Guidance on the Remediation of Ethanol Fuel Releases: A Conceptual Model Approach Kirk O’Reilly1, Ravi Kolhatkar2, and Tim Buscheck3 1. Exponent, Bellevue, WA; 2. Chevron Energy Technology Co. Houston, TX; 3. Chevron Energy Technology Co. Richmond, CA.

Abstract As the use of ethanol as a fuel additive has increased, so has research on its behavior in the subsurface. This research has highlighted similarities and differences between releases of traditional gasoline, ethanol-blended gasoline (EBG), and fuel grade ethanol (FGE). Understanding the subsurface behavior of various fuel components is critical to developing successful strategies for the assessment and remediation of EBG and FGE release sites. An important difference between ethanol and the hydrocarbon components of these fuels is their distribution between the vadose zone, capillary fringe, and groundwater following a release. In the absence of ethanol, gasoline and water are present as two separate phases, and equilibrium partitioning limits the mass of fuel constituents that dissolve in groundwater. But the interaction of fuel and water changes with ethanol concentrations. At low concentrations, ethanol can be drawn out of EBG by soil moisture and water in the capillary fringe, while at the high ethanol concentrations found in FGE, hydrocarbons, ethanol, and water are present as a single phase until dilution causes phase separation of fuel hydrocarbons. FGE can also cause mobilization and redistribution of residual NAPL from prior hydrocarbon releases. Relative biodegradability and its influence on environmental fate are additional factors that differentiate ethanol and hydrocarbons. The API will soon publish a guidance document to assist project managers in developing successful assessment and remediation strategies. It uses a conceptual model approach to explain the subsurface behavior of ethanol-based fuels and to set the technical basis for site management decisions. This paper presents these conceptual models, and also summarizes key recommendations and suggestions for site remediation.

Introduction As a result of a number of legal and economic factors, use and transport of ethanol fuel blends in the United States has increased. Even with upgraded underground storage tanks, the release of ethanol-containing fuels is possible. Because of significant differences in the chemistry of ethanol compared to hydrocarbon gasoline constituents, ethanol behaves differently in the environment (Powers et al. 2001). Effectively assessing and remediating ethanol fuel releases requires an understanding of these differences. In this document, a conceptual model approach is used to describe hydrocarbon and ethanol transport behavior and explain their expected distribution within the subsurface environment. These conceptual models provide the basis for discussing how the presence of ethanol may affect assessment and remediation methods.

The API will soon publish a guidance document to assist project managers in developing successful assessment and remediation strategies. Chapter I of the API Guidance includes a review of key finding and recommendations, as well as conceptual models of EBG and FGE releases. Detailed guidance on assessment and remediation follows. Chapter II provides the technical basis for the conceptual models. The goal of this paper is to summarize finding presented in the API guidance. Domestically, ethanol is typically blended into gasoline in the range of 5 to 20% (E5 to E20), with 10% ethanol blends (E10) being the most common. In this document, these fuels are called ethanol-blended gasoline (EBG). The ethanol used to make EBG is denatured with 5% gasoline (E95), while a number of vehicles can run on an 85% ethanol blend (E85). In this document, E85–E95 are called fuel grade ethanol (FGE). Retail sites primarily handle EBG. At these sites either small chronic or in rare cases large releases from a catastrophic tank failure may occur. Because ethanol is added just prior to transport from terminals, such storage and blending facilities handle large volumes of FGE. While unlikely, larger catastrophic FGE releases can occur. Another potential source of a large volume release is during transportation by truck or rail. Such releases may be either EBG or FGE. Guidance for dealing with these various scenarios is provided.

Conceptual Models of Ethanol Fuel Releases Ethanol Blended Gasoline As a gasoline constituent, ethanol is fundamentally different from hydrocarbon compounds. Hydrocarbons are non-polar and so repel water. This results in a separate NAPL phase, low aqueous solubility, and only minor changes in the bulk composition over time due to dissolution. In contrast, ethanol is highly polar and thus infinitely soluble in water. This polarity causes even small amounts of water to draw ethanol out of gasoline. A means of evaluating potential impacts of ethanol on site management decisions is to compare conceptual models for the behavior of traditional gasoline to that for ethanol blended fuels. Following a release, gasoline moves down through the vadose zone as a liquid. Residual NAPL remains associated with the vadose zone soils. With sufficient volume, the NAPL displaces water in the capillary fringe and pools at the water table. If the depth of the water table fluctuates, a smear zone of NAPL is created. A smear zone is defined as the vertical extent over which both residual and above-residual NAPL exists in the aquifer. For traditional fuels, the primary source of dissolved contaminants of concern (COCs) is partitioning from NAPL in the smear zone beneath the water table. Additional loading of COCs can occur as a result of infiltration of water that contacts the NAPL above the water table. Studies have shown that soil moisture is sufficient to draw ethanol out of EBG. (McDowell and Powers 2003). So as fuel moves through the vadose zone, it creates a trail of

ethanol-enriched soil moisture, while the bulk fuel composition becomes depleted in ethanol. When EBG comes in contact with the capillary fringe, more of the ethanol partitions into the aqueous phase. Little of the ethanol gets beyond the capillary fringe. As a limited amount of the ethanol actually reaches the groundwater, its concentration in the mobile aqueous phase is expected to be less than a few percent. The solubility of hydrocarbons in an ethanol/water mixture is greater than in pure water because of cosolvency (Heermann and Powers, 1998; Rixey et al., 2005). The effect increases with ethanol concentration, and is greatest for compounds with lower aqueous solubility. While the maximum dissolved concentration of individual constituents in a gasoline / water system can be predicted using Raoult’s law, it becomes more difficult in the presence of ethanol. As cosolvency is inversely related to solubility, it has less of an effect on benzene than other less soluble hydrocarbon fuel components (Corseuil et al. 2004). Because aqueous ethanol concentrations exceeding 10% are required to see significant effects, cosolvency is expected to have a limited impact on effective solubility of benzene in groundwater. (Rixey et al. 2005). More BTEX may be retained above the water table because of elevated ethanol concentrations in the vadose zone moisture. (ÖsterreicherCunha et al. 2009) While the total volume is low relative to the volume of groundwater, it may serve as a source of dissolved benzene because of infiltration. With sufficient soil moisture, reduction in surface tension may result in a mobile aqueous ethanol solution that is enriched in dissolved hydrocarbons (Österreicher-Cunha et al. 2009).

Figure 1- Conceptual model of an ethanol blended gasoline release. Ethanol partitions into soil moisture and the capillary fringe. While little ethanol is expected to reach groundwater, degradation products such as acetate (white arrows) may be present. The vapor phase can also result in the transfer of ethanol and other constituents from EBG to aqueous phase. Pure ethanol has a relatively low vapor pressure given its molecular weight, because of the hydrogen bonding between polar molecules. But when ethanol is added to a non-polar matrix, its vapor pressure is much higher (Kovarik and Hermes 2005). This results in a vadose zone that can be elevated in ethanol vapors, thus serving as a

transport pathway. Because of the low Henry’s constant, the ethanol in the vapor phase can partition into the soil moisture. Figure 2 presents a conceptual model of the mass distribution within the subsurface of ethanol, BTEX, and the less soluble bulk fuel. With or without ethanol, a majority of the mass of the fuel hydrocarbons and aromatic compounds remain as part of the source or are lost to volatilization in the vadose zone. Compared to the non-aromatic fractions of the gasoline, more of the BTEX is lost to dissolution in vadose zone soil moisture, the capillary fringe, or groundwater, but it is still only a small fraction of the mass. Ethanol-induced cosolvency can increase the mass of dissolved BTEX, but this will have little impact on overall mass distribution. Because of its affinity for water, little of the ethanol will remain associated with the source. In the absence of infiltration, most of the ethanol will remain dissolved in soil moisture or the capillary fringe. Infiltration can serve to transport ethanol to the groundwater.

Figure 2- Conceptual model of the mass distribution of EBG constituents (small circles) following a release. The large circle represents the makeup of EBG. While most of the hydrocarbons remain associated with NAPL or are lost to volatilization, ethanol partitions to soil moisture and the capillary fringe. To assess whether the presence of ethanol in groundwater is a potential or growing concern, we evaluated groundwater monitoring data that had been submitted to the State of California. The state’s water resource board files these data in the publicly accessible GeoTracker database. Data from three urban counties, Los Angeles, San Diego, and Contra Costa, were used. Results of ethanol analysis (detection limits ≤ 100 µg/L) were available from 13,000 individual samples taken from 2,280 wells at 207 sites collected between 2000 and 2009. Ethanol was detected in 101 (0.78%) of the samples. Of these, six (0.05%) had an ethanol concentration greater than 10,000 µg/L and 21 (0.16%) had a concentration greater than 1,000 µg/L. Ethanol was detected more than once in only three of the wells. The maximum detected ethanol concentration was 96,000 µg/L.

Fuel Grade Ethanol Fuel grade ethanol is a very different material from traditional gasoline. While gasoline and water are always two phases in the absence of ethanol, FGE mixes as a single phase with both. As FGE contacts water, the ethanol-gasoline-water mixture remains as a single phase until dilution with water causes the ethanol concentration to drop below about 70% m/m. At lower ethanol concentrations, the gasoline separates as a NAPL phase. So if FGE is released in the subsurface, much of the gasoline denaturant will end up where the FGE first contacts the capillary fringe. Even with this high concentration of ethanol, little is expected to reach the groundwater in the absence of infiltration or water table fluctuation. Under these conditions, there may be enough ethanol to result in cosolvency. But as the effect of cosolvency is inversely related to a compound’s solubility, it will have less impact on benzene than on other gasoline constituents.

Figure 3- Conceptual model of a fuel grade ethanol release. The gasoline in FGE phase separates as a result of mixing with water in the capillary fringe. Most of the ethanol remains above the water table. The results of experimental controlled releases suggest that ethanol is mostly confined to the capillary fringe, and mixing with underlying saturated zone water is minimal (McDowell and Powers 2003; Stafford 2007). In the absence of infiltration or pumping, a zone of elevated ethanol concentration is found above the water table and little of the ethanol reaches the groundwater. Stafford et al. (2009) measured a maximum saturated zone ethanol concentration of 0.08% or 630 mg/L following an experimental release of neat ethanol even though the capillary zone concentration exceeded 85%. With infiltration, water with elevated ethanol and cosolvency-controlled BTEX concentrations may impact groundwater. Because FGE is stored at locations where previous fuel releases may have occurred, it is important to understand how an FGE release may impact the mobility of existing NAPL. Because fuel hydrocarbons are soluble in ethanol, the FGE acts as a solvent, both dissolving and remobilizing the NAPL (Falta 1998). As the highly concentrated (> 70%) ethanol fuel mix migrates downward, preexisting NAPL is dissolved and migrates vertically with the bulk fuel (McDowell et al. 2003). When the bulk fuel reaches the capillary fringe it reduces the

capillary fringe height (Powers 2001; McDowell et al. 2003), spreads within the capillary fringe, and migrates according to the hydraulic gradient (Stafford 2007). Dilution of ethanol by pore water causes NAPL to partition out of solution, generating a new NAPL source nearer to the water table (McDowell et al. 2003; Stafford 2007; Stafford et al. 2009).

Figure 4- Conceptual model of a fuel grade ethanol release at a site with pre-existing NAPL. Residual NAPL (A) is mobilized by the FGE (B). Redistribution and increased mobility was the likely cause of increased NAPL thickness measured in monitoring wells following a 20,000-gallon FGE release at a fuel terminal (Buscheck 2003). Monitoring wells may not give an accurate indication of the NAPL thickness in the presence of FGE. If FGE gets into a well, the ethanol will dissolve, leaving any fuel that had been dissolved in the FGE as a separate phase in the well. The amount of NAPL found in the well can be a function of the relative fuel volume in the FGE and the volume of FGE that entered the well. Benzene concentrations increased in some wells near the release. While the ethanol concentrations as high as a few percent were detected in groundwater, ethanol was depleted over the first couple of years following the release. Generation of elevated methane concentrations following a catastrophic release of FGE can be a long-term concern. Because of a combination of inhibition of microbial activity at high ethanol concentrations, the time required to deplete soluble electron acceptors, and the slow growth rate of methanogens, it can be many months or more after an FGE release before methane concentrations begin to increase in groundwater. While methanogenesis occurs in the aqueous phase, methane can off-gas and build up in the vadose zone if concentrations exceed water solubility. Methane-saturated groundwater and elevated vadose concentrations have been detected following large FGE releases at terminals and those caused by derailment of tank cars (Buscheck 2003; Spalding 2009). The gas produced by methanogenesis has about a 1:1 ratio of methane and carbon dioxide. So while the concentration of CH4 can reach 50%, it shouldn’t be explosive in the

vadose zone because of the lack of oxygen and the inability of a flame to propagate through a porous medium. An explosive or fire risk is possible if the methane builds up in confined spaces where oxygen is present. In the presence of air, the explosive range of methane is between 5 and 15%. Monitoring of areas, such as basements, adjacent to an FGE release should be considered.

Remediation at Ethanol Release Sites Critical remedial drivers at gasoline release sites will continue to be dissolved and vapor phase aromatic compounds. Because ethanol can result in zones of elevated dissolved BTEX concentrations and somewhat longer plumes, it may increase the need for or extent of remediation at some sites. Concerns caused by elevated methane concentrations may also drive the need for remediation. The primary purpose of remediation is to reduce the risk to human health and the environment, usually by reducing the mass of contaminants in the subsurface to appropriate levels. This can be accomplished either by removing NAPL or by extracting water or vapors containing COCs. As the presence of ethanol can change the behavior of NAPL and phase transfer characteristics, it can impact the effectiveness of remedial technologies. The impact of ethanol on commonly employed remediation techniques for gasoline release sites are described below: Excavation—Ethanol should have little impact on the effectiveness of excavation as a remedial technology. Given the reduced surface tension caused by the ethanol, the amount of soil impacted may be higher relative to a release of a similar volume of traditional gasoline, but the residual NAPL concentration would be lower. Because of the transfer of ethanol from NAPL to soil moisture, vadose zone soils that are not directly impacted by NAPL may contain some ethanol. Soil Vapor Extraction (SVE)—SVE works by drawing air through the vadose zone to remove volatile components. Ethanol does not impact the volatility of hydrocarbons in NAPL. And when mixed in gasoline, ethanol has sufficient vapor pressure for removal by SVE. Given the high biodegradability of ethanol and its products, the oxygen introduced into the vadose zone by SVE will promote biodegradation. Thermal and Catalytic Systems—Thermal and catalytic systems for offgas treatment will be effective for ethanol-rich air streams. Dual Phase Extraction—In dual or multi-phase extraction, water is pumped at a sufficient rate to create a cone of depression under mobile NAPL. The purpose is to remove both fuel and dissolved phase constituents. While an EBG release should have little impact on the basic operation of an extraction system, it might result in some mixing of ethanol into the groundwater phase. An FGE release may result in

additional recovery of NAPL as a result of increases in mobility, but the presence of ethanol may interfere with oil water separation and could potentially result in biological fouling. Air Sparging—In air sparging, air is injected below the water table. It removes contaminants by stripping dissolved constituents from groundwater and by introducing oxygen to stimulate biodegradation. While ethanol should have minimal impact on the stripping efficiency of hydrocarbons, the stripping of either ethanol or acetate is limited. Sparging is likely to reduce the concentration of dissolved methane. Oxygen introduced into groundwater through sparging will stimulate degradation of ethanol and acetate. It could reduce the electron acceptor depleted ‘anaerobic shadow’ caused by ethanol degradation, but because of the high oxygen demand the influence may be limited. The growth of ethanol degrading bacteria may result in fouling. Pump and Treat—Pump and Treat works by removing dissolved constituents from areas within the capture zone. The greatest impact of ethanol will be on the water treatment system. Ethanol and its products can stimulate bacterial growth, which may result in biofouling, especially wherever oxygen is introduced. This can lead to fouling and plugging of water treatment units used to treat dissolved BTEX, such as activated carbon canisters and air stripping towers. While biologically-based treatment is effective for ethanol, activated carbon, air strippers, or separators are not. Ethanol may induce corrosion in steel parts. Monitored Natural Attenuation (MNA)—Natural attenuation relies on the activity of naturally occurring organisms present in the subsurface to degrade compounds of concern. While ethanol is readily degraded, its presence can inhibit the degradation of aromatic compounds such as benzene by preferential utilization of limited electron acceptors. Natural attenuation may still be a suitable technology at some sites, but reduced efficacy can result in elongated plumes and a longer time frame for treatment. In the absence of ethanol, geochemical changes related to electron acceptor concentrations are used as evidence of the degradation of BTEX. But because ethanol degradation results in similar effects, geochemical changes alone may not provide evidence of BTEX biodegradation. MNA Tier 1 data indicating a stable or shrinking BTEX plume are required. Biostimulation—Like MNA, biostimulation relies on the activity of naturally occurring organisms present in the subsurface to degrade compounds of concern. The difference with biostimulation is that environmental conditions are modified to increase the degradation rate. At hydrocarbon-impacted sites, the most common method for stimulated degradation is to add a source of electron acceptors. While oxygen is typically added, stimulants including nitrate and sulfate have been tried.

Sources of oxygen include air, pure oxygen, aerated water, peroxide, and ozone. Even in the absence of ethanol, the challenge is to adequately deliver and distribute the added oxygen or other stimulant. Ethanol can negatively impact BTEX treatment because it results in an increased electron acceptor demand. This reduces the availability and transport of added electron acceptors. Bacterial growth supported by ethanol degradation may result in fouling of wells or injection points.

Acknowledgment We appreciate the input of Brent Stafford of Shell Global Solutions. References Buscheck, T.E. 2003. Answers to frequently asked questions about ethanol impacts to groundwater. API Soil & Groundwater Research Bulletin No. 20. Corseuil, H.X., B.I.A. Kaipper, and M. Fernandes. 2004. Cosolvency effect in subsurface systems contaminated with petroleum hydrocarbons and ethanol. Water Res. 38:1449 1456. Falta, R. 1998. Using phase diagrams to predict the performance of cosolvent floods for NAPL remediation. Groundwater Monitor. Remed. 18:94 102. Heermann, S.E., and S.E. Powers (1998). Modeling the partitioning of BTEX in waterreformulated gasoline systems containing ethanol. J. Contam. Hydrol. 34:315–341. Kovarik, W. and M.E. Hermes. 2005. Fuels and society: www.chemcases.com/converter/converter-24.htm.

Back to ethanol.

McDowell, C.J., and S.E. Powers. 2003. Mechanisms affecting the infiltration and distribution of ethanol-blended gasoline in the vadose zone. Environ. Sci. Technol. 37:1803 1817. Österreicher-Cunha, P., E. do Amaral Vargas, Jr., J.R.D. Guimarães, G.P. Lago, F. dos Santos Antunes, and M.I.P. da Silva. 2009. Effect of ethanol on the biodegradation of gasoline in an unsaturated tropical soil. Int. Biodeterior. Biodegrad. 63:208 216. Powers, S.E. 2001. Transport and dissolution of ethanol and ethanol-blended gasoline in the subsurface. Workshop on the Increased Use of Ethanol and Alkylates in Automotive Fuels in California. Hosted by Lawrence Livermore National Laboratory, April 10-11, 2001. Powers, S.E., D.W. Rice, B. Dooher, and P.J.J. Alvarez. 2001. Will ethanol-blended gasoline affect groundwater quality? Environ. Sci. Technol. 35:24A 30A.

Rixey, W.G., X. He, and B.P. Stafford. 2005. The impact of gasohol and fuel-grade ethanol on BTX and other hydrocarbons in groundwater: Effect of concentrations near a source. API Soil & Groundwater Research Bulletin No. 23. Spalding, R.F. 2009. Impact of ethanol releases: long term monitoring results. Presented at the Interstate Technology and Regulatory Council’s Biofuels Team Meeting, Cincinnati, OH, April 22-24, 2009. Stafford, B.P. 2007. Impacts to groundwater from releases fuel grade ethanol: Source behavior. Dissertation. Environmental Engineering Graduate Program, University of Houston. Stafford, B.P., N.L., Capiro, P.J., Alvarez, and W.G. Rixey. 2009 . Pore water characteristics following a release of neat ethanol onto a pre-existing NAPL. Groundwater Monit. Remed. 29(3):93-104. Yu, S., J.G. Freitas, A.J.A. Ungar, J.F. Barker, and J. Chatzis. 2009. Simulating the evolution of an ethanol and gasoline source zone within the capillary fringe. J. Contam. Hydrol. 105:1 17.

Authors Kirk O’Reilly, Ph.D., J.D. Kirk O’Reilly is a Managing Scientist in Exponent’s Environmental Sciences practice. He has more than 20 years of experience investigating the interaction between environmental and biological chemistry. He is a recognized expert in bioremediation, environmental chemistry, and innovative remedial technologies. Specific contaminants studied include crude oil, refined products, chlorinated solvents, wood treatment compounds, pesticides, and fertilizers. He has developed innovative methods for monitoring the transformation and assessing the risk of petroleum, and has participated in collaborative research projects with regulators at the federal, state, and local levels. Dr. O’Reilly is a member of the Washington State Bar. Ravi Kolhatkar, Ph.D. Ravi Kolhatkar is a Staff Environmental Hydrogeologist with Chevron. He is a member of the API Soil and Groundwater Technical Task Force and has been working on fuel oxygenates remediation, natural and enhanced attenuation, and vapor intrusion issues for more than 10 years. Dr. Kolhatkar has a Ph.D. in Chemical Engineering from the University of Tulsa, Oklahoma, and an MBA from the University of Chicago, Illinois. Tim Buscheck, P.E. Tim Buscheck is a Consulting Hydrogeologist in the Groundwater Team of the Health, Environment and Safety Group. He consults with various Chevron Operating Companies on site assessments and remediation for marketing, chemical, and refining facilities throughout the United States and internationally. Mr. Buscheck manages a Remediation Strategic Research program. He has authored papers on the subjects of compound specific isotope analysis, natural attenuation, ethanol fate and transport, and multi-site plume studies. He has an M.S. in Geological Engineering from the University of California, Berkeley, and a B.S. in Chemical Engineering from Lafayette College.