The MECO experiment A search for lepton flavor violation in muonic atoms

a r X i v :h e p -e x /0101017v 1 13 J a n 2001The MECO experiment:A search for lepton flavor violation in muonic atomsJames L.Popp for the MECO CollaborationDepartment of Physics,New York University,New York,NY 100031IntroductionLepton-flavor-violating transition searches are likely to be sensitive probes of super-unified theories since the predicted rates are often nearest experimental limits.Although pre-cise predictions of theories depend on the specific model,the physi-cal mechanisms that lead to lepton flavor violations are generic to su-persymmetric quark-lepton unifica-tion.Many extensions to the Stan-dard Model that unify quarks and leptons,including supersymmetric theories, e.g.,[1],suggest that in muonic atoms the branching frac-tion for coherent conversion of a muon into an electron relative to muon capture in the nucleus R µe ∼10−14−10−17over much of the param-eter space.In terms of decay rates,with mass number A and atomic number Z ,R µe =Γ(µ−N(A,Z )→e −+N(A,Z ))/Γ(µ−N(A,Z )→νµ+N(A,Z −1)),where the nu-merator only involves the internal nuclear ground state and denomina-tor includes all possible nuclear final states.The MECO experiment [2]will search for direct evidence for this lepton-flavor-violating process.The signature of coherent muon-electron conversion is a two-body final state with a mono-energetic E o electron.When the internal state of the nucleus remains in the ground state during the tran-sition the electron recoils coher-ently offthe entire nucleus,result-ing in a strong enhancement in the conversion rate.The electron en-ergy E o ≃E µ+E 2µ/2M N with E µ=m µ+BE,the muon mass plus Coulomb binding energy BE.The Schr¨o dinger equation gives good es-timates of BE for light nuclei;ne-glecting the size of the nucleus in alu-minum,the MECO stopping target material,BE(1s)∼−0.48MeV and E o∼105.0MeV.MECO will look for conversion electrons in the energy window103.6-105.1MeV,where the signal to background ratio is∼20.The dominant background as in previous experiments[3,4]is muon decay in orbitµ−N→νµ+¯νe+e−+ N;the electron energy spectrum falls rapidly near the endpoint at exactly the conversion energy,∝(E o−E e)5. Thus background rejection improves rapidly with energy resolution.The MECO tracker probes this crucial re-gion more deeply than ever before.The current measured limit set by Sindrum II[3]is Rµe<6.1×10−13 at90%confidence in titanium,with sensitivity expected to improve soon to2×10−14.This experiment is limited by beam intensity,not back-ground.With a significantly scaled-up approach to muon production and transport,∼1.8×1011µ−/s at the stopping target and at least 104greater than existing low-energy (p<100MeV/c)muon beams, MECO is expected to improve on these limits by at least three to four orders of magnitude Rµe<5×10−17. 2The experimentThe MECO setup is shown in Figure1and consists of three large superconducting magnets:the muon production,transport,and detec-tor solenoids.The clear bores of the magnets are connected forming a single cavity,held in vacuum.The magneticfield is continuous through-out the cavity.In the4m long pro-duction solenoid,the axialmagneticFig.1.The MECO experimentfield intensity B is graded from3.6T to2.3T.Thefield follows the12m S-shaped transport solenoid,falling only in the straight sections from 2.3T to2.0T.In the10.5m detec-tor solenoid B is graded from2.0T at the entrance to1.0T between the stopping target and tracking detec-tor,and constant thereafter.Both the trigger and tracker are high-rate detectors with their detection mate-rial positioned far from the solenoid axis to intercept conversion electron helices and avoid beam particle in-teractions.The primary proton beam is pro-vided by the Alternating Gradient Synchrotron(AGS),with∼1sec machine cycle for40×1012pro-tons.The choice of beam energy, 7-8GeV,optimizesπand minimizes ¯p(a potential background source) production.The beam(not shown) enters the production solenoid on the right at10o to the axis and ex-its to the left.A tungsten target is bombarded to produceπ±which subsequently decay toµ±,following helical trajectories along the mag-net axis.The gradedfield reflects left-moving charged particles with momenta outside the30o(half an-gle)loss cone,toward the transport entrance.Since particle energy andp2⊥/B are constant,as charges move to regions of lower B the angleχbe-tween the momentum and solenoid axis decreases.Pions with transverse momenta p⊥<180MeV/c travel within the30cm inner radius of the magnet.Mostπdecays occur in the production region.A pulsed beam with good extinc-tionεbetween pulses is critical to prompt background rejection and thus success for MECO.The mean lifetime forµ−in aluminum atoms τAlµ=880ns sets the pulse sepa-ration time scale.Thus the pulse spacing is microseconds and given by τpulse+τdelay+τobs.The actual pro-ton pulse timeτpulse≪τAlµ.Prompt backgrounds are reduced to an ac-ceptable level by,first,introducting a delayτdelay between each proton pulse and the detection windowτobs so that all beam particles have left the detector region beforeτobs;and second,by requiring that the ratio of protons crossing the production target during the observation win-dow to that during the pulse be ε∼10−9.AGS extinction studies are ongoing with current measured value ε∼10−7.The transport solenoidfilters the particleflux producing a momentum-and charge-selected muon beam, with good reduction in contami-nation from e±,µ+,π±,p,and¯p. Transport entrance and exit collima-tors with15cm inner radius further limit the beam.For sufficiently low momenta,centers of particle orbits follow the S-shape of the solenoid; however,the curvature of the mag-neticfield in the two bent sections gives rise to a drift out of the plane of the magnet.In thefirst section pos-itives drift up and negatives down, and in the second directions are re-versed.By vertically limiting the aperture of a narrower collimator in the center straight section with a downward offset positives and high-energy negatives are absorbed from the beam.Low-energy antiprotons originating in the production region have long transport crossing times, and as already mentioned the po-tential to create backgrounds at the detectors.A thin beryllium window in the center section reduces the ¯p component to a negligable level. The shallow grades in B in each straight section further prevents late arrivals.At the transport exit e+ andµ+have p<40MeV/c,most e−have p<80MeV/c with none above100MeV/c.Theµ−beam has 14MeV/c<p<100MeV/c.Onlyµ−below50MeV/c stop in the target.Negative muons come to rest in matter by slowing to thermal-like velocities through in-elastic atomic collisions and falling into a Coulomb orbit about a nu-cleus.Excited states cascade to lower ones with time scale10−13s pro-ducing X-rays and Auger electrons. Approximately60%are captured by the nucleus and40%decay in orbit.The stopping target design op-timizes the probability for muon stopping and conversion electron de-tection,and minimizes energy loss of exiting signal e−and the number of decay-in-orbit electrons reaching the detectors.Background rates are re-duced by minimizing stopping target mass.The target has17parallel disks 0.02cm thick with5.0cm spacing and radii from8.3-6.5cm;total mass is159g.Simulations yield a stoppingefficiency0.0025µ−per proton,i.e., 1011Hz stop rate.The target is centered in the gradedfield between1.63-1.37T; the gradient reflects e−emitted up-stream back to the detectors,re-sulting in∼60%of all conversions hitting the tracker.Conversion elec-trons with p⊥>90MeV/c(120o>χ>60o)are swept forward by the decreasing magneticfield into the range75MeV/c<p⊥<90MeV/c (45o<χ<60o)at the track-ing detector.Beam particles with p⊥<90MeV/c that do not scatter in the target pass down the center of the solenoid without intercepting the detectors.Placing the target in a gradedfield also ensures that e−originating upstream of the gradient with105MeV/c arrives at the detec-tors with p⊥<75MeV/c,eliminat-ing many potential backgrounds.On average the stopping target produces 2n,2γ,and0.1p per muon cap-ture,displacing the detectors>1m downstream from the target greatly reduces acceptance for n andγ.The straw tube drift chamber tracking detector has large accep-tance for and measures with good efficiency the helix parameters of conversion electrons in the uniform field.Good energy resolution is es-sential to distinguish conversion elec-trons from decay in orbit,<∼900keV FWHM.Resolution is dominated by scattering in the tracker,and to a much lesser extent by pattern recog-nition errors.Energy loss in the tar-get,proton absorbers(not shown), and tracker broadens the peak of the resolution function for conversions, and introduces a small mean energy loss and low-energy tail.The tracker has16rectangular planes,8forming an octagon cen-tered on the solenoid axis and8 positioned at each corner extending radially.This geometry gives≥3 plane crossings per conversion e−helix orbit;3m long detector planes guarantee≥2helix orbits in the de-tector,a powerful advantage for pat-tern recognition.With tubes parallel to the length of a plane,a slight tilt of each plane,<1o,prevents a con-version e−from entering the same tube on different orbits.Each plane has three layers of 0.25cm radius circular tubes in a close-packed arrangment;each tube contains a central anode wire.Drift time information and wire positions combine to give the track-detector plane crossing position perpendicu-lar to the wires and the track pro-jection angle(valuable for pattern recognition)in a plane perpendic-ular to the wires and wire planes. Position along the wires is measured with capacitively-coupled cathode foils,∼1cm wide,running perpen-dicular to the wires on both sides of a plane.Resolutions from wire infor-mation are highly angle-dependent with average position and angle res-olutions,expressed as root mean square(RMS)quantities,0.014cm and1.0o.The expected RMS pad resolution is∼0.05cm.The purpose of the trigger calor-imeter is to select with high effi-ciency conversion electrons that pass through the tracker,while mini-mizing triggers from lower-energy events.A high degree of segmen-tation limits signal pile-up.Since background rises exponentially as the trigger threshold decreases,it isimportant to have E threshold as highas possible.The calorimeter should also have good energy and position resolution to provide an independent measurement of these quantities for the helix,a feature shown to have high discriminant value to suppress background.Trigger design is currently un-der development.The current design (not shown)using single3×3×12cm3 crystals of BGO(300ns EM shower decay constant)and avalanche pho-todiodes show that E threshold= 80MeV gives87%efficiency for conversions.This results in an ac-ceptable0.2kHz trigger rate.The RMS position and energy resolutions for105.0MeV electrons are1.0cm and5.3MeV.Research continues on other materials with shorter decay times,such as GSO and PbWO4.A conversion electron candi-date event has:(1)detector signals within the observation time win-dow,(2)calorimeter shower energy ≥E threshold,(3)detector plane cross-ings found by pattern recognition consistent with a helix of≥2helix or-bits,(4)a helix which when extended back intersects the stopping target and extended forward is consistent with the calorimeter shower posi-tion,(5)p⊥of the helix in the range 75MeV/c<p⊥<90MeV/c,(6)he-lix energy consistent with calorime-ter shower energy,and(7)helix en-ergy between103.6-105.1MeV.3ConclusionTable1summarizes calculations of primary background contribu-tions in the energy window103.6-µ−DIO0.25Pattern Recognition<0.006µ−RC<0.005⋆µ−DIF(a)<0.03⋆µ−DIF(b)0.04⋆π−RC0.07⋆π−DIF<0.001⋆beam e−<0.04π−RC0.001¯p induced0.007CR induced0.004 Table1Expected MECO backgrounds(abbre-viations in text)forε=10−9and107s running time.Stars indicate extinction dependence.105.1MeV for a running time of one-third of a year.Decay in orbit(DIO) and radiative capture(RC)ofµ−are intrinsic to muon decay,both are made negligible with precise energy measurement.Starred entries arise from out-of-time proton crossings at the production target which are pro-portional to the extinction:Decay in flight(DIF)ofµ−without(a)and with(b)scattering in the stopping target,radiative pion capture,DIF forπ−,and beam e−.The next two backgrounds in the table are from source particles with long transit times in the transport solenoid.Ac-ceptable cosmic ray(CR)induced background rejection for MECO re-quires passive and active shielding with a modest improvement over previous[3,4].Inefficiency in the CRµ−stopped/p0.0025P(µ−capture)0.60F(τobs,µ−capture)0.49Trigger efficiency0.90Fitting/selection0.19Rµe10−16Table2Expected MECO sensitivity.shield veto is expected to be10−4. Assumingε∼10−9the total num-ber of background events expected is <0.45.Table2shows that a four-month run and Rµe=10−16yields∼5 muon-electron conversion events. The number of muon captures is just the product of thefirst seven en-tries of Table2,where P(µ−capture) is the probability ofµ−capture, F(τobs,µ−capture)the fraction of µ−capture in the observation win-dow,and the overallfitting and event selection criteria efficiency is given. References[1]R.Barbieri,L.Hall,Phys.Lett.B338(1994)212;R.Barbieri,L.Hall,and A.Strumia,Nucl.Phys.B445(1995)219.[2]MECO Collaboration,Rare Sym-metry Violating Processes’RSVP’:A Proposal to the National ScienceFoundation to Construct theMECO and KOPIO Experiments,(1999),/.[3]Sindrum II Collaboration,Proc.of6th Conference on Intersectionsof Particle and Nuclear Physics,T.Donnelly,ed.(AIP,New York,1997),p.34; C.Dohmen et al.,Phys.Lett.B317(1993)631. [4]S.Ahmad et al.,Phys.Rev.D38(1988)2102.。