When you toss together a salad or grab a bag of pre-washed spinach, it’s easy to take for granted that your greens are clean and safe to eat. But keeping leafy greens free of harmful bacteria like certain E. coli is one of the trickiest steps in food production.

Between 2009 and 2018, 139 foodborne outbreaks were linked to fresh vegetables, and over 50% were caused by the consumption of leafy vegetables. (Centers for Disease Control and Prevention, 2022) That’s because they’re often eaten raw and go through minimal processing.

To reduce risk, most produce processors rely on chlorine-based washes. But organic matter from cut leaves reacts with chlorine, quickly breaking it down and reducing its effectiveness. This forces producers to use more chlorine, which can lead to the formation of chemical byproducts that aren’t good for our health or the environment.

How can we keep our salads clean without overloading them with chemicals?

The Research: Cleaner Greens with Smarter Systems

A team of AIFS scientists at the University of Illinois at Urbana–Champaign set out to optimize how leafy greens are cleaned using machine learning.

Ultrasound cleaning creates tiny shock waves that loosen bacteria from leaf surfaces. When paired with a traditional chlorine wash, it helps improve product safety without relying on higher chemical levels.

Machine learning helped the researchers determine which washing factors — such as agitation speed, sanitizer concentration, vegetable load, and cut size — had the greatest impact on safety. It then predicted the most effective combinations of ultrasound and chlorine.

To test their approach, the team built a specialized washing system that simulated real processing conditions. They evaluated spinach, romaine lettuce, and iceberg lettuce in different cut sizes under a wide range of wash settings, completing 127 sanitation trials.

Key Findings That Could Change Your Salad

1. More surface area = less effective chlorine. When greens are chopped into smaller pieces, they leak more juices into the wash water. This organic matter reacts with chlorine, reducing its ability to kill bacteria.
2. Not all greens behave the same. Iceberg lettuce, for example, released more organic material than spinach, which means it required more chlorine to stay clean. This shows us that sanitation systems need to be tailored to the specific type and cut of the green.
3. Ultrasound helps, but balance is key. High ultrasound power helped clean bacteria from leaf surfaces, especially in whole spinach leaves. But it also stirred up more organic matter, which can weaken chlorine if not managed carefully.
4. Minimum chlorine thresholds are essential. The team found they needed to start with at least 20 mg/L of free chlorine to keep bacteria from surviving in the wash water. Too little meant a higher risk of contamination. Too much could lead to chemical buildup.

Why It Matters to You

If you ever buy bagged salad, eat pre-cut lettuce, or serve leafy greens at home, this research affects you directly. It helps ensure the food you eat is:

• Safer: Reduces the risk of foodborne illness from contaminated greens.
• Cleaner: Removes more bacteria from produce surfaces without relying solely on chemicals.
• Smarter: Applies data and AI to improve food safety without compromising on sustainability.

The study also gives produce processors practical tools for designing customized cleaning protocols for different types of greens. That means better outcomes without unnecessary waste or chemical overuse.

What’s Next?

The research team sees this work as a step toward more intelligent sanitation systems. By combining traditional experimentation techniques with artificial intelligence, they’ve shown it’s possible to optimize fresh produce cleaning with fewer trial-and-error experiments — and better results.

In the future, this kind of approach could lead to automated sanitation systems that adjust in real time based on the type of produce, the amount of organic matter in the water, and other factors.

“Our goal is to protect public health while also reducing the environmental footprint of food processing,”
— Dr. Hao Feng, lead researcher

Research Team:

Mengyi Dong, Manan Mehta, Caroline Lee, Melannie Kavannaugh, Chenhui Shao, Matthew J. Stasiewicz, and Hao Feng

(University of Illinois at Urbana-Champaign and North Carolina Agricultural and Technical State University)

Published in:

ACS Sustainable Chemistry & Engineering, 2024

Machine Learning and Taguchi DOE Combined Approach for Modeling Dynamic Ultrasound-Assisted Fresh-Cut Leafy Green Sanitation

References:

Centers for Disease Control and Prevention. (2022, May 4). National Outbreak Reporting System (NORS) dashboard