Status of the LHCb experiment

Andrei Golutvin ( Imperial & ITEP & CERN )on behalf of the LHCb collaboration

Outline:

Introduction Detector Subsystems Experimental Area Commissioning Cost and funding issues Collaboration matters

RRB April 2009 1

2

List of the 2008-2009 shut-down activities is nearly completed

Full Experimental System Test (FEST) started in December 2008

Exercise Online and Offline systems,Trigger, Monitoring, Data Quality checking and prompt (online)reconstruction

Strategy for trigger and subsequent physics analysis of 2009-2010 Run is being prepared LHCb is fully operational for Physics Run

RRB April 2009

RRB April 2009 3

Beam Pipe (CERN)23 m long beam pipe consisting of 3 Be and 1 stainless steel sections

• Current identified leak fixed by varnishing. The order for a replacement of the third sub-section (UX85/3) was placed, the delivery is expected in 2010

• Development of beam pipe supports with reduced material budget

Magnet (CF)Warm dipole magnet with 4 Tm bending power

along the beam axis

• Since September 2008 the magnet has not been switched on

• Perform magnet test in June 2009 and verify improved position stability of the Trigger Tracker and RICH1

VErtex LOcator (CH,DE,GB,NL,US,RU) 2×21 pairs of Si sensors arranged in 2 halves; each pair consists of one sensor with R-

and one sensor with -strips

4RRB April 2009

• Improvements since last RRB

– Software development – Monitoring – Availability of Spare TELL1 (FPGA) boards– Detector Safety / Interlock System– High rate test is underway– Replacement Modules and Mechanics

• No outstanding problems

SUMMARY :

VELO Monitoring

• Online– Extended Diagnostics

• All in place

– Closing Manager• NEW realtime vertex

monitoring

• Offline– Diagnostic at pit

• 5 fold improvement in speed

• Enhanced reliability and ease of use

5RRB April 2009

VELO TELL1 boards

• TELL1 (84 for VELO)– Receives the trigger and synchronizes

the ADCs to the incoming data. The

data is digitized and processed

• During operation in 2008 a major issue was

availability of spare TELL1 boards and their

reliability

• Reliability & Spares– No new TELL1 failures since 8/2008– Now have ~ 12% spare (10/84)– Additional 3 spares in test beam– Additional 3 used for the spare VELO

module production

6RRB April 2009

VELO Interlock system

• Final “hardware” safety– Prevents operation in modes that could

damage the detector • Triggers on temperature, pressure,

radiation, cooling• Acts on HV, LV, • Manual Emergency button

– Functional (hot) spare existing– Full system up and tested

Detector safety remains of the paramountimportance to VELO

7RRB April 2009

VELO Replacement(The VELO modules will require replacement after about 3 years of running at

full luminosity, due to radiation damage)

• Rebuild VELO under way– Delivery Q3 2010– Modules

(started)– Hoods & transport (started)– Bases (started)– Cables (ordered)– Evaporators– Vacuum Flanges

(ordered)

All hybrids and mechanical components completed ~45% of sensors received

Attain planned output rate by June 20098RRB April 2009

Hybrid 206, first fully bonded complete module with silicon sensors and quartz pitch adaptors.

Silicon sensorsQuartz Pitch Adaptors

Chips

9RRB April 2009

RRB April 2009 10

Outer Tracker – OT (CERN,CF,CN,DE,NL,PL)Three stations with each 4 stereo layers of straw tubes 5 mm diameter and 5m length; 55k channels

OT

CaloMuon

• The installation phase completed in summer 2008 The focus has been on analyzing cosmic and beam data and preparing the detector to 2009/2010 Run

• Production and test of FE electronics completed

• The anti-ageing treatment of detector modules at 40 degrees is carried on. At present the entire C-side and two out of 6 C-Frames on the A-side have been treated. Investigations of the treatment results are ongoing. Final report to collaboration in May 2009

SUMMARY :

Outer Tracker Ageing

The phenomenon is understood:• Rapid gain loss due to Araldite AY103 glue• Characteristic croissant shape due to formation of

O3: Aging downstream is prevented– Confirmed with O3-like gas in input: NO2 prevents ageing

Relative gain loss:

Module length (cm)

M

odul

e w

idth

(ch

.nr) Gas flow

High voltage (V)

O3 concentration at output:

11RRB April 2009

Outer Tracker Ageing: strategyStrategy unchanged:• Flushing removes glue-vapors (slowly)

– Continous flushing since 1.5 year

• Warming the detector increases outgassing– Continue warming detector in situ

• O2 reduces ageing rate– Decide to add few percent O2 to counting gas

• HV training (sometimes) repairs damage

FlushingO2

HV training

Trained half

Flushing

O2

12RRB April 2009

Outer Tracker Ageing: strategyStrategy unchanged:• Flushing removes glue-vapors (slowly)

– Continous flushing since 1.5 year

• Warming the detector increases outgassing– Continue warming detector in situ

• O2 reduces ageing rate– Decide to add few percent O2 to counting gas

• HV training (sometimes) repairs; no damage of the wire surface seen

HV training

Before After

0% O2

2.5% O2

286 mV

319 mVO2

Small (~15%) gain lossNo effect on drift speed

13RRB April 2009

After

Outer Tracker Ageing: in situ test

Heating in situ• Heating at 30-40 0C with elec. blankets• Heated all 6 C-frames on C-side• Heated 2 C-frames on A-side

Monitoring in situ• Installed scanning frame in situ:

T2-Q13-XU

T1-Q13-XU

14RRB April 2009

ST: Trigger Tracker & Inner Tracker(CERN,CH,DE,ES,UA)

TT covers area of 1.41.2 m2; 4 stereo layers with ladders consisting of 3 or 4 chained Si- sensors with strip pitch 183 micron; 143k channelsIT: 3 stations with 4 boxes each arranged around beam pipe; each box has 4 stereo layers x-u-v-x, modules with one or two chained Si-sensors; strip pitch 198 micron; 130k channels

RRB April 200915

• Commissioning of both TT and IT detectors is making good progress

• Broken bonds for TT: majority of the problems occurred a few weeks after module installation and no further problems developed since last RRB - Slow creeping of the pitch adapters to hybrids - Preparing for the production of spares for the case this spreads to other hybrids

• Further progress in the development of detector control and data quality monitoring

SUMMARY :

IT Status

• Service boxes with faults on backplane replaced

• Replacement of weak optical links (low power optical transmitters) ongoing. To be completed in April

• After fixes: ~ 99.8 % detector functional [0.5 % with higher noise]

Working

Faulty fixed[not tested]

Faulty

Noisy fixed[not tested]

Noisy

16RRB April 2009

TT Status

71 low power optical transmitters successfully

replaced [10 more to be done next week ]

Majority of remaining problems related to broken bonds

Will be fixed before start of this years run

One module shows HV problem [breakdown @ 160 V ]

> 99 % of the detector working for this years run

17RRB April 2009

TT Broken Bonds

Innermost row of bonds between hybrid + Pitch Adaptor broken on 8 hybrids • Problem not spread to new modules since

October• Enough spares in hand to replace broken

modules• In affected modules # broken bonds increases• One module, Kapton + pitch adaptor fixed

with araldite, problem does not develop further

Bondwire 1 pitch adapter side • One being used to establish a repair procedure• One placed in burn-in stand + thermally cycled

[effect not reproduced]• All relevant information on production being

collected• Offer for new hybrid production obtained:

- Waiting for better understanding before order placed

Two modules removed and taken to Zurich:

18RRB April 2009

RICH (CERN,CF,GB,IT)RICH1 and RICH2 with 3 radiators covers momentum range 2-100 GeV; RICH1: 5cm aerogel with n=1.03 & 4m3 C4F10 with n=1.0014; RICH2: 100m3 CF4 with n=1.0005;

~500 HPD to readout

RRB April 200919

• Both RICH1 and RICH2 are complete and routinely taking continuous laser data

• Further investigation performed to characterize the effects due to vacuum degradation of some HPDs. Degraded tubes have been replaced or are going to be replaced with spares

• Installation of the Magnetic Distortion Monitoring System for RICH1 has been completed

• RICH reconstruction software, data quality monitoring and calibration procedures are well advanced

SUMMARY :

20

RICH Test pattern

• Regular array of light spots projected

over HPD plane in situ in RICH2 (similar results achieved in RICH1 using motorized stage with LEDs)

• Nicely uniform response and very low

noise for almost all the HPDs. A few show noisy behaviour, peaking around the central axis of the tube

• Due to degraded vacuum quality in

those tubes photoelectrons ionize residual gas Ions then accelerated back to the photocathode producing

further p.e. ion feedback

RRB April 2009

21

Ion feedback

• Ion feedback rate determined using fraction of large clusters (≥ 5 hits)Rate measured regularly for HPDs over the last 18 months

• Most show a linear increase of ion feedback with time, with low gradientThe noisy tubes have a higher gradient

• The bad tubes eventually start to glow, but only after ion feedback rate > 5%

Glow light

RRB April 2009

22

HPD repair

• Failure of tubes can be accurately predicted from the ion feedback measurements

~ 11 HPDs / year predicted to require replacement over the lifetime of the experiment(i.e. 2% / year)

• Removed tubes are successfully repaired by DEP-Photonis. Opened, cleaned, photocathode reapplied: ion feedback gradient is low after repair

Age (days)

RRB April 2009

RICH2 – Status Map (before intervention)

8_181 9_101 9_120 9_115 00 8_127 8_153

9_2 9_110 01 8_121

9_10 9_59 02 8_107

9_71

8_179 03 8_112

8_60 9_114 04

8_147 8_177 05

9_127 06

9_155 9_148 8_135 9_181 07 8_133 8_124 9_189

08 9_208

09 9_137 9_12 8_122

9_265 9_69 10 8_160 9_142 8_197

9_56

9_3 11 8_110 8_134

8_154

12 8_156

9_146 13 8_157 8_178

9_15 9_125 9_250

8_176 14 8_119

9_104

8_126 8_141 15 8_130 9_123 9_11 8_108

C0 C1 C2 C3 C4 C5 C6 C7 C8 A0 A1 A2 A3 A4 A5 A6 A7 A8

27 glowing

10 not glowing IFB>5% already

9 not glowingIFB>5% in 2009

11 not glowingIFB>5% in 2010

57 HPDs

23RRB April 2009

HPD RICH2 replacement summary (intervention march 2009)

• 40 HPDs removed and shipped to Photonis-DEP for repair

• 28 replacement tubes installed– 22 repaired– 1 reference tube– 5 spare tubes from RICH2 spare column

• 17 tubes “at risk” for 2009 and 2010 to be replaced with the new repaired from Photonis-DEP (if back before August)

24RRB April 2009

25

RICH Alignment & calibration• After all corrections, reconstructed

light spot positions match precisely with regular ~ 10 mm grid pattern

• Residual contribution to resolution

s ~ 0.6 mm << pixel size (2.5 mm) correction of magnetic distortion and misalignment is under control

x (mm)

y (

mm

)

RRB April 2009

Calorimeters (CERN,CF,ES,FR,RO,RU,UA) PS/SPD: 12k scint. tiles readout by WLS; ECAL:

6k shashlik cells; HCAL: TILE Calo, 1.5k channels

26RRB April 2009

• The calorimeter system was operational for data taking in September 2008 Detector elements were aligned in time within 3 ns

• In order to improve stability of the phototube response and minimize the noise for ECAL the Cockroft Walton bases of 6000 PMT have been modified during winter shut-down. The stability tests using LED monitoring system have demonstrated much improved performance of ECAL

• The HCAL calibration system with Cs source is used regularly to monitor gain calibration

• Monitoring system for PS/SPD has been commissioned

• Serializers for the trigger path have been modified and exchanged successfully

SUMMARY :

Electromagnetic calorimeter ECAL

6000 Ecal PMT bases modified:

• Improve PMT noise

• Improve radiation resistance

• Improve PMT stability

first hour excluded

(stabilization)

99% of cells are stable

within 1%

1%

Modified detector fully operational from mid-March 200927RRB April 2009

before: rms = 2.9 after: rms = 1.2

Calorimeters calibration

HCAL calibration

• 6 Cesium source runs done• Reproducibility : ~0.5%• Intercalibration ~2%

ECAL calibration

• Derived from LED signal width• Expected pre-calibation gain level ~10%

( precision on HV)

24 cells

Adjusted calibrations will come from 1st data analysis

28RRB April 2009

Muon (CERN,CF,IT,RU)Arranged in 5 SuperLayers; M1 consists of 12 double triple GEM

chambers and 264 MWPC’s; M2-M5 consists of 1104 MWPCs

29RRB April 2009

• M2-M5 stations have been fully commissioned using cosmic rays Performance agrees well with the expectations

• Intense debugging activity to fix a number of problematic readout channels (~4% of the total). All the problems understood and most of them fixed (currently ~0.4% problematic channels)

• Good performance of the HV system. All M2-M5 chambers are at the nominal voltage

• Very detailed schedule of the M1 installation has been devised and is being followed strictly in order to complete in time. The tests of installed M1 chambers are progressing well

SUMMARY :

Commissioning of M2-M5(using ~250k cosmic ray tracks and 19 TED shots in August 2008)

30RRB April 2009

• The TED events were recorded with

HV @2.3 kV, well below the MWPC

plateau for safety reasons

• Useful to:– test the DAQ chain– check the front-end channels,

especially in the inner regions,

poorly illuminated by cosmic rays

• Several missing channels can be seen

in the channel maps debugging/recovery

A snapshot of the muon detectorilluminated by TED events

Status of M2-M5 debugging

• After the cancellation of the LHC run we started a systematic campaign of debugging to fix the problems left open in M2-M5 stations

30/10/2008 30/03/2009 Today

Tot log. channels

Bad log. channels

fraction Bad log. channels

fraction Bad log. channels

fraction

Side A 8352 390 4.7% 65 0.8% 65 0.8%

Side C 8352 257 3.1% 17 0.2% 9 0.1%

Total 16704 647 3.9% 82 0.5% 74 0.4%

Fixed 573 89%

• A few chambers with HV problems have been replaced and retested: all OK

31RRB April 2009

Performance with cosmic rays• Cosmic ray data were collected at a different working point wrt to nominal:

– Higher thresholds– no CF4 (lower gain)– HV at 2.5 kV (instead of 2.65 – 2.70 kV ideal working point)

• Performance is to be compared with that expected @ 2.45 kV Expectations:

– σ(t) =5.3-6.0 ns– Cluster size in 20ns = 1.02-1.1– total eff. >0.99 (in an “infinite” time window)

32RRB April 2009

Not enough statistics for R1

Time resolution from CR data (ns)2.45 kV

Comparison with test beam data

agree very well withcosmic data !!!

Status of M1 installation• Very tight schedule. Exceptional effort by the groups to provide manpower

• First milestone met: all chambers of the 1st layer installed by April 10

• Some delay in the testing/debugging of 1st layer

• Next milestones:

– complete alignment and testing of 1st layer by May 15

– complete the full installation by 15th of July

• Complete test of M1 chambers: gas, LV, HV, connectivity, noise

33RRB April 2009

E lec tronic T es ts

23

22

33

30

10

11

3

6

36

36

33

33

S ide-A F ront

S ide-A B ack

S ide-C F ront

S ide-C B ack

C ompleted T es ting Not Ins talled Today:

• 50% chambersinstalled

• Out of them 78%chambers tested

RRB April 2009 34

RRB April 2009 35

Trigger & Online & Computing(CERN,CF,CH,DE,FR,ES,GB,IT,NL)

• Full scale HLT testing in realistic environment using FEST • Change of strategy for the last phase of HLT (HLT2) Replace “OR” of many exclusive channels with more inclusive approach

• CPU Farm is being completed to allow the full 1 MHz readout at nominal event size. The final upgrade is foreseen for beginning of 2010. The readout network will be completed to full capacity during summer 2009

• Completing DIRAC3 commissioning for both production and user analysis activities. Continue development of data access system

• Regular FEST weeks are essential to test data distribution, reconstruction and data quality checks

Global Commissioning(Status of LHCb detector alignment)

Collection of survey data in geometry/conditions database done

Extensive tests of track-based alignment algorithms on MC still ongoing

Main emphasis right now: improve alignment of LHCb detector with first 'real' data

Cosmics

- relatively small yield ('horizontal cosmics'), but still 'millions' of events

- alignment of IT, OT (enormous progress), CALO and MUON

TED events

- short runs (August, September 2008), high track density

- about 1500 VELO tracks and as many good IT tracks

- alignment of VELO, TT and IT Other efforts: FEST, express stream trigger

36RRB April 2009

SUMMARY :

VELO alignment with TED data(TED tracks perfect for VELO alignment: cross detector almost parallel to z)

Translation in X of VELOmodules extracted fromtwoTED runs - good agreement between

two runs (change with respect to survey less than 20 micron)

reso

luti

on in

μ

pitch in μ

z of module [cm]

Resolution estimated fromVELO hit residuals agreeswell with expectations

Further improvement possible

37RRB April 2009

R strips Φ strips

IT Alignment• Studies used data from the LHC synchronization tests - occupancy 20 times that of normal LHCb running - ghost rate ~ 10 % [not ideal for alignment]• Alignment down to granularity of ladders performed

Bias 30 mCore ~120 m

Bias 3 mCore 96 mAlign

Survey - pre-alignment in x Box alignment TxTyRzLayer TxRz, Ladder Tx

Tail due to ghosts

(For 10 GeV tracks expect 72 m)

Unbiasedresiduals

38RRB April 2009

Typical example layer IT2X1:

TT-Velo and TT-IT alignment

TED tracks in IT and VELO can also be matched to TT

Residuals of TT hits with respect to tracks in the IT before and after internal alignment of IT

remaining shift is possibly due to an IT shearing

Residuals of TT hits with respect to tracks in the VELO before and after a global alignment of TT

using TT ladder surveynot using TT ladder survey

next step: relative alignment of 3

systems using 'long' tracks

before

after

39RRB April 2009

FEST and express stream(VELO alignment / monitoring procedure tested as part of FEST)

Set values of velo positioning system such that data 'looks' misaligned Use express stream data to extract new alignment constants

Final aim: Use calibration and alignment farm to run alignment within hours after

data taking Update alignment before GRID event processing if monitoring tells that

old alignment was not good enough

difference in X coordinate of primary vertex reconstructed in left and right VELO as seen by monitoring before and after alignment with express stream

40RRB April 2009

RRB April 2009 41

Cost and Funding

Overall cost remains unchanged at 75 MCHFThe underfunding in 2005 of 2.6 MCHF for the DAQ CPU Farm, hasbeen covered through extra contributions (BMBF Brazil, France,MPI Germany, Spain, UK, US and CERN)

Collaboration Matters

• Two new groups from UK accepted:

- University of Manchester led by Prof. David Bailey - University of Warwick led by Prof. Tim Gershon

Service contribution of both groups was extensively discussed and well defined

RRB April 2009 42

LHCb Collaboration today:

In total: 702 members 15 countries 52 institutes

RRB April 2009 43

Possible Physics with 2009/2010 dataWith 0.3 fb-1 LHCb should be able to improve Tevatron results(expected for 9 fb-1) on key observables in flavour physics:

90% C.L. exclusion limits at 8 TeV CMB

R(B

s0 μ+

μ- ) (

x10-9

)

Mixing phase ϕs BR(Bs μμ)

RRB April 2009 44

Conclusion

LHCb is well prepared for the 2009/2010 Physics Run

We need luminosity in order to be competitive with Tevatron