LHC Upgrades Physics Detectors Marc Weber RAL Why

LHC Upgrades – Physics & Detectors Marc Weber (RAL) • Why LHC Upgrades? • Schedule and Technical Challenges • UK ATLAS and CMS Plans and Interests • Selected R&D Highlights Many thanks to numerous ATLAS and CMS colleagues for their input to this talk, in particular G. Hall and P. Allport

What is the LHC upgrade ? • Increase of luminosity by factor 10 by ~2020 SLHC Energy upgrade is not practical or affordable on this time scale First step of luminosity upgrade is factor 3 increase by ~2015 • Requires upgrade or replacement of all pre-accelerators ~400 proton-proton collisions every 50 ns • Price tag: ~ 1 billion € (for accelerator, from CERN budget) • Requires changes to LHC inner detectors, triggers and electronics • Highest priority of European particle physics community (see Particle Physics European Strategy Roadmap, CERN Council 2006) Marc Weber, PPAP, July 2009 2

Why upgrade the LHC ? After all the LHC will have made significant discoveries already. • Consolidation of LHC discoveries (e. g. Higgs boson, Supersymmetry, Z’) • Extended discovery mass reach by 30% or 1 Te. V for direct searches (e. g. heavy scalar quarks or gluinos) • Increased precision and access to rare processes (e. g. multi-gauge boson production, flavor-changing top decays, H to µµ) Potential has been studied for various physics and machine scenarios and will be refined over the next few years. No doubt that case will become even stronger with specific information from LHC. Marc Weber, PPAP, July 2009 3

Other considerations • Upgrade is “best value for money” and fully exploits current LHC investments • No other affordable new accelerator could cover or extend the SLHC effective energy range • Upgrade assures continued European leadership in particle physics Marc Weber, PPAP, July 2009 4

Schedule and organization • Schedule is driven by machine and desire to collect ~700 fb-1 of data before changes. Luminosity time profile is the big unknown. Expect continuous luminosity increase with two distinct phases preceded by shutdowns • Phase I - 3 x 1034 cm-2 s-1 by ~2015 • Phase II 1035 cm-2 s-1 by ~2020 • Both collaborations have started major R&D efforts and set up organizational structures • ATLAS plans: Letter of intent (LOI) 2010; Technical proposal (TP) 2011/12; Technical Design Report (2012/13) • CMS plans: LOI 2010; TDR 2012 Marc Weber, PPAP, July 2009 5

Technical challenges • SLHC detectors must equal LHC detector performance in a tougher environment LHC: ~1000 particles every 25 ns 30 overlap events SLHC: > 10, 000 particles every 50 ns 400 overlap events • Most (outer) detectors and infrastructure will survive and perform well at SLHC, but • • Need to replace inner trackers (pixels and strips) Need to modify/rebuild triggers CMS and possibly ATLAS require Level-1 track based trigger Need to modify electronics Marc Weber, PPAP, July 2009 6

Biggest tracker challenges: reduction of power consumption and detector mass Strong correlation between parameters 160 Mb/s electrical 5 Gb/s optical links Key challenge for LHC was: radiation hardness and scale of detectors

UK is strongly represented in ATLAS/CMS upgrade organization ATLAS Strips Tracker Upgrade (Allport), Module coordinator (Affolder) Optolinks (Issever), Powering (Weber) Radiation and Shielding (Dawson), Stave 09 deputy (Robinson) Thermal Management (Viehauser) Tracker Layout (Tseng) 3 D ATLAS R&D convener (Da Via) AFP convener (Watts), Computing (Jones) Review office (Tyndel), Trigger/DAQ (Gee) CMS Upgrade PM Tracker Trigger simulation (Nash) (Hall) (Foudas) (Newbold)

UK ATLAS Interests AFP Forward Physics Spectrometer at 420 m from IP Higgs quantum numbers, diffractive Physics, ɤp Tracker Upgrade Short strip silicon barrel, generic R&D, forward pixel discs Level-1 Calorimeter Trigger Track Trigger Higher-Level Trigger (HLT)/ DAQ Computing Interests overlap very strongly with UK contributions for LHC detector UK ATLAS aims for 10% contribution to upgrade Marc Weber, PPAP, July 2009 9

UK CMS Interests Simulation of detector design Tracker layout and L 1 track trigger performance Level-1 Track Trigger Conceptual design and prototyping; design of generic trigger and data distribution board Outer Tracker read-out electronics Design of silicon strip read-out chip (successor of APV 25); off-detector electronics (successor of FED) Pixels Replacement of Phase I pixel detector Tracking layer construction ECAL Endcap Vaccum phototriode replacement Again, interests consistent with contributions for LHC detector Marc Weber, PPAP, July 2009 10

Why Tracker input to CMS Level 1 trigger ? • Single µ and e L 1 trigger rates will greatly exceed 100 k. Hz – similar behaviour for jets • increase latency to 6. 4 µs but maintain 100 k. Hz for compatibility with existing systems, and depths of memory buffers Single electron trigger rate <p. T> ≈ few Ge. V/bx/trigger tower Isolation criteria alone are insufficient to reduce rate at L= 1035 cm-2 s-1 L = 2 x 1033 5 k. Hz @ 1035 L = 1034 muon L 1 trigger rate

CMS Track Trigger approaches Need to send reduced data volume from detector further logic readout electrodes (Aim for reduction by factor 20 with p. T > few Ge. V/c) a) Cluster width information to eliminate low p. T tracks – simple, but thinner sensors may limit b) Recognize track stubs in neighbouring layers – p. T cut set by angle of track in layer Pass ~1 mm ~100 μm Marc Weber, PPAP, July 2009 Fail Upper Sensor Lower Sensor 12

A CMS stacked trigger layer concept 3 D integration features • • • Short data path through interposer Information available regionally, close to where needed Requires Chip Through-Silicon-Vias (3 D) Direct Oxide Bonding to Large Area device High rate data transmission without disturbing analogue performance Sensor 1 ASIC 1 Interposer ASIC 2 Sensor 2 Marc Weber, PPAP, July 2009 13

Stacked Layer Algorithm Performance • Sensor separation is an effective cut on p. T • Width of transition region increases with separation due to: - pixel pitch - sensor thickness - charge sharing - track impact point Increasing separation + + 1 mm separation 2 mm 3 mm 4 mm • Max. efficiency decreases with sensor separation due to larger column (z) windows Performance of a stacked layer at R = 25 cm 10, 000 di-muon events with smearing Marc Weber, PPAP, July 2009 14

CMS Silicon Strip Readout IC • Fast front end in 130 nm CMOS – – comparator binary pipeline no ADC Very simple & lowest power FE amp comp. digital pipeline digital MUX vth off-chip O/P driver vth • Features retained from APV 25 – simple synchronous • no timestamps – constant data volume • no trigger to trigger variation simulated pulse shapes (CSENSOR = 5 p. F) vth digital CBC = CMS Binary Chip Binary unsparsified readout Power target: 0. 5 m. W/channel slow control, bias, test pulse, ……

CMS ECAL Endcaps at SLHC: Photodetectors • • • Current vacuum phototriodes will suffer significant performance loss at SLHC: - Degradation of photocathode - Radiation darkening of tube window There is interest in CMS in developing a new generation of robust photodetectors. Replacing Pb. WO 4 with a brighter scintillator (LYSO? ) would ease this problem. Silicon devices: Neutron damage high leakage currents amplifier noise? ‘Nuclear counter effect’ (direct sensitivity to shower leakage particles >> sensitivity to scintillation light high energy tail on energy measurement) APD? Vacuum devices: - Good match to bi-akali photocathode - Internal gain not necessary Vacuum photodiode? Marc Weber, PPAP, July 2009 16

Partial Endcap Upgrade ? At VPTs: Dose(h =2. 2)/Dose(h =3. 0) ~1/10 (neutron fluence ~1/3) ~25% (18/71) Supercrystals are at h >2. 2 Only replace detectors/crystals at small radius ? Very challenging because of complexity of EE construction and high radiation levels Needs detailed study

ATLAS Forward Physics Broad QCD and photo-production physics programme Plenty of diffractive events (SD and DPE) Central Exclusive Production of Higgses and determination of quantum numbers LHC as a photon-photon collider – photon-photon physics CDF ar. Xiv 0902. 1271 Quantum number selection rule. High precision mass measurement independent of decay channel See few events => JPC = 0++ cf. High energy photon collisions at the LHC – CERN April 2008 Production very large. Well-known cross sections for SM and BSM processes: SUSY production and anomalous couplings

AFP Detector Requirements • Silicon pixel detector to measure the 2 leading protons (1 μrad angular resolution) • Timing system (10 -20 ps resolution) to identify primary vertex • Beam proximity • Trigger capability • Radiation hardness Detectors integrated into beam line High precision mass spectrometer 420 m 30 7 x 8 mm 2 sensors per tracking station; 4 stations. Marc Weber, PPAP, July 2009 220 m Twice the number of sensors. 19

ATLAS Silicon Tracker (short strips) UK funding since 2 years. Intense phase of prototyping. Short-term deliverable is a 1. 2 m long integrated silicon supermodule ABCN-25 Readout chip (0. 25 µm CMOS) - UK initiated project and contributed to specs - UK first proved chip to be functional and developed DAQ - now wafer probing and contributing to 130 nm specs Hybrid - UK designed and built first ATLAS prototype hybrid - Fully functional and low noise - Vehicle for DAQ development, powering R&D and module design Module design - First prototype module built in UK Marc Weber, PPAP, July 2009 20

ATLAS Tracker Upgrade Module Concept 1. 2 m Marc Weber, PPAP, July 2009 21

Sensor R&D (generic R&D) n-on-p sensors Sensor R&D - UK driving R&D on radiation-hard planar n-on-p sensors (of interest for both short strips and pixels) - UK is leading and coordinating 3 D pixel sensor R&D - 3 D pixels are key technology for AFP and of interest for ATLAS inner pixel layers Vertical structure allows operation at reduced HV; Challenge are vias Marc Weber, PPAP, July 2009 Max. strip dose: 9× 1014 n/cm 2 Max. pixel dose: 2. 5× 1016 n/cm 2 14 µm holes, 200 µm deep 22

Services (generic R&D) Services (power, electrical signals, optolinks, cooling) are massive and have been major source of trouble at LHC. Services could be show-stopper at SLHC major R&D efforts and integration in supermodule Power distribution R&D - UK leading this new and popular field - Serial powering very promising; reduction of ~50 in number of power cables - Got full custom serial powering circuitry (with SPi and ABCN-25 chips) - System design of serial powered super-module is advanced Some SCT cables (cable length > 100 m) Marc Weber, PPAP, July 2009 23

Ideas for a hardware track trigger in ATLAS Cannot readout the whole tracker at 40 MHz n Within the L 1 latency ¨ Processing in parallel with L 1 Calo & L 1 Muon at 40 MHz n Requires substantial ondetector trigger logic to reduce bandwidth (CMS double layer coincidences) ¨ Regional tracker readout driven by L 1 Calo/Muon n L 1 Muon/Calo reduce rate from 40 MHz to ~400 k. Hz, then request data from small regions in the tracker ¨ Would require longer L 1 latency than today (6. 4 s? ) Marc Weber, PPAP, July 2009 n Outside the L 1 latency ¨ L 1. 5: if regional readout cannot fit in L 1 latency… n On L 1 A, data moves to buffers in the FEs, but does not get read out until a further Trigger decision is sent a few s later ¨ At the limit of very large buffers on FEs, the above scheme “approaches” current L 2 processing in Regions-of-Interest (Ro. I) 24

Level-1 Calorimeter Trigger: Phase I • Current L 1 Calo uses towers of Δ xΔ = 0. 1 x 0. 1 • Phase I upgrade uses the same data, and adds simple topology: – New firmware, faster backplane links, one new crate of electronics – Simple topologies – e. g. muon isolation, gap Current trigger tower • Studies underway: – Monte Carlo algorithm simulation with pileup – Bit error rates measured to be <10 -10 at 160 MHz – System architectures, including evaluation options in existing hardware Marc Weber, PPAP, July 2009 Backplane eye diagram at 160 MHz, 4 x design frequency

Level-1 Calorimeter Trigger: Phase II • Digitisation on detector. L 1 gets data with full readout precision (data rate: 160, 000 Gbits/sec !) • Higher granularity trigger towers, incl. fine strips in LAr 1 st layer -Separate depth samples in e. m. , smaller “minitowers”perhaps 0. 05 x 0. 05 -Send detailed object information to new global processor ( , , ET, quality…) LAr cell structure. One tower (red outline) could become ~10 minitowers -Incorporate similar info from L 1 Muon & Track -Run L 2 -like algorithms at L 1 • Requires complete rebuild of warm on-detector electronics and all of L 1 trigger Marc Weber, PPAP, July 2009 26

ATLAS Higher Level Trigger Phase I: Rolling replacement of HLT farms; continuing improvements and optimisations of tracking and selection software due to: increased occupancy; new B-pixel layer; etc. Phase II: Substantial changes in readout-system and network infrastructure; new architecture; more exclusive selections at HLT. Exploit and cope with new detectors: tracker, track trigger, L 1 calorimeter trigger Phase II HLT will be very different. While HLT input information will be more detailed, events are much more complicated. Some of the current reduction strategies will already have been applied before HLT. Marc Weber, PPAP, July 2009 27

Summary • LHC Upgrades will secure exciting physics programme at the energy frontier beyond 2018. • Technical challenges for experiments are significant, but current R&D and LHC lessons strongly suggest they will be met. • UK is very well positioned in ATLAS and CMS upgrade organization and expects to contribute to upgrades at a level comparable to LHC. Marc Weber, PPAP, July 2009 28

Backup Marc Weber, PPAP, July 2009 29

Pre. Processor 124 modules Readout data Marc Weber, PPAP, July 2009 0. 2 Calorimeter signals (~7200) 0. 1 Digitized ET Readout Driver 14 modules Backplane Cluster Processor 56 modules Merging 8 modules Backplane Jet/Energy 32 modules Merging 4 modules New Topology Processor CTP L 1 Muon Readout Region-of. Driver interest data 6 modules 30

Organisation • • • L 1 Calo Eo. I circulated within ATLAS – 3 extra institutes now in L 1 Calo, so we have 4 UK + 3 European + 2 USA – Birmingham, Cambridge, QMUL, RAL (PPD + ID) – Heidelberg, Mainz, Stockholm – ANL, MSU Original L 1 Calo was UK-led throughout construction and up to the present. UK provided 50% of effort and capital. ATLAS-internal TDAQ R&D system proposal contents agreed, expect to submit soon to ATLAS for internal review Marc Weber, PPAP, July 2009 31

STFC RAL CMS ECAL Endcaps at SLHC: Crystals Lead tungstate (Pb. WO 4) crystals suffer two types of optical damage under irradiation: - Ionising radiation produces colour centres. The colour centres self-anneal at room temperature. Under irradiation, a dynamic equilibrium is established, thus the density of colour centres and hence the decrease in light yield depends principally on the dose rate (rather than the integrated dose). The impact on detector performance would be substantial, but manageable at SLHC - Charged hadrons and neutrons. In addition to the ionisation damage, hadrons collide with Pb and W nuclei causing ‘star’ formation. The resulting damage to the crystal lattice produces optical scattering centres and light is lost through Rayleigh scattering. Self annealing is very slow at room temperature and the light yield decreases progressively with increasing dose. The performance loss at small radii (large | |) is expected to be unacceptable at SLHC Scintillators with lower-z nuclei are expected to be more resistant to lattice damage. This has been verified with Ce. F 3 LYSO is under study. CMS ECAL at SLHC 14/0709 R M Brown - RAL 32

Phase 1 Upgrade – ATLAS – Potential UK interest • “ PHASE 1” Insertable B-Layer (IBL) Project Kick-off meeting 8 July 2009. Cost 9. 6 MCHF Sensors 642 k. CHF Install in 2014. Possible technologies – 3 Dsilicon/new planar silicon/ Diamond Pre-production with two technologies 2010. Choose technology 2011. New Front-End Readout Chip FE-I 4 - > 200 Mrad. 4 times area of FE-I 3 First chips available mid/late 2010. This will have impact on Pixel Detector Replacement for Phase 2 Upgrades. • ATLAS Forward Physics (AFP) - “Phase 1” Internal ATLAS review process in final stages. Referees questions answered. Some R&D issues to resolve in next 2/3 years– mainly MCP-PMT lifetime of timing detectors. Silicon trackers at 220 m and 420 m from IP. Need and use 3 D silicon technology. Install 220 first then 420 (needs new connection cryostat installed) 220 m 2011/2012 and full AFP operation 2014. Needs upgraded L 1 ECAL trigger for

IBL Performance - From Heinz Pernegger, CERN • • IP res Z: 100 m -> ~60 m IP res R : 10 m -> 7 m B-tagging: Light Jet rejection factor improves by factor ~2 To maintain Pixel Detector performance with inserted layer, material budget is critical. Component Pad size in Z: 250 m % X 0 beam-pipe 0. 6 New-BL @ R=3. 5 cm 1. 5 Old BL @ R=5 cm 2. 7 L 1 @ R=8 cm 2. 7 L 2 + Serv. @ R=12 cm 3. 5 Total 11. 0