Creating a fibrous structure from soy proteins requires an understanding of the extrusion process. What happens to the proteins along the way, and what determines their structure? Sam Kuijpers investigated these questions during his doctoral research at Wageningen University. Using MRI, he tracked the formation of structures ranging from small protein assemblies to larger layers. His research offers insights for product developers and process technologists working on plant-based meat alternatives.
Kuijpers studied high-moisture extrusion of soy proteins. During this process, the extruder heats and kneads a mixture of proteins and water. The hot mass then flows under pressure through a cooling die.
To track structure formation, he stopped a pilot-scale extruder. He collected samples from different sections of the equipment. Using MRI and complementary measurement techniques, he visualized the protein structures.
Small clusters of proteins bundle into thin protein fibrils inside the extruder. These align primarily in the transition zone and the initial section of the cooling die. Protein-rich and water-rich regions are also already visible in the transition zone. The layered structure then continues to grow as the material passes through the cooling die.
The pH affects how the proteins come together and form structures. Near the isoelectric point, the proteins carry very little electrical charge. At this pH, a network of particles forms without a clear preferred orientation.
At other pH values tested, the protein clusters form elongated structures. These can align with the flow in the extruder. Excessive electrical charge, however, increases repulsion between them. This reduces the alignment of both the protein fibrils and the larger layers.
Kuijpers also developed an MRI setup for measurements during extrusion. This allowed him to measure flow velocities directly inside the cooling die. The protein mass slipping along the wall played a dominant role in these flow patterns. The measurement data provide input for computer models that predict this flow.
Source: Wageningen University