4/10/2014
Frank Calaprice
Department of Physics
Princeton University
Stanford Univsersity Seminar May 7 2014 1

• Chlorine experiment: (Pontecorvo, Alvarez, Davis)
– First solar neutrino detector was the chlorine radiochemical experiment.
– Technique avoided the intense source of radiological backgrounds by detecting
37 Ar from the reaction 37 Cl(ν,e) 37 Ar on a target of chlorine atoms.
• Gallium radiochemical experiments: (Gallex, SAGE)
– Used simliar radiochemical technique to measure pp neutrinos: 37 Ga(ν,e) 37 Ge
• Water Cerenkov: (Kamiokande, Super-K, and SNO)
– Detected high enegy 8 B neutrinos (> 5 MeV ) to avoid radiogenic backgrounds
• Borexino Liquid Scintillator
– First experiment to directly detect low energy neutrinos in the midst of intense radiogenic background below 3 MeV.
– A breakthrough made possible by “brute force” low-background technology.
4/10/2014 Stanford Univsersity Seminar May 7 2014 2

• Borexino is a large liquid scintillator detector operating in the Gran
Sasso laboratory in Italy.
• It the only operating experiment capable of direct detection of low energy solar neutrinos.
• It’s signature feature is unmatched ultra-low background achieved at start-up of Phase 1 in 2007, and improving in Phase 2 and 3.
• Phase 1 2007-2010 yielded data on 7 Be, pep, and 8 B solar neutrinos, geo-neutrinos, and other results.
• Phase 2 2010-2014 achieved lower background by scintillator repurification and aims to measure pp neutrinos.
• Phase 3 2015-201? will exploit improved scintillator re-purification methods for additional reduction of background toward measuring
CNO neutrinos and addressing solar metallicity problem.
4/10/2014 Stanford Univsersity Seminar May 7 2014 3

4/10/2014
The pp cycle The CNO cycle
<0.42 MeV
1.44 MeV
0.86 MeV
<15 MeV
<1.2 MeV <1.7 MeV
<1.7 MeV
Stanford Univsersity Seminar May 7 2014 4

4/10/2014 Stanford Univsersity Seminar May 7 2014 5

Neutrino Energy Spectrum Neutrino-Electron Elastic Scattering
Energy Spectrum
Ev e nt s / (d a y x 1 0 0 to n s x 1 0 p . e. )
4/10/2014 Stanford Univsersity Seminar May 7 2014 6

Neutrino Energy Spectrum Neutrino-Electron Elastic Scattering
Energy Spectrum
Ev e nt s / (d a y x 1 0 0 to n s x 1 0 p . e. )
4/10/2014 Stanford Univsersity Seminar May 7 2014
X2 for keV
7

• Place unavoidable radioactivity (PMT’s, etc.) far away and totally shielded from detector.
– Use active liquid shields (water, liquid scint.).
– Define fiducial mass far from surfaces: use event position.
• Use a detector that can be purified in situ to remove internal radioactivity.
– Liquids and gasses are good. Solids more challenging.
• Develop effective purification method to reduce radioactivity in detector and in shielding.
– Distillation, water extraction on liq. scintillator work well.
4/10/2014 Stanford Univsersity Seminar May 7 2014 8

• Water and Scintillator Shielding
– Purified for ultra-low internal radioactivity.
• Scintillator Containment Vessel
– Nylon balloon with small mass and low radioactivity.
• Scintillator Purification System
– Pseudocumene (PC) & 1.5 g/l PPO
– Distillation, water extraction, and N
2 gas stripping.
4/10/2014 Stanford Univsersity Seminar May 7 2014 9

(Mostly Active Shielding)
• Shielding Against Ext. Backgnd.
– Water:
– Buffer zones:
2.25m
2.5 m
– Outer scintillator zone: 1.25 m
• Main backgrounds: in Liq. Scint.
– 14 C/ 12 C
• 10 -18 g/g. cf. 10 -11 g/g in air CO
2
– U, Th impurities
– Cosmogenic 11 C (t
1/2
= 20 min)
– 222 Rn daught ( 210 Pb, 210 Bi, 210 Po)
– 85 Kr (air leak)
• Light yield (2200 PMT’s)
– Detected: 500 p e
/MeV (~5%)
• Pulse shape discrimination.
– Alpha-beta separation
4/10/2014 Stanford Univsersity Seminar May 7 2014 10

• Simple two-component liquid scintillator
– Pseudocumene + 1.5 g/liter PPO
• Chose two-component without secondary wavelength shifter to simplify purification and re-purification.
• High light yield (11,000 ph/MeV) with fast timing and excellent alpha-beta pulse shape discrimination.
• Aggressive to acrylic and many plastics, but very safe with nylon.
• Not safest and most environmentally friendly, but not bad.
4/10/2014 Stanford Univsersity Seminar May 7 2014 11

Fabricated in special Princeton Low-Radon Cleanroom
John Bahcall
First hermetically sealed cleanroom with low-radon air was developed to avoid surface radioactivity due to 222 Rn daughters:
 210 Pb (22 yr).
Fabrication time: > 1 yr
Low-radon cleanrooms are becoming common low-background research.
4/10/2014 Stanford Univsersity Seminar May 7 2014 12

238
232
Radon in air deposits 210 Pb (22 yr) on nylon foil, which later contaminates scintillator with 210 Bi (1 MeV b ) and 210 Po (5 MeV a ).
4/10/2014 Stanford Univsersity Seminar May 7 2014 13

4/10/2014 Stanford Univsersity Seminar May 7 2014 14

Distillation, Stripping, Water Extraction Columns
“Precision cleaned”, then assembled in low-Rn cleanroom
Stripping column
Low radon clean room
Radio-pure (LAK) nitrogen (Heidelberg)
Simgen, Heusser, Zusel,
Appl. Rad. 61, 213 (2004)
4/10/2014 Stanford Univsersity Seminar May 7 2014
Distillation column
Assembly of distillation & stripping columns
In Princeton Low-Radon Cleanroom.
15

• PPO was found to have a high level of K that could be reduced by water extraction.
• A concentrated solution of
PPO in PC was purified by water extraction then distillation.
• Recent studies show that
LNGS de-ionized water has significant 210 Pb and 210 Po.
• The studies also indicate that simple distillation done is ineffective for removing 210 Po.
Afer water extraction, the PPO purification plant used the above simple evaporator to distill the PPO-PC “master solution”.
Water extraction may have introducedf 210 Po that was not removed by distillation.
4/10/2014 Stanford Univsersity Seminar May 7 2014 16

Borexino Energy Spectra
PRL 107 141302 (2011)
Data are based on 740.7 live days between
May 16, 2007 and May 8, 2010.
Prominent backgrounds are:
210 Po 210 Bi 85 Kr, 11 C & 14 C (not shown)
CNO obscured mainly by 210 Bi due to similar shape.
The 210 Po alpha rate was high, but rejected by alpha/beta pulse shape discrimination.
The pep was measured by applying cuts to reduce the 11 C. (muon track, neutron, other) pp
CNO
7 7Be pep
4/10/2014 Stanford Univsersity Seminar May 7 2014 17

Solar Neutrinos
✓ 7 Be 46.0 cpd/100t ± 5%. PRL 2011
✓ 8 B (> 3 MeV) 0.22 cpd/100t
✓ Pep 3.1 cpd/100t
± 19% PRD 2010
± 22% PRL 2012
✓ CNO limit
✓ 7 Be day/night asy.
< 7.9 cpd/100t
A = 0.001 ± 0.014
✓ 7 Be annual modukation
PRL
PLB
PLB
2012
2012
2012
Geo-neutrinos
 Geo-neutrinos 14.3 ± 3.4 eV/(613 t-yr) PLB 2013
Rare Processes
 Test of Pauli Exclusion Principle in Nuclei
 Solar axion upper limit
PRC 2010
PRD 2012
4/10/2014 Stanford Univsersity Seminar May 7 2014 18

11 C
210 Po
210 Bi
85 Kr
CNO
4/10/2014 Stanford Univsersity Seminar May 7 2014 19

7
R

46

1 .
5
( stat )

1 .
5

1 .
6
( syst )
R no oscillatio n

74

5 .
2 cpd / 100 t cpd / 100 t
5 s evidence of oscillation
 n e flux reduction 0.62 +- 0.05
 electron neutrino survival probability 0.51 +- 0.07
• Search for a day night effect:
• not expected for 7 Be in the LMA-MSW model
• Large effect expected in the “LOW” solution (excluded by solar exp+Kamland)
A
DN

( N
N

D

D ) / 2

0 .
001

0 .
012 ( stat )

0 .
007 ( sys )
G. Bellini et al., Borexino Collaboration, Phys. Lett. B707 (2012) 22.
4/10/2014 Stanford Univsersity Seminar May 7 2014 20

MSW theory of neutrino oscillations in the Sun predicts a transition in survival probability of n e from vacuum oscillations below 2 MeV to matter effect oscillation at higher energy.
The effect has been observed
Reducing the uncertainty in pep neutrino rate will improve the confirmation of this MSW feature, and tighten constraints on non-standard interactions.
pp
7 7 Be pep
Uncertainty too large.
8 B
Figure from Haxton, et al
Ann. Rev. Astron. Astrophys. 2013
4/10/2014 Stanford Univsersity Seminar May 7 2014 21

• Technical goals:
– Reduce scintillator backgrounds with loop purification by water extraction. ☐
• 210 Bi ( 210 Pb) ☐
• 85 Kr by nitrogen stripping ☐
• Measurement goals:
– pp neutrino measurement ☐
– Improve pep, 7 Be, 8 B measurement ☐
– CNO neutrinos detection or lower limit (?) ☐
4/10/2014 Stanford Univsersity Seminar May 7 2014 22

• Lower backgrounds essential for Phase 2 solar program.
– 210 Bi obscures CNO and pep neutrinos.
– 85 Kr interferes with 7 Be neutrinos
• Purification of the scintillator by “water extraction” and “nitrogen stripping” was carried out 2010-2011.
– Backgrounds were reduced significantly.
• Lower background is still necessary.
– Refinements in water extraction are being developed
4/10/2014 Stanford Univsersity Seminar May 7 2014 23

• “Loop” purification is achieved by draining fluid from bottom of vessel, passing it through purification system, and returning to the top.
• Processes available are:
– Water extraction or distillation
– Nitrogen stripping ( 85 Kr)
4/10/2014 Stanford Univsersity Seminar May 7 2014 24

• Contacting high purity water with scintillator can remove radioactive impurities from the scintillator.
– Works best if impurities have higher affinity for water, e.g., polar species, but can also be effective if not.
• For water extraction, we use LNGS ground water purified by the following:
– Reverse osmosis and Ion-exchange (deionization water plant)
– Single stage evaporator distillation.
• Ground water has high levels of radioactivity.
– ICPMS studies show that 238 U, 232 Th, 40 K are removed effectively by de-ionization.
– 222 Rn is high (10,000 Bq/ton), but can removed by de-gasification with N
2
.
– Radon daughters 210 Pb, 210 Bi, and 210 Po studied recently (see below).
The water extraction system.
N
2 stripping column used in series
4/10/2014 Stanford Univsersity Seminar May 7 2014 25

•
•
• 85 Kr:
• 238 U ( 226 Ra):
214 Bi214 Po
• 232 Th:
212 Bi212 Po
210
210
Bi:
Po:
30 cpd/100t →
[(530 ± 50) →
< 5 cpd/100t
< 8 x 10 -20 gU/g
Reduction factor > 77 (< 0.8 count/100t/yr).
[(3.8 ± 0.8 → < 1.0] x 10 -18 gTh/g
Reduction factor > 3. ( < 0.8 count./100t/yr)
70 cpd/100t → 20 ± 5 cpd/100t
Essentially not reduced!! WHY???
Rates before purifcation are based on 153.6 ton-yr exposure taken in
740.7 d between May 16, 2007 and May 8, 2010. See Borexino Coll. arXiv 1308.0443v1.
4/10/2014 Stanford Univsersity Seminar May 7 2014 26

• Technical goals:
– Reduce scintillator backgrounds with loop purification ≈ ✓
• 210 Bi ( 210 Pb) ≈ ✓ (incomplete)
• 85 Kr by nitrogen stripping 
• Measurement goals
– pp neutrino measurement (to be published) 
– Improve pep, 7 Be, 8 B measurement (ongoing) ☐
– CNO neutrinos detection or new limit (on-going) ☐
4/10/2014 Stanford Univsersity Seminar May 7 2014 27

• The low levels of U and Th (< 1 c/y/100ton) are quite likely the lowest levels ever achieved in a counting experiment.
• Results show promise for accurate pep neutrino measurement with future deeper detectors, free of cosmogenic 11 C.
• Also promising is use of liquid scintillators for neutrino-less double beta decay: Kamland-Zen &
SNO+.
– Background from 2448 keV 214 Bi and 2614 keV 208 Tl gamma rays could be absent for multi-ton detectors.
4/10/2014 Stanford Univsersity Seminar May 7 2014 28

2
Region sensitive to CNO & 210 Bi
Before re-purification 2008-2010
Rates in parentheses are in cpd/100t.
Without 11 C cuts. See arXiv1308.0443v1.
After re-purification 2012-2013
(with 11 C cuts)
4/10/2014 Stanford Univsersity Seminar May 7 2014 29

210
210
• Why did we have so much 210 Po at start of Borexino in
2007 compared to 210 Bi ( 210 Pb)? They should be in secular equilibrium, but they are far from it.
– 210 Bi initial rate: ~15 cpd/100t (increased later to ~40 )
– 210 Po initial rate: ~8000 cpd/100t (decaying with t
1/2
= 138d)
• Why did water extraction work so well for U and Th, but not as well for Pb, and not at all for Po?”
– We have reduced 210 Po with water extraction in 1996 CTF studies, but with 210 Po at higher concentrations.
• Is the water used for water extraction really free of radioactivity, especially radon daughters?
4/10/2014 Stanford Univsersity Seminar May 7 2014 30

238
232
A lower 210 Bi rate compared to 210 Po implies a chemical separation process that removes Pb more efficiently than Po.
4/10/2014 Stanford Univsersity Seminar May 7 2014 31

• The puzzling results on 210 Po and 210 Bi ( 210 Pb) motivated a series of off-line studies at LNGS and
Princeton for the past 3 years, eventually producing surprising results for radioactivity in “purified water”.
• Due to limited time, we won’t discuss details today, but summarize results and discuss impact on further purification.
• Designs for an upgrade of the water extraction system will then be described.
4/10/2014 Stanford Univsersity Seminar May 7 2014 32

• The water is obtained from natural drainage from the rock into the laboratory.
• The water has 222 Rn at a concentration of 10,000 Bq/m 3 , typical for many sites.
– Typical concentrations of 210 Pb and 210 Po are ~ 10 Bq/m 3
• Water purification system based on reverse osmosis and ion exchange is used to remove radioactive impurities.
– Nitrogen stripping removes 222 Rn.
– [U], [Th] are very low: < 10 -15 g/g by ICPMS.
• The water in then distilled before being used for water extraction.
4/10/2014 Stanford Univsersity Seminar May 7 2014 33

And Radio-purity of Water
• Water extraction of scintillator was very effective in removing the radioactive impurites in the U and Th chains.
– This indicates that the water itself had low levels of these elements.
– 238 U: Measured by 214 Bi214 Po delayed coincidences.
• The results indicate low concentration in water of elements above 214 Bi in the U decay chain, in particular 226 Ra, which feeds the A=214 isotopes.
– As an alkaline earth, 226 Ra should be removed well by ion exchange processes. As example, Mg reduction is ~10 6.
4/10/2014 Stanford Univsersity Seminar May 7 2014 34

1. Water purification systems based on reverse osmosis and ion exchange do a poor job of removing 210 Pb and 210 Po.
– Measured reductions of ~1000 (Pb) and ~10 (Po).
• Methods: ICPMS on 208 Pb and alpha counting on 210 Po.
– Results imply activities > mBq/m 3 before distillation.
2. Polonium in ground water is processed biologically by microorganism into a volatile dimethyl polonium compound.
– Recent publications were found by student Brooke Russell.
– With an estimated boiling point of 138 C, 210 Po is difficult to remove from water by single stage evaporator distillation.
4/10/2014 Stanford Univsersity Seminar May 7 2014 35

210
• Small-scale fractional distillation system designed and tested at
Princeton to remove dimethyl polonium from well water.
– 6-foot tall column with structured packing and high reflux designed for x1000 reduction.
• Princeton well water has x5 more
210 Po than LNGS.
• Results:
– Operated as simple evaporator, high 210 Po in product.
– Operated with high-reflux column, no 210 Po in product.
– Reduction factor > 300.
Prototype distillation system with two 6-foot columns: 2 l/hr capacity.
Students : B.Russell, C. Aurup, W. Taylor
4/10/2014 Stanford Univsersity Seminar May 7 2014 36

Brooke Russell ‘11 found papers on volatile Po
& directed team of rising seniors and technical specialist Allan to develop and test a new distillation system that worked beautifully to remove 210 Po.
Brooke Russell
Allan Nelson
Will Taylor
For a Professor it doesn’t get better than that!
4/10/2014 Stanford Univsersity Seminar May 7 2014
Christian Aurup
37

• Due to limited effectiveness of reverse osmosis, ion exchange, and distillation, the water used for water extraction had relatively high levels of
210 Pb and 210 Po.
• Distillation removes most of the 210 Pb, but leaves significant 210 Po in the water.
4/10/2014 Stanford Univsersity Seminar May 7 2014 38

210
210
1. The initial high level of 210 Po with low level of
210 Pb( 210 Bi) is due to water extraction of PPO.
– Water extraction and distillation were used to purify the
PPO master solution
• Water extraction introduced 210 Po and 210 Pb.
• Distillation removed 210 Pb, but not 210 Po due to volatile dimethyl polonium compound.
• Predicted 210 Po rate agrees roughly with what we observed.
2. Failure of water extraction to reduce 210 Po in scintillator.
– This is due to 210 Po in water at roughly same level as lowest concentration in scintillator achieved.
4/10/2014 Stanford Univsersity Seminar May 7 2014 39

• Reduce 210 Pb and 210 Po radioactivity in water used for water extraction.
– Fractional distillation system was designed and tested for removing 210 Po from well water.
• Upgrades to existing system have been designed and changes are modest.
– Two new columns will be added to existing evaporators.
4/10/2014 Stanford Univsersity Seminar May 7 2014 40

Current & Proposed Upgrade with 2 Fractional Distillation Columns
Proposed
System
4/10/2014
Make-up water system not shown.
Stanford Univsersity Seminar May 7 2014
Make-up water system not shown.
41

• The Solar Metallicity:
– Measurement of CNO neutrinos (±10%??)
– Accurate measurement of 7 Be neutrinos (±3%?)
– The 750 keV mono-energetic 51 Cr neutrino source for SOX will provide good energy calibration for measurement of the
862 keV 7 Be neutrino, reducing a major systematic error.
• Sterile Neutrino Searches: The “SOX” Experiment.
– Intense neutrino sources will be placed in tunnel under
Borexino to search for short-baseline neutrino oscillations.
– 10 MCi 51 Cr 750 keV mono-energetic neutrinos and 100 kCi
144 Ce anti-neutrino source (Emax = 3 MeV) to be used.
4/10/2014 Stanford Univsersity Seminar May 7 2014 42

Note 10% difference in 7 Be rates
Note ~30% difference but 15% errors.
Accurate measurements of CNO (+/-10%) and 7 Be(+/- 3%) neutrinos, coupled with improved nuclear reaction cross sections, especially 3 He( a,g ) 7 Be (LUNA), will probe metallicity in Sun’s core.
4/10/2014
Table from Haxton, et al
Stanford Univsersity Seminar May 7 2014
Ann. Rev. Astron. Astrophys. 2013

• In 1998 the metallicity (abundance of elements heavier than 4 He) determined from line spectra in Sun’s atmosphere agreed well with other data.
– Standard solar model based on uniform composition.
– Helioseismology data
– Solar neutrino data ( 8 B by SNO)
• Improvements were made in the analysis of solar atmospheric spectra over next 10 years (3D model, etc.)
– A 2009 assessment of data resulted in a lower metallicity.
• Z /X = metal/hydrogen ratio = 0.024 (GS98)  0.018 (AGSS09).
– The new results are in conflict with helio-seismic data that probe the
Sun’s composition at greater depths.
• This is a serious problem for stellar models because it implies that the chemical composition is not uniform.
– New models involving accretion and planetary formation by Haxton et. al.
4/10/2014 Stanford Univsersity Seminar May 7 2014 44

e
4/10/2014 Stanford Univsersity Seminar May 7 2014 45

51
144
144
4/10/2014 Stanford Univsersity Seminar May 7 2014 46

• Continuing reduction of backgrounds toward ~1 event/100 tons/yr opens new opportunities for future research.
– New solar neutrino measurements will improve oscillation data in the vacuum to matter-effect transition.
– pep neutrinos will be improved, but would be easier to measure with high accuracy at SNOlab.
– Precision 7 Be (3%) and CNO (10%) may resolve the metallicity problem. Deeper site is better for CNO.
– Low backgrounds demonstrate possibilities in other rare-event experiments, such as searches for dark matter or neutrino-less double beta decay.
• Ultra-low background is possible, but requires persistence, investment in big detectors, and time to acquire data.
– Demonstrating 1 c/yr/100t requires a 100-ton detector and >1 yr
4/10/2014 Stanford Univsersity Seminar May 7 2014 47

4/10/2014 Stanford Univsersity Seminar May 7 2014 48