Measurement of the Diffractive Structure Function F_2^D(4) at HERA

TL;DR

This work reports the first measurement of diffractive DIS with the final-state proton detected in the ZEUS LPS, enabling direct access to the momentum transfer $t$ and extending the $M_X$ range up to about 35 GeV. The authors extract the diffractive structure functions $F_2^{D(4)}(β,Q^2,x_{I\!P},t)$ in a restricted kinematic region and obtain $F_2^{D(3)}$ by integrating over $t$, finding an $x_{I\!P}$-dependence consistent with $(1/x_{I\!P})^{a}$ with $a\approx 1.0$, across $eta$ bins. The measured $t$-distribution yields a slope $b\approx 7.2\,\text{GeV}^{-2}$, and the data require both hard and soft pomeron components to describe the $eta$-dependence, with subleading reggeon contributions possible at higher $x_{I\!P}$. Overall, the results provide a more precise, background-controlled view of diffraction in DIS and reinforce Regge-based interpretations of diffractive DIS at HERA.

Abstract

This paper presents the first analysis of diffractive photon dissociation events in deep inelastic positron-proton scattering at HERA in which the proton in the final state is detected and its momentum measured. The events are selected by requiring a scattered proton in the ZEUS leading proton spectrometer (LPS) with $\xl>0.97$, where $\xl$ is the fraction of the incoming proton beam momentum carried by the scattered proton. The use of the LPS significantly reduces the contamination from events with diffractive dissociation of the proton into low mass states and allows a direct measurement of $t$, the square of the four-momentum exchanged at the proton vertex. The dependence of the cross section on $t$ is measured in the interval $0.073<|t|<0.4$~$\gevtwo$ and is found to be described by an exponential shape with the slope parameter $b=\tslopeerr$. The diffractive structure function $\ftwodfour$ is presented as a function of $\xpom \simeq 1-\xl$ and $β$, the momentum fraction of the struck quark with respect to $\xpom$, and averaged over the $t$ interval $0.073<|t|<\ftwodfourtmax$~$\gevtwo$ and the photon virtuality range $5<Q^2<20~\gevtwo$. In the kinematic range $4 \times 10^{-4} < \xpom < 0.03$ and $0.015<β<0.5$, the $\xpom$ dependence of $\ftwodfour$ is fitted with a form $\xpoma$, yielding $a= \ftwodfouraerr$. Upon integration over $t$, the structure function $\ftwod$ is determined in a kinematic range extending to higher $\xpom$ and lower $β$ compared to our previous analysis; the results are discussed within the framework of Regge theory.

Figures (11)

• Figure 1: Geometric LPS acceptance for tracks with $\hbox{$x_L$}=1$ in the $p_X^{LPS},p_Y^{LPS}$ plane. The lines indicate the regions where the acceptance is greater than $5\%,50\%$ and $95\%$.
• Figure 2: Observed event distributions as a function of (a) $Q^2$, (b) $x$, (c) $W$, and (d) $y_{JB}$ of the reconstructed data (dots) compared to the Monte Carlo model (solid line). The Monte Carlo is the weighted sum of the RAPGAP and $\rho^0$ samples. The errors, shown as vertical bars, are statistical only.
• Figure 3: Observed event distributions as a function of (a) $\eta_{max}$ , (b) $M_X$, (c) $\ln M_X^2$ for $50< W< 120~\hbox{$\rm GeV$}$, (d) $\ln M_X^2$ for $120 < W <270 ~\hbox{$\rm GeV$}$, (e) $\hbox{$x_{_{I\space P}}$}$ and (f) $\beta$ of the reconstructed data (dots) compared to the Monte Carlo (solid line). The Monte Carlo is the weighted sum of the RAPGAP and $\rho^0$ samples. The errors are statistical only.
• Figure 4: Observed event distributions as a function of (a) $\hbox{$x_L$}$ , (b) $|t|$, (c) $p_X^{LPS}$ and (d) $p_Y^{LPS}$ of the reconstructed data (dots) compared to the Monte Carlo (solid line). The Monte Carlo is the weighted sum of the RAPGAP and $\rho^0$ samples. The errors are statistical only.
• Figure 5: