As part of one of the Molecular Visualization Principles, I am attempting to design a simulation of molecular water with correct scale, density, and an approximation of brownian motion.
I began by creating a very simple water molecule model, ensuring that the dimensions were accurate. Okay, the bond angles might be very slightly out, but who would ever be able to tell? I knew there were going to be a huge number of copies, so keeping the poly count low was a must.
When making the simulation, I decided to ignore hydrogen bonding interactions, but I did use the bonding radius for the collision radius, which created a natural looking separation between the molecules.
In order to determine how the simulation would look once rendered, I made a small cube’s worth of molecules at the size, speed, and resolution I would be using for the final animation. This is the reason it takes up only a small part of the frame. If the result was a fuzzy blurry mess, I would know that changes were required.
A used a hand animated sphere as a collider and included noisy turbulence to induce brownian motion. The final animation will include a lateral wipe to show and hide the water, so I made sure to include a simple test of that. The result is below, and I think it is successful enough to move on to the full scale simulation.
Measured bond length: 0.9584 Angstroms
Measured bond angle: 107.9 degrees