Placing two layers of solvent relative to one another is a simple matter of using the editconf module. We will define a unit cell of desired dimensions and place the cyclohexane layer at a specific location within it. Once satisfied with the placement, we can fill the rest of the box with water.

Following the equilibration of my 466-molecule cyclohexane layer, the box dimensions were 4.30795 nm in each dimension. The goal of this exercise will be to build the biphasic system such that it has a water layer of equivalent size. Thus, we will keep the x- and y-dimensions fixed, and double the system size in z. The trick here is that if we define such a box for our cyclohexane system, editconf will automatically center the cyclohexane layer in it. While not incorrect (since, because of periodicity, the proper interfaces will still form), such a position is not advantageous if we want to add, for example, a small protein or peptide into the water layer. A divided layer will not easily allow us to do this.

Fortunately, editconf allows us to manually specify a center for our system. Since we want to keep the cyclohexane layer essentially in the same place while we just increase the box, we will specify a system center at (x/2, y/2, z/4), which is normal centering within the original box (x/2, y/2, z/2), since GROMACS builds all boxes from the coordinate origin.

gmx editconf -f chx_10ns.gro -o chx_newbox.gro -box 4.30795 4.30795 8.6159 -center 2.153975 2.153975 2.153975

We now have a system aligned in the following way:

Now, the only remaining task is to solvate the empty space with water. This is easily done with solvate:

gmx solvate -cp chx_newbox.gro -cs spc216.gro -p chx.top -o chx_solv.gro

The solvate module will place some water molecules within the cyclohexane layer, so make a copy of the file vdwradii.dat (normally in $GMXLIB) in your working directory. Change the value of the C radius from 0.17 to 0.35 and run solvate again. There may be a few stray water molecules still in the cyclohexane layer but these will come out during equilibration. Do not be concerned with water molecules that appear at the "bottom" of the box, as these will be part of the continuous solvent layer across the periodic boundary in the x-y plane.