ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/stokes/stokes.tex
Revision: 3788
Committed: Thu Dec 15 16:19:36 2011 UTC (12 years, 9 months ago) by gezelter
Content type: application/x-tex
File size: 44224 byte(s)
Log Message:

File Contents

# User Rev Content
1 gezelter 3769 \documentclass[11pt]{article}
2     \usepackage{amsmath}
3     \usepackage{amssymb}
4     \usepackage{setspace}
5     \usepackage{endfloat}
6     \usepackage{caption}
7     %\usepackage{tabularx}
8     \usepackage{graphicx}
9     \usepackage{multirow}
10     %\usepackage{booktabs}
11     %\usepackage{bibentry}
12     %\usepackage{mathrsfs}
13     %\usepackage[ref]{overcite}
14     \usepackage[square, comma, sort&compress]{natbib}
15     \usepackage{url}
16     \pagestyle{plain} \pagenumbering{arabic} \oddsidemargin 0.0cm
17     \evensidemargin 0.0cm \topmargin -21pt \headsep 10pt \textheight
18     9.0in \textwidth 6.5in \brokenpenalty=10000
19    
20     % double space list of tables and figures
21     \AtBeginDelayedFloats{\renewcommand{\baselinestretch}{1.66}}
22     \setlength{\abovecaptionskip}{20 pt}
23     \setlength{\belowcaptionskip}{30 pt}
24    
25     %\renewcommand\citemid{\ } % no comma in optional reference note
26     \bibpunct{[}{]}{,}{n}{}{;}
27     \bibliographystyle{achemso}
28    
29     \begin{document}
30    
31 gezelter 3786 \title{Velocity Shearing and Scaling RNEMD: a minimally perturbing
32 gezelter 3788 method for simulating temperature and momentum gradients}
33 gezelter 3769
34     \author{Shenyu Kuang and J. Daniel
35     Gezelter\footnote{Corresponding author. \ Electronic mail: gezelter@nd.edu} \\
36     Department of Chemistry and Biochemistry,\\
37     University of Notre Dame\\
38     Notre Dame, Indiana 46556}
39    
40     \date{\today}
41    
42     \maketitle
43    
44     \begin{doublespace}
45    
46     \begin{abstract}
47 skuang 3779 We present a new method for introducing stable nonequilibrium
48     velocity and temperature gradients in molecular dynamics simulations
49 gezelter 3788 of heterogeneous and interfacial systems. This method conserves both
50     the linear momentum and total energy of the system and improves
51     previous reverse non-equilibrium molecular dynamics (RNEMD) methods
52     while retaining equilibrium thermal velocity distributions in each
53     region of the system. The new method avoids the thermal anisotropy
54 gezelter 3786 produced by previous methods by using isotropic velocity scaling and
55 gezelter 3788 shearing (VSS) on all of the molecules in specific regions. To test
56     the method, we have computed the thermal conductivity and shear
57 skuang 3785 viscosity of model liquid systems as well as the interfacial
58 gezelter 3786 friction coeefficients of a series of solid / liquid interfaces. The
59     method's ability to combine the thermal and momentum gradients
60     allows us to obtain shear viscosity data for a range of temperatures
61     from a single trajectory.
62 gezelter 3769 \end{abstract}
63    
64     \newpage
65    
66     %\narrowtext
67    
68     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
69     % BODY OF TEXT
70     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
71    
72     \section{Introduction}
73 gezelter 3786 One of the standard ways to compute transport coefficients such as the
74     viscosities and thermal conductivities of liquids is to use
75     imposed-flux non-equilibrium molecular dynamics
76     methods.\cite{MullerPlathe:1997xw,ISI:000080382700030,kuang:164101}
77     These methods establish stable non-equilibrium conditions in a
78     simulation box using an applied momentum or thermal flux between
79     different regions of the box. The corresponding temperature or
80     velocity gradients which develop in response to the applied flux is
81     related (via linear response theory) to the transport coefficient of
82     interest. These methods are quite efficient, in that they do not need
83     many trajectories to provide information about transport properties. To
84     date, they have been utilized in computing thermal and mechanical
85     transfer of both homogeneous liquids as well as heterogeneous systems
86     such as solid-liquid
87 skuang 3785 interfaces.\cite{garde:nl2005,garde:PhysRevLett2009,kuang:AuThl}
88 skuang 3770
89 gezelter 3786 The reverse non-equilibrium molecular dynamics (RNEMD) methods utilize
90     additional constraints that ensure conservation of linear momentum and
91     total energy of the system while imposing the desired flux. The RNEMD
92     methods are therefore capable of sampling various thermodynamically
93     relevent ensembles, including the micro-canonical (NVE) ensemble,
94     without resorting to an external thermostat. The original approaches
95     proposed by M\"{u}ller-Plathe {\it et
96 skuang 3771 al.}\cite{MullerPlathe:1997xw,ISI:000080382700030} utilize simple
97 gezelter 3786 momentum swapping moves for generating energy/momentum fluxes. The
98     swapping moves can also be made compatible with particles of different
99     identities. Although the swapping moves are simple to implement in
100     molecular simulations, Tenney and Maginn have shown that they create
101     nonthermal velocity distributions.\cite{Maginn:2010} Furthermore, this
102     approach is not particularly efficient for kinetic energy transfer
103     between particles of different identities, particularly when the mass
104     difference between the particles becomes significant. This problem
105     makes applying swapping-move RNEMD methods on heterogeneous
106     interfacial systems somewhat difficult.
107 skuang 3770
108 gezelter 3786 Recently, we developed a somewhat different approach to applying
109 gezelter 3788 thermal fluxes in RNEMD simulation using a non-isotropic velocity
110     scaling (NIVS) algorithm.\cite{kuang:164101} This algorithm scales all
111 gezelter 3786 atomic velocity vectors in two separate regions of a simulated system
112     using two diagonal scaling matrices. The scaling matrices are
113     determined by solving single quartic equation which includes linear
114     momentum and kinetic energy conservation constraints as well as the
115     target thermal flux between the regions. The NIVS method is able to
116     effectively impose a wide range of kinetic energy fluxes without
117     significant perturbation to the velocity distributions away from ideal
118     Maxwell-Boltzmann distributions, even in the presence of heterogeneous
119     interfaces. We successfully applied this approach in studying the
120     interfacial thermal conductance at chemically-capped metal-solvent
121 skuang 3771 interfaces.\cite{kuang:AuThl}
122 gezelter 3769
123 gezelter 3786 The NIVS approach works very well for preparing stable {\it thermal}
124     gradients. However, as we pointed out in the original paper, it has
125     limited application in imposing {\it linear} momentum fluxes (which
126     are required for measuring shear viscosities). The reason for this is
127     that linear momentum flux was being imposed by scaling random
128     fluctuations of the center of the velocity distributions. Repeated
129     application of the original NIVS approach resulted in temperature
130     anisotropy, i.e. the width of the velocity distributions depended on
131     coordinates perpendicular to the desired gradient direction. For this
132     reason, combining thermal and momentum fluxes was difficult with the
133     original NIVS algorithm. However, combinations of thermal and
134     velocity gradients would provide a means to simulate thermal-linear
135     coupled processes such as Soret effect in liquid flows. Therefore,
136     developing improved approaches to the scaling imposed-flux methods
137     would be useful.
138 gezelter 3769
139 gezelter 3786 In this paper, we improve the RNEMD methods by introducing a novel
140     approach to creating imposed fluxes. This approach separates the
141     shearing and scaling of the velocity distributions in different
142     spatial regions and can apply both transformations within a single
143     time step. The approach retains desirable features of previous RNEMD
144     approaches and is simpler to implement compared to the earlier NIVS
145 gezelter 3788 method. In what follows, we first present the velocity shearing and
146     scaling (VSS) RNEMD method and its implementation in a
147     simulation. Then we compare the VSS-RNEMD method in bulk fluids to
148     previous methods. We also compute interfacial frictions are computed
149     for a series of heterogeneous interfaces.
150 gezelter 3769
151     \section{Methodology}
152 gezelter 3786 In an approach similar to the earlier NIVS method,\cite{kuang:164101}
153     we consider a periodic system which has been divided into a series of
154     slabs along a single axis (e.g. $z$). The unphysical thermal and
155     momentum fluxes are applied from one of the end slabs to the center
156     slab, and thus the thermal flux produces a higher temperature in the
157     center slab than in the end slab, and the momentum flux results in a
158     center slab moving along the positive $y$ axis (see Fig. \ref{cartoon}).
159     The applied fluxes can be set to negative values if the reverse
160     gradients are desired. For convenience the center slab is denoted as
161     the {\it hot} or {\it ``H''} slab, while the end slab is denoted {\it
162     ``C''} (or {\it cold}).
163 skuang 3770
164 gezelter 3786 \begin{figure}
165     \includegraphics[width=\linewidth]{cartoon}
166 gezelter 3788 \caption{The VSS-RNEMD approach imposes unphysical transfer of both
167     linear momentum and kinetic energy between a ``hot'' slab and a
168     ``cold'' slab in the simulation box. The system responds to this
169     imposed flux by generating both momentum and temperature gradients.
170     The slope of the gradients can then be used to compute transport
171     properties (e.g. shear viscosity and thermal conductivity).}
172 gezelter 3786 \label{cartoon}
173     \end{figure}
174    
175     To impose these fluxes, we periodically apply a set of operations on
176     the velocities of particles {$i$} within the cold slab and a separate
177     operation on particles {$j$} within the hot slab.
178 skuang 3770 \begin{eqnarray}
179     \vec{v}_i & \leftarrow & c\cdot\left(\vec{v}_i - \langle\vec{v}_c
180 gezelter 3788 \rangle\right) + \left(\langle\vec{v}_c\rangle +
181     \vec{a}_c\right) \label{eq:xformc} \\
182 skuang 3770 \vec{v}_j & \leftarrow & h\cdot\left(\vec{v}_j - \langle\vec{v}_h
183 gezelter 3788 \rangle\right) + \left(\langle\vec{v}_h\rangle + \vec{a}_h\right) \label{eq:xformh}
184 skuang 3770 \end{eqnarray}
185 gezelter 3786 where $\langle\vec{v}_c\rangle$ and $\langle\vec{v}_h\rangle$ denote
186     the instantaneous average velocity of the molecules within slabs $C$
187     and $H$ respectively. When a momentum flux $\vec{j}_z(\vec{p})$ is
188     present, these slab-averaged velocities also get corresponding
189     incremental changes ($\vec{a}_c$ and $\vec{a}_h$ respectively) that
190     are applied to all particles within each slab. The incremental
191     changes are obtained using Newton's second law:
192 skuang 3770 \begin{eqnarray}
193 gezelter 3786 M_c \vec{a}_c & = & -\vec{j}_z(\vec{p}) \Delta t \label{eq:newton1} \\
194 skuang 3770 M_h \vec{a}_h & = & \vec{j}_z(\vec{p}) \Delta t
195 gezelter 3786 \label{eq:newton2}
196 skuang 3770 \end{eqnarray}
197 skuang 3781 where $M$ denotes total mass of particles within a slab:
198 skuang 3770 \begin{eqnarray}
199     M_c & = & \sum_{i = 1}^{N_c} m_i \\
200     M_h & = & \sum_{j = 1}^{N_h} m_j
201     \end{eqnarray}
202 skuang 3781 and $\Delta t$ is the interval between two separate operations.
203 skuang 3770
204 gezelter 3786 The operations in Eqs. \ref{eq:newton1} and \ref{eq:newton2} already
205     conserve the linear momentum of a periodic system. To further satisfy
206     total energy conservation as well as to impose the thermal flux $J_z$,
207     the following constraint equations must be solved for the two scaling
208     variables $c$ and $h$:
209 skuang 3770 \begin{eqnarray}
210     K_c - J_z\Delta t & = & c^2 (K_c - \frac{1}{2}M_c \langle\vec{v}_c
211     \rangle^2) + \frac{1}{2}M_c (\langle \vec{v}_c \rangle + \vec{a}_c)^2 \\
212     K_h + J_z\Delta t & = & h^2 (K_h - \frac{1}{2}M_h \langle\vec{v}_h
213 gezelter 3786 \rangle^2) + \frac{1}{2}M_h (\langle \vec{v}_h \rangle + \vec{a}_h)^2.
214     \label{constraint}
215 skuang 3770 \end{eqnarray}
216 gezelter 3786 Here $K_c$ and $K_h$ denote the translational kinetic energy of slabs
217     $C$ and $H$ respectively. These conservation equations are sufficient
218     to ensure total energy conservation, as the operations applied in our
219     method do not change the kinetic energy related to orientational
220     degrees of freedom or the potential energy of a system (as long as the
221     potential energy is independent of particle velocity).
222 skuang 3770
223 gezelter 3786 Equations \ref{eq:newton1}-\ref{constraint} are sufficient to
224     determine the velocity scaling coefficients ($c$ and $h$) as well as
225     $\vec{a}_c$ and $\vec{a}_h$. Note that there are usually two roots
226     respectively for $c$ and $h$. However, the positive roots (which are
227     closer to 1) are chosen so that the perturbations to the system are
228     minimal. Figure \ref{method} illustrates the implementation of this
229     algorithm in an individual step.
230 skuang 3770
231 skuang 3772 \begin{figure}
232 gezelter 3786 \centering
233     \includegraphics[width=5in]{method}
234 gezelter 3788 \caption{A single step implementation of the VSS algorithm starts with
235     the velocity distributions for the two slabs in a shearing fluid
236     (solid lines). Equations \ref{eq:xformc} and \ref{eq:xformh} are
237     used to scale and shear velocities in both slabs to mimic the effect
238     of both a thermal and a momentum flux. Gaussian distributions are
239 gezelter 3786 preserved by both the scaling and shearing operations.}
240 skuang 3772 \label{method}
241     \end{figure}
242    
243 gezelter 3786 By implementing these operations at a fixed frequency, stable thermal
244     and momentum fluxes can both be applied and the corresponding
245     temperature and momentum gradients can be established.
246 skuang 3770
247 gezelter 3788 Compared to the previous NIVS method, the VSS-RNEMD approach is
248 gezelter 3786 computationally simpler in that only quadratic equations are involved
249     to determine a set of scaling coefficients, while the NIVS method
250     required solution of quartic equations. Furthermore, this method
251     implements {\it isotropic} scaling of velocities in respective slabs,
252     unlike NIVS, which required extra flexibility in the choice of scaling
253     coefficients to allow for the energy and momentum constraints. Most
254     importantly, separating the linear momentum flux from velocity scaling
255     avoids the underlying cause of the thermal anisotropy in NIVS. In
256     later sections, we demonstrate that this can improve the calculated
257     shear viscosities from RNEMD simulations.
258 gezelter 3769
259 gezelter 3788 The VSS-RNEMD approach has advantages over the original momentum
260 gezelter 3786 swapping in many respects. In the swapping method, the velocity
261     vectors involved are usually very different (or the generated flux
262     would be quite small), thus the swapping tends to create strong
263     perturbations in the neighborhood of the particles involved. In
264 gezelter 3788 comparison, the VSS approach distributes the flux widely to all
265 gezelter 3786 particles in a slab so that perturbations in the flux generating
266     region are minimized. Additionally, momentum swapping steps tend to
267     result in nonthermal velocity distributions when the imposed flux is
268     large and diffusion from the neighboring slabs cannot carry momentum
269     away quickly enough. In comparison, the scaling and shearing moves
270 gezelter 3788 made in the VSS approach preserve the shapes of the equilibrium
271 gezelter 3786 velocity disributions (e.g. Maxwell-Boltzmann). The results presented
272     in later sections will illustrate that this is quite helpful in
273     retaining reasonable thermal distributions in a simulation.
274 gezelter 3769
275 skuang 3772 \section{Computational Details}
276 skuang 3773 The algorithm has been implemented in our MD simulation code,
277 gezelter 3786 OpenMD.\cite{Meineke:2005gd,openmd} We will first compare the method
278     with previous RNEMD methods and equilibrium MD in homogeneous
279     fluids (Lennard-Jones and SPC/E water). We have also used the new
280     method to simulate the interfacial friction of different heterogeneous
281     interfaces (Au (111) with organic solvents, Au(111) with SPC/E water,
282 gezelter 3788 and the ice Ih - liquid water interface).
283 gezelter 3769
284 skuang 3773 \subsection{Simulation Protocols}
285 gezelter 3786 The systems we investigated were set up in orthorhombic simulation
286 skuang 3784 cells with periodic boundary conditions in all three dimensions. The
287 gezelter 3786 $z$-axes of these cells were typically quite long and served as the
288     temperature and/or momentum gradient axes. The cells were evenly
289     divided into $N$ slabs along this axis, with $N$ varying for the
290     individual system. The $x$ and $y$ axes were of similar lengths in all
291     simulations. In all cases, before introducing a nonequilibrium method
292     to establish steady thermal and/or momentum gradients, equilibration
293     simulations were run under the canonical ensemble with a Nos\'e-Hoover
294     thermostat\cite{hoover85} followed by further equilibration using
295     standard constant energy (NVE) conditions. For SPC/E water
296     simulations, isobaric-isothermal equilibrations\cite{melchionna93}
297     were performed before equilibration to reach standard densities at
298     atmospheric pressure (1 bar); for interfacial systems, similar
299     equilibrations with anisotropic box relaxations are used to relax the
300     surface tension in the $xy$ plane.
301 skuang 3772
302 gezelter 3786 While homogeneous fluid systems can be set up with random
303     configurations, interfacial systems are typically prepared with a
304     single crystal face presented perpendicular to the $z$-axis. This
305     crystal face is aligned in the x-y plane of the periodic cell, and the
306     solvent occupies the region directly above and below a crystalline
307     slab. The preparation and equilibration of butanethiol covered
308     Au(111) surfaces, as well as the solvation and equilibration processes
309     used for these interfaces are described in detail in reference
310     \cite{kuang:AuThl}.
311 skuang 3773
312 gezelter 3786 For the ice / liquid water interfaces, the basal surface of ice
313     lattice was first constructed. Hirsch and Ojam\"{a}e
314     \cite{doi:10.1021/jp048434u} explored the energetics of ice lattices
315     with all possible proton ordered configurations. We utilized Hirsch
316     and Ojam\"{a}e's structure 6 ($P2_12_12_1$) which is an orthorhombic
317 gezelter 3788 cell that produces a proton-ordered version of ice Ih. The basal face
318     of ice in this unit cell geometry is the $\{0~0~1\}$ face. Although
319 skuang 3784 experimental solid/liquid coexistant temperature under normal pressure
320 gezelter 3786 are close to 273K, Bryk and Haymet's simulations of ice/liquid water
321     interfaces with different models suggest that for SPC/E, the most
322     stable interface is observed at 225$\pm$5K.\cite{bryk:10258}
323     Therefore, our ice/liquid water simulations were carried out at an
324     average temperature of 225K. Molecular translation and orientational
325     restraints were applied in the early stages of equilibration to
326     prevent melting of the ice slab. These restraints were removed during
327     NVT equilibration, well before data collection was carried out.
328 gezelter 3769
329     \subsection{Force Field Parameters}
330 gezelter 3788 For comparing the VSS-RNEMD method with previous work, we retained
331 skuang 3784 force field parameters consistent with previous simulations. Argon is
332     the Lennard-Jones fluid used here, and its results are reported in
333 gezelter 3786 reduced units for purposes of direct comparison with previous
334     calculations.
335 gezelter 3769
336 gezelter 3786 For our water simulations, we utilized the SPC/E model throughout this
337     work. Previous work for transport properties of SPC/E water model is
338     available\cite{Bedrov:2000,10.1063/1.3330544,Medina2011} so direct
339     comparison with previous calculation methods is possible.
340 gezelter 3769
341 skuang 3774 The Au-Au interaction parameters in all simulations are described by
342     the quantum Sutton-Chen (QSC) formulation.\cite{PhysRevB.59.3527} The
343     QSC potentials include zero-point quantum corrections and are
344 gezelter 3769 reparametrized for accurate surface energies compared to the
345 skuang 3775 Sutton-Chen potentials.\cite{Chen90} For gold/water interfaces, the
346     Spohr potential was adopted\cite{ISI:000167766600035} to depict
347     Au-H$_2$O interactions.
348 gezelter 3769
349 skuang 3784 For our gold/organic liquid interfaces, the small organic molecules
350     included in our simulations are the Au surface capping agent
351 gezelter 3786 butanethiol as well as liquid hexane and liquid toluene. The
352     United-Atom
353 skuang 3774 models\cite{TraPPE-UA.thiols,TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes}
354 gezelter 3786 for these components were used in this work for computational
355 skuang 3774 efficiency, while maintaining good accuracy. We refer readers to our
356     previous work\cite{kuang:AuThl} for further details of these models,
357     as well as the interactions between Au and the above organic molecule
358     components.
359 gezelter 3769
360 skuang 3774 \subsection{Thermal conductivities}
361 gezelter 3786 When the linear momentum flux $\vec{j}_z(\vec{p})$ is set to zero and
362 gezelter 3788 the target $J_z$ is non-zero, VSS-RNEMD imposes kinetic energy
363     transfer between the slabs, which can be used for computation of
364     thermal conductivities. As with other RNEMD methods, we assume that we
365 gezelter 3786 are in the linear response regime of the temperature gradient with
366     respect to the thermal flux. The thermal conductivity ($\lambda$) can
367 gezelter 3788 then be calculated using the imposed kinetic energy flux and the
368     measured thermal gradient:
369 skuang 3775 \begin{equation}
370     J_z = -\lambda \frac{\partial T}{\partial z}
371     \end{equation}
372     Like other imposed-flux methods, the energy flux was calculated using
373     the total non-physical energy transferred (${E_{total}}$) from slab
374 gezelter 3788 {\it C} to slab {\it H}, which was recorded throughout the simulation, and
375 gezelter 3786 the total data collection time $t$:
376 skuang 3775 \begin{equation}
377 gezelter 3786 J_z = \frac{E_{total}}{2 t L_x L_y}.
378 skuang 3775 \end{equation}
379 gezelter 3788 Here, $L_x$ and $L_y$ denote the dimensions of the plane in the
380     simulation cell that is perpendicular to the thermal gradient, and a
381     factor of two in the denominator is necessary as the heat transport
382     occurs in both the $+z$ and $-z$ directions. The average temperature
383     gradient ${\langle\partial T/\partial z\rangle}$ can be obtained by a
384     linear regression of the temperature profile, which is recorded during
385     a simulation for each slab in a cell. For Lennard-Jones simulations,
386 skuang 3775 thermal conductivities are reported in reduced units
387     (${\lambda^*=\lambda \sigma^2 m^{1/2} k_B^{-1}\varepsilon^{-1/2}}$).
388    
389 skuang 3774 \subsection{Shear viscosities}
390 gezelter 3786 Alternatively, when the linear momentum flux $\vec{j}_z(\vec{p})$ is
391     non-zero in either the $x$ or $y$ directions, the method can be used
392     to compute the shear viscosity. For isotropic systems, the direction
393     of $\vec{j}_z(\vec{p})$ on the $xy$ plane should not matter, but the
394     ability to arbitarily specify the vector direction in our method could
395 gezelter 3788 be convenient when working with anisotropic interfaces.
396 skuang 3775
397 gezelter 3786 In a manner similar to the thermal conductivity calculations, a linear
398     response of the momentum gradient with respect to the shear stress is
399     assumed, and the shear viscosity ($\eta$) can be obtained with the
400     imposed momentum flux (e.g. in $x$ direction) and the measured
401     velocity gradient:
402 skuang 3775 \begin{equation}
403     j_z(p_x) = -\eta \frac{\partial v_x}{\partial z}
404     \end{equation}
405     where the flux is similarly defined:
406     \begin{equation}
407     j_z(p_x) = \frac{P_x}{2 t L_x L_y}
408     \end{equation}
409     with $P_x$ being the total non-physical momentum transferred within
410 skuang 3784 the data collection time. Also, the averaged velocity gradient
411 skuang 3775 ${\langle\partial v_x/\partial z\rangle}$ can be obtained using linear
412 skuang 3784 regression of the $x$ component of the mean velocity ($\langle
413     v_x\rangle$) in each of the bins. For Lennard-Jones simulations, shear
414     viscosities are also reported in reduced units
415 skuang 3775 (${\eta^*=\eta\sigma^2(\varepsilon m)^{-1/2}}$).
416    
417 gezelter 3786 Although $J_z$ may be switched off for shear viscosity simulations,
418 gezelter 3788 the VSS-RNEMD method allows the user the ability to {\it
419     simultaneously} impose both a thermal and a momentum flux during a
420     single simulation. This creates configurations with coincident
421     temperature and a velocity gradients. Since the viscosity is
422     generally a function of temperature, the local viscosity depends on
423     the local temperature in the fluid. Therefore, in a single simulation,
424     viscosity at $z$ (corresponding to a certain $T$) can be computed with
425     the applied shear flux and the local velocity gradient (which can be
426     obtained by finite difference approximation). This means that the
427     temperature dependence of the viscosity can be mapped out in only one
428 gezelter 3786 trajectory. Results for shear viscosity computations of SPC/E water
429 gezelter 3788 will demonstrate VSS-RNEMD's efficiency in this respect.
430 skuang 3784
431 gezelter 3786 \subsection{Interfacial friction and slip length}
432     Shear stress creates a velocity gradient within bulk fluid phases, but
433     at solid-liquid interfaces, the effects of a shear stress depend on
434     the molecular details of the interface. The interfacial friction
435     coefficient, $\kappa$, relates the shear stress (e.g. along the
436 gezelter 3788 $x$-axis) with the relative velocity of the fluid tangent to the
437     interface:
438 skuang 3775 \begin{equation}
439 gezelter 3786 j_z(p_x) = \kappa \left[v_x(fluid) - v_x(solid)\right]
440 skuang 3775 \end{equation}
441 gezelter 3786 where $v_x(fluid)$ and $v_x(solid)$ are the velocities measured
442     directly adjacent to the interface. Under ``stick'' boundary
443     condition, $\Delta v_x|_\mathrm{interface} \rightarrow 0$, which leads
444     to $\kappa\rightarrow\infty$. However, for ``slip'' boundary
445     conditions at a solid-liquid interface, $\kappa$ becomes finite. To
446     characterize the interfacial boundary conditions, the slip length,
447     $\delta$, is defined by the ratio of the fluid-phase viscosity to the
448     friction coefficient of the interface:
449 skuang 3775 \begin{equation}
450     \delta = \frac{\eta}{\kappa}
451     \end{equation}
452 gezelter 3788 In ``stick'' boundary conditions, $\delta\rightarrow 0$, so $\delta$
453     is a measure of how ``slippery'' an interface is. Figure
454     \ref{slipLength} illustrates how this quantity is defined and computed
455     for a solid-liquid interface.
456 gezelter 3769
457 skuang 3775 \begin{figure}
458     \includegraphics[width=\linewidth]{defDelta}
459 gezelter 3788 \caption{The slip length ($\delta$) can be obtained from a velocity
460     profile along an axis perpendicular to the interface. The data shown
461     is for an Au / hexane interface -- the crystalline region (Au) is
462     moving as a block (lower dots), while the measured velocity gradient
463     in the hexane phase is discontinuous at the interface.}
464 skuang 3775 \label{slipLength}
465     \end{figure}
466 gezelter 3769
467 gezelter 3788 Since the VSS-RNEMD method can be applied for interfaces as well as
468     for bulk materials, the shear stress is applied in an identical manner
469     to the shear viscosity computations, e.g. by applying an unphysical
470     momentum flux, $j_z(\vec{p})$. With the correct choice of $\vec{p}$
471     in the $x-y$ plane, one can compute friction coefficients and slip
472     lengths for a number of different dragging vectors on a given slab.
473     The corresponding velocity profiles can be obtained as shown in Figure
474 gezelter 3786 \ref{slipLength}, in which the velocity gradients within the liquid
475     phase and the velocity difference at the liquid-solid interface can be
476     easily measured from saved simulation data.
477 skuang 3775
478 skuang 3776 \section{Results and Discussions}
479     \subsection{Lennard-Jones fluid}
480 gezelter 3786 Our orthorhombic simulation cell for the Lennard-Jones fluid has
481 gezelter 3788 identical parameters to previous work to facilitate
482     comparison.\cite{kuang:164101} Thermal conductivities and shear
483     viscosities were computed with the new algorithm applied to the
484     simulations. The results of thermal conductivity are compared with our
485     previous NIVS algorithm. However, since the NIVS algorithm produces
486     temperature anisotropy for shear viscocity computations, these results
487     are instead compared to the previous momentum swapping
488     approaches. Table \ref{LJ} lists these values with various fluxes in
489     reduced units.
490 skuang 3770
491 skuang 3776 \begin{table*}
492     \begin{minipage}{\linewidth}
493     \begin{center}
494 gezelter 3769
495 skuang 3776 \caption{Thermal conductivity ($\lambda^*$) and shear viscosity
496     ($\eta^*$) (in reduced units) of a Lennard-Jones fluid at
497     ${\langle T^* \rangle = 0.72}$ and ${\rho^* = 0.85}$ computed
498     at various momentum fluxes. The new method yields similar
499     results to previous RNEMD methods. All results are reported in
500     reduced unit. Uncertainties are indicated in parentheses.}
501    
502     \begin{tabular}{cccccc}
503     \hline\hline
504     \multicolumn{2}{c}{Momentum Exchange} &
505     \multicolumn{2}{c}{$\lambda^*$} &
506     \multicolumn{2}{c}{$\eta^*$} \\
507     \hline
508 gezelter 3786 Swap Interval & Equivalent $J_z^*$ or $j_z^*(p_x)$ &
509 skuang 3785 NIVS\cite{kuang:164101} & This work & Swapping & This work \\
510 skuang 3776 \hline
511     0.116 & 0.16 & 7.30(0.10) & 6.25(0.23) & 3.57(0.06) & 3.53(0.16)\\
512     0.232 & 0.09 & 6.95(0.09) & 8.02(0.56) & 3.64(0.05) & 3.43(0.17)\\
513     0.463 & 0.047 & 7.19(0.07) & 6.48(0.15) & 3.52(0.16) & 3.51(0.08)\\
514     0.926 & 0.024 & 7.19(0.28) & 7.02(0.13) & 3.72(0.05) & 3.79(0.11)\\
515     1.158 & 0.019 & 7.98(0.33) & 7.37(0.23) & 3.42(0.06) & 3.53(0.06)\\
516     \hline\hline
517     \end{tabular}
518     \label{LJ}
519     \end{center}
520     \end{minipage}
521     \end{table*}
522 gezelter 3769
523 skuang 3776 \subsubsection{Thermal conductivity}
524     Our thermal conductivity calculations with this method yields
525     comparable results to the previous NIVS algorithm. This indicates that
526 skuang 3785 the thermal gradients introduced using this method are also close to
527 skuang 3776 previous RNEMD methods. Simulations with moderately higher thermal
528     fluxes tend to yield more reliable thermal gradients and thus avoid
529     large errors, while overly high thermal fluxes could introduce side
530     effects such as non-linear temperature gradient response or
531     inadvertent phase transitions.
532 gezelter 3769
533 skuang 3776 Since the scaling operation is isotropic in this method, one does not
534     need extra care to ensure temperature isotropy between the $x$, $y$
535 skuang 3785 and $z$ axes, while for NIVS, thermal anisotropy might happen if the
536     criteria function for choosing scaling coefficients does not perform
537     as expected. Furthermore, this method avoids inadvertent concomitant
538 skuang 3776 momentum flux when only thermal flux is imposed, which could not be
539     achieved with swapping or NIVS approaches. The thermal energy exchange
540 skuang 3778 in swapping ($\vec{p}_i$ in slab ``c'' with $\vec{p}_j$ in slab ``h'')
541 skuang 3776 or NIVS (total slab momentum conponemts $P^\alpha$ scaled to $\alpha
542 skuang 3785 P^\alpha$) would not achieve this effect unless thermal flux vanishes
543     (i.e. $\vec{p}_i=\vec{p}_j$ or $\alpha=1$ which do not contribute to
544 gezelter 3786 applying a thermal flux). In this sense, this method aids in achieving
545 skuang 3785 minimal perturbation to a simulation while imposing a thermal flux.
546 skuang 3776
547     \subsubsection{Shear viscosity}
548 skuang 3785 Table \ref{LJ} also compares our shear viscosity results with the
549     momentum swapping approach. Our calculations show that our method
550     predicted similar values of shear viscosities to the momentum swapping
551 skuang 3776 approach, as well as the velocity gradient profiles. Moderately larger
552     momentum fluxes are helpful to reduce the errors of measured velocity
553     gradients and thus the final result. However, it is pointed out that
554     the momentum swapping approach tends to produce nonthermal velocity
555     distributions.\cite{Maginn:2010}
556    
557     To examine that temperature isotropy holds in simulations using our
558     method, we measured the three one-dimensional temperatures in each of
559     the slabs (Figure \ref{tempXyz}). Note that the $x$-dimensional
560 skuang 3785 temperatures were calculated after subtracting the contribution from
561     bulk velocities of the slabs. The one-dimensional temperature profiles
562 skuang 3776 showed no observable difference between the three dimensions. This
563     ensures that isotropic scaling automatically preserves temperature
564     isotropy and that our method is useful in shear viscosity
565     computations.
566    
567 gezelter 3769 \begin{figure}
568 skuang 3776 \includegraphics[width=\linewidth]{tempXyz}
569 gezelter 3788 \caption{Unlike the previous NIVS algorithm, the VSS method does not
570     produce thermal anisotropy under shearing stress. Temperature
571     differences along the different box axes were significantly smaller
572     than the error bars. Note that the two regions of elevated
573     temperature are caused by the shear stress. This is the same
574     frictional heating reported previously by Tenney and
575     Maginn.\cite{Maginn:2010}}
576 skuang 3776 \label{tempXyz}
577 gezelter 3769 \end{figure}
578    
579 skuang 3776 Furthermore, the velocity distribution profiles are tested by imposing
580     a large shear stress into the simulations. Figure \ref{vDist}
581     demonstrates how our method is able to maintain thermal velocity
582     distributions against the momentum swapping approach even under large
583     imposed fluxes. Previous swapping methods tend to deplete particles of
584     positive velocities in the negative velocity slab (``c'') and vice
585 skuang 3785 versa in slab ``h'', where the distributions leave notchs. This
586 skuang 3776 problematic profiles become significant when the imposed-flux becomes
587     larger and diffusions from neighboring slabs could not offset the
588 skuang 3785 depletions. Simutaneously, abnormal peaks appear corresponding to
589     excessive particles having velocity swapped from the other slab. These
590     nonthermal distributions limit applications of the swapping approach
591     in shear stress simulations. Our method avoids the above problematic
592 skuang 3776 distributions by altering the means of applying momentum
593     fluxes. Comparatively, velocity distributions recorded from
594     simulations with our method is so close to the ideal thermal
595 skuang 3785 prediction that no obvious difference is shown in Figure
596     \ref{vDist}. Conclusively, our method avoids problems that occurs in
597 skuang 3776 previous RNEMD methods and provides a useful means for shear viscosity
598     computations.
599 gezelter 3769
600 skuang 3776 \begin{figure}
601 gezelter 3788 \centering
602     \includegraphics[width=5.5in]{velDist}
603     \caption{Velocity distributions that develop under VSS more closely
604     resemble ideal Maxwell-Boltzmann distributions than earlier
605     momentum-swapping RNEMD approaches. This data was obtained from
606 skuang 3779 Lennard-Jones simulations with $j_z^*(p_x)\sim 0.4$ (equivalent to a
607 skuang 3777 swapping interval of 20 time steps). This is a relatively large flux
608 gezelter 3788 which demonstrates some of the non-thermal velocity distributions
609     that can develop under swapping RNEMD.}
610 skuang 3776 \label{vDist}
611     \end{figure}
612 gezelter 3769
613 skuang 3776 \subsection{Bulk SPC/E water}
614 skuang 3785 We extend our applications of thermal conductivity and shear viscosity
615     computations to a complex fluid model of SPC/E water. A simulation
616     cell with 1000 molecules was set up in the similar manner as in
617     \cite{kuang:164101}. For thermal conductivity simulations,
618     measurements were taken to compare with previous RNEMD methods; for
619     shear viscosity computations, simulations were run under a series of
620     temperatures (with corresponding pressure relaxation using the
621     isobaric-isothermal ensemble\cite{melchionna93}), and results were
622     compared to available data from EMD
623     methods\cite{10.1063/1.3330544,Medina2011}. Besides, a simulation with
624     both thermal and momentum gradient was carried out to map out shear
625     viscosity as a function of temperature to see the effectiveness and
626     accuracy our method could reach.
627 skuang 3777
628 skuang 3776 \subsubsection{Thermal conductivity}
629 skuang 3778 Table \ref{spceThermal} summarizes our thermal conductivity
630     computations under different temperatures and thermal gradients, in
631     comparison to the previous NIVS results\cite{kuang:164101} and
632     experimental measurements\cite{WagnerKruse}. Note that no appreciable
633     drift of total system energy or temperature was observed when our
634     method is applied, which indicates that our algorithm conserves total
635 skuang 3785 energy well for systems involving electrostatic interactions.
636 gezelter 3769
637 skuang 3778 Measurements using our method established similar temperature
638     gradients to the previous NIVS method. Our simulation results are in
639     good agreement with those from previous simulations. And both methods
640     yield values in reasonable agreement with experimental
641     values. Simulations using moderately higher thermal gradient or those
642     with longer gradient axis ($z$) for measurement seem to have better
643     accuracy, from our results.
644    
645     \begin{table*}
646     \begin{minipage}{\linewidth}
647     \begin{center}
648    
649     \caption{Thermal conductivity of SPC/E water under various
650     imposed thermal gradients. Uncertainties are indicated in
651     parentheses.}
652    
653     \begin{tabular}{ccccc}
654     \hline\hline
655     $\langle T\rangle$ & $\langle dT/dz\rangle$ & \multicolumn{3}{c}
656     {$\lambda (\mathrm{W m}^{-1} \mathrm{K}^{-1})$} \\
657     (K) & (K/\AA) & This work & Previous NIVS\cite{kuang:164101} &
658     Experiment\cite{WagnerKruse} \\
659     \hline
660     300 & 0.8 & 0.815(0.027) & 0.770(0.008) & 0.61 \\
661     318 & 0.8 & 0.801(0.024) & 0.750(0.032) & 0.64 \\
662     & 1.6 & 0.766(0.007) & 0.778(0.019) & \\
663 skuang 3779 & 0.8 & 0.786(0.009)\footnote{Simulation with $L_z$
664     twice as long.} & & \\
665 skuang 3778 \hline\hline
666     \end{tabular}
667     \label{spceThermal}
668     \end{center}
669     \end{minipage}
670     \end{table*}
671    
672 skuang 3776 \subsubsection{Shear viscosity}
673 skuang 3778 The improvement our method achieves for shear viscosity computations
674     enables us to apply it on SPC/E water models. The series of
675     temperatures under which our shear viscosity calculations were carried
676     out covers the liquid range under normal pressure. Our simulations
677     predict a similar trend of $\eta$ vs. $T$ to EMD results we refer to
678 skuang 3779 (Table \ref{spceShear}). Considering subtlties such as temperature or
679     pressure/density errors in these two series of measurements, our
680     results show no significant difference from those with EMD
681     methods. Since each value reported using our method takes only one
682     single trajectory of simulation, instead of average from many
683     trajectories when using EMD, our method provides an effective means
684     for shear viscosity computations.
685 gezelter 3769
686 skuang 3778 \begin{table*}
687     \begin{minipage}{\linewidth}
688     \begin{center}
689    
690     \caption{Computed shear viscosity of SPC/E water under different
691     temperatures. Results are compared to those obtained with EMD
692     method[CITATION]. Uncertainties are indicated in parentheses.}
693    
694     \begin{tabular}{cccc}
695     \hline\hline
696     $T$ & $\partial v_x/\partial z$ & \multicolumn{2}{c}
697     {$\eta (\mathrm{mPa}\cdot\mathrm{s})$} \\
698 skuang 3785 (K) & (10$^{10}$s$^{-1}$) & This work & Previous
699     simulations\cite{Medina2011} \\
700 skuang 3778 \hline
701 skuang 3785 273 & 1.12 & 1.218(0.004) & 1.282(0.048) \\
702     & 1.79 & 1.140(0.012) & \\
703     303 & 2.09 & 0.646(0.008) & 0.643(0.019) \\
704     318 & 2.50 & 0.536(0.007) & \\
705     & 5.25 & 0.510(0.007) & \\
706     & 2.82 & 0.474(0.003)\footnote{Simulation with $L_z$ twice
707     as long.} & \\
708     333 & 3.10 & 0.428(0.002) & 0.421(0.008) \\
709     363 & 2.34 & 0.279(0.014) & 0.291(0.005) \\
710     & 4.26 & 0.306(0.001) & \\
711 skuang 3778 \hline\hline
712     \end{tabular}
713     \label{spceShear}
714     \end{center}
715     \end{minipage}
716     \end{table*}
717 gezelter 3769
718 skuang 3785 A more effective way to map out $\eta$ vs $T$ is to combine a momentum
719     flux with a thermal flux. Figure \ref{Tvxdvdz} shows the thermal and
720     velocity gradient in one such simulation. At different positions with
721     different temperatures, the velocity gradient is not a constant but
722     can be computed locally. With the data provided in Figure
723     \ref{Tvxdvdz}, a series of $\eta$ is calculated as in Figure
724     \ref{etaT} and a linear fit was performed to $\partial v_x/\partial z$
725     vs. $z$ so that the resulted $\eta$ can be present as a curve as
726     well. For comparison, other results are also mapped in the figure.
727    
728     \begin{figure}
729 gezelter 3788 \centering
730     \includegraphics[width=5.5in]{tvxdvdz}
731 skuang 3785 \caption{With a combination of a thermal and a momentum flux, a
732 gezelter 3788 simulation can exhibit both a temperature (top) and a velocity
733     (middle) gradient. $\partial v_x/\partial z$ depends on the local
734     temperature, so it varies along the gradient axis. This derivative
735     can be computed using finite difference approximations (lower) and
736     can be used to calculate $\eta$ vs $T$ (Figure \ref{etaT}).}
737 skuang 3785 \label{Tvxdvdz}
738     \end{figure}
739    
740     From Figure \ref{etaT}, one can see that the generated curve agrees
741     well with the above RNEMD simulations at different temperatures, as
742     well as results reported using EMD
743     methods\cite{10.1063/1.3330544,Medina2011} in much of the temperature
744     range simulated. However, this curve has relatively large error in
745     lower temperature regions and has some difference in predicting $\eta$
746     near 273K. Provided that this curve only takes one trajectory to
747     generate, these results are of satisfactory efficiency and
748     accuracy. Since previous work already pointed out that the SPC/E model
749     tends to predict lower viscosity compared to experimental
750     data,\cite{Medina2011} experimental comparison are not given here.
751    
752     \begin{figure}
753     \includegraphics[width=\linewidth]{etaT}
754 gezelter 3788 \caption{The temperature dependence of the shear viscosity generated
755     by a {\it single} VSS-RNEMD simulation. Applying both thermal and
756     momentum gradient predicts reasonable values in much of the
757     temperature range being tested.}
758 skuang 3785 \label{etaT}
759     \end{figure}
760    
761 skuang 3779 \subsection{Interfacial frictions and slip lengths}
762 skuang 3785 Another attractive aspect of our method is the ability to apply
763     momentum and/or thermal flux in nonhomogeneous systems, where
764     molecules of different identities (or phases) are segregated in
765     different regions. We have previously studied the interfacial thermal
766     transport of a series of metal gold-liquid
767     surfaces\cite{kuang:164101,kuang:AuThl}, and would like to further
768     investigate the relationship between this phenomenon and the
769 skuang 3779 interfacial frictions.
770 gezelter 3769
771 skuang 3779 Table \ref{etaKappaDelta} includes these computations and previous
772     calculations of corresponding interfacial thermal conductance. For
773     bare Au(111) surfaces, slip boundary conditions were observed for both
774     organic and aqueous liquid phases, corresponding to previously
775 skuang 3785 computed low interfacial thermal conductance. In comparison, the
776     butanethiol covered Au(111) surface appeared to be sticky to the
777     organic liquid layers in our simulations. We have reported conductance
778     enhancement effect for this surface capping agent,\cite{kuang:AuThl}
779     and these observations have a qualitative agreement with the thermal
780     conductance results. This agreement also supports discussions on the
781     relationship between surface wetting and slip effect and thermal
782     conductance of the
783     interface.\cite{PhysRevLett.82.4671,doi:10.1080/0026897031000068578,garde:PhysRevLett2009}
784 skuang 3776
785 gezelter 3769 \begin{table*}
786     \begin{minipage}{\linewidth}
787     \begin{center}
788    
789 skuang 3779 \caption{Computed interfacial friction coefficient values for
790     interfaces with various components for liquid and solid
791     phase. Error estimates are indicated in parentheses.}
792 gezelter 3769
793 skuang 3779 \begin{tabular}{llcccccc}
794 gezelter 3769 \hline\hline
795 skuang 3779 Solid & Liquid & $T$ & $j_z(p_x)$ & $\eta_{liquid}$ & $\kappa$
796     & $\delta$ & $G$\footnote{References \cite{kuang:AuThl} and
797     \cite{kuang:164101}.} \\
798 skuang 3785 surface & molecules & K & MPa & mPa$\cdot$s &
799     10$^4$Pa$\cdot$s/m & nm & MW/m$^2$/K \\
800 gezelter 3769 \hline
801 skuang 3785 Au(111) & hexane & 200 & 1.08 & 0.197(0.009) & 5.30(0.36) &
802     3.72 & 46.5 \\
803     & & & 2.15 & 0.141(0.002) & 5.31(0.26) &
804     2.76 & \\
805     Au-SC$_4$H$_9$ & hexane & 200 & 2.16 & 0.286(0.019) & $\infty$
806     & 0 & 131 \\
807     & & & 5.39 & 0.320(0.006) & $\infty$
808     & 0 & \\
809 gezelter 3769 \hline
810 skuang 3785 Au(111) & toluene & 200 & 1.08 & 0.722(0.035) & 15.7(0.7) &
811     4.60 & 70.1 \\
812     & & & 2.16 & 0.544(0.030) & 11.2(0.5) &
813     4.86 & \\
814     Au-SC$_4$H$_9$ & toluene & 200 & 5.39 & 0.980(0.057) &
815     $\infty$ & 0 & 187 \\
816     & & & 10.8 & 0.995(0.005) &
817     $\infty$ & 0 & \\
818 skuang 3779 \hline
819 skuang 3785 Au(111) & water & 300 & 1.08 & 0.399(0.050) & 1.928(0.022) &
820 skuang 3779 20.7 & 1.65 \\
821 skuang 3785 & & & 2.16 & 0.794(0.255) & 1.895(0.003) &
822 skuang 3779 41.9 & \\
823     \hline
824 skuang 3785 ice(basal) & water & 225 & 19.4 & 15.8(0.2) & $\infty$ & 0 & \\
825 gezelter 3769 \hline\hline
826     \end{tabular}
827 skuang 3779 \label{etaKappaDelta}
828 gezelter 3769 \end{center}
829     \end{minipage}
830     \end{table*}
831    
832 skuang 3779 An interesting effect alongside the surface friction change is
833     observed on the shear viscosity of liquids in the regions close to the
834 skuang 3785 solid surface. In our results, $\eta$ measured near a ``slip'' surface
835     tends to be smaller than that near a ``stick'' surface. This may
836     suggest the influence from an interface on the dynamic properties of
837     liquid within its neighbor regions. It is known that diffusions of
838     solid particles in liquid phase is affected by their surface
839     conditions (stick or slip boundary).\cite{10.1063/1.1610442} Our
840     observations could provide a support to this phenomenon.
841 gezelter 3769
842 skuang 3779 In addition to these previously studied interfaces, we attempt to
843     construct ice-water interfaces and the basal plane of ice lattice was
844 skuang 3785 studied here. In contrast to the Au(111)/water interface, where the
845     friction coefficient is substantially small and large slip effect
846 skuang 3779 presents, the ice/liquid water interface demonstrates strong
847 skuang 3785 solid-liquid interactions and appears to be sticky. The supercooled
848     liquid phase is an order of magnitude more viscous than measurements
849     in previous section. It would be of interst to investigate the effect
850     of different ice lattice planes (such as prism and other surfaces) on
851     interfacial friction and the corresponding liquid viscosity.
852 gezelter 3769
853 skuang 3776 \section{Conclusions}
854 skuang 3779 Our simulations demonstrate the validity of our method in RNEMD
855     computations of thermal conductivity and shear viscosity in atomic and
856     molecular liquids. Our method maintains thermal velocity distributions
857     and avoids thermal anisotropy in previous NIVS shear stress
858     simulations, as well as retains attractive features of previous RNEMD
859     methods. There is no {\it a priori} restrictions to the method to be
860     applied in various ensembles, so prospective applications to
861     extended-system methods are possible.
862 gezelter 3769
863 skuang 3785 Our method is capable of effectively imposing thermal and/or momentum
864     flux accross an interface. This facilitates studies that relates
865     dynamic property measurements to the chemical details of an
866     interface. Therefore, investigations can be carried out to
867     characterize interfacial interactions using the method.
868 gezelter 3769
869 skuang 3779 Another attractive feature of our method is the ability of
870 skuang 3785 simultaneously introducing thermal and momentum gradients in a
871     system. This facilitates us to effectively map out the shear viscosity
872     with respect to a range of temperature in single trajectory of
873     simulation with satisafactory accuracy. Complex systems that involve
874     thermal and momentum gradients might potentially benefit from
875     this. For example, the Soret effects under a velocity gradient might
876     be models of interest to purification and separation researches.
877 gezelter 3769
878     \section{Acknowledgments}
879     Support for this project was provided by the National Science
880     Foundation under grant CHE-0848243. Computational time was provided by
881     the Center for Research Computing (CRC) at the University of Notre
882     Dame.
883    
884     \newpage
885    
886 skuang 3770 \bibliography{stokes}
887 gezelter 3769
888     \end{doublespace}
889     \end{document}