Sensitivity of neutrino mass measurements with microcalorimeters

Transcription

1 Sensitivity of neutrino mass measurements with microcalorimeters A. Nucciotti, Nucciotti E. Ferri, O.Cremonesi Dipartimento di Fisica G. Occhialini,Università di Milano-Bicocca INFN - Sezione di Milano-Bicocca direct neutrino mass measurement introduction spectrometers vs. calorimeters calorimeters calorimeter statistical sensitivity analytic approach montercarlo approach calorimeter systematics arxiv: v1 INT Workshop - The Future of Neutrino Mass Measurements: Terrestrial, Astrophysical, and Cosmological Measurements in the Next Decade Seattle WA, USA, February 8-11, 2010

2 Experimental approaches for direct measurements Spectrometers: source detector 3 e H source counter analyzer differential or integral spectrometer: s from the 3H spectrum E are magnetically and/or electrostatically selected and transported to the counter Calorimeters: source detector e source excitation energies e calorimeter ideally measures all the energy E released in the decay except for the e energy: E=E0 E 2

3 Calorimetry of beta sources calorimeters measure the entire spectrum at once use low E0 decaying isotopes to achieve enough statistics near the end-point best choice 187Re: Re E0 = 2.47 kev F( E=10 ev)~( E/E0)3= m m = 0 m = 20 ev Calorimetry advantages E = 0 fpile-up= 0 no backscattering no energy losses in the source no atomic/molecular final state effects no solid state excitation Calorimetry drawbacks limited statistics systematics due to pile-up other systematics... E = 30 ev fpile-up= Pile-up time unresolved superposition of decays for a source activity A, a time resolution R and an energy resolution function R(E) N exp(e) (N(E) RA N(E) N(E)) R(E) pile-up fraction: fpile-up = R A E F E E0 3 3

4 187 Re calorimetric experiment statistical sensitivity / 1 resolving time R analysis interval E source activity A number of detectors Ndet pile-up fraction fpile-up= RA experimental exposure tm=t Ndet N (E,m ) N E,m 3 2 E 0 E 3 E0 2 1 m 2 E 0 E E0 F E m =A N det N E,m de E 0 E fpile-upn (E,0) N (E,0) F E 0 A N det pile-up E 3 3 E0 F E m F E 0 1 E 3m 2 2 E 2 E0 N (E,m =15 ev) N (E,m =0) pp F E R A2 N det E 0 E R A N det signal = N (E,m =0) N (E,m =15 ev) N E,0 N E,0 de E E0 F E m F E 0 t M signal = =1.7 for 90% C.L. pp background F E 0 t M F E t M 4

5 Calorimetric experiment statistical sensitivity / 2 2 signal F E m F E 0 t M = = t M pp bkg F E 0 t M F E t M 90 m 1.13 E0 4 N ev [ A N det 2 4 =1.7 for 90% C.L. E E A N R det E0 E 30 E 3 f E 0 10 pile up E 0.89 E E FWHM 3 E0 E f pile up= R A 2 E0 90 m 3 E 3m A N det 3 E 0 2 E 2 ] Optimal energy interval E 1/ 4 E = max 0.55E 0 R A, E FWHM pile-up is negligible 3 E 0 E A t M experimental challenges energy resolution EFWHM time resolution R exposure tm = Ndet T single channel activity A 5

6 thermometer T V T = E/C electro-thermal link G temperature Cryogenic detectors as calorimeters particle absorber E T time complete energy thermalization = C/G 1 mg of 100 mk C ~ T 3 (Debye) C ~ J/K Erms ~ 1 ev 6 kev x-ray T ~ 10 mk G ~ W/K = C / G ~ 10 ms (ionization, excitation heat) calorimetry T = E/C with C total thermal capacity (phonons, electrons, spins...) phonons: C ~ T 3 (Debye law) in dielectrics or superconductors below Tc low T (i.e. T 1K) 2 1/2 Erms = (kb T C) due statistical fluctuations of internal energy E T(t) = E/C e-t/ with = C/G and G thermal conductance 6

7 Montecarlo simulations: statistical sensitivity generate many ( ) simulated experiments calculate total spectrum S(E) = (Nev (N (E,0) + fppn (E,0) N (E 0)) + b(e)) R(E) Nev total statistics N (E,0) normalized 187Re spectrum for m = 0 fpp fraction of unresolved pile-up events b(e) background (usually constant) R(E) detector energy response function (usually gaussian) generate spectra introducing Poisson fluctuations in S(E) fit the spectra with standard technique (m 2, E, Nev, fpp, b(e) free) 0 obtain 90% C.L. m sensitivity 90(m ) from (1.7 ) of m 2 distribution Montecarlo input parameters vs. real experiment parameters Nev = Ndet tm A fpp R A ( R rise) 7

8 Sub-eV m statistical sensitivity Montecarlo analytic 1st order analytic 2nd order 8

9 Sub-eV m statistical sensitivity / 2 Nev= 1014 Montecarlo analytic 1st order analytic 2nd order fpp = fpp = 10-3 fpp = 10-4 fpp = 10-5 fpp =

10 Sub-eV m statistical sensitivity / 3 MC formula events tm = year detector 10

11 Effect of background on statistical sensitivity 90 m 1.13 E0 4 N ev [ E0 E0 E 3 f pp b 10 E 0 E A Optimal energy interval E E = max E 0 ] 1/ 4 b bkg counts/kev/s/det E0 3 f pp b, E FWHM 10 A Mibeta S/B S/B=Nev/Nbkg Nbkg=bE0T 11

12 Statistical sensitivity summary exposure required for 0.2 ev m sensitivity A R Nev E exposure [Hz] [ s] [ev] b=0 [counts] [det year] 8 arrays 5000 pixels 10 years 400 g Rhenium exposure required for 0.1 ev m sensitivity A R Nev E exposure [Hz] [ s] [ev] [counts] [det year] arrays pixels 10 years 3.2 kg Rhenium 12

13 Montecarlo simulations: analysis of systematics Assessing systematic uncertainties with Montecarlo simulations effects due to incomplete/incorrect data modeling generate simulated experimental spectra with systematic effect analyze spectra without effect obtain (m ) and m 2 as function of effect size 90 uncertainty due to experimental parameter finite accuracy generate simulated experimental spectra with randomly fluctuated parameter analyze spectra with fixed average parameter obtain (m ) and m 2 as function of uncertainty magnitude 90 systematic uncertainties analyzed for Nev=1014, EFWHM=1.5 ev and fpp=10-6 two main classes of systematics source related systematic effects instrumental systematic uncertainties 13

14 Source related systematic uncertainties: summary electron surface escape investigation with MC methods N'(E ) = N(E ) (1 a E/E0) esc for 1mg Re crystal a esc 187 Re decay spectral shape improve theoretical description of electron spectrum N'(E ) = N(E ) (1 + a E + a E 2) 1 2 from Dvornicky-Simkovic (Medex09) f(e)= E E E condensed matter effects: BEFS observed in Re and AgReO4: improve modeling and parametrization pile-up spectrum spectral shape energy dependent rejection efficiency: investigation with MC methods eff = f (, A /A ) N' (E ) = N (E ) f (E,f ) R R 1 2 pp pp corr pp source of uncertainty electron surface escape correction to quadratic spectral shape correction to pile-up spectral shape maximum quantity effect for describing the m 2<0.01 ev2 effect aesc 10-5 a1 (a2=0) 10-9 ev-1 a2 (a1=0) ev-2 fpp

15 Electron escape systematic uncertainties N' (E )=N(E )fesc(e ) fesc(e )= 1-2x10-5 E/E0 Geant4 MC simulation for 1mg Re crystal E = 1.5 ev; fpp = 10-6; Nev = 1014 systematic effect with escape neglected 15

16 Beta spectrum shape systematic uncertainties N'(E ) = N(E ) (1 + a1 E + a2 E 2) a2 m 2 on z axis a1 E = 1.5 ev; fpp = 10-6; Nev =

17 BEFS: Re vs. AgReO4 BEFS: Beta Environmental Fine Structure Modulation of the electron emission probability due to the atomic and molecular surrounding of the decaying nucleus: it is explained by the wave structure of the electron (analogous of EXAFS) 17

18 Systematics from BEFS Nev = 1010 E = 5 ev fpp= 10-5 expected end-point spectral deformation systematic shift with BEFS neglected 18

19 Pile-up spectrum systematic uncertainties / 1 t / A t =A e decay e t / rise t = 15 ms t = 10 ms t = 5 ms t = 3 ms Example 2 pulses with: rise= 1.5 ms decay= 10 ms A2/A1 = 0.5 Montecarlo with simulated pulses rise 2 ms = 10 samples resolving time eff = ( rise, A2/A1 ) source of systematics new MC tools and new algorithms A2/A1 F. Fontanelli et al., NIM A 421 (1999) 464 t 19

20 missed Pile-up spectrum systematic uncertainties / 2 energy dependent pile-up resolving time eff f corr E,f pp =10 1 e E E 0 / 480 ev 1 constant pile-up resolving time R MC simulation 20

21 Pile-up spectrum systematic uncertainties / 3 E = 1.5 ev; fpp = 10-6; Nev =

22 Systematics from instrumental uncertainties: summary error on energy resolution E quantity describing the uncertainty err( E)/ E maximum uncertainty for m 2<0.01 ev tail in response function ( =0.2eV-1) Atail 10-4 error on single pixel energy calibration K spread in energy resolution E in the array (K)/K spread( E)/ E hidden constant background Nev/Nbkg 108 background linear deviation (bt=105c/ev) b1 0.1 source of uncertainty

23 Instrumental uncertainties: large arrays E = 1.5 ev; fpp = 10-6; Nev = 1014 error on single pixel energy calibration error on global response function width spread in single pixel response function width spectrum calibration accuracy improves with statistics technological issue 23

24 Detector response function Mn K Mn K Cr K Ti K Ca K Ti K Ca K K, Au, Sn, Pb Cl K Cl K Al K Al K 2168 hours mg with fluorescence source open calibration gives the energy scale and the response function Pb M 1 M Pb M X-ray peaks have tails on low energy side 1~6 kev X-rays in AgReO4 have an attenuation length < 2 μm are the response functions for X-rays and for s from 187Re decay the same? need for a good phenomenological description of the X-ray peak shape 24

25 Instrumental uncertainties: response function tail = 1 ev = 0.1 ev-1 [ ][ T E =A tail exp E E erf E E ] E = 1.5 ev; fpp = 10-6; Nev = 1014 calibration statistics 25

26 Background (MIBETA) Pb-Re L-L escape 187Re pile-up? unshielded 55 Fe calibration source 55Fe Inner-Bremsstrahlung (QIB = 232 kev, AIB = 12 kbq) causes too high background fluorescence from surroundings Re X-ray escape peaks continuum lead shielded 55 Fe calibration source remaining background to be understood and reduced cosmic rays environmental radioactivity the hidden background is a source of systematic uncertainties Go underground? 26

27 Instrumental uncertainties: constant background E = 1.5 ev; fpp = 10-6; Nev = 1014 unknown constant background below the spectrum S/B=4x108 S/B=Nev/Nbkg Milano experiment S/B 3x104 bt 106 c/ev on this plot 27

28 Instrumental uncertainties: background linear term E = 1.5 ev; fpp = 10-6; Nev = 1014 B E =bt 1 b1 E0 2E 0 E 2E0 28

29 Conclusions thermal calorimetry of 187Re decay can give sub-ev sensitivity on m preliminary systematics estimate by Montecarlo methods shows source related systematics require more investigations Beta Environmental Fine Structure 187 Re spectral shape... instrumental systematic uncertainties seem to be more controllable more systematics will show up increasing the statistics in the spectra intermediate size experiments are crucial 29

30 Backups... 30

31 MIBETA: Measurement of response function (2004) external X-rays probe only detector surface escape peaks allow internal calibration (6 kev) 3 m in AgReO4 (70 kev) 400 m escape peaks are broad because of natural widths of atomic transitions Mn K 1 + K 2 Re 71.7 kev E > 71.7 kev internal K 1=2.5eV E =36eV calibration with 44Ti Re X 44 Ti - Re K 1 esc K 1=47eV E =75eV no tail! e kev -X escape peaks have only Re K natural width ( ReK~47 ev) the response function is a possible source of systematic uncertainties in calorimetric neutrino mass experiments 31

32 MIBETA: BEFS analysis (2005) BEFS: Beta Environmental Fine Structure Modulation of the electron emission probability due to the atomic and molecular surrounding of the decaying nucleus: it is explained by the wave structure of the electron (analogous of EXAFS) BEFS experimental evidence in 187Re decay in AgReO4 less pronounced than in metallic rhenium fit =0 =1 BEFS k e =F s lexafs F p lexafs lexafs k e = = 1 l N 2k e2 2n B nl k e,rn e n=1 sin 2k e R n 0l nl Fp = 0.84 ± 0.30 BEFS is a possible source of systematic uncertainties in 187Re neutrino mass experiments AgReO4 C. Arnaboldi et al., Phys. Rev. Lett. 96 (2006) EXAFS measurements better models 32

33 MC analysis of systematics: BEFS fit MIBETA Nev = 1010 E = 20 ev fpp=10-4 experiment with AgReO4 Nev = spectrum without BEFS spectrum with BEFS rhenium E = 1 ev fpp=

34 MIBETA experiment: 2002/03 heat sink G: Al bonding wires 17µm thermistor epoxy glue joint AgReO4 crystal X-ray calibration source Si-implanted thermistors (ITC-irst) AgReO single crystals 4 187Re 0.6 years live time (0.45 years only ) Re decays above 700 ev m 2 = -96 ± 189stat ± 63sys ev2 m 15.2 ± 2.0sys ev (90 % C.L.) activity A = 0.54 dec/mg/s 10 microcalorimeter array magreo = 271 g 4 A = 0.15 decay/s mtot = 2.71 mg EFWHM = 28.5 ev rise = 490 s fpile-up C. Arnaboldi et al., Phys. Rev. Lett. 91 (2003) M. Sisti et al, NIM A 520 (2004) 125 Angelo Nucciotti A. Nucciotti, NuMass2010, 8-11 Feb 2010, INT, Seattle 35 35

35 MARE extensions: 163Ho electron capture measurement Ho + e- 163Dy* + νe 163 calorimetric measurement of non-radiative Dy 8x8 array atomic de-excitations (Coster-Kronig, Auger...) fraction of events at end-point may be as high as for 187Re: depends on QEC ( 2.5 kev) QEC? fewer active nuclei are needed ( 4000 y) can be implanted in any suitable absorber first implantation tests at ISOLDE are encouraging new NASA/Goddard TES arrays ( E = 2eV) can be implanted with 163 Ho 36

36 163 Ho end-point statistics Re 163 Ho 3 H

37 163 Ho MC sensitivity estimate 38

38 163 Ho spectrum simulation 39

39 163 Ho pile-up spectrum De Rujula and Lusignoli, Phys. Lett. 118B (1982)

40 MARE and the cosmological relic neutrino background MARE-2: detectors, 20 mg each 650 g of 187Re counts/year... A. G. Cocco, G. Mangano and M. Messina, arxiv:hep-ph/ v2 41