ELE 2612 – Draft of Lecture 2.

On the Electromagnetic Fields in Cylindrically

Layered Uniaxially-Anisotropic Media
Part I
Guilherme Simon da Rosa

(Rev. 2020.10.09)

Contents
1 Electromagnetic Fields in Homogeneous Anisotropic Media 2
1.1 Axial Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Solving the Wave Equation via the Method of Separa-
tion of Variables . . . . . . . . . . . . . . . . . . . . . 3
1.2 Transversal Fields . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Electromagnetic Fields Along Layered Anisotropic Media 8

2.1 Reflection and Transmission from a Two-Layer Medium . . . . 8
2.1.1 Outgoing wave in region 1 . . . . . . . . . . . . . . . . 9
2.1.2 Ingoing wave in region 2 . . . . . . . . . . . . . . . . . 10
2.1.3 Summary of the local R and T matrices . . . . . . . . 11

References 12

1
1 Electromagnetic Fields in Homogeneous
Anisotropic Media
The Maxwell’s equations in a linear homogeneous and anisotropic medium
are
∇ × E = iω µ̄ ¯ · H, (1)
∇ × H = −iω ¯ · E + J, (2)
∇ · (¯ · E) = %, (3)
¯ · H) = 0,
∇ · (µ̄ (4)
where E and H are the electric and the magnetic fields, respectively, due to
the impressed volumetric electric current and charge densities J and %. Our
media is characterized by the complex permeability
¯ = diag(µs , µs , µz ), with µ{s,z} = µ0 µ{s,z}
µ̄ (5)
and permittivity
¯ = diag(s , s , z ), with {s,z} = 0 r{s,z} + iσ{s,z} /ω (6)
tensors, represented in cylindrical coordinates. We have decomposed the
tensors into axial (along z) and transverse components, with subscripts z
and s, respectively. We have also expressed the constitutive parameters in
terms of the usual vacuum constants 0 and µ0 .
We next consider the fields in a source-free region and decompose the
problem into axial and transversal components as
F = Fs + ẑFz , (7)
in which F = E or H, Fz = Ez or Hz , and
∂
∇ = ∇s + ẑ
, (8)
∂z
where ∇s denotes the transverse two-dimensional nabla operator given by
∂ 1 ∂
∇s = ρ̂ + φ̂ . (9)
∂ρ ρ ∂φ

1.1 Axial Fields

Similarly to the decomposition presented in Lecture 1.3, for Cartesian coor-
dinates, we have
z ∂ 2 Ez
∇2s Ez + + ω 2 µs z Ez = 0, (10)
s ∂z 2

2
 µz ∂ 2 Hz
∇2s Hz + + ω 2 µz s Hz = 0, (11)
µs ∂z 2
in which the two-dimensional Laplacian operator in polar coordinates is de-
fined by

∂2
∇2s = ∇2 − (12)
∂z 2
= ∇s · ∇s (13)
1 ∂ ∂ 1 ∂2
= ρ + . (14)
ρ ∂ρ ∂ρ ρ2 ∂φ2

1.1.1 Solving the Wave Equation via the Method of Separation of

Variables
It is clear that (10) and (11) are homogeneous Helmholtz wave equations for
Ez and Hz whose solutions can be constructed by the method of separation
of variables [1, Ch. 5]. The Helmholtz equation in cylindrical coordinates

1 ∂2 pz ∂ 2
 
1 ∂ ∂ 2
ρ + + + ω pz p̃s Fz = 0, (15)
ρ ∂ρ ∂ρ ρ2 ∂φ2 ps ∂z 2

where Fz = {Ez , Hz }, p = {, µ} and p̃ = {µ, }, can be solved by the

elementary wave functions. Following the method of separation of variables,
we seek to find solutions of the form

Fz = R(ρ)Φ(φ)Z(z). (16)

Substitution of (16) into (15), division by Fz , and noting that the partial
derivatives now becomes total derivatives, allow us to derive

1 d2 Φ pz 1 d2 Z
 
1 d dR
ρ + 2 + + ω 2 pz p̃s = 0, (17)
ρR dρ dρ ρ Φ dφ2 ps Z dz 2

where the third term is explicitly independent of ρ and φ. This term must
also be independent of z because the equation is to sum to zero for all ρ, φ, z.
Hence, we can define
1 d2 Z
2
= −kz2 , (18)
Z dz
where kz is a constant. Substitution of the above into (17) multiplied by ρ2
gives
1 d2 Φ
   
ρ d dR 2 pz 2 2
ρ + + ω pz p̃s − kz ρ = 0. (19)
R dρ dρ Φ dφ2 ps

3
Now, the second term of the above is independent of ρ and z, and the re-
maining terms are independents of φ. Hence, we can properly define
1 d2 Φ
= −n2 , (20)
Φ dφ2
where n is a constant. Now, the preceding equation gives rise to
   
ρ d dR 2 2 pz 2 2
ρ − n + ω pz p̃s − kz ρ = 0. (21)
R dρ dρ ps
The wave equation is now separated by means of (18), (20) and (21). To
summarize, we can introduce
kρ2 = ω 2 ps p̃s − kz2 , (22)
and also the wavenumber related to the transversal anisotropy given by
ks2 = ω 2 ps p̃s = ω 2 s µs . (23)
We can write the separated equations as
  "r 2 #
d dR pz
ρ ρ + kρ ρ − n2 R = 0, (24)
dρ dρ ps
d2 Φ
+ n2 Φ = 0, (25)
dφ2
d2 Z
+ kz2 Z = 0. (26)
dz 2
The expression in (24) is a Bessel differential equation of order n [2,
p. 358], whose general solution is given by
r  r  r 
pz pz pz
Rn kρ ρ = AJn kρ ρ + BYn kρ ρ , (27)
ps ps ps
where Jn and Yn are fisrt and second kind Bessel functions, respectively. The
constants A and B are determined by the boundary conditions. Another
commonly used solution to Bessel’s equation employ the first and second
kind Hankel functions of order n:
r  r  r 
pz (1) pz (2) pz
Rn kρ ρ = AHn kρ ρ + BHn kρ ρ , (28)
ps ps ps
where
Hn(1) (x) = Jn (x) + iYn (x), Hn(2) (x) = Jn (x) − iYn (x). (29)

4
 (1) (2)
In general, any two of the functions Jn , Yn , Hn and Hn are linearly inde-
pendent and can be used as a fundamental of solutions of (24). Note that
only the function Jn is nonsingular at ρ = 0. Hence, if a field is to be finite
at the radial origin, we have Rn (·) = AJn (·). In contrast, for an unbounded
(1)
radial domain, we must select Rn (·) = AHn (·) in order to satisfy the Som-
merfeld radiation condition.
Equations (25) and (26) are one-dimensional non-homogeneous Laplace’s
equations whose solutions are given by harmonic functions such as

Φ(nφ) = C cos(nφ) + D sin(nφ) or Ceinφ + De−inφ , (30)

Z(kz z) = E cos(kz z) + F sin(kz z) or Eeikz z + F e−ikz z , (31)

where C, D, E and F are constants to be determined by enforcing the bound-

ary conditions. The azimuthal coordinate is an angle, and for this reason,
we need to restrict the choice of Φ(nφ). If the domain of our problem is
over 0 ≤ φ < 2π, in order to achieve single-valued fields, it is required that
Fz (φ) = Fz (φ+2πp), p = 0, 1, 2, . . . . Thus, we have immediately that n must
be an integer. Finally, the elementary solution for (15) can be written as
r 
pz
Fz,kρ ,n,kz = Rn kρ ρ Φ(nφ)Z(kz z). (32)
ps

Linear combination of the elementary wave functions are also a solution

to (15) [1, p. 200], such that the general solution is in the shape of
XZ
Fz = dkρ C(kρ , n) Fz,kρ ,n,kz , or (33)
n kρ
XX
Fz = Ckρ ,n Fz,kρ ,n,kz , (34)
n kρ

where the continuous function C(kρ , n) or the discrete constants Cρ,n are
given by a source boundary condition.
It is important to emphasize that the longitudinal dependence of fields
render ∂ 2 /∂z 2 → −kz2 , for both Ez and Hz . As a consequence, the separation
equation in (22) is

ks2 = kρ2 + kz2 , (35)

again, for both Ez and Hz . The radial derivatives, contrast, will be propor-
tional to the anisotropic factor (pz /ps )1/2 kρ .

5
 The forward-propagating axial fields can be written in a compact fashion
as   ∞ X
∞  
Ez X ez (ρ) inφ+ikz,np z
= e . (36)
Hz hz (ρ)
n=−∞ p=1

Notice that the sum over the azimuthal indices is taken through the positive
and negative integer values of n due to the choice of the exponential form
exp(inφ). Another commonly used form is a linear combination of sine and
cosine functions, but in this case, the sum could be restricted over the pos-
itive integer values of n due the parity of the harmonic sinusoidal functions
in relation to its arguments. Also, as we intend to express the sources con-
tribution into the axial direction, the form exp(ikz,np z) appears appropriated
indeed. Thus, we prefer the form shown in (36) because the fields can be
represented in a compacted shape.
The modal propagation constant in the z-direction is kz,np , and the radial
propagation constant kρ,np satisfies
2
kρ,np = ks2 − kz,np
2
, ks2 = ω 2 µs s . (37)
The azimuthal dependence of the fields is associated to index n, while p
refers to its radial dependency. Due to phase matching, the fields assume the
exp(ikz z) dependence throughout all layers over the transverse cross-section.
In order to simplify the notation, we will temporarily drop the modal
subscript np and also the argument of ez and hz , restoring them later as
needed. The radially-dependent coefficients in (36) are given by a linear
combination of first-kind Hankel and Bessel function of integer order n. In a
compact notation, we can write them as
 
ez ¯ (k ρ)ā + J¯ (k ρ)b̄,
= H̄zn ρ zn ρ (38)
hz
where
Gn (kρe ρ)
 
¯ 0
Ḡzn (kρ ρ) = , (39)
0 Gn (kρh ρ)
(1)
with Gn representing Hn or Jn . The field amplitudes ā and b̄ are vectors
2 × 1 to be determined from the boundary conditions, and are given by
 e  e
a b
ā = h , b̄ = h . (40)
a b
For simplicity, we have introduced kρe,h = αe,h kρ in which the anisotropic
coefficients αe,h are given by
r r
e z h µz
α = , α = . (41)
s µs

6
 Under some conditions, the formulas in (38) suffer from the numerical
overflow of the Bessel and Hankel functions due extreme arguments and/or
orders. Recent studies have shown that this problem can be circumvented
through suitable rescaling of various expressions involved in the computa-
tional chain [3, 4], as well as expressing (38) by the ratios of Bessel or Hankel
functions [5]. For extreme conductivity media the techniques reported in
[3, 4, 5] need to be used.

1.2 Transversal Fields

The transversal fields components can be expressed as a combinations of the
axial ones, as will be shown in the following. The Maxwell’s curl equations
(1) and (2) in a source-free region can now be rewritten as

∇s × (ẑEz ) ± ikz ẑ × Es = iωµs Hs (42a)

∇s × (ẑHz ) ± ikz ẑ × Hs = −iωs Es (42b)

∇s × Es = iωµz ẑHz (43a)

∇s × Hs = −iωz ẑEz , (43b)
where we have assumed that the axial dependence of all field vectors match
with those in (36), i.e., exp(ikz z). In the above derivation we have replaced
∂/∂z by ikz .
Initially, by the vector multiplication of ẑ on both sides of the equations
in (42a) and (42b), and using the identities [6, p. 165]

ẑ × (∇s × ẑFz ) = ∇s Fz , and (44)

ẑ × (ẑ × Fs ) = −Fs , (45)

and after a few manipulations, we can express the transversal fields as a
combinations of the axial ones via
1
Es = [ikz ∇s Ez + iωµs ∇s × (ẑHz )] , (46)
kρ2
1
Hs = 2 [ikz ∇s Hz − iωs ∇s × (ẑEz )] . (47)
kρ

In a compact shape, the ρ-dependent transversal fields can be written as

 
eφ ¯ (k ρ)ā + J¯ (k ρ)b̄,
= H̄φn ρ φn ρ (48)
hφ

7
  
eρ ¯ (k ρ)ā + J¯ (k ρ)b̄,
= H̄ρn ρ ρn ρ (49)
hρ
where
−nkz Gn (kρe ρ) −iωµs kρh ρG0n (kρh ρ)
 
¯ 1
Ḡφn (kρ ρ) = 2 , (50)
kρ ρ iωs kρe ρG0n (kρe ρ) −nkz Gn (kρh ρ)
1 ikz kρe ρG0n (kρe ρ) −nωµs Gn (kρh ρ)
 
¯
Ḡρn (kρ ρ) = 2 , (51)
kρ ρ nωs Gn (kρe ρ) ikz kρh ρG0n (kρh ρ)

where the 0 represents the first derivative with respect to argument of the
(1)
cylindrical functions. Similarly to previous definitions, Gn stands for Hn in
the matrices associated to ā. In addition, Gn stands for Jn in the matrices
associated to b̄. The field components in (38) and (48) reduces to ones found
in [7] and [8]1 for αh = 1, and reduces to the usual isotropic solution found in
[9, Ch. 3] for isotropic media (αe,h = 1). Note that in the above expressions
for the transverse fields, the φ and z dependent exponential factor has been
omitted for simplicity. Also, the full expression involves the sums over n and
p indices, as expressed in (36).

2 Electromagnetic Fields Along Layered

Anisotropic Media
2.1 Reflection and Transmission from a Two-Layer
Medium
Consider a cylindrical wave propagating in region 1, that is incident on region
2 at ρ = r1 . At the interface, the following boundary conditions should be
observed:

ρ̂ × E1 = ρ̂ × E2 (52)
ρ̂ × H1 = ρ̂ × H2 . (53)

Due to the translational invariance of the medium in the azimuthal and

longitudinal directions (along φ and z), field solutions for all layers must have
the einφ+ikz z variation, i.e., n = nj and kz = kjz for each j = {1, 2, 3, . . . }
layer. As a result, we can write the component α = {ρ, φ, z} of a field at
1
We found a misprint in the equation for the azimuthal electric field in [8, eq. 3]: the
contribution of the axial electric field should be multiplied by ikz .

8
the jth region as the superposition of radially forwardly- and backwardly-
propagating fields, namely,
h i
¯ ¯
F̄jα = H̄αn (kjρ ρ)āj + Jαn (kjρ ρ)b̄j g(nφ, kz z) (54)
= f¯jα (ρ)g(nφ, kz z). (55)

We can verify that pure TMz and TEz fields no longer can satisfy the
boundary conditions at a cylindrical interface for n 6= 0, and we should
consider hybrid fields in what follows. By following the approach presented
in [9, Ch. 3], we the reflection and transmission of waves at a cylindrical
interface we be accounted by considering first a radially outgoing wave in
region 1, and then a radially ingoing wave in region 2.

2.1.1 Outgoing wave in region 1

Consider a cylindrical outgoing wave present in region 1. This wave will
experience a reflection at interface ρ = r1 , and the total field in region 1 can
be expressed as
¯ (k ρ)ā + J¯ (k ρ)b̄
f¯1α = H̄ (56)
1α 1ρ 1 1α 1ρ 1
h i
= H̄¯ (k ρ) + J¯ (k ρ)R̄
¯ (ρ) ā , with α = {ρ, φ, z}. (57)
1α 1ρ 1α 1ρ 12 1

In region 2, only radially outgoing fields are present, and we can express
them via
¯ (k ρ)ā
f¯2α = H̄ (58)
2α 2ρ 2
¯ (k ρ)T̄¯(ρ) ā , with α = {ρ, φ, z}.
= H̄ (59)
2α 2ρ 12 1

2
Notice that kjρ = ±(kjs − kz )1/2 , with =m(kjρ ) ≥ 0. The enforcement of
boundary conditions (52) and (53) at ρ = r1 requires the continuity of the
tangential components of the electric and magnetic field

f¯1z |ρ=r1 = f¯2z |ρ=r1 (60)

f¯1φ |ρ=r1 = f¯2φ |ρ=r1 . (61)

In view of (57) and (59), the above becomes

h i
¯ (k r ) + J¯ (k r )R̄
H̄ ¯ (ρ) ā = H̄
¯ (k r )T̄¯(ρ) ā (62)
1z 1ρ 1 1z 1ρ 1 12 1 2z 2ρ 1 12 1
h i
¯ (k r ) + J¯ (k r )R̄
H̄ ¯ (k r )T̄¯(ρ) ā
¯ (ρ) ā = H̄ (63)
1φ 1ρ 1 1φ 1ρ 1 12 1 2φ 2ρ 1 12 1

9
The above coupled matrix equations can be used to solve R̄ ¯ and T̄¯ , yield-
12 12
ing
h i
¯ (ρ) = D̄
R̄ ¯ −1 H̄¯ H̄ ¯ − ¯ H̄
H̄ ¯ H̄ ¯ −1 H̄¯
12 1a φ2,1 z1,1 φ2,1 z2,1 φ2,1 φ1,1 , (64)
h i
T̄¯12 = D̄¯ −1 J¯ H̄ ¯ ¯ ¯ ¯−1 ¯
(ρ)
1b φ1,1 z1,1 − Jφ1,1 Jz1,1 Jφ1,1 H̄φ1,1 , (65)
¯ = H̄
¯ H̄ ¯ ¯ −1 ¯ ¯ ¯
D̄1a φ2,1 z2,1 H̄φ2,1 Jφ1,1 − H̄φ2,1 Jz1,1 , (66)
¯ = J¯ H̄ ¯ ¯ ¯ ¯−1 ¯
D̄ 1b φ1,1 z2,1 − Jφ1,1 Jz1,1 Jφ1,1 H̄φ2,1 . (67)
In the above we have employed a shorthand notation, such that
¯
B̄ ¯
αi,j = B̄αn (kiρ rj ), (68)
¯ = H̄
where B̄ ¯ or J¯ and α = {φ, z}.
αn αn αn

2.1.2 Ingoing wave in region 2

Consider a cylindrical ingoing wave present in region 2. This wave will ex-
perience a reflection at interface ρ = r1 , and the total field in region 2 can
be expressed as
h i
f¯2α = J¯2α (k2ρ ρ) + H̄
¯ (k ρ)R̄
2α 2ρ
¯ (ρ) ā , with α = {ρ, φ, z}.
21 2 (69)

In region 1, only radially ingoing fields are present, and we can express them
via
f¯ = J¯ (k ρ)T̄¯ ā , with α = {ρ, φ, z}.
(ρ)
1α 1α 1ρ 21 2 (70)
The enforcement of boundary conditions (52) and (53) at ρ = r1 requires
f¯1z |ρ=r1 = f¯2z |ρ=r1 (71)
f¯1φ |ρ=r1 = f¯2φ |ρ=r1 , (72)
or
h i
J¯1z (k1ρ r1 )T̄¯21 ā2 = J¯2z (k2ρ r1 ) + H̄
¯ (k r )R̄¯ (ρ) ā
(ρ)
2z 2ρ 1 21 2 (73)
h i
J¯1φ (k1ρ r1 )T̄¯21 ā2 = J¯2φ (k2ρ r1 ) + H̄
¯ (k r )R̄¯ (ρ) ā
(ρ)
2φ 2ρ 1 21 2 (74)
¯ and T̄¯ , yield-
The above coupled matrix equations can be used to solve R̄21 21
ing
h i
¯ (ρ) = D̄
R̄ ¯ −1 J¯ J¯ J¯−1 J¯ − J¯ J¯
21 1b φ1,1 z1,1 φ1,1 φ2,1 φ1,1 z2,1 , (75)
h i
¯ (ρ) ¯ −1 ¯ ¯ ¯ −1 ¯ ¯ ¯
T̄21 = D̄1a H̄φ2,1 H̄z2,1 H̄φ2,1 Jφ2,1 − H̄φ2,1 Jz2,1 , (76)

10
2.1.3 Summary of the local R and T matrices
The local reflection and transmission matrices can be succinctly written as
h i
¯ (ρ) = D̄
R̄ ¯ −1 H̄¯ ¯ − H̄
H̄ ¯ ¯ ¯ −1 H̄ ¯
j,j+1 ja φj+1,j zj,j H̄ H̄
φj+1,j zj+1,j φj+1,j φj,j , (77a)
h i
¯ (ρ) = D̄
R̄ ¯ −1 J¯ J¯ J¯−1 J¯ − J¯ J¯ , (77b)
j+1,j jb φj,j zj,j φj,j φj+1,j φj,j zj+1,j
h i
T̄¯j,j+1 = D̄
¯ −1 J¯ H̄ ¯ ¯ ¯ ¯−1 ¯
(ρ)
jb φj,j zj,j − Jφj,j Jzj,j Jφj,j H̄φj,j , (77c)
h i
T̄¯j+1,j = D̄¯ −1 H̄¯ ¯ ¯ −1 ¯ ¯ ¯
(ρ)
ja φj+1,j H̄zj+1,j H̄φj+1,j Jφj+1,j − H̄φj+1,j Jzj+1,j , (77d)

where j = 1, 2, . . . , N − 1 and
¯ = H̄
¯ ¯ ¯ −1 ¯ ¯ ¯
D̄ja φj+1,j H̄zj+1,j H̄φj+1,j Jφj,j − H̄φj+1,j Jzj,j , (78a)
¯ = J¯ H̄ ¯ ¯ ¯ ¯−1 ¯
D̄ jb φj,j zj+1,j − Jφj,j Jzj,j Jφj,j H̄φj+1,j . (78b)

The notation employed here is very similar to that introduced in [9]. How-
ever, we employ the extra superscript (ρ) in all scattering matrices relating
the radial discontinuities.
The matrices R̄¯ (ρ) and T̄¯(ρ) in (77) can be reduced exactly to those shown
in [10, eq. 31]. Despite the fact that the latter expressions are more compact,
we adopt the forms shown above because, as it will be shown later, poles
stemming from matrix inversions can be more easily addressed based on D̄ ¯
ja
¯ ¯ ¯
and D̄jb . Also note that for isotropic media H̄zi,j and Jzi,j become diagonal
matrices, and (77) reduces to the well-know expressions for isotropic media
[9, Ch. 3], as expected. For further details, please see [11, 12, 13, 14].

11
References
[1] R. F. Harrington, Time-harmonic electromagnetic fields. New York,
NY, USA: McGraw-Hill, 1961.

[2] M. Abramowitz and I. A. Stegun, Handbook of Mathematical Functions:

With Formulas, Graphs, and Mathematical Tables. New York, NY,
USA: Dover Publications, 1964.

[3] K. Sainath, F. L. Teixeira, and B. Donderici, “Robust computation of

dipole electromagnetic fields in arbitrarily anisotropic, planar-stratified
environments,” Phys. Rev. E, vol. 89, p. 013312, Jan. 2014. [Online].
Available: http://link.aps.org/doi/10.1103/PhysRevE.89.013312

[4] H. Moon, F. L. Teixeira, and B. Donderici, “Stable pseudoanalytical

computation of electromagnetic fields from arbitrarily-oriented dipoles
in cylindrically stratified media,” J. Comp. Phys., vol. 273, pp. 118–142,
Sep. 2014.

[5] D. Hong, W. F. Huang, and Q. H. Liu, “Radiation of arbitrary magnetic

dipoles in a cylindrically layered anisotropic medium for well-logging
applications,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 11, pp.
6362–6370, Nov. 2016.

[6] J. Kong, Electromagnetic Wave Theory. New York, NY, USA: Wiley,
1986.

[7] G.-S. Liu, F. L. Teixeira, and G.-J. Zhang, “Analysis of directional log-
ging tools in anisotropic and multieccentric cylindrically-layered earth
formations,” IEEE Trans. Antennas Propag., vol. 60, no. 1, pp. 318–327,
Jan. 2012.

[8] H. Wang, P. So, S. Yang, W. Hoefer, and H. Du, “Numerical modeling of

multicomponent induction well-logging tools in the cylindrically strat-
ified anisotropic media,” IEEE Trans. Geosci. Remote Sens., vol. 46,
no. 4, pp. 1134–1147, Apr. 2008.

[9] W. C. Chew, Waves and Fields in Inhomogeneous Media. New York,

NY, USA: John Wiley & Sons, 1995.

[10] H. Moon, B. Donderici, and F. L. Teixeira, “Stable evaluation of Green’s

functions in cylindrically stratified regions with uniaxial anisotropic lay-
ers,” J. Comp. Phys., vol. 325, pp. 174–200, 2016.

12
[11] G. S. Rosa, J. R. Bergmann, and F. L. Teixeira, “A robust
mode-matching algorithm for the analysis of triaxial well-logging
tools in anisotropic geophysical formations,” IEEE Trans. Geosci.
Remote Sens., vol. 55, no. 5, pp. 2534–2545, May 2017, DOI:
10.1109/TGRS.2016.2646839.

[12] ——, “A perturbation method to model electromagnetic well-logging

tools in curved boreholes,” IEEE Trans. Geosci. Remote Sens., vol. 56,
no. 4, pp. 1979–1993, Apr. 2018, DOI: 10.1109/TGRS.2017.2771723.

[13] ——, “Full-wave pseudoanalytical methods for cylindrically symmetric

electromagnetic structures: Applications in sensing and telemetry for
geophysical exploration,” in Antennas, Antenna Arrays and Microwave
Devices: Electromagnetic Analysis and Applications, T. Bauernfeind,
Ed. World Scientific Publishing Co, 2020, ch. 10.

[14] G. S. Rosa, “Pseudo-analytical modeling for electromagnetic well-

logging tools in complex geophysical formations,” Ph.D. dissertation,
Dept. Elect. Eng., Pontifical Catholic University of Rio de Janeiro, Rio
de Janeiro, RJ, Brazil, 2017, DOI: 10.17771/PUCRio.acad.30559.

13