The design and performance of the ATLAS jet trigger

Transcription

1 th International Conference on Computing in High Energy and Nuclear Physics (CHEP) IOP Publishing Journal of Physics: Conference Series () doi:.88/7-696/// he design and performance of the ALAS jet trigger Shima Shimizu on behalf of the ALAS collaboration Kobe University, Rokkodai-cho -, Nada-ku, Kobe 67-8, Japan Abstract. he ALAS jet trigger is an important element of the event selection process, providing data samples for studies of Standard Model physics and searches for new physics at the LHC. he ALAS jet trigger system has undergone substantial modifications over the past few years of LHC operations, as experience developed with triggering in a high luminosity and high event pileup environment. In particular, the region-of-interest based strategy has been replaced by a full scan of the calorimeter data at the third trigger level, and by a full scan of the level- trigger input at level- for some specific trigger chains. Hadronic calibration and cleaning techniques are applied in order to provide improved performance and increased stability in high luminosity data taking conditions. In this note we discuss the implementation and operational aspects of the ALAS jet trigger during and data taking periods at the LHC.. Introduction A jet is a collimated spray of hadrons. It is produced by a scattered quark or gluon with high transverse energy (E ) and is an important signature in the collider physics. he LHC provided proton-proton collisions at the centre-of-mass energies of 7 ev with luminosity up to.6 cm s in and of 8 ev with luminosity up to 7.7 cm s in, with a frequency of MHz. In the operation of the ALAS detector [], the ALAS trigger performs the first selection of interesting collision events at the LHC. he ALAS jet trigger [] is designed to tag jets with high flexibility, to adapt to the LHC beam conditions and to allow a variety of physics analyses using the ALAS data.. Overview of the ALAS jet trigger system he ALAS trigger system consists of three levels, Level (L), Level (L) and Event Filter (EF), where L is hardware-based and L and EF are software-based. he system can deal with a maximum bunch-crossing rate of MHz. L is designed to give a trigger rate rate of 7 khz (7 khz was the actual peak value in ) with a fixed latency less than. µs, L to give. khz within ms ( khz in 7 ms in ) and EF to have an output rate of Hz within s (7 Hz in about s in ). he ALAS jet trigger is allocated about % of the total bandwidth of the ALAS trigger. he schematic illustration of the ALAS jet trigger, which includes upgrades in, is shown in Figure. At L, jet finding is performed with a sliding window algorithm applied on L calorimeter towers, which have a granularity of.. in pseudo-rapidity η and azimuthal angle φ. he algorithm looks for local energy maxima in the ALAS calorimeter. he detector region where a maximum is located is called Region of Interest (ROI) and Content from this work may be used under the terms of the Creative Commons Attribution. licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by IOP Publishing Ltd

2 th International Conference on Computing in High Energy and Nuclear Physics (CHEP) IOP Publishing Journal of Physics: Conference Series () doi:.88/7-696/// the information of the ROI is passed to L. In the original implementation of the L jet trigger algorithm, jets were reconstructed with the cone algorithm in ROIs using calorimeter cells. Starting from, a new Full Scan algorithm was added to the L system. he new algorithm, L Full Scan, can perform reconstruction of anti-k t [] jets (as implemented in the Fast package []) using L calorimeter towers with a granularity of.. in η φ, across the entire ALAS detector. In this algorithm, electromagnetic energies of jets can be corrected to compensate for calorimeter response to hadrons and for energy loss in dead materials. he reconstructed jets at this layer are called LFS jets or L. jets. hese LFS jets and ROI from L can be used as inputs at L for further jet reconstruction with anti-k t or cone algorithms using calorimeter cells. At EF, topological clusters [] are created from calorimeter cells and reconstruction of anti-k t jets is done using topological clusters across the entire detector. All the offline jet algorithms and calibration can be applied at EF. he ALAS jet trigger has a variety of possible configurations and this variety is increased especially by the newly implemented L Full Scan. Not Figure. Schematic illustration of the ALAS jet trigger system including upgrades in. only single jet triggers and multijet triggers but combined triggers with other physics signature, such as missing transverse energy, are also possible, with several choices in jet reconstruction such as constituents, jet algorithm and jet calibration.. Pileup and noise suppression Due to the high instant luminosity provided by the LHC, calorimeter energy deposits in an interesting event can be affected by other additional interactions (called pileup) in the same bunch crossing or in close bunch crossings. In the operation, the average number of interactions per bunch crossing was.7 [6]. A noise suppression tool is implemented at L and EF and only calorimeter cells with energy depositions above a certain threshold are considered in the jet reconstruction at L and EF since. Figure shows efficiencies of single jet triggers at L, L and EF measured using data. Pileup suppression improves the jet energy resolution and hence steepens the turn-on curves of the efficiencies.

3 th International Conference on Computing in High Energy and Nuclear Physics (CHEP) IOP Publishing Journal of Physics: Conference Series () doi:.88/7-696/// ALAS Preliminary Data anti k R=. η <.8 L E > GeV No pile up noise suppression L E > GeV EF E > GeV With pile up noise suppression L E > GeV EF E > GeV Offline E [GeV]..8 ALAS Preliminary Data anti k R=..6< η <..6 L E > GeV No pile up noise suppression. L E > GeV EF E > GeV With pile up noise suppression. L E > GeV EF E > GeV [GeV] Offline E Figure. Efficiencies of single jet triggers at L, L and EF are shown for anti-k t jets with a jet size of R =. in the central region of η <.8 (left) and in the forward region of.6 < η <. (right), as a function of offline jet E. For L and EF, efficiencies with and without noise suppression are compared by full markers (with suppression) and open markers (without suppression). [7] Normalised entries.... ALAS Preliminary rigger run offline L (.. towers) L. (.. towers) L. (.. towers) L (all cells) s = 7 ev Data E,offline > GeV η <.8 offline.. η η online offline Figure. angular resolution in η is shown for L jets, LFS jets (denoted as L.) with different tower sizes in inputs and L jets. he trigger algorithms are run offline and η differences with offline jets are plotted. he algorithms use a sliding window of.8.8 for L, anti-k t with a jet size of R =. for LFS and a threeiteration cone with a jet size of R =. using inputs from the L ROI for L. [7]. Performance of L Full Scan trigger Performance of the new L Full Scan trigger was tested by running the algorithm offline on the proton-proton collision data in. Figure shows η differences between online jets and corresponding offline jets. It shows that jet angular resolution is quite improved in LFS jets compared to L jets. he processing time needed for LFS is also checked. he time taken to read-out and to find jets for LFS was measured during lead-lead collisions at the LHC, where multiplicity in the ALAS detector is extremely high. hey are shown in Figure. he total latency is within the nominal time limit at L of about ms.. of L Full Scan in the operation LFS was in use in the operation and Figure shows the efficiencies of single jet triggers at LFS. In the plot, efficiency of a trigger requiring a jet with the electromagnetic (EM) transverse energy of GeV and that of a trigger requiring a jet with the calibrated transverse energy of GeV, where hadronic energy calibration is applied to EM energy, are shown. While both

4 th International Conference on Computing in High Energy and Nuclear Physics (CHEP) IOP Publishing Journal of Physics: Conference Series () doi:.88/7-696/// Entries / ms 6 ALAS rigger Operations Data (Pb Pb) L. (.. towers), <t>~.7 ms L. (.. towers), <t>~7. ms Entries / ms ALAS rigger Operations Data (Pb Pb) L. (.. towers), <t>~.8 ms L. (.. towers), <t>~.6 ms Read out time [ms] finding time [ms] Figure. he time taken to read out the L calorimeter towers (left) and the time taken for jet finding using the anti-k t algorithm with a jet size of R =. (right) are shown using towers with granularity of.. and.. as input, at the L Full Scan (denoted as L.). hey were measured in lead-lead collisions in. [7]..8.6 ALAS preliminary Data η <.8 anti k R=. ALAS preliminary Data 6 offline jets anti k R=. η <.8 jet.. LFS EM J LFS EM+JES J Leading jet p (EM+JES) [GeV],offline Figure. Efficiencies of single jet triggers at LFS for anti-k t jets with a jet size of R =. in the central region of η <.8, as a function of offline jet p. wo triggers are compared, one requiring a jet with an electromagnetic energy of GeV (EM J) and one requiring a jet with a calibrated energy of GeV (EM+JES J), where hadronic energy calibration is applied. [7]. L 6j LFS 6j th Offline 6 jet E [GeV] Figure 6. of multi jet triggers requiring six jets at L (L 6j) and LFS (LFS 6j) measured in events with six offline anti-k t jets for a jet size of R =. with E > GeV and η <.8, as a function of E of the sixth jet. he events are preselected using a trigger with four jets requirement. [7] triggers give efficiency higher than 99 % for offline jets with p > 6 GeV, the one with hadronic calibration (EM+JES) shows a steeper turn-on curve due to better jet energy resolution than the other. In terms of trigger rate, this hadronically calibrated trigger has a smaller trigger rate by 8 %. Figure 6 shows the efficiencies of multijet triggers requiring six jets, at L and LFS. he efficiencies are measured against events with six offline anti-k t jets with a jet size of R =., where each jet should have E > GeV and η <.8 after the offline calibration. hese events are preselected using a trigger requiring four jets. While the L multijet trigger has an inefficiency of more than % even at E = GeV of sixth jet due to the different geometry

5 th International Conference on Computing in High Energy and Nuclear Physics (CHEP) IOP Publishing Journal of Physics: Conference Series () doi:.88/7-696/// between L sliding windows and offline jets, the efficiency is recovered in the LFS multijet trigger. It shows one of the advantages of LFS, which performs jet reconstruction across the entire detector, not only in ROIs given by L. 6. Summary he ALAS jet trigger was designed to tag jets with high transverse energies in proton-proton collisions provided by the LHC with high instantaneous luminosity and collision rates up to MHz. It consists of three levels, L, L and EF, where the output rate is reduced at each level with longer latencies at higher levels. A new algorithm called LFS was implemented at L in. It enables reconstruction of anti-k t jets using L calorimeter towers across the entire ALAS detector, in addition to jet reconstruction within ROIs identified at L. he processing time of LFS is within the requirement of L even in lead-lead collisions. It provides jet finding at L with good angular resolution and also provides more flexibility in the jet trigger configurations. As seen in the improved jet trigger efficiencies in the data-taking, the ALAS jet trigger has achieved excellent performance in the ALAS operation. References [] ALAS Collaboration 8 JINS S8 [] ALAS Collaboration 8 CERN-OPEN-8- (Preprint arxiv:9.) [] Cacciari M, Salam G and Soyez G 8 J. High Energy Phys. 8 6 (Preprint arxiv:8.89) [] Cacciari M, Salam G P and Soyez G [] Lampl W et al. 8 AL-LARG-PUB-8- [6] [7]