Particle.news
Download on the App Store
3 ARTICLES
·
3d ago

UC Berkeley Builds 16‑Sensor Electronic Nose to Detect Food Spoilage and Allergens

A lab prototype that pairs a tiny carbon‑nanotube sensor array with machine learning will move into real‑world testing to check performance outside controlled chambers.

An “electronic nose” created by UC Berkeley researchers can detect the gases emitted by spoiled food and food allergens better than human noses. (Credit: Brandon Sánchez-Mejia/UC Berkeley)

Overview

  • The research team published their results in Science Advances, with the study released June 17, 2026, describing a proof‑of‑concept chip that reads airborne chemical fingerprints to identify specific foods and spoilage stages.
  • The device is a fingernail‑sized chip with 16 different gas sensors built on carbon nanotube conductors that operate at room temperature and can be made by simple drop‑casting of sensing films.
  • Researchers trained machine‑learning models on lab data to classify seven foods and to tell fresh from 24‑ and 48‑hour spoiled milk, eggs, and raw chicken, reporting about 92.6% overall accuracy and 99.0% accuracy on spoilage when models were specialized.
  • In controlled tests the nose detected trace allergens, sensing as little as 0.05 grams of walnut, and the team has built a portable iPhone‑connected prototype for follow‑up trials.
  • Key limits remain: all results come from sealed‑chamber lab tests using dry carrier air, so the team must now test mixed odors, humidity, refrigerator conditions, model generalization, and manufacturability before the technology can be used in homes or food industry settings.