Groups and chunks are different ways of listing atoms/particles.
Groups are used by many parts of LAMMPS, and are therefore implemented in a computationally efficient way. The implementation uses bitmasks, which restrict the use of a maximum of 32 groups (because there are 32 bits in the integers used to store the group flags of each atom).
Chunks are explicit lists of atoms, and are meant to be used primarily for analysis purposes (although you can use them to affect dynamics in a few ways). Looping over the lists is slower than checking the atom bitmasks, so chunks are meant to provide more flexibility than groups. Bin can be many things, but I presume you refer to the fractions of space that distinguish chunks.
What you are trying to do may be done equally well with groups or chunks, but you are encouraged to check the documentation of both to see what is the best option.