arXiv:2311.10683v1 [math.

AP] 17 Nov 2023

Solvability of Vekua-type periodic operators

and applications to classical equations
Alexandre Kirilov, Wagner Augusto Almeida de
Moraes and Pedro Meyer Tokoro

Abstract. In this note, we investigate Vekua-type periodic operators

of the form P u = Lu − Au − B ū, where L is a constant coefficient
partial differential operator. We provide a complete characterization of
the necessary and sufficient conditions for the solvability and global
hypoellipticity of P . As an application, we provide a comprehensive
characterization of Vekua-type operators associated with classical wave,
heat, and Laplace equations.
Mathematics Subject Classification (2010). Primary 35B10 Secondary
35E20, 30G20.
Keywords. Vekua-type operators, Solvability, Global hypoellipticity, pe-
riodic solutions.

1. Introduction
In [15], the Georgian mathematician I. N. Vekua introduced the theory of
generalized analytic functions, which are the solutions to equations of the
form
∂z̄ u + Au + B ū = F,
where ∂z̄ = (∂x + i∂y )/2 and the coefficients A and B belong to a suitable
function space in the complex plane. This theory is closely related to the
theory of holomorphic functions, assuming some regularity on the coefficients
to apply the well-known similarity principle. Vekua also applied this theory
to explore problems in the membrane theory of shells and their connections
with the problem of infinitesimal bendings of surfaces.

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de

Nı́vel Superior - Brasil (CAPES) - Finance Code 001. The first and second authors were
supported in part by CNPq - Brasil (grants 316850/2021-7 and 423458/2021-3).
2 Kirilov, Moraes and Tokoro

In [11], V. Kravchenko extended this theory by replacing the Cauchy-

Riemann operator with more general complex-valued vector fields. This ex-
tension allowed the study of well-known equations from mathematical physics,
such as the Schrödinger, Dirac, and Maxwell equations, among others. Fur-
thermore, connections between second-order elliptic equations and Vekua-
type equations were established.
The contributions of Vekua, Kravchenko, and several other authors have
found wide applications, including studies in boundary value problems in
elasticity theory, hydrodynamics, electric potential, mechanics, and more.
In this paper we investigate the solvability of the operator P on the
n-dimensional torus Tn ≃ Rn /2πZn , defined by
P u = Lu − Au − B ū, (1)
∞ n ∞ n
where A, B ∈ C, and L : C (T ) → C (T ) is a partial differential operator.
Our main contribution lies in establishing both necessary and sufficient
conditions for the solvability of the operator P . Furthermore, we establish
the equivalence between solvability and global hypoellipticity. In conclusion,
we apply this result to study Vekua-type operators associated with classical
operators, characterizing their solvability and global hypoellipticity.
Our approach is inspired by results in [4] and [5]. Additional insights
about solvability on the torus can be obtained in [1, 2, 3, 8, 14] and related
works. In the broader context of compact Lie groups, some initial results are
presented in [7, 9, 10]. Additional references regarding infinitesimal deforma-
tions of surfaces connected with solvability of Vekua-type operators include
[6, 12, 13] and references therein

2. Necessary and sufficient conditions for solvability

Consider the operator P : C ∞ (Tn ) → C ∞ (Tn ) defined by
P u = Lu − Au − B ū, (2)
where A, B ∈ C, and L is a differential operator of the form
X
Lu = cα ∂ α ,
0<|α|6m

where cα ∈ C, for all α ∈ Nn0

satisfying 0 < |α| 6 m, with symbol
X
σL (ξ) = i|α| cα ξ α , ξ ∈ Zn .
0<|α|6m
P
By the continuity of P , if u(x) = ξ∈Zn u b(ξ)eiξ·x is a distribution in
′ n
D (T ) then
X
X
P u(x, t) = (L − A) u b(ξ)eiξ·x − B b(ξ)eiξ·x
u
ξ∈Zn ξ∈Zn
X X
= b(ξ)eiξ·x − B
(σL (ξ) − A) u b(−ξ)e−iξ·x .
u
ξ∈Zn ξ∈Zn
 Solvability of Vekua-type periodic operators 3

Therefore, the Fourier coefficients of any solution of P u = f must satisfy

the following equation:
(σL (ξ) − A)bu(ξ) − Bb u(−ξ) = fb(ξ).
Taking the conjugate of the previous equation for −ξ ∈ Zn , we obtain
fb(−ξ) = (σL (−ξ) − Ā)b
u(−ξ) − B̄b
u(ξ)
which gives us, for each ξ ∈ Zn , the following linear system:
(
(σL (ξ) − A) u
b(ξ) − Bb u(−ξ) = fb(ξ)
−B̄bu(ξ) + (σL (−ξ) − Ā)bu(−ξ) = fb(−ξ)
Solving this system for u
b(ξ), we obtain
b(ξ) = (σL (−ξ) − Ā)fb(ξ) + B fb(−ξ),
∆ξ u (3)
where the discriminant
∆ξ = (σL (ξ) − A) · (σL (−ξ) − Ā) − |B|2 . (4)
n
Observe that ∆ξ = ∆−ξ , for all ξ ∈ Z . In particular, ∆ξ = 0 if, and
only if, ∆−ξ = 0.
We consider the same notion of solvability used in [4] and [5].
Definition 1. We say that an operator P : C ∞ (Tn ) → C ∞ (Tn ) is solvable if
exists a subspace F ⊂ C ∞ (Tn ) of finite codimension such that, for all f ∈ F ,
exists u ∈ C ∞ (Tn ) such that P u = f .
Theorem 2. The operator P is solvable if and only if the following Diophan-
tine condition holds: there exists γ > 0 such that
kξk > γ ⇒ |∆ξ | > kξk−γ . (DC)
Proof. This demonstration follows the same procedure as the proof of Theo-
rem 1 given in [5]. Let us start by assuming that the condition (DC) holds.
In this context, the set Λ = {ξ ∈ Zn : kξk < γ} is finite, and the discriminant
∆ξ is non-zero for all ξ ∈ Zn \ Λ. Additionally, we have ∆−1 γ
ξ 6 kξk , for every
n
ξ ∈ Z \ Λ.
If the Fourier coefficients of f ∈ C ∞ (Tn ) satisfy the compatibility con-
dition
(σL (−ξ) − Ā)fb(ξ) + B fb(−ξ) = 0, ξ ∈ Λ (5)
By (3), we set
(σL (−ξ) − Ā)fb(ξ) + B fb(−ξ)
u
b(ξ) = ,
∆ξ
P
and thus u = ξ∈Zn \Λ u b(ξ)eiξ·x is a solution of P u = f .
Note that the sequence {fb(ξ)} decays rapidly, implying the rapid decay
of {bu(ξ)}, and consequently, u ∈ C ∞ (Tn ).
Finally, given the finiteness of Λ, the number of compatibility conditions
on f ∈ C ∞ (Tn ) ensuring the existence of a solution u ∈ C ∞ (Tn ) to P u = f
is finite. Therefore, according to our definition, P is solvable.
4 Kirilov, Moraes and Tokoro

On the other hand, let us assume that (DC) does not hold. Thus, for
each ℓ ∈ N, there exists a ξℓ ∈ Zn such that
kξℓ k > ℓ and |∆ξℓ | < kξℓ k−ℓ .
It is worth noting that kξℓ k → ∞, since kξℓ k > ℓ for all ℓ ∈ N. Denoting
ξℓ = (ξ1ℓ , . . . , ξnℓ ), and by passing to a subsequence if necessary, we may
assume that there is a coordinate j ∈ {1, . . . , n} such that ξjℓ is non-negative
and maintains the same sign, for all ℓ ∈ N.
Now, consider the set Ω = {ξℓ ∈ Zn : ℓ ∈ N}. The previous choice of the
sequence {ξℓ }ℓ∈N implies that, if ξ ∈ Ω, then −ξ ∈ / Ω. Particularly, for any
Ω0 ⊂ Ω, if ξ ∈ Ω0 , then −ξ ∈ / Ω0 .
Case 1: ∆ξℓ = 0 for infinitely many indices ℓ ∈ N.
Upon considering a subsequence, we can assume that ∆ξℓ = 0 for all
ℓ ∈ N. Consequently, by (3), we have

(σL (−ξ) − Ā)fb(ξ) + B fb(−ξ) = 0, ξ ∈ Ω. (6)

In the scenario where B 6= 0, utilizing (4), we obtain
(σL (ξℓ ) − A) · (σL (−ξℓ ) − Ā) = |B|2 , ℓ ∈ N.
As |B|2 6= 0, we can conclude that σL (ξℓ ) − A 6= 0 and σL (−ξℓ ) − Ā 6= 0
for all ℓ ∈ N. In this context, (6) implies the existence of infinitely many
compatibility conditions for the Fourier coefficients of f ∈ C ∞ (Tn ) to satisfy
P u = f . Consequently, P is not solvable according to our definition.
Now, let us consider the case where B = 0. In this scenario, the Fourier
coefficients of u and f must satisfy, for all ξ ∈ Zn ,
u(ξ) = fb(ξ)
(σL (ξ) − A)b u(−ξ) = fb(−ξ)
and (σL (−ξ) − Ā)b
Since
0 = ∆ξℓ = [σL (ξℓ ) − A] · [σL (−ξℓ ) − Ā], ℓ ∈ N
it implies that, for each ℓ ∈ N, at least one of the following is satisfied:
σL (ξℓ ) − A = 0 or σL (−ξℓ ) − Ā = 0,
which further implies that fb(ξ) = 0 or fb(−ξ) = 0, for infinitely many ξ ∈ Zn .
Therefore, in this case, there exist infinitely many compatibility condi-
tions for the Fourier coefficients of f such that P u = f has a smooth solution.
Consequently, P is not solvable.
Case 2: ∆ξℓ = 0 for a finite number of indices ℓ ∈ N.
By passing to a subsequence, we may assume ∆ξℓ 6= 0 for all ℓ ∈ N, i.e.,
0 < |∆ξℓ | < kξℓ k−ℓ , ℓ ∈ N. (7)
Assume B 6= 0 and consider an infinite subset Ω0 ⊂ Ω. In this case,
X
f (x) = ∆ξ eiξ·x
ξ∈Ω0

defines a smooth function f ∈ C (Tn ), due to (7).

∞
 Solvability of Vekua-type periodic operators 5

If u ∈ D ′ (Tn ) is a solution of P u = f , the projection of u on the

subspace of D ′ (Tn ) generated by the frequencies ±Ω0 is
X X
v(x) = (σL (−ξ) − Ā)eiξ·x + Be−iξ·x .
ξ∈Ω0 ξ∈Ω0

In fact, if ξ ∈ Ω0 , then fb(ξ) = ∆ξ and fb(−ξ) = 0. It follows from (3)

that
b(ξ) = [σL (−ξ) − Ā]fb(ξ) = [σL (−ξ) − Ā]∆ξ ⇒ u
∆ξ u b(ξ) = σL (−ξ) − Ā.

If −ξ ∈ Ω0 , then fb(ξ) = 0 and fb(−ξ) = ∆−ξ = ∆ξ . It follows from (3)

that
∆ξ ub(ξ) = B∆ξ ⇒ u b(ξ) = B.
′ n ∞ n
Observe that v ∈ D (T ) \ C (T ), implying u ∈ / C ∞ (Tn ). Since this
construction is valid for any infinite subset Ω0 ⊂ Ω, we obtain infinitely
many linearly independent functions f ∈ C ∞ (Tn ) for which there are no
corresponding u ∈ C ∞ (Tn ) solutions to P u = f . Therefore, P is not solvable.
Now we assume B = 0 and consider an infinite subset Ω0 ⊂ Ω. It follows
from (7) that
|∆ξℓ | = |σL (ξℓ ) − A| · |σL (−ξℓ ) − A| 6 kξℓ k−ℓ , ℓ ∈ N.
This implies that,
|σL (ξℓ ) − A| 6 kξℓ k−ℓ/2 or |σL (−ξℓ ) − A| 6 kξℓ k−ℓ/2 , ℓ ∈ N.
Let us suppose that |σL (ξℓ ) − A| 6 kξℓ k−ℓ/2 for infinitely many ℓ ∈ N.
Passing to a subsequence, if necessary, we may assume that
|σL (ξℓ ) − A| 6 kξℓ k−ℓ/2 , ℓ ∈ N. (8)
It follows from (8) that
X
f (x) = (σL (ξ) − A)eiξ·x
ξ∈Ω0
∞ n
defines a function f ∈ C (T ).
If u ∈ D ′ (Tn ) is a solution of P u = f , the projection of u on the
subspace of D ′ (Tn ) generated by the frequencies Ω0 is
X
v(x) = eiξ·x .
ξ∈Ω0

In fact, if ξ ∈ Ω0 , (3) implies

[σL (ξ) − A] · [σL (−ξ) − Ā] u
b(ξ) = [σL (−ξ) − Ā] · [σL (ξ) − A] .
| {z } | {z }
=∆ξ =fb(ξ)

Hence, ub(ξ) = 1, for any ξ ∈ Ω0 . Thus v ∈ D (T ) \ C ∞ (Tn ), which

′ n

/ C ∞ (Tn ).
implies u ∈
If we have |σL (−ξℓ ) − A| 6 kξℓ k−ℓ/2 for all ℓ ∈ N (passing to a subse-
quence), take f as in the previous case with ξ ∈ −Ω0 .
6 Kirilov, Moraes and Tokoro

Again, observe that this construction is valid for all infinite subsets
Ω0 ⊂ Ω. Then, we obtain infinitely many linearly independent functions
f ∈ C ∞ (Tn ) such that there is no u ∈ C ∞ (Tn ) solution of P u = f . Therefore,
P is not solvable. 
Corollary 3. P is solvable if, and only if, P is globally hypoelliptic.
Proof. Let u ∈ D ′ (Tn ) be a solution to P u = f ∈ C ∞ (Tn ). Since P is
solvable, the condition (DC) holds. Therefore, there exists γ > 0 such that
(σL (−ξ) − Ā)fb(ξ) + B fb(−ξ)
kξk > γ =⇒ u
b(ξ) = .
∆ξ
Consequently, the rapid decay of the sequence (b u(ξ)) is a consequence
b
of the rapid decay of (f (ξ)). This implies that u ∈ C (Tn ) and P is globally
∞

hypoelliptic.
On the other hand, assume that P is not solvable. It follows from The-
orem 2 that there exists f ∈ C ∞ (Tn ) and u ∈ D ′ (Tn ) \ C ∞ (Tn ) such that
P u = f . Therefore, P is not globally hypoelliptic. 
Theorem 4. If L is an elliptic differential operator, then P is solvable and
globally hypoelliptic.
Proof. Given that L − A is elliptic for any A ∈ C, we can assume, without
loss of generality, that P is of the form P u = Lu − B ū, with B ∈ C.
Let m be the order of L. Due to the ellipticity of L, there are positive
constants R0 and M such that
kξk > R0 ⇒ |σL (ξ)| > M kξkm .
For kξk > R0 , we have
|∆ξ | = |σL (ξ) · σL (−ξ) − |B|2 | > M 2 kξk2m − |B|2 .
Choosing R > R0 sufficiently large so that M 2 R2 > |B|2 + 1, we ensure
that kξk > R implies |∆ξ | > 1.
Note that the set Λ = {ξ ∈ Zn : kξk < R} is finite, and, in particular,
Λ0 = {ξ ∈ Zn : ∆ξ = 0} ⊂ Λ is also finite.
Let u ∈ D ′ (Tn ) be a solution to P u = f , and suppose fb(ξ) = 0 for all
ξ ∈ Λ0 . In this case, we have
σL (−ξ)fb(ξ) + B fb(−ξ)
u
b(ξ) = , ξ ∈ Λ0 .
∆ξ
u(ξ)) decays rapidly, resulting in u ∈ C ∞ (Tn ).
Therefore, the sequence (b
It is crucial to note that we have only finitely many compatibility conditions
over f ∈ C ∞ (Tn ) for the equation P u = f to admit a solution u ∈ C ∞ (Tn ).
This finiteness arises from the fact that Λ0 is finite. Consequently, P is solv-
able. The global hypoellipticity is a direct consequence of Corollary 3. 
Pn
Example (Laplace operator). If L = j=1 ∂ 2 /∂x2j then the operator P given
by P u = Lu − Au − B ū, with A, B ∈ C is solvable and globally hypoelliptic.
 Solvability of Vekua-type periodic operators 7

3. Applications
In this section, we present results regarding the solvability and global hypoel-
lipticity of some classical examples of constant-coefficients operators.
P
Theorem 5 (Heat Operator). If L = ∂/∂t − η 2 nj=1 ∂ 2 /∂x2j , where η > 0,
then P is solvable and globally hypoelliptic.
Proof. To prove that the discriminant associated with P satisfies the condi-
tion (DC) of Theorem 2, we begin by noting that
σL (τ, ξ) = iτ + η 2 kξk2 = σL (−τ, −ξ),
for all (τ, ξ) ∈ Zn+1 . By (4) we obtain
∆τ,ξ = η 4 kξk4 − 2Re(A)η 2 kξk2 − τ 2 + |A|2 − |B|2 + 2iτ (η 2 kξk2 − Re(A)).
For τ = 0, there exists γ1 > 0 such that
kξk > γ1 =⇒ |∆0,ξ | >|Re(∆0,ξ )| (9)
4 4 2 2 2 2
= η |ξ| − 2Re(A)η |ξ| + |A| − |B| > 1.
On the other hand, when τ 6= 0, we find
|∆τ,ξ | > |Im(∆τ,ξ )| = 2|τ | η 2 kξk2 − Re(A) .
Note that the condition η 2 |ξ|2 − Re(A) = 0 holds for at most a finite
number of indices (τ, ξ) ∈ Zn+1 . Consequently, there exists γ2 > 0 such that
η 2 kξk2 − Re(A) > 1/2, whenever kξk > γ2 . Thus,
kξk > γ2 =⇒ |∆τ,ξ | > 1. (10)
Now, let us consider the remaining case where (τ, ξ) ∈ Zn+1 satisfies
η 2 kξk2 = Re(A). In this scenario, we have
|∆τ,ξ | > |Re(∆τ,ξ )| = τ 2 + Re(A)2 + |B|2 − |A|2 ,
indicating that |Re(∆τ,ξ )| = 0 if and only if
p
τ = ± −Re(A)2 + |A|2 − |B|2 .
Therefore, |∆τ,ξ | vanishes for at most two integers τ ∈ Z. So, let γ3 > 0
be such that
|τ | > γ3 =⇒ |∆τ,ξ | > τ 2 + Re(A)2 + |B|2 − |A|2 > 1. (11)
.
Define γ = 2 max{γ1 , γ2 , γ3 }. Consequently, when kξk + |τ | > γ, we can
ensure that either kξk > γ/2 or |τ | > γ/2.
For kξk > γ/2, if τ = 0, then |∆0,ξ | ≥ |Re(∆0,ξ )| > 1 since γ ≥ γ1 .
Similarly, for τ 6= 0, we have |∆τ,ξ | ≥ |Im(∆τ,ξ )| > 1 because γ/2 ≥ γ2 .
Now, for |τ | > γ/2, if ηkξk2 − Re(A) = 0, then |∆τ,ξ | > |Re(∆τ,ξ )| > 1
since γ/2 > γ3 . In the case where ηkξk2 − Re(A) 6= 0, we have γ/2 > 0,
implying |τ | > 1. Consequently,
|∆τ,ξ | ≥ |Im(∆τ,ξ )| > 2|η 2 kξk2 − Re(A)| > C > 0,
where C = inf{|η 2 kξk2 − Re(A)| : ξ ∈ Zn and η 2 kξk2 − Re(A) 6= 0}. Notice
that C > 0 because the set {ξ ∈ Zn : η 2 kξk2 − Re(A) = 0} is closed.
8 Kirilov, Moraes and Tokoro

Then, if kξk + |τ | > γ, |∆τ,ξ | > C > 0, implying that (DC) holds. This,
in turn, leads to the conclusion from Theorem 2 that P is solvable. The global
hypoellipticity follows from Corollary 3. 
The hypoellipticity of the heat operator is a well-established result in the
literature. The preceding theorem establishes that the global hypoellipticity
of the heat operator remains stable under both zero-order perturbations and
zero-order conjugate perturbations.
P
Theorem 6 (Wave operator). If L = ∂ 2 /∂t2 − η 2 nj=1 ∂ 2 /∂x2j , where η > 0,
then P is solvable (and globally hypoelliptic) if and only if one of the following
conditions holds:
(i) |B| < |Im(A)|;
(ii) |A| = |B|, Re(A) = 0 and η is an irracional non-Liouville number;
(iii) (DC) holds.
Proof. Let us prove that the discriminant ∆τ,ξ satisfies the condition (DC)
of Theorem 2. First, observe that the symbol
σL (τ, ξ) = σL (−ξ, −τ ) = −τ 2 + η 2 kξk2 ,
for all (τ, ξ) ∈ Zn+1 .
It follows from (4) that
 2
∆τ,ξ = τ 2 − η 2 kξk2 + Re(A) + Im(A)2 − |B|2 , (τ, ξ) ∈ Zn+1 .
Assuming condition (i) holds, we have
|∆τ,ξ | = ∆τ,ξ > Im(A)2 − |B|2 > 0,
for all (τ, ξ) ∈ Zn+1 . Therefore, (DC) holds and P is solvable.
Now, assume that condition (ii) holds, then
|∆τ,ξ | = | − τ 2 + η 2 kξk2 |2 = |τ − ηkξk|2 · |τ + ηkξk|2 .
Given that η is an irrational non-Liouville number, there exist positive
constants C and γ0 such that
τ ± ηkξk > C(kξk + |τ |)−γ0 ,
for all (τ, ξ) ∈ (Zn \ {0}) × Z. Consequently, we obtain
|∆τ,ξ | > C 4 (kξk + |τ |)−γ0 = C 4 (kξk + |τ |)γ0 (kξk + |τ |)−5γ0 .
Take γ1 > 0 such that kξk + |τ | > γ1 implies C 4 (kξk + |τ |)γ0 > 1, and
set γ = max{γ1 , 5γ0 }. Consequently,
kξk + |τ | > γ =⇒ |∆τ,ξ | > (kξk + |τ |)−γ .
Hence, (DC) holds, and P is solvable.
Lastly, if condition (iii) holds, then P is solvable directly.
On the other hand, if none of the conditions (i)-(iii) holds, then, in
particular, (DC) does not hold. Consequently, according to Theorem 2, P
is not solvable. Finally, the global hypoellipticity prevails, as indicated by
Corollary 3. 
 Solvability of Vekua-type periodic operators 9

Finally, we recover the following result concerning complex vector fields

from [4].
Theorem 7 (Complex vector fields). The operator P : C ∞ (T2 ) → C ∞ (T2 )
given by
∂u ∂u
Pu = +C − Au − B̄u,
∂t ∂x
with A, B, C ∈ C, is solvable if and only if one of the following situations
occurs:
(i) |B| > |A|;
(ii) Im(C) 6= 0;
(iii) |B| < |A| andpRe(A) 6= 0;
(iv) The pair (C, |A|2 − |B|2 ) belongs to R2 and satisfies (DC).

References
[1] A. P. Bergamasco, P. D. Cordaro, and G. Petronilho. Global solvability for
a class of complex vector fields on the two-torus. Commun. Partial Differ.
Equations, 29(5-6):785–819, 2004.
[2] A. P. Bergamasco, P. L. Dattori da Silva, and R. B. Gonzalez. Existence and
regularity of periodic solutions to certain first-order partial differential equa-
tions. J. Fourier Anal. Appl., 23(1):65–90, 2017.
[3] A. P. Bergamasco, P. L. Dattori da Silva, R. B. Gonzalez, and A. Kirilov.
Global solvability and global hypoellipticity for a class of complex vector fields
on the 3-torus. J. Pseudo-Differ. Oper. Appl., 6(3):341–360, 2015.
[4] A. P. Bergamasco, P. L. Dattori da Silva, and A. Meziani. Solvability of a first
order differential operator on the two-torus. J. Math. Anal. Appl., 416(1):166–
180, 2014.
[5] M. de Almeida and P. L. Dattori da Silva. Solvability of a class of first order
differential operators on the torus. Result. Math., 76(2):17, 2021. Id/No 104.
[6] B. de Lessa Victor and A. Meziani. Infinitesimal bendings for classes of two-
dimensional surfaces. Complex Variables and Elliptic Equations, 0(0):1–23,
2022.
[7] W. A. A. de Moraes. Regularity of solutions to a Vekua-type equation on
compact Lie groups. Ann. Mat. Pura Appl. (4), 201(1):379–401, 2021.
[8] J. Hounie. Globally hypoelliptic and globally solvable first-order evolution
equations. Trans. Amer. Math. Soc., 252:233–248, 1979.
[9] A. Kirilov, W. A. A. de Moraes, and M. Ruzhansky. Partial Fourier series on
compact Lie groups. Bull. Sci. Math., 160:27, 2020. Id/No 102853.
[10] A. Kirilov, W. A. A. de Moraes, and M. Ruzhansky. Global hypoellipticity
and global solvability for vector fields on compact Lie groups. J. Funct. Anal.,
280(2):39, 2021. Id/No 108806.
[11] V. V. Kravchenko. Applied pseudoanalytic function theory. Front. Math. Basel:
Birkhäuser, 2009.
[12] A. Meziani. Solvability of planar complex vector fields with applications to
deformation of surfaces. In Complex analysis, Trends Math., pages 263–278.
Birkhäuser/Springer Basel AG, Basel, 2010.
10 Kirilov, Moraes and Tokoro

[13] A. Meziani. Nonrigidity of a class of two dimensional surfaces with positive

curvature and planar points. Proc. Amer. Math. Soc., 141(6):2137–2143, 2013.
[14] G. Petronilho. Global hypoellipticity, global solvability and normal form for a
class of real vector fields on a torus and application. Trans. Am. Math. Soc.,
363(12):6337–6349, 2011.
[15] I. N. Vekua. Generalized analytic functions. Transl. ed. by Ian N. Sneddon.
International Series of Monographs on Pure and Applied Mathematics. Vol.
25. Oxford-London-New York-Paris: Pergamon Press 1962; Reading, Mass.-
London: Addison-Wesley Publ. Co., Inc. xxvi, 668 p. (1962)., 1962.

Alexandre Kirilov
e-mail: akirilov@ufpr.br
Departamento de Matemática
Universidade Federal do Paraná
CP 19096, CEP 81531-990, Curitiba
Brasil
Wagner Augusto Almeida de Moraes
Departamento de Matemática
Universidade Federal do Paraná
Caixa Postal 19096
CEP 81531-990, Curitiba, Paraná
Brasil
e-mail: wagnermoraes@ufpr.br
Pedro Meyer Tokoro
Programa de Pós-Graduação em Matemática
Universidade Federal do Paraná
Caixa Postal 19096
CEP 81531-990, Curitiba, Paraná
Brasil
e-mail: tokoro.p@gmail.com