1. Measurement of the neutron detection efficiency of the KLOE calorimeter P.Gauzzi (Universita’ La Sapienza e INFN – Roma) for the KLONE Group 10 th ICATPP Conference October 8-12, 2007 – Como M.Anelli, G.Battistoni, S.Bertolucci, C.Bini, P.Branchini, C.Curceanu, G.De Zorzi, A.Di Domenico, B.Di Micco, A.Ferrari, S.Fiore, P.G., S.Giovannella,F.Happacher, M.Iliescu, M.Martini, S.Miscetti, F.Nguyen, A.Passeri, A.Prokofiev, P.Sala, B.Sciascia, F.Sirghi

2. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 2 The KLOE calorimeter Pb - scintillating fiber sampling calorimeter of the KLOE experiment at DA  NE (LNF): 1 mm diameter sci.-fi. (Kuraray SCSF-81 and Pol.Hi.Tech 0046) –Core: polystyrene,  =1.050 g/cm 3, n=1.6, peak ~ 460 nm 0.5 mm groved lead foils Lead:Fiber:Glue volume ratio = 42:48:10 X 0 = 1.6 cm  =5.3 g/cm 3 Calorimeter thickness = 23 cm Total scintillator thickness ~ 10 cm 1.2 mm 1.35 mm1.0 mm Lead

3. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 3 The KLOE calorimeter Operated from 1999 to 2006 with good performance and high efficiency for electron and photon detection, and also good capability of π/  /e separation Energy resolution:  E /E=5.7%/  E(GeV)  t =54 ps/  E(GeV)  147 ps Time resolution (see KLOE Collaboration, NIM A482 (2002),364) (  K S K L ; K S  π + π ─ ; K L  2π 0 ) 44

4. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 4 Why neutrons at KLOE ? Detection of n of few to few hundreds of MeV is traditionally performed with organic scintillators (elastic scattering of n on H atoms produces protons detected by the scintillator itself)  efficiency scales with thickness  1%/cm Use of high-Z material improves the neutron efficiency (see C.Birattari et al., NIM A297 (1990) and NIM A338 (1994) and also T.Baumann et al., NIM B192 (2002)) Preliminary estimate with KLOE data (n produced by K  interactions in the apparatus) showed a high efficiency (  40%) for neutrons with E n < 20 MeV, confirmed by the KLOE MC n detection is relevant for the DA  NE-2 program at LNF; two proposals: – search for deeply bounded kaonic nuclei (AMADEUS) – measurement of the neutron time-like form factors (DANTE) A test has been performed with the neutron beam of the The Svedberg Laboratory of Uppsala (October 2006 and June 2007)

5. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 5 The neutron beam @ TSL E KIN (MeV) 5.31 m KLOE calorimeter module (  2 cm) Neutrons produced in the reaction 7 Li(p,n) 7 Be Proton beam energy from 180 MeV to  20 MeV Neutron energy spectrum peaked at max energy (at 180 MeV  f p =42% of n in the peak) Tail down to termal neutrons

6. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 6 Experimental setup 1. Old KLOE prototype: total length  60 cm 3  5 cells (4.2 cm  4.2 cm) read out at both ends by Hamamatsu/Burle PMTs 2. Beam Position Monitor: array of 7 scintillating counters 1 cm thickness 3.Reference counter : NE110; 5 cm thick; 10  20 cm 2 area (in June 2007  two other NE110 counters 2.5 cm thick) All mounted on a rotating frame allowing for vertical (data taking with n beam) and horizontal (for calibration with cosmic rays) positions (1) (3) (2)

7. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 7 Trigger & DAQ Trigger: No beam extraction signal available Scintillator trigger: Side 1 – Side 2 overlap coincidence Calorimeter trigger: analog sum of the signals of the first 12 cells (4 planes out of 5)   A  B overlap coincidence Trigger signal is phase locked with the RF signal (45 - 54 ns) DAQ: Simplified version of the KLOE experiment DAQ system (VME standard) Max DAQ rate : 1.7 kHz - Typical run: 10 6 events For each configuration/energy: scans with different trigger thresholds Three data-sets: –E peak = 180 MeV -- October 2006 - two weeks –E peak = 46.5 MeV -- June 2007 –E peak = 21.8 MeV -- “ 4 days n Y X Z

8. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 8 Efficiency measurement Global efficiency measurement: integrated over all the energy spectrum f live = fraction of DAQ live time  = acceptance (assuming the beam fully contained in the calorimeter surface    1) Sizeable beam halo at low peak energy  Rate(trigger) must be corrected E n = 180 MeV f live

9. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 9 Neutron rate Absolute flux of neutrons measured after the collimator; 2 monitors of beam intensity (see A.Prokofiev et al., PoS (FNDA2006) 016) : –Ionization Chamber Monitor (7 cm  ): online monitor, not position sensitive –Thin-Film Breakdown Counter (1 cm  ): offline monitor; used to calibrate the ICM by measuring the neutron flux at the collimator exit Rate(n) = Rate(ICM)  K  πr 2 / f p r = collimator radius (1 cm) K = calibration factor (TFBC to ICM) f p = fraction of neutrons in the peak  accuracy: 10% at higher peak energy (180 MeV) 20% at lower peak energy (20 – 50 MeV)

10. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 10 Scintillator calibration Trigger threshold calibration in MeV eq.el.en.: ADC counts  source to set the energy scale in MeV: 90 Sr  ─ endpoint = 0.56 MeV 90 Y  ─ endpoint = 2.28 MeV 25 keV/ADC count ADC counts Thr. [mV] ADC counts  6 counts/mV Events

11. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 11  (%) E peak = 180 MeV Scintillator efficiency Agrees with the “thumb rule” (1%/cm) at thresholds above 2.5 MeV el.eq.en. Check of the method and of the beam monitor accuracy EnEn  (%)/ cm of scintillator Agrees with previous measurements in the same energy range after rescaling for the thickness

12. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 12 Calorimeter calibration Cell response equalized: MIP peak at  550 ADC counts Trigger threshold calibration ADC counts Thr. [mV] Energy scale calibration with the MIP/MeV conversion factor from KLOE ( 1 MIP in one calorimeter cell  35 MeV eq. en.) ( see KLOE Collaboration, NIM A354 (1995),352 )  6 counts/mV Events

13. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 13  (%) Comparison with our scintillator normalized to the same active material thickness Calorimeter efficiency Stable for different run conditions Very high efficiency w.r.t. the naive expectation ( ~ 10% @ 2 MeV thr.) E peak = 180 MeV

14. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 14 Energy spectrum from TOF n Rephasing is needed, since the trigger is phase locked with the RF (45 ns period) From TOF   spectrum of the neutrons Assuming the neutron mass  kinetic energy spectrum Energy spectrum can be reconstructed from TOF

15. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 15 Efficiency vs energy Fast MC to test the sensitivity of the time distribution to the shape of the efficiency curve Better agreement with an efficiency decreasing with energy Some discrepancy in the low energy part of the spectrum T cell (ns) MC ● Data E kin (MeV) ε(%) T cell (ns) MC ● Data E kin (MeV) ε(%)

16. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 16  (%)/ cm of scintillator  (%) - scint. In June 2007 we took data at lower peak energies: 21.8 and 46.5 MeV Large errors: big uncertainty in the beam halo evaluation worse accuracy of the beam monitors Correction factor for beam halo  0.9  0.1 Low energy data

17. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 17 The ratio is almost independent from the halo (Ratio normalized to the same active material thickness)  (%) - calorimeter Calorimeter efficiency Very high efficiency at low threshold Agreement with the high energy measurements Correct. factor for beam halo  0.8  0.1

18. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 18 MC simulation A detailed simulation of the calorimeter structure and of the beamline (source, collimator and concrete shielding) has been carried out with the FLUKA Code Neutron fluence along the beamline (n/cm 2 /29000n) collimator concrete shielding KLOECAL 7 Li target GLUE FIBERS base module replicas 200 layers LEAD

19. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 19 Preliminary results of MC n Some discrepancy in the low energy part of the spectrum No cut in released energy No trigger simulation  Upper limit on   (%) E n (MeV) Data - Low thresh.

20. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 20 Conclusions The first measurement of the detection efficiency for neutrons of 20 - 180 MeV of a high sampling Pb-sci.fi. calorimeter has been performed at the The Svedberg Laboratory in Uppsala. Measurement of the n efficiency of a NE110 scintillator agrees with published results in the same energy range. The calorimeter efficiency, integrated over the whole neutron energy spectrum, ranges between 30-50 % at the lowest trigger threshold. Study of the efficiency as a function of n energy is in progress. Full simulation with FLUKA is in progress. Further test foreseen for beginning 2008 at Louvain-la-Neuve (E n = 10 – 70 MeV; larger interbunch time)

21. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 21 Spares

22. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 22 Time structure 4.2 ms 2.4 ms 40 ns 41 ns  5 ns FWHM

23. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 23 Preliminary results of MC Preliminary results of MC Simulated neutron beam: E kin = 180 MeV Each primary neutron has a high probability to have elastic/inelastic scattering in Pb In average 5.4 secondaries per primary neutron are generated, counting only neutrons above 19.6 MeV. target P el (%) P inel (%) Pb 32.6 31.4 fibers 10.4 7.0 glue 2.3 2.2 neutrons above 19.6 MeV 62.2% photons 26.9% protons6.8% He-43.2% deuteron 0.4% triton 0.2% He-3 0.2% neutrons 94.2% protons 4.7% photons 1.1% Secondaries created in interactions of low energy neutrons (below 19.6 MeV) are - in average - 97.7 particles per primary neutron.

24. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 24 Preliminary results of MC The enhancement of the efficiency appears to be due to the large inelastic production of neutrons in Pb. These secondary neutrons: - are produced isotropically; - are associated with a non negligible fraction of e.m. energy and of protons, which can be detected in the nearby fibers; - have low energy and then have a large probability to do new interactions in the calorimeter with neutron/proton/γ production. n Z(cm) p n1n1 n2n2 n3n3 n4n4 X(cm)

25. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 25 Beam halo TSL beam experts measured a sizeable beam halo at low peak energy (21.8 and 46.5 MeV): –TFBC scan of the area near the collimator  integrated flux over the ICM area  5% of the “core” flux (with large uncertainty)  halo shape also measured Confirmed by our background counters Our calorimeter is larger than the projection of ICM area By integrating over the calorimeter we obtain an estimate of the halo contribution to the trigger rate of (20  10)% Only 10% on the reference scintillator due to the smaller area

26. P.Gauzzi10 th ICATPP Conference - October 9 th 2007 – Como 26  (%) KLOE vs others Comparison with other calorimeters: –KLOE: 23 cm thick –Crystall Ball: NaI 40.7 cm thick (NIMA462(2001),463) –GRAAL: BGO 24 cm thick (NIMA562(2006),85)