1. A New Charged Lepton Flavor Violation Experiment: Muon-Electron Conversion at FNAL

3. muon converts to electron in the presence of a nucleus
What is μe Conversion?
muon converts to electron in the presence of a nucleus
Charged Lepton Flavor Violation (CLFV)
Related Processes:
μ or τ → eγ, e+e-e, KL→μe, and more 3

4. Experimental Signal A Single Monoenergetic Electron
If N = Al, Ee = 105. MeV
energy depends on Z
will use this many times, so take a moment

5. Overview Of Processes μ- stops in thin Al foil μ- in 1s state
Al Nucleus
~4 fm
μ- in 1s state
the Bohr radius is ~ 10 fm, so the μ- sees the nucleus
total disappearance rate =0.864 μ sec
NORMALIZATION
60% captured
μ- emits ν
Al turns into Mg
40% decay-in-orbit
decays by
normal process
but can recoil
off nucleus
BACKGROUND
2.2
μ sec
1.4

6. Why Normalize to Capture?
Al turns into Mg
Nuclear wavefunctions “cancel,” calculation simpler
As muon cascades to 1s, X-rays give stop rate
and Mg →Al yields a 2.6 MeV β followed by γ that can be used to measure capture rate
μ- emits ν
Al turns into Mg
NORMALIZATION

8. Neutrino Oscillations and LFV

11. “Model-Independent” Picture
“Loops”
“Contact Terms”
how can we use both experiments?
Exchange of a new, massive particle
Supersymmetry and Heavy Neutrinos
Contributes to μ→eγ
Does not produce μ→eγ

12. μe Conversion and μ→eγ κ Λ (TeV) 1) Mass Reach to 104 TeV
2) x10 beyond MEG in loop-dominated physics
Project X Mu2e
Mu2e
higher mass scale
generally acknowledge for experimental reasons mu e gamma will not get much better
MEG
MEGA
SINDRUM κ

14. Supersymmetry and Mu2e in Minimal SU(5)
J. Hisano, T. Moroi, K. Tobe and M. Yamaguchi, Phys. Lett. B 391, 341 (1997). [Erratum-ibid. B397, 357 (1997).]

15. Supersymmetry in Tau LFV
L. Calibbi, A. Faccia, A. Masiero, S. Vempati hep-ph/
Neutrino-Matrix Like (PMNS)
Minimal Flavor Violation(CKM)
neutrino mass via the see-saw mechanism, analysis is performed in an SO(10) framework

16. ….and Muon-Electron Conversion
Neutrino-Matrix Like (PMNS)
Minimal Flavor Violation(CKM)
Mu2e Phase I
complementarity between Lepton Flavor Violation
(LFV) and LHC experiments!

17. Contributions to μe Conversion
note both loop processes and new particle exchange, will come back to this next slide

18. Design of Mu2e How can we do better? Examine previous best experiment
What were the limitations?
limitations from prompts
limitations from Decay-in- Orbit
How can we do better?

19. Previous Best Experiment
SINDRUM-II
Rμe < 6.1 x in Au
Want to probe to or better
≈104 improvement

20. SINDRUM II Results Final SINDRUM-II on Au
Note Two Background Events past Signal Region
signal region
radiative π’s?
W. Bertl et al, Eur. Phys. J. C 47, (2006)

21. What Limited SINDRUM-II?
DC Beam
no time separation between
signal and prompt background
point to radiative pi capture in b), explain degrader and veto. We use a different method, beam extinction,
radiative π capture

22. How Can We Do Better? Requiring
>103 increase in muon intensity from SINDRUM
Requiring
Pulsed Beam to Eliminate prompt backgrounds like radiative π capture
protons out of beam pulse/ protons in beam-pulse < and we must measure it

23. Advantage of Pulsed Beam
target foils: muon converts here
Recall:
Muon-electron conversion signal is a
single,monoenergetic electron
pulsed beam lets us wait until after prompt backgrounds disappear
delayed 105 MeV electron

24. Pulsed Beam Structure Tied to prompt rate and machine: FNAL “perfect”
Want pulse duration << τμ, pulse separation ≈ τμ
FNAL Accumulator has circumference 1.7μsec !
Extinction between pulses < needed
= # protons out of pulse/# protons in pulse
10-9 based on simulation of prompt backgrounds

25. achieving 10-9 is hard; normally get 10-2 – 10-3
Extinction Scheme
Eliminate protons in beam in-between pulses:
“Switch” dipole timing to switch signal and background: accept only out-of-time protons for direct measurement of extinction
Other schemes under investigation
Measurement: off-angle collimators and telescope?
achieving 10-9 is hard; normally get 10-2 – 10-3

26. FNAL Beam Delivery FNAL has unique, major strength: Multiple Rings
no interference with NOvA neutrino oscillation experiment
reuse existing rings with only minor modifications
big strength of FNAL: can run without interference

27. 8 GeV Proton Beam MI/Recycler ¯ were used for p
After TeVatron shut-down, Accumulator, Debuncher, and Recycler no longer needed for antiprotons
were used for p 27

28. Detector and Solenoid Tracking and Calorimeter
Decay into muons and transport to stopping target
S-curve eliminates backgrounds and sign-selects
Production: Magnetic bottle traps backward-going π that can decay into accepted μ’s

29. Production Solenoid:
Protons enter opposite to outgoing muons – this is a central idea to remove prompt background
Protons leave through thin window
π’s are captured, spiral around and decay
muons exit to right
Target Shielding
Proton Target
Pions
Protons enter here
4 m X 0.75 m

30. Transport Solenoid
Curved solenoid eliminates line-of-sight transport of photons and neutrons
Curvature drift and collimators sign and momentum select beam
occasional μ+

31. Detector Solenoid octagonal tracker surrounding central region:
radius of helix proportional to momentum
low momentum particles and almost all DIO background passes down center
so background goes down center and signal spirals out to tracker
Al foil stopping target
signal events pass through octagon of tracker
and produce hits

32. Graded Fields Production Solenoid: graded from ~5.0 to 2.5T
to (a) capture backwards-going pions
and allow them to decay and (b) “reflect” backward-going muons
Transport Solenoid:
graded from ~2.5 to 2.0T
to accelerate muons along beamline
reflection nearly 100% efficient near detector, doubling statistics
Detector Solenoid:
graded from ~2.0 to 1T
to “reflect” backwards-going electrons
and send them into detector

33. Choice of Stopping Material: rate vs wait
Stop muons in target (Z,A)
Physics sensitive to Z: with signal, can switch target to probe source of new physics
Why start with Al?
rate normalized to Al
2.5
Rate
1.0 Z
Kitano, et al., PRD 66, (2002)
shape governed by relative conversion/capture rate, form factors, ...

34. Prompt Background and Choice of Z
choose Z based on tradeoff between rate and lifetime: longer lived reduces prompt backgrounds
Nucleus
Rμe(Z) / Rμe(Al)
Bound Lifetime
Conversion Energy
Fraction >700 ns
Al(13,27)
1.0
864 nsec
MeV
0.45
Ti(22,~48)
1.7
328 nsec
MeV
0.16 Au
(79,~197) ~
72.6 nsec
MeV
negligible

35. Detector σ = 200 μ transverse, 1.5 mm axially
Octagon and Vanes of Straw Tubes
Immersed in solenoidal field, so particle follows near-helical path
up to dE/dx, scattering, small variations in field
Particles with pT < 55 MeV do not pass through detector, but down the center
Followed by Calorimeter
2800 axial straw tubes, 2.6 m by 5 mm, 25μ thick
use return yoke as CR shield
Calorimeter/Trigger:
σ /E = 5%, × 3.5 × 12 cm PbWO4

36. Beam’s Eye View of Tracker
Octagon and Vanes of Straw Tubes
Immersed in solenoidal field
Below pT = 55 MeV, electron stays inside tracker and is not seen; about 60° at MeV
Looking for helix as particle propagates downstream
target
target
Only ~ 0.3% of DIO’s are even accepted

37. Signal and Background
Rμe = 10-16

38. Backgrounds...
bd = albedo from beam stop (after calorimeter): splashback, extra hits confusing pattern recognition

39. Two Classes of Backgrounds
Prompt
Decay-In-Orbit
Source
Mostly π’s produced in target
Physics Background nearly indistinguishable from signal
Solution
Design of Muon Beam,
formation, transport, and time structure
Spectrometer Design:
resolution and pattern recognition

40. Decay-in-Orbit Background
High Rate
Peak 52.8 MeV
Detector insensitive to these
Fraction within 3 MeV of signal is 5 x 10-15
Rate falls as (Emax- E)5Drives Resolution Requirement
Zero energy neutrinos and coherent scatter off nucleus put DIO’s at conversion energy
Rate falls as (Emax- E)5
Fraction within 2 MeV of signal is 1.2 x 10-15
fifth power expression approximate, see Shanker

41. Final Backgrounds For Rμe = 10-16 expect ~5 events / 0.5 bkg
Extinction factor of 10-9
Source
Number/
4 x 1020
DIO
0.25
Radiative π capture
0.08
μ decay-in-flight
Scattered e-
0.04
π decay in flight
<0.004

44. Time Schedule

45. Cost Estimate

46. INFN-Pisa
and
Mu2e Experiment

47. R. Carosi,F. Cervelli,M.Incagli, T. Lomtadze, L.Ristori, F. Scuri
Participants:
R. Carosi,F. Cervelli,M.Incagli,
T. Lomtadze, L.Ristori, F. Scuri
and C. Vannini
(staff members only)

48. Medium Term Milestone: TDR

49. A light supporting structure
Mechanical Project
A light supporting structure

50. Choice: LSO vs PbWO4 vs Kloe (?)
Calorimeter
Proposal:
× 3.5 × 12 cm PbWO4
Choice: LSO vs PbWO4 vs Kloe (?)

51. Choice: Long vs Transv vs 4p (?)
Tracker
target
Octagon and Vanes of Straw Tubes
Immersed in solenoidal field
Below pT = 55 MeV, electron stays inside tracker and is not seen; about 60° at MeV
Looking for helix as particle propagates downstream
Only ~ 0.3% of DIO’s are even accepted
target
Choice: Long vs Transv vs 4p (?)

52. Alternative Tracker T-tracker (T for transverse): 260 sub-planes
sixty 5 mm diameter conducting straws
length from cm
total of 13,000 channels
T-Tracker Pattern Recognition Difficult

53. Trigger Design and related Electronics
Trigger: Calorimeter + (Tracker) + (TOF)

54. Backgrounds Studies
bd = albedo from beam stop (after calorimeter): splashback, extra hits confusing pattern recognition

55. A contribution to Solenoid (?)
Production
Transportation
Detection

56. Conclusion A difficult experiment looking for NP
cLFV experiment at the High Intensity Frontier:
A difficult experiment looking for NP
Enforcing collaboration between INFN and Fermilab
Nothing is frozen: other participants are welcome
(expecially from Meg and other Flavour Physics…)

57. BACKUPS
R. Bernstein, FNAL

58. Better than MECO because of better beam structure
if MECO could handle rates, Mu2e at FNAL can as well:
pre-project X or with Project X
MECO
Mu2e Booster
Mu2e
Project X, no expt. upgrade
Mu2e Project X,
expt. upgrade
protons/sec
40x1012 (design)
18x1012
70x1012
160x1012
average beam power
50 kW (design)
23 kW
90 kW
200 kW
duty factor
0.5 s on, 0.5 s off, 50%
75-90%
instantaneous rate
80x1012 (design)
20x1012
77x1012
220x1012
short term beam power
100 kW (design)
25 kW
100 kW
220 kW
Beam pulse period, msec
1.35
1.65
Data collection time interval msec

59. André de Gouvêa, Project X Workshop Golden Book
μe Conversion and μ→eγ
Λ (TeV)
1) Mass Reach to 104 TeV
2) x10 beyond MEG in loop-dominated physics
Project X Mu2e
Mu2e
higher mass scale
generally acknowledge for experimental reasons mu e gamma will not get much better
MEG
MEGA
SINDRUM κ
André de Gouvêa, Project X Workshop Golden Book

60. Prompt Background and Choice of Z
choose Z based on tradeoff between rate and lifetime: longer lived reduces prompt backgrounds
Prompt Background and Choice of Z
Nucleus
Rμe(Z) / Rμe(Al)
Bound Lifetime
Conversion Energy
Fraction >700 ns
Al(13,27)
1.0
864 nsec
MeV
0.45
Ti(22,~48)
1.7
328 nsec
MeV
0.16 Au
(79,~197) ~
72.6 nsec
MeV
negligible

61. Quick Fermilab Glossary
Booster:
The Booster accelerates protons from the 400 MeV Linac to 8 GeV
Accumulator:
momentum stacking successive pulses of antiprotons now, 8 GeV protons later
Debuncher:
smooths out bunch structure to stack more p now; rebunch for mu2e
Recycler:
holds more p than Accumulator can manage, “store” here _ _

62. Cost and Schedule
A detailed cost estimate of the MECO experiment performed just before RSVP was cancelled: (in Actual Year $, including inflation)
Solenoids and cryogenics: $59M
Remainder of experimental apparatus: $21M
Additional Fermilab costs have not been worked out in detail
accelerator and civil construction costs are being worked out
Estimate for contingency, overhead, etc then yields $120M before beamline and civil costs

63. Schedule: 2016 for commissioning
Based on the original MECO proposal, we believe the experiment could be operational within 3-4 years of “CD-2/3a” = begin large, long-lead time purchases
Use NOνA experience for time for DOE Approval Process
Use MECO schedule for Technical Issues, especially solenoid construction
Aggressive but possible

64. Rates at Beginning of > 700 nsec Live Window, so these are highest
≈ 2 hits per straw during beam flash
Rates are manageable: (1/4 of MECO instantaneous)
Rates In Tracker

65. Final Backgrounds For Rμe = 10-16 expect ~5 events / 0.5 bkg
Extinction factor of 10-9
Source
Number/
4 x 1020
DIO
0.25
Radiative π capture
0.08
μ decay-in-flight
Scattered e-
0.04
π decay in flight
<0.004

66. Mu2e Phase II Signal? Yes No 1. Change Z of Target
to determine source of new physics
2. Need Project X to provide statistics
1. Probe additional two orders of magnitude made possible by Project X
2. Need upgrades to muon transport and detector

67. available 8 GeV Power for intensity frontier
20 kW
(current)
200 kW
(Project X)
2000 kW
(Project X Upgrades)
Mu2e and Project X
Project X is required for the next step
Needed whether first phase sees a signal or sets a limit
Well timed for Mu2e first phase, late this decade or early next

68. Conclusions With Mu2e we would either:
Reduce the limit for Rμe by more than four orders of magnitude (Rμe 90% C.L.)
Discover unambiguous proof of Beyond Standard Model physics
In a second phase (Project X)
Extend the limit by up to two orders of magnitude
Study the details of new physics

69. Z-Dependence
For a small nucleus compared to the extent of the muon wf, the Schroedinger equation gives hydrogenic wavefunctions which scale like Z**3 at small radius. There are Z protons in the nucleus, so the probability of ordinary capture goes like Z**4.
For a conversion process, the cross section is coherent and therefore goes like Z**2 (or A**2) rather than Z in the case of ordinary capture. So overall the process goes like Z**5. Strictly speaking this overlap argument only works well for short-range forces like the weak force, since we assume that the probability of reaction is proportional to the overlap between the nuclear and the muon wavefunctions. For the EM force it will not work as well.
Therefore Rμe is proportional to Z4/Z3 = Z for small Z.
As Z increases, the finite size of the nucleus becomes important. The muon wavefunction is inside the nucleus and does not see the full Z, reducing the wavefunction overlap. Also, there are relativistic effects which reduce the Z dependence, and a few other effects.

70. Outline The search for muon-electron conversion Experimental Technique
Fermilab Accelerator
Project X Upgrades and Mu2e

71. Outline The search for muon-electron conversion Experimental Technique
Fermilab Accelerator
Project X Upgrades and Mu2e

72. Outline The search for muon-electron conversion Experimental Technique
Fermilab Accelerator
Project X Upgrades and Mu2e

73. History of CLFV Searches
supersymmetry may turn on around here in many models, and this plot does not include tau channels, discussed later
supersymmetry ~ 10-15

74. Current and Planned Lepton Flavor Violation Searches
Neutrino Oscillations!
τ LFV current limits at 10-7 for τ→μγ
MEG and μ→eγ
Mu2e:
Strengths of muon-electron conversion
Complementarity to other processes
what’s going on in the world now?

75. “Model-Independent” Picture
“Loops”
“Contact Terms”
how can we use both experiments?
Exchange of a new, massive particle
Supersymmetry and Heavy Neutrinos
Contributes to μ→eγ
Does not produce μ→eγ
Quantitative Comparison?

76. Alternative Tracker T-Tracker Pattern Recognition Difficult but
T-tracker (T for transverse):
260 sub-planes
sixty 5 mm diameter conducting straws
length from cm
total of 13,000 channels
T-Tracker Pattern Recognition Difficult but
Kalman Filter is promising

77. L-Tracker vs. T-Tracker
T-Tracker: straws perp to beam
More prone to pattern recognition errors?
L-Tracker vs. T-Tracker
L-Tracker: straws along beam
Conceptually simpler tracking
Basis of MECO
Where does support/infrastructure go? Material in electron path
Can anyone build straws 0.5 cm × 2.6m in vacuum?
Active Investigation:
kalman filter, applied to both on same events
work just beginning
help welcome!

78. Background Rates vs. Time
nsec
700 nsec
beam e-
Rate (15 MHz/wire)
μ DIF
divide by 4
FNAL/BNL
this is why we wait 700 nsec
Protons in stopping tgt
nsec
Rate (560 kHz/wire)

79. Expected Resolution We must understand resolution
this side lowers signal acceptance
We must understand resolution
Measure resolution with special runs varying target foils, field, location of stopping target
this side smears background into signal
relevant tail for DIO
σ ~110 keV

80. Proposed Site

83. History of Lepton Flavor Violation Searches1
10-2
K0 +e-
K+ + +e-
- N  e-N
+  e+
+  e+ e+ e-
10-4
10-6
10-8
10-10
SINDRUM II
10-12
Initial MEG Goal 
10-14
Initial mu2e Goal 
10-16

86. L-Tracker vs. T-Tracker
T-Tracker: straws perp to beam
More prone to pattern recognition errors?
L-Tracker vs. T-Tracker
L-Tracker: straws along beam
Conceptually simpler tracking
Basis of MECO
Where does support/infrastructure go? Material in electron th
Can anyone build straws 0.5 cm × 2.6m in vacuum?
Active Investigation:
kalman filter, applied to both on same events
work just beginning
help welcome!

87. Rates In Tracker
Rates at Beginning of > 700 nsec Live Window, so these are highest
≈ 2 hits per straw during beam flash
Rates are manageable: (1/4 of MECO instantaneous)

88. Expected Resolution We must understand resolution
this side lowers signal acceptance
We must understand resolution
Measure resolution with special runs varying target foils, field, location of stopping target
this side smears background into signal
relevant tail for DIO
σ ~110 keV

89. Background Studies
Rμe = 10-16

90. Two Classes of Backgrounds
Prompt
Decay-In-Orbit
Source
Mostly π’s produced in target
Physics Background nearly indistinguishable from signal
Solution
Design of Muon Beam,
formation, transport, and time structure
Spectrometer Design:
resolution and pattern recognition

91. Decay-in-Orbit Background
High Rate
Peak 52.8 MeV
Detector insensitive to these
Fraction within 3 MeV of signal is 5 x 10-15
Rate falls as (Emax- E)5Drives Resolution Requirement
Zero energy neutrinos and coherent scatter off nucleus put DIO’s at conversion energy
Rate falls as (Emax- E)5
Fraction within 2 MeV of signal is 1.2 x 10-15
fifth power expression approximate, see Shanker

92. A Contribution to Solenoid Construction (?)
Decay into muons and transport to stopping target
S-curve eliminates backgrounds and sign-selects
Production: Magnetic bottle traps backward-going π that can decay into accepted μ’s

93. Production Solenoid:
Protons enter opposite to outgoing muons – this is a central idea to remove prompt background
Protons leave through thin window
π’s are captured, spiral around and decay
muons exit to right
Target Shielding
Proton Target
Pions
Protons enter here
4 m X 0.75 m

94. Transport Solenoid
Curved solenoid eliminates line-of-sight transport of photons and neutrons
Curvature drift and collimators sign and momentum select beam
occasional μ+

95. Detector Solenoid octagonal tracker surrounding central region:
radius of helix proportional to momentum
low momentum particles and almost all DIO background passes down center
so background goes down center and signal spirals out to tracker
Al foil stopping target
signal events pass through octagon of tracker
and produce hits

98. Contributions from Pisa

99. Better than MECO because of better beam structure
if MECO could handle rates, Mu2e at FNAL can as well:
pre-project X or with Project X
MECO
Mu2e Booster
Mu2e
Project X, no expt. upgrade
Mu2e Project X,
expt. upgrade
protons/sec
40x1012 (design)
18x1012
70x1012
160x1012
average beam power
50 kW (design)
23 kW
90 kW
200 kW
duty factor
0.5 s on, 0.5 s off, 50%
75-90%
instantaneous rate
80x1012 (design)
20x1012
77x1012
220x1012
short term beam power
100 kW (design)
25 kW
100 kW
220 kW
Beam pulse period, msec
1.35
1.65
Data collection time interval msec

100. Rates at Beginning of > 700 nsec Live Window, so these are highest
≈ 2 hits per straw during beam flash
Rates are manageable: (1/4 of MECO instantaneous)
Rates In Tracker