PHYSICAL REVIEW A 95, 063839 (2017)

Light scattering by dielectric bodies in the Born approximation

A. S. Bereza,1,2 A. V. Nemykin,1,2 S. V. Perminov,3 L. L. Frumin,1,2 and D. A. Shapiro1,2,*

1
Institute of Automation and Electrometry, Russian Academy of Sciences, Siberian Branch, 1 Koptjug Avenue, Novosibirsk 630090, Russia
2
Novosibirsk State University, 2 Pirogov Street, Novosibirsk 630090, Russia
3
Rzhanov Institute of Semiconductor Physics, Russian Academy of Sciences, Siberian Branch, 13 Lavrent’yev Avenue,
Novosibirsk 630090, Russia
(Received 23 April 2017; published 23 June 2017)

Light scattering is one of the most important elementary processes in near-field optics. We build up the Born
series for scattering by dielectric bodies with sharp boundaries. The Green’s function for a two-dimensional
homogeneous dielectric cylinder is obtained. As an example, the formulas are derived for a scattered field of two
parallel cylinders. The polar diagram is shown to agree with the numerical calculation by the known methods of
discrete dipoles and boundary elements.

DOI: 10.1103/PhysRevA.95.063839

I. INTRODUCTION The second-order approximation is needed for numerical

reconstruction of a shallow buried object by the scattered
In the past decades substantial progress has been achieved
amplitude [19]. The resonant-state expansion approximation
in nano-optics [1–3]. However, a significant methodological
uses second-order terms to find eigenfrequencies in an optical
deficiency still persists even for basic problems, like scattering
fiber waveguide [20]. However, the traditional Born series is
by nanosized bodies. Unlike “macroscopic” optics, where
not applicable to a system of dielectric bodies with sharp edges,
transverse waves (for instance, plane or spherical) are very
as it does not satisfy the boundary conditions.
useful to study, say, diffraction and interference, in the
The main goal of the present work is to construct a
subwavelength region the treatment of these phenomena
modified Born approximation for a set of dielectric bodies. The
becomes much more complicated. The reason is evanescent
integral relations are derived and the series for two dielectric
waves near the boundary of illuminated objects. Such waves
cylinders is obtained. We managed to account for the first
usually can be neglected in optical processes with large
cylinder exactly by means of the special Green’s function for
scatterers, but in nano-optics this is not the case. Strong
a cylindrical dielectric, which intrinsically includes multiple
coupling via evanescent waves is the key feature that most
scattering processes with this cylinder. Thus, another aim of
practical nanophotonics tasks focus on. They include a light
our paper is to derive that special Green’s function.
energy concentration within a few-nanometer range [4], high-
The Born series is constructed in Sec. II. Scattering by two
efficiency broad-band solar cells [5], light-induced forces at
cylinders, considered in Sec. III, illustrates the application
the nanoscale [6,7], surface-enhanced Raman spectroscopy
of the developed approach. The obtained formulas are in
[8], and the tomographic reconstruction of a nanostructure [9].
agreement with numerical calculations using surface integral
Only a few problems allow analytical solutions in photon-
equations and the discrete dipole approximation (DDA). The
ics. Along with the classical papers on one cylinder [10,11],
Green’s function is derived in Appendix A: the expressions for
the scattering from two cylinders [12] and two perfectly
a source point inside and outside the dielectric are given for the
conducting spheres [13] can be found in the quasistatic limit
cases of both p and s waves. The boundary element method
using bipolar coordinates; a perfectly conducting cylinder
(BEM) has been discussed in our works [21–23] devoted to the
near a surface was considered in terms of cylindrical wave
scattering by cylinders on a dielectric substrate. The formulas
expansion [14]. In more complicated cases, numerical or
for the two-dimensional discrete dipole method are derived in
semianalytical methods become the only ones capable of
Appendix B.
calculating electromagnetic fields in both near and far regions,
for instance, in a system of several cylinders or a periodic chain
of them [15–17]. II. BORN SERIES
Analytical approximations are very useful for understand-
ing the scattering properties of a structure, at least for testing The Helmholtz equations for magnetic fields inside and
numerical methods. There is a universal method to derive the outside the dielectric (denoted by the subscripts “in” and “out,”
formulas based on the Born approximation. It consists in taking respectively) are
the incident field in place of the total field at each point inside    
 + k12 Hin (r) = 0,  + k02 Hout (r) = 0, (1)
the scattering potential. If the scatterer is not sufficiently weak,
the next approximations are exploited. Several recent optical where  is the two-dimensional Laplace operator with respect
research studies have been devoted to high-order terms of the
approximation. In optical diffusion tomography high orders √ the x and y variables, wave numbers k0 = ω/c and k1 =
to
εω/c, c is the speed of light, ω is the frequency, and ε is the
are necessary for solving the nonlinear inverse problem [18]. dielectric permittivity. The field Hout in free space is slightly
changed due to a weak perturbation, which is small enough
(i.e., k1 a  1, where a is its size) and/or low-polarizable
*
shapiro@iae.nsk.su (|ε − 1|  1). The internal field Hin can be quite different.

2469-9926/2017/95(6)/063839(6) 063839-1 ©2017 American Physical Society

BEREZA, NEMYKIN, PERMINOV, FRUMIN, AND SHAPIRO PHYSICAL REVIEW A 95, 063839 (2017)

The integral over  can be calculated explicitly by the known

Γ relation for the Wronskian determinant [24]:

 π
(1) 
D+ −
kρ
H0 (kρ) + i cos ϕH1(1) (kρ) eikρ cos ϕ dϕ
4i −π
γ −

D = 2π kρ J1 (kρ)H0(1) (kρ) − J0 (kρ)H1(1) (kρ) = 1.

FIG. 1. Domains of integration D+ and D− for Eq. (3). The
boundary of D− consists of γ = ∂D+ (solid line) and the external
infinitely remote contour  (dashed line). Arrows indicate external
normals n to ∂D− .
The Green’s function obeys the equation
( + k 2 )G = δ(r − r ).
Then this integral reproduces the field of a plane incident wave
(0)
Hout = H0 eikr . Equations (6) are similar to boundary integral
equations; the only difference is the absence of the factor 1/2
in the terms outside the integral. These terms are given within
the external or internal limit, in contrast to boundary equations,
where they are determined directly at the contour [26].
The successive approximation series can be built up for
(2) both external and internal fields:

Here k is the corresponding value: k = k0 or k = k1 . (0) (1)

We use Eqs. (1) and (2) to derive the relations between Hout = Hout + Hout + · · · , Hin = Hin
(0)
+Hin
(1)
+ ··· . (7)
amplitudes at the boundary:
 Then from (6) we get the recurrent relations:
Hin (r) = [Hin (r )Gp − Gp Hin (r )]dS  ,
D+ 

(j ) 
n Hout (r )∇Gp − εGp ∇Hout (r ) dl  ,
(j )
Hin
(j )
(r) =
Hout (r) = [Hout (r )G − GHout (r )]dS  , (3) γ
D−  
(j +1) G (j )  (j ) 

where dS is the element of integration over the dielectric Hout (r) = n ∇Hin (r ) − Hin (r )∇G dl  . (8)
ε
domain D+ or free-space domain D− (Fig. 1). The Green’s γ

function G(r,r ) describes free space, and Gp (r,r ) is a similar

function that corresponds to a dielectric with permittivity ε. The approximation exactly takes into account the boundary
The Green’s function of free space is the solution of Eq. (2) conditions, what is distinct from the Born approach in quantum
with k = k0 and can be written as mechanics. This is to emphasize that the shape of the contour
γ can be arbitrary; the circular cylinder (considered in the next
1 (1)
G(r,r ) = H (k|r − r |) section) is, basically, just the simplest example. The dielectric
4i 0 region D− could be unconnected; in that case the contour γ is
∞ 
1  im(ϕ−ϕ  ) Hm(1) (kr )Jm (kr  ), r > r  , the sum of all the boundaries of dielectric domains.
= e
4i m=−∞ Hm(1) (kr  )Jm (kr ), r  > r ,
(4) III. SCATTERING BY TWO CYLINDERS

where Jm , and Hm(1) are Bessel and Hankel functions [24]. Let us now consider two cylinders (see Fig. 2). There
The boundary conditions are are three domains, with different dielectric permittivities.
√ The
  Helmholtz equation (2), is valid for k0 = ω/c, k1 = εω/c,
1 ∂Hin  ∂Hout 
Hin |γ = Hout |γ , = , (5)
ε ∂r γ ∂r γ

where γ is the contour separating the D− and D+ domains. ỹ

Applying Green’s theorem [25] to Eq. (3) we can reduce the
surface integral as y r̃

r
R x̃
Hin (r) = n[Hout (r )∇Gp − εGp ∇Hout (r )]dl  , E0 εp , b
γ α
  r̃
Hout (r) = −  1
n Hin (r )∇G − G∇Hin (r ) dl 
 H0k0 ε, a
x
γ ε r

+ n[Hout (r )∇G − G∇Hout (r )]dl  , (6) FIG. 2. Scheme of p-wave scattering by two parallel cylinders.

The dielectric permittivity and the radius are indicated in the first (left)
where dl  is the element of path, n is the unit vector along and second (right) cylinders. The external infinitely remote contour
the external normal, and  is some remote contour (Fig. 1).  is not shown.

063839-2
LIGHT SCATTERING BY DIELECTRIC BODIES IN THE . . . PHYSICAL REVIEW A 95, 063839 (2017)

90° 90°
105° 75° 105° 75°
120° 60° 120° 60°

135° 45° 135° 45°

150° 30° 150° 30°

165° 15° 165° 15°

180° 0 180° 0

-3
195° 10 345° 195° 10
-3 345°

210° 2·10
-3 330° 210° 330°
2·10-3

225° 315° 225° 315°

3·10-3
240° 300° 240° 3·10
-3
300°
255° 285° 255° 285°
270° 270°

FIG. 3. Polar diagram of scattering by a pair of equal dielectric FIG. 4. Polar diagram of the scattered field with the same
cylinders at a = b = 0.1 μm, εp = ε = 2.25, and R = 0.3 μm at parameters as in Fig. 3: third Born approximation (dotted line), BEM
incidence angle α = −π/4, wavelength λ = 1.5 μm, and distance (solid line), and DDA (dashed line). Inset: Magnifcation of the area
between the observation point and the center of the first cylinder of the main plot indicated by the box.
r = 2λ. First Born approximation (dotted line), second Born approx-
imation (dashed line), and BEM (solid line).
accuracy, we compare it with numerical solutions obtained by
known well-studied approaches, namely, the discrete dipole
√ approximation and boundary element methods.
or kp = εp ω/c, and condition (5) is The boundary equations are analogous to (6). When the
  perturbation is weak, expansion (7) yields
1 ∂Hin  ∂Hout  
Hin |γ = Hout |γ , = ,
(j ) 
ε ∂r γ ∂r γ Hp(j ) (r̃) = n Hout (r )∇Gp − εp Gp ∇Hout (r ) d l˜ ,
(j )
 
1 ∂Hp  ∂Hout  
γp

Hp |γp = Hout |γp , = . (9)
εp ∂ r̃ γp ∂ r̃ γp (j +1)
Hout (r) = n
G
∇Hp(j ) (r̃ ) − Hp(j ) (r̃ )∇G d l˜ . (10)
γp εp
We treat the second cylinder as the perturbation here. Let us
obtain a number of successive approximations for the whole Here Gp = Gp (r̃,r̃ ), r̃ = r − R, n = nγp . The recurrence
complicated configuration shown in Fig. 2. We exploit the relations are valid for arbitrary shape of the perturber with
Green’s function for cylindric geometry (A12). Using this a sharp boundary, provided its layout is in the external region
function makes it possible to account for the first cylinder of the main cylinder. The integral over the boundary of the
exactly including the multiple scattering. The second cylinder perturber can be calculated. The final relation is a Fourier
is described in terms of the Born approximation. To find series in azimuthal angle with a shift due to the axis offset.
the number of terms sufficient to get the field with a given The coefficients of series are:

 ∞

π kp b √ 
 
Dpm(0) = [Jm (k0 b)Hm (kp b) − εp Jm (k0 b)Hm (kp b)] i e m ik0 R cos α
+ n i(n−m)α
i e Cn Hn−m (k0 R) ,
2i n=−∞

π kp b √
D̃out [Hm (k0 b)Hm (kp b) − εp Hm (k0 b)Hm (kp b)]
m(j )
Dpm(j ) =
2i
∞

 √
Hn−m (k0 R)[Jm (k0 b)Hm (kp b) − εp Jm (k0 b)Hm (kp b)] ,
n(j ) i(n−m)α
+ Dout e
n=−∞

∞ 
m(j +1) π k0 b  1 
Dout =− Cm n(j ) i(n−m)α
Dp e Hm−n (k0 R) Jn (kp b)Jn (k0 b) − √ Jn (kp b)Jn (k0 b) ,
2i n=−∞
εp

m(j +1) π k0 b m(j ) 1
D̃out =− Dp Jm (kp b)Jm (k0 b) − √ Jm (kp b)Jm (k0 b) .
2i εp

063839-3
BEREZA, NEMYKIN, PERMINOV, FRUMIN, AND SHAPIRO PHYSICAL REVIEW A 95, 063839 (2017)

Hereafter the upper index (1) of the Hankel function is omitted; The coefficient
π 1/2π is found from the delta-function normal-
the prime indicates the derivative of a cylindric function with ization −π δ(ϕ) dϕ = 1.
respect to its argument. Expanding the Green’s function in partial waves
Figure 3 shows the angular dependence of the scattered
∞

field square |Hsc /H0 |2 . As the figure demonstrates, the first 

approximation gives a rather correct qualitative description of G(r,ϕ; r  ,ϕ  ) = gm (r,r  )eim(ϕ−ϕ ) (A3)
m=−∞
the diagram, with a deviation of 15%. The error of the second
order is nearly 3%. Figure 4 shows the comparison of the third and substituting into (2), we get an ordinary equation for each
Born approximation with calculations by the BEM and DDA. m:
The deviation appears to be about 1%.
 
d 2 gm 1 dgm m2 1
+ + k − 2 gm =
2
δ(r − r  ). (A4)
IV. CONCLUSIONS dr 2 r dr r 2π r
The Green’s function for a dielectric cylinder is found in the
At r > r  or r < r  , the corresponding solutions can be
cases of p and s waves with source points inside and outside the
expressed through the combinations of Bessel and Hankel
cylinder. High-order Born approximations of two dielectrics
functions, k = k0 :
with sharp boundaries are reduced to recurrence relations. This
technique is analytically applied to the scattering by a pair of ⎧
⎪
⎨am Jm (k1 r), r < a,
cylinders. The first approximation demonstrates its qualitative
agreement in shape with numerical results. The second and gm = Am Jm (kr) + Bm Hm (kr), a < r < r  , (A5)
⎪
⎩d H (kr), 
third approximations are shown to agree quantitatively with m m r < r.
calculations by boundary elements and discrete dipoles.
Conditions (5) are the continuity of the magnetic field
ACKNOWLEDGMENTS and its weighted normal derivative at the interface between
dielectric and free space, r = a,
The authors are grateful to O. V. Belai for helpful discus- 
sions. This work was supported by the Russian Foundation for 1 dgm
Basic Research (16-02-00511) and the Council for grants of [gm ]r=a = 0, = 0, (A6)
ε dr r=a
President of Russian Federation (NSh-6898.2016.2).
where the brackets denote a jump of the corresponding value.
APPENDIX A: SCALAR GREEN’S FUNCTION The conditions are written for a p wave, where the magnetic
field is parallel to the z axis. The next pair of conditions follows
Let us consider a cylinder whose axis is along the z direc- from the continuity of the Green’s function and the jump of its
tion, as shown in Fig. 5. We are looking for the scalar Green’s first derivative at r = r  :
function G(r,r ), which is the solution to the inhomogeneous 
two-dimensional Helmholtz equation (2), with k = k0 in free dgm 1
space and k = k1 in the dielectric. [gm ]r=r  = 0, = . (A7)
dr r=r  2π r 
Rewrite the delta function (2), in polar coordinates,
1 We omit the Hankel function in the first line and Bessel
δ(r − r  ) = δ(r − r  )δ(ϕ − ϕ  ), (A1) function in the third line in Eq. (A5) on the basis of regularity at
r r → 0 and the Sommerfeld radiation requirement at r → ∞.
where r, ϕ and r  ,ϕ  are the polar coordinates of the source Substituting (A5) into boundary conditions (A6) and (A7)
and observation points and decompose the angular factor into we get the set for the coefficients:
the Fourier series:
∞
1  im(ϕ−ϕ  ) Am Jm (ka) + Bm Hm (ka) = am Jm (k1 a), (A8)
δ(ϕ − ϕ  ) = e . (A2)
2π m=−∞
1
Am Jm (ka) + Bm Hm (ka) = √ am Jm (k1 a),
ε
y Am Jm (kr  ) + Bm Hm (kr  ) = dm Hm (kr  ),
1
Am Jm (kr  ) + Bm Hm (kr  ) = dm Hm (kr  ) − . (A9)
E0 2π kr 
(x, y)
From (A8) we get Bm = αm Am , where
x √
k Jm (ka)Jm (k1 a) − εJm (ka)Jm (k1 a)
H0 αm = √  . (A10)
εHm (ka)Jm (k1 a) − Hm (ka)Jm (k1 a)

FIG. 5. Right: Cross section of an infinite cylinder in homo- Then the determinant of the set (A9), for coefficients Am and
geneous space. The green circle indicates the observation point dm is {Jm + αm Hm ,Hm } = 2i/π kr  , where the curly bracket
r = (x,y). Left: p polarization. stands for the Wronskian {f,g} = f g  − f  g at r = r  .

063839-4
LIGHT SCATTERING BY DIELECTRIC BODIES IN THE . . . PHYSICAL REVIEW A 95, 063839 (2017)

Let us summarize the formulas for the partial Green’s In Eq. (B1) we implicitly isolate the term that includes the
function. At r  < a they are singularity of the Green’s tensor at r = r by means of a small
⎧1 volume V0 , for which the point r is internal. Then we rewrite
⎪ J (k r )Z1 (k1 r  ), 0 < r < r ,
⎨ 4i m 1 this term, introducing the following quantities:
gm = 4i1 Jm (k1 r  )Z1 (k1 r), r < r < a,  
⎪
⎩ 
M(V0 ,r) = d r Gαβ (r − r )
2 
βm Hm (k0 r)Jm (k1 r  ), a < r, V0

Hm (k1 a)Hm (k0 a)
−ε ν
Hm (k1 a)Hm (k0 a) 4Rα Rβ − 2δαβ R 2
Cm = − , (A11) − χ (r )E(r ) (B5)
 R4
where Z1 (z) = Cm Jm (z) + Hm (z), βm = εν /2π k1 a. and
At r  > a the formulas are 
⎧  4Rα Rβ − 2δαβ R 2
β J (k r )Hm (k0 r  ), r < a, L(V0 ,r) = − d 2r  χ (r )E(r ). (B6)
⎪
⎨ m m 1 V0 R4
gm = 4i1 Hm (k0 r  )Z2 (k0 r), a < r < r ,  is free from the singularity, thus M  → 0 with
⎪
⎩1 Note that M
H (k r )Z2 (k0 r  ), a < r  < r,
4i m 0 V0 → 0. The fraction under integration is, basically, the static
Jm (k1 a)Jm (k0 a) − εν Jm (k1 a)Jm (k0 a) limit (at k → 0) of the Green’s tensor. Also, we need to
Cm = − , (A12) discretize the whole scattering volume V into parts Vj (in

such a way that V0 coincides with one of them). With the use
 = Jm (k1 a)Hm (k0 a) − εν Jm (k1 a)Hm (k0 a), of (B5) and (B6), Eq. (B1) becomes
where Z2 (z) = Jm (z) + Cm Hm (z), βm = 1/2π k0 a. The for- 
E(ri ) = Einc (ri ) +  i − r )χ (r )E(r )]
d 2 r  [G(r
mulas with ν = −1/2,1/2 refer to the case of p and s waves, Vj
j =i
respectively.
 i ,ri ) − 
+ M(V L(Vi ,ri ), (B7)
APPENDIX B: DDA
where ri denotes a point lying inside the volume Vi .
Below we briefly recall the two-dimensional DDA approach Up to this line, the equations are fully correct as exact
[27] to obtain the particular relationships we used in our consequences of the initial wave equation. Now we make two
calculations. Let us have some scattering body, with volume approximations: the first is that E(r ) and χ (r ) are constant
V (which is per unit length along the z direction in the within the volume Vj ; the second approximation assumes that
two-dimensional case) and permittivity ε (which is constant 
1  i ,r ) = G(r
 i ,rj ).
within the body), placed in vacuum. From the Helmholtz d 2 r  G(r (B8)
Vj Vj
equation we obtain the integral equation for an isotropic
medium, Condition (B8) is intrinsically contained in all DDA formula-
 tions [28], which initially deal with replacing the scatterer with
E(r) = Einc (r) + 
d 2 r  [G(r,r
)χ (r )E(r )] a set of point dipoles. If the volumes Vi are square cells (we
V \V0 should keep in mind that we are treating the two-dimensional

 case), then we can place the points ri in the center of the
+ d 2 r  [G(r,r
)χ (r )E(r )], (B1)
V0
corresponding squares.
Below, we neglect M,  as most authors do, thus choosing
where V0 is a small volume around the singularity point R = the simpler (or “weak”) DDA formulation [28,29]. Integrating
r − r → 0, V \V0 is the volume of the dielectric without the (B6) we transform (B7) into its final form,
singular part, Einc (r) is the given field of the incident wave, 
χ (r) ≡ (ε − 1)/4π is the polarizability, and the Green’s tensor αi−1 = Ei,inc +
di  i − rj )dj ,
G(r (B9)

G(r,r 
) is the solution to the Maxwell equations: j =i

rot rot G 2
 − k G = 4π k δ(r − r ).2 
(B2) where we denote, for simplicity, the dependence on ri (and rj )
by the corresponding subscript; di = Vi χi Ei is the polarization
The Green’s tensor obeying (B2) can be expressed [1,26] in of the volume Vi (basically, its dipole moment, as we took χi
terms of the scalar Green’s function g, which satisfies Eq. (2), and Ei to be constant within Vi ); and αi is the polarizability
 
Gαβ = 4π k 2 δαβ + ∇α ∇β g, (B3) tensor, defined as

where α and β are Cartesian indices. Then, it is well known a 2 ε − 1

αi = 
 IVi χi (1 + 2π χi )−1 ≡ I. (B10)
that the Green’s tensor actually depends on the difference R. 2 ε+1
Finally, in the two-dimensional case we have
 The last term is the known quasistatic dipole polarizability of
iπ k Rα Rβ a thin cylinder (two-dimensional dipole) with the cross section
Gαβ (R) = A(kR)δαβ − B(kR) 2 ,
R R π a 2 equal to Vi .
(B4) Thus, the calculations consisted in finding the dipole
A(x) = xH0 (x) − H1 (x), B(x) = xH0 (x) − 2H1 (x), moments di by solving (B9) with (B10) and (B4). From them,
where R = r − r , H0 (x), H1 (x) are Hankel functions of the all the quantities of interest can be obtained. In our case, we
first kind. calculate the scattered magnetic field.

063839-5
BEREZA, NEMYKIN, PERMINOV, FRUMIN, AND SHAPIRO
[1] L. Novotny and B. Hecht, Principles of Nano-Optics
(Cambridge University Press, Cambridge, UK, 2006).
[2] C. Girard, Rep. Prog. Phys. 68, 1883 (2005).
[3] S. Kawata and V. M. Shalaev (eds.), Nanophotonics with Surface
Plasmons (Elsevier, Oxford, UK, 2007).
[4] M. I. Stockman, Phys. Today 64, 39 (2011).
[5] S. Y. Chou and D. Wei, Opt. Express 21, A60 (2013).
[6] S. V. Perminov, V. P. Drachev, and S. G. Rautian, Opt. Lett. 33,
2998 (2008).
[7] D. Shapiro, D. Nies, O. Belai, M. Wurm, and V. Nesterov,
Opt. Express 24, 15972 (2016).
[8] K. A. Willets and R. P. Van Duyne, Annu. Rev. Phys. Chem. 58,
267 (2007).
[9] T. Yamaoki, H. Hamada, and O. Matoba, Appl. Opt. 55, 6874
(2016).
[10] L. Rayleigh, Philos. Mag. Ser. 6 36, 365 (1918).
[11] J. R. Wait, Can. J. Phys. 33, 189 (1955).
[12] P. E. Vorobev, Sov. Phys. JETP 110, 193 (2010).
[13] I. E. Mazets, Sov. Phys. Tech. Phys. 45, 1238 (2000).
[14] R. Borghi, F. Frezza, G. Schettini, F. Gori, and M. Santarsiero,
J. Opt. Soc. Am. A 13, 483 (1996).
[15] K. J. Schaudt, N.-H. Kwong, and J. D. Garcia, Phys. Rev. A 44,
4076 (1991).
[16] S. Belan and S. Vergeles, Opt. Mater. Express 5, 130 (2015).
063839-6
PHYSICAL REVIEW A 95, 063839 (2017)
[17] S.-C. Lee, J. Quant. Spectrosc. Radiat. Transfer 182, 119
(2016).
[18] G. Y. Panasyuk, V. A. Markel, P. Scott Carney, and J. C.
Schotland, Appl. Phys. Lett. 89, 221116 (2006).
[19] M. Salucci, G. Oliveri, A. Randazzo, M. Pastorino, and A.
Massa, J. Opt. Soc. Am. A 31, 1167 (2014).
[20] M. B. Doost, Phys. Rev. A 93, 023835 (2016).
[21] O. V. Belai, L. L. Frumin, S. V. Perminov, and D. A. Shapiro,
Opt. Lett. 36, 954 (2011).
[22] O. V. Belai, L. L. Frumin, S. V. Perminov, and D. A. Shapiro,
Europhys. Lett. 97, 10007 (2012).
[23] L. L. Frumin, A. V. Nemykin, S. V. Perminov, and D. A. Shapiro,
J. Opt. 15, 085002 (2013).
[24] F. W. J. Ovler, D. W. Lozier, R. F. Boisvert, and C. W.
Clark, NIST Handbook of Mathematical Functions (Cambridge
University Press, New York, 2010).
[25] J. D. Jackson, Classical Electrodynamics (Wiley, New York,
1999).
[26] A. M. Kern and O. J. F. Martin, J. Opt. Soc. Am. A 26, 732
(2009).
[27] O. J. F. Martin and N. B. Piller, Phys. Rev. E 58, 3909 (1998).
[28] M. Yurkin and A. Hoekstra, J. Quant. Spectrosc. Radiat. Transfer
106, 558 (2007).
[29] A. Lakhtakia, Int. J. Mod. Phys. C 03, 583 (1992).