1: % This file is part of the APS files in the REVTeX 4 distribution.
2: % Version 4.0 of REVTeX, August 2001
3: %
4: % Copyright (c) 2001 The American Physical Society.
5: %
6: % See the REVTeX 4 README file for restrictions and more information.
7: %
8: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
9: % as well as the rest of the prerequisites for REVTeX 4.0
10: %
11: % See the REVTeX 4 README file
12: % It also requires running BibTeX. The commands are as follows:
13: %
14: % 1) latex apssamp.tex
15: % 2) bibtex apssamp
16: % 3) latex apssamp.tex
17: % 4) latex apssamp.tex
18: %
19: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
20: \documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
21:
22: % Some other (several out of many) possibilities
23: %\documentclass[preprint,aps]{revtex4}
24: %\documentclass[preprint,aps,draft]{revtex4}
25: %\documentclass[prb]{revtex4}% Physical Review B
26:
27: \usepackage{graphicx}% Include figure files
28: \usepackage{dcolumn}% Align table columns on decimal point
29: \usepackage{bm}% bold math
30:
31: %\nofiles
32:
33: \begin{document}
34:
35: %\preprint{APS/123-QED}
36:
37: \title{Equilibrium properties of highly asymmetric star-polymer mixtures}% Force line breaks with \\
38:
39: \author{Christian Mayer}
40: %\email{mayer@thphy.uni-duesseldorf.de}
41: \author{Christos N. Likos}
42: \author{Hartmut L\"owen}
43: \affiliation{Institut f{\"u}r Theoretische Physik II,
44: Heinrich-Heine-Universit{\"a}t D{\"u}sseldorf,
45: Universit{\"a}tsstra{\ss}e 1, D-40225 D{\"u}sseldorf, Germany
46: }
47:
48:
49:
50: \date{{\bf \today}, submitted to {\sl Physical Review E}}
51:
52: \begin{abstract}
53: We employ effective interaction potentials to study the equilibrium
54: structure and phase behavior of highly asymmetric mixtures of star
55: polymers. We consider in particular the influence of the addition of
56: a component with a small number of arms and a small size on a
57: concentrated solution of large stars with a high functionality.
58: By employing liquid integral equation theories we examine the evolution
59: of the correlation functions of the big stars upon addition of the
60: small ones, finding a loss of structure that can be attributed to
61: a weakening of the repulsions between the large stars due to the presence
62: of the small ones. We analyze this phenomenon be means of a generalized
63: depletion mechanism which is supported by computer simulations.
64: By applying thermodynamic perturbation theory we draw the phase
65: diagram of the asymmetric mixture, finding that the addition of small
66: stars melts the crystal formed by the big ones. A systematic comparison
67: between the two- and effective one-component descriptions of the
68: mixture that corroborates the reliability of the
69: generalized depletion picture is also carried out.
70: \end{abstract}
71:
72: \pacs{61.20.-p, 61.20.Gy, 64.70.-p}
73:
74: \maketitle
75:
76: \section{Introduction}
77:
78: Mixtures whose constituent particles show a high asymmetry in sizes
79: are quite common in soft matter physics. As a matter of fact, all
80: soft-matter systems are at least two-component mixtures, as they
81: are typically suspensions or dispersions of mesoscopically-sized
82: colloidal particles in a microscopic solvent. For many practical purposes,
83: though, it suffices to model the solvent as a continuous medium and
84: then an effective, one-component description of the suspended colloidal
85: particles is sufficient. The phenomenology is much richer when more
86: than one species of colloids is dispersed in the solvent and also there
87: the asymmetry in the sizes of the two kinds of colloidal particles can
88: be much higher than the one encountered in atomic or molecular fluids.
89: In the recent past, a great deal of attention has been paid to the
90: investigation of model colloid-polymer mixtures, in which the two
91: species are hard colloidal spheres and soft, flexible polymer
92: chains \cite{poon:review}.
93: The bulk of the theoretical analysis of such systems is carried out
94: within the framework of the Asakura-Oosawa (AO)
95: model \cite{ao:54, ao:56, vrij:76}, in which the
96: polymers are modeled as ideal, interpenetrating spheres that experience
97: a hard repulsion towards the colloids. Another popular system that
98: has attracted a lot of attention recently are binary hard-sphere
99: mixtures of various size ratios \cite{biben91, biben91a, md1, md2, md3}.
100: In both of those cases, attention is
101: usually focussed on the influence of the smaller component on the
102: structural and phase behavior of the larger one \cite{fuchs:review}.
103: Demixing phase
104: transitions and their competition to the crystallization transition
105: of the large hard spheres have been an issue of intensive investigations
106: in the past \cite{ilett:95, louis:99, dijkstra:99, matthias:prl:00, fuchs:00, matthias:jpcm:02, matthias:jcp:02, louis:prl:02}
107: with current research steering in the direction
108: of the study of interfacial and wetting properties of such
109: mixtures \cite{evans:04,brader:00, brader:01, brader:02, dijkstra:02, wijting:03, moncho:03, wessels:04},
110: as well as the influence of the additives on the vitrification
111: transition of the hard spheres \cite{bergenholtz:99, johner:01, foffi:02, puertas:03, bergenholtz:03, pham:04, zacca:04}.
112:
113: A convenient concept that has helped shed light into the phenomenology
114: of such asymmetric mixtures is that of the effective, {\it depletion
115: interaction} between the hard spheres, which is mediated by the
116: smaller component \cite{likos01, louis:jpcm:01}.
117: In the case of the AO-mixture, the depletion interaction
118: is purely attractive and has the range of the size of the added
119: polymer. For binary hard-sphere mixtures, the effective depletion
120: potential displays oscillatory behavior due to
121: correlation effects \cite{goetz:99, roth:00}.
122: Interpolating between the two extremes of ideal and hard additives
123: are star polymers of varying functionality, whose depleting effects on
124: hard spheres have been investigated both
125: theoretically \cite{dzubiella02, dzubiella02a} and
126: experimentally \cite{joe:pre}.
127:
128: The notion of depletion is almost exclusively invoked whenever the
129: large particles are hard colloids. Nevertheless, it can be expanded
130: in its interpretation to account for the modification of the properties
131: of the large particles in the presence of smaller ones also for arbitrary
132: kinds of interactions between the constituent particles. There is
133: relatively little done in this direction, however, with the
134: exception of the derivation of effective potentials in
135: Yukawa mixtures \cite{roth:02}
136: and in mixtures of star polymers and linear chains \cite{stiakakis02}.
137: In the last case,
138: it has been shown that the depletion mechanism of the chains on the
139: stars can account for the experimentally observed melting of the
140: star-polymer gel upon addition of linear polymer. In this paper, we
141: turn our attention to two-component mixtures in which all particle
142: species interact by means of soft potentials and, in particular,
143: to mixtures of two kinds of star polymers: large ones with a high
144: number of arms and small ones will a low arm number. All species
145: interact via logarithmic-Yukawa pair potentials. We find that in this
146: case the depletion mechanism of the small stars on the big ones
147: has the effect of reducing the repulsive potential between the latter
148: and thus it brings about a melting of the colloidal crystal formed
149: by the large stars. Concomitant to this effect is a partial loss
150: of correlations between the centers of the big stars, manifested in
151: a drastic lowering of the peak height of their partial structure
152: factor. Upon addition of a sufficiently large quantity of depleting
153: agents, even an effective {\it attraction} between the large stars
154: shows up, resulting in a demixing spinodal between the two species.
155:
156: The rest of the paper is organized as follows: In Sec.\ \ref{twoc:sec}
157: we present the pair potentials and the full, two-component description
158: of the mixture, examining the effects of the depletants on the
159: structural correlations of the big stars. In Sec.\ \ref{onec:sec},
160: we formally trace out the small stars and examine the resulting
161: effective, one-component interactions between the big ones. This
162: effective potential is employed, in turn,
163: in order to draw the phase diagram of the system in Sec.\ \ref{phase:sec},
164: where thermodynamic perturbation theory is used for the calculation
165: of the Helmholtz free energies of the fluid and solid phases.
166: In Sec.\ \ref{compare:sec} we carry out a comparison between the
167: one- and full two-component descriptions of the mixture and demonstrate
168: the validity of the former, whereas in Sec.\ \ref{concl:sec} we summarize
169: and draw our conclusions.
170:
171: \section{Two-component description}
172: \label{twoc:sec}
173:
174: We consider binary mixtures of star-polymers which differ in
175: terms of their sizes and arm numbers (functionalities).
176: The system consists of $N_1$ stars of corona diameter $\sigma_1$ and
177: functionality $f_1$ and $N_2$ stars,
178: characterized by $\sigma_2$ and $f_2$, in a volume $V$.
179: We first calculate the properties of the binary fluid. To obtain
180: information about the pair correlations between the constituent particles,
181: we describe the system using the full two-component picture
182: for the mixture of the two different star-polymer species.
183: The structural quantities we calculate
184: are used as input for the mapping onto an effective
185: one-component system in Sec.\ \ref{onec:sec}.
186: We define the size ratio of the different species as
187: $q=\sigma_2/\sigma_1 < 1$.
188: Let $\rho_i=N_i/V$ ($i=1,\;2$) be the
189: partial number densities of the two species.
190:
191: We start from the effective pair potentials between the mesoscopic
192: particles, having traced out the monomer and solvent degrees of freedom.
193: The effective interaction between
194: the star-polymers diverges logarithmically
195: with the center-to-center distance $r$ as $r\to0$,
196: as derived by Witten and Pincus \cite{witten86}.
197: A full expression for identical star-polymers,
198: which is valid for all star separations,
199: has been derived and verified by neutron scattering and monomer resolved
200: molecular simulation \cite{likos98, jusufi99}.
201: The pair potential is given by an ultrasoft
202: interaction which shows logarithmic behavior for
203: small distances and an exponential Yukawa-type
204: decay at large star--star separation \cite{likos98, watzlawek99}.
205:
206: In the case of mixtures we need an expression for the
207: effective interaction between star polymers in an athermal solvent that differ
208: in their sizes $\sigma_1$, $\sigma_2$
209: \footnote{The $\sigma_i$ are proportional to the corresponding radii of gyration, as described in Ref.\ \cite{vonferber00}.}
210: and functionalities
211: $f_1$, $f_2$, as a function of their center-to-center separation $r$.
212: In this work we use the effective pair potential which was put forward
213: by means of field-theoretical arguments
214: and confirmed by molecular dynamics computer
215: simulations in Ref.\ \cite{vonferber00}, namely:
216: \begin{widetext}
217: \begin{equation}\label{eq.pot1}
218: \beta V_{ij}=\Theta_{ij}
219: \left\{
220: \begin{aligned}
221: &-\ln\left(\frac{r}{\sigma_{ij}}\right)+\frac{1}{1+\sigma_{ij}\kappa_{ij}}&\qquad&\text{for $r\leq\sigma_{ij}$;}\\
222: &\frac{1}{1+\sigma_{ij}\kappa_{ij}}\left(\frac{\sigma_{ij}}{r}\right)\exp(\sigma_{ij}\kappa_{ij}-r\kappa_{ij}) &\qquad&\text{else,}
223: \end{aligned}
224: \right.
225: \end{equation}
226: \end{widetext}
227: where $\sigma_{ij}=(\sigma_i+\sigma_j)/2$, $1/\kappa_{ij}=\sigma_i/\sqrt{f_i}+\sigma_j/\sqrt{f_j}$ and
228: \begin{equation}
229: \Theta_{ij}=\frac{5}{36}\frac{1}{\sqrt{2}-1}\left[(f_i+f_j)^{3/2}-(f_i^{3/2}+f_j^{3/2})\right].
230: \end{equation}
231: Moreover, $\beta=(k_\mathrm{B}T)^{-1}$
232: is the inverse temperature, with $k_\mathrm{B}$
233: being Boltzmann's constant. Since all three interactions are purely entropic,
234: the $\beta V_{ij}(r)$ are independent of the temperature.
235: For $i=j$ the potential reduces to the interaction of
236: identical star-polymers which was introduced in \cite{likos98}.
237: In what follows, we fix the functionality of the large stars to
238: $f_1 = 263$ in order to make contact with recently performed
239: experiments \cite{stiakakis02} in which smaller
240: polymeric entities were used as additives in gelated solutions of
241: the large stars in order to examine their overall influence on the
242: rheology of the mixture. This functionality is large enough for
243: the star polymers to crystallize into a fcc-structure roughly
244: at their overlap concentration \cite{watzlawek99}. For the small
245: stars, we considered functionalities $f_2 = 16$ and $32$ and
246: size ratios $q$ in the range between $0.1$ and $0.3$.
247:
248: The pair structure of the mixture can now be calculated
249: using the Ornstein-Zernike (OZ) equations for binary mixtures
250: together with the two-component Rogers-Young (RY)
251: closure.
252: The pair correlations of the system are described by three
253: independent total
254: correlation functions $h_{ij}(r)$, $i\leq j=1,2$,
255: since the symmetry with respect to the exchange of indices
256: dictates $h_{ij}(r)=h_{ji}(r)$ for $i \ne j$. In addition, we
257: have the same number of direct correlation functions $c_{ij}(r)$.
258: The Fourier transforms of these quantities are denoted
259: by $\tilde{h}_{ij}(k)$ and $\tilde{c}_{ij}(k)$,
260: respectively.
261:
262: For multicomponent mixtures, the OZ relation
263: takes the form \cite{lebowitz64, likos01}
264: \begin{equation}\label{eqOZmult}
265: \mathbf{\tilde{H}}(k)=\mathbf{\tilde{C}}(k)+\mathbf{\tilde{C}}(k)\cdot\mathbf{D}\cdot\mathbf{\tilde{H}}(k),
266: \end{equation}
267: where $\mathbf{\tilde{H}}(k)$ and $\mathbf{\tilde{C}}(k)$ are symmetric $\nu\times\nu$ matrices with
268: \begin{equation}
269: [\mathbf{\tilde{H}}(k)]_{ij}=\tilde{h}_{ij}(k) \qquad\text{and}\qquad [\mathbf{\tilde{C}}(k)]_{ij}=\tilde{c}_{ij}(k).
270: \end{equation}
271: $\mathbf{D}$ is a diagonal $\nu\times\nu$ matrix with
272: \begin{equation}
273: [\mathbf{D}]_{ij}=\rho_i\delta_{ij}.
274: \end{equation}
275:
276: From Eq.\ (\ref{eqOZmult}) we obtain three independent equations
277: for the six unknown functions $\tilde{h}_{ij}(k)$ and $\tilde{c}_{ij}(k)$,
278: $i,j = 1,2$.
279: In order to obtain a solvable system, we need three additional
280: {\it closure equations} between these functions.
281: The Rogers-Young closure for multicomponent
282: mixtures reads as \cite{likos01}
283: \begin{equation}\label{eq.RYmult}
284: g_{ij}(r)=\exp[-\beta V_{ij}(r)]\left[1+\frac{\exp[\gamma_{ij}(r)f_{ij}(r)]-1}{f_{ij}(r)}\right],
285: \end{equation}
286: where $g_{ij}(r)=h_{ij}(r)+1$, $\gamma_{ij}(r)=h_{ij}(r)-c_{ij}(r)$ and $V_{ij}(r)$ is the pair interaction between particles of species $i$ and $j$.
287: The ``mixing function'' $f_{ij}(r)$ is
288: defined as
289: \begin{equation}\label{eqmixing}
290: f_{ij}(r)=1-\exp(-\alpha_{ij}r).
291: \end{equation}
292: Usually, the same self-consistency parameter
293: $\alpha=\alpha_{ij}$ is used for all components of the mixture.
294: This allows to
295: fulfill one thermodynamic consistency requirement,
296: namely the equality between the ``virial'' and ``fluctuation''
297: total compressibilities of the mixture.
298: Multi-parameter versions have also been proposed \cite{biben91},
299: invoking thermodynamic consistency for the partial
300: compressibilities of each species.
301: For $\alpha\to 0$ Eq.\ (\ref{eq.RYmult}) reduces to
302: the Percus-Yevick (PY) and for $\alpha\to\infty$
303: to the hypernetted chain (HNC) multicomponent closures.
304: When dealing with star-polymers, which feature a soft repulsion of relatively short range,
305: neither the PY nor the HNC closure are adequate to capture the details of the correlation functions
306: with high accuracy, therefore employing the full RY closure is essential \cite{watzlawek98}.
307:
308: In our work we solve the OZ equation with the RY-closure
309: [Eqs.\ (\ref{eqOZmult})--(\ref{eqmixing})] for the two-component mixture.
310: The effective interactions
311: between the star polymers are given by Eq.\ (\ref{eq.pot1}).
312: The thermodynamic consistency of the RY closure
313: was obtained by using a single parameter $\alpha$.
314: The structure of the binary mixture can be described either by
315: the partial radial distribution functions $g_{ij}(r) = h_{ij}(r) + 1$
316: in real space or by the
317: three partial static structure factors
318: $S_{ij}(k)=\delta_{ij}+\sqrt{\rho_i\rho_j}\tilde{h}_{ij}(k)$ in
319: wavenumber space.
320: The structure factors are relevant in comparing with experiments,
321: because they can be measured via scattering techniques.
322:
323: \begin{figure}
324: \centering
325: \includegraphics[width=6cm, clip=true]{fig1a.eps}
326: \includegraphics[width=6cm, clip=true]{fig1b.eps}
327: \caption{The partial (a) radial distribution functions and
328: (b) static structure factors for species 1 (big stars) in a mixture
329: with small ones. The density of species 1 is $\rho_1\sigma_1^3=0.05$.
330: The structure has been calculated using the two component OZ equations with
331: the RY closure. The plotted lines are for different densities $\rho_2$,
332: as indicated in the legend. In this case, $q=0.1$ and $f_2=32$.
333: The partial structure factor grows for $k\to 0$ with increasing $\rho_2$,
334: as we approach the demixing spinodal line of the system.}
335: \label{figspinodal}
336: \end{figure}
337:
338:
339: \begin{figure}
340: \centering
341: \includegraphics[width=6cm, clip=true]{fig2a.eps}
342: \includegraphics[width=6cm, clip=true]{fig2b.eps}
343: \caption{Same as Fig.\ \ref{figspinodal} but now for big star density
344: $\rho_1\sigma_1^3=0.3$.
345: By increasing
346: the density of the smaller component,
347: the structure of the fluid diminishes.}
348: \label{fig.hansenverlet}
349: \end{figure}
350:
351: In Fig.\ \ref{figspinodal}(a) we show results for the radial distribution
352: function $g_{11}(r)$ between the large stars in a dilute solution, and
353: its evolution upon increasing the concentration of small additives with
354: $f_2 = 32$ and $q = 0.1$. Although for very small concentration
355: of smaller stars the function
356: $g_{11}(r)$ has a relatively structureless shape, it rapidly
357: develops a pronounced peak when $\rho_2$ is further increased. This is
358: a first indication of clustering of big stars, which has its physical
359: origin in some effective attraction induced by the small component.
360: One physically expects that when this attraction becomes sufficiently
361: strong, a demixing transition between the two species will take place.
362: This hypothesis is corroborated by the evolution of the corresponding
363: structure factor $S_{11}(k)$, shown in Fig.\ \ref{figspinodal}(b).
364: A fluid--fluid demixing binodal is indicated by
365: the divergence of all partial structure factors in
366: the long-wavelength limit $k\to 0$. As can be seen in
367: Fig.\ \ref{figspinodal}(b),
368: a growth of the $k \to 0$-limit occurs upon
369: increasing $\rho_2$. The existence of
370: a demixing spinodal will be confirmed in Sec.\ \ref{phase:sec}
371: where we draw the phase diagram of the mixture.
372:
373: We now examine the effect of the additives at the complementary
374: regime of high concentration of large stars, and in particular
375: slightly above their overlap concentration, in which the latter
376: are in a thermodynamically stable
377: crystalline state \cite{watzlawek99} or in a dynamically
378: arrested gel state \cite{foffi03}.
379: We derive the partial structure factor $S_{11}(k)$
380: of the (metastable) fluid in the absence of small stars and monitor
381: its evolution as $\rho_2$ is increased. Representative results are
382: shown in Fig.\ \ref{fig.hansenverlet}. In Fig.\ \ref{fig.hansenverlet}(a),
383: already a loss of correlations in real space can be discerned,
384: as witnessed by the broadening and lowering of the coordination
385: peaks in $g_{11}(r)$. Moreover, the large stars approach closer to
386: each other upon an increase of $\rho_2$, an effect that can be
387: interpreted as a weakening of the strength of their mutual repulsion.
388: As can be seen in Fig.\ \ref{fig.hansenverlet}(b),
389: the principle peak height of the structure factor
390: of species 1 {\it diminishes} as the density of the smaller
391: component is increased. The Hansen-Verlet criterion \cite{hansen69}
392: states that a fluid solidifies when the maximum
393: of the structure factor exceeds the threshold value
394: $S_{\rm th}(k_{\rm max}) = 2.85$.
395: Therefore, the diminishing of structure in the
396: system is a first indication for the melting
397: of the crystal of big star-polymers by addition
398: of the smaller species. In Fig.\ \ref{fig.hansenverlet}(b)
399: it can be seen that the first peak of the structure
400: factor is bigger than 3 for $\rho_2\sigma_2^3=0$.
401: Already small densities of the smaller component
402: lead to a drastic decline of the peak height, a finding that is in line
403: with recent experimental and theoretical results on mixtures of
404: star polymers with {\it linear} chains \cite{stiakakis02}.
405:
406: \section{Effective one-component description}
407: \label{onec:sec}
408:
409: We now wish to put the assumptions regarding the influence of the
410: additives on the effective interaction of the big stars into a concrete
411: test, by calculating an effective potential $V_{\rm eff}(r)$ between
412: the latter in the presence of the former. To this end,
413: we carry out the mapping of the two-component system
414: onto an effective one-component description, in which the degrees
415: of freedom of the smaller star-polymers
416: have been traced out.
417: The interactions
418: cause spatial correlations of the density of small stars in the vicinity
419: of two big ones, influencing thereby the shape of the resulting
420: generalized depletion interaction.
421: There are different methods to obtain these effective interactions.
422: All of them omit many-body forces and reduce the interaction
423: to an additive pair potential, we will confirm however
424: that many-body effects only play a minor role.
425:
426: Instead of the so-called
427: `system representation',
428: in which the two densities $\rho_1$ and $\rho_2$ in the mixture are given,
429: we now switch into the more convenient `reservoir representation' $\rho_1$
430: and $\rho_2^\mathrm{r}$.
431: Since the effective interaction between the large stars depends rather on the
432: chemical potential $\mu_2$ of the small ones rather than on
433: their density $\rho_2$, this description is
434: more convenient \cite{md3}. The reservoir is a system consisting of pure
435: small stars and their density there, $\rho_2^\mathrm{r}$, is determined
436: by the requirement that the partial chemical potentials $\mu_2$ in the
437: real system and $\mu_2^\mathrm{r}$ in the reservoir are equal. Clearly, due to the
438: finite value of the density $\rho_1$ in the system, it must hold
439: $\rho_2 \ne \rho_2^\mathrm{r}$. The mapping between the two densities,
440: depending parametrically on the big star density $\rho_1$, will be
441: carried out in Sec.\ \ref{compare:sec}.
442:
443: \begin{figure}
444: \centering
445: \includegraphics[width=7cm, clip=true]{fig3a.eps}
446: \includegraphics[width=7cm, clip=true]{fig3b.eps}
447: \includegraphics[width=7cm, clip=true]{fig3c.eps}
448: \caption{Comparison of Monte Carlo simulation (MC),
449: inversion of Ornstein-Zernike equation (OZ),
450: and superposition approximation (SA) results for the depletion
451: forces between the big stars. (a) Results for $q=0.3$, $f_2=32$
452: and $\rho_2^\mathrm{r}\sigma_2^3=0.027$; (b) $q$ and $f_2$ same as in (a)
453: but $\rho_2^\mathrm{r}\sigma_2^3=0.081$; (c) for $q=0.2$, $f_2=32$
454: and $\rho_2^\mathrm{r}\sigma_2^3=0.08$.}
455: \label{fig.depl}
456: \end{figure}
457:
458: \subsection{Simulations}\label{subsec.sim}
459:
460: The most accurate way to obtain the effective interactions
461: between the big star-polymers is a computer simulation
462: of the mixture \cite{damico97, allahyarov98, allahyarov01}.
463: We place two big stars of species 1 at fixed positions
464: $\mathbf{R}_1$ and $\mathbf{R}_2$ along the diagonal
465: of the simulation cube, so that their common center
466: coincides with the center of the cube.
467: They are surrounded by the smaller species that move according to the
468: forces dictated by the effective interactions of Eq.\ (\ref{eq.pot1}).
469: Since we have only two big star-polymers in our simulation box,
470: the density is $\rho_1\to 0$. Therefore the simulation provides
471: directly the sought-for effective force as a function of the
472: reservoir density $\rho_2^\mathrm{r}$.
473:
474: We use standard $NVT$ Monte Carlo simulation with periodic
475: boundary conditions and minimum image convention.
476: The length of the cubic simulation box
477: is $L=5\sigma_1$, so that the number of small
478: stars in the simulation results which
479: are shown in Fig.\ \ref{fig.depl} is between 125 and 1250.
480: For each particle up to 5 million
481: Monte Carlo steps are calculated, where the maximum
482: displacement of the particles is chosen
483: in such a way that half the steps will be accept.
484: The force is then measured after every 1000 simulation steps.
485: Due to the presence of the
486: second big star, the density distribution around
487: each star is not spherically symmetric.
488: This leads to an average nonvanishing force between them,
489: which is mediated by the small
490: star-polymers and parallel to the vector
491: $\mathbf{R}_{12}=\mathbf{R}_2-\mathbf{R}_1$.
492: Due to the symmetry of the system, the components perpendicular
493: to $\mathbf{R}_{12}$ have to vanish.
494: The force $\mathbf{F}_1$ acting on the particle
495: at $\mathbf{R}_1$ can then be calculated by averaging
496: over the simulation results, namely:
497: \begin{equation}
498: \mathbf{F}_1(R_{12})=\left<-\sum_{j=1}^{N_2}\nabla_{\mathbf{R}_1}V_{12}(|\mathbf{R}_1-\mathbf{r}_j|)\right>_{\mathbf{R}_{12}},
499: \end{equation}
500: where $\mathbf{r}_j$ are the star-polymer positions
501: of species 2 and $\left<\cdots\right>_{\mathbf{R}_{12}}$
502: denotes the statistical average, taken under the constraint
503: of constant $\mathbf{R}_{12}$. Clearly, the effective force
504: satisfies the relation $\mathbf{F}_2(R_{12})=-\mathbf{F}_1(R_{12})$.
505: We further define the {\it depletion} force ${\bf F}_\mathrm{dep}(R_{12})$
506: as the difference between $\mathbf{F}_2(R_{12})$ and the direct
507: force ${\bf F}_{\rm {dir}}(R_{12})$ between the two stars due to their
508: direct interaction potential $V_{11}(R_{12})$.
509: The magnitude of the depletion force $F_\mathrm{dep}(R_{12})$ is
510: then given by
511: \begin{equation}
512: F_\mathrm{dep}(R_{12})=
513: \frac{\mathbf{R}_{12}}{R_{12}}\cdot\mathbf{F}_{\rm dep}(R_{12}).
514: \end{equation}
515: Accordingly,
516: the total effective interaction between the big star-polymers
517: in a sea of the smaller species is the sum of their
518: interaction potential $V_{11}(r)$ and the depletion potential
519: $V_\mathrm{dep}(r)$:
520: \begin{equation}
521: V_{\rm eff}(r) = V_{11}(r) + V_\mathrm{dep}(r).
522: \label{vdep:eq}
523: \end{equation}
524:
525: A large number of long simulation runs is required to
526: to obtain accurate depletion forces with good statistics, which renders
527: this approach inefficient if ones needs to calculate $V_{\rm eff}(r)$
528: for arbitrary values of $q$, $f_2$ and $\rho_2^{\rm r}$. Thus, we
529: resorted to
530: approximative theoretical methods to calculate
531: the effective interactions and used
532: the simulation results at selected parameter combinations in order to
533: put the theoretical approximations into test. The two theoretical
534: approaches invoked in this work are the inversion of the
535: Ornstein-Zernike equation and the superposition approximation,
536: which are presented below.
537:
538: \subsection{Inversion of the Ornstein-Zernike equation}\label{subsecinvOZ}
539:
540: The effective potential can be obtained by inversion
541: of the two-component OZ equation results in the
542: limit of low density of big stars \cite{dzubiella02, mendez00, konig01}.
543: It can be shown from diagrammatic expansions in the theory of
544: liquids \cite{hansen86} that the radial distribution
545: function $g(r)$ of any fluid whose constituent particles
546: interact via the pair potential $V(r)$, reduces to the Boltzmann factor
547: $g(r)=\exp[-\beta V(r)]$ in the low-density limit.
548: The effective interaction between the big stars
549: depends on the reservoir density $\rho_2^{\rm r}$ of the
550: smaller component. The interaction can be obtained by
551: solving the full two-component OZ equations with the RY closure
552: for different small-component densities $\rho_2$
553: in the limit $\rho_1 \to 0$; due to the latter limit, it then also
554: holds $\rho_2 = \rho_2^{\rm r}$.
555: The radial distribution function $g_{11}(r)$ can then be
556: inverted to yield the effective potential as
557: \begin{equation}
558: \beta V_\mathrm{eff}(r)=
559: -\lim_{\rho_1\to 0}\ln[g_{11}(r;\rho_1,\rho_2^\mathrm{r})].
560: \end{equation}
561: Thereafter, the depletion force $F_\mathrm{dep}(r)$ can be calculated as
562: $F_\mathrm{dep}(r)=-\partial[V_\mathrm{eff}(r)-V_{11}(r)]/{\partial r}$
563: and compared them to the simulation results of the preceding subsection.
564: Selected comparisons are shown in
565: Fig.\ \ref{fig.depl}, where it can be seen that the inversion of the
566: OZ relation yields very reliable results. We emphasize here that the
567: approximate character of the OZ-inversion technique lies exclusively
568: in the approximations involved in solving the two-component integral
569: equation theories, i.e., in the Rogers-Young (or any other chosen)
570: closure relation. Otherwise, the method is based on the exact statement
571: that the radial distribution function of a one-component system at
572: low densities is equal to the Boltzmann factor of the associated
573: pair potential and hence the agreement of the inversion method with
574: the simulation results comes as no surprise. It rather corroborates
575: the fact that the two-component RY-closure is very accurate whenever
576: one deals with soft, repulsive interactions, a result already seen
577: in the case of mixtures between hard spheres and
578: star polymers \cite{dzubiella02a}.
579:
580: \subsection{Superposition approximation}
581:
582: Another possibility to derive the effective
583: interaction is the \textit{superposition approximation}
584: (SA) \cite{attard89}. If the density distribution
585: $\rho_2(\mathbf{r}; \mathbf{R}_1,\mathbf{R}_2)$ of the small stars
586: around two big stars held fixed at positions ${\bf R}_1$ and ${\bf R}_2$
587: is known, then the depletion force
588: in the low-density limit can be calculated by a simple integration.
589: The density $\rho_2(\mathbf{r}; \mathbf{R}_1,\mathbf{R}_2)$
590: is proportional to three-body distribution function
591: $g_{112}(\mathbf{R}_1, \mathbf{R}_2, \mathbf{r})$,
592: which expresses the probability density of finding a particle of
593: species 2 at position ${\bf r}$, given that two particles of
594: species one are fixed at positions ${\bf R}_1$ and
595: ${\bf R}_2$. This
596: function is in general unknown; a usual procedure is to
597: approximate it by the product of pair distribution
598: functions \cite{attard89}.
599:
600: We consider two big stars at the positions
601: $\mathbf{R}_1$ and $\mathbf{R}_2$ and
602: choose, without loss of generality, $\mathbf{R}_1=0$.
603: Let the distance between the particles be $R_{12}$.
604: The surrounding smaller star-polymers have the
605: density $\rho_2(\mathbf{r}; \mathbf{R}_1,\mathbf{R}_2)$.
606: By taking the average for fixed $R_{12}$ we obtain the depletion force as
607: \begin{eqnarray}
608: F_\mathrm{dep}(R_{12})&=&-2\pi\int_0^\infty r^2\frac{\mathrm{d}V_{12}(r)}{\mathrm{d}r}\mathrm{d}r \nonumber \\
609: &&\int_{-1}^1\rho_2(\mathbf{r}; \mathbf{R}_1,\mathbf{R}_2)\omega \mathrm{d}\omega,
610: \end{eqnarray}
611: where $\omega=\cos\theta$.
612:
613: \begin{figure}[h]
614: \includegraphics[width=6cm, clip=true]{fig4.eps}
615: \caption{A sketch of two big stars
616: at a distance $R_{12}$ with
617: $\rho_2(\mathbf{r}; \mathbf{R}_1, \mathbf{R}_2)$
618: denoting the density of the smaller stars at $\mathbf{r}$.
619: The density distribution
620: depends on
621: the positions of the two big star-polymers.}
622: \end{figure}
623:
624: Since $\rho_2(\mathbf{r}; \mathbf{R}_1,\mathbf{R}_2)$
625: is in general not known, at this point the exact
626: density distribution has to be approximated. The density distribution
627: around two big stars is replaced by the product of the
628: density distributions around two isolated star-polymers
629: at the positions $\mathbf{R}_1$ and $\mathbf{R}_2$,
630: respectively. The SA then reads as
631: \begin{equation}
632: \rho_2(\mathbf{r}; \mathbf{R}_1,\mathbf{R}_2)\approx\rho_2^\mathrm{r}g_{12}(|\mathbf{r}-\mathbf{R}_1|)g_{12}(|\mathbf{R}_2-\mathbf{r}|),
633: \end{equation}
634: where $\rho_2^\mathrm{r}$ is the reservoir density, again identical to
635: the system density for the situation at hand, since only two big stars
636: are considered in the thermodynamic limit and thus $\rho_1 = 0$.
637: The functions $g_{12}(|\mathbf{R}_i-\mathbf{r}|)$
638: are the radial distribution functions of small star polymers surrounding a
639: single large one.
640: Therefore, they can be obtained in the $\rho_1 \to 0$-limit
641: of the two-component OZ equations.
642: Using simple geometrical considerations,
643: we obtain $|\mathbf{r}-\mathbf{R}_1|=\sqrt{R_{12}^2+r^2-R_{12}r\omega}$.
644: Finally we obtain for the depletion force in the SA the expression:
645: \begin{eqnarray}
646: \nonumber
647: F_\mathrm{dep}(R_{12})&=&-2\pi\rho_2^\mathrm{r}\int_0^\infty r^2\frac{\mathrm{d}V_{12}(r)}{\mathrm{d}r}g_{12}(r)\mathrm{d}r
648: \\
649: \nonumber
650: &\times& \int_{-1}^1g_{12}\left(\sqrt{R_{12}^2+r^2-R_{12}r\omega}\right)\omega \mathrm{d}\omega.
651: \\
652: & &
653: \end{eqnarray}
654:
655: The results we obtain by this method are compared
656: to Monte Carlo and to those derived from inversion of the
657: Ornstein-Zernike equation in Fig.\ \ref{fig.depl}. The results are
658: very similar to the ones we obtain by inverting the OZ equation
659: and both approximations yields reasonable agreement with the
660: simulation data. Therefore, it is
661: possible to choose the results of either approximation
662: for calculating the phase diagrams and we expect that only minor
663: quantitative differences will be seen by employing the one or the
664: other theoretical approach. We have chosen to work with the effective
665: interactions that result from the inversion of the OZ-equation since
666: the latter is based on an exact statement, whereas the SA has an
667: approximate nature.
668:
669: \subsection{Effective interactions}\label{subsec.effint}
670:
671: \begin{figure*}
672: \begin{center}
673: \includegraphics[width=6cm, clip=true]{fig5a.eps}
674: \includegraphics[width=6cm, clip=true]{fig5b.eps}
675: \includegraphics[width=6cm, clip=true]{fig5c.eps}
676: \includegraphics[width=6cm, clip=true]{fig5d.eps}
677: \includegraphics[width=6cm, clip=true]{fig5e.eps}
678: \includegraphics[width=6cm, clip=true]{fig5f.eps}
679: \end{center}
680: \caption{The effective potential $V_\mathrm{eff}(r)$ between two big
681: star polymers in the presence of a sea of small ones. The various
682: combinations of parameters regarding the density, functionality and
683: size of the additives are shown in the legends. Notice the development
684: of a strong attractive part in the interaction for the case $q = 0.1$.}
685: \label{figveff}
686: \end{figure*}
687:
688: We have chosen to employ the effective
689: interactions that result from the inversion of the OZ-equation since
690: the latter is based on an exact statement, whereas the SA has an
691: approximate nature.
692: Representative results are shown
693: in Fig.\ \ref{figveff}. For the lowest size ratio, $q = 0.1$,
694: we see that irrespective of the functionality of the depletants
695: ($f_2 = 16$ or $32$), the following scenario materializes: as
696: $\rho_2^{\rm r}$ increases, first a weakening of the repulsions takes
697: place, followed by the development of an attraction between the
698: stars at sufficiently high reservoir densities, see
699: Figs.\ \ref{figveff}(a) and \ref{figveff}(b). These findings
700: provide a possible physical realization of the recently-proposed
701: model ultrasoft repulsion potentials that are accompanied by
702: an attractive part \cite{loverso:03}.
703: This attraction is more pronounced for $f_2=32$ than for $f_2=16$,
704: if one compares two systems with equal
705: density $\rho_2^\mathrm{r}\sigma_2^3$. This result is not surprising,
706: since the $f_2 = 32$-stars exert a higher osmotic pressure on the
707: large ones than the $f_2 = 16$-stars and can therefore reduce the
708: direct repulsions and induce attractions more efficiently.
709:
710: Novel features in the effective potential appear for higher size ratios,
711: $q = 0.2$ and $q = 0.3$. As can be seen in Fig.\ \ref{figveff}(c)-(f),
712: an oscillatory structure appears in the effective potential, which
713: is akin to that seen for hard-sphere mixtures of two different sizes.
714: Contrary to this case, however, a deep attraction between the big
715: stars does not develop and, therefore, it seems that a demixing
716: transition between the two species does not exist when the sizes
717: of the two stars become more and more similar.
718: In all cases, however, the range of the repulsion decreases
719: due to the depletion effect, i.e., the
720: big star polymers appear, in the presence of the small ones,
721: to be softer than they are in a pure solvent. Another possible interpretation,
722: to be elaborated on in what follows, is that the big stars appear to be
723: `smaller', i.e., they acquire a reduced
724: effective hard sphere packing fraction as a result of the depletants.
725: Since the star-polymers then need less space, they
726: become more mobile so the solid can melt.
727: This property will be discussed
728: in more detail in Sec.\ \ref{subsec.phaseresults}.
729:
730: \section{Phase diagrams}
731: \label{phase:sec}
732:
733: \subsection{Hard sphere mapping}
734:
735: \begin{figure}
736: \centering
737: \includegraphics[width=7.2cm, clip=true]{fig6.eps}
738: \caption{The two possible effective hard sphere diameters $d$ and $d_0$
739: pertaining to the effective interaction $V_{\rm eff}(r)$ between big
740: stars, against the reservoir density $\rho_2^{\rm r}$. The parameter
741: combination here is $f_2=32$ and $q=0.1$.
742: As explained in the text, $d$ is calculated using the full
743: effective interaction and $d_0$ only for the reference part.}
744: \label{fig.diameter}
745: \end{figure}
746:
747: In order to trace out the phase diagram of the mixture in the
748: $(\rho_1, \rho_2^{\rm r})$-representation, we first perform a
749: mapping of the effective one-component interaction $V_{\rm eff}(r)$
750: between the big stars onto an effective hard-sphere system of
751: diameter $\sigma$. Clearly, the latter depends on the reservoir density
752: $\rho_2^{\rm r}$ as well as on the system parameters $q$ and $f_2$.
753: For the purposes of performing the mapping in a physically meaningful
754: way, we distinguish between two cases.
755:
756: First, we consider the case in which $V_{\rm eff}(r)$ is either
757: free of attractive parts or positive definite or, at most, it
758: contains negative parts not exceeding a small fraction of $k_{\rm B}T$
759: in magnitude. In this case, it is physically meaningful to identify
760: $\sigma$ with the Barker-Henderson hard sphere diameter of the
761: {\it full} effective interaction $V_{\rm eff}(r)$, $d$, defined
762: as \cite{barker67}:
763: \begin{equation}
764: d=\int_0^\infty\left\{1-\exp[-\beta V_{\rm eff}(r)]\right\}\mathrm{d}r.
765: \label{eq.barkerhend}
766: \end{equation}
767: Most of the curves shown in Fig.\ \ref{figveff} fall into this category.
768: An important exception are the curves pertaining to
769: $\rho_2^{\rm r}\sigma_2^3 = 0.1$ in Fig.\ \ref{figveff}(a) and
770: to $\rho_2^{\rm r}\sigma_2^3 = 0.05$ in Fig.\ \ref{figveff}(b). For
771: these combinations, and also for all others at even higher
772: reservoir densities, a deep negative minimum appears in $V_{\rm eff}(r)$
773: and application of Eq.\ (\ref{eq.barkerhend}) to such cases would
774: lead to unphysically small and even negative effective hard sphere
775: diameters. For such combinations, it is physically appealing to
776: separate the effective potential $V_{\rm eff}(r)$ into a purely
777: repulsive part $V_0(r)$ and a perturbation part $V_{\rm {pert}}(r)$,
778: by truncating and shifting upwards the full interaction at the
779: deepest minimum \cite{barker67}. In this second case, it is pertinent
780: to define another effective hard sphere diameter, $d_0$, that is
781: associated with $V_0(r)$ only and is calculated again from the
782: Barker-Henderson recipe, namely
783: \begin{equation}
784: d_0=\int_0^\infty\left\{1-\exp[-\beta V_0(r)]\right\}\mathrm{d}r.
785: \label{eq.barkerhend0}
786: \end{equation}
787:
788: In attempting to choose and match between the two possible hard sphere
789: diameters, $d$ and $d_0$, we are confronted with a technical difficulty.
790: The evolution of the potential $V_{\rm eff}(r)$ with $\rho_2^{\rm r}$
791: is continuous and the appearance of negative minima is in general
792: accompanied by a soft repulsive barrier after the minimum. The effective
793: hard sphere diameter, on the other hand, has to be a continuous
794: function of $\rho_2^{\rm r}$, so as to avoid unphysical jumps of the
795: phase boundaries in the phase diagram. In Fig.\ \ref{fig.diameter}
796: we show a typical result for the dependence of $d$ and $d_0$
797: on $\rho_2^{\rm r}$. For low values of $\rho_2^{\rm r}$, where
798: $V_{\rm eff}(r)$ is purely repulsive, $d$ is a meaningful measure
799: of the effective hard-sphere diameter. On the other hand, at high
800: values of $\rho_2^{\rm r}$, where a deep attraction between the big stars
801: effectively sets in, it is $d_0$ that most realistically captures
802: the physics of the repulsions. The two curves cross at some point
803: and, in order to guarantee both the continuity of $\sigma$ as a
804: function of $\rho_2^{\rm r}$ and its correct asymptotic behavior for
805: small and large values of $\rho_2^{\rm r}$, we choose
806: \begin{equation}
807: \sigma = \max\{d,d_0\}.
808: \label{sigma:eq}
809: \end{equation}
810: It is then clear from the discussion above that for the perturbation
811: part, $V_{\rm {pert}}(r)$, of the interaction, it holds
812: \begin{equation}
813: V_{\rm {pert}}(r) = 0 \qquad{\rm {if}\,} d > d_0.
814: \label{pert:eq}
815: \end{equation}
816:
817: The phase diagrams can now be calculated using standard
818: first-order perturbation theory \cite{hansen86} and taking
819: Eqs.\ (\ref{sigma:eq}) and (\ref{pert:eq}) into account.
820: We do not take higher orders into account because we
821: are mainly interested in the qualitative behaviour of the freezing line
822: for small densities $\rho_2^\mathrm{r}$.
823: Denoting by $F_0$ the Helmholtz free energy of the reference
824: hard sphere system (effective hard sphere diameter $\sigma$),
825: the total Helmholtz free energy $F$ of the one-component system
826: consisting of $N_1$ big star polymers is approximated by
827: \begin{equation}
828: \frac{\beta F}{N_1}=\frac{\beta F_0}{N_1}
829: +\frac{1}{2}\beta\rho_1\int g_0(r)V_\mathrm{pert}(r)\mathrm{d}^3r.
830: \label{f0:eq}
831: \end{equation}
832: In Eq.\ (\ref{f0:eq}) above, $g_0(r)$ denotes the radial distribution
833: function of the reference hard-sphere system in the fluid phase
834: and its angle-averaged counterpart in the solid phase, as
835: defined in Ref.\ \cite{kincaid77}.
836:
837: We note here that
838: more accurate methods for the treatment of
839: potentials with a soft cores have also been proposed \cite{ben03},
840: but are not used here since we are only interested in
841: the basic topology of the phase diagrams.
842: For the free energy of the reference hard sphere system we used
843: the equations of state of the Carnahan-Starling \cite{carnahan69}
844: and Hall \cite{hall70} for the solid and fluid phases, respectively.
845: For the calculation of pair distribution
846: functions we use the expressions of Henderson and Grundke
847: \cite{henderson75} for the fluid and
848: of Kincaid and Weis \cite{kincaid77} for the fcc-solid.
849: We only considered the fcc-solid because this crystal structure
850: appears at the fluid--solid transition in one component
851: star-polymer solutions with an arm number $f_1 = 263$ \cite{watzlawek99}.
852:
853: \subsection{Results}\label{subsec.phaseresults}
854:
855: \begin{figure*}
856: \centering
857: \includegraphics[width=7.1cm, clip=true]{fig7a.eps}
858: \includegraphics[width=7.1cm, clip=true]{fig7b.eps}
859: \includegraphics[width=7.1cm, clip=true]{fig7c.eps}
860: \includegraphics[width=7.1cm, clip=true]{fig7d.eps}
861: \caption{Phase diagrams for star polymer mixtures
862: for different parameter combinations. The big star functionality is
863: fixed at $f_1 = 263$.
864: The circles denote the phase boundaries as calculated by the hard-sphere
865: mapping including the perturbation part $V_{\rm {pert}}(r)$ and the
866: lines are a guide to the eye.
867: It can be seen that we only find a fluid--fluid demixing for $q=0.1$.
868: This binodal line is metastable with respect to the formation
869: of the fcc-solid. The kink in the phase boundaries for $q=0.1$
870: is an artifact and is caused by the method used
871: to split the effective pair potential
872: into reference and perturbation part. The symbol F stands for the fluid
873: and the symbol S for the solid regions.}
874: \label{figphases}
875: \end{figure*}
876:
877: In Fig.\ \ref{figphases} we show
878: the resulting phase diagrams for size ratios
879: $q=0.1$ and $q=0.2$ and
880: functionalities of the smaller species $f_2=16$ and $f_2=32$, as obtained
881: by the procedure described in the preceding subsection.
882: The kinks for $q=0.1$ at about $\rho_1\sigma_1^3=0.8$
883: and $\rho_2\sigma_2^3=0.1$ and
884: $\rho_1\sigma_1^3=0.6$ and $\rho_2\sigma_2^3=0.05$, respectively,
885: are an artifact of the choice (\ref{sigma:eq}) and are associated with
886: the sudden appearance of the $V_{\rm {pert}}(r)$-term in
887: Eq.\ (\ref{f0:eq}), once we cross over from the case
888: $d > d_0$ to the case $d < d_0$ (cf.\ also Fig.\ \ref{fig.diameter}).
889: Since we are primarily interested
890: in the behavior for small densities $\rho_2^\mathrm{r}$ and the
891: influence of the additives on crystallization, on the one hand,
892: and on the possible {\it existence} of a spinodal line, on the other,
893: a more sophisticated approach to the problem is at this stage not
894: necessary. The liquid--solid coexistence region
895: obtained by this approach is rather wide, due to the mapping on the
896: effective hard-sphere system. Accurate calculations of the phase
897: diagram of star polymers reveal that
898: the coexistence region is much more narrow \cite{watzlawek99}, yet
899: the shape and evolution of the freezing lines as a function of
900: $\rho_2^{\rm r}$ are not influenced by the width of the density gap
901: between the fluid and the solid phases. Finally, we note that
902: we have shifted the freezing line to higher densities by
903: an amount $\Delta\rho_1\sigma_1^3=0.06$, in order
904: to obtain the same density values for the crystallization
905: as in the accurately known one component case \cite{watzlawek99}.
906:
907: For size ratio $q=0.1$,
908: the effective potential $V_\mathrm{eff}(r)$ develops a strong attraction.
909: This leads to a broadening of the coexistence area between the solid
910: and fluid phase and eventually to a demixing binodal.
911: This binodal is found, however, to be metastable
912: with respect to the crystallization. The sharp kinks that show
913: up in Figs.\ \ref{figphases}(a) and \ref{figphases}(b) are
914: artifacts of the way in which the effective hard sphere diameter
915: $\sigma$ was determined and, in particular, of the fact that
916: the attractive perturbation part $V_{\rm pert}(r)$ of the effective
917: potential is absent in the treatment for small reservoir densities
918: below the kink and present above it. In reality, we expect the
919: phase coexistence lines to `turn around' smoothly, i.e., without
920: the aforementioned artificial kink. However, the topology of the
921: phase diagram and in particular the positive slope of the freezing
922: lines and the subsequent broadening of the coexistence region into
923: a `gas-crystal' phase separation is not expected to be affected by
924: these approximations.
925: For larger values of $q$,
926: a strong attraction
927: does not emerge, so there is no fluid-fluid demixing.
928: From Fig.\ \ref{figphases} it can be seen that
929: less stars with $f_2=32$ than with $f_2=16$ are needed to achieve
930: similar effect. This is in agreement with the
931: properties of star-polymer--colloid mixtures
932: which were investigated in \cite{dzubiella02a}.
933: The metastable binodal is closer to the stable region
934: of the phase diagram for the smaller functionality of the
935: depletant. The same trends were also observed in star-polymer--colloid
936: mixtures \cite{dzubiella02a},
937: where stable binodals were only found for small
938: depletant functionalities such as
939: $f_2 = 2$ and $f_2 = 6$.
940: For the star-polymer mixtures we consider here, the existence of a
941: binodal is less likely than in star-polymer--colloid mixtures,
942: because for star polymers the depletion force
943: has to overcome the Yukawa-type repulsion between them before an
944: effective attraction sets in.
945:
946: A striking effect is the melting of the crystal of the big stars upon
947: addition of the small component, as can be seen from the positive
948: slope of the freezing- and melting lines in Fig.\ \ref{figphases}.
949: No effective attraction between the star-polymers is needed for this
950: effect.
951: As can be seen from Fig.\ \ref{fig.diameter}, the effective hard sphere
952: diameter $\sigma$ of the system decreases with increasing depletant
953: density. Therefore, the effective packing fraction
954: $\eta_{\rm {HS}} = (\pi/6)\rho_1\sigma^3$ for the big stars gets smaller
955: with increasing $\rho_2^{\rm r}$.
956: Since this
957: quantity determines the location of the freezing transition,
958: the latter shifts to higher values of $\rho_1$.
959: The melting is therefore caused by the
960: apparent shrinkage of the stars due
961: to the depletion effects.
962: The counterintuitive behavior of melting of a crystal although the
963: overall polymer concentration is increased has its physical roots in
964: the {\it soft depletion} mechanism and the associated reduction of
965: the range of the repulsive potential between the big stars.
966:
967: In real experimental systems of star polymers,
968: usually the formation of a glass is observed
969: instead of crystallization. These glass transition lines are usually parallel
970: to the freezing lines of the phase diagram, a
971: property explicitly confirmed for
972: the glass line of one-component star polymer solutions \cite{foffi03}.
973: Therefore, we expect that the kinetic phase diagram of the mixture will
974: have a topology running parallel to that of the equilibrium one that
975: we traced in this study. We predict, therefore, that addition of small
976: stars will bring about a melting of the colloidal glass (or gel)
977: formed by the large ones. This has been shown to be the case for
978: mixtures of star polymers with linear homopolymer chains \cite{stiakakis02}.
979:
980: \section{Comparison between one- and two-component descriptions}
981: \label{compare:sec}
982:
983: \subsection{Chemical potentials}
984:
985: By calculating the partial chemical potentials
986: of the two-component system, we now map the reservoir
987: representation $(\rho_1, \rho_2^\mathrm{r})$ on the real
988: physical system $(\rho_1, \rho_2)$, so as to make contact with
989: experimental work, in which $\rho_2^{\rm r}$ has no direct relevance.
990: The densities $\rho_2^\mathrm{r}$ and $\rho_2$
991: are linked by the condition that the chemical potential of the reservoir
992: $\mu_2^\mathrm{r}$ has to be the same as the
993: partial chemical potential in the system $\mu_2$.
994:
995: The chemical potential in the reservoir
996: of density $\rho_2^{\rm r}$ can easily be calculated
997: if the fluid structure is known. The Helmholtz free energy density
998: of the small stars in the reservoir,
999: $F/V = f(\rho_2^\mathrm{r})$, can be split into an ideal and an excess part:
1000: \begin{equation}
1001: f(\rho_2^\mathrm{r})=f_\mathrm{id}(\rho_2^\mathrm{r})
1002: +f_\mathrm{ex}(\rho_2^\mathrm{r}),
1003: \end{equation}
1004: where $f_\mathrm{id}(z)=\rho\ln(z\tau^3)-z$ and $\tau$ is an arbitrary
1005: length scale.
1006: The second derivative of the excess
1007: part is connected to the structure via the relation \cite{likos95}
1008: \begin{equation}
1009: f_\mathrm{ex}^{\prime\prime}(\rho_2^\mathrm{r})
1010: =\frac{1}{\rho_2^\mathrm{r} S(k=0;\rho_2^\mathrm{r})}
1011: -\frac{1}{\rho_2^\mathrm{r}},
1012: \end{equation}
1013: where $S(k)$ is the structure factor of the small stars
1014: in the reservoir (one-component system).
1015: The free energy can then be calculated by two integrations, with
1016: the integration constants determined
1017: by the conditions $f_\mathrm{ex}^\prime(0)=0$ and
1018: $f_\mathrm{ex}(0)=0$. The chemical potential of the reservoir is given by
1019: \begin{equation}\label{chempot.eq}
1020: \mu_2^\mathrm{r}=f^\prime(\rho_2^\mathrm{r}).
1021: \end{equation}
1022:
1023: In order two calculate the partial chemical potential $\mu_2$
1024: in the two-component system representation,
1025: we employ the so-called concentration structure factor \cite{bhatia70}
1026: \begin{equation}
1027: S_\mathrm{con}(k)=x_1x_2^2S_{11}(k)+x_1^2x_2S_{22}(k)-(x_1x_2)^{3/2}S_{12}(k),
1028: \end{equation}
1029: where $x_i=\rho_i/(\rho_1+\rho_2)$. Thermodynamic properties
1030: can then be calculated by using the equation \cite{biben91a, bhatia70}
1031: \begin{equation}\label{gibbs.eq}
1032: \lim_{k\to 0}S_\mathrm{con}(k)=\left[\frac{\partial^2 g(x_2, P, T)}{\partial x_2^2}\right]^{-1},
1033: \end{equation}
1034: where $g(x_2, P, T)=G(x_2, N, P, T)/N$ is
1035: the Gibbs free energy per particle of the
1036: two-component mixture and $P$ its total pressure.
1037: Eq.\ (\ref{gibbs.eq}) can then be integrated as
1038: described in Ref.\ \cite{dzubiella02a}
1039: to yield the sought-for Gibbs free energy per particle $g(x_2, P, T)$.
1040: Once the Gibbs free energy is known,
1041: the partial chemical potentials can be calculated using the equations
1042: \begin{equation}\label{chempot1.eq}
1043: g^\prime(x_2)=\mu_2-\mu_1
1044: \end{equation}
1045: and
1046: \begin{equation}\label{chempot2.eq}
1047: g(x_2)-x_2g^\prime(x_2)=\mu_1.
1048: \end{equation}
1049: Reverting from the pair of variables $(x_2, P)$
1050: back to $(\rho_1, \rho_2)$ for the mixture
1051: and using Eqs.\ref{chempot.eq}, \ref{chempot1.eq} and \ref{chempot2.eq},
1052: the mapping
1053: $\rho_2^\mathrm{r}) \rightarrow (\rho_1, \rho_2)$
1054: can be carried out.
1055: \begin{figure}
1056: \centering
1057: \includegraphics[width=6cm, clip=true]{fig8.eps}
1058: \caption{Mapping between the system- and reservoir densities
1059: $\rho_2$ and $\rho_2^{\rm r}$ of the small stars
1060: for $f_2=32$ and $q=0.1$, and for various different values of the
1061: big star density $\rho_1$.}
1062: \label{fig.mapping}
1063: \end{figure}
1064: Representative results are shown in Fig.\ \ref{fig.mapping}.
1065: Clearly, the mapping depends parametrically
1066: on the big star density $\rho_1$ in the mixture.
1067: For $\rho_1=0$ one recovers $\rho_2=\rho_2^\mathrm{r}$.
1068: In all cases, we obtain $\rho_2<\rho_s^\mathrm{r}$
1069: because all interactions are purely repulsive.
1070: The difference between the reservoir and system densities grows with
1071: increasing $\rho_1$.
1072:
1073: \subsection{Structure}
1074:
1075: \begin{figure*}
1076: \centering
1077: \includegraphics[width=6cm, clip=true]{fig9a.eps}
1078: \includegraphics[width=6cm, clip=true]{fig9b.eps}
1079: \includegraphics[width=6cm, clip=true]{fig9c.eps}
1080: \includegraphics[width=6cm, clip=true]{fig9d.eps}
1081: \includegraphics[width=6cm, clip=true]{fig9e.eps}
1082: \includegraphics[width=6cm, clip=true]{fig9f.eps}
1083: \caption{Comparison of the fluid structure in the
1084: two- and effective one-component case.
1085: The densities are denoted in the diagrams.
1086: The structure of the mixture
1087: was calculated using the two-component RY closure,
1088: in the one component the usual RY closure was used.
1089: All plots are for $q=0.1$ and $f_2=32$. In panels (a), (e) and (f)
1090: the solid and dashed lines fall on top of each other, so that the
1091: latter cannot be distinguished from the former.}
1092: \label{figcompare}
1093: \end{figure*}
1094:
1095: We now consider the spatial correlations between the
1096: big stars in the fluid phase. We compare the
1097: correlation functions calculated in the two-component
1098: description with those arising from
1099: the one-component description using the effective
1100: interactions in the presence of the smaller species.
1101: The comparison has to be carried out for parameter
1102: combinations such that the reservoir- and system
1103: partial chemical potentials of the smaller species
1104: are equal to one another.
1105: As the one-component description reduces the effective
1106: interaction to pair potentials, the comparison of
1107: the structural properties allows to estimate the magnitude of
1108: many body effects on the depletion.
1109:
1110: In
1111: Fig.\ \ref{figcompare} we show the radial distribution
1112: functions and static structure factors of the big stars
1113: calculated with both different approaches.
1114: For the one-component case, the RY closure can be used
1115: because for the densities we consider here
1116: the effective interactions remain purely repulsive.
1117: The cases we consider here
1118: are for small densities $\rho_2$ but a wide range of values of $\rho_1$.
1119: The two component description includes many-body forces between the
1120: big stars which are caused by the smaller component.
1121: These effects are neglected
1122: in the effective one-component description \cite{md3}.
1123: In Fig.\ \ref{figcompare}, one can see excellent agreement
1124: of the results obtained by using these two
1125: different descriptions of the physical system. This additionally
1126: corroborates the
1127: validity of our approximation of the effective interaction. It can be
1128: also be concluded from the plots that the three-body forces
1129: are indeed much weaker than the pair interactions and can be
1130: safely neglected \cite{melchionna00, goulding01}.
1131:
1132: \section{Summary and conclusions}
1133: \label{concl:sec}
1134:
1135: We have analyzed the structural and phase behavior of highly asymmetric
1136: mixtures of star polymers, with the asymmetry characterizing both their
1137: sizes and functionalities. The most striking phenomenon predicted
1138: by our investigations is the counter-intuitive {\it melting} of the
1139: colloidal crystal of the big stars upon addition of small ones. Though
1140: this finding appears paradoxical at first sight, since addition of
1141: smaller stars increases the overall polymer concentration of the
1142: solution, its physical explanation can be traced to the effects of
1143: {\it soft depletion}. Whereas the depletants of hard spheres simply
1144: superimpose an effective attraction on a hard potential, when the
1145: big particles are themselves repulsive the depletion attraction
1146: is superimposed on a (soft) repulsion. In this way, a repulsive
1147: potential of reduced strength and/or range results and the effective,
1148: reduced repulsion is not any more sufficient to maintain the stability
1149: of the crystal, which therefore melts. When the depletant concentration
1150: becomes sufficiently high, the attractive depletion force dominates
1151: over the soft repulsion, leading to a (possibly metastable) demixing
1152: transition between the two species.
1153:
1154: In real experimental systems of star polymers with high functionality,
1155: crystallization is hindered by the vitrification (gelation) transition
1156: and the large stars become structurally arrested in a glassy state
1157: above the overlap concentration. The next step would be then to
1158: investigate the role and influence of smaller star additives on
1159: the glass transition of larger stars, a problem of significant
1160: importance for the control of the rheological behavior of soft matter
1161: through additives. Work along these lines is currently under way
1162: and the presentation of the results of this investigation will be
1163: the subject of a future publication.
1164:
1165: \begin{acknowledgments}
1166: We thank Dimitris Vlassopoulos, Joachim Dzubiella, Martin Konieczny,
1167: Emanuela Zaccarelli, Francesco Sciortino, and
1168: Piero Tartaglia for helpful discussions. This work has been supported
1169: by the Marie Curie European Network MRTN-CT-2003-504712 and by the DFG within SFB TR 6.
1170: \end{acknowledgments}
1171:
1172: \begin{thebibliography}{99}
1173:
1174: \bibitem{poon:review} See, e.g., W. C. K. Poon, J. Phys.: Condens. Matter
1175: {\bf 14}, R859 (2002) and references therein.
1176:
1177: \bibitem{ao:54} S. Asakura and F. Oosawa, J. Chem. Phys. {\bf 22}, 1255
1178: (1954).
1179:
1180: \bibitem{ao:56} S. Asakura and F. Oosawa, J. Polymer Sci. {\bf 33}, 183 (1956).
1181:
1182: \bibitem{vrij:76} A. Vrij, Pure Appl. Chem. {\bf 48}, 471 (1976).
1183:
1184: \bibitem{biben91a}
1185: T. Biben and J.-P. Hansen, Phys. Rev. Lett. \textbf{66}, 2215 (1991).
1186:
1187: \bibitem{biben91}
1188: T. Biben and J.-P. Hansen, J. Phys.: Condens. Matter \textbf{3}, F65 (1991).
1189:
1190: \bibitem{md1} M. Dijkstra, R. van Roij, and R. Evans,
1191: Phys. Rev. Lett. {\bf 81}, 2268 (1998).
1192:
1193: \bibitem{md2} M. Dijkstra, R. van Roij, and R. Evans,
1194: Phys. Rev. Lett. {\bf 82}, 117 (1999).
1195:
1196: \bibitem{md3} M. Dijkstra, R. van Roij, and R. Evans,
1197: Phys. Rev. E {\bf 59}, 5744 (1999).
1198:
1199: \bibitem{fuchs:review} See, e.g., M. Fuchs and K. S. Schweizer,
1200: J. Phys.: Condens. Matter {\bf 14}, R239 (2002) and references therein.
1201:
1202: \bibitem{ilett:95} S. M. Ilett, A. Orrock, W. C. K. Poon, and
1203: P. N. Pusey, Phys. Rev. E {\bf 51}, 1344 (1995).
1204:
1205: \bibitem{louis:99} A. A. Louis, R. Finken, and J.-P. Hansen,
1206: Europhys. Lett. {\bf 46}, 741 (1999).
1207:
1208: \bibitem{dijkstra:99} M. Dijkstra, J. M. Brader, and R. Evans,
1209: J. Phys.: Condens. Matter {\bf 11}, 10079 (1999).
1210:
1211: \bibitem{matthias:prl:00} M. Schmidt, H. L{\"o}wen, J. M. Brader, and
1212: R. Evans, Phys. Rev. Lett. {\bf 85}, 1934 (2000).
1213:
1214: \bibitem{fuchs:00} M. Fuchs and K. S. Schweizer, Europhys. Lett. {\bf 51},
1215: 621 (2000).
1216:
1217: \bibitem{matthias:jpcm:02} M. Schmidt, H. L{\"o}wen, J. M. Brader, and
1218: R. Evans, J. Phys.: Condens. Matter {\bf 14}, 9353 (2002).
1219:
1220: \bibitem{matthias:jcp:02} M. Schmidt and M. Fuchs, J. Chem. Phys. {\bf 117},
1221: 6308 (2002).
1222:
1223: \bibitem{louis:prl:02} P. G. Bolhuis, A. A. Louis, and J.-P. Hansen,
1224: Phys. Rev. Lett. {\bf 89}, 128302 (2002).
1225:
1226: \bibitem{evans:04} For a recent review on inhomogeneous colloid-polymer
1227: mixtures, see J. M. Brader, R. Evans, and M. Schmidt, Mol. Phys. {\bf 101},
1228: 3349 (2003).
1229:
1230: \bibitem{brader:00} J. M. Brader and R. Evans, Europhys. Lett. {\bf 49},
1231: 678 (2000).
1232:
1233: \bibitem{brader:01} J. M. Brader, M. Dijkstra, and R. Evans,
1234: Phys. Rev. E {\bf 63}, 041405 (2001).
1235:
1236: \bibitem{brader:02} J. M. Brader, R. Evans, M. Schmidt, and H. L{\"o}wen,
1237: J. Phys.: Condens. Matter {\bf 14}, L1 (2002).
1238:
1239: \bibitem{dijkstra:02} M. Dijkstra and R. van Roij, Phys. Rev. Lett. {\bf 89},
1240: 208303 (2002).
1241:
1242: \bibitem{wijting:03} W. K. Wijting, N. A. M. Besseling, and M. A. Cohen Stuart,
1243: Phys. Rev. Lett. {\bf 90}, 196101 (2003).
1244:
1245: \bibitem{moncho:03} A. Moncho-Jorda, B. Rothenberg, and A. A. Louis,
1246: J. Chem. Phys. {\bf 119}, 12667 (2003).
1247:
1248: \bibitem{wessels:04} P. P. F. Wessels, M. Schmidt, and H. L{\"o}wen,
1249: J. Phys.: Condens. Matter {\bf 16}, L1 (2004).
1250:
1251: \bibitem{bergenholtz:99} J. Bergenholtz and M. Fuchs, Phys. Rev. E
1252: {\bf 59}, 5706 (1999).
1253:
1254: \bibitem{johner:01} A. Johner, J.-F. Joanny, S. D. Orrite, and
1255: J. B. Avalos, Europhys. Lett. {\bf 56}, 549 (2001).
1256:
1257: \bibitem{foffi:02} G. Foffi, K. A. Dawson, S. V. Buldyrev, F. Sciortino,
1258: E. Zaccarelli, and P. Tartaglia, Phys. Rev. E {\bf 65}, 050802 (2002).
1259:
1260: \bibitem{puertas:03} A. M. Puertas, M. Fuchs, and M. E. Cates,
1261: Phys. Rev. E {\bf 67}, 031406 (2003).
1262:
1263: \bibitem{bergenholtz:03} J. Bergenholtz, W. C. K. Poon, and M. Fuchs,
1264: Langmuir {\bf 19}, 4493 (2003).
1265:
1266: \bibitem{pham:04} K. N. Pham, S. U. Egelhaaf, P. N. Pusey, and
1267: W. C. K. Poon, Phys. Rev. E {\bf 69}, 011503 (2004).
1268:
1269: \bibitem{zacca:04} E. Zaccarelli, H. L{\"o}wen, P. P. F. Wessels,
1270: F. Sciortino, P. Tartaglia, and C. N. Likos, Phys. Rev. Lett.,
1271: \textbf{92}, 225703 (2004).
1272:
1273: \bibitem{likos01}
1274: C. N. Likos, Phys. Rep. \textbf{348}, 261 (2001).
1275:
1276: \bibitem{louis:jpcm:01} A. A. Louis and R. Roth, J. Phys.: Condens. Matter
1277: {\bf 13}, L777 (2001).
1278:
1279: \bibitem{goetz:99} B. G{\"o}tzelmann, R. Roth, S. Dietrich, M. Dijkstra,
1280: and R. Evans, Europhys. Lett. {\bf 47}, 398 (1999).
1281:
1282: \bibitem{roth:00} R. Roth, R. Evans, and S. Dietrich, Phys. Rev. E
1283: {\bf 62}, 5360 (2000).
1284:
1285: \bibitem{dzubiella02}
1286: J. Dzubiella, C. N. Likos, and H. L\"owen,
1287: Europhys. Lett. \textbf{58}, 133 (2002).
1288:
1289: \bibitem{dzubiella02a}
1290: J. Dzubiella, C. N. Likos, and H. L\"owen, J. Chem. Phys. \textbf{116},
1291: 9518 (2002).
1292:
1293: \bibitem{joe:pre} J. Dzubiella, A. Jusufi, C. N. Likos, C. von Ferber,
1294: H. L{\"o}wen, J. Stellbrink, J. Allgaier, D. Richter, A. B. Schofield,
1295: P. A. Smith, W. C. K. Poon, and P. N. Pusey, Phys. Rev. E {\bf 64},
1296: 010401(R) (2001).
1297:
1298: \bibitem{roth:02} A. A. Louis, E. Allahyarov, H. L{\"o}wen, and R. Roth,
1299: Phys. Rev. E {\bf 65}, 061407 (2002).
1300:
1301: \bibitem{stiakakis02}
1302: E. Stiakakis, D. Vlassopoulos, C. N. Likos, J. Roovers, and G. Meier,
1303: Phys. Rev. Lett. \textbf{89}, 208302 (2002).
1304:
1305: \bibitem{witten86}
1306: T. A. Witten and P. A. Pincus, Macromolecules \textbf{19}, 2509 (1986).
1307:
1308: \bibitem{likos98}
1309: C. N. Likos, H. L\"owen, M. Watzlawek, B. Abbas, O. Jucknischke,
1310: J. Allgaier, and D. Richter, Phys. Rev. Lett. \textbf{80}, 4450 (1998).
1311:
1312: \bibitem{jusufi99}
1313: A. Jusufi, M. Watzlawek, and H. L\"owen, Macromolecules \textbf{32},
1314: 4470 (1999).
1315:
1316: \bibitem{watzlawek99}
1317: M. Watzlawek, C. N. Likos and H. L\"owen, Phys. Rev. Lett. \textbf{82},
1318: 5289 (1999).
1319:
1320: \bibitem{vonferber00}
1321: C. von Ferber, A. Jusufi, M. Watzlawek, C. N. Likos, and H. L\"owen,
1322: Phys. Rev. E \textbf{62}, 6949 (2000).
1323:
1324: \bibitem{lebowitz64}
1325: J. L. Lebowitz and J. S. Rowlinson, J. Chem. Phys. \textbf{41}, 133 (1964).
1326:
1327: \bibitem{watzlawek98}
1328: M. Watzlawek, H. L\"owen, and C. N. Likos, J. Phys.: Condens. Matter \textbf{10}, 8189 (1998).
1329:
1330: \bibitem{foffi03}
1331: G. Foffi, F. Sciortino, P. Tartaglia, E. Zaccarelli, F. Lo Verso,
1332: L. Reatto, K. A. Dawson, and C. N. Likos, Phys. Rev. Lett. \textbf{90},
1333: 238301 (2003).
1334:
1335: \bibitem{hansen69}
1336: J.-P. Hansen and L. Verlet, Phys. Rev. \textbf{184}, 151 (1969).
1337:
1338: \bibitem{mendez00}
1339: J. M. M\'endez-Alcaraz and R. Klein, Phys. Rev. E \textbf{61}, 4095 (2000).
1340:
1341: \bibitem{konig01}
1342: A. K\"onig and N. W. Ashcroft, Phys. Rev. E \textbf{63}, 041203 (2001).
1343:
1344: \bibitem{damico97}
1345: I. D'Amico and H. L\"owen, Physica A \textbf{235}, 25 (1997).
1346:
1347: \bibitem{allahyarov98}
1348: E. Allahyarov, I. D'Amico, and H. L\"owen, Phys. Rev. Lett. \textbf{81},
1349: 1334 (1998).
1350:
1351: \bibitem{allahyarov01}
1352: E. Allahyarov and H. L\"owen, J. Phys.: Condens. Matter \textbf{13},
1353: L277 (2001).
1354:
1355: \bibitem{hansen86}
1356: J.-P. Hansen and I. R. McDonald, \textit{Theory of Simple Liquids},
1357: 2nd ed.\ (Academic, London, 1986).
1358:
1359: \bibitem{attard89}
1360: P. Attard, J. Chem. Phys. \textbf{91}, 3083 (1989).
1361:
1362: \bibitem{loverso:03} F. Lo Verso, M. Tau, and L. Reatto,
1363: J. Phys.: Condens. Matter {\bf 15}, 1505 (2003).
1364:
1365: \bibitem{barker67}
1366: J. A. Barker and D. Henderson, J. Chem. Phys. \textbf{47}, 4714 (1967).
1367:
1368: \bibitem{ben03}
1369: D. Ben-Amotz and G. Stell, J. Chem. Phys. \textbf{119}, 10777 (2003).
1370:
1371: \bibitem{kincaid77}
1372: J. M. Kincaid and J. J. Weis, Mol. Phys. \textbf{34}, 931 (1977).
1373:
1374: \bibitem{carnahan69}
1375: N. F. Carnahan and K. E. Starling, J. Chem. Phys. \textbf{51}, 635 (1969).
1376:
1377: \bibitem{hall70}
1378: K. R. Hall, J. Chem. Phys. \textbf{57}, 2252 (1970).
1379:
1380: \bibitem{henderson75}
1381: D. Henderson and E. W. Grundke, J. Chem. Phys. \textbf{63}, 601 (1975).
1382:
1383: \bibitem{likos95}
1384: C. N. Likos and G. Senatore, J. Phys.: Condens. Matter \textbf{7}, 6797 (1995).
1385:
1386: \bibitem{bhatia70}
1387: A. B. Bhatia and D. Thornton, Phys. Rev. B \textbf{2}, 3004 (1970).
1388:
1389: \bibitem{melchionna00}
1390: S. Melchionna and J.-P. Hansen, Phys. Chem. Chem. Phys. \textbf{2},
1391: 3465 (2000).
1392:
1393: \bibitem{goulding01}
1394: D. Goulding and S. Melchionna, Phys. Rev. E \textbf{64}, 011403 (2001).
1395:
1396: \end{thebibliography}
1397:
1398: \end{document}
1399: