IAP weekly specialised seminars / 26 January 2024
Sonia El Hedri (Laboratoire AstroParticule et Comologie, Université Paris Cité, France)
Core-collapse supernovae (CCSNe) are important actors in the dynamics of galaxies but their underlying mechanism is still only partially understood. The detection of neutrinos from SN1987A confirmed that these particles play a key role in supernova explosions and that their detection would provide a complete recording of the behavior of the core of the star during the collapse. Moreover, CCSN neutrinos could be detected minutes to hours before the electromagnetic signal, thus providing an advance warning for telescopes. However, the CCSN detection range of current and upcoming neutrino experiments only covers our Galaxy and its immediate neighborhood, where supernovae take place only 2-3 times per century. The rarity and unpredictability of close-by supernovae has motivated the development of CCSN search strategies at a wide variety of neutrino detectors, sometimes extending their nominal energy range. Furthermore, a centralized real time analysis system, the Supernova Early Warning System (SNEWS), has been developed in order to combine observations from multiple experiments and send relevant information to telescopes within minutes after a CCSN neutrino detection. This colloquium will illustrate the current challenges of CCSN neutrino detection through two topics. First, I will describe the CCSN search program of the KM3NeT neutrino experiment, originally aimed at studying neutrino oscillations and high-energy neutrinos from cosmic ray sources. Second, I will describe how to locate a supernova using neutrino observations, and will show how exploiting novel features of next-generation neutrino and dark matter experiments could improve SNEWS’s current strategies.
Okay uh good morning everybody so let’s start uh today we have a pleasure of welcoming Sonia elri so uh Sonia did her uh PhD in Stanford in slack on Dark Matter phenomenology uh then she continued uh with postd ducts in Gutenberg University in mines and NF in Amsterdam still working on Dark Matter
Phenomenology uh so she comes from particle astrophysics and then she came back to France uh to do a p in labor l in Palo and then she switch to working on nutrino astrophysics uh the field in which she continues to working now and in uh 2021 she was recruited as a s
Researcher in APC so not far from here she’s is a member of the antaris and Kret uh collaborations and she’s working on N trino astrophysics and specifically astrophysical sources of nutrino of nrus for example core collap Supernova on which he’s going to talk uh today so thank you
Sonia uh thank you Arena uh is the microphone working perfect so um thank you everyone for joining and good morning uh Asa said I’m going to talk to you about how to get straight to the core of collapsing star using neutrinos about why it’s quite difficult and why you should do it
Anyway a collapse Supernova is a slight not working of course it has to a COR collap Supernova is what happens at the end of life of a massive star of a heavy star more than eight solar masses after it has gone through its uh main sequence nuclear fusion steps
As well as all the like nuclear fusion steps of hydrogen into helium as well as all the subsequent steps involving heavier and heavier elements down all the way down to iron so at this point of its life the Dynamics of the star is quite difficult to prop directly because the radius of
The star itself can reach up to hundreds of million kilometers While most of the action happens in this car which is as you can see the size of a planet however we can get uh we can infer what happens in the core of the Star based on
The laws of physics in particular we can infer that after uh we have reached iron nuclear fusion cannot counter um gravitation gravitational collapse anymore in lighter Stars gravitational collapse can still be countered by the firmy pressure of electrons in the core of the star however Beyond this eight
Solar mass threshold this pressure is reduced by the fact that electrons get captur on protons at this point nothing hinders the collapse of the cor the St or gravitational forces and you get a supernova the Supernova itself is a rapid succession of extremely complex phenomena taking uh a fraction of a
Second total so as I said it’s extremely difficult to observe however we can uh understand what happens during a supernova fr simulations and in particular a robust prediction of simulations is that supanova occurs in three step the first step is the formation of what we call a proton
Neutron star due to collapsing matter that is an extremely dense region inside the core where nucleons are packed against each other a density 10 to the 12 gram per cimer cubed this region is so dense that the rest of the collap collapsing matter is going to bounce on
It leading to the formation of a shock wave which propagates outwards this is the official beginning of of the Supernova this phase lasts a few tens of milliseconds at that point you might think okay the star is just going to explode but actually in most simulations
The shock wave is going to stop deep inside the core like this and the star enters in what is called the accretion phase where matter continues collapsing onto the proton neutron star passing through the shock wave while matter is passing through the shock wave a nuclei get dissociated into three nucleons so
Around the proton neutron star you have this extremely hot Mel with a temperature of 100 billion degrees uh and three nucleons flying around this is what we call in scientific terms a gigantic mess so this what is going to happen in this gigantic mess is going to
Determine whether the star under goes a complete collapse into a black hole or whether somehow the shock wave restarts and you get an explosion which leaves behind a neutron star notice the somehow here because in fact the behavior of this mantal is so complex that in current Supernova simulations it is
Extremely hard to get an explosion and most of the times the explosions you are going to get are not as powerful as what is actually being observed in the cosmos as I mentioned observing directly the core of the star using poons is not possible but and here you see what I’m
Coming at during a supernova um you expect the emission of a short but also extremely intense burst of neutrinos these neutrinos are expected to play a key role in the explosion by depositing a small fraction a few percent of their energy below the shock wave creating a kind of a pressure cooker
Effect in addition to being playing a key role in the explosion nutrino rates and energy Spectra could give you a complete recording of what is happening in the core of the star during the Supernova to understand this let’s retrace the different steps I exposed earlier but from the point of view of
Nutrino production uh well the idea is that the core of the star is too dense for the shock wave to be able to lift all the upper layers of the star this is a matter of this is a problem of mechanics like the star is pushing like it’s trying to counter the
Gravitational collapse and it’s not succeeding at it so you need to find an additional source of energy to power this shock wave and this is a combination of complex Factor like nutrio energy deposition and hydrodynamical instability that could bring concentrate the energy in one point for example but in general you
Don’t have one factor you have several of them um so let’s start for neutral production let’s start with the initial shock propagation when it starts propagating the shock wave is going to dissociate nuclei into free nucleons in particular you’re going to have three protons which will capture the electrons that are flying around
This leads to the production of an intense burst of electron neutrinos with this very characteristic Peak then during the accretion phase you have a similar captured phenomenon but this time this capture is going to involve both electrons and positrons because of this PA production mechanism due to proton conversion in the hot mle
So now you are going to have production dominantly of electron neutrinos and electron anti neutrinos and as you see this production has a very complex profile if you look at the Luminosity of neutrinos it’s correlated with the accretion rate of matter onto the core and you have this oscillatory profiles
Which with oscillations which correspond to hydrodynamical instabilities I can tell you more about it later uh and then you start having these other nutrino flavors like M and to that we write using an X here and they come from the deepest regions of the star which are cooling down through neutral
Current interactions you have a nucleon here producing pairs of neutrinos and anti nutrino of all flavors this equal production of all flavors through cooling can be seen after an explosion when all is Left Behind is the neutron star and you have this exponentially um decreasing profile so the scale here is not the
Same so as you can see we have gone from having this Photon information from the explosion which is extremely indirect to having a direct record and complete recording of the core of the with millisecond Precision on three channels corresponding to this three neutral species this is extremely informative so
How many times did we actually make this observation uh the answer is once and I’m not going to tell you how many years ago because this was my year of birth so in 1987 uh star exploded in the large maelani Cloud around 50 kilo par away
From us you can see the explosion here invisible light and this uh explosion was associated with the observation of 25 neutrinos in three experiments cand 2 IMB and ban 25 neutrinos in three experiments is really not a lot but from their energy and time profiles we can
Already um learn a few important things about the behavior of the star for example you see that this time profile extends up to around 10 or 11 seconds which is consistent with the existence of a cooling phase and here you have this High concentration of higher concentration of neutrinos consistent
With the existence of this acction phase so here we demonstrate that there indeed Supernova for which the sock wave is stalling inside the core in addition looking at the energies neutrinos are thermal and they T of MV energies reflect the temperature of the star score in addition to this interesting
Physics uh SN 1987 a gives us a few lesson on how to U uh to observe the next galactic Supernova using the next Supernova using neutrinos the first lesson is that current and upcoming nutrino experiments will be mainly sensitive to supern no occurring in our immediate neighborhood our galaxy or the
Large marinan cloud and these supern no are expected to occur only two or three times per Century in addition to being extremely rare this phenomenon is extremely hard to predict and this can have quite scary consequences for example here on the right you can see an event record of um
Neutrinos at cand 2 around the time of SN 1987a this event record has the peak here the Nobel prize winning Peak but it also has a gap here that occurred around two minutes before this peak this Gap corresponding to a schedule detector maintenance and the fact that this
Maintenance did not occur at the time the neutros arrived is pure luck we had no way to know so the lesson here is that the best experiment to observe super neutrinos is not the one that is the biggest it’s not the one that has the highest resolution
It’s the one that is actually working when your Supernova neutrinos are arriving from this observation we can um we can infer the current Supernova observation strategy which uh rests on two requirements the first requirement is to involve as many experiments as possible even experiments that have not been made initially to
Detect supera neutrinos and even experiments that have not been made initially to detect neutrinos at all like for example dark matter experiments in addition to this requirement we also require that these experiments communicate with each other in real time analyze their data as fast as possible and communicate their
Findings to telescopes and I’m going to expand on this requirement more during this presentation this is why right now we’re having uh experiment all all over the world having this Supernova nuto detection program I’m only showing a few examples here and these experiments are sending their data in real time to what
We call the Supernova early warning system which is a centralized uh neutrino data analysis Network you can see the website here we call it SNS so during this presentation I’m going to uh present illustrate a bit these two aspects of the Supernova nutrino observation strategy by presenting how to detect Anova neutrinos
In an experiment that was not initially designed for it the C freet nutrino detector and by showing you how to how and why to pull the together the data from several different nuto experiments including Next Generation experiments let me start with the detection of supanova neutrinos in the cfret
Experiment camet is a neutrino detector an extremely large scale neutrino detector located in the Mediterranean why is the Mediterranean Sea because Kret is looking for this blue chair rank of light emitted by relativ sck charged particles traveling in water including above all neutrino interaction products this light is detected by these photo
Multipliers which are arranged in spherical structured called digital optical modules so dog I’m going to talk a lot about it so I’m going to use this acronym A Lot these dos are arranged along with lines these strings here and the lines are put together to form a
Gigantic detection array and here on the right you can see a video of an animation of what is happening to this array like you have a charge particle passing by and the photo multipliers being activated as they receive the light this structure has two important advantages first it is highly modular it
Means that you can start taking taking data even when your detector is not finished you can have a few lines installed somewhere and then in install a few other ones um while you are taking data and the second Advantage is that just by varying the spacing between the
Optical modules you can be sensitive to different energy ranges in practice Kret exploits this fact by having two different detectors two different arrays Orca of the two long coast and ARA near sic here you see schematic Vision to scale of what the ARA and orca arrays
Are going to look like what is important to know here is that each of these blocks here called bending blocks has the same number of optical modules so what is changing between this one and this one is the spacing Ora is well as you can see the densest detector it’s sensitive to what
We call low energy which is like GV scale neutrinos from uh produced in the atmosphere and it’s for oscillation studies ARA is the least dense array it’s targeting much higher energies and it’s aimed at the detection of astrophysical neutrinos a final build a building block is going to be 115 lines in hopefully
2027 right now about 12% of the detector is in place 28 lines in ARA 18 lines in Orca and periodically as you can see in the video above we are going out with a boat I mean we colleagues are going out with a boat and they are loading new
Lines so here is a picture of the line being loaded uh since a line can be very long it can be one kilometer long it’s rolled up inside this sphere if you look um you can see the optical modules uh here so this spere is anchored to the ground
With this yellow structure where you do all the branching cabling Etc and it’s lowered to the sea floor using a crane and uh on the sea floor there is a robot which plugs in everything and unforce the sphere so we take the the scaffolding back and the line is just
Floating so Kret is an impressive structure but it has one problem for Supernova neutrino detection which is that its lower energy threshold it’s still several orders of magnitude higher than the t of M of superan neutrinos so to try to detect superan neutrinos at Kret we have to go off the beaten path
So first let’s consider let’s notice that the dominant neutrino interaction mode for these energies is the inverse pad decay of electron anti neutrinos and protons so who says Supernova detection with water chering of detector says electron anti neutrinos this inverse beta DEC produces among other things oppositon which emits
Chank of light first good news the difficulty here is that at this energies chank amount of chank of light emitted will be so localized that in a best case scenario it will activate a single is digital optical module so at the scale of CET at Zero’s order what we get is a
Single blip without reconstruction being possible this is not necessarily a problem because the Supernova signal is expected to be concentrated in time around half a second and uh it’s expected to be extremely intense so what we could see is a rise of the number of detected of activated Optical modules
Above the ambient nose and this rise and the shape of this rise can be actually informative already you can learn about the start of the Supernova its outcome the existence of hydrodynamical instabilities and the rate of the collapse for example however whether or not you are going to get this
Information depends on the level and stability of this noise down there and the difficulty here is that we are in a natural environment that is extremely difficult to control also it is darker and less lively than here so when faced with an extremely difficult question
Like this I tried to be modern and first asked CH GPT what it was thinking about Supernova nutrino detection in the Mediterranean Sea and am gave me quite creative answers first it noticed that the Mediterranean Sea is located near several regions of the sky that are expected to contain many Supernova
Explosions so bring sunscreen I guess uh however while the mediteran sea is a large body of water it does not have the detectors needed to Det superv neutrinos so too bad and additionally it’s hard to detect neutrinos there because the sea is not located underground also I think
Being located under water is pretty nice uh so AI has declared defeat and I went back to science which told me that at the energies of super neutrinos there are two main types of backgrounds radioactivity mainly from potasium 4 decays in Sea Water and Light from living organisms so this these two backgrounds
Total 200 Kilz noise per Optical module this is huge and the bioluminesence background is unstable so here you can see an event record at the anes experiment the the precursor of CET which so this short burst on time scales which are relatively similar to the Supernova neutral emission time
Scales however well even with the stable components that you can see on this plot when or and ARA will be finished we expect to see 600 ion background events in a supernova analysis time window so that is what nature brings you I guess so here we’ll have to be a bit more
Subtle we’ll have to do a bit of reconstruction after all and to do reconstruction we are going to exploit the fact that a CET Optical module is actually made of 31 3inch photo multipliers since these photo multipliers are close to each other within a sphere of 44 CM diameter it is
Definitely possible for the chank of life of supernova neutrino interaction products to activate several photo multipliers so instead of having a single blip that you cannot reconstruct you have a pattern which you can analyze and the simplest thing you can actually do is to count the number of
Auto multiplier Heats within a Time window of a few T of nanc to ensure that they all come from the same particle so this is an observable we call the multiplicity and this multiplicity is ploted here uh Multiplicity distribution is ploted here uh for data you can see the black points
I’m sorry I have lost my mouse and uh for background simulations the first good news is that for the backgrounds that we have been worried about by luminiscence and radioactivity are mainly concentrated at the single heat level when we increase multiplicity when we like go higher like this radioactivity backgrounds Falls
Extremely sharply this fall is mitigated at by the fact that at some point you run into backgrounds from meons produced by Cosmic interactions in the atmosphere however this muons are high energy in general so they leave very obvious signatures in the detector you can see it on the right here with each
Ball representing a digital optical module they live is tracks and these tracks are most of the times Well identified by cem fret triggers high energy triggers these findings form the two main ingredients of the published Supernova analysis at CET uh this analysis uh rests mainly on
The existence of a muon veto where we remove the optical modules associated with Cina triggers in time and space and on a selection on this multiplicity observable here we defined it as a number of Heats in a 10 nond window so we narrowed it a bit compared to what I explained
Before you can see here on the right uh background and Signal distributions of this multiplicity in a final KET building block of 115 lines um so here we have A5 second analysis window uh the blue light blue triangles show what you expect as a background in Orca so dark blue
Triangles are what you expect in ARA and then the field histograms are for three different signals with three different progenitor masses in units of solar masses this is at 10 kilo second away from us so here you can see there is a definite switch spot and we can characterize this switch spot by
Maximizing the detection Horizon of CET and we find that with this switch spot is a region of with multiplicities 7 to 11 what is the sensitivity associated with this me Vito plus multiplicity selection well here this plot shows the multiplicity as a function of the distance to us in kilop par Supernova
Distance to us uh the sensitivity as a function of the distance uh for the three different models so the green one the 11 solar mass represent a quite pessimistic model but also quite realistic since light progenitors are more frequent the gray one with 27 solar masses represents quite an optimistic
Model and uh the pink one represents uh an extreme model where you have an extremely heavy progenitor and black hole production you see that for extreme models we can reach up to the large maganic cloud and in all when we introduce when we integrate over prior
And the super no distance and a prior on the progenitor Mass we get five Sigma sensitivity to 96% of potential Galactic Supernova so here CET can definitely be an actor in the hunting for the next Galactic Supernova however there is a bit of room for improvement not only to turn this
96% into 100% not only to take to gain sensitivity in the large maganic Cloud but also in order to improve our current sens itivity because there is only 12% of the detector in place and a supernova could occur tomorrow for what we know so to improve research there is one
Uh clear path forward which is to go beyond merely counting photo multiplier Heats and instead like analyzing like the shape of this pattern or its position for example so here this was a work initiated by gwel deas at APC who is now at ucil Lua and I took it over
Working with Isabelle go and miam D who is now in Naples what we did is that we sat down and we started writing as many observables as possible that could help us characterize the pattern of activated photo multipliers so in practice we obtained 20 observables which could be um
Classified into four categories one of them is how intense the overall signal is the second one is the position mean position of the heat cluster on the optical module third one is a spatial extension of your heat cluster and fourth one is the time extension how
Spread out it is in time and then we use Boosty decision trees to both select a minimal set of best observables and maximize the signal or background ratio so both observable selection and signal and classification with this boosted decision we are able to prove that a full description of low energy
Signatures at cfret could be made with five observables two of them are related to the intensity of the signal this is our multiplicity and the total charge deposited and then the three others are in the one in each of the remaining categories so we have one variable for the heat cluster concentration spatial
Concentration one variable for its height on the optical module and another variable for the average time spread of the Heats and these five observables are enough so to show you how they are working and why they are working uh on the right you can see the distribution of the height of the
Heat cluster and of the overall heat concentration for K3 net data taken like a bit a few years earlier and here to show the role of observables other than the multiplicity I have fixed the multiplicity to be seven so we have quite a complex pattern a seven pixel
Image to analyze and here you see that this distribution is definitely Bodel on one end of the distribution you have radioactivity radioactivity is very low energy signal so only vertices close to the optical module are going to be detected which means that we are going to get the beginning of the chank of
Cone so very concentrated signature in addition the bottom of the optical module is more instrumented you can see the cap on the top that is uring the view uh so there is going to be a strong preference for this uh bottom part and additionally you are going to have low
Charge deposited and very few hits muons have completely opposite properties they are coming from above so they are going to leave primarily deposits on the upper hemisphere of the optical module and they have long trajectory so they will deposit a high amount of light and their
Signal is going to be spread out both in space and time and the advantage of having these opposite properties is that in the middle you have a switch spot and the signal fits right into it because it looks like radioactivity its topology is similar to radioactivity but its energy
Is higher so it’s going to be less concentrated and it can creep into the upper part of the optical module this is why these observables are working uh we formalized these kind of qualitative observations by uh incorporating these five obser into a boosted decision Tree in fact we had one
Boosted decision tree for each multiplicity we trained it on a simulation for the signal aant for simulation and on data for background this led to systematics that we dealt with accordingly you can see here that a test of this Boosty decision Tree in an upcoming configuration of the of our
Detectors with 29 lines for ARA and 24 lines for Orca uh I say upcoming it was supposed to be there for icrc 2023 but um it um we we had we had a few delays uh so here you can see the orcan ARA backgrounds and this blue histogram represents the total background you
Expect and then you have our three supernova genitors for a signal at the galactic center So currently at CET we’re able to see any Supernova potential Supernova at the galactic center to have a more global view we also ploted the sensitivity as a function of the distance for this upcoming archan Orca
Configuration here the dash lines are when we use the multiplicity only so the old analysis and the solid line is what we get uh when we introduce our boosted decision trees in practice we have a 23% increase of the distance Horizon at CET with our new method what it means is
That for the most pessimistic but also the most realistic model the amount of stars in the galaxy of potential supern progenitors who are going to probe goes from 33% to 52% so this is actually quite a considerable Improvement and Kret with this new analysis will be able
To even now probe I mean or even in a few months prob the majority of potential core collapse Supernova in the galaxy okay so now we have shown that K freet even now or even like next year could probe um most of supernova in the
Galaxy and could be an actor in The Hunt for the next Galactic cor collap Supernova so of course we’re going to want to brag about it as fast as possible so this is not quite the reason I will explain it more later but indeed the cuts that I have presented there uh
Are incorporated into a real time analysis systems that is processing ARA and orca data and sending alerts within 20 seconds of the data reception to the Supernova early Warning Systems news so the Supernova selections that I have presented there is in this blue part there we count the number of
Optical modules passing a certain threshold uh like a selection fres that have presented earlier and if this number of optical module is beyond the expected background and if the P value is small enough we issue an alert of course we can also receive alerts from s
News and dig into our database to see if we have anything interesting and another part that is extremely important for Supernova analysis is the buffering part the idea here is that each time we issue a supernova alert we also save 10 minutes of data around the time of this alert with extremely loose
Cuts so you can see here what uh the time distribution of this data would look like in the final configuration of ARA for a supernova at 5 Kil SEC this is extremely noisy but could be able to track the evolution of the Supernova neutrino rates as a function of
Time now why tracking this evolution is important here I think I made my point I explained why uh we want to use neutrinos to dissect corap Supernova but still why do we want to make this dissection in real time and why do we want to share it with the public
Uh this is not only a matter of ego in fact this is also because inside the star neutrinos are going to travel considerably faster than the shock wave causing the explosion for this reason on Earth we expect to see the Supernova neutrinos several minutes to several hours before
We actually see the explosion in Optical x-ray infrared telescopes and this delay depends only on the size of the star so what this delay would allow us to do would be to use neutrinos to actually locate the Supernova and send the relevant information to telescope so that they
Can get as prompt information as prompt as possible about the explosion so one first step in locating a supernova is finding its angular position in the sky this can be done by estimating the Supernova detection time at several experiments so for example you can take the rate curve like the
Noisy rate curve that I present um two slides ago you can smooth it out and fit it to have an estimation of the Supernova detection time at cam freet with a final ARA configuration you will get an 8 millisecond uncertainty for a supernova the galactic center with
Detectors that have less noise that are more sensitive to cor collapse Supernova you could go down to a millisecond uh resolution such a high res such a uh precise resolution would allow us to compare the detection times of supernova neutrinos at different experiments located at different locations on Earth
And locate the Supernova in the sky using triangulation there are other ways to locate a supernova in some experiments you can measure the neutrino Direction but triangulation has the advantage of involving any type of experiment it’s the location that matters and of being extremely fast
So in 2020 there has been a study by KERO Al about how precise triangulation could be so here they compar they try to shift the rising parts of the superar rate curves between different experiments and see when they matched and uh they performed a global study involving Next Generation
Experiments such that the full configuration of cinet ARA Ice Cube generation 2 Juno and Hyper Candi and here you on the right you see some examples of the 90% uh confidence region in the sky for supera localization they show that this 90% confident region could be shrunk down to 140 degrees
Squared and here this result will be obtained within minutes after detection now what does 140 degrees Square means for telescopes here I have shown um the field of view of different telescopes associated with their Optical depths in magnitude and you can see that the field of view of the telescope
Varries a lot and that the narrower the field of view the deeper the telescope is so in fact um the te the optimization of the telescope observation strategy will depend not only on the angular position of the Supernova in the sky it will also depend on where it’s located actually in
The Galaxy how how far it is away from us for example here you can see a map of the optical magnitude of the Supernova as a function of the position in the Galaxy so Center of this map is the galactic center on the left you can see
The Earth in this dark blow region and you see that near the the galactic center so expected visibility of the Supernova varies uh like by orders of magnitude over distances of a few kilo so all in all angular position is not enough we need the Supernova distance who says measuring the distance
Of a COS an astrophysical object means finding a standard candle that is an observable whose initial value we know and uh whose final value it’s on Earth we can infer in the case of a supernova um an adequate standard candle can be uh the early nutrino rates that is a
Nutrino rates emitted during the few first tens of milliseconds after the beginning of the explosion after the beginning of the shock wave propagation in the case of electron neutrinos at the emission point you can see here that the size of the emission Peak initial emission Peak corresponding to shock wave propagation is extremely
Stable with respect to changes in the progenitor depend progenitor properties and changes in the features of the simulation the way it is modeled in general we don’t have access to this peak for example in water chank of detectors we have anti neutrinos but we could still measure the number of
Neutrinos emitted during the first 50 milliseconds of the explosion and get a relatively stable estimate of the Supernova distance and recently study by zet Al in 2021 has shown that if you supplement this rate in the first 50 millisecond by an observable called f Delta that includes the nutrino rates in
100 to 150 millisecond window you could get an estimate of the Supernova distance down to a few percent of precision for a supernova in the Galaxy this is extremely encouraging however these distance measurement algorithms assume that you know the nutral properties and that you are sure that they are described by the
Standard model and this is not necessarily the case so to understand why let’s track a nutrino from its emission Point down to its detection in for example hyper cand here W CH of detector so first even in the standard model neutrinos are going to undergo matter effect which will change their
Flavor composition so in particular one matter effect inside the star is the mikf smov wenin effect and even in the standard model it means that your um emitted neutrino rates will depend uh on the mass ordering in the standard model there is currently a twofold ambiguity
It can be either normal or inverted a twofold ambiguity is not a huge problem but in addition to this you could have new physics effects so here let me focus on uh propagation effects with heavy neutrino Masen States undergoing two body decays into the lightest massagen State plus an invisible particle in this
Case looking at the orange curve you go from solid to Dash to the dash line so you have an overall increase in the detecting nuto rates which could be mistaken from seeing the Supernova closer than it actually is conversely in the inverted Mass ordering you could see the Supernova further away than it
Actually is so this could interfere sizeably interfere with with your Supernova distance estimates of course it’s totally legitimate to assume that there is no new physics when you are making measurements otherwise you would you wouldn’t know um you would have to deal with zillion of models but in within the
During the next 20 years it might be possible to actually stabilize have have distance estimates which are robust against new physics effects the reason why is that currently the Supernova detection uh like the bulk of the Supernova searches are um are taken by water chair large scale water CH of
Detectors which are sensitive mainly to electron anti neutrinos during the next 20 years these detectors are going to be better they’re going to be bigger but they’re also going to be accompanied by liquid argon detectors which will be sensitive to other nutrino flavors notably we have June sensitive to
Electron neutrinos and dark matter detectors sensitive to the sum of all nutrino flavors such as dark side 20K and a bigger version of it called Argo and to understand the impact of these detectors let’s look again at the impact of neutrino decays on the Supernova rates but now in three
Experiments junee hyper camand and Argo so here the blue lines are the standard model and the oranged green dashed and solid lines are four different neutrino DEC case scenarios here to describe neutrino de case I use two parameters arar the Supernova distance over the characteristic Decay length which is
Zero in the standard model and which becomes larger as more and more neutrinos Decay on their way to Earth and second parameter is Zeta which is the branching GR ratio to active neutrinos so if D is zero then you get only sterile neutrinos as ZK products and you can get very low obser
Fluxes so here you see that in hyper camand water CH of detector you have a near degeneration between a change in DK parameters and a change in the Supernova distance this is a bad news but you can see that in June and Argo even though you have less statistics than in hyper K
You can have heighten sensitivity to neutrino decays due to the heavy change due to the significant change in the shape of the rate curve like the change in the height of the peak notably so if you combine the statistics of large scale water chunk of detectors with the shape sensitivity of the other
Experiments you could be able to measure simultaneously the Supernova distance and possible presence of new physics to formalize this we performed the study um as collaboration between different APC group and uh the Supernova modeling group at um a im at and for this study we used the same
Supernova models as the one used for the standard Supernova distance measurement algorithms so 149 Model 9 to 120 solar masses we model detection rates using snoy and snow globes tools and for the time windows we first considered the same time Windows as in the standard super res distance measurements so 0er
To 50 millisecond 100 to 150 but we split the 0 to 50 millisecond window further to resolve the neutronization peak I mean the initial Peak region and then while we have different time window we have the nutrino event counts in these different windows so we can perform a likelihood Feit of the Supernova
Distance so when we perform such a likelihood fit and in order to be uh to understand the effect of neutrino decays on distance measurements we simulated uh an observation of a supernova with an 11 solar mass progenitor a different distance with R equals 5 and Zeta equals 1 it
Means in practice that most of your heavy neutr States will have already decayed to active neutrinos on the way to us so here I’m presenting the measured distance as a function of a true distance uh for the best fit and for the 90% confidence intervals the gray bond
Is what you obtain when you assume that neutal properties are fully described by the standard model and you don’t have additional deg of Freedom so here you can see that indeed we have a bias compared to the dash line which is the identity we can mitigate this bias by introducing additional degrees of
Freedom in the likelihood to allow for new physics but of course when you do this you mitigate this bias but you get terrible error bonds this is not nice however when you combine complimentary experiments like June and Hyper camand here suddenly you gain not you not only get an accurate estimate with this
Additional degrees of freedom but you also gain back most of your Precision to have a more global view of it uh we perform this study for more experiments for a supernova are at 10 kilo per sec and we just looked at let me see if I can get the mouse back no uh
And we just looked at the size of the confidence interval so uh here is the size of the confidence interval so the position is normalized to one for 10 kilo SEC uh the dashed rectangle is what you get when the observation is like from a standard model NEOS obeying standard model
Properties and your likelihood assume that neutrinos are obeying standard model properties so this is the case where you get both maximal precision and maximal accuracy then the light rectangle is when you have the standard model for the observation but you have introduced additional degrees of freedom in the
Likelihood and then the dark rectangle is what you uh is the case that I presented before when you have Decay and you have introduced your additional degrees of freedom in the likelihood so on the right on the left you have IND individual experiments like June Ice Cube hyper camand and Argo
Where you see your error bars blowing up as was seen before and on the right you have pairs of detectors and you see that combining Dune with Hyper camand or Dune with Ice Cube allows you to gain back almost all your Precision like with a 25 up to 25% difference in practice
This result for the was for the inverted Mass ordering going to the normal Mass ordering we get a similar result except that here we can get a huge Precision gain even involving the Argo experiments so all three experiments can help you get back both your precision and your accuracy for the
Supernova distance the important Point here is the complimentarity sorry I don’t have a measurement of time that I’m just just to conclude I have taken you through a rather packed journey through Supernova nuto detection and characterization in current and upcoming experiments the bottom line is that the next Galactic
Corcap Supernova will be a once in A- lifetime opportunity to probe the physics of the star and the physics of the neutros themselves but it requires a multipronged approach and it requires including as many experiments as possible so here we have to be creative we have seen how to successfully
Integrate high energy neutrino experiments like m freet to the Supernova analysis program and we have seen how Next Generation experiments could kind of open New Horizons allowing us to simultaneously estimate the Supernova distance and constrain neutrino properties and just to finish uh this kind of study uh also featured um an
Important technological um aspect of Cam fret which is this Optical modules with multiple photo multipliers this study at C fret would not have been possible without these modules so and it represents also a testing round for future detectors as modules are going to be installed maybe in hyper and in Ice
Cube generation to thank you for your attention thank you very much son for this talk um so any questions thank you for the very nice talk just back to the last two um slides I didn’t understand why the particular combination of experiments that gives the tightest constraints like what what
Is so specific about these experiments and there is no km3 net in the in the combination so so why only all two experiments have been considered yes uh okay so first maybe why there is no CET in the combination it’s the quickest one uh it’s because um what I presented that
Kret can detect a supernova in the galaxy pretty much at any place in the Galaxy but it’s hard for Kret to do anything beyond this I can show you like some what you get it’s sorry this is what you get for example this is a neutrino rate curve that you
Expect for a 20 solar mass progenitor at 5 kilo per se so for very close by Supernova and you already see that you have something very noisy so doing shape analysis for a galactic supernova especially when at 10 kilo par SEC is very hard with ceret so here we consider
The experiments that have the smallest noise level and the highest statistics and then uh the next question I think is why uh these particular combinations of experiments well I think this bolts down to the complementarity between these experiments uh they are sensitive to three different combinations of nutrino flavors electron anti neutrinos for
Water CH of detectors electron neutrinos for June and the sum of all nuto flavors for these dark matter detectors so when you and uh when you bring all of them together you can um you can pinpoint uh new physics effects which affect different nutrino flavors differently
Which say very frequently do and this is how you can break the degeneracy between a change in between the presence of neutral decays and uh the Supernova distance estimate yes let me go back to this uh yes so the reason is the size of the time window because here we are trying
To uh count the number of photo multiplier Heats that come from the same event with don’t want to have pile up so to avoid having several events contributing to this uh to this photo to this Optical module activation we are counting the number of activated photo multipliers within an extremely narrow
Window 20 here 20 nond for the final analysis 10 nond and uh living organisms they emit a lot of light but they emit it at a very low rate compared to the to this 20 nond window it’s extremely unlikely that a living organism is going
To emit two photons within like 10 or 20 NS from each other that’s that’s the idea why does it help to have many detectors in this regime because you made the point that the value of this system is that it had many photo multipliers within the box or within the
Sphere uh well it’s because um if you have only one photo multiplier if you have only one photo multiplier activated it’s much more difficult to distinguish this single heat bioluminesence from um like a nuto interaction product that activates that that is going to give a to emit photons at much higher
Rates um so you could count the number of photo electrons deposited but this is less precise than having actually different photo multipliers recording different hits the redund uh I not sure what you mean by redundancy yeah it’s a repeat it’s the fact that several photons are emitted in a short time
Interval uh okay so I have a question online uh Joe please go ahead yeah um thank you for a beautiful talk uh my question concerns the possibility of neutrino energy resolution uh from The Source because this could be a way of um you know distinguishing different equations of
State for neutron stars Etc so is that something that’s feasible in the future okay so um well there’s first uh since the study I presented involve involved also other experiments and C freet there are quite a few experiments for which you can um which have a low energy threshold and in these
Experiments you can actually measur the energy with with a reasonable Precision like 10% or less and in experiments reconstructing the energy spectrum is quite an important part of supernova nutrino analysis the reason why I have insisted Less on this is because in experiments like Ice Cube Argo or cfret
This is much more difficult so here I have an example of how this is working out at CET at Kret the current energy reconstruction algorithm assumes that Supernova neutrinos have a spectral distribution which is described by a pinched fa distribution I can tell you more about it later it’s like it starts
As fidak and propagation effects are pinching it uh so you parameterize it using um uh the normalization a mean energy and a pinching parameter Alpha so three parameters and in the current analysis they use just the multiplicity H to look for possible hardening of the spectrum and they made
The Feit of the multiplicity distribution and with this feat uh they found this um they found this these um areas performing a k square feet I think 90% confidence region on the signal scale versus the minan energy and you can see that it varies a lot depending on what
You assume for the pinching parameter Alpha so you can get an estimate of the overall uh of the global spectral parameters of the neutros with cemet but it’s not going to be a precise estimate and it’s definitely not going to be event by event okay thank you okay there
Questions I may have a knife question from nonexpert what about the neutrinos that come from the other side of the Earth how would they in this range be affected by crossing the Earth before getting into the detectors okay yes um thank you that’s very interesting um so
I don’t have plots to illustrate this I can give you references later people have studied this uh they won’t have okay for Kret this kind of matter effects in the Earth are not going to be seen um I mean for the dek studies they’re not going to affect them because
These are really um small effects compared to what I have shown here uh in some experiments like I think hyper camand maybe Juno when it’s built you when these experiments are built I think you could pinpoint them but I don’t remember what are the distances is uh
But yeah this could uh this could be seen in principle but they are negligible compared to effects for propagation inside the Sun if this was your question uh okay other questions uh I have maybe U small there here I have a small question in your um analysis now for the distance um
Determination you if I correctly you assume there is you only rely on the nutrino signal uh but what I mean what if you also detect electromagnetic signal or maybe gravitation wave signal how does this enter into the analysis okay yes so indeed so gravitational waves are not
Going to be delayed they are going to arrive uh around the same time as the neutrinos but uh okay this I’m not an expert there so I’m not sure if there are distance estimates possible with gravitational waves maybe we could it could help for triangulation but yeah
For the distance I don’t know it really depends on whether you have a kind of a standard candle for the amplitude of the gravitational waves and this I have no idea yeah I think the signal is pretty uncertain for the gravitational waves but uh it still will be some additional
Information so I was wondering if it can can somehow help yes I agree the question is whether there is a lower limit on the expected gravitational wave flux and uh is there one or uh well I guess there is a low some lower limit I’m not sure how informative it in fact
The models are quite vary quite a lot but yeah if if they converge that will be yeah because if there is a lower limit on the expected flux you could at least say that the Supernova has to be at a distance of at least this and this could supplement the nuton observations
But true that the uncertainties are if the uncertainties are quite high it’s a bit difficult so but maybe we could look into this okay um so if there not other questions let’s uh sorry of the experiment which fraction of the time is it taking data uh in the
Case of Kret I don’t have a number here because the detector is still not quite stable so it’s um but what are the expectation for the future I mean is it a large fraction of the time oh it’s expected to be a large fraction of the
Time I I don’t have a I just don’t have the the exact number in my mind but yes 80% at least what is the probability to have another experiment being active at the same time in order to well I mean giving the at the same time I yeah well given
The number of experiments that are observing cor collapse Supernova that are aimed at observing cor collap Supernova I think right now we can comfortably say there will be at least one experiment working when uh when there is a supernova taking place for example this news program right now it has 10 experiments being
Registered and many of these experiments are actually quite stable in super cand for example the duty cycle is extremely high because it’s really Supernova is among the main part of his program the detector is well known so it’s something like shifters have have to react extremely fast if there is a data acquisition
Problem I mean I remember one time I was on shift when I was a postdoc and the data acquisition failed and by the time I managed to reach Japan it was it stayed 20 minutes like this due to a connection issue and that was already a
Huge anguish so that gives you an idea of what the stakes are okay uh last uh last question no uh okay well let’s thank Sonia again sure should they connect