Nucleon Helicity Parton Distribution Function in the Continuum Limit with Self-Renormalization

TL;DR

This paper reports the first lattice QCD calculation of the nucleon helicity PDF using LaMET with self-renormalization, continuum extrapolation, and systematic error reduction, providing insights into nucleon spin structure.

Contribution

It introduces a novel lattice calculation of the nucleon helicity PDF employing self-renormalization and systematic error control techniques, advancing the precision of nucleon structure studies.

Findings

Systematic errors greatly reduced by RGR and LRR methods.

Final PDFs show a small positive antiquark helicity distribution.

Results exhibit some tension with global experimental fits.

Abstract

We present the first lattice calculation of the nucleon isovector helicity parton distribution function (PDF) in the framework of large-momentum effective theory (LaMET) that uses the hybrid scheme with self-renormalization. We use ensembles generated by the MILC collaboration at lattice spacings $a=\{0.1207,0.0888,0.0582\}$ fm, with $N_f=2+1+1$ flavors of highly improved staggered quarks at sea pion mass of $M_{\pi}\approx 315$ MeV. We use clover-improved action for our valence quarks with nucleon boost momentum $P_z\approx 1.75$ GeV and high-statistics measurements for the LaMET matrix elements. We perform an extrapolation to the continuum limit and improve the handling of systematic errors using renormalization-group resummation (RGR) and leading-renormalon resummation (LRR). Our final nucleon helicity PDF is renormalized in the $\overline{\text{MS}}$ scheme at energy scale $\mu=2.0$ GeV. We compare our results with and without the two systematic improvements of RGR and LRR at each lattice spacing as well as the continuum limit, and we see that the application of RGR and LRR greatly reduces the systematic errors across the whole $x$ range. Our continuum results with both RGR and LRR show a small positive antiquark region for the nucleon helicity PDF as well as a change of as much as a factor of two in the central values compared to results with neither RGR or LRR. By contrast, the application of RGR and LRR only changes the central values by about 5\% in the quark region. We compare our lattice results with the global fits by the JAM, NNPDF and DSSV collaborations, and we observe some tension between our results.