Resonance contributions to HBT correlation radii

We study the effect of resonance decays on intensity interferometry for heavy ion collisions. Collective expansion of the source leads to a dependence of the two-particle correlation function on the pair momentum K. This opens the possibility to reconstruct the dynamics of the source from the K-dependence of the measured ``HBT radii''. Here we address the question to what extent resonance decays can fake such a flow signal. Within a simple parametrization for the emission function we present a comprehensive analysis of the interplay of flow and resonance decays on the one- and two-particle spectra. We discuss in detail the non-Gaussian features of the correlation function introduced by long-lived resonances and the resulting problems in extracting meaningful HBT radii. We propose to define them in terms of the second order q-moments of the correlator C(q,K). We show that this yields a more reliable characterisation of the correlator in terms of its width and the correlation strength $\lambda$ than other commonly used fit procedures. The normalized fourth-order q-moments (kurtosis) provide a quantitative measure for the non-Gaussian features of the correlator which, in the class of models studied here, are dominated by resonance decay contributions. We show that by comparing the K-dependences of the so defined HBT radius and kurtosis parameters one can separate effects from expansion flow and resonance decays.