Non-cubical dipoles

[GoogleCodeExporter]

There is some evidence that using dipoles with non-equal sides (rectangular 
parallelepipeds) can be beneficial (smaller number of dipoles to reach the same 
accuracy) in some cases, e.g. for very oblate/prolate scatterers. The latest 
papers about this are:
1. E. Massa et al., “Discrete-dipole approximation on a rectangular cuboidal 
point lattice: considering dynamic depolarization,” J. Opt. Soc. Am. A 31(1), 
135–140 (2014) [doi:10.1364/JOSAA.31.000135].
2. Y. O. Agha et al., “Near-field properties of plasmonic nanostructures with 
high aspect ratio,” Prog. Electromag. Res. 146, 77–88 (2014). 
[http://www.jpier.org/PIER/pier.php?paper=14012904]

So implementing such functionality is a good ides, although it may be hard to 
integrate it with all existing features of ADDA. Also, choosing an optimal 
aspect ratio of dipoles for a particular problem can be non-trivial task by 
itself.

Original issue reported on code.google.com by yurkin on 20 May 2014 at 3:49

[GoogleCodeExporter]

Original comment by yurkin on 22 Jul 2014 at 6:07

[GoogleCodeExporter]

Other relevant papers are:
- D. Gutkowicz-Krusin and B.T. Draine, “Propagation of electromagnetic waves 
on a rectangular lattice of polarizable points,” 2004, 
<http://arxiv.org/abs/astro-ph/0403082>.
Provides an LDR-like derivation for point dipoles on a rectangular lattice.

- P.C. Chaumet, A. Sentenac, and A. Rahmani, “Coupled dipole method for 
scatterers with large permittivity,” Phys. Rev. E 70, 036606 (2004) 
[doi:10.1103/PhysRevE.70.036606].
Mentions the possibility of extending IGT to rectangular dipoles. Probably the 
corresponding Fortran code, currently used in ADDA, can be modified to compute 
the integral (interaction term) over the non-cubical dipoles.

Original comment by yurkin on 3 Aug 2014 at 6:18