Thin-thick surface phase coexistence and boundary tension of the square-

well fluid on a weak attractive surface

Jayant K. Singh, Gautam Sarma, and Sang K. Kwak

Citation: J. Chem. Phys. 128, 044708 (2008); doi: 10.1063/1.2824503
View online: http://dx.doi.org/10.1063/1.2824503
View Table of Contents: http://jcp.aip.org/resource/1/JCPSA6/v128/i4
Published by the AIP Publishing LLC.

Additional information on J. Chem. Phys.
Journal Homepage: http://jcp.aip.org/
Journal Information: http://jcp.aip.org/about/about_the_journal
Top downloads: http://jcp.aip.org/features/most_downloaded
Information for Authors: http://jcp.aip.org/authors
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
Thin-thick surface phase coexistence and boundary tension
of the square-well uid on a weak attractive surface
Jayant K. Singh
a
and Gautam Sarma
Department of Chemical Engineering, Indian Institute of Technology, Kanpur 208016, India
Sang K. Kwak
Division of Chemical and Biomolecular Engineering, School of Chemical and Biomedical Engineering,
Nanyang Technological University, Singapore 637459, Singapore
Received 27 August 2007; accepted 21 November 2007; published online 28 January 2008
Prewetting transition is studied for the square-well uid of attractive-well diameter
f f

f f
=1.5 in the
presence of a homogeneous surface modeled by the square-well potential of attractive well from
0.8
f f
to 1.8
f f
. We investigate surface phase coexistence of thin-thick lm transition using
grand-canonical transition matrix Monte Carlo GC-TMMC and histogram reweighting techniques.
Molecular dynamics MD and GC-TMMC are utilized to predict the properties of the uid for
various surface uid afnities. Occurrences of prewetting transition with the variation of surface
afnity are observed for a domain of reduced temperature from T
*
=0.62 to 0.75. We have used MD
and GC-TMMC+finite size scaling FSS simulations to calculate the boundary tension as a
function of temperature as well as surface afnity. Boundary tensions via MD and GC-TMMC
+FSS methods are in good agreement. The boundary tension increases with the decrease of
wall-uid afnity. Prewetting critical properties are calculated using rectilinear diameter approach
and scaling analysis. We found that critical temperature and density increase with the decrease of
wall-uid afnity. 2008 American Institute of Physics. DOI: 10.1063/1.2824503
I. INTRODUCTION
Morphological transition in the presence of surfaces
plays an important role in various industrial applications.
Relative strengths of surface-uid and uid-uid interactions
can lead to various phase transitions such as prewetting, lay-
ering, and capillary condensation.
1
Coexistence phases of va-
por and liquid in contact with an attractive solid surface can
induce two different kinds of phenomena. For strong attrac-
tive surface the uid molecules spread across the surface in
the form of thick lm, leading to complete wetting of the
surface. On the other hand, a nonwetting behavior is exhib-
ited, if the uid molecules are weakly adsorbed on the sur-
face, so that the interactions between the uid molecules
becomes stronger than those between the wall and the uid
molecules.
2
In such case partial wetting occurs in the ad-
sorbed lm, which consists of a layer of vapor bubbles and
liquid drops. Wetting transition is closely associated with a
temperature called wetting temperature T
w
at which adsorp-
tion state transforms from partial wetting to complete wet-
ting. Below the wetting temperature the thickness of the lm
adsorbed on the surface remains nite at all pressures, which
are below the bulk saturation pressure. Above wetting tem-
perature, prewetting transition occurs when there is a rst
order transition between the thin and thick lms adsorbed on
the surface. This prewetting transition terminates at prewet-
ting critical point T
cpw
, where thin and thick lms surface
states become indistinguishable. In 1977, Cahn
2
predicted
the existence of wetting transition through a two-phase mix-
ture of uids near a third phase, surface. Independently, Eb-
nar and Saam
3
also predicted wetting and prewetting transi-
tion of argon lm adsorbed onto a weakly attractive solid
carbon dioxide surface. The authors used the density func-
tional theory DFT for the prediction of wetting transition.
Since then, DFT has been widely used to study wetting
phenomena.
4
Experimental evidences, which came much
later, support the prediction of prewetting transitions. Ex-
amples can be found for helium adsorption on Cs Refs. 5
and 6 and Rb,
7
liquid hydrogen on various substrates,
8
and
acetone on graphite.
9
Phase transition in the presence of surfaces has been
studied using various molecular simulation techniques in ad-
dition to theory and experiments. Finn and Monson
10
were
rst to study the prewetting transition of argon molecules on
solid carbon dioxide system using isothermal-isobaric Monte
Carlo MC method. Subsequently, few more groups have
utilized different molecular simulation methodologies
1114
to
understand the prewetting transition. Phase equilibria calcu-
lations of the thin-thick lm have become much easier with
the advent of simulation techniques such as transitional ma-
trix Monte Carlo TMMC,
15
multicanonical sampling,
16
and
histogram reweighting.
17
Errington recently revisited the
prewetting transitions of model argon on solid carbon diox-
ide and successfully demonstrated the use of grand-canonical
transitional matrix Monte Carlo GC-TMMC to predict the
wetting temperature
18
and boundary tension.
19
In this work,
we investigate prewetting transitions of a model system
based on square-well potential and demonstrate the equiva-
lence of MC and molecular dynamics MD methods in the
calculation of boundary tension. Also, GC-TMMC method is
a
Author to whom correspondence should be addressed. Electronic mail:
jayantks@iitk.ac.in.
THE JOURNAL OF CHEMICAL PHYSICS 128, 044708 2008
0021-9606/2008/1284/044708/8/$23.00 2008 American Institute of Physics 128, 044708-1
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
used to study the rst order wetting behavior of a square-
well uid in the extreme range of potential parameters. In
prediction of boundary tension of the thin-thick lm, the
concept of nite size scaling FSS method has attracted less
attention and hence forms a secondary object of our study.
We utilize MD and GC-TMMC simulations to explore the
thermodynamics of thin-thick lm phases on a surface and
present comparison of the results via the above two methods.
The paper is organized as follows. Section II describes the
models used in this study as well as our simulation methods.
Section III presents the results of phase coexistence of thin-
thick lms with the variation of the strength of substrate
potential and the details of boundary tension calculation us-
ing MD as well as GC-TMMC simulations. Section IV con-
cludes our study.
II. MODEL AND METHODOLOGY
One of the simple pair potentials containing both the
attractive and repulsive forces is square-well SW potential.
Owing to its simplicity, this model has been widely used to
study phenomena related to phase coexistence and interfacial
properties of simple uid.
2027
Henderson and Swol
28
used
SW model to understand wetting transition. The SW poten-
tial is also used as a model for the colloidal
2931
and globular
protein solutions.
32,33
In this work uid-uid interaction is
modeled as SW potential, which is represented as
U
f f
r = if r
f f
=
f f
if
f f
r
f f

f f
=0 if r
f f

f f
1
where
f f
,
f f
, and
f f
are hard sphere diameter, potential
well depth, and well diameter of uid-uid potential, respec-
tively. Fluid-wall interaction is also modeled by the follow-
ing SW-type potential:
U
wf
z = if z 0.8
f f
=
wf
if 0.8
f f
z
wf

f f
=0 if z
wf

f f
, 2
where
wf
and
wf
are corresponding parameters for uid-
substrate potential. In this work, units are adopted such that

f f
and
f f
are unity and
f f
=1.5 and
wf
=1.8. Wall-uid
interaction
wf
is varied from 2 to 5 in units of
f f
. All
lengths are in units of core diameter
f f
.
In grand-canonical ensemble, for a xed temperature T,
chemical potential , and volume V, the probability of vis-
iting a microstate s with particles Ns and conguration en-
ergy Es probability is given by the following expression:
s =
1

V
Ns

3Ns
Ns!

exp EsexpNs, 3
where is grand-canonical partition function, is de Bro-
glie wavelength, =1/ k
B
T, and k
B
is Boltzmanns constant.
We adopted the methodology as shown in Ref. 18 to get
the macrostate probability from the probability transition ma-
trix. We employed the possible state changes by deletion,
addition, and displacement of MC moves. For the moves S
S, SS1, and SS+1, the detailed balance can be
simplied to give the macrostate probability as
ln S + 1 = ln s ln

PS + 1 S
PS S + 1

. 4
To enhance the probability of transition from one phase to
another phase, we utilize multicanonical sampling
16
with a
weighting function S =ln S. We modify the accep-
tance criteria in the presence of the bias as follows:
as t = min

1,
expTt
expSs

. 5
Histogram reweighting techniques
17
is used to obtain the
chemical potential that gives the phase coexistence of the
thin-thick lm.
We use Binder formalism
34
to calculate innite size
boundary tension, between thin and thick lms, from a series
of nite size calculations. If
L
is interfacial tension for a
system of box length L, then the boundary tension for
innite system is given by

L
=
F
L
2L
= C
1
1
L
+ C
2
lnL
L
+ , 6
where c
1
and c
2
are constants and F
L
represents the free
energy of the thin-thick interface for a nite system size L.
Helmholtz potential, for the thin-thick lm on a substrate
see Fig. 1, for the innitesimal and reversible transforma-
tion is given by
dF = PHdA PAdH SdT + dN + dL
y
, 7
where F is the Helmholtz free energy, A is the substrate area,
H is the distance between the upper wall and the substrate, S
is the entropy, is the boundary tension between thin and
thick lms on the substrate, P is the pressure, and T is the
temperature. Expression above has taken into account im-
plicitly that the change in the density is across the x direc-
tion.
For a constant substrate area A, H, the total number of
particles N, and T, the boundary tension is expressed as fol-
lows:
=
F
L
y

A,H,N,T
. 8
Above equation can be related to pressure components by
simple mathematical manipulation,
FIG. 1. Schematic representation of the thin-thick lm over a solid surface.
044708-2 Singh, Sarma, and Kwak J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

F
L
y

A,H,N,T
= HL
x

F
V

L
x
,H,N,T

F
V

L
y
,H,N,T

= HL
x
P
xx
P
yy
. 9
Hence, thin-thick boundary tension is given by
=
HL
x
P
xx
P
yy

2
. 10
The factor 2 in Eq. 10 accounts for the two interfaces in
the simulations. P
xx
is the pressure tensor component in the x
direction, which is perpendicular to the thin-thick lm inter-
face see Fig. 1. P
yy
is the pressure tensor component in the
y direction, which is parallel to the thin-thick lm interface.
Pressure-tensor components, in this work, are obtained
from the virial formalism.
35
For pairwise-additive potentials
the expression is
p

= k
B
T +
1
V

i=1
N1

ji
N
r
ij

f
ij

, 11
where is the number density, r
ij
is the vector between the
center of mass of molecules i and j, and f
ij
=u
ij
is the
force between them; the angle brackets indicate ensemble or
time average. For discontinuous potential such as those used
in this study, the forces are impulsive, having innite mag-
nitude but acting for an innitesimal time. When integrated
over time each collision contributes a well-dened amount to
the average in the following expression:
p

= k
B
T +
1
Vt
sim

collisions
r
ij

p
ij

, 12
where t
sim
is the total simulation time and the sum is over all
collisions occurring in this time; p
ij
is the impulse associ-
ated with the collision between molecules i and j.
Prewetting critical temperature has been calculated in
the literature using various techniques.
11,36
In this work, we
have estimated the critical parameters of the thin-thick lm
by using the coexistence data and the least squares t of the
following scaling law:
N/A
l
N/A
v
= C1
T
T
cpw

c
, 13
where N/ A
l
and N/ A
v
are surface densities of thick and
thin lms, respectively, and C and
cpw
are tting param-
eters. The critical temperature T
cpw
estimated from Eq. 13
is used to calculate the critical density N/ A
c
from the least
squares t of the following equation:
N/A
l
+ N/A
v
2
= N/A
c
+ DT T
c
, 14
where D is a tting parameter.
One side of the simulation box was chosen as a surface
plane with attractive square-well potential. The opposite side
was chosen to be repulsive wall to keep the uid inside the
simulation box. The height of the box was set to a xed
value of 20 in z direction see Fig. 1 to avoid the effect of
repulsive wall on the attractive surface. The other sides of
the box L
x
and L
y
were equal in length in GC-TMMC simu-
lations. The lateral variation of the box in these two dimen-
sions represents varied system sizes with surface areas of
L
x
L
y
. GC-TMMC simulations for calculating surface thin-
thick saturated densities and pressures are conducted on
AMD Opteron CPUs. Although simulations can be done us-
ing single CPU, we utilized the parallizable nature of transi-
tion matrix and used 820 CPUs to calculate the data. Four
independent runs were performed to calculate the statistical
error.
MD has been previously used to determine the prewet-
ting transition.
14
In current investigation, we have utilized
the method for the calculation of boundary tension and com-
pared the performance with that of GC-TMMC. Initially, we
place molecules on the attractive surface of the simulation
box with L
z
constant same as in GC-TMMC such that ini-
tial surface density is slightly higher than coexistence surface
density of the thick lm. Second step is to create a vacuum
by expanding the simulation cell in x direction such that L
x
=3L
y
. Periodic boundary condition is applied only in x and y
directions. MD simulations are conducted using ETOMICA
37
package. Equilibration period is performed in a canonical
NVT ensemble. During equilibration, the temperature was
kept constant by momentum scaling, with all momenta mul-
tiplied by an appropriate factor at the end of each time step
such that the total kinetic energy of the system is consistent
with the equipartition value of the temperature.
35
Production
period is conducted in microcanonical NVE ensemble. The
reduced time step t
*
in units of

m/ was xed at 0.04.

The simulations were conducted with the system size of
2000 particles. We have taken 10
5
time steps as equilibration
period and equal number for production steps.
III. RESULTS AND DISCUSSIONS
We rst present the adsorption isotherms calculated us-
ing GC-TMMC simulation and histogram reweighting tech-
nique. The isotherms were calculated for a system corre-
sponding to simulation box size of 9920 for which
maximum number of particles in the GCMC simulations was
kept at 300. Figure 2 presents the surface density number of
FIG. 2. Adsorption isotherm plot of surface density vs chemical potential at
T
*
=0.68. Lines from left to right correspond to the wall-uid interaction
strengths,
wf
=4.0, 3.0, 2.5, and 2.0, respectively. The dotted line represents
the bulk coexistence chemical potential.
044708-3 Thin-thick surface phase coexistence J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
molecules per unit surface area as a function of chemical
potential for various wall-uid afnity at a constant tempera-
ture T
*
=kT/
f f
=0.68. At a very low chemical potential a
thin vapor lm is deposited on the surface. As we increase
the chemical potential, we observe a jump in the surface
density. This jump occurs at a certain value of chemical po-
tential lesser than that for the bulk vapor-liquid coexistence
phase shown as dashed line. Further increase in the chemi-
cal potential toward bulk vapor-liquid chemical potential
leads to a state, where liquid is spread on the surface. Similar
behavior is observed for different wall-uid afnities, viz.,

wf
=2.54.0. This discontinuous change in the density rep-
resents the prewetting transition and implicitly indicates the
value of the prewetting chemical potential. From Fig. 2, it is
evident that the prewetting chemical potential increases with
the decrease in the wall-uid afnity. For weaker attractive
surface,
wf
=2.0, prewetting transition disappears. For rela-
tively stronger attractive surface than
wf
=4.0, layering be-
havior in the adsorption isotherms is observed, as shown for

wf
=5.0 in the Fig. 2. Number of layers increase with the
increase in the wall-uid interaction not shown in the g-
ure. Prewetting transition is also found to be sensitive with
respect to the temperature of the system. Figure 3 shows the
adsorption isotherms for
wf
=3.0 for different temperatures.
As the temperature increases, the density of thick lm de-
creases and that of thin lm increases. Prewetting chemical
potential increases with the increase of temperature.
Figure 4 shows the thin-thick coexistence surface den-
sity probability distribution for
wf
=3.0 for different tem-
peratures. Such distribution is evaluated at prewetting chemi-
cal potential. To determine prewetting chemical potential
from the probability distribution generated at chemical po-
tential away from the coexistence chemical potential, the dis-
tribution data are reweighted until areas under two peaks,
corresponding to thin- and thick-lm phases, are equal.
Figure 5 presents the results of surface phase coexistence
of the thin-thick lm as a function of temperature for differ-
ent wall-uid interaction energies
wf
using GC-TMMC.
These calculations were made using simulation box size of
9920. Maximum number of molecules for the simula-
tion was kept at 300. Effect of wall-uid interaction seems
insensitive on the thin-lm density at lower temperature.
However, as the temperature increases, the thin-lm density
is distinctly increased for higher
wf
. Contrary to the thin-
lm case, the effect of wall-uid interaction is prominent on
the thick-lm density. There are distinguishable differences
in the densities of thick lms even at lower temperature.
Simulation data presented in Fig. 5 indicate that at a given
temperature surface density of thick lm is found to be
higher for the case of weak attractive surface. To understand
the effect of wall-uid afnity on the structure of thin and
thick lms, MD simulations are conducted for various wall-
uid interaction strengths. Figure 6 presents the density pro-
les of thin-thick lm phases for T
*
=0.68. Two peaks are
observed in both phases for different wall-uid afnities. The
peaks are observed at L
z
=0.8 and 1.8, respectively. Density
for the two layers formed for the surface lm is higher for
FIG. 3. Plot of adsorption isotherms of surface density vs chemical potential
for wall-uid interaction potential,
wf
=3.0. The isotherms from left to right
are for T
*
=0.62, 0.65, 0.68, and 0.75, respectively. Dotted lines represent
the corresponding bulk coexistence chemical potential for different
temperatures.
FIG. 4. Coexistence surface density probability distribution of thin-thick
lm for
wf
=3.0. The lines from the bottom are for the temperatures T
*
=0.62, 0.65, 0.68, 0.7, and 0.72.
FIG. 5. Temperature against surface density plot of thin-thick lm of SW
uid for wall-uid interaction strengths,
wf
=2.5, 3.0, and 4.0. Open sym-
bols represent data calculated from GC-TMMC+histogram reweighting
analysis. Filled symbols correspond to the estimated critical points.
044708-4 Singh, Sarma, and Kwak J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
stronger wall-uid afnity; consequently beyond the wall-
uid interaction range, L
z
=1.8, the molecules for the lower
wall-uid afnity system tend to stick together, leading to
higher density compared to higher wall-uid afnity system.
It is the density at position larger than
wf
=1.8 from the wall
appears to make a considerable impact on the overall number
density of the thin and thick lms. As wall-uid interaction
increases, the thick-lm density decreases.
The effect of wall-uid afnity is much stronger at
higher temperature, consequently, change in the surface criti-
cal behavior is observed with the change of wall-uid inter-
action strength see Fig. 5. Phase coexistence data of Table
I are tted to Eqs. 13 and 14 to obtain the critical points.
The values of critical parameters for various wall-uid inter-
actions, shown by lled symbols in Fig. 5, are listed in Table
II. Finite size effect is found to be prominent for tempera-
tures greater than T
*
=0.74, hence to remove the system size
effect, the number of particles were increased to 500, which
led to increase the simulation box size to 202020. Criti-
cal exponent calculated in this work via least squares t of
the coexistence data using Eq. 13 is far from two-
dimensional Ising value, 1/ 8. It is argued that nite size
scaling
38,39
is necessary for such calculation as shown by
Mller and MacDowell for thin-thick lm,
36
however, a thor-
ough analysis of critical properties of the current system via
nite size scaling is reserved for a future study.
One of the objectives of this work was to use MD to
obtain phase densities, pressure tensor components, and
boundary tension. Figure 7 presents the pressure components
prole along the x direction for wall-uid interaction
wf
=3.0 at T
*
=0.62 and 0.68. Pressure components P
xx
and P
yy
are observed to agree with each other within the thin- and
thick-lm phases. Noise level in the simulated data is more
for the case of thick-lm as expected due to higher surface
density. Near the thin-thick interface we observe a sharp
change in the sign of P
yy
. The difference in the two compo-
nents near the interface contributes to the boundary tension
see Eq. 10. Consequently, increase in the difference of
pressure components increases the boundary tension. In Fig.
7, we can observe that the difference is greater at lower
temperature. We further analyzed the contribution to the
boundary tension as a function of the distance perpendicular
to the substrate i.e., along the z direction. We observed that
the behavior gure not shown is similar to the nature of
density proles see Fig. 6. Contribution to the boundary
tension is signicant only by the rst two layers of the inho-
FIG. 6. Plot of density prole of thin-thick lm for varied wall-uid inter-
action strengths at temperature T
*
=0.68. Upper region presents the density
prole of thick lm, whereas the lower region displays the density prole of
thin lm.
TABLE I. Thin-thick lm coexistence data obtained from GC-TMMC simu-
lations of SW uid with varied wall-uid afnity,
wf
=2.5, 3.0, and 4.0.
Surface coexistence densities for thick N/ A
l
and thin N/ A
v
lms are
calculated for a system size L
x
L
y
=81. For higher temperatures T
*
0.74, a system size of L
x
L
y
=400 was used for coexistence density
calculations. The errors are not indicated, where the error is an order of
magnitude smaller than the last signicant digit.
T
*
N/ A
v
N/ A
l
P
*

wf
=2.5
0.62 0.028 031 1.38181 0.00 76
0.65 0.042 916 1.35229 0.001 16
0.68 0.064 734 1.31497 0.001 72
0.7 0.085 233 1.280212 0.002 205
0.72 0.113 11 1.24147 0.002 785
0.74 0.154 0120 1.185416 0.003 495
0.75 0.180 436 1.153339 0.003 895
0.76 0.223 435 1.101014 0.004 35010
0.77 0.259 123 1.061824 0.004 84010
0.78 0.323 828 0.99251 0.005 375

wf
=3.0
0.62 0.026 4813 1.331823 0.000 695
0.65 0.040 9310 1.290601 0.001 065
0.68 0.063 0408 1.239610 0.001 58
0.7 0.085 3201 1.193310 0.002 025
0.72 0.118 0541 1.136711 0.002 57
0.74 0.178 44130 1.052719 0.003 25
0.75 0.220 8994 0.996727 0.003 6451
0.76 0.283 081617 0.9138194 0.004 085

wf
=4.0
0.62 0.025 407 1.316214 0.000 655
0.65 0.039 354 1.262813 0.000 995
0.68 0.061 7113 1.199110 0.001 465
0.7 0.085 2321 1.144528 0.001 875
0.72 0.123 00144 1.075320 0.002 39
0.74 0.205 35920 0.9540122 0.003 035
0.75 0.270 142912 0.8863104 0.003 42010
TABLE II. Prewetting critical temperature T
cpw
*
, critical surface density
N/ A
c
, and critical exponent
c
for SW uid with respect to wall-uid
afnity
wf
. Critical parameters are calculated using Eqs. 13 and 14.

wf
T
cpw
*
N/ A
c

c
2.5 0.78686 0.65564 0.228762
3.0 0.765829 0.59986 0.2324155
4.0 0.755816 0.572239 0.241213
044708-5 Thin-thick surface phase coexistence J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
mogeneous lm. Increase in the temperature T
*
T
*
cpw

does reduce the contribution from each layer of the lm;

however, the behavior is still akin to the density prole at
higher temperature. Figure 8 presents the pressure compo-
nent perpendicular P
*
xx
=P
*
=P
3
/
f f
to the thin-thick lm
interface i.e., saturation pressure via MD and compares
with that from GC-TMMC. The data are being plotted in a
Clausius-Clayperon form. The results from MD and GC-
TMMC are in excellent agreement. Equilibrium pressure de-
creases as the wall afnity increases. Slope of the curve,
however, is not sensitive to the wall-uid afnity.
Figure 9 presents the surface density prole N/ A ver-
sus inhomogeneous axis i.e., L
x
for the two coexistence
phases. Isotherms are being plotted for a constant wall-uid
interaction. Akin to vapor-liquid transition, we observe a
similar behavior in the density prole of thin-thick lm tran-
sition. Sharpness of the interface decreases with the increase
in the temperature, which indicates the increase in the inter-
facial width. Similar to vapor-liquid interfacial tension, one
can connect the interfacial width with the boundary tension;
boundary tension should go to zero as the interfacial width
approaches to innity at prewetting critical temperature.
Figure 10 presents the surface density prole for various
wall-uid afnities at a constant temperature. Effect on the
interfacial width is not as drastic as we observed in Fig. 9.
Nevertheless, we can see that thick-lm density decreases
with increase of the wall-uid afnity, as also seen from Fig.
5 of the surface phase coexistence. Because of the change in
the thick-lm density, interfacial width appears to decrease
calculation is not shown with the increase in the wall-uid
afnity.
We took two approaches to investigate the inuence of
wall-uid afnity on the behavior of boundary tension. Bind-
ers formalism is employed to calculate the boundary tension
of the innite system size from a set of nite system size
calculations. Since the wall-uid afnity,
wf
=3, exhibits the
prewetting phenomenon, the value of
wf
=3 is selected to
explore size effects of the system. Figure 11 presents the
nite size scaling plot of boundary tension values for tem-
peratures T
*
=0.68 and 0.72. Scaling behavior is found to be
linear as also observed by Errington for model argon on solid
carbon dioxide system.
19
Figure 12 presents the boundary tension
*
=
f f
/
f f

FIG. 7. Pressure proles P

xx
x and P
yy
x of SW uid for wall-uid inter-
action strength,
wf
=3.0 at T
*
=0.62 and 0.68.
FIG. 8. Saturation pressure curve for thin-thick lm of SW uid for two
values of wall-uid afnity,
wf
=2.5 and 4.0. Equilibrium pressure values
from MD simulation are shown as lled symbols and those from GC-
TMMC are shown as open symbols with dotted lines, which are guides to an
eye.
FIG. 9. Thin-thick lm density prole obtained from MD simulation of SW
uid for wall-uid afnity,
wf
=3.0, with varied temperatures T
*
=0.62,
0.68, and 0.72.
FIG. 10. Thin-thick lm density prole obtained from MD simulation of
SW uid for temperature T
*
=0.68 with varied wall-uid afnity,
wf
=2.5,
3.0, and 4.0.
044708-6 Singh, Sarma, and Kwak J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
of thin-thick lms for two different wall-uid strengths. The
data summarized in Table III are generated using
GC-TMMC+FSS and MD. The agreement between two
methods is excellent. Boundary tension of thin-thick SW
lms decreases with increase of the temperature, similar to
vapor-liquid bulk surface tension; however, at a constant
temperature boundary tension increases with decrease in the
wall-uid afnity. The decreasing value of boundary tension
can be attributed to lower free energy of the two-phase sys-
tem with respect to the interface as the wall-uid afnity
increases at a constant temperature. We have observed that,
for the current set of parameters, interfacial free energy does
not scale with the system size as dramatically as for the bulk
system, leading to the requirement of very precise calcula-
tion extensive sampling for the nite size scaling. Hence, in
practice, for the current system, MD method is found to be
more effective than GC-TMMC+FSS for calculating interfa-
cial properties especially at lower temperature.
IV. CONCLUSION
We studied the rst order wetting transition of a SW
uid system using GC-TMMC and MD methods. In this
work we investigate the effects of wall-uid interaction on
the prewetting transition and boundary tension. Surface
phase-coexistence calculation was done using GC-TMMC.
Saturation pressure and boundary tension via both methods
are in good agreement. Prewetting transition is being ob-
served for
wf
in the range of 2.54.0. We observe increase
in the prewetting critical temperature and density with the
decrease of the surface afnity. Layering behavior is ob-
served at higher wall-uid afnity. We further investigated
the usefulness of two molecular simulation methods for cal-
culating the boundary tension. We observed that MD route is
computationally more suitable for the calculation of bound-
ary tension, for the present system, especially at lower tem-
perature. Boundary tension increases with the decrease in
temperature. Increase in the surface wall afnity reduces the
boundary tension.
ACKNOWLEDGMENTS
This work was supported by the Department of Science
and Technology, Government of India Grant No. SR/S3/
CE10/2006. We acknowledge further computational support
from Computer Center, Indian Institute of Technology Kan-
pur.
1
L. D. Gelb, K. E. Gubbins, R. Radhakrishnan, and M. S. Bartkowiak,
Rep. Prog. Phys. 62, 1573 1999.
2
J. W. Cahn, J. Chem. Phys. 66, 3667 1977.
3
C. Ebner and W. F. Saam, Phys. Rev. Lett. 38, 1486 1977.
4
H. T. Davis, Statistical Mechanics of Phases, Interfaces and Thin Films
Wiley-VCH, New York, 1996.
5
J. E. Rutledge and P. Taborek, Phys. Rev. Lett. 69, 937 1992.
6
R. B. Hallock, J. Low Temp. Phys. 101, 31 1995.
7
J. A. Phillips, D. Ross, P. Taborek, and J. E. Rutledge, Phys. Rev. B 58,
3361 1998.
8
G. M. E. Cheng, H. C. Lee, M. H. W. Chan, M. W. Cole, C. Carraro, W.
F. Saam, and F. Toigo, Phys. Rev. Lett. 70, 1854 1993.
9
F. Kruchten and K. Knorr, Phys. Rev. Lett. 91, 085502 2003.
10
J. E. Finn and P. A. Monson, Phys. Rev. A 39, 6402 1989.
11
W. Shi, X. Zhao, and J. Johnson, Mol. Phys. 100, 2139 2002.
12
M. B. Sweatman, Phys. Rev. E 65, 011102 2001.
13
S. Sokolowski and J. Fischer, Phys. Rev. A 41, 6866 1990.
FIG. 11. Finite size boundary tension as a function of the system size for
temperatures T
*
=0.68 upper plot and 0.72 lower plot. Open symbols
represent the simulation data and the solid line shows an extrapolation of
boundary tension at innite system size. Simulation data are collected for
system sizes with 20L40.
FIG. 12. Comparison of the thin-thick lm boundary tension values from
MD and GC-TMMC+FSS simulations as a function of temperature. Open
circles are the line tension values from MC simulation for
wf
=3.0. MD
simulation data are represented by triangles.
TABLE III. MD simulation results of boundary tension for SW uids over
a homogeneous solid surface with wall-uid afnities
wf
=3.0 and 4.0.

wf
=3.0
wf
=4.0
T
*

*
T
*

*
0.6299 0.49446 0.6271 0.42694
0.6617 0.38115 0.6503 0.32836
0.6829 0.2819 0.6807 0.19113
0.7007 0.2204 0.7023 0.1255
044708-7 Thin-thick surface phase coexistence J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
14
M. J. P. Nijmeijer, C. Bruin, and A. F. Bakker, Mol. Phys. 72, 927
1991.
15
M. Fitzgerald, R. R. Picard, and R. N. Silver, Europhys. Lett. 46, 282
1999.
16
B. A. Berg and T. Neuhaus, Phys. Rev. Lett. 61, 9 1992.
17
A. M. Ferrenberg and R. H. Swendsen, Phys. Rev. Lett. 61, 2635 1988.
18
J. R. Errington, Langmuir 20, 3798 2004.
19
J. R. Errington and D. W. Wilbert, Phys. Rev. Lett. 95, 226107 2005.
20
H. Liu, S. Garde, and S. Kumar, J. Chem. Phys. 123, 174505 2005.
21
F. del Rio, E. Avalos, R. Espindola, L. F. Rull, G. Jackson, and S. Lago,
Mol. Phys. 100, 2531 2002.
22
J. R. Elliot and L. Hu, J. Chem. Phys. 110, 3043 1999.
23
G. Orkoulas and A. Z. Panagiotopoulos, J. Chem. Phys. 110, 1581
1999.
24
J. K. Singh, D. A. Kofke, and J. R. Errington, J. Chem. Phys. 119, 3405
2003.
25
L. Vega, E. de Miguel, L. F. Rull, G. Jackson, and I. A. McLure, J. Chem.
Phys. 96, 2296 1992.
26
P. Orea, Y. Duda, and J. Alejandre, J. Chem. Phys. 118, 5635 2003.
27
J. K. Singh and S. K. Kwak, J. Chem. Phys. 126, 024702 2007.
28
J. R. Henderson and F. V. Swol, Mol. Phys. 56, 1313 1985.
29
P. Bolhuis, M. Hagen, and D. Frenkel, Phys. Rev. E 50, 4880 1994.
30
N. Asherie, A. Lomakin, and G. B. Benedek, Phys. Rev. Lett. 77, 4832
1996.
31
E. Zaccarelli, G. Fof, K. A. Dawson, F. Sciortino, and P. Tartaglia, Phys.
Rev. E 63, 031501 2001.
32
A. Lomakin, N. Asherie, and G. B. Benedek, J. Chem. Phys. 104, 1646
1996.
33
Y. Zhou, M. Karplus, J. M. Wichert, and C. K. Hall, J. Chem. Phys. 107,
10691 1997.
34
K. Binder, Phys. Rev. A 25, 1699 1982.
35
M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids Clar-
endon, Oxford, 1987.
36
M. Muller and L. G. MacDowell, Macromolecules 33, 3902 2000.
37
D. A. Kofke and B. C. Mihalick, Fluid Phase Equilib. 194197, 327
2002.
38
A. D. Bruce and N. B. Wilding, Phys. Rev. Lett. 68, 193 1992.
39
N. B. Wilding, J. Phys.: Condens. Matter 9, 585 1997.
044708-8 Singh, Sarma, and Kwak J. Chem. Phys. 128, 044708 2008
Downloaded 23 Sep 2013 to 14.139.38.10. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions