Separation and Purification Technology 63 (2008) 444–451 Contents lists available at ScienceDirect Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur Extraction of ethanol with higher alcohol solvents and their toxicity to yeast Richard D. Offeman ∗ , Serena K. Stephenson 1 , Diana Franqui, Jessica L. Cline, George H. Robertson, William J. Orts U.S. Department of Agriculture, Western Regional Research Center, 800 Buchanan Street, Albany, CA 94710, USA article info Article history: Received 12 September 2007 Received in revised form 4 June 2008 Accepted 5 June 2008 Keywords: Solvent extraction Ethanol Distribution coefficients Yeast toxicity abstract In a solvent extraction screening study, several -branched alcohols in the 14–20 carbons range show improved extractive performance to recover ethanol from aqueous solutions compared to commonly studied solvents such as oleyl alcohol and 1-dodecanol. These -branched alcohols were selected for screening based on extrapolation of results in earlier work with lower molecular weight aliphatic alcohol solvents, that indicated higher separation factors should be realized when hydroxyl position is mid-chain, and there is branching. Solvent toxicity to a commercial yeast commonly used in fuel ethanol production also was evaluated for these as well as several lower molecular weight alcohols. For the alcohols studied, those containing 12 or fewer carbons were toxic or inhibitory to the yeast; those containing 14 or more carbons were non-toxic and non-inhibitory. Published by Elsevier B.V. 1. Introduction The use of renewable feedstocks for conversion to transporta- tion fuels is growing rapidly. For fuel ethanol, following high-starch grains, the next large source of feedstock is lignocellulosic biomass [1]. A difficulty with this source is that the fermentable C6 sug- ars that result from hydrolysis are typically less concentrated than those that are derived from high starch grains. Significant amounts of soluble hemicellulose drive up the viscosity in the fermentor. In addition, fermentation inhibitors may be present, though this depends on the pretreatment hydrolysis method [2–4]. For these reasons, lignocellulose-based fermentations are more dilute than high starch grain-based fermentations, and the alcohol product is necessarily significantly less concentrated [5,6]. This is a prob- lem when distillation is used as the alcohol recovery method. As distillation feed concentration drops (i.e. for lignocellulosic feed- stocks), distillation energy use (and cost) rises exponentially [7,8]. In a typical fuel ethanol plant in the U.S., the fermentation broth is distilled in a system using a beer column and a rectification column, then dehydrated by pressure-swing adsorption using molecular sieve adsorbents. A well-integrated process requires approximately 15,600 BTU/US gallon (4350 MJ/m 3 ) of anhydrous ethanol [8–10]. This is equivalent to ∼20% of the energy content of the ethanol pro- duced, based on the lower heating value (LHV) of 75,700 BTU/US gal ∗ Corresponding author. Tel.: +1 510 559 6458; fax: +1 510 559 5818. E-mail address: Richard.Offeman@ars.usda.gov (R.D. Offeman). 1 Present address: Dow Chemical Company, Freeport, TX 77541, USA. (21,100 MJ/m 3 ). The LHV assumes that the latent heat of vaporiza- tion of water in the fuel and the reaction products is not recovered. It is useful in comparing fuels where condensation of the combustion products is impractical, as in automobile engines. There are a variety of alternatives to distillation for recover- ing ethanol from aqueous solutions such as fermentation broths [11,12]. These include membrane permeation, vacuum stripping, gas stripping, solvent extraction, adsorption and various hybrid processes. Depending on the concentration of the feed solution and other factors, some of these methods have the potential to be less energy intensive than distillation. For instance, Othmer and Ratcliffe [13] showed substantial energy savings for solvent extrac- tion over distillation for recovering ethanol from dilute feeds of 2% ethanol, and lesser savings for 5% feeds. A reduction in this sepa- ration energy requirement would increase the net energy ratio for ethanol production, reduce carbon emissions, and reduce operating costs. This paper describes a search for high performance solvents to be used in liquid–liquid solvent extraction of ethanol from dilute aqueous solutions. There are many important criteria for select- ing a liquid–liquid extraction solvent, and these criteria can differ depending on the extraction methodology. For instance, in an extractive fermentation process, the solvent contacts the fermenta- tion broth or a filtered portion of the broth, and after disengagement from the solvent, the aqueous phase is returned to the fermentor. In this case, some important criteria are [14]: (1) good extraction per- formance (i.e., partitioning of the product between the solvent and aqueous phases), (2) low solvent solubility in the aqueous phase, (3) low solvent toxicity (to workers, to the environment, to fer- 1383-5866/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.seppur.2008.06.005