Andrew I. Strutynsky, P.E., STRATIGRAPHICS, Inc., Glen Ellyn, Illinois Timothy J. Sainey, P.G., ERM-Midwest, Inc., Columbus, Ohio
ABSTRACT: Piezometric Cone Penetration Testing (CPTU) and penetrometer groundwater sampling were used in locating a volatile organic contaminant plume at an industrial site in southern Ohio. Nine CPTU tests (soundings) were performed to determine site hydrostratigraphy in real-time. On-site chemical analysis of penetrometer groundwater samples provided near real-time detection of contaminants. These results were used to define subsequent exploration points. Using this investigation approach, drilling operations to set monitor wells began as penetrometer exploration ended. Program quality increased, while exploration costs decreased by using this combination penetrometer-drilling rig approach.
Tricholoroethene (TCE), an industrial solvent, along with other volatile organic compounds, was identified in the groundwater at a manufacturing site in rural southern Ohio. Contamination was initially found in four monitor wells, located around a suspected source. The immediate soils consist of mixed sand, silt, and clay, overlying granular strata, which overlie bedrock. The water table is located at EL 550 ft, or 8 to 20 ft below the surface.
To define the limits of contamination, a soil gas survey was performed using a hand-held hammer drill to drive probes 5 ft into the ground. Probing traced contamination to an off-site agricultural field and defined the general direction of contaminant movement (Figure 1). As the survey expanded, indications of volatiles abruptly stopped. It was suspected that a thicker clay unit was masking volatiles that might be present in the groundwater. The limits of contamination still had to be determined, but additional exploration was subject to the following constraints:
The needs of the off-site property owner had to be met while still allowing exploration to proceed. His concerns were addressed by negotiating permission to set a maximum of nine wells to monitor the extent of contamination, with assurances of minimal surficial disturbance to future crop cultivation.
A possible solution to the constraints of the agreement was to use a drilling rig to split-spoon sample soils, and to sample groundwater by using either a Hydropunch sampler or wellpoints. The exploration boreholes would be drilled and plugged until the extent of contamination had been determined. At this point, the monitor wells would be set. However, the cost of drilling and plugging numerous exploratory boreholes, and the overall ineffectiveness of drilling rigs to advance the Hydropunch in dense sands and gravels, made the use of a drill rig less than optimal.
Another solution was to use a cone penetrometer rig to explore subsurface conditions followed by conventional drilling to set monitor wells. This plan had the following advantages:
Use of the penetrometer exploration/drill rig well installation approach was decided upon as the most cost- and performance-effective solution for this project An analytical laboratory was set up on-site allowing for successive exploration points to be optimally chosen on the basis of contaminant plume detection, rather than on an arbitrary grid pattern.
PENETROMETER TECHNIQUE AND EQUIPMENT
Penetrometer methods are being used with increasing frequency, and they provide significant advantages for geo-environmental exploration. The technique uses a large hydraulic ram to push small (1.5 to 2.5 inch) diameter probes into the ground without drilling a borehole (Figure 2). Instrumented probes, called penetrometers, provide semi-direct and direct information on geotechnical, hydrogeological, and geochemical site conditions. Penetrometer samplers are used to obtain physical samples of subsurface materials.
Penetrometer methods are used to their greatest advantage in sand, silt and clay deposits. Penetrometer profiling (Figure 3) is continuous and accurate, and measurements (sounding logs) are reliably interpreted for definition of aquifers and confining layers. Lateral continuity of layers is readily apparent from a series of adjacent penetrometer soundings.
Site disturbance and waste material disposal is minimized because no cuttings or drilling fluids are generated during penetrometer operations. Downhole equipment is steam cleaned during retrieval. Personnel exposure to contaminants is much less than exposures during drilling. Heaving sands pose little problem, as no open hole actually exists during penetrometer advance.
Three hundred to 900 ft of geotechnical soundings can be performed in a day, depending on access and project requirements. The special demands of geo-environmental investigations decrease daily footage, but productivity is still significantly higher than that using drilling rigs. Sounding depths in excess of 100 ft can be achieved, depending on site stratigraphy.
The penetrometer is mounted at the end of a series of heavy-wall sounding rods. A hydraulic ram is used to push the penetrometer into the ground at a rate of 4 ft per minute. Electronic signals from downhole sensors are transmitted by a cable, strung through the sounding rods, to a computer inside the penetrometer rig. Data are recorded at depth intervals of 3/8 to 3/4 inch. A real-time data display is monitored for evaluation of test performance and for immediate definition of site conditions. At the end of a sounding, the penetrometer and sounding rods are retrieved and decontaminated. Open hole can be grouted where cross-contamination between layers may occur. Open hole was allowed to collapse at the end of each sounding for this project, due to the lack of confining layers at the site.
A special truck is used to house, transport, and deploy the penetrometer. Twenty tons of truck weight and ballast are used to counteract the thrust of the hydraulic ram. The work area is enclosed and includes heating and air conditioning. Computers, penetrometers, samplers, electrical power, lighting, compressed air, grout and water pumps, steam cleaner for equipment decontamination, 275 gal water tank, tools and spare parts are all included within the penetrometer truck, providing for self-contained operations.
The basic electronic penetrometer consists of two separate soil shear resistance sensors, and is used to acquire information on soil strength and stratigraphy. Tests conducted with this penetrometer are called Cone Penetration Tests (CPT) and are specified by ASTM Standard D3441. Electronic CPT has been used for geotechnical engineering projects for more than 25 years, while less sophisticated, uninstrumented versions of the test (Dutch cone test) have been used since the 1930s.
Two laboratory grade, strain gage loadcells, mounted inside the penetrometer, are used to measure the soil shear resistance to penetration acting on the conical tip and along the cylindrical sides of the CPT penetrometer. CPT measurements are continuous, accurate and repeatable. The tip or cone end bearing resistance can respond to soil seams as thin as 2 to 4 inches. The side or friction resistance measurement has an isolation of about 6 inches.
A pressure transducer is added to the CPT penetrometer to additionally measure the soil pore water pressure response to penetration; this is called a Piezometric Cone Penetration Test (CPTU). CPTU piezometric data allow for the evaluation of soil saturation, water tables, potentiometric surfaces, and soil horizontal permeability in both aquifers and aquitards. The CPTU piezometric measurement has a resolution of about 1 inch.
Another penetrometer has been used to measure the shear resistance, piezometric response, electrical conductivity and temperature of penetrated soils. This penetrometer is useful in detecting free hydrocarbon product on groundwater. This test type is termed CPTU-EC. Additional details on penetrometer instrumentation can be found in Strutynsky, et al., 1985 and 1991.
The Hydropunch and BAT penetrometer samplers were used to sample groundwater for this project. The Hydropunch sampler (Figure 4) consists of a stainless steel, shielded wellpoint and sample barrel assembly (Edge and Cordry, 1989); it is deployed using the heavy-wall penetrometer sounding rod. The shield prevents contamination of the sampler while pushing. When the shield is retracted, groundwater flows under in situ pressure conditions into the 500 ml sample barrel.
A water level indicator can be lowered to the top of the Hydropunch in order to determine sampler filling. The sample is isolated in the barrel by two ball check valves. The entire sampler is retrieved to the surface to pour off the sample and for sampler decontamination. The tip of the Hydropunch sampler must be at least 4 ft below the water table to acquire a sample.
The BAT sampling system (Figure 5) consists of a wellpoint that is internally sealed with a septum (Torstensson, 1984). After pushing the wellpoint to depth, an evacuated 35 or 70 ml vial, also sealed with a septum, is wirelined down the casing. A double-ended hypodermic needle, mounted in an adapter below the vial, pierces both the wellpoint and the sample vial septa, and allows fluids to flow into the vial. The septa seal as the sample vial is retrieved to the surface, maintaining the sample at in situ pressure conditions. This procedure may be speared to develop the wellpoint and to obtain increased sample volumes.
The BAT Enviroprobe and Mk. 2 wellpoints were used during this study. The Enviroprobe is shielded, while the Mk. 2 tip is unshielded. The Mk. 2 filter is exposed to possible cross-contamination during deployment. A thin-wall AQ drill rod is used to deploy the BAT samplers, as the heavy-wall penetrometer sounding rods are too small to pass the BAT sample vial.
PENETROMETER DATA INTERPRETATION
Correlations between penetrometer data and soil type have been developed from observational criteria on adjacent CPT soundings and drilled and sampled boreholes (Douglas and Olsen, 1981). CPT soil classifications reflect the shear response of a soil to penetration. Soil shear response is not entirely controlled by grain size distribution. However, soil types evaluated from CPT data generally agree with those classifications based on soil grain size distribution methods, such as the Unified Soil Classification System (USCS).
The CPT cone end bearing resistance increases exponentially with increases in grain size. The cone end bearing resistance in dense sands ranges from about 150 to 300 tons per square foot (TSF), while the cone end bearing resistance in a stiff clay ranges from about 7 to 15 TSF. The proportion of CPT friction to cone end bearing resistance, termed friction ratio, is related to the fines content of a soil. The friction ratio is low in sands and high in clays. CPT measurements are computer analyzed using CPT soil classification charts (Figure 6) to quickly define site stratigraphy.
FIGURE 4 - HYDROPUNCH GROUNDWATER SAMPLER
FIGURE 5 - BAT GROUNDWATER SAMPLING SYSTEM
FIGURE 6 - CORRELATION CHART FOR CPT SOIL TYPES
The penetration of a saturated soil generates a localized pore water pressure field, in excess of equilibrium, around the penetrating probe. This generated pressure field dissipates quickly in soils of high permeability, so only the equilibrium pore water pressure field is measured in clean sands and gravels during Piezometric Cone Penetration Testing (CPTU). In low permeability soils, excess pore pressures require a significant amount of time to dissipate (Saines, et al., 1989). The dissipation of excess pore water pressure can be recorded as a function of time by pausing in the penetration process; this is termed a CPTU dissipation test. If the pauses are sufficiently long for all excess pressures to dissipate, measurements can be obtained of equilibrium potentiometric surfaces at multiple depths.
A CPTU dissipation test is somewhat similar to a falling head slug test, and can bc used to calculate a value of soil horizontal permeability. However, tens to hundreds of feet of excess pore water pressure are induced in low permeability soils by penetrometer advance, as compared to several feet of head induced in a well during slug testing. The greater pressure changes during CPTU dissipation testing require soil compressibility effects to be included in analyses. The CPT cone end bearing resistance provides an index of compressibility for permeability computations.
A Hewlett-Packard HP 589OA gas chromatograph (GC) with two electron capture detectors (ECD) was selected as the on-site analytical instrument for this project. It has the advantage of:
For the analysis of volatile organics, U.S. EPA SW-846 Method 3810 Headspace was used. Calibration procedures were performed using standard mixtures to establish the internal standard curve. Sample analysis consisted of pouring collected water samples into a 1.0 ml syringe, and injecting this sample into a nitrogen purged 40 ml septum vial. The sample was then heated in a 70 deg C water bath for 15 minutes; 100 ml of headspace was drawn from the center of the septum vial and injected into the GC for analysis.
For quality control (QC), every 10th unknown was run, per a modified U.S. EPA Method 3810. Modifications are to allow samples to equilibrate to 70 deg C, to use 40 ml vials with Teflon face septa, and to inject 100 ml of headspace gas. For each 10th unknown, four samples were injected to determine the quality of data. These consisted of a:
In addition to the field analyses, select duplicate samples were sent to an off-site analytical laboratory for verification. Sample preparation, standards and QC were assisted by an analytical subcontractor.
CPTU soundings were performed at nine locations (Figure 7) during the period of February 6 to 13, 1990, for a total of 699 ft of test. Only partial days were worked on February 6 and 9, and no field work was done on February 10 when bulldozer support for access to locations in the wet, recently plowed field was not available. Twenty seven groundwater sampling attempts, with 22 successfully recovered samples, were also performed during this period. The sequence of penetrometer operations was as follows:
The CPTU soundings revealed site soil conditions to have general lateral continuity, as can be seen on a stratigraphic cross-section (Figure 9). Site stratigraphy is summarized in Table l.
FIGURE 7 - LOCATION OF CPT TESTS,
FIGURE 8 - INTERPRETED CPT SOUNDING LOG C-3 WITH SAMPLING DEPTHS
FIGURE 9 - STRATIGRAPHIC CROSS-SECTION DEVELOPED FROM CPT DATA
A highly detailed characterization of the site geology was obtained from the CPTU sounding logs. Split-spoon sampling during drilled investigations often results in poor recovery in sands and gravels, which make detailed description impossible. In contrast, the CPTU data provided a continuous record of the material being penetrated, thus allowing even minor changes to be recorded accurately (Figures 3, 8 and 9). The high resolution of the CPTU data was important in finding zones of higher permeability for future groundwater remediation and also verified the lack of confining layers at the site. The assumption of a thick, surficial clay zone interfering with the soil gas survey was verified by soundings C-3, C-7 and C-8 (Figures 3, 8 and 9).
Groundwater conditions were measured using CPTU piezometric data (Figures 8 and 9). Water table measurements were also taken using a water level indicator lowered into the open hole left after a sounding. The potentiometric surface at one groundwater sampling depth was measured by allowing the sounding rod string to fill through the Hydropunch sampler.
The Hydropunch sampler was successfully used in obtaining 22 groundwater samples out of 27 attempts. In general, the Hydropunch sampler worked well, especially in regards to depth capacity. Some disadvantages of the Hydropunch sampler were:
The BAT groundwater samplers were used during two sampling attempts at the site. The thin-wall deployment casing buckled while pushing the Mk. 2 wellpoint through the shallow gravelly sands at location C-2, which precluded taking a sample. The Enviroprobe was successfully used at location C-5 and illustrated some of the features of the BAT system, including the retrieval of multiple samples and rapid feedback as to wellpoint shield opening.
Groundwater samples were analyzed at the on-site laboratory for four organic volatiles that had been identified in the groundwater. TCE was found to be the most prevalent contaminant. Several samples were sent to an off-site laboratory for verification of the field results. On-site analytical results were consistently lower than results generated in the off-site laboratory. However, the on-site laboratory results were sufficiently accurate to guide the location of successive exploration locations.
The approach of using a truck-mounted penetrometer rig, CPTU soundings, penetrometer groundwater samplers and an on-site GC laboratory to conduct geo-environmental site characterization studies, in combination with drill rigs to set monitor wells, proved to be highly advantageous in terms of:
Based on the results of the penetrometer investigation, monitor wells were installed using drill rigs along the downstream perimeter of the contaminant plume. Little split-spoon sampling was performed during well installation as stratigraphy had already been defined to a high degree by the CPTU sounding data. Several rounds of monitor well sampling have verified the accuracy of penetrometer groundwater sampling and the use of field analytical testing.
Edge, R.W., K. Cordry, 1989. The Hydropunch: An In Situ Sampling Tool for Collecting Ground Water from Unconsolidated Sediments. Ground Water Monitoring Review, Vol. IX (3), pp 177-183.
Douglas, B.J., R.S. Olsen, 1981. Soil Classification Using Electric Cone Penetrometer. Cone Penetration Testing and Experience, ASCE, pp 209-227.
Saines, M., A.I. Strutynsky and G. Lytwynyshyn, 1989. Use of Piezometric Cone Penetration Testing in Hydrogeologic Investigations. Presented at the First USA/USSR Hydrogeology Conference, Moscow, USSR.
Strutynsky, A.I., B.J. Douglas, L.J. Mahar, G.F. Edmonds, and E. Hencey, 1985. Arctic Penetration Test Systems. Civil Engineering in the Arctic Offshore, ASCE, pp 162-168.
Strutynsky, A.I., R. Sandoford, and D. Cavaliere, 1991. Use of Piezometric Cone Penetration Testing with Electrical Conductivity Measurements (CPTU-EC) for Detection of Hydrocarbon Contamination in Saturated Granular Soils. Accepted for publication, Ground Water and Vadose Zone Investigations, ASTM.
Torstensson, B-A, 1984. A New System for Groundwater Monitoring. Ground Water Monitoring Review, Vol. IV (4), pp 131-138.
Andrew I. Strutynsky, P.E., the Technical Director of STRATIGRAPHICS, is responsible for all penetrometer operations at the company. He designs custom penetrometer equipment and performs research and development on new techniques. He was a speaker at the First International Symposium on Penetration Testing (1988), and conducted a Piezometric ConePenetration Test (CPTU) workshop at the Third Outdoor Action Conference (1989), sponsored by the NWWA. He has over 10 years of subsurface exploration experience, eight of which have been exclusively devoted to penetrometer exploration projects. He has worked throughout the United States, in the Far East, the Middle East, Central America, Europe, and the High Arctic. He has performed penetrometer research for a major oil company and both the U.S. Geological Survey and. the National Science Foundation. He has authored various papers on penetrometer equipment and interpretation of penetrometer data He has the B.S. and M.S degrees from the University of Illinois.
Timothy J Sainey P.G., is Senior Project Geologist at ERM-Midwest, Inc. With more than 17 years of geologic experience, his specialties include OVA/GC instrumentation, soil gas surveys and subsurface geophysical testing methods. While at ERM, Mr. Sainey has managed numerous environmental assessment, site remediation and closure projects, involving soil and groundwater sampling, well construction installation and monitoring, and the evaluation of remedial actions. Previously, he managed a geotechnical laboratory that specialized in ASTM testing. Also, he worked as an exploratory geologist for an oil company. Mr. Sainey received his B.S. and M.S. in geology from Ohio University in 1969 and 1973.