As the size of the nanocrystallite decreases, the conduction band (CB) upshifts and the valence band (VB) downshifts. That is your stripped down version of the quantum confinement model (QCM). This model found validation when the optical transitions observed in the 1980s revealed a blue shift with decreasing crystallite sizes. These optical experiments however specify band energies relative to each other and yield useful information mainly for the band edges. The VB photoemission reported by the authors was the first of its kind, and a non-optical technique to probe the QCM.
The experiment was carried out on CdS quantum dot (QD) bonded to a metal (Al or Au) via a short bridging alkane chain. The photoelectric emission edge was observed to shift systematically with size. The authors fit their data to a simple 1/R2 formula alongwith a constant shift due to polarization. The latter is necessary since the final state in photoelectric emission is ionized.
We have looked in vain for similar work by this or other groups in the years following this publication. Perhaps the task of bonding the QD to the metal is too daunting. Or the Alladin Synchrotron source at Wisconsin is not available. We do not know.