Chemical Reactor Engineering

Ethanol production through continuous fermentation (2006-present)

• Background and motivation.
• Reactor Productivity.
• Product Yield.
• References.
• Published papers.

Background and motivation

The interest in biofuels has renewed since the Kyoto Protocol, where many industrialized countries agreed to reduce their carbon dioxide emissions and greenhouse gas production. Renewable fuels have the potential to greatly reduce Australia's reliance on expensive imported oil. One of the proposed fuel alternatives is ethanol, which can be produced from many renewable sources such as lignocellulosic waste/materials (Ward and Singh, 2002). The largest national ethanol fuel industries exist in Brazil, where almost 50% of all cars are able to use 100% ethanol fuels and gasoline sold contains at least 20% ethanol (Reel; 2006).

Ethanol has a number of attractive features as a fuel and ethanol blends are increasingly being used worldwide; more than 10 percent of all gasoline sold in the US in 2002 contained ethanol (United Sates Environmental Protection Agency, 2002). It is a much cleaner fuel than gasoline, being biodegradable without having harmful effects on the environment. It provides high octane at low cost, acting as an alternative to harmful fuel additives; ethanol blends can be used in gasoline engines without modifications. Ethanol's high oxygen content reduces carbon monoxide levels by 25-30\% according to the US EPA and dramatically reduces emissions of hydrocarbons, a major contributor to the depletion of the ozone layer. Many car manufacturing companies (GM, Ford, Chrysler, Toyota, Honda) are now developing hybrid vehicle that run on an ethanol mix.

In the US the Energy Independence and Security Act of 2007 (EISA) required the use of 16 billion gallons of cellulosic biofuels by 2022 (U.S. Department of Energy, 2008). According to a study by the U.S. Department of Agriculture bioethanol generates 35% more energy than it takes to produce (Shapouri et al 2002). In addition, it is a renewable fuel as it may be made from plants.

We have investigated the continuous production of ethanol. The biochemical model used was originally proposed by Ghommidh et al (1989), to account for oscillations observed during the continuous fermentation of ethanol using Zymononas mobilis, and extended by Jarzebski (1992), to explain oscillations observed during continuous fermentation using Zymononas mobilis. Jarzebski estimated parameter values using data obtained from the continuous fermentation of sugar-cane blackstrap molasses a temperature of 37^oC by Pergo et al (1985). The model contains five variables: the concentration of substrate, product (ethanol) and biomass (viable cells, non-viable cells and dead cells). The biomass is Zymomonas mobilis. The equation for the concentration of dead cells uncouples from the system, so that there are four equations in a single reactor. The biochemical kinetic model accounts for both product inhibition and substrate limitation. We assume that only the feed contains only substrate.

A combination of steady state analysis and path following methods is used. The performance of the reaction scheme in one tank is used as benchmark for comparing the performance of multiple tanks. We have investigated ethanol production in both a single tank (Watt et al, 2007a) and in reactor cascades (Watt et al, 2007b; Sidhu et al, 2008; Watt et al, 2010). In these investigations there was no recycle in the system.

Reactor Productivity

Our results for a single reactor include:

• We have established the conditions for washout to occur in a single reactor (Watt et al (2007a)).
• We have established in a single reactor that the steady-state diagram contains no hopf bifurcations if the substrate concentration in the feed is below 108 gl^-1, two hopf bifurcations if the substrate concentration in the feed (S₀) is in the range 108 gl^-1 < S₀ < 122 gl^-1 and one hopf bifurcation point if the substrate concentration in the feed satisfies 122 gl^-1 < S₀. (Watt et al (2007a))
• Our analysis shows that although the system can exhibit interesting behaviour, such as oscillatory states and a period doubling root to chaos, the productivity of the system is optimised by operating in a steady-state regime. (Watt et al (2007a)).
• The optimal productivity for a single tank was only very weakly dependent upon the substrate concentration in the feed and was 3.8 gl^-1h^-1. (Watt et al; 2007b)

Our results for a cascade of two reactors include:

• For sufficiently low total residence times the cascade system never out-performs the optimised single reactor. (Watt et al; 2007b)
• When the total residence time is greater than approximately eight hours, an optimally designed cascade outperforms the optimal single reactor. In fact, the best two-reactor configuration has a total residence time of 15 hours (residence time of 6.03 hours in the first reactor and 8.97 hours in the second) yielding an optimal productivity of 5.08 gl^-1h^-1. This is an increase in productivity of 34% compared to the optimal single tank. (S₀= 160 gl^-1). (Watt et al; 2007b)
• The optimal productivity of a two-reactor cascade with equal residence times in both reactors is 4.82 gl^-1h^-1. This is increase in productivity of 27% compared to the optimal single tank. (S₀= 160 gl^-1). (Watt et al; 2007b)

  The optimal productivity of a two-reactor cascade with different residences time in both reactors increased from 3.99 gl^-1h^-1 (S₀= 100 gl^-1) to 5.08 gl^-1h^-1 (S₀= 160 gl^-1) (Watt et al; 2010).

• The optimal productivity of a three-reactor cascade with different residences time in both reactors is 5.181 gl^-1h^-1 (Watt et al; 2010). (S₀= 160 gl^-1). This is an improvement of 36% over that of the single reactor system and 2% over the optimal two-reactor cascade. It is likely that the small gain that can be achieved from going from a cascade of two reactors to one of three us not warranted given the uncertainty in the biochemical parameters.

Our results for a cascade of three reactors include:

• The optimal productivity of a three-reactor cascade with equal residence times in both reactors is 4.87 gl^-1h^-1. This is an increase of 28% compared to the optimal single reactor. (S₀= 160 gl^-1). (Watt et al; 2007b)

Our results for a cascade of four or five reactors:

• With equal residence times in each reactor we found that increasing the number of tanks in the cascade to four or five decreased the maximum productivity: 4.285 gl^-1h^-1 (four reactors) and 3.776 gl^-1h^-1 (five reactors). (Watt et al; 2010).

The principal contribution of our investigations into ethanol productivity is not in determining the particular increases that can be gained from a cascade using this particular biochemical mechanism. Instead, the main contribution is to show that path following are efficient tools to investigate the performance of a single reactor, any two-reactor cascade and any reactor cascade with equal residence times in each reactor.

Product Yield

• The maximum yield obtained in a single stage continuous Zymomonas fermenter is significantly lower than the yields currently obtained in the US corn to ethanol industry, which are currently around 0.48 (Wallace et al; 2005). (Sidhu et al; 2008).
• In a two stage continuous fermenter it was possible to obtained yields higher than 0.48 by having a total residence time of 15.23 hours and having different residence times in each stage. To obtain the same yield in a two-reactor cascade having equal residence times in each stage a total residence time of 16.22 hours was required. (Sidhu et al; 2008).
• In a three stage continuous fermenter, with equal residence times in each stage, it was possible to obtained a yield of 0.48 by having a total residence time of 16.29 hours. (Sidhu et al; 2008).

References

1. A.B. Jarzebski. (1992). Modelling of oscillatory behaviour in continuous ethanol fermentation. Biotechnology Letters, 14(2), 137-142. http://dx.doi.org/10.1007/BF01026241.
2. C. Ghommidh, J. Vaija, S. Bolarinwa and J.M. Navarro. Oscillatory behaviour of Zymomonas in continuous cultures: A simple stochastic model. Biotechnology Letters, 2(9), 659-664.
3. L. Pergo, J.M.C.D. Dias, L.H. Koshimizu, M.R.D. Cruz, W. Borzani and M.L.R. Vairo. (1985). Influence of temperature, dilution rate and sugar concentration on the establishment of steady-state in continuous ethanol fermentation of molasses. Biomass 6(3), 247-256.
4. M. Reel (2006). Brazil's road to energy independence. The Washington Post, August 19th.
5. Shapouri, H., Duffield, J.A., and Wang, M. The energy balance of corn ethanol: An update. United States Department of Agriculture. Agricultural Economic Report Number 813, 16 pages, 2002.
6. R. Wallace, K. Ibsen, A. McAloon and W. Yee. (2005). Feasibility Study for Co-Locating and Integrating Ethanol Production Plants from Corn Starch and Lignocellulosic Feedstocks. US Department of Agriculture and US Department of Energy.
7. Ward, O.P., and Singh, A. Bioethanol technology: developments and perspectives. Advances in Applied Microbiology 51, 53-80.
8. U.S. Department of Energy. Biofuels & Greenhouse Gas Emissions: Myths versus facts. 2008. www.eere.energy.gov/biomass/biomass_basics_faqs.html. Accessed 16th October 2008.
9. United Sates Environmental Protection Agency. Clean Alternative Fuels: Ethanol. 2002. http://www.eere.energy.gov/afdc/pdfs/epa_ethanol.pdf. Accessed 16th October 2008.

Ethanol production: Published papers

Referred journal papers

1. S.D. Watt, H.S. Sidhu, M.I. Nelson and A.K. Ray. (2007a) Analysis of a model for ethanol production through continuous fermentation. ANZIAM Journal E (EMAC2007), 49, C85-C99, 2007. http://anziamj.austms.org.au/ojs/index.php/ANZIAMJ/article/view/322 .
2. S. Watt, H.S. Sidhu, M.I. Nelson, and A.K. Ray. Analysis of a model for ethanol production through continuous fermentation: Ethanol productivity. International Journal of Chemical Reactor Engineering, 8: Article A52, 2010. http://www.bepress.com/ijcre/vol8/A52.

Referred conference papers

3. S.D. Watt, H.S. Sidhu, M.I. Nelson and A.K. Ray. (2007b) Improving ethanol production through continuous fermentation. In Proceedings of the 35th Australasian Chemical Engineering Conference, CHEMECA 2007, pages 1862-1869 (on CDROM), Engineers Australia, 2007. ISBN 0-858-25844-7.
4. H.S. Sidhu, J. Kavanagh, S.D. Watt and M.I. Nelson. Performance Evaluation of Ethanol Production Through Continuous Fermentation. In Proceedings of the 36th Australasian Chemical Engineering Conference, CHEMECA 2008, pages 590-599 (on CDROM), Engineers Australia, 2008. ISBN 85825-823-4.