Topological analysis on grid

Alexey · 2025-09-25 17:34:34

Hello, I'm working with electron density on a fairly dense grid (more than 40,000 points at a size of 1 x 1 x 4 bohr box). When I try to perform a topological analysis, I can't generate a bond path between attractors properly. At first, I thought it was my grid, maybe it was too coarse, but it turned out to be related to the bond path generation settings. If I change the connection path generation settings, it might appear. I wanted to ask what settings are recommended for bond path generation and critical point detection when working with grid data. Thank you.

sobereva · 2025-09-26 00:11:39

Usually the default settings is OK, and I didn't find evident problems for locating BCPs for common molecules. I don't know how did you exactly do, in principle, the denser the grid the better, at least grid spacing of 0.2 Bohr should be used. Also make sure that "iuserfunc" has been set to -3 so that user-defined function corresponds to 3D B-spline interpolation.

Note that even if BCP has been located, if NCP corresponding to the bond path is missing, the bond path can also not be generated. If you find it is difficult to locate NCPs, you can also manually add the NCPs. Please always carefully check existing CPs.

Alexey · 2025-09-26 21:38:04

Dear Tian Lu, thank you for your reply. However, let me clarify my situation. I work with intermolecular interactions and want to search for critical points only between certain atoms. I'll give a simple example - the picture shows my blue box with a grid for atoms 4 and 15 (the Z axis is along atoms 4 and 15), it is 1 by 1 by 4 bohr in size, I changed the number of points in it from 32*32*32, 48*48*48 to 64*64*64, which seems to me to be a fairly dense grid. The thing is, with the standard generation settings, the bond path between atoms 4 and 15 appears only if I make 128*128*128 points in each direction of the box, which is too many. However, if I change the bond path generation settings to Stepsize, current: 0.2000, Stop generation if distance to any CP is smaller than: 0.2000, the bond path appears for a 64*64*64 grid and even 48*48*48 grids (see the other figure). My question is: is the bond path not displayed due to insufficient grid density or due to incorrect bond path generation settings for intermolecular interactions?

sobereva · 2025-09-27 04:19:22

The topology path generation algorithm was not optimized specifically for grid based density, you can try different settings until satisfactorily result is obtained. To successfully generate paths, the variation of interpolated electron density should be smooth enough, which requires small enough grid spacing and fine enough electron density. Without your file I cannot make test and provide more comments. BTW: Before performing topology analysis based on your grid data, I suggest first using main function 0 to check isosurface of the grid data (gradually changing isovalue) to ensure that variation of electron density in the interaction region is indeed quite smooth.

Whenever possible, using wavefunction file as input is always highly recommended. For isolated systems there are many freely available quantum chemistry codes can be used; for periodic systems, using CP2K to generate .molden file is highly suggested.

Multiwfn forum

#1 2025-09-25 17:34:34

Topological analysis on grid

#2 2025-09-26 00:11:39

Re: Topological analysis on grid

#3 2025-09-26 21:38:04

Re: Topological analysis on grid

#4 2025-09-27 04:19:22

Re: Topological analysis on grid

Board footer