J. Rossabi, B. B. Looney, C. A. Eddy-Dilek, B. D. Riha, and D. G. Jackson
Westinghouse Savannah River Company
Aiken, SC 29808

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

This report has been reproduced directly from the best available copy.

Available for sale to the public, in paper, from:  U.S. Department of Commerce, National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161,  phone: (800) 553-6847,  fax: (703) 605-6900,  email:  orders@ntis.fedworld.gov   online ordering:  http://www.ntis.gov/support/ordering.htm

Available electronically at  http://www.osti.gov/bridge/

Available for a processing fee to U.S. Department of Energy and its contractors, in paper, from: U.S. Department of Energy, Office of Scientific and Technical Information, P.O. Box 62, Oak Ridge, TN 37831-0062,  phone: (865 ) 576-8401,  fax: (865) 576-5728,  email:  reports@adonis.osti.gov

In natural subsurface systems dominated by heterogeneity, the delineation and even the detection of sparingly soluble, dense contaminants can be extremely difficult. The performance assessment of cleanup at these sites is therefore more complex. Several technologies for the characterization of sites contaminated with dense non aqueous phase liquids (DNAPLs) have recently been developed. These include geophysical techniques, tracer tests, and direct sampling or sensing methods. The innovative methods provide some significant advances over conventional sampling-based approaches but the real value of these methods is in their addition to a toolbox approach to DNAPL characterization. The toolbox approach recognizes that all characterization methods contribute to the conceptual model of the site. The strategic selection of technology and results from each application must contribute to the evolution of this conceptual model. The ultimate goal is the absolute knowledge of the contamination at the site. This, of course, can never be reached but using the right suite of tools and comprehensive integration of the data, the most accurate understanding is obtained.

The evolving conceptual model and toolbox approach is simply a framework for the ideal understanding of a site. It can be used for the characterization of any site, not just DNAPL suspect sites. The principles involved are generally just good scientific and engineering practice. Unfortunately these principles are often overlooked in the face of complex systems or external pressures. Generally characterization methods should progress from less invasive to more invasive methods and from less expensive to more expensive. Simple, long-term observations should also not be overlooked (e.g. water level fluctuations in existing wells) and can often provide inexpensive and important keys to understanding a site. Finally it is useful to borrow from the experience of the geochemical exploration professionals – lots of inexpensive measurements (with slightly larger error bands) and fewer expensive (usually due to excessive quality assurance requirements) measurements. Of course, these are general guidelines and should not prevent the use of a very expensive but very useful method at the beginning of an investigation if it is justified for a particular site.

Despite the complexity of the situation or the encroaching time and economic pressures, the current conceptual model of the site must be logical, explainable, and defensible. A useful intellectual tool to isolate pressures that may cloud prudent decisions is to personalize the perspective - what would I do if this were my property? This mode of thinking maximizes the responsibility of the investigator because they are now thinking of a site in reference to scenarios where the protection of their family and their interests is involved. It also maximizes the attention to practical details such as cost and long term goals (e.g., remediation method and extent) for the site. The available technology is usually important in selecting characterization methods but the technology must be selected in the context of the remediation goals and requirements. If your planned remediation is vapor extraction you do not need to know the temperature gradient in the soil, at least initially.

DNAPL contaminated sites tend to be more complex than light non aqueous phase liquid (LNAPL) or aqueous contaminated sites because the physical and chemical characteristics of dense, sparingly soluble contaminants add additional complexity to heterogeneous geology and hydrogeology at most sites. While LNAPL contamination is usually constrained to the top of the water table and above, and aqueous phase contamination follows the hydrology of the site, DNAPL movement is controlled by gravity and capillary pressure of sediments, and can move against the hydraulic gradient.

The conceptual model is the snapshot or realization of the site at a particular time. It is optimally constructed through historical process information, the complete set of physical, chemical, geologic, and hydrogeologic data and understanding of the dynamic processes, experiential knowledge of similar sites, the risk scenarios related to the site and its contaminants, and the type and scope of remediation envisioned. The model is updated and refined by subsequent data and information and the process is iterative.

At the Savannah River Site (SRS), we have used disposal records (amounts, types, timing), construction diagrams of drains, sewers, and basins (e.g., unlined manholes every 50 yards), aerial photos regional geologic and weather information, and we talked to the personnel who worked in the operations and were responsible for it. All of this information is not available at every site but much more is available then is generally used. A very important resource is the verbal record of site personnel. This information may not always be precise but it can be used with other information or to open an investigation path. Often the best information is obtained by allowing the personnel to explain what they did, saw, smelled, and tasted in their own narrative account with few directed questions that may influence the testimony. Understanding the daily operations at a site is also critical. How often were degreasing units used? How often were solvents purchased and in what quantities? How much process water was released per day? How often were machines serviced and what happened to the used lubricating oils? Were acids and caustics also released? Each of these questions can dramatically impact the deposition of contaminants at a site.

At SRS we have had the pleasure of hosting many different researchers and vendors with innovative technologies. We have found them to be nearly universally sincere in their desire to provide real solutions rather than carelessly promote themselves or their technology. One thing that we have experienced rarely, though often enough to mention, is the visitor to the site who believes he understands the host’s problems better than the host does. A visitor may have some useful insight and ideas based on their experience in similar situations but the host knows his site best. We have learned this lesson and apply it when we visit other sites to help with their environmental problems.

DNAPL sites are rarely sought, they are generally stumbled upon based on unusually high concentrations found in the analytical results from a nearby aqueous or gas sample. The first suspicion of DNAPL arises from rules of thumb (e.g., 1 or 10% of aqueous solubility or vapor pressure) applied to these samples that are contaminated via dissolution and/or diffusion from a source. The value of the data from these routine samples should not be dismissed. Besides implying the presence of DNAPL, the diffusion-based samples can help focus more direct searches for DNAPL by concentration differences between aqueous or gas phase samples.

Baseline and modified baseline soil sampling is still the most reliable method for directly detecting DNAPL (Eddy Dilek et al., 1999). Using equilibrium partitioning rules, the presence of DNAPL in a soil sample can be determined (Cohen and Mercer, 1992). Some important modifications that should be incorporated include higher depth resolution in soil sampling. Because residual DNAPL at sites is often in small, dispersed blobs correlated to the soil type, the ideal depth resolution of sampling should be comparable to the resolution of the geologic heterogeneity. Traditional sampling generally prescribes a regular grid with samples collected at intervals of 5 feet at best. We have found that collecting samples every 1 to 2 feet and at lithologic changes has been significantly better than rigid 5-foot intervals. The cost of analysis for these samples can be defrayed by using less costly analytical techniques such as field screening methods or even EPA SW 846 methods (like heated headspace method 5021) without rigidly applying regulatory protocol or incurring the EPA Certified Lab Program paperwork. Of course, the local regulators should be consulted when pursuing this strategy. When continuously coring a drilled hole, collecting additional samples is very inexpensive. If the cost of analysis is prohibitive, a subset of the samples need only be analyzed. The rest can be cold stored for analysis later if desired. The literature has shown that VOC samples are stable for longer than the regulatory limit of 14 days, especially when the samples contain very high concentrations or DNAPL(West et al., 1996). If possible, aqueous samples should also be taken at higher vertical resolution by using depth-discrete samplers or setting wells with smaller screen zones. Heterogeneity and the ability to access and analyze at the scale of the heterogeneity limit subsurface characterization. The most common access techniques (drilling and direct penetration) have inherent lateral limitations but fewer vertical limitations. We should take better advantage of the accessibility to finer vertical resolution.

Despite the availability of more sophisticated methods of characterization, the ultimate performance assessment in the remediation of a site relies on the continued clean analytical results from down-gradient wells.

Earlier we recommended the general strategy of progressing from non- or minimally invasive to more invasive. The most obvious choice for non-invasive methods beyond evaluating historical records and using existing infrastructure are the geophysical techniques. Acoustic and electromagnetic probing of the subsurface have been used successfully to detect reservoirs of petroleum or natural gas, or on a smaller scale, buried drums and other anomalous objects. They are also useful for detecting changes at a site through time (Daily et al., 1992). They probe volumes rather then measuring at discrete points, which can be an advantage in characterizing a site with small, dispersed pockets of contaminants. An important limitation of these methods, however, is that the resolving power is generally too low to detect the disparate blobs of DNAPL due to the inherent noise caused by subsurface heterogeneity. If pores were predominately NAPL filled and were contiguous over a large enough volume (usually on the order of a cubic meter), some geophysical techniques could directly detect DNAPL. A far more important use of the geophysical measurements is to characterize and differentiate the geologic units that would be likely to harbor DNAPL or control its movement. Like the lower resolution, diffusion based measurements, geophysics is very important for telling you where to look. The information provided from geophysics has even greater value when combined with other data from the conceptual model. At SRS, surface seismic and electromagnetic measurements were integrated to laterally map confining zones and determine their thickness.

In addition to surface geophysics, existing boreholes can often be used to enhance the resolution of surface techniques. While evaluating the conductivity logs at a known DNAPL site, anomalous patterns were evident in some of the wells near the release point (Nelson, 1996). Large quantities of caustics and acids were disposed with process water as a dense aqueous phase liquid (DAPL) at the same location as the solvents. The caustic solution presented a sharp electrical conductivity contrast to the normally low conductivity formation water at the site. The conductivity anomalies coincided with higher concentrations of organics indicating that the DAPL may have traveled in the same path as the DNAPL until dilution of the caustic solution rendered it neutrally buoyant. These boreholes may also be suitable for long term measurements of the subsurface. Simple equipment can be used monitor water table fluctuations or chemical concentrations through time using the existing infrastructure. The time concentration profile may provide indications of the location of DNAPL (Rossabi, 1999).

Direct push methods are invasive, however they are less disruptive than conventional drilling. Some advantages of direct push methods such as the cone penetrometer are rapid inexpensive access to the subsurface, minimal investigation derived waste, the ability to mount multiple sensors on the access platform and the ability to collect very high-resolution data with depth. These methods are becoming the method of choice at sites where the use of direct push is possible. The principal disadvantages are they can only be used in unconsolidated materials and they are ultimately limited in depth. The cone penetrometer truck has a standard suite of sensors (tip stress, sleeve friction, pore pressure, and electrical conductivity) which can provide a real time indication of soil type with depth at centimeter scale resolution. At SRS we’ve used these basic tools to produce a map of the topography of an important confining unit that constrains DNAPL movement. We have also combined the soil type sensors with tools to collect depth discrete gas, water, and soil samples. These more targeted samples help focus DNAPL investigations.

Some of the most promising tools for DNAPL characterization have been incorporated with the cone penetrometer test (CPT). Optical probing through a hardened window in the CPT push rod has been used to detect DNAPL either directly by Raman spectroscopy (Rossabi et al., 2000) or inferentially using induced fluorescence measurements of co-constituents dissolved in the DNAPL (Kram, 1998; Rossabi and Nave, 1998). Visual detection of DNAPL has also been accomplished with a high resolution video microscope mounted in a CPT probe (Lieberman and Knowles, 1998). Another important tool developed for a Geoprobeä system and modified for cone penetrometer use is the Membrane Interface Probeä (MIP). Using a selectively permeable membrane permitting entry of only volatile organic compounds (VOCs), the tool provides a continuous profile of VOCs through depth (Christy, 1998) Finally, the Ribbon NAPL Sampler (RNS) by FLUTe can be used in conjunction with direct push or drilling methods. It consists of a hydrophobic sorbent liner with an impregnated indicator dye that is deployed in a borehole directly contacting the formation (Riha et al., 2000). DNAPL is wicked into the liner and causes a color change in the dye. The liner is retrieved from the borehole and colored spots and their depths are logged as DNAPL detect locations. This has been the most consistently robust DNAPL characterization technique that we have used to date.

Chemicals that strongly partition into an organic phase may be used with conservative tracers to determine the presence of NAPLs and, in some cases, the NAPL saturation (Jin et al., 1995). During this test, partitioning and conservative tracers are injected into the formation though a well and recovered in a different well(s). By analyzing the center of mass arrival times of the conservative and reactive tracers, the amount of residual NAPL may be estimated because the retardation of the reactive tracer is attributed to partitioning to the NAPL. This method has been successfully used at many NAPL contaminated sites and has been particularly useful in comparing the performance of clean up methods by pre and post testing. The advantages of this method include the ability to probe a large volume, relatively simple formulas for interpretation, and that they can be used both above and below the water table. The most important issue with this technology is that the tracers can only infer NAPL that is in the tracer flowpath. This is particularly critical in the vadose zone where organic contaminants from old releases tend to reside in the fine grain soils not easily accessed by tracer gases which prefer to flow through more permeable zones. Other issues with this technology include permitting and waste handling of the generally large volume of tracers required and limiting the error in the empirically determined, partitioning coefficients of the tracers into the NAPL. A recently developed complementary technology that has increased the utility of this method is interfacial area partitioning tracer tests (Kim et al., 1997). This helps determine the surface area of NAPL available for the NAPL partitioning tracer.

Another test using a solubilizing fluid injected through a cone penetrometer rod was conducted at a DNAPL site on the Cape Canaveral Air Station (CCAS). In this test, a small volume (less than 4 liters) of potable water is injected and recovered at a specific depth through a cone penetrometer system to determine hydraulic parameters and baseline concentrations. A similar volume of an alcohol solution or other NAPL solubilizing fluid is then injected and recovered at the same location. The concentrations of the target organic contaminant during the two tests are then compared. If NAPL is encountered, the recovered solubilizing fluid will have significantly greater concentrations of the contaminant than the water recovery. If only aqueous contaminant is present both tests will have similar concentrations in the recovered samples. During the CCAS test, higher concentrations of TCE were measured in the recovered fluid after alcohol injection than in the fluid recovered following water injection, suggesting TCE DNAPL in the test zone of influence. In addition, concentrations of cis dichloroethylene (existing only in the aqueous phase as a byproduct of reductive dechlorination of TCE) remained the same following both injections. This test takes advantage of the high resolution geologic data that the cone penetrometer provides to focus the investigation and minimize waste. Unfortunately this also results in a smaller volume probed and the test has similar NAPL flowpath contact issues as the partitioning interwell tracer test (PITTä ).

Each technology has advantages and disadvantages in its application. Geophysical techniques image a volume and are noninvasive but may lack the resolution to directly characterize a DNAPL-contaminated site. These methods are quite useful in mapping subsurface units that may control the movement of DNAPL. Partitioning or solubilizing tracer tests also probe a large subsurface volume and can detect small pockets of separate phase contaminants but the contaminants must be in the flowpath of the tracers. There may also be permitting and waste issues associated with these methods. Direct sampling or sensing, particularly when applied with direct penetration tools, can offer positive identification of DNAPL at very high vertical resolution but low lateral resolution because the methods do not probe beyond the radius of the borehole. They also require additional boreholes, which increases the cost of characterization.

Ideally, many techniques would be used in a characterization effort because of the complementary information they provide but a limited budget often precludes that choice. The unique features of a specific site will dictate and narrow the list of appropriate tools and the cost of the technologies will further constrain the selection. Finally the order of technology application is important, generally progressing from low cost, less invasive techniques to more specific technologies. By selecting the right combination of technologies, both conventional and innovative, and using these in the frame of an evolving conceptual model of the site, the most effective DNAPL characterization can be obtained.

1. Cohen, R. M., and J. W. Mercer. 1993. DNAPL Site Evaluation, C. K. Smoley-CRC Press, Boca Raton, Florida.
2. Christy, T. 1998. "A Permeable Membrane Sensor for the Detection of Volatile Compounds in Soil", in Proceedings of the Symposium on the Application of Geophysics to Engineering and Environmental Problems, Chicago, IL.
3. Daily, W., A. Ramirez, D. LaBrecque, and J. Nitao. 1992. "Electrical Resistivity Tomography of Vadose Water Movement." Water Resour. Res. 28 (5): 1429-1442.
4. Eddy-Dilek, C. A., B. D. Riha, D. Jackson, and J. Rossabi. 1999. DNAPL Source Zone Characterization of Launch Complex 34, Cape Canaveral Air Station, Florida. WSRC-TR-99-00024, Westinghouse Savannah River Co., Aiken, SC 29808.
5. Jin, M., M. Delshad, V. Dwarakanath, D. C. McKinney, G. A. Pope, K. Sepehrnoori, C. Tilburg, and R. E. Jackson. 1995. "Partitioning tracer test for detection, estimation and remediation performance assessment of subsurface nonaqueous phase liquids." Water Resour. Res. 31(5):1201-1211.
6. Kim, H., P. S. C. Rao, and M. D. Annable. 1997. "Determination of effective air-water interfacial area in partially saturated porous media using surfactant adsorption." Water Resour. Res. 33(12):2705-2711.
7. Kram, Mark. 1998. "Use of SCAPS petroleum hydrocarbon sensor technology for real-time indirect DNAPL detection.", Journal of Soil Contamination, 17(1), 73-86, 1998.
8. Lieberman, S.H., and D.S. Knowles. 1998. "Cone penetrometer deployable in situ video microscope for characterizing sub-surface soil properties." Field Analytical Chemistry and Technology 2(2):127-132.
9. Nelson, P.H. and J.E. Kibler. 1996. Geophysical Logs and Groundwater Geochemistry in the A/M Area, Final Report, Savannah River Site, South Carolina. USGS Open-File Report 96-75.
10. Riha, B. D., J. Rossabi, C. A. Eddy-Dilek, and D. Jackson. 2000. "DNAPL Characterization Using the Ribbon NAPL Sampler: Methods and Results", in Proceedings of the Second International Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, May 22-25, 2000.
11. Rossabi, J. and S. E. Nave. 1998. Characterization of DNAPL using Fluorescence Techniques (U), WSRC-TR-98-00125, Westinghouse Savannah River Company, Aiken, SC 29808.
12. Rossabi, J. 1999. The influence of atmospheric pressure variations on subsurface soilgas and the implications for environmental characterization and remediation, Ph.D. thesis, Clemson University, University of Michigan Press.
13. Rossabi, J., B. D. Riha, J. Haas, C. A. Eddy-Dilek, A. Lustig, M. Carrabba, W. K. Hyde, and J. Bello. 2000. "Field Tests of a DNAPL Characterization System Using Cone Penetrometer-based Raman Spectroscopy." Ground Water Monitoring and Remediation, in press.
14. West, O.R., D. W. Bottrell, C. K. Bayne, R. L. Siegrist, and W. H. Holden. 1996. "A Scientific Basis for Revising Regulatory Holding Times for VOC Water Samples." Am. Envrion. Lab., pp. 15-16.