How old is the water in your drinking glass? What about the ice cubes floating in it? Any answer is bound to make reference to the water cycle (evaporate, rain, repeat). Still, for most practical purposes, water is both eternal and constantly replenished.

But when water flows underground or freezes into glacial ice, a clock starts ticking from the moment it loses contact with the air. In order to read the time, scientists must trap and count almost infinitesimally small quantities of the radioactive isotope krypton 81. Physicist Zheng-Tian Lu and his team at the U.S. Department of Energy’s Argonne National Laboratory outside of Chicago have honed the technique, called Atom Trap Trace Analysis (ATTA), over more than a decade and successfully used it to map the flow of water in the Nubian Aquifer, two miles beneath the Sahara Desert.

The Nubian Aquifer and ages of water. Courtesy Argonne National Laboratory

If scientists can determine how long water has been in an underground aquifer, they’ll know how fast it travels and how fast it can be replenished in certain areas – critical information in desert climates where the population depends on groundwater. Krypton 81 can help tell water’s story. Dating glacial ice, a new frontier for the technology, can tell scientists about the atmosphere hundreds of thousands of years ago.

As the technology improves, radio-krypton dating is set to become part of the scientist’s toolkit, going far beyond the range of carbon 14 dating to pinpoint the age of samples 150,000 – 1.5 million years old. But the journey to this discovery has been slow.

“In 1999, when we first demonstrated the principle of ATTA, it would take millions of liters of water for a krypton 81 analysis,” Lu said. “Basically it was impossible.”

These days, sample sizes are 100 liters but Lu hopes to get it down to 1 liter within his career.

Lu stumbled on the krypton-dating problem in 1996 when he attended a lecture by Dr. Walter Kutschera at Vienna University. The more common carbon 14 dating technique, which is used for fossils and is standard in groundwater studies, can only reach back 60,000 years. The Krypton 81 isotope decays much more slowly, with a half-life of 229,000 years.

The isotope is produced when krypton is bombarded by cosmic rays and, while it’s scarce, it is evenly distributed throughout the atmosphere and is constantly replenished.

“This exchange process stops when the water goes underground,” Lu said. Thus, it’s possible to figure out when a sample of water disappeared into the aquifer by counting the number of isotopes it has left.

The Atom Trap Trace Analysis (ATTA-3) apparatus. Courtesy Argonne National Laboratory

The problem is that there are so few krypton 81 isotopes — the name means “hidden one” — to count. How few? There are 3.34 X 10 to the 25th water molecules in a liter of water, but only 1,000 Krypton 81 atoms. (Imagine dropping a contact lens from a plane flying over the Sahara. Now find it.)

Kutschera’s solution was to count the krypton 81 atoms in an accelerator, a technique used for carbon dating called Accelerator Mass Spectrometry. In 2000, Kutschera published the first radio-Krypton dating of water from the Great Artesian Basin of Australia. But it was difficult to replicate with other studies: large accelerators are expensive to operate and difficult to access – Kutschera used the National Superconducting Cyclotron Laboratory in East Lansing, Mich. – and they needed about 16 tons of water to sample.

Lu wanted to use the atom trap technique pioneered by Energy Secretary Steven Chu (he shared a Nobel Prize in physics for developing the technology). Scientists use six lasers in a 6-and-a-half-foot-long table-top apparatus to trap krypton 81 atoms by matching the lasers’ frequency to the isotope’s resonance. The trapped atoms become excited, glow brighter and can be captured and counted by a camera.

Lu realized that the technology was well-suited for dating water but it needed to be refined on the quantitative level. “It was this mental leap that we can actually push this thing down 10 orders of magnitude,” he said.

Zheng-Tian Lu at work. Courtesy Argonne National Laboratory

By 2003, his team had tweaked the analysis so it required only 1,000 liters of water. In conjunction with Neil Sturchio of the University of Illinois at Chicago and a team from the University of Bern, Switzerland, the team extracted water from six sites in Egypt and determined the samples were anywhere between 210,000 – 1 million years old and moving anywhere from 20 cm – 2 meters per year. The Nubian Aquifer study was a milestone for Lu, but he wants to continue pushing the technology. The lab is currently testing samples from Brazil in partnership with the Isotope Hydrology Section of the International Atomic Energy Agency.

The other research front is refining the technique’s efficiency to the point where it can analyze small samples of deep ice cores from Greenland and Antarctica. If scientists can figure out how old an ice sample is, they can understand the composition of the atmosphere at the time it froze from the bubbles within it.

Only 10 samples had been radio-Krypton dated in the past 10 years — four by Kutschera’s team for Australia and six from the Nubian Aquifer.

“Starting Nov. 1, we have started doing Krypton 81 samples at the rate of two per week,” Lu said.

Matthew Van Dusen is the editor of Txchnologist. He tweets @matthewvandusen