1: \documentclass[figures]{epl}
2: \usepackage{epsfig}
3: \title{Scaling and hydrodynamic effects in lamellar ordering}
4: \shorttitle{Lamellar ordering}
5: \author{Aiguo Xu\inst{1}\thanks{Present address: Dept. of Physics,
6: Yoshida-south Campus, Kyoto University, Sakyo-ku, Kyoto, 606-8501 Japan}
7: \and G. Gonnella\inst{1} \and A. Lamura\inst{2}
8: \and G. Amati\inst{3} \and F. Massaioli\inst{3}}
9: \shortauthor{Aiguo Xu \etal}
10: \institute{
11: \inst{1} Dipartimento di Fisica, Universit\`a di Bari {\rm and}
12: Center of Innovative Technologies for Signal Detection and Processing
13: (TIRES) {\rm and} INFN (Istituto Nazionale di Fisica Nucleare), Sez. di Bari
14: {\rm and} Istituto Nazionale per la Fisica della Materia,\\
15: via Amendola 173, 70126 Bari, Italy\\
16: \inst{2} Istituto Applicazioni Calcolo, CNR, Sezione di Bari,\\
17: via Amendola 122/D, 70126 Bari, Italy\\
18: \inst{3} CASPUR, Via dei Tizii 6/b, 00185 Rome, Italy
19: }
20: \pacs{64.75.+g}{Solubility, segregation, and mixing; phase separation}
21: \pacs{05.70.Ln}{Nonequilibrium and irreversible thermodynamics}
22: \pacs{47.20.Hw}{Morphological instability; phase changes}
23: \pacs{82.35.Jk}{Copolymers, phase transitions, structure}
24:
25: \begin{document}
26:
27: \maketitle
28:
29: \begin{abstract}
30: We study the kinetics of domain growth of fluid mixtures quenched
31: from a disordered to a lamellar phase. At low viscosities, in two
32: dimensions, when
33: hydrodynamic modes become important, dynamical scaling is verified in the form
34: $C(\vect{k}, t) \sim L^{\alpha} f[(k-k_M)L]$ where $C$ is the
35: structure factor with maximum at $k_M$ and
36: $L$ is a typical length of the system with
37: logarithmic growth at late times.
38: The presence of extended defects can explain the behavior of $L$.
39: Three-dimensional simulations confirm that diffuse grain boundaries
40: inhibit complete ordering of lamellae.
41: Applied shear flow alleviates frustration and
42: gives power-law growth at all times.
43: \end{abstract}
44:
45: \section{Introduction}
46: The kinetics of the growth of ordered phases after a quench from a
47: disordered state continues to provide interesting physical
48: questions. While the process is reasonably well understood for the
49: case of binary mixtures, where dynamical scaling occurs and
50: domains grow with power-law behavior \cite{B94}, basic questions
51: remain to be clarified for more complex systems \cite{Y01}. In
52: this Letter we consider the case of fluid mixtures where, due to
53: competing interactions, the system would arrange itself in
54: stripes. Examples are di-block copolymer melts, where chains of
55: type A and B, covalently bonded end-to-end in pairs, segregate at
56: low temperatures in regions separated by a stack of lamellae
57: \cite{Bat}, or ternary mixtures where surfactant form interfaces
58: between oil and water \cite{GS94}. Other systems with lamellar
59: order are smectic liquid crystals \cite{Deg}, dipolar \cite{SD}
60: and supercooled liquids \cite{KKZNT95}, chemically reactive binary
61: mixtures \cite{GC}.
62: Lamellar patterns are also observed in Raleigh-B\'enard cells
63: above the convective
64: threshold \cite{CH}.
65:
66: The ordering of lamellar systems is characterized by the presence
67: of frustration on large scales. This affects the late time
68: evolution which, as discussed below, can be very slow or also
69: frozen. Relations with the dynamics of structural glasses have
70: been also considered in recent literature \cite{BV02,SW00}. One
71: can expect that in real systems the effects of the velocity field,
72: inducing motion around local or extended defects, can be relevant.
73: Our purpose here is to analyze the process of lamellar ordering
74: taking into account the effects of hydrodynamics.
75: We consider a model
76: based on a Ginzburg-Landau free-energy, where dynamics is
77: described by convection-diffusion and Navier-Stokes equations \cite{yeo}.
78: Our results for two-dimensional systems show that, when
79: hydrodynamical modes are effective (at sufficiently low
80: viscosities), dynamical scaling holds with a complex dependence
81: of characteristic lengths on time, analogous to that of systems
82: with quenched disorder \cite{RC} not present here. An intermediate
83: faster growth is followed by a slower logarithmic evolution at
84: later times.
85: The slowing-down of the dynamics is attributed to the formation
86: of extended
87: defects between domains of lamellae
88: with different orientation, almost perpendicular each other.
89: Our simulations show that the same phenomenon occurs in
90: three-dimensional systems.
91: These results are not expected to depend on the specific model
92: considered and could be relevant for a broad class of systems with
93: lamellar order. We also verified that applied shear flows
94: alleviate frustration giving power-law growth at all times.
95:
96: \section{The model}
97: We consider the free-energy model \cite{B75,Lei80}
98: \begin{equation}
99: {\cal F}\{\varphi\} = \int d^d x
100: \{\frac{a}{2} \varphi^2 + \frac{b}{4} \varphi^4
101: + \frac{\kappa}{2} \mid \nabla \varphi \mid^2 +\frac {c}{2} (\nabla^2
102: \varphi)^2\}
103: \label{eqn1}
104: \end{equation}
105: where $\varphi$ is the order parameter field representing the
106: local concentration difference between the two components of the
107: mixture. The parameters $b$ and $ c $ have to be positive in order
108: to ensure stability. For $a<0$ and $\kappa
109: > 0 $ the two homogeneous phases with $\varphi = \pm \sqrt{-a/b}$ coexist.
110: A negative $\kappa$ favors the presence of interfaces and a
111: transition into a lamellar phase can occur. In single mode
112: approximation,
113: assuming a profile like $A \sin k_0 x$ for the direction transversal to the
114: lamellae,
115: one finds the transition ($|a|=b$) at $ a \approx - 1.11 \kappa^2 /c $
116: where $ k_0 = \sqrt{ -\kappa /2c }$, $A^2 = 4 (1 +
117: \kappa^2/4cb)/3$.
118: The expression (1) can be
119: also written for negative $\kappa$ as ${\cal F}\{\varphi\} = \int
120: d^d x \{\frac{\tau}{2} \varphi^2 + \frac{b}{4} \varphi^4 +
121: \frac{c}{2}[(\nabla ^2 + k_0^2) \varphi)^2]\}$, $\tau = a - c
122: k_0^4 $, and in this form is generally used to describe di-block
123: copolymers in the weak segregation limit\cite{Lei80}.
124:
125: The dynamical equations are \cite{Y01}
126: \begin{equation}
127: \frac {\partial v_{\alpha}}{\partial t} + \vect{v} \cdot \vect{\nabla}
128: v_{\alpha} = - \frac{1}{\varrho} \frac {\partial P_{\alpha
129: \beta}} {\partial x_{\beta}} + \nu \nabla^2 v_{\alpha} \quad,
130: \label{motion1}
131: \end{equation}
132: \begin{equation}
133: \frac {\partial \varphi}{\partial t}+ \vect{\nabla} \cdot (\varphi
134: \vect{v}) = \Gamma \nabla^2 \frac {\delta {\cal F}} {\delta \varphi}
135: \quad, \label{motion2}
136: \end{equation}
137: where $v_{\alpha} $
138: are the components
139: of the velocity field, $\varrho$ is the total density of the
140: mixture and the incompressibility condition $\vect{\nabla} \cdot
141: \vect{v} = 0 $ is considered. $\nu $ is the kinematic viscosity and
142: $\Gamma$ is a mobility coefficient. The pressure tensor is the sum
143: of the usual hydrodynamical part and of a tensor $P_{\alpha
144: \beta}^{chem} $ depending on $ \varphi$, with a functional form
145: obtainable from the free energy containing off-diagonal terms,
146: whose expression can be found in \cite{yeo}. It can be shown that
147: $\delta \mathcal {F} /\delta \varphi = 0 $ implies
148: $\partial_{\alpha} P_{\alpha \beta}^{chem} = 0$. The laplacian in
149: the r.h.s. of Eq.~(\ref{motion2}) ensures the conservation of the
150: order parameter. We will simulate these equations by using a finite
151: difference scheme for the convection-diffusion equation and a
152: Lattice Boltzmann Method (LBM) for the Navier-Stokes equation.
153: Advantages of this method \cite{det} with respect to other LBM
154: schemes for fluid mixtures are that spurious terms
155: appearing in Eq.(3) \cite{Swift96} are avoided
156: and the numerical efficiency is increased. In $D=3$ we used a
157: parallel version of the LBM code which {\it fuses} the
158: streaming and the collision processes to reduce bandwidth
159: requirements \cite{giorgio}.
160:
161: \begin{figure}[t]
162: \begin{center}
163: %\epsfig{file=fig1.eps,bbllx=104 pt,bblly= 242 pt,bburx= 523
164: %pt,bbury= 710 pt,width=0.4\textwidth,clip=}
165: % \includegraphics*[width=0.51\textwidth]{fig2.eps}\\*
166: Fig.~1\hspace{0.5\textwidth}Fig.~2
167: \end{center}
168: \vskip -0.5cm
169: \caption{Configurations at
170: different times of a $256 \times 256$ portion of the lattice.}
171: \caption{Time behavior of characteristic size $L$
172: and $C_M \equiv C(k_M,t)$ on log-log (inset) and log-linear
173: scales. The straight line in the inset has slope $0.3$.}
174: \end{figure}
175:
176: We can summarize now what is known on the ordering properties of
177: lamellar fluids. Previous two-dimensional studies of the full
178: model (1-3) \cite{yeo}, without quantitative analysis due to the
179: small size of the lattice considered ($128^2$), showed the
180: relevance of hydrodynamics for obtaining well ordered domains on
181: the scale of the system simulated. More results exist for variants
182: of the model (1) with long-range interactions, without
183: hydrodynamics \cite{LG89,STY96,MZF98} or neglecting the inertial
184: terms of the l.h.s. of Eq.(2) \cite{YS02}. Eq.(3) without
185: advection and with non-conserved order parameter,
186: corresponding to the Swift-Hohenberg model for
187: Raleigh-B\'enard convection \cite{SH,EVG92,HSG97,CM95,BV01}, has
188: been also largely studied. Regimes with dynamical scaling have
189: been found with the order parameter structure factor behaving as
190: $ C(\vect{k}, t) \equiv <|\varphi_{\vect{k}}(t)|^2>
191: \sim t^z f[(k - k_0) t^z]$ with different values for the exponent $z$,
192: $ \varphi_{\vect{k}}(t) $ being the Fourier transform of
193: $\varphi(\vect{x}, t)$
194: \cite{EVG92,CM95,HSG97,CB98,QM02,YS02,Har00,BV01}.
195: Slower evolution with frozen states and grain boundary pinning has
196: been also observed \cite{BV02,HSG97} for very deep quenches. In $D=3$
197: the few existing simulations have not considered dynamical scaling
198: \cite{coveney,ohta}.
199:
200: \section{Ordering dynamics}
201: We run our two-dimensional simulations on
202: a $1024^2$ lattice starting from disordered configurations. We
203: fixed $|a| = b $ so that the minimum of the potential part of the
204: free-energy is at $\varphi = \pm 1$. We checked on small systems
205: ($128^2$) for different $\kappa $ in the lamellar phase that the
206: expected equilibrium state was reached. The
207: value of $c$ was fixed in such a way that the period is about 10
208: times the lattice spacing and $|a| = 2 \times 10^{-4}, \kappa = -3
209: \times |a|, c = 3.8 \times |a| , \Gamma = 25$
210: for the cases presented. Details on the LBM part
211: of the code can be found in \cite{Xu03}.
212:
213: After the quench, lamellae start to form evolving until the
214: equilibrium wavelength is locally reached. This part of the
215: ordering process does not depend on the value of viscosity.
216: Later, however, the system continues to order only at sufficiently
217: low viscosities ($\nu \stackrel{<}{\sim} 0.1$). Otherwise, local defects
218: dominate
219: \cite{yeo} and, at times of order $ t \sim
220: 3000$, the system results frozen in tangled configurations. An
221: example of evolution at low viscosity ($ \nu = 8.33 \times 10^{-3}
222: $) is shown in Fig.1. Different kinds of defects can be observed.
223: The annihilation of a couple of dislocations is put in evidence.
224: We have measured the relative velocity of the dislocations and we
225: found it constant. At the last time of the figure the system can
226: be observed to be ordered on larger scales.
227:
228: A quantitative description of the ordering process comes from the
229: analysis of the structure factor $C(\vect{k}, t)$. After spherical
230: average we plotted $C(k,t)$ at different times. In the early
231: regime, with $t \stackrel{<}{\sim} 3000$, $C(k,t)$ develops a maximum at a
232: momentum $k_M$ which decreases with time until the equilibrium
233: value $k_0$ is reached. Then the peak of $C(k,t)$, remaining at
234: $k_0$, continues to grow while the width decreases indicating an
235: increase of order in the system. A characteristic length $L$ for
236: this process can be extracted from the structure factor in the
237: usual way by measuring the full width $ \delta k$ of
238: $C$ at half maximum and defining $L(t) = 2 \pi / \delta k $. In
239: the inset of Fig.2 we plotted $L(t)$ on log-log scale, calculated
240: by averaging over 5 different runs. For more than one decade a
241: behavior consistent with the power-law $L \sim t^{z} $ can be
242: observed with $z = 0.30 \pm 0.02$. The growth becomes slower at
243: later times and we find $ L \sim \ln t$ as the straight line in
244: the main frame of Fig.2 suggests. We checked that the slowing down
245: cannot be attributed to finite size effects (at the latest
246: times considered, $L(t)$ is less than 1/10 of the size of the
247: system).
248:
249: \begin{figure}[t]
250: \begin{center}
251: %\epsfig{file=oldfig3.eps,bbllx=84 pt,bblly= 2 pt,bburx= 670
252: %pt,bbury= 590 pt,width=0.4\textwidth,clip=}
253: % \includegraphics*[width=0.33\textwidth]{fig3.eps}
254: % \includegraphics*[width=0.66\textwidth]{fig4.eps}\\*
255: Fig.~3\hspace{0.5\textwidth}Fig.~4
256: \end{center}
257: \vskip -0.5cm
258: \caption{Configuration of a $128^3$ system at $t=1000000$.
259: The following parameters were used: $a = - b = - 0.026$,
260: $\kappa = - 0.005$,
261: $c = 0.0025$, $\Gamma = 0.25$,
262: and $\nu = 0.1$.}
263: \caption{The middle horizontal sections of a $512^3$
264: systems at times $t=600000$ and $t=700000$. Parameters of this
265: simulations are the same of Fig.~3.
266: Observe how the vertical stripes in the central
267: bottom part of the figures are distorted by the pressure of
268: horizontal domains.}
269: \end{figure}
270:
271: The behavior of $L$ can be related to the role played in late
272: time dynamics by extended defects (grain boundaries) between
273: domains of differently oriented lamellae. A decrease in the
274: growing rate of $L$ was observed at times $t\approx 25000$ when
275: the average value of the angle between the normals of neighboring
276: domains becomes close to $90^{\circ}$. Isolated grain boundaries
277: can be shown to be stable defects in lamellar systems described by
278: Eq.(\ref{eqn1}) \cite{BV02}.
279: A quantitative description of
280: ordering in lamellar systems can be obtained by considering the
281: evolution equation for a single defect. The position
282: $\textit{x}_{gb}$ of the grain boundary evolves according to
283: $\dot{\textit{x}}_{gb} = A \mathcal{C}^2(t) - B \cos(2 k_0
284: \textit{x}_{gb} + \phi)$ where $ \mathcal{C}$ is the curvature of
285: lamellae parallel to the boundary and $ A,B,\phi$ are
286: model-dependent parameters \cite{BV02}. If one assumes that
287: dynamical scaling is verified, a typical length can be defined as
288: the inverse of curvature, and the previous equation implies a
289: growth like $t^{1/3}$ until a critical value for curvature is
290: reached. This analysis, elaborated for one single defect, suggests
291: that in the ordering of a large system, in absence of other
292: processes and without hydrodynamics, defects will become pinned at
293: a certain time with the system frozen in a configuration with many
294: grain boundaries. In our case the velocity field helps the system
295: in continuing the ordering process also at very late times but
296: slower. We measured $L \sim \ln t$ which is the growth expected
297: for systems with metastable states. An extension of the analytical
298: arguments of \cite{BV02,HV04} in presence of hydrodynamics is not
299: available. However, it is interesting to observe that our
300: simulations suggest an intermediate regime with a power law
301: consistent with that of \cite{BV02} followed by a logarithmic
302: asymptotic growth, both results explainable as due to grain
303: boundary dynamics.
304:
305: We also considered three dimensional systems.
306: Simulations ranging from $64^3$ to $512^3$ lattices were performed: The
307: largest grid asked for about 20 days of sustained computations, using
308: a 8-CPU vector machine NEC SX6 and more than 32 GB of memory.
309: While complete order was reached on $64^3$ lattices,
310: larger systems confirm the role of grain boundaries
311: in lamellar ordering.
312: An example is given in Fig.3 where a single grain boundary is shown: After
313: $t \sim 10^{7}$ no clear 2D ordering was observed.
314: On the $512^3$ lattice we studied the behavior of $L(t)$,
315: defined as in $D=2$. After the formation of lamellae,
316: the intermediate
317: and late time regimes were found shorter and it was not possible
318: to deduce quantitative behaviors. However, the role of grain
319: boundaries can be appreciated looking for example at the
320: configurations of Fig.4. Here the central horizontal sections of
321: the system at $t=600000$ and $t=700000$ are shown. Many grain
322: boundaries with three main orientations can be seen. After $t
323: \sim 600000$, in correspondence of a sudden decrease observed in
324: the growing rate of $L(t)$, the further evolution of the system
325: proceeds through the distortion of the grain boundaries existing
326: at that time. This can be seen for example in the central bottom
327: part of the configuration at $t=700000$. Different domains, trying to
328: increase their size, push themselves reciprocally, helped by the
329: velocity field. This gives an unexpected late-time increase of
330: curvature, also observed in $D=2$. The role of hydrodynamics is
331: important in this process: The system can reach a higher degree of order
332: only by developing a strong velocity field which bends the existing flat
333: grain boundaries. This will eventually bring later to a configuration with
334: less defects. This picture is confirmed by the
335: increase of kinetic energy observed at $t\sim 600000$.
336:
337: \section{Dynamical scaling}
338: In order to analyze dynamical
339: scaling, together with $L(t)$, we also considered the behavior of
340: the peak $C_M \equiv C(k_M,t)$ of the structure factor. In
341: two-dimensional simulations, from Fig.2, one sees that $C_M$ grows
342: similarly to $L(t)$ but with a different rate. This suggests to
343: consider scaling behavior in the general form $C(k,t) \sim
344: L^{\alpha} f[(k-k_M)L]$ \cite{CRZ94}. This differs from the usual
345: expression considered in lamellar ordering ($C \sim t^{z}
346: f[(k-k_M)t^{z}]$) for the introduction of $\alpha$ and
347: the use of $L$ even if not given by a power law.
348: As illustrated in the inset of Fig.5, we evaluated $\alpha= 1.25$.
349: Then, by using this value, we found, as shown in Fig.5, that
350: rescaled structure factors overlap {\it at all times} after $t \sim 3000$
351: on a single
352: curve, confirming
353: our scaling assumption and suggesting an analogy with kinetic
354: behavior in systems with quenched disorder. In Random Field Ising
355: Models (RFIM), for example, diffusive growth changes over to
356: logarithmic behavior but dynamical scaling holds at all times
357: with the same scaling function as in standard Ising model
358: \cite{RC}.
359: Here a similar
360: scenario occurs but without quenched disorder.
361: This is interesting also in relation with the existence of an
362: equilibrium glass transition in systems with lamellar order
363: \cite{SW00}.
364:
365: \begin{figure}[t]
366: \begin{center}
367: %\epsfig{file=fig5.eps,bbllx=41 pt,bblly= 168 pt,bburx= 511
368: %pt,bbury= 650 pt,width=0.42\textwidth,clip=}
369: % \includegraphics*[width=0.53\textwidth]{fig6.eps}\\*
370: Fig.~5\hspace{0.5\textwidth}Fig.~6
371: \end{center}
372: \vskip -0.5cm
373: \caption{Rescaled structure factors at different times. Symbols
374: $\triangle, \circ, \star, \bullet, \ast, \diamond$ refer to times
375: $4600$, $14100$, $40800$, $61100$, $78100$, $123100$,
376: respectively. In the inset $C_M$ is plotted on log-log scale as a
377: function of $L$; the slope of the line is 1.25.}
378: \caption{Evolution of $L_x$ and $L_y$
379: for $\dot{\gamma} = 10^{-4} $. The straight line has slope 0.6. At
380: the latest times of the figure the system is almost completely
381: ordered, as shown in the $200 \times 200$ portion of the inset.}
382: \end{figure}
383:
384: Concerning the exponent $\alpha$, its value is related to the
385: compactness of domains and is $\alpha = D$ in asymptotic growth
386: of binary mixtures, $\alpha = 0 $ in early time regime when
387: interfaces are forming \cite{CRZ94}. Our results with $\alpha
388: > 1 $ show that the argument for $\alpha
389: =1$ based on the existence of one scaling dimension for $2D$
390: lamellar systems \cite{EVG92} does not hold in presence of
391: hydrodynamics.
392: We found dynamical scaling also in 3D simulations but for a
393: shorter time interval.
394:
395: \section{Shear effects}
396: We also studied lamellar ordering in presence of
397: shear flow.
398: Shear was applied with bounce-back conditions on rigid walls
399: moving with velocity $\pm v_{wall}$ at the top and bottom of the
400: simulation volume \cite{Xu03}. Ordering is favored since
401: interfaces want to orientate with the flow \cite{On97}. Breaking
402: and recombination of domains, induced by shear, makes the
403: elimination of topological defects faster \cite{prl2}. As a
404: result, logarithmic growth disappears and ordering occurs also
405: asymptotically with power-law behavior even in cases, at high
406: viscosity, where without shear freezing would have been observed.
407:
408: An example of this behavior ($\nu = 7.5$) is shown in Fig.6.
409: Growth was measured by the quantities $L_{\alpha} = \int d\vect{k}
410: C(\vect{k},t) / \int d\vect{k} |k_{\alpha}| C(\vect{k},t)$ ($\alpha
411: =x,y$).
412: $L_x$ grows with exponent
413: $z_x = 0.6$ while $L_y$ relaxes to the equilibrium value.
414: Similar results were found for other values of $\nu $;
415: a detailed exposition of our results with shear will appear
416: elsewhere.
417:
418: \section{Conclusions}
419: To conclude, we have studied lamellar ordering in fluid mixtures
420: with competitive interactions. At low viscosities, when
421: hydrodynamic modes are effective, dynamical scaling holds in the
422: form $C(\vect{k} , t) \sim L^{\alpha} f[(k-k_M)L]$ where $L$ is a
423: characteristic length with complex behavior and logarithmic growth
424: at late times. This scaling, similar to that found in systems
425: with quenched disorder, is shown here for the first time in 2D
426: systems with lamellar order. 3D simulations also show the relevance
427: of grain boundaries defects for this dynamics.
428: Shear flow removes extended defects giving power-law behavior at
429: all times.
430: \acknowledgments
431: We warmly thank F. Corberi, E. Orlandini, M. Zannetti and J.
432: Yeomans for helpful discussions. We acknowledge
433: support by MIUR (PRIN-2004).
434:
435: \begin{thebibliography}{99}
436:
437: \bibitem{B94}
438: See, e.g.,
439: \Name{Bray A. J.}
440: \REVIEW{Adv. Phys.}{43}{1994}{357}.
441:
442: \bibitem{Y01}
443: See, e.g.,
444: \Name{Yeomans J. M.}
445: \REVIEW{Ann. Rev. Comput. Phys.}{VII}{2000}{61}.
446:
447: \bibitem{Bat}
448: \Name{Bates F. S. \And Fredrickson G.}
449: \REVIEW{Ann. Rev. Phys. Chem.}{41}{1990}
450: {525}.
451:
452: \bibitem{GS94}
453: \Name{Gompper G. \And Schick M.}
454: \Book{Phase transitions and critical phenomena, Vol. 16}
455: \Publ{Academic, New York}\Year{1994}.
456:
457: \bibitem{Deg}
458: See, for example,
459: \Name{de Gennes P. G.}
460: \Book{The Physics of Liquid Crystals}
461: \Publ{Clarendon, Oxford}\Year{1974}.
462:
463: \bibitem{SD}
464: \Name{Roland C. \And Desai R. C.}
465: \REVIEW{Phys. Rev. B}{42}{1990}{6658}.
466:
467: \bibitem{KKZNT95}
468: \Name{Kivelson D., Kivelson S. A., Zhao X., Nussinov Z. \And Tarjus G.}
469: \REVIEW{Physica A}{219}{1995}{27};
470: \Name{Grousson M., Krakoviack V., Tarjus G. \And Viot P.}
471: \REVIEW{Phys. Rev. E}{66}{2002}{026126}.
472:
473: \bibitem{GC}
474: \Name{Glotzer S. C., Di Marzio E. A. \And Muthukumar M.}
475: \REVIEW{Phys. Rev. Lett.}{74}{1995}{2034}.
476:
477: \bibitem{CH}
478: \Name{Cross M. C. \And Hohenberg P. C.}
479: \REVIEW{Rev. Mod. Phys.}{65}{1993}{851}.
480:
481: \bibitem{BV02}
482: \Name{Boyer D. \And Vi\~{n}als J.}
483: \REVIEW{Phys. Rev. E}{65}{2002}{046119}.
484:
485: \bibitem{SW00}
486: \Name{Schmalian J. \And Wolynes P. G.}
487: \REVIEW{Phys. Rev. Lett.}{85}{2000}{836};
488: \Name{Grousson M., Tarjus G. \And Viot P}
489: \REVIEW{Phys. Rev. Lett.}{86}{2001}{3455}.
490:
491: \bibitem{yeo}
492: \Name{Gonnella G., Orlandini E. \And Yeomans J. M.}
493: \REVIEW{Phys. Rev. Lett.}{78}{1997}{1695};
494: \REVIEW{Phys. Rev. E}{58}{1998}{480}.
495:
496: \bibitem{RC}
497: \Name{Rao M. \And Chakrabarti A.}
498: \REVIEW{Phys. Rev.Lett.}{71}{1993}{3501}.
499:
500: \bibitem{B75}
501: \Name{Brazovskii S. A.}
502: \REVIEW{Sov. Phys. JETP}{41}{1975}{85}.
503:
504: \bibitem{Lei80}
505: \Name{Leibler L.}
506: \REVIEW{Macromolecules}{13}{1980}{1602};
507: \Name{Ohta T. \And Kawasaki K.}
508: \REVIEW{Macromolecules}{19}{1986}{2621};
509: \Name{Fredrickson G. H. \And Helfand E.}
510: \REVIEW{J. Chem. Phys.}{87}{1987}{697}.
511:
512: \bibitem{det}
513: Details about the method will appear elsewhere.
514:
515: \bibitem{Swift96}
516: \Name{Swift M. R., Orlandini E., Osborn W. R. \And Yeomans J. M.}
517: \REVIEW{Phys. Rev. E}{54}{1996}{5041}.
518:
519: \bibitem{giorgio}
520: \Name{Amati G. \And Massaioli F.}
521: {\it Proceeding of IBM Scicomp9, Bologna},\\
522: http://www.spscicomp.org/ScicomP9/Presentations/MassaioliAmatiScicomp09.pdf,
523: (2004).
524:
525: \bibitem{LG89}
526: \Name{Liu F. \And Goldenfeld n.}
527: \REVIEW{Phys. Rev. A}{39}{1989}{4805}.
528:
529: \bibitem{STY96}
530: \Name{Shiwa Y., Taneike T. \And Yokojima Y.}
531: \REVIEW{Phys. Rev. Lett.}{77}{1996}{4378}.
532:
533: \bibitem{MZF98}
534: \Name{Maurits N. M., Zvelindovsky A. V., Sevink G. I. A., van
535: Vlimmeren B. A. C. \And Fraaije J. G. E. M.}
536: \REVIEW{J. Chem. Phys.}{108}{1998}{9150}.
537:
538: \bibitem{YS02}
539: \Name{Yokojima Y. \And Shiwa Y.}
540: \REVIEW{Phys. Rev. E}{65}{2002}{056308}.
541:
542: \bibitem{SH}
543: \Name{Swift J. \And Hohenberg P. C.}
544: \REVIEW{Phys. Rev. A}{15}{1977}{319}.
545:
546: \bibitem{EVG92}
547: \Name{Elder K. R., Vi\~nals J. \And Grant M.}
548: \REVIEW{Phys. Rev. Lett.}{68}{1992}{3024}.
549:
550: \bibitem{CM95}
551: \Name{Cross M. C. \And Meiron D. I.}
552: \REVIEW{Phys. Rev. Lett.}{75}{1995}{2152}.
553:
554: \bibitem{HSG97}
555: \Name{Hou Q., Sasa S. \And Goldenfeld N.}
556: \REVIEW{Physica A}{239}{1997}{219}.
557:
558: \bibitem{BV01}
559: \Name{Boyer D. \And Vi\~nals J.}
560: \REVIEW{Phys. Rev. E}{64}{2001}{050101(R)}.
561:
562: \bibitem{CB98}
563: \Name{Christensen J. J. \And Bray A. J.}
564: \REVIEW{Phys. Rev. E}{58}{1998}{5364}.
565:
566: \bibitem{QM02}
567: \Name{Qian H. \And Mazenko G. F.}
568: \REVIEW{Phys. Rev. E}{67}{2003}{036102}.
569:
570: \bibitem{Har00}
571: \Name{Harrison C. \etal}
572: \REVIEW{Science}{290}{2000}{1558};
573: \REVIEW{Phys. Rev. E}{66}{2002}{011706}.
574:
575: \bibitem{coveney}
576: \Name{Gonzalez-Segredo N. \And Coveney P. V.}
577: \REVIEW{Phys. Rev. E}{69}{2004}{061501}.
578:
579: \bibitem{ohta}
580: \Name{Nonomura M., Yamada K. \And Ohta T.}
581: \REVIEW{J. Phys.:Cond. Matt.}{15}{2003}{L423}.
582:
583: \bibitem{Xu03}
584: \Name{Xu Aiguo, Gonnella G. \And Lamura A.}
585: \REVIEW{Phys. Rev. E}{67}{2003}{056105}.
586:
587: \bibitem{HV04}
588: \Name{Huang Z. F. \And Vi\~nals J.}
589: preprint cond-mat/0411284, to be published in Phys. Rev. E.
590:
591: \bibitem{CRZ94}
592: \Name{Coniglio A., Ruggiero P. \And Zannetti M.}
593: \REVIEW{Phys. Rev. E}{50}{1994}{1046}.
594:
595: \bibitem{On97}
596: See, e.g.,
597: \Name{Onuki A.}
598: \REVIEW{J. Phys.:Cond. Matt.}{9}{1997}{6119}.
599:
600: \bibitem{prl2}
601: \Name{Corberi F., Gonnella G. \And Lamura A.}
602: \REVIEW{Phys. Rev. Lett.}{83}{1999}{4057};
603: \REVIEW{ibid.}{81}{1998}{3852}.
604:
605: \end{thebibliography}
606:
607: \end{document}
608: