1 #LyX 2.4 created this file. For more info see https://www.lyx.org/
5 \save_transient_properties true
8 \use_default_options true
9 \maintain_unincluded_children false
11 \language_package default
14 \font_roman "default" "TeX Gyre Pagella"
15 \font_sans "default" "default"
16 \font_typewriter "default" "default"
17 \font_math "auto" "auto"
18 \font_default_family default
19 \use_non_tex_fonts false
23 \font_typewriter_osf false
24 \font_sf_scale 100 100
25 \font_tt_scale 100 100
27 \use_dash_ligatures true
29 \default_output_format default
31 \bibtex_command default
32 \index_command default
33 \float_placement class
34 \float_alignment class
35 \paperfontsize default
38 \pdf_author "Marek Nečada"
40 \pdf_bookmarksnumbered false
41 \pdf_bookmarksopen false
42 \pdf_bookmarksopenlevel 1
50 \use_package amsmath 1
51 \use_package amssymb 1
54 \use_package mathdots 1
55 \use_package mathtools 1
57 \use_package stackrel 1
58 \use_package stmaryrd 1
59 \use_package undertilde 1
61 \cite_engine_type default
65 \paperorientation portrait
77 \paragraph_separation indent
78 \paragraph_indentation default
80 \math_numbering_side default
85 \paperpagestyle default
87 \tracking_changes false
100 \begin_layout Subsection
104 \begin_layout Standard
105 The basic idea of MSTMM is quite simple: the driving electromagnetic field
106 incident onto a scatterer is expanded into a vector spherical wavefunction
107 (VSWF) basis in which the single scattering problem is solved, and the
108 scattered field is then re-expanded into VSWFs centered at the other scatterers.
109 Repeating the same procedure with all (pairs of) scatterers yields a set
110 of linear equations, solution of which gives the coefficients of the scattered
111 field in the VSWF bases.
112 Once these coefficients have been found, one can evaluate various quantities
113 related to the scattering (such as cross sections or the scattered fields)
118 \begin_layout Standard
119 The expressions appearing in the re-expansions are fairly complicated, and
120 the implementation of MSTMM is extremely error-prone also due to the various
121 conventions used in the literature.
122 Therefore although we do not re-derive from scratch the expressions that
123 can be found elsewhere in literature, we always state them explicitly in
127 \begin_layout Subsection
128 Single-particle scattering
131 \begin_layout Standard
132 In order to define the basic concepts, let us first consider the case of
133 EM radiation scattered by a single particle.
134 We assume that the scatterer lies inside a closed sphere
135 \begin_inset Formula $\particle$
138 , the space outside this volume
139 \begin_inset Formula $\medium$
142 is filled with an homogeneous isotropic medium with relative electric permittiv
144 \begin_inset Formula $\epsilon(\vect r,\omega)=\epsbg(\omega)$
147 and magnetic permeability
148 \begin_inset Formula $\mu(\vect r,\omega)=\mubg(\omega)$
151 , and that the whole system is linear, i.e.
152 the material properties of neither the medium nor the scatterer depend
153 on field intensities.
154 Under these assumptions, the EM fields
155 \begin_inset Formula $\vect{\Psi}=\vect E,\vect H$
159 \begin_inset Formula $\medium$
162 must satisfy the homogeneous vector Helmholtz equation together with the
163 transversality condition
166 \left(\nabla^{2}+k^{2}\right)\Psi\left(\vect r,\vect{\omega}\right)=0,\quad\nabla\cdot\vect{\Psi}\left(\vect r,\vect{\omega}\right)=0\label{eq:Helmholtz eq}
172 \begin_inset Note Note
175 \begin_layout Plain Layout
176 frequency-space Maxwell's equations
179 \nabla\times\vect E\left(\vect r,\omega\right)-ik\eta_{0}\eta\vect H\left(\vect r,\omega\right) & =0,\\
180 \eta_{0}\eta\nabla\times\vect H\left(\vect r,\omega\right)+ik\vect E\left(\vect r,\omega\right) & =0.
191 \begin_inset Note Note
194 \begin_layout Plain Layout
196 \begin_inset Formula $\Psi$
199 , mention transversality
205 \begin_inset Formula $k=k\left(\omega\right)=\omega\sqrt{\mubg(\omega)\epsbg(\omega)}/c_{0}$
208 , as can be derived from the Maxwell's equations
209 \begin_inset CommandInset citation
212 key "jackson_classical_1998"
221 \begin_layout Subsubsection
225 \begin_layout Standard
227 \begin_inset CommandInset ref
229 reference "eq:Helmholtz eq"
236 can be solved by separation of variables in spherical coordinates to give
245 vector spherical wavefunctions (VSWFs)
246 \begin_inset Formula $\vswfrtlm{\tau}lm\left(k\vect r\right)$
250 \begin_inset Formula $\vswfouttlm{\tau}lm\left(k\vect r\right)$
253 , respectively, defined as follows:
256 \vswfrtlm 1lm\left(k\vect r\right) & =j_{l}\left(kr\right)\vsh 1lm\left(\uvec r\right),\nonumber \\
257 \vswfrtlm 2lm\left(k\vect r\right) & =\frac{1}{kr}\frac{\ud\left(krj_{l}\left(kr\right)\right)}{\ud\left(kr\right)}\vsh 2lm\left(\uvec r\right)+\sqrt{l\left(l+1\right)}\frac{j_{l}\left(kr\right)}{kr}\vsh 3lm\left(\uvec r\right),\label{eq:VSWF regular}
265 \vswfouttlm 1lm\left(k\vect r\right) & =h_{l}^{\left(1\right)}\left(kr\right)\vsh 1lm\left(\uvec r\right),\nonumber \\
266 \vswfouttlm 2lm\left(k\vect r\right) & =\frac{1}{kr}\frac{\ud\left(krh_{l}^{\left(1\right)}\left(kr\right)\right)}{\ud\left(kr\right)}\vsh 2lm\left(\uvec r\right)+\sqrt{l\left(l+1\right)}\frac{h_{l}^{\left(1\right)}\left(kr\right)}{kr}\vsh 3lm\left(\uvec r\right),\label{eq:VSWF outgoing}\\
267 & \tau=1,2;\quad l=1,2,3,\dots;\quad m=-l,-l+1,\dots,+l,\nonumber
273 \begin_inset Formula $\vect r=r\uvec r$
277 \begin_inset Formula $j_{l}\left(x\right),h_{l}^{\left(1\right)}\left(x\right)=j_{l}\left(x\right)+iy_{l}\left(x\right)$
280 are the regular spherical Bessel function and spherical Hankel function
281 of the first kind, respectively, as in [DLMF §10.47], and
282 \begin_inset Formula $\vsh{\tau}lm$
287 vector spherical harmonics
292 \vsh 1lm\left(\uvec r\right) & =\frac{1}{\sqrt{l\left(l+1\right)}}\nabla\times\left(\vect r\ush lm\left(\uvec r\right)\right)=\frac{1}{\sqrt{l\left(l+1\right)}}\nabla\ush lm\left(\uvec r\right)\times\vect r,\nonumber \\
293 \vsh 2lm\left(\uvec r\right) & =\frac{1}{\sqrt{l\left(l+1\right)}}r\nabla\ush lm\left(\uvec r\right),\nonumber \\
294 \vsh 3lm\left(\uvec r\right) & =\uvec r\ush lm\left(\uvec r\right).\label{eq:vector spherical harmonics definition}
299 In our convention, the (scalar) spherical harmonics
300 \begin_inset Formula $\ush lm$
303 are identical to those in [DLMF 14.30.1], i.e.
306 \ush lm=\left(\frac{\left(l-m\right)!\left(2l+1\right)}{4\pi\left(l+m\right)!}\right)^{\frac{1}{2}}e^{im\phi}\dlmfFer lm\left(\cos\theta\right)
311 where importantly, the Ferrers functions
312 \begin_inset Formula $\dlmfFer lm$
315 defined as in [DLMF §14.3(i)] do already contain the Condon-Shortley phase
317 \begin_inset Formula $\left(-1\right)^{m}$
321 \begin_inset Note Note
324 \begin_layout Plain Layout
325 TODO názornější definice.
333 \begin_layout Standard
334 The convention for VSWFs used here is the same as in
335 \begin_inset CommandInset citation
337 key "kristensson_spherical_2014"
342 ; over other conventions used elsewhere in literature, it has several fundamenta
343 l advantages – most importantly, the translation operators introduced later
346 \begin_inset CommandInset ref
348 reference "eq:translation op def"
355 are unitary, and it gives the simplest possible expressions for power transport
356 and cross sections without additional
357 \begin_inset Formula $l,m$
360 -dependent factors (for that reason, we also call our VSWFs as
365 Power-normalisation and unitary translation operators are possible to achieve
366 also with real spherical harmonics – such a convention is used in
367 \begin_inset CommandInset citation
369 key "kristensson_scattering_2016"
377 \begin_layout Standard
378 \begin_inset Note Note
381 \begin_layout Plain Layout
382 Its solutions (TODO under which conditions? What vector space do the SVWFs
383 actually span? Check Comment 9.2 and Appendix f.9.1 in Kristensson)
391 \begin_layout Standard
392 \begin_inset Note Note
395 \begin_layout Plain Layout
396 TODO small note about cartesian multipoles, anapoles etc.
397 (There should be some comparing paper that the Russians at META 2018 mentioned.)
405 \begin_layout Subsubsection
409 \begin_layout Standard
411 \begin_inset Formula $\vswfrtlm{\tau}lm\left(k\vect r\right)$
414 constitute a basis for solutions of the Helmholtz equation
415 \begin_inset CommandInset ref
417 reference "eq:Helmholtz eq"
425 \begin_inset Formula $\openball 0{R^{>}}$
429 \begin_inset Formula $R^{>}$
432 and center in the origin; however, if the equation is not guaranteed to
433 hold inside a smaller ball
434 \begin_inset Formula $B_{0}\left(R\right)$
437 around the origin (typically due to presence of a scatterer), one has to
438 add the outgoing VSWFs
439 \begin_inset Formula $\vswfrtlm{\tau}lm\left(k\vect r\right)$
442 to have a complete basis of the solutions in the volume
443 \begin_inset Formula $\openball 0{R^{>}}\backslash B_{0}\left(R\right)$
447 \begin_inset Note Note
450 \begin_layout Plain Layout
451 Vnitřní koule uzavřená? Jak se řekne mezikulí anglicky?
459 \begin_layout Standard
460 The single-particle scattering problem at frequency
461 \begin_inset Formula $\omega$
464 can be posed as follows: Let a scatterer be enclosed inside the ball
465 \begin_inset Formula $B_{0}\left(R\right)$
468 and let the whole volume
469 \begin_inset Formula $\openball 0{R^{>}}\backslash B_{0}\left(R\right)$
472 be filled with a homogeneous isotropic medium with wave number
473 \begin_inset Formula $k\left(\omega\right)$
477 Inside this volume, the electric field can be expanded as
478 \begin_inset Note Note
481 \begin_layout Plain Layout
482 doplnit frekvence a polohy
490 \vect E\left(\omega,\vect r\right)=\sum_{\tau=1,2}\sum_{l=1}^{\infty}\sum_{m=-l}^{+l}\left(\rcoefftlm{\tau}lm\vswfrtlm{\tau}lm\left(k\vect r\right)+\outcoefftlm{\tau}lm\vswfouttlm{\tau}lm\left(k\vect r\right)\right).\label{eq:E field expansion}
495 If there was no scatterer and
496 \begin_inset Formula $B_{0}\left(R_{<}\right)$
499 was filled with the same homogeneous medium, the part with the outgoing
500 VSWFs would vanish and only the part
501 \begin_inset Formula $\vect E_{\mathrm{inc}}=\sum_{\tau lm}\rcoefftlm{\tau}lm\vswfrtlm{\tau}lm$
504 due to sources outside
505 \begin_inset Formula $\openball 0R$
509 Let us assume that the
510 \begin_inset Quotes eld
514 \begin_inset Quotes erd
517 is given, so that presence of the scatterer does not affect
518 \begin_inset Formula $\vect E_{\mathrm{inc}}$
521 and is fully manifested in the latter part,
522 \begin_inset Formula $\vect E_{\mathrm{scat}}=\sum_{\tau lm}\outcoefftlm{\tau}lm\vswfouttlm{\tau}lm$
526 We also assume that the scatterer is made of optically linear materials,
527 and hence reacts on the incident field in a linear manner.
528 This gives a linearity constraint between the expansion coefficients
531 \outcoefftlm{\tau}lm=\sum_{\tau'l'm'}T_{\tau lm}^{\tau'l'm'}\rcoefftlm{\tau'}{l'}{m'}\label{eq:T-matrix definition}
537 \begin_inset Formula $T_{\tau lm}^{\tau'l'm'}=T_{\tau lm}^{\tau'l'm'}\left(\omega\right)$
540 are the elements of the
546 \begin_inset Formula $T$
550 It completely describes the scattering properties of a linear scatterer,
551 so with the knowledge of the
552 \begin_inset Formula $T$
555 -matrix, we can solve the single-patricle scatering prroblem simply by substitut
556 ing appropriate expansion coefficients
557 \begin_inset Formula $\rcoefftlm{\tau'}{l'}{m'}$
560 of the driving field into
561 \begin_inset CommandInset ref
563 reference "eq:T-matrix definition"
571 The outgoing VSWF expansion coefficients
572 \begin_inset Formula $\outcoefftlm{\tau}lm$
575 are the effective induced electric (
576 \begin_inset Formula $\tau=2$
580 \begin_inset Formula $\tau=1$
583 ) multipole polarisation amplitudes of the scatterer, and this is why we
584 sometimes refer to the corresponding VSWFs as the electric and magnetic
587 \begin_inset Note Note
590 \begin_layout Plain Layout
591 TODO mention the pseudovector character of magnetic VSWFs.
599 \begin_layout Standard
600 \begin_inset Note Note
603 \begin_layout Plain Layout
604 TOOD H-field expansion here?
612 \begin_layout Standard
613 \begin_inset Formula $T$
616 -matrices of particles with certain simple geometries (most famously spherical)
617 can be obtained analytically
618 \begin_inset CommandInset citation
620 key "kristensson_scattering_2016,mie_beitrage_1908"
625 , but in general one can find them numerically by simulating scattering
626 of a regular spherical wave
627 \begin_inset Formula $\vswfouttlm{\tau}lm$
630 and projecting the scattered fields (or induced currents, depending on
631 the method) onto the outgoing VSWFs
632 \begin_inset Formula $\vswfrtlm{\tau}{'l'}{m'}$
636 In practice, one can compute only a finite number of elements with a cut-off
638 \begin_inset Formula $L$
641 on the multipole degree,
642 \begin_inset Formula $l,l'\le L$
646 We typically use the scuff-tmatrix tool from the free software SCUFF-EM
648 \begin_inset CommandInset citation
650 key "reid_efficient_2015,SCUFF2"
656 Note that older versions of SCUFF-EM contained a bug that rendered almost
658 \begin_inset Formula $T$
661 -matrix results wrong; we found and fixed the bug and from upstream version
662 xxx onwards, it should behave correctly.
666 \begin_layout Subsubsection
667 T-matrix compactness, cutoff validity
670 \begin_layout Standard
672 \begin_inset Formula $T$
675 -matrix elements depends heavily on the scatterer's size compared to the
678 \begin_inset Formula $T$
681 -matrix of a bounded scatterer is a compact operator [REF???], so from certain
682 multipole degree onwards,
683 \begin_inset Formula $l,l'>L$
686 , the elements of the
687 \begin_inset Formula $T$
690 -matrix are negligible, so truncating the
691 \begin_inset Formula $T$
694 -matrix at finite multipole degree
695 \begin_inset Formula $L$
698 gives a good approximation of the actual infinite-dimensional itself.
699 If the incident field is well-behaved, i.e.
700 the expansion coefficients
701 \begin_inset Formula $\rcoefftlm{\tau'}{l'}{m'}$
704 do not take excessive values for
705 \begin_inset Formula $l'>L$
708 , the scattered field expansion coefficients
709 \begin_inset Formula $\outcoefftlm{\tau}lm$
713 \begin_inset Formula $l>L$
716 will also be negligible.
717 \begin_inset Note Note
720 \begin_layout Plain Layout
721 TODO when it will not be negligible
729 \begin_layout Standard
730 A rule of thumb to choose the
731 \begin_inset Formula $L$
735 \begin_inset Formula $T$
738 -matrix element accuracy
739 \begin_inset Formula $\delta$
742 can be obtained from the spherical Bessel function expansion around zero,
747 \begin_layout Subsubsection
751 \begin_layout Standard
752 For convenience, let us introduce a short-hand matrix notation for the expansion
753 coefficients and related quantities, so that we do not need to write the
754 indices explicitly; so for example, eq.
756 \begin_inset CommandInset ref
758 reference "eq:T-matrix definition"
766 \begin_inset Formula $\outcoeffp{}=T\rcoeffp{}$
770 \begin_inset Formula $\rcoeffp{},\outcoeffp{}$
773 are column vectors with the expansion coefficients.
774 Transposed and complex-conjugated matrices are labeled with the
775 \begin_inset Formula $\dagger$
781 \begin_layout Standard
782 With this notation, we state an important result about power transport,
783 derivation of which can be found in
784 \begin_inset CommandInset citation
787 key "kristensson_scattering_2016"
794 \begin_inset Formula $\openball 0{R^{>}}\backslash B_{0}\left(R\right)$
798 \begin_inset CommandInset ref
800 reference "eq:E field expansion"
808 Then the net power transported from
809 \begin_inset Formula $B_{0}\left(R\right)$
813 \begin_inset Formula $\openball 0{R^{>}}\backslash B_{0}\left(R\right)$
816 via by electromagnetic radiation is
819 P=\frac{1}{2k^{2}\eta_{0}\eta}\left(\Re\left(\rcoeffp{}^{\dagger}\outcoeffp{}\right)+\left\Vert \outcoeffp{}\right\Vert ^{2}\right)=\frac{1}{2k^{2}\eta_{0}\eta}\rcoeffp{}^{\dagger}\left(\Tp{}^{\dagger}\Tp{}+\frac{\Tp{}^{\dagger}+\Tp{}}{2}\right)\rcoeffp{}.\label{eq:Power transport}
824 In realistic scattering setups, power is transferred by radiation into
825 \begin_inset Formula $B_{0}\left(R\right)$
828 and absorbed by the enclosed scatterer, so
829 \begin_inset Formula $P$
832 is negative and its magnitude equals to power absorbed by the scatterer.
835 \begin_layout Subsubsection
839 \begin_layout Standard
840 In many scattering problems considered in practice, the driving field is
843 \begin_inset Formula $\vect k\cdot\vect E_{0}=0$
846 ) plane wave propagating in direction
847 \begin_inset Formula $\uvec k$
850 with (complex) amplitude
851 \begin_inset Formula $\vect E_{0}$
854 can be expanded into regular VSWFs
855 \begin_inset CommandInset citation
858 key "kristensson_scattering_2016"
866 \vect E_{\mathrm{PW}}\left(\vect r,\omega\right)=\vect E_{0}e^{ik\uvec k\cdot\vect r}=\sum_{\tau,l,m}\rcoeffptlm{}{\tau}lm\left(\vect k,\vect E_{0}\right)\vswfrtlm{\tau}lm\left(k\vect r\right),
871 with expansion coefficients
874 \rcoeffptlm{}1lm\left(\vect k,\vect E_{0}\right) & = & 4\pi i^{l}\vshD 1lm\left(\uvec k\right),\nonumber \\
875 \rcoeffptlm{}2lm\left(\vect k,\vect E_{0}\right) & = & -4\pi i^{l+1}\vshD 2lm\left(\uvec k\right).\label{eq:plane wave expansion}
881 \begin_inset Formula $\vshD{\tau}lm$
885 \begin_inset Quotes eld
889 \begin_inset Quotes erd
892 vector spherical harmonics defined by duality relation with the
893 \begin_inset Quotes eld
897 \begin_inset Quotes erd
900 vector spherical harmonics
903 \iint\vshD{\tau'}{l'}{m'}\left(\uvec r\right)\cdot\vsh{\tau}lm\left(\uvec r\right)\,\ud\Omega=\delta_{\tau'\tau}\delta_{l'l}\delta_{m'm}\label{eq:dual vsh}
908 (complex conjugation not implied in the dot product here).
909 In our convention, we have
912 \vshD{\tau}lm\left(\uvec r\right)=\left(\vsh{\tau}lm\left(\uvec r\right)\right)^{*}=\left(-1\right)^{m}\vsh{\tau}{l-}m\left(\uvec r\right).
920 \begin_layout Subsubsection
921 Cross-sections (single-particle)
924 \begin_layout Standard
926 \begin_inset Formula $T$
929 -matrix and expansion coefficients of plane waves in hand, we can state
930 the expressions for cross-sections of a single scatterer.
931 Assuming a non-lossy background medium, extinction, scattering and absorption
932 cross sections of a single scatterer irradiated by a plane wave propagating
934 \begin_inset Formula $\uvec k$
937 and (complex) amplitude
938 \begin_inset Formula $\vect E_{0}$
942 \begin_inset CommandInset citation
945 key "kristensson_scattering_2016"
953 \sigma_{\mathrm{ext}}\left(\uvec k\right) & = & -\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\Re\left(\rcoeffp{}^{\dagger}\outcoeffp{}\right)=-\frac{1}{2k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\rcoeffp{}^{\dagger}\left(\Tp{}+\Tp{}^{\dagger}\right)\rcoeffp{},\label{eq:extincion CS single}\\
954 \sigma_{\mathrm{scat}}\left(\uvec k\right) & = & \frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\left\Vert \outcoeffp{}\right\Vert ^{2}=\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\rcoeffp{}^{\dagger}\left(\Tp{}^{\dagger}\Tp{}\right)\rcoeffp{},\label{eq:scattering CS single}\\
955 \sigma_{\mathrm{abs}}\left(\uvec k\right) & = & \sigma_{\mathrm{ext}}\left(\uvec k\right)-\sigma_{\mathrm{scat}}\left(\uvec k\right)=-\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\left(\Re\left(\rcoeffp{}^{\dagger}\outcoeffp{}\right)+\left\Vert \outcoeffp{}\right\Vert ^{2}\right)\nonumber \\
956 & & =\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\rcoeffp{}^{\dagger}\left(\Tp{}^{\dagger}\Tp{}+\frac{\Tp{}^{\dagger}+\Tp{}}{2}\right)\rcoeffp{},\label{eq:absorption CS single}
962 \begin_inset Formula $\rcoeffp{}=\rcoeffp{}\left(\vect k,\vect E_{0}\right)$
965 is the vector of plane wave expansion coefficients as in
966 \begin_inset CommandInset ref
968 reference "eq:plane wave expansion"
976 \begin_layout Subsection
978 \begin_inset CommandInset label
980 name "subsec:Multiple-scattering"
987 \begin_layout Standard
988 If the system consists of multiple scatterers, the EM fields around each
989 one can be expanded in analogous way.
991 \begin_inset Formula $\mathcal{P}$
994 be an index set labeling the scatterers.
995 We enclose each scatterer in a ball
996 \begin_inset Formula $B_{\vect r_{p}}\left(R_{p}\right)$
999 such that the balls do not touch,
1000 \begin_inset Formula $B_{\vect r_{p}}\left(R_{p}\right)\cap B_{\vect r_{q}}\left(R_{q}\right)=\emptyset;p,q\in\mathcal{P}$
1004 \begin_inset Note Note
1007 \begin_layout Plain Layout
1008 TODO bacha, musejí být uzavřené!
1013 so there is a non-empty volume
1014 \begin_inset Note Note
1017 \begin_layout Plain Layout
1024 \begin_inset Formula $\openball{\vect r_{p}}{R_{p}^{>}}\backslash B_{\vect r_{p}}\left(R_{p}\right)$
1027 around each one that contains only the background medium without any scatterers.
1028 Then the EM field inside each such volume can be expanded in a way similar
1030 \begin_inset CommandInset ref
1032 reference "eq:E field expansion"
1039 , using VSWFs with origins shifted to the centre of the volume:
1040 \begin_inset Formula
1042 \vect E\left(\omega,\vect r\right) & =\sum_{\tau=1,2}\sum_{l=1}^{\infty}\sum_{m=-l}^{+l}\left(\rcoeffptlm p{\tau}lm\vswfrtlm{\tau}lm\left(k\left(\vect r-\vect r_{p}\right)\right)+\outcoeffptlm p{\tau}lm\vswfouttlm{\tau}lm\left(k\left(\vect r-\vect r_{p}\right)\right)\right),\label{eq:E field expansion multiparticle}\\
1043 & \vect r\in\openball{\vect r_{p}}{R_{p}^{>}}\backslash B_{\vect r_{p}}\left(R_{p}\right).\nonumber
1048 Unlike the single scatterer case, the incident field coefficients
1049 \begin_inset Formula $\rcoeffptlm p{\tau}lm$
1052 here are not only due to some external driving field that the particle
1053 does not influence but they also contain the contributions of fields scattered
1056 all other scatterers
1059 \begin_inset Formula
1061 \rcoeffp p=\rcoeffincp p+\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\tropsp pq\outcoeffp q\label{eq:particle total incident field coefficient a}
1067 \begin_inset Formula $\rcoeffincp p$
1070 represents the part due to the external driving that the scatterers can
1072 \begin_inset Formula $\tropsp pq$
1077 translation operator
1079 defined below in Sec.
1081 \begin_inset CommandInset ref
1083 reference "subsec:Translation-operator"
1090 , that contains the re-expansion coefficients of the outgoing waves in origin
1092 \begin_inset Formula $\vect r_{q}$
1095 into regular waves in origin
1096 \begin_inset Formula $\vect r_{p}$
1100 For each scatterer, we also have its
1101 \begin_inset Formula $T$
1104 -matrix relation as in
1105 \begin_inset CommandInset ref
1107 reference "eq:T-matrix definition"
1115 \begin_inset Formula
1117 \outcoeffp q=T_{q}\rcoeffp q.
1123 \begin_inset CommandInset ref
1125 reference "eq:particle total incident field coefficient a"
1132 , this gives rise to a set of linear equations
1133 \begin_inset Formula
1135 \outcoeffp p-T_{p}\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\tropsp pq\outcoeffp q=T_{p}\rcoeffincp p,\quad p\in\mathcal{P}\label{eq:Multiple-scattering problem}
1140 which defines the multiple-scattering problem.
1142 \begin_inset Formula $p,q$
1145 -indexed vectors and matrices (note that without truncation, they are infinite-d
1146 imensional) are arranged into blocks of even larger vectors and matrices,
1147 this can be written in a short-hand form
1148 \begin_inset Formula
1150 \left(I-T\trops\right)\outcoeff=T\rcoeffinc\label{eq:Multiple-scattering problem block form}
1156 \begin_inset Formula $I$
1159 is the identity matrix,
1160 \begin_inset Formula $T$
1163 is a block-diagonal matrix containing all the individual
1164 \begin_inset Formula $T$
1168 \begin_inset Formula $\trops$
1171 contains the individual
1172 \begin_inset Formula $\tropsp pq$
1175 matrices as the off-diagonal blocks, whereas the diagonal blocks are set
1179 \begin_layout Standard
1180 In practice, the multiple-scattering problem is solved in its truncated
1181 form, in which all the
1182 \begin_inset Formula $l$
1185 -indices related to a given scatterer
1186 \begin_inset Formula $p$
1190 \begin_inset Formula $l\le L_{p}$
1194 \begin_inset Formula $N_{p}=2L_{p}\left(L_{p}+2\right)$
1198 \begin_inset Formula $\tau lm$
1202 The truncation degree can vary for different scatterers (e.g.
1203 due to different physical sizes), so the truncated block
1204 \begin_inset Formula $\tropsp pq$
1208 \begin_inset Formula $N_{p}\times N_{q}$
1211 , not necessarily square.
1213 \begin_inset Note Note
1216 \begin_layout Plain Layout
1217 Such truncation of the translation operator
1218 \begin_inset Formula $\tropsp pq$
1221 is justified by the fact on the left, TODO
1229 \begin_layout Standard
1230 If no other type of truncation is done, there remain
1231 \begin_inset Formula $2L_{p}\left(L_{p}+2\right)$
1235 \begin_inset Formula $\tau lm$
1239 \begin_inset Formula $p$
1242 -th scatterer, so that the truncated version of the matrix
1243 \begin_inset Formula $\left(I-T\trops\right)$
1246 is a square matrix with
1247 \begin_inset Formula $\left(\sum_{p\in\mathcal{P}}N_{p}\right)^{2}$
1251 The truncated problem
1252 \begin_inset CommandInset ref
1254 reference "eq:Multiple-scattering problem block form"
1261 can then be solved using standard numerical linear algebra methods (typically,
1262 by LU factorisation of the
1263 \begin_inset Formula $\left(I-T\trops\right)$
1266 matrix at a given frequency, and then solving with Gauss elimination for
1267 as many different incident
1268 \begin_inset Formula $\rcoeffinc$
1274 \begin_layout Standard
1275 Alternatively, the multiple scattering problem can be formulated in terms
1276 of the regular field expansion coefficients,
1277 \begin_inset Formula
1279 \rcoeffp p-\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\tropsp pqT_{q}\rcoeffp q & =\rcoeffincp p,\quad p\in\mathcal{P},\\
1280 \left(I-\trops T\right)\rcoeff & =\rcoeffinc,
1285 but this form is less suitable for numerical calculations due to the fact
1286 that the regular VSWF expansion coefficients on both sides of the equation
1287 are typically non-negligible even for large multipole degree
1288 \begin_inset Formula $l$
1291 , hence the truncation is not justified in this case.
1293 \begin_inset Note Note
1296 \begin_layout Plain Layout
1305 \begin_layout Subsubsection
1306 Translation operator
1307 \begin_inset CommandInset label
1309 name "subsec:Translation-operator"
1316 \begin_layout Standard
1318 \begin_inset Formula $\vect r_{1},\vect r_{2}$
1321 be two different origins; a regular VSWF with origin
1322 \begin_inset Formula $\vect r_{1}$
1325 can be always expanded in terms of regular VSWFs with origin
1326 \begin_inset Formula $\vect r_{2}$
1332 \begin_layout Standard
1333 \begin_inset Formula
1335 \vswfrtlm{\tau}lm\left(k\left(\vect r-\vect r_{1}\right)\right)=\sum_{\tau'l'm'}\tropr_{\tau lm;\tau'l'm'}\left(k\left(\vect r_{2}-\vect r_{1}\right)\right)\vswfrtlm{\tau'}{l'}{m'}\left(\vect r-\vect r_{2}\right),\label{eq:regular vswf translation}
1340 where an explicit formula for the (regular)
1342 translation operator
1345 \begin_inset Formula $\tropr$
1350 \begin_inset CommandInset ref
1352 reference "eq:translation operator"
1357 For singular (outgoing) waves, the form of the expansion differs inside
1358 and outside the ball
1359 \begin_inset Formula $\openball{\left\Vert \vect r_{2}-\vect r_{1}\right\Vert }{\vect r_{1}}:$
1363 \begin_inset Formula
1365 \vswfouttlm{\tau}lm\left(k\left(\vect r-\vect r_{1}\right)\right) & = & \begin{cases}
1366 \sum_{\tau'l'm'}\trops_{\tau lm;\tau'l'm'}\left(k\left(\vect r_{2}-\vect r_{1}\right)\right)\vswfouttlm{\tau'}{l'}{m'}\left(\vect r-\vect r_{2}\right), & \vect r\in\openball{\left\Vert \vect r_{2}-\vect r_{1}\right\Vert }{\vect r_{1}}\\
1367 \sum_{\tau'l'm'}\tropr_{\tau lm;\tau'l'm'}\left(k\left(\vect r_{2}-\vect r_{1}\right)\right)\vswfrtlm{\tau'}{l'}{m'}\left(\vect r-\vect r_{2}\right), & \vect r\notin\closedball{\left\Vert \vect r_{2}-\vect r_{1}\right\Vert }{\left|\vect r_{1}\right|}
1368 \end{cases},\label{eq:singular vswf translation}
1373 where the singular translation operator
1374 \begin_inset Formula $\trops$
1377 has the same form as
1378 \begin_inset Formula $\tropr$
1382 \begin_inset CommandInset ref
1384 reference "eq:translation operator"
1388 except the regular spherical Bessel functions
1389 \begin_inset Formula $j_{l}$
1392 are replaced with spherical Hankel functions
1393 \begin_inset Formula $h_{l}^{(1)}$
1398 \begin_inset Note Note
1401 \begin_layout Plain Layout
1402 TODO note about expansion exactly on the sphere.
1410 \begin_layout Standard
1411 As MSTMM deals most of the time with the
1413 expansion coefficients
1416 \begin_inset Formula $\rcoeffptlm p{\tau}lm,\outcoeffptlm p{\tau}lm$
1419 in different origins
1420 \begin_inset Formula $\vect r_{p}$
1423 rather than with the VSWFs directly, let us write down how
1427 transform under translation.
1428 Let us assume the field can be in terms of regular waves everywhere, and
1429 expand it in two different origins
1430 \begin_inset Formula $\vect r_{p},\vect r_{q}$
1434 \begin_inset Formula
1436 \vect E\left(\vect r,\omega\right)=\sum_{\tau,l,m}\rcoeffptlm p{\tau}lm\vswfrtlm{\tau}lm\left(k\left(\vect r-\vect r_{p}\right)\right)=\sum_{\tau',l',m'}\rcoeffptlm q{\tau'}{l'}{m'}\vswfrtlm{\tau}{'l'}{m'}\left(k\left(\vect r-\vect r_{q}\right)\right).
1441 Re-expanding the waves around
1442 \begin_inset Formula $\vect r_{p}$
1445 in terms of waves around
1446 \begin_inset Formula $\vect r_{q}$
1450 \begin_inset CommandInset ref
1452 reference "eq:regular vswf translation"
1457 \begin_inset Formula
1459 \vect E\left(\vect r,\omega\right)=\sum_{\tau,l,m}\rcoeffptlm p{\tau}lm\sum_{\tau'l'm'}\tropr_{\tau lm;\tau'l'm'}\left(k\left(\vect r_{q}-\vect r_{p}\right)\right)\vswfrtlm{\tau'}{l'}{m'}\left(\vect r-\vect r_{q}\right)
1464 and comparing to the original expansion around
1465 \begin_inset Formula $\vect r_{q}$
1469 \begin_inset Formula
1471 \rcoeffptlm q{\tau'}{l'}{m'}=\sum_{\tau,l,m}\tropr_{\tau lm;\tau'l'm'}\left(k\left(\vect r_{q}-\vect r_{p}\right)\right)\rcoeffptlm p{\tau}lm.\label{eq:regular vswf coefficient translation}
1476 For the sake of readability, we introduce a shorthand matrix form for
1477 \begin_inset CommandInset ref
1479 reference "eq:regular vswf coefficient translation"
1484 \begin_inset Formula
1486 \rcoeffp q=\troprp qp\rcoeffp p\label{eq:reqular vswf coefficient vector translation}
1491 (note the reversed indices; TODO redefine them in
1492 \begin_inset CommandInset ref
1494 reference "eq:regular vswf translation"
1499 \begin_inset CommandInset ref
1501 reference "eq:singular vswf translation"
1505 ? Similarly, if we had only outgoing waves in the original expansion around
1507 \begin_inset Formula $\vect r_{p}$
1511 \begin_inset Formula
1513 \rcoeffp q=\tropsp qp\outcoeffp p\label{eq:singular to regular vswf coefficient vector translation}
1518 for the expansion inside the ball
1519 \begin_inset Formula $\openball{\left\Vert \vect r_{q}-\vect r_{p}\right\Vert }{\vect r_{p}}$
1523 \begin_inset Note Note
1526 \begin_layout Plain Layout
1533 \begin_inset Formula
1535 \outcoeffp q=\troprp qp\outcoeffp p\label{eq:singular to singular vswf coefficient vector translation-1}
1543 \begin_layout Standard
1544 In our convention, the regular translation operator can be expressed explicitly
1545 as (TODO CHECK CAREFULLY FOR POSSIBLE
1546 \begin_inset Formula $(-1)^{m'}$
1549 AND SIMILAR FACTORS AND REWRITE IN TERMS OF SPHERICAL HARMONICS)
1550 \begin_inset Note Note
1553 \begin_layout Plain Layout
1554 Teďka jsou tam zkopírovány výrazy pro C a D z Kristenssona, chybějí fase
1560 \begin_inset Formula
1562 \tropr_{\tau lm;\tau l'm'}\left(\vect d\right)=\frac{\left(-1\right)^{m+m'}}{2}\sum_{\lambda=\left|l-l'\right|}^{l+l'}\left(-1\right)^{\frac{l'-l+\lambda}{2}}\left(2\lambda+1\right)\sqrt{\frac{\left(2l+1\right)\left(2l'+1\right)\left(\lambda-\left(m-m'\right)\right)!}{l\left(l+1\right)l'\left(l'+1\right)\left(\lambda+\left(m-m'\right)\right)!}}\times\\
1563 \times\begin{pmatrix}l & l' & \lambda\\
1565 \end{pmatrix}\begin{pmatrix}l & l' & \lambda\\
1567 \end{pmatrix}\left(l\left(l+1\right)+l'\left(l'+1\right)-\lambda\left(\lambda+1\right)\right)\times\\
1568 \times j_{\lambda}\left(d\right)P_{\lambda}^{m-m'}\left(\cos\theta_{\uvec d}\right)e^{i\left(m-m'\right)\phi_{\uvec d}},\\
1569 \tropr_{\tau lm;\tau'l'm'}\left(\vect d\right)=-i\frac{\left(-1\right)^{m+m'}}{2}\sum_{\lambda=\left|l-l'\right|+1}^{l+l'}\left(-1\right)^{\frac{l'-l+\lambda+1}{2}}\left(2\lambda+1\right)\sqrt{\frac{\left(2l+1\right)\left(2l'+1\right)\left(\lambda-\left(m-m'\right)\right)!}{l\left(l+1\right)l'\left(l'+1\right)\left(\lambda+\left(m-m'\right)\right)!}}\times\\
1570 \times\begin{pmatrix}l & l' & \lambda-1\\
1572 \end{pmatrix}\begin{pmatrix}l & l' & \lambda\\
1574 \end{pmatrix}\sqrt{\lambda^{2}-\left(l-l'\right)^{2}}\sqrt{\left(l+l'+1\right)^{2}-\lambda^{2}}\times\\
1575 \times j_{\lambda}\left(d\right)P_{\lambda}^{m-m'}\left(\cos\theta_{\uvec d}\right)e^{i\left(m-m'\right)\phi_{\uvec d}},\qquad\tau\ne\tau'.\label{eq:translation operator}
1580 The singular operator
1581 \begin_inset Formula $\trops$
1584 for re-expanding outgoing waves into regular ones has the same form except
1585 the regular spherical Bessel functions
1586 \begin_inset Formula $j_{l}$
1589 in are replaced with spherical Hankel functions
1590 \begin_inset Formula $h_{l}^{(1)}=j_{l}+iy_{l}$
1596 \begin_layout Standard
1597 In our convention, the regular translation operator is unitary,
1598 \begin_inset Formula $\left(\tropr_{\tau lm;\tau'l'm'}\left(\vect d\right)\right)^{-1}=\tropr_{\tau lm;\tau'l'm'}\left(-\vect d\right)=\tropr_{\tau'l'm';\tau lm}^{*}\left(\vect d\right)$
1602 \begin_inset Note Note
1605 \begin_layout Plain Layout
1606 todo different notation for the complex conjugation without transposition???
1611 or in the per-particle matrix notation,
1612 \begin_inset Formula
1614 \troprp qp^{-1}=\troprp pq=\troprp qp^{\dagger}\label{eq:regular translation unitarity}
1620 Note that truncation at finite multipole degree breaks the unitarity,
1621 \begin_inset Formula $\truncated{\troprp qp}L^{-1}\ne\truncated{\troprp pq}L=\truncated{\troprp qp^{\dagger}}L$
1624 , which has to be taken into consideration when evaluating quantities such
1625 as absorption or scattering cross sections.
1626 Similarly, the full regular operators can be composed
1627 \begin_inset Note Note
1630 \begin_layout Plain Layout
1637 \begin_inset Formula
1639 \troprp ac=\troprp ab\troprp bc\label{eq:regular translation composition}
1644 but truncation breaks this,
1645 \begin_inset Formula $\truncated{\troprp ac}L\ne\truncated{\troprp ab}L\truncated{\troprp bc}L.$
1651 \begin_layout Subsubsection
1652 Cross-sections (many scatterers)
1655 \begin_layout Standard
1656 For a system of many scatterers, Kristensson
1657 \begin_inset CommandInset citation
1660 key "kristensson_scattering_2016"
1665 derives only the extinction cross section formula.
1666 Let us re-derive it together with the many-particle scattering and absorption
1668 First, let us take a ball circumscribing all the scatterers at once,
1669 \begin_inset Formula $\openball R{\vect r_{\square}}\supset\particle$
1674 \begin_inset Formula $\openball R{\vect r_{\square}}$
1677 , we can describe the EM fields as if there was only a single scatterer,
1678 \begin_inset Formula
1680 \vect E\left(\vect r\right)=\sum_{\tau,l,m}\left(\rcoeffptlm{\square}{\tau}lm\vswfrtlm{\tau}lm\left(\vect r-\vect r_{\square}\right)+\outcoeffptlm{\square}{\tau}lm\vswfouttlm{\tau}lm\left(\vect r-\vect r_{\square}\right)\right),
1686 \begin_inset Formula $\rcoeffp{\square},\outcoeffp{\square}$
1689 are the vectors of VSWF expansion coefficients of the incident and total
1690 scattered fields, respectively, at origin
1691 \begin_inset Formula $\vect r_{\square}$
1695 In principle, one could evaluate
1696 \begin_inset Formula $\outcoeffp{\square}$
1699 using the translation operators (REF!!!) and use the single-scatterer formulae
1701 \begin_inset CommandInset ref
1703 reference "eq:extincion CS single"
1708 \begin_inset CommandInset ref
1710 reference "eq:absorption CS single"
1715 \begin_inset Formula $\rcoeffp{}=\rcoeffp{\square},\outcoeffp{}=\outcoeffp{\square}$
1718 to obtain the cross sections.
1719 However, this is not suitable for numerical evaluation with truncation
1720 in multipole degree; hence we need to express them in terms of particle-wise
1722 \begin_inset Formula $\rcoeffp p,\outcoeffp p$
1726 The original incident field re-expanded around
1727 \begin_inset Formula $p$
1730 -th particle reads according to
1731 \begin_inset CommandInset ref
1733 reference "eq:regular vswf translation"
1741 \begin_inset Formula
1743 \rcoeffincp p=\troprp p{\square}\rcoeffp{\square}\label{eq:a_inc local from global}
1748 whereas the contributions of fields scattered from each particle expanded
1749 around the global origin
1750 \begin_inset Formula $\vect r_{\square}$
1754 \begin_inset CommandInset ref
1756 reference "eq:singular vswf translation"
1764 \begin_inset Formula
1766 \outcoeffp{\square}=\sum_{q\in\mathcal{P}}\troprp{\square}q\outcoeffp q.\label{eq:f global from local}
1772 \begin_inset CommandInset ref
1774 reference "eq:regular translation unitarity"
1782 \begin_inset CommandInset ref
1784 reference "eq:regular translation composition"
1792 \begin_inset Formula
1794 \rcoeffp{\square}^{\dagger}\outcoeffp{\square} & =\rcoeffincp p^{\dagger}\troprp p{\square}\troprp{\square}q\outcoeffp q=\rcoeffincp p^{\dagger}\sum_{q\in\mathcal{P}}\troprp pqf_{q}\nonumber \\
1795 & =\sum_{q\in\mathcal{P}}\left(\troprp qp\rcoeffincp p\right)^{\dagger}f_{q}=\sum_{q\in\mathcal{P}}\rcoeffincp q^{\dagger}f_{q},\label{eq:atf form multiparticle}
1800 where only the last expression is suitable for numerical evaluation with
1801 truncated matrices, because the previous ones contain a translation operator
1802 right next to an incident field coefficient vector (see Sec.
1805 \begin_inset Formula
1807 \left\Vert \outcoeffp{\square}\right\Vert ^{2} & =\outcoeffp{\square}^{\dagger}\outcoeffp{\square}=\sum_{p\in\mathcal{P}}\left(\troprp{\square}p\outcoeffp p\right)^{\dagger}\sum_{q\in\mathcal{P}}\troprp{\square}q\outcoeffp q\nonumber \\
1808 & =\sum_{p\in\mathcal{P}}\sum_{q\in\mathcal{P}}\outcoeffp p^{\dagger}\troprp pq\outcoeffp q.\label{eq:f squared form multiparticle}
1814 \begin_inset CommandInset ref
1816 reference "eq:atf form multiparticle"
1824 \begin_inset CommandInset ref
1826 reference "eq:f squared form multiparticle"
1834 \begin_inset CommandInset ref
1836 reference "eq:scattering CS single"
1844 \begin_inset CommandInset ref
1846 reference "eq:absorption CS single"
1853 , we get the many-particle expressions for extinction, scattering and absorption
1854 cross sections suitable for numerical evaluation:
1855 \begin_inset Formula
1857 \sigma_{\mathrm{ext}}\left(\uvec k\right) & = & -\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\Re\sum_{p\in\mathcal{P}}\rcoeffincp p^{\dagger}\outcoeffp p=-\frac{1}{2k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\Re\sum_{p\in\mathcal{P}}\rcoeffincp p^{\dagger}\left(\Tp p+\Tp p^{\dagger}\right)\rcoeffp p,\label{eq:extincion CS multi}\\
1858 \sigma_{\mathrm{scat}}\left(\uvec k\right) & = & \frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\sum_{q\in\mathcal{P}}\outcoeffp p^{\dagger}\troprp pq\outcoeffp q\nonumber \\
1859 & & =\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\sum_{q\in\mathcal{P}}\rcoeffp p^{\dagger}\Tp p^{\dagger}\troprp pq\Tp q\rcoeffp q,\label{eq:scattering CS multi}\\
1860 \sigma_{\mathrm{abs}}\left(\uvec k\right) & = & -\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\Re\left(\outcoeffp p^{\dagger}\left(\rcoeffincp p+\sum_{q\in\mathcal{P}}\troprp pq\outcoeffp q\right)\right).\nonumber \\
1861 \label{eq:absorption CS multi}
1867 \begin_inset Note Note
1870 \begin_layout Plain Layout
1871 \begin_inset Formula $=\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\mbox{TODO}.$
1879 An alternative approach to derive the absorption cross section is via a
1880 power transport argument.
1881 Note the direct proportionality between absorption cross section
1882 \begin_inset CommandInset ref
1884 reference "eq:absorption CS single"
1891 and net radiated power for single scatterer
1892 \begin_inset CommandInset ref
1894 reference "eq:Power transport"
1902 \begin_inset Formula $\sigma_{\mathrm{abs}}=-\eta_{0}\eta P/2\left\Vert \vect E_{0}\right\Vert ^{2}$
1906 In the many-particle setup (with non-lossy background medium, so that only
1907 the particles absorb), the total absorbed power is equal to the sum of
1908 absorbed powers on each particle,
1909 \begin_inset Formula $-P=\sum_{p\in\mathcal{P}}-P_{p}$
1913 Using the power transport formula
1914 \begin_inset CommandInset ref
1916 reference "eq:Power transport"
1924 \begin_inset Formula
1926 \sigma_{\mathrm{abs}}\left(\uvec k\right)=-\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\left(\Re\left(\rcoeffp p^{\dagger}\outcoeffp p\right)+\left\Vert \outcoeffp p\right\Vert ^{2}\right)\label{eq:absorption CS multi alternative}
1931 which seems different from
1932 \begin_inset CommandInset ref
1934 reference "eq:absorption CS multi"
1942 \begin_inset CommandInset ref
1944 reference "eq:particle total incident field coefficient a"
1951 , we can rewrite it as
1952 \begin_inset Formula
1954 \sigma_{\mathrm{abs}}\left(\uvec k\right) & =-\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\Re\left(\outcoeffp p^{\dagger}\left(\rcoeffp p+\outcoeffp p\right)\right)\\
1955 & =-\frac{1}{k^{2}\left\Vert \vect E_{0}\right\Vert ^{2}}\sum_{p\in\mathcal{P}}\Re\left(\outcoeffp p^{\dagger}\left(\rcoeffincp p+\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\tropsp pq\outcoeffp q+\outcoeffp p\right)\right).
1960 It is easy to show that all the terms of
1961 \begin_inset Formula $\sum_{p\in\mathcal{P}}\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\outcoeffp p^{\dagger}\tropsp pq\outcoeffp q$
1964 containing the singular spherical Bessel functions
1965 \begin_inset Formula $y_{l}$
1969 \begin_inset Note Note
1972 \begin_layout Plain Layout
1973 TODO better formulation
1979 \begin_inset Formula $\sum_{p\in\mathcal{P}}\Re\left(\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\outcoeffp p^{\dagger}\tropsp pq\outcoeffp q+\outcoeffp p^{\dagger}\outcoeffp p\right)=\sum_{p\in\mathcal{P}}\Re\left(\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\outcoeffp p^{\dagger}\troprp pq\outcoeffp q+\outcoeffp p^{\dagger}\outcoeffp p\right)=\sum_{p\in\mathcal{P}}\Re\left(\sum_{q\in\mathcal{P}\backslash\left\{ p\right\} }\outcoeffp p^{\dagger}\troprp pq\outcoeffp q+\outcoeffp p^{\dagger}\troprp pp\outcoeffp p\right)=\sum_{p\in\mathcal{P}}\sum_{q\in\mathcal{P}}\outcoeffp p^{\dagger}\troprp pq\outcoeffp q,$
1982 proving that the expressions in
1983 \begin_inset CommandInset ref
1985 reference "eq:absorption CS multi"
1993 \begin_inset CommandInset ref
1995 reference "eq:absorption CS multi alternative"