If electrons spin not in 2d orbits but 3d orbitals, how is it that in the magnetic domain explanation, we say that electron's motion around the nucleus is like a current loop and hence we get the magnetic field...etc...
If it is a 3d motion, then the direction of the normal of the instantaneous current loop thus approximated would change continuously.
Then how should one explain the net magnetic north-south behaviour of atoms?
The main source of ferromagnetism is the spin magnetic moment of unpaired electrons rather than the orbital angular momentum.
Having said this, the orbital angular momentum of an electron in a $d$ orbital is not zero. All orbitals are 3D in the sense that they fill all the space around the nucleus (though there are nodal surfaces in which they are zero). The angular momentum of the more complicated orbitals like the $d$, $f$, etc orbitals can't be simply understood by a classical analogy of the electron orbiting the nucleus.
How much easier would electromagnetism be to understand if the electron were given not only its intrinsic charge but also its intrinsic magnetic dipole. The discovery of an electrically charged particle - easily separated by a potential difference and easily stored by an insulator between the electric poles - has obscured the fact that subatomic particles are also natural magnetic dipoles.
The movement of electrons in orbitals around the atomic nucleus from the time of Niels Bor, Samuel Goudsmit and George Uhlenbeck in 1925 as a justification for the magnetic field is obsolete. Your question about the change in direction of the normal of the instantaneous current loop and thus the change in direction of the magnetic dipole is a good argument against the gyration of the electrons as the cause the magnetic dipole of the electron.
