Here are some numbers on the precision that could be reached on sin(2beta) with the STT included or not in the L2 trigger. Part of these results are included in the NSF talk I gave at the beginning of the year and that you can find in http://www-d0.fnal.gov/~stichel/talks/talk_nsf98.ps Especially, pages 12, 13, 18, and 19 show how I obtained the error on sin(2beta). For the J/psi->e+e- trigger, I also looked at Paul Grannis note #3506. The facts related to the trigger efficiency are then the following: J/psi -> mu+mu- (described in STT proposal) =============== -1) With a single muon trigger corresponding to p_T > 4 GeV and |eta| < 2, the standard trigger ntuples give a trigger rate of 40 Hz at L1. I don't apply any additional cut at L2. -2) The corresponding efficiency for the signal is 24%. -3) Given that the trigger rate is based on a low statistics of event, this rate could be subject to a large underestimation. The STT would then be the best way to survive: Trigger configuration | rate reduction at L2 | signal eff. (%) -------------------------|-----------------------|----------------- p_T > 6 GeV | 2.9 | 10.8% ask for 2 CFT tracks | 1.8 | 13.5% ask STT with S_b > 2.5 | 2.9 | 17.3% ask STT with S_b > 5.0 | 10. | 14.3% -4) For a dimuon trigger, it would be impossible to reduce the threshold to 2 GeV on both muons without the STT: Trigger configuration | L2 rate | eff. ------------------------------|------------|----------- * (2 GeV, 4 GeV) - Loose tag | 110 Hz | 15% * (2 GeV, 4 GeV) - Medium tag | 1 Hz | 10% * (2 GeV, 2 GeV) - Tight tag | 270 Hz | 20% * (2 GeV, 2 GeV) + STT | 8 Hz | 15% The (2 GeV, 4 GeV) is not very useful as the triggered sample is already contained in the pt>4 GeV single muon trigger. -5) With the STT, I can combine the single muon and dimuon triggers to obtain a total efficiency of 32% and a trigger rate of 45 Hz. By playing with the STT criterium, it would be possible to reduce this rate by a factor 4 while preserving a signal efficiency of 23%. Without the STT, the best we can hope is probably 24% with the pt> 4 GeV single muon trigger. If the rate of this trigger becomes larger than expected, the only thing we could do is to increase the pt threshold to 6 GeV, resulting in an efficiency decrease to 11%. Part of that efficiency drop would be recovered by the (2 GeV, 4 GeV) dimuon trigger and we may end up with an efficiency around 15% to 20%. -6) So, the bottom line is: * With STT ==> eff(J/psi->mu+mu-) = 32% * W/out STT ==> eff(J/psi->mu+mu-) = 24% J/psi -> e+e- =============== Meenakshi has calculated the efficiency of the low pt central electron trigger included by Jerry and Bornali in the analysis package. She used events containing 2 electrons with pt > 1.5 GeV and |eta| < 2. These acceptance cuts are identical to those used for mu+mu- such that the efficiencies can directly be compared. - Level 1: * one electron with electromagnetic energy in the calorimeter CALEM > 3 GeV and eta < 1.2 * one CFT track with pt > 3 GeV matching FPS + CALO information - Level 2: * additional matching with CFT tracks Using this trigger, Meenakshi obtains a 10% efficiency for the signal, without making use of the STT. If the CFT matching is replaced by the STT at L2, the efficiency drops to 5.7%. According to Bornali's analysis, the trigger rate for the central electron trigger with a 3 GeV threshold is 150/75 Hz with CFT/STT matching at L2 when using L = 1x 10^32. She also tried to reduce the threshold to 2.0 GeV but then the rates become very large. Here is a table showing the L1/L2 rates versus CALEM threshold. In the last column, I also add the numbers obtained by Paul using a very similar trigger (as I understand it): | L1 (Bornali)| L2 (same)| L1 (Grannis) -------------------|-------------|----------|------------- CALEM = 2.0 GEV | 8.0 kHz | 500 Hz | 7.8 kHz CALEM = 2.5 GEV | --- | --- | 2.5 KHZ CALEM = 3.0 GEV | 1.0 kHz | 150 Hz | --- Based on these numbers, I doubt we could reduce the CALEM threshold below 3 GeV unless we apply very sophisticated criteria like Paul did. Thus, the best efficiency we can hope for is 10%. Meenakshi will try to improve these numbers before her departure on the 26th December and I could update the sin(2beta) precision with her new numbers. sin(2beta) precision ==================== Here is an estimation of the number of fully reconstructed events in 2 fb-1 corresponding to the trigger efficiencies obtained before. (For comparison, CDF quotes 10,000 J/psi->mu+mu- and 5,000 J/psi->e+e- events for 2 fb-1) J/psi ==> mu+mu- with STT : Nreco = 8500 J/psi ==> mu+mu- w/out STT : Nreco = 6400 J/psi ==> e+e- w/out STT : Nreco = 2600 Then, using a flavor tagging efficiency eff_tag * D_tag^2 = 5% (see NSF talk), I obtain the following sin(2beta) errors: * J/psi -> mu+mu-/e+e- without STT: error = 0.142 * J/psi -> mu+mu-/e+e- with STT: error = 0.128 This uncertainty is larger than CDF expectations (0.09 according to their baseline) but, given that sin(2beta) is constrained between 0.56 and 0.94 at 95% C.L. based on indirect measurements, that uncertainty would correspond to a direct measurement of CP violation at 4 sigma at least. So, as in the other channels, the STT doesn't bring a dramatic improvement to the physics expectations for Run II. However, it is a powerful tool to reduce the rates of the low pt triggers so that they would fit into the limited allocated bandwith. It would a pity to disregard the potential of the STT for B physics in Run II. The main problem is probably to put the right spin on it in order to convince the PAC panel.