In general, each of the codes you names shoud offer the SAME accuracy in evaluating certain physical quantities when convergence has been achieved. The main difference is in the ease of use and eventually on the computational cost. In this view, you should band gaps that are similar with QE or ABINIT. Indeed, the accuracy of the band gap is not tied with the actual code but with the level of approximation you decide to use.

For example, standard DFT (LDA, PBE) is known for underestimating the band gap. More accurate methods, like hybrid functionals or GW methods, improve on this. But again, any code using these theories should give you the same accuracy once the parameters of those calculations are at convergence.

So you should build your chaintool more on how comfortable you are in using those tools and the kind of calculations you plan to run as they all should provide the same accuracy.

Dear Roberto D'Agosta

Thank you very much for your detailed and clear explanation! Your answer helped me to better understand that the key point is not the code itself, but the level of theory and convergence parameters. I really appreciate you sharing your experience and advice. Best regards, Mohlaroyim

The code selection is based on the application (the system you want to study). There, computational efficiency might vary significantly. You might think of this situation like this example: A Ferrari is surely an exhilarating drive on Monaco windy roads, but if you are on dirty roads you might prefer something else.

Roberto D'Agosta

In studies of radiation-induced changes in the band gap of semiconductor metal oxides, which simulation tools (e.g., Quantum ESPRESSO, ABINIT, VASP, etc.) are most suitable for capturing both the structural and electronic modifications? Are there any codes or methods that are considered more accurate or widely used for this application?

I am sorry I am not an expert on this kind of applications. Did you try to see if there are other calculations and how they are performed? A large community use VASP but its licence is expensive.

Regards,

Roberto

Dear Roberto D'Agosta

Thank you very much for your response and advice. I have reviewed several studies on band gap changes and have tried out QE, ABINIT, and SIESTA simulators, but unfortunately, I do not have access to a VASP license. I will continue testing different methods within the available codes and try to choose the most suitable one for my research. Thank you again for your recommendation!

As I said, I am expecting the different codes you have named will give essentially the same results. What could change is the further analysis you might want to do as those suites might offer different post-processing tools.

Roberto D'Agosta

Thank you for clarifying this point! That makes sense — I will also pay attention to the post-processing options and tools provided by each code as I move forward with my calculations. Your comments have been very helpful for me. Thank you again for your support!