carlos o. escobar fermilab dune workshop at ... -...
TRANSCRIPT
1. Why use a light signal in a LNG TPC?
2. How to detect the scintillation light in LAr?
2a: Light guides with WLS coating, two technologies
2b: Arapuca: a light trap using dichroic filters
3. Looking in the Near-Infrared (NIR)
4. A personal wish list of R&D activities in LNG scintillation
1
Motivations:The simultaneous read-out of the charge and light signals in LNG’s detectors is an important tool for many experiments ranging from DM to large LAr TPC’s. Charge and light are complementary. Light signal providing t0 (start clock) for non-beam physics such as SNB, nucleon decay. Also helps with particle ID:Q = Q0exp-(t-t0)/ , see next slide.
So far the light signal used (or planned to be used) comes from the VUV scintillation with wavelengths ranging from 78 nm (LNe) to 175 nm(LXe).
2
Scintillation mechanismLight emitted by excited diatomic molecules formed by two different mechanisms: 1) impact excitation and 2) recombination:e- + Ar --> Ar* + e- impact excitation
Ar* + Ar --> Ar2* excimer formation ( = 0 electr. exc.)
Ar2* + Ar --> Ar2* + Ar relaxation (collisional but also IR)
Ar2* --> Ar + Ar + h VUV emission
The last step occurs from one of two electronic excited states, a singlet one or a triplet one, denoted 1u
+ and 3u+ respectively, to
a dissociative ground state. is the same but is vastly different for LAr: 1.6 s for the triplet and 6 ns for the singlet.
2) The recombination path
e- + Ar --> Ar+ + 2e- ionization
Ar+ + 2 Ar --> Ar+2 + Ar
e- + Ar+2
__> Ar** + Ar recombination
Ar** + Ar --> Ar* + Ar + heat
Ar* + Ar + Ar --> Ar2* + Ar + heat
Ar2* --> Ar + Ar + h VUV emission
Last step is the same as before, same lifetimes and wavelength but under electric field recombination is
suppressed.
Motivations
VUV detection presents many challenges:
1. Use of WLS (long term stability? )
2. Complicated schemes for collecting the light
3. Singlet/triplet ratio by PSD needs excellent light collection ( see DEAP Coll.) and delayed light emission from TPB does not help it .
4. Recent estimates of Rayleigh scattering length puts it at 55 cm (Grace and Nikkel)- start having pernicious effects especially for extremely large LAr TPC’s (DUNE)
5. Attempting to improve charge collection (recovering charge lost by recombination) by doping with a photosensitive chemical such as TMG (Icarus -1995) kills the VUV light .
6. VUV poorly reflected from almost any surface/material; more opaque wire meshes.
3
PROTODUNE/DUNE VUV LIGHT DETECTION: SOME PROPOSALS
Light Guides inserted behind wire mesh in the APA:
2 approaches: Indiana University & MIT-FNAL
Another concept: the Arapuca light trap
The idea at the basis of the ARAPUCA is to trap photons inside a box with highly reflective internal surfaces, so that the detection efficiency of trapped photons is high even with a limited active coverage of its internal surface
FAPESP DUNE WORKSHOP JUNE 1, 2017
• The core of the device is a dichroic filter. It is a multilayer acrylic film - same technology used to produce reflective plastic foils like 3M VIKUITI or VM2000.
• It has the property of being highly transparent for wavelength below a cutoff and highly reflective above it.
12
Reflectance
cutoff
Transmittance
cutoff
CPAD 2016 - Calthec – 8-10 Oct 13
0
10
20
30
40
50
60
70
80
90
100
325 345 365 385 405 425 445 465 485
Tra
nsm
issi
on (
%)
Wavelength (nm)
400nm Dichroic Shortpass Filter FOR REFERENCE ONLY
Average Polarization
S-Polarization
P-Polarization
p-Therphenylemission spectrum
P-Terphenyl on the external side
Filter cut-off at 400 nm
Deposited by vacuum evaporation
0
50
100
150
200
250
300
350
400
450
350 375 400 425 450 475 500 525 550 575 600
wavelength (nm)
inte
nsity
(A.U
.)TPB emission spectrum
TPB on the internal side
CPAD 2016 - Calthec – 8-10 Oct 14
Filter cut-off at 400 nm
Deposited by vacuum evaporation
FAPESP DUNE WORKSHOP JUNE 1, 2017
• Arapuca with 5x5 cm2 acceptance window;
• Box with dimensions 5x5x0.6 cm3
• Read-out by 2 SiPM 0.6cm X 0.6cm active area each SensL MicroFC-60035-SMT (courtesy of Cormac Campbell - SensL Technologies Ltd.)
• Dichroic filter (Quantum Design- cutoff @ 400 nm – substrate fused silica)
• The device has been installed inside a liquid argon cryostat (SCENE) and exposed to an alpha source.
• Alpha source is a 1Ci 241Am that produces 5.4 MeV monochromatic particles
• Two different runs have been performed and two different read-out electronics have been tested
FAPESP DUNE WORKSHOP JUNE 1, 2017
a source
FAPESP DUNE WORKSHOP JUNE 1, 2017
0
25
50
75
100
125
150
175
200
225
0 10 20 30 40 50 60 70 80 90 100
P1 197.5
P2 8.401
P3 42.51
P4 98.15
P5 22.61
P6 48.30
P7 38.41
charge (mVolt nsec)
am
pli
tud
e (A
.U.)
0
100
200
300
400
500
600
0 50 100 150 200 250 300
P1 566.4
P2 101.9
P3 18.96
Signal integral (phel)
am
pli
tud
e (A
.U.)
• Encouraging results: we measured an efficiency of ~1% (Seq ~ 0.25 cm2) in this first test• It can be significantly improved considering that:
Low quality of the evaporations Thicknesses of the films non-optimized Internal reflectivity probably not at its maximum (cleanliness, quality of the
material, thickness of the box walls)
Single electron spectrum Source spectrum
Thanks to:
The Fermilab LDRD program.
Eileen Hahn, Jim Freeman, Ron Davis, Bill Miner, Brian Rebel and several others.
19
Picture of tallbo Picture of arapuca rack
FAPESP DUNE WORKSHOP JUNE 1, 201720
Some with TPB coated VikuitiSome with non coated Vikuiti
Dichroic filterAll have P-terphenyl on filter side (external)Some coated with TPB on glass (internal
ARAPUCA tests at TallBo: Jan and March 2017
The January test taught us a lot of good things about TallBo, ARAPUCA, the SSP electronics (ANL) and data analysis.
We had 6 ARAPUCAS with 3 brands of filters (Omega, Asahi, and Edmunds) and 3 sizes (5cm x 5cm, 5cm x 7cm and 2.5cm x 3.5cm (doubles)).
But it was a failure because the TPB coatings and some P-ter coating were damaged by thermal stress. Coatings too thick.
For the March test we figured that the coatings adhere better to the filter side, so we kept all the P-terphenyl coatings (200ug/cm2 and 500ug/cm2).
Some filters were coated on the glass side with 250ug/cm2 of TPB.
Some filters were single coated with P-terphenyl and the TPB was deposited on Vikuiti reflector at the bottom of the ARAPUCA.
We collected data from scintillation light from a Am241 source.5.1 MeV Alphas produce about 200K PEs, part of the light is quenched,
The actual number of PEs generated is unknown because there’s a thin gold layer on the source that may reduce the nominal number.
We also tested a 3 SensL array actively ganged: 4x4 (3mm x 3mm)
and we used a single array next to the Alpha source as a reference.
FAPESP DUNE WORKSHOP JUNE 1, 201721
• Below ARAPUCA number 3 for 4 different positions of the Am source.
• Maximum light at Z=6000
FAPESP DUNE WORKSHOP JUNE 1, 201722
of Am241 source
Z = 5200 Z = 5600 Z = 6000 Z = 6400
(plot by Ernesto Kemp)
(plots and analysis Adam Para)
Average signal charge = 23.4
Average number of photons detected = 23.4/2.149 = 10.89
From the gaussian fit : average =23.1, sigma = 8.65
Number of photons:
From average charge = 10.89 (fast and slow component included)
From the peak value = 9.35 (most of the slow component included due to the shaping time)
Single avalanche @ 12.74Average number of avalanches per primary signal = 15.57/12.74 = 1.22
Raw data
An estimate of channel 3 efficiency
Alpha source: peak energy 4.7 MeV ( 241Am
5.4 MeV, gold foil
# of 128nm photons (all): {(4.7e6)eV/19.5eV}x0.72
≈173,000 photons
Acceptance of channel 3 at y=5,600: 0.025
# of photons at the window of ch3: 4,300
Efficiency: 11/4,300 = 0.26% (bars typically have 0.1%)
At λ=255 nm P-terphenyl is at least 50% less efficient than at 128nm
A SETI UV-LED from was calibrated using a 818-UV/DB
NIST calibrated detector from Newport Lab. The detector
is connected to a Newport-Oriel 70310 calibrated meter,
with 1pW resolution.
LAr (128nm) relative absorption?
For a constant bias point the LED response shows to be a quite linear function of the pulse width down to 100ns pulses. A 100ns at 1KHz repetition rate pulse delivers 25pW optical power as measured by the detector.
Other calibration measurements were performed, varying the bias point of the UV-LED.
FAPESP DUNE WORKSHOP JUNE 1, 2017
• The LED is powered by a Tektronix pulser in current source mode with pulses of several widths and currents.
• The sensor uses an external OD3 attenuator with a responsivity of 2.5e-4 A/W at 255nm.
We collected data using a 5cm x 5cm ARAPUCA with a double side coated Omega filter.
We collected data at several pulse lengths and bias points of the UV-LED.
FAPESP DUNE WORKSHOP JUNE 1, 201728
By measuring darks the 1PE gain seems to be at 12 ADUs (needs verification)
Number of photons collected by ARAPUCA = fitted mean/1PE gainNumber of photons incident to ARAPUCA=Popt (J/s)/E(1PE,225nm)Efficiency = NPE collected/(NPE incident * P-terphenyl efficiency)
Still trying to figure out several things but it is on the order of 1% ±0.2%
20
Combined test: IU bars and ARAPUCAs side by side in TallBo:
August 20 to end of September
Adherence studies of TBP on sand blasted and etched Omega (5x5cm2) filters (glass side)
More optical tests to measure efficiency. Parallel test at Campinas with a 128nm lamp
New active ganging with optimized amplifier.
6/2/2017Presenter | Presentation Title29
NIR light: History and BackgroundNIR light emission from noble gases has been known since the late 40’s with lines around 1,300 nm.
In condensed noble gases NIR emission was seen in transient optical absorption experiments via electron excitation in late 60’s and early to mid 70’s by various groups. Next slide gives one example for Ar.
Bressi and collaborators pursued for many years the study of NIR emission in both gaseous and liquid states for xenon and argon: clear evidence for NIR emission from gas but much lower LY in liquid. 4
History and background cont.example of spectrum obtained in argon:
from Suemoto et al (1977)Top: solid Middle: liquid Bottom: gas
Same results were obtained by Arai and co-workers evidence for excited neutral rare Gases molecules(1.3 eV = 954 nm)
5
Recent investigationsTwo groups have pursued the investigation of the
NIR scintillation in LAr very vigorously in recent yearsNovosibirski (Bondar, Buzulutskov et al.) and TU Munich (Ulrich, Schoenert and various students and postdocs).
Both use table-top setups(cubic centimeter volumes)and very intense, pulsed low energy beams (12 keVelectrons Munich and pulsed X-rays between 30 and 40 keV-Bondar et al.)
Next slides show pictures of their apparatuses.6
Summary of the results from the two groups
1st results from the Munich group indicated NIR emission from LAr:
No absolute light yield but spectral information
9
recent investigations cont.Results from the Novosibirski group:
510+/-90 photons/MeV from 400nm to 1000nm
10
Recent investigations: codaBut, more recently (end of 2014 and beginning of 2015) the Munich group revisited their NIR results in LAr, claiming that impurities had caused the 1st result. Their new line of investigation is mixing small parts of xenon into argon (10 ppm produces the best results: 10,000 NIR photons/MeV at 1,180 nm)
Bondar et alHave not reconsideredtheir results.
11
Conclusions so farA vigorous, systematic, experimental program is still needed, one that would use larger volumes of LAr with
1)rigorous control of purity 2)spectral information3)determination of the time structure of an eventual NIR emission4) determination of the light yield5) “standard” ways of exciting the LAr (radioactive sources; cosmic rays; particle test beams)
12
In the mean timeTest using the Scene cryostat at Fermilab:Use CPTA SiPM’s for NIR detection (4.4 mm*2)Tagged 0.511 MeV gammas from Na22 source (1Ci)Tag with liquid scintillator and PMT.Good purity of LAr.Expect very low rate of events due to: source intensity; geometrical acceptance; small energy loss of gammas in the LAr (mostly Comptons). But would serve as a Yes or No test
13
Results
We trigger on our best SiPM (operate it at a bias voltage of 44 V- trigger on a fraction of a single p.e.) and look for the PMT pulse in a time window that spans negative (PMT fires before) and positivetimes (PMT signal after SiPM). Expect, if SiPM is triggering on scintillation light, that there are more early PMT pulses than late ones.
This is what we see:
15
Results cont
time distribution:
Not clearwhat thelong tailis.Fast comp.~ 200nstime const
Rate of events 0.02Hz confirmed by GEANT4Simulation of geometry and energy loss in the LAr. Rate is compatible with light with λ in the very end of the SiPM pde curve
λ > 900 nmTook data on gas and rate is compatibly lower (small energy loss).
17
A new test using SCENE-summer 2016
Same CPTA’s SiPMs, same read-out electronics but now using an alpha source Am 241, 1C (37,000 dps)
Work done with summer intern at Fermilab,Alex Jamieson-Binnie from University of St. Andrews, UK.
Next slides are from his final presentation at Fermilab
18
Experimental Design – Setup
715 nm Filter
• Transmits photons with wavelengths greater than 715 nm
47FAPESP DUNE WORKSHOP JUNE 1, 2017
Analysis uses the triple to double coincidence ratio method
We trigger on double coincidence and look for a triple coincidence, the ratio of triples to doubles give the light yield times the acceptance times the pde of the detector
22
Analysis - Signals
-0,1
-0,05
0
0,05
0,1
0,15
0,2
0,25
200 220 240 260 280 300
Vo
lta
ge
(V
)
Time (ns)
Detector 1
Detector 2
Detector 3
49FAPESP DUNE WORKSHOP JUNE 1, 2017
Analysis – Photon Detection Efficiency
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
400 500 600 700 800 900 1000
Pe
rce
nta
ge
Wavelength (nm)
Relative Intensity
50FAPESP DUNE WORKSHOP JUNE 1, 2017
Analysis – Photon Detection Efficiency
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
400 500 600 700 800 900 1000
Pe
rce
nta
ge
Wavelength (nm)
Relative Intensity Filter Transmittance
51FAPESP DUNE WORKSHOP JUNE 1, 2017
Analysis – Photon Detection Efficiency
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
400 500 600 700 800 900 1000
Pe
rce
nta
ge
Wavelength (nm)
Filtered Intensity
52FAPESP DUNE WORKSHOP JUNE 1, 2017
Analysis – Photon Detection Efficiency
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
400 500 600 700 800 900 1000
Pe
rce
nta
ge
Wavelength (nm)
PDE Filtered Intensity
𝒑𝒅𝒆 = 𝟏% − 𝟐𝟎%
53FAPESP DUNE WORKSHOP JUNE 1, 2017
Analysis – Photon Yield
0,01
0,1
1
10
100
1000
0,1% 1,0% 10,0% 100,0%
NIR
ph
oto
ns p
er
alp
ha
PDE in %
Detector 1
Detector 2
Detector 3
Viable PDE
54FAPESP DUNE WORKSHOP JUNE 1, 2017
September-October, 2016
Taking a ride with the Indiana University group (Stuart Mufson, Denver Whittington and Bruce Howard)-see Bruce’s talk
29
Using a Hamamatsu NIR PMT R669, at none of the two borosilicate glass windows of the cryostat.
photocatodearea: 46 mm AreaxQ.E. larger than red sensitive CPTASiPMs.A 715 nm filter in front of PMT, reduces area to 25.4mm, still 5 times More sensitivity thanSiPM
30
No quantitative results so far but PMT pulses were seen correlated with the hodoscope trigger. A time stamp for the triggers will be provided by the IU group and compared to our time stamp. Ex. of wave forms
31
Conclusions so far
NIR scintillation from LNG is a promising alternative for the much needed light signal, one that avoids problems associated with the detection of the VUV light and opens the door for simultaneously improving the charge collection in LAr TPC’s. Simultaneous detection of VUV and NIR in smaller volumes (of the scale of DM experiments) could help separate nuclear recoils from the electromagnetic background.
32
A new experiment: A dedicated NIR Cryostat at PAB(with Paul Rubinov ) Just started cryogenic work, early May. Two glass chambers (Pyrex), 3 MgF2 observation windows, LN2 cooling
Main challenges in this program:
Determine the light yield in the NIR
Obtain spectral information
Determine time structure of the signal
Imperative to turn this into a useful tool:
1. Good and affordable NIR photodetector with single-photon sensitivity.
2. Efficient and affordable light collection system
34
A wish list of R&D topics in Scintillation Light from Liquid Noble Gases
Measure Rayleigh Scattering Length in LAr and LXe
Determine Reflectivity of VUV light from commonly used materials in LNG TPC’s (SS, Au, Al, PTFE, G-10)
Aging studies of WLS (TPB, pTp, BIS-MSB…) coated elements in LNG’s
Measure attenuation length of VUV light in LNGs (role of impurities, see recent ArDM result-surprises in store?)
VUV Light Yield: pin down LY (fractional LY for fast/slow components; LY as function of E, LY as function of LET)
NIR: LY, , LY vs LET
NIR efficient and affordable detection: new red sensitive SiPMs, InGaAs and HgCdTe avalanche photodiodes, graphene based photon detectors
NIR light collection (nanotechnology, photovoltaic industry)
Emission spectrum of LAr as measured by the TUM group (electron excitation)
from M. Hofmann’s thesis at TUM, Germany
A message from Prof. Andreas Ulrich, TUM (9/16/15)
As you say in your slide about “future plans”, spectroscopic information will be important to learn about the origin of the NIR signals which you presented so clearly. I had asked Alexander Neumeier to go over his files to see what he might have besides the data in our publications. He found 4 spectra from gaseous and liquefied argon with a 1% nitrogen admixture. Those spectra actually show several lines between 500 and 1200nm. The data were recorded with a not very sensitive PbS detector. So the light may be not so weak. Unfortunately, these spectra are so poor quality and we have no conclusive assignment for the lines that I am not sending them at the time being.
A wish list of R&D topics in Scintillation Light from Liquid Noble Gases
Measure Rayleigh Scattering Length in LAr and LXe
Determine Reflectivity of VUV light from commonly used materials in LNG TPC’s (SS, Au, Al, PTFE, G-10)
Aging studies of WLS (TPB, pTp, BIS-MSB…) coated elements in LNG’s
Measure attenuation length of VUV light in LNGs (role of impurities, see recent ArDM result-surprises in store?)
VUV Light Yield: pin down LY (fractional LY for fast/slow components; LY as function of E, LY as function of LET)
NIR: LY, , LY vs LET
NIR efficient and affordable detection: new red sensitive SiPMs, InGaAs and HgCdTe avalanche photodiodes, graphene based photon detectors
NIR light collection (nanotechnology, photovoltaic industry)