ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/trunk/langevinHull/langevinHull.tex
Revision: 3668
Committed: Thu Nov 4 16:01:58 2010 UTC (13 years, 7 months ago) by gezelter
Content type: application/x-tex
File size: 35625 byte(s)
Log Message:
modified the parallel figure

File Contents

# User Rev Content
1 gezelter 3640 \documentclass[11pt]{article}
2     \usepackage{amsmath}
3     \usepackage{amssymb}
4     \usepackage{setspace}
5     \usepackage{endfloat}
6     \usepackage{caption}
7     \usepackage{graphicx}
8     \usepackage{multirow}
9     \usepackage[square, comma, sort&compress]{natbib}
10     \usepackage{url}
11     \pagestyle{plain} \pagenumbering{arabic} \oddsidemargin 0.0cm
12     \evensidemargin 0.0cm \topmargin -21pt \headsep 10pt \textheight
13     9.0in \textwidth 6.5in \brokenpenalty=10000
14    
15     % double space list of tables and figures
16     %\AtBeginDelayedFloats{\renewcomand{\baselinestretch}{1.66}}
17     \setlength{\abovecaptionskip}{20 pt}
18     \setlength{\belowcaptionskip}{30 pt}
19    
20     \bibpunct{[}{]}{,}{s}{}{;}
21 gezelter 3667 \bibliographystyle{achemso}
22 gezelter 3640
23     \begin{document}
24    
25     \title{The Langevin Hull: Constant pressure and temperature dynamics for non-periodic systems}
26    
27 kstocke1 3644 \author{Charles F. Vardeman II, Kelsey M. Stocker, and J. Daniel
28 gezelter 3640 Gezelter\footnote{Corresponding author. \ Electronic mail: gezelter@nd.edu} \\
29     Department of Chemistry and Biochemistry,\\
30     University of Notre Dame\\
31     Notre Dame, Indiana 46556}
32    
33     \date{\today}
34    
35     \maketitle
36    
37     \begin{doublespace}
38    
39     \begin{abstract}
40     We have developed a new isobaric-isothermal (NPT) algorithm which
41     applies an external pressure to the facets comprising the convex
42 gezelter 3665 hull surrounding the system. A Langevin thermostat is also applied
43     to facets of the hull to mimic contact with an external heat
44     bath. This new method, the ``Langevin Hull'', performs better than
45     traditional affine transform methods for systems containing
46     heterogeneous mixtures of materials with different
47 gezelter 3652 compressibilities. It does not suffer from the edge effects of
48     boundary potential methods, and allows realistic treatment of both
49     external pressure and thermal conductivity to an implicit solvent.
50     We apply this method to several different systems including bare
51 gezelter 3665 metal nanoparticles, nanoparticles in an explicit solvent, as well
52     as clusters of liquid water. The predicted mechanical properties of
53     these systems are in good agreement with experimental data and
54     previous simulation work.
55 gezelter 3640 \end{abstract}
56    
57     \newpage
58    
59     %\narrowtext
60    
61     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
62     % BODY OF TEXT
63     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
64    
65    
66     \section{Introduction}
67    
68 gezelter 3641 The most common molecular dynamics methods for sampling configurations
69 gezelter 3667 from an isobaric-isothermal (NPT) ensemble maintain a target pressure
70     in a simulation by coupling the volume of the system to a {\it
71     barostat}, which is an extra degree of freedom propagated along with
72     the particle coordinates. These methods require periodic boundary
73     conditions, because when the instantaneous pressure in the system
74     differs from the target pressure, the volume is reduced or expanded
75     using {\it affine transforms} of the system geometry. An affine
76     transform scales the size and shape of the periodic box as well as the
77     particle positions within the box (but not the sizes of the
78     particles). The most common constant pressure methods, including the
79     Melchionna modification\cite{Melchionna1993} to the
80     Nos\'e-Hoover-Andersen equations of
81     motion,\cite{Hoover85,ANDERSEN:1980vn,Sturgeon:2000kx} the Berendsen
82     pressure bath,\cite{ISI:A1984TQ73500045} and the Langevin
83     Piston,\cite{FELLER:1995fk,Jakobsen:2005uq} all utilize scaled
84 gezelter 3665 coordinate transformation to adjust the box volume. As long as the
85 gezelter 3667 material in the simulation box has a relatively uniform
86     compressibility, the standard affine transform approach provides an
87     excellent way of adjusting the volume of the system and applying
88     pressure directly via the interactions between atomic sites.
89 gezelter 3652
90 gezelter 3665 One problem with this approach appears when the system being simulated
91     is an inhomogeneous mixture in which portions of the simulation box
92     are incompressible relative to other portions. Examples include
93     simulations of metallic nanoparticles in liquid environments, proteins
94     at ice / water interfaces, as well as other heterogeneous or
95 gezelter 3652 interfacial environments. In these cases, the affine transform of
96     atomic coordinates will either cause numerical instability when the
97 gezelter 3665 sites in the incompressible medium collide with each other, or will
98     lead to inefficient sampling of system volumes if the barostat is set
99     slow enough to avoid the instabilities in the incompressible region.
100 gezelter 3652
101 gezelter 3640 \begin{figure}
102 gezelter 3641 \includegraphics[width=\linewidth]{AffineScale2}
103 gezelter 3667 \caption{Affine scaling methods use box-length scaling to adjust the
104     volume to adjust to under- or over-pressure conditions. In a system
105     with a uniform compressibility (e.g. bulk fluids) these methods can
106     work well. In systems containing heterogeneous mixtures, the affine
107     scaling moves required to adjust the pressure in the
108     high-compressibility regions can cause molecules in low
109     compressibility regions to collide.}
110 gezelter 3640 \label{affineScale}
111     \end{figure}
112    
113 gezelter 3653 One may also wish to avoid affine transform periodic boundary methods
114     to simulate {\it explicitly non-periodic systems} under constant
115     pressure conditions. The use of periodic boxes to enforce a system
116 gezelter 3665 volume requires either effective solute concentrations that are much
117 gezelter 3653 higher than desirable, or unreasonable system sizes to avoid this
118 gezelter 3665 effect. For example, calculations using typical hydration shells
119 gezelter 3653 solvating a protein under periodic boundary conditions are quite
120     expensive. [CALCULATE EFFECTIVE PROTEIN CONCENTRATIONS IN TYPICAL
121     SIMULATIONS]
122 gezelter 3640
123 gezelter 3665 \subsection*{Boundary Methods}
124 gezelter 3667 There have been a number of approaches to handle simulations of
125     explicitly non-periodic systems that focus on constant or
126     nearly-constant {\it volume} conditions while maintaining bulk-like
127     behavior. Berkowitz and McCammon introduced a stochastic (Langevin)
128     boundary layer inside a region of fixed molecules which effectively
129     enforces constant temperature and volume (NVT)
130     conditions.\cite{Berkowitz1982} In this approach, the stochastic and
131     fixed regions were defined relative to a central atom. Brooks and
132     Karplus extended this method to include deformable stochastic
133     boundaries.\cite{iii:6312} The stochastic boundary approach has been
134     used widely for protein simulations. [CITATIONS NEEDED]
135 gezelter 3640
136 gezelter 3653 The electrostatic and dispersive behavior near the boundary has long
137 gezelter 3665 been a cause for concern when performing simulations of explicitly
138     non-periodic systems. Early work led to the surface constrained soft
139     sphere dipole model (SCSSD)\cite{Warshel1978} in which the surface
140     molecules are fixed in a random orientation representative of the bulk
141     solvent structural properties. Belch {\it et al.}\cite{Belch1985}
142     simulated clusters of TIPS2 water surrounded by a hydrophobic bounding
143     potential. The spherical hydrophobic boundary induced dangling
144     hydrogen bonds at the surface that propagated deep into the cluster,
145     affecting most of molecules in the simulation. This result echoes an
146     earlier study which showed that an extended planar hydrophobic surface
147     caused orientational preference at the surface which extended
148     relatively deep (7 \r{A}) into the liquid simulation
149     cell.\cite{Lee1984} The surface constrained all-atom solvent (SCAAS)
150     model \cite{King1989} improved upon its SCSSD predecessor. The SCAAS
151     model utilizes a polarization constraint which is applied to the
152     surface molecules to maintain bulk-like structure at the cluster
153     surface. A radial constraint is used to maintain the desired bulk
154     density of the liquid. Both constraint forces are applied only to a
155     pre-determined number of the outermost molecules.
156 gezelter 3640
157 gezelter 3665 Beglov and Roux have developed a boundary model in which the hard
158     sphere boundary has a radius that varies with the instantaneous
159     configuration of the solute (and solvent) molecules.\cite{beglov:9050}
160     This model contains a clear pressure and surface tension contribution
161     to the free energy which XXX.
162 gezelter 3640
163 gezelter 3665 \subsection*{Restraining Potentials}
164 gezelter 3653 Restraining {\it potentials} introduce repulsive potentials at the
165     surface of a sphere or other geometry. The solute and any explicit
166 gezelter 3665 solvent are therefore restrained inside the range defined by the
167     external potential. Often the potentials include a weak short-range
168     attraction to maintain the correct density at the boundary. Beglov
169     and Roux have also introduced a restraining boundary potential which
170     relaxes dynamically depending on the solute geometry and the force the
171     explicit system exerts on the shell.\cite{Beglov:1995fk}
172 gezelter 3653
173 gezelter 3665 Recently, Krilov {\it et al.} introduced a {\it flexible} boundary
174     model that uses a Lennard-Jones potential between the solvent
175     molecules and a boundary which is determined dynamically from the
176     position of the nearest solute atom.\cite{LiY._jp046852t,Zhu:xw} This
177     approach allows the confining potential to prevent solvent molecules
178     from migrating too far from the solute surface, while providing a weak
179     attractive force pulling the solvent molecules towards a fictitious
180     bulk solvent. Although this approach is appealing and has physical
181     motivation, nanoparticles do not deform far from their original
182     geometries even at temperatures which vaporize the nearby solvent. For
183     the systems like this, the flexible boundary model will be nearly
184 gezelter 3653 identical to a fixed-volume restraining potential.
185    
186 gezelter 3665 \subsection*{Hull methods}
187 gezelter 3653 The approach of Kohanoff, Caro, and Finnis is the most promising of
188     the methods for introducing both constant pressure and temperature
189     into non-periodic simulations.\cite{Kohanoff:2005qm,Baltazar:2006ru}
190     This method is based on standard Langevin dynamics, but the Brownian
191     or random forces are allowed to act only on peripheral atoms and exert
192     force in a direction that is inward-facing relative to the facets of a
193     closed bounding surface. The statistical distribution of the random
194     forces are uniquely tied to the pressure in the external reservoir, so
195     the method can be shown to sample the isobaric-isothermal ensemble.
196     Kohanoff {\it et al.} used a Delaunay tessellation to generate a
197     bounding surface surrounding the outermost atoms in the simulated
198     system. This is not the only possible triangulated outer surface, but
199     guarantees that all of the random forces point inward towards the
200     cluster.
201    
202     In the following sections, we extend and generalize the approach of
203     Kohanoff, Caro, and Finnis. The new method, which we are calling the
204     ``Langevin Hull'' applies the external pressure, Langevin drag, and
205 gezelter 3667 random forces on the {\it facets of the hull} instead of the atomic
206     sites comprising the vertices of the hull. This allows us to decouple
207     the external pressure contribution from the drag and random force.
208     The methodology is introduced in section \ref{sec:meth}, tests on
209     crystalline nanoparticles, liquid clusters, and heterogeneous mixtures
210     are detailed in section \ref{sec:tests}. Section \ref{sec:discussion}
211     summarizes our findings.
212 gezelter 3653
213 gezelter 3640 \section{Methodology}
214 gezelter 3653 \label{sec:meth}
215 gezelter 3640
216 gezelter 3665 The Langevin Hull uses an external bath at a fixed constant pressure
217     ($P$) and temperature ($T$). This bath interacts only with the
218     objects on the exterior hull of the system. Defining the hull of the
219     simulation is done in a manner similar to the approach of Kohanoff,
220     Caro and Finnis.\cite{Kohanoff:2005qm} That is, any instantaneous
221     configuration of the atoms in the system is considered as a point
222     cloud in three dimensional space. Delaunay triangulation is used to
223     find all facets between coplanar
224     neighbors.\cite{delaunay,springerlink:10.1007/BF00977785} In highly
225     symmetric point clouds, facets can contain many atoms, but in all but
226     the most symmetric of cases the facets are simple triangles in 3-space
227     that contain exactly three atoms.
228 gezelter 3640
229 gezelter 3652 The convex hull is the set of facets that have {\it no concave
230 gezelter 3665 corners} at an atomic site.\cite{Barber96,EDELSBRUNNER:1994oq} This
231     eliminates all facets on the interior of the point cloud, leaving only
232     those exposed to the bath. Sites on the convex hull are dynamic; as
233     molecules re-enter the cluster, all interactions between atoms on that
234     molecule and the external bath are removed. Since the edge is
235     determined dynamically as the simulation progresses, no {\it a priori}
236     geometry is defined. The pressure and temperature bath interacts only
237 gezelter 3660 with the atoms on the edge and not with atoms interior to the
238     simulation.
239 gezelter 3640
240 gezelter 3662 \begin{figure}
241     \includegraphics[width=\linewidth]{hullSample}
242     \caption{The external temperature and pressure bath interacts only
243     with those atoms on the convex hull (grey surface). The hull is
244 gezelter 3667 computed dynamically at each time step, and molecules can move
245     between the interior (Newtonian) region and the Langevin hull.}
246 gezelter 3662 \label{fig:hullSample}
247     \end{figure}
248    
249 gezelter 3665 Atomic sites in the interior of the simulation move under standard
250 gezelter 3660 Newtonian dynamics,
251 gezelter 3640 \begin{equation}
252 gezelter 3652 m_i \dot{\mathbf v}_i(t)=-{\mathbf \nabla}_i U,
253     \label{eq:Newton}
254 gezelter 3640 \end{equation}
255 gezelter 3652 where $m_i$ is the mass of site $i$, ${\mathbf v}_i(t)$ is the
256     instantaneous velocity of site $i$ at time $t$, and $U$ is the total
257     potential energy. For atoms on the exterior of the cluster
258     (i.e. those that occupy one of the vertices of the convex hull), the
259     equation of motion is modified with an external force, ${\mathbf
260     F}_i^{\mathrm ext}$,
261 gezelter 3640 \begin{equation}
262 gezelter 3652 m_i \dot{\mathbf v}_i(t)=-{\mathbf \nabla}_i U + {\mathbf F}_i^{\mathrm ext}.
263 gezelter 3640 \end{equation}
264    
265 gezelter 3665 The external bath interacts indirectly with the atomic sites through
266     the intermediary of the hull facets. Since each vertex (or atom)
267     provides one corner of a triangular facet, the force on the facets are
268     divided equally to each vertex. However, each vertex can participate
269     in multiple facets, so the resultant force is a sum over all facets
270     $f$ containing vertex $i$:
271 gezelter 3640 \begin{equation}
272     {\mathbf F}_{i}^{\mathrm ext} = \sum_{\begin{array}{c}\mathrm{facets\
273     } f \\ \mathrm{containing\ } i\end{array}} \frac{1}{3}\ {\mathbf
274     F}_f^{\mathrm ext}
275     \end{equation}
276    
277 gezelter 3652 The external pressure bath applies a force to the facets of the convex
278     hull in direct proportion to the area of the facet, while the thermal
279 gezelter 3660 coupling depends on the solvent temperature, viscosity and the size
280     and shape of each facet. The thermal interactions are expressed as a
281     standard Langevin description of the forces,
282 gezelter 3640 \begin{equation}
283     \begin{array}{rclclcl}
284     {\mathbf F}_f^{\text{ext}} & = & \text{external pressure} & + & \text{drag force} & + & \text{random force} \\
285     & = & -\hat{n}_f P A_f & - & \Xi_f(t) {\mathbf v}_f(t) & + & {\mathbf R}_f(t)
286     \end{array}
287     \end{equation}
288 gezelter 3665 Here, $A_f$ and $\hat{n}_f$ are the area and (outward-facing) normal
289     vectors for facet $f$, respectively. ${\mathbf v}_f(t)$ is the
290     velocity of the facet centroid,
291 gezelter 3652 \begin{equation}
292     {\mathbf v}_f(t) = \frac{1}{3} \sum_{i=1}^{3} {\mathbf v}_i,
293     \end{equation}
294 gezelter 3660 and $\Xi_f(t)$ is an approximate ($3 \times 3$) resistance tensor that
295     depends on the geometry and surface area of facet $f$ and the
296     viscosity of the fluid. The resistance tensor is related to the
297     fluctuations of the random force, $\mathbf{R}(t)$, by the
298     fluctuation-dissipation theorem,
299 gezelter 3640 \begin{eqnarray}
300     \left< {\mathbf R}_f(t) \right> & = & 0 \\
301     \left<{\mathbf R}_f(t) {\mathbf R}_f^T(t^\prime)\right> & = & 2 k_B T\
302 gezelter 3652 \Xi_f(t)\delta(t-t^\prime).
303     \label{eq:randomForce}
304 gezelter 3640 \end{eqnarray}
305    
306 gezelter 3665 Once the resistance tensor is known for a given facet, a stochastic
307 gezelter 3660 vector that has the properties in Eq. (\ref{eq:randomForce}) can be
308 gezelter 3665 calculated efficiently by carrying out a Cholesky decomposition to
309     obtain the square root matrix of the resistance tensor,
310 gezelter 3652 \begin{equation}
311     \Xi_f = {\bf S} {\bf S}^{T},
312     \label{eq:Cholesky}
313     \end{equation}
314     where ${\bf S}$ is a lower triangular matrix.\cite{Schlick2002} A
315     vector with the statistics required for the random force can then be
316     obtained by multiplying ${\bf S}$ onto a random 3-vector ${\bf Z}$ which
317     has elements chosen from a Gaussian distribution, such that:
318     \begin{equation}
319     \langle {\bf Z}_i \rangle = 0, \hspace{1in} \langle {\bf Z}_i \cdot
320     {\bf Z}_j \rangle = \frac{2 k_B T}{\delta t} \delta_{ij},
321     \end{equation}
322     where $\delta t$ is the timestep in use during the simulation. The
323     random force, ${\bf R}_{f} = {\bf S} {\bf Z}$, can be shown to
324     have the correct properties required by Eq. (\ref{eq:randomForce}).
325 gezelter 3640
326 gezelter 3660 Our treatment of the resistance tensor is approximate. $\Xi$ for a
327     rigid triangular plate would normally be treated as a $6 \times 6$
328 gezelter 3653 tensor that includes translational and rotational drag as well as
329 gezelter 3660 translational-rotational coupling. The computation of resistance
330 gezelter 3653 tensors for rigid bodies has been detailed
331 gezelter 3663 elsewhere,\cite{JoseGarciadelaTorre02012000,Garcia-de-la-Torre:2001wd,GarciadelaTorreJ2002,Sun:2008fk}
332 gezelter 3653 but the standard approach involving bead approximations would be
333     prohibitively expensive if it were recomputed at each step in a
334     molecular dynamics simulation.
335    
336 gezelter 3665 Instead, we are utilizing an approximate resistance tensor obtained by
337     first constructing the Oseen tensor for the interaction of the
338     centroid of the facet ($f$) with each of the subfacets $\ell=1,2,3$,
339 gezelter 3653 \begin{equation}
340 gezelter 3665 T_{\ell f}=\frac{A_\ell}{8\pi\eta R_{\ell f}}\left(I +
341     \frac{\mathbf{R}_{\ell f}\mathbf{R}_{\ell f}^T}{R_{\ell f}^2}\right)
342 gezelter 3653 \end{equation}
343 gezelter 3665 Here, $A_\ell$ is the area of subfacet $\ell$ which is a triangle
344     containing two of the vertices of the facet along with the centroid.
345     $\mathbf{R}_{\ell f}$ is the vector between the centroid of facet $f$
346     and the centroid of sub-facet $\ell$, and $I$ is the ($3 \times 3$)
347     identity matrix. $\eta$ is the viscosity of the external bath.
348 gezelter 3653
349     \begin{figure}
350     \includegraphics[width=\linewidth]{hydro}
351 gezelter 3660 \caption{The resistance tensor $\Xi$ for a facet comprising sites $i$,
352     $j$, and $k$ is constructed using Oseen tensor contributions between
353     the centoid of the facet $f$ and each of the sub-facets ($i,f,j$),
354     ($j,f,k$), and ($k,f,i$). The centroids of the sub-facets are
355     located at $1$, $2$, and $3$, and the area of each sub-facet is
356 gezelter 3653 easily computed using half the cross product of two of the edges.}
357     \label{hydro}
358     \end{figure}
359    
360 gezelter 3665 The tensors for each of the sub-facets are added together, and the
361     resulting matrix is inverted to give a $3 \times 3$ resistance tensor
362     for translations of the triangular facet,
363 gezelter 3653 \begin{equation}
364     \Xi_f(t) =\left[\sum_{i=1}^3 T_{if}\right]^{-1}.
365     \end{equation}
366 gezelter 3667 Note that this treatment ignores rotations (and
367 gezelter 3660 translational-rotational coupling) of the facet. In compact systems,
368     the facets stay relatively fixed in orientation between
369     configurations, so this appears to be a reasonably good approximation.
370    
371 gezelter 3652 We have implemented this method by extending the Langevin dynamics
372 gezelter 3665 integrator in our code, OpenMD.\cite{Meineke2005,openmd} At each
373     molecular dynamics time step, the following process is carried out:
374     \begin{enumerate}
375     \item The standard inter-atomic forces ($\nabla_iU$) are computed.
376 gezelter 3667 \item Delaunay triangulation is carried out using the current atomic
377 gezelter 3665 configuration.
378     \item The convex hull is computed and facets are identified.
379     \item For each facet:
380     \begin{itemize}
381     \item[a.] The force from the pressure bath ($-PA_f\hat{n}_f$) is
382     computed.
383     \item[b.] The resistance tensor ($\Xi_f(t)$) is computed using the
384     viscosity ($\eta$) of the bath.
385     \item[c.] Facet drag ($-\Xi_f(t) \mathbf{v}_f(t)$) forces are
386     computed.
387     \item[d.] Random forces ($\mathbf{R}_f(t)$) are computed using the
388     resistance tensor and the temperature ($T$) of the bath.
389     \end{itemize}
390     \item The facet forces are divided equally among the vertex atoms.
391     \item Atomic positions and velocities are propagated.
392     \end{enumerate}
393     The Delaunay triangulation and computation of the convex hull are done
394     using calls to the qhull library.\cite{Qhull} There is a minimal
395     penalty for computing the convex hull and resistance tensors at each
396     step in the molecular dynamics simulation (roughly 0.02 $\times$ cost
397     of a single force evaluation), and the convex hull is remarkably easy
398     to parallelize on distributed memory machines (see Appendix A).
399 gezelter 3652
400 gezelter 3640 \section{Tests \& Applications}
401 gezelter 3653 \label{sec:tests}
402 gezelter 3640
403 gezelter 3663 To test the new method, we have carried out simulations using the
404     Langevin Hull on: 1) a crystalline system (gold nanoparticles), 2) a
405 gezelter 3665 liquid droplet (SPC/E water),\cite{Berendsen1987} and 3) a
406     heterogeneous mixture (gold nanoparticles in a water droplet). In each
407     case, we have computed properties that depend on the external applied
408     pressure. Of particular interest for the single-phase systems is the
409     isothermal compressibility,
410 gezelter 3660 \begin{equation}
411     \kappa_{T} = -\frac{1}{V} \left ( \frac{\partial V}{\partial P} \right
412     )_{T}.
413     \label{eq:BM}
414     \end{equation}
415    
416     One problem with eliminating periodic boundary conditions and
417     simulation boxes is that the volume of a three-dimensional point cloud
418     is not well-defined. In order to compute the compressibility of a
419     bulk material, we make an assumption that the number density, $\rho =
420 gezelter 3665 \frac{N}{V}$, is uniform within some region of the point cloud. The
421 gezelter 3660 compressibility can then be expressed in terms of the average number
422     of particles in that region,
423     \begin{equation}
424 gezelter 3665 \kappa_{T} = -\frac{1}{N} \left ( \frac{\partial N}{\partial P} \right
425 gezelter 3660 )_{T}
426     \label{eq:BMN}
427     \end{equation}
428 gezelter 3663 The region we used is a spherical volume of 10 \AA\ radius centered in
429     the middle of the cluster. $N$ is the average number of molecules
430     found within this region throughout a given simulation. The geometry
431     and size of the region is arbitrary, and any bulk-like portion of the
432 gezelter 3665 cluster can be used to compute the compressibility.
433 gezelter 3660
434 gezelter 3665 One might assume that the volume of the convex hull could simply be
435     taken as the system volume $V$ in the compressibility expression
436     (Eq. \ref{eq:BM}), but this has implications at lower pressures (which
437     are explored in detail in the section on water droplets).
438 gezelter 3660
439 gezelter 3663 The metallic force field in use for the gold nanoparticles is the
440     quantum Sutton-Chen (QSC) model.\cite{PhysRevB.59.3527} In all
441     simulations involving point charges, we utilized damped shifted-force
442     (DSF) electrostatics\cite{Fennell06} which is a variant of the Wolf
443     summation\cite{wolf:8254} that has been shown to provide good forces
444     and torques on molecular models for water in a computationally
445     efficient manner.\cite{Fennell06} The damping parameter ($\alpha$) was
446     set to 0.18 \AA$^{-1}$, and the cutoff radius was set to 12 \AA. The
447     Spohr potential was adopted in depicting the interaction between metal
448     atoms and the SPC/E water molecules.\cite{ISI:000167766600035}
449    
450 gezelter 3667 \subsection{Compressibility of gold nanoparticles}
451 gezelter 3640
452 gezelter 3665 The compressibility is well-known for gold, and it provides a good first
453 gezelter 3663 test of how the method compares to other similar methods.
454    
455 gezelter 3640 \begin{figure}
456 gezelter 3665 \includegraphics[width=\linewidth]{P_T_combined}
457     \caption{Pressure and temperature response of an 18 \AA\ gold
458     nanoparticle initially when first placed in the Langevin Hull
459     ($T_\mathrm{bath}$ = 300K, $P_\mathrm{bath}$ = 4 GPa) and starting
460     from initial conditions that were far from the bath pressure and
461     temperature. The pressure response is rapid, and the thermal
462     equilibration depends on both total surface area and the viscosity
463     of the bath.}
464 gezelter 3640 \label{pressureResponse}
465     \end{figure}
466    
467     \begin{equation}
468     \kappa_T=-\frac{1}{V_{\mathrm{eq}}}\left(\frac{\partial V}{\partial
469     P}\right)
470     \end{equation}
471    
472     \begin{figure}
473     \includegraphics[width=\linewidth]{compress_tb}
474     \caption{Isothermal Compressibility (18 \AA gold nanoparticle)}
475     \label{temperatureResponse}
476     \end{figure}
477    
478     \subsection{Compressibility of SPC/E water clusters}
479    
480 gezelter 3660 Prior molecular dynamics simulations on SPC/E water (both in
481     NVT~\cite{Glattli2002} and NPT~\cite{Motakabbir1990, Pi2009}
482     ensembles) have yielded values for the isothermal compressibility that
483     agree well with experiment.\cite{Fine1973} The results of two
484     different approaches for computing the isothermal compressibility from
485     Langevin Hull simulations for pressures between 1 and 6500 atm are
486     shown in Fig. \ref{fig:compWater} along with compressibility values
487     obtained from both other SPC/E simulations and experiment.
488     Compressibility values from all references are for applied pressures
489     within the range 1 - 1000 atm.
490 kstocke1 3649
491 gezelter 3640 \begin{figure}
492 gezelter 3659 \includegraphics[width=\linewidth]{new_isothermalN}
493 kstocke1 3649 \caption{Compressibility of SPC/E water}
494 gezelter 3660 \label{fig:compWater}
495 gezelter 3640 \end{figure}
496    
497 gezelter 3660 Isothermal compressibility values calculated using the number density
498     (Eq. \ref{eq:BMN}) expression are in good agreement with experimental
499     and previous simulation work throughout the 1 - 1000 atm pressure
500     regime. Compressibilities computed using the Hull volume, however,
501     deviate dramatically from the experimental values at low applied
502     pressures. The reason for this deviation is quite simple; at low
503     applied pressures, the liquid is in equilibrium with a vapor phase,
504     and it is entirely possible for one (or a few) molecules to drift away
505     from the liquid cluster (see Fig. \ref{fig:coneOfShame}). At low
506     pressures, the restoring forces on the facets are very gentle, and
507     this means that the hulls often take on relatively distorted
508     geometries which include large volumes of empty space.
509 kstocke1 3649
510 gezelter 3660 \begin{figure}
511     \includegraphics[width=\linewidth]{flytest2}
512     \caption{At low pressures, the liquid is in equilibrium with the vapor
513     phase, and isolated molecules can detach from the liquid droplet.
514 gezelter 3665 This is expected behavior, but the volume of the convex hull
515     includes large regions of empty space. For this reason,
516 gezelter 3662 compressibilities are computed using local number densities rather
517     than hull volumes.}
518 gezelter 3660 \label{fig:coneOfShame}
519     \end{figure}
520 kstocke1 3649
521 gezelter 3665 At higher pressures, the equilibrium strongly favors the liquid phase,
522     and the hull geometries are much more compact. Because of the
523     liquid-vapor effect on the convex hull, the regional number density
524     approach (Eq. \ref{eq:BMN}) provides more reliable estimates of the
525 gezelter 3667 compressibility.
526 kstocke1 3649
527 gezelter 3665 In both the traditional compressibility formula (Eq. \ref{eq:BM}) and
528     the number density version (Eq. \ref{eq:BMN}), multiple simulations at
529     different pressures must be done to compute the first derivatives. It
530     is also possible to compute the compressibility using the fluctuation
531     dissipation theorem using either fluctuations in the
532     volume,\cite{Debenedetti1986},
533 kstocke1 3649 \begin{equation}
534 gezelter 3665 \kappa_{T} = \frac{\left \langle V^{2} \right \rangle - \left \langle
535     V \right \rangle ^{2}}{V \, k_{B} \, T},
536 kstocke1 3649 \end{equation}
537 gezelter 3665 or, equivalently, fluctuations in the number of molecules within the
538     fixed region,
539     \begin{equation}
540     \kappa_{T} = \frac{\left \langle N^{2} \right \rangle - \left \langle
541     N \right \rangle ^{2}}{N \, k_{B} \, T},
542     \end{equation}
543     Thus, the compressibility of each simulation can be calculated
544     entirely independently from all other trajectories. However, the
545     resulting compressibilities were still as much as an order of
546 gezelter 3667 magnitude larger than the reference values. However, compressibility
547 gezelter 3665 calculation that relies on the hull volume will suffer these effects.
548     WE NEED MORE HERE.
549 kstocke1 3649
550     \subsection{Molecular orientation distribution at cluster boundary}
551    
552 gezelter 3665 In order for non-periodic boundary conditions to be widely applicable,
553     they must be constructed in such a way that they allow a finite system
554 gezelter 3667 to replicate the properties of the bulk. Early non-periodic
555     simulation methods (e.g. hydrophobic boundary potentials) induced
556     spurious orientational correlations deep within the simulated
557     system.\cite{Lee1984,Belch1985} This behavior spawned many methods for
558     fixing and characterizing the effects of artifical boundaries
559     including methods which fix the orientations of a set of edge
560     molecules.\cite{Warshel1978,King1989}
561 kstocke1 3649
562 gezelter 3667 As described above, the Langevin Hull does not require that the
563     orientation of molecules be fixed, nor does it utilize an explicitly
564     hydrophobic boundary, orientational constraint or radial constraint.
565     Therefore, the orientational correlations of the molecules in a water
566     cluster are of particular interest in testing this method. Ideally,
567     the water molecules on the surface of the cluster will have enough
568     mobility into and out of the center of the cluster to maintain a
569     bulk-like orientational distribution in the absence of orientational
570     and radial constraints. However, since the number of hydrogen bonding
571     partners available to molecules on the exterior are limited, it is
572     likely that there will be some effective hydrophobicity of the hull.
573 kstocke1 3649
574 gezelter 3667 To determine the extent of these effects demonstrated by the Langevin
575     Hull, we examined the orientationations exhibited by SPC/E water in a
576     cluster of 1372 molecules at 300 K and at pressures ranging from 1 -
577     1000 atm. The orientational angle of a water molecule is described
578 kstocke1 3649 \begin{equation}
579 gezelter 3640 \cos{\theta}=\frac{\vec{r}_i\cdot\vec{\mu}_i}{|\vec{r}_i||\vec{\mu}_i|}
580     \end{equation}
581 gezelter 3667 where $\vec{r}_{i}$ is the vector between molecule {\it i}'s center of
582     mass and the cluster center of mass and $\vec{\mu}_{i}$ is the vector
583     bisecting the H-O-H angle of molecule {\it i} (See
584     Fig. \ref{fig:coords}).
585 gezelter 3640 \begin{figure}
586 kstocke1 3649 \includegraphics[width=\linewidth]{g_r_theta}
587 gezelter 3667 \caption{Orientation angle of the water molecules relative to the
588     center of the cluster. Bulk-like distributions will result in
589     $\langle \cos \theta \rangle$ values close to zero. If the hull
590     exhibits an overabundance of externally-oriented oxygen sites the
591     average orientation will be negative, while dangling hydrogen sites
592     will result in positive average orientations.}
593     \label{fig:coords}
594 kstocke1 3649 \end{figure}
595    
596 gezelter 3667 Fig. \ref{fig:pAngle} shows the distribution of $\cos{\theta}$ values
597     for molecules in the interior of the cluster (squares) and for
598     molecules included in the convex hull (circles).
599 kstocke1 3649 \begin{figure}
600 gezelter 3640 \includegraphics[width=\linewidth]{pAngle}
601 gezelter 3667 \caption{Distribution of $\cos{\theta}$ values for molecules on the
602     interior of the cluster (squares) and for those participating in the
603     convex hull (circles) at a variety of pressures. The Langevin hull
604     exhibits minor dewetting behavior with exposed oxygen sites on the
605     hull water molecules. The orientational preference for exposed
606     oxygen appears to be independent of applied pressure. }
607     \label{fig:pAngle}
608 gezelter 3640 \end{figure}
609    
610 gezelter 3667 As expected, interior molecules (those not included in the convex
611     hull) maintain a bulk-like structure with a uniform distribution of
612     orientations. Molecules included in the convex hull show a slight
613     preference for values of $\cos{\theta} < 0.$ These values correspond
614     to molecules with oxygen directed toward the exterior of the cluster,
615     forming a dangling hydrogen bond acceptor site.
616 gezelter 3640
617 gezelter 3667 In the absence of an electrostatic contribution from the exterior
618     bath, the orientational distribution of water molecules included in
619     the Langevin Hull will slightly resemble the distribution at a neat
620     water liquid/vapor interface. Previous molecular dynamics simulations
621     of SPC/E water \cite{Taylor1996} have shown that molecules at the
622     liquid/vapor interface favor an orientation where one hydrogen
623     protrudes from the liquid phase. This behavior is demonstrated by
624     experiments \cite{Du1994} \cite{Scatena2001} showing that
625     approximately one-quarter of water molecules at the liquid/vapor
626     interface form dangling hydrogen bonds. The negligible preference
627     shown in these cluster simulations could be removed through the
628     introduction of an implicit solvent model, which would provide the
629     missing electrostatic interactions between the cluster molecules and
630     the surrounding temperature/pressure bath.
631 kstocke1 3649
632 gezelter 3667 The orientational preference exhibited by hull molecules in the
633     Langevin hull is significantly weaker than the preference caused by an
634     explicit hydrophobic bounding potential. Additionally, the Langevin
635     Hull does not require that the orientation of any molecules be fixed
636     in order to maintain bulk-like structure, even at the cluster surface.
637 kstocke1 3649
638 gezelter 3640 \subsection{Heterogeneous nanoparticle / water mixtures}
639    
640 gezelter 3665 \section{Discussion}
641     \label{sec:discussion}
642 gezelter 3640
643 gezelter 3667 The Langevin Hull samples the isobaric-isothermal ensemble for
644     non-periodic systems by coupling the system to an bath characterized
645     by pressure, temperature, and solvent viscosity. This enables the
646     study of heterogeneous systems composed of materials of significantly
647     different compressibilities. Because the boundary is dynamically
648     determined during the simulation and the molecules interacting with
649     the boundary can change, the method and has minimal perturbations on
650     the behavior of molecules at the edges of the simulation. Further
651     work on this method will involve implicit electrostatics at the
652     boundary (which is missing in the current implementation) as well as
653     more sophisticated treatments of the surface geometry (alpha
654     shapes\cite{EDELSBRUNNER:1994oq,EDELSBRUNNER:1995cj} and Tight
655     Cocone\cite{Dey:2003ts}). The non-convex hull geometries are
656     significantly more expensive ($\mathcal{O}(N^2)$) than the convex hull
657     ($\mathcal{O}(N \log N)$), but would enable the use of hull volumes
658     directly in computing the compressibility of the sample.
659    
660 gezelter 3663 \section*{Appendix A: Computing Convex Hulls on Parallel Computers}
661 gezelter 3640
662 gezelter 3666 In order to use the Langevin Hull for simulations on parallel
663     computers, one of the more difficult tasks is to compute the bounding
664     surface, facets, and resistance tensors when the processors have
665     incomplete information about the entire system's topology. Most
666     parallel decomposition methods assign primary responsibility for the
667     motion of an atomic site to a single processor, and we can exploit
668     this to efficiently compute the convex hull for the entire system.
669    
670 gezelter 3667 The basic idea involves splitting the point cloud into
671     spatially-overlapping subsets and computing the convex hulls for each
672     of the subsets. The points on the convex hull of the entire system
673     are all present on at least one of the subset hulls. The algorithm
674     works as follows:
675 gezelter 3666 \begin{enumerate}
676     \item Each processor computes the convex hull for its own atomic sites
677 gezelter 3668 (left panel in Fig. \ref{fig:parallel}).
678 gezelter 3666 \item The Hull vertices from each processor are passed out to all of
679     the processors, and each processor assembles a complete list of hull
680     sites (this is much smaller than the original number of points in
681     the point cloud).
682 gezelter 3668 \item Each processor computes the global convex hull (right panel in
683 gezelter 3667 Fig. \ref{fig:parallel}) using only those points that are the union
684     of sites gathered from all of the subset hulls. Delaunay
685     triangulation is then done to obtain the facets of the global hull.
686 gezelter 3666 \end{enumerate}
687    
688     \begin{figure}
689     \begin{centering}
690 gezelter 3668 \includegraphics[width=\linewidth]{parallel}
691 gezelter 3666 \caption{When the sites are distributed among many nodes for parallel
692     computation, the processors first compute the convex hulls for their
693 gezelter 3668 own sites (dashed lines in left panel). The positions of the sites
694 gezelter 3666 that make up the convex hulls are then communicated to all
695 gezelter 3668 processors (middle panel). The convex hull of the system (solid line in right panel) is the convex hull of the points on the hulls for all
696     processors.}
697     \label{fig:parallel}
698 gezelter 3666 \end{centering}
699     \label{fig:parallel}
700     \end{figure}
701    
702     The individual hull operations scale with
703 gezelter 3667 $\mathcal{O}(\frac{n}{p}\log\frac{n}{p})$ where $n$ is the total
704     number of sites, and $p$ is the number of processors. These local
705     hull operations create a set of $p$ hulls each with approximately
706     $\frac{n}{3pr}$ sites (for a cluster of radius $r$). The worst-case
707     communication cost for using a ``gather'' operation to distribute this
708     information to all processors is $\mathcal{O}( \alpha (p-1) + \frac{n
709     \beta (p-1)}{3 r p^2})$, while the final computation of the system
710     hull scales as $\mathcal{O}(\frac{n}{3r}\log\frac{n}{3r})$.
711 gezelter 3666
712 gezelter 3667 For a large number of atoms on a moderately parallel machine, the
713     total costs are dominated by the computations of the individual hulls,
714     and communication of these hulls to so the Langevin hull sees roughly
715     linear speed-up with increasing processor counts.
716    
717 gezelter 3663 \section*{Acknowledgments}
718 gezelter 3640 Support for this project was provided by the
719     National Science Foundation under grant CHE-0848243. Computational
720     time was provided by the Center for Research Computing (CRC) at the
721     University of Notre Dame.
722    
723     \newpage
724    
725     \bibliography{langevinHull}
726    
727     \end{doublespace}
728     \end{document}