J. Evol. Equ.

© 2020 Springer Nature Switzerland AG

Journal of Evolution
[Link] Equations

Asymptotics for a parabolic equation with critical exponential

nonlinearity

Michinori Ishiwata, Bernhard Ruf, Federica Sani and Elide Terraneo

Abstract. We consider the Cauchy problem:


∂t u = u − u + λ f (u) in (0, T ) × R2 ,
u(0, x) = u 0 (x) in R2 ,

where λ > 0,

2
f (u):=2α0 ueα0 u , for some α0 > 0,

with initial data u 0 ∈ H 1 (R2 ). The nonlinear term f has a critical growth at infinity in the energy
space H 1 (R2 ) in view of the Trudinger-Moser embedding. Our goal is to investigate from the initial data
u 0 ∈ H 1 (R2 ) whether the solution blows up in finite time or the solution is global in time. For 0 < λ < 2α1 ,
0
we prove that for initial data with energies below or equal to the ground state level, the dichotomy between
finite time blow-up and global existence can be determined by means of a potential well argument.

1. Introduction and main results

Model parabolic problem. We consider the Cauchy problem for a two-space

dimensional parabolic equation with exponential-type nonlinearity; more precisely,
we focus the attention on the following model problem:

∂t u = u − u + λ f (u) in (0, T ) × R2 ,
(1.1)
u(0, x) = u 0 (x) in R2 ,

where λ > 0,

f (u) := 2α0 ueα0 u , for some α0 > 0,

2

and we consider initial data in the energy space H 1 (R2 ), i.e.,

u 0 ∈ H 1 (R2 ).
 M. Ishiwata et al. J. Evol. Equ.

In this framework, energy refers to the functional associated with the stationary prob-
lem:

1
I (v) := v H 1 − λ
2
F(v) dx,
2 R2

where
 v
 1
f (η) dη = eα0 v − 1.
2
v H 1 := ∇v2L 2 + v2L 2 2 , and F(v) :=
0

The above functional is well defined in H 1 (R2 ), and the nonlinear term f that we are
considering has critical growth in the energy space in view of the Trudinger–Moser
embedding [1,32].
Concerning local existence and uniqueness for (1.1), Ibrahim, Jrad, Majdoub and
Saanouni [14] proved that, for any u 0 ∈ H 1 (R2 ), the Cauchy problem (1.1) has a local
in time solution
 
u ∈ C [0, T ]; H 1 (R2 )

for some finite time T > 0 (see Definition 2.1 and Remark 2.2), and the solution
is unique. Then, the smoothing effect of the heat kernel implies that u is a classical
solution; in fact, it belongs to the class
∞
     
u ∈ L loc (0, T ]; L ∞ (R2 ) ∩ C 1 (0, T ); L 2 (R2 ) ∩ C 1,2 (0, T ) × R2 ,

see [20, Remark 4.1].

We define the maximal existence time T∗ of the solution u as
  
T∗ := sup T > 0 : the problem (1.1) admits a solution u ∈ C [0, T ]; H 1 (R2 ) ∈ (0, +∞].

If T∗ < +∞, then the L ∞ -norm of the solution blows up, i.e.,

if T∗ < +∞ then lim sup u(t) L ∞ = +∞,

t→T∗

see, e.g., [5, Section 5.3]. In view of the definition of T∗ , it is natural to try to understand
whether T∗ < +∞ yields also the blow-up of the H 1 -norm of the solution. This
problem is related to the dependence of the local existence time of the solution to (1.1)
from the size of the initial data u 0 ∈ H 1 (R2 ); this aspect is emphasized in Sect. 2 in
comparison with the energy subcritical problem. For the energy subcritical problem,
the local existence time is uniform with respect to the H 1 -norm, while for the energy
critical Cauchy problem (1.1), we can find a uniform local existence time for small
initial data only, and we quantify the smallness condition in Theorem 2.6.
As a consequence of Theorem 2.6, we deduce that if the H 1 -norm of the solution u to
(1.1) is sufficiently small then u is a global solution, see Corollary 2.8. Indeed, our aim
is to find sufficient conditions in order to determine from the initial data u 0 ∈ H 1 (R2 )
 Asymptotics for a parabolic equation

whether the solution blows up in finite time (i.e., T∗ < +∞) or the solution is global
in time (i.e., T∗ = +∞). The same problem for nonlinear hyperbolic and parabolic
equations with polynomial nonlinearities has been widely studied via the potential
well argument starting from the seminal papers by Sattinger [35], Tsutsumi [39], Ishii
[21], and Payne and Sattinger [30]. Let us recall the central idea of this method in the
parabolic case following the presentation given in [31].
Polynomial case. Let  ⊂ R N , N ≥ 3, be a smooth bounded domain, and let us
consider
⎧
⎪
⎨∂t u = u + |u|
p−1 u in (0, T ) × ,
⎪
u(t, x) = 0 in (0, T ) × ∂, (1.2)
⎪
⎪
⎩u(0, x) = u (x) in ,
0

with 1 < p ≤ 2∗ − 1, and 2∗ = N2N −2 . For any initial data in the energy space
H0 (), there exists
1
 some finite time T > 0 and a local in time solution u belonging to
C [0, T ]; H01 () (this is a consequence of the L p+1 -existence result in [4] for any
1 < p ≤ 2∗ − 1 and of the smoothing effect of the heat kernel). In this framework,
the energy functional is given by
1 1 p+1
I p (v) := ∇v2L 2 − v L p+1 .
2 p+1
Let v ∈ H01 ()\{0}, and let us analyze the energy of the function σ v for any σ ≥ 0.
By an easy computation, one can show that
σ2 σ p+1 p+1
I p (σ v) = ∇v2L 2 − v L p+1
2 p+1
attains its unique maximum at a point σ̄ = σ̄ (v) > 0, and v̄ := σ̄ v satisfies
p+1
∇ v̄2L 2 − v̄ L p+1 = 0.

Therefore, the energy I (σ v) has the structure of a potential well, and every ray σ v,
for any σ > 0 and for v ∈ H01 ()\{0}, has a unique intersection with the Nehari
manifold
p+1
N = {v ∈ H01 ()\{0} : ∇v2L 2 − v L p+1 = 0}.

The depth of the well is given by the lowest pass over the ridge defined by all possible
maps σ → I p (σ v) as v ranges over H01 ()\{0}, namely

d p := inf max I p (σ v).

v∈H01 ()\{0} σ ≥0

It is well known that d p can be characterized as

p−1
d p = inf I p (v), and also dp = 2( p+1)/( p−1)
,
v∈N 2( p + 1)
 M. Ishiwata et al. J. Evol. Equ.

where = p+1 () is the best constant in the Sobolev embedding H01 () ⊂
L p+1 (), i.e.,
∇v L 2
= inf .
v∈H01 ()\{0} v L p+1

If 1 < p < 2∗ − 1, then d p is the energy level of ground state solutions.

The potential well associated with the Cauchy problem (1.2) is the set (stable set)
 
p+1
W p := v ∈ H01 () : I p (v) < d p , ∇v2L 2 − v L p+1 > 0 ∪ {0},

and the exterior of the potential well (unstable set) is

 
p+1
V p := v ∈ H01 () : I p (v) < d p , ∇v2L 2 − v L p+1 < 0 .

The sets V p and W p are both invariant under the flow associated with (1.2). Concerning
the stable set, if 1 < p < 2∗ − 1, any solution which enters the stable set W p exists
globally in time. This result is a direct consequence of the fact that, in the subcritical
case, the time T of local existence of the solution to (1.2) depends only on the size of
the norm of the initial data in H01 (), and for any v ∈ W p the Dirichlet norm ∇v L 2
is uniformly bounded (see [39]). Similar results have also been proved for p = 2∗ − 1,
where the situation is different because the local existence time of the solution to (1.2)
depends on the specific initial data rather than its size (see [19,21,22], and [23,38]).
On the other side, if 1 < p ≤ 2∗ − 1, then any solution which intersects the unstable
set V p blows up in finite time (see [30] and [21]). Related studies can be found in
[6,18,28,29]. For the case p = 2∗ − 1 and  = R N , N ≥ 3, we refer to [17] and to
the recent result [8] in which the authors completely describe the dynamics near the
ground state.
Related results on the asymptotic behavior of solutions for the Cauchy problem

∂t u = u − u + |u| p−1 u in (0, T ) × R N ,
u(0, x) = u 0 (x) in R N ,

when N ≥ 3 and for the subcritical 1 < p < 2∗ − 1 case can be found in [7,9,11,12]
and references therein. Our forthcoming work [24] also contains results for the problem
above with subcritical and critical nonlinearity.
When N = 2, any power nonlinearity is allowed, and the critical nonlinearity
seems to be of exponential type as in (1.1). In the same spirit of the previous results,
we show that for the Cauchy problem (1.1) if the energy is below the ground state
level the dichotomy between blow-up and global existence is determined by means of
a potential well argument.
The stationary problem. It is not difficult to show that the stationary problem asso-
ciated with (1.1), i.e.,

− v + v = λ f (v) in R2 , (1.3)
 Asymptotics for a parabolic equation

has no non-trivial H 1 -solution if λ ≥ 1

2α0 . Therefore, from now on, we will assume

1
0<λ< . (1.4)
2α0
The existence of ground state solutions for (1.3) with λ in the range (1.4) is proved in
[33]. From [33], we also know that the mountain pass level
       
c := inf sup I γ (s) , := γ ∈ C [0, 1]; H 1 (R2 ) : γ (0) = 0, I γ (1) < 0 .
γ ∈ s∈[0,1]

(1.5)

coincides with the ground state energy level, and ground state solutions can be char-
acterized as minimizers of I on a suitable constraint, i.e.,

1
c = inf I (v) : v ∈ H (R )\{0}, v L 2 − λ
1 2 2
F(v) dx = 0 . (1.6)
2 R2

Moreover,
2π
0<c< . (1.7)
α0
Another useful characterization of the mountain pass level c can be obtained by
means of the Nehari functional

J (v) := d I (v), v = v2H 1 − λ v f (v) dx. (1.8)
R2

Let
 
d := inf I (v) : v ∈ H 1 (R2 )\{0}, J (v) = 0 ,

then the existence of a mountain pass solution v ∈ H 1 (R2 )\{0} to (1.3) implies
I (v) = c and d I (v) ≡ 0; therefore, d ≤ I (v) = c. The opposite inequality also
holds; hence,

c = d, (1.9)

and this can be deduced from the geometry of J and I in the energy space. In particular,
(1.9) is a consequence of the following property which gives also the potential well
structure of the energy functional I .
Proposition 1.1. Assume that λ is as in (1.4). For any v ∈ H 1 (R2 )\{0}, there exists
a unique σ = σ (v) > 0 such that
⎧
⎪
⎨> 0 if 0 < σ < σ ,
⎪
J (σ v) = 0 if σ = σ , (1.10)
⎪
⎪
⎩< 0 if σ > σ .
 M. Ishiwata et al. J. Evol. Equ.

Moreover,
lim I (σ v) = −∞, (1.11)
σ →+∞

and σ is the unique maximum point of the map σ → I (σ v) on [0, +∞).

The proof of Proposition 1.1 follows by simple computations (see also Lemma 7.4
and Lemma 7.5 with a = 1 and b = 0). From Proposition 1.1, it is easy to deduce that
c ≤ d by comparison with the auxiliary level
c̃ = inf sup I (σ v),
v∈H 1 (R2 )\{0} σ >0

see Proposition 7.1 with a = 1 and b = 0.

Stable and unstable sets. In view of Proposition 1.1, for any fixed v ∈ H 1 (R2 )\{0},
the function σ → I (σ v) has the shape of a potential well. The idea of the potential
well method is to trap the solution to (1.1) in the well to the left of σ (v) in order to
guarantee global existence. To ensure that the solution to (1.1) is trapped, we have
to find the lowest pass over the ridge defined by all possible I (σ v) as v ranges over
H 1 (R2 )\{0}. The height of the lower pass over the ridge is the mountain pass level c̃
and c̃ = d.
Therefore, the potential well argument suggests to consider the splitting of the d-
sublevel set of the energy I determined by the Nehari functional J . More precisely,
we consider the unstable set V and the stable set W defined, respectively, by
 
V := v ∈ H 1 (R2 ) : I (v) < d, J (v) < 0 ,

and
 
W := v ∈ H 1 (R2 ) : I (v) < d, J (v) ≥ 0
 
= v ∈ H 1 (R2 ) : I (v) < d, J (v) > 0 ∪ {0}.
 
Theorem 1.2. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with λ
as in (1.4), and u 0 ∈ H 1 (R2 ).
(i) If u(t0 ) ∈ V for some t0 ∈ [0, T∗ ), then T∗ < +∞.
(ii) There exists t0 ∈ [0, T∗ ) such that u(t0 ) ∈ W if and only if
T∗ = +∞, and lim u(t) H 1 = 0. (1.12)
t→+∞

The first part of Theorem 1.2 complements the blow-up result obtained in [14] for
non-positive energies.
 
Theorem 1.3. ([14, Theorem 2.1.3]). Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal
solution to (1.1) with
1
0<λ≤ , (1.13)
2α0
 
and u 0 ∈ H 1 (R2 ). If I u(t0 ) ≤ 0 and u(t0 ) = 0 for some t0 ∈ [0, T∗ ) then T∗ < +∞.
 Asymptotics for a parabolic equation

Theorem 1.3 is proved in [14] in the particular case α0 = 1 and λ = 21 , but the
arguments of the proof in [14] can be adapted to cover the general case with α0 > 0
and λ in the range (1.13), see Remark 4.4.
Up to our knowledge, Theorem 1.2 is a new application of the potential well argu-
ment to heat equations with critical exponential nonlinearities in the two-space dimen-
sional case. The same problem with subcritical exponential nonlinearities is studied
in [10] and [34].
It is important to mention that similar results for dispersive equations are already
available in the literature, for example, see [2] and [25] for the subcritical exponential
case and see [16] for the critical exponential case.
Differently from the dispersive framework, the energy associated with heat equa-
tions decreases along solutions, and this monotonicity property enables us to easily
determine the dichotomy between blow-up and global existence also at the ground
state energy level d.
 
Theorem 1.4. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with λ
 
as in (1.4), and u 0 ∈ H 1 (R2 ). Assume that I u(t0 ) = d for some t0 ∈ [0, T∗ ).
 
(i) If J u(t0 ) < 0 then u(t) ∈ V for any t ∈ (t0 , T∗ ).
(ii) If J  u(t0 ) > 0 then u(t) ∈ W for any t ∈ (t0 , T∗ ).
(iii) If J u(t0 ) = 0 then u(t0 ) is a stationary ground state solution, and u(t) = u(t0 )
for any t ∈ [t0 , +∞).
Outline of the paper. In Sect. 2, we discuss the dependence of the local existence
time of the solution to (1.1) from the H 1 -norm of the initial data, and we obtain a
sufficient condition for global existence (see Corollary 2.8).
In Sect. 3, we collect some basic properties of the solution to (1.1) which will be
crucial to prove the instability of the set V and the stability of the set W .
Section 4 is devoted to the study of the unstable set V and more precisely to the
proof of Theorem 1.2(i). The proof is based on the classical concavity method due to
Levine [26] (see Lemma 4.2) which applies to (1.1) due to the fact that the Nehari
functional J along solutions entering V is—not only negative but—bounded away
from zero by a negative constant (see Proposition 4.3).
Section 5 is devoted to the study of the stable set W , and more precisely to the proof
of Theorem 1.2(ii). This second part of the statement of Theorem 1.2 is more accurate
with respect to the first part concerning the instability in V ; in fact Theorem 1.2(ii)
gives a characterization of W in terms of the necessary and sufficient condition (1.12).
The proof of the stability of the set W is mainly based on Corollary 2.8. In order
to show that solutions entering W satisfy the assumptions of Corollary 2.8, it was
important to realize that the following inclusion holds:
 
W ⊆ v ∈ H 1 (R2 ) : I (v) < d, P(v) ≥ 0 , (1.14)

where

1
P(v) := v2L 2 − λ F(v) dx (1.15)
2 R2
 M. Ishiwata et al. J. Evol. Equ.

is the so-called Pohozaev functional, i.e., the functional appearing in the characteri-
zation of the ground state energy level (1.6), as developed in [33]. The validity of the
inclusion (1.14) is the idea underlying the argument of the proof of Proposition 5.2.
The positivity of the Nehari functional J near the origin of H 1 (R2 ) (see Theo-
rem 5.4) is crucial to show that (1.12) is a sufficient condition for the solution to (1.1)
to enter W .
To show that the H 1 -norm of solutions entering W must decay to zero as time
tends to infinity, we need a compactness result, see Proposition 5.6. In the proof of
Proposition 5.6, we will consider the following auxiliary growth functions

f˜(u) := 2α0 u(eα0 u − 1), F̃(u) := eα0 u − 1 − α0 u 2 ,

2 2
and (1.16)

satisfying

f (u) = f˜(u) + 2α0 u, and F(u) = F̃(u) + α0 u 2 .

Both u f˜(u) and F̃(u) are critical in the energy space with respect to the Trudinger–
Moser inequality [1,32], but these auxiliary growth functions are not affected by the
lack of compactness at spatial infinity (i.e., the lack of compactness of the embedding
H 1 (R2 ) → L 2 (R2 )), in fact

u f˜(u) F̃(u)
lim = lim = 0.
u→0 u 2 u→0 u 2

The description of the asymptotics at the ground state energy level given by Theo-
rem 1.4 is developed in Sect. 6.
The validity of (1.14) may raise questions about the role of the splitting of the d-
sublevel set of the energy I determined by the sign of the Pohozaev functional P with
respect to the flow associated with the Cauchy problem (1.1). Indeed, the Pohozaev
functional P and the Nehari functional J determine the same splitting below the
ground state energy level d, as already observed in [16], see also [25]. In [16], a
slightly different critical exponential nonlinearity is considered, and in Sect. 7, we
show that the argument in [16] can be adapted to the energy functional associated with
(1.1).

2. Uniform local existence time and blow-up alternative

Let  ⊆ R2 be any smooth domain, and let us consider the more general Cauchy
problem
⎧
⎪
⎨∂t u = u + g(u) in (0, T ) × ,
⎪
u(t, x) = 0 on (0, T ) × ∂, (2.1)
⎪
⎪
⎩u(0, x) = u (x) in ,
0

where u 0 ∈ H01 (), and g ∈ C 1 (R, R) satisfies

 Asymptotics for a parabolic equation

(g1 ) g(0) = 0, and

(g2 ) there exists α0 > 0 such that for any ε > 0 we have
 2 
|g(s1 ) − g(s2 )| ≤ Cε |s1 − s2 | eα0 (1+ε)s1 + eα0 (1+ε)s2 , s1 , s2 ∈ R,
2

for some positive constant Cε .

Under the above assumptions on the nonlinear term g, the Cauchy problem (2.1)
includes the model problem (1.1) as a particular case. Note that, by assuming con-
dition (g2 ), we take into account nonlinear terms with square exponential growth at
infinity, which are critical in the energy space. For any initial data in H01 (), the
argument introduced in [14] gives the  local existence and uniqueness of the solution
u to (2.1) in the class of functions C [0, T ]; H01 () , for some T > 0.

 u 0 ∈ H0 (). We say that u is a (mild) solution to (2.1) if u ∈

Definition 2.1. Let 1

C [0, T ]; H0 () , and u verifies the integral equation
1

 t
u(t) = et u 0 + e(t−s) g(u(s))ds.
0

Remark 2.2. As proved in [20, Proposition 4.1] and [13], u is a (mild) solution to (2.1)
if and only if u satisfies

∂t u = u + g(u)

in the sense of distributions.

Combining the arguments of [14] with [20, Remark 4.1], we have the following
result.

Theorem 2.3. ([14,20]). Let g ∈ C(R, R) satisfy (g1 ) and (g2 ), and assume u 0 ∈
H01 (). There exist T = T (u 0 ) > 0 and a unique solution u ∈ C [0, T ]; H01 () to
(2.1). Moreover,
∞
 
u ∈ L loc (0, T ]; L ∞ (R2 ) .

Let us introduce the maximal existence time of the solution u to (2.1) as

  
T∗ := sup T > 0 : the problem (2.1) admits a solution u ∈ C [0, T ]; H01 () ∈ (0, +∞].
(2.2)

Under the assumptions of Theorem 2.3, if the maximal existence time defined by (2.2)
satisfies T∗ < +∞, then

lim sup u(t) L ∞ = +∞,

t→T∗

see, e.g., [5, Section 5.3].

 M. Ishiwata et al. J. Evol. Equ.

In view of the definition of T∗ , the following question arises:

Does T∗ < +∞ imply lim sup u(t) H 1 = +∞? (2.3)

t→T∗

The above question remains open: the critical exponential nonlinearity that we con-
sider could have an effect on the blow-up alternative (2.3), and in analogy with the
critical polynomial case (see [36]), our guess  is that there could
 exist initial data in
H01 () for which T∗ < +∞ even if u ∈ L ∞ [0, T∗ ), H01 () .
As mentioned in Sect. 1, the above question about the blow-up alternative (2.3) is
related to the dependence of the local existence time of the solution to (1.1) from the
size of the initial data in H01 (): if one could find a local existence time T > 0 which
is uniform with respect to the H 1 -norm of the initial data, i.e., T = T (u 0  H 1 ), then
the blow-up alternative (2.3) would hold.
To explain this point of view, let us compare the energy critical problem with
the subcritical and supercritical cases. To take into account nonlinear terms with
subcritical or supercritical growth in the energy space, it is enough to replace (g2 ),
respectively, with:
(g2 )sub for any α0 > 0 there exists Cα0 > 0 such that
 2 
|g(s1 ) − g(s2 )| ≤ Cα0 |s1 − s2 | eα0 s1 + eα0 s2 , s1 , s2 ∈ R;
2

(g2 )sup there exists γ > 2 and α0 > 0 such that

|g(s)|
lim inf γ > 0.
s→+∞ eα0 s
The subcritical, critical or supercritical behavior of g affects the local existence time
of the solution to the Cauchy problem (2.1). In the supercritical case, we have a
non-existence result for (2.1).

Theorem 2.4. Let g ∈ C 1 (R, R) satisfy (g1 ) and (g2 )sup , and assume that g ≥ 0 on
R. There exists u 0 ∈ H 1 (R2 ), u 0 ≥ 0, such that for any
 T > 0 the Cauchy problem

∞ (0, T ), L ∞ (R2 ) .
(2.1) has no nonnegative solution in C [0, T ), H 1 (R2 ) ∩ L loc

Proof. Let γ > 2 be as in (g2 )sup , and define

⎧ 1  
⎨ 1 γ
log |x| 1
log log |x| |x| ≤ 1e ,
u 0 (x) :=
⎩
0 |x| > 1e .

Then, u 0 ∈ H01 (B1/e (0)), and arguing as in [20, Section 3] it is not difficult to deduce
non-existence. 

In the subcritical case, the solution to (2.1) exists up to some finite time which
depends only on the size of the initial data in H01 ().
 Asymptotics for a parabolic equation

Theorem 2.5. Let g ∈ C 1 (R, R) satisfy (g1 ) and (g2 )sub , and let M > 0. There exists
T = T (M) > 0 such that, for any u 0 ∈ H 1 (R2 ) with u H 1 ≤ M, the Cauchy
problem (2.1) has a unique solution u ∈ C [0, T ]; H01 () .

We omit the proof of Theorem 2.5, since it can be obtained by means of a standard fixed
point argument by exploiting the integral representation formula and the smoothing
effect of the heat kernel (see also the proof of Theorem 2.6). In the subcritical case, it
is clear that the blow-up alternative holds:

if T∗ < +∞ then lim sup u(t) H 1 = +∞. (2.4)

t→T∗

Indeed, if not we could extend the solution to (2.1) beyond the time T∗ < +∞, using
Theorem 2.5, and reach a contradiction.
In the critical case, from [14], we cannot deduce that the local existence time T > 0
is bounded away from zero by a positive constant depending only on the H 1 -norm of
the initial data, and we expect that the smallness of the local existence time T depends
on the specific initial data and not only on its size. Nevertheless, if we consider small
initial data, we can find a uniform local existence time for the solution to (2.1).

Theorem 2.6. Let g ∈ C 1 (R, R) satisfy (g1 ) and (g2 ). Let 0 < m < 4π
α0 , and M > 0.
There exists T = T (m, M) > 0 such that, for any u 0 ∈ H (R ) with
1 2

∇u 0 2L 2 ≤ m and u 0 2L 2 ≤ M, (2.5)

 
the Cauchy problem (1.1) has a unique solution u ∈ C [0, T ]; H01 () .

The smallness condition (2.5) with 0 < m < 4π α0 comes from the following scale
invariant form of the Trudinger–Moser inequality in H 1 (R2 ).

Theorem 2.7. ([1]). If α ∈ (0, 4π ), then there exists a constant Cα > 0 such that

(eαv − 1) dx ≤ Cα v2L 2 , for any v ∈ H 1 (R2 ) with ∇v L 2 ≤ 1, (2.6)
2

R2

and the above inequality fails if α ≥ 4π .

Proof
 of Theorem 2.6. In order to prove the existence of a unique solution u ∈
C [0, T ]; H0 () , let us first write the equation in (1.1) in the equivalent integral
1

formulation (see [20, Proposition 4.1] and [13] for a justification of this equivalence)
 t
u(t) = et u 0 + e(t−s) g(u(s))ds. (2.7)
0

Since 0 < m < 4π

α0 there exists ε ∈ (0, 1) such that m = α0 (1
4π
− ε). Let us consider
the set
 M. Ishiwata et al. J. Evol. Equ.

X = X (m, M)
 
  4π  ε
= u ∈ L ∞ (0, T ), H 1 (R2 ) : sup ∇u(t)2L 2 ≤ 1− ; sup u(t)2L 2 ≤ 2M .
t∈[0,T ] α0 2 t∈[0,T ]

This set endowed with the distance

d(u, w) = sup ∇u(t) − ∇w(t) L 2 + sup u(t) − w(t) L 2

t∈[0,T ] t∈[0,T ]

is a complete metric space. We show that if T > 0 is small enough the map
 t
(u)(t) = e u 0 +
t
e(t−s) g(u(s))ds
0

is a contraction from X into itself. We remark that u 0 satisfies

 
√ 4π 4π ε √
∇u 0  L 2 ≤ m = (1 − ε) < (1 − ), u 0  L 2 ≤ M.
α0 α0 2

Let us first prove that  maps X into itself. Indeed, thanks to property (g2 ) for any
u ∈ X (m, M) and for any t ∈ [0, T ] we obtain
 t
(u)(t) L 2 ≤ et u 0  L 2 + e(t−s) g(u(s)) L 2 ds
0
 t   
 (t−s) 
|u(s)| eα0 (1+ε)u (s) + 1  2 ds
2
≤ u 0 ] L 2 + Cε e
0 L
√  t
≤ M + 2Cε e(t−s) |u(s)| L 2 ds
0
 t   
 (t−s) 
|u(s)| eα0 (1+ε)u (s) − 1  2 ds
2
+ Cε e
0 L
√ √  t   
1  
eα0 (1+ε)u (s) − 1  r ds,
2
≤ M + 2Cε t 2M + Cε,r 1 1 |u(s)|
0 (t − s) r − 2 L

where 1 < r < 2 will be chosen later. If we could prove that there exists a constant
C = C(m, M) such that for any s ∈ [0, T ] we have
  r
 
u(s) eα0 (1+ε)u (s) − 1  r ≤ C
2
(2.8)
L

then we would obtain

√ √  t 1
(u)(t) L 2 ≤ M + 2Cε t 2M + C̃ε,r 1 1
ds
0 (t − s) r − 2
√ √
= M + 2Cε t 2M + C̃ε,r t 3/2−1/r .

Therefore, if T is sufficiently small depending only on m, M then

√
(u)(t) L 2 ≤ 2M.
 Asymptotics for a parabolic equation

The estimate (2.8) can be obtained via the scale invariant Trudinger–Moser inequality
(2.6). Indeed, for p, q > 1, 1p + q1 = 1 and provided that r p ≥ 2 we have
    
u(s) eα0 (1+ε)u (s) − 1 rL r ≤ |u(s)|r eα0 r (1+ε)u (s) − 1 dx
2 2

R2
 1/ p    1/q
α0 rq(1+ε)u 2 (s)
≤ |u(s)| dx
rp
e − 1 dx
R2 R2
   1/q
eα0 rq(1+ε)u
2 (s)
≤ Cu(s)rH 1 − 1 dx .
R2

4π(1−ε)
Now choosing q = 1 + ε2 , r = 1 + ε4 , and since α0 = m , we can estimate
    1/q
  1/q 4π(1−ε8 )

u(s)
√
2
α0 rq(1+ε)u 2 (s)
e − 1 dx = e m − 1 dx
R2 R2
 1/q  1/q
u(s)2L 2 2M
≤C ≤C .
m m
2
We remark that with this choice of q we obtain p = 1+ε
ε2
and r p ≥ 2. Therefore, we
obtain
  r
 
u(s) eα0 (1+ε)u (s) − 1  r ≤ C.
2

Next for any u ∈ X and for any t ∈ [0, T ], thanks to (2.8), we obtain
 t
∇(u)(t) L 2 ≤ et ∇u 0  L 2 + ∇e(t−s) g(u(s)) L 2 ds
0
 √
t   
1  
 t − s∇e(t−s) |u(s)| eα0 (1+ε)u (s) + 1  2 ds
2
≤ ∇u 0 ] L 2 + Cε √
0 t −s L
 t
√ 1 √
≤ m + 2Cε √  t − s∇e(t−s) |u(s)| L 2 ds
0 t −s
 t √   
1  
 t − s∇e(t−s) |u(s)| eα0 (1+ε)u (s) − 1  2 ds
2
+ Cε √
0 t − s L
 t   
√ √√ 1  
eα0 (1+ε)u (s) − 1  r ds
2
≤ m + 2Cε t 2M + Cε,r 1 |u(s)|
0 (t − s) r L

√ √√ t 1
≤ m + 2Cε t 2M + C̃ε,r 1
ds
0 (t − s) r
√ √√ 1
≤ m + 2Cε t 2M + C̃ε,r t 1− r ,

with the same 1 < r < 2 chosen above. Therefore, if T is sufficiently small depending
only on m, M then

4π  ε
sup ∇(u)(t) L 2 ≤ 1− .
t∈[0,T ] α0 2

In a similar way it is possible to prove that for T = T (m, M) small enough the
map  is a contraction on X . Finally, by using the standard regularizing properties
 M. Ishiwata et al. J. Evol. Equ.

of the heat kernel it is possible to prove that the fixed point u ∈ X of  satisfies
u ∈ C [0, T ], H 1 (R2 ) . 
From Theorem 2.6, we deduce a sufficient condition for global existence.
Corollary 2.8. Let g ∈ C 1 (R, R) satisfy (g1 ) and (g2 ), and let u 0 ∈ H01 (). If the
 
maximal solution u ∈ C [0, T∗ ); H 1 (R2 ) to (2.1) satisfies

4π
sup ∇u(t)2L 2 < , and sup u(t)2L 2 < +∞, for some t 0 ∈ [0, T∗ ),
t∈[t0 ,T∗ ) α0 t∈[t0 ,T∗ )

then u is global in time, i.e., T∗ = +∞.

Proof. As in the subcritical case, if we assume T∗ < +∞ then we can apply Theo-
rem 2.6 to extend the solution u beyond the maximal existence time T∗ , and reach a
contradiction. 

3. Basic properties of the solution to the model problem (1.1)

 
Let u 0 ∈ H 1 (R2 ). Let u ∈ C [0, T ]; H 1 (R2 ) be the local in time solution to (1.1)
 T = T (u 0 ) > 0is the local time of existence. We already pointed
found in [14], where
∞ (0, T ]; L ∞ (R2 ) , see [20, Remark 4.1]. Moreover, by using the
out that u ∈ L loc
integral formulation of the equation and the growth property of the nonlinearity it is
possible to prove (see also [5, Chapter 5]) that
   
u ∈ C (0, T ], L 2 (R2 ) , and u ∈ C 1 (0, T ], L 2 (R2 ) .

Then, by standard arguments, u is a classical solution for (1.1), i.e.,

 
u ∈ C 1,2 (0, T ) × R2 .

Proposition 3.1. For any t ∈ (0, T ), we have

d  
∂t u(t)2L 2 = − I u(t) , (3.1)
dt 
1 d  
u(t)2L 2 = −u(t)2H 1 + λ u(t) f u(t) dx, (3.2)
2 dt R2

and
 
|d I u(t) , ϕ| ≤ ∂t u(t) L 2 ϕ L 2 , for any ϕ ∈ H 1 (R2 ). (3.3)

Proof. The monotonicity of the energy (3.1) follows by multiplying the equation in
(1.1) by ∂t u, integrating over R2 , and applying density arguments as in [5, Lemma
5.4.5].      
Since u ∈ C [0, T ]; H 1 (R2 ) ∩C 1 (0, T ]; L 2 (R2 ) , and u ∈ C (0, T ], L 2 (R2 ) ,
(3.2) follows by multiplying the equation in (1.1) by u and integrating over R2 . Finally,
 Asymptotics for a parabolic equation

to deduce (3.3) we multiply the equation in (1.1) by ϕ ∈ H 1 (R2 ), and we integrate

over R2 , obtaining
 
 
∂t u(t)ϕ dx = − ∇u(t) · ∇ϕ + u(t)ϕ dx
R 2 R2

   
+λ f u(t) ϕ dx = −d I u(t) , ϕ.
R2


We complete this section with the following continuity result that can be proved
arguing as in [14, Proposition 3.6].
 
Lemma 3.2. If T > 0 and u ∈ C [0, T ], H 1 (R2 ) , then
   
F(u) ∈ C [0, T ], L 1 (R2 ) , and u f (u) ∈ C [0, T ], L 1 (R2 ) .
 
Hence, J (u) ∈ C [0, T ], R .

4. Blow-up in V

If v ∈ V , and σ = σ (v) > 0 is given by Proposition 1.1, then σ ∈ (0, 1), and
hence,

2d ≤ 2I (σ v) ≤ σ v2H 1 < v2H 1 . (4.1)

To prove the invariance of the set V under the flow associated with (1.1), it is crucial
to recall that, from (1.7) and (1.9), we know that

d > 0. (4.2)
 
Lemma 4.1. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with λ as
in (1.4), and u 0 ∈ H 1 (R2 ). If u(t0 ) ∈ V for some t0 ∈ [0, T∗ ), then u(t) ∈ V for any
t ∈ [t0 , T∗ ).
 
Proof. In view of the monotonicity
  of the energy (3.1), and since J(u) ∈ C [0, T∗ ); R ,
it is enough to prove that J u(t) = 0 for any t ∈ (t0 , T∗ ). If J u(t) = 0 for some
t ∈ (t0 , T∗ ), then there exists t1 ∈ (t0 , T∗ ) such that
   
J u(t) < 0 for any t ∈ [t0 , t1 ), and J u(t1 ) = 0.

Therefore, u(t) ∈ V for any t ∈ [t0 , t1 ), and

 
• either u(t1 ) = 0. Hence d ≤ I u(t1 ) , which is not possible due to the mono-
tonicity of the energy (3.1);
• or u(t1 ) = 0 which yields

lim u(t) H 1 = u(t1 ) H 1 = 0,

t→t1−
 M. Ishiwata et al. J. Evol. Equ.

and this contradicts (4.1) and (4.2).



In order to prove that solutions entering V blow up in finite time, we will apply the
following blow-up Lemma containing the classical idea of the concavity method due
to Levine [26].

Lemma 4.2. ([26]). There exists no non-negative and increasing function y ∈

C 2 (t, +∞), with t ∈ R, such that, for some β > 0,

y(t)y  (t) ≥ (β + 1)[y  (t)]2 on (t, +∞),

and

lim y(t) = +∞. (4.3)

t→+∞

Proof. For the sake of completeness, we briefly sketch the proof. By contradiction, we
assume that such a function y exists. In view of (4.3), h(t) := y −β (t) is well defined
on the half-line (t  , +∞), for some t  ≥ t sufficiently large. Moreover,

lim h(t) = 0. (4.4)

t→+∞

For any t > t  , we can compute

h  (t) = −β[y(t)]−β−1 y  (t) < 0,

and
 
h  (t) = β[y(t)]−β−2 (β + 1)[y  (t)]2 − y(t)y  (t) ≤ 0.

Therefore, h is concave and decreasing on (t  , +∞), and this contradicts (4.4). 

The concavity method works in our setting due to the fact the Nehari functional
along solutions entering V is bounded away from zero by a strictly negative constant.
 
Proposition 4.3. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with
λ as in (1.4), and u 0 ∈ H 1 (R2 ). If u(t0 ) ∈ V for some t0 ∈ [0, T∗ ) then there exists
ε > 0 such that J u(t) < −ε for any t ∈ [t0 , T∗ ).

Proof. Let
 
d  := inf H (v) : v ∈ H 1 (R2 )\{0}, J (v) ≤ 0 ,

where
  
1 1
H (v) := I (v) − J (v) = λ v f (v) − F(v) dx. (4.5)
2 R2 2
 Asymptotics for a parabolic equation

Then, d = d  . In fact, clearly d  ≤ d, and in order to deduce that d ≤ d  , it is enough

to show that

d ≤ H (v) for any v ∈ H 1 (R2 )\{0} with J (v) < 0.

Let v ∈ H 1 (R2 )\{0}, and let σ = σ (v) > 0 be as in Proposition 1.1. If J (v) < 0,
then σ v ∈ (0, 1), and we can estimate
   +∞ k

1 α
d ≤ I (σ v) = λ σ v f (σ v) − F(σ v) dx = λ 0
(k − 1)σ v2k
L 2k
R2 2 k!
k=2
+∞ k
 +∞ k

α α
=λ 0
(k − 1)σ 2k v2k
L 2k
≤λ 0
(k − 1)v2k
L 2k
k! k!
k=2 k=2
  
1
=λ v f (v) − F(v) dx = H (v).
R2 2

With the above characterization of d, it is easy to show that for any ε > 0
  ε
dε := inf I (v) : v ∈ H 1 (R2 )\{0}, J (v) = −ε ≥ d − . (4.6)
2
In fact, by direct computations
 ε ε 
dε = inf I (v) + − : v ∈ H 1 (R2 )\{0}, J (v) = −ε
2 2
 1 ε 
= inf I (v) − J (v) − : u ∈ H 1 (R2 )\{0}, J (v) = −ε
2 2
 1  ε
= inf I (v) − J (v) : v ∈ H 1 (R2 )\{0}, J (v) = −ε −
 2  ε 2
ε
= inf H (v) : v ∈ H (R )\{0}, J (v) = −ε − ≥ d − .
1 2
2 2
Next, we assume that the maximal solution u to (1.1) satisfies u(t0 ) ∈ V for some
t0 ∈ [0, T∗ ). Then, there exists ε > 0 such that
    
min d − I u(t0 ) , −J u(t0 ) > ε.

In view of (4.6) and the monotonicity of the energy (3.1), we get

ε    
dε ≥ d − > I u(t0 ) ≥ I u(t) , for any t ∈ [t0 , T∗ ). (4.7)
2
   
Assume that J u(t1 ) = −ε for some t1 ∈ (t0 , T∗ ). Then, dε ≤ I u(t1 ) , which
contradicts (4.7).
   
Summarizing, we have J u(t0 ) < −ε, and J u(t) = −ε for any t ∈ [t0 , T∗ ).
Therefore, the proof is complete in view of the continuity of J along the solution, see
Lemma 3.2. 
 M. Ishiwata et al. J. Evol. Equ.

Proof of Theorem 1.2.(i). We argue by contradiction assuming that the solution u is

global, i.e., T∗ = +∞, and we apply the blow-up Lemma 4.2 to the non-negative and
increasing C 2 -function defined by

1 t
y(t) := u(s)2L 2 ds, t ∈ [t0 , +∞). (4.8)
2 t0

In view of (3.2), we have

1 d  
y  (t) = u(t)2L 2 = −J u(t) > ε, t ∈ (t0 , +∞), (4.9)
2 dt
where ε > 0 is given by Proposition 4.3. From (4.9), we deduce that

lim y  (t) = lim y(t) = +∞. (4.10)

t→+∞ t→+∞

Let f˜ and F̃ be as in (1.16). Since there exists θ > 2 such that

θ F̃(s) ≤ s f˜(s), for any s ∈ R,

we can estimate
  
   
y  (t) = −J u(t) = − ∇u(t)2L 2 + (1 − 2α0 λ)u(t)2L 2 + λ u(t) f˜ u(t) dx
R2
  
 
≥ − ∇u(t)2L 2 + (1 − 2α0 λ)u(t)2L 2 + λθ F̃ u(t) dx
R2
  
  θ
≥ −θ I u(t) + − 1 ∇u(t)2L 2 + (1 − 2α0 λ)u(t)2L 2 (4.11)
2
 
≥ −θ I u(t) + C y  (t),

where C = C(θ, α0 , λ) := (θ − 2)(1 − 2λα0 ) > 0.

Using (3.2), we get
 t
   
− I u(t) = ∂s u(s)2L 2 ds − I u(t0 ) , (4.12)
t0

and hence,
   
 θ t t  
y(t)y (t) ≥ u(s)2L 2 ds ∂s u(s)2L 2 ds + y(t) C y  (t) − θ I u(t0 )
2 t0 t0
 t   2 
θ  
≥ u(s)∂s u(s) dx ds + y(t) C y  (t) − θ I u(t0 )
2 t0 R2
 2 (4.13)
θ t
1 d   
≥ u(s)2L 2 ds + y(t) C y  (t) − θ I u(t0 )
2 t0 2 ds
θ  2   
= y (t) − y  (t0 ) + y(t) C y  (t) − θ I u(t0 ) .
2
 Asymptotics for a parabolic equation

In view of (4.10), for any β ∈ (0, 1) there exists tβ > 1 such that
θ
y(t)y  (t) ≥ β[y  (t)]2 , for any t ≥ tβ . (4.14)
2
If we choose β > 0 such that θ2 < β < 1, then we are in the framework of the blow-up
Lemma 4.2, and we reach a contradiction. 
Remark 4.4. The proof of Theorem 1.3 in [14] is given in the particular case α0 = 1
and λ = 21 , but it can be adapted to cover the general case
1
0<λ≤ , for some α0 > 0. (4.15)
2α0
In fact, as showed in [14], it is easier to apply the concavity method of Levine if the
energy of the solution becomes negative. For completeness, we briefly show how to
modify the previous arguments to prove  Theorem
 1.3 in the general case (4.15).
First, assume that u 0 ∈ H (R ), I u(t0 ) = 0 and u(t0 ) = 0 for some t0 ∈ [0, T∗ ),
1 2

then there exists t1 ∈ (t0 , T∗ ) such that I u(t1 ) < 0. If not, then the monotonicity of
 
the energy (3.1) yields I u(t) = 0 for any t ∈ [t0 , T∗ ), and u(t) = u(t0 ) a.e. in R2 ,
 any t ∈ [t0 , T∗ ). Therefore, u(t0 ) solves the stationary problem (1.3), in particular
for
J u(t0 ) = 0. Since
  
1 1
I (v) = J (v) + λ v f (v) − F(v) dx
2 R2 2
1  
≥ J (v) + α02 λv4L 4 , for any v ∈ H 1 (R2 ), (4.16)
2
we deduce that u(t0 ) = 0, which is not possible.
 
Next, assume that u 0 ∈ H 1 (R2 ), and I u(t0 ) < 0 for some t0 ∈ [0, T∗ ). Following
the proof of Theorem 1.2.(i), we argue by contradiction assuming that T∗ = +∞, and
we consider the function y defined in (4.8). Combining (4.9) with (4.16) and the
monotonicity of the energy (3.1), we get
     
y  (t) = −J u(t) ≥ −2I u(t) ≥ −2I u(t0 ) > 0, t ∈ (t0 , +∞);
therefore, it is enough to obtain (4.14) to reach a contradiction. From (4.11), and
recalling that θ > 2 and 1 − 2α0 λ ≥ 0, we get
 
y  (t) ≥ −θ I u(t) .
 
Moreover, since I u(t0 ) < 0, (4.12) yields
 t
 
−I u(t) > ∂s u(s)2L 2 ds,
t0

and hence, arguing as in (4.13), we conclude that

θ  2
y(t)y  (t) ≥ y (t) − y  (t0 ) ,
2
which gives (4.14).
 M. Ishiwata et al. J. Evol. Equ.

5. Global existence and destiny of the orbits in W

The uniqueness of the solution to (1.1) plays a role in the proof of the invariance of
the set W under the flow associated with (1.1).
 
Lemma 5.1. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with λ as
in (1.4), and u 0 ∈ H 1 (R2 ). If u(t0 ) ∈ W for some t0 ∈ [0, T∗ ), then u(t) ∈ W for any
t ∈ [t0 , T∗ ).

Proof. We argue  by contradiction

 assuming
  u(t1 ) ∈ V for some t1 ∈ (t0 , T∗ ).
that
Since J (u) ∈ C [0, T
 ∗ ); R , we have J u(t 2 = 0 for some t2 ∈ [t0 , t1 ). Therefore,
)
either d ≤ I u(t2 ) or u(t2 ) = 0.
   
The monotonicity of the energy (3.1) yields I u(t2 ) ≤ I u(t0 ) < d , and hence
u(t2 ) = 0. Therefore, by uniqueness, u(t) = 0 for any t ∈ [t2 , T∗ ), which contradicts
u(t1 ) ∈ V . 

In order to prove that solutions entering W are global in time, the idea is to apply
Theorem 2.6 in view of the following property of W in the energy space.

Proposition 5.2. If λ is as in (1.4) then, for any v ∈ W , we have ∇v2L 2 < 2d.

Proof. If v ∈ W then in particular


1 1
∇v2L 2 + v2L 2 − λ F(v) dx < d,
2 2 R2

and to complete the proof it is enough to show that


1
P(v) := v L 2 − λ
2
F(v) dx ≥ 0.
2 R2

Note that the auxiliary functional P is strictly related to the definition of the set W ; in
fact, we already pointed out in (1.6) and (1.9) that
 
d = inf I (v) : v ∈ H 1 (R2 )\{0}, P(v) = 0 . (5.1)

It is not difficult to obtain the analogue of (1.10) for the functional P and show that
for any v ∈ H 1 (R2 )\{0} there exists a unique σ̃ = σ̃ (v) > 0 such that
⎧
⎪
⎨> 0 if 0 < σ < σ̃ ,
⎪
P(σ v) = 0 if σ = σ̃ , (5.2)
⎪
⎪
⎩< 0 if σ > σ̃ .

In fact, P(σ v) = 0 if and only if


2λ
v2L 2 = 2 F(σ v) dx,
σ R2
 Asymptotics for a parabolic equation

and the function

 +∞ k

2λ α
h(σ ) := 2 F(σ v) dx = 2λ 0
σ 2(k−1) v2k
σ R2 k! L 2k
k=1

satisfies
lim h(σ ) = 2α0 λv2L 2 , lim h(σ ) = +∞, and h  > 0 on (0, +∞).
σ →0+ σ →+∞

Using (5.2), we prove that if v ∈ W \{0} then P(v) ≥ 0. If this was not true, then
P(v) < 0, and σ̃ = σ̃ (v) ∈ (0, 1). Hence, the characterization of the level d given by
(5.1) yields
d ≤ I (σ̃ v). (5.3)
The point is that (5.3) cannot happen. In fact, since v ∈ W \{0}, Proposition 1.1
implies
d 1
I (σ v) = J (σ v) > 0, for any σ ∈ (0, 1],
dσ σ
and in particular
I (σ̃ v) < I (v) < d.

The set W is stable, and more precisely
 
Theorem 5.3. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with λ
as in (1.4), and u 0 ∈ H 1 (R2 ). If u(t0 ) ∈ W for some t0 ∈ [0, T∗ ) then T∗ = +∞.
Proof. Without loss of generality, we may assume that u(t) = 0 for any t ∈ [t0 , T∗ ).
From Lemma 5.1, we see that u(t) ∈ W for any t ∈ [t0 , T∗ ). On the one hand, (3.2)
yields for any t ∈ (t0 , T∗ )
1 d  
u(t)2L 2 = −J u(t) < 0, (5.4)
2 dt
and
sup u(t) L 2 < +∞.
t∈[t0 ,T∗ )

On the other hand, from Proposition 5.2, we get

sup ∇u(t)2L 2 ≤ 2d, (5.5)
t∈[t0 ,T∗ )

and it is crucial to recall that from (1.7) and (1.9), we know that
4π
2d < . (5.6)
α0
Therefore, we are under the assumptions of Corollary 2.8 which guarantees that T∗ =
+∞. 
 M. Ishiwata et al. J. Evol. Equ.

Theorem 5.4. Assume that λ is as in (1.4). There exists m = m(α0 , λ) > 0 such that

J (v) > 0 for any v ∈ H 1 (R2 )\{0} with ∇v L 2 ≤ m, (5.7)

 of the origin in H (R ). Therefore, if the maximal

1 2
and W contains a neighborhood
solution u ∈ C [0, T∗ ); H (R ) to (1.1) with u 0 ∈ H (R2 ) is global (i.e., T∗ = +∞)
1 2 1

and

lim u(t) H 1 = 0,
t→+∞

then there exists t0 ∈ [0, +∞) such that u(t) ∈ W for any t ∈ [t0 , +∞).

Proof. The relevant part of the proof is to show that (5.7) holds. In fact, it is clear that
(5.7) implies that W contains a neighborhood of the origin in H 1 (R2 ): if we set
  √ 
Sδ := v ∈ H 1 (R2 ) : v H 1 < δ , with 0 < δ < min 2d, m(α0 , λ) ,

then we have for any v ∈ Sδ

1 1
I (v) ≤ v2H 1 < δ 2 < d,
2 2
and (5.7) yields J (v) > 0 provided v = 0. Therefore, Sδ ⊆ W .
In the second part of the statement of Theorem 5.4, since the maximal solution u is
global and

lim u(t) H 1 = 0,
t→+∞

we have u(t) ∈ Sδ for any t > 0 sufficiently large, and we get the desired conclusion.
In order to prove (5.7), we will follow the argument developed in [16, Lemma 2.1],
and we begin by recalling the Gagliardo–Nirenberg inequality:
q q−2
v L q ≤ Cq ∇v L 2 v2L 2 , for any v ∈ H 1 (R2 ), with q ≥ 2. (5.8)

For any v ∈ H 1 (R2 ), we can estimate

   
v 2 (eα0 v − 1) dx
2
λ v f (v) dx = 2α0 λ v2L 2 +
R2 R2
   1   1 
q  2 q
≤ 2α0 λ v2L 2 + |v|2q (eα0 q v − 1) dx
R2 R2
  1
2− 2 2
 2 q
≤ 2α0 λv2L 2 + C∇v L 2 q v Lq 2 (eα0 q v − 1) dx ,
R2

where we used Hölder’s inequality with q, q  > 1 satisfying 1

q + 1
q = 1, and the
Gagliardo–Nirenberg inequality (5.8).
 Asymptotics for a parabolic equation


If 0 < m < 2π
α0 q  , then we can apply the scale invariant Trudinger–Moser inequal-
ity (2.6) to any v ∈ H 1 (R2 ) with ∇v L 2 ≤ m and get
   α q  2
 2
(eα0 q v − 1) dx = (e2π 2π v − 1) dx ≤ Cv2L 2 ,
0

R2 R 2

where the constant C = C(α0 , q) > 0 is independent of m.

C(α0 , λ, q) > 0 such
Summarizing, for any q > 1, there exists a constant C = 
that if v ∈ H (R ) satisfies ∇v L 2 ≤ m, for some 0 < m < α2π
1 2
0q
 , then we have

 2

λ v f (v) dx ≤ 2α0 λv2L 2 + C∇v Lq 2 v2L 2
R2
2
≤ 2α0 λv2L 2 + Cm q  v2L 2 ,

and hence,
2
J (v) ≥ ∇v2L 2 + (1 − 2α0 λ)v2L 2 − Cm q  v2L 2
 
2
≥ (1 − 2α0 λ) − Cm q  v2L 2 .

Since 1 − 2α0 λ > 0, and the constant C > 0 is independent of m, if we choose m

sufficiently small, then we reach the desired conclusion. 
Actually, all the solutions entering W (which are global, i.e., T∗ = +∞, in view of
Theorem 5.3) have the same destiny in the following sense.
 
Theorem 5.5. Let u ∈ C [0, T∗ ); H 1 (R2 ) be the maximal solution to (1.1) with
u 0 ∈ H 1 (R2 ). If u(t0 ) ∈ W for some t0 ∈ [0, T∗ ), then

lim u(t) H 1 = 0.
t→+∞

In order to prove Theorem 5.5, we will use the following convergence result.
Proposition 5.6. Let {vn }n ⊂ W be such that

M := sup vn  L 2 < +∞. (5.9)

n

If

lim J (vn ) = 0, (5.10)

n→+∞

and

lim I (vn )=:I∞ ∈ (−∞, d), (5.11)

n→+∞

then

lim vn  H 1 = 0, and I∞ = 0. (5.12)

n→+∞
 M. Ishiwata et al. J. Evol. Equ.

Proof. Let H be the functional defined by (4.5), i.e.,

  
1 1
H (v) := I (v) − J (v) = λ v f (v) − F(v) dx, for any v ∈ H 1 (R2 ),
2 R2 2

and recall that in the proof of Proposition 4.3, we emphasized that the level d can be
characterized as
 
d = inf H (v) : v ∈ H 1 (R2 )\{0}, J (v) ≤ 0 . (5.13)

Let f˜ be as in (1.16). If we show that


lim vn f˜(vn ) dx = 0, and lim H (vn ) = 0, (5.14)
n→+∞ R2 n→+∞

then the proof is complete. Indeed, we can rewrite


J (vn ) = ∇vn 2L 2 + (1 − 2λα0 )vn 2L 2 − λ vn f˜(vn ) dx,
R2

or equivalently

∇vn 2L 2 + (1 − 2λα0 )vn 2L 2 = J (vn ) + λ vn f˜(vn ) dx,
R2

and hence combining (5.10) with (5.14), we deduce that

lim ∇vn 2L 2 + (1 − 2λα0 )vn 2L 2 = 0.

n→+∞

Since 1 − 2α0 λ > 0, this is enough to obtain the first part of (5.12). The second part
of (5.12) (i.e., I∞ = 0) is a direct consequence of (5.10), (5.14), and the following
identity
1
I (vn ) = H (vn ) + J (vn ).
2
The rest of the proof is dedicated to showing (5.14), and we begin by summarizing
some properties of the sequence {vn }n which will be useful to obtain (5.14). Since
{vn }n ⊂ W , we have

J (vn ) > 0, for any n ≥ 1, (5.15)

and as a consequence of Proposition 5.2, we also have

m := sup ∇vn 22 ≤ 2d. (5.16)

n

In particular, recalling (5.6), we know that

4π
m< , (5.17)
α0
 Asymptotics for a parabolic equation

and this strict inequality will be crucial in the proof of the convergence result expressed
by (5.14).
In order to prove (5.14), we consider the Schwarz symmetrized sequence {vn∗ }n ⊂
Hrad (R2 ), i.e., the sequence of the non-negative spherically symmetric and decreasing
1

rearrangements of {vn }n (see e.g. [27, Chapter 3]). In view of the properties of Schwarz
symmetrization, we have
 
vn f˜(vn ) dx = vn∗ f˜(vn∗ ) dx, and H (vn ) = H (vn∗ ),
R2 R2

and to obtain (5.14), it is enough to show that


lim vn∗ f˜(vn∗ ) dx = 0, and lim H (vn∗ ) = 0. (5.18)
n→+∞ R2 n→+∞

Using again the properties of Schwarz symmetrization, together with (5.16) and
(5.9), we get

sup ∇vn∗ 2L 2 ≤ m, and sup vn∗  L 2 = M < +∞.

n n

In particular, up to subsequences, vn∗  w in H 1 (R2 ), and vn∗ → w a.e. in R2 . We

divide the proof of (5.18) into two steps: first,
1. we show that
 
∗ ˜ ∗
lim vn f (vn ) dx = w f˜(w) dx, and lim H (vn∗ ) = H (w),
n→+∞ R2 R2 n→+∞
(5.19)

and finally,
2. we deduce that w = 0.
Step 1. The proof of (5.19) is a direct consequence of the compactness result [15,
Theorem 1.5(C)] related to the new scale invariant Trudinger–Moser inequality with
the exact growth condition. For the sake of completeness, we show that (5.19) can
be deduced as well by the classical compactness lemma of Strauss [37, Compactness
Lemma 2] (see also [3, Theorem A.I]). From (5.17), we have the existence of ε ∈ (0, 1)
such that
4π
m= (1 − ε), (5.20)
α0
and according to the notations used in [3, Theorem A.I], we introduce the auxiliary
growth function

Q(s) := eα0 (1+ε)s − 1.

2

If P : R → R is a continuous function satisfying

 M. Ishiwata et al. J. Evol. Equ.

P(s) P(s)
(i) lim = 0, and lim = 0,
|s|→+∞
 Q(s) s→0 Q(s)
(ii) sup Q(vn∗ ) dx < +∞,
n R2
(iii) P(vn∗ ) → P(w) a.e. in R2 , and
(iv) vn∗ (x) → 0 as |x| → +∞ uniformly with respect to n,
then the compactness lemma of Strauss guarantees that
P(vn∗ ) → P(w) in L 1 (R2 ). (5.21)
To see that (ii) holds, we renormalize each vn∗ by setting
v∗
wn := √n ,
m
so that (5.16) yields ∇wn  L 2 ≤ 1. In view of the scale invariant Trudinger–Moser
inequality (2.6), we can estimate
  
 α (1+ε)[v ∗ ]2   α (1+ε)mw2 
Q(vn∗ ) dx = e 0 n − 1 dx = e 0 n − 1 dx
R2
R R2
2

 4π(1−ε2 )w2 
= e n − 1 dx
R2
vn∗ 2L 2 M2
≤ Cε wn 2L 2 = Cε ≤ Cε ,
m m
where we also used (5.20) and (5.9).
The a.e.-convergence of the sequence {vn∗ }n and the continuity of P yield (iii).
Moreover, the radial symmetry and the boundedness of {vn∗ }n in H 1 (R2 ) give the
uniform decay at infinity expressed by (iv).
Finally, if we set
1
either P(s) := s f˜(s) or P(s) := s f (s) − F(s)
2
then also the assumption (i) of the compactness lemma of Strauss is satisfied, and
hence, (5.21) holds. This completes the proof of (5.19).
Step 2. First, we show that J (w) ≤ 0. Since 1−2α0 λ > 0, using the weak convergence
vn∗  w in H 1 (R2 ), (5.19), and (5.10), we can estimate

J (w) = ∇w2L 2 + (1 − 2α0 λ)w2L 2 − λ w f˜(w) dx
R2
  
∗ 2 ∗ 2 ∗ ˜ ∗
≤ lim inf ∇vn  L 2 + (1 − 2α0 λ)vn  L 2 − λ vn f (vn ) dx
n→+∞ R2
= lim inf J (vn∗ ) ≤ lim inf J (vn ) = 0.
n→+∞ n→+∞

Next, we argue by contradiction assuming that w = 0. On the one hand, since

w = 0 and J (w) ≤ 0, (5.13) yields
d ≤ H (w).
 Asymptotics for a parabolic equation

On the other hand, from (5.19) and (5.15), we deduce that

H (w) = lim H (vn∗ ) = lim H (vn )

n→+∞ n→+∞
 1 
= lim I (vn ) − J (vn ) ≤ lim I (vn ) = I∞ .
n→+∞ 2 n→+∞

Since by assumption (5.11), we have I∞ < d then

H (w) < d,

and we reach a contradiction. 

Proof of Theorem 5.5. As in the proof of Theorem 5.3, we have the monotonicity of
the L 2 -norm of the solution (5.4) which ensures both

sup u(t) L 2 < +∞ (5.22)

t∈[t0 ,+∞)

and

lim u(t) L 2 exists, (5.23)

t→+∞

and we also have (5.5) and (5.6), i.e.,

4π
sup ∇u(t)2L 2 ≤ 2d < . (5.24)
t∈[t0 ,+∞) α0

Moreover, the monotonicity of the energy (3.1) and [14, Theorem 2.1.3] (see also
Theorem 1.3) guarantee that
 
lim I u(t) = I∞ ∈ [0, d). (5.25)
t→+∞

Next, we find a sequence tn ∈ [t0 , +∞) satisfying

 
lim tn = +∞, and lim J u(tn ) = 0. (5.26)
n→+∞ n→+∞

We point out that (5.25), and in particular the fact that I∞ ≥ 0, implies
d  
lim sup I u(t) = 0. (5.27)
t→+∞ dt
Combining (5.27) with the identity (3.1), we deduce the existence of a sequence
tn ∈ [t0 , +∞) such that

lim tn = +∞, and lim ∂t u(tn ) L 2 = 0.

n→+∞ n→+∞

The sequence {u(tn )}n is a Palais–Smale sequence for the energy functional at the
level I∞ . More precisely, there exists {εn }n ⊂ R+ satisfying
 
|d I u(tn ) , ϕ| ≤ εn ϕ L 2 for any ϕ ∈ H 1 (R2 ), and lim εn = 0. (5.28)
n→+∞
 M. Ishiwata et al. J. Evol. Equ.

In fact, recalling (3.3), it is enough to choose εn := ∂t u(tn ) L 2 . Since u(tn ) ∈ W for
any n ≥ 1, we have J (u n ) > 0, and as a particular case of (5.28), we deduce that
    
0 < J u(tn ) = d I u(tn ) , u(tn  ≤ εn u(tn ) L 2 .
Therefore, with the help of (5.22), we get
 
lim J u(tn ) = 0.
n→+∞

We are now in position to prove that u(t) H 1 → 0 as t → ∞. Recalling (5.22)

and (5.25), since u(tn ) ∈ W satisfies (5.26), then we can apply Proposition 5.6 to
conclude that
lim u(tn ) H 1 = 0, and I∞ = 0. (5.29)
n→+∞

In particular,
lim u(tn ) L 2 = 0,
n→+∞

and hence, in view of (5.23), we have

lim u(t) L 2 = 0.
t→+∞

If we set
u(t)
w(t) := √ , for any t ∈ [t0 , +∞),
2d
then (5.24) yields ∇w(t) L 2 ≤ 1. Applying the scale invariant Trudinger–Moser
inequality (2.6), we can estimate
 
   2dα [w(t)]2 
F u(t) dx = e 0 − 1 dx ≤ Cd w(t)2L 2
R2 R2
u(t)2L 2
= Cd , for any t ∈ [t0 , +∞).
2d
Note that it is possible to apply (2.6) in view of the fact that 2dα0 < 4π , see (5.24).
Therefore,

 
lim F u(t) dx = 0,
t→+∞ R2

and
  
   
lim u(t)2H 1 = lim 2I u(t) + 2λ F u(t) dx = 2I∞ .
t→+∞ t→+∞ R2
But from (5.29), we know that I∞ = 0, and hence the proof of the theorem is complete.

Remark 5.7. The validity of the Palais–Smale compactness condition for the energy
α0 ) is still an open question. This problem will be
functional I in the region (−∞, 2π
addressed in a forthcoming paper.
 Asymptotics for a parabolic equation

6. Asymptotics at the ground state level

In this section, we prove Theorem 1.4 concerning the asymptotic behavior of the
solution for initial data with the same energy as the ground state solution. The key
property is expressed by the following Lemma.
Lemma 6.1. Assume that there exists t0 ∈ [0, T∗ ) such that

I (u(t0 )) = d and J (u(t0 )) = 0.

Then, I (u(t)) < d for any t ∈ (t0 , T∗ ).

Proof. By contradiction, let us assume that there exists t1 ∈ (t0 , T∗ ) such that
I (u(t1 )) = d. Then, by the monotonicity of the energy I (u(t)) = d, for any t ∈ [t0 , t1 ].
Moreover,
d
∂t u(t)2L 2 = − I (u(t)) = 0,
dt
for any t ∈ (t0 , t1 ), and hence u(t) = u(t0 ) a.e. in R2 for any t ∈ [t0 , t1 ]. There-
fore, u(t0 ) is a stationary solution and d I (u(t0 )) ≡ 0. In particular, J (u(t0 )) =
d I (u(t0 )), u(t0 ) = 0 and we reach a contradiction. 
Proof of Theorem 1.4. The results in (i) and (ii) follow directly from Lemma 6.1 and
the arguments in Lemma 4.1 and in Lemma 5.1. Let us now prove (iii). First, we remark
that if v ∈ H 1 (R2 )\{0} is such that I (v) = d and J (v) = 0, then v is a critical point
for the energy functional. Indeed thanks to the property J (v) = 0, we deduce that the
map σ → I (σ v), for σ > 0, attains its unique maximum at σ = 1. In particular,

I (σ v) < I (v) = d, ∀ σ ∈ (0, 1) ∪ (1, +∞). (6.1)

From (5.2), we deduce the existence of a unique σ̃ = σ̃ (v) > 0 such that P(σ̃ v) =
0, where P is the Pohozaev functional defined in Sect. 5. Therefore, if we use the
characterization (1.6) of the ground state energy level d in terms of the Pohozaev
functional, we get

d ≤ I (σ̃ v).

In view of (6.1), the above inequality holds if and only if σ̃ = 1.

By using again the characterization (1.6) of the level d, we deduce that v is a
minimizer; therefore, there exists a Lagrange multiplier θ ∈ R such that
   
∇v · ∇ϕ dx = θ vϕ dx − λ f (v)ϕ dx , for any ϕ ∈ H 1 (R2 ).
R2 R2 R2

In particular,

∇v2L 2 ∇v2L 2
θ=  = = −1.
R2 v dx − λ R2 f (v)v dx J (v) − ∇v2L 2
2
 M. Ishiwata et al. J. Evol. Equ.

Hence, for any ϕ ∈ H 1 (R2 ) we have

  
d I (v), ϕ = ∇v · ∇ϕ dx + vϕ dx − λ f (v)ϕ dx = 0.
R2 R2 R2

Therefore, if there exists t0 ∈ [0, T∗ ) such that

I (u(t0 )) = d, and J (u(t0 )) = 0,

then u(t0 ) is a stationary solution (more precisely a ground state), and by uniqueness,
u is global and it coincides with the ground state u(t0 ) a.e. in R2 after the time t0 . 

7. Indistinguishable splittings

In this section, we adapt the arguments of [16] to show that the Nehari functional
J defined by (1.8) and the Pohozaev functional P defined by (1.15) determine the
same splitting below the ground state energy level d. In particular, we will follow the
arguments of the variational part of the paper [16], and we point out that the energy
critical nonlinearity that we are considering does not fulfill the hypothesis required in
[16].
Given two parameters (a, b) ∈ R2 satisfying

a ≥ 0, and b ≥ 0, with (a, b) = (0, 0), (7.1)

we define the functional

 
Ja,b (v) := a∇v2L 2 + (a + b)v2L 2 −a λv f (v) dx − 2b λF(v) dx.
R2 R2

The above functional is relevant for the two-parameter rescaling function

σ
 
va,b (x) := σ a v σ −b x , σ > 0, x ∈ R2 , (7.2)

in fact
d  σ  1  σ 
I va,b = Ja,b va,b , σ > 0. (7.3)
dσ σ
Note that if a = 1 and b = 0, then

J1,0 (v) := v2H 1 − λ v f (v) dx
R2

is the Nehari functional J defined by (1.8). If a = 0 and b = 1, then


J0,1 (v) := v L 2 − 2λ
2
F(v) dx
R2
 Asymptotics for a parabolic equation

is related to the Pohozaev functional P defined by (1.15), and more precisely J0,1 =
2P. Therefore, Ja,b interpolates between the Pohozaev and Nehari functionals in the
following sense

Ja,b (v) = a J (v) + 2b P(v). (7.4)

If we consider the constrained minimization problem

 
da,b := inf I (v) : v ∈ H 1 (R2 )\{0}, Ja,b (v) = 0 ,

then [16, Lemma 2.6] suggests that da,b is positive and independent of a and b.
Proposition 7.1. Assume that λ is as in (1.4), and (a, b) as in (7.1). Then, the level
da,b is independent of a and b, and more precisely da,b = c, where c is the mountain
pass level (1.5). Hence,
2π
0 < da,b < . (7.5)
α0
In view of (1.7), it is clear that (7.5) is a direct consequence of the first part of the
statement of Proposition 7.1.
Next, we consider the sets
 
Wa,b := v ∈ H 1 (R2 ) : I (v) < da,b , Ja,b (v) ≥ 0 ,

and
 
Va,b := v ∈ H 1 (R2 ) : I (v) < da,b , Ja,b (v) < 0 .

Adapting the arguments of the proof of [16, Lemma 2.9] to our framework, we show
that the functionals Ja,b define the same splitting below the ground state energy level
da,b = d independently of a and b in the range (7.1).
Proposition 7.2. Assume that λ is as in (1.4), and (a, b) as in (7.1). The sets Wa,b
and Va,b are independent of a and b.
The proof of Proposition 7.1 and Proposition 7.2 can be found in Sects. 7.2 and 7.3,
respectively. More precisely, the remaining part of this section is organized as follows.
In Sect. 7.1, we study the geometry of the functionals Ja,b and I along the rescaling
(7.2). The results of Sect. 7.1 will enable us to prove Proposition 7.1 (in Sect. 7.2) and
Proposition 7.2 (in Sect. 7.3).
Remark 7.3. We mention that in view of the results in [16] and [25], we expect that
Proposition 7.1 and Proposition 7.2 hold also when the parameters (a, b) are in the
range

a ≥ 0, and a + b ≥ 0, with (a, b) = (0, 0).

However, for simplicity, we restrict the attention to the range (7.1).

 M. Ishiwata et al. J. Evol. Equ.

7.1. Geometry of the functionals along the rescaling

The energy functional has a unique global maximum point on the rescaling σ ∈
σ .
(0, +∞) → va,b
Lemma 7.4. Assume that λ is as in (1.4), and (a, b) as in (7.1) with a > 0. For any
v ∈ H 1 (R2 )\{0}, there exists σ = σ (v) > 0 such that
⎧
⎪
⎨> 0 if 0 < σ < σ ,
⎪
σ
Ja,b (va,b ) = 0 if σ = σ , x
⎪
⎪
⎩< 0 if σ > σ .
 σ 
Moreover, the map σ ∈ (0, +∞) → I va,b is monotone strictly increasing on (0, σ ),
stricly decreasing on (σ , +∞), and attains its unique maximum at σ .
Proof. Note that it is enough to prove the first part of the statement; then, the second
part is a direct consequence of (7.3).
We have
 σ 
Ja,b va,b = aσ 2a ∇v2L 2 + (a + b)σ 2(a+b) v2L 2
 
   
−λσ 2b aσ a v f σ a v + 2bF σ a v dx.
R2

First, we can rewrite

+∞ k
 α
as f (s) + 2bF(s) = 2 0
(ka + b)s 2k , s ∈ R,
k!
k=1

and
 
    +∞ k
 α0  2k
aσ a v f σ a v + 2bF σ a v dx = 2 (ka + b)σ a v  L 2k
R2 k!
k=1
+∞ k
 α
= 2σ 2a 0
(ka + b)σ 2(k−1)a v2k
L 2k
.
k!
k=1

Therefore,
 σ 
Ja,b va,b = aσ 2a ∇v2L 2 + (a + b)σ 2(a+b) v2L 2
+∞ k
 α
−2λσ 2(a+b) 0
(ka + b)σ 2(k−1)a v2k
L 2k
k!
k=1
= aσ ∇v L 2 + (a + b)σ 2(a+b) (1 − 2λα0 )v2L 2
2a 2

+∞ k
 α0
−2λσ 2(a+b)
(ka + b)σ 2(k−1)a v2k
L 2k
.
k!
k=2
 Asymptotics for a parabolic equation

Next, we distinguish two cases according to b > 0 or b = 0. First, we assume

b > 0, and we rewrite
 σ 
Ja,b va,b = σ 2(a+b) h(σ ),

where

h(σ ) := aσ −2b ∇v2L 2 + (a + b)(1 − 2α0 λ)v2L 2

+∞ k
 α
−2λ 0
(ka + b)σ 2(k−1)a v2k
L 2k
.
k!
k=2
 σ 
In particular, the map σ ∈ (0, +∞) → Ja,b va,b has the same sign as h. Since b > 0,
and by assumption we also have a > 0, it is clear that

lim h(σ ) = +∞, and lim h(σ ) = −∞,

σ →0+ σ →+∞

and
+∞ k

 −2b−1 α
h (σ ) = −2abσ ∇v2L 2 − 2λ 0
(ka + b)2(k − 1)aσ 2(k−1)a−1 v2k
L 2k
< 0.
k!
k=2

Therefore, there exists σ = σ (v) > 0 such that

⎧
⎪
⎨> 0 if 0 < σ < σ ,
⎪
h(σ ) = 0 if σ = σ ,
⎪
⎪
⎩< 0 if σ > σ .

To complete the proof, it remains to consider the case b = 0. In this case, we rewrite
 σ 
Ja,b va,b = σ 2a h̃(σ ),

where
  +∞ k
 α0
h̃(σ ) := a ∇v2L 2 + (1 − 2α0 λ)v2L 2 − 2λ (ka + b)σ 2(k−1)a v2k
L 2k
.
k!
k=2

Since a > 0, we deduce that

h̃(0) > 0, lim h̃(σ ) = −∞, and h̃  (σ ) < 0,

σ →+∞

and the proof is complete. 

Lemma 7.5. Assume that λ is as in (1.4), and (a, b) as in (7.1) with a > 0. For any
v ∈ H 1 (R2 )\{0}, we have
 σ 
lim I va,b = −∞.
σ →+∞
 M. Ishiwata et al. J. Evol. Equ.

Proof. We can rewrite


 σ  1  2a   
I va,b = σ ∇v2L 2 + σ 2(a+b) v2L 2 − λσ 2b F σ a v dx
2 R2

1 2a  +∞ k
 α0 2(k−1)a
= σ ∇v2L 2 + σ 2(a+b) v2L 2 − λσ 2(a+b) σ v2k
L 2k
2 k!
k=1
1  2a  α2
≤ σ ∇v2L 2 + σ 2(a+b) v2L 2 − λσ 2(a+b) 0 σ 2a v4L 4 ,
2 2
and the right-hand side goes to −∞ as σ → +∞, provided a > 0. 
7.2. The level da,b is independent of (a, b) in the range (7.1)

This section is devoted to the proof of Proposition 7.1. As already mentioned in

Sect. 1, from the existence result in [33], we know that the level d0,1 coincides with
the mountain pass level c associated with the energy functional I , i.e.,

d0,1 = c, (7.6)

see (1.6). Moreover, [33, Theorem 4, Propositions 1 and 2] gives the existence of
v ∈ H 1 (R2 )\{0} satisfying

I (v) = c, P(v) = 0, and J (v) = 0. (7.7)

If a = 0 and b > 0, then J0,b (v) = 2b P(v) for any v ∈ H 1 (R2 ), and hence,

d0,b = c, for any b > 0.

To complete the proof of Proposition 7.1, we will show that

da,b = c for any a > 0 and b ≥ 0, (7.8)

and from now on, we assume that a > 0 and b ≥ 0.

Combining (7.4) with (7.7), we get

da,b ≤ I (v) = c (7.9)

Next, we introduce the auxiliary level

 σ 
ca,b := inf max I va,b .
v∈H 1 (R2 )\{0} σ >0

The proof of (7.8) is complete if we show that

c ≤ ca,b , and ca,b = da,b .

Step 1: c ≤ ca,b . Let v ∈ H 1 (R2 )\{0}. From Lemma 7.5, we deduce the existence of
σ̃ = σ̃ (v) > 0 such that
 σ̃ 
I va,b < 0.
 Asymptotics for a parabolic equation

 
We define γ ∈ C [0, 1], H 1 (R2 ) as

s σ̃ if 0 < s ≤ 1,
va,b
γ (s) :=
0 if s = 0,

so that
     
γ ∈ := γ ∈ C [0, 1]; H 1 (R2 ) : γ (0) = 0, I γ (1) < 0 .

Note that the continuity of γ at s = 0 is a consequence of the conditions on (a, b). In

fact, since a > 0 and b ≥ 0, then we have
 σ 2  
lim va,b  1 = lim σ 2a ∇v2 2 + σ 2(a+b) v2 2 = 0.
H L L
(7.10)
σ →0+ σ →0+

By construction,
   σ 
c ≤ max I γ (s) ≤ max I va,b ,
s∈[0,1] σ >0

and hence,

c ≤ ca,b .

Step 2: da,b ≤ ca,b . For any v ∈ H 1 (R2 )\{0},

 applying Lemma 7.4, we have the
σ
existence of σ = σ (v) > 0 such that Ja,b va,b = 0, and
 σ   σ 
da,b ≤ I va,b = max I va,b .
σ >0

Therefore,

da,b ≤ ca,b .

Step 3: da,b ≥ ca,b . Let v ∈ H 1 (R2 )\{0} be such that Ja,b (v) = 0. From Lemma 7.4,
we deduce that σ = σ (v) = 1 and
 σ   1 
ca,b ≤ max I va,b = I va,b = I (v),
σ >0

and hence,

ca,b ≤ da,b .

7.3. The sets Wa,b and Va,b are independent of (a, b) in the range (7.1)

This section is devoted to the proof of Proposition 7.2. First, following [16, Lemma
2.1], we show that the functional Ja,b is positive near the origin of H 1 (R2 ).
 M. Ishiwata et al. J. Evol. Equ.

Lemma 7.6. Assume that λ is as in (1.4), and (a, b) as in (7.1). There exists m =
m(α0 , λ, a + b) > 0 such that

Ja,b (v) > 0 for any v ∈ H 1 (R2 )\{0} with ∇v L 2 ≤ m, (7.11)

and Wa,b contains a neighborhood of the origin in H 1 (R2 ).

Proof. As in Theorem 5.4, the relevant part of the proof is to show (7.11). Since
2F(s) ≤ s f (s) for any s ∈ R, and (a, b) is in the range (7.1), we can estimate the
nonlinear part of the functional Ja,b as
  
a λv f (v) dx + 2b λF(v) dx ≤ λ(a + b) v f (v) dx, for any v ∈ H 1 (R2 ).
R2 R2 R2

Therefore, arguing as in the proof of Theorem 5.4, for any q > 1, we get the existence of
a constant C = C(α 0 , λ, q, a, b) > 0 such that if v ∈ H (R ) satisfies ∇v L 2 ≤ m,
1 2

for some 0 < m < 2π

α0 q  , then

  2

a λv f (v) dx + 2b λF(v) dx ≤ 2α0 λ(a + b)v2L 2 + C∇v Lq 2 v2L 2
R2 R2
2
≤ 2α0 λ(a + b)v2L 2 + Cm q  v2L 2 ,

and hence,
 
2
Ja,b (v) ≥ (a + b)(1 − 2α0 λ) − Cm q  v2L 2 .

Since (a + b)(1 − 2α0 λ) > 0, and the constant C > 0 is independent of m, if we

choose m sufficiently small, then we reach the desired conclusion. 

We point out that

 
Wa,b = v ∈ H 1 (R2 ) : I (v) < da,b , Ja,b (v) > 0 ∪ {0}. (7.12)

In fact, on the one hand, clearly

 
v ∈ H 1 (R2 ) : I (v) < da,b , Ja,b (v) > 0 ∪ {0} ⊆ Wa,b .

On the other hand, if v ∈ Wa,b \{0}, then Ja,b (v) > 0. If not, then Ja,b (v) = 0; hence,
da,b ≤ I (v), and we easily reach a contradiction.
Next, following [16, Lemma 2.9], we show that the set Wa,b is path connected.

Lemma 7.7. Assume that λ is as in (1.4), and (a, b) as in (7.1) with a > 0. The set
σ .
Wa,b \{0} is contracted to {0} by the rescaling σ ∈ (0, 1] → va,b
 Asymptotics for a parabolic equation

σ 
Proof. Let v ∈ Wa,b \{0}. Arguing as in (7.10), we have that va,b H 1 → 0 as
+
σ → 0 , and hence, it is enough to show that
σ
va,b ∈ Wa,b , for any σ ∈ (0, 1).

Since v ∈ Wa,b \{0}, we have v ∈ H 1 (R2 )\{0}, I (v) < da,b , and Ja,b (v) > 0.
Applying Lemma 7.4, we deduce that σ = σ (v) > 1, and
 σ 
Ja,b va,b > 0, for any σ ∈ (0, 1).
 σ 
Moreover, the map σ ∈ (0, 1] → I va,b is monotone strictly increasing, and in
particular
 σ   1 
I va,b < I va,b = I (v) < da,b .

To complete the proof of Proposition 7.2, we follow closely [16, Lemma 2.9].

Proof of Proposition 7.2. In view of Proposition 7.1, the union of the disjoint sets
Wa,b and Va,b is independent of a and b. Therefore, it is enough to show that Wa,b is
independent of a and b.
By definition, the set Va,b is open in H 1 (R2 ). Also, Wa,b is open in H 1 (R2 ): in
fact, we have (7.12), and Lemma 7.6 guarantees that Wa,b contains a neighborhood
of the origin in H 1 (R2 ).
Let a  > 0 and b ≥ 0, then the set Wa  ,b is connected, see Lemma 7.7. For any
(a, b) in the range (7.1), we have Wa  ,b = (Wa  ,b ∩ Wa,b ) ∪ (Wa  ,b ∩ Va,b ), and
hence,

Wa  ,b = Wa  ,b ∩ Wa,b ⊆ Wa,b .

In particular, the set Wa,b is independent of (a, b) if a > 0 and b ≥ 0. Therefore, we

set

W := Wa,b for any (a, b) in the range (7.1) with a > 0.

If (a, b) is in the range (7.1) with a = 0, i.e., a = 0 and b > 0, then there
exists a sequence {an }n of positive real numbers converging to a = 0. More precisely,
{an }n ⊂ R+ and

lim an = 0.
n→+∞

We know that d0,b = dan ,b , and since an > 0, Wan ,b = W . Clearly, for any fixed
v ∈ H 1 (R2 ), we have

lim Jan ,b (v) = J0,b (v),

n→+∞
 M. Ishiwata et al. J. Evol. Equ.

and hence,

W0,b = Wan ,b = W.
n

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.

REFERENCES
[1] S. Adachi, K. Tanaka. Trudinger type inequalities in R N and their best exponents. Proc. Am. Math.
Soc. 128 (2000), 2051–2057.
[2] C. O. Alves, M. M. Cavalcanti. On existence, uniform decay rates and blow up for solutions of the
2-D wave equation with exponential source. Calc. Var. Partial Differ. Equ. 34 (2009), 377–411.
[3] H. Berestycki, P. L. Lions. Nonlinear scalar field equations. I. Existence of a ground state. Arch.
Ration. Mech. Anal. 82 (1983), 313–345.
[4] H. Brezis, T. Cazenave. A nonlinear heat equation with singular initial data. J. Anal. Math. 68
(1996), 277–304.
[5] T. Cazenave, A. Haraux. An introduction to semilinear evolution equations. Translated from the 1990
French original by Yvan Martel and revised by the authors. Oxford Lecture Series in Mathematics
and its Applications, 13. The Clarendon Press, Oxford University Press, New York, 1998.
[6] T. Cazenave, P. L. Lions. Solutions globales d’équations de la chaleur semi linéaires [Global
solutions of semilinear heat equations]. Commun. Partial Differ. Equ. 9 (1984), 955–978.
[7] R. Chill, M. A. Jendoubi. Convergence to steady states of solutions of non-autonomous heat equa-
tions in R N . J. Dyn. Differ. Equ. 19 (2007), 777–788.
[8] C. Collot, F. Merle, P. Raphaël. Dynamics near the ground state for the energy critical nonlinear
heat equation in large dimensions. Commun. Math. Phys. 352 (2017), 215–285.
[9] C. Cortázar, M. del Pino, M. Elgueta. The problem of uniqueness of the limit in a semilinear heat
equation. Commun. Partial Differ. Equ. 24 (1999), 2147–2172.
[10] H. Dai, H. Zhang. Energy decay and nonexistence of solution for a reaction-diffusion equation with
exponential nonlinearity. Bound. Value Probl. 2014, 9 pp.
[11] E. Fašangová, E. Feireisl. The long-time behaviour of solutions to parabolic problems on unbounded
intervals: the influence of boundary conditions. Proc. R. Soc. Edinb. Sect. A 129 (1999), 319–329.
[12] E. Feireisl, H. Petzeltová. Convergence to a ground state as a threshold phenomenon in nonlinear
parabolic equations. Differ. Integral Equ. 10 (1997), 181–196.
[13] G. Furioli, T. Kawakami, B. Ruf, E. Terraneo. Asymptotic behavior and decay estimates of the
solutions for a nonlinear parabolic equation with exponential nonlinearity. J. Differ. Equ. 262
(2017), 145–180.
[14] S. Ibrahim, R. Jrad, M. Majdoub, T. Saanouni. Local well posedness of a 2D semilinear heat
equation. Bull. Belg. Math. Soc. Simon Stevin 21 (2014), 535–551.
[15] S. Ibrahim, N. Masmoudi, K. Nakanishi. Trudinger–Moser inequality on the whole plane with the
exact growth condition. J. Eur. Math. Soc. (JEMS) 17 (2015), 819–835.
[16] S. Ibrahim, N. Masmoudi, K. Nakanishi. Scattering threshold for the focusing nonlinear Klein–
Gordon equation. Anal. PDE 4 (2011), 405–460.
[17] R. Ikehata, M. Ishiwata, T. Suzuki. Semilinear parabolic equation in R N associated with critical
Sobolev exponent. Ann. I. H. Poincaré 27 (2010) 877–900.
[18] R. Ikehata, T. Suzuki. Stable and unstable sets for evolution equations of parabolic and hyperbolic
type. Hiroshima Math. J. 26 (1996), 475–491.
[19] R. Ikehata, T. Suzuki. Semilinear parabolic equations involving critical Sobolev exponent: local
and asymptotic behavior of solutions. Differ. Integral Equ. 13 (2000), 869–901.
[20] N. Ioku, B. Ruf, E. Terraneo. Existence, non-existence and uniqueness for a heat equation with
exponential nonlinearity in R2 . Math. Phys. Anal. Geom. 18 (2015), 19.
 Asymptotics for a parabolic equation

[21] H. Ishii. Asymptotic stability and blowing up of solutions of some nonlinear equations. J. Differ.
Equ. 26 (1977), 291–319.
[22] M. Ishiwata. Existence of a stable set for some nonlinear parabolic equation involving critical
Sobolev exponent. Discrete Contin. Dyn. Syst. 2005, suppl., 443–452.
[23] M. Ishiwata. Asymptotic behavior of strong solutions of semilinear parabolic equations with critical
Sobolev exponent. Adv. Differ. Equ. 13 (2008), 349–366.
[24] M. Ishiwata, B. Ruf, F. Sani, E. Terraneo. Energy methods for a polynomial heat equation. Work
in preparation.
[25] L. Jeanjean, S. Le Coz. Instability for standing waves of nonlinear Klein–Gordon equations via
mountain-pass arguments. Trans. Am. Math. Soc. 361 (2009), 5401–5416.
[26] A. H. Levine. Instability and nonexistence of global solutions to nonlinear wave equations of the
form Pu tt = −Au + F (u). Trans. Am. Math. Soc. 192 (1974), 1–21.
[27] E. H. Lieb, M. Loss, Analysis. Second edition. Graduate Studies in Mathematics, 14. American
Mathematical Society, Providence, RI, 2001.
[28] P. L. Lions. Asymptotic behavior of some nonlinear heat equations. Physica D 5 (1982), 293–306.
[29] M. Ôtani. Existence and asymptotic stability of strong solutions of nonlinear evolution equations
with a difference term of subdifferentials. Qualitative theory of differential equations, Vol. I, II
(Szeged, 1979), pp. 795–809, Colloq. Math. Soc. János Bolyai, 30, North-Holland, Amsterdam-
New York, 1981.
[30] L. E. Payne, D. H. Sattinger. Saddle points and instability of nonlinear hyperbolic equations. Israel
J. Math. 22 (1975), 273–303.
[31] P. Quittner, P. Souplet. Superlinear Parabolic Problems. Blow-up, Global Existence and Steady
States. Birkhäuser Verlag, Basel (2007).
[32] B. Ruf. A sharp Trudinger-Moser type inequality for unbounded domains in R2 . J. Funct. Anal.
219 (2005), 340–367.
[33] B. Ruf, F. Sani. Ground states for elliptic equations in R2 with exponential critical growth. Geo-
metric properties for parabolic and elliptic PDE’s, 251–267, Springer, Milan, 2013.
[34] T. Saanouni. A note on the inhomogeneous nonlinear heat equation in two space dimensions.
Mediterr. J. Math. 13 (2016), 3651–3672.
[35] D. H. Sattinger. On global solution of nonlinear hyperbolic equations. Arch. Ration. Mech. Anal.
30 (1968), 148–172.
[36] R. Schweyer. Type II blow-up for the four dimensional energy critical semi linear heat equation. J.
Funct. Anal. 263 (2012), 3922–3983.
[37] W. A. Strauss. Existence of solitary waves in higher dimensions. Commun. Math. Phys. 55 (1977),
149–162.
[38] T. Suzuki Semilinear parabolic equation on bounded domain with critical Sobolev exponent. Indiana
Univ. Math. J. 57 (2008), 3365–3395.
[39] M. Tsutsumi. On solutions of semilinear differential equations in a Hilbert space. Math. Japan 17
(1972), 173–193.
M. Ishiwata et al. J. Evol. Equ.

Michinori Ishiwata
Department of Systems Innovation, Graduate School
of Engineering Sciences
Osaka University
Osaka 560-8531
Japan
E-mail: ishiwata@[Link]

Bernhard Ruf and Elide Terraneo

Dipartimento di Matematica
Università di Milano
Via C. Saldini 50
20133 Milan
Italy
E-mail: [Link]@[Link]

Elide Terraneo
E-mail: [Link]@[Link]

Federica Sani
Dipartimento di Scienze Fisiche, Informatiche e
Matematiche
Università di Modena e Reggio Emilia
Via G. Campi 213, a
41125 Modena
Italy
E-mail: [Link]@[Link]

Accepted: 30 October 2020