1. GlueX Luminosity Limits Richard Jones, University of Connecticut GlueX Collaboration Meeting, Newport News, May 8-10, 2008 1.Design luminosity 2.Physics possibilities at higher luminosities 3.Limiting factors in current design

2. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 2 Design Luminosity Goal – produce sufficient samples of exclusive reations to be systematics-limited (maximum sensitivity to weak exotic waves) in amplitude analysis for key channels. Translation – when that occurs depends on the final state, ie. specific backgrounds, PID demands, … Rule of thumb: 10 7 events is sufficient for a decent PWA Consider a hypothetical case:  = 50 nb BR = 30%  = 25%

3. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 3 Design Luminosity L =.071 x 30 cm x 6.0 10 23 x 10 -33 x I beam g cm 3 1g1g cm 2 nb = 1.3 10 -9 nb -1 x I beam At I beam = 10 7  /s, it would take 57 khr (~ 20 years) to collect these statistics. At I beam = 10 8  /s, it would take 6 khr (~ 2 years) to collect these statistics. Result: 10 8  /s is sufficient to complete the hybrid spectroscopy program. But is it optimal ?

4. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 4 Design Luminosity tagger figure of merit Define: tagger figure of merit Factor that rescales the amount of run time needed to reach a given level of statistical error in a tagged histogram. Reference for FOM shown is the GlueX tagged beam under nominal conditions at 9 GeV, but with no mistags. 1.Assumes detector identifies correct beam bucket 100% of the time. 2.Shows some gains up to 3 10 8 Hz. 3.Gains are only about 25% for factor 3 in backgrounds.

5. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 5 Design Luminosity For 25% more statistics, what do we lose? x3 radiation damage in FCal x3 accidentals in the TOF and Start x3 pileup in the FDC, extra tracks, etc. x3 in channel count in the microscope – $$$ x3 in radiator thickness – reduced polarization

6. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 6 Design Luminosity If this argument was not made before, what was the basis of the design goal of 10 8  /s? the intuitive criterion of 50% accidental tags evidence from Monte Carlo simulation that detector backgrounds are going to preclude higher luminosities 1.FCal radiation damage – already an issue at 10 8 2.TOF occupancy – within a factor of 3-5 of ceiling 3.FDC pileup and extra tracks – within factor of 3-5

7. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 7 Physics at Higher Luminosity What physics might make this interesting? inverse DVCS – looks feasible threshold J/  – statistically difficult Cascade baryons – needs kaon PID

8. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 8 Beam Limiting Factors Tagging near the end-point no polarization no significant collimation amorphous radiator – factor 100 more luminosity available (if untagged) current tagger design has full coverage over 9-11.4 GeV, designed to run up to 50 MHz / GeV. at 50 MHz / GeV in end-point region, detector backgrounds are comparable to nominal conditions with polarized beam at 10 8  /s.

9. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 9 Detector Limiting Factors FCal radiation damage Inner blocks could be shielded, giving up low- angle  acceptance, ok for some physics. FTOF occupancy ditto. FDC pile-up – will be ultimate limiting factor. essential for just about any physics no effective means to shield them

10. GlueX Collaboration Meeting, Newport News, May 8-10, 2008 10 Conclusions Design luminosity is optimized for carrying out the hybrid spectroscopy program. Nominal high-intensity running conditions are consistent with tagging at 50 MHz / GeV at the end-point. The photon source will produce as much intensity as the experiment can handle in any scenario. With a dedicated end-point tagger, one can tag effectively up to 250 MHz, provided the detector can trigger. With 250 MHz on 11 < E  < 12 GeV, detector background would be x5 nominal, probably an upper limit. FDC pile-up will be the limiting factor – how to estimate it?