Hello everyone,<div><br></div><div>I am attempting to use the NEGF implementation in CP2K for transport calculations. At present I have a monolayer between two aluminum [110] contacts (details not too important as it is just a test case).</div><div><br></div><div>I have been using the examples provided online <a href="https://github.com/schulkov/cp2k/blob/negf_samples/tests/QS/regtest-negf-fft/au111_c6h4s2_kpoints_0.50V.inp">here</a> from Dr. Chulkov as my example and I have done very few modifications to it--only the coordinates/lattice. I have comprised each electrode of 16 Al layers, with with the bulk contact being 8 layers, the screening region being 4, and the the remaining 4 being part of the scattering region. The subsystems used for the electrodes are layers 5-8 (from outside in). This is consistent with how Chulkov used atoms of moltype L2 and R2 in the example for his example script. However, I continue to get the following error:</div><div><br></div><div><p> *******************************************************************************</p>
<p> * ___ *</p>
<p> * / \ *</p>
<p> * [ABORT] *</p>
<p> * \___/ One lattice vector of the contact unit cell (FORCE_EVAL section 1) *</p>
<p> * | must be parallel to the direction of the contact 1. *</p>
<p> * O/| *</p>
<p> * /| | *</p>
<p> * / \ negf_env_types.F:258 *</p>
<p> *******************************************************************************</p>
which confuses me. Looking at the source, I see that the direction_vector is that between the atoms of the bulk and screening regions of the electrode. Looking at my structure in VESTA, everything looks to be parallel. Also by calculating the origin of L2+L3 and L1, and finding the vector between them, I get [ 1.77635684e-15, -3.55271368e-15, 8.07786000]. So I'm wondering if maybe I misunderstand what needs to be parallel.</div><div><br></div><div>I'd like to know (a) how to fix this error, and (b) I'm also interested to know why, in the example script, the two electrode subsystems (Force_eval 1 and 2), only contain molecules L2 and R2, rather than L2+L3 and R2+R3, which would correspond to the whole bulk region of the contact.</div><div><br></div><div>If anyone has assistance (maybe Dr. Chulkov is on this mailing list), it would be much appreciated. I've attached my input file, and the various include files used with it, and also a cif file of my full structure.</div><div><br></div><div>Best,</div><div>Nick Winner </div>