Potential Enhancement: Surface Defects

[kavanase]

Previously some draft code for modelling defects at surfaces was added to the surface-defects branch, but this was based on much older versions of doped and so has been deleted. Defect generation in doped is now sufficiently fast that we should be able to just provide a surface slab to doped, and it will efficiently enumerate all possible symmetry-inequivalent defects.

Key points:

• Previously the pymatgen Slab.symmetrically_remove_atoms and Slab.symmetrically_add_atom methods were used so that defects were initialised symmetrically on both sides of the slab -- is this still desired / standard practice for surface defect studies?
• In future, it would be worth adding a tutorial on defect generation for surfaces (and their charge-corrections) to the tutorials. Likely to be driven by a specific research project (let us know!).