Merzon et al.

Boundary Value Problems (2025) 2025:26

https://doi.org/10.1186/s13661-025-02022-w

RESEARCH Open Access

Model boundary value problem for the

Helmholtz equation in a nonconvex angle
with periodic boundary data
A. Merzon1* , P. Zhevandrov2 , M.I. Romero Rodríguez3 and J.E. De la Paz Méndez4
*
Correspondence:
anatolimx@gmail.com Abstract
1
Instituto de Física y Matemáticas,
Universidad Michoacana, Morelia, In the present work, we solve the Dirichlet boundary problem for the Helmholtz
Michoacán, México equation in an exterior angle with periodic boundary data. We prove the existence
Full list of author information is and uniqueness of solution in an appropriate functional class and give an explicit
available at the end of the article
formula for it in the form of the Sommerfeld integral. The method of complex
characteristics (Komech and Merzon in Stationary Diffraction by Wedges, Lect. Notes
Math., vol. 2249, Springer, Berlin, 2019) is used.
Keywords: Helmholtz equation; Nonconvex angles; Periodic boundary conditions

1 Introduction
We consider the following model boundary value problem (BVP) for the Helmholtz equa-
{︁ }︁
tion in a plane angle Q of magnitude Φ > π with a complex frequency ω ∈ C+ = Im ω > 0 :
⎧
⎨ (–Δ – ω )u(x) = 0, x ∈ Q
⎪ 2

⃓ (1.1)
⎪
⎩ ⃓
Bl u(x)⃓ = fl (x), l = 1, 2.
Ql

{︁ }︁
Here Ql for l = 1, 2 are the sides of the angle Q = Φ < θ < 2π , θ is the polar angle, φ =
∂
2π – Φ, Bl = I or Bl = (nl is the exterior normal to Ql ), fl are given functions, which
∂nl
can be distributions, see Fig. 1.
Problems of this type arise in many areas of mathematical physics. We list some of them.
Firstly, such BVPs describe the diffusion of a desintegrating gas [40, Ch. VII, §1].
Secondly, diffraction problems by the wedge W = R2 \ Q are reduced to such problems
for a lossy medium [39] or for a slightly conducting medium [8].
Thirdly, time-dependent diffraction by wedges [2, 10, 11, 28, 32, 35–38] is reduced to
a problem of this type after the Fourier–Laplace (F–L) transform t → ω with respect to
time [9, 13, 15, 27].
Let us describe this scattering problem in more detail since it was precisely this problem
which was the starting point of the present paper.

© The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives
4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format,
as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and
indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived
from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/
4.0/.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 2 of 44

Figure 1 BVP in exterior angle

Figure 2 Point source

Let us consider the following time-dependent scattering problem:

⃓
Utt (y, t) = ΔU(y, t) + δ(y – y∗ )e–iω0 t , y ∈ Q, t ≥ 0, ω0 ≥ 0 ⃓⃓
⃓
⃓ ⃓ (1.2)
⃓ ⃓
U(y, t)⃓ = 0. ⃓
∂Q

Here y∗ ∈ Q. After the F–L transform t → ω : U(y, t) −→ Û(y, ω) (1.2) becomes equivalent
to
⃓
i ⃓
2
–ω Û(y, ω) = ΔÛ(y, ω) + δ(y – y ) ∗
, ω0 > 0 ⃓⃓
ω – ω0 ⃓ ω ∈ C+
⃓ ⃓ (1.3)
⃓ ⃓
Û(y, ω)⃓ = 0, ⃓
∂Q

see Fig. 2.
We reduce this problem to a homogeneous Helmholtz equation with nonhomogeneous
boundary conditions. Let E (y, ω) be s.t.

iδ(y – y∗ )
–ΔE (y, ω) – ω2 E (y, ω) = , y ∈ R2 , Im ω > 0, ω0 > 0. (1.4)
ω – ω0

Passing to the Fourier transform y → η in (1.4), we obtain

∗
ieiη·y
E˜ (η, ω) = .
(|η| – ω2 )(ω – ω0 )
2

Hence,

∫︂ ∗
i e–iη(y–y )
E (y, ω) = dη.
(ω – ω0 )(2π)2 |η|2 – ω2
R2
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 3 of 44

Define

Us (y, ω) := Û(y, ω) – E (y, ω), y ∈ Q, ω ∈ C+ .

By (1.3), (1.4) this implies that Us (y, ω) satisfies a problem of type (1.1):
⎧ ⃓
⃓
⎨ –ΔUs (y, ω) – ω Us (y, ω) = 0, y ∈ Q ⃓
⎪ 2
⃓
⃓ ⃓ ⃓ ω ∈ C+ . (1.5)
⎪ ⃓
⎩ Us (y, ω)⃓ = –E (y, ω)⃓ ⃓ ⃓
⃓
∂Q ∂Q

Let us calculate E (y, ω) on ∂Q. We have

⃓ ∫︂
⃓
–E (y, ω)⃓ = –E (y1 , 0, ω) = C(ω) e–iη1 y1 C1 (η, ω, y∗ ) dη,
Q1
R2

eiη·y
∗ ⃓
i ⃓
where C(ω) = ; C1 (η, ω, y∗ ) = . Hence, –E (y, ω)⃓ =
∫︂ (ω – ω0 )(2π) 2 |η|2 – ω2 Q2

C(ω) e–i(η1 cot Φ+η2 )y2 C1 (η, ω, y∗ ) dη. Thus (1.5) is equivalent to
R2

⎧
⎪
⎪ –ΔUs (y, ω) – ω2 Us (y, ω) = 0, y ∈ Q
⎪
⎪
⎪
⎪ ∫︂
⎪
⎪
⎨ U (y , 0, ω) = C(ω) e–iη1 y1 C (η, ω, y∗ ) dη, y > 0
s 1 1 1
(1.6)
⎪
⎪ R2
∫︂
⎪
⎪
⎪
⎪ e–i(η1 cot Φ+η2 )y2 C1 (η, ω, y∗ ) dη,
⎪
⎪ U (y
s 2 cot φ, y2 , ω) = C(ω) y2 > 0.
⎩
R2

Therefore, it seems natural to solve first the following model problem, corresponding to
problem (1.6). In polar coordinates y1 = ρ cos θ , y2 = ρ sin θ , this problem takes the form
⎧
⎨ (–Δ – ω2 )U(ρ, θ , ω) = 0, (ρ, θ ) ∈ Q
(1.7)
⎩ –ik1 ρ –ik2 ρ
U(ρ, 2π) = e , U(ρ, 2π – Φ) = e ,

where k1 , k2 ∈ R. Without loss of generality, we can assume that

k1 > k2 > 0. (1.8)

Obviously, to solve problem (1.7) it suffices, by linearity, to consider the following two
problems with homogeneous boundary conditions on Q1 or Q2 :
⎧
⎨ –Δu1 (ρ, θ ) – ω2 u1 (ρ, θ ) = 0, (ρ, θ ) ∈ Q
(1.9)
⎩ –ik1 ρ
u1 (ρ, 2π) = e , u1 (ρ, 2π – Φ) = 0, ρ>0

and
⎧
⎨ –Δu2 (ρ, θ ) – ω2 u2 (ρ, θ ) = 0, (ρ, θ ) ∈ Q
(1.10)
⎩ –ik2 ρ
u2 (ρ, 2π) = 0 , u2 (ρ, 2π – Φ) = e , ρ > 0.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 4 of 44

Moreover, if we can solve the problem (1.9) with any k1 , the solution of the problem (1.10)
is written as

u2 (ρ, θ ) = u1 (ρ, –θ + 4π – Φ), (1.11)

where u1 is a solution of (1.9) with k1 = k2 . Then the solution of (1.7) is the sum of u1 and
u2 ,

U(ρ, θ ) = u1 (ρ, θ ) + u2 (ρ, θ ). (1.12)

Clearly, the solutions of problems (1.9) and (1.10) have worse regularity properties than
the solution of (1.7), since the boundary conditions in (1.9)–(1.10) do not satisfy the com-
patibility condition U(0, 2π) = U(0, 2π – Φ) as in (1.7) (for example, the derivatives of u1
have singularities of the type 1/ρ and hence their traces on the boundary are not locally
integrable). Nevertheless, the final formula (1.12) for the solution of (1.7) does not have
these singularities. The procedure of splitting of (1.7) into (1.9) and (1.10) is due to the
technical advantage of the latter problems from the point of view of application of the
method of automorphic functions (MAF), see below.
Note that this problem is similar to the BVP in [13, (23)], arising in the problem of time-
dependent diffraction of a plane wave incident on the wedge W of magnitude φ < π/2 at
an angle α < φ.
⎧ (︁ )︁
⎪
⎨ – Δ – ω2 U(ω, y) = 0, y∈Q
⎪ (︁ )︁ (1.13)
⎩ U(ω, y) = –g(ω)eiωy1 cos α , y ∈ Q –iωy2 cos(α+Φ)
U(ω, y) = –g(ω)e sin Φ , y ∈ Q2 .
1

However, there are substantial differences. The exponents on the right-hand side of this
problem are complex (Im ω > 0) and, thus, the corresponding functions decrease expo-
nentially when y1,2 −→ +∞. In contrast, problem (1.7) has periodic boundary conditions
that are independent of the first equation. This results in the fact that the corresponding
difference equation cannot be solved as easily as in the previous case, except for the case
when Φ = 3π/2, when the solution of the corresponding difference equation is easy to
guess (see Sect. 5).
This paper is devoted to solving the model problem (1.7). However the main technical
part of the paper relates to solving problem (1.9). This is due to the fact that it is technically
easier to use MAF for this problem. We obtain an explicit solution of problem (1.9) and
we calculate its asymptotics at y = 0. Then, using (1.12) we obtain a solution of (1.7) and
we prove its uniqueness in a certain functional class.
Note that the BVP in a right angle Q or in its complement and in other particular angles
whose magnitudes are commensurate with π were considered in many papers [3–7, 22–
26, 33].
In those papers, exact results were obtained by means of operator methods. Bound-
ary data in those papers belong to Sobolev spaces Hs (R), s > 0. We consider another
type of boundary data, namely, periodic functions. We obtain exact solutions in explicit
form, namely, in the form of Sommerfeld type integrals. We use the method of automor-
phic functions (MAF) on complex characteristics [17]. This method was developed by
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 5 of 44

A. Komech for Φ < π in [12] and then was extended to Φ > π in [17, Sect. 1.2 and Part 2].
It allows us to find all distributional solutions of the BVPs for the Helmholtz equation at
arbitrary angles with general boundary conditions. It was applied, in particular, to time-
dependent diffraction problems by angles in [16, 18, 19, 27, 29, 30].
It should be noted that there is a very effective Sommerfeld–Malyzhinetz method of
constructing solutions of diffraction problems in angles; by means of this method many
important results were obtained [1]. This method allows one to obtain the solution in
the form of the Sommerfeld integral. However, this method does not allow one to prove
uniqueness which usually is proved on the basis of physical considerations. We also obtain
solution in the Sommerfeld integral form, using the MAF, which also allows us to prove
uniqueness in an appropriate functional space (see, e.g., [22]).
The choice of this functional space is nontrivial because, as already noted, the solution
to (1.9) has singularities at the origin and we do not know whether the solution that we
construct is unique in the corresponding class of functions possessing these singularities.
The MAF that we use for the construction of the solution to (1.9) turns out not to be
applicable for the proof of uniqueness because the continuation by zero of functions pos-
sessing such singularities is nonunique. Problem (1.7), in contrast, does not present such
difficulties (see the Appendix) because of the compatibility conditions at the origin.
In paper [20], the Wiener–Hopf method is used for the problem of a penetrable wedge
Note that in some sense the Wiener–Hopf method resembles MAF because it also relies
on the double Fourier transform and on the analytic continuation of the solution to some
complex region.
The paper is organized as follows: in Sect. 2, we formulate the main result, Sects. 3–9 are
devoted to its proof. In Sect. 3, we reduce boundary value problem to a difference equation
and we prove the necessary and sufficient conditions for the existence of its solution. In
Sects. 4 and 5, we find the solution of the difference equation for Φ ̸= 3π/2 and Φ = 3π/2,
respectively. In Sects. 6 and 7, we solve problem (1.9). In Sect. 8, we study the properties
of the solution to (1.7). In Sect. 9, we prove the uniqueness of solution. In the Appendix,
we prove some technical assertions.

2 Main result
We will construct a unique (in some class) solution of (1.7) in the form (1.12), where the
solution u1 has the form of the well-known Sommerfeld integral

∫︂
e–ωρ sinh w v(w + iθ ) dw, (2.1)
C

where C is a certain contour on the complex plane and the integrand v is such that (2.1)
satisfies the boundary conditions in (1.9).
To formulate the main result, we need to describe the integrand v(w) of the Sommerfeld
integral. The construction of this integrand is the main contents of this paper and the main
difficulty of the problem.
Consider v̂1 (w) given by (3.21), where v̂11 (w) is given by (4.27) for Φ ̸= 3π/2 and by (5.13)
for Φ = 3π/2 with Ĝ given by (3.21).
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 6 of 44

(︁ )︁
Definition 2.1 For α ∈ (0, 1), let Eα be the space of C(Q) ∩ C 1 Q \ {0} -functions u(ρ, θ )
bounded with their first derivatives in Q \ {0} and satisfying

|u(ρ, θ )| ≤ C, |∇u(ρ, θ )| ≤ C1 ρ –1+α + C2 , (ρ, θ ) ∈ Q \ {0}. (2.2)

Our main result consists in the following statement.

Theorem 2.2 Let ω ∈ C+ , k1 > k2 > 0. There exists a solution U(ρ, θ ), (ρ, θ ) ∈ Q to problem
(1.7), belonging to Eα with α = π/Φ given by (1.12), where u1 admits a Sommerfeld integral
representation,
∫︂
1
u1 (ρ, θ ) = e–ωρ sinh w v̂1 (w + iθ )dw, (2.3)
4π sin Φ
C

and v̂1 is constructed according to the algorithm presented below. The function u2 (ρ, θ ) is
given by (1.11) with v̂1 constructed according to the same algorithm with k1 replaced by k2 .
Moreover, the solution U is unique in Eα , α ∈ (0, 1). The contour C in (2.3) is defined by
(6.1).

Remark 2.3 The integral in (2.3) converges absolutely for ρ > 0 and ω ∈ C+ , since the
infinite part of C belongs to the region of the superexponential decrease of e–ωρ sinh w (see
Fig. 11) and v̂1 admits expansion (8.1). The boundary values are understood in the sense
of distributions. Note that, by (1.11), the function u2 is given by
∫︂
1 (︁ )︁
u2 (ρ, θ ) = e–ωρ sinh w v̂1 w + (–iθ + 4π – Φ) dw, (2.4)
4π sin Φ
C

where, as mentioned above, vˆ1 is constructed in the same way as in (2.3) by replacing k1
by k2 . To simplify the notation we omit the dependence of v1 on k1 and k2 .

Remark 2.4 Note that the solution u1 also admits a slightly different representation, where
several different Sommerfeld-type contours are used (see (7.4)).

3 Reduction to a difference equation. Necessary and sufficient conditions for

the Neumann data
Consider problem (1.9). The MAF permits us to reduce this problem to finding the Neu-
mann data of solution u1 , and it consists of several steps. In the following subsections we
present these steps. {︂ ⃓ }︂
⃓
We assume that the solution u ∈ S′ (Q) := u⃓ , u ∈ S′ (R2 ) . The first step of the MAF is
Q
to reduce the problem to the complement of the first quadrant and to extend the solution
u to the plane, see [15, 17].

3.1 First step: extension of the Cauchy data to the whole plane R2
Consider the linear transformation

y2
L(y) : x1 = y1 + y2 cot Φ, x2 = – , (3.1)
sin Φ
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 7 of 44

which sends the angle Q to the right angle K , which is the complement of the first quadrant
{x1 ≥ 0, x1 ≥ 0}. This transformation reduces system (1.9) to problem (3.3a)–(3.3c) in K
for

(︁ )︁
v(x) := u1 L–1 (y) (3.2)
⎧
⎪
⎪ H (D)v(x) = 0, x∈K (3.3a)
⎪
⎨
v(x1 , 0) = e–ik1 x1 , x1 > 0 (3.3b)
⎪
⎪
⎪
⎩
v(0, x2 ) = 0, x2 > 0, (3.3c)

where

1 [︂ ∂ 2 ]︂
H (D) = – 2
Δ – 2 cos Φ – ω2 . (3.4)
sin Φ ∂x1 ∂x2

The next lemma shows that in order to obtain a solution of (3.3a) it is sufficient to solve
the following problem on the whole plane R2 (see [17], Lemma 8.4).

Lemma 3.1 Let v0 (x) ∈ S ′ (K) and

H (D)v0 (x) = γ (x), x ∈ R2 , (3.5)

where

1 [︂
γ (x) = δ(x2 )v11 (x1 ) + δ ′ (x2 )v01 (x1 ) + δ(x1 )v12 (x2 ) + δ ′ (x1 )v02 (x2 )
sin2 Φ (3.6)
]︂
– 2δ(x2 ) cos Φ ∂x1 v01 (x1 ) – 2δ(x1 ) cos Φ ∂x2 v02 (x2 ) ,

{︂ }︂
vl (xl ) ∈ S ′ (R+ ) := v ∈ S ′ (R) : supp v ⊂ R+ . Then
β

⃓
⃓
v(x) := v0 (x)⃓ ∈ C ∞ (K),
K

⃓
β ⃓
v(x) satisfies (3.3a), and vl (xl )⃓ are the Cauchy data of the function v:
R+

β
v1 (x1 ) = ∂x2 v(x1 , 0–), x1 > 0
(3.7)
β
v2 (x2 ) = ∂x1 v(0–, x2 ), x2 > 0.

The limits here are understood in the sense of S ′ (R+ ).

We will use the extension of the Fourier transform F defined on S(R) ⊂ S′ (R2 ),
ϕ(x1 , x2 ) → ϕ̃(z1 , z2 ), ϕ ∈ S(R2 ) to S′ (R2 ) by continuity:

∫︂∫︂
[︁ ]︁ [︁ ]︁
Fx→z ϕ (z) = F ϕ(x) (z) = ϕ̃(z1 , z2 ) := eiz1 x1 +iz2 x2 ϕ(x1 , x2 ) dx1 dx2 , (3.8)
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 8 of 44

and denote this extension by the tilde, ṽ(z) = Fx→z [v(x)], v ∈ S′ (R2 ). Applying this trans-
form to (3.5) and using the fact that H (z) ̸= 0, z ∈ R2 , we obtain

γ̃ (z)
ṽ0 (z) = , z ∈ R2 , ṽ0 ∈ S′ (R2 ). (3.9)
H (z)

Hence,
[︄ ]︄
⃓
γ̃ (z) ⃓
–1
v0 (x) = Fx→z , x ∈ R2 , v(x) = v0 (x)⃓ . (3.10)
H (z) K

Here γ̃ (z) is the Fourier transform of (3.6), and for z ∈ R2

1 [︂ 1 (︁ )︁ (︁ )︁]︂
γ̃ (z) = 2
ṽ1 (z1 ) – ṽ01 (z1 ) iz2 – 2 cos Φ iz1 + ṽ12 (z2 ) – ṽ02 (z2 ) iz1 – 2 cos Φ iz2 ,
sin Φ
(3.11)

β β β
where ṽl (zl ) are the Fourier transforms of vl (xl ). Thus, if we know vl (xl ), we know v by
β
(3.10), and problem (3.3a)–(3.3c) is reduced to finding the four functions ṽl (xl ), l = 1, 2, β =
0, 1. The representation (3.6) is easily obtained if the solution of (3.3a) belongs to C ∞ (K). In
our case, the solution to this problem does not belong to C ∞ (K) (it belongs only to C ∞ (K \
0)). However, it turns out that formula (3.6) remains true for distributional solutions.
β
As noted, in order to construct the solution we need to know the Cauchy data vl . Obvi-
0
ously, the boundary condition (3.3b) and (3.3c) provide the values of vl (up to δ-functions
supported at the origin). Indeed, let v ∈ S′ (K) be a solution to (3.3a)–(3.3c) and v0 ∈ S ′ (K)
be its extension by 0 satisfying (3.5); then, by (3.7) and Lemma 3.1,
⃓ ⃓
⃓ ⃓
v01 (x1 )⃓ = e–ik1 x1 , x1 > 0, v02 (x2 )⃓ = 0, x2 > 0.
R+ R+

Since supp v0l (xl ) ⊂ R+ , by the distribution theory, we have, generally speaking,

∑︁ j (j) ∑︁m j (j)

v01 (x1 ) = e–ik1 x1 Θ(x) + j=0m c1 δ (x1 ), v02 (x2 ) = j=0 c2 δ (x2 )

for some m ≥ 0. Here, Θ is the Heaviside function. We will find a solution to (3.3a)–(3.3c)
j j
for c1 = c2 = 0, j = 0, 1, . . . m. Thus, we put

v01 (x1 ) = e–ik1 x1 Θ(x) ∈ S′ (R), v02 (x2 ) = 0. (3.12)

Substituting these Dirichlet data v01 , v02 in (3.5), we obtain

H (D)v0 (x) = γ (x), (3.13)

with γ containing only two unknown functions v11 and v12 (Neumann data).
The MAF gives the necessary and sufficient conditions for finding the functions v11 and
1
v2 in an explicit form. Substituting these functions in (3.13), we obtain v0 (and so v) by
(3.10), (3.6). In what follows we consider equation (3.13).
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 9 of 44

3.2 Second step: Fourier–Laplace transform and the lifting to the Riemann
surface. Connection equation
In addition to the (real) Fourier transform (3.8) we will use the complex Fourier transform
(or Fourier–Laplace (F–L) transform). Let
{︂ }︂
f ∈ S′ (R+ ) := f ∈ S′ (R) : supp f ⊂ R+ .

(︁ )︁ [︁ ]︁
Then by the Paley–Wiener theorem [34] see also [10, Theorem 5.2] , ˜ (z1 ) = F f (z1 ), z1 ∈
f
(︁ )︁ {︂ }︂
R, admits an analytic continuation f˜ (z) ∈ H C+ , C+ := z ∈ C : Im z > 0 and lim f˜ (z1 +
iτ1 ) = f˜ (z1 ) in S′ (R), as τ1 → 0+. Since vl (xl ) ∈ S′ (R+ ), there exist their F–L transforms
β

β (︁ )︁
ṽl (zl ) ∈ H C+ , l = 1, 2; β = 0, 1. (3.14)

In particular, from (3.12), we have

i
ṽ01 (z1 ) = , z1 ∈ C+ , (3.15)
z1 – k1

where for z1 ∈ R, ṽ01 (z1 ) = lim ṽ01 (z1 + iτ1 ) in S′ (R). Hence, using (3.11) we obtain (since
τ1 →0+
v02 ≡ 0)
[︄ ]︄
1 z2 – 2 cos Φ z1
γ̃ (z) = 2 1
ṽ (z1 ) + 1
+ ṽ2 (z2 ) , z ∈ R2 . (3.16)
sin Φ 1 z1 – k1

In the MAF, the Riemann surface of complex zeros of the symbol of the operator (3.4)
plays an essential role, since a necessary condition for γ̃ (z), ensuring the existence of the
solution, can be written in terms of this surface. The symbol of operator (3.4) is the poly-
nomial

1 (︂ 2 2 )︂
H (z) = 2
z1 + z2 – 2z1 z2 cos Φ – ω2 , (z1 , z2 ) ∈ C2 .
sin Φ

Obviously, H (z) does not have real zeros, but it does have complex ones. Denote the
Riemann surface of the complex zeros of H by
{︂ }︂
V := z ∈ C2 : H (z) = 0 .

It is convenient to parameterize the complex surface V introducing the parameter w ∈ C.

The Riemann surface V admits a universal covering V̂ , which is isomorphic to C (see
[17, Ch. 15]). Let w be a parameter on V̂ ∼
= C. Then the formulas

z1 = z1 (w) = –iω sinh w, z2 = z2 (w) = –iω sinh(w + iΦ) (3.17)

describe an infinitely sheeted covering of V . {︂

β
Let us “lift” the functions ṽl (zl ), zl ∈ C+ to V̂ . For this we must identify Vl+ := z ∈ C2 :
}︂
Im zl > 0 with regions on V̂ . This can be done in several ways. For example, define, for
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 10 of 44

ω ∈ C+ ,
{︂ (︂ ω )︂⃓ }︂
1 ⃓
Γ0 = Γ0 (ω) := w1 + i arctan tanh w1 ⃓w1 ∈ R . (3.18)
ω2

(︁ )︁ (︂ ω )︂ ω1
1
Obviously, for w ∈ Γ0 , Im z1 (w) = 0. Moreover, arctan tanh w1 −−−−−→ ± arctan .
ω2 w1 →±∞ ω2
For ν ∈ R, define

Γν = Γν (w) := Γ0 (w) + iν

and for δ < ν, define the strip lying between the curves Γδ and Γν :
{︄ }︄
(︂ ω )︂ (︂ ω )︂
1 1
V̂δν := w ∈ C : arctan tanh w1 + δ < Im w < arctan tanh w1 + ν .
ω2 ω2

{︂ }︂
For l = 1, 2, let us “lift” Vl+ to V̂ . Denote this lifting by V̂l+ = w ∈ V̂ : zl (w) ∈ Vl+ . Then

∞
⋃︂ ∞
⋃︂
V̂1+ = (2k+1)π
V2kπ , V̂1– = 2kπ
V(2k+1)π , V2± = V1± – 2iΦ.
k=–∞ k=–∞

(︁ )︁
Note that ±Im zl (ω, w) > 0 for w ∈ V̂l± . We choose the connected component of V̂l+ cor-
responding to the condition Im zl > 0 as V̂1+ := V0π , V̂2+ = V–Φ
π –Φ
(see Fig. 3, where all the
contours Γν (w) are shown for ω1 ≥ 0).

Figure 3 Connection equation

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 11 of 44

β
Now we “lift” ṽl (zl ) to V̂l+ ,l = 1, 2, β = 0, 1, using (3.17). We obtain from (3.15), (3.12)

i
v̂01 (w) = , w ∈ V̂1+ , v̂02 (w) = 0, w ∈ V̂2+ . (3.19)
–iω sinh w – k1

Further, v̂1l (w) are analytic functions in V̂l+ by (3.14). Our aim is to find them. Having found
these functions, we obtain γ̃ (z) and the solution v0 (x) by (3.10), (3.9).
Note that in the case Φ > π the function γ̃ (z) given by (3.16) cannot be lifted to V̂ since
(︁ )︁
γ̂ (w) := γ̃ z1 (w), z2 (w) is not defined at any point of V̂ . In fact, v̂11 (w) is not defined in
V̂2+ , v̂12 (w) is not defined in V̂1+ since V̂ ∗ := V̂1+ ∩ V̂2+ = ∅, see Fig. 3. In the case Φ < π this
intersection is not empty and such lifting to V̂ ∗ is possible [17]. Thus, in that case there
exists a connection between v̂11 and v̂12 generated by (3.9), since H (z) has zeros in V̂ ∗ and
γ̂ (w) must vanish for w ∈ V̂ ∗ .
Nevertheless, a similar relation between v̂11 and v̂12 exists in the case Φ > π too (see [17,
Chap. 21]). Let us describe the corresponding construction. The function γ̂ (z) is naturally
splitted into two summands, each of which is extended to V̂1+ and V̂2+ , respectively. Namely,

sin2 Φ γ̃ (z1 , z2 ) = ṽ1 (z1 , z2 ) + ṽ2 (z1 , z2 ),

where

z2 – 2 cos Φ z1
v1 (z1 , z2 ) := ṽ11 (z1 ) + , v2 (z1 , z2 ) := ṽ12 (z2 ), (z1 , z2 ) ∈ R2 .
z1 – k1

By the Paley–Wiener theorem, the function ṽ1 (z1 , z2 ) admits an analytic continuation{︂ ⃓ to
C+z1 × C, and ṽ2 (z1 , z2 ) admits an analytic continuation to C+ × C+z2 , where C+zl = zl ⃓ Im zl >
}︂
0 . Now we can “lift” ṽ1 and ṽ2 to the Riemann surface V̂ by formulas (3.17).
We obtain

v̂1 (w) = v̂11 (w) + ω sinh(w – iΦ) v̂01 (w), w ∈ V̂1+ , v̂2 (w) = v̂12 (w), w ∈ V̂2+ . (3.20)

Then, by (3.20), (3.19),

iω sinh(w – iΦ)
v̂1 (w) = v̂11 (w) – Ĝ(w), w ∈ V̂1+ , Ĝ(w) := , w ∈ C. (3.21)
iω sinh w + k1

In the case Φ < π , v̂1 (w) and v̂2 (w) have a common domain V̂ ∗ , which is not empty, and
thus the connection equation has the following form:

v̂1 (w) + v̂2 (w) = 0, w ∈ V̂ ∗ . (3.22)

(see [17, Chap. 10]).

In the case Φ > π , the domain V̂ ∗ = ∅ (see Fig. 3). Nevertheless, it turns out that in
this case, there exists a connection between v̂1 and v̂2 such that (3.22) holds in a slightly
different sense, namely this equation holds for analytic continuations of v̂1 and v̂2 . Let us
formulate the corresponding theorem.

Definition 3.2 Denote V̂Σ := V̂1+ ∪ V̂2+ ∪ V̂ ∗ = V̂–Φ

π
, where V̂ ∗ = Vπ0–Φ ; V̂Σ,δ := V̂–Φ+δ
π –δ
. (see
Fig. 3).
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 12 of 44

Theorem 3.3 (Connection equation in the case Φ > π ) [14, Sect. 6.3]. Let v ∈ S′ (K) be any
distributional solution to (3.3a). Then functions (3.20) admit analytic continuations [v̂l ]
along the Riemann surface V̂ from V̂l+ to V̂Σ (see Fig. 3) and
(a)

[︁ ]︁ [︁ ]︁
v̂1 (w) + v̂2 (w) = 0, w ∈ V̂Σ , (3.23)

(b)

(︁ )︁q
|[v̂1 (w)]| ≤ Cδ 1 + e|w| , w ∈ VΣ,δ , δ ∈ (0, Φ/2 + π/2). (3.24)

Remark 3.4 Using the connection equation (3.23) we will find v̂1 . The solution u1 of prob-
lem (1.9) is then given by (2.3) (see [13, Theorem 9.1]). In Sects. 7 and 8, we will prove the
Sommerfeld representation (2.3) directly.

3.3 Third step: reduction to a difference equation

From (3.23), (3.21) it follows that v̂11 (w) and v̂12 (w) admit meromorphic continuations to
V̂Σ and

v̂11 (w) + v̂12 (w) = Ĝ(w), w ∈ V̂Σ . (3.25)

(︂ )︂
We will use the following automorphisms on V̂ see [17, Ch. 13] and [13, (73)] :

h1 w = –w + πi, h2 w = –w + πi – 2iΦ, w ∈ C, (3.26)

π π
which are symmetries with respect to i and i – iΦ, respectively.
2 2 (︂ )︂
Sometimes, we will use the notation f hl (w) := f hl (w) , l = 1, 2.
The functions v̂11 and v̂12 are automorphic functions with respect to h1 and h2 , respec-
tively,

v̂1h1 1 1
1 (w) = v̂1 (–w + πi) = v̂1 (w), w ∈ V̂1+ , (3.27)

v̂1h2 1 1
2 (w) = v̂2 (–w + πi – 2iΦ) = v̂2 (w), w ∈ V̂2+ , (3.28)

as follows from the fact that ṽ1l (zl ) depend only on zl and hence their liftings v̂l (w) to V̂l+
satisfy (3.27), (3.28), since sinh w satisfies (3.27) and sinh(w + iΦ) satisfies (3.28).
Due to this automorphy we can eliminate one unknown function in the undetermined
equation (3.25) and reduce it to an equation with a shift, see [13]. The idea of this technique
is due to Malyshev [21].

Lemma 3.5 Let v ∈ S′ (R2 ) satisfy (3.3a)-(3.3c) and v1l (xl ), l = 1, 2, be its Neumann data.
Then their liftings to V̂ , v̂1l (w), w ∈ V̂l+ , admit meromorphic continuations to C (which we
also denote v̂1l ) such that for w ∈ C

v̂11 (w) + v̂12 (w) = Ĝ(w) (3.29)

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 13 of 44

holds and v11,2 are hl -automorphic functions,

(︁ )︁ (︁ )︁
v̂11 h1 (w) = v̂11 (w), v̂12 h2 (w) = v̂12 (–w + πi – 2iΦ) = v̂12 (w). (3.30)

The proof of this lemma is given in Appendix A.1.

Now we reduce system (3.29)–(3.30) to a difference (or shift) equation. This reduction
is the part of MAF which was introduced in [21] for difference equations in angles. It uses
β
the automorphy of v̂l on V̂ under the automorphisms hl and the term MAF is due to this
observation.
Define, for w ∈ C,

(︂ )︂ iω sinh(w – iΦ) iω sinh(w + 3iΦ)

Ĝ2 (w) := Ĝ(w) – Ĝ h2 (w) = – . (3.31)
iω sinh w + k1 iω sinh(w + 2iΦ) + k1

Here and everywhere below, for a region U in C, we will denote by M(Ω) the set of mero-
morphic functions on Ω.

Lemma 3.6 Let v ∈ S′ (K) satisfy (3.3a)-(3.3b). Then the connection equation (3.23) holds,
the function v̂11 belongs to M(C) ∩ H(V̂1+ ), satisfies the difference equation

v̂11 (w) – v̂11 (w + 2iΦ) = Ĝ2 (w), (3.32)

and the automorphy condition (3.30).

The proof of this lemma is given in Appendix A.2.

(︁ )︁
Our goal is to find v̂11 (w) ∈ M(C) ∩ H V̂1+ s.t. (3.32), (3.30) and the condition v̂1 (w) ∈
(︁ )︁
H V̂Σ hold.
Here v̂1 is given by (3.20). In its turn, this condition is equivalent to the condition

(︁ )︁
v̂1 (w) = v̂11 (w) – Ĝ(w) ∈ H V̂Σ (3.33)

by (3.21).
In the next section, we find the necessary and sufficient conditions for v̂11 such that con-
dition (3.33) holds.

3.4 Necessary and sufficient condition for v̂11

The analyticity condition (3.33), which follows from the connection equation (3.23), im-
poses certain necessary conditions for the poles of the function v̂11 , more exactly, of its
continuation obtained in Lemma 3.5. This section is devoted to the derivation of these
conditions and to the proof of the fact that they are also sufficient for (3.23) to hold.
Denote
{︂ }︂ (︂ ik )︂
1
P := p1 , –p1 ± πi, p1 + 2πi , p1 := sinh–1 ∈ Γ0 . (3.34)
ω

(See Fig. 4, where the positions of the curves Γν correspond to the case Re ω > 0. We will
always assume in the following that this is the case; in the case Re ω < 0 the construction
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 14 of 44

is similar.) Introduce the two following important parameters:

sinh(p1 + iΦ) sinh(p1 – iΦ)

r1 := res Ĝ := – , r2 := res Ĝ := . (3.35)
–p1 ±π i cosh p1 p1 cosh p1

In the next proposition, we give a necessary and sufficient condition for v̂11 guaranteeing
that condition (3.31) hold.
(︁ )︁
Proposition 3.7 Let v̂11 ∈ M(C) ∩ H(V̂1+ ) satisfy (3.32) and (3.30). Then v̂1 ∈ H V̂Σ if and
only if
(︁ Φ+π )︁
v̂11 ∈ H V–Φ \P (3.36)

and

res v̂11 = r2 , res v̂11 = r1 . (3.37)

p1 –p1 –π i

Proof First, let us prove the necessity of condition (3.37). By (3.21), the poles of Ĝ(w) in
C are p1 + 2kπi, –p1 – πi + 2kπi, k ∈ Z. Of all of these poles only p1 , –p1 – πi belong
to(︂V̂Σ (see Fig. 4). Hence
)︂ formulas (3.37)(︂follow from (3.33), (3.35). Further,
)︂ since v̂11 ∈
H V–Φ \ {p1 , –p1 – πi} by (3.33), v̂1 ∈ H V0 \ {p1 + 2πi, –p1 + πi} by (3.30). Hence
π 1 π +Φ

(3.36) also holds.

Let us prove the sufficiency of conditions (3.36), (3.37). From (3.37)(︂ and (3.35) it follows
)︂
(︁ )︁
that res v̂1 (w) = res v̂1 (w) = 0. Hence, v̂1 ∈ H V̂Σ since v̂11 ∈ H V̂Σ \ {p1 , –p1 – πi}
w=p1 w=–p1 –π i
by (3.36) and Ĝ(w) belongs to this space too. The proposition is proven. □

Figure 4 Necessary conditions

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 15 of 44

(︁ )︁
Remark 3.8 Condition (3.36) implies v̂11 ∈ H V̂1+ .

4 h1 -Automorphic solution of difference equation (3.32),  ̸= 3π /2

In this section, we construct an h1 -automorphic solution of difference equation (3.32)
satisfying all the conditions of Proposition 3.7 for Φ ̸= 3π/2. This limitation is related to
the method of obtaining a solution, which uses the Cauchy-type integral. The kernel of
this integral must be analytic on the integration contour. In turns out that it is possible to
find such a kernel only when Φ = 3π/2. Fortunately, the case Φ = 3π/2 does not need an
integral of the Cauchy type, since the difference equation (3.32) is solved by elementary
methods in this case (see Sect. 5).

4.1 Poles of Ĝ2 and its expansion when Re w → ∞

In this subsection, we give the properties of G2 , given by (3.31), which are necessary for
the solution of the main problem for Φ ̸= 3π/2. Denote q1 = –p1 – πi + 2iΦ.

Lemma 4.1 i) The poles of Ĝ2 belonging to V ππ –Φ

–Φ
are –p1 – πi for Φ ≥ 3π
2
and – q1 +
2
πi for Φ ≤ 3π
2
and res Ĝ2 = res Ĝ2 = r1 .
–p1 –π i –q1 +π i

π –Φ are p1 – 2πi for Φ ≥ 3π/2, –p1 + πi – 2iΦ for

ii) The poles of Ĝ2 belonging to V––Φ
2
Φ ≤ 3π/2 and res Ĝ2 = res Ĝ2 = r2 .
p1 –2π i –p1 +π i–2iΦ
iii) The function Ĝ2 admits the expansion

Ĝ2 (w) = ∓2i sin Φ + g1± e∓w + g2± e∓2w + · · · , ±Re w ≥ 2Re p1 (4.1)

uniformly with respect to Im w.

iv)
(︂ )︂ (︂ π )︂
Ĝ2 h2 (w) = –Ĝ2 (w), Ĝ2 – iΦ = 0. (4.2)
2
Remark 4.2 The condition in (4.1) is necessary since the function G2 is analytic for
|Re w| ≥ 2Re p1 by (3.34) and the same function G2 with the change of k1 to k2 also is
analytic in the same region because k1 < k2 , see Theorem 2.2.

Proof The first three assertions follow directly from (3.31). The last assertion (4.2) follows
(︂ πi )︂ πi
from the fact that h2 – iΦ = – iΦ. □
2 2

Remark 4.3 In the case Φ = 3π/2, we have g1± = 0. However, this will not affect the final
results.

4.2 Reductionπof problem (3.32), {︂ (3.30) to a conjugate

}︂ problem
+Φ
Denote Π̂ := V π2 –Φ , Π̂± = w ∈ Π̂ : ±Re w > 0 , ∂ Π̂+ = β̂ ∪ γ̂ ∪ (β̂ + 2iΦ), where
{︂ 2 }︂ {︂ ⃓ [︂ ]︂}︂
β̂ := w ∈ Γ π2 –iΦ : Re w ≥ 0 , γ̂ := w⃓Re w = 0, Im w ∈ π2 – Φ, π2 + Φ . We will look
for a solution of the following problem: to find an analytic function in Π̂+ , whose boundary
values on ∂ Π̂+ , â1 (w + i0), w ∈ β̂; â1 (w + 2iΦ – i0), w ∈ β̂ + 2iΦ; â1 (w + 0), w ∈ γ̂ exist and
are such that they satisfy the following conditions of conjugation:

â1 (w + i0) – â1 (w + 2iΦ – i0) = Ĝ2 (w), w ∈ Γπ /2–Φ ; â1 (w + 0) = â1 (–w + πi + 0), w ∈ γ̂ ,
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 16 of 44

Figure 5 Reduction to conjugate problem

Figure 6 Riemann–Hilbert problem

(4.3)

see Fig. 5.

4.3 Solution of the conjugate problem (4.3) for  ̸= 3π /2

We start solving the problem (4.3). We will reduce this problem to a Riemann–Hilbert
problem. To this end we map Π̂+ conformally onto Π̌ := C∗ \ β̌, where C∗ is the Riemann
sphere, β̌ := t(β̂). For example, define

(︂ π (︂ πi )︂)︂
w ⃓→ t = t(w) = coth2 w– , w ∈ Π̂+ . (4.4)
2Φ 2

Denote the inverse transform w(t) : Π̌+ to Π̂+ .

Note that when w ∈ Π̂+ tends to β̂, t(w) tends to β̌ from above, and when w ∈ Π̂+ tends
to β + 2iΦ, t(w) tends to β̌ from below. Obviously,

(︂ (︂ π )︂)︂ (︂ πi )︂
t(γ̂ ) =: γ̌ = [–∞, 0], t ±i + Φ = 0, t(∞) = 1, t =∞
2 2

(see Fig. 6).

(︁ )︁
Hence problem (4.3) is equivalent to the Riemann–Hilbert problem for ǎ1 (t) := â1 w(t) ,
t ∈ Π̌+ , which at the same time is the saltus problem (see [17, Ch. 16, 18])

ǎ1 (t + i0) – ǎ1 (t – i0) = Ǧ2 (t), t ∈ β̌. (4.5)

(︁ )︁
Here Ǧ2 (t) := Ĝ2 (w(t)), t ∈ β̌, ǎ1 (t) := ǎ1 w(t) , t ∈ Π̌; ǎ1 (t ± i0) = lim ǎ1 (t ± iε) and for
ε→0+
′
t ∈ β̌, w(t) ∈ β̂. From (4.4) and (4.2), it follows that Ǧ2 (t) and Ǧ2 (t) are continuous on β̌
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 17 of 44

and

Ǧ2 (0) = 0, Ǧ2 (1) = –2i sin Φ. (4.6)

It is well known that a particular solution of (4.5) is given by the Cauchy-type integral

∫︂
1 Ǧ2 (t ′ ) ′
ǎ1 (t) = dt , t ∈ Π̌. (4.7)
2πi t′ – t
β̌

Obviously, ǎ1 (t) ∈ H(Π̌+ ) and ǎ1 (t) −−−→ 0. Moreover, there exist ǎ1 (t ± i0), satisfying
t→∞
(4.5) and by (4.6) there exists lim ǎ1 (t), t ∈/ β̌.
t→0
In the following lemma, we establish an expansion of (4.7) at t = 1; it plays an impor-
tant role in describing the fact that the solution belongs to a certain class and hence is
unique. We introduce the parameter α := π/Φ, which plays an important role in further
calculations. Since π < Φ < 2π , we have

1/2 < α < 1. (4.8)

Throughout the following (up to Sect. 8) α will represent this number.

Everywhere in what follows, we will understand the expression “expansion into a con-
vergent series of functions in a certain region”, as a uniformly convergent series in the
indicated region.
We will need the following important technical statement related to the behavior of
the Cauchy-type integrals at the endpoints of an interval, see [31, Ch. 4, §30]. Let β̌ε :=
β̌ ∩ Bε (1), where Bε := {t ∈ C : |t – 1| < ε}.

Lemma 4.4 Let g(t) be an analytic in Bε (1), 1 < λ < 2 and

∫︂
1 (1 – t ′ )λ g(t ′ ) ′
Ω(t) := dt ,
2πi β̌ε t′ – t

where (1 – t ′ )λ = exp(λ ln |1 – t ′ | + i arg(1 – t ′ )), arg(1 – t ′ ) ∈ (–π, π). Then

Ω(t) = c0 (1 – t)λ + Ψ(t), t ∈ Bε (1) \ β̌,

where Ψ(t) is analytic in Bε (1).

The proof of this Lemma is a slight modification of the proof of formula (30.2) from [31],
where a similar statement is proved for g ≡ 1 and –1 < Re λ < 0.

Lemma 4.5 Let Φ ̸= 3π/2. There exists ε > 0 depending only on k1 such that the function
ǎ1 (t) can be expanded in the following convergent series in Bε (1) \ β̌

sin Φ 1
ǎ1 (t) = – ln(1 – t) + c0 + c1 (1 – t) + cα (1 – t) α + c2 (1 – t)2
π (4.9)
k+ α1
k
+ ck (1 – t) + ck+α (1 – t) + ck+2 (1 – t) k+2
+ ··· , t ∈ Bε (1) \ β̌,
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 18 of 44

where ln(1 – t), (1 – t)(1/α) = e(1/α) ln(1–t) are analytic in Bε \ β̌, ln z ∈ R for z > 0 and the
coefficients in (4.9) depend only on Ĝ2 .

Proof From (4.4) it follows that there exists ε > 0 such that |t – 1| < ε implies |Re w| >
2Re p1 . Moreover, expanding ekw in a series in powers of (t – 1) in this region, we obtain
from (4.1) that Ǧ2 (t) admits the following expansion in the neighborhood Bε (1) of the
point t = 1:

Ǧ2 (t) = –2i sin Φ + (1 – t)1/α g(t), (4.10)

where g is analytic in Bε (1) and for (1 – t)1/α the same branch as in Lemma 4.4 is chosen.
Further, from (4.7), we have:

∫︂
1 Ǧ2 (t ′ ) ′
ǎ1 (t) = dt + A(t), (4.11)
2πi β̌ε t′ – t

where A(t) is analytic in Bε (1)

Substituting (4.10) into the Cauchy-type integral in (4.11), using the expansions of the
integral into convergent series
∫︂
dt ′ ′
dt = ln(1 – t) + c0 + c1 (1 – t) + ...ck (1 – t)k + · · ·
t′ – t (4.12)
β̌

and Lemma 4.4 we obtain (4.9) by (4.8). □

Now, we are able to find a solution to problem (3.32). First, we define this solution in Π̂
and then we extend it to C. Let us define â1 (w), w ∈ Π̂+ by the formula

(︁ )︁
â1 (w) := ǎ1 t(w) , w ∈ Π̂+ , (4.13)

where ǎ1 (t) is given by (4.7). Moreover, the same formula (4.13) defines the analytic func-
tion â1 (w) in Π̂, since â1 (w) satisfies the second condition in (4.3). Obviously, (4.4) implies
that

â1 (w) = â1 (–w + πi), w ∈ Π̂. (4.14)

Moreover, â1 (w) satisfies the first condition in (4.3). In fact, for w ∈ β̂, Re w > 0 this follows
from (4.5), and for w ∈ β̂, Re w < 0 this follows from (4.5) and (4.2). Further, we extend
π +3Φ
â1 (w) to V π2 –3Φ by the “continuation” formulas corresponding to the difference equation
2
(3.32):

â1 (w) = â1 (w – 2iΦ) – Ĝ2 (w – 2iΦ), w ∈ Π̂ + 2iΦ, Π̂ + 4iΦ, . . . (4.15)

and

â1 (w) = â1 (w + 2iΦ) + Ĝ2 (w), for w ∈ Π̂ – 2iΦ, Π̂ – 4iΦ, . . . (4.16)
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 19 of 44

This extension is meromorphic by (4.3) and still has property (4.14). Let us prove this.
Let w ∈ Π̂ – 2iΦ (see Fig. 7, 8). Then w + πi ∈ Π̂ + 2iΦ. By (4.16) and (4.15) we have
â1 (w) = â1 (w + 2iΦ) + Ĝ2 (w), â1 (–w + πi) = â1 (–w + πi – 2iΦ) – Ĝ2 (–w + πi – 2iΦ). But
â1 (w + 2iΦ) = â1 (–w + πi – 2iΦ) by (4.14) and Ĝ2 (w) = –Ĝ2 (–w + πi – 2iΦ) by (4.2). Hence
â1 (w) = â1 (–w + πi) in this case. Similarly, this is true for w ∈ Π̂ + 2iΦ.Moreover, â1 (w)
admits a meromorphic extension to C, which satisfies (3.32) and (3.30).

Proposition 4.6 For Φ ̸= 3π/2,

i) There exists a meromorphic in C and analytic in Π̂ solution â1 of problem (3.32), (3.30)
given by (4.15), (4.16).
ii) The solution â1 has poles in V–ππ+Φ–Φ for Φ > 3π/2 only at q1 := –p1 – πi + 2iΦ, –q1 + πi =
2
p1 + 2πi – 2iΦ and p1 – 2πi with residues

res â1 = –r1 , res â1 = r1 , res â1 = r2 . (4.17)

q1 –q1 +π i p1 –2π i

For Φ ≤ 3π/2, â1 has poles in V–ππ+Φ–Φ only at p1 + 2πi, –p1 – πi and –p1 + πi – 2iΦ with
2
residues

res â1 = –r1 , res â1 = r1 , res â1 = r2 . (4.18)

p1 +2π i –p1 –π i –p1 +π i–2iΦ

iii) The function â1 (w) admits the following expansion into a convergent series:

sin Φ
â1 (w) = ± w + c± ± –αω
0 + c1,α e + c±
1e
–w
+ c±
2,α e
–2αw
+ c±
2e
–2w
+ ···
Φ (4.19)
+c±
k,α e
–kαw
+ c±
ke
–kw
+ ··· , ±Re w ≥ 2Re p1 .

Proof Statement i) is proven above. Statement ii) follows from the difference equation
(3.32), h1 -automorphy of â1 , Lemma A.1 and Lemma 4.1, since the function â1 is analytic
in Π̂ by i), see Fig. 7, 8.
The expansion (4.19) in Π̂+ ∩{Re w > 2Re p1 } follows from (4.9), (4.4) and (4.13). For Π̂– ∩
{Re w < –2Re p1 } this follows from h1 - automorphy of â1 , see item i). Finally the expansion
(4.19) for any w ∈ C, with |Re w| > 2Re p1 follows from the “continuation” formulas (4.15)
and the expansion (4.1). □

In the following section, we render the meromorphic solution â1 into an analytic one.

4.4 Automorphic solution of difference equation for  ̸= 3π /2 satisfying

necessary and sufficient conditions
We want to modify â1 into v̂11 , which will satisfy all the conditions of Proposition 3.7. To
this end, for Φ > 3π/2 we add a function T1 to â1 in such a way that it removes the poles
q1 and –q1 + πi (see (4.17)), since by this proposition v̂11 must be analytic at these points. It
turns out that it is possible to construct T1 in such a way that it produces the pole –p1 – πi
with the desired residue, as the same proposition requires.
Second, we add a function T2 producing the pole p1 with the desired residue according
to the same proposition.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 20 of 44

Figure 7 Φ > 3π /2

Consider
(︄ )︄
π (︂ π(w – q ) )︂ (︂ π(–w + πi – q ) )︂
1 1
T1 (w) = r1 coth + coth , (4.20)
2Φ 2Φ 2Φ

where r1 is given by (3.37), and q1 is defined in Proposition 4.6.

It is easy to see that the function T1 satisfies the following conditions:

T1 (–w + πi) = T1 (w), T1 (w + 2iΦ) = T1 (w). (4.21)

π +Φ
The poles of T1 belonging to V–Φ are

q1 , –p1 – πi, –q1 + πi, p1 + 2πi, (4.22)

(see Fig. 8), and

res T1 = res T1 = r1 ; res T1 = res T1 = –r1 . (4.23)

q1 –p1 –π i p1 +2π i p1 +2π i

Further, we define
(︄ )︄
π (︂ π(w – p ) )︂ (︂ π(–w + πi – p ) )︂
1 1
T2 (w) := r2 coth + coth , (4.24)
2Φ 2Φ 2Φ
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 21 of 44

Figure 8 Φ < 3π /2

where r2 is defined by (3.35). Obviously, T2 also satisfies (4.21).

π +Φ
The poles of T2 in V–Φ are only

p1 , –p1 + πi and res T2 = r2 , res T2 = –r2 . (4.25)

p1 –p1 +π i

Note that (4.20) and (4.24) imply that T1 and T2 admit the expansion into convergent
series of the following type:

T1,2 (w) = c±
1e
∓αw
+ c±
2e
∓2αw
+ c±
3e
∓3αw
+ · · · , ±Re w ≥ 2Re p1 . (4.26)

Finally, we define

⎧
⎨ â1 (w) + T1 (w) + T2 (w), Φ > 3π/2,
v̂11 (w) := (4.27)
⎩
â1 (w) + T2 (w), Φ < 3π/2,

where â1 (w) is given in Proposition 4.6.

Theorem 4.7 Let Φ ̸= 3π/2.

i) The function v̂11 satisfies all the hypothesis of Proposition 3.7.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 22 of 44

(︂ 3π )︂
ii) v̂11 (w) ∈ H Vπ 2 \ Γπ and has a unique pole –p1 + πi on Γπ with residue

res v̂11 = –r2 . (4.28)

–p1 +π i

(︂ )︂
iii) v̂11 (w) ∈ M V––Φ
π –Φ and has a unique pole at p1 – 2πi for Φ > 3π/2,
2

res v̂11 = r2 . (4.29)

p1 –2π i

iv) v̂11 admits the following expansion into a convergent series of type (4.19):

sin Φ
v̂11 (w) = ± w + c± ± ∓αw
0 + cα e + c±
1e
∓w
+ c±
2α e
∓2αw
+ ··· , ±Re w ≥ 2Re p1 . (4.30)
Φ

Proof i) v̂11 satisfies (3.32) and (3.30) by Proposition 4.6, (4.27) and (4.21).
Let us prove (3.36) and (3.37). Consider Φ > 3π/2. {︂ By Proposition
}︂ 4.6, (4.23), (4.22) and
π +Φ
(4.25), the possible poles of v̂11 in V–Φ belong to q1 , –q1 + πi ∪ P, where P is given by
(3.34). Moreover, res v̂11 = res v̂11 = 0 by (4.17), (4.23) and (4.25). Hence, v̂11 satisfies (3.36).
q1 –q1 +π i
Moreover, by (4.23), res v̂11 = r1 , res v̂11 = r2 . Thus, (3.37) is proven for Φ > 3π/2, and v̂11
–p1 –π i p1
satisfies all the hypotheses of Proposition 3.7 in this case (see Fig. 7).
For the case Φ < 3π/2, (3.36) is obtained using (4.27), and (3.37) follows from Lemma
A.1 and the analyticity of â1 in Π̂.
(︂ π +Φ )︂ (︂ 3π )︂
ii) By Proposition 4.6, â1 ∈ H V–2π –Φ , which implies â1 ∈ H Vπ 2 . By (4.22), T1 has
2
π +Φ
poles in V–Φ only at q1 , –p1 – πi, –q1 + πi, p1 + 2πi. For Φ > 3π/2 none of these poles
3π 3π
belong to Vπ 2 . T2 has a pole in V–Φπ +Φ
only at p1 , –p1 + πi by (4.24). Further, p1 ∈/ Vπ 2 ,
–p1 + πi ∈ Γπ and hence ii) holds by (4.27), (4.25).
iii) Consider V––Φ
π –Φ . By Proposition 4.6, â1 has poles here at p1 – 2πi for Φ > 3π /2 and
2
at –p1 + πi – 2Φi for Φ < 3π/2 with residues (4.17) and (4.18).
From (4.20) and (4.24) it follows that T1 does not have poles in V––Φ π –Φ and T2 has a
2
unique pole at –p1 + πi – 2iΦ there for Φ ≤ 3π/2 and res T2 = –r2 . Hence v̂11 has a
–p1 +π i–2iΦ
pole at p1 – 2πi for Φ > 3π/2 and a possible pole at –p1 + πi – 2iΦ for Φ < 3π/2. Then
from (4.27) we obtain res v̂11 = r2 . Similarly, res v̂11 = 0. Therefore iii) is proven.
p1 –2π i –p1 +π i–2iΦ
iv) Finally, expansion (4.30) follows from (4.27), (4.19) and (4.26). Theorem 4.7 is proven.
□

Corollary 4.8 For Φ ̸= 3π/2 a unique pole of the Sommerfeld-type integral kernel v̂1 be-
3π
longing to V– 2π –Φ is –p1 + πi and
2

res v̂1 = 2i sin Φ. (4.31)

–p1 +π i

3π
Proof A unique pole of v̂11 (w) in Vπ 2 is only –p1 + πi by Theorem 4.7 ii) with residue (4.28).
3π
Hence, v̂1 has a unique pole at –p1 + πi in Vπ 2 and (4.31) follows from (3.33), (4.28) and
π
(3.35). The function v̂1 is analytic in V̂Σ = V–Φ by Proposition 3.7, since v̂11 satisfies all the
hypothesis of this Proposition by Theorem 4.7.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 23 of 44

1
It remains only to prove that v̂1 is analytic in V––Φ
π –Φ . By Theorem 4.7 the function v̂1
2
has a unique pole at p1 – 2πi in V––Φ
π –Φ with residue (4.29) and the function Ĝ also has a
2
unique pole at this point with residue –r2 by (3.35). Hence, the function v̂1 is analytic in
π –Φ for Φ ̸= 3π/2 and the Corollary is proven.
V––Φ □
2

5 h1 -Automorphic solution of the difference equation (3.32) for  = 3π /2

In the previous sections, we have constructed a solution to problem (3.32), (3.30) satisfying
all the conditions of Proposition 3.7 for Φ ̸= 3π/2.
It is possible to construct a solution for Φ = 3π/2 using the same method. A slight tech-
nical difficulty in this case arises from the fact that the function Ǧ2 (t) has a pole on β̌.
Nevertheless, one can obtain a solution with the properties indicated in Theorem 4.7.
However, we prefer to find a solution of the problem in the case Φ = 3π/2 by another
method. The point is that in this case it is easy to find a solution of the difference equation
(3.32) in an explicit form without using the Cauchy-type integral.
Using the Liouville theorem, it is easy to show that this elementary solution coincides
with the solution obtained by the Cauchy-type integral.
In this section, we give a meromorphic h1 -invariant solution of (3.32).

5.1 Meromorphic solution of the difference equation for  = 3π /2

In this case, the construction of a meromorphic solution of difference equation (3.32) is
simpler than in the case Φ ̸= 3π/2 and v̂11 is expressed through elementary functions. By
(3.31), for Φ = 3π/2, we have

iω2 sinh 2w
G2 (w) = . (5.1)
ω2 sinh2 w + k12

Let us solve difference equation (3.32) in this case.

Remark 5.1 Comparing the right-hand sides G2 in (3.31) and (5.1) of the difference equa-
tion (3.32), we see that in the case G2 (Φ = 3π/2) the poles of G2 acquire additional sym-
metry. As we will see a little further, this allows us to guess a solution to the difference
equation explicitly in elementary functions without using a Cauchy-type integral.

First, we solve (3.32) in the class of meromorphic functions. It is easy to guess a solution,
using the 3πi-periodicity of G2 . Let us define

iw G2 (w)
m1 (w) := .
3π

Then, by (5.1), m1 satisfies (3.32). Of course, this solution is not unique. All the other
solutions differ from it by a 3πi-periodic function. Similarly to the case Φ ̸= 3π/2, we will
modify this solution in such a way that it will satisfy all the conditions of Proposition 3.7.
Function (5.1) is not automorphic with respect to ĥ1 . Let us symmetrize it.
Define

m1 (w) + m1 (–w + πi)

m(w) := . (5.2)
2
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 24 of 44

Then

π + 2iw
m(w) = G2 (w). (5.3)
6π

Lemma 5.2 i) The function m is an h1 -automorphic solution to (3.32).

5π
2
ii) m has poles in V –5π only at the points of the set
2

{︂ }︂
P1 := ± p1 , p1 ± πi, –p1 ± πi, p1 ± 2πi, –p1 ± 2πi , (5.4)

(see Fig. 9), and

res m = m1 , res m = m2 , res m = m3 , res m = m4 , res m = m5 ,

p1 p1 +π i p1 –π i p1 +2π i p1 –2π i
(5.5)
res m = –m2 , res m = –m1 , res m = –m4 , res m = –m3 , res m = m6 ,
–p1 –p1 +π i –p1 –π i –p1 +2π i –p1 –2π i

where

p1 i p1 i p1 i p1 i
m1 = – + , m2 = – – , m3 = – + , m4 = – – ,
3π 6 3π 6 3π 2 3π 2
(5.6)
p1 5i p1 5i
m5 = – + , m6 = + .
3π 6 3π 6

Proof i) The assertion follows from a direct substitution of (5.3) into (3.32), and (3.30)
follows from (5.2).
ii) The zeros of ω2 sinh2 w + k 2 are ±p1 + 2kπi, ±p1 + πi + 2kπi, k ∈ Z, where p1 is
5π
2
defined by (3.34). Obviously, only the poles from P1 belong to V , see Fig. 9.
– 5π
2
Formulas (5.5) follow from (5.3) and Lemma A.1. □

Now we modify the function m(w) in such a way that it will satisfy the conditions (3.36),
(3.37) of Proposition 3.7. To this end, we add to m an appropriate 3πi-periodic function.
(︂ 5π
Since for Φ = 3π/2, r1 = r2 = i, conditions (3.36), (3.37) take the form v̂11 (w) ∈ H V– 23π \
)︂ 2

P , where P is given by (3.34), and res v̂11 (w) = i, res v̂11 (w) = i.
p1 –p1 –π i

5.2 Automorphic solution of the difference equation for  = 3π /2 satisfying

necessary and sufficient conditions
5π
By (5.4), the function m has 8 poles in V–23 π belonging to P1 with residues (5.5). We modify
2
m so that (3.36) and (3.37) hold. To this end, we first add to m functions T1 and T3 which
“correct” the residues –m1 at the point p1 and –m4 at the point –p1 – πi by i. Then we add
T2 , which annihilates the poles –p1 , p1 + πi and p1 – πi, –p1 + 2πi.
So, consider the following functions (the 3πi - periodic supplements)
[︄ ]︄
i – m1 w – p1 w – (–p1 + πi)
T1 (w) := coth – coth , (5.7)
3 3 3
[︄ ]︄
m2 w – (p1 – πi) w – (–p1 )
T2 (w) := – coth – coth , (5.8)
3 3 3
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 25 of 44

Figure 9 Poles of the function m, Φ = 3π /2.

[︄ ]︄
m3 w – (p1 – πi) w – (–p1 – πi)
T3 (w) := – coth – coth , (5.9)
3 3 3

where m1,2,3 are given by (5.6). Obviously, the functions T1,2,3 are 3πi-periodic, and ĥ1 -
automorphic:

(︁ )︁
T1,2,3 (w + 3πi) = T1,2,3 (w), T1,2,3 ĥ1 w = T1,2,3 (w). (5.10)

Finally, define

T (w) := T1 (w) + T2 (w) + T3 (w). (5.11)

5π
From (5.7)–(5.11), it follows directly that the set of the poles of T in V 2
is P1 given by
– 5π
2
(5.4) (see Fig. 9) and

res T = m6 , res T = res T = –m2 , res T = res T = ∓m3 ,

p1 p1 +π i p1 –2π i ±p1 –π i ±p1 +2π i
(5.12)
res T = res T = –m6 , res T = m2 .
–p1 –2π i –p1 +π i –p1

Define

v̂11 (w) := m(w) + T (w), (5.13)

where m is given by (5.3) and T is given by (5.11).

Theorem 5.3 i) For Φ = 3π/2 the function v̂11 satisfies all the conditions of Proposition 3.7.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 26 of 44

3π
ii) The poles of v̂11 in V 2
are
– 5π
2

–p1 – πi, –p1 + πi, p1 , p1 – 2πi (5.14)

with the following residues

res v̂11 = i, res v̂11 = –i, res v̂11 = i, res v̂11 = i. (5.15)
–p1 –π i –p1 +π i p1 p1 –2π i

Proof i) Equations (3.32) and (3.30) follow from Lemma 5.2 and (5.13). From (5.5), (5.6)
and (5.12), we obtain (3.37).
5π
Let us prove (3.36). Since all the poles of v̂11 in V– 23π belong to P1 by Lemma 5.2, it suffices
{︂ 2 }︂
to prove that v̂11 is analytic in P1 \ P = – p1 , p1 ± πi, –p1 + 2πi .
From (5.5), (5.12), (3.30) and Lemma A.1, we obtain res v̂11 = res v̂11 = res v̂11 =
w=–p1 p1 +π i –p1 +2π i
res v̂11 = 0. Thus (3.36) and i) are proven.
p1 –π i
3π
ii) By (5.4) and (5.11) the poles of m and T in V 2
are ±p1 , p1 ± πi, ±p1 – 2πi, –p1 ± πi.
– 5π2
From (5.5), (5.6), (5.12) and (5.13) it follows that v̂11 (w)
does not have poles at –p1 , p1 + πi,
–p1 – 2πi and has poles (5.14) with residues (5.15). Statement ii) is also proven. □

Now we establish an important property of the function v̂1 , similar to Corollary 4.8 for
the case Φ = 3π/2. We recall that this function plays a crucial role in the construction of
the Sommerfeld-type representation for the solution of the main problem. This represen-
tation will be given in the following section.

Corollary 5.4 For Φ = 3π/2 the function v̂1 given by (3.33), just as in the case Φ ̸= 3π/2,
3π
2
has a unique pole at –p1 + πi belonging to V , and res v̂1 = –2i.
– 5π
2 –p1 +π i

(︁ )︁ (︁ )︁
Proof The function v̂1 ∈ H V̂Σ = H V–π3π by Proposition 3.7 and Theorem 5.3. It suf-
2
3π 3π – 3π
fices to analyze its behavior in V 2
\ V̂Σ = Vπ 2 ∪ V 2
.
– 5π
2 – 5π
2
3π
By Theorem 5.3 ii), (3.21), and (2πi)-periodicity of Ĝ, unique poles of v̂11 and Ĝ in Vπ 2
– 3π
2
are –p1 + πi and p1 , and a unique pole in V of the same functions is p1 – 2πi with
– 5π
2
residues (5.15) and (3.35). Hence the statement follows from (3.33). □

6 Solution to problem (1.9)

In this section, we give a Sommerfeld-type representation of solution to problem (1.9).
This representation was obtained by A. Sommerfeld for diffraction by a half-plane and
is widely used in Mathematical Diffraction Theory [38]. This representation is an inte-
gral with a specially chosen integrand along a Sommerfeld-type contour. This contour
has double-loop form as in Fig. 11.
We define first this curvilinear contour depending on ω ∈ C+ (in contrast to the Som-
merfeld contour), and then we reduce it to the rectilinear contour that coincides with the
Sommerfeld contour C (see Fig. 11)). Define

C (ω) = C1 (ω) ∪ C2 (ω), (6.1)

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 27 of 44

where
{︂ }︂ {︂ }︂
C2 (ω) := w ∈ Γ– 5π (ω), w1 ≤ –b ∪ γ2 (w) ∪ w ∈ Γ– π2 (ω), w1 ≤ –b ,
2

{︁ }︁
γ2 (ω) is the segment of the line – b + iw2 , w2 ∈ R lying between Γ– 5π and Γ– π2 , C1 (ω) =
2
–C2 (ω) – 3πi and b ≥ 2Re p1 (see Fig. 10).
In our case, the integrand is the Sommerfeld exponential e–ωρ sinh w multiplied by the
kernel v̂1 , which was constructed in the previous sections.

Solution to problem (1.9) Let us prove that the first equation in (1.9) holds. First, we write
the Sommerfeld integral (2.3) with C (ω) instead of C . We keep the notation u1 for this
integral because we will see later that these two integrals coincide.
So, let
∫︂
1
u1 (ρ, θ ) = e–ωρ sinh w v̂1 (w + iθ ) dw, (ρ, θ ) ∈ Q (6.2)
4π sin Φ
C (ω)

where C (ω) is defined above, and v̂1 is given by (3.32) with v11 constructed in Sects. 3-5.

Remark 6.1 The function v̂1 was obtained as a result of solving problem (3.3a)–(3.3c).
However, the solution v(x) of this problem does not interest us. We are only interested in
its Neumann datum, which is an important part of the kernel v̂1 .

Here and below, we will use the following estimate: for ρ > 0, τ ∈ [τ0 , π – τ0 ] with 0 <
τ0 ≤ π/2, w = w1 + iw2 ∈ Γ0 and ω ∈ C+ , we have

⃓ –ωρ sinh(w–iτ ) ⃓
⃓e ⃓ ≤ e–C(ω,τ0 )ρ cosh w1 , (6.3)

where C(ω, τ0 ) > 0. The proof of this estimate is given in Appendix A.4. Hence the integral
(6.2) converges by the expansion (8.1), since v̂1 (w + iθ ) does not have poles on C (ω) by
Corollaries 4.8 and 5.4 (see Fig. 10, where the exponential decreases superexponentially
in the shaded regions).
Let us prove that u1 (ρ, θ ) satisfies the first equation of (1.9). To this end we rewrite (6.2)
as
∫︂
1
u1 (ρ, θ ) = e–ωρ sinh(w–iθ) v̂1 (w)dw.
4π sin Φ
C (ω)+iθ

Let us fix ρ > 0, θ0 ∈ (φ, 2π). By the Cauchy theorem

∫︂
1
u1 (ρ, θ ) = e–ωρ sinh(w–iθ) v̂1 (w)dw
4π sin Φ
C (ω)+iθ0

for any θ0 sufficiently close to θ .

Now the differentiation in (ρ, θ ) under the sign of the integral is possible and the first
(︁ )︁
equation in (1.9) follows from the formula Δ + ω2 e–ωρ sinh(w–iθ) = 0.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 28 of 44

Figure 10 Sommerfeld double-loop contour C(ω).

Figure 11 Sommerfeld two-loop contour C = C(i)

Finally, boundary conditions (3.3b) and (3.3c) are verified in the next section. The inte-
gral (6.2) is transformed into the integral (2.3) over the contour C := C (i), which no longer
depends on ω (see Fig. 11). □
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 29 of 44

7 Proof of the boundary conditions (1.9)

7.1 Decomposition of the solution into a plane wave and a wave diffracted by the
vertex
In this section, we decompose the solution of problem (1.9) given by (6.2) into two parts:
the first part is the plane wave generated by the first boundary condition (1.9) and the
second part is the wave diffracted by the edge of the wedge.
3π
To give this decomposition, we recall that a unique pole of v̂1 (w) lying in V– 2π –Φ is
2

–p1 + πi and res v̂1 (w) = 2i sin Φ, (7.1)

–p1 +π i

as follows from Corollaries 4.8 and 5.4.

Define the plane wave generated by the first boundary condition in (1.9):

up (ρ, θ ) := e–ωρ sinh(p1 +iθ) , ρ > 0, θ ∈ R, (7.2)

where p1 is given by (3.34), and a “diffracted” wave

∫︂
1
ud (ρ, θ ) := e–ωρ sinh w v̂1 (w + iθ )dw, θ ∈ (2π – Φ, 2π]. (7.3)
4π sin Φ
Γ 5π ∪ Γ– π
– 2 2
–––––→ ←––––

The integrand here coincides with the integrand in (2.3), but the contour of integration
differs from C (see Fig. 12).
It turns out that the solution u1 is decomposed into the sum of up and ud and the cor-
responding decomposition is more convenient for the proof of boundary conditions.

Figure 12 Decomposition of the solution

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 30 of 44

Theorem 7.1 The solution of problem (1.9) given by (2.3) admits the following represen-
tation:

⎧ ∫︂ [︂ )︂
⎪ 1
⎪
⎪ e–ωρ sinh w v̂1 (w + iθ )dw, θ ∈ 2π – Φ, 3π/2 ,
⎪
⎪ 4π sin Φ
⎪
⎪ Γ 5π ∪ Γ– π
⎪
⎨ – 2 2
–––––→ ←––––
u1 (ρ, θ ) =
⎪ ∫︂ (︂ ]︂
⎪
⎪ 1
⎪
⎪ e–ωρ sinh w v̂1 (w + iθ )dw + up (ρ, θ ), θ ∈ 3π/2, 2π .
⎪
⎪ 4π sin Φ
⎪
⎩ Γ 5π ∪ Γ– π
– 2 2
–––––→ ←––––

(7.4)

Proof By the Cauchy theorem, u1 defined by (6.2) admits the representation

∫︂
1
u1 (ρ, θ ) = e–ωρ sinh w v̂1 (w + iθ )dw
4π sin Φ
Γ 5π ∪Γ– π
– 2 2
–––––→ ←––––
∫︂
1
– e–ωρ sinh w v̂1 (w + iθ )dw, (7.5)
4π sin Φ
γ (ω)

where γ (ω) is the contour bounded by γ– (ω), γ+ (ω) and Γ– π2 , Γ– 5π (see Fig. 12).
2
Let us find the poles of v̂1 (w + iθ ), w ∈ Ω, for any θ ∈ (2π – Φ, 2π), where Ω is the region
3π
bounded by γ (ω). Since v̂1 (ω) has a unique pole (7.1) belonging to V– 2π –Φ , the function
2
3π –iθ
2
v̂1 (w + iθ ) has a unique pole –p1 + πi – iθ in V . This pole belongs to Ω if and only if
– π2 –Φ–iθ
θ ∈ (3/2π, 2π). Calculating the second integral in (7.5) with the help of residues, we obtain
(7.4). □

Remark 7.2 It may seem that formula (7.4) gives a discontinuous solution on the ray θ =
3π/2, but this is not the case, since (7.4) coincides with (6.2) by construction. Nevertheless,
it is possible to prove this independently (see Appendix A.5).

7.2 Boundary values of the solution u1

We continue to solve problem (1.9).

Proposition 7.3 The solution u1 (ρ, θ ) given by (2.3) is a solution to (1.9).

Proof The fact that u1 satisfies the Helmholtz equation in (1.9) has been proven in Sect. 6.
Let us prove the boundary conditions in (1.9). First, we prove the first condition (1.9),
which in polar coordinates takes the form

u1 (ρ, 2π) = e–ik1 ρ , ρ > 0.

Since by (7.2), (3.34), up (ρ, 2π) = e–ik1 ρ , it suffices to prove that ud (ρ, θ ) satisfies the homo-
geneous conditions. From (3.33), (7.3) and 2π -periodicity of Ĝ, we obtain that it suffices
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 31 of 44

to prove that
∫︂
e–ωρ sinh w v̂11 (w + 2πi)dw = 0. (7.6)
Γ 5π ∪ Γ– π
– 2 2
–––––→ ←––––

Making the change of the variable w′ = w + 2πi, w′ ∈ Γ– π2 ∪ Γ 3π , we obtain (7.6) since the
–––→ ←–2–
integrand (after the change) is an h1 -automorphic function and h1 Γ– π2 = Γ 3π .
–––→ ←–2–
Let us prove the second boundary condition in (1.9), (7.4): u1 (ρ, 2π – Φ) = 0, ρ > 0. From
(3.23), (3.20) (see Remark 3.4) we have
∫︂
1
u1 (ρ, 2π – Φ) = – e–ωρ sinh w v̂12 (w + 2πi – iΦ)dw.
4π sin Φ
Γ 5π ∪ Γ– π
– 2 2
–––––→ ←––––

Making the change of the variable w′ = w + 2πi – iΦ, we obtain

∫︂
1
u1 (ρ, 2π – Φ) = – e–ωρ sinh(w–2π i+iΦ) v̂12 (w)dw = 0,
4π sin Φ
Γ– π –Φ ∪ Γ 3π
2 – –Φ
–––––––→ ←–––2–––––

since v̂12 is an h2 -automorphic function and h2 Γ– π2 –Φ = Γ 3π –Φ . □

–––––→ ←–2––––

8 The solution U to problem (1.7) belongs to Eα , α = π /

8.1 Expansion of v̂1 at infinity
We will need to prove (2.2). To this end, we have to find the expansion of the integrand
v̂1 (w) at infinity.

Lemma 8.1 For any Φ ∈ (π, 2π), the function v̂1 admits expansion into the convergent
series

sin Φ
v̂1 (w) = ± w + c±
0 + R1 (w), (8.1)
Φ
∞
∑︂
R1 (w) := c±
α e∓αw
+ a±
ke
∓φk (α)w
, ±Re w ≥ 2Re p1 , (8.2)
k=1

where φk (α) ≥ 1.

Moreover,

|R1 (w)| ≤ Ce–α|Re w1 | , |Re w1 | ≥ 2Re p1 . (8.3)

Proof For Φ ̸= 3π/2i, v̂11 (w) admits the expansion (4.30), which remains true for the case
Φ = 3π/2 as follows from (5.13). Obviously, G(w) admits the expansion into the convergent
series

Ĝ(w) = e∓iΦ + c±
1e
∓w
+ c±
2e
–2w
+ ··· , ±Re w ≥ 2Re p1 .
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 32 of 44

Hence (8.1), (8.2) follow for Φ ̸= 3π/2. For Φ = 3π/2 (8.1), (8.2) is checked directly using
(5.13), (5.3) and (5.1). Finally, estimate (8.3) follows from (8.2). □

8.2 Behavior of the function U at infinity {︂ }︂

Lemma 8.2 The solution u1 is a C ∞ -function in Q \ {0}, bounded in Q ∩ (ρ, θ ), ρ ≥ ε > 0
with all its first derivatives.

Proof This follows from the superexponential decay of the exponential e–ωρ sinh w in the
integral (2.3) (see shaded region in Fig. 11), analyticity of v̂1 (w) for |Re w| ≥ 2Re p1 (see
Corollaries 4.8, 5.4), and the expansion (8.1).

8.3 Behavior of the function U at the origin

We continue to prove the main theorem, namely, we prove that U given by (1.12) belongs
to Eα . Represent the contour C as

(︁ )︁ (︁ )︁
C := C+ ∪ γ+ ∪ C– ∪ γ– , (8.4)

where

(︁ )︁ (︁ )︁
C+ := Γ+ ∪ Γ+ – 2πi , C– := Γ– ∪ Γ– – 2πi , (8.5)
←– – → ←– – →

and the contours Γ± , γ± are shown in Fig. 11. Note that the “finite” part of the integral
(2.3), has a “good” asymptotics
∫︂
e–ωρ sinh w v̂1 (w + iθ ) dw = C(θ ) + o(1), ρ→0 (8.6)
γ+ ∪γ–

since v̂1 (w) is an analytic function for ±Re w ≥ 2Re p1 by its construction. We will need to
derive the explicit form of C(θ ) in (8.6), because we need to prove that U = u1 + u2 (see
(1.12)) is continuous at 0.
For this, we will use simple but important properties of the Sommerfeld integral based
on the symmetry of the contour C– ∪ C+ and automorphy of sinh w with respect to –3/2iπ .
These relations will allow us to “remove” the “bad” leading terms in the asympotocs of U
and ∇U at 0.
In fact, the function sinh w is a 2πi-periodic function and is automorphic whith respect
to –3πi/2. The contour C– passes into C+ under the automorphism w → (–w – 3πi). Hence
∫︂ ∫︂ (︂ )︂
e–ωρ sinh w dw = e–ωρ sinh w ± (w + iθ ) dw
C+ ∪C– C+ ∫︂
∪C–
(︂ )︂ (8.7)
= e–ωρ sinh w (sinh w) ± (w + iθ ) dw = 0.
C+ ∪C–

Here and below, the sign “±” in the integral over the union of two contours means that
the “+” sign is taken for the right-hand side of the union (i.e., for Re w > 0) and the “–” sign,
for the left-hand side (i.e., for Re w < 0).
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 33 of 44

Lemma 8.3 The function u1 admits the following asymptotics at the origin:
[︄ ∫︂ ∫︂ ]︄
1
u1 (ρ, θ ) = (iθ ) dw – dw + c1 + o(1), ρ → 0, (8.8)
4π sin Φ
γ+ γ–

where c1 does not depend on θ and o(1) is uniform with respect to θ . Moreover,
(︄ )︄
iω sin θ iω cos θ
∇u1 (ρ, θ ) = ,– + O(ρ –1+α ), (8.9)
ρΦ ρΦ

where O(ρ –1+α ) is uniform with respect to θ .

Remark 8.4 Equality (8.8) shows that u1 is not continuous at 0; this is due to the inconsis-
tency of the boundary conditions in (1.9). However, as we will see later, the solution U of
the problem (1.7)) is continuous.

Proof Let R1 be given by (8.2), and

∫︂
1
u1,R (ρ, θ ) := e–ωρ sinh w R1 (w + iθ )dw. (8.10)
4π sin Φ
C

Then u1,R (ρ, θ ) is continuous at 0, namely

u1,R (ρ, θ ) = c0 + o(1), ρ → 0, (8.11)

where c0 does not depend on θ and o(1) is uniform with respect to θ .

In fact, changing w + iθ = w and using the expansion (8.2), we obtain

∫︂ ∫︂
1 1
u1,R (ρ, θ ) = e–ωρ sinh(w–iθ) R1 (w)dw → R1 (w)dw, ρ → 0.
4π sin Φ 4π sin Φ
C +iθ C +iθ

(8.12)

Hence, by the analyticity of R1 (w) in {|Re w| ≥ 2p1 }, estimate (8.3) and the homotopy of
∫︁
the contours C + iθ for any θ , we obtain (8.11) with c0 := 4π sin
1
Φ
R1 (w)dw.
C
Let us prove that the convergence in (8.12) is uniform with respect to θ . Obviously, it
suffices to prove it for the integrals over C+ ∪ C– , which in turn comes down to the proof
of it over, for example, the contour Γ+ . Let ω := ω1 + iω2 , w = w1 – iπ/2 + iθ . Then the
←–
difference of the integrals in (8.12) takes the form (up to a factor)

∫︂∞
[eiω1 ρ cosh w1 e–ω2 ρ cosh w1 – 1]R1 (w1 – iπ/2 + iθ )dw1 , ρ → 0. (8.13)
2p1

By the Lebesgue Theorem and the estimate (8.3) we see that there exists an integrable
majorant, which does not depend on θ . Hence the uniform convergence follows.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 34 of 44

Let us prove (8.8) and (8.9). By (2.3), (8.1), (8.10), (8.4), (8.7), and 2π -periodicity of sinh w,
we get
∫︂ (︃ )︃
sin Φ
u1 (ρ, θ ) = e–ωρ sinh w ± w + c±
0 dw
Φ
γ+ ∪γ–
∫︂
1
+ e–ωρ sinh w (±iθ )dw + u1,R (ρ, θ )
4π sin Φ
γ+ ∪γ–

∫︂ (︃ )︃
sin Φ ±
Hence, (8.8) follows by (8.11) with c1 = e–ωρ sinh w
± w + c0 dw + c0 .
Φ
γ+ ∪γ–
Let us prove (8.9). We have by (2.3):
∫︂
ω
∂ρ u1 (ρ, θ ) = – e–ωρ sinh w (sinh w)v̂1 (w + iθ )dw. (8.14)
4π sin Φ
C

Similarly, differentiating under the sign of the integral in (2.3), integrating by parts and
using the superexponential decrease of the exponent (see Remark 2.3), we obtain
∫︂
iωρ
∂θ u1 (ρ, θ ) = e–ωρ sinh w (cosh w)v̂1 (w + iθ )dw. (8.15)
4π sin Φ
C

Let us use the asymptotics of these functions as ρ → 0 (see Appendix A.7)

∂ρ u1 (ρ, θ ) = c1 (ω, θ )ρ –1+α + O(1), ρ → 0 (8.16)

1
∂θ u1 (ρ, θ ) = – + O(ρ α ), ρ → 0, (8.17)
Φ

where O(·) are uniform with respect to θ . Then, since

(︂(︁ )︁ sin θ (︁ )︁ cos θ )︂

∇ul (ρ, θ ) = cos θ ∂ρ ul – ∂θ ul (ρ, θ ), sin θ ∂ρ ul + ∂θ ul (ρ, θ ) , l = 1, 2,
ρ ρ
(8.18)

we obtain (8.9). □

Corollary 8.5 The solution U to problem (1.7) belongs to the space Eα , α = π/Φ.

Proof Note that by (1.8), the contour of integration for u2 can be taken equal to the contour
C (i) from (6.1) Similarly to (8.8), by (1.11), we have
[︄ ∫︂ ∫︂ ]︄
1 (︁ )︁
u2 (ρ, θ ) = – iθ + 4πi – 4iΦ dw – dw + O(1), ρ → 0.
4π sin Φ
γ+ γ–

Hence, by (1.12), (8.8) we obtain

U(ρ, θ ) = u1 (ρ, θ ) + u2 (ρ, θ ) = c0 + o(1), ρ → 0,

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 35 of 44

where c0 does not depend on θ . Therefore, U ∈ C(Q̄).

Let us prove that

∇U = O(ρ –1+α ) (8.19)

uniformly with respect to θ . Since the leading terms of asymptotics (8.16) and (8.17) do
not depend on k1 and (8.17) does not depend on θ , we have by (1.11) and (2.4)

1
∂ρ u2 (ρ, θ ) = c2 (ω, θ )ρ –1+α + O(1), ∂θ u2 (ρ, θ ) = + O(ρ –1+α ), ρ → 0.
Φρ

Hence, using (8.18), we have:

(︂ iω sin θ iω cos θ )︂
∇u2 (ρ, θ ) = – , + O(ρ –1+α ) (8.20)
Φρ Φρ

Thus, (8.9) and (8.20) give (8.19) which completes the proof of the corollary, since

∇U = ∇u1 + ∇u2 . □

9 Uniqueness
Let U, U1 be solutions to (1.7) belonging to the space Eα (see Definition 2.1).
Then Uh := U – U1 ∈ Eα and it satisfies the following homogeneous problem
⎧ (︁ )︁
⎨ – Δ – ω2 Uh (ρ, θ ) = 0, (ρ, θ ) ∈ Q,
⎩
Uh (ρ, 2π) = Uh (ρ, 0) = 0, ρ > 0.

Define σ (ρ, θ ) := Uh (L(y)), see (3.1). Obviously, σ ∈ Eα (K) (Eα (K) is defined similarly to
Eα on the domain K ). Similarly to (3.3a)–(3.3c), we obtain
⎧
⎨ H (D)σ (ρ, θ ) = 0, x ∈ K,
(9.1)
⎩
σ (x1 , 0) = σ (0, x2 ) = 0, x1,2 > 0.
{︂ }︂ ⃓
⃓
For p > 0 let us denote Kp := (x1 , x2 ) ∈ K : x1 < –p or x2 < –p , σp = σ ⃓ . The function
Kp
σ ∈ C ∞ (K) since it is a solution to an elliptic equation in K . Denote
⎧
⎨ σ (x), x ∈ K p,
σp (x) =
⎩
0, x ∈/ Kp .

Then, σp → σ in L1loc (R2 ) as p → 0+ since σ ∈ E. Therefore,

σp (x) → σ (x) in S (R2 ), ρ →0+. (9.2)

Since σp is a discontinuous function and σ is a solution of the homogeneous Helmholtz

equation in K , we have

H (w, D)σp (x) = dp (x), x ∈ R2 ,

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 36 of 44

where

1 [︂
dp (x) = δ(x2 + p)Θ(x1 – p)∂x2 σ (x1 , –p) + σ ′ (x2 + p)Θ(x1 – p)σ (x1 , –p)+
sin2 Φ
+ σ (x1 + p)Θ(x2 + p)∂x σ (–p, x2 ) + σ ′ (x1 + p)Θ(x2 + p)σ (–p, x2 )–
(9.3)
– 2 cos Φ δ(x2 + p)Θ(x1 + p)∂x1 σ (x1 , –p)–
]︂
– 2 cos Φ δ(x1 + p)Θ(x2 + p)∂x2 σ (–p, x2 ) ,

and Θ is the Heaviside function. Equation (9.2) implies that

dp → H σ in S (R2 ), ρ →0+.

Let us calculate the limit d0 of dp as p → 0.

1. The continuity of σ in K and the boundary condition in (9.1) imply that

σ (x1 , –p) → σ10 (x1 ) = 0. (9.4)

By (2.2), we have

|σ (x1 , –p)| ≤ C, p ≥ 0, x1 > 0.

Therefore, (9.4) implies that

Θ(x1 + p)σ (x1 , –p) −→ Θ(x1 )σ10 (x1 ) = 0, in S ′ (R2 ), p→0+.

Hence we have

δ ′ (x2 + p)Θ(x1 + p)σ (x1 , –p) −→ 0, in S ′ (R2 ), p → 0. (9.5)

Similarly, we have

δ ′ (x1 + p)Θ(x2 + p)v0p (–p, x2 ) −→ δ ′ (x1 )Θ(x2 )v02 (x2 ) = 0, in S ′ (R2 ), p→0+.

(9.6)

2. The continuity of ∇σ in K \ {0} implies that

∂x2 σ (x1 , –p) −→ ∂x2 σ (x1 ), x1 > 0.

By (2.2) we have

|∂x2 σ (x1 , –p)| ≤ C1 ρ –1+α + C2 , x1 > 0.

Hence

Θ(x1 + p)∂x2 σ (x1 , –p) −→ ∂x2 σ (x1 , 0) in S ′ (R2 ), ρ → 0+

Merzon et al. Boundary Value Problems (2025) 2025:26 Page 37 of 44

by the Lebesgue theorem, and

δ ′ (x2 + p)Θ(x1 + p)∂x2 σ (x1 , –p) −→ δ ′ (x2 )σ11 (x1 ) in S ′ (R2 ), ρ →0+. (9.7)

Similarly, we have

⃓
δ ′ (x1 + p)Θ(x2 + p)∂x1 σ (–p, x2 ) −→ δ ′ (x1 )σ21 (x2 ) ⃓
⃓
⃓
⃓
–2 cos Φσ (x2 + p)Θ(x1 + p)∂x1 σ (x1 , –p) −→ –2 cos Φ∂x1 v01 (x1 ) = 0 ⃓⃓ p → 0+
⃓
⃓
′
–2 cos Φδ(x1 + p)Θ(x2 + p)∂x2 σ (x2 , –p) −→ 0, in S (R ).
2 ⃓

(9.8)

Finally, (9.3) (9.5), (9.6), (9.7), (9.8) imply that

1 [︂ 1 1
]︂
d0 = δ(x 2 )σ 1 (x 1 ) + δ(x 1 )σ 2 (x 2 ) . (9.9)
sin2 Φ

Applying the Fourier transform to (9.9), we obtain (cf. (3.16))

1 [︂ 1 ]︂
d̃0 (z1 , z2 ) = 2
σ̃1 (z1 ) + σ̃21 (z2 ) , z ∈ R2 .
sin Φ

By the Paley–Wiener theorem the functions ṽ1 (z1 ) and ṽ2 (z2 ) admit an analytic contin-
uation to C+z1 × C and C × C+z2 . Now we can lift ṽ11 (z1 ) and ṽ12 (z2 ) to the Riemann surface
V̂ by formulas (3.17) and use Theorem 3.3. We obtain
[︂ ]︂ [︂ ]︂
σ̂11 (w) + σ̂21 (w) = 0, w ∈ V̂Σ ,

see Definition 3.2, (3.23). Further, we implement all the steps described in Sect. 3.3 and
obtain the difference equation for σ̂11 (w)
⃓
σ̂11 (w) – σ̂11 (w + 2iΦ) = 0, ⃓
⃓
⃓ w ∈ C. (9.10)
⃓
σ̂ (–w + πi) = σ̂ (w) ⃓
1
1
1
1

Moreover, σ̂11 (w) is analytic in C and satisfies the estimate

|σ̂11 (w)| ≤ Cem|Re w| . (9.11)

[︁ ]︁
In fact, by (3.24) σ̂11 (w) satisfies this estimate in VΣ,δ since v̂1 (w) = σ̂11 (w). Moreover, this
estimate is extended to C by the automorphy σ̂11 (–w + πi) = σ̂11 (w), see (9.10). Applying
the reasoning of Sect. 3.3 to the function σ̂11 with zero right-hand side G2 , we obtain the
function σ̌11 (t), which is analytic in Π̌ = C∗ \ {1} where C∗ = C ∪ {∞}, since σ11 is 2iΦ-
periodic by the first equation in (9.10). Further, the point t = 1 is a unique pole of the
function v̌11 by (9.11) and (4.4).
It remains only to prove that σ̌11 (t) −−→ 0. Then, by the Liouville Theorem, σ̌11 (t) =
t→1
σ̂11 (w) = σ̃11 (z1 ) = σ11 (x1 ) = 0, which implies that Uh = 0, and we are done. Let us prove
that indeed σ̌11 (t) → 0, as t → 1.
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 38 of 44

Since σ ∈ Eα , σ11 (x1 ) satisfies the second estimate in (2.2). Hence its Fourier–Laplace
transform satisfies the estimate

(1 + Im z1 )1–α
|σ̃11 (z1 )| ≤ C .
Im z1

Rewriting this estimate in the variable w and using (3.17), we get

[1 + Im (–iω sinh w)]1–α

|σ̂11 (w)| ≤ C , w ∈ V̂11 .
Im (–iω sinh w)

Substituting in this inequality w = w1 + iπ/2 with w1 ≥ 0 (which belongs to V̂11 ), we obtain

(1 + ω cosh w)1–α
|σ̂11 (w)| ≤ C , w1 ≥ 0,
ω2 cosh w1

and therefore σ̂11 (w1 + iπ/2) = O(e–αw1 ), w1 → ∞. Hence σ̌11 (t) → 0 as t → 1 by (4.4). □

10 Conclusion
As is known, an angle is one of the few regions where the boundary value problems for
the Helmholtz equation admit an explicit solution. As far as we know, this has always
been done for decreasing boundary data, where the operator methods are normally used
with the exception of a very specific boundary value problem associated with the plane
incident wave (Sommerfeld’s diffraction problem [38]). In the present work, we solve the
Dirichlet boundary problem not related to the incidence of a plane wave and we obtain an
explicit solution in the form of the Sommerfeld integral. The method proposed is suitable
for the Neumann (NN) and Dirichlet-Neumann (DN) boundary conditions, and for angles
smaller than π . We hope that the method is suitable for solving such problems with a real
wave number in the Helmholtz operator and also for nonstationary problems, e.g., for
the proof of the limiting absorption principle. Finally, we note that the obtained explicit
formulas for the solution allow one to carry out numerical experiments.

Appendix
A.1 Proof of Lemma 3.5
Since v̂11 (w) is meromorphic in V̂Σ by (3.25) and is automorphic with respect to πi/2, we
extend v̂11 by symmetry with respect to πi/2 to h1 V̂Σ = V̂0π +Φ . Namely, define

⎧
⎨ v̂11 (w), w ∈ V̂Σ ,
v11 (w) =
⎩
v̂11 (–w + πi), w ∈ V̂ππ +Φ .

Obviously, v11 (w) is meromorphic on Γπ since v̂11 is meromorphic on Γ0 . We will still use
the notation v̂11 for v11 . Thus (3.30) holds for this extension too.
Since Ĝ(w) is meromorphic in C, by (3.25), v̂12 admits a meromorphic continuation onto
π +Φ
V̂–Φ which we also denote v̂12 . Hence

v̂11 (w) + v̂12 (w) = Ĝ(w), w ∈ V̂–Φ

π +Φ
, (A.1)
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 39 of 44

)︂ continuation. Let us extend v̂2 (w), w ∈ V̂–π , to V̂–π ∪

1 π +Φ π +Φ
by the uniqueness(︂ of analytic
π +Φ π –Φ 1 1 π –Φ
h2 V̂–π = V–3Φ see (3.26) . Similarly to v̂1 , v̂2 is a meromorphic function in V–3Φ .

)︂ to V–3Φ and obtain v̂1 (w) + v̂2 (w) = Ĝ(w), w ∈ V–3Φ . Hence v̂l ∈
π –Φ 1 1 π –Φ 1
Now we extend
(︂ (A.1)
H(V̂l+ ) ∩ M V–3Φ
π –Φ
.
Continuing the process of extension of v̂11 and v̂12 by symmetries (3.26), (3.27), we can
extend equation (A.1) to C and obtain (3.29)–(3.30). □

A.2 Proof of Lemma 3.6

Let us apply the automorphism h2 to equation (3.29). Since, by (3.21), (3.28),

iω sinh(w + 3iΦ)
Ĝ(h2 w) = ,
iω sinh(w + 2iΦ) + k

(3.25) gives

iω sinh(w + 3iΦ)
v̂11 (h2 w) + v̂12 (h2 w) = Ĝ(h2 w) = , w ∈ C. (A.2)
iω sinh(w + 2iΦ) + k

Hence, subtracting equation (A.2) from equation (3.25), we obtain v̂11 (w) – v̂11 (h2 w) =
Ĝ2 (w), where Ĝ2 is given by (3.31).
Using (3.27), we can represent v̂11 (h2 w) as a function with shifted argument. Applying
(3.26) we have

v̂11 (h2 w) = v̂11 (h1 h2 w) = v̂11 (hw),

where h(w) is a shift since hw = h1 h2 w = (–h2 w) + πi = (w – πi + 2iΦ) + πi = w + 2iΦ. Hence,

by (3.31), v̂11 (w) – v̂11 (w + 2iΦ) = Ĝ2 (w). □

A.3 An auxiliary statement

Lemma A.1 Let A ∈ M(C) be h1 - automorphic function, see (3.26). Then res A(w) =
w=w0
– res A.
w=–w0 +π i

The proof of this statement is trivial.

A.4 Proof of estimate (6.3)

For w ∈ Γ0 , ω = ω1 + iω2 , ω2 > 0, consider e–ωρ sinh(w–iτ ) , w ∈ Γ0 , 0 < τ0 ≤ τ <
π – τ0 . We have e–ωρ sinh w cos τ = e–iω sinh w(–iρ cos τ ) = ez1 (w)·(–iρ cos τ ) , where z1 (w) is given
⃓ –ωρ sinh w cos τ ⃓
by (3.17). For w ∈ Γ0 , we have[︂ that z1 (w) ∈ R by (3.18). Hence ⃓e ⃓ = 1 and,
]︂ ⃓ –ωρ sinh(w–iτ ) ⃓
since –ωρ sinh(w – iτ ) = –ωρ sinh w cos τ – i cosh w sin τ , we obtain ⃓e ⃓=
⃓ iωρ cosh w sin τ ⃓ ⃓ –ω ρ sin τ cosh w cos w ⃓ ⃓ –ω ρ sin τ sinh w sin w ⃓
⃓e ⃓ = ⃓e 2 1 2 ⃓ · ⃓e 1 1 2 ⃓.

Since for w ∈ Γ0 , w2 = arctan( ω2 tanh w1 ), (see (3.18)), cos w2 ≥ C(ω) > 0, we have for
ω1

τ ∈ [τ0 , π – τ0 ]
⃓ –ω ρ sin τ cosh w cos w ⃓
⃓e 2 1 2 ⃓ ≤ e–C(ω,τ0 )ρ cosh w1 , C(ω, τ0 ) > 0, w2 ∈ Γ0 , ω2 > 0.

sinh2 w1
⃓ ⃓ ⃓⃓ –ω1 ρ sin τ ⃓
2 √︂
⃓
Moreover, for w ∈ Γ0 we have ⃓e–ω1 ρ sin τ cosh w1 sin w2 ⃓ = ⃓e ω22 cosh2 w1 +ω12 sinh2 w2
⃓ ≤ 1.
Hence, (6.3) follows. □
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 40 of 44

A.5 Analysis of the solution near the ray θ = 3π /2

We prove that u1 (ρ, 3π/2 + δ) is continuous in δ for small δ. By (7.4) it suffices to prove
that
∫︂
1
I(ρ, δ) := e–ωρ sinh w v̂1 (w + 3πi/2 + iδ)dw (A.3)
4π sin Φ
Γ– π
2
←––––

satisfies (see (7.2))

(︂ )︂
I(ρ, 0–) = I(ρ, 0+) + up ρ, 3π/2 (A.4)

since v̂1 does not have poles on Γ–π by (7.1).

3πi
Making the change of the variable w′ → w + in (A.3) it is easy to see that
2
∫︂ (︁ )︁
1 v̂1 (w) w – (–p1 + πi)
I(ρ, 0+) = lim eiωρ cosh(w) dw. (A.5)
δ→0 4π sin Φ w – (–p1 + πi) + iδ
Γπ
←––

Using (7.1), and the Sokhotski–Plemelj theorem, we obtain

1
I(ρ, 0+) == – eiωρ cosh p1 + V .P.,
2

where V .P. here and below denotes the principal value of the integral (A.5) with δ = 0.
Similarly,

1
I(ρ, 0–) = eiωρ cosh p1 + V .P.
2
(︂ )︂
Hence (A.4) follows, since up ρ, 3π/2 = eiωρ cosh p1 by (7.2). □

A.6 A lemma
Lemma A.2 Let ω ∈ C+ ,

∫︂∞
A(ρ) := eiωρ cosh w (cosh w)e–βw dw, (A.6)
0

0 < β < 1. Then

A(ρ) = C(ω)ρ –1+β + o(ρ –1+β ), ρ → 0, (A.7)

Proof Replacing in (A.6) cosh w by (ew )/2, it is not difficult to check that

A(ρ) = A1 (ρ) + A2 (ρ),

where

∫︂∞
w /2
A1 (ρ) = eiωρe (ew /2)e–βw dw
0
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 41 of 44

and A2 , which is the difference between A and A1 , admits a “good” asymptotics

A2 (ρ) = C(ω) + o(1), ρ → 0.

Making the change of the variable ρew = s, we obtain

∫︂∞
1
A1 (ρ) = ρ –1+β eiωs/2 s1–β ds.
2
ρ

Hence (A.7) follows. □

A.7 Proof of (8.16) and (8.17)

Let us prove (8.16). From (8.14) and (8.4), we have
[︄ ∫︂
ω
∂ρ u1 (ρ, θ ) = – e–ωρ sinh w (sinh w)v̂1 (w + iθ )dw
4π sin Φ
γ+ ∪γ–
∫︂ ]︄
–ωρ sinh w
+ e (sinh w)v̂1 (w + iθ )dw .
C+ ∪C–

Similarly to (8.6),
∫︂
e–ωρ sinh w (sinh w)v̂1 (w + iθ )dw = O(1), ρ → 0. (A.8)
γ1 +γ2

By (8.1), and getting rid of the constants c±

0 by the 2πi-periodicity of the functions sinh
and cosh and the definition of the contours C+ and C– , we obtain
∫︂ ∫︂ [︂ sin Φ ]︂
e–ωρ sinh w (sinh w)v̂1 (w + iθ )dw = e–ωρ sinh w (sinh w) ± (w + iθ ) +
Φ
C+ ∪C∫︂1 C+ ∪C–
–ωρ sinh w
+ e (sinh w) R1 (w + iθ )dw
C+ ∪C–

with R1 given by (8.2). Taking into account (8.7) (A.8) and Lemma A.2, we obtain (8.16).
Let us to prove (8.17). By (8.15), (8.4), (8.1) getting rid of the constants c±
0 and iθ similarly
to the previous reasoning and estimating the integral over the finite part of the contour
γ+ ∪ γ– as in (A.8), we get
[︄ ∫︂
iωρ
∂θ u1 = e–ωρ sinh w (cosh w)vˆ1 (w + iθ )dw
4π sin Φ
γ+ ∪γ–
∫︂ ]︄
+ e–ωρ sinh w
(cosh w)vˆ1 (w + iθ )dw
C+ ∪C–
∫︂ (︂ )︂
iωρ
= O(1) + e–ωρ sinh w (cosh w) ± w + R1 (w + iθ ) dw. (A.9)
4πΦ
C+ ∪C–
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 42 of 44

Making the change of variable w → (–w – 3πi) and using automorphy of sinh w with re-
spect to –3πi/2, 2π -periodicity of sinh w, cosh w, and (8.5), we obtain
∫︂ ∫︂
iωρ iωρ
e–ωρ sinh w (cosh w)(–w)dw = e–ωρ sinh w (cosh w)w dw
4πΦ 4πΦ
C– C+
∫︂
ωρ
=– e–ωρ sinh w cos wdw.
2Φ
Γ+ –2π i

Adding to this integral the integral

∫︂
iωρ
e–ωρ sinh w (cosh w)wdw,
4πΦ
C+

reducing by 2πi - periodicity the latter to the integral over contour Γ+ –2πi and calculating
it, we obtain

∫︂ ∫︂
iωρ –ωρ sinh w ωρ eiωρ
e (cosh w)(±w)dw = – e–ωρ sinh w cosh w dw = –
4πΦ Φ Φ
C+ ∪C– Γ+ –2π i

(A.10)

Finally, by (8.3) and Lemma A.2

∫︂
iωρ iω · c
e–ωρ sinh w cosh w R1 (w + iθ )dw = ρ α + o(ρ α ), ρ→0
4πΦ 4π sin Φ
C+ ∪C–

and by (8.15), (A.9) and (A.10), we finish the proof of (8.17) □

Author contributions
All authors participated in the writing of the manuscript.

Funding
PZ, AM, and JEDM were partially financially supported by Sistema Nacional de Investigadores (grants 14536, 21641 and
102939), and MIRR was partially financially supported by Vicerrectoría de la Investigación de la UMNG (grant INV-CIAS
3917).

Data Availability
No datasets were generated or analysed during the current study.

Declarations
Competing interests
The authors declare no competing interests.

Author details
1
Instituto de Física y Matemáticas, Universidad Michoacana, Morelia, Michoacán, México. 2 Facultad de Ciencias
Físico-Matemáticas, Universidad Michoacana, Morelia, Michoacán, México. 3 Facultad de Ciencias Básicas y Aplicadas,
Universidad Militar Nueva Granada, Bogotá, Colombia. 4 Facultad de Matemáticas II, Universidad Autónoma de Guerrero,
Cd. Altamirano, Guerrero, México.

Received: 19 June 2024 Accepted: 12 February 2025

References
1. Babich, V.M., Lyalinov, M.A., Gricurov, V.E.: The Sommerfeld-Malyuzhinets Tecnique in Diffraction Theory. Alpha
Science International, Oxford (2007)
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 43 of 44

2. Bernard, J.M.L., Pelosi, G., Manara, G., Freni, A.: Time domains scattering by an impedance wedge for skew incidence.
In: Proceeding Conference ICEAA, pp. 11–14 (1991)
3. Castro, L.P., Kapanadze, D.: Dirichlet-Neumann impedance boundary-value problems arising in rectangular wedge
diffraction problems. Proc. Am. Math. Soc. 136, 2113–2123 (2008)
4. Castro, L.P., Kapanadze, D.: Wave diffraction by a 45 degree wedge sector with Dirichlet and Neumann boundary
conditions. Math. Comput. Model. 48(1/2), 114–121 (2008)
5. Castro, L.P., Kapanadze, D.: Wave diffraction by a 270 degrees wedge sector with Dirichlet, Neumann and impedance
boundary conditions. Proc. A. Razmadze Math. Inst. 155, 96–99 (2011)
6. Castro, L.P., Speck, F.-O., Teixeira, F.S.: On a class of wedge diffraction problems posted by Erhard Meister. Oper. Theory,
Adv. Appl. 147, 213–240 (2004)
7. Castro, L.P., Speck, F.-O., Teixeira, F.S.: Mixed boundary value problems for the Helmholtz equation in a quadrant.
Integral Equ. Oper. Theory 56, 1–44 (2006)
8. Dos Santos, A.F., Teixeira, F.S.: The Sommerfeld problem revisited: solution spaces and the edge conditions. J. Math.
Anal. Appl. 143, 341–357 (1989)
9. Esquivel, N.A., Merzon, A.E.: An explicit formula for the nonstationary diffracted wave scattered on a NN-wedge. Acta
Appl. Math. 136(1), 119–145 (2014)
10. Kay, I.: The diffraction of an arbitrary pulse by a wedge. Commun. Pure Appl. Math. 6, 521–546 (1953)
11. Keller, J., Blank, A.: Diffraction and reflection of pulses by wedges and corners. Commun. Pure Appl. Math. 4(1), 75–95
(1951)
12. Komech, A.I.: Elliptic boundary value problems on manifolds with piecewise smooth boundary. Math. USSR Sb. 21(1),
91–135 (1973)
13. Komech, A.I., Mauser, N.J., Merzon, A.E.: On Sommerfeld representation and uniqueness in scattering by wedges.
Math. Methods Appl. Sci. 28, 147–183 (2004)
14. Komech, A.I., Merzon, A.E.: General boundary value problem in regions with corners. Oper. Theory, Adv. Appl. 57,
171–183 (1992)
15. Komech, A.I., Merzon, A.E.: Limiting amplitude principle in the scattering by wedges. Math. Methods Appl. Sci. 29(10),
1147–1185 (2006)
16. Komech, A.I., Merzon, A.E.: On uniqueness and stability of Sobolev’s solution in scattering by wedges. Z. Angew.
Math. Phys. 66(5), 2485–2498 (2015)
17. Komech, A.I., Merzon, A.E.: Stationary Diffraction by Wedges. Lect. Notes Math., vol. 2249. Springer, Berlin (2019)
18. Komech, A.I., Merzon, A.E., De La Paz Méndez, J.E.: Time-dependent scattering of generalized plane waves by wedges.
Math. Methods Appl. Sci. 38(18), 4774–4785 (2015)
19. Komech, A.I., Merzon, A.E., Esquivel Navarrete, A., De La Paz Méndez, J.E., Villalba Vega, T.J.: Sommerfeld’s solution as
the limiting amplitude and asymptotics for narrow wedges. Math. Methods Appl. Sci. 42(15), 4957–4970 (2019)
20. Kunz, D., Assier, R.: Diffraction by a right-angled no-contrast penetrable wedge revisited: a double Wiener-Hopf
approach. SIAM J. Appl. Math. 82(4), 1495–1519 (2022)
21. Malyshev, V.A.: Random Walks, Wiener-Hopf Equations in the Quadrant of Plane, Galois Automorphisms. Moscow
University, Moscow (1970). (in Russian)
22. Meister, E.: Some solved and unsolved canonical problems of diffraction theory. Lect. Notes Math. 1285, 320–336
(1987)
23. Meister, E., Passow, A., Rottbrand, K.: New results on wave diffraction by canonical obstacles. Oper. Theory, Adv. Appl.
110, 235–256 (1999)
24. Meister, E., Penzel, F., Speck, F.-O., Teixeira, F.S.: Some interior and exterior boundary-value problems for the Helmholtz
equations in a quadrant. Proc. R. Soc. Edinb. A 123(2), 275–294 (1993)
25. Meister, E., Penzel, F., Speck, F.-O., Teixeira, F.S.: Two canonical wedge problems for the Helmholtz equation. Math.
Methods Appl. Sci. 17, 877–899 (1994)
26. Meister, E., Speck, F.-O., Teixeira, F.S.: Wiener-Hopf-Hankel operators for some wedge diffraction problems with mixed
boundary conditions. J. Integral Equ. Appl. 4(2), 229–255 (1992)
27. Merzon, A.E., De la Paz Méndez, J.E.: DN-scattering of a plane wave by wedges. Math. Methods Appl. Sci. 34(15),
1843–1872 (2011)
28. Merzon, A.E., De La Paz Méndez, J.E., Villalba Vega, T.J.: On the Keller-Blank solution to the scattering problem of
pulses by wedges. Math. Methods Appl. Sci. (2014)
29. Merzon, A.E., Komech, A.I., De la Paz Méndez, J.E., Villalba Vega, T.J.: On the Keller-Blank solution to the scattering
problem of pulses by wedges. Math. Methods Appl. Sci. 38(10), 2035–2040 (2015)
30. Merzon, A.E., Zhevandrov, P.N., De La Paz Mendez, J.E.: On the behavior of the edge diffracted nonstationary wave in
scattering by wedges near the front. Russ. J. Math. Phys. 22(4), 491–503 (2015)
31. Muskhelishvili, N.I.: Cauchy integrals near ends of the line of integration. In: Singular Integral Equations. Springer,
Dordrecht (1958). https://doi.org/10.1007/978-94-009-9994-7_4
32. Oberhettinger, F.: On the diffraction and reflection of waves and pulses by wedges and corners. J. Res. Natl. Bur.
Stand. 61(2), 343–365 (1958)
33. Penzel, F., Teixeira, F.S.: The Helmholtz equation in a quadrant with Robin’s conditions. Math. Methods Appl. Sci. 22,
201–216 (1999)
34. Rid, M., Saymon, B.: Methods of Modern Mathematical Physics II: Fourier Analysis, Self-Adjointness. Academic, New
York (1975)
35. Rottbrand, K.: Exact solution for time-dependent diffraction of plane waves by semi-infinite soft/hard wedges and
half-planes (1998). Preprint 1984 Technical University Darmstadt
36. Rottbrand, K.: Time-dependent plane wave diffraction by a half-plane: explicit solution for Rawlins’ mixed initial
boundary value problem. Z. Angew. Math. Mech. 78(5), 321–335 (1998)
37. Sobolev, S.L.: Theory of diffraction of plane waves. Proc. Seismol. Inst. Acad. Sci. 41, 605–617 (1934). Russian Leningrad
38. Sommerfeld, A.: Mathematical Theory of Diffraction. Progress in Mathematical Physics, vol. 35. Birkhäuser, Basel
(2004). R.J. Nagem, M. Zampolli, G. Sandri (translators)
39. Teixeira, F.S.: Diffraction by a rectangular wedge: Wiener-Hopf-Hankel formulation. Integral Equ. Oper. Theory 14,
436–454 (1991)
40. Tikhonov, A.N., Samarskii, A.A.: Equations of Mathematical Physics. Dover, New York (1990)
Merzon et al. Boundary Value Problems (2025) 2025:26 Page 44 of 44

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