Файл: Doicu A., Wriedt T., Eremin Y.A. Light scattering by systems of particles (OS 124, Springer, 2006.pdf

ВУЗ: Не указан

Категория: Не указан

Дисциплина: Не указана

Добавлен: 28.06.2024

Просмотров: 818

Скачиваний: 0

ВНИМАНИЕ! Если данный файл нарушает Ваши авторские права, то обязательно сообщите нам.

292 C Computational Aspects in E ective Medium Theory

Performing the integration and using the relation

 

 

 

 

 

 

 

1

 

 

 

 

ejksz

 

 

 

 

 

 

 

 

 

 

1

 

 

 

 

 

ejksz ,

Pn

 

 

 

 

= (

1)n

 

2n + 1

 

 

 

jks ∂z

 

ks2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ks2

2

 

 

 

we obtain

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2π

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1,z

 

 

 

 

 

 

 

 

 

 

 

2n + 1

 

 

 

 

Imm n =

 

 

 

 

 

jn +1

 

 

 

 

 

 

 

 

ejksz0l

δmm .

k2

(K

s

k )

 

2

 

 

 

 

 

 

 

 

 

 

 

s

 

 

 

s

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The final result is

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Imm1 n = Imm1,R n + Imm1,z n

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16πR3

 

 

 

 

 

 

 

2n + 1

 

 

 

 

 

=

 

 

 

 

 

 

 

 

 

jn

 

 

 

 

 

 

 

 

 

 

 

ejKsz0l δmm Fn (Ks, ks, R)

2

 

 

 

 

2

 

 

 

 

 

2

 

 

 

 

 

(ksR) (KsR)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2π

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

+

 

 

 

 

 

 

jn +1

 

2n + 1

ejksz0l δmm .

 

 

(C.2)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2

 

 

 

 

 

 

 

ks2 (Ks − ks)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

C.2 Computation of the Integral Imm2 n

The integral Imm2 n is

Imm2 n = [g (rlp) 1] ejKe·r0p u3m −mn (ksrlp) dV (r0p) ,

Dp

where Ke = Ksez , and the domain of integration contains the half-space z0p 0 less a sphere of radius 2R centered at r0l. For a spherical symmetric pair distribution function, g(rlp) = g(rlp), the function g(rlp)1 tends to zero for a few 2R-values. Therefore, if the point r0l is at least several diameters deep in the scattering medium, the volume of integration can be extended to infinity (Fig. C.2). Taking into account that r0p = r0l + rlp, and replacing the variables rlp, θlp and ϕlp by r, θ and ϕ, respectively, we obtain

Imm2 n = ejKe·r0l 2π π ∞ [g(r) 1] ejKe·r h(1)n (ksr) Pn|m −m| (cos θ)

00 2R

×ej(m −m)ϕr2 sin θ dr dθ dϕ .

Using the plane wave expansion (C.1), the orthogonality relations of the associated Legendre functions and the equation Ke · r0l = Ksz0l, we derive

 

n

 

 

 

 

 

 

 

2

2n + 1

e

jKsz0l

 

(1)

2

 

Imm n = 4πj

 

 

 

δmm

2R [g(r) 1] jn (Ksr) hn

(ksr) r

dr.

 

2

 

Changing the variables from r to x = r/(2R), we find that


C.3 Computation of the Terms S11nn and S12nn

293

Op

 

2R

S

 

θlp

 

 

rlp

SR

 

 

Ol

 

R

Fig. C.2. Illustration of the domain of integration bounded by the spherical surface SR and a sphere situated in the far-field region

 

 

2n + 1

 

 

 

Imm2 n = 32πR3jn

 

 

ejKsz0l δmm Gn (Ks, ks, R) ,

(C.3)

2

where

Gn (Ks, ks, R) = [g (2Rx) 1] h(1)n (2ksRx) jn (2KsRx) x2dx.

1

C.3 Computation of the Terms S11nn and S12nn

The terms S11nn

and S12nn are given by

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2n

 

+ 1

 

 

 

 

 

 

1

 

 

n

 

 

 

 

1

 

 

 

 

 

S1nn

=

 

(1)

 

 

 

2

 

I1nn n ,

 

 

 

 

n =0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2n

 

+ 1

 

 

 

 

 

 

2

 

 

n

 

 

 

 

2

 

 

 

 

 

S1nn

=

 

(1)

 

 

 

 

 

I1nn n ,

 

 

 

 

 

 

 

2

 

 

 

 

 

n =0

 

 

 

 

 

 

 

 

 

 

 

where

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I1

=

π π1

(β)π1

(β) + τ 1(β)τ 1

(β) Pn

 

(cos β) sin β dβ,

 

1nn n

 

n

n

 

 

n

 

n

 

 

 

 

 

 

 

0

 

 

 

 

 

 

 

 

 

 

 

 

 

I12nn n

= π πn1 (β)τn1 (β) + τn1(β)πn1 (β) Pn

(cos β) sin β dβ.

 

 

 

0

 

 

 

 

 

 

 

 

 

 

 

 

 

To compute S11nn , we consider the function

 

 

 

 

 

 

 

 

 

fnn (β) = πn1 (β)πn1 (β) + τn1(β)τn1 (β)

(C.4)


294 C Computational Aspects in E ective Medium Theory

for fixed values of the indices n and n . This function can be expanded in terms of Legendre polynomials

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fnn (β) =

 

ann ,n Pn (cos β) ,

 

 

 

 

 

(C.5)

 

 

 

 

n =0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

where the expansion coe cients are given by

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ann ,n = π fnn (β)Pn

(cos β) sin β dβ = I11nn n .

0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Using the special value of the Legendre polynomial at β = π,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pn (cos π) = (

 

 

 

1)n

 

2n + 1

,

 

 

 

 

 

 

 

 

 

 

 

 

 

2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we see that (C.5) leads to

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2n + 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

( 1)

n

 

 

 

 

 

 

 

1

 

n

 

 

=

 

 

1

 

.

 

fnn (π) =

 

 

 

 

 

 

 

 

 

 

 

I1nn

 

 

 

 

1nn

 

 

 

n =0

 

 

 

 

 

 

 

 

 

2

 

 

 

 

 

 

 

 

 

S

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

On the other hand, since

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

πn1 (π) =

(1)n−1

 

 

 

 

 

,

 

 

 

 

n(n + 1) (2n + 1)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

22

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

τ 1(π) =

(1)n

 

 

 

 

.

 

 

 

 

 

 

n(n + 1) (2n + 1)

 

 

 

 

 

 

 

n

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2

2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(C.4) implies that

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fnn (π) = snn =

(1)n+n

 

 

 

 

 

(C.6)

n(n + 1) (2n + 1)

n (n

+ 1) (2n + 1)

4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

and therefore,

S11nn = snn .

Analogously, we find that

S12nn = −snn .


D

Completeness

of Vector Spherical Wave Functions

The completeness and linear independence of the systems of radiating and regular vector spherical wave functions on closed surfaces have been established by Doicu et al. [49]. In this appendix we prove the completeness and linear independence of the regular and radiating vector spherical wave functions on two enclosing surfaces. A similar result has been given by Aydin and Hizal [5] for both the regular and radiating vector spherical wave functions with nonzero divergence, i.e., the solutions to the vector Helmholtz equation.

Before formulating the boundary-value problem for the Maxwell equations we introduce some normed spaces which are relevant in electromagnetic scattering theory. With S being the boundary of a domain Di we denote by [39, 40]

Ctan(S) = {a/a C(S), n · a = 0}

the space of all continuous tangential fields and by

Ctan0(S) = a/a C0(S), n · a = 0 , 0 < α ≤ 1 ,

the space of all uniformly H¨older continuous tangential fields equipped with the supremum norm and the H¨older norm, respectively. The space of uniformly H¨older continuous tangential fields with uniformly H¨older continuous surface divergence ( s) is defined as

0

 

0

(S), s · a C

0

(S)

 

Ctan,d

(S) =

a/a Ctan

 

, 0 < α ≤ 1 ,

while the space of square integrable tangential fields is introduced as

L2tan(S) = a/a L2(S), n · a = 0 .

Now, let S1 and S2 be two closed surfaces of class C2, with S1 enclosing S2 (Fig. D.1). Let the doubly connected domain between S1 and S2 be denoted by Di,1, the domain interior to S2 by Di,2, the domain interior to S1 by Di,


296 D Completeness of Vector Spherical Wave Functions

n1

n2

εi,µi O

Di,2

S2

S1

Di,1

 

Ds

Fig. D.1. The closed surfaces S1 and S2

and the domain exterior to S1 by Ds. We assume that the origin is in Di,2, and denote by n1 and n2 the outward normal unit vectorsto S1 and S2, respectively. The wave number in the domain Di,1 is ki = k0 εiµi, where k0 is the wave number in free space, and εi and µi are the relative permittivity and permeability of the domain Di,1. We introduce the product spaces

Ctan(S1,2) = Ctan(S1) × Ctan(S2),

C0tan(S1,2) = Ctan0(S1) × Ctan0(S2) ,

C0tan,d(S1,2) = Ctan0,d(S1) × Ctan0,d(S2) ,

L2tan(S1,2) = L2tan(S1) × L2tan(S2),

and define the scalar product in L2tan(S1,2) by

- x1

 

y1

.

x2

,

y2

= x1, y1 2,S1 + x2, y2 2,S2 .

 

 

 

2,S1,2

The Maxwell boundary-value problem for the doubly connected domain Di,1 has the following formulation.

Find a solution Ei, Hi C1(Di,1) ∩ C(Di,1) to the Maxwell equations in

Di,1

× Ei = jk0µiHi , × Hi = jk0εiEi,

satisfying the boundary conditions

n1 × Ei = f 1

on

S1 ,

n2 × Ei = f 2

on

S2 ,

where f 1 and f 2 are given tangential vector fields.

Denoting by σ(Di,1) the spectrum of eigenvalues of the Maxwell boundaryvalue problem, we assume that for f 1 Ctan0,d(S1) and f 2 Ctan0,d(S2), both