ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/interfacial/interfacial.tex
Revision: 3736
Committed: Mon Jul 11 22:34:42 2011 UTC (12 years, 11 months ago) by skuang
Content type: application/x-tex
File size: 42932 byte(s)
Log Message:
add a figure. fix some citations.

File Contents

# User Rev Content
1 gezelter 3717 \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 skuang 3727 %\renewcommand\citemid{\ } % no comma in optional reference note
26 gezelter 3717 \bibpunct{[}{]}{,}{s}{}{;}
27     \bibliographystyle{aip}
28    
29     \begin{document}
30    
31     \title{Simulating interfacial thermal conductance at metal-solvent
32     interfaces: the role of chemical capping agents}
33    
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 3725
48 skuang 3732 With the Non-Isotropic Velocity Scaling algorithm (NIVS) we have
49     developed, an unphysical thermal flux can be effectively set up even
50     for non-homogeneous systems like interfaces in non-equilibrium
51     molecular dynamics simulations. In this work, this algorithm is
52     applied for simulating thermal conductance at metal / organic solvent
53     interfaces with various coverages of butanethiol capping
54     agents. Different solvents and force field models were tested. Our
55     results suggest that the United-Atom models are able to provide an
56     estimate of the interfacial thermal conductivity comparable to
57     experiments in our simulations with satisfactory computational
58     efficiency. From our results, the acoustic impedance mismatch between
59     metal and liquid phase is effectively reduced by the capping
60     agents, and thus leads to interfacial thermal conductance
61     enhancement. Furthermore, this effect is closely related to the
62     capping agent coverage on the metal surfaces and the type of solvent
63     molecules, and is affected by the models used in the simulations.
64 skuang 3725
65 gezelter 3717 \end{abstract}
66    
67     \newpage
68    
69     %\narrowtext
70    
71     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
72     % BODY OF TEXT
73     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
74    
75     \section{Introduction}
76 skuang 3725 Interfacial thermal conductance is extensively studied both
77 skuang 3733 experimentally and computationally, due to its importance in nanoscale
78     science and technology. Reliability of nanoscale devices depends on
79     their thermal transport properties. Unlike bulk homogeneous materials,
80     nanoscale materials features significant presence of interfaces, and
81     these interfaces could dominate the heat transfer behavior of these
82     materials. Furthermore, these materials are generally heterogeneous,
83     which challenges traditional research methods for homogeneous systems.
84 gezelter 3717
85 skuang 3733 Heat conductance of molecular and nano-scale interfaces will be
86     affected by the chemical details of the surface. Experimentally,
87     various interfaces have been investigated for their thermal
88     conductance properties. Wang {\it et al.} studied heat transport
89     through long-chain hydrocarbon monolayers on gold substrate at
90     individual molecular level\cite{Wang10082007}; Schmidt {\it et al.}
91     studied the role of CTAB on thermal transport between gold nanorods
92     and solvent\cite{doi:10.1021/jp8051888}; Juv\'e {\it et al.} studied
93     the cooling dynamics, which is controlled by thermal interface
94     resistence of glass-embedded metal
95     nanoparticles\cite{PhysRevB.80.195406}. Although interfaces are
96     commonly barriers for heat transport, Alper {\it et al.} suggested
97     that specific ligands (capping agents) could completely eliminate this
98     barrier ($G\rightarrow\infty$)\cite{doi:10.1021/la904855s}.
99    
100     Theoretical and computational studies were also engaged in the
101     interfacial thermal transport research in order to gain an
102 skuang 3734 understanding of this phenomena at the molecular level. Hase and
103     coworkers employed Non-Equilibrium Molecular Dynamics (NEMD)
104     simulations to study thermal transport from hot Au(111) substrate to a
105     self-assembled monolayer of alkylthiolate with relatively long chain
106 skuang 3736 (8-20 carbon atoms)\cite{hase:2010,hase:2011}. However,
107     emsemble average measurements for heat conductance of interfaces
108     between the capping monolayer on Au and a solvent phase has yet to be
109     studied. The relatively low thermal flux through interfaces is
110     difficult to measure with Equilibrium MD or forward NEMD simulation
111     methods. Therefore, the Reverse NEMD (RNEMD) methods would have the
112     advantage of having this difficult to measure flux known when studying
113     the thermal transport across interfaces, given that the simulation
114 skuang 3734 methods being able to effectively apply an unphysical flux in
115     non-homogeneous systems.
116    
117 skuang 3725 Recently, we have developed the Non-Isotropic Velocity Scaling (NIVS)
118     algorithm for RNEMD simulations\cite{kuang:164101}. This algorithm
119     retains the desirable features of RNEMD (conservation of linear
120     momentum and total energy, compatibility with periodic boundary
121     conditions) while establishing true thermal distributions in each of
122     the two slabs. Furthermore, it allows more effective thermal exchange
123     between particles of different identities, and thus enables extensive
124 skuang 3734 study of interfacial conductance under steady states.
125 skuang 3725
126 skuang 3734 Our work presented here investigated the Au(111) surface with various
127     coverage of butanethiol, a capping agent with shorter carbon chain,
128     solvated with organic solvents of different molecular shapes. And
129     different models were used for both the capping agent and the solvent
130     force field parameters. With the NIVS algorithm applied, the thermal
131     transport through these interfacial systems was studied and the
132     underlying mechanism for this phenomena was investigated.
133 skuang 3733
134 skuang 3734 [WHY STUDY AU-THIOL SURFACE; MAY CITE SHAOYI JIANG]
135    
136 skuang 3721 \section{Methodology}
137     \subsection{Algorithm}
138 skuang 3727 [BACKGROUND FOR MD METHODS]
139 skuang 3721 There have been many algorithms for computing thermal conductivity
140     using molecular dynamics simulations. However, interfacial conductance
141     is at least an order of magnitude smaller. This would make the
142     calculation even more difficult for those slowly-converging
143     equilibrium methods. Imposed-flux non-equilibrium
144     methods\cite{MullerPlathe:1997xw} have the flux set {\it a priori} and
145     the response of temperature or momentum gradients are easier to
146     measure than the flux, if unknown, and thus, is a preferable way to
147     the forward NEMD methods. Although the momentum swapping approach for
148     flux-imposing can be used for exchanging energy between particles of
149     different identity, the kinetic energy transfer efficiency is affected
150     by the mass difference between the particles, which limits its
151     application on heterogeneous interfacial systems.
152    
153     The non-isotropic velocity scaling (NIVS)\cite{kuang:164101} approach in
154     non-equilibrium MD simulations is able to impose relatively large
155     kinetic energy flux without obvious perturbation to the velocity
156     distribution of the simulated systems. Furthermore, this approach has
157     the advantage in heterogeneous interfaces in that kinetic energy flux
158     can be applied between regions of particles of arbitary identity, and
159     the flux quantity is not restricted by particle mass difference.
160    
161     The NIVS algorithm scales the velocity vectors in two separate regions
162     of a simulation system with respective diagonal scaling matricies. To
163     determine these scaling factors in the matricies, a set of equations
164     including linear momentum conservation and kinetic energy conservation
165 skuang 3736 constraints and target momentum / energy flux satisfaction is
166 skuang 3721 solved. With the scaling operation applied to the system in a set
167 skuang 3736 frequency, corresponding momentum / temperature gradients can be
168     built, which can be used for computing transport properties and other
169     applications related to momentum / temperature gradients. The NIVS
170 skuang 3721 algorithm conserves momenta and energy and does not depend on an
171 skuang 3736 external thermostat.
172 skuang 3721
173 skuang 3727 \subsection{Defining Interfacial Thermal Conductivity $G$}
174     For interfaces with a relatively low interfacial conductance, the bulk
175     regions on either side of an interface rapidly come to a state in
176     which the two phases have relatively homogeneous (but distinct)
177     temperatures. The interfacial thermal conductivity $G$ can therefore
178     be approximated as:
179     \begin{equation}
180     G = \frac{E_{total}}{2 t L_x L_y \left( \langle T_\mathrm{hot}\rangle -
181     \langle T_\mathrm{cold}\rangle \right)}
182     \label{lowG}
183     \end{equation}
184     where ${E_{total}}$ is the imposed non-physical kinetic energy
185     transfer and ${\langle T_\mathrm{hot}\rangle}$ and ${\langle
186     T_\mathrm{cold}\rangle}$ are the average observed temperature of the
187     two separated phases.
188 skuang 3721
189 skuang 3727 When the interfacial conductance is {\it not} small, two ways can be
190     used to define $G$.
191    
192     One way is to assume the temperature is discretely different on two
193     sides of the interface, $G$ can be calculated with the thermal flux
194     applied $J$ and the maximum temperature difference measured along the
195     thermal gradient max($\Delta T$), which occurs at the interface, as:
196     \begin{equation}
197     G=\frac{J}{\Delta T}
198     \label{discreteG}
199     \end{equation}
200    
201     The other approach is to assume a continuous temperature profile along
202     the thermal gradient axis (e.g. $z$) and define $G$ at the point where
203     the magnitude of thermal conductivity $\lambda$ change reach its
204     maximum, given that $\lambda$ is well-defined throughout the space:
205     \begin{equation}
206     G^\prime = \Big|\frac{\partial\lambda}{\partial z}\Big|
207     = \Big|\frac{\partial}{\partial z}\left(-J_z\Big/
208     \left(\frac{\partial T}{\partial z}\right)\right)\Big|
209     = |J_z|\Big|\frac{\partial^2 T}{\partial z^2}\Big|
210     \Big/\left(\frac{\partial T}{\partial z}\right)^2
211     \label{derivativeG}
212     \end{equation}
213    
214     With the temperature profile obtained from simulations, one is able to
215     approximate the first and second derivatives of $T$ with finite
216     difference method and thus calculate $G^\prime$.
217    
218     In what follows, both definitions are used for calculation and comparison.
219    
220     [IMPOSE G DEFINITION INTO OUR SYSTEMS]
221     To facilitate the use of the above definitions in calculating $G$ and
222     $G^\prime$, we have a metal slab with its (111) surfaces perpendicular
223     to the $z$-axis of our simulation cells. With or withour capping
224     agents on the surfaces, the metal slab is solvated with organic
225     solvents, as illustrated in Figure \ref{demoPic}.
226    
227     \begin{figure}
228     \includegraphics[width=\linewidth]{demoPic}
229     \caption{A sample showing how a metal slab has its (111) surface
230     covered by capping agent molecules and solvated by hexane.}
231     \label{demoPic}
232     \end{figure}
233    
234     With a simulation cell setup following the above manner, one is able
235     to equilibrate the system and impose an unphysical thermal flux
236     between the liquid and the metal phase with the NIVS algorithm. Under
237     a stablized thermal gradient induced by periodically applying the
238     unphysical flux, one is able to obtain a temperature profile and the
239     physical thermal flux corresponding to it, which equals to the
240     unphysical flux applied by NIVS. These data enables the evaluation of
241     the interfacial thermal conductance of a surface. Figure \ref{gradT}
242     is an example how those stablized thermal gradient can be used to
243     obtain the 1st and 2nd derivatives of the temperature profile.
244    
245     \begin{figure}
246     \includegraphics[width=\linewidth]{gradT}
247     \caption{The 1st and 2nd derivatives of temperature profile can be
248     obtained with finite difference approximation.}
249     \label{gradT}
250     \end{figure}
251    
252 skuang 3730 [MAY INCLUDE POWER SPECTRUM PROTOCOL]
253    
254 skuang 3727 \section{Computational Details}
255 skuang 3730 \subsection{Simulation Protocol}
256 skuang 3736 Our MD simulation code, OpenMD\cite{Meineke:2005gd,openmd}, has the
257     NIVS algorithm integrated and was used for our simulations. In our
258     simulations, Au is used to construct a metal slab with bare (111)
259     surface perpendicular to the $z$-axis. Different slab thickness (layer
260     numbers of Au) are simulated. This metal slab is first equilibrated
261     under normal pressure (1 atm) and a desired temperature. After
262     equilibration, butanethiol is used as the capping agent molecule to
263     cover the bare Au (111) surfaces evenly. The sulfur atoms in the
264     butanethiol molecules would occupy the three-fold sites of the
265     surfaces, and the maximal butanethiol capacity on Au surface is $1/3$
266     of the total number of surface Au atoms[CITATIONs]. A series of
267 skuang 3727 different coverage surfaces is investigated in order to study the
268     relation between coverage and conductance.
269    
270 skuang 3736 [COVERAGE DISCRIPTION]
271     In the initial configurations for each coverage precentage,
272     butanethiols were distributed evenly on the Au(111) surfaces. However,
273     since the interaction descriptions between surface Au and butanethiol
274     is non-bonded in our simulations, the capping agent molecules are
275 skuang 3727 allowed to migrate to an empty neighbor three-fold site during a
276     simulation. Therefore, the initial configuration would not severely
277     affect the sampling of a variety of configurations of the same
278     coverage, and the final conductance measurement would be an average
279     effect of these configurations explored in the simulations. [MAY NEED FIGURES]
280    
281     After the modified Au-butanethiol surface systems are equilibrated
282 skuang 3736 under canonical ensemble, organic solvent molecules are packed in the
283     previously vacuum part of the simulation cells and guarantees that
284     short range repulsive interactions do not disrupt the
285     simulations\cite{packmol}. Two solvents are investigated, one which
286     has little vibrational overlap with the alkanethiol and plane-like
287     shape (toluene), and one which has similar vibrational frequencies and
288     chain-like shape ({\it n}-hexane). [MAY EXPLAIN WHY WE CHOOSE THEM]
289 skuang 3727
290 skuang 3730 The spacing filled by solvent molecules, i.e. the gap between
291     periodically repeated Au-butanethiol surfaces should be carefully
292     chosen. A very long length scale for the thermal gradient axis ($z$)
293     may cause excessively hot or cold temperatures in the middle of the
294     solvent region and lead to undesired phenomena such as solvent boiling
295     or freezing when a thermal flux is applied. Conversely, too few
296     solvent molecules would change the normal behavior of the liquid
297     phase. Therefore, our $N_{solvent}$ values were chosen to ensure that
298     these extreme cases did not happen to our simulations. And the
299     corresponding spacing is usually $35 \sim 60$\AA.
300    
301 skuang 3728 The initial configurations generated by Packmol are further
302     equilibrated with the $x$ and $y$ dimensions fixed, only allowing
303     length scale change in $z$ dimension. This is to ensure that the
304     equilibration of liquid phase does not affect the metal crystal
305     structure in $x$ and $y$ dimensions. Further equilibration are run
306     under NVT and then NVE ensembles.
307    
308 skuang 3727 After the systems reach equilibrium, NIVS is implemented to impose a
309     periodic unphysical thermal flux between the metal and the liquid
310 skuang 3728 phase. Most of our simulations are under an average temperature of
311     $\sim$200K. Therefore, this flux usually comes from the metal to the
312 skuang 3727 liquid so that the liquid has a higher temperature and would not
313     freeze due to excessively low temperature. This induced temperature
314     gradient is stablized and the simulation cell is devided evenly into
315     N slabs along the $z$-axis and the temperatures of each slab are
316     recorded. When the slab width $d$ of each slab is the same, the
317     derivatives of $T$ with respect to slab number $n$ can be directly
318     used for $G^\prime$ calculations:
319     \begin{equation}
320     G^\prime = |J_z|\Big|\frac{\partial^2 T}{\partial z^2}\Big|
321     \Big/\left(\frac{\partial T}{\partial z}\right)^2
322     = |J_z|\Big|\frac{1}{d^2}\frac{\partial^2 T}{\partial n^2}\Big|
323     \Big/\left(\frac{1}{d}\frac{\partial T}{\partial n}\right)^2
324     = |J_z|\Big|\frac{\partial^2 T}{\partial n^2}\Big|
325     \Big/\left(\frac{\partial T}{\partial n}\right)^2
326     \label{derivativeG2}
327     \end{equation}
328    
329 skuang 3725 \subsection{Force Field Parameters}
330 skuang 3728 Our simulations include various components. Therefore, force field
331     parameter descriptions are needed for interactions both between the
332     same type of particles and between particles of different species.
333 skuang 3721
334     The Au-Au interactions in metal lattice slab is described by the
335 skuang 3736 quantum Sutton-Chen (QSC) formulation\cite{PhysRevB.59.3527}. The QSC
336 skuang 3721 potentials include zero-point quantum corrections and are
337     reparametrized for accurate surface energies compared to the
338     Sutton-Chen potentials\cite{Chen90}.
339    
340 skuang 3736 Figure \ref{demoMol} demonstrates how we name our pseudo-atoms of the
341     organic solvent molecules in our simulations.
342 skuang 3730
343 skuang 3736 \begin{figure}
344     \includegraphics[width=\linewidth]{demoMol}
345     \caption{Denomination of atoms or pseudo-atoms in our simulations: a)
346     UA-hexane; b) AA-hexane; c) UA-toluene; d) AA-toluene.}
347     \label{demoMol}
348     \end{figure}
349    
350 skuang 3728 For both solvent molecules, straight chain {\it n}-hexane and aromatic
351     toluene, United-Atom (UA) and All-Atom (AA) models are used
352     respectively. The TraPPE-UA
353     parameters\cite{TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes} are used
354     for our UA solvent molecules. In these models, pseudo-atoms are
355     located at the carbon centers for alkyl groups. By eliminating
356     explicit hydrogen atoms, these models are simple and computationally
357 skuang 3729 efficient, while maintains good accuracy. However, the TraPPE-UA for
358     alkanes is known to predict a lower boiling point than experimental
359     values. Considering that after an unphysical thermal flux is applied
360     to a system, the temperature of ``hot'' area in the liquid phase would be
361     significantly higher than the average, to prevent over heating and
362     boiling of the liquid phase, the average temperature in our
363 skuang 3730 simulations should be much lower than the liquid boiling point. [MORE DISCUSSION]
364 skuang 3729 For UA-toluene model, rigid body constraints are applied, so that the
365 skuang 3730 benzene ring and the methyl-CRar bond are kept rigid. This would save
366     computational time.[MORE DETAILS]
367 skuang 3721
368 skuang 3729 Besides the TraPPE-UA models, AA models for both organic solvents are
369 skuang 3730 included in our studies as well. For hexane, the OPLS-AA\cite{OPLSAA}
370     force field is used. [MORE DETAILS]
371 skuang 3729 For toluene, the United Force Field developed by Rapp\'{e} {\it et
372     al.}\cite{doi:10.1021/ja00051a040} is adopted.[MORE DETAILS]
373 skuang 3728
374 skuang 3729 The capping agent in our simulations, the butanethiol molecules can
375     either use UA or AA model. The TraPPE-UA force fields includes
376 skuang 3730 parameters for thiol molecules\cite{TraPPE-UA.thiols} and are used for
377     UA butanethiol model in our simulations. The OPLS-AA also provides
378     parameters for alkyl thiols. However, alkyl thiols adsorbed on Au(111)
379     surfaces do not have the hydrogen atom bonded to sulfur. To adapt this
380     change and derive suitable parameters for butanethiol adsorbed on
381 skuang 3736 Au(111) surfaces, we adopt the S parameters from Luedtke and
382     Landman\cite{landman:1998} and modify parameters for its neighbor C
383     atom for charge balance in the molecule. Note that the model choice
384     (UA or AA) of capping agent can be different from the
385     solvent. Regardless of model choice, the force field parameters for
386     interactions between capping agent and solvent can be derived using
387     Lorentz-Berthelot Mixing Rule:[EQN'S]
388 skuang 3721
389 skuang 3730
390 skuang 3721 To describe the interactions between metal Au and non-metal capping
391 skuang 3730 agent and solvent particles, we refer to an adsorption study of alkyl
392     thiols on gold surfaces by Vlugt {\it et
393     al.}\cite{vlugt:cpc2007154} They fitted an effective Lennard-Jones
394     form of potential parameters for the interaction between Au and
395     pseudo-atoms CH$_x$ and S based on a well-established and widely-used
396 skuang 3736 effective potential of Hautman and Klein\cite{hautman:4994} for the
397     Au(111) surface. As our simulations require the gold lattice slab to
398     be non-rigid so that it could accommodate kinetic energy for thermal
399 skuang 3730 transport study purpose, the pair-wise form of potentials is
400     preferred.
401 skuang 3721
402 skuang 3730 Besides, the potentials developed from {\it ab initio} calculations by
403     Leng {\it et al.}\cite{doi:10.1021/jp034405s} are adopted for the
404     interactions between Au and aromatic C/H atoms in toluene.[MORE DETAILS]
405 skuang 3725
406 skuang 3730 However, the Lennard-Jones parameters between Au and other types of
407     particles in our simulations are not yet well-established. For these
408     interactions, we attempt to derive their parameters using the Mixing
409     Rule. To do this, the ``Metal-non-Metal'' (MnM) interaction parameters
410     for Au is first extracted from the Au-CH$_x$ parameters by applying
411     the Mixing Rule reversely. Table \ref{MnM} summarizes these ``MnM''
412     parameters in our simulations.
413 skuang 3729
414 skuang 3730 \begin{table*}
415     \begin{minipage}{\linewidth}
416     \begin{center}
417     \caption{Lennard-Jones parameters for Au-non-Metal
418     interactions in our simulations.}
419    
420     \begin{tabular}{ccc}
421     \hline\hline
422 skuang 3732 Non-metal atom & $\sigma$ & $\epsilon$ \\
423     (or pseudo-atom) & \AA & kcal/mol \\
424 skuang 3730 \hline
425     S & 2.40 & 8.465 \\
426     CH3 & 3.54 & 0.2146 \\
427     CH2 & 3.54 & 0.1749 \\
428     CT3 & 3.365 & 0.1373 \\
429     CT2 & 3.365 & 0.1373 \\
430     CTT & 3.365 & 0.1373 \\
431     HC & 2.865 & 0.09256 \\
432     CHar & 3.4625 & 0.1680 \\
433     CRar & 3.555 & 0.1604 \\
434     CA & 3.173 & 0.0640 \\
435     HA & 2.746 & 0.0414 \\
436     \hline\hline
437     \end{tabular}
438     \label{MnM}
439     \end{center}
440     \end{minipage}
441     \end{table*}
442 skuang 3729
443    
444 skuang 3730 \section{Results and Discussions}
445     [MAY HAVE A BRIEF SUMMARY]
446     \subsection{How Simulation Parameters Affects $G$}
447     [MAY NOT PUT AT FIRST]
448     We have varied our protocol or other parameters of the simulations in
449     order to investigate how these factors would affect the measurement of
450     $G$'s. It turned out that while some of these parameters would not
451     affect the results substantially, some other changes to the
452     simulations would have a significant impact on the measurement
453     results.
454 skuang 3725
455 skuang 3730 In some of our simulations, we allowed $L_x$ and $L_y$ to change
456     during equilibrating the liquid phase. Due to the stiffness of the Au
457     slab, $L_x$ and $L_y$ would not change noticeably after
458     equilibration. Although $L_z$ could fluctuate $\sim$1\% after a system
459     is fully equilibrated in the NPT ensemble, this fluctuation, as well
460     as those comparably smaller to $L_x$ and $L_y$, would not be magnified
461     on the calculated $G$'s, as shown in Table \ref{AuThiolHexaneUA}. This
462     insensivity to $L_i$ fluctuations allows reliable measurement of $G$'s
463     without the necessity of extremely cautious equilibration process.
464 skuang 3725
465 skuang 3730 As stated in our computational details, the spacing filled with
466     solvent molecules can be chosen within a range. This allows some
467     change of solvent molecule numbers for the same Au-butanethiol
468     surfaces. We did this study on our Au-butanethiol/hexane
469     simulations. Nevertheless, the results obtained from systems of
470     different $N_{hexane}$ did not indicate that the measurement of $G$ is
471     susceptible to this parameter. For computational efficiency concern,
472     smaller system size would be preferable, given that the liquid phase
473     structure is not affected.
474    
475     Our NIVS algorithm allows change of unphysical thermal flux both in
476     direction and in quantity. This feature extends our investigation of
477     interfacial thermal conductance. However, the magnitude of this
478     thermal flux is not arbitary if one aims to obtain a stable and
479     reliable thermal gradient. A temperature profile would be
480     substantially affected by noise when $|J_z|$ has a much too low
481     magnitude; while an excessively large $|J_z|$ that overwhelms the
482     conductance capacity of the interface would prevent a thermal gradient
483     to reach a stablized steady state. NIVS has the advantage of allowing
484     $J$ to vary in a wide range such that the optimal flux range for $G$
485     measurement can generally be simulated by the algorithm. Within the
486     optimal range, we were able to study how $G$ would change according to
487     the thermal flux across the interface. For our simulations, we denote
488     $J_z$ to be positive when the physical thermal flux is from the liquid
489     to metal, and negative vice versa. The $G$'s measured under different
490     $J_z$ is listed in Table \ref{AuThiolHexaneUA} and [REF]. These
491     results do not suggest that $G$ is dependent on $J_z$ within this flux
492     range. The linear response of flux to thermal gradient simplifies our
493     investigations in that we can rely on $G$ measurement with only a
494     couple $J_z$'s and do not need to test a large series of fluxes.
495    
496     %ADD MORE TO TABLE
497 skuang 3725 \begin{table*}
498     \begin{minipage}{\linewidth}
499     \begin{center}
500     \caption{Computed interfacial thermal conductivity ($G$ and
501 skuang 3731 $G^\prime$) values for the 100\% covered Au-butanethiol/hexane
502     interfaces with UA model and different hexane molecule numbers
503     at different temperatures using a range of energy fluxes.}
504 skuang 3730
505 skuang 3731 \begin{tabular}{cccccccc}
506 skuang 3730 \hline\hline
507 skuang 3731 $\langle T\rangle$ & & $L_x$ & $L_y$ & $L_z$ & $J_z$ &
508     $G$ & $G^\prime$ \\
509 skuang 3732 (K) & $N_{hexane}$ & \multicolumn{3}{c}{(\AA)} & (GW/m$^2$) &
510 skuang 3730 \multicolumn{2}{c}{(MW/m$^2$/K)} \\
511     \hline
512 skuang 3731 200 & 266 & 29.86 & 25.80 & 113.1 & -0.96 &
513     102() & 80.0() \\
514     & 200 & 29.84 & 25.81 & 93.9 & 1.92 &
515     129() & 87.3() \\
516     & & 29.84 & 25.81 & 95.3 & 1.93 &
517     131() & 77.5() \\
518     & 166 & 29.84 & 25.81 & 85.7 & 0.97 &
519     115() & 69.3() \\
520     & & & & & 1.94 &
521     125() & 87.1() \\
522     250 & 200 & 29.84 & 25.87 & 106.8 & 0.96 &
523     81.8() & 67.0() \\
524     & 166 & 29.87 & 25.84 & 94.8 & 0.98 &
525     79.0() & 62.9() \\
526     & & 29.84 & 25.85 & 95.0 & 1.44 &
527     76.2() & 64.8() \\
528 skuang 3730 \hline\hline
529     \end{tabular}
530     \label{AuThiolHexaneUA}
531     \end{center}
532     \end{minipage}
533     \end{table*}
534    
535     Furthermore, we also attempted to increase system average temperatures
536     to above 200K. These simulations are first equilibrated in the NPT
537     ensemble under normal pressure. As stated above, the TraPPE-UA model
538     for hexane tends to predict a lower boiling point. In our simulations,
539     hexane had diffculty to remain in liquid phase when NPT equilibration
540     temperature is higher than 250K. Additionally, the equilibrated liquid
541     hexane density under 250K becomes lower than experimental value. This
542     expanded liquid phase leads to lower contact between hexane and
543     butanethiol as well.[MAY NEED FIGURE] And this reduced contact would
544     probably be accountable for a lower interfacial thermal conductance,
545     as shown in Table \ref{AuThiolHexaneUA}.
546    
547     A similar study for TraPPE-UA toluene agrees with the above result as
548     well. Having a higher boiling point, toluene tends to remain liquid in
549     our simulations even equilibrated under 300K in NPT
550     ensembles. Furthermore, the expansion of the toluene liquid phase is
551     not as significant as that of the hexane. This prevents severe
552     decrease of liquid-capping agent contact and the results (Table
553     \ref{AuThiolToluene}) show only a slightly decreased interface
554     conductance. Therefore, solvent-capping agent contact should play an
555     important role in the thermal transport process across the interface
556     in that higher degree of contact could yield increased conductance.
557    
558 skuang 3731 [ADD Lxyz AND ERROR ESTIMATE TO TABLE]
559 skuang 3730 \begin{table*}
560     \begin{minipage}{\linewidth}
561     \begin{center}
562     \caption{Computed interfacial thermal conductivity ($G$ and
563 skuang 3731 $G^\prime$) values for a 90\% coverage Au-butanethiol/toluene
564     interface at different temperatures using a range of energy
565     fluxes.}
566 skuang 3725
567     \begin{tabular}{cccc}
568     \hline\hline
569     $\langle T\rangle$ & $J_z$ & $G$ & $G^\prime$ \\
570     (K) & (GW/m$^2$) & \multicolumn{2}{c}{(MW/m$^2$/K)} \\
571     \hline
572 skuang 3731 200 & -1.86 & 180() & 135() \\
573     & 2.15 & 204() & 113() \\
574     & -3.93 & 175() & 114() \\
575     300 & -1.91 & 143() & 125() \\
576     & -4.19 & 134() & 113() \\
577 skuang 3725 \hline\hline
578     \end{tabular}
579     \label{AuThiolToluene}
580     \end{center}
581     \end{minipage}
582     \end{table*}
583    
584 skuang 3730 Besides lower interfacial thermal conductance, surfaces in relatively
585     high temperatures are susceptible to reconstructions, when
586     butanethiols have a full coverage on the Au(111) surface. These
587     reconstructions include surface Au atoms migrated outward to the S
588     atom layer, and butanethiol molecules embedded into the original
589     surface Au layer. The driving force for this behavior is the strong
590     Au-S interactions in our simulations. And these reconstructions lead
591     to higher ratio of Au-S attraction and thus is energetically
592     favorable. Furthermore, this phenomenon agrees with experimental
593     results\cite{doi:10.1021/j100035a033,doi:10.1021/la026493y}. Vlugt
594     {\it et al.} had kept their Au(111) slab rigid so that their
595     simulations can reach 300K without surface reconstructions. Without
596     this practice, simulating 100\% thiol covered interfaces under higher
597     temperatures could hardly avoid surface reconstructions. However, our
598     measurement is based on assuming homogeneity on $x$ and $y$ dimensions
599     so that measurement of $T$ at particular $z$ would be an effective
600     average of the particles of the same type. Since surface
601     reconstructions could eliminate the original $x$ and $y$ dimensional
602     homogeneity, measurement of $G$ is more difficult to conduct under
603     higher temperatures. Therefore, most of our measurements are
604 skuang 3732 undertaken at $\langle T\rangle\sim$200K.
605 skuang 3725
606 skuang 3730 However, when the surface is not completely covered by butanethiols,
607     the simulated system is more resistent to the reconstruction
608     above. Our Au-butanethiol/toluene system did not see this phenomena
609 skuang 3734 even at $\langle T\rangle\sim$300K. The Au(111) surfaces have a 90\% coverage of
610 skuang 3730 butanethiols and have empty three-fold sites. These empty sites could
611     help prevent surface reconstruction in that they provide other means
612     of capping agent relaxation. It is observed that butanethiols can
613     migrate to their neighbor empty sites during a simulation. Therefore,
614     we were able to obtain $G$'s for these interfaces even at a relatively
615     high temperature without being affected by surface reconstructions.
616 skuang 3725
617 skuang 3730 \subsection{Influence of Capping Agent Coverage on $G$}
618     To investigate the influence of butanethiol coverage on interfacial
619     thermal conductance, a series of different coverage Au-butanethiol
620     surfaces is prepared and solvated with various organic
621     molecules. These systems are then equilibrated and their interfacial
622     thermal conductivity are measured with our NIVS algorithm. Table
623     \ref{tlnUhxnUhxnD} lists these results for direct comparison between
624 skuang 3731 different coverages of butanethiol. To study the isotope effect in
625     interfacial thermal conductance, deuterated UA-hexane is included as
626     well.
627 skuang 3730
628 skuang 3731 It turned out that with partial covered butanethiol on the Au(111)
629     surface, the derivative definition for $G$ (Eq. \ref{derivativeG}) has
630     difficulty to apply, due to the difficulty in locating the maximum of
631     change of $\lambda$. Instead, the discrete definition
632     (Eq. \ref{discreteG}) is easier to apply, as max($\Delta T$) can still
633     be well-defined. Therefore, $G$'s (not $G^\prime$) are used for this
634     section.
635 skuang 3725
636 skuang 3731 From Table \ref{tlnUhxnUhxnD}, one can see the significance of the
637     presence of capping agents. Even when a fraction of the Au(111)
638     surface sites are covered with butanethiols, the conductivity would
639     see an enhancement by at least a factor of 3. This indicates the
640     important role cappping agent is playing for thermal transport
641     phenomena on metal/organic solvent surfaces.
642 skuang 3725
643 skuang 3731 Interestingly, as one could observe from our results, the maximum
644     conductance enhancement (largest $G$) happens while the surfaces are
645     about 75\% covered with butanethiols. This again indicates that
646     solvent-capping agent contact has an important role of the thermal
647     transport process. Slightly lower butanethiol coverage allows small
648     gaps between butanethiols to form. And these gaps could be filled with
649     solvent molecules, which acts like ``heat conductors'' on the
650     surface. The higher degree of interaction between these solvent
651     molecules and capping agents increases the enhancement effect and thus
652     produces a higher $G$ than densely packed butanethiol arrays. However,
653     once this maximum conductance enhancement is reached, $G$ decreases
654     when butanethiol coverage continues to decrease. Each capping agent
655     molecule reaches its maximum capacity for thermal
656     conductance. Therefore, even higher solvent-capping agent contact
657     would not offset this effect. Eventually, when butanethiol coverage
658     continues to decrease, solvent-capping agent contact actually
659     decreases with the disappearing of butanethiol molecules. In this
660     case, $G$ decrease could not be offset but instead accelerated.
661 skuang 3725
662 skuang 3731 A comparison of the results obtained from differenet organic solvents
663     can also provide useful information of the interfacial thermal
664     transport process. The deuterated hexane (UA) results do not appear to
665     be much different from those of normal hexane (UA), given that
666     butanethiol (UA) is non-deuterated for both solvents. These UA model
667     studies, even though eliminating C-H vibration samplings, still have
668     C-C vibrational frequencies different from each other. However, these
669 skuang 3732 differences in the infrared range do not seem to produce an observable
670 skuang 3731 difference for the results of $G$. [MAY NEED FIGURE]
671 skuang 3730
672 skuang 3731 Furthermore, results for rigid body toluene solvent, as well as other
673     UA-hexane solvents, are reasonable within the general experimental
674     ranges[CITATIONS]. This suggests that explicit hydrogen might not be a
675     required factor for modeling thermal transport phenomena of systems
676     such as Au-thiol/organic solvent.
677    
678     However, results for Au-butanethiol/toluene do not show an identical
679     trend with those for Au-butanethiol/hexane in that $G$'s remain at
680     approximately the same magnitue when butanethiol coverage differs from
681     25\% to 75\%. This might be rooted in the molecule shape difference
682     for plane-like toluene and chain-like {\it n}-hexane. Due to this
683     difference, toluene molecules have more difficulty in occupying
684     relatively small gaps among capping agents when their coverage is not
685     too low. Therefore, the solvent-capping agent contact may keep
686     increasing until the capping agent coverage reaches a relatively low
687     level. This becomes an offset for decreasing butanethiol molecules on
688     its effect to the process of interfacial thermal transport. Thus, one
689     can see a plateau of $G$ vs. butanethiol coverage in our results.
690    
691     [NEED ERROR ESTIMATE, MAY ALSO PUT J HERE]
692 skuang 3725 \begin{table*}
693     \begin{minipage}{\linewidth}
694     \begin{center}
695 skuang 3732 \caption{Computed interfacial thermal conductivity ($G$) values
696     for the Au-butanethiol/solvent interface with various UA
697     models and different capping agent coverages at $\langle
698     T\rangle\sim$200K using certain energy flux respectively.}
699 skuang 3725
700 skuang 3731 \begin{tabular}{cccc}
701 skuang 3725 \hline\hline
702 skuang 3732 Thiol & \multicolumn{3}{c}{$G$(MW/m$^2$/K)} \\
703     coverage (\%) & hexane & hexane(D) & toluene \\
704 skuang 3725 \hline
705 skuang 3732 0.0 & 46.5() & 43.9() & 70.1() \\
706     25.0 & 151() & 153() & 249() \\
707     50.0 & 172() & 182() & 214() \\
708     75.0 & 242() & 229() & 244() \\
709     88.9 & 178() & - & - \\
710     100.0 & 137() & 153() & 187() \\
711 skuang 3725 \hline\hline
712     \end{tabular}
713 skuang 3730 \label{tlnUhxnUhxnD}
714 skuang 3725 \end{center}
715     \end{minipage}
716     \end{table*}
717    
718 skuang 3730 \subsection{Influence of Chosen Molecule Model on $G$}
719     [MAY COMBINE W MECHANISM STUDY]
720    
721 skuang 3732 In addition to UA solvent/capping agent models, AA models are included
722     in our simulations as well. Besides simulations of the same (UA or AA)
723     model for solvent and capping agent, different models can be applied
724     to different components. Furthermore, regardless of models chosen,
725     either the solvent or the capping agent can be deuterated, similar to
726     the previous section. Table \ref{modelTest} summarizes the results of
727     these studies.
728 skuang 3725
729 skuang 3732 [MORE DATA; ERROR ESTIMATE]
730 skuang 3725 \begin{table*}
731     \begin{minipage}{\linewidth}
732     \begin{center}
733    
734     \caption{Computed interfacial thermal conductivity ($G$ and
735 skuang 3732 $G^\prime$) values for interfaces using various models for
736     solvent and capping agent (or without capping agent) at
737     $\langle T\rangle\sim$200K.}
738 skuang 3725
739 skuang 3732 \begin{tabular}{ccccc}
740 skuang 3725 \hline\hline
741 skuang 3732 Butanethiol model & Solvent & $J_z$ & $G$ & $G^\prime$ \\
742     (or bare surface) & model & (GW/m$^2$) &
743     \multicolumn{2}{c}{(MW/m$^2$/K)} \\
744 skuang 3725 \hline
745 skuang 3732 UA & AA hexane & 1.94 & 135() & 129() \\
746     & & 2.86 & 126() & 115() \\
747     & AA toluene & 1.89 & 200() & 149() \\
748     AA & UA hexane & 1.94 & 116() & 129() \\
749     & AA hexane & 3.76 & 451() & 378() \\
750     & & 4.71 & 432() & 334() \\
751     & AA toluene & 3.79 & 487() & 290() \\
752     AA(D) & UA hexane & 1.94 & 158() & 172() \\
753     bare & AA hexane & 0.96 & 31.0() & 29.4() \\
754 skuang 3725 \hline\hline
755     \end{tabular}
756 skuang 3732 \label{modelTest}
757 skuang 3725 \end{center}
758     \end{minipage}
759     \end{table*}
760    
761 skuang 3732 To facilitate direct comparison, the same system with differnt models
762     for different components uses the same length scale for their
763     simulation cells. Without the presence of capping agent, using
764     different models for hexane yields similar results for both $G$ and
765     $G^\prime$, and these two definitions agree with eath other very
766     well. This indicates very weak interaction between the metal and the
767     solvent, and is a typical case for acoustic impedance mismatch between
768     these two phases.
769 skuang 3730
770 skuang 3732 As for Au(111) surfaces completely covered by butanethiols, the choice
771     of models for capping agent and solvent could impact the measurement
772     of $G$ and $G^\prime$ quite significantly. For Au-butanethiol/hexane
773     interfaces, using AA model for both butanethiol and hexane yields
774     substantially higher conductivity values than using UA model for at
775     least one component of the solvent and capping agent, which exceeds
776     the upper bond of experimental value range. This is probably due to
777     the classically treated C-H vibrations in the AA model, which should
778     not be appreciably populated at normal temperatures. In comparison,
779     once either the hexanes or the butanethiols are deuterated, one can
780     see a significantly lower $G$ and $G^\prime$. In either of these
781     cases, the C-H(D) vibrational overlap between the solvent and the
782     capping agent is removed. [MAY NEED FIGURE] Conclusively, the
783     improperly treated C-H vibration in the AA model produced
784     over-predicted results accordingly. Compared to the AA model, the UA
785     model yields more reasonable results with higher computational
786     efficiency.
787 skuang 3731
788 skuang 3732 However, for Au-butanethiol/toluene interfaces, having the AA
789     butanethiol deuterated did not yield a significant change in the
790     measurement results.
791     . , so extra degrees of freedom
792     such as the C-H vibrations could enhance heat exchange between these
793     two phases and result in a much higher conductivity.
794 skuang 3731
795 skuang 3732
796     Although the QSC model for Au is known to predict an overly low value
797     for bulk metal gold conductivity[CITE NIVSRNEMD], our computational
798     results for $G$ and $G^\prime$ do not seem to be affected by this
799     drawback of the model for metal. Instead, the modeling of interfacial
800     thermal transport behavior relies mainly on an accurate description of
801     the interactions between components occupying the interfaces.
802    
803 skuang 3730 \subsection{Mechanism of Interfacial Thermal Conductance Enhancement
804     by Capping Agent}
805 skuang 3732 %OR\subsection{Vibrational spectrum study on conductance mechanism}
806 skuang 3730
807 skuang 3732 [MAY INTRODUCE PROTOCOL IN METHODOLOGY/COMPUTATIONAL DETAIL, EQN'S]
808 skuang 3730
809 skuang 3725 To investigate the mechanism of this interfacial thermal conductance,
810     the vibrational spectra of various gold systems were obtained and are
811     shown as in the upper panel of Fig. \ref{vibration}. To obtain these
812     spectra, one first runs a simulation in the NVE ensemble and collects
813     snapshots of configurations; these configurations are used to compute
814     the velocity auto-correlation functions, which is used to construct a
815 skuang 3732 power spectrum via a Fourier transform.
816 skuang 3725
817 skuang 3732 The gold surfaces covered by
818     butanethiol molecules, compared to bare gold surfaces, exhibit an
819     additional peak observed at a frequency of $\sim$170cm$^{-1}$, which
820     is attributed to the vibration of the S-Au bond. This vibration
821     enables efficient thermal transport from surface Au atoms to the
822     capping agents. Simultaneously, as shown in the lower panel of
823     Fig. \ref{vibration}, the large overlap of the vibration spectra of
824     butanethiol and hexane in the all-atom model, including the C-H
825     vibration, also suggests high thermal exchange efficiency. The
826     combination of these two effects produces the drastic interfacial
827     thermal conductance enhancement in the all-atom model.
828    
829     [MAY NEED TO CONVERT TO JPEG]
830 skuang 3725 \begin{figure}
831     \includegraphics[width=\linewidth]{vibration}
832     \caption{Vibrational spectra obtained for gold in different
833     environments (upper panel) and for Au/thiol/hexane simulation in
834     all-atom model (lower panel).}
835     \label{vibration}
836     \end{figure}
837    
838 skuang 3732 [COMPARISON OF TWO G'S; AU SLAB WIDTHS; ETC]
839     % The results show that the two definitions used for $G$ yield
840     % comparable values, though $G^\prime$ tends to be smaller.
841    
842 skuang 3730 \section{Conclusions}
843 skuang 3732 The NIVS algorithm we developed has been applied to simulations of
844     Au-butanethiol surfaces with organic solvents. This algorithm allows
845     effective unphysical thermal flux transferred between the metal and
846     the liquid phase. With the flux applied, we were able to measure the
847     corresponding thermal gradient and to obtain interfacial thermal
848     conductivities. Our simulations have seen significant conductance
849     enhancement with the presence of capping agent, compared to the bare
850     gold/liquid interfaces. The acoustic impedance mismatch between the
851     metal and the liquid phase is effectively eliminated by proper capping
852     agent. Furthermore, the coverage precentage of the capping agent plays
853     an important role in the interfacial thermal transport process.
854 skuang 3725
855 skuang 3732 Our measurement results, particularly of the UA models, agree with
856     available experimental data. This indicates that our force field
857     parameters have a nice description of the interactions between the
858     particles at the interfaces. AA models tend to overestimate the
859     interfacial thermal conductance in that the classically treated C-H
860     vibration would be overly sampled. Compared to the AA models, the UA
861     models have higher computational efficiency with satisfactory
862     accuracy, and thus are preferable in interfacial thermal transport
863     modelings.
864 skuang 3730
865 skuang 3732 Vlugt {\it et al.} has investigated the surface thiol structures for
866     nanocrystal gold and pointed out that they differs from those of the
867     Au(111) surface\cite{vlugt:cpc2007154}. This difference might lead to
868     change of interfacial thermal transport behavior as well. To
869     investigate this problem, an effective means to introduce thermal flux
870     and measure the corresponding thermal gradient is desirable for
871     simulating structures with spherical symmetry.
872 skuang 3730
873 skuang 3732
874 gezelter 3717 \section{Acknowledgments}
875     Support for this project was provided by the National Science
876     Foundation under grant CHE-0848243. Computational time was provided by
877     the Center for Research Computing (CRC) at the University of Notre
878 skuang 3730 Dame. \newpage
879 gezelter 3717
880     \bibliography{interfacial}
881    
882     \end{doublespace}
883     \end{document}
884