Dear Prof. Gale,
Could you please clarify if the final lattice parameters (a,b,c) in the output file of example 6 Lithium substituted in MgO - Defect energy calculation, is with the defect taken into account or not?
I don’t think you’ve understood the Mott-Littleton method for defect calculations. I’d recommend reading about the method in the manual & literature, then the answer should be obvious.
This approximation
is only thus valid for small perturbations and also requires that the bulk lattice
has been optimised prior to the defect calculation.
From the above line in the manual, I consider that it is of the bulk lattice, means without the defect taken into account.
Could you confirm, please.
My kind request is please do not say that I have not understood it without knowing the origin of the doubt. I tried to replicate the lattice parameters of Ba alloyed SrTaO2N from the reported experimental literature through GULP and it was exactly the opposite.
Hence I got the doubt!
Moreover, I dont think anyone has studied lattice parameter variation upon alloying in GULP.
If you think you have understood the Mott-Littleton method, then can you summarise how it works here as a precursor to discussing your question, if you haven’t already spotted the answer from the basics of the theory.
You’ll also find there are previous studies of lattice parameter variation with mixing and alloying using GULP in the literature, but you are talking about example6 which is the use of the Mott-Littleton method for defects.
I am talking about GULP unable to reproduce experimental results as pointed in my previous message.
It would be better if you use separate threads for different topics - this one is about example 6 and your question about that.
Anyways, since my past threads I am not getting any inputs. And my hope is less, I wont get any answer even I put it as a separate thread. I am not finding this useful anyways. I hope in this small circle of computational research group, peer groups must motivate people instead of the opposite.
Surely the excitement of learning new methods is sufficient motivation, without peers having to motivate you to read the literature? Anyway, that’s a deviation from the question on which the thread is based and why the asking of the question suggests that either didn’t read the manual/literature or understand what was read. All of this can be explained through the following story that closely parallels this situation:
Once upon a time, an astronomer created a piece of software that calculated the properties of planets which they made available for free with a manual and references to the literature that explained how things were calculated. Many years later, a user asked the question what is the length of the edge of the Earth that separates its corners? The astronomer suggested that perhaps the user hadn’t read or understood the theory behind the model being used, to which the user asked how they could possibly know whether they had understood the model or not? The answer from the astronomer, that you can perhaps anticipate (with apologies to members of the Flat Earth society), was because the defining feature of the model was that it assumed that the Earth was a sphere not a cube & so the question of the distance between corners made no sense if you knew what the model was.
Now apply what you have learnt from the above to the Mott-Littleton model for simulating defects at infinite dilution, as described in the manual and literature, and perhaps you will have a lightbulb moment.
As for the other question, it’s up to you whether you wish to post it or not, but make sure you discuss it with your supervisor so that they can explain the difference between a theory, a model and a code. Most importantly, they need to explain why when something doesn’t agree with experiment it’s usually not the last one of these 3 that is the problem.