LBL Comet-Mu2E Workshop January 24, 2009

Mu-e Conversion Backgrounds and Sensitivities– from proposal

to measurement

Doug Bryman

"Wishing does not make a poor man rich." Arabian Proverb

LFV Experiments Limit Reached

Goal; (Result/Goal)

“Comments”

Badertscher et al. 1982 µ->e

7x10-11

TRIUMF TPC

Ahmad et al. 1987

µ->e

4.6x10-12 2x10-12

(2)

Data collection took 5x as long as originally guessed

(1 month!)

SINDRUM II

Bertl et al. 2006

µ->e

7x10-13

Au

4.3x10-12

(6x10-13)

Ti

“10-14” (1987)->

3x10-14(1993)

“engineering”

Ti (>60)

Flux lower by 10; pion suppression device didn’t work; unanticipated high electron bkg.; shorter running.

MEGA

Ahmed et al. 2002

µ->eγ

1.2x10-11 0.9->4x10-13

“engineering”

(133-35)

Death by a thousand blows to acceptance

N.B.: Every one of these experiments was carefully reviewed by experts!

MEGA at LAMPF

SINDRUM II PSI

• Proposed 108 stops; (muE1) beam was only 107

• Designed “PMC” to kill pions; simulated; swamped unexpectedly by electrons; solenoid took years longer to obtain.

• Eventually went to very low momentum (50 MeV/c) killing pions by range; pion background persisted.

• Final result obtained in a couple of months; group had dispersed….” could have done better”….

MEG PSI• Proposal goal (1999): <2x10-14

in 2.2x107 s.• Engineering…lost factor 17• Current goal (2008): <1.7x10-13

in 4.5x107 s.

“ Goals “ 2008 (1999)

Measured

Simulated

Gamma energy % 4.5 – 5.0 (1.4)

Gamma Timing (ns) 0.15 (0.1)

Gamma Position (mm) 4.5 – 9.0 (1.7)

Gamma Efficiency (%) >40 (70)

e+ Timing (ns) 0.1

e+ Momentum (%) 0.8 (0.3)

e+ Angle (mrad)10.5 (10)

e+ Efficiency (%) 65 (95)

Muon decay Point (mm) 2.1

Muon Rate (108/s) 0.3 (1.0)

Running Time (weeks) 100 (49)

Single Event Sens (10-

13)0.5 (0.094)

Accidental Rate (10-13) 0.1 – 0.3

# Accidental Events 0.2 - 0.5

90% CL Limit (10-13) 1.7 (0.2)

COMET?----------------Mu2E?

0.047

0.07x0.6=0.42

0.41

0.34

s.e.s: 2.3 x 10-17

*

*

***Kaons?

Sensitivity/Background Estimate Questions

• What are the uncertainties and risk factors in the background, acceptance estimates?

• What processes are missing?

• How are the backgrounds to be measured?

• How is a blind analysis to be done?

• What would make a believable signal?

• Optimistic resolutions (DIO) – contamination?• Optimistic acceptances – extra losses due to

cuts• Missing background sources due to high energy

production… e.g. more electrons (from where?); previous experiments used low energy sources.

• Combinations of cosmics and beam-related effects?

• Fill in your own….

How to lose a factor 10 (100)?

Advance Background-related Measurements?

• Cosmic rays – could be done in a test setup• Extinction – could be done in advance• Radiative pion capture -> 100 MeV electrons?• Neutrons? Pbars?• …

The bifurcation method uses two uncorrelated high-rejection cuts to estimate each background

A

B

C

DCU

T1

CUT2

The signal region A is not directly examined

Measure B:Invert CUT1 and apply CUT2

Measure C+D:Invert CUT2 Estimate A:

A = BC/D

Measure D:Invert CUT2 andapply CUT1

Background estimate = A!

Digression: E787/E949 Blind Analysis and Background Estimate Methodology

Background estimates are performed using data different from those used to develop

cuts

EntireData Set

CutDevelopment

BackgroundEstimates

1/3 2/3

Compare background

estimates

•Estimate and then measure the backgrounds near the signal region to verify estimates.

Conclusions/Recommendations

• Perform risk analysis to get best estimates of ultimate sensitivity

• Devise measurements to test assumptions

• Make Conservative claims – then do better!

The 2005 HEPAP review of MECO stated that a minimal goal of achieving sensitivity significantly less 10-15 than was essential to justify the considerable effort and expense.