We present an extended analysis of the data for the pion-photon transition form factor from different experiments, CELLO, CLEO, and BaBar, and discuss various theoretical approaches which try to reason from them. We focus on the divergent behavior of the BaBar data for the pion and those for the η(η ′ ) pseudoscalar mesons and comment on recently proposed explanations for this discrepancy. We argue that it is not possible at present to accommodate these data within the standard QCD framework self-consistently.
Keywords:Meson transition form factors, pion distribution amplitude, QCD (light-cone) sum rules 1. Two-photon processes for π 0 and η(η ′ ) in QCD At the basis of applications of Quantum Chromodynamics (QCD), which governs the interactions of quarks and gluons, is the property of factorization of corresponding amplitudes in hard processes. Proving the property of factorization means that a hadronic process at large momentum transfer can be dissected in a product of distinct quantities each associated with separate regimes of dynamics. Then, subprocesses developing at short distances and involving partons-quarks and gluons-can be accurately described in a systematic way within perturbation theory. On the other hand, the dynamical (mainly nonperturbative) features of the hadron binding effects are encoded in correlators, which contain quark and gluon field operators, and can be parameterized in terms of light-cone wave functions and parton distribution functions. These quantities are process-independent and describe the universal interpolation between hadrons and their quark and gluon degrees of freedom as distributions over the fractions of * Corresponding author * * Speaker of the talk at the "Psi-to-Phi" Conference Email address: stefanis@tp2.ruhr-uni-bochum.de (N. G. Stefanis) the longitudinal and intrinsic transverse momenta carried by the partons. They have to be determined by nonperturbative methods (models), lattice calculations, or from the data.The (spacelike) transition form factors (TFFs) of pseudoscalar mesons, in particular the pion, have been extensively studied within QCD, because in leading order they are purely electromagnetic processes with the binding QCD effects being factorized out into the pion distribution amplitude (DA)-see [1] for a review and [2] for recent references. This means that for a highly virtual photon with the four-momentum transfer Q 2 and a quasi-real photon with q 2 → 0, the TFF can be cast as the convolution of the twist-two hard-scattering amplitude T (Q 2