aa r X i v : . [ h e p - e x ] O c t th International Conference on High Energy Physics, Philadelphia, 2008
Early measurements using W/Z in ATLAS
D. Prieur
STFC Rutherford Appleton Laboratory, Harwell Science & Innovation Campus, Didcot, OX11 0QX, UK
On behalf of the ATLAS Collaboration
The LHC experiments are close to collecting collision data. An overview of first physics measurements of the W and Z production cross-sections is presented. The electron and muon decay channels are considered. Emphasis will begiven to data-driven approaches.
1. INTRODUCTION
The study of the W and Z production is a fundamental area for the ATLAS [1] early running. These processesare very well understood theoretically and with high expected counting rates they will provide stringent tests ofQCD. They will help to understand the detector performance and be used for: calibration of the electromagneticcalorimeter (EM), alignment of the muon spectrometer system, as well as for extracting the lepton identificationefficiencies. The first measurements will consist in determining the W / Z and W / Z +jets cross-sections [2]. Increasedstatistics will give access to fundamental electroweak parameters. In this note, the emphasis is put on the earlycross-section measurements assuming a LHC peak luminosity of L = 10 cm − s − . W / Z CROSS-SECTION MEASUREMENT IN ELECTRON FINAL STATES
The reconstruction of electrons is based on clusters in the EM calorimeter, with a matching track in the InnerDetector [1]. The identification of isolated high- p T electrons is then based on the shapes of the EM showers, andon track reconstruction information. Three sets of identification criteria with different degrees of tightness (Loose,Medium, Tight) are used [2]. The selection of W → eν events proceeds as follows. The trigger selects events withat least one electron candidate with E T >
20 GeV. The analysis procedure selects events with exactly one electroncandidate satisfying E T >
25 GeV, | η | < .
37 or 1 . < | η | < . /E T >
25 GeV, and the transverse massof the ( l, ν ) system should satisfy M T >
40 GeV. The resulting transverse mass distribution is shown in Fig. 1. Jetevents constitute the largest background component. The jet production cross-section and fragmentation propertiesat the LHC are not well known and induce a significant uncertainty on the magnitude of this background. A data-driven method to monitor the jet background is applied and is presented in section 4. A MC study of this channel for R L dt = 50 pb − gave a measured cross-section of σ = 20530 ± ± ± σ = 20510 pb [3, 4].The analysis of Z → ee selects at least two electron candidates with E T >
10 GeV at the trigger level. The presenceof two loosely identified isolated electrons with E T >
15 GeV and | η | < . W → eν analysis, the jet background is estimated by a data-driven method. The signal and background fractions are estimated simultaneously, via a fit to both contributions.The signal is described by the convolution of a Breit-Wigner and a Gaussian resolution function, and the background,completely dominated by jet events, by an exponential function. A MC study of this channel for R L dt = 50 pb − gave a measured cross-section of σ = 2016 ± ± ± σ = 2015 pb. 14 th International Conference on High Energy Physics, Philadelphia, 2008
Invariant Mass Mee (GeV)0 20 40 60 80 100 120 140 160 180 200 E v e n t s / G e V -1
50 pbExtrapolated BackgroundSignal 50) × QCD MC stat (
ATLAS [GeV] TW M0 20 40 60 80 100 120 e v e n t s / G e V -2 -1 e ν WeQCD τ ντ WZeeATLAS -1
50 pb
Figure 1: Left: Transverse mass distribution in the W → eν channel for signal and background, for R L dt = 50 pb − . Right:Di-electron invariant mass distribution in the Z → ee channel, for signal and background, for R L dt = 50 pb − . [GeV] TW M0 20 40 60 80 100 120 140 E v e n t s / G e V µ νµ W τ ντ W µ µ Z tt QCD
ATLAS [GeV] µµ m40 50 60 70 80 90 100 110 120 A r b i t r a r y u n i t s ATLAS µµ→ Z νµ→ W µµ→ bb µµ→ tt ττ→ Z Figure 2: Left: Transverse mass distribution in the W → µν channel, for signal and background, for R L dt = 50 pb − . Right:Di-muon invariant mass distribution in the Z → µµ channel, for signal and background, for R L dt = 50 pb − . W / Z CROSS-SECTION MEASUREMENT IN MUON FINAL STATES
The analysis of W → µν selects events with at least one muon candidate with E T >
20 GeV at the triggerlevel. The events are further selected requiring exactly one muon track candidate, identified in the muon andinner detector tracking system, satisfying | η | < . p T >
25 GeV. The energy deposited in the calorimeter,in a cone of radius ∆ R = 0 . /E T >
25 GeV and M T >
40 GeV. Figure 2 shows the corresponding W transverse mass distribution before the M T cut. In contrast to the electron channels, the jet background is less important here. The dominant backgroundscome from W → τ ν and Z → µµ events. These processes are well understood theoretically and can be safelyestimated based on simulation. A MC study of this channel for R L dt = 50 pb − gave a measured cross-section of σ = 20530 ± ± ± σ = 20510 pb.The Z → µµ analysis uses the 10 GeV single muon trigger. The data sample is further reduced by requiringtwo offline tracks with opposite charges, in the muon spectrometer only, with | η | < . p T >
20 GeV. Theinvariant mass of the muon pair is required to fulfil | M µµ | <
20 GeV. The corresponding invariant massdistribution before the mass cut is shown in Fig. 2. In this channel, the dominant background originates from t ¯ t events. The jet background is expected to be smaller, but is theoretically not well known. Other backgrounds aresmaller, theoretically well known, and contribute negligibly to the overall background uncertainty. A MC study ofthis channel for R L dt = 50 pb − gave a measured cross-section of σ = 2016 ± ± ± σ = 2015 pb.
4. DATA-DRIVEN BACKGROUND ESTIMATION FOR W → eν The principle is to measure the normalisation and shape of the jet background before the /E T cut, in a sufficientlypure jet sample. This sub-sample is then used to evaluate the rejection of the /E T cut, allowing a realistic estimation24 th International Conference on High Energy Physics, Philadelphia, 2008 [GeV] T E10 15 20 25 30 35 40 45 50 E v en t s / . G e V ATLAS
Figure 3: Comparison of the jet background (points with error bars) and the fitted background (rectangles), for an integratedluminosity of R L dt = 50 pb − . of the jet background in the W → eν selection. The jet background control sample is selected using a single photontrigger with E T >
20 GeV, and subsequent calorimeter only based electron identification. Simulation studies showthat these selections provide a sample almost entirely composed of jet events, even at high values of /E T , and thatthe shape of the /E T distribution is identical, within the statistical precision, to that of the jet background in the W → eν sample (see Fig. 3). Above /E T >
10 GeV, the slope can be described with the convolution of an exponentialand a second degree polynomial function. After the subtraction of the estimated background from the signal sample,the analysis then proceeds by applying the /E T selection mentioned above. W / Z + JETS The production of W / Z +jets events is an interesting measurement in itself. In addition this process is a backgroundto many other Standard Model and beyond Standard Model physics channels. Furthermore these channels will testjet reconstruction techniques. Compared to the W / Z inclusive production, more statistics is needed and the analysisdone is based on an integrated luminosity of 1 fb − . Selections are similar to the inclusive W / Z production, exceptthat one, two or three jets with E T >
40 GeV are required. At larger jet multiplicities, the dominant backgroundarises from top quark events. The jet energy scale is the largest source of systematic error on the cross section. Theinitial uncertainty on the jet energy scale is expected to be 5-10%. A jet energy scale with precision better than 10 %is required to distinguish between the LO/NLO predictions of the different Monte-Carlo generators.
6. CONCLUSIONS