-
|
Is there any online (or offline) documentation on how to run embedded-atom method (EAM) calculations with QUIP? The docs say this is possible (it's actually the first item in the list of featured methods). How is it invoked, how to find potential files, etc? |
Beta Was this translation helpful? Give feedback.
Replies: 6 comments 8 replies
-
|
it's called "Glue" model. you can see how it's used in the silicon PRX potential. the glue.xml has three parts, a pair potential, and a radial density, and an embedding function (energy as a function of density). for the si prx only the pair potential is used, the rest set to zero, but you can see the format in the file. |
Beta Was this translation helpful? Give feedback.
-
|
OK, thanks. I have dealt with glue potentials in the past, but just those with the pair term, not the density dependent term. Hopefully I can do the parametrization of that part. |
Beta Was this translation helpful? Give feedback.
-
|
Glue != EAM, though We also have EAM_ErcolAd in QUIP. I wouldn't particularly recommend using it in preference to EAM from ASE, matscipy or LAMMPS, unless you need to combine it with GAP. |
Beta Was this translation helpful? Give feedback.
-
|
I guess one could recreate most EAMs as a glue, no? which EAMs are not Glues? |
Beta Was this translation helpful? Give feedback.
-
|
I've never taken the time to look at the differences, just pointing out we have both implemented so presumably they're not exactly equivalent. |
Beta Was this translation helpful? Give feedback.
-
|
Both are the sum of a pair potential and a nonlinear ("embedding") function of a spherically symmetric "density" linearly summed over the neighbors. The differences in the QUIP implementation are in the level of generality that are allowed. I think the Glue allows arbitrary splined curves, and so can encompass any EAM, but not 100% sure of this. |
Beta Was this translation helpful? Give feedback.
-
|
yes, all three 1d functions are represented as splines in glue |
Beta Was this translation helpful? Give feedback.
-
|
(I'm going to press 'convert to discussion' on this issue, since it doesn't report a fault in the code) |
Beta Was this translation helpful? Give feedback.
-
|
Thanks and sorry I put it as an issue, I wasn't aware of these "discussions". |
Beta Was this translation helpful? Give feedback.
-
|
No worries - this is the very first discussion, we just turned them on today. |
Beta Was this translation helpful? Give feedback.
-
|
Looks good, like Stack Overflow with unlimited comment space, could get out of hand easily! |
Beta Was this translation helpful? Give feedback.
-
Except without scoring |
Beta Was this translation helpful? Give feedback.
-
|
I can score your comment! You just get paid in thumbs up! |
Beta Was this translation helpful? Give feedback.
it's called "Glue" model. you can see how it's used in the silicon PRX potential. the glue.xml has three parts, a pair potential, and a radial density, and an embedding function (energy as a function of density). for the si prx only the pair potential is used, the rest set to zero, but you can see the format in the file.