1: \documentclass[12pt]{article}
2: \usepackage{fullpage,amssymb,amsmath}
3: \usepackage[dvips]{epsfig}
4:
5:
6: %for 3-vectors/dyadics
7: \def\##1{{\bf #1}}
8: \def\=#1{\underline{\underline{#1}}}
9:
10: %for 6-vectors/dyadics
11: \def\+#1{\underline{\bf #1}}
12: \def\*#1{\underline{\underline{\bf #1}}}
13:
14: %\def\_#1{\underline{#1}}
15:
16: \def\r#1{(\ref{#1})}
17: \def\l#1{\label{#1}}
18: \def\c#1{\cite{#1}}
19:
20: \def\le{\left(}
21: \def\ri{\right)}
22: \def\les{\left[}
23: \def\ris{\right]}
24: \def\lec{\left\{}
25: \def\ric{\right\}}
26:
27: \def\.{\mbox{ \tiny{$^\bullet$} }}
28:
29: \def\epso{\epsilon_{\scriptscriptstyle 0}}
30: \def\lambdao{\lambda_{\scriptscriptstyle 0}}
31: \def\muo{\mu_{\scriptscriptstyle 0}}
32: \def\etao{\eta_{\scriptscriptstyle 0}}
33: \def\kappao{k_{\scriptscriptstyle 0}}
34: \def\ko{k_{\scriptscriptstyle 0}}
35: \def\co{c_{\scriptscriptstyle 0}}
36:
37: \def\eps{\epsilon}
38: \def\epsR{\epsilon_R}
39: \def\epsI{\epsilon_I}
40: \def\muR{\mu_R}
41: \def\muI{\mu_I}
42: \def\hcm{\mbox{\tiny{HCM}}}
43: \def\cm{\mbox{\tiny{CM}}}
44: \def\Br{\mbox{\tiny{Br}}}
45: \def\MG{\mbox{\tiny{MG}}}
46:
47: \def\eBr{\eps^{\Br}_{\hcm}}
48: \def\eMGa{\eps^{\MG }_{\hcm/a}}
49: \def\eMGb{\eps^{\MG }_{\hcm/b}}
50:
51: \def\kR{k_R}
52: \def\kI{k_I}
53:
54: \def\ct{\cos\theta}
55: \def\st{\sin\theta}
56:
57: \def\I{\*I}
58: \def\curl{\nabla\times}
59: \def\ux{\hat{\#u}_x}
60: \def\uy{\hat{\#u}_y}
61: \def\uz{\hat{\#u}_z}
62:
63: %\renewcommand{\baselinestretch}{1.6}
64:
65: %\pagestyle{myheadings}
66:
67: \begin{document}
68:
69: \LARGE
70: \begin{center}
71: {\bf A limitation of the Bruggeman formalism for homogenization}
72:
73: \vspace{10mm} \large
74:
75: Tom G. Mackay\footnote{Corresponding Author. Fax: + 44 131
76: 650 6553; e--mail: T.Mackay@ed.ac.uk.}\\
77: {\em School of Mathematics,
78: University of Edinburgh, Edinburgh EH9 3JZ, UK}\\
79: \vspace{3mm}
80: Akhlesh Lakhtakia\footnote{Fax:+1 814 865 99974; e--mail: akhlesh@psu.edu}\\
81: {\em CATMAS~---~Computational \& Theoretical
82: Materials Sciences Group\\ Department of Engineering Science and
83: Mechanics\\ Pennsylvania State University, University Park, PA
84: 16802--6812, USA}
85:
86: \end{center}
87:
88:
89: \vspace{4mm}
90:
91: \normalsize
92:
93:
94: \begin{abstract}
95:
96: The Bruggeman formalism provides an estimate of the effective
97: permittivity of a particulate composite medium comprising two component mediums.
98: The Bruggeman estimate is required to lie within the Wiener bounds and
99: the Hashin--Shtrikman bounds. Considering
100: the homogenization of
101: weakly dissipative component mediums
102: characterized by relative
103: permittivities with real parts of opposite signs,
104: we show that the
105: Bruggeman estimate may not be not physically reasonable when the component mediums
106: are weakly dissipative; furthermore,
107: both the Wiener bounds and the Hashin--Shtrikman bounds exhibit strong
108: resonances.
109:
110:
111:
112: \end{abstract}
113:
114:
115: \noindent {\bf Keywords:} homogenization; negative permittivity;
116: Bruggeman formalism; Maxwell Garnett formalism; Hashin--Shtrikman bounds; Wiener
117: bounds
118:
119: \section{Introduction}
120:
121: Metamaterials in the form of particulate composite mediums are currently of
122: considerable scientific and technological interest \c{Walser}.
123: Provided that wavelengths are sufficiently long compared with the
124: length scales of inhomogeneities, such a metamaterial may be
125: envisaged as a homogenized composite medium (HCM), arising from
126: two homogeneous component mediums \c{L96, Mackay03}. HCMs with especially
127: interesting properties may be conceptualized if the real parts of
128: the
129: relative permittivities (and/or relative permeabilities) of the two
130: component mediums have opposite signs \c{Lijimmw}.
131: This possibility arises for metal--in--insulator
132: dielectric composites \c{Sherwin, MLW01} and
133: has recently become feasible with the fabrication
134: of dielectric--magnetic materials displaying a negative index of
135: refraction in the microwave frequency range \c{Shelby,Smith}.
136:
137: Over many years, several theoretical formalisms have been developed in order
138: to estimate the effective constitutive parameters of particulate composite mediums \c{L96}.
139: In particular, the Maxwell Garnett and the Bruggeman homogenization formalisms
140: have been widely used \c{Ward}. Generally, the Maxwell Garnett formalism
141: is seen to hold only for dilute composite mediums \c{MG}.\footnote{The restriction on the applicability of
142: the Maxwell Garnett formalism to
143: dilute composite mediums
144: generally emerges from comparison with experimental data \c{Ward}. As the particulate volume fraction increases,
145: the distribution of particles begins to lose the randomness which is inherent to the theory
146: of the Maxwell Garnett formalism \c{Lijaem}. However, the restriction
147: could be bypassed if the distribution of particles in a composite medium continues to lack order even under densification,
148: which thought underlies the {\em random unit cell\/} approach developed by Smith
149: and colleagues \c{Sm1, Niklasson,Sm2}. An anonymous reviewer has suggested that self--assembly
150: techniques \c{Zhang} can yield randomness even at large particulate volume fractions, and could therefore extend
151: the applicability of the Maxwell Garnett formalism.}
152: More
153: widely applicable is the Bruggeman formalism that was initially
154: founded on the intuition that the total
155: polarization field is zero throughout the HCM \c{Brugge}.
156: A rigorous basis for the Bruggeman formalism is also available,
157: within the framework of the strong--permittivity--fluctuation theory (SPFT)
158: \c{TK81, MLW00}.
159:
160: Estimates of HCM constitutive parameters generated by
161: homogenization formalisms may
162: be required to lie within certain bounds. In particular, the
163: Wiener bounds \c{Wiener, Aspnes} and the Hashin--Shtrikman bounds \c{HS} are often invoked.
164: The Hashin--Shtrikman bounds coincide with the constitutive
165: parameter estimates of the Maxwell Garnett homogenization
166: formalism \c{Aspnes}. The applicability of theoretical bounds on the HCM
167: permittivity has recently been the focus of attention for composites
168: specified by relative permittivities with
169: positive--valued
170: real parts \c{Sihvola}.
171:
172: In this communication, we consider the application of the
173: Bruggeman formalism, together with the Wiener and
174: Hashin--Shtrikman bounds, to isotropic dielectric HCMs which arise
175: from component mediums characterized by
176: complex--valued relative
177: permittivities whose real parts have opposite signs.
178: This is scenario is typical of metal--in--insulator HCMs \c{Aspnes, Milton}, for example.
179: By duality, our analysis extends to
180: isotropic magnetic HCMs. It also extends to
181: isotropic dielectric--magnetic HCMs, because the permeability and the permittivity are
182: then independent of each other in the Bruggeman formalism \c{Kampia} (as also in the
183: Maxwell Garnett formalism \c{Lak92}). Therefore, our findings
184: are very relevant to the application of
185: homogenization formalisms \c{Lijimmw} to mediums
186: displaying negative index of refraction \c{LMW03}, for example.
187: Furthermore, the implications of our mathematical study extend beyond
188: the Bruggeman formalism to the SPFT
189: as well \c{Mackay03}.
190:
191: A note on notation: An $\exp(-i\omega t)$ time--dependence is
192: implicit in the following sections; and
193: the real and imaginary parts of
194: complex--valued quantities are denoted by
195: $\mbox{Re} \le \. \ri$ and $\mbox{Im} \le \. \ri$, respectively.
196:
197:
198: \section{Theory}
199:
200: \subsection{Bruggeman equation}
201:
202: Consider the homogenization of two isotropic dielectric component
203: mediums labelled $a$ and $b$. Let their
204: relative permittivities
205: be denoted by $\eps_a$ and $\eps_b$, respectively. For later
206: convenience, we define
207: \begin{equation}
208: \delta = \left\{
209: \begin{array}{ccc}
210: \displaystyle
211: \frac{\eps_a}{\eps_b} & \quad \mbox{if} \quad & \eps_a, \eps_b \in
212: \mathbb{R},\\ && \\
213: \displaystyle
214: \frac{\mbox{Re} \le \eps_a \ri }{\mbox{Re} \le \eps_b \ri} & \quad
215: \mbox{if} \quad & \eps_a, \eps_b \in
216: \mathbb{C}.
217: \end{array}
218: \right.
219: \end{equation}
220:
221: The Bruggeman estimate of the HCM relative permittivity, namely
222: $\eBr$, is provided implicitly via the relation \c{Ward}
223: \begin{eqnarray}
224: && \eBr = \frac{f_a \eps_a \le \eps_b + 2 \eBr \ri + f_b \eps_b
225: \le \eps_a + 2 \eBr \ri }{f_a \le \eps_b + 2 \eBr \ri + f_b \le
226: \eps_a + 2 \eBr \ri}, \l{Br}
227: \end{eqnarray}
228: wherein $f_a$ and $f_b$ are the respective volume fractions of
229: component mediums $a$ and $b$, and the particles of both component
230: mediums are assumed to be spherical. The Bruggeman equation \r{Br}
231: emerges naturally within the SPFT framework \c{Mackay03}. A rearrangement of
232: \r{Br} gives the quadratic equation
233: \begin{equation}
234: 2 \le \eBr \ri^2 + \eBr \les \eps_a \le f_b - 2f_a \ri + \eps_b
235: \le f_a - 2f_b \ri \ris - \eps_a \eps_b = 0. \l{quadratic}
236: \end{equation}
237: Only those $\eBr$--solutions of \r{quadratic} are valid under the principle of causality as encapsulated
238: by the Kramers--Kronig relations \c{BH} which conform to the restriction $\mbox{Im}\,\le
239: \eBr \ri \ge 0$.
240:
241:
242: Let $\Delta$ be the discriminant of the quadratic equation \r{quadratic}; i.e.,
243: \begin{equation}
244: \Delta = \les \eps_a \le f_b - 2f_a \ri + \eps_b \le f_a - 2f_b
245: \ri \ris^2 + 8 \eps_a \eps_b . \l{Delta1}
246: \end{equation}
247: Since $f_b = 1 - f_a$, we may express \r{Delta1} as
248: \begin{equation}
249: \Delta = 9 f^2_a \le \eps_a - \eps_b \ri^2 -6 f_a \le \eps^2_a -3
250: \eps_a \eps_b + 2 \eps^2_b \ri + \le \eps_a + 2 \eps_b \ri^2.
251: \end{equation}
252: An insight into the applicability of the Bruggeman formalism may
253: be gained by considering the $f_a$--roots of the equation $\Delta =
254: 0$; these are as follows:
255: \begin{equation}
256: \left.
257: f_a \right|_{\Delta = 0} = \frac{\eps^2_a -3 \eps_a
258: \eps_b + 2 \eps^2_b \pm 2 \sqrt{2} \sqrt{ - \eps_a \eps_b \le
259: \eps_a - \eps_b \ri^2}}{3 \le \eps_a - \eps_b \ri^2}.
260: \end{equation}
261: On restricting attention to nondissipative component
262: mediums (i.e., $\eps_{a,b} \in \mathbb{R}$), it is clear that $ \left. f_a \right|_{\Delta=0}$ are
263: complex--valued if $\delta
264: > 0$. Consequently, $\Delta > 0$ which implies that
265: $\eBr \in \mathbb{R}$. On the other hand,
266: $ \left. f_a \right|_{\Delta=0}$
267: are real--valued if $\delta < 0$. Thus, the Bruggeman estimate
268: $\eBr$ for $\delta < 0 $ may be complex--valued with nonzero imaginary part,
269: even though neither component medium is dissipative.
270:
271:
272:
273: \subsection{Bounds on the HCM relative permittivity}
274:
275:
276: Various bounds on the HCM relative permittivity have been
277: developed. Two of the most widely used are the Wiener bounds
278: \c{Wiener, Aspnes}
279: \begin{equation}
280: \left.
281: \begin{array}{l}
282: W_\alpha = \le \frac{f_a}{\eps_a} + \frac{f_b}{\eps_b} \ri^{-1} \\
283: \vspace{-4mm} \\
284: W_\beta = f_a \eps_a + f_b \eps_b
285: \end{array}
286: \right\}\,\l{Wiener}
287: \end{equation}
288: and the Hashin--Shtrikman
289: bounds \c{HS}
290: \begin{equation}
291: \left.
292: \begin{array}{l}
293: {HS}_\alpha = \eps_b + \frac{3 f_a \eps_b \le \eps_a - \eps_b
294: \ri}
295: {\eps_a + 2 \eps_b - f_a \le \eps_a - \eps_b \ri} \\ \vspace{-2mm} \\
296: {HS}_\beta = \eps_a + \frac{3 f_b \eps_a \le \eps_b - \eps_a
297: \ri} {\eps_b + 2 \eps_a - f_b \le \eps_b - \eps_a \ri}
298: \end{array}
299: \right\}\,. \l{Hashin}
300: \end{equation}
301: While both the Wiener bounds and the Hashin--Shtrikman bounds were
302: originally derived for real--valued constitutive parameters,
303: generalizations to complex--valued constitutive parameters have
304: been established \c{Milton}.
305:
306: The Hashin--Shtrikman bound $\mbox{HS}_\alpha$ is
307: equivalent to the Maxwell Garnett estimate of the HCM relative
308: permittivity $\eMGa$ based on spherical particles of component
309: medium $a$ embedded in the host component medium $b$. Similarly,
310: $\mbox{HS}_\beta$ is equivalent to the Maxwell Garnett estimate
311: of the HCM relative permittivity $\eMGb$ based on spherical
312: particles of component medium $b$ embedded in the host component
313: medium $a$. The estimate $\eMGa$ is valid for $f_a \lesssim 0.3$,
314: whereas the estimate $\eMGb$ is valid for $f_b \lesssim 0.3$; but see the
315: footnote in Section 1.
316:
317: To gain insights into the asymptotic behaviour of the Wiener and
318: Hashin--Shtrikman bounds, let us again restrict attention to the
319: case of nondissipative component mediums (i.e., $\eps_{a,b} \in
320: \mathbb{R}$). From \r{Wiener}, we see that $W_\beta$ remains
321: finite for all values of $\delta$, but $W_\alpha$ may become
322: infinite for $\delta < 0$ since
323: \begin{equation}
324: | W_\alpha | \rightarrow \infty \qquad \mbox{as} \qquad \delta
325: \rightarrow - \frac{f_a}{f_b}\, . \l{Wa_lim}
326: \end{equation}
327: In a similar vein, from \r{Hashin} we find that
328: \begin{equation}
329: | HS_\alpha | \rightarrow \infty \qquad \mbox{as} \qquad \delta
330: \rightarrow \frac{f_b - 3}{f_b}\,; \l{HSa_lim}
331: \end{equation}
332: thus, for all values of $\delta < -2 $ there exists a value of
333: $f_b \in (0,1)$ at which $HS_\alpha$ is unbounded. Analogously,
334: \begin{equation}
335: | HS_\beta | \rightarrow \infty \qquad \mbox{as} \qquad \delta
336: \rightarrow \frac{f_a}{f_a - 3}; \l{HSb_lim}
337: \end{equation}
338: so we can always
339: find a value of $f_a \in (0,1)$ at which $HS_\beta$ is unbounded, provided that $\delta \in ( -\frac{1}{2}, 0 ) $.
340:
341:
342:
343: \section{Numerical results}
344:
345: Let us now present, calculated values of the HCM relative
346: permittivity $\eBr$, along with the corresponding values of the
347: bounds $W_{\alpha, \beta}$ and $HS_{\alpha, \beta}$, for some representative examples. Both nondissipative
348: and dissipative HCMs are considered for $\delta = \pm 3$.
349:
350: \subsection{Nondissipative component mediums}
351:
352: The effects
353: of dissipation may be very clearly appreciated through first
354: considering the idealized situation wherein the components mediums
355: are nondissipative \c{vandeHulst}. Furthermore, although the absence of dissipation is unphysical due to
356: the dictates of causality \c{BH},
357: weak dissipation in a particular spectral regime
358: is definitely possible and is then often ignored \cite[Sec.2.5]{Ward}.
359:
360: Thus, it is instructive to begin with the commonplace
361: scenario wherein both $\eps_a > 0$ and $\eps_b > 0$.
362: For example, let
363: $\eps_a = 6$ and $\eps_b = 2$.
364: In Figure~1, $\eBr$ is plotted against $f_a$, along with the corresponding
365: Wiener bounds
366: $W_{\alpha, \beta}$
367: and Hashin--Shtrikman bounds $HS_{\alpha, \beta}$. The latter bounds are stricter than the former
368: bounds in the sense that
369: \begin{equation}
370: W_\alpha < HS_\alpha < \eBr < HS_\beta < W_\beta. \l{inequ}
371: \end{equation}
372: The close agreement between $\eBr$ and the lower Hashin--Shtrikman
373: bound $HS_\alpha$ at low volume fractions $f_a$ is indicative of
374: the fact that $HS_\alpha \equiv \eMGa$.
375: Similarly,
376: $\eBr$ agrees closely with the upper
377: Hashin--Shtrikman bound $HS_\beta$ at high values of $f_a$ since
378: $ HS_\beta \equiv \eMGb$.
379:
380: A markedly different situation develops if the real--valued $\eps_a $ and $\eps_b $ have opposite signs.
381: For example, the values of $\eBr$ calculated for $\eps_a = -6$ and $\eps_b = 2$ are graphed against $f_a$ in
382: Figure~2, together with the corresponding Wiener and
383: Hashin--Shtrikman bounds. The Bruggeman estimate $\eBr$ is
384: complex--valued with nonzero imaginary part for $f_a \lesssim
385: 0.82$. This estimate
386: is not physically reasonable. The Bruggeman homogenization formalism~---~unlike the
387: SPFT which is its natural generalization~---~has no mechanism for
388: taking coherent scattering losses into account. Furthermore, no account has been taken
389: in the Bruggeman equation \r{Br} for the finite size of the particles \c{Lijaem,PLS,Shan96}. Therefore, the
390: Bruggeman estimate of the HCM relative permittivity is required to
391: be real--valued if the component mediums are nondissipative.
392:
393: While $\eBr$ in Figure~2 is complex--valued,
394: the Wiener bounds
395: and the Hashin--Shtrikman bounds are both real--valued. In
396: accordance with \r{Wa_lim}, we see that $| W_\alpha | \rightarrow
397: \infty$ as $f_a \rightarrow \frac{3}{4}$. Similarly, $|
398: HS_\alpha | \rightarrow \infty$ in the limit $f_a \rightarrow
399: \frac{1}{4}$, as may be anticipated from \r{HSa_lim}. Furthermore, since
400: $HS_\alpha \equiv \eMGa$, the Maxwell Garnett formalism is clearly
401: inappropriate here. We also observe that the inequalities \r{inequ}
402: which hold for $\delta > 0$, do not hold for $\delta < 0$.
403:
404:
405: \subsection{Weakly dissipative component mediums}
406:
407: Let us now investigate $\eBr$ and its associated bounds when the
408: component mediums are dissipative; i.e., $\eps_{a,b} \in
409: \mathbb{C}$. We begin with those cases for which $\delta > 0$:
410: for
411: example, we take $\eps_a = 6 + 0.3i$ and $\eps_b = 2 + 0.2i$. In
412: Figure~3, $\eBr$ is plotted against $f_a$, and the associated Wiener bounds
413: $W_{\alpha, \beta}$ and the Hashin--Shtrikman bounds $HS_{\alpha,
414: \beta}$ are also presented. The behaviour of the real parts of
415: $\eBr$, $W_{\alpha, \beta}$ and $HS_{\alpha, \beta}$
416: closely resembles that displayed in the nondissipative example
417: of Figure~1. In fact,
418: the
419: following generalization of \r{inequ} holds:
420: \begin{equation}
421: \mbox{Re} \le W_\alpha \ri < \mbox{Re} \le HS_\alpha
422: \ri < \mbox{Re} \le \eBr \ri < \mbox{Re} \le HS_\beta \ri <
423: \mbox{Re}
424: \le W_\beta \ri .
425: \l{inequ2}
426: \end{equation}
427: However, this ordering \r{inequ2} does not extend to
428: the imaginary parts of $\eBr$, $W_{\alpha, \beta}$ and
429: $HS_{\alpha, \beta}$.
430:
431: Turning to the cases for $\delta < 0$, we let $\eps_a = -6 + 0.3i$ and
432: $\eps_b = 2 + 0.2i$, for example. The corresponding Bruggeman
433: estimate $\eBr$ is graphed as function of $f_a$, along with the
434: Wiener bounds $W_{\alpha, \beta}$ and the Hashin--Shtrikman bounds
435: $HS_{\alpha, \beta}$ in Figure~4. Since $\mbox{Im} \le \eps_{a,b}
436: \ri \neq 0$, the real parts of $W_\alpha$ and $HS_\alpha$ remain
437: finite, unlike in the corresponding nondissipative scenario
438: presented in Figure~2.
439:
440: However, the real and imaginary parts of
441: $W_\alpha$ and $HS_\alpha$ exhibit strong resonances in the
442: vicinity of $f_a = \frac{3}{4}$ (for $W_\alpha$) and $f_a =
443: \frac{1}{4}$ (for $HS_\alpha$). These resonances become
444: considerably more pronounced if the degree of dissipation
445: exhibited by the component mediums is reduced. For example, in
446: Figure~5 the graphs corresponding to Figure~4 are reproduced for
447: $\eps_a = -6 +
448: 0.003i$ and $\eps_b = 2 + 0.002i$. We observe in particular that
449: $\mbox{Im} \le \eBr \ri > 1$ for $ 0.05 \lesssim f_a \lesssim 0.8$.
450: Thus, the Bruggeman estimate $\eBr$ vastly exceeds both the
451: Wiener bounds $W_{\alpha, \beta}$ and the Hashin--Shtrikman bounds
452: $HS_{\alpha, \beta}$ for a wide range of $f_a$.
453: Since $\mbox{Im} \le \eps_{a,b} \ri
454: \leq 0.003$, the estimates of ${\rm Im}(\eBr)$ are clearly
455: unreasonable.
456:
457: Furthermore, since the real and imaginary parts of
458: $HS_{\alpha} \equiv \eMGa $ exhibit sharp resonances at $f_a = \frac{1}{4}$,
459: we may infer that the Maxwell Garnett formalism is inapplicable for $\delta < 0$.
460:
461: \subsection{Highly dissipative component mediums}
462:
463:
464: On comparing Figures~4 and 5, we conclude that the Bruggeman
465: formalism, the Weiner bounds and the Hashin--Shtrikman bounds become increasing
466: inappropriate as the degree of dissipation decreases towards zero. This means that
467: all three {\em could\/} be applicable rather well when the dissipation is not weak.
468:
469: Therefore, let us examine the scenario wherein the real and imaginary parts of the relative permittivities of the
470: component medium are of the same order of magnitude;
471: i.e., we take $\eps_a = -6 + 3i$ and $\eps_b = 2 + 2i$. The
472: corresponding plots of
473: the Bruggeman estimate $\eBr$ together with the
474: Wiener bounds $W_{\alpha, \beta}$ and the Hashin--Shtrikman bounds
475: $HS_{\alpha, \beta}$ are presented in Figure~6.
476: The real and imaginary parts of the Bruggeman estimate are
477: physically plausible, and both lie within the Hashin--Shtrikman
478: bounds. The Hashin--Shtrikman bounds themselves do not exhibit
479: resonances, and the Weiner bounds do not exhibit strong resonances.
480: Accordingly, we conclude that many previously published results are
481: not erroneous, but caution is still advised.
482:
483:
484:
485: \section{Discussion}
486:
487: The Bruggeman homogenization formalism is well--established in the
488: context of isotropic dielectric HCMs, as well as more generally
489: \c{L96}. However, this formalism was shown in Section 3.2 to be
490: inapplicable for HCMs which arise from two
491: isotropic dielectric
492: component mediums, characterized by relative permittivities $\eps_a$
493: and $\eps_b$, with
494: \begin{itemize}
495: \item[(i)]
496: $\mbox{Re} \le \eps_a \ri$ and $\mbox{Re} \le
497: \eps_b \ri$
498: having opposite signs; and
499: \item[(ii)] $| \mbox{Re} \le \eps_{a,b} \ri |$ $ \gg $
500: $| \mbox{Im} \le \eps_{a,b} \ri |$.
501: \end{itemize}
502: Since the Bruggeman formalism provides the comparison medium
503: which underpins the SPFT, it may be inferred that the SPFT is
504: likewise not applicable to the scenarios of (i) with (ii).
505:
506: It is also demonstrated in Section 3.2 that both
507: the Wiener bounds and the Hashin--Shtrikman bounds
508: can exhibit strong resonances when the component mediums are
509: characterized by (i) with (ii). In the vicinity of resonances,
510: these bounds clearly do not constitute tight bounds on the HCM
511: relative permittivity. As a direct consequence, the Maxwell
512: Garnett homogenization formalism, like the Bruggeman
513: homogenization formalism, is inapplicable to the scenarios of (i)
514: with (ii). This limitation also extends to the recently developed
515: incremental \c{IMG} and differential \c{DMG} variants of the
516: Maxwell Garnett formalism.
517:
518:
519: If the component mediums are sufficiently dissipative then
520: the Bruggeman formalism and the Hashin--Shtrikman bounds (and
521: therefore
522: also the Maxwell Garnett formalism) provide physically plausible
523: estimates, despite the real parts of the component medium
524: relative permittivities
525: having opposite signs~---~as shown in Section 3.3. The explicit delineation
526: of the appropriate parameter range(s) for the Bruggeman formalism and
527: the Hashin--Shtrikman bounds is a matter for
528: future investigation.
529:
530: Bounds can, of course, be violated by a formalism if the underlying conditions for
531: the formalism are in conflict with those used for deriving the bounds. Sihvola \c{Sihvola}
532: has catalogued the following conflicts:
533: \begin{itemize}
534: \item[(a)] Bounds derived for nondissipative component mediums can be invalid for
535: the real parts of either $\eMGa$ or $\eMGb$ for a composite medium containing dissipative component mediums.
536: \item[(b)] Percolation cannot be cannot be captured by the Maxwell Garnett formalism \c{Sherwin,Sasta}. Hence, the Hashin--Shtrikman
537: bounds, being based on the Maxwell Garnett formalism, can be violated by
538: the Bruggeman estimate $\eBr$ for a percolative composite medium.
539: \item[(c)] The derivations of bounds generally assume that the particles in a composite medium
540: have simple shapes. If the particle shapes are complicated, the composite medium
541: may display properties not characteristic of the either of the
542: component mediums. For instance, magnetic properties can be displayed when the particles in a composite medium
543: have complex shapes \cite{Lbel,PHRS}, even though the component mediums are nonmagnetic. Clearly,
544: the magnetic analogs of $W_\alpha$, $W_\beta$, ${HS}_\alpha$ and ${HS}_\beta$
545: are then inapplicable.
546: \item[(d)] $W_\alpha$, $W_\beta$, ${HS}_\alpha$, and ${HS}_\beta$ as well as their magnetic
547: analogs are also invalid {\em prima facie\/} when the component
548: mediums exhibit magnetoelectric properties
549: \c{Mackay03,Lijaem,Mich00}.
550: \item[(e)] Bounds derived for electrically small particles become inapplicable with increasing frequency, due to
551: the emergence
552: of finite--size effects \c{PLS}. Even the
553: concept of homogenization becomes questionable with increasing electrical size \cite[p. xiii]{L96}.
554: \end{itemize}
555: In contrast, the bounds and the homogenization formalisms studied in this paper share the same premises; yet, a conflict
556: arises in certain situations because the bounds exhibit resonance while the homogenization estimates do not.
557:
558:
559: \section{Concluding remarks}
560: As several conventional approaches to homogenization are not
561: appropriate to the HCMs arising from component mediums characterized
562: by (i) with (ii),
563: there is a requirement for new theoretical
564: techniques to treat this case.
565: This
566: requirement is all the more pressing, given the growing scientific
567: and technological importance of new types of
568: metamaterials
569: \c{Walser, LMW03}.
570:
571: \bigskip\bigskip
572: \noindent {\bf Acknowledgement.} We thank two anonymous reviewers for comments that led
573: to the improvement of this paper.
574:
575:
576: \begin{thebibliography}{99}
577:
578:
579: \bibitem{Walser}
580: R.M. Walser,
581: %Metamaterials: an introduction,
582: in: W.S. Weiglhofer, A. Lakhtakia (Eds.),
583: Introduction to Complex Mediums for Optics and
584: Electromagnetics, SPIE,
585: Bellingham, WA, USA, 2003, pp.295--316.
586:
587:
588: \bibitem{L96}
589: A. Lakhtakia (Ed.), Selected Papers on Linear Optical
590: Composite Materials, SPIE,
591: Bellingham, WA, USA, 1996.
592:
593: \bibitem{Mackay03}
594: T.G. Mackay,
595: in: W.S. Weiglhofer, A. Lakhtakia (Eds.),
596: Introduction to Complex Mediums for Optics and
597: Electromagnetics, SPIE,
598: Bellingham, WA, USA, 2003, pp.317--345.
599:
600: \bibitem{Lijimmw}
601: A. Lakhtakia,
602: Int. J. Infrared Millim. Waves 23 (2002) 813.
603:
604: \bibitem{Sherwin}
605: J.A. Sherwin, A. Lakhtakia, B. Michel,
606: Opt. Commun. 178 (2000) 267.
607:
608: \bibitem{MLW01}
609: T.G. Mackay, A. Lakhtakia, W.S. Weiglhofer,
610: Opt. Commun. 197 (2001) 89.
611:
612: \bibitem{Shelby}
613: R.A. Shelby, D.R. Smith, S. Schultz,
614: %``Experimental verification of a negative index of refraction''
615: Science 292 (2001) 77.
616:
617: \bibitem{Smith}
618: D.R. Smith, W.J. Padilla, D.C. Vier, S.C. Nemat--Nasser, S. Schultz,
619: %``Composite medium with simultaneously negative permeability and
620: %permittivity'',
621: Phys. Rev. Lett. 84 (2000) 4184.
622:
623:
624: \bibitem{Ward}
625: L. Ward, The Optical Constants of Bulk Materials and Films,
626: 2nd Edition,
627: Adam Hilger, Bristol, UK, 1995; Chap. 8.
628:
629: \bibitem{MG}
630: J.C. Maxwell Garnett,
631: %``Colours in metal glasses and in metallic
632: %films'',
633: Phil. Trans. R. Soc. Lond. A 203 (1904) 385.
634: (Reproduced in \c{L96}).
635:
636:
637: \bibitem{Lijaem}
638: A. Lakhtakia, B. Shanker,
639: Int. J. Appl. Electromag. Mater. 4 (1993) 65.
640:
641:
642: \bibitem{Sm1}
643: G.B. Smith,
644: J. Phys. D: Appl. Phys. 10 (1977) L39.
645:
646: \bibitem{Niklasson}
647: G.A. Niklasson, C.G. Granqvist, O. Hunderi,
648: Appl. Opt. 20 (1981) 26.
649:
650: \bibitem{Sm2}
651: G.B. Smith, in: W.S. Weiglhofer, A. Lakhtakia (Eds.),
652: Introduction to Complex Mediums for Optics and
653: Electromagnetics, SPIE,
654: Bellingham, WA, USA, 2003, pp.421--446.
655:
656: \bibitem{Zhang}
657: S. Zhang,
658: Materials Today 6(5) (2003), 20; May issue.
659:
660: \bibitem{Brugge}
661: D.A.G. Bruggeman,
662: %Berechnung verschiedener physikalischer
663: %Konstanten von heterogenen Substanzen, I.
664: %Dielektrizit\"ats\-konstanten und Leit\-f\"ahig\-keiten der
665: %Misch\-k\"orper aus isotropen Substanzen,
666: Ann. Phys. Lpz.
667: 24 (1935) 636. (Reproduced in \c{L96}).
668:
669:
670: \bibitem {TK81}
671: L. Tsang, J.A. Kong,
672: %Scattering of electromagnetic waves from
673: %random media with strong permittivity fluctuations,
674: Radio
675: Sci. 16 (1981) 303. (Reproduced in \c{L96}).
676:
677:
678: \bibitem{MLW00}
679: T.G. Mackay, A. Lakhtakia, W.S. Weiglhofer,
680: %Strong--property--fluctuation theory for homogenization of
681: %bianisotropic composites: formulation,
682: Phys. Rev. E
683: 62 (2000) 6052; Erratum 63 (2001) 049901.
684:
685:
686: \bibitem{Wiener}
687: O. Wiener,
688: %``Die Theorie des Mischk\"orpers f\"ur das Feld der
689: %Station\"aren Str\"omung",
690: Abh. Math.--Phys. Kl. S\"achs.
691: 32 (1912) 507. (Reproduced in \c{L96}).
692:
693: \bibitem{Aspnes}
694: D.E. Aspnes,
695: %Local-field effects and effective--medium theory: A microscopic
696: %perspective,
697: Am. J. Phys. 50 (1982) 704. (Reproduced in \c{L96}).
698:
699: \bibitem{HS}
700: Z. Hashin, S. Shtrikman,
701: %A variational approach to the theory of
702: %the effective magnetic permeability of multimedium materials,
703: J. Appl. Phys. 33 (1962) 3125.
704:
705: \bibitem{Sihvola}
706: A.H. Sihvola,
707: IEEE Trans. Geosci. Remote Sens. 40 (2002) 880.
708:
709:
710: \bibitem{Milton}
711: G.W. Milton,
712: %Bounds on the complex dielectric constant
713: %of a composite material,
714: Appl. Phys. Lett. 37 (1980) 300.
715:
716: \bibitem{Kampia}
717: R.D. Kampia, A. Lakhtakia,
718: %Bruggeman model for chiral particulate composites,
719: J. Phys. D: Appl. Phys. 25 (1992) 1390.
720:
721: \bibitem{Lak92}
722: A. Lakhtakia,
723: %General theory of Maxwell--Garnett model for particulate
724: %composites with bi-isotropic host materials,
725: Int. J. Electron. 73 (1992) 1355.
726:
727:
728: \bibitem{LMW03}
729: A. Lakhtakia, M.W. McCall, W.S. Weiglhofer,
730: in: W.S. Weiglhofer, A. Lakhtakia (Eds.),
731: Introduction to Complex Mediums for Optics and
732: Electromagnetics, SPIE,
733: Bellingham, WA, USA, 2003, pp.347--363.
734:
735:
736: \bibitem{vandeHulst}
737: H.C. van de Hulst,
738: Light Scattering by Small Particles,
739: Dover, New York, NY, USA, 1981; Chap. 10.
740:
741: \bibitem{BH}
742: C.F. Bohren, D.R. Huffman,
743: Absorption and Scattering of Light by Small Particles,
744: Wiley, New York, NY, USA, 1983; Sec. 2.3.2.
745:
746: \bibitem{PLS}
747: M.T. Prinkey, A. Lakhtakia, B. Shanker,
748: Optik 96 (1994) 25.
749:
750: \bibitem{Shan96}
751: B. Shanker,
752: J. Phys. D: Appl. Phys. 29 (1996) 281.
753:
754:
755: \bibitem{IMG}
756: A. Lakhtakia,
757: % ``Incremental Maxwell Garnett formalism for
758: %homogenizing particulate composite media",
759: Microw. Opt. Technol. Lett. 17 (1998) 276.
760:
761:
762: \bibitem{DMG}
763: B. Michel, A. Lakhtakia, W.S. Weiglhofer, T.G. Mackay,
764: %``Incremental and differential Maxwell Garnett formalisms for
765: %bi--anisotropic composites",
766: Compos. Sci. Technol. 61 (2001) 13.
767:
768: \bibitem{Sasta}
769: A. Sihvola, S. Saastamoinen, K. Heiska,
770: Remote Sens. Rev. 9 (1994) 39.
771:
772: \bibitem{Lbel}
773: A. Lakhtakia,
774: Beltrami Fields in Chiral Media,
775: World Scientific, Singapore, 1994; Chap. 3.
776:
777: \bibitem{PHRS}
778: J.B. Pendry, A.J. Holden, D.J. Robbins, W.J. Stewart,
779: IEEE Trans. Microw. Theory Tech. 47 (1999) 2075.
780:
781: \bibitem{Mich00}
782: B. Michel,
783: in: O.N. Singh, A. Lakhtakia (Eds.),
784: Electromagnetic Fields in Unconventional Materials and
785: Structures, Wiley, New York, NY, USA, 2000, pp.39-82.
786:
787:
788: \end{thebibliography}
789:
790: \newpage
791:
792: \begin{figure}[!ht]
793: \centering \psfull \epsfig{file=figure1.eps,width=5.0in}
794: \caption{\label{fig1}
795: The Bruggeman estimate $\eBr$ (solid line) plotted against $f_a$
796: for $\eps_a = 6$ and $\eps_b = 2$. Also plotted are the Wiener
797: bounds, $W_\alpha$ (thick dashed line) and $W_\beta$ (thin dashed line), and the
798: Hashin--Shtrikman bounds, $HS_\alpha$ (thick broken dashed
799: line) and $HS_\beta$ (thin broken dashed
800: line).
801: }
802: \end{figure}
803:
804:
805: \newpage
806:
807: \begin{figure}[!ht]
808: \centering \psfull \epsfig{file=figure2a.eps,width=5.0in}
809: \epsfig{file=figure2b.eps,width=5.0in}
810: \caption{\label{fig2}
811: The real (top) and imaginary (bottom) parts of the Bruggeman
812: estimate $\eBr$ (solid line) plotted against $f_a$ for $\eps_a =
813: -6$ and $\eps_b = 2$. Also plotted are the real parts of the
814: Wiener
815: bounds, $W_\alpha$ (thick dashed line) and $W_\beta$ (thin dashed line), and the
816: Hashin--Shtrikman bounds, $HS_\alpha$ (thick broken dashed
817: line) and $HS_\beta$ (thin broken dashed
818: line).
819: The imaginary parts of $W_{\alpha, \beta}$ and
820: $HS_{\alpha, \beta}$ are null--valued. }
821: \end{figure}
822:
823: \newpage
824:
825: \begin{figure}[!ht]
826: \centering \psfull \epsfig{file=figure3a.eps,width=5.0in}
827: \epsfig{file=figure3b.eps,width=5.0in}
828: \caption{\label{fig3}
829: The real (top) and imaginary (bottom) parts of the
830: Bruggeman estimate $\eBr$ (solid line) plotted against $f_a$
831: for $\eps_a = 6 + 0.3i$ and $\eps_b = 2 + 0.2i$.
832: Also plotted are the Wiener
833: bounds, $W_\alpha$ (thick dashed line) and $W_\beta$ (thin dashed line), and the
834: Hashin--Shtrikman bounds, $HS_\alpha$ (thick broken dashed
835: line) and $HS_\beta$ (thin broken dashed
836: line).
837: }
838: \end{figure}
839:
840: \newpage
841:
842: \begin{figure}[!ht]
843: \centering \psfull \epsfig{file=figure4a.eps,width=5.0in}
844: \epsfig{file=figure4b.eps,width=5.0in}
845: \caption{\label{fig4}
846: As Figure~3 but for
847: $\eps_a = -6 + 0.3i$ and $\eps_b = 2 + 0.2i$.
848: }
849: \end{figure}
850:
851: \newpage
852:
853: \begin{figure}[!ht]
854: \centering \psfull \epsfig{file=figure5a.eps,width=5.0in}
855: \epsfig{file=figure5b.eps,width=5.0in}
856: \caption{\label{fig5}
857: As Figure~3 but for
858: $\eps_a = -6 + 0.003i$ and $\eps_b = 2 + 0.002i$.
859: }
860: \end{figure}
861: \newpage
862:
863: \begin{figure}[!ht]
864: \centering \psfull \epsfig{file=figure6a.eps,width=5.0in}
865: \epsfig{file=figure6b.eps,width=5.0in}
866: \caption{\label{fig6}
867: As Figure~3 but for
868: $\eps_a = -6 + 3i$ and $\eps_b = 2 + 2i$.
869: }
870: \end{figure}
871:
872:
873: \end{document}
874: