MEASUREMENTS OF αs FROM HADRONIC EVENT SHAPES IN e + e − ANNIHILATION

Most experimental measurements of αs are limited in precision by theoretical errors arising from missing higher order terms and uncertainties in the hadronisation corrections which are normally derived from the phenomenological Monte Carlo models. Recently, improved perturbative QCD predictions for the event shape observables, C-parameter and the Jet Broadenings (BW , BT ), have become available to add to those already produced for Thrust (T) and Heavy Jet Mass (M2H) . Leading non-perturbative (1/Q, where Q = √ s) power corrections have also been calculated for these observables. They are based on an effective strong coupling, α0(μI), at an infra-red matching scale, μI (usually set to 2 GeV). This coupling is expected to be approximately universal but must be derived from experiment. Once evaluated, power corrected distributions can be compared directly with experiment leaving the Monte Carlo models for detector corrections only. Recently, a 2-loop analysis of these 1/Q power corrections has been performed 4 for Thrust rescaling the original prediction by a so-called Milan factor (M) which has been shown to apply equally to the other aforementioned observables . The power corrections basically shift the perturbative predictions for the differential spectra of each observable linearly in α0 although additional logarithmic terms are thought to be required for the jet broadenings . In a contribution to this conference, ALEPH fit the power corrected perturbative predictions in the 3-jet regions of the 1-T, C, M2H and BW distributions to the measurements at 5 values of √ s from 91.2 to 183 GeV. The perturbative predictions areO(αs) with ln(R) (or R) matched NLL terms. The perturbative renormalisation scale parameter, xμ(=μ/ √ s), is set to 1. They extract simultaneously values for α0(μI) and αs(MZ) from each observable. The fits are reasonably satisfactory except to the BW distributions where the results are far from those obtained for the other observables with either matching scheme. The bad quality of these fits disfavours the concept of a logarithmically enhanced shift . DELPHI 7 Table 1: Values of αs(MZ ) and α0(μI ) as obtained from the fits. † Results from C and (1-T) are consistent; ‡ values are compatible only at 30% level.