Gas-solid fluidized beds consist of a vertical chamber with a perforated base plate through which air, or some other gas, is blown with sufficient velocity to fluidize the substrate particles (Fig. 9.5(a)). It is necessary to design the bioreactor with sufficient height to allow for expansion of the bed upon fluidization. Also, in order to facilitate separation of the solids, the upper regions of the bioreactor need to be somewhat wider than the fluidization region. Due to the greater cross-sectional area for flow, the superficial velocity of the air falls below the minimum fluidiza-tion velocity and the particles in this region therefore settle. It may be necessary to incorporate a mechanical mixer slightly above the base plate to help to break up any unfluidized agglomerates that may deposit there. In this type of bioreactor it is a relatively simple matter to make additions to the substrate bed. Water, or nutrient or pH correcting solutions can be sprayed onto the top of the substrate bed.
It may be interesting to recycle the process air, in order to reduce the air preparation costs involved with heating and humidification, although in an aerobic process care must be taken not to allow the O2 level to fall too low and the CO2 level to rise too high. Such bioreactors can be used for anaerobic processes if N2 is used for fluidization, but recycling is then essential in order to minimize process costs.
The ability to use this type of bioreactor depends on the substrate properties. There are two potential difficulties. Firstly, large agglomerates will form if sticky particles are used and these agglomerates will not fluidize. Secondly, if the substrate particles have different sizes, then some particles might fluidize while others
o w disengagement space o
Fig. 9.5. Bioreactors in which the mixing action is provided by the gas stream. (a) Gassolid fluidized bioreactors, in which the whole bed is fluidized. (b) A spouted bed in which only particles that fall into the central region are fluidized (Silva and Yang 1998)
might not. Even with a non-sticky substrate of uniform size, fluidized-bed operation would also be expected to face challenges given that the properties of the substrate particle can change markedly during a fermentation due to the consumption of nutrients within the particle by the microorganism and consequent loss of particle mass in the form of CO2.
Matsuno et al. (1993) describe the use of two air-solid fluidized beds by the soy sauce company Kikkoman in the 1970s:
• a 16-L bioreactor 2 m high, with a diameter of 20 cm in the lower fluidization region and a diameter of 28 cm in the upper disengagement region;
• an 8000-L bioreactor 8 m high, with a diameter of 1.5 m in the lower fluidization region and a diameter of 2 m in the upper disengagement region. This bioreactor had a capacity for 833 kg of wheat bran at 40% moisture content.
According to Matsuno et al. (1993), Kikkoman claimed that the air-solid fluidized bed gave a higher productivity for the production of proteases and amylases by Aspergillus sojae on wheat bran powder than either static-bed SSF systems or submerged liquid culture. However, detailed information about this bioreactor and its operation is not available in the literature.
Gas-solid fluidized beds also received interest in the 1980s for the production of ethanol. A pilot-scale bioreactor of 55 cm diameter was built by Rottenbacher et al. (1987). In this case the system had some differences from "typical" SSF processes. Given that ethanol production requires anaerobic conditions, N2 was used as the fluidizing gas. It was recycled through the bioreactor, with ethanol being condensed from the gas before it was returned to the bioreactor. The idea was that, in continuously stripping ethanol from the system, this strategy would minimize the inhibitory effects of ethanol and maximize ethanol yields. Another difference was that the solid phase was not a nutrient phase but rather consisted of pellets of compressed yeast. The bioreactor had a capacity for 20 kg of yeast pellets. A glucose solution was sprayed onto the bed surface, therefore each pellet received fresh nutrients as it circulated through the bed.
None of the workers who have used fluidized beds have mentioned any problems with temperature control. This is not unexpected, since the high flow rates required for fluidization should provide sufficient convective cooling capacity. In fact, due to the ease of temperature control, mathematical models that have been developed for fluidized bed operation (Rottenbacher et al. 1987; Bahr and Menner 1995) do not include energy balances. Further, due to the good homogeneity of the bed, they tend to be concerned with intra-particle phenomena.
A variant of the fluidized bed is the "spouted bed" (Fig. 9.5(b)). The major difference is that air is blown upwards only along the central axis of the bed, such that only part of the bed is fluidized at any one time. There is a continuous cycling of particles as the solids slip down the sloped sides at the bottom of the bioreactor.
Silva and Yang (1998) built a spouted-bed bioreactor of 7.6 cm diameter and 73 cm height. However, in their experiments, in which they grew Aspergillus oryzae on rice, the bed height was only 9 cm, which means that ratio of overall bioreactor volume to bed volume was quite large. Further, they used an aeration rate of around 250 L min-1. This represents an aeration rate of 625 vvm (volumes of air per volume of bed per minute), which is unlikely to be practical to maintain at large scale. In this particular fermentation, continual spouting led to poorer growth and lower enzyme levels than in beds that were operated as packed beds (at an aeration rate of around 50 L min-1) for most of the time and only spouted intermittently at 1 or 4 h intervals (at an aeration rate of around 250 L min-1), presumably due to the shear damage caused by the continuous motion (Silva and Yang 1998). Note that in intermittently spouted beds, the aeration of the bed will not be uniform during the periods of packed-bed operation, since air is not introduced evenly across the whole section of the bed. It is not clear how suitable spouted operation will be for large-scale bioreactors.
Was this article helpful?