1: \documentstyle[psfig,twocolumn,amssymb,pre,aps]{revtex}
2: %\documentstyle[psfig,twocolumn,amssymb,pre,aps]{revtex}
3:
4: %\tighten
5: %\draft
6:
7: \newcommand{\alo}{\mbox{${\rm Al}_2{\rm O}_3$}}
8:
9: \begin{document}
10:
11: \wideabs{
12: \title{
13: Partial Resonances of Three-Phase Composites at Long Wavelengths }
14:
15: \author{N. A. Nicorovici and R. C. McPhedran}
16: \address{School of Physics, University of Sydney, NSW 2006, Australia}
17: \author{G. W. Milton}
18: \address{Department of Mathematics, University of Utah,
19: Salt Lake City, UT 84112, USA}
20: \author{L. C. Botten}
21: \address{Department of Mathematical Sciences, University of
22: Technology, Sydney, Broadway, P O Box 123, NSW 2007, Australia\\~\\}
23:
24: \date{\today}
25: \maketitle
26:
27: \begin{abstract}
28: We investigate the behaviour of a three-phase composite,
29: structured as a hexagonal array of coated cylinders, when the sum
30: of the dielectric constants of the core and shell equals zero. In
31: such cases, the absorption of the composite is the same as the
32: absorption of a periodic array of solid cylinders of core material
33: and radius equal to the outer radius of the original coated
34: cylinder. When the sum of the dielectric constants of the shell
35: and matrix equals zero, the composite has the same absorption as a
36: periodic array of solid cylinders of core material, and radius
37: exceeding the shell radius.
38: %We show that this effect is also
39: %related with the anomalous absorptance of metallic cylinders.
40: \end{abstract}
41: }
42:
43: \section{Introduction}
44: \label{intro}
45:
46: Previous studies of the effective dielectric constant of a
47: periodic array of coated cylinders, in the quasistatic limit, have
48: revealed some unexpected and interesting
49: results\cite{cc1,cc2,cc3}.
50: Specifically, we have used the method of Lord
51: Rayleigh\cite{ray} to investigate the analytic properties of the
52: effective dielectric constant of the array, as a function of the core
53: dielectric constant ($\varepsilon_c$) and the shell dielectric constant
54: ($\varepsilon_s$), while keeping the matrix dielectric constant
55: ($\varepsilon_m$) fixed.
56:
57: We have shown that when
58: $\varepsilon_c+\varepsilon_s=0$, the composite has exactly the
59: same effective dielectric constant as a periodic array of solid
60: cylinders with dielectric constant
61: $\varepsilon_c$ and radius equal to the
62: outer radius of the original coated cylinder. We have also shown
63: that when
64: $\varepsilon_s+\varepsilon_m=0$,
65: the composite has exactly the same effective
66: dielectric constant as a periodic array of solid cylinders with
67: dielectric constant $\varepsilon_c$,
68: and radius exceeding the shell radius.
69: In both cases, the core is {\em magnified} and it may be possible for
70: a system with a vanishingly small concentration
71: to exhibit properties like those of a concentrated system.
72: This is what we have called a {\em partially resonant} system,
73: since the system has
74: a finite, rather than an infinite, response to the applied field.
75:
76: The same effect appears in the case of a random mixture
77: of coated cylinders with nonlinear cores and linear
78: shells\cite{ohad}.
79: Under conditions of partial resonance, the system behaves as a
80: mixture of nonlinear solid cylinders of radii equal to, or
81: larger than, those of the original coated cylinders.
82:
83: Here, we show how these results carry over to problems of
84: electromagnetic scattering by periodic arrays of
85: coated cylinders, and elucidate the relation between
86: partial resonances and
87: anomalous absorptance of metallic wires \cite{anomaly}.
88: From the many possible ways of packing cylinders in
89: regular arrays in two dimensions we concentrate on the hexagonal
90: lattice that is isotropic for fields applied in the plane
91: perpendicular to the cylinder axes (see Fig.~\ref{WSC}).
92: However, the method we use
93: may be applied equally well to any periodic
94: array of coated cylinders.
95:
96: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
97: \begin{figure}
98: \vspace{6ex}
99: \centerline{\psfig{file=WSC.eps,height=3.0in}}
100: \vspace{6ex}
101: \caption{The Wigner-Seitz cell (red) with the central inclusion
102: (coated cylinder with $r_c=0.03d$, $r_s=0.31d$ and $d=1$) and the
103: nearest neighbours (located along the dashed circle) for a
104: hexagonal array. The centers of the nearest neighbours are given
105: by the lattice vectors ${\bf R}_p$ for
106: $p\in\{(-1,0),(-1,1),(0,-1),(0,1),(1,-1),(1,0)\}$. We also show
107: the fundamental translation vectors ${\bf e}_1=d(1, 0)$ and ${\bf
108: e}_2=d(1/2, \sqrt{3}/2)$.} \label{WSC}
109: \end{figure}
110: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
111:
112: Recent work by Pendry {\em et al}. \cite{pendry} has heightened interest in the
113: extraordinary and sometimes paradoxical behaviour which can arise
114: in composite systems, in which one component has a dielectric constant
115: (or a magnetic permeability) which takes approximately a real and
116: negative value.
117: In such a composite we may observe strong focusing of light at a plane
118: interface, or ultra-refraction \cite{gralak}, or a dilute composite
119: {\em masquerading} as a concentrated one \cite{cc1,cc2,cc3}.
120: Again, the use of metal-coated spheres in photonic crystals has resulted
121: in robust band gaps, independent of sphere-packing geometries \cite{pingsheng}.
122:
123:
124:
125: \section{The Field Identity}
126: \label{raysec}
127: \subsection{Field Expansions}
128: \label{fieldssec}
129:
130: We consider the electromagnetic modes of a three--phase composite
131: consisting of an array of coated cylinders of infinite length, each
132: aligned parallel to the $z$ axis,
133: inserted into a matrix of dielectric constant $\varepsilon_m$ and
134: magnetic permeability $\mu_m$.
135: For each cylinder, the core and shell are characterized by the
136: radii $r_c$ and $r_s$, dielectric constants
137: $\varepsilon_c$ and $\varepsilon_s$, and magnetic permeabilities
138: $\mu_c$ and $\mu_s$, respectively.
139: We also assume that the electric and magnetic fields depend on
140: time through a factor $\exp{(- i \omega t)}$.
141: The core and shell of all cylinders, as well as the matrix are
142: homogeneous so that, the fields satisfy the Maxwell equations:
143: \begin{eqnarray}
144: \nabla \times {\bf H} = & - & i \omega \varepsilon({\bf r}) {\bf E} ~ , ~ ~
145: \nabla \cdot {\bf E} = 0 ~ ,
146: \label{a1} \\
147: \nabla \times {\bf E} = & & i \omega \mu({\bf r}) {\bf H} ~ , ~ ~
148: \nabla \cdot {\bf H} = 0 ~ ,
149: \label{a2}
150: \end{eqnarray}
151: with
152: \begin{equation}
153: \varepsilon({\bf r}) = \left\{
154: \begin{array}{lcl}
155: \varepsilon_m & ~~ {\rm for} ~~ & {\bf r} \in \Omega_m \, , \\
156: \varepsilon_s & ~~ {\rm for} ~~ & {\bf r} \in \Omega_s \, , \\
157: \varepsilon_c & ~~ {\rm for} ~~ & {\bf r} \in \Omega_c \, , \\
158: \end{array} \right.
159: \label{a3}
160: \end{equation}
161: and
162: \begin{equation}
163: \mu({\bf r}) = \left\{
164: \begin{array}{lcl}
165: \mu_m & ~~ {\rm for} ~~ & {\bf r} \in \Omega_m \, , \\
166: \mu_s & ~~ {\rm for} ~~ & {\bf r} \in \Omega_s \, , \\
167: \mu_c & ~~ {\rm for} ~~ & {\bf r} \in \Omega_c \, . \\
168: \end{array} \right.
169: \label{a4}
170: \end{equation}
171: Here, in the $xy-$plane, $\Omega_c$ denotes the union of all the
172: domains occupied by cylinder cores, while $\Omega_s$
173: denotes the union of all the domains occupied by the cylinder shells.
174: Thus, $\Omega_c \bigcup \Omega_s$ represents the region occupied by
175: all the cylinders in the array and $\Omega_m$ is the region between
176: the cylinders (the matrix region).
177:
178: By eliminating ${\bf E}$ or ${\bf H}$ between (\ref{a1}) and (\ref{a2}),
179: we obtain the Helmholtz equations for the components of the fields:
180: \begin{equation}
181: \left( \nabla^2 + \kappa^2 \right)
182: \left\{ \begin{array}{c} {\bf E} \\ {\bf H} \end{array} \right\} = 0
183: \, ,
184: \label{a5}
185: \end{equation}
186: where $\kappa^2=\omega^2 \varepsilon \mu$ is again a function of
187: position.
188: The axes of the cylinders in the array are parallel to the $z-$axis, and
189: the fields are taken to depend on $z$ through the factor $\exp(i \beta z)$,
190: with $\beta$ representing the propagation constant along the $z$
191: direction, in all the regions $\Omega_c$, $\Omega_s$ and $\Omega_m$.
192: Also, in cylindrical coordinates ($r$, $\theta$, $z$), the field
193: components $E_z$ and $H_z$ determine the other four field components
194: through the relations\cite{simon}:
195: \begin{eqnarray}
196: E_r & = &
197: \frac{i}{\kappa^2 - \beta^2}
198: \left[ \beta \, \frac{\partial E_z}{\partial r} +
199: \frac{\omega \mu}{r} \, \frac{\partial H_z}{\partial \theta}
200: \right]
201: % =
202: %\frac{i}{\kappa^2 - \beta^2}
203: %\left[ \beta \, \frac{\partial E_z}{\partial \nu} +
204: %\omega \mu \, \frac{\partial H_z}{\partial \tau}
205: %\right]
206: \, ,
207: \label{a7} \\
208: E_\theta & = &
209: \frac{i}{\kappa^2 - \beta^2}
210: \left[ \frac{\beta}{r} \, \frac{\partial E_z}{\partial \theta} -
211: \omega \mu \, \frac{\partial H_z}{\partial r} \right]
212: % =
213: %\frac{i}{\kappa^2 - \beta^2}
214: %\left[ \beta \, \frac{\partial E_z}{\partial \tau} -
215: %\omega \mu \, \frac{\partial H_z}{\partial \nu} \right]
216: \, ,
217: \label{a8} \\
218: H_r & = &
219: \frac{- i}{\kappa^2 - \beta^2}
220: \left[ \frac{\omega \varepsilon}{r} \,
221: \frac{\partial E_z}{\partial \theta} -
222: \beta \, \frac{\partial H_z}{\partial r}
223: \right]
224: %=
225: %\frac{- i}{\kappa^2 - \beta^2}
226: %\left[ \omega \varepsilon \,
227: %\frac{\partial E_z}{\partial \tau} -
228: %\beta \, \frac{\partial H_z}{\partial \nu}
229: %\right]
230: \, ,
231: \label{a9} \\
232: H_\theta & = &
233: \frac{i}{\kappa^2 - \beta^2}
234: \left[ \omega \varepsilon \, \frac{\partial E_z}{\partial r} +
235: \frac{\beta}{r} \, \frac{\partial H_z}{\partial \theta} \right]
236: % =
237: %\frac{i}{\kappa^2 - \beta^2}
238: %\left[ \omega \varepsilon \, \frac{\partial E_z}{\partial \nu} +
239: %\beta \, \frac{\partial H_z}{\partial \tau} \right]
240: \, .
241: \label{a10}
242: \end{eqnarray}
243: The field components $E_z$ and $H_z$ satisfy the Helmholtz equations:
244: \begin{equation}
245: \left( \nabla_\perp^2 + k^2 \right)
246: \left\{ \begin{array}{c} E_z \\ H_z \end{array} \right\} = 0 \, ,
247: \label{a11}
248: \end{equation}
249: where $\nabla_\perp^2$ is the transverse Laplacian and
250: \begin{equation}
251: k^2 = \kappa^2 - \beta^ 2 \, ,
252: \label{a12}
253: \end{equation}
254: represents the transverse (in-plane) propagation constant.
255:
256: For the cylinder centered at the origin
257: of coordinates, we represent the electric and magnetic fields
258: $E_z$ and $H_z$ (denoted here by $V$),
259: by series expansions in terms of cylindrical harmonics:
260: \begin{equation}
261: V(r, \theta) =
262: \left\{
263: \begin{array}{l}
264: {\displaystyle \sum_{l=-\infty}^\infty A_l^{c} J_l(k_c r)
265: \, e^{i l \theta}} \, ,
266: \\
267: % && \\
268: {\displaystyle \sum_{l=-\infty}^\infty
269: \left[ A_l^{s} J_l(k_s r) +
270: B_l^{s} Y_l(k_s r) \right] e^{i l \theta}}\, ,
271: \\
272: % && \\
273: {\displaystyle \sum_{l=-\infty}^\infty
274: \left[ A_l^{m} J_l(k_m r) +
275: B_l^{m} Y_l(k_m r) \right] e^{i l \theta}} \, ,
276: \end{array}
277: \right.
278: \label{fields}
279: \end{equation}
280: where $J$ and $Y$ represent the Bessel functions of the first
281: and second kind. The three forms of the series expansions
282: in (\ref{fields}) correspond to
283: $0 \leq r \leq r_c$ (i.e., inside the core),
284: $r_c \leq r \leq r_s$ (i.e., inside the shell) and
285: $r \geq r_s$ (i.e., in the matrix), respectively.
286: Also, the superscripts $c$, $s$ and $m$ label the fields inside the
287: cylinder core, cylinder shell, and in the matrix, respectively.
288: Thus, we have $k_c^2=\omega^2 \varepsilon_c \mu_c$,
289: $k_s^2=\omega^2 \varepsilon_s \mu_s$, and
290: $k_m^2=\omega^2 \varepsilon_m \mu_m$.
291:
292:
293: \subsection{Boundary Conditions}
294: \label{bcdsec}
295:
296: The boundary conditions express
297: the continuity of the tangential components of the electric
298: ($E_z$, $E_\theta$)
299: and magnetic ($H_z$ and $H_\theta$) fields across the core
300: and shell surfaces.
301: When the crystal momentum is perpendicular
302: to the axes of the cylinders in the array, we have $\beta=0$,
303: and the problem can be reduced to solving two independent
304: problems\cite{panofsky}:
305: (i) for TM or $s$ polarization (when $H_z=0$ and the transverse
306: parts of ${\bf E}$ and ${\bf H}$ are generated by $\nabla_\perp E_z$), and
307: (ii) for TE or $p$ polarization (when $E_z=0$ and
308: $\nabla_\perp H_z$ gives the transverse components of ${\bf E}$ and
309: ${\bf H}$).
310: Thus, depending on the polarization, we obtain:
311: \begin{eqnarray}
312: \left. E_z^{(c)} \right|_{r=r_c} & = & \left. E_z^{(s)}
313: \right|_{r=r_c},
314: \label{a13x} \\
315: \left. \frac{1}{\mu_c} \, \frac{\partial E_z^{(c)}}{\partial r}
316: \right|_{r=r_c}
317: & = &
318: \left. \frac{1}{\mu_s} \, \frac{\partial E_z^{(s)}}{\partial r}
319: \right|_{r=r_c} ,
320: \label{a14x} \\
321: \left. E_z^{(s)} \right|_{r=r_s} & = & \left. E_z^{(m)}
322: \right|_{r=r_s},
323: \label{a13} \\
324: \left. \frac{1}{\mu_s} \, \frac{\partial E_z^{(s)}}{\partial r}
325: \right|_{r=r_s}
326: & = &
327: \left. \frac{1}{\mu_m} \, \frac{\partial E_z^{(m)}}{\partial r}
328: \right|_{r=r_s} ,
329: \label{a14}
330: \end{eqnarray}
331: for TM polarization, and:
332: \begin{eqnarray}
333: \left. H_z^{(c)} \right|_{r=r_c} & = & \left. H_z^{(s)}
334: \right|_{r=r_c},
335: \label{a15x} \\
336: \left. \frac{1}{\varepsilon_c} \, \frac{\partial H_z^{(c)}}{\partial r}
337: \right|_{r=r_c}
338: & = &
339: \left. \frac{1}{\varepsilon_s} \, \frac{\partial H_z^{(s)}}{\partial r}
340: \right|_{r=r_c} ,
341: \label{a16x} \\
342: \left. H_z^{(s)} \right|_{r=r_s} & = & \left. H_z^{(m)}
343: \right|_{r=r_s},
344: \label{a15} \\
345: \left. \frac{1}{\varepsilon_s} \, \frac{\partial H_z^{(s)}}{\partial r}
346: \right|_{r=r_s}
347: & = &
348: \left. \frac{1}{\varepsilon_m} \, \frac{\partial H_z^{(m)}}{\partial r}
349: \right|_{r=r_s} ,
350: \label{a16}
351: \end{eqnarray}
352: for TE polarization.
353:
354: In the case of TM polarization, for the cylinder centered at the origin
355: of coordinates, we use the series expansions (\ref{fields}) for
356: $E_z$, and
357: by means of the boundary conditions (\ref{a13x}-\ref{a14})
358: we obtain relations of the form:
359: \begin{eqnarray}
360: \left[ \begin{array}{c} A_l^m \\ B_l^m \end{array} \right]
361: & = &
362: \left[ \begin{array}{rr}
363: J_l(k_m r_s) & Y_l(k_m r_s) \\
364: Z_m^{-1} J_l'(k_m r_s) & Z_m^{-1} Y_l'(k_m r_s) \end{array}
365: \right]^{-1}
366: \nonumber \\ & \times &
367: \left[ \begin{array}{rr}
368: J_l(k_s r_s) & Y_l(k_s r_s) \\
369: Z_s^{-1} J_l'(k_s r_s) & Z_s^{-1} Y_l'(k_s r_s) \end{array}
370: \right]
371: \label{TMbcd} \\ & \times &
372: \left[ \begin{array}{rr}
373: J_l(k_s r_c) & Y_l(k_s r_c) \\
374: Z_s^{-1} J_l'(k_s r_c) & Z_s^{-1} Y_l'(k_s r_c) \end{array}
375: \right]^{-1}
376: \nonumber \\ & \times &
377: \left[ \begin{array}{rr}
378: J_l(k_c r_c) & Y_l(k_c r_c) \\
379: Z_c^{-1} J_l'(k_c r_c) & Z_c^{-1} Y_l'(k_c r_c) \end{array}
380: \right]
381: \left[ \begin{array}{c} A_l^c \\ 0 \end{array} \right]
382: \, , \nonumber
383: \end{eqnarray}
384: or
385: \begin{equation}
386: A_l^m= - M_l B_l^m \, .
387: \label{a21}
388: \end{equation}
389: In (\ref{TMbcd}), the prime indicates the derivative of the
390: corresponding function and $Z_i=\sqrt{\mu_i/\varepsilon_i}$
391: ($i = m, s, c$) represent the impedances of the matrix, shell
392: and core, respectively.
393: For TE polarization we obtain the relation between $A_l^m$
394: and $B_l^m$ by changing $Z_i\rightarrow1/Z_i$ in (\ref{TMbcd}).
395:
396: The structure of (\ref{TMbcd}) allows us a generalization to
397: multicoated cylinders, and for this reason we have added the
398: second column in the last matrix. However, this last matrix is
399: multiplied by the vector $[A_l^c, 0]^T$ corresponding to the field
400: expansion inside the core (i.e., the first form in
401: (\ref{fields})).
402: Note that the first two matrices in the right side of
403: (\ref{TMbcd}) arise from the boundary conditions at the interface
404: between the shell and matrix ((\ref{a13}) and
405: (\ref{a14})), while the next two matrices come from the boundary
406: conditions at the core -- shell interface ((\ref{a13x}) and
407: (\ref{a14x})). Consequently, in the case of a multicoated cylinder,
408: we may infer the form of a general term representing the interface
409: $i$ (between the $i^{\rm th}$ and $(i+1)^{\rm th}$ media)
410: characterized by the radius $r_i$:
411: \begin{eqnarray}
412: {\bf M}_l^{i, i+1} & = &
413: \left[ \begin{array}{rr}
414: J_l(k_{i} r_i) & Y_l(k_{i} r_i) \\
415: Z_{i}^{-1} J_l'(k_{i} r_i) & Z_{i}^{-1} Y_l'(k_{i} r_i) \end{array}
416: \right]^{-1}
417: \nonumber \\ & \times &
418: \left[ \begin{array}{rr}
419: J_l(k_{i+1} r_i) & Y_l(k_{i+1} r_i) \\
420: Z_{i+1}^{-1} J_l'(k_{i+1} r_i) & Z_{i+1}^{-1} Y_l'(k_{i+1} r_i) \end{array}
421: \right] .
422: \label{bcdterm}
423: \end{eqnarray}
424: Here, we have counted the layers from the exterior to interior and
425: thus, for $n-1$ shells, the matrix is indexed by $i=0$ and the core
426: (the innermost medium) by $i=n$. Now, Eq.~(\ref{TMbcd}) takes the form
427: \begin{equation}
428: \left[ \begin{array}{c} A_l^0 \\ B_l^0 \end{array} \right]
429: =
430: \left( \prod_{i=0}^{n-1} {\bf M}_l^{i, i+1} \right)
431: \left[ \begin{array}{c} A_l^{n+1} \\ 0 \end{array} \right] \, .
432: \label{TMbcdgen}
433: \end{equation}
434: We also obtain the expression for the TE polarization by
435: substituting $Z_i\rightarrow1/Z_i$.
436:
437:
438:
439: \subsection{The Generalized Rayleigh Identity}
440: \label{GRIsec}
441:
442: For each polarization (TM or TE), we may write
443: a Generalized Rayleigh Identity of the form\cite{cnr,nrb1}:
444: \begin{equation}
445: M_l B_l
446: + \sum_{n=-\infty}^\infty
447: S_{l-n}^Y(k_m, {\bf k}_0) \, B_n = 0 \, ,
448: \label{gri}
449: \end{equation}
450: with $M_l$ from (\ref{a21}) and $-\infty < l < \infty$.
451: The quantities $S_l^Y$ are dynamic lattice sums, which may be evaluated
452: using accelerated summation over the reciprocal array\cite{cnr}:
453: \begin{eqnarray}
454: && S_l^Y(k_m, {\bf k}_0) J_{l+n}(k_m \xi) =
455: \nonumber \\
456: && -
457: \left[ Y_n(k_m \xi) + \frac{1}{\pi} \sum_{q=1}^n
458: \frac{(n-q)!}{(q-1)!} \left( \frac{2}{k_m \xi} \right)^{n - 2 q +2} \right]
459: \delta_{l,0} \nonumber \\
460: && - i^l \frac{4}{A} \sum_{h} \left( \frac{k_m}{Q_h} \right)^n
461: \frac{J_{l+n}(Q_h \xi)}{Q_h^2 - k_m^2}
462: e^{i l \theta_h} .
463: \label{rls}
464: \end{eqnarray}
465: Here, $n$ is an arbitrary non--negative integer, $\xi$ is the length of
466: an arbitrary vector inside the unit cell (shorter than the shortest
467: line connecting two of its vertices), and
468: $A$ represents the area of the unit cell. The vector ${\bf k}_0$
469: is the in--plane component of the ``crystal momentum''
470: ${\bf k}_B = {\bf k}_0 + \beta {\bf {\widehat{z}}}$,
471: with ${\bf {\widehat{z}}}$ the unit vector along the $z-$axis.
472: Note that, in this paper we consider only the case $\beta=0$,
473: so that ${\bf k}_B = {\bf k}_0$.
474:
475: The quantities ${\bf Q}_h$ and $\theta_h$ from (\ref{rls}) depend
476: on the type of the array. In our case, the hexagonal array is
477: defined by the fundamental translation vectors ${\bf e}_1=d(1, 0)$
478: and ${\bf e}_2=d(1/2, \sqrt{3}/2)$, where $d$ is the array
479: constant (see Fig.~\ref{WSC}). Hence, the centers of the cylinders
480: are located at the points given by the radius vectors
481: \begin{equation}
482: {\bf R}_p = p_1 {\bf e}_1 + p_2 {\bf e}_2 ,
483: \end{equation}
484: with $p$ denoting the pair of integers
485: $(p_1, p_2)\in{\mathbb Z}^2$.
486: The fundamental vectors of the reciprocal array are given by the
487: relations
488: \begin{eqnarray}
489: {\bf u}_1 & = &
490: \frac{2\pi}{A}{\bf e}_2 \times {\bf e}_3
491: = \frac{2\pi}{d} \left( 1, - \frac{1}{\sqrt{3}} \right) ,
492: \label{u1} \\
493: {\bf u}_2 & = &
494: \frac{2\pi}{A} {\bf e}_3 \times {\bf e}_1
495: = \frac{2\pi}{d} \left( 0, \frac{2}{\sqrt{3}} \right) ,
496: \label{u2}
497: \end{eqnarray}
498: where $A=\|{\bf e}_1 \times {\bf e}_2\|$
499: and
500: ${\bf e}_3=({\bf e}_1 \times {\bf e}_2)/A$
501: is a dimensionless unit vector, along the $z$ axis.
502: With the vectors (\ref{u1}) and (\ref{u2}) we form the
503: reciprocal lattice vectors
504: \begin{equation}
505: {\bf K}_h = h_1 {\bf u}_1 + h_2 {\bf u}_2 ,
506: \end{equation}
507: where $h=(h_1,h_2)\in{\mathbb Z}^2$.
508: Finally, we have
509: ${\bf Q}_h = {\bf k}_0 + {\bf K}_h$
510: and $\theta_h = \arg{({\bf Q_h})}$.
511:
512: The eigenfrequencies ($\omega$) of the Helmholtz
513: equation (\ref{a5}) are obtained from
514: the zeros of the determinant of (\ref{gri}).
515: In the coordinate system $\omega$ versus ${\bf k}_B$,
516: the photonic band diagrams are given by the trajectories of the
517: eigenfrequencies $\omega$
518: when ${\bf k}_B$ follows the boundary of one of the
519: irreducible regions in the first Brillouin zone
520: (see Fig.~\ref{IBZ}).
521:
522: In the case of metallic media, ${\bf k}_B$
523: becomes complex. It is then simpler to use a
524: summation method \cite{jmpls} different from (\ref{rls}),
525: and a technique, based on Bloch's theorem,
526: developed originally by McRae\cite{mcrae} in low energy electron
527: diffraction, and applied recently to photonic crystals by
528: Gralak {\em et al.} \cite{gralak}.
529: We consider the array as an infinite
530: stack of gratings and calculate the scattering
531: matrices for a single grating in the array \cite{josa12}.
532: Then, the complex
533: crystal momentum ${\bf k}_B$ is found by solving an eigenvalue problem
534: \cite{josa3}.
535:
536:
537: \section{The Long Wavelength Limit}
538: \label{longwavelentgh}
539:
540: \subsection{Partial Resonances}
541: \label{prlwl}
542:
543: The solution of the problem of electromagnetic scattering by
544: a solid cylinder, at normal incidence and in the long wavelength
545: limit\cite{ray2}, revealed the occurrence of
546: a resonance in scattering when the relative dielectric
547: constant of the cylinder was -1. Later, the existence of this resonance
548: was proved experimentally in studies of radio reflection by meteor
549: trails\cite{herlofson}, and of plasma columns\cite{tonks,romell}.
550: Also, Wait\cite{wait1,wait2} has shown that in the long wavelength
551: limit the scattering resonances given by the formula
552: \begin{equation}
553: \left( \mu_1 + \mu_2 \right)
554: \left( \varepsilon_1 + \varepsilon_2 \right) = 0,
555: \label{wait}
556: \end{equation}
557: where the label 1 refers to the cylinder and 2 refers to the
558: homogeneous medium outside the cylinder, are independent of the
559: polarization and angle of incidence.
560: Note that, in electrostatics the condition
561: $(\varepsilon_1 + \varepsilon_2)=0$ defines the accumulation point
562: for the poles and zeros of the effective dielectric constant of a
563: two--phase composite\cite{bergman,ross},
564: with $(\mu_1 + \mu_2)=0$ corresponding to magnetostatics.
565:
566: We applied the method developed by Wait\cite{wait1} to solve
567: the problem of scattering of a plane wave by a coated cylinder at
568: oblique incidence. In this case, in the long wavelength limit we
569: obtained the resonance condition
570: \begin{equation}
571: \left( \mu_c + \mu_s \right)
572: \left( \mu_s + \mu_m \right)
573: \left( \varepsilon_c + \varepsilon_s \right)
574: \left( \varepsilon_s + \varepsilon_m \right)
575: = 0,
576: \label{us}
577: \end{equation}
578: which describes precisely the electrostatic partial resonances of
579: a three--phase composite\cite{cc1,cc2,cc3}, plus their magnetostatic
580: counterpart.
581: The resonance condition (\ref{us}) also suggests a generalization
582: to a $n$--phase composite, of the form
583: \begin{equation}
584: \prod_{i=0}^{n-1}
585: \left( \mu_i + \mu_{i+1} \right)
586: \left( \varepsilon_i + \varepsilon_{i+1} \right) = 0,
587: \end{equation}
588: for the case of a multi-coated cylinder.
589: Here, $\mu_{n}$ and $\varepsilon_{n}$ denote the magnetic permeability
590: and dielectric constant of the innermost phase (core), while
591: $\mu_{0}$ and $\varepsilon_{0}$ represent the magnetic permeability
592: and dielectric constant of the medium outside the composite.
593:
594:
595: \subsection{TE Polarization}
596: \label{tepr}
597:
598: We concentrate now on non-magnetic materials for which
599: $\mu_m=\mu_s=\mu_c=\mu_0$, so that
600: $\varepsilon_m=n_m^2 \varepsilon_0$,
601: $\varepsilon_s=n_s^2 \varepsilon_0$ and
602: $\varepsilon_c=n_c^2 \varepsilon_0$, where $\varepsilon_0$
603: is the dielectric constant of free space, and $n_i$
604: ($i = m, s, c$) represent the refractive indexes of the matrix, shell
605: and core, respectively.
606: In the long wavelength limit ($k_m\rightarrow 0$),
607: we use the expressions of the
608: Bessel functions for small arguments\cite{AS} in
609: the boundary conditions coefficients $M_l$ from (\ref{a21}),
610: for TE polarization.
611: For $l\neq0$ we obtain
612: \begin{equation}
613: M_l = M_{-l} =
614: - \frac{1}{\pi} \left( \frac{2}{k_m r_s} \right)^{2 l}
615: l! (l-1)! \, \gamma_l \, ,
616: \label{eq6ad}
617: \end{equation}
618: where
619: \begin{equation}
620: \gamma_l =
621: \frac{r_c^{2l} (\varepsilon_s - \varepsilon_c)
622: (\varepsilon_m - \varepsilon_s) + r_s^{2l}
623: (\varepsilon_s + \varepsilon_c)
624: (\varepsilon_m + \varepsilon_s)}
625: {r_c^{2l} (\varepsilon_s - \varepsilon_c)
626: (\varepsilon_m + \varepsilon_s) + r_s^{2l}
627: (\varepsilon_s + \varepsilon_c)
628: (\varepsilon_m - \varepsilon_s)} \, ,
629: \label{eq6a}
630: \end{equation}
631: while for $l=0$, we obtain a completely different form
632: \begin{equation}
633: M_0 = - \frac{1}{\pi} \left( \frac{2}{k_m r_s} \right)^4
634: \frac{\varepsilon_m}{(\varepsilon_s-\varepsilon_c)
635: (r_c/r_s)^4 + (\varepsilon_m-\varepsilon_s)} \, .
636: \label{tem0}
637: \end{equation}
638:
639: To relate the long wavelength limit of the dynamic problem with
640: the corresponding problem in electrostatics we apply the same
641: method as in Ref.~\cite{jewatem}. Thus, the boundary conditions
642: (\ref{a15x}-\ref{a16}) correspond to an electrostatic problem in which
643: the inverse of the dielectric constants
644: ($\varepsilon_c\rightarrow 1/\varepsilon_c$,
645: $\varepsilon_s\rightarrow 1/\varepsilon_s$ and
646: $\varepsilon_m\rightarrow 1/\varepsilon_m$) have to be considered.
647: This will also change
648: $\gamma_l\rightarrow -\gamma_l$. Now, the boundary conditions
649: (\ref{a21}) for $l\neq0$ can be written in the form
650: \begin{equation}
651: A_l^m = - \frac{1}{\pi} \left( \frac{2}{k_m} \right)^{2 l}
652: l! (l-1)! \, \gamma_l \,
653: \frac{B_l^m}{r_s^{2l}} .
654: \label{tepr1}
655: \end{equation}
656:
657: In electrostatics, the corresponding relationship between the
658: coefficients $A_l$ and
659: $B_l$ which controls the response of a coated cylinder to an
660: external field, has the form\cite{cc1}
661: \begin{equation}
662: A_l' = \gamma_l \, \frac{B_l'}{r_s^{2l}} \, .
663: \label{statics}
664: \end{equation}
665: Note that, due to the symmetry properties of the electrostatic
666: potential, $l$ cannot
667: take even values in the case of a square array, or values which are
668: multiples of 3 in the case of a hexagonal array\cite{perrins}.
669: By comparing (\ref{tepr1}) with (\ref{statics}) we may
670: deduce the relation between static and dynamic multipole coefficients
671: \begin{equation}
672: B_l' = - \frac{1}{\pi} \left( \frac{2}{k_m} \right)^{2 l}
673: l! (l-1)! \, B_l^m \, .
674: \label{tepr2}
675: \end{equation}
676: for $l\neq0$.
677: The dynamic field in the matrix ($H_z^{(m)}$) may be approximated
678: by\cite{jewatem,siam}
679: \begin{equation}
680: H_z^{(m)}({\bf r}) = A_0 \exp{[ i k_B V_m({\bf r}) ]}
681: + {\mathcal O}(k_B^2) \, ,
682: \label{tepr3}
683: \end{equation}
684: where $V_m$ is the solution of the static problem.
685: Note that the derivation of (\ref{tepr3}) does not involve the
686: boundary conditions coefficient $M_0$ from (\ref{tem0}).
687:
688: In electrostatics, the partial resonances of a three--phase
689: composite consisting of coated cylinders are defined by the
690: equations
691: \begin{eqnarray}
692: \varepsilon_c + \varepsilon_s & = & 0 \qquad ({\rm core--shell~
693: resonance}),
694: \label{cspr} \\
695: \varepsilon_s + \varepsilon_m & = & 0 \qquad ({\rm shell--matrix~
696: resonance}),
697: \label{smpr}
698: \end{eqnarray}
699: when (\ref{statics}) becomes
700: \begin{equation}
701: A_l' = \frac{\varepsilon_m + \varepsilon_c}{\varepsilon_m -
702: \varepsilon_c} \, \frac{B_l'}{r_s^{2l}} \, ,
703: \label{cspres}
704: \end{equation}
705: or
706: \begin{equation}
707: A_l' = \frac{\varepsilon_m + \varepsilon_c}{\varepsilon_m -
708: \varepsilon_c} \, \frac{B_l'}{(r_s^2/r_c)^{2l}} \, ,
709: \label{smpres}
710: \end{equation}
711: respectively.
712: In the first case, the field inside the coated cylinder is exactly
713: the same as would be found within a solid cylinder of radius $r_s$
714: and dielectric constant $\varepsilon_c$, while the potential
715: outside the coated cylinder, in the matrix, is precisely the same
716: as that outside the solid cylinder\cite{cc1}.
717: The second case corresponds to an effective dielectric constant of
718: the array of coated cylinders, identical with that of an array of
719: solid cylinders of radius $r_s^2/r_c > r_s$ (the geometrical image of
720: the core boundary with respect to the shell outer boundary), and
721: dielectric constant $\varepsilon_c$. Now, the field external to the coated
722: cylinder {\em and beyond the radius} $r_s^2/r_c$ is the same as that
723: external to the solid cylinder\cite{cc1}.
724:
725: Since it is the relationship between the coefficients $A_l$ and
726: $B_l$ which controls the response of a coated cylinder to an
727: external field,
728: equations (\ref{statics}) and (\ref{tepr1}) show that this
729: response is determined by $\gamma_l$ in electrostatics as well
730: as in the long wavelengths limit of electrodynamics.
731: The limiting process is smooth and,
732: therefore, we expect a resonant behaviour of the array of
733: coated cylinders, for nonzero frequencies, when one of the
734: conditions (\ref{cspr}) or (\ref{smpr}) is satisfied.
735:
736:
737:
738: \subsection{TM Polarization}
739: \label{tmpr}
740:
741: The behaviour of the array of coated cylinders is completely
742: different in the case of TM polarization.
743: Now, in the long wavelength limit,
744: the boundary conditions coefficients $M_l$ from (\ref{a21})
745: take the form
746: \begin{eqnarray}
747: && M_l = M_{-l} \nonumber \\
748: && =
749: \frac{1}{\pi} \left( \frac{2}{k_m r_s} \right)^{2 l + 2}
750: \frac{2 \, l! (l+1)! \, \varepsilon_m}{(\varepsilon_c - \varepsilon_s)
751: (r_c/r_s)^{2l+2} + (\varepsilon_s - \varepsilon_m)} ,
752: \label{tmpr1}
753: \end{eqnarray}
754: for $l\neq0$, and
755: \begin{equation}
756: M_0 =
757: \frac{1}{\pi} \left( \frac{2}{k_m r_s} \right)^{2}
758: \frac{\varepsilon_m}{(\varepsilon_c - \varepsilon_s)
759: (r_c/r_s)^{2} + (\varepsilon_s - \varepsilon_m)} \, .
760: \label{tmpr2}
761: \end{equation}
762: Following the same method as in Ref.\cite{jewatem} we consider the
763: central equation ($l=0$) of the Rayleigh identity (\ref{gri})
764: \begin{equation}
765: \left[ M_0 + S_0^Y(k_m, {\bf k}_0) \right] B_0 + \sum_{n\neq0}
766: S_{-n}^Y(k_m, {\bf k}_0) B_n = 0 \, ,
767: \label{gri0}
768: \end{equation}
769: and employ the forms of lattice sums for small
770: $k_m$\cite{jewatem}
771: \begin{eqnarray}
772: S_0^Y & \sim & \frac{4}{A k_0^2 (\alpha^2 -1)} \, , \\
773: S_1^Y & \sim & \frac{4 i}{A k_0^2 \alpha (\alpha^2 -1)}
774: \, e^{i \theta_0} \, , \\
775: S_n^Y & \sim & {\mathcal O}(k_0^{-n}) \, , \quad {\rm for}~l\geq2
776: \, , \quad
777: \left( S_{-n}^Y = \overline{S_n^Y} \right) ,
778: \end{eqnarray}
779: where the superposed bar denotes complex conjugation,
780: $A$ is the area of the Wigner-Seitz cell,
781: $\theta_0=\arg{({\bf k}_0)}$ and $\alpha=k_m/k_0$.
782: We also introduce the relations between dynamic ($B_n$) and static
783: ($B_n'$) multipole coefficients\cite{jewatem}
784: \begin{eqnarray}
785: B_n & \sim & (k_m r_s)^{n+2} \, B_n' \, , \\
786: B_{-n} & \sim & (k_m r_s)^{n+2} \, \overline{B_n'} \, ,
787: \end{eqnarray}
788: for $n\geq0$. Here, the static multipole
789: coefficients are, to leading order, independent of $k_m$.
790: Finally, the equation (\ref{gri0}) becomes
791: \begin{eqnarray}
792: \left[
793: \frac{\varepsilon_m}{(\varepsilon_c - \varepsilon_s)
794: (r_c/r_s)^{2} + (\varepsilon_s - \varepsilon_m)}
795: - \frac{\pi \alpha^2 r_s^2}{A (1 - \alpha^2)} \right] B_0'
796: \nonumber \\
797: + \sum_{n\neq0} B_n' \, {\mathcal O}(k_m) = 0 .
798: \label{tmpr3}
799: \end{eqnarray}
800: From all the other equations ($l\neq0$) of (\ref{gri}) we obtain
801: $B_l'=0$ so that, in the long wavelengths limit,
802: to leading order, the only non-trivial solution
803: of the Rayleigh identity (\ref{gri}) is
804: $B_0'\neq0$ and $B_n'=0$ ($\forall n\neq0$), if
805: \begin{equation}
806: \frac{\varepsilon_m}{\alpha^2} =
807: f_c \varepsilon_c + f_s \varepsilon_s + f_m \varepsilon_m \, ,
808: \label{tmpr4}
809: \end{equation}
810: with
811: $f_c=\pi r_c^2/A$, $f_s=\pi(r_s^2-r_c^2)/A$ and $f_m=1-\pi
812: r_s^2/A$,
813: denoting the area fractions of the phases inside the unit cell.
814: From (\ref{fields}) we derive the form of the electric
815: field in the matrix:
816: \begin{equation}
817: E_z^{(m)} \sim \frac{\varepsilon_m}{(\varepsilon_c - \varepsilon_s)
818: (r_c/r_s)^{2} + (\varepsilon_s - \varepsilon_m)} B_0' \, .
819: \label{tmpr5}
820: \end{equation}
821:
822: Equation (\ref{tmpr4}) represents the linear mixing formula
823: \cite{wiener}
824: for the effective dielectric constant of an array of coated
825: cylinders subject to a uniform field applied along the cylinder
826: axis. Note that
827: there are no terms of the form $\varepsilon_c + \varepsilon_s$ or
828: $\varepsilon_s + \varepsilon_m$ in (\ref{tmpr3}) or (\ref{tmpr4})
829: to indicate a core--shell or shell--matrix partial resonance.
830: Again, the limiting process is smooth and,
831: consequently, we do not expect a resonant behaviour of the array of
832: coated cylinders, for any frequency, in the case of TM polarization.
833:
834: \section{Numerical Results}
835: \label{numerics}
836:
837: In this section we discuss the behaviour of a three--phase
838: composite consisting of aluminium oxide (\alo), whose refractive index
839: depends on the frequency, and two pure dielectric phases:
840: one with a refractive index $n_d=1.37$, and vacuum ($n_0=1$).
841:
842: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
843: \begin{figure}[h]
844: \vspace{1ex}
845: \centerline{\psfig{file=IBZ.eps,height=1.5in}}
846: \vspace{3ex}
847: \caption{ The irreducible part (green triangle) of the first
848: Brillouin zone.} \label{IBZ}
849: \end{figure}
850: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
851:
852: Note that, all the defining equations for partial resonances
853: and the boundary conditions coefficients, in the long wavelengths
854: limit or in electrostatics, are homogeneous functions of
855: dielectric constants.
856: Also, note that the phases are non--magnetic materials.
857: Hence, in this section we will denote by
858: $\varepsilon$ the relative dielectric constant
859: $\varepsilon/\varepsilon_0=n^2$ (the square of the refractive index).
860: With this notation, we use
861: the experimental data for the real and imaginary parts
862: of the refractive index of aluminium oxide in the region
863: $0.3\mu{\rm m}\le\lambda\le20.0\mu{\rm m}$\cite{palik},
864: to evaluate the relative dielectric constant of aluminium oxide
865: ($\varepsilon_{\alo}=n_{\alo}^2$).
866: Also, the relative dielectric constant of the second phase is
867: $\varepsilon_d=1.876$.
868:
869: We assume that
870: the composite is structured as a hexagonal array
871: (array constant $d$=1$\mu$m) of coated
872: cylinders (core radius $r_c=0.03\mu$m,
873: cylinder radius (core+coating) $r_s=0.31\mu$m)
874: in vacuum.
875: The Wigner--Seitz cell is shown in Fig.~\ref{WSC}, while
876: the irreducible part of the first Brillouin zone (K$\Gamma$MK)
877: is shown Fig.~\ref{IBZ}.
878: For this region of the first Brillouin zone, the coordinates of the
879: corners are
880: $${\rm K} \left( \frac{2\pi}{3 d}, \frac{2\pi}{d \sqrt{3}} \right),
881: \quad \Gamma (0, 0) ,
882: \quad {\rm M} \left( 0, \frac{2\pi}{d \sqrt{3}} \right).$$
883:
884:
885: \subsection{Partial Resonances}
886: \label{pr}
887:
888: In the case of an absorbing (metallic--type) core and
889: a dielectric shell,
890: the constant, the positive value of the
891: shell dielectric constant
892: ($\varepsilon_s=n_d^2$)
893: makes impossible a relation of the type
894: $n_d^2+n_m^2=0$ corresponding to a shell--matrix
895: partial resonance.
896: Hence, we expect only core--shell partial resonances, which are
897: defined by the equation
898: \begin{equation}
899: n_{\alo}^2 + n_d^2 =0 \, .
900: \label{cs01}
901: \end{equation}
902:
903: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
904: \begin{figure}[h]
905: \centerline{\psfig{file=al2o3eps-cs.eps,width=3.3in}}
906: \vspace{2ex}
907: \caption{Real (solid curve) and imaginary (dashed curve) part of
908: $\alo$ relative dielectric constant versus wavelength. The
909: horizontal red line shows $-n_d^2$, while the vertical magenta
910: lines mark the two solutions $\lambda_1$ and $\lambda_2$, of
911: (\protect\ref{realcs01}).} \label{al2o3eps}
912: \end{figure}
913: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
914:
915: Note that, $n_d^2$ is real and so we will solve
916: the real part of (\ref{cs01}):
917: \begin{equation}
918: \Re e \left( n_{\alo}^2 + n_d^2 \right) =0 \, ,
919: \label{realcs01}
920: \end{equation}
921: and then check the magnitude of the imaginary part.
922:
923: In the wavelengths domain $0.3\mu{\rm m}\le\lambda\le20.0\mu{\rm
924: m}$, Eq.~(\ref{realcs01}) has two solutions
925: \begin{equation}
926: \lambda_1 \approx 12.861\mu{\rm m},
927: %\qquad
928: %\omega_1 d/(2 \pi c) \approx 0.0778,
929: \label{pr01}
930: \end{equation}
931: giving a value of
932: \begin{equation}
933: n_{\alo}^2 + n_d^2 \approx 10^{-5} + 0.329\, i \, ,
934: \label{pr02}
935: \end{equation}
936: when substituted in (\ref{cs01}), and
937: \begin{equation}
938: \lambda_2 \approx 17.333\mu{\rm m},
939: %\qquad
940: %\omega_2 d/(2 \pi c) \approx 0.0577,
941: \label{pr03}
942: \end{equation}
943: giving a value of
944: \begin{equation}
945: n_{\alo}^2 + n_d^2 \approx 10^{-4} + 32.72\, i \, ,
946: \label{pr04}
947: \end{equation}
948: when substituted in (\ref{cs01}). They are marked by the magenta
949: vertical lines in Fig.~\ref{al2o3eps}. Actually, the magnitude of
950: the imaginary part of (\ref{cs01}) shows that we may have a
951: partial core--shell resonance at $\lambda_1$, while the second
952: root appears only as the result of an anomaly of $n_{\alo}^2$
953: around $\lambda_2$.
954:
955: In the case of dielectric core and absorbing (metallic--type)
956: shell we expect the coated cylinders to exhibit both types of
957: partial resonances: core--shell $$n_d^2+n_{\alo}^2=0$$ and
958: shell--matrix $$n_{\alo}^2 + n_m^2=0.$$
959: The wavelengths at which the first type of partial resonances
960: occurs are given by Eq.~(\ref{realcs01}), while
961: the second type, is identical with the classical
962: resonance\cite{bergman,ross} exhibited by solid cylinders
963: from $\alo$ in vacuum.
964:
965: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
966: \begin{figure}[h]
967: \vspace{2ex}
968: \centerline{\psfig{file=al2o3eps-ms.eps,width=3.3in}}
969: \vspace{3ex}
970: \caption{Real (solid curve) and imaginary (dashed curve) part of
971: $\alo$ relative dielectric constant versus wavelength. The
972: horizontal red line shows $-n_m^2=-1$, while the vertical magenta
973: lines mark the two solutions $\widetilde{\lambda}_1$,
974: $\widetilde{\lambda}_2$, of (\protect\ref{realsm01}).}
975: \label{al2o3eps1}
976: \end{figure}
977: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
978:
979: Thus, for the shell--matrix resonances, in the range
980: $0.3\mu{\rm m}\le\lambda\le20.0\mu{\rm m}$,
981: the equation
982: \begin{equation}
983: \Re e \left( n_{\alo}^2 + n_m^2 \right) =0 \, ,
984: \label{realsm01}
985: \end{equation}
986: has two solutions (see Fig.~\ref{al2o3eps1}):
987: \begin{equation}
988: \widetilde{\lambda}_1 \approx 12.151\mu{\rm m},
989: %\qquad
990: %\widetilde{\omega}_1 d/(2 \pi c) \approx 0.0823,
991: \label{pr05}
992: \end{equation}
993: giving a value of
994: \begin{equation}
995: n_{\alo}^2 + n_m^2 \approx 10^{-15} + 0.234\, i \, ,
996: \label{pr06}
997: \end{equation}
998: and
999: \begin{equation}
1000: \widetilde{\lambda}_2 \approx 17.344\mu{\rm m},
1001: %\qquad
1002: %\widetilde{\omega}_2 d/(2 \pi c) \approx 0.0577,
1003: \label{pr07}
1004: \end{equation}
1005: giving a value of
1006: \begin{equation}
1007: n_{Al_2O_3}^2 + n_m^2 \approx 10^{-13} + 32.611\, i \, .
1008: \label{pr08}
1009: \end{equation}
1010: These wavelengths are marked by magenta lines in
1011: Fig.~\ref{al2o3eps1}. Again, the magnitude of the imaginary part
1012: of (\ref{pr06}) and (\ref{pr08}) shows that we may have a partial
1013: core--shell resonance at $\widetilde{\lambda}_1$, while the second
1014: root appears only as the result of the anomaly of $n_{\alo}^2$
1015: around $\widetilde{\lambda}_2$.
1016:
1017: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1018: \begin{figure}[h]
1019: \vspace{3ex}
1020: \centerline{\psfig{file=RTA-TE-cc-co.eps,width=3.3in}}
1021: \vspace{3ex}
1022: \caption{TE polarization: Reflectance (red),
1023: transmittance (green) and absorptance (blue) for a stack of solid
1024: cylinders of refractive index $n_c=n_{\alo}$ and radius
1025: $r_c=0.03\mu$m (dashed curves), and for a stack of coated
1026: cylinders (solid curves). In both cases the stack consists of 30
1027: gratings of cylinders, and represents a ``slice'' from a
1028: 2-dimensional hexagonal array. The vertical magenta lines mark the
1029: two solutions $\lambda_1$ and $\lambda_2$, of
1030: (\protect\ref{realcs01}).}
1031: \label{rtateccco}
1032: \end{figure}
1033: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1034:
1035:
1036: \subsection{Absorbing Core and Dielectric Shell}
1037: \label{mcdsresults}
1038:
1039: Here, we consider that the coated cylinders have
1040: an aluminium oxide core ($n_c=n_{\alo}$),
1041: a dielectric coating ($n_s=n_d$), and are embedded in vacuum ($n_m=1$).
1042: The coating is a lossless, dielectric material of
1043: constant (i.e., independent of frequency)
1044: refractive index $n_s$=1.37.
1045: This is essentially the same as the refractive index
1046: of MgF$_2$ for
1047: $1\mu{\rm m} \lesssim \lambda \lesssim 2.5\mu{\rm m}$\cite{palik,duncanson}.
1048:
1049:
1050:
1051: \subsubsection{TE Polarization}
1052: \label{teres}
1053:
1054: In Fig.~\ref{rtateccco} we show
1055: a comparison between two types of stacks of gratings:
1056: a stack of 30 gratings of solid cylinders having a refractive index
1057: $n_c=n_{\alo}$ and radius $r_c=0.03\mu{\rm m}$ (dashed curves),
1058: and for a stack of 30 gratings of coated cylinders defined
1059: at the beginning of Sec.~\ref{mcdsresults} (solid curves).
1060: In both cases, the period of the gratings is $d=1\mu{\rm m}$.
1061: Actually, we compare the stack of coated cylinders with a stack
1062: in which the cylinder coatings have been removed (but keeping al the
1063: other parameters unchanged), to show the
1064: influence of the shell on the optical properties of the stack
1065: (reflectance, transmittance and absorptance), for normal incidence
1066: and TE polarization.
1067:
1068: We can see in Fig.~\ref{rtateccco} an enhancement in absorptance
1069: for the stack of coated cylinders at $\lambda_1$, while the stack
1070: of solid cylinders exhibits an anomalous absorptance at $\lambda_c
1071: \approx 12.151\mu{\rm m}<\lambda_1$, which corresponds to the
1072: classical resonance of a two--phase composite
1073: $n_{\alo}^2+n_m^2=0$\cite{bergman,ross} (in fact, the wavelength
1074: for this classical resonance, $\lambda_c=\widetilde{\lambda}_1$,
1075: is given by (\ref{realsm01})) Thus, in the case of TE polarization
1076: there is a partial resonance of the type core--shell at $\lambda_1
1077: = 12.861\mu{\rm m}$. At this wavelength, the anomalous increase of
1078: the absorptance of coated cylinders shows a lossy (or metallic)
1079: type behaviour, which is a characteristic of aluminium oxide and
1080: is never exhibited by the dielectric shell (lossless media). Note
1081: that the core radius is ten times smaller than the shell radius.
1082: We can also see in Fig.~\ref{rtatccecdetail} that, at $\lambda_1$
1083: the absorptance of the stack of coated cylinders (blue solid
1084: curve) is approximatively equal with the absorptance of a stack of
1085: solid cylinders, made from Al$_2$O$_3$, and having the radius
1086: $r_s=0.31\mu$m (blue dashed curve).
1087:
1088: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1089: \begin{figure}[h]
1090: \centerline{\psfig{file=RTA-TE-cc-ec-detail.eps,width=3.3in}}
1091: \vspace{3ex}
1092: \caption{TE polarization: Reflectance (red),
1093: transmittance (green) and absorptance (blue) for a stack of solid
1094: cylinders of refractive index $n_c=n_{\alo}$ and radius
1095: $r_s=0.31\mu$m (dashed curves), and for a stack of coated
1096: cylinders (solid curves). In both cases the stack consists of 30
1097: parallel gratings of cylinders, and represents a ``slice'' from a
1098: 2-dimensional hexagonal array. The vertical magenta lines mark the
1099: two solutions $\lambda_1$ and $\lambda_2$, of
1100: (\protect\ref{realcs01}).}
1101: \label{rtatccecdetail}
1102: \end{figure}
1103: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1104:
1105: Note the oscillations of the absorptance and transmittance, for
1106: the stack of solid cylinders of radius $r_s$ and refractive index
1107: $n_c$ (Fig.~\ref{rtatccecdetail}). In contrast with this
1108: behaviour, the absorptance and the transmittance of the stack of
1109: coated cylinders are almost constant, except a small region around
1110: $\lambda_1$. If we label by $cc$ the reflectance ($R_{cc}$),
1111: transmittance ($T_{cc}$) and absorptance ($A_{cc}$) for the stack
1112: of coated cylinders, and by $sc$ the corresponding quantities
1113: ($R_{sc}$, $T_{sc}$, $A_{sc}$) for the stack of solid cylinders,
1114: then, in the neighbourhood of the core--shell partial resonance,
1115: we have:
1116: \begin{equation}
1117: R_{cc} \approx T_{sc} \approx 0, \quad
1118: T_{cc} \approx R_{sc}, \quad
1119: A_{cc} \approx A_{sc} .
1120: \label{scRTA}
1121: \end{equation}
1122: Equation (\ref{scRTA}) shows that the {\em equivalence} between the
1123: two types of stacks is actually a sort of {\em duality}in the sense that
1124: they have similar absorptance, but the resonant system resembles a
1125: material which absorbs and transmits the incident radiation, while
1126: the stack of \alo cylinders absorbs and reflects incident
1127: radiation.
1128:
1129: In conclusion, at $\lambda_1$, for normal incidence and TE
1130: polarization, the stack of coated cylinders with a very small
1131: metallic--type core, has really the same absorptance as a stack of
1132: solid metallic cylinders of an enlarged radius $r_s=0.31\mu$m. It
1133: looks like the core and the shell {\em are replaced} by a new
1134: material which has the same absorptance as the core material, an
1135: effect identical with the core--shell partial resonance from
1136: electrostatics\cite{cc1,cc2,cc3}.
1137:
1138: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1139: \begin{figure}[h]
1140: \vspace{1ex}
1141: \centerline{\psfig{file=RTA-TM-cc-co.eps,width=3.3in}}
1142: \vspace{3ex}
1143: \caption{TM polarization: Reflectance (red),
1144: transmittance (green) and absorptance (blue) for a stack of solid
1145: cylinders of refractive index $n_c=n_{\alo}$ and radius $r_c$
1146: (dashed curves), and for a stack of coated cylinders (solid
1147: curves). In both cases the stack consists of 30 gratings of
1148: cylinders, and represents a ``slice'' from a 2-dimensional
1149: hexagonal array. The vertical magenta lines mark the two solutions
1150: $\lambda_1$ and $\lambda_2$, of (\protect\ref{realcs01}).}
1151: \label{rtatmccco}
1152: \end{figure}
1153: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1154:
1155: \subsubsection{TM Polarization}
1156: \label{tmres}
1157:
1158: There is no partial resonance for a stack of coated cylinders, for
1159: TM polarization.
1160: Fig.~\ref{rtatmccco} shows no change of optical characteristics
1161: around $\lambda_1$. The second solution of
1162: (\ref{realcs01}) is $\lambda_2 \approx 17.333\mu{\rm m}$, and at
1163: this wavelength the stack of 30 gratings of coated cylinders
1164: exhibits an anomalous absorptance (see Fig.~\ref{rtatmccco}),
1165: which corresponds to the large imaginary part of (\ref{pr04}). The
1166: stack of solid cylinders having a refractive index
1167: $n_c=n_{\alo}$ and radius $r_c=0.03\mu{\rm m}$, exhibits an
1168: anomalous absorptance (see Fig.~\ref{rtatmccco}) at
1169: $\widetilde{\lambda}_2$, corresponding to (\ref{pr08}).
1170:
1171: In both cases, the anomalous high absorptance is not a resonance
1172: effect, and it is produced by an anomalous behaviour of the
1173: relative dielectric constant of $\alo$. At
1174: $\lambda_2\approx\widetilde{\lambda}_2$ the real part of the
1175: relative dielectric constant of $\alo$ is relatively small, while
1176: the imaginary part reaches a maximum value of $\approx 30$ (see
1177: Fig.~\ref{al2o3eps}). Therefore the absorptance of the core of
1178: coated cylinders, or of the thin cylinders of core material only
1179: becomes large (see Fig.~\ref{rtatmccco}). At the same time, for
1180: long wavelengths and TM polarization, the effective dielectric
1181: constant of an array of cylinders is given by the linear mixing
1182: formula (\ref{tmpr4}). Thus, for an array of coated cylinders the
1183: relative effective dielectric constant is
1184: \begin{equation}
1185: \varepsilon^* =
1186: f_c \frac{\varepsilon_c}{\varepsilon_m} +
1187: f_s \frac{\varepsilon_s}{\varepsilon_m} + f_m \, ,
1188: \label{cclinmix}
1189: \end{equation}
1190: while for the array of solid cylinders of core material and radius
1191: $r_c$ we have
1192: \begin{equation}
1193: \varepsilon^* =
1194: f_c \frac{\varepsilon_c}{\varepsilon_m} + f_m \, .
1195: \label{colinmix}
1196: \end{equation}
1197: Hence, for a pure dielectric shell ($\varepsilon_s=$constant),
1198: in both cases, the behaviour of
1199: $\varepsilon^*$ is determined by $\varepsilon_c$ and this explains
1200: the common maximum in absorptance shown in
1201: Fig.~\ref{rtatmccco}, at the same position as the maximum in the
1202: imaginary part of the relative dielectric constant of $\alo$
1203: (see Fig.~\ref{al2o3eps}).
1204: Also, around the wavelength associated with this maximum, the real part of
1205: the relative dielectric constant of $\alo$ exhibits a sudden
1206: jump from negative to positive values, which makes almost equal
1207: the roots of the equations (\ref{realcs01}) and (\ref{realsm01}).
1208: Actually, the anomalous absorptance exhibited by
1209: stacks of thin, solid cylinders has been described in
1210: Ref.~\cite{anomaly}.
1211:
1212:
1213: \subsection{Dielectric Core and Absorbing Shell}
1214: \label{dcmsresults}
1215:
1216: Here, we consider the complementary structure with coated cylinders, having
1217: a dielectric core ($n_c=n_d$) and
1218: an aluminium oxide coating ($n_s=n_{\alo}$),
1219: embedded in vacuum ($n_m=1$).
1220: Now, the core is a lossless, dielectric material of
1221: constant $\varepsilon_c$.
1222: In this case, we expect both types of partial resonances to occur,
1223: and in both cases the coated cylinder has to exhibit the characteristics
1224: of a solid cylinder made from the core material.
1225: Absorptance is a characteristic feature of the metallic phase.
1226: Therefore, we expect a decrease of the total absorptance for
1227: the stack of gratings,
1228: around $\lambda_1$ and $\widetilde{\lambda}_1$.
1229:
1230:
1231:
1232: \subsubsection{TE Polarization}
1233: \label{teres1}
1234:
1235: For TE polarization the reflectance, transmittance and absorptance
1236: of the stack of coated cylinders with dielectric core and metallic
1237: shell, are almost the same as in the case of a stack of solid
1238: cylinders of radius $r_s$, made from aluminium oxide (see
1239: Fig.~\ref{rtatecc1}). Thus, in the neighbourhood of the
1240: shell--matrix resonance at $\widetilde{\lambda}_1 \approx
1241: 12.151\mu{\rm m}$ the two stacks have the same optical
1242: characteristics. In fact, Eq.(\ref{smpres}) shows that the coated
1243: cylinders must have optical characteristics similar to those of
1244: solid cylinders made from core material and of an extended radius
1245: $a=r_s^2/r_c$. In our case this means {\em equivalent} dielectric
1246: cylinders of radius $a=3.2\mu{\rm m}>\!\!>d/2=0.5\mu{\rm m}$, i.e.
1247: we have a stack of intersecting cylinders forming a homogeneous,
1248: dielectric slab, and we cannot apply the formulation from
1249: Sec.~\ref{raysec}. In the neighbourhood of the core--shell
1250: resonance at $\lambda_1 \approx 12.861\mu{\rm m}$, the reflectance
1251: of the stack of coated cylinders is diminished, while the
1252: absorptance is increased, compared with a stack of aluminium oxide
1253: cylinders (Fig.~\ref{rtatecc1}). At
1254: $\lambda_2\approx\widetilde{\lambda}_2$ the two types of stacks
1255: exhibit identical optical characteristics, and the local maximum
1256: in absorptance appears due to the anomalous behaviour of
1257: $\varepsilon_{\alo}$ around this wavelength (see
1258: Fig.~\ref{al2o3eps1}).
1259:
1260: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1261: \begin{figure}[h]
1262: \vspace{2ex}
1263: \centerline{\psfig{file=dmvTE003.eps,width=3.3in}}
1264: \vspace{2ex}
1265: \caption{TE polarization: Reflectance (red), transmittance (green)
1266: and absorptance (blue) for a stack of solid cylinders of
1267: refractive index $n_c=n_{\alo}$ and radius $r_s=0.31$ (dashed
1268: curves), and for a stack of coated cylinders with $r_c=0.03$ and
1269: $r_s=0.31$ (solid curves). In both cases the stack consists of 30
1270: gratings of cylinders in a hexagonal array. The vertical magenta
1271: lines mark the solutions $\lambda_1$ and $\lambda_2$ of
1272: (\protect\ref{realcs01}), and $\widetilde{\lambda}_1$ and
1273: $\widetilde{\lambda}_2$ of (\protect\ref{realsm01}).}
1274: \label{rtatecc1}
1275: \end{figure}
1276: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1277:
1278: To avoid such situations when the resonant system cannot be
1279: modelled by the formulation from Sec.~\ref{raysec}, we consider a
1280: stack of coated cylinders with a larger core $r_c=0.21$. In this
1281: case, a shell--matrix partial resonance will lead us to an
1282: equivalent dielectric cylinder, having a radius $a=0.46\mu{\rm
1283: m}<d/2$. Now, in the neighbourhood of the shell--matrix resonance
1284: at $\widetilde{\lambda}_1$, the stack of coated cylinders exhibits
1285: a drop in absorptance, an increased transmittance and a very small
1286: reflectance, while a stack of aluminium oxide cylinders having the
1287: radius $r_s$ has a zero transmittance and equal reflectance and
1288: absorptance (see Fig.~\ref{rtatecc2}). The equivalent stack,
1289: formed by dielectric cylinders of radius $a$, has a zero
1290: absorptance by definition. For long wavelengths ($\lambda>d$) it
1291: has a low reflectance and a high transmittance, therefore, the
1292: only indication of the existence of a shell--matrix partial
1293: resonance is the nonzero transmittance exhibited by the coated
1294: cylinders.
1295:
1296: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1297: \begin{figure}[h]
1298: \centerline{\psfig{file=dmvTE021.eps,width=3.3in}}
1299: \vspace{2ex}
1300: \caption{TE polarization: Reflectance (red), transmittance (green)
1301: and absorptance (blue) for a stack of solid cylinders of
1302: refractive index $n_c=n_{\alo}$ and radius $r_s=0.31$ (dashed
1303: curves), and for a stack of coated cylinders with $r_c=0.21$ and
1304: $r_s=0.31$ (solid curves). In both cases the stack consists of 30
1305: gratings of cylinders in a hexagonal array. The vertical magenta
1306: lines mark the solutions $\lambda_1$ and $\lambda_2$ of
1307: (\protect\ref{realcs01}), and $\widetilde{\lambda}_1$ and
1308: $\widetilde{\lambda}_2$ of (\protect\ref{realsm01}).}
1309: \label{rtatecc2}
1310: \end{figure}
1311: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1312:
1313: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1314: \begin{figure}[h]
1315: \vspace{3ex}
1316: \centerline{\psfig{file=dmvTM003.eps,width=3.3in}}
1317: \vspace{2ex}
1318: \caption{TM polarization: Reflectance (red), transmittance (green)
1319: and absorptance (blue) for a stack of solid cylinders of
1320: refractive index $n_c=n_{\alo}$ and radius $r_s=0.31$ (dashed
1321: curves), and for a stack of coated cylinders with $r_c=0.03$ and
1322: $r_s=0.31$ (solid curves). In both cases the stack consists of 30
1323: gratings of cylinders in a hexagonal array. The vertical magenta
1324: lines mark the solutions $\lambda_1$ and $\lambda_2$ of
1325: (\protect\ref{realcs01}), and $\widetilde{\lambda}_1$ and
1326: $\widetilde{\lambda}_2$ of (\protect\ref{realsm01}).}
1327: \label{rtatmcc1}
1328: \end{figure}
1329: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1330:
1331: In the neighbourhood of the core--shell resonance (at
1332: $\lambda_1$), Fig.~\ref{rtatecc2} shows a reduced absorptance for
1333: the stack of coated cylinders. Again, the equivalent stack,
1334: formed by dielectric cylinders of radius $r_s$, predicted by
1335: (\ref{cspres}), has a zero absorptance by definition, and so the
1336: comparison with a stack of aluminium oxide cylinders of radius $r_s$,
1337: seems to be more natural. In this case, it is interesting to
1338: remark the appearance of the duality mentioned in
1339: Sec.~\ref{teres}.
1340: The two stacks have the same absorptance and the reflectance of
1341: solid cylinders is equal with the transmittance of coated
1342: cylinders
1343: \begin{equation}
1344: R_{cc} \approx T_{sc} \approx 0, \quad
1345: T_{cc} \approx R_{sc}, \quad
1346: A_{cc} \approx A_{sc} .
1347: \label{scRTA1}
1348: \end{equation}
1349:
1350: \subsubsection{TM Polarization}
1351: \label{tmres1}
1352:
1353: As in the case of TE polarization, when the core radius of the
1354: coated cylinders is small (so that $a=r_s^2/r_c>d/2$),
1355: for TE polarization the reflectance, transmittance and
1356: absorptance of the stack of coated cylinders
1357: with dielectric core and metallic--type shell, are almost the same
1358: as in the case of a stack of solid cylinders of radius $r_s$,
1359: made from aluminium oxide (see Fig.~\ref{rtatmcc1}).
1360: In fact, there is no resonant behaviour in the whole range
1361: $10\mu{\rm m}<\lambda<20\mu{\rm m}$.
1362:
1363:
1364: In the case of a larger core of coated cylinders, the effect of
1365: shell--matrix partial resonance is small for TM polarization.
1366: Around $\lambda=\widetilde{\lambda}_1$ the absorptance of the
1367: coated cylinders is almost equal with their transmittance, and the
1368: reflectance is close to zero. The curves displayed in
1369: Fig.~\ref{rtatmcc2} show no sudden change around
1370: $\widetilde{\lambda}_1$ (shell--matrix partial resonance) and
1371: $\lambda_1$ (core--shell partial resonance), but it is interesting
1372: to remark the duality (\ref{scRTA1}) at $\lambda_1$.
1373:
1374: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1375: \begin{figure}[h]
1376: \vspace{3ex}
1377: \centerline{\psfig{file=dmvTM021.eps,width=3.3in}}
1378: \vspace{3ex}
1379: \caption{TM polarization: Reflectance (red), transmittance (green)
1380: and absorptance (blue) for a stack of solid cylinders of
1381: refractive index $n_c=n_{\alo}$ and radius $r_s=0.31$ (dashed
1382: curves), and for a stack of coated cylinders with $r_c=0.21$ and
1383: $r_s=0.31$ (solid curves). In both cases the stack consists of 30
1384: gratings of cylinders in a hexagonal array. The vertical magenta
1385: lines mark the solutions $\lambda_1$ and $\lambda_2$ of
1386: (\protect\ref{realcs01}), and $\widetilde{\lambda}_1$ and
1387: $\widetilde{\lambda}_2$ of (\protect\ref{realsm01}).}
1388: \label{rtatmcc2}
1389: \end{figure}
1390: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1391:
1392:
1393: \section{Conclusions}
1394:
1395: We have analyzed the partial resonances of a three--phase
1396: composite, structured as a hexagonal array of coated cylinders, at
1397: long wavelengths. For such structures there are two types of
1398: partial resonances: shell--matrix and core--shell. We have shown
1399: that the shell--matrix resonance has a very small effect. In
1400: contrast, the core--shell resonance appears for TE polarization in
1401: both cases of coated cylinders metal--dielectric--vacuum and
1402: dielectric--metal-- vacuum. Note that, for such resonances the
1403: stack of coated cylinders has a nonzero absorptance, so we cannot
1404: compare it with a stack of dielectric cylinders (in the case of
1405: dielectric--metal--vacuum composite). However, the comparison with
1406: a stack of solid, metallic cylinders of radius $r_s$ shows equal
1407: absorptance and a reflectance--transmittance duality, for the
1408: core--shell resonance.
1409:
1410: An interesting behaviour is exhibited by the composite in the
1411: vicinity of shell--matrix ($\widetilde{\lambda}_1$) and
1412: core--shell ($\lambda_1$) resonances (see
1413: Figs.~\ref{rtatccecdetail}, \ref{rtatmccco}, \ref{rtatecc2} and
1414: \ref{rtatmcc2}). In these regions, the composite has a very low
1415: reflectance ($R_{cc} \approx 0$), while the transmittance and
1416: absorptance take similar values.
1417:
1418: Generally, a band gap is defined by R=1, and T=0, so that sudden
1419: changes in optical characteristics of an array of coated cylinders
1420: will change the photonic band diagram. In our calculations the
1421: partial resonances occur at long wavelengths (i.e., in the region
1422: of the acoustic band) and we expect the photonic band diagram of
1423: the composite to show horizontal lines intersecting the acoustic
1424: band.
1425: %We are in progress to develop a numerical method
1426: %\cite{josa3} capable to produce the photonic band diagrams for a
1427: %complex crystal momentum ${\bf k}_B$, and thus enabling the study
1428: %of changes in the shape of the acoustic band for multi--phase
1429: %composites containing metallic phases.
1430: These ultrafast changes in the behaviour of a composite,
1431: due to the dependence of the refractive indexes on the wavelength,
1432: can provide the basis of future optical switches in photonic
1433: integrated circuits.
1434:
1435:
1436:
1437: \acknowledgments The Australian Research Council supported this
1438: work. Helpful discussions with Professor J. B. Pendry are acknowledged.
1439:
1440:
1441: \begin{references}
1442:
1443: \bibitem{cc1} N. A. Nicorovici, R. C. McPhedran and G. W. Milton,
1444: Proc.\ R.\ Soc.\ Lond.\ A {\bf 442}, 599 (1993).
1445:
1446: \bibitem{cc2} N. A. Nicorovici, R. C. McPhedran and G. W. Milton,
1447: Phys.\ Rev.\ B\ {\bf 49}, 8479 (1994).
1448:
1449: \bibitem{cc3} N. A. Nicorovici, D. R. McKenzie and R. C. McPhedran,
1450: Opt.\ Comm.\ {\bf 117} 151 (1955).
1451:
1452: \bibitem{ray} Lord Rayleigh,
1453: Philos.\ Mag.\ {\bf 34}, 481 (1892).
1454:
1455: \bibitem{ohad} L. Ohad,
1456: J.\ Appl.\ Phys.\ {\bf 77}, 1696 (1995).
1457:
1458: \bibitem{anomaly} R. C. McPhedran, N. A. Nicorovici, L. C. Botten,
1459: C. M. de Sterke, P. A. Robinson and A. A. Asatryan,
1460: %{\em Anomalous Absorptance by Stacked Metallic Cylinders},
1461: Opt.\ Comm.\ {\bf 168}, 47 (1999).
1462:
1463: \bibitem{pendry} J. B. Pendry,
1464: ``Intense focusing of light using metals'',
1465: in
1466: {\em Photonic Crystals and Light Localization in the 21st Century},
1467: ed. C. M. Soukoulis,
1468: NATO SCIENCE SERIES: C: Mathematical and Physical Sciences,
1469: Vol. {\bf 563}
1470: (Kluwer, Dordrecht, The Netherlands, 2001).
1471:
1472: \bibitem{gralak} B. Gralak, S. Enoch and G. Tayeb,
1473: J.\ Opt.\ Soc.\ Am.\ A {\bf 17}, 1012 (2000).
1474:
1475: \bibitem{pingsheng} W. Y. Zhang, X. Y. Lei, Z. L. Wang, D. G.
1476: Zheng, W. Y. Tam, C. T. Chan, and Ping Sheng,
1477: Phys.\ Rev.\ Lett.\ {\bf 84}, 2853 (2000).
1478:
1479: \bibitem{simon} R. Simon, J. R. Whinnery and T. van Duzer,
1480: {\em Fields and waves in communication electronics}
1481: (Wiley, New York, 1984).
1482:
1483: \bibitem{panofsky} W. K. H. Panofsky and M. Phillips,
1484: {\em Classical electricity and magnetism}, 2nd edn.
1485: (Addison-Wesley, Reading, Mass., 1962).
1486:
1487: \bibitem{cnr} S. K. Chin, N. A. Nicorovici and R. C. McPhedran,
1488: Phys.\ Rev.\ E\ {\bf 49}, 4590 (1994).
1489:
1490: \bibitem{nrb1} N. A. Nicorovici, R. C. McPhedran and L. C. Botten,
1491: Phys.\ Rev.\ E\ {\bf 52}, 1135 (1995).
1492:
1493: \bibitem{jmpls}
1494: R. C. McPhedran, N. A. Nicorovici, L. C. Botten and K. A. Grubits,
1495: %{\em Lattice Sums for Gratings and Arrays},
1496: J.\ Math.\ Phys.\ {\bf 41}, 7808 (2000).
1497:
1498: \bibitem{mcrae} E. G. McRae,
1499: Surface Science {\bf 11}, 479 (1968);
1500: Surface Science {\bf 11}, 492 (1968).
1501:
1502: \bibitem{josa12}
1503: L. C. Botten, N. A. Nicorovici, A. A. Asatryan, R. C. McPhedran,
1504: C. M. de Sterke and P. A. Robinson,
1505: %{\em Electromagnetic Scattering and Propagation through Gratings Stacks
1506: %of Metallic and Dielectric
1507: %Cylinders for Photonic Crystal Calculations.
1508: %(Part 1: Method, Part 2: Properties and implementation)},
1509: J.\ Opt.\ Soc.\ Am.\ A {\bf 17}, 2165 (2000);
1510: J.\ Opt.\ Soc.\ Am.\ A {\bf 17}, 2177 (2000).
1511:
1512: \bibitem{josa3} L. C. Botten, N. A. Nicorovici, R. C. McPhedran,
1513: C. M. de Sterke, A. A. Asatryan,
1514: ``Photonic band structure calculations using scattering matrices'',
1515: submitted to Phys.\ Rev.\ E.
1516:
1517: \bibitem{ray2} Lord Rayleigh,
1518: Philos.\ Mag.\ {\bf 36}, 365 (1918).
1519:
1520: \bibitem{herlofson} N. Herlofson,
1521: Arkiv.\ Fysik.\ {\bf 3}, 247 (1951).
1522:
1523: \bibitem{tonks} L. Tonks,
1524: Phys.\ Rev.\ {\bf 38}, 1219 (1931).
1525:
1526: \bibitem{romell} D. Romell,
1527: Nature\ {\bf 167}, 243 (1951).
1528:
1529: \bibitem{wait1} J. R. Wait,
1530: Can.\ J.\ Phys.\ {\bf 33}, 189 (1955).
1531:
1532: \bibitem{wait2} J. R. Wait,
1533: Can.\ J.\ Phys.\ {\bf 43}, 2212 (1965).
1534:
1535: \bibitem{bergman} D. J. Bergman,
1536: J.\ Phys.\ C: Solid State Phys.\ {\bf 12}, 4947 (1979).
1537:
1538: \bibitem{ross} R. C. McPhedran and D. R. McKenzie,
1539: Appl.\ Phys.\ {\bf 23}, 223 (1980).
1540:
1541: \bibitem{AS} M. Abramowitz and I. A. Stegun,
1542: {\em Handbook of mathematical functions}
1543: (Dover, New York, 1972).
1544:
1545: \bibitem{jewatem} R. C. McPhedran, N. A. Nicorovici and L. C. Botten,
1546: %{\em The TEM Mode and Homogenisation of Doubly Periodic Structures},
1547: J.\ Electromagn.\ Waves Appl.\ {\bf 11}, 981 (1997).
1548:
1549: \bibitem{perrins} W. T. Perrins, D. R. McKenzie and R. C. McPhedran,
1550: Proc.\ R.\ Soc.\ Lond.\ A{\bf 369}, 207 (1979).
1551:
1552: \bibitem{siam} C. G. Poulton, L. C. Botten, R. C. McPhedran,
1553: N. A. Nicorovici and A. B. Movchan,
1554: %{\em Noncommuting Limits in Electromagnetic Scattering:
1555: %Asymptotic Analysis for an Array of Highly Conducting Inclusions},
1556: SIAM J.\ Appl.\ Math.\ {\bf 61}, 1706 (2001).
1557:
1558: \bibitem{wiener} O. Wiener,
1559: Abhandlungen der mathematisch--physischen Klasse der K\"{o}niglich
1560: S\"{a}chsischen Gesellschaft der Wissenschaften
1561: {\bf 32}, 509 (1912).
1562:
1563: \bibitem{palik} E. D. Palik,
1564: {\em The handbook of optical constants of solids}
1565: (Academic, New York, 1993).
1566:
1567: \bibitem{duncanson}
1568: A. Duncanson and R. W. Stevenson,
1569: Proc.\ Phys.\ Soc.\ Lond. 72 (1958) 1001.
1570:
1571:
1572: %\bibitem{seamouse}
1573: %A. R. Parker, R. C. McPhedran, D. R. McKenzie, L. C. Botten
1574: %and N. A. Nicorovici,
1575: %{\em Aphrodite's iridescence},
1576: %Nature (2000), in press.
1577:
1578: \end{references}
1579:
1580: \end{document}
1581: