- QQbar X-sections at 161 and 172 GeV ----------------------------------- Marie-Noelle has given a detailed description of the hadronic x-section measurements and the evaluation of systematic errors. For the individual numbers (backgrounds, efficiencies and x-sections) it is referred to the tables on the Web and to a note by Marie-Noelle (to be submitted asap). After a simple preselection which requires at least 5 tracks the main contributions to the cross section are: gamma-gamma, qqbar, WW Gamma-gamma : ------------- The sample corresponds to the sum of untagged and tagged events. Taking the low Mvis (<40 GeV) region, she can obtain the normalization of this contribution from the data. In the Monte Carlo (Pythia,Herwig) one finds a smaller x-section, therefore the expected x-sections are rescaled to give the observed ones, and the difference of the expected and the rescaled values is taken as systematic error, i.e. 161 : 11.9 pb --- rescale ---> 13.35 pb +- 12 % sys 172 : 10.9 pb --- rescale ---> 12.53 pb +- 15 % sys This systematic error is unproblematic, as the final background-contribution from gamma-gamma to the high-energy sample is very small. Note : M.N. corrects for the calibration of the SICAL energy by applying a down-scale factor of 1.23 (this is related to bunch #2) to the data. Dieter suggests that this should be made public. ISR-gamma tagging : ------------------- First jets are formed with JADE ycut=0.008. Then she looks at the electromagnetic content of the jet (photon or electron-type from Eflow), which should exceed 95 % for ISR. A final cut of 50 GeV is applied to the energy sum of all tagged jets. In the data one finds 1.21+-0.11 times more ISR than in the MC (with a x-section of 148.5 pb). A particular problem is found in the cos(theta)-distr. for ISR-photons in Pythia. Around cos(theta)=0 a dip is observed when taking only energetic photons (E>40 GeV). This dip is not there when using a special MC production by P.Janot (PYTH04 with MSUB 19. In this case the hard process e+e- -> Z gamma is generated, instead of e+e- -> Z. For high pt this should give a better description, however, one has to impose a pt cut at low pt (1 GeV) of the photon from the hard process to avoid singularities). If removing the Pythia events with Egamma>3 GeV (in the Z-return region) and replacing them with events from the special MC, the ratio Data-MC is improved to 1.03+-0.11, and the angular distribution matches. The difference between the MCs will be taken as systematic error on the efficiency. Efficiency calculation : ----------------------- The efficiency is defined as Nevent with : Ntrack>4, Mvis(without gamma)>50 GeV; (s' from jet direction) Eps(s') = --------------------------------------------------------------------------- Nevent , s' = M(qqbar) In the calculation of s' also LCAL objects are removed. This definition treats FSR as ISR, evaluated to be 1.1 %. Note that up to now we have given the inclusive x-section for sqrt(s'/s)>0.122, however, we have no acceptance for sqrt(s'/s)<0.375! The final efficiencies are taken as the mean values between the values found with the standard MC and the rescaled one (for ISR effects). 161 : Eps = 0.903 +- 0.9 % 172 : Eps = 0.912 +- 1.0 % Marie-Noelle hasn't looked in detail at KORALZ, as we have a MC production only at 161 GeV. She has found that KORALZ gives too many ISR-photons at large angles. Dieter suggests that we should generate additional MC sets with KORALZ. Guenther will take care of this. If looking at the efficiencies for the Z-return area, she finds a difference of 2% between Pythia and the Janot-model, and in addition 15% more measured photons in the apparatus. Altogether this results in a systematic error of 0.9 % on the MC efficiency. The error from the E-scale is calculated by taking the known uncertainty on charged momenta, delta(Ecal)=1.5% and delta(Ehad)=2.5%. The main remaining background from WW is taken from the SM expectation, with the nominal value calculated taking Mw from the CDF/D0 measurements. At 172 GeV she takes the difference of the WW x-section from CC03 and all 4-fermion. Ian thinks that this is an overestimation. As Zfitter has the 4-fermion processes built in, we should take CC03 as background. As a further test of double-radiative events she looks at the number of events with Mvis/Elep<0.7. For example, at 172 GeV one excepts 6.3% of all events, we observe 7.1+-0.96 %. This is in agreement, however, she takes these 0.96% as additional systematic error. In order to test FSR and the ISR/FSR interference she has computed the efficiency by switching the FSR on and off. Guenther pointed out that Pythia has no interference built in at all, therefore this method is somehow vague, and it is not clear how the corresponding systematic error should be evaluated. Future projects : ---------------- One should try to minimize the error on the x-section as a function of the s'-cut. Note that Delphi and L3 use a s'-cut of 0.85, whereas we and OPAL have s'>0.9. SM-expectations : ---------------- Calculated with Zfitter. Marie-Noelle doesn't know yet how to estimate the systematic error of the predictions related to FSR.
- Leptonic X-sections and asymmetries at 161 and 172 GeV ------------------------------------------------------ Elsa has shown a summary of her measurements and systematic studies of x-sections and asymmetries for dimuons and ditaus, as well as x-section results for dielectrons. For the detailed tables of x-sections, efficiencies etc. it is referred to a note (again, to appear asap) and the corresponding Web page. There is also a link to the transparencies shown by Elsa. Single Muons : ------------- She uses the standard digital selection which was already applied at the Z and at 133 GeV. Furthermore she uses a calorimetric selection. The .OR. of both is taken for the final muon identification. Dimuons : -------- The two most energetic muon candidates with opposite charges are selected, with the further conditions sum(momenta) > 60 GeV/c, Invariant Mass(dimuons) > 60 GeV. Systematics : ------------- Apart from the statistical uncertainty there are systematic uncertainties on backgrounds arising from an error on the --> x-section of e+e--->tau tau --> W x-section (from W-mass uncertainty) --> difference (Data-MC) of number of events for gamma-gamma, estimated in an invariant mass range of [15,50] GeV. The background from gamma-gamma --> tau tau is negligible. The efficiency has an uncertainty due to a different single-muon identification efficiency in data and MC. This is estimated from events at the Z peak (1996 runs) : Z data : 99.00 +- 0.30 % MC-2mu at the Z peak : 99.98 +- 0.01 % Maybe the efficiency map in the MC is not correct. Another explanation would be further tau background. Elsa will check this hypothesis. Other effects such as stability of cuts or charge misidentification are below the 1%-level. As is the case for qqbar, also here we have no acceptance at sqrt(s'/s)<0.122. It has been suggested to check if events with only a single muon seen are o.k. Muon Asymmetries : ----------------- Two methods have been tried: 1) Counting method (as used in the 133 paper) 2) A maximum likelood-fit using the function P(2) * (1 + 8/3 * P(1) * X + X**2) P(1) = asymmetry, P(2) = normalization There has been a long discussion about the applicability of method 2. If the asymmetry gets large, namely larger than 75%, then the cross section goes negative near cos(theta)=-1, if it is not constrained. Although this method has more statistical power, care should be taken if the fitted values of P(1) come out to be close to 75%, which is indeed the case for dimuons. For a later combination of the measurements of other collaborations, the counting method is to be preferred. It was observed that when looking at the measurements at both energies, the agreement with the SM is rather bad. Ditaus : ------ For the selection jets are formed with JADE ycut=0.0008, with a number of charged tracks required to be between 1 and 8. Further cuts are applied on the invariant mass, the momentum of the leading track of each jet, and the openeing angle of each jet (largest angle between two Eflow-tracks of the jet). The Bhabha-rejection has been improved. After an electron-id events are removed if one has one electron identified in each hemisphere, or only one electron/positron in one hemisphere and Ejet>0.6*Elep. Bhabha is the main background. Systematics from the background are estimated from the errors of the x-sections for dimuon,qqbar production and Bhabha, and from the difference between data and MC for gamma-gamma --> tau tau. Systematics of the efficiencies are studied by checking the stability w.r.t. to the selection cuts, using the known detector resolutions for momenta and energies. Tau asymmetries : ---------------- After the same selection as applied for the x-section measurement, one further asks for opposite charges of the two jets. It looks like there is a problem in the last bin of the cos(theta) distr, where data are always higher than MC (also observed for muons). In the tau-sample this could arise from remaining Bhabha background, however, in the muon-case it is not clear where it comes from. It has to be checked. Note that this bin is not counted for in the asymmetry measurement (one takes -0.9 < cos(theta) < 0.9). Here the fit gives asymmetry values which are well below 75%,therefore method 2 seems to be more justified. Dielectrons : ------------ First the two most energetic charged tracks are kept in CLAS15 events. The selection is based on the sum of their momenta and the sum of their ECAL energies. This rejects mainly dimuons and ditaus. The HCAL energy of tracks which pass through cracks and the energy of radiated photons are taken care of. Next the pseudorapidity of the tracks has to be lower than 0.3, and tracks identified as muons via QMUIDO are rejected. Backgrounds are negligible in this case. The systematic uncertainty is studied via checks of the stability of the cuts after smearing momenta and energies according to the resolutions: charged tracks : delta(p) = 6.2x10^-4 p^2 ECAL energy : delta(E) = 1.5 % * E HCAL energy : delta(E) = 8.0 % * E Furthermore the contamination from double radiative events is estimated.
- Discussion ----------
- Next meeting ---> In about two weeks. Exact date yet to be decided.
Guenther Dissertori
Thu May 22 19:21:38 MET DST 1997