A theory of spin-lattice relaxation is presented in which the modulated crystalline potential does not act directly, but entails modulation of orientation and polarization of the orbitals. These factors are responsible for a modulation of the effective magnetic field acting on the electron spin, which supplies the spin transitions. Three consequences result from this formulation: (a) The spin Zeeman term does not play its traditional role in the relaxation process, the reason for this being that the effective dynamic magnetic field in Kramers salts is related to $\lambda \overrightarrow{\mathrm{L}}·\overrightarrow{\mathrm{S}}$ and $\beta \overrightarrow{\mathrm{L}}·\overrightarrow{\mathrm{H}}$, and not to $2\beta \overrightarrow{\mathrm{S}}·\overrightarrow{\mathrm{H}}$. In other words, the relaxation results from the modulation of the anisotropic $g$ factor. (b) The rotational modes of the crystalline complex may play an important role, even though the amplitude of vibration is weaker than that of the vibration modes. We will see indeed that the rotating motion of the orbitals can be essential in the relaxation process; this motion is generated by all the vibration modes (which also entail polarizations), but in particular by the purely rotational modes of the complex. (c) Another effect will result from the modulation of the orbital energy; this effect will be studied in an addendum [Phys. Rev. (to be published)].

• Received 14 October 1968

DOI:https://doi.org/10.1103/PhysRevB.1.4261