The LHCb VErtex LOcator q Introduction What is
The LHCb VErtex LOcator q Introduction: – What is LHCb ? – Physics goals – Experimental needs q Designing the VELO: – Role of VELO – LHCb requirements – Machine integration constraints q VELO sensors: – Design – Operational conditions – Performance, latest-beam q Integration with LHC machine – – q Cooling Positioning Vacuum RF Outlook CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi / CERN On behalf of VELO group & LHCb Collaboration Massimiliano Ferro-Luzzi 1
B production, LHCb overview q B B q q LHCb => precision study of B-meson decays LHC: large production of BB pairs at small and correlated polar angles Build LHCb as a forward spectrometer Muons ECAL HCAL Tracker B B IP 8 at LHC VELO CERN Detector Seminar Magnet 8 -April-2005 CERN Massimiliano Ferro-Luzzi 2
LHCb goal q LHCb will look for (abnormal? ) CP violation and rare decays by precision studies of the B-meson systems q q CP violation in one line: Rate( i f ) For many channels, measuring the proper-time = L m. B / p. B of the decaying B-meson is required ( oscillations ! ), hence measure momentum p. B and flight distance L. proper time dependent rates (asymmetries) Rate( i ) – Rate( i ) Example: f = f ACP = Rate( i ) + Rate( i ) CERN Detector Seminar 8 -April-2005 CERN = Adir cos( m ) + Amix sin( m ) Massimiliano Ferro-Luzzi 3
B events and proper time q q q B life time 1. 5 ps Momenta = 10 … 100 Ge. V/c Flight distance = 1 … 10 mm + Bs Ds p p Flavour tag - Bq D 0 Fast oscillations for Bs ! require excellent vertex resolution K+ K - = 38 fs 0. 25 ps => 140 um at 10 Ge. V/c CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 4
Primary Vertex Resolution q Primary vtx resolutions z 50 um q x, y 8 um B-decay vtx resolution ~ 150 um (depending on decay channel) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 5
Impact parameter and momentum resolutions B ~150 um ~10 um CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 6
Mass resolution q Mass resolution depends also on VELO performance, especially for lower momenta CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 7
Triggering on B meson events Signal rate vs Minimum Bias rate: q “visible” inelastic event rate 10 MHz q “reconstructable” BB event rate 15 k. Hz q “interesting” channels: BR ~ 10 -3 … 10 -9 !! How to find these B events at run time ? LHCb n = 1 At 2 x 1032 cm-2 s-1 !! x 1/1000 or less… Use B signatures: 1. B-hadrons are heavy high PT 2. B hadrons are “long-lived” tracks with large impact parameter relative to primary vertex n = 2 n = 3 PT of charged hadron with highest PT in the event n = 4 n = 5 CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 8
An event with B hadrons … q q RZ view in VELO (R>0 by definition) Green = all charged MC particles with mass < 1 Ge. V Two primary vertices… lots of tracks… CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 9
… and the same event with an impact parameter cut q q Red = charged MC particles from B hadrons, with mass<1 Ge. V Cyan = tracks with IP > 0. 1 mm CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 10
Role of VELO in LHCb 40 MHz Level-0 PT, PU, multiplicity 1 MHz Level-1 “software” VELO PT, impact param 40 k. Hz High-Level Trigger Everything (except RICH) 2 k. Hz q 20 ms 1 ms 1 us “hardware” TRIGGER – Find 1 st prim vtx, then 2 nd prim vtx ? – Veto events with “large” 2 nd vtx – Reject high multiplicity events q Reconstruct primary vertices (2 D tracking, RZ) Find high impact parameter tracks w. r. t. prim vtx Extrapolate these in 3 D to Trigger Tracker (for PT) Match with muon candidates q Vertex reconstruction: q q q – Primary, secondary vertices – Measure precisely flight distances q Tracking before spectrometer – Determine track angles for momentum measurement – Tracks used in RICH 1 reconstruction: VELO performance affects RICH 1 performance ! Offline CERN Detector Seminar Pile-Up rejection (RZ, backward stations): 8 -April-2005 CERN Massimiliano Ferro-Luzzi 11
VELO Design « Guidelines » CONSTRAINTS / DEMANDS q IMPLICATIONS Best possible resolution: – As close as possible to vertices – Minimize amount of material q Put sensors in vacuum Develop fast, low-noise ASIC, optimize sensor/strip geom Use R-Phi strip geometry Level-1: – – q Cluster efficiency > 99% Spill-Over(+25 ns) < 30% Noise occupancy < 0. 1% Fast computation of prim vertices and impact parameters LHC: – Injection clearance Ø > 54 mm – RF fields, impedance budget – Dynamic vacuum phenomena: § Ion/ photon-induced desorption § electron multipacting – Large radiation fluences CERN Detector Seminar 8 -April-2005 CERN Detector must be retractable Guide mirror charge (wake field suppression, RF shielding) Separate beam vacuum from detector vacuum, NEG coating Use « Rad hard » detectors, cool down sensors Massimiliano Ferro-Luzzi 12
A very brief comparison Area (m 2) Chann. (106) Trigger Precision (um) Beam at (cm) Dose ATLAS SCT strips CMS Pixels ILC CCD LHCb VELO 60 6 No 20 30 0. 2 2 33 No 12 4 3. 0 0. 3 800 N/A 4 1. 5 0. 01 0. 3 0. 2 Yes! 5 0. 7 1. 3 (1014 neq yr-1 cm-2) => VELO: small, yes, but … high precision at high fluence! CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 13
Detector: Chosen Layout d ra 0 m q 36 IP 8 q q p 21 stations Each station measures R and Phi +2 « Pile-Up » stations (R only) Analogue read-out Left-Right staggered in Z to allow for overlap r-sensors • 4 x 45 o sectors • 2048 strips • Pitch: 40 to 100 um 42 mm Retractable by 30 mm • 10 o-20 o stereo angle • Inner and outer region • 2048 strips • Pitch: 35 to 100 um p CERN Detector Seminar 8 mm -sensors 35 78 39 97 pit um ch Both use 2 nd metal layer 8 -April-2005 CERN Massimiliano Ferro-Luzzi 14
Strip Pitches, Lengths, Occupancies TDR RF foil => truncate Now Segmentation: helps pattern recognition and gives rough measure of phi (trigger) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 15
Radiation damage effects q Change of depletion voltage – Complicated time and temperature dependence q Increase of leakage current , I = x Fluence x Volume – Increased noise – Possible thermal runaway Decrease of charge collection efficiency Depletion voltage [V] q Fluence R. Wunstorf, Ph. D. thesis, DESY FH 1 K-92 -01 M. Moll DESY-THESIS-1999 -040 CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 16
The environment q Radiation damage mostly from p-p collisions (i. e. pions, few Ge. V … few 100 Ge. V) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 17
Time dependence of depletion voltage q q Curves: expectation from a simple model Data: measured with an early VELO prototype Design goal At 107 seconds/year (2 fb-1/yr) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 18
VELO: Full-depletion voltage after irradiation q q Very inhomogeneous irradiation across VELO sensor ! Most crucial strips for precision (small R) are the most exposed ! Middle station Far station an example Vdep, 200 um? Not type–inverted Vdep q Typeinverted After one year, will have to operate part of the sensors under- or overdepleted R/cm CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 19
Effect of irradiation on microstrip sensors (p-on-n) q Fully depleted q Underdepleted – undepleted region is next to the implants – generate less charge – spread over more strips CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 20
n side signal n+ implants CERN Detector Seminar 8 -April-2005 CERN Resolution / um Effect of Irradiation on Microstrip Sensors: p-on-n versus n-on-n Massimiliano Ferro-Luzzi p-side n-side 21
And now with second metal layer … q q As expected, charge collection decreases with increasing irradiation But in addition: charge seen in 2 nd metal layer … A lot ! Up to 20 % ! CERN Detector Seminar 8 -April-2005 CERN => choose n-on-n Massimiliano Ferro-Luzzi 22
Detector module design light-weight, radhard, high-vacuum compatible R sensor 300 um Kapton hybrids ~200 um Phi sensor 300 um Carbon fiber 2 x 200 um TPG 500 um CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 23
Detector Module Prototype q q q R silicon sensor Beetle FE chips Picth adapters CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 24
Test Beam Nov/2004 q Measure: – – q Signal/Noise Efficiencies Resolutions Etc. Telescope: – VA 2 readout – n-on-n PR 01 sensors – 300 um R/Phi sc 2 Phi/R R/Phi Phi/R sc 1 Beam 120 Ge. V pions Test sensor CERN Detector Seminar Compare 200 um vs 300 um thick sensors 8 -April-2005 CERN Massimiliano Ferro-Luzzi 25
Telescope performance CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 26
Front-end chip LHCb: 160 x 25 ns deep, read out in 900 ns => SCTA not OK Design a new chip => the Beetle: q 0. 25 um CMOS ASIC q Used in LHCb by VELO, Pile-Up system, Silicon tracker CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 27
Beetle pulse shape from test beam data q q Asynchronous beam particles Mean ADC versus TDC => pulse shape typ. ~15 ns ADC TDC / ns CERN Detector Seminar 8 -April-2005 TDC / ns CERN Massimiliano Ferro-Luzzi 28
Noise and signal distributions R-sensor 300 um R-sensor 200 um separation S/N VELO requirements: q Cluster efficiency > 99% q Spill-Over < 30% q Noise occupancy < 0. 1% CERN Detector Seminar 8 -April-2005 Can we meet these ? CERN Massimiliano Ferro-Luzzi 29
Detector Performance PRELIMINARY q S/N cut: – Efficiency puts an upper limit on S/N cut value – Noise and spill-over put a lower limit q Spill-over stronger constraint than noise ? Efficiency Spill-over/350 Noise occupancy Spill-over for 300 um yet to be finalized R 200 um CERN Detector Seminar 8 -April-2005 R 300 um CERN Massimiliano Ferro-Luzzi 30
Charge sharing for perpendicular tracks q q L R Eta = (signal Left strip) / (signal Left strip + signal Right strip) Pos = (distance to Left strip) / pitch large pitch Pos Pos fine pitch Prototype “PR 04” Under investigation - analysis bug ? - integration time? - actual det angle ? - … ? ? Prototype “PR 01” Eta Eta CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 31
Charge sharing for inclined tracks PRELIMINARY q q q Most important for LHCb: R-sensors B-track average polar angle ~5 deg For R-sensors, tracks will be inclined ! 200 um CERN Detector Seminar 8 -April-2005 CERN 300 um Massimiliano Ferro-Luzzi 32
Detector integration q q Cooling Mechanics Vacuum Wake fields CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 33
Cooling Block q q In the detector vacuum Coolant: mixed-phase CO 2 – Excellent for capillaries – Radiation friendly 1. 5 mm stainless steel capillary CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 34
Temperature Gradients q TSi Tcapillary + 8 K + 4 K Substrate “transverse” resistance Substrate “longitudinal” resistance Substrate to silicon -10 C 24 W of heat -21 C Capillary at about -30 C Aluminium blocks Carbon block CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 35
Positioning the VELO Detector mounted to movable support Flanges fixed to vacuum vessel m 30 m q q Trigger requires “axial symmteric” geometry ! RZ tracking ! Determine position of beams relative to each detector half using collisions and VELO tracks Move halves accordingly Center the luminous region to better than 100 um CERN Detector Seminar 8 -April-2005 CERN m 30 m be am s +/-5 mm Massimiliano Ferro-Luzzi 36
Large rectangular bellow q q q Custom-made First full-size prototype finished Even leak-tight ! CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 37
Detector vacuum box More than just a vacuum box… It must q be transparent… massless … non -existing… q allow overlap of left sensors with right sensors q shield sensors+electronics against RF q smoothly guide the beams mirror charge q be LHC-UHV compatible q suppress dynamic vacuum phenomena q suppress electron multipacting CERN Detector Seminar 8 -April-2005 CERN ~120 cm Massimiliano Ferro-Luzzi 38
Making the box ~ 300 um thick ~ 150 um Extensive R&D q Forming the foil (Al. Mg 3): – Superplastic deformation – Cold forming – Hot-gas forming q q q Leak-tightness of foil Leak-tightness of welded box Stiffness, rupture pressure q Applying coatings: – Detector side: insulation (Torlon) – Beam side: reduce secondary electron yield, desorption (NEG) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 39
Wake Field Suppression q q q Resonant modes of the VELO structures Simulated VELO shielding scheme in MAFIA Varied corrugation depth d Determined acceptable depth d < ~20 mm 100 W heat for a 0. 5 A beam CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 40
Attenuation of high-frequency fields How thin can the box be ? LHC beam (simplified): fields at the wall surface (beam side) Fourier decomposition (frequency spectrum) Through the wall: q Main component is q Tangential electric field ~ Al@40 MHz: ~ 10 -5 For 100 um, attenuation ~ 4. 5 x 10 -4 from 27 A/m down to 0. 01 A/m Tangential electric from 0. 1 V/m to 5 x 10 -5 V/m CERN Detector Seminar 8 -April-2005 CERN low frequencies … Massimiliano Ferro-Luzzi 41
Smooth conducting structures… RF box corrugations Wake field suppressor 70 um Cu. Be CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 42
The first prototype box q q CERN Detector Seminar 8 -April-2005 CERN Al. Mg 3 Welded 300 um Al on 300 um Al Shape homogeneity measured to be within ~200 um over ~ 1 m Measured/calculated differential pressure deformation (elastic) Massimiliano Ferro-Luzzi 43
Vacuum system Maintain pressure difference < 5 mbar NEG coating !! Do not contaminate! CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 44
Deflections due to pressure difference Measurements: FEA <-> measurements typically agree to ~50% FEA calculations: During venting/pumping down, we control the pressure difference to < 5 mbar defined nominal distance to Si (1. 2 mm) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 45
A rather complex shape Redesigned Si sensor corners to match best possible bending radius of detector box CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 46
Material in the acceptance = pseudorapidity = -ln(tan( /2)) = polar angle = azimuthal angle CERN Detector Seminar Material traversed by particles in 2 < < 5 (…): material before the 1 st measured hit 8 -April-2005 CERN Massimiliano Ferro-Luzzi 47
The VELO (still virtual) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 48
The VELO (getting real) CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 49
The VELO group Lausanne CERN Detector Seminar 8 -April-2005 CERN Massimiliano Ferro-Luzzi 50
- Slides: 50