It seems to be limit on the surface, however the thickness of metal layer and how deep was the ozone treatment are important to determine the final configuration. I guess the best way to corroborate the cross-section band gap distribution of your sample is by theoretical simulations of DOS.

I think the work fuction increse in overall since metal oxide behave like semiconductor which posses work function higher than that of ITO

By definition the work function is a surface property. What you may mean is that the electronic properties of your oxide film will change also in the bulk. Here the work function is not the best parameter to look after. You should consider the change of the onset of the valence band, the band alignment and if possible a band bending. All these Parameters including the work function are accessable by UPS/XPS if your sample is prepared in situ in UHV. If your sample is transfered through air you must consider a strategy for cleaning your sample to remove e.g. water and hydrocarbons as they easily change the work function and the band Bendig.

Dear,

Is this not a theoretical question: the work function has only importance in the contact with the neighbouring material (i.e. determines the potential gap for free carriers going from one material to the other). Hence, even if the ozone treatment only changes the work function of the outer layers, the effect will be the same.

Dear Jaehyun Kim,

It is well established that oxidization will increase the electron work function of metals, metal alloys and metal oxides. Work function measurements give informations about the surface not the bulk.