ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/interfacial/interfacial.tex
Revision: 3763
Committed: Tue Sep 27 21:02:48 2011 UTC (12 years, 9 months ago) by skuang
Content type: application/x-tex
File size: 52212 byte(s)
Log Message:
revision according to reviewer comments.

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 3740 \bibpunct{[}{]}{,}{n}{}{;}
27     \bibliographystyle{achemso}
28 gezelter 3717
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 gezelter 3761 With the Non-Isotropic Velocity Scaling (NIVS) approach to Reverse
48     Non-Equilibrium Molecular Dynamics (RNEMD) it is possible to impose
49     an unphysical thermal flux between different regions of
50     inhomogeneous systems such as solid / liquid interfaces. We have
51     applied NIVS to compute the interfacial thermal conductance at a
52     metal / organic solvent interface that has been chemically capped by
53     butanethiol molecules. Our calculations suggest that the acoustic
54     impedance mismatch between the metal and liquid phases is
55     effectively reduced by the capping agents, leading to a greatly
56     enhanced conductivity at the interface. Specifically, the chemical
57     bond between the metal and the capping agent introduces a
58     vibrational overlap that is not present without the capping agent,
59     and the overlap between the vibrational spectra (metal to cap, cap
60     to solvent) provides a mechanism for rapid thermal transport across
61     the interface. Our calculations also suggest that this is a
62     non-monotonic function of the fractional coverage of the surface, as
63     moderate coverages allow convective heat transport of solvent
64     molecules that have been in close contact with the capping agent.
65 gezelter 3717 \end{abstract}
66    
67     \newpage
68    
69     %\narrowtext
70    
71     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
72     % BODY OF TEXT
73     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
74    
75     \section{Introduction}
76 gezelter 3761 Due to the importance of heat flow (and heat removal) in
77     nanotechnology, interfacial thermal conductance has been studied
78     extensively both experimentally and computationally.\cite{cahill:793}
79     Nanoscale materials have a significant fraction of their atoms at
80     interfaces, and the chemical details of these interfaces govern the
81     thermal transport properties. Furthermore, the interfaces are often
82 gezelter 3751 heterogeneous (e.g. solid - liquid), which provides a challenge to
83 gezelter 3761 computational methods which have been developed for homogeneous or
84     bulk systems.
85 gezelter 3717
86 gezelter 3761 Experimentally, the thermal properties of a number of interfaces have
87     been investigated. Cahill and coworkers studied nanoscale thermal
88 skuang 3755 transport from metal nanoparticle/fluid interfaces, to epitaxial
89 gezelter 3761 TiN/single crystal oxides interfaces, and hydrophilic and hydrophobic
90 skuang 3755 interfaces between water and solids with different self-assembled
91     monolayers.\cite{Wilson:2002uq,PhysRevB.67.054302,doi:10.1021/jp048375k,PhysRevLett.96.186101}
92 gezelter 3761 Wang {\it et al.} studied heat transport through long-chain
93     hydrocarbon monolayers on gold substrate at individual molecular
94     level,\cite{Wang10082007} Schmidt {\it et al.} studied the role of
95     cetyltrimethylammonium bromide (CTAB) on the thermal transport between
96     gold nanorods and solvent,\cite{doi:10.1021/jp8051888} and Juv\'e {\it
97     et al.} studied the cooling dynamics, which is controlled by thermal
98     interface resistance of glass-embedded metal
99 gezelter 3751 nanoparticles.\cite{PhysRevB.80.195406} Although interfaces are
100     normally considered barriers for heat transport, Alper {\it et al.}
101     suggested that specific ligands (capping agents) could completely
102     eliminate this barrier
103     ($G\rightarrow\infty$).\cite{doi:10.1021/la904855s}
104 skuang 3733
105 skuang 3737 Theoretical and computational models have also been used to study the
106     interfacial thermal transport in order to gain an understanding of
107     this phenomena at the molecular level. Recently, Hase and coworkers
108     employed Non-Equilibrium Molecular Dynamics (NEMD) simulations to
109     study thermal transport from hot Au(111) substrate to a self-assembled
110 skuang 3738 monolayer of alkylthiol with relatively long chain (8-20 carbon
111 gezelter 3751 atoms).\cite{hase:2010,hase:2011} However, ensemble averaged
112 skuang 3737 measurements for heat conductance of interfaces between the capping
113 gezelter 3751 monolayer on Au and a solvent phase have yet to be studied with their
114     approach. The comparatively low thermal flux through interfaces is
115 skuang 3755 difficult to measure with Equilibrium
116     MD\cite{doi:10.1080/0026897031000068578} or forward NEMD simulation
117 skuang 3750 methods. Therefore, the Reverse NEMD (RNEMD)
118 gezelter 3761 methods\cite{MullerPlathe:1997xw,kuang:164101} would be advantageous
119     in that they {\it apply} the difficult to measure quantity (flux),
120     while {\it measuring} the easily-computed quantity (the thermal
121     gradient). This is particularly true for inhomogeneous interfaces
122     where it would not be clear how to apply a gradient {\it a priori}.
123 gezelter 3751 Garde and coworkers\cite{garde:nl2005,garde:PhysRevLett2009} applied
124     this approach to various liquid interfaces and studied how thermal
125 gezelter 3761 conductance (or resistance) is dependent on chemical details of a
126 skuang 3762 number of hydrophobic and hydrophilic aqueous interfaces. {\bf And
127     Luo {\it et al.} studied the thermal conductance of Au-SAM-Au
128     junctions using the same approach, with comparison to a constant
129     temperature difference method\cite{Luo20101}. While this latter
130     approach establishes more thermal distributions compared to the
131     former RNEMD methods, it does not guarantee momentum or kinetic
132     energy conservations.}
133 skuang 3734
134 gezelter 3751 Recently, we have developed a Non-Isotropic Velocity Scaling (NIVS)
135 skuang 3725 algorithm for RNEMD simulations\cite{kuang:164101}. This algorithm
136     retains the desirable features of RNEMD (conservation of linear
137     momentum and total energy, compatibility with periodic boundary
138     conditions) while establishing true thermal distributions in each of
139 skuang 3737 the two slabs. Furthermore, it allows effective thermal exchange
140     between particles of different identities, and thus makes the study of
141     interfacial conductance much simpler.
142 skuang 3725
143 skuang 3737 The work presented here deals with the Au(111) surface covered to
144     varying degrees by butanethiol, a capping agent with short carbon
145     chain, and solvated with organic solvents of different molecular
146 skuang 3763 properties. {\bf To our knowledge, few previous MD inverstigations
147     have been found to address to these systems yet.} Different models
148     were used for both the capping agent and the solvent force field
149 skuang 3762 parameters. Using the NIVS algorithm, the thermal transport across
150     these interfaces was studied and the underlying mechanism for the
151     phenomena was investigated.
152 skuang 3733
153 skuang 3721 \section{Methodology}
154 gezelter 3761 \subsection{Imposed-Flux Methods in MD Simulations}
155 gezelter 3751 Steady state MD simulations have an advantage in that not many
156 skuang 3749 trajectories are needed to study the relationship between thermal flux
157 gezelter 3751 and thermal gradients. For systems with low interfacial conductance,
158     one must have a method capable of generating or measuring relatively
159     small fluxes, compared to those required for bulk conductivity. This
160     requirement makes the calculation even more difficult for
161     slowly-converging equilibrium methods.\cite{Viscardy:2007lq} Forward
162     NEMD methods impose a gradient (and measure a flux), but at interfaces
163     it is not clear what behavior should be imposed at the boundaries
164     between materials. Imposed-flux reverse non-equilibrium
165     methods\cite{MullerPlathe:1997xw} set the flux {\it a priori} and
166     the thermal response becomes an easy-to-measure quantity. Although
167 skuang 3749 M\"{u}ller-Plathe's original momentum swapping approach can be used
168     for exchanging energy between particles of different identity, the
169     kinetic energy transfer efficiency is affected by the mass difference
170     between the particles, which limits its application on heterogeneous
171     interfacial systems.
172 skuang 3721
173 gezelter 3751 The non-isotropic velocity scaling (NIVS) \cite{kuang:164101} approach
174     to non-equilibrium MD simulations is able to impose a wide range of
175 skuang 3737 kinetic energy fluxes without obvious perturbation to the velocity
176     distributions of the simulated systems. Furthermore, this approach has
177 skuang 3721 the advantage in heterogeneous interfaces in that kinetic energy flux
178 gezelter 3761 can be applied between regions of particles of arbitrary identity, and
179 skuang 3737 the flux will not be restricted by difference in particle mass.
180 skuang 3721
181     The NIVS algorithm scales the velocity vectors in two separate regions
182 gezelter 3761 of a simulation system with respective diagonal scaling matrices. To
183     determine these scaling factors in the matrices, a set of equations
184 skuang 3721 including linear momentum conservation and kinetic energy conservation
185 skuang 3737 constraints and target energy flux satisfaction is solved. With the
186     scaling operation applied to the system in a set frequency, bulk
187     temperature gradients can be easily established, and these can be used
188     for computing thermal conductivities. The NIVS algorithm conserves
189     momenta and energy and does not depend on an external thermostat.
190 skuang 3721
191 gezelter 3751 \subsection{Defining Interfacial Thermal Conductivity ($G$)}
192    
193     For an interface with relatively low interfacial conductance, and a
194     thermal flux between two distinct bulk regions, the regions on either
195     side of the interface rapidly come to a state in which the two phases
196     have relatively homogeneous (but distinct) temperatures. The
197     interfacial thermal conductivity $G$ can therefore be approximated as:
198 skuang 3727 \begin{equation}
199 gezelter 3751 G = \frac{E_{total}}{2 t L_x L_y \left( \langle T_\mathrm{hot}\rangle -
200 skuang 3727 \langle T_\mathrm{cold}\rangle \right)}
201     \label{lowG}
202     \end{equation}
203 gezelter 3751 where ${E_{total}}$ is the total imposed non-physical kinetic energy
204     transfer during the simulation and ${\langle T_\mathrm{hot}\rangle}$
205     and ${\langle T_\mathrm{cold}\rangle}$ are the average observed
206 gezelter 3756 temperature of the two separated phases. For an applied flux $J_z$
207     operating over a simulation time $t$ on a periodically-replicated slab
208     of dimensions $L_x \times L_y$, $E_{total} = J_z *(t)*(2 L_x L_y)$.
209 skuang 3721
210 skuang 3737 When the interfacial conductance is {\it not} small, there are two
211 skuang 3752 ways to define $G$. One common way is to assume the temperature is
212     discrete on the two sides of the interface. $G$ can be calculated
213     using the applied thermal flux $J$ and the maximum temperature
214     difference measured along the thermal gradient max($\Delta T$), which
215 gezelter 3761 occurs at the Gibbs dividing surface (Figure \ref{demoPic}). This is
216 skuang 3755 known as the Kapitza conductance, which is the inverse of the Kapitza
217     resistance.
218 skuang 3752 \begin{equation}
219     G=\frac{J}{\Delta T}
220     \label{discreteG}
221     \end{equation}
222 skuang 3727
223 skuang 3745 \begin{figure}
224     \includegraphics[width=\linewidth]{method}
225     \caption{Interfacial conductance can be calculated by applying an
226     (unphysical) kinetic energy flux between two slabs, one located
227     within the metal and another on the edge of the periodic box. The
228 gezelter 3761 system responds by forming a thermal gradient. In bulk liquids,
229     this gradient typically has a single slope, but in interfacial
230     systems, there are distinct thermal conductivity domains. The
231     interfacial conductance, $G$ is found by measuring the temperature
232     gap at the Gibbs dividing surface, or by using second derivatives of
233     the thermal profile.}
234 skuang 3745 \label{demoPic}
235     \end{figure}
236    
237 skuang 3727 The other approach is to assume a continuous temperature profile along
238     the thermal gradient axis (e.g. $z$) and define $G$ at the point where
239 gezelter 3751 the magnitude of thermal conductivity ($\lambda$) change reaches its
240 skuang 3727 maximum, given that $\lambda$ is well-defined throughout the space:
241     \begin{equation}
242     G^\prime = \Big|\frac{\partial\lambda}{\partial z}\Big|
243     = \Big|\frac{\partial}{\partial z}\left(-J_z\Big/
244     \left(\frac{\partial T}{\partial z}\right)\right)\Big|
245     = |J_z|\Big|\frac{\partial^2 T}{\partial z^2}\Big|
246     \Big/\left(\frac{\partial T}{\partial z}\right)^2
247     \label{derivativeG}
248     \end{equation}
249    
250 gezelter 3751 With temperature profiles obtained from simulation, one is able to
251 skuang 3727 approximate the first and second derivatives of $T$ with finite
252 gezelter 3751 difference methods and calculate $G^\prime$. In what follows, both
253     definitions have been used, and are compared in the results.
254 skuang 3727
255 gezelter 3751 To investigate the interfacial conductivity at metal / solvent
256     interfaces, we have modeled a metal slab with its (111) surfaces
257     perpendicular to the $z$-axis of our simulation cells. The metal slab
258     has been prepared both with and without capping agents on the exposed
259     surface, and has been solvated with simple organic solvents, as
260 skuang 3746 illustrated in Figure \ref{gradT}.
261 skuang 3727
262 skuang 3737 With the simulation cell described above, we are able to equilibrate
263     the system and impose an unphysical thermal flux between the liquid
264     and the metal phase using the NIVS algorithm. By periodically applying
265 gezelter 3751 the unphysical flux, we obtained a temperature profile and its spatial
266     derivatives. Figure \ref{gradT} shows how an applied thermal flux can
267     be used to obtain the 1st and 2nd derivatives of the temperature
268     profile.
269 skuang 3727
270     \begin{figure}
271     \includegraphics[width=\linewidth]{gradT}
272 gezelter 3761 \caption{A sample of Au (111) / butanethiol / hexane interfacial
273     system with the temperature profile after a kinetic energy flux has
274     been imposed. Note that the largest temperature jump in the thermal
275     profile (corresponding to the lowest interfacial conductance) is at
276     the interface between the butanethiol molecules (blue) and the
277     solvent (grey). First and second derivatives of the temperature
278     profile are obtained using a finite difference approximation (lower
279     panel).}
280 skuang 3727 \label{gradT}
281     \end{figure}
282    
283     \section{Computational Details}
284 skuang 3730 \subsection{Simulation Protocol}
285 skuang 3737 The NIVS algorithm has been implemented in our MD simulation code,
286 gezelter 3751 OpenMD\cite{Meineke:2005gd,openmd}, and was used for our simulations.
287     Metal slabs of 6 or 11 layers of Au atoms were first equilibrated
288     under atmospheric pressure (1 atm) and 200K. After equilibration,
289     butanethiol capping agents were placed at three-fold hollow sites on
290     the Au(111) surfaces. These sites are either {\it fcc} or {\it
291     hcp} sites, although Hase {\it et al.} found that they are
292     equivalent in a heat transfer process,\cite{hase:2010} so we did not
293     distinguish between these sites in our study. The maximum butanethiol
294 skuang 3747 capacity on Au surface is $1/3$ of the total number of surface Au
295     atoms, and the packing forms a $(\sqrt{3}\times\sqrt{3})R30^\circ$
296 skuang 3749 structure\cite{doi:10.1021/ja00008a001,doi:10.1021/cr9801317}. A
297 gezelter 3751 series of lower coverages was also prepared by eliminating
298     butanethiols from the higher coverage surface in a regular manner. The
299     lower coverages were prepared in order to study the relation between
300     coverage and interfacial conductance.
301 skuang 3727
302 skuang 3737 The capping agent molecules were allowed to migrate during the
303     simulations. They distributed themselves uniformly and sampled a
304     number of three-fold sites throughout out study. Therefore, the
305 gezelter 3751 initial configuration does not noticeably affect the sampling of a
306 skuang 3737 variety of configurations of the same coverage, and the final
307     conductance measurement would be an average effect of these
308 gezelter 3751 configurations explored in the simulations.
309 skuang 3727
310 gezelter 3751 After the modified Au-butanethiol surface systems were equilibrated in
311     the canonical (NVT) ensemble, organic solvent molecules were packed in
312     the previously empty part of the simulation cells.\cite{packmol} Two
313 skuang 3737 solvents were investigated, one which has little vibrational overlap
314 gezelter 3751 with the alkanethiol and which has a planar shape (toluene), and one
315     which has similar vibrational frequencies to the capping agent and
316     chain-like shape ({\it n}-hexane).
317 skuang 3727
318 gezelter 3751 The simulation cells were not particularly extensive along the
319     $z$-axis, as a very long length scale for the thermal gradient may
320     cause excessively hot or cold temperatures in the middle of the
321 skuang 3730 solvent region and lead to undesired phenomena such as solvent boiling
322     or freezing when a thermal flux is applied. Conversely, too few
323     solvent molecules would change the normal behavior of the liquid
324     phase. Therefore, our $N_{solvent}$ values were chosen to ensure that
325 gezelter 3751 these extreme cases did not happen to our simulations. The spacing
326 skuang 3760 between periodic images of the gold interfaces is $45 \sim 75$\AA in
327     our simulations.
328 skuang 3730
329 skuang 3746 The initial configurations generated are further equilibrated with the
330 gezelter 3751 $x$ and $y$ dimensions fixed, only allowing the $z$-length scale to
331     change. This is to ensure that the equilibration of liquid phase does
332     not affect the metal's crystalline structure. Comparisons were made
333     with simulations that allowed changes of $L_x$ and $L_y$ during NPT
334     equilibration. No substantial changes in the box geometry were noticed
335     in these simulations. After ensuring the liquid phase reaches
336     equilibrium at atmospheric pressure (1 atm), further equilibration was
337     carried out under canonical (NVT) and microcanonical (NVE) ensembles.
338 skuang 3728
339 gezelter 3751 After the systems reach equilibrium, NIVS was used to impose an
340     unphysical thermal flux between the metal and the liquid phases. Most
341     of our simulations were done under an average temperature of
342     $\sim$200K. Therefore, thermal flux usually came from the metal to the
343 skuang 3727 liquid so that the liquid has a higher temperature and would not
344 gezelter 3751 freeze due to lowered temperatures. After this induced temperature
345 gezelter 3761 gradient had stabilized, the temperature profile of the simulation cell
346     was recorded. To do this, the simulation cell is divided evenly into
347 gezelter 3751 $N$ slabs along the $z$-axis. The average temperatures of each slab
348 skuang 3747 are recorded for 1$\sim$2 ns. When the slab width $d$ of each slab is
349     the same, the derivatives of $T$ with respect to slab number $n$ can
350 gezelter 3751 be directly used for $G^\prime$ calculations: \begin{equation}
351     G^\prime = |J_z|\Big|\frac{\partial^2 T}{\partial z^2}\Big|
352 skuang 3727 \Big/\left(\frac{\partial T}{\partial z}\right)^2
353     = |J_z|\Big|\frac{1}{d^2}\frac{\partial^2 T}{\partial n^2}\Big|
354     \Big/\left(\frac{1}{d}\frac{\partial T}{\partial n}\right)^2
355     = |J_z|\Big|\frac{\partial^2 T}{\partial n^2}\Big|
356     \Big/\left(\frac{\partial T}{\partial n}\right)^2
357     \label{derivativeG2}
358     \end{equation}
359    
360 gezelter 3751 All of the above simulation procedures use a time step of 1 fs. Each
361     equilibration stage took a minimum of 100 ps, although in some cases,
362     longer equilibration stages were utilized.
363 skuang 3747
364 skuang 3725 \subsection{Force Field Parameters}
365 gezelter 3751 Our simulations include a number of chemically distinct components.
366     Figure \ref{demoMol} demonstrates the sites defined for both
367     United-Atom and All-Atom models of the organic solvent and capping
368     agents in our simulations. Force field parameters are needed for
369 skuang 3744 interactions both between the same type of particles and between
370     particles of different species.
371 skuang 3721
372 skuang 3736 \begin{figure}
373 gezelter 3740 \includegraphics[width=\linewidth]{structures}
374     \caption{Structures of the capping agent and solvents utilized in
375     these simulations. The chemically-distinct sites (a-e) are expanded
376     in terms of constituent atoms for both United Atom (UA) and All Atom
377 gezelter 3761 (AA) force fields. Most parameters are from References
378     \protect\cite{TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes,TraPPE-UA.thiols}
379 skuang 3755 (UA) and \protect\cite{OPLSAA} (AA). Cross-interactions with the Au
380     atoms are given in Table \ref{MnM}.}
381 skuang 3736 \label{demoMol}
382     \end{figure}
383    
384 skuang 3744 The Au-Au interactions in metal lattice slab is described by the
385 gezelter 3751 quantum Sutton-Chen (QSC) formulation.\cite{PhysRevB.59.3527} The QSC
386 skuang 3744 potentials include zero-point quantum corrections and are
387     reparametrized for accurate surface energies compared to the
388 gezelter 3751 Sutton-Chen potentials.\cite{Chen90}
389 skuang 3744
390 gezelter 3751 For the two solvent molecules, {\it n}-hexane and toluene, two
391     different atomistic models were utilized. Both solvents were modeled
392     using United-Atom (UA) and All-Atom (AA) models. The TraPPE-UA
393 skuang 3728 parameters\cite{TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes} are used
394 skuang 3744 for our UA solvent molecules. In these models, sites are located at
395     the carbon centers for alkyl groups. Bonding interactions, including
396     bond stretches and bends and torsions, were used for intra-molecular
397 gezelter 3751 sites closer than 3 bonds. For non-bonded interactions, Lennard-Jones
398     potentials are used.
399 skuang 3721
400 gezelter 3751 By eliminating explicit hydrogen atoms, the TraPPE-UA models are
401     simple and computationally efficient, while maintaining good accuracy.
402 gezelter 3761 However, the TraPPE-UA model for alkanes is known to predict a slightly
403 gezelter 3751 lower boiling point than experimental values. This is one of the
404     reasons we used a lower average temperature (200K) for our
405     simulations. If heat is transferred to the liquid phase during the
406     NIVS simulation, the liquid in the hot slab can actually be
407     substantially warmer than the mean temperature in the simulation. The
408     lower mean temperatures therefore prevent solvent boiling.
409 skuang 3744
410 gezelter 3751 For UA-toluene, the non-bonded potentials between intermolecular sites
411     have a similar Lennard-Jones formulation. The toluene molecules were
412     treated as a single rigid body, so there was no need for
413     intramolecular interactions (including bonds, bends, or torsions) in
414     this solvent model.
415 skuang 3744
416 skuang 3729 Besides the TraPPE-UA models, AA models for both organic solvents are
417 skuang 3752 included in our studies as well. The OPLS-AA\cite{OPLSAA} force fields
418     were used. For hexane, additional explicit hydrogen sites were
419 skuang 3744 included. Besides bonding and non-bonded site-site interactions,
420     partial charges and the electrostatic interactions were added to each
421 skuang 3752 CT and HC site. For toluene, a flexible model for the toluene molecule
422     was utilized which included bond, bend, torsion, and inversion
423     potentials to enforce ring planarity.
424 skuang 3728
425 gezelter 3751 The butanethiol capping agent in our simulations, were also modeled
426     with both UA and AA model. The TraPPE-UA force field includes
427 skuang 3730 parameters for thiol molecules\cite{TraPPE-UA.thiols} and are used for
428     UA butanethiol model in our simulations. The OPLS-AA also provides
429     parameters for alkyl thiols. However, alkyl thiols adsorbed on Au(111)
430 gezelter 3751 surfaces do not have the hydrogen atom bonded to sulfur. To derive
431     suitable parameters for butanethiol adsorbed on Au(111) surfaces, we
432     adopt the S parameters from Luedtke and Landman\cite{landman:1998} and
433     modify the parameters for the CTS atom to maintain charge neutrality
434     in the molecule. Note that the model choice (UA or AA) for the capping
435     agent can be different from the solvent. Regardless of model choice,
436     the force field parameters for interactions between capping agent and
437     solvent can be derived using Lorentz-Berthelot Mixing Rule:
438 skuang 3738 \begin{eqnarray}
439 gezelter 3751 \sigma_{ij} & = & \frac{1}{2} \left(\sigma_{ii} + \sigma_{jj}\right) \\
440     \epsilon_{ij} & = & \sqrt{\epsilon_{ii}\epsilon_{jj}}
441 skuang 3738 \end{eqnarray}
442 skuang 3721
443 gezelter 3751 To describe the interactions between metal (Au) and non-metal atoms,
444     we refer to an adsorption study of alkyl thiols on gold surfaces by
445     Vlugt {\it et al.}\cite{vlugt:cpc2007154} They fitted an effective
446     Lennard-Jones form of potential parameters for the interaction between
447     Au and pseudo-atoms CH$_x$ and S based on a well-established and
448     widely-used effective potential of Hautman and Klein for the Au(111)
449     surface.\cite{hautman:4994} As our simulations require the gold slab
450     to be flexible to accommodate thermal excitation, the pair-wise form
451     of potentials they developed was used for our study.
452 skuang 3721
453 gezelter 3751 The potentials developed from {\it ab initio} calculations by Leng
454     {\it et al.}\cite{doi:10.1021/jp034405s} are adopted for the
455     interactions between Au and aromatic C/H atoms in toluene. However,
456     the Lennard-Jones parameters between Au and other types of particles,
457     (e.g. AA alkanes) have not yet been established. For these
458     interactions, the Lorentz-Berthelot mixing rule can be used to derive
459     effective single-atom LJ parameters for the metal using the fit values
460     for toluene. These are then used to construct reasonable mixing
461     parameters for the interactions between the gold and other atoms.
462     Table \ref{MnM} summarizes the ``metal/non-metal'' parameters used in
463     our simulations.
464 skuang 3725
465 skuang 3730 \begin{table*}
466     \begin{minipage}{\linewidth}
467     \begin{center}
468 gezelter 3741 \caption{Non-bonded interaction parameters (including cross
469     interactions with Au atoms) for both force fields used in this
470     work.}
471     \begin{tabular}{lllllll}
472 skuang 3730 \hline\hline
473 gezelter 3741 & Site & $\sigma_{ii}$ & $\epsilon_{ii}$ & $q_i$ &
474     $\sigma_{Au-i}$ & $\epsilon_{Au-i}$ \\
475     & & (\AA) & (kcal/mol) & ($e$) & (\AA) & (kcal/mol) \\
476 skuang 3730 \hline
477 gezelter 3741 United Atom (UA)
478     &CH3 & 3.75 & 0.1947 & - & 3.54 & 0.2146 \\
479     &CH2 & 3.95 & 0.0914 & - & 3.54 & 0.1749 \\
480     &CHar & 3.695 & 0.1003 & - & 3.4625 & 0.1680 \\
481     &CRar & 3.88 & 0.04173 & - & 3.555 & 0.1604 \\
482     \hline
483     All Atom (AA)
484     &CT3 & 3.50 & 0.066 & -0.18 & 3.365 & 0.1373 \\
485     &CT2 & 3.50 & 0.066 & -0.12 & 3.365 & 0.1373 \\
486     &CTT & 3.50 & 0.066 & -0.065 & 3.365 & 0.1373 \\
487     &HC & 2.50 & 0.030 & 0.06 & 2.865 & 0.09256 \\
488     &CA & 3.55 & 0.070 & -0.115 & 3.173 & 0.0640 \\
489     &HA & 2.42 & 0.030 & 0.115 & 2.746 & 0.0414 \\
490     \hline
491 skuang 3744 Both UA and AA
492     & S & 4.45 & 0.25 & - & 2.40 & 8.465 \\
493 skuang 3730 \hline\hline
494     \end{tabular}
495     \label{MnM}
496     \end{center}
497     \end{minipage}
498     \end{table*}
499 skuang 3729
500 gezelter 3751
501 gezelter 3754 \section{Results}
502     There are many factors contributing to the measured interfacial
503     conductance; some of these factors are physically motivated
504     (e.g. coverage of the surface by the capping agent coverage and
505     solvent identity), while some are governed by parameters of the
506     methodology (e.g. applied flux and the formulas used to obtain the
507     conductance). In this section we discuss the major physical and
508     calculational effects on the computed conductivity.
509 skuang 3746
510 gezelter 3754 \subsection{Effects due to capping agent coverage}
511 skuang 3747
512 gezelter 3754 A series of different initial conditions with a range of surface
513     coverages was prepared and solvated with various with both of the
514     solvent molecules. These systems were then equilibrated and their
515 skuang 3755 interfacial thermal conductivity was measured with the NIVS
516 gezelter 3754 algorithm. Figure \ref{coverage} demonstrates the trend of conductance
517     with respect to surface coverage.
518    
519     \begin{figure}
520     \includegraphics[width=\linewidth]{coverage}
521 gezelter 3761 \caption{The interfacial thermal conductivity ($G$) has a
522     non-monotonic dependence on the degree of surface capping. This
523     data is for the Au(111) / butanethiol / solvent interface with
524     various UA force fields at $\langle T\rangle \sim $200K.}
525 gezelter 3754 \label{coverage}
526     \end{figure}
527    
528 gezelter 3756 In partially covered surfaces, the derivative definition for
529     $G^\prime$ (Eq. \ref{derivativeG}) becomes difficult to apply, as the
530     location of maximum change of $\lambda$ becomes washed out. The
531     discrete definition (Eq. \ref{discreteG}) is easier to apply, as the
532     Gibbs dividing surface is still well-defined. Therefore, $G$ (not
533     $G^\prime$) was used in this section.
534 gezelter 3754
535 gezelter 3756 From Figure \ref{coverage}, one can see the significance of the
536     presence of capping agents. When even a small fraction of the Au(111)
537     surface sites are covered with butanethiols, the conductivity exhibits
538 gezelter 3761 an enhancement by at least a factor of 3. Capping agents are clearly
539 gezelter 3756 playing a major role in thermal transport at metal / organic solvent
540     surfaces.
541 gezelter 3754
542 gezelter 3756 We note a non-monotonic behavior in the interfacial conductance as a
543     function of surface coverage. The maximum conductance (largest $G$)
544     happens when the surfaces are about 75\% covered with butanethiol
545     caps. The reason for this behavior is not entirely clear. One
546     explanation is that incomplete butanethiol coverage allows small gaps
547     between butanethiols to form. These gaps can be filled by transient
548     solvent molecules. These solvent molecules couple very strongly with
549     the hot capping agent molecules near the surface, and can then carry
550     away (diffusively) the excess thermal energy from the surface.
551 gezelter 3754
552 gezelter 3756 There appears to be a competition between the conduction of the
553     thermal energy away from the surface by the capping agents (enhanced
554     by greater coverage) and the coupling of the capping agents with the
555     solvent (enhanced by interdigitation at lower coverages). This
556     competition would lead to the non-monotonic coverage behavior observed
557     here.
558 gezelter 3754
559 gezelter 3756 Results for rigid body toluene solvent, as well as the UA hexane, are
560     within the ranges expected from prior experimental
561     work.\cite{Wilson:2002uq,cahill:793,PhysRevB.80.195406} This suggests
562     that explicit hydrogen atoms might not be required for modeling
563     thermal transport in these systems. C-H vibrational modes do not see
564     significant excited state population at low temperatures, and are not
565     likely to carry lower frequency excitations from the solid layer into
566     the bulk liquid.
567 gezelter 3754
568 gezelter 3756 The toluene solvent does not exhibit the same behavior as hexane in
569     that $G$ remains at approximately the same magnitude when the capping
570     coverage increases from 25\% to 75\%. Toluene, as a rigid planar
571     molecule, cannot occupy the relatively small gaps between the capping
572     agents as easily as the chain-like {\it n}-hexane. The effect of
573     solvent coupling to the capping agent is therefore weaker in toluene
574     except at the very lowest coverage levels. This effect counters the
575     coverage-dependent conduction of heat away from the metal surface,
576     leading to a much flatter $G$ vs. coverage trend than is observed in
577     {\it n}-hexane.
578 gezelter 3754
579     \subsection{Effects due to Solvent \& Solvent Models}
580 gezelter 3756 In addition to UA solvent and capping agent models, AA models have
581     also been included in our simulations. In most of this work, the same
582     (UA or AA) model for solvent and capping agent was used, but it is
583     also possible to utilize different models for different components.
584     We have also included isotopic substitutions (Hydrogen to Deuterium)
585     to decrease the explicit vibrational overlap between solvent and
586     capping agent. Table \ref{modelTest} summarizes the results of these
587     studies.
588 gezelter 3754
589     \begin{table*}
590     \begin{minipage}{\linewidth}
591     \begin{center}
592    
593 skuang 3755 \caption{Computed interfacial thermal conductance ($G$ and
594 gezelter 3754 $G^\prime$) values for interfaces using various models for
595     solvent and capping agent (or without capping agent) at
596 gezelter 3761 $\langle T\rangle\sim$200K. Here ``D'' stands for deuterated
597     solvent or capping agent molecules; ``Avg.'' denotes results
598     that are averages of simulations under different applied
599     thermal flux $(J_z)$ values. Error estimates are indicated in
600     parentheses.}
601 gezelter 3754
602     \begin{tabular}{llccc}
603     \hline\hline
604     Butanethiol model & Solvent & $J_z$ & $G$ & $G^\prime$ \\
605     (or bare surface) & model & (GW/m$^2$) &
606     \multicolumn{2}{c}{(MW/m$^2$/K)} \\
607     \hline
608     UA & UA hexane & Avg. & 131(9) & 87(10) \\
609     & UA hexane(D) & 1.95 & 153(5) & 136(13) \\
610     & AA hexane & Avg. & 131(6) & 122(10) \\
611     & UA toluene & 1.96 & 187(16) & 151(11) \\
612     & AA toluene & 1.89 & 200(36) & 149(53) \\
613     \hline
614     AA & UA hexane & 1.94 & 116(9) & 129(8) \\
615     & AA hexane & Avg. & 442(14) & 356(31) \\
616     & AA hexane(D) & 1.93 & 222(12) & 234(54) \\
617     & UA toluene & 1.98 & 125(25) & 97(60) \\
618     & AA toluene & 3.79 & 487(56) & 290(42) \\
619     \hline
620     AA(D) & UA hexane & 1.94 & 158(25) & 172(4) \\
621     & AA hexane & 1.92 & 243(29) & 191(11) \\
622     & AA toluene & 1.93 & 364(36) & 322(67) \\
623     \hline
624     bare & UA hexane & Avg. & 46.5(3.2) & 49.4(4.5) \\
625     & UA hexane(D) & 0.98 & 43.9(4.6) & 43.0(2.0) \\
626     & AA hexane & 0.96 & 31.0(1.4) & 29.4(1.3) \\
627     & UA toluene & 1.99 & 70.1(1.3) & 65.8(0.5) \\
628     \hline\hline
629     \end{tabular}
630     \label{modelTest}
631     \end{center}
632     \end{minipage}
633     \end{table*}
634    
635 gezelter 3756 To facilitate direct comparison between force fields, systems with the
636     same capping agent and solvent were prepared with the same length
637     scales for the simulation cells.
638 gezelter 3754
639 gezelter 3756 On bare metal / solvent surfaces, different force field models for
640     hexane yield similar results for both $G$ and $G^\prime$, and these
641     two definitions agree with each other very well. This is primarily an
642     indicator of weak interactions between the metal and the solvent, and
643     is a typical case for acoustic impedance mismatch between these two
644     phases.
645 gezelter 3754
646 gezelter 3756 For the fully-covered surfaces, the choice of force field for the
647 gezelter 3761 capping agent and solvent has a large impact on the calculated values
648 gezelter 3756 of $G$ and $G^\prime$. The AA thiol to AA solvent conductivities are
649     much larger than their UA to UA counterparts, and these values exceed
650     the experimental estimates by a large measure. The AA force field
651     allows significant energy to go into C-H (or C-D) stretching modes,
652     and since these modes are high frequency, this non-quantum behavior is
653     likely responsible for the overestimate of the conductivity. Compared
654     to the AA model, the UA model yields more reasonable conductivity
655     values with much higher computational efficiency.
656 skuang 3755
657     \subsubsection{Are electronic excitations in the metal important?}
658 gezelter 3756 Because they lack electronic excitations, the QSC and related embedded
659     atom method (EAM) models for gold are known to predict unreasonably
660     low values for bulk conductivity
661     ($\lambda$).\cite{kuang:164101,ISI:000207079300006,Clancy:1992} If the
662     conductance between the phases ($G$) is governed primarily by phonon
663     excitation (and not electronic degrees of freedom), one would expect a
664     classical model to capture most of the interfacial thermal
665     conductance. Our results for $G$ and $G^\prime$ indicate that this is
666     indeed the case, and suggest that the modeling of interfacial thermal
667     transport depends primarily on the description of the interactions
668     between the various components at the interface. When the metal is
669     chemically capped, the primary barrier to thermal conductivity appears
670     to be the interface between the capping agent and the surrounding
671     solvent, so the excitations in the metal have little impact on the
672     value of $G$.
673 gezelter 3754
674     \subsection{Effects due to methodology and simulation parameters}
675    
676 gezelter 3756 We have varied the parameters of the simulations in order to
677     investigate how these factors would affect the computation of $G$. Of
678     particular interest are: 1) the length scale for the applied thermal
679     gradient (modified by increasing the amount of solvent in the system),
680     2) the sign and magnitude of the applied thermal flux, 3) the average
681     temperature of the simulation (which alters the solvent density during
682     equilibration), and 4) the definition of the interfacial conductance
683     (Eqs. (\ref{discreteG}) or (\ref{derivativeG})) used in the
684     calculation.
685 skuang 3725
686 gezelter 3756 Systems of different lengths were prepared by altering the number of
687     solvent molecules and extending the length of the box along the $z$
688     axis to accomodate the extra solvent. Equilibration at the same
689     temperature and pressure conditions led to nearly identical surface
690     areas ($L_x$ and $L_y$) available to the metal and capping agent,
691     while the extra solvent served mainly to lengthen the axis that was
692     used to apply the thermal flux. For a given value of the applied
693     flux, the different $z$ length scale has only a weak effect on the
694     computed conductivities (Table \ref{AuThiolHexaneUA}).
695 skuang 3725
696 gezelter 3756 \subsubsection{Effects of applied flux}
697     The NIVS algorithm allows changes in both the sign and magnitude of
698     the applied flux. It is possible to reverse the direction of heat
699     flow simply by changing the sign of the flux, and thermal gradients
700     which would be difficult to obtain experimentally ($5$ K/\AA) can be
701     easily simulated. However, the magnitude of the applied flux is not
702 gezelter 3761 arbitrary if one aims to obtain a stable and reliable thermal gradient.
703 gezelter 3756 A temperature gradient can be lost in the noise if $|J_z|$ is too
704     small, and excessive $|J_z|$ values can cause phase transitions if the
705     extremes of the simulation cell become widely separated in
706     temperature. Also, if $|J_z|$ is too large for the bulk conductivity
707     of the materials, the thermal gradient will never reach a stable
708     state.
709 skuang 3755
710 gezelter 3756 Within a reasonable range of $J_z$ values, we were able to study how
711     $G$ changes as a function of this flux. In what follows, we use
712     positive $J_z$ values to denote the case where energy is being
713     transferred by the method from the metal phase and into the liquid.
714     The resulting gradient therefore has a higher temperature in the
715     liquid phase. Negative flux values reverse this transfer, and result
716     in higher temperature metal phases. The conductance measured under
717     different applied $J_z$ values is listed in Tables
718     \ref{AuThiolHexaneUA} and \ref{AuThiolToluene}. These results do not
719     indicate that $G$ depends strongly on $J_z$ within this flux
720     range. The linear response of flux to thermal gradient simplifies our
721     investigations in that we can rely on $G$ measurement with only a
722     small number $J_z$ values.
723 skuang 3730
724 skuang 3725 \begin{table*}
725     \begin{minipage}{\linewidth}
726     \begin{center}
727 gezelter 3761 \caption{In the hexane-solvated interfaces, the system size has
728     little effect on the calculated values for interfacial
729     conductance ($G$ and $G^\prime$), but the direction of heat
730     flow (i.e. the sign of $J_z$) can alter the average
731     temperature of the liquid phase and this can alter the
732     computed conductivity.}
733 skuang 3730
734 skuang 3738 \begin{tabular}{ccccccc}
735 skuang 3730 \hline\hline
736 gezelter 3756 $\langle T\rangle$ & $N_{hexane}$ & $\rho_{hexane}$ &
737 skuang 3738 $J_z$ & $G$ & $G^\prime$ \\
738 gezelter 3756 (K) & & (g/cm$^3$) & (GW/m$^2$) &
739 skuang 3730 \multicolumn{2}{c}{(MW/m$^2$/K)} \\
740     \hline
741 gezelter 3756 200 & 266 & 0.672 & -0.96 & 102(3) & 80.0(0.8) \\
742     & 200 & 0.688 & 0.96 & 125(16) & 90.2(15) \\
743     & & & 1.91 & 139(10) & 101(10) \\
744     & & & 2.83 & 141(6) & 89.9(9.8) \\
745     & 166 & 0.681 & 0.97 & 141(30) & 78(22) \\
746     & & & 1.92 & 138(4) & 98.9(9.5) \\
747 skuang 3739 \hline
748 gezelter 3756 250 & 200 & 0.560 & 0.96 & 75(10) & 61.8(7.3) \\
749     & & & -0.95 & 49.4(0.3) & 45.7(2.1) \\
750     & 166 & 0.569 & 0.97 & 80.3(0.6) & 67(11) \\
751     & & & 1.44 & 76.2(5.0) & 64.8(3.8) \\
752     & & & -0.95 & 56.4(2.5) & 54.4(1.1) \\
753     & & & -1.85 & 47.8(1.1) & 53.5(1.5) \\
754 skuang 3730 \hline\hline
755     \end{tabular}
756     \label{AuThiolHexaneUA}
757     \end{center}
758     \end{minipage}
759     \end{table*}
760    
761 gezelter 3756 The sign of $J_z$ is a different matter, however, as this can alter
762     the temperature on the two sides of the interface. The average
763     temperature values reported are for the entire system, and not for the
764     liquid phase, so at a given $\langle T \rangle$, the system with
765     positive $J_z$ has a warmer liquid phase. This means that if the
766     liquid carries thermal energy via convective transport, {\it positive}
767     $J_z$ values will result in increased molecular motion on the liquid
768     side of the interface, and this will increase the measured
769     conductivity.
770    
771 gezelter 3754 \subsubsection{Effects due to average temperature}
772    
773 gezelter 3756 We also studied the effect of average system temperature on the
774     interfacial conductance. The simulations are first equilibrated in
775     the NPT ensemble at 1 atm. The TraPPE-UA model for hexane tends to
776     predict a lower boiling point (and liquid state density) than
777     experiments. This lower-density liquid phase leads to reduced contact
778     between the hexane and butanethiol, and this accounts for our
779     observation of lower conductance at higher temperatures as shown in
780     Table \ref{AuThiolHexaneUA}. In raising the average temperature from
781 skuang 3760 200K to 250K, the density drop of $\sim$20\% in the solvent phase
782 gezelter 3761 leads to a $\sim$40\% drop in the conductance.
783 skuang 3730
784 gezelter 3756 Similar behavior is observed in the TraPPE-UA model for toluene,
785     although this model has better agreement with the experimental
786     densities of toluene. The expansion of the toluene liquid phase is
787     not as significant as that of the hexane (8.3\% over 100K), and this
788 skuang 3760 limits the effect to $\sim$20\% drop in thermal conductivity (Table
789 gezelter 3756 \ref{AuThiolToluene}).
790 skuang 3730
791 gezelter 3756 Although we have not mapped out the behavior at a large number of
792     temperatures, is clear that there will be a strong temperature
793     dependence in the interfacial conductance when the physical properties
794     of one side of the interface (notably the density) change rapidly as a
795     function of temperature.
796    
797 skuang 3730 \begin{table*}
798     \begin{minipage}{\linewidth}
799     \begin{center}
800 gezelter 3761 \caption{When toluene is the solvent, the interfacial thermal
801     conductivity is less sensitive to temperature, but again, the
802     direction of the heat flow can alter the solvent temperature
803     and can change the computed conductance values.}
804 skuang 3725
805 skuang 3738 \begin{tabular}{ccccc}
806 skuang 3725 \hline\hline
807 skuang 3738 $\langle T\rangle$ & $\rho_{toluene}$ & $J_z$ & $G$ & $G^\prime$ \\
808     (K) & (g/cm$^3$) & (GW/m$^2$) & \multicolumn{2}{c}{(MW/m$^2$/K)} \\
809 skuang 3725 \hline
810 skuang 3745 200 & 0.933 & 2.15 & 204(12) & 113(12) \\
811     & & -1.86 & 180(3) & 135(21) \\
812     & & -3.93 & 176(5) & 113(12) \\
813 skuang 3738 \hline
814 skuang 3745 300 & 0.855 & -1.91 & 143(5) & 125(2) \\
815     & & -4.19 & 135(9) & 113(12) \\
816 skuang 3725 \hline\hline
817     \end{tabular}
818     \label{AuThiolToluene}
819     \end{center}
820     \end{minipage}
821     \end{table*}
822    
823 gezelter 3756 Besides the lower interfacial thermal conductance, surfaces at
824     relatively high temperatures are susceptible to reconstructions,
825     particularly when butanethiols fully cover the Au(111) surface. These
826     reconstructions include surface Au atoms which migrate outward to the
827     S atom layer, and butanethiol molecules which embed into the surface
828     Au layer. The driving force for this behavior is the strong Au-S
829     interactions which are modeled here with a deep Lennard-Jones
830 gezelter 3761 potential. This phenomenon agrees with reconstructions that have been
831 gezelter 3756 experimentally
832     observed.\cite{doi:10.1021/j100035a033,doi:10.1021/la026493y}. Vlugt
833     {\it et al.} kept their Au(111) slab rigid so that their simulations
834     could reach 300K without surface
835     reconstructions.\cite{vlugt:cpc2007154} Since surface reconstructions
836     blur the interface, the measurement of $G$ becomes more difficult to
837     conduct at higher temperatures. For this reason, most of our
838     measurements are undertaken at $\langle T\rangle\sim$200K where
839     reconstruction is minimized.
840 skuang 3725
841 skuang 3730 However, when the surface is not completely covered by butanethiols,
842 gezelter 3756 the simulated system appears to be more resistent to the
843 skuang 3760 reconstruction. Our Au / butanethiol / toluene system had the Au(111)
844 gezelter 3756 surfaces 90\% covered by butanethiols, but did not see this above
845     phenomena even at $\langle T\rangle\sim$300K. That said, we did
846     observe butanethiols migrating to neighboring three-fold sites during
847 skuang 3763 a simulation. Since the interface persisted in these simulations, we
848 gezelter 3756 were able to obtain $G$'s for these interfaces even at a relatively
849     high temperature without being affected by surface reconstructions.
850 skuang 3725
851 gezelter 3754 \section{Discussion}
852 skuang 3748
853 gezelter 3756 The primary result of this work is that the capping agent acts as an
854     efficient thermal coupler between solid and solvent phases. One of
855     the ways the capping agent can carry out this role is to down-shift
856     between the phonon vibrations in the solid (which carry the heat from
857     the gold) and the molecular vibrations in the liquid (which carry some
858     of the heat in the solvent).
859    
860 gezelter 3754 To investigate the mechanism of interfacial thermal conductance, the
861     vibrational power spectrum was computed. Power spectra were taken for
862     individual components in different simulations. To obtain these
863 gezelter 3756 spectra, simulations were run after equilibration in the
864     microcanonical (NVE) ensemble and without a thermal
865     gradient. Snapshots of configurations were collected at a frequency
866 gezelter 3761 that is higher than that of the fastest vibrations occurring in the
867 gezelter 3756 simulations. With these configurations, the velocity auto-correlation
868     functions can be computed:
869 gezelter 3754 \begin{equation}
870     C_A (t) = \langle\vec{v}_A (t)\cdot\vec{v}_A (0)\rangle
871     \label{vCorr}
872     \end{equation}
873     The power spectrum is constructed via a Fourier transform of the
874     symmetrized velocity autocorrelation function,
875     \begin{equation}
876     \hat{f}(\omega) =
877     \int_{-\infty}^{\infty} C_A (t) e^{-2\pi it\omega}\,dt
878     \label{fourier}
879     \end{equation}
880 skuang 3725
881 gezelter 3756 \subsection{The role of specific vibrations}
882 skuang 3747 The vibrational spectra for gold slabs in different environments are
883     shown as in Figure \ref{specAu}. Regardless of the presence of
884 gezelter 3756 solvent, the gold surfaces which are covered by butanethiol molecules
885     exhibit an additional peak observed at a frequency of
886 skuang 3759 $\sim$165cm$^{-1}$. We attribute this peak to the S-Au bonding
887 gezelter 3756 vibration. This vibration enables efficient thermal coupling of the
888     surface Au layer to the capping agents. Therefore, in our simulations,
889     the Au / S interfaces do not appear to be the primary barrier to
890     thermal transport when compared with the butanethiol / solvent
891 skuang 3763 interfaces. {\bf This confirms the results from Luo {\it et
892     al.}\cite{Luo20101}, which reported $G$ for Au-SAM junctions
893     generally twice larger than what we have computed for the
894     thiol-liquid interfaces.}
895 skuang 3732
896 skuang 3725 \begin{figure}
897     \includegraphics[width=\linewidth]{vibration}
898 gezelter 3761 \caption{The vibrational power spectrum for thiol-capped gold has an
899     additional vibrational peak at $\sim $165cm$^{-1}$. Bare gold
900     surfaces (both with and without a solvent over-layer) are missing
901     this peak. A similar peak at $\sim $165cm$^{-1}$ also appears in
902     the vibrational power spectrum for the butanethiol capping agents.}
903 skuang 3747 \label{specAu}
904 skuang 3725 \end{figure}
905    
906 gezelter 3756 Also in this figure, we show the vibrational power spectrum for the
907     bound butanethiol molecules, which also exhibits the same
908 skuang 3759 $\sim$165cm$^{-1}$ peak.
909 gezelter 3756
910     \subsection{Overlap of power spectra}
911 skuang 3755 A comparison of the results obtained from the two different organic
912     solvents can also provide useful information of the interfacial
913 gezelter 3756 thermal transport process. In particular, the vibrational overlap
914     between the butanethiol and the organic solvents suggests a highly
915     efficient thermal exchange between these components. Very high
916     thermal conductivity was observed when AA models were used and C-H
917     vibrations were treated classically. The presence of extra degrees of
918     freedom in the AA force field yields higher heat exchange rates
919     between the two phases and results in a much higher conductivity than
920 skuang 3763 in the UA force field. {\bf Due to the classical models used, this
921     even includes those high frequency modes which should be unpopulated
922     at our relatively low temperatures. This artifact causes high
923     frequency vibrations accountable for thermal transport in classical
924     MD simulations.}
925 skuang 3732
926 gezelter 3756 The similarity in the vibrational modes available to solvent and
927     capping agent can be reduced by deuterating one of the two components
928     (Fig. \ref{aahxntln}). Once either the hexanes or the butanethiols
929     are deuterated, one can observe a significantly lower $G$ and
930     $G^\prime$ values (Table \ref{modelTest}).
931    
932 skuang 3755 \begin{figure}
933 gezelter 3756 \includegraphics[width=\linewidth]{aahxntln}
934     \caption{Spectra obtained for all-atom (AA) Au / butanethiol / solvent
935     systems. When butanethiol is deuterated (lower left), its
936     vibrational overlap with hexane decreases significantly. Since
937     aromatic molecules and the butanethiol are vibrationally dissimilar,
938     the change is not as dramatic when toluene is the solvent (right).}
939     \label{aahxntln}
940     \end{figure}
941    
942     For the Au / butanethiol / toluene interfaces, having the AA
943     butanethiol deuterated did not yield a significant change in the
944     measured conductance. Compared to the C-H vibrational overlap between
945     hexane and butanethiol, both of which have alkyl chains, the overlap
946     between toluene and butanethiol is not as significant and thus does
947     not contribute as much to the heat exchange process.
948    
949     Previous observations of Zhang {\it et al.}\cite{hase:2010} indicate
950     that the {\it intra}molecular heat transport due to alkylthiols is
951     highly efficient. Combining our observations with those of Zhang {\it
952     et al.}, it appears that butanethiol acts as a channel to expedite
953     heat flow from the gold surface and into the alkyl chain. The
954     acoustic impedance mismatch between the metal and the liquid phase can
955     therefore be effectively reduced with the presence of suitable capping
956     agents.
957    
958     Deuterated models in the UA force field did not decouple the thermal
959     transport as well as in the AA force field. The UA models, even
960     though they have eliminated the high frequency C-H vibrational
961     overlap, still have significant overlap in the lower-frequency
962     portions of the infrared spectra (Figure \ref{uahxnua}). Deuterating
963     the UA models did not decouple the low frequency region enough to
964     produce an observable difference for the results of $G$ (Table
965     \ref{modelTest}).
966    
967     \begin{figure}
968 skuang 3755 \includegraphics[width=\linewidth]{uahxnua}
969 gezelter 3761 \caption{Vibrational power spectra for UA models for the butanethiol
970     and hexane solvent (upper panel) show the high degree of overlap
971     between these two molecules, particularly at lower frequencies.
972     Deuterating a UA model for the solvent (lower panel) does not
973     decouple the two spectra to the same degree as in the AA force
974     field (see Fig \ref{aahxntln}).}
975 skuang 3755 \label{uahxnua}
976     \end{figure}
977    
978 skuang 3730 \section{Conclusions}
979 gezelter 3756 The NIVS algorithm has been applied to simulations of
980     butanethiol-capped Au(111) surfaces in the presence of organic
981     solvents. This algorithm allows the application of unphysical thermal
982     flux to transfer heat between the metal and the liquid phase. With the
983     flux applied, we were able to measure the corresponding thermal
984     gradients and to obtain interfacial thermal conductivities. Under
985     steady states, 2-3 ns trajectory simulations are sufficient for
986     computation of this quantity.
987 skuang 3747
988 gezelter 3756 Our simulations have seen significant conductance enhancement in the
989     presence of capping agent, compared with the bare gold / liquid
990 skuang 3747 interfaces. The acoustic impedance mismatch between the metal and the
991 gezelter 3756 liquid phase is effectively eliminated by a chemically-bonded capping
992 gezelter 3761 agent. Furthermore, the coverage percentage of the capping agent plays
993 skuang 3747 an important role in the interfacial thermal transport
994 gezelter 3756 process. Moderately low coverages allow higher contact between capping
995     agent and solvent, and thus could further enhance the heat transfer
996     process, giving a non-monotonic behavior of conductance with
997     increasing coverage.
998 skuang 3725
999 gezelter 3756 Our results, particularly using the UA models, agree well with
1000     available experimental data. The AA models tend to overestimate the
1001 skuang 3732 interfacial thermal conductance in that the classically treated C-H
1002 gezelter 3756 vibrations become too easily populated. Compared to the AA models, the
1003     UA models have higher computational efficiency with satisfactory
1004     accuracy, and thus are preferable in modeling interfacial thermal
1005     transport.
1006    
1007     Of the two definitions for $G$, the discrete form
1008 skuang 3747 (Eq. \ref{discreteG}) was easier to use and gives out relatively
1009     consistent results, while the derivative form (Eq. \ref{derivativeG})
1010     is not as versatile. Although $G^\prime$ gives out comparable results
1011     and follows similar trend with $G$ when measuring close to fully
1012 gezelter 3756 covered or bare surfaces, the spatial resolution of $T$ profile
1013     required for the use of a derivative form is limited by the number of
1014     bins and the sampling required to obtain thermal gradient information.
1015 skuang 3730
1016 gezelter 3756 Vlugt {\it et al.} have investigated the surface thiol structures for
1017     nanocrystalline gold and pointed out that they differ from those of
1018     the Au(111) surface.\cite{landman:1998,vlugt:cpc2007154} This
1019     difference could also cause differences in the interfacial thermal
1020     transport behavior. To investigate this problem, one would need an
1021     effective method for applying thermal gradients in non-planar
1022     (i.e. spherical) geometries.
1023 skuang 3730
1024 gezelter 3717 \section{Acknowledgments}
1025     Support for this project was provided by the National Science
1026     Foundation under grant CHE-0848243. Computational time was provided by
1027     the Center for Research Computing (CRC) at the University of Notre
1028 gezelter 3754 Dame.
1029 skuang 3762
1030     \section{Supporting Information}
1031     This information is available free of charge via the Internet at
1032     http://pubs.acs.org.
1033    
1034 gezelter 3754 \newpage
1035 gezelter 3717
1036     \bibliography{interfacial}
1037    
1038     \end{doublespace}
1039     \end{document}
1040