Talk:Mod:Hunt Research Group/sdfs generate

Generating Spatial Distribution Functions using TRAVIS

M. Brehm and B. Kirchner; J. Chem. Inf. Model. 2011, 51, 2007-2023

In general, spatial distribution functions (SDFs) may be thought of as 3D enhancements to radial distribution functions (RDFs). Remember that RDFs indicate the probability of finding a particle/atom within a radially averaged distance to another particle/atom.

Thus, SDFs depict the probability of finding a ‘particle/atom’ at a particular position in space around a fixed reference system consisting of other particles/atoms.

To generate an SDF we must

Select a reference molecule. For example, we select the bmim cation as the reference molecule, provided we want to locate the average position of Cl anions relative to a bmim cation.

Moreover, to set up the reference frame, three atoms (which must not be located in one line) are chosen within the reference molecule. TRAVIS places the first atom at the origin, the second atom is located on the positive X axis, and the third atom is put into the X-Y plane (with positive Y value). This is done to allow for averaging of probabilities in the space around the reference molecule. More details can be obtained from the TRAVIS reference. Very Important: As 3 atoms are defined to setup a reference molecule no single ions (i.e. Cl-) can be used as a reference.

Select the atom/particle to observe using either INTRAMOLECULAR or INTERMOLECULAR

For example to select Cl relative to bmim

I recommend leaving selections including the RADIUS OF THE SDF and the BINNING RESOLUTION at the default values.

Two selections, degree of smoothing and the values at which the iso-surfaces are typically given i.e. the unit of particle density (nm-3 or pm-3), are important.

These are important as the SDF probability in 3-D space is a 4-D function, which is obviously hard to visualize as it is. It is necessary to reduce the dimensionality to three (or even lower). A common and reasonable method for doing so is to visualize iso-surfaces on the data, which can be easily drawn in three-dimensional space. Isosurfaces are surfaces that pass through all areas with the same probability. The values of iso-surfaces are typically given in the unit of particle density (nm-3 or pm-3), specifying the particle density along the iso-surface.

Moreover, when analysing short or medium size trajectories, the sampling of data is usually too poor for direct SDF creation. This results in poor SDFs, the appearance of which can be improved by smoothing the underlying data. More details can be found in the TRAVIS paper.the cations is observed.

I recommend using 3 smoothing degrees and either of the iso-value scales.

Visualising the SDF using VMD

1. Open VMD

2. Load the reference XYZ coordinates generated by TRAVIS as a NEW MOLECULE

                  Use FILE –– New Molecule

3. Load generated SDF .cube file for the reference coordinates. Make sure you are aware of the original and smoothed versions (include .sx.; i.e. original = name.cube, S1 = name.s1.cube, S3 = name.s3.cube).

4. Generate Graphical representations

Graphics –– Representations

5. I like to change the drawing method to CPK for the coordinates.

6. Create a representation for the isosurface.

Drawing method = Isosurface, Vol = the file you have loaded, Isovalue = what you require (NB 1 is bulk) – in the example I have used 2.5, which represents 2.5 times bulk (higher probability than average), Draw = Solid Surface, Show = Isosurface. Colouring method = whatever you want!!!