Estimation of irradiation doses in chicken samples using a reaction-based fingerprinting method

Area of Science:

• Food Science
• Analytical Chemistry
• Radiation Chemistry

Background:

• Industrial-scale food irradiation requires rapid, cost-effective dose estimation methods.
• Optical sensing using dose-dependent indicator reactions is an emerging solution.
• Accurate dosimetry is crucial for ensuring food safety and quality after irradiation.

Purpose of the Study:

• To develop and validate a reaction-based optical sensing strategy for estimating radiation doses in food.
• To assess the efficacy of carbocyanine dye reactions coupled with machine learning for dosimetry.
• To investigate the feasibility of dose estimation across different chicken producers.

Main Methods:

• Raw chicken breast samples were irradiated with 1 MeV accelerated electrons.
• Samples were extracted and reacted with carbocyanine dyes, hydrogen peroxide, or hypochlorite.
• Absorbance and fluorescence were measured, and machine learning was applied for data analysis.

Main Results:

• Supervised machine learning accurately discriminated between irradiated (250, 1000, 5000 Gy) and non-irradiated chicken samples from the same producers.
• The method showed promise for dose estimation, with potential for broader application through database construction.
• Variability between producers necessitates a comprehensive database for unknown sample analysis.

Conclusions:

• Reaction-based optical sensing coupled with machine learning offers a viable method for food irradiation dose estimation.
• The developed technique demonstrates high accuracy for samples from known producers.
• Expanding the producer database is key to achieving reliable dose estimation for diverse food sources.