I intend to observe the interaction of a water droplet with the surface of a mineral particle. To do so, I have made a slab model of the mineral including 8 layers with a vaccuum of 100 A. My first question is about this model:

1. When I optimize the geometry - top three layers relaxed, charges are assigned with Qeq and ewald summation for electrostatic and VdW - the total energy has a positive value. So in another attempt I constrained all the layers, put the water cluster on the surface, and optimized the geometry; this time the energy took a negative value. Is there anything wrong with this procedure? or I could go on to the dynamics calculation with this optimized structure?

The second question is about the interaction of water molecules with the surface: 

2. At the end of a run of 500 ps, a very slight spreading of the droplet is observed if the atoms based summation method is applied, and a complete spreading happens when Ewald is used. Neither of these two are correct according to the experimental values. Which summation method should I utilize? Does the thermostat play any role here? 

Your help will be truly apprecaited.