Clays and Clay Minerals, Vol. 43, No. 1, 85-91, 1995. SEPARATION OF SUB-MICRON PARTICLES FROM SOILS AND SEDIMENTS WITHOUT MECHANICAL DISTURBANCE G. J. CHURCHMANAND D. A. WEISSMANN CSIRO Division of Soils, and Australian Cooperative Research Centre for Soil and Land Management Private Bag No. 2, Glen Osmond, South Australia 5064 Abstract--A method is described for the separation of the finest particles from soils and sediments without mechanicaldisturbance. Particles are separated through the inductionof osmotic stress. Generally, samples are treated with a concentrated sodium salt solution and then exposed to water by diffusion. Naturally sodic samples are simply exposed to water. Solid samples and the swollen and dispersed material they produce are confined by dialysis tubing. Examples show that the method gives a size gradient of particles in a vertical column of suspension. The compositions of particles can vary with size. The method can be used to show the effects on separated particles of ions other than Na § and also of other physico- chemical treatments of soils and sediments. It is inexpensive and requires little labor. Key Words--Clays, Dialysis, Dispersion, Mobility, Particles, Sediments, Separation, Size, Soils, Sub- micron, Swelling. INTRODUCTION The finest clay material almost invariably dominates both chemical and physical properties of soils even though the coarser sand and silt fractions are often predominant by weight or volume (e.g., Robert and Chenu 1992). Characterisation of the fine material of- ten requires its separation from the whole soil. Sepa- ration by centrifugation is common. However, me- chanical disturbance of the soil inevitably accompanies centrifugation and may change the nature and distri- bution of separated particles. A method described here which uses dialysis tubing enables the separation of sub-micron particles from soils and also from sedi- ments. The method avoids mechanical disturbance, is inexpensive and requires little labor. It also provides a visual demonstration of swelling, dispersion and the subsequent diffusion of fine particles in soils and sed- iments and enables the measurement of extent of oc- currence of these phenomena. PRINCIPLE OF THE METHOD In this method, particles are separated by inducing sufficient osmotic stress between them to cause dis- persion. Subsequent diffusion against gravity produces a size gradient of particles for sampling. The use of osmotic stress as a "chemical hammer" (Clapp and Emerson 1965) to assess the strength of bonds between soil aggregates was initiated by Emer- son (1954). By its use, degree of dispersion of a sample is measured when the easily exchangeable cations are replaced by Na § ions and electrolyte concentration is reduced. In the new method, osmotic stress is normally ap- plied to samples by treating them with a concentrated Copyright 1995, The Clay Minerals Society solution of a sodium salt to replace exchangeable cat- ions with Na + and then exposing the Na+-exchanged sample to water. Osmotic stress can also be induced in soils and sediments which are naturally sodic, and hence dispersive (Gupta and Abrol 1990), by their con- tact with water without additions of sodium. PROCEDURE The gadgetry used in conjunction with dialysis tub- ing has undergone evolution since the inception of the method. The complexity of the procedure to be fol- lowed depends on the aim of the application of the method. Procedure for demonstration purposes A concentrated solution of NaC1 is added slowly to an air-dried whole sample in a dry dialysis tube which is sealed at both ends by tying them tightly with rubber bands. The tube containing the treated sample is hung vertically in a cylinder overnight. The cylinder is then filled with deionised water which is replaced regularly by siphoning until enough electrolyte has been re- moved from the equilibrated solution that it shows no precipitate with mgNO 3 solution. As the electrolyte concentration is decreased the sample may swell, with dispersion generally beginning before final dilution. In the example shown in Figure 1, a sample ofa Mollisol (Oxic Argiudoll) or Prairie soil (Stace et al 1968) from Glen Innes, New South Wales, was treated with 1 M NaC1 and dispersion was allowed to occur for 34 days. Procedure for analytical purposes The outfit we now use is illustrated in Figure 2, and shown in operation with a saline sediment from Chow- illa, South Australia in Figure 3. 85