Estimate of the Background Rate at Low Energies in Nai Crystal from Stainless Steel Of

Estimate of the background rate at low energies in NaI crystal from stainless steel of the vessel

V.A. Kudryavtsev

University of Sheffield

27 June 2010

To estimate the background rate induced by gamma-rays originated in the stainless steel vessel of the NaI detector, I have used the results of Monte Carlo simulations carried out by Matt Robinson (Sheffield) and published in Kudryavtsev et al., Astroparticle Physics 33 (2010) 91. The simulations were carried out with GEANT4 v9.2 for the DAMA/LIBRA detector setup. Fig. 1 shows the spectra of background events seen in NaI due to different sources: U/Th decay chains and 40K decay in PMT envelopes, as well as 60Co decay in the copper vessel.

For those simulations the concentrations of 30 ppb U and Th in 200 g of the 2 PMT envelopes were used resulting in a factor of 2.5 higher rate of electron recoils (Compton electrons) from U than from Th. The measurements of radio-purity of stainless steel for DM-Ice (presented by Darren Grant at the EVO-meeting on 22 June 2010) showed the concentrations of Th around 0.6 ppb and of U around 0.07 ppb. As there is almost 10 times more Th than U, obviously the contribution of Th will be dominant. 40K can be neglected unless natural potassium is 103 times more abundant than both U and Th.

The average rate of events at 1-10 keV from Th on Fig. 1 is about 4.5´10-3 events/kg/day/keV. The rate includes only single electron recoils which in fact dominate in the DAMA background. If we assume 0.6 ppb of Th in 200 kg of stainless steel at the same distance from the crystal as PMTs, then we can scale the rate as: 4.5´10-3 / 0.2 kg ´ 200 kg / 30 ppb ´ 0.6 ppb = 0.09 events/kg/day/keV. Certainly a significant part of the vessel will be further from the crystals than PMTs, so this value can be considered as an upper limit, although 30-40% increase is foreseen to take into account gamma-rays from U and the fact that only single events will be recorded in DM-Ice whereas the set up in the simulations contained closely packed crystals allowing the detection of multiple hit events which were rejected. Additional suppression of gamma-rays is provided by stainless steel itself (similar to a factor of 2-3 suppression of radiation from PMTs by light guides). In any case the value is much smaller than the rate reported for NAIAD crystals of 7-8 events/kg/day/keV and smaller than the rate in NaI crystals reported by DAMA.

Another way of estimating the background rate is to use the results for 60Co. The number of gammas per decay of 60Co (2) is similar to that from Th decay chain. Fig. 1 shows that the rate of about 1.8´10-2 events/kg/day/keV is expected from 10 mBq/kg of Cu in the vessel. The concentration of 0.6 ppb of Th corresponds to the decay rate of 2.4 mBq/kg. If the mass of stainless steel vessel in DM-Ice is 10 times bigger than that of the copper vessel in DAMA/LIBRA, then the rate can be scaled as: 1.8´10-2 ´ 10 (mass factor) / 10 mBq ´ 2.4 mBq = 0.043 events/kg/day/keV. (We assume here that copper shielding does not contribute to the events rate in DAMA/LIBRA since it is further away from crystals.) Again this rate is much smaller than the measured rate in NAIAD and DAMA crystals.

The difference in the two estimates can serve as a rough estimate of the uncertainty which is about a factor of 2-3. Even with the highest possible value for the estimated rate it is still a factor of 3-4 smaller than the DAMA/LIBRA rate.

In conclusion, I think that for our first detector, the vessel can be made of stainless steel since we will probably be dominated by large background from the crystal. For future deployments accurate Monte Carlo of radioactive background should be carried out.