Do supernovae generate neutrinos or antineutrinos?

[Ken G]

It was a rethorical question ;) SN1987A is the most recent close enough supernovae.

Ah, I see. Yes, we have a dearth of data here....

Read about the actual experiements yourself instead.
- Kamiokande II
- IMB
- Baksan (BNO)

That's a bit of work! It will be nice when AI answers can actually be trusted...

 

[malawi_glenn]

That's a bit of work! It will be nice when AI answers can actually be trusted...

Hard work is the key to success and understanding :)

 

[Vanadium 50]

You cannot actually invert beta decay

As @malawi_glenn said, that's what it's called. You don't complain that you drive on a parkway and park on a driveway, do you?

 

[Vanadium 50]

I tried ChatGPT

Which is about as reliable as a fortune cookie,

 

[Ken G]

It's better than that, but it certainly can't be trusted. It can give a very nice answer to a difficult question, but it can also mess up a simple one! It's actually not a bad place to start, but it's a terrible place to finish. Much like everything else about the internet!

 

[Vanadium 50]

That's interesting, it sounds like you are saying the nuclei interact differently with neutrinos and antineutrinos,

They do. If a (anti-)neutrino interacts with a nucelus, a (positron) electron is emitted, and the buclear charge (decreases) increases by one. So event rates are determined by the nuclear properties of the target nucleus and the adjacent nuclei.

As a practical matter, only a few nuclei are considered. You need kilotons of target or so, so it needs to be cheap, and since most detectors rely at least partially on optical signals, it's better if it were transparent. That leaves water (oxygen), argon, in principle nitrogen (but anything nitrogen can do argon can do better), and a few others. Art MacDonald managed to borrow tons of deuterated water from the Canadian government - there is no way he could afford to buy it.

IceCube has a thresholds much, much higher than typical SN neutrinos. Its job isn't so much to study normal SNe, but the few that are thought to drive cosmic ray production, AGNs, etc, It is approximately twice as sensitive to neutrinos as antineutrinos.

 

[Ken G]

Thus capture of an antineutrino (resulting in simultaneous emission of three photons of fixed energies) is a much more distinctive event than capture of a neutrino (resulting in emission of just one electron of no fixed energy).

This is not a hijack, it's an effort to answer the question I raised earlier-- since it was suggested that neutrinos seen from 1987A were antineutrinos, why is it easier to detect an antineutrino than a neutrino. That suggested answer was specific to the electron neutrino, and I believe Cerenkov detectors get all three types of neutrinos because there are decay channels that involve other things than electrons and positrons (and malawi_glenn mentioned muons). But it's a reasonable point, related to what I mentioned earlier about the fact that if you create antimatter in a detector containing regular matter, you will get certain telltale signs of annihilation, which might help signal that an antineutrino entered the detector. However, Cerenkov detectors don't focus on light created by annihilations, they focus on Cerenkov light, so at the moment I'm still not even clear that the Cerenkov detectors do detect more antineutrinos than neutrinos from supernovae, since such sources should emit way more neutrinos than antineutrinos, and the signal we get might be dominated by processes that cannot distinguish.

 

[Vanadium 50]

Good question, there is a supernova about every 100 years in each galaxy,

People estimate 30 for the Milky Way. So we're due,,

One might look at M31 and M33 and try and draw conclusions. The problem is that the last SN in M31 was in 1885 and there has never been one observed in M33.

Thing is M31 has a low star formation rate, and M33, while a little better, is a small galaxy. The Milky Way is thought to have a high star formation rate, but 85% of it is obscured by dust.

 

[Ken G]

Many answers are in a paper in 1987 by Krauss, which is behind a pay wall I believe.

[Another reference to ChatGPT redacted by the Mentors]

It is true what was said in the thread that the neutrinos from 1987A were probably antineutrinos (based on theoretical flux expectations and the fact that the cross section for the antineutrino-nucleus interactions is about 70 times higher than for neutrino-electron scattering, as well as what was said that antineutrinos react better with these nuclei), and it is also true that none of these detections can directly tell neutrinos from antineutrinos, even though the processes that create the Cerenkov light are very different. The big thing that is clear from this paper is that the theoretical expectation is that the total neutrino+antineutrino production rate is quite a bit larger than the excess neutrinos you get from lepton number conservation. (The paper points out that the positive lepton excess comes from deep in the supernova where neutrinos do not easily escape, so perhaps the extra leptons do not generally come out as neutrinos at all, and even if they do, it's not a large fraction of the total.)

So a core collapse supernova is now thought to produce similar numbers of neutrinos and antineutrinos, but since the nuclei in the detectors have a much higher cross section for interacting with antineutrinos, and that is also much higher than neutrino-electron scattering cross sections, the theoretical expectation is that all the neutrinos detected were probably antineutrinos, in stark contradiction to ChatGPT (but at least it was right about the different detection channels). The theoretical expectations fit the data well, so it's good evidence that the model passed a test here, but we are just shy of being able to say we directly detected antineutrinos rather than regular neutrinos. (But we are pretty close-- it would be very strange indeed if somehow there were at least 70 times more neutrinos than antineutrinos, that would totally put neutrino physics on its ear.)

I probably should have just done this work in the first place, but there you have it.

 

[berkeman]

Thread closed for Moderation...

 

[berkeman]

After some cleanup, the thread is reopened.

 

[malawi_glenn]

This is a good read on the Ice Cube (theoretical) performance on supernovae neutrinos if anyone is interested https://arxiv.org/abs/1106.6225

 

[Ken G]

On the issue of why nuclei have larger cross sections for antineutrinos, it seems to me this should not be a nucleus-dependent issue, it should have simply to do with the fact that an antineutrino can interact with a proton in a way not available to neutrinos. It's very much the antimatter aspect, though not so much an issue of annihilation-- a proton has a positive charge, so it can make a positron (or antimuon, or antitauon) when something with negative lepton number bashes into it, hence the antineutrino but not the neutrino. This cross section is way larger than neutrino-electron scattering cross sections (available to both neutrinos and antineutrinos because it is just a scattering by the weak force), so the mere fact that our detectors contain lots of protons and not antiprotons is the reason we detect antineutrinos so much more easily. Core collapse supernovae are copious sources of both neutrinos and antineutrinos, so until we detect a very large number of neutrinos in the Cerenkov detectors (which will require a supernova in our galaxy), we will not detect neutrinos from supernovae.

Alternatively, perhaps we have already seen just as many neutrinos as antineutrinos, and we simply don't have the statistics or the theoretical accuracy to know if we are getting the fluxes wrong by a factor of 2 (because as the Fermilab site above said, we don't really know we have the neutrino physics correct). The Fermilab site intimated that maybe neutrinos will end up being their own antiparticles ("Majorana fermions"), in which case, regular neutrinos should be just as able to make positrons (and antimuons and antitauons) as antineutrinos can. I'm puzzled as to why our laboratory experiments don't already know if that is happening or not, but I guess it's just hard to know what kind of neutrino you have, if you don't know how they differ.

 

[snorkack]

If neutrinos can oscillate to antineutrinos, should we have seen solar neutrinos oscillated into antineutrinos?

 

[Ken G]

Yes you'd think this would have been tested by now, by using a source you think should have few antineutrinos and lots of regular neutrinos, and a detector that mostly detects antineutrinos, and see if you get way more of them than you were expecting. For some reason, this has not yet been done in the Fermilab neutrino experiments, or they would know by now. (I don't think the issue is oscillation, it is simply if neutrinos and antineutrinos are the same particle.)

 

[Vanadium 50]

Reason 1 is that even a neutrino beam contains a few percent of antineutrinos. It's easier to say "create a pure beam" tnan it is to actually do it.

The second reason is that if you allow the decay \pi^- \rightarrow e^- \nu in addition to the normal antineutrino, it goes about 20,000x faster. So even a tiny mixing between neutrinos and antineutrinos would be immediately obvious.

In the past you've argued, yeahbut they didn't measure this exact thing. Science witha capital S says we should do itl, It is hard to get support to mount an experiment that is 100-1000x less sensitive than existing measurements because in some unspecufied way maybe this time thinghs are just different.

 

[Ken G]

Well, not quite-- even the Fermilab site explicitly dangles the possibility that neutrinos are Majorana fermions! (OK, maybe they are just looking for funding, but it's legitimate scientific curiosity, not something that has been ruled out yet.) As it happens supernovae are expected to be a pretty equal source of both neutrinos and antineutrinos, so if we think we are only detecting antineutrinos, and neutrinos are the same particle, we have an observation that is only a factor of 2 too high. The combined low statistics with theory uncertainties make it difficult to be able to notice that. (25 counts, many of which are thought to be noise, means that if we think only 20 were real counts as Strauss suggests, and we think that means the supernova should have given us 40 of which we could only see 20, it seems to us like all is well. But maybe a repeat of the experiment would have seen 30 by chance, and that's all the supernova gave us because the theory was high by 30%, then we would have been fooled by Majorana fermions). That's why we have to keep track of what we have actually established by observation, to high confidence.

If we do get a supernova in our own galaxy, we'll have plenty of statistics, and maybe we'll have better theory and will be able to use a factor of 2 difference between what is observed and what we expect to discover that neutrinos actually are Majorana fermions, and we would immediately stop saying we knew all along that what we detected before were bona fide antineutrinos. So often this has happened in the history of science, we should know better by now!

 

[snorkack]

On the issue of why nuclei have larger cross sections for antineutrinos, it seems to me this should not be a nucleus-dependent issue, it should have simply to do with the fact that an antineutrino can interact with a proton in a way not available to neutrinos. It's very much the antimatter aspect, though not so much an issue of annihilation-- a proton has a positive charge, so it can make a positron (or antimuon, or antitauon) when something with negative lepton number bashes into it, hence the antineutrino but not the neutrino.

It is a heavily nucleus-dependent issue. Consider oxygen instead of protium...
In case of O-16, F-16 is unbound, meaning that a neutrino has to remove a neutron from O-16 and leave behind O-15 (halflife 2 min) - a threshold of 15 MeV or so. An antineutrino would produce N-16 (halflife 7 s).
In case of O-18, F-18 half-life is 110 min and N-18, while bound, has half-life 0,6 s. Since neutrino absorption in O-18 is inverse electron capture of F-18, it has threshold of about 1,66 MeV.
Etc., etc.. The thresholds and cross-sections are going to depend on the specifics of daughter nuclei.

What is a bias for antineutrino: what do you get when you absorb a neutrino? A fast electron. Which looks much like a fast electron emitted by beta decay.
Absorb an antineutrino? Sure, a positron emitted by antineutrino looks much the same as a positron emitted by positron decay. But positron emitting isotopes are somewhat less common in nature than electron emitting ones. (For example K-40 emits both but far fewer positrons than electrons). It is not so much that antineutrino absorption has higher cross-section but that it seems to have lower background noise of similar looking but different events.

This cross section is way larger than neutrino-electron scattering cross sections (available to both neutrinos and antineutrinos because it is just a scattering by the weak force), so the mere fact that our detectors contain lots of protons and not antiprotons is the reason we detect antineutrinos so much more easily.

Our detectors contain a lot of neutrons, though.

Core collapse supernovae are copious sources of both neutrinos and antineutrinos, so until we detect a very large number of neutrinos in the Cerenkov detectors

There are two basic ways of neutrino/antineutrino interacting.
One is absorption. This has flavour specific information and also receives the whole energy and momentum of the incoming particle.
And the other is elastic scattering. This is flavour unspecific, and while it is constricted in terms of energy and momentum, it does not take the full momentum information of the incoming particle (because it moves on with unknown energy).
Both of these usually produce a rapid lepton. (The obvious exception is events of elastic scattering off baryons, but those are hard to detect anyway). The absorption also produces altered/unstable nuclei.
Detecting the rapid lepton is already the next step. Cherenkov detectors have high energy threshold, but have some directional information. Scintillation has far lower threshold, so catches lower energy events, but seems to lose the direction information.

 

[Ken G]

OK, thanks, now I understand what @Vanadium 50 meant. I did not recognize that an antineutrino knocking a positron off a proton is not really all that different from a neutrino knocking an electron out of a neutron. To the weak force, a proton and a neutron must not look all that different, so one has to consider the more detailed issues you are talking about, and that @Vanadium 50 alluded to earlier. Since we always have lots of water, it will probably always be easier to detect antineutrinos, so it looks like the best scenario for using astrophysical sources to test if neutrinos are Majorana fermions is to wait for a supernova in our galaxy and count on the factor 2 difference in detectable neutrinos to decide the issue, assuming our models can be relied on at the factor 2 level.

 

[PeterDonis]

Moderator's note: Discussion of Majorana neutrinos has been moved to a separate thread:

https://www.physicsforums.com/threads/majorana-neutrinos.1050693/

 

[Ken G]

In summary resolution of the issues raised, it seems that the main difference between what is called a "neutrino" vs. an "antineutrino" is its opposite chirality. So that is likely what the question was about, not the semantics of what label is most appropriate. Since we expect similar numbers of neutrinos of opposite chirality to be released from a SN, the type that gets detected depends on the detecting material. Since this is generally water, and water detects better the chirality we call antineutrinos, this is what we see, which is what @Vanadium 50 said earlier. So I think it's all clear now, but perhaps the additional issues that came up served some purpose.