Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Boundary element method for surface nonlinear optics of nanoparticles: erratum

Open Access Open Access

Abstract

We report a correction to the numerical procedure, in which the source vector lacked a factor 1/2 and the integration in Eq. (19) was incorrect. The errors are inconsequential for the main results.

© 2013 Optical Society of America

We wish to point out small errors in the mathematical details of our paper (Ref. [1]). The errors are related to the technical details and are inconsequential for the main results.

The operators 𝒦l involve an integration over a strong singularity, so that the integral exists by means of the Cauchy principal value:

𝒦lf(r)=Vl[Gl(r,r)]×f(r)dS=12f(r)×nl+𝒦lf(r),
where
𝒦lf(r)=lima0VlD(r,a)[Gl(r,r)]×f(r)dS
where D(r, a) is a disk of radius a around r and it is assumed that the boundary ∂Vl is smooth. Thus the proper way to write Eqs. (7) and (8) of [1] would be
(𝒟1J1S+𝒦1M1S𝒟2J2S𝒦2M2S)tan=12εSPnS,(𝒦1J1S+1η12𝒟1M1S+𝒦2J2S1η22𝒟2M2S)tan=i12ΩPS×n.
Although the Eqs. (7) and (8) are correct, the method of moments procedure should be applied to the operators involving the principal values. This was not noticed in the implementation of the numerical procedure. Consequently, the first two elements of the source vector b lacked a factor 1/2. This factor does not appear in the PMCHWT formulation for linear scattering due to the continuity of E tan and H tan.

Furthermore, there is a mistake in the source integral (19). For if PnS is constant in each triangle, the integral still vanishes for RWG functions. This may be overcome by first expanding SPnS in RWG basis as

SPnS=l=1Nplfl.
where pl can be obtained in the same way as bmn1 by integration by parts. Then the coefficients bmn2 can be computed directly as
bmn2=1εlplSmSlfmn×fldS,
where l runs over indices for which SmSl ≠ ∅. In [2], an alternative distributional approach / was presented for evaluating the coefficients bmn1 and bmn2.

Finally, we point out a curious observation, that these two mistakes very accurately cancelled each other out in the numerical results, as is apparent from the comparison to the multipole solution. In general this cancellation is not to be expected. Through use of the corrected expressions we verified the validity of the results in Fig. 5.

References and links

1. J. Mäkitalo, S. Suuriniemi, and M. Kauranen, “Boundary element method for surface nonlinear optics of nanoparticles,” Opt. Express 19, 23386–23399 (2011) [CrossRef]   [PubMed]  

2. C. Forestiere, A. Capretti, and G. Miano, “Surface Integral Method for the Second Harmonic Generation in Metal Nanoparticles,” arXiv:1301.1880 [physics.optics] (2013)

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Equations (5)

Equations on this page are rendered with MathJax. Learn more.

𝒦 l f ( r ) = V l [ G l ( r , r ) ] × f ( r ) d S = 1 2 f ( r ) × n l + 𝒦 l f ( r ) ,
𝒦 l f ( r ) = lim a 0 V l D ( r , a ) [ G l ( r , r ) ] × f ( r ) d S
( 𝒟 1 J 1 S + 𝒦 1 M 1 S 𝒟 2 J 2 S 𝒦 2 M 2 S ) tan = 1 2 ε S P n S , ( 𝒦 1 J 1 S + 1 η 1 2 𝒟 1 M 1 S + 𝒦 2 J 2 S 1 η 2 2 𝒟 2 M 2 S ) tan = i 1 2 Ω P S × n .
S P n S = l = 1 N p l f l .
b m n 2 = 1 ε l p l S m S l f m n × f l d S ,
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.