

k
+
(2)
k
)

2
u

x
k
x

H
1
(R
N
)
c
2
[f[
H
1
(R
N
)
. (6.38)
We can write
z

(x)

k
(x)
u

x
k
(x) +
2
(

k
(x) +
(2)
k
)

2
u

x
k
x

(x)
=
_
||
1
u

()e
ix
[

1
(x; )

1
(x; 0) ip
(0)

k
(x)
k
i
+ p
(0)
(

k
(x) +
(2)
k
)
2

]d

_
||>
1
ip
(0)

k
(x)
k
u

()e
ix
d
+
_
||>
1
p
(0)
(

k
(x) +
(2)
k
)
2

()e
ix
d. (6.39)
The analysis of the right-hand side of (6.39) is similar to that of (6.30). The new
information needed is the following:
_
_
_
1
(; )
1
(; 0) ip
(0)

k
()
k
+p
(0)
(
k
() +
(2)
k
)
k

_
_
_
H
1
(Y )
c[[
3
(6.40)
for [[ , which is a simple consequence of Proposition 1.10. The proof is concluded
via a simple application of Lemma 2.3.
Acknowledgments. This research stems from a couple of visits by R. Orive and
M. Vanninathan to the Center of Mathematical Modelling and to the Department of
Mathematical Engineering of the University of Chile, supported by the FONDAPs
Programme on Mathematical-Mechanics. These authors warmly thank these institu-
tions.
1198 C. CONCA, R. ORIVE, AND M. VANNINATHAN
REFERENCES
[1] G. Allaire, Homogenization and two-scale convergence, SIAM J. Math. Anal., 23 (1992),
pp. 14821518.
[2] G. Allaire and C. Conca, Bloch wave homogenization for a spectral problem in uid-solid
structures, Arch. Ration. Mech. Anal., 135 (1996), pp. 197257.
[3] G. Allaire and C. Conca, Boundary layers in the homogenization of a spectral problem in
uid-solid structures, SIAM J. Math. Anal., 29 (1998), pp. 343379.
[4] M. Avellaneda, L. Berlyand, and J.-F. Clouet, Frequency-dependent acoustics of compos-
ites with interfaces, SIAM J. Appl. Math., 60 (2000), pp. 21432181.
[5] A. Bensoussan, J.-L. Lions, and G. Papanicolaou, Asymptotic Analysis for Periodic Struc-
tures, North-Holland, Amsterdam, 1978.
[6] F. Bloch,

Uber die quantenmechanik der electronen in kristallgittern, Z. Phys., 52 (1928),
pp. 555600.
[7] H. Brezis, Analyse Fonctionnelle, Theorie et Applications, Collection Mathematiques Ap-
pliquees pour la Matrise, Masson, Paris, 1983.
[8] C. Conca, Bloch waves, in Encyclopaedia Mathematics, M. Hazewinkel et al., eds., Kluwer
Academic Publishers, Amsterdam, 1999, pp. 7274.
[9] C. Conca, R. Orive, and M. Vanninathan, Bloch Approximation in Bounded Domains,
manuscript, 2001.
[10] C. Conca, J. Planchard, and M. Vanninathan, Fluids and Periodic Structures, Rech. Math.
Appl., 38, John Wiley/Masson, New York, Paris, 1995.
[11] C. Conca and M. Vanninathan, Homogenization of periodic structures via Bloch decomposi-
tion, SIAM J. Appl. Math., 57 (1997), pp. 16391659.
[12] C. Conca and M. Vanninathan, On uniform H
2
-estimates in periodic homogenization, Proc.
Roy. Soc. Edinburgh Sect. A, 131 (2001), pp. 499517.
[13] G. Floquet, Sur les equations dierentielles lineaires ` a coecients periodiques, Ann. Ecole
Norm. Ser. 2, 12 (1883), pp. 4789.
[14] P. G erard, P.-A. Markowich, N.-J. Mauser, and F. Poupaud, Homogenization limits and
Wigner transforms, Comm. Pure Appl. Math., 50 (1997), pp. 323379.
[15] T. Kato, Perturbation Theory for Linear Operators, Springer-Verlag, Berlin, 1966.
[16] F. Murat and L. Tartar, H-convergence, in Topics in the Mathematical Modeling of Compos-
ite Materials, Progr. Nonlinear Dierential Equations Appl. 31, A. Cherkaev and R. Kohn,
eds., Birkhauser, Boston, 1997, pp. 2144.
[17] G. Nguetseng, A general convergence result for a functional related to the theory of homoge-
nization, SIAM J. Math. Anal., 20 (1989), pp. 608623.
[18] F. Odeh and J. Keller, Partial dierential equations with periodic coecients and Bloch
waves in crystals, J. Math. Phys., 5 (1964), pp. 14991504.
[19] M. Reed and B. Simon, Methods of Modern Mathematical Physics, Academic Press, New
York, 1978.
[20] F. Santosa and W. W. Symes, A dispersive eective medium for wave propagation in periodic
composites, SIAM J. Appl. Math., 51 (1991), pp. 9841005.
[21] L. Tartar, Probl`emes dhomogeneisation dans les equations aux derivees partielles, in Cours
Peccot au Coll`ege de France, 1977.
[22] C. Wilcox, Theory of Bloch waves, J. Anal. Math., 33 (1978), pp. 146167.