Berkeley, California - A 20-month experiment to detect dark matter from the depths of an abandoned gold mine has come to an end, narrowing the search for, but failing to find, an elusive particle that should be everywhere.

The Large Underground Xenon dark matter experiment, or LUX, operated beneath a mile of rock at the Sanford Underground Research Facility in the former Homestake gold mine in the Black Hills of South Dakota. It’s mission: to detect so-called weakly interactive massive particles, or WIMPS, that some physicists think comprise the mysterious dark matter that makes up 85 percent of all the matter in the universe.

If the WIMP hypothesis is correct, billions of these particles pass through our bodies every second, as well as Earth and everything on it. But because WIMPs interact so weakly with ordinary matter, this ghostly traverse goes entirely unnoticed.

Despite the LUX team’s non-discovery, they have narrowed the mass-range in which the putative particle can exist, focusing future searches within that range.

“The extra experiment gave us more sensitivity in searching for high-mass dark matter particles,” said Daniel McKinsey, a UC Berkeley professor of physics and senior faculty scientist at Lawrence Berkeley National Laboratory, and co-spokesperson for LUX. “Many theorists think that the dark matter mass will be high, around a thousand times the mass of a proton.”

Overall, the last 20-month experiment provided about four times better sensitivity to high-mass particles than earlier underground experiments that began in 2013.

LUX’s sensitivity far exceeded the original goals for the project, according to McKinsey, which makes the team confident that if dark matter particles had interacted with the LUX’s xenon target, the detector would almost certainly have seen it. That enables scientists to confidently eliminate many potential models for dark matter particles, offering critical guidance for the next generation of dark matter experiments.

First constructed in 2011, LUX completed its final dark matter search in May, and will now be upgraded to a new and more sensitive experiment called LUX-ZEPLIN that draws in researchers who built a similar underground dark matter detector in the UK called ZEPLIN. The new detector will be funded by the Department of Energy.

The new research results were described today at an international dark matter conference in Sheffield, UK, and detailed on the LUX collaboration’s website.

Key to the team’s ability to improve the sensitivity of the experiment was a calibration technique developed in McKinsey’s research group at Yale University and now UC Berkeley that allowed the researchers to remove sources of noise caused by electrons building up on the inner Teflon coating of the tank holding a third-of-a-ton of cooled liquid xenon. If a WIMP collided with of a xenon atom within the tank, powerful sensors inside would detect the tiny flash of light and electrical charge created.

“Teflon is a terrific reflector of light, but it builds up electrons, which distort the electric field we use to detect the effects of dark matter,” McKinsey said. “So we relied on our calibrations to compensate. That was fairly tricky and took a lot of effort.”

The data analysis alone took more than 1,000 computer nodes at Brown University’s Center for Computation and Visualization (CCV) and the advanced computer simulations at Lawrence Berkeley National Laboratory’s National Energy Research Scientific Computing Center (NERSC).

The work was conducted by a group of researchers from around the world, including Berkeley teams totaling 21 people: two led by McKinsey and physicist Robert Jacobsen at UC Berkeley, and teams led by Murdock Gilchriese, Kevin Lesko, Simon Fiorucci, Peter Sorensen and Mike Witherell at Berkeley Lab.

“We worked hard and stayed vigilant over more than a year and a half to keep the detector running in optimal conditions and maximize useful data time,” said Fiorucci, a physicist and science coordination manager for the experiment. “The result is unambiguous data we can be proud of and a timely result in this very competitive field — even if it is not the positive detection we were all hoping for.”

While the LUX experiment successfully eliminated a large swath of mass ranges and interaction-coupling strengths where WIMPs might exist, the WIMP model itself “remains alive and viable,” said Rick Gaitskell, a professor of physics at Brown University and co-spokesperson for the LUX experiment. “LUX was racing over the last three years to get first evidence for a dark matter signal. We will now have to wait and see if the new run this year at the Large Hadron Collider at CERN will show evidence of dark matter particles, or if the discovery occurs in the next generation of larger direct detectors.”