A systematic molecular dynamics study shows that the contact angle of a water droplet on graphite changes significantly as a function of the water−carbon interaction energy. Together with the observation that a linear relationship can be established between the contact angle and the water monomer binding energy on graphite, a new route to calibrate interaction potential parameters is presented. Through a variation of the droplet size in the range from 1000 to 17 500 water molecules, we determine the line tension to be positive and on the order of 2 × 10-10 J/m. To recover a macroscopic contact angle of 86°, a water monomer binding energy of −6.33 kJ mol-1 is required, which is obtained by applying a carbon−oxygen Lennard-Jones potential with the parameters εCO = 0.392 kJ mol-1 and σCO = 3.19 Å. For this new water−carbon interaction potential, we present density profiles and hydrogen bond distributions for a water droplet on graphite.

A conversion error has been detected in the analysis of the line tension (Figure 4); the units of the abscissa axis should be Å -1 . The subsequent analysis of the line tension on page 1349 is therefore in error by a factor of 10. Thus, the magnitude of the line tension can be estimated from the slopes of the fits in Figure 1, compare eq 3, which are -0.94 (case 14), -3.33 (case 1), and -3.72 Å (case 10), respectively. For a surface tension of water of γ LV ) 72 mN/m, the line tension τ is found to be 0.7 × 10 -11 (case 14), 2.4 × 10 -11 (case 1), and 2.7 × 10 -11 J/m (case 10). This error has no implication for the remaining analysis and conclusions presented in the paper. A corrected version of Figure 4 is given below.

We study the structural properties of water surrounding a carbon nanotube using molecular dynamics simulations. The interaction potentials involve a description of the carbon nanotube using Morse, harmonic bending, torsion, and Lennard-Jones potentials. The water is described by the flexible Simple Point Charge (SPC) model by Teleman et al., 1 and the carbon-water interactions include a carbon-oxygen Lennard-Jones potential, and an electrostatic quadrupole moment acting between the carbon atoms and the charge sites of the water. Vibration of the breathing mode of the carbon nanotube in water is inferred from the oscillations in carbon-carbon van der Waals energy, and the inverse proportionality between the radius of the carbon nanotube and the breathing frequency is in good agreement with experimental values. The results indicate, that under the present conditions, the presence of the water has a negligible influence on the breathing frequency. The water at the carbon-water interface is found to have a HOH plane nearly tangential to the interface, and the water radial density profile exhibits the characteristic layering also found in the graphite-water system. The average number of hydrogen bonds decreases from a value of 3.73 in the bulk phase to a value of 2.89 at the carbon-water interface. The inclusion of the carbon quadrupole moment is found to have a negligible influence on the structural properties of the water. Energy changes that occur by the process of introducing a carbon nanotube in water are calculated. The creation of a cavity in the bulk water to accommodate the nanotube constitutes the largest energy contribution. Results include an analysis of surface structure and energy values for planar and for concave cylindrical surfaces of water.

Carbon nanotube (CNT) membranes hold the promise of extraordinary fast water transport for applications such as energy efficient filtration and molecular level drug delivery. However, experiments and computations have reported flow rate enhancements over continuum hydrodynamics that contradict each other by orders of magnitude. We perform large scale molecular dynamics simulations emulating for the first time the micrometer thick CNTs membranes used in experiments. We find transport enhancement rates that are length dependent due to entrance and exit losses but asymptote to 2 orders of magnitude over the continuum predictions. These rates are far below those reported experimentally. The results suggest that the reported superfast water transport rates cannot be attributed to interactions of water with pristine CNTs alone.

No abstract

In this section we present the details on simulations we performed. Initially N + 2 graphene layers of size L × L are placed along z axis with d = 3.35Å spacing between each other. Followed is the water block of size L × L × W and N + 2 more graphene layers of the same size. The resulting system is mirrored with respect to a XY plane and shifted along Z such that distance d between graphene layers of the original system and its symmetric image is maintained. The schematic of the setup is shown in Fig. 1 of the Main Text.Each simulation consists of three stages: equilibration to isothermal-isobaric (N, P, T ) ensemble; applying temperature gradient at canonical ensemble (N, V, T ); collecting the statistics at canonical ensemble while the temperature gradient is maintained. To achieve the temperature gradient, a high temperature Nosé-Hoover heat bath is applied to four central graphene layers and a low temperature Nosé-Hoover heat bath is applied to four outermost graphene layers (two leftmost and two rightmost). Periodic boundary conditions are applied in all directions. Equilibration always takes 400 ps. Applying temperature gradient takes from 1 ns to 3 ns depending on the number of used layers, and collecting the statistics takes from 2 ns to 3 ns.Water was modeled with flexible simple point charge (SPC) model [1]. We also perform a simulation using rigid SPC water model [2]. The result are almost identical to the corresponding simulation with flexible water model: see Table S1.Typical pressure evolution versus time is shown in the Fig. S3. High pressure oscillations are observed due to the high stiffness of both water and graphene. However, the average pressure remains constant. DATA EXTRACTIONKapitza resistance R K was calculated as R K = ∆T /J, where ∆T is the temperature jump at the solid-liquid interface, and J is the heat flux through the interface. The heat flux is defined as the conducted energy from the high temperature heat bath to the low temperature heat sink per unit time across unit area. Due to the symmetry of our simulation setup (Fig. 1), the heat flux can be computed as half of the slope of the energy change with respective to time in the heat bath: J = 0.5 d dt ∆E(t), where ∆E is induced energy in heat bath (Fig. S2). In our calculations, we average the local temperature in two symmetric copies. As shown in Fig. S1, ∆T is computed from the solid and liquid temperature at the interface. Solid temperature at the interface is determined by a linear fitting of the temperatures in different graphene layers. Water is divided into bins of 0.1Å in thickness along z, thus liquid temperature at the interface is determined by fitting the temperatures of water in different bins on a straight line. The error in Kapitza resistance is given by errors of two linear fits.Important to note that with the described procedure, we calculate two Kapitza resistances -one at a high temperature interface, and the other at a low temperature interface (Fig. S1). In all our simulation, the low temperature Kapitza resis...

We study the behavior of water droplets confined in a carbon nanotube by means of parallel molecular dynamics simulations. We report radial density profiles, radial hydrogen bond distributions, and contact angles for tube radii ranging from 12.5 to 37.5 Å and for droplets containing up to 4632 water molecules. Our results indicate nonwetting behavior of the pristine CNT at room temperatures.

We study the thermophoretic motion of water nanodroplets confined inside carbon nanotubes using molecular dynamics simulations. We find that the nanodroplets move in the direction opposite the imposed thermal gradient with a terminal velocity that is linearly proportional to the gradient. The translational motion is associated with a solid body rotation of the water nanodroplet coinciding with the helical symmetry of the carbon nanotube. The thermal diffusion displays a weak dependence on the wetting of the water-carbon nanotube interface. We introduce the use of the moment scaling spectrum (MSS) in order to determine the characteristics of the motion of the nanoparticles inside the carbon nanotube. The MSS indicates that affinity of the nanodroplet with the walls of the carbon nanotubes is important for the isothermal diffusion and hence for the Soret coefficient of the system.