cond-mat0107192/w.tex
1: %%%%%%%%%%%% First-Principles Approach to Electrorotation Assay %%%%%%%%%%%
2: % =========================================================================
3: % Dr. K. W. Yu
4: % THE CHINESE UNIVERSITY OF HONG KONG
5: % Department of Physics, Shatin, New Territories, Hong Kong
6: % email: kwyu@phy.cuhk.edu.hk  phone: (852) 2609 6349  fax: (852) 2603 5204
7: % =========================================================================
8: %%%%%%%%%%%% First-Principles Approach to Electrorotation Assay %%%%%%%%%%%
9: % w.tex (written 12 May 2001; last updated 10 December 2001)
10: \documentstyle[preprint,aps,epsfig]{revtex}
11: 
12: \def\be{\begin{equation}}
13: \def\ee{\end{equation}}
14: \def\ep{\epsilon}
15: \def\t{\tilde}
16: \def\sg{\sigma}
17: \def\et{\tilde{\epsilon}}
18: 
19: \begin{document}
20: \draft
21: \title{First-Principles Approach to Electrorotation Assay}
22: \author{J. P. Huang and K. W. Yu}
23: \address{Department of Physics, The Chinese University of Hong Kong,
24:  Shatin, NT, Hong Kong}
25: \maketitle
26: 
27: \begin{abstract}
28: We have presented a theoretical study of electrorotation assay based on 
29: the spectral representation theory. We consider unshelled and shelled 
30: spheroidal particles as an extension to spherical ones.
31: From the theoretical analysis, we find that the coating can change the 
32: characteristic frequency at which the maximum rotational angular velocity 
33: occurs.
34: The shift in the characteristic frequency is attributed to a change in the 
35: dielectric properties of the bead-coating complex with respect to those of 
36: the unshelled particles. 
37: By adjusting the dielectric properties and the thickness of the coating, 
38: it is possible to obtain good agreement between our theoretical 
39: predictions and the assay data.
40: 
41: \end{abstract}
42: \vskip 5mm \pacs{PACS Number(s): 82.70.-y, 87.22.Bt, 77.22.Gm, 77.84.Nh}
43: 
44: \section{Introduction}
45: 
46: When a suspension of colloidal particles or biological cells is exposed to 
47: an external electric field, the analysis of the frequency-dependent response
48: yields valuable information on various processes, like the structural 
49: (Maxwell-Wagner) polarization effects \cite{Gimsa,Gimsa99}.
50: The polarization is characterized by a variety of the characteristic 
51: frequency-dependent changes.
52: While the polarization of biological cells can be investigated by the method
53: of dielectric spectroscopy \cite{Asami80}, conventional dielectrophoresis
54: and electrorotation (ER) analyze the frequency dependence of translations 
55: and rotations of single cells in an inhomogeneous and rotating external
56: field, respectively \cite{Fuhr,Gimsa91}.
57: With the recent advent of experimental techniques such as automated video 
58: analysis \cite{Gasperis} as well as light scattering methods \cite{Gimsa99}, 
59: the cell movements can be accurately monitored.
60: In ER, the frequency dependence of the polarization leads to a phase shift 
61: between the induced dipole moment and the rotating field, giving rise to a
62: torque acting on the particle which causes the rotation of the individual 
63: cell.
64: 
65: The phenomenon of ER can be developed into a useful technique known as 
66: ER assay. The ER assay combines antibody technology with ER to detect 
67: analytes in aqueous solutions. 
68: The analyte to be detected is bound to a latex bead of known dielectric
69: properties to form the analyte-bead complex, which causes a change in the 
70: dielectric properties. 
71: The change can be detected by ER technique, thus allowing the rapid and 
72: accurate detection of analytes in aqueous solutions.
73: This method can be used to detect various analytes, the selection of 
74: which can be controlled by the proper choice of binding agents. 
75: 
76: In this work, we propose the use of the spectral representation 
77: \cite{Bergman} for analyzing the electrorotation of particles in 
78: suspensions.
79: The spectral representation is a rigorous mathematical formalism of 
80: the effective dielectric constant of a two-phase composite material 
81: \cite{Bergman}. 
82: It offers the advantage of the separation of material parameters 
83: (namely the dielectric constant and conductivity) from the cell
84: structure information, thus simplifying the study. 
85: From the spectral representation, one can readily derive the dielectric 
86: dispersion spectrum, with the dispersion strength as well as the 
87: characteristic frequency being explicitly expressed in terms of the 
88: structure parameters and the materials parameters of the cell suspension 
89: (see section II.B below). 
90: The actual shape of the real and imaginary parts of the permittivity
91: over the relaxation region can be uniquely determined by the Debye relaxation
92: spectrum, parametrized by the characteristic frequencies and the dispersion
93: strengths. So, we can study the impact of these parameters on the dispersion
94: spectrum directly. The same formalism has been used recently to study the 
95: dielectric behavior of cell suspensions \cite{Lei}.
96: 
97: In this connection, we mention alternative methods, namely, by solving 
98: the Laplace's equation directly and Gimsa's approach based on equivalent 
99: circuits \cite{Gimsa,Gimsa99}. To our knowledge, none of these methods 
100: separate microstructure parameters from material information.
101: 
102: \section{FORMALISM}
103: 
104: We regard a suspension as a composite system consisting of 
105: spherical or spheroidal particles of complex dielectric constant 
106: $\t{\ep}_1$ dispersed in a host medium of $\t{\ep}_2$.
107: A uniform electric field ${\bf E}_0=E_0 \hat{\bf z}$ is applied to the
108: composites along the $z$-axis. 
109: We briefly review the spectral representation theory of the effective 
110: dielectric constant to establish notations.
111: 
112:  \subsection{Spectral representation}
113: 
114: The spectral representation is a mathematical transformation of the
115: complex effective dielectric constant $\t{\ep}_e$. 
116: In its original form \cite{Bergman}, a two-phase composite material is 
117: considered, in which inclusions of complex dielectric constant 
118: $\t{\ep}_1$ and volume fraction $p$ are randomly embedded in a host 
119: medium of $\t{\ep}_2$.
120: The complex effective dielectric constant $\t{\ep}_e$ will in general 
121: depend on the constituent dielectric constants, the volume fraction 
122: of inclusions as well as the detailed microstructure of the 
123: composite materials.
124: 
125: The essence of the spectral representation is to define the following 
126: transformations. If we denote a complex material parameter
127:  \be
128: \t{s}=\left(1 - {\t{\ep}_1\over \t{\ep}_2} \right)^{-1},
129:  \ee%1
130: then the reduced effective dielectric constant
131:  \be
132: w(\t{s})=1 - {\t{\ep}_e\over \t{\ep}_2}, 
133:  \ee%2
134: can be written as
135:  \be
136: w(\t{s}) = \sum_n {F_n \over \t{s}-s_n},
137: \label{Fs}
138:  \ee%3
139: where $n$ is a positive integer, i.e., $n=1, 2, ...$, and $F_n$ and $s_n$, 
140: are the $n$-th microstructure parameters of the composite materials 
141: \cite{Bergman}. In Eq.(\ref{Fs}), $0\le s_n<1$ is a real number,
142: while $F_n$ satisfies a sum rule \cite{Bergman}:
143:  \be
144: \sum_n F_n=p,
145:  \ee%4
146: where $p$ is the volume fraction of the suspended cells.
147: 
148: As a result, the spectral representation is a useful theory which helps 
149: separate the material property from the geometric information. 
150: In what follows, we illustrate the spectral representation by the 
151: capacitance of simple geometry. 
152: In particular, a parallel-plate capacitor is considered as an example. 
153: We will discuss two cases, namely, the series combination and the 
154: parallel combination.
155: 
156: In the first case, if one inserts a dielectric slab of dielectric constant 
157: $\t{\ep}_1$ and thickness $h_1$, as well as a dielectric of $\t{\ep}_2$ and 
158: thickness $h_2$ (both of the same area $A$), 
159: into a parallel-plate capacitor of total thickness $h=h_1+h_2$, 
160: the overall capacitance $C$ is given by
161: $$
162: C^{-1}=C_1^{-1}+C_2^{-1}
163: $$
164: where $C_1=\t{\ep}_1 A/h_1$, and $C_2=\t{\ep}_2 A/h_2$, $A$ being the area of a 
165: plate. 
166: On the other hand, we may define the equivalent capacitance as 
167: $C=\t{\ep}_e A/h$, where $\t{\ep}_e$ is the effective dielectric constant.
168: That is, we replace the composite dielectric by a homogeneous dielectric
169: of dielectric constant $\t{\ep}_e$.
170: 
171: Let $\t{\ep}_1=\t{\ep}_2(1-1/\t{s})$, we can express $C$ in the spectral 
172: representation, 
173: $$
174: C=\frac{A \t{\ep}_2}{h}-\frac{A\t{\ep}_2 h_1/h^2}{\t{s}-h_2/h}.
175: $$
176: In accord with the spectral representation, one may introduce 
177: $w(\t{s})=1-\t{\ep}_e/\t{\ep}_2$, which is in fact the same as $w(\t{s})=1-C/C_0$, 
178: where $C_0$ is the capacitance when the plates are all filled with a 
179: dielectric material of $\t{\ep}_2$, namely, $C_0=\t{\ep}_2 A/h$. 
180: Thus we obtain
181: $$
182: w(\t{s})=\frac{h_1/h}{s-h_2/h}.
183: $$
184: from which we find that the material parameter is separated from the 
185: geometric parameter.
186: The comparison of $w(\t{s})$ with Eq.(\ref{Fs}) yields
187:  $$
188:  F_1=h_1/h,\ \ \ s_1=h_2/h.
189:  $$
190: We should remark that $F_1$ obtained herein is just equal to the 
191: volume fraction of the dielectric of $\t{\ep}_1$, and that $s_1$ satisfies 
192: $0\le s_1<1$, as required by the spectral representation theory.
193:   
194: Next we consider the parallel combination.
195: If one inserts a material of dielectric constant $\t{\ep}_1$ and area 
196: $w_1$ as well as a dielectric of $\t{\ep}_2$ and area $w_2$
197: (both of the same thickness $h$), 
198: into a parallel-plate capacitor of total area $A=w_1+w_2$, 
199: the overall capacitance $C$ is given by
200: $$
201: C=C_1+C_2
202: $$
203: where $C_1=\t{\ep}_1 w_1/h$, and $\t{\ep}_2=\t{\ep}_2 w_2/h$. Similarly, after
204: introducing the effective dielectric constant $\t{\ep}_e$, we may define 
205: the overall capacitance as $C=\t{\ep}_e A/h$.
206:    
207: Again, in the spectral representation, let $\t{\ep}_1=\t{\ep}_2(1-1/\t{s})$, then 
208: $$
209: C=\frac{\t{\ep}_2 A}{h}-\frac{\t{\ep}_2 w_1}{h \t{s}}.
210: $$
211: Writing $w(\t{s})=1-C/C_0$, we obtain
212: $$
213: w(\t{s})=\frac{w_1/A}{\t{s}}.
214: $$
215: From this equation, the material parameter is also found to be separated 
216: from the geometric parameter.
217: It is clear that $F_1=w_1/A$, i.e., the volume fraction of the 
218: dielectric of $\t{\ep}_1$, and $s_1=0$. 
219: 
220:  \subsection{Shell-spheroidal model}
221:  
222: For inclusions of arbitrary shape, the spectral representation can only be
223: solved numerically \cite{Bergman}. However, analytic solutions can be 
224: obtained for isolated spherical and ellipsoidal particles.
225: For dilute suspensions of prolate spheroidal particles, the particles 
226: can be regarded as noninteracting.
227: The problem is simplified to the calculation of $s_n$ with a single 
228: particle, which can be solved exactly.
229:  
230: In fact, the suspension of shell-spheroidal particles dispersed in a host 
231: medium is a three-component system. Although the spectral representation is 
232: generally valid for two-component composites, we have recently shown that 
233: it also applies to composites of shelled spheres randomly embedded in a host 
234: medium \cite{Yuen}.
235: Similarly, we will show that the spectral representation also applies to 
236: the suspension of spheroidal particles of complex dielectric constant 
237: $\t{\ep}_1$ coated with a shell of $\t{\ep}_s$ dispersed in a host 
238: medium of $\t{\ep}_2$, where
239:  \be
240:  \t{\ep}=\ep+\sigma/i2\pi f\nonumber
241:  \ee %5
242: where $f$ is the frequency of the applied field. In what follows, we will 
243: show that from the spectral representation, one can obtain the analytic 
244: expressions for the characteristic frequency at which the maximum 
245: electrorotation velocity occurs.
246: The depolarization factors of the spheroidal particles will be described by 
247: a sum rule
248:  \be
249:  L_z+2L_{xy}=1\nonumber
250:  \ee %6
251: where $0<L_z\le 1/3$ and $L_{xy}$ are the depolarization factors along the 
252: $z$- and $x$- (or $y$-) axes of the spheroidal particle, respectively. 
253: In fact, $L_z=L_{xy}=1/3$ just indicates a spherical particle.
254: 
255: The phenomenon of electrorotation is based on the interaction between a 
256: time-varying electric field ${\bf E}$ and the induced dipole moment 
257: ${\bf M}$. The dipole moment of the particle arises from the induced 
258: charges that accumulate at the interface of the particle.
259: As the prolate spheroidal particles are easily oriented along their long 
260: axes by the rotating field, we consider the orientation in which
261: the long axis lies within the field plane \cite{Gimsa2001}.
262: In fact, the extension of our theory to deal with oblate spheroid is 
263: straightforward. In doing so, it suffices to consider $1/3<L_z<1$ and
264: the formlism will remain unchanged.
265: 
266: The angle between ${\bf M}$ and ${\bf E}$ is denoted by $\theta$, where 
267: $\theta=\omega \times$ time and $\omega=2\pi f$ is angular velocity of 
268: the rotating electric field. 
269: The torque acting on the particle is given by the vector cross product 
270: between the electric field and the dipole moment, so that only the 
271: imaginary part of the dipole moment contributes to the electrorotation 
272: response.
273: In the steady state, the frequency-dependent rotation speed $\Omega(f)$, 
274: which results from the balance between the torque and the viscous drag, 
275: is given by
276:    \begin{eqnarray}
277: \Omega(f) &=& -F(\ep_2, E,\eta)Im(\t{b}_z \langle\cos^2\theta \rangle
278:  + \t{b}_{xy} \langle\sin^2 \theta\rangle), \nonumber \\
279:   &=& -F(\ep_2, E,\eta)Im(\t{b}_z/2
280:  + \t{b}_{xy}/2), 
281:    \end{eqnarray}%7
282: where $E$ is the strength of the applied electric field and $F$ is a 
283: coefficient which is inversely proportional to the dynamic viscosity 
284: $\eta$ of the host medium. For spherical cells, $F=-\ep_2 E^2/2\eta$.
285: In Eq.(7), $Im(\cdots)$ indicates the imaginary parts of $(\cdots)$, and 
286: the angular brackets denote a time average. 
287: Since the angular velocity of the rotating field is much greater than the 
288: electrorotation angular velocity, i.e., $\omega \gg \Omega$, the time 
289: averages are just equal to 1/2. For a single shelled spheroidal particle, 
290: the dipole factor $\t{b}_z$ is given by \cite{Gao}
291:  \be
292: \t{b}_z=\frac{1}{3}\frac{(\t{\ep}_s-\t{\ep}_2)
293:  [\et_s+L_z(\et_1-\et_s)]+(\et_1-\et_s)y[\et_s+L_z(\et_2-\et_s)]}
294:  {(\et_s-\et_1)(\et_2-\et_s)yL_z(1-L_z)+[\et_s+(\et_1-\et_s)L_z]
295:  [\et_2+(\et_s-\et_2)L_z]}
296:  \ee %8
297: where $y$ is the volume ratio of the core to the whole shelled spheroid, 
298: while $b_{xy}$ can be obtained by replacing $L_z$ with $L_{xy}$ in Eq.(8).
299:   
300: We are now in a position to represent $b_z$ and $b_{xy}$ in the spectral 
301: representation. Let $\et_1=\et_2(1-1/\t{s})$, and assume
302: $x=\et_s/\et_2$, we obtain
303:    \be
304:    \t{b}_z=N.P.+\frac{F_1}{\t{s}-s_1}
305:    \ee %9
306: where $N.P.$ denotes the nonresonant part \cite{Yuen} which vanishes in the 
307: limit of unshelled spheroidal cells. In Eq.(9), the various quantities 
308: are given by
309:  \begin{eqnarray}
310:  s_1&=&-\frac{\beta}{\gamma},\\
311:  F_1&=&\frac{-x^2y}{\alpha \gamma},\\
312:  N.P.&=&\frac{-L_z+L_zy+x(-1+2L_z+y-2L_zy)+x^2(1-L_z-y+L_zy)}{\alpha}\\
313: \alpha&=&3L_z-3L_z^2-3L_zy+3L_z^2y+x(3-6L_z+6L_z^2+6L_zy-6L_z^2y)\nonumber\\
314:       & &x^2(3L_z-3L_z^2-3L_zy+3L_z^2y)\nonumber \\
315: \beta&=&-L_z+L_z^2+L_zy-L_z^2y+x(-L_z^2-L_zy+L_z^2y)\nonumber\\
316: \gamma&=&L_z-L_z^2-L_zy+L_z^2y+x(1-2L_z+2L_z^2+2L_zy-2L_z^2y)\nonumber\\
317:       & &+x^2(L_z-L_z^2-L_zy+L_z^2y)\nonumber
318:  \end{eqnarray}%10--12
319:  
320: Note that we have assumed $x$ to be a real number, which will be 
321: justified below. 
322: After substituting $\et=\ep+\sg/i2\pi f$ into Eq.[9], we rewrite $\t{b}_z$ 
323: after simple manipulations
324:  \be
325:  \t{b}_z=(N.P.+\frac{F_1}{s-s_1})+\frac{\delta\ep_1}{1+if/f_{c1}}
326:  \ee
327: with $s=(1-\ep_1/\ep_2)^{-1}$ and $t=(1-\sg_1/\sg_2)^{-1}$, where
328:  \begin{eqnarray}
329:  \delta\ep_1&=&F_1\frac{s-t}{(t-s_1)(s-s_1)},\nonumber\\
330: f_{c1}&=&\frac{1}{2\pi}\frac{\sg_2}{\ep_2}\frac{s(t-s_1)}{t(s-s_1)}.\nonumber
331:  \end{eqnarray}
332: Similarly, we may rewrite $\t{b}_{xy}$ as 
333:  \be
334:  \t{b}_{xy}=(N.P.'+\frac{F_2}{s-s_2})+\frac{\delta\ep_2}{1+if/f_{c2}}.
335:  \ee %14
336:  Therefore, we obtain
337:  
338:  \begin{eqnarray}
339:  \delta\ep_2&=&F_2\frac{s-t}{(t-s_2)(s-s_2)},\nonumber\\
340:  f_{c2}&=&\frac{1}{2\pi}\frac{\sg_2}{\ep_2}\frac{s(t-s_2)}{t(s-s_2)}.\nonumber
341:  \end{eqnarray}
342: Note that $N.P.'$, $s_2$ and $F_2$ are obtained by replacing $L_z$ with 
343: $L_{xy}$ in the expressions for $N.P.$, $s_1$ and $F_1$, respectively.
344: 
345: We have thus predicted that two characteristic frequencies may appear for 
346: unshelled or shelled spheroidal particles.
347: Previous theories were often limited to spherical particles,
348: i.e., $L_z=L_{xy}=1/3$. 
349: Therefore, only one characteristic frequency exists.
350: We should remark that even though two characteristic frequencies are 
351: predicted, only one of them is dominant (see below).
352:  
353:  \section{numerical calculations}
354: 
355: The model put forward in the previous section applies to various 
356: situations such as biological cells and polystyrene beads, etc. 
357: Here we perform numerical calculations to investigate the characteristic 
358: frequency. Let $s=1.1$, $t=-0.005$, and $\ep_2=80\ep_0$, where $\ep_0$ is 
359: the dielectric constant of the vacuum.
360: In Fig.1, we investigate the effect of particle shape on $s_1$ and 
361: $f_{c1}$ (upper panels), and $s_2$ and $f_{c2}$ (lower panels), for $z=2$ 
362: and $\sg_2=2.9 \times 10^{-5} Sm^{-1}$, where
363: $z=1/y$, and $z^{1/3}>1$ reflects the thickness 
364: of the shell (or coating). As is evident from the figure, an increase
365: in the dielectric constant ratio $x$ leads to a red-shift of the
366: characteristic frequency. 
367: For a certain $x$, a small depolarization factor, i.e., the particle is 
368: largely deviated from spherical shape, may yield a red-shift too.
369: 
370: In Fig.2, we investigate the effect of the depolarization factor on the 
371: quantities $-Im[b_z]/2$ and $-Im[b_z/2+b_{xy}/2]$. 
372: It is evident that the effect of $b_{xy}$ on the peak is small as $L_z$ is 
373: small, whereas it gives a large effect as $L_z$ is large. 
374: Generally speaking, the dipole moment along the x-(or y-) axis strongly 
375: affects both the location and the magnitude of the peak of rotation speed.
376: For spheroidal particles, only one peak is found, and the other peak 
377: predicted by the theory may be too small to observe.
378: 
379: In order to validate our theory, here we considered the spherical particles 
380: as a limiting case of our model.
381: In Fig.3, $s_1$ and $f_c$ are plotted versus $x$ for a spherical bead 
382: ($L_z=1/3$), for different $z$ as $\sg_2=2.9\times 10^{-5} Sm^{-1}$. 
383: For large $x$, the shell thickness has only a minor effect on the 
384: characteristic frequency.
385: Moreover, a thick shell leads to a red shift (blue shift) of the 
386: characteristic frequency appearing when $x>1$ ($x<1$).
387: Also, all the $f_c$ predicted by different $z$ are smaller (larger) than 
388: that predicted one at $z=1$ (i.e., unshelled bead) for $x>1$ ($x<1$).
389: 
390: At $x=1$, i.e., the shell has the same dielectric constant as the host, 
391: all the characteristic frequencies predicted by different thickness of 
392: shell are the same. Similar conclusions can be obtained not only for 
393: spherical shape, but also for prolate spheroidal shape 
394: (not shown here).
395: 
396: We attempt to fit our theoretical predictions with experimental data,
397: which are extracted from an assay \cite{Burt}. 
398: In this assay, three cases were studied, all dealing with spherical 
399: particles: that is, unshelled beads, 
400: beads coated with an antibody with specificity for Giardia (Shell 1),
401: and beads coated with an antibody with specificity for Cryptosporidium 
402: (Shell 2).
403: The bead diameter is $6 \mu m$ according to Burt et al \cite{Burt}, 
404: while Giardia and Cryptosporidium are both $0.8\mu m$ in diameter. 
405: Let $\sg_2=2.18 \times 10^{-5} Sm^{-1}$, $z=6$, $x=7.06$ (Shell 1) and 
406: $4.63$ (Shell 2), and $F=0.353$ (Without shell), $1.629$ (Shell 1) and 
407: $2.278$(Shell 2). Good agreement between our theoretical predictions 
408: and the assay data is shown in Fig.4.
409: From our theory, it is easy to find the corresponding characteristic 
410: frequencies at which maximum rotational angular velocity occurs, 
411: $f_c=4.755\times 10^5Hz$ (Without shell), $10^5Hz$ (Shell 1) and 
412: $1.415\times 10^5Hz$ (Shell 2).
413: From this, we find that the coating leads to a red shift of the 
414: characteristic frequency.
415: It is because the dielectric properties of the bead-coating complex have 
416: been changed. 
417: 
418: We used $x>1$ in our fitting, i.e., $x=7.06$ and 4.63 for bead coated 
419: with an antibody with specificity for Giardia and Cryptosporidium, 
420: respectively. 
421: It is known that, for biological cells, such 
422: as Giardia and Cryptosporidium, they have structures such as the cell 
423: wall, plasma membrane and the cytoplasm, among which the cell wall has 
424: a larger conductivity than the suspending medium. Hence, we may safely 
425: take $x>1$. 
426: 
427: \section{Discussion and conclusion}
428: 
429: Here we would like to add some comments.
430: We would like to clarify the assumptions with our model in more detail. 
431: In fact, there is only one (rather than two) peak for each polarization 
432: in our theory. It is because we assumed the ratio of the shell to host 
433: dielectric constant $x$ to be a real and positive number. 
434: If we had retained the (indeed small) imaginary part of $x$ in our 
435: calculation, then we would have two peaks for each polarization. 
436: The conductivity-dominated peak would have occured at substantially lower
437: frequency. Thus, the neglect of the imaginary part of $x$ is to drop the
438: lower frequency peak.
439: Moreover, according to our calculations, there is one dominant pole 
440: associated with $b_z$ and two (degenerate) sub-dominant poles associated 
441: with $b_{xy}$ in the spectral representation. 
442: Thus, for shelled particles in the present work, only one peak has been 
443: shown.
444: 
445: We developed simple equations to describe the electrorotation of 
446: particles in a suspension from the spectral representation. 
447: These equations serve as a basis which describe the parameter dependence 
448: of the polarization and thereby enhances the applicability of various 
449: cell models for the analysis of the polarization mechanisms. 
450: In this connection, the shell-spheroidal cell model may readily be 
451: extended to multi-shell cell model.
452: However, we believe that the multi-shell nature of the cell may have 
453: a minor effect on the electrorotation spectrum.
454: 
455: We have considered the isolated cell case, which is a valid assumption
456: for low concentration of cells. However, for a higher concentration of 
457: cells, we should consider the mutual interaction between cells. For a
458: randomly dispersed cells suspension, we may replace the dielectric constant
459: of the host medium by the effective dielectric constant of the whole 
460: suspension.
461: 
462: When a strong rotating electric field is applied to a suspension,
463: the induced dipole moment will induce an overall attractive force between 
464: the polarized cells, leading to rapid formation of sheet-like structures 
465: in the plane of the rotating field. In reality there is a phase shift 
466: between the induced dipole moment of the structure and the applied field, 
467: and this can lead to electrorotation. However, the situation will be much 
468: more difficult than the single-cell case that we have studied because 
469: the many-body as well as multipolar interactions between the particles 
470: will produce a complicated electrorotation spectrum. 
471: Fortunately, our recently developed integral equation formalism \cite{Yu2000} 
472: can definitely help to solve for the Maxwell-Wagner relaxation spectrum.
473: 
474: In summary,
475: we have presented a theoretical study of electrorotation assay based on 
476: the spectral representation theory. We consider unshelled and shelled 
477: spheroidal particles as an extension to spherical ones.
478: From the theoretical analysis, we find that the coating can change the 
479: characteristic frequency at which maximum rotational angular velocity occurs.
480: By adjusting the dielectric properties and the thickness of the coating, 
481: it is possible to obtain good agreement between our theoretical 
482: predictions and the assay data.
483: 
484: \section*{Acknowledgments}
485: This work was supported by the Research Grants Council of the Hong Kong 
486: SAR Government under grant CUHK 4245/01P. 
487: J. P. H. is grateful to Dr. L. Gao and Dr. C. Xu for fruitful discussion.
488: K. W. Y. acknowledges the hospitality of Prof. Hong Sun when he visited 
489: the University of California at Berkeley and useful discussion with Prof. 
490: G. Q. Gu. 
491: 
492: \begin{references}
493: 
494: \bibitem{Gimsa} For a review, see J. Gimsa and D. Wachner,
495:  Biophys. J. {\bf 77}, 1316 (1999).
496: \bibitem{Gimsa99} J. Gimsa, Ann. NY Acad. Sci. {\bf 873}, 287 (1999).
497: \bibitem{Asami80} K. Asami, T. Hanai and N. Koizumi, Jpn. J. Appl. Phys.
498:  {\bf 19}, 359 (1980).
499: \bibitem{Fuhr} G. Fuhr, J. Gimsa and R. Glaser, Stud. Biophys. {\bf 108},
500:  149 (1985).
501: \bibitem{Gimsa91} J. Gimsa, P. Marszalek, U. Lowe and T. Y. Tsong, 
502:  Biophys. J. {\bf 73}, 3309 (1991).
503: \bibitem{Gasperis} G. De Gasperis, X.-B. Wang, J. Yang, F. F. Becker 
504:  and P. R. C. Gascoyne, Meas. Sci. Technol. {\bf 9}, 518 (1998).
505: 
506: \bibitem{Bergman} D. J. Bergman, Phys. Rep. {\bf 43}, 379 (1978).  
507: \bibitem{Lei} J. Lei, Jones T. K. Wan, K. W. Yu and Hong Sun,
508:  Phys. Rev. E {\bf 64}, 012903 (2001).
509: \bibitem{Yuen} K. P. Yuen and K. W. Yu, J. Phys.: Condens. Matter 
510:  {\bf 9}, 4669 (1997).
511: \bibitem{Gimsa2001} It is not always necessary that prolate particles
512:  orient with their longest axis in the field plane. 
513:  Depending on frequency, object and medium properties also a perpendicular 
514:  orientation can be observed, see e.g.,
515:  J. Gimsa, Bioelectrochemistry {\bf 54}, 23 (2001).
516: \bibitem{Gao} L. Gao, Jones T. K. Wan, K. W. Yu and Z. Y. Li, J. Phys.: 
517: Condens. Matter {\bf 12}, 6825 (2000).
518: \bibitem{Burt} J. P. H.  Burt, K. L. Chan, D. Dawson, A. Parton and 
519:  R. Pethig, Ann. Biol. Clin. {\bf 54}, 253 (1996); see also 
520:  http://www.ibmm.informatics.bangor.ac.uk/pages/science/rot.htm
521:  for the basic science of electrorotation.
522: 
523: \bibitem{Yu2000} K. W. Yu, Hong Sun and Jones T. K. Wan, 
524:  Physica B {\bf 279}, 78 (2000).
525: \end{references}
526: 
527: \begin{figure}[h]
528: \caption{Upper panels: $s1$ and $f_{c1}$ plotted against $x$ for 
529: different $L_z$ at $z=2$.
530: Lower panels: $s2$ and $f_{c2}$ plotted against $x$ for different 
531: $L_z$ at $z=2$.}
532: \end{figure}
533: 
534: \begin{figure}[h]
535: \caption{$-$Im$(b_z/2+b_{xy}/2)$ and $-$Im$b_z/2$ are plotted against
536: frequency $\omega$ for different $L_z$ at $z=6$, $x=2$ and 
537: $\sigma_2=2.9 \times 10^{-5}Sm^{-1}$.}
538: \end{figure}
539: 
540: \begin{figure}[h]
541: \caption{$s1$ and $f_c$ are plotted versus $x$ for different $z$ as 
542: $L_z=1/3$ (i.e., spherical shape).}
543: \end{figure}
544: 
545: \begin{figure}[h]
546: \caption{Curve fitting for $L_z=1/3$ (i.e., spherical particles).}
547: \end{figure}
548: 
549: \newpage
550: \centerline{\epsfig{file=f1.eps,width=\linewidth}}
551: \centerline{Fig.1}
552: 
553: \newpage
554: \centerline{\epsfig{file=f11.eps,width=\linewidth}}
555: \centerline{Fig.2}
556: 
557: \newpage
558: \centerline{\epsfig{file=f2.eps,width=\linewidth}}
559: \centerline{Fig.3}
560: 
561: \newpage
562: \centerline{\epsfig{file=f3.eps,width=\linewidth}}
563: \centerline{Fig.4}
564: 
565: \end{document}
566: %%%%%%%%%%%% First-Principles Approach to Electrorotation Assay %%%%%%%%%%%
567: