Prompt photon processes

The quantity $x_\gamma$ is defined as the fraction of the photon energy taking part in the hard QCD subprocess, and is a powerful tool in characterising high energy photoproduction processes [13,14]. For direct processes its value is by definition unity. Experimentally, its distribution is expected to differ markedly between dijet processes and prompt photon processes. An ``observed" value of $x_\gamma$ may be evaluated, using the definitions of [13] and [6], as $\sum_{jets}(E-p_Z)/\sum_{event}(E-p_Z)$, summing over particles (or calorimeter cells) in the jets (or the jet plus the photon) and in the whole event, for dijet events or prompt photon events as appropriate.

 

Figure: Distributions of $x_\gamma$ [6] for (a) dijet events, (b) prompt photon events as observed in an idealised HERA detector. Solid curve = direct contribution; dashed curve = resolved; dots = total. (c) Preliminary ZEUS results [15]: dashes = resolved + direct, dots = resolved.

Figure 2 shows the difference between the distributions of the ``observed" value of $x_\gamma$ between dijet events and events with a prompt photon and an accompanying high-E_T jet, subject to the typical kinematic constraints of a HERA experiment. High E_T jets have been reconstructed (E_T > 5 GeV) from the four-vectors of HERWIG simulations of direct and resolved events, with jets and prompt photons (also E_T > 5 GeV) accepted in the pseudorapidity range $-1.5 \le \eta \le 1.7$.First experimental results from ZEUS have been reported at the Rome DIS 96 Workshop [15]. In round figures, an integrated luminosity of 6 pb^-1 gives approximately 50 direct events with a prompt photon in the ZEUS barrel calorimeter and 30 resolved events, an accompanying jet also being observed.

The Direct Compton diagram gives a good measurement of the quark content of the proton, but an even more important aim is to measure the photon structure by means of the resolved events. With an integrated luminosity of 1000 pb^-1 we may thus expect to record around 5000 resolved events and 7500 direct events with a prompt photon and a jet. This should suffice to give a reasonable measurement of the photon quark density and distinguish between present models which currently give cross sections differing among themselves by typically 10-20% [6]. Such measurements would be noticeably degraded if the total integrated luminosity were less, say, by a factor of 4. Use of photons detected in the rear calorimeter would give a small improvement, but not in the numbers of resolved events. To improve the resolved statistics we would require a better understanding of photon detection in the forward calorimeter. This is technically difficult, however, and its viability needs further study.

An alternative possibility would be to use inclusive prompt photon distributions. These [4] appear to give a better sensitivity to the photon structure, different models varying by as much as 40%. The inclusive cross section for prompt photons within a pseudorapidity range of $\vert\eta\vert \le 1$ is given by [4] as 24-34 pb for the resolved contribution, with a similar figure for the direct. An integrated luminosity of 1000 pb^-1 will then give around 30k events in each category, the gain coming by an avoidance of the need to detect the jet. However to take advantage of these statistics, it would still be necessary to distinguish somehow between direct and resolved events. One way to do this is to model the photon remnant in terms of energy detected (outside any high E_T jets) in the rear regions of the detector, with the aim of reconstructing $x_\gamma$ approximately for each event. First investigations have been made [16] but the technique is clearly more difficult than if the outgoing jet is detected. For this reason, it seems necessary at present to aim for an $x_\gamma$ measurement, for which the highest possible luminosity is required.