15 Oct 2009 HAW09 meeting The TWIST experiment at TRIUMF (a status report) Jean-Michel POUTISSOU TRIUMF.

15 Oct 200915 Oct 2009 HAW09 meetingHAW09 meeting

The TWIST experiment at The TWIST experiment at TRIUMF (a status report)TRIUMF (a status report)

Jean-Michel POUTISSOUJean-Michel POUTISSOU

TRIUMFTRIUMF

Muon DecayMuon Decay

Purely leptonic weak processPurely leptonic weak process Most precisely studied weak processMost precisely studied weak process Theoretically known up to second order Theoretically known up to second order

radiative corrections (A.Czarnecki, U of radiative corrections (A.Czarnecki, U of Alberta)Alberta)

Large Experimental effort :Large Experimental effort :

Lifetime (PSI,RAL), Michel Lifetime (PSI,RAL), Michel parameters (TRIUMF,PSI), Magnetic parameters (TRIUMF,PSI), Magnetic moment (g-2 @BNL), New g-2 moment (g-2 @BNL), New g-2 Rare decays ( PSI,J-PARC,FNAL?) Rare decays ( PSI,J-PARC,FNAL?)

15 Oct 200915 Oct 2009 HAW09 meetingHAW09 meeting

Muon decayGeneral 4-fermion interaction:

@ 90% C.L.96.0gVLL

Experimentally:

Standard Model (“V-A”):

1gVLL 90% C.L. from

a global analysis, all others zero

1

rate ~ x2 3 3x 2

3 4x 3 3xo

1 x

x

P cose 1 x

2

3 4x 3

e+ spectrum in x, cose

(L. Michel, A. Sirlin)

maxe

e

E

Ex e

p e

s

Spectral shape in x, cose is characterized in terms of four parameters -- , , , P is the muon polarization

xo me

Eemax

Eemax

m2 me

2

2m

Muon decay parameters in terms of Non SM weak Muon decay parameters in terms of Non SM weak couplingscouplings

P

SM PDG

3/4 = 0.75180.0026 1969

3/4 = 0.74860.00260.0028 1988

1.0 = 1.00270.0026 1987

0.0 = 0.0070.013 1985

1.0 > 0.99682, CL = 90% 1986

status prior to TWIST--status prior to TWIST--

TWIST measured , , in two steps -- 10-3 in 2004; ~2x10-4 in 2006/7

TWIST measured , , in two steps -- 10-3 in 2004; ~2x10-4 in 2006/7

Requirements for High precision Requirements for High precision Michel Parameter measurementMichel Parameter measurement High (Known) muon polarization High (Known) muon polarization

must account for depolarisation in muon must account for depolarisation in muon transport, injection, stoppingtransport, injection, stopping

Excellent energy resolutionExcellent energy resolution Uniform and known acceptanceUniform and known acceptance Uniform tracking efficiencyUniform tracking efficiency Repeatability of measurement under varying Repeatability of measurement under varying

conditions.conditions.

Experimental solutionsExperimental solutions

M13 beam ,small dp/p selection.M13 beam ,small dp/p selection. High field solenoidal magnet (2T)High field solenoidal magnet (2T) Special injection optics and tracking TECSpecial injection optics and tracking TEC High purity metal targets (Al, Ag)High purity metal targets (Al, Ag) Low mass, High precision detectorLow mass, High precision detector Large solid angle and uniform acceptanceLarge solid angle and uniform acceptance High performance tracking softwareHigh performance tracking software Detailed,as built geant 3 simulationDetailed,as built geant 3 simulation

15 Oct 200915 Oct 2009 HAW09 meetingHAW09 meeting

TWIST experiment: muon deliveryStopping

targetNucl. Instr. and Meth. A566 (2006) 563-574

6

The The TWISTTWIST Spectrometer Spectrometer

Use highly Use highly polarized polarized ++ beam.beam.

Stop them in a Stop them in a very symmetric very symmetric detector.detector.

Decay Decay ee++ are are tracked through tracked through uniform, well-uniform, well-known field.known field.

Analysis MethodAnalysis Method

Extract energy and angle distributions for data:Extract energy and angle distributions for data:• apply (unbiased) cuts on muon variables.apply (unbiased) cuts on muon variables.• reject fast decays and backgrounds.reject fast decays and backgrounds.• calibrate calibrate ee++ energy to kinematic end point at 52.8 MeV. energy to kinematic end point at 52.8 MeV.

Fit to identically derived distributions from Fit to identically derived distributions from simulation:simulation:• GEANT3 geometry contains virtually all detector GEANT3 geometry contains virtually all detector

components.components.• simulate detector response in detail (clusters of ionization).simulate detector response in detail (clusters of ionization).• realistic, measured beam profile and divergence.realistic, measured beam profile and divergence.• extra muon and beam positron contamination included.extra muon and beam positron contamination included.• output into digitized format, identical to real data.output into digitized format, identical to real data.• fit to hidden variables with blind analysis method.fit to hidden variables with blind analysis method.• Validation by special data sets ( upstream stops) Validation by special data sets ( upstream stops)

Determination of the difference between data and MC data

Event Analysis

• Event classification (31 types)

• Helix fit to events within fiducial volume

• Extract e+ momentum and angle spectrum in bins of x and cose

Geant data are produced with Hidden decay parameters

Geant data spectrum

Twist data are collected

Twist data spectrum

, ,

Fit

Evaluation of Systematic UncertaintiesEvaluation of Systematic Uncertainties

TWISTTWIST relies on a fit to a Geant 3 relies on a fit to a Geant 3 simulation:simulation:• Simulation must be validated.Simulation must be validated.• Reconstruction systematics eliminated if Reconstruction systematics eliminated if

simulation is perfect.simulation is perfect.

General method:General method:• exaggerate a condition (in data or MC) that may exaggerate a condition (in data or MC) that may

cause error.cause error.• measure effect by fitting, using correlated sets measure effect by fitting, using correlated sets

where practical.where practical.• scale results according to variance in a data set.scale results according to variance in a data set.• Linearity? Double counting?Linearity? Double counting?

Evaluating Systematic Errors

Event Analysis

• Event classification (31 types)

• Helix fit to events within fiducial volume

• Extract e+ momentum and angle spectrum in bins of x and cose

nominal data collected

nominal data spectrum

special data runs collected

special data spectrum

, ,

Fit

Blind analysisBlind analysis

The base MC set is generated The base MC set is generated using unknown, hidden values of using unknown, hidden values of decay parameters (decay parameters (MCMC).).

All systematic uncertainties as All systematic uncertainties as well as offsets well as offsets are confirmed are confirmed prior to revealing hidden values.prior to revealing hidden values.

Open box to reveal hidden Open box to reveal hidden values of values of ..

Obtain final values of decay Obtain final values of decay parameters.parameters.

Fitting the data distributionsFitting the data distributions Michel distribution is Michel distribution is

linear in linear in , , , , , and , and , so a fit to first , so a fit to first order expansion is order expansion is exact.exact.

Fit data (Fit data (datadata) to sum ) to sum of a base MC of a base MC distribution (distribution (MCMC) plus ) plus MC-generated MC-generated derivative derivative distributions times distributions times fitting parameters fitting parameters (() representing ) representing deviations from base deviations from base MC.MC.

Can also fit data to Can also fit data to data and MC to MC data and MC to MC for systematic tests.for systematic tests.

15 Oct 200915 Oct 2009 HAW09 meetingHAW09 meeting

Summary of Pμξ final uncertaintiesStatistical

μ+ beam

External

Momentum calibration

Resolution

e+ interactions

Alignment

Chamber response

Beam intensity

Magnetic field map

Stopping material

Production target

Background μ+

Pμ uncertainties

uncertaintiesfrom e+

reconstruction

Opening the BOXOpening the BOX

TBA….. soonTBA….. soon

ConclusionsConclusions

TWIST will meet its objectives of an TWIST will meet its objectives of an order of magnitude improvements in order of magnitude improvements in the precision of the determination of the precision of the determination of the Michel parameters.the Michel parameters.

So far no inconsistencies with the So far no inconsistencies with the Standard ModelStandard Model

New constraints on physics beyond New constraints on physics beyond the SM are providedthe SM are provided

TWISTTWIST ParticipantsParticipantsTRIUMFTRIUMF

Ryan Bayes Ryan Bayes Yuri DavydovYuri DavydovJaap DoornbosJaap DoornbosWayne FaszerWayne FaszerMakoto FujiwaraMakoto FujiwaraDavid GillDavid GillAlex GrossheimAlex GrossheimPeter GumplingerPeter GumplingerRobert HendersonRobert HendersonAnthony HillairetAnthony HillairetJingliang HuJingliang HuJohn A. Macdonald John A. Macdonald xxGlen MarshallGlen MarshallDick MischkeDick MischkeMina NozarMina NozarKonstantin Olchanski Konstantin Olchanski Art Olin Art Olin Robert OpenshawRobert OpenshawTracy Porcelli Tracy Porcelli Jean-Michel PoutissouJean-Michel PoutissouRenRenéée Poutissoue PoutissouGrant ShefferGrant ShefferBill Shin Bill Shin

AlbertaAlbertaAndrei Gaponenko Andrei Gaponenko

¤¤Peter KitchingPeter KitchingRobert MacDonald Robert MacDonald

¤¤Maher QuraanMaher QuraanNate Rodning Nate Rodning xxJohn SchaapmanJohn SchaapmanGlen StinsonGlen Stinson

British ColumbiaBritish ColumbiaBlair Jamieson Blair Jamieson ¤¤Mike HasinoffMike HasinoffJames BuenoJames Bueno

MontrMontrééalalPierre DepommierPierre Depommier

ReginaReginaTed MathieTed MathieRoman TacikRoman Tacik

Kurchatov InstituteKurchatov Institute

Vladimir SelivanovVladimir Selivanov

Vladimir TorokhovVladimir Torokhov

Texas A&MTexas A&M

Carl GagliardiCarl Gagliardi

Jim Musser Jim Musser ¤¤

Bob Tribble Bob Tribble

Maxim VasilievMaxim Vasiliev

ValparaisoValparaiso

Don Koetke Don Koetke

Paul NordPaul Nord

Shirvel Stanislaus Shirvel Stanislaus

¤ ¤ Graduated studentGraduated student

x x deceaseddeceased

Supported by NSERC(Canada), Supported by NSERC(Canada), TRIUMF, DOE(USA) and Russian TRIUMF, DOE(USA) and Russian

ministry of Scienceministry of Science

The end The end