Re: Periodic Boundary Conditions changes Peak width

[Bachir Aoun]

Hi Domenica,

you can set boundary conditions either in the pdb file or using the engine method set_boundary_conditions

the boundary conditions tells fullrmc about the periodicity of your system if there is any. the unit used is the angstrom, with this being said, your boundary condition is obviously wrong because you set a cube of 2A side

where the first vector is (2.0,0) the second (0,2,0) and the thrid is (0,0,2) 

REMARK    Boundary Conditions: 2.0  0.0  0.0  0.0  2.0  0.0  0.0  0.0  2.0

Before setting your boundary conditions, fullrmc considered your system as isolated and computed a shape function to compensate for the density loss. Unless you are simulating an isolated system this is wrong. 

you must specify the good boundary condition for your simulation to work right.

thanks 

[Bachir Aoun]

Hi Domenica,

There is no silly questions, don't worry about that.

I recommend you visualizing your system using Engine.visualize helper method. This requires vmd to be installed, it's a free software from illinois' university. visualize method by default will draw your boundary conditions, this way you can judge if those are visually correct or not. 

Looking at your plot, it seems to me that your high r ordering is very pronounced (high peaks) this is a bit controversial because intensity will drop with r (disordering will happen at high r due to perturbation even with highly ordered crystals) which is not the case in your experimental data.

if you believe that your experimental data is accurate and that your system is a well ordered crystal then i recommend you using other technics such as the traditional Rietveld refinement using GSAS for instance.

fullrmc is meant to study disordering in crystalline atomic systems or amorphous and molecular systems. Nevertheless, fullrmc should work perfectly fine with pure crystals but you will have hard time extracting symmetry out of a fullrmc simulation if that makes sense.

regards