Total energy of Si-doped Tio2 film

[zhySarah]

Hei,

I have built a 2*1*1 TiO2 supercell and substituted a Ti atom with Si. I then add optimizegeometry and total energy together to the calculator to do the geometry optimization and got one value for total energy (TE1=-9287eV). Afterwards I used the optimized Si-doped TiO2 for DFT calculations where band structure, DOS, total energy and others were added. I hereby got another value for total energy(TE2=-8437). My confusion is why TE1 and TE2 are different. I wanted to further calculate the formation energy using E(formation)=E(Si-tio2)-E(tio2)-E(si)+E(Ti), which of the two values (TE1 or TE2) should I use for E(Si-tio2)?

Thanks in advance!
   

[Anders Blom]

Probably your second total energy is from the unoptimized structure. My guess is you did

config = nlread("file.nc", BulkConfiguration)[0]
total_energy = TotalEnergy(config)

where file.nc is the output file from the optimization run. It should be

config = nlread("file.nc", BulkConfiguration)[-1]
total_energy = TotalEnergy(config)

because in that NC file, the first config [ 0 ] is the unoptimized structure, the second one [1] is the optimized one. The notation [-1] mean the LAST object, which is most likely the optimized (if you ran several times the same script, with different settings, you may have more than 2 objects in the file).

[zhySarah]

Probably your second total energy is from the unoptimized structure. My guess is you did

config = nlread("file.nc", BulkConfiguration)[0]
total_energy = TotalEnergy(config)

where file.nc is the output file from the optimization run. It should be

config = nlread("file.nc", BulkConfiguration)[-1]
total_energy = TotalEnergy(config)

because in that NC file, the first config [ 0 ] is the unoptimized structure, the second one [1] is the optimized one. The notation [-1] mean the LAST object, which is most likely the optimized (if you ran several times the same script, with different settings, you may have more than 2 objects in the file).

Thanks Anders!

FYI, for the second total energy calculation, I sent the optimized geometry (gID001)to the builder and added a new calculator and other analysis. Therefore there is no such 'nlread' command in the script.

Regards

[Anders Blom]

Then I can only assume you used different parameters. Else post scripts, otherwise there is no way to  help.