The SINDRUM-I Experiment

,


History -how it all began
In the fall of 1976 rumors spread about an experiment performed at SIN for the search of the decay µ → eγ.A debate was going on, whether or not the decay had been observed.The rumors traveled from SIN via email to R. Eichler at Stanford and from him to a graduate student in the lecture-class of James Bjorken.The next week, J. Bjorken in turn gave the students an exercise to compute the decay rate and also confronted his colleague Steven Weinberg with the rumor.It took a few weeks after Weinberg's talk at the APS meeting to reach the New York Times.There it read on February 8 th 1977: Experimenters in Switzerland have reportedly observed an "impossible" transmutation of atomic particles.This has thrown the world community of theoretical physicists into a frenzy of speculations, calculations and publications (S.Weinberg).This inspired R. Hofstadter of Stanford to initiate an experiment at LAMPF for µ + → e + γ to try to resolve the dispute around the SIN experiment.
The results from the SIN experiment were finally published as an upper limit for the muon decay µ → eγ.However, all these speculations triggered a wider range of searches of muon flavour violating decays at LAMPF and SIN, and these activities continue presently at PSI, Fermilab and J-PARC.

The lepton flavour violating process µ
In the Standard Model (SM), charged lepton-flavour-violating reactions (LFV) are forbidden at tree level and can only be induced by lepton mixing through higher-order diagrams.One of the dominant contributions, the mixing through loop diagrams with massive neutrinos, see Not all of these models have survived with equal popularity today.However, modern models also include new particles such as Higgs particles or doubly charged Higgs particles, Rparity-violating scalar neutrinos, supersymmetric particles and new heavy vector bosons.The main focus of the SINDRUM I experiment was the search for the decay µ + → e + e − e + [2][3][4], with the aim to improve the sensitivity substantially beyond the then existing limits of B µ→3e < 1.9 × 10 −9 at 90% C.L. [5].
The unique kinematic topology of the 3-body decay was exploited in the analysis, namely been able to observe a handful of events (≤ 7 events).Thus, this was the first statistically significant observation of the µ + → e + e − e + ν e νµ decay.

Measurement of π
In the decays π + → e + ν e γ and π + → e + ν e e − e + , both the vector-and axial-vector weak hadronic currents contribute to the decay amplitudes and are parameterized by the vector and axial vector form factors F V and F A , respectively.There is a firm prediction for the value of F V .The conserved vector current rule connects F V with the π 0 lifetime so that |F V | = 0.0255, but the sign is undetermined.Contrary to the case of π + → e + ν e γ , the ratio of F A /F V is unambiguously measurable in the decay π + → e + ν e e − e + and the result of [6] excludes a possible negative value of F A /F V from the π + → e + ν e γ experiments.In the high statistics run of SINDRUM-I [7] the first determination of was achieved, where the first error is the statistical uncertainty and the second error is due to the uncertainty of the form factors.This B π + →e + ν e e − e + still holds as of this writing.By fixing the value F V =0.0255 the form factor F A = 0.019 ± 0.008 was determined.

Search for light particles produced in muon-or pion decays
Many theories beyond the Standard Model predict "massless" Nambu-Goldstone bosons arising from the breaking of an underlying symmetry.Examples are the "familon" for a broken family hierarchy, the "axion" for a broken axial baryon number proposed to solve the strong CP problem, the majoron, and neutral scalar bosons.
In the search for a light Higgs h in the decay π + → e + ν e h, where the Higgs decays in h → e + e − , the same selection criteria as for the analysis of the pion form factors were applied [7].Higgs particles with a decay length less than the vertex resolution of the SINDRUM detector should be visible in the decay π + → e + ν e e − e + as a peak in the e + e − -invariant mass distribution.No such signal was observed for Higgs masses 2m e < m h < 110 MeV/c 2 .
A similar search was made for an axion-like neutral particle produced in both µ or π decays, µ + → e + φ and π + → e + νφ, with a subsequent decay φ → e + e − .No candidates were found, and therefore upper limits for the branching ratios were determined as a function of the φ masses and lifetimes.For φ lifetimes below 10 −10 s limits on B down to 2 × 10 −12 were obtained [8].
In the decay π 0 → e + e − γ, the hadronic structure of the pion is parameterized by a form factor F = 1/(1 − a x) with x = m e + e − /m π 0 .The SINDRUM-I analysis of the Dalitz plot distribution measured the value as a = 0.02 ± 0.02 ± 0.04 [11] with the uncertainties being statistical and systematic, respectively.This value is consistent with the prediction of vector meson dominance of a ≈ 0.03.

General description of the SINDRUM-I Apparatus
A schematic view of the SINDRUM spectrometer is given in Figure 7.

The low mass multiwire proportional chamber (MWPC)
A main issue of concern for the design of SINDRUM was multiple scattering of the low-energy electrons.A very low mass for the target and the tracking chambers was a real challenge.
The spectrometer was equipped with five very thin cylindrical MWPCs, three of which had cathode strip readouts.Each chamber consisted of two concentric Kapton/Rohacell sandwich cylinders, which were assembled on steel mandrels.Glass-fiber epoxy rings were glued to the ends of the cylinders supporting printed circuit rings onto which the 20µm anode wires, resistors, condensors, and multipin connectors were soldered.The cathodes of chambers 1, 3, and 5 consisted of strips of aluminum evaporated on Kapton having an angle of ±45 • for the outer and inner cathodes, respectively.The strips were connected to end printed circuit boards with conductive paint.The strips of chamber 1 were divided in the middle and read out at both ends of the chamber to reduce the rate per strip.The chambers were operated with a gas mixture of 49.9% Ar, 49.9% C 2 H 6 and 0.2% freon at a gas gain of ∼ 5 × 10 4 .The chamber electrodes were connected through 1 m long 75Ω coaxial cables to the amplifiers mounted around the circumference of the magnet.The spatial resolution of the ϕ-measurement was limited by the wire spacing of 2 mm (σ 0.6 mm) and the z-resolution was determined with cosmic rays to be σ 0.3 mm.The chambers were successfully operated throughout

Summary
The highlight of the SINDRUM-I experiment is clearly the improvement of the sensitivity on the rare decay µ + → e + e − e + by three order of magnitudes, reaching an upper limit BR < 1.0× 10 −12 at 90% C.L. The experiment was statistically limited and was not suffering from backgrounds.
However, to gain another order of magnitude in precision, a much higher intensity of the muon beam would have been required.Thus, the successor experiment, SINDRUM-II, concentrated on the complementary muon-electron conversion process.As the SINDRUM-I detector resolution was not sufficient for competitive µ − e-conversion measurements, a major upgrade of the detector was done, followed by measurements achieving best upper limits for µ − econversion [12,13].

Figure 7 .
Figure 7.1a, is strongly suppressed in the SM with a predicted branching ratio B below the level of 10 −50 [1] .Thus, the decay µ + → e + e − e + potentially provides very high sensitivity to LFV reactions in various models of physics Beyond the Standard Model, in which the couplings are mediated by completely new particles.At the time of the SINDRUM-I experiment, lepton flavour violation in the neutral lepton sector (neutrino oscillations) were not yet established, and theories were focused on extensions of the SM by introducing different new heavy particles that can mediate charged LFV either in virtual loops (Figure 7.1b), at tree level (see Figure 7.1c), or in box diagrams.These new models included right-handed bosons, additional Higgs doublets, neutral scalar singlets, familons, extended technicolor gauge bosons, doubly charged so-called "heptons", various "horizontal" models, and notably supersymmetric (SUSY) models with scalar leptons.An example is Figure 7.1b, in which a γ/Z-penguin diagram is shown with new SUSY particles running in a loop.These loop contributions are important for all models where new particle couplings to electrons and muons are introduced.

Furthermore, a search 7 . 3 . 5
for weakly interacting neutral bosons (X) produced in π − p interactions at rest and decaying into e + e − pairs was performed with the SINDRUM detector.The data sample searched contained 98400 π 0 → e + e − γ decays and 27200 π − p → ne + e − events, each with an e + e − invariant mass between 25 and 139 MeV/c.Upper limits for the branchingratios Γ (π 0 → X γ, X → e + e − )/Γ (π 0 → al l) and Γ (π − p → X n, X → e + e − )/Γ (π − → al l) forX lifetimes between 10 −23 s and 10 −11 s were obtained.Upper limits at 90% C.L. range from 10 −3 at an invariant e + e − mass of 25 MeV/c 2 to 10 −5 at 100 MeV/c 2[9].Measurement of the decay π 0 → e + e − and π 0 → e + e − γ The large helicity suppression of the electromagnetic amplitude of the decay π 0 → e + e − has led to speculations that additional contributions might be important.Anomalous quark-lepton couplings could lead to significant enhancements of the value for this branching ratio.A branching ratio above the unitarity value would be a sign of CP violating neutral currents.The reaction π − p → π 0 n at rest was used as a source of tagged mono -energetic π 0 in a search for the decay π 0 → e + e − with the SINDRUM I spectrometer.The measurement resulted in[10]

3 ,
with the coordinate system shown.With the help of the evacuated solenoid S, a surface muon beam with momentum 25 MeV/c and intensity 7 × 10 6 s −1 (produced by a 120 µA proton current extracted from the cyclotron) was refocussed from the entrance collimator to the target T, where it stopped.The target was a hollow double-cone shaped body of 58 mm diameter and 220 mm length made of Rohacell 1 with a thickness of 1 mm (11 mg/cm 2 ).The cylindrical magnet with a normal conducting coil M produced a homogeneous (∆B/B < 1%) magnetic field of up to 0.6 T parallel to the symmetry axis (z-axis) in a volume of 110 cm length × 75 cm diameter.Tracks of decay particles were measured with five concentric self-supporting cylindrical multiwire proportional chambers C of low mass density.Three of them were equipped with cathode strips in order to obtain z-coordinates for three-dimensional reconstruction of tracks.For a field of B = 0.334 T, as used in the experiment, the momentum resolution is ∆p/p = (12.0±0.5)% and (8.5 ± 0.5)% (FWHM) for p = 50 MeV/c and 20 MeV/c, respectively.The angular resolution at the target is ∆θ = (65 ± 3) mrad (FWHM) for tracks of 20 MeV/c momentum.Fast timing signals were obtained from the cylindrical scintillator hodoscope H placed between the coil M and the chambers C. The 64 hodoscope elements were viewed at both ends by photomultipliers P. A time resolution of ∆t = 0.57 ns (FWHM) between two hodoscope counters was obtained after correcting for walk and time of flight.The solid angle covered by the spectrometer was 0.73 of 4π.

Figure 7 . 3 :
Figure 7.3: The SINDRUM I detector in the horizontal operating orientation.

Figure 7 . 4 :
Figure 7.4: The assembly of the SINDRUM I detector in the vertical orientation.The MWPC are being lowered into the setup by (clockwise from top left) Erwin Hermes (technician UZH), Norbert Kraus (PhD student UZH), Nik Lordong (Technician PSI), and within the setup Michael Doser (Master student ETHZ).