The DEAP-1 Detector at SNOLAB Chris Jillings, SNOLAB/Laurentian U. For the DEAP/CLEAN Collaboration
Jan 27, 2016
The DEAP-1 Detector at SNOLAB
Chris Jillings, SNOLAB/Laurentian U.
For the DEAP/CLEAN Collaboration
The DEAP-1 Detector
Nuclear recoil
Electron recoil
DEAP-1 at Queen’s
arXiv:0904.2930
Demonstrated a pulse shape discrimination between electron recoils and nuclear recoils at ~4x10-8
Detector stability(120-240 pe)
Measured at 511 keV2.9
2.7
2.8
DEAP-1 moved to SNOLAB in 2007
• Runs underground: – December 4, 2007 to January 2008– v1 clean chamber: July 4, 08 to Dec 6, 08– v2 clean chamber: March 19, 09 to Dec 10, 09– v3 clean chamber: March 25, 10
• PSD improved to ~ 10-8
• Light yield increased using HQE PMTs• Backgrounds in WIMP energy ROI greatly
reduced
Clean v1 chamber
Glove box preparation of inner chamber (reduce Rn adsorption/implantation on surfaces)
222Rn introduced from gas bottle, settles to about 25 decays per day
alpha
222Rn in DEAP-1 (Gen 1)
DEAP-1 Gen 2 chamber
• DEAP-1 inner chamber redesigned, teflon as reflector instead of TiO2 paint
• Radon trap installed for filling
Gen 1 chamberGen 2, no Rn spike and ~10 times cleaner
Stability (Generation 2)
Stable to 10% over 150 days
Gen 2 data taken with new DAQ
Gen 3: Improved light yield
v3, ~4.7 pe/keV
v2, ~2.5 pe/keV
60 keV gammas from 241Am in AmBe neutron calibration runs
Hamamatsu R5912 HQE PMTs
• Qualified two of each candidate 8” PMT• Evaluate gain, relative efficiency, dark rate,
timing, late pulsing, after pulsing, prepulsing, magnetic field sensitivity ....
R5912 HQE will be used for DEAP-3600.
PMT in testing facility at Queen’s
5912 SPE
<75 ppb U/Th forR5912
Manufacturer Model No. Eff. Relative to R1408
Hamamatsu R1408 (SNO) 1.00
Hamamatsu R5912 1.30
Hamamatsu R5912 HQE 1.40
Photonis XP1806 1.18
Electron Tubes ET9354KB 1.20
Background rates in DEAP-1 versus time
v3 data being analyzed
120-240 pe region
Background Questions
• Given the efforts at surface cleaning between Gen 2 and Gen 3 yielded small results, is there a source of low-energy backgrounds we are missing?
• Is the WIMP-region background caused by radon in the bulk?
• Or quantitatively: what is the event rate in the WIMP region induced by radon in the bulk?
• A sample of radon extracted from approximately 100 litres of air, after corrections for efficiencies, should add Bq levels of radon.
Radon Spike From Air
Procedures and equipment from SNO.
InletNaOH
Water trap(coils at -60C)
ChromaSorbTrap at -110C(ethanol slush)
Lucas cellDEAP Rn tube
Radon Spike
• Use high-flow trap with chromasorb at -110C to trap 222Rn.
• Oxygen, nitrogen and argon pass through trap.• Transfer radon with cryopumping to small trap• Volume expand radon into Lucas cell and Rn tube.• Count Lucas cell to measure Rn spike.• Next day: install on inlet to argon system.
• As long as only a few standard cc’s of contaminant gas, our SAES purifier will purify.
• Concentrating the radon in 1m3 of air is not considered a “source” by SNOLAB.
PSD Underground
• PSD is a huge data-reduction effort• Depends low-noise electronics• We have 27 TeraBytes of MIDAS data.
DAQ SamplingData Rate
Ev/sec
Data Rate
Mbyte/secBottleneck
Scope500 MHz
10 sec<~150 /s 1 Scope readout
V1720 &
MIDAS
250 MHz
16 sec~350 /s 8
Source strength
Sample PSD Data
Background To PSD
• The detector high-Fprompt background rates have some probability of being coincident with a valid tag as described in the DEAP-1 Surface paper (arXiv:0904.2930).
• Depends on rate of tags and the time window imposed in analysis.
• We expect:
RunPSD
Entries
Expected #
pile-up events
Surface 17 M 0.26
U/G 2008 (scope) 22 M 0.16 (preliminary)
U/G 2009 (MIDAS) 70 M 0.13 (preliminary)
Total 109 M 0.45
Analyzed PSD
Future PSD
• Surface, Gen 1 and Gen 2 data u/g had the same light yield. Analysis of Gen 3 PSD will allow the relationship between energy and PSD to be explored as well as effects of photon counting.
• Would like few x 109 events background free.• Requires
– Optimized tagging– Stronger source
New 22Na source
Place source in bicron BC-490 plastic scint in mold.
Double-tag:
1- positron in plastic
2- back 511keV1 cmPMT
12
Source in design stages. Early testing with BC-490 successful.
Neutron-Shielding/M.C. Tests
• A series of runs were taken with the SNO AmBe neutron source behind various thicknesses of plastic
• Model test– Neutron spectrum from AmBe source (Neutron energy
spectrum from AmBe source depends on the grain sizes.)– neutron shielding Monte-Carlo calculations CLEAN nuclear-recoil quenching factor.
• Analysis ongoing
Fixed source holder
Frame holds from0.25” to 2” HDPP
Some Notes About Analysis
• Switching PMT and base circuit forced change in baseline algorithm.
• PMT SPE mean charge was determined using a mean charge over a restricted integration window. We have developed fits to Polya functions.
• Software noise-reduction techniques developed.• Re-analysis of all SNOLAB data just underway.• Goal: submit manuscripts for publication in timely
way.
DEAP-1 to DEAP-3600
• Light yield in DEAP-1 + Monte Carlo Light yield in DEAP-3600 > 8 pe/keV (with R5912 HQE PMTs)
• Stability of DEAP-1 suggests that continuous purification of Argon not needed in DEAP-3600 (but it is available)
• PSD data are consistent with surface results: PSD model used holds up.– Detailed analysis of Gen 3 PSD underway. This is important
because PSD depends on statistics of photon counting and energy.
• PMT/Electronics for DEAP-3600 prototyped on DEAP-1– We are likely to go to a tapered base to improve signal linearity.
• Measured backgrounds in DEAP-1 allow for DEAP-3600 with reduced FV to be useful.
• Re-assembly of DEAP-1 in J-drift after with cleaned plumbing and new chamber.
Next 12 Months
• Move to J drift• Gen 4 acrylic chamber
– Better control of neck events– Wash all argon plumbing lines– Small improvements to cooling system
• Hotter source for PSD with improved time tag
SNOLAB
• SNOLAB has provided – Services (IT, logistics …)
– LN2,
– technical staff, – engineering support, – URAs,– funds for new source development – extra shifts, …
People
• DEAP-1 slides shown here are drawn from work by many including people at…– LU/SNOLAB (incl 9+ URAs in past three years)– Queen’s– Alberta– TRIUMF– Carleton– Yale– U. North Carolina– U. New Mexico– LANL
Position reconstruction
Size of DEAP-1
Very good position reconstruction, useful for identifying surface background events
Background rates in DEAP-1 (120-240 pe)
Date Background Rate (in WIMP ROI)
Configuration Improvements forthis rate
April 2006 20 mBq First run (Queen’s) Careful design with input from materials assays (Ge couting)
August 2007 7 mBq Water shield (Queen’s) Water shielding, some care in surface exposure (< a few days in lab air)
January 2008 2 mBq Moved to SNOLAB 6000 m.w.e. shielding
August 2008 0.4 mBq Clean v1 chamber at SNOLAB
Glove box preparation of inner chamber (reduce Rn adsorption/implantation on surfaces)
March 2009 0.15 mBq Clean v2 chamber at SNOLAB
Sandpaper assay/selection, improved purging, PTFE instead of BC-620 reflector (from Rn emanation measurements), Rn diffusion mitigation, UP water in glove box, documented procedures;Rn Trap@SNOLAB for filling.
March 2010 ? Clean v3 chamber at SNOLAB
Acrylic monomer purification for coating chamber. TPB purification.
Table from Mark Boulay
Alpha backgrounds
• Are very high energy• Non-linear energy response must be
calibrated out.
Clipping of Prompt Light
Average alpha
Average low energy recoil scaled to alpha energy
• Protection diodes clip the pulse
• Clipping is necessary to observe alphas and low energy recoils in the same run for DEAP-1 (clipping will be rare in DEAP 3600)
• New energy scale required for alphas
Energy Non-linearity
• Each PMT sees a >50% change in light based on event vertex position
• With clipped pulses, the effective gain may be highly non-linear over this range
• Methods to deal with this:1.Correct for clipping (currently gives ~10% energy resolution)2.Develop independent alpha energy scale (currently gives ~3% energy resolution)
Radon Daughter Coincidence Tags
232Th Chain238U chain
• Timing coincidences for alpha decays give calibration points for the alpha energy scale
Radon 220 Coincidences
Fit T½ = 0.15 ± .02 sReal T½ = 0.15 s
220Rn
216Po
Polonium 214 Coincidences
Fit T½ = 163 ± 27 usReal T½ = 164 s
214Po
214Bi
Correcting for Nonlinearity
Correcting for Nonlinearity
Corrected Prompt = Total Prompt _1 – 0.05 PromptZ + 1.33 PromptZ2
PromptZ = Prompt0 – Prompt1 Prompt0 + Prompt1
Calibrated Alpha Spectrum
Daughter Constrained Fit Unconstrained Fit222Rn 267 ± 14 325 ± 54218Po 267 ± 14 214 ± 20214Po 41 ± 7 42 ± 9 210Po 35 ± 18 2 ± 58220Rn 68 ± 7 123 ± 35216Po 68 ± 7 54 ± 9212Po 20 ± 10 20 ± 10
Χ2/dof 83/60 67/58
All widths are set at 2.9%