Add two extra neutrons to the lightest element and hydrogen becomes radioactive, earning the name tritium. Even before the Three Mile Island accident in 1979 regulators worried that this ubiquitous by-product of nuclear reactors could pose a threat to human health. The U.S. Environmental Protection Agency (EPA) was only seven years old when it put the first rules on the books for tritium in 1977. But a lot has happened in the intervening decades, and it is not just a longer list of nuclear accidents.
The Chernobyl and Fukushima meltdowns let loose plenty of tritium, but so have a seemingly endless series of leaks at aging reactors in the U.S. and elsewhere. Such leaks have prompted the EPA to announce on February 4 plans to revisit standards for tritium that has found its way into water—so-called tritiated water, or HTO—along with risk limits for individual exposure to radiation and nuclear waste storage, among other issues surrounding nuclear power.
The agency’s recent announcement in the Federal Register notes that tritium levels as high as 3.2 million picocuries per liter (pCi/L) in ground water have been reported to the U.S. Nuclear Regulatory Commission (NRC) at some nuclear facilities. (A curie is a unit of radiation emission; a picocurie is one trillionth of a curie.) That is 160 times higher than the standard set back in 1977 by the fledgling EPA—and the NRC has made measurements even higher at some nuclear facilities. "Because of these releases to groundwater at these sites, and related investigations, the agency considers it prudent to reexamine its initial assumption in 1977 that the water pathway is not a pathway of concern," the EPA stated in its filing.
This new evaluation is likely to prove challenging, however, as tritium is difficult to get a grip on from both a radiological and human health perspective. On the one hand, there is evidence that the risk from tritium is negligible and current standards are more than precautionary. On the other, there is also some evidence that tritium could be more harmful than originally thought.
Or, as a health physicist who has studied tritium for years observes, in the 1970s, the EPA did not rely on any health studies in setting its original standards. Instead, the EPA back-calculated acceptable levels of tritium in water from the radiation exposure delivered by already extant radionuclides from nuclear weapons testing in surface waters. "It's not a health-based standard, it's based on what was easily achievable," remarks David Kocher of the Oak Ridge Center for Risk Analysis, who has evaluated health risks from tritium and spent 30 years at Oak Ridge National Laboratory. The standard of 20,000 pCi/L of drinking water made compliance easy. "No drinking water anywhere was anywhere close, so it cost nothing to meet."
By the EPA's calculations, the 1977 standard should result in an extra radiation dose of less than four millirems, or 40 microsieverts per year, about the amount from a chest X-ray.  (A rem is a dosage unit of x-ray and gamma-ray radiation exposure; one sievert equals 100 rems.) But the standard begs the question: is tritium safe to drink?
Natural background
The EPA will have to take into account complex but sparse data about tritium exposures in formulating new standards. Calculations of exposure levels must take into account not just the levels in waters around nuclear plants but also how much drinking water exposure there is, as well as radiation from natural sources.
High in the atmosphere cosmic rays produce four million curies worth of tritium each year. This atmospheric tritium rains out into surface waters. Nuclear power plants the world over produce roughly the same amount annually, although production (and releases) vary among facilities. For example, the Beaver Creek nuclear power facility in Pennsylvania is the biggest producer of tritiated water in the U.S., per NRC records, churning out roughly 1.5 curies worth per megawatt of electricity produced. Even more escapes in steam from power plants like Palo Verde in Arizona, whose three reactors combine to billow out more than 2,000 curies worth of tritiated steam per year.
But both nuclear power plants and cosmic rays are outweighed by orders of magnitude by the legacy of nuclear bomb testing. Using tritium triggers to explode thermonuclear bombs aboveground produced copious quantities of atmospheric tritium. For every megaton of nuclear blast, roughly seven megacuries of tritium resulted. Despite an end to aboveground testing, leading to a peak in tritium production in 1963, bomb-made tritium lingers, decaying away over a half-life of 12 years. For tritium levels to reach under 1 percent of the original amount released by nuclear weapons testing will thus take seven half-lives, or 84 years. "Setting off all those hydrogen bombs aboveground sent a tremendous pulse into the atmosphere," notes Kocher, who is also a member of the National Council on Radiation Protection and Measurement. "It's basically everywhere."
In fact, everyone drinks tritiated water. "People are exposed to small amounts of tritium every day, since it is widely dispersed in the environment and in the food chain," as the EPA notes in its public information on the radionuclide.
That bomb-made tritium will eventually decay away completely (presuming the test ban holds), leaving power plants and cosmic rays as the major sources, along with minor contributions from the tritium in photoluminescent signs and the like. But nuclear power plants have not done a good job of containing tritium, whether from steam or water leaks at U.S. plants. In 2005 a group of farmers in Illinois successfully sued utility Exelon for tritiated water escaping from the Braidwood nuclear power plant that had contaminated their wells, even though the levels were below those set by the EPA.
And there is at least 400,000 cubic meters of tritiated water now in storage at Japan’s wrecked Fukushima Daiichi nuclear power complex, which suffered multiple meltdowns after the 2011 earthquake and subsequent tsunami. A suite of technologies there filter out 62 different radioactive particles created by the Fukushima meltdowns—leaving out only tritium, largely because it is difficult and expensive to separate water from water. Companies such as Kurion, which already helps filter out radionuclides like cesium, suggest that they have a solution if the Japanese want to eliminate the tritium as well. "It's up to TEPCO [the utility] and the Japanese people to decide what they want to do with that water," says materials scientist Gaetan Bonhomme, vice president of strategic planning and initiatives at Kurion. "It is a radionuclide and it does cause public concern."
The Kurion process concentrates the radionuclide in a small volume of water. A proprietary material then captures the tritium and stores it—and will not release it until heated above 500 degrees Celsius. "It's stable in an accident," Bonhomme notes.
The technology could be applied wherever tritium is produced, including aging nuclear reactors in the U.S. It is the hope of Bonhomme and others that by offering a solution for tritium and other nuclear wastes, they can help ease fears of fission as a source of electricity. But any treatment will be more expensive than simply dumping tritiated water. "If it was really all about science, we would be releasing most of tritium from nuclear power in the water stream, because that's the best way to dilute it," Bonhomme admits.
So the question becomes: Is treating for tritium worth it? And that answer depends on the risk.
The big C
Cancer is the main risk from humans ingesting tritium. When tritium decays it spits out a low-energy electron (roughly 18,000 electron volts) that escapes and slams into DNA, a ribosome or some other biologically important molecule. And, unlike other radionuclides, tritium is usually part of water, so it ends up in all parts of the body and therefore can, in theory, promote any kind of cancer. But that also helps reduce the risk: any tritiated water is typically excreted in less than a month.
Some evidence suggests the kind of radiation emitted by tritium—a so-called beta particle—is actually more effective at causing cancer than the high-energy radiation such as gamma rays, even though skin can block a beta particle. The theory is that the low-energy electron actually produces a greater impact because it doesn't have the energy to travel as far and spread its impact out. At the end of its atomic-scale trip it delivers most of its ionizing energy in one relatively confined track rather than shedding energy all along its path like a higher-energy particle. This is known as density of ionization, and has been shown with the similar form of radiation called an alpha particle.
Ionization is what makes radiation dangerous for human health. Essentially, the radioactive particle smashes into the atom or molecule and pushes out an electron or other particle, leaving that atom or molecule in a charged or ionized state. These charged molecules can then cause other damage as they interact with other atoms and molecules. That includes damage to DNA, genes and other cellular mechanisms. Over time this DNA instability results in a higher chance of cancer. As a result, scientists work under the assumption that any amount of radiation poses a health risk.
Density of ionization suggests tritium exposure may have an increased risk of causing cancer. The National Institute for Occupational Safety and Health calculates compensation due energy workers who develop cancers that may have been caused by exposure to ionizing radiation with such enhanced biological effectiveness of tritium in mind as does the fund for the 200,000 or so personnel who served at nuclear test sites, the atomic veterans (although few had any tritium exposure).
But there is no definitive epidemiological study to assess the true risk of tritium, and animal studies are also lacking. The cancer rates in Japanese survivors of the nuclear bombs dropped on Hiroshima and Nagasaki can reveal little because they were not exposed to tritium either. "You need huge study populations to have any chance of seeing anything," Kocher notes, and that money is simply unavailable. "There is no compelling need to spend the money required to do this."
To make matters even more tricky, tritium's radioactivity is difficult to detect. Because the electron tritium spits out is not a penetrating or high-energy particle, it is hard for radiation monitoring devices to even detect. That makes measuring the radiation dose from tritium difficult. "Dosimetry has been a problem," Kocher notes. "I think a definitive epidemiological study is probably impossible."
In fact, the current National Research Council effort to determine cancer risk from living near a nuclear power plant in the U.S. will not examine the specific risk from tritium leaks. "Our study will not be examining the cancer risks from the leaks as separate events, so it will not be a useful source of information for the purpose of linking cancer occurrence or death from cancer with tritium ingestion," noted Ourania Kosti, director of the ongoing study and a senior program officer at the Institute of Medicine of the National Academies, in an e-mail response.
This lack of data may complicate the EPA’s new rulemaking. Federal regulators might choose to maintain existing standards (as has been done after re-evaluations in the past) or look at what individual states have done, although everywhere the picture remains clouded by uncertainties.
Some states, such as Colorado and California, have set lower goals for the tritium in drinking water. For example, the U.S. Department of Energy has agreed to clean surface waters surrounding its former nuclear weapons production facility Rocky Flats in Colorado to the level of 500 pCi/L. By comparison, the levels of tritiated water found in a monitoring well near the leaking Oyster Creek nuclear power plant in New Jersey reached 4.5 million pCi/L, although no tritiated water has been detected off-site as yet.
At Braidwood in Illinois the tritiated water had spread via leaks in a plume, reaching levels of 1,600 pCi/L in the groundwater under nearby farm fields. If consumed for an entire year, tritiated water at that level would result in an extra dose of radiation of roughly 0.3 millirem. That is 1,000 times smaller than the amount of radiation from natural sources absorbed by the average American in a year and 12 times smaller than the dose absorbed during a single flight across the U.S. For comparison, one chest x-ray, which also falls into the class of radiation that appears to be more biologically effective, results in a dose of four millirem.
The potential innocuousness raises the question of whether more stringent standards are really needed—which is the determination the EPA made the last time it revisited these standards at the end of the 20th century. "I think the levels of tritium in drinking water today are low enough that I wouldn't worry," Kocher says. "The good news about tritium is that: even if you inhale or ingest an awful lot, it is going to flush out of your body." He adds: "Just have a few beers and you're done."