Partial-wave analysis of proton Compton scattering data below the pion-production threshold

Abstract: Low-energy Compton scattering off the proton is used for determination of the proton polarizabilities. However, the present empirical determinations rely heavily on the theoretical description(s) of the experimental cross sections in terms of polarizabilities. The most recent determinations are based on either the fixed-$t$ dispersion relations (DR) or chiral perturbation theory in the single-baryon sector ($\chi$PT). The two approaches obtain rather different results for proton polarizabilities, most notably for $\beta_{M1}$ (magnetic dipole polarizability). We attempt to resolve this discrepancy by performing a partial-wave analysis of the world data on proton Compton scattering below threshold. We find a large sensitivity of the extraction to a few "outliers", leading us to conclude that the difference between DR and $\chi$PT extraction is a problem of the experimental database rather than of "model-dependence". We have specific suggestions for new experiments needed for an efficient improvement of the database. With the present database, the difference of proton scalar polarizabilities is constrainedto a rather broad interval: $\alpha_{E1}-\beta_{M1} = (6.8\, \ldots \, 10.9)\times 10^{-4}$ fm$^3$, with their sum fixed much more precisely [to $14.0(2)$] by the Baldin sum rule.