Search for neutrino oscillations on a long base-line at the CHOOZ nuclear power station

41 Pages·3.5 MB·English

by  M. Apollonio

EPJ manuscript No.
(will be inserted by the editor)
Search for neutrino oscillations on a long base-line at the
CHOOZ nuclear power station
M. Apollonio3, A. Baldini2, C. Bemporad2, E. Caffau3, F. Cei2, Y. D´eclais5aH. de Kerret6, B. Dieterle8, A. Etenko4,
3
L. Foresti3, J. George8, G. Giannini3, M. Grassi2, Y. Kozlov4, W. Kropp7, D. Kryn6, M. Laiman5, C. E. Lane1,
0
B. Lefi`evre6, I. Machulin4, A. Martemyanov4, V. Martemyanov4, L. Mikaelyan4, D. Nicolo`2, M. Obolensky6,
0
2 R. Pazzi2, G. Pieri2, L. Price7, S. Riley7, R. Reeder8, A. Sabelnikov4, G. Santin3, M. Skorokhvatov4, H. Sobel7,
J. Steele1, R. Steinberg1, S. Sukhotin4, S. Tomshaw1, D. Veron7, V. Vyrodov7
n
a 1 DrexelUniversity
J 2 INFN and University of Pisa

3 3 INFN and University of Trieste
1 4 Kurchatov Institute
5 LAPP-IN2P3-CNRSAnnecy
1 6 PCC-IN2P3-CNRS Coll`ege de France
v 7 University of California, Irvine
7
8 University of New Mexico, Albuquerque
1
0
1 Received: 18 November2002 / Revised version:
0
3 Abstract. ThisfinalarticleabouttheCHOOZexperimentpresentsacompletedescriptionoftheνe source
0
and detector, the calibration methods and stability checks, the event reconstruction procedures and the
/
x Monte Carlo simulation. The data analysis, systematic effects and the methods used to reach our conclu-
e sions are fully discussed. Some new remarks are presented on the deduction of the confidence limits and
- on thecorrect treatment of systematic errors.
p
e
h
: 1 Introduction δm2 = m2−m2 as follows:
v 2 1
i (cid:12) (cid:12)
X (cid:12) (cid:12) 1.27δm2(eV2)L(m)
Neutrino oscillation experiments are sensitive probes of P(ν →ν )=1−sin22θ sin2 .
ar the possible existence of a finite neutrino mass and pro- e e (cid:18) Eν(MeV) (cid:19)
vide a way to study physics beyond the Standard Model (1)
of electroweak interactions [1]. In fact, lepton flavour vi- Atmospheric neutrino results give a δm2 from 10−2 to
olation and the existence of nonzero neutrino masses can 10−3eV2. Long base-line (L-B) reactor neutrino experi-
give rise to neutrino oscillations, as first pointed out by ments [15] have been one of the most powerful ways to
Pontecorvo[2,3] and Maki et al. [4]. Severalexperiments, investigate νe → νµ neutrino oscillations (or, more gen-
studyingsolar[5,6,7,8]oratmosphericneutrinos[9,10,11, erally, νe → νx oscillations). The CHOOZ [16,17] and
12,13,14], have measured fluxes consistently lower than PALO VERDE[18] experiments utilized the high inten-
expectations. This can be interpreted as due to various sity and purity of the reactor core flux to achieve high
forms of neutrino oscillations. In particular the so-called sensitivity.
“atmospheric neutrino anomaly” is the observation of a TheCHOOZexperimenthadanaveragevalueofL/E ∼
ν /ν ratio which is roughly one half of what expected 300 (L ∼ 1km, E ∼ 3MeV), an intense and nearly pure
µ e
and its possible explanation might be due to either os- neutrino flavour composition (∼ 100%νe) and an inten-
cillation of ν ↔ ν or to ν ↔ ν . In a model with two sity known to better than 2%. It could therefore make a
µ τ µ e
neutrinoeigenstatesofmassm andm whichmixtoform definitive contribution to solving the problem of the at-
1 2
twoflavourstates,apurebeamofelectron–flavouredneu- mospheric neutrino anomaly. CHOOZ removedthe possi-
trinos has a survival probability which oscillates due to bility of explaining the atmospheric neutrino anomaly by
the m1−m2 mass difference.For asingle neutrinoenergy νe ↔νµ oscillations and Super–Kamiokande showed that
Eν(MeV) and a distance from the source L (meters), the νµ ↔ντ caused the effect [19].
survivalprobabilitycanbe written intermsofthe mixing Theexperimentwasdesignedtodetectreactorν ’svia
e
parametersin22θ andthedifferenceofthesquaredmasses the inverse β-decay reaction
a Present address: IPNL-IN2P3-CNRSLyon νe+p→e++n (2)
2
Thesignatureisadelayedcoincidencebetweentheprompt since the ν emissionrate is lowerthan the one of ν by a
e e
e+ signal (boosted by the two 511− keV annihilation γ factor >105 and may therefore be discarded [20].
rays)andthe signalfromthe neutroncapture.Thetarget
material is a Hydrogen-rich (free protons) paraffin–based
liquidscintillatorloadedwithGadolinium,whichwascho- 2.1 Description
sen due to its large neutron capture cross section and to
the high γ-ray energy released after n-capture (∼8MeV, The CHOOZ nuclear power plant is located in the village
well above the natural radioactivity). of the same name, standing on the banks of the River
Inthisfinalpaperwepresentacompletedescriptionof Meuse in the north of France close to the border with
the experiment, calibration methods and stability checks, Belgium.
eventreconstructionproceduresandtheMonteCarlosim- Thepowerplantconsistsoftwotwinpressurized-water
ulation. (PWR) reactors belonging to a recently developed gener-
Threedifferentanalyses,whichhavealreadybeenpublisheadt[i1o7n],(N4) in France. The main innovation consists in an
are more extensively discussed in this paper. The main improved power yield (4.25 GWth, 1.45 GWe at full op-
one is based on all the available information: the mea- erating conditions), larger than any other PWR reactor.
sured number of positron events as a function of energy, The first reactorreachedfull power in May 1997,the sec-
separately obtained from each reactor. It uses the two ond in August 1997. Tab. 1 summarizes data acquisition
spectral shapes, as well as the absolute normalizations. (from April 7, 1997 to July 20, 1998). The schedule was
The second result is based only on the comparison of the
positronspectrafromthe two,different-distancereactors.
Table 1. Summary of the CHOOZ data acquisition cycle be-
Thisanalysisislargelyunaffectedbytheabsolutevalueof
tween April 1997 and July 1998.
theν flux,thecrosssection,thenumberoftargetprotons
e
andthedetectorefficiencies,andisthereforedominatedby
statistical errors. The sensitivity in this case is limited to Time (h) Wdt (GWh)
δm2∼>2·10−3eV2 due to the small distance, ∆L=116.7 Run 8761.7 R
m, between the reactors. The explored (δm2,sin22θ) pa- Livetime 8209.3
Dead time 552.4
rameter space still matches well the region of the atmo-
Reactor 1 only ON 2058.0 8295
spheric neutrino anomaly. The third analysis is similar to
Reactor 2 only ON 1187.8 4136
the first, but does not include the knowledge of the abso-
Reactors 1 & 2 ON 1543.1 8841
lute normalizations.
Reactors 1 & 2 OFF 3420.4
Finally, some new remarks are presented, concerning
the deduction of the confidence limits and the correct
treatment of the systematic errors.
quitesuitableforseparatingthe individualreactorcontri-
butionsandfordeterminingthereactor-OFFbackground.
2 The νe source and the expected interaction The core of both reactors consists of an assembly of
205fuelelements bound to the socketplate of the reactor
rate
vessel.The vesselis filledwithpressurizedwater(p=155
bars) at a temperature ranging from 280◦C at the en-
Nuclear reactors were the first sources used to search for trance to about 320◦C at the exit. The water,which acts
neutrino oscillations and, in general, for systematic stud- as a neutron moderator and cooling element, circulates
ies of neutrino properties. In past experiments (i.e. up to throughfour independent systems. Eachof these involves
the1980s),theknowledgeofthereactorneutrinofluxand a primary pump and a steam generator comprising 5416
spectrum(≈10%accurateatthattime)limitedthesensi- tubes immersed in the water of a secondary loop at a
tivity to oscillations. Oscillation tests were performed by much lower pressure (56 bars) than primary loop pres-
comparing the neutrino event rate at different distances sure. As soon as the primary water passes through these
(up to ∼ 100m) from the reactor. This approach elimi- tubes, the secondary water is vaporized and the steam
nated the systematic uncertainty associated with the ab- producedtravels to the turbine-alternatorunit connected
solute neutrino flux, often much greater than the experi- to the electric system [21].
ment statistical error. Since then, our knowledge base of The water in the primary loop is slightly doped by
fissionreactors,inparticularofpressurizedwaterreactors Boron, a strong absorber of thermal neutrons. Boron ac-
(PWR), has much improved,thanks to the previous reac- tivityyieldsinformationonthetrendofthechainreaction
tor experiments and to direct studies of fission reactions in the core.
on several elements. Present uncertainty on the neutrino
flux and spectrum induces systematic errors on the event
rate which are lower than 3%. It is therefore possible to 2.2 Reactor power monitor
performexperiments,atone fixeddistance froma reactor
only, relying on the adequate knowledge of the neutrino TwomethodsweredevelopedbytheE.D.F.technicalstaff
source. The reactor ν flux is perfectly isotropic and es- to monitor the total thermal power produced by each re-
e
sentiallywithnocontaminationfromotherneutrinotypes, actor. The first one is based on the heat balance of the
2. The νe source and the expected interaction rate 3
steam generators. Measurements are made of parameters are changed at the end of each cycle. The remainder is
(water flow rate, vapour pressure and temperature in the displaced towards the centre and new fuel elements are
secondary loop) needed to determine the amount of heat arrangedintheouterpartofthecore,soastogetthefuel
suppliedtothesteamgenerators.Theresultingvaluesare burning as uniformly as possible. For the start-up of the
periodicallyavailableonaspecialcomputernetworkwhich CHOOZ reactors, brand-new fuel rods were used. In or-
recordsthe data onanExcelfile. The overallprecisionon dertoreproducethebehaviourofreactorsatequilibrium,
the calculated thermal power is claimed to be 0.6%. where partially burned-up fuel is used, 137 less enriched
The second set of thermal power data is provided by elements are located at the centre (68 at 2.4% and 69 at
the external neutron flux measurement. This flux is ex- 1.8%,whilethestandardenrichmentis3.1%).Aschematic
pected to be directly proportional to the fission rate in- map of the reactor core is drawn in Fig. 2.
side the core and, as a consequence, to the released ther-
mal power. For each reactor, four different neutron de-
tectors (one proportional counters plus three ionization Y
214 mm
chambers) are located at the opposite sides of the reac-
8 20 8
tor vessel. This method is less precise than the other one
16 20 20 16
because of the spread in the energy release per fission of
20 12 24 12 20
the different fissile isotopes. Accuracy is estimated to be 20 12 16 16 12 20
about 1.5%. However, this method has the advantage of 16 12 12 16 12 12 16
operating continuously and is used simultaneously as a 8 12 12 12 12 12 12 8
20 16 12 16 12 16 20
power monitor and sensor for the reactor safety system.
20 24 16 16 16 16 24 20 X
Neutrondetectoroutputsarefedtothecomputernetwork
20 16 12 16 12 16 20
and written twice a day (or more frequently if the power 8 12 12 12 12 12 12 8
rampsupordown)to anotherExcelfile.The computeris 16 12 12 16 12 12 16
alsoprogrammedtodriveadirectcurrentgeneratorwhose 20 12 16 16 12 20
20 12 24 12 20
amplitude is proportional to the neutron detection rate.
16 20 20 16
SuchasignalwasmeasuredintheCHOOZexperimentby
8 20 8
aCAMACvoltmeteracrossa5Ωresistor(connectedinse-
ries to each loop). This provided a direct, instantaneous
information about the thermal power for both reactors. Region 1: 1.8 % enrichment
The voltmeter calibration is shown in Fig. 1, where ther-
Region 2: 2.4 % enrichment
mal power values, provided by E.D.F., are plotted versus
thecorrespondingvoltmeterdata.Thecalibrationparam- Region 3: 3.1 % enrichment
eters obtained by a linear fit are also given.
Fig.2. Schematicviewofthefuelrodsinthecoreforthefirst
cycle of the CHOOZ reactors. The number of Boron poison
rods assembled with each fuel element is also indicated.
Reactor 1 Reactor 2
W)th4500 W)th4500
M4000 Wth = p0 + p1*V M4000 Wth = p0 + p1*V
Power (33050000 pp01 == ((-1145.63734.2 – – 0 0.0.90)8 M) MWWthth mV-1 Power (33050000 pp01 == ((-1144.87347.2 – – 0 0.0.80)8 M) MWWthth mV-1 enriAchsesdheolwemneinntsFaigr.e2s,urtrhoeunfudeeldabsysemabnluiemsboefrtohfe1m2.7o%st
2500 2500 Boron-doped steel rods, termed “poison” rods. The rods
absorbthethermalneutronexcess,thusaccumulating11B,
2000 2000
whichhasanegligibleneutronabsorptioncrosssection,in-
1500 1500
side the core. Therefore the neutron absorption power of
1000 1000
the poison rods diminishes with time partly compensat-
500 500
ingthefuelburn-up.Thenumberofrodsperfuelelement
0 100 150 200 250 300 350 400 0 100 150 200 250 300 350 400 varies according to the fuel loading as well as to the el-
voltage (mV) voltage (mV)
ement position in the core. This was taken into account
Fig. 1. Voltmetercalibrationforreactor1(left)and2(right).
when computing fuel evolution.
2.4 Fuel evolution
2.3 Map of the reactor core
The unit used to describe the aging of the fuel at nuclear
Thenuclearfuelconsistsof110Tofuraniumoxidetablets reactors is the MWd/T, which measures the amount of
(diameter = 8.2mm) enriched with 235U and stacked in energy per ton extracted from the nuclear fuel after its
4m long, 1cm wide assemblies. The standard enrichment introduction into the reactor core. This quantity is called
for this type of reactor is 3.1%. Each fuel element con- “burn-up”andiscloselyrelatedtothefissileisotopecom-
tains 264 assemblies. About 1/3 of the 205 fuel elements position of the fuel.
4
For any PWR reactor, the procedure to compute the Table 2. Energyreleaseperfissionofthemainfissileisotopes
evolution of the fuel in the core needs daily information (from ref. [22]).
provided by the reactor technical staff. This includes:
isotope energy (MeV)
– the daily cumulative burn-up, given as 235U 201.7±0.6
238U 205.0±0.9
β(t)≡ 1 tW (t′)dt′, (3) 239Pu 210.0±0.9
MU Z0 th 241Pu 212.4±1.0
W being the thermal power,t the time since the start
of reactor operation and M = 110.26 T the total
U
amount of uranium in the core: 242Pu, etc. amount to less than 0.1% and are therefore
– the burn-up βi and the relative contribution αi to the neglected.
power from the i-th fuel element, at several stages of To obtain the source spectrum, in addition to the av-
the reactor cycle (an example is shown in Fig. 3). eragefission rate N of each of the four isotopes, we need
k
the corresponding differential neutrino yield per fission
S (E ). The next section explains how these spectra are
k ν
b = 1000 MW d/T evaluated.
1 1.12
2111.0104 1 1.15 2.5 The expected neutrino spectra
1100 1100
1 1.15 2 1.22 1 1.19 We usedthe so-called“conversion”approachwhich is the
1100 1200 1200 mostreliableandrecentmethodtodeterminetheνe spec-
2 1.11 1 1.13 2 1.21 1 1.14
trum at reactors. This method utilizes measurements of
1100 1100 1200 1100 theelectronspectrumemittedbyalayeroffissilematerial
1 1.06 2 1.08 1 1.10 2 1.14 1 1.03
activatedbythermalneutrons.Theexperimentalelectron
1000 1100 1100 1100 1000
3 1.05 1 1.02 2 1.09 1 1.02 2 0.92 3 1.11 spectrumisthenconvertedintotheν one[23,24,25].The
e
1000 1000 1100 1000 900 1100 most recent and precise measurements were reported by
1 0.96 2 0.94 1 0.96 3 0.97 3 0.90 3 0.66
Schreckenbachetal.[26,27,28].Inthelattercasethinfoils
2100.0808 3 19.0103 3100.0805 3100.0602 900 700 Contribution to power enrichedwiththemainfissilenuclei(about1mg)wereex-
posed to an intense thermal neutron flux (≈3·1014s−1).
900 1200 900 600 Element burnup
3 0.79 3 0.65 A high-resolution (δp/p=3.5·10−4) β-spectrometer was
900 700 usedtomeasurethemomentumoftheemergingelectrons.
The β− spectrum for each fissile isotope is approxi-
Region index
matedbythesuperpositionofasetofhypotheticalallowed
Fig. 3. Power distribution and burn-up values for the fuel branches with amplitude a and end-point Ei:
i 0
elementsin an octant oftheCHOOZreactor core at acertain
step(β =1000ofthefirstcycle).Thecontributiontothepower S (E )= a Si[E ,Ei,Z(Ei)], (5)
β β i β β 0 0
of each element is normalized to have a mean value equal to
i
one. X
where Si is the spectrum shape of the i-th branch and
β
the summation is overthe brancheswith end-point larger
Thesetwoinputsdeterminethedailyburn-upofeachfuel thanEβ.Thespectrumweaklydependsalsoonthecharge
element. An example of the tables providedby E.D.F. for Z (averagechargeoftheβ-decayingnucleiwithend-point
these quantities is given in Fig. 3. E.D.F. also provides Ei) because of the Coulombinteractionin the final state.
0
another set of tables (at severalburn-up stages) in which Themeasuredelectronspectrumisthenusedtodetermine
the relative power contribution fi from the k-th fissile the set of values {a ,Ei} by means of a fit procedure.
k i 0
isotope for the i-th fuel element is given. Eq. (5), with the introduction of the best-fit parameters,
The number ni of fissions per second for the i-th ele- reproduces the measured spectrum to better than 1%.
k
ment for each isotope k can then be computed with the Theβ−spectrumforeachindividualhypotheticalbranch
equation is then converted into the correlated ν spectrum under
e
α (β)fi(β)W(t) the assumption that both the electron and the antineu-
ni(β)= i k , (4)
k fi(β)E trino share the total available energy E0i. Thus, for each
k k branch with end-point Ei, the probability of emitting an
k 0
X electronwithenergyE isequaltotheprobabilityofhav-
β
where Ek is the energy releaseper fissionfor the k-th iso- ing a νe of energy E0i −Eβ. Inserting the fit parameters
tope,whosevaluesarelistedinTab.2.Addingthecontri- into (5), one obtains
bution of all the fuel elements yields the average number
Nk of fissions per second for the k-th isotope. Contribu- S (E )= a Si[(Ei −E ),Ei,Z(Ei)] (6)
tions from other fissioning isotopes, such as 236U, 240Pu, ν ν i β 0 ν 0 0
i
X
2. The νe source and the expected interaction rate 5
These yields contain a normalization error of 1.9% stem- of characteristics known at a few percent level. The sta-
ming from the error on the neutron flux and from the tistical accuracy obtained in these experiments makes it
absolute calibrationuncertainty of the spectrometer. The possible to use these results to discriminate between the
conversion procedure also introduces a global shape un- existing models of reactor neutrino spectra.
certainty into the neutrino spectrum, beyond the inher- The Bugey 3 collaboration[32] measuredthe positron
ent experimental errors. The main sources of this addi- energy spectrum at 15 and 40m from the reactor core
tional uncertainty, ranging from 1.34% at 3MeV to 9.2% and compared its data with the results of a Monte Carlo
at 8MeV, are the scattering in the nuclear charge distri- simulation using the neutrino spectrum models proposed
bution and the higher-order corrections (higher Coulomb by [27,33,34]. As can be seen in Fig. 5, the data per-
terms and weak magnetism, for which an uncertainty of fectly fit with the measurements made at Institute Laue
the order of the correction term itself was assumed). LangevinatGrenoble(ILL),whereasthereisalowercom-
Thismethodwasappliedtoobtainthe neutrinoyields patibility with other models.
of the 235U, 239Pu and 241Pu fissions. The resulting spec-
tra are presented in Fig. 4. Unfortunately, no experimen-
taldata is availablefor238Uwhichcannotbe fissionedby
thermal neutrons. We must therefore rely on the theoret- C 1.2
ical predictions [29] to estimate the contribution to the M 1.1 Klapdor et al.
νe spectrum by the 238U fissionproducts. Although these data/ 1
predictions are less reliable than direct measurements, it 0.9
shouldbenotedthatthecontributiontothenumberoffis- 0.80 1 2 3 4 5 6 7
sions,due to this isotope, is quite stable andnever higher
than 8%. Thus any possible discrepancy between the pre- 1.2
Tengblad et al.
dictedandtherealspectrumshouldnotleadtosignificant
1
errors.
0.8
0 1 2 3 4 5 6 7
1.2
1.1 Schreckenbach et al.
-1s) 10 1
s
-1V fi 1 00..890 1 2 3 4 5 6 7
Me positron energy (MeV)
s
ount 10-1 Fig. 5. Comparison of Bugey 3 data with three different re-
c actorspectrum models. Theerrorbars includeonly statistical
(
Sn uncertainties. The dashed lines are the quadratic sum of the
10-2 235U quoted error of the models and the error due to the energy
calibration.
239Pu
10-3 241Pu
Animprovedmeasurementoftheintegralneutrinoflux
-4 wasperformedin1993.Themeasurementusedanintegral
10
type detector previously employed at the Rovno nuclear
powerplant[35]andsubsequentlymovedtoBugey[36].In
10-5 that apparatus only neutrons were detected in 3He coun-
0 1 2 3 4 5 6 7 8 9 10
E (MeV) ters,whilepositronswerenot.Theapparatuswasinstalled
n
at 15m from the reactor core. About 3 × 105 neutrino
Fig. 4. Neutrino yield per fission of the listed isotopes, as events were collected so that the reaction rate was deter-
determined by convertingthe measured β spectra [27,28]. minedwith0.67%statisticalaccuracy.Theneutrinoevent
rate,n ,correspondstoacertainaveragefuelcomposition
ν
and is related to the cross section per fission σ and the
f
number of target protons N by
p
1 W
2.6 Systematic uncertainties of the neutrino spectrum th
n = N εσ , (7)
ν 4πR2hE i p f
f
In past experiments at reactors, e.g. at G¨osgen [30] and
Bugey[31],morethan3×104neutrinoeventswererecorded The fuel composition, the thermal power W, the average
atseveralreactor–detectordistancesR.Sincenoevidence energy per fission hE i absorbed in the reactor core and
f
for oscillations was observed and rates were found to be the distance R were provided by the E.D.F.–Bugey tech-
consistent with a 1/R2 law, these experiments can be in- nical staff. The efficiency of the neutron detection ε was
terpreted as a check that a reactor is a neutrino source carefullymeasured;the overallaccuracywasestimated to
6
be 1.4%. The experimental result was then compared to 2.7 Neutrino spectrum time relaxation and residual
the expected neutrino flux, which can be inferred by in- neutrino emission
troducing into (7) the neutrino spectra S obtained at
k
ILL[26,27,28], the cross section for the reaction (2) and Anotherpossiblesourceofuncertaintyoftheneutrinoflux
the reactor parameters in (7). The expected cross section is related to the residual emission due to the β− decay
per fission for reaction (2) is given by of long-livedfission fragments. Taking this further contri-
bution into account, the linear relation (7) between the
∞
σfexp = σ(Eν)S(Eν)dEν promptνe interactionrateandthe currentthermalpower
Z0 ∞ (8) mWinnedoalotnILgeLrwheorledds.erNiveevderatfhteerleasbs,outthe1.5νedesxppecotsruaredteitmere-,
= f σ(E )S (E )dE = f σ
k ν k ν ν k k so that neutrinos from fission fragment decays of longer
Xk Z0 Xk life are not included. The expected neutrino rate based
where f refers to the contribution of the main fissile on this model may thus be underestimated with respect
k
nuclei to the total number of fission, S to their corre- to the experimental data.Fortunately the maximum neu-
k
spondingν spectrumandσ totheircrosssectionperfis- trino energy is above the reaction (2) threshold only in a
e k
sion.The result of the measurementwasσmeas =5.752× fewcasessince,ascanbeexpected,thelongerthelifetime,
f
10−19barns/fission±1.4%,whichperfectlyagreeswiththe the lower the Q-value of the decay1. This effect has been
expectedvalue(σexp =5.824×10−19barns/fission±2.7%), evaluated by using the cumulative yields of the known
f long-lived fission fragments [37]; the results for 235U and
and is twice as accurate as the predictions based on the
239PuaresummarizedinTab.3.Inparticular,thereaction
knowledge of the neutrino spectrum.
We couldthereforeadopttheconversionprocedurefor
the shape of the neutrino spectra but normalize the to-
Table3.Timeevolutionofneutrinospectrarelativetoinfinite
tal cross section per fission to the Bugey measurement,
irradiation time (from [37]).
i.e. , after taking all the different reactors conditions into
account.
Infigure6(left) the ILLcrosssection(8)andthe com- Eν (MeV) 235U 239Pu
bined(ILL+Bugey)crosssectionobtainedfortheCHOOZ 104 s 1.5 d 107 104 s 1.5 d 107
reactors are plotted vs. the reactor burn-up. The average 1.5 0.837 0.946 0.988 0.861 0.949 0.990
ratio of the two curves amounts to 0.987. By combining 2 0.897 0.976 0.992 0.904 0.968 0.986
the uncertainty on the neutrino spectra, on the cross sec- 2.5 0.925 0.981 0.990 0.939 0.975 0.986
tion for the reaction (2) and on the fission contributions 3 0.963 0.997 1.000 0.967 0.989 0.993
f (which are of the order of 5%), we obtained the rel- 3.5 0.967 1.000 1.000 0.979 0.997 1.000
k
ative error on the neutrino detection rate, as a function
of the fuel burn-up. As shown in Fig. 6(right), the aver-
age error decreases from 2.4% (ILL data alone) to 1.6%
(ILL+Bugey). Other minor sources of errors come from cross section σf computed by using (8) is probably lower
than the effective one by ≈ 0.3%. This systematic shift
theresidualneutrinoemissionfromlong-livedfissionfrag-
ments (dealt with in the nextSection). Howeverthe reac- affects the accuracyonthe reactioncrosssection. We will
tor source can be considered to be known at a 2% level. thenassumeanoverallandconservative1.9%uncertainty
on the integral neutrino rate.
2.8 The inverse beta-decay reaction
-19-1cross section (10b fiss)6666....672468 IILLLL+Bugey %relative error ()12..2355 IILLLL+Bugey Tth-heeirtdehaetacetsciottinhoen(2ho)if.ghrTeehasictstcoirsrotashsnetsiemncetouisottrnisnuinoitsatihbselueespnureaorlgclyeysbsraassniengdceeoo:nf
5.8 reactor antineutrinos:
5.6 1 - provides a convenient time correlated pair of positron
5.4 and neutron signals, which allows us to reject most of
0.5
5.2 the background.
50 2000 4000 6000 8000 10000 12000 14000 00 2000 4000 6000 8000 10000 12000 14000
b (MW d/ton) b (MW d/ton) The antineutrino and the positron energy are related by
Fig. 6. Comparison of the combined (ILL+Bugey) reaction
crosssectionwiththeILLcrosssection(left)andtheirrelative Eνe =Ee+ +(Mn−Mp)+O(Eνe/Mn), (9)
error (right) as a function of the first CHOOZ reactor cycle
1 This effect might be much more relevant in the case of
burn-up.
experimentslookingforneutrinoelasticscatteringinteractions
(i.e. , measurement of the neutrino magnetic moment) and
needsa more careful treatment.
3. The experiment 7
where the infinitesimal term corresponds to the neutron
rlcoeowcmosiinla.gnTaahncutcsiunreaauttmeriendaoest.uerTremhmeinenathttiroeonsfhotohfldethfpeoorseintthererognryeeaoncfetritoghnye (ian2l)-- -1-1V fiss) 101
is 1.804MeV, equal to the nucleon mass difference plus Me
the positron mass. In the low energy limit, the cross sec- ounts 10-1
tion for the reaction (2) may be written as a function of (c
Sn
outgoing positron energy as follows: 10-2 b = 0
2π2¯h3
σ(E )= p E (1+δ +δ ) (10) 10-3 b = 7000
e+ m5fτ e+ e+ rad WM
e n
10-4
Thetransitionmatrixelementhasbeenexpressedinterms
ofthefreeneutrondecayphase-spacefactorf =1.71465(15)
[38]andlifetimeτn =(886.7±1.9)s[39].Twohigher-order 10-50 1 2 3 4 5 6 7 8 9 10
correctionterms(bothof1%orderofmagnitude)arealso En (MeV)
included: Fig. 7. Comparison between the neutrino spectrum at the
beginning and duringthe first cycle of the CHOOZ reactors.
(i) a radiative correction of the order of α, including an
internalbremsstrahlungcontribution,whichcanbeap-
proximated by
positrons coming from (2) is essentially the antineutrino
δ (E )=11.7×10−3(E −m )−0.3 (11)
rad e+ e+ e spectrumshifted inenergyand weightedby the crosssec-
tion (10). So, following (7), each positron is assigned a
with the positron energy expressed in MeV [40] and
weight given by
(ii) acorrectionforweakmagnetism,arisingfromthedif-
ference µ = µ −µ = −4.705890(2)µ between the
n p N N
p
anomalous magnetic moment of the neutron and the Se+(Te+)= 4πd2σ(Eν)Sν(Eν), (13)
proton
wheredisthedistancefromtheneutrinoproductionpoint
µλ
δWM(Ee+)=−21+3λ2(Ee+ +∆βpe+)/Mp, (12) tothedetectorandthepositronkineticenergyTe+ isgiven
by (9). Fig. 8 showsthe positronyield obtained by gener-
ating the neutrino spectra drawn in Fig. 7 in one day of
where λ = g /g = 1.2601±0.0025 is the ratio of
A V
datatakingwithbothreactorsatfullpower.Althoughthe
axial-vector and vector coupling constants and ∆ is
generatedneutrinoluminosityisthesame,thedecreaseof
the nucleon mass difference [41].
the positron yield with the reactor operating time is evi-
Theknowledgeofthecrosssectionisthereforemuchmore dent.The evolutionofthe positronspectrummustbe fol-
accurate than the one of the νe spectrum, the major lim- lowedveryaccurately in order to reproducethe hardware
itation being related to the uncertainty on τn (whose rel- threshold effects on the positron detection. As an imme-
ative error is at most 0.3%). diate consequence, also the integral neutrino interaction
rate is expected to vary significantly during the reactor
fuel cycle. A decrease of about 10% has been forecast for
2.9 Simulation of the νe spectrum the cross section per fission (which is linear with the in-
teraction rate, according to (7)) during the first cycle of
The expected neutrino spectrum was obtained by means the CHOOZ reactors, as shown in Fig. 9. The measured
of a Monte Carlo simulation of the reactor core in which neutrino rate as a function of the burn-up will be shown
wefoldedalltheingredientsdescribedintheprevioussec- and compared to the expected behaviour, under the no-
tions, that is: oscillation hypothesis.
– thedailyvariationofthefluxduetothefissileelements
burn-up and to the live time of the apparatus;
– thecontributionoftheindividualfuelelementsaccord- 3 The experiment
ing to their position inside the reactor core;
– the individual contributions of the different fissile ele-
3.1 The site
ments in each fuel element to the neutrino flux.
Fig.7showstheneutrinospectrumobtainedatCHOOZ The detector was located in an underground laboratory
on the start-up day (β = 0) and at an intermediate step about 1km from the neutrino source (see Fig. 10). The
(β = 7000) of the first reactor cycle. Due to the decrease 300MWE rock overburden reduced the external cosmic
of the 235U concentration, a reduction of the neutrino raymuonfluxbyafactorof∼300toavalueof0.4m−2s−1,
interaction rate is observed and a softening of the neu- significantly decreasing the most dangerous background,
trino spectrum is expected. The energy spectrum of the whichiscausedbyfastneutronsproducedbymuon–induced
8
-1KeV) 1.2 NuclearC hPooowze rB Station
0 2 x 4200 MWth
0
-1ay (1 1 bb == 07000
d
nts 0.8
ou n
c
0.6 distance = 1.0 km
0.4
Depth
300 mwe
0.2
neutrino
target
0
0 1 2 3 4 5 6 7 8
e+ kinetic energy (MeV)
Fig. 8. Positron spectra at the startup and during the first
cycle of the CHOOZ reactors at maximum daily neutrino lu- Chooz Underground Neutrino Laboratory
Ardennes, France
minosity.
Fig. 10. Overview of the experiment site with indication of
thesource-detector distance and rock overburden.
-1b fiss) 78
-190 todeterminetheexactlocationoftheundergroundexper-
1
n ( 6 imentwithrespecttothetworeactorsanditsorientation.
ctio All the information obtained in the preliminary studies
oss se 5 wtheeredeutseecdtotrosgimuiudleatiimonp.rovements in the detector and in
r
c 4
The cosmic ray measurements were made for several
dayswithasystemofsixResistivePlateChambers(RPC),
3 235U
eachofarea1×1m2.Thecomparisonoftheexperimental
239Pu
2 and expected angular distributions was fairly good, but
some discrepancies persisted. A further geological study
1 revealedtheexistenceofseveralveryhighdensityrocklay-
223481UPu ers(3.1g/cm3thenormaldensitybeing2.8g/cm3),whose
00 2000 4000 6000 8000 10000 12000 14000 positionsandorientationsfully explainedthe observedef-
b (MW d/ton)
fects. The natural radioactivity spectrum was measured,
at the position where the experiment had to be installed,
Fig. 9. Cross section per fission as a function of the reactor
by a 3×3inches2 NaI crystal. Natural radioactivity was
burn-up.Thecontributionofeachfissileisotopeisalsoshown.
ratherhighandthespectrumshowsthenormalnaturalra-
dioactivitylines but alsosome artificialradioactivitycon-
tributions (see Fig. 11).
nuclearspallationsinthematerialssurroundingthedetec- The magnetic field was measured by a rotating coil,
tor.Thiscosmicrayshieldingwasanimportantfeatureof pressurized-airdevice, and was found to be B =0.178±
k
the CHOOZ site. As shown in Fig. 12, the rock shielding 0.007Gand B =0.388±0.007G.
⊥
preserved the signal to noise ratio of previous reactor ex-
periments,inspite ofa reductionby a factor∼100ofthe
neutrinoflux due tothe largerdistancefromthe reactors. 3.2 The detector
The characteristics of the site were thoroughly assessed
before and after the set–up of the experiment. We mea- Thedetector(Fig.13)wasinstalledinaweldedcylindrical
suredthecosmicrayfluxandangulardistributionandthe steel vessel 5.5m in diameter and 5.5m deep. The inter-
results were comparedwith predictions based onthe rock nal walls of the vessel were painted with high–reflectivity
structureaboveandaroundthesite.Thenaturalradioac- white paint. The vessel was placed in a pit 7m in diame-
tivity background, the radon level in the tunnel and the terand7mdeep.Toprotectthedetectorfromthenatural
intensity and orientation of the local terrestrial magnetic radioactivity of the rock, the steel vessel was surrounded
fieldwerealsomeasured.Ageodesicsurveywasperformed by 75cm of low radioactivity sand (Comblanchien from
3. The experiment 9
RADIOACTIVITY SPECTRUM AT CHOOZ SITE whileitsbehaviourcouldbewellchecked.Sixlaserflashers
214Bi were installed in the three regions together with calibra-
137Cs
-1-1V Kg)NaI105 214Bi40K tTiohnepdiepteescttooralcloouwldthbeeinrterloiadbulcytiosinmouflarateddioabcytitvheesoMurocnetse.
-1NTS (days Ke 110034 214Pb 228Ac60Co 208Tl 214Bi208Tl CtciolalianrTltcoohidrme.eneTttachheregobednet.etwruetegreiinnonothdceoetpnetrcaotiminoepndtwpaaosGsibtdra-olsoneadsdiogendnatllihqgeuendideerlsaacytieendd-
U
CO 214Bi by reaction (2), boosted by the annihilation γ-rays, and
102 thesignalassociatedwiththeγ-rayemissionfollowingthe
neutron capture reaction
10 232Th Family
238U Family n+Gd→Gd⋆ →Gd+ γ (14)
i
Artificial Radioactivity
i
1 X
0 1000 2000 3000 4000
ENERGY (KeV) The choice of a Gd-doping was to maximize the neutron
Fig. 11. Natural radioactivity spectrum recorded by a NaI capture efficiency; Gadolinium has the highest thermal
crystal in thewell hosting thedetector. neutron cross section. Moreover, the large total γ-ray en-
ergy (≈ 8MeV, as shown in Tab. 4) easily discriminates
(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)t(cid:0)(cid:0)h(cid:0)(cid:0)(cid:0)(cid:0)e(cid:0)(cid:0)(cid:0)(cid:0)n(cid:0)(cid:0)(cid:0)e(cid:0)(cid:0)(cid:0)u(cid:0)(cid:0)(cid:0)t(cid:0)(cid:0)r(cid:0)(cid:0)(cid:0)o(cid:0)(cid:0)(cid:0)n(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)c(cid:0)(cid:0)a(cid:0)(cid:0)(cid:0)p(cid:0)(cid:0)(cid:0)(cid:0)t(cid:0)(cid:0)u(cid:0)(cid:0)(cid:0)r(cid:0)(cid:0)e(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)f(cid:0)(cid:0)r(cid:0)(cid:0)o(cid:0)(cid:0)(cid:0)m(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)t(cid:0)(cid:0)(cid:0)h(cid:0)(cid:0)(cid:0)e(cid:0)(cid:0)(cid:0)(cid:0)n(cid:0)(cid:0)(cid:0)(cid:0)a(cid:0)(cid:0)(cid:0)t(cid:0)(cid:0)u(cid:0)(cid:0)(cid:0)r(cid:0)(cid:0)(cid:0)a(cid:0)(cid:0)(cid:0)l(cid:0)(cid:0)(cid:0)r(cid:0)(cid:0)a(cid:0)(cid:0)(cid:0)d(cid:0)(cid:0)(cid:0)(cid:0)i(cid:0)o(cid:0)(cid:0)(cid:0)a(cid:0)(cid:0)(cid:0)c(cid:0)(cid:0)(cid:0)t(cid:0)(cid:0)i(cid:0)(cid:0)v(cid:0)(cid:0)(cid:0)i(cid:0)t(cid:0)(cid:0)(cid:0)y(cid:0)(cid:0),(cid:0)(cid:0)(cid:0)w(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)h(cid:0)(cid:0)(cid:0)o(cid:0)(cid:0)(cid:0)s(cid:0)(cid:0)e(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)(cid:0)
energy does not exceed 3.5MeV.
Table 4. Abundancesandthermalneutron capturecross sec-
tions for theGd isotopes.
Gd iEγi Abundance Cross section Relative
isotope (KeV) (%) (barns) intensity
P
152 6247 0.20 735 3·10−5
154 6438 2.18 85 3.8·10−5
155 8536 14.80 60900 0.1848
156 6360 20.47 1.50 6·10−6
157 7937 15.65 254000 0.8151
158 5942 24.84 2.20 1.1·10−5
Fig. 12. Cosmic muon flux compared to the neutrino flux at 160 5635 21.86 0.77 3·10−6
the different underground experimental sites. In the CHOOZ
case the lower neutrino flux is compensated by the reduction
of the muon flux.
Region II was filled with an undoped high-flash point
liquid scintillator. It provided a high-efficiency contain-
mentofthee.m.energydeposit;thiswashigherthan99%
Burgundy in France) and covered by 14cm of cast iron. for positrons from ν -interactions in Region I. The con-
e
The detector comprised three concentric regions: tainment of the γ-rays due to the neutron capture on Gd
was(onaverage)slightlylowerthan95%foranenergyde-
– a central 5–ton target in a transparent Plexiglas con-
posit E >6MeV. The intermediate volume was bounded
tainer (total mass = 117 kg) filled with a 0.09% Gd–
by the “geode”, an opaque plastic structure serving as a
loaded scintillator (“Region I”);
support for the 192 inward-looking photomultiplier tubes
– an intermediate 17–ton region (70cm thick) equipped
(PMT from now on).
with 192 eight–inch PMT’s (15% surface coverage, ∼
The outer volume, also filled with the undoped scin-
130 photoelectrons/MeV), used to protect the target
tillator of Region II, was the “Veto” region (Region III).
from PMT radioactivity and to contain the gamma
An additional 48 PMT’s, arranged in two circular rings
rays from neutron capture (“Region II”);
located at the top and the bottom of the main tank, de-
– an outer 90–ton optically separated active cosmic–ray
tectedthescintillationlightassociatedwiththrough-going
muonvetoshield(80cmthick)equippedwithtworings
cosmicmuons.TheVeto signalwasusedtotagandreject
of 24 eight–inch PMT’s (“Region III”).
thismajorbackgroundsource.Theouterscintillatorlayer
The apparatus was conceived as a liquid scintillator low wasalsothickenoughtoshieldtheneutrinotargetagainst
energy, high-resolution calorimeter. The detector geome- the natural radioactivity from the surrounding materials.
try (a centralvolume of scintillatorsurroundedby photo- The inner detectorvolumewasseparatedfromRegion
multipliers) is common to the Borexino, LSND and SNO IIbyatransparent8mm-thickvessel,averticalcylindrical
detectors. The detector was simple and easily calibrated, surface closed by two hemispherical end-caps. The outer
10
radius of the cylinder and of the end-caps was 90cm, the 3.3 The liquid scintillators
height of the cylinder was 100cm; the inner volume was
5.555m3, while the mass was 150kg (empty). The vessel About 5 tons of Gd-loaded scintillator and 107 tons of
was made of an acrylic polymer (Altuglass), chosen for unloaded scintillator were used in the experiment. The
its excellent optical and mechanical properties and for its mainpropertiesofthetwoscintillatorsarelistedinTab.5.
chemical resistance to aromatic compounds in the scin-
tillator. The upper part of the vessel was fitted with a
chimney (diameter φ = 70mm) to allow passage of fill-
ing pipes, calibration sources, temperature and pressure Table5. Mainpropertiesoftheliquidscintillatorsusedinthe
experiment.
sensors.
Thegeodehadthesameshapeofthetargetvessel,but Gd-loaded unloaded
alargersize;thecylinderheightwasthesame,whereasthe Chemical content:
outer radius was 160cm. The volume between the geode basic Norpar-15 Mineral oil
andthe targetwas19.6m3.The geodesurface (a drawing (50% vol.) (92.8% vol.)
of which is shown in Fig. 13) had a total area of 42m2 aromatics, alcohols IPB+hexanol IPB
segmented into 32 panels; each panel was equipped with (50% vol.) (7.2% vol.)
6 8′′ PMT’s detecting the scintillation light produced in wavelength shifters p-PTP+bis-MSB PPO + DPA
RegionsIandII.TheglobalPMTcoveragewasthen15%. (1 g/l) (1.5 g/l)
Unliketheacrylicinnervessel,thegeodewasopaquesoas Atomic mass
composition:
H 12.2% 13.3%
C 84.4% 85.5%
Gd 0.1%
others 3.3% 1.2%
compatibility acrylic, Teflon
◦
density (20 C) 0.846 g/ml 0.854 g/ml
◦ ◦
Flash point 69 C 110 C
Scintillation yield 5300 ph/MeV (35% of anthracene)
Optical attenuation 4 m 10 m
length
Refractive index 1.472 1.476
Neutron capture 30.5µs 180µs
time
Neutron capture ∼6cm ∼40cm
path length
Capture fraction 84.1%
on Gd
ThesolutionoftheGadoliniumsaltGd(NO ) inhex-
3 3
anolas wellas the mixing of the basic and aromaticcom-
pounds wasperformedin a dedicatedhallclose to the en-
tranceoftheundergroundtunnel.TheamountofGadolin-
ium (0.09 % in weight) was chosen to optimize neutron
capture time and neutron detection efficiency. The mea-
suredvaluesfortheaveragecapturetime,pathlengthand
captureefficiencyarelistedinTab.5.Ahigherconcentra-
Fig. 13. Mechanicaldrawingof thedetector;thevisibleholes tion would have required more alcohol, which could have
on thegeode are for thePMT housing (from [42]). loweredthehighflashpointofthesolutionbelowthelimit
imposed by safety regulations. Moreover, as we shall see,
the presence of the nitrate ions in solution progressively
deteriorated the optical properties of the scintillator. A
higherconcentrationwouldhavefurthercompromisedthe
toopticallyshieldtheinnerregionsfromthevetoscintilla- chemical stability of the scintillator.
tion light. The externalsurface was white-coatedin order A fundamental quantity for normalizing the neutrino
to enhance the light collection in Region III and improve event rate is the number of free protons in the target.
the Veto rejectionefficiency, while the black inner surface An accurate evaluation of this number relies on precise
reduced light reflections, which could degrade the vertex measurements of the density and of the Hydrogen con-
resolution in the detector[43]. tent of the Gd-loaded scintillator. The Hydrogen content