search

Yufeng-shen / TJ2GBE

Reconstruct grain boundary energy from triple junction geometries
BSD 3-Clause "New" or "Revised" License
1 stars 2 forks source link

Request for MgO data used for Fig9a #10

Open sgbaird opened 3 years ago

sgbaird commented 3 years ago

Hi @Yufeng-shen, we have some renewed interest in looking at experimental TJ data and pairing that with a 5DOF interpolation scheme (which also imposes some regularization). Before, I was having trouble getting a reasonable interpolation for Ni data (Li 2009, DOI: 10.1016/j.actamat.2009.06.004). It was OK in some regions, but pretty poor across the board. I tried many different regularization strengths, but unfortunately, the best I could get still had pretty poor overall accuracy (I treated the TJ2GBE out GBEs as "true" values). I've been under the impression that it's due to either large experimental characterization errors or presence of metastable states (i.e. microscopic degrees of freedom). Do you have the MgO data available that was used for the paper? (i.e. the triples.dat input file and the .gbdat output file). This would really help us to troubleshoot whether we messed something up in our data pipeline by allowing us to compare directly against the MgO results from your paper (Fig 9b). If you have it available of course.

Thank you for all of your help so far.

Yufeng-shen commented 3 years ago

Currently I don't have the data at my hand, I can try to dig it. Can you contact Prof. Rohrer first? It is his data.

sgbaird commented 3 years ago

Dr. Rohrer got back to me and directed me to the Grain Boundary Data Archive for MgO triples.txt.zip file. This was enough for me to reproduce Fig9a comparable to the paper.