Volume 43 Number 8
A Dealer’s Perspective: Reducing MTBE & BTEX in Private Wells
Summary: Water treatment dealers serving private water supplies may be confronted with the daunting task of remediating MTBE and BTEX compounds due to gasoline contamination. As important as initial water quality analysis and system design is follow-up metering and monitoring. Here’s a good primer on the topic.
Methyl tertiary butyl ether (MTBE) and benzene, toluene, ethylbenzene and xylenes (BTEX compounds) are volatile organic compounds (VOCs) associated with gasoline contamination of water. MTBE is a fuel oxygenate added in concentrations of 11-15 percent in reformulated gasoline to promote complete burning and reduced emissions of carbon monoxide and organic combustion products. In the 1990s, MTBE was the third most produced organic chemical in the United States—18 billion pounds per year (less only than ethylene and propylene).
MTBE and BTEX compounds can enter water supplies from leaking underground storage tanks, automobile tanks, recreational boating and spillage from lawnmowers and garden equipment. In areas where leakage has occurred, concentrations found in water supplies can be very high.
Table 1 provides an overview of maximum contaminant levels (MCLs), odor thresholds, health effects, contamination sources and recommended treatment approaches for MTBE and BTEX compounds.
MTBE is highly soluble in water and less readily absorbed by soils. Thus, it tends to "lead the plume" in waters contaminated from these sources and can spread rapidly in groundwater long before other BTEX compounds are detected. The U.S. Geological Survey recently reported finding MTBE in 20 percent of groundwater in areas where it is used, versus only 2 percent in areas that don’t use reformulated gasoline.1 Its high solubility also makes MTBE more difficult to remove from water via conventional treatment processes than BTEX compounds.
Adsorption of MTBE and BTEX compounds by granular activated carbon (GAC) is a complex process involving the nature of the contaminants themselves, water and other contaminants it may contain, and activated carbon used in the application. Contaminants of lower solubility and volatility are generally more readily adsorbed.
Time and amount
All organic compounds exhibit a degree of attractive force to the surface of activated carbon. This can be determined by measuring how much of a contaminant is removed in a specific time or how long it takes to remove a specified amount (see Factoring isotherms). The rate of adsorption is driven by the equilibrium concentration of the contaminant in the water. As the contaminant concentration is reduced as it passes through the carbon, the driving force that pushes it onto the carbon is also reduced. For example, if it takes "x" time for the carbon to reduce 50 percent of the contaminant, it will take twice the time to reduce 75 percent of the contaminant and three times as long to reduce 87.5 percent and 10 times the time to reduce 99.9 percent (see Chart 1). So the effectiveness of activated carbon is related to how well the contaminant is attracted to the carbon, the contaminant concentration in the raw water and the amount of time the water containing the contaminant is in contact with the activated carbon. This is known as "empty bed contact time" (EBCT). EBCT is calculated as follows:
While carbon manufacturers may conservatively recommend 15 minutes of EBCT, field experience indicates an EBCT of 1.5 to 3 minutes is often effective. Depending on the degree of removal required, the EBCT can be increased by using more carbon or lowering the flow rate through the carbon. Using a smaller GAC mesh size has the effect of increasing the EBCT relative to a larger size because the surface area of the carbon is more available for adsorption, but it will also exhibit higher pressure drop at a given flow rate.
Entering the MTZ
Figure 1 represents a single packed GAC adsorption column. As the influent contaminant(s) progress through the column for a specific number of gallons, the upper part of the carbon bed (shaded) becomes saturated, and the contaminant(s) concentration in the raw water is at equilibrium with the concentration of contaminant(s) adsorbed onto the carbon’s surface, i.e., no new adsorption is taking place. The next zone down is called the mass transfer zone (MTZ) where new adsorption of the contaminant(s) is taking place in accordance with the "half life" adsorption rate driven by influent contaminant concentration (see Chart 2). The length of the MTZ is a function of the adsorption rate, the volume and mesh size of carbon in the column and the flow rate of the contaminated water through the column. The greater the length of the MTZ, the sooner the column will show contaminant breakthrough in the effluent. This length of MTZ also represents a portion of the total carbon volume that’s yet to reach saturation. Thus, more contaminant(s) could be adsorbed onto the carbon, although greater amounts of breakthrough would be seen in the effluent as complete carbon bed saturation was reached, until finally the effluent concentration reached the influent concentration. In single column applications where contaminant breakthrough isn’t desired—the longer the MTZ, the less efficient is carbon bed utilization.
Using two or more adsorption columns in a series is a technique commonly employed to increase carbon bed utilization and or to provide a margin of safety. This way the second bed catches the breakthrough of the first bed until it reaches saturation. A safer, more conservative approach when only two tanks are used is to only operate the first column until breakthrough begins, then replace the first column with the second column and add a new second column. To get the benefits of both maximum bed utilization and a safety margin, three columns in a series are used. When column one reaches exhaustion, it’s removed, each column moves up one spot and a new third column is put into service. This approach is best when MTZ length is very long, safety is critical and maximum bed utilization is desired. A multiple column approach can also be a good way to pilot plant applications, whereby performance data can be obtained over time. The installation can then be optimized based on real data.
System design & installation
A basic design for a residential or small commercial treatment system is shown in Figure 2. The carbon tanks are fed water filtered to a minimum of 20 microns, unless other water quality parameters such as hardness, iron, manganese or bacteria dictate additional pre-treatment. The water then passes through carbon beds in series, is again filtered to 5 microns and passed through an ultraviolet (UV) disinfection unit to help ensure microbial potability before going to service. Remember the pre-treatment requirement for effective UV application when considering additional pretreatment options. Using quick connects with flexible tubing for GAC tanks will make replacement easier. Sampling ports and pressure gauges should be incorporated for influent water quality and the effluent of each GAC tank to facilitate monitoring GAC tank performance and pressure drop. Additional shut-off and drain valves should be employed to facilitate equipment servicing. UV systems may require UV meters, solenoid valves, wipers and alarms.
Field experience has shown that GAC tank sizes ranging from 1 cubic foot (9 × 40-inches) to 2 cubic feet (12 × 42-inches) will handle flow rates from 4.5-to-8 gpm. These exchange tank sizes are often equipped with rubber bases and bump bands for frequent transportation. Their smaller size makes them easier to transport, more flexible in tight installations and provide difficult access parameters. Tank handling is easier if the water is blown out before transport. Larger flow rates or higher concentrations can be handled by splitting the water flow into two or more parallel trains of carbon tanks in a series, as the larger carbon volume increases the EBCT.
The carbon tanks are typically not automatically backwashed to help preserve the integrity of the MTZ and to prevent introduction of contaminants into the bottom of the bed. There also may be problems associated with discharging un-adsorbed contaminants to drain. Generally, it’s best to follow the carbon manufacturer’s recommendation in preparing the carbon beds prior to replacing the tanks in the field. This usually involves filling the clean tanks with carbon, soaking the carbon overnight, then backwashing the carbon to remove fines, rinsing the carbon and finally purging the freeboard water capacity in the tank to make it lighter. Disposal of spent carbon should be done in accordance with state regulations.
All new installations should be tested for hardness, iron, manganese, pH, total dissolved solids, sulfur, coliform, total plate count, iron/sulfur bacteria, total organic carbon (TOC) and radon. Annual monitoring of these items is a good idea as water quality can change relatively easily in a well contaminated with gasoline. Shallow well sources contaminated with gasoline are especially susceptible, since they’re frequently under the influence of surface water. Iron and manganese concentrations often appear to increase over time as the gasoline compounds and associated bacteria produce redox conditions in the water supply.
Initial monitoring testing should be done for MTBE and BTEX compounds on a frequent basis, usually monthly, to ensure performance. These samples are taken on the influent raw water and the effluent of each tank, after several bed volumes have been purged at peak flow rate. Meter readings are also taken to help evaluate performance. Coliform monitoring should also be conducted to monitor UV performance. Over time it may be possible to reduce sampling frequency to quarterly or semi-annual readings, as historical performance data are obtained. These data provide the signal to change out carbon tanks (always take meter readings when changing tanks), so design changes like adding more tanks or trains can be evaluated. Carbon tanks should be changed at least annually, even if monitoring does not indicate the need. This helps ensure the GAC doesn’t get loaded up with too many hydrocarbons, bacteria slime, radon, iron, odors and so forth. Longer service life can be evaluated in those instances where water quality is less problematic. Cartridge filters should be changed at least every six months (or as pressure drop requires) and UV lamps should be changed as recommended by the manufacturer.
Due diligence is required in the proper design, installation, monitoring and maintenance of GAC systems for MTBE and BTEX reduction in gasoline contaminated water. The enormous number of variables affecting carbon performance and the relative severity of the potential health threat from both drinking and bathing when these compounds are present dictate taking a conservative, well-thought-out approach that provides an adequate margin of safety for the end user.
1. Kahler, David, "Going the Extra Mile: Identifying and Eliminating MTBE," WC&P, pp. 94-96, March 2001.
2. Snoeyink, Vernon L., "Adsorption of Organic Compounds," Water Quality and Treatment, 4th Edition, American Water Works Association, McGraw-Hill Inc., 1990.
3.Michaud, C.F., "GAC To Become ‘Workhorse’ of Water Purification," WC&P, p. 20, June 1988.
4.Michaud, C.F., "Granular Activated Carbon: If At First You Don’t Succeed…Try, Try Again," WC&P, p. 38, July 1988.
5.Michaud, C.F., "Granular Activated Carbon: Putting It All Together," WC&P, p. 36, August 1988.
About the author
Most carbon manufacturers can provide isotherms—charts representing changes of volume or pressure under conditions of constant temperature—developed in the lab for specific compounds, using a specific activated carbon challenged with a specific contaminant in distilled water. Results of an adsorption isotherm are expressed in terms of the carbon’s capacity (usually as milligrams of contaminant per gram of carbon) for a given contaminant at a specified equilibrium concentration. The capacity of the carbon for a specific compound increases with contaminant concentration.
While these isotherms are a good place to start, actual field performance usually varies because of the many variables that affect carbon performance:
• There’s often more than one contaminant present, competing with varying degrees of attraction for the carbon’s surface. Even a few milligrams per liter of naturally occurring background organics such as tannins can reduce GAC capacity by 50 percent or more.
• Contaminant concentration can fluctuate over time (see Chart 2).
• Microbial degradation can help reduce the concentrations of adsorbed BTEX compounds, but MTBE isn’t readily broken down by bacteria.
• The carbon itself may not be of the same type, mesh size and pore structure; there are even differences between batches of the same carbon.
• The water’s pH, temperature, ionic loading and presence of foulants (such as turbidity, sediment, iron, manganese and bacterial loading) can all affect the performance of the carbon.
There are so many variables to consider that pilot testing on the water source is often the only reasonable means of accurately predicting performance in smaller scale applications such as private water sources.
Of particular concern are those situations where desorption of the contaminants that have been adsorbed can occur, such as:
• A sudden decrease in the influent contaminant concentration(s), especially near bed exhaustion.
• Improved contaminant solubility due to a pH shift in raw water.
• Increased volatility of contaminants due to temperature increase. Contaminants with boiling points under 65°C are more prone to desorption from temperature increase.
Click here to view additional information: beauchamp.pdf