New device uses carbon nanotubes to detect marijuana in human breath

As the number of states legalizing marijuana grows, so do concerns that users may drive under the influence of the drug. A breathalyzer-type instrument—like the ones used to test for alcohol impairment—would permit a fast, accurate, and noninvasive means to test drivers’ breath for the main psychoactive compound in marijuana, tetrahydrocannabinol (THC). Now, researchers have designed a sensitive, handheld prototype THC detector that contains a sensor made of carbon nanotubes (ACS Sens. 2019, DOI: 10.1021/acssensors.9b00762).

Current efforts to detect THC in breath involve analytical techniques such as colorimetric detection or miniaturized mass spectrometry. But THC pharmacodynamics expert Marilyn A. Huestis of Thomas Jefferson University, who was not involved in the study, says carbon nanotube–based sensors have the potential to offer unprecedented sensitivity. Carbon nanotubes have a large surface area to bind molecules, making them excellent sensors. They are also stable in the face of high temperatures and humidity.

These properties motivated Alexander Star of the University of Pittsburgh and colleagues to fabricate a sensor out of semiconducting carbon nanotubes. They built a 3-D printed case around it, creating a device similar to an alcohol breathalyzer. With their prototype, the researchers can measure the change in the nanotubes’ electrical resistance when molecules bind to them and get released.

To be practical, the sensor would need to differentiate THC from the mixture of other chemicals in human breath. So Star’s team exposed the sensor to such a mix, including compounds like carbon dioxide, water, ethanol, and acetone. They discovered they could distinguish THC from these other compounds by monitoring how long it took the sensor to return to its normal voltage after the bound THC molecules were released from the nanotubes. Different species were released at different rates over time, but THC stayed bound the longest. They trained a machine learning algorithm to sift through these recovery traces, sorting their lab-made vapor samples into those that contained THC and those that did not.

The researchers then tested their breathalyzer prototype using samples taken from a human volunteer and showed that it performed comparably to mass spectrometry, the gold standard for detecting breath components, though it’s much less portable.

Huestis says the next step is to ensure the device is specific for THC and doesn’t get confused between other drugs like cocaine or nonpsychoactive cannabinoids with similar chemical structures.

The prototype cannot yet deduce exact concentrations of THC from random mixtures of unknown compounds, but Huestis points out that US lawmakers have not instituted a standardized legal limit for marijuana, so currently there is not a target concentration for detection. And because THC affects chronic and daily users differently, “there is no one number that differentiates ‘cannabis-impaired’ from ‘non-impaired,’” she says. She advises that THC breathalyzers be paired with behavioral tests to gauge intoxication.

Carbon nanotube expert Meyya Meyyappan of NASA’s Ames Research Center, who was not involved in the study, says the high sensitivity of the researchers’ prototype instrument is a great start. “This is a good story,” he says. “It can go places.”