1

Luminosity measurement in the ATLASLuminosity measurement in the ATLASexperiment at the LHCexperiment at the LHC

26 November 2008Per GrafstromCERN

2

3

Electroweak symmetry breakingThe HiggsSuper symmetryCP-violationExtra dimensions, black holes

ATLAS

415-April-2008 ATLAS RRB4

ATLAS Collaboration

(Status April 2008)

37 Countries 167 Institutions 2200 Scientific Authors total(1750 with a PhD, for M&O share)

Albany, Alberta, NIKHEF Amsterdam, Ankara, LAPP Annecy, Argonne NL, Arizona, UT Arlington, Athens, NTU Athens, Baku, IFAE Barcelona, Belgrade, Bergen, Berkeley LBL and UC, HU Berlin, Bern, Birmingham, UAN Bogota, Bologna, Bonn, Boston, Brandeis,

Bratislava/SAS Kosice, Brookhaven NL, Buenos Aires, Bucharest, Cambridge, Carleton, Casablanca/Rabat, CERN, Chinese Cluster, Chicago, Chile, Clermont-Ferrand, Columbia, NBI Copenhagen, Cosenza, AGH UST Cracow, IFJ PAN Cracow, DESY, Dortmund, TU Dresden, JINR

Dubna, Duke, Frascati, Freiburg, Geneva, Genoa, Giessen, Glasgow, Göttingen, LPSC Grenoble, Technion Haifa, Hampton, Harvard, Heidelberg, Hiroshima, Hiroshima IT, Indiana, Innsbruck, Iowa SU, Irvine UC, Istanbul

Bogazici, KEK, Kobe, Kyoto, Kyoto UE, Lancaster, UN La Plata, Lecce, Lisbon LIP, Liverpool, Ljubljana, QMW London, RHBNC London, UC London, Lund, UA Madrid, Mainz, Manchester, Mannheim, CPPM Marseille, Massachusetts, MIT, Melbourne, Michigan, Michigan SU, Milano,

Minsk NAS, Minsk NCPHEP, Montreal, McGill Montreal, FIAN Moscow, ITEP Moscow, MEPhI Moscow, MSU Moscow, Munich LMU, MPI Munich, Nagasaki IAS, Nagoya, Naples, New Mexico, New York, Nijmegen,

BINP Novosibirsk, Ohio SU, Okayama, Oklahoma, Oklahoma SU, Oregon, LAL Orsay, Osaka, Oslo, Oxford, Paris VI and VII, Pavia, Pennsylvania, Pisa, Pittsburgh, CAS Prague, CU Prague, TU Prague, IHEP Protvino, Regina, Ritsumeikan, UFRJ Rio de Janeiro, Rome I,

Rome II, Rome III, Rutherford Appleton Laboratory, DAPNIA Saclay, Santa Cruz UC, Sheffield, Shinshu, Siegen, Simon Fraser Burnaby, SLAC, Southern Methodist Dallas, NPI Petersburg, Stockholm, KTH Stockholm, Stony Brook, Sydney,

AS Taipei, Tbilisi, Tel Aviv, Thessaloniki, Tokyo ICEPP, Tokyo MU, Toronto, TRIUMF, Tsukuba, Tufts, Udine/ICTP, Uppsala, Urbana UI, Valencia, UBC Vancouver, Victoria, Washington, Weizmann Rehovot, FH Wiener Neustadt, Wisconsin, Wuppertal, Yale, Yerevan

5

The very first Event in ATLASThe very first Event in ATLAS

6

Motivation-why we need to measure the Motivation-why we need to measure the luminosityluminosity

Measure the cross sections for “Standard “ processes Top pair production Jet production ……

New physics manifesting in deviation of x BR

relative to the Standard Model predictions. Precision measurement becomes more

important if new physics not directly seen. (characteristic scale too high!)

Important precision measurements Higgs production x BR tan measurement for MSSM Higgs ……. Relative precision on the measurement of HBR for various

channels, as function of mH, at Ldt = 300 fb–1. The dominant uncertainty is from Luminosity: 10% (open symbols), 5% (solid symbols).

(ATLAS Physics TDR , May 1999)

Theoretically known to ~ 10 %

Higgs coupling

Instantaneous LuminosityorLdt Integrated Luminosity

7

Absolute versus Relative measurementAbsolute versus Relative measurement Relative measurements or Luminosity Monitoring

Using suitable observables in existing detectors Beam condition monitor Current in Tile calorimeter PM’s Minimum bias scintillators

Using dedicated luminosity monitor LUCID

Absolute measurements Several different methods-next slide

Strategy:

1. Measure the absolute luminosity with a precise method at optimal conditions2. Calibrate luminosity monitor with this precise measurement and then use the calibrated monitor at all conditions

8

Absolute Luminosity MeasurementsAbsolute Luminosity Measurements Goal: Measure L with ≲ 3% accuracy (long term goal)

How? Three major approachesLHC Machine parameters

Rates of well-calculable processes:e.g. QED (like LEP), EW and QCD

Elastic scatteringOptical theorem: forward elastic rate + total inelastic rate:

Luminosity from Coulomb Scattering

Hybrids

Use tot measured by others

Combine machine luminosity with optical theorem

We better pursue all options

9

OutlineOutline

Methods for Absolute Measurement of Luminosity Processes with known cross sections

Machine Parameters

Elastic scattering

Methods for Relative Measurement of Luminosity LUCID

10

Two photon production of muon pairs-Two photon production of muon pairs-QEDQED

p

p

• Pure QED• Theoretically well understood• No strong interaction involving the muons• Proton-proton re-scattering can be controlled• Cross section known to better than 1 %

Muon pairs

11

Two photon production of muon pairsTwo photon production of muon pairs

+

-

Pt 3 GeV to reach

the muon chambers

Pt 6 GeV to maintain

trigger efficiency andreasonable rates

Centrally produced 2.5

Pt() 10-50 MeV

Close to back to backin (background suppression)

Muon pairs

12

BackgroundsBackgrounds

Strong interaction of a single proton

Strong interaction between colliding proton

Di-muons from Drell-Yan production

Muons from hadron decay

Muon pairs

13

Event selection-two kind of cutsEvent selection-two kind of cuts

Kinematic cuts Pt of muons are equal within 2.5 σ

of the measurement uncertainty

Good Vertex fit and no other charged track

Suppress Drell-Yan background and hadron decays

Suppresses efficiently proton excitationsand proton-proton re-scattering

Muon pairs

14

What are the difficulties ?What are the difficulties ? The resolution

The pt resolution has to be very good in order to use the Pt() 10-50 MeV cut. The rate

The kinematical constraints σ 1 pbA typical 1033/cm2/sec year 6 fb -1 and 150 fills 40 events fill Luminosity MONITORING excludedWhat about LUMINOSITY calibration?1 % statistical error more than a year of running

Efficiencies Both trigger efficiency and detector efficiency must be known

very precisely. Non trivial. Pile-up

Running at 1034/cm2/sec “vertex cut” and “no other charged track cut”will eliminate many good events

CDF result First exclusive two-photon observed in e+e-. …. but….16 events for 530 pb-1 for a σ of 1.7 pb overall efficiency 1.6 %

Summary – Muon PairsCross sections well known and thus a potentially precise method.However it seems that statistics will always be a problem.

Muon pairs

15

W and Z countingW and Z counting

∣y∣W ±

∣η∣ l±

W and Z

16

W and Z countingW and Z counting Constantly increasing precision of QCD calculations makes counting of

leptonic decays of W and Z bosons a possible way of measuring luminosity. In addition there is a very clean experimental signature through the leptonic decay channel.

The Basic formula

Ldt is the integrated luminosityN is the number of W or Z candidatesB is the number of back ground events is the efficiency for detecting W or Z decay productsa is the acceptance

th is the theoretical inclusive cross section

W and Z

17

Uncertainties on Uncertainties on thth

th is the convolution of the Parton Distribution Functions (PDF) and of the partonic cross section

The uncertainty of the partonic cross section is available to NNLO in differential form with estimated scale uncertainty below 1 % (Anastasiou et al PRD 69, 94008.)

PDF’s more controversial and complex

W and Z

18

NNLO Calculations

Bands indicate the uncertainty from varying the renormalization (R) and factorization (F) scales in the range: MZ/2 < (R = F) < 2MZ

At LO: ~ 25 - 30 % x-s error At NLO: ~ 6 % x-s error At NNLO: < 1 % x-s error

Anastasiou et al., Phys.Rev. D69:094008, 2004

Perturbative expansion is stabilizing and renormalization and factorization scales reduces to level of 1 %

W and Z

19

Sensistive to x values10-1 > x > x10-4

Sea quarks and antiquark dominatesgqqbar

Gluon distribution at low x

HERA result important

x and Q2 range of PDF’s at LHCW and Z

20

Sea(xS) and gluon (xg) PDF’sSea(xS) and gluon (xg) PDF’s

PDF uncertainties reduced enormously with HERA.Most PDF sets quote uncertainties implying errorin the W/Z cross section 5 %However central values for different sets differs sometimes more !

W and Z

21

Uncertainties in the acceptance aUncertainties in the acceptance aThe acceptance uncertainty depends on QCD theoretical error.

Generator needed to study the acceptance

The acceptance uncertainty depends on PDF,s , Initial State Radiation, intrinsic kt…..

Uncertainty estimated to about 2 -3 %

Uncertainties on

Uncertainty on trigger efficiency for isolated leptons

Uncertainty on offline reconstruction efficiencies

Estimation using tag and probe method 50 pb-1 of data 3 %

1 fb-1 < 1 %

W and Z

22

Summary – W and ZSummary – W and Z

W and Z production has a high cross section and clean experimental signature making it a good candidate for luminosity measurements.

The biggest uncertainties in the W/Z cross section comes from the PDF’s. This contribution is sometimes quoted as big as 8 % taking into account different PDF’s sets .Adding the experimental uncertainties we end up in the 10 % range.

The precision might improve considerably if the LHC data themselves can help the understanding of the differences between different parameterizations

The PDF’s will hopefully get more constrained from early LHC data .

The experimental errors will also diminish with time

Aiming at 3-5 % error in the error on the Luminosity from W/Z cross section after a couple of years of data taking.

W and Z

23

Luminosity from Machine parametersLuminosity from Machine parameters Luminosity depends exclusively on beam parameters:

Depends on frev revolution frequency

nb number of bunches

N number of particles/bunch * beam size or rather

overlap integral at IP

Machine parameters

•The luminosity is reduced if there is a crossing angle ( 300 µrad )•1 % for * = 11 m and 20% for * = 0.5 m

24

Luminosity accuracy limited by

extrapolation of x, y (or , x*, y*) from measurements of beam profiles elsewhere to IP; knowledge of optics

Precision in the measurement of the bunch current

beam-beam effects at IP, effect of crossing angle at IP, …

Machine parameters

25

What means special effort?What means special effort?

Calibration runs with simplified LHC conditions

Reduced intensity Fewer bunches No crossing angle Larger beam size ….

Simplified conditions that will optimize the condition for an accurate determination of both the beam sizes (overlap integral) and the bunch current.

Calibrate the relative beam monitors of the experiments during those dedicated calibration runs.

Machine parameters

26

Determination of the overlap integralDetermination of the overlap integral(pioneered by Van der Meer @ISR)(pioneered by Van der Meer @ISR)

Machine parameters

27

Example LEP Example LEP Machine parameters

28

Summary – Summary – MachineMachine parameters parameters

Special calibration runs will improve the precision in the determination of the overlap integral . In addition it is also possible to improve on the measurement of N (number of particles per bunch). Parasitic particles in between bunches complicate accurate measurements. Calibration runs with large gaps will allow to kick out parasitic particles.

Calibration run with special care and controlled condition has a good potential for accurate luminosity determination.

Less than ~5 % might be in reach at the LHC (will take some time !)

Ph.D student in the machine department is working on this (supervisor Helmut Burkhardt)

Machine parameters

29

Elastic scattering and luminosityElastic scattering and luminosity

Elastic scattering has traditionally provided a handle on luminosity at colliders.

Optical theorem

σtot = 4π Im fel (0)

The basis for this is the Optical Theorem which relates the total cross-section to the forward elastic scattering amplitude

The optical theorem is a general law of wave scattering theory derived from conservation of probability using quantum mechanics. It is a reflection of the fact that the very existence of scattering requires scattering in the forward direction in order to interfere with the incident wave, and thereby reduce the probability

current in this direction.

30

The Optical theorem can be used in several ways for luminosity determination.

Method 1Extrapolate elastic scattering to t=0 (the optical point) and in addition measure the total rate.

Method 2 Extrapolate elastic scattering to t=0 and use a measurement

of total from another experiment.

Method 3Measure elastic scattering as such small angles that

the cross section is also sensitive to the Coulomb part.

Optical theorem

31

Optical theorem

σtot = 4π Im fel (0)

Ntot = tot · L

where = Re fel (t=0)/Im fel(t=0)

Thus we need to

•Extrapolate the elastic cross section to t=o•Measure the total rate•Use best estimate of ( ~ 0.13 +- 0.02 0.5 % in L/L )

Method 1: Extrapolate elastic scattering to t=0 (the optical point) and in addition measure the total rate.

32

Extrapolate to t=0

Exponential region

Optical theorem

33

Measure the total rateMeasure the total rate

Ntotal=Nel +Ninel

Ninel = Nnd

+Nsd+Ndd

Trigger efficiencies with MBTS

ATLAS covers an range up to 5

Optical theorem

ND 99 %DD 54%SD 45 %

34

Measure the total rate (cont.)Measure the total rate (cont.)

PHOJET 7 % of the total rate not triggered

PHYTIA12.5 % of the total rate not triggered

Taking this as extremes we would have a 5 % error in estimates of the total rate and this dominates and gives us about 10 % error in Luminosity

Optical theorem

Tuning PHOJET or PYTHIA with real data (here ALFA will be important)will improve this considerably….however difficult to be quantitative withoutsimulations. Theses studies have started.

If the diffractive rate could be estimated at the 10% level the total rateestimate would be at the level of 2% and the luminosity number could be better than 5 %

35

* uncertainty 0.13 +- .02 0.5 % in L*extrapolation to t =0 1 % in L * tot from Totem 1-2 % 2-4 % in L

5 % in luminosity possible

Method 2: Extrapolate to t=0 and use the optical theorem in combination with and independent measurement of tot

Optical theorem

36

Method 3: Optical theorem and measuring of elastic scattering at very small angles

Measure elastic scattering at such small t-values that the cross section becomes sensitive to the Coulomb amplitude

Effectively a normalization of the luminosity to the exactlycalculable Coulomb amplitude

No total rate measurement needed

UA4 used this method to determine the luminosity to 2-3 %

Coulomb

37

Elastic scattering at very small anglesElastic scattering at very small anglesCoulomb

38

ATLAS Roman PotsATLAS Roman Pots

• Absolute• Luminosity• For • ATLAS

Coulomb

39

What is needed for the small angle elastic What is needed for the small angle elastic scattering measurements?scattering measurements?

Special beam conditions

“Edgeless” Detector

Compact electronics

Precision Mechanics in the form of Roman Pots to approach the beam

Coulomb

40

Mechanics - the pot itselfMechanics - the pot itself

ATLASspecific

Roman Pot concept

Coulomb

Beam

41

Mechanics - the Roman Pot unitMechanics - the Roman Pot unit

Profiting fromAB collimation group

Reproducible and precisemovements required

Coulomb

42

The detectors-fiber trackerThe detectors-fiber tracker

Choice of technology:• minimum dead space• no sensitivity to EM induction from beam• resolution ~ 30 m

Concept• 2x10 U planes 2x10 V planes• Scintillating fibers 0.5 mm2 squared• 64 fibers per plane• Staggered planes• MAPMT readout

Coulomb

43

Test beam campaigns-DESY and CERNTest beam campaigns-DESY and CERN

ALFA

A number of prototypes withlimited amount of fibers were tested

Main results:• Light yield ~ 4 p.e.• resolution ~ 25 m• non-active edge 100 m

Coulomb

44

The ALFA electronicsThe ALFA electronics

ATLAS

TDAQ

Coulomb

45

The ALFA electronics – front end - PMFThe ALFA electronics – front end - PMF

Adjustable amplifier-MAPMT Adjustable threshold High speedSmall cross talk Compact

Requirements

Coulomb

46

The beam conditionsThe beam conditionsNominal divergence of LHC is 32 rad We are interested in angles ~ x 10 smaller high beta optics and small emittance(divergence / β* )

y*

y*

parallel-to-point focusingydet

IP Leff

Zero crossing angle fewer bunches

Insensitive to vertex smearinglarge effective lever arm Leff

To reach the Coulomb interference region we willuse an optics with β* ~ 2.6 km and N ~ 1 m rad

β [m

]

D [m

]

Compatiblewith TOTEM

High β* and few bunches low luminosity

Coulomb

Amount of beam halo very important

47

Summary - CoulombSummary - Coulomb Getting the Luminosity through Coulomb normalization will be extremely

challenging due to the small angles and the required closeness to the beam.

Main challenge is not in the detectors but rather in the required beam properties

Will the optics properties of the beam be know to the required precision?

Will it be possible to decrease the emittance as much as we need?

Will the beam halo allow approaches in the mm range?

No definite answers before LHC start up

UA4 achieved a precision using this method at the level of 2-3 %but at the LHC it will be harder .....

Coulomb

48

Measure the LHC luminosity.

Count the number of charged particles per BX, pointing to the primary pp interactions.

LUCID in ATLASLUCID in ATLAS

Monte Carlo simulationTwo symmetrical arms at17 m from the pp interaction

region.

Relative measurement

49

Relative measurement

50

LUCID locationLUCID location

Expected dose: 7 Mrad/year @ higest luminosity (1034 cm-2s-1)

Relative measurement

51

LUCID detectorLUCID detector

Array of mechanically polished Aluminum tubes in a Cherenkov gas (C4F10).

CC44FF1010 pressure mantained at 1.25/1.5 bar (Leak <10 mbar/day). pressure mantained at 1.25/1.5 bar (Leak <10 mbar/day).

coverage: 5.6, 6.0

Relative measurement

52

Optical fibers (PUV700) via Winston Cone to multi-anode PMT (Hamamatsu H7546B).

Better for high luminosity runs (MAPMT not exposed to high radiation doses).

Direct coupling to Photo-Multiplier Tubes (PMT, Hamamatsu R762).PMT must be radiation hard.

Cherenkov Tube PMT

PMT quartz window (1 mm)

Cherenkov Tube

MAPMTWinston cone

Fiber bundle

15 mm

15 mm

Relative measurement

Read-Out schemeRead-Out scheme

53

Overall conclusionsOverall conclusions

To start with we will work with LHC Machine parameters In the beginning 20-25 % precision.

Special calibration runs with simplified conditions will improve :

maybe 10 % after some time and then gradually towards 5 %

Rates of well-calculable processes will be next step: Two photon production of muon pairs have a low cross section

EW processes like W/Z production are good candidates:

high cross section and clean signature

QCD NNLO corrections to the partonic cross section have been

calculated and the scale error is less than 1 %

PDF’s more problematic 5-8 %

Taking into account experimental error a precision of 10 % in L might be reached

quite early

Aiming at 3- 5 % after some time - LHC data itself might constrain the PDF

Ultimate goal in ATLAS: Measure L with ~ 3 % accuracy

How do we get there?

54

Overall conclusion (cont.)Overall conclusion (cont.) Third step : Elastic scattering

Elastic scattering without reaching Coulomb regionMeasurements of the total rate in combination with the t-dependence of the elastic

cross section is a well established and potentially powerful method for luminosity calibration and measurement of tot

10 % rather straight forward – 5% after some time

Optical theorem in combination with independent measurement of tot

Depends on the error on tot that will be provided but total error on Luminosity of 5 % seems

reasonable

Coulomb Interference measurement

Getting the Luminosity through Coulomb normalization will be extremely challenging due to

the small angles and the required closeness to the beam .However there is a potential for

3 % precision in luminosity

55

Back up

56

Overall conclusion (cont.)Overall conclusion (cont.) ATLAS proposes to use scintillating fibers in Roman Pots to measure

Elastic Scattering at small angles.

The ultimate goal is to reach the Coulomb interference region. This will be extremely challenging due to the small angles and the required closeness to the beam.

Main challenge is not so much in the detectors but rather in the required beam properties. Especially the level of beam halo at small distances will be important.

The ALFA project will be an important part of the luminosity determination in ATLAS even if the Coulomb Interference region is not completely reached.

In addition, ALFA opens up the door for a Forward Physics program in ATLAS. Experience with working close to the beam will prepare us for expanding towards a Forward Physics program as a future upgrade.

57

Hit pattern and acceptanceHit pattern and acceptanceCoulomb

58

Error evaluation as in csc notes and gueillemin article Error evaluation for pdf’s Tords comment

59

Performance simulationPerformance simulationCoulomb

60

Systematic errorsSystematic errorsCoulomb

61

Statistical results from fitStatistical results from fit

• 10 M events corresponding

to 100 hours at L= 1027

• Edge 1.5 mm from beam

dNdt

∣t≈0=Lπ∣f Cf N∣2≈Lπ∣−

2aEM

∣t∣

σtot

4π iρ e

−b∣t∣

2∣2

4.30.140.15

0.2517.918B (Gev-2)

0.9101.1101tot ( mb )

1.88.158.10L (1026 cm-2 s-1 )

Error (%)Lin.fitInput

Coulomb

62

Machine induced backgroundMachine induced background

Multiple Sources

• Beam gas scattering in the arcs• Local beam gas• Inefficiencies of collimation system

Summary: 2kHz integrated above 10

Important and difficult• determines how close we can come• backgrounds estimate often wrong

Coulomb

63

Machine induced background (cont.)Machine induced background (cont.)

Main handles:

Elastic signature:• left –right coincidence• acollinearity cut

Vertex cut

Background reduced to 2 % of the elastic signal

Single diffractive background

(generated with PYTHIA )negligible : 1 permille

Coulomb

64

Vertex reconstruction for background Vertex reconstruction for background rejectionrejection

Coulomb

65

The overlap The overlap detectordetector concept conceptCoulomb

66

MAROC2 chip bonded at CERN

67

Test beam-this summerTest beam-this summer

Complete detector for one Roman Pot i.e. 1460 channels

Coulomb

68

The ATLAS DetectorThe ATLAS Detector

GeV

10%( , ) 0.3%

/E eE E

GeV

60mrad

/E

GeV

4 ns

/t E

GeV

50%( ) 3%

/E

E E

GeV

50%( jet) 2%

/E

E E

Calorimetry:

TAS

R

50 1 2 43 6 7 8

Barrel

FCAL LUCIDTracking

EndCapRP

ZDC/TAN

Diffraction/Proton Tagging Region

y109

-chambers

GeV(Inner Det) (0.03 / 1.2)%TT

pp

Tracking:

GeV(IDet+ ) (0.009 / 1.4)%TT

pp

69

Luminosity Measurement (cont.)Luminosity Measurement (cont.)

Relative precision on the measurement of HBR for various channels, as function of mH, at Ldt = 300 fb–1. The dominant uncertainty is from Luminosity: 10% (open symbols), 5% (solid symbols).

(ATLAS-TDR-15, May 1999)

Higgs coupling tan measurement

ExamplesExamples

Systematic error dominated by luminosity (ATLAS Physics TDR )