Dark matter experiment XENON1T sees an excess at low energies

[mfb]

TL;DR Summary

XENON1T sees an excess of events at low energies that doesn't have a good explanation at the moment.

Press release
arXiv

~3.5 sigma significance and it could be caused by tritium contamination - although it's unclear where the tritium would come from. If it is really new physics then solar axions or a magnetic moment of neutrinos would fit well.

The key plot, figure 5 in the arXiv version, the excess is right above the threshold where they can find events:

 

[George Jones]

For a good commentary on this, see Jester's blog,

http://resonaances.blogspot.com/2020/06/hail-xenon-excess.html.

 

[dlgoff]

... they cannot exclude the presence of a small amount of tritium in the detector ...

I'm wondering where the tritium may come from during the processes of making the detector?

 

[Vanadium 50]

I'm wondering where the tritium may come from during the processes of making the detector?

A. Tritium is everywhere.
B. XENON100, the predecessor, used tritated methane to calibrate. XENON1T did not, but it has many of the same people, same labs, etc.

(I'm not arguing that the signal is tritium. Just answering the question.)

 

[eloheim]

For a good commentary on this, see Jester's blog,

http://resonaances.blogspot.com/2020/06/hail-xenon-excess.html.

Holy dang that's a name I haven't heard in a while. Last blog post before this one was June 20...2018! Says something a little sad about the state of high energy particle physics today though.

 

[Vanadium 50]

I don't think the output of one blogger is a good indication of this, or any, field.

 

[dlgoff]

(I'm not arguing that the signal is tritium. Just answering the question.)

Thank you sir.

 

[ohwilleke]

Another paper has identified plausible sources of unaccounted for background in addition to tritium, and astrophysical constraints also disfavor the solar axion hypothesis.

One of the authors of the paper in a statement made to the New York Times clarifies:

“We want to be very clear that all we are reporting is observation of an excess (a fairly significant one) and not a discovery of any kind,” said Evan Shockley of the University of Chicago in an email.

This very public statement undermining the cymbal crash that is getting attention for this paper, and the self-contradictory language in the abstract and press release, strongly suggests that the 163 scientists in the collaboration were sharply divided over whether such a highly speculative and ill supported claim of new physics should be included in the abstract and press release accompanying the paper. Shockley is doing his best not to have his professional reputation tarnished by this paper by making this statement.

The pre-print of the new paper by the XEON1T dark matter detection experiment (press release available here) didn't discovery anything cool and basically admits that in its abstract, which should have read as follows :

We report results from searches for new physics with low-energy electronic recoil data recorded with the XENON1T detector. With an exposure of 0.65 tonne-years and an unprecedentedly low background rate of 76±2stat events/(tonne×year×keV) between 1-30 keV. . . .

The excess can . . . be explained by β decays of tritium, which was initially not considered, at 3.2σ significance with a corresponding tritium concentration in xenon of (6.2±2.0)×10^−25 mol/mol. Such a trace amount can be neither confirmed nor excluded with current knowledge of production and reduction mechanisms. . . .

This analysis also sets the most restrictive direct constraints to date on pseudoscalar and vector bosonic dark matter for most masses between 1 and 210 keV/c2.

Why did the authors go further an explore BSM physics explanations in the face of a reasonable alternative possibility, and the fact that it was impossible to prove or disprove the tritium contamination hypothesis because "the tiny levels of tritium in question here would be impossible to perfectly screen out. And with XENON1T now taken apart to build a bigger future experiment, it’s impossible to go back and check."?

Normally, fairly speculative observations like the hypotheses proposed in the main paper about axions would have been omitted, and the physicists who were advocated of that explanation would have published their own hep-ph paper without putting the collaboration's entire staff's reputation on the line by including it in the main paper to propose an idea like this one.

I think that the most likely reason that this didn't happen was out of personal homage for the late Roberto Peccei and his surviving family members.

Roberto Peccei, the physicist at UCLA who first proposed the axion as a hypothetic particle to explain the lack of CP violation in the strong force called the axion back in 1977 died on June 1, 2020 at age 78 of non-COVID related causes. According to Lubos Motl, graduate students working on the experiment had been aware of the result for about a year before it was published. Given how the academic grapevine works, Peccei himself had probably been informed of the result by someone associated with the collaboration prior to its publication.

It isn't implausible that the collaboration allowed the axion claim to be highlighted despite only marginal support for it in this experiment, to honor and call attention to his recent passing. As a towering figure in high energy physics, Peccei certainly deserved more of a tribute than his passing initially received.

 

[mfb]

Another paper has identified plausible sources of unaccounted for background in addition to tritium

Do they? Their proposed backgrounds fit worse, and if you allow both (figure 4) then the best fit value has tritium and zero Sb/Pb. It even looks as if negative Sb/Pb contributions would fit even better, but that's an unphysical region of course.

This very public statement undermining the cymbal crash that is getting attention for this paper, and the self-contradictory language in the abstract and press release, strongly suggests that the 163 scientists in the collaboration were sharply divided over whether such a highly speculative and ill supported claim of new physics should be included in the abstract and press release accompanying the paper.

I don't see how you interpret that into the public releases. XENON1T is a dark matter experiment, after all. They see an unexplained excess, so they study possible explanations, dark matter and others, and report both in the abstract.
Of course there was a lot of internal discussion, but that's true for every excess. The Higgs discovery papers got a lot of internal discussion, too, even though the interpretation was pretty clear.

It's nearly impossible to disprove the tritium contamination hypothesis because the experiment is gone, but it's also unlikely. Where would the tritium come from? They didn't worry about tritium in advance for a good reason: No one found a way how relevant amounts of tritium would have come into the experiment.

It isn't implausible that the collaboration allowed the axion claim to be highlighted despite only marginal support for it in this experiment, to honor and call attention to his recent passing.

I think it's a very implausible idea.

 

[ohwilleke]

It's nearly impossible to disprove the tritium contamination hypothesis because the experiment is gone, but it's also unlikely. Where would the tritium come from? They didn't worry about tritium in advance for a good reason: No one found a way how relevant amounts of tritium would have come into the experiment. I think it's a very implausible idea.

Per the original paper, the amount of Tritium contamination required to produce the effect seen, with a substance that wasn't tightly screened for because it wasn't pertinent to the primary mission of the experiment to detect WIMP dark matter (which it didn't), is truly tiny. It is roughly one tritium atom (each of which has a mass of 3.016 atomic mass units) per 1,000,000,000,000,000,000,000,000 xenon atoms (each of which has a mass of 131.293 atomic mass units). So, it would take roughly a mere 26 tiny tritium atoms interspersed into each ten kilograms of xenon, to produce the effect observed.

Many ways that could have found its way in have been proposed.

 

[mfb]

Naturally tritium has a concentration of ~10^-18 in the atmosphere/near surface. With atmospheric/surface water sources they would have needed a 10^-6 concentration of water in their xenon, with underground water sources it would have had to be even more. Safe to say this didn't happen.

 

[Vanadium 50]

While I am not a fan of this paper, or even it's ilk (theorists, usually from the "subcontinent" who try and explain what experimenters did wrong), I think the odds of tritium contamination are higher than you think.

First, these experiments are extraordinarily sensitive. LZ can see a single weld made with the wrong kind of welding rods. CDMS (I think) could see a single solder joint made with ordinary solder and not special solder made with ancient lead.

Second, tritium is everywhere, and can be concentrated by unexpected processes. While I don't think the krypton removal concentrates tritium, I wouldn't want to exclude it without testing it.

Third, XENON100 used tritiated methane to calibrate. XENON1T didn't, but it was built largely be the same people in the same places. There was some tritium around.

 

[mfb]

I don't say it can't be tritium, but writing out large numbers isn't an argument for it. If it is tritium then it must have an artificial origin or it must have been enriched by some method. It's not "ah of course, there is tritium in water and we had a bit of water, problem solved".

 

[Vanadium 50]

Well, before I would conclude that they have discovered solar axions - and discovered them in a region excluded by other observations, I'd want more evidence that it is not tritium than "Tritium! You besmirch my honor by saying it's tritium! It shall be pistols at dawn!"

I've been to a few seminars on this and walked into them thinking "this can't be tritium". Now I am not so sure.

1. It doesn't take much tritium to cause this. (about 5000 atoms)
2. The evidence from xenon light yield and drift properties suggests that H2 and H2O contamination is low, but a process that either enriched this in HT or HTO is not excluded. Nor are other molecules, such as CH2T. Krypton removal is a process that separates out a chemically identical but physically identical material.
3. The signal is largest at the start of the run. That's exactly what you expect if you have a contaminant that takes a while to purge.

In short, the experiment says they can't exclude tritium, and I believe them.

My own experience is that this is hard. I recall an incident with a liquid helium plant. You don't want water in the helium because it ices up. So we designed the plant so that any device in contact with helium is not in contact with water and vice versa. Even the piping is never next to each other. But we were getting water ice. Not from the air (no nitrogen ice). Contaminants can sneak in. Fact of scientific life.

 

[mfb]

A statistical fluctuation stays likely, too.

This was 0.65 t*year. XENONnT should quickly exceed that and I guess they'll pay more attention to tritium now.