About all the compounds calculated by GGA+U


#1

How can I figure out all the compounds calculated by GGA+U? A full scan may find out all the compounds, but, if any simpler way to achieve this? And, have you calculated those compounds without U, just PBE? Hope for your reply! THANKS


#2

Hi again @henkekao,

Using the API, there’s an is_hubbard property for each material that you can query which is a boolean true/false value, and then a hubbards array with the U values used. This is all in the API documentation. If you’re just using the website, you can scroll down to the bottom of the materials details page, and that will give all the individual calculations used, and it will say if they’re GGA or GGA+U there.

Originally, all calculations were performed with both GGA and GGA+U, but I believe more recently we only do one or the other (whichever is appropriate). Maybe another team member can clarify this however.

Hope this helps :slight_smile:

Matt


#3

Thanks for your reply.
But how can I get those GGA data of the those compounds which calculated by GGA+U? For example, how can I get the total energy of FeO calculated by GGA?


#4

Hi Henkekao,

You can find “task data”, i. e. the data corresponding to individual VASP runs using the MPRester.get_task_data method. For example, if you want to get task data for everything that was run using GGA, you should be able to do so using the following snippet.

from pymatgen import MPRester
mpr = MPRester() # or with your API key
data = mpr.get_task_data("FeO")
data = [d for d in data if not d['is_hubbard']]

Note that none of the data that relies on aggregation or postprocessing (e. g. e_above_hull) is present.


#5

thanks @Joseph_Montoya
So, all the data obtained by this method are full relaxed and static calculated?


#6

Yes, I believe so. They might be a mixture of static and relaxation tasks, but the energies for those should be very similar. I’ll double check and get back to you.