Samuel A. Dinkins, Environmental Specialist

Jason P. Heath, Water Quality Monitoring and Assessment Programs Manager

Ohio River Valley Water Sanitation Commission

5735 Kellogg Avenue, Cincinnati, OH 45228

The Ohio River Valley Water Sanitation Commission (ORSANCO) is currently involved in a sampling program to quantify concentrations of dioxin in the Ohio River. Water quality standards and typical ambient water column concentrations are well below the current analytical detection limit of 10 parts per quadrillion (ppq). High volume water sampling, however, is a new technique which, for the first time, allows for the direct measurement of in-stream dioxin levels. This is accomplished by drawing a large volume of water first through glass fiber filters, which separate and collect the suspended solids. The filtered water then passes through two XAD-2 resin columns that extract the dioxin in the dissolved phase. The filters and columns are then analyzed separately to quantify dioxin levels in both the particulate and dissolved phases.

ORSANCO has successfully used high volume water sampling to quantify ambient water column concentrations on the Ohio and Kanawha Rivers. In each sampling event, 1000 liters of water were filtered over a twelve-hour period. Dissolved dioxin concentrations were slightly above the US EPA water quality criterion of 0.013 ppq. Concentrations in the particulate phase were consistently more than an order of magnitude greater than levels detected in the dissolved phase. These results demonstrate that high volume water sampling is an effective means of attaining detection levels below the water quality standard for dioxin, and that the sampling method could prove to be extremely valuable in quantifying in-stream levels of other pollutants.

In 1995, the Ohio River Valley Water Sanitation Commission, an interstate water pollution control agency serving the Ohio River and its member states, developed and initiated the Ohio River Watershed Pollutant Reduction Program. The program was initiated to address pollutants known to inhibit the beneficial uses of the Ohio River and its tributaries. The long-term goal of the program is to generate the information necessary to achieve water quality objectives utilizing a pollutant reduction strategy. Dioxin was selected as the first pollutant to be addressed under the program due to elevated levels observed in fish tissue collected from the Ohio River and the Kanawha River in West Virginia. The concentrations of dioxin found in the fish suggested a possible water quality problem.

The presence of dioxins in the environment has attracted considerable attention in recent years from the public and scientific community. These compounds are not intentionally produced, but a variety of industrial and combustion sources have been identified. The toxic nature of dioxins at extremely low concentrations poses a very difficult problem for scientists and regulators to address. Typical analytical detection limits for water samples collected using conventional sampling methods ranges from one to ten parts per quadrillion (ppq). The US EPA, however, established a water quality criterion for the protection of human health for 2,3,7,8 tetrachlorodibenzo-p-dioxin (the only dioxin congener with a water quality standard) at 0.013 ppq (EPA 1984). Due to the limitations of the analytical methods, a stream could have a concentration that is two orders of magnitude greater than the water quality standard and still not be detected. As a result of the inability to accurately quantify dioxin concentrations in surface waters that were at levels of concern, an alternative sampling method was necessitated.

The term dioxin refers to a group of 210 different polychlorinated dibenzodioxins and polychlorinated dibenzofurans. Only 17 of these congeners have dioxin-like toxicity, with the best known and most toxic congener being 2,3,7,8 tetrachlordibenzodioxin (TCDD). These compounds are unwanted by-products of various combustion and chemical processes, and have never been intentionally produced with the exception of small quantities synthesized for scientific research. A variety of sources have been identified including pulp and paper mills, incinerators, certain chemical manufacturing processes, wastewater treatment plants, PCB transformer fires, wood treating facilities, automobile exhaust, and possibly forest fires.

Dioxins have been found to cause a wide array of physiological responses in animals and humans. Chloracne, a severe form of acne, is the only response in humans clearly attributed to dioxin exposure. Several epidemiological studies, however, have suggested dioxins may cause cancer (Bertazzi et al 1993), decreased growth (Guo et al 1994), decreased birth weights (Lucier 1991), delayed developmental milestones (Rogan et al 1988), and decreased testis size (Egeland et al 1994). Based on the collective results of various laboratory studies and limited epidemiological studies, EPA classifies 2,3,7,8 TCDD as a probable human carcinogen (EPA 1996).

EPA (1989) developed a standardized method to evaluate the toxicity of complex mixtures of dioxin-like compounds. This procedure assigns a Toxicity Equivalency Factor (TEF) to each of the 17 dioxin and furan congeners with chlorine substitution in the 2,3,7, and 8 positions. Since the 2,3,7,8 TCDD congener is the most toxic, its TEF is set at one. The TEFs of the other congeners are based on their toxicity relative to the 2,3,7,8 TCDD congener (see Table 1). To determine the risk posed by a mixture of dioxins and furans, the concentration of each congener is multiplied by its corresponding TEF. The resulting concentration is expressed in terms of 2,3,7,8 TCDD equivalents (TEQ). The summation of all the TEQs give the estimated total concentration in terms of a 2,3,7,8 TCDD equivalent concentration.

The ultimate fate of dioxins in the environment is their accumulation in aquatic sediments. Dioxins deposited to soil surfaces strongly bind to particulate matter and are either buried in place, resuspended into air, or are transported to surface waters through erosion. Once in water, dioxins primarily sorb to suspended solids. These particles can be transported considerable distances downstream before settling to the bottom (EPA 1987).

Current analytical detection limits for dioxin congeners for water samples collected using conventional sampling techniques (e.g. discrete water sampling with a bailer or Kemmerer) is approximately two orders of magnitude greater than EPA’s water quality standard of 0.013 ppq. Considering that a body of water could have concentrations of dioxins exceeding water quality standards and yet be undetectable, it became obvious that a new sampling technique that allowed for concentrating dioxins from large volumes of water was necessary. The basic concept of concentrating samples is not new to environmental sampling. This technique has been used for several years to sample pollutant levels in ambient air. The difficult task in concentrating large volume samples is capturing the pollutants in both the particulate and dissolved phases without allowing significant break-through of the contaminants. In order to accomplish this, two different pollutant removal mechanisms must be employed. Pollutants bound to the particulate phase can be removed via a filtering system that physically removes all particulate matter. Those pollutants in the dissolved phase, however, must be extracted from the water utilizing a substance that attracts the pollutants, and subsequently binds to the target analytes.

Axys Environmental Services of Sidney, British Columbia, Canada, designed the sampling unit used by ORSANCO to collect high volume water samples. Water samples are collected utilizing a positive displacement pump that draws river water up through an intake line and pushes the water through the sampling unit. The intake line is made of Teflon, and is set at a depth several feet below the surface of the water. A fine mesh screen is placed at the end of the intake to prevent very large debris from clogging the line. The water drawn into the unit, first passes through an inline filter that removes particles greater than 140 m m. This is necessary to avoid large particles from damaging the pump head. Considering dioxins prefer to bind to small particulate matter, the amount of dioxins sorbed to these large particles is believed to be negligible.

The water then is pushed through a four-inch glass fiber filter (GFF) that effectively removes all particles greater than one micron. The GFF is seated in a stainless steel canister with a pressure gauge mounted to the bottom. As the filter begins to clog with particulates, the pressure gradually increases. Once the pressure reaches 15 pounds per square in (psi), the filter is clogged and a new filter must be inserted. The high volume water sampling unit is designed with two GFF canisters arranged in parallel paths. Valves on each side of the canisters permit the water to be directed through one canister at a time. This feature allows the filters to be changed quickly with minimal disruption of the sampling. To change the GFF, the operator simply turns the pump off, switches the valves to direct the water through the other canister containing a GFF, and then immediately resumes pumping. The used filter is then removed from the filter holder, and is wrapped in foil, labeled, placed in a sample container, and promptly placed on ice.

After passing through the GFF, the filtered water runs through Amberlite XAD-2 resin columns. The number and size of the resin columns can vary depending on the users needs. The sampling unit used by ORSANCO was designed to accommodate the use of two 75 gram (g) Teflon columns arranged in parallel paths. Thus by pushing the water through two columns simultaneously, the time to filter the desired volume of water is cut in half.

The resin columns dictate the rate at which water can be pushed through the system. Adequate contact time is necessary to allow for efficient extraction of the dioxins and other organic pollutants from the water. The XAD-2 resin is a cross-linked polymer of styrene and divinylbenzene. The extremely hydrophobic nature of the resin attracts other hydrophobic organic compounds such as dioxins. The dioxins prefer to attach to the resin rather than to remain dissolved in the water. Thus, the resin effectively extracts hydrophobic organics out of the water phase and into the solid phase (Axys 1991).

The filtered and extracted water exits the resin columns, and passes through a volume totalizer and flow meter. A digital display unit allows the user to monitor the total volume filtered, as well as, the rate at which the water is passing through the system. When using two 75 g resin columns, 1.6 liters per minute (l/min) is the maximum flow rate at which the resin columns can efficiently extract the dioxins from the water.

High volume water sampling offers several advantages over the use of conventional sampling methods. The most significant benefit of the high volume water sampling method is the ability to attain extremely low detection limits by concentrating pollutants from very large samples. The method currently provides the only practical means of detecting dioxins at concentrations below EPA’s water quality standard of 0.013 ppq. A very large volume water sample (e.g. 1000 liters) could be collected using conventional sampling methods and concentrated in a lab to provide comparable detection limits. This type of sampling, however, would be impractical due to the difficulty and costs associated with the transport of the sample to the lab. Also the liquid/liquid extraction would require a tremendous amount of solvent, thus creating the possibility of significant solvent blank contamination. The solvent impurities may interfere with the analysis, and thereby reducing the accuracy of the analytical results (Axys 1991).

Another useful feature of the high volume water sampling method is that the unit can be programmed to operate on a timer. When sampling in a protected area, such that the sampling equipment could be left unattended, the control unit can be programmed to turn on and off at designated times. This provides a means of compositing samples over long periods of time, while minimizing personnel time in the field. Also, the high volume water sampling unit can be used for monitoring a variety of pollutants. The XAD-2 resin columns effectively remove hydrophobic organics such as dioxins, chlorinated pesticides, and PCBs. Metals including copper, iron, lead, cobalt, cadmium, nickel, manganese, and zinc can be monitored using a trace metal resin column (Axys 1991). Other types of resins for other parameters will be developed in the future.

There are some disadvantages to using the high volume water sampling method. The equipment needed to collect samples of this nature is rather expensive. The high volume water sampling unit uses only Teflon and high grade stainless steel for all parts coming in contact with the sample water. The resin columns, which are reusable, are also either made of Teflon or stainless steel. The columns must also be repacked with clean resin after each use, which can cost several hundred dollars each time. The initial equipment setup for an intensive high volume water sampling program can cost well over $20,000. Also, depending on the sampling locations, field sampling can prove to be resource intensive. When sampling in protected areas where the equipment can be safely setup and left unattended, sampling cost are minimized. Sampling a large river from a boat, however, requires constant supervision of the sampling unit.

In 1997, ORSANCO conducted high volume water sampling at four locations (see Figure 1). The three sites located on the Ohio River were sampled three times each, while five sampling events were conducted at the Kanawha River site. Special emphasis was placed on sampling the Kanawha River due to the presence of several potential and confirmed sources of dioxin located upstream. Each full round of sampling entailed sampling one site per day for four consecutive days. One thousand liters of water were drawn through the sampling unit at each location over a 12-hour period . Sampling was conducted from a boat that was anchored just outside of the navigational channel. A second boat was used to collect total suspended solids (TSS) samples throughout the sampling period.

The filters and resin columns were analyzed separately thus providing results for the particulate and dissolved phases individually. It should be noted that some portion of the dioxins captured by the resin columns may be attributed to dioxins bound to particles less than one micron in size that pass through the glass fiber filter. As expected, considering dioxin congeners have a strong affinity for particulate matter, the concentrations in the particulate phase were considerably higher than that of the dissolved phase. Three of the fourteen high volume water samples had undetectable levels of 2,3,7,8 TCDD in the dissolved phase with detection limits below 0.001 ppq. Dissolved concentrations for 2,3,7,8 TCDD ranged from less than 0.001 ppq to 0.020 ppq. TEQ dioxin concentrations in the dissolved phase ranged from 0.0073 ppq on the Ohio River to 0.242 ppq on the Kanawha River. Particulate concentrations were consistently an order of magnitude greater than dissolved for 2,3,7,8 TCDD and TEQ dioxin concentrations. Total 2,3,7,8 TCDD concentrations (particulate phase plus dissolved phase) exceeded the water quality standard in 12 of the 14 samples (see Figure 2), while total TEQ concentrations ranged from 0.133 ppq to 0.906 ppq (see Figure 3).

Having successfully completed high volume water sampling at four locations in 1997, ORSANCO has expanded its 1998 field sampling plan to include sampling at 14 locations. Each location will be sampled three times, preferably under different river flow conditions. Upon completion of this sampling program, concentrations of dioxins in the upper three hundred miles of the Ohio River will be characterized. Also, contributions of dioxins from potential sources on the Kanawha River will be quantified. These sources will be bracketed with a sampling location upstream and downstream of the source.

To reduce the length of the sampling period, larger XAD-2 resin columns will be used for 1998’s sampling activities. These newly designed columns are made of stainless steel, and hold 250 grams of resin. The larger columns have a maximum flow rate of 2.2 l/min, and will effectively reduce sampling times from nearly 12 hours using two 75 gram Teflon columns, to approximately eight hours using one 250 gram column.

High volume water sampling has proven to be an effective means of directly measuring in-stream concentrations of dioxins at levels below EPA’s water quality standard of 0.013 ppq. The sampling method allows for concentrations in both the particulate and the dissolved phases to be measured independently. There are a variety of applications of the sampling technique including monitoring drinking water supplies and effluent discharges from industries and municipalities. The development of this technique provides a method by which sources could be positively identified, and their pollutant contributions quantified. The versatility of the high volume water sampling method to be used for a variety of pollutants other than dioxin, including PCBs, chlorinated pesticides, and several metals, makes it an extremely valuable technique that should continue to be refined and expanded upon.