Exploring Muon Decay with TWIST

1
Exploring Muon DecaywithTWIST

• Carl A. Gagliardi
• Texas AM University
• for the TWIST Collaboration
• Outline

• Physics of muon decay
• TWIST experiment
• TWIST results to date
• How will we do better?

2
Muon decay matrix element

• Most general local, derivative-free,
  lepton-number conserving muon decay matrix
  element
• In the Standard Model, gVLL 1, all others are
  zero
• Pre-TWIST global fit results (all 90 c.l.)
• Theoretical constraints have recently been
  derived on the LR and RL terms from neutrino mass
  limits (hep-ph/0608163)

3
Muon decay spectrum

• The energy and angle distributions of positrons
  following polarized muon decay obey the Michel
  spectrum
• Pre-TWIST accepted values for the Michel
  parameters

where
SM ? 0.7518 0.0026 3/4 ?
-0.007 0.013 0 P?? 1.0027 0.0079
0.0030 1 ? 0.7486 0.0026 0.0028
3/4 P?(??/?) gt 0.99682 (90 c.l.) 1
4
Goal of TWIST

• Search for new physics that can be revealed by
  order-of-magnitude improvements in our knowledge
  of ?, d, and Pµ?
• Model-independent limit on muon handedness
• Left-right symmetric models
• ..

Two examples
5
What is required?

• Must
• Understand sources of muon depolarization
• -- Pµ and ? come as a product
• Determine spectrum shape
• -- All three parameters
• Measure forward-backward asymmetry
• -- For Pµ? and d
• to within a few parts in 104

6
Surface muon beam
7
TWIST spectrometer
8
Typical events

• Use pattern recognition (in position and time) to
  sort hits into tracks, then fit to helix
• Must also recognize beam positrons, delta tracks,
  backscattering tracks

9
2-d momentum-angle spectrum
In angular fiducial
In momentum fiducial
Acceptance of the TWIST spectrometer
10
Fitting the data distributions
aMC hidden ? blind analysis

• Fit data to sum of a MC base spectrum plus
  MC-generated derivative distributions.
• Decay distribution is linear in the Michel
  parameters, so this is exact, no matter what
  values (aMC) are used in the MC base spectrum.

11
Physics data sets

• Fall 2002
• Test data-taking procedures and develop analysis
  techniques
• First physics results ? and d
• Graphite-coated Mylar target not suitable for Pµ?
• Fall 2004
• Al target and Time Expansion Chamber enabled
  first Pµ? measurement
• Improved determinations of ? and d are underway
• 2006-07
• Achieve ultimate TWIST precision for ?, d, and
  Pµ?

12
Fitting the 2002 data to determine ? and d
Normalized residuals (Data-Fit)/sigma of the
2-d momentum-angle fit Fit describes the data
well, even when extrapolated far outside the
fiducial region
Angle-integrated results
13
Fitting the 2004 data to determine Pµ?
Separating the asymmetry into components
14
Results to date

• From Fall, 2002 run
• ? 0.75080 0.00032 (stat) 0.00097 (syst)
  0.00023 (?)
• PRL 94, 101805
• d 0.74964 0.00066 (stat) 0.00112 (syst)
• PRD 71, 071101
• New global analysis (PRD 72, 073002) using the ?
  and d results, together with previous
  measurements and recent e transverse
  polarization measurements (PRL 94, 021802)
• Significant improvements in the limits for
  gS,V,TLR
• ? -0.0036 0.0069
• From Fall, 2004 run (so far)
• Pµ? 1.0003 0.0006 (stat) 0.0038 (syst)
• PRD 74, 072007
• Factors of 2-3 improvements on pre-TWIST
  precisions

15
New limits in left-right symmetric models
General LRS model
Restricted (manifest) LRS model
Initial TWIST measurements already provide
significant new limits
16
Results to date

• From Fall, 2002 run
• ? 0.75080 0.00032 (stat) 0.00097 (syst)
  0.00023 (?)
• PRL 94, 101805
• d 0.74964 0.00066 (stat) 0.00112 (syst)
• PRD 71, 071101
• New global analysis (PRD 72, 073002) using the ?
  and d results, together with previous
  measurements and recent e transverse
  polarization measurements (PRL 94, 021802)
• Significant improvements in the limits for
  gS,V,TLR
• ? -0.0036 0.0069
• From Fall, 2004 run (so far)
• Pµ? 1.0003 0.0006 (stat) 0.0038 (syst)
• PRD 74, 072007
• Factors of 2-3 improvements on pre-TWIST
  precisions

17
Systematics in the previous measurements
The same effects tend to dominate the systematic
uncertainties for all three parameters.
18
Reducing the leading systematics

• Issues that were unique to 2002 data
• Stopping target thickness uncertainty
• Chamber orientation uncertainty with respect to
  magnetic field
• For all three Michel parameters
• Chamber response
• Improved gas system regulation and monitoring
• Improved determination of foil geometry
• Improved treatment of drift chamber behavior
• Positron interactions
• Specific to Pµ?
• Muon depolarization when crossing fringe field
• Muon depolarization in the stopping target

19
Controlling positron interaction uncertainties
upstream stops

• Our Monte Carlo must simulate positron
  interactions properly
• Need a well-understood test beam to validate the
  Monte Carlo
• We use the Michel spectrum!

20
Validating the Monte Carlo with upstream stops
(Downstream fit result) (upstream fit result)
Bremsstrahlung
Scattering

• The TWIST Monte Carlo provides an excellent
  description of the hard interaction physics
• Took 50 times more upstream stop events in 2004
  than in 2002

21
Muon depolarization across the fringe field
Use Time Expansion Chamber (TEC) to measure and
optimize the muon beam

• First installed for 2004 run
• Found vertical beam offset now corrected
• Now take frequent beam characterizations
• Have techniques to identify when the beam changes
  between TEC measurements

2004 muon beam spot
22
Muon depolarization after stopping

• Observed significant depolarization in 2004 data
• New techniques allow us to veto muons that stop
  outside the metal stopping target
• More sensitive analysis procedures to determine
  the residual depolarization rate
• Now taking data with a Ag target to explore
  material dependence

23
Conclusions

• The initial TWIST measurements have improved our
  knowledge of the Michel parameters ?, d, and Pµ?
  by factors of 2-3
• Improvements by additional factors of 3-5 are
  anticipated
• Stay tuned!

24
TWIST Participants

• TRIUMF
• Ryan Bayes y
• Yuri Davydov
• Jaap Doornbos
• Wayne Faszer
• Makoto Fujiwara
• David Gill
• Alex Grossheim
• Peter Gumplinger
• Anthony Hillairet y
• Robert Henderson
• Jingliang Hu
• John A. Macdonald x
• Glen Marshall
• Dick Mischke
• Mina Nozar
• Konstantin Olchanski
• Art Olin y

• Alberta
• Andrei Gaponenko
• Peter Kitching
• Robert MacDonald
• Maher Quraan
• Nate Rodning x
• John Schaapman
• Glen Stinson
• British Columbia
• James Bueno
• Mike Hasinoff
• Blair Jamieson
• Montréal
• Pierre Depommier

Kurchatov Institute Vladimir
Selivanov Vladimir Torokhov Texas AM Carl
Gagliardi Jim Musser Bob Tribble Maxim
Vasiliev Valparaiso Don Koetke Paul
Nord Shirvel Stanislaus Graduate student
Graduated y also U Vic z also Manitoba zz also
Saskatchewan x deceased
Supported under grants from NSERC (Canada) and
DOE (USA). Additional support from TRIUMF, NRC
(Canada), and the Russian Ministry of
Science. Computing facilities of WestGrid are
gratefully acknowledged.
25
(No Transcript)
26
Coupling constants and Michel parameters

• The Michel parameters are bilinear combinations
  of the coupling constants

27
Detector array

• 56 low-mass high-precision planar chambers
  symmetrically placed around thin target foil
• Measurement initiated by single thin
  scintillation counter at entrance to detector
• Beam stop position controlled by variable He/CO2
  gas degrader

28
Analysis method

• Extract energy and angle distributions for data
• Apply (unbiased) cuts on muon variables.
• Reject fast decays and backgrounds.
• Calibrate e energy to kinematic end point at
  52.83 MeV.
• Fit to identically derived distributions from
  simulation
• GEANT3 geometry contains virtually all detector
  components.
• Simulate chamber response in detail.
• Realistic, measured beam profile and divergence.
• Extra muon and beam positron contamination
  included.
• Output in digitized format, identical to real
  data.

29
Data set by data set 2-d fit results
2002
?
2004
Pµ?
d
30
Recent muon decay global analysis
PRD 72, 073002
These improvements arise from the TWIST ? and d
measurements.

• Fit also finds ? -0.0036 0.0069, a factor of
  2 more precise than the previously accepted
  value, -0.007 0.013.
• Significant improvement in ? comes from new
  measurements of the transverse polarization of
  the e in muon decay (PRL 94, 021802), but the
  TWIST ? and d measurements also play an important
  part.

31
Validating the soft physics simulationwith
upstream stops
Multiple Scattering
dE/dx

• Multiple scattering is well reproduced
• Small differences in dE/dx are seen between data
  and MC
• Exploring differences between GEANT3 and GEANT4

32
Tracking (in)efficiency

• Measured in upstream stop events
• Reconstruct track upstream, then ask if
  downstream also reconstructs
• Overestimates true inefficiency not all tracks
  reach the downstream half

Difference between inefficiencies in data and
Monte Carlo events