Following Nike and Tyberius's suggestions, I rewrite my comment into an answer.
Firstly, it is obvious that when the volume of the simulation box is sufficiently large, then a significant fraction of the box must be gaseous. Secondly, note that the simulation box has a finite size, so that when the equilibrium pressure of the gas is sufficiently low, it may well be that there is not even a single gaseous molecule during the whole simulation, and all molecules cluster together as a solid, a liquid, or a mixture of solid and liquid.
But even if this happens, it does not mean that the real material is a very porous, uniform mixture of small particles/droplets and vacuum. The reason why in your simulation the material behaves as such, is that you have employed the PBC. When the PBC is active, any particle that is much smaller than the size of the box cannot meet and merge with its periodic image (otherwise the PBC is violated), so that once all molecules in a box forms a single cluster, the condensing process stops there. While in reality, there is no such thing as the PBC, and the particles/droplets are free to find each other and merge, until you end up with macroscopic rain drops or snowflakes. Unfortunately, in calculations you must either only consider a microscopic number of molecules, or use the PBC, due to computational cost considerations, which means that you have to live with such artifacts of the simulation.