Padé approximants, Borel transforms and renormalons: the Bjorken sum rule as a case study

We prove that Padé approximants yield increasingly accurate predictions of higher-order coefficients in QCD perturbation series whose high-order behaviour is governed by a renormalon. We also prove that this convergence is accelerated if the perturbative series is Borel transformed. We apply Padé approximants and Borel transforms to the known perturbative coefficients for the Bjorken sum rule. The Padé approximants reduce considerably the renormalization-scale dependence of the perturbative correction to the Bjorken sum rule. We argue that the known perturbative series is already dominated by an infra-red renormalon, whose residue we extract and compare with QCD sum-rule estimates of higher-twist effects. We use the experimental data on the Bjorken sum rule to extract αs(MZ2) = 0.116−0.006+0.004, including theoretical errors due to the finite order of available perturbative QCD calculations, renormalization-scale dependence and higher-twist effects.