1 |
chrisfen |
1453 |
%\documentclass[prb,aps,twocolumn,tabularx]{revtex4} |
2 |
gezelter |
1463 |
\documentclass[11pt]{article} |
3 |
|
|
\usepackage{endfloat} |
4 |
chrisfen |
1453 |
\usepackage{amsmath} |
5 |
|
|
\usepackage{epsf} |
6 |
|
|
\usepackage{berkeley} |
7 |
gezelter |
1463 |
\usepackage{setspace} |
8 |
|
|
\usepackage{tabularx} |
9 |
chrisfen |
1453 |
\usepackage{graphicx} |
10 |
gezelter |
1463 |
\usepackage[ref]{overcite} |
11 |
|
|
\pagestyle{plain} |
12 |
|
|
\pagenumbering{arabic} |
13 |
|
|
\oddsidemargin 0.0cm \evensidemargin 0.0cm |
14 |
|
|
\topmargin -21pt \headsep 10pt |
15 |
|
|
\textheight 9.0in \textwidth 6.5in |
16 |
|
|
\brokenpenalty=10000 |
17 |
|
|
\renewcommand{\baselinestretch}{1.2} |
18 |
|
|
\renewcommand\citemid{\ } % no comma in optional reference note |
19 |
chrisfen |
1453 |
|
20 |
|
|
\begin{document} |
21 |
|
|
|
22 |
chrisfen |
1905 |
\title{Computational free energy studies of a new ice polymorph which |
23 |
|
|
exhibits greater stability than Ice $I_h$} |
24 |
chrisfen |
1453 |
|
25 |
gezelter |
1463 |
\author{Christopher J. Fennell and J. Daniel Gezelter \\ |
26 |
chrisfen |
1905 |
Department of Chemistry and Biochemistry\\ |
27 |
|
|
University of Notre Dame\\ |
28 |
chrisfen |
1453 |
Notre Dame, Indiana 46556} |
29 |
|
|
|
30 |
|
|
\date{\today} |
31 |
|
|
|
32 |
gezelter |
1463 |
\maketitle |
33 |
chrisfen |
1453 |
%\doublespacing |
34 |
|
|
|
35 |
|
|
\begin{abstract} |
36 |
chrisfen |
1905 |
The absolute free energies of several ice polymorphs were calculated |
37 |
|
|
using thermodynamic integration. These polymorphs are predicted by |
38 |
|
|
computer simulations using a variety of common water models to be |
39 |
|
|
stable at low pressures. A recently discovered ice polymorph that has |
40 |
|
|
as yet {\it only} been observed in computer simulations (Ice-{\it i}), |
41 |
|
|
was determined to be the stable crystalline state for {\it all} the |
42 |
|
|
water models investigated. Phase diagrams were generated, and phase |
43 |
|
|
coexistence lines were determined for all of the known low-pressure |
44 |
|
|
ice structures. Additionally, potential truncation was shown to play |
45 |
|
|
a role in the resulting shape of the free energy landscape. |
46 |
chrisfen |
1453 |
\end{abstract} |
47 |
|
|
|
48 |
|
|
%\narrowtext |
49 |
|
|
|
50 |
|
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
51 |
|
|
% BODY OF TEXT |
52 |
|
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
53 |
|
|
|
54 |
|
|
\section{Introduction} |
55 |
|
|
|
56 |
chrisfen |
1459 |
Water has proven to be a challenging substance to depict in |
57 |
gezelter |
1463 |
simulations, and a variety of models have been developed to describe |
58 |
|
|
its behavior under varying simulation |
59 |
gezelter |
1477 |
conditions.\cite{Stillinger74,Rahman75,Berendsen81,Jorgensen83,Bratko85,Berendsen87,Caldwell95,Liu96,Berendsen98,Dill00,Mahoney00,Fennell04} |
60 |
gezelter |
1463 |
These models have been used to investigate important physical |
61 |
gezelter |
1475 |
phenomena like phase transitions, transport properties, and the |
62 |
chrisfen |
1471 |
hydrophobic effect.\cite{Yamada02,Marrink94,Gallagher03} With the |
63 |
|
|
choice of models available, it is only natural to compare the models |
64 |
|
|
under interesting thermodynamic conditions in an attempt to clarify |
65 |
chrisfen |
1905 |
the limitations of |
66 |
|
|
each.\cite{Jorgensen83,Jorgensen98b,Clancy94,Mahoney01} Two important |
67 |
|
|
properties to quantify are the Gibbs and Helmholtz free energies, |
68 |
|
|
particularly for the solid forms of water as these predict the |
69 |
|
|
thermodynamic stability of the various phases. Water has a |
70 |
|
|
particularly rich phase diagram and takes on a number of different and |
71 |
|
|
stable crystalline structures as the temperature and pressure are |
72 |
|
|
varied. It is a challenging task to investigate the entire free |
73 |
|
|
energy landscape\cite{Sanz04}; and ideally, research is focused on the |
74 |
chrisfen |
1471 |
phases having the lowest free energy at a given state point, because |
75 |
gezelter |
1475 |
these phases will dictate the relevant transition temperatures and |
76 |
chrisfen |
1905 |
pressures for the model. |
77 |
chrisfen |
1459 |
|
78 |
chrisfen |
1905 |
The high-pressure phases of water (ice II - ice X as well as ice XII) |
79 |
|
|
have been studied extensively both experimentally and |
80 |
|
|
computationally. In this paper, standard reference state methods were |
81 |
|
|
applied in the {\it low} pressure regime to evaluate the free energies |
82 |
|
|
for a few known crystalline water polymorphs that might be stable at |
83 |
|
|
these pressures. This work is unique in that one of the crystal |
84 |
|
|
lattices was arrived at through crystallization of a computationally |
85 |
|
|
efficient water model under constant pressure and temperature |
86 |
|
|
conditions. Crystallization events are interesting in and of |
87 |
|
|
themselves\cite{Matsumoto02,Yamada02}; however, the crystal structure |
88 |
|
|
obtained in this case is different from any previously observed ice |
89 |
|
|
polymorphs in experiment or simulation.\cite{Fennell04} We have named |
90 |
|
|
this structure Ice-{\it i} to indicate its origin in computational |
91 |
|
|
simulation. The unit cell of Ice-{\it i} and an axially-elongated |
92 |
|
|
variant named Ice-{\it i}$^\prime$ both consist of eight water |
93 |
|
|
molecules that stack in rows of interlocking water tetramers as |
94 |
|
|
illustrated in figures \ref{unitcell}A and \ref{unitcell}B. These |
95 |
|
|
tetramers form a crystal structure similar in appearance to a recent |
96 |
|
|
two-dimensional surface tessellation simulated on silica.\cite{Yang04} |
97 |
|
|
As expected in an ice crystal constructed of water tetramers, the |
98 |
|
|
hydrogen bonds are not as linear as those observed in ice $I_h$, |
99 |
|
|
however the interlocking of these subunits appears to provide |
100 |
|
|
significant stabilization to the overall crystal. The arrangement of |
101 |
|
|
these tetramers results in surrounding open octagonal cavities that |
102 |
|
|
are typically greater than 6.3 \AA\ in diameter |
103 |
|
|
(Fig. \ref{iCrystal}). This open structure leads to crystals that |
104 |
|
|
are typically 0.07 g/cm$^3$ less dense than ice $I_h$. |
105 |
gezelter |
1463 |
|
106 |
chrisfen |
1460 |
\begin{figure} |
107 |
chrisfen |
1905 |
\centering |
108 |
gezelter |
1463 |
\includegraphics[width=\linewidth]{unitCell.eps} |
109 |
chrisfen |
1905 |
\caption{(A) Unit cells for Ice-{\it i} and (B) Ice-{\it i}$^\prime$. |
110 |
|
|
The spheres represent the center-of-mass locations of the water |
111 |
|
|
molecules. The $a$ to $c$ ratios for Ice-{\it i} and Ice-{\it |
112 |
|
|
i}$^\prime$ are given by 2.1214 and 1.785 respectively.} |
113 |
|
|
\label{unitcell} |
114 |
chrisfen |
1460 |
\end{figure} |
115 |
gezelter |
1463 |
|
116 |
chrisfen |
1460 |
\begin{figure} |
117 |
chrisfen |
1905 |
\centering |
118 |
gezelter |
1463 |
\includegraphics[width=\linewidth]{orderedIcei.eps} |
119 |
chrisfen |
1905 |
\caption{A rendering of a proton ordered crystal of Ice-{\it i} looking |
120 |
|
|
down the (001) crystal face. The presence of large octagonal pores |
121 |
|
|
leads to a polymorph that is less dense than ice $I_h$.} |
122 |
|
|
\label{iCrystal} |
123 |
chrisfen |
1460 |
\end{figure} |
124 |
chrisfen |
1459 |
|
125 |
gezelter |
1465 |
Results from our previous study indicated that Ice-{\it i} is the |
126 |
chrisfen |
1905 |
minimum energy crystal structure for the single point water models |
127 |
|
|
investigated (for discussions on these single point dipole models, see |
128 |
|
|
our previous work and related |
129 |
|
|
articles).\cite{Fennell04,Liu96,Bratko85} Our earlier results |
130 |
|
|
considered only energetic stabilization and neglected entropic |
131 |
chrisfen |
1812 |
contributions to the overall free energy. To address this issue, we |
132 |
gezelter |
1475 |
have calculated the absolute free energy of this crystal using |
133 |
chrisfen |
1905 |
thermodynamic integration and compared it to the free energies of ice |
134 |
|
|
$I_c$ and ice $I_h$ (the common low density ice polymorphs) and ice B |
135 |
|
|
(a higher density, but very stable crystal structure observed by |
136 |
|
|
B\`{a}ez and Clancy in free energy studies of SPC/E).\cite{Baez95b} |
137 |
|
|
This work includes results for the water model from which Ice-{\it i} |
138 |
|
|
was crystallized (SSD/E) in addition to several common water models |
139 |
|
|
(TIP3P, TIP4P, TIP5P, and SPC/E) and a reaction field parametrized |
140 |
|
|
single point dipole water model (SSD/RF). The axially-elongated |
141 |
|
|
variant, Ice-{\it i}$^\prime$, was used in calculations involving |
142 |
|
|
SPC/E, TIP4P, and TIP5P. The square tetramers in Ice-{\it i} distort |
143 |
|
|
in Ice-{\it i}$^\prime$ to form a rhombus with alternating 85 and 95 |
144 |
|
|
degree angles. Under SPC/E, TIP4P, and TIP5P, this geometry is better |
145 |
|
|
at forming favorable hydrogen bonds. The degree of rhomboid |
146 |
|
|
distortion depends on the water model used, but is significant enough |
147 |
|
|
to split a peak in the radial distribution function which corresponds |
148 |
|
|
to diagonal sites in the tetramers. |
149 |
chrisfen |
1459 |
|
150 |
chrisfen |
1453 |
\section{Methods} |
151 |
|
|
|
152 |
chrisfen |
1454 |
Canonical ensemble (NVT) molecular dynamics calculations were |
153 |
chrisfen |
1905 |
performed using the OOPSE molecular mechanics program.\cite{Meineke05} |
154 |
gezelter |
1465 |
All molecules were treated as rigid bodies, with orientational motion |
155 |
chrisfen |
1812 |
propagated using the symplectic DLM integration method. Details about |
156 |
chrisfen |
1471 |
the implementation of this technique can be found in a recent |
157 |
gezelter |
1468 |
publication.\cite{Dullweber1997} |
158 |
chrisfen |
1454 |
|
159 |
chrisfen |
1905 |
Thermodynamic integration was utilized to calculate the Helmholtz free |
160 |
|
|
energies ($A$) of the listed water models at various state points |
161 |
|
|
using the OOPSE molecular dynamics program.\cite{Meineke05} |
162 |
|
|
Thermodynamic integration is an established technique that has been |
163 |
|
|
used extensively in the calculation of free energies for condensed |
164 |
|
|
phases of |
165 |
|
|
materials.\cite{Frenkel84,Hermens88,Meijer90,Baez95a,Vlot99}. This |
166 |
|
|
method uses a sequence of simulations during which the system of |
167 |
|
|
interest is converted into a reference system for which the free |
168 |
|
|
energy is known analytically ($A_0$). The difference in potential |
169 |
|
|
energy between the reference system and the system of interest |
170 |
|
|
($\Delta V$) is then integrated in order to determine the free energy |
171 |
|
|
difference between the two states: |
172 |
chrisfen |
1458 |
\begin{equation} |
173 |
chrisfen |
1905 |
A = A_0 + \int_0^1 \left\langle \Delta V \right\rangle_\lambda d\lambda. |
174 |
chrisfen |
1458 |
\end{equation} |
175 |
chrisfen |
1905 |
Here, $\lambda$ is the parameter that governs the transformation |
176 |
|
|
between the reference system and the system of interest. For |
177 |
|
|
crystalline phases, an harmonically-restrained (Einsten) crystal is |
178 |
|
|
chosen as the reference state, while for liquid phases, the ideal gas |
179 |
|
|
is taken as the reference state. |
180 |
chrisfen |
1458 |
|
181 |
chrisfen |
1905 |
In an Einstein crystal, the molecules are restrained at their ideal |
182 |
|
|
lattice locations and orientations. Using harmonic restraints, as |
183 |
|
|
applied by B\`{a}ez and Clancy, the total potential for this reference |
184 |
|
|
crystal ($V_\mathrm{EC}$) is the sum of all the harmonic restraints, |
185 |
chrisfen |
1471 |
\begin{equation} |
186 |
|
|
V_\mathrm{EC} = \frac{K_\mathrm{v}r^2}{2} + \frac{K_\theta\theta^2}{2} + |
187 |
|
|
\frac{K_\omega\omega^2}{2}, |
188 |
|
|
\end{equation} |
189 |
|
|
where $K_\mathrm{v}$, $K_\mathrm{\theta}$, and $K_\mathrm{\omega}$ are |
190 |
|
|
the spring constants restraining translational motion and deflection |
191 |
|
|
of and rotation around the principle axis of the molecule |
192 |
chrisfen |
1555 |
respectively. These spring constants are typically calculated from |
193 |
|
|
the mean-square displacements of water molecules in an unrestrained |
194 |
|
|
ice crystal at 200 K. For these studies, $K_\mathrm{r} = 4.29$ kcal |
195 |
|
|
mol$^{-1}$, $K_\theta\ = 13.88$ kcal mol$^{-1}$, and $K_\omega\ = |
196 |
|
|
17.75$ kcal mol$^{-1}$. It is clear from Fig. \ref{waterSpring} that |
197 |
|
|
the values of $\theta$ range from $0$ to $\pi$, while $\omega$ ranges |
198 |
|
|
from $-\pi$ to $\pi$. The partition function for a molecular crystal |
199 |
gezelter |
1465 |
restrained in this fashion can be evaluated analytically, and the |
200 |
|
|
Helmholtz Free Energy ({\it A}) is given by |
201 |
chrisfen |
1454 |
\begin{eqnarray} |
202 |
|
|
A = E_m\ -\ kT\ln \left (\frac{kT}{h\nu}\right )^3&-&kT\ln \left |
203 |
|
|
[\pi^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{A}kT}{h^2}\right |
204 |
|
|
)^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{B}kT}{h^2}\right |
205 |
|
|
)^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{C}kT}{h^2}\right |
206 |
|
|
)^\frac{1}{2}\right ] \nonumber \\ &-& kT\ln \left [\frac{kT}{2(\pi |
207 |
|
|
K_\omega K_\theta)^{\frac{1}{2}}}\exp\left |
208 |
|
|
(-\frac{kT}{2K_\theta}\right)\int_0^{\left (\frac{kT}{2K_\theta}\right |
209 |
|
|
)^\frac{1}{2}}\exp(t^2)\mathrm{d}t\right ], |
210 |
|
|
\label{ecFreeEnergy} |
211 |
|
|
\end{eqnarray} |
212 |
chrisfen |
1471 |
where $2\pi\nu = (K_\mathrm{v}/m)^{1/2}$, and $E_m$ is the minimum |
213 |
|
|
potential energy of the ideal crystal.\cite{Baez95a} |
214 |
gezelter |
1463 |
|
215 |
chrisfen |
1456 |
\begin{figure} |
216 |
chrisfen |
1905 |
\centering |
217 |
|
|
\includegraphics[width=4in]{rotSpring.eps} |
218 |
chrisfen |
1456 |
\caption{Possible orientational motions for a restrained molecule. |
219 |
|
|
$\theta$ angles correspond to displacement from the body-frame {\it |
220 |
|
|
z}-axis, while $\omega$ angles correspond to rotation about the |
221 |
chrisfen |
1814 |
body-frame {\it z}-axis. $K_\theta$ and $K_\omega$ are spring |
222 |
chrisfen |
1456 |
constants for the harmonic springs restraining motion in the $\theta$ |
223 |
|
|
and $\omega$ directions.} |
224 |
|
|
\label{waterSpring} |
225 |
|
|
\end{figure} |
226 |
chrisfen |
1454 |
|
227 |
chrisfen |
1471 |
In the case of molecular liquids, the ideal vapor is chosen as the |
228 |
|
|
target reference state. There are several examples of liquid state |
229 |
|
|
free energy calculations of water models present in the |
230 |
|
|
literature.\cite{Hermens88,Quintana92,Mezei92,Baez95b} These methods |
231 |
|
|
typically differ in regard to the path taken for switching off the |
232 |
|
|
interaction potential to convert the system to an ideal gas of water |
233 |
chrisfen |
1905 |
molecules. In this study, we applied one of the most convenient |
234 |
gezelter |
1475 |
methods and integrated over the $\lambda^4$ path, where all |
235 |
|
|
interaction parameters are scaled equally by this transformation |
236 |
|
|
parameter. This method has been shown to be reversible and provide |
237 |
|
|
results in excellent agreement with other established |
238 |
|
|
methods.\cite{Baez95b} |
239 |
chrisfen |
1471 |
|
240 |
chrisfen |
1905 |
Near the cutoff radius ($0.85 * r_{cut}$), charge, dipole, and |
241 |
|
|
Lennard-Jones interactions were gradually reduced by a cubic switching |
242 |
|
|
function. By applying this function, these interactions are smoothly |
243 |
|
|
truncated, thereby avoiding the poor energy conservation which results |
244 |
|
|
from harsher truncation schemes. The effect of a long-range |
245 |
|
|
correction was also investigated on select model systems in a variety |
246 |
|
|
of manners. For the SSD/RF model, a reaction field with a fixed |
247 |
|
|
dielectric constant of 80 was applied in all |
248 |
|
|
simulations.\cite{Onsager36} For a series of the least computationally |
249 |
|
|
expensive models (SSD/E, SSD/RF, TIP3P, and SPC/E), simulations were |
250 |
|
|
performed with longer cutoffs of 10.5, 12, 13.5, and 15 \AA\ to |
251 |
|
|
compare with the 9 \AA\ cutoff results. Finally, the effects of using |
252 |
|
|
the Ewald summation were estimated for TIP3P and SPC/E by performing |
253 |
|
|
single configuration Particle-Mesh Ewald (PME) |
254 |
|
|
calculations~\cite{Tinker} for each of the ice polymorphs. The |
255 |
|
|
calculated energy difference in the presence and absence of PME was |
256 |
|
|
applied to the previous results in order to predict changes to the |
257 |
|
|
free energy landscape. |
258 |
chrisfen |
1454 |
|
259 |
chrisfen |
1905 |
\section{Results and Discussion} |
260 |
chrisfen |
1454 |
|
261 |
chrisfen |
1905 |
The calculated free energies of proton-ordered variants of three low |
262 |
|
|
density polymorphs ($I_h$, $I_c$, and Ice-{\it i} or Ice-{\it |
263 |
|
|
i}$^\prime$) and the stable higher density ice B are listed in Table |
264 |
|
|
\ref{freeEnergy}. Ice B was included because it has been |
265 |
|
|
shown to be a minimum free energy structure for SPC/E at ambient |
266 |
|
|
conditions.\cite{Baez95b} In addition to the free energies, the |
267 |
|
|
relevant transition temperatures at standard pressure are also |
268 |
|
|
displayed in Table \ref{freeEnergy}. These free energy values |
269 |
|
|
indicate that Ice-{\it i} is the most stable state for all of the |
270 |
|
|
investigated water models. With the free energy at these state |
271 |
|
|
points, the Gibbs-Helmholtz equation was used to project to other |
272 |
|
|
state points and to build phase diagrams. Figure \ref{tp3PhaseDia} is |
273 |
|
|
an example diagram built from the results for the TIP3P water model. |
274 |
|
|
All other models have similar structure, although the crossing points |
275 |
|
|
between the phases move to different temperatures and pressures as |
276 |
|
|
indicated from the transition temperatures in Table \ref{freeEnergy}. |
277 |
|
|
It is interesting to note that ice $I_h$ (and ice $I_c$ for that |
278 |
|
|
matter) do not appear in any of the phase diagrams for any of the |
279 |
|
|
models. For purposes of this study, ice B is representative of the |
280 |
|
|
dense ice polymorphs. A recent study by Sanz {\it et al.} provides |
281 |
|
|
details on the phase diagrams for SPC/E and TIP4P at higher pressures |
282 |
|
|
than those studied here.\cite{Sanz04} |
283 |
chrisfen |
1454 |
|
284 |
chrisfen |
1456 |
\begin{table*} |
285 |
|
|
\begin{minipage}{\linewidth} |
286 |
|
|
\begin{center} |
287 |
chrisfen |
1905 |
\caption{Calculated free energies for several ice polymorphs along |
288 |
|
|
with the calculated melting (or sublimation) and boiling points for |
289 |
|
|
the investigated water models. All free energy calculations used a |
290 |
|
|
cutoff radius of 9.0 \AA\ and were performed at 200 K and $\sim$1 atm. |
291 |
|
|
Units of free energy are kcal/mol, while transition temperature are in |
292 |
|
|
Kelvin. Calculated error of the final digits is in parentheses.} |
293 |
|
|
\begin{tabular}{lccccccc} |
294 |
gezelter |
1463 |
\hline |
295 |
chrisfen |
1905 |
Water Model & $I_h$ & $I_c$ & B & Ice-{\it i} & Ice-{\it i}$^\prime$ & $T_m$ (*$T_s$) & $T_b$\\ |
296 |
gezelter |
1463 |
\hline |
297 |
chrisfen |
1905 |
TIP3P & -11.41(2) & -11.23(3) & -11.82(3) & -12.30(3) & - & 269(4) & 357(2)\\ |
298 |
|
|
TIP4P & -11.84(3) & -12.04(2) & -12.08(3) & - & -12.33(3) & 266(5) & 354(2)\\ |
299 |
|
|
TIP5P & -11.85(3) & -11.86(2) & -11.96(2) & - & -12.29(2) & 271(4) & 337(2)\\ |
300 |
|
|
SPC/E & -12.87(2) & -13.05(2) & -13.26(3) & - & -13.55(2) & 296(3) & 396(2)\\ |
301 |
|
|
SSD/E & -11.27(2) & -11.19(4) & -12.09(2) & -12.54(2) & - & *355(2) & -\\ |
302 |
|
|
SSD/RF & -11.51(2) & -11.47(2) & -12.08(3) & -12.29(2) & - & 278(4) & 349(2)\\ |
303 |
chrisfen |
1456 |
\end{tabular} |
304 |
|
|
\label{freeEnergy} |
305 |
|
|
\end{center} |
306 |
|
|
\end{minipage} |
307 |
|
|
\end{table*} |
308 |
chrisfen |
1453 |
|
309 |
chrisfen |
1456 |
\begin{figure} |
310 |
|
|
\includegraphics[width=\linewidth]{tp3PhaseDia.eps} |
311 |
|
|
\caption{Phase diagram for the TIP3P water model in the low pressure |
312 |
chrisfen |
1812 |
regime. The displayed $T_m$ and $T_b$ values are good predictions of |
313 |
chrisfen |
1456 |
the experimental values; however, the solid phases shown are not the |
314 |
chrisfen |
1812 |
experimentally observed forms. Both cubic and hexagonal ice $I$ are |
315 |
chrisfen |
1456 |
higher in energy and don't appear in the phase diagram.} |
316 |
chrisfen |
1905 |
\label{tp3PhaseDia} |
317 |
chrisfen |
1456 |
\end{figure} |
318 |
gezelter |
1463 |
|
319 |
chrisfen |
1905 |
Most of the water models have melting points that compare quite |
320 |
|
|
favorably with the experimental value of 273 K. The unfortunate |
321 |
|
|
aspect of this result is that this phase change occurs between |
322 |
|
|
Ice-{\it i} and the liquid state rather than ice $I_h$ and the liquid |
323 |
|
|
state. These results do not contradict other studies. Studies of ice |
324 |
|
|
$I_h$ using TIP4P predict a $T_m$ ranging from 214 to 238 K |
325 |
|
|
(differences being attributed to choice of interaction truncation and |
326 |
|
|
different ordered and disordered molecular |
327 |
chrisfen |
1466 |
arrangements).\cite{Vlot99,Gao00,Sanz04} If the presence of ice B and |
328 |
|
|
Ice-{\it i} were omitted, a $T_m$ value around 210 K would be |
329 |
chrisfen |
1812 |
predicted from this work. However, the $T_m$ from Ice-{\it i} is |
330 |
chrisfen |
1806 |
calculated to be 265 K, indicating that these simulation based |
331 |
|
|
structures ought to be included in studies probing phase transitions |
332 |
chrisfen |
1812 |
with this model. Also of interest in these results is that SSD/E does |
333 |
chrisfen |
1905 |
not exhibit a melting point at 1 atm but does sublime at 355 K. This |
334 |
|
|
is due to the significant stability of Ice-{\it i} over all other |
335 |
|
|
polymorphs for this particular model under these conditions. While |
336 |
|
|
troubling, this behavior resulted in the spontaneous crystallization |
337 |
|
|
of Ice-{\it i} which led us to investigate this structure. These |
338 |
|
|
observations provide a warning that simulations of SSD/E as a |
339 |
|
|
``liquid'' near 300 K are actually metastable and run the risk of |
340 |
|
|
spontaneous crystallization. However, when a longer cutoff radius is |
341 |
|
|
used, SSD/E prefers the liquid state under standard temperature and |
342 |
|
|
pressure. |
343 |
chrisfen |
1456 |
|
344 |
chrisfen |
1458 |
\begin{figure} |
345 |
|
|
\includegraphics[width=\linewidth]{cutoffChange.eps} |
346 |
chrisfen |
1806 |
\caption{Free energy as a function of cutoff radius for SSD/E, TIP3P, |
347 |
|
|
SPC/E, SSD/RF with a reaction field, and the TIP3P and SPC/E models |
348 |
chrisfen |
1905 |
with an added Ewald correction term. Error for the larger cutoff |
349 |
|
|
points is equivalent to that observed at 9.0\AA\ (see Table |
350 |
|
|
\ref{freeEnergy}). Data for ice I$_c$ with TIP3P using both 12 and |
351 |
|
|
13.5 \AA\ cutoffs were omitted because the crystal was prone to |
352 |
chrisfen |
1834 |
distortion and melting at 200 K. Ice-{\it i}$^\prime$ is the form of |
353 |
|
|
Ice-{\it i} used in the SPC/E simulations.} |
354 |
chrisfen |
1458 |
\label{incCutoff} |
355 |
|
|
\end{figure} |
356 |
|
|
|
357 |
chrisfen |
1905 |
For the more computationally efficient water models, we have also |
358 |
|
|
investigated the effect of potential trunctaion on the computed free |
359 |
|
|
energies as a function of the cutoff radius. As seen in |
360 |
|
|
Fig. \ref{incCutoff}, the free energies of the ice polymorphs with |
361 |
|
|
water models lacking a long-range correction show significant cutoff |
362 |
|
|
dependence. In general, there is a narrowing of the free energy |
363 |
|
|
differences while moving to greater cutoff radii. As the free |
364 |
|
|
energies for the polymorphs converge, the stability advantage that |
365 |
|
|
Ice-{\it i} exhibits is reduced. Adjacent to each of these plots are |
366 |
|
|
results for systems with applied or estimated long-range corrections. |
367 |
|
|
SSD/RF was parametrized for use with a reaction field, and the benefit |
368 |
|
|
provided by this computationally inexpensive correction is apparent. |
369 |
|
|
The free energies are largely independent of the size of the reaction |
370 |
|
|
field cavity in this model, so small cutoff radii mimic bulk |
371 |
|
|
calculations quite well under SSD/RF. |
372 |
|
|
|
373 |
|
|
Although TIP3P was paramaterized for use without the Ewald summation, |
374 |
|
|
we have estimated the effect of this method for computing long-range |
375 |
|
|
electrostatics for both TIP3P and SPC/E. This was accomplished by |
376 |
|
|
calculating the potential energy of identical crystals both with and |
377 |
|
|
without particle mesh Ewald (PME). Similar behavior to that observed |
378 |
|
|
with reaction field is seen for both of these models. The free |
379 |
|
|
energies show reduced dependence on cutoff radius and span a narrower |
380 |
|
|
range for the various polymorphs. Like the dipolar water models, |
381 |
|
|
TIP3P displays a relatively constant preference for the Ice-{\it i} |
382 |
chrisfen |
1812 |
polymorph. Crystal preference is much more difficult to determine for |
383 |
|
|
SPC/E. Without a long-range correction, each of the polymorphs |
384 |
|
|
studied assumes the role of the preferred polymorph under different |
385 |
chrisfen |
1905 |
cutoff radii. The inclusion of the Ewald correction flattens and |
386 |
|
|
narrows the gap in free energies such that the polymorphs are |
387 |
|
|
isoenergetic within statistical uncertainty. This suggests that other |
388 |
|
|
conditions, such as the density in fixed-volume simulations, can |
389 |
|
|
influence the polymorph expressed upon crystallization. |
390 |
chrisfen |
1456 |
|
391 |
chrisfen |
1905 |
So what is the preferred solid polymorph for simulated water? The |
392 |
|
|
answer appears to be dependent both on the conditions and the model |
393 |
|
|
used. In the case of short cutoffs without a long-range interaction |
394 |
|
|
correction, Ice-{\it i} and Ice-{\it i}$^\prime$ have the lowest free |
395 |
|
|
energy of the studied polymorphs with all the models. Ideally, |
396 |
|
|
crystallization of each model under constant pressure conditions, as |
397 |
|
|
was done with SSD/E, would aid in the identification of their |
398 |
|
|
respective preferred structures. This work, however, helps illustrate |
399 |
|
|
how studies involving one specific model can lead to insight about |
400 |
|
|
important behavior of others. In general, the above results support |
401 |
|
|
the finding that the Ice-{\it i} polymorph is a stable crystal |
402 |
|
|
structure that should be considered when studying the phase behavior |
403 |
|
|
of water models. |
404 |
chrisfen |
1453 |
|
405 |
chrisfen |
1905 |
We also note that none of the water models used in this study are |
406 |
|
|
polarizable or flexible models. It is entirely possible that the |
407 |
|
|
polarizability of real water makes Ice-{\it i} substantially less |
408 |
|
|
stable than ice $I_h$. However, the calculations presented above seem |
409 |
|
|
interesting enough to communicate before the role of polarizability |
410 |
|
|
(or flexibility) has been thoroughly investigated. |
411 |
chrisfen |
1458 |
|
412 |
chrisfen |
1905 |
Finally, due to the stability of Ice-{\it i} in the investigated |
413 |
|
|
simulation conditions, the question arises as to possible experimental |
414 |
|
|
observation of this polymorph. The rather extensive past and current |
415 |
|
|
experimental investigation of water in the low pressure regime makes |
416 |
|
|
us hesitant to ascribe any relevance to this work outside of the |
417 |
|
|
simulation community. It is for this reason that we chose a name for |
418 |
|
|
this polymorph which involves an imaginary quantity. That said, there |
419 |
|
|
are certain experimental conditions that would provide the most ideal |
420 |
|
|
situation for possible observation. These include the negative |
421 |
|
|
pressure or stretched solid regime, small clusters in vacuum |
422 |
gezelter |
1465 |
deposition environments, and in clathrate structures involving small |
423 |
chrisfen |
1905 |
non-polar molecules. For experimental comparison purposes, example |
424 |
|
|
$g_{OO}(r)$ and $S(\vec{q})$ plots were generated for the two Ice-{\it |
425 |
|
|
i} variants (along with example ice $I_h$ and $I_c$ plots) at 77K, and |
426 |
|
|
they are shown in figures \ref{fig:gofr} and \ref{fig:sofq} |
427 |
|
|
respectively. |
428 |
chrisfen |
1459 |
|
429 |
chrisfen |
1467 |
\begin{figure} |
430 |
chrisfen |
1905 |
\centering |
431 |
chrisfen |
1467 |
\includegraphics[width=\linewidth]{iceGofr.eps} |
432 |
chrisfen |
1905 |
\caption{Radial distribution functions of ice $I_h$, $I_c$, and |
433 |
|
|
Ice-{\it i} calculated from from simulations of the SSD/RF water model |
434 |
|
|
at 77 K. The Ice-{\it i} distribution function was obtained from |
435 |
|
|
simulations composed of TIP4P water.} |
436 |
chrisfen |
1467 |
\label{fig:gofr} |
437 |
|
|
\end{figure} |
438 |
|
|
|
439 |
gezelter |
1469 |
\begin{figure} |
440 |
chrisfen |
1905 |
\centering |
441 |
gezelter |
1469 |
\includegraphics[width=\linewidth]{sofq.eps} |
442 |
chrisfen |
1479 |
\caption{Predicted structure factors for ice $I_h$, $I_c$, Ice-{\it i}, |
443 |
|
|
and Ice-{\it i}$^\prime$ at 77 K. The raw structure factors have |
444 |
|
|
been convoluted with a gaussian instrument function (0.075 \AA$^{-1}$ |
445 |
|
|
width) to compensate for the trunction effects in our finite size |
446 |
|
|
simulations.} |
447 |
gezelter |
1469 |
\label{fig:sofq} |
448 |
|
|
\end{figure} |
449 |
|
|
|
450 |
chrisfen |
1453 |
\section{Acknowledgments} |
451 |
|
|
Support for this project was provided by the National Science |
452 |
|
|
Foundation under grant CHE-0134881. Computation time was provided by |
453 |
chrisfen |
1458 |
the Notre Dame High Performance Computing Cluster and the Notre Dame |
454 |
|
|
Bunch-of-Boxes (B.o.B) computer cluster (NSF grant DMR-0079647). |
455 |
chrisfen |
1453 |
|
456 |
|
|
\newpage |
457 |
|
|
|
458 |
|
|
\bibliographystyle{jcp} |
459 |
|
|
\bibliography{iceiPaper} |
460 |
|
|
|
461 |
|
|
|
462 |
|
|
\end{document} |