(239c) Lignocellulosic Biomass to Ethanol: Process Design and Economics for Corn Stover

Introduction

Bioethanol production worldwide is continuing to grow at a rapid pace.  In the United States, ethanol production has more than tripled over the past 5 years, from 2.8 billion gallons per year in 2003 to over 9 billion gallons per year in 2008 (RFA, 2008)¹. The Energy Independence and Security Act of 2007 (EISA 2007)² mandates 36 billion gallons per year of renewable fuel by 2022, of which 15 billion gallons are traditional or corn-based, and 16 billion are derived from cellulosic biomass. In order for cellulosic biofuel production to ramp in 2012 as dictated by the EISA, it becomes increasingly important to understand its production economics. This information is used to predict the cost-competitiveness of cellulosic biofuels with petroleum-derived fuels and corn ethanol.  A number of different conversion technologies exist for the conversion of cellulosic biomass to a liquid fuel. This paper is focused on biochemical conversion to ethanol, in which biocatalysts such as enzymes and microorganisms are used along with heat and chemicals to first convert biomass to an intermediate mixed sugar stream and then to ethanol or other fermentative biofuel. National Renewable Energy Laboratory (NREL) researchers have described such processes in technical and economic detail in two previous design reports (Aden et al. 2002³; Wooley et al. 1999⁴). This paper updates the ongoing process design with the most current data and integration efforts at NREL as well as research funded by the US Department of Energy (DOE), and other sources (e.g., industry, academia). The result is a so-called techno-economic model that reasonably estimates a production cost for this pre-commercial process.

Methods

Developing process economics of this sort requires experimental data, modeling toolsets, vendor equipment information, and engineering expertise. The first step is to develop a conceptual process design from information on the products, feeds, and proposed chemical and physical processing steps (Douglas 1989)⁵. The effective use of conceptual design methods in the early stages of process design can have a large impact on overall process design and development. The conceptual designs for the cellulosic ethanol process described here were largely developed by NREL in collaboration with a number of research partners.

Material and energy balance and flowrate information for a given process design are then generated using process simulation software packages. For these particular applications, Aspen Plus (Aspen Plus 2006)⁶ was used. This software contains physical property and thermodynamic data for a large number of chemical compounds. NREL has further developed customized physical property data for biomass constituents such as cellulose, lignin, and xylan (Wooley, Putsche 1996)⁷.  The material and energy balance data generated by these models are used to size and cost process equipment, which are fed into spreadsheets built for capital and operating cost estimation. Using a published engineering methodology (Peters and Timmerhaus 1991)⁸, a discounted cash flow rate of return analysis is performed using these capital and operating costs to determine the minimum ethanol selling price (MESP, $/gallon) required to obtain a net present value (NPV) of zero for a 10% internal rate of return (IRR).  The MESP is therefore slightly different from a true cost of production.

Results and Discussion

Process Design Overview. This design uses dilute-acid pretreatment followed by enzymatic hydrolysis and pentose/hexose co-fermentation with recombinant Zymomonas mobilis. The biomass (corn stover) is first treated with dilute sulfuric acid catalyst at a high temperature (190 ˚C) for a short time (average at 2 minutes), liberating the hemicellulose sugars and other compounds. Before enzymatic hydrolysis, alkaline conditioning is required to detoxify the hydrolyzate for the fermenting organisms.

Historically, the pretreated biomass was separated into solid and liquid fractions and lime was used to condition the liquor only, which was then remixed with the separately-washed solids from pretreatment before enzymatic hydrolysis. This required an additional separation to remove gypsum from the liquor, during which some of the sugar was invariably lost with the solid. The current design uses ammonium hydroxide solution to condition the hydrolyzate instead, avoiding this sugar loss. The high miscibility of liquid ammonium hydroxide additionally permits treatment of the whole hydrolyzate slurry and eliminates the solid-liquid separation. While ammonia is considerably more expensive than lime, the economic benefits of reduced sugar loss and reduced capital cost make ammonia the more economic alternative by a narrow margin. Enzyme is added to the hydrolyzate at an optimized temperature for enzyme activity. In cases with higher temperature for saccharification than for fermentation, a cooling step is required to ensure growth of the fermenting organism Zymomonas mobilis at anaerobic condition. Three to seven days are required to convert most of the cellulose and xylose to ethanol. The ?beer? with ~4-8 wt% of ethanol is then sent to the recovery unit to purify ethanol to fuel grade. The solids remaining after fermentation are combusted in a fluidized bed combustor to produce high pressure steam for electricity credits and process heat.

Process Flow Diagram. The process flow diagram (PFD) includes nine areas: feedstock handling, pretreatment and detoxification, saccharification and fermentation, onsite enzyme production, distillation and ethanol purification, waste water treatment, boiler and turbo generator, and utility. The schematic PFD from the 2002 design report is shown in Fig. 1.

Figure 1.  Biochemical Conversion Process. (Aden et al., 2002)

Cost Estimation and Process Economics.  The PFD and rigorous material and energy balances generated by the Aspen Plus simulator are used to define detailed equipment sizes and costs using an in-house spreadsheet model for capital cost. Proper choice of equipment design and configuration is critical, so NREL processing technology research was combined with outside engineering firm consulting. Vendor proposals were obtained for the larger unit operations. Costs for secondary equipment were drawn from Harris Group's proprietary cost database. Total project investment, variable, fixed costs were developed first, then a discounted cash flow analysis was applied to determine ethanol production cost when the net present value of the project is zero. The cost of ethanol production is used either to assess its potential in the marketplace with absolute numbers from targeted research and highlights areas in which economic improvements are needed.

Sensitivity Analysis. Alternatives for each process area and many unit operations are discussed in detail in this paper. For instance, different feedstocks (or compositional variation of the same feedstock) can have a significant impact on equipment design, raw material utilization, utilities, and overall process economics. A comparison between onsite enzyme production and offsite or purchased enzyme is made. In addition, the economic impact of uncertainty surrounding equipment design and installation and construction costs will impact the economics is discussed. Of the remaining process targets, some can be met through core research (such as yield), while others are not as controllable (such as feedstock composition, availability, and cost). Therefore, the sensitivity analysis provides an understanding of the impact of likely process variables and how these might be controlled to a definable degree.

Conclusions

A conceptual process design for biochemical ethanol from corn stover has been developed based on current research at NREL and other sources. The process design reflects the best available conversion technology for a fermentation-based process. Process economics of ethanol production follow from this design through the use of rigorous material and energy balance calculations with Aspen Plus, capital and project cost estimation using vendor proposals and engineering cost databases, and discounted cash flow analyses. The result is a final calculation of minimum ethanol selling price that can be used to predict the future market competiveness of cellulosic ethanol. Sensitivity analysis of process variables and process alternatives benefit future engineering designs. 

Acknowledgement.  NREL would like to thank the US Department of Energy Office of the Biomass Program for its continued leadership, support and collaboration in the biofuels arena.

References

(1)     RFA. Changing the Climate, Ethanol Industry Outlook. 2008.

(2)     EISA. EISA of 2007 Calls for Additional Production of Biofuels. http://www.renewableenergyworld.com/rea/partner/stoel-rives-6442/news/article/2008/01/eisa-of-2007-calls-for-additional-production-of-biofuels-51063. 2007.

(3)     Ibsen, K., Jechura J, Neeves, K., Sheehan, J., Wallace, R. Lignocellulosic Biomass to Ethanol Process Design and Economics Utilizing Co-Current Dilute Acid Prehydrolysis and Enzymatic Hydrolysis For Corn Stover. 2002. NREL Report NREL/TP-510-32438, http://www.nrel.gov/docs/fy02osti/32438.pdf

(4)     Wooley, R., Ruth, M., Glassner, D., Sheehan, J. Process Design and Costing of Bioethanol Technology: a Tool for Determining the Status and Direction of Research and Development. Biotechnology Progress, 1999, 15(5), 794-803.

(5)     Douglas, J. M., Conceptual Design of Chemical Processes. McGraw-Hill. 1989.

(6)     Aspen Plus^TM. Release 2006.5, Aspen Technology Inc., Cambridge MA. 2006.

(7)     Wooley, R., Putsche, V. Development of an ASPEN PLUS Physical Property Database for Biofuels Components.  National Renewable Energy Laboratory, Golden CO.  1996. NREL Report No. MP-425-20685.

(8)     Peters, M. S., Timmerhaus, K. D., Plant Design and Economics for Chemical Engineers, 4th edition, McGraw-Hill, New York, 1991.