1 |
\documentclass[11pt]{article} |
2 |
\usepackage{amsmath} |
3 |
\usepackage{amssymb} |
4 |
\usepackage{setspace} |
5 |
\usepackage{endfloat} |
6 |
\usepackage{caption} |
7 |
%\usepackage{tabularx} |
8 |
\usepackage{graphicx} |
9 |
\usepackage{multirow} |
10 |
%\usepackage{booktabs} |
11 |
%\usepackage{bibentry} |
12 |
%\usepackage{mathrsfs} |
13 |
%\usepackage[ref]{overcite} |
14 |
\usepackage[square, comma, sort&compress]{natbib} |
15 |
\usepackage{url} |
16 |
\pagestyle{plain} \pagenumbering{arabic} \oddsidemargin 0.0cm |
17 |
\evensidemargin 0.0cm \topmargin -21pt \headsep 10pt \textheight |
18 |
9.0in \textwidth 6.5in \brokenpenalty=10000 |
19 |
|
20 |
% double space list of tables and figures |
21 |
\AtBeginDelayedFloats{\renewcommand{\baselinestretch}{1.66}} |
22 |
\setlength{\abovecaptionskip}{20 pt} |
23 |
\setlength{\belowcaptionskip}{30 pt} |
24 |
|
25 |
%\renewcommand\citemid{\ } % no comma in optional reference note |
26 |
\bibpunct{[}{]}{,}{n}{}{;} |
27 |
\bibliographystyle{achemso} |
28 |
|
29 |
\begin{document} |
30 |
|
31 |
\title{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 |
|
48 |
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 |
|
65 |
\end{abstract} |
66 |
|
67 |
\newpage |
68 |
|
69 |
%\narrowtext |
70 |
|
71 |
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
72 |
% BODY OF TEXT |
73 |
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
74 |
|
75 |
\section{Introduction} |
76 |
Due to the importance of heat flow in nanotechnology, interfacial |
77 |
thermal conductance has been studied extensively both experimentally |
78 |
and computationally.\cite{cahill:793} Unlike bulk materials, nanoscale |
79 |
materials have a significant fraction of their atoms at interfaces, |
80 |
and the chemical details of these interfaces govern the heat transfer |
81 |
behavior. Furthermore, the interfaces are |
82 |
heterogeneous (e.g. solid - liquid), which provides a challenge to |
83 |
traditional methods developed for homogeneous systems. |
84 |
|
85 |
Experimentally, various interfaces have been investigated for their |
86 |
thermal conductance. Cahill and coworkers studied nanoscale thermal |
87 |
transport from metal nanoparticle/fluid interfaces, to epitaxial |
88 |
TiN/single crystal oxides interfaces, to hydrophilic and hydrophobic |
89 |
interfaces between water and solids with different self-assembled |
90 |
monolayers.\cite{Wilson:2002uq,PhysRevB.67.054302,doi:10.1021/jp048375k,PhysRevLett.96.186101} |
91 |
Wang {\it et al.} studied heat transport through |
92 |
long-chain hydrocarbon monolayers on gold substrate at individual |
93 |
molecular level,\cite{Wang10082007} Schmidt {\it et al.} studied the |
94 |
role of CTAB on thermal transport between gold nanorods and |
95 |
solvent,\cite{doi:10.1021/jp8051888} and Juv\'e {\it et al.} studied |
96 |
the cooling dynamics, which is controlled by thermal interface |
97 |
resistence of glass-embedded metal |
98 |
nanoparticles.\cite{PhysRevB.80.195406} Although interfaces are |
99 |
normally considered barriers for heat transport, Alper {\it et al.} |
100 |
suggested that specific ligands (capping agents) could completely |
101 |
eliminate this barrier |
102 |
($G\rightarrow\infty$).\cite{doi:10.1021/la904855s} |
103 |
|
104 |
Theoretical and computational models have also been used to study the |
105 |
interfacial thermal transport in order to gain an understanding of |
106 |
this phenomena at the molecular level. Recently, Hase and coworkers |
107 |
employed Non-Equilibrium Molecular Dynamics (NEMD) simulations to |
108 |
study thermal transport from hot Au(111) substrate to a self-assembled |
109 |
monolayer of alkylthiol with relatively long chain (8-20 carbon |
110 |
atoms).\cite{hase:2010,hase:2011} However, ensemble averaged |
111 |
measurements for heat conductance of interfaces between the capping |
112 |
monolayer on Au and a solvent phase have yet to be studied with their |
113 |
approach. The comparatively low thermal flux through interfaces is |
114 |
difficult to measure with Equilibrium |
115 |
MD\cite{doi:10.1080/0026897031000068578} or forward NEMD simulation |
116 |
methods. Therefore, the Reverse NEMD (RNEMD) |
117 |
methods\cite{MullerPlathe:1997xw,kuang:164101} would have the |
118 |
advantage of applying this difficult to measure flux (while measuring |
119 |
the resulting gradient), given that the simulation methods being able |
120 |
to effectively apply an unphysical flux in non-homogeneous systems. |
121 |
Garde and coworkers\cite{garde:nl2005,garde:PhysRevLett2009} applied |
122 |
this approach to various liquid interfaces and studied how thermal |
123 |
conductance (or resistance) is dependent on chemistry details of |
124 |
interfaces, e.g. hydrophobic and hydrophilic aqueous interfaces. |
125 |
|
126 |
Recently, we have developed a Non-Isotropic Velocity Scaling (NIVS) |
127 |
algorithm for RNEMD simulations\cite{kuang:164101}. This algorithm |
128 |
retains the desirable features of RNEMD (conservation of linear |
129 |
momentum and total energy, compatibility with periodic boundary |
130 |
conditions) while establishing true thermal distributions in each of |
131 |
the two slabs. Furthermore, it allows effective thermal exchange |
132 |
between particles of different identities, and thus makes the study of |
133 |
interfacial conductance much simpler. |
134 |
|
135 |
The work presented here deals with the Au(111) surface covered to |
136 |
varying degrees by butanethiol, a capping agent with short carbon |
137 |
chain, and solvated with organic solvents of different molecular |
138 |
properties. Different models were used for both the capping agent and |
139 |
the solvent force field parameters. Using the NIVS algorithm, the |
140 |
thermal transport across these interfaces was studied and the |
141 |
underlying mechanism for the phenomena was investigated. |
142 |
|
143 |
\section{Methodology} |
144 |
\subsection{Imposd-Flux Methods in MD Simulations} |
145 |
Steady state MD simulations have an advantage in that not many |
146 |
trajectories are needed to study the relationship between thermal flux |
147 |
and thermal gradients. For systems with low interfacial conductance, |
148 |
one must have a method capable of generating or measuring relatively |
149 |
small fluxes, compared to those required for bulk conductivity. This |
150 |
requirement makes the calculation even more difficult for |
151 |
slowly-converging equilibrium methods.\cite{Viscardy:2007lq} Forward |
152 |
NEMD methods impose a gradient (and measure a flux), but at interfaces |
153 |
it is not clear what behavior should be imposed at the boundaries |
154 |
between materials. Imposed-flux reverse non-equilibrium |
155 |
methods\cite{MullerPlathe:1997xw} set the flux {\it a priori} and |
156 |
the thermal response becomes an easy-to-measure quantity. Although |
157 |
M\"{u}ller-Plathe's original momentum swapping approach can be used |
158 |
for exchanging energy between particles of different identity, the |
159 |
kinetic energy transfer efficiency is affected by the mass difference |
160 |
between the particles, which limits its application on heterogeneous |
161 |
interfacial systems. |
162 |
|
163 |
The non-isotropic velocity scaling (NIVS) \cite{kuang:164101} approach |
164 |
to non-equilibrium MD simulations is able to impose a wide range of |
165 |
kinetic energy fluxes without obvious perturbation to the velocity |
166 |
distributions of the simulated systems. Furthermore, this approach has |
167 |
the advantage in heterogeneous interfaces in that kinetic energy flux |
168 |
can be applied between regions of particles of arbitary identity, and |
169 |
the flux will not be restricted by difference in particle mass. |
170 |
|
171 |
The NIVS algorithm scales the velocity vectors in two separate regions |
172 |
of a simulation system with respective diagonal scaling matricies. To |
173 |
determine these scaling factors in the matricies, a set of equations |
174 |
including linear momentum conservation and kinetic energy conservation |
175 |
constraints and target energy flux satisfaction is solved. With the |
176 |
scaling operation applied to the system in a set frequency, bulk |
177 |
temperature gradients can be easily established, and these can be used |
178 |
for computing thermal conductivities. The NIVS algorithm conserves |
179 |
momenta and energy and does not depend on an external thermostat. |
180 |
|
181 |
\subsection{Defining Interfacial Thermal Conductivity ($G$)} |
182 |
|
183 |
For an interface with relatively low interfacial conductance, and a |
184 |
thermal flux between two distinct bulk regions, the regions on either |
185 |
side of the interface rapidly come to a state in which the two phases |
186 |
have relatively homogeneous (but distinct) temperatures. The |
187 |
interfacial thermal conductivity $G$ can therefore be approximated as: |
188 |
\begin{equation} |
189 |
G = \frac{E_{total}}{2 t L_x L_y \left( \langle T_\mathrm{hot}\rangle - |
190 |
\langle T_\mathrm{cold}\rangle \right)} |
191 |
\label{lowG} |
192 |
\end{equation} |
193 |
where ${E_{total}}$ is the total imposed non-physical kinetic energy |
194 |
transfer during the simulation and ${\langle T_\mathrm{hot}\rangle}$ |
195 |
and ${\langle T_\mathrm{cold}\rangle}$ are the average observed |
196 |
temperature of the two separated phases. For an applied flux $J_z$ |
197 |
operating over a simulation time $t$ on a periodically-replicated slab |
198 |
of dimensions $L_x \times L_y$, $E_{total} = J_z *(t)*(2 L_x L_y)$. |
199 |
|
200 |
When the interfacial conductance is {\it not} small, there are two |
201 |
ways to define $G$. One common way is to assume the temperature is |
202 |
discrete on the two sides of the interface. $G$ can be calculated |
203 |
using the applied thermal flux $J$ and the maximum temperature |
204 |
difference measured along the thermal gradient max($\Delta T$), which |
205 |
occurs at the Gibbs deviding surface (Figure \ref{demoPic}). This is |
206 |
known as the Kapitza conductance, which is the inverse of the Kapitza |
207 |
resistance. |
208 |
\begin{equation} |
209 |
G=\frac{J}{\Delta T} |
210 |
\label{discreteG} |
211 |
\end{equation} |
212 |
|
213 |
\begin{figure} |
214 |
\includegraphics[width=\linewidth]{method} |
215 |
\caption{Interfacial conductance can be calculated by applying an |
216 |
(unphysical) kinetic energy flux between two slabs, one located |
217 |
within the metal and another on the edge of the periodic box. The |
218 |
system responds by forming a thermal response or a gradient. In |
219 |
bulk liquids, this gradient typically has a single slope, but in |
220 |
interfacial systems, there are distinct thermal conductivity |
221 |
domains. The interfacial conductance, $G$ is found by measuring the |
222 |
temperature gap at the Gibbs dividing surface, or by using second |
223 |
derivatives of the thermal profile.} |
224 |
\label{demoPic} |
225 |
\end{figure} |
226 |
|
227 |
The other approach is to assume a continuous temperature profile along |
228 |
the thermal gradient axis (e.g. $z$) and define $G$ at the point where |
229 |
the magnitude of thermal conductivity ($\lambda$) change reaches its |
230 |
maximum, given that $\lambda$ is well-defined throughout the space: |
231 |
\begin{equation} |
232 |
G^\prime = \Big|\frac{\partial\lambda}{\partial z}\Big| |
233 |
= \Big|\frac{\partial}{\partial z}\left(-J_z\Big/ |
234 |
\left(\frac{\partial T}{\partial z}\right)\right)\Big| |
235 |
= |J_z|\Big|\frac{\partial^2 T}{\partial z^2}\Big| |
236 |
\Big/\left(\frac{\partial T}{\partial z}\right)^2 |
237 |
\label{derivativeG} |
238 |
\end{equation} |
239 |
|
240 |
With temperature profiles obtained from simulation, one is able to |
241 |
approximate the first and second derivatives of $T$ with finite |
242 |
difference methods and calculate $G^\prime$. In what follows, both |
243 |
definitions have been used, and are compared in the results. |
244 |
|
245 |
To investigate the interfacial conductivity at metal / solvent |
246 |
interfaces, we have modeled a metal slab with its (111) surfaces |
247 |
perpendicular to the $z$-axis of our simulation cells. The metal slab |
248 |
has been prepared both with and without capping agents on the exposed |
249 |
surface, and has been solvated with simple organic solvents, as |
250 |
illustrated in Figure \ref{gradT}. |
251 |
|
252 |
With the simulation cell described above, we are able to equilibrate |
253 |
the system and impose an unphysical thermal flux between the liquid |
254 |
and the metal phase using the NIVS algorithm. By periodically applying |
255 |
the unphysical flux, we obtained a temperature profile and its spatial |
256 |
derivatives. Figure \ref{gradT} shows how an applied thermal flux can |
257 |
be used to obtain the 1st and 2nd derivatives of the temperature |
258 |
profile. |
259 |
|
260 |
\begin{figure} |
261 |
\includegraphics[width=\linewidth]{gradT} |
262 |
\caption{A sample of Au-butanethiol/hexane interfacial system and the |
263 |
temperature profile after a kinetic energy flux is imposed to |
264 |
it. The 1st and 2nd derivatives of the temperature profile can be |
265 |
obtained with finite difference approximation (lower panel).} |
266 |
\label{gradT} |
267 |
\end{figure} |
268 |
|
269 |
\section{Computational Details} |
270 |
\subsection{Simulation Protocol} |
271 |
The NIVS algorithm has been implemented in our MD simulation code, |
272 |
OpenMD\cite{Meineke:2005gd,openmd}, and was used for our simulations. |
273 |
Metal slabs of 6 or 11 layers of Au atoms were first equilibrated |
274 |
under atmospheric pressure (1 atm) and 200K. After equilibration, |
275 |
butanethiol capping agents were placed at three-fold hollow sites on |
276 |
the Au(111) surfaces. These sites are either {\it fcc} or {\it |
277 |
hcp} sites, although Hase {\it et al.} found that they are |
278 |
equivalent in a heat transfer process,\cite{hase:2010} so we did not |
279 |
distinguish between these sites in our study. The maximum butanethiol |
280 |
capacity on Au surface is $1/3$ of the total number of surface Au |
281 |
atoms, and the packing forms a $(\sqrt{3}\times\sqrt{3})R30^\circ$ |
282 |
structure\cite{doi:10.1021/ja00008a001,doi:10.1021/cr9801317}. A |
283 |
series of lower coverages was also prepared by eliminating |
284 |
butanethiols from the higher coverage surface in a regular manner. The |
285 |
lower coverages were prepared in order to study the relation between |
286 |
coverage and interfacial conductance. |
287 |
|
288 |
The capping agent molecules were allowed to migrate during the |
289 |
simulations. They distributed themselves uniformly and sampled a |
290 |
number of three-fold sites throughout out study. Therefore, the |
291 |
initial configuration does not noticeably affect the sampling of a |
292 |
variety of configurations of the same coverage, and the final |
293 |
conductance measurement would be an average effect of these |
294 |
configurations explored in the simulations. |
295 |
|
296 |
After the modified Au-butanethiol surface systems were equilibrated in |
297 |
the canonical (NVT) ensemble, organic solvent molecules were packed in |
298 |
the previously empty part of the simulation cells.\cite{packmol} Two |
299 |
solvents were investigated, one which has little vibrational overlap |
300 |
with the alkanethiol and which has a planar shape (toluene), and one |
301 |
which has similar vibrational frequencies to the capping agent and |
302 |
chain-like shape ({\it n}-hexane). |
303 |
|
304 |
The simulation cells were not particularly extensive along the |
305 |
$z$-axis, as a very long length scale for the thermal gradient may |
306 |
cause excessively hot or cold temperatures in the middle of the |
307 |
solvent region and lead to undesired phenomena such as solvent boiling |
308 |
or freezing when a thermal flux is applied. Conversely, too few |
309 |
solvent molecules would change the normal behavior of the liquid |
310 |
phase. Therefore, our $N_{solvent}$ values were chosen to ensure that |
311 |
these extreme cases did not happen to our simulations. The spacing |
312 |
between periodic images of the gold interfaces is $45 \sim 75$\AA. |
313 |
|
314 |
The initial configurations generated are further equilibrated with the |
315 |
$x$ and $y$ dimensions fixed, only allowing the $z$-length scale to |
316 |
change. This is to ensure that the equilibration of liquid phase does |
317 |
not affect the metal's crystalline structure. Comparisons were made |
318 |
with simulations that allowed changes of $L_x$ and $L_y$ during NPT |
319 |
equilibration. No substantial changes in the box geometry were noticed |
320 |
in these simulations. After ensuring the liquid phase reaches |
321 |
equilibrium at atmospheric pressure (1 atm), further equilibration was |
322 |
carried out under canonical (NVT) and microcanonical (NVE) ensembles. |
323 |
|
324 |
After the systems reach equilibrium, NIVS was used to impose an |
325 |
unphysical thermal flux between the metal and the liquid phases. Most |
326 |
of our simulations were done under an average temperature of |
327 |
$\sim$200K. Therefore, thermal flux usually came from the metal to the |
328 |
liquid so that the liquid has a higher temperature and would not |
329 |
freeze due to lowered temperatures. After this induced temperature |
330 |
gradient had stablized, the temperature profile of the simulation cell |
331 |
was recorded. To do this, the simulation cell is devided evenly into |
332 |
$N$ slabs along the $z$-axis. The average temperatures of each slab |
333 |
are recorded for 1$\sim$2 ns. When the slab width $d$ of each slab is |
334 |
the same, the derivatives of $T$ with respect to slab number $n$ can |
335 |
be directly used for $G^\prime$ calculations: \begin{equation} |
336 |
G^\prime = |J_z|\Big|\frac{\partial^2 T}{\partial z^2}\Big| |
337 |
\Big/\left(\frac{\partial T}{\partial z}\right)^2 |
338 |
= |J_z|\Big|\frac{1}{d^2}\frac{\partial^2 T}{\partial n^2}\Big| |
339 |
\Big/\left(\frac{1}{d}\frac{\partial T}{\partial n}\right)^2 |
340 |
= |J_z|\Big|\frac{\partial^2 T}{\partial n^2}\Big| |
341 |
\Big/\left(\frac{\partial T}{\partial n}\right)^2 |
342 |
\label{derivativeG2} |
343 |
\end{equation} |
344 |
|
345 |
All of the above simulation procedures use a time step of 1 fs. Each |
346 |
equilibration stage took a minimum of 100 ps, although in some cases, |
347 |
longer equilibration stages were utilized. |
348 |
|
349 |
\subsection{Force Field Parameters} |
350 |
Our simulations include a number of chemically distinct components. |
351 |
Figure \ref{demoMol} demonstrates the sites defined for both |
352 |
United-Atom and All-Atom models of the organic solvent and capping |
353 |
agents in our simulations. Force field parameters are needed for |
354 |
interactions both between the same type of particles and between |
355 |
particles of different species. |
356 |
|
357 |
\begin{figure} |
358 |
\includegraphics[width=\linewidth]{structures} |
359 |
\caption{Structures of the capping agent and solvents utilized in |
360 |
these simulations. The chemically-distinct sites (a-e) are expanded |
361 |
in terms of constituent atoms for both United Atom (UA) and All Atom |
362 |
(AA) force fields. Most parameters are from |
363 |
Refs. \protect\cite{TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes,TraPPE-UA.thiols} |
364 |
(UA) and \protect\cite{OPLSAA} (AA). Cross-interactions with the Au |
365 |
atoms are given in Table \ref{MnM}.} |
366 |
\label{demoMol} |
367 |
\end{figure} |
368 |
|
369 |
The Au-Au interactions in metal lattice slab is described by the |
370 |
quantum Sutton-Chen (QSC) formulation.\cite{PhysRevB.59.3527} The QSC |
371 |
potentials include zero-point quantum corrections and are |
372 |
reparametrized for accurate surface energies compared to the |
373 |
Sutton-Chen potentials.\cite{Chen90} |
374 |
|
375 |
For the two solvent molecules, {\it n}-hexane and toluene, two |
376 |
different atomistic models were utilized. Both solvents were modeled |
377 |
using United-Atom (UA) and All-Atom (AA) models. The TraPPE-UA |
378 |
parameters\cite{TraPPE-UA.alkanes,TraPPE-UA.alkylbenzenes} are used |
379 |
for our UA solvent molecules. In these models, sites are located at |
380 |
the carbon centers for alkyl groups. Bonding interactions, including |
381 |
bond stretches and bends and torsions, were used for intra-molecular |
382 |
sites closer than 3 bonds. For non-bonded interactions, Lennard-Jones |
383 |
potentials are used. |
384 |
|
385 |
By eliminating explicit hydrogen atoms, the TraPPE-UA models are |
386 |
simple and computationally efficient, while maintaining good accuracy. |
387 |
However, the TraPPE-UA model for alkanes is known to predict a slighly |
388 |
lower boiling point than experimental values. This is one of the |
389 |
reasons we used a lower average temperature (200K) for our |
390 |
simulations. If heat is transferred to the liquid phase during the |
391 |
NIVS simulation, the liquid in the hot slab can actually be |
392 |
substantially warmer than the mean temperature in the simulation. The |
393 |
lower mean temperatures therefore prevent solvent boiling. |
394 |
|
395 |
For UA-toluene, the non-bonded potentials between intermolecular sites |
396 |
have a similar Lennard-Jones formulation. The toluene molecules were |
397 |
treated as a single rigid body, so there was no need for |
398 |
intramolecular interactions (including bonds, bends, or torsions) in |
399 |
this solvent model. |
400 |
|
401 |
Besides the TraPPE-UA models, AA models for both organic solvents are |
402 |
included in our studies as well. The OPLS-AA\cite{OPLSAA} force fields |
403 |
were used. For hexane, additional explicit hydrogen sites were |
404 |
included. Besides bonding and non-bonded site-site interactions, |
405 |
partial charges and the electrostatic interactions were added to each |
406 |
CT and HC site. For toluene, a flexible model for the toluene molecule |
407 |
was utilized which included bond, bend, torsion, and inversion |
408 |
potentials to enforce ring planarity. |
409 |
|
410 |
The butanethiol capping agent in our simulations, were also modeled |
411 |
with both UA and AA model. The TraPPE-UA force field includes |
412 |
parameters for thiol molecules\cite{TraPPE-UA.thiols} and are used for |
413 |
UA butanethiol model in our simulations. The OPLS-AA also provides |
414 |
parameters for alkyl thiols. However, alkyl thiols adsorbed on Au(111) |
415 |
surfaces do not have the hydrogen atom bonded to sulfur. To derive |
416 |
suitable parameters for butanethiol adsorbed on Au(111) surfaces, we |
417 |
adopt the S parameters from Luedtke and Landman\cite{landman:1998} and |
418 |
modify the parameters for the CTS atom to maintain charge neutrality |
419 |
in the molecule. Note that the model choice (UA or AA) for the capping |
420 |
agent can be different from the solvent. Regardless of model choice, |
421 |
the force field parameters for interactions between capping agent and |
422 |
solvent can be derived using Lorentz-Berthelot Mixing Rule: |
423 |
\begin{eqnarray} |
424 |
\sigma_{ij} & = & \frac{1}{2} \left(\sigma_{ii} + \sigma_{jj}\right) \\ |
425 |
\epsilon_{ij} & = & \sqrt{\epsilon_{ii}\epsilon_{jj}} |
426 |
\end{eqnarray} |
427 |
|
428 |
To describe the interactions between metal (Au) and non-metal atoms, |
429 |
we refer to an adsorption study of alkyl thiols on gold surfaces by |
430 |
Vlugt {\it et al.}\cite{vlugt:cpc2007154} They fitted an effective |
431 |
Lennard-Jones form of potential parameters for the interaction between |
432 |
Au and pseudo-atoms CH$_x$ and S based on a well-established and |
433 |
widely-used effective potential of Hautman and Klein for the Au(111) |
434 |
surface.\cite{hautman:4994} As our simulations require the gold slab |
435 |
to be flexible to accommodate thermal excitation, the pair-wise form |
436 |
of potentials they developed was used for our study. |
437 |
|
438 |
The potentials developed from {\it ab initio} calculations by Leng |
439 |
{\it et al.}\cite{doi:10.1021/jp034405s} are adopted for the |
440 |
interactions between Au and aromatic C/H atoms in toluene. However, |
441 |
the Lennard-Jones parameters between Au and other types of particles, |
442 |
(e.g. AA alkanes) have not yet been established. For these |
443 |
interactions, the Lorentz-Berthelot mixing rule can be used to derive |
444 |
effective single-atom LJ parameters for the metal using the fit values |
445 |
for toluene. These are then used to construct reasonable mixing |
446 |
parameters for the interactions between the gold and other atoms. |
447 |
Table \ref{MnM} summarizes the ``metal/non-metal'' parameters used in |
448 |
our simulations. |
449 |
|
450 |
\begin{table*} |
451 |
\begin{minipage}{\linewidth} |
452 |
\begin{center} |
453 |
\caption{Non-bonded interaction parameters (including cross |
454 |
interactions with Au atoms) for both force fields used in this |
455 |
work.} |
456 |
\begin{tabular}{lllllll} |
457 |
\hline\hline |
458 |
& Site & $\sigma_{ii}$ & $\epsilon_{ii}$ & $q_i$ & |
459 |
$\sigma_{Au-i}$ & $\epsilon_{Au-i}$ \\ |
460 |
& & (\AA) & (kcal/mol) & ($e$) & (\AA) & (kcal/mol) \\ |
461 |
\hline |
462 |
United Atom (UA) |
463 |
&CH3 & 3.75 & 0.1947 & - & 3.54 & 0.2146 \\ |
464 |
&CH2 & 3.95 & 0.0914 & - & 3.54 & 0.1749 \\ |
465 |
&CHar & 3.695 & 0.1003 & - & 3.4625 & 0.1680 \\ |
466 |
&CRar & 3.88 & 0.04173 & - & 3.555 & 0.1604 \\ |
467 |
\hline |
468 |
All Atom (AA) |
469 |
&CT3 & 3.50 & 0.066 & -0.18 & 3.365 & 0.1373 \\ |
470 |
&CT2 & 3.50 & 0.066 & -0.12 & 3.365 & 0.1373 \\ |
471 |
&CTT & 3.50 & 0.066 & -0.065 & 3.365 & 0.1373 \\ |
472 |
&HC & 2.50 & 0.030 & 0.06 & 2.865 & 0.09256 \\ |
473 |
&CA & 3.55 & 0.070 & -0.115 & 3.173 & 0.0640 \\ |
474 |
&HA & 2.42 & 0.030 & 0.115 & 2.746 & 0.0414 \\ |
475 |
\hline |
476 |
Both UA and AA |
477 |
& S & 4.45 & 0.25 & - & 2.40 & 8.465 \\ |
478 |
\hline\hline |
479 |
\end{tabular} |
480 |
\label{MnM} |
481 |
\end{center} |
482 |
\end{minipage} |
483 |
\end{table*} |
484 |
|
485 |
|
486 |
\section{Results} |
487 |
There are many factors contributing to the measured interfacial |
488 |
conductance; some of these factors are physically motivated |
489 |
(e.g. coverage of the surface by the capping agent coverage and |
490 |
solvent identity), while some are governed by parameters of the |
491 |
methodology (e.g. applied flux and the formulas used to obtain the |
492 |
conductance). In this section we discuss the major physical and |
493 |
calculational effects on the computed conductivity. |
494 |
|
495 |
\subsection{Effects due to capping agent coverage} |
496 |
|
497 |
A series of different initial conditions with a range of surface |
498 |
coverages was prepared and solvated with various with both of the |
499 |
solvent molecules. These systems were then equilibrated and their |
500 |
interfacial thermal conductivity was measured with the NIVS |
501 |
algorithm. Figure \ref{coverage} demonstrates the trend of conductance |
502 |
with respect to surface coverage. |
503 |
|
504 |
\begin{figure} |
505 |
\includegraphics[width=\linewidth]{coverage} |
506 |
\caption{Comparison of interfacial thermal conductivity ($G$) values |
507 |
for the Au-butanethiol/solvent interface with various UA models and |
508 |
different capping agent coverages at $\langle T\rangle\sim$200K.} |
509 |
\label{coverage} |
510 |
\end{figure} |
511 |
|
512 |
In partially covered surfaces, the derivative definition for |
513 |
$G^\prime$ (Eq. \ref{derivativeG}) becomes difficult to apply, as the |
514 |
location of maximum change of $\lambda$ becomes washed out. The |
515 |
discrete definition (Eq. \ref{discreteG}) is easier to apply, as the |
516 |
Gibbs dividing surface is still well-defined. Therefore, $G$ (not |
517 |
$G^\prime$) was used in this section. |
518 |
|
519 |
From Figure \ref{coverage}, one can see the significance of the |
520 |
presence of capping agents. When even a small fraction of the Au(111) |
521 |
surface sites are covered with butanethiols, the conductivity exhibits |
522 |
an enhancement by at least a factor of 3. Cappping agents are clearly |
523 |
playing a major role in thermal transport at metal / organic solvent |
524 |
surfaces. |
525 |
|
526 |
We note a non-monotonic behavior in the interfacial conductance as a |
527 |
function of surface coverage. The maximum conductance (largest $G$) |
528 |
happens when the surfaces are about 75\% covered with butanethiol |
529 |
caps. The reason for this behavior is not entirely clear. One |
530 |
explanation is that incomplete butanethiol coverage allows small gaps |
531 |
between butanethiols to form. These gaps can be filled by transient |
532 |
solvent molecules. These solvent molecules couple very strongly with |
533 |
the hot capping agent molecules near the surface, and can then carry |
534 |
away (diffusively) the excess thermal energy from the surface. |
535 |
|
536 |
There appears to be a competition between the conduction of the |
537 |
thermal energy away from the surface by the capping agents (enhanced |
538 |
by greater coverage) and the coupling of the capping agents with the |
539 |
solvent (enhanced by interdigitation at lower coverages). This |
540 |
competition would lead to the non-monotonic coverage behavior observed |
541 |
here. |
542 |
|
543 |
Results for rigid body toluene solvent, as well as the UA hexane, are |
544 |
within the ranges expected from prior experimental |
545 |
work.\cite{Wilson:2002uq,cahill:793,PhysRevB.80.195406} This suggests |
546 |
that explicit hydrogen atoms might not be required for modeling |
547 |
thermal transport in these systems. C-H vibrational modes do not see |
548 |
significant excited state population at low temperatures, and are not |
549 |
likely to carry lower frequency excitations from the solid layer into |
550 |
the bulk liquid. |
551 |
|
552 |
The toluene solvent does not exhibit the same behavior as hexane in |
553 |
that $G$ remains at approximately the same magnitude when the capping |
554 |
coverage increases from 25\% to 75\%. Toluene, as a rigid planar |
555 |
molecule, cannot occupy the relatively small gaps between the capping |
556 |
agents as easily as the chain-like {\it n}-hexane. The effect of |
557 |
solvent coupling to the capping agent is therefore weaker in toluene |
558 |
except at the very lowest coverage levels. This effect counters the |
559 |
coverage-dependent conduction of heat away from the metal surface, |
560 |
leading to a much flatter $G$ vs. coverage trend than is observed in |
561 |
{\it n}-hexane. |
562 |
|
563 |
\subsection{Effects due to Solvent \& Solvent Models} |
564 |
In addition to UA solvent and capping agent models, AA models have |
565 |
also been included in our simulations. In most of this work, the same |
566 |
(UA or AA) model for solvent and capping agent was used, but it is |
567 |
also possible to utilize different models for different components. |
568 |
We have also included isotopic substitutions (Hydrogen to Deuterium) |
569 |
to decrease the explicit vibrational overlap between solvent and |
570 |
capping agent. Table \ref{modelTest} summarizes the results of these |
571 |
studies. |
572 |
|
573 |
\begin{table*} |
574 |
\begin{minipage}{\linewidth} |
575 |
\begin{center} |
576 |
|
577 |
\caption{Computed interfacial thermal conductance ($G$ and |
578 |
$G^\prime$) values for interfaces using various models for |
579 |
solvent and capping agent (or without capping agent) at |
580 |
$\langle T\rangle\sim$200K. (D stands for deuterated solvent |
581 |
or capping agent molecules; ``Avg.'' denotes results that are |
582 |
averages of simulations under different applied thermal flux |
583 |
values $(J_z)$. Error estimates are indicated in |
584 |
parentheses.)} |
585 |
|
586 |
\begin{tabular}{llccc} |
587 |
\hline\hline |
588 |
Butanethiol model & Solvent & $J_z$ & $G$ & $G^\prime$ \\ |
589 |
(or bare surface) & model & (GW/m$^2$) & |
590 |
\multicolumn{2}{c}{(MW/m$^2$/K)} \\ |
591 |
\hline |
592 |
UA & UA hexane & Avg. & 131(9) & 87(10) \\ |
593 |
& UA hexane(D) & 1.95 & 153(5) & 136(13) \\ |
594 |
& AA hexane & Avg. & 131(6) & 122(10) \\ |
595 |
& UA toluene & 1.96 & 187(16) & 151(11) \\ |
596 |
& AA toluene & 1.89 & 200(36) & 149(53) \\ |
597 |
\hline |
598 |
AA & UA hexane & 1.94 & 116(9) & 129(8) \\ |
599 |
& AA hexane & Avg. & 442(14) & 356(31) \\ |
600 |
& AA hexane(D) & 1.93 & 222(12) & 234(54) \\ |
601 |
& UA toluene & 1.98 & 125(25) & 97(60) \\ |
602 |
& AA toluene & 3.79 & 487(56) & 290(42) \\ |
603 |
\hline |
604 |
AA(D) & UA hexane & 1.94 & 158(25) & 172(4) \\ |
605 |
& AA hexane & 1.92 & 243(29) & 191(11) \\ |
606 |
& AA toluene & 1.93 & 364(36) & 322(67) \\ |
607 |
\hline |
608 |
bare & UA hexane & Avg. & 46.5(3.2) & 49.4(4.5) \\ |
609 |
& UA hexane(D) & 0.98 & 43.9(4.6) & 43.0(2.0) \\ |
610 |
& AA hexane & 0.96 & 31.0(1.4) & 29.4(1.3) \\ |
611 |
& UA toluene & 1.99 & 70.1(1.3) & 65.8(0.5) \\ |
612 |
\hline\hline |
613 |
\end{tabular} |
614 |
\label{modelTest} |
615 |
\end{center} |
616 |
\end{minipage} |
617 |
\end{table*} |
618 |
|
619 |
To facilitate direct comparison between force fields, systems with the |
620 |
same capping agent and solvent were prepared with the same length |
621 |
scales for the simulation cells. |
622 |
|
623 |
On bare metal / solvent surfaces, different force field models for |
624 |
hexane yield similar results for both $G$ and $G^\prime$, and these |
625 |
two definitions agree with each other very well. This is primarily an |
626 |
indicator of weak interactions between the metal and the solvent, and |
627 |
is a typical case for acoustic impedance mismatch between these two |
628 |
phases. |
629 |
|
630 |
For the fully-covered surfaces, the choice of force field for the |
631 |
capping agent and solvent has a large impact on the calulated values |
632 |
of $G$ and $G^\prime$. The AA thiol to AA solvent conductivities are |
633 |
much larger than their UA to UA counterparts, and these values exceed |
634 |
the experimental estimates by a large measure. The AA force field |
635 |
allows significant energy to go into C-H (or C-D) stretching modes, |
636 |
and since these modes are high frequency, this non-quantum behavior is |
637 |
likely responsible for the overestimate of the conductivity. Compared |
638 |
to the AA model, the UA model yields more reasonable conductivity |
639 |
values with much higher computational efficiency. |
640 |
|
641 |
\subsubsection{Are electronic excitations in the metal important?} |
642 |
Because they lack electronic excitations, the QSC and related embedded |
643 |
atom method (EAM) models for gold are known to predict unreasonably |
644 |
low values for bulk conductivity |
645 |
($\lambda$).\cite{kuang:164101,ISI:000207079300006,Clancy:1992} If the |
646 |
conductance between the phases ($G$) is governed primarily by phonon |
647 |
excitation (and not electronic degrees of freedom), one would expect a |
648 |
classical model to capture most of the interfacial thermal |
649 |
conductance. Our results for $G$ and $G^\prime$ indicate that this is |
650 |
indeed the case, and suggest that the modeling of interfacial thermal |
651 |
transport depends primarily on the description of the interactions |
652 |
between the various components at the interface. When the metal is |
653 |
chemically capped, the primary barrier to thermal conductivity appears |
654 |
to be the interface between the capping agent and the surrounding |
655 |
solvent, so the excitations in the metal have little impact on the |
656 |
value of $G$. |
657 |
|
658 |
\subsection{Effects due to methodology and simulation parameters} |
659 |
|
660 |
We have varied the parameters of the simulations in order to |
661 |
investigate how these factors would affect the computation of $G$. Of |
662 |
particular interest are: 1) the length scale for the applied thermal |
663 |
gradient (modified by increasing the amount of solvent in the system), |
664 |
2) the sign and magnitude of the applied thermal flux, 3) the average |
665 |
temperature of the simulation (which alters the solvent density during |
666 |
equilibration), and 4) the definition of the interfacial conductance |
667 |
(Eqs. (\ref{discreteG}) or (\ref{derivativeG})) used in the |
668 |
calculation. |
669 |
|
670 |
Systems of different lengths were prepared by altering the number of |
671 |
solvent molecules and extending the length of the box along the $z$ |
672 |
axis to accomodate the extra solvent. Equilibration at the same |
673 |
temperature and pressure conditions led to nearly identical surface |
674 |
areas ($L_x$ and $L_y$) available to the metal and capping agent, |
675 |
while the extra solvent served mainly to lengthen the axis that was |
676 |
used to apply the thermal flux. For a given value of the applied |
677 |
flux, the different $z$ length scale has only a weak effect on the |
678 |
computed conductivities (Table \ref{AuThiolHexaneUA}). |
679 |
|
680 |
\subsubsection{Effects of applied flux} |
681 |
The NIVS algorithm allows changes in both the sign and magnitude of |
682 |
the applied flux. It is possible to reverse the direction of heat |
683 |
flow simply by changing the sign of the flux, and thermal gradients |
684 |
which would be difficult to obtain experimentally ($5$ K/\AA) can be |
685 |
easily simulated. However, the magnitude of the applied flux is not |
686 |
arbitary if one aims to obtain a stable and reliable thermal gradient. |
687 |
A temperature gradient can be lost in the noise if $|J_z|$ is too |
688 |
small, and excessive $|J_z|$ values can cause phase transitions if the |
689 |
extremes of the simulation cell become widely separated in |
690 |
temperature. Also, if $|J_z|$ is too large for the bulk conductivity |
691 |
of the materials, the thermal gradient will never reach a stable |
692 |
state. |
693 |
|
694 |
Within a reasonable range of $J_z$ values, we were able to study how |
695 |
$G$ changes as a function of this flux. In what follows, we use |
696 |
positive $J_z$ values to denote the case where energy is being |
697 |
transferred by the method from the metal phase and into the liquid. |
698 |
The resulting gradient therefore has a higher temperature in the |
699 |
liquid phase. Negative flux values reverse this transfer, and result |
700 |
in higher temperature metal phases. The conductance measured under |
701 |
different applied $J_z$ values is listed in Tables |
702 |
\ref{AuThiolHexaneUA} and \ref{AuThiolToluene}. These results do not |
703 |
indicate that $G$ depends strongly on $J_z$ within this flux |
704 |
range. The linear response of flux to thermal gradient simplifies our |
705 |
investigations in that we can rely on $G$ measurement with only a |
706 |
small number $J_z$ values. |
707 |
|
708 |
\begin{table*} |
709 |
\begin{minipage}{\linewidth} |
710 |
\begin{center} |
711 |
\caption{Computed interfacial thermal conductivity ($G$ and |
712 |
$G^\prime$) values for the 100\% covered Au-butanethiol/hexane |
713 |
interfaces with UA model and different hexane molecule numbers |
714 |
at different temperatures using a range of energy |
715 |
fluxes. Error estimates indicated in parenthesis.} |
716 |
|
717 |
\begin{tabular}{ccccccc} |
718 |
\hline\hline |
719 |
$\langle T\rangle$ & $N_{hexane}$ & $\rho_{hexane}$ & |
720 |
$J_z$ & $G$ & $G^\prime$ \\ |
721 |
(K) & & (g/cm$^3$) & (GW/m$^2$) & |
722 |
\multicolumn{2}{c}{(MW/m$^2$/K)} \\ |
723 |
\hline |
724 |
200 & 266 & 0.672 & -0.96 & 102(3) & 80.0(0.8) \\ |
725 |
& 200 & 0.688 & 0.96 & 125(16) & 90.2(15) \\ |
726 |
& & & 1.91 & 139(10) & 101(10) \\ |
727 |
& & & 2.83 & 141(6) & 89.9(9.8) \\ |
728 |
& 166 & 0.681 & 0.97 & 141(30) & 78(22) \\ |
729 |
& & & 1.92 & 138(4) & 98.9(9.5) \\ |
730 |
\hline |
731 |
250 & 200 & 0.560 & 0.96 & 75(10) & 61.8(7.3) \\ |
732 |
& & & -0.95 & 49.4(0.3) & 45.7(2.1) \\ |
733 |
& 166 & 0.569 & 0.97 & 80.3(0.6) & 67(11) \\ |
734 |
& & & 1.44 & 76.2(5.0) & 64.8(3.8) \\ |
735 |
& & & -0.95 & 56.4(2.5) & 54.4(1.1) \\ |
736 |
& & & -1.85 & 47.8(1.1) & 53.5(1.5) \\ |
737 |
\hline\hline |
738 |
\end{tabular} |
739 |
\label{AuThiolHexaneUA} |
740 |
\end{center} |
741 |
\end{minipage} |
742 |
\end{table*} |
743 |
|
744 |
The sign of $J_z$ is a different matter, however, as this can alter |
745 |
the temperature on the two sides of the interface. The average |
746 |
temperature values reported are for the entire system, and not for the |
747 |
liquid phase, so at a given $\langle T \rangle$, the system with |
748 |
positive $J_z$ has a warmer liquid phase. This means that if the |
749 |
liquid carries thermal energy via convective transport, {\it positive} |
750 |
$J_z$ values will result in increased molecular motion on the liquid |
751 |
side of the interface, and this will increase the measured |
752 |
conductivity. |
753 |
|
754 |
\subsubsection{Effects due to average temperature} |
755 |
|
756 |
We also studied the effect of average system temperature on the |
757 |
interfacial conductance. The simulations are first equilibrated in |
758 |
the NPT ensemble at 1 atm. The TraPPE-UA model for hexane tends to |
759 |
predict a lower boiling point (and liquid state density) than |
760 |
experiments. This lower-density liquid phase leads to reduced contact |
761 |
between the hexane and butanethiol, and this accounts for our |
762 |
observation of lower conductance at higher temperatures as shown in |
763 |
Table \ref{AuThiolHexaneUA}. In raising the average temperature from |
764 |
200K to 250K, the density drop of ~20\% in the solvent phase leads to |
765 |
a ~65\% drop in the conductance. |
766 |
|
767 |
Similar behavior is observed in the TraPPE-UA model for toluene, |
768 |
although this model has better agreement with the experimental |
769 |
densities of toluene. The expansion of the toluene liquid phase is |
770 |
not as significant as that of the hexane (8.3\% over 100K), and this |
771 |
limits the effect to ~20\% drop in thermal conductivity (Table |
772 |
\ref{AuThiolToluene}). |
773 |
|
774 |
Although we have not mapped out the behavior at a large number of |
775 |
temperatures, is clear that there will be a strong temperature |
776 |
dependence in the interfacial conductance when the physical properties |
777 |
of one side of the interface (notably the density) change rapidly as a |
778 |
function of temperature. |
779 |
|
780 |
\begin{table*} |
781 |
\begin{minipage}{\linewidth} |
782 |
\begin{center} |
783 |
\caption{Computed interfacial thermal conductivity ($G$ and |
784 |
$G^\prime$) values for a 90\% coverage Au-butanethiol/toluene |
785 |
interface at different temperatures using a range of energy |
786 |
fluxes. Error estimates indicated in parenthesis.} |
787 |
|
788 |
\begin{tabular}{ccccc} |
789 |
\hline\hline |
790 |
$\langle T\rangle$ & $\rho_{toluene}$ & $J_z$ & $G$ & $G^\prime$ \\ |
791 |
(K) & (g/cm$^3$) & (GW/m$^2$) & \multicolumn{2}{c}{(MW/m$^2$/K)} \\ |
792 |
\hline |
793 |
200 & 0.933 & 2.15 & 204(12) & 113(12) \\ |
794 |
& & -1.86 & 180(3) & 135(21) \\ |
795 |
& & -3.93 & 176(5) & 113(12) \\ |
796 |
\hline |
797 |
300 & 0.855 & -1.91 & 143(5) & 125(2) \\ |
798 |
& & -4.19 & 135(9) & 113(12) \\ |
799 |
\hline\hline |
800 |
\end{tabular} |
801 |
\label{AuThiolToluene} |
802 |
\end{center} |
803 |
\end{minipage} |
804 |
\end{table*} |
805 |
|
806 |
Besides the lower interfacial thermal conductance, surfaces at |
807 |
relatively high temperatures are susceptible to reconstructions, |
808 |
particularly when butanethiols fully cover the Au(111) surface. These |
809 |
reconstructions include surface Au atoms which migrate outward to the |
810 |
S atom layer, and butanethiol molecules which embed into the surface |
811 |
Au layer. The driving force for this behavior is the strong Au-S |
812 |
interactions which are modeled here with a deep Lennard-Jones |
813 |
potential. This phenomenon agrees with reconstructions that have beeen |
814 |
experimentally |
815 |
observed.\cite{doi:10.1021/j100035a033,doi:10.1021/la026493y}. Vlugt |
816 |
{\it et al.} kept their Au(111) slab rigid so that their simulations |
817 |
could reach 300K without surface |
818 |
reconstructions.\cite{vlugt:cpc2007154} Since surface reconstructions |
819 |
blur the interface, the measurement of $G$ becomes more difficult to |
820 |
conduct at higher temperatures. For this reason, most of our |
821 |
measurements are undertaken at $\langle T\rangle\sim$200K where |
822 |
reconstruction is minimized. |
823 |
|
824 |
However, when the surface is not completely covered by butanethiols, |
825 |
the simulated system appears to be more resistent to the |
826 |
reconstruction. O ur Au / butanethiol / toluene system had the Au(111) |
827 |
surfaces 90\% covered by butanethiols, but did not see this above |
828 |
phenomena even at $\langle T\rangle\sim$300K. That said, we did |
829 |
observe butanethiols migrating to neighboring three-fold sites during |
830 |
a simulation. Since the interface persisted in these simulations, |
831 |
were able to obtain $G$'s for these interfaces even at a relatively |
832 |
high temperature without being affected by surface reconstructions. |
833 |
|
834 |
\section{Discussion} |
835 |
|
836 |
The primary result of this work is that the capping agent acts as an |
837 |
efficient thermal coupler between solid and solvent phases. One of |
838 |
the ways the capping agent can carry out this role is to down-shift |
839 |
between the phonon vibrations in the solid (which carry the heat from |
840 |
the gold) and the molecular vibrations in the liquid (which carry some |
841 |
of the heat in the solvent). |
842 |
|
843 |
To investigate the mechanism of interfacial thermal conductance, the |
844 |
vibrational power spectrum was computed. Power spectra were taken for |
845 |
individual components in different simulations. To obtain these |
846 |
spectra, simulations were run after equilibration in the |
847 |
microcanonical (NVE) ensemble and without a thermal |
848 |
gradient. Snapshots of configurations were collected at a frequency |
849 |
that is higher than that of the fastest vibrations occuring in the |
850 |
simulations. With these configurations, the velocity auto-correlation |
851 |
functions can be computed: |
852 |
\begin{equation} |
853 |
C_A (t) = \langle\vec{v}_A (t)\cdot\vec{v}_A (0)\rangle |
854 |
\label{vCorr} |
855 |
\end{equation} |
856 |
The power spectrum is constructed via a Fourier transform of the |
857 |
symmetrized velocity autocorrelation function, |
858 |
\begin{equation} |
859 |
\hat{f}(\omega) = |
860 |
\int_{-\infty}^{\infty} C_A (t) e^{-2\pi it\omega}\,dt |
861 |
\label{fourier} |
862 |
\end{equation} |
863 |
|
864 |
\subsection{The role of specific vibrations} |
865 |
The vibrational spectra for gold slabs in different environments are |
866 |
shown as in Figure \ref{specAu}. Regardless of the presence of |
867 |
solvent, the gold surfaces which are covered by butanethiol molecules |
868 |
exhibit an additional peak observed at a frequency of |
869 |
$\sim$165cm$^{-1}$. We attribute this peak to the S-Au bonding |
870 |
vibration. This vibration enables efficient thermal coupling of the |
871 |
surface Au layer to the capping agents. Therefore, in our simulations, |
872 |
the Au / S interfaces do not appear to be the primary barrier to |
873 |
thermal transport when compared with the butanethiol / solvent |
874 |
interfaces. |
875 |
|
876 |
\begin{figure} |
877 |
\includegraphics[width=\linewidth]{vibration} |
878 |
\caption{Vibrational power spectra for gold in different solvent |
879 |
environments. The presence of the butanethiol capping molecules |
880 |
adds a vibrational peak at $\sim$165cm$^{-1}$. The butanethiol |
881 |
spectra exhibit a corresponding peak.} |
882 |
\label{specAu} |
883 |
\end{figure} |
884 |
|
885 |
Also in this figure, we show the vibrational power spectrum for the |
886 |
bound butanethiol molecules, which also exhibits the same |
887 |
$\sim$165cm$^{-1}$ peak. |
888 |
|
889 |
\subsection{Overlap of power spectra} |
890 |
A comparison of the results obtained from the two different organic |
891 |
solvents can also provide useful information of the interfacial |
892 |
thermal transport process. In particular, the vibrational overlap |
893 |
between the butanethiol and the organic solvents suggests a highly |
894 |
efficient thermal exchange between these components. Very high |
895 |
thermal conductivity was observed when AA models were used and C-H |
896 |
vibrations were treated classically. The presence of extra degrees of |
897 |
freedom in the AA force field yields higher heat exchange rates |
898 |
between the two phases and results in a much higher conductivity than |
899 |
in the UA force field. |
900 |
|
901 |
The similarity in the vibrational modes available to solvent and |
902 |
capping agent can be reduced by deuterating one of the two components |
903 |
(Fig. \ref{aahxntln}). Once either the hexanes or the butanethiols |
904 |
are deuterated, one can observe a significantly lower $G$ and |
905 |
$G^\prime$ values (Table \ref{modelTest}). |
906 |
|
907 |
\begin{figure} |
908 |
\includegraphics[width=\linewidth]{aahxntln} |
909 |
\caption{Spectra obtained for all-atom (AA) Au / butanethiol / solvent |
910 |
systems. When butanethiol is deuterated (lower left), its |
911 |
vibrational overlap with hexane decreases significantly. Since |
912 |
aromatic molecules and the butanethiol are vibrationally dissimilar, |
913 |
the change is not as dramatic when toluene is the solvent (right).} |
914 |
\label{aahxntln} |
915 |
\end{figure} |
916 |
|
917 |
For the Au / butanethiol / toluene interfaces, having the AA |
918 |
butanethiol deuterated did not yield a significant change in the |
919 |
measured conductance. Compared to the C-H vibrational overlap between |
920 |
hexane and butanethiol, both of which have alkyl chains, the overlap |
921 |
between toluene and butanethiol is not as significant and thus does |
922 |
not contribute as much to the heat exchange process. |
923 |
|
924 |
Previous observations of Zhang {\it et al.}\cite{hase:2010} indicate |
925 |
that the {\it intra}molecular heat transport due to alkylthiols is |
926 |
highly efficient. Combining our observations with those of Zhang {\it |
927 |
et al.}, it appears that butanethiol acts as a channel to expedite |
928 |
heat flow from the gold surface and into the alkyl chain. The |
929 |
acoustic impedance mismatch between the metal and the liquid phase can |
930 |
therefore be effectively reduced with the presence of suitable capping |
931 |
agents. |
932 |
|
933 |
Deuterated models in the UA force field did not decouple the thermal |
934 |
transport as well as in the AA force field. The UA models, even |
935 |
though they have eliminated the high frequency C-H vibrational |
936 |
overlap, still have significant overlap in the lower-frequency |
937 |
portions of the infrared spectra (Figure \ref{uahxnua}). Deuterating |
938 |
the UA models did not decouple the low frequency region enough to |
939 |
produce an observable difference for the results of $G$ (Table |
940 |
\ref{modelTest}). |
941 |
|
942 |
\begin{figure} |
943 |
\includegraphics[width=\linewidth]{uahxnua} |
944 |
\caption{Vibrational spectra obtained for normal (upper) and |
945 |
deuterated (lower) hexane in Au-butanethiol/hexane |
946 |
systems. Butanethiol spectra are shown as reference. Both hexane and |
947 |
butanethiol were using United-Atom models.} |
948 |
\label{uahxnua} |
949 |
\end{figure} |
950 |
|
951 |
\section{Conclusions} |
952 |
The NIVS algorithm has been applied to simulations of |
953 |
butanethiol-capped Au(111) surfaces in the presence of organic |
954 |
solvents. This algorithm allows the application of unphysical thermal |
955 |
flux to transfer heat between the metal and the liquid phase. With the |
956 |
flux applied, we were able to measure the corresponding thermal |
957 |
gradients and to obtain interfacial thermal conductivities. Under |
958 |
steady states, 2-3 ns trajectory simulations are sufficient for |
959 |
computation of this quantity. |
960 |
|
961 |
Our simulations have seen significant conductance enhancement in the |
962 |
presence of capping agent, compared with the bare gold / liquid |
963 |
interfaces. The acoustic impedance mismatch between the metal and the |
964 |
liquid phase is effectively eliminated by a chemically-bonded capping |
965 |
agent. Furthermore, the coverage precentage of the capping agent plays |
966 |
an important role in the interfacial thermal transport |
967 |
process. Moderately low coverages allow higher contact between capping |
968 |
agent and solvent, and thus could further enhance the heat transfer |
969 |
process, giving a non-monotonic behavior of conductance with |
970 |
increasing coverage. |
971 |
|
972 |
Our results, particularly using the UA models, agree well with |
973 |
available experimental data. The AA models tend to overestimate the |
974 |
interfacial thermal conductance in that the classically treated C-H |
975 |
vibrations become too easily populated. Compared to the AA models, the |
976 |
UA models have higher computational efficiency with satisfactory |
977 |
accuracy, and thus are preferable in modeling interfacial thermal |
978 |
transport. |
979 |
|
980 |
Of the two definitions for $G$, the discrete form |
981 |
(Eq. \ref{discreteG}) was easier to use and gives out relatively |
982 |
consistent results, while the derivative form (Eq. \ref{derivativeG}) |
983 |
is not as versatile. Although $G^\prime$ gives out comparable results |
984 |
and follows similar trend with $G$ when measuring close to fully |
985 |
covered or bare surfaces, the spatial resolution of $T$ profile |
986 |
required for the use of a derivative form is limited by the number of |
987 |
bins and the sampling required to obtain thermal gradient information. |
988 |
|
989 |
Vlugt {\it et al.} have investigated the surface thiol structures for |
990 |
nanocrystalline gold and pointed out that they differ from those of |
991 |
the Au(111) surface.\cite{landman:1998,vlugt:cpc2007154} This |
992 |
difference could also cause differences in the interfacial thermal |
993 |
transport behavior. To investigate this problem, one would need an |
994 |
effective method for applying thermal gradients in non-planar |
995 |
(i.e. spherical) geometries. |
996 |
|
997 |
\section{Acknowledgments} |
998 |
Support for this project was provided by the National Science |
999 |
Foundation under grant CHE-0848243. Computational time was provided by |
1000 |
the Center for Research Computing (CRC) at the University of Notre |
1001 |
Dame. |
1002 |
\newpage |
1003 |
|
1004 |
\bibliography{interfacial} |
1005 |
|
1006 |
\end{doublespace} |
1007 |
\end{document} |
1008 |
|