1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: % File: OLpagelength.tex %
3: % VERSION: 1.1 %
4: % Date: May 15, 2004 [sdinee] %
5: % %
6: % For assistance, contact Joseph Richardson, %
7: % jricha@osa.org %
8: % %
9: % LaTeX template and instructions for %
10: % length check and submission of OSA %
11: % Optics Letters manuscripts %
12: % %
13: % %
14: % \documentclass[10pt,letterpaper,twocolumn]{article} %
15: % \usepackage{OL} %
16: % %
17: % (c) 2004 Optical Society of America %
18: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
19:
20: \documentclass[10pt,letterpaper,twocolumn]{article} %% two column, final layout
21:
22: %\documentclass[12pt]{article} % single column, double spaced
23: %\usepackage[tablesfirst,notablist,nomarkers]{endfloat} %% float figs. to back
24:
25: \usepackage{ol}
26: \usepackage{hyperref}
27: \usepackage{amsmath}
28:
29: \begin{document}
30:
31: \twocolumn[ %% activate for two-column option
32:
33: \title{Localized vibrational modes in optically bound structures}
34: \author{Jack Ng and C.T. Chan}
35: \address{Department of Physics, Hong Kong University of Science and Technology, Clearwater Bay, Hong Kong, China}
36:
37: % Do not use \email or \homepage here. E-mail and URL can be given just before references.
38:
39: \begin{abstract}
40: We show, through analytical theory and rigorous numerical calculations, that
41: optical binding can organize a collection of particles into stable
42: one-dimensional lattice. This lattice, as well as other optically-bound
43: structures, are shown to exhibit spatially localized vibrational eigenmodes.
44: The origin of localization here is distinct from the usual mechanisms such
45: as disorder, defect, or nonlinearity, but is a consequence of the
46: long-ranged nature of optical binding. For an array of particles trapped by
47: an interference pattern, the stable configuration is often dictated by the
48: external light source, but our calculation revealed that inter-particle
49: optical binding forces can have a profound influence on the dynamics.
50: \end{abstract}
51:
52:
53: \ocis{140.7010,220.4610,220.4880,999.9999(Optical Binding)}
54:
55: ] %% activate for two-column option
56:
57: \noindent Since its introduction many years ago,\cite{Ashkin:1970} optical
58: manipulation has evolved into a major technique for manipulating small
59: particles, and recently, simultaneous manipulations of multi-particles have
60: been demonstrated.\cite{See:2003}$^{ }$It is known that in addition to the
61: well-known one-body force such as the gradient force that depends on the
62: intensity profile, there is an optical binding (OB) force that couples the
63: particles together.\cite{Burns:1990}$^{,}$\cite{Lin:2005} Nevertheless,
64: for an extended array of particles, the nature of OB is not fully
65: understood, although some theoretical efforts were devoted to small
66: clusters.\cite{Lin:2005}$^{,}$\cite{Chaumet:2001} As the principles
67: underlining these inter-particle forces are different from that of the
68: traditional light-trapping, we expect some new and interesting applications.
69:
70: In this paper, we demonstrate an interesting consequence of OB in a
71: spatially extended structure bound by light: the existence of spatially
72: localized VEM (vibrational eigenmodes). We illustrate the physics by
73: considering a one-dimensional ``lattice'' bound by light. Wave localization
74: is known to occur in defect or impurity sites of an otherwise ordered
75: lattice. In solids, the ``defect'' can be impurity atoms that localize
76: phonons, and in the intrinsic localized modes, the ``defect'' is derived
77: from the nonlinearly excited particles.\cite{Campbell:2004} Here the
78: localization occurs in the linear dynamics regime in an ordered array of
79: identical particles without defect or disorder.
80:
81: Optically bound structures have been investigated in a number of recent
82: experiments. Stable cluster configurations had been
83: realized\cite{Lin:2005}$^{,}$\cite{Tatarkova:2002}$^{,}$\cite{Singer:2003}$^{,}$\cite{Black:2003}$^{,}$\cite{Garces:2005}$^{,}$\cite{Mellor:2006}
84: and vibrational motions were observed.\cite{Tatarkova:2002} In particular,
85: the most commonly observed geometry is an one-dimensional array of
86: particles, bound by a pair counterpropagating
87: beams\cite{Tatarkova:2002}$^{,}$\cite{Singer:2003}$^{,}$\cite{Black:2003}
88: or evanescent waves.\cite{Garces:2005}$^{,}$\cite{Mellor:2006}
89:
90: Consider a linear chain of $N$ evenly spaced spheres in air. The particles have
91: mass density $\rho $=1,050 kg/m$^{3}$, dielectric constant $\varepsilon
92: $=2.53 ($\sim $polystyrene), and radii $a = \lambda / 10 = 52$ nm, so that
93: they are small compare to the incident light's wavelength $\lambda $=520 nm.
94: The particles are illuminated by the standing wave formed by a pair of
95: counterpropagating plane waves
96:
97:
98: \begin{equation}
99: \label{eq1}
100: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
101: {E}} _{in}
102: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
103: {r}} ) = 2E_0 \cos \left( {kz} \right)\hat {x},
104: \end{equation}
105:
106:
107:
108: \noindent
109: where $k$ is the wavenumber, and the intensity for each beam is set to be 0.01
110: W/$\mu $m$^{2}$.\cite{Burns:1990}$^{,}$\cite{Lin:2005}
111:
112: To calculate the optical force acting on the particles, we employ the
113: rigorous and highly accurate multiple scattering and Maxwell stress tensor
114: (MS-MST) formalism,\cite{Lin:2005} which requires no approximation
115: and subject only to numerical truncation errors (we use multipoles up to
116: $L$=6). The optical force tends to drive small particles to the region of
117: strong light intensity. For an array of $N$ evenly spaced particles aligned
118: along the $z$-axis, one expects a stable one-dimensional lattice with a lattice
119: constant of $\lambda $/2:
120:
121:
122: \begin{equation}
123: \label{eq2}
124: {\begin{array}{*{20}c}
125:
126: {\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
127: {R}} _n = \left( {0,0,n\lambda / 2} \right),} \hfill & {n = 1,2\ldots ,N}
128: \hfill \\
129: \end{array} },
130: \end{equation}
131:
132:
133:
134: \noindent
135: where
136: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
137: {R}} _n $ is the equilibrium position for the \textit{n-th} particle. Indeed, we found
138: that the geometry defined in (\ref{eq2}) corresponds to a zero-force configuration
139: and the configuration is proven to be stable by using linear stability
140: analysis.\cite{Lin:2005}$^{,}$\cite{The:1} The longitudinal trapping
141: (along the $z$-axis) is mainly provided by the gradient force of the incident
142: beam, and it is further enhanced by OB.\cite{This:1} On the other hand, the
143: transverse stability (on the \textit{xy}-plane) is solely induced by OB. We note that
144: there are other beam configurations, other than that specified by (\ref{eq1}), that
145: can stabilize a linear chain as demonstrated by recent experiments.$^{
146: }$\cite{Tatarkova:2002}$^{,}$\cite{Singer:2003}$^{,}$\cite{Mellor:2006}$^{,}$\cite{Chowdhury:1985}
147:
148: The VEMs are obtained by diagonalizing the force matrix
149: $(\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
150: {K}} )_{jk} = \partial
151: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
152: {f}} _{light} )_j / \partial (\Delta
153: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
154: {x}} )_k $,\cite{Lin:2005} which is found by linearizing the optical
155: force near the
156: equilibrium: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
157: {f}} _{light} \approx
158: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
159: {K}} \Delta
160: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
161: {x}} $, where $\Delta
162: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
163: {x}} $ is the displacement vector of the \textit{i-th} particle away from its equilibrium
164: configuration. The vibration profile of the VEM is described by the
165: eigenvectors
166: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
167: {V}} ^{(i)}$
168: of $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
169: {K}} $, and the natural vibrational frequency is $\Omega _{0i} = ( - K_i /
170: m)^{1 / 2}$ where $K_{i}$ is an eigenvalue of
171: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
172: {K}} $ and $m$ is the mass of a sphere. Due to the reflection symmetry, the
173: modes fall into three separate branches (each of $N$ modes), corresponding
174: to the vibrations along the three Cartesian directions (see
175: Fig. \ref{fig1}(e)). We shall denote the branches as the
176: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
177: {k}} $-branch,
178: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
179: {E}} $-branch, and
180: $(\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
181: {k}} \times
182: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
183: {E}} )$-branch, corresponding respectively to particle displacements along
184: the incident
185: wavevector$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
186: {k}} = \pm k\hat {z}$, the incident polarization ($x$-axis), and the $y$-axis.
187:
188: The degree of localization of the modes can be quantified by calculating the
189: inverse participation ratio\cite{Book:1}
190:
191:
192: \begin{equation}
193: \label{eq3}
194: \left( {I.P.R.} \right)_i = \left( {\sum\nolimits_{n = 1}^N {\left| {\left(
195: {\Delta X_n^{(i)} ,\Delta Y_n^{(i)} ,\Delta Z_n^{(i)} } \right)} \right|^4}
196: } \right)^{ - 1},
197: \end{equation}
198:
199: \noindent
200: which indicates the number of particles participating the vibration. Here,
201: the index $i$ stands for the \textit{i-th} eigenmode and $\Delta X_n^{(i)} $ is the
202: vibration amplitude of the \textit{n-th} particle along the $x$-axis. A small value of
203: $I.P.R.$ indicates a localized mode, while $I.P.R.\sim N$ indicates a
204: delocalized mode. Fig. \ref{fig1} shows the $I.P.R.$
205: computed by the MS-MST formalism. For comparison, the $I.P.R.$ for an
206: ordinary ``ball and spring'' model is also plotted in
207: Fig. \ref{fig1}(d), where a lattice of 100 particles are
208: connected to its nearest neighbors by a Hooke spring. As expected, the ball
209: and spring model supports only propagating modes in which the displacement
210: of the \textit{n-th} particle $\sim e^{inq\Delta }$, where $q$ is the phonon wavevector and
211: $\Delta $ is the lattice constant. Depending on whether $q\Delta $ is an
212: integer multiple of $\pi$, $I.P.R.$ takes either $\sim $200/3 or $\sim $100.
213:
214:
215: In general, the VEMs of the optically-bound lattice are more localized than
216: the propagating modes, especially for the
217: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
218: {k}} $-branch. A few modes selected from the
219: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
220: {k}} $-branch is shown in Fig. \ref{fig2}. The
221: high-frequency modes are highly localized near the center of the lattice
222: (e.g. Fig. \ref{fig2}(c)), while those with a lower
223: vibrational frequency are less localized (e.g. Fig. \ref{fig2}(d)-(e)). For very low frequencies, the modes are further delocalized
224: spatially (e.g. Fig. \ref{fig2}(f)), with the vibration
225: being stronger on both ends. The evolution of a VEM as the number of
226: particles increases is also depicted in Fig. \ref{fig2}(a)-(c); clearly the overall profile of the modes are getting more and
227: more localized as the number of particle increases.
228:
229: The physics of the localized mode (LM) can be captured qualitatively by a
230: simple potential energy model (P.E. model)$_{.}$\cite{Lin:2005} For
231: small ($a \ll \lambda )$ lossless dielectric particles placed in a
232: standing wave of light, one may define an approximate potential energy for the
233: light-induced mechanical interaction as
234:
235: \begin{equation}
236: \label{eq4}
237: {\begin{array}{*{20}c}
238: {U = - \sum\limits_{n = 1}^N {\left( {\alpha / 4} \right)\vert
239: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
240: {E}} _{in}
241: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
242: {r}} _n )\vert ^2 - \alpha ^2 / 2} } \hfill \\
243: {\times \sum\limits_{n = 1}^N {\sum\limits_{m < n}
244: {\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
245: {E}} _{in}
246: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
247: {r}} _m )Re\left\{
248: {\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
249: {G}}
250: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
251: {r}} _n -
252: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
253: {r}} _m )} \right\}} }
254: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
255: {E}} _{in}
256: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
257: {r}} _n )} \hfill \\
258: \end{array} },
259: \end{equation}
260:
261: \noindent
262: where $\alpha = 4\pi \varepsilon _0 a^3(\varepsilon - 1) / (\varepsilon +
263: 2)$ and
264:
265: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
266: {G}}
267: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
268: {R}} ) = e^{ikR} / 4\pi \varepsilon _0 R^3\left[ {{\begin{array}{*{20}c}
269: {( - k^2R^2 - 3ikR + 3)\hat {R}\hat {R}^T} \hfill \\
270: { + (k^2R^2 + ikR -
271: 1)\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
272: {I}} } \hfill \\
273: \end{array} }} \right]$. To leading orders, the
274: force matrices for the three branches, evaluated using the P.E. model, are
275:
276: %$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
277: %{G}}
278: %(\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
279: %{R}} ) = e^{ikR} / 4\pi \varepsilon _0 R^3\left[ {\left( { - k^2R^2 - 3ikR +
280: %3} \right)\hat {R}\hat {R}^T + \left( {k^2R^2 + ikR - 1 }\right)
281: %\mbox{\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
282: %{I}} }} \right]$ is the electric dipole propagator. To leading orders, the
283: %force matrices for the three branches, evaluated using the P.E. model, are
284:
285: \begin{equation}
286: \label{eq5}
287: \begin{array}{l}
288:
289: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
290: {K}}
291: _{\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
292: {k}} - branch} )_{lq} = \\
293: \left\{ {{\begin{array}{*{20}c}
294: {K_{local} (l) - \beta \sum\nolimits_{n = 1,n \ne l}^N {\left( {\vert l -
295: n\vert \pi } \right)^{ - 1}} } \hfill & {\left( {l = q} \right)} \hfill \\
296: {\beta \left( {\vert l - q\vert \pi } \right)^{ - 1}} \hfill & {\left( {l
297: \ne q} \right)} \hfill \\
298: \end{array} }} \right. \\
299: \end{array},
300: \end{equation}
301:
302: \begin{equation}
303: \label{eq6}
304: \begin{array}{l}
305:
306: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
307: {K}}
308: _{(\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
309: {k}} \times
310: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
311: {E}} ) - branch} )_{lq} = \\
312: \left\{ {{\begin{array}{*{20}c}
313: { - \beta \sum\nolimits_{n = 1,n \ne l}^N {\left[ {\begin{array}{l}
314: 2\left( {\vert l - n\vert \pi } \right)^{ - 3} \\
315: - 3\left( {\vert l - n\vert \pi } \right)^{ - 5} \\
316: \end{array}} \right]} } \hfill & {\left( {l = q} \right)} \hfill \\
317: {\beta \left[ {2(\vert l - q\vert \pi )^{ - 3} - 3(\vert l - q\vert \pi )^{
318: - 5}} \right]} \hfill & {\left( {l \ne q} \right)} \hfill \\
319: \end{array} }} \right. \\
320: \end{array},
321: \end{equation}
322:
323:
324:
325: \noindent
326: and
327:
328:
329: \begin{equation}
330: \label{eq7}
331: \begin{array}{l}
332:
333: (\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over
334: {K}}
335: _{\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
336: {E}} - branch} )_{lq} = \\
337: \left\{ {{\begin{array}{*{20}c}
338: { - \beta \sum\nolimits_{n = 1,n \ne l}^N {\left[ {\begin{array}{l}
339: 4\left( {\vert l - n\vert \pi } \right)^{ - 3} \\
340: - 9\left( {\vert l - n\vert \pi } \right)^{ - 5} \\
341: \end{array}} \right]} } \hfill & {\left( {l = q} \right)} \hfill \\
342: {\beta \left[ {4(\vert l - q\vert \pi )^{ - 3} - 9(\vert l - q\vert \pi )^{
343: - 5}} \right]} \hfill & {\left( {l \ne q} \right)} \hfill \\
344: \end{array} }} \right. \\
345: \end{array},
346: \end{equation}
347:
348: \noindent
349: where
350:
351: $K_{local} (l) = - 2k^2\alpha E_0^2 - \beta \sum\nolimits_{n = 1,n \ne l}^N
352: {\left( {\vert l - n\vert \pi } \right)^{ - 1}} ,$
353:
354:
355:
356: \noindent
357: $\beta = k^5\alpha ^2E_0^2 / 2\pi \varepsilon _0 ,$ and $l$ and $q$ are particle
358: indices. The $I.P.R.$ computed using the P.E. model is plotted in
359: Fig. \ref{fig1} as dotted lines, which are surely not quantitative compared
360: with the exact result, but nevertheless captures the salient features of the rigorous
361: calculations.
362:
363: It is evident from (\ref{eq6}) and (\ref{eq7}) that the modes of the
364: $(\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
365: {k}} \times
366: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
367: {E}} )$-branch and the
368: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
369: {E}} $-branch are similar, because the leading terms are essentially an
370: action-reaction couplings between every pair of particles, with the coupling
371: strength being proportional to inverse-cubic distance. These two branches
372: are more localized than those of the ball and spring model because the
373: interaction has a longer range.\cite{The:1976} The
374: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
375: {k}} $-branch is the most localized and interesting. Its force matrix consists
376: of two components, the long range (inverse distance) action-reaction
377: coupling and $K_{local} (l)$ which acts like a spring that
378: ties the \textit{l-th} particle to its equilibrium position. The first term of $K_{local}
379: (l)$ is caused by the incident beam and is the same for each particle. This
380: term gives a frequency gap at low frequency (e.g. between 0 and 4.7 MHz in
381: Fig. \ref{fig1}(a)), while the second term is induced by OB.
382: One may define an intrinsic vibration frequency for every individual
383: particle as
384:
385:
386: \begin{equation}
387: \label{eq9}
388: \Omega _{\mbox{intrinsic}} (l) = \sqrt { - K_{local} (l) / m} ,
389: \end{equation}
390:
391:
392:
393: \noindent
394: plotted in Fig. \ref{fig2}(g). We note that the first term of
395: $K_{local} (l)$ contributes a constant to $\Omega _{\mbox{intrinsic}} (l)$,
396: while the term due to OB gives a position dependent contribution that makes
397: $\Omega _{\mbox{intrinsic}} (l)$ higher (lower) near the center (ends) of
398: the lattice. It is the variation of $\Omega _{\mbox{intrinsic}} (l)$ along
399: the chain that elicits the enhanced localization: only particles near the
400: center (both ends) participate in the high (low) frequency vibrations, see
401: Fig. \ref{fig2}(c) (Fig. \ref{fig2}(f)).
402:
403: We now consider the strength of the OB. As revealed by recent
404: theoretical\cite{Lin:2005}$^{,}$\cite{Antonoyiannakis:1997} and
405: experimental\cite{Burns:1990}$^{,}$\cite{Tatarkova:2002}$^{,}$\cite{Singer:2003}$^{,}$\cite{Garces:2005}$^{,}$\cite{Mellor:2006}
406: works, the optical force on microspheres can dominate over other relevant
407: interactions such as gravity, van der Waals, and thermal fluctuations. For
408: the lattice consists of smaller spheres defined in (\ref{eq2}), the potential energy
409: per particle $U / N$ for $N$=1, 10, 50, 100 are respectively -9.6, -10.5,
410: -11.3, -11.7 $k_{B}T_{Room}$, and the chain should thus be thermally stable
411: at the assumed intensity. Furthermore, $U / N$ is enhanced by more than
412: 20{\%} as $N$ is increased from 1 to 100, implying that the OB carries a
413: non-negligible contribution.
414:
415: We have showed that OB can bind a collection of particles into a 1D lattice
416: that is stable in all three dimensions. We shall emphasize that the
417: localization discussed here is a general phenomena for optically-bound
418: structures that are spatially extended, and it is not restricted to the
419: particular geometry or incident wave considered here. We found that LMs are
420: also observed in other structures such as the photonic cluster made from
421: microspheres shown in Fig. 4(\textbf{g}) of reference
422: \onlinecite{Lin:2005}, and also another lattice configuration defined by
423: ${\begin{array}{*{20}c}
424:
425: {\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
426: {R}} _n = \left( {0,n\lambda ,0} \right),} \hfill & {n = 1,2\ldots ,N}
427: \hfill \\
428: \end{array} }.$ A difference between this lattice configuration and (\ref{eq2}) is
429: that the lattice constant of the later (former) is dictated by optical
430: binding (trapping). It is the long-ranged OB that induces the variation
431: of $\Omega _{\mbox{intrinsic}} (l)$, which in turn induce the localization.
432:
433: It is worth to note that in the case of the 1D array specified by (\ref{eq2}), the
434: stable configuration is defined by optical traps produced by the incident
435: wave rather than OB, yet OB plays a crucial role on the dynamics. The
436: quasi-stable dynamics that arises from the nonconservative nature of the
437: optical forces,\cite{Lin:2005} and the LMs considered here,
438: could be major causes of the inconsistencies between the vast amount of
439: light-trapping experiments and theoretical predictions where OB is
440: neglected.\cite{See:2003} A deeper investigation into the subject
441: would be an interesting and important research topic for the future.
442:
443: Support by CA02/03.SC05 is gratefully acknowledged. We thank Kin-Hung Fung
444: for useful discussions. C.T. Chan's e-mail address is phchan@ust.hk.
445:
446:
447: \begin{figure}[htbp]
448: \centerline{\includegraphics[width=3.34in]{LocalizedModes1.eps}}
449: \caption{(Color online) Natural vibration frequencies $\Omega _{0i} $ versus the
450: inverse participation ratio, for the 1D lattice with $N=100$. Panel (a), (b),
451: and (c) correspond respectively to
452: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
453: {k}} $-branch,
454: $(\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
455: {k}} \times
456: \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
457: {E}} )$-branch, and
458: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
459: {E}} $-branch. The open circles are obtained by the rigorous MS-MST
460: formalism and the dotted line is that of the P.E. model (\ref{eq4}). Panel (d) shows
461: the $I.P.R.$ for the ``ball and spring'' model. Panel (e) shows
462: schematically the direction of the particles' displacements for the three
463: branches.}
464: \label{fig1}
465: \end{figure}
466:
467: \begin{figure}[htbp]
468: \centerline{\includegraphics[width=3.34in]{LocalizedModes2.eps}}
469: \caption{(Color online) The profiles of a few selected VEMs in the
470: $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\rightharpoonup$}}\over
471: {k}} $-branch. Dashed lines show the equilibrium positions. Each circle
472: represents one particle. Panel (a)-(c): the highest frequency mode for a
473: lattice containing (a): $N=10$ particles, (b): $N=50$, and (c) $N=100$. Panel
474: (c)-(f): the VEMs for $N=100$, with (c) showing the highest
475: frequency mode, and (d)-(e) correspond to two intermediate frequencies, and
476: (f) shows the lowest frequency mode. Note that the interparticle separation
477: and the size of the particles are not drawn to scale. Panel (g) shows
478: $\Omega _{\mbox{intrinsic}} $ with $N=100$, see (\ref{eq9}).}
479: \label{fig2}
480: \end{figure}
481:
482: \begin{thebibliography}{16}
483: \bibitem{Ashkin:1970} A. Ashkin, Phys. Rev. Lett. \textbf{24}, 156 (1970).
484: \bibitem{See:2003} See e.g. D.G. Grier, Nature \textbf{424}, 810 (2003).
485: \bibitem{Burns:1990} M.M. Burns, J.M. Fournier, and J.A. Golovchenko, Science \textbf{249}, 749 (1990).
486: \bibitem{Lin:2005} J. Ng, Z.F. Lin, C.T. Chan, and P. Sheng, Phys. Rev. B \textbf{72}, 085130 (2005); \textit{ibid}, \textit{Opt. Lett.} \textbf{30}, 1956 (2005).
487: \bibitem{Chaumet:2001} P.C. Chaumet and M. Nieto-Vesperinas, Phys. Rev. B \textbf{64}, 035422 (2001).
488: \bibitem{Campbell:2004} D.K. Campbell, S. Flach, and Y.S. Kivshar, Physics Today \textbf{57}, \textit{No.} 1, 43 (2004).
489: \bibitem{Tatarkova:2002} S.A. Tatarkova, A.E. Carruthers, and K. Dholakia, Phys. Rev. Lett. \textbf{89}, 283901 (2002).
490: \bibitem{Singer:2003} W. Singer, M. Frick, S. Bernet, and M. Ritsch-Marte, J. Opt. Soc. Am. B \textbf{20}, 1568 (2003).
491: \bibitem{Black:2003} A.T. Black, Hilton, W. Chan, and V. Vuletic, Phys. Rev. Lett. \textbf{91}, 203001 (2003).
492: \bibitem{Garces:2005} V. Garces-Chavez and K. Dholakia, Appl. Phys. Lett. \textbf{86}, 031106 (2005).
493: \bibitem{Mellor:2006} C.D. Mellor and C.D. Bain, ChemPhysChem \textbf{7}, 329 (2006).
494: \bibitem{The:1} The stable configurations calculated by the MS-MST formalism deviate from (\ref{eq2}) by less than 0.003$\lambda$.
495: \bibitem{This:1} This is so because, on every sphere, the path difference between the incident field and the scattered field from the other spheres, are roughly an integer multiple of 2$\pi $, which enhances the stability.
496: \bibitem{Chowdhury:1985} A. Chowdhury and B. Ackerson, Phys. Rev. Lett. \textbf{55}, 833 (1985).
497: \bibitem{Book:1} N.E. Cusack, The Physics of Structurally Disordered Matter: An Introduction (A. Hilger, Philadelphia, 1987), p. 239.
498: \bibitem{The:1976} The finite coherent length of real laser will effectively set an upper limit on \textit{N}. In the hypocritical case where $N \to \infty $, the modes for (\ref{eq6})-(\ref{eq7}) become extended modes, whereas (\ref{eq5}) diverges.
499: \bibitem{Antonoyiannakis:1997} M.I. Antonoyiannakis and J.B. Pendry, Phys. Rev. B \textbf{60}, 613 (1997).
500: \end{thebibliography}
501:
502: \end{document}