ardf calculates the radial distribution function of sites, binning the results by the angle defined by vectors on each target molecule. The result, then, is a standard rdf split up into functions of individual vector angles. Such functions can be useful when studying, for instance, the variation in approach distance of planar molecules.
ardf <HISTORYfile> <OUTPUTfile> <sp1> <sp2>
<HISTORYfile>is the name of the DL_POLY HISTORY file
<OUTPUTfile>is the name of the DL_POLY OUTPUT file
<sp1>is the integer index of the species to consider as the central one
<sp2>is the integer index of the species to compare axes with and form the radial distribution function
Specify the angle histogram bin width to use in the calculation (default = 5.0).
-axis sp x1 x2 y1 y2
Define the axis system for species index sp from the four supplied atom indices. The atom indices are local atom indices for the species in question (i.e. range from 1 to the number of atoms in the molecule). See the section on Defining Axes for the pdens code for more information.
Specify the histogram bin width to use in the calculation (default = 0.1).
Request calculation and output of cross-axis correlations (i.e. xy, xz, yz etc.).
Discard n frames at the start of the trajectory.
Stop calculating after processing n frames.
Read in trajectory header from the specified DL_POLY HIStory file. Useful when the desired target frames are present in a restarted (and not appended) HIStory file.
-restrict axis angMin angMax distMin distMax
By default all surrounding molecules are used to form the ardfs, but the
-restrict allows a dependency on one axis correlation to be introduced. For instance, one may wish to calculate the distribution of the y and x axis angles if the z axes are aligned more or less parallel. The axis should be the lowercase letter indicating the axis by which to restrict the calculation, which is then followed by the minimum and maximum allowable angles for this like-axis correlation, and those for the distance.
Defines a site on sp1 that should be used in calculation of distances instead of the origin of the defined axis system. The atoms should be a comma-separated list of local atom indices in the molecule, the avarage position of which will be the new centre.
Defines a site on sp2 that should be used in calculation of distances instead of the origin of the defined axis system. The atoms should be a comma-separated list of local atom indices in the molecule, the avarage position of which will be the new centre.
Given an input HISTORY file
results.HISu, then output files are as follows:
results.ardfNNMMAABB : angle RDF between the AA and BB axes of species NN and MM (column 2 = normalised data, column 3 = unnormalised data)
results.ardfNNMMAABB.sanity : summed angle RDF between the AA and BB axes of species NN and MM, which should be equivalent to the standard RDF between (written on for AA == BB)
Since we are splitting up an average over all molecular orientations into ones that depends on the orientation of two vectors, the resulting histogram for a given angle bin must be normalised to account for solid angle effects. Consider two molecules that have their Z axes at 90 degrees to one another. Rotating one vector about the other traces a circle with radius equal to the vector's magnitude. If the angle is now made successively smaller, the circle traced out by rotating the second vector becomes smaller, until at zero degrees (when the vectors overlap) the circle has become an infinitesimally small point in space. This illustrates the 'degeneracy' of the possible ways to orient two free vectors in space - there are more 'ways' of orienting the vectors at higher angles (approaching 90 degrees) than there are at lower angles. For this reason, histograms related to vector angles nearer 90 degrees will have more points contributing to them than will those towards the extremes of zero or 180. As such, a sine function normalisation is applied to the histograms to re-weight them in the correct manner, so that peak heights may be properly compared.
NOTE: For restricted calculations the normalisation is not currently correct!