First, please make sure that the structure periodicity is correct. Is this a 45-rotated rectangular lattice or hesagonal lattice?
If it is rotated rectangular periodicity, you may not need to rotate it, except there is a necessary.
DIverging can be due to several different issues, such as large diffraction angle (then use Steep Angle PML with 64 layers), PML too close to structure (increase z span); It is not necessarily to use the Stabilized PML, which may stablize the simulation but result can be incorrect. Please refer this
Large than 1 usually means the simulation does not run long enough, the autoshutoff min is not reached etc. Most likely this current configuration has very large diffraction angle. Please use visualizer to check the farfield with assuming periods 10 by 10. or use grating analysis: