ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/stokes/stokes.tex
Revision: 3782
Committed: Mon Dec 12 19:39:38 2011 UTC (12 years, 9 months ago) by skuang
Content type: application/x-tex
File size: 36821 byte(s)
Log Message:
add a figure, plus a few edits.

File Contents

# Content
1 \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 \title{ENTER TITLE HERE}
32
33 \author{Shenyu Kuang and J. Daniel
34 Gezelter\footnote{Corresponding author. \ Electronic mail: gezelter@nd.edu} \\
35 Department of Chemistry and Biochemistry,\\
36 University of Notre Dame\\
37 Notre Dame, Indiana 46556}
38
39 \date{\today}
40
41 \maketitle
42
43 \begin{doublespace}
44
45 \begin{abstract}
46 We present a new method for introducing stable nonequilibrium
47 velocity and temperature gradients in molecular dynamics simulations
48 of heterogeneous systems. This method conserves the linear momentum
49 and total energy of the system and improves previous Reverse
50 Non-Equilibrium Molecular Dynamics (RNEMD) methods and maintains
51 thermal velocity distributions. It also avoid thermal anisotropy
52 occured in NIVS simulations by using isotropic velocity scaling on
53 the molecules in specific regions of a system. To test the method,
54 we have computed the thermal conductivity and shear viscosity of
55 model liquid systems as well as the interfacial frictions of a
56 series of metal/liquid interfaces.
57
58 \end{abstract}
59
60 \newpage
61
62 %\narrowtext
63
64 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
65 % BODY OF TEXT
66 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
67
68 \section{Introduction}
69 [REFINE LATER, ADD MORE REF.S]
70 Imposed-flux methods in Molecular Dynamics (MD)
71 simulations\cite{MullerPlathe:1997xw} can establish steady state
72 systems with a set applied flux vs a corresponding gradient that can
73 be measured. These methods does not need many trajectories to provide
74 information of transport properties of a given system. Thus, they are
75 utilized in computing thermal and mechanical transfer of homogeneous
76 or bulk systems as well as heterogeneous systems such as liquid-solid
77 interfaces.\cite{kuang:AuThl}
78
79 The Reverse Non-Equilibrium MD (RNEMD) methods adopt constraints that
80 satisfy linear momentum and total energy conservation of a system when
81 imposing fluxes in a simulation. Thus they are compatible with various
82 ensembles, including the micro-canonical (NVE) ensemble, without the
83 need of an external thermostat. The original approaches by
84 M\"{u}ller-Plathe {\it et
85 al.}\cite{MullerPlathe:1997xw,ISI:000080382700030} utilize simple
86 momentum swapping for generating energy/momentum fluxes, which is also
87 compatible with particles of different identities. Although simple to
88 implement in a simulation, this approach can create nonthermal
89 velocity distributions, as discovered by Tenney and
90 Maginn\cite{Maginn:2010}. Furthermore, this approach to kinetic energy
91 transfer between particles of different identities is less efficient
92 when the mass difference between the particles becomes significant,
93 which also limits its application on heterogeneous interfacial
94 systems.
95
96 Recently, we developed a different approach, using Non-Isotropic
97 Velocity Scaling (NIVS) \cite{kuang:164101} algorithm to impose
98 fluxes. Compared to the momentum swapping move, it scales the velocity
99 vectors in two separate regions of a simulated system with respective
100 diagonal scaling matrices. These matrices are determined by solving a
101 set of equations including linear momentum and kinetic energy
102 conservation constraints and target flux satisfaction. This method is
103 able to effectively impose a wide range of kinetic energy fluxes
104 without obvious perturbation to the velocity distributions of the
105 simulated systems, regardless of the presence of heterogeneous
106 interfaces. We have successfully applied this approach in studying the
107 interfacial thermal conductance at metal-solvent
108 interfaces.\cite{kuang:AuThl}
109
110 However, the NIVS approach limits its application in imposing momentum
111 fluxes. Temperature anisotropy can happen under high momentum fluxes,
112 due to the nature of the algorithm. Thus, combining thermal and
113 momentum flux is also difficult to implement with this
114 approach. However, such combination may provide a means to simulate
115 thermal/momentum gradient coupled processes such as freeze
116 desalination. Therefore, developing novel approaches to extend the
117 application of imposed-flux method is desired.
118
119 In this paper, we improve the NIVS method and propose a novel approach
120 to impose fluxes. This approach separate the means of applying
121 momentum and thermal flux with operations in one time step and thus is
122 able to simutaneously impose thermal and momentum flux. Furthermore,
123 the approach retains desirable features of previous RNEMD approaches
124 and is simpler to implement compared to the NIVS method. In what
125 follows, we first present the method to implement the method in a
126 simulation. Then we compare the method on bulk fluids to previous
127 methods. Also, interfacial frictions are computed for a series of
128 interfaces.
129
130 \section{Methodology}
131 Similar to the NIVS method,\cite{kuang:164101} we consider a
132 periodic system divided into a series of slabs along a certain axis
133 (e.g. $z$). The unphysical thermal and/or momentum flux is designated
134 from the center slab to one of the end slabs, and thus the thermal
135 flux results in a lower temperature of the center slab than the end
136 slab, and the momentum flux results in negative center slab momentum
137 with positive end slab momentum (unless these fluxes are set
138 negative). Therefore, the center slab is denoted as ``$c$'', while the
139 end slab as ``$h$''.
140
141 To impose these fluxes, we periodically apply different set of
142 operations on velocities of particles {$i$} within the center slab and
143 those of particles {$j$} within the end slab:
144 \begin{eqnarray}
145 \vec{v}_i & \leftarrow & c\cdot\left(\vec{v}_i - \langle\vec{v}_c
146 \rangle\right) + \left(\langle\vec{v}_c\rangle + \vec{a}_c\right) \\
147 \vec{v}_j & \leftarrow & h\cdot\left(\vec{v}_j - \langle\vec{v}_h
148 \rangle\right) + \left(\langle\vec{v}_h\rangle + \vec{a}_h\right)
149 \end{eqnarray}
150 where $\langle\vec{v}_c\rangle$ and $\langle\vec{v}_h\rangle$ denotes
151 the instantaneous bulk velocity of slabs $c$ and $h$ respectively
152 before an operation is applied. When a momentum flux $\vec{j}_z(\vec{p})$
153 presents, these bulk velocities would have a corresponding change
154 ($\vec{a}_c$ and $\vec{a}_h$ respectively) according to Newton's
155 second law:
156 \begin{eqnarray}
157 M_c \vec{a}_c & = & -\vec{j}_z(\vec{p}) \Delta t \\
158 M_h \vec{a}_h & = & \vec{j}_z(\vec{p}) \Delta t
159 \end{eqnarray}
160 where $M$ denotes total mass of particles within a slab:
161 \begin{eqnarray}
162 M_c & = & \sum_{i = 1}^{N_c} m_i \\
163 M_h & = & \sum_{j = 1}^{N_h} m_j
164 \end{eqnarray}
165 and $\Delta t$ is the interval between two separate operations.
166
167 The above operations already conserve the linear momentum of a
168 periodic system. To further satisfy total energy conservation as well
169 as to impose the thermal flux $J_z$, the following equations are
170 included as well:
171 [SUPPORT INFO OR APPENDIX MIGHT BE NECESSARY TO PUT EXTRA MATH IN]
172 \begin{eqnarray}
173 K_c - J_z\Delta t & = & c^2 (K_c - \frac{1}{2}M_c \langle\vec{v}_c
174 \rangle^2) + \frac{1}{2}M_c (\langle \vec{v}_c \rangle + \vec{a}_c)^2 \\
175 K_h + J_z\Delta t & = & h^2 (K_h - \frac{1}{2}M_h \langle\vec{v}_h
176 \rangle^2) + \frac{1}{2}M_h (\langle \vec{v}_h \rangle + \vec{a}_h)^2
177 \end{eqnarray}
178 where $K_c$ and $K_h$ denotes translational kinetic energy of slabs
179 $c$ and $h$ respectively before an operation is applied. These
180 translational kinetic energy conservation equations are sufficient to
181 ensure total energy conservation, as the operations applied in our
182 method do not change the kinetic energy related to other degrees of
183 freedom or the potential energy of a system, given that its potential
184 energy does not depend on particle velocity.
185
186 The above sets of equations are sufficient to determine the velocity
187 scaling coefficients ($c$ and $h$) as well as $\vec{a}_c$ and
188 $\vec{a}_h$. Note that there are two roots respectively for $c$ and
189 $h$. However, the positive roots (which are closer to 1) are chosen so
190 that the perturbations to a system can be reduced to a minimum. Figure
191 \ref{method} illustrates the implementation sketch of this algorithm
192 in an individual step.
193
194 \begin{figure}
195 \includegraphics[width=\linewidth]{method}
196 \caption{Illustration of the implementation of the algorithm in a
197 single step. Starting from an ideal velocity distribution, the
198 transformation is used to apply the effect of both a thermal and a
199 momentum flux from the ``c'' slab to the ``h'' slab. As the figure
200 shows, thermal distributions can preserve after this operation.}
201 \label{method}
202 \end{figure}
203
204 By implementing these operations at a certain frequency, a steady
205 thermal and/or momentum flux can be applied and the corresponding
206 temperature and/or momentum gradients can be established.
207
208 Compared to the previous NIVS method, this approach is computationally
209 more efficient in that only quadratic equations are involved to
210 determine a set of scaling coefficients, while the NIVS method needs
211 to solve quartic equations. Furthermore, this method implements
212 isotropic scaling of velocities in respective slabs, unlike the NIVS,
213 where an extra criteria function is necessary to choose a set of
214 coefficients that performs a scaling as isotropic as possible. More
215 importantly, separating the means of momentum flux imposing from
216 velocity scaling avoids the underlying cause to thermal anisotropy in
217 NIVS when applying a momentum flux. And later sections will
218 demonstrate that this can improve the performance in shear viscosity
219 simulations.
220
221 The advantages of this approach over the original momentum swapping
222 one lies in its nature of preserve a Maxwell-Boltzmann distribution
223 (mathimatically a Gaussian function). However, because the momentum
224 swapping steps tend to result in a nonthermal distribution, when an
225 imposed flux is relatively large, diffusions from the neighboring
226 slabs could no longer remedy this effect, problematic distributions
227 would be observed. Results in later section will illustrate this
228 effect in more detail.
229
230 \section{Computational Details}
231 The algorithm has been implemented in our MD simulation code,
232 OpenMD\cite{Meineke:2005gd,openmd}. We compare the method with
233 previous RNEMD methods or equilibrium MD (EMD) methods in homogeneous fluids
234 (Lennard-Jones and SPC/E water). And taking advantage of the method,
235 we simulate the interfacial friction of different heterogeneous
236 interfaces (gold-organic solvent and gold-SPC/E water and ice-liquid
237 water).
238
239 \subsection{Simulation Protocols}
240 The systems to be investigated are set up in a orthorhombic simulation
241 cell with periodic boundary conditions in all three dimensions. The
242 $z$ axis of these cells were longer and was set as the gradient axis
243 of temperature and/or momentum. Thus the cells were divided into $N$
244 slabs along this axis, with various $N$ depending on individual
245 system. The $x$ and $y$ axis were usually of the same length in
246 homogeneous systems or close to each other where interfaces
247 presents. In all cases, before introducing a nonequilibrium method to
248 establish steady thermal and/or momentum gradients for further
249 measurements and calculations, canonical ensemble with a Nos\'e-Hoover
250 thermostat\cite{hoover85} and microcanonical ensemble equilibrations
251 were used to prepare systems ready for data
252 collections. Isobaric-isothermal equilibrations are performed before
253 this for SPC/E water systems to reach normal pressure (1 bar), while
254 similar equilibrations are used for interfacial systems to relax the
255 surface tensions.
256
257 While homogeneous fluid systems can be set up with random
258 configurations, our interfacial systems needs extra steps to ensure
259 the interfaces be established properly for computations. The
260 preparation and equilibration of butanethiol covered gold (111)
261 surface and further solvation and equilibration process is described
262 as in reference \cite{kuang:AuThl}.
263
264 As for the ice/liquid water interfaces, the basal surface of ice
265 lattice was first constructed. Hirsch {\it et
266 al.}\cite{doi:10.1021/jp048434u} explored the energetics of ice
267 lattices with different proton orders. We refer to their results and
268 choose the configuration of the lowest energy after geometry
269 optimization as the unit cells of our ice lattices. Although
270 experimental solid/liquid coexistant temperature near normal pressure
271 is 273K, Bryk and Haymet's simulations of ice/liquid water interfaces
272 with different models suggest that for SPC/E, the most stable
273 interface is observed at 225$\pm$5K. Therefore, all our ice/liquid
274 water simulations were carried out under 225K. To have extra
275 protection of the ice lattice during initial equilibration (when the
276 randomly generated liquid phase configuration could release large
277 amount of energy in relaxation), restraints were applied to the ice
278 lattice to avoid inadvertent melting by the heat dissipated from the
279 high enery configurations.
280 [MAY ADD A FIGURE HERE FOR BASAL PLANE]
281
282 \subsection{Force Field Parameters}
283 For comparison of our new method with previous work, we retain our
284 force field parameters consistent with the results we will compare
285 with. The Lennard-Jones fluid used here for argon , and reduced unit
286 results are reported for direct comparison purpose.
287
288 As for our water simulations, SPC/E model is used throughout this work
289 for consistency. Previous work for transport properties of SPC/E water
290 model is available\cite{Bedrov:2000,10.1063/1.3330544,Medina2011} so
291 that unnecessary repetition of previous methods can be avoided.
292
293 The Au-Au interaction parameters in all simulations are described by
294 the quantum Sutton-Chen (QSC) formulation.\cite{PhysRevB.59.3527} The
295 QSC potentials include zero-point quantum corrections and are
296 reparametrized for accurate surface energies compared to the
297 Sutton-Chen potentials.\cite{Chen90} For gold/water interfaces, the
298 Spohr potential was adopted\cite{ISI:000167766600035} to depict
299 Au-H$_2$O interactions.
300
301 The small organic molecules included in our simulations are the Au
302 surface capping agent butanethiol and liquid hexane and toluene. The
303 United-Atom
304 models\cite{TraPPE-UA.thiols,TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes}
305 for these components were used in this work for better computational
306 efficiency, while maintaining good accuracy. We refer readers to our
307 previous work\cite{kuang:AuThl} for further details of these models,
308 as well as the interactions between Au and the above organic molecule
309 components.
310
311 \subsection{Thermal conductivities}
312 When $\vec{j}_z(\vec{p})$ is set to zero and a target $J_z$ is set to
313 impose kinetic energy transfer, the method can be used for thermal
314 conductivity computations. Similar to previous RNEMD methods, we
315 assume linear response of the temperature gradient with respect to the
316 thermal flux in general case. And the thermal conductivity ($\lambda$)
317 can be obtained with the imposed kinetic energy flux and the measured
318 thermal gradient:
319 \begin{equation}
320 J_z = -\lambda \frac{\partial T}{\partial z}
321 \end{equation}
322 Like other imposed-flux methods, the energy flux was calculated using
323 the total non-physical energy transferred (${E_{total}}$) from slab
324 ``c'' to slab ``h'', which is recorded throughout a simulation, and
325 the time for data collection $t$:
326 \begin{equation}
327 J_z = \frac{E_{total}}{2 t L_x L_y}
328 \end{equation}
329 where $L_x$ and $L_y$ denotes the dimensions of the plane in a
330 simulation cell perpendicular to the thermal gradient, and a factor of
331 two in the denominator is present for the heat transport occurs in
332 both $+z$ and $-z$ directions. The temperature gradient
333 ${\langle\partial T/\partial z\rangle}$ can be obtained by a linear
334 regression of the temperature profile, which is recorded during a
335 simulation for each slab in a cell. For Lennard-Jones simulations,
336 thermal conductivities are reported in reduced units
337 (${\lambda^*=\lambda \sigma^2 m^{1/2} k_B^{-1}\varepsilon^{-1/2}}$).
338
339 \subsection{Shear viscosities}
340 Alternatively, the method can carry out shear viscosity calculations
341 by switching off $J_z$. One can specify the vector
342 $\vec{j}_z(\vec{p})$ by choosing the three components
343 respectively. For shear viscosity simulations, $j_z(p_z)$ is usually
344 set to zero. Although for isotropic systems, the direction of
345 $\vec{j}_z(\vec{p})$ on the $xy$ plane should not matter, the ability
346 of arbitarily specifying the vector direction in our method could
347 provide convenience in anisotropic simulations.
348
349 Similar to thermal conductivity computations, linear response of the
350 momentum gradient with respect to the shear stress is assumed, and the
351 shear viscosity ($\eta$) can be obtained with the imposed momentum
352 flux (e.g. in $x$ direction) and the measured gradient:
353 \begin{equation}
354 j_z(p_x) = -\eta \frac{\partial v_x}{\partial z}
355 \end{equation}
356 where the flux is similarly defined:
357 \begin{equation}
358 j_z(p_x) = \frac{P_x}{2 t L_x L_y}
359 \end{equation}
360 with $P_x$ being the total non-physical momentum transferred within
361 the data collection time. Also, the velocity gradient
362 ${\langle\partial v_x/\partial z\rangle}$ can be obtained using linear
363 regression of the $x$ component of the mean velocity, $\langle
364 v_x\rangle$, in each of the bins. For Lennard-Jones simulations, shear
365 viscosities are reported in reduced units
366 (${\eta^*=\eta\sigma^2(\varepsilon m)^{-1/2}}$).
367
368 \subsection{Interfacial friction and Slip length}
369 While the shear stress results in a velocity gradient within bulk
370 fluid phase, its effect at a solid-liquid interface could vary due to
371 the interaction strength between the two phases. The interfacial
372 friction coefficient $\kappa$ is defined to relate the shear stress
373 (e.g. along $x$-axis) and the relative fluid velocity tangent to the
374 interface:
375 \begin{equation}
376 j_z(p_x)|_{interface} = \kappa\Delta v_x|_{interface}
377 \end{equation}
378 Under ``stick'' boundary condition, $\Delta v_x|_{interface}
379 \rightarrow 0$, which leads to $\kappa\rightarrow\infty$. However, for
380 ``slip'' boundary condition at the solid-liquid interface, $\kappa$
381 becomes finite. To characterize the interfacial boundary conditions,
382 slip length ($\delta$) is defined using $\kappa$ and the shear
383 viscocity of liquid phase ($\eta$):
384 \begin{equation}
385 \delta = \frac{\eta}{\kappa}
386 \end{equation}
387 so that $\delta\rightarrow 0$ in the ``no-slip'' boundary condition,
388 and depicts how ``slippery'' an interface is. Figure \ref{slipLength}
389 illustrates how this quantity is defined and computed for a
390 solid-liquid interface. [MAY INCLUDE A SNAPSHOT IN FIGURE]
391
392 \begin{figure}
393 \includegraphics[width=\linewidth]{defDelta}
394 \caption{The slip length $\delta$ can be obtained from a velocity
395 profile of a solid-liquid interface simulation. An example of
396 Au/hexane interfaces is shown. Calculation for the left side is
397 illustrated. The right side is similar to the left side.}
398 \label{slipLength}
399 \end{figure}
400
401 In our method, a shear stress can be applied similar to shear
402 viscosity computations by applying an unphysical momentum flux
403 (e.g. $j_z(p_x)$). A corresponding velocity profile can be obtained as
404 shown in Figure \ref{slipLength}, in which the velocity gradients
405 within liquid phase and velocity difference at the liquid-solid
406 interface can be measured respectively. Further calculations and
407 characterizations of the interface can be carried out using these
408 data.
409
410 \section{Results and Discussions}
411 \subsection{Lennard-Jones fluid}
412 Our orthorhombic simulation cell of Lennard-Jones fluid has identical
413 parameters to our previous work\cite{kuang:164101} to facilitate
414 comparison. Thermal conductivitis and shear viscosities were computed
415 with the algorithm applied to the simulations. The results of thermal
416 conductivity are compared with our previous NIVS algorithm. However,
417 since the NIVS algorithm could produce temperature anisotropy for
418 shear viscocity computations, these results are instead compared to
419 the momentum swapping approaches. Table \ref{LJ} lists these
420 calculations with various fluxes in reduced units.
421
422 \begin{table*}
423 \begin{minipage}{\linewidth}
424 \begin{center}
425
426 \caption{Thermal conductivity ($\lambda^*$) and shear viscosity
427 ($\eta^*$) (in reduced units) of a Lennard-Jones fluid at
428 ${\langle T^* \rangle = 0.72}$ and ${\rho^* = 0.85}$ computed
429 at various momentum fluxes. The new method yields similar
430 results to previous RNEMD methods. All results are reported in
431 reduced unit. Uncertainties are indicated in parentheses.}
432
433 \begin{tabular}{cccccc}
434 \hline\hline
435 \multicolumn{2}{c}{Momentum Exchange} &
436 \multicolumn{2}{c}{$\lambda^*$} &
437 \multicolumn{2}{c}{$\eta^*$} \\
438 \hline
439 Swap Interval $t^*$ & Equivalent $J_z^*$ or $j_z^*(p_x)$ &
440 NIVS & This work & Swapping & This work \\
441 \hline
442 0.116 & 0.16 & 7.30(0.10) & 6.25(0.23) & 3.57(0.06) & 3.53(0.16)\\
443 0.232 & 0.09 & 6.95(0.09) & 8.02(0.56) & 3.64(0.05) & 3.43(0.17)\\
444 0.463 & 0.047 & 7.19(0.07) & 6.48(0.15) & 3.52(0.16) & 3.51(0.08)\\
445 0.926 & 0.024 & 7.19(0.28) & 7.02(0.13) & 3.72(0.05) & 3.79(0.11)\\
446 1.158 & 0.019 & 7.98(0.33) & 7.37(0.23) & 3.42(0.06) & 3.53(0.06)\\
447 \hline\hline
448 \end{tabular}
449 \label{LJ}
450 \end{center}
451 \end{minipage}
452 \end{table*}
453
454 \subsubsection{Thermal conductivity}
455 Our thermal conductivity calculations with this method yields
456 comparable results to the previous NIVS algorithm. This indicates that
457 the thermal gradients rendered using this method are also close to
458 previous RNEMD methods. Simulations with moderately higher thermal
459 fluxes tend to yield more reliable thermal gradients and thus avoid
460 large errors, while overly high thermal fluxes could introduce side
461 effects such as non-linear temperature gradient response or
462 inadvertent phase transitions.
463
464 Since the scaling operation is isotropic in this method, one does not
465 need extra care to ensure temperature isotropy between the $x$, $y$
466 and $z$ axes, while thermal anisotropy might happen if the criteria
467 function for choosing scaling coefficients does not perform as
468 expected. Furthermore, this method avoids inadvertent concomitant
469 momentum flux when only thermal flux is imposed, which could not be
470 achieved with swapping or NIVS approaches. The thermal energy exchange
471 in swapping ($\vec{p}_i$ in slab ``c'' with $\vec{p}_j$ in slab ``h'')
472 or NIVS (total slab momentum conponemts $P^\alpha$ scaled to $\alpha
473 P^\alpha$) would not obtain this result unless thermal flux vanishes
474 (i.e. $\vec{p}_i=\vec{p}_j$ or $\alpha=1$ which are trivial to apply a
475 thermal flux). In this sense, this method contributes to having
476 minimal perturbation to a simulation while imposing thermal flux.
477
478 \subsubsection{Shear viscosity}
479 Table \ref{LJ} also compares our shear viscosity results with momentum
480 swapping approach. Our calculations show that our method predicted
481 similar values for shear viscosities to the momentum swapping
482 approach, as well as the velocity gradient profiles. Moderately larger
483 momentum fluxes are helpful to reduce the errors of measured velocity
484 gradients and thus the final result. However, it is pointed out that
485 the momentum swapping approach tends to produce nonthermal velocity
486 distributions.\cite{Maginn:2010}
487
488 To examine that temperature isotropy holds in simulations using our
489 method, we measured the three one-dimensional temperatures in each of
490 the slabs (Figure \ref{tempXyz}). Note that the $x$-dimensional
491 temperatures were calculated after subtracting the effects from bulk
492 velocities of the slabs. The one-dimensional temperature profiles
493 showed no observable difference between the three dimensions. This
494 ensures that isotropic scaling automatically preserves temperature
495 isotropy and that our method is useful in shear viscosity
496 computations.
497
498 \begin{figure}
499 \includegraphics[width=\linewidth]{tempXyz}
500 \caption{Unlike the previous NIVS algorithm, the new method does not
501 produce a thermal anisotropy. No temperature difference between
502 different dimensions were observed beyond the magnitude of the error
503 bars. Note that the two ``hotter'' regions are caused by the shear
504 stress, as reported by Tenney and Maginn\cite{Maginn:2010}, but not
505 an effect that only observed in our methods.}
506 \label{tempXyz}
507 \end{figure}
508
509 Furthermore, the velocity distribution profiles are tested by imposing
510 a large shear stress into the simulations. Figure \ref{vDist}
511 demonstrates how our method is able to maintain thermal velocity
512 distributions against the momentum swapping approach even under large
513 imposed fluxes. Previous swapping methods tend to deplete particles of
514 positive velocities in the negative velocity slab (``c'') and vice
515 versa in slab ``h'', where the distributions leave a notch. This
516 problematic profiles become significant when the imposed-flux becomes
517 larger and diffusions from neighboring slabs could not offset the
518 depletion. Simutaneously, abnormal peaks appear corresponding to
519 excessive velocity swapped from the other slab. This nonthermal
520 distributions limit applications of the swapping approach in shear
521 stress simulations. Our method avoids the above problematic
522 distributions by altering the means of applying momentum
523 fluxes. Comparatively, velocity distributions recorded from
524 simulations with our method is so close to the ideal thermal
525 prediction that no observable difference is shown in Figure
526 \ref{vDist}. Conclusively, our method avoids problems happened in
527 previous RNEMD methods and provides a useful means for shear viscosity
528 computations.
529
530 \begin{figure}
531 \includegraphics[width=\linewidth]{velDist}
532 \caption{Velocity distributions that develop under the swapping and
533 our methods at high flux. These distributions were obtained from
534 Lennard-Jones simulations with $j_z^*(p_x)\sim 0.4$ (equivalent to a
535 swapping interval of 20 time steps). This is a relatively large flux
536 to demonstrate the nonthermal distributions that develop under the
537 swapping method. Distributions produced by our method are very close
538 to the ideal thermal situations.}
539 \label{vDist}
540 \end{figure}
541
542 \subsection{Bulk SPC/E water}
543 Since our method was in good performance of thermal conductivity and
544 shear viscosity computations for simple Lennard-Jones fluid, we extend
545 our applications of these simulations to complex fluid like SPC/E
546 water model. A simulation cell with 1000 molecules was set up in the
547 same manner as in \cite{kuang:164101}. For thermal conductivity
548 simulations, measurements were taken to compare with previous RNEMD
549 methods; for shear viscosity computations, simulations were run under
550 a series of temperatures (with corresponding pressure relaxation using
551 the isobaric-isothermal ensemble[CITE NIVS REF 32]), and results were
552 compared to available data from Equilibrium MD methods[CITATIONS].
553
554 \subsubsection{Thermal conductivity}
555 Table \ref{spceThermal} summarizes our thermal conductivity
556 computations under different temperatures and thermal gradients, in
557 comparison to the previous NIVS results\cite{kuang:164101} and
558 experimental measurements\cite{WagnerKruse}. Note that no appreciable
559 drift of total system energy or temperature was observed when our
560 method is applied, which indicates that our algorithm conserves total
561 energy even for systems involving electrostatic interactions.
562
563 Measurements using our method established similar temperature
564 gradients to the previous NIVS method. Our simulation results are in
565 good agreement with those from previous simulations. And both methods
566 yield values in reasonable agreement with experimental
567 values. Simulations using moderately higher thermal gradient or those
568 with longer gradient axis ($z$) for measurement seem to have better
569 accuracy, from our results.
570
571 \begin{table*}
572 \begin{minipage}{\linewidth}
573 \begin{center}
574
575 \caption{Thermal conductivity of SPC/E water under various
576 imposed thermal gradients. Uncertainties are indicated in
577 parentheses.}
578
579 \begin{tabular}{ccccc}
580 \hline\hline
581 $\langle T\rangle$ & $\langle dT/dz\rangle$ & \multicolumn{3}{c}
582 {$\lambda (\mathrm{W m}^{-1} \mathrm{K}^{-1})$} \\
583 (K) & (K/\AA) & This work & Previous NIVS\cite{kuang:164101} &
584 Experiment\cite{WagnerKruse} \\
585 \hline
586 300 & 0.8 & 0.815(0.027) & 0.770(0.008) & 0.61 \\
587 318 & 0.8 & 0.801(0.024) & 0.750(0.032) & 0.64 \\
588 & 1.6 & 0.766(0.007) & 0.778(0.019) & \\
589 & 0.8 & 0.786(0.009)\footnote{Simulation with $L_z$
590 twice as long.} & & \\
591 \hline\hline
592 \end{tabular}
593 \label{spceThermal}
594 \end{center}
595 \end{minipage}
596 \end{table*}
597
598 \subsubsection{Shear viscosity}
599 The improvement our method achieves for shear viscosity computations
600 enables us to apply it on SPC/E water models. The series of
601 temperatures under which our shear viscosity calculations were carried
602 out covers the liquid range under normal pressure. Our simulations
603 predict a similar trend of $\eta$ vs. $T$ to EMD results we refer to
604 (Table \ref{spceShear}). Considering subtlties such as temperature or
605 pressure/density errors in these two series of measurements, our
606 results show no significant difference from those with EMD
607 methods. Since each value reported using our method takes only one
608 single trajectory of simulation, instead of average from many
609 trajectories when using EMD, our method provides an effective means
610 for shear viscosity computations.
611
612 \begin{table*}
613 \begin{minipage}{\linewidth}
614 \begin{center}
615
616 \caption{Computed shear viscosity of SPC/E water under different
617 temperatures. Results are compared to those obtained with EMD
618 method[CITATION]. Uncertainties are indicated in parentheses.}
619
620 \begin{tabular}{cccc}
621 \hline\hline
622 $T$ & $\partial v_x/\partial z$ & \multicolumn{2}{c}
623 {$\eta (\mathrm{mPa}\cdot\mathrm{s})$} \\
624 (K) & (10$^{10}$s$^{-1}$) & This work & Previous simulations[CITATION]\\
625 \hline
626 273 & & 1.218(0.004) & \\
627 & & 1.140(0.012) & \\
628 303 & & 0.646(0.008) & \\
629 318 & & 0.536(0.007) & \\
630 & & 0.510(0.007) & \\
631 & & & \\
632 333 & & 0.428(0.002) & \\
633 363 & & 0.279(0.014) & \\
634 & & 0.306(0.001) & \\
635 \hline\hline
636 \end{tabular}
637 \label{spceShear}
638 \end{center}
639 \end{minipage}
640 \end{table*}
641
642 [MAY COMBINE JZPX AND JZKE TO RUN ONE JOB BUT GET ETA=F(T)]
643 [PUT RESULTS AND FIGURE HERE IF IT WORKS]
644 \subsection{Interfacial frictions and slip lengths}
645 An attractive aspect of our method is the ability to apply momentum
646 and/or thermal flux in nonhomogeneous systems, where molecules of
647 different identities (or phases) are segregated in different
648 regions. We have previously studied the interfacial thermal transport
649 of a series of metal gold-liquid
650 surfaces\cite{kuang:164101,kuang:AuThl}, and attemptions have been
651 made to investigate the relationship between this phenomenon and the
652 interfacial frictions.
653
654 Table \ref{etaKappaDelta} includes these computations and previous
655 calculations of corresponding interfacial thermal conductance. For
656 bare Au(111) surfaces, slip boundary conditions were observed for both
657 organic and aqueous liquid phases, corresponding to previously
658 computed low interfacial thermal conductance. Instead, the butanethiol
659 covered Au(111) surface appeared to be sticky to the organic liquid
660 molecules in our simulations. We have reported conductance enhancement
661 effect for this surface capping agent,\cite{kuang:AuThl} and these
662 observations have a qualitative agreement with the thermal conductance
663 results. This agreement also supports discussions on the relationship
664 between surface wetting and slip effect and thermal conductance of the
665 interface.[CITE BARRAT, GARDE]
666
667 \begin{table*}
668 \begin{minipage}{\linewidth}
669 \begin{center}
670
671 \caption{Computed interfacial friction coefficient values for
672 interfaces with various components for liquid and solid
673 phase. Error estimates are indicated in parentheses.}
674
675 \begin{tabular}{llcccccc}
676 \hline\hline
677 Solid & Liquid & $T$ & $j_z(p_x)$ & $\eta_{liquid}$ & $\kappa$
678 & $\delta$ & $G$\footnote{References \cite{kuang:AuThl} and
679 \cite{kuang:164101}.} \\
680 surface & molecules & K & MPa & mPa$\cdot$s & Pa$\cdot$s/m & nm
681 & MW/m$^2$/K \\
682 \hline
683 Au(111) & hexane & 200 & 1.08 & 0.20() & 5.3$\times$10$^4$() &
684 3.7 & 46.5 \\
685 & & & 2.15 & 0.14() & 5.3$\times$10$^4$() &
686 2.7 & \\
687 Au-SC$_4$H$_9$ & hexane & 200 & 2.16 & 0.29() & $\infty$ & 0 &
688 131 \\
689 & & & 5.39 & 0.32() & $\infty$ & 0 &
690 \\
691 \hline
692 Au(111) & toluene & 200 & 1.08 & 0.72() & 1.?$\times$10$^5$() &
693 4.6 & 70.1 \\
694 & & & 2.16 & 0.54() & 1.?$\times$10$^5$() &
695 4.9 & \\
696 Au-SC$_4$H$_9$ & toluene & 200 & 5.39 & 0.98() & $\infty$ & 0
697 & 187 \\
698 & & & 10.8 & 0.99() & $\infty$ & 0
699 & \\
700 \hline
701 Au(111) & water & 300 & 1.08 & 0.40() & 1.9$\times$10$^4$() &
702 20.7 & 1.65 \\
703 & & & 2.16 & 0.79() & 1.9$\times$10$^4$() &
704 41.9 & \\
705 \hline
706 ice(basal) & water & 225 & 19.4 & 15.8() & $\infty$ & 0 & \\
707 \hline\hline
708 \end{tabular}
709 \label{etaKappaDelta}
710 \end{center}
711 \end{minipage}
712 \end{table*}
713
714 An interesting effect alongside the surface friction change is
715 observed on the shear viscosity of liquids in the regions close to the
716 solid surface. Note that $\eta$ measured near a ``slip'' surface tends
717 to be smaller than that near a ``stick'' surface. This suggests that
718 an interface could affect the dynamic properties on its neighbor
719 regions. It is known that diffusions of solid particles in liquid
720 phase is affected by their surface conditions (stick or slip
721 boundary).[CITE SCHMIDT AND SKINNER] Our observations could provide
722 support to this phenomenon.
723
724 In addition to these previously studied interfaces, we attempt to
725 construct ice-water interfaces and the basal plane of ice lattice was
726 first studied. In contrast to the Au(111)/water interface, where the
727 friction coefficient is relatively small and large slip effect
728 presents, the ice/liquid water interface demonstrates strong
729 interactions and appears to be sticky. The supercooled liquid phase is
730 an order of magnitude viscous than measurements in previous
731 section. It would be of interst to investigate the effect of different
732 ice lattice planes (such as prism surface) on interfacial friction and
733 corresponding liquid viscosity.
734
735 \section{Conclusions}
736 Our simulations demonstrate the validity of our method in RNEMD
737 computations of thermal conductivity and shear viscosity in atomic and
738 molecular liquids. Our method maintains thermal velocity distributions
739 and avoids thermal anisotropy in previous NIVS shear stress
740 simulations, as well as retains attractive features of previous RNEMD
741 methods. There is no {\it a priori} restrictions to the method to be
742 applied in various ensembles, so prospective applications to
743 extended-system methods are possible.
744
745 Furthermore, using this method, investigations can be carried out to
746 characterize interfacial interactions. Our method is capable of
747 effectively imposing both thermal and momentum flux accross an
748 interface and thus facilitates studies that relates dynamic property
749 measurements to the chemical details of an interface.
750
751 Another attractive feature of our method is the ability of
752 simultaneously imposing thermal and momentum flux in a
753 system. potential researches that might be benefit include complex
754 systems that involve thermal and momentum gradients. For example, the
755 Soret effects under a velocity gradient would be of interest to
756 purification and separation researches.
757
758 \section{Acknowledgments}
759 Support for this project was provided by the National Science
760 Foundation under grant CHE-0848243. Computational time was provided by
761 the Center for Research Computing (CRC) at the University of Notre
762 Dame.
763
764 \newpage
765
766 \bibliography{stokes}
767
768 \end{doublespace}
769 \end{document}