Dear Prof. Alexandru, Thank you for sharing the informative video. It arrived at the perfect time as I'm currently investigating the stability of defect clusters in wustite. I'm particularly interested in applying convex hull analysis to this study. I understand that this topic isn't extensively covered in textbooks, and I believe a deeper understanding of convex hull application would be invaluable to my research. Would you be able to share some resources on this subject? Thank you for your time and consideration. Best, Emmanuel
I have a conceptional question. Lets say I am studying a material which contains A and B , and A in its pure form is fcc and B is hcp . By this I mean, that a total energy for A is minimal for the crystal structure fcc , but for B it is hcp . The formation energy you defined is the total energy of A_n B_m minus n * atomic energy of A minus m * atomic energy of B. First question : What reference atomic energies do I choose ? From my understanding I have to perform a energycalculation for fcc A bulk, divide it by the number of atoms. And for the other element , I perform energy calculation for hcp B bulk. Is this right ? The second issue : Consider that I calculate the total energy of A_n B_m for the whole composition space, but only for one crystal structure. This is not the convex hull , right ? In fact, I have to calculate for all possible crystal structures the total energy over the whole composition space . However this is way to expensive. So is there a way to improve this calculation? In other words, if I know pure B is hcp , and pure A is fcc, then probably I dont have to run a calculation with fcc A_n B_m compound for large amount of B . But how to set the limit ? For this I would have to know all the possible phases..
in all cases you pick the lowest energy state you have available. and often this will include iterating over multiple possible structures, unfortunately, for each possible composition. as always, with theory, you are not guaranteed to obtain the correct experimental answer, but it's one technique you can use. high throughput computational libraries like materials project and OQMD do this for you, and there are other tools as well, such as ATAT and icet .
Thank you for your clear explanation
Ah, typos sorry: at 4:40 E_AB should be E_A3B. At 5:19 should be lowest energy structure for given stoichiometry.
Dear Prof. Alexandru,
Thank you for sharing the informative video. It arrived at the perfect time as I'm currently investigating the stability of defect clusters in wustite. I'm particularly interested in applying convex hull analysis to this study.
I understand that this topic isn't extensively covered in textbooks, and I believe a deeper understanding of convex hull application would be invaluable to my research. Would you be able to share some resources on this subject?
Thank you for your time and consideration.
Best,
Emmanuel
Thank you very much. Indeed very helpful. Also, would love to explore your group.
here it is! georgescu.lab.indiana.edu/
I have a conceptional question. Lets say I am studying a material which contains A and B , and A in its pure form is fcc and B is hcp . By this I mean, that a total energy for A is minimal for the crystal structure fcc , but for B it is hcp . The formation energy you defined is the total energy of A_n B_m minus n * atomic energy of A minus m * atomic energy of B. First question : What reference atomic energies do I choose ? From my understanding I have to perform a energycalculation for fcc A bulk, divide it by the number of atoms. And for the other element , I perform energy calculation for hcp B bulk. Is this right ?
The second issue : Consider that I calculate the total energy of A_n B_m for the whole composition space, but only for one crystal structure. This is not the convex hull , right ? In fact, I have to calculate for all possible crystal structures the total energy over the whole composition space . However this is way to expensive. So is there a way to improve this calculation? In other words, if I know pure B is hcp , and pure A is fcc, then probably I dont have to run a calculation with fcc A_n B_m compound for large amount of B . But how to set the limit ? For this I would have to know all the possible phases..
in all cases you pick the lowest energy state you have available. and often this will include iterating over multiple possible structures, unfortunately, for each possible composition. as always, with theory, you are not guaranteed to obtain the correct experimental answer, but it's one technique you can use. high throughput computational libraries like materials project and OQMD do this for you, and there are other tools as well, such as ATAT and icet .
@@AlexandruGeorgescuB thank you very much
Thank you for this video! I understand what's happening in a convex hull plot and how they did it now 😁
very happy you found it useful!