Investigating the Effect of Texture on the Relative Benefit of Electroosmosis in Soil Remediation


The purpose of this experiment was to determine the influence of texture on the effect of electroosmosis (EO) on a hydraulic gradient imposed flow of an ionic solution through soil. It was hoped that by determining this, it would be possible to determine the effect of texture on the remedial potential of electroosmosis in a particular soil. A number of soil samples were gathered to provide a variety of attributes for a basis of comparison and analysis of trends. Soil samples were characterized for texture, pore space, particle density, and organic matter.

Using PVC pipe, a test chamber was designed to maintain an open-flow arrangement with a bottle used at the side of the anode to maintain a constant and relatively small external hydraulic gradient. The ionic solution used was a sodium chloride solution (3g/L).

For each test, an air-dried sample was loaded into the test chamber and the hydraulic gradient applied. Measurements of flow output were taken at ten-minute intervals for a total of four hours. Electroosmosis tests were run with a direct current of eight volts applied and control tests were run in the absence of electric current.

Of all variables, texture was found to be most directly related to the efficacy electroosmosis, with clay influencing EO negatively; the relationship identified substantiated the trend identified by the previous year’s study and contradicted the theoretical basis for electroosmotic efficacy.

Also available, in condensed + stylized form, in PDF.


Electroosmosis is one of a group of several “electrokinetic” phenomena which relate to the motion of charged electrolyte solutions and the motion of charged particles in solution or suspension; it is defined as “the movement of a liquid relative to a stationary charged surface (e.g., a capillary or porous plug) by an electric field.” (Probstein 195).

Since the discovery of these phenomena by F.F. Reuss in 1809, the potential of electroosmosis for various applications has been demonstrated numerous times, with electroosmosis being first utilized in the dewatering and stabilizing of soils in the 1930s. Electroosmosis has since been shown to have applicability in the separation of organic and inorganic contaminants, in construction of leak detection systems in clay liners, in the diversion of waste “plumes,” in the injection of nutrients or microorganisms into subsoil layers, in the increasing of pile strength, in the dewatering of foams, sludges, and dredgings, and in systems and barriers for lowering ground-water. The focus of this project, however, was the application of electroosmosis to the remediation of contaminated soil.

Electroosmosis and Soil

When in contact with an ionic solution the mineral particles present in soil give rise to a negative charge referred to as the soil’s zeta potential. This potential arises for many reasons, including chemical and physical adsorption within the soil and lattice imperfections. The negative charged surface of the particles attracts positive ions from the solution, creating a super-thin layer of water (sometimes referred to as double-layer or Debye sheath) in the pore walls. This layer of water moves much like a single ion under the influence of the electric field, and the viscous drag of the water causes the water in the pores to also gain a velocity component, generally in the direction of the cathode.

Soil Remediation

Despite the growing ecological awareness in society today, contamination of soil (and the resulting ground-water contamination) by both industrial and consumer chemicals is becoming a more prevalent concern. Minimization of chemical by-products and pollutants is obviously an important step in minimizing ecological impact, but identification of better methods of soil remediation is also an important focus given that conditions of our society will continue to favor the escape of certain levels of chemical pollutants into the environment regardless of what measures are taken to limit them.

Identification of methods for contaminant removal which could be most effectively be utilized given the conditions of a site and considering the resources available must take into account the nature of the pollutant and soil as well as economic factors, since prohibitive costs can curtail the implementation of a method regardless of the benefits of the method.

Though the study of electroosmosis for its potential in soil remediation has been largely limited to laboratory studies and very limited field tests, indications are that electroosmosis may prove to be an extremely valuable addition to conventional methods of soil remediation for both its efficacy and cost.

Application of Electroosmosis to Soil Remediation

The focus of this project was electroosmosis; however, in applying electroosmosis to soil remediation other electrokinetic phenomena are involved, and characterization of the movement of the contaminant only in terms of electroosmotic convection and electromigration is still inadequate to fully describe the motion of the contaminant. The chemical reactions that take place at the electrodes, in the fluid itself, desorption, dissolution, and chemical interactions within the soil itself all have bearing upon the contaminant’s motion.

In effecting the removal of soluble or slightly soluble contaminants from soil both electroosmosis and electromigration would be involved, electroosmosis effecting the movement of the purge solution through the soil and electromigration effecting the movement of the contaminant within the purge solution. In cases where electroosmosis would be the dominant mechanism, the percentage of contaminant removal would be directly correlated to pore volumes of purge solution flushed through the soil.

One group of weakly ionized compounds commonly associated with ground water contamination and which to which electroosmosis has great applicability is the aromatics, which includes such compounds as benzene, toluene, and xylene. These compounds are known to ionize within a range that should give rise to a sufficient zeta potential on the soil for electroosmosis to take place. Electromigration, unlike electroosmosis, is independent of zeta potential and soil particle size and depends primarily on the charge of the contaminant itself.

For compounds such as heavy hydrocarbons, which are essentially not ionized, the primary benefit of electroosmosis is restricted to purging the compound indirectly by pushing it with a water “front” in the soil unless surfactants can be utilized within the purge solution to solubize the contaminants.

Electroosmosis research has been largely limited to the laboratory, and only in certain locations have large-scale experiments been conducted in a natural setting. The majority of laboratory experiments concerning the remediary potential of electroosmosis has been limited to tests on kaolinite clay samples; one of the main goals of this project was to extend tests of electroosmosis past uniform samples to tests on soils of various texture.

Advantages of Electroosmosis

Although electroosmosis (in theory) should be more effective in soils with a lower hydraulic conductivity (i.e., soils with a greater percentage of micropores and that are more difficult to move water through by means of a pressure-driven device), this research sought to determine how great an effect texture has on the efficacy of electroosmosis by experimental means.

Conventional pressure-driven methods inevitably channel water through the larger-sized pores; electroosmosis-driven flow, however, should allow for a more uniform channeling of water (or purge solution) through pores and a much greater degree of control of the motion exhibited by the purge solution. Additionally, conventional methods of soil remediation rely primarily on various extraction processes which can range in cost from $50 to $1,500 per cubic yard of contaminated soil, compared to $2.00/ton of effluent removed by EO (Shapiro and Probstein 1993).


Quite a number of laboratory studies have been conducted in regards to the determination of the feasibility of electroosmosis as a method of soil remediation. Whereas a large proportion of these studies published in scientific journals have dealt with the nature of various contaminants, very few were concerned with the textural properties of the soil. Most used a kaolinite clay sample for use in electroosmosis tests, and those that did involve different textures were limited to very pure silt and clay mixtures. There is a large deal of theoretical basis for the assumption that electroosmosis would have a greater efficacy in soils of low hydraulic permeabilities (i.e., clays), however, the effect of electroosmosis is still veiled in a certain level of practical uncertainty, with most studies still restricted to laboratory studies.

Phase One

The first phase of this research conducted during 1997-1998 involved studying what relationships might exist between soil texture and the efficacy of electroosmosis in effecting the removal of organic contaminants from soil. This research attempted to identify a relationship by tracing the movement of tannic acid through test samples under the influence of electroosmosis. Tannic acid was chosen as the model for the experiment since it is relatively safe and more importantly because it is structurally similar to the aromatics, a group of organic chemicals for which there exists a potentially wide range of applicability for electroosmosis.

The boundary conditions for flow in this experiment were a closed electrode configuration, with a slight current applied across the test sample by means of a DC power supply. Samples were taken over time at three points along the test chamber in order to establish the change in concentration over time as effected by electroosmosis.

The data produced by this phase of the research indicated several potential relationships involving electroosmotic efficacy in the removal of organic contaminants from soil. Significant among these trends is the positive relationship of sand to electroosmotic efficacy; that is, as sand content increases, electroosmotic efficacy was found to also increase. This contradicts the theoretical range of applicability of electroosmosis, which is in fine-grained soils of low hydraulic permeabilities (i.e., soils with high clay content). To explore this contradiction, a second phase of research was devised.

Phase Two

The second and current phase of research was devised with several new considerations in addition to those made in the first phase of experimentation. One consideration was the process of adsorption of the contaminant by the soil; because concentration readings of a contaminant present in a soil sample could be drastically altered by adsorption of the compound, it was decided that a more accurate measure of electroosmotic efficacy could be determined by looking at electroosmosis-induced flow rather than contaminant migration. By removing the variable of the contaminant, it was hypothesized that a more accurate comparison of electroosmosis in various soils could be made with a wider range of applicability of the results.

This phase of the project attempted to measure electroosmotic efficacy by flow rate in order to obtain a more reliable measure of how texture affects electroosmosis. Also of interest is the apparent trend that the first phase of this research established, which seemed to contradict theory concerning the soils in which electroosmosis would have greatest applicability; this research seemed to indicate a correlation between sandy soils and greater electroosmosis efficacy, which was a primary reason for a continued and intensified investigation of this process. It was hoped that the data provided by additional research could better quantify the relationships indicated by the first phase of research.


Design and Construction of Test Chamber

The experimental set-up was designed to allow the effects of electroosmosis to be observed separate from other the electrokinetic phenomena associated with the processes that would be involved in the electroosmotic purging of a contaminant through soil; for this reason, the purpose of the test chamber was to allow measurement of water flow through the soil. The test chamber was designed to maintain an open-flow arrangement, with a bottle used at the side of the anode to maintain a constant and relatively small external hydraulic gradient.

The main body of the test chamber was designed to separate into two pieces to allow easy loading of samples and was based on a 2″ PVC pipe. Also comprising the main body were a cap, coupler, and reducer (2″ to ¾”). The hydraulic gradient from the elevated salt solution was maintained from a bottle and 2″ coupler through a series of connecting pieces of PVC pipe. The entire apparatus was also designed with the intention that unused salt solution in the bottle at the conclusion of a test could be reclaimed for use in later tests.

An outlet for the water was drilled through the cap at the cathode end of the electroosmosis test chamber with the hydraulic gradient being supplied to the sample through a reducer on the anode side of the chamber. Electrodes for the chamber consisted of iron washers that were placed in the cap (at the cathode end) and the reducer (at the anode end). These electrodes were connected to the external power supply (8 volts, DC) by means of metal screws that contacted the washers and breached the walls of the test chamber.


Characterization of Soil Samples

Soil samples with a wide range of attributes for a basis of comparison and analysis of trends were gathered. Each sample was characterized by its textural grouping, pore space, bulk and particle density, and organic matter content.

Texture was determined for each sample by means of a texture test kit which separated the soil into its three basic mineral fractions of sand, silt, and clay based on the amount of time required for each to settle in solution.

Fifteen milliliters of a soil sample were added to one 50-mL soil separation tube, and the tube was tapped firmly on a hard surface to eliminate air spaces. One milliliter of Texture Dispersing Reagent was added to the tube, and the sample was then diluted to 45 mL with water. The tube was capped and shaken for two minutes, and then allowed to settle for 30 seconds. The solution from the tube was poured into another tube and allowed to settle for 30 minutes. The amount of sediment in the first tube was divided by the initial amount of soil (15 mL) to calculate the percentage of sand in the soil. The amount of sediment in the second tube was then divided by the initial amount of soil (15 mL) to calculate the percentage of silt in the soil. The percentage of clay in the sample was calculated by subtracting the sum of the other two percentages from 100 %. This was done to obtain a more accurate reading of clay percentage than the measurement of the volume of clay, as the colloidal nature of clay causes it to swell in water. This procedure was repeated for each soil sample.

Pore space was determined by gradually adding water to a dried soil sample of a known volume. The volume of water used by the point at which the soil reached saturation was noted, and this volume was divided by the total volume of the soil sample to determine percent pore space.

Density for each sample was determined by weighing a known volume of a soil sample. Bulk density was calculated as the sample’s overall density while particle density was calculated by factoring out the volume of the pore space in the soil.

Percent organic matter content for each soil sample was determined by heating a small oven-dried sample of the soil over a Bunsen burner to a constant mass. Though the values obtained by means of this method are approximate, the technique is accurate enough to give a fairly good relative indication of organic matter present in the samples.

Testing of Soil Samples

A double-chamber set-up was devised to allow simultaneous testing of electroosmosis-affected flow and control flow, which consisted of an identical set-up without the current applied.

Each half of the test-chamber was loaded and the two halves were joined. To minimize leakage through the coupler, electrical tape was utilized to seal the chamber. A weak purge effluent of sodium chloride was prepared to a concentration of 3 g/L (~ 0.05 M). The hydraulic gradient was then applied; for the electroosmosis tests, a direct current of 8 volts was applied via the external power supply, while the control tests maintained the hydraulic gradient in the absence of electric current. Volume readings were taken at ten-minute intervals to determine the amount of effluent discharged at the cathode and individual tests were carried out for a total of four hours. The comparatively short duration of the tests was intentional to minimize the chemical interactions within the soil during the tests.


Many factors are known to have an effect on the potential benefit of electroosmosis in soil remediation, and for this reason many recent studies have investigated numerous variables involved in the process including electric potential; contaminant type, concentration, and distribution; and electrode material, arrangement, and configuration. The focus of this investigation was soil texture, a variable given little consideration or attention outside the range of what is theoretically assumed to be the applicable range of electroosmosis (i.e., very fine-grained soils-specifically, soils with high clay content). Other variables examined in this investigation for their potential influence on electroosmosis were pore space, particle density, and organic matter. Several variables of significance to the electroosmotic flow which were kept constant during this experiment were voltage, concentration of the ionic purge solution, temperature of soil sample, hydraulic head, and electrode configuration.

This experiment was designed with the intent that relative electroosmotic efficacy in various soils could be determined by measuring the change in flow rate of an ionic solution through a sample of the soil; it was expected that electroosmosis would impose an additional velocity vector on the advection of the effluent through the test chamber. It was hoped that this focus would yield a more quantitative understanding of the relationships existing between soil texture and electroosmosis than the previous year’s research, which focused on contaminant migration.

One difficulty in interpreting the data was that the flow varied greatly even within several tests of a single sample. Despite this variability, it was found that third-order approximations for the data of total flow over time had high R2 values: of 67 tests conducted overall, R2 values for third-order polynomial trendlines of only three tests fell below 0.98. Another observation of note to this research was that a slight disturbance of the test chamber during testing could result in a radical change of the flow rate; this suggests that minor changes within the structural character of a soil sample could have disproportionately large influences on flow through the medium, contributing to the idea that radical variations of flow through various tests of a single soil sample might be expected. This variability of flow rate was one reason that a large number of trials for a relatively small number of soil samples was favored over a smaller number of trials for a larger number of soil samples; while a larger number of soil samples would allow for a more diverse texture base, it was not felt that the results obtained would reflect the true relationships as well as the average of a large number of trials of a smaller number of soil samples would.

Flow rates achieved during electroosmosis tests averaged significantly less than those of the control tests did, and overall flow rates for both control and EO tests decreased over time. Additionally, the decrease in flow rate for EO tests was initially found to average significantly greater than the decrease found in the control tests (i.e., flow rate in EO tests decreased more rapidly than flow rate in control tests). The overall decrease in flow rate for both sets of tests can be partially attributed to the saturation of the soils over time by the effluent solution. The additional decrease in flow rate for the EO tests presented a difficulty for analyzing the data collected, since it was originally hypothesized that the relative strength of electroosmotic influence on flow rate in various soils could be determined by calculating the proportion of increase in flow rather than decrease. This decrease could be attributed to the effects of the electrode reactions on the pH of the soil (to decrease pH at the anode and increase pH at the cathode), which would tend to have a negative effect on the flow rate over time (Shapiro and Probstein 1993).

While this offers an explanation for the initially more rapid decrease in flow rate (with both soil saturation with the purge solution and electrochemical changes within the soil contributing to a greater decrease in flow rate than just soil saturation), this does not account for the difference between control and EO flow rates initially. The significance of this observation is also uncertain, and additional experimentation may be necessary in order to explain the origins of this effect. Hamed et al. noted a delay in the initiation of EO-induced flow rate in experiments designed to measure EO-induced flow over time, though reasons for this delay were not hypothesized (1991).

For the purposes of this investigation, comparative values for trend analyses were obtained by averaging the differences between control and EO flow rates and change in flow rates (referred to in this paper as “rate change index” [RCI]). The convention used for this was to subtract the EO values for rate and rate change from the control values.

Textural composition of each of the samples was represented in several ways, in an attempt to give an accurate analysis of the correlations existing between texture and electroosmosis. Texture was broken into its distinct particle size groupings and the rate change index was plottted against each of these individual fractions (i.e., separate graphs were created that plotted rate change index against percent sand, percent silt, and percent clay). For an additional graph, samples were given numerical rankings based on their location on the textural triangle, with sand being given a ranking of 1 and sandy clay loam being given a ranking of 4; following this convention, however, resulted in data being grouped into only three categories (2, 3, and 4 corresponding to loamy sand, sandy loam, and sandy clay loam).

In an effort to consolidate data and provide means for better analyzing trends, averages were obtained for final RCI values (taken from the last five data points of 200, 210, 220, 230, and 240 minutes) and plotted against texture. Lower [negative] values would correspond to data points for which EO rate change exceeded control rate change and EO flow decreased less rapidly than control flow. Lower values would therefore seem to indicate samples for which flow rate was most strongly influenced by electroosmosis.

The apparent correlation between sand content and RCI index was of a lesser RCI index with higher sand content; this would suggest that electroosmosis had a stronger influence on the flow rate established by advection in soils of high sand content. Additionally, negative relationships were indicated between silt and RCI index and clay and RCI index, suggesting that higher clay and silt content would have an adverse effect on EO in application to soil remediation. The R2 value for the correlation between clay content and the RCI index was significantly greater than the correlation between silt and the index.

While the R2 values produced for these relationships are relatively low, it is significant that they reproduce the results indicated by the first phase of this research, which also indicated a correlation between sand content and electroosmotic efficacy as well as between clay content and a decrease in electroosmotic efficacy. Theoretical basis for the EO phenomenon is of higher efficacy in fine-grained soils, a trend opposite that indicated by the data produced by this research. For this reason, the data produced by this research appear to contradict the theoretical basis for EO efficacy. Additional significance can be found in the duplication of these results in two separate experiments of different experimental design (phase one and phase two).

Several possibilities have been identified by this researcher for the existence of this contradiction found in the data.

While previous studies have been devised to investigate the effects of a variation of texture within a limited range of silt and clay mixtures, none have studied the effects of texture within the range studied by this project. For this reason, it is possible that EO efficacy increases at the extremes of the textural triangle. In order to investigate this possibility, it may be helpful to obtain larger quantities of soil samples for purposes of mixing; with only a minimally larger sample base of soil samples, it would then be possible to create a much wider range of textures for purposes of testing by mixing samples.

The possibility also exists that the discrepancy between the data and the theoretical applicability of electroosmosis is linked to the relatively short duration of trials in this experiment. While future research may want to focus on flow over a more extended time period, it was felt that for the purposes of this particular experiment more exact measures of rate and change in rate (obtained by more frequent readings of flow volume) would be desirable over less frequent and less accurate measures of rate and change in rate.

Finally, there exists the possibility that theoretical models of electroosmotic efficacy in application to soil remediation do not serve as accurate representations of the processes at work. Though many studies have been conducted recently to investigate electroosmotic phenomenon, the processes at work are still not very well understood and the potential exists that present theoretical models may be in error.

The remaining variables in addition to texture were considered by this research were systematically assumed to be the independent variables in order to establish whether a stronger correlation existed between the strength of EO-induced flow and a variable other than texture (table follows). Based on the data, however, clay content remained the variable with the greatest likelihood of having a direct effect on EO-flow.


Based on the findings of the previous year’s research, this phase of experimentation sought to refute or quantify the correlations between texture and electroosmotic efficacy that were indicated. It is significant that both phase one and two of this research indicated the same potential trend of clay content having a negative influence upon electroosmotic efficacy. Additionally, it is important to note that this trend is contrary to the theoretical range of greater electroosmotic-applicability, which is of clay content to greater efficacy. Three possibilities were suggested to explain this contradiction-a greater efficacy with extremes of sand and clay content, a strong time dependence which differs over more extended periods of time greatly from the initially established trends, or an inaccuracy of theoretical models at explaining the true processes taking place under the influence of electroosmosis.


Acar, Y. et. al. Phenol removal from kaolinite by electrokinetics. J. Geotechnical Eng. 118: 1837; 1992.

Anderson, J.; Idol, W. Electroosmosis through pores with nonuniformly charged walls. Chem. Eng. 38: 93-106; 1985.

Bouma, J., Brown, R.B., and P.S.C. Rao. “Movement of Water: Basics of Soil-Water Relationships – Part III.” University of Florida Cooperative Extension Service. (11-Nov-97).

Bruell, C. et. al. Electroosmotic removal of gasoline hydrocarbons and TCE from clay. J. Geotechnical Eng. 118:68; 1992.

Buttler, Tasha, Martinkovic, Wendel, and O. Nesheim. “Factors Influencing Pesticide Movement to Ground Water.” University of Florida Cooperative Extension Service. (11-Nov-97).

Catsimpoolas, Nicholas, ed. Isoelectric Focusing. New York: Academic Press, 1976.

Chapelle, Francis H. The Hidden Sea. Tucson, AZ: Geoscience, 1997.

Hamed, J. et. al. PB(II) removal from kaolinite by electrokinetics. J. Geotechnical Eng. 117: 241-271; 1991.

“Introduction to Soil Water Potential.” Department of Agronomy and Horticulture, New Mexico State University. (14-Nov-97).

M.Mizoguchi, T.Ito and K.Matsukawa. Movement of water and ions in frozen clay by electroosmosis.

Probstein, Ronald F. Physicochemical Hydrodynamics. New York: Wiley, 1994.

Removal of contaminants from soils by electric fields. Science. 260:498; 1993.

Segall, B. et. al. Electroosmotic contaminant-removal processes. J. Env. Engineering. 118: 84-100; 1992.

Shapiro, A. and R. Probstein. Removal of Contaminants from Saturated Clay by Electroosmosis. Envirn. Sci. Technol. 27: 283-291; 1993.

“Soil Bulk Density.” (12-Nov-97).

“Soil Texture.” (12-Nov-97).

Webber, M.D. and S.S. Singh. “Contamination of Agricultural Soils.” Soil Health. (12-Nov-97).


(This is more or less the list of acknowledgments posted with my project for the 1998-1999 fair)
I will say that I feel that my independent science projects over my four years of high school have been an enormous success and that I have quite a number of people to thank who have make this possible:

  • I would first and foremost like to thank my parents for, well, everything.
  • Thanks must also go out to my grandparents for support among other things.
  • Profuse thanks and applause to my science-fair adviser of two years, Mrs. Patricia Wee, whose amazing encouragement and influence have been of more help to me than I could possibly imagine, I am sure.
  • Thanks also go out to Mr. Larry Hess, science department head at my high school, whose support and encouragement have also led my ideas to become realized through my science fair projects.
  • I would like to thank Mr. Wilmer Nolt, whose seemingly endless energy and encouragement have also served as an inspiration to me and allowed me to make my projects a reality.
  • I would be remiss if I were not to acknowledge the help of Colleen LeFevre, whose insight and ideas have helped to direct me in my studies of electroosmosis.

Additional thanks to (in no particular order):

Wendy Zwally; Mr. James Collier; Mr. Bruce Lasala; Mr. Wayne Boggs; Mr. Michael Vavreck; Mr. Pat Ross; Mrs. Tamme Westbrooks; Mr. Michael Eyster; Mrs. Filitea Dean; The numerous teachers who have had enough foresight and understanding to allow classes to be missed “in the name of science”; Numerous colleages/classmates, especially fellow independent science “associates”; Science surplus catalogs; A hearty thanks to velcro, foam-board, double-stick tape, and formica.