correct erronous structures that fail DFT-opt

[mpougin]

@danieleongari, after DFT-opt I inspected structures that fail convergence/have negative BG. I corrected the errors I spotted. Structures that couldn't be fixed within a reasonable amount of time were discarded

• 21750N2: remove overlaps
• 22050N2: only PI-COF published (relabel framework)
• 22080N2: wrong connectivity, for correction 2x1x1 supercell has to be created
• 22104N2: add missing hydrogen on benzene
• 22120N2: wrong atomic distances and missing atoms, discard
• 22121N2: wrong atomic distances and missing atoms, discard
• 22210N2: wrong atomic distances and missing atoms, discard
• 22233N2: wrong atomic distances, discard
• 22452N2: overlaps and floating atoms, discard
• 2245nN2: overlaps, discard

[danieleongari]

Thanks @mpougin. The CI seems to fail due to name "pi cof" already present: if you see similar cases I add the surname of the first author of the paper to distinguish COFs with the exact same name. A suitable name in this case would be, e.g., PI-COF-Skorjanc

[mpougin]

thanks for the heads-up. I corrected the label, sry, I forgot to run the checks locally. Should be fine to merge now