jump to navigation

Neutrino Telescopes Day 1 note March 11, 2009

Posted by dorigo in cosmology, news, physics, science.
Tags: , , , , , , , ,
trackback

Below are some notes I collected today during the first day of the “Neutrino Telescopes” conference in Venice. I have to warn you, dear readers, that my superficial knowledge of most of the topics discussed today makes it very likely to certain that I have inserted some inaccuracy, or even blatant mistakes, in this summary. I am totally responsible for the mistakes, and I apologize in advance for whatever I have failed to report correctly. Also, please note that because of the technical nature of this conference, and the specialized nature of the talks, I have decided to not even try to simplify the material: this is thus only useful for experts.

In general, the conference is always a pleasure to attend. The venue, Palazzo Franchetti, is located on the Canal Grande in Venice. To top that, today was a very nice and sunny day. I skipped the first few “commemorative” talks, and lazily walked to the conference hall in time for coffee break. The notes I took refer to only some of the talks, those which I managed to follow closely.

Art Mc Donald, SNO and the new SNOLAB

This was a discussion of the SNO experiment and a description of the new telescopes that will start to operate in the expansion of the SNO laboratory. SNO is an acrylic vessel, 12m in diameter, containing 1000 tonnes of deuterium (D_2 O), with some additional 1700 tonnes of water for inner shielding, and 5300 tonnes for outer shielding. 9500 photomultiplier tubes watch it, quick to record the faint neutrino signals.

The detector is located deep underground, in the Creighton mine near Sudbury, Ontario, Canada. The depth makes for smaller cosmic-ray backgrounds than other neutrino detectors, at a depth where muons from neutrino interactions start to compete with primary ones.

SNO was designed to observe neutrinos in three different reactions:

  1. In the charged-current weak interaction of a neutrino with a deuterium nucleus the neutrino becomes an electron, emitting a W boson which turns the nucleus into a pair of protons. This reaction has a energy threshold of 1.4 MeV, and the electron can be measured by the Cherenkov light it yields in the liquid.
  2. Neutral-current interactions -where the neutrinos interact with matter by exchanging virtual Z bosons, are possible with all kinds of neutrinos, and they provide a signature of a neutron and a proton freed from the nucleus, if the incoming neutrino has an energy above 2.2 MeV.
  3. Finally, elastic scattering both in water and deuterium can occur between neutrinos and the electrons of the medium.

SNO tries three neutron detection methods, which are “systematically different”: they rely on different physical processes and have thus different measurement systematics. First of all, in pure heavy water one can detect neutrons by capturing them into deuterium, with the emission of a 6.25 MeV photon.

Putting salt in the detector allows to get more gamma rays from neutron capture, because the sodium chloride allow neutron capture in 35Cl, and neutral current events can be separated from charged-current events using event isotropy.

In phase III they put in an array of long tubes of ultrapure Helium-3, and observe neutron capture and measure neutral current rates with an entirely different detection system.

Measurements showed that CC and NC reactions were not the same, fluxes were in a ratio of R(CC/NC)=0.34 \pm 0.023^{+0.029}_{-0.031}.

Phase III consists in inserting 40 strings on a 1-meter spaced grid in the vessel, for a total of 440 meters of proportional counters filled with 3He. The signal collected in phase III amounts to 983+-77 events.

Combined with the results of the KamLAND and Borexino experiments, the fit to SNO data constrains the angle \theta_12 to 34.4+-1.2 degrees, and \delta m^2 = (7.59^{+0.19}_{-0.21})\times 10^{-5} eV^2.

The future for SNO is to have it filled with liquid scintillator doped with Neodimium for double beta decay studies. 150-Nd is one of the most favourable candidates for double beta, with large phase space due to its high endpoint energy (3.37 MeV). It provides for a long attenuation length, and it is stable for more than 2 years. For double beta decays they expect to get to 0.1 eV sensitivity with a 1000 ton-mass detector.

Atsuto Suzuki: KamLAND

Atsuto discussed the history and the results of the KamLand experiment. There was a first proposal of the detector in 1994, a full budget approval in 1997 by the Japanese. In April 1998 the construction started, and in 1999 US-Kamland was approved by DOE. Data taking began in 2002. In August 2009 there will be a new budget proposal, for double beta decay studies.

Kamland consists in a Xenon-filled vessel, with an outside one filled with Gadolinium. Kamland wants detects neutrino oscillation with >100 Km baseline, exploiting the many nuclear reactors in Japan. The second goal is to search for geo-neutrinos: these are potential anti-neutrinos coming from fusion processes which could hypothetically take place at the center of the Earth.

Many reactors provide the source of neutrinos, a total of 70GW (12% of global nuclear power) at an average 175+-35 km distance from KamLAND. The largest systematic for reactor neutrino detection come from the knowledge of the fiducial volume (4.7%), the energy threshold (2.3%), the antineutrino spectrum (2.5%), for a total of 6.5%.

The experiment observed neutrino disappearance, measured the parameters of neutrino oscillations, and also put an upper limit of 6.4 TW for geo-neutrinos. Theoretical models, which predict the power at 3 TW, have not been excluded yet.

Gianluigi Fogli:  SNO, KamLAND and neutrino oscillations: theta_13.

Gianluigi started his talk with a flash-back: four slides which were shown at NO-VE 2008, the former instantiation of this conference. This came after the KamLAND 08 release, but before the SNO 2008 release of results.

What one would like to know is the hierarchy (normal or inverted), the CP asymmetry in the neutrino sector, and the \theta_{13} mixing. Some aspects of this picture are currently hidden below the 1-sigma level. A recent example is the slight preference for \sin^2_{13} = 0.01 from the combination of solar and reactor 2008 data. They are consistent with zero but their combination prefers a value 1-sigma different.

In the second slide from 2008, the reason was discussed. A disagreement comes from the difference between solar data, SNO-dominated, and the kamLAND data at \theta_{13}=0. The disagreement is reduced for \theta_{13} >0. A choice of \sin^2 \theta_{13}=0.03 (instead of zero) gives a better fit of the two sets of data. It is a tiny effect, but with some potential for improvement, once final SNO data and further Kamland data will be available.

The content of Fogli’s talk was organized as a time-table of eight events, in two acts.

First: in May 2008 the effect was discussed independently by Balantekin and Yilmaz. Then, in May, SNO-III data was released. In June, our analysis giving \sin^2_{13} = 0.021 \pm 0.017 went to PRL, and then an independent analysis of S+K was given in August.

Concerning atmospheric and long-baseline neutrinos, there were results yielding 0.016+-0.010 from all data in our analysis, then comments on the atmospheric hint by Maltoni and Shwetz, then a new three-flavor atmospheric analysis from SK. Finally, just a month ago we saw the first MINOS results on electron neutrino appearance.

Act one: solar and kamland hint for \theta_{13}>0: Balantekin and Yilmaz discussed it. The release of SNO-III data saw a strong improvement in the data, and the result is slightly lower cc/nc ratio, so a slightly lower value of \sin^2_{13} is preferred. Fogli here noted that the new data are ok from a model-independent viewpoint, that is, there is an internal consistency among SNO and SK. Also, there is consistency among neutral-current measurements and the standard solar model of 2005. On the other hand, also kamland data have their own internal consistency: they reconsruct the oscillation pattern through one full period. The fact that the solar and kamland datasets are ok, but they disagree on theta_12, unless theta_13>0, is thus intriguing.

Event 3: the hints of theta_13>0 from the global analysis. We have the hint plotted in the plane of the two mixing angles, and you see that the solar and Kamland region in sintheta_13 vs sintheta_12. the agreement is reached only if \sin \theta_{13} is larger than zero. When they are combined, they find a best fit more than one sigma away from zero, 0.021+-0.017. The reason of the different correlation of the two mixing angles relies on the relative sign of mixings in the expression for the survival probability of neutrinos in SNO and Kamland. At low energy, in the vacuum the survival probability is given with an anticorrelation of \sin^2 \theta_{12} and \sin^2 \theta_{13}. At high energy, adiabatic MSW (SNO), the sign is opposite.

Complementarity: solar and kamland data taken separately prefer theta_13=0. Combined they are 1.2 sigma away from zero.

Event 4 in the list given above was the analysis by Schwetz and Tortola and Valle: they also found a preference for \theta_{13}>0  at a slightly higher confidence level.

In conclusion, a weak preference for \theta_{13}>0 is accepted at 1.2-1.5 sigma. Is this preference also supported by atmospheric and acceleratr data ? In Fogli’s paper (0806.2649) they used as independent support for a nonzero value of the angle, an older hint coming from their analysis of atmospheric data with CHOOZ and long-baseline results.

The complication comes out in Act 2: event 5 is the older but persisting hint for \theta_{13}>0. It comes from the 3-neutrino analysis of atmospheric, LBL, and CHOOZ data. There one has to go in detail, by considering what one means when one talks of an excess of electron events induced by three-neutrino sub-leading effects. The calculations are based on a numerical evolution of the Hamiltonian along the neutrino path in the atmosphere and in the known Earth layers. However, semianalytical approximations can be useful. An important observable is the excess of expected electron events compared to the no-oscillation case.

The excess is given by a formula,N_e/N_0-1 = (P(ee)-1)+rP(e \mu), where P(ee) and P(e\mu) are the oscillation probabilities, and R is the ratio of fluxes. The excess is zero when both \theta_{13} and \delta m^2 are both zero, but can have contributions otherwise.

We have two kinds of matter effects that take place in the propagation. If one assumes a constant density approximation, and with a normal hierarchy, the three quantities can be given, where one can distinguish the theta_13, the delta_m, and the interference terms. All three effects can singularly dominate. The different terms help fitting the small electron excess in sub-Gev and multi-Gev data.

The atmospheric three-neutrino analyses by the SK Collaboration (in hep-ex/0604011) and Schwetz, Tortola, and Valle in 0808.2016 cannot directly compare with the one of Fogli, because they do not include the two sub-leading solar terms, since they make the assumption of one-mass-scale-dominance.

Sticking to his own analysis, Fogli continued by taking the two hints from solar+kamland results on one side, and atmospheric neutrinos+chooz+lbl on the other: they indicate a 1.6 sigma discrepancy from zero of theta_13. Combining all data together, sin^2 theta_{13} = 0.016+-0.010. This is the result of 0806.2649. Below are the results for the two angles together, showing their anticorrelation in the two simultaneous determinations.

Event 6: rather recent, in December of last year Maltoni and schwetz published 0812.3161, which includes the discussion of the preliminary Superkamiokande-II data. Using SK-I data they find at most 0.5 sigma from atmospheric neutrinos plus chooz data. This is weaker than Fogli’s 0.9 sigma, but shows similar qualitative features.

Event 7: a discussion of the data of SK-I, SK-II, and maybe SK-III, even if all these things are not yet published. There eists ongoing three-flavor analyses as reported in recent PhD theses using SK I+II data. Wendell, Takenaga. Unfortunately, none of the above analyses allows both theta_13 and \Delta M>0, and thus they do not include interference effects linear in theta_13, which may play some non trivial role.

Concerning the sub-Gev electron excess, effects persist in phases I and II, but slight excess present of upgoing multi-Gev evens is present in SK I but not in SK II. This downward fluctution may disfavor a non-zero value of theta_13, as noted by Maltoni and Schwetz.

From SK-III two distributions presented at neutrino 2008 by J.Raaf show that a slight excess of upgoing multi-Gev seems to be back, together with a persisting excess of sub-Gev data.

So the question is: SK-III shows both effects. Can this be interpreted away from statistical fluctuations ? This requires a refined statistical analysis with a complete set of data coming from SuperKamiokande.

Currently, there is an impressive number of bins in energy and angle, and 66 sources of systematics. These need to be handled carefully. Such a level of refinement is difficult to reproduce outside of the collaboration, In other words, independent analyses of atmospheric data searching for small effects at the level of 1-sigma are harder to perform now. So, it will be important to see the next official SK data release and especially the SK oscillation analysis, hopefully including a complete treatment of three flavor oscillation withboth parameters allowed to go larger than 0.

In the meantime, Fogli noted that he does not have compelling reasons to revise his 0.9-sigma hint of theta_13 coming from published SK-I data.

Finally, Event 8: this last one is very recent, concerns the first MINOS results on electron-neutrino appearance. These preliminary results have been released too recently, and it would be unfair to anticipate results and slides that will be shown later in this workshop, but

Fogli could not help noticing that the MINOS best fit for theta_13 sits around the chooz limit, and is away from zero at 90% C.L.

If we see the glass half-full, then we might have two independent 90% C.L. hints in favor of theta_13>0: one coming from Fogli’s global analysis of 2008, and one coming from MINOS, that can be roughly symmetrized and approximated in the form \sin^2 \theta_{13} = 0.05 \pm 0.03. A combination at face value gives a value of 0.02 +- 0.01, an indication at 2-sigma of a non-zero value of this important angle. In other words, the odds against a null theta_13 are now 20 to 1.

G.Testera:  Borexino

Borexino is a liquid scintillator detector. The active volume is filled by 270 liters of liquid scintillator contained in a thin nylon vessel. Light emitted is seen by Photomultiplier tubes. The outer volume is filled by the same organic material, but with a quencher in the buffer region. Water used as shield. The tubes are looking at Cherenkov light. Used for active muon veto. Borexino is a simple detector, but in practice the requirements needed for the radiopurity are tough to comply with.

The physics goals are a measurement in real time of flux and spectrum of solar neutrinos in MeV or sub-MeV range. Why measure solar neutrinos of low energy ? The LMA-MSW model predicts a specific behavior for the neutrino survival probability for the various types of neutrinos emitted in the sun. The shape of the prediction as a function of energy shows a larger survival probability at lower energy.

All data before Borexino measured higher energy. So Borexino wants to measure shape of survival probability as a function of energy, going lower. Measurement can constrain additional oscillation models. If we asssume that neutrinos oscillate and we take data of the survival probability, we get the absolute neutrino flux, and we might be able to measure the component of the CNO source in the neutrino flux, this can help constrain the solar models.

Borexino can also see antineutrinos (geo-nu), in gran sasso this will be relatively easy because background from reactor antineutrinos is small. We need statistics, several years to collect significant data. The signal to noise ratio provided by the apparatus is of 1.2. The detector has also sensitivity to supernova neutrinos. Borexino is thus entering the SNEW community.

Results of Borexino will be complementary to others. Taking data since mid march 2007. They have about 450 days of live time so far. The process of neutrino detection is elastic scattering on electrons. High scintillator yield of 500 photoelectrons per MeV, a high energy resolution, and a low threshold. No information on the direction of neutrinos, however. Scintillator is fast, can reconstruct the position with time measurements. Different answers to alpha and beta particles can distinguish the two. The shape of the energy spectrum allows to distinguish them. The energy spectrum is the only sign they can recognize.

The story of the cleanliness of Borexino encompasses 15 years of work. Careful selection of construction materials, special procedures for fluid procurement, scintillator and buffer purification during filling. Background from U and Th is very small, smaller than the initial goals. The purity of the liquid scintillator is very high.

If there is only a neutrino signal, the simulation shows that the Beryllium 7 neutrino signal is very well distinguishable, it shows a flat spectrum with an upper edge at 350 MeV. 14C is at smaller energy. 11C at high energy cannot be eliminated. Can be tagged some way, but not completely eliminated. At further higher energy there is the signal from Carbon-10.

In 192 days of lifetime there is a big Polonium peak and the edge of the Beryllium region, together with a contribution from Kripton. Data indicates also the presence of Bismuth-210. The rate of neutrinos from 7Be is of 49 counts per 100 T. They see an oscillation from these neutrinos because otherwise they would see 75 counts +- 4. The no-oscillation hypothesis is rejected at 4-sigma level. This is the first real-time measurement of oscillation of 7Be neutrinos.

Largest errors are coming from the fiducial mass ratio, and detector response function. These amount to 6% each.

Neutrino interactions in the earth could lead to regeneration of neutrinos: solar nu flux higher in the night than in the day, due to geometry. In the energy region of 7Be, they expect a very small effect. A larger effect would be expected in the low-solution, now excluded.

A new preliminary result: the day-night asymmetry for 7Be solar neutrinos. 422 days of live time are used for this. In the region where neutrinos contribute, there is no asymmetry seen.

Flux of Boron-8 neutrinos with low threshold: Borexino goes lower in energy threshold. In Borexino they go down to 2-3 MeV. After subtracting the muon contribution they see the oscillation of 8B neutrinos. By putting them together with 7Be, more points can be added to the survival probability plot. They are describing well the curve as a function of energy.

In conclusion, Borexino claims a first real time detection of the 7Be flux.

M.Nakahata: Superkamiokande results in neutrino astrophysics.

Kamiokande from 1983 to 1996 was a 16m high, 15.6m diameter tank with more than a thousand large photomultiplier  tubes. SK started in 1996. A 50,000T water tank, 32,000 T photosensitive volume.

After the accident they took data at SKII, then 2006 was SKIII and new electronics, and since September 2008 it is SK-IV. The original purpose of Kamiokande was a search for proton decay. Protons could be though to decay to positron plus neutral pion; but they wanted to measure different branching ratios. They made a detector with large coverage.

telescope, the advantage was directionality, provided by the imaging Cherenkov detector. And the The large photocollection efficiency is useful also for detecting low-energy neutrinos. As a second item is energy information. The number of Cherenkov photons is proportional to the energy of the particle. Another advantage is the particle identification. From the diffuseness of the ring pattern they can distinguish electron from muon events. The misidentification probability is less than 1%, very important when discussing atmospheric neutrinos.

The first solar neutrino plot at Kamiokande came from 450 days of exposure. E>9.3 MeV threshold. Saw an excess of neutrinos coming from the sun, but could not say much about the size. In Superk they had larger number, 22400 solar neutrino events, 14.5 per day, with very precise flux, with stat accuracy of 1% and syst of about 4%. SK info gave 8B flux and \nu_\mu and \nu_\tau fluxes.

SK will measure the survival probability of solar neutrinos as a function of energy, from 4 MeV down, and measure their spectrum distorsion.

From the supernova SN1987, SK observed 11 events in 13 seconds. Other 11 events were seen in that case from Baksan and IMB3. ASsuming we now got a new Supernova at 10 kpc, SK could measure directly energy information from the reaction. The event rate would discriminate models.

Adding Gadolinium in water can reduce backgrounds, because n capture yields a gamma ray, which gives 8 MeV energy, and the time is correlated (30 msec delay). If invisible muon backgrounds can be reduced by a factor of five using this neutron tagging, with 10 years of SK the signal will amount to 33 events, 27 from backgrounds, in energy of 10 to 30 MeV: they can thus see SN relic neutrinos. But they must first study water transparency, corrosion in the tank, etcetera, due to the addition of Gadolinium.

Atmospheric neutrino anomaly in Kamiokande: mu-e decay ratio was the first evidence. Data from 1983 to 1985 allowed to measure the ratio, 60% of the expectation in mu/e ratio. A paper was published in 1988. In 1994 they obtained a zenith angle distribution for multi-GeV events. In superK they got a much better result, and got sub-GeV electron-like and muon-like events.

Oscillation agreed very well with observed data. The latest plot of two-flavor oscillation analysis gives a \delta m^2 = 0.0021 eV^2, and angle theta consistent with 1.0.

And that is all for today!

Comments

1. Marcos - March 11, 2009

Very nice post! I was hoping to get time (and funding) to attend this conference, I’ll be entering the world of nu-detectors this year as a PhD student and that conference offered a good panorama of the current status of the field. Thanks for posting on this.

2. Singlino - March 11, 2009

How come the uncertainty in R(CC/NC) is more accurate than the average value?

3. dorigo - March 11, 2009

You are welcome Marcos. Please note that the conference site has the slides of all talks, in close to real time (or even ahead of time, as this morning!):

Singlino, I am not sure I understand what you mean. Do you refer to the SNO measurement ? I did not quote the individual values, but the slides do. There, on page 8, you find (in units of 10^6 cm^2 s^{-1}), a CC flux of \Phi_{CC}=1.68 \pm 0.06 \pm 0.09, and a NC flux of \Phi_{NC} = 4.94 \pm 0.21 \pm 0.38. I do not see a problem in taking a ratio to R=0.34 \pm 0.023 \pm 0.031, given that some systematics may be correlated.

Cheers,
T.

4. Andrea Giammanco - March 11, 2009

Singlino, when you measure the CC cross section you are counting CC-like events, and convert this number into a cross section. Ditto for NC. But if you are only interested in the ratio of the two cross sections (as in this case), your formula simplifies because you can cancel out a few factors (like the normalization over a given flux, that you usually need for estimating a cross section) which have a large uncertainty and which are absolutely the same for NC and CC.
This is quite standard practice: ratios are very often much more precise than numerator or denominator alone.

5. Luboš Motl - March 11, 2009

Dear Singlino, for another example of what Andrea is saying, see e.g. this near-extremal rotating black hole GRS 1915+105

http://arxiv.org/abs/astro-ph/0606076

Its ratio “a*” of the angular momentum and the squared mass is between 0.98 and 1.00 (the bound!), nevertheless, the mass is only known up to a factor of two or so.

Again, the overall normalization may be very uncertain but ratios are kind of “intensive” quantities that may be more accurate. Similarly with your eyes, you can distinguish a mostly red light from a mostly blue light, even if you’re not able to estimate the intensity of the light itself.😉

Tommaso, do the artists in McDonald’s actually draw these “telescopes”?😉 Just kidding.

6. Singlino - March 13, 2009

Hi,
what i meant is the following
instead of

R=0.34\pm 0.023 \pm 0.031

I thought it shold have been something like

R=0.34x\pm 0.023 \pm 0.031

where x is some integer 0-9

7. Singlino - March 13, 2009

Hi Lubos,
does our mind take ratios of some kind which eleminate the uncertainties in intensities? I am not sure I fully understood the color analogy, could you elaborate a bit?


Sorry comments are closed for this entry

%d bloggers like this: