Widespread radioactive plume in 2017 likely came from Russia

An accident during the reprocessing of spent nuclear fuel in southwestern Russia is the likely source of a plume of radioactive ruthenium detected throughout Europe in 2017, according to a new study. The analysis, published last month in Proceedings of the National Academy of Sciences (PNAS), reinforces the conclusions of a previous report that singled out the Mayak Production Association, which had been producing a radioactive antineutrino source for a particle-physics experiment. Mayak and Russian officials deny that they are responsible.

The concentrations of ruthenium-106 weren’t high enough to endanger the public. Still, researchers hope to definitively identify the cause so that they can work toward preventing something similar from happening again. “It’s not that we want to point fingers at Russia,” says study coauthor Georg Steinhauser, a radioecologist at the Institute of Radioecology and Radiation Protection in Hannover, Germany. “We want to learn lessons.” His call comes as Russian officials downplay a recent release of radioactive material at one of their military bases.

Narrowing down the culprit

The first hint that something was amiss came on 2 October 2017, when a laboratory in Milan, Italy, reported airborne ¹⁰⁶Ru levels on the order of millibecquerels per cubic meter; its detectors are sensitive to levels as low as 0.1 μBq m^–3. The Milan lab is part of a network of European laboratories established in the aftermath of the 1986 Chernobyl accident that share precise measurements of radionuclides in the environment. Within the next few days, 27 other European countries detected concentrations up to about 150 mBq m^–3. The ¹⁰⁶Ru levels were the highest reported for that isotope in Europe since Chernobyl.

Despite the significant contamination, no country or institution claimed responsibility, and some, including Mayak, specifically denied they were the source. That left the radiation safety community to do some detective work. For a report published in November 2017, researchers at the French Institute of Radioprotection and Nuclear Safety (IRSN) plugged the environmental measurements into an atmospheric dispersion model and narrowed down the site of the release to the Volga and Ural regions in southwestern Russia.

Radioactive isotopes of ruthenium form only via nuclear fission (see sidebar). The fact that no other radioactive species were detected ruled out a mishap at a nuclear reactor, and the large concentration eliminated small-scale sources such as medical devices. That left one likely cause: an accident during the manufacture of a sizable radioactive source using spent reactor fuel, which contains ¹⁰⁶Ru along with many other radionuclides. At the time of the detections, the Mayak plant, located outside Chelyabinsk in the Ural region, was extracting chemicals from spent fuel to create a cerium-144 source for the SOX experiment at Italy’s Gran Sasso National Laboratory.

The IRSN researchers encouraged the radioprotection community to investigate further. Russian officials again denied responsibility, and the Russian Academy of Sciences convened a commission of international researchers to conduct an investigation. Several months after the release, the SOX experiment was canceled because Mayak was unable to meet the project’s specifications.

In the new study, Steinhauser, Olivier Masson of IRSN, and colleagues confirmed the 2017 IRSN report’s conclusions. The researchers reviewed 1300 measurements taken across Europe from late September through mid October 2017 to track the movement of the radioactive ruthenium plume. They relied heavily on air-filter readings from Romania, which has several dozen monitoring sites that record concentrations multiple times a day, as compared with weekly readings at most other European sites. The Romanian data revealed a short-duration plume that moved westward across the country from 29 September to 3 October.

The researchers analyzed about 200 samples of rainwater along with grass and soil, where radionuclides can collect after falling with precipitation. Four sites within 20 kilometers of Mayak had high concentrations of ¹⁰⁶Ru during the last week of September 2017. Two months later, a French nongovernmental organization conducted a soil analysis just outside a restricted zone around Mayak. Only one of eight samples contained significant levels of ¹⁰⁶Ru. But that one sample was taken from west-southwest of the facility—along the likely trajectory of the plume based on air measurements.

How it probably happened

Using forensic chemistry, the PNAS authors then deduced what might have triggered the release. Along with ¹⁰⁶Ru, which has a 372-day half-life, some air-monitoring stations also detected ¹⁰³Ru (39 days). The difference in decay rates enabled the researchers to calculate when the ruthenium species was produced, which coincides with the time elapsed since the nuclear fuel had been used in a reactor. They came up with about a year and a half, an incredibly short amount of time. Most reprocessing facilities wait at least 4 years, and usually more like 10, to handle spent fuel, Steinhauser says.

What was SOX?

The SOX (Short Distance Oscillations with Borexino) experiment was designed to search for evidence of a fourth neutrino flavor known as a sterile neutrino. The plan was to use the long-running Borexino particle detector (pictured above) at the underground Gran Sasso National Laboratory in Italy to clock antineutrinos from a source located meters away. The source was to be made of the radioisotope cerium-144; its two-step decay includes emission of a high-energy antineutrino. The researchers hoped to measure enough particles over 18 months to evaluate whether the three known neutrino flavors were sufficient to account for the particles’ oscillations between source and detector. However, Mayak proved unable to make a source that would emit enough antineutrinos, which led France’s CEA-Saclay and Italy’s National Institute of Nuclear Physics to cancel SOX in February 2018.

Why would anyone handle such radioactive fuel? The specifications of the detector for the neutrino experiment provided the researchers with an answer. To achieve the desired flux of antineutrinos, SOX physicists had asked for a high-purity ¹⁴⁴Ce source with very high radioactivity (see sidebar). After a facility in France turned the collaboration down, Mayak agreed to chemically extract the ¹⁴⁴Ce from spent fuel and produce the source. However, ¹⁴⁴Ce’s 285-day half-life makes it difficult to produce such a source because of the increasing difficulty over time in packing enough of the radionuclide into a target small enough for the physicists’ needs. “The only way they could possibly do it,” Steinhauser says, “was to reduce the cooling time of the nuclear fuel,” a dangerous proposition considering the radioactivity levels.

The ruthenium release itself, which the researchers estimate at about 250 terabecquerels, may have occurred via multiple scenarios. For example, gaseous ruthenium in the form of RuO₄ could have escaped via a flaw as subtle as a pinhole in a pipe. Molecules of RuO₄ are reduced in the atmosphere to form insoluble RuO₂. Yet to their surprise, the researchers found that some of the ruthenium samples they tested were highly soluble, which suggests an additional means of ¹⁰⁶Ru release. Though the researchers don’t speculate in the paper, Steinhauser says he suspects that a fire or explosion heated and ultimately volatilized ruthenium-containing ionic compounds that had been separated from the fuel.

At an impasse

Rosatom, the Russian state corporation that oversees Mayak, said in a statement that national and international experts have found no evidence that the ¹⁰⁶Ru originated at the site, nor have they seen any signs of an accident or human exposure. The spokespeople for the SOX experiment pointed to a February 2018 statement about the project’s cancellation. In a December 2017 email shown to a journalist from Science News, SOX leaders reported to their collaborators that “unexpected problems occurred” during the purification process at Mayak.

Leonid Bolshov, chair of the Russian Academy commission, says that although he agrees with the data sets used in the study, he does not accept the conclusions. He points to committee-commissioned measurements of soils around Mayak that revealed low levels of contamination. And he says there aren’t enough data to dismiss alternative explanations, in particular that a satellite powered by a radioisotope thermoelectric generator (RTG) broke up in the atmosphere. But the IRSN and PNAS reports say that is extremely unlikely: International officials confirmed that no objects with RTGs reentered Earth’s atmosphere during that time, and the airborne measurements are far more consistent with a ground source.

Jean-Luc Lachaume, commissioner of the French Nuclear Safety Authority and former IRSN director delegate for crisis management, presented the IRSN findings in the two meetings at the Russian Academy, in January and April 2018. He says that all the international members of the committee—from France, Finland, Sweden, Germany, Norway, and the UK—agreed that the source was likely located within the Mayak region, but the Russian members dissented. “That’s why there’s no clear conclusion,” Lachaume says. Considering such a fundamental disagreement, he does not expect the commission to reconvene or to issue a final report.

That’s unfortunate, Steinhauser says, because of the seriousness of what occurred. The release likely qualifies as a level 4 or 5 on the International Atomic Energy Agency’s International Nuclear and Radiological Event Scale, according to Lachaume. For comparison, the 1979 Three Mile Island disaster was a level 5; Chernobyl was a level 7. However, the IAEA doesn’t assign ratings until a member state reports an event within its territory. Members are encouraged to share information on any event more serious than level 2.