Getting the first 7Be n detection: scintillator purification, detector response and data analysis in Borexino

1
Getting the first 7Be n detectionscintillator
purification, detector response and data analysis
in Borexino

• Marco Pallavicini
• Università di Genova INFN
• On behalf of the Borexino Collaboration

2
Contents

• Physics goals, detector design, construction
  filling
• Design guidelines
• Radiopurity issues
• Plants and Filling
• Detector response Data analysis
• Event selection
• Detector response
• Background content
• Spectral fits

3
Borexino Collaboration
Genova
Princeton University
APC Paris
Virginia Tech. University
Munich (Germany)
Dubna JINR (Russia)
Kurchatov Institute (Russia)
Jagiellonian U. Cracow (Poland)
Heidelberg (Germany)
4
Abruzzo, Italy 120 Km from Rome
Laboratori Nazionali del Gran Sasso Assergi
(AQ) Italy 3500 m.w.e
External Labs
Borexino Detector and Plants
5
Detection principles and n signature

• Borexino detects solar n via their elastic
  scattering off electrons in a volume of highly
  purified liquid scintillator
• Mono-energetic 0.862 MeV 7Be n are the main
  target, and the only considered so far
• Mono-energetic pep n , CNO n and possibly pp n
  will be studied in the future
• Detection via scintillation light
• Very low energy threshold
• Good position reconstruction
• Good energy resolution
• BUT
• No direction measurement
• The n induced events cant be distinguished from
  other b events due to natural radioactivity
• Extreme radiopurity of the scintillator is a
  must!

Typical n rate (SSMLMABorexino)
6
Detector layout and main features
Stainless Steel Sphere 2212 PMTs 1350 m3
Scintillator 270 t PCPPO in a 150 mm thick
nylon vessel
Nylon vessels Inner 4.25 m Outer 5.50 m

• Water Tank
• g and n shield
• m water C detector
• 208 PMTs in water
• 2100 m3

Carbon steel plates
20 legs
7
15 years of work in three slides (I)

• Detector Plants
• All materials carefully and painfully selected
  for
• Low intrinsic radioactivity
• Low Rn emanation
• Good behaviour in contact with PC
• Pipes, vessels, plants
• electropolished, cleaned with detergent(s),
  pickled and passivated with acids, rinsed with
  ultra-pure water down to class 20-50
• The whole plant is vacuum tight
• Leak requirements lt 10-8 atm/cc/s
• Critical regions (pumps, valves, big flanges,
  small failures) were protected with additional
  nitrogen blanketing

• PMTs (2212)
• Sealing PC and water tolerant
• Low radioactivity glass
• Light cones (Al) for uniform light collection in
  fiducial volume
• Time jitter 1.1 ns (for good spatial resolution,
  mu-metal shielding)
• 384 PMTs with no cones for m id
• Nylon vessels
• Material selection for chemical mechanical
  strength
• Low radioactivity to get lt1 c/d/100 t in FV
• Construction in low 222Rn clean room
• Never exposed to air

8
Picture gallery (I)
Pmt sealing PC Water proof
2000
Nylon vessels installation (2004)
PMT installation in SSS
2002
9
15 years of work in three slides (II)

• Water ( production rate 1.8 m3/h)
• RO, CDI, filters, N2 stripping
• U, Th lt 10-14 g/g
• 222Rn 1 mBq/m3
• 226Ra lt0.8 mBq/m3
• 18.2-18.3 MW/cm typical _at_ 20C
• Scintillator
• IV PCPPO (1.5 g/l)
• OV Buffer PCDMP (5 g/l)
• PC Distillation (all PC)
• 6 stages distillation
• 80 mbar, 90 C
• Vacuum stripping with low Ar-Kr N2
• Humidified with water vapor 60-70

• PPO purification
• PPO is solid.
• A concentrated solution (120 g/l) in PC is done
  first (master solution)
• Master solution was purified with
• Water extraction ( 4 cycles)
• Filtration
• Single step distillation
• N2 stripping with LAKN
• Filling operations
• Purging of the SSS volume with LAKN (early 06)
• Water filling (Aug. 06 ? Nov. 06)
• Replacement of water with PCPPO or PCDMP (Jan.
  07 ? May. 07)
• Mixing online
• DATA TAKING from May 15, 2007

10
Picture gallery (II)
Water Plant
Storage area and Plants
CTF and Plants
11
Low Argon Krypton Nitrogen
Specification 222Rn ?
7 µBq/m3 Ar ? 0.4 ppm
Kr ? 0.2
ppt Expected signal from 39Ar, 85Kr and 222Rn
in the Borexino FV ? 1 cpd (for each isotope)

• LAKN developed for
• IV/OV inflating/flushing
• scintillator purification
• blanketing and cleaning
• Production rate reaches 100 m3/h (STP)

Achieved results
High Purity Nitrogen 222Rn lt 0.3 µBq/m3
222Rn 8 ?Bq/m3 Ar 0.01 ppm
Kr 0.02 ppt
Details discussed by G. Zuzel Low-level
techniques applied in the expe- riments looking
for rare events, Wed. 12.09, Solar Low BG
Techniques.
1 ppb Ar in N2 1.4 nBq/m3 for 39Ar 0.1 ppt Kr
in N2 0.1 µBq/m3 for 85K
12
15 years of work in three slides (III)
RadioIsotope RadioIsotope Concentration or Flux Concentration or Flux Strategy for Reduction Strategy for Reduction
Name Source Typical Required Hardware Software Achieved
m cosmic 200 s-1 m-2 10-10 Underground Cherenkov signal lt10-10
    at sea level   Cherenkov detector PS analysis (overall)
Ext. g rock     Water Tank shielding Fiducial Volume negligible
Int. g PMTs, SSS     Material Selection Fiducial Volume negligible
  Water, Vessels     Clean constr. and handling    
14C Intrinsic PC/PPO 10-12 10-18 Old Oil, check in CTF Threshold cut 10-18
238U Dust 10-5-10-6 g/g lt 10-16 g/g Distillation, Water Extraction   lt 10-17
232Th Organometallic (?) (dust) (in scintillator) Filtration, cleanliness   lt 10-17
7Be Cosmogenic (12C) 3 10-2 Bq/t lt 10-6 Bq/ton Fast procurement, distillation Not yet measurable ?
40K Dust, 2 10-6 g/g lt 10-14 g/g scin. Water Extraction Not yet measurable ?
  PPO (dust) lt 10-11 g/g PPO Distillation    
210Pb Surface contam.     Cleanliness, distillation Not yet measurable ?
  from 222Rn decay       (NOT in eq. with 210Po)  
210Po Surface contam.     Cleanliness, distillation Spectral analysis 60
  from 222Rn decay       a/b stat. subtraction 0.01 c/d/t
222Rn air, emanation from 10 Bq/l (air) lt 1 c/d/100 t Water and PC N2 stripping, Delayed coincidence lt 0.02 c/d/t
  materials, vessels 100 Bq/l (water) (scintillator) cleanliness, material selection    
39Ar Air (nitrogen) 17 mBq/m3 (air) lt 1 c/d/100 t Select vendor, leak tightness Not yet measurable ?
85Kr Air (nitrogen) 1 Bq/m3 in air lt 1 c/d/100 t Select vendor, leak tightness Spectral fit 0.2
      lt0.01 ppt  (learn how to measure it) fast coincidence lt0.35
13
Whats important of previous table

• 238U and 232Th content in the scintillator and in
  the nylon vessels meet specifications or
  sometimes are even below specs
• GOAL lt 10-16 g/g (lt 10 c/d/FV) ACHIEVED lt 10-17
  g/g
• 14C is 10-18 g/g as expected (2.7 10-18 g/g
  measured)
• Muon rejection is fine lt 10-4
• Two main backgrounds are still above specs,
  although are managable
• Off equilibrium 210Po as (no evidence of 210Pb or
  210Bi at that level)
• Some 85Kr contamination, probably due to a small
  air leak during filling

14
Finally, May 15th, 2007
15
Our first result (astro-ph 0708.2251v2)

• We have detected the scattering rate of 7Be solar
  ns on electrons
• 7Be n Rate 47 7STAT
  12SYS c/d/100 t

How did we get here ?
16
The starting point no cut spectrum
14C dominates below 200 KeV
210Po NOT in eq. with 210Pb
Arbitrary units
Mainly external gs and ms
Photoelectrons
Statistics of this plot 1 day
17
m cuts
Outer detector efficiency
Preliminary
m with OD tag

• m are identified by the OD and by the ID
• OD eff 99
• ID analysis based on pulse shape variables
• Deutsch variable ratio between light in the
  concentrator and total light
• Pulse mean time, peak position in time
• Estimated overall rejection factor
• gt 104 (still preliminary)

No OD tag lt 1
A muon in OD
m track
ID efficiency
18
Spectrum after m cut (above 14C)
No cuts
After m cut

• After cuts, m are not a relevant background for
  7Be analysis
• Residual background lt 1 c/d/100 t

19
Position reconstruction

• Position reconstruction algorythms (we have 4
  codes right now)
• time of flight fit to hit time distribution
• developed with MC, tested and validated in CTF
• cross checked and tuned in Borexino with
  214Bi-214Po events and 14C events

z vs Rc scatter plot
Resolution
214Bi-214Po (800 KeV) 142 cm 14C (100
KeV) 414 cm
20
Fiducial volume cut

• External background is large at the periphery of
  the IV
• g from materials that penetrate the buffer
• They are removed by a fiducial volume cut
• R lt 3.276 m (100 t nominal mass)
• Another volumetric cut, z lt 1.8 m, was done to
  remove some Rn events caused by initial
  scintillator termal stabilization

Preliminary
Radial distribution
z vs Rc scatter plot
R2
gauss
FV
21
Spectrum after FV cut
Clear 7Be shoulder
No cuts
After FV cuts
11C
No ms

• External background is the dominant background
  component in NW, except in the 210Po peak region

22
11C and neutrons after muons

• ms may produce 11C by spallation on 12C
• n are also produced 90 of the times
• Only the first neutron after a muon can be
  currently detected
• Work in progress to try to improve this
• Events that occur within 2 ms after a m are
  rejected

Preliminary
Neutron Capture Time
Neutron spatial distribution
t 210 ms
23
Final spectrum after all cuts
Understanding the final spectrum main components
210Po (only, not in eq. with 210Pb!)
14C
85Kr7Be n
11C
Last cut 214Bi-214Po and Rn daughters removal
24
Energy calibration and stability

• We have not calibrated with inserted sources
  (yet)
• Planned for the near future
• So far, energy calibration determined from 14C
  end point spectrum
• Energy stability and resolution monitored with
  210Po a peak
• Difficult to obtain a very precise calibration
  because
• 14C intrinsic spectrum and electron quenching
  factor poorly known

Light yield determined from 14C fit
Light yield monitored with 210Po peak position
25
238U and 232Th content
212Bi-212Po
Assuming secular equilibrium, 232Th and 238U are
measured with the delayed concidences
232Th Events are mainly in the south vessel
surface (probably particulate)
212Bi-212Po
214Bi-214Po
Only 3 bulk candidates
238U lt 2. 10-17 g/g
232Th lt 1. 10-17 g/g
26
a/b discrimination
Full separation at high energy
Small deformation due to average SSS light
reflectivity
a particles
b particles
ns
250-260 pe near the 210Po peak
200-210 pe low energy side of the 210Po peak
2 gaussians fit
2 gaussians fit
a/b Gatti parameter
a/b Gatti parameter
27
7Be signal fit without a/b subtraction

• Strategy
• Fit the shoulder region only
• Use between 14C end point and 210Po peak to limit
  85Kr content
• pep neutrinos fixed at SSM-LMA value
• Fit components
• 7Be n
• 85Kr
• CNO210Bi combined
• very similar in this limited energy region
• Light yield left free

210Po peak not included in this fit
7Be n
CNO 210Bi
85Kr
These bins used to limit 85Kr content in fit
28
7Be signal fit a/b subtraction of 210Po peak
2 gaussians fit

• The large 210Po background is subtracted in the
  following way
• For each energy bin, a fit to the a/b Gatti
  variable is done with two gaussians
• From the fit result, the number of a particles in
  that bin is determined
• This number is subtracted
• The resulting spectrum is fitted in the energy
  range between 270 and 800 KeV
• A small 210Po residual background is allowed in
  the fit
• Results are totally consistent with those
  obtained without the subtraction

a
b
The two analysis yield fully compatible results
29
Comments on errors

• Statistical
• Right now, it includes combined the effect of
  statistics itself, the lack of knowledge of 85Kr
  content, and the lack of a precise energy
  calibration
• These components are left free in the final fit,
  and contribute to the statistical error
• Systematic
• Mostly due to fiducial volume determination
• With 45 days of data taking, and without an
  internal source calibration, we estimate an upper
  limit of 25 for this error
• Can be much improved even without internal
  calibration with more statistics and better
  understanding of the detector response

30
Conclusions

• Borexino has performed the first real time
  detection of sub/MeV solar neutrinos
• Quite surprising even for us, after just two
  months of data
• A clear 7Be neutrino signal is visible after a
  few cuts
• We made no attempt to under-estimate the errors.
• Better results to come in the near future
• The central value is well in agreement with
  MSW/LMA.
• Significant improvements are expected shortly

In memory of Cristina Arpesella, Martin
Deutsch, Burkhard Freudiger, Andrei Martemianov
and Sandro Vitale