| 1 |
%\documentclass[prb,aps,twocolumn,tabularx]{revtex4} |
| 2 |
\documentclass[preprint,aps,endfloats]{revtex4} |
| 3 |
%\documentclass[11pt]{article} |
| 4 |
%\usepackage{endfloat} |
| 5 |
\usepackage{amsmath} |
| 6 |
\usepackage{epsf} |
| 7 |
\usepackage{berkeley} |
| 8 |
%\usepackage{setspace} |
| 9 |
%\usepackage{tabularx} |
| 10 |
\usepackage{graphicx} |
| 11 |
%\usepackage[ref]{overcite} |
| 12 |
%\pagestyle{plain} |
| 13 |
%\pagenumbering{arabic} |
| 14 |
%\oddsidemargin 0.0cm \evensidemargin 0.0cm |
| 15 |
%\topmargin -21pt \headsep 10pt |
| 16 |
%\textheight 9.0in \textwidth 6.5in |
| 17 |
%\brokenpenalty=10000 |
| 18 |
|
| 19 |
%\renewcommand\citemid{\ } % no comma in optional reference note |
| 20 |
|
| 21 |
\begin{document} |
| 22 |
|
| 23 |
\title{A Free Energy Study of Low Temperature and Anomolous Ice} |
| 24 |
|
| 25 |
\author{Christopher J. Fennell and J. Daniel Gezelter{\thefootnote} |
| 26 |
\footnote[1]{Corresponding author. \ Electronic mail: gezelter@nd.edu}} |
| 27 |
|
| 28 |
\address{Department of Chemistry and Biochemistry\\ University of Notre Dame\\ |
| 29 |
Notre Dame, Indiana 46556} |
| 30 |
|
| 31 |
\date{\today} |
| 32 |
|
| 33 |
%\maketitle |
| 34 |
%\doublespacing |
| 35 |
|
| 36 |
\begin{abstract} |
| 37 |
\end{abstract} |
| 38 |
|
| 39 |
\maketitle |
| 40 |
|
| 41 |
\newpage |
| 42 |
|
| 43 |
%\narrowtext |
| 44 |
|
| 45 |
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
| 46 |
% BODY OF TEXT |
| 47 |
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
| 48 |
|
| 49 |
\section{Introduction} |
| 50 |
|
| 51 |
\section{Methods} |
| 52 |
|
| 53 |
Canonical ensemble (NVT) molecular dynamics calculations were |
| 54 |
performed using the OOPSE (Object-Oriented Parallel Simulation Engine) |
| 55 |
molecular mechanics package. All molecules were treated as rigid |
| 56 |
bodies, with orientational motion propogated using the symplectic DLM |
| 57 |
integration method. Details about the implementation of these |
| 58 |
techniques can be found in a recent publication.\cite{Meineke05} |
| 59 |
|
| 60 |
Thermodynamic integration was utilized to calculate the free energy of |
| 61 |
several ice crystals at 200 K using the TIP3P, TIP4P, TIP5P, SPC/E, |
| 62 |
SSD/RF, and SSD/E water models. Liquid state free energies at 300 and |
| 63 |
400 K for all of these water models were also determined using this |
| 64 |
same technique, in order to determine melting points and generate |
| 65 |
phase diagrams. All simulations were carried out at densities |
| 66 |
resulting in a pressure of approximately 1 atm at their respective |
| 67 |
temperatures. |
| 68 |
|
| 69 |
A single thermodynamic integration involves a sequence of simulations |
| 70 |
over which the system of interest is converted into a reference system |
| 71 |
for which the free energy is known. This transformation path is then |
| 72 |
integrated in order to determine the free energy difference between |
| 73 |
the two states: |
| 74 |
\begin{equation} |
| 75 |
\begin{center} |
| 76 |
\Delta A = \int_0^1\left\langle\frac{\partial V(\lambda |
| 77 |
)}{\partial\lambda}\right\rangle_\lambda d\lambda, |
| 78 |
\end{center} |
| 79 |
\end{equation} |
| 80 |
where $V$ is the interaction potential and $\lambda$ is the |
| 81 |
transformation parameter. Simulations are distributed unevenly along |
| 82 |
this path in order to sufficiently sample the regions of greatest |
| 83 |
change in the potential. Typical integrations in this study consisted |
| 84 |
of $\sim$25 simulations ranging from 300 ps (for the unaltered system) |
| 85 |
to 75 ps (near the reference state) in length. |
| 86 |
|
| 87 |
For the thermodynamic integration of molecular crystals, the Einstein |
| 88 |
Crystal is chosen as the reference state that the system is converted |
| 89 |
to over the course of the simulation. In an Einstein Crystal, the |
| 90 |
molecules are harmonically restrained at their ideal lattice locations |
| 91 |
and orientations. The partition function for a molecular crystal |
| 92 |
restrained in this fashion has been evaluated, and the Helmholtz Free |
| 93 |
Energy ({\it A}) is given by |
| 94 |
\begin{eqnarray} |
| 95 |
A = E_m\ -\ kT\ln \left (\frac{kT}{h\nu}\right )^3&-&kT\ln \left |
| 96 |
[\pi^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{A}kT}{h^2}\right |
| 97 |
)^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{B}kT}{h^2}\right |
| 98 |
)^\frac{1}{2}\left (\frac{8\pi^2I_\mathrm{C}kT}{h^2}\right |
| 99 |
)^\frac{1}{2}\right ] \nonumber \\ &-& kT\ln \left [\frac{kT}{2(\pi |
| 100 |
K_\omega K_\theta)^{\frac{1}{2}}}\exp\left |
| 101 |
(-\frac{kT}{2K_\theta}\right)\int_0^{\left (\frac{kT}{2K_\theta}\right |
| 102 |
)^\frac{1}{2}}\exp(t^2)\mathrm{d}t\right ], |
| 103 |
\label{ecFreeEnergy} |
| 104 |
\end{eqnarray} |
| 105 |
where $2\pi\nu = (K_\mathrm{v}/m)^{1/2}$.\cite{Baez95a} In equation |
| 106 |
\ref{ecFreeEnergy}, $K_\mathrm{v}$, $K_\mathrm{\theta}$, and |
| 107 |
$K_\mathrm{\omega}$ are the spring constants restraining translational |
| 108 |
motion and deflection of and rotation around the principle axis of the |
| 109 |
molecule respectively (Fig. \ref{waterSpring}), and $E_m$ is the |
| 110 |
minimum potential energy of the ideal crystal. In the case of |
| 111 |
molecular liquids, the ideal vapor is chosen as the target reference |
| 112 |
state. |
| 113 |
\begin{figure} |
| 114 |
\includegraphics[scale=1.0]{rotSpring.eps} |
| 115 |
\caption{Possible orientational motions for a restrained molecule. |
| 116 |
$\theta$ angles correspond to displacement from the body-frame {\it |
| 117 |
z}-axis, while $\omega$ angles correspond to rotation about the |
| 118 |
body-frame {\it z}-axis. $K_\theta$ and $K_\omega$ are spring |
| 119 |
constants for the harmonic springs restraining motion in the $\theta$ |
| 120 |
and $\omega$ directions.} |
| 121 |
\label{waterSpring} |
| 122 |
\end{figure} |
| 123 |
|
| 124 |
Charge, dipole, and Lennard-Jones interactions were modified by a |
| 125 |
cubic switching between 100\% and 85\% of the cutoff value (9 \AA ). By |
| 126 |
applying this function, these interactions are smoothly truncated, |
| 127 |
thereby avoiding poor energy conserving dynamics resulting from |
| 128 |
harsher truncation schemes. The effect of a long-range correction was |
| 129 |
also investigated on select model systems in a variety of manners. For |
| 130 |
the SSD/RF model, a reaction field with a fixed dielectric constant of |
| 131 |
80 was applied in all simulations.\cite{Onsager36} For a series of the |
| 132 |
least computationally expensive models (SSD/E, SSD/RF, and TIP3P), |
| 133 |
simulations were performed with longer cutoffs of 12 and 15 \AA\ to |
| 134 |
compare with the 9 \AA\ cutoff results. Finally, results from the use |
| 135 |
of an Ewald summation were estimated for TIP3P and SPC/E by performing |
| 136 |
calculations with Particle-Mesh Ewald (PME) in the TINKER molecular |
| 137 |
mechanics software package. TINKER was chosen because it can also |
| 138 |
propogate the motion of rigid-bodies, and provides the most direct |
| 139 |
comparison to the results from OOPSE. The calculated energy difference |
| 140 |
in the presence and absence of PME was applied to the previous results |
| 141 |
in order to predict changes in the free energy landscape. |
| 142 |
|
| 143 |
\section{Results and discussion} |
| 144 |
|
| 145 |
The free energy of proton ordered Ice-{\it i} was calculated and |
| 146 |
compared with the free energies of proton ordered variants of the |
| 147 |
experimentally observed low-density ice polymorphs, $I_h$ and $I_c$, |
| 148 |
as well as the higher density ice B, observed by B\`{a}ez and Clancy |
| 149 |
and thought to be the minimum free energy structure for the SPC/E |
| 150 |
model at ambient conditions (Table \ref{freeEnergy}).\cite{Baez95b} |
| 151 |
Ice XI, the experimentally observed proton ordered variant of ice |
| 152 |
$I_h$, was investigated initially, but it was found not to be as |
| 153 |
stable as antiferroelectric variants of proton ordered or even proton |
| 154 |
disordered ice$I_h$.\cite{Davidson84} The proton ordered variant of |
| 155 |
ice $I_h$ used here is a simple antiferroelectric version that has an |
| 156 |
8 molecule unit cell. The crystals contained 648 or 1728 molecules for |
| 157 |
ice B, 1024 or 1280 molecules for ice $I_h$, 1000 molecules for ice |
| 158 |
$I_c$, or 1024 molecules for Ice-{\it i}. The larger crystal sizes |
| 159 |
were necessary for simulations involving larger cutoff values. |
| 160 |
|
| 161 |
\begin{table*} |
| 162 |
\begin{minipage}{\linewidth} |
| 163 |
\renewcommand{\thefootnote}{\thempfootnote} |
| 164 |
\begin{center} |
| 165 |
\caption{Calculated free energies for several ice polymorphs with a |
| 166 |
variety of common water models. All calculations used a cutoff radius |
| 167 |
of 9 \AA\ and were performed at 200 K and $\sim$1 atm. Units are |
| 168 |
kcal/mol. *Ice $I_c$ is unstable at 200 K using SSD/RF.} |
| 169 |
\begin{tabular}{ l c c c c } |
| 170 |
\hline \\[-7mm] |
| 171 |
\ \quad \ Water Model\ \ & \ \quad \ \ \ \ $I_h$ \ \ & \ \quad \ \ \ \ $I_c$ \ \ & \ \quad \ \ \ \ B \ \ & \ \quad \ \ \ Ice-{\it i} \ \quad \ \\ |
| 172 |
\hline \\[-3mm] |
| 173 |
\ \quad \ TIP3P & \ \quad \ -11.41 & \ \quad \ -11.23 & \ \quad \ -11.82 & \quad -12.30\\ |
| 174 |
\ \quad \ TIP4P & \ \quad \ -11.84 & \ \quad \ -12.04 & \ \quad \ -12.08 & \quad -12.33\\ |
| 175 |
\ \quad \ TIP5P & \ \quad \ -11.85 & \ \quad \ -11.86 & \ \quad \ -11.96 & \quad -12.29\\ |
| 176 |
\ \quad \ SPC/E & \ \quad \ -12.67 & \ \quad \ -12.96 & \ \quad \ -13.25 & \quad -13.55\\ |
| 177 |
\ \quad \ SSD/E & \ \quad \ -11.27 & \ \quad \ -11.19 & \ \quad \ -12.09 & \quad -12.54\\ |
| 178 |
\ \quad \ SSD/RF & \ \quad \ -11.51 & \ \quad \ NA* & \ \quad \ -12.08 & \quad -12.29\\ |
| 179 |
\end{tabular} |
| 180 |
\label{freeEnergy} |
| 181 |
\end{center} |
| 182 |
\end{minipage} |
| 183 |
\end{table*} |
| 184 |
|
| 185 |
The free energy values computed for the studied polymorphs indicate |
| 186 |
that Ice-{\it i} is the most stable state for all of the common water |
| 187 |
models studied. With the free energy at these state points, the |
| 188 |
temperature and pressure dependence of the free energy was used to |
| 189 |
project to other state points and build phase diagrams. Figures |
| 190 |
\ref{tp3phasedia} and \ref{ssdrfphasedia} are example diagrams built |
| 191 |
from the free energy results. All other models have similar structure, |
| 192 |
only the crossing points between these phases exist at different |
| 193 |
temperatures and pressures. It is interesting to note that ice $I$ |
| 194 |
does not exist in either cubic or hexagonal form in any of the phase |
| 195 |
diagrams for any of the models. For purposes of this study, ice B is |
| 196 |
representative of the dense ice polymorphs. A recent study by Sanz |
| 197 |
{\it et al.} goes into detail on the phase diagrams for SPC/E and |
| 198 |
TIP4P in the high pressure regime.\cite{Sanz04} |
| 199 |
\begin{figure} |
| 200 |
\includegraphics[width=\linewidth]{tp3PhaseDia.eps} |
| 201 |
\caption{Phase diagram for the TIP3P water model in the low pressure |
| 202 |
regime. The displayed $T_m$ and $T_b$ values are good predictions of |
| 203 |
the experimental values; however, the solid phases shown are not the |
| 204 |
experimentally observed forms. Both cubic and hexagonal ice $I$ are |
| 205 |
higher in energy and don't appear in the phase diagram.} |
| 206 |
\label{tp3phasedia} |
| 207 |
\end{figure} |
| 208 |
\begin{figure} |
| 209 |
\includegraphics[width=\linewidth]{ssdrfPhaseDia.eps} |
| 210 |
\caption{Phase diagram for the SSD/RF water model in the low pressure |
| 211 |
regime. Calculations producing these results were done under an |
| 212 |
applied reaction field. It is interesting to note that this |
| 213 |
computationally efficient model (over 3 times more efficient than |
| 214 |
TIP3P) exhibits phase behavior similar to the less computationally |
| 215 |
conservative charge based models.} |
| 216 |
\label{ssdrfphasedia} |
| 217 |
\end{figure} |
| 218 |
|
| 219 |
\begin{table*} |
| 220 |
\begin{minipage}{\linewidth} |
| 221 |
\renewcommand{\thefootnote}{\thempfootnote} |
| 222 |
\begin{center} |
| 223 |
\caption{Melting ($T_m$), boiling ($T_b$), and sublimation ($T_s$) |
| 224 |
temperatures of several common water models compared with experiment.} |
| 225 |
\begin{tabular}{ l c c c c c c c } |
| 226 |
\hline \\[-7mm] |
| 227 |
\ \ Equilibria Point\ \ & \ \ \ \ \ TIP3P \ \ & \ \ \ \ \ TIP4P \ \ & \ \quad \ \ \ \ TIP5P \ \ & \ \ \ \ \ SPC/E \ \ & \ \ \ \ \ SSD/E \ \ & \ \ \ \ \ SSD/RF \ \ & \ \ \ \ \ Exp. \ \ \\ |
| 228 |
\hline \\[-3mm] |
| 229 |
\ \ $T_m$ (K) & \ \ 269 & \ \ 265 & \ \ 271 & 297 & \ \ - & \ \ 278 & \ \ 273\\ |
| 230 |
\ \ $T_b$ (K) & \ \ 357 & \ \ 354 & \ \ 337 & 396 & \ \ - & \ \ 349 & \ \ 373\\ |
| 231 |
\ \ $T_s$ (K) & \ \ - & \ \ - & \ \ - & - & \ \ 355 & \ \ - & \ \ -\\ |
| 232 |
\end{tabular} |
| 233 |
\label{meltandboil} |
| 234 |
\end{center} |
| 235 |
\end{minipage} |
| 236 |
\end{table*} |
| 237 |
|
| 238 |
Table \ref{meltandboil} lists the melting and boiling temperatures |
| 239 |
calculated from this work. Surprisingly, most of these models have |
| 240 |
melting points that compare quite favorably with experiment. The |
| 241 |
unfortunate aspect of this result is that this phase change occurs |
| 242 |
between Ice-{\it i} and the liquid state rather than ice $I_h$ and the |
| 243 |
liquid state. These results are actually not contrary to previous |
| 244 |
studies in the literature. Earlier free energy studies of ice $I$ |
| 245 |
using TIP4P predict a $T_m$ ranging from 214 to 238 K (differences |
| 246 |
being attributed to choice of interaction truncation and different |
| 247 |
ordered and disordered molecular arrangements). If the presence of ice |
| 248 |
B and Ice-{\it i} were omitted, a $T_m$ value around 210 K would be |
| 249 |
predicted from this work. However, the $T_m$ from Ice-{\it i} is |
| 250 |
calculated at 265 K, significantly higher in temperature than the |
| 251 |
previous studies. Also of interest in these results is that SSD/E does |
| 252 |
not exhibit a melting point at 1 atm, but it shows a sublimation point |
| 253 |
at 355 K. This is due to the significant stability of Ice-{\it i} over |
| 254 |
all other polymorphs for this particular model under these |
| 255 |
conditions. While troubling, this behavior turned out to be |
| 256 |
advantagious in that it facilitated the spontaneous crystallization of |
| 257 |
Ice-{\it i}. These observations provide a warning that simulations of |
| 258 |
SSD/E as a ``liquid'' near 300 K are actually metastable and run the |
| 259 |
risk of spontaneous crystallization. However, this risk changes when |
| 260 |
applying a longer cutoff. |
| 261 |
|
| 262 |
\begin{figure} |
| 263 |
\includegraphics[width=\linewidth]{cutoffChange.eps} |
| 264 |
\caption{Free energy as a function of cutoff radius for (A) SSD/E, (B) |
| 265 |
TIP3P, and (C) SSD/RF. Data points omitted include SSD/E: $I_c$ 12 |
| 266 |
\AA\, TIP3P: $I_c$ 12 \AA\ and B 12 \AA\, and SSD/RF: $I_c$ 9 |
| 267 |
\AA\. These crystals are unstable at 200 K and rapidly convert into a |
| 268 |
liquid. The connecting lines are qualitative visual aid.} |
| 269 |
\label{incCutoff} |
| 270 |
\end{figure} |
| 271 |
|
| 272 |
Increasing the cutoff radius in simulations of the more |
| 273 |
computationally efficient water models was done in order to evaluate |
| 274 |
the trend in free energy values when moving to systems that do not |
| 275 |
involve potential truncation. As seen in Fig. \ref{incCutoff}, the |
| 276 |
free energy of all the ice polymorphs show a substantial dependence on |
| 277 |
cutoff radius. In general, there is a narrowing of the free energy |
| 278 |
differences while moving to greater cutoff radius. This trend is much |
| 279 |
more subtle in the case of SSD/RF, indicating that the free energies |
| 280 |
calculated with a reaction field present provide a more accurate |
| 281 |
picture of the free energy landscape in the absence of potential |
| 282 |
truncation. |
| 283 |
|
| 284 |
To further study the changes resulting to the inclusion of a |
| 285 |
long-range interaction correction, the effect of an Ewald summation |
| 286 |
was estimated by applying the potential energy difference do to its |
| 287 |
inclusion in systems in the presence and absence of the |
| 288 |
correction. This was accomplished by calculation of the potential |
| 289 |
energy of identical crystals with and without PME using TINKER. The |
| 290 |
free energies for the investigated polymorphs using the TIP3P and |
| 291 |
SPC/E water models are shown in Table \ref{pmeShift}. TIP4P and TIP5P |
| 292 |
are not fully supported in TINKER, so the results for these models |
| 293 |
could not be estimated. The same trend pointed out through increase of |
| 294 |
cutoff radius is observed in these results. Ice-{\it i} is the |
| 295 |
preferred polymorph at ambient conditions for both the TIP3P and SPC/E |
| 296 |
water models; however, there is a narrowing of the free energy |
| 297 |
differences between the various solid forms. In the case of SPC/E this |
| 298 |
narrowing is significant enough that it becomes less clear cut that |
| 299 |
Ice-{\it i} is the most stable polymorph, and is possibly metastable |
| 300 |
with respect to ice B and possibly ice $I_c$. However, these results |
| 301 |
do not significantly alter the finding that the Ice-{\it i} polymorph |
| 302 |
is a stable crystal structure that should be considered when studying |
| 303 |
the phase behavior of water models. |
| 304 |
|
| 305 |
\begin{table*} |
| 306 |
\begin{minipage}{\linewidth} |
| 307 |
\renewcommand{\thefootnote}{\thempfootnote} |
| 308 |
\begin{center} |
| 309 |
\caption{The free energy of the studied ice polymorphs after applying |
| 310 |
the energy difference attributed to the inclusion of the PME |
| 311 |
long-range interaction correction. Units are kcal/mol.} |
| 312 |
\begin{tabular}{ l c c c c } |
| 313 |
\hline \\[-7mm] |
| 314 |
\ \ Water Model \ \ & \ \ \ \ \ $I_h$ \ \ & \ \ \ \ \ $I_c$ \ \ & \ \quad \ \ \ \ B \ \ & \ \ \ \ \ Ice-{\it i} \ \ \\ |
| 315 |
\hline \\[-3mm] |
| 316 |
\ \ TIP3P & \ \ -11.53 & \ \ -11.24 & \ \ -11.51 & \ \ -11.67\\ |
| 317 |
\ \ SPC/E & \ \ -12.77 & \ \ -12.92 & \ \ -12.96 & \ \ -13.02\\ |
| 318 |
\end{tabular} |
| 319 |
\label{pmeShift} |
| 320 |
\end{center} |
| 321 |
\end{minipage} |
| 322 |
\end{table*} |
| 323 |
|
| 324 |
\section{Conclusions} |
| 325 |
|
| 326 |
The free energy for proton ordered variants of hexagonal and cubic ice |
| 327 |
$I$, ice B, and recently discovered Ice-{\it i} where calculated under |
| 328 |
standard conditions for several common water models via thermodynamic |
| 329 |
integration. All the water models studied show Ice-{\it i} to be the |
| 330 |
minimum free energy crystal structure in the with a 9 \AA\ switching |
| 331 |
function cutoff. Calculated melting and boiling points show |
| 332 |
surprisingly good agreement with the experimental values; however, the |
| 333 |
solid phase at 1 atm is Ice-{\it i}, not ice $I_h$. The effect of |
| 334 |
interaction truncation was investigated through variation of the |
| 335 |
cutoff radius, use of a reaction field parameterized model, and |
| 336 |
estimation of the results in the presence of the Ewald summation |
| 337 |
correction. Interaction truncation has a significant effect on the |
| 338 |
computed free energy values, but Ice-{\it i} is still observed to be a |
| 339 |
relavent ice polymorph in simulation studies. |
| 340 |
|
| 341 |
\section{Acknowledgments} |
| 342 |
Support for this project was provided by the National Science |
| 343 |
Foundation under grant CHE-0134881. Computation time was provided by |
| 344 |
the Notre Dame High Performance Computing Cluster and the Notre Dame |
| 345 |
Bunch-of-Boxes (B.o.B) computer cluster (NSF grant DMR-0079647). |
| 346 |
|
| 347 |
\newpage |
| 348 |
|
| 349 |
\bibliographystyle{jcp} |
| 350 |
\bibliography{iceiPaper} |
| 351 |
|
| 352 |
|
| 353 |
\end{document} |