Step 7 - Refined Data
With the 1000 iterations finished the structure factors should look something like this:
Refined total structure factors (dashed lines) compared to the experimental data (solid lines).
The total r-factor should be around a value of 4.8×10-4. This is a significant improvement from where we began with our reference forcefield, and the agreement for all datasets is now pretty good. We can of course be critical and say, for example, that the H3D8-H sample (green data in the above plot) has the worst agreement with the “doublet” ion-ion peak too pronounced in the simulation. However, what we can say is that the current simulation is overall a lot more consistent with the experimental data than the reference forcefield was, and so whatever properties we calculate from it are a better reflection of the measured reality.