Well-posedness for the initial value problem associated

to the Zakharov–Kuznetsov (ZK) equation in

asymmetric spaces
Émile Deléage, Felipe Linares

To cite this version:

Émile Deléage, Felipe Linares. Well-posedness for the initial value problem associated to the Zakharov–
Kuznetsov (ZK) equation in asymmetric spaces. SN Partial Differential Equations and Applications,
2023, 4 (2), pp.9. �10.1007/s42985-023-00223-5�. �hal-04071783�

HAL Id: hal-04071783

https://hal.science/hal-04071783v1
Submitted on 17 Apr 2023

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.

Distributed under a Creative Commons Attribution 4.0 International License

 WELL-POSEDNESS FOR THE INITIAL VALUE PROBLEM
ASSOCIATED TO THE ZAKHAROV-KUZNETSOV (ZK)
EQUATION IN ASYMMETRIC SPACES.

ÉMILE DELÉAGE1 AND FELIPE LINARES2

Abstract. We study well-posedness for Zakharov-Kuznetsov and modified

Zakharov-Kuznetsov equations in asymmetric spaces. In order to do so, we
extend a theory initiated by Kato for the Korteweg-de Vries equation to higher
dimensions n ≥ 2. As an application, we prove a result concerning dispersive
blow-up for the modified Zakharov-Kuznetsov in dimension 2.

1. Introduction and main results

The Zakharov-Kuznetsov (ZK) equation was first formally derived by Zakharov
and Kuznetsov in [23], as an asymptotic limit of the Euler-Poisson system, in the
setting of the “cold plasma” approximation. This equation describes motion of
plasma in a uniform magnetic field, in a long wave small-amplitude limit, and can
be stated as

(1) ∂t u + ∂1 ∆u + u∂1 u = 0, (t, x) ∈ R × Rn .

In [15], this asymptotic limit was rigorously justified. In [7], this equation was
shown to be an asymptotic limit for the Vlasov-Poisson system. In the case n = 1,
this equation becomes the well-known Korteweg-de Vries (KdV) equation, which
describes waves on shallow water surfaces. Thus equation (1) can be seen as a
generalization of the KdV equation in higher dimensions. Note that (1) is not
integrable (see [5]). However, it possesses conserved quantities (cf. [5] for instance).
These equations belongs to the larger class of nonlinear dispersive equations (see
[18] for an introduction to the subject). We will focus on the properties of the
initial value problem (IVP) associated to (1), that is
(
∂t u + ∂1 ∆u + u∂1 u = 0 (t, x) ∈ R∗+ × Rn ,
(2)
u|t=0 = u0 ,
and to the IVP associated to the generalized Zakharov-Kuznetsov equation which
can be written as
(
∂t u + ∂1 ∆u + uk ∂1 u = 0 (t, x) ∈ R∗+ × Rn ,
(3)
u|t=0 = u0 ,

Date: April 17, 2023.

2010 Mathematics Subject Classification. Primary 35Q53; Secondary 35B65.
Key words and phrases. Zakharov-Kuznetsov, Asymmetric Spaces, Dispersive Blow-up.
The second author was partially supported by CNPq and FAPERJ/Brazil.
1
2 É. DELÉAGE AND F. LINARES

where k ≥ 2. These IVPs were studied by many authors for an initial data u0 ∈
H s (Rn ). In [5], Faminskii showed local well posedness for (2) in dimension 2, in
the setting H s , s ∈ Z+ . Ever since, a lot of advancements have been made. Still
in the two dimensional case, Linares and Pastor proved local well-posedness of (3)
with k = 2 for s > 3/4 by using smoothing effects in [16]. The Fourier restriction
method was also used by Molinet and Pilod in [20] and by Grünrock and Herr in
[6] to extend local well-posedness of (2) to s > 1/2.
In dimension 3, Molinet and Pilod [20] and Ribaud and Vento [21] proved local
and global well-posedness for (2) when s > 1. We also mention the recent works
of Kinoshita [13] and Herr and Kinoshita [8] in which well-posedness for (2) was
obtained with the Picard iteration method in the best possible setting: s > −1/4
in dimension 2 and s > (n − 4)/2 when n ≥ 3.
To describe our results, we define the solution of the linear problem associated to
the IVPs (2) and (3) by using a group of unitary operators {V (t)}t∈R . This group
is given explicitely by the formula V (t)u0 = exp(−t∂1 ∆)u0 , or with the Fourier
transform by V\ (t)u0 (t, ξ) = exp(itξ1 |ξ|2 )c
u0 (ξ).
In [11], Kato studied well-posedness of the IVP associated to the KdV equation
(dimension n = 1) for an initial datum in H s ∩ L2b , where, if b ∈ Rn ,
 Z 
(4) L2b = f : Rn → R, f 2 (x)e2b·x dx < +∞ .
Rn
The key property that Kato used for this particular space is that, after pointwise
multiplication by ebx where b > 0, the unitary group of evolution V (t) becomes
parabolic. More precisely, there exists a parabolic semigroup {Ub (t)} such that
ebx V (t) = Ub (t)ebx (cf. Section 9 in [11]). Among other results, he proved that for
initial data φ ∈ H s ∩L2b , s ≥ 2, there exists a unique solution u ∈ C([0, ∞), H s ∩L2b )
of the IVP associated to the KdV equation, with the map φ 7→ u being continuous in
the associated topologies. Furthermore, he proved that the KdV equation possesses
a smoothing property for solution with initial data in this space: for any u0 ∈
H 2 ∩ L2b , there exists a unique corresponding solution u of KdV such that
||(d/dt)n ebx u(t)||H s ≤ Ct−(s+3n)/2 , 0 < t ≤ T < +∞,
s ≥ 0, n = 0, 1, 2, . . .
(cf. Theorem 11.1 in [11]). In particular, the solution u belongs to C ∞ (R∗+ × R).
For further properties of the solutions of the KdV equation, see [10] and [14].
Here, we generalize the results of [11] for the Zakharov-Kuznetsov and modified
Zakharov-Kuznetsov equations in dimension n ≥ 2, by using a similar method. Our
first result covers well-posedness in H s ∩ L2b for (2) and (3):
Pn
Theorem 1. Let n ≥ 2, u0 ∈ H s0 ∩L2b for some s0 > n/2 and b1 > ( k=2 b2k /3)1/2 .
Then there is a unique solution to (2) or (3) such that u ∈ C([0, ∞); H s0 ∩ L2b )
with the map u0 7→ u continuous in the associated topologies. Moreover, eb·x u ∈
C((0, ∞), H s ) for any s < s0 + 2.
We also extend the smoothing property discovered by Kato in this particular
setting. For the Zakharov-Kuznetsov and the generalized Zakharov-Kuznetsov
equation, we obtain the following result:
Theorem 2. Let n ≥ 2, k ≥ 1, s0 > n/2 and u ∈ C([0, ∞), H s0 ) be the solution to
(2) if k = 1 and to (3) if k ≥ 2. If u0 ∈ H s0 ∩ L2b , with b ∈ Rn as in Theorem 1,
 ZK IN ASYMETRIC SPACES 3

then eb·x u ∈ C((0, ∞), H ∞ ) with the following estimates: for any T > 0, s ≥ 0 and
β > nk/2,

(5) keb·x u(t)kH s ≤ Ct−βs/2 , 0 < t ≤ T,

and for every α ≥ 3β, α > β(1 + kn/2),

(6) k(d/dt)l eb·x u(t)kH s ≤ Ct−(βs+αl)/2 , l = 1, 2, 3, . . .

Nonlinear dispersive equations are also known to exhibit what is called a dispersive
blow-up : a smooth and bounded initial datum with finite energy can result in
a solution which develops pointwise singularities in finite time. This focusing
phenomenon is caused by the linear operator which possesses an unbounded dispersion
speed. In an unbounded domain, it is then possible that infinitely many waves,
initially spatially dispersed, come all together at the same point after a finite
time, resulting in a blow-up. Bona and Saut initiated the mathematical study of
dispersive blow-up for generalized KdV in [3]. We mention [19] for an improvement
of their result, and [17] for a more recent study.
Dispersive blow-up was also studied for other nonlinear dispersive equations. In
[2], Bona, Ponce, Saut, and Sparber studied dispersive blow-up for the nonlinear
Schrödinger equation. In [9], the pointwise notion of dispersive blow-up is extended
in higher dimension n ≥ 2 to larger sets such as lines or spheres for the nonlinear
Schrödinger equation.
As an application of our previous results, we exhibit an example of dispersive
blow-up in the setting n = 2, k = 2. This example extends Theorem 1.3 from [17].

Theorem 3. Let p > 2 and s ∈ N, s ≥ 2. For any t∗ ∈ R∗ , there exists u0 ∈

H s (R2 ) ∩ W s,p (R2 ) ∩ C ∞ (R2 ) such that the corresponding solution of (3) with
k = 2 is global in time and satisfies
(1) u ∈ C(R, H s (R2 )) and
(2) ku(t∗ )kW s,p blows up on the whole line {0} × R, in the following sense: for
any U ⊂ R2 neighborhood of some (0, y) ∈ {0} × R, k∂xs u(t∗ )kLp (U) = +∞.
(3) The non linear part of the solution stays bounded, i.e. for every t ∈ R,
ku(t) − V (t)u0 kW s,p < +∞.

Note that dispersive blow-up was initially defined for the W s,∞ (R2 ) norm caused
by the highest derivative, cf for instance [3]. However, our proof here only works in
the setting W s,p (R2 ) with p < ∞. In fact, we can almost prove the same theorem
in the W s,∞ (R2 ) setting. The missing part is that in the latter case, the smoothing
effect on the nonlinear term defined hereafter in the proof is not sufficient to prove
that the blow-up is caused by the linear part of the solution (third property of the
solution in Theorem 3).
The paper is organised as follows: in section 2, we give the notations and state
a set of useful results that we will need. In section 3, we prove some preliminary
results concerning the space L2b . In particular, we show that the linear group of
evolution operators {V (t)} becomes parabolic after multiplication by an exponential
function. Section 4 is devoted to the proof of Theorem 1. In section 5, we prove
Theorem 2. Section 3, 4 and 5 are greatly inspired of sections 9, 10 and 11 in
[11]. In section 6, we give two examples of linear dispersive blow-up for the group
{V (t)}, and then we prove Theorem 3.
4 É. DELÉAGE AND F. LINARES

2. Notations and some helpful results

Notations: Let n ∈ N∗ . If x ∈ Rn , we denote x = (x1 , . . . , xn ). If 1 ≤ j ≤ n, we
denote ∂j = ∂/∂xj the partial derivative relative to xj . We denote the Laplacian
operator by ∆ = ∂12 + · · · + ∂n2 and the gradient operator ∇ = (∂i )1≤i≤n . If s ∈ R,
H s (Rn ) or H s is the Sobolev space of order s, endowed with the norm k · kH s . If
1 ≤ p ≤ ∞, we denote Lp (Rn ) or Lp the Lebesgue associated with p, endowed with
the norm k · kp . If b = (b1 , . . . , bn ) ∈ Rn and x = (x1 , . . . , xn ) ∈ Rn , we denote
b · x = b1 x1 + · · · + bn xn .
Preliminary results: we use the following propositions to estimate products in
Sobolev spaces (see [1] and [22], [12] respectively):
Proposition 1. Let s, s1 , s2 ∈ R and n ∈ N.
• s≥0
If s ≥ 0, min(s1 , s2 ) ≥ s and s1 + s2 − s > n/2 then there exists a
constant C > 0 such that for any (u, v) ∈ H s1 × H s2 the pointwise product
uv belongs to H s with kuvkH s ≤ CkukH s1 kvkH s2 .
• s<0
If s < 0, 0 > min(s1 , s2 ) ≥ s, s1 + s2 − s > n/2 and s1 + s2 ≥ 0, then
there exists a constant C > 0 such that for any (u, v) ∈ H s1 × H s2 the
pointwise product uv belongs to H s with kuvkH s ≤ CkukH s1 kvkH s2 .
Proposition 2. If s ≥ 0 and f, g ∈ H s ∩ L∞ , then f g ∈ H s with
kf gkH s ≤ C (kf kH s kgk∞ + kf k∞ kgkH s ) .
The following lemma is an oscillatory integral estimate (see Lemma 2.3 in [16]).
Lemma 1. Let n = 2. For any t 6= 0 and u0 ∈ L1 (R2 ), the following estimate
holds:
kV (t)u0 k∞ ≤ |t|2/3 ku0 k1 .
Finally, this nonlinear smoothing effect comes from Proposition 1.4 of [17]:
Proposition 3. Fix k ≥ 2. Let v0 ∈ H s (R2 ), s ∈ N∗ and v ∈ C([−T, T ], H s (R2 )
the local solution of (3). Then
Z t
z(t) = V (t − t′ )v k ∂1 v(t′ )dt′ ∈ C([−T, T ], H s+1 (R2 )).
0

3. The space L2b

Here we follow the proof of Kato in [11] for the dimension 1 and try to adapt
it for higher dimensions. Kato’s proof has three main steps. The first step uses
a commutation property (see below) combined with Duhamel’s principle to derive
smoothing effects for the solutions of the KdV equation. The second step combines
these smoothing properties with an energy estimate to obtain the well-posedness
for the KdV equation in H s ∩ L2b (see Section 4). Finally, the third step improves
the well-posedness result, by using again the smoothing effects, to show that the
solution has an higher regularity (see Section 5).
Kato’s proof is based on the following commutation property: for f ∈ D′ , eb·x ∂i f =
(∂i −bi )eb·x f . Hence, in dimension 1, the operator ∂13 in (1) becomes parabolic when
b > 0. The following lemma extends this property to higher dimensions.
 ZK IN ASYMETRIC SPACES 5

Pn
Lemma 2. Let b ∈ Rn such that b1 > 0 and k=2 b2k < 3b21 . Define the semigroup
" n
#
X
(7) Ub (t) = exp −t(∂1 − b1 ) (∂k − bk )2 , t ≥ 0.
k=1

Then {Ub (t) : t > 0} is an infinitely differentiable semigroup on H s for each s ∈ R,

with
(8) k∂ α Ub (t)k ≤ Cα t−|α|/2 exp(b1 |b|2 t), α ∈ Nn ,

(9) k(d/dt)Ub (t)k ≤ C(t−3/2 + 1) exp(b1 |b|2 t).

′
Ub (t) is bounded from H s to H s , with
′
(10) kUb (t)kB(H s ,H s′ ) ≤ C(t−(s −s)/2 + 1) exp(b1 |b|2 t), s ≤ s′ .
Proof. In Fourier, Ub (t) acts like a multiplication by the factor
" n
#
X
2
λ(t, ξ) = exp −t(iξ1 − b1 ) (iξk − bk ) .
k=1

Developing the products gives

2
|λ(t, ξ)| = eb1 |b| t exp[−tθ(ξ)]
with θ(ξ) = 2ξ1 ξ · b + b1 |ξ|2 . Denote ξ̃ = (ξ2 . . . , ξn ) and b̃ = (b2 , . . . , bn ). For every
b1 > ǫ > 0, we get using Cauchy-Schwarz inequality

˜ 2 + 2ξ1 ξ˜ · b̃ ≥ 3b1 ξ 2 + b1 |ξ|

θ(ξ) = 3b1 ξ12 + b1 |ξ| ˜ 2 − 2|ξ||
˜ b̃||ξ1 |
1
˜ 2 − (b1 − ǫ)|ξ|
≥ 3b1 ξ 2 + b1 |ξ| ˜ 2 − ξ 2 |b̃|2 /(b1 − ǫ)
1 1

˜2 + 3b21 2
− |b̃| − 3b1 ǫ 2
= ǫ|ξ| ξ1 .
b1 − ǫ
By choosing 0 < 3b1 ǫ < 3b21 − |b̃|2 , we obtain that there is C > 0 such that
θ(ξ) ≥ C|ξ|2 . Hence Ub is a parabolic semigroup (see [4], Section 7.4 for instance),
and the results follow. 
Lemma 3. Let
eb·x u ∈ L∞ ([0, T ]; L2), eb·x f ∈ L∞ ([0, T ], H −1 ),

(11) du/dt + ∂1 ∆u = f, 0 < t < T.

Then one has
eb·x u ∈ C([0, T ], L2) ∩ C((0, T ], H s ) for every s < 1,

Z t
b·x b·x
e u(t) = Ub (t)e u(0) + Ub (t − t′ )eb·x f (t′ )dt′ .
0
and the following estimate, for 0 < s < 1 and 0 < t ≤ T :


(12) keb·x ukH s ≤ Ct−s/2 keb·x ukL∞ ([0,T ],L2 ) + keb·x f kL∞ ([0,T ],H −1 ) .
6 É. DELÉAGE AND F. LINARES

Proof. By multiplying Equation (11) by eb·x and using the commutation property,
we obtain
n
X
(13) (d/dt)eb·x u + (∂1 − b1 ) (∂k − bk )2 eb·x u = eb·x f.
k=1
This leads to the integral form of the equation. We can then use Lemma 2 to obtain
that kUb (t − t′ )kB(H −1 ,H s ) ≤ C(1 + (t − t′ )(−1−s)/2 ), which is integrable if s < 1.
We can then bound the nonlinear integral term by
Z t Z t
′ b·x ′ ′
k Ub (t − t )e f (t )dt ks ≤ CT + C (t − t′ )(−1−s)/2 dt′
0 0
and the change of variables r = t′ /t shows that the last integral is bounded on [0, T ]
as a function of t. The estimate follows. 
Lemma 4. Let n ≥ 2, T > 0 and

du/dt + ∂1 ∆u + a(t)∂1 u = 0, 0 ≤ t ≤ T,
where a ∈ C([0, T ], H s0 ) for some s0 > n/2. If eb·x u ∈ L∞ ([0, T ], L2 ), then eb·x u ∈
C((0, T ], H s ) for any s < s0 + 2, with the estimate
′
keb·x ukH s ≤ C(keb·x ukL∞ ([0,T ],L2 ) , kakL∞ ([0,T ],H s0 ) )t−s /2 , 0 < t ≤ T.
Proof. We want to apply the previous lemma, with f = −a(t)∂1 u. Since eb·x u ∈
L∞ ([0, T ], L2 ), we get that eb·x ∂1 u ∈ L∞ ([0, T ], H −1), hence by Proposition 1, since
s0 > n/2, eb·x f = −a(t)eb·x ∂1 u ∈ L∞ ([0, T ], H −1 ). The previous lemma then gives
that eb·x u ∈ C((0, T ], H s ) for every s < 1 with
Z t
b·x b·x
e u(t) = Ub (t)e u(0) − Ub (t − t′ )eb·x a(t′ )∂1 u(t′ )dt′
0
and we obtain the estimate for every s < 1. Now for fix s ∈ [1/2, s0 + 3/2), we will
show that if the result is true for s − 1/2, then it is also true for s. We first note
that ts/2 u solves
d s/2 s
t u − ts/2−1 u + ∂1 ∆ts/2 u + ts/2 a(t)∂1 u = 0.
dt 2
We want to apply again Lemma 3 with
s
f = ts/2−1 u − ts/2 a(t)∂1 u.
2
The term eb·x f is not bounded in H −1 because of the factor ts/2−1 in front of
u (recall that s could be less than 2). However, for every ǫ > 0, we see that
eb·x f ∈ L∞ ([ǫ, T ], H −1 ). We apply Lemma 3 on the interval [ǫ, T ] to get that
ts/2 eb·x u(t) − Ub (t − ǫ)ǫs/2 eb·x u(ǫ)
Z t hs i
= Ub (t − t′ ) (t′ )s/2−1 eb·x u(t′ ) − (t′ )s/2 a(t′ )eb·x ∂1 u(t′ ) dt′ .
ǫ 2
We then let ǫ go to zero to obtain the following integral equation, for 0 < t ≤ T ,
valid a priori in H −1 (note that s ≥ 1/2, hence the unbounded term (t′ )s/2−1 is
integrable):
Z t hs i
s/2 b·x
t e u(t) = Ub (t − t′ ) (t′ )s/2−1 eb·x u(t′ ) − (t′ )s/2 a(t′ )eb·x ∂1 u(t′ ) dt′ .
0 2
 ZK IN ASYMETRIC SPACES 7

Hence by using the properties of the semigroup Ub we obtain

Z th
ts/2 keb·x ukH s ≤C (t − t′ )−1/4 (t′ )s/2−1 keb·x ukH s−1/2
0
i
+(t − t′ )−3/4 (t′ )s0 /2 ka(t′ )eb·x ∂1 u(t′ )kH s−3/2 dt′ .
Now by hypothesis the first term can be estimated by
(t − t′ )−1/4 (t′ )s/2−1 keb·x ukH s−1/2 ≤ C(t − t′ )−1/4 (t′ )−3/4
which is integrable, with an integral bounded for t ∈ [0, T ], and the second one by
(t − t′ )−3/4 (t′ )s/2 ka(t′ )eb·x ∂1 u(t′ )kH s−3/2
≤C(t − t′ )−3/4 (t′ )s/2 ka(t′ )kH s keb·x ∂1 ukH s−3/2
≤C(t − t′ )−3/4 (t′ )s/2 keb·xukH s−1/2 ≤ C(t − t′ )−3/4 (t′ )1/4
which is also integrable, with a bounded integral. Here we used again Proposition
1 and the hypothesis on keb·x ukH s−1/2 . This concludes the proof. 

4. Proof of Theorem 1
Pn
Let s0 > n/2 and b ∈ Rn such that b1 > 0, and b1 > ( k=2 b2k /3)1/2 . In this
section, we show well-posedness of (2) and (3) in H s0 ∩ L2b . As a consequence of the
well posedness theory in H s0 , we already know that there exists a unique solution in
C([0, T ], H s0 ), which is global if the norm of the initial data is sufficiently small. To
simplify computations, we will restrict ourselves to the setting of global solutions.
However, one could adapt the following proof to show the local well-posedness in
H s0 ∩ L2b when the solution of (2) or (3) in H s0 is only local in time. According to
Lemma 4, it is enough to show that eb·x u(t) ∈ L2 . In the following, we fix k ≥ 1.
Again, we follow Kato ([11]) and introduce the bounded weight functions
eb·x eb·x
q(x) = , r(x) = , p(x) = q(x)2
(1 + ǫe2b·x )1/2 1 + ǫe2b·x
depending on a parameter ǫ > 0. Both q and r are L∞ functions with the L∞
norm proportional to ǫ−1/2 , and both tend monotonically to eb·x as ǫ ↓ 0. We note
several properties of these functions required in the sequel:
(14)
∂i p = 2bi r2 , |∂i ∂j p| < 4|bi bj |r2 , |∂i ∂j ∂k p| < 8|bi bj bk |r2 , |∂i r| < |bi |r.
We now take u the solution of the problem in H s , multiply equation (1) (or (3))
by 2pu and integrate over Rn to obtain
Z Z
d
pu2 = −2 pu(∂1 ∆u + uk ∂1 u).
dt
Integrations by parts give that
Z Z
1
puk+1 ∂1 u = − (∂1 p)uk+2
k+2
and Z Z
1
pu∂1 ∆u = − (∂1 ∆p)u2 − 2(∂1 u)∇p · ∇u − (∂1 p)|∇u|2 .
2
Now, using (14) leads to
8 É. DELÉAGE AND F. LINARES

Z Z
−2(∂1 u)∇p · ∇u − (∂1 p)|∇u|2 = −2 r2 [b1 |∇u|2 + 2(∂1 u)b · ∇u].

Note that b1 |∇u|2 + 2(∂1 u)b · ∇u = θ(∇u) ≥ C|∇u|2 , where C > 0, in virtue of the
condition on b (see the proof of Lemma 2 for the definition and properties of θ).
Using again (14) and putting everything together yields
Z Z Z Z
d 2 2 2 2 4 2 k+2
pu < 8b1 |b| r u + b1 r u − C r2 |∇u|2 .
dt k+2
Since u ∈ H s0 with s0 > n/2, u ∈ L∞ . Finally, we get
d
kquk22 ≤ K(kukH s0 )kruk22 ≤ K(kukH s0 )kquk22 .
dt
Since K is independent of ǫ, going to the limit ǫ ↓ 0 gives
keb·x u(t)k22 ≤ eKt ku0 k22 , 0≤t≤T
with K = K(kukL∞([0,T ],H s0 ) ). Since a(t) = u ∈ C([0, T ], H s0 ), we can apply
Lemma 4 to obtain that eb·x u ∈ C([0, T ], L2 ) ∩ C((0, T ], H s ) for any s ≤ s0 + 2.
Thus we have proved the main part of Theorem 1.
It remains to prove the continuous dependence u0 7→ u. Since this is known for
the H s0 norm by the H s0 theory, it suffices to show that the map eb·x u0 7→ eb·x u
is continuous in the L2 norms, uniformly in t ∈ [0, T ]. This can be seen by the
following integral equation satisfied by v(t) = eb·x u(t):
Z t
v(t) = Ub (t)v0 − W (t, t′ )v(t′ )dt′ ,
0

where
W (t, t′ ) = (∂1 − b1 )Ub (t − t′ )uk (t′ )/(k + 1)
is an operator valued kernel such that kW (t, t′ )kB(L2 ) ≤ C(t − t′ )−1/2 (because
u ∈ H s0 and s0 > n/2). This equation is obtained by Lemma 1 with f = −uk ∂1 u.
It should be noted that W (t, t′ ) depends on u and hence on u0 , but the dependence
is known to be continuous in the H s0 norm which is weaker than the H s0 ∩ L2b
norm.

5. Proof of Theorem 2
We present here the proof of Theorem 2.

Proof. We start by proving the first inequality. By Theorem 1 and the estimate of
Lemma 4, we already know that it is true for any s < s0 +2 (note that β > nk/2 ≥ 1,
hence the estimate of Lemma 4 for s < s0 + 2 is stronger than the one that we need
to prove). Now we fix δ > 0 and show that if the estimate holds for some s − δ with
s ≥ 1/2, then it also holds for s. We write again the integral equation satisfied by
tβs/2 eb·x u(t):
Z t  
βs/2 b·x ′ βs ′ βs/2−1 b·x ′ ′ βs/2
t e u(t) = Ub (t − t ) (t ) e u(t ) − (t ) u(t ) e ∂1 u(t ) dt′ .
′ k b·x ′
0 2
 ZK IN ASYMETRIC SPACES 9

Hence by using the properties of the semigroup Ub we obtain

Z th
tβs/2 keb·xukH s ≤C (t − t′ )−δ/2 (t′ )βs/2−1 keb·x ukH s−δ
0
i
+(t − t′ )−(δ+1)/2 (t′ )βs/2 ku(t′ )k eb·x ∂1 u(t′ )kH s−1−δ dt′ .

Now by hypothesis the first term can be estimated by

(t − t′ )−δ/2 (t′ )βs/2−1 keb·x ukH s−δ ≤ C(t − t′ )−δ/2 (t′ )βδ/2−1 ,
and the integral of this term is finite and bounded in t ≤ T whenever β ≥ 1 and
0 < δ < 2 (to prove this, one can again make the change of variables r = t′ /t).
For the second term, we write that
1
ku(t′ )k eb·x ∂1 u(t′ )kH s−δ−1 = k(∂1 −b1 )eb·x u(t′ )k+1 kH s−δ−1 ≤ Ckeb·x u(t′ )k+1 kH s−δ .
k+1
To estimate the norm of eb·x uk+1 = (eb·x/(k+1) u)k+1 we use Proposition 2. By
′
induction on k, we obtain the following generalized version: for any v ∈ H s ∩ L∞
and k ≥ 0,
kv k+1 kH s′ ≤ Ckvkk∞ kvkH s′ .
We use this last inequality with v = eb·x/(k+1) u and s′ = s − δ, combined with the
Sobolev embedding theorem, to obtain
keb·x u(t′ )k+1 kH s−δ ≤ Ckeb·x/(k+1) u(t′ )kH s−δ keb·x/(k+1) u(t′ )kk∞
≤ Ckeb·x/(k+1) u(t′ )kH s−δ keb·x/(k+1) u(t′ )kkH s1 ,
with s1 > n/2. Now we use the hypothesis for s − δ, and the estimate of Lemma 4
for s1 , to obtain that
(t − t′ )−(δ+1)/2 (t′ )βs/2 ku(t′ )k eb·x ∂1 u(t′ )kH s−1−δ ≤ C(t − t′ )−(δ+1)/2 (t′ )(βδ−ks1 )/2 .
The integral of this expression is finite and bounded in t ≤ T if we take 1 > δ > 0
and s1 > n/2 such that β ≥ 1 + (ks1 − 1)/δ (once again, this bound comes from
the change of variables r = t′ /t). The hypothesis β > nk/2 ensures that we can
find such s1 and δ.
Note that we are allowed to use the property for eb·x/(k+1) u instead of eb·x u
1−1/(k+1) b·x 1/(k+1)
because eb·x/(k+1) u0 ∈ L2 , with keb·x/(k+1) u0 k22 ≤ ku0 k2 ke u0 k2
(Hölder). The homogeneous condition verified by b is also true for b/(k + 1), hence
we can use the hypothesis for eb·x/(k+1) u instead of eb·x u. Hence the decay (5) is
valid for every s ≥ 0.
Now we prove the second inequality by induction on l. For l = 0, it is known by
(5). Assuming that it has been proved for all s ≥ 0 up to a given l, we prove it for
l + 1. Again using (13) with f = −uk ∂1 u, we obtain on differentiation
n
X
(15) (d/dt)l+1 eb·x u = −(∂1 − b1 ) (∂k − bk )2 (d/dt)l eb·x u − (d/dt)l eb·x uk ∂1 u.
k=1
s
The H norm of the first term on the right is dominated by
k(d/dt)l eb·x ukH s+3 ≤ Ct−(β(s+3)+αl)/2 ≤ Ct−(βs+α(l+1))/2
by induction hypothesis. This gives the required estimate.
10 É. DELÉAGE AND F. LINARES

For the second term in (15), we have as above

k(d/dt)l eb·x uk ∂1 ukH s ≤Ck(d/dt)l eb·x uk+1 kH s+1
X
≤C k(dl1 v) . . . (dlk+1 v)kH s+1 ,
l1 +···+lk+1 =l

where we have written v = eb·x/(k+1) u and d = d/dt for simplicity. Using Proposition
2 multiple times again, we obtain
k+1
X Y
k(dl1 v) . . . (dlk+1 v)kH s+1 ≤ C kdli vkH s+1 kdlj vk∞ .
i=1 j6=i
s1 ∞
By induction hypothesis and H ֒→ L , where s1 = (α/β − 1)/k (since we know
that α > β(kn/2 + 1)), this is dominated by Ct−m/2 , where
X
m = β(s + 1) + αli + (βs1 + αlj ) = β(s + 1) + αl + kβs1 = βs + (l + 1)α.
j6=i

This is the required estimate. 

6. Application: dispersive blow-up in dimension n ≥ 2

6.1. Linear dispersive blow-up. In this section, we construct an initial datum
for the linear problem associated to (1) such that the linear evolution exhibits a
singularity at a given time, on a linear subspace of dimension d < n. More precisely,
we state the following
Proposition 4. Let n ≥ 2 and 1 ≤ d ≤ n. For t ∈ R, let V (t) = e−t∂1 ∆ . Recall
that (V (t))t∈R is a group of evolution operators that preserves H s norms. For any
t∗ ∈ R∗ , there exists u0 ∈ H 1 (Rn ) ∩ C ∞ (Rn ) such that:
(1) For every t ∈ R − {t∗ }, u(t) = V (t)u0 ∈ H 1 (Rn ) ∩ C ∞ (Rn ).
(2) u(t, x) ∈ C 0 (R × Rn ) ∩ C ∞ (R × Rn − {t∗ } × Vd ), where
Vd = {(x1 , . . . , xn ) ∈ Rn , x1 = · · · = xd = 0}.
(3) |Du(t, x)| → +∞ when (t, x) → (t∗ , x∗ ), for every x∗ ∈ Vd .
Proof. For x ∈ Rn , let us write x = (y, z) where y = (x1 , . . . , xd ) and z =
2
(xd+1 , . . . , xn ). Let φ(x) = e−2π|y|e−π|z| . Note that φ has an exponential decay,
which will enable to use the Pn smoothing properties of Lemma 2. Take any b ∈ Rn
such that 1 ≥ b1 > 0 and k=2 bk < 3b1 . Then eb·x φ belongs to L2 (Rn ). Note that
2 2

eb·x V (t)φ = Ub (t)eb·x φ, where Ub (t) is defined as in Lemma 2. By the smoothing

properties of Ub (t) stated in Lemma 2, for any t > 0, the function Ub (t)eb·x φ belong
to H ∞ (Rn ), hence is smooth. Thus V (t)φ is also a smooth function.
For negative times, use the fact that e−b·x φ also belongs to L2 (Rn ), and
e−b·x V (−t)φ = U−b (−t)e−b·x φ.
Reversing the proof of Lemma 2 shows that U−b is parabolic backwards in time.
Hence here again e−b·x V (−t)φ and then V (−t)φ are smooth functions, for any
t > 0.
The candidate for Proposition 4 is thus u0 = V (−t∗ )φ. By the previous arguments,
V (t)u0 is smooth for any t 6= t∗ . We then show that u0 ∈ H 1 (Rn ). By the properties
 ZK IN ASYMETRIC SPACES 11

of V (t), it is enough to check that φ ∈ H 1 (Rn ). The Fourier transform of φ is given

by
2
e−π|ξz |
φ̂(ξy , ξz ) = C := fˆ(ξy )ĝ(ξz )
(1 + |ξy |2 )(d+1)/2
where C > 0 is a constant. Since g ∈ S(Rn−d ) and f ∈ H 1+d/2− (Rd ), φ ∈
H 1+d/2− (Rn ). Note that, for any s ≥ 1,
Z
2 |ξy |s
|Dys φ(0, z)| = C e−π|z| d+1 dξy = ∞.
Rd (1 + |ξy |2 ) 2

We also state the following example in the case n = 2:

Proposition 5. Let s ∈ N∗ . For any t∗ ∈ R∗ , there exists u0 ∈ H s (R2 ) ∩ C ∞ (R2 )
such that:
(1) For every t ∈ R − {t∗ }, u(t) = V (t)u0 ∈ H s (R2 ) ∩ C ∞ (R2 ).
(2) u(t, x) ∈ C s−1 (R × R2 ).
(3) |∂xs u(t, x)| → +∞ when (t, x) → (t∗ , x∗ ), ∀x∗ ∈ {(x, y) ∈ R2 , x = 0}.
2 2
Proof. Let p > 2 and φp (x, y) = |x|s−1/p e−x −y . Then φp ∈ H s ∩L2b for any b ∈ R2 .
The proof of the previous proposition shows that V (t)φp is smooth for any t 6= 0.
2 2
Note that ∂xs φp (x, y) = Csgn(x)s |x|−1/p e−x −y + g(x, y), where g is a continuous
function with exponential decay. In particular, ∂xs φp ∈ L2 and |∂xs φp (x, y)| → ∞
for any y when x goes to zero. Taking u0 = V (−t∗ )φp again enables to end the
proof. 

6.2. Non linear dispersive blow-up on a line. We give here the proof of
Theorem 4.

Proof. We use here a proof very similar to the one of Theorem 1.3 in [17]. Consider
φp as in the proof of Proposition 5 and define u0 = V (−t∗ )φp . We write
Z t
u(t) = V (t)u0 + V (t − t′ )u2 ∂x u(t′ )dt′ := V (t)u0 + z(t)
0

the solution of (3) with n = k = 2. Since {V (t)}t∈R is an unitary group in H s

and φp ∈ H s (R2 ), u0 ∈ H s (R2 ). Up to multiplying φp by a small constant, we can
suppose that u(t) is globally defined and u ∈ C(R, H s ). By the Proposition 1.4 of
[17] (cf. Proposition 3), z(t) ∈ H s+1 (R2 ) ⊂ W s,p (R2 ) for all times. By the proof
of Proposition 5, V (t∗ )u0 = φp ∈ W s,1 (R2 ). By Lemma 1, for any t 6= t∗ ,

kV (t)u0 kW s,∞ (R2 ) ≤ C|t − t∗ |−2/3 kφp kW s,1 (R2 ) .

Hence for any t 6= t∗ , V (t)u0 ∈ W s,∞ (R2 ) ∩ H s (R2 ) ⊂ W s,p (R2 ). Thus the solution
u(t) = V (t)u0 + z(t) belongs to W s,p whenever t 6= t∗ . Finally, by Proposition 5,
u0 ∈ C ∞ (R2 ).
2 2
But for any y ∈ R, |∂xs φp (x, y)| ∼ C|x|−1/p e−x −y as x goes to zero. Hence the
Lp norm of ∂xs u(t∗ ) = ∂xs φp + ∂xs z(t∗ ) blows up on any open subset U ⊂ R2 such
that U ∩ ({0} × R) 6= ∅. 
12 É. DELÉAGE AND F. LINARES

Declarations
• Conflict Of Interest statement: there is no competing interest.
• Data availability statement: there is no associated data.

References
[1] A. Behzadan and M. Holst. Multiplication in Sobolev spaces, revisited. Arkiv för Matematik,
59:275–306, 2015.
[2] J. Bona, G. Ponce, J.-C. Saut, and C. Sparber. Dispersive blow-up for nonlinear Schrödinger
equations revisited. Journal de Mathématiques Pures et Appliquées, 102(4):782–811, 2014.
[3] J. Bona and J. Saut. Dispersive blowup of solutions of generalized Korteweg-de Vries
equations. Journal of Differential Equations, 103(1):3–57, 1993.
[4] L. Evans. Partial Differential Equations. Graduate Series in Mathematics. American
Mathematical Society, 2010.
[5] A. Faminskii. The Cauchy problem for the Zakharov-Kuznetsov equation. Differential
Equations, 31:1002–1012, 1995.
[6] A. Grünrock and S. Herr. The Fourier restriction norm method for the Zakharov-Kuznetsov
equation. Discrete and Continuous Dynamical Systems - A, 34(5):2061–2068, 2014.
[7] D. Han-Kwan. From Vlasov–Poisson to Korteweg–de Vries and Zakharov–Kuznetsov.
Communications in Mathematical Physics, 324(3):961–993, 2013.
[8] S. Herr and S. Kinoshita. The Zakharov–Kuznetsov equation in high dimensions: small initial
data of critical regularity. Journal of Evolution Equations, 21(2):2105–2121, 2021.
[9] Y. Hong and M. Tasković. On dispersive blow-ups for the nonlinear Schrödinger equation.
Differential and Integral Equations, 29(9/10):875 – 888, 2016.
[10] P. Isaza, F. Linares, and G. Ponce. On decay properties of solutions of the k-generalized KdV
equation. Communications in Mathematical Physics, 324:129–146, 2012.
[11] T. Kato. On the Cauchy problem for the (generalized) Korteweg-de Vries equation. Advances
in Mathematics Supplementary Studies, 8:93–128, 1983.
[12] T. Kato and G. Ponce. Commutator estimates and the Euler and Navier-Stokes equations.
Communications on Pure and Applied Mathematics, 41:891–907, 1988.
[13] S. Kinoshita. Global well-posedness for the Cauchy problem of the Zakharov-kuznetsov
equation in 2d. Annales de l’Institut Henri Poincaré C, Analyse non linéaire, 38(2):451–
505, 2021.
[14] S. N. Kruzhkov and A. V. Faminskiı̆. Generalized solutions of the Cauchy problem for the
Korteweg-de Vries equation. Mathematics of the USSR-Sbornik, 48(2):391–421, 1984.
[15] D. Lannes, F. Linares, and J.-C. Saut. The Cauchy problem for the Euler-Poisson system and
derivation of the Zakharov-Kuznetsov equation. Progress in Nonlinear Differential Equations
and Their Application, 84:181–213, 2012.
[16] F. Linares and A. Pastor. Well-posedness for the two-dimensional modified
Zakharov–Kuznetsov equation. SIAM J. Math. Analysis, 41:1323–1339, 01 2009.
[17] F. Linares, A. Pastor, and J. Drumond Silva. Dispersive blow-up for solutions of the zakharov-
kuznetsov equation. Annales de l’Institut Henri Poincaré C, Analyse non linéaire, 38(2):281–
300, 2021.
[18] F. Linares and G. Ponce. Introduction to Nonlinear Dispersive Equations. Universitext.
Springer New York, 2015.
[19] F. Linares, G. Ponce, and D. Smith. On the regularity of solutions to a class of nonlinear
dispersive equations. Mathematische Annalen, 369:797–837, 2017.
[20] L. Molinet and D. Pilod. Bilinear strichartz estimates for the Zakharov–Kuznetsov equation
and applications. Annales de l’Institut Henri Poincaré C, Analyse non linéaire, 32(2):347–
371, 2015.
[21] F. Ribaud and S. Vento. Well-posedness results for the three-dimensional Zakharov–
Kuznetsov equation. SIAM Journal on Mathematical Analysis, 44(4):2289–2304, 2012.
[22] T. Tao. Nonlinear Dispersive Equations: Local and Global Analysis. Conference Board of the
Mathematical Sciences regional conference series in mathematics. American Mathematical
Society, 2006.
[23] V. Zakharov and E. Kuznetsov. Three-dimensional solitons. Soviet Physics JETP, 29:594–
597, 1974.
 ZK IN ASYMETRIC SPACES 13

1 ENS de Lyon, 35 parvis René Descartes, 69007 Lyon, France

Email address: emile.deleage@ens-lyon.fr
2 IMPA, Estrada Dona Castorina 110, Rio de Janeiro 22460-320, RJ, Brazil
Email address: linares@impa.br