The significance of the studies summarized above can be discussed in the context of increasing bioreactor productivity in synthesis-gas fermentations. Metabolic modeling can be used to calculate the maximum theoretical yields possible from synthesis gas or combined with experimental data to calculate the fluxes of carbon, electron, and ATP through the branched metabolic pathways. Such information is useful in elucidating the patterns of regulation in response to environmental variables (e. g., pH). However, it is unlikely that manipulation of environmental variables alone will be sufficient to achieve the desired yields. Consequently, additional metabolic-engineering approaches will likely be necessary. Examples include isolation of mutants with altered genetic regulation patterns, elimination of enzyme(s) in unwanted pathways, and overexpression of rate-limiting enzymes in desired pathways. Implementation of these strategies will require development of recombinant tools (e. g., plasmids) for the microbes of interest.

Significant increases in productivity can be achieved via reactor engineering, as well, particularly through increasing cell concentration and increasing the rate of synthesis-gas mass transfer. The cell-immobilization studies indicated that B. methylotrophicum readily attaches to a variety of support materials that would be well-suited for industrial fermentations, and that these support materials do not inhibit cell growth or product formation. The cell recycle approach was highly successful. Both cell and product concentrations were increased several-fold over values obtained without cell recycle (17,19). Moreover, even during runs in excess of 1000 h, virtually no membrane fouling was observed.

The experimental and modeling results to date suggest that microbubble dispersions are well-suited for enhancing synthesis-gas fermentations. Extremely high Кьа values have been measured for microbubbles without mechanical agitation. When the surfactant concentration in the bulk liquid was low, these coefficients approached the theoretical values, suggesting that the mass-transfer resistance of the surfactant shell can be maintained at low levels. The power — consumption rate to produce the microbubbles is projected to be quite low. Because the microbubbles would be produced in a relatively small vessel and pumped to the bioreactor, only the minimal amount of power input required to maintain sufficient mixing would be required in the bioreactor. Consequently, energy-efficient, pneumatically mixed configurations, such as the airlift could be used. Several surfactants have been identified that do not interfere with the growth and product formation yet form high-quality microbubbles. The dynamic microbubble model has been used to evaluate the influence of bubble shrinkage, surfactant-shell resistance, and changes in gas pressure and composition on the mass-transfer efficiency of microbubbles and to help interpret experimental results. Experiments are currently underway in our laboratory to evaluate the suitability of microbubble mass transfer in long-term synthesis-gas fermentations.

Although this paper has focused primarily on issues related to bioreactor productivity, there are also important separations issues related to synthesis-gas fermentations. First, the products are currently produced in relatively low concentrations, so cost-effective methods to separate them from dilute fermentation broths are needed. Second, the acids and alcohols produced in synthesis-gas fermentations become inhibitory as they accumulate. The effects of these products on the growth and stationary-phase product formation in B. methylotrophicum have been measured (77). Cell growth was found to be inhibited at alchohol concentrations on the order of 5 g/L, even though the stationary-phase CO metabolism was unaffected by such levels. Simultaneous fermentation and separation approaches would be expected to be useful in this situation, such as pervaporation membranes, which are selective for alcohols in their transport properties. Flux through the membrane is facilitated by the use of a vacuum, whereby components that diffuse through the membrane are immediately removed by evaporation. The resulting vapor, which is enriched in the alcohols, would then be condensed prior to further purification steps (e. g., distillation). Third, is will likely be necessary to recover and reuse the surfactant when microbubbles are used. However, it may be possible to rely on the surfactants that are naturally produced in the fermentation to form the microbubbles. This approach has been demonstrated in bench-scale yeast fermentations (33).

This paper has identified several engineering issues that currently limit the commercial prospects of synthesis-gas fermentations and has summarized recent research that addresses these issues. Such research, combined with complementary biocatalyst-development efforts, may make bioconversion of biomass-derived synthesis gas into fuels and chemicals a commercial reality.