Neutrino2002Poster Session

LLarge VVolume

DDetector at LNGS

and eutrino oscillationMarco Selvi

Bologna University & INFNOn behalf of the LVD collaboration

Physics with Physics with SupernovaeSupernovae

Detector descriptionDetector description

CNGS beam MonitorCNGS beam Monitor

Rock Mass 1969 kt

Nominal CNGS intensity

4.5 x 1019 pot/year

CC interaction rate 5.85 x 10-17 CC/pot/year

CC in rock 5.18 x 106 / year

Number of in LVD

33600 / year

LVD efficiency 72%

Detected in LVD ~ 120 / day

The possibility to detect neutrinos in different channels makes LVD sensitive to different scenarios for properties, such as normal or inverted mass hierarchy, and adiabatic or non adiabatic MSW resonances associated to Ue3.

Thanks to its large area (~130 m2) LVD could act as a very efficient muon monitor for the CNGS beam. A mean number of 120 muons per day is expected at nominal beam intensity. A 3% statistical error is achievable in 9 days. The measurement is pratically background free.

The Large Volume Detector (LVD) in the INFN Gran Sasso National Laboratory, Italy, consists of an array of 840 scintillator counters, 1.5 m3 each. These are interleaved by streamer tubes, and arranged in a compact and modular geometry. The active scintillator mass is M=1000 t.

There are two subsets of counters: the external ones (43%), operated at energy threshold h 7 MeV, and inner ones (57%), better shielded from rock radioactivity and operated at h 4 MeV.

In order to tag the delayed pulse due to n-capture, all counters are equipped with an additional discrimination channel, set at a lower threshold, l 1 MeV.

Relevant features of the detector are:

• good event localization;

• accurate absolute and relative timing: tabs = 1 s, trel = 12.5 ns;

• short dead time (2 s for each counter)

• uptime greater than 99%

• energy resolution: (E)/E = 0.07 + 0.23 (E/MeV)-0.5.

Number of events expected in LVD, in the reaction e p,n e+, as a function of . The dashed line represents the no-oscillation case, while full and dotted lines represent the oscillation case, adiabatic and non adiabatic, respectively. A large increase due to mixing is clearly visible, with respect to the no-oscillation case. It should be noted that the number of e p,n e+ events is practically the same both for adiabatic and non-adiabatic conditions, since, for normal mass hierarchy, MSW effect takes place in the neutrino sector only. Quite a different picture would appear, if we were to assume inverse mass hierarchy.

eeT T

Number of events expected in LVD, in NC interactions with 12C, as a function of . They are, of course, insensitive to mixing. However, the number of carbon de-excitations can test the temperature of neutrinospheres at the source, and therefore could be used in combination with CC data to overcome theoretical uncertainties on the temperature.

eT

•The observable neutrino reactions are:

• e p,n e+ observed through a prompt signal from e+ above threshold h

(detectable energy Ed Ee - 1.8 MeV + 2 me c2), followed by the signal from the n p,d capture (E = 2.2 MeV), above l and with a mean delay t 180 s.

• e 12C, 12N e-, observed through two signals: the prompt one due to the e-above h

(detectable energy Ed Ee - 17.8 MeV)

followed by the signal, above h , from the decay of 12N (mean life time = 15.9 ms).

• e 12C, 12B e+, observed through two signals: the prompt one due to the e+ above h

(detectable energy Ed Ee - 13.9 MeV + 2 me c2), followed by the signal, above h , from the decay of 12B (mean life time = 29.4 ms).

• l 12C, l 12C* , whose signature is the monochromatic photon from carbon de-excitation (E= 15.1 MeV), above h .

• l e-, l e-, which yields a single signal, above h , due to the recoil electron. The signature of this channel is not as clear as the other ones, and the number of expected events is low; therefore, we disregard it in the following.

(-) (-)

(-) (-)

In spite of the lack of a “standard'' model of the gravitational collapse of a massive star, some features of its dynamics and, in particular, of the correlated neutrino emission appear to be well established.

At the end of its burning phase a massive star (M > 8 Mo) explodes into a supernova (SN), originating a neutron star which cools emitting its binding energy EB ~ 3 x 1053 erg mostly in neutrinos.

The largest part of this energy, almost equipartitioned among neutrino and antineutrino species, is emitted in the cooling phase: Eanti-e ~ Ee ~ Ex ~ EB/6 (x denotes generically , anti- , , anti-).

The energy spectra are approximatively Fermi-Dirac, but with different temperatures, since e , anti-e , x have different couplings with the stellar matter: Te, < Tanti-e < Tx .

These features are common to all existing stellar collapse models, and lead to rather model independent expectations for supernova neutrinos. The observable signal is then sensitive to intrinsic properties, as oscillation of massive neutrinos. Indeed, oscillations change significantly the expected number of events in LVD.

In the study of supernova neutrinos , are indistinguishable, both in the star and in the detector, because their energy is below the charge lepton production threshold; consequently, in the frame of three-flavor oscillations, the relevant parameters are just m2

sol , Ue22)

and (m2atm , Ue3

2).

We will adopt the following numerical values: m2sol ,= 5 x

10-5 eV2 , m2atm = 2.5 x 10-3 eV2 , Ue2

2 = 0.33; the selected solar parameters m2

sol , Ue22) describe a LMA solution,

favored by recent analyses.

For a normal mass hierarchy scheme, neutrinos (not anti-neutrinos) cross two resonance layers: one at higher density (H), which corresponds to m2

atm , and the other at lower density (L), corresponding to m2

sol , (for inverted mass hierarchy, transitions at the higher density layer occur in the anti-neutrino sector, while at the lower density layer they occur in the neutrino sector. Anyway, both in case of normal and inverted mass hierarchy, the dynamics of collapse is not affected, since these layers are located far outside the core of the star).

Given the energy range of supernova neutrinos (5 MeV E 50 MeV), and considering a star density profile 1/r3, the adiabaticity condition is always satisfied at the L resonance for any LMA solution, while at the H resonance, this depends on the value of Ue3

2.

When Ue32 5 x 10-4 the conversion is completely adiabatic,

meaning that e are completely converted into the mass eigenstate 3 (detected at the Earth mainly as and ).

Therefore, the SN neutrino signal could feel the effect of Ue32

(and could also help to discriminate the type of mass hierarchy).

F. Vissani, G. Nurzia & LVD collaboration

Reference paper: see TAUP 2001 Proceedings (astro-ph/0112312)

We calculated the number of events expected in the various reaction in the cases of no-oscillation and oscillation, under the following hypotheses:

• We assumed a supernova exploding at D=10 kpc, with an energy release Etot = 3x1053 erg, pure Fermi-Dirac time integrated spectrum, energy equipartition, and neutrinospheres temperatures as Te = Tanti-e = Tx /2.

• We included the active mass of the detector and the energy thresholds. We used the following values of detection efficiencies above threshold: e p,n e+) = 95% and n p, d ) = 50% , e 12C, 12N e-) = 85% , e 12C, 12B e+) = 70% and l 12C, l 12C*) = 55% .

• In the oscillation case, we used two extreme values for Ue32 : Ue3

2 = 10-2 and Ue32 =10-6, and

the above mentioned mixing parameters (normal mass hierarchy, LMA solution).

• We did not include Earth matter effects (“open sky” neutrino burst).

Number of events expected in LVD, in CC interactions with 12C, due to both e and e, as a function of : the dashed line represents the no-oscillation case, while full and dotted lines represent the oscillation case, adiabatic and non adiabatic, respectively. The mixing results in an increase of the number of events, either for adiabatic or for non adiabatic conditions: in case of adiabaticity the increase is larger, and this is solely due to e interactions.

eeT T

“Beyond material, mass and depth, a Supernova neutrino telescope must have:

• buffers adequate to handle high throughoutput,

• short deadtime

• accurate absolute and relative timing

• good energy resolution

• low maintenance cost and a high duty cycle

(A. Burrows, 1992)

LVD detector fulfills all requests

Neutrino interactions in scintillator

Energy threshold

Number of events

e+ p n + e+ CC 1.8 390

i+ e- i + e- CC-NC 13 (3)

e+ 12C 12N + e-

CC 17.8 1

e+ 12C 12B + e+

CC 13.9 1

i+12C i + 12C*12C* 12C +

NC 15.11 22 (17)

e p,n e+ CC with 12C

NC with 12C

• Each cluster (sequence of k events, tk > t1, duration t= tk-t1) is taken into account (after rejection)

• Unique request: tmax = 200 s (in order to be model-independent)

• Given a standard acquisition rate, we compute the probability for each cluster (k, t) to be generated from background.

Experimental background measured in the fiducial volume of LVD and expected e+ spectrum. In a 10 s burst, about 5-10 background events are expected.

LVD

is

mem

ber

of

SN

EW

: S

up

ern

ova E

arl

y W

arn

ing

Syst

em

Event display of a CNGS muon in LVD. Side and top view of the detector are shown. The green area means energy loss in scintillator between 200 and 300 MeV (mean muon loss in LVD tanks), while yellow area stands for E between 300 and 500 MeV , i.e. muon has undergone some radiative energy loss.

The CNGS beam from CERN to the Gran Sasso Underground Laboratory (LNGS), over a distance of 732 km, is a wide-band high-energy beam optimized for appearance experiments. Such a beam provides a large number of interactions at Gran Sasso (about 2600 CC/kt/year at nominal beam intensity).

In principle the experiments forseen at LNGS could provide monitor informations by counting the number of CC interactions. Unfortunately, this could take months to accumulate due to their limited mass.

In order to monitor the performance of the CNGS beam, it has been suggested to implement, in one of the LNGS halls, a wide area simple apparatus capable of detecting the muons induced by neutrino interactions in the upstream rock and emerging into the experimental hall.

In this analysis we show the capabilities, in this respect, of the LVD detector whose beam-orthogonal surface is 13 x 10 m2 , greately much larger than the other forseen CNGS experiments.

Energy spectrum of muons when they reach LVD. The energy loss in rock, from the interaction point and the detector, has been subtracted using a full 3-d MonteCarlo simulation in which ionization, pair production and bremsstrahlung are taken into account.

The angle in space between muons and the main hall axis is the convolution of 3 contributes:

• The (fixed) beam angle w.r.t. the horizon (3.2O).

• The angle between and in the CC interaction.

• Multiple scattering in rock and other radiative processes (pair production, bremsstrahlung, ...).

Background estimateThe main background sources are cosmic muons. The rate in the full LVD detector (3 towers) is 8600 muons/day (6 per minute).

If we ask an energy loss greater than 200 MeV per tank, 22% of them survive, that is 1900 per day.

The use of the informations from the CNGS beam spill (10.5 s of spill lenght and 50 ms inter-spill gap) allows a reduction of the number of cosmic muons of a factor 104, i.e. about 0.5 cosmic muons per day.

The mean energy loss in LVD scintillator is 1.56 Mev/cm, the tank lenght is 1.5 m, so the mean energy loss in each tank is about 230 MeV. This allow to define a muon criteria, requiring at least one tank with energy > 200 MeV. The resulting efficiency is 72%.

A complete analysis of selected clusters tests their consistency with a neutrino burst, based on:

• the study of topological distribution of pulses inside LVD,

• the energy spectrum,

• the time distribution of delayed low energy pulses (due to neutron capture following the anti-e interaction).

(-) (-)

(-) (-)

A full simulation of the interaction, the muon transport in rock and the LVD detector response has been developed, in order to estimate muon tagging efficiencies CC interactions have been uniformly generated in a rock volume larger than the transverse LVD dimensions, in order to take into account also laterally impinging ).

The time-coincidence with the CNGS beam spill, makes this measurement pratically background free.

no oscillatio

nnon

adiabatic

adiabatic

no oscillatio

n

adiabatic

non

adiabatic