The GW Experimental Nuclear Physics Group has been a major contributor to the instrumentation in Hall B from the outset. We have designed, built, installed, tested, and maintained (with others) the focal-plane detector array of 61 T-counters and 384 E-counters for the Tagged Photon Facility. We also are responsible for the design, acquisition, installation, and testing (with others) of the goniometer for the coherent bremsstrahlung beam of linearly polarized photons. Now we are engaged in developing the polarimeter with which to measure the polarization of that beam so that one does not have to rely on calculations to determine the beam polarization. Dr. Feldman has been leading this effort.

(Feldman, O'Rielly, Welch)

The GW group has been a major participant in the development of a polarized photon facility in Hall B. The polarized photons are produced by the process of coherent bremsstrahlung from a diamond crystal. We have designed, procured, installed and tested the goniometer which sets the angular position of the diamond. A parallel effort has gone into the design and testing of a polarimeter to measure directly the polarization of the photon beam, which will complement theoretical calculations of the beam polarization.

The polarimeter is based on triplet photoproduction. In this process, the incident photon produces an electron-positron pair in the Coulomb field of an atomic electron, and the azimuthal asymmetry of the recoil electron is directly related to the polarization of the incident photon [Bol72, Bol74]. This method has several advantages – the reaction has a sizable analyzing power, the recoil electron is scattered at large angles, so that it is relatively easy to detect away from the beam, and the associated e⁺e- pair permits a multiple-coincidence condition to be used as a trigger, thus providing strong background rejection. This method was presented at a Polarized Photon Polarimetry Workshop held at Jefferson Lab [Fel98], and the technique was approved for further study.

The prototype polarimeter was obtained from the Saskatchewan Accelerator Laboratory (SAL). Using this device, an initial one-week test run of the triplet photoproduction technique was performed at LEGS using nearly 100% polarized (tagged) photons of energy E_g = 220-330 MeV. In the test run, a thin (2 mm) plastic scintillator paddle was used as an active target, and four recoil paddles (up, down, left, right) were used to detect the azimuthal distribution of the recoil electrons. Two overlapping plastic scintillator paddles downstream from the target were used to detect the e⁺e- pair. A coincidence between the target scintillator, the e⁺e- pair counters and one of the recoil paddles provided a rather stringent requirement for the experimental trigger.

The results of the LEGS test run are shown in Fig. 1. The asymmetry is determined by summing the up and down recoil paddles to get a yield in the vertical plane (N_v) and summing the left and right paddles to get a yield in the horizontal plane (N_h) – then the asymmetry is calculated as (N_v- N_h)/(N_v+N_h). The plane of polarization of the LEGS photon beam can be varied, so we obtained asymmetry data at 8 distinct polarization angles. The cycle of angles was then repeated to check the reproducibility of the measurements. The data analysis revealed a

clear cos(2f) azimuthal angular dependence (see Fig. 1), but the asymmetry (2.66%) measured in the test run was smaller than the predicted asymmetry from a theoretical calculation of the triplet photoproduction process. The results of this experiment have been presented in preliminary form at an APS meeting [Ori00]. The reduced experimental asymmetry observed in the data had been tentatively attributed to multiple scattering of the low-energy recoil electrons.

A naive Monte-Carlo simulation of the triplet polarimeter detector system was developed in order to model the effect of multiple scattering and determine the impact of such an effect. This simulation took into account scattering effects along the entire path length of the low-energy recoil electrons, as they exit the target, traverse a certain distance through the air and enter the recoil detectors. The results of the simulation indicated that the idealized asymmetry expected for our geometry (~12%) is significantly reduced by multiple scattering, down to about 5.6%. Despite this substantial reduction, it was still not in agreement with our measured asymmetry of 2.7%. More work was required to understand the origin of the discrepancy between the measurement and the simulation.

In Fall 2000, we established a new collaboration with a group of Ukrainian theorists (Yu. Peresunko and I. Shapoval) who are experts in the QED processes of pair and triplet photoproduction. They have developed a new Monte-Carlo simulation of the triplet and pair production processes based on GEANT, which includes other physics processes that could contribute to a dilution of the idealized triplet recoil electron asymmetry. In addition to the "obvious" effect of multiple-scattering, a major contributor to the diluted asymmetry is the effect of delta electrons. In this case, a pair is created at the reaction vertex, and then either the positron or the electron strikes another electron during its trajectory and "knocks" it into the recoil detectors. This two-step process will be detected as a three-particle final state (the pair plus the "knock-on" electron) and unfortunately is indistinguishable experimentally from a true triplet (one-step) reaction. The inclusion of the delta electrons reduces the asymmetry in the new Monte Carlo simulation to about 3-4%, which is in reasonable agreement with the measured value of 2.7% from the LEGS test run. Thus, at this stage, it is believed that the low measured asymmetry from the LEGS run is understood. A publication describing the results of the LEGS test run and the comparison with the Monte Carlo simulation is being prepared [Per03].

Much of the multiple scattering and delta-ray production occurs in the target, which was a 2-mm piece of plastic scintillator. We have obtained a thinner plastic scintillator (0.5 mm) and the GEANT simulation predicts an improvement of the asymmetry to the 7-9% level. The simulation also provides insight into possible experimental cuts that might help further enhance the asymmetry -- for example, selection of the recoils by energy. To this end, we have fabricated new recoil detectors that are 5 cm thick, sufficient to stop recoil electrons of energies up to 10 MeV. This will enable a total-energy determination of the recoils, such that analysis cuts can be made on the full deposited energy. A graduate student (K. Conor Welch) is working on the recoil detectors, as well as analyzing results produced by the Monte-Carlo simulation to determine the optimal set of analysis cuts.

The original polarimeter stand from SAL has been recast in a more portable form at GW to facilitate convenient transport and setup at other laboratories. This modification was accomplished by the machine shop at GW. In addition, the recoil paddles have been mounted on a rotating plate, such that the azimuthal angle of the paddles can be changed relative to the polarization direction of the photon beam. This flexibility is necessary to check systematics in the event that the photon polarization plane cannot be varied.

Further GEANT simulations are being conducted to determine optimal conditions for obtaining the asymmetry. A future test experiment will be performed at LEGS (E_g ~ 250 MeV) or at GRAAL (E_g ~ 1.5 GeV), so both energies are being checked in the simulation. A thinner active target (0.5 mm) would be included in this upcoming test run. It is hoped that this improvement, along with the thicker stopping detectors for the recoil electrons, will enable a larger asymmetry to be measured. The Monte-Carlo simulations with the thinner target and the thick recoil paddles are currently being analyzed.

This site is best viewed in Internet Explorer 6.0 or higher. Send mail to Web Master with questions or comments about this web site.
Copyright © 2004 GW Experimental Nuclear Physics Research Group
Last modified: 03/31/05