USGS Home
Contact USGS
Search USGS

 Home  Resources  Research  Publications  About  Contact Us  Groundwater Information

Monitoring A Permeable Reactive Iron Wall Installation In Unconsolidated Sediments By Using A Cross-Hole Radar Method*

By John W. Lane, Jr., Peter K. Joesten **, and Jennifer G. Savoie

ABSTRACT

Cross-hole common-depth (CD) radar scanning was used at the Massachusetts Military Reservation (MMR) in Cape Cod, Massachusetts, to monitor pilot-scale testing of a hydraulic-fracturing method to install permeable reactive zero-valent iron walls in unconsolidated sediments. The pilot-scale study was undertaken to assess the feasibility of using zero-valent iron to remediate ground water that is contaminated with chlorinated solvents at depths exceeding the range of conventional iron wall installation methods. The pilot-scale test was conducted at the site near the source area of Chemical Spill 10 (CS-10), a chlorinated-solvents plume that underlies the MMR. Two iron walls 5 meters (m) apart and 12 m long were designed to intersect the contaminated ground water at depths ranging from 24 to 37 m below land surface.

A series of post-installation cross-hole CD radar-scanning surveys were conducted in boreholes installed on opposite sides of the walls. The presence of iron significantly reduces the radar-pulse amplitude and can be identified using CD radar scanning. Significant decreases in cross-hole radar-pulse amplitude were observed in field data after the iron walls were installed. Changes in cross-hole radar-pulse amplitudes observed in the field data were compared to results of two-dimensional finite-difference time-domain models used to predict the effects of holes in the wall and wall edges. Analysis of these data from the south wall indicates the presence of an irregularly shaped wall about 8 m wide, extending from about 27 to 37 m below land surface. Analysis of data from the north wall is presently underway.

INTRODUCTION

Zero-valent iron has been used to remediate ground water that is contaminated with chlorinated solvents at more than a dozen sites in North America and Europe (Gillham and O’Hannesin, 1994). At many of these sites, a ‘funnel and gate’ technology has been implemented utilizing impermeable barriers to channel ground water through a permeable zone that contains zero-valent iron filings (Starr and Cherry, 1994, Smyth and others, 1996). An impermeable barrier, which often consists of sheet pilings, channels ground water towards a ‘gate’, which is filled with granular iron. At some sites, contaminated ground water is too deep to use the ‘funnel and gate’ method.

At the Massachusetts Military Reservation (MMR) at Cape Cod, Massachusetts, contaminated ground water near the source area for the Chemical Spill 10 (CS-10) plume (fig. 1), which contains tetrachloroethylene (PCE) and trichloroethylene (TCE), is located between 24 and 37 meters (m) below land surface. Because these depths exceed the practical limits of the normal ‘funnel and gate’ construction approach, the CS-10 site required an alternative installation approach. The method of installation that was selected was hydraulic fracturing (Hubble and others, 1997). The hydraulic-fracturing method uses a proprietary injection technology to install permeable reactive iron walls in unconsolidated sediments (Grant Hocking and S.L. Wells, Golder Sierra LLC, written commun., 1997). A pilot-scale test of their installation method was conducted near the CS-10 plume source area at MMR in which two iron walls were designed to intersect the contamination between 24 and 37 m below land surface.

Figure 1. Location of the study area and the CS-10 plume, Massachusetts Military Reservation (MMR), Cape Cod, Massachusetts.

The geometry of the iron walls was estimated using borehole-radar surveys conducted in 21 boreholes installed on opposite sides of the iron walls (fig. 2). The presence of the iron significantly reduces the radar-pulse amplitude, and can be identified using CD radar scanning. A two-dimensional finite-difference time-domain model (Xiao and others, 1998) was used to predict the effects of holes in the wall and wall edges on cross-hole radar amplitudes. This paper presents results of numerical modeling and analysis of cross-hole CD radar field data at the south wall.

Figure 2. Map view of the 7-centimeter plastic-cased observation boreholes at the MMR reactive wall site. 250 MHz cross-hole CD radar scans collected over the south wall after installation are shown.

NUMERICAL MODELING

A cross-sectional two-dimensional finite-difference time-domain model (Xiao and others, 1998) was used to predict the effects of holes in a wall and wall edges on cross-hole radar amplitudes. The simulation grid contains 5-centimeter (cm) by 5-cm cells, and spans 12.8 m (256 cells) in the horizontal direction perpendicular to the wall, and 12.8 m (256 cells) in the vertical direction. A 2.5-m (50 cell) super-absorbing boundary surrounds the model to reduce model-edge reflection effects. The electromagnetic properties of the model interior are consistent with field measurements at the MMR site. The model simulates a saturated sand with a relative dielectric permittivity (e_r) of 25, an electrical conductivity (s) of 0.01 siemens per meter (S/m), and a relative magnetic permeability (m_r) of 1. Due to the electrical conductivity of iron (1 x 107 S/m) (Carmichael, 1989), the wall was approximated as a 10-cm thick near-perfect reflector. The cross-hole radar pulse was simulated using a 250-Megahertz (MHz) Ricker wavelet, with a wavelength of 24 cm in the saturated sand, which is considered a reasonable approximation of the field data. Model responses were sampled at 5270 MHz using 1024 samples per waveform.

Models were used to predict the effects of holes in the wall on cross-hole radar data. The modeled wall contains holes having apertures ranging from 0- to 60-cm (0 to 2.50 wavelengths) in 5-cm (0.208 wavelength) increments. In the simulations, the transmitting and receiving antennas are 6.30 m apart, centered over the modeled hole in the wall (fig. 3). 

Figure 3. Model wall with hole in the wall ranging from 0 to 60 cm (0 to 2.50 wavelengths).

Numerical modeling results demonstrate the effect of the wall hole-size on radar-pulse amplitude. Radar-pulse amplitude increases as hole-size increases (fig. 4). Minor differences in radar-pulse amplitude observed for holes smaller than 10 cm suggest that the lower limit of hole-size detection is about 40 percent of the radar-pulse wavelength.

Figure 4. Radar model waveform data as a function of time with holes ranging in aperture from 0 to 60 cm (0 to 2.50 wavelengths).

Models also were constructed to predict the effects of wall edges on cross-hole radar data. These models contain a wall that extends vertically from the center of the model to the model boundary (fig. 5). The modeled separation between the radar transmitter and receiver ranges from 5.00 to 7.25 m, which is consistent with field conditions. The models simulate a CD cross-hole scan by stepping the transmitter and receiver location along the model at 10-cm intervals. Results of these simulations over the range of transmitter-receiver separations modeled show that the top of the wall consistently correlated to a depth where the pulse amplitude had decreased to about 43 percent of the normalized ‘background’ wave amplitude (fig. 6).

Figure 5. Model geometry used to determine the effects on radar waves due to the edges of the wall.

Figure 6. Model amplitudes normalized to unaffected amplitudes, for transmitter (TX) – receiver (RX) separations ranging from 5.00 to 7.25 meters plotted from 3 m above top edge to 3 m below top edge of the wall.

Results of the ‘hole in the wall’ models and ‘edge’ effect models shown in figures 4 and 6 can be used to help interpret field radar data to identify discontinuities in the wall and estimate wall dimensions.

FIELD METHODS

CD cross-hole radar surveys were conducted after iron wall installation at the MMR. Radar data were collected using a Mala GeoScience RAMAC [1] borehole radar system with electric-dipole antennas with center frequencies of 100 and 250 MHz. Electromagnetic waves of cross-hole radar are sensitive to the electrical conductivity of iron. The amplitude of an electromagnetic wave propagating between the radar transmitter and receiver that intersects a region containing iron will decrease with respect to ‘background’ radar-pulse amplitudes. CD cross-hole scans were conducted from boreholes on opposite sides of the wall (fig. 2). CD cross-hole scans were acquired by lowering the transmitter and receiver at 10-cm increments from the top of the 7-cm plastic casing to the bottom of the boreholes.

DATA ANALYSIS

The CD cross-hole radar-data was analyzed to identify decreases in radar-pulse amplitude induced by iron absorption. Radar-pulse amplitudes were normalized to the average amplitude of cross-hole scans below 40 m, where a normalized value of 1.00 implies a ‘background’ value. The scans below 40 m were used to determine the mean ‘background’ amplitude because these scans are several meters below the expected bottom of the wall. Figure 7 shows normalized CD cross-hole data between boreholes RW-20 and RW-13 (fig. 2). Amplitude changes caused by the iron wall and the results of numerical modeling are shown (fig. 7). The numerical model results were fit to the field data by changing the length of the wall (fig. 8). Based on this analysis, the section of wall between boreholes RW-20 and RW-13 is about 3.10 m in length, and extends from 31.7 m to 34.8 m below the top of casings.

Figure 7. Comparison of normalized model wall amplitude for a 3.10-meter wall and the normalized field data amplitude from the CD radar survey between boreholes RW-20 and RW-13, MMR, Cape Cod, Massachusetts.

Figure 8. Model configuration to determine the extent of the wall.

After normalization, all of the field data were combined into contoured cross-section plots. Figure 9 shows data along the cross-sectional extent of the south wall, by use of the ray paths that cross the south wall (fig. 2). Normalized amplitudes below the value of 84 percent, which is the median value of the normalized amplitude data, outline an irregularly shaped zone about 9 m wide and 12 m high. Horizontal bands extend outward from the main zone. These ‘stringers’ could be a result of the presence of salt water, suspension gel used in the hydraulic-fracturing process, or iron particles moving into higher permeability zones within the saturated sediments.

Figure 9. Cross-section plot of the normalized transmission amplitude data collected over the south iron filing wall, MMR, Cape Cod, Massachusetts. The 0.84 contour represents the median value of the normalized amplitude data.

The outline of the south wall using the results of the ‘edge’ model simulations are superimposed in figure 10. The wall edge is defined as the 43-percent normalized wave amplitude contour. The field data are complex, containing small-scale structures that could be interpreted as isolated small zones or stringers of iron; this demonstrates the need for additional modeling and analysis of the CD cross-hole data. However, based on numerical modeling results to date and the analysis of cross-hole radar field data from the south wall, the hydraulic fracturing produced an irregularly shaped wall that is about 8 m wide and 10 meters high and extends from about 27 to 37 m below land surface.

Figure 10. Cross-section plot of the normalized transmission amplitude data collected over the south iron filing wall with the estimated edges of the wall superimposed, MMR, Cape Cod, Massachusetts. The 0.43 contour represents the edges of the wall based on numerical modeling.

CONCLUSIONS

Cross-hole CD radar scanning was used at the Massachusetts Military Reservation in Cape Cod, Massachusetts, to monitor the pilot-scale testing of a hydraulic-fracturing method to install permeable reactive iron walls in unconsolidated sediments. The pilot-scale study was undertaken to assess the feasibility of using zero-valent iron to remediate ground water that is contaminated with chlorinated solvents at depths exceeding the range of conventional iron wall installation methods. The pilot-scale test site was conducted near the source area of Chemical Spill 10 (CS-10), a chlorinated-solvents plume that extends across the MMR. Two 12-m-long iron walls were designed to intersect contamination detected between about 24 and 37 m below land surface.

A series of post-installation cross-hole common-depth radar scanning surveys were conducted in rows of boreholes installed on opposite sides of the walls. Significant decreases in cross-hole radar-pulse amplitude are observed in field data after installation of the iron walls. Changes in cross-hole radar-pulse amplitude observed in field data were compared to results of two-dimensional finite-difference models that predict the effects of holes in the wall and wall edges on radar-pulse amplitudes. Based on numerical modeling results, preliminary analysis of cross-hole radar field data from the south wall indicate the hydraulic fracturing produced an irregularly shaped wall about 8 m wide, extending from about 27 to 37 m below land surface. Interpretation of small-scale structures that may be stringers of iron will require additional modeling and analysis of the CD cross-hole data.

ACKNOWLEDGMENTS

The authors thank the Air Force Center for Environmental Excellence at the Massachusetts Military Reservation and the Army National Guard for funding the work on this project. The authors thank Lei Xiao and Dr. Lanbo Liu of the University of Connecticut for the use of the finite-difference model and for their guidance with the modeling.

REFERENCES

Carmichael, R.S., 1989, Practical handbook of Physical Properties of Rocks and Minerals: Boca Raton, Florida, CRC Press, Inc., p. 373.

Gillham, R.W., and O’Hannesin, S.F., 1994, Enhanced degradation of halogenated aliphatics by zero-valent iron: Ground Water, v. 32, no. 6, p. 958-967.

Hubble, D.W., Gillham, R.W., and Cherry, J.A., 1997, Emplacement of zero-valent metal for remediation of deep contaminant plumes: in International Containment Technology Conference, St. Petersburg, Florida, February 9-12, 1997, Proceedings: Tallahassee, Florida, Florida State University, p. 872-878.

Smyth, D.J.A., Shikaze, S.G., and Cherry, J.A., 1996, Hydraulic performance of permeable barriers for in situ treatment of contaminated groundwater: Rumer, R.R. and Mitchell, J.K, eds., Assessment of Barrier Containment Technologies: A Comprehensive Treatment for Environmental Remediation Applications: Springfield, Virginia, National Technical Information Service, p. 881-886.

Starr, R.C., and Cherry, J.A., 1994, In Situ Remediation of Contaminated Ground Water: The Funnel-and-Gate System: Ground Water, v. 32, no. 3, p. 465-476.

Xiao, Lei, Liu, Lanbo, and Cormier, V.F., 1998, Two dimensional finite-difference time-domain solution for Maxwell’s equations using pseudo-spectral method: in Seventh International Conference on Ground Penetrating Radar (GPR’98), Lawrence, Kansas, May 27-30, 1998, Proceedings: Lawrence, Kansas, Radar Systems and Remote Sensing Laboratory, p. 585-589.

AUTHOR INFORMATION

John W. Lane, Jr. and Peter K. Joesten, U.S. Geological Survey, Storrs Mansfield, Connecticut (jwlane@usgs.gov, pjoesten@usgs.gov)

Jennifer G. Savoie, U.S. Geological Survey, Marlborough, Massachusetts (jsavoie@usgs.gov)

[1] The use of trade names in this report is for identification purposes only and does not constitute an endorsement by the U.S. Geological Survey.

Citation: Lane, John W., Jr., Joesten, Peter J., and Savoie, Jennifer, 1999, Monitoring a permeable reactive iron wall installation in unconsolidated sediments by using a cross-hole radar method, in Morganwalp, D.W. and Buxton, H.T., eds., U.S. Geological Survey Toxic Substances Hydrology Program -- Proceedings of the Technical Meeting, Charleston, South Carolina, March 8-12, 1999: USGS Water-Resources Investigations Report 99-4018C, v. 3, p. 747-756.