The accuracy of ATK electronic structure calculation

[mkalam]

We have a received a comment from a reviewer as follows regarding the ATK (which I think is a serious comment on the software capability):

"As we all know, ATK is a software for calculating transport properties, and it is not accurate for the calculation of electronic structure. Can the authors repeat the main results in the article using VASP or QE? "

Is he/she right? Is there any scientific implementation limitation in ATK compared to VASP/QE? He is suggesting to redo all the calculations using VASP or QE. This doesnt make sense. Just for the info, in our work, we did HSE06 functional based bandstructure calculation to get better accuracy over PBE. Is there any method available in VASP/QE better than ATK for electronic structure calculation?

Could anybody please clarify this? Is there any benchmarking / comparative study of ATK with VASP/QE for known systems?

Thank you very much.

[Anders Blom]

To be honest, this comment more reflects the reviewer's bias and ignorance, and might be based on some experience from many years ago.

First of all, you can do plane-wave calculations in QuantumATK and the results are just as accurate as in any other code. Now, you cannot do transport simulations with planewaves but for a huge amount of cases where we have benchmarked the LCAO basis sets, the results are very close to planewave. Also, we are confident that the HSE implementation in QuantumATK is actually more accurate than in some other codes.