Soil Wettability As Determined from Using Low-Field Nuclear Magnetic

May 7, 2003 - parts of the world and occurs naturally or as a result of a fire or pollution event. ... important and maybe dominant factor, is not eno...
0 downloads 0 Views 173KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 2701-2706

Soil Wettability As Determined from Using Low-Field Nuclear Magnetic Resonance F L O R E N C E P . M A N A L O , †,‡ A P O S T O L O S K A N T Z A S , * ,†,‡ A N D COOPER H. LANGFORD§ Department of Chemical and Petroleum Engineering, Department of Chemistry, and TIPM Laboratory, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada T2N 1N4

The molarity of ethanol droplet and water drop penetration time methods are commonly used to determine soil wettability because these tests are quick and easy to perform. However, these tests do not provide reproducible results on the same sample. Low-field nuclear magnetic resonance (NMR) is shown as an alternative tool to determine soil wettability. Addition of small amounts of water in dry wettable porous media produces predominant amplitude peaks at transverse relaxation times (T2) of 100 ms or less while addition of water in dry water-repellent porous media with the same pore structure produce predominant amplitude peaks at T2 values near 1000 ms. The geometric mean of T2 (T2gm) from water-repellent samples immediately after the addition of water is greater than 1000 ms, which is close to that of bulk water, while T2gm from wettable samples immediately after the addition of water is significantly less than 1000 ms. Measurements over time show that water-repellent samples eventually reach the same equilibrium end point as its corresponding wettable sample when continually exposed to water. This paper will show that NMR can be used to formulate a screening criterion for quickly determining wettability. The advantage of using NMR is that the results are reproducible provided the sample is prepared and analyzed in a systematic manner.

Introduction Soil water repellency is defined as poor soil wettability (1). Water-repellent soil is characterized by lack of plant growth, extra dusty appearance, and consistency due to the lack of aggregates in the soil (2). This phenomenon may increase the likelihood of groundwater contamination (3). Soil water repellency is encountered in diverse ecosystems in many parts of the world and occurs naturally or as a result of a fire or pollution event. Pollution-induced water repellency usually develops after prolonged exposure of soil to liquid- or vaporphase petroleum hydrocarbons (1). Water repellency in soil has been typically attributed to the presence of hydrophobic organic substances forming a coating over the surface of soil surfaces (1, 4). It is important * Corresponding author e-mail: [email protected]; phone: (403)220-8907; fax: (403)282-5060. † Department of Chemical and Petroleum Engineering. ‡ TIPM Laboratory. § Department of Chemistry. 10.1021/es0259685 CCC: $25.00 Published on Web 05/07/2003

 2003 American Chemical Society

to remember that organic matter content, although it is an important and maybe dominant factor, is not enough on its own to predict the persistence of water repellency. The quality of organic matter must also be known (4-6). The surface coverage density and conformation of hydrophobic organic compounds may be parameters that help determine how readily water spreads on soil surfaces. Several researchers have proposed that the interfacial conformation of sorbed water-repellent substances changes in response to compositional changes of the interstitial pore fluids. Hydrophilic functional groups in organic matter (i.e., -OH, -COOH, and -NH2) interact with water molecules when the soil is wet but interact with each other when the soil is dry (1, 2, 7). Repeated observations of reversible water repellency suggest that mechanisms other than removal of causative agents can also contribute to reductions in soil water repellency (1). There is no satisfactory method available for measuring wettability in the field, making it necessary to estimate wettability from laboratory measurements. Methods for measuring soil wettability include the molarity of ethanol droplet (MED) and water drop penetration time (WDPT) tests. Each of these is easy to perform, but the results obtained are questionable. Results obtained from repeated MED tests on the same soil do not usually bear a high correlation coefficient because of sample heterogeneity and the fact that soil/water/ air systems cannot be directly compared to soil/aqueous ethanol/air systems without accounting for differences in the nature of their solid-liquid molecular interactions. Some organic and inorganic compounds on soil particle surfaces are more soluble in ethanol than they are in water (1). A problem with the WDPT test is that, like the MED test, this test uses arbitrary time scales that are chosen for convenience and have no specific physical meaning (5, 8). This paper proposes measuring soil wettability using low-field nuclear magnetic resonance (NMR) as an alternative to the MED and WDPT tests for measuring soil wettability. NMR occurs when the nuclei of certain atoms are under the influence of a static magnetic field and exposed to a second oscillating magnetic field (9). The magnetic moments of nuclei, similar to the earth’s magnetic axis, are aligned in an external magnetic field and the nuclei are spinning around their own axes, like spinning tops. A second magnetic field in the form of a radiofrequency pulse is applied perpendicular to the static magnetic field. This causes the magnetization vector to move perpendicular to its original orientation while it is still spinning at a characteristic frequency. Over time, the system will return to equilibrium, and the time constant for this process is known as the transverse relaxation time constant (T2). Each proton-containing fluid, when measured as a bulk fluid phase, has a characteristic T2 value. Fluids contained within porous media experience restriction of proton movement due to binding onto organic matter or clays, residing in smaller pore spaces or residing in a viscous environment. Consequently, the presence of these factors leads to faster relaxation and, thus, lower T2 values. NMR has rapidly progressed to become one of the most powerful nondestructive analytical methods and is used in many fields, including medicine, chemistry, and the petroleum industry (11). Advantages of using this tool are that results are obtained quickly, are obtained easily, and are reproducible; only a small representative sample is required for analysis, and it is possible that these measurements can be taken on-site without having to transfer the sample to laboratory facilities. VOL. 37, NO. 12, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2701

TABLE 1. Chemical Composition of Sand and Clays substance

sand (%)

IRC (%)

KRC (%)

silica alumina potassium oxide iron oxide titanium oxide magnesium oxide barium oxide calcium oxide sodium oxide

93.5 4.5

72.7 15.6 2.2 1.2 0.6 0.5 0.4 0.2 0.1

49.0 33.1 0.6 1.1 1.2 0.2

0.3 0.3 0.4 0.8

0.2 0.1

Experimental Section Materials. The porous media used included sand, clays, fulvic acids (FA), and soils. Commercial topsoil loam and a heavy crude oil from Cold Lake, AB, were used to extract organics for the purpose of coating sands in order to create additional samples for analysis. The sand was obtained from Target Products Limited in Calgary, AB. The clays were supplied by Plainsman Clays in Medicine Hat, AB. The illite-rich clay (IRC) is mined from Ravenscrag, SK, while the kaolinite-rich clay (KRC) is mined in Troy, ID. Composition of the sand and clays are shown in Table 1 with the balance being organics lost on ignition (12). The Laurentian FA were extracted using the procedure formerly endorsed as the standard by the International Humic Substances Society (IHSS) (13). Distilled water was added to all of the samples in these experiments. Wettable and water-repellent soils from an agricultural field near Devon, AB, (DCW and DNW, respectively) were used in this program. This site was contaminated as a result of a crude oil well blow-out in 1948. Despite remediation efforts, the residual oil contents of the water-repellent and the wettable soils from this site are 6.5 ( 0.3 and 1.5 ( 0.1 g of oil/kg of soil, respectively, on an oven-dry soil basis (determined by a 24-h Soxhlet extraction using dichloromethane as the extractant). Wettable and water-repellent soils were also collected from a site near Bruderheim, AB, that was contaminated with crude oil in 1982. This soil consists of dune sand with a very high quartz content. The residual oil contents for the wettable (BCW) and waterrepellent (BNW) soils are 0.1 ( 0.0 and 2.3 ( 0.2 g of oil/kg of soil (on an oven-dry soil basis), respectively. Like the Devon site, the oil-contaminated soil at Bruderheim only began to manifest severe water repellency and structural degradation a decade or more after oil contamination. Pristine wettable soil (i.e., never contaminated with petroleum hydrocarbons) was collected from the Ellerslie research station near Edmonton, AB. This soil (denoted as EPW) was sampled well outside the oil-contaminated area at the Ellerslie site (1). Instrumentation. Particle size distributions were performed in duplicate with the Malvern Instruments Mastersizer 2000 particle analyzer, a laser diffraction analyzer capable of detecting particles in the range of 0.02-2000 µm. NMR measurements were made with the Corespec 1000 at a frequency of 1 MHz (field strength is 0.024 T). While quantitative analysis by NMR has traditionally used highfield measurements where the frequency is on the order of 100 MHz, low-field NMR has since been found to be more accurate than high-field NMR in measuring water and oil volumes in media that possess large internal magnetic gradients. Higher magnetic fields result in a stronger signal; hence, measurements are completed in a shorter period of time. However, this same field causes larger internal gradients, which may lead to the degradation of the measurements. Another disadvantage of high-field NMR for analyzing fluids within porous media is that large internal magnetic field gradients can obscure relaxation curves for bulk fluid or surface relaxation, which prevents researchers from 2702

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 12, 2003

determining the relative importance of surface relaxivity on fluid-solid interactions (14). Duplicate runs were performed during each measurement in order to ensure that results were genuine and not a result of experimental artifacts. The data were processed using EchoFit v. 3.02, which is a software package created by NUMAR Corporation, a subsidiary of Halliburton. This software program employed the nonnegative least-squares method to create the NMR spectra used for analysis. Organic Matter Extraction and Sand Coating. Humic acids (HA) extracted from commercial topsoil loam and asphaltenes extracted from heavy oil were used to coat separate subsamples of the sand. This method of extracting organics and coating sands with the extracted substances has proven to be useful by allowing researchers to control factors such as composition to some extent and particle size distribution. As a result, closer focus on the mechanism of wettability alteration has been possible (7). The procedure used to extract HA from the topsoil is a slight variation to the extraction procedure endorsed by the IHSS (13). The variation was made because these experiments did not require HA of very high purity. Purity levels as high as those obtained from the IHSS extraction procedure would have severely restricted the rate at which the HA could be extracted. Furthermore, heterogeneous solutions are acceptable in this program since it more accurately represents organic matter found in soils. HA were extracted using 0.1 N NaOH under a nitrogen blanket to ensure that the oxygen does not degrade any of the organics. The solution was vigorously agitated for 4 h and then allowed to settle for at least 12 h or overnight. The liquor was decanted off, and any floating solids were skimmed. HCl was added to bring the pH of the liquid down to 2 and to allow the HA to precipitate while keeping the FA in solution. The HA were redissolved in 0.1 N NaOH, and 2 kg of sand was added for every liter of liquid solution. The solution was brought to a pH of 2 once again over a time span of 1 h, and then the slurry was dried. Obtaining a heterogeneous sample for coating sands was considered acceptable since field conditions are more closely represented that way (7). The procedure used to extract asphaltenes from heavy oil is a slight variation from the method endorsed by the American Society for Testing and Materials (ASTM) (15). The solvent used was n-pentane instead of n-heptane, and consequently, more resinous material was obtained as compared to material acquired from using n-heptane. A heavy crude oil from Cold Lake, AB, was heated to a temperature of 80 °C, and then 2 g of sand was added for every gram of heavy oil. The slurry was mixed thoroughly and then left to age overnight. The solvent n-pentane was added to the slurry in the ratio of 100 mL of solvent/g of sample the next morning. The mixture was stirred thoroughly under a fume hood for several minutes and then placed in the oven for 15 min. While the sample was in the oven, a funnel was placed inside an Erlenmeyer flask, and a piece of filter paper was placed into the funnel. The slurry from the oven was slowly poured into the funnel. The filter paper was replaced as necessary, and care was taken to keep all the solids that had collected on the filter paper. The solids that rested on the filter paper were rinsed with additional n-pentane, then transferred to a clean watch glass, and placed in the oven to dry. This procedure allows for the organic matter to completely coat the sand, thus allowing the sample to behave identically to isolated asphaltenes. Sample Preparation for Analysis. A representative subsample of each porous medium was analyzed in the Mastersizer 2000 to determine particle size distributions. All porous media were air-dried prior to sample preparation for NMR analysis. The CoreSpec 1000 was tuned at least daily using water doped with 0.002 M CuSO4 (the relaxation time for this solution was approximately 240 ms). This was to

TABLE 2. Saturation Levels of Samples Immediately after Water Addition sample ID

saturation (wt %)

FA IRC KRC original HA AC

60 45 50 30 40 25

sample ID

saturation (wt %)

BCW BNW DCW DNW EPW

25 20 40 25 40

FIGURE 2. Particle size distribution of clay samples.

FIGURE 1. Particle size distribution of soil samples. Soils collected from the same site, although they have different wettabilities, have the same particle size distribution. This shows that a contaminated soil that has experienced wettability alteration did not achieve this through physical changes of the soil particles. ensure that differences in NMR spectra over time were indeed due to changes in the soil-water-air system and not due to partial proton excitation. Each unconsolidated porous medium was poured into a glass vial 3.5 cm in diameter and 5.0 cm in height. The sample was packed slightly so that the column was slightly less than 1.5 cm in height, which is the length of the homogeneous region of the magnet used in these experiments. Distilled water was added at a rate of approximately 1 drop/s until there was a thin film of fluid on top of the sample. Water was added in this manner to simulate rainfall. The presence of the film above the sample was desired so that the NMR spectra would include an amplitude peak corresponding to bulk water. The amount of water added to each sample is indicated in Table 2. Measurements conducted with the CoreSpec 1000 show that distilled water alone produces an amplitude peak at 2500 ms, which corresponds well to published values (16, 17). The lid of the vial was placed and wrapped several times with Teflon tape in order to minimize water loss as a result of evaporation. Each sample was tested before water was added, every half hour for the 4 h immediately after adding water, then once an hour for 4 h afterward, then once daily for 20 days or until evaporation losses became significant (i.e., more than 5%).

Results and Discussion The particle size distribution of each soil analyzed in this program can be found in Figure 1, and the corresponding plot for the clay samples is included in Figure 2. The coated sand samples did not undergo particle size analysis because sample preparation includes removing the organic material from the particles and removing the coatings would give the original sand. It is assumed that the coating is not thick enough to significantly affect the particle size distribution of the sand samples. Figure 1 shows that soils from the same

FIGURE 3. Normalized NMR spectra of distilled water in DCW over time. Note the predominant amplitude peaks at times less than 10 ms and that there is very little change in amplitude peak location over 168 h. These characteristics are typical in NMR spectra of water in wettable porous media. site have the same particle size distribution regardless of wettability. While one may say that this is due to removing the organics from the particles, this also implies that wettability is not due to physical differences of the particles. All samples were indeed dry prior to adding water because the NMR spectra obtained before the addition of water (not shown) depict only noise, indicating the absence of protons in the liquid phase. Duplicate measurements were taken, and variation in the total amplitude for the two successive runs was less than 4% for all the samples. A standard of 0.002 M CuSO4 was measured before each NMR test to ensure that the relaxometer excited the protons to the same extent each time. Total amplitude readings from the standard varied by less than 5%. Figure 3 shows NMR spectra obtained after adding water to DCW over time. Water in a wettable sample produces an amplitude peak at low T2 values (i.e., less than 10 ms), which indicates the presence of water as thin films or in small pores. An amplitude peak at T2 values between 10 and 100 ms indicates the presence of water in mid-sized pores. The amplitude peak on the right-hand side of the T2 distribution is indicative of the bulk water present as a film above the sample. While bulk water alone produces an amplitude peak at 2500 ms, the bulk water peak for DCW is at values of approximately 1000 ms and less. This difference is due to fast energy exchange between water protons bound to the surface and water protons forming the film of water above the sample, which does not occur in a bulk water sample. Another feature in Figure 3 that is typical of water in wettable samples is that the locations of the amplitude peaks do not VOL. 37, NO. 12, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2703

FIGURE 4. Normalized NMR spectra of distilled water in IRC over time. These spectra contain the same characteristics as wettable soil (see Figure 3). The amplitude peak at approximately 1500 ms is due to the bulk water present as a film above the sample.

FIGURE 5. Normalized NMR spectra of distilled water in DNW over time. Note the predominant amplitude peaks at 2000 ms for zero hours and the shift of this peak toward lower relaxation times over 168 h. This is typical in NMR spectra of water in water-repellent porous media. Another typical characteristic is the appearance of amplitude peaks at lower relaxation times (between 2 and 20 ms in this figure) after prolonged exposure to water change significantly over time (17). This suggests that water uptake in these samples is very fast, perhaps almost immediately after water is introduced. The slight differences in Figure 3 may be due to power fluctuations during measurements, small diffusion effects, and variability in water uptake due to fluid redistribution. Figure 4 shows the spectra of distilled water in IRC over time. One can see from this figure that the trends apparent in Figure 3 (i.e., location of amplitude peaks at all orders of magnitude of T2, lack of movements of amplitude peaks over time) are also included in Figure 4. A significant difference is the location of the amplitude peak on the left-hand side of the spectra. The water relaxed faster in the clay than in the soil (1 ms for IRC as compared to 2-3 ms for DCW), which agrees with other researchers’ findings that clay-bound water will have extremely fast relaxation times (10). Results obtained here indicate that an amplitude peak at 1 ms is indicative of the presence of illite clays in a sample and agrees with other researchers’ results (17-19). The NMR spectra of water in DNW over time are shown in Figure 5. Water in DNW produces three amplitude peaks as it does in DCW, but Figure 5 shows that only the peak indicating the presence of bulk water is significant (i.e., the maximum of the amplitude peak at T2 ∼2500 ms is significantly greater than the maximum of the other amplitude 2704

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 12, 2003

FIGURE 6. Mass of water in samples, determined using NMR data and mass balance calculations. The two methods correlate very well, and the decrease over time is due to evaporation of water in the sample. peaks). This amplitude peak is located at approximately 2500 ms immediately after water addition, which is typical for water-repellent samples (17). Another feature in Figure 5 that is typical of water in water-repellent samples is that the amplitude peaks move to lower T2 values over time. At time equal to zero hours, the peaks at T2 values of less than 100 ms had a maximum of less than 5%, but the local maximum increased to almost 10% after 168 h. This verifies that water uptake in these samples is significantly slower in waterrepellent samples than in wettable samples. Amplitude peaks at approximately 2 ms, which appear in DNW 1 week after the addition of water, corresponds to the location of the amplitude peaks for DCW, the neighboring wettable soil. This suggests that water-repellent samples can reach the same equilibrium end point as its corresponding wettable sample if it remains exposed to water for a long enough period of time. A method to determine the mass of water in the sample based on the total signal amplitude obtained during an NMR measurement was developed in-house and compared to mass of water from mass balance calculations. The NMR method is based on the fact that the total signal amplitude is directly proportional to the number of protons, thus in this case the amount of fluid, in a sample. The amount of water in a sample can be quantitatively determined either by performing a water mass balance or by using NMR data to calculate the amplitude index, which is the NMR amplitude signal of each fluid per unit mass of the same fluid. This calculation method is explained in detail elsewhere (16, 17, 20-22). Figure 6 shows water weights calculated using the two different methods for some of the samples analyzed in this program. One can see from these figures that the water weights calculated using the two different methods are very similar. Variation between the values from the two different methods is less than 5%. Another trend visible in Figure 6 is that the water weights for any sample slightly decreases over time. This is due to evaporation, and NMR can apparently detect changes in the mass of water that are as small as 0.01 g. Water loss for BCW and BNW appears to be significant in Figure 6 as compared to the other samples. This is because there is little organic matter in the soils from Bruderheim as compared to the soils collected from other sites. As such, water is not bound as tightly to the sand grains as it is to the clay or organic material. Another possible explanation for the significant water loss is that the seal on these two samples were not as good as the seal on the other samples. Independent tests (23) show fluid uptake in soils to be a two-component process: a fast process, which requires a maximum of 24 h to complete, and a slow process, which

FIGURE 7. T2gm immediately after water addition. The water-repellent samples produce T2gm values of 1500 ms and more, which is the same order of magnitude as bulk water, while the wettable samples produce T2gm values that are significantly lower than 1000 ms. This finding can be used as a preliminary screening criterion to determine the wettability of a sample. requires more that 20 d to approach an equilibrium-like state. The same mechanism was noted in soil components, although the time frames for each of the two processes are much shorter. The fast process is due to wetting and welling of dry organics to form “gels”. This gel formation due to swelling produces micropores, and water in these pores relax very quickly. Additionally, thick organic coating on mineral particles factilitates an increase in the quantity of water entering new micropores in this time frame. This explains the presence of amplitude peaks at low T2 values for the water-wetting materials. A thinner organic coating, on the other hand, requires the disruption of the interactions between the organic coating and the supporting clay particles in order for gel formation to occur. The slow wetting process takes more than 10 d and is due to water penetration between the clay-organic matter interface. This occurs only after the interactions between the clay and the organic matter have been broken. However, there are not as many new sorption sites at this point, thus less water is absorbed during this process. This is why there is relatively little water uptake after the first 24 h of the experiment (23). Another parameter from NMR data that can be analyzed to determine porous media wettability is the geometric mean of T2 (T2gm), which is a weighted average of T2. Figure 7 shows the value of T2gm immediately after water addition for all samples analyzed. When measured immediately after water addition, T2gm values are 1 order of magnitude lower for wettable samples than for water-repellent samples, despite the fact that the terms from bulk water are included in the calculation of T2gm. Figure 7 shows that HA-coated sand (HA) is similar in wettability to uncoated sand (original) since they have similar T2gm values immediately after water addition. This suggests that coating the sand with humic acids extracted from commercial topsoil did not significantly change the wettability of the sand. On the other hand, T2gm for the asphaltene-coated sand (AC) is on the same order of magnitude as the water-repellent soils. This suggests that the substances causing water repellency in soils that have been contaminated with petroleum hydrocarbons include the resinous material that is insoluble to n-pentane. Additional work is required to verify or refute this finding. Figure 8 shows T2gm values over time for the soil samples, and Figure 9 shows the same parameter over time for the clays, FA, and uncoated and coated sands. These two figures show that the T2gm values for corresponding wettable and water-repellent samples will eventually converge. This is because of the trends apparent in both Figures 8 and 9, which is that T2gm appears

FIGURE 8. Change in T2gm over time for soils. The values from the wettable samples do not change significantly over the course of the experiment, but the values from the water-repellent samples decrease dramatically over time. These trends can also be used when determining sample wettability.

FIGURE 9. Change in T2gm over time for clays, FA, and sands. Note that soil components show the same behaviors as whole soils. This suggests that the use of coated sands is useful in determining solid-fluid interactions between water and soil because the synthetic systems closely mimic the behavior of the field samples. An advantage of using coated sands is that the composition can be controlled.

to remain relatively constant for preferentially water-wet samples with water while T2gm for water-repellent samples with water decreases with time. This holds true despite significant mass loss due to evaporation (see BCW and BNW). These findings have been used to develop the following as preliminary screening criteria, although it is acknowledged that additional work is necessary to refine these rules: (i) A sample is water-repellent if T2gm immediately after water addition is greater than 1000 ms and this parameter decreases by 1 order of magnitude over 3 weeks. (ii) A sample is wettable if T2gm immediately after water addition is significantly less than 1000 ms and this value does not change significantly over 3 weeks.

Acknowledgments The authors thank Dr. Tiona Todoruk from the Department of Chemistry at the University of Calgary for her help in obtaining the materials used as well as Dr. Michael Aikman for his help in extracting HA and coating sands. Provision of the Mastersizer 2000 and training for use of this equipment from Brenda Mottle of the Earth Science Program at the University of Calgary is genuinely appreciated. NSERC and VOL. 37, NO. 12, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2705

Imperial Oil are gratefully acknowledged for financial support of this project.

Nomenclature AB

Alberta

AC

asphaltene coated

ASTM

American Society for Testing and Materials

BCW

wettable soil from Bruderheim, AB

BNW

water-repellent soil from Bruderheim, AB

DCW

wettable soil from Devon, AB

DNW

water-repellent soil from Devon, AB

EPW

pristine wettable soil from Ellerslie, AB

FA

fulvic acids

HA

humic acids

ID

Idaho

IHSS

International Humic Substances Society

IRC

illite-rich clay

KRC

kaolinite-rich clay

MED

molarity of ethanol droplet

NMR

nuclear magnetic resonance

SK

Saskatchewan

T2

transverse relaxation time, ms

T2gm

geometric mean of transverse relaxation time, ms

WDPT

water drop penetration time

Literature Cited (1) Roy, J. L. Ph.D. Thesis, University of Alberta, 1999. (2) Terry, J. P.; Shakesby, R. A. Earth Surf. Processes Landforms 1993, 18, 519.

2706

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 12, 2003

(3) Dekker, L. W.; Ritsema, C. J. J. Environ. Qual. 1995, 24, 324. (4) Ma’shum, M.; Tate, M. E.; Jones, G. P.; Oades, J. M. J. Soil Sci. 1988, 39, 99. (5) Dekker, L. W.; Ritsema, C. J. Water Resour. Res. 1994, 30I (9), 2507. (6) Tschapek, M. Z. Pflanzenernahr. Dueng Bodenkd. 1984, 147 (2), 137. (7) Aikman, M. J. L. Ph.D. Thesis, University of Calgary, 2001. (8) Ritsema, C. J.; Dekker: L. W. J. Contam. Hydrol. 1998, 31, 295. (9) Hornak, J. P. http://www.cis.rit.edu/htbooks/nmr/bnmr.htm (accessed July 1999). (10) Coates, G. R.; Xiao, X.; Prammer, M. G. NMR Logging: Principles and Applications; Halliburton Energy Services: TX, 1999. (11) Bellieveau, S. M. M.Sc. Thesis, University of Calgary, 1996. (12) Plainsman. http://ceramicsearch.com/plainsman/data/ native.htm (accessed February 2001). (13) International Humic Substances Society. http:// www.ihss.gatech.edu/ (accessed June 2000). (14) Straley, C.; Rossini, D.; Vinegar, H.; Tutunjian, P.; Morriss, C. Log Anal. 1997, 38 (2), 84. (15) Annual Book of ASTM Standards, Section 4, Vol. 04.03; ASTM D3279-97; American Society for Testing and Materials: Philadelphia, 2001. (16) Mirotchnik, K.; Kantzas, A. J. Can. Pet. Technol. 1999, 37 (11), 41. (17) Manalo, F. P. M.Sc. Thesis, University of Calgary, 2001. (18) Matteson, A.; Tomanic, J.; Herron, M.; Allen, D.; Kenyon, W. SPE Reservoir Eval. Eng. 2000, 3 (5), 408. (19) Prammer, M.; Drack, E.; Bouton, J.; Gardner, J.; Coates, G.; Chandler, R.; Miller, M. Proceedings of the 1996 SPE Annual Technical Conference and Exhibition; Paper SPE 36522. (20) Mirotchnik, K.; Kantzas, A.; Starosud, A.; Aikman, M. J. Can. Pet. Technol. 2001, 40 (7), 38. (21) Mirotchnik, K.; Kubika, P.; Randall, L.; Starosud, A.; Allsopp, K.; Kantzas, A. Proceedings of the 1998 Society of Core Analysts International Symposium; Paper 98-15. (22) Mirotchnik, K.; Allsopp, K.; Kantzas, A. Proceedings of the 1997 Society of Core Analysts International Symposium; Paper 97-03. (23) Todoruk, T. R. Ph.D. Thesis, University of Calgary, 2003.

Received for review July 14, 2002. Revised manuscript received March 10, 2003. Accepted March 15, 2003. ES0259685