physics0104025/ms.tex
1: %% ****** Start of file template.aps ****** %
2: %%
3: %%
4: %%   This file is part of the APS files in the REVTeX 4 distribution.
5: %%   Version 4.0 beta 4 of REVTeX, May 24, 2000.
6: %%
7: %%
8: %%   Copyright (c) 2000 The American Physical Society.
9: %%
10: %%   See the REVTeX 4 README file for restrictions and more information.
11: %%
12: %
13: % This is a template for producing files for use with REVTEX 4.0 beta
14: % Copy this file to another name and then work on that file.
15: % That way, you always have this original template file to use.
16: %
17: % Group addresses by affiliation; use superscriptaddress for long
18: % author lists, or if there are many overlapping affiliations.
19: % For Phys. Rev. appearance, change preprint to twocolumn.
20: %  Add 'draft' option to mark overfull boxes with black boxes
21: %  Add 'showpacs' option to make PACS codes appear
22: %\documentclass[aps,preprint,groupedaddress]{revtex4}
23: %\documentclass[aps,preprint,superscriptaddress]{revtex4}
24: 
25: 
26: \documentclass[aps,twocolumn,groupedaddress]{revtex4}
27: %\setlength{\topmargin}{0.2in}
28: 
29: \usepackage[oztex]{graphicx}
30: \begin{document}
31: % You should use BibTeX and apsrev.bst for references
32: %\bibliographystyle{jpc}
33: 
34: % Use the \preprint command to place your local institutional report
35: % number on the title page in preprint mode.
36: % Multiple \preprint commands are allowed.
37: \preprint{LA-UR-01-1530}
38: 
39: %Title of paper
40: \title{Quasi-chemical
41: Theory for the Statistical Thermodynamics of the Hard Sphere Fluid}
42: % Optional argument for running titles on pages
43: %\title[]{}
44: 
45: % repeat the \author .. \affiliation  etc. as needed
46: % \email, \thanks, \homepage, \altaffiliation all apply to the current
47: % author. Explanatory text should go in the []'s, actual e-mail
48: % address or url should go in the {}'s for \email and \homepage.
49: % Please use the appropriate macro for the type of information
50: 
51: % \affiliation command applies to all authors since the last
52: % \affiliation command. The \affiliation command should follow the
53: % other information
54: 
55: \author{Lawrence  R.  Pratt}
56: %\email[]{Your e-mail address}
57: %\homepage[]{Your web page}
58: %\thanks{}
59: %\altaffiliation{}
60: \affiliation{Theoretical Division, Los Alamos National Laboratory, 
61: Los Alamos, NM 87545, USA}
62: 
63: \author{Randall A. LaViolette}
64: %\email[]{Your e-mail address}
65: %\homepage[]{Your web page}
66: %\thanks{}
67: %\altaffiliation{}
68: \affiliation{Idaho National Engineering and Environmental Laboratory, 
69: PO Box 1625, Idaho Falls, ID 83415-2208, USA}
70: 
71: \author{Maria A. Gomez}
72: %\email[]{Your e-mail address}
73: %\homepage[]{Your web page}
74: %\thanks{}
75: %\altaffiliation{}
76: \affiliation{Department of Chemistry, Vassar College, Poughkeepsie, NY
77: 12603, USA}
78: 
79: \author{Mary E. Gentile}
80: %\email[]{Your e-mail address}
81: %\homepage[]{Your web page}
82: %\thanks{}
83: %\altaffiliation{}
84: \affiliation{Department of Chemistry, Vassar College, Poughkeepsie, NY
85: 12603, USA}
86: 
87: 
88: 
89: 
90: %Collaboration name if desired (requires use of superscriptaddress
91: %option in \documentclass). \noaffiliation is required (may also be
92: %used with the \author command).
93: %\collaboration{}
94: %\noaffiliation
95: 
96: \date{\today}
97: 
98: \begin{abstract}
99: % insert abstract here
100: We develop a quasi-chemical theory for the study of packing
101: thermodynamics in dense liquids.  The situation of hard-core
102: interactions is addressed by  considering the binding of solvent
103: molecules to a precisely defined `cavity' in order to assess the
104: probability that the `cavity' is entirely evacuated. The primitive
105: quasi-chemical approximation corresponds to a extension of the Poisson
106: distribution used as a default model in an information theory approach.
107: This primitive quasi-chemical theory is in good qualitative agreement
108: with the observations for the hard sphere fluid of  occupancy
109: distributions that are central to quasi-chemical theories but begins to
110: be quantitatively erroneous for the equation of state in the dense
111: liquid regime of $\rho d^3\!>$0.6. How the quasi-chemical approach can be
112: iterated to treat correlation effects is addressed.  Consideration of
113: neglected correlation effects leads to a simple model for the form of
114: those contributions neglected by the primitive quasi-chemical
115: approximation. These considerations, supported by simulation
116: observations,  identify a `break away' phenomena that requires special
117: thermodynamic consideration for the zero (0) occupancy case as distinct
118: from the rest of the distribution.  A empirical treatment leads to a one
119: parameter model occupancy distribution that accurately  fits the hard
120: sphere equation of state and observed distributions.
121: \end{abstract}
122: % insert suggested PACS numbers in braces on next line
123: \pacs{}
124: 
125: %\maketitle must follow title, authors, abstract and \pacs
126: \maketitle
127: 
128: % body of paper here - Use proper section commands
129: % References should be done using the \cite, \ref, and \label commands
130: %\label{}
131: 
132: \section{Introduction}
133: The quasi-chemical theory\cite{Pratt:MP:98,Pratt:ES:99,Hummer:CPR:2000}
134: is a fresh attack on the molecular statistical thermodynamic theory of
135: liquids.  It is intended to be specifically appropriate in describing
136: liquids of genuinely chemical interest.  But, in view of its generality,
137: the quasi-chemical theory must be developed and tested for its
138: description of the paradigmatic hard sphere fluid.  In addition to 
139: the conceptual point, these developments are expected to be helpful in
140: subsequent applications of the quasi-chemical theory to real solutions.
141: 
142: 
143: The foundational virtues of the hard sphere fluid for the theory of
144: liquids are widely recognized\cite{WCA} and the interest in this system
145: continues to evolve\cite{Altenberger:96,Robles:98,Yelash:99,Parisi:00}. Recent
146: developments of the theory of hydrophobic
147: effects\cite{Pohorille:JACS:90,Pratt:PNAS:92,Palma,Hummer:PNAS:96,%
148: Garde:PRL:96,Pratt:ECC,Hummer:PNAS:98,Hummer:JPCB:98,Pohorille:PJC:98,%
149: Pratt:NATO:99,Gomez:99,Garde:99,Hummer:CPR:2000}, in addition the
150: related quasi-chemical theory, have emphasized again the significance of
151: packing issues in a realistic molecular description of complex liquids.
152: This paper studies the hard sphere fluid and develops default
153: models with utility in recent information theory
154: approaches\cite{Hummer:PNAS:96,%
155: Garde:PRL:96,Pratt:ECC,Hummer:PNAS:98,Hummer:JPCB:98,Pohorille:PJC:98,%
156: Pratt:NATO:99,Gomez:99,Garde:99,Hummer:CPR:2000,Crooks:PRE:97}.  
157: 
158: A compact derivation requires several preliminary results, including
159: brief specifications of the potential distribution theorem,  of the
160: expression of chemical equilibrium, and of the quasi-chemical
161: formulation. Additionally, the notation here is not elsewhere
162: standardized because these ideas are unconventional.  The plan of the
163: paper is thus to collect the necessary preliminary results in
164: Appendix~\ref{aprelim} so that the conceptual argument needn't be
165: interrupted. Then we derive the new equation of state format, learn what
166: we can by comparison of the primitive quasi-chemical
167: approximation with Monte Carlo simulation results, study correlation
168: contributions to propose an improved equation of state format, and
169: finally examine how this improved format works.
170: 
171: Interestingly, though the some of these basic considerations are
172: regarded as `preliminary,' Eqs.~\ref{hsqca} or \ref{qcarule}, and the
173: formal identification of the equibrium ratio Eq.~\ref{K},  have been
174: given before and have wide generality.
175: 
176: 
177: \begin{figure}
178: \begin{center}
179: \leavevmode
180: \includegraphics[scale=0.7]{knobs.01.eps}
181: \end{center}
182: \caption{An example of an AS$_7$ cluster considered in the text.  The
183: nucleus (A) is visible in the center.  Each of the ligands (S) overlaps
184: the nucleus but no other ligand.}
185: \label{knobs}
186: \end{figure}
187: 
188: 
189: \section{A Quasi-chemical View of the Solvation Free Energy of
190: Hard Core Solutes}
191: The preliminary results of Appendix~\ref{aprelim} permit an attack
192: on the solvation thermodynamics of hard core species built upon a simple
193: device.   Let's consider a solute A that doesn't interact with the
194: solvent S molecules at all.  We will consider formation of AS$_n$
195: complexes and Fig.~\ref{knobs} depicts such a cluster.  The interaction
196: contribution $\beta\Delta\mu_\mathrm{A}$ is zero and the quasi-chemical
197: Eq.~\ref{gqca} expresses
198: \begin{eqnarray}
199: \ln\left[
200: \left\langle \left\langle \prod\limits_j {(1-b_{\mathrm{A} j})}
201: \right\rangle\right\rangle_0  \right] = \ln x_0~.
202: \label{trick} \end{eqnarray}
203: But the left side here is a test particle average for solute that
204: rigidly excludes solvent molecules from the region defined by the
205: indicator function $b_{\mathrm{A} j}$. If the region is taken as
206: defining a physically interesting molecule pair excluded volume, then
207: the right side of Eq.~\ref{trick} gives the negative of the excess
208: chemical potential for the hard core solute defined by $b_{\mathrm{A}
209: j}$. This is an example of the well known relation for hard core solutes
210: $\ln p_0=-\beta\Delta\mu_\mathrm{HC}$ with `HC' denoting `hard
211: core'\cite{Pohorille:JACS:90,Pratt:PNAS:92,Palma,Hummer:PNAS:96,%
212: Garde:PRL:96,Pratt:ECC,Hummer:PNAS:98,Hummer:JPCB:98,Pohorille:PJC:98,%
213: Pratt:NATO:99,Gomez:99,Garde:99,Hummer:CPR:2000}.  This observation
214: sheds light on the compensation of inner and outer sphere contributions
215: to the quasi-chemical Eq.~\ref{gqca} but is not surprising. We then
216: consider `chemical' equilibria for binding of S molecules to the A
217: molecule. Of course, there is no interaction between the A
218: molecule and the solvent molecules. The binding is just the occupancy by
219: a solvent molecules of the `cavity' defined by $b_{\mathrm{A}
220: j}$. Combining these considerations gives
221: \begin{eqnarray}
222: \beta\Delta\mu_\mathrm{HC}  =\ln\left[{1 +  \sum\limits_{m\ge 1} K_m{}
223: \rho_\mathrm{S}{}^m }\right]~.
224: \label{hsqca}
225: \end{eqnarray}
226: The K$_m$ are well-defined but typically computationally demanding; see
227: Eq.~\ref{K}. The evaluation of K$_m{}^{(0)}$ will require few-body
228: integrals over excluded volumes as is discussed in Appendix~\ref{aa}. 
229: The primitive quasi-chemical approximation is
230: \begin{eqnarray}
231: \beta\Delta\mu_\mathrm{HC} \approx \ln\left[1 +  \sum\limits_{m\ge 1} K_m{}^{(0)}
232: \rho_\mathrm{S}{}^m\lambda{}^m \right]
233: \label{pqca}
234: \end{eqnarray}
235: with $\lambda$ a `mean field' factor that achieves the self-consistency
236: condition $\sum_n n K_n^{(0)}\rho_\mathrm{S}{}^n\lambda^n =
237: \rho_\mathrm{S} K_1^{(0)}\sum_n K_n^{(0)}\rho_\mathrm{S}{}^n\lambda^n$. 
238: This amounts to an extension of the Poisson distribution  for use  in
239: an information theory procedure\cite{Pratt:NATO:99}.  Here
240: $\rho_\mathrm{S}$K$_1^{(0)}$=$<$n$>$ is the expected occupancy of the
241: volume stenciled by $b_{\mathrm{A} j}$.  Thus the multiplicative factors
242: of $\rho_\mathrm{S}$ in $x_n^{(0)}\propto K_{n}{}^{(0)}\rho
243: _\mathrm{S}{}^n$ are augmented by a self-consistent `mean field'
244: $\lambda$\footnote{This point has a twist for the non-profound one
245: dimensional problem:  This primitive quasi-chemical approximation is
246: exact for the one dimensional `hard plate' system.  But as a distribution
247: without evaluation of the mean field factors, this distribution is an
248: exceedingly weak theory.  Evaluation of the mean field
249: produces the exact answer because the number of statistical
250: possibilities is only 2.  The situation for the continuum analog, the
251: Poisson distribution, is different.  It is not accurate as a
252: distribution and the same information theory interpretation still gives
253: an incorrect result for the one dimensional hard plate system.}.
254: 
255: 
256: For reuse below, we summarize the technical results of this argument for
257: hard core solutes, writing
258: \begin{eqnarray}
259: \left\langle {\left\langle {e^{-\beta \Delta U_\mathrm{HC} }} \right\rangle }
260: \right\rangle _0 = {1\over 1 +  \sum\limits_{m\ge 1} K_m{}
261: \rho_\mathrm{S}{}^m }~. \label{qcarule}
262: \end{eqnarray}
263: This combines Eq.~\ref{hsqca} and the potential distribution theorem Eq.~\ref{mu} for
264: this problem.
265: 
266: \section{Primitive Quasi-chemical Approximation}
267: 
268: We can give a simple demonstration of the quantitative results of the
269: primitive quasi-chemical theory by considering the hard disk (2d) and
270: hard sphere fluids (3d). Table~\ref{tab:kn} gives Monte Carlo estimates
271: of the K$_n{}^{(0)}$ for those cases. The predicted distributions x$_n$
272:  for two densities are in Figs.~\ref{compare:fig} and
273: \ref{compare.04:fig}.  Equation of state results
274: $\beta\Delta\mu(\rho)$ for these systems predicted by this primitive
275: quasi-chemical theory are shown in Figs.~\ref{2dHD:fig} and
276: \ref{3dHS:fig}.  The primitive quasi-chemical approximation is
277: remarkably successfully in all qualitative respects, particularly in
278: view of its simplicity.  In particular, the predicted occupancy
279: distributions such as shown in Fig.~ \ref{compare.04:fig} are remarkably
280: faithful to the data. Nevertheless, the equation of state predictions
281: begin progressively to incur serious quantitative errors at liquid
282: densities $\rho d^3\!>$ 0.6, (Fig.~\ref{3dHS:fig}).
283: 
284: \begin{table}
285: \caption{`Hit-or-Miss' Monte Carlo estimates \cite{HH}, as 
286: described in Appendix \ref{aa}, of $\ln$K$_n{}^{(0)}$ for unit diameter
287: hard spheres and disks, respectively.  K$_1{}^{(0)}$=${ 4\pi\over 3}$ ($\pi$) and
288: K$_2{}^{(0)}$=${17\pi^2\over 36}$ (${3 \sqrt{3}\pi \over 8}$).  The sample size was 24
289: G-configurations and the results are believed to be accurate to the number
290: of significant figures given. }
291: \label{tab:kn}
292: \begin{tabular}{|c|c|c|}
293: \hline
294: n &  spheres (3d) & disks (2d) \\ 
295: \hline
296: 1 & 1.43241 & 1.14473 \\
297: 2 & 1.53915 & 0.71321 \\
298: 3 & 0.56585 & -1.190 \\
299: 4 & -1.4697 & -5.241 \\
300: 5 & -4.684  & -13.77\\
301: 6 & -9.168  & - \\
302: 7 & -15.46  & - \\
303: \hline
304: \end{tabular}
305: \end{table}
306: 
307: 
308: 
309: \begin{figure}
310: \begin{center}
311: \leavevmode
312: \includegraphics[scale=0.8]{compare.05.eps}
313: \end{center}
314: \caption{For the hard sphere fluid  at $\rho_\mathrm{S}d^3$=0.277,
315: comparison for n$\le$5 of the Poisson distribution (solid curve) with
316: primitive quasi-chemical distribution (dashed curve) implemented with
317: the information theory constraint on the first moment $\sum n
318: x_n=4\pi\rho_S d^3/3$.  The dots are the results of Monte Carlo
319: simulation\cite{Gomez:99} as discussed in Appendix~\ref{amc}.  The
320: primitive quasi-chemical default model depletes the probability of
321: high-n and low-n constellations and enhances the probability near the
322: mode.}
323: \label{compare:fig}
324: \end{figure}
325: 
326: 
327: \begin{figure}
328: \begin{center}
329: \leavevmode
330: \includegraphics[scale=0.8]{compare.04.eps}
331: \end{center}
332: \caption{As in Fig.~\ref{compare:fig} but for $\rho_\mathrm{S}d^3$=0.8.
333: The error bars indicate the statistical uncertainty by showing the 67\%
334: confidence interval.}
335: \label{compare.04:fig}
336: \end{figure}
337: 
338: 
339: \begin{figure}
340: \begin{center}
341: \leavevmode
342: \includegraphics[scale=0.8]{2dHD.eps}
343: \end{center}
344: \caption{$\beta\Delta\mu(\rho)$ for the two dimensional hard disk
345: fluid on the basis of the primitive quasi-chemical approximation (dashed lined). The 
346: Ree-Hoover 3,3 Pad\'e approximant\cite{RH:64} is the solid line and
347: the dash-dot line is the first virial coefficient approximation.}
348: \label{2dHD:fig}
349: \end{figure}
350: 
351: \begin{figure}
352: \begin{center}
353: \leavevmode
354: \includegraphics[scale=0.8]{3dHS.eps}
355: \end{center}
356: \caption{$\beta\Delta\mu(\rho)$ for the three dimensional hard sphere
357: fluid on the basis of the primitive quasi-chemical approximation (dashed
358: line).  The solid line is the prediction of the Carnahan-Starling
359: equation of state, taken as the accurate basis for comparison, and
360: the dash-dot line is the first virial coefficient approximation.}
361: \label{3dHS:fig}
362: \end{figure}
363: 
364: 
365: 
366: 
367: 
368: \section{Test of the Equilibrium Ratios} 
369: As a direct check on the primitive quasi-chemical mechanism, we can
370: focus on testing ideal populations Eq.~\ref{cluster-var.0} as
371: approximations to formally correct populations Eq.~\ref{cluster-var}. 
372: It is then natural to consider the ratios $x_j/x_0={K_{n}\rho _\mathrm{S}{}^n}$.  Consideration of
373: these ratios corresponds to shifting the curves of Figs.~\ref{compare:fig}
374: and \ref{compare.04:fig} so that the initial point is at the common
375: value (0,1).   A specific example is shown in Fig.~\ref{approx.04:fig}.
376: Compared with this normalization, it is clear that the observed
377: equilibrium ratios K$_n$ are greater than the ideal ratios
378: K$_n{}^{(0)}$.
379: 
380: \begin{figure}
381: \begin{center}
382: \leavevmode
383: \includegraphics[scale=0.8]{approx.04.eps}
384: \end{center}
385: \caption{$\ln\left[{(x_n/x_0)_{qca} \over (x_n/x_0)_{sim}}\right]$ {\it
386: vs.} n comparing for the hard sphere fluid the primitive quasi-chemical
387: approximate populations with those observed by Monte Carlo simulation
388: for $\rho d^3$=0.8. This normalization focuses on the relative sizes of
389: K$_n$ and K$_n{}^{(0)}$ suggesting that K$_n\!>$K$_n{}^{(0)}$ even after the
390: maxent reweighting. The variations are modest {\it except} for the
391: change between n=0 and n=1.    The error bars indicate the statistical
392: uncertainty by showing the 67\% confidence interval. In the middle of
393: the distribution the statistical uncertainty is mostly from the
394: uncertainty in the denominator factor x$_0$.}
395: \label{approx.04:fig}
396: \end{figure}
397: 
398: \section{Correlations} 
399: 
400: A point of view here is that the geometric weighting   with the
401: $\lambda$'s of Eq.~\ref{pqca} establishes a mean field that adapts to the
402: prescribed density.  We now consider how to go beyond that mean field
403: description. One idea is to extract the features of the summand of
404: Eq.~\ref{qcarule} that would give purely geometric weighting and then to
405: analyze what remains.   To this end, we define $\zeta =
406: \exp(\beta\Delta\mu)$ and consult the formal identification of the
407: equilibrium ratios Eqs.~\ref{K}.  Thus we can rewrite Eq.~\ref{qcarule}
408: as
409: \begin{eqnarray}
410: \zeta = 1 + \sum_{m\ge 1} x_{0/m}K_m{}^{(0)}\zeta^m \rho_\mathrm{S}{}^m~,
411: \label{f-exact}
412: \end{eqnarray}
413: with 
414: \begin{eqnarray}
415:  x_{0/n} \equiv{ \left\langle {\left\langle e^{-\beta \Delta U_n}
416: \right\rangle } \right\rangle _0 \over \left\langle  e^{-\beta \Delta U}
417:  \right\rangle _0 }~.
418: \label{ratio}
419: \end{eqnarray}
420: The remarkable Eq.~\ref{f-exact} is formally exact and hasn't been given
421: before.  The correlation factors $x_{0/m}$ might, in principle, be
422: investigated on the basis of simulation data and information theory
423: analysis.  That is likely to a specialized nontrivial activity except of
424: the lower density cases where the primitive quasi-chemical
425: approximation is satisfactory.
426: 
427: 
428: \subsection{Iterating the Quasi-Chemical Analysis}
429: 
430: Alternatively, the quasi-chemical rules suggest natural
431: theoretical approximation for the equilibrium ratios given
432: formally by  Eq.~\ref{K}. Applying
433: the rule Eq.~\ref{qcarule}, for n$>$0,
434: \begin{eqnarray}
435: K_n  & = & \left\langle {\left\langle e^{-\beta \Delta U_n}
436: \right\rangle } \right\rangle _0 K_n{}{}^{(0)}\zeta^{n+1} \nonumber \\
437:  & = & { K_n{}^{(0)}\zeta^{n+1} \over 1 + \sum\limits_{m\ge 1} K_{m/n}
438:  \rho_\mathrm{S}{}^m  } \nonumber \\
439:  & \approx & { K_n{}^{(0)}\zeta^{n+1} \over 1 + \sum\limits_{m\ge 1}
440:  K_{m/ n}{}^{(0)} \rho_\mathrm{S}{}^m  }~.
441: \label{iterate}
442: \end{eqnarray}
443: The  K$_{m/n}$ can be understood by considering the chemical equilibrium
444: \begin{eqnarray} \mathrm{AS'_{n}S_{m=0}\ +\ mS \rightleftharpoons AS'_{n}S_m}~,
445: \label{rereaction} \end{eqnarray}
446: {\it i.e.\/} the original $\mathrm{AS'_{n}}$ cluster is the solute and it
447: provides a nucleus for a constellation of m S particles, different in type
448: for the S$'$ species.   How to address the calculation of the  K$_{m/ n}{}^{(0)}$
449: is discussed in Appendix~\ref{ab}.
450: 
451: It is still helpful to focus on the populations even though more coefficients
452: are involved now.   To do this we consider 
453: \begin{eqnarray}
454: \zeta =1+\zeta \sum\limits_{m\ge 1} {\left\langle {\left\langle
455: {e^{-\beta \Delta U_m}} \right\rangle } \right\rangle _0K_m{}^{(0)}\rho_\mathrm{S}{}
456: ^m\zeta ^m}
457: \end{eqnarray}
458: and, to accomodate the additional factor of $\zeta$ multiplying
459: the terms m$\ge$1, rearrange  so that
460: \begin{eqnarray}
461: \zeta =1+{{\sum\limits_{m\ge 1} {\left\langle {\left\langle {e^{-\beta
462: \Delta U_m}} \right\rangle } \right\rangle _0K_m{}^{(0)}\rho_\mathrm{S}{} ^m\zeta ^m}}
463: \over {1-\sum\limits_{m\ge 1} {\left\langle {\left\langle {e^{-\beta
464: \Delta U_m}} \right\rangle } \right\rangle _0K_m{}^{(0)}\rho_\mathrm{S}{} ^m\zeta
465: ^m}}}~.
466: \label{iqca}
467: \end{eqnarray}
468: This last equation is significant particularly because it suggests that a principal
469: consequence of correlations can be a uniform reweighting of all coefficients
470: m$\ge$1.  Strikingly, that is
471: exactly the suggestion of Fig.~\ref{approx.04:fig}.
472: 
473: 
474: We can use this insight to push the argument further:  the fact that the
475: {\em primitive} quasi-chemical populations for m$\ge$1 are correct relative to
476: each other means that the quantities $\left\langle {\left\langle
477: {e^{-\beta \Delta U_m}} \right\rangle } \right\rangle _0$ are nearly
478: exponentially dependent on m.  For, in the first place, when the density
479: is high, almost all the population is in the center of the distribution,
480: and the Lagrange multipliers are negligibly affect by the relative
481: reweighting of the m=0 term.  Then the alteration of the original geometric weighting is
482: literally irrelevant.  In the second place, when the density is
483: sufficiently low, these correlation factors are nearly unity anyway.
484: So we can accurately write
485: \begin{eqnarray}
486: \zeta \approx 1+A(\rho_\mathrm{S}{})\sum\limits_{m\ge 1}
487: K_m{}^{(0)}\rho_\mathrm{S}{}^m\lambda{}^m~. \label{iqca.1}
488: \end{eqnarray}
489: x$_0$ `breaks away' from the rest of the distribution and
490: requires individual consideration when the density is high enough that x$_0$ is sufficiently small due to 
491: correlation effects.  Nevertheless
492: \begin{eqnarray}
493: A(\rho_\mathrm{S}{}) \approx  {\zeta -1 \over \zeta_0 -1}~,
494: \label{fit-eq}
495: \end{eqnarray}
496: where $\zeta_0$ is the primitive quasi-chemical approximate value. Thus,
497: when the primitive quasi-chemical approximation is sufficiently
498: accurate, the difficulty of evaluating the corrections should be much
499: reduced.
500: 
501: Though it would be interesting to calculate correlation corrections on
502: the basis of Eq.~\ref{iterate} and Appendix~\ref{ab}, a simpler,
503: empirical approach suffices for our present purposes.  This is because
504: the discrepancies seen in Fig.~\ref{3dHS:fig} are substantial but not
505: problematic and, therefore, the required A($\rho_\mathrm{S}{}$) is
506: simple.  In particular, the form
507: \begin{eqnarray}
508: \beta \Delta \mu =\ln \left( {1+e^{7.361\rho_\mathrm{S}{}
509: ^4}\sum\limits_{m\ge 1} {K_m{}^{(0)}\rho_\mathrm{S}{}^m\lambda{}^m}}
510: \right)  \nonumber  \\
511: \label{supertheory}
512: \end{eqnarray}
513: conforms accurately to the Carnahan-Starling equation of state;  see
514: Fig.~\ref{dis3d:fig}.  The literal coefficient in Eq.~\ref{supertheory}
515: was obtained by fitting on the basis of the Eq.~\ref{fit-eq} to minimize
516:  the discrepancy with the Carnahan-Starling equation of state.
517: 
518: \begin{figure}
519: \begin{center}
520: \leavevmode
521: \includegraphics[scale=0.77]{dis3d.02.eps}
522: \end{center}
523: \caption{Equation of state with the empirical correlation model,
524: Eq.~\ref{supertheory}.  (Lower panel) The solid line is the
525: Carnahan-Starling equation of state and the dashed line is the model of
526: Eq.~\ref{supertheory}. (Upper panel) Discrepancy: the empirical
527: correlation model (Eq.~\ref{supertheory}) less the Carnahan-Starling
528: value. The mean absolute discrepancy against Carnahan-Starling equation
529: of state is about 1\% and the maximum discrepancy is less than 3\%,
530: nearly as good conformance to the Carnahan-Starling model as that model
531: to simulation data.  When the final empirical parameter was fitted using
532: only $\rho_\mathrm{S}d^3\le$ 0.3, the mean absolute discrepancy hardly
533: changed but the maximum discrepancy doubled. }
534: \label{dis3d:fig}
535: \end{figure}
536: 
537: The occupancies predicted by this empirical model are depicted in
538: Fig.~\ref{iqca.04:fig}.
539: 
540: 
541: \begin{figure}
542: \begin{center}
543: \leavevmode
544: \includegraphics[scale=0.8]{iqca.04.eps}
545: \end{center}
546: \caption{Predicted occupancies with the empirical correlation model,
547: Eq.~\ref{supertheory}, for $\rho_\mathrm{S}d^3$=0.8. Compare to
548: Fig.~\ref{compare.04:fig}.  Again the solid line is the Poisson
549: distribution.}
550: \label{iqca.04:fig}
551: \end{figure}
552: 
553: \section{Concluding Discussion}
554: Our first goal was to work-out how the quasi-chemical
555: theory, a fresh attack on the the statistical thermodynamic theory of
556: liquids, applies to the paradigmatic hard sphere fluid.  The second goal
557: was to work-out theoretical approximation procedures that might assist
558: in describing dense liquids of non-spherical species.  The new
559: fundamental results here apply generally to `hard core' molecular models.
560: The primitive quasi-chemical approximation, the procedure for
561: iterating the quasi-chemical analysis, and the recognition of the `break away'
562: phenomenon of Fig.~\ref{approx.04:fig} are likely to be helpful in
563: understanding packing in dense molecular liquids, beyond the hard
564: sphere fluid. For the hard sphere system specifically, we have obtained
565: a simple occupancy model, Eq~\ref{supertheory}, that is likely to be
566: helpful in a variety of other situations.
567: 
568: One  situation is the description of packing restrictions
569: when the quasi-chemical theory is used to treat genuinely chemical
570: interactions, for example in the study of hydration of atomic ions in
571: water\cite{Rempe:JACS:2000,Rempe:FPE:2001}.  The issue of `context
572: hydrophobicity' associated with many molecular solutes, including
573: molecular ions, in water can also be addressed on the basis of
574: quasi-chemical calculations and the
575: developments here.
576: 
577: Another situation of current interest is the theory of primitive hydrophobic effects
578: that has recently been
579: reborn\cite{Pohorille:JACS:90,Pratt:PNAS:92,Palma,Hummer:PNAS:96,%
580: Garde:PRL:96,Pratt:ECC,Hummer:PNAS:98,Hummer:JPCB:98,Pohorille:PJC:98,%
581: Pratt:NATO:99,Gomez:99,Garde:99,Hummer:CPR:2000}. An historical view has
582: been that the initial issue of hydrophobic effects was the hydration structures and
583: thermodynamics following from volume exclusion by nonpolar
584: molecules in liquid water.  The balance of attractive forces that 
585: might produce drying phenomena was a
586: secondary concern, except that `drying' was always present in the scaled particle
587: models\cite{Stillinger:73}.  With the convincing clarification of
588: the first of these
589: problems\cite{Pohorille:JACS:90,Pratt:PNAS:92,Palma,Hummer:PNAS:96,%
590: Garde:PRL:96,Pratt:ECC,Hummer:PNAS:98,Hummer:JPCB:98,Pohorille:PJC:98,%
591: Pratt:NATO:99,Gomez:99,Garde:99,Hummer:CPR:2000}, the issue of
592: drying phenomena has been now taken up more
593: enthusiastically\cite{Hummer:98,LCW,Hummer:CPR:2000}.  In this context, we note
594: that the striking success of the two-moment information models and the
595: Pratt-Chandler
596: theory\cite{Pohorille:JACS:90,Pratt:PNAS:92,Palma,Hummer:PNAS:96,%
597: Garde:PRL:96,Pratt:ECC,Hummer:PNAS:98,Hummer:JPCB:98,Pohorille:PJC:98,%
598: Pratt:NATO:99,Gomez:99,Garde:99,Hummer:CPR:2000,PC} is due, in part, to
599: a fortuitous balance of a `gaussian' approximation in the
600: theory\cite{Pratt:NATO:99} and a compensating disregard for
601: drying possibilities\cite{Pratt:80}; both of these compensating approximations are
602: expected to be benign for small molecule solutes\footnote{Another factor is that the dimensionless densities
603: relevant for the first hydrophobic applications aren't as large as the
604: most difficult cases considered here.}.  One ingredient in a
605: better understanding of this situation is a careful solution to the case
606: where drying phenomena are entirely absent.  That ingredient is better
607: in hand with the results above.
608: 
609: \begin{acknowledgments}This work was supported by the US Department of
610: Energy under contract W-7405-ENG-36 and the LDRD program at Los Alamos.
611: LA-UR-01-1530. Work at the INEEL was supported by the Office of
612: Environmental Management, U.S. Department of Energy, under DOE-ID
613: Operations Office Contract DE-AC07-99ID13727.
614: % put your acknowledgments here.
615: \end{acknowledgments}
616: 
617: 
618: 
619: \appendix
620: 
621: \section{Preliminary Results}\label{aprelim}
622: \subsection{Potential Distribution Theorem}
623: The potential distribution theorem\cite{Pratt:ECC,Pratt:MP:98,Pratt:ES:99,Widom:JPC:82}
624: may be expressed as 
625: 
626: \begin{eqnarray}
627: \rho _\sigma = \left\langle {\left\langle {e^{-\beta \Delta U}}
628: \right\rangle } \right\rangle _0z_\sigma \left( {q_\sigma /V} \right)
629: \label{double}
630: \end{eqnarray}
631: where $\rho_\sigma$ is the density of molecules of type
632: $\sigma$ (the `solute' under consideration), $z_\sigma =
633: \exp(\beta\mu_\sigma)$ is the absolute activity of that species,
634: q$_\sigma$ is the single molecule partition function for that species,
635: and V is the volume.  The double brackets
636: $\left<\left<\ldots\right>\right>_0$ indicate the average over the
637: thermal motion of the solute {\em and} the solvent under the conditions
638: of no interaction between them, and the averaged quantity is the
639: Boltzmann factor of those interactions. The average indicated here is
640: the ratio of the activity of an isolated solute, $\rho_\sigma
641: V/{q_\sigma}$, divided by the absolute activity, $z_\sigma$, of the
642: actual solute.  Thus
643: \begin{eqnarray}
644: \beta\mu_\sigma = \ln \left[{V \rho _\sigma \over \left\langle
645: {\left\langle {e^{-\beta \Delta U}} \right\rangle } \right\rangle _0
646: q_\sigma }\right]~.
647: \label{mu}
648: \end{eqnarray}
649: This is a formal result to the extent that evaluation of the quantities
650: on the right side typically will  involve nontrivial calculations on
651: many-body systems.
652: 
653: \subsection{Chemical Equilibrium}
654: The traditional chemical thermodynamic consideration of a chemical transformation such as
655: \begin{eqnarray}
656: \mathrm{n_\mathrm{A} A + n_\mathrm{B} B \rightleftharpoons  n_\mathrm{C} C + n_\mathrm{D} D }
657: \end{eqnarray}
658: with the formal result of Eq.~\ref{mu} leads to the formal expression
659: \begin{eqnarray}
660: K & \equiv & {\rho _\mathrm{C}{}^{n_\mathrm{C}} \rho _\mathrm{D}{}^{n_\mathrm{D}}\over \rho _\mathrm{A}{}^{n_\mathrm{A}} \rho _\mathrm{B}{}^{n_\mathrm{B}} } \nonumber \\
661: & = &
662: {
663: \left(\left\langle\left\langle e^{-\beta \Delta U_\mathrm{C}} \right\rangle \right\rangle _0
664: {q_\mathrm{C}\over V} \right){}^{n_\mathrm{C}} 
665: \left(\left\langle\left\langle e^{-\beta \Delta U_\mathrm{D}} \right\rangle \right\rangle _0
666: {q_\mathrm{D}\over V} \right){}^{n_\mathrm{D}}\over 
667: \left(\left\langle\left\langle e^{-\beta \Delta U_\mathrm{A}} \right\rangle \right\rangle _0
668: {q_\mathrm{A}\over V} \right){}^{n_\mathrm{A}} 
669: \left(\left\langle\left\langle e^{-\beta \Delta U_\mathrm{B}} \right\rangle \right\rangle _0
670: {q_\mathrm{B}\over V} \right){}^{n_\mathrm{B}} 
671: }~.
672: \label{K}
673: \end{eqnarray}
674: This should be compared to the textbook result for ideal gas
675: systems\cite{mcq}. That comparison shows that the single molecule
676: partition functions are multiplicatively augmented by the test particle
677: averages\footnote{A more conventional view of these results is that the
678: concentrations are multiplicatively augmented by the test particle
679: averages and that serves to identify activity coefficients.  That point
680: of view has no special utility here.}.  But otherwise the structure of
681: this important result is unchanged. The conclusion here is that the
682: equilibrium ratios are well-defined objects though formal to the extent
683: that nontrivial computational effort would be required to evaluate them
684: on the basis of molecular information.
685: 
686: \subsection{Quasi-chemical Theory}
687: The quasi-chemical develop starts from consideration of a distinguished
688: molecule in the solution and seeks to evaluate the chemical potential on
689: the basis of events occurring within a defined `inner sphere.' For a
690: species of type A, that definition is codified by specifying a function
691: $b_{\mathrm{A} j}$ that is equal to one (1) when solution molecule j is
692: inside the defined region and zero (0) otherwise. Our starting point
693: can be \cite{Pratt:ES:99}
694: \begin{eqnarray}
695: \beta\Delta\mu_\mathrm{A} &  = & \ln x_0  \nonumber \\
696: & - & \ln\left[
697: \left\langle\left\langle {e^{-\beta
698: \Delta U_\mathrm{A}}}\prod\limits_j {(1-b_{\mathrm{A} j})}
699: \right\rangle\right\rangle_0  \right],
700: \label{gqca} \end{eqnarray}
701: where $x_0$ is the fraction of A solute species with zero (0) neighbors
702: in the defined region.  $\Delta U_\mathrm{A}$ is the interaction energy
703: of the solvent with the solute A that is treated as a test particle.
704: The potential distribution theory perspective on Eq.~(\ref{gqca}) is
705: \begin{eqnarray}
706: x_0 & = & \left\langle \prod\limits_j {(1-b_{\mathrm{A} j})} \right\rangle
707: \nonumber \\ 
708: & = &{ \langle\langle{e^{-\beta \Delta U_\mathrm{A}}}\prod\limits_j
709: {(1-b_{\mathrm{A} j})}\rangle\rangle_0 \over \langle\langle {e^{-\beta \Delta U_\mathrm{A}}}
710: \rangle\rangle_0} .
711: \label{x0}
712: \end{eqnarray}
713: 
714: The first, or {\em chemical} term, of Eq.~\ref{gqca} can be analyzed with chemical
715: concepts associated with the reactions
716: \begin{eqnarray} \mathrm{AS_{n=0}\ +\ nS \rightleftharpoons AS_n}
717: \label{reaction} \end{eqnarray}
718: Here the indicated complexes are
719: composed of $n$ solvent (S) molecules within the defined
720: region.  Remember that the
721: A molecule is a `distinguished' solute molecule considered at the
722: lowest non-zero concentration \cite{Pratt:MP:98}. The fractional amount of
723: A species with a given solvation number $n$ can be described by a
724: chemical equilibrium ratio
725: \begin{eqnarray}
726: K_n={\rho_\mathrm{AS_n} \over \rho_\mathrm{AS_{n=0}}
727: \rho_\mathrm{S}{}^n }~.
728: \label{Kn-ob}
729: \end{eqnarray}
730: The $\rho_\sigma$ are the number densities and, in particular,
731: $\rho_\mathrm{S}$ is the bulk number density of solvent molecules since the
732: distinguished A molecule is infinitely dilute. This notation
733: permits the normalized re-expression
734: \begin{eqnarray}
735: x_n={{K_{n}\rho _\mathrm{S}{}^n} \over {1 + \sum\limits_{m\ge 1}
736: {K_{m}\rho _\mathrm{S}{}^m}}} .
737: \label{cluster-var}
738: \end{eqnarray}
739: Since this yields
740: \begin{eqnarray}
741: x_0{}^{-1}={1 + \sum\limits_{m\ge 1}
742: {K_{m}\rho _\mathrm{S}{}^m}} ,
743: \label{x0.f}
744: \end{eqnarray}
745: the original Eq.~(\ref{gqca}) can be re-expressed as 
746: \begin{eqnarray}
747: \beta\Delta\mu_\mathrm{A}  &=&  -\ln \left[1 +  \sum\limits_{m\ge 1}
748: {K_{m}\rho _\mathrm{S}{}^m}\right] \nonumber\\&&- \ln\left[
749: \left\langle\left\langle {e^{-\beta
750: \Delta U}}\prod\limits_j {(1-b_{\mathrm{A} j})}
751: \right\rangle\right\rangle_0  \right].
752: \label{gqca.f} \end{eqnarray}
753: The virtue of these rearrangements is that the natural first approximation is
754: \begin{eqnarray}
755: x_n\approx x_n{}^{(0)} = {{K_{n}{}^{(0)}\rho _\mathrm{S}{}^n} \over
756: {1 + \sum\limits_{m\ge 1} {K_{m}{}^{(0)}\rho _\mathrm{S}{}^m}}} .
757: \label{cluster-var.0}
758: \end{eqnarray}
759:  The K$_{n}{}^{(0)}$ are equilibrium ratios for the chemical reaction
760: (\ref{reaction}) in an ideal gas.  This formulation and the
761: approximation of Eq.~(\ref{cluster-var.0}) are closely related
762: \cite{Pratt:ES:99} to the quasi-chemical (or cluster-variation)
763: approximations of Guggenheim \cite{Guggenheim:35}, Bethe
764: \cite{Bethe:35}, and Kikuchi \cite{brush}.
765: 
766: This approach should have greatest utility where the chemical balances of
767: Eq.~\ref{reaction} are dominated by inner sphere chemistry that can be
768: captured with computations on clusters.  Such chemical interactions are
769: often much larger than the {\em outer sphere\/} contribution, the
770: right-most term of Eq.~\ref{gqca}.  
771: 
772: But that outer sphere contribution remains and can't be neglected
773: forever.  An interesting example based on simulation of liquid water was
774: discussed recently\cite{Hummer:CPR:2000}.  There the x$_0$ was estimated
775: from molecular dynamics results and the remainder, the outer-sphere
776: contributions to $\beta\Delta\mu$, were positive, suggesting domination
777: of those outer-sphere contributions  by the packing constraints studied
778: here. A principal goal of the present work is the development of a
779: reasonable approach for describing the packing issues necessary for
780: treating those outer sphere contributions.
781: 
782: Reiss and Merry\cite{Reiss:81} analyzed population relations analogous to
783: Eq.~\ref{cluster-var} but with activities appearing in the place of
784: densities and with coefficients, here the equilibrium ratios K$_n$,
785: appropriately different.  The additional formal point here is the
786: replacement of the activity by the density that permits the
787: identification of the K$_n$ in Eq.~\ref{hsqca}, and then further permits
788: consideration of the mean field treatment Eq.~\ref{pqca} on the basis of an
789: information theory constraint when  K$_n{}^{(0)}$ will be used.  At this
790: stage, the quasi-chemical approximation achieves a particularly
791: primitive character and deviates from the goal of bounding these
792: thermodnamic quantities that was pursued by Reiss and
793: Merry\cite{Reiss:81}.
794: 
795: 
796: 
797: \section{Calculation of the K\lowercase{$_n{}^{(0)}$} for hard spheres and hard
798: disks}\label{aa} The K$_n{}^{(0)}$ sought for reaction
799: Eq.~\ref{reaction} are given by
800: \begin{eqnarray}
801: K_n{}^{(0)}  = 
802: { {q_\mathrm{AS_n}}  \over 
803: \left({q_\mathrm{S}/ V}\right)^n 
804: {q_\mathrm{AS_{n=0}}}} ~.
805: \label{K0n}
806: \end{eqnarray}
807: (See Eq.~\ref{K}.) For this problem, $ q_\mathrm{S} = V/\Lambda_S{}^3$
808: with $ \Lambda_S$ a thermal deBroglie wavelength for S but these momentum
809: integrals cancel perfectly and are irrelevant as usual.  Therefore,
810: \begin{eqnarray}
811: n!K_n^{(0)}=\int\limits_A {d^3r_1\ldots \int\limits_A {d^3r_n\left(
812: {\prod\limits_{j>i=1}^n {e(i,j)}} \right)}}~.
813: \label{integral}
814: \end{eqnarray}
815: The notation $\int_A {d^3r_k}$ indicates the three-dimensional spatial
816: integral over the volume of the A-ball, a sphere of radius 1.  The
817: indicated integrand is thus 3n dimensional. The integrand is zero (0) if
818: $\vert\bf{r}_i-\bf{r}_j\vert<1$ (overlap) for any (ij) and one (1)
819: otherwise.  Thus the integral can be estimated by sampling n-point
820: uniform placments  in the A-ball and scoring the fraction of such
821: placements that are free from overlaps between the n unit diameter
822: S-spheres.  This approach fails for n larger than those presented in
823: Table~\ref{tab:kn}. But larger clusters were not observed in our
824: simulation of the fluid, so our approach should be regarded as
825: satisfactory.
826: 
827: The analogous two dimensional procedure was used for the hard disk results.
828: 
829: \section{Calculation of the K\lowercase{$_{m/n}{}^{(0)}$}}\label{ab}
830: In contrast to Appendix \ref{aa}, here the ratio sought is
831: \begin{eqnarray}
832: K_{m/n}{}^{(0)}  = 
833: { {q_\mathrm{AS_n'S_m}}  \over 
834: \left({q_\mathrm{S}/ V}\right)^m 
835: {q_\mathrm{AS_{n}'}}} ~,
836: \label{K0mn}
837: \end{eqnarray}
838: corresponding to the reaction Eq.~\ref{rereaction}. Again, the explicit
839: factors of V, the momentum integrals, and the factor of $n!$ all cancel
840: perfectly so that
841: \begin{eqnarray}
842: m!K_{m/n}^{(0)}=\left\langle {\int\limits_{AS'_n} {d^3r_1\ldots
843: \int\limits_{AS'_n} {d^3r_m\left( {\prod\limits_{j>i=1}^m {e(i,j)}}
844: \right)}}} \right\rangle~.\nonumber \\
845: \label{ave-int} 
846: \end{eqnarray}
847: Here the notation $\int_{AS_n'} {d^3r_k}$ indicates an integral over the
848: excluded volume of an AS$_n'$ complex to an S ligand.  The SS excluded
849: volume, the integrand, is the same as before.  But the structure of
850: the AS$_n'$ complex fluctuates and the volumes obtained for specific
851: structures are averaged over these fluctuations.  The brackets $\langle
852: \ldots \rangle$ indicate the average over the structures of the isolated
853: AS$_n'$ complex. This averaging is permitted and governed by the
854: non-trivial denominator that appears in Eq.~\ref{K0mn}.
855: 
856: Operationally, the calculation can be much as in Appendix \ref{aa}
857: except for (a)  averaging utlizing a Metropolis Monte Carlo calculation for
858: the n ligand spheres in the star AS$_n'$; and (b) random placements of
859: the m additional points are into a sphere of radius two (2) since that
860: would fully enclose any conformation of the cluster.
861: 
862: 
863: \section{Calculation of \lowercase{x$_{n}$} for the Hard Sphere Fluid}
864: \label{amc}
865:  
866: The probability that there are $\mathrm{n-1}$ points in a sphere of
867: radius $r$, x$_{n-1}(r)$, can be obtained from the
868: distribution, $4 \pi \rho_S r ^2\mathcal{D}_{n}(r)$,  of the distance $r$ to the
869: nth nearest neighbor of an arbitrary point. The probability that there
870: are no more than $\mathrm{n-1}$ molecules in the void is equal to the
871: probability that the nth nearest neighbor is at least $r$ away
872: from the void center
873:  
874: \begin{equation}
875: \sum_{m=0}^{n-1} x_{m}(r) = 4 \pi \rho_S\int_{r}^{\infty} \mathcal{D}_{n}(y)
876: y ^2 dy ~.
877: \end{equation}
878: Isobaric-isothermal Monte Carlo can be used to calculate
879: $\mathcal{D}_{n}(r)$.  x$_n(r)$ for a
880: range of $r$ can be obtained from the distributions
881: $\mathcal{D}_{n}(r)$. To increase the accuracy of the estimated
882: $\mathcal{D}_{n}(r)$ for rarely observed $r$, small and large, a
883: specific point in the simulation volume was chosen, and the sampling
884: probability was reweighted by $\left[4\pi \rho_\mathrm{S}r_1{}^2
885: e^{-4\pi \rho_\mathrm{S}r_1{}^3/3}+ C\right]^{-1}$ where $r_j$ is the
886: distance from the chosen point to the jth nearest center for each
887: configuration and C is an empirically chosen, dimensional constant. This
888: importance sampling
889: 
890: \begin{equation}
891: 4 \pi \rho_S r^2\mathcal{D}_{n}(r)=
892: {
893: \left<\left[4\pi \rho_\mathrm{S}r_1{}^2 e^{-4\pi \rho_\mathrm{S}r_1{}^3/3}+
894: C\right] \delta(r-r_n)\right>
895: \over
896: \left<\left[4\pi \rho_\mathrm{S}r_1{}^2 e^{-4\pi \rho_\mathrm{S}r_1{}^3/3}+
897: C\right]\right>
898: }
899: \label{importance}
900: \end{equation}
901: is based upon the idea that $\mathcal{D}_{1}^{(0)}(r)=e^{-4\pi
902: \rho_\mathrm{S}r^3/3}$ is the function\cite{Chandrasekhar}
903: appropriate for a random distribution of spheres. This idea attempts to
904: make the observed distribution of the distance to the nearest particle
905: more nearly uniform. The constant C was included to avoid an unbounded
906: weighting function.  The denominator of Eq. D2 is just a normalizing
907: factor on the distribution. The denominator of Eq.~\ref{importance}
908: merely provides a normalizing factor.
909: 
910: Isobaric-isotermal ensembles of 108 and 256 hard spheres were
911: sufficient.  The Carnahan-Starling equation was used to find the $\beta$p
912: needed for a hard sphere simulation at each specific density.
913: 
914: % If in two-column mode, this environment will change to single-column
915: % format so that long equations can be displayed. Use
916: % sparingly.
917: %\begin{widetext}
918: % put long equation here
919: %\end{widetext}
920: 
921: % figures should be put into the text as floats.
922: % Use the graphics or graphicx packages (distributed with LaTeX2e).
923: % See the LaTeX Graphics Companion by Michel Goosens, Sebastian Rahtz,
924: % and Frank Mittelbach for instance.
925: %
926: % Here is an example of the general form of a figure:
927: % Fill in the caption in the braces of the \caption{} command. Put the label
928: % that you will use with \ref{} command in the braces of the \label{} command.
929: %
930: % \begin{figure}
931: % \includegraphics{}%
932: % \caption{}
933: % \label{}
934: % \end{figure}
935: 
936: % tables follow here or maybe be put in the text
937: %
938: % Here is an example of the general form of a table:
939: % Fill in the caption in the braces of the \caption{} command. Put the label
940: % that you will use with \ref{} command in the braces of the \label{} command.
941: % Insert the column specifiers (l, r, c, d, etc.) in the empty braces of the
942: % \begin{tabular}{} command.
943: % The ruledtabular enviroment adds doubled rules to table and sets a
944: % nice set of default table settings.
945: % Use the table* environment to get a full-width table in two-column
946: % \begin{table}
947: % \caption{}
948: % \label{}
949: % \begin{ruledtabular}
950: % \begin{tabular}{}
951: % \end{tabular}
952: % \end{ruledtabular}
953: % \end{table}
954: 
955: % If you have acknowledgments, this puts in the proper section head.
956: 
957: % Create the reference section using BibTeX:
958: %\bibliography{ms}
959: \begin{thebibliography}{10}
960: \expandafter\ifx\csname bibnamefont\endcsname\relax
961:   \def\bibnamefont#1{#1}\fi
962: \expandafter\ifx\csname bibfnamefont\endcsname\relax
963:   \def\bibfnamefont#1{#1}\fi
964: \expandafter\ifx\csname url\endcsname\relax
965:   \def\url#1{\texttt{#1}}\fi
966: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
967: \providecommand{\bibinfo}[2]{#2}
968: \providecommand{\eprint}[2][]{\url{#2}}
969: 
970: \bibitem{Pratt:MP:98}
971: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}} \bibnamefont{and}
972:   \bibinfo{author}{\bibfnamefont{R.~A.} \bibnamefont{LaViolette}},
973:   \bibinfo{journal}{Mol. Phys.} \textbf{\bibinfo{volume}{94}},
974:   \bibinfo{pages}{909} (\bibinfo{year}{1998}).
975: 
976: \bibitem{Pratt:ES:99}
977: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}} \bibnamefont{and}
978:   \bibinfo{author}{\bibfnamefont{S.~B.} \bibnamefont{Rempe}}, in
979:   \emph{\bibinfo{booktitle}{Simulation and Theory of Electrostatic Interactions
980:   in Solution. Computational Chemistry, Biophysics, and Aqueous Solutions}},
981:   edited by \bibinfo{editor}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}}
982:   \bibnamefont{and} \bibinfo{editor}{\bibfnamefont{G.}~\bibnamefont{Hummer}}
983:   (\bibinfo{publisher}{American Institute of Physics},
984:   \bibinfo{address}{Melville, NY}, \bibinfo{year}{1999}), vol.
985:   \bibinfo{volume}{492} of \emph{\bibinfo{series}{AIP Conference Proceedings}},
986:   pp. \bibinfo{pages}{172--201}.
987: 
988: \bibitem{Hummer:CPR:2000}
989: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
990:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
991:   \bibinfo{author}{\bibfnamefont{A.~E.} \bibnamefont{Garc\'{\i}a}},
992:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
993:   \bibinfo{journal}{Chem. Phys.} \textbf{\bibinfo{volume}{258}},
994:   \bibinfo{pages}{349} (\bibinfo{year}{2000}).
995: 
996: \bibitem{WCA}
997: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Chandler}},
998:   \bibinfo{author}{\bibfnamefont{J.~D.} \bibnamefont{Weeks}}, \bibnamefont{and}
999:   \bibinfo{author}{\bibfnamefont{H.~C.} \bibnamefont{Andersen}},
1000:   \bibinfo{journal}{Science} \textbf{\bibinfo{volume}{220}},
1001:   \bibinfo{pages}{787} (\bibinfo{year}{1983}).
1002: 
1003: \bibitem{Altenberger:96}
1004: \bibinfo{author}{\bibfnamefont{A.~R.} \bibnamefont{Altenberger}}
1005:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.~S.}
1006:   \bibnamefont{Dahler}}, \bibinfo{journal}{Phys. Rev. E}
1007:   \textbf{\bibinfo{volume}{54}}, \bibinfo{pages}{6242} (\bibinfo{year}{1996}).
1008: 
1009: \bibitem{Robles:98}
1010: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Robles}},
1011:   \bibinfo{author}{\bibfnamefont{M.~L.} \bibnamefont{deHaro}},
1012:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Santos}}, \bibnamefont{and}
1013:   \bibinfo{author}{\bibfnamefont{S.~B.} \bibnamefont{Yuste}},
1014:   \bibinfo{journal}{J. Chem. Phys.} \textbf{\bibinfo{volume}{108}},
1015:   \bibinfo{pages}{1290} (\bibinfo{year}{1998}).
1016: 
1017: \bibitem{Yelash:99}
1018: \bibinfo{author}{\bibfnamefont{L.~V.} \bibnamefont{Yelash}},
1019:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Kraska}}, \bibnamefont{and}
1020:   \bibinfo{author}{\bibfnamefont{U.~K.} \bibnamefont{Deiters}},
1021:   \bibinfo{journal}{J. Chem. Phys.} \textbf{\bibinfo{volume}{110}},
1022:   \bibinfo{pages}{3079} (\bibinfo{year}{1999}).
1023: 
1024: \bibitem{Parisi:00}
1025: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Parisi}} \bibnamefont{and}
1026:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Slanina}},
1027:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{62}},
1028:   \bibinfo{pages}{6554} (\bibinfo{year}{2000}).
1029: 
1030: \bibitem{Pohorille:JACS:90}
1031: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Pohorille}} \bibnamefont{and}
1032:   \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1033:   \bibinfo{journal}{J. Am. Chem. Soc.} \textbf{\bibinfo{volume}{112}},
1034:   \bibinfo{pages}{5066} (\bibinfo{year}{1990}).
1035: 
1036: \bibitem{Pratt:PNAS:92}
1037: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}} \bibnamefont{and}
1038:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Pohorille}},
1039:   \bibinfo{journal}{Proc. Natl. Acad. Sci. USA} \textbf{\bibinfo{volume}{89}},
1040:   \bibinfo{pages}{2995} (\bibinfo{year}{1992}).
1041: 
1042: \bibitem{Palma}
1043: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}} \bibnamefont{and}
1044:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Pohorille}}, in
1045:   \emph{\bibinfo{booktitle}{Proceedings of the {EBSA} 1992 International
1046:   Workshop on Water-Biomolecule Interactions}}, edited by
1047:   \bibinfo{editor}{\bibfnamefont{M.~U.} \bibnamefont{Palma}},
1048:   \bibinfo{editor}{\bibfnamefont{M.~B.} \bibnamefont{Palma-Vittorelli}},
1049:   \bibnamefont{and} \bibinfo{editor}{\bibfnamefont{F.}~\bibnamefont{Parak}}
1050:   (\bibinfo{publisher}{Societ\'{a} Italiana de Fisica},
1051:   \bibinfo{address}{Bologna}, \bibinfo{year}{1993}), pp.
1052:   \bibinfo{pages}{261--268}.
1053: 
1054: \bibitem{Hummer:PNAS:96}
1055: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
1056:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
1057:   \bibinfo{author}{\bibfnamefont{A.~E.} \bibnamefont{Garc{\'{\i}}a}},
1058:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Pohorille}},
1059:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1060:   \bibinfo{journal}{Proc. Natl. Acad. Sci. USA} \textbf{\bibinfo{volume}{93}},
1061:   \bibinfo{pages}{8951} (\bibinfo{year}{1996}).
1062: 
1063: \bibitem{Garde:PRL:96}
1064: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
1065:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
1066:   \bibinfo{author}{\bibfnamefont{A.~E.} \bibnamefont{Garc{\'{\i}}a}},
1067:   \bibinfo{author}{\bibfnamefont{M.~E.} \bibnamefont{Paulaitis}},
1068:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1069:   \bibinfo{journal}{Phys. Rev. Lett.}
1070:   \textbf{\bibinfo{volume}{77}}(\bibinfo{number}{24}), \bibinfo{pages}{4966}
1071:   (\bibinfo{year}{1996}).
1072: 
1073: \bibitem{Pratt:ECC}
1074: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1075:   \emph{\bibinfo{title}{Encyclopedia of Computational Chemistry}}
1076:   (\bibinfo{publisher}{John Wiley \& Sons}, \bibinfo{address}{Chichester},
1077:   \bibinfo{year}{1998}), pp. \bibinfo{pages}{1286--1294},
1078:   \bibinfo{note}{hydrophobic effects}.
1079: 
1080: \bibitem{Hummer:PNAS:98}
1081: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
1082:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
1083:   \bibinfo{author}{\bibfnamefont{A.~E.} \bibnamefont{{Garc\'{\i}a}}},
1084:   \bibinfo{author}{\bibfnamefont{M.~E.} \bibnamefont{Paulaitis}},
1085:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1086:   \bibinfo{journal}{Proc. Natl. Acad. Sci. USA} \textbf{\bibinfo{volume}{95}},
1087:   \bibinfo{pages}{1552} (\bibinfo{year}{1998}).
1088: 
1089: \bibitem{Hummer:JPCB:98}
1090: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
1091:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
1092:   \bibinfo{author}{\bibfnamefont{A.~E.} \bibnamefont{Garcia}},
1093:   \bibinfo{author}{\bibfnamefont{M.~E.} \bibnamefont{Paulaitis}},
1094:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1095:   \bibinfo{journal}{J. Phys. Chem. B} \textbf{\bibinfo{volume}{102}},
1096:   \bibinfo{pages}{10469} (\bibinfo{year}{1998}).
1097: 
1098: \bibitem{Pohorille:PJC:98}
1099: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Pohorille}},
1100:   \bibinfo{journal}{Pol. J. Chem.} \textbf{\bibinfo{volume}{72}},
1101:   \bibinfo{pages}{1680} (\bibinfo{year}{1998}).
1102: 
1103: \bibitem{Pratt:NATO:99}
1104: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Pratt}},
1105:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}}, \bibnamefont{and}
1106:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}}, in
1107:   \emph{\bibinfo{booktitle}{New Approaches to Problems in Liquid State
1108:   Theory}}, edited by
1109:   \bibinfo{editor}{\bibfnamefont{C.}~\bibnamefont{Caccamo}},
1110:   \bibinfo{editor}{\bibfnamefont{J.-P.} \bibnamefont{Hansen}},
1111:   \bibnamefont{and} \bibinfo{editor}{\bibfnamefont{G.}~\bibnamefont{Stell}}
1112:   (\bibinfo{publisher}{Kluwer}, \bibinfo{address}{Netherlands},
1113:   \bibinfo{year}{1999}), vol. \bibinfo{volume}{NATO Science Series 529}, pp.
1114:   \bibinfo{pages}{407--420}.
1115: 
1116: \bibitem{Gomez:99}
1117: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Gomez}},
1118:   \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1119:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}}, \bibnamefont{and}
1120:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}}, \bibinfo{journal}{J.
1121:   Phys. Chem. B} \textbf{\bibinfo{volume}{103}}, \bibinfo{pages}{3520}
1122:   (\bibinfo{year}{1999}).
1123: 
1124: \bibitem{Garde:99}
1125: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
1126:   \bibinfo{author}{\bibfnamefont{A.~E.} \bibnamefont{Garcia}},
1127:   \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}}, \bibnamefont{and}
1128:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
1129:   \bibinfo{journal}{Biophys. Chem.} \textbf{\bibinfo{volume}{78}},
1130:   \bibinfo{pages}{21} (\bibinfo{year}{1999}).
1131: 
1132: \bibitem{Crooks:PRE:97}
1133: \bibinfo{author}{\bibfnamefont{G.~E.} \bibnamefont{Crooks}} \bibnamefont{and}
1134:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Chandler}},
1135:   \bibinfo{journal}{Phys. Rev. E}
1136:   \textbf{\bibinfo{volume}{56}}(\bibinfo{number}{4}), \bibinfo{pages}{4217}
1137:   (\bibinfo{year}{1997}).
1138: 
1139: \bibitem{HH}
1140: \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{Hammersley}}
1141:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{D.~C.}
1142:   \bibnamefont{Handscomb}}, \emph{\bibinfo{title}{Monte Carlo Methods}}
1143:   (\bibinfo{publisher}{Chapman and Hall}, \bibinfo{address}{London},
1144:   \bibinfo{year}{1964}).
1145: 
1146: \bibitem{RH:64}
1147: \bibinfo{author}{\bibfnamefont{F.~H.} \bibnamefont{Ree}} \bibnamefont{and}
1148:   \bibinfo{author}{\bibfnamefont{W.~G.} \bibnamefont{Hoover}},
1149:   \bibinfo{journal}{J. Chem. Phys.} \textbf{\bibinfo{volume}{40}},
1150:   \bibinfo{pages}{939} (\bibinfo{year}{1964}).
1151: 
1152: \bibitem{Rempe:JACS:2000}
1153: \bibinfo{author}{\bibfnamefont{S.~B.} \bibnamefont{Rempe}},
1154:   \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1155:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}},
1156:   \bibinfo{author}{\bibfnamefont{J.~D.} \bibnamefont{Kress}},
1157:   \bibinfo{author}{\bibfnamefont{R.~L.} \bibnamefont{Martin}},
1158:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Redondo}},
1159:   \bibinfo{journal}{J. Am. Chem. Soc.} \textbf{\bibinfo{volume}{122}},
1160:   \bibinfo{pages}{966} (\bibinfo{year}{2000}).
1161: 
1162: \bibitem{Rempe:FPE:2001}
1163: \bibinfo{author}{\bibfnamefont{S.~B.} \bibnamefont{Rempe}} \bibnamefont{and}
1164:   \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}},
1165:   \bibinfo{journal}{Fluid Phase Equilibrium} \textbf{\bibinfo{volume}{xxx}},
1166:   \bibinfo{pages}{yyy} (\bibinfo{year}{2001}), \bibinfo{note}{in press,
1167:   LA-UR-00-2309, http://xxx.lanl.gov/abs/physics/0006026}.
1168: 
1169: \bibitem{Stillinger:73}
1170: \bibinfo{author}{\bibfnamefont{F.~H.} \bibnamefont{Stillinger}},
1171:   \bibinfo{journal}{J. Soln. Chem.} \textbf{\bibinfo{volume}{2}},
1172:   \bibinfo{pages}{141} (\bibinfo{year}{1973}).
1173: 
1174: \bibitem{Hummer:98}
1175: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Hummer}} \bibnamefont{and}
1176:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Garde}},
1177:   \bibinfo{journal}{Phys. Rev. Letts.} \textbf{\bibinfo{volume}{80}},
1178:   \bibinfo{pages}{4193} (\bibinfo{year}{1998}).
1179: 
1180: \bibitem{LCW}
1181: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Lum}},
1182:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Chandler}}, \bibnamefont{and}
1183:   \bibinfo{author}{\bibfnamefont{J.~D.} \bibnamefont{Weeks}},
1184:   \bibinfo{journal}{J. Phys. Chem. B} \textbf{\bibinfo{volume}{103}},
1185:   \bibinfo{pages}{4570} (\bibinfo{year}{1999}).
1186: 
1187: \bibitem{PC}
1188: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}} \bibnamefont{and}
1189:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Chandler}},
1190:   \bibinfo{journal}{J. Chem. Phys.} \textbf{\bibinfo{volume}{67}},
1191:   \bibinfo{pages}{3863} (\bibinfo{year}{1977}).
1192: 
1193: \bibitem{Pratt:80}
1194: \bibinfo{author}{\bibfnamefont{L.~R.} \bibnamefont{Pratt}} \bibnamefont{and}
1195:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Chandler}},
1196:   \bibinfo{journal}{J. Chem. Phys.} \textbf{\bibinfo{volume}{73}},
1197:   \bibinfo{pages}{3434} (\bibinfo{year}{1980}).
1198: 
1199: \bibitem{Widom:JPC:82}
1200: \bibinfo{author}{\bibfnamefont{B.}~\bibnamefont{Widom}}, \bibinfo{journal}{J.
1201:   Phys. Chem.} \textbf{\bibinfo{volume}{86}}, \bibinfo{pages}{869}
1202:   (\bibinfo{year}{1982}).
1203: 
1204: \bibitem{mcq}
1205: \bibinfo{author}{\bibfnamefont{D.~A.} \bibnamefont{McQuarrie}},
1206:   \emph{\bibinfo{title}{Statistical Mechanics}} (\bibinfo{publisher}{Harper \&
1207:   Row}, \bibinfo{address}{New York}, \bibinfo{year}{1976}),
1208:   \bibinfo{note}{chapter 9}.
1209: 
1210: \bibitem{Guggenheim:35}
1211: \bibinfo{author}{\bibfnamefont{E.~A.} \bibnamefont{Guggenheim}},
1212:   \bibinfo{journal}{Proc. Rov. Soc. London A} \textbf{\bibinfo{volume}{148}},
1213:   \bibinfo{pages}{304} (\bibinfo{year}{1935}).
1214: 
1215: \bibitem{Bethe:35}
1216: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Bethe}},
1217:   \bibinfo{journal}{Proc. Rov. Soc. London A} \textbf{\bibinfo{volume}{150}},
1218:   \bibinfo{pages}{552} (\bibinfo{year}{1935}).
1219: 
1220: \bibitem{brush}
1221: \bibinfo{author}{\bibfnamefont{S.~G.} \bibnamefont{Brush}} \bibnamefont{and}
1222:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Kikuchi}},
1223:   \emph{\bibinfo{title}{Lattice Models for Cooperative Phenomena. {I}. Survey
1224:   and Comparison of Approximate Methods}}, \bibinfo{type}{Tech. Rep.}
1225:   \bibinfo{number}{UCRL-14287}, \bibinfo{institution}{University of California,
1226:   Lawrence Radiation Laboratory, Livermore, CA} (\bibinfo{year}{1965}).
1227: 
1228: \bibitem{Reiss:81}
1229: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Reiss}} \bibnamefont{and}
1230:   \bibinfo{author}{\bibfnamefont{G.~A.} \bibnamefont{Merry}},
1231:   \bibinfo{journal}{J. Phys. Chem.} \textbf{\bibinfo{volume}{85}},
1232:   \bibinfo{pages}{3313} (\bibinfo{year}{1981}).
1233: 
1234: \bibitem{Chandrasekhar}
1235: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Chandrasekhar}},
1236:   \bibinfo{journal}{Rev. Mod. Phys.} \textbf{\bibinfo{volume}{15}},
1237:   \bibinfo{pages}{1} (\bibinfo{year}{1943}).
1238: 
1239: \end{thebibliography}
1240: 
1241: \end{document}
1242: %
1243: % ****** End of file template.aps ******
1244: 
1245: