On the model discriminating power of mu -> e conversion in nuclei

We assess the model discriminating power of a combined phenomenological analysis of mu -> e gamma and mu -> e conversion on different target nuclei, including the current hadronic uncertainties. We find that the theoretical uncertainties can be largely reduced by using input from lattice QCD and do not constitute a limiting factor in discriminating models where one or at most two underlying operators (dipole, scalar, vector) provide the dominant source of lepton flavor violation. Our results show that a realistic discrimination among underlying mechanisms requires a measurement of the ratio of conversion rates at the 5% level (two light nuclei) or at the 20% level (one light and one heavy nucleus). We have illustrated these main conclusions also in the context of a supersymmetric model.