Accompanied by extensive studies, meat analogues have gained significant attention, evolving from niche alternatives into popular options for flexitarians. To further improve their appeal, continued product development is necessary. This includes enhancing sensory properties, such as taste and texture, as well as diversifying the raw materials used. Meeting these demands requires a mechanistic understanding of the process in order to perform targeted product design instead of relying on trial-and-error or experience.
The twin-screw extrusion process is the key technology for producing meat analogues. Despite extensive research, the mechanistic understanding required for targeted product design remains incomplete. In particular, the interplay between local process conditions and resulting structure is not yet fully understood. One reason is the limited availability of methods capable of examining both the process and the material properties under extrusion-relevant conditions.
This thesis contributed to closing this gap by using three complementary methods to investigate and link local processing conditions with material behavior under extrusion: the structure modulation method, the ramp test method, and the Moving-Particle Semi-Implicit (MPS) simulation method. ... mehr
To study the extrusion process, which is complex and involves many modifiable parameters affecting both process conditions and product properties, a model process was required to generate reproducible model products. The first step was therefore to establish a method that allows modification of the final product structure without changing either the process parameters or the protein source. This requirement was fulfilled by the structure modulation method presented in chapter 2. In this method, the pH of the liquid phase was systematically lowered by adding acids. It is well known that changes in pH affect the surface charge of proteins and thereby influence their folding and intermolecular interactions. Accordingly, pH adjustment was expected to alter network formation during extrusion and thus change product properties. Material screening of soy protein isolate (SPI) and wheat gluten doughs in a closed cavity rheometer confirmed these expected effects. Lowering the pH increased the complex viscosity of SPI, whereas gluten displayed altered reaction behavior. Extrusion trials reflected these findings. SPI extrudates became denser and more brittle with decreasing pH, while gluten extrudates became more elastic and dough-like. Mechanical and solubility analyses further revealed that SPI formed a higher density of physical cross-links, while gluten formed fewer disulfide bonds. Based on these results, SPI processed with the structure modulation method was selected as the model system for this thesis, as it produced a controllable set of model products for systematic investigation.
To gain a mechanistic understanding of the extrusion process, the local process conditions must be known. Since these conditions strongly influence product properties but can only be measured to a limited extent, chapter 3 presents a method for determining them using numerical flow simulation. Grid-based simulation approaches are restricted to small sections of the screw and generally assume a fully filled domain. To overcome these limitations, the particle-based MPS method was introduced. Since this method had not been applied to the screw section of extruders before, the first step was to evaluate the effect of particle size on simulation outputs, such as residence time and degree of fill. Data on material parameters were taken from literature on a SPI dough with 55 wt% water. The results showed that particle sizes must be smaller than the gap between screw and barrel to obtain results independent of particle resolution. Based on this, simulations were performed for three kneading block configurations to determine local residence time, degree of fill, and shear rate distributions. The simulated fill levels were compared qualitatively with dead-stop experiments and showed good agreement in the kneading block. Finally, an extended simulation including the energy equation was carried out, demonstrating that a local temperature increase of more than 10 °C can occur in the kneading block, which is an effect not detectable experimentally.
To understand the extrusion process, it is essential not only to know the local process conditions but also how the material behaves under these conditions. Shear viscosity is the key material property describing this behavior. Although several methods exist to determine shear viscosity, none are suitable for protein doughs used in meat analogue production. These mixtures are highly viscoelastic, may undergo reactions at elevated temperatures, and can suffer from water loss or evaporation. The closed cavity rheometer (CCR) can overcome these limitations due to its sealed and pressurized measuring cavity, which enables rheological measurements at extrusion-relevant temperatures without water loss. The rheometer can perform a single controlled rotation within a defined time, the so called ramp test, which should provide access to shear viscosity at steady shear.
To verify the ramp test and assess the accuracy of the resulting shear viscosity data, three commercial polyolefin polymers were investigated in chapter 4: low density polyethylene, linear low density polyethylene, and polybutadiene. Their complex viscosities were measured and compared with shear viscosities obtained using a capillary rheometer and the CCR. Time temperature superposition master curves for the complex viscosity and the CCR-based shear viscosity were developed for linear low density polyethylene and polybutadiene. The influence of the cavity sealing on measurement accuracy was examined using finite element simulations, showing an effect of less than 12%. Overall, the results demonstrated that the ramp test performed with the CCR is a practical method for determining reliable and reproducible shear viscosity data across a temperature range from 50 to 180 °C and for materials covering viscosities from 1,600 to 480,000 Pa*s.
After validating the method with polymers, the ramp test was transferred to soy protein isolate dough with a water content of 55 wt% in chapter 5. Measurements were carried out at temperatures from 95 to 140 °C. In parallel, the protein doughs were examined using a capillary rheometer to identify potential flow instabilities. However, the capillary rheometer data did not align with the ramp test results. The ramp test produced higher consistency indices (K = 3,412–32,237 Pa*s^n}) and lower flow behavior indices (n = 0–0.43) compared to the uncorrected capillary rheometer data (K = 708–13,013 Pa*s^n, n = 0.10–0.48). Visual inspection revealed gel fracture at 95 °Cand 110 °C wall-slip effects, and browning at 140 °C, indicating protein reactions. Overall, chapter 5 presents two approaches for determining shear viscosity of protein douhgs at extrusion-relevant temperatures, identifies their limitations, and discusses how these observations may influence structure formation of meat analogues.
In chapter 6, the ramp test and MPS simulation were then applied to interpret observations made using the structure modulation method on SPI. When examining torque, die pressure, and material temperature, contradictory results were observed. The ramp test showed that lowering the pH increases shear viscosity, while increasing temperature decreases shear viscosity, which is consistent with complex viscosity measurements presented in chapter 2. However, these trends could not explain the unusual torque and pressure behavior. Using the MPS method with shear viscosity data from the ramp test for water-based dough and a pH-reduced dough at 140 °C revealed that the extruder operated at a very low degree of filling. Shear rate distributions remained unchanged, and residence times differed increased slightly with decreasing pH. These results suggest that the unexpected torque and pressure behavior is primarily caused by lubrication effects in the screw section and slippage in the die, rather than by changes in the local flow field. Under low filling conditions, the material temperature is dominated by heat exchange with the barrel rather than by viscous energy dissipation from mechanical input, which explains the nearly constant material temperature.
While the ramp test and MPS simulation helped interpret the observations made with the structure modulation method, each method also provides valuable insights on its own. The structure modulation method allows systematic adjustment of protein networks and targeted modification of meat analogue texture. The ramp test provides direct access to shear viscosity at extrusion-relevant temperatures, supporting material characterization and model development. The MPS simulation approach enables realistic analysis of partially filled screw sections, offering insight into local processing conditions.
Together, these methods contribute to a deeper mechanistic understanding of plant protein extrusion and provide a foundation for more targeted and efficient product and process design.
1