SWF
SWF
SWF
243246 in B. Gustafson, L. Kolokolova, and G. Videen (eds) Electromagnetic and Light Scattering by Nonspherical Particles Army Research Laboratory, Adelphi, Maryland (2002) Presented at: 6th Conference on Electromagnetic and Light Scattering by Nonspherical Particles: Theory, Measurements, and Applications Gainesville, Florida, USA (2002)
Abstract
Vector spherical wavefunction expansions of an incident radiation eld are essential for theoretical light scattering methods such as the T-matrix method and generalized Lorenz-Mie theory (GLMT). In general, it is theoretically straightforward to nd a vector spherical wavefunction representation of an arbitrary radiation eld. For example, a simple formula results in the useful case of an incident plane wave. Laser beams present some difculties. These problems are not a result of any deciency in the basic process of spherical wavefunction expansion, but are due to the fact that laser beams, in their standard representations, are not radiation elds, but approximations of radiation elds. We show that both integral and point-matching methods can be used to nd vector spherical wavefunction expansions of laser beams, including strongly focussed beams.
1 Introduction
Representation of an incident radiation eld in terms of vector spherical wavefunctions (VSWFs) is required for theoretical scattering methods such as the T-matrix method [1, 2, 3] and generalized Lorenz-Mie theory [4]. Any monochromatic radiation eld with all sources and sinks at r = can be expressed as a sum of regular vector spherical wavefunctions [3]: Einc (r) = where RgNmn (kr) =
n=1 m=n
(1)
Cmn ( , ) =
m m Yn ( , ) Yn ( , ), sin 1
m Bmn ( ) = r Yn ( , ) =
(5) (6)
m jn (kr) are spherical Bessel functions, Yn ( , ) are normalised scalar spherical harmonics and Nn = 1/ n(n + 1) are normalisation constants. If the wavenumber of the eld and the expansion coefcients a mn and bmn (also called beam shape coefcients) are known, the eld can be found at any point. It is theoretically straightforward, though possibly numerically and computationally difcult, to nd the VSWF expansion of an arbitrary radiation eld [5, 6]. From the general formula for an orthogonal eigenfunction transform (a.k.a. generalised Fourier transform), we obtain
amn =
(7)
Greens theorem can be usefully employed to convert the volume integral to a surface integral. In the case of an incident plane wave E(r) = E 0 exp(ik r), with k in the direction ( , ), this gives a simple formula for the expansion coefcients [2, 3]: amn = 4 i n Nn Cmn E0 exp(im), bmn = 4 i n1 Nn Bmn E0 exp(im). (8)
Note that the amplitude vector E 0 contains the information regarding the polarisation and phase of the wave, and can be complex. While the case of plane wave illumination is useful for a wide range of scattering problems, many applications of scattering involve laser beams. In particular, laser trapping requires strongly focussed beams. As scattering calculations allow the optical forces and torques with the optical trap to be determined [7, 8], and the T-matrix method is particularly useful for the repeated calculations typically required, spherical harmonic expansion of laser beams, including strongly focussed beams, is extremely useful. Unfortunately, the VSWF representation of laser beams presents some difculties. These problems are not a result of any deciency in the basic process of expansion of radiation elds in terms of VSWFs, but are due to the fact that standard representations of laser beams are not radiation elds. Since the individual spherical wavefunctions satisfy the vector Helmholtz equation
E k2 E = 0,
(9)
only solutions of this equation have VSWF expansions. Standard laser beam representations do not satisfy this equation, but instead satises the paraxial wave equation. Some method must be used to approximate the standard laser beam with a real radiation eld. We can note that there exists a signicant and useful body of work on spherical harmonic representation of beams [9, 10, 11, 12, 13, 14, 15]. However, while satisfactorily efcient and accurate methods exist for weakly focussed beams, when the deviations from paraxiality are small, strongly focussed beams, as required for trapping, remain problematic.
2.1
Choice of surface
Fortunately, there are two suitable natural choices of the surface at which the elds can be matched: the far eld and the focal plane (gure 1).
focal plane
Figure 1: Natural choices of surface for eld matching: focal plane and far eld. Perhaps the most natural choice is the focal plane; those who work with strongly focussed beams tend to be familiar with the structure of beams near the focal plane. This also has the convenience being similar to paraxial beams in the region where trapping will occur, allowing easy comparison of results. If we only consider the amplitude and phase of the electric eld in a single plane, we can calculate a partial set of expansion coefcients from this incomplete information. Then, using the fact that the direction of propagation of the beam is known, the remaining expansion coefcients can be found. The far eld of the beam can also be used. Since the incoming eld can be readily specied, the amplitude and phase at a particular (large) r, along with the direction of propagation, can again be used to nd the expansion coefcients. Since, in practice, the beam will pass through nite apertures (especially the nal focussing lens), use of only the incoming eld allows the beam to be easily truncated, as long as the aperture is in the far eld of the beam. Focal plane matched and far eld matched beams will be compared, and the differences between them and paraxial beams noted.
2.2
The VSWF transform (equation 7) can be used to calculate expansion coefcients one at a time. Checks for convergence can be readily implemented, as coefcients can be calculated until they are no longer signicant. This is not necessarily the fastest method. Since the required number of wavefunctions to include in the expansion can be estimated beforehand, especially for focal plane matching, point-matching methods can prove useful. Since the sum of the nite number of spherical wavefunctions is equal to the incident beam on the matching surface, the choice of suitable discrete points at which they can be compared yields a system of linear equations. This linear system can be solved. Since, in general, it will not be possible to match the elds exactly with only a nite number of wavefunctions in the expansion, it can be useful to generate an overdetermined system of equations, and thus obtain a least-squares match between the two elds. Other methods are also possible, such as Fourier expansion of the focal plane eld, followed by using the plane wave expansion formula (equation 8). These methods will be compared for accuracy and computational efciency.
3 Conclusion
A reasonable solution to the problem of vector spherical wavefunction representation of strongly focussed laser beams is presented. Although the VSWF expansions are not identical to the parax3
ial beams from which they are derived, they are related in a natural manner. These beams are of particular interest for optical trapping and scattering by single particles within optical traps.
References
[1] P. C. Waterman, Symmetry, unitarity and geometry in electromagnetic scattering,Physical Review D 3, 825839 (1971). [2] L. Tsang, J. A. Kong and R. T. Shin, Theory of microwave remote sensing, (John Wiley, New York, 1983). [3] M. I. Mishchenko, Light scattering by randomly oriented axially symmetric particles, Journal of the Optical Society of America A 8, 871882 (1991). [4] J. A. Lock, Improved Gaussian beam-scattering algorithm, Applied Optics 34, 559570 (1995). [5] C. F. Bohren and D. R. Huffman, Absorption and scattering of light by small particles (John Wiley, New York, 1983). [6] K. F. Ren, G. Gouesbet and G. Gr han, Integral localized approximation in generalized Lorenze Mie theory, Applied Optics 35, 42184225 (1998). [7] T. A. Nieminen, H. Rubinsztein-Dunlop and N. R. Heckenberg, Calculation and optical measurement of laser trapping forces on non-spherical particles, Journal of Quantitative Spectroscopy and Radiative Transfer 70, 627637 (2001). [8] H. Polaert, G. Gr han and G. Gouesbet, Forces and torques exerted on a multilayered spherical e particle by a focused Gaussian beam, Optics Communications 155, 169179 (1998). [9] A. Doicu and T. Wriedt, Computation of the beam-shape coefcients in the generalized LorenzMie theory by using the translational addition theorem for spherical vector wave functions, Applied Optics 36, 29712978 (1997). [10] G. Gouesbet, J. A. Lock and G. Gr han, Partial-wave representations of laser beams for use in e light-scattering calculations Applied Optics 34, 21332143 (1995). [11] G. Gouesbet, Partial-wave expansions and properties of axisymmetric light beams, Applied Optics 35, 15431555 (1996). [12] J. A. Lock and J. T. Hodges Far-eld scattering of an axisymmetric laser beam of arbitrary prole by an on-axis spherical particle, Applied Optics 35, 42834290 (1996). [13] J. A. Lock and J. T. Hodges, Far-eld scattering of a non-Gaussian off-axis axisymmetric laser beam by a spherical particle, Applied Optics 35, 66056616 (1996). [14] H. Polaert, G. Gr han and G. Gouesbet, Improved standard beams with application to reverse e radiation pressure, Applied Optics 37, 24352440 (1998). [15] Y. Han and Z. Wu, Scattering of a spheroidal particle illuminated by a Gaussian beam, Applied Optics 40, 25012509 (2001). [16] L. W. Davis, Theory of electromagnetic beams, Physical Review A 19, 11771179 (1979).