Abstract
See also ibid., vol.5, p.1385 (1975). Calculations of the wavevector dependent susceptibility chi (q) and dielectric functions epsilon (q) for paramagnetic nickel are reported. The calculations are based on a simple interacting band model described earlier. Owing to the highly anisotropic nature of the nickel Fermi surface, different Houston averaging schemes have been studied with the view to obtaining convergence in chi (q). It is shown that a nine-directional Houston averaging scheme is appropriate for paramagnetic nickel. The calculated chi (q) shows a broad hump around q=0.5 au which is attributed to Umklapp processes. Inclusion of matrix elements weakens this hump. The calculated xi (q) and epsilon (q) are not in agreement with those of earlier workers. The reasons for this disagreement are critically discussed.