BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/1 Hadron Production Cross-Sections and Secondary Particle Yields from 2 to 15 GeV using Neutrino Beam Targets The HARP Experiment –Physics goals and motivations –Summary of the experimental program –Detector overview and performance The fist physics analysis: pion yields from K2K target –Goals –Results Alessandra Tonazzo, Università Roma Tre and INFN On behalf of the HARP Collaboration

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/2 HARP physics goals Precise (~2-3% error) measurement of d 2  /dp T dp L for secondary HAdRon Production by incident p and  ± with –Beam momentum from 1.5 to 15 GeV/c –Large range of target materials, from Hydrogen to Lead ►Acceptance over the full solid angle ►Final state particle identification

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/3 Input for prediction of neutrino fluxes for the MiniBooNE and K2K experiments Pion/Kaon yield for the design of the proton driver and target system of Neutrino Factories and SPL- based Super-Beams Input for precise calculation of the atmospheric neutrino flux (from yields of secondary ,K) Input for Monte Carlo generators (GEANT4, e.g. for LHC or space applications) HARP Physics Motivations

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/4 Data taking summary K2K: AlMiniBoone: BeLSND: H 2 O 5% 50% 100% Replica 5% 50% 100% Replica 10% 100% GeV/c+8.9 GeV/c+1.5 GeV/c SOLID: CRYOGENIC: EXP: HARP took data at the CERN PS T9 beamline in Total: 420 M events, ~300 settings

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/5 The HARP Experiment 124 physicists24 institutes

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/6 The HARP Experiment

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/7 The HARP detector layout TRACKING + PARTICLE ID at Large angle and Forward

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/8 Beam detectors Beam tracking with MWPCs : –96% tracking efficiency using 3 planes out of 4 –Resolution <100  m MiniBoone target TOF-A CKOV-A CKOV-B TOF-B 21.4 m T9 beam MWPCs

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/9 Beam particle selection Beam TOF: –separate  /K/p at low energy over 21m flight distance –time resolution 170 ps after TDC and ADC equalization –proton selection purity >98.7% Beam Cherenkov: –Identify electrons at low energy,  at high energy, K above 12 GeV –~100% eff. in e-  tagging 3 GeV/c beam  K 12.9 GeV/c beam  K p d

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/10  Status of TPC  p T /p T p dE/dx Elastic scattering of 3 GeV p and  on H 2 target Missing mass m x 2 = ( p beam + p target – p TPC ) Red: using dE/dx for PID pp  pp pppp

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/11 Analysis for K2K: motivations Computation of fluxes at SK is based on near/far ratio R –For pointlike source (no oscillations), R~1/r 2 –If the near detector does not see a pointlike source, R depends on E Current K2K computation of R is based on MC –Confirmed by  production measurement (pion monitor) at E >1 GeV –Extrapolated to E <1 GeV : the interesting region for oscillations ! beam 250km

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/12 Analysis for K2K: motivations

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/13 Analysis for K2K: interesting region K2K needs measurement of pions with  E>1 GeV  E<4.5 GeV   <300 mrad ►Forward region Main tracking detectors: drift chambers PID detectors: TOF, Cherenkov, calorimeter Dip from neutrino oscillations in K2K they come from decay of these pions:

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/14 P > 1 GeV K2K interest K2K interest Forward acceptance dipole NDC1 NDC2 B x z A particle is accepted if it reaches the second module of the drift chambers

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/15 Forward Tracking: NDC Reused NOMAD Drift Chambers 12 planes per chamber, wires at 0°,±5° w.r.t. vertical Hit efficiency ~80% (limited by non-flammable gas mixture) –stable in time –lowered by high particle flux –recovered between spills –correctly reproduced in the simulation Alignment with cosmics and beam muons  drift distance resolution ~340  m Plane efficiencies Side modules

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/16 Forward tracking dipole magnet NDC1 NDC2 B x z NDC5 beam target Top view 1 2 NDC3 NDC4 Plane segment 3 3 track types depending on the nature of the matching object upstream the dipole  Track-Track  Track-Plane segment  Track-Target/vertex Aim: recover as much efficiency as possible and avoid dependencies on track density in 1st NDC module (hadron model dependent)

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/17 The momentum and angular resolutions are well within the K2K requirements MC data 1 type No vertex constraintincluded MC momentum resolution angular resolution Forward tracking: resolution

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/18 Only particles with no track downstream the dipole do not enter into the game (~2%). –This 2% should be quite independent of the track density (hadron model dependent) because tracks downstream the dipole are uncorrelated Downstream tracking efficiency ~98% Up-downstream matching efficiency ~75% We have been working to improve it by: 1.Alignment 2.Optimizing matching cuts Tracking efficiency  track is known at the level of 5%

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/19 Total tracking efficiency as a function of p(left) and  x (right) computed using MC (2 hadron generators) properly scaled by data Total tracking efficiency Green: type 1 Blue: type 2 Red: type 3 Black: sum of normalized efficiency for each type

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/20 Forward PID: Cherenkov Separate  /p at large momenta 31 m 3 filled with C 4 F 10 (n=1.0014) Light collection: mirrors+Winston cones → 38 PMTs in 2 rows LED flashing system for calibration  mass is a free parameter nominal threshold N pe → 21 p (GeV/c) Number of photoelectrons e+e+ ++ 3 GeV beam particles p p ++ 5 GeV beam particles N phel data

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/21 Forward PID: TOF Wall Calibration / equalization –Cosmic ray runs (every 2-3 months) –Laser (continuous: monitor stability) TOF time resolution ~160 ps =>3  separation of  /p (K/  ) up to 4.5 (2.4) GeV/c =>7  separation of  /p at 3 GeV/c Separate  /p (K/  ) at low momenta 42 slabs of fast scintillator read at both ends by PMTs 3 GeV beam particles data  p

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/22 Forward PID: Calorimeter Pb/fibre: 4/1 –EM1: 62 modules, 4 cm thick –EM2: 80 modules, 8 cm thick Total 16 X 0 Reused from CHORUS Calibration with cosmic rays: –Measurement of attenuation length in fibers –Module equalization Energy resolution 23%/sqrt(E) intrinsic resolution 15%/sqrt(E) convoluted with beam spread at detector entrance electrons pions 3 GeV data

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/23 tof cerenkov calorimeter momentum distribution Using the Bayes theorem: 1.5 GeV 3 GeV 5 GeV data we use the beam detectors to establish the “true” nature of the particle Forward PID:  efficiency and purity Iterative approach: dependence on the prior removed after few iterations  j  -(t) = N j  -true-obs / N j  -true  j  -(t) = N j  -true-obs / N j  -obs  j  -(t) /  j  -(t)

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/24 pion efficiency The 3 types of tracks must be treated separately because of the different momentum resolution pion purity pion yield tracking efficiency migration matrix (not computed yet) acceptance depend on momentum resolution i = bin of true (p,  ) j = bin of recosntructed (p,  ) The cross section

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/25 To be decoupled from absorption and reinteraction effects we have used a thin target p-e/  misidentification background K2K replica target 5% Al target 200% Al target Raw data p > 0.2 GeV/c |  y | < 50 mrad 25 < |  x | < 200 mrad Pion yield

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/26 5% Al target p > 0.2 GeV/c |  y | < 50 mrad 25 < |  x | < 200 mrad Pion yield After PID correction Afterefficiencycorrection

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/27 Pion yield Afteracceptancecorrection 5% Al target p > 0.2 GeV/c |  y | < 50 mrad 25 < |  x | < 200 mrad Systematics are still to be evaluated: tracking efficiency know to 5% expect small effect from PID

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/28 Conclusions The HARP Experiment has collected data for hadron production measurements with a wide range of beam energies and targets Status of detector –Forward region: good tracking and PID –Large angle: much recent progress First physics results are available: thin (5% ) K2K target –Using forward region of the detector To do –Compute data deconvolution and migration matrix –Evaluate systematic errors –Analyse empty target data for background subtraction –Investigate  =0 region: saturation effects due to beam particle removal –Introduce normalization for absolute cross-section (using min.bias trigger) –… go on to full statistics, and to the rest of the data !

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/29 Backup

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/30 The efficiency of NDC2 and NDC5 is flat within ~5%. The efficiency of the lateral modules (3 and 4) is flat within 10% The combined efficiency is not sensitive to these variations. NDC 2 NDC 5 NDC 4 NDC 3 data NDC module efficiency NDC2 NDC5 NDC3 NDC4 dipole

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/31 NDC2 NDC5 NDC3 NDC4 MC dipole moduleacceptancemoduleefficiency Module acceptance / Downstream eff.

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/32 data MC x 2,  x2 Up-downstream matching efficiency We need to convert x 2 and  x2 to p and . For that we use the MC

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/33 Probability of matching a downstream track with the other side of the dipole dipole magnet NDC1 NDC2 B x z NDC5 beam target Top view 1 NDC3 NDC4 2D segment 2 3 MC and data agree within ~3% in their shapes We tune to the DATA the absolute scale of each track type MCdata + Up-down matching efficiency

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/34 CalorimeterCkov TOF The different pid detectors are independent of each other Using the Bayes theorem Measured quantities Particle types Forward PID: continuous probability

BEACH04 Chicago, June 28-July A.Tonazzo - HARP pion yields for neutrino beams/35 P( p, |  ) from TOF = m 2 / p 2 with P( p,N phe |  ) from Cherenkov P( p, E 1, E 2 ) from Calorimeter Line : true PID Colored histograms: reconstructed PID 3 GeV  p eK Step 1Step 2 Step 3 Step 4 Step 5 Product of conditional probabilities and relative abundances Iterative Bayesian approach: dependence on the prior removed after few iterations Forward PID: iterative procedure