| 1 |
|
%\documentclass[prb,aps,twocolumn,tabularx]{revtex4} |
| 2 |
|
\documentclass[11pt]{article} |
| 3 |
– |
%\documentclass[11pt]{article} |
| 3 |
|
\usepackage{endfloat} |
| 4 |
|
\usepackage{amsmath} |
| 5 |
|
\usepackage{epsf} |
| 126 |
|
common water models (TIP3P, TIP4P, TIP5P, and SPC/E) and a reaction |
| 127 |
|
field parametrized single point dipole water model (SSD/RF). The |
| 128 |
|
extruded variant, Ice-{\it i}$^\prime$, was used in calculations |
| 129 |
< |
involving SPC/E, TIP4P, and TIP5P due to its enhanced stability with |
| 130 |
< |
these models. There is typically a small distortion of proton ordered |
| 131 |
< |
Ice-{\it i}$^\prime$ that converts the normally square tetramer into a |
| 132 |
< |
rhombus with alternating approximately 85 and 95 degree angles. The |
| 133 |
< |
degree of this distortion is model dependent and significant enough to |
| 134 |
< |
split the tetramer diagonal location peak in the radial distribution |
| 135 |
< |
function. |
| 129 |
> |
involving SPC/E, TIP4P, and TIP5P. These models exhibit enhanced |
| 130 |
> |
stability with Ice-{\it i}$^\prime$ because of their more |
| 131 |
> |
tetrahedrally arranged internal charge distributions. Additionally, |
| 132 |
> |
there is often a small distortion of proton ordered Ice-{\it |
| 133 |
> |
i}$^\prime$ that converts the normally square tetramer into a rhombus |
| 134 |
> |
with alternating approximately 85 and 95 degree angles. The degree of |
| 135 |
> |
this distortion is model dependent and significant enough to split the |
| 136 |
> |
tetramer diagonal location peak in the radial distribution function. |
| 137 |
|
|
| 138 |
|
Thermodynamic integration was utilized to calculate the free energies |
| 139 |
|
of the listed water models at various state points using a modified |
| 141 |
|
This calculation method involves a sequence of simulations during |
| 142 |
|
which the system of interest is converted into a reference system for |
| 143 |
|
which the free energy is known analytically. This transformation path |
| 144 |
< |
is then integrated in order to determine the free energy difference |
| 144 |
> |
is then integrated, in order to determine the free energy difference |
| 145 |
|
between the two states: |
| 146 |
|
\begin{equation} |
| 147 |
|
\Delta A = \int_0^1\left\langle\frac{\partial V(\lambda |
| 281 |
|
each of the polymorphs studied assumes the role of the preferred |
| 282 |
|
polymorph under different cutoff conditions. The inclusion of the |
| 283 |
|
Ewald correction flattens and narrows the sequences of free energies |
| 284 |
< |
so much that they often overlap within error, indicating that other |
| 285 |
< |
conditions, such as cell volume in microcanonical simulations, can |
| 284 |
> |
such that they often overlap within error, indicating that other |
| 285 |
> |
conditions, such as the density in fixed volume simulations, can |
| 286 |
|
influence the chosen polymorph upon crystallization. |
| 287 |
|
|
| 288 |
|
So what is the preferred solid polymorph for simulated water? The |
