The Synthesis of Hydrocarbons for Fuel and Lubricants via Ceramic Membrane Reactor
In order for biomass based fuel and chemicals to have a significant market penetration they need to compete economically with petroleum based streams, thus finding ways to reduce the cost of the biomass conversion is essential. As part of a project to produce infrastructure compatible fuel and chemicals Ceramatec is developing an electrochemical process, adapted from the Kolbe electrolysis. This process converts fatty acids from a biomass origin into long chain hydrocarbons, which can be further converted into lubricants and fuel using existing infrastructure.
The conventional Kolbe electrolysis has been investigated since 1849 when Kolbe applied it to the synthesis of different hydrocarbons. The reaction pathway is generally considered to follow a mechanism in which a one electron oxidation at an electrode surface generates a radical and CO2. The radical can then undergo homo- or hetero- coupling with other radicals generated at the surface of the electrode. While it has been extensively studied, the Kolbe electrolysis has never been widely commercialized because of low efficiency and yield caused by the high over potentials and reactivity of the electrolyte that are conventionally used.
To circumvent these shortcomings, Ceramatec’s decarboxylation process uses a two compartment electrochemical reactor, where the compartments are separated by a ceramic membrane. The separation afforded by the membrane (NaSelect®) is such that Na-ions are exclusively (>99% efficiency) transferred between the compartments, allowing different electrolytes to be used for the anolyte and catholyte. This affords: 1) the anolyte to be customized for the electrolysis of interest, and 2) the catholyte to be designed to have high conductivity and a low reduction potential. The latter helps reduce the overall cell potential, improving the energy efficiency of the process. The use of this membrane requires the anolyte to contain sodium salts of carboxylic acids, thus the anodic reaction can be represented generically by the reaction below.
2RCO2Na → R-R + 2CO2 + 2e -+ 2Na+
The Na-ions that are produced in the anolyte are then transferred across the membrane to the cathode compartment, where the corresponding reduction reaction occurs, as shown below using water as an example.
2H2O + 2e- +2Na+ → 2NaOH + H2
The sodium hydroxide produced in the cathode compartment can be used to saponify the fatty acid feed stock, making it a closed loop system for sodium. Also, the use of the two compartment reactor causes the distance between the anode and cathode to be separated into three regions: 1) the distance between the anode and the membrane, 2) the membrane thickness, and 3) the distance between the membrane and the cathode. This separation decreases the length of diffusion through the anolyte, permitting the anolyte conductivity to be sacrificed for improved oxidative stability.
We have used these benefits to optimize the decarboxylation process, obtaining saturated hydrocarbons with yields and electrical efficiency over 80% while maintaining a low power consumption of 1.8 kWh/L ($0.18/L). These oxygen free hydrocarbons are produced using a modular process without the use of hydrogen. This process has also been scaled up in a self-contained portable pilot unit designed to produce over 1 L/day of saturated hydrocarbons from a tubular electrochemical reactor. The pilot unit is being used to address scale-up issues at a modular level and obtain data that is needed to help determine the economic feasibility of the process on an industrial scale.
The results of this reactor optimization and scale-up will be discussed. Also, results using the membrane reactor to produce different high value chemicals will be shared.
This material is based upon research supported by the National Institute of Food and Agriculture, U.S. Department of Agriculture, under Agreement No. 2012-10008-20263.
 H. J. Schafer, Topics in Current Chemistry, Electrochemistry IV, 1990, 152, 91.
 M. Karanjikar, S. Bhavaraju, A. V. Joshi, P. Chitta, D. J. Hunt, U.S. Patent 8,647,492, Feb. 2014.