17 March 2011

Reactor Data Suggest Sterile Neutrinos

Many people expected that physicists would discover a new particle in 2011. But, most people thought it would happen at the newly opened Large Hadron Collider, not in obscure neutrino experiments.

Researchers at the Neutel Conference have announced that it looks like they have discovered new non-Standard Model particles, most likely two of them given their data so far, called sterile neutrinos. The result isn't rock solid, but it does appear for the first time, that the experimental data make the existence of sterile neutrinos the most likely possibility rather than a mere theoretical pipe dream with some marginal data to support it.

The last times that similarly Standard Model changing developments have been announced with such relatively weak theoretical consensus anticipating the discovery were when the first of the third generations fermions was discovered (the tau in 1974) and when the charm quark was discovered in 1970, respectively. The other third generation fermion discoveries and the other boson discoveries were widely expected by a consensus of physicists by that time that their experimental detection was announced.

Similarly, a discovery of a light Higgs boson at the LHC, even it made, would be a discovery that was widely anticipated. This discovery, while it has been discussed for many years, was not nearly so widely expected in the community of physicists, who have been focused on the possible discovery of a Higgs boson and possibly a lightest supersymmetric particle.

What is a sterile neutrino?

Unlike three flavors (aka generations) of up quarks, down quarks and electrons in Standard Model, each of which come in four varieties (left handed, right handed, left handed state anti-particles, and right handed state anti-particles), the three flavors of neutrinos come in only two varieties each (left handed state neutrinos and right handed state anti-neutrinos). Thus, one might expect sterile neutrinos to fill that slots in the table of Standard Model fermions for right handed and left handed anti-neutrinos.

Fermions basically fit an ordinary lay understanding a "matter" while the other kind of Standard Model particles called "bosons" such as photons, gluons and proposed gravitons, are massless, with the exception of the ephmeral and massive weak nuclear force bosons and composite "glueballs," made of multiple gluons; bosons are the particles that carry non-contact forces between fermionos in the Standard Model. Within the taxonomy of fermions, there are quarks (which are the constituent components of protons and neutrons) and leptons (electrons and neutrinos).

The "sterile neutrino" experimenters are searching for is defined as a "light" (i.e. lepton mass sized), electrically neutral, left handed state anti-neutrino (which is the anti-particle of a right handed neutrino).

Why call them "sterile" neutrinos?

1. All neutrinos lack electrical charge (hence the root of the name meaning "neutral"), so they don't interact electo-magnetically.

2. The weak nuclear force interaction acts only on left-handed particles (and right-handed state antiparticles). So right handed neutrinos and their left handed state anti-particles wouldn't interact with it.

3. The strong nuclear force interacts only with quarks and gluons, not leptons or the carrier bosons of the electro-weak interactions (the photon, the W+ boson, the W- boson and the Z boson).

Thus, sterile neutrinos interact only with gravity and even then, they are lighter than all other particles that have mass.

In the sterile neutrino model used by the neutrino researchers I discuss today, normal neutrinos can oscillate into sterile neutrinos (and thus become invisible to devices designed to observe normal neutrinos through weak force interactions), presumably via the weak nuclear force.

What new experimental data about neutrinos was announced today?

Reanalysis of data from experiments studying neutrinos generated in beta decay at nuclear fission reactors show that previous studies showing experiment to match theory was wrong. Instead, 3% fewer neutrinos were detected than theory predicted. So, alternative theories that could explain this data have been considered. What is the bottom line? (Punctuation and grammar errors by the non-native English speaker author of the quoted post corrected without notation in the block quote).

For the first time, a neutrino model that includes sterile neutrinos provides a better fit to the global world data on neutrino oscillations than the conventional three active neutrino model. . . . Furthermore, the model with two sterile neutrinos is definitely better than the model with just one. More steriles or other exotic ingredients, like non-standard neutrino interactions or CPT violation, don't seem to be necessary. . . . Piece by piece, the evidence in favor of sterile neutrinos are not that great (they are considered anomalies rather than evidences), but the overall picture now favors sterile neutrinos.


How heavy are the predicted sterile neutrinos?

At a 95% confidence level, the combined experimental evidence best fits sterile neutrinos with the sum of their masses being less than 0.70 eV. Thus, they would be lighter than any non-neutrino fermion, but much heavier than ordinary neutrinos.

Early in the conference, the best available data put the rest mass of tau neutrinos is on the order of 0.04 eV and the rest mass of ordinary neutrinos is on the order of 0.0009 eV (the experimental data only directly measured the mass differences between neutrino types, rather than the absolute values, so these are only order of magnitude approximations).

Upper limits on these rest masses in absolute terms could be significantly higher if they have similar masses, perhaps as great as 0.28 eV for ordinary neutrinos based on galactic level measurements reported in the summer of 2010.

How many kinds of sterile neutrinos have been observed?

At a 95% confidence level, the combined experimental evidence best fits 1.62 (0.7 to 2.54) kinds of sterile neutrinos, in addition to the three ordinary ones and their antiparticles.

What could that raw data concerning the best fit number of sterile neutrinos mean?

Of course, we actually expect the number of kinds of sterile neutrinos to be a whole number like 0, 1, 2, 3 or some larger number, rather than a decimal, so reporting the data in decimal form is a bit odd, despite being standard practice for particle searches in physics. I presume that particle physicists simply aren't very comfortable with statistical distributions designed to force probabilities into discrete whole number bins like the Poisson distribution does, relative to the tried and true normal distributions that they are more familiar with using and understand better, even when the theoretical motivation for using a discrete set of data bins is very strong. After all, physicists are already overwhelmed with mathematics pre-requisites and have little time for "extras."

The data fit a model with two sterile neutrinos better than any other number, although it is only slightly favored relative to a one sterile neutrino model. A two sterile neutrino result is about 0.83 sigma from the mean, while a one sterile neutrino result is about 1.34 sigma from the mean.

This is surprising.

The most natural expectation would be that there would be three kinds of sterile neutrinos, exactly filling the gap in the Standard Model fermion table, just as there are exactly three generations of every other known kind of fermion. The next most common theoretical expectation had been that there would be just one kind of sterile neutrino analogous to, but not quite the same as a fourth generation neutrino (a true fourth generation neutrino would not be sterile).

The hypothesis that there are sterile neutrinos, relative to the null hypothesis that there are no sterile neutrinos, is supported by the data at roughly a 3.5 sigma level - not quite the 5 sigma that physicists use as a rule of thumb to say that they are really sure that they have discovered new physics, but well about the 2 sigma level that physicists routinely ignore as fluke results unless they have very strong theoretical motivations to suspect otherwise. A hypothesis that there are three sterile neutrinos is disfavored by this data at roughly a 3.0 sigma level - more likely than no neutrinos, but only barely.

However, this probability analysis is really skewed a little against a finding of three sterile neutrinos and towards finding zero sterile neutrinos, for at least two reasons.

First, because conservation of mass-energy suppressed oscillation into heavier neutrinos if the reaction that is generating them isn't energetic enough, but no similar constraint suppresses data in favor of a zero sterile neutrino hypothesis.

Second, since the probability of third generation fermion production is always lower than the probability of first and second generation fermion production, production of higher generation fermions is flukier than production of lower generation fermions. In other words, the law of averages keeps the expected result proportionately closer to the average result for more common events than it does for less common events. So, it is harder to get statistically significant numbers of events to confirm the existence of a third generation of any kind of fermion than it is to confirm the existence of the first two generations of those fermions. Put another way, this report's use of the same upside and downside variation from the mean to establish a confidence interval, as the presentation at the Neutel conference did, is actually a sloppy way to state the statistical confidence interval that isn't entirely accurate. Done more rigorously, the low end confidence interval is almost certainly smaller than the high end confidence interval.

We know from historical experience that it is much easier to detect that first two generations of any given kind of fermion experimentally than it is to find experimental evidence for a third generation of the same fermion. Indeed, we have just recently established with confidence experimentally that there really is indeed a third generation of neutrino at all, in 2000, despite the fact that it has been assumed theoretically to exist ever since the top quark, bottom quark (1977) and tau (1974) were discovered. Even in the case of regular neutrinos, where we are pretty confident that there are three different kinds of them, some experimental papers at this year's Neutel conference fit their data to a two neutrino oscillation model because the math is easier and it doesn't change the end result much.

To make a long story short, while the best fit to this round of data in the first paper that really strongly shows the existence of sterile neutrinos at all shows that two sterile neutrinos are most likely, on theoretical grounds and given the history of particle physics, there is good reason to believe that further analysis and experiments will show either that (1) there are no sterile neutrinos and that there was some flaw in the reactor data analysis that now undermines this hypothesis, or (2) there are three sterile neutrinos. A single sterile neutrino is probably the third most likely possibility, on theoretical grounds and because the data only weakly prefer a two sterile neutrino model over a one sterile neutrino model. Thus, it is not unlikely that the initial finding that there are probably two sterile neutrinos will not survive the test of time.

Further, Less Neutral, Analysis

Forced to bet on the matter, I would bet that there are three sterile neutrinos, in electron, muon and tau generations, with a similar mass hierarchy but heavier than ordinary neutrinos, that differ from ordinary neutrinos and their anti-particles only in mass.

This would be a departure from the rest of the Standard Model, in which left handed, right handed, left handed state antiparticles, and right handed state antiparticles all have precisely the same masses. But, since neutrinos lack charge, which is present in quarks and electrons, if sterile neutrinos and ordinary neutrinos had the same mass, it would otherwise be impossible to distinguish a left hand state anti-sterile neutrino and a left hand state ordinary neutrino. Similarly, if neutrinos and anti-neutrinos had the same mass, it would otherwise be impossible to distinguish a right hand state ordinary anti-neutrino from a right hand sterile neutrino.

The estimated mass of these sterile neutrinos is a departure from the expectation of one of the most popular theories to explain mass generation in neutrinos, call the see-saw mechanism for generating Majorana mass:

The minimal scenario is the type I See-Saw in which three RH neutrinos are responsible for the smallness of the light neutrino masses. In the standard type I See-Saw, the mass of these RH neutrinos should be very large, more than 10^9 GeV.


Thus, if these sterile neutrinos are indeed the missing Standard Model right handed neutrinos, both particles would probably have Dirac mass (like all other fermions) rather than Majorana mass and they would not be their own antiparticles, even though we have no widely accepted reason to explain why they are so much lighter than other fermions.

A general analysis of neutrino mass issues from 2002, can be found here.

While the neutrino was proposed theoretically in 1930, and discovered experimentally in 1956, neutrino mass wasn't definitely established by experiment until 1998 (prior to that point in time the Standard Model assumed that neutrinos had no mass).

Also, if electro-weak interactions have something to do with mass generation in other leptons, it wouldn't be surprising for sterile neutrinos, which don't experience those interactions, to have a different mass hierarchy. We don't have any generally accepted and experimentally supported theories for how fermions get the particular masses that they do in the Standard Model anyway. Yes, it happens via a Higgs field interaction, but that doesn't tell us, for example, why a top quark has precisely the rest mass that it does relative to an electron. In other words, it doesn't tell us why particular particles interact with a scalar Higgs field at the strength that they do.

Furthermore, supposing that there are sterile neutrinos in a place where there is a possible hole in the fermion table in the Standard Model, is on the whole, far less dramatic a conclusion than a proposal that involves a new kind of fermion that doesn't have a place on the Standard Model fermion chart at all, or CPT symmetry violation. Indeed, we have not only a Standard Model fermion table hole, we also have compelling data from astronomy to suggest that dark matter exists, and sterile neutrinos can fill part of that whole.

Another possibility is that we are seeing signatures of fourth and fifth generation ordinary neutrinos. The data come from missing neutrinos in detectors, rather than direct observation. So, it is conceivable that the detectors simply aren't tuned property to see the signal that fourth and fifth generation ordinary neutrinos are generating. This interpretation would be even more profound.

Sterile Neutrinos As A Dark Matter Candidate

Indeed, if weak nuclear force decay can generate sterile neutrinos, which don't interact except through gravity, and there is no corresponding force to turn sterile neutrinos into non-sterile matter, we have a natural candidate to produce a steadily increasing amount of dark matter in the universe, and a natural reason for that dark matter, which is produced in nuclear reactions such as those taking place in stars, to have the kind of strictly defined distribution relative to the luminous matter in a galaxy, since the luminous matter is made up of stars whose nuclear reactions are an important source of that dark matter.

It should be possible with back of napkin calculations based on the empirically observed result that sterile neutrino production appears to account for a 3% reduction from the expected value of neutrino production in nuclear reactors, the age of the universe, and pre-existing baryogenesis and leptogenesis models in cosmology, to make a pretty good estimate of the amount of dark matter in the universe today that can be expected to have arisen from nuclear reactions in that time period, and also to estimate the amount of leptogensis in the very early moments after the Big Bang that created sterile neutrinos. This, in turn, should make it possible to infer what percentage of the dark matter in the universe can be attributed to sterile neutrinos.

Once mainstream dark matter proportion estimates are adjusted to reflect sterile neutrinos and previous great underestimates of the amount of ordinary matter in elliptical galaxies (that bring the ratio of ordinary matter to dark matter to something closer to 50-50) the target amount of remaining missing mass will be easier to determine.

Previous speculative research has largely downplayed sterile neutrinos as plausible dark matter candidate, but that was before we had a sufficiently experiementally well defined particle that acts like dark matter to work with when making that analysis. Of course, the existence of sterile neutrino dark matter doesn't imply that there can't be other kinds of dark matter in existence as well. Indeed, I strongly suspect that the back of napkin calculations that I suggest above would not produce enough sterile neutrinos to enitrely fill the dark matter gap by themselves.

Footnote on flavor-changing neutral currents

As I noted above, the weak nuclear force acts only on left handed particles and their right handed anti-particles. The strong nuclear force, the electro-magnetic force and gravity, in contrast, do not change the flavor (aka generation) of a fermion. Bosons don't appear to have different flavors. Photons of a particular frequency/wave length can be more or less energetic in multiple of Plank's constant, but we don't consider these different energy states to be different flavors of a photon.

So far as I know, left handed and right handed quarks, electrons, muons and taus are equally common (although neutrinos are all lefted handed), although I don't know this with any great certainty.

Also, we know that second and third generation quarks and electrons decay into lighter particles, even if they are right handed particles. There is no such thing as a stable, right handed top quark, charm quark, bottom quark, strange quark, tau, or muon. So far as I know, left handed and right handed second and third generation quarks and electrons decay into lighter particles at precisely the same rate.

This is an apparent contradiction. If a right handed top quark doesn't interact via the weak nuclear force, how does it change flavors?

I believe, but I am not really expert enough to know, that these kinds of events are called Flavor changing neutral currents. But, clearly, I need a better understanding of the weak nuclear nuclear force in relation to parity and flavor changing.

Other News

The probability of the rate of electron neutrinos directly oscillating to tau neutrinos, which is determined by sin squared theta 13, is in the vincinity of 0.1 to 0.3.

MINOS is showing apparent differences between the mass of neutrinos and anti-neutrinos.

A new study at Fermilab has found a two standard deviation difference between the mass of the top quark and its anti-quark. CPT symmetry would expect the masses of the top quark and the anti-top quark to be identical.

1 comment:

Andrew Oh-Willeke said...

Some cosmology data also favors five rather than three types of neutrino.