Strangeness in the proton from W+charm production and SIDIS data

TL;DR

This work addresses how much strangeness is present in the proton by performing a global QCD analysis that simultaneously fits unpolarized PDFs and FFs, incorporating new $W$+$c$ production data from the LHC and SIDIS hadron production data from COMPASS. The authors employ collinear factorization at fixed order, a flexible PDF/FF parameterization, and a Bayesian Monte Carlo framework to merge DIS, DY, jets, SIDIS, and SIA data, achieving an overall fit quality of $\chi^2_{\text{red}}\approx 0.99$ across ~5853 data points. They find that including both SIDIS and $W$+$c$ data reduces the uncertainty on the strange PDF $s^+$ over $0.01 \lesssim x \lesssim 0.3$ and yields a suppressed strange-to-nonstrange ratio $R_s$ that decreases with increasing $x$, with $R_s$ values of $0.72^{+0.52}_{-0.34}$ at $x=0.01$, $0.46^{+0.30}_{-0.20}$ at $x=0.04$, and $0.32^{+0.23}_{-0.15}$ at $x=0.1$ at $Q^2=4$ GeV$^2$, while the strange asymmetry $s-\bar s$ remains compatible with zero. The results are broadly consistent with other global analyses but emphasize the complementary roles of SIDIS and $W$+$c$ data in constraining the strange sector; the study also outlines avenues for future improvements, including extended FF constraints, lattice QCD inputs, parity-violating DIS, and future facilities like an Electron-Ion Collider.

Abstract

We perform a global QCD analysis of unpolarized parton distribution functions (PDFs) in the proton, including new $W +$\,charm production data from $pp$ collisions at the LHC and semi-inclusive pion and kaon production data in lepton-nucleon deep-inelastic scattering, both of which have been suggested for constraining the strange quark PDF. Compared with a baseline global fit that does not include these datasets, the new analysis reduces the uncertainty on the strange quark distribution over the range $0.01 < x < 0.3$, and provides a consistent description of processes sensitive to strangeness in the proton. Including the new datasets, the ratio of strange to nonstrange sea quark distributions is $R_s = (s+\bar s)/(\bar u+\bar d) = \{0.72^{+0.52}_{-0.34},\, 0.46^{+0.30}_{-0.20},\, 0.32^{+0.23}_{-0.15}\}$ for $x = \{ 0.01, 0.04, 0.1 \}$ at $Q^2 = 4$~GeV$^2$. The data place more stringent constraints on the strange asymmetry $s-\bar s$, which is found to be consistent with zero in this range.

Figures (11)

• Figure 1: Kinematic coverage of datasets used in this analysis, with the scale $Q^2$ versus the Bjorken scaling variable $x_i=x_B$ for DIS/SIDIS, and $x_i=x_1$ or $x_2$ for Drell-Yan/weak boson/jet production data (upper panel), and the fragmentation variable $z_h = p \cdot p_h /p \cdot q$ for SIDIS and $z_h = 2 \, p_h \cdot q/Q^2$ for SIA data (lower panel).
• Figure 2: $\chi^2_{\rm res}$ and $Z$-score values of all datasets used in the current analysis, including for the baseline datasets (top panel), $W$+ charm data from CMS and ATLAS (middle panel), and SIDIS $\pi^\pm$, $K^\pm$, and $h^\pm$ production from COMPASS (bottom panel). The reduced $\chi^2$ for individual datasets and averages over datasets are indicated by the blue and red points, respectively, and an additional open circle shows the reduced $\chi^2$ and $Z$-score of the HERA data using for $Q^2 > 3.5$ GeV$^2$ cut.
• Figure 3: Comparison of CMS and ATLAS $W$+ charm data with the JAM24 global analysis as a function of lepton pseudorapidity, $|\eta_\ell|$, with 95% credible interval uncertainty bands. (Left): Sum (top) and ratio (middle) at 7 TeV CMS:2013wql, and sum (bottom) at 13 TeV CMS:2018dxg of $W^+ + \bar{c}$ and $W^- + c$ cross sections from CMS. (Right): Differential cross section of $W^- + c$ (top) and $W^+ + \bar{c}$ (middle) from ATLAS at 7 TeV ATLAS:2014jkm. The cuts on the transverse momentum of the final state lepton $p_T^\ell$ are indicated on the panels.
• Figure 4: SIDIS multiplicities $\dd M^h/\dd z_h$ from COMPASS COMPASS:2016xvm as a function of $z_h$ for $h=\pi^+$ (dashed lines) and $h=\pi^-$ (dotted lines) compared with the JAM24 fit result. The data and curves are scaled by a factor of $2^i$ ($i=0,\ldots, 4$) to more clearly separate and isolate the various $x_B$ and $y$ bins.
• Figure 5: As in Fig. \ref{['fig:sidis-p']}, but for $K^\pm$ data.