From thermal randomness to ordered structures: : the effect of mechanical gradients on the thermodynamics of phase separation during extrusion processes

Media type: E-Book

Title: From thermal randomness to ordered structures : the effect of mechanical gradients on the thermodynamics of phase separation during extrusion processes

Contributor: Sandoval Murillo, José Luis [Author]; Hiermaier, Stefan [Degree supervisor]

Corporation: Albert-Ludwigs-Universität Freiburg, Fakultät für Angewandte Wissenschaften

Published: Freiburg: Universität, 2021

Extent: Online-Ressource

Language: English

DOI: 10.6094/UNIFR/194035

Identifier:

Keywords: Thermodynamics ; Spinodale Entmischung ; Entmischung ; Simulation ; Cahn-Hilliard-Gleichung ; Spektralmethode ; Gitter-Boltzmann-Methode ; Finite-Punktmengen-Methode ; Strukturbildung ; Strömungsmechanik ; Material point method ; Extrusion cooking ; Phase separation ; Texturization ; Extrusion simulation ; (local)doctoralThesis

Origination:

University thesis: Dissertation, Universität Freiburg, 2020

Footnote:

Description: Abstract: This thesis presents a novel numerical approach to model the texturization process during high-moisture extrusion cooking (HMEC) of meat substitutes. Pie-protein based meat substitutes manufactured by means of HMEC are some of the most successful vegetarian food products in the market. The reason why this products are so well suitable as an alternative to meat is their unique anisotropic texture which remarkably emulate the fibrous structures of meat. The characteristic texture is formed during the final stage of manufacture, where the extrudate flows through a cooling channel following the hot extrusion of an originally homogeneous mixture of water and pea proteins. A great amount of experimental work has been carried out in order to obtain the optimal texture which best resembles meat. Interestingly, the desired fibrous structures can only be achieved within a very narrow range of process parameters, namely extrusion speed and temperature. However, a completely clear explanation why and how texture formation takes place during HMEC has not been yet found. To fully understand this texturization mechanism would enable the production of extrudates that emulate the desired sort of meat at will just by wisely choosing the corresponding parameters of HMEC. Thus, reducing or even eliminating expensive try-and-error procedures. The current state of knowledge is that shear flow and gradual cooling play an essential role in texture formation as protein denaturation changes the molecular structure and alignment along the flow direction. On the other hand, experimental analysis have shown that the meat-like structures consist of water-rich and protein-rich domains. Based on these facts, this work presents a novel mechanism which could well explain the process of texturization during HMEC. The proposed model suggests that the water-rich and protein-rich domains form due to a phase separation process which is extremely sensitive to changes in shear rate and, most importantly, in temperature. This means that there is only a narrow range in the combination of these parameters where the phase separation leads to the desired fibrous, meat-like structures. Any variation outside that range would inhibit the texturization process as intended. This mechanism is successfully verified by means of numerical simulations using a state-of-the-art, mesh-free method, namely the Material Point Method (MPM). MPM is a continuum-based approach that incorporates both Eulerian and Lagrangian aspects. This property of the method enables the simulation of large deformations and, at the same time, multi-phasic or multi-component material behavior, \ie phase separation. Both, are essential features to implement the mathematical models involved in the proposed texturization mechanism in HMEC. In this regard, the Cahn-Hilliard equation was chosen to describe the phase separation. This model describes a process also known as spinodal decomposition by which the components of a binary substance separate spontaneously and form domains rich in either one of the two components. To account for the influence of shear and temperature, this equation is coupled with the constitutive equations of motion and heat transfer, correspondingly. The complete model is implemented in an own-written MPM-code which incorporates a Fourier spectral method that, in combination with a semi-implicit time integration, dramatically increases the efficiency when solving the Cahn-Hilliard equation. Further modifications to the original MPM-algorithm improve the description of the velocity field as well. Simulation results demonstrate that fiber-like, lamellar structures are obtained when the rates of flow, heat transfer and phase separation are matched in a narrow window, which is in good agreement with experimental observations. The findings highlight that the dominant mechanism leading to the meat-like texture is a temperature gradient driving the phase separation. Indeed, the gradient direction determines, predominantly, the orientation of the fibrous structures. Therefore, the attention is then focused on texturization driven by phase separation under temperature gradients, regardless of the extrudate's complex flow behavior. For that purpose, the Lattice-Boltzmann method (LBM) is used. LBM is a class of the computational-fluid-dynamics (CFD) methods based on statistic mechanics. The use of this method is justified by two main reasons. First, its versatility to simulate multicomponent flows makes it better suited to solve the Cahn-Hilliard equation. Secondly, to prove that the anisotropic structures obtained in the MPM-simulations are not a numerical artifact attributed to the periodic boundary conditions associated to the Fourier spectral method. LBM-simulations were carried out to find a correlation between physical properties, \ie thermal and phase separation parameters, and the anisotropy of the extrudate's texture. Using a computational tool for image processing together with the concept of nematic order parameter from the liquid-crystal theory, a technique is developed here to quantify texture anisotropy of the simulated extrudates with a single scalar number. This order parameter is correlated with a newly defined, dimensionless number that determines the rate ratio of thermal diffusion to phase separation. Finally, a parameter study is carried out where different mathematical models are analyzed, \eg power-law and saturation functions. The aim was to find a possible governing equation of texturization anisotropy as a function of the new dimensionless number mentioned above. The findings and modeling methodologies presented here could be used in the future to predict and asses texturization processes. Applications can be found not only in HMEC but also in other fields of physics and engineering, such as binary alloys and polymers

Access State: Open Access

Search in field:

Recently searched for: