Summary of E835 meeting 4/27/00 Present: Menichetti, Pallavicini, Mussa, Obertino, Baldini, Negrini, Cibinetto, Lasio, Stancari(M), Stancari(G), Kasper, Gollwitzer, Graham, Seo, Pordes, Garzoglio Status: The accelerators are meant to start to come up sometime on Friday (tomorrow). The pbar source will be in controlled access that day. The first thing to do will be to re-establish beam on the pbar target; the transfer line from the Main Injector has been changed with B3 magnets replacing a lambertson. It may well take a shift or so to get the line working. A problem in the helium supply to AP-30 which provides liquid helium to the stochastic cooling has made it necessary to decontaminate the helium system in AP-30. This means that the stochastic cooling system is not presently cold enough for effective stacking. With some luck, we may be in a position to stack at a rate of 1mA/hr by late Saturday by which time, with some more luck, beam will be getting effectively to the pbar target. Saturday morning, then, has been scheduled for pbar studies with reverse protons. >>> (Since writing this, the helium situation has improved and if we get beam on the target, (previous paragraph), we should be able to stack efficiently). The air-conditioner in the pit - which was found not to be working on Wednesday- has been fixed. An oil leak was repaired and the compressor is now functioning properly. Discussion of next stack and coming week: We had a long discussion on the program for the next stack and this next week. For the next stack(s), suggested energies included: chi-0 peak (for gamma-gamma signal), chi-0 background (for gamma-gamma background), chi-0 shoulder (for chi-0 width) psi' peak (for electron weight calibration (cerenkov mainly)) psi' scan (for reference orbit setting) 1P1 point (for 1P1 search). There was discussion on the relative importance of the total width compared to the gamma-gamma branching ratio. So far we have the equivalent of 7pb-1 on the peak and about 4pb-1 on the background. If one does a direct background subtraction, assuming one knows the form of the background, the present data give a ~35% measurement of the gamma-gamma branching ratio (ie factor of 2 from lowest (65) to highest (135) value). 4pb-1 on the background would improve this to about 30%. 3pb-1 on the peak would further improve this to 28%. For the gamma-gamma measurement, it seems one really wants to do both peak and background. For the chi-0 width, it was not clear how much data were needed though it seems clear that a second point on the resonance and below the peak is needed. A (sort of) quantitative Monte Carlo test of the effect of a point at 3408 with luminosities of 1.5 pb-1 and 2.5 pb-1 was proposed. The idea is to simulate such a point assuming three widths, the present fit width and +/- 1 sigma, add the point to the current data set and see how it affects the answer. Stephen claimed that the best energy to use is Gamma/sqrt(2) from the peak. The psi' was introduced into the discussion for two considerations. The first mention was by Keith who pointed out that the existing psi' scan lacks luminosity information for more than half the scan and has truncated BPM measurements for the first half. To establish a correct reference orbit, he says we need to redo the scan - at some time. The prospect of more psi' data encouraged a request for a lot of data at the peak. This is to allow for calibration, in particular, of the large angle Cerenkov counter - which was running on CO2 for the psi', not Freon as it now does. It was suggested that one stack could provide adequate data for this plus a point on the chi-0 shoulder. Starting the 1P1 search was pushed because it is not known how long we will be able to take beam beyond May. If we can take beam after May 1, the argument goes we can come back to the chi-0, and if we cannot, then we need to maximize our chances of finding the 1P1 for which we were approved. The absence of a scan plan makes the 1P1 peak the energy for the first point. The ability to keep the beam energy constant is a key requirement. (This would not be so critical for a point at the edge of the scan) (I include Giovanni Borreani's preliminary note on a 1 month scan at the end). It became clear that there were two issues to resolve. The first issue was whether we spend a) no time this coming week on the 1p1 or b) we spend some time. Just as an expression of opinion, 3 people present (1 claiming a proxy) supported a) and 12 people supported b). Given this, we assumed at most one stack not devoted to the 1P1. The expressions of opinion were as below: 1P1 (5) (ie go to the 1P1 immediately on May 1- some of this because less than 1 week on the chi-0 was not considered useful) chi-0 peak (0) chi-0 background (2) chi-0 width point (0) psi' peak plus chi-0 shoulder (5) psi' scan (0) It is expected that the discussion will continue on Friday. (Given the good news about the helium, there is the real possibility of a stack before May 1. Stephen believes there is a consensus to ask for this stack to be 45 mA or so, spend about 10 mA on the psi' peak, and the rest on the chi0 shoulder point around 3407) ------------------------------------------------------------------------ >From Giovanni Borreani 4/26/00 Dear colleagues I have investigated a possible scheme for the first month of 1P1 search and simulated the possible outcome on a statistical basis under the following assumptions: 1) In a month running a total of 12 stacks will be collected for a total luminosity of 24 pb^-1 2) The beam width is dE_cm= 0.35 MeV (average from chi0 scan) 3) The resonance search is restricted to the interval 1.4 MeV wide, to cover +/-2*(0.15+0.2)MeV ( 2 standard deviation from the published mass) 4) The resonance parameters are assumed: mass=3526.2 MeV Br(pbar p)xBr(J/psi pi0)xBr(J/psi->e+e-) = 0.84 10**(-8) S_bkgd = 2.5 pb (corrected for 0.73 analysis + trig eff.) Gamma_tot = .44 MeV or 1.0 MeV 5) The scanning of the resonance is assumed to explore the following 7 points: 3523 (3526.2-0.67) (3526.2-0.33) 3526.2 (3526.2+0.33) (3526.2+0.67) 3529 6) two different shares of luminosity have been explored (units pb^-1) for a total of 24 pb^-1: W i d t h 3523 3525.53 3525.87 3526.2 3526.53 3526.87 3529 G=0.44 G=1.0 4 2 4 4 4 2 4 Exp.d1 Exp.e1 2 2 5 6 5 2 2 Exp.d2 Exp.e2 also the possible intermediate (14 pb^-1) was studied: 2 2 2 2 2 2 2 Exp.d1_i Exp.d2_i Notice that the center of gravity is deliberately left off the search: because of lack of time we are forced to restrict ourselves to confirming or not confirming what we have seen in 1991. Notice also that the assumed branching ratio is half that of e760 for Gamma_tot=0.44 MeV and 0.74 that of e760 for Gamma_tot=1.0 MeV (tomorrow I'll try 1.0 MeV with a reduced branching ratio, to keep number of events equal on average to that of the simulation with Gamma_tot=0.44 MeV and study the effect of the width at constant number of events). The background points should be at least +-3 MeV from peak to be off resonance: in this way for a resonance 1 MeV wide, the convolution beam-resonance is <0.028 on the background points and 0.76 on the peak. I use the method described in my letter of april 7 with title '1p1 outlook' and refer to it for more details. I must only add that, since then, I have defined the following criterion for accepting an experiment: I generate a number (500) of experiments where the resonance hypothesis(H1) is true: the experiments are considered as candidates for a discovery if the error on the fitted branching ratio dbr/br is smaller than 0.5 (dbr is the parabolic error given by MINUIT). It turns out that this cut converts into a sharp cut on the likelihood ratio -2ln(L_H0/L_H1) as you will see in the dbr/br vs lamd plot. The fraction of experiments that have fitted mass within +-0.5 MeV from 3526.2 (cut $6) and 03525.7.and.mass<3526.7 cut $7 = dbr/br>0.and.dbr/br<0.5 The results are summarized in the table: exper. L_tot Nexp $6 $6.$7 power d1 24 500 68.1 546 352 0.70 d2 24 500 75.0 462 341 0.68 d1_i 14 500 39.1 413 229 0.46 -------------------------------------------------------- e1 24 500 85.2 447 427 0.85 e2 24 500 96.1 451 399 0.80 e1_i 14 500 49.8 414 334 0.67 In conclusion it appears that: 1) if the signal is half that of e760 (for G=0.44 MeV), we have 46% chance of seeing something, in this scheme, after collecting 14 pb^-1 2) when other 10 pb^-1 are added near the peak the chance becomes 69% 3) if the signal is 0.75 that of e760 (for G=1 MeV) we have a 67% chance of seeing something after 14 pb-1 and 80-85% after adding 10 pb^-1 This situation will worsen if the resonance is away from where we found it by more than 350 keV and/or the branching ratio is smaller than what assumed. These effects have not been estimated. Regards Giovanni