The separation of glycinin and β-conglycinin from soy is of increasing commercial interest. Methods for separation in the literature generally involve the aqueous extraction of protein from defatted soy flour followed by the precipitation of glycinin by acidification producing a glycinin-rich precipitate phase and a β -conglycinin-rich supernatant phase. The fundamental driving force for this separation is the differential solubility between glycinin and β-conglycinin. However, there is a lack of comprehensive solubility data for the proteins in the literature. The focus of this project was on studying the extraction of protein from soy flour and on acquiring a set of comprehensive solubility data under a wide variety of conditions so that a protein extraction and separation scheme can be rationally designed. Precipitate formation was studied in detail at the conditions that separate glycinin from β-conglycinin as this has not previously been reported.
In the aqueous extraction of protein from soy flour, only an estimated 60 - 70 % of protein was extractable regardless of the solvent to flour ratio used. Repeated extraction of the spent soy flour extracted little additional protein. The constant protein yield in the extraction meant that the lower the solvent to flour ratio, the higher the extract concentration. The composition of the extract was also similar for all of the extraction ratios studied with ~ 30% β-conglycinin, ~50% glycinin and ~ 20% other proteins.
The pH-concentration profiles of total soy protein and the components glycinin and β-conglycinin all followed a U-shaped trend with a minimum at pH 4 - 5. Surprisingly, increasing the initial protein concentration of the aqueous extraction increased the residual protein concentration. The increase in residual concentration was almost proportional to the increase in initial concentration in all cases. This unexpected behaviour could not be fully explained. Several likely contributing factors were explored and hypotheses were put forward for future consideration. Due to this unusual solubility behaviour, the data are, strictly speaking, not thermodynamic solubility data. Therefore the term 'residual concentration' has been used instead.
Between pH 5.8 - 7 is a region where glycinin precipitated while β-conglycinin remained in solution thus protein separation is achieved. The most suitable pH range is between 6 - 6.2 for a robust separation. Sodium chloride (NaCl) had a salting-in effect on soy proteins and modulated the residual concentration of glycinin and β- conglycinin such that the window for separation was eliminated. Addition of NaCl should therefore be avoided. Temperature had a statistically significant effect on the residual concentration of β-conglycinin only. When combined with the glycinin data, which had no temperature effects, the residual concentration of total protein also showed no temperature effects.
Remarkably, precipitation between pH 5.8 and 6.4 was in fact a binary liquid-liquid separation. The secondary liquid phase appeared as droplets of 1 - 10 µm in size. They coalesced upon centrifugation to form a uniform bottom layer. The secondary phase was a protein-rich phase containing 20 - 30 wt% protein, ten times more concentrated than the supernatant phase. It was 80 - 99% pure in glycinin. The concentrated liquid protein phase is stable (did not crystallise or precipitate when stored at 4°C for up to one month). This presents a unique opportunity for a new soy product development. The concentrated liquid phase protein could allow food manufacturers to incorporate soy protein into food products where it has previously been difficult with solid phase soy protein such as protein isolates.
Analysis of the droplet size distribution showed that the secondary phase formed by both nucleation of new droplets as well as growth on existing ones. A small amount of droplet coalescence was observed under a variety of mixing speed experiments but no droplet breakage could be observed.
The protein extraction and separation scheme recommended involves using a 10: 1 extraction ratio for protein extraction from soy flour, stirring for 5 - 10 minutes at room temperature (21 °C). The aqueous protein extract is then separated from the spent soy flour by centrifugation. A glycinin-enriched protein-rich liquid phase is produced when the extract is acidified to between pH 6 - 6.2, stirring for 10 minutes at room temperature. The protein-rich phase separated from the supernatant by centrifugation. A separation process flowsheet is provided.