Modeling of Solid-State Fermentation for Ethanol Production From Sweet Sorghum Stalks

 

Introduction

Biofuels production worldwide continues to grow at a very rapid pace.  In 2007, China's fuel ethanol output reached 1.6 million tons, of which 80% used corn as feedstocks, with more than four million tons of corn consumed [1]. However, corn-based ethanol manufactures have ceased for further capacity expansion due to the new government policies [1]. Therefore, cellulosic ethanol has become an inevitable trend for further biofuel development. Sweet sorghum has the potential to be used as a renewable energy crop and has become a viable candidate for ethanol production, so it was chosen as the cellulosic feedstock for this study. Biochemical conversion of sweet sorghum stalks to fuel ethanol can be achieved by solid-state fermentation (SSF) or liquid-state fermentation (LSF). Only recently, with emerging attention to biofuel demand, has SSF technology been applied extensively in the biofuels field for simplified process design and potential reduction of fuel production cost.  During the SSF process, the absence of free water leads to poor heat removal characteristics, and it is not easy to mix a bed of solid substrate particles well.  Because of this, heat removal is a major challenge in the design and operation of large-scale SSF bioreactors.  Therefore, to overcome these disadvantages effectively and to solve these engineering barriers, the design of the fermenter is among the most critical aspects for feasibility and eventual scale-up of SSF for commercial biofuel production. Among these several types of SSF reactors, rotating drum bioreactors (RDB) provide relatively gentle and uniform mixing by improving baffle design, since there is no agitator within the substrate bed. Currently, the main reason for the limited industrial application of SSF is the lack of engineering data and knowledge about the design and scale-up of solid-state fermenters. Fortunately, the significant improvement in understanding of how to design, operate and scale up SSF bioreactors has been possible through application of mathematical modeling techniques to describe the biological and transport phenomena within the system [2]. Mathematical models are then viewed as important tools for guiding the simulation, design, and operation of large-scale bioreactors for optimum performance.

In this paper, we have developed a mathematical model for discounted RDB operation, including kinetics and heat and mass transport. The model not only describes several key variables changing with time, i.e., biomass concentration, substrate concentration, and main product (ethanol) concentration, but also gives the radial temperature profile in the substrate bed. Finally, the model is validated with pilot unit testing for a 5 m³ RDB using milled sweet sorghum stalk for the production of bioethanol. Through our mathematical modeling approach, significant advances in understanding the biological and mass and heat transfer phenomena can be made toward development of quantitative scale-up strategies for SSF bioreactors.

Methods

The yeast Saccharomyces cerevisiae TSH-SC-1 was provided by our laboratory (New Resources Graduate School, Institute of Nuclear and New Energy Technology, Tsinghua University) and was used in the bench scale and pilot scale fermentation experiments. The sweet sorghum stalks were collected from a local rural area of Wuyuan county of Inner Mongolia, China. Leaves were first stripped from the fresh sweet sorghum stalks by hand, and then stalks were chopped to small particles before being loaded in the substrate bed of the rotating drum bioreactor. Ethanol concentrations were determined by a Shimadzu GC-14C gas chromatography system equipped with a flame ionization detector. 

Model Development

The RDB model includes growth kinetics of biomass, sugar consumption rate, ethanol production rate, and mass and energy balances. Compared with aerobic solid-state fermentation, oxygen is not needed for growth of our microorganisms in the SSF, so there is no sterilized fresh air charged into the RDB. But for both aerobic and anaerobic solid-state fermentation, CO2 is produced by microorganisms as a byproduct of metabolism and will be discharged continuously from the RDB. For anaerobic SSF, the headspace gas is mainly composed of CO₂, and the remaining amount of CO₂ could be assumed to be a constant since RDB usually is operated under constant pressure. The solid substrate bed was segregated into multiple layers from the RDB wall to the center of the RDB. Each layer was assumed to be well-mixed and therefore assumed to have a single temperature. Each layer has a material and energy balance equation, in addition to equations for heat transfer between the solid layers and interface with the headspace gas. A fourth-order Runge Kutta method was used and the model was programmed in C++ consisting of subroutine blocks for the simulation.

Pilot Plant SSF Validation

Batch fermentation experiments were carried out in a 5 m³ stainless steel rotating drum using yeast TSH-SC-1 to produce fuel ethanol from sweet sorghum stalks. The experiments were conducted for 48 hours hold time without temperature control. The drum rotated for 20 minutes every 5 hours. After each time interval and at the end of fermentation, samples were taken for cell, sugar, and ethanol analyses.

The cylindrical RDB is shown in the following figure.

Results and Discussion

The models were experimentally validated using the above pilot plant RDB experimental procedure. These experimental and model data indicate that first, cellulosic materials such as sweet sorghum stalks can be fermented well and converted to ethanol using yeast Saccharomyces cerevisiae TSH-SC-1 under a relatively short fermentation time (within 20 hours); second, cell growth, sugar reduction, and ethanol production are synchronized well in both the experimental results and the modeling prediction, implying good experimental work and model development in this case; and third, further improvement of the overall process can be focused on different cellulosic feedstocks, higher yield in biomass and ethanol production, etc., using similar approaches. Model validation with experiments data also is applied to the radial temperature profile.

Conclusions

Although the SSF bioreactor is a promising option for biofuel production, scaling it up is a major challenge because the absence of free water leads to poor heat removal characteristics, and it is not easy to mix a bed of solid substrate particles well. The main obstacles lie in heterogeneously distributed physiological, physical, and chemical environments in the substrate bed. This causes difficulties in fermentation control and heat buildup occurs due to weak heat transfer in an SSF bioreactor, especially in an RDB. Therefore, a mathematical model is an important tool with which to study the true nature of this complex system for further application in bioethanol production. In this paper, we developed a mathematical model to study the microorganism growth, sugar consumption, and ethanol production, along with heat generation, heat removal, and heat accumulation, in rotating drum bioreactors during the fermentation process. Validation experiments were conducted in a 5 m³ pilot-plant fermenter for production of fuel ethanol from chopped sweet sorghum stalks. The mathematical model agrees with the experimental data very well, especially in the microorganism growth period. Microorganism growth, sugar content reduction, and ethanol production were predicted and measured in the mathematical model and experimental results. The maximum rising temperature of the substrate bed was also predicted and matched very well in both experimental and modeling results, and the temperature range studied is acceptable for yeast to propagate and grow. Since the model-predicted temperature variation with fermentation time fits well with small scale and pilot scale experimental results, a similar model approach can be further utilized for process design, development, and optimization purposes in larger scale ethanol production using sweet sorghum stalk or other cellulosic materials.

 

Acknowledgement.  This work is carried out under the project 2006BAD07A01of the National Science & Technology Pillar Program in the Eleventh Five-year Plan Period, and the project 06YFGZSH02700 of the Science and Technology Development Program of Tianjin, China. We also thank China Oil and Food Corporation Limited (COFCO) for financial and other support of the pilot-plant RDB experiments

References

[1]. China Bioenergy Industry Report 2007-2008,

http://en.ec.com.cn/article/enindustry/enenergy/eneenews/200804/603886_1.html

[2]. C. Krishna, Solid-state fermentation systems, Critical Reviews in Biotechnology. 25 (2005) 1¨C30.