Sensors composed of "frozen smoke" and that use artificial intelligence (AI) techniques to detect formaldehyde in real time have been developed by researchers from the University of Cambridge.

Capable of real-time detection of formaldehyde at concentrations as low as eight parts per billion — far surpassing the sensitivity of most indoor air quality sensors — the 'frozen smoke' sensors are composed of aerogels with precisely engineered holes. According to the researchers, thanks to these specially designed holes, the sensors were able to detect the fingerprint of formaldehyde, which is a common indoor air pollutant.

The researchers explained that the proof-of-concept sensors could be fine-tuned to detect a range of hazardous gases, including formaldehyde, which is a volatile organic compound (VOC) typically emitted by household items — such as pressed wood products, wallpapers and paints, and assorted synthetic fabrics.

"VOCs such as formaldehyde can lead to serious health problems with prolonged exposure even at low concentrations, but current sensors don't have the sensitivity or selectivity to distinguish between VOCs that have different impacts on health," explained the researchers. "We wanted to develop a sensor that is small and doesn't use much power, but can selectively detect formaldehyde at low concentrations."

Basing their sensors on ultra-light aerogels often referred to as "liquid smoke," because they are more than 99% air by volume, the researchers explained that the open structure of aerogels lets gases move in and out of the structure easily. Further, the precisely engineered shape of the holes enable the aerogels to function as effective sensors.

To create the sensors, the researchers 3D printed lines of a paste composed of graphene, which is a two-dimensional form of carbon. The lines of graphene paste were then freeze-dried to form the holes in the aerogel structure. In addition to the holes, the aerogels also contain tiny semiconductors called quantum dots.

In the lab, the researchers found that the sensors detected formaldehyde at concentrations as low as eight parts per billion at room temperature, unlike traditional gas sensors that need to be heated up.

Eventually, the researchers added machine learning algorithms into the sensors, which were trained to detect the 'fingerprint' of different gases, thereby enabling the sensor to distinguish formaldehyde’s fingerprint from other VOCs.

The sensors are detailed in the article, “Real-time, noise and drift resilient formaldehyde sensing at room temperature with aerogel filaments,” which appears in the journal Science Advances.