8832 J. Phys. Chem.

C 2007, 111, 8832-8842

Molecular Structure of the Chloroform-Water and Dichloromethane-Water Interfaces†

Dennis K. Hore,‡ Dave S. Walker, Libby MacKinnon,§ and Geraldine L. Richmond*

Department of Chemistry, UniVersity of Oregon, Eugene, Oregon 97403
ReceiVed: October 31, 2006; In Final Form: December 19, 2006

Using equilibrium molecular dynamics simulations, we investigate the structure of chloroform-water and
dichloromethane-water interfaces. These systems are analyzed in terms of the orientation of water, chloroform,
and dichloromethane molecules as a function of distance from the Gibbs surface. We also calculate order
parameters for all molecules across the interface. The results show that the structures may be described in the
context of a few distinct regions of the interface, where organic and water molecules generally arrange
themselves either to maximize hydrogen-bonding interactions or to minimize the net dipole moment. In cases
where molecular arrangements promote hydrogen-bonding interactions, they often do not provide complete
dipole compensation. The orientation of organic and water molecules then creates a field across the interface.
For the CHCl3-water interface, the field has a uniform direction with the positive end of the dipole pointing
toward the bulk water phase. The interfacial field is more complicated in the case of CH2Cl2-water but is
oriented in the same direction close to the bulk water side.

1. Introduction important contributor to surfactant structure. At biological

The interface between water and chlorinated organic liquids interfaces, for example, the structure of many proteins is
is of central importance to a broad category of scientific, stabilized by association with solvated ions. The historic opinion
industrial, biological, and environmental disciplines. Chloroform has been that ions are repelled from the interface between water
and dichloromethane (DCM) in particular have received a lot and a hydrophobic medium.32,33 This view has been challenged
of attention, primarily because of their widespread use as recently with detailed molecular dynamics studies that show
solvents but also because the small size of these molecules is an enhancement of polarizable anions at the air-water inter-
amenable to computational study. Because chloroform is often face,35,36 and this provides some justification for the observed
considered a model hydrophobic polar liquid, there have been negative charge at many hydrophobic interfaces with water.37
many experimental and simulation studies of molecules adsorbed It has been demonstrated recently that the origin of this effect
at the chloroform-water interface. These include ions of various may lie in a field created by oriented interfacial water
sizes,1,2 proteins,4,5 phospholipids,7 nanocrystals,8 polymer- molecules.38 This field would then attract hydroxide ions, either
enzyme complexes,9,10 and polyelectrolytes.7 Given this wide from the autodissociation of water or from an additional source,
range of applications, it is surprising that there have not been to the interface. The current study entails detailed investigations
more studies of the neat chloroform-water interface, with the of the orientation of both water and organic molecules at the
exception of early interfacial tension measurements11,12 and interface. We will therefore use our findings on the structure to
studies of chloroform adsorbed to ice.13,14 The dichloromethane- additionally speculate on the origin, size, and polarity of any
water interface has a similar spectrum of applications, such as orientation-induced field at the interface and discuss its conse-
self-assembly of polymer surfactants,16-18 solvent dechlorina- quence on the surface charge.
tion,20 ion partitioning,21 and microemulsions of these two We will begin with a brief overview of the details of our
liquids.22,23 There is also biological interest in this liquid-liquid molecular dynamics simulations, followed by a description of
interface with studies of adsorbed phosopholipids24 and the primary analyses performed with the resulting data. Two
proteins.25-28 The neat dichloromethane-water interface has systems will then be discussed in turn, starting with the
been studied by means of interfacial tension measurements30 chloroform-water interface, followed by the dichloromethane-
and molecular dynamics simulation.31 In nearly all of the above water interface. A general comparison of these two systems then
studies where structural information is sought, it is either of a follows.
surfactant at the liquid-liquid interface or of the water
molecules. This paper addresses the interfacial structure by 2. Simulation Details
examining the mutual orientation of the water and organic Equilibrium molecular dynamics simulations were performed
molecules as a function of depth through the interface. with the Amber 7 package39 (using the Amber 7 force field)
The nature of the charge at a liquid interface is an important and an integration time of 1 fs. Initial configurations consisted
consideration, both from a fundamental standpoint and as an of separate (40 Å)3 boxes, each containing 2135 POL339 water,
† Part of the special issue “Kenneth B. Eisenthal Festschrift”.
475 chloroform, or 601 dichloromethane molecules. These were
* Corresponding author. E-mail: richmond@uoregon.edu. Phone: 541- subject to energy minimization and were then equilibrated for
346-4635. Fax: 541-346-5859. 200 ps. Temperature was controlled by weak coupling to a heat
‡ Present address: Department of Chemistry, University of Victoria,
bath at 300 K; molecular geometries were constrained using
Victoria, British Columbia, V8W 3V6, Canada.
§ Department of Chemistry, California Polytechnic State University, San the SHAKE algorithm; long-range interactions were cut off at
Luis Obispo, CA 93407. 8 Å using the Particle Mesh Ewald technique. The partial atomic
10.1021/jp067176t CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/22/2007
Chloroform-Water and Dichloromethane-Water J. Phys. Chem. C, Vol. 111, No. 25, 2007 8833

Figure 1. Organic and water boxes, showing the configuration of the

system for molecular dynamics simulations. Values of n depend on
the density of the organic liquid. z ) 0 corresponds to the lower Gibbs
surface. For all analyses, molecules above the system center are treated
as if they originated in a corresponding location in the lower half of
the box. The dotted lines depict 1-Å-thick slabs that were used to obtain
the statistics for the analyses.

charges that were used to construct the organic models were

used from the Amber 7 library for chloroform (same as Dang’s
polarizable model40) and from the literature for dichlo-
romethane.41 Water-organic boxes were then assembled as
illustrated in Figure 1. Such configurations allowed for easier
control of the center of mass while instituting periodic boundary
conditions in all three directions. The energy of the entire system
was then minimized and allowed to reach equilibrium by
following the dynamics for 2 ns. Snapshots of all of the
coordinates were then recorded every 50 fs for 10 ns, thereby Figure 2. In the laboratory frame, the Gibbs dividing surface is in the
collecting 200 000 configurations. It was necessary to collect xy plane, so z is the normal to this plane. (a) In the frame of the water
molecule, c is defined to be along the C2 symmetry axis, pointing away
such a large amount of data (nearly 1 Tb of files on disk)
from the hydrogens. The molecule lies in the ac plane. The Euler angle,
because the sought information on the molecular orientation θ, is defined as the angle between the z and c unit vectors. The twist
would be obtained through the generation and analysis of joint angle, ψ, parametrizes the rotation about the molecular c axis. (b) For
tilt-twist angle histograms. It has been recognized that con- chloroform, the C-H vector defines the molecular c axis. (c) The
structing such plots with useful resolution typically requires dichloromethane axes are defined in an analogous manner to those of
about an order of magnitude more data than would be used for the water molecule, with H-C-H defining the ac plane, and the c
one-dimensional analyses.42,43 All calculations and analyses were axis bisecting the H-C-H angle.
performed on our 16-processor AthlonMP/Opteron cluster,
running Fedora Core 3 linux, and using the LAM message- 3.1. Quantitative Description of Molecular Orientation.
passing libraries. There are two useful metrics for describing the orientation of
molecules at a liquid interface. One is the extent to which the
3. Results collection of molecules is ordered with respect to a designated
direction in the laboratory frame; the other is the actual
We now examine all of the data produced by the molecular
orientation of the molecules, specified by Euler angles (θ, φ,
dynamics simulations to arrive at a structural description of the
ψ) that relate the molecule-fixed to laboratory frames. The first
interface between water and the two chlorinated organic liquids.
description is achieved by specifying three biaxial order
First, we will describe the two categories of quantities that we
parameters.44-46 One is for the tilt angle, θ, defined as the angle
have calculated, namely, order parameters and orientation
between the molecular c axis and the interface normal, z. (The
histograms. Next, we will outline the results of each system in
relevant molecular frame axes for each molecule are defined in
turn, beginning with the chloroform-water interface. For each
Figure 2.)
system, we will first show the order parameters for water and
This order parameter is
the organic species to gain a qualitative understanding of
interfacial features. This will be followed with an analysis of 1
the orientation histograms to arrive at a detailed picture of the S1 ) 〈3 cos2 θ - 1〉. (1)
2
water and organic structure. This data will then be summarized
with structural diagrams to illustrate the orientation of both For chloroform, we will assume that there is very weak (if any)
species throughout the interface. Finally, we use knowledge of orientational preference for rotation about the C-H axis, so S1
this orientation to comment on the dipolar interactions between will completely describe the order. For water and dichlo-
water and organic species to generate net fields perpendicular romethane, however, we must consider that a twist preference
to the plane of the interface. about the molecular c axis may be significant. For those systems,
8834 J. Phys. Chem. C, Vol. 111, No. 25, 2007 Hore et al.

we therefore additionally determine a twist (ψ) order parameter

〈sin θ cos 2ψ〉

S2 ) . (2)
〈sin θ〉
A three-dimensional orientation normally requires the additional
specification of an azimuthal (φ) orientation, and corresponding
order parameter S3. In the case of a liquid, however, there can
be no anisotropy about the surface normal, so S3 ) 0 for all
molecules in the bulk and throughout the interface. The ranges
for these order parameters are -0.5 < S1 < 1 and -1 < S2 <
1. S1 ) 1 indicates that all molecules are perfectly aligned with
their C2 axes parallel (or antiparallel) to the interface normal;
S1 ) -0.5 indicates that they are perfectly aligned perpendicular
to the interface normal, with their symmetry axes in the plane
of the interface. Likewise, for S2 a value of 1 indicates complete
alignment of all molecules with ψ ) 0° or 180°; a value of -1
indicates that the corresponding angle is 90° or 270°. An
isotropic distribution would have all order parameters equal to
zero.
Although the order parameters provide an insightful descrip-
tion of the extent of the molecular orientation at various depths,
z, with respect to the Gibbs surface (GS, z ) 0 Å), they do not
suffice to describe the actual orientation of the molecules. To
study this orientation directly, we constructed joint tilt-twist
θ-ψ angular histograms for 20 1-Å-thick slabs ranging from z
) -11 Å (on the bulk organic side of the GS) to z ) +8 Å (on
the bulk water side). The ranges of the Euler angles in the
current definition are 0° e θ e 180° and 0° e ψ e 360°. As Figure 3. Tilt S1 and twist S2 order parameters for water (red) and
a result of the molecular symmetry (C2V for water and dichlo- chloroform (blue) at the chloroform-water interface. The behavior of
romethane; C∞V for chloroform’s C-H bond) and the isotropic both molecules throughout the interface may be grouped into three
distribution in the azimuthal angle φ, the twist angle is unique distinct regions (A-C), described in the text.
only in the range 0° e ψ e 90°. We have chosen to plot 0°e
ψ e 180° and so this symmetry is apparent. Because the data directed toward the bulk water phase, and the other O-H bond
is noisy in regions where there are few molecules (water toward the bulk chloroform) and in-plane molecules (with the
molecules in the organic phase, or organic molecules in the H-O-H plane parallel to the Gibbs surface). Straddling water
water phase), this repeated range for ψ allows for an easy visual appears as S2 > 0 (region B) and in-plane water appears as S2
discrimination of noise artifacts from real trends. < 0 (region C) in the bottom part of Figure 3.
Even with a sample size that consists of 200 000 configura- A more complete description of the orientation requires
tions of the entire system, the resulting histograms had a fair studying the orientation histograms shown for water in Figure
amount of noise. We have resolved this issue by applying a 15 4 and chloroform in Figure 6. For water, these are shown as
× 15 pixel (degree) median filter. Median filtering is very joint θ-ψ histograms assembled at 1 Å slices through the
common in data48 and image49 analysis. It generally does a better interface. Bold inset numbers indicate distance (in angstroms)
job than a Gaussian filter and has two primary advantages over from the Gibbs dividing surface, located at z ) 0 Å. The bulk
a simple mean filter. First, the median is more robust than the chloroform phase occurs at z < 0 Å; the bulk water phase at z
mean because it is not as sensitive to outliers. Second, because > 0 Å. Blue colors indicate low populations, and red colors
the median value is indeed one of the histogram values, no new indicate large populations, relative to that of an isotropic
counts are created in the histogram images. We have verified distribution of molecules (such as found in the bulk water phase).
that in all cases the median filtering does not alter the location, The slices in the top row (z ) -11 to -8 Å) appear uniformly
relative intensity, or shape of the features. The detailed features blue because there are few water molecules this deep in the
of the results for each system will now be described in turn. chloroform, and they do not have any preferred orientation with
3.2. Chloroform-Water Interface. Tilt and twist order respect to the surface. Likewise, the slices in the bottom row (z
parameters for the chloroform-water interface are shown in ) 5-8 Å) are red because there are many water molecules on
Figure 3. Classifying the interfacial structure according to the the bulk water side of the interface. The last slice (z ) 8 Å)
sign of the order parameters results in three regions. In the first has a near-uniform intensity, indicating that the water molecules
one, labeled A in Figure 3, S1 and S2 > 0 for water and S1 ≈ have no preferred orientation this far from the interface.
0 for chloroform. This indicates that water molecules are aligned Beginning at z ) -6 Å (region A) on the chloroform side of
such that their symmetry axes are closer to the interface normal the interface, a bright stripe along θ ) 0° is present, evidence
than to the plane of the interface, and there is neglible orientation for water aligned with its symmetry axis along the interface
of the chloroform here. In regions B and C, S1 < 0 for water normal. Additionally, these histograms reveal the “sense” of
and S1 > 0 for chloroform. In this region, water molecules have this orientation: the water molecules here have their oxygens
their symmetry axes aligned parallel to the plane of the interface. pointing toward the bulk water phase, and hydrogens toward
The chloroform C-H vector is aligned along the interface the bulk chloroform. A typical slice in the center of region B
normal. With this water axis orientation, we still need to may seen at z ) -4 Å, where the distribution of water symmetry
distinguish between straddling molecules (with one O-H bond axes is centered at θ ) 70° from the interface normal, and the
Chloroform-Water and Dichloromethane-Water J. Phys. Chem. C, Vol. 111, No. 25, 2007 8835

Figure 4. Snapshots of water molecule orientation at the chloroform-water interface. Blue indicates the smallest population; red indicates the
largest population. Bold inset numbers indicate distance (in angstroms) from the Gibbs dividing surface. Negative values are on the bulk organic
side; positive values are on the bulk water side. Shaded gray background is used to distinguish regions A-C, classified according to the order
parameters in Figure 3.

twist angle distribution is centered at ψ ) 0, 180°, indicative preference for both orientations. The histogram at z ) +2 Å is
of straddling water molecules with one O-H bond directed representative of strongly oriented water molecules in region
toward the chloroform and the other toward the bulk water. This C. Here we can see that the orientation distribution is centered
geometry is illustrated in Figure 5a and results in a fairly narrow about θ ) 90°, ψ ) 90° describing in-plane water molecules
angular distribution for the OH oscillator directed toward the (Figure 5b). A fine feature in these histograms, not observed in
organic phase (the so-called “free OH”) centered at 18° from the order parameters, is the onset of a second straddling
the interface normal. The transition region between dominant distribution at z > 1 Å. However, unlike the one that comprises
straddling and in-plane orientations observed as S2 ) 0 at z ≈ the major water orientation on the organic side of the GS
-1.5 Å may be studied in the z ) -2 Å and z ) -1 Å slices (labeled “up”), this one has the water symmetry axis tipped
of Figure 4. Between these distances, we can see an equal downward to θ ) 120°. This species (illustrated in Figure 5c,
8836 J. Phys. Chem. C, Vol. 111, No. 25, 2007 Hore et al.

Figure 6. Orientation distribution of the chloroform C-H axis,

additionally binned according to distance from the Gibbs surface. Blue
colors indicate small population; red colors indicate large population
of chloroform relative to the bulk isotropic distribution.

Figure 7. Hydrogen-bonding geometry corresponding to the maximum

orientation of chloroform molecules. Dashed blue arrows indicate
hydrogen bonds between water and chloroform; green arrows show
Figure 5. Dominant orientations of water molecules at the chloroform- the direction of the chloroform permanent dipole moment.
water interface. (a) Deepest into the organic phase (regions A and B),
straddling water molecules have their symmetry axes tilted upward, maximum extent of in-plane water molecule orientation (nega-
toward the bulk water. (b) Moving past the GS toward the bulk water tive peak of S2 < 0 for water in the bottom part of Figure 3).
phase, the dominant orientation in region C becomes in-plane water
molecules. (Note that the axes in this frame are x and y.) (c) Coexisting
Figure 7 shows that this arrangement is optimal for hydrogen
with in-plane water molecules in region C, there is a lesser population bonding between water and chloroform. This may also account
of straddling species. Here the symmetry axes are tipped away from for the reason that the distribution of C-H bond vectors are
the bulk water phase. Azimuthal rotation φ is indicated as a reminder not more tightly constrained about the interface normal: a tilt
that the orientation is unformly distributed about z. of the C-H bond creates a stronger H bond with in-plane water
molecules. Water molecules in this orientation have their
labeled “down”) has a broad distribution of free OH angles permanent dipoles in the plane of the interface. For the
centered at 68° from the interface normal. chloroform molecules, however, there is an out-of-plane com-
The chloroform orientation distribution may be described in ponent of the permanent dipole moment, indicated by the green
a more compact manner in Figure 6 because we are tracking arrows in Figure 7.
only the tilt angle of the C-H bond. We can see that the strong Now that we have a quantitative understanding of the
alignment of this bond parallel to the interface normal has a orientation of water and chloroform molecules throughout the
polarity such that the hydrogen points toward the bulk water interface, it is interesting to consider the combined effect of
phase. The orientation distribution has a peak between 0° and number density and orientation on the field across the interface.
20° from the interface normal, a width (at half max) of about Figure 8 shows the projection of the permanent dipole moment
10 Å, centered around 2 Å into the bulk water phase. Greater for the water (red) and chloroform (blue) molecules across the
than ca. 6 Å into the bulk water phase, the histogram appears interface. The total out-of-plane dipole moment (water and
noisy because there are so few chloroform molecules in this chloroform together) is plotted in black. In our convention,
region. dipoles oriented such that their positive end (“plus” side of the
Considering the water and organic orientations together, it is arrow in the figure) is directed up toward the bulk water are
evident that the maximum extent of chloroform orientation (seen assigned positive values of the polarization on the y axis. If the
as the peak of the blue chloroform curve, 2 Å into the bulk negative end of the dipole (arrowhead) is oriented in this
water phase in the top part of Figure 3) coincides with the direction, then the polarization is given a negative sign. On the
Chloroform-Water and Dichloromethane-Water J. Phys. Chem. C, Vol. 111, No. 25, 2007 8837

Figure 8. The z-component of the net dipole moment for 1 Å slabs

through the chloroform-water interface. Contributions from water
molecules are shown in red, chloroform in blue, total out-of-plane dipole
moment in black. The sign of the polarization indicates the direction
of the net dipole moment, depicted with arrows.

basis of what we have seen in our orientation analysis, it is no

surprise that water molecules on the organic side of the interface
contribute to setting up a field that is negative toward the bulk
water and positive toward the bulk organic. This results from
the water dipole axes being, on average, tilted slightly up so
the more electronegative oxygens are directed toward the bulk
water phase (Figure 5a). On the bulk water side of the Gibbs
surface, however, the dominant orientation for water molecules Figure 9. Tilt S1 and twist S2 order parameters for water (red) and
is their dipolar axes in the plane of the surface. We have, dichloromethane (blue) at the dichloromethane-water interface. The
however, described a smaller population that is also straddling behavior of both molecules throughout the interface may be grouped
into four distinct regions (A-D), described in the text.
the interface here, but with its dipolar axes tipped toward the
bulk organic phase (Figure 5c). It is these molecules that create histograms for water (Figure 10) and dichloromethane (Figure
a field with the opposite polarity (positive in our convention). 12), we can conclude that both molecules are generally oriented
We have seen that the chloroform molecules are pointed in the with their molecular c axes (defined in Figure 2) pointing toward
same direction throughout the region in which they are oriented. the bulk water phase. This same feature was observed for the
This creates a field with the same sign as that from the water chloroform-water interface, but the extent of the orientation is
molecules on the bulk water side of the GS. In the region where much greater for dichloromethane as evidenced by the magni-
the water polarization is negative, the positive chloroform tude of the order parameter and the orientation profiles in the
polarization has a greater magnitude. As a result, the overall histograms. Figure 13a illustrates this orientation and shows that
net dipole moment is oriented such that the negative end is on water and DCM molecules are arranged to maximize their
the bulk organic side of the interface and the positive end is on hydrogen-bonding interactions. It is interesting to note, however,
the bulk water side. that this arrangement does not offer any cancellation of the
3.3. Dichloromethane-Water Interface. We will investigate permanent dipole moments of either molecule. We should
the structure of the dichloromethane-water interface using the therefore expect a large interfacial field on the organic side of
same treatment we applied to the chloroform water interface. the Gibbs surface. A close examination of the organic orientation
As an overview, tilt and twist order parameters for the DCM- histograms in Figure 12 reveals a second, minor population of
water interface are shown in Figure 9. The behavior of these DCM molecules that have their chlorine atoms pointing toward
molecules is slightly more complicated than that for the chloro- the bulk organic phase and hydrogen atoms pointing up toward
form water interface, and so the analyses will be categorized in the bulk water phase. Figure 13b illustrates that this does not
terms of four regions, labeled A-D. The actual orientation of create any H-bonding opportunities with water molecules in the
the water molecules will be shown in the joint θ-ψ histograms same region. Even water molecules in higher layers (closer to
in Figure 10. Again, the same general features observed for the the Gibbs surface) will not H-bond favorably with such DCM
chloroform-water interface will be identified here, but the orientations because those also have their oxygens pointing up
appearance of the data is quite different on account of the more toward the bulk water. However, the dipole arrows in Figure
pronounced oxygen-up water orientation, to be described below. 13b show that the motivation for this arrangement may in fact
Because the orientation of dichloromethane molecules also be compensation of the water and organic permanent dipole
necessitates the specification of tilt and twist angles, this will moments.
be described by additional θ-ψ histograms in Figure 12. We In region B, which occurs between z ) -5.5 Å and -2.5 Å,
now describe features of the DCM-water interface, moving S1 < 0 and S2 > 0 for water while S1 > 0 and S2 > 0 for
from the layers deepest in the organic phase (region A) toward dichloromethane. This region contains straddling water mol-
the bulk water side of the interface (region D). ecules, with their H-O-H plane perpendicular to the plane of
Examining the order parameters in Figure 9 shows that in the Gibbs surface. There is a gradual change from the oxygen-
region A, S1 and S2 > 0 for water; S1 > 0 and S2 < 0 for up orientation (shown in Figure 11a) of the previous layers to
dichloromethane. Together with an inspection of the joint θ-ψ the straddling orientations observed here. During this progres-
8838 J. Phys. Chem. C, Vol. 111, No. 25, 2007 Hore et al.

Figure 10. Snapshots of water molecule orientation at the dichloromethane-water interface. Blue indicates the smallest population; red indicates
the largest population. Bold inset numbers indicate distance (in angstroms) from the Gibbs dividing surface. Negative values are on the bulk
organic side; positive values are on the bulk water side. Shaded gray background is used to distinguish regions A-D, classified according to the
order parameters in Figure 9.
sion, the tilt angle of the water symmetry axis increases from population of organic molecules that have their hydrogen atoms
37° at z ) -5 Å (free OH centered at 15°; Figure 11b) to 58° in the plane of the interface and their chlorine atoms straddling
at z ) -3 Å (free OH centered at 6°; Figure 11c) to 63° at z ) this plane.
-2 Å (free OH centered at 11°; Figure 11d). As a result of the In region C, the small distance between z ) -2.5 Å and z )
azimuthal isotropy, these angles describe cones about the -1 Å, dichloromethane molecules are oriented as in the previous
interface normal. Looking at the organic orientation for the region, only to a lesser degree. The water molecules, however,
corresponding slices in Figure 12 shows that there are now two have begun to make their transition to an in-plane orientation.
populations of dichloromethane. The dominant one is oriented At this point, both straddling and in-plane orientations of water
as it was in region A, with chlorine atoms pointing up toward molecules have nearly equal populations, as can be seen between
the bulk water phase. There is now an additional small the -2 Å and -1 Å slices in Figure 10. In this region, an
Chloroform-Water and Dichloromethane-Water J. Phys. Chem. C, Vol. 111, No. 25, 2007 8839

molecules show S1 < 0 and S2 > 0. The corresponding

histograms in Figures 10 and 12 show that this is to be
interpreted in terms of in-plane water molecules and organics
with their hydrogens straddling the plane (chlorine atoms in
the plane of the Gibbs surface). Figure 13d illustrates that this
arrangement is motivated by good hydrogen-bonding geometry.
Because the permanent dipole moments of both species are
largely in the plane of the interface, we should not expect a
large overall out-of-plane dipole moment for this region of the
interface.
With the above knowledge of the water and dichloromethane
orientations, we now consider the net dipole moment across
the dichloromethane-water interface. Figure 14 shows the out-
of-plane dipole moment of water (red), dichloromethane (blue),
and the combined total (black). The water polarization profile
has the same general shape as it did for the chloroform-water
interface, but here the negative region on the organic side of
the interface has about three times the magnitude of the positive
region on the bulk water side. The DCM polarization is negative
throughout the interfacial region and makes a very small
contribution to the total dipole moment on the bulk water side
of the interface. Overall, the net dipole changes sign through
the interface, with positive ends oriented in the direction of both
bulk phases, and a negative region localized about 2 Å into the
bulk water phase. The area under the net dipole moment curve
(black in Figure 14) that also corresponds to regions with
significant bulk water concentration are filled with a gray hatch
pattern. It may be noticed that the positive and negative hatched
regions are nearly equal in area, additionally minimizing the
net field in the water region of the interface.

4. Discussion
There have been many detailed structural investigations from
simulations of the neat air-water interface50-52 and the interface
between water and other organic liquids such as carbon
tetrachloride,54,43 dichloromethane,31 dichloroethane,56 and ben-
zene.57 Of particular relevance to the current study is the detailed
water structure at a variety of interfaces investigated by joint
tilt-twist histograms by Jedlovszky et al.42,43 The authors have
used slices through the interfacial region to identify straddling
water molecules near the organic side of the Gibbs surface and
in-plane water molecules near the bulk water side. Because of
Figure 11. Orientation of water molecules on the organic side of the the small size of their system and short simulations, however,
GS for the dichloromethane-water interface. (a) Deepest into the this was the extent of the detail that could be extracted from
organic layer in region A, water molecules are aligned with their their orientation histograms, an important contribution at the
symmetry axes normal to the GS. By definition, this necessitates the time. In the current study, the statistics afforded by a large
twist angles being uniformly distributed (ψ and φ are degenerate here). number of molecules, large area of the interface, and length of
(b-d) Moving toward the GS in region B, the symmetry axes of the the simulations enabled high-resolution maps of the orientation
straddling “up” water molecules become closer to the plane of interface.
(In-plane and straddling “down” molecules in regions C and D are not of both water and organic molecules. For the water structure,
illustrated and have orientations similar to those shown in there are two primary additional features that can therefore be
Figure 5.) observed in our data. The first is the preference of the straddling
water population on the organic side of the GS to point its
interesting subpopulation of DCM molecules develops, with the symmetry axis toward the bulk water phase. The second is the
Cl-C-Cl plane tilted approximately 120° from the interface identification of a second straddling water population, on the
normal. Figure 13c shows that such an orientation of DCM water side of the GS, coexisting with the dominant in-plane
molecules may favorably H bond with the emerging in-plane orientation in that region, and with their symmetry axes pointing
orientation of water. Figure 13c also illustrates that the toward the bulk organic side of the interface. The important
diminishing straddling water molecules in this region (tilted at consequence of these water molecules, observed in both the
120°) are not able to make strong H-bonds to DCM in this chloroform-water and DCM-water systems, is that they result
orientation. However, these organic and straddling water in a net water dipole component normal to the interface, with
molecules have components of their permanent dipole moments negative charge pointing toward the GS on the organic side and
that cancel each other within this region of the interface. negative charge pointing away from the GS on the water side.
Finally in region D of Figure 9 (-1 Å < z < 12 Å), water Had this subset of straddling water molecules on the water side
molecules exhibit S1 and S2 < 0, while dichloromethane not been observed, it would have been difficult to account for
8840 J. Phys. Chem. C, Vol. 111, No. 25, 2007 Hore et al.

Figure 12. Snapshots of dichloromethane orientation at the dichloromethane-water interface. Blue indicates the smallest population; red indicates
the largest population. Bold inset numbers indicate distance (in angstroms) from the Gibbs dividing surface. Negative values are on the bulk
organic side; positive values are on the bulk water side. Shaded gray background is used to distinguish regions A-D, classified according to the
order parameters in Figure 9.

such a sizable out-of-plane dipole moment in this region from Considering all of these structural features for both the water
primarily in-plane water molecules. In the case of the chloro- and organic molecules allows us to evaluate the net out-of-plane
form-water interface, this subset of the water population dipole moment across the interface. Results of the chloroform-
actually dominates the out-of-plane response of all of the water water interface are summarized in Figure 15a, illustrating that
molecules (Figure 8). For the DCM-water interface, the same a net dipole moment perpendicular to the interface is observed
features are observed, but the out-of-plane dipole moment on from z ) -7 Å in the chloroform to z ) +7 Å in the bulk
the organic side of the interface (with opposite polarity, Figure water phase. The interfacial dipole is oriented such that it would
14) is larger. These important observations are possible as a attract negative ions, such as hydroxide, from the bulk water
result of our large and lengthy simulations and the clarity of phase to create a negatively charged interface, in agreement with
the histogram images imparted by the median filtering. what is generally observed for water-hydrophobic liquid or
Chloroform-Water and Dichloromethane-Water J. Phys. Chem. C, Vol. 111, No. 25, 2007 8841

Figure 14. The z component of the net dipole moment for 1 Å slabs
through the dichloromethane-water interface. Contributions from water
molecules are shown in red, dichloromethane in blue, total out-of-plane
dipole moment in black. The sign of the polarization indicates the
direction of the net dipole moment, depicted with arrows. The hatched
area under the total dipole moment curve corresponds to regions with
significant water density.

Figure 15. (a) For the chloroform-water interface, there is a net out-
of-plane dipole moment oriented such that hydroxide ions from the
bulk water phase would be attracted to the interface. (b) For the
dichloromethane-water interface, there is a net dipole moment over a
larger region, but it is not unidirectional. Negative ions would be
attracted to the topmost layer, closest to the bulk water; positive ions
would accumulate in a narrow region close to the Gibbs surface.

the bulk water phase. Considering that the adjacent layer of

water molecules closer to the interface has a straddling orienta-
tion, it is reasonable to suspect that an oxygen-up orientation
would maximize the number of water-water hydrogen bonds.
However, this far into the bulk organic phase, it would normally
be difficult for water molecules to exhibit any strong orienta-
tional preference. We suspect that this water molecule orienta-
Figure 13. Predominant orientations of DCM and water molecules tion is stabilized by DCM as a result of its own large permanent
(a,b) furthest into the organic liquid (regions A and B), (c) close to the dipole moment. Below this region, extending to a depth of z )
Gibbs surface in region C, and (d) furthest into the bulk water in region
D. Dashed blue arrows indicate hydrogen bonds between water and -10 Å in the DCM, there is a large out-of-plane dipole moment
DCM; green arrows show the direction of the permanent dipole in the opposite direction. As a result, there is a small region (a
moments. few angstroms wide) centered at the Gibbs surface that is likely
to attract cations from the bulk water phase. It would be
solid interfaces.38,58 Our results predict a more complicated interesting to compare these predictions with experimental
dipole pattern for the water-DCM interface as illustrated in measurements of the interfacial potential. However, although
Figure 15b. First we observe a much smaller dipole moment in liquid-surface potentials may readily be measured,59 the differ-
the same direction on the bulk water side of the DCM-water ence between internal (Galvani) potentials for liquid-liquid
interface, from z ) 2 -7 Å, likely attracting fewer anions from interfaces cannot be measured directly.60
bulk water. From our structural results, we can see the origin It is important to note that a more rigorous calculation of the
of this dipole moment in the large population of water molecules surface polarization must take into account contributions from
on the DCM side of the GS with their oxygens pointing toward not only the permanent dipole moments of the molecules but
8842 J. Phys. Chem. C, Vol. 111, No. 25, 2007 Hore et al.

also the induced dipole moments and local field effects. This (14) Picaud, S.; Hoang, P. N. M. Phys. Chem. Chem. Phys. 2004, 6,
would necessitate more sophisticated methods of simulation such 1970-1974.
(15) Holmes, N. S.; Sodeau, J. R. J. Phys. Chem. A 1999, 103, 4673-
as quantum or mixed quantum-classical molecular dynamics and 4679.
the incorporation of flexible water and organic geometries and (16) Boury, F.; Ivanova, T.; Panaı̈otov, I.; Proust, J. E.; Bois, A.; Richou,
force fields. Nevertheless, the component of the permanent J. J. Colloid Interface Sci. 1995, 169, 380-392.
(17) Balashev, K.; Bois, A.; Proust, J. E.; Ivanova, T.; Petkov, I.;
dipole moment normal to the interface is an important and Masuda, S.; Panaı̈otov, I. Langmuir 1997, 13, 5362-5367.
significant contributor to the interfacial polarization. Our results (18) Malzert, A.; Boury, F.; Saulnier, P.; Ivanova, T.; Panaı̈otov, I.;
indicate that in both the chloroform-water and DCM-water Benot, J. P.; Proust, J. E. J. Colloid Interface Sci. 2003, 259, 398-407.
systems the balance between the organic and water dipole (19) Babak, V. G.; Boury, F. Colloids Surf., A 2004, 243, 33-42.
(20) Wan, C.; Chen, Y. H.; Wei, R. EnViron. Toxicol. Chem. 1999, 18,
moments to create the net dipole moment in each layer is not 1091-1096.
a subtle one: orientations of the molecules create strongly (21) Nagy, P. I.; Takács-Novák, K. J. Am. Chem. Soc. 2000, 122, 6583-
enhancing dipole contributions in many regions of the interface. 6593.
It is anticipated that, although future simulations will be able (22) Berthod, A.; Georges, J. J. Colloid Interface Sci. 1985, 106, 194-
202.
to measure these effects in a more quantitative manner, the (23) Boury, F.; Olivier, E.; Proust, J. E.; Benot, J. P. J. Colloid Interface
overall trends would remain the same. Sci. 1994, 163, 37-48.
(24) Saulnier, P.; Boury, F.; Malzert, A.; Heurtault, B.; Ivanova, T.;
5. Conclusions Cagna, A.; Panaı̈otov, I.; Proust, J. E. Langmuir 2001, 17, 8104-8111.
(25) Sah, H. J. Controlled Release 1999, 58, 143-151.
Equilibrium molecular dynamics simulations have been used (26) van de Weert, M.; Hoechsteter, J.; Hennink, W. E.; Crommelin,
to investigate the structure of chloroform-water and dichlo- D. J. A. J. Controlled Release 2000, 68, 351-359.
(27) Pérez, C.; Griebenow, K. J. Pharm. Pharmacol. 2001, 53, 1217-
romethane-water interfaces. A detailed study of orientation of 1226.
water and organic molecules has been performed as a function (28) Malzert, A.; Boury, F.; Saulnier, P.; Benot, J. P.; Proust, J. E.
of distance from the Gibbs surface. The results have been Langmuir 2002, 18, 10248-10254.
(29) Sah, H.; Bahl, Y. J. Controlled Release 2005, 106, 51-61.
presented in terms of tilt and twist angle order parameters and (30) Bahramian, A.; Danesh, A. Fluid Phase Equilib. 2004, 221, 197-
orientation histograms. In these two interfacial systems, both 205.
water and organic molecules may reorient to minimize the (31) Dang, L. X. J. Chem. Phys. 1999, 110, 10113-10122.
surface free energy. It appears that the minimum energy structure (32) Adam, N. K. The Physics and Chemistry of Surfaces; Oxford
University Press: London, 1941.
is a balance between dipole compensation and hydrogen-bonding (33) Bikerman, J. J. Surface Chemistry: Theory and Applications;
geometry optimization. In cases where creating hydrogen bonds Academic Press: New York, 1958.
is more energetically favorable, a significant net dipole moment (34) Chattoraj, D. K.; Birdi, K. S. Adsorption and the Gibbs Surface
perpendicular to the plane of the interface may exist. For the Excess; Plenum: New York, 1984.
(35) Jungwirth, P.; Tobias, D. J. J. Phys. Chem. B 2002, 106, 6361-
chloroform-water system, this net dipole is oriented such that 6373.
a slight positive charge occurs on the bulk water side of the (36) Mucha, M.; Frigato, T.; Levering, L. M.; Allen, H. C.; Tobias, D.
interface. This may be responsible for the purported enhance- J.; Dang, L. X.; Jungwirth, P. J. Phys. Chem. B 2005, 109, 7617-7623.
(37) Kreuzer, H. J.; Wang, R. L. C.; Grunze, M. J. Am. Chem. Soc.
ment of hydroxide ions at water-hydrophobic interfaces. 2003, 125, 8384-8389.
However, when the same analysis is applied to the dichlo- (38) Zangi, R.; Engberts, J. B. F. N. J. Am. Chem. Soc. 2005, 127, 2272-
romethane-water system, a more complex field pattern is 2276.
observed with a much smaller dipole moment on the bulk water (39) Case, D. A. et al. Amber 7; University of California: San Francisco,
CA. 2002.
side of the interface. We aspire that our results will lead to a (40) Chang, T.-M.; Dang, L. X. J. Phys. Chem. B 1997, 101, 10518-
better understanding of the adsorption of large and small electro- 10526.
lytes, biomolecules, and other surfactants at these interfaces. (41) Dang, L. X. J. Phys. Chem. B 2001, 105, 804-809.
(42) Jedlovszky, P.; Vincze, A.; Horvai, G. J. Mol. Liq. 2004, 109, 99-
108.
Acknowledgment. We thank the Department of Energy, (43) Jedlovszky, P.; Vincze, A Ä .; Horvai, G. Phys. Chem. Chem. Phys.
Basic Energy Sciences (DEFG02-96ER45557) for the financial 2004, 6, 1874-1879.
support of this project. L.M. was supported by the NSF REU (44) Buffeteau, T.; Lagugné, Labarthet, F.; Sourisseau, C.; Kostromine,
program. S.; Bieringer, T. Macromolecules 2004, 37, 2880-2889.
(45) Straley, J. P. Phys. ReV. A 1974, 10, 1881-1887.
(46) Photinos, D. J.; Bos, P. J.; Doane, J. W.; Neubert, M. E. Phys.
References and Notes ReV. A 1979, 20, 2203-2211.
(1) Schnell, B.; Schurhammer, R.; Wipff, G. J. Phys. Chem. B 2004, (47) Collings, P. J.; Photinos, D. J.; Bos, P. J.; Ukleja, P.; Doane, J. W.
108, 2285-2294. Phys. ReV. Lett. 1979, 42, 996-999.
(2) Schurhammer, R.; Engler, E.; Wipff, G. J. Phys. Chem. B 2001, (48) Stone, D. C. Can. J. Chem. 1995, 73, 1573-1581.
105, 10700-10708. (49) Meyer-Arendt, J. R. Two-Dimensional Digital Signal Processing;
(3) Muzet, N.; Engler, E.; Wipff, G. J. Phys. Chem. B 1998, 102, Springer-Verlag: Secaucus, NJ, 1981; Vol. 2.
10772-10788. (50) Morita, A.; Hynes, J. T. Chem. Phys. 2000, 258, 371-390.
(4) Miller, R.; Fainerman, V. B.; Makievski, A. V.; Krägel, J.; (51) Morita, A.; Hynes, J. T. J. Phys. Chem. B 2002, 106, 673-685.
Grigoriev, D. O.; Kazakov, V. N.; Sinyachenko, O. V. AdV. Colloid Interface (52) Perry, A.; Ahlborn, H.; Space, B.; Moore, P. B. J. Chem. Phys.
Sci. 2000, 86, 39-82. 2003, 118, 8411-8419.
(5) Nichols, M. R.; Moss, M. A.; Reed, D. K.; Hoh, J. H.; Rosenberry, (53) DeVane, R.; Space, B.; Perry, A.; Neipert, C.; Ridley, C.; Keyes,
R. L. Biochemistry 2005, 44, 165-173. T. J. Chem. Phys. 2004, 121, 3688-3701.
(6) Lu, G.; Chen, H.; Li, J. Colloids Surf., A 2003, 215, 25-32. (54) Chang, T.-M.; Dang, L. X. J. Chem. Phys. 1996, 104, 6772-6783.
(7) Urzúa, M. D.; Ros, H. E. Polym. Int. 2003, 52, 783-789. (55) Moreira, N. H.; Skaf, M. S. Prog. Colloid Polym. Sci 2004, 128,
(8) Liu, B.; Qian, D. J.; Huang, H.-X.; Wakayama, T.; Hara, S.; Huang, 81-85.
W.; Nakamura, C.; Miyake, J. Langmuir 2005, 21, 5079-5084. (56) Jedlovszky, P.; Vincze, A.; Horvai, G. J. Chem. Phys. 2002, 117,
(9) Wang, L.; Zhu, G.; Wang, P.; Zhang Newby, B.-m. Biotechnol. 2271-2280.
Prog. 2005, 21, 1321-1328. (57) Jedlovszky, P.; Kereszturi, A.; Horvai, G. Faraday Discuss. 2005,
(10) Zhu, G.; Wang, P. J. Am. Chem. Soc. 2004, 126, 11132-11133. 129, 35-46.
(11) Speakman, J. J. Chem. Soc. 1933, 1440. (58) Schurhammer, R.; Wipff, G. New J. Chem. 1999, 23, 381-391.
(12) Ikeda, T.; Ozaki, T. Bull. Chem. Soc. Jpn. 1950, 23, 43-45. (59) Peterson, I. R. ReV. Sci. Instrum. 1999, 70, 3418-3424.
(13) Aoki, M.; Ohashi, Y.; Masuda, S. Surf. Sci. 2003, 532-535, 137- (60) Markin, V. S.; Volkov, A. G. AdV. Colloid Interface Sci. 1990, 31,
141. 111-152.