Directly detecting dark matter would be one of the most important scientific discoveries of all time – but it's no easy feat to find something that's invisible and barely interacts with regular matter. Now the results are in from one of the most comprehensive dark matter experiments ever run, and while the stuff was again a no-show, the study helps scientists zero in on where it might be hiding.

In theory, dark matter is all around us, but you wouldn't know it because it doesn't interact with light at all and only rarely with other matter. The logical question to ask then is why do scientists think it exists at all? The answer dates back to the 1930s, when astronomers realized that the movements of stars and galaxies indicated they had far more mass than could be seen. Research over the 85 years since then has continuously backed up the observation, suggesting dark matter makes up over 80 percent of the mass of the universe.

For decades, scientists have been trying to directly detect the rare instances where dark matter interacts with the regular stuff. The problem is, ordinary matter interacts with other ordinary matter all the time, so spotting dark matter interactions among all that background noise is a complicated task. In an attempt to get around this, dark matter detectors are usually set up miles underground to avoid interference from cosmic rays, and designed to be incredibly sensitive instruments to pick up the tiniest of disturbances.

These experiments are usually focused on specific candidate particles, and although none of them have directly borne fruit yet, they've each helped narrow the search. The HADES particle detector ruled out "dark photons" in 2014, the LUX facility dismissed some weakly-interacting massive particles (WIMPs) in 2016, and last year the nEDM experiment ruled out certain kinds of axions.

This newest study comes from an experiment known as XENON1T, which is the largest and most sensitive detector of its type. Similar to the LUX detector, XENON1T is studying WIMPs as a dark matter candidate by watching for their interactions with atoms of liquid xenon. The detector contains an active target volume of 1,300 kg (2,866 lb) of xenon, and to keep interference to a minimum this is housed inside a cryo-chamber, which is submersed in a tank of water and located deep beneath a mountain.

The idea is that dark matter particles will collide with the nuclei of xenon atoms in the tank, giving off a flash of light that the instrument can detect. This is an incredibly rare event, so to increase their chances of witnessing it the researchers used a huge target volume of xenon, and watched it for 279 days.

Although it was calculated that two events should be detected in that time, the XENON1T team reports seeing no signals at all. A null-result isn't a complete washout though – instead, it tells the scientists that dark matter particles must be smaller than previously suspected, below the range of the instrument's sensitivity.

In this case, the experiment places the upper limit on the effective size of dark matter to be 4.1 x 10-47 sq cm – or a trillionth of a trillionth of a centimeter.

"We now have the tightest limit for what is known as 'the WIMP-nucleon cross section,' which is a measure of the effective size of dark matter, or how strongly it interacts with normal matter," says Ethan Brown, a researcher on the project. "With these results, we have now tested many new theoretical models of dark matter and placed the strongest constraints on these models to date."

Future iterations of the instrument may be more sensitive, helping to pin down this ever-elusive particle.