Fast pyrolysis followed by bio-oil upgrading into a hydrocarbon blendstock is considered to be a near-term opportunity for the production of biofuels that could mitigate the impact of petrochemical transportation fuels, a sector that accounts for 70% of all petroleum consumption in the United States. Understanding conversion performance as a function of feedstock properties is a critical element of both the design and successful operation of an integrated biorefinery. In addition, this understanding would allow for commoditization of biomass, allowing producers to optimize quality versus quantity trade-offs for specific feedstocks.

To date, several studies have investigated the yields of bio-oil from fast pyrolysis and reported liquid mass yields ranging from 36 to 62% (dry basis). The yield of upgraded hydrocarbon fuels from these bio-oils is being studied but remains far less understood. The wide range of reported yields coupled with unknown upgrading performance presents a challenge for the bioenergy industry in successfully designing and operating a fast pyrolysis biorefinery.

—Howe et al.

In fast pyrolysis, biomass is rapidly heated in an inert environment at temperatures ranging from 450 to 550 °C to produce a mixture of vapors and liquid aerosols (bio-oil), solids (char), and noncondensable gases such as CO₂, CO, and CH₄. Organic liquid yields vary widely, depending upon the amount of total ash present in the feedstock, amounts of specific inorganic compounds such as alkali and alkali earth metals, syringyl/guaiacyl (S/G) ratios of the feedstocks, and processing conditions such as temperature, residence time, and particle size.

Because raw fast pyrolysis oil is corrosive and thermally unstable, blending or direct insertion into refinery operations at levels greater than 5% is challenging. Therefore, raw fast pyrolysis bio-oil typically is considered to be an intermediate product that can be upgraded via processes such as catalytic hydrogenation, hydrotreating, and hydrocracking. (Earlier post.)

In the study, eight different feedstocks—six pure feedstocks and two blends—were prepared and characterized at INL: clean (no bark) pine; whole-tree (including bark) pine; tulip poplar; hybrid poplar; switchgrass; and corn stover; and two blends (equal weight percentages whole-tree pine/tulip poplar/switchgrass and whole-tree pine/clean pine/hybrid poplar).

The feedstocks then were processed via fast pyrolysis at NREL to generate bio-oils that were subsequently characterized. The bio-oils were then hydrotreated at PNNL to generate an upgraded hydrocarbon fuel blendstock. This blendstock was characterized and analyzed to determine overall yields from feedstock to upgraded fuel, as well as potential fuel product distributions (gasoline, diesel, jet fuel).

Results showed overall fuel yields of:

• 17% (switchgrass)
• 20% (corn stover)
• 24% (tulip poplar, blend 1, blend 2)
• 25% (whole-tree pine, hybrid poplar)
• 27% (clean pine)

Simulated distillation of the upgraded oils indicated that the gasoline fraction varied from 39% (clean pine) to 51% (corn stover), while the diesel fraction ranged from 40% (corn stover) to 46% (tulip poplar). Little variation was seen in the jet fuel fraction at 11–12%

Hydrogen consumption during hydrotreating, a major factor in the economic feasibility of the integrated process, ranged from 0.051 g/g dry feed (tulip poplar) to 0.070 g/g dry feed (clean pine). The blends required less H₂ than any of the woody feedstocks with the exception of the tulip poplar.

These results illustrate three possible advantages of blends over single-material feedstock materials. First, blends may have conversion performance higher than that of the sum of their constituents because of nonlinear conversion, i.e., chemical interactions related to composition may result in conversion performance that is not accurately predicted by simply summing the weighted fractions of the pure components. Second, blends may consume less H₂ than would be expected based on the sum of their constituents. Third, because of the economics of supply and demand, large quantities of blended feedstocks will likely be available for lower costs than single-feedstock materials.

… Another important consideration is that nonlinearities in conversion and upgrading processes may cause blended materials to perform better or worse than separated feed streams. Specifically, results obtained from clean pine, whole-tree pine, and other samples in this work indicate that the effects of small changes in key properties, such as mineral content, may have profound effects on conversion performance. However, larger changes in key properties do not have correspondingly larger effects on conversion, indicating that the relationship between feedstock attributes and conversion properties is nonlinear. These results also demon- strate that the product distributions of upgraded fuels can vary widely depending on the feedstock.

—Howe et al.

Funding was provided by the Department of Energy’s Bioenergy Technologies Office.

Resources

• Daniel Howe, Tyler Westover, Daniel Carpenter, Daniel Santosa, Rachel Emerson, Steve Deutch, Anne Starace, Igor Kutnyakov, and Craig Lukins (2015) “Field-to-Fuel Performance Testing of Lignocellulosic Feedstocks: An Integrated Study of the Fast Pyrolysis–Hydrotreating Pathway” Energy & Fuels doi: 10.1021/acs.energyfuels.5b00304