I think in cyclic voltametry you can measure it. 

Band alignments are notoriously difficult to measure reliably, mostly because the position-dependent band energy is not a physical observable.  Most of the techniques used for macroscopic epitaxial samples require a good ohmic contact technology, and even then are very indirect measurements, typically finding a built-in voltage from which the band alignment is inferred using oversimplified theoretical models. 

The most reliable approach is to use (X-ray) photoemission spectroscopy to find the energies of well-defined core states with respect to the valence-band edge in the bulk materials, and then do measurements on samples with heterolayers that are sufficiently thin that core-level emission from both the constituents can be observed.

Of course this is a technique that works best if you have placeable and orientable samples.