CAMBRIDGE, Mass. — Researchers at the Massachusetts Institute of Technology’s Media Lab have dumped wastewater underground in apparent violation of a state environmental regulation, according to documents and interviews, potentially endangering local waterways in and near the town of Middleton.

Nitrogen levels from the lab’s wastewater registered more than 20 times above the legal limit, according to documents provided by a former Media Lab employee. When water contains large amounts of nitrogen, it can kill fish and deprive infants of oxygen.

Nine months ago, the Massachusetts Department of Environmental Protection began asking questions, but MIT’s health and safety office failed to provide the required water quality reports, according to documents obtained by ProPublica and WBUR. This triggered an ongoing state investigation.

After ProPublica and WBUR contacted MIT for comment, an institute official said the lab in question was pausing its operations while the university and regulators worked on a solution. Tony Sharon, an MIT deputy vice president who oversees the health and safety office, didn’t comment on the specific events described in the documents.

The state’s investigation adds to recent scrutiny of the Media Lab for accepting donations from Jeffrey Epstein, a convicted sex offender who was charged with trafficking minors before he died in jail last month. Joichi Ito, the director of the Media Lab, has resigned, and students have called for the resignation of MIT President L. Rafael Reif, who signed off on at least one of Epstein’s gifts.

The lab responsible for the dumping is the Open Agriculture Initiative, one of many research projects at the Media Lab. Led by principal research scientist Caleb Harper, who was trained as an architect, the initiative has been under fire for overhyping its “food computers”: boxes that could supposedly be programmed to grow crops, but allegedly didn’t work as promised.

Throughout early 2018, the lab’s research site in Middleton, about 20 miles north of the main MIT campus in Cambridge, routinely drained hundreds of gallons of water with nitrogen into an underground disposal well, at concentrations much higher than the lab’s permit allowed, according to documents and interviews. The nitrogen came from a fertilizer mix used to grow plants hydroponically.

The information comes from dozens of emails and lab results shared by Babak Babakinejad, a former researcher in Harper’s lab. Babakinejad said he decided to speak out because he’s worried about the health and environmental impacts of the dumping. Babakinejad’s account of the lab’s actions was confirmed by two other sources with knowledge of the experiments, who asked for anonymity.

Babakinejad told ProPublica and WBUR that he warned Harper and MIT’s Environment, Health and Safety Office (EHS) about the situation after he realized their hydroponic solution exceeded their environmental permit, which limited the wastewater to concentrations of 10 parts per million (ppm) for nitrogen.

EHS is responsible for health and safety throughout the institute, from environmental sustainability to the proper handling of toxic chemicals in research labs.

“Our base fertilizer regiment is at 150 ppm Nitrogen… way above the required limit,” Babakinejad wrote in an April 2018 email to Harper, other Media Lab employees and senior staffers at EHS. “I am looking forward to discuss available options such as diluting our waste water… or apply for an appropriate license.”

Harper responded to Babakinejad within the hour, scolding him for emailing health and safety officials: “Writing emails directly to Senior EHS / Facilities teams at MIT, especially those that effect [sic] our groups ability to do research, without asking [the project’s assistant director] or I to review, comment and approve is inappropriate… If emails are directed to you regarding our teams [sic] EHS responsibilities please redirect them to me until further notice.”

This followed prior emails when Babakinejad had questioned Harper about whether the lab’s food computers could really do what Harper claimed. In news reports about this question, Harper did not address allegations about the project’s shortcomings.

Babakinejad said he later spoke to the Massachusetts Department of Environmental Protection (MassDEP) in the fall of 2018, prompting the agency to take a closer look at the lab’s wastewater disposal permit.

For more than five months, a MassDEP scientist tried to get basic information from MIT’s EHS office about how the lab disposed of its wastewater. This June, the scientist expressed frustration in an email to a senior EHS official:

“MassDEP is concerned about the time that it is taking to provide what should be easy to obtain information regarding the (disposal well) discharges and other on-site discharges,” he wrote. “MassDEP is concerned that MIT still hasn’t indicated to MassDEP its long term solution to the management of spent growing solution wastewater containing unacceptably high concentrations of total nitrogen.”

In a statement, MassDEP spokesman Edmund Coletta stated the agency was “concerned about the wastewater discharge issue connected to the Open Agriculture Initiative’s facility in Middleton (MA) and we are investigating the issue further. However, as this is a potential enforcement matter, I cannot offer any other comments.”

Harper provided a statement through his lawyer, David Siegal: “Mr. Harper and his lab are, and have always been, deeply committed to protecting the environment. He has been and will continue to be fully cooperative with and responsive to MIT’s Department of Environmental Health and Safety and the Massachusetts Department of Environmental Protection in their efforts to make sure the lab conforms to all environmental laws and regulations,” Siegal said.

At this point there is no evidence that the discharge from Harper’s lab has reached local drinking water or the nearby Ipswich River.

Excess nitrogen, when ingested by infants under four months old, can prevent blood from carrying oxygen, which can be fatal if left untreated. Municipal water systems routinely check for contaminants, but homes and businesses that use private drinking water wells are responsible for monitoring their own water. ProPublica and WBUR did not obtain any of those testing results.

Pamela Templer, a Boston University professor who studies biogeochemistry, said nitrogen is an essential component of all living things.

“But at high concentrations, it can become what we consider too much of a good thing,” she said. “In waterways, it can lead to phenomena like harmful algal blooms, which can be toxic to people and pets.”

The type of disposal well used by MIT is part of an Environmental Protection Agency program that handles industrial and municipal waste, said Carl Reeverts, former deputy director of the EPA’s Drinking Water Protection Division. There are more than 650,000 of these “Class V” wells across the country. They are designed to protect underground sources of drinking water, but only if the well is properly built, maintained and regularly inspected.

The wells are considered a lower priority for enforcement than others that store hazardous waste from mining, oil and gas, Reeverts said. In general, Class V wells are “most likely to be mismanaged… It’s the one that may be monitored least of all.”

Wastewater was less of a concern when the initiative was launched in 2015 on MIT’s Cambridge campus, which is connected to a municipal sewer system with a wastewater treatment plant that could handle some nitrogen. But with plans to expand to the school’s more rural Middleton facility, which lacks a public sewer system, questions arose about how to dispose of the water.

In August 2016, a consultant emailed the Media Lab’s director of facilities to explain that the best option was a disposal well if the nitrogen in the lab’s hydroponic water stayed below 10 ppm.

The setup would be easy, requiring just a one-time registration to install it with the EPA’s Underground Injection Control (UIC) program, he wrote.

But if the water showed higher nitrogen concentrations, regulations would be more stringent. “The water will need to be treated as sanitary waste (piped to municipal sewer, a septic system/field, or use a holding tank for monthly pickup by a waste management company),” the consultant wrote.

Samples from the months before and after that email showed a huge range of concentrations, as high as 276 ppm, according to documents provided, indicating that some staff knew they could exceed the nitrogen limits if they built a well.

The lab had a well installed, and in December 2017, Massachusetts regulators granted a permit with restrictions. The permit lists Harper as the well operator and the head of MIT’s EHS Office as the well owner. As part of the permit, MIT can only accept about 1,300 gallons of water per month, and must notify regulators within 10 days if it exceeded the 10 ppm nitrogen limit. Finally, the lab was required to provide monthly reports throughout 2018 showing the nitrogen content of the water discharged into the well.

Babakinejad said he joined Harper’s lab about half a year before it got the permit. He had a PhD in neuroscience and nanotechnology from London’s Imperial College, and saw the Open Agriculture Initiative as a chance to work on food science projects that could improve health care.

He began spending time at the Middleton site, called Bates, in October 2017, overseeing research on cotton and basil. The plants were set up in two shipping containers, each filled with 10 to 12 racks of plants floating in pools of water enriched with fertilizer. All together, the experiments could hold more than 500 gallons of the nitrogen-water mix at a time.

The water had to be changed regularly, both to run new experiments and to prevent the tanks from filling with algae, Babakinejad said. A valve on the bottom of each tank allowed scientists to drain the solution into the well, before replacing it with a new fertilizer mix. Lab workers took regular water samples to track the experiments’ progress. The samples were sent to an outside lab, which analyzed the water for nitrogen and other compounds.

Emails and lab notes from early 2018 show the experiments were in full swing. They were changing the water every two weeks, including on March 23, draining it to “flush” the crops. Documents show samples taken that day had nitrogen levels reaching 222 ppm, which is 22 times the allowable concentration.

Babakinejad said the water, once drained, had to go into the well, because there was no other approved disposal method and nowhere to store hundreds of gallons of wastewater.

He first emailed Harper about his environmental concerns in April 2018: “Our license only allows for 10 ppm [of nitrogen] to be discharged as waste however the nitrogen concentration in fertilisers and sanitation materials is significantly higher than what our registration notice allows.”

Babakinejad repeated the warning in an April 16 email to Harper and EHS officials, prompting Harper’s reply that any emails to EHS should go through him first.

The next day, Phyllis Carter, senior program manager at the EHS office, emailed Harper, Babakinejad and other lab employees, explaining that a sample from the previous week had registered 140 ppm nitrogen. “You are correct in that discharge at these levels is not allowed,” she wrote.

Babakinejad said lab officials met to discuss the problem, but never resolved it. He left in mid-2018, disillusioned both by the nitrogen pollution and concerns that Harper had oversold the lab’s capabilities to funders, when it was struggling with a basic ability to grow plants. He said he felt pushed out, and that Harper retaliated against him for expressing concerns by giving him a work improvement plan that required him to document, in 30-minute increments, how he was using his time. Harper did not comment on the allegations of retaliation or circumstances of Babakinejad’s departure.

Babakinejad said he was particularly disappointed by what he saw as the health and safety department’s failure to enforce the permit.

“This is not about Open Agriculture, per se, or Caleb Harper,” he said. “This is a bigger issue… I took every action I could, to go through the right channels to address it. I came to a point that I realized that the institution, apparently, has made a decision not to address this.”

In January 2019, Joseph Cerutti, a DEP employee who handles its disposal well program, emailed Carter, the EHS officer, asking for the monthly reports her office was required to send to his agency the previous year. Carter had told him the lab hadn’t discharged anything into the well from April through June of 2018, but there were still nine months of missing reports.

After a month without a response, Cerutti wrote back with a terse reminder, adding Harper to the email. If Cerutti didn’t get answers within the next two weeks, he would issue a notice of noncompliance, followed by possible fines and revocation of the permit.

Harper responded quickly, writing, “We have been following the protocol agreed with EHS which was for any agricultural effluent was to be spread in the open field and NOT put into the UIC system.”

Cerutti seemed unaware of this. The lab’s permit only allowed MIT researchers to use the well. “When was the protocol to exclusively discharge the hydroponic growing solution to the open field rather than to the UIC well implemented?” he wrote back.

After a phone call with Carter in April, Cerutti was still left with basic questions. In June, he asked for copies of all nitrogen water sample results since January 2018. Carter responded in early July, attaching results since July 2018, but not the samples from March that frequently showed concentrations more than 10 times the limit.

State regulators did an on-site inspection of the facility in July. The investigation is ongoing.

Sharon, the MIT deputy vice president, issued a statement, saying EHS is “committed to working constructively with MassDEP to find a solution that enables OpenAg’s research at Bates to continue and meets their requirements.”