pdens calculates the spatial probability densities of molecules around other molecules, and amounts to performing 3-dimensional histogram binning (into volume elements or 'voxels') as opposed to 1-dimensional histogram binning in the case of the radial distribution function (averaging the spatial probability density spherically gives the standard radial distribution function). pdens uses the defined axis origins to represent the positions of the surrounding molecules or, if one has not beeb given, the centre-of-mass is used instead. Different points on the surrounding molecules may be defined with the
For the central species, a 3D histogram is constructed for each molecular species in the system, extending a finite number of voxels of a given size in each of the Cartesian axes (controlled by the
-grid keywords respectively). The intramolecular spatial probabilty density is also calculated for the central species, and provides information on the conformational space explored by the central molecules (and 'constrained' by the defined axis system).
If the probability of finding specific orientations of surrounding molecules is needed, this can be achieved by using the
-orient keyword. This permits surrounding molecules to be excluded based on the angular difference between a specific axis on the central and surrounding molecules.
Note that, since binning is performed in three dimensions, many more frames are required compared to rdf in order to get good statistics (and hence better-looking surfaces). A post-processing utility to perform Gaussian smoothing of the data, pdensgauss, is available.
pdens <HISTORYfile> <OUTPUTfile> <sp> <othersp>
<HISTORYfile>is the name of the DL_POLY HISTORY file
<OUTPUTfile>is the name of the DL_POLY OUTPUT file
<sp>is the integer index of the species to consider as the central one, around which the probability densities will be calculated
<othersp>is a comma-separated list of indices specifying which species to consider as the surrounding ones
Although this is enough to run pdens, axis definitions for at least the central species must be provided (see the section on Defining Axes below).
-atoms sp list
For species sp, use the coordinates of each individual atom index in the supplied (and comma-separated) list as a point in the surrounding species, with each contributing individually to the final probability density. These coordinates will be used in preference to the coordinate centre derived from -axis and -otheraxis specifications. This option is mutually-exclusive with -cog.
-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 more information.
For the specified species, use basic Cartesian axes (i.e. a set of axes invariant with molecular orientation).
-cog sp list
For species sp, use the average centre-of-geometry of the atom indices in the supplied (and comma-separated) list as the point to consider in the surrounding species. This centre will be used in preference to the coordinate centre derived from -axis and -otheraxis specifications. This option is mutually-exclusive with -atoms.
Set the size of the intermolecular histogram bin voxel to be d along each side (default = 0.5).
Stop calculating when frame n is reached.
Specify that the intermolecular voxel histogram should extend npoints in each positive and negative Cartesian direction (default = 20). Note that the overall extent of the volume covered also depends on the voxel size (see the
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.
Turn on calculation of the intramolecular probability distribution.
Set tha maximum allowable distance between the central molecule and the surrounding molecule to r.
Set tha minimum allowable distance between the central molecule and the surrounding molecule to r.
Instead of considering all molecules of the central species from each frame, the supplied file dictates specific indices of the central molecule in each frame that should be considered.
Turn off calculation of the intermolecular probability distributions.
-orient sp axis angle delta
By default all surrounding molecules are binned, regardless of their orientation. The
-orient keyword allows deviations in angle between the central and surrounding molecule to be defined for each axis, outside of which the molecules are omitted. See the section on Selecting Molecule Orientations for more information.
-otheraxis sp x1 x2 y1 y2
-otheraxis keyword allows a separate set of axes to be defined for a given species which will be used to define the centre and orientation of surrounding molecules, distinct from the axes defining the central molecule. These axes, if specified, will be used in orientation calculations rather than those provided by -axis. Note that if either -atoms or -cog is given for the same species, those will be used in preference to the axes centre inferred here.
Set the size of the intramolecular histogram bin voxel to be d along each side (default = 0.15).
Specify that the intramolecular voxel histogram should extend npoints in each positive and negative Cartesian direction (default = 50). Note that the overall extent of the volume covered also depends on the voxel size (see the
-select sp pdensfile atoms cutoff
Restricts how the current pdens for species sp is constructed, by making it dependent on the pdens of a previous calculation on the same species. For the target species sp, the existing pdensfile is loaded and used as a reference. The point used to calculate this existing density must be specified using the comma-separated list of atom indices in atoms - the geometric mean of these atom coordinates is used. When calculating the new probability density, a point is added to the average only if the corresponding point specified by atoms in the existing density is equal to or above the cutoff specified. In this manner, a distribution of another point on a molecule, for instance, can be constructed relative to the probabilty of finding another point on the same molecule at some density.
Start calculating on trajectory frame n (default = 1).
Given an input HISTORY file results.HISu, then output files are as follows:
For any calculation of the spatial density around a central molecule, an axes definition must be provided that uniquely defines the local orientation of the molecule (or at least part of it). This means that the central species must contain at least three atoms in a non-linear arrangement. The exception to this rule is when using Cartesian (invariant) axes for the central molecule (see the
-cartesian keyword above), which will work even for atomic species.
Four atom indices are provided in order to define a set of axes - the first two, x1 and x2 define the direction of the x axis and also the origin of the axis system, taken to be the average of the coordinates of the two atoms. Because of this, the two indices x1 and x2 [b]must[/b] be different. The y axis is then formed from the axis origin to the average coordinates of the two atom indices y1 and y2. Note that y1 and y2 may be the same atom index, and also that they do not have to represent [i]exactly[/i] the direction of the y axis since the vector is orthogonalised with respect to the already defined x axis. The z axis is then formed from the cross product of the x and y axes.
The atoms used to define the axis system effectively 'fixes' those atoms as a point of reference for the central molecule - unless the molecule is quite rigid, the other atoms may 'average out' into a tangled ball in the corresponding intramolecular spatial density functions. Furthermore, for non-rigid molecules, the further one moves away from the atoms used to define the axis system, the more 'diffuse' and averaged the spatial density is likely to become. As such, for molecules that possess more than one site or functional group of interest it is prudent to run separate calculations with different axis definitions. For instance, for the methanol molecule we may wish to focus on both the methyl and hydroxyl ends of the molecule, and so might make the following axis definitions (performing a separate pdens calculation for each):
This shows two possible axis definitions for a methanol molecule, with the middle image focussing on the hydroxyl group (
-axis 1 1 3 2 2) and the right-hand image focussing on the methyl group (
-axis 1 6 5 4 4).
By default all molecules contribute to the spatial density, provided they are within the range of the defined grid. In some circumstances, however, it is desirable to construct the spatial density of only those molecules with specific orientations relative to the central one. In pdens this can be achieved by defining one or more angle deviations between the axes of the central and surrounding molecule, outside of which the second molecule will be ignored.
Consider the case of benzene, where we have defined the axes as follows:
As an example, let us suppose that we wish to find out in which positions co-planar molecules sit around a central molecule - in other words, how do molecules sit parallel to each other in close proximity. Since the z-axis is perpendicular to the plane of the benzene ring, we can specify an angular deviation away from this axis to request only parallel molecules are considered. Within pdens, this amounts to considering the dot product of the z axis of the central molecule with that of the candidate molecule:
We may request that only such parallel orientations are considered by passing
-orient 1 3 0.0 10.0, which would ask that species 1 (benzene) molecules should only be considered if the angle between the third axes (the z axes) is within 10.0 degrees of 0.0. However, the present example is actually a special case, since the XY plane is also a plane of symmetry, and so when the angle theta tends towards 180 degrees this also corresponds to a parallel orientation. In such circumstances a negative delta can be given, which basically indicates that a mirror plane exists for the axis in question.