Richard B. Reid and Dr. Larry Baxter, Chemical Engineering
The search for sustainable energy sources and technologies is one of the world’s primary concerns in the 21st century. Biomass energy is projected to be an important component of a diverse and sustainable energy portfolio. In order to utilize this source effectively there must be accurate composition data and predictive correlations for material properties. Though the literature provides some empirical correlations of fuel properties such as heating value [Jenkins et al 1996], fundamental chemical and structural analyses have not been explicitly correlated with such properties. This project has sought to establish relationships between fuel properties and chemical functional group contributions in various biomass sources using a database of 5,960 samples previously gathered at Brigham Young University.
Biomass fuels come from a diverse array of plant matter but contain common functional groups and structural components. The chief structural components in these fuels are the organic polymers cellulose and lignin. On a sub-molecular scale, components are made up of chains of aliphatic and aromatic hydrocarbons with alcohol and ester side groups. One measure of a fuel’s aliphatic vs. aromatic content that has predictive potential for fuel properties is the hydrogen-tocarbon ratio plotted against the oxygen-to-carbon ratio. As can be seen in Figure 1, a plot of H:C vs. O:C for biomass materials from the BYU biomass composition database follows a fairly linear trend. Most of the fuels of significance for energy processing lie within the region bounded by lignin on the oxygen-poor side and cellulose on the oxygen-rich side.
The first task in this project was to evaluate the predictive correlativity between these two ratios and heating value, one material property of significance. For this evaluation, the method of Principal Component Analysis was used. The analysis showed that the H:C and O:C ratios accounted for over 80% of the variation in heating value for straws some of the woods. However, this promising result did not hold for each of the fuel types. It was hoped that these two variables coupled with the nitrogen content (representing the small amount of proteins present in the fuels) would provide a description of the aliphatic/aromatic/acidic character of each fuel and provide a basis for correlations of material properties. However, the analysis showed that such a basis would likely exhibit as much error in predictions as other empirical correlations already in the literature.
The next task was to obtain a chemical description of each fuel based on its lignin, cellulose, and protein content. These are difficult to obtain experimentally, so thermodynamic and material balances were used to predict these. The method required accurate data for thermodynamic properties of lignin, cellulose, and protein. This was problematic because of the difficulty in finding such data in the literature. Cellulose is a compound of known chemical structure, but lignin is not one compound but rather a class of similar polymeric compounds. The proteins found in the fuel arise from various fundamental amino acids in different amounts, creating another potential source of error. These uncertainties and the lack of complete and accurate data for the thermodynamic properties stalled the progress of the project. A complete literature review will be necessary in the future to ensure that the data being used for analysis are valid.
However, it was possible obtain a structural description of each fuel based on its lignin and cellulose content, using the ratio data from the plot in Figure 1. This predictive method was verified with 108 samples from the database that included data for lignin and cellulose. When the predicted and actual contents were compared, 33 agreed within 2 percent and 77 agreed within 10 percent. The method of prediction was based on average assumed values of H:C and O:C for lignin in the hardwoods, softwoods, and grasses. Use of more precise estimates of these values for lignin in various biomass groups is expected to further increase the predictive accuracy. For each fuel, the predicted lignin and cellulose content can then be used to correlate important fuel properties accurately.
This project was rigorous and challenging, but instructive. The major frustrations were in the lack of available data for fundamental properties and chemical structures of the biomass components. This, coupled with the need to learn PCA and its implementation within the R package, extended the project longer than anticipated. At present it is still a work in progress, but the future result is anticipated to be a peer-reviewed publication detailing the relationships between fundamental chemical components and fuel properties.
Sources
- Jenkins, B. M., R. Baker, J. Gilmer and L. L. Baxter (1996). Combustion Behavior of Leached Straw. Developments in Thermochemical Biomass Conversion, Banff, Canada.
- Jenkins, B. M., L. L. Baxter, T. R. M. Jr and T. R. Miles (1996). Combustion Properties of Biomass. Engineering Foundation Conference on Industrial and Utility Use of Biomass, Snowbird, UT.