Extraction

Overview

After passing through the external cell separator, the liquid stream from the bioreactor will contain dissolved salts, PHB granules, unused volatile fatty acids and other solutes. Commercially, PHB extraction is most commonly carried out using solvent extraction. However, there are many disadvantages involved with solvent extraction including using up large volumes of toxic and volatile solvents such as chloroform (Kunasundari & Sudesh, 2011). Solvent extraction, therefore, would not be feasible on Mars due to the large volumes of solvents which would have to be shipped up there and the toxicity of those solvents.

We partly tackled the challenge of PHB extraction by engineering the PHB-producing E.coli to secrete the PHB granules. However, the next challenge was getting the secreted PHB granules out of the media. Based on literature review we anticipated PHB granules to be 20-60 nm in size (Rahman, Linton, Hatch, Sims & Miller, 2013). Given the size of the particles, we expected that the first step in the extraction process would be agglomeration/flocculation to get larger particles which could then be separated using centrifugation or sedimentation. Following that, we explored a number of methods including dissolved air flotation, chemical coagulation and electrocoagulation, to agglomerate the PHB nanoparticles and subsequently separate them from the media.

Design Options

Flocculation and Centrifugation

One of the first design options we tested in the lab was chemical coagulation using calcium chloride. It was shown in literature that PHB particles have a zeta potential of pH 3.5, thereby in our supernatant where pH was measured to be around pH 5, PHB would be negatively charged (van Hee, Elumbaring, van der Lans & Van der Wielen, 2006). By supplying positively charged ions it would be possible to neutralize the charge on the PHB nanoparticles and have them agglomerate to form larger particles. Previously, experiments with PHB released from E.coli cells found that dications, especially calcium ions work well to agglomerate PHB (Resch et al., 1998); hence, we used calcium chloride as the coagulant in our lab experiments. We demonstrated that addition of calcium chloride followed by centrifugation increases the amount of PHB recovered as opposed to centrifugation by itself. More about these experiments can be found here.

The large-scale process that we envisioned would have had the liquid stream from the PHB fermentation unit pass into a mechanical flocculator where it would be mixed with calcium chloride before being passed into a centrifugation unit. The PHB would be settled out and the remaining liquid could be used to recover water. However, we encountered several huge setbacks when planning to use this process on Mars. Firstly, there was the issue of obtaining calcium chloride on Mars. It would either have to be shipped up there or produced in-situ. Currently, there has been some investigation into producing metal oxides and oxygen from Martian soil (Badescu, 2014); this process could potentially be modified to produce calcium chloride, however that investigation was beyond the scope of this project.

Unfortunately, a crucial problem with the flocculator and centrifugation design was the potential of having calcium chloride remaining bound to the PHB particles. We spoke to Dr. Nashaat Nassar, who is a professor in the Chemical Engineering department at University of Calgary, and he confirmed that it was most likely that the calcium chloride would remain bound to the PHB after centrifugation. He suggested we wash the recovered PHB with acidified water to remove the calcium chloride. However, this introduced a number of challenges including risk of resuspending the PHB once again and breaking up the PHB agglomerates, essentially undoing all the work we did to separate the PHB.

Hence, for these reasons we decided not to pursue chemical coagulation as our method of separation.

Electrocoagulation

Next, we looked at electrocoagulation (EC). In principle it works similarly to chemical coagulation, but in this case the ions causing the coagulation are supplied by sacrificial anodes which are usually made from iron or aluminium (Tian et al., 2017). EC has a number of advantages such as use of simple equipment, easy operation, a shortened reactive retention time and no chemical additions (Ozyonar & Karagozoglu, 2017).

Our lab experiments with electrocoagulation which are described here, showed that while it was possible to settle out PHB using electrocoagulation, we also obtained a lot sludge when attempting it with a synthetic feces sample. It proved very difficult to extract PHB from this sludge. Since electrocoagulation was not selective for PHB in our synthetic feces samples we decided not to pursue electrocoagulation for our process on Mars.

Dissolved Air Flotation

Literature has demonstrated the use of dissolved air flotation (DAF) to separate PHAs with an overall yield of 86 ± 0.3% in lab-scale experiments (van Hee, Elumbaring, van der Lans & Van der Wielen, 2006). In DAF, air bubbles are created by injection of air-oversaturated water. With our proposed process, compressed Martian air will be passed into a dissolved air vessel where it will be dissolved in water before being injected into a flotation column. The fluid stream from the bioreactor will also be fed into the flotation column. In it, the passage of air bubbles would cause PHB nanoparticles to float up to the top. Eventually the PHB-rich top phase will be removed and passed into a dryer similar to ones currently being used in the industrial production of PHB.

DAF has several advantages, the most significant of them being that it does not use up large quantities of toxic chemicals (Kunasundari & Sudesh, 2011). Every other design option we examined would require shipping of chemicals such as coagulants or solvents. DAF would also work well in conjunction with the secretion system which produces PHB nanoparticles that would otherwise need to be agglomerated before separation via either sedimentation or centrifugation. However, there are also several drawbacks with DAF. First of all, it would have to run in batch mode, perhaps even requiring several consecutive flotation steps (van Hee, Elumbaring, van der Lans & Van der Wielen, 2006). The process also has not been used in scaled-up applications yet and PHB yield is comparatively low when compared to solvent extraction. We would also need to recycle the water we recover in our process into the dissolved air vessel.

Recommendations for further experiments would be optimising batch duration and designing a pilot large-scale system. When designing a large-scale system, experiments with increasing pressure in the dissolved air vessel could be conducted; higher pressure would lead to higher dissolved air concentration and increase the volume fraction of air bubbles (van Hee, Elumbaring, van der Lans & Van der Wielen, 2006). This in turn could potentially improve the yield of PHB. The batch flotation system could also be turned into a continuous process with the PHB-rich top phase being removed continually. It is expected that in continuous flotation that top phase PHB concentration will be higher (van Hee, Elumbaring, van der Lans & Van der Wielen, 2006).