Abstract
An expression for the Knight shift, accurate to zero order in the pseudopotential, is derived and tested by calculations for Li, Na, Al, Ga, and Cs. The calculations take at most a few hours, and can be used to decide rapidly for any metal whether the pseudopotential method is likely to prove fruitful in general for that metal. In particular the results indicate whether in a more accurate calculation of the Knight shift or change in Knight shift with temperature, pressure or alloy composition, the pseudopotential approach would be appropriate, and whether relativistic, orbital or core polarisation effects might be significant.