Effect of Freezing on the Swelling of Clay Minerals

It has been known for some time that when a gel of montmorillonite is freeze-dried the residual clay retains the volume and shape of the gel, and forms a skeleton of low density and some rigidity. Because there is no apparent change in the gel structure during freeze-drying, the dry residues have been studied in the past in an effort to determine the structure of the gels. However, X-ray diffraction measurements have shown that, as a result of freeze-drying, the distance between the elementary layers of Na-montmorillonite drops from high spacings (> 30/~) to zero. The present investigation demonstrates that the main reduction in interlayer spacing takes place on freezing, with a further small collapse on drying. Na-montmorillonite gels, swollen oriented aggregates of Na-montmorillonite and Na-nontronite, and swollen single crystals of Li-vermiculite and butylammonium-vermiculite, all behave in the above manner. INTRODUCTION Weiss, Fahn and Hofmann (1952) reported that when a thixotropic gel of Na-montmoriUonite is frozen and the water then sublimated under vacuum (freeze-dried), the solid particles of clay remain as a rigid skeleton of a very low density (0.05 g/ml), occupying the same volume as the gel. This result was interpreted as direct evidence of the existence of rigid structures in clay gels, thus providing an explanation of the phenomena associated with thixotropy. Small angle X-ray diffraction measurements by Norrish (1954) showed that when Na-montmorillonite is allowed to swell in water or in dilute solutions of sodium chloride the elementary silicate layers move apart to distances greater than 30 A. I f the interpretation of Weiss, Fahn and Hofmann (1952) were correct, i.e., if the water from the swollen montmorillonite could be removed by freezedrying without altering the arrangement of the silicate sheets, there would be practical advantages in making the X-ray diffraction measurements on the freeze-dried residues rather than on the gels. The intensities of diffraction should increase because of reduced absorption and also because of the extra contrast between the clay and its medium. Also, it should be possible to extend the diffraction measurements to higher angles where they are normally confused by diffraction from water. 10 K. NORRISH AND J. A. RAUSELL-COLOM Accordingly, low-density structures without loss of volume were obtained by freeze-drying Na-montmorillonite which had been swollen in water and in dilute solutions of sodium chloride. When the diffraction pattern from the freeze-dried residues, originating from either thixotropic gels or oriented aggregates, was examined, it was found that the inter-particle distances present in the gels (>3fl A) were not retained but dropped to zero (basal spacings of 10 A). This was in agreement with the previous work of Call (1953) --later confirmed by Van Olphen (1956)--who obtained surface areas, by the standard Brunauer, Emmett and Teller (B.E.T.) method, not differing significantly from those of the original powdered montmorillonite. All these observations indicate that although the gels do not change in volume on freeze-drying, there is collapse in local regions, with the necessary development of large pores. The aim of this investigation was to establish how the open structure of the gel collapses to that present in the freeze-dried residues. EXPERIMENTAL A few grams of bentonite from Wyoming, U.S.A., and of nontronite from Spokane, Washington, U.S.A., were saturated with Na + by several treatments with a 2 N solution of sodium chloride, followed by dialysis until the clay was free of C1-. The Na-saturated clays were then diluted with distilled water to a concentration of 0.1 per cent. and passed through a Sharpies centrifuge to select the particles of an equivalent spherical diameter of less than 0"lt,. Thin oriented aggregates were prepared by evaporation of the gel on a flat glass surface. A thin flake cleaved from a crystal of vermiculite from Kenya was cut into pieces of approximately 0-5 • 10 mm, some of which were treated with a 1 N solution of lithium chloride and others with a 4 per cent. solution of butylammonium chloride, until saturation with the corresponding cations was complete. The following specimens were mounted in thin walled plastic tubes (0.9 mm int. diam.): (a) a thixotropic gel (concentration 4.5 per cent.) of Na-montmorillonite; (b) an oriented aggregate of Na-montmorillonite swollen in 0.1 N sodium chloride solution; (c) an oriented aggregate of Na-nontronite in 0-1 N sodium chloride solution; (d) a flake of Li-vermiculite in 0-1 N lithium chloride solution; (e) a flake of butylammonium-vermiculite in water. After sealing the ends of the tubes to prevent evaporation these specimens were mounted in a Norelco diffractometer and, diffraction FREEZE-DRYING OF CLAY MINERALS 11 traces were obtained over the ranges of 20 from 0.5 ~ to 2 ~ (low angles) and from 3.5 ~ to 10 ~ (high angles). Then, while each specimen was still mounted in the diffractometer, a stream of cold carbon dioxide was directed at it. The counter was positioned at 1 ~ during the cooling process. After some 10 to 20 seconds the specimen froze suddenly, and simultaneously there was a sharp drop in the diffracted intensity. The diffraction pattern of the frozen specimen was then recorded for both angular ranges. Finally, for the two specimens of vermiculite, the flow of cold carbon dioxide was stopped and the specimens allowed to thaw. A sudden increase in the intensity of diffraction at 1 ~ was observed at the moment of thawing. The diffraction patterns corresponding to this final stage were also obtained. These results are reproduced in Figs. 1 and 2; CoKa radiation was used for all the traces. For the measurements at low angles slit apertures were small (1/30 ~ divergent, 0.075 m m receiving), in order to reduce the background near the direct beam. In the high angle region the slit apertures were bigger (1 ~ divergent, 0.15 m m receiving), and therefore the diffracted intensities are not to be compared with those of the low angle region. Quantitative measurements of the angular distribution of diffracting particles in each specimen were obtained from the decrease of intensity that occurred when the specimen was rotated with the counter kept stationary at a value of 20 where there was significant diffraction (RauseU-Colom and Norrish, 1962). These measurements are plotted in Fig. 3 as curves of orientation, giving, for each specimen, the number of diffracting particles at different angles from the plane of orientation. Arbitrarily, the total number of diffracting particles at zero degrees of rotation is made equal to unity for all specimens. DISCUSSION The diffraction traces obtained before freezing (see Figs. 1 and 2) are characteristic of swollen specimens of day minerals. The intensities follow the theoretical curve F2.~_~*, but there are deviations corresponding to the distribution of interlayer distances in the gel (Norrish, 1954). As these distances are greater thart 40/~, there are no deviations from F2..'~ -, in the range from 3.5 ~ to 10 ~ From 0.5 ~ to 2 ~ the diffraction effects are better defined for Li-vermiculite than for rnontmorillonite and nontronite. Because the interlayer spacings in the swollen butylammonium-vermiculite are very high (,~540 A, according to Walker, 1960) sharp diffraction effects cannot be observed on that trace, as they would appear at angles smaller than 0.5 ~ where scatter from the direct beam makes observation impracticable. *FL ~ represents the product of the square of the modulus of the structure factor of montmorillonite (its component in the direction of c*) by the function giving the angular correction. 12 K. NORRISH AND J. A. RAUSELL-COLOM