1 |
chrisfen |
1459 |
|
2 |
chrisfen |
1453 |
%\documentclass[prb,aps,twocolumn,tabularx]{revtex4} |
3 |
|
|
\documentclass[preprint,aps,endfloats]{revtex4} |
4 |
|
|
%\documentclass[11pt]{article} |
5 |
|
|
%\usepackage{endfloat} |
6 |
|
|
\usepackage{amsmath} |
7 |
|
|
\usepackage{epsf} |
8 |
|
|
\usepackage{berkeley} |
9 |
|
|
%\usepackage{setspace} |
10 |
|
|
%\usepackage{tabularx} |
11 |
|
|
\usepackage{graphicx} |
12 |
|
|
%\usepackage[ref]{overcite} |
13 |
|
|
%\pagestyle{plain} |
14 |
|
|
%\pagenumbering{arabic} |
15 |
|
|
%\oddsidemargin 0.0cm \evensidemargin 0.0cm |
16 |
|
|
%\topmargin -21pt \headsep 10pt |
17 |
|
|
%\textheight 9.0in \textwidth 6.5in |
18 |
|
|
%\brokenpenalty=10000 |
19 |
|
|
|
20 |
|
|
%\renewcommand\citemid{\ } % no comma in optional reference note |
21 |
|
|
|
22 |
|
|
\begin{document} |
23 |
|
|
|
24 |
chrisfen |
1459 |
\title{A Free Energy Study of Low Temperature and Anomalous Ice} |
25 |
chrisfen |
1453 |
|
26 |
|
|
\author{Christopher J. Fennell and J. Daniel Gezelter{\thefootnote} |
27 |
|
|
\footnote[1]{Corresponding author. \ Electronic mail: gezelter@nd.edu}} |
28 |
|
|
|
29 |
|
|
\address{Department of Chemistry and Biochemistry\\ University of Notre Dame\\ |
30 |
|
|
Notre Dame, Indiana 46556} |
31 |
|
|
|
32 |
|
|
\date{\today} |
33 |
|
|
|
34 |
|
|
%\maketitle |
35 |
|
|
%\doublespacing |
36 |
|
|
|
37 |
|
|
\begin{abstract} |
38 |
chrisfen |
1459 |
The free energies of several ice polymorphs in the low pressure regime |
39 |
|
|
were calculated using thermodynamic integration of systems consisting |
40 |
|
|
of a variety of common water models. Ice-{\it i}, a recent |
41 |
|
|
computationally observed solid structure, was determined to be the |
42 |
|
|
stable state with the lowest free energy for all the water models |
43 |
|
|
investigated. Phase diagrams were generated, and melting and boiling |
44 |
|
|
points for all the models were determined and show relatively good |
45 |
|
|
agreement with experiment, although the solid phase is different |
46 |
|
|
between simulation and experiment. In addition, potential truncation |
47 |
|
|
was shown to have an effect on the calculated free energies, and may |
48 |
|
|
result in altered free energy landscapes. |
49 |
chrisfen |
1453 |
\end{abstract} |
50 |
|
|
|
51 |
|
|
\maketitle |
52 |
|
|
|
53 |
|
|
\newpage |
54 |
|
|
|
55 |
|
|
%\narrowtext |
56 |
|
|
|
57 |
|
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
58 |
|
|
% BODY OF TEXT |
59 |
|
|
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
60 |
|
|
|
61 |
|
|
\section{Introduction} |
62 |
|
|
|
63 |
chrisfen |
1459 |
Molecular dynamics has developed into a valuable tool for studying the |
64 |
|
|
phase behavior of systems ranging from small or simple |
65 |
chrisfen |
1462 |
molecules\cite{Matsumoto02andOthers} to complex biological |
66 |
chrisfen |
1459 |
species.\cite{bigStuff} Many techniques have been developed in order |
67 |
|
|
to investigate the thermodynamic properites of model substances, |
68 |
|
|
providing both qualitative and quantitative comparisons between |
69 |
|
|
simulations and experiment.\cite{thermMethods} Investigation of these |
70 |
|
|
properties leads to the development of new and more accurate models, |
71 |
|
|
leading to better understanding and depiction of physical processes |
72 |
|
|
and intricate molecular systems. |
73 |
|
|
|
74 |
|
|
Water has proven to be a challenging substance to depict in |
75 |
|
|
simulations, and has resulted in a variety of models that attempt to |
76 |
|
|
describe its behavior under a varying simulation |
77 |
chrisfen |
1462 |
conditions.\cite{Berendsen81,Jorgensen83,Bratko85,Berendsen87,Liu96,Mahoney00,Fennell04} |
78 |
|
|
Many of these models have been used to investigate important physical |
79 |
|
|
phenomena like phase transitions and the hydrophobic |
80 |
|
|
effect.\cite{evenMorePapers} With the advent of numerous differing |
81 |
|
|
models, it is only natural that attention is placed on the properties |
82 |
|
|
of the models themselves in an attempt to clarify their benefits and |
83 |
|
|
limitations when applied to a system of interest.\cite{modelProps} One |
84 |
|
|
important but challenging property to quantify is the free energy, |
85 |
|
|
particularly of the solid forms of water. Difficulty in these types of |
86 |
|
|
studies typically arises from the assortment of possible crystalline |
87 |
|
|
polymorphs that water that water adopts over a wide range of pressures |
88 |
|
|
and temperatures. There are currently 13 recognized forms of ice, and |
89 |
|
|
it is a challenging task to investigate the entire free energy |
90 |
|
|
landscape.\cite{Sanz04} Ideally, research is focused on the phases |
91 |
|
|
having the lowest free energy, because these phases will dictate the |
92 |
|
|
true transition temperatures and pressures for their respective model. |
93 |
chrisfen |
1459 |
|
94 |
|
|
In this paper, standard reference state methods were applied to the |
95 |
|
|
study of crystalline water polymorphs in the low pressure regime. This |
96 |
|
|
work is unique in the fact that one of the crystal lattices was |
97 |
|
|
arrived at through crystallization of a computationally efficient |
98 |
|
|
water model under constant pressure and temperature |
99 |
|
|
conditions. Crystallization events are interesting in and of |
100 |
chrisfen |
1462 |
themselves\cite{Matsumoto02,Yamada02}; however, the crystal structure |
101 |
chrisfen |
1459 |
obtained in this case was different from any previously observed ice |
102 |
|
|
polymorphs, in experiment or simulation.\cite{Fennell04} This crystal |
103 |
|
|
was termed Ice-{\it i} in homage to its origin in computational |
104 |
|
|
simulation. The unit cell (Fig. \ref{iceiCell}A) consists of eight |
105 |
|
|
water molecules that stack in rows of interlocking water |
106 |
|
|
tetramers. Proton ordering can be accomplished by orienting two of the |
107 |
|
|
waters so that both of their donated hydrogen bonds are internal to |
108 |
|
|
their tetramer (Fig. \ref{protOrder}). As expected in an ice crystal |
109 |
|
|
constructed of water tetramers, the hydrogen bonds are not as linear |
110 |
|
|
as those observed in ice $I_h$, however the interlocking of these |
111 |
|
|
subunits appears to provide significant stabilization to the overall |
112 |
|
|
crystal. The arrangement of these tetramers results in surrounding |
113 |
|
|
open octagonal cavities that are typically greater than 6.3 \AA\ in |
114 |
|
|
diameter. This relatively open overall structure leads to crystals |
115 |
|
|
that are 0.07 g/cm$^3$ less dense on average than ice $I_h$. |
116 |
chrisfen |
1460 |
\begin{figure} |
117 |
|
|
\includegraphics[scale=1.0]{unitCell.eps} |
118 |
chrisfen |
1461 |
\caption{Unit cells for (A) Ice-{\it i} and (B) Ice-2{\it i}, the elongated variant of Ice-{\it i}. For Ice-{\it i}, the $a$ to $c$ relation is given by $a = 2.1214c$, while for Ice-2{\it i}, $a = 1.7850c$.} |
119 |
chrisfen |
1460 |
\label{iceiCell} |
120 |
|
|
\end{figure} |
121 |
|
|
\begin{figure} |
122 |
|
|
\includegraphics[scale=1.0]{orderedIcei.eps} |
123 |
|
|
\caption{Image of a proton ordered crystal of Ice-{\it i} looking |
124 |
|
|
down the (001) crystal face. The rows of water tetramers surrounded by |
125 |
|
|
octagonal pores leads to a crystal structure that is significantly |
126 |
|
|
less dense than ice $I_h$.} |
127 |
|
|
\label{protOrder} |
128 |
|
|
\end{figure} |
129 |
chrisfen |
1459 |
|
130 |
|
|
Results in the previous study indicated that Ice-{\it i} is the |
131 |
|
|
minimum energy crystal structure for the single point water models |
132 |
|
|
being studied (for discussions on these single point dipole models, |
133 |
|
|
see the previous work and related |
134 |
|
|
articles\cite{Fennell04,Ichiye96,Bratko85}). Those results only |
135 |
|
|
consider energetic stabilization and neglect entropic contributions to |
136 |
|
|
the overall free energy. To address this issue, the absolute free |
137 |
|
|
energy of this crystal was calculated using thermodynamic integration |
138 |
|
|
and compared to the free energies of cubic and hexagonal ice $I$ (the |
139 |
|
|
experimental low density ice polymorphs) and ice B (a higher density, |
140 |
|
|
but very stable crystal structure observed by B\`{a}ez and Clancy in |
141 |
|
|
free energy studies of SPC/E).\cite{Baez95b} This work includes |
142 |
|
|
results for the water model from which Ice-{\it i} was crystallized |
143 |
|
|
(soft sticky dipole extended, SSD/E) in addition to several common |
144 |
|
|
water models (TIP3P, TIP4P, TIP5P, and SPC/E) and a reaction field |
145 |
|
|
parametrized single point dipole water model (soft sticky dipole |
146 |
|
|
reaction field, SSD/RF). In should be noted that a second version of |
147 |
|
|
Ice-{\it i} (Ice-2{\it i}) was used in calculations involving SPC/E, |
148 |
|
|
TIP4P, and TIP5P. The unit cell of this crystal (Fig. \ref{iceiCell}B) |
149 |
|
|
is similar to the Ice-{\it i} unit it is extended in the direction of |
150 |
|
|
the (001) face and compressed along the other two faces. |
151 |
|
|
|
152 |
chrisfen |
1453 |
\section{Methods} |
153 |
|
|
|
154 |
chrisfen |
1454 |
Canonical ensemble (NVT) molecular dynamics calculations were |
155 |
|
|
performed using the OOPSE (Object-Oriented Parallel Simulation Engine) |
156 |
|
|
molecular mechanics package. All molecules were treated as rigid |
157 |
chrisfen |
1459 |
bodies, with orientational motion propagated using the symplectic DLM |
158 |
chrisfen |
1454 |
integration method. Details about the implementation of these |
159 |
|
|
techniques can be found in a recent publication.\cite{Meineke05} |
160 |
|
|
|
161 |
|
|
Thermodynamic integration was utilized to calculate the free energy of |
162 |
chrisfen |
1456 |
several ice crystals at 200 K using the TIP3P, TIP4P, TIP5P, SPC/E, |
163 |
|
|
SSD/RF, and SSD/E water models. Liquid state free energies at 300 and |
164 |
|
|
400 K for all of these water models were also determined using this |
165 |
|
|
same technique, in order to determine melting points and generate |
166 |
|
|
phase diagrams. All simulations were carried out at densities |
167 |
|
|
resulting in a pressure of approximately 1 atm at their respective |
168 |
|
|
temperatures. |
169 |
chrisfen |
1454 |
|
170 |
chrisfen |
1458 |
A single thermodynamic integration involves a sequence of simulations |
171 |
|
|
over which the system of interest is converted into a reference system |
172 |
|
|
for which the free energy is known. This transformation path is then |
173 |
|
|
integrated in order to determine the free energy difference between |
174 |
|
|
the two states: |
175 |
|
|
\begin{equation} |
176 |
|
|
\Delta A = \int_0^1\left\langle\frac{\partial V(\lambda |
177 |
|
|
)}{\partial\lambda}\right\rangle_\lambda d\lambda, |
178 |
|
|
\end{equation} |
179 |
|
|
where $V$ is the interaction potential and $\lambda$ is the |
180 |
chrisfen |
1459 |
transformation parameter that scales the overall |
181 |
|
|
potential. Simulations are distributed unevenly along this path in |
182 |
|
|
order to sufficiently sample the regions of greatest change in the |
183 |
|
|
potential. Typical integrations in this study consisted of $\sim$25 |
184 |
|
|
simulations ranging from 300 ps (for the unaltered system) to 75 ps |
185 |
|
|
(near the reference state) in length. |
186 |
chrisfen |
1458 |
|
187 |
chrisfen |
1454 |
For the thermodynamic integration of molecular crystals, the Einstein |
188 |
|
|
Crystal is chosen as the reference state that the system is converted |
189 |
|
|
to over the course of the simulation. In an Einstein Crystal, the |
190 |
|
|
molecules are harmonically restrained at their ideal lattice locations |
191 |
|
|
and orientations. The partition function for a molecular crystal |
192 |
|
|
restrained in this fashion has been evaluated, and the Helmholtz Free |
193 |
|
|
Energy ({\it A}) is given by |
194 |
|
|
\begin{eqnarray} |
195 |
|
|
A = E_m\ -\ kT\ln \left (\frac{kT}{h\nu}\right )^3&-&kT\ln \left |
196 |
|
|
[\pi^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{A}kT}{h^2}\right |
197 |
|
|
)^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{B}kT}{h^2}\right |
198 |
|
|
)^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{C}kT}{h^2}\right |
199 |
|
|
)^\frac{1}{2}\right ] \nonumber \\ &-& kT\ln \left [\frac{kT}{2(\pi |
200 |
|
|
K_\omega K_\theta)^{\frac{1}{2}}}\exp\left |
201 |
|
|
(-\frac{kT}{2K_\theta}\right)\int_0^{\left (\frac{kT}{2K_\theta}\right |
202 |
|
|
)^\frac{1}{2}}\exp(t^2)\mathrm{d}t\right ], |
203 |
|
|
\label{ecFreeEnergy} |
204 |
|
|
\end{eqnarray} |
205 |
|
|
where $2\pi\nu = (K_\mathrm{v}/m)^{1/2}$.\cite{Baez95a} In equation |
206 |
|
|
\ref{ecFreeEnergy}, $K_\mathrm{v}$, $K_\mathrm{\theta}$, and |
207 |
|
|
$K_\mathrm{\omega}$ are the spring constants restraining translational |
208 |
|
|
motion and deflection of and rotation around the principle axis of the |
209 |
|
|
molecule respectively (Fig. \ref{waterSpring}), and $E_m$ is the |
210 |
|
|
minimum potential energy of the ideal crystal. In the case of |
211 |
|
|
molecular liquids, the ideal vapor is chosen as the target reference |
212 |
|
|
state. |
213 |
chrisfen |
1456 |
\begin{figure} |
214 |
|
|
\includegraphics[scale=1.0]{rotSpring.eps} |
215 |
|
|
\caption{Possible orientational motions for a restrained molecule. |
216 |
|
|
$\theta$ angles correspond to displacement from the body-frame {\it |
217 |
|
|
z}-axis, while $\omega$ angles correspond to rotation about the |
218 |
|
|
body-frame {\it z}-axis. $K_\theta$ and $K_\omega$ are spring |
219 |
|
|
constants for the harmonic springs restraining motion in the $\theta$ |
220 |
|
|
and $\omega$ directions.} |
221 |
|
|
\label{waterSpring} |
222 |
|
|
\end{figure} |
223 |
chrisfen |
1454 |
|
224 |
chrisfen |
1456 |
Charge, dipole, and Lennard-Jones interactions were modified by a |
225 |
chrisfen |
1462 |
cubic switching between 100\% and 85\% of the cutoff value (9 \AA |
226 |
|
|
). By applying this function, these interactions are smoothly |
227 |
|
|
truncated, thereby avoiding poor energy conserving dynamics resulting |
228 |
|
|
from harsher truncation schemes. The effect of a long-range correction |
229 |
|
|
was also investigated on select model systems in a variety of |
230 |
|
|
manners. For the SSD/RF model, a reaction field with a fixed |
231 |
|
|
dielectric constant of 80 was applied in all |
232 |
|
|
simulations.\cite{Onsager36} For a series of the least computationally |
233 |
|
|
expensive models (SSD/E, SSD/RF, and TIP3P), simulations were |
234 |
|
|
performed with longer cutoffs of 12 and 15 \AA\ to compare with the 9 |
235 |
|
|
\AA\ cutoff results. Finally, results from the use of an Ewald |
236 |
|
|
summation were estimated for TIP3P and SPC/E by performing |
237 |
chrisfen |
1456 |
calculations with Particle-Mesh Ewald (PME) in the TINKER molecular |
238 |
chrisfen |
1462 |
mechanics software package.\cite{Tinker} TINKER was chosen because it |
239 |
|
|
can also propagate the motion of rigid-bodies, and provides the most |
240 |
|
|
direct comparison to the results from OOPSE. The calculated energy |
241 |
|
|
difference in the presence and absence of PME was applied to the |
242 |
|
|
previous results in order to predict changes in the free energy |
243 |
|
|
landscape. |
244 |
chrisfen |
1454 |
|
245 |
chrisfen |
1456 |
\section{Results and discussion} |
246 |
chrisfen |
1454 |
|
247 |
chrisfen |
1456 |
The free energy of proton ordered Ice-{\it i} was calculated and |
248 |
|
|
compared with the free energies of proton ordered variants of the |
249 |
|
|
experimentally observed low-density ice polymorphs, $I_h$ and $I_c$, |
250 |
|
|
as well as the higher density ice B, observed by B\`{a}ez and Clancy |
251 |
|
|
and thought to be the minimum free energy structure for the SPC/E |
252 |
|
|
model at ambient conditions (Table \ref{freeEnergy}).\cite{Baez95b} |
253 |
|
|
Ice XI, the experimentally observed proton ordered variant of ice |
254 |
|
|
$I_h$, was investigated initially, but it was found not to be as |
255 |
|
|
stable as antiferroelectric variants of proton ordered or even proton |
256 |
|
|
disordered ice$I_h$.\cite{Davidson84} The proton ordered variant of |
257 |
|
|
ice $I_h$ used here is a simple antiferroelectric version that has an |
258 |
|
|
8 molecule unit cell. The crystals contained 648 or 1728 molecules for |
259 |
|
|
ice B, 1024 or 1280 molecules for ice $I_h$, 1000 molecules for ice |
260 |
|
|
$I_c$, or 1024 molecules for Ice-{\it i}. The larger crystal sizes |
261 |
|
|
were necessary for simulations involving larger cutoff values. |
262 |
chrisfen |
1454 |
|
263 |
chrisfen |
1456 |
\begin{table*} |
264 |
|
|
\begin{minipage}{\linewidth} |
265 |
|
|
\renewcommand{\thefootnote}{\thempfootnote} |
266 |
|
|
\begin{center} |
267 |
|
|
\caption{Calculated free energies for several ice polymorphs with a |
268 |
|
|
variety of common water models. All calculations used a cutoff radius |
269 |
|
|
of 9 \AA\ and were performed at 200 K and $\sim$1 atm. Units are |
270 |
|
|
kcal/mol. *Ice $I_c$ is unstable at 200 K using SSD/RF.} |
271 |
|
|
\begin{tabular}{ l c c c c } |
272 |
|
|
\hline \\[-7mm] |
273 |
|
|
\ \quad \ Water Model\ \ & \ \quad \ \ \ \ $I_h$ \ \ & \ \quad \ \ \ \ $I_c$ \ \ & \ \quad \ \ \ \ B \ \ & \ \quad \ \ \ Ice-{\it i} \ \quad \ \\ |
274 |
|
|
\hline \\[-3mm] |
275 |
|
|
\ \quad \ TIP3P & \ \quad \ -11.41 & \ \quad \ -11.23 & \ \quad \ -11.82 & \quad -12.30\\ |
276 |
|
|
\ \quad \ TIP4P & \ \quad \ -11.84 & \ \quad \ -12.04 & \ \quad \ -12.08 & \quad -12.33\\ |
277 |
|
|
\ \quad \ TIP5P & \ \quad \ -11.85 & \ \quad \ -11.86 & \ \quad \ -11.96 & \quad -12.29\\ |
278 |
|
|
\ \quad \ SPC/E & \ \quad \ -12.67 & \ \quad \ -12.96 & \ \quad \ -13.25 & \quad -13.55\\ |
279 |
|
|
\ \quad \ SSD/E & \ \quad \ -11.27 & \ \quad \ -11.19 & \ \quad \ -12.09 & \quad -12.54\\ |
280 |
|
|
\ \quad \ SSD/RF & \ \quad \ -11.51 & \ \quad \ NA* & \ \quad \ -12.08 & \quad -12.29\\ |
281 |
|
|
\end{tabular} |
282 |
|
|
\label{freeEnergy} |
283 |
|
|
\end{center} |
284 |
|
|
\end{minipage} |
285 |
|
|
\end{table*} |
286 |
chrisfen |
1453 |
|
287 |
chrisfen |
1456 |
The free energy values computed for the studied polymorphs indicate |
288 |
|
|
that Ice-{\it i} is the most stable state for all of the common water |
289 |
|
|
models studied. With the free energy at these state points, the |
290 |
|
|
temperature and pressure dependence of the free energy was used to |
291 |
|
|
project to other state points and build phase diagrams. Figures |
292 |
|
|
\ref{tp3phasedia} and \ref{ssdrfphasedia} are example diagrams built |
293 |
|
|
from the free energy results. All other models have similar structure, |
294 |
|
|
only the crossing points between these phases exist at different |
295 |
|
|
temperatures and pressures. It is interesting to note that ice $I$ |
296 |
|
|
does not exist in either cubic or hexagonal form in any of the phase |
297 |
|
|
diagrams for any of the models. For purposes of this study, ice B is |
298 |
|
|
representative of the dense ice polymorphs. A recent study by Sanz |
299 |
|
|
{\it et al.} goes into detail on the phase diagrams for SPC/E and |
300 |
|
|
TIP4P in the high pressure regime.\cite{Sanz04} |
301 |
|
|
\begin{figure} |
302 |
|
|
\includegraphics[width=\linewidth]{tp3PhaseDia.eps} |
303 |
|
|
\caption{Phase diagram for the TIP3P water model in the low pressure |
304 |
|
|
regime. The displayed $T_m$ and $T_b$ values are good predictions of |
305 |
|
|
the experimental values; however, the solid phases shown are not the |
306 |
|
|
experimentally observed forms. Both cubic and hexagonal ice $I$ are |
307 |
|
|
higher in energy and don't appear in the phase diagram.} |
308 |
|
|
\label{tp3phasedia} |
309 |
|
|
\end{figure} |
310 |
|
|
\begin{figure} |
311 |
|
|
\includegraphics[width=\linewidth]{ssdrfPhaseDia.eps} |
312 |
|
|
\caption{Phase diagram for the SSD/RF water model in the low pressure |
313 |
|
|
regime. Calculations producing these results were done under an |
314 |
|
|
applied reaction field. It is interesting to note that this |
315 |
|
|
computationally efficient model (over 3 times more efficient than |
316 |
|
|
TIP3P) exhibits phase behavior similar to the less computationally |
317 |
|
|
conservative charge based models.} |
318 |
|
|
\label{ssdrfphasedia} |
319 |
|
|
\end{figure} |
320 |
|
|
|
321 |
|
|
\begin{table*} |
322 |
|
|
\begin{minipage}{\linewidth} |
323 |
|
|
\renewcommand{\thefootnote}{\thempfootnote} |
324 |
|
|
\begin{center} |
325 |
|
|
\caption{Melting ($T_m$), boiling ($T_b$), and sublimation ($T_s$) |
326 |
|
|
temperatures of several common water models compared with experiment.} |
327 |
|
|
\begin{tabular}{ l c c c c c c c } |
328 |
|
|
\hline \\[-7mm] |
329 |
|
|
\ \ Equilibria Point\ \ & \ \ \ \ \ TIP3P \ \ & \ \ \ \ \ TIP4P \ \ & \ \quad \ \ \ \ TIP5P \ \ & \ \ \ \ \ SPC/E \ \ & \ \ \ \ \ SSD/E \ \ & \ \ \ \ \ SSD/RF \ \ & \ \ \ \ \ Exp. \ \ \\ |
330 |
|
|
\hline \\[-3mm] |
331 |
|
|
\ \ $T_m$ (K) & \ \ 269 & \ \ 265 & \ \ 271 & 297 & \ \ - & \ \ 278 & \ \ 273\\ |
332 |
|
|
\ \ $T_b$ (K) & \ \ 357 & \ \ 354 & \ \ 337 & 396 & \ \ - & \ \ 349 & \ \ 373\\ |
333 |
|
|
\ \ $T_s$ (K) & \ \ - & \ \ - & \ \ - & - & \ \ 355 & \ \ - & \ \ -\\ |
334 |
|
|
\end{tabular} |
335 |
|
|
\label{meltandboil} |
336 |
|
|
\end{center} |
337 |
|
|
\end{minipage} |
338 |
|
|
\end{table*} |
339 |
|
|
|
340 |
|
|
Table \ref{meltandboil} lists the melting and boiling temperatures |
341 |
|
|
calculated from this work. Surprisingly, most of these models have |
342 |
|
|
melting points that compare quite favorably with experiment. The |
343 |
|
|
unfortunate aspect of this result is that this phase change occurs |
344 |
|
|
between Ice-{\it i} and the liquid state rather than ice $I_h$ and the |
345 |
|
|
liquid state. These results are actually not contrary to previous |
346 |
|
|
studies in the literature. Earlier free energy studies of ice $I$ |
347 |
|
|
using TIP4P predict a $T_m$ ranging from 214 to 238 K (differences |
348 |
|
|
being attributed to choice of interaction truncation and different |
349 |
|
|
ordered and disordered molecular arrangements). If the presence of ice |
350 |
|
|
B and Ice-{\it i} were omitted, a $T_m$ value around 210 K would be |
351 |
|
|
predicted from this work. However, the $T_m$ from Ice-{\it i} is |
352 |
|
|
calculated at 265 K, significantly higher in temperature than the |
353 |
|
|
previous studies. Also of interest in these results is that SSD/E does |
354 |
|
|
not exhibit a melting point at 1 atm, but it shows a sublimation point |
355 |
|
|
at 355 K. This is due to the significant stability of Ice-{\it i} over |
356 |
|
|
all other polymorphs for this particular model under these |
357 |
|
|
conditions. While troubling, this behavior turned out to be |
358 |
chrisfen |
1459 |
advantageous in that it facilitated the spontaneous crystallization of |
359 |
chrisfen |
1456 |
Ice-{\it i}. These observations provide a warning that simulations of |
360 |
|
|
SSD/E as a ``liquid'' near 300 K are actually metastable and run the |
361 |
|
|
risk of spontaneous crystallization. However, this risk changes when |
362 |
|
|
applying a longer cutoff. |
363 |
|
|
|
364 |
chrisfen |
1458 |
\begin{figure} |
365 |
|
|
\includegraphics[width=\linewidth]{cutoffChange.eps} |
366 |
|
|
\caption{Free energy as a function of cutoff radius for (A) SSD/E, (B) |
367 |
|
|
TIP3P, and (C) SSD/RF. Data points omitted include SSD/E: $I_c$ 12 |
368 |
|
|
\AA\, TIP3P: $I_c$ 12 \AA\ and B 12 \AA\, and SSD/RF: $I_c$ 9 |
369 |
|
|
\AA\. These crystals are unstable at 200 K and rapidly convert into a |
370 |
|
|
liquid. The connecting lines are qualitative visual aid.} |
371 |
|
|
\label{incCutoff} |
372 |
|
|
\end{figure} |
373 |
|
|
|
374 |
chrisfen |
1457 |
Increasing the cutoff radius in simulations of the more |
375 |
|
|
computationally efficient water models was done in order to evaluate |
376 |
|
|
the trend in free energy values when moving to systems that do not |
377 |
|
|
involve potential truncation. As seen in Fig. \ref{incCutoff}, the |
378 |
|
|
free energy of all the ice polymorphs show a substantial dependence on |
379 |
|
|
cutoff radius. In general, there is a narrowing of the free energy |
380 |
chrisfen |
1459 |
differences while moving to greater cutoff radius. Interestingly, by |
381 |
|
|
increasing the cutoff radius, the free energy gap was narrowed enough |
382 |
|
|
in the SSD/E model that the liquid state is preferred under standard |
383 |
|
|
simulation conditions (298 K and 1 atm). Thus, it is recommended that |
384 |
|
|
simulations using this model choose interaction truncation radii |
385 |
|
|
greater than 9 \AA\. This narrowing trend is much more subtle in the |
386 |
|
|
case of SSD/RF, indicating that the free energies calculated with a |
387 |
|
|
reaction field present provide a more accurate picture of the free |
388 |
|
|
energy landscape in the absence of potential truncation. |
389 |
chrisfen |
1456 |
|
390 |
chrisfen |
1457 |
To further study the changes resulting to the inclusion of a |
391 |
|
|
long-range interaction correction, the effect of an Ewald summation |
392 |
|
|
was estimated by applying the potential energy difference do to its |
393 |
|
|
inclusion in systems in the presence and absence of the |
394 |
|
|
correction. This was accomplished by calculation of the potential |
395 |
|
|
energy of identical crystals with and without PME using TINKER. The |
396 |
|
|
free energies for the investigated polymorphs using the TIP3P and |
397 |
|
|
SPC/E water models are shown in Table \ref{pmeShift}. TIP4P and TIP5P |
398 |
|
|
are not fully supported in TINKER, so the results for these models |
399 |
|
|
could not be estimated. The same trend pointed out through increase of |
400 |
chrisfen |
1459 |
cutoff radius is observed in these PME results. Ice-{\it i} is the |
401 |
chrisfen |
1457 |
preferred polymorph at ambient conditions for both the TIP3P and SPC/E |
402 |
|
|
water models; however, there is a narrowing of the free energy |
403 |
|
|
differences between the various solid forms. In the case of SPC/E this |
404 |
|
|
narrowing is significant enough that it becomes less clear cut that |
405 |
|
|
Ice-{\it i} is the most stable polymorph, and is possibly metastable |
406 |
|
|
with respect to ice B and possibly ice $I_c$. However, these results |
407 |
|
|
do not significantly alter the finding that the Ice-{\it i} polymorph |
408 |
|
|
is a stable crystal structure that should be considered when studying |
409 |
|
|
the phase behavior of water models. |
410 |
chrisfen |
1456 |
|
411 |
chrisfen |
1457 |
\begin{table*} |
412 |
|
|
\begin{minipage}{\linewidth} |
413 |
|
|
\renewcommand{\thefootnote}{\thempfootnote} |
414 |
|
|
\begin{center} |
415 |
chrisfen |
1458 |
\caption{The free energy of the studied ice polymorphs after applying |
416 |
|
|
the energy difference attributed to the inclusion of the PME |
417 |
|
|
long-range interaction correction. Units are kcal/mol.} |
418 |
chrisfen |
1457 |
\begin{tabular}{ l c c c c } |
419 |
|
|
\hline \\[-7mm] |
420 |
|
|
\ \ Water Model \ \ & \ \ \ \ \ $I_h$ \ \ & \ \ \ \ \ $I_c$ \ \ & \ \quad \ \ \ \ B \ \ & \ \ \ \ \ Ice-{\it i} \ \ \\ |
421 |
|
|
\hline \\[-3mm] |
422 |
|
|
\ \ TIP3P & \ \ -11.53 & \ \ -11.24 & \ \ -11.51 & \ \ -11.67\\ |
423 |
|
|
\ \ SPC/E & \ \ -12.77 & \ \ -12.92 & \ \ -12.96 & \ \ -13.02\\ |
424 |
|
|
\end{tabular} |
425 |
|
|
\label{pmeShift} |
426 |
|
|
\end{center} |
427 |
|
|
\end{minipage} |
428 |
|
|
\end{table*} |
429 |
|
|
|
430 |
chrisfen |
1453 |
\section{Conclusions} |
431 |
|
|
|
432 |
chrisfen |
1458 |
The free energy for proton ordered variants of hexagonal and cubic ice |
433 |
|
|
$I$, ice B, and recently discovered Ice-{\it i} where calculated under |
434 |
|
|
standard conditions for several common water models via thermodynamic |
435 |
|
|
integration. All the water models studied show Ice-{\it i} to be the |
436 |
|
|
minimum free energy crystal structure in the with a 9 \AA\ switching |
437 |
|
|
function cutoff. Calculated melting and boiling points show |
438 |
|
|
surprisingly good agreement with the experimental values; however, the |
439 |
|
|
solid phase at 1 atm is Ice-{\it i}, not ice $I_h$. The effect of |
440 |
|
|
interaction truncation was investigated through variation of the |
441 |
|
|
cutoff radius, use of a reaction field parameterized model, and |
442 |
|
|
estimation of the results in the presence of the Ewald summation |
443 |
|
|
correction. Interaction truncation has a significant effect on the |
444 |
chrisfen |
1459 |
computed free energy values, and may significantly alter the free |
445 |
|
|
energy landscape for the more complex multipoint water models. Despite |
446 |
|
|
these effects, these results show Ice-{\it i} to be an important ice |
447 |
|
|
polymorph that should be considered in simulation studies. |
448 |
chrisfen |
1458 |
|
449 |
chrisfen |
1459 |
Due to this relative stability of Ice-{\it i} in all manner of |
450 |
|
|
investigated simulation examples, the question arises as to possible |
451 |
|
|
experimental observation of this polymorph. The rather extensive past |
452 |
|
|
and current experimental investigation of water in the low pressure |
453 |
|
|
regime leads the authors to be hesitant in ascribing relevance outside |
454 |
|
|
of computational models, hence the descriptive name presented. That |
455 |
|
|
being said, there are certain experimental conditions that would |
456 |
|
|
provide the most ideal situation for possible observation. These |
457 |
|
|
include the negative pressure or stretched solid regime, small |
458 |
|
|
clusters in vacuum deposition environments, and in clathrate |
459 |
|
|
structures involving small non-polar molecules. |
460 |
|
|
|
461 |
chrisfen |
1453 |
\section{Acknowledgments} |
462 |
|
|
Support for this project was provided by the National Science |
463 |
|
|
Foundation under grant CHE-0134881. Computation time was provided by |
464 |
chrisfen |
1458 |
the Notre Dame High Performance Computing Cluster and the Notre Dame |
465 |
|
|
Bunch-of-Boxes (B.o.B) computer cluster (NSF grant DMR-0079647). |
466 |
chrisfen |
1453 |
|
467 |
|
|
\newpage |
468 |
|
|
|
469 |
|
|
\bibliographystyle{jcp} |
470 |
|
|
\bibliography{iceiPaper} |
471 |
|
|
|
472 |
|
|
|
473 |
|
|
\end{document} |