Abstract
The ionization energy for a metallic cluster of radius R is I(R)=We2/R+O(R2), where W is the work function. Two different classical values for α have been derived: 1/2 from the spherical-capacitor approach, and 3/8 from the image-potential approach. We present a ‘‘classical jellium’’ or Wigner-crystal model in which α=1/2 and W=9/10e2/rs. In the image-potential approach to the ‘‘perfect-conductor’’ model, we discover an additional work 1/8e2/R, resolving the contradiction in favor of α=1/2. The experimental deviation from α=1/2 is a quantum effect.