Last week, NASA’s Perseverance rover shot for a new milestone in the search for extraterrestrial life: drilling into Mars to extract a plug of rock, which will eventually get fired back to Earth for scientists to study. Data sent to NASA scientists early on August 6 indicated a victory—the robot had indeed drilled into the red planet, and a photo even showed a dust pile around the borehole.

“What followed later in the morning was a rollercoaster of emotions,” wrote Louise Jandura, chief engineer for sampling and caching at NASA’s Jet Propulsion Laboratory, in a blog post yesterday describing the attempt. While data indicated that Perseverance had transferred a sample tube into its belly for storage, that tube was in fact empty. “It took a few minutes for this reality to sink in, but the team quickly transitioned to investigation mode,” Jandura wrote. “It is what we do. It is the basis of science and engineering.”

By now, the team has a few indications of what went wrong in what Katie Stack Morgan, deputy project scientist of the Mars 2020 mission, calls “the case of the missing core.”

“We’ve successfully demonstrated the sample caching process, yet we have a tube with no core in it,” she says. “How could it be possible that we have carried out all of these steps perfectly and successfully, yet there is no rock—and no anything—in the tube?”

One theory, of course, was that the rover had simply dropped the core sample. But there were no broken pieces on the surface. Also, Stack Morgan says, the tube was “very clean, not even dusty, suggesting that there was perhaps nothing that had ever gotten into the tube.”

NASA scientists now think that the core was actually pulverized in the drilling process, then scattered around the borehole. “That would explain why we don’t see any pieces in the hole and why we don’t see any pieces on the ground because they have basically become part of the cutting,” says Stack Morgan. “So we started to think about why that happened because that is not a behavior that the engineers saw in the very extensive test set of rocks that they cored prior to launch.”

Perseverance is drilling in Jezero Crater, which used to cradle a lake, and therefore may have been home to ancient microbial life. (It’s been relying on the Mars helicopter, Ingenuity, to scout ahead for spots to dig.) By digging into the rock instead of just sampling dust at the surface, the rover will provide vital clues about the geological history of the planet. The Curiosity rover, which landed on Mars in 2012, also drilled, but it was designed to grind the rock instead of extracting cores. This time, NASA engineers want samples that let them observe the rock as it was laid down so they can analyze it for hallmarks of life—some microbes, for instance, leave behind characteristic minerals.

For Perseverance, the drilling process actually begins inside the rover in a section called the adaptive caching assembly. Here, a robotic arm takes a tube out of storage and inserts it into the “bit carousel,” a storage container for all of Perseverance’s coring bits. The carousel then rotates, presenting the tube—which is about the same shape and size as a laboratory test tube—to the 7-foot-long arm that will actually do the drilling. “We pick up that coring bit, and that has the tube inside,” said Jessica Samuels, surface mission manager for Perseverance, in an interview before the first drilling attempt. “And now at that time we’re ready to actually acquire the sample.”

To get that rock, the drill on the larger robotic arm both rotates into the ground (the way you’d use an apple corer) and hammers into it. All the while, the rover is sensing its progress as it drills. This data feeds into an algorithm that automatically adjusts the drilling, for instance adding more or less hammering. Once the robot has bored far enough, it has to break the rock sample off, so it will actually shift the drill. “It causes the tube inside the coring bit to actually shift to the side to cause that core-break motion,” said Samuels.

Ideally, the robot will come up with a chalk-sized piece of Mars. Perseverance will actually repeat this process many more times, taking multiple samples from the crater. Think of it like drawing a blood sample: The phlebotomist swaps tubes in and out as they fill up, only Perseverance swaps the containers as they fill with rock.

Once a tube is full, the drilling arm then docks it back in the bit carousel within the adaptive caching assembly. Now the smaller arm picks up the sample and shuttles it around to different stations. There’s a probe, for example, that measures the volume of the sample and a camera that snaps photos of the tube. Then it’s off to a dispenser that plops a seal into the tube, and then yet another station that pushes down on the seal to activate it. The camera takes a few more pictures of the sample, just to make sure everything looks good, and finally it is sent back to temporary storage in the robot’s belly.

The robot is expected to collect about three dozen samples as it rolls around Mars. “We drive around with these tubes until we’re ready to drop them off in a collective cache,” said Samuels. The tubes will wait in this cache until a future Mars sample return mission picks them up and ferries them to Earth. “The science team is looking for all different types of rocks—sedimentary, igneous—to be able to bring back because they’re going to tell us different things about Mars,” she continued. Once the retrieval mission returns, scientists from many different institutions will be able to study the geology of the red planet.

The robot is doing this autonomously. Like its sibling rovers, Perseverance can’t rely on a human on Earth to constantly pilot it around Mars—it takes up to 20 minutes for radio signals to travel between the two planets. So Perseverance is largely a set-it-and-forget-it kind of science machine. “It is completely hands-off, from the beginning where the sample tube is taken out of storage, and all the way through the sample acquisition process, all the way to the point where it goes back into storage,” said Samuels. “All of that is autonomous.”

And while the first drilling attempt didn’t exactly go as planned, what initially seemed like a problem might actually provide vital clues about the Martian geology. Going into the maneuver, Stack Morgan and other NASA scientists reckoned the rock was either a sedimentary or a basalt (crystalized magma). Given how the rock behaved when drilled, now they are leaning towards basalt, which crystallizes at depth to form coarse grains. “When we started to core this rock, it basically broke up along these kind of disintegrating grain boundaries,” says Stack Morgan.

This is exciting because Perseverance is drilling in a former lake bed. If it can drill into sedimentary rock—layers of muck laid down by the lake—that could potentially provide signatures of microbial life. But igneous rock like basalt provides a timeline: Scientists can date when the magma turned into hard rock.

In other words, Perseverance may have stumbled onto something exhilarating. “Honestly, the best-case scenario would have been that we successfully cored this rock,” says Stack Morgan. “But the next-best scenario is that we have potentially discovered a sequence of rocks where we have the opportunity both to explore the habitability of this area while also providing those age constraints that tell us exactly when Jezero Crater was habitable.”

NASA hasn’t yet released a date for Perseverance’s next move, but chief engineer Louise Jandura wrote in her blog post that the rover will leave the first borehole behind and continue to the next sampling location, which the Ingenuity helicopter has identified as likely to be sedimentary rock “that we anticipate will align better with our Earth-based test experience.”

“The hardware performed as commanded, but the rock did not cooperate this time,” she continued. “It reminds me yet again of the nature of exploration. A specific result is never guaranteed, no matter how much you prepare.”

This story originally appeared on wired.com.