Separation of isoflavones form okara : process mechanisms & synthesis

By-product utilisation, more efficient use of resources, and more sustainable processing have become of the utmost importance for society and the food industry. During soymilk production, a by-product called okara is produced in great quantities. Despite being a by-product, okara contains many nutrients, which could be utilised for human consumption. Isoflavones are one example of the components present in soy, which are also found in okara. Isoflavones are a subclass of flavonoids, a group of the many polyphenols that exist, and are believed to have a positive effect in the prevention of hormone related cancers, cardiovascular disease, osteoporosis and obesity. One of the main challenges when extracting isoflavones from okara are their low concentration in okara and the strong water binding capacity of the biopolymer matrix of okara. Besides, isoflavones comprise several classes of components having different properties with respect to for example their solubility. This makes the separation and purification of those components in a cost-efficient manner with a non-toxic solvent route challenging.

The aim of this thesis is to provide insight for the development of a polyphenol separation process on the case of okara. The presence of a partly solid matrix and its impact on the separation process was investigated, and the opportunities and challenges for isoflavone separation from okara are presented in this thesis. By means of this case, the applicability of a process synthesis methodology for separation processes was investigated.

In chapter 2, the use of ethanol and water was investigated as relatively non-toxic solvents for extraction of the isoflavones. High extraction yields were obtained with 50-70% ethanol. However, different optima in the range of 0%-100% ethanol were obtained dependent on the isoflavone group, which highly differ in hydrophobicity. The glucosides had an optimum at 50%-60% ethanol, the malonyl-glucosides between 30% and 60% ethanol, and the aglycone yield increased monotonously with the ethanol concentration. The solvent choice does not just determine the yield of isoflavones in the extract, but also the swelling of the matrix, important for the separation of the solvent and the okara, and the purity in the extract. While finding a good solvent for a range of components is one issue, a second challenge in the utilisation of okara is its high moisture content, because this limits the ethanol concentrations that can be used in the extraction solvent when extracting the crude okara. Nevertheless, the same extraction yields were obtained as with the dried material, which led to the conclusion that drying of the starting material, being an energy intensive operation, should be omitted.

The consequences of using the wet, non-dried okara were evaluated using an exergy analysis. The use of the water in okara poses a challenge if a high fraction of a co-solvent such as ethanol is needed, since that increases the total liquid-to solid ratio. Therefore, we conceptually designed the extraction process and evaluated the efficiency and sustainability of different options (chapter 3). Exergy analysis can quantify and combine effects of solvent consumption and physical energy thermodynamically. In addition, it can indicate the resource efficiency of a process and can be used to compare streams with different solvents such as ethanol and water. Often used for process optimisation, we investigated the use of exergy for process synthesis, which delivered the information for further process design and decision-making. A drying step was found to be less detrimental than an increased solvent use. The use of ethanol, its loss and distillation, and the loss of the extracted residue were identified as the most inefficient steps within the conceptual process, and with this, guidelines were given how to improve the systems sustainability.

The analysis performed in chapter 3 showed that a significant step towards better sustainability is made when ethanol is fully omitted as solvent in the extraction step. Furthermore, it was shown in chapter 2 that part of the isoflavones are rather water soluble due to their glucosidic nature. Therefore, the extraction and solubilisation of isoflavones in water was investigated in detail in chapter 4. Besides, the co-extraction of proteins in the water environment and their effect on the isoflavone extraction was investigated. The temperature did not influence the extraction, but an increased liquid-to-solid ratio and the pH did have a clear influence on the extraction yield. Okara may also contain the less polar aglycones, which are in general not present naturally in the soybeans, but which are products of hydrolysis of the glycosides due to certain processing conditions. The ionisation (dissociation) of the components at higher pH can modify their solubility and interaction with the matrix; therefore, by adjusting the pH the aglycones could be solubilised in water as well.

Further purification based on affinity separation was shown to be more feasible with water as solvent, since a water-ethanol mixture would be such a good solvent that the adsorption on the column would be hindered. The common way of regenerating such components includes the use of another solvent than is used for extraction, leading to an additional evaporation step. The higher affinity of isoflavones in an aqueous environment to a resin, and the possibility to omit an additional preparation step demonstrated the suitability of water as solvent in the primary extraction step: the extracts are more suitable for chromatographic purification, while ethanol or aqueous ethanol could be used as eluent. This demonstrates that it is important not to design a processing step by itself, but to optimise an internally coherent processing system as a whole.

Chapter 5 reports on the adsorption of the isoflavones on polyvinyl polypyrrolidone (PVPP) with special regards to the effect of proteins that are co-suspended in the extract. The adsorption efficiency was only negatively influenced when accounting for the isoflavones that are lost with protein that precipitates at lower pH. The affinity of the isoflavone groups could be described with constant partition coefficients resulting from a simple model. Furthermore, the adsorption of the isoflavones onto PVPP with the entire okara matrix present at floating pH could be described with a model using the partition coefficients describing the affinity to the PVPP and to the matrix. The model suggests that it is possible to ‘pull’ the isoflavones from the okara by concurrent adsorption of the isoflavones to an adsorption resin that is present in the same solution, thus increasing both yield and purity.

The three main isoflavone groups present in okara behave differently throughout the entire study, due to their different properties. The order of their adsorption affinities was the reverse of their water extractability. A more sustainable integrated recovery process can be designed for polyphenolic components, by identifying the right balance between the polarity and hydrophobicity of the component of interest, the solvent, and the adsorbent.

The investigated process synthesis methodology, and the design options that followed by applying this methodology are discussed in detail in chapter 6. A simplified simultaneous extraction-adsorption separation process, which resulted from the process synthesis is presented. Two base case flow diagrams that can be investigated in depths are further discussed in chapter 7.

Since the relevant properties and phenomena underlying the results presented in this thesis are quite generic, they can be translated to other separation processes, and the same approach can be used to develop more sustainable separation processes for polyphenols or other useful compounds from other food-by-products. By using the sustainability and efficiency as a prime objective within process design in combination with a good understanding of the underlying phenomena of the system, a new step in process design can be made: Sustainable process development requires proper understanding of the complex systems we are dealing with, the matrix itself, but also the thermodynamic behaviour of the components, and their behaviour in the matrix and during processing.