Progress in Particle and

Nuclear Physics

PERGAMON Progress in Particle and Nuclear Physics 48 (2002) S-20 http:Nwww.elsevier.com/locate/npe

Solar Neutrino Results from the Sudbury Neutrino Observatory

I. T. LAWSON*

Department of Physics, University qf Guelph. Guelph, Canada NlG 2WI

November 14, 2001

Abstract

Solar neutrinos from the decay of ‘B have been detected at the Sudbury Neutrino Observatory (SNO) via the charged-current (CC) reaction on deuterium and by the elastic scattering (ES) of electrons. The CC reaction is sensitive exclusively to electron neutrinos while the ES reaction also has a small sensitivity to muon and tau neutrinos. In addition, SNO has the capability to provide a measurement of the flux of all active neutrino flavoun via the neutral-current (NC) reaction on deuterium. The flux of electron neutrinos from ‘B decays measured by the CC reaction and the flux of the ES reaction, assuming no flavour transformation, will be presented. These flux measurements provide evidence that there is a non-electron flavour active neutrino component in the solar flux. The total flux of active sB neutrinos will be presented and shown to be in good agreement with predictions of solar models.

1 Introduction

For more than 30 years, solar neutrino experiments [l, 2, 3,4, 5, 61 have been observing fewer neutrinos than what are predicted by the detailed models of the Sun [i’, 81. The observed solar neutrino fluxes for these experiments is shown in table 1. As can be observed, the experimental results are less than the theoretical expectations for each experiment even though each experiment probes different aspects of the solar neutrino energy spectrum and have an energy dependence on the observed solar neutrino

*for the Sudbury Neutrino Observatory Collaboration: Q.R. Ahmad, R.C. Allen, T.C. Anderson, J.D. Anglin, G. Biihler, J.C. Barton, E.W. Beier, M. Bercovitch, J. Bigu, S. Biller, R.A. Black, I. Blevis, R.J. Boardman, J. Boger, E. Bonvin, M.G. Boulay, M.G. Bowler, T.J. Bowles, S.J. Brice, M.C. Browne, T.V. Bullard, T.H. Burritt, K. Cameron, J. Cameron, Y.D. Ghan, M. Chen, H.H. Chen, X. Chen, M.C. Chon, B.T. Cleveland, E.T.H. Clifford, J.H.M. Cowan, D.F. Cowan, G.A. Cox, Y. Dai, X. Dai, F. Dalnoki-Veress, W.F. Davidson, P.J. Doe, G. Doucas, M.R. Dragowsky, C.A. Duba, F.A. Duncan, J. Dunmore, E.D. Earle, S.R. Elliott, H.C. Evans, G.T. Ewan, 3. Farine, H. Fergani, A.P. Fer- raris, R.J. Ford, M.M. Fowler, K. Frame, E.D. Frank, W. Frati, J.V. Germani, S. Gil;, A. Goldschmidt, D.R. Grant, R.L. Hahn, A.L. Hallin, E.D. Hallman, A. Hamer, A.A. Hamian, R.U. Haq, C.K. Hargrove, P. J. Harvey, R. Hazama, R. Heaton, K.M. Heeger, W.J. Heintzelman, J. Heise, R.L. Helmer, J.D. Hepburn, H. Heron, J. Hewett, A. Hime, M. Howe, J.G. Hykawy, M.C.P. Isaac, P. Jagam, N.A. Jelley, C. Jillings, G. Jonkmans, J. Karn, P.T. Keener, K. Kirch, J.R. Klein, A.B. Knox, R.J. Komar, R. Kouzes, T. Kutter, C.C.M. Kyba, J. Law, LT. Lawson, M. Lay, H.W. Lee, K.T. J_esko, J.R. Leslie, I. Levine, W. Locke, M.M. Lowry, S. Luoma, J. Lyon, S. Majerus, H.B. Mak, A.D. Marino, N. McCauley, A.B. McDonald, D.S. McDonald, K. McFarlane, G. McGregor, W. McLatchie, R. Meijier Drees, H. Mes, C. Mifflin, G.G. Miller, G. Milton, B.A. Moffat, M. Moorhead, C.W. Nally, M.S. Neubauer, F.M. Newcomer, H.S. Ng, A.J. Noble, E.B. Norman, V.M. Novikov, M. O’Neill, C.E. Okada, R.W. Ollerhead, M. Omori, J.L. Orrell, S.M. Oser, A.W.P. Peon, T.J. Radcliffe, A. Roberge, B.C. Robertson, R.G.H. Robertson, J.K. Rowley, V.L. Rusu, E. Saettler, K.K. SchafIer, A. Schuelke, M.H. Schwendener, H. Seifert, M. Shatkay, J.J. Simpson, D. Sinclair, P. Skensved, A.R. Smith, M.W.E. Smith, N. Starinsky, T.D. Steiger, R.G. Stokstad, R.S. Story, B. Sur, R. Taflrout, N. Tagg, N.W. Tanner, R.K. Taplin, M. Thorman, P. Thornewell, P.T. tint, Y.I. Tserkovnyak, R. Van Berg, R.G. Van de Water, C.J. Virtue, C.E. Waltham, J.-X. Wang, D.L. Wark, N. West, J.B. Wiihelmy, J.F. Wilkerson, J. Wilson, P. Wittich, J.M. Wouters, and M. Yeh

0146-6410/02/$ - see front matter 0 2002 Elsevier Science BV. All rights reserved.

PII: SOl46-64lO(O2)00103-5

6 I. Z Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20

Table 1: A summary of the solar neutrino observations at different solar neutrino detectors. The neutrino production processes are shown in decreasing order of magnitude. The systematic and statistical uncertainties are added in quadrature for each experiment. Experiment Measured Flux SSM Flnx[7] Production Process Ref.

Homestake 2.56 f 0.23 SNU 7.6!1.3 SNU 1.1 8B, ‘Be, PP, PIP PI SAGE 67.2:;:; SNU 12829, SNU PP, PIP, ‘Be, sB [31 Gallcx 77.5:;:; SNU 128:; SNU PP, pep, 7Be, ‘B ]41 GNO 65.8+;!:: SNU 12829, SNU PP, pep, ‘Be, sB I51 Kamiokande 2.80 f 0.38 x lo6 cm-%-r 5.05(1?$:~) x lo6 cm-2s-1 sB ]21 Super-Kamiokande 2.32’:::: x lo6 cmWasW1 5.05(1+!:7:) x lOa cm-%-r 8B I61

Figure 1: The standard solar neutrino energy spectra for the different processes that produce the electrons neutrinos in the sun. The spectra are taken from reference [7] and references therein.

&‘2 a,, l1

~16'O 109 10 8 10' 106 105 104 103 10 2

10 -’ 1 10 E, (MeV)

flux. Figure 1 shows the energy spectra for the different reactions in the sun which produce electron neutrinos. One explanation of this long-standing neutrino flux deficit would be the transformation of the Sun’s electron neutrinos into another flavour while travelling to the Earth.

The Sudbury Neutrino Observatory (SNO) is 8n imaging water Cerenkov detector that was con- structed to resolve this solar neutrino anomaly. It can make simultaneous measurements of the electron- type neutrino (Y,) flux from the sB decay in the Sun and the flux of all active neutrino flavours through the following three reactions:

ue + d -+ p+p+e- (CC)

vz + d + p+n+v, (NC)

v, + e- -9 v,+e- (ES)

The charged-current (CC) reaction on the deuteron is sensitive exclusively to v,, and the neutral-current (NC) reaction has equal sensitivity to all active neutrino flavours (vz, x = e, p, T). The elastic scattering (ES) reaction is also sensitive to all active flavours, but with a reduced sensitivity to vP and or. Each of these interactions is detected by the SNO detector when one or more electrons produced during the reaction emit Cerenkov light that impinges on a phototube array. The ES reaction is highly directional, and establishes the sun ss the source of the detected neutrinos. The CC reaction produces an electron with an energy highly correlated with that of the neutrino. This reaction is sensitive to the energy spectrum of v, and hence to deviations from the parent spectrum. The NC reaction is detected after the neutron is absorbed by the deuterium giving a 6.25 MeV photon. The photon subsequently Compton scatters imparting enough energy into electrons to create Cerenkov light.

I. I? Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20 7

A comparison of the solar neutrino flux inferred from the reaction rates of these three interaction

channels under the assumption of no oscillations can provide evidence for flavour-changing neutrino oscillations. If v,‘s from the Sun transform into another active flavour, then the solar neutrino flux

deduced from the CC reaction rate (a”(~=)) must be less than those deduced from the ES reaction

rate or the NC reaction rate, such that Qcc(v.) < QES(v2) or (a”(~,) < aNC(vZ).

This paper presents recent results [9] from the first measurement of the solar sB neutrino flux by the SNO detector using the CC and ES reactions. The measured (Pss(~,) is consistent with the

high precision ES measurement by the Super-Kamiokande Collaboration [6]. However, the measured Gee (ve) at SNO is significantly smaller and is therefore inconsistent with the null hypothesis of a pure v, constituent in the solar neutrino flux. This indicates that the electron-type neutrinos emitted by the Sun are transformed into the other active neutrino species (v,, and r+) as they travel to the Earth.

2 The SNO Detector

The SNO detector [lo] is an imaging water Cerenkov detector located in the International Nickel Company (INCO) Creighton mine near Sudbury, Ontario, Canada. The excavation for the detector consisted of a barrel shaped cavity with a height of 34 m and a diameter of 22 m at a depth of 2092 m (or 6010 meters of water equivalent). Figure 2 shows a cross-sectional view of the SNO detector. It contains 1000 tonnes of 99.92% isotopically pure DzO contained inside a 12 m diameter acrylic sphere. The acrylic vessel (AV) is constructed out of 122 ultraviolet transmitting acrylic panels. This sphere is surrounded by 7000 tonnes of ultra-pure Hz0 contained in the cavity. The Hz0 shields the detector from high energy y-rays and neutrons originating from the cavity wall. A 17.8 m diameter stainless steel structure supports 9456 20-cm inward-facing photomultiplier tubes (PMTs). A non-imaging light concentrator is mounted on each PMT, and the total photocathode coverage is about 55% within 7 m of the center of the detector. An additional 91 PMTs are mounted facing outwards on the support structure to serve as a cosmic veto. To remove the vertical components of the terrestrial magnetic field, 14 horizontal magnetic compensation coils were built into the cavity wall. The maximum residual field at the PMT array is less than 19pT, and the reduction in photo-detection efficiency is about 2.5% from the zerofield value.

Physics event triggers are generated in the detector when there are 18 or more PMTs exceeding a threshold of about 0.25 photo-electrons within a coincidence time window of 93 ns. All the PMT hits registered in the approximate 420 ns window after the start of the coincidence time window are recorded in the data stream. This wider time window allows scattered and reflected Cerenkov photons to be included in the event. The mean noise rate of the PMTs is about 500 Hz, which results in about 2 noise PMT hits in this 420 ns window. The instantaneous trigger rate is about 15-20 Hz, of which 6-8 Hz are physics triggers. The remaining triggers are diagnostic triggers for monitoring the well-being of the detector. The trigger efficiency reaches 100% when the PMT multiplicity (Ni& in the event window is 2 23. For every event trigger, the time and charge responses of each participating PMT are recorded.

3 Physics Analysis Program

The solar neutrino physics analysis program is designed to exploit the unique NC capability of the SNO detector. Since the result of this NC measurement is a definitive statement on the oscillation of solar neutrinos, the SNO experiment intends to make three separate NC measurements of the total sB active neutrino flux.

The first NC measurement will be made with a pure DzO target; this phase is complete and the analysis is ongoing. The free neutron from the NC interaction is thermal&d, and about 30% of the time, a 6.25 MeV y-ray is emitted following the neutron capture by the deuteron. A large amount of the 6.25 MeV photopeak is below the neutrino analysis threshold thus requiring large statistics. The

8 I. T. Lawson / Prog. Part. Nucl. Phys. 48 (2002) S-20

Figure 2: A cross-sectional view of the SNO detector. The outer geodesic structure is the PMT support structure which surrounds the acrylic vessel. The light and heavy water regions are located where indicated.

second NC measurement will be made with NaCl added to the D,O; this phase is currently underway, in May 2001 two tonnes of NaCl were added to the DzO. In this configuration, the free neutron is readily captured by 35C1, and a cascade of y-rays with a total energy of 8.6 MeV follow. The neutron detection efficiency is significantly enhanced and approximately 45% of the NC events have a detectable signal above the analysis threshold. In the third NC measurement, discrete 3He proportional counters will be installed inside the DzO volume [ll]. The neutron detection efficiency of the proportional counter array is about 37%. In this detector configuration, the detection of the CC and NC signals are decoupled, and the covariance of the CC and NC signals that appear in the first two detector configurations is eliminated.

SNO can also make contributions to other areas of physics. For example, one can search for the relic supernova neutrinos integrated over all past supernovae. For the relic supernova neutrinos, the interaction of ve on protons in the deuterons produces a Cerenkov signal plus two neutrons, which makes a clean signature. Other topics that SNO will study include the flavour composition of the atmospheric neutrino flux, searches for certain types of dark matter and nucleon decay. The observatory is designed to handle the high data rates induced by an intense burst of neutrinos from a supernova explosion ss close as one kiloparsec away.

4 Solar Neutrino Analysis

The data reported in this paper were recorded from November 2, 1999 to January 15, 2001 and give a live time is 240.95 days. The target media was pure D20 throughout this period. A summary of the analysis procedure is shown in Fig. 3. The data were divided into two sets, with approximately 70% of the data used to establish the data analysis procedures and the remaining 30% of the data reserved

I. C Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20 9

Data Selection

+ PMT, Charge, Time

Calibration

- Inst. Bkg Removal I Inst. Bkg Removal II

+ +

R’econstruction I Reconstruction II

+

Calibrations I

T r Fiducial Volume I Fiducial Volume II

Energy ThLsho*cl cuts Energy &old Cuts

v Signal Extraction

Figure 3: A simplified flow chart of the solar neutrino analysis at SNO.

for a blind test of the statistical bias of these analysis procedures. AnalyM of the open and blind data sets employing the same techniques showed no statistically significant differences. In the following subsections, the analysis of the combined data sets are outlined.

4.1 Instrumental Background Removal

The analysis begins by removing the instrumental backgrounds from the data set. Electrical pickup or interference can create false PMT hits and electrical discharges in the PMTs or insulating materials may produce false light. These events have characteristic PMT time and charge distributions that are significantly different from Cerenkov light, and can be removed using cuts based on these distributions. For example, the light emitted from a PMT undergoing an electrical discharge is detected across the detector about 70 11s after the initial discharge is registered. Some of these light-emitting instrumental backgrounds are localized near the water piping near the top of the detector. Veto PMTs were installed in this region in order to enhance the rejection efficiency of these non-Cerenkov events. Most of the observed electronic channel charges in the interference events are near the pedestal, and can be removed by a cut on the mean charge of the fired PMTs. Some of these electrical discharge or electronic interference background events also have different event-to-event time correlations from physics events, and time correlation cuts are used to remove these events. Two independent instrumental background rejection schemes are used. An event-by-event comparison of the data sets reduced by these two schemes shows a difference of < 0.2%.

The efficiency and residual backgrounds of these instrumental background requirements are studied using a triggered 16N 6.13 MeV y-ray source [12] and a triggered *Li 13 MeV endpoint /J source [13] deployed in the DgO and Hz0 volumes in various locations. Further tests of the Nhits dependence on the cuts are performed with an isotropic light source at various wavelengths. The efficiency on the physics data after the instrumental background cuts are applied, weighted over the fiducial volume,

10 I. T. Lawson / Pmg. Part. Nucl. Phys. 48 (2002) 5-20

Progression of Instrumental Cuts

~~

li !&_ y.q; I._ :, ‘L;‘.‘, .*. :., - +c-___ ____ I._. i.

!, ‘; -..vp, . -.:,:. ‘C., ‘J”C”‘_.,_,; ___

-: “.;L _,! ‘: 3; .I 1.-.-*:,.-.,_ .(_ I...

D Nhits

Figure 4: Instrumental background removal, the different selection criteria are shown on the plot.

is measured to be 0.9967+::$$$. The residual instrumental background rejection contamination is less than 1%. Fig. 4 shows the progression of the instrumental cuts as they are applied to the raw data set.

In addition, cosmic ray induced neutrons and high-energy /?-decay nuclei from spallations can form a significant background to the solar neutrino signal. Since the cosmic muon event rate is sufficiently low, a 20-second coincidence window after each cosmic muon event is used to eliminate any contamination of the neutrino signal from the spallation products.

4.2 Event Reconstruction

For events passing the instrumental background cuts, all events with Nhits > 30 (-3.5 MeV electron energy) are reconstructed. The calibrated times and positions of the hit PMTs are used to reconstruct the vertex position and direction of the particle. Two different reconstruction algorithms were used, an ,event-by-event comparison shows good agreement between the data sets reconstructed by these two algorithms. The analysis described in this paper used a maximum likelihood technique which uses both the time and angular characteristics of Cerenkov light. Vertex reconstruction accuracy and resolution for electrons are measured using Compton electrons from the 16N T-ray source, and their energy dependence is verified by the 8Li p source. Compton scattered electrons from a 6.13 MeV y-ray are preferentially scattered in the forward direction relative to the incident y-ray direction. In order to minimize the effect of finite vertex resolution on this angular resolution measurement, only r6N events that are reconstructed to more than 150 cm from the source are used in this measurement. At these energies the vertex reconstruction resolution is 16 cm and the angular resolution is 26.7”. Reconstruction related systematic uncertainties to the solar neutrino flux measurements is about 4%.

4.3 Detector Calibration

The calibration of the PMT time and charge pedestals, slopes, offsets, charge versus time dependencies and second order rate dependencies are performed using electronic pulsers and pulsed light sources. Optical calibrations are obtained using a near-isotropic source of pulsed laser light (laser source) at

I. I: Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20

l NWD&

1

n WDdm

.

-MOllbCWb

4 APTData ccMontmcuk

% L (D E 10

z

1

406080 loo 120 140 160 1M 200 220 Numberk PNlT Hit8

Figure 5: Comparison of the Monte Carlo predicted responses to different calibrated sources. The sources are shown on the plot and the solid histograms are the Monte Carlo simulations of each source.

337, 365, 386, 420, 500 and 620 run with variable intensity at repetition rates from near 0 to 45 Hz [14, 151. The light source is deployed to locations accessible by the source manipulator system on the two orthogonal planes in the DzO and on a linear grid in the HzO. Optical parameters of different optical media in the detector are obtained at these wavelengths [16]. The attenuation lengths in DsO and Hz0 are found to be near the Raleigh scattering limit.

The absolute energy scale and uncertainties are measured with a triggered 16N source (predominantly 6.13 MeV T-rays) deployed in the same regions as the laser source. The detector energy response to the photopeak of this source provides a normalization to the PMT photon collection efficiency used in the Monte Carlo, and establishes the absolute energy calibration. A long-term stability study of the detector response to the 16N source shows a linear drift of -2.2f0.3% per year. The cause of this effect is under investigation, and a drift correction is applied to the event-by-event energy estimator.

The resulting Monte Carlo is then used to make predictions for the energy response to different calibration sources. The pT source generates 19.8 MeV T-rays through the 3He(p,y)4He reaction [17], and is used to check the linearity of the energy response beyond the endpoint of the *B neutrino energy spectrum. The z52Cf fission source provides an extended distribution of 6.25 MeV y-rays from d(n,r)t. Fig. 5 shows a comparison of the Monte Carlo predictions and the detector responses to these sources.

The energy response estimator uses the same input parameters as the Monte Carlo. It assigns an effective kinetic energy T,* to esch event based upon its position, direction and the number of hit PMTs within the prompt (unscattered) photon peak. For an electron of total energy E,, the derived energy response can be parameterized by a Gaussian:

I?(&, Eel = &i~(Ee) exp [-1(%?)‘] 3

where Ed = T,PI + m,, and the energy resolution is given by

ue(Ee) = -0.4620 + 0.547Ofi + O.O08722E, MeV. (2)

The systematic uncertainty on this absolute energy calibration is found to be f1.4%, which results in a neutrino flux uncertainty about 4 times larger. This is the most significant uncertainty in the flux

12 I. T. Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20

m 1.8 CD

0.6 -

0.4

l ch~ovwmh

0.2 0 Iwirume~~WB~

0 ~“~‘~~~~“~~1”~~~““~‘~‘~“~“~‘~~~~”~I”’~~~- 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 I

ITR

Figure 6: Separation of instrumental backgrounds and Cerenkov light events using the high level cuts (0,) and the (ITR).

measurement. Other energy related systematic uncertainties to the flux include the energy resolution and the energy scale linearity, and each contributes to no more than 0.5% uncertainty to the flux measurement. A second energy estimator using Nfits is employed for validation purposes. These two energy estimators give consistent results in the neutrino flux measurement.

4.4 Additional Background Requirements

Further instrumental background rejection requirements are applied to the data once the event recon- struction information becomes available. These high level cuts test the hypothesis that each event has the properties of single electron Cerenkov light. The reconstruction figure-of-merit cuts test for the consistency between the time and angular expectations for an event fitted to the location of the reconstructed vertex and that based on the properties of the Cerenkov light and the detector response.

The cerenkov characteristics of each event can is also parameterized using the the average opening angle between two hit PMTs (e,), measured from the reconstructed vertex, and the in-time ratio (ITR) which is the ratio of the number of hit PMTs within an asymmetric time window around the prompt light peak to the number of calibrated PMTs in the event. Fig. 6 shows the correlations between Bij and ITR for instrumental backgrounds and neutrino candidate events. As shown in the figure, this two dimensional cut has a very high instrumental background rejection efficiency.

The total signal loss from the different instrumental background cuts are calibrated with the 16N and the ‘Li sources. For the fiducial volume (radial distance R 5 550 cm) and the energy threshold (effective electron kinetic energy Tee 2 6.75 MeV) used in this analysis, the volume-weighted neutrino signal loss is determined to be 1.4+$%.

The residual instrumental background contamination in the neutrino signal after the background cuts is estimated using a bifurcated analysis, in which the signal contamination is obtained from cross calibrating the background leakage of two groups of orthogonal cuts. For the same fiducial volume and

I. 1: Lawson / Pmg. Part. Nucl. Phys. 48 (2002) 5-20

Table 2: Data reduction steps. Analysis step Number of events

Total event triggers 355 320 964

Neutrino data triggers 143 756 178

PMT hit multiplicity (Nhib) 2 30 6 372 899

Instrumental background (Pass 0) cuts 1 842 491

Muon followers 1 809 979

High level cuts 923 717

Fiducial volume cut 17 884

Energy threshold cut 1 169

Total events 1 169

13

energy thresholds, the instrumental background contamination is estimated to be less than 0.2% of the final neutrino candidate data set. Table 2 summarizes the sequence of cuts that are used to reduce the raw data set to 1169 neutrino candidate events.

4.5 Physics Backgrounds

In the remaining events above a threshold of T d 2 6.75 MeV, there are contributions from CC in the DzO, ES in the DzO and HzO, and the residual tail of neutron capture events (which can be NC or backgrounds from photodisintegration of the deuteron). The data set also includes a high energy tail of the internal radioactivity background, and high energy y-rays from the cavity wall. The radial distribution, (R/&v)‘, where &v = 600 cm is the radius of the acrylic vessel, of the simulated neutrino signals, weighted by the results from the signal extraction, is shown in Fig. 7. The data show a clear neutrino signal within the DsO. For the Hz0 region ((R/&&V)3 > l), the background contribution rises until it reaches the acceptance cutoff of the PMT light concentrators at R -7 m. A fiducial volume cut is applied at R = 550 cm to reduce backgrounds exterior to the DzO and to minimize systematic uncertainties associated with optics and reconstruction near the acrylic vessel.

Additional background can come from the & decays of the 20sT1 and 214Bi, which are daughters in the natural Th and U chains. These & radionuclei can also emit y-rays with sufhcient energy to photodisintegrate the deuteron. The free neutron from this break-up is indistinguishable from the NC signal. However, this neutron background can be subtracted from the total neutron signal in the detector if the internal radioactivity level of the detector is known. In this analysis, most of the Cerenkov signals from the ,&y decays are removed by the high energy threshold imposed. Fig. 8 shows a Monte Carlo radial distribution for the CC, NC and ES signals and backgrounds from the 2esT1 and 214Bi decays with low and high energy thresholds, it is clearly observed that the high threshold requirement removes nearly all of these background events inside the fiducial volume of 550 cm. Internal radioactivity levels in the DzO and Hz0 are measured by regular low level radio-assays of U and Th chain daughters. The light isotropy parameter B,, is also used to provide an in situ monitoring of these backgrounds. Both techniques show that the U and Th radioactivity levels in the DzO and Hz0 are either at or below the target level of one disintegration per day per tonne of water.

There are also /3~ contributions from the construction materials in the PMT support structure and the PMTs to the low energy background. Monte Carlo simulations predict that these contributions are insignificant to the flux measurement. This was verified by the deployment of an encapsulated Th source in the vicinity of the PMT support structure. Contributions from all sources of low energy backgrounds to the neutrino flux measurements is < 0.2%.

High energy y-rays from the cavity wall are also attenuated by the Hz0 shield. A limit on their leakage into the fiducial volume is estimated by deploying the 16N source in the vicinity of the PMT support structure. The contribution of these y-rays in the event candidate set is found to be < 10 events (68% CL.), or a 1.9% uncertainty to the ES flux and a 0.8% uncertainty to the CC flux.

I. Z Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20 14

60

OoL 0.5 5

Figure 7: Radial distribution of event candidates with Tea 2 6.75 MeV ss a function of the volume- weighted radial variable (R/&v)‘. The Monte Carlo simulation of the signals, weighted by the results from the signal extraction, is shown as the histogram. The dotted line indicates the fiducial volume used in this analysis.

4.6 Solar Neutrino Signal Extraction

The final data set contains 1169 events after the fiducial and kinetic energy cuts. The extended max- imum likelihood method is used in extracting the CC, ES and neutron contributions in the candidate data set. Data distributions in T&, (R/R*“)’ and cos& are simultaneously fitted to the probability density functions (PDFs) generated from Monte Carlo simulations assuming no flavour transformation and the shape of *B spectrum from Ortiz et al. [18] (hep neutrinos are not included in the fit). The quantity cosBo is the angle between the reconstructed direction of the event and the instantaneous direction from the Sun to the Earth and is shown in Fig. 9. The forward peak (cos& N 1) arises from the strong directionality in the ES reaction. The cos& distribution for the CC reaction, before accounting for the detector response, is expected to be (1 - 0.34Ocos&,) [19]. The extraction yields 975.4 f 39.7 CC events, 106.1 f 15.2 ES events and 87.5 l 24.7 neutron events for Td 2 6.75 MeV and R < 550 cm, where the uncertainties given are statistical only. The dominant sources of systematic uncertainty in this signal extraction are the energy scale uncertainty and reconstruction accuracy, see table 3 for the complete list of the systematic uncertainties. Independent analyses using Nhjts as an energy estimator, or in various fiducial volumes up to 620 cm with the inclusion of background PDFs in the signal extraction give consistent results.

The sB neutrino flux can be determined from normalizing the observed integrated event rate above the energy threshold. Assuming the 8B spectrum from [18], the flux deduced from the CC and ES reactions are:

(a$$o(~,) = 1.75 f O.O’l(stat.)f~:~:(sys.) f O.OB(theor.) x 106cm-2s-’ (3)

(a&o(Y,) = 2.39 f 0.34(stat.)‘$$(sys.) x 106cm-2s-1, (4)

where the theoretical uncertainty is the CC cross section uncertainty [20]. Radiative corrections to the CC cross section have not been applied to the CC cross section, but they are expected to decrease

I. T. Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20

on od I1 01 a0 Id

Figure 8: The left plot shows Monte Carlo simulations of the expected CC, NC and ES signals along with the backgrounds from the aosTl and 214Bi decays with Nh,, > 45 for the Radial distribution. The

right plot shows the same distributions but after the Arkits cut was increased to 65 (An electron energy of 6.75 MeV). It can be observed that most of the backgrounds are now removed from the Monte Carlo samples and that a fiducial cut of 550 cm removes nearly all of the remaining background.

O-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6ci;80 1

sun

Figure 9: The ~0~8~ distribution from the candidate event set. The Monte Carlo simulation of the signals, weighted by the results from the signal extraction, is shown as the histogram.

I. Z Lawson / Pmg. Part. Nucl. Phys. 48 (2002) 5-20

Table 3: Systematic uncertainties on the fluxes. Error source CC error ES error

Energy scale

(percent) (percent)

-5.2, +6.1 -3.5, +5.4

Energy resolution f0.5 f0.3

Energy scale non-linearity f0.5 f0.4

Vertex accuracy f3.1 53.3 Vertex resolution f0.7 f0.4

Angular resolution f0.5 f2.2

High energy y’s -0.8, +O.O -1.9, +o.o

Low energy background -0.2, +o.o -0.2, +o.o

Instrumental background -0.2, +o.o -0.6, +O.O

%igger efficiency 0.0 0.0 Live time fO.l fO.l

Cut acceptance -0.6, +0.7 -0.6, +0.7

Earth orbit eccentricity fO.l fO.l 1’0. ‘80 0.0 0.0

the measured $$o(ve) by up to a few percent [21]. The difference between @g&(z+,) and @g&v,) is 0.64f0.40 x 106cm-2s-1, or 1.60. For reference, the ratio of act sno(v.J to the predicted sB solar neutrino flux from BPBOl solar model [7] is 0.347 f 0.029, where all uncertainties are added in quadrature.

The Super-Kamiokande experiment has made a high precision measurement of the 8B solar neutrino flux by the ES reaction[6]:

a%(~=) = 2.32 f O.O3(stat.)+~$sys.) x 10scm-zs-l. (5)

It is observed that @&(vz) and a,“,( v, are consistent with each other. Assuming that the systematic ) uncertainties are normally distributed, the difference is,

iaL - @tio(va) = 0.57 f 0.17 X 106cm-2s-1, (6)

or 3.3~7. The probability that the observed (9 g$o(oJ is a 2 3.3~ downward fluctuation of Oz(,(v,) is 0.04%.

If v,,‘s from 8B decays in the Sun oscillate solely to sterile neutrinos, the SNO CC-derived 8B flux with Tea 2 6.75 MeV would be consistent with the integrated Super-Kamiokande ES derived flux above a threshold of 8.5 MeV [22]. The difference between these derived fluxes after adjusting for the ES threshold [S] is 0.53 f 0.17 x 10gcm-2s-1, or 3.la. The probability of a 2 3.10 downward fluctuation is 0.13%. Therefore, the results presented here are evidence for a non-electron type active neutrino component in the solar neutrino flux. These data are also inconsistent with the “Just-Soa” parameters for neutrino oscillations [23].

The CC energy spectrum can be extracted from the data by repeating the signal extraction with the CC energy spectral constraint removed. This is shown in Fig. 10. There is no evidence for spectral distortion under the no-oscillation hypothesis.

5 8B Neutrino Flux

Using the high precision ES messurement Og(4,(v,) and the pure u, flux from @$o(vJ, one can infer the flux of the non-electron flavour active neutrino a(~~~) to be

@E = @(ve) + O.l54@(v,,). (7)

I. T. Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-20

-I I

8 9 10

Figure 10: SNO CC energy spectrum. The left plot shows the extracted CC kinetic energy spectrum from a shape-unconstrained fit of events with R 5 550 cm and Td 2 6.75 MeV. The error bars are statistical only. The expected undiiorted sB spectrum, derived from reference [18], is shown ss the histogram. The right plot shows the ratio of the extracted CC spectrum to the expected kinetic energy spectrum. The band at each energy bin represents the la uncertainty derived from the most significant energy-dependent systemic uncertainties. The uncertainties in the sB spectrum have not been included.

This is shown in Fig. 11, in which @(vP7) is shown versus @(v.). The two data bands are the 1~ measurements of @go and Og, and the error ellipses are 68%, 95% and 99% joint probability contours for a(~.) and a(~,,~). The best fit to a(~,,~) is

@(vP7) = 3.69 f 1.13 x 106cm-2s-1, (8)

and the flux of the v. neutrino is

a’(~,) = 1.75 f 0.13 x 106cm-2s-‘. (9)

The total *B flux derived from the SNO and Super-Kamiokande experiments (@EK+SNo) is shown ss the diagonal band in Fig. 11. The agreement with the standard solar model prediction (QisM) is good. The total flux of active sB neutrinos is found to be

a(~=) = 5.44 f 0.99 x 106cm-2s-1.

This is the first determination of the total flux of ‘B neutrinos generated by the Sun.

(10)

6 Cosmological Implications

Assuming that the oscillation of massive neutrinos explains both the evidence for electron neutrino flavour change presented here and the atmosphericneutrino data of the Super-Kamiokande collab oration [24], two separate splittings of the squares of the neutrino mess eigenvalues are indicated:

I. I: Lawson / Prog. Part. Nucl. Phys. 48 (2002) 5-i-20

(relative to BPBOl) o . 2 $2; . 0.6 0.8 1 1.2

0 1 2 3 4 5 6

$(v,) ( lo6 Cm-2i’)

Figure 11: Flux of non-electron flavour active sB solar neutrinos (a(~~~)) versus @z$o and @g. The diagonal solid band is derived from the SNO and the Super-Kamiokande results (@zK+SNo) while the dashed band shows the BPBOl prediction [7] (QzSM); these are in good agreement. The intercepts of these bands with the axes represents the fla errors.

(Amev,,)2 or (Ame,)” are < 10e3 eV2 for the solar sector [7, 251 and (Arr~,,)~ N 3.5 x 10e3 eV2 for

atmospheric neutrinos. These results, together with the beta spectruni of tritium [26], limit the sum of maes eigenvalues to be between 0.05 and 8.4 eV, corresponding to a constraint 0.001 < 52, < 0.18 for the contribution to the critical density of the universe [27, 281.

7 Conclusion and Future Prospects

In conclusion, the results presented here are the first direct evidence that there is an active non-electron flavour neutrino component in the solar neutrino flux. This is also the first experimental determination of the total flux of active ‘B solar neutrinos, which is in good agreement with the solar model predictions. The SNO Collaboration is now analysing additional data from the completed pure DzO phase and with a lowered energy threshold. Efforts are devoted to understanding the low energy background contribution to the NC measurement.

The second phase of SNO began with the deployment of NaCl to enhance the NC capability on May 28, 2001. Fig. 12 shows the detector background level observed in the Cerenkov data before, during and after the NaCl addition. The increase in the event rate during the injection is attributed to 24Na, which were activated by neutrons from the cavity wall when the NaCl brine was stored in the underground laboratory. After the addition process finished, one observes the decay of the =Na with a characteristic half-life of 15 hours. The background level in the detector returned to the original values after several days. This experiment phase will last for about’ 8 months of neutrino live time, after which the 3He counters will be installed.

I. T. Lawson / Pmg. Part. Nucl. Phys. 48 (2002) 5-20 19

100

4 J 150 Level Switches Change

‘d

Injection Ends

0 100 100

Th Ih salt Idctim mara)

Figure 12: The event rate in a low energy background monitoring window before, during and after the NaCl injection. The gaps in this plot indicate detector down time, when detector hardware changes necessary to salt deployment were made. The dotted line is an exponential fit of the event rate. The fit is consistent with r1j2 = 15 hours, which is the half-life of %Na.

8 Acknowledgements

This research was supported by the Natural Sciences and Engineering Research Council of Canada, Industry Canada, National Research Council of Canada, Northern Ontario Heritage Fund Corpora- tion, the Province of Ontario, the United States Department of Energy, and in the United Kingdom by the Science and Engineering Research Council and the Particle Physics and Astronomy Research Council. Further support was provided by INCO, Ltd., Atomic Energy of Canada Limited (AECL), Agra-Monenco, Canatom, Canadian Microelectronics Corporation, AT&T Microelectronics, Northern Telecom, and British Nuclear Fuels, Ltd. The heavy water was loaned by AECL with the cooperation of Ontario Power Generation.

References

[II

PI

[31

[41

151

161 171

[Sl

B.T. Cleveland et al., Astm. J. 496 (1998) 505.

KS. Hirata et al., Phys. Rev. Lett. 65 (1990) 1297; K.S. Hirata et al., Phys. Rev. D 44 (1991) 2241; KS. Hirata et al., Phys. Rev. D 45 (1992) 2170E; Y. Fukuda et al.Phys. Rev. Lett. 77 (1996) 1683. This is the combined results from Kamiokande II and III.

J.N. Abdursshitov et al., Phys. Rev. C 60 (1999) 055801.

W. Hampel et al., Phys. Lett. B 447 (1999) 127. This is the combined results from’ GALLEX I, II, III and IV.

M. Altmann et al., Phys. Lett. B 490 (2000) 16.

S. .Fukuda et al., Phys. Rev. Lett. 86 (2001) 5651.

J.N. Bahcall, M.H. Pinsonneault and S. Basu, Ast. J. 555 (2001) 990.

A.S. Brun, S. Tnrck-Chibe and J.P. Zahn, Ast. J. 525 (1999) 1032; S. Turck-Chieze et al., Ap. J. Lett. 555 (2001) L69.

20 I. T. Lawson /Prog. Part. Nucl. Phys. 48 (2002) 5-20

[9] The SNO Collaboration, Phys. Rev. Lett. 87 (2001) 071301.

[lo] The SNO Collaboration, Nucl. Inst. and Meth. A 449 (2000) 172.

[ll] M.C. Browne et al., IEEE tins. on NucE. Sci. 46 (1999) 873.

[12] M.R. Dragowsky et al., nucl-ex/0109011(2001), accepted for publication in Nucl. Inst. and Meth. A.

P31

1141

1151

P61

1171

P81

P91

I201

1211

I221

[231

I241

[251

1261

I271

[281

N. Tagg, The sLi Calibration Source and Through-Going Muon Analysis in the Sudbuy Neutrino Observatory, Ph.D. thesis, The University of Guelph (2001).

R.T. Ford, Nitrogen/Dye Laser System for the Optical Calibration of SNO, M.Sc. thesis, Queens University (1993).

R.T. Ford, Calibration of SNO for Detection of ‘8 Neutrinos, Ph.D. thesis, Queen’s University (1998).

B.A. Moffat, The Optical Calibration of the Sudbury Neutrino Observatory, Ph.D. thesis, Queen’s University (2001).

A.W.P. Poon et al., Nucl. Inst. and Meth. A 452 (2000) 115.

C.E. Ortiz et al., Phys. Rev. Lett. 85 (2000) 2909.

J.F. Beacom and P. Vogel, Phys. Rev. D 60 (1999) 053003.

S. Nakamura, T. Sato, V. Gudkov and K. Kubodera, Phys. Rev. C 63 (2001) 034617; M. Butler, J.-W. Chen and X. Kong, Phys. Rev. C 63 (2001) 035501; G. ‘t Hooft, Phys. Lett. B 37 (1971) 195. The Butler et aE. cross section with &,A = 5.6 fm3 is used.

I.S. Towner, J. Beacom and S. Parke, private communication; I.S. Towner,Phys. Rev. C 58 (1998) 1288; J. Beacom and S. Parke, Phys. Rev. D 64 (2001) 091302; J.N. BahcaIl, M. Kamionkowski and A. Sirlin, Phys. Rev. D 51 (1995) 6146.

G.L. Fogli, E. Lisi, A. Palazzo and F.L. Villa&e, Phys. Rev. D 63 (2001) 113016; F.L. Villa&e, G. Fiorentini and E. Lisi, Phys. Rev. D 59 (1999) 013006; G.L. Fogli, E. Lisi, D. Montanino and A. Palazzo, Phys. Rev. D 64 (2001) 093007.

J.N. Bahcall, P.I. Krastev and A. Yu. Smirnov, J. High Energy Phys. 05 (2001) 015.

T. Toshito et al., hegex/0105023.

M. Apollonio et al., Phys. Lett. B 466 (1999) 415.

J. Bonn et al., Nucl. Phys. Proc. B Suppl. 91 (2001) 273.

Allen’s Astrophysical Quantities, 4th ed., ed. Arthur Cox (New York: Springer-Veriag) 2000; Eu- ropean Physical Journal Cl5 (2000) 1.

H. Pas and T.J. Weiler Phys. Rev. D 63 (2001) 113015.