Thursday, May 31, 2018

The MiniBoone Anomaly

A pre-print from May 30, 2018 from the MiniBooNE neutrino oscillation reactor experiment's collaboration reports a big anomaly. Lubos also discusses the announcement and he draws some conclusions that the paper does not. The abstract is as follows:
The MiniBooNE experiment at Fermilab reports results from an analysis of νe appearance data from 12.84×1020 protons on target in neutrino mode, an increase of approximately a factor of two over previously reported results. A νe charged-current quasi-elastic event excess of 381.2±85.2 events (4.5σ) is observed in the energy range 200<EQEν<1250~MeV. Combining these data with the ν¯e appearance data from 11.27×1020 protons on target in antineutrino mode, a total νe plus ν¯e charged-current quasi-elastic event excess of 460.5±95.8 events (4.8σ) is observed. If interpreted in a standard two-neutrino oscillation model, νμνe, the best oscillation fit to the excess has a probability of 20.1% while the background-only fit has a χ2-probability of 5×107 relative to the best fit. The MiniBooNE data are consistent in energy and magnitude with the excess of events reported by the Liquid Scintillator Neutrino Detector (LSND), and the significance of the combined LSND and MiniBooNE excesses is 6.1σ. All of the major backgrounds are constrained by in-situ event measurements, so non-oscillation explanations would need to invoke new anomalous background processes. Although the data are fit with a standard oscillation model, other models may provide better fits to the data.
Basically, it is seeing more electron neutrinos and electron anti-neutrinos than expected; twice the rate of previously reported results.

A similar anomaly was seen at the LSDN experiment, leading to the "reactor anomaly" in neutrino physics, but this had almost been ruled out by a variety of experiments and revised estimates of the number of neutrinos that should be produced from a reactor source, when the MiniBooNE result dropped in our lap, arguably supporting the hypothesis of the reactor anomaly that there is a fourth kind of neutrino out there.

For my money, I expect that the MiniBooNE anomaly will be resolved as well, just like the LSDN one was, due to some previously unconsidered source of additional neutrinos, particularly as MiniBooNE seems to be reporting different data than it did earlier in its run. 

For example, maybe the reactor that is the main source of its neutrinos was operated differently in some manner that the MiniBooNE experiment wasn't told about causing it to produce more neutrinos, such as having a different fuel mix. Or, perhaps somebody tuned up the neutrino detectors better causing them to capture a larger percentage of events than it had in the past and this wasn't considered in determining the expected number of events when evaluating this anomaly.

But, for now, this is the hot new mystery in physics land.


Cherry picking the LSDN and MiniBooNE anomalies without also including data from other reactor experiments is statistically unsound, so the 6.1 sigma anomaly quoted isn't meaningful. If you include all available evidence, pro- and con- that is known and available, as any valid statistical probability estimate should, the case for a fourth neutrino type that oscillates with the three Standard Model neutrinos is greatly diminished.

A review of the pre-anomaly evidence that there is not a sterile neutrino that oscillates with the three active neutrinos can be found at this prior post at this blog (as does 2014 data from the Daya Bay and JUNO reactor neutrino experiments). Some highlights from that post:
[A]s of 2015, the constraint with Planck data and other data sets was 3.04 ± 0.18 (even in 2014 cosmology ruled out sterile neutrinos). Neff equal to 3.046 in a case with the three Standard Model neutrinos and neutrinos with masses of 10 eV or more not counting in the calculation. So, the four neutrino case is ruled out at a more than 5.3 sigma level already, which is a threshold for a scientific discovery that there are indeed only three neutrinos with masses of 10 eV or less, ruling out the sterile neutrino hypothesis for a stable sterile neutrino of under 10 eV (when a best fit of potential anomalies from reactors predicts a sterile neutrino mass of about 1 eV also here). A 2015 pre-print on notes that:
The 95% allowed region in parameter space is Neff < 3.7, meff s < 0.52 eV from PlanckTT + lowP + lensing + BAO. This result has important consequences for the sterile neutrino interpretation of short-baseline anomalies. It has been shown that a sterile neutrino with the large mixing angles required to explain reactor anomalies would thermalize rapidly in the early Universe, yielding ∆Neff = 1. The preferred short-baseline solution then corresponds to ms of about 1 eV and ∆Neff = 1 and is strongly excluded (more than 99% confidence) by the above combination of Planck and BAO data.
The MINOS and MINOS+ reactor experiments rule out a light sterile neutrino, confirming the cosmology result. The abstract of a new pre-print on their results states that:
"A simultaneous fit to the charged-current muon neutrino and neutral-current neutrino energy spectra in the two detectors yields no evidence for sterile neutrino mixing using a 3+1 model. The most stringent limit to date is set on the mixing parameter sin2θ24 for most values of the sterile neutrino mass-splitting Δm241>104 eV2."

The MINOS data explores a range of values for Δm41 between the lightest mass state and the sterile neutrino mass state of 10 meV to 32,000 meV, where the bounds on the sum of the three neutrino masses from cosmology in the currently experimentally preferred normal hierarchy is 60 meV to 110 meV. For example, the MINOS data shows that:
At fixed values of ∆m241 the data provide limits on the mixing angles θ24 and θ34. At ∆m241 = 0.5 eV2, we find sin2θ24 less than [0.0050 (90% C.L.), 0.0069 (95% C.L.)] and sin2θ34 less than [0.16 (90% C.L.), 0.21 (95% C.L.)].
Weak boson decays have long ago ruled out the possibility of a number of weakly interacting neutrinos different than three. The number of weakly interacting neutrinos of less than 45 GeV upon Z boson decay is 2.992 ± 0.007 (with a mean value 1.14 sigma from 3) which is consistent with the Standard Model, in a quantity that must have an integer value. The two neutrino and four neutrino hypotheses are ruled out at the 140+ sigma level, when a mere 5 sigma result is considered scientifically definitive.
The Daya Bay and Juno paper abstract states that:
In this work, we show that the high-precision data of the Daya Bay experiment constrain the 3+1 neutrino scenario imposing upper bounds on the relevant active-sterile mixing angle sin 2 2 θ14 . 0 .06 at 3 σ confidence level for the mass-squared difference ∆ m 2 41 in the range (10 − 3 , 10 − 1 ) eV 2 . The latter bound can be improved by six years of running of the JUNO experiment, sin2 2θ14 . 0.016, although in the smaller mass range ∆m2 41 ∈ (10 − 4 , 10 − 3 ) eV 2 . We have also investigated the impact of sterile neutrinos on precision measurements of the standard neutrino oscillation parameters θ13 and ∆ m 2 31 (at Daya Bay and JUNO), θ12 and ∆ m 2 21 (at JUNO), and most importantly, the neutrino mass hierarchy (at JUNO). We find that, except for the obvious situation where ∆ m 2 41 ∼ ∆ m 2 31, sterile states do not affect these measurements substantially. 
Further data from Daya Bay in 2017 further disfavors this hypothesis. The abstract of this paper notes that:
The Daya Bay experiment has observed correlations between reactor core fuel evolution and changes in the reactor antineutrino flux and energy spectrum. Four antineutrino detectors in two experimental halls were used to identify 2.2 million inverse beta decays (IBDs) over 1230 days spanning multiple fuel cycles for each of six 2.9 GWth reactor cores at the Daya Bay and Ling Ao nuclear power plants. Using detector data spanning effective 239Pu fission fractions, F239, from 0.25 to 0.35, Daya Bay measures an average IBD yield, σ¯f, of (5.90±0.13)×1043 cm2/fission and a fuel-dependent variation in the IBD yield, dσf/dF239, of (1.86±0.18)×1043 cm2/fission. 
This observation rejects the hypothesis of a constant antineutrino flux as a function of the 239Pu fission fraction at 10 standard deviations. The variation in IBD yield was found to be energy-dependent, rejecting the hypothesis of a constant antineutrino energy spectrum at 5.1 standard deviations. While measurements of the evolution in the IBD spectrum show general agreement with predictions from recent reactor models, the measured evolution in total IBD yield disagrees with recent predictions at 3.1σ. 
This discrepancy indicates that an overall deficit in measured flux with respect to predictions does not result from equal fractional deficits from the primary fission isotopes 235U, 239Pu, 238U, and 241Pu. Based on measured IBD yield variations, yields of (6.17±0.17) and (4.27±0.26)×1043 cm2/fission have been determined for the two dominant fission parent isotopes 235U and 239Pu. A 7.8% discrepancy between the observed and predicted 235U yield suggests that this isotope may be the primary contributor to the reactor antineutrino anomaly.


Graham Dungworth said...

"For my money, I expect that the MiniBooNE anomaly will be resolved as well..."
Those combined excesses of 6.1 sigma represent a ca. 1 in 10^9 chance of being false, Andrew.

"...twice the rate of previously reported results."

It's fascinating. Actually a factor 2 increase is interesting.
Perhaps the easiest test would be cosmological, to re-analyse
the cmb continuum profile.

That "4th sterile" species may masquerade as a doubling of the 3 neutrino species. That a kack handed Left hand chiral standard model also has an extant Right handed representation; I'm left handed too. This may be due to a possible excess radiation profile peaking at 0.89 Kelvin (0.000383eV)on the 2.725Kelvin continuum that may have been subtracted as ir radiation from "dust".

Yes- it is hot news; we may live in exciting times but these particle physics experiments take decades of work.

andrew said...

"Those combined excesses of 6.1 sigma represent a ca. 1 in 10^9 chance of being false, Andrew."

By that reasoning the superluminal neutrinos detected by OPERA were true too.

Estimates of statistical error are reasonably accurate (although they make an assumption of a Gaussian distribution of error which past experience shows does not accurately characterize actual statistical errors in many HEP experiments).

Estimates of systemic error only consider the sources of systemic error that are quantified by the investigators so those estimates are only accurate if you assume that nobody missed a source of systemic error and accurately characterized the situation theoretically, and by definition miss the theoretical and systemic error sources that the authors didn't think about. When an extraordinary claim is made that is contested by many other investigations that did not reach the same result, the contradiction of two claims that we have in this case can't be evaluated in a vacuum.

This is particularly true when the experiment itself is self-contradictory, reporting one result early on and a different one later in the experiment. This is a red flag for a possible failure to identify a source of systemic error.

There are a great many ways you could tweak the theory to match this result, but figuring out how to do that without contradicting other contrary reactor neutrino data which is copious is the problem.

Marni Sheppeard said...

Hi Graham. Good to see you here. This one will be tricky for the standard paradigm, and of course we have a lot of work to do on our cosmological Higgs mechanism, but the data seems to be working out nicely.

andrew said...

More on the MiniBooNE anomaly here:

Tom Andersen said...

"Basically, it is seeing more electron neutrinos and electron anti-neutrinos than expected; twice the rate of previously reported results."

They collected twice as much data, the signal rate is the same.

"FIG. 12: The neutrino event rates versus time over the 15 year MiniBooNE running period, including neutrino- mode, antineutrino-mode, and beam-dump-mode data. The neutrino-mode event rate of 1 event per 1015 POT has been stable to < 2% over 15 years."