Early searches for new physics with the atlas detector E. V. Sedykh

Several extensions beyond the SM predict new particle states with heavy masses decaying to two energetic partons visible as a pair of energetic jets (dijet) in the detector. An excited quark (q^*), which is the natural consequence of quark composite models, is considered as a benchmark model in ATLAS dijet searches because it predicts relatively large cross sections and has been searched for in many different experiments in the past.

Many extensions of the SM predict the existence of new heavy gauge bosons. It is common to name W′ any charge ±1, spin 1 particle outside the SM. Here we report on the ATLAS search for a W′ boson decaying to an electron and a neutrino.

1. 
   ATLAS detector
   

ATLAS is a multi-purpose detector with a forward-backward symmetric cylindrical geometry with almost 4π coverage in solid angle [1].

The inner detector consists of a pixel detector, a silicon tracker, and a transition radiation detector and is the subsystem closest to the interaction point. A thin superconducting solenoid magnet surrounds the inner detector and provides a field of 2 T. The calorimeter system consists of electromagnetic, hadronic, and forward calorimeters and is situated outside of the solenoid. The muon spectrometer is the largest subsystem and defines the size of the whole ATLAS with length of 44 meters and diameter of 25 meters. Large toroid magnets create magnetic field inside the muon spectrometer.

A 3-level trigger system is used for the selection of events of interest for physics and efficient use of storage facilities: Level-1 (L1), Level-2 (L2), and event filter. The L1 trigger gets reduced-granularity information from a subset of detectors to be able to make decisions within 2.5 μs. The L2 trigger uses more information from the detector and simplified reconstruction is carried out providing a decision time of 40 ms. The EF trigger makes full offline event reconstruction with an average event processing time of order 4 seconds.

2. 
   Search for dijet resonances
   

The analysis technique consisted of a search for a dijet mass resonance on top of a smooth and rapidly falling spectrum and relied on the observed dijet invariant mass m_jj distribution [2]. In the absence of observed new physics signals, upper limits were determined on the product of cross section (σ) and detector acceptance (A) for several q^* masses.

The 3.15 pb⁻¹ sample of 7 TeV pp collisions data was collected during periods of stable running in 2010 using a trigger configuration requiring the L1 hardware-based calorimeter jet trigger to be fired with a nominal energy threshold of 55 GeV [3]. This trigger had an efficiency of > 99% for events with at least one jet with transverse energy >150 GeV.

Jets were reconstructed using the anti-k_T jet clustering algorithm with a radius parameter R = 0.6 [4]. The inputs to this algorithm were clusters of calorimeter cells seeded by those with energy well above the measured noise.

Fig. 1. The data (D) m_jj distribution (filled points) fitted using a binned background (B) distribution (histogram). The predicted signals for m_q* of 600, 900, and 1500 GeV/c² are overlaid, and the bin-by-bin significance of the D-B difference is shown.

Fig. 2. The 95% CL upper limit on σA as a function of m_jj (black filled circles). The black dotted curve shows the expected 95% CL upper limit and the light and dark shaded bands represent the 68% and 95% credibility intervals of the expected limit, respectively. The dashed curves represent excited-quark σA predictions for different MC tunes and different PDF sets.

The principal observable in this analysis, the dijet invariant mass m_jj, was used to select the final event sample by requiring m_jj > 350 GeV in order to eliminate any potential bias in the m_jj distributions from the selection requirements on the jet candidates.

MC signal events were generated using the excited quark () production model. The excited quark q^* was assumed to have quark-like couplings to the SM.

The background shape was determined by fitting the observed spectrum with the smooth function

where , and p_{0,1,2,3} are free parameters constrained such that f(1) = 0 and f(0) = +∞. The function in Eqn. 1 has been shown to fit the m_jj observable well in Pythia [5], Herwig [6], and NLO pQCD predictions for pp¯ collisions at = 1.96 TeV [7]. Studies using Pythia and GEANT4 [8, 9] based detector simulation scaled to the data integrated luminosity were performed to demonstrate that the smooth and monotonic form of Eqn. 1 described QCD-predicted m_jj distributions arising in 7 TeV pp collisions. There was good agreement between the MC prediction and fitted parameterization, as evidenced by a χ²/NDF of 27.0/22 over the dijet mass range 200 < m_jj < 1900 GeV (Fig. 1).

The presence or absence of detectable resonances in this distribution was determined by performing several statistical tests. The results of all these tests supported the background-only hypothesis and did not yield any evidence of a resonance.

The dominant sources of systematic uncertainty encountered in this analysis were the absolute jet energy scale, the time-integrated luminosity, the background fit parameters, and the jet energy resolution. These uncertainties were incorporated as nuisance parameters into the likelihood function, and then marginalized by integrating over each variable. Fig. 2 depicts the resulting 95% credibility-level (CL) upper limits on σA as a function of m_jj. The corresponding observed 95% CL excited quark mass exclusion region was found to be 400 < m_q* < 1265 GeV using MRST2007 PDFs [10], extending limits 260 < m_q* < 870 GeV from Tevatron experiments [11, 12].

3. 
   Search for states with an electron and missing transverse energy
   

This analysis used reconstructed electron candidates in the electromagnetic part of the calorimeter and missing transverse energy reconstructed using the information from the whole calorimeter.

Energy clusters were reconstructed in the electromagnetic compartment with a sliding window algorithm and then identified as electrons by matching them with inner detector tracks. The electron energy was obtained from the cluster and its direction from the track. ATLAS defines three levels of electron identification using a set of shower-shape, track matching and inner detector track variables: loose, medium and tight [13]. This study makes use of medium level for which shower-shape and track-matching criteria give about 95% identification efficiency for electrons with p_T > 200 GeV and a rate of 1/5000 to falsely identify jets as electrons.

The neutrino is not detected directly but the transverse components of its momentum are taken to be the missing E_T, i.e. the transverse energy required to balance the other objects reconstructed in the event. The missing E_T is obtained from a vector sum over calorimeter cells associated with so called topological clusters [13]:

Fig. 3. m_T spectra after the final selection. Points are ATLAS data and the filled histograms show the MC background from QCD, that plus , and that plus W and Z boson contributions. Open histograms are W′ signals added to the background. All MC is normalized to the data integrated luminosity using the calculated cross sections.

7 TeV pp collisions data for this study were collected during stable run periods and only luminosity blocks without problems in the inner detector, calorimeter, trigger, or in the measurement of the beam position or luminosity were used in the analysis. The integrated luminosity for the data used in this study was 317 nb⁻¹ [14].

Events were required to have a primary vertex reconstructed from at least three tracks with p_T above 150 MeV and |z|< 15 cm from the center of the collision region to reject cosmic and beam-related backgrounds. Spurious tails in may appear because of calorimeter noise, and are suppressed by checking the quality of reconstructed jets.




Fig. 4. 95% CL limits for 6 mass points on W′ production (filled circles) and the Pythia SSM predictions (open circles) for the range of expected limits for fluctuations up to 1σ (light gray) and 2σ (gray) standard deviations in background level.
Events were required to have exactly one candidate electron defined as follows. A candidate electron is one reconstructed with E_T> 20 GeV, |η| < 1.37 or 1.52 < |η| < 2.47 (excluding electromagnetic calorimeter crack region) and satisfying medium electron requirement described above. In addition, the inner detector track associated with the electron is required to be close to the primary vertex, specifically with a transverse distance of approach satisfying |r_PV0| < 1 mm and longitudinal distance at this point |z_PV0| < 5 mm.

The kinematic variable used to identify the W′ is the transverse mass , where φ is the angle between the transverse components of the electron momentum and the missing momentum. The transverse mass distribution for the signal has a Jacobian peak which falls sharply above the for W′ mass.

The major background for this search is the irreducible tail of the SM W boson decay but there are also contributions from and other QCD sources. To suppress the latter, we require the electron to be well isolated, defining isolation by and requiring < 0.05. Here is the electron transverse momentum and the sum in the numerator is over the transverse momenta of the inner detector tracks with > 1.0 GeV in a cone < 0.30 () around the direction of the electron. For the final selection, we additionally require  > 25 GeV. The final m_T spectrum is shown in Fig. 3. The agreement between data and MC is good over the entire m_T spectrum.

Source of uncertainty	Size (%)
Identification, material, fiducial cuts	8.0
Electron energy scale	2.0
Mass dependence, scale and PDF variation	7.0
QCD scale factor for mT>300 GeV	<5.0
Total integrated luminosity	11.0

Table . The systematic uncertainties used for the calculation of the limits in W′ searches.
The Pythia [5] signal model used as a benchmark for W′ searches is the Sequential Standard Model (SSM). In this model, the new heavy gauge bosons have the same couplings as the SM W. Details on MC samples generation, the propagation of particles and the response of the detector may be found in Ref. [14].The p_T and m_T spectra show no evidence for the existence of a W′ and the data are used to set limits on σB for a series of W′ masses ranging from 150 to 600 GeV. Limits are obtained by counting the number of events with m_T > 0.7m_W´.

Table 1 shows the systematic uncertainties used for the calculation of the limits. Besides the luminosity uncertainty, the main uncertainties are those coming from electron identification, material effects and fiducial cuts. The effect of the large QCD scale factor uncertainty decreases as the m_T cut value increases. The impact of event pileup is negligible with respect to the other uncertainties considered.

Taking into account all mentioned above uncertainties ATLAS excludes an SSM W′ with mass less than 465 GeV at 95% CL (Fig. 4). This result is consistent with the limit set at the Tevatron experiments (1.0 TeV) [15, 16].
References
[1] ATLAS Collaboration, JINST, 3, S08003 (2008).

[2] ATLAS Collaboration, Phys. Rev. Lett. 105, (2010), 161801.

[3] ATLAS Collaboration, ATLAS-CONF-2010-093 (unpublished, 2010).

[4] S. G. P. Cacciari and G. Soyez, JHEP, 04, 063 (2008).

[5] S. M. T. Sjostrand and P. Skands, JHEP 05 (2006) 026.

[6] G. Corcella et al., JHEP, 01 (2001) 010.

[7] T. Aaltonen et al. (CDF Collaboration), Phys. Rev. D, 79, 112002 (2009).

[8] ATLAS Collaboration, (unpublished, 2010), arXiv:1005:4568v1 [physics.ins-det].

[9] S. Agostinelli et al. (GEANT4), Nucl. Instrum. Meth., A506, 250 (2003).

[10] A. Sherstnev and R. S. Thorne, Eur. Phys. J., C55, 553 (2008), arXiv:0711.2473 [hep-ph].

[11] T. Aaltonen et al. (CDF Collaboration), Phys. Rev. D79, 112002 (2009).

[12] V. M. Abazov et al. (D0 Collaboration), Phys. Lett. B693, 531 (2010), arXiv:1002.4594 [hep-ex].

[13] ATLAS Collaboration, CERN-OPEN-2008-020 (2009), arXiv:0901.0512. [hep-ex]

[14] ATLAS Collaboration, ATLAS-CONF-2010-089 (unpublished, 2010).