The synergy of the golden and silver channels at the Neutrino Factory Pasquale Migliozzi INFN - Napoli Work done in collaboration with D.Autiero, G.De Lellis, A.Donini, M.Komatsu, PM, D.Meloni, R.Petti, L.Scotto Lavina, F.Terranova (hep-ph/ )

Aim of the work

Outline The golden channel The signal, the background and the magnetized iron detector The silver channel The signal, the background and the ECC detector Systematic uncertainties on the signal and on the background Results from the synergy of the golden and the silver channels Conclusion

Caveat As a first step we restrict ourself to the (  13,  ) ambiguity In particular  23 = 45°: no (  23,p/2 –  23 ) ambiguity The sign of  m 2 atm has been kept fixed The other parameters Solar sector  12 = 33° ;  m 2 sol = 1.0 x eV 2 Atmospheric sector  23 = 45° ;  m 2 atm = 2.9 x eV 2 Matter parameter: A = 1.1 x eV 2 /GeV The way how golden and silver channels can be used to solve the other ambiguities will be discussed in the A.Donini talk

The golden channel at the Neutrino Factory

The magnetized iron detector and the kinematical variables  p  /p  as a function of p  40 kton

Signal and background Antineutrino CC Antineutrino NC

Results for golden  at L=3000 km

Comments Even with ten years data taking (five for each polarity) clone regions show-up  use silver channels to eliminate them In the past, systematic errors both on the signal and on the background were not taken into account We will see the effects on the sensitivity of systematic errors in the following

The silver channel at the Neutrino Factory

Golden vs silver events

The Emulsion Cloud Chamber (ECC) Emulsions for tracking, passive material astarget Established technique charmed “X-particle” first observed in cosmic rays (1971) DONUT/FNAL beam-dump experiment: 7  observed (2000) <  m space res. mass Pb  1 mm Emulsion layers track segments Experience with emulsions and/or  searches : E531, CHORUS, NOMAD and DONUT  m 2 = O (10 -3 eV 2 )   M target ~ 2 kton modular structure (“bricks”): basic performance is preserved large detector  sensitivity, complexity required: “industrial” emulsions, fast automatic scanning

8 m Target Trackers Pb/Em. target Electronic detectors  select  interaction brick A “hybrid” experiment at work Emulsion analysis  vertex search Extract selected brick Pb/Em. brick 8 cm Pb 1 mm Basic “cell” Emulsion  decay search  spectrometer  e/  ID, kinematics   ID, charge and p  (DONUT)

Event reconstruction with an ECC High precision tracking (  x < 1  m ;  < 1mrad) Kink decay topology Electron and  /  0 identification Energy measurement Multiple Coulomb Scattering Track counting (calorimetric measurement) Ionization (dE/dx meas.)  /  separation e/  0 separation Topological and kinematical analysis event by event 5X 0 ( ~ ½ brick) 1 mm 5 cm ECC exposure at CERN-PS 01mm 

Muon identification      The charge mis-assignment in the energy range of interest for the background (>20GeV) is %

 “Long” decays (   ) kink angle  kink > 20 mrad  “Short” decays (   ) impact parameter I.P. > 5 to 20  m kink  kink Long decays Pb (1 mm) plastic base I.P. Short decays emulsion layers Pb (1 mm) Exploited  decay topologies The overall efficiency, including BR, is about 5%

Expected background Anti-neutrino charm production Decay in flight of mesons and punch-through Events from      + from the muonic decay of the  + from the leading oscillation channel anti-   anti- . Suppressed by good charge measurement. Other backgrounds: charm-production,  wrongly matched to hadrons, ass. charm-production, large angle  scattering

Signal & background vs E 732 km 3000 km signal charm decay in flight and punch-through ++

Scanning load How big the ECC detector could be? It depends on the scanning load. Notice that a large fraction of the scanning load is given by events that can be easily rejected afterwards with the ECC analysis Boundary conditions: Detector located 732 km from the source 5 years data taking Four classes of events to be scanned (1 kton): Silver Signal: ~ 30 events and Golden Signal: ~ 300 events Background from oscillations: ~ 2000 events Charm background: ~ 100 events Decay of h and punch-through: ~ 5000 events Overall less than 1x10 4 events in 5 years have to be scanned A 5kton ECC detector is feasible from the scanning point of view

! Warning ! 1 kton ECC  1 kton Pb Not 1 kton ECC  1 kton Emulsions

Signal uncertainty Golden channel Flux and cross-section are very well known Mainly related to the detection efficiency and to the calibration We assume 10%, taking into account the coarse granularity of the detector and the absence of an in situ calibration beam line (i.e. like in NuTeV) Silver channel Main sources are the knowledge of the emulsion scanning efficiency and the cross-section ratio     We assume 15% according to the calculation quoted in the OPERA Proposal Note that the silver channel is dominated by statistics. NB The signal uncertainty has a larger impact on the golden than on the silver channel (see in the following)

Background uncertainty (I) The background sources are common to both channels Anti-neutrino charm-production Very few data available  40% at present However a near detector at a NuFact will collect about 1 million charm events  We assume (conservatively) 10% error

Background uncertainty (II) Events from     For this background we assume the same uncertainty as for the silver signal: 15% Decay in flight and punch-through This background is dominated by the poor knowledge of the hadronization and of the hadronic re-interaction processes. We conservatively assume a 50% uncertainty NB The background uncertainty has a larger impact on the silver than on the golden channel

Statistical approach Since signal and background have a different energy dependence, the number of expected events is divided into several bins The number of bins has been optimized by maximizing the sensitivity of the experiment For each point (  13,  ) and each bin we extract several sets of data according to the signal and background Poisson distributions with mean values corresponding to the expected ones. Furthermore, the mean Poisson values are smeared assuming a gaussian width corresponding to the quoted uncertainties In total we have 14 bins (4 ECC-bins and 5 iron-bins for each polarity) They are combined using the frequentist approach A maximum likelihood fit is performed to the simulated data to extract the signal content

Expected signal and background Number of background and golden muon signal events expected in the iron detector at L = 3000 km including our estimate of the systematic effects. The binning definition refers to the reconstructed energy. Number of background and ”silver” signal events expected in the OPERA-like detector at 732 km and 3000 km. The binning definition refers to the reconstructed energy. The reconstructed energy bins have been obtained by smearing the true neutrino energy with a resolution of 20% NB the smearing strongly affects the S/N ratio of the bins in the silver channel  worse sensitivity and  13 = 1° dominated by statistical fluctuations of the background

Combining 732km and 3000km No clone regions for  13 >1°, for  13 =1° they show-up in less than 10% of the exps ECC + IronIron 3000 km Allowed regions on the parameter extracted from the analysis of the simulated data for  13 = 1°,  = 90°. The best fit corresponds to  13 = 0.9°,  = 80°.

Comments Since the expected backgrounds and the silver signal are small (<10 evts) in all bins, the systematic uncertainties on this component have a minor effect on the overall sensitivity However, this is not the case for the golden signal What about the use of dE/dx to reduce the back. in the ECC? Effect of the systematic uncertainty on the signal in the iron detector for the 99% CL intervals obtained for  13 =5°,  =0°, when combining the 5 Kton ECC detector at L = 732 km and the 40 Kton MID at L = The best fit for the default configuration (i.e., with a 10% systematic uncertainty on the golden muon signal at the MID) corresponds to  13 =5.1°,  =2°. Effect of a background reduction by a factor of two for the OPERA-like detector at 732 km and the simulated point  13 =5°,  =90°. The best fit for the standard configuration corresponds to  13 =5.0°,  =86°.

Combining ECC and 3000km A location of the OPERA-like detector at 3000 km would reduce considerably most background contributions (proportional to 1/L 2 ), with the exclusion of the muonic decay of  + events from anti-   anti-  oscillations. Furthermore, the scanning load reduces by about a factor 20

Comments For  13 =1° a clone region appears at 90% C.L.!!! This is mainly due to the tiny signal contribution from the silver channel (close to the the intrinsic limit of the experimental sensitivity) By increasing the number of simulated experiments for  13 =1° and  =90° with 732 km, in few cases ( less than 10%) we observe the presence of clone regions We were just unlucky in simulating the configuration with both detectors at 3000 km In the real life there is a few percent risk that for  13 =1° clone regions show-up

Possible improvements The lighter scanning load (only few thousand events) would allow some potential improvements Remove more than one brick per event. This would allow an increase of the signal detection efficiency by about 20% (mainly due to an increase in the brick finding efficiency) Inclusion of the anti- e  anti-  channel when the beam is run with opposite polarity. The background is larger by about a factor of two with respect to the e   channel, while the signal is reduced by the same amount due to the anti-neutrino cross-section. This channel would slightly increase (about 10%) the sensitivity of the measurement in spite of the worse S/B ratio One could also, but this is at the present quite ambitious, increase the mass of the detector and/or the exposure

Conclusion e   is the Golden Channel at the Neutrino Factory A good signal with high statistics and low background e   is the Silver Channel at the Neutrino Factory A difficult-to-signal with low statistics; however, it can be crucial to solve the (  13,  ) ambiguity An ECC based detector is mandatory to exploit this channel With the presently available know-how it is possible to think about a 5 Kton ECC detector for the Neutrino Factory A realistic estimate of the systematic error both on the signal and on the background is mandatory in order to have a reliable evaluation of the golden and silver channels synergy For  13 =1° we are at the limit of the experimental sensitivity  sensitive to the large statistical fluctuations on both signal and background To explore values of  13 smaller than 1 and solve the other ambiguities, one should consider also other channels (see A.Donini talk)