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: 847

Скачиваний: 0

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

3.6 Multiple Particles

235

Z

 

z1

L

x1

 

O1

−L

X

 

z2 O

L

x2

−L

O2

 

z1

 

 

r

2l

β1

x1

O1

y1

α 1

 

Fig. 3.55. Geometry of a system of four spheres

DSCS

100

 

 

TMATMULT - parallel

10−1

TMATMULT - perpendicular

TMATMULTSPH - parallel

 

TMATMULTSPH - perpendicular

10−2

10−3

10−4

10−5

10−6

60

120

180

240

300

360

0

Scattering Angle (deg)

Fig. 3.56. Normalized di erential scattering cross-sections of a system of four spheres


236 3 Simulation Results

<DSCS>

100

 

TMATMULT - parallel

 

TMATMULT - perpendicular

10−1

TMATMULTSPH - parallel

TMATMULTSPH - perpendicular

10−2

10−3

10−4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0

30

60

90

120

150

180

 

 

 

 

 

Scattering Angle (deg)

 

 

 

 

Fig. 3.57. Averaged di erential scattering cross-sections of a system of four spheres

N = kf

Rg

Df

(3.14)

 

 

,

 

rp

 

 

 

 

where the radius of gyration is determined from the position of each primary spherule ri to the geometrical center of the cluster r0,

N

Rg2 = N1 |ri − r0|2 i=1

,N

and r0 = (1/N ) i=1 ri. Equation (3.14) is important because the values of parameters are linked to real physical processes. If Df 1.8 the process belongs to a di usion limited aggregation and if Df 2.1 it belongs to a reaction limited aggregation. For simulating di usion limited aggregation, we developed a Fortran program and used the Cartesian coordinates of the spherule as input parameters for the TMULTSPH routine. This program is based on the fast algorithm described by Filippov et al. [66] and generates aggregates by using two di erent methods. In the first method, each new primary spherule will be stuck on the mother aggregate after touching one of the existing spherules. This method is known as the di usion limited algorithm (DLA) and is used for aggregates not larger than about N = 150 spherules. With the second method, complete small aggregates are stuck on the mother aggregate. The small aggregates are generated with the DLA algorithm and consists of about 20 or more primary spherules. This method is called di usion limited cluster cluster aggregation (DLCCA) and gives more realistic aggregates. An example of a monodisperse aggregate representing


3.6 Multiple Particles

237

Fig. 3.58. Monodisperse aggregate with Df = 1.8, N = 130 and rp = 10 nm

Fig. 3.59. Polydisperse aggregate with Df = 1.8, N = 203 and rp between 10 and 20 nm

a soot particle from a combustion processes is shown in Fig. 3.58, while a polydisperse aggregate is shown in Fig. 3.59. For the monodisperse aggregate, the scattering characteristics are shown in Fig. 3.60, and the essential parameters controlling the convergence process are Nrank = 8 and Mrank = 6 for the system T -matrix, and Nrank = 4 and Mrank = 3 for the primary spherules. The di erential scattering cross-sections of the polydisperse aggregate are plotted in Fig. 3.61, and the parameters of calculation for the system T -matrix increase to Nrank = 12 and Mrank = 10. The program described above has been used by Riefler et al. [204] to characterize soot from a flame by analyzing the measured scattering patterns.

238 3 Simulation Results

10−1

 

 

 

 

 

 

 

10−2

 

 

 

 

 

 

 

10−3

 

parallel

 

 

 

 

 

 

perpendicular

 

 

 

 

DSCS

 

 

 

 

 

 

 

10−4

 

 

 

 

 

 

 

10−5

 

 

 

 

 

 

 

10−6

0

30

60

90

120

150

180

 

 

 

Scattering Angle (deg)

 

 

Fig. 3.60. Normalized di erential scattering cross-sections of a monodisperse aggregate

 

10−3

 

 

 

 

 

 

 

 

 

 

 

 

parallel

 

 

 

 

10−4

 

 

 

perpendicular

 

 

 

10−5

 

 

 

 

 

 

 

DSCS

10−6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10−7

 

 

 

 

 

 

 

 

10−8

 

 

 

 

 

 

 

 

10−9

0

60

120

180

240

300

360

 

 

 

 

Scattering Angle (deg)

 

 

Fig. 3.61. Normalized di erential scattering cross-sections of a polydisperse aggregate

3.7 Composite Particles

Electromagnetic scattering by axisymmetric, composite particles can be computed with the TCOMP routine. An axisymmetric, composite particle consists of several nonenclosing, rotationally symmetric regions with a common axis of symmetry. In contrast to the TMULT routine, TCOMP is based on a formalism which avoid the use of any local origin translation. The scattering



3.7 Composite Particles

239

z

r

 

 

l

 

L

z1

 

 

x

z3

 

Fig. 3.62. Geometry of a composite particle consisting of three identical cylinders

characteristics can be computed with localized or distributed sources, and the program supports calculations for dielectric particles. Particle geometries included in the library are half-spheroids with o set origins and three cylinders.

In Fig. 3.62, we show a composite particle consisting of three identical cylinders of radius ksr = 2 and length ksl = 4. The axial positions of the first and third cylinder are given by ksz1 = 4 and ksz3 = 4, and the relative refractive indices are chosen as mr1 = mr3 = 1.5 and mr2 = 1.0. Because mr2 = 1.0, the scattering problem is equivalent to the multiple scattering problem of two identical cylinders. On the other hand, the composite particle can be regarded as an inhomogeneous cylinder with a cylindrical inclusion. Therefore, the scattering characteristics are computed with three independent routines: TCOMP, TMULT, and TLAY. Localized and distributed sources are used for T -matrix calculations with the TCOMP routine, while distributed sources are used for calculations with the TLAY routine. Figure 3.63 shows the di erential scattering cross-sections for a fixed orientation of the composite particle (αp = βp = 45), while Fig. 3.64 illustrates the numerical results for a random orientation. The behavior of the far-field patterns are quite similar.

In order to demonstrate the capability of the code to compute the electromagnetic scattering by composite particles with elongated and flattened regions, we consider the geometry depicted in Fig. 3.65. For this application, the sources are distributed on the real and imaginary axes. The parameters specifying the particle shape are: ksr = 1, ksR = 3, ksl = 6 and ksL = 7, while the relative refractive indices are identical mr1 = mr2 = mr3 = 1.5. For the prolate cylinders, Nrank = 10 sources are distributed along the axis of symmetry, and for the oblate cylinder, Nrank = 10 sources are distributed