Hello,

I have been modeling EELS/dielectric functions of organic molecules with no crystallographic structure files in my possession, and thus I create xyz files in matlab manually.

The present molecule I am working with is a 2D (x,y) polymer chain. I would like to create an input file using the molecule attribute, Cartesian atomic coordinates, a vacuum in the y and z direction, and x,y basis vectors that are of the length and width of the molecule such that (assuming periodicity is still calculated) the neighboring cells repeat in the x direction,

I'm not sure if that is readable so let me explain how I believe this works and someone can correct me if I am wrong or confirm I am right.

When molecule is selected, the atomic coordinates are specified in cartesian coordinates in units of Bohrs. These coordinates are independent of the basis vectors chosen. The basis vectors then define the unit cell? So if I specify say a single atom at coordinate 5 5 1 with vacuum=0.00 and I choose basis vectors of

10 0 0

0 10 0

0 0 6

The atom will be modeled at a Cartesian location of 5 5 1 bohrs, In a 10x10x6 bohr unit cell that will act as a vacuum of the same vector magnitudes around the atom, due to periodicity. for example, in the x direction the atom has 5 bohrs of empty space (vacuum) next to it in the unit cell, and then an additional 5 from the next unit cell.

This boils down to a few questions:

1) if the molecule attribute is used, is periodicity still calculated? for example if the basis vectors were of lengths that correspond with the dimensions of the molecule, would the end atoms still interact? Or is a vacuum distance needed in all directions?

2) In both the standard inputs and molecular style inputs, If the basis vectors are made large enough in a direction, is it then isolated from neighboring cells (assuming effects from periodicity are calculated)? If so, how is this distance calculated? 2x nearest neighbor?

Generally, I would like to know how most and especially this DFT package handle periodicity. I believe I understand that inside the interstitial region the effects atoms impose on each other is modeled by a plane wave, but is there a distance for which no effects are calculated?

If these questions are best answered by published papers, please do feel free to recommend one that an experimentalist can read.

Appreciatively,

Rob