[ase-users] potential issue in electron transport calculator

[Caleb Carlin]

Greetings,

I believe there is an issue with the electron transport code where using
non-identical leads with non-orthonormalized basis functions leads to an
error.  From what I can see, it is assumed that s2 is either an identity
matrix or equal to s1.  In practice I have systems where this assumption
doesn't hold, but I may be overlooking something or using the code wrong.
Is there a reason this assumption was made?

The code in question from ase/transport/calculators.py  line 159

if p['s2'] == None and not identical_leads:
    p['s2'] = np.identity(len(p['h2']))  # Orthonormal basis for lead 2
else:  # Lead2 is idendical to lead1
    p['s2'] = p['s1']

Thank you

Dr. Caleb Carlin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20161031/8dae7728/attachment.html>