Wettability Changes in Trichloroethylene-Contaminated Sandstone

University of Greenwich, Pembroke, Chatham Maritime,. Kent, ME4 4AW, U.K., British Geological Survey, Crowmarsh. Gifford, Wallingford, Oxon, OX10 8BB,...
0 downloads 0 Views 602KB Size
Environ. Sci. Technol. 2001, 35, 1504-1510

Wettability Changes in Trichloroethylene-Contaminated Sandstone GAVIN HARROLD,† DAREN C. GOODDY,‡ DAVID N. LERNER,§ AND S T E P H E N A . L E H A R N E * ,† School of Earth and Environmental Science, University of Greenwich, Pembroke, Chatham Maritime, Kent, ME4 4AW, U.K., British Geological Survey, Crowmarsh Gifford, Wallingford, Oxon, OX10 8BB, U.K., and Groundwater Protection and Restoration Group, Department of Civil and Structural Engineering, Sheffield University, Mappin Street, Sheffield S1 3JD, U.K.

It is usually assumed that chlorinated solvent nonaqueous-phase liquids (NAPLs) are nonwetting with respect to water-saturated porous media. The focus of this work was to examine whether this supposition is appropriate for used trichloroethylene (TCE) samples. In this work, the term “used” indicates that the sample has been employed industrially and therefore contains solutes and breakdown products related to its previous use. The data obtained in this study indicate that exposure of initially water wet quartz slides to industrially used solvents can cause a contact angle change, measured through the aqueous phase, of 100° with a maximum stable contact angle of 170° (indicative of strong NAPL wetting characteristics) being recorded. The work on quartz slides was complemented by the use of sandstone cores. Wettability was measured using the Amott test. Used TCE again proved able to alter the wetting properties of sandstone to neutral wetting. The complexity of the industrially used samples precluded any realistic attempt to examine the agents causing these wetting changes. The data captured in these experiments were compared with laboratory grade TCE, and some attempts were made to synthesize known mixtures in order to replicate wetting changes. These experiments resulted in contact angle changes but did not alter the overall wettability of the quartz slides or sandstone cores. Finally the work reported here also demonstrates that increasing the duration of exposure to solvent has an important impact upon measured contact angle.

Introduction It is now recognized, on the basis of recent research conducted in the U.K. (1-3), that here, as in many other countries, there has been widespread and long-term contamination of groundwater resources by chlorinated solvents. This is of little surprise since the potential for groundwater contamination from these compounds is high due to a combination of their wide variety of usessin a large number * Corresponding author telephone: +44 (0)208 331 9565; fax: +44 (0)208 331 9805; e-mail: [email protected]. † University of Greenwich. ‡ British Geological Survey. § Sheffield University. 1504

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

of industriessoften sited over important U.K. aquifers and previously poor industrial and disposal practices. With migration and hence distribution of non-aqueous-phase liquids (NAPLs) being controlled by a combination of physicochemical properties and geological structure, a greater understanding of these two facets is essential. It is evident from the literature that there is a gap in our knowledge relating to changes in the physicochemical properties of chlorinated solvents. Thus, while a large number of studies have dealt with spillage’s of laboratory grade solvents into aquifers, these are not the only forms of solvent that may be introduced into the environment. Indeed any chlorinated solvent that does find its way into the subsurface is likely to contain quantities of solutes related to its previous or intended use (4). Indeed, attention has recently been drawn to the changes in aquifer wettability that might be anticipated due to the presence of various additives and solubilized oils and greases (5). Moreover, when one considers the frequency of handling and quantities of solvent used by industry and the often long times to detect spills, some consideration of the length of prior exposure is needed. Thus, this paper seeks to further our understanding of chlorinated solvent transport and fate in the subsurface via an investigation of wettability changes in Permo-Triassic Sherwood Sandstone, following exposure to various waste trichloroethylene (TCE) samples. Capillary Pressure, Wettability, and Its Influences. Capillary pressure has been defined as the property that causes a porous media to draw in a wetting fluid and repel a nonwetting fluid (6). Because of the spacial variability of this property within a natural porous media, it exerts a considerable influence on the shape of any spill as the DNAPL seeks the path of least capillary resistance (7). Capillary pressure exists whenever a DNAPL-water interface is curved, with the pressure across the interface increasing to balance the interfacial tension forces. In a smooth bore capillary tube, this pressure is related to the interfacial tension between the two liquids, the contact angle, and the capillary aperture, i.e.

Pc )

2σ cos θ r

(1)

where Pc is the capillary pressure (Pa); σ is the interfacial tension between the DNAPL and water (N m-1); r is the radius of the capillary tube (m), and θ is the contact anglesthe angle made by the fluid/fluid boundary at the three phase contact line. The contact angle is a measure of substrate wettability, i.e., its preference for one of the two liquids, and since it appears in the equation as a cosine, it clearly has a pivotal influence upon the direction in which the capillary pressure acts. The experience of the petroleum industry suggests that, although the wettability of reservoir rocks is partly defined by mineralogy, it is the presence of molecules in the aqueous or organic phases that are able to sorb onto the solid surface that exert the largest influence on wetting preference. The most significant changes occur through the adsorption of high molecular weight polar components present in crude oil that adsorb onto the reservoir rock surface through hydrogen bonding and cation-exchange mechanisms (8). More recently, it has been demonstrated that additives such as dodecylamine and polybutene amine, which are commonly added to petrol as detergents, can also have a considerable impact upon wetting (9). Research evaluating wettability changes in nonpetroleum NAPL systems has been relatively scarce. In general, it has 10.1021/es0000504 CCC: $20.00

 2001 American Chemical Society Published on Web 03/01/2001

TABLE 1. Uncontaminated Porewater Concentration for Porewaters Extracted from Western Quarries between 30 and 110 ma n median mean max min a

SEC

pH

Na

K

Ca

Mg

Si

DOC

HCO3

SO4

Cl

NO3-N

19 904 1030 2520 467

19 8.1 7.9 8.7 6.6

19 6.96 7.23 19.3 1.97

19 0.14 0.20 0.54 0.09

19 0.44 2.60 25.0 0.12

19 0.24 0.73 4.98 0.03

19 0.14 0.16 0.28 0.09

19 0.21 0.22 0.31 0.16

19 1.03 1.21 2.74 0.07

19 1.54 1.76 7.93 0.91

19 3.04 9.03 76.1 1.67

19 0.17 0.17 0.25 0.13

Units are in mmol/L.

been assumed that, in a DNAPL-water system, water is perfectly wetting (10). However, recent research has concluded that this may not always be the case. Powers et al. (11) demonstrated that both creosote and coal tars can be the wetting phase in a DNAPL-water-quartz system. Similarly, Barranco et al. (12) demonstrated that coal tars wet quartz media under coal tar receding conditions at acidic to neutral pH. Both groups concluded that it is the presence of high molecular mass compounds such as asphaltenes in these complex mixtures which are responsible for the wetting changes. With regard to chlorinated solvents, Thakur et al. (13) conducted a series of waterflood experiments with both laboratory grade 1,1,2-trichloroethane (TCA) and a complex DNAPL mixture recovered from the Petro Processors Superfund site. This was a mixture composed of chlorinated hydrocarbons, aromatic hydrocarbons, oils, and dissolved plastics. The results of the experiments showed that the field DNAPL displaced water from a water-saturated sand pack more efficiently than the TCA. Similarly, a greater flow time and quantity of water were required to displace the waste. These flow differences were attributed to viscosity and mineral wettability effects and demonstrate the danger in assuming that field samples will possess the same physicochemical properties as laboratory grade solvents. Additionally, Thakur also noted that the passage of the solvent phase through the sand pack altered the chemical composition of the field DNAPL.

Materials and Methods For this investigation, two methods were used to assess wettability. The first involved the measurement of three phase contact angles, while the second employed the Amott method (14). The Amott method examines the ratios of spontaneous drainage and imbibition to forced drainage and imbibition in a rock core. While the usefulness of such an idealized system may be the subject of debate, it was felt that information gained from this examination would be useful in the interpretation of the core wettability measurements. Both quartz microscope slides (Quartz Scientific Inc.) and a series of cores subsampled from two sandstone field cores, originating from the Western Quarries on the Runcorn Peninsula (Cheshire), were used in conjunction with five TCE solvent mixtures: (i) Laboratory grade TCE (99.8% purity, stabilized with 0.02% triethylamine; (Fisher Scientific), interfacial tension (IFT) 21.6 mN m-1 (ii) A waste sample supplied by Distillex, a solvent recovery company; IFT 11.33 mN m-1 (iii) A field sample recovered from a site in Canada; IFT 22.9 mN m-1 (iv) Laboratory grade TCE mixed with Castrol LM grease; IFT 22.6 mN m-1 (v) Laboratory grade TCE mixed with Castrol molybdenum grease (moly); IFT 25.4 mN m-1 All five solvent samples were used for the contact angle analysis and for the drainage and imbibition cycles involving the cores. The latter two mixtures were synthesized to examine whether commonly used automotive greases have any impact upon wettability and because the used solvent

sample seemed to contain large quantities of grease. TCE is, of course, commonly used as a degreaser. The aqueous phase used in all the contact angle measurements was distilled water. As the porewater from the cores (see Table 1) has relatively low ionic concentrations and was thus unlikely to have any major first-order effects, it was not considered necessary to use a synthetic groundwater. Instead a 2 mM solution of sodium bicarbonate was utilized to minimize cement solubilization from the sandstone matrix and buffer pH. Interfacial tension was measured by drop shape analysis using the DSA 10 instrument supplied by Kru ¨ ss Instruments, Germany. Contact Angle Measurement. Prior to each series of measurements, both the quartz slide and glass cell, which was used to house the slide and aqueous solution, were cleaned in chromic acid to remove any organic material. They were then rinsed with a copious amount of distilled water. The slides were left wet with water following this procedure and immediately subjected to the exposure process. This consisted simply of immersing the slide in the desired solvent for the required time, which ranged between 1 min and 24 h. Upon withdrawal from the solvent, the slide was rinsed with distilled water. Subsequently, the slide was then placed in a glass cell containing the aqueous-phase solution. Approximately 8 drops of TCE, each 15 µL in volume, were then placed on the slide, and the system was left for a period of 20 h to reach equilibrium. Images of individual drops were captured using a digital camera with the lens, in all cases positioned at 90° (in the horizontal plain) to the leading edge of the cell, thus ensuring that the angles photographed for all drops were the same. Images were subsequently magnified and printed, and both the left and right contact angles were measured. Imbibition and Drainage Measurement. Six separate imbibition and drainage cycles were completed involving the use of 28 different cores. Cycles 1 and 2 were duplicates. The sandstone cores were dried in an oven at 60 °C for 48 h and then resaturated as detailed in Table 2. It should be noted that each core letter/number combination refers to a new physical core sample. Cores 1a-5b were saturated with the sodium bicarbonate solution. Following this cores 3a-5b were centrifuged under their respective TCE samples at a capillary pressure of 68.95 M Pa, for a period of 8 h in order to drive solvent into the core for the aging process. These cores were subsequently removed from the centrifuge and stored in their TCE sample for the allotted time of prior exposure. Note: prior exposure refers to the length of time the TCE sample was allowed to stay in contact with the core before the wettability assessment was made. Following oven drying, cores 6a-f were left in the laboratory to pick up ambient atmospheric moisture, resulting in an aqueous saturation of approximately 2%. They were subsequently saturated with their respective solvent samples and stored as necessary. Once saturated, cores 1a-2d were placed in separate 250mL polycarbonate centrifuge tubes and immersed in their respective solvent samples for a period of 2 days to allow for spontaneous drainage. The volume of aqueous phase expelled, which could be found floating on the TCE, was VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1505

TABLE 2. Summary of Core Treatmentsa prior exposure (days) yes core initially saturated with:

solvent sample

no

20

30

40

2 mM sodium bicarbonate soln 2 mM sodium bicarbonate soln 2 mM sodium bicarbonate soln 2 mM sodium bicarbonate soln 2 mM sodium bicarbonate soln pure TCE waste TCE

waste TCE field TCE lab grade TCE TCE + LM TCE + moly pure TCE waste TCE

1a, 2a 1b, 2b 1c, 2c 1d, 2d 1e, 2e 6a, 6b 6c, 6d

3a 3b 3c 3d 3e

4a 4b 4c 4d 4e

5a 5b 6e 6f

a Cores were initially saturated with the liquid shown in column 1. The experiments run were commenced using samples shown in column 2. These experiments may have involved no prior exposure; this information is given in column 3. Columns 4-6 indicate those cores in which the solvent sample was allowed to reside in the core for a period of 20 (column 4), 30 (column 5), or 40 (column 6) days, respectively.

measured using a Hamilton microliter syringe. Subsequently, the four cores were centrifuged at incrementally increasing pressures until a capillary pressure of 68.95 MPa was reached. The total volume of the aqueous phase expelled was measured. Following this, the reverse process was completed, with the cores immersed in the aqueous phase for 2 days to allow for spontaneous imbibition and then centrifuged to a capillary pressure of 68.95 MPa. To allow the denser TCE to drain freely and to prohibit its imbibition, each core was placed in a core holder, which was constructed of sealed stainless steel tubing. This raised the core 5 mm off the bottom of the tube. The volume of solvent expelled was then measured. A similar process was followed for cores 3a-6b; however, as they were initially solvent-saturated, the experiments started with the aqueous-phase drive. Results are expressed as (a) the displacement by organic ratio (δo), which is the ratio of water volume displaced by spontaneous organic imbibition alone (Vwsp), to the total displaced by organic imbibition and centrifugal displacement (Vwt):

δo )

Vwsp Vwt

(2)

and (b) the displacement by water (aqueous-phase) ratio (δw), which is the ratio of the organic-phase volume displaced by spontaneous water imbibition (Vosp) to the total organicphase volume displaced by imbibition and centrifugal displacement (Vot):

δw )

Vosp Vot

FIGURE 1. Measured contact angles for waste TCE.

FIGURE 2. Measured contact angles for TCE and LM.

(3)

A preferentially water wetting core would have a positive δw and a zero value for the δo, with its water wetness increasing as δw approaches a value of 1. Conversely, an organic wetting core would have a positive δo and a zero δw. In the case of a neutraly wetting core, both ratios would be zero, and two positive ratios would be indicative of fractional wetting (14).

Results and Discussion

FIGURE 3. Measured contact angles for TCE and moly.

The results of the contact angle measurements conducted on the quartz slides are shown in Figures 1-5, where the line joins the average value for each set of data representing a particular duration of prior exposure. A number of interesting observations can be made. The most striking of these being the change in the contact angle of the waste sample (Figure 1) from an average of 18.8° on a “clean” quartz slide to averaged 112° on the same slide after it has been immersed in the solvent sample for just 1 min. In other words, the slide’s surface has changed from strongly water wetting to weakly solvent wetting. This weakly solvent wetting behavior

is maintained as the period or prior exposure is increased. The large variations in the data ranges for each averaged data point may suggest problems with the precision of the results. However, close inspection of the data reveals that the variation in the measurements for all the solvents on an unexposed surface is of the order of (3° and for the majority of the laboratory grade TCE measurements is (5°. This tends to suggest that precision is not an issue. Instead the large variation in values for the waste (and others) is an indication of the heterogeneous nature of the

1506

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

FIGURE 4. Measured contact angles for field TCE.

FIGURE 5. Measured contact angles for pure TCE. surface that has been produced by an uneven pattern of sorption of materials from the solvent to the surface. Close observation of representative drops illustrates that they do not spread uniformly, as demonstrated in Figure 6. Indeed the measured contact angle, especially with the waste solvent sample where the effect of prior exposure is most apparent, is dependent on which side of the drop is facing the camera, with one drop displaying many different contact angles as it spreads across the surface. Thus, each data series illustrate the range of angles that could be expected, perhaps even on a single drop, although here they reflect measurements across a minimum of 16 drops. Such a range of contact angles will be of little surprise to those who are aware of similar work conducted within the oil industry. Indeed it has been noted that on smooth, high-energy mineral surfaces (such as a quartz slide) finite contact angles arise when some form of surface deposition, such as adsorption, alters the surface of the mineral. However under these circumstances, such systems will not give a single valued contact angle, instead a range of stable or slowly changing contact angles may be observed that are bound by advancing and receding values (15). The effect of prior exposure to the two synthetic TCE wastes is much less marked. In the case of the LM grease mixture (Figure 2), the initial contact angle is 18.7° while the angle following solvent exposure for 24 h is 36.1°. Similarly, the molybdenum grease/solvent system (Figure 3) has an initial contact angle of 20.1° and 40.3° after an exposure of 24 h. In the case of the field sample (Figure 4), the change is more pronounced with the values being 19.3° and 51.7°, respectively. For the laboratory grade solvent sample (Figure 5), the contact angle remains much the same until the length of prior exposure exceeds 24 h, and then it significantly increases to a value of 43.2°. Finally, in the case of all five solvent samples, the various solutes present appear to have minimal effect on initial contact angle as the variation between all samples is only 1.4°, a value that is well within the limits of experimental error. It is proposed that these wetting changes arise from the sorption of material to the quartz surface. It is possible that this behavior could be due to differences in water/solvent

IFTs. The range of IFT measurements reported in the Materials and Methods however would appear to preclude this conjecture. Moreover, measurements of changes in IFT with time established that initial and equilibrium IFT values differed from each other by, at most, 4.5 mN m-1. An interesting phenomenon that emerged from the digitally captured images that were used to compute the IFT values was the formation of an interfacial film around the waste droplets, as shown in Figure 7. Similar interfacial films have been reported for crude oil (16), coal tars (17), and diesel (11). In the case of crude oil, these films are the result of the adsorption of high molecular weight polar molecules at the crude oil-water interface (16). Investigations by Nelson et al. (18) concluded that the formation of an interfacial film in the case of coal tars was due to weak, reversible bonding between water molecules and the coal tar constituents. Additionally, it was noted that similar bonding was observed in coal tar-water emulsions. No investigation of this film has been carried out, and at present it is possible that either of these two mechanisms could be responsible for the film formation given the solutes that may be present in the waste solvent sample. Imbibition and Drainage Measurements. The results from the series of imbibition and drainage experiments can be found in Table 3. In the case of the unexposed cores 1b-e and 2c-e, the high δw indicates that the majority of the aqueous-phase imbibition is spontaneous; conversely, the zero values for δo signifies that there is no spontaneous drainage. If a core contains a low wetting-phase saturation, as is the case at the end of the drainage run, and the wetting preference of the rock is sufficiently strong, capillary forces will cause the wetting fluid to imbibe spontaneously. This process is driven by the favorable surface energy change that occurs as the wetting fluid replaces the nonwetting. In the case of a core with a strong preference, the wetting fluid imbibes rapidly due to the large decrease in the surface free energy. Thus, as the Amott classification denotes, all these cores are strongly water wetting. In the case of core 2b, where the aqueous phase is displaced by field TCE, there is both a small degree of spontaneous drainage, as indicated by the low δo, and a large amount of spontaneous imbibition, denoting that areas of the core are displaying different wetting preferences. While this behavior differed from the repeat 1b, the large δw suggests that the overall preference of the core remains strongly water wetting, as with 1b. Similarly in cores 1a and 2a where the aqueous phase is displaced by waste TCE, there are positive values for both δw and δo, again the core is fractionally wetting. However, unlike 2b, the fact that the two ratios are much closer in value is symptomatic of a core where there is no obvious preference for one liquid over another for the majority of the pore space. The process of imbibition is most sensitive to wetting changes. This has been established by a number of authors including Morrow (19), who looked at uniformly wetting Teflon cores. He concluded that while the drainage process was insensitive to contact angles which were less than 50°, the process of imbibition was only insensitive below 22°. Similarly, it can be inferred from work reported by Demond et al. (20) that the drainage process becomes sensitive to contact angles larger than 35-50°, while the process of imbibition is affected at contact angles greater than 15-25°. Knowing this, why is the process of imbibition different in the waste core when the contact angles for all the solvent mixtures measured on uncontaminated quartz slide are all similar (about 20°)? The answer lies in the effect of ongoing exposure to the solvent. With the drainage run being conducted first, a procedure that took at least 1 week to complete, there is time for the wettability of these cores to change, especially if it is recalled that only 1 min of prior exposure to the waste effected the wettability of the slides. VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1507

FIGURE 6. Waste TCE droplet spreading on a surface previously exposed to the solvent. Ongoing exposure is less likely to have an effect on the results involving the cores that had been subjected to a period of prior exposure. This is due both to the smaller ratio of ongoing exposure to prior exposure and the fact that the imbibition (aqueous) run was conducted first. Thus, in the majority of cases, the bulk of the solvent is very quickly removed from the cores after their period of prior exposure by the spontaneous imbibition of the aqueous phase. Moving on to consider these results in more depth, it is apparent that the process of prior exposure has had a varying effect on the wettability of the cores. In the case of the synthesized solvent mixtures, TCE/moly and TCE/LM, there is no evidence that there has been any change with the displacement by water ratios (δw) remaining very high for all cores. Similar observations were made for the field TCE sample. However, in the case of aqueous-phase/laboratory grade solvent0saturated cores 3c, 4c, and 5a, the situation is somewhat different. Although there appears to be no little change in the δw after 20 days (2c), where the value is 80.8, as compared with 100 (1c) and 97.8 (2c), after 30 days (3c) the δw has dropped to a value of 38.7, indicating that the water wetting preference of the core has diminished quite 1508

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

considerably. This is again repeated in core 5a, which was subjected to 40 days prior exposure where the δw has a value of 35. This may be due to the adsorption of the solvent stabilizer (triethylamine) to the quartz surface. The combination of the ionizable amine group with the nonpolar ethyl groups should confer on the molecule some surface activity. The sequestration of a proton, from solution, by the amine group will result in the formation of a cationic species that can readily alter wetting through adsorption to the quartz surface (9). This forms the basis of ongoing research in our laboratories. The aqueous-phase/waste saturated cores 3a, 4a, and 5b show the least change when compared with the equivalent unexposed cores (1a and 2a). Again these three cores display fractional wetting characteristics with both spontaneous imbibition and drainage occurring. This points to the fact that the sorption of wettability altering molecules is less time dependent with this sample, something that is reflected in the results of the contact angle study. When considering the solvent-only saturated cores, it is noticeable that the δw is significantly higher for cores 6c and 6d than the equivalent sodium bicarbonate saturated cores

FIGURE 7. Interfacial skin formation around waste TCE droplet. 1 and 2a, something that is explainable by the ongoing exposure that cores 1 and 2a are subjected to. The most dramatic change in wettability occurs in the waste saturated core 6f, where there is no spontaneous imbibition but there is a degree of spontaneous drainage (the core has become weakly solvent wetting), this is in comparison to the fractional wetting found in the waste/sodium bicarbonate saturated cores. The much stronger wetting change is possibly due to the differences in the initial water saturations of the cores leading to a much thinner water film on the surface of core 6f. Thus this film is easier to destabilize and allows the adsorbable materials in the NAPL to come into contact with the core’s surface and alter the wetting through sorption. This more pronounced wetting change is again displayed by the TCE saturated core 6e, where the δw is reduced to 19. As a final point to this discussion, it is worth noting that a considerable amount of previously dissolved solute can remain behind in the core once the solvent phase has been removed. Indeed upon drying one of the cores, which had been contaminated with the waste sample, 35.6% of the weight of the waste remained. Similar observations were also made for both the synthesized solvent mixtures as well. Significance. The results obtained in these experiments challenge assumptions that porous geological media will always be water wetting in a water-chlorinated solventDNAPL system. While it is not suggested that every chlori-

nated solvent introduced into sandstone has the ability to change the wettability of the system, the results illustrate that such changes are possible, even when laboratory grade solvents are released. It has been established that the residence time of the solvent in the media over a short period of time is influential in altering wetting behavior. It should be anticipated that longer residence times might have a still greater effect, although this still needs to be investigated experimentally. It is important to stress that these changes can only occur when the solvent has come into contact with the mineral surface for a period of time. Thus, for a first time spill one would anticipate that the advancing contact angle of the solvent front should be around 20°, no matter the solute load in the solvent phase. On the other hand, the receding contact angle of the spill tail can be considerably larger. In the case of a repeat spill, one might expect that preferential pathways created by the previous invasion of a wetting changing solvent would become important in determining the transportation of a subsequent spill. It is perhaps worth noting at this point that any changes in the vadose zone are likely to be even more pronounced than those in the saturated zone. These data therefore serve to suggest that there is greater uncertainty in predicting the location of such contaminants and perhaps suggest a need to use the experimental techniques used in this study to assess the wettability of VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1509

TABLE 3. Summary of Results from Imbibition/Drainage Measurements displacement by

core

solvent

1a 1b 1c 1d 1e 2a 2b 2c 2d 2e 3a 3b 3c 3d 3e 4a 4b 4c 4d 4e 5a 5b 6a 6b 6c 6d 6e 6f

waste TCE field TCE pure TCE TCE + LM TCE + moly waste TCE field TCE pure TCE TCE + LM TCE + moly waste TCE field TCE pure TCE TCE + LM TCE + moly waste TCE field TCE Pure TCE TCE + LM TCE + moly pure TCE waste TCE pure TCE pure TCE waste TCE waste TCE pure TCE waste TCE

length of water organic prior ratio ratio exposure (δw)a (δo)a none 10.6 none 93.1 none 100 none 97.9 none 83.3 none 7.8 none 80.8 none 97.8 none 92.6 none 74.3 20 days 13.5 20 days 73.2 20 days 80.1 20 days 91.6 20 days 72.5 30 days 29.8 30 days 60.9 30 days 38.7 30 days 62.4 30 days 76.8 40 days 35 40 days 21 none 94 none 95.6 none 84.5 none 65.9 40 days 19 40 days 0

2.6 0 0 0 0 3 2 0 0 0 2.5 0 0 0 0 2.8 0 0 0 0 0 2.6 0 0 3.9 4.3 0 11

Acknowledgments wettability fractional wetting water wetting water wetting water wetting water wetting fractional wetting fractional wetting water wetting water wetting water wetting fractional wetting water wetting water wetting water wetting water wetting fractional wetting water wetting water wetting water wetting water wetting water wetting fractional wetting water wetting water wetting fractional wetting fractional wetting water wetting organic wetting

a Both ratios have been expressed in terms of a percentage, i.e., the percentage of imbibition/drainage that is spontaneous.

carefully preserved samples from spill sites. With regards to remediation, a move to more neutraly wetting conditions could be desirable, as it may be easier to displace a larger proportion of the DNAPL. However if the solvent contains a large proportion of solutes, then one could envisage that technologies such as vacuum extraction or certain thermal treatments might leave behind large quantities of undesirable residues. This would of course be dependent on the chemical properties of the incorporated solutes. Additionally the effectiveness of surfactant or alcohol flushes in dealing with such contaminants is uncertain. Finally with regard to the deposition of previously dissolved solutes, one would expect to see reductions in both the porosity and permeability of

1510

9

the geologic unit with certain fractures, pore network, and pores becoming blocked. These areas may subsequently act as absorptive capture zones for repeat spills.

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

This work forms part of a wide ranging study into the penetration depths of chlorinated hydrocarbon solvents in the Triassic Sandstone aquifers of the U.K. The authors are grateful not only for funding from the Engineering and Physical Sciences Research Council (Grant GR/L59191/01) and the Environment Agency but also to our solvent suppliers for samples. D.C.G. publishes with the permission of the Director of the British Geological Survey.

Literature Cited (1) Rivett, M. O.; Lerner, D. N.; Lloyd, J. W. J. Chart. Inst. Water Environ. Manage. 1990, 4, 242-250. (2) Burston, M. W.; Nazari, M. M.; Bishop, P. K. J. Hydrol. 1993, 149, 137-161. (3) Lawrence, A. R.; Stuart, M. E.; Barker, J. A.; Tester, D. J. J. Chart. Inst. Water Environ. Manage. 1996, 10, 263-272. (4) Verschueren, K. Handbook of Environmental Data on Organic Chemicals; Van Nostrand Reinhold: New York, 1983. (5) Jackson, R. E.; Dwarakanath, V. Ground Water Monit. Remed. 1999, 102-110. (6) Bear, J. Dynamics of Fluids in Porous Material; Dover Publications Inc: New York, 1972. (7) Mercer, D. M.; Cohen, M. J. Contam. Hydrol. 1990, 6, 107-163. (8) Dubey, S. T.; Waxman. M. H. Proceedings: 1989 SPE International Symposium on Oilfield Chemistry; Society of Petroleum Engineers: 1989. (9) Powers, S. E.; Tamblin, M. E. J. Contam. Hydrol. 1995, 19, 105125. (10) Kueper, B. H.; McWhorter, D. B. Ground Water 1991, 29, 716728. (11) Powers, S. E.; Anckner, W. H.; Seacord, T. F. J. Environ. Eng. 1996, 122, 889-896. (12) Barranco, F. T.; Dawson, H. E. Environ. Sci. Technol. 1999, 33, 1598-1603. (13) Thakur, S.; Schenewerk, P.; Wolcott, J.; Groves, F., Jr. J. Environ. Sci. Health 1995, 30, 1105-1118. (14) Anderson, W. G. J. Petrol. Technol. 1986, 38, 1125-1144. (15) Morrow, N. R. J. Can. Petrol. Technol. 1975, 14, 49-69. (16) Dasgupta, P. K.; Hwang, H. Anal. Chem. 1985, 18, 1009-1012. (17) Luthy, R. G.; Ramaswami, A.; Ghoshal, S.; Merkel, W. Environ. Sci. Technol. 1993, 27, 2914-2918. (18) Nelson, E. C.; Ghoshal, S.; Edwards, J. C.; Marsh, G. X.; Luthy, R. L. Environ. Sci. Technol. 1996, 30, 1014-1023. (19) Morrow, N. R. J. Can. Petrol. Technol. 1976, 15, 42-53. (20) Demond, A. H.; Roberts, P. V. Water Resour. Res. 1991, 27, 423437.

Received for review March 9, 2000. Revised manuscript received December 11, 2000. Accepted January 10, 2001. ES0000504