Grantee Research Project Results
Final Report: The Effect of Clay on DNAPL Behavior During Alcohol Flushing
EPA Grant Number: R827120Title: The Effect of Clay on DNAPL Behavior During Alcohol Flushing
Investigators: Hayden, Nancy J.
Institution: University of Vermont
EPA Project Officer: Aja, Hayley
Project Period: December 1, 1998 through November 11, 2001 (Extended to June 21, 2003)
Project Amount: $375,240
RFA: Exploratory Research - Environmental Engineering (1998) RFA Text | Recipients Lists
Research Category: Sustainable and Healthy Communities , Land and Waste Management , Safer Chemicals
Objective:
The overall objective of the research project was to investigate the effect of clays (and clay colloids) on the distribution of dense nonaqueous phase liquids (DNAPLs) in porous media and the subsequent effect on DNAPL (tetrachloroethylene [PCE]) dissolution and interfacial phenomena during alcohol (isopropyl alcohol [IPA]) flushing. The specific objectives were to: (1) investigate the relationship between clay type and clay fraction, mixed in a sandy media, on permeability and residual DNAPL saturation and the mechanisms involved in this relationship; and (2) determine the effect that clay has on the dissolution and mobilization of the trapped DNAPL within a porous media during alcohol flushing. A series of experiments and analyses were performed at the micro-scale, pore-scale, and column-scale levels to meet the objectives. Both swelling and nonswelling clays were used in batch, micromodel, and column experiments. Initial characterization of the clays included x-ray diffraction, surface area, cation exchange capacity, and particle size distribution. Batch experiments were used to investigate the effect of clay colloids on DNAPL interfacial phenomena and clay flocculation behavior. Micromodel visualization experiments coupled with pore-scale modeling also were performed. Column experiments were used to determine pressure-saturation relationships for water-DNAPL systems, permeabilities with and without a residual saturation, residual saturation and trapping, dissolution, and remobilization of the DNAPL during water and alcohol flushes.
Summary/Accomplishments (Outputs/Outcomes):
Washed Ottawa white Accusand and two source clays were used to create three distinct types of porous media: (1) sand, (2) a mixture of kaolinite with sand by weight (10 percent K), and (3) a mixture of Ca-montmorillonite with sand by weight (10 percent M). Approximately 89 percent of the montmorillinite and 50 percent of the kaolinite were smaller than 0.002 mm. Batch flocculation studies showed that 100 percent IPA acted as a flocculent with the two swelling clays but not with the kaolinite. The IPA was having an effect on the surface charge of the swelling clays. We also found that the d-001 spacing of Ca-montmorillonite was reduced in 100 percent IPA solutions, which would affect its ability to swell. The interaction of kaolinite with PCE also was noted as blobs of PCE coated with a thin film of kaolinite. This contrasted with the sharp interface that developed between PCE and the montmorillonite clays. The difference in behavior of kaolinite/PCE as compared to montmorillonite/PCE may be caused by the fact that kaolinite has a much lower surface charge than the swelling clays, thus allowing an interaction between the clays and the hydrophobic PCE.
Micromodels were used to observe pore-scale phenomena in beads and sandy porous media with and without clay. The micromodels consisted of two glass microscope slides filled with porous media and the edges sealed with an epoxy. More details of the micromodels are presented in Matmon and Hayden (2003). Figures 1 and 2 show some typical images of the pores of clay-containing media under dry, water-saturated, and PCE-saturated conditions. All four images in each figure represent a single location in the micromodel under the various conditions. The wet kaolinite aggregates coat the grains, and concentrate in the small pores and narrow wedges of large pores. Many large pores are still visible, and they appear similar to those observed in the sand micromodel (not shown). The wet montmorillonite behaves differently and forms a gel-like layer that spreads over the sand grains in a more continuous fashion, undoubtedly caused by the swelling nature of this clay. The montmorillonite covers and fills entire pores. This will dramatically affect fluid movement and the effective porosity. The movement of DNAPL into pore spaces (see Figures 1d and 2d) is clearly influenced by the presence of the clay. In the kaolinite-containing media (see Figure 1d), the DNAPL can move into areas where the pores are relatively large and open. The dark red of the two large central pores shows that the PCE has penetrated. In other areas, where the clay is filling most of the pore, the PCE has not penetrated. In the montmorillonite containing media, PCE was observed also to fill the larger pores (see Figure 2d). The large coating of montmorillonite in the central portion of the figure excluded the DNAPL.
Figure 1. Ten Percent Kaolinite in Sand Micromodel Showing: (a) Dry; (b) Water Saturated; (c) Clay/Grain Surface Enlarged; and (d) PCE in the Pore Space.
Figure 2. Ten Percent Montmorillonite in Sand Micromodel Showing: (a) Dry; (b) Water Saturated; (c) Clay/Grain Surface Enlarged; and (d) PCE in the Pore Space.
A conceptual model of clays and NAPLs at the pore scale was developed from a mathematical unit cell model and direct micromodel observation and measurement of clay-containing porous media. The mathematical model uses a unit cell concept with uniform spherical grains for simulating the sand in the sand-clay matrix (approximately 10-percent clay). A detailed analysis of the pore scale geometry, pore size distribution, NAPL entry pressures, and the effect of clay on this geometry were determined. Interesting NAPL saturation profiles were observed as a result of the complexity of the pore-space geometry with the different packing angles and the presence of clays. Details of the modeling approach and results are outlined in Matmon and Hayden (2003). The unit cell approach has applications for enhancing the mechanistic understanding and conceptualization, both visually and mathematically, of pore-scale processes, such as NAPL and clay distribution.
The effect of clay content on permeability is shown in Figure 3. As expected, the increase in clay content resulted in a decrease in the permeability for each clay type, even though the overall or total porosity of the packing did not change with the different media. The clay in the media created small pores, but did not change the overall total volume of pores, although the effective porosity was changed. Montmorillonite content resulted in a greater reduction in permeability than the kaolinite content. This is easy to understand when looking at the micrographs of these clays. In the case of the kaolinite (see Figures 1b and 1c), considerable clay can be seen in and around the contact points of sand grains. This area is less important in affecting permeability because most of the water flows through the largest pores. In the case of the 10 percent M (see Figures 2b and 2c), a gel-type layer has formed and in some cases has completely covered large pores, making them inaccessible for flow. This had a dramatic effect on permeability.
Figure 3. The Effect of Clay Content and Type on Hydraulic Conductivity (Points Represent Averages of Multiple Columns)
Capillary pressure-saturation relationships were determined for the various media to quantify and compare the differences in PCE saturation, entry pressure, and pore size distribution in media of the same total porosity. As expected, there is considerable difference in the pressure-saturation behavior for the different media. This is exhibited as differences in the shape of the curves, in the residual water saturation, and in the pressure needed to start drainage of water from the column, the entry pressure, Pe. The presence of the clay, especially the clay type, had a significant effect on these parameters. Table 1 presents a summary of the range or average for various parameters determined from the column studies. Although these results do not present any surprising trends, they do provide a more quantitative analysis than presently is found in the literature. It is interesting to note that increasing the clay content (beyond some small amount) does not show a similar increase in DNAPL residual saturation. At higher clay contents, it is more difficult to drain the water from small pores; therefore, most of the pores remain inaccessible to the DNAPL. The DNAPL that does remain seems to exist as more complicated, larger ganglia, and this accounts for the residual saturations that are higher than those of sand.
Sediment in Column | Sand | 5 percent K | 10 percent K | 20 percent K | 5 percent M | 10 percent M |
K (m/s) | 1.8 x 10-4 | 1.3 x 10-4 | 4.2 x 10-5 | 1.8 x 10-5 | 9.0 x 10-6 | 5.2 x 10-7 |
Pe (mbar) | 15-20 | 15-20 | 20-30 | 80-110 | 80-110 | 60-110 |
Irreducible water saturation (%) | 4-12 | 4-10 | 25-45 | 30-45 | 60-70 | Never reached |
Residual DNAPL saturation | 11 | 15 | 26 | 28 | 20 | 24 |
Kpce/Kinitial | 0.92 | 0.21 | 0.14 | 0.10 | 0.30 | 0.34 |
Kafter IPA/Kinitial | 1.09 | 0.32 | 0.74 | 0.23 | 2.0 | 4.6 |
One of the most interesting things to note from alcohol flushing was the effect of the IPA flushing on hydraulic conductivity as shown by the ratio of K after IPA flushing to K initial (Kclean/Kinitial). For the sand column, it is approximately the same; for the kaolinite-containing media, it is consistently lower. This means that, after IPA flushing, the kaolinite-containing media became less permeable. The PCE and the alcohol may result in movement of the clay that might result in smaller pores or pore throats. For the montmorillonite-containing media, the conductivity of the soil increases. This seems reasonable based on the XRD and bottle test results presented earlier. IPA affects the swelling and flocculation behavior of the montmorillonite, and it tends to shrink and agglomerate, thus allowing more flow through the pores. Dissolution studies of the PCE showed mass transfer limitations in the columns containing the clay. This was undoubtedly because of the more complicated DNAPL distribution in these columns as a result of the fingering and bypassing that occurred. The flowing water did not come into contact with much of the residual PCE and therefore did not achieve equilibrium. At higher IPA concentrations, however, some movement of the PCE did occur. Mobilization was noted for a wide variety of trapping numbers. This fact counters earlier hopes that one trapping number value may be used to characterize the potential for mobilization of a residual PCE saturation. In this study, the type of clay and clay content was shown to affect mobilization in media containing a clay/sand mixture.
Pore-scale observations of clay and DNAPLs were performed using micromodels. These models provided insight into the behavior of the different clays in the pore space and the effect on DNAPL movement and distribution. These observations led to the development of a unit-cell approach to model NAPL in clay-containing media that is useful in characterizing and depicting DNAPL in the pore space of clay-containing media. The pore-scale observations also were compared to quantitative results from column studies that were used to investigate hydraulic conductivity, DNAPL entry pressures, residual DNAPL saturation, and hydraulic conductivity after establishing residual saturation. Although the trends were as expected, the study did provide quantitative results for these parameters, given a variety of clay contents, and for two different clay types, and showed the differences between a swelling and nonswelling clay in the pore space. Among the most interesting results were the changes in hydraulic conductivity that occurred after the alcohol-flushing process. The montmorillonite-containing media became more permeable. Alcohol-flushing results showed the more complicated dissolution and mobilization behavior of PCE in clay-containing media. The larger, and often more isolated, PCE ganglia were a result of smaller pores within the clay media. These were often more likely to mobilize at lower trapping number (lower IPA concentrations).
Journal Articles on this Report : 2 Displayed | Download in RIS Format
Other project views: | All 8 publications | 2 publications in selected types | All 2 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Matmon D, Hayden NJ. Pore space analysis of NAPL distribution in sand-clay media. Advances in Water Resources 2003;26(7):773-785. |
R827120 (2001) R827120 (2002) R827120 (Final) |
not available |
|
Hayden N, Jiebold J, Farrell C, Laible J, Stacey R. Characterization and removal of DNAPL from sand and clay layered media.JOURNAL OF CONTAMINANT HYDROLOGY2006;86(1-2):53-71. |
R827120 (Final) |
Exit |
Supplemental Keywords:
groundwater, soil, solvents, dense nonaqueous phase liquid, DNAPL, NAPL, remediation, environmental chemistry, cleanup., RFA, Scientific Discipline, Air, Toxics, Waste, Remediation, Environmental Chemistry, HAPS, chemical mixtures, Hazardous Waste, Hazardous, Engineering, Chemistry, & Physics, Electron Microscopy, hazardous waste treatment, DNAPL, alcohol flushing, infrared spectroscopy sensor, interfacial phenomena, mass transfer, electrophoretic studies, hazardous chemicals, restoration, clayProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.