Researchers have trained an e-nose, which detects chemicals in the air, to identify the origin of oil based on the proportion of volatile organic compounds. Since oil is a mixture of different hydrocarbons present in varying proportions depending on its source, the e-nose could be used to locate soil-contaminating oil spills, monitor the environment at refineries, and conduct oil field studies.
The e-nose is 20 times less expensive than sophisticated devices currently used to measure the composition of gas mixtures. Its low cost will allow it to be used in a wider range of applications, including field studies. The research was published in the Journal of Hazardous Materials.
The e-nose uses the latest technology to detect volatile chemicals in gas mixtures such as air. Like the human nose, it analyzes the odor in its entirety, with all the compounds it contains, and should be trained to identify the target odor. This distinguishes it from tandem gas chromatograph-mass spectrometers, which laboratories use to determine chemical composition by testing each component of the mixture individually.
Despite their high accuracy and sensitivity, tandem gas chromatograph-mass spectrometers are too expensive and bulky to be used in the field, for example in environmental monitoring, to detect soil contamination by oil or to analyze the composition of oil and oil reservoirs at production sites. Although the e-nose needs to be properly trained and tuned, it is mobile and compact, making chemical analysis much cheaper and easier.
Researchers from the Skolkovo Institute of Science and Technology in Moscow and their colleagues proposed using the e-nose to determine the components of oil. They took nine samples of lighter and heavier oil with different proportions of volatile hydrocarbons from three oil fields in Kazakhstan.
Lighter oil evaporates faster and is therefore harder to detect after some time after a spill, while heavier oil evaporates slowly and escapes into the environment in small amounts, making it difficult to immediately pinpoint the source.
The researchers added oil samples to a soil sample and held them under a stream of air in a gas analyzer for 30 minutes. The stream of air containing volatile compounds was then analyzed using the e-nose.
Customization of e-nose algorithms. © Timur Sabirov
The team used an AI-powered system called a “random forest” model to train the e-nose. The model’s algorithm analyzes the effects of a series of related decisions to arrive at a result. The researchers tuned the processing of data from eight sensors so that the e-nose could accurately identify the origin of the oil, no matter how volatile. In addition, the e-nose was able to detect oil in the soil even 12 hours after sampling, when some of the oil had already evaporated.
The e-nose is 20 times less expensive than a tandem gas chromatograph-mass spectrometer that takes up several square meters of laboratory space, while the e-nose is portable and as compact as a paperback, making it suitable for a variety of applications, such as detecting oil contamination in soil, verifying compliance of emissions from oil refineries, and much more.
“Our approach can be used to monitor reservoir productivity in oil fields. The e-nose will determine the chemical properties of volatile oil compounds in the air based on their smell. I suppose the e-nose can also help to identify the exact locations of new oil fields,” says Valery Zaitsev, a Ph.D. student at Skoltech’s Laboratory of Nanomaterials and a participant in the project.
“We will continue our research at a Skoltech startup to provide industries with a device and an algorithm for monitoring emissions and training the e-nose for specific tasks. Since our device is expected to be used by industrial companies, our immediate goal is to implement it as a commercial product.
“In the meantime, we will try to teach the e-nose to perceive odors like humans do, for example, to determine whether an odor is pleasant or unpleasant and to what extent,” says Fedor Fedorov, an assistant professor at the Laboratory of Nanomaterials at Skoltech’s Photonics Center and the head of the project.
Other organizations involved in the study include the L.N. Gumilev Eurasian National University (Kazakhstan), the A.V. Topchiev Institute of Petrochemical Synthesis of RAS (Moscow), Xi’an University of Architecture and Technology (China) and Bauman Moscow State Technical University (Moscow).
More information:
Valeriy Zaytsev et al, Coding smell patterns of crude oil by the electronic nose: A soil pollution case, Journal of Hazardous Materials (2024). DOI: 10.1016/j.jhazmat.2024.135838
Provided by
Skolkovo Institute of Science and Technology
Citation:
AI-powered e-nose can detect oil spills efficiently (2024, November 11)
