It’s amazing that so much hard work (and such high levels of stress) can be condensed down so much… 5 pages, 3 plots and a table – and the new world leading limit on the WIMP-nucleon spin-independent elastic scattering cross section, of course.
Yes, the LUX Run 3 reanalysis results are finally out. It’s been in the works for over a year, and it has been a genuinely wonderful experience to watch this paper grow – and seeing my own plot in there has felt like sending forth a child into the world!
As much as I worked to improve our signal efficiency at low energies, the real star of the LUX reanalysis show was the “D-D” calibration – D-D standing for deuterium-deuterium. We calibrated the detector’s response to nuclear recoils (which we expect WIMP dark matter to cause) with something that sounds like it is out of science fiction, a D-D generator. This generator uses the fusion of deuterium (think heavy hydrogen – one proton, one neutron) to generate neutrons that are focussed into a beam and sent into the detector.
Quick LUX 101 – LUX is a dark matter search experiment. Dark matter is that mysterious dark, massive substance that makes up 27% of our universe. How does LUX look for dark matter? Well, it is a ‘dual phase xenon TPC’ detector, and it lives 4850 feet underground at the Sanford Underground Research Facility. It must be underground to shield it from as much cosmic radiation as possible – as it is looking for a very rare, weakly interacting dark matter particle called a WIMP. LUX is basically a big tank of liquid xenon, with a gas layer on top. It is sensitive to particles that enter this xenon – photons and electrons cause what we call an electron recoil (think of them bouncing off an atomic electron) whilst neutrons cause a nuclear recoil (bounce off a xenon nucleus). We expect that WIMPs will interact with the atomic nuclei too, just incredibly rarely – so understanding the detector response to these nuclear recoils is of utmost importance. Both these electron recoils and nuclear recoils, inside the liquid xenon cause a flash of light, a signal we call “S1”, the scintillation signal. Any light in LUX is picked up by two arrays of photomultiplier tubes, 122 in total. Recoils can cause ionisation of electrons; electrons are ‘knocked off’ their atoms by the collision. If you place an electric field over the xenon volume, you can actually push these electrons along, instead of letting them recombine with their atoms. In LUX, the electrons are pushed all the way to the top, and into the gaseous xenon later. They then cause a second flash of light via scintillation in the gas, “S2”, the ionisation signal (as its source is the ionised electrons). Two signals mean two things – discrimination between electron recoils (background) and nuclear recoils (possible dark matter signal!) due to the differing distribution of energy between S1 and S2 for each recoil, and secondly, 3D position reconstruction. XY coordinates can be determined from looking at which photomultiplier tubes light up, whilst the time between the S1 and S2 tells us the depth of the interaction. This XYZ position is very important; we use the xenon to shield itself from radiation from the detector materials itself, or from the surrounding rock. If we have the 3D position of all our events, we can only look in the very inner region of the detector, where it is very quiet, for those rare dark matter interactions.
Back to the deuterium-deuterium fusion neutron gun – it’s actually a wonderfully simple but extremely clever idea. We fire a beam of neutrons into our detector, all at the same energy (monoenergetic or monochromatic), at a set position. We then select events in our data along that beam, and look for those neutrons that scattered a second time in the detector. Because of that XYZ position reconstruction, if we have signals from two different scatters, we can actually determine the angle of scattering. As the initial energy is known, allows the energy of the recoil to then be calculated, via simple kinematics. Matching the recoil energy with the size of the two signals allows us to calibrate the nuclear recoil response of the detector extremely well. The light yield (in S1), tougher to measure than the charge yield (in S2), as we are talking about individual photons, was measured as low as 1.1keV. (keV are kiloelectronvolts, or 1000x the energy of a single electron moved across a potential difference of 1V. In other words, a tiny quantity. 1keV is only 1.6×10-16 joules!) The charge yield was measured below 1keV. In the previous LUX results, we had assumed a conservative hard cut off – ie we would measure no light for recoils below 3 keV. Now we know that isn’t the case, and can extend our sensitivity to lower energies – which corresponds to lighter WIMPs.
This improvement in low energy calibration, as well as a more streamlined and improved analysis framework, has led to a huge improvement in LUX’s low WIMP mass limit. In the plot above, which shows the WIMP mass against the probability of interaction with a nucleus, everything above the black line is now ruled out. If you’ve been following the WIMP dark matter saga, you will know that a few experiments were claiming hints of signals in that low mass region, but this new result definitely lays those signals to rest.
Getting a paper ready for publication has turned out to be far harder work than I expected. It requires a lot of teamwork, perseverance, brain power and very importantly, the ability to take on criticism and use it to improve. I must have remade the LUX efficiency plot over 100 times, and a fair few of those times it was because someone didn’t quite like the way I’d formatted it. In a collaboration, you have to be willing to learn from others and compromise. In the last few days before we finished I did not benefit at all from my UK time zone, as I stayed up later and later to finish things off. But – it was worth it! Now, if I search for my name on arXiv, I come up 4 times (3 LUX papers and the LZ Conceptual Design Report). As pathetic as it sounds, this is actually quite exciting for me, and is what I hope to be the foundations of a long career in physics.
The new LUX results are obviously nowhere near as exciting as an actual WIMP discovery, but it’s another step on the way there. LUX Run 4 is in full swing, where we will obtain over 3 times more data, increasing our sensitivity even further , and who knows – those WIMPs might just finally show their faces.