Supplementary Information for Stable Liquid Water Droplet on Water Monolayer Formed at Room Temperature on Ionic Model Substrates Chunlei Wang1,2 , Hangjun Lu 1, 2, Zhigang Wang1, Peng Xiu1, Bo Zhou1,2 , Guanghong Zuo3, Rongzheng Wan1, Jun Hu1 and Haiping Fang1,3† 1 Shanghai Institute of Applied Physics, Chinese Academy of Sciences, P.O. Box 800- 204, Shanghai 201800, China 2 Graduate School of the Chinese Academy of Sciences, Beijing 100080, China 3 T-Life Research Center, Department of Physics, Fudan University, Shanghai 200433, China PS1. Discussions on the unrecognized possibilities associated with existing materials and potentials for new materials that can have an adsorbed water film which does not wet water. The charge distribution of our model solid structure closely resembles some surfaces such as InSb(110). In reality, there are many methods of modifying the surface to produce charged-group in the appropriate positions, such as intentional doping [S1]. Here we show the results from molecular dynamics simulation of a III-V compound InSb(110)-like surface. As shown in Fig S1, both In and Sb atoms, are fixed coplanar and all atoms prefer sp2-like bonding saturation with three nearest- neighbor atoms of a bond length of 2.80 Å [S2,S3], which is very close to the distance between every couple of positive and negative charges, 2.84 Å, in our model system. The Lennard-Jones parameters of the In atoms and Sb atoms are set εIn= 0.599 kcal/mol, σIn = 3.976 Å and εSb =0.449 kcal/mol, σSb = 3.938 Å, respectively, from UFF force field [S4]. There are positive and negative charges with a same magnitude on the In atom and Sb atoms. The explicit value of the magnitude is not clear, which is expected to be close to 1.0 e [S3]. In total, the solid surface is neutral. Numerically, a water droplet on water monolayer is clearly observed for q = 1.2 e as shown in Fig. S2. FIG. S1 Geometry of the InSb(110)-like surface. In and Sb atoms are marked by red and blue colors, respectively. FIG. S2 Side view snapshot of a water droplet on water monolayer at T= 300K. In the simulations, the initial simulation box size was Lx= 6.395 nm, Ly= 6.816 nm, and Lz = 20.0 nm, whereas Lz was large enough to eliminate the image effect due to periodic boundary condition. The periodic boundary conditions were applied in all directions. During simulation, the InSb(110)-like surface with 504 atoms was fixed. Initially, 1006 water molecules, with a water film of about 1.1 nm, completely covered the surface. A constant-temperature and constant-volume MD simulation was performed with a time step 1.0 fs using Gromacs 3.3.1. The selection of vapor-liquid coexistence system could maintain the ambient condition. A Berendsen thermostat was used to keep the temperature of the water at 300 K. The particle-mesh Ewald method with a real space cutoff of 10Å was used to treat the long-range electrostatic interactions, whereas a 10Å cutoff was applied to the van der Waals interactions. W e used molecular dynamics to simulate the water molecules for 4 ns for each value of q, and the last 2 ns were collected for analysis We note that the InSb(110) surface usually relaxes such that the In atom moves inwards and the Sb atom moves outwards. We have neglected the buckling of the InSb(110) in the simulation system here. It should be noted that, Maier et al. proposed that a water layer adsorbed on the hydrogenated diamond surface might be responsible for the substantial surface conductivity [S5,S6]. However, the existence of the water layer was questioned by the AFM measurement of the contact angle of the water droplet on the surface of the hydrogenated diamond, showing that the surface was extremely hydrophobic [S7]. Conventionally, Most of scientists believe that a water layer cannot exist on an extremely hydrophobic surface. This was pointed out again by Angus and colleagues very recently [S8]. It is clear that this paradox might be solved if the surface of the hydrogenated diamond has a similar effect on water as our model surface. In our model, if there were not enough water molecules, water monolayer would fully cover the surface. More water molecules result in the phenomenon of water droplet on water monolayer where the contact angle of the water droplet can be measured by AFM. In summary, there is a high possibility for the existing materials and for the successful fabrication of new materials that can have the phenomenon of “water droplet on water monolayer”. PS2: Method to compute the contact angles The contact angle of a water droplet was calculated from the density profile according to the method presented in Ref. S9, S10. The water molecules below 4.0 Å from the solid surface were not taken into account. A common approach to extract the contact angle from nanoscale droplets is to fit the time-averaged liquid/vapor interface to a circle as the example shown in Fig. S3. The liquid/vapor interface was defined as where the water number density fell to about one-half that of the bulk. Through fitting the liquid/vapor profile to a circle by averaging 8000 frames, we obtained the water contact angle θ. FIG. S3 A typical profile of the simulated water/vapor interface, fitted with a circle for the water contact angle calculation with q = 1e. PS3: More details of water monolayer configuration Fig. S4a shows the probability distribution of the dipole orientation Ψ of the water molecules of the outside-droplet-monolayer, where Ψ is the angle between a water dipole and the z axis. It is clear that there are two peaks at Ψ ≈ 100o and Ψ ≈ 40o. We find that they correspond to the two binding states of water as the typical examples shown in Fig. S4b. The water molecule in state 1 has one -OH pointing to surface plane and the other –OH pointing to another water molecule; the water molecule in state 2 has both -OHs pointing to neighboring water molecules. It is found that the probability of state 1 is larger than that of state 2. This structure with two binding states are quite similar to the H-down configuration of the water monolayer binding on Pt(111) surfaces at ~135 K [S11,S12]. a b FIG. S4 a, Probability distribution of the dipole orientation Ψ of the water molecules of the water monolayer outside the droplet, where Ψ is the angle between a water dipole and the z axis. There are two peaks at Ψ = 100o and Ψ = 40o, corresponding to the state 1 and state 2, respectively. b, Two typical stable binding states of the water molecules of the water monolayer outside the droplet. PS4: Discussion on the simulation of water layer on the model system at very low temperature (~135K) We adopt here the same model surface as described in main text (see Fig 1a) but the number of water molecules is only 414 water molecules. This number of water molecules corresponds to the number of water molecules fully covering the solid surface in the form of water monolayer at room temperature. Fig. S5 shows the probability distribution of the angle φ of water molecules on our model surface at T=135K, for q=0.2e, 0.3e, 0.4e, respectively. Only for q = 0.3e and 0.4 e, there are clear peaks at 0o, 120o and 240o, showing the ordered structures of the water layer at those charge values. We note that the SPC/E water model is usually not suitable for the simulation of water molecules at very low temperature since it does not predict reasonable solidification behaviour. Here, for the comparison of the behviour of the water layer for different temperature and we do not consider the solidification since only one layer of water above the solid surface is considered. FIG. S5. Probability distribution of the angle φ between the x-y plane projection of a water dipole and a crystallographic direction of monolayer water molecules at T=135K, for q=0.2e, 0.3e, 0.4e, respectively. PS5: Discussion on the simulation of water on a Pt(111)-like surface at room temperature. The van de Waals parameters of the atoms in the solid surface are also important for hydrophobic behavior. For comparison, we configured a new solid surface by removing all the uncharged atoms from the model structure shown in Fig. 1A and then substituting a neutral atom at the center of each hexagon containing the charged atoms. This model solid surface is quite similar to the structure of Pt(111). At room temperature, if the van de Waals parameters are assumed to be the same as those used previously, we found that the critical q value required forming a stable hydrophobic monolayer increased to ~0.7e. Fig. S6 shows the probability distribution of the angle φ of water molecules on a Pt-like surface for different Lennard-Jones potential length scale σ for q= 1.0e. It is clear that when σ ≤ 2.7 Å, the water layer does not show peaks at 0o, 120o and 240o, and thus no hydrophobic water monolayer exists. FIG. S6. Probability distribution of the angle φ between the x-y plane projection of a water dipole and a crystallographic direction of monolayer water molecules, for a Pt- like surface for different Lennard-Jones potential length scale σ. References S1. D. C. Wei et al. Nano Lett. Doi: 10.1021/nl803279t (2009). S2. D. W. Palmer, www.semiconductors.co.uk, 02 (2006). S3. A. Höglund, PhD thesis: Electronic Structure Calculations of Point Defects in Semiconductors. Uppsala Universitet (2007). S4. A. K. Rappé, C. J. Casewit, K. S. Colwell, III. W. A. Goddard and W. M. J. Skiff, J. Am. Chem. Soc. 114, 10024 (1992). S5. F. Maier, M. Riedel, B. Mantel, J. Ristein and L. Ley, Phys. Rev. Lett. 85, 3472 (2000). S6. A. P. Sommer, D. Zhu and K. Brühne, Cryst. Growth Des. 7, 2298 (2007). S7. Y. Kaibara, K. Sugata, M. Tachiki, H. Umezawa and H. Kawarada, Diamond Relat. Mater. 12, 560 (2003). S8. V. Chakrapani, J. C. Angus, A. B. Anderson, S. D. Wolter, B. R. Stoner and G. U. Sumanasekera, Science 318, 1424 (2007). S9. T. Werder, J. H. Walther, R. L. Jaffe, T. Halicioglu and P. Koumoutsakos, J. Phys. Chem. B 107, 1345 (2003). S10. X. Li, J. Y. Li, M. Eleftheriou and R. H. Zhou, J. Am. Chem. Soc.128, 12439 (2006). S11. G. A. Kimmel,N. G. Petrik, Z. Dohnálek and B. D. Kay, Phys. Rev. Lett. 95, 166102 (2005). S12. H. Ogasawara, B. Brena, D. Nordlund, M. Nyberg, A. Pelmenschikov, L. G. M. Pettersson and A. Nilsson, Phys. Rev. Lett. 89, 276102 (2002).