You are on page 1of 25

EXISTENCE AND NONEXISTENCE RESULTS FOR CRITICAL

GROWTH BIHARMONIC ELLIPTIC EQUATIONS


FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA
Abstract. We prove existence of nontrivial solutions to semilinear fourth order problems at critical growth in some contractible domains which are perturbations of small
capacity of domains having nontrivial topology. Compared with the second order case,
some difficulties arise which are overcome by a decomposition method with respect to
pairs of dual cones. In the case of Navier boundary conditions, further technical problems have to be solved by means of a careful application of concentration compactness
lemmas. Also the required generalization of a Struwe type compactness lemma needs a
somehow involved discussion of certain limit procedures.
A Sobolev inequality with optimal constant and remainder term is proved, which may
be of interest not only as a technical tool.
Finally, also nonexistence results for positive solutions in the ball are obtained, extending a result of Pucci and Serrin on so called critical dimensions to Navier boundary
conditions.

1. Introduction
We consider the following biharmonic critical growth problem
(1)

2 u = u + |u|2 2 u

in

either with Navier boundary conditions


(2)

u = u = 0

on

or with Dirichlet boundary conditions


(3)

u = u = 0

on .

2n
is the
Here Rn (n > 5) is an open bounded smooth domain, > 0 and 2 = n4
2
n
2
n
critical Sobolev exponent for the embedding H (R ) , L (R ). We are interested in
existence and nonexistence of nontrivial solutions of (1) with one of the above boundary
conditions.
We consider first the case where = 0, namely the equation

(4)

2 u = |u|8/(n4) u

in .

It is well known that (4)(2) and (4)(3) admit no positive solutions if is star shaped (see
[24, Theorem 3.3] and [26, Corollary 1]). Clearly, these are not exhaustive nonexistence
results as they only deal with positive solutions, see also Section 3 below for further
comments. On the other hand, by combining the Pohozaev identity [28] with the unique
continuation property [19, Section 3], it is known that one indeed has nonexistence results
Date: April 23, 2002.
1991 Mathematics Subject Classification. 31B30; 35J40; 58E05.
Key words and phrases. Semilinear biharmonic elliptic problems, Navier boundary conditions, critical
exponent, best Sobolev constant.
The first author was supported by MURST project Metodi Variazionali ed Equazioni Differenziali non
Lineari.
1

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

of any nontrivial solution for the correponding second order equation u = |u|4/(n2) u
in star shaped domains. This suggests that, in order to obtain existence results for (4), one
should either add subcritical perturbations or modify the topology or the geometry of the
domain . For general subcritical perturbations we refer to [4, 11] and references therein.
Domains with nontrivial topology are studied in [3] for Dirichlet and in [9] for Navier
boundary conditions. Here we are interested in topologically simple but geometrically
complicated domains. As far as we are aware, the case of strange contractible domains
has not been tackled before and this is precisely the first scope of the present paper. We
show that (4)(2) admits a positive solution in a suitable contractible domain . This
result is the exact counterpart of [24, Theorem 3.3] where it is shown that (4)(2) admits
no positive solution if is star shaped. For the Dirichlet boundary condition we obtain
a weaker result: we show that (4)(3) admits a nontrivial solution in the same kind
of contractible domain. Clearly, this still leaves open some problems concerning Dirichlet
boundary conditions, see Section 3. On one hand our proof is inspired by strong arguments
developed by Passaseo [27], on the other hand we have to face several hard difficulties,
especially (and somehow unexpectedly) under Navier boundary conditions. One of the
crucial steps in the approach by Passaseo is to prove that sign changing solutions of (4)
double the energy of the associated functional. For the second order problem this may
be shown by the usual technique of testing the equation with the positive and negative
parts of the solution. Of course, this technique fails for (4) where higher order derivatives
are involved and we overcome this difficulty thanks to the decomposition method in dual
cones developed in [12]. This method enables us to bypass the lack of nonexistence results
for nodal solutions of (4) in star shaped domains. Moreover, when dealing with Navier
boundary conditions, the required generalization of the Struwe compactness lemma [33]
turns out to be very delicate because of the second boundary datum and the lack of a
uniform extension operator for H 2 H01 functions in families of domains. See Lemma
5.4 and its proof in Section 8. The same problem arises in Lemma 6.1 where a uniform
lower bound for an enlarged optimal Sobolev constant has to be found in a suitable class
of domains.
Next, we restrict our attention to the case where > 0 and = B, the unit ball
in Rn . We are interested in positive radially symmetric solutions. Let 1 > 0 be the
first eigenvalue of 2 with homogeneous Dirichlet boundary conditions. A celebrated
result by PucciSerrin [30] shows that (1)(3) admits a nontrivial radially symmetric
solution for all (0, 1 ) if and only if n > 8 (n is the space dimension). If n = 5, 6, 7
the range 0 < < 1 is no longer correct since there exists (0, 1 ) such that
(1)(3) admits no nontrivial radial solution whenever (0, ]; moreover, there exists
[ , 1 ) (presumably = ) such that (1)(3) admits even a positive radially
symmetric solution for all ( , 1 ). In other words, the well known [6] nonexistence
result for radially symmetric solutions of the second order problem for small in dimension
n = 3 bifurcates for (1)(3) to the dimensions n = 5, 6, 7: PucciSerrin call these
dimensions critical. Furthermore, the space dimensions n for which one has nonexistence
of positive radially symmetric solutions of (1)(3) for in some (right) neighborhood of
0 are called weakly critical. This latter notion was introduced in [15] in order to study
higher order polyharmonic equations for which critical dimensions seem more difficult to
handle; clearly, critical dimensions are also weakly critical. In Theorem 3.1 below we prove
that the dimensions n = 5, 6, 7 are weakly critical also for Navier boundary conditions.
Therefore, critical dimensions seem not to depend on the boundary conditions. This result
is perhaps related to the fact that critical dimensions may have an explanation in terms

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

of the summability of the fundamental solution corresponding to the differential operator


2 , see [17, 23].
Finally, when = B we show that (4)(3) admits no nontrivial radially symmetric
solution; note that no positivity assumption on the solution is made. This result was
already obtained in [14, Theorem 3.11]; we give here a different (and simpler) proof.
2. Existence results
n

n
By D2,2
R (R ) we2 denote the completion of the space Cc (R ) with respect to the norm
2
kf k2,2 = Rn |f | ; this is a Hilbert space when endowed with the scalar product
Z
(f, g)2,2 =
f g dx.
Rn

By solution of (1) we mean here a function u H 2 H01 () (when dealing with (2)) or a
function u H02 () (when dealing with (3)) which satisfies
Z
Z 

u dx =
u + |u|8/(n4) u dx

H 2 H01 ()

H02 ()).

for all
(resp.
For Dirichlet boundary conditions this variational
formulation is standard, while for Navier boundary data we refer to [4, p.221]. By well
known regularity results (see [22, Theorem 1] and [38, Lemma B3]), if C 4, , then
the solutions u of (1) satisfy u C 4, () and solve (1) in the classical sense.
Definition 2.1. Let K Rn bounded. We say that u D2,2 (Rn ) satisfies u > 1 on K in
the sense of D2,2 (Rn ) if there exists a sequence (uh ) in Cc2 (Rn ) such that uh > 1 on K for
each h N and uh u in D2,2 (Rn ). Analogously, u 6 1 and u = 1 on K are defined.
Definition 2.2. We define the (2, 2)capacity of K (cap K) as
Z

2
2,2
n
|u| dx : u = 1 on K in the D (R ) sense .
cap K = inf
Rn

We set cap := 0.
Since the nonempty set


u D2,2 (Rn ) : u = 1 on K in the D2,2 (Rn ) sense
is closed and convex, there exists a unique function zK D2,2 (Rn ) such that zK = 1 on
K and
Z
|zK |2 dx = cap K.
Rn

Finally we make precise what we mean by set deformations:


e Rn and let H, .
e We say that H cannot be deformed in
Definition 2.3. Let
e
into a subset of if there exists no continuous function
e
H : H [0, 1]
such that H (x, 0) = x and H (x, 1) for all x H.
Our first result states the existence of positive solutions for the critical growth equation
(4) with Navier boundary conditions. Combined with the already mentioned nonexistence result in star shaped domains [24], this shows that the existence of positive solution
strongly depends on the geometry of the domain.

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

e be a smooth bounded domain of Rn (n > 5) and let H be a closed


Theorem 2.4. Let
e Then there exists > 0 such that if
e is a smooth domain
subset contained in .
e
e
with cap( \ ) < and such that H cannot be deformed in into a subset of then
there exists a positive solution of
(
2 u = u(n+4)/(n4) in
(5)
u = u = 0
on .
We will prove this result in Section 6.
Next we turn to Dirichlet boundary conditions. In this case, we merely show the
existence of nontrivial solutions to (4). The lack of information about the sign of the
solution is due to the lack of information about the sign of the corresponding Green
function. Moreover, in general domains sign change has even to be expected. See [16] for
a survey on this feature and for further references.
e be a smooth bounded domain of Rn (n > 5) and let H be a closed
Theorem 2.5. Let
e Then there exists > 0 such that if
e is a smooth domain
subset contained in .
e \ ) < and such that H cannot be deformed in
e into a subset of then
with cap(
there exists a nontrivial solution of
(
2 u = |u|8/(n4) u in
(6)
u = u = 0
on .
The proof of this result is simpler than the one of Theorem 2.4. As we will explain
below, for (5), one has to study very carefully the behaviour of suitable sequences uh
H 2 H01 (h ) for varying domains h . In contrast with the spaces H02 (h ), there is no
trivial extension operator into H 2 (Rn ). Only as positivity of solutions is concerned, the
situation with respect to Dirichlet boundary conditions is more involved than with respect
to Navier boundary conditions. Here we have no positivity statement of the solution
because Lemma 5.3 below does not seem to hold.
If is an annulus and if one restricts to radial functions, the problem is no longer
critical and then both (5) and (6) admit a nontrivial solution. For the second order
problem u = |u|4/(n2) u in , u = 0 on , it was shown with help of very subtle
methods by BahriCoron [2] that this result extends to domains with nontrivial topology.
Recently, corresponding results have been found also in the higher order case: see [9]
for problem (5) and [3] for the polyharmonic analogue of (6) under Dirichlet boundary
conditions. However, the solutions which are constructed in these papers and here are not
e satisfies the assumptions of [9] or [3]. If u is the
related. To explain this, assume that
solution of (5) or (6), then by exploiting the proof of Theorem 2.4 one has that u tends
e \ ) 0. Hence, any nontrivial solution in
e
to disappear via concentration as cap(
may not be obtained as limit of the solutions u in .
To conclude the section, we observe that there are also contractible domains , which
satisfy the assumptions of the preceding theorems. In the following example we describe
precisely such a situation. Further examples may be adapted to the biharmonic setting
from [27, pp. 3941].
e Rn with n > 5, where
Example 2.6. We consider an annular shaped domain as
we drill a sufficiently thin cylindrical hole along a straight line in order to obtain the
smooth contractible subdomain .

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

e \ is contained in a cylinder
To be more precise: we assume that for small enough,
n1
withbasis B R
and fixed height. Then by simple scaling arguments one finds that
n5
e
cap \ = O(
) 0 as 0, provided that the dimension satisfies n > 5.
e which can certainly not be deformed into
As H, we take a spherical hypersurface in ,
a subset of . This can e.g. be seen by looking at the degree of mapping d(H ( . , t), H, 0)
for t [0, 1], where H is assumed to exist according to Definition 2.3.
Instead of the above mentioned segment of fixed length, one may consider any bounded
piece of a fixed generalized plane, provided that its codimension is at least 5.
3. Nonexistence results
We collect here a number of known as well as of new nonexistence results for the equation
(7)

2 u = u + |u|8/(n4) u in

under either Navier or Dirichlet boundary conditions. Here, we include the linear term u.
In the borderline case = 0, where the purely critical equation has to be considered, the
behaviour of boundary value problems for (7) switches and, in contrast with the second
order case where the principal part is the Laplacian, it is indeed difficult to decide to which
of the two cases < 0 or > 0 it is more similar.
3.1. Under Navier boundary conditions. It turns out that nonexistence questions in
this case are relatively hard and that only restricted results are available.
The case = 0
In this case, no positive solution to (5) may exist, provided is star shaped. See [37,
Theorem 3.10] and also [24, Theorem 3.3].
The key ingredient is a Pohozaev type identity [29], which reads in this case:
Z
u (u)
(8)
x dS = 0.


Here / denotes the directional derivative with respect to the exterior unit normal
= (x). From the assumption u > 0, by applying twice the maximum principle for ,

one gets
(u) < 0 as well as u
< 0 on the boundary and hence a contradiction with
(8). By means of techniques developed by Pucci and Serrin in [30], such a nonexistence
result may be also shown for nontrivial radial solutions in the ball.
More difficult seems this nonexistence result for any nontrivial solution of (5). For the
second order equation u = |u|4/(n2) u (with u = 0 on ), the Pohozaev identity
yields u/ = 0 on ; moreover, from the equation itself we infer 2 u/ 2 = 0. Then,
by applying in a suitable fashion the unique continuation property (see [19, Section 3]) one
obtains that u 0. Returning to (5), in order to apply the unique continuation property
[31] one should have
3u
u
=
=0
on

3
but, to this end, (8) is not powerful enough.
The case > 0
In Section 7 we prove the following result.

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

Theorem 3.1. Let n {5, 6, 7} and assume that = B Rn is the unit ball. Then
there exists a number > 0 such that a necessary condition for a positive solution to
(
2 u = u + u(n+4)/(n4) in
(9)
u = u = 0
on
with > 0 to exist is > .
By [36], positive solutions to (9) are radially symmetric.
Theorem 3.1 answers a question which was left open by van der Vorst in [39, Theorem
3]. Moreover, in [39, Theorem 1] complementary existence results were shown implying
that such a nonexistence result can hold at most in the dimensions 5, 6, 7.
These are the critical dimensions for problem (9), at least, when restricting to positive
solutions in the ball. Theorem 3.1 can also be shown just for nontrivial radial solutions:
here one has to combine methods of [30] with some techniques, which appear in the proof
of Theorem 3.2 below.
Again one observes an unsymmetry between existence and nonexistence results: all
nonexistence results are restricted to the class of positive or at least radial functions.
Nevertheless, one would expect nonexistence of any nontrivial solution. A proof, however,
seems to be out of reach.
The case < 0
In the preceding nonexistence result, the positivity of a solution, the nonexistence of which
had to be shown, is intensively exploited. For this we use in the proof of Theorem 3.1 that
the right hand side of (9) maps R+ R+ (when > 0) and that the solution operator
for the biharmonic operator under Navier conditions is positivity preserving.
For negative, we can only argue as before, when is close enough to 0, see [18]. Then
we can still conclude that u > 0 in . But when << 0, the solution operator for
2 is no longer positivity preserving, see [32, pp. 99/100] and also [7]. Hence, the
framework of positive solutions is no longer adequate for (7), and the argument in [37, pp.
390/391] and [39, Theorem 3] breaks down. Instead, one has to look for nonexistence of
any nontrivial solution. But such a result seems to be still unknown, even for nontrivial
radial solutions in the ball.
3.2. Under Dirichlet boundary conditions. In this situation, existence results for
positive solutions are more involved than under Navier boundary conditions (see [13]) and
cannot even be expected in general domains. On the other hand, we know much more
about nonexistence for relatively large classes of solutions.
The case < 0
Nonexistence of any nontrivial solution is shown in [29], provided the domain is star
shaped. This proof is based on a generalized Pohozaev identity and covers also the general
polyharmonic problem with ()K as linear principal part.
The case > 0
In [30] we already find the following result, which gives a stronger statement than the
analogous Theorem 3.1 above for Navier boundary conditions:

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

Let n {5, 6, 7} and assume that is the unit ball B. Then there exists
a number > 0 such that a necessary condition for a nontrivial radial
solution to (6) to exist is > .
As this special behaviour of the critical growth Dirichlet problem (6) may be observed only
in space dimensions 5, 6 and 7, Pucci and Serrin call these dimensions critical. For the critical dimension phenomenon for general polyharmonic problems with Dirichlet boundary
conditions we refer to [15].
The case = 0
Also in this case, the nonexistence of any nontrivial solution to (6) has to be left open,
only more restricted results are available. Oswald [26] proved the nonexistence of positive
solutions in strictly star shaped domains. His proof is based on a Pohozaev type identity
(see e.g. [24, (2.6)]) and gives, after integration by parts
Z
(10)
(u)2 (x )dS = 0.

Since x > 0 on , (10) entails u = 0 on and therefore


(11)

2u
=0
2

on .

As the solution is assumed to be positive and to satisfy homogeneous Dirichlet boundary


conditions, one directly concludes that u > 0 in and obtains a contradiction with
the Hopf boundary point lemma for superharmonic functions. For nodal solutions of the
biharmonic problem (6), the second order boundary condition (11) is not enough to apply
the unique continuation property [31]; one would also need
(12)

3u
=0
3

on .

Therefore, if one could prove that any solution u of (6) satisfies (12), then we would have
the desired nonexistence result in star shaped domains.
Moreover, we can also exclude nontrivial radial solutions:
Theorem 3.2. Let B Rn (n > 5) be the unit ball. Then problem
( 2
u = |u|8/(n4) u in B,
(13)
u = u = 0
on B
admits no nontrivial radial solution.
This result can be found in [14, Theorem 3.11]. The proof there is based on a refined
integral identity of Pucci and Serrin [30] and some Hardy type embedding inequalities. In
Section 7 we will give a more elementary and geometric proof.
Unfortunately neither this proof nor the one given in [14, Theorem 3.11] carry over to the
general polyharmonic problem, where the discussion of nonexistence of nontrivial radial
solutions in the borderline case = 0 remains essentially open.

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

4. A Sobolev inequality with optimal constant and remainder term


Sobolev embeddings with critical exponents and the corresponding optimal Sobolev
constants will play a crucial role in the proof of our existence results.
Let us set
(14)

S=

inf

uD2,2 (Rn )\{0}

kuk22
.
kuk22

When working in the whole Rn , this optimal Sobolev constant is attained and concerning
the minimizers the following result is known.
Lemma 4.1. The constant S in (14) is a minimum and (up to nontrivial real multiples)
it is attained only on the functions
(15)

u,x0 (x) =

[(n 4)(n 2)n(n + 2)](n4)/8 (n4)/2


,
(2 + |x x0 |2 )(n4)/2

for any x0 Rn and each > 0. Moreover, the functions u,x0 are the only positive
solutions of the equation
(16)

2 u = u(n+4)/(n4)

in Rn .

Proof. See [10, Theorem 2.1], [34, Theorem 4] and [20, Theorem 1.3].

Similarly one can define S() for any Rn . It is well known that S() is not attained
if is bounded [10] and that it does not depend neither on the boundary conditions [38]
nor on the domain. The following extension of the embedding inequality corresponding to
(14) with optimal constant S gives a quantitative formulation for its nonattainment.
n
Theorem 4.2. Let Rn be open and of finite measure and p [1, n4
). Then there
2
1
exists a constant C = C(n, p, ||) > 0, such that for every u H H0 (), one has
Z
1
(17)
|u|2 dx > S kuk22 + kuk2p .
C

Here, || denotes the n-dimensional Lebesgue measure of . Moreover, if > 0, then


(17) holds with a constant C = C(n, p, ) for all with || 6 .
For the smaller class H02 () a more refined result can be found in [12]. The proof there
bases on an extension and decomposition method (see also Section 5 below), by means of
which analogues to (17) for the spaces H0m () of any order can be also obtained. This
proof, however, does not seem to carry over directly to the space H 2 H01 due to the lack
of information at the boundary. In Section 7 we sketch a technical proof of Theorem 4.2,
which is based on nonexistence results like in Theorem 3.1 and on Talentis symmetrization
result.
5. Preliminary results
In the previous section we have briefly discussed energy minimizing solutions of (16).
These are necessarily of one sign (this is obtained as a byproduct of the following lemma)
and of the form given in Lemma 4.1. The next results show that nodal solutions of (4)
double the energy level of the corresponding free energy functional (observe that here,
we are working with a constrained functional). This will allow us to bypass the lack of
nonexistence results for nodal solutions of (4) in star shaped domains.

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

Lemma 5.1. Let = Rn and u D2,2 (Rn ) be a nodal solution of the equation
2 u = |u|8/(n4) u

(18)

in Rn .

Then kuk22 > 24/n Skuk22 .


Proof. Consider the convex closed cone


K = u D2,2 (Rn ) : u > 0 a.e. in Rn ,
and its dual cone

K 0 = u D2,2 (Rn ) :

(u, v)2,2 6 0 v K

Let us show that K 0 K . For each h Cc (Rn ) K consider the solution uh of the
problem
2 uh = h in Rn .
Then by the positivity of the fundamental solution of 2 in Rn , there results uh K and
thus if v K 0 ,
Z
Z

n
h Cc (R ) K :
hv dx =
2 uh v dx = (uh , v)2,2 6 0.
Rn

Rn

By a density argument one obtains v 6 0 a.e. in Rn , i.e. v K . Now, by a result of


Moreau [25], for each u D2,2 (Rn ) there exists a unique pair (u1 , u2 ) in K K 0 such
that
(19)

u = u1 + u2 ,

(u1 , u2 )2,2 = 0.

Let u be a nodal solution of (18) and let u1 K and u2 K 0 be the components of u


according to this decomposition. It results that ui 6 0 and
(20)

|u(x)|2 2 u(x)ui (x) 6 |ui (x)|2 ,

i = 1, 2,

for a.e. x Rn . Indeed, if i = 1 and u(x) 6 0 then (20) is trivial, while if u(x) > 0, since
u2 K one has u(x) = u1 (x) + u2 (x) 6 u1 (x) and again (20) holds. The case i = 2 is
similar. By combining the Sobolev inequality with (19) and (20), we get for i = 1, 2
Z
Z
2
2
Skui k2 6 kui k2,2 =
uui dx =
2 uui dx =
Rn
Rn
Z
Z
=
|u|2 2 uui dx 6
|ui |2 dx = kui k22 ,
Rn

Rn

which implies kui k22 > S (n4)/4 for i = 1, 2. Hence, again by (19), one obtains
Z
 Z
2/2 Z
4/n
kuk22
2
2
2
=
|u| dx
|u| dx
=
|u| dx
=
kuk22
Rn
Rn
Rn
4/n
Z
Z
2
2
=
|u1 | dx +
|u2 | dx
>
Rn

Rn


4/n
> Sku1 k22 + Sku1 k22
> 24/n S ,
which concludes the proof.
In the case of the half space we have a similar result.

10

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

Lemma 5.2. Let = {xn > 0} be the half space and assume that u D2,2 () solves the
equation
2 u = |u|8/(n4) u

(21)

in {xn > 0}

with boundary data either (2) or (3). Then kuk22 > 24/n Skuk22 .
Proof. Notice first that by [24, Theorem 3.3], equation (21) with (2) does not admit
positive solutions. A similar nonexistence result holds for boundary conditions (3), see
[26, Corollary 1] which may be easily extended to unbounded domains.
Therefore, any nontrivial solution of (21) (with (2) or (3)) necessarily changes sign.
We obtain the result if we repeat the argument of the proof of Lemma 5.1. With Navier
boundary conditions this is straightforward, while with Dirichlet boundary data one may
invoke Boggios principle [5] in the half space.

In a completely similar fashion, one may extend the previous result to any domain in
case of Navier boundary conditions. For Dirichlet boundary conditions this seems not
possible due to the lack of information about the positivity of the corresponding Green
function.
Lemma 5.3. Let be a bounded domain of Rn and assume that u solves (5) and changes
sign. Then kuk22 > 24/n Skuk22 .
As pointed out in the introduction, the H 2 H01 framework is more involved than the
H02 case. Therefore, from now on we restrict our attention to the first situation.
Consider the functional f : V () R
Z
(22)
f (u) =
|u|2 dx

constrained on the manifold


V () =

(23)

uH

H01 ()

Z
:


|u| dx = 1 .
2

We say that (uh ) V () is a PalaisSmale sequence for f at level c if


lim f (uh ) = c,
h

lim kf 0 (uh )k(Tu


h

V ())

where Tuh V () denotes the tangent space to the manifold V () at uh .


We can now state a global compactness result for the biharmonic operator, in the spirit
of [33]. Even if the proof is on the lines of that of Struwe, some difficulties arise. In the
appendix we give an outline of the proof emphasizing the main differences with respect to
second order equations.
Lemma 5.4. Let (uh ) V () be a PalaisSmale sequence for f at level c R. Then, for
a suitable subsequence, one has the following alternative: either (uh ) is relatively compact
in H 2 H01 () or there exist k nonzero functions u
b1 , ..., u
bk D2,2 , solving either (18) or
(21) and a solution u
b0 H 2 H01 () of (5) such that

1/2
k Z
X
uh * u
b0
(24)
|b
uj |2 dx
in H 2 H01 ()
j=0

and
(25)

2/2
k Z
k Z
X
X
lim f (uh ) =
|b
uj |2 dx
|b
uj |2 dx
.
h

j=0

j=0

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

11

The domain of integration for u


b0 is just , while for u
b1 , ..., u
bk , it is either a half space or
n
the whole R .
We believe that nontrivial solutions of (21) do not exist and that the previous lemma
may be strengthened (see Section 3.1). Nevertheless, Lemma 5.4 is sufficient to prove the
following compactness property in the second critical energy range.
Lemma 5.5. Let (uh ) V () be a PalaisSmale sequence for f at level c (S, 24/n S).
Then, up to a subsequence, (uh ) strongly converges in H 2 H01 ().
Proof. Assume by contradiction that the sequence (uh ) is not relatively compact in H 2
H01 (). Then, by Lemma 5.4 one finds functions u
b0 H 2 H01 () and u
b1 , ..., u
bk
2,2
n
D (R ) satisfying (24) and (25). Assume first that all u
b1 , . . . , u
bk are positive solutions of
(18). Then, by Lemma 4.1 each u
bj is of type (15) and attains the best Sobolev constant,
i.e.
Z
2/2
Z
Z
2
2
2
|b
uj | dx =
|b
uj | dx = S
|b
uj | dx
(26)
, j = 1, . . . k,
Rn

which implies

Rn
Rn

Rn

|b
uj |2 dx = S n/4 for j = 1, ..., k. In particular, one obtains
4/n
Z
.
|b
u0 |2 dx + kS n/4
lim f (uh ) =
h

If u
b0 0, it results f (uh )
Z

k 4/n S,

while if u
b0 6 0 for each k > 1 one has
4/n
2
n/4
|b
u0 | dx + kS
> (k + 1)4/n S.

In any case we have a contradiction with S < lim f (uh ) < 24/n S.
h

We now consider the case in which at least one u


bj is sign changing or a solution in the
half space {xn > 0}. By Lemmas 5.1 and 5.2 we then have for these u
bj :
2/2
Z
Z
Z
(27)
,
|b
uj |2 dx > 24/n S
|b
uj |2 dx
|b
uj |2 dx =
Rn

Rn

Rn

|b
uj |2 dx > 2S n/4 , while for the remaining u
bj , (26) holds true. In any case,
4/n
n/4
4/n
we have lim f (uh ) > 2 S
= 2 S, again a contradiction.


and hence

Rn

6. Proof of Theorems 2.4 and 2.5


The proof of Theorem 2.4 is by far more involved than the proof of Theorem 2.5, because
the trivial extension of any function u H 2 H01 () by 0 does not yield a function in
H 2 (Rn ). In particular, for the space H02 () Lemma 6.1 easily follows by this extension
argument. A further difference is that the positivity conclusion of Lemma 5.3 does not
hold in the Dirichlet boundary value case. But in the latter case, one simply has to drop
this argument. Therefore we only deal with the proof of Theorem 2.4.
e let : V () Rn be the barycenter map
For all smooth
Z
x|u(x)|2 dx.
(u) =

e is smooth one finds re > 0 such that


e is a deformation retract of
Since
n
o
e+ = x Rn : d(x, )
e < re .

12

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

First we show that the energy or, equivalently, the optimal Sobolev constants will remain
relatively large if we prevent the functions from concentrating too close to their domain
of definition.
Lemma 6.1.
= inf

n n
o
o
e+ , smooth subset of
e > S.
inf f (u) : u V (), (u) 6

e and
Proof. Assume by contradiction that for each > 0 there exists a smooth
2
1

u H H0 C ( ) such that
Z
(28)
|u |2 dx = 1

Z
(29)

|u |2 dx 6 S +

e+ .
(u ) 6

(30)

Let U H 2 (Rn ) be any entire extension of u . As is smooth, the existence of such


an extension is well known. We emphasize that the quantitative properties of U outside
will not be used (and are expected to blow up for & 0). Further, let 1 be the
characteristic function of .
Step I. We claim that 1 U 0 in L2 (Rn ) as 0.
By Lemma 4.2 there exists C > 0 independent of such that
u H 2 H01 ( ) : kuk22 > Skuk22 +

(31)

1
kuk21 .
C

Putting together (28), (29) and (31), for each > 0 it results
Z
1
1
S+>
|u |2 dx > Sku k22 + ku k21 = S + k1 U k21
C
C

so that k1 U k1 0 as 0. By (28) and classical Lp interpolation, we also get


k1 U k2 0 as 0.
Step II. The weak* limit (in the sense of measures) of 1 |U |2 is a Dirac mass.
Integration by part shows that ku k2L2 ( ) 6 ku kL2 ( ) ku kL2 ( ) . Then by Step I
and (29), we infer that
k1 U kL2 (Rn ) = ku kL2 ( ) 0.
Therefore, taking any Cc (Rn ), as 0 it results that
Z
Z
2
(32)
1 |(U )| dx =
1 ||2 |U |2 dx + o(1) .
Rn

Rn

L1 (Rn )

Now, by the
boundedness of the sequences (1 |U |2 ) and (1 |U |2 ) there exist
two bounded non negative measures , on Rn such that
(33)

1 |U |2 * ,

1 |U |2 *

in the sense of measures. By the Sobolev inequality in we have


Z
2/2 Z
2
2
> 0 : S
1 |U | || dx
6
1 |(U )|2 dx
Rn

Rn

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

and therefore, by (32) and (33), letting 0 yields


Z
2/2 Z
2
S
|| d
6

||2 d.

By (28) and (29) we also know that


Z
S

d.

Rn

13

Rn

2/2
d

Z
=
Rn

Rn

e and > 0 such that = x . Again from


Hence by [21, Lemma I.2] there exist x
(28) we finally see that = 1.
The contradiction follows by Step II, since (u ) x, against (30). This completes the
proof.

According to the previous lemma, we may choose such that S < < min{24/n S, }.
Let Cc (B1 (0)) be such that
Z
Z
2
(34)
|| dx = 1,
||2 dx < .
B1 (0)

Define for each > 0 and y

B1 (0)

Rn

the function ,y : Rn R by setting




xy
,y (x) =

where is set to zero outside B1 (0) . It is readily seen that there exists > 0 such that
e and hence ,y C ()
e for each y H . For all (
e let z D2,2 (Rn ) be
B (y)
c
e
such that z = 1 in \
and
Z
e
|z |2 dx = cap(\),
Rn

see what follows Definition 2.1. Note that one has


(1 z )Ty () H 2 H01 () y H,
where Ty () :=

,y
k,y k2 .

Moreover, for each > 0 there exists > 0 with


sup kz Ty ()kH 2 H01 () <

yH

e
whenever cap(\)
< . Then one gets
k(1 z )Ty ()kL2 () 6= 0 y H,
if is sufficiently small. So, if we define the map : H V () by
(y) =

(1 z ) Ty ()
.
k(1 z ) Ty ()k2

Taking into account (34), we find > 0 such that


(35)

sup{f ( (y)) : y H} <

e
provided that cap(\)
< . From now on is fixed subject to the previous restrictions.
e
e
Let be such that cap(\)
< and r (0, re) be such that is a deformation
retract of


+ = x Rn : d(x, ) < r .

14

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

As in Lemma 6.1, one obtains


inf {f (u) : u V (), (u) 6 + } > S.

(36)
Notice that
(37)

inf {f (u) : u V (), (u) 6 + } 6 sup {f ( (y)) : y H}

e given by
otherwise the map R : H [0, 1]
R(y, t) = (1 t)y + t( (y)) y H, t [0, 1]
e into a subset of + and then into a subset of ( is a deformation
would deforme H in
e we used
retract of + ) contradicting the assumption. Here, in order to see R(y, t) ,
e
that supp (y) B (y) and, as > 0, that ( (y)) B (y) .
Therefore, by combining (35), (36) and (37) one gets
(38)

S < inf {f (u) : u V (), (u) 6 + } 6 sup {f ( (y)) : y H}


n
o
e+ .
< < 6 inf f (u) : u V (), (u) 6

In what follows we need to find two appropriate levels, such that the corresponding sublevel
sets
f c = {u V () : f (u) 6 c}
cannot be deformed into each other. For this purpose let c1 , c2 > S be such that
c1 < inf{f (u) : u V (), (u) 6 + } ,
c2 = sup{f ( (y)) : y H}.
Assume by contradiction that there exists a deformation of f c2 into f c1 , i.e.
: f c2 [0, 1] f c2 ,

( . , 0) = Idf c2 and (f c2 , 1) f c1 .

e by setting for each x H


We define H : H [0, 1]

(1 3t)x + 3t( (x))


H (x, t) = %e((( (x), 3t 1)))

%e (% ( (( (x), 1)) , 3t 2))

if t [0, 1/3]
if t [1/3, 2/3]
if t [2/3, 1]

e+
e is a retraction and % : + [0, 1] + is a continuous map with
where %e :
e one should
%(x, 0) = x and %(x, 1) for all x + . In order to see that H (x, t) ,
e+ .
observe that c2 < and ( (x), 3t 1) f c2 , hence (( (x), 3t 1))
c
1
As ( (x), 1) f for each x H and
c1 < inf {f (u) : u V (), (u) 6 + } ,
then for each x H
(( (x), 1)) +
e into a subset of , in contradiction with our assumption.
and H is a deformation of H in
Then the sublevel set
f c2 = {u V () : f (u) 6 c2 }
cannot be deformed into
f c1 = {u V () : f (u) 6 c1 } .
Hence, by combining Lemma 5.5 with the standard deformation lemma one obtains a
constrained critical point u such that
S < f (u ) 6 sup{f ( (y)) : y H} < < 24/n S.

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

Finally, u does not change sign by Lemma 5.3.

15

7. Proof of Theorems 3.1, 3.2 and 4.2


Proof of Theorem 3.1. We assume that = B = B1 (0) is the unit ball and that (9)
with > 0 has a positive solution u. As we have already mentioned above, the result
by Troy [36, Theorem 1] shows that u is radially symmetric. As the right hand side in
(9) is positive, an iterated application of the maximum principle for yields that u
and u are strictly positive and strictly radially decreasing. We will exploit the following
Pohozaev type identity, which may be taken e.g. from [37, Lemma 3.9]:

Z
Z  

u
2
(u) dS.
(39)
2
u dx =
r
r
B
B
As u is radially symmetric we may proceed as follows with the term on the right hand
side; here, en denotes the n-dimensional volume of the unit ball B and c(n) denotes a
generic positive constant which may vary from line to line:

Z  
u

(u) dS = n en (1)
(u)(1)
r
r
r
r
B
Z 

 Z 


1
u

dS
(u) dS
n en
r
B r
ZB
 Z

1
(40)
=
(u) dx
2 u dx .
n en
B
B
The first integral needs some further consideration. First of all, by making use of homogeneous Navier boundary conditions, we find:
Z
Z
n Z
X
 2
2
(2xj )(u)xj dx = 2 n
(u) dx.
1 |x| u dx =
B

j=1

Furthermore, as we have already mentioned, u and by (9) also 2 u are positive and strictly
radially decreasing. This fact can be used to get rid of the degenerate weight factor 1|x|2
in the previous term:
Z
Z
Z
2 u dx 6
2 u dx +
2 u dx
|x|61/2

1/26|x|61

 
1
6
u dx + c(n) u
2
|x|61/2
Z
Z
6
2 u dx + c(n)
2 u dx
|x|61/2
|x|61/2
Z

6 c(n)
1 |x|2 2 u dx.
Z

Combining this estimate with the previous identity, we obtain from (39) and (40):
Z
2
Z
2
2
(41)

u dx > c(n)
u dx = c(n)k2 uk21 .
B

By a duality argument and using elliptic estimates, we find


(42)

kuk2 6 c(n)k2 uk1 , provided n < 8.

16

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

Observe that L2 estimates for the biharmonic operator under homogeneous Navier boundary conditions follow immediately from the L2 estimates for the Laplacian.
Combining (41) and (42), we see that for any positive radial solution u of (9), we have
Z
Z
2

u dx > c(n)
u2 dx;
B

where the constant c(n) is in particular independent of u. As u > 0, we have necessarily


> c(n).

Proof of Theorem 3.2. Assume by contradiction that u is a nontrivial radial solution of
(13). According to the remarks in Section 3.2 on the nonexistence of positive solutions, we
may assume that u is sign changing. Further, u cannot have infinitely many oscillations
d
near B, because in this case, as u C 4 (B) we would also have dr
u = 0 on B and
n
could extend u by 0 to the whole of R as a solution of the differential equation. The
unique continuation property [31] would give the desired contradiction (see also (12)).
As one may replace u with u if necessary, we may thus assume that there exists a
number a (0, 1) such that
u(x) = 0 for |x| = a,

u(x) > 0 for a < |x| < 1;

moreover, by (11),
u(x) = u(x) = 2 u(x) = 0 for |x| = 1.

(43)
Clearly

d
dr u(a)

> 0. We define
f (x) := |u(x)|2 2 u(x)

and compare u with the solution v of


( 2
v(x) = f (x)

for a < |x| < 1,

v(x) = v(x) = 0 for |x| {a, 1}.


Obviously v > 0 on a < |x| < 1. We now make use of comparison results, which hold true
for radial solutions of biharmonic inequalities in the annulus {x : a < |x| < 1} and can
e.g. be found in [8, Section 2]. First one obtains
d2
v(1) > 0.
dr2

(44)

d
Next, as u and v satisfy the same differential equation, if dr
u(a) = 0 it follows that u v
d
d
while if dr u(a) > 0 = dr v(a) we find u > v in a < |x| < 1. In both cases (44) gives

d2
u(1) > 0
dr2
and hence a contradiction with (43).

n
Proof of Theorem 4.2. We keep p [1, n4
) fixed as in the theorem. Let B = B1 (0)
denote the unit ball, centered at the origin. By scaling and Talentis comparison principle
[35, Theorem 1] it suffices to prove that there exists CB > 0 with

(45)

kuk22 > Skuk22 +

1
kuk2p
CB

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

17

for each u H 2 H01 (B), which is radially symmetric and radially decreasing with respect
to the origin (and hence positive). Let us set
S,p =

inf

1 (B)
uH 2 H0

u radially decreasing

kuk22 kuk2p
.
kuk22

If by contradiction (45) does not hold, one gets


> 0 : S,p < S.

(46)
Let us fix some (0, 1,p ), where
(47)

1,p :=

kuk22
2
uH 2 H01 (B)\{0} kukp
inf

is the first positive eigenvalue of


(
p2 u
2 u = kuk2p
p |u|
u = u = 0

in B
on B.

Arguing along the lines of [6] and [13], one finds a smooth positive radial strictly decreasing
solution of
(
p2 u in B
2 u = u(n+4)/(n4) + kuk2p
p |u|
(48)
u = u = 0
on B.
On the other hand, for any positive radially decreasing solution of (48) one has the following variant of Pohozaevs identity:



Z  
n(p 2)
u

(49)
2
kuk2p =
(u) dS.
2p
r
r
B
As above in the proof of Theorem 3.1, we conclude that
(50)

kuk2p > c(n, p)k2 uk21 > c(n, p) kuk2p ,

as we have assumed that p < n/(n 4). For > 0 sufficiently close to 0, we obtain a
contradiction.

8. Appendix: Proof of Lemma 5.4
We prove the corresponding statement for the free functional
Z
Z
1
1
E (u) =
|u|2 dx
|u|2 dx
2
2
which is defined on the whole space H 2 H01 (). More precisely, we have
Lemma 8.1. Let (uh ) H 2 H01 () be a PalaisSmale sequence for E at level c R.
Then either (uh ) is relatively compact in H 2 H01 () or there exist k > 0 nonzero functions
u
bj D2,2 (0,j ), j = 1, . . . , k, with 0,j either the whole Rn or a half space, solving either
(18) or (21) and a solution u
b0 H 2 H01 () of (5) such that, as h :
uh * u
b0 in H 2 H01 (),
kuh k22 kb
u0 k22 +

k
X
j=1

kb
uj k22 ,

E (uh ) E (b
u0 ) +

k
X
j=1

E0,j (b
uj ).

18

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

Step I. Reduction to the case uh * 0.


By well known arguments [6] we know that there exists u H 2 H01 () such that
0 (u) = 0, namely u solves (5). Moreover, by the
uh * u (up to a subsequence) and E
BrezisLieb lemma we infer that (uh u) is again a PalaisSmale sequence for E. Therefore
we may assume that
(51)

uh * 0,

E(uh ) c >

2 n/4
S .
n

Indeed, if c < n2 S n/4 , then we are in the compactness range of E and by arguing as in [6],
uh 0 up to a subsequence. In this case the statement follows with k = 0.
Assuming (51) and arguing as for the proof of [33, (3.2), p.187] we deduce
Z
(52)
|uh |2 dx > S n/4 + o(1).

Let L N be such that B2 (0) is covered by L balls of radius 1. By continuity of the maps
Z
Z
2
R 7 sup
|uh | dx,
y
7
|uh |2 dx
y

BR1 (y)

BR1 (y)

and (52), for h large enough one finds Rh > 2/diam() and xh such that
Z
Z
1 n/4
2
(53)
|uh | dx = sup
|uh |2 dx =
S .
2L
y B 1 (y)
B 1 (xh )
R

When passing to a suitable subsequence, three cases may now occur:


Case I. Rh % + and (Rh d(xh , )) is bounded;
Case II. (Rh d(xh , )) +;
Case III. (Rh ) is bounded.
Step II. Preliminaries for Case I.
For every x Rn let us denote by x0 its projection onto Rn1 , so that x = (x0 , xn ).
Since d(xh , ) 0, up to a subsequence xh x0 and %h := Rh d(xh , ) %.
Moreover, for sufficiently large h (say h > h) there exists a unique y h such that
d(xh , ) = |y h xh |.
For all h > h, up to a rotation and a translation, we may assume that y h = 0 and that
the tangent hyperplane H to at 0 has equation xn = 0 so that xh y h H. Then,
by the smoothness of , there exist an (n 1)dimensional ball B (0) of radius > 0
(independent of h) and smooth maps h : B (0) R such that in local orthonormal
coordinate systems over the tangent hyperplanes at yh , we have:
 

(54)
B (0) [, ] = (x0 , h (x0 )) : x0 B (0) .
Furthermore, there exists C > 0 (independent of h) such that
(55)

|h (x0 )| 6 C|x0 |2 ,

|h (x0 )| 6 C|x0 |,

|D2 h (x0 )| 6 C

We now rescale the domain by setting



h = Rh xh ,

x0 B (0).

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

19

so that xh is mapped into the origin 0 while the origin is mapped into (0, %h ) and (54)
becomes

h BRh (0) [Rh %h , Rh %h ] =



 0
x
0
0
%h : x BRh (0) .
x , R h h
Rh
Let us set
(
h =

Rh h

x0
Rh

(x0 , x) Rn : x0 BRh (0),


%h < xn < Rh h

x0
Rh


%h + R h

where > 0 and sufficiently small to have h h . Consider the injective map
n
h : BRh (0) [Rh , Rh ] R

 0

(x0 , xn ) 7
x0 , Rh h Rxh + xn %h



and its inverse 1
.
We
observe
that

B
(0)

[0,
R
]
= h and that h is bijective
R

h
h
h
h
on these sets. Thanks to (55), after some computations, one sees that (h ) converges in
2 ({x > 0}) to the translation by the vector (0, %). Therefore,
Cloc
n


0 = (x0 , xn ) Rn : xn > %
is the local uniform limit of (h ). In particular, for every Cc (0 ) we also have that
Cc (h ) for sufficiently large h. This will be used below. Let us now set


4n
x
(56)
vh (x) = Rh 2 uh xh +
,
Rh
so that vh H 2 H01 (h ). By boundedness of (uh ) we infer that there exists C > 0 such
that kvh kH 2 (h ) 6 C. Let 1h denote the characteristic function of h , then the sequence
(1h vh ) is bounded in D1,2n/(n2) (Rn ) (and in L2 (Rn )) so that, up to a subsequence, we
have
(57)

1h vh * v0

in D1,2n/(n2) L2 (Rn )

where supp(v0 ) 0 and v0 |xn =% = 0. Moreover, as (1h D2 vh ) is bounded in L2 (Rn ),


by weak continuity of distributional derivatives, we deduce
Z
Z
2
2
Dij vh dx
Dij
v0 dx
h

for all Cc (0 ) and i, j = 1, . . . , n. In particular, v0 has in 0 second order weak


derivatives.
Step III. The limiting function v0 in (57) solves (5) in 0 .
Fix Cc (0 ), then for h large enough we have supp() h . Define h Cc ()
by setting


n4
h (x) = Rh 2 Rh (x xh ) .

20

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

Therefore, (D2 h ) being bounded in L2 () and taking into account (56), one obtains
Z


n
0
2
uh Rh (x xh ) dx
o(1) =E (uh )(h ) = Rh

Z


n4
Rh 2
|uh |2 2 uh Rh (x xh ) dx

Z
Z
=
vh dx
|vh |2 2 vh dx
n
n
ZR
ZR
=
v0 dx
|v0 |2 2 v0 dx + o(1).
Rn

Rn

Then v0 H01 H 2 (0 ) solves (5) in distributional sense. The delicate point is to see that
v0 = 0 on 0 . To this end, let Cc2 (Rn ) with = 0 on 0 and define

h (x) = 1
h (x) (0, %) .
Notice that for h large enough, also h Cc2 (Rn ) with h = 0 on h . Taking into
2
n
account that (1
h ) tends to the translation by (0, %), it results h in C (R ) and
Z
Z
(v0 |v0 |2 2 v0 ) dx = lim
(vh h |vh |2 2 vh h ) dx
h
0
h


 n4
Z 
n4
h
2 2
h
2
2
= lim
uh Rh h (Rh (x x )) dx
uh Rh h (Rh (x x )) |uh |
h

 n4

0
h
2
= lim E (uh ) Rh h (Rh (x x )) = 0.
h

Therefore, by extending any given H01 H 2 (0 ) oddly with respect to xn % as a


function of H 2 (Rn ) and then by approximating it by a sequence of C 2 functions (k ) with
k = 0 on 0 , we get
Z
Z
1
2
H0 H (0 ) :
v0 dx =
|v0 |2 2 v0 dx,
0

which, as pointed out in [4, p. 221] implies that v0 is also a strong solution of (5) in 0 .
Step IV. The limiting function v0 in (57) is nontrivial.
Let Cc (Rn ) such that 0 supp 6= ; then for h large enough, we may define

(58)
ve0,h (x) = v0 1
(x)

(0,
%)
.
ve0,h : h supp() R,
h
Since v0 C 2 (0 ), v0 = v0 = 0 on 0 and (h ) converges to the translation by the
vector (0, %), we get as h +
(59)

k10 v0 1h ve0,h kL2 (supp()) = o(1),


k10 v0 1h e
v0,h kL2 (supp()) = o(1),

k10 v0 1h ve0,h kL2 (supp()) = o(1)

k10 v0 1h e
v0,h kL2 (supp()) = o(1).

To symplify the notations, in what follows we omit the 1h in front of vh , vh , vh and 10


in front of v0 , v0 , v0 . Then by (59) and some computations, one obtains, as h +
Z
Z
2
|(v0 vh )| dx =
|(e
v0,h vh )|2 dx + o(1)
Rn

h supp()

!2/2
2

|(e
v0,h v0 )| dx

>S
h supp()

Z
+ o(1) > S

|(v0 vh )| dx
Rn

2/2
+ o(1).

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

21

By compact embedding one has vh v0 in L2loc (Rn ) and thanks to an integration by parts
one gets
k((vh v0 ))k2 6 k(vh v0 )k2 k((vh v0 ))k2
and therefore vh v0 in L2loc (Rn ). Hence the previous inequality yields
2/2
Z
Z
+ o(1),
2 |(v0 vh )|2 dx > S
||2 |v0 vh |2 dx
Rn

Rn

which implies
Z

Z

d > S

Rn

2/2

|| d

Rn

where d and d denote respectively the weak limits of |(v0 vh )|2 and |v0 vh |2 in
the sense of measures. By Lions principle, there exists an at most countable set J, points
2/2
(xj )jJ 0 , nonnegative real numbers (j ) and (j ) with j > Sj such that
Z
Z
X
2
(60)
|vh | dx
|v0 |2 dx +
j (xj )
h

|vh |2 dx

(61)

for every
obtain

Cc (Rn ).

|v0 |2 dx +

Let

0
E
(vh )(vh )
h

jJ

Cc (Rn )

Z
=

j (xj )

jJ

0 (v )(v ). Thanks to (56), we


and compute E
h
h
h

(|uh (x)|2 |uh (x)|2 )(Rh (x xh )) dx

+2
uh (x)uh (x) ((Rh (x xh ))) dx

Z
+
uh (x)uh (x)((Rh (x xh ))) dx.

0 (v )(v ) = E 0 (u )(
In particular, putting
eh = (Rh (x xh )), it results E
h
h
h eh uh ).
h
Then, taking into accont that kuh
eh kH 2 H01 = o(1), by (60) and (61) we get as h +
Z
Z
Z
0
2
2
o(1) = Eh (uh )(
eh uh ) =
|v0 | dx
|v0 | dx + 2
v0 v0 dx
0
0
0
Z
X
X
+
v0 v0 dx +
j (xj )
j (xj ) + o(1).
0

jJ

jJ

By arguing as in the proof of [4, Lemma 3.3] one then infers that the set J is finite.
Assume by contradiction that v0 0. Choose 0 < 6 12 such that
xj 6 B1+ (0) \ B1 (0),

j J.

Then, for any Cc (0 ) with supp() B1+ (0) \ B1 (0) it results, as h +


Z
Z
2
|vh | dx = o(1),
|vh |2 dx = o(1).
h B1+ (0)\B1 (0)

Let

Cc (Rn ),

h B1+ (0)\B1 (0)

0 6 6 1 with
(
1 on B1+/3 (0)
=
0 on Rn \ B1+2/3 (0)

22

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

and put h = vh . Then one has


Z
(62)

|h |2 dx = o(1).

h B1+2/3 (0)\B1+/3 (0)

With the same computations as in the proof of [1, Lemma 2], it follows
Z
 4 !Z
n4
n
|h |2 dx
|h |2 dx 6 o(1).
(63)
1 S n4
h

On the other hand, in view of (62) and (53), it results


Z
Z
Z
2
2
|h | dx =
|h | dx =
h

h B1+2/3 (0)

|vh |2 dx + o(1) =

6
B3/2 (0)

|uh |2 dx + o(1)
3/2 R1
h

2R1
h

(xh )

(xh )

6
B

h B1+/3 (0)

|h |2 dx + o(1)

|uh | dx + o(1) 6 L
B

1
|uh |2 dx + o(1) = S n/4 + o(1).
2
h
1 (x )
h

Therefore, by (63) we infer that, as h +


Z
|h |2 dx = o(1).
h

But then we immediately get, again exploiting (53)


Z
Z
2
o(1) =
|vh | dx =
h B1+/3 (0)

|uh |2 dx
(1+/3)R1
h

|uh |2 dx =

>
B

R1
h

(xh )

(xh )

1 n/4
S ,
2L

a contradiction. Then it has to be v0 6= 0.


Step V. In Case II, a solution of (18) appears.
This follows by arguing as in [1], see also [33]. After rescaling as in (56), (vh ) converges
to a solution of (18).
Step VI. Case III cannot occur.
By contradiction, assume that (Rh ) is bounded. Then, being Rh > 2/diam(), we may
assume that, up to a subsequence,
xh x0 ,

Rh R0 > 0.

Let us set 0 = R0 ( x0 ) and


4n
2

vh = R0


uh

Therefore, vh H 2 H01 (0 ) and


Z
Z
|vh |2 dx =
|uh |2 dx,
0

H2

x
x0 +
R0
Z


.

|vh |2 dx =

0
H01 (0 ).

|uh |2 dx

so that, up to a subsequence, vh * v0

Moreover, it is readily seen that


(vh ) is a PalaisSmale sequence for the functional
Z
Z
1
1
2
E0 (u) =
|u| dx
|u|2 dx
2 0
2 0

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

23

at the same level c. By weak continuity of E00 we get E00 (v0 ) = 0, namely v0 solves (5) in
0 . Arguing as in Step IV, we infer that v0 6= 0 which is absurd since uh * 0.
Step VII. Conclusion.
If (uh ) is a PalaisSmale sequence for E , then by Step I its weak limit u
b0 solves (5).
By Steps III, IV and V the remaining part (uh u
b0 ) suitably rescaled gives rise to a
nontrivial solution v0 of (18) (if Case II occurs) or (21) (if Case I occurs). With help of
this solution, we construct from (uh u
b0 ) a new PalaisSmale sequence (wh ) for E in
H 2 H01 () at a strictly lower energy level. In the case where we find the solution v0 in
the whole space, one proceeds precisely as in [33], cf. also [1]. In the case that v0 is a
solution of (21) in a half space 0 we again have to use the locally deformed versions ve0,h
in h of v0 . These have been defined in (58). Let Cc (Rn ) be any cutoff function
with 0 6 6 1, = 1 in B1 (0) and = 0 outside B2 (0). We put

 p

(n4)/2
wh (x) := (uh u
b0 )(x) Rh
ve0,h Rh (x xh )
Rh (x xh ) .

First we remark that wh is indeed well defined as = 0 for Rh |x xh | > 2 Rh and as the
domain of definition of ve0,h grows at rate Rh . Further we notice that h and 1
h converge
uniformly to translations even on BRh (0). For this reason we have



ve0,h ()
v0 in D2,2 (0 )
Rh
and, of course, also in D1,2n/(n2) (0 ) and in L2 (0 ). Hence, we have
E (wh ) = E (uh ) E (b
u0 ) E0 (v0 ) + o(1)


0 (w ) 0 strongly in H 2 H 1 () . Now, the same procedure from Steps II to


and E
h
0
VI is applied to (wh ) instead of (uh u
b0 ). As v0 6 0 and hence
2
E0 (v0 ) > S n/4 ,
n
this procedure stops after finitely many iterations.


Acknowledgement. The authors are grateful to Nicola Garofalo for raising their attention
on the paper [19].
References
` Positive solutions of a fourth-order semilinear problem involving critical
[1] C.O. Alves, J.M. do O,
growth, Preprint, 2001.
[2] A. Bahri, J.M. Coron, On a nonlinear elliptic equation involving the critical Sobolev exponent: the
effect of the topology of the domain, Comm. Pure Appl. Math. 41 (1988), 253294.
[3] T. Bartsch, T. Weth, M. Willem, A Sobolev inequality with remainder term and critical equations
on domains with nontrivial topology for the polyharmonic operator, Preprint, 2002.
[4] F. Bernis, J. Garca Azorero, I. Peral, Existence and multiplicity of nontrivial solutions in
semilinear critical problems of fourth order, Adv. Differential Equations 1 (1996), 219240.
[5] T. Boggio, Sulle funzioni di Green dordine m, Rend. Circ. Mat. Palermo 20 (1905), 97135.
[6] H. Brezis, L. Nirenberg, Positive solutions of nonlinear elliptic equations involving critical Sobolev
exponents, Comm. Pure Appl. Math. 36 (1983), 437477.
[7] C.V. Coffman, C.L. Grover, Obtuse cones in Hilbert spaces and applications to partial differential
equations, J. Funct. Anal. 35 (1980), 369396.
[8] R. Dalmasso, Elliptic equations of order 2m in annular domains, Trans. Amer. Math. Soc. 347
(1995), 35753585.
[9] F. Ebobisse, M.O. Ahmedou, On a nonlinear fourth order elliptic equation involving the critical
Sobolev exponent, Preprint, 2001.

24

FILIPPO GAZZOLA, HANSCHRISTOPH GRUNAU, MARCO SQUASSINA

[10] D.E. Edmunds, D. Fortunato, E. Jannelli, Critical exponents, critical dimensions and the biharmonic operator, Arch. Ration. Mech. Anal. 112 (1990), 269289.
[11] F. Gazzola, Critical growth problems for polyharmonic operators, Proc. Royal Soc. Edinburgh Sect.
A 128 (1998), 251263.
[12] F. Gazzola, H.Ch. Grunau, Critical dimensions and higher order Sobolev inequalities with remainder terms, NoDEA Nonlinear Differential Equations Appl. 8 (2001), 3544.
[13] H.Ch. Grunau, Positive solutions to semilinear polyharmonic Dirichlet problems involving critical
Sobolev exponents, Calc. Var. Partial Differential Equations 3 (1995), 243252.
[14] H.Ch. Grunau, Polyharmonische Dirichletprobleme: Positivit
at, kritische Exponenten und kritische
Dimensionen, Habilitationsschrift, Universit
at Bayreuth, (1996).
[15] H.Ch. Grunau, On a conjecture of P. Pucci and J. Serrin, Analysis 16 (1996), 399403.
[16] H.Ch. Grunau, G. Sweers, Positivity for perturbations of polyharmonic operators with Dirichlet
boundary conditions in two dimensions, Math. Nachr. 179 (1996), 89-102.
[17] E. Jannelli, The role played by space dimension in elliptic critical problems, J. Differential Equations
156 (1999), 407426.
[18] B. Kawohl, G. Sweers, On a conjecture of McKenna and Walter and related anti-eigenvalues,
preprint 2001.
[19] C. Kenig, Restriction theorems, Carleman estimates, uniform Sobolev inequalities and unique continuation, Harmonic analysis and partial differential equations, (El Escorial, 1987), 6990, Lect. Notes
Math. 1384, Springer, (1989).
[20] C.S. Lin, A classification of solutions of a conformally invariant fourth order equation in Rn , Comment.
Math. Helv. 73 (1998), 206231.
[21] P.L. Lions, The concentrationcompactness principle in the calculus of variations. The limit case. I,
Rev. Mat. Iberoamericana 1 (1985), 145201.
[22] S. Luckhaus, Existence and regularity of weak solutions to the Dirichlet problem for semilinear
elliptic systems of higher order J. Reine Angew. Math. 306 (1979), 192207.
[23] E. Mitidieri, On the definition of critical dimension, unpublished manuscript, (1993).
[24] E. Mitidieri, A Rellich type identity and applications, Comm. Partial Differential Equations 18
(1993), 125151.
[25] J.J. Moreau, Decomposition orthogonale dun espace hilbertien selon deux c
ones mutuellement polaires, C. R. Acad. Sci. Paris 255 (1962), 238240.
[26] P. Oswald, On a priori estimates for positive solutions of a semilinear biharmonic equation in a ball,
Comment. Math. Univ. Carolinae 26 (1985), 565577.
[27] D. Passaseo, The effect of the domain shape on the existence of positive solutions of the equation

u + u2 1 = 0, Topol. Meth. Nonlinear Anal. 3 (1994), 2754.


[28] S.I. Poho
zaev, Eigenfunctions of the equation u + f (u) = 0, Soviet. Math. Dokl. 6 (1965), 1408
1411.
[29] P. Pucci, J. Serrin, A general variational identity, Indiana Univ. Math. J. 35 (1986), 681703.
[30] P. Pucci, J. Serrin, Critical exponents and critical dimensions for polyharmonic operators, J. Math.
Pures Appl. 69 (1990), 5583.
[31] R.N. Pederson, On the unique continuation theorem for certain second and fourth order elliptic
equations, Comm. Pure Appl. Math. 11 (1958), 6780.
der, Operator inequalities, Math. Science Engineering 147, Academic Press, New York,
[32] J. Schro
(1980).
[33] M. Struwe, A global compactness result for elliptic boundary value problems involving limiting
nonlinearities, Math Z. 187 (1984), 511517.
[34] C.A. Swanson, The best Sobolev constant, Applicable Anal. 47 (1992), 227239.
[35] G. Talenti, Elliptic equations and rearrangements, Ann. Scuola Norm. Sup. Pisa Cl. Sci. 3 (1976),
697718.
[36] W.C. Troy, Symmetry properties in systems of semilinear elliptic equations, J. Differential Equations
42 (1981), 400413.
[37] R.C.A.M. van der Vorst, Variational identities and applications to differential systems, Arch.
Ration. Mech. Anal. 116 (1991), 375398.
[38] R.C.A.M. van der Vorst, Best constant for the embedding of the space H 2 H01 () into
L2N/(N 4) (), Differential Integral Equations 6 (1993), 259276.
[39] R.C.A.M. van der Vorst, Fourth order elliptic equations with critical growth, C. R. Acad. Sci.
Paris, Serie I 320 (1995), 295299.

BIHARMONIC EQUATIONS AT CRITICAL GROWTH

Dipartimento di Scienze e Tecnologie Avanzate,


Corso Borsalino 54, I15100 Alessandria, Italy.
E-mail address: gazzola@unipmn.it
t fu
r Mathematik, Otto-von-Guericke Universita
t,
Fakulta
Postfach 4120, D39016 Magdeburg, Germany.
E-mail address: Hans-Christoph.Grunau@mathematik.uni-magdeburg.de
Dipartimento di Matematica e Fisica
Via Musei 41, I25121 Brescia, Italy.
E-mail address: squassin@dmf.bs.unicatt.it

25