Glueballs and topology with O(a)–improved lattice QCD.

Glueballs and topology with O(a)–improved lattice QCD. We present evidence for unquenching effects in N f = 2, 16 3 32 ensembles by comparing with 'equivalent' quenched data atˆr0 ≃ 5.0. A (small) VEV for torelons signals (weak) string breaking. A 15 − 20% reduction in the scalar glueball mass relative to quenched is argued to be (in part at least) a discretisation effect. We find a chiral suppression of the topological susceptibility consistent with expectations, and agreement between fermionic and gluonic methods for measuring the topological charge. To study the effects of vacuum polarisation in QCD, we should vary the sea quark mass, m q , having first removed discretisation effects. With confident continuum extrapolations of many lattice quantities unavailable, however, we must instead try to 'freeze' lattice artefacts. UKQCD does this by attempting to fix the lattice Som-mer scale, ˆ r 0 ≃ 5, as m q is changed. Variations due to residual changes inˆr 0 are further reduced by the the leading order discretisation errors for the O(a)–improved action being quadratic, rather than linear, in the lattice spacing, a. An 'equiva-lent' value ofˆr 0 is found in the quenched, 'gluody-namical' theory at coupling β Wil = 5.93. Details of scales, parameters and matching can be found in [1]. We use as a measure of m q the square of the pseudoscalar 'pion' mass. Circumflexes denote dimensionless lattice quantities. The failure of experiments to detect unam-biguous glueball states is believed to be a consequence of mixing between the light glueballs and q ¯ q states. A major goal of lattice QCD is to predict the mixing matrix elements, but the uncertainties in attempts so far [2,3] mean, however, that phenomenological attempts to describe the content (gluonic or q ¯ q) of the scalar sector glue-ball candidates lead to widely differing results [4]. As a preliminary to a full mixing analysis [5], we study the light glueballs and winding flux tubes on UKQCD dynamical configurations and find, in common with [6], that accurate mass estimates are possible on moderately sized ensembles (4000 − 8000 HMC trajectories). Full results can be found in [7], along with a discussion of the applicability of the term 'glueball' when mixing is induced by vacuum polarisation. On a spatial torus, Polyakov loops (PL) couple to colour flux tubes that close upon themselves through the periodic boundary with unit winding number. In the confining phase …