WETTING TRANSITIONS IN ALKANE-WATER NANODROPLETS
G. WILEMSKI1, F. HRAHSHEH1,2 and P. NEUPANE1
1
Department of Physics, Missouri University of Science and Technology,
Rolla, Missouri, 65409, United States.
2
Department of Physics, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia.
Keywords: Nanodroplets, Wetting, Droplet Morphology, MD Simulations.
INTRODUCTION
Interest in the morphology of liquid-liquid phase separated aerosol droplets has grown considerably of
late. Qiu and Molinero, (2015) present of large list of experimental and theoretical investigations that have
helped to advance our understanding of this subject. Most of these prior investigations have focused on the
effects of droplet composition and relative humidity on droplet morphology at a single temperature. In
contrast, our efforts have used molecular dynamics simulations to study the temperature dependent
wetting behavior of nonane-water nanodroplets and its influence on droplet structure. We use the SPC/E
water model and a unified atom 6-12 Lennard-Jones (LJ) model for alkanes. Water-alkane interactions are
treated using a LJ potential with energy and size parameters adjusted to reproduce experimentally
observed wetting behavior for nonane at 295 K. For reasonable values of the LJ parameters, core-shell
structures of nonane-water nanodroplets are observed at low temperatures, T < 235 K, indicating that
nonane perfectly wets supercooled SPC/E water with a zero contact angle at low temperatures. At higher
T, imperfect wetting is found as a nonane lens forms on the water drop with a finite, non-zero contact
angle yielding a partially engulfed Russian Doll structure. We also studied planar alkane-water interfaces.
For these systems, our simulation results are consistent with the occurrence of low temperature and high
temperature wetting transitions. Within the imperfect wetting region at low temperature, the contact angle
formed by alkanes on water initially increases with temperature until a maximum dewetting temperature
(MDT) is reached. Beyond the MDT, the contact angle decreases as the system approaches the usual high
temperature wetting transition.
COMPUTATIONAL MODELS AND METHODS
All simulations were performed with either the GROMACS (Berendsen et al., 1995; Hess et al., 2008) or
LAMMPS (Plimpton, 1995) molecular dynamics simulation packages. Long time simulations of 10-50 ns
were carried out using the (NVT) canonical ensemble with periodic boundary conditions and a time step
of 2 fs. The Nosé-Hoover algorithms (Nosé, 1984a, 1984b; Hoover, 1985) were used to control the system
temperature with a coupling time constant of 0.2 ps. Water was treated with the SPC/E potential
(Berendsen et al., 1987) Methyl and methylene groups of the alkanes were treated in the united atom
approximation. All short-range intermolecular interactions were handled with Lennard-Jones (LJ)
potentials with a cut-off radius of 1.5 nm. The long-range coulomb interaction between water molecules
was also cut-off at 1.5 nm. Both of the short-range and long-range interactions were handled with twinrange cutoffs and the neighbor list was updated every fifth time step (10 fs). For the alkane models, we
used force field parameters from Yi and Rutledge (2009). In our earlier work (Hrahsheh and Wilemski,
2013), we used the geometric mean combining rule to determine the LJ parameters (ε,σ) between unlike
atoms, but these values produced much larger contact angles for nonane drops on water compared to
thermodynamic estimates using the best measured values for the nonane-water interfacial tension (Goebel
and Lunkenheimer, 1997). Because our simulated surface tensions for pure water and nonane are
systematically somewhat lower than the experimental values, we chose to adjust the (ε,σ) values not to fit
the measured nonane-water interfacial tension, but rather to approximate the spreading coefficient S for
nonane (N) on water (W). The spreading coefficient is defined as the following difference of interfacial
tensions γ, S = γW – (γN + γNW). While adjusting the (ε,σ) values, we discovered that the simulated contact
angles were rather sensitive to small changes in either parameter. A similar observation has been reported
by Qiu and Molinero (2015). In addition to the nanodroplet simulations, we also performed many
simulations of alkane-water planar interfaces. Two types of planar simulation systems were used: (1)
liquid (alkane or water) slab inserted in vapor and (2) a slab of water sandwiched between a pair of
identical alkane slabs surrounded by vapor. From these simulations, the interfacial tensions for liquidvapor and liquid-liquid interfaces could be determined. These values could then be used with the rigorous
thermodynamic formula, (Rowlinson and Widom, 1982)
cos 𝜗 =
2
2
𝛾𝑊
− (𝛾𝑁2 + 𝛾𝑁𝑊
)
2𝛾𝑁 𝛾𝑁𝑊
to calculate the contact angle 𝜗. Similar values of 𝜗 would also be obtained from the simpler Young’s
equation, as shown by Qiu and Molinero (2015).
RESULTS AND CONCLUSIONS
Nonane-water nanodroplets containing 1000 water molecules and 500 nonane molecules were simulated at
T = 230 K and 295 K using the adjusted cross interaction parameters ε = 0.65 kJ/mol and σ = 0.362 nm.
For comparison, the values used in our earlier simulations were ε = 0.55 kJ/mol and σ = 0.359 nm.
Snapshots of cross-sections of the simulated nanodroplets are shown in Fig. 1. The low temperature
droplet clearly has a core-shell structure indicative of perfect wetting with a contact angle of 0o. In
contrast, the higher temperature droplet has assumed a Russian doll structure characteristic of imperfect
wetting. The nonane lens has a nonzero contact angle on water.
Figure 1. Cross-sectional snapshots of simulated water (blue) and nonane (yellow) nanodroplets at T =230
K (left) and T = 295 K (right). The core-shell structure is evident on the left, while the warmer droplet on
the right has a Russian doll structure characteristic of imperfect wetting.
The occurrence of perfect wetting at low temperatures is consistent with the unusual behavior found
previously by Hrahsheh and Wilemski (2013) in which 𝜗 increased with T. We explored this further by
calculating 𝜗 from interfacial tensions obtained from simulations of planar bulk interfaces using cross
interaction parameters ε = 0.65 kJ/mol and σ = 0.359 nm. The results obtained so far for nonane and water
are shown in Figure 2.
Figure 2. Temperature dependence of the nonane-water contact angle found from MD simulations of planar
bulk interfaces. Two sets of results are shown: The older results of Hrahsheh were found using GROMACS;
the newer results of Neupane (with error bars) were obtained with LAMMPS. The two sets are in good
agreement. The solid line is based on fits to the simulated interfacial tensions.
Calculated contact angles are quite sensitive to small fluctuations in the simulated values of the interfacial
tensions. To reduce this sensitivity, we also calculated 𝜗 using fits to the simulated interfacial tensions.
This results in a much smoother variation of 𝜗 with T, as can be seen in Figure 2. Using the fitted surface
tensions, the calculated contact angles show that both an upper and lower wetting transition temperature
occurs for the nonane-water and octane-water (not shown here) systems. In contrast to conventional
expectations, the simulated contact angles formed by octane and nonane on water at low temperature
(T≈250K) tend to increase with temperature until reaching a maximum dewetting temperature (MDT).
Similar behavior was observed in nonane-water nanodroplet simulations but only at low T (< 300) where
droplet simulations can be performed. For the planar systems, after the MDT is passed, the contact angle
decreases, as expected, until perfect wetting is achieved at a higher temperature. Although lower wetting
transition temperatures have been reported for partially miscible liquids with lower consolute points, the
presence of a low-T wetting transition in the essentially immiscible alkane-water systems is highly unusual
and has not been reported before. In future work, we will explore whether this low-T wetting transition is
found only for the SPC/E water model or if it occurs for other popular models such as TIP5P (Mahoney
and Jorgensen, 2000) or monatomic water (Molinero and Moore, 2009).
ACKNOWLEDGEMENTS
This work was supported in part by the National Science Foundation under Grant No. CBET-1033387.
REFERENCES
Berendsen, H. J. C., J. P. M. Postma, W. F. van Gunsteren, A. DiNola, and J. R. Haak (1984). Molecular
dynamics with coupling to an external bath, J. Chem.Phys. 81, 3684.
Berendsen, H. J. C., J. R. Grigera, and T. P. Straatsma (1987) The missing term in effective pair
potentials, J. Phys. Chem. 91, 6269.
Berendsen, H. J. C., D. van der Spoel, and R. van Drunen (1995). GROMACS: A message-passing
parallel molecular dynamics implementation, Comp. Phys. Comm. 91, 43.
Goebel, A. and K. Lunkenheimer (1997). Interfacial tension of the water/n-alkane interface, Langmuir 13,
369.
Hess, B., C. Kutzner, D. van der Spoel, and E. Lindahl (2008). GROMACS 4: Algorithms for highly
efficient, load-balanced, and scalable molecular simulation, J. Chem. Theory Comput. 4, 435.
Hoover, W. G. (1985). Canonical dynamics: Equilibrium phase-space distributions, Phys. Rev. A 31,
1695.
Hrahsheh, F. and G. Wilemski (2013). Fluctuating structure of aqueous organic nanodroplets, in
Nucleation and Atmospheric Aerosols AIP Conf. Proc. 1527, (Melville, New York), 63.
Hrahsheh, F. (2015). Unpublished work.
Mahoney, M.W. and W. L. Jorgensen (2000). A five-site model for liquid water and the reproduction of
the density anomaly by rigid, nonpolarizable potential functions, J. Chem.Phys. 112, 8912.
Molinero, V. and E. B. Moore (2009). Water modeled as an intermediate element between carbon and
silicon, J. Phys. Chem. B 113, 4008.
Nosé, S. (1984a). A unified formulation of the constant temperature molecular dynamics methods, J.
Chem. Phys. 81, 511.
Nosé, S. (1984b). A molecular dynamics method for simulations in the canonical ensemble, Mol. Phys.
52, 255.
Plimpton, S. (1995). Fast parallel algorithms for short-range molecular dynamics, J. Comp. Phys.117, 1.
Qiu, Y. and V. Molinero (2015). Morphology of liquid-liquid phase separated aerosols, J. Am. Chem. Soc.
137, 10642.
Rowlinson, J.S. and B. Widom (1982). Molecular Theory of Capillarity, (Oxford University Press, New
York, N.Y.), p. 211.
Yi, P. and G. C. Rutledge (2009). Molecular simulation of crystal nucleation in n-octane melts, J. Chem.
Phys. 131, 134902.