UKDMC Dark Matter Search with inorganic scintillators Vitaly A. Kudryavtsev University of Sheffield IDM2000 York, UK September 19, 2000

The UK Dark Matter Collaboration N. J. C. Spooner, T. Gamble, V. A. Kudryavtsev, T. B. Lawson, M. J. Lehner, P. K. Lightfoot, R. Lüscher, J. E. McMillan, J. W. Roberts, M. Robinson, D. R. Tovey Department of Physics and Astronomy, University of Sheffield, Sheffield N. J. T. Smith, P.F. Smith, G. J. Alner, S. P. Hart, J. D. Lewin, R. M. Preece Particle Physics Department, Rutherford Appleton Laboratory, Chilton, Oxon T. J. Sumner, J. J. Quenby, B. Ahmed, A. Bewick, D. Davidge, J. V. Dawson, A. S. Howard, I. Ivaniouchenkov, W. G. Jones, M. K. Joshi, V. Lebedenko, I. Liubarsky Blackett Laboratory, Imperial College of Science, Technology and Medicine, London J. C. Barton Physics Department, Birkbeck College, London

Outline 1. UK Dark Matter experiments Technique Anomalous fast events in NaI(Tl) detectors 2. Recent test with CsI(Tl) crystal Spectrum of fast component Decrease of the rate of fast events 3. New array of unencapsulated crystals (NAIAD - NaI Advanced Detector) Design Preliminary results from the first module 4. Conclusions.

UK Dark Matter experiments At present several NaI(Tl) detectors are running in the underground laboratory in the Boulby mine. Each crystal is viewed by two PMTs. Integrated pulses from the PMTs are digitised using a LeCroy oscilloscope driven by a Labview-based DAQ software. Pulse shape analysis is used to discriminate between the the pulses due to nuclear and electron recoils. Nuclear recoil pulses can be caused by WIMP interactions or neutrons, while electron recoil pulses are caused by gamma background. Time profiles of integrated pulses from the PMTs are fitted to the exponential to derive pulse parameters, such as amplitude (energy) and time constant. Time constant distributions of observed events reveal the presence of anomalous fast population of events. The mean time constant of this population is smaller than gamma/electron time constant and even smaller than nuclear recoil time constant.

Typical integrated pulses from NaI(Tl) crystals

Anomalous fast events (‘bump’) Time constant distributions from DM46: E=35-40 keV Data Compton Calibration

Spectra of anomalous fast events in various NaI(Tl) crystals

DM26 (NaI) spectrum: fast component (low energies) + alphas Spectrum of anomalous fast events in DM26 2kg crystal from data covering several energy spans. The MeV range peaks correspond to the expected alpha spectrum from U/Th Graph from P. F. Smith et al. Phys. Rep., 307 (1998) 275

Alpha contamination hypothesis Alpha contamination looks like the best explanation (P. F. Smith et al. Phys. Rep., 307 (1998) 275) Crucial point: there should be many alphas depositing small energy in the crystal. But internal bulk U, Th level is too low to account for the high rate at low energies. External incoming alphas from U, Th outside the crystal (PTFE?) needs ~1 ppm and very fine tuning of dead layers and NaI efficiency. Moreover, time constant of external alphas is not matched well to that of fast component. Internal (surface) contamination of crystal by alpha emitting isotope(s) (N. J. T. Smith et al. Phys. Lett. B 485 (2000) 9). Nuclear recoils from Rn decay can be implanted into the crystal surface. This creates a thin ( microns) alpha emitting layer. Alternatively, crystal surface can be contaminated by U/Th. In both cases high concentration of alpha emitting isotope is needed (0.1-1 ppm). Test of surface alpha hypothesis - polishing crystal surface and run experiment again.

Test with CsI(Tl) detector Why CsI(Tl)? Non-hygroscopic crystal, can be handled much easier than NaI crystals, can be re-polished Better discrimination between nuclear and electron recoils than in NaI(Tl) (this suggest better discrimination between ‘fast’ events and ‘normal’ electron recoils) NaI(Tl) CsI(Tl)

CsI(Tl) crystal - before polishing Time constant distribution in CsI(Tl) crystal before polishing - E=30-50 keV Time constant, ns Number of events

Rate of fast events in CsI(Tl) detector

CsI(Tl) crystal - after polishing Time constant distribution in CsI(Tl) crystal after polishing - E=30-50 keV Time constant, ns Number of events

Rate of fast events in CsI(Tl) detector after polishing The first two points (arrows) show the upper limit on the rate

Design of an unencapsulated detector

Unencapsulated vertical detector

Preliminary results from the first NAIAD module Time constant distributions from DM74: E=35-40 keV

Preliminary results from the first NAIAD module Exposure kg  days keV11-12 keV keV13-14 keV

Preliminary limits from the first NAIAD module Spin-independentSpin-dependent (higgsino) Blue - Na, green - I, red - NaI Halo parameters:  =0.3 GeV/cm 3, v o =220 km/s, v Earth =232 km/s, v esc =650 km/s

Conclusions Anomalous fast events observed in the UK Dark Matter experiments can be explained by the contamination of the crystal surface by the products of radon decay. It is unknown, however, how and when radioactive nuclei were implanted into NaI(Tl) crystals. Results from test experiment with CsI(Tl) crystal show that polishing crystal surface removes major part of fast events. UKDMC has started a new program to run unencapsulated NaI(Tl) crystals (NAIAD array) in liquid (paraffin) and in nitrogen atmosphere. The crystals can be taken out from the detectors and be re-polished. First NAIAD module does not show contamination from anomalous fast events. Preliminary limits on WIMP- nucleon cross-section are presented.