BNL AGS Precision Measurement of the Muon (g 2) Value To (continue to) be or not to be? B. Lee Roberts roberts@bu.edu http://physics.bu.edu/roberts.html Department of Physics Boston University B. Lee Roberts, HEPAP 9 February 2002 p.1/27
Outline Introduction Results from E821 Comparison with Theory Work in Progress on the Theory Summary and Conclusions B. Lee Roberts, HEPAP 9 February 2002 p.2/27
Muon (g 2) Collaboration Institutes Boston University, Brookhaven National Laboratory, Budker Institute of Nuclear Physics - Novosibirsk, Cornell University, KEK, KVI and Rijksunviversiteit - Groningen, University of Heidelberg, University of Illinois, University of Minnesota, RIKEN, (Yale) B. Lee Roberts, HEPAP 9 February 2002 p.3/27
Why am I here? The E821 Collaboration has finished the analysis of all of our data. We have reached ±0.7 ppm accuracy both for µ + and µ. B. Lee Roberts, HEPAP 9 February 2002 p.4/27
Why am I here? The E821 Collaboration has finished the analysis of all of our data. We have reached ±0.7 ppm accuracy both for µ + and µ. Our combined ±0.5 ppm result still differs with the standard model. B. Lee Roberts, HEPAP 9 February 2002 p.4/27
Why am I here? The E821 Collaboration has finished the analysis of all of our data. We have reached ±0.7 ppm accuracy both for µ + and µ. Our combined ±0.5 ppm result still differs with the standard model. We are statistics limited. For 2001 data set: σ stat = 0.66 ppm; σ syst = 0.27 ppm! B. Lee Roberts, HEPAP 9 February 2002 p.4/27
Why am I here? The E821 Collaboration has finished the analysis of all of our data. We have reached ±0.7 ppm accuracy both for µ + and µ. Our combined ±0.5 ppm result still differs with the standard model. We are statistics limited. For 2001 data set: σ stat = 0.66 ppm; σ syst = 0.27 ppm! We believe that our final result warrants additional work. B. Lee Roberts, HEPAP 9 February 2002 p.4/27
We received support from DOE to run in 2001 We asked for funds to run also in 2002, and it was suggested that we finish our data analysis before asking for additional resources. B. Lee Roberts, HEPAP 9 February 2002 p.5/27
We received support from DOE to run in 2001 We asked for funds to run also in 2002, and it was suggested that we finish our data analysis before asking for additional resources. This enabled us to focus on the key questions, and to put all of our attention on the data analysis. B. Lee Roberts, HEPAP 9 February 2002 p.5/27
We received support from DOE to run in 2001 We asked for funds to run also in 2002, and it was suggested that we finish our data analysis before asking for additional resources. This enabled us to focus on the key questions, and to put all of our attention on the data analysis. We now understand the experiment much better, and can improve on many aspects, further reducing the systematic errors. B. Lee Roberts, HEPAP 9 February 2002 p.5/27
Education and Training on (g 2) EDUCATION and training provided for 75 young physicists: 22 post-doctoral fellows 14 Ph.D. students, 6 other graduate students 33 undergraduate students. B. Lee Roberts, HEPAP 9 February 2002 p.6/27
A Brief Look at the Experiment Polarized muons produced from a 3 GeV/c pion beam are injected and stored in a superferric storage ring. The spin precesses relative to the momentum: ω a = dθ R dt = e mc ( [a µ B a µ 1 ) γ 2 1 ] β E B. Lee Roberts, HEPAP 9 February 2002 p.7/27
The Storage Ring: ω a = e mc a µb B. Lee Roberts, HEPAP 9 February 2002 p.8/27
And < B > φ for 2001 is: B. Lee Roberts, HEPAP 9 February 2002 p.9/27
Measure e ± from µ eν ν The highest energy e ± from decays carry the spin information. B. Lee Roberts, HEPAP 9 February 2002 p.10/27
The Detector Geometry muon momentum e muon spin Sci Fi Calorimeter module Measures Energy and time p spin forward, more high energy e spin backward, less high energy e 400 MHz digitizer B. Lee Roberts, HEPAP 9 February 2002 p.11/27
Time Spectrum, 4 10 9 e, E > 1.8GeV, σ stat 0.7 ppm f(t) = N 0 e λt [1 + A cos(ω a t + φ)] electron time spectrum (2001) Million Events per 149.2ns 10 1 10-1 32 100 100 200 200 300 µs µs µs 10-2 300 400 400 500 µs µs 10-3 500 600 µs 0 20 40 60 80 100 Time modulo 100µs [µs] B. Lee Roberts, HEPAP 9 February 2002 p.12/27
Where we came from: (10 ppm) (9.4 ppm) CERN CERN µ + µ ~1983 116 590 000 116 591 000 Theory 116 592 000 116 593 000 116 594 000 116 595 000 X 10 11 a µ B. Lee Roberts, HEPAP 9 February 2002 p.13/27
Today: with e + e based theory: (9.4 ppm) CERN (10 ppm) CERN µ (13 ppm) E821 (97) µ + (5 ppm) (1.3 ppm) (0.7 ppm) (0.7 ppm) E821 (98) E821 (99) E821 (00) E821 (01) µ + µ + µ + µ µ + 116 590 000 116 591 000 DEHZ03 e + e 116 592 000 a µ 116 593 000 116 594 000 a µ = 11 659 214(8)(3) 10 10 116 595 000 X 10 11 B. Lee Roberts, HEPAP 9 February 2002 p.14/27
E821 Measurements - 11659000 10 10 230 220 210 + µ - µ Avg. [τ] a µ 200 190 + - [e e ] 180 170 160 150 Experiment Theory Consistency using a blind analysis! B. Lee Roberts, HEPAP 9 February 2002 p.15/27
New Physics Beyond the SM? If the experimental value of a µ does not equal the SM value, a µ (NP) = a µ (Measured) a µ (SM) B. Lee Roberts, HEPAP 9 February 2002 p.16/27
New Physics Beyond the SM? If the experimental value of a µ does not equal the SM value, a µ (NP) = a µ (Measured) a µ (SM) 466 citations on SPIRES to our 2001 paper and 204 citations to our 2002 paper (featured on the PRL cover), most speculating on new physics. Our papers are in the top 1% of most-cited papers. B. Lee Roberts, HEPAP 9 February 2002 p.16/27
New Physics Beyond the SM? If the experimental value of a µ does not equal the SM value, a µ (NP) = a µ (Measured) a µ (SM) 466 citations on SPIRES to our 2001 paper and 204 citations to our 2002 paper (featured on the PRL cover), most speculating on new physics. Our papers are in the top 1% of most-cited papers. e.g. the SUSY contribution depends on m and tan β, so a µ is useful, even after (if) the mass spectrum is measured. B. Lee Roberts, HEPAP 9 February 2002 p.16/27
Theory for Muon (g 2) Q E D H a d µ W e a k γ γ h 692.4 (6.2) X 10 µ γ µ γ W γ µ γ e e + h µ h γ h 10 10 8.5 ( 4.0) X 10 10 W ν µ + 11 658 470.57(.29) X 10 10 γ 10.1 (.6) X 10 µ Z µ +38.9 19.4 < 0.1 1st + 2nd Order Weak = γ µ e µ µ γ e + γ 15.1 (.4) X 10 γ + higher order terms µ H µ 10 + higher order terms γ h µ γ µ γ h B. Lee Roberts, HEPAP 9 February 2002 p.17/27
a µ (Had) from Dispersion Theory µ γ γ h e+ e γ h (0) a µ (had; 1) = ( αm µ 3π )2 4m 2 π ds s 2 K(s)R(s) (0) where R(s) = σ(e+ e hadrons) σ(e + e µ + µ ) B. Lee Roberts, HEPAP 9 February 2002 p.18/27
τ for 2 (π)? What are the issues which might make the τ-decay data complicated to compare with e + e data. µ γ γ h e+ e CVC? It s not perfect. h π + π τ W isospin violation (including ρ mass differences? γ ν τ h π 0 π experimental problems (normalization)? B. Lee Roberts, HEPAP 9 February 2002 p.19/27
CVC Tests: Br(τ ν τ π π 0 ) e+ e ν τ τ γ W π π + π π 0 e + e CVC 24.52 ± 0.32 23 24 25 26 27 B(τ ν τ π π o ) (in %) CLEO 25.42 ± 0.12 ± 0.42 OPAL 25.44 ± 0.17 ± 0.29 L3 25.44 ± 0.16 ± 0.10 ALEPH 25.47 ± 0.10 ± 0.09 τ Average 25.46 ± 0.10 preliminary BR comparison with the prediction using CVC, lepton universality and e + e π + π as input. Discrepancy is 2.9σ. From Davier, Eidelman, Höcker, Zhang, Eur.Phys.J.C31, 503 (2003) B. Lee Roberts, HEPAP 9 February 2002 p.20/27
F π 2 from τ-decay and e + e τ ν τ π π 0 vrs. e + e π + π 0.3 0.3 ( F π 2 [ee] F π 2 [τ]) / F π 2 [τ] 0.2 0.1 0-0.1 τ Average preliminary CMD-2 Aug. 2003 CMD OLYA DM1 ( F π 2 [ee] F π 2 [τ]) / F π 2 [τ] 0.2 0.1 0-0.1 τ Average preliminary CMD-2 Aug. 2003 CMD OLYA DM1-0.2-0.2-0.3 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 s (GeV 2 ) -0.3 0.48 0.5 0.52 0.54 0.56 0.58 0.6 0.62 0.64 0.66 0.68 0.7 s (GeV 2 ) From Davier, Eidelman, Höcker, Zhang, Eur.Phys.J.C31, 503 (2003) With the differences between τ and e + e, the conservative theoretical path is to take e + e data only. B. Lee Roberts, HEPAP 9 February 2002 p.21/27
Worldwide effort on a µ (had)! B. Lee Roberts, HEPAP 9 February 2002 p.22/27
Future Goals We are seeking a path to an improved uncertainty of 0.2 ppm, to further clarify the discrepancy/agreement with theory. B. Lee Roberts, HEPAP 9 February 2002 p.23/27
Future Goals We are seeking a path to an improved uncertainty of 0.2 ppm, to further clarify the discrepancy/agreement with theory. We have 9 billion e + e thus far, so that means about 50 billion in an upgraded experiment. B. Lee Roberts, HEPAP 9 February 2002 p.23/27
Future Goals We are seeking a path to an improved uncertainty of 0.2 ppm, to further clarify the discrepancy/agreement with theory. We have 9 billion e + e thus far, so that means about 50 billion in an upgraded experiment. We need to improve the beamline, as well as detectors and electronics. B. Lee Roberts, HEPAP 9 February 2002 p.23/27
Future Goals We are seeking a path to an improved uncertainty of 0.2 ppm, to further clarify the discrepancy/agreement with theory. We have 9 billion e + e thus far, so that means about 50 billion in an upgraded experiment. We need to improve the beamline, as well as detectors and electronics. Systematic errors have improved substantially with time: B. Lee Roberts, HEPAP 9 February 2002 p.23/27
Errors vs. Time Year Field (ω p ) Spin (ω a ) σ syst σ stat (ppm) (ppm) (ppm) (ppm) 1999 0.4 0.3 0.5 1.28 2000 0.24 0.31 0.39 0.62 2001 0.17 0.21 0.27 0.66 B. Lee Roberts, HEPAP 9 February 2002 p.24/27
Time and Money We think that in 2+ years from now we could rebuild and be ready to run again. The cost would be several $ M for upgrades, and running costs of about $6 M per 5 month run with RHIC. There is a tradeoff between R&D on more muons, and running time. We have working groups which are studying all systems, in order to come up with a plan for the improved experiment. To date we have spent $78 M: construction = $25M; running = $53M. B. Lee Roberts, HEPAP 9 February 2002 p.25/27
Summary E821 was statistics limited. B. Lee Roberts, HEPAP 9 February 2002 p.26/27
Summary E821 was statistics limited. The experiment worked very well, and we understand how to improve it. B. Lee Roberts, HEPAP 9 February 2002 p.26/27
Summary E821 was statistics limited. The experiment worked very well, and we understand how to improve it. Driven by our success, Novosibirsk, KLOE, BaBar, and Belle are all working on data to improve the hadronic contribution. B. Lee Roberts, HEPAP 9 February 2002 p.26/27
Summary E821 was statistics limited. The experiment worked very well, and we understand how to improve it. Driven by our success, Novosibirsk, KLOE, BaBar, and Belle are all working on data to improve the hadronic contribution. KLOE will announce a final result this week which confirms the CMD2 results, using a very different technique, strengthening the e + e evaluation! B. Lee Roberts, HEPAP 9 February 2002 p.26/27
Conclusions For the health and future of the field, we as a community must make sure there is room for small projects with potential large payoffs, as well as the large projects. B. Lee Roberts, HEPAP 9 February 2002 p.27/27
Conclusions For the health and future of the field, we as a community must make sure there is room for small projects with potential large payoffs, as well as the large projects. While HEPAP is not a program committee, you determine the program in HEP. It s crucial for the continued viability of the field that you insist on a balanced program. B. Lee Roberts, HEPAP 9 February 2002 p.27/27
Conclusions For the health and future of the field, we as a community must make sure there is room for small projects with potential large payoffs, as well as the large projects. While HEPAP is not a program committee, you determine the program in HEP. It s crucial for the continued viability of the field that you insist on a balanced program. We believe an upgraded muon (g 2) experiment has an important role to play in that program, and hope that HEPAP will support it. B. Lee Roberts, HEPAP 9 February 2002 p.27/27