Pitt Researchers Develop Marijuana Breathalyzer Technology

Associate professor of electrical and computer engineering Ervin Sejdic, left, and professor of chemistry Alexander Star show the prototype of the THC Breathalyzer developed using their interdisciplinary research. Medical and recreational marijuana sales in the United States are poised to reach more than $12 billion, according to the 2019 Marijuana Business Factbook. As states across the country consider ways to introduce the product into their markets, they are also considering ways to detect intoxication and establish guidelines for safe use.

A new device from an interdisciplinary team at the Dietrich School of Arts and Sciences Department of Chemistry and the Swanson School of Engineering could be a critical tool in helping law enforcement accurately detect when a driver is under the influence of marijuana. The team, led by professor of chemistry Alexander Star and associate professor of electrical and computer engineering Ervin Sejdic, has developed a device capable of measuring tetrahydrocannabinol (THC), the psychoactive compound in marijuana, in a user’s breath.

Current drug testing methods rely on blood, urine or hair samples and therefore cannot be done in the field. They also only reveal that the user has recently inhaled the drug, not that they are currently under the influence.

“In legal states, you’ll see road signs that say, “Drive High, get a DUI,’ but there has not been a reliable and practical way to enforce that,” said Star. “There are debates in the legal community about what levels of THC would amount to a DUI, but creating such a device is an important first step toward making sure people don’t partake and drive.”

The breathalyzer was developed using carbon nanotubes, tiny tubes of carbon 100,000 times smaller than a human hair. The THC molecule, along with other molecules in the breath, bind to the surface of the nanotubes and change their electrical properties. The speed at which the electrical current’s recover then signals whether THC is present.

“The semiconductor carbon nanotubes that we are using weren’t available even a few years ago,” said Sean Hwang, coauthor of the project’s research paper and a doctoral candidate in chemistry working in Star Research Group. “We used machine learning to ‘teach’ the breathalyzer to recognize the presence of THC based on the electrical currents recovery time, even when there are other substances, like alcohol, present in the breath.”

Said Star: “We can use nanotube sensors to detect THC at levels that are comparable to or better than current measurements that are considered to be the gold standard, such as mass spectrometry."

The prototype looks similar to a breathalyzer for alcohol, with a plastic casing, protruding mouthpiece, and digital display. It was tested in the lab and was shown to be able to detect the THC in a breath sample that also contained components like carbon dioxide, water, ethanol, methanol, and acetone. The researchers will continue to test the prototype but hope it will soon move to manufacturing and be available for use.

“Creating a prototype that would work in the field was a crucial step in making this technology applicable,” says Sejdic. “It took a cross-disciplinary team to turn this idea into a usable device that’s vital for keeping the roads safe.” 

Star said the next step is to use the prototype device to measure “as many volunteers as we can” to see how accurately the machine learning model is able to distinguish between THC and non-THC components.

He said it’s up to legislators to determine what level of THC detection equals intoxication, but the device is an important step in helping to make that decision.

With legalization of marijuana, the medical research community and legal community are playing catch up. We think our device gives a tool to researchers, users and the legal community to use to build that knowledge,” Star said.

In July, the paper, “Tetrahydrocannabinol (THC) Detection Using Semiconductor-Enriched Single-Walled Carbon Nanotube Chemiresistors,” was published in ACS Sensors.

In addition to Star, Sejdic and Hwang, the work was coauthored by Long Bian, David White, Seth Burkert, Raymond Euler, Brett Sopher and Miranda Vinay from the Department of Chemistry, and Nicholas Franconi, Michael Rothfuss and Kara Bocan from the Department of Electrical and Computer Engineering.