1. University of Arizona Physics Colloquium, March 7, 2008 E Prebys A Muon to Electron Experiment at Fermilab Eric Prebys* For the Mu2e Collaboration 1 *U of A ’84 (Eng. Phys.)

2. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 2 Mu2e Collaboration Currently: 50 Scientists 11 Institutions *Co-contact persons R.M. Carey, K.R. Lynch, J.P. Miller*, B.L. Roberts Boston University Y. Semertzidis, P. Yamin Brookhaven National Laboratory Yu.G. Kolomensky University of California, Berkeley C.M. Ankenbrandt, R.H. Bernstein, D. Bogert, S.J. Brice, D.R. Broemmelsiek,D.F. DeJongh, S. Geer, M.A. Martens, D.V. Neuffer, M. Popovic, E.J. Prebys*, R.E. Ray, H.B. White, K. Yonehara, C.Y. Yoshikawa Fermi National Accelerator Laboratory D. Dale, K.J. Keeter, J.L. Popp, E. Tatar Idaho State University P.T. Debevec, D.W. Hertzog, P. Kammel University of Illinois, Urbana-Champaign V. Lobashev Institute for Nuclear Research, Moscow, Russia D.M. Kawall, K.S. Kumar University of Massachusetts, Amherst R.J. Abrams, M.A.C. Cummings, R.P. Johnson, S.A. Kahn, S.A. Korenev, T.J. Roberts, R.C. Sah Muons, Inc. R.S. Holmes, P.A. Souder Syracuse University M.A. Bychkov, E.C. Dukes, E. Frlez, R.J. Hirosky, A.J. Norman, K.D. Paschke, D. Pocanic University of Virginia

3. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Acknowledgement This effort has benefited greatly (and plagiarized shamelessly) from over a decade of voluminous work done by the MECO collaboration, not all of whom have chosen to join the current collaboration. 3

4. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Outline Theoretical Motivation Experimental Technique Making Mu2e work at Fermilab Sensitivities Future Upgrades Conclusion 4

5. University of Arizona Physics Colloquium, March 7, 2008 E Prebys General The study or rare particle decays allows us to probe mass scales far beyond those amenable to direct searches. Among these decays, rare muon decays offer the possibility of experimentally clean and unambiguous evidence of physics beyond the current Standard Model. Such searches are a natural part of the “Intensity Frontier”, which is being proposed for Fermilab after the end of the current collider program. In the case of muon conversion, we can take advantage of a great deal of work that has already been done in the planning of the Muon to Electron Conversion Experiment (MECO), which was proposed at Brookhaven. 5

6. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Lepton Number Conservation The concept of Lepton Number Conservation dates back to the earliest experiments and models for the Weak Interaction, originally involving only electrons and electron neutrinos. Example: After the discovery of the muon, it was discovered that Lepton number was separately conserved for each lepton generation: These conservation laws were an important constraint in formulating what is now the “Standard Model” 6

7. University of Arizona Physics Colloquium, March 7, 2008 E Prebys The Standard Model In the Standard Model, both quarks and leptons are arranged in generations. In weak eigenspace, the weak interaction causes transition within generations Because the mass eigenstates are superpositions of the weak eigenstates, transitions between physical generations can occur, iff  The mixing element is nonzero  The masses are nonzero (otherwise unitarity will force the amplitude to sum to zero) Thus, to first order (where neutrinos are equally massless), generational transtions are  Allowed for quarks  Forbidden for leptons 7

8. University of Arizona Physics Colloquium, March 7, 2008 E Prebys  ->e CLFV in the SM Forbidden in Standard Model Observation of neutrino mixing shows this can occur at a very small rate Photon can be real (  ->e  ) or virtual (  N->eN) Standard model B.R. ~ O (10 -50 ) 8 First Order FCNC: Higher order dipole “penguin”: Virtual mixing

9. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Beyond the Standard Model Because extensions to the Standard Model couple the lepton and quark sectors, lepton number violation is virtually inevitable. Charged Lepton Flavor Violation (CLFV) is a nearly universal feature of such models, and the fact that it has not yet been observed already places strong constraints on these models. CLFV is a powerful probe of multi-TeV scale dynamics: complementary to direct collider searches Among various possible CLFV modes, rare muon processes offer the best combination of new physics reach and experimental sensitivity 9

10. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Generic Beyond Standard Model Physics ? ? ? Flavor Changing Neutral Current Mediated by massive neutral Boson, e.g.  Leptoquark  Z’  Composite Approximated by “four fermi interaction” Dipole (penguin) Can involve a real photon Or a virtual photon ? ? ?

11. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 11 Muon-to-Electron Conversion:  +N  e+N Similar to  e  with important advantages:  No combinatorial background  Because the virtual particle can be a photon or heavy neutral boson, this reaction is sensitive to a broader range of BSM physics Relative rate of  e  and  N  eN  is the most important clue regarding the details of the physics  105 MeV e - When captured by a nucleus, a muon will have an enhanced probability of exchanging a virtual particle with the nucleus. This reaction recoils against the entire nucleus, producing the striking signature of a mono-energetic electron carrying most of the muon rest energy

12. University of Arizona Physics Colloquium, March 7, 2008 E Prebys  e Conversion vs.  e  12 Courtesy: A. de Gouvea ? ? ? Sindrum II MEGA MEG proposal We can parameterize the relative strength of the dipole and four fermi interactions. This is useful for comparing relative rates for  N  eN and  e 

13. University of Arizona Physics Colloquium, March 7, 2008 E Prebys History of Lepton Flavor Violation Searches 1 10 -2 10 -16 10 -6 10 -8 10 -10 10 -14 10 -12 1940 1950 1960 1970 1980 1990 2000 2010 Initial mu2e Goal   - N  e - N  +  e +   +  e + e + e - K 0    + e - K +    +  + e - SINDRUM II Initial MEG Goal  10 -4 10 -16

14. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Example Sensitivities* Compositeness Second Higgs doublet Heavy Z’, Anomalous Z coupling Predictions at 10 -15 Supersymmetry Heavy Neutrinos Leptoquarks *After W. Marciano

15. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 15 Sensitivity (cont’d) Examples with  >>1 (no  e  signal):  Leptoquarks  Z-prime  Compositeness  Heavy neutrino SU(5) GUT Supersymmetry:  << 1 Littlest Higgs:   1 Br(  e  ) Randall-Sundrum:   1 MEG mu2e 10 -12 10 -14 10 -16 10 -11 10 -13 10 -15 R(  Ti  eTi) 10 -13 10 -11 10 -9 Br(  e  ) 10 -16 10 -10 10 -14 10 -12 10 -10 R(  Ti  eTi)

16. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Decay in Orbit (DIO) Backgrounds: Biggest Issue Very high rate Peak energy 52 MeV Must design detector to be very insensitive to these. Nucleus coherently balances momentum Rate approaches conversion (endpoint) energy as (E s -E) 5 Drives resolution requirement. 16 N Ordinary: Coherent:

17. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 17 Previous muon decay/conversion limits (90% C.L.) Rate limited by need to veto prompt backgrounds!  >e Conversion: Sindrum II LFV  Decay: High energy tail of coherent Decay-in-orbit (DIO)

18. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 18 Mu2e (MECO) Philosophy Eliminate prompt beam backgrounds by using a primary beam with short proton pulses with separation on the order of a muon life time Design a transport channel to optimize the transport of right-sign, low momentum muons from the production target to the muon capture target. Design a detector to strongly suppress electrons from ordinary muon decays ~100 ns ~1.5  s Prompt backgrounds live window

19. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 19 Signal Single, monoenergetic electron with E=105 MeV, coming from the target, produced ~1  s (   Al ~ 880ns) after the “  ” bunch hits the target foils Need good energy resolution: ≲ 0.200 MeV Need particle ID Need a bunched beam with less than 10 -9 contamination between bunches

20. University of Arizona Physics Colloquium, March 7, 2008 E Prebys negligible 95.56 MeV10.08 MeV.0726  s~0.8-1.5Au(79,~197) 0.16 0.45 Prob decay >700 ns 104.18 MeV 104.97 MeV Conversion Electron Energy 1.36 MeV.328  s1.7Ti(22,~48) 0.47 MeV.88  s1.0Al(13,27) Atomic Bind. Energy(1s) Bound lifetime R  e (Z) / R  e (Al) Nucleus  Aluminum is nominal choice for Mu2e Choosing the Capture Target Dipole rates are enhanced for high-Z, but Lifetime is shorter for high-Z  Decreases useful live window Also, need to avoid background from radiative muon capture  Want M(Z)-M(Z-1) < signal energy

21. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 21 mu2e Muon Beam and Detector MECO spectrometer design for every incident proton 0.0025   ’s are stopped in the 17 0.2 mm Al target foils

22. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 22 Production Region Axially graded 5 T solenoid captures low energy backward and reflected pions and muons, transporting them toward the stopping target Cu and W heat and radiation shield protects superconducting coils from effects of 50kW primary proton beam Superconducting coils Production Target Heat & Radiation Shield Proton Beam 5 T 2.5 T

23. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 23 Transport Solenoid Curved solenoid eliminates line-of-sight transport of photons and neutrons Curvature drift and collimators sign and momentum select beam dB/ds < 0 in the straight sections to avoid trapping which would result in long transit times Collimators and pBar Window 2.5 T 2.1 T

24. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 24 Detector Region 1 T 2 T Axially-graded field near stopping target to sharpen acceptance cutoff. Uniform field in spectrometer region to simplify momentum analysis Electron detectors downstream of target to reduce rates from  and neutrons Stopping Target Foils Straw Tracking Detector Electron Calorimeter

25. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Magnetic Field Gradient 25 Production Solenoid Transport Solenoid Detector Solenoid

26. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Transported Particles 26 E~3-15 MeV Vital that e- momentum < signal momentum

27. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Tracking Detector/Calorimeter 27 Coherent Decay-in- orbit, falls as (E e -E) 5

28. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 28 A long time coming 1992MELC proposed at Moscow Meson Factory 1997 MECO proposed for the AGS at Brookhaven as part of RSVP (at this time, experiment incompatible with Fermilab) 1998-2005 intensive work on MECO technical design: magnet system costed at $58M, detector at $27M July 2005RSVP cancelled for financial reasons 2006 MECO subgroup + Fermilab physicists work out means to mount experiment at Fermilab June 2007mu2e EOI submitted to Fermilab October 2007LOI submitted to Fermilab Fall 2008mu2e submits proposal to Fermilab 2010technical design approval: start of construction 2014first beam

29. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 29 Enter Fermilab  Fermilab  Built ~1970  200 GeV proton beams  Eventually 400 GeV  Upgraded in 1985  900GeV x 900 GeV p-pBar collisions  Most energetic in the world ever since  Upgraded in 1997  Main Injector-> more intensity  980 GeV x 980 GeV p-pBar collisions  Intense neutrino program  Will become second most energetic accelerator (by a factor of seven) when LHC comes on line ~2009  What next???

30. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 30 The Fermilab Accelerator Complex MiniBooNE/BNB NUMI

31. University of Arizona Physics Colloquium, March 7, 2008 E Prebys microBooNE, August 20 th, 2007 - Prebys 31 Preac(cellerator) and Linac “Preac” - Static Cockroft-Walton generator accelerates H- ions from 0 to 750 KeV. “Old linac”(LEL)- accelerate H- ions from 750 keV to 116 MeV “New linac” (HEL)- Accelerate H- ions from 116 MeV to 400 MeV

32. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 32 Booster Accelerates the 400 MeV beam from the Linac to 8 GeV Operates in a 15 Hz offset resonant circuit Sets fundamental clock of accelerator complex From the Booster, 8 GeV beam can be directed to The Main Injector The Booster Neutrino Beam (MiniBooNE) A dump. More or less original equipment

33. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 33 Main Injector/Recycler The Main Injector can accept 8 GeV protons OR antiprotons from Booster The anti-proton accumulator The Recycler (which shares the same tunnel and stores antiprotons) It can accelerate protons to 120 GeV (in a minimum of 1.4 s) and deliver them to The antiproton production target. The fixed target area. The NUMI beamline. It can accelerate protons OR antiprotons to 150 GeV and inject them into the Tevatron.

34. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 34 Present Operation of Debuncher/Accumulator Protons are accelerated to 120 GeV in Main Injector and extracted to pBar target pBars are collected and phase rotated in the “Debuncher” Transferred to the “Accumulator”, where they are cooled and stacked Not used for NOvA

35. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 35 Producing ~10 18   6 batches x 4x10 12 /1.33 s x 2x10 7 s/yr = 3.6x10 20 protons/yr mu2e Note: 8 GeV booster energy is the optimal energy for mu2e muon beam

36. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 36 Proposed Location Requires new building. Minimal wetland issues. Can tie into facilities at existing experimental hall.

37. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 37 What we Get Proton flux1.8x10 13 p/s Running time2x10 7 s Total protons3.6x10 20 p/yr   stops/incident proton0.0025   capture probability0.60 Time window fraction0.49 Electron trigger efficiency0.90 Reconstruction and selection efficiency0.19 Detected events for R  e = 10 -16 4.5

38. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 38 Three Types of Backgrounds Muon decay in orbit:   → e  E e < m  c 2 – E NR – E B N  (E 0 - E e ) 5 Fraction within 3 MeV of endpoint  5x10 -15 Defeated by good energy resolution Radiative muon capture:   Al →  Mg  endpoint 102.5 MeV 10 -13 produce e - above 100 MeV 1. Stopped Muon Induced Backgrounds

39. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 39 Backgrounds (continued) 2. Beam Related Backgrounds Suppressed by minimizing beam between bunches –Need ≲ 10 -9 extinction –Get 10 -3 for free Muon decay in flight:   → e  Since E e 77 GeV/c Radiative   capture:   N → N* ,  Z → e  e  Beam electrons Pion decay in flight:   → e  e 3. Asynchronous Backgrounds Cosmic rays suppressed by active and passive shielding

40. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 40 The Bottom Line Roughly half of background is beam related, and half interbunch contamination related Total background per 3.4x10 20 protons, 2x10 7 s:0.43 events Signal for R  e = 10 -16 :5 events Single even sensitivity: 2x10 -17 90% C.L. upper limit if no signal:6x10 -17 Blue text: beam related.

41. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Possible Future: “Project X” Three 5 Hz pulses every 1.4 s Main Injector cycle = 2.3MW at 120 GeV This leaves four pulses (~200 kW) available for 8 GeV physics These will be automatically stripped and stored in the Recycler, and can also be rebunched there. 41

42. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 42 Experimental Challenges for Increased Flux Achieve sufficient extinction of proton beam.  Current extinction goal directly driven by total protons Upgrade target and capture solenoid to handle higher proton rate  Target heating  Quenching or radiation damage to production solenoid Improve momentum resolution for the ~100 MeV electrons to reject high energy tails from ordinary DIO electrons.  Limited by multiple scattering in target and detector planes  Requirements at or beyond current state of the art. Operate with higher background levels.  High rate detector Manage high trigger rates All of these efforts will benefit immensely from the knowledge and experience gained during the initial phase of the experiment. If we see a signal a lower flux, can use increased flux to study in detail  Precise measurement of R  e  Target dependence  Comparison with  e  rate

43. University of Arizona Physics Colloquium, March 7, 2008 E Prebys However, the future has some uncertainty! 43 (from Dep. Director Y-K Kim)

44. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 44 Conclusions We have proposed a realistic experiment to measure Single event sensitivity of R  e =2x10 -17 90% C.L. limit of R  e <6x10 -17 ANY signal = Beyond Standard Model physics This represents an improvement of more than four orders of magnitude compared to the existing limit, or over a factor of ten in effective mass reach. For comparison –TeV -> LHC = factor of 7 –LEP 200 -> ILC = factor of 2.5 Potential future upgrades could increase this sensitivity by one or two orders of magnitude ANY signal would be unambiguous proof of physics beyond the Standard Model The absence of a signal would be a very important constraint on proposed new models.

45. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Backup Slides 45

46. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Project X Linac 46 β=1 Modulator β=1 Modulator 36 Cavites / Klystron ILC LINAC 8 Klystrons 288 Cavities in 36 Cryomodules 1300 MHz β=1 β<1 ILC LINAC 2 Klystrons 96 Elliptical Cavities 12 Cryomodules 1300 MHz 0.1-1.2 GeV β=1 Modulator β=1 Modulator β=1 Modulator β=1 Modulator β=1 Modulator β=1 Modulator 10 MW ILC Multi-Beam Klystrons 48 Cavites / Klystron β=.81 Modulator β=.81 8 Cavites / Cryomodule 0.5 MW Initial 8 GeV Linac 11 Klystrons (2 types) 449 Cavities 51 Cryomodules “PULSED RIA” Front End Linac 325 MHz 0-110 MeV H-RFQMEBTRTSRSSRDSR Single 3 MW JPARC Klystron Multi-Cavity Fanout at 10 - 50 kW/cavity Phase and Amplitude Control w/ Ferrite Tuners DSR β=.47 Modulator β=.47β=.61 or… 325 MHz Spoke Resonators Elliptical Option Modulator 10 MW ILC Klystrons

47. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Helical Cooling Channel 47 A helical cooling channel (similar to a “Siberian Snake”) provides transverse cooling of muon beam: This, together with an ionizing degrader could allow the forward muons to be used, for a much higher efficiency.

48. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 48 The Big Picture: Goals of Experiment Initial Phase:  Exploit post-collider accelerator modifications at Fermilab to mount a  ->e conversion experiment patterned after proposed MECO experiment at BNL 4x10 20 protons in ~2 years Measure Single event sensitivity of R  e =2x10 -17 90% C.L. limit of R  e <6x10 -17 ANY signal = Beyond Standard Model physics Ultimate goal  Take advantage of intense proton source being developed for Fermilab (“Project X”) as well as muon collider R&D If no signal: set limit R  e <1x10 -18 If signal: measure target dependence, etc

49. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 49 Beam Related Rates Cut ~700 ns after pulse to eliminate most serious prompt backgrounds.

50. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Proton Timeline: Now and Post-Collider In order to increase protons to the NOvA neutrino experiment after the collider program ends, protons will be “stacked” in the Recycler while the Main Injector is ramping, thereby eliminating loading time. 50 15 Hz Booster cycles Present Operation: “wasted” loading time

51. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 51 Available Protons: NOvA Timeline Roughly 6*(4x10 12 batch)/(1.33 s)*(2x10 7 s/year)=3.6x10 20 protons/year available MI uses 12 of 20 available Booster Batches per 1.33 second cycle Preloading for NOvA Available for 8 GeV program Recycler Recycler  MI transfer 15 Hz Booster cycles MI NuMI cycle (20/15 s)

52. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 52 Delivering Protons: “Boomerang” Scheme Deliver beam to Accumulator/Debuncher enclosure with minimal beam line modifications and no civil construction. Recycler (Main Injector Tunnel) MI-8 -> Recycler done for NOvA New switch magnet extraction to P150 (no need for kicker)

53. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 53 Momentum Stacking Inject a newly accelerated Booster batch every 67 mS onto the low momentum orbit of the Accumulator The freshly injected batch is accelerated towards the core orbit where it is merged and debunched into the core orbit Momentum stack 3-6 Booster batches T<133ms T=134ms T=0 Energy 1 st batch is injected onto the injection orbit 1 st batch is accelerated to the core orbit T<66ms 2nd Batch is injected T=67ms 2 nd Batch is accelerated 3 rd Batch is injected

54. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 54 Rebunching in Accumulator/Debuncher Momentum stack 6 Booster batches directly in Accumulator (i.e. reverse direction) Capture in 4 kV h=1 RF System. Transfer to Debuncher Phase Rotate with 40 kV h=1 RF in Debuncher Recapture with 200 kV h=4 RF system  t ~40 ns

55. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 55 Resonant Extraction Exploit 29/3 resonance Extraction hardware similar to Main Injector  Septum: 80 kV/1cm x 3m  Lambertson+C magnet ~.8T x 3m

56. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 56 Beam Extinction Need 10 -9 Get at least ~10 -3 from beam bunching Remainder from AC Dipole in beam line Working with Osaka (FNAL+US-Japan funds) to develop AC dipole design, as well as explore measurement options

57. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 57 Expected Background (from MECO TDR) For 4x10 20 protons on target: SourceEventsComments  decay in orbit0.25 S/N = 20 for R  e = 10 -16 Tracking errors< 0.006 Radiative  decay< 0.005 Beam e - < 0.04  decay in flight< 0.03Without scattering in stopping target  decay in flight 0.04With scattering in stopping target  decay in flight< 0.001 Radiative  capture0.07From out of time protons Radiative  capture0.001From late arriving pions Anti-proton induced0.007Mostly from   Cosmic ray induced0.004Assuming 10 -4 CR veto inefficiency Total Background0.45Assuming 10 -9 inter-bunch extinction Signal Events5For R  e = 10 -16

58. University of Arizona Physics Colloquium, March 7, 2008 E Prebys 58 Cost and Time Scale A detailed cost estimate of the MECO experiment had been done just before it was cancelled*  Solenoids and cryogenics: $58M  Remainder of experimental apparatus: $27M Additional Fermilab costs have not been worked out in detail, but are expected to be on the order of $10M. Hope to begin Accelerator work along with NOvA upgrades  ~2010 (or 2011 if Run II extended) Based on the original MECO proposal, we believe the experiment could be operational within five years from the start of significant funding  Driven by magnet construction.  ~2014 With the proposed beam delivery system, the experiment could collect the nominal 4x10 20 protons on target in about one to two years, with no impact on NOvA  NOvA rate limited by Main Injector * Costs in 2005 dollars, including contingency

59. University of Arizona Physics Colloquium, March 7, 2008 E Prebys Tracking Detector/Calorimeter 3000 2.6 m straws   (r,  ) ~ 0.2 mm 17000 Cathode strips   z) ~ 1.5 mm 1200 PBOW4 cyrstals in electron calorimeter   E/E ~ 3.5% Resolution:.19 MeV/c 59