[Q-e-developers] contribution of subroutines for evaluation of topological numbers

[H. Lee]

Dear QE developers:

I am Hyungjun Lee, a postdoc in Yazyev's group in EPFL.
Recently, there are much theoretical and experimental interest in materials
with non-trivial topology and in the theoretical point of view, evaluation
of topological numbers (TNs) characterizing non-trivial topology of real
materials is very important.

In general, calculation of TN is very similar to calculation of Berry
phases and I remember there are some packages for this purpose.

For my research, from late 2014, I maintained my codes for high-throughput
search of topological materials and they are implemented in my modified
version of QE and Vasp.

Now, if allowed, I would like to contribute my implementation to QE main
package.
There are already relevant subroutines in QE package, for example,
bp_c_phase.f90, and basically, my implementation relies heavily on them.

The main features are as follows:
(1) My implementation, in principle, can evaluate all TNs; however,
currently, input flags exit only for Z2 invariants and (mirror) Chern
numbers with spin-orbit interaction. It is easy to extend to further cases.
(2) Use of crystal symmetry in order to consider only non-equivalent k
points. In many cases, it reduces computing time significantly.
(3) Compatible with USPP and PAW with spin-orbit
(4) Evaluation of TNs is paralleled over strings.
(5) Only relevant eigenstates considered, for instance, low-lying states
originating from semi-core states are excluded since they are isolated.
(6) Use of custom FFT grid without degradation of accuracy in order to
reduce computing load

I believe that my implementation is highly suitable to, especially,
high-throughput search of topological materials.

I look forward to receiving your reply.

Sincerely,

Hyungjun Lee
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20161026/4217a6c1/attachment.html>