WATER AM) ION MOVEMENT BY ELECTROCSM3SIS IN WYOMING BENTONITE By HYDE S. JACOBS A THESIS Submitted to the School for AdvancedGraduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR CF PHILOSOPHY Department of Soil Science 195? ProQuest Number: 10008582 All rights reserved INFO RM ATIO N TO ALL USERS The quality o f this reproduction is dependent upon the quality of the copy subm itted. In the unlikely event that the author did not send a com plete m anuscript and there are m issing pages, these will be noted. Also, if m aterial had to be removed, a note will indicate the deletion. uest. ProQ uest 10008582 Published by ProQ uest LLC (2016). C opyright of the Dissertation is held by the Author. All rights reserved. This w ork is protected against unauthorized copying under Title 17, United States Code M icroform Edition © ProQ uest LLC. ProQ uest LLC. 789 East E isenhow er Parkway P.O. Box 1346 Ann Arbor, Ml 4 8 1 0 6 - 1346 WATER AND ION rOVEMENT EY ELECTROCS ICS IS IN WYOMING BENTONITE by ityde S. Jacobs AN ABSTRACT Submitted to the School of Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Soil Science Year Approved ^ 1957 The effect of electrical treatment of soil has been of interest in the drainage, reclamation, and stabilization of soils for some time* This study was undertaken to characterize more coupletely the effect of an electric field on water and ion movement in Wfyoming Bentonite* The clay was hydrogen saturated and freed from anions by resin treatment. Neutralization of the hydrogen clay was acconplished by add­ ing the appropriate hydroxide or oxide. The bentonite-sand mixtures used in this stucty’ contained 5 per cent clay and 95 per cent sand. Electroosmosis was carried out in cells made of lucite tubing and plexiglass which were designed so that both water and ion movement could be determined. Silver;silver chloride electrodes were used to study the initial flow rates, and platinum electrodes were used to stucty water and ion movement under continuous current flow. In sodium and calcium bentonite-sand mixtures the water flow rate increased when the percentage moisture was increased. The effect of the kind of exchangeable bases and the per cent base saturation on the flow of water was marked. Higher moisture contents were required in a sanple containing 0.76 symmetry of sodium than in a sample containing 0.98 sym­ metry of sodium in order to produce the same rate of flow. At equal moisture contents the flow rate in sanples containing exchangeable cal­ cium were higher than in sanples containing exchangeable sodium. When the systems contained complementary ions (other than hydrogen or alumi­ num) , the rate of flow in the early stages of electroosmosis was much higher than if the sauple contained only a single species of ion. The comparative rates of removal of the ions under electroosmosis was as follows: sodium > potassium >raagnesium ^calcium. The rate of removal of sodium and potassium was found to be directly proportional to the amount of these ions remaining in the system. In the early stages of electroosmosis the transference numbers of sodium and potassium in unsaturated systems were very near unity, where­ as the transference number of calcium in both saturated and unsaturated systems was near that of the calcium in 0,01 normal calcium chloride. Evidence is presented which indicates that calcium is bonded by two different energies of adsorption by bentonite. That calcium held be­ tween zero and 30 per cent saturation is held strongly and is removedby the electroosmotic process very slowly. calcium held by the ben­ That tonite when the per cent saturation exceeds 30 per cent is held less strongly and is removed at a faster rate. The experimental work involving the movement of ions indicated that there was differential movement of ions through the bentonite-sand mix­ ture. It was demonstrated that this differential movement tends to sep­ arate the ions or to deplete the cations in bands between the electrodes. ACKNOWLEDGEMENTS The author wishes to express his sincere thanks to Dr. M. M. Mortland under whose able supervision, inspiration and en­ couragement this stu) are taken to prevent it. Usually this would not be practical on a large scale basis. Helmholtz is credited with being the first to work out a math­ ematical theory explaining electroosmosis on the basis of the electric double layer. Gouy later made important contributions in clarifying ^-Numbers in parenthesis refer to literature cited in the bibliography. the nature of the double layer by pointing outits diffuse character, Taylor and Glasstone (33) have discussed these theories with respect to electroosmosis and Bolt(6) has discussed them with respect to cation exchange reactions* Taylor and Glasstone use the following equation to describe electroosmotic flows (I) v = f n i .. li7T fj L* Where V is the volume ofliquid transported per second, f is potential, D is the dielectric the zeta constant in the double layer, I is the c u r r e n t , i s the viscosity, the L 1 is the specific conductivity of the liquid in the capillaries. Units are in the Centimeter-Gram-Second system. According to the equation the amount of the flow is independent of the length or radius of the capillaries, and is directly proportion­ al to the zeta potential and current but inversely proportional to viscosity and specific conductivity. In another form (3U) equation (I) can be written: (II) V = 0 AirTp. Where Q is the cross section of a bundle of capillaries or diaphragm, 1 is the distance between electrodes, and E is the potential. From equation (I) and (II) it can be seen that either voltage or current can be utilized to describe electroosmotic flow. Winterkorn (37) points out that values for the dielectric constant, D, and the viscosity, , in the vicinity of an electrical charge are likely to vary considerably from values given in the literature. Since the zeta potential cannot be determinded by itself (3U) but is depend- h ent on the values chosen for D and , Winterkorn suggests that it would be more scientific and practical to use the factor D g as a 5 /J whole and to e:xpress it as = ke. The equation for electroosmotic ITTi flow derived by Winterkorn is given belows (III) Vp « n k E Where Vp is the volume of liquid moved in unit time through a unit cross section of a porous system, n is the porosity of the system and k « kg. Winterkorn points out that equation (III) for describing electro­ osmotic flow is of the same form developed by Darcy describing flow under a hydrostatic head and it is therefore of particular interest from the standpoint of those familiar with hydraulic flow through porous media, Casagrande (9) presents an equation describing electroosmosis which is based on the Helmholtz theory of the double layer. This equation is similar to that of Winterkorn and is given below: (DO V « K ie A Where K is the coefficient of electroosmotic permeability, ie is the electric gradient, and A is the cross sectional area of the soil times the per cent porosity. The value of K for most materials was found to be about 5> X 10”^ cm/sec, per volt/cm. Casagrande also noted the development of fissures in samples subjected to electroosmosis. This was attributed to stresses developed by the water although Bernatzik (it) disagrees with this. The cracks spread outward from the cathode where water accumulated during the electroosmotic process. Schaad (32) felt that the formulas deduced by Helmholtz neglected several important influences. He states: The former and later experiments of numerous physicists (e, g.) Quincke, Jllig and Schonfeldt) on the electroosmotic flow through capillaries and diaphrams show, that this flow decreases with increasing diameter of capillaries or voids. The formulas of Helmholtz mentioned in L, Casagrandes report give the opposite effect. From a capillary of infinite diameter an infinite electro osmotic discharge would result. This effect is contradictory to every test result. Geuze and his associates (16) and Schaad (32) have done experiment al work and developed formulas for determining the coefficient of electroosmotic permeability, K. After prolonged electroosmosis the value of K decreases. In a later paper Winterkorn (38) pointed out various shortcomings of the Helroholtz-Smoluchowski theories for describing electroosmotic flow when the effective pore radii are very small. In this same paper Winterkorn tested experimentally equations for describing elec­ troosmosis based on the Schmid theory for a microporous system. This theory is based on the assumption that the counterions are distributed evenly throughout the entire liquid volume rather than in a simple or diffuse double layer. The equation for describing electroosmotic flow derived by Winterkorn for a swelling system based on Schmid*s theory is as follows: (V) = a (1-n)^/^ n3 Where kw is the volume of liquid moved per second across a square centimeter section of system under an electric potential of 1 volt per centimeter, and A is a constant which is a function of the vis­ cosity, exchange capacity, and form of the particles. According to the theory ky. increases with increasing porosity to a maximum in the vicinity of n = 0,8 and then falls rapidly with further increases in the porosity. Winterkorn demonstrated experiment­ ally that the k^^-n relationship showed maxima for kaolinite and bentonite in accordance with Schmid1s theory but in contradiction to that of Helmholtz. However, hydration of the clay is not taken into account in the theory so that the experimental data deviated somewhat from the theoretical. Electroosmosis has been applied to soils for various purposes. Endell and Hoffman (15) utilized an aluminum anode and a copper cathode for electrodes in the electrical treatment of a clay sarrple. After electroosmosis had been performed, the sample neither swelled nor disintegrated during several months immersion in water. The chemical changes were thought to be due to exchange reactions with aluminum, hydrogen, and sodium. Bauxite (A^OgOHgO) was formed on the aluminum anode. Dawson and McDonald (lIj) studied the influence of electroosmosis on the consolidation of soil materials. It X\ras found that the passage of direct current through the soil accelerates the rate and amount of consolidation. Geuze et. al. (16) points out that hardening occurs when aluminum or iron electrodes are used and that little hardening occurs when inert electrodes such as platinum or carbon are used. Bernatzik (U) reports that hardening of the soil material is a result of chemical changes rather than the movement of water, but that hard­ ening of the soil is possible only if a supply of the pore water is prevented. Marwick and Dobson (21) reviews the application of electroosmosis to stabilize soil for excavations incidental to the building of a 7 railroad track and a U-boat pen. The electrodes were placed in such a way as to direct the water flow away from the excavation. Preece (26) also reviews various applications of electroosmosis to practical situations encountered in engineering problems. No reports in the literature concerning the removal of ions by electroosmosis have been found, but the process is similar in many ways to that of electrodialysis which has been investigated more thoroughly than electroosmosis. However, in electroosmosis the ions have to travel relatively large distances through the soil mass before being released at thecathode whereas in electrodialysis the release is effected quickly due to the short distances the ions must travel and the large amount of solution that is present. In this regard Lo’ddesol (19) found that the initial amount of ions removed from a small sample was greater than the amount removed from a large sample. Bradfield (7) and rfettson(23) have shown that electrodialysis removes mainly the exchangeable ions and Puri(2Q) has charted the relative rates of release of various ions by electrodialysis. Several authors (2U) (27) (35>) (36) have found that less magnesium is removed by electrodialysis than by replacement with various salts. (2I4.) attributed this to the stability of the silicate. Jfettson Contrary to these results, Salagado and Chapman (31) obtained a similar amount of magnesium release by both electrodialysis and by replacement with normal ammonium acetate. ?■ One of the drawbacks to electrical drainage of soils is the relative high cost (9) of performing electroosmosis. Barber (2) reports the electrical requirements for controlling groundwater around several excavations during construction. Woodward and Miller (39) review the literature on electroosmosis for those interested in electrical drainage of soil. They list some of the points which need further clarification before the value cultural drainage can be determined. of electroosmosis for agri­ These points are summarized below: 1. What are the economic costs and optimum power requirements? 2. Is it necessary to place the entire poorly drained area in the electric field or can an electroosmotic pressure be established which will create a hydraulic gradient outside the electric field? 3* Will the electrical potential increase or decrease the hydraulic permeability over a long time basis. U. What should the electrode composition be, the duration of the run, current consumption, etc.? Additional laboratory and field experiments are required before these questions can be fully answered. PREPARATION CF THE ELECTROOSMOTIC MEDIA A 1.5 per cent suspension of Wyoming Bentonite was passed through a column of Araberlite Resin IR-120 to remove the exchangeable bases and to saturate the clay particles with hydrogen* thus prepared was about 2*8. The pH of the suspension The low pH was caused by small amounts of hydrochloric and sulfuric acid which were present* To remove these acids and any other free anions present the hydrogen saturated clay was then shaken with Amberlite Resin IR-U-B. After this treatment, the pH rose to about 3.25 and the conductance fell to about 200 raicromhos/centimetec The anion free and hydrogen saturated bentonite was then centrifuged in a Sharpies centrifuge to remove excess water* Neutralization was carried out by adding the appropriate oxide or hydroxide. Sodium and potassium hydroxides were added in solution but powdered calcium hydroxide and magnesium oxide were utilized for neutralization by these cations* The neutralization reaction was very rapid when calcium hydroxide was added but took approximately two days when magnesium oxide was used. The ra­ pidity of the neutralization reaction ■was judged by following the pH changes which took place. The electroosmotic media was prepared by mixing the resin treated and neutralized bentonite with Ottawa silica sand so that 5> per cent ben­ tonite was present in the mixture. The bentonite was kept moist at all times once it had been resin treated. The cation exchange capacity of the clay as it was effected by var­ ious treatments and neutralizing agents is shown in Table I. All values 10 were determined conductimetrically with a Serfass Conductivity Bridge model RCM15* Note that the cation exchange capacity of the resin treat­ ed clays was 88 milliequivalents/100 grams of clay whether the hydrogen saturated clay was titrated with sodium .hydroxide or the barium satura­ ted clay was titrated with magnesium sulfate* The barium saturated clay which had neither been resin treated nor electrodialized also had an ex­ change capacity of 88 milliequivalents/100 grams of clay* This is very Table I The effect of method of preparation and neutralizing agent on the exchange capacity of looming Bentonite* Exchange Capacity NaOH CO CD H-saturated by resin; Neutralized with Na, Ca and Mg; H-saturated by TCI. Neutralizing Agent * Method of Preparation H-saturated by resin; Ba-saturated by BaClp. MgSO^ 88 Ba-saturated by BaClg MgSO^ 88 H-saturated by electrodialysis NaOH 105 H-saturated by electrodialys is Ca(OH) 2 125 H-saturated by resin; Aged 3 months. NaOH 5U H-saturated by resin; Aged 3 months. Ca(OH) 2 56 ■^Milliequivalents/100 grams of bentonite. interesting in view of the fact that the electrodialyzed clay gave very different exchange capacities when neutralized with calcium hydroxide (125 milliequivalents/100 grams) and sodium hydroxide (105 milliequivalents/lOO grams) . 11 Aging the hydrogen saturated clay for approx innately three months materially reduced the exchange capacity as determined by neutralization with sodium hydroxide and calcium hydroxide. At the end of the three month period the exchange capacity of the clay was 5>6 milliequivalents/ 100 grams and 5h milliequivalents/100 grams when titrated with calcium hydroxide and sodium hydroxide respectively* The potentiometric and conductimetric titration curves for the first sanple reported in Table I are shown in Figure 1* The titrations were carried out immediately after the sanple had been washed free of chlorides sifter being hydrogen saturated with hydrochloric acid. 280 10 flj •P C o ft) ■g-Sil6o C to o o o x 20 6o 100 12 Milliequivalents of Na/lOO grams Mill of clay Figure 1. Potentiometric and conductimetric titrations of a resin prepared bentonite which had been neutralized with a mixture of Wa, Ca, and Mg hydrox­ ides and then subsequently H-saturated with HOI. ' A 12 single end point was obtained for both conductimetric and potentiometric titrations in this case. Upon aging the hydrogen saturated clay becomes appreciably satura-r ted with aluminum as is shown by the conductimetric and potentiometric titration curves in Figures 2 and 3* These sanples were hydrogen satu­ rated by resin treatment and then aged three months. The presence of aluminum on the exchange complex is indicated by the appearance of an additional end point in the titration curves. After aging three months these samples were approximately 80 per cent saturated with aluminum and 20 per cent saturated with hydrogen. 10 9 8 7 6 5 h 20 U0 60 80 Milliequivalents Na/100 gram of clay Figure 2 0 20 1*0 60 80 Milliequivalents Ca/100 gram of clay Figure 3 Figures 2 and 3. Conductimetric and potentiometric titrations of H-bentonite with NaOH and Ca(0H)2 respectively three months after clay preparation. 3 EXPERIMENTAL PROCEDURE AND APPARATUS The basic cell used throughout this study is pictured diagram ma tic ally in Figure It, The cell was made from lucite tubing which had an opening 5.5 centimeters in diameter and walls 0,5 centimeter thick. The sample ring was 2.5 centimeters deep and the rings forming the anode and cathode chambers were also 2.5 centimeters deep. ple holder was 59. U milliliters. The volume of the sam­ The anode and cathode chambers were cemented to perforated plexiglass plates and the sample ring was bolted o CD CD CD CD CD CD Figure U. Diagrammatic sketch of cell used for electroosmosis. A and A*s siphons; B and B 1: rubber stoppers; C and C 1: cathode and anode chambers respectively; D and D 1: outlet tubes fitted with stopcocks; E and E*s cathode and anode respectively; F and F 1: 5/32 inch holes in plastic plate; Q and G*: filter paper membranes; H: sauple chamber; Is lucite ring. between them. The moist sanple was packed into the sample ring, H, with a spoon so as to exclude as many air bubbles and to obtain as much uniformity as possible* The samples used were bentonite-sand mixtures containing 5 per cent bentonite* The mixture was held in the sanple ring by membranes, G and G», made out of Whatman No* 50 filter paper. The filter paper mem­ branes were held rigid by clamping them between the sample ring and the 0.25 inch plexiglass plates, F and F 1, which were perforated with 5/32 inch holes to allow the passage of solution. The membrane proved very- satisfactory except at moisture values high enough so that the mixture would flow. The sanple holder was sealed to the plexiglass plate by brushing hot paraffin around the edges of the sample ring and allowing it to cool. The electrode chambers were conpletely filled with liquid so that any water flowing into the cathode chamber, C, would result in a like amount of water flowing into a weighing bottle through the siphon, A. The siphon, A 1, from the anode chamber was immersed in a beaker large enough so that any change in the hydraulic head during the run would be negligi­ ble. The electrode chambers were fitted with stopcocks and stoppers at D and D* to facilitate filling and gas removal. The electroosmotic flow was measured by weighing the over flow from the cathode chamber siphon. For uniformity the overflow was allowed to drip from the siphon and all samplings were ended immediately after the falling of a drop. It was observed in this laboratory that the flow in both closed and open end capillaries was often erratic and for this rea­ son weighing the overflow was thought to be a better method for measur­ ing electroosmotic flow. Biefer and IYfeison (5) have pointed out that there is iipedence of the bubble in a capillary flowmeter and Ballou (1) 15 used open end capillaries to overcome this effect. In some of his ex­ periments Ballou determined the amount of flow by weighing the overflow from the cathode chamber. The capillary flowmeter can only be used for very short runs due to the small volume of the capillary whereas by us­ ing the overflow method the volume of flow is not limited. A diagram of the circuit used to supply direct current to the cell is shown in Figure 5* All determinations were run in duplicate with the duplicate cells connected in series as is shown in the circuit diagram. The direct current power supply was obtained from E. H. Sargent and Co. D C POWER SUPPLY 0 - 4 0 0 VOLTS O-IOO MA Figure 5* Electrical circuit used to supply direct current for electroosmosis. R^: 10,000 ohm resistance, R2? 1,000 ohm resistance, and C2: electroosmotic cells, At milliameter. and the voltage could be varied from zero to ij.00 volts. rent output of the unit was 200 millianperes. The maximum cur­ By means of varying the voltage or the resistance across the shunt it was relatively simple to keep the rate of current flow constant if this was desired. Short Runs: Water Movement Only For short run water flow studies in itfhich the total electricity drawn did not exceed 5 coulombs, silver;silver chloride electrodes were utilized so that the flow of water could be determined accurately with­ out interference from the products of electrolysis. During the short runs the amount of current drawn was measured with a milliameter and the current was kept constant at 0.5 millianpere. The silver;silver chloride electrodes were made from silver screen by connecting the screen as the anode in 0.1 normal hydrochloric acid with a platinum cathode and drawing 50 milliamperes of current for three hours. To keep the electrodes reversible and to minimize chemical changes on the clay during the short runs, the electrode chambers were filled with 0*001 norml electrolyte. the The electrolyte was chosen so that cation would be the same as that adsorbed on the clay and the anion was always the chloride. During the short runs the electrode chambers were sealed except for the siphon openings to insure that the water flowing from the siphon would be equal to that flowing into the cathode chamber from electrosmosis. The terrperature was kept at 2ii° centigrade plus or minus one de­ gree. Long Runs; Water and Ion Movement In the longer runs where both the water flow and the removal of cat­ ions were being studied, platinum electrodes were used. The platinum e- lectrodes used were perforated sheets of platinum cut in a circular shape so that they would just fit inside the lucite ring which formed the elec­ trode chamber. The openings at the top of the electrode chambers, D and D 1, were left open to the air to avoid any gas pressure build up duetoelectrolysis. Because the electrode chambers were not closed, the siphons were arranged to maintain the water level at the top of the cell. The electrode chambers were filled with distilled water for nearly all runs. In a few cases dilute hydrochloric acid was placed in the electrode chambers for special reasons that will be noted later. The a- mount of cation movement was measured by chemical analyses of the solu­ tion in the cathode chamber after it had been combined with the overflow and rinse water. The cells were rinsed by refilling the chamber and al­ lowing the rinse water to drain. The solution in each electrode chamber was renewed at each sampling period but that from the anode chamber was discarded xjithout any analysis being performed. When the removal of calcium or magnesium was being studied, boiled distilled water was used in the electrode chambers, the opening to the cell was protected from the entrance of carbon dioxide by a column of ascarite, and dilute hydrochloric acid was used in one of the rinses if necessary to assure complete recovery of the calcium or magnesium. In a few cases dilute hydrochloric acid was used to fill the electrode cham­ bers instead of water. These precautions were necessary due to the low solubility of the hydroxides of these elements. Sodium and potassium were determined on a model 52 A Perkin Elmer Flame Photometer. Calcium and magnesium were determined volumetricallv according to a procedure outlined in the U.S.D.A. handbook, Saline and Alkali Soils (29). Exchange capacities of the mixtures were determined conductimetrically using a method modified from the one worked out by Mortland and Mellor (2ij) . In some cases the sauples were leached with 0.5 normal hydrochloric acid instead of barium chloride and the titra­ tion carried out with sodium hydroxide instead of magnesium sulfate. This procedure facilitated the determination of sodium and potassium on the flame photometer. The circuit which was used to supply direct current for electroos­ mosis was the same as is given in Figure 5 except that an iodine 18 coulometer (12) -was placed in series with the electroosmotic cells. In the coulometer iodine is liberated according to the amount of current passing through a potassium iodide solution. The amount of current drawn ■was determined by draining the coulometer and titrating the iodine with standard sodium thiosulfate. This method for determining the amount of electrical equivalents was very satisfactory. current flow was allowed to fluctuate In these runs the rate of somewhat but was usually main­ tained between 1.0 and 2.0 milliamperes. WATER MOVEMENT Short Runs: Effect of Various ftfoisture Contents Three different benton ite-sand mixtures were used to observe the effect of varying moisture contents on electroosmotic flow. these the clay was 8b per cent saturated with calcium. In one of The other two sairples were sodium-bentonites, one 76 per cent saturated and the other 98 par cent saturated. Hereafter, the per cent saturation of the sairples will be referred to with respect to their symmetry concentration, S. A symmetry is defined as a concentration of ions equal to the exchange ca­ pacity. For exairple, a sanple 8U per cent saturated with calcium would be designated 0.8b S calcium. The effect of moisture content on the initial electroosmotic flow rates (milliliters of flow/coulomb) for the three samples are depicted graphically in Figures 6, 7, and 8. tension for each run is given. The per cent moisture and moisture The flow data from which these curves were constructed appear in Tables VI, VII, and VIII of the appendix. per cent moisture these tables. The and the conductivity of the samples is also given in The moisture tension of the samples was obtained from a graph constructed from the data given in Table IX of the appendix. Examination of the flow rate curves in these figures shows an in­ crease in the flow rate for each increase in moisture content or decrease in moisture tension. An increase in per cent moisture would be accom­ panied by an increase in the zeta potential (18) and a decrease in the viscosity which would act to increase the rate of flow. 2? Figures 6, 7, and 8 also show that the flow rate tends to change as the run proceeds. This effect is most-pronounced in the 0.76 S sodi­ um and the 0 .8U S calcium samples. In these samples the flow rate in- Va&erj ur s'ue riA mc 3.6;£.::cm m TrfK-£!ilOn; _ 2 3 Coulombs Figure 6. >Jater flow rate in a 98 per cent Na-saturated sandbentonite mixture containing 5 per cent bentonite at various moisture levels. The per cent moisture and the moisture ten­ sion are given for each run. creased x-nth the passage of current when free water was present or when the moisture tension in the sample was low. The rate decreased from its initial value at moisture tensions greater than 20 centimeters of water for the 0.76 S sodium sanple and at values greater than 120 centimeters for the O.8J4 S calcium sanple. In the 0.98 S sodium sample the flow rate increased over the ini­ tial value with the passage of current for all moisture levels studied. 21 Flow Rate (ml/coulomb) y 9 $ & % mter:: ^eer-Sfal,er Prt raei '&rmK)\% Wat:ir:: 20 icmJ Tens! 62.0 % I'ater Yly oil .10 0 1 2 3 h Coulombs 5 Figure 7. Uhter flow rate in a 76 per cent Na-saturated sandbentonite mixture containing 5 per cent bentonite at various moisture levels. The per cent moisture and the moisture tension are given for each run. Flow Rate (ml/coulomb) .22 0 :cm .10 cm cm. Coulombs Figure 8. l^barflow rate in a 81* per cent Ca-saturated sandbentonite mixture containing 5 par cent bentonite at various moisture levels. The per cent moisture and the moisture tension are given for each run. However, in each of the different bentonite-sand mixtures used an equi librium flow rate was reached after the passage of about 3*5 coulombs. These changes in the flow rates are probably due to small changes in the zeta potential and specific conductivity which occur with the passage of the first few coulombs. The attainment of equlibrium after the passage of only 3*5 coulombs shows that a steady state is quickly reached. The relationship between the flow rates and the per cent moisture is shown in Figure 9. For the 0.8u calcium sample the increase in flow rate with increasing moisture content is linear over the entire range tested which was from 18 to 46 per cent moisture. The flow rates in the benton­ ite—sand mixtures containing exchangeable calcium was erratic at moisture 3k 30 o.26 -£.18 10 n .06 10 20 60 Per Cent Moisture 70 80 100 Figure 9. The relationship between the per cent moisture and the water flow rate at 3*5 coulombs in benton ite-sand mixtures containing various exchangeable bases. values above U6 per cent. This was probably due to the fact that the bentonite-sand mixture was not conpletely stable at moisture contents higher than U6 per cent moisture. In the 0*98 S sodium and the 0.76 S sodium samples the relationship departed from linearity when free water was present. In the 0.98 S sod­ ium sanple the water flow increased at a faster rate when the amount of free water in the system was increased than when the water present in the system was under tension and the moisture content was increased. At the highest moisture content, 90.9 per cent, the sample contained enough free water to flow slightly. In the 0.76 S sodium sanple the water flow rate appeared to be ap­ proaching a maximum at the highest moisture content which was 90.6 per cent moisture. The mixture would flow slightly at this high moisture content as was true in the case of the O .98 S sodium sample. Schaad (32) and Winterkorn (38) have noted that the flow of water increases with in­ creasing per cent moisture, although they found that a maximum is reach­ ed after which the rate of flow decreases. The only sanple used in this study in which a maximum was approached was the 0.76 S sodium sanple. The sodium bentonites were noticeably more viscous than the calcium bentonites. As a result a similar water flow rate could be obtained for both the calcium and sodium samples only when the mixtures containing sodium were at a much higher moisture content than for the mixtures containg calcium. considerably. The per cent saturation also effected the rate of flow Much higher moisture contents were required in the 0.76 S sodium sanples than in the 0.98 S sodium samples in order to produce the same rate of flow. There is some evidence (3) that the viscosity of unsaturated sodium-clays is greater than that of saturated sodium-clays• A greater viscosity in the 0.76 S-sodium samples than in the 0.98 S.sod­ ium samples would account for the lower flow in the O.76 S sodium samples when the moisture content in the two samples was the same. There is no direct experimental evidence for this conclusion, however since the vis­ cosity of the samples was not determined. 22 <—t ■ lU s. .10 Ca .06 0 200 300 100 Tension (centimeters of water) Uoo Figure 10. The electroosmotic flow rates in bentonitesand mixtures containing exchangeable sodium and calcium as a function of the tension holding the water in the mixture and the per cent saturation. The effect of degree of saturation is also shown in Figure 10 where the rate of electroosmotic flow is plotted against the tension with which the water is held in the mixture. The data in Figure 10 show that the rate of flow in the 0.98 S sodium samples decreased slowly with in­ creasing tension, whereas the rate in the 0.76 S sodium samples decreas­ ed very rapidly with increasing tension. It is of note that in each sample with a negative tension holding the water in the soil that the rate of flow decreased linearly when the tension was increased. The flow rate at 25 centimeters of tension was about 0.17 milliliter/coulomb for all three samples. At greater 25 tensions than 25 centimeters the flow rate at a given tension varied greatly depending on the kind of ion and the degree of saturation. Long Runs: Effect of Exchangeable Ions Water Flow in Sodium, Potassiurn, or Calcium Systems The water flow data and ion removal data from the long runs reported in Tables X, XI, XII, XIII, XIV, and XV in the appendix were taken from the same samples. The viscosity of the sanples is dependent on the moisture content, the kind and proportions of ex­ changeable ions present. Since both the kind and amount of exchange­ able ion was varied, the various samples were brought to approximately the same consistency judged by their handling qualities before each electroosmotic run. Because of this, water flow data from the various treatments can be compared to a better advantage than if a single moisture level had been used for all of the sanples. The flow rates for the samples containing monovalent ions are shown in Figure 11 and those from samples containing calcium in Figure 12. By extrapolating the flow rates of the 0.86 S calcium sanple and the 0.78 S sodium sanple to zero coulombs, their flow rates can be conpared to similar samples for which the initial flow rate was actually determined. The extrapolated value at zero coulombs for the 0.86 S calcium sanple is 0.15 milliliter/coulomb. contained 38 per cent moisture. This sanple The corresponding value for a 0.8U S calcium sanple containing 38 per cent moisture can be obtained from Figure 9. This value is 0.16 milliliter/coulomb. Extrapolation for the 0.78 S sodium sanple shows the initial flow rate to be 0.12 milliliter/coulomb. At a corresponding moisture 26 level of £2 per cent, a 0.7U S sodium sanple had an initial flow rate of 0.09 mil iliter/coulomb. In view of the slight differences in the per cent saturation of the samples compared, it was felt that very good agreement was obtained by the two methods of determining the initial flow. It can also be concluded that the flow rates as obtained by the use of the cell with platinum electrodes compare very well with those from the cell with silver;silver chloride electrodes which was designed especially to measure the initial flow rate accurate­ ly* The flow rates for the systems containing sodium or potassium are given in Figure 11. Except in the 1.11 S sodium sanple where the data is quite erratic, the flow rate decreases continuously as electro­ osmosis proceeds and the cation is depleted. When about 0.10 symmetry of the cation is left in the sample the rate levels off and becomes con­ stant. -p 11 100 300 t 1. ^00 Coulombs 700 900 Figure 11. Electroosmotic flow rate through bentonite-sand mixtures containing exchangeable sodium or potassium. The symmetry concentration of cations remaining in the mixture is indicated at various points along the rate curve. Figure 12 gives the flow rates for the calcium systems. The flow rate patterns from the calcium systems are considerably different from those of systems containing monovalent ions and are also different from the systems containing a mixture of ions shown in Figure 13. The flow rates of all the calcium systems increases as electroosmosis proceeds ■6------ o 1.00 S Ca 1*------ • 0.86 S Ca .3 : i::► -0..0.,. 86 ,s -wCa ' ..j .. -m - .1 700 100 900 Coulombs Figure 12. Electroosmotic flow rate through bentonite-sand mixtures containing exchangeable calcium. The symmetry con­ centration of cations remaining in the mixture is indicated at various points along the rate curve. "*One thousandth normal hydrochloric acid in electrode chambers. until the amount of calcium left in the system is close to 0.£ symmetry. The rates then decrease until the cation content in the cores is between 0.3 and 0.U symmetry and then becomes more or less constant. It should be remembered that this calcium left in the core is not distributed homogeneously throughout the mixture. The bentonite-sand mixture nearest the anode will be almost conpletely hydrogen saturated while that near the cathode will contain about the same amount of calcium as it did originally. Rollins (30) has suggested that under such conditions, i.e. hydrogen clay at the anode and calcium clay at the cathode., that o° <£U non-homogeneous flow would occur. He showed that the flow rate in a hydrogen saturated kaolinite was less than in a calcium saturated kaoUnite, He reasoned therefore, that the inflow of water through the hydrogen clay would limit the flow of water through the entire system, (The flow rate from the calcium saturated side of the core could exceed that for the hydrogen saturated side for a limited time, but this would have the effect of drying the core.) A few calculations readily show that the flow observed in the calcium sanples cannot be explained satisfactorily on this basis. Consider the 0.86 S calcium saiple in Figure 12, after the passage of 900 coulombs, the sarple is 30 per cent calcium saturated and the flow rate has dropped to 0.11 milliliter/coulomb. Further clectroosmosis showed that the flow rate of the hydrogen saturated mixture in this saiple would approach 0.08 milliliter/coulomb. This then should be the rate governing the flow of water core if we assume that the rate of flow through the hydrogen clay limits the floxj into the entire saiple. At the beginning of the run the core contained 38 per cent moisture which would equal 29.1 milliliters of water. If the flow through the hydrogen clay at the anode controlled the amount of water entering the cell, then the amount of water entering the cell during the passage of 285 coulombs of electricity would be (0.08) (285) or 22.8 milliliters. During this period then the maximum water available for outflow would be equal to the inflow plus that already in the cell or a total of 51.9 milliliters. The actual amount of water that could flow out of the cell would be much less than 51.9 milliliters because if this amount of water flowed out the core would be left coipletely dry provided 22.8 milliliters was all that flowed in. Actual measurements 29 showed that 62.2 milliliters of water flowed out of the cell during this interval so it is evident that the flow rate through the hydrogen clay could not be the major factor controlling the flow of water through the entire sample over this current range. The evidence for systems other than the calcium systems is not clear cut. Similar calculations show that it would be theoretically possible for the flow through the hydrogen clay to be limiting the amount of flow into the cells in these systems although considerable reduction in moisture content of the core would be required in most cases. Since the moisture content of the cores at the end of the electroosmotic runs were not determined, it was not possible to come to any definite conclusions for any systems except those containing calcium. Further examination of the flow rates for the calcium samples shows that the shape of the flow rate curves are very similar to the calcium removal rates for these same systems. 20). (Compare Figures 12 and Both the rate of calcium removal and the rate of flow increase until the amount of calcium remaining in the cell is reduced to about 0.5 symmetry. Then in each case a decrease is observed which continues until the calcium content in the cores is reduced to between 0.3 and O.Ij. symmetry where both the calcium removal rate and the flow rate become more or less constant. Evidently the water removal rate is closely connected with the amount of current carried by the calcium in these systems. Water Flow in Mixed-Ionic Systems The flow rates for the mixed ionic systems were constructed from data appearing in Tables XV*I, XVII, and XVIII, According to Figure 13 the flow rates in the systems containing a mixture of ions decrease much more rapidly than in the mono-ionic systems* This decrease corresponds to the rapid decrease of sodium or sodium and potassium 0 100 700 330 900 Coulombs Figure 13* Electroosmotic flow rates through bentonite-sand mixtures containing complementary basic ions* The symmetry concentration of cations remaining in the mixture is indicated at various points along the rate curve* *Qne hundredth normal hydrochloric acid in electrode chambers• in these systems which is induced by the presence of complementary ions* In those systems containing sodium and calcium; or sodium, calcium, and magnesium a constant rate of flow is observed when the cations in the cell have been reduced to a symmetry concentration of about 0.5* At this point 98 per cent and 87 per cent of the sodium has been removed from the respective sanples (see Figures 16 and 17). The rapid reduction in flow irate prior to this is most likely due to 31 the change in zeta potential which accompanies the removal of the sodium. In the sanple containing a mixture of sodium and potassium the reduction in flow rate is continuous over the entire range and doesn*t become constant until the clay is almost coupletely hydrogen saturated. In this system there is no sharp change in the zeta potential due to the removal of one of the ions as in the mixed systems containing both sodium and calcium* Therefore, there is no sharp break in the flow rate although the change is more rapid in the early stages of electro­ osmosis where the rate of ion removal is the most rapid* CATION MOVEMENT This part of the experiment was carried out to study the removal of basic ions from Wyoming Bentonite tinder conditions of electroosraosis. The electroosmotic media used was a Wyoming Bentonite-sand mixture containing 5 par cent bentonite prepared as outlined in Section III. Ionic systems containing calcium, sodium or potassium were studied first. These systems contained but a single basic ion on the exchange coiplex except vrhen uns&turated sanples were used in which case hydrogen and aluminum were also present. In later experiments samples containing mixtures of exchangeable sodium and calcium; sodium and potassium; and sodium, calcium and magnesium were evaluated. From the results of these studies it appeared that if the path in the bentonite-sand plug were long enough, the passage of current would tend to separate the calcium and sodium ions due to differences in their mobilities. To determine the degree of separation which would take place hydrogen saturated bentonite-sand plugs were utilized. One set of plugs was 2.5 centimeters long and another set 5 centimeters long. For the 2.5 centimeter plugs 1.5 milliequivalents of calcium and 1.5 mi H i equivalents of sodium were placed in the anode chamber and the rate of movement across the plug followed by analyzing the con­ tents of the cathode chamber at various intervals. This process was repeated with the 5 centimeter plugs except that 2.5 mill {equivalents of calcium and 2.5 mill iequivalents of sodium were placed in the anode chamber. Cation Movement in Sodium, Calcium, or Potassium Systems In Figure ll; the per cent of the cation removed is plotted against the total number of coulombs flowing through the cell. The actual milliequivalents of the various ions removed and the current flowing 100 m nr 20 10 200 600 Coulombs 1000 Figure lH. The per cent of basic cations removed from a _ bentonite-sand mixture containing calcium, potassium or sodium during electroosmosis* The ion species and symmetry concentre tion for each system is shown. ^ ^ "?0ne thousandth normal hydro chloric acid in electrode 'chambers. 3k during their removal are shown in Tables X, XI, XII, XIII, XIV, and XV in the appendix. The electroosmotic flow from these sanples is discussed in an earlier section of this paper. Data in Figure lk show that the monovalent ions of potassium and sodium are removed much more quickly and efficiently from the exchange conplex than is calcium. For exanple after the passage of one hundred coulombs of electricity, 37 to I4.6 per cent of the sodium and potassium had been removed; whereas only 1? to 16 per cent of the calcium had been removed at this point. The higher mobilities of sodium and potassium would account for their greater removal along with the fact that bentonite has a greater bonding energy for calcium than for sodium and potassium* In the unsaturated potassium and sodium sanples, potassium was removed more quickly than sodium; but this was probably due to the fact that the potassium mixture contained 10 per cent more water than the sodium sanple. The mean free path would be shorter in the potassium sanple due to the higher moisture and hence the potassium would move more rapidly even though the mobility of sodium is higher than potassium under similar conditions. When sodium and potassium were present in the same sanple the sodium was removed at a rate greater than that for potassium (see Figure 15>) • The effect of the hydrochloric acid in the electrode chambers on the removal of calcium can be seen by a conparison of the 1.0 S, 0.86 S, and 0.86* S calcium sanples in Figure ll*. The electrolyte had little effect on the initial removal rate, however in the later stages of electro osmosis the amount of calcium removed was reduced noticeably. This can be explained as follox^s; 35 In the early stages of electroosmosis the electrolyte from the electrode chambers has not penetrated the cell appreciably or changed the per cent saturation to any large degree. The anode chamber contained about 0.05 milliequivalents of hydrochloric acid which was renewed at each sampling period. The amount of hydrogen ion in the anode chamber during any one saiplinc period was sufficient to change the per cent saturation only 1.5 per cent. Hence in the early stage of electro osmosis the rate of calcium removal was not affected greatly by the free electrolyte in the electrode chambers. The proportion of current conducted by the calcium in the mixture was about the same as in the cells containing just water. In the later stages of electroosmosis as Ca was removed and after the hydrogen had had time to penetrate the cell, a greater proportion of the current was carried by the electrolyte. This acted to reduce the amount of calcium removed per unit of current flowing. Eff ect of Comp 1ementary Ions To study the effect of complementary ions on ionic removal three systems were utilized in addition to the systems containing only one basic ion on the exchange system. The removal pattern for the sodium-potassium system is shown in Figure 15; that for the sodium-calcium system in Figure 16; and that for the sodium-calciummagnesium system in Figure 17. The actual mill iequivalents of the ions removed and the total current drawn are given in Tables XVI, XVII and XVIII in the appendix. In each case the addition of a cor.plementary basic ion increased the percentage removal of the original ion as well as that of the Figure 15>. The per cent removal of sodium and potassium from a 106 per cent saturated bentonite-sand mixture by electroosmosis* The symmetrv concentrations were 0.65 for sodium and Q.tl for potassium* Figure 16. The per cent removal of sodium and calcium from a 116 per cent saturated bentonite-sand mixture by electroosmosis. The symmetry concentrations were 0.59 for sodium and 0.57 for calcium. -o Per Cent of Cation Removed 100 ■m 80 6o 20 0 to 0 0 Coulombs o 200 Coulombs too Figure 17. The per cent removal of sodium, magnes­ ium, and calcium from a lit per cent saturated bent­ onite-sand mixture by electroosmosis. The symmetry concentrations were O.tt for sodium, 0.30 for mag­ nesium, and O.tO for calcium. The electrode chambers contained 0.01 N hydrochloric acid. 100 80 60 o o t0 20 0 200 6oo 8 00 37 complementary ion added. This is shown more clearly in Table II where the percentage of ions removed after the passage of 100 coulombs of electricity is given. Table II The amount of calcium, sodium, potassium and magnesium removed after the passage of 100 coulombs during electroosmosis. Clay Treatment 07 /° Ha K Cations Removed at 100 Coulombs Milliequivalent of % % 1 1a, K, Mg, and Ca Ca Mg Ua-l.il 5 U6 1.16 Ha-0.78 S 37 0.86 U6 K-0.79 S 0.79 Ca-1.0 S li.u 0.1)0 Ca-0.86 s 12.2 0.36 Ca-0.86 S* 12.6 0.36 Ha-0.65 S K-0.U1 S 70 Na-0.59 S Ca-0.57 S 92.5 1.78 5 5 .5 16.2 1.67 Ha-0.U5 S** 1.51 2 5 .5 3U.5 Ca-O.iiC S 77.5 Mq-0.30 S *0ne thousandth normal hydrochloric acid in electrode chambers. ‘'^One hundredth normal hydrochloric acid in electrode chambers. Care should be taken in conparing the values for the unsaturated sairples with those that contain free ions, but in each case the percent­ age removal of an ion is increased when a cortplementary basic ion is present. For exauple in the calcium saturated system 12.5 percent of the calcium had been removed at 100 coulombs, whereas in the sodiumcalcium system 16.2 per cent of the calcium had been removed and in the 38 sodium-calcium-magnesium system 25.5 per cent of the calcium had been removed* The increase in percentage sodium removed due to the pre­ sence of a complementary ion is even more marked than it is for calcium. Forty six per cent of the sodium had been removed at 100 coulombs from the 1.11 S sodium system, 77.5 per cent from the sodium-calcium-magnesium system, and 92.5 per cent from the sodiumcalcium system. It appears that the greater the bonding energy of the bentonite for the conpl ementary ion, the greater will be the per cent replacement of the other ion in the system per unit of electricity passing through the cell. Data in Table II also show the mi 11 iequivalents of sodium, potassium, calcium, and magnesium removed at 100 coulombs. The total amount of these cations removed by 100 coulombs of electricity from systems containing complementary basic ions is greater than the amounts removed from the sodium, potassium or calcium systems. In the sodium system which was 111 per cent saturated, 1.16 mi 11 iequivalents of sodium was removed by 100 coulombs of electricity; ‘ while in the sodium-potassium system which was 106 per cent saturated, 1.78 mill iequivalents of sodium and potassium was removed which is an increase of over 50 per cent. Of this later amount 1.17 milli- equivalentswas sodium and 0*6l mi 11 iequivalents was potassium. The increased cation removal in the sodium-potassium system over that of the sodium system was not due to moisture differences since the sodium system contained 81 per cent moisture and the sodium-potassium system contained only 68 per cent. Values given in the table shows that the amount of cations removed per unit of electricity passing through the cell was higher in all systems containing complementary basic ions than in any of the sodium, potassium or calcium systems. 39 Sodium and Potassium Removal The removal rate for a given ion species depends on the amount of current conducted by that species in the media. From an inspection of the removal curves it appeared that the amount of monovalent ions appearing in the cathode chamber per electrical equivalent was a function of the amount of these ions remaining in the bentonite-sand mixture at a given time: (VI) d K = kCMo- ^ 13 ctq Where M is the per cent of monovalent ion drawn into the cathode chamber* is the per cent of M in cell at q = 0* q is the coulombs of electricity necessary to remove M into the cathode chamber and k is a constant. and M are defined in terms of per cent because the removal patterns for all of the systems are given in per cent* but these values could be defined in any suitable concentration units just as well. When defined in per cent* (IIq-TO may be found by sub­ tracting the per cent cation removed* M* from 100. Assuming m = 1 and rearranging for integration gives: (VII) dM « k(dq) and upon integration: (VIII) ln(]^-J0 - -kq ♦ C where C is the constant of integration. If the rate of ion removal is dependent upon the percentage of the ion remaining in the cell* a plot of the log of (Mo-i4) versus q should yield a straight line with a negative slope. The data from each of the runs containing a monovalent ion (see Figures lU* 15* 16* and 17) are plotted in this manner in Figures 18 and 19* The data deviates but little from the theoretical for sodium and potassium* Uo 10090- — ol.l S Na .-•65 S l\la iUl S K - - ~ .-q* LI; S Wa-J4.S Ch .30 S Mg ‘— .59S N a ^ S C a 100 9065 S Na o 70Ol 60 - £*Uo- -p 30- (X 20- 20- 10 0 100 200 300 q (Coulombs) Figure 18. iiOO 0 100 200 300 q (Coulombs) 1*00 Figure 19. Figures 18 and 19* The log of the per cent sodium and potassi um respectively remaining in the cell as a function of the amount of current in coulombs. The legend shows the symmetry concentration and the basic ion species originally present in each system. -ff One hundredth normal hydrochloric acid in electrode chambers• hence it was concluded that equation (VI) adequately describes the removal of the monovalent ions under the electroosmotic conditions used. (VI). The removal of calcium and magnesium did not conform to equation hi It is evident that cation exchange reactions do not interfere greatly with the removal of sodium or potassium, or the removal rate would not be proportional to the concentration remaining in the cell. However* there is some evidence that cation exchange reactions do affect the removal rate to a degree. The addition of a complementary ion with a high bonding force such as calcium or magnesium speeds the removal of the monovalent sodium and potassium ions by reducing the number of exchange reactions that they can undergo. The greater the bonding energy of the complementary ion the fewer exchange reactions the sodium or potassium will be able to enter into and the greater will be their rate of removal. This is shown in Figures 18 and 19* The slope of the lines is an indication of the removal rate. steeper the slope, k, the more rapid the removal. The In the presence of calcium which has a high bonding energy the removal rate of sodium is greatest. The rate gets progressively smaller as complementary ions with lower bonding energies are used in conjunction with the sodium. However, as shown in Figure 19 this same tendency prevails when the removal rate of potassium is considered in conjunction with sodium. Bray (8) determined the "relative ease of release'*, (f) values, of various exchangeable cations by determining the amount of cation released when the soil was shaken with very dilute hydrochloric acid. The value for calcium is taken as 1.0, and the "relative ease of release11 values for the other ions is calculated by dividing the percentage release for each by the percentage release of calcium. Such a value provides a way to numerically compare the "relative ease of release" of cations from the various systems. Data in h2 Table III show a coitparison of the values Bray obtained with those calculated from the amount of cations removed by elcctroosmosis from the sodium-calcium-magnesium system (Figure 17) and the sodium-calcium system (Figure 16), The percentage of ions removed at 20 coulombs was used as the basis for the calculations, A comparison of the (f) values obtained by Bray and those obtained from the sodium-calcium-magnesium system shows good agreement. However, the "ease of release'1 of sodium increased markedly when the amount of calcium in the system was increased. Bray found in the dilute soil- electrolyte systems he worked with that the "relative ease of relase" was a constant for a given ion and a given clay. In such systems there was little hindrance to the removal process since the ions once Table III A conparison of the "relative ease of release" of cations by replacement with electrolyte and by re­ placement due to electroosmosis. Cation (f) Values from Bray (f) Values from Na-Ca-Pflg System (f) Values from PJa-Ca System Calcium 1.0 1.0 1.0 P^Lgnes ium 1.6 1.7 --- Sodium 6.2 1.5 12.2 exchanged moved directly into solution. In electroosmotic systems where the cations have to travel through the plug before being re­ leased into the cathode chamber, there is much more opportunity to enter into cation exchange reactions. The cations with low "ease of release" or high bonding energy (calcium and magnesium) are absorbed preferentially which reduces greatly the number of exchange reactions U3 into which the cations with low bonding energy can enter* The affect of the complementary ion is therefore much greater under electroosraotic conditions than thoseimposed byBray, increase in the "easeofrelease" and this accounts for the marked when the percentage calcium is in­ creased* Calcium.Removal An interesting phenomena in the removal of calcium from bentonite under electroosmotic conditions was the marked change in the rate of removal that occurred about midway in the removal process • Initially the rate of calcium removal was much lower than that for sodium or Table IV The basic cation removal rates of the sodium, calcium, or potassium systems at various per cent saturation values* Clay Treatment p r , rL bas e Saturation Ca Ca Ca ha ha K 1*° s ■- 0* 86 5 0.86 S* 1.11 S 0.78 S 0*79 S - . -- ...— .. _.. . Removal Rate (Hi 11 iequivalents/coulombX lo3) 30 -5f 3.5 70 h*h h.h 3.6 12.6 60 k .5 5.5 3.6 10.6 5o h.5 4.0 3.2 8.8 7.9 9.8 ho 1.6 3.0 0.6 8.1 6.6 9.2 30 0.6 0*9 7.1 6.0 7.6 20 6.5 5.5 7.2 10 5.9 3.8 3.0 One thousandth normal hydrochloric acid in electrode chambers hh potassium but the rate remained fairly constant whereas the rate for sodium and potassium fell continuously until they were completely re­ moved. The removal rates at various degrees of saturation for all systems containing but a single basic ion are presented in Table IV. The data show that the calcium rate began to fall between Uo and >0 per cent saturation and decreased until at about 30 per cent saturation the rate had fallen to about 0.6 mi H i equivalent s/coulomb X loO. The rate then remained constant over a wide range of current input although the per cent saturation did not change much because the removal rate was so lot*. Other studies indicate that the release of calcium from bentonite at saturation values below 30 per cent saturation is sufficiently low to adversely affect plant uptake of calcium. Chu and Turk (10) utilized bentonite-sand mixtures for fertility studies. At saturation values above 30 per cent the calcium content of oats and rye grown on these mixtures was quite constant. When the saturation was reduced from 30 to 1$ per cent, the calcium content of the oats and lye grown was re­ duced as much as £0 per cent* Puri (28) found a correlation between the Ca/TJa ratio in the electrodialysate and the yield of wheat in some stud­ ies conducted on soils, in India. In Figure 20 the removal rates of the three calcium systems are pre­ sented. The rate of removal is plotted against the total coulombs or electricity passed through the sairple during the electroosmotic process. If the removal rate for calcium was dependent solely on the concen­ tration in a manner similar to the monovalent ions, the removal rate would be expected to decrease continuously. However, the two distinct plateaus in the removal curves at which the removal rate is more or less constant is taken as evidence that calcium is bonded by bentonite x^rith two different energies of absorption. That calcium held when the per cent saturation is at approximately 30 per cent or below is held rela­ Ila &> r~t PC 2 ::S* 0 200 iiOO 6 00 Coulombs 800 1000 Figure 20. The rate of x'smoval of calcium from sand-bentonite mixtures containing 5 per cent bentonite under conditions of electroosmosis. "One thousandth normal hydrochloric acid in electrode chambers. tively strongly and hence the rate of release is very low. The calcium bonded at per cent saturation values between 30 and 100 per cent is held less strongly and is removed by the current at a higher rate. As this calcium held with the weaker bond is removed, the rate remains constant for a time and then as the supply of this calcium is depleted the rate falls until only the more tightly bound calcium is left. The rate then becomes more or less constant again. Calcium is bonded strongly by the bentonite and is the most diffi­ cult to displace from the clay of the ions studied. Once the calcium is displaced it will have a great tendency to enter into cation ex­ change reactions and be readsorbed by the clay. These factors are probably the main ones governing the irate of movement of calcium into k6 the cathode chamber. Over the two ranges where the release rate is con­ stant, cation exchange reactions may well be the limiting factor which holds the calcium release rate constant. In the electroosmotic cell the current must be carried by ions. The cations moving to the cathode would be those that are either free in solution or exchangeable plus hydrogen which is created in each system by ionization of water, or electrolysis. The addition of a com­ plementary basic ion apparently acts to reduce the proportion of current carried by the hydrogen and to increase the proportion carried by the other cations in the system. Differential Ion ffovement Data presented thus far indicate that the ions are drawn through the bentonite at various rates in the relative order of sodium>potas­ sium > magnesium > calcium. Puri (27) has also noted that the cations move differentially-due to electrodialysis although he found that calcium moved more readily than magnesium. The effect of electroosmosis on a homogeneous mixture of cations in a clay system then would be to make the system heterogeneous due to the differences in the rates of move­ ment. Under electroosmosis the ions of a given species would tend to be depleted in bands between the electrodes. The sodium depletion band would extend farthest from the anode and the calcium band would remain closest to the anode with potassium and magnesium intermediate. This effect of differential movement of ions has been demonstrated by Ifenecke (20) using inorganic cations and a cation resin as the ex­ change material. He was able to obtain sharp separations of the cations in a column 8 centimeters long with an electric potential applied across the resin column. hi The separation of calcium and sodium when drawn through a bentonitesand mixture during electroosmosis is shown under three different con­ ditions in Figures 21, 22, and 23. These graphs were constructed from the data which appear in Tables XVII, XIX, and XX of the appendix. In Figure 21 the removal rates are plotted to show the removal pattern of a bentonite-sand sairple containing both sodium and calcium adsorbed on the clay. When all of the sodium had been removed, 52 per cent of the calcium still remained on the clay. The data in Figure 22 represents the rate at which sodium and cal­ cium ions appeared in the cathode chamber after passing from the anode chamber through a 2.5 centimeter hydrogen saturated bentonite-sand core. Most of the sodium moved across the cell in a sharp wave and had been expelled into the cathode before appreciable calcium began to appear. Sixty one per cent of the sodium and 3 per cent of the calcium had cross­ ed through the bentonite-sand column after the passage of 200 coulombs of electricity. The removal of sodium and calcium through a 5 centimeter hydrogen saturated bentonite-sand core from the anode chamber is plotted in Figure 23. In the longer core the wave of sodium is spread out over a greater current interval but at 700 coulombs, 65 per cent of the sod­ ium had crossed the core along with 13 per cent of the calcium. At 260 coulombs where the highest rate of sodium removal occurred,22 per cent of the sodium had been removed but only 1 per cent of the calcium. If electroosmosis were to be used in the reclamation of salinealkali or alkali soils, the differential movement of ions xvould be an important consideration. Poor physical conditions due to excess ex­ changeable sodium is usually a major problem in these soils and practi- •2,11* r—| X ■§12 10 o 3d 0 200 Coulombs Figure 21. The removal rates of sodium and calcium from a 116 per cent saturated benton­ ite sand mixture by electrosmo sis. The symmetry concentra­ tions were 0.59 for sodium and 0.57 for calcium. 0« 0 200 Coulombs Figure 22. The rate at which sodium and calcium was drawn into the cathode chamber though a hydrogen saturated bentonite-sand core 2.5 cent­ imeters in length during electroosmosis. as o d5 •—1 i-H x ""Removed Ik i -W Potassium >Ivfe.gnesium >Calcium# The addition of a couplementaiy ion such as sodium, potass ium, magnesium, or calcium to a system containing one of these ions increased the percentage removal of the original ion as well as that of the complementary ion# Wot only was the percentage removal of each of the complementary ions increas­ ed but the total concentration removed per coulomb of electricity drawn was also greater# It was found that the greater the bonding energy of bentonite for the complementary ion, the greater was the per cent replace­ ment of the other ion in the system per unit of electricity drawn through the cell# The rate of removal of sodium and potassium in all of the systems containing these ions was directly proportional to the amount of these ions remaining in the bentonite-sand mixture* Transference numbers could be calculated roughly for the cations in those systems which did not con­ tain cations in excess of the cation exchange capacity. During the ini­ tial sampling period, the transference numbers of sodium and potassium in the bentonite-sand mixtures were very near unity* This indicates that during this time, essentially all of the current was carried by these ions. The transference numbers of calcium in the calcium systems during this initial sampling period were nearly the same as those for the calcium in a 0.01 normal calcium chloride solution. This indicated that in the early stages of electroosmosis, the calcium present in a bentonite-sand mixture carries the same proportion of current that it would if the cal­ cium had been in 0.01 normal calcium chloride. Evidence is presented which indicates that calcium is bonded by bent­ onite with two different energies of adsorption. That calcium held by the 56 bentonite when the per cent saturation is between zero and 30 per cent is held strongly, and is removed by the electroosmotic process very slow­ ly. That calcium held by the bentonite when the per cent saturation ex­ ceeds 30 per cent is held less strongly and is removed at a Taster rate. All of the experimental work involving the movement of ions indicat­ ed that there was differential movement of ions through the bentonitesand mixture. It was demonstrated using sodium and calcium that this differential movement tends to separate the ions or to deplete the ions in bands between the electrodes depending upon the length of the core and the time interval of the electroosmotic run. The quantity of electricity required to move a given amount of water or exchangeable bases is a variable. The data presented herein show that the amount of water moved per ionit of electricity is dependent on the moisture content of the system, the kind, and proportion in which the exchangeable ions are present. The amount of exchangeable bases re­ moved per unit of electricity is dependent upon the kind and proportion in which the exchangeable bases are present. For exairple during the passage of 100 coulombs of electricity through base unsaturated systems containing but a single base, the sodium and potassium removed was about twice that of the calcium. BIBLIOGRAPHY Ballou, E. V. Electroosmotic Flow in Homo ionic Kaolinite. Coll. Sci. 10:l;50-60, 1955* Barber, E. S. Discussion of Geotechnics. 27, pg. U17, 19U7. Jour. Proc. High. Res. Bd. Vol. Baver, L. D. The Effect of the Amount and Nature of Exchangeable Cations on the Structure of a Colloidal Clay. Univ. Mo. Res. Bui. 129, 1929. Bernatzik, W. Contribution to the Problem of Seepage Pressure in Electroosmosis. Proc. Sec. Int. Conf. Soil Mech. and Found. Eng. Vol. Ill, pp. 6f>-66, 19U8. Biefer, G. J. and hhson, S. G* Electroosmosis and Streaming in Natural and synthetic Fibers. Jour. Coll. Sci. 9:20-35, 195U. Bolt, G. H. Ion Adsorption by Clays. Soil Sci. 79:267-76, 1955* Bradfield, R. The Use of Electrodialysis in Physico-Chemical In­ vestigations of Soils. Proc. First Int. Cong. Soil Sci. Vol. II, pp. 261|-270, 1927. Bray, R. H. Ionic Competition in Base Exchange Reactions. Chem. Soc. 6ki95b-£>3> 19U2. Jour. Am. Casagrande, L. Electroosmosis. Proc. Sec. Int. Conf. Soil Mech. and Found. Eng. Vol. I, pp. 218-23, 19U3. Chu, T. S. and Turk, L. M. Growth and Nutrition of Plants as Effected by Degree of Base Saturation of Different Types of Clay Minerals. Mich. State Coll. Tech. Bui. 211;, 19h9* Daniels, F. Outlines of Physical Chemistry. Inc., New York, p. 1±13, 19L8. John Hi ley and Sons, _______3 m, thews, J. H., and Williams, J. W. Experimental Physical Chemistry. 3^rd Ed., McGraw Hill Book Co., New York, pp• 263- Tstfrml. _______. and Alberty, R. H. Physical Chemistry. Sons, Inc., New York, pg. 363, 1955. John Wiley and Dawson, R. F. and McDonald, R. W. Some Effects of Electric Current on the Consolidation Characteristics of a Soil. Proc. Sec. Int. Conf. Soil Mech. and Found. Eng. Vol. V, pp. 51-57, 19UQ. 52 (15) End ell, K. and Hoffman, V. Electrochemical Hardening of Clay. Proc. Int. Conf. Soil Mech. and Found. Eng. Vol. I, p.. 273, 191*6. (16) Geuze, E. C. W. A., deBruyn, C. M. A., and Joustra, K. Results of Laboratory Investigations of the Electrical Treatment of Soils. Proc. Sec. Int. Conf. Soil Mech. and Found. Eng. Vol. Ill, pp. 153-157, 191*8. (17) Harward, M. E. and Coleman, 1'J. T. Some Properties of H- and AlClays and Exchange Resins. Soil Sci. 75:131-88, 1951*. (.18) Jenny, H. and Reitmeier, R. F. Ionic Exchange in Relationship to the Stability of Colloidal Systems. Jour. Phy. Chem. 3*9:59360U, 1935- (19) Loddesol, A. Factors Effecting the Amount of Electrodalysable Ions Liberated from some Soils. Soil Sci. 33:187-211, 1932. (20) 1'fe.neke, G. Trennung von lonen in Ionenaust&ushchersaulen durch lonophorese. haturewissenschaften 39:62-63, 1952. (21) Mhrwick, A. H. D. and Dobson, A. F. Application of Electroosmosis to Soil Drainage. Engineering 163:121-23, 191*7. (22) others, A. C. and Coleman, IT. T. Ion Exchange Reactions Involving Aluminum and hydrogen. Agron. Abs., pg. 13, 1956. (Seen in abstract only) (23) I'feittson, S. Electrodialysis of the Colloidal Soil Materials and the E j^changeable Bases. Jour. Agr. Res. 33:553-67, 1926. (2l*) Mattson, S. The Laws of Colloidal Behavior: XI. Electrodialysis in Relation to Soil Processes. Soil Sci. 36:11*9-163, 1933(25) Mortland, M. M. and Mellor, J. L. Conductimetric Titration of Soils for Cation Exchange Capacity. Soil Sci. Soc. Am. Proc. 18:363364, 1951*. (26) Preece, E. F. Geotechnics and Geotechnical Research. Res. Bd. Vol. 27, pp. 38U-U16, 191*7. (27) Puri, A. N. and Hoon, R. C. Studies in the Electrodialysis of Soils: I. Electrodialysis by the Rotating Electrode. Soil Sci. 1*3: 305-309, 1937/ (28) Puri, A. N. Soils: Their Physics and Chemistry. ing Corp., Hew York, pp. .120-1*0, 1959- (29) Richards, L. A. (Ed). Diagnosis and Improvement of Saline and AdkaH Soils. Ag. Handbook ho. 60, U.S.D.A., pp. 85-80; l95L7^ (30) Rollins, R. L. The Development of hon Homogeneous Flow Condition During Electroosmosis in a Saturated Clay Mineral. Proc. High. Res. Bd. 35:686-92, 1956. Proc. High. Reinhold Publish­ 59 (31) Salgado, M. L. M. and Chapman, G. W. A Siirple Electrodialysis Cell for the Routine Determination of Exchangeable Bases. Soil Sci. 32s199-213, 1931. (32) Schaad, ¥. Electrical Treatment of Soils. Proc. Sec. Int. Conf. Soil Mech. and Found. Eng. Vol. VI, pp. 35-86, 191+8. (33) Taylor, H. S. and Glasstone, S. A Treatise on Physical Chemistry. Vol. 2, D. Van Nostrand Co., Inc., New York, pp. 628-38, 1952. (3U) Weiser, H. B. Colloid Chemistiy. 2nd Ed. John Wiley and Sons, Inc., New York, pp. 235-36, 298, 1950. (35) Hi Ison, B. D. Exchangeable Cations in Soils as Determined by Means of Normal Ammonium Chloride and Electrodialysis. Soil Sci. 26:1+07-1+19, 1928. (36) Wilson, B. D. Extraction of Adsorbed Cations from Soil by Electro­ dialysis. Soil Sci. 28:1+11-21, 1929. (37)" Winterkorn, H. F. Fundamental Similarities between Electro-osmotic and Thermo-osmotic Phenomena. Proc. High. Res. Bd. Vol. 27, pp. (38) (39) 1+1+3-51+, 19U7. . Surface Chemical Properties of Clay Minerals and Soils From Theoretical and E:xperimental Developments in Electroosmo­ sis. A.S.T.M. Spec. Pub. No. li+2, pp. 1+1+-52, 1952. Woodward, G. 0. and Miller, ¥. M. A study of Soil-Water Movement by Electroosmosis. Ag. Eng. 3U:29-33, 1933. APPENDIX 61 Table VI The flow of water and electric current m a bentonite-sand mixture containing 0.8U symmetry of calcium at various moisture and conductivity levels. The mixture contained 5 per cent bentonite. Run No. 95-96 97-98 101-102 105-106 103-101; Sauple No. 1 Total Water Flow (ml) 0.15 Coulombs of E lectricity 0.79 2 0.33 1.79 3 0.50 2.65 k 0.69 3.62 1 0.12 0.82 2 0.27 1.70 3 o.U5 2.83 h 0.58 3.56 1 0.13 0.97 2 0.25 1.97 3 0.3U 2.6U h 0.1;7 3.67 5 0.6U 5.00 1 0.08 0.85 2 0.19 2.00 3 0.27 3.01 h 0.33 3.72 1 0.09 1.31 2 ---- 2.81 3 0.15 3.81 Per Cent Moisture k6.h Conductance (micromhos/ cm 275 39.3 275 27.8 25U 21.9 180 17.7 86 62 Table VII The flow of water and. electric current in a bentonite-sand mixture con­ taining 0.76 symmetry of sodium at various moisture and conductivity levels. The mixture contained 5> per cent bentonite. Run No. 36-37 38-39 %2 £0-£l U2-U3 hk-bS Sauple Mo. 1 Total Water Flow (ml) 0.21 Coulombs of Electricity 1.07 2 0 .1*1 1.89 3 0.61* 2.83 k 0.81 3-51 1 0.17 0.91 2 0.36 1.80 3 0.55 2.70 k 0.76 3.73 1 0.20 0.89 2 0.36 1.79 3 0.52 2.68 h 0.69 3.67 1 0.16 0.97 2 0.31 1.92 3 0 .1*3 2.76 U 0.55 3.62 1 O.ll* 0.99 2 0.27 2.03 3 0 .1*7 3.71* 1 0.10 1.39 2 0.20 2.87 3 0.32 1*.73 Per Cent Moisture 90.6 Conductance (micromhos/cm --- 71*.5 --- 65.0 --- 62.0 --- 56.2 --- 1*8.6 — — .. _ 63 Table VIII The -flow of water and electric current in a bentonite-sand mixture con* taining 0*98 symmetry of sodium at various moisture and conductivity levels# The mixture contained 5 per cent bentonite# Total Water Flow (ml) Coulombs of Electricity Run Wo. Sample Wo. 69-70 1 0 .2 1 0.77 2 0 .5 2 1.70 3 0.83 2.67 k 1.13 3.53 5 1.1*0 h*3h 6 1.72 5.23 1 0 .2 5 0 .9 8 2 o .5o 1.90 3 0.79 2.92 k 0.97 3.55 5 1.12 U.07 1 0.25 1.11 2 0.14; 1.95 3 0 .6 1 2 .6 8 h 0.85 3.7U 1 0.17 1 .0 0 2 0 .3 5 2.07 3 0 . 14.8 2.79 h 0 . 6U 3.71 6 7 -6 8 71-72 73-7U Per Cent Moisture Conductance (micromhos/cm) 90.9 485 83.0 696 73.3 735 5 8.6 780 6k Table VIII (Continued) Total Water Flow (ml) Coulombs of Electricity Run Wo. SampIs Wo. 75-76 1 0 .0 8 0.72 2 0.21 1.8U 3 0 .3U 2.95 k 0.U2 3*70 1 0 .08 1.07 2 o .ij 2.10 3 0.23 3.12 k 0.31 14.08 77-73 Per Cent Moisture Conductance (mi cromhos/cm) 39.8 780 3 2 .2 600 Table IX The per cent moisture, symmetry values of exchangeable bases, and the moisture ten­ sion of three bentonite-sand sanples used in short electroosmotic runs. Each con­ tained 5 per cent bentonite* Tension (Centimeters of water) Clay Treatment Ha Na Ca 0.98 S 0.81+ S 0.76 S Per Cent Ffoisture 10 1+0.3 71.5 61.7 20 37.lt 6h.9 58.2 30 33'-2 50.6 5U.7 l+o 3h.9 57.2 53.5 50 ---- 53.7 ---- 60 33.2 ---- 50.6 3UU 13.3 ---- 3U.3 517 10.6 ---- 26.6 Table X The effect of electric current on the removal of exchangeable sodium and water flow in a bentonite-sand mixture containing 5 per cent bentonite and 1.11 symmetry of sodium. Run Wo: 115-116* Sarpls5 Wo. 1 2 3 k % 6 7 8 Coulombs of Electricity 76 L80 270 370 503 566 601 802 Total Water Flow (ail) 12.8 113.2 138.8 Amount of Wa Removed (m.e.) 0.96 27.5 kk.% 1.71 2.26 63.6 2.2*7 98.8 83.9 2.52 2.53 2.53 2.53 ■^he exchange capacity of the mixture in the cell was 2.21* milliequivalents, the initial moisture content 81.1 per cent, and the initial conductivity 880 raicrorahos/cm. Table XI The effect of electric current on the removal of exchangeable sodium and water flow in a bentonite-sand mixture containing 5 per cent bentonite and 0.783 symmetry of sodium. Run Wo 1 129-130* Sample Wo. Coulombs of Electricity 1 2 3 50 128 216 Total Water Flow (ml) 5.6 Amount of Wa Removed (m. e.) 0.53 12.0 1.07 18.5 1.62 2* 292 21*.U 2.02 380 30.7 2.22 6 7 8 *81 550 630 37.1 2.29 2*2.1 2.32 1*7.8 2.35 ■**The exchange capacity of the mixture in the cell was 3.05 milliequivalents, the initial moisture content 51.6 per cent, and the initial conductivity 93% micromhos/cm. 6? Tabic XII The effect of electric current on the removal of exchangeable potassium and ‘ water flow in a bentonite-sand mixture containing 5 per cent benton­ ite and 0.792 symmetry of potassium. Run Mo.i 117-118* Sanple Wo. 1 2 3 1* 5 6 7 8 Coulombs of Electricity 72 162 295 bOS k7l 591 7l*S 867 Total Water Flow (ml) 16.9 100.1 112.9 Amount of K Removed (m*e.) 0.71 32.6 $9. 8 1*7.7 1.1*2* 1.81 1.95 69.7 83.5 2.02 2.08 2.12 2.11* *The exchange capacity of the mixture in the cell was 3*63 milliequivalents, the initial moisture content 6l.£ per cent, and the initial conductivity 600 micromhos/cm. Table XIII The effect of electric current on the removal of exchangeable calcium and water flow in a bentonite-sand mixture containing 5 per cent bentonite and 1.002 symmetry of calcium. Run Wo.: 113-llU* Sarnple Wo. Coulombs of Electricity Total Water Flow (ml) Amount of Ca Removed (m.e.) 1 2 3 k 21*0 U21 S69 820 963 11U.1 128.0 3 2 .6 1 .0 6 6 5 .3 1.90 95.8 2.25 2.36 6 7 8 L065 1213 1323 2.1+9 11+2.6 2.60 158.1+ 2.7C 169.9 2.78 *The exchange capacity of the mixture in the cell was 3*^9 milliequivalents, the initial moisture content 37*1 per cent, and the initial conductivity 2l*2 micromhos/cm. Table XIV The effect of electric current on the removal of exchangeable calcium and water flow in a bentonite-sand mixture containing 5 per cent bentonite and 0.861; symmetry of calcium. Run No.s 111-112* Sampl e Wo. Coulombs of Electricity Total Water Flow (ml) Amount of Ca Removed (m.e.) 1 2 3 k 5 6 7 8 160 270 390 £98 785 911 L132 I31U 117.8 132.5 1U5.8 33.3 0.70 61.3 1.3U 86.li 1.67 1.89 2.11; 2.05 161;.6 iQl.h 2.25 2.33 ^The exchange capacity of the mixture in the cell was 3.U3 milliequivalents, the initial moisture content 37*6 per cent, and the initial conductivity 2S>1 micromhos/cm. Table XV The effect of electric current on the removal of exchangeable calcium and water f lo^ in a bentonite-sand mixture containing 5 per cent bentonite and 0.861; symmetry of calcium. 91-92** Run Wo.s Sample Wo. Coulombs of Electricity Total Water Flow (ml) 1 2 3 h 78 157 266 355 8.9 19.9 37.U 53-1 5 m 68.6 6 7 8 592 735 L007 88.6 108.8 IU4.3 Amount of Ca 1.70 1.21 1.36 1.35 1.U7 0.60 0.91; 0.27 Removed (m.e.) ^The liquid in the electrode chambers was 0.001 W hydrochloric acid. -x-Sfrhe exchange capacity of the mixture in the cell was 3*32 milliequivalents, the initial moisture content 1;0.1 per cent, and the initial conductivity 2l;5 micromhos/cm. Table XVI The effect of electric current on sodium removal, potassium removal, and water flow in a bentonite-sand mixture containing 5 per cent bentonite, 0.65 symmetry of sodium, and 0.1*1 symmetry of potassium. Run No.: 121-122* Sample No. Coulombs of Electricity 1 2 3 18 60 120 k 197 5 6 7 299 1*31 U8I4 Total Water Flow (nil) 7 .9 Na Removed (m.e.) 0 .3 0 0.71* 1.23 1 . 1*1* 1.50 1.53 1.55 K Removed (m*e.) 0.1*4 0.3U 0.65 0.85 0.91 0 .9 6 0.98 1 8 . 1* 3 1 .8 iti^* i* 57.0 70.9 81.9 *The exchange capacity of the mixture in the cell was 2.U5 nilliequivalents, the initial moisture content 68.0 per cent, and the initial conductivity 938 micromhos/cm. Table XVII The effect of electric current on sodium removal, calcium removal, and water flow in a bentonite-sand mixture containing 5 per cent bentonite, 0.59 symmetry of sodium, and 0.57 symmetry of calcium. Run No.: 123-121*'* Sanple No. 1 2 3 U 5 6 7 8 Coulombs of Electricity 5 39 86 178 21*1 296 362 1*7U Total Water Flow (ml) 9 .0 18.3 28.5 Na Removed (m.e.) 0.1*1* 0.96 1.1*1 1.57 1.59 1.59 Ca Removed (m.e.) o.o 3 0 .09 0.20 0 .53 0.79 0.95 6 57.0 6 8 .1 89. I 10 9.1 ---- ----- 1.13 The exchange capacity or tne mixouit; m ^ ~ --- ^ equivalents, the initial moisture content 60 per cent, and the initial conductivity 25U micromhos/cm. 1.25 Table XVIII The effect of electric current on sodium removal, magnesium removal, calciunremoval, and water flow in a bentonite-sand mixture containing 5 per cent bentonite, 0.1*1; symmetry of sodium, 0,30 symmetry of magnesium and 0.1*0 symmetry of calciurf. Run No.: 119-120** Sarapl