PHOBOS TOF Results from AGS Test Beam '97
The time difference distribution from a localized source provides a measure of the intrinsic timing resolution of the TOF slats. This figure shows the time difference measured with the PMTs on either end of a horizontal slat in coincidence with a vertical slat behind it for a 1 GeV/c NEG beam. Thus the active area effectively becomes about 1 cm X 1 cm. The timing resolution for MIPs is better than 80 ps (1 sigma) even without slew correction.
The 9.5 cm separation between the horizontal and vertical TOF slats in the test beam setup provides a 300 ps flight path to measure TOF resolution independent of the start detector. This figure shows the time difference distribution measured between a horizontal and vertical TOF slat again for a 1 GeV/c NEG beam (note the centroid of this TOF distribution is 298 ps). The timing resolution recorded in this manner is approximately 100 ps (1 sigma) without slew correction, assuming the same resolution for each channel involved in this measurement.
This figure shows the time difference measured with the PMTs on either end of a horizontal slat in coincidence with adjacent vertical slats behind it for a 2 GeV/c NEG beam. The centroid of these distributions clearly shifts as a cut is placed on each different 1 cm wide vertical slat, providing information on the position where the beam particles hit the horizontal slat. The position resolution for MIPs is better than 1.5 cm (1 sigma) based on this time difference. The position resolution for MIPs is better than 3.0 cm (1 sigma) based on pulse heights (not shown here).
The cross talk for TOF signals with amplitudes expected at RHIC can be determined by examining the correlation of pulse heights in adjacent channels. An example is shown in the scatter plots L2E vs L1E and L3E vs L1E which show the pulseheights measured in adjacent and diagonal channels of the multi-anode PMT. respectively. Small signals are produced in adjacent channels for large pulse heights (this situation is less pronounced in the diagonal case). Note that the range of pulse heights in the TOF slats for beam particles is relatively narrow.
To determine where the cross talk comes in time, the time dependence of the pulse heights in a single channel (L1E vs L1T) is compared to that for pulse heights in an adjacent channel (L2E vs L1T). The slew effect is clearly evident in the L1E vs L1T scatter plot. Because the cross talk is associated with a narrow range of large pulse heights in an adjacent channel, the right panel shows this cross talk is cleanly separated in time from the actual signals. Higher discriminator thresholds during actual running conditions will prevent this cross talk from entering the data stream, so cross talk shouldn't present problems at RHIC using the R5900 multi-anode PMTs.
The best possibility for particle ID would be for data taken with a 500 MeV/c NEG beam. Positive beamline magnet settings could not be run, because the radiation safety interlocks in the cave tripped for settings below 2 GeV/c POS. The figures show the time-of-flight distributions measured with the PMTs on either end of a horizontal slat in coincidence with a vertical slat (TOF H2 cut on V2) and on either end of a vertical slat in coincidence with a horizontal slat (TOF V2 cut on H2). The time of flight over the 1.7 m path length to a horizontal slat for particles traveling nearly at the speed of light is 5.69 ns. The TOF to a vertical slat for these particles is 6.00 ns. The 92 ps time of flight difference between muons and pions can not be resolved with the present setup, but the 140 ps width of these distributions is significantly broader than expected for a single particle type.
Time-of-flight spectra obtained with a 5 GeV/c NEG beam. The figures show the time-of-flight distributions measured with the PMTs on either end of a horizontal slat in coincidence with a vertical slat (TOF H2 cut on V2) and on either end of a vertical slat in coincidence with a horizontal slat (TOF V2 cut on H2). A second peak in the trailing vertical TOF slats, which becomes progressively more pronounced as the beam momentum increases, is attributed to the production of secondaries in the leading horizontal TOF slats. Note that the width of the prompt peak is narrower than the corresponding cases for the 500 MeV/c beam, further supporting the claim that more than a single particle type is present at the lower momentum.