1: \documentclass[debug,overfull]{epl}
2: \usepackage{amsmath,graphicx}
3:
4: \title{Self-consistent variational theory for globules}
5:
6: \author{Arti Dua and Thomas A. Vilgis}
7: \institute{ Max-Planck-Institute for Polymer Research,
8: Ackermannweg 10, 55128 Mainz, Germany
9: }
10: \pacs{36.20.-r}{Macromolecules and polymer molecules}
11: \pacs{61.25.Hq}{Macromolecular and polymer solutions; polymer melts; swelling}
12: \pacs{82.35.Rs}{Polyelectrolytes}
13:
14: \begin{document}
15:
16: \maketitle
17:
18: \begin{abstract}
19: A self-consistent variational theory for globules based on the uniform
20: expansion method is presented. This method, first introduced by Edwards
21: and Singh to estimate the size of a self-avoinding chain, is restricted to a
22: good solvent regime, where two-body repulsion leads to chain swelling.
23: We extend the variational method to a poor solvent regime where the balance
24: between the two-body attractive and the three-body repulsive interactions leads to
25: contraction of the chain to form a globule. By employing the Ginzburg criterion, we
26: recover the correct scaling for the $\theta$-temperature. The introduction
27: of the three-body interaction term in the variational scheme recovers the
28: correct scaling for the two important length scales in the globule---its
29: overall size $R$, and the thermal blob size $\xi_{T}$. Since these two
30: length scales follow very different statistics---Gaussian on length scales
31: $\xi_{T}$, and space filling on length scale $R$---our approach extends the
32: validity of the uniform expansion method to non-uniform contraction
33: rendering it applicable to polymeric systems with attractive interactions.
34: We present one such application by studying the Rayleigh instability of polyelectrolyte globules in poor solvents. At a critical fraction
35: of charged monomers, $f_c$, along the chain backbone, we observe a clear
36: indication of a first-order transition from a globular state at small $f$
37: to a stretched state at large $f$; in the intermediate regime the bistable
38: equilibrium between these two states shows the existence of a pearl-necklace
39: structure.
40: \end{abstract}
41:
42:
43: \section{Introduction}
44:
45: The uniform expansion method of Edwards and Singh is a well known
46: self-consistent variational approach to study the size and the probability
47: distribution of the end-to-end distance of a chain in a good solvent
48: \cite{edwards1,doi}.
49: %The method is based on introducing an unknown step length $b_1$ in such a way
50: %that the end-to-end distance of the self-avoiding chain of $N$ segments
51: %follows Gaussian statistics, $R = N^{1/2} b_{1}$.
52: The method is based on the uniform expansion of a chain in terms of the
53: expansion in an unknown step length $b_{1}$ such that the size of the
54: self-avoiding chain consisting of $N$ segments is governed by Gaussian
55: statistics, $ {R} = N^{1/2} b_1$. The variational parameter $b_1$ is then
56: determined self-consistently by a perturbative calculation, usually truncated at first order. In particular, the method recovers the correct scaling,
57: $\left< {\bf R}^2 \right> \approx b^{2} (v/b^{3})^{2/5} N^{6/5} $, for the size of an
58: excluded volume chain, a result which is consistent with the classical Flory
59: theory \cite{flory}. In contrast to other analytical approaches---for
60: instance, the self-consistent field theory by Edwards \cite{edwards2}, de
61: Gennes \cite{degennes}, and the more rigorous renormalization group treatment
62: by Freed \cite{freed,freed1}---the advantage of the uniform expansion method
63: is its mathematical simplicity in treating the complex excluded volume
64: problem, which renders it applicable to more complicated systems with
65: additional interactions. It has become a standard method in polymer physics to
66: understand the equilibrium behaviour of wide variety of polymeric systems.
67:
68: %polyelectrolytes \cite{yethiraj}, polyamphholytes \cite{joanny}, semiflexible polymers \cite{muthu2}, chains in disordered media \cite{thiru2}, critical fluids \cite{binny}, to name a few.
69:
70: The uniform expansion method though applicable to wide variety of problems is
71: only restricted to the behavior of polymers in a good solvent regime, where
72: the two-body repulsive interaction swells the chain. As the temperature is
73: reduced from the $\Theta$ temperature to a poor solvent regime, the two-body
74: interaction becomes attractive; therefore, any attempt to extend this approach
75: to a poor solvent regime requires the stabilizing influence of the three-body
76: repulsive interaction. In a simple scaling form, the free energy is given by
77: \begin{equation}
78: \label{f-energy}
79: \beta{\cal {F}} \simeq \frac{R^{2}}{Nb^{2}} - v \frac{N^2} {R^3} + w \frac{N^3}{R^6},
80: \end{equation}
81: where $v \sim \tau b^3$ and $w \sim b^6$ are the strength of the two and three
82: body interactions respectively. The reduced temperature $\tau$ describes the
83: distance from the $\Theta$ temperature, i.e., $\tau = |T- \Theta|/ \Theta$.
84: Close to the $\Theta$ temperature, $v \approx 0$, and the chain is essentially
85: Gaussian, $R \approx N^{1/2}b$; the $\Theta$ temperature can thus be determined
86: by the balance of the first two terms in the above equation, which yields $v
87: N^2/(N^{1/2}b)^3 \approx 1$ implying $\tau \approx N^{-1/2}$. The latter is
88: the Ginzburg criterion which determines the temperature at which the entropic
89: effects (fluctuations) become less important. As the solvent quality becomes
90: poor, the polymer conformation is determined by the balance of the two-body
91: attraction and the three-body repulsion; the comparison of the two opposing
92: interactions (which amounts to minimizing the above free energy with respect
93: to $R$ and then comparing the second and the third term) gives rise to a
94: globular structure of size $R \approx b \tau^{-1/3}N^{1/3}$, where $b$ is the
95: step length. Thus the size of the globule depends not only on the number of
96: segments, $N$, but also on the reduced temperature, $\tau$.
97:
98: Apart from the overall size of the globule there is another important length
99: scale in the problem---the correlation length of the density fluctuations
100: which defines the thermal blob size $\xi_{T}$ \cite{degennes1}. For length
101: scales smaller than $\xi_T$, the chain is unaffected by the attractive
102: interaction and essentially retains the Gaussian statistics, i.e., $\xi_{T}
103: \approx b {g_{T}}^{1/2}$, where $g_{T} \approx 1/\tau^2$ is the number of
104: monomers in the thermal blob of size $\xi_{T} \approx b/ \tau$. On length
105: scales larger than $\xi_{T}$ the chain begins to feel the attractive influence
106: of the two-body interaction and the thermal blobs are space filling, that is,
107: $\rho \xi_{T}^3 \approx 1$; the overall size of the globule is given by $R
108: \approx {\xi_{T}} {n_{T}}^{1/3}$, where $n_{T} = N/g_{T}$ is the number of
109: thermal blobs in a globule. In contrast to the ``uniform'' expansion in a good
110: solvent regime, the presence of the two different length scales in a globule
111: require a ``non-uniform'' contraction of the Gaussian chain. In what follows we
112: extend this variational method to a poor solvent regime by including the
113: repulsive three-body interaction along with the attractive two-body
114: interaction term.
115:
116: Let us first review some of the main results of the uniform expansion method,
117: which is essentially based on defining a new step length, $b_1 \gg b$, such
118: that the mean square end-to-end distance of the chain in presence of the
119: excluded volume is Gaussian and is governed by $\left< {\bf R}^2 \right> = N
120: b_1^2$. This condition amounts to replacing the original Hamiltonian ${\cal
121: H}_{0} = \frac{3}{2b^2} \int _{0}^{N} ds {\dot{\bf r}}(s)^2$ with the reference
122: Hamiltonian ${\cal H}_{1} = \frac{3}{2b_1^2} \int _{0}^{N} ds {\dot {\bf r}}(s)^2$.
123: %such that the distribution function ${\cal P} \propto \exp[-{\cal H}_{1}]$
124: %remains Gaussian, but contains un unknown step length.
125: This step can easily be carried out by adding and subtracting ${\cal H}_{1}$
126: from ${\cal H}_{0}$ to give the following expression for the mean square
127: end-to-end distance:
128: \begin{equation}
129: \label{Ham}
130: \left< {\bf R}^2 \right> =
131: \frac{\int {\cal D}[r(s)]{\bf R}^2 \exp[-({\cal H}_{1}+
132: ({\cal H}_{0}-{\cal H}_{1})+{\cal H}_{2})]}{\int {\cal D}[r(s)]
133: \exp[-({\cal H}_{1}+ ({\cal H}_{0}-{\cal H}_{1})+{\cal H}_{2})]},
134: \end{equation}
135: where ${\cal H}_{2} = \frac{|v|}{2}\int_{0}^{N} ds \int_{0}^{N} ds^{\prime}
136: \delta[{\bf r}(s)-{\bf r}(s^\prime)]$ represents the two-body interaction, and
137: $v = \tau b^3$ is the strength of the interaction. The idea is to expand
138: $\left< {\bf R}^2 \right>$ in a perturbative series about the reference
139: Hamiltonian ${\cal H}_{1}$ such that to the first order correction in $v$ one
140: obtains the following variational equation for the unknown parameter $b_1$:
141: \begin{equation}
142: \label{eqn1}
143: \left< {\bf R}^2 \right>_{1} \left< {\cal H}_{0}-{\cal H}_{1}+{\cal H}_2 \right>_{1} - \left< {\bf R}^2 ({\cal H}_{0}-{\cal H}_{1}+{\cal H}_2) \right>_{1} = 0,
144: \end{equation}
145: where $\left< \cdots \right>$ denotes the average with respect to the
146: probability distribution
147: $$
148: {\cal P}[{\bf r}(s)] \propto \exp \left[{-\frac{3}{2b_1^2} \int _{0}^{N} ds
149: {\dot{\bf r}}(s)^2}\right].
150: $$
151: The Gaussian nature of the probability
152: distribution makes the evaluation of the above averages simple. The
153: details of the calculation can be found in References (1) and (2); here we
154: simply present the result:
155: \begin{equation}
156: \label{eqn2}
157: \alpha^{5} - \alpha^3 - 2 v \left({\frac{6}{\pi^3}}\right)^{1/2}\frac{N^{1/2}}{b^3} = 0,
158: \end{equation}
159: where $\alpha = R/R_{0}$ is the expansion factor and $R_{0} = N^{1/2}b$. For a
160: swollen chain, $\alpha \gg 1$, the above equation can be solved to give $R
161: \approx |v|^{1/5}b^{2/5}N^{3/5}$, which recovers the correct scaling for the
162: excluded volume chain. It is interesting to note that inclusion of the higher
163: order terms only changes the numerical factors without altering the overall
164: structure of $R$ \cite{edwards1,doi}.
165:
166: In presence of the attractive two-body interaction term Eq. (\ref{eqn1}) has
167: to be modified to include the effects of the three-body repulsive interaction.
168: The resulting expression is given by
169: \begin{equation}
170: \label{eqn3}
171: \left< {\bf R}^2 \right>_{1} \left< {\cal H}_{0}-{\cal H}_{1}-{\cal H}_2 + {\cal H}_3\right>_{1} - \left< {\bf R}^2 ({\cal H}_{0}-{\cal H}_{1} - {\cal H}_2 + {\cal H}_3) \right>_{1} = 0.
172: \end{equation}
173: where $w = b^6$ is the strength of the three-body interaction and
174: $${\cal
175: H}_{3} = \frac{w}{6}\int_{0}^{N} ds \int_{0}^{N} ds^{\prime} \int_{0}^{N}
176: ds^{\prime\prime} \delta[{\bf r}(s)-{\bf r}(s^\prime)] \delta[{\bf
177: r}(s^\prime)-{\bf r}(s^{\prime\prime})].$$
178: The quantity we are interested in
179: evaluating is $\left< {\bf R}^2 \right>_{1}\left< {\cal H}_3\right>_{1} -
180: \left< {\bf R}^2 {\cal H}_3 \right>_{1}$, where, as before, the average with
181: respect to the Gaussian distribution function can easily be performed to give
182: \begin{eqnarray}
183: \label{3body1}
184: \left< {\bf R}^2 \right>_{1}\left< {\cal H}_3\right>_{1} - \left< {\bf R}^2 {\cal H}_3 \right>_{1} &=& \frac{w b_{1}^4}{9 (2\pi)^6}\int_0^{\infty} dq\int_{0}^{\infty} dk \int_{0}^{N} ds \int_{0}^{s} ds^{\prime} \int_{0}^{s^\prime} ds^{\prime\prime} ~~[(s-s^\prime)^2 k^4 q^2 \nonumber\\
185: & & ~~~~~~~~~~~~~+ (s^{\prime}-s^{\prime\prime})^2 k^2 q^4]e^{-k^2 b_{1}^2(s-s^\prime)/6} e^{-q^2 b_{1}^2(s^{\prime}-s^{\prime\prime})/6}.
186: \end{eqnarray}
187: It is clear from the structure of the above expression that the integration
188: over $k$ and $q$ followed by integrations over the internal coordinates of the
189: chain (e.g., $s^{\prime\prime}$, $s^\prime$ and $s$) leads to a divergence. This
190: divergence is a common feature of excluded-volume problems in polymer physics,
191: and has it origins in the use of delta function pseudopotentials. This
192: divergence, though present even in the case of the two-body interaction, cancels
193: in the overall expression for $\left<{\bf R}^2\right>$ (see References (1) and
194: (2) for more details). This is not the case when the three-body interaction is
195: included, and the evaluation of above integrations turns out to be
196: non-trivial. However, at the present level of the mean-field description,
197: it is possible to make reasonable approximations to probe the relevant length
198: scales in a globule.
199:
200: The idea is to first evaluate the $k$ integration in the first term and the
201: $q$ integration in the second term of Eq. (\ref{3body1}) followed by
202: integrations over the internal coordinates of the chain ($s^{\prime\prime}$,
203: $s^\prime$ and $s$). This order of integration is important to be able to
204: perform all but the last integration without encountering any divergence. The
205: result of such an operation can be expressed as
206: \begin{equation}
207: \label{3body2}
208: \left< {\bf R}^2 \right>_{1}\left< {\cal H}_3\right>_{1} - \left< {\bf R}^2 {\cal H}_3 \right>_{1} = \frac{w N}{ {b_{1}}^4}\left(\frac{81}{(2\pi)^6}\right) \left[ \left(\frac{2}{3}\right)^{3/2}\int_{0}^{\infty}dq ~~N^{1/2}b_{1} - \frac{\pi^{3/2}}{2} \right],
209: \end{equation}
210: which diverges as $q \rightarrow \infty$. However, within the mean-field
211: description, the divergence can be removed by introducing an upper cut-off for
212: the wave number $q$. This is not a serious approximation since density
213: fluctuations in a globule are extremely small, and the mean-field description
214: suffices. Thus by introducing an upper cut-off, $q_{max} = 2\pi/\Lambda$, one
215: can probe the relevant length scales, $\Lambda$, in a globule. The result of
216: the above integration when substituted into Eq. (\ref{eqn3}), produces the
217: following equation for $\alpha$:
218: \begin{equation}
219: \label{eqn4}
220: \alpha^{5} - \alpha^3 + \tau N^{1/2} - \frac{N^{1/2}b}{ \alpha^2 \Lambda} = 0,
221: \end{equation}
222: where we have used $v = \tau b^3$ and $w = b^6$. For convenience, we have
223: ignored all the numerical coefficients. First, we begin by asking for the
224: solution of the above equation for length scale of the size of the chain,
225: $\Lambda = R$. Since the only known length scale for the average size of the
226: chain is $R = N^{1/2} b_{1}$, it is reasonable to substitute $\Lambda \approx
227: N^{1/2}b_{1}$ into the above equation to yield
228: \begin{equation}
229: \label{eqn5}
230: \alpha^{5} - \alpha^3 + \tau N^{1/2} - \frac{1}{ \alpha^3}= 0.
231: \end{equation}
232: In presence of the attractive potential, the chain adopts a globular
233: conformation, and entropic effects become less important. We can, therefore,
234: use the Ginzburg criterion to estimate the $\Theta$ temperature by comparing
235: the first term (containing effects due to entropy) with the second term
236: (containing effects due to the attractive potential) in the above equation for
237: the condition $\alpha \ll 1$. Such a comparison yields $\tau \approx
238: N^{-1/2}$, a result which is consistent with the scaling analysis. The balance
239: between the second and the third term, on the other hand, yields the size of
240: the globule, $R = b\tau^{-1/3}N^{1/3}$.
241:
242: Second, for length scales smaller than the size of the globule but larger than the step length $b$, we look for the density correlations of the order of the size of the thermal blob, i.e., $\Lambda \approx \xi_{T}$. The substitution of the latter expression into Eq. (\ref{eqn4}) yields
243: \begin{equation}
244: \label{eqn5}
245: \alpha^{5} - \alpha^3 + \tau N^{1/2} - \frac{N^{1/2}b}{ \alpha^2 \xi_{T}}= 0.
246: \end{equation}
247: Since the statistics inside a thermal blob is Gaussian, we determine $\xi_{T}$
248: for the condition $\alpha = 1$. The latter condition produces $\xi_{T} =
249: b/\tau$, which, as discussed in the Introduction, represents the size of the
250: thermal blob. Moreover, the Gaussian statistics for $\xi_{T}$ implies $\xi_{T}
251: = {g_{T}}^{1/2} b$, where $g_{T}$ is the number of monomers inside the thermal
252: blob. The latter result along with $\xi_{T} = b/\tau$ leads to $g_{T} =
253: 1/\tau^2$. When the size of the thermal blob $\xi_{T}$ is of the order of the
254: step length $b$, we recover the condition, $\tau \approx 1$, defining the most
255: compact globule of size given by $R \approx b N^{1/3}$.
256:
257: In contrast to the scaling theories, the present approach provides an alternative way to estimate the blob size. In presence of the repulsive three-body interaction, the resulting variational equation needs to be regularized, a requirement that leads to a nonuniform contraction of the chain. The presence of the three body interaction is, therefore, essential to compute the overall size as well as the internal structure of the globule.
258:
259: Thus our extension of the method of Edwards and Singh to the nonuniform case recovers all the essential features of a globular structure, and yields a simple variational scheme which can be applied to systems with attractive interactions. In the next section we use this approach to study the Rayleigh instability of charged globules in poor solvents.
260:
261:
262:
263: \section{Polyelectrolytes in poor solvents}
264:
265:
266: In addition to the two-body attractive and three-body repulsive interactions,
267: the conformation of polyelectrolytes in poor solvents is also governed by the
268: long-range electrostatic interaction between charged monomers. With the
269: increase in the fraction of charges, $f$, along the chain backbone, a conformational transition occurs from a compact globular state at small $f$
270: to a highly extended rod-like state at large $f$. This transition is due the
271: repulsive nature of the Coloumb interaction, which at a critical fraction of
272: charges makes the large globular structure highly unstable; the system gains
273: energy by splitting into smaller charged globules connected by a narrow string
274: to form what is commonly known as a pearl-necklace---a string of locally
275: collapsed pearls (globules) \cite{avd,thomas}. This instability of a charged globule
276: is believed to be similar to the Rayleigh instability of a charged liquid
277: droplet which occurs as soon as its electrostatic energy, $Q^2k_{B}Tl_{B}/e^2
278: R$, exceeds the surface energy, $\gamma R^2$, i.e., $Q > (\gamma R^3/k_{B}T
279: l_{B})^{1/2} e$ \cite{kantor1,kantor2}, where $Q$ is the charge on the droplet
280: of size $R$, $e$ is the elementary charge, $l_{B} = e^2/4\pi\epsilon k_{B}T$
281: is the Bjerrum length, $\epsilon$ is the dielectric constant of the solvent
282: and $\gamma$ is the surface tension. As a result of this instability a charged
283: droplet spontaneously breaks up into smaller droplets, placed infinitely apart, each with a charge less than the critical one. However, due
284: to the constraint of chain connectivity the preferred equilibrium conformation
285: of a charged globule is that of a pearl-necklace, which with the increase in
286: $f$ results in a stretched state. Although scaling theories do provide some
287: evidence of the abrupt transition at a critical fraction of the charge from a
288: globule to a stretched state, a clear indication of its being first-order is
289: still missing. In what follows we study in detail the nature of the transition
290: that brings about the Rayleigh instability in a charged globule; the approach
291: we use is the variational method discussed in the last section.
292:
293: The repulsion between charged monomer is described by the Coulomb interaction
294: given by ${\cal H}_{c} = \frac{v_{c}}{2} \int_{0}^{N} ds \int_{0}^{N}
295: ds^{\prime} |{\bf r}(s) - {\bf r}(s^{\prime})|^{-1}$, where $v_{c} = f^2
296: l_{B}$ is the strength of the Coloumb interaction. This electrostatic
297: interaction when included into Eqs. (\ref{Ham}) and (\ref{eqn3}) yields the
298: following equation:
299: \begin{equation}
300: \label{eqn-c}
301: \left< {\bf R}^2 \right>_{1} \left< {\cal H}_{0}-{\cal H}_{1}-{\cal H}_2 + {\cal H}_3 + {\cal H}_c\right>_{1} - \left< {\bf R}^2 ({\cal H}_{0}-{\cal H}_{1} - {\cal H}_2 + {\cal H}_3 + {\cal H}_c) \right>_{1} = 0.
302: \end{equation}
303: The additional term that needs to be calculated is given by $\left< {\bf R}^2 \right>_{1}\left< {\cal H}_c\right>_{1} - \left< {\bf R}^2 {\cal H}_c \right>_{1}$. The details of the calculation will be presented elsewhere \cite{a-t}; here we simply present the result:
304: \begin{equation}
305: \label{eqn-c1}
306: \left< {\bf R}^2 \right>_{1}\left< {\cal H}_c\right>_{1} - \left< {\bf R}^2 {\cal H}_c \right>_{1} = - \frac{4}{45}\left(\frac{6}{\pi}\right)^{1/2} f^2 l_{B}N^{5/2}b_{1}
307: \end{equation}
308: \begin{figure}
309: \rotatebox{270}{\includegraphics[width=8cm,height=14cm]{newdata1.ps}}
310: \caption{ The numerical solution of Eq. (\ref{eqn-c2}) for $\alpha$ is plotted
311: as a function of the strength of the electrostatic energy ${\bar f} = f^2 l_{B}N/b$ for $\tau = 0.5$ and $N = 1000$. The region between the dotted lines is the
312: region of coexistence between a globule and an extended state. The bistable
313: equilibrium between these two states is represented by the dashed line,
314: where $R$ indicates the rod-like extended state and $G$ indicates the
315: globular state; the unstable part of the region is represented by the curve
316: $U$. ${\bar f}_c$ represents the critical fraction of charge at which the
317: first-order transition from a globular to an extended state occurs. The
318: parallel horizontal lines represent regions $A_1$ and $A_2$ of equal area.}
319: \end{figure}
320: The above expression when substituted into Eq. (\ref{eqn-c}) gives the
321: following equation for $\alpha$:
322: \begin{equation}
323: \label{eqn-c2}
324: \alpha^{5} - \alpha^3 + \tau N^{1/2} - \frac{1}{ \alpha^3} - \left(\frac{f^2 l_{B}}{b}\right)N^{3/2}\alpha^2= 0.
325: \end{equation}
326: The above equation for $\alpha$ can be solved in the following two limiting cases: when
327: $f^2 l_{B}N/b \ll \tau$, the attractive influence of the two-body interaction
328: dominates, and the comparison of the third and the fourth term yields a
329: globule of size $R = b \tau^{-1/3} N^{1/3}$; in the opposite limit of $f^2
330: l_{B}N/b \gg \tau$, the repulsive effects of the electrostatic interaction
331: overcomes the attractive interaction, and the comparison of the first and the
332: last term yields $\alpha = (f^2l_{B}/b)^{1/3} N^{1/2}$ implying a stretched
333: state, $R = bN(f^2l_{B}/b)^{1/3}$. In terms of the electrostatic blob picture
334: the latter represents the linear arrangement of $N/g_{el}$ (electrostatic)
335: blobs of size $\xi_{el} = (f^2 l_{B}/b)^{-1/3}$, where $g_{el} =
336: (l_{B}f^2/b)^{-2/3}$ is the number of monomers in a blob. Therefore, the
337: expressions obtained from the variational approach are in complete agreement
338: with the scaling predictions \cite{avd,thomas}.
339:
340: In the intermediate regime, ${\bar f} = f^2 l_{B}N/b \simeq \tau$, Eq. (\ref{eqn-c2}) can only be solved numerically. The result is presented in the form of Fig. (1),
341: which is a plot of $\alpha$ as a function of the strength of the electrostatic
342: energy. In the intermediate regime (the region between the dotted lines
343: represented by ${\bar f}_{min}$ and ${\bar f}_{max}$) the curve folds back on itself to give a bistable equilibrium between a globular state ($\alpha < 1$) and a
344: rod-like state ($\alpha > 1$). The coexistence of these two states in the
345: region between the two dotted lines in Fig.1 is a clear indication of the
346: existence of the pearl-necklace structure in this region. Moreover, the abrupt
347: transition from a globular to an extended state is evidently first-order in
348: nature. Therefore, we use the Maxwell equal-area construction to determine the
349: critical fraction of charge ${\bar f}_{c}$ at which the spontaneous Rayleigh
350: splitting of the charged globule occurs. The point ${\bar f}_{c}$ in Fig. 1 is the point which ensures that the area under the two curves (represented by
351: $A_1$ and $A_2$) is equal. The intermediate regime corresponds to the balance between the forces close to the Rayleigh splitting, i.e., when the electrostatic force $k_{B}Tl_{B}f^{2}N^{2}/R^{2}_{pn}$ becomes equal to the line tension $k_{B}T\tau/b$. The chain extension then scales as $R_{pn} = (l_{B}f^{2}/b\tau)^{1/2} N b$, where $R_{pn}$ represents the mean size of the pearl-necklace structure.
352:
353: The pearl-necklace structure, as seen in simulations, is a highly fluctuating
354: structure both in terms of the position and the size of pearls \cite{limbach}.
355: In the mean-field description, the existence of such a structure is evident
356: through a bistable equilibrium between a globular and a stretched state, as is
357: clear from the coexistence region, ${\bar f}_{min} \leq {\bar f} \leq {\bar f}_{max}$, in Fig. 1. The scaling theories usually estimates the average size of a pearl-necklace by comparing its surface energy with that of the electrostatic energy. In a
358: future publication, we will present a detailed analysis of the dependence of
359: $R$, in the coexistence region, on $f$, $N$ and $\tau$ to make comparison with
360: the scaling predictions. The effects of the variation of $\tau$ and $N$ on the
361: coexistence region (pertaining to pearl-necklaces) and $f_c$ will also be
362: discussed \cite{a-t}.
363:
364: \section{Conclusions}
365:
366: We have extended the uniform expansion method of Edwards and Singh to a poor solvent regime by accounting for the three-body repulsive interaction in addition to the two-body attractive potential. The self-consistent variational approach recovers the two important length scales in a globule---its overall size and the thermal blob size. This method can be applied to several different polymeric systems where the effective two-body potential is negative. To illustrate the generality of our approach, we applied this method to polyelectrolytes in poor solvents to study the Rayleigh instability of charged globules in detail. The presence of the additional long-range repulsion due to the electrostatic interaction between the charged monomers destabilizes the globular structure as the fraction of charged monomers along the chain backbone is increased. At a certain critical fraction of charges the globule splits into locally collapsed monomers joined by a linear string to form a pearl-necklace. Using the present approach, we observe a clear indication of a first-order transition, at a critical fraction of charged monomers, from a globular to an extended state. The intermediate regime of bistable equilibrium between these two states is the region where the pearl-necklace structure is stable.
367:
368: In contrast to flexible polymers, the behaviour of semiflexible polymers in poor solvents is not well understood. Given the simplicity of the present approach, it holds out the possibility of addressing many such issues, including the equilibrium behaviour of biopolymers with specific interactions \cite{a-t}.
369:
370:
371:
372: \begin{thebibliography}{0}
373: \bibitem{edwards1} S. F. Edwards and P. Singh, J. Chem. Soc. Faraday Trans II {\bf 75}, 1001 (1979).
374: \bibitem{doi} M. Doi and S. F. Edwards, The Theory of Polymer Dynamics, Clarendon Press, Oxford (1986).
375: \bibitem{flory} P. J. Flory, J. Chem. Phys. {\bf 17}, 303 (1949).
376: \bibitem{edwards2} S. F. Edwards, Proc. Phys. Soc. {\bf 85}, 613 (1965).
377: \bibitem{degennes} P. G. de Gennes, Phys. Letters A {\bf 38}, 339 (1972).
378: \bibitem{freed} M. K. Kosmas and K. F. Freed, J. Chem. Phys. {\bf 68}, 4878 (1978).
379: \bibitem{freed1} K. F. Freed, Renormalization Group Theory of Macromolecules, Wiley, New York (1987).
380: \bibitem{degennes1} P. G. de Gennes, Scaling Concepts in Polymer Physics, University Press: Ithaca, NY (1985).
381: \bibitem{avd} A. V. Dobrynin, M. Rubinstein and S. P. Obukhov, Macromolecules {\bf 29}, 2974 (1996).
382: \bibitem{thomas} T. A. Vilgis, A. Johner and J. F. Joanny , Eur. Phys. J. E {\bf 2}, 289 (2000).
383: \bibitem{kantor1} Y. Kantor and M. Kardar, Europhys. Lett. {\bf 27}, 643 (1994).
384: \bibitem{kantor2} Y. Kantor and M. Kardar, Phys. Rev. E {\bf 51}, 1299 (1995).
385: \bibitem{limbach} H. J. Limbach, C. Holm and K. Kremer, Europhys. Lett. {\bf 60}, 566 (2002).
386: \bibitem{a-t} A. Dua and T. A. Vilgis (unpublished).
387:
388: \end{thebibliography}
389: \end{document}