Hall D Internal Note

Photoproduction with Quasi-Real Photons

Richard Jones
July 15, 1998

At the recent CMU workshop we discussed briefly the possibility of doing Eight+ physics with low-Q electron scattering in the place of a real photon beam. In my talk at CMU I showed a comparison between the virtual and real photon beams in terms of flux and polarization. There are several advantages of virtual over real photons. The primary concern that was raised in our discussions was with the additional backgrounds, particularly Moller scattering, that are present with electron beams.

I have now done a study of these backgrounds, and have come to the conclusion that a significant slice of recoil electron kinematics exists in the far-forward region where backgrounds are modest and a large fraction of the total hadronic photoproduction cross section can be tagged. The forward region at high energy can be separated into three regions: very low pT (electron mass scale) where bremsstrahlung dominates, modest pT (diffractive scale ~300MeV), and low pT (in between the two at ~30MeV). The angular distribution of recoil electrons from hadronic photoproduction processes goes like 1/(pT) whereas the cross section for background processes (bremsstrahlung, Mott scattering, Moller scattering) is dropping like 1/(pT). What happens at higher CEBAF energies is that this crossover where background drops below the signal rate gets pushed to small angles of order 1 where it becomes feasible to build a dedicated tagging spectrometer that does not interfere too much with coverage of the hadronic detector.

CLAS has become interested in this possibility, where they would capture the electrons which exit through the small hole in the forward region. The study that I presented at the Hall B session at the recent JLab workshop showed that they could tag approximately 15% of the total hadronic photoproduction rate in the endpoint region just covering the region out to 2 with 6GeV electrons. At 6GeV the backgrounds are modest in this angular region. At 12GeV they are still better. It occurs to me that this would be an excellent alternative to bremsstrahlung beams for Eight+. The advantages for us, as I see them, are as follows.

1. In addition to reading out the energy of the virtual photon, the plane polarization of the photon is also recorded by the tagger. All of the photons in this virtual beam are polarized. The polarization is the same as the theoretical maximum for coherent bremsstrahlung at extreme collimation, i.e. plane polarization of all photons in the virtual beam is equal to the polarization of the coherent component from crystal bremsstrahlung at its maximum right at the edge of the peak. This means that effectively the plane polarization can be an order of magnitude higher for virtual than for coherent beams, if you average over the region from 80% to endpoint. Note that this does NOT diminish the strength of our argument for 12GeV beams. The polarization is still going to zero at the endpoint. But in this case the average polarization one can actually achieve in the endpoint region goes to zero linearly like (k-E) instead of quadradically like (k-E).
2. Only the photons that actually produce hadronic final states in the target produce counts in the tagger. Compare that with the design with real photons where only 1/10000 hits in the tagger correspond to actual hadronic events in the target. This effectively sets aside the present limitations from tagging which required us to limit fluxes to 3e8/s, roughly comparable to production rates at a hadronic beam facility.
3. Target can be as small. With a real tagged beam your target has to be extended in order to get enough hadronic rate, because the photon flux is limited by the tagging rate. With a virtual beam the scattered electron flux in the tagger is proportional to the hadronic rate, so you can make the target thin and compensate with beam current, with no change to the accidentals situation.

Number 3 is particularly interesting from the point of view of TOF, fast online track reconstruction, and general compactness of the barrel components. It also makes building a target much simpler and makes a polarized target thinkable.

 
 
 
 
 
 
 
 
I have included copies of the transparencies from my Hall B talk (presented at the Workshop) of how this would work with CLAS. In this study I compared backgrounds from the bremsstrahlung tail, Mott scattering and Moller scattering in the target, and found that the backgrounds are substantial at 4GeV, getting better at 6GeV, and really looking very tame at 12GeV. In talking with others around here (Hall B), I am beginning to suspect that some years ago people looked into this possibility for experiments at lower energies and decided that the backgrounds were too high, and just forgot about it. Things get better at higher energies. I am suggesting that perhaps we are overlooking something by assuming that we should carry on at 12GeV the same way that we did at 6GeV because that was the best way to go at 6GeV. Clearly HERA has proved that photoproduction physics can be done with high-energy electrons. What I am finding out is that it seems to work quite well at 12GeV as well. Before we invest anything in replicating the photon beamline from Hall B, why don't we look into this possibility?

It also important implications for the larger detector design. In particular the barrel tracking has to avoid the Moller cone. Things are much better with a solenoid than with a field like CLAS where the target sits at a null in the magnetic field, but it has to be looked at. Beyond 10 the Mollers carry pT of 6MeV or less, so they will not get far from the axis, but there are lots of them. At lower angles the transverse momenta are somewhat higher but the intensities are lower. This would probably be the limiting factor in how high a flux we could run at. Experience with CLAS shows that they could run with electron beam currents high enough to gain signficantly in photoproduction rate at higher energies over bremsstrahlung tagging. With our solenoidal field, things can only be better. One of the soft spots in our proposal that has emerged so far has been the plausibility of tagging at 300MHz. At this point it is too early to draw conclusions, but there are hints that with electrons this limit can be overcome without stretching things.

Colour copies of these transparencies can be downloaded from the web site at http://zeus.phys.uconn.edu/hallB/virtual-6-98.