On 29 Jun 2007, at 17:36, Synge Todo wrote:
From: Javier Vijande vijande@usal.es Date: Wed, 27 Jun 2007 10:17:33 +0200
I have another question, I would like to include a term in the hamiltonian to be solved in a 2D 4x4 square lattice of the form
H_s= -alpha * (x_i y_j - x_j y_i)
where the x and y reflect the position of the nearest neighbors, i=(x_i,y_i) and j=(x_j,y_j). I was thinking to use a term in the model definition like
<BONDTERM type="1" source="i" target="j"> -alpha*(x(i)*y(j)-x(j)*y(i))) </BONDTERM>
Is this correct or I need to define x and y and modify the lattice somewhere?.
No, currently site coordinate in BONDTERM is not supported. Although you can still use symbol 'x', 'y', etc, if you add
<INHOMOGENEOUS><EDGE/></INHOMOGENEOUS>
in your lattice definition (see, e.g., "inhomogeneous square lattice"), these symbols just return the coordinate of the bond center (not the coordinates of ends).
We can add the function x(i) and y(j) in the next release if there is a need and this cannot be done differentky. x and y are the bond0center coordinates. would they also be fine?
Matthias