Probing the sea quark content of the proton with one-particle-inclusive processes

We investigate the feasibility of constraining parton distribution functions in the proton through a comparison with data on semi-inclusive deep-inelastic lepton-nucleon scattering. Specifically, we reweight replicas of these distributions according to how well they reproduce recent, very precise charged kaon multiplicity measurements and analyze how this procedure optimizes the determination of the sea quark densities and improves their uncertainties. The results can help to shed new light on the long standing question on the size of the flavor and charge symmetry breaking among quarks of radiative origin. An iterative method is proposed and adopted to account for the inevitable correlation with what is assumed about the parton-to-hadron fragmentation functions in the reweighting procedure. It is shown how the fragmentation functions can be optimized simultaneously in each step of the iteration. As a first case study, we implement this method to analyze kaon production data.

Figure 1

NLO results for charged kaon multiplicities in SIDIS computed with the DSS 17 set of FFs [23] and the MMHT 14 set of PDFs [34] (solid lines) for both proton and deuteron targets compared to data from the Hermes collaboration [13]. The light shaded (grey) bands correspond to the uncertainty estimates for the FFs, and the dark shaded (blue) bands represent the uncertainties of the MMHT 14 set of PDFs, both at the 90% confidence level (C.L.). The dashed lines are the NLO results obtained with a reweighted set of PDFs and a refined set of FFs, see text.

Figure 2

Same as in Fig. 1 but now for the positively charged kaon multiplicities on a deuteron target obtained by Compass [18]; see text.

Figure 3

Same as in Fig. 2 but for negatively charged kaons.

Figure 4

Distribution of ${\chi }^2$ per data point for the combined kaon multiplicity data sets of Hermes and Compass before and after reweighting, upper and lower left panel, respectively, the weights $w_k$ (upper right panel) and of $\mathcal{P}(\alpha )$ (lower right panel). Each distribution is computed neglecting the error coming from the FFs (dark blue), including the 68% C.L. estimates of Ref. [23] (light blue), and an intermediate C.L. value adjusted such that $\mathcal{P}(\alpha )$ peaks at one (green); see text.

Figure 5

Reweighting of the strange quark distribution (upper left panel) and for the PDF combinations sensitive to charge (upper right panel) and flavor (lower panels) symmetry breaking using the DSS 17 set of kaon FFs that is based on the MMHT 14 set of PDFs; see text. The dashed light blue and black lines and the hatched areas represent the results of one iteration of the reweighting procedure and the corresponding uncertainty bands, respectively; see text. All results are shown at a scale of $Q^2=5{\mathrm{GeV}}^2$.

Figure 6

Same as Fig. 5 but now for the nonstrange light quark PDF flavors, again normalized to the corresponding mean value of the NNPDF 3.0 replicas.

Figure 7

Same as Fig. 5 but now starting from a variant of the DSS 17 set of kaon FFs, extracted using the NNPDF 3.0 PDF set as input; see text.

Figure 8

Same as Fig. 6 but now starting from a variant of the DSS 17 set of kaon FFs, extracted using NNPDF 3.0 PDF set as input; see text.

Figure 9

Reweighting of the strange quark distribution as in Figs. 5 and 7 (upper left panels) but now using CT14 Monte Carlo PDF replicas [37] as the starting point; see text. All results are normalized to the mean strangeness density of the NNPDF 3.0 replicas.

Figure 10

Parton-to-kaon FFs at a scale $Q^2=10{\mathrm{GeV}}^2$ as obtained in the original DSS 17 global analysis and based on fits adopting the CT14 and NNPDF 3.0 set of PDFs. The dashed line represents the result of one iteration in the reweighting procedure of the MMHT 14 based DSS 17 set of FFs. The shaded bands illustrate the 68 and 90% uncertainty bands of the global analysis.