NYL!” WYN x IN MENU UNCLASSIFIED ORNL 573 OR ND-d23 - OCT 30 1964. * USE OF NEGATIVE SORPTION IN STUDTES OF ION FIXATION BY HYDROBIOTITEL by Donald G. Jacobs Busintetico Health Physics Division Oak Ridge National laboratory Oak Ridge, Tenne8860 ABSTRACT The exclusion of anions relative to tritiated water from the nega- tively charged interlamellar spacing of hydrobiotites having a limited degree of lattice expansion permits measurement of the interlamellar volume. Measurement of the changes in the interlamellar volume and in the exchange capacity when hydrobiotite lattices are collapsed by potassium treatment or expanded by treatment with sodium tetraphenyl boron provides a means for estimating the distribution of the surface charge densities of these interstratified minerals. In the hydrobiotite sample studied (zonolite from Traveler's Rest, South Carolina), the biotite fraction was estimated to have a surface charge density of 2.0 + 0.14 x 10° 'meg me and the vermiculite fraction 1.5 + 0.51 x 10°'meq cm2. The charge density, and its relative distri- bution, was reflected in the tendency of these mineral fractions to collapse upon potassium treatment and to entrap trace concentrations of cesium in the interlayer spacing. .. . . ..... .. . Research sponsored by the U. S. Atomic Energy Commission under con- tract; with the Union Carbide Corporation, ... , : : TEGAL NOTICE To repertorium w we r d wa. Hatter the O, www maulana mand WS, mo ' N ., . . INTRODUCTION Clay minerals generally have a net negative charge which cannot be satisfied locally, since the counterions (ions having the opposite sign of charge from that of the surface) cannot penetrate the denso lato tice. Thus the charge can be thought to effectively reside on the surface of the clay lattice, and an electric double layer 18 established with an accumulation of counterions and a deficit, or negative sorption, of co-ions (ions having the same sign of charge as the surface). The variation of the electric potential, v, with distance from the surface, X, 18 givèn by the standard Gowy-Chapman theory (Kruyt, 1952, p. 129): - (92% no kop_)2 sinh ( where n = the number of ions per cubic centimeter at a point far removed from the charged surface, k = the Boltzman constant, = the absolute temperature, the dielectric constant of the medium, = the valency of the ions, and e = the charge on an electron. - -- The local concentration of a given type of lon 18 given by the Boltzman expression: C = C10 exp () (2) When these equations are combined and integrated over the electric double layer, the amount of co-lon exclusion from the double layer can - be calculated. Schofield and Tallbuddin (1948) have shown that the integrated expression can be given approximately as: Tot ziens (3) V ZBC where 1 . the surface deficit of anions, & factor determined by the ratio of valency of the counterion and the co-ion (and 18 equal to two for symmetrical salts), Sao P = 1.06 x 1045 cm-meg-2 when water 18 used as the solvent at 25°c, and = the Faraday, = the dielectric constant of the medium, . the molar gas constant, absolute temperature, and = the charge density of the surface in meq-cmº?. F € R I The value 1 a/c obtained from equation (3) has units of length and 18 the effective distance from which co-fons are excluded from the charged sur- face. The internal exclusion volume can be calculated, if the surface area 18 know, or it can be determined experimentally from the intercept of the linear portion of the curve when exclusion volume is plotted versus qlzBC. When two successive electric double layers overlap, a correction must be applied. If the surfaces are sufficiently close, however, the exclusion of anions in the region of overlap can be depicted adequately by Donnan membrane theory (Schofield and Talibuddin, 1948). EXPERIMENTAL PROCEDURE Sodium-saturated hydrobiotite was used to fill ion exchange zonulite from Traveler's Rest, South Carolina columns. Hydrobiotite exists as a mixed-layer mineral, having a rather random mixing of collapsed biotite-type layers with no interlamellar water, and vermiculite-type layers having two layers of interlamellar water. The spacing between the charged faces of the sodiu-saturated vermiculite layers 18 5.55 A (Grim, 1953, p. 74); thus Donnan membrane theory can be used to describe the chloride exclusion from the inter- lamellar region. The columns were pre-equilibrated with untagged NaCl solution of the desired concentration. Tritiated water solutions of Naci of the same concentration and tagged with 1001 were passed through the column at a constant flow rate. Each run consisted of a saturation and a leach- ing step, and the same column was used for all runs. Samples were counted simultaneously for 'n and ci, using a Packard Instruments . . .. . . . Tri-Carb counter, .. .. .. A computer program, written by George Atta, Mathematics Division, . . -7 .'5. Oak Ridge National Laboratory, was used to correct the overlap of the two beta spectra and to fit the experimental data to a chromatographic breakthrough curve of the type (Glueckauf, 1955; Hashimoto, et al, 1964; Rifai, et al, 1956): X - V-V C/C. - 1/2 {-ers (vp vetem ). (*) - where c/cthe fraction breakthrough of the tracor, P = the Poc.let, number for the column V = the throughput volume, and Ñ - the equivalent column volwe. SP where N 18 the number of theoretical plates in the notation of Glueckauf (1955) and where ✓ 18 the linear pore velocity of the solution, 1 18 the column length, and D 18 the dispersion coefficient in the notation of Rifai et al (1956). CHLORIDE EXCLUSION : Exper:mental data showed that the apparent pore volume for tritium remains relatively constant with changes in NaCl concentration. For chloride there was a steady drop in the apparent pore volume with de- creasing Naci concentration. If the charged surfaces of the hydrobiotite had been sufficiently separated to permit full development of the electric double layers, a plot of the measured exclusion volume versus the right hand side of equation (3) would have yielded a straight line with an intercept of zero and a slope corresponding to the surface area. The linear portion of the curve (Figure 1) has a slope corresponding to 1.15 m2/8, but the intercept 18 0.127 m/8, which 18 a measure of the Interlamellar volume where the electric double layers overlap. The measured exchange capacity of the sodium treated bydrobiotite 18 0.753 meq/8, and the total accessible internal and external surface area 18 calculated from the experimental data to be 459 m/8; thus the mean su face charge density 18 1.64 x 10° meg/cm². Using this value, the second term of equation (3) 18 evaluated as 2.3 x 10°°cm, and the volume charge density of the internal pore volwe is 5.93 N. To ensure that the above explanation of internal and external effects was correct, another series of experiments was conducted in which the hydrobiotite was subjected to potassium treatment for collapse of the vermiculite layers (Barbhad, 1948) or to sodium tetra- phenylboron treatment for expansion of the biotite layers (De Mumbrum, 1959). In this series 0.01 M NaCl solutions were used in the determination of the chloride exclusion volumes. The results, compiled in Table 1, provide cvidence that the model does provide an accurate description of - - - - the system. As the biotite expands the exclusion volume increases and - - - as the vermiculite collapses the exclusion volume decreases. - ... - - . .. , i - * . . . . LE . After the hydrobiotite had been subjected to five treatments with • sodium tetraphenyl boron, the interlamellar exclusion volume was esti- mated to be 0.200 ml/g. This 18 the value that would be expected if all of the interlamellar spacing is filled with water, except for the volune occupied by the sodium ions. Thus, it seems that this treatment caused complete expansion of the biotite fraction. From this value it 18 further estimated that there 18 a total internal surface area of about 720 m/s. It was noted also that the breakthrough curves for chloride were steeper than those for tritium. This 18 reasonable since the tritium Putriticissitärin We st . . must penetrate the restricted interlamellar spacing from which the SAYS: chloride ion is excluded. It was assumed that differences in the experimental plate heights (obtained from the slopes of the '8 and 3001 breakthrough curves) would be due to a slow diffusion of tritium into the interlamellar spacing. Using Glueckauf's simplified model for the contribution of particle diffusion to the total plate height (Holfferich, 1962, p. 453), the partiole diffusion coefficient for tritiun in the interlamellar spacing was estimated to be 4.6 + 2.7 x 10° cm/sec. These values are of the same order of magnitude as those obtained by Keay and wila (1962) for self-diffusion of barium ions in vermiculite. They also agree quite well with later estimates made using 1sotopic exchange data for sodium in similar columns of hydro- biotite which yielded values of about 3 x 107cm/sec. A : CESIUM FIXATION - BY HYDROBIOTITE The sorption of low concentrations of cesium by sodium chloride treated hydrobiotite 18 primarily due to the collapsei biotite layers (Jacobs and Tamura, 1960; Jacobs, 1963). When the concentration of cerium is raised beyond the point where all of the edge fixation sites have been satisfied, part of the additional cesium is sorbed at the basal surface of the vermiculite layers. When sufficient cesium 18 sorbed at the basal surface of a vermiculite layer, collapse of that layer occurs. Collapse of the vermiculite layers cause entrapment of the cesium sorbed at the basal exchange sites in a process of interlayer fixation. Desorption Studies Desorption of cesium entrapped by interlayer fixation is quite dif- ficult (Table 2, column 1). The interlayer cesium is almost completely unexchanged by stable cesium salt, since the excess cesium tends to main- tain the lattice in its collapsed state. Only slight amounts of the interlayer cesium are exchanged by acid, even when sufficient acid 18 used to cause apparent degradation of the hydrobiotite. The cesium held at the edges of the collapsed biotite layers (Table 2, column 2) 18 more easlly exchanged hy acid. A greater quantity 1s exchanged also by the first leach with CsNO, but the small additional quantities leached with succes- sive leaches suggest that the Can', has been effective in collapsing the vermiculite lattices and entrapping the remaining cesium. When the hydro- biotite had been treated with sodium tetraphenyl boron to open the biotite layers, edge fixation of cesium was prevented and the sorbed cesium was readily exchanged by hydrochloric acid. Again desorption with C8NO, sub- gested that col.lapse of tła lattices prevented desorption beyond the first leach cycle. Column Studies ix Sodium chloride treated hydrobiotite (10 g) was placed in 1/2 in. diameter columns and 7.5 liters of 0.5 M Nano, containing 1.7 x 10'M CsNO, and various concentrations of potassium was run through the columns at a constant flow rate of 1.1 ml mincm?. Concentrations of potassium of less than 4 x 10" M in the influent solution had a slight depressing effect on the sorption of cesium (Fig. 2). At higher concentrations of 2 potassium, however, the potassium caused lattice collapse, with a decrease in exchange capacity and interlayer fixation of the trace quantities of ceslum. The peak cesium loading in this series occurred at about 0.04 M potassium; higher concentrations of potassium caused more lattice collapse but provided more competition for cesium sorption at the basal exchange sites. When the columns were filled with hydrobiotite treated with sodium :-- tetraphenyl boron, two distinct peaks occurred in the interlayer cesium - fixation curve as the concentration of potassium in the influent solu- tion is increased (Fig. 3), rather than the single fixation peak found for the sodium chloride treated hydrobiotite. Each peak is accompanied by a distinct inflection in the exchange capacity curve. The sharp peak occurring between 10") and 10-2 M KNO, reflects the recollapse of the opened biotite layers, while the broader peak occurring between 10°C and 10°- M KNO, 18 due to the collapse of the vermiculite layers. From the data presented in Table 1, it can be estimated that the average interlayer charge density 1s 2.00 + 0.14 x 10°' meq/cm for the biotite layers and 1.50 + 0.51 x 10°7 meq/cm2 for the vermiculite layers. It seems that the position and width of the cesium fixation curves reflect the charge deneity distribution of the collapsible lattices. As noted by Barshad (1948, 1949, 1950) and Weaver (1958), the higher the charge density the more susceptible 18 the lattice to collapse; hence a smaller K*/Na* ratio is required to induce collapse. The distinct peaks arising from the biotite and vermiculite type layers indicate that this hydrobiotite does not consist of an even gradation from low-charge density vermiculite to high-charge density biotite, but that there are two distinct mineral species in the sample. Differences in the behavior of the biotite and vermiculite layers of the Zonolite were observed also in time studies of potassium release in which it was noted that the collapsed vermiculite lattices had a much greater tendency for re-expansion than the biotite lattices (Jacobs, 1963). South African vermiculite, supplied by the Phosphate Development Corporation (Pty) Ltd., Phalaborwa, Transvaal, South Africa, was also studies for cesium fixation. In this case, a potassium concentration of of 0.2 M was required to obtain optimum interlayer fixation of cesium from 0.5 M NaNO, (Fig. 4). The results of negative sorption of chloride indicate that the surface charge density of this material is slightly less than that of the sodium chloride treated Zonolite. The measured exchange capacity of the African vermiculite was 1.08 meq/8. From nega- tive sorption data the external surface area was estimated to be 1.16 m