Studying Two-Photon Exchange in Deep Inelastic Scattering with the HERA Data

TL;DR

The paper investigates whether higher-order QED effects, notably two-photon exchange (TPE), bias inclusive DIS cross sections and the extraction of parton distributions. It analyzes the beam-charge ratio $R_{\pm}=\frac{\sigma_{e^+}}{\sigma_{e^-}}$ using the final combined H1+ZEUS inclusive NC DIS data at $\sqrt{s}=319$ GeV and SLAC measurements, across ranges in $Q^2$ and $\varepsilon$, with radiative corrections applied by HERACLES and no TPE correction applied to preserve sensitivity. The results show no significant deviation from unity for $Q^2\le 300~\mathrm{GeV}^2$, with a global average around $R_{\pm}=1.0019\pm0.0018$, indicating no large DIS TPE effect and a tension with elastic-scattering parameterizations. The findings support the robustness of inclusive DIS analyses against higher-order QED contamination within current uncertainties and suggest possible cancellations between TPE and quark-lepton radiation interference; they also highlight the need for theoretical work on quark-level radiation contributions and additional measurements to fully resolve TPE in DIS.

Abstract

Two-photon exchange (TPE) is one of the leading explanations for discrepancies in measurements of the proton electromagnetic form factors. It has been proposed that TPE could impact not only elastic scattering, but also the cross sections for both inclusive deep inelastic scattering (DIS) and semi-inclusive DIS, thereby affecting the interpretation of DIS structure functions in terms of parton distributions. It is expected that higher-order QED effects such as TPE should manifest as a deviation from unity in the ratio of \epp and \emp DIS cross sections. We use the existing inclusive $e^{\pm}p$ DIS data from HERA and SLAC to constrain higher-order QED effects on inclusive DIS.

Figures (6)

• Figure 1: Example TPE diagram contributing to inclusive DIS.
• Figure 2: HERA neutral current DIS data points in the kinematic plane of $x_{\mathrm{Bj}}$ and $Q^2$. This subset of the data corresponds to the bins where the $e^+p$ and $e^-p$ cross sections are presented with the same binning. Only data points with $Q^2\xspace\leq3000~\mathrm{GeV}^2\xspace$ are shown. The color scale shows the value of $\varepsilon$ at the kinematic point.
• Figure 3: Top: Linear fits $R_{\pm}(\varepsilon)=A(1-\varepsilon)+B$ to the data in different bins of $Q^2$. Bottom: Extracted fit parameters. The existence of a negative slope, $A$, and an intercept, $B$, less than one indicate the effect of electroweak interference.
• Figure 4: Linear fits of $R_{\pm}$ as a function of $\varepsilon$. Left: Fit to the HERA data only, in the range $60\leq Q^2\xspace\leq300\ \mathrm{GeV}^2\xspace$. Right: Global fit including the SLAC results from Refs. Rochester:1975fkFancher:1976ea. Note the different color scales between panels.
• Figure 5: Linear fits of $R_{\pm}$ as a function of $Q^2$. Left: HERA data only, points at the same $Q^2$ are offset horizontally for clarity. Right: Global fit including the SLAC Fancher:1976eaRochester:1975fk measurements.