TiO partition function is problematic

[ajwheeler]

The Barklem and Collet partition function for TiO (the one in Korg) doesn't agree well with the state-of-the-art ExoMol values:

The ExoMol TiO paper compares the partition function to other values and finds much better agreement (table 9), so I think this is unambiguously a case where I just need to upgrade things.