ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/trunk/nivsRnemd/nivsRnemd.tex
Revision: 3569
Committed: Tue Mar 9 23:14:21 2010 UTC (14 years, 3 months ago) by gezelter
Content type: application/x-tex
File size: 23157 byte(s)
Log Message:
Added a figure

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{booktabs}
10 %\usepackage{bibentry}
11 %\usepackage{mathrsfs}
12 \usepackage[ref]{overcite}
13 \pagestyle{plain} \pagenumbering{arabic} \oddsidemargin 0.0cm
14 \evensidemargin 0.0cm \topmargin -21pt \headsep 10pt \textheight
15 9.0in \textwidth 6.5in \brokenpenalty=10000
16
17 % double space list of tables and figures
18 \AtBeginDelayedFloats{\renewcommand{\baselinestretch}{1.66}}
19 \setlength{\abovecaptionskip}{20 pt}
20 \setlength{\belowcaptionskip}{30 pt}
21
22 \renewcommand\citemid{\ } % no comma in optional referenc note
23
24 \begin{document}
25
26 \title{A gentler approach to RNEMD: Non-isotropic Velocity Scaling for computing Thermal Conductivity and Shear Viscosity}
27
28 \author{Shenyu Kuang and J. Daniel
29 Gezelter\footnote{Corresponding author. \ Electronic mail: gezelter@nd.edu} \\
30 Department of Chemistry and Biochemistry,\\
31 University of Notre Dame\\
32 Notre Dame, Indiana 46556}
33
34 \date{\today}
35
36 \maketitle
37
38 \begin{doublespace}
39
40 \begin{abstract}
41
42 \end{abstract}
43
44 \newpage
45
46 %\narrowtext
47
48 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49 % BODY OF TEXT
50 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
51
52
53
54 \section{Introduction}
55 The original formulation of Reverse Non-equilibrium Molecular Dynamics
56 (RNEMD) obtains transport coefficients (thermal conductivity and shear
57 viscosity) in a fluid by imposing an artificial momentum flux between
58 two thin parallel slabs of material that are spatially separated in
59 the simulation cell.\cite{MullerPlathe:1997xw,ISI:000080382700030} The
60 artificial flux is typically created by periodically ``swapping'' either
61 the entire momentum vector $\vec{p}$ or single components of this
62 vector ($p_x$) between molecules in each of the two slabs. If the two
63 slabs are separated along the z coordinate, the imposed flux is either
64 directional ($j_z(p_x)$) or isotropic ($J_z$), and the response of a
65 simulated system to the imposed momentum flux will typically be a
66 velocity or thermal gradient. The transport coefficients (shear
67 viscosity and thermal conductivity) are easily obtained by assuming
68 linear response of the system,
69 \begin{eqnarray}
70 j_z(p_x) & = & -\eta \frac{\partial v_x}{\partial z}\\
71 J & = & \lambda \frac{\partial T}{\partial z}
72 \end{eqnarray}
73 RNEMD has been widely used to provide computational estimates of thermal
74 conductivities and shear viscosities in a wide range of materials,
75 from liquid copper to monatomic liquids to molecular fluids
76 (e.g. ionic liquids).\cite{ISI:000246190100032}
77
78 RNEMD is preferable in many ways to the forward NEMD methods because
79 it imposes what is typically difficult to measure (a flux or stress)
80 and it is typically much easier to compute momentum gradients or
81 strains (the response). For similar reasons, RNEMD is also preferable
82 to slowly-converging equilibrium methods for measuring thermal
83 conductivity and shear viscosity (using Green-Kubo relations or the
84 Helfand moment approach of Viscardy {\it et
85 al.}\cite{Viscardy:2007bh,Viscardy:2007lq}) because these rely on
86 computing difficult to measure quantities.
87
88 Another attractive feature of RNEMD is that it conserves both total
89 linear momentum and total energy during the swaps (as long as the two
90 molecules have the same identity), so the swapped configurations are
91 typically samples from the same manifold of states in the
92 microcanonical ensemble.
93
94 Recently, Tenney and Maginn\cite{ISI:000273472300004} have discovered
95 some problems with the original RNEMD swap technique. Notably, large
96 momentum fluxes (equivalent to frequent momentum swaps between the
97 slabs) can result in ``notched'', ``peaked'' and generally non-thermal momentum
98 distributions in the two slabs, as well as non-linear thermal and
99 velocity distributions along the direction of the imposed flux ($z$).
100 Tenney and Maginn obtained reasonable limits on imposed flux and
101 self-adjusting metrics for retaining the usability of the method.
102
103 In this paper, we develop and test a method for non-isotropic velocity
104 scaling (NIVS-RNEMD) which retains the desirable features of RNEMD
105 (conservation of linear momentum and total energy, compatibility with
106 periodic boundary conditions) while establishing true thermal
107 distributions in each of the two slabs. In the next section, we
108 develop the method for determining the scaling constraints. We then
109 test the method on both single component, multi-component, and
110 non-isotropic mixtures and show that it is capable of providing
111 reasonable estimates of the thermal conductivity and shear viscosity
112 in these cases.
113
114 \section{Methodology}
115 We retain the basic idea of Muller-Plathe's RNEMD method; the periodic
116 system is partitioned into a series of thin slabs along a particular
117 axis ($z$). One of the slabs at the end of the periodic box is
118 designated the ``hot'' slab, while the slab in the center of the box
119 is designated the ``cold'' slab. The artificial momentum flux will be
120 established by transferring momentum from the cold slab and into the
121 hot slab.
122
123 Rather than using momentum swaps, we use a series of velocity scaling
124 moves. For molecules $\{i\}$ located within the cold slab,
125 \begin{equation}
126 \vec{v}_i \leftarrow \left( \begin{array}{ccc}
127 x & 0 & 0 \\
128 0 & y & 0 \\
129 0 & 0 & z \\
130 \end{array} \right) \cdot \vec{v}_i
131 \end{equation}
132 where ${x, y, z}$ are a set of 3 scaling variables for each of the
133 three directions in the system. Likewise, the molecules $\{j\}$
134 located in the hot slab will see a concomitant scaling of velocities,
135 \begin{equation}
136 \vec{v}_j \leftarrow \left( \begin{array}{ccc}
137 x^\prime & 0 & 0 \\
138 0 & y^\prime & 0 \\
139 0 & 0 & z^\prime \\
140 \end{array} \right) \cdot \vec{v}_j
141 \end{equation}
142
143 Conservation of linear momentum in each of the three directions
144 ($\alpha = x,y,z$) ties the values of the hot and cold bin scaling
145 parameters together:
146 \begin{equation}
147 P_h^\alpha + P_c^\alpha = \alpha^\prime P_h^\alpha + \alpha P_c^\alpha
148 \end{equation}
149 where
150 \begin{eqnarray}
151 P_c^\alpha & = & \sum_{i = 1}^{N_c} m_i \left[\vec{v}_i\right]_\alpha \\
152 P_h^\alpha & = & \sum_{j = 1}^{N_h} m_j \left[\vec{v}_j\right]_\alpha
153 \label{eq:momentumdef}
154 \end{eqnarray}
155 Therefore, for each of the three directions, the hot scaling
156 parameters are a simple function of the cold scaling parameters and
157 the instantaneous linear momentum in each of the two slabs.
158 \begin{equation}
159 \alpha^\prime = 1 + (1 - \alpha) p_\alpha
160 \label{eq:hotcoldscaling}
161 \end{equation}
162 where
163 \begin{equation}
164 p_\alpha = \frac{P_c^\alpha}{P_h^\alpha}
165 \end{equation}
166 for convenience.
167
168 Conservation of total energy also places constraints on the scaling:
169 \begin{equation}
170 \sum_{\alpha = x,y,z} K_h^\alpha + K_c^\alpha = \sum_{\alpha = x,y,z}
171 \left(\alpha^\prime\right)^2 K_h^\alpha + \alpha^2 K_c^\alpha
172 \end{equation}
173 where the kinetic energies, $K_h^\alpha$ and $K_c^\alpha$, are computed
174 for each of the three directions in a similar manner to the linear momenta
175 (Eq. \ref{eq:momentumdef}). Substituting in the expressions for the
176 hot scaling parameters ($\alpha^\prime$) from Eq. (\ref{eq:hotcoldscaling}),
177 we obtain the {\it constraint ellipsoid equation}:
178 \begin{equation}
179 \sum_{\alpha = x,y,z} a_\alpha \alpha^2 + b_\alpha \alpha + c_\alpha = 0
180 \label{eq:constraintEllipsoid}
181 \end{equation}
182 where the constants are obtained from the instantaneous values of the
183 linear momenta and kinetic energies for the hot and cold slabs,
184 \begin{eqnarray}
185 a_\alpha & = &\left(K_c^\alpha + K_h^\alpha
186 \left(p_\alpha\right)^2\right) \\
187 b_\alpha & = & -2 K_h^\alpha p_\alpha \left( 1 + p_\alpha\right) \\
188 c_\alpha & = & K_h^\alpha p_\alpha^2 + 2 K_h^\alpha p_\alpha - K_c^\alpha
189 \label{eq:constraintEllipsoidConsts}
190 \end{eqnarray}
191 This ellipsoid equation defines the set of cold slab scaling
192 parameters which can be applied while preserving both linear momentum
193 in all three directions as well as kinetic energy.
194
195 The goal of using velocity scaling variables is to transfer linear
196 momentum or kinetic energy from the cold slab to the hot slab. If the
197 hot and cold slabs are separated along the z-axis, the energy flux is
198 given simply by the decrease in kinetic energy of the cold bin:
199 \begin{equation}
200 (1-x^2) K_c^x + (1-y^2) K_c^y + (1-z^2) K_c^z = J_z\Delta t
201 \end{equation}
202 The expression for the energy flux can be re-written as another
203 ellipsoid centered on $(x,y,z) = 0$:
204 \begin{equation}
205 x^2 K_c^x + y^2 K_c^y + z^2 K_c^z = K_c^x + K_c^y + K_c^z - J_z\Delta t
206 \label{eq:fluxEllipsoid}
207 \end{equation}
208 The spatial extent of the {\it flux ellipsoid equation} is governed
209 both by a targetted value, $J_z$ as well as the instantaneous values of the
210 kinetic energy components in the cold bin.
211
212 To satisfy an energetic flux as well as the conservation constraints,
213 it is sufficient to determine the points ${x,y,z}$ which lie on both
214 the constraint ellipsoid (Eq. \ref{eq:constraintEllipsoid}) and the
215 flux ellipsoid (Eq. \ref{eq:fluxEllipsoid}), i.e. the intersection of
216 the two ellipsoids in 3-dimensional space.
217
218 \begin{figure}
219 \includegraphics[width=\linewidth]{ellipsoids}
220 \caption{Scaling points which maintain both constant energy and
221 constant linear momentum of the system lie on the surface of the
222 {\it constraint ellipsoid} while points which generate the target
223 momentum flux lie on the surface of the {\it flux ellipsoid}. The
224 velocity distributions in the hot bin are scaled by only those
225 points which lie on both ellipsoids.}
226 \label{ellipsoids}
227 \end{figure}
228
229 One may also define momentum flux (say along the x-direction) as:
230 \begin{equation}
231 (1-x) P_c^x = j_z(p_x)\Delta t
232 \label{eq:fluxPlane}
233 \end{equation}
234 The above {\it flux equation} is essentially a plane which is
235 perpendicular to the x-axis, with its position governed both by a
236 targetted value, $j_z(p_x)$ as well as the instantaneous value of the
237 momentum along the x-direction.
238
239 Similarly, to satisfy a momentum flux as well as the conservation
240 constraints, it is sufficient to determine the points ${x,y,z}$ which
241 lie on both the constraint ellipsoid (Eq. \ref{eq:constraintEllipsoid})
242 and the flux plane (Eq. \ref{eq:fluxPlane}), i.e. the intersection of
243 an ellipsoid and a plane in 3-dimensional space.
244
245 To summarize, by solving respective equation sets, one can determine
246 possible sets of scaling variables for cold slab. And corresponding
247 sets of scaling variables for hot slab can be determine as well.
248
249 The following problem will be choosing an optimal set of scaling
250 variables among the possible sets. Although this method is inherently
251 non-isotropic, the goal is still to maintain the system as isotropic
252 as possible. Under this consideration, one would like the kinetic
253 energies in different directions could become as close as each other
254 after each scaling. Simultaneously, one would also like each scaling
255 as gentle as possible, i.e. ${x,y,z \rightarrow 1}$, in order to avoid
256 large perturbation to the system. Therefore, one approach to obtain the
257 scaling variables would be constructing an criteria function, with
258 constraints as above equation sets, and solving the function's minimum
259 by method like Lagrange multipliers.
260
261 In order to save computation time, we have a different approach to a
262 relatively good set of scaling variables with much less calculation
263 than above. Here is the detail of our simplification of the problem.
264
265 In the case of kinetic energy transfer, we impose another constraint
266 ${x = y}$, into the equation sets. Consequently, there are two
267 variables left. And now one only needs to solve a set of two {\it
268 ellipses equations}. This problem would be transformed into solving
269 one quartic equation for one of the two variables. There are known
270 generic methods that solve real roots of quartic equations. Then one
271 can determine the other variable and obtain sets of scaling
272 variables. Among these sets, one can apply the above criteria to
273 choose the best set, while much faster with only a few sets to choose.
274
275 In the case of momentum flux transfer, we impose another constraint to
276 set the kinetic energy transfer as zero. In another word, we apply
277 Eq. \ref{eq:fluxEllipsoid} and let ${J_z = 0}$. After that, with one
278 variable fixed by Eq. \ref{eq:fluxPlane}, one now have a similar set
279 of equations on the above kinetic energy transfer problem. Therefore,
280 an approach similar to the above would be sufficient for this as well.
281
282 \section{Computational Details}
283 Our simulation consists of a series of systems. All of these
284 simulations were run with the OpenMD simulation software
285 package\cite{Meineke:2005gd} integrated with RNEMD methods.
286
287 A Lennard-Jones fluid system was built and tested first. In order to
288 compare our method with swapping RNEMD, a series of simulations were
289 performed to calculate the shear viscosity and thermal conductivity of
290 argon. 2592 atoms were in a orthorhombic cell, which was ${10.06\sigma
291 \times 10.06\sigma \times 30.18\sigma}$ by size. The reduced density
292 ${\rho^* = \rho\sigma^3}$ was thus 0.85, which enabled direct
293 comparison between our results and others. These simulations used
294 velocity Verlet algorithm with reduced timestep ${\tau^* =
295 4.6\times10^{-4}}$.
296
297 For shear viscosity calculation, the reduced temperature was ${T^* =
298 k_B T/\varepsilon = 0.72}$. Simulations were first equilibrated in canonical
299 ensemble (NVT), then equilibrated in microcanonical ensemble
300 (NVE). Establishing and stablizing momentum gradient were followed
301 also in NVE ensemble. For the swapping part, Muller-Plathe's algorithm was
302 adopted.\cite{ISI:000080382700030} The simulation box was under
303 periodic boundary condition, and devided into ${N = 20}$ slabs. In each swap,
304 the top slab ${(n = 1)}$ exchange the most negative $x$ momentum with the
305 most positive $x$ momentum in the center slab ${(n = N/2 + 1)}$. Referred
306 to Tenney {\it et al.}\cite{ISI:000273472300004}, a series of swapping
307 frequency were chosen. According to each result from swapping
308 RNEMD, scaling RNEMD simulations were run with the target momentum
309 flux set to produce a similar momentum flux and shear
310 rate. Furthermore, various scaling frequencies can be tested for one
311 single swapping rate. To compare the performance between swapping and
312 scaling method, temperatures of different dimensions in all the slabs
313 were observed. Most of the simulations include $10^5$ steps of
314 equilibration without imposing momentum flux, $10^5$ steps of
315 stablization with imposing momentum transfer, and $10^6$ steps of data
316 collection under RNEMD. For relatively high momentum flux simulations,
317 ${5\times10^5}$ step data collection is sufficient. For some low momentum
318 flux simulations, ${2\times10^6}$ steps were necessary.
319
320 After each simulation, the shear viscosity was calculated in reduced
321 unit. The momentum flux was calculated with total unphysical
322 transferred momentum ${P_x}$ and data collection time $t$:
323 \begin{equation}
324 j_z(p_x) = \frac{P_x}{2 t L_x L_y}
325 \end{equation}
326 And the velocity gradient ${\langle \partial v_x /\partial z \rangle}$
327 can be obtained by a linear regression of the velocity profile. From
328 the shear viscosity $\eta$ calculated with the above parameters, one
329 can further convert it into reduced unit ${\eta^* = \eta \sigma^2
330 (\varepsilon m)^{-1/2}}$.
331
332 For thermal conductivity calculation, simulations were first run under
333 reduced temperature ${T^* = 0.72}$ in NVE ensemble. Muller-Plathe's
334 algorithm was adopted in the swapping method. Under identical
335 simulation box parameters, in each swap, the top slab exchange the
336 molecule with least kinetic energy with the molecule in the center
337 slab with most kinetic energy, unless this ``coldest'' molecule in the
338 ``hot'' slab is still ``hotter'' than the ``hottest'' molecule in the ``cold''
339 slab. According to swapping RNEMD results, target energy flux for
340 scaling RNEMD simulations can be set. Also, various scaling
341 frequencies can be tested for one target energy flux. To compare the
342 performance between swapping and scaling method, distributions of
343 velocity and speed in different slabs were observed.
344
345 For each swapping rate, thermal conductivity was calculated in reduced
346 unit. The energy flux was calculated similarly to the momentum flux,
347 with total unphysical transferred energy ${E_{total}}$ and data collection
348 time $t$:
349 \begin{equation}
350 J_z = \frac{E_{total}}{2 t L_x L_y}
351 \end{equation}
352 And the temperature gradient ${\langle\partial T/\partial z\rangle}$
353 can be obtained by a linear regression of the temperature
354 profile. From the thermal conductivity $\lambda$ calculated, one can
355 further convert it into reduced unit ${\lambda^*=\lambda \sigma^2
356 m^{1/2} k_B^{-1}\varepsilon^{-1/2}}$.
357
358 Another series of our simulation is to calculate the interfacial
359 thermal conductivity of a Au/H${_2}$O system. Respective calculations of
360 water (SPC/E) and gold (QSC) thermal conductivity were performed and
361 compared with current results to ensure the validity of
362 NIVS-RNEMD. After that, the mixture system was simulated.
363
364 \section{Results And Discussion}
365 \subsection{Shear Viscosity}
366 Our calculations (Table \ref{shearRate}) shows that scale RNEMD method
367 produced comparable shear viscosity to swap RNEMD method. In Table
368 \ref{shearRate}, the names of the calculated samples are devided into
369 two parts. The first number refers to total slabs in one simulation
370 box. The second number refers to the swapping interval in swap method, or
371 in scale method the equilvalent swapping interval that the same
372 momentum flux would theoretically result in swap method. All the scale
373 method results were from simulations that had a scaling interval of 10
374 time steps. The average molecular momentum gradients of these samples
375 are shown in Figure \ref{shearGrad}.
376
377 \begin{table*}
378 \begin{minipage}{\linewidth}
379 \begin{center}
380
381 \caption{Calculation results for shear viscosity of Lennard-Jones
382 fluid at ${T^* = 0.72}$ and ${\rho^* = 0.85}$, with swap and scale
383 methods at various momentum exchange rates. Results in reduced
384 unit. Errors of calculations in parentheses. }
385
386 \begin{tabular}{ccc}
387 \hline
388 Series & $\eta^*_{swap}$ & $\eta^*_{scale}$\\
389 \hline
390 20-500 & 3.64(0.05) & 3.76(0.09)\\
391 20-1000 & 3.52(0.16) & 3.66(0.06)\\
392 20-2000 & 3.72(0.05) & 3.32(0.18)\\
393 20-2500 & 3.42(0.06) & 3.43(0.08)\\
394 \hline
395 \end{tabular}
396 \label{shearRate}
397 \end{center}
398 \end{minipage}
399 \end{table*}
400
401 \begin{figure}
402 \includegraphics[width=\linewidth]{shearGrad}
403 \caption{Average momentum gradients of shear viscosity simulations}
404 \label{shearGrad}
405 \end{figure}
406
407 \begin{figure}
408 \includegraphics[width=\linewidth]{shearTempScale}
409 \caption{Temperature profile for scaling RNEMD simulation.}
410 \label{shearTempScale}
411 \end{figure}
412 However, observations of temperatures along three dimensions show that
413 inhomogeneity occurs in scaling RNEMD simulations, particularly in the
414 two slabs which were scaled. Figure \ref{shearTempScale} indicate that with
415 relatively large imposed momentum flux, the temperature difference among $x$
416 and the other two dimensions was significant. This would result from the
417 algorithm of scaling method. From Eq. \ref{eq:fluxPlane}, after
418 momentum gradient is set up, $P_c^x$ would be roughly stable
419 ($<0$). Consequently, scaling factor $x$ would most probably larger
420 than 1. Therefore, the kinetic energy in $x$-dimension $K_c^x$ would
421 keep increase after most scaling steps. And if there is not enough time
422 for the kinetic energy to exchange among different dimensions and
423 different slabs, the system would finally build up temperature
424 (kinetic energy) difference among the three dimensions. Also, between
425 $y$ and $z$ dimensions in the scaled slabs, temperatures of $z$-axis
426 are closer to neighbor slabs. This is due to momentum transfer along
427 $z$ dimension between slabs.
428
429 Although results between scaling and swapping methods are comparable,
430 the inherent temperature inhomogeneity even in relatively low imposed
431 exchange momentum flux simulations makes scaling RNEMD method less
432 attractive than swapping RNEMD in shear viscosity calculation.
433
434 \subsection{Thermal Conductivity}
435
436 Our thermal conductivity calculations also show that scaling method results
437 agree with swapping method. Table \ref{thermal} lists our simulation
438 results with similar manner we used in shear viscosity
439 calculation. All the data reported from scaling method were obtained
440 by simulations of 10-step exchange frequency, and the target exchange
441 kinetic energy were set to produce equivalent kinetic energy flux as
442 in swapping method. Figure \ref{thermalGrad} exhibits similar thermal
443 gradients of respective similar kinetic energy flux.
444
445 \begin{table*}
446 \begin{minipage}{\linewidth}
447 \begin{center}
448
449 \caption{Calculation results for thermal conductivity of Lennard-Jones
450 fluid at ${\langle T^* \rangle = 0.72}$ and ${\rho^* = 0.85}$, with
451 swap and scale methods at various kinetic energy exchange rates. Results
452 in reduced unit. Errors of calculations in parentheses.}
453
454 \begin{tabular}{ccc}
455 \hline
456 Series & $\lambda^*_{swap}$ & $\lambda^*_{scale}$\\
457 \hline
458 20-250 & 7.03(0.34) & 7.30(0.10)\\
459 20-500 & 7.03(0.14) & 6.95(0.09)\\
460 20-1000 & 6.91(0.42) & 7.19(0.07)\\
461 20-2000 & 7.52(0.15) & 7.19(0.28)\\
462 \hline
463 \end{tabular}
464 \label{thermal}
465 \end{center}
466 \end{minipage}
467 \end{table*}
468
469 \begin{figure}
470 \includegraphics[width=\linewidth]{thermalGrad}
471 \caption{Temperature gradients of thermal conductivity simulations}
472 \label{thermalGrad}
473 \end{figure}
474
475 During these simulations, molecule velocities were recorded in 1000 of
476 all the snapshots. These velocity data were used to produce histograms
477 of velocity and speed distribution in different slabs. From these
478 histograms, it is observed that with increasing unphysical kinetic
479 energy flux, speed and velocity distribution of molecules in slabs
480 where swapping occured could deviate from Maxwell-Boltzmann
481 distribution. Figure \ref{histSwap} indicates how these distributions
482 deviate from ideal condition. In high temperature slabs, probability
483 density in low speed is confidently smaller than ideal distribution;
484 in low temperature slabs, probability density in high speed is smaller
485 than ideal. This phenomenon is observable even in our relatively low
486 swapping rate simulations. And this deviation could also leads to
487 deviation of distribution of velocity in various dimensions. One
488 feature of these deviated distribution is that in high temperature
489 slab, the ideal Gaussian peak was changed into a relatively flat
490 plateau; while in low temperature slab, that peak appears sharper.
491
492 \begin{figure}
493 \includegraphics[width=\linewidth]{histSwap}
494 \caption{Speed distribution for thermal conductivity using swapping RNEMD.}
495 \label{histSwap}
496 \end{figure}
497
498 \begin{figure}
499 \includegraphics[width=\linewidth]{histScale}
500 \caption{Speed distribution for thermal conductivity using scaling RNEMD.}
501 \label{histScale}
502 \end{figure}
503
504 \subsection{Interfaciel Thermal Conductivity}
505
506 \section{Acknowledgments}
507 Support for this project was provided by the National Science
508 Foundation under grant CHE-0848243. Computational time was provided by
509 the Center for Research Computing (CRC) at the University of Notre
510 Dame. \newpage
511
512 \bibliographystyle{jcp2}
513 \bibliography{nivsRnemd}
514 \end{doublespace}
515 \end{document}
516