Discussion of Proposed mini-TimeCube UH Team: Michelle Alderman, Steve Dye, John Learned, Shige Matsuno, Marc Rosen, Michinari Sakai, Stefanie Smith, Gary Varner

20 January 2010 mini-TimeCube presentation to NGA 2 Idea Build small (2 liter) liquid scintillation detector which uses very fast timing to recognize neutrino interactions (inverse beta decay signature) in the face of heavy noise. Timing replaces optics, hence “TimeCube”. Employ new (MCP) PMTs, only recently available, which have many small pixels (mm) and very fast timing (~100ps or less) to reconstruct images, and reject background on the fly. Prospect of 1m^2 panels (ANL, Chicago, HI) of such detectors for factor of more than ten less than PMTs will be a game changer in neutrino detection. The mini-TimeCube is a start, with important application in reactor monitoring. Arradiance availability?

20 January 2010 mini-TimeCube presentation to NGA 3 58mm edge 126mm cube 133mm V = 2L A = 953cm 2 Max PMTs = 24 (~75% coverage) V = 1.5L A = 626cm 2 Max PMTs = 12 (~54% coverage) Two geometries under consideration: dodecahedron and cube Cube favored for prototype

20 January 2010 mini-TimeCube presentation to NGA 4 The photo-sensor: Photonis XP85012 (64 channel MCP)

20 January 2010 mini-TimeCube presentation to NGA 5 408nm laser 100 Photo-Electrons Conclusions: Gain is 40mV/100= 0.4mV/PE (25  m) at 2100 V 5mV/100= 50  V/PE (10  m) at 2500V 10  m somewhat faster rise time, longer trailing edge, presumably due to 4 pads connected together. The rise time does NOT depend upon the amplitude Signals

20 January 2010 mini-TimeCube presentation to NGA 6 Data processing card cPCI crate cPCI CPU Data Acquisition System (DAQ) based on cPCI format 3Gbs fiber link x3 (= 24 PMTs) x1

20 January 2010 mini-TimeCube presentation to NGA 7 Mini-TimeCube Sensitivity for nominal 12.6 cm cube with 16 Photonis 64 chanel PMTs Rate: 15 anti-neutrino events/day at 25m from 3.3GW power reactor. (Times some yet ill determined efficiency factor, maybe 50%) Very rough cost estimate: $300K, with electronics and several spares, no labor. ($8500/PMT, and $100/channel for electronics, inclusive) Cube area = cm^2 PC Area (16 PMTs) = cm^2 Fractional coverage = 47.4% (75% with 24 PMTs) 1024 pixels (1536 for 24 PMTs) Sensitivity, assuming 10,000 quanta/MeV and 25% quantum efficiency of PMTs, is thus = 1200 PE/MeV energy deposition (electron equivalent). Typical reactor anti-neutrino, with ~2 MeV positron energy then would yield about 2400 PE, and all channels typically with several PE. For location of tracks, scintillator will be dominating constraint. Assume for here that we have 1 ns scintillator decay constant. Assuming 100ps first hit time resolution, this time corresponds to 20mm in light travel distance in the scintillator. With 1 ns scintillator, 10% of hits will be within first 100ps, or 120 PE/MeV prompt hits. (How well can we use these in a Fermat weighted reconstruction? TBD)

20 January 2010 mini-TimeCube presentation to NGA 8 Impulse dark noise vs HV Conclusion: At full efficiency (25  m 2000V, 10  m 2400V), dark counts rates are: 25Hz (25  m) 20Hz (10  m)

20 January 2010 mini-TimeCube presentation to NGA 9 Neutrino Vertex Resolution With few thousand hits in 1024 channels, some hits in every neighborhood close to Fermat time. With 120 PE/MeV on Fermat surface, we can use time-space fit to get to roughly 20mm/sqrt(120*E/MeV) = 1.3 mm for 2 MeV. Track length is of order 2cm*E/MeV, so 4cm track for 2 MeV deposit gives 4cm positron track. Given small dimensions of cube, most positron annihilation gammas will not interact but leave detector, so we only need reconstruct positron track vertex. Technique of time weighted moments (center-of-time) for early hits was shown to yield point near start of track, easy algorithm. Center of charge, yields estimate of track center, and combination yields estimate of positron Direction. Positron direction somewhat interesting, but most importantly initiation point gives origin of neutron. Summary for prompt detection: find vertex to several mm, direction to few degrees (?).

20 January 2010 mini-TimeCube presentation to NGA 10 From Hiroko Watanabe

20 January 2010 mini-TimeCube presentation to NGA 11

20 January 2010 mini-TimeCube presentation to NGA 12

20 January 2010 mini-TimeCube presentation to NGA 13

20 January 2010 mini-TimeCube presentation to NGA 14

20 January 2010 mini-TimeCube presentation to NGA 15

20 January 2010 mini-TimeCube presentation to NGA 16

20 January 2010 mini-TimeCube presentation to NGA 17

20 January 2010 mini-TimeCube presentation to NGA 18

20 January 2010 mini-TimeCube presentation to NGA 19

20 January 2010 mini-TimeCube presentation to NGA 20 Neutron Detection assuming 6Li loading. Net energy from 6Li is 680 KeV equivalent => 816 PE ! Position resolution thus quite good… ~ 3mm! Thus angular resolution on neutron vector (with length around 4.4cm) very good, order 5 degrees. Neutrino resolution should be mainly from weak interaction kinematics. May do better than above calculations using more information (both vectors). + Pulse shape discrimination adds to signal discrimination.

20 January 2010 mini-TimeCube presentation to NGA 21

20 January 2010 mini-TimeCube presentation to NGA 22 Noise Rates Talk that was given in Trieste last summer by Battaglieri from Genoa. Their idea is to build a segmented detector of square logs of scintillator with 3 inch PMTs on the ends, eventually aiming at a detector around 1m3 but they operated a smaller prototype. In this attached report are some nice measurements of various scintillators they evaluated. They chose NE110 plastic. In their prototype measurements they had troubles at the (apparently not so well shielded old Romanian reactor) reactor they visited, which had a huge difference between reactor on and off in backgrounds (unlike San Onofre). Their idea was to use no shielding, so this is directly applicable to our case. They did not do so well in rate because their time window was tau = 330 microsec, and singles rates with reactor on were R1 = 120/s. With a dumb trigger of two hits in this window the net trigger rate of R12*tau = 4.75/s So, this is most encouraging, that even with their much larger mass and sensitive volume of 40x30x30 cm3 (not initially sure how to scale this... by volume I suppose, so something like 18 times our volume. If we scale by surface area it would be more like 10x. If we take the more modest 10x factor in singles rates, and we take a more reasonable loaded scintillator capture time, let us say 1/10 th of theirs, our raw two fold random rate would be down by 1000 from theirs, and hence totally trivial. (Of course this is not trivial compared to the neutrino rate of a few per day, but it is trivial compared to what we can easily harvest and chew upon in our leisure.) In any event, this looks very nice for us. I have a hard time believing it could be so good.... Next we will have to use GEANT to figure out how much rejection we can get.

20 January 2010 mini-TimeCube presentation to NGA 23 Elastic Scatters of Neutrons Looks better than in earlier notes… proton Birks factor much less than for alpha, hence 20 KeV kinetic energy transfer, may be on order of 5PE equivalent. Can we discern this between prompt and delayed? Probably. Can we reconstruct anything to help? Looks doubtful. Needs study.

20 January 2010 mini-TimeCube presentation to NGA 24 Items for study Need GEANT Simulations of mini-TimeCube Study response to various processes, such as stopping muon and decay. Study of liquid scintillators… find shortest time, best Li loading. LS compatibility issues (fallback, use quartz container) Study use of boron loaded plastic from Eljen. Effectiveness of pulse shape discrimination for neutron capture? Studies of backgrounds in reactor circumstances: random hits from external and internal gammas + thermal neutrons = potential fake neutrino signature. (But looks very good) Check on this due to wrong order of prompt and delayed signals (self measured background rate). Can we do anything with neutron elastic scattering? More….