Abstract
Anharmonic phonon frequency shifts Delta omega (T) and damping Gamma (T) in alkali metals are calculated using a previously described pseudopotential model. The theory is in good agreement with the most accurate measurements for Li. For Na, K and Rb agreement usually lies within the experimental errors. Relative shifts delta = Delta omega / omega and relative damping eta = Gamma / omega increase linearly with T when T>or approximately=(0.3-0.5) theta D, the coefficients being close for different alkali metals. Values of 5Gd and eta vary significantly in the Brillouin zone but remain small up to the melting point for all the phonons except those in the vicinity of the soft Sigma 4 branch. Phonon frequencies for this branch at large mod K mod increase rather than decrease with T. The sensitivity of the results to the model and calculation approximations has been studied. Comparison with other authors' calculations appears to imply that 'self-consistent phonon'-type approximations are unreliable in the description of anharmonic effects.