Branching out: ASU engineering leads new cross-enterprise food safety project

Outbreaks of E. coli and other sources of food contamination are frustratingly common. For example, the Centers for Disease Control and Prevention documented three outbreaks of salmonella in mushrooms, peaches and onions across the United States during 2020, and seven similar outbreaks in 2019. In total, more than 48 million Americans are sickened by foodborne illnesses each year, costing the economy more than $15 billion.

To combat this persistent problem, the U.S. Food and Drug Administration and other government agencies are prioritizing improved safety measures across all sectors of food production, processing, distribution and preparation. A core element of these efforts is better technology-enabled food traceability, so a new, multidisciplinary project funded by the U.S. Department of Agriculture and led by Arizona State University is working on a groundbreaking solution.

Michael Kozicki, a professor of electrical engineering in the Ira A. Fulton Schools of Engineering at ASU, is directing an effort to create dendritic tags as a means to securely identify food at any point in the supply chain. Dendrites are tree-like shapes that occur abundantly in the natural world. Examples include the branches of trees, but also the streams and tributaries of river systems or the blood vessels and nerves in the human body.

“These patterns form with a high degree of entropy, so no two dendrites are exactly the same,” Kozicki said. “And since dendrites are relatively easy to produce electrochemically or photochemically, we can cheaply manufacture dendritic tags or labels offering truly singular identities that are effectively impossible to forge or duplicate, unlike a bar code or QR code.”

Application of this dendritic technology could include, for example, labeling every head of commercially grown lettuce with the identity of the farm, field and row from which it is sourced. Such precision could enable a level of traceability that dramatically reduces the impact of a contamination incident. A small batch of tainted lettuce could be more quickly identified and isolated in the supply chain, preventing human illness and sparing tons of safe but suspect food that currently is destroyed out of caution.

This novel opportunity is particularly remarkable because it seems serendipitous. Kozicki says he has no professional experience in the agriculture industry. His pioneering work with dendrites relates to the development of next-generation computer memory through research in the School of Electrical, Computer and Energy Engineering, one of the six Fulton Schools.

“If I had been asked about using dendrites to create unique, unclonable ‘fingerprints’ or marks of authenticity, I might have thought about expensive aviation electronics or defense industry equipment,” Kozicki said. “Not fruits and vegetables.”

But encouragement and seed money from the Fulton Schools’ Office of the Vice Dean for Research and Innovation, ASU’s Swette Center for Sustainable Food Systems and The Biomimicry Center supported Kozicki’s investigation of the idea and his successful proposal for the National Institute of Food and Agriculture grant now funding the two-year project.

Kozicki and Yago Gonzalez Velo, an assistant research professor of electrical engineering in the Fulton Schools, have started working with students in the lab to improve and scale the currently manual and time-consuming process of dendritic fabrication using an electrolyte solution. Alongside tag production, they will test their output with stretching, bending, abrasion, heat, humidity and other factors that represent the rigors of the food supply chain.

Making the resulting identity tags or labels function for their intended purpose will be the work of a collaborator with image processing expertise at Northern Arizona University. Abolfazl Razi, an assistant professor of electrical engineering and computer science, will develop the algorithm and the reading system necessary to verify tag data using cell-phone and cloud-based software platforms.

“On the network side, we’ll need a reference library of images with which we can develop the algorithm to authenticate these dendritic tags,” Razi said. “We’ll be implementing techniques like graph theory and also deep learning methods to reconstruct an image of what is actually a 3D shape and then verify its legitimacy down to the nano scale. Additionally, we need to develop a cell phone adapter device and an app to make this system easy to use by industry and consumers.”

Another important aspect of real-world application is integrating this innovative technology with current food systems processing materials and equipment. This part of the project will be led by Mark Manfredo, a professor of agribusiness for the Morrison School of Agribusiness in the W. P. Carey School of Business at ASU.

“We look forward to working with our local industry contacts to help test what is being developed,” Manfredo said. “We’re already engaging with a large grower of organic greens, and we also hope to work with a melon grower in the state. We need to learn more about the supply chains for these commodities and ultimately evaluate the new tags in commercial settings.”

Manfredo says the project team additionally needs to consider the most economically feasible place within the supply chain to adopt these dendritic identifiers in the context of current systems.

“Is it with the growers? Or the processors? Or with retailers?” he asked. “So, we’ll look at all the incremental costs of implementing the tags at different stages.”

Manfredo also says there is incremental value to consider in adopting this innovation.

“What is the economic value of applying these unclonable tags? Certainly, there is value in waste reduction. But the data also represent marketing opportunities,” he said. “And, of course, the public health value is just enormous.”