arXiv:hep-ph/0009044v1 5 Sep 2000 bu htw hudd nteftr.Iwl be- neutrinos. abruptly will switch extragalactic I then to and future. neutrinos thinking the solar in our with do gin guide should we help of what may development about which the subject of our aspects those emphasize them in discuss topics not today. will other I many but say astrophysics, and to neutrino things subjects, grand in these of nor about lots nucleosynthesis are role Bang There the Big discuss cosmology. not in also neutrinos that will of I stars calcula- for exploding. not the processes are cooling in neutrino and of enor- neutrinos detec- tions the supernova and about of prediction anything tion the say in not achievements mous will I neu- extragalactic trinos. and neutrinos solar topics: cific future. the for not goals did our I about if anything even spoke say and I fast, if unintelligibly even fast, cover very to too material just is important There much century. 20th neutrino twenty- the of of in all astrophysics intelligently impossible discuss be to would minutes five so it to that worried. realized say really I I got should as I I friends, But, knowledgeable what many accepted. about readily thinking Feel- I started Beyond.” honored, Astro- and very Neutrino Century ing “ 20th the subject in the physics on talks opening Introduction 1. nex the in neutrinos solar with and neutrinos extragalactic www.sns.ias.edu/ E-mail: a Lecturer Centennial IUPAP BAHCALL, N. BEYOND J. AND CENTURY 20th NEUTRINOS: ASTROPHYSICAL nttt o dacdSuy rneo,N 84,USA 08540, NJ Princeton, Study, Advanced for Institute umrz h rtfu eae fslrnurn research neutrino solar of decades four first the summarize I iltk oehthsoia prahand approach historical somewhat a take will I spe- two to remarks my limit to decided I So, the of one give to McDonald Art by asked was I ∼ jnb a decade. t da beti h u. iuecuts fS. of courtesy Figure an sun.’ possibilities, . the detection Bilenky is of object view ideal of ‘From point 1967: in the wrote Pontecorvo Bruno 1. Figure n ugs htmyb osbet er with learn to possible be may what suggest and 1 2

2. Solar neutrinos 2.1. Bruno Pontecorvo and Ray Davis I want to begin by paying tribute to two of the great scientists and pioneers of neutrino as- trophysics, Ray Davis and Bruno Pontecorvo. Bruno first suggested using chlorine as a detec- tor of neutrinos in a Chalk River report writ- ten in 1946. Ray followed Bruno’s suggestion and the careful unpublished feasibility study of Louie Alvarez. Using with care and skill a chlo- rine detector and reactor neutrinos, Ray showed in 1955-1958 that νe andν ¯e were different. About a decade later, Ray first detected solar neutrinos, laying the foundation for the studies we shall hear about today. In 1967, one year before the first results of Ray’s chlorine solar neutrino experiment were an- Figure 2. The energy Spectrum of neutrinos from nounced, Bruno published a prophetic paper en- the pp chain of interactions in the Sun, as pre- titled: ‘Neutrino Experiments and the Problem dicted by the standard solar model. Neutrino of Conservation of Leptonic Charge’ [Zh. Exp. fluxes from continuum sources (such as p − p and Teor. Fiz. 53, 1717 (1967)]. In this paper, Bruno 8B) are given in the units of counts per cm2 per suggested many different experiments that could second. The percentage errors are the calculated test whether leptonic charge was conserved. The 1σ uncertainties in the predicted fluxes. The p−p grandchildren of most of these experiments are chain is responsible for more than 98% of the en- being discussed in this conference, Neutrino 2000. ergy generation in the standard solar model. Neu- Bruno included a short section in his paper that trinos produced in the carbon-nitrogen-oxygen he called ‘Oscillations and Astronomy.’ In this CNO chain are not important energetically and section, Bruno wrote: “From the point of view of are difficult to detect experimentally. The arrows detection possibilities, an ideal object is the sun,” at the top of the figure indicate the energy thresh- What a wonderfully contemporary statement! olds for the ongoing neutrino experiments. This Bruno, like most particle physicists of the spectrum is from BP98: J. N. Bahcall, S. Basu, 1960’s and perhaps 1970’s and 1980’s, did not be- and M. H. Pinsonneault, Phys. Lett. B, 433, 1 lieve astrophysical calculations could be reliable. (1998). He wrote in this same section on oscillations and astronomy: “Unfortunately, the weight of the var- ious thermonuclear reactions in the sun, and the central temperature of the sun are insufficiently What can we learn from this bit of history? well known in order to allow a useful comparison When Ray and I wrote our PRL papers arguing of expected and observed solar neutrinos, from that a chlorine detector of 600 tons could observe the point of view of this article.” [This was 30 solar neutrinos, we never discussed the possibility years before the precise confirmation of the stan- of using neutrinos to learn about particle physics. dard solar model by helioseismology.] To support The only motivation we gave was “...to see into his claim, Bruno referenced only his 1946 Chalk the interior of a star and thus verify directly the River report, which mentioned the sun in just two hypothesis of nuclear energy generation in stars.” sentences. Bruno did cite our calculations of the [PRL 12, 300 (1964)]. solar neutrino fluxes elsewhere in his 1967 paper, Why did we not discuss using neutrinos for par- but they seem not to have affected his thinking. ticle physics? Frankly, because we never thought 3 about it. And even if we had, we would have eigenfrequencies. known better than to mention it to our parti- cle physics friends. Bruno had the insight and the vision and indeed the courage to argue that astronomical neutrinos could potentially give us unique information about neutrino characteris- tics. His paper is all the more remarkable because it was published a year before the first results of the chlorine experiment showed that the rate Ray observed was less than our calculated rate. We learn from these events that pioneering ex- periments can lead to important results in areas that are unanticipated. We will come back to this conclusion at the end of this talk.

2.2. Standard Model Predictions Figure 2 shows the calculated solar neutrino spectrum predicted by the Standard solar model. The percentage errors are the calculated 1σ un- certainties in the predicted fluxes, based upon Figure 3. Comparison of measured rates and the published errors of the measured quantities standard-model (BP98) predictions for six so- and on many calculations of standard solar mod- lar neutrino experiments. The unit for the ra- els. As you will hear from the talks in the later diochemical experiments (chlorine and gallium) parts of this morning session, the total intensities is SNU (10−36 interactions per target atom per and the energy spectra shown in Fig. 2 are now sec); the unit for the water-Cerenkov experiments widely used to interpret, and indeed to plan, so- (Kamiokande and Super-Kamiokande) is the rate lar neutrino experiments such as those discussed predicted by the standard solar model plus stan- in today’s sessions: chlorine, Super-Kamiokande, dard electroweak theory. The experimental re- SNO, SAGE, GALLEX, GNO, and BOREXINO. sults are described by Lande, Suzuki, Gavrin, and Figure 3 compares the calculated versus the Belotti in these proceedings. measured rates for the six solar neutrino experi- ments for which results have been reported. As- suming nothing happens to the neutrinos after they are created, the measured rates range from Could the solar model calculations be wrong by 33% ±5% of the calculated rate (for chlorine) to enough to explain the discrepancies between pre- 58% ±7%. As is now well known, the observed dictions and measurements shown in Fig. 3? He- rates cannot be fit (at a C.L. of about 99%) with lioseismology, which confirms predictions of the any linear combination of undistorted solar neu- standard solar model to high precision, suggests trino energy spectra. that the answer is “No.” Today we know that there are three reasons Figure 4 shows the fractional differences be- that the calculations of solar neutrino fluxes tween the most accurate available sound speeds are robust: 1) the availability of precision mea- measured by helioseismology and sound speeds surements and precision calculations of input calculated with our best solar model (with no data that have been gradually refined over four free parameters). The horizontal line corresponds decades; 2) the intimate connection between neu- to the hypothetical case in which the model pre- trino fluxes and the measured solar luminosity; dictions exactly match the observed values. The and 3) the measurement of the helioseismological rms fractional difference between the calculated frequencies of the solar pressure-mode (p-mode) and the measured sound speeds is 1.1 × 10−3 for 4 the entire region over which the sound speeds are trinos are present. Remember, we believe we can measured, 0.05R⊙

2.3. Summing up and looking ahead I want now to look back and then look ahead. I will begin by giving my view of the principal ac- complishments of solar neutrino research to date (Sect. 2.3.1). Then I will discuss two of the ex- pected highlights of the next decade of solar neu- trino research, the measurement of the neutral current to charge current ratio for 8B neutrinos ((Sect. 2.3.2) and the detection of solar neutrinos with energies less than 1 MeV ((Sect. 2.3.3).

Figure 4. Predicted versus Measured Sound 2.3.1. Principal achievements Speeds. This figure shows the excellent agree- What are the principal achievements of the first ment between the calculated (solar model BP98, four decades of solar neutrino research? I give Model) and the measured (Sun) sound speeds, a below my personal list of the ‘top three achieve- fractional difference of 0.001 rms for all speeds ments.’ • Solar neutrinos have been detected. measured between 0.05R⊙ and 0.95R⊙. The vertical scale is chosen so as to emphasize that The chlorine, Kamiokande, Super-Kamiokande, the fractional error is much smaller than generic GALLEX, SAGE, GNO, and SNO experiments changes in the model, 0.09, that might signifi- have all measured solar neutrino events. This is cantly affect the solar neutrino predictions. The the most important achievement. The detection measured sound speeds are from S. Basu et al., of solar neutrinos shows empirically that the sun Mon. Not. R. Astron. Soc. 292, 234 (1997); The shines by the fusion of light elements. • figure is taken from BP98. Evidence for new physics has been found. For more than thirty years, beginning with the fact that Ray’s first measurements in 1968 indicated a flux lower than the standard The arrow in Fig. 4 shows how different the model predictions, we have had evidence for new solar model sound speeds would have to be from physics in the solar neutrino arena. This evi- the observed sound speeds if one wanted to use dence has steadily deepened as new solar neu- solar physics to reduce the 7Be neutrino flux. The trino experiments have confirmed and extended position of the arrow is fixed by artificially reduc- the neutrino discrepancies and helioseismology ing the predicted 7Be neutrino flux that is not has confirmed the standard solar model. The observed in the gallium experiments, SAGE and fact that neutrino oscillations have now been ob- GALLEX plus GNO (see Fig. 3). if the p−p neu- served in atmospheric neutrino phenomena fur- 5 ther strengthens the case that oscillations occur for solar neutrinos. We are still looking for a ‘smoking gun’ single effect that shows up in just one solar neutrino experiment, rather than com- bining the results of two or more different exper- iments. I will discuss some possibilities below. • Neutrino fluxes and energy spectra are approximately as predicted by the stan- dard solar model. If you had told me in 1964 that six solar neutrino experiments would give re- sults within a factor of three of the predicted stan- dard model results, I would have been astonished and delighted. This is especially so considering that the crucial 8B neutrino flux depends upon the 25th power of the central temperature of the sun. This agreement exists without making any corrections for neutrino oscillations. If we correct the observed solar neutrino event Figure 5. The neutral current to charged current rates for the effects of neutrino oscillations us- double ratio, [NC]/[CC] . The standard model value ing the six currently allowed two-neutrino oscil- for [NC]/[CC] is 1.0. The figure shows, for a 5 MeV lation scenarios, the inferred 8B neutrino flux at threshold for the CC measurement, the predicted the source is rather close to the best-estimate pre- double ratio of Neutral Current to Charged Current dicted flux. At the 99% CL, one infers (see hep- for different neutrino scenarios. The solid error bars ph/9911248): represent the 99% C.L. for the allowed regions of the six currently favored neutrino oscillation solutions. 8 The dashed error bar labeled “Measure 3σ” repre- 0.55 ≤ φ( B)/(Standard prediction) ≤ 1.32, (1) sents the net estimated uncertainty in interpreting the measurements, including the energy resolution, 8 which is a slightly tighter range than the 3σ pre- energy scale, B neutrino energy spectrum, neutrino diction of the standard solar model. cross section, counting statistics, and the hep flux. This is Fig. 7a of Bahcall, Krastev, and Smirnov, 2.3.2. SNO and the [NC]/[CC] ratio hep-ph/0002293. Figure 5 shows the predictions of the currently allowed neutrino oscillation solutions for the dou- ble ratio, [NC]/[CC], of neutral current to charged current event rates in the deuterium detector 2.3.3. Solar neutrinos below 1 MeV SNO. Art McDonald will describe later this morn- More than 98% of the calculated standard ing the experimental characteristics of this great model solar neutrino flux lies below 1 MeV. The 8 observatory and outline for us the extensive pro- rare B neutrino flux is the only solar neutrino gram of SNO measurements. The important mes- source for which measurements of the energy have 8 sage of Fig. 5 is that all of the currently allowed been made, but B neutrinos constitute a fraction −4 oscillation solutions for active neutrinos predict a of less than 10 of the total solar neutrino flux. value for the double ratio that is different from The great challenge of solar neutrino astron- the no oscillation value of 1.0 by at least nine omy is to measure neutrino fluxes below 1 MeV. times the estimated non-statistical measurement We must develop experiments that will measure 7 uncertainty. the Be neutrinos (energy of 0.86 MeV) and the We all eagerly look forward to this crucial and fundamental p-p neutrinos (< 0.43 MeV). A num- decisive measurement. ber of promising possibilities were discussed at the LowNu workshop that preceded this confer- 6 ence. The BOREXINO observatory, which can detect ν − e scattering, is the only approved solar neutrino experiment which can measure energies less than 1 MeV. The p-p neutrinos are overwhelmingly the most abundant source of solar neutrinos, carrying about 91% of the total flux according to the stan- dard solar model. The 7Be neutrinos constitute about 7% of the total standard model flux. ½ We want to test and to understand neutrino oscillations with high precision using solar neu- trino sources. To do so, we have to measure the neutrino-electron scattering rate with 7Be neutri- nos, as will be done with the BOREXINO experi- ment, and also the CC (neutrino-absorption) rate with 7Be neutrinos (no approved experiment). With a neutrino line as provided by 7Be electron- Figure 6. Survival probabilities for MSW solu- capture in the sun, unique and unambiguous tests tions. The figure presents the yearly-averaged of neutrino oscillation models can be carried out survival probabilities for an electron neutrino that if one measures both the charged-current and the is created in the sun to remain an electron neu- neutral current reaction rates. trino upon arrival at the Super-Kamiokande de- I believe that we have calculated the flux of p-p tector. neutrinos produced in the sun to an accuracy of ±1%. This belief should be tested experimentally. Unfortunately, we do not yet have a direct mea- ergy region, ∼ 7 MeV, where the Kamiokande surement of this flux. The gallium experiments and Super-Kamiokande data are best. The sur- only tell us the rate of capture of all neutrinos vival probability shows a strong change with en- with energies above 0.23 MeV. ergy below 1 MeV for all the solutions, whereas The most urgent need for solar neutrino re- in the region above 5 MeV (accessible to Super- search is to develop a practical experiment to Kamiokande and to SNO) the energy dependence measure directly the p-p neutrino flux and the of the survival probability is at best modest. energy spectrum of electrons produced by weak The p-p neutrinos are the gold ring of solar neu- interactions with p-p neutrinos. Such an exper- trino astronomy. Their measurement will con- iment can be used to test the precise and fun- stitute a simultaneous and critical test of stellar damental standard solar model prediction of the evolution theory and of neutrino oscillation solu- p-p neutrino flux. Moreover, the currently favored tions. neutrino oscillation solutions all predict a strong influence of oscillations on the low-energy flux of 3. Extragalactic neutrinos νe. Figure 6 shows the calculated neutrino sur- Experimentalists often like to describe the vival probability as a function of energy for three power of their experiments in terms of the ex- global best-fit MSW oscillation solutions. You pected or observed number of events per year and can see directly from this figure why we need ac- L/E, where L is the distance between the accel- curate measurements for the p-p and 7Be neu- erator and the detector and E is the beam energy. trinos. The currently favored solutions exhibit The quantity L/E determines, together with the their most characteristic and strongly energy de- square of the mass difference, the survival proba- pendent features below 1 MeV. Naturally, all of bility for vacuum neutrino oscillations. More gen- the solutions give similar predictions in the en- erally, L/E represents the time of flight in the rest 7

bilities for detecting GRB neutrinos, which will be discussed in more detail in these proceedings by Eli Waxman. I believe that GRBs offer the best chance for detecting extragalactic neutrinos among all the known sources of astronomical pho- tons. The phenomenology of the photons observed from gamma-ray bursts is now relatively well un- derstood. Many different types of observations have been carried out and the results are well summarized by the expanding fireball model. Us- ing this model, one can work out the flux of neu- trinos from shocks. Figure 8 shows the neutrino energy spectra that Waxman and I have estimated to be produced by GRBs, both from the direct burst (energies ∼ 106 GeV) and from the afterglow (energies ∼ 108 19 Figure 7. Very longbaseline neutrino oscillation ex- GeV to ∼ 10 GeV). The observed population of 2 periments. The figure shows that experiments such GRBs should give rise to ∼ 10 events per km per as ANTARES, BAIKAL, ICECUBE, and NESTOR, year from neutrinos with characteristic energies which may detect high-energy neutrinos from dis- of order 1014 eV. We shall hear on the last day tant gamma-ray bursts, have extraordinary sensitiv- 5 of this conference that the calculated GRB flux ity to vacuum neutrino oscillations. Neutrinos of 10 may be detectable in ANTARES, ICECUBE, or GeV from gamma-ray bursts located at cosmolog- NESTOR. The fundamental assumption used in ical distances were used to locate the positions of calculating the GRB neutrino flux is that GRBs ANTARES, BAIKAL, ICECUBE, and NESTOR in produce the observed flux of high-energy cosmic the figure. rays, an assumption for which Eli Waxman has provided a strong plausibility argument. GRBs occur at modest to large redshifts. We frame of the particle, the time for rare events to know the time of the explosion to an accuracy occur. ∼ 10 sec (from the gamma rays). Therefore, Figure 7 shows the extraordinary sensitiv- GRBs can be used to test special relativity to ity to neutrino oscillation of experiments like an accuracy of 1 part in 1016 and to test the ANTARES, BAIKAL, ICECUBE, and NESTOR weak equivalence principle to an accuracy of 1 that can detect neutrinos from distant extragalac- part in 106. If special relativity is right, the pho- tic sources. The accelerator experiments that will tons and the neutrinos should arrive at the same be discussed at Neutrino 2000 lie in the left-hand time (to an accuracy of about 10 sec, the dura- side of Fig. 7, L/E < 104 km/GeV . Solar neu- tion of the burst). If the weak equivalence princi- trino experiments like Super-Kamiokande, SNO, ple is valid, the arrival times of neutrinos (which and BOREXINO can reach to 1010 km/GeV traverse significant gravitational potentials) from and, for the lower energy experiments, even 1011 distant sources should be independent of neutrino km/GeV. Extragalactic sources such as gamma- flavor. ray-bursts (GRBs) have such a long baseline (∼ GRBs can also be used to probe the weak in- 1010 lyrs) that the new generation of extragalactic teractions to an extraordinary level of precision. experiments, ANTARES, BAIKAL, ICECUBE, Gamma-ray bursts are expected to produce only and NESTOR will extend to the right-hand side νe and νµ. The large area detectors of extra- of Fig. 7, to L/E > 1018 km/GeV. galactic neutrinos are in principle sensitive to vac- I want to say a few words about the possi- uum neutrino oscillations with mass differences as 8

−5 0 10 HZ97 ting π decay to the observed gamma-ray back- M95B ground. −6 P97 10

−7 4. Goals for Astrophysical neutrinos: 10 WB Limit 2000-2010 s sr] 2 −8 10 GRB (burst) It seems to me that we have three principal −9 goals for this next decade. 10 [GeV/cm ν Φ

2 ν −10

E 10 • Determine the mixing angles −11 10 and mass differences that are GRB (afterglow) −12 important for solar neutrino 10 4 6 8 10 10 10 10 10 phenomena. E [GeV] ν • Test precisely stellar evolution by observing p − p and 7Be neutrinos, and by determining 8 Figure 8. The Waxman-Bahcall upper bound on the total flux of B neutrinos. muon neutrino intensities (νµ +ν ¯µ). This figure • Discover extragalactic is from Bahcall and Waxman, hep-ph/9902383. neutrinos, perhaps from The numerical value of the bound assumes that gamma-ray bursts. 100% of the energy of protons is lost to π+ and π0 and that the π+ all decay to muons that also From time to time, friends ask me to compare produce neutrinos. The dot–dash line gives the the search for solar neutrinos with the search for upper bound corrected for neutrino energy loss neutrinos from GRBs. They are very different. due to redshift and for the maximum known evo- From photon studies, we know more observation- lution (QSO or star-formation evolution). The ally about the sun than about any other astro- lower line is obtained assuming no evolution. The nomical source, certainly much more than about solid curves show the predictions of representa- the mysterious GRBs. Moreover the sun is in the tive AGN jet models taken from the earlier pa- simplest stage of stellar evolution, in quasi-static pers of Mannheim (marked M95B in the figure), equilibrium with a characteristic time scale for 9 16 Protheroe (P97), and Halzen and Zas (HZ97). evolution of 10 yr (10 s ). We do not even The AGN models were normalized so that the know the energy source of GRBs. We do know calculated gamma-ray flux from π0 decay fits the that GRBs are far from equilibrium, evolving ex- −3 observed gamma-ray background. plosively on a time scale of order 10 s. We want to do extragalactic neutrino astron- omy because it is truly an exploration of the uni- verse. We do solar neutrino astronomy to test 2 −17 2 small as ∆m ≥ 10 eV (from νµ → ντ ). fundamental theories of physics and astronomy. Not everything is encouraging in Fig. 8. The But, perhaps solar neutrino research and extra- figure also shows the upper limit that is allowed galactic neutrino research may in the end share a for astrophysical neutrino production from (γ, π) fundamental characteristic: surprise. Remember, interactions on high energy protons. The upper that we undertook solar neutrino research to test limit is established by using the observed cos- stellar evolution and unexpectedly (at least for mic ray flux of high energy protons. Prior to everybody except Bruno Pontecorvo) we found the recognition of this limit a number of authors evidence for new neutrino physics. had suggested much more optimistic models (also In a sense, we are returning to our original shown in the figure), that were normalized by fit- goal in neutrino astronomy, but by a round-about 9 path. We must first understand neutrino oscilla- tion phenomena in order to be able to use solar neutrino observations to test precisely the the- ory of stellar evolution, our original goal. Per- haps with extragalactic astronomy we will partici- pate in a similar cycle of astronomical exploration and physical clarification. T. S. Elliot in ‘The Four Quartets’ described the cycle succinctly and beautifully:

We shall not cease from exploration And the end of all of our exploring Will be to arrive where we started And know the place for the first time.

Acknowledgments I acknowledge support from NSF grant #PHY95-13835.