1. ATLAS P. Morettini - ATLAS Collaboration 3/9/20131Paolo Morettini - ICNFP 2013

2. ATLAS Outline In this talk, we will briefly review the performance of the ATLAS detector in its present configuration and we will illustrate the upgrade plan:  ATLAS performance 2010-2012  Motivations for the upgrade  Phase 0 upgrade – LS1 i.e. now  Phase 1 upgrade – LS2 – 2018-2019  Phase 2 upgrade – LS3 – 2022-2023 An obvious disclaimer: experience teaches us that many things can change in ten years, so, especially for Phase 2, there are many unknowns (from the scientific and also from the financial point of view) that can modify our plans. 3/9/2013 Paolo Morettini - ICNFP 2013 2

3. ATLAS Muon spectrometer  tracking MDT (Monitored drift tubes) CSC (Cathode Strip Chambers) RPC (Resistive Plate Chamber) Trigger TGC (Thin Gas Chamber) Trigger Toroid Magnet 3 Level Trigger system L1 – hardware – 100 kHz 2.5  s latency L2 – software – 3-4 kHz 10 ms latency EF – software – 100 Hz 1-2 s latency Inner Detector (ID) Tracking Silicon Pixels 50 x 400  m 2 Silicon Strips (SCT) 80  m stereo Transition Radiation Tracker (TRT) up to 36 points/track 2T Solenoid Magnet Calorimeter system EM and Hadronic energy Liquid Ar (LAr) EM barrel and end-cap LAr Hadronic end-cap Tile calorimeter (Fe – scintillator) hadronic barrel The ATLAS Detector 4/7/20123 Paolo Morettini - iWoRID 2012 Tile Calorimeter Liquid Argon calorimeter TRTPixel DetectorSCTSolenoid Magnet Muon Detector Toroid Magnet ATLAS Performance

4. ATLAS 6/5/2013 Paolo Morettini 4 ATLAS Trigger and DAQ 10 8 channels on-detector (40 MHz readout) L2 trigger – 10 kHz – 20 ms On L1 selected geometrical regions. Software. L1 trigger – 100 kHz – 25 us Then extraction from detector to readout system Event Filter – 400 Hz – 2 s Software. ATLAS Performance

5. ATLAS LHC Performance 2010-2012  LHC performed extremely well in the three years of the first run.  In 2012, peak luminosity has been systematically above 7 10 33 cm -2 s- 1  Total accumulated luminosity was 23.3 fb -1, enough to firmly establish the Higgs discovery and its basic parameters. 3/9/2013 Paolo Morettini - ICNFP 2013 5 ATLAS Performance 2012 2011 2010

6. ATLAS ATLAS Performance 2010-2012  ATLAS performance was as well satisfactory, with an average data taking efficiency exceeding 93% and a fraction of active channels of 95%  It is important to note that most of 2012 data was taken with an average number of pile-up events per crossing around 35, due to the 50 ns spacing. A data-taking environment more challenging than the one expected at design luminosity and 25 ns spacing. 3/9/2013 Paolo Morettini - ICNFP 2013 6 ATLAS Performance

7. ATLAS Trigger performance The ATLAS trigger system demonstrated to be robust and flexible enough to follow the rapid LHC luminosity increase maintaining full efficiency. Many algorithms (more than 500 trigger items) were developed to guarantee optimal selection. Stability vs pile-up was a concern, but no problem observed so far. 3/9/2013 Paolo Morettini - ICNFP 2013 7 ATLAS Performance Single e, Et > 25 GeV Efficiency vs pile-up

8. ATLAS Higgs physics perspectives As recognized in the recent European Strategy symposium, the priority after the consolidation of the Higgs discovery is to obtain precise measurement of the parameters of the new particle:  Mass and width  Quantum numbers  Couplings and self-coupling  Comparison with the SM LHC has been recognized as the natural facility to complete these studies, but obviously an increase in luminosity would significantly enhance the achievable precision. 3/9/2013 Paolo Morettini - ICNFP 2013 8 Motivations for the upgrade

9. ATLAS Beyond the Standard Model An increase in luminosity would as well be beneficial to extend the range of the searches for SUSY particle and for other “exotic” processes. 3/9/2013 Paolo Morettini - ICNFP 2013 9 Motivations for the upgrade SUSY particles at HL-LHC 3 TeV for squarks ~ 2.5 TeV for gluinos 400 GeV rise in sensitivity wrt the L=300 fb -1 case ttbar resonances at HL-LHC 6.7 TeV L=3000 fb -1 leptons + jets 5.6 TeV L=3000 fb -1 dileptons 4.3 TeV L=300 fb -1 leptons + jets 4 TeV L=300 fb -1 dileptons

10. ATLAS LHC upgrade plan 3/9/2013 Paolo Morettini - ICNFP 2013 10 Motivations for the upgrade

11. ATLAS Based on the experience of the first LHC run, we can say that ATLAS as is can successfully operate at 10 34 cm -2 s -1 and possibly more. However, the ultimate goal of accumulating 3000 fb -1 in more than 15 years at 5 x 10 34 cm -2 s -1 requires some intervention:  Several detectors, especially close to the beam line, will be damaged by the accumulated dose  The large number of interactions per crossing ( up to 150) will saturate read-out links and generate large occupancies.  A more and more selective trigger will be necessary to efficiently isolate the few interesting events 3/9/2013 Paolo Morettini - ICNFP 2013 11 Motivations for the upgrade 10 33 cm -2 s -1 5 10 34 cm -2 s -1

12. ATLAS ATLAS detector upgrade plan 3/9/2013 Paolo Morettini - ICNFP 2013 12 LS1 ↠ PHASE 0 ℒ = 10 34 cm -2 s -1 =24 100 fb -1 (2014-2017) LS2 ↠ PHASE 1 ℒ = 2x10 34 cm -2 s -1 =50 350 fb -1 (2019-2021) LS3 ↠ PHASE 2 ℒ =5x10 34 cm -2 s -1 =140 3000 fb -1 (2023-2030) Installation of the 4 th Pixel Detector Layer Pixel Detector improvements DBM beam monitor Silicon tracker cooling system replacement Muon EE chambers completion New Muon Small Wheel detector Upgrade of the central L1 trigger processor Topological L1 triggers L1 Calo granularity increase New “All Silicon” tracker New L0-L1 trigger schema Inclusion of track info at L1 Upgrade of the calorimeter readout Upgrade of the muon spetrometer Phase 0 In Progress…

13. ATLAS Pixel detector upgrade Need to cure progressive radiation damage and mitigate inefficiencies due to pile-up effects. Two substantial interventions are in progress during LS1:  Replacement of service distribution panels, to cure malfunctioning channels, increase accessibility and bandwidth.  Installation of a 4 th layer (IBL), close to the beam pipe (33-38 mm from the beam line). 3/9/2013 Paolo Morettini - ICNFP 2013 13 Phase 0 New SQP “old” Beam Pipe R = 29 mm IBL setup 25 mm 31-40 mm

14. ATLAS IBL technologies  Originally thought as a LS2 intervention, the Pixel upgrade was anticipated to guarantee a more robust tracking system and a less radioactive working environment.  New technologies prototyping Phase 2 upgrade:  New FE chip (FE-I4) in 130 nm CMOS, with smaller cells (50 x 250  m 2 ) and faster output links 3/9/2013 Paolo Morettini - ICNFP 2013 14 Phase 0 Pixel Read-out inefficiency vs LHC Luminosity FE-I3 FE-I4 b-tagging rejection vs pile-up With IBL Without IBL

15. ATLAS IBL technologies  Originally thought as a LS2 intervention, the Pixel upgrade was anticipated to guarantee a more robust tracking system and a less radioactive working environment.  New technologies prototyping Phase 2 upgrade:  New FE chip (FE-I4) in 130 nm CMOS, with smaller cells (50 x 250  m 2 ) and faster output links  3D sensors in the forward region (lower bias voltage, immunity to bulk defects), planar, slim edge, n-in-n in the central region. 3/9/2013 Paolo Morettini - ICNFP 2013 15 Phase 0

16. ATLAS More LS1 interventions…  As a part of the IBL installation, we will replace the beam pipe: the central part will be in beryllium, the outer part in aluminium.  The evaporative cooling system of Pixel and Strips will be replaced.  More Muon End-cap Extension chambers (EE) will be installed, to improve coverage in the 1.0 < |n| < 1.3 region.  Add specific neutron shielding  Diamond Beam Monitor (DBM) diamond pixel detector with IBL readout Then, in this long shutdown, as in the following ones, many small interventions will be performed here and there to cure problems that cannot be addressed in a regular shutdown or to replace obsolete components (e.g. power supplies, readout elements, …) 3/9/2013 Paolo Morettini - ICNFP 2013 16 Phase 0 EE

17. ATLAS ATLAS detector upgrade plan 3/9/2013 Paolo Morettini - ICNFP 2013 17 LS1 ↠ PHASE 0 ℒ = 10 34 cm -2 s -1 =24 100 fb -1 (2014-2017) LS2 ↠ PHASE 1 ℒ = 2x10 34 cm -2 s -1 =50 350 fb -1 (2019-2021) LS3 ↠ PHASE 2 ℒ =5x10 34 cm -2 s -1 =140 3000 fb -1 (2023-2030) Installation of the 4 th Pixel Detector Layer Pixel Detector improvements DBM beam monitor Silicon tracker cooling system replacement Muon EE chambers completion New Muon Small Wheel detector Upgrade of the central L1 trigger processor Topological L1 triggers L1 Calo granularity increase New “All Silicon” tracker New L0-L1 trigger schema Inclusion of track info at L1 Upgrade of the calorimeter readout Upgrade of the muon spetrometer Phase 1 Preparing TDR

18. ATLAS Muon Small Wheel upgrade The innermost layer of the muon endcap is extremely sensitive to beam background. While working fine at the moment, the existing detector would produce an excessive fake L1 rate at luminosities above 10 34 cm -2 s -1. Will be replaced with a new detector with higher position resolution (100  m) and direction reconstruction capability (1 mrad) to select tracks pointing to the primary vertex. 3/9/2013 Paolo Morettini - ICNFP 2013 18 Phase 1

19. ATLAS New Small Wheel impact From the detector technology point of view, NSW will use two solutions:  Small strip Thin Gas Chambers (sTGC) for L1 trigger  Micromegas (MM) for precision tracking The new detector will ensure a strong reduction of single muon rates with a reasonable safety margin up to 5 x 10 34 cm -2 s -1 3/9/2013 Paolo Morettini - ICNFP 2013 19 Phase 1 Muon L1 Rates vs p t threshold NOW with NSW x3 reduction for p T (μ)>20 GeV at 10 34 cm ‐2 s ‐1 sTGC MM sTGC

20. ATLAS Calorimetric trigger  The granularity of the EM L1 trigger will be increased, to exploit shower longitudinal and transvers shape.  Better electron-jet separation will be achievable.  Will allow un-prescaled single electron triggers at Et ~ 25 GeV above 10 34 cm -2 s -1  Together with an update of the central L1 trigger processor, topological L1 triggers will be available. 3/9/2013 Paolo Morettini - ICNFP 2013 20 Phase 1

21. ATLAS FTK FTK is a track trigger processor. It can produce tracks with a quality similar to the off-line in ~25  s.  Based on CDF experience  Pattern recognition is done using associative memories, track fit with a FPGA processor.  In ATLAS, FTK uses L2 trigger data, but the output is ready at the beginning of L2 software processing.  Better HLT algorithms for b- tagging,  identification and lepton isolation will be available. 3/9/2013 Paolo Morettini - ICNFP 2013 21 Phase 1

22. ATLAS ATLAS detector upgrade plan 3/9/2013 Paolo Morettini - ICNFP 2013 22 LS1 ↠ PHASE 0 ℒ = 10 34 cm -2 s -1 =24 100 fb -1 (2014-2017) LS2 ↠ PHASE 1 ℒ = 2x10 34 cm -2 s -1 =50 350 fb -1 (2019-2021) LS3 ↠ PHASE 2 ℒ =5x10 34 cm -2 s -1 =140 3000 fb -1 (2023-2030) Installation of the 4 th Pixel Detector Layer Pixel Detector improvements DBM beam monitor Silicon tracker cooling system replacement Muon EE chambers completion New Muon Small Wheel detector Upgrade of the central L1 trigger processor Topological L1 triggers L1 Calo granularity increase New “All Silicon” tracker New L0-L1 trigger schema Inclusion of track info at L1 Upgrade of the calorimeter readout Upgrade of the muon spetrometer Phase 2 LoI submitted

23. ATLAS New “All Silicon” tracker To face the challenge of HL LHC ATLAS will need a new tracker:  Progressive radiation damage will make the old detector inefficient  More granularity and more bandwidth is needed to operate at 5 x 10 34 cm -2 s -1, due to the large pileup. The layout proposed in the LoI provides 14 points/track to |n| < 2.7  Pixel: 4 layers + 5 disks, 25 x 150 (in) / 50 x 150 (out)  m 2  Strips: 5 layers + 7 disks stereo 3/9/2013 Paolo Morettini - ICNFP 2013 23 Phase 2

24. ATLAS Tracker layout Clearly there are many options under investigation, and the details of the tracker layout will be fixed later. The challenge is always the same:  Produce a mechanical support with the best possible thermal and mechanical characteristics and, at the same time, as light as possible.  Find room and power dissipation capabilities for the on-detector electronics and the links needed to transmit the enormous amount of data. Present estimate 2500-3000 lpGBT links at 9.6 Gb/s 3/9/2013 Paolo Morettini - ICNFP 2013 24 Phase 2

25. ATLAS TDAQ upgrade One interesting possibility open by the complete redesign of the tracker and by modern data transmission technologies is the inclusion of tracks at L1. This requires, however, some deep modification of the trigger strategy, as it is impossible to build tracks at full rate in few  s. The idea is to use a calorimetric and muon pre-trigger (called L0) to select events were tracks will be reconstructed. L0 signal will be sent only to tracker modules in selected space regions. 3/9/2013 Paolo Morettini - ICNFP 2013 25 Phase 2 Level-0 Rate 500 kHz, Lat. 6  s Muon + Calo Level-1 Rate 200 kHz, Lat. 20  s Muon + Calo + Tracks

26. ATLAS Calorimeter upgrade  No upgrade is required for the EM and Hadronic calorimeter to run at HL LHC.  The FE electronics (both LAr and Tile) needs however to be replaced. The idea is to move off-detector data from each collision (no on-detector buffering) and do L0, L1 processing off detector.  Hadronic endcap is designed for 1000 fb -1. A replacement is considered.  Forward calorimeter (3.2 < |n| < 4.9) may have overheating problems at high luminosities. A complete new system could be installed or just a new calorimeter in front of the existing one. 3/9/2013 Paolo Morettini - ICNFP 2013 26 Phase 2

27. ATLAS Muon spectrometer upgrade 3/9/2013 Paolo Morettini - ICNFP 2013 27 Phase 2 Extra layers with more resolution High resolution forward chambers New trigger chambers

28. ATLAS Summary A sophisticated apparatus like ATLAS needs constant care to operate efficiently. This is even more true if optimal performance has to be maintained over 20 years with a luminosity increase of a factor of 5 (and possibly more) compared to the original design. ATLAS has a three stage upgrade plan, following the LHC upgrade, with the emphasis on:  Replacing detectors damaged by radiation or saturated by the luminosity increase.  Add flexibility to the trigger system and bandwidth to the readout.  Replace obsolete components with newer and more maintainable technologies. 3/9/2013 Paolo Morettini - ICNFP 2013 28