ICT ..,.•..!:.•. ■,,...- .: , ■'-""■ LIBRARY OF THE UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN (oZQ no. 13 - \(o CONF. ROOM 'W5INFFRIN3 *®#Wy f EB 27 1978 The person charging this material is re- sponsible for its return to the library from which it was withdrawn on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN P$ .". JNTERLIBRAR* ENGIILW yFfa lqah iy L161— O-1096 Digitized by the Internet Archive in 2013 http://archive.org/details/effectofsoilprop14sues Tits' c WGIRKW$€kUBRAR7 ~-t * • ( *»- UNIVERSITY OF ILLINOIS CIVIL ENGINEERING SVUIM$u.in«s i SANITARY ENGINEERING SERIES NO. 14 0&* OCT 2 0$ THE EFFECT OF SOIL PROPERTIES ON THE ADSORPTION OF ALKYL BENZENE SULFONATE By MICHAEL JAKOB SUESS Supported By DIVISION OF WATER SUPPLY AND POLLUTION CONTROL U. S. PUBLIC HEALTH SERVICE RESEARCH PROJECT WP-1 8 DEPARTMENT OF CIVIL ENGINEERING UNIVERSITY OF ILLINOIS URBANA, ILLINOIS JANUARY, 1963 EFFFXT OF SOIL PROPERTIES ON TIE ADSORPTION OF ALKYL BENZEI^ SULFONATE by MICHAEL Jo SUESS Supported by Water Supply and Pollution Control Division U. So Public Health Service Research Project WP-18 Department of Civil Engineering The University of Illinois Urbana, Illinois January, 1963 ii ABSTRACT This research was conducted in order to study the effect of soil properties, especially grain size and shape, surface area and mineralogical composition, on the adsorption of alkyl benzene sulfonate. For the purpose of this study siliceous, calcareous and silty clay soils were used. It was found that the most effective factor was the surface area, which was obviously related to the nrain size and shape. However, although the adsorption of ABS per unit weight of soil was increasing with decreasing grain size, the adsorption per unit surface area (adsorption intensity) was higher on the coarser soils. It was also found that the adsorption per unit weight on ground limestones was somewhat higher than the adsorption on sand- stones, but the adsorption intensity of both remained in the same range. It was shown that the retardation of ABS flow in an aquifer increased with decreasing grain size and adsorption intensity. Because of the generally poor retention of ABS on typical aquifer soils, the conclusion was reached that ABS movement in ground could only be delayed but not stopped, as the practice has shown. iii ACKNOWLEDGMENTS This report is submitted as a thesis in partial fulfillment of the requirements for the degree of Master of Science in Sanitary Engineering under the direction of Dr. Ben B. Ewing, Professor of Sanitary Engineering, University of IllinoiSe The research was supported by a research grant WP-18(C2) from the Water Supply and Pollution Control Division, Public Health Service and carried out at the Sanitary Engineering Laboratory of the University of Illinoiso The radioactive alkyl benzene sulfonate used in this study was supplied gratis by the California Research Corporation. The Ottawa Silica used in this research was obtained from The Ottawa Silica Company free of charge. I wish to extend my deep gratitude to Dr. John P. Kempton, Illinois State Geological Survey, for the supply of many of the soils used in this study and for his advice in the identification of the soils, and to Dr. Walter Parham, Illinois State Geological Survey, for preparing and analysing the X-ray diffraction patterns of the soils. My sincere thanks to Mr. Shankha K. Banerji and Mr c Sambhunath Ghosh, both co-workers in the ABS project, for their help. But above all, I wish to thank Dr<, Ben B. Ewing for his guidance and assistance in perfoming this research, and to express my deep gratitude for spending many hours advising me in my work. Michael J. Suess Urbana, Illinois IV CONTENTS Page ACKNOWLEDGMENTS iii LIST OF TABLES v LIST OF FIGURES vi INTRODUCTION 1 CHAPTER 1: THEORY OF ADSORPTION 5 Id Adsorption and affecting factors 5 1.2 Specific surface and adsorption 10 1.3 Polar compounds 13 CHAPTER 2: SOIL STRICTURE 16 2.1 Definition of nature and size 16 2.2 Sandstone and limestone minerals 17 2.3 Clay minerals 18 2.4 Description of soils used in this study 23 CHAPTER 3: PROCEDURE OF STUDY 26 3.1 Determination of soil properties 26 3.2 Determination of soil surface area 28 3.3 Determination of ABS retention 36 CHAPTER 4: RESULTS OF STUDY 39 4.1 Grain size, shape, density and surface area 39 4.2 Morphology 39 4 e 3 Minera logical and chemical composition 46 4o4 ABS adsorption 46 CHAPTER 5: DISCUSSION OF RESULTS 52 5.1 Hydrometer test and swelling of clay 52 5.2 Size and weight properties of soils 53 5.3 Mineralogical composition 59 5.4 ABS adsorption 60 CONCLUSIONS 70 REFERENCES 72 LIST OF TABLES Table Page 1 Approximate values of sand shape factors 11 2 Source and origin of the soils 25 3 Sieve openings size 26 4 Studied fractions of soils 40 5 Orain size, shape, density and surface area of the soils 41 6 Mineralogical composition of the soils 47 7 Chemical analysis of some soils 48 8 ABS adsorption on the soils 50 9 ABS adsorption on the soils at concentrations of 5 and 16 mg/1 61 10 Comparison of ABS adsorption on sandstone and limestone 68 11 The retardation of ABS in soils 69 VI LIST OF FIGURES Figure Page 1 Polar and non-polar molecules between two montmorillonite sheets 15 2 Diagramatic sketches of structural units of clay minerals 20 3 Schematic representation of clay minerals structure 20 4 The Blaine air permeability apparatus 35 5 The batch method equipment - boxes and tubes 38 6 The hydrometer test 30 7 Grain size distribution curve - Hydrometer test 42 8 a-d Photomicrographs of the soil grains 43 8 e-h Photomicrographs of the soil grains 45 9 The change of ABS adsorption with concentration 51 10 Change of soil surface area with grain size 58 11 Intensity of ABS adsorption at concentration of 5 mg/1 62 12 Intensity of ABS adsorption at concentration of 16 mg/1 63 13 Change of ABS adsorption with soil surface area at concentration of 5 mg/1 67 INTRODUCTION Surface active agents have created numerous problems in the disposal of wastewaters and the re-usage of treated effluents. Synthetic detergents amount to about 75 percent of the entire cleaning agents used in the United States and to 90 percent of the soaps and detergents used in households. In volumetric terms, the synthetic detergents amount to 3.5 billion pounds per annum and will surely grow higher (1) (2)*i A package of household detergent product is composed of about 30 percent surfactant - a surface active agent, about 30 percent of phosphate builders and the remainder miscellaneous builders and minor ingredients. Detergents can be divided into four categories: Soap, and cationic, nonionic and anionic synthetic surfactants. While soaps and cationic sur- factants are easily removed from waste, the first by microbial degradation and the second by anionic surfactants present, some of the nonionic and anionic surfactants have been found many times to resist microbial degra- dation and removal from sewage. The fact that only a small portion of detergents on the market is cationic and a much bigger part is anionic enables the possible removal of all the cationic surfactants with an equivalent amount of anionic surfactants. On the other hand, the main part of anionic surfactant ions remains without cationic ions to act with and precipitate together; therefore, the anionic ions cannot be removed in this way. That such a possibility does exist in principle, is shown by a recent study (3). The addition of a cationic detergent to a solution of an anionic detergent produces a turbidity in solution, caused by a colloidal precipitate of anionic and cationic surfactant ions. This turbidity may *1 A figure in brackets indicates a reference number (see list of References). then be removed by a conventional treatment processo The nonionics are not yet very common, but the anionic surfactants are most commonly used in households and laundryo About 70 percent of the anionic surfactants are in the form of Alkyl Benzene Sulfonate (ABS) o What is ABS? ABS is made by polymerizing molecules of propylene gas to form a chain of about 12 carbon atoms associated with hydrogen atoms. This chain is usually branched rather than straight. Next, a molecule of benzene is attached to the chain probably not on the end carbon. Then, the whole is reacted with sulfuric acid or SO- to give it water-soluble characteristics, and finally neutralized with sodium hydroxide to form the sodium salt of alkyl benzene sulfonateo Every molecule of ABS does not have exactly the same structural arrangement or molecular weight, but consists of closely related variants having the same properties (1) (see also section 3.3) o ABS compounds are not readily decomposed by biochemical treatment because of the disability of the bacteria to break down branched chains; therefore, having the property of being surface active, they cause a foam formation first in the wastewater and then in the treated ground or surface water for domestic use. Many surveys and investigations on ABS contamination in untreated and treated sewage, stream waters and water for domestic use are already mentioned in previous publications (1) (4) (5) and there is no need in mentioning them again. However, it is of interest to mention one study, conducted by the Illinois State Water Survey, which has investigated the sources and the extent of movement of ABS in the Peoria, Illinois water supply (6). By examining ABS concentrations in ground water during artificial recharge of water from the Illinois River, the study finds correlation between ABS concentration in the river to its concentration in ground water. A previous survey of ABS concentration in Illinois streams points out a maximum level in highly polluted sewage effluents and surface waters up to 15 and 5 ppm respectively, although the concentration in most of the streams is under 1 ppm (7). Although some of the ABS is adsorbed by the soil, when entering ground water by recharge or infiltration, a significant part remains in the water when pumped out for re-utilization. The regular water treatment including filtration will lower the ABS concentration only slightly; but even though not hazardous to human health, concentrations as low as 1-2 ppm might cause a foaming nuisance and unplea- sant taste* 2 (8) (9). The first report on ground water contamination comes from Long Island, ten years after the use of synthetic detergents started. Many other reports follow. Ground water in Minnesota, Michigan, Wisconsin, Rhode Island and many other states are continuously contaminated by synthetic detergents and result with contaminated ground water, mostly from shallow wells, located not far from small scaled private sewage treatment units. In general, one can conclude that water pollution by detergents from domestic wastewater represents a growing problem, and even the natural filtration of soil is not serving any more as a screen for ABS infiltration into ground waters. In an unsewered area served by private wells and indi- vidual sewage disposal systems of the soil adsorption type, it was found that the water from 11 percent of about 6000 examined wells was contaminated by surfactants of sewage origin (11) c From this study it appears that one of the factors influencing ground water pollution is the character of the soil *2 USPHS Drinking Water Standards recommends a limit of o 5 ppm ABS in drinking water (10) „ Many of the published reports on ground water contamination and previous studies with Ottawa sand in our laboratory show that some ABS, although not all, is retained on the granular media. The front of ABS is much slower than the water front when traveling underground and some of it is retained by the granular media, either in the pores between the grains or by adsorption on the particleso Because of the interest to know more about the retention on soils, and because their character does affect the underground movement of ABS, it is found necessary to study in detail the effect of the soil properties and in particular their mineralogical composition and particle size on the adsorption of ABS The present study is one part of a more extensive research program on ABS conducted by the sanitary engineering laboratory at the University of Illinois (4) (5) CHAPTER 1: TIEORY OF ADSORPTION 1.1 Ad sor pt i o n and affecting factors *3 The adsorption ' of a liquid on a solid phase is defined as the concentration in the form of a film of molecules of a dissolved substance upon the surface of a solid. The surface of a liquid is in a state of strain and that of a solid has similarly a residual field of force; there will consequently be a tendency for the free energy of any surface to decrease, and it is this tendency which is ultimately responsible for the phenomenon of adsorption. The term adsorption refers strictly to the existence of a higher concentration of any particular component at the surface of a liquid or solid phase than is present in the bulk (12). The adsorption at a liquid-solid surface is possibly influenced by a number of complex factors, which can be divided into four groups: a. Solid and liquid surface factors, as surface tension resulting from mass and electrical properties of the two phases; b. liquid factors, as the bulk composition and concentration, relative size of molecules and pH; c. solid factors, as composition of mineral, its crystalline nature and binding forces within the crystal lattice, particle size and surface area; do environmental factors, as temperature, pressure and time. Not all the factors are of the same magnitude of importance and some of them are not yet well understoodo *3 The term "adsorption" was first used by H. Kayser (1081) at the suggestion of E c du Bois-Reymond (12). 1.1.1 Solid-liquid factors Very little information is available concerning the tensions at solid-liquid surfaceso Two aspects may be involved: one is the aspect of chemical forces or molecular attraction (as in the adsorption of gases) and the second, of physical character, involves a surface of large area at which a solute capable of lowering the surface tension may accumulate. The first type of adsorption is also known as chemisorption, since it involves forces of chemical nature similar to those concerned in chemical combina- tions. Chemisorption is indicated by the heats of adsorption which are of the order of 20 to 100 kcal per mole. The bonds formed are almost as strong as those existing in stable stoichiometric compounds and are therefore seldom reversibleo Chemisorption appears especially at higher temperatures (above water boiling point) and may therefore, for our purposes, be ignored. Physical adsorption is due to the interaction of forces between the solid surface and the adsorbate molecules. The adsorption involves Van der Waals forces, which exhibit attraction dimminuation according to the seventh power of the distance between the solid surface and the adsorbate molecules, and makes the Van der Waals forces a dominating factor when distances become very smalU Physical adsorption, compared to chemisorption, is characterized by only small heats of adsorption, about 5 kcal per mole or less (12). Other forces which may be involved are the electrostatic Coulomb forces, whose attraction falls off with the square root of the distance and becomes of minor importance for very small distances. Certain repulsion forces have possibly some influence on the force component as well and diminish rapidly with distance*. 101.2 Liquid factors The composition of a liquid is no doubt a main factor in the adsorption phenomenon, as its properties clearly affect the liquid surface tension. Some liquids, as water and several organic compounds, are known to be composed of polar molecules. The separate discussion on those liquids, which are of special interest to this study, is given in section 2.3. The process of adsorption from a solution increases with concen- tration and reaches a maximum, changing no more with further increase of the solution's concentration. However, this process is mostly reversible, and a definite equilibrium is reached, depending on the concentration of the solution and the quantity of the adsorbent. The classical adsorption isotherm (at constant temperature), also known as Freundlich isotherm, is represented by the empirical equation: 7T=KC m (1.1) m where: X is the mass of solution in gram adsorbed on M grams of adsorbent, C is the concentration of solution at equilibrium in gram per liter, K and m are constant for a given system and temperature. On a log-log scale, the equation takes the form: log -| = log K + m log C (1.2) which gives a straight line. Both molecular size and pH seem to affect adsorption, but as yet, *4 not very much is known . Size may influence the penetration ability of the molecule into clay sheets as well as the mass of the molecule and its electrostatic properties (13). *4 The effect of Ph on the adsorption of ABS on soils i sanitary engineering laboratory of the University of Illinois, as a r>l*>* nf ~- ADC _«...,!.. „-„„-„_ soils is studied at the rv of the Univer part of an ABS study program. 8 1.1.3 Solid factor s The mineralogical composition of an adsorbent has some effect on the adsorption, which increases when clay minerals are involved. Some other influences may be observed as well, as the results of this study indicate (see chapters 5 and 6) . A most important factor to affect adsorption from solution is the available surface of the adsorbent. The surface area of a unit weight of adsorbent increases rapidly with decreasing particle size and is affected to a limited extent by the shape of the particle too. Because of the great importance of the solid geometrical properties, a special discussion is given in section 1.2 . 1.1.4 Temperature and time factors The adsorption from solution decreases with increasing temperature hich shows that heat is evolved in the process (12) . Only few data on the subject are available. A study conducted on bentonite proves the decrease of adsorption with thermal treatment however (14). Time affects the adsorption as long as equilibrium is not reached because of the kinetics of the three mechanisms which are involved in the dispersion of molecules of a solute in the liquid and between the adsorbent particleso One mechanism is dispersion, by which the molecules infiltrate in between the particles with the liquid flow. This mechanism moves the molecules forward into the soil medium and disperses them as well in the form of branches between the grains resulting in both longitudinal and lateral dispersiono The second mechanism is eddy diffusion as a result of turbulance in the flow of the liquid phase. This mechanism mixes the molecules of dissolved substances in the liquid phase and tends to provide a uniform solution As a w result, the probability of adsorption of solute molecules on soil throughout the solution is increased. The last mechanism is molecular diffusion as a result of concentration differences. Even in the absence of mechanical mixing, the concentration of substances that are dissolved in water will eventually become uniform through- out a given volume. Fick's law of diffusion states that the rate of diffusion in time across an area is proportional to the concentration gradient of the substance from a point of higher concentration to one of lower concentrationo This process is extremely slow. As the concentration at a particle surface is essentially zero, the molecules have the tendency to diffuse towards the particle surface, where forces of attraction will then come into actiono When the batch method was used, the molecular diffusion mechanism became very important. (The tubes were mixed each day by hand shaking See section 3.3). Although no mixing, dispersion or eddy diffusion take place between mixing (once a day for five minutes); it is the molecular diffusion which moves the absorbate molecules to the surface of the particleso It is probable that the process of penetrating the fine clay particles by big organic molecules like ABS is especially slow. Therefore, the time needed to reach equilibrium changes from a few hours and less when coarse sand is concerned to many days and even weeks when clays are used as the adsorbent. 1.1.5 Molecular layers The thickness of the adsorbent layer depends on the interface forceso As pointed out by Langmuir, the rapid falling off of intermolecular forces with distance probably causes no more than a monomolecular layer to be adsorbed (15). Even though this point of view is widely accepted, there is an evidence that adsorbed molecules can hold other molecules by Van der Waals' 10 forces, so that multimolecular layers are possible (12) . From one study it appears that one to four molecular water layers are possible (16) (see also section 1.3). 1.2 Specific surface and adsorption The specific surface of a soil is signified by the square centimeters of surface area per gram of soil particles. For very high specific surfaces (10 cm /g and higher), it is more convenient to use m /g as a unit for specific surface. 1.2.1 The effect of shape The shape of soil grains is of great importance, as the surface area per unit volume changes with the shape of the particles (17). A specific grain having a shape of a sphere has the smallest surface area, but will exhibit the greatest surface area when deformed into an extremely thin disc 3 or sheet, e.g., a grain of 1 cm volume, shaped as a sphere (r * 0.62 cm) 2 has a surface area of 4.07 cm ; when shaped as a cube (a * 1 cm), its surface 2 area rises to 6.0 cm ; when shaped as a plate with a thickness of 0.1 cm, 2 (radius of disc ■ 1.73 cm) the surface increases to 21.12 cm , and when the thickness decreases to 0.0001 cm = 1 /j, (radius of disc * 56.42 era), the surface area is as high as 2 x 10 cm or 2 m , still having a volume of 1 cm only. If the surface area S of a particle is defined as: S - ad 2 (1.3) where d is the measured diameter, and a is the surface area shape factor, and the volume V of the particle is defined as: V - £d 3 (1.4) 11 where [3 is the volume shape factor, then, the ratio of surface area to volume becomes: Sal n r v 3 Knowing the volume of W grams of sand with a specific weight y g/cm , finding the mean diameter of the grains d cm, and estimating the shape factor 2 a/p, the surface area S cm can be calculated: Values of a/[3 are given in Table 1 (10). Table 1: Approximate values of sand shape factors Type of sand Spherical Rounded Worn Sharp Angular a/[3 6.0 5.5 5o7 6.2 6.9 In nature one hardly finds perfect sphere-shaped soil grains, but some sands, grains of which have been shaped endlessly by river waters, have a fairly sphere-like rounded shape, like the Ottawa sand from the Illinois River bedo When sandstones, limestones or any other rocks are broken down into small grains by a mechanical device, the formed grains will get a sharp or angular shape. If those grains are then left to natural forces as wind and water, they will worn out in time becoming more rounded. On the other hand, clay minerals have the shape of thin sheets, for which a shape factor can only roughly be estimated. 12 1.2.2 The effect of diameter The strong effect of the diameter of a grain on the specific surface of the soil can easily be seen from the equation 1.6, where the surface area S is inversely proportional to the diameter d, i.e., the specific surface area of a soil is tenfolded when the mean diameter of its grains is reduced by ten and so on, assuming the shape factor remains the same. 1.2.3 The effec t of crystalline structure It must be realized, that the crystalline structure of a mineral influences the specific surface by affecting the shape of the grain. This can be demonstrated by the fact, that a sodium chloride crystal, being cubic, retains its shape even when broken into smaller particles. Thus, the crystal- line structure of clay minerals is the cause for their plate-like shape (see section 2.3.2) . On the other hand, the structure of a mineral affects the net adsorption at least partly, by affecting the penetration of the molecules of a solute into the crystal lattice. The montmorillonite clay mineral is built of book- like sheets, the surface area of which is regarded as internal surface area. The penetration of solute molecules in between the sheets causes the mineral to swell in the f c' axis (section 2o3.3)o On the other hand, the crystalline structure of ill ite and other clay minerals prevents the penetration of such molecules in between the layers and no swelling is observed (section 2o3.4). The phenomena of not being readily wetted is observed for example on hydro- phobic silica gel (19) „ Therefore, specific surface, as such, is not always a good indication of rate of adsorption, and two soils with the same specific surface may differ in the amount of solute they adsorb. 13 1.2.4 The retardation of ABS movement in soils It is observed that the movement of a solute front in an aquifer is always slower than the movement of the water front itself. The ratio of the water front velocity to the solute front velocity depends on the soil and solution properties (4): v Ap, — -i+cF (1 - 7) s where v , is the velocity of the water front w J v is the velocity of the solute front s J A is the adsorption capacity of the soil in /xg/g C is the concentration of the solution in mg/1 3 p. is the bulk density of the soil in g/cm f is the porosity of the soilo By knowing the ABS adsorption capacity of an aquifer soil at a certain equilibrium concentration of ABS, the retardation of ABS movement can be calculated and the approximate time needed for ABS to travel from the disposal point to a well can be evaluated. 1.3 Polar compounds Polar compounds are defined by the dipole character of their molecules, which possess positive and negative charges the centers of which do not coincide. The water molecule is polar, and by far the most important polar compound sorbed by clay minerals. It was on this basis that initial water adsorption was explained (20). Organic compounds, as ethylene glycol (21), glycerol (22), ABS and many others have been shown to be polar compounds *5 Information obtained from Prof. D. Curtin, Department of Chemistry, University of Illinois. 14 tooo It has been shown that the dipole character of molecules is a result of the lack of symmetry of electron distribution within individual moleculeso Such molecules act as if they carried centers of both positive and negative charges. Since the surface of the clay particles is normally negatively charged, the positive ends of the molecules are considered to lie towards the clay surface with the negative ends extending outwards. The initial layer of oriented-dipole molecules may form another surface of negative charges on which another layer can be built. Theoretically, this process of building up layers could be continued, but due to thermal energy, the molecules are in continuous motion which opposes the regular orientation. At the actual clay mineral surface, the molecules are highly oriented, but the degree of orientation decreases with the increasing distance from the surface, as the relative effect of thermal movement becomes greater (23). However, it has been shown that not only the electrostatic Coulomb forces are holding the molecules to the surface, but also Van der Waals 1 forces act between the molecules and the surface, and cause stronger ad- sorption (24). The organic dipolar molecules are found to be oriented between the sheet surfaces of the clay minerals in a position as flat as possibleo In between two montmorillonite sheets, each sheet tends to adsorb a layer of dipoles on both its surfaces, so that between two such sheets a two-layer structure is formed, whereas non polar liquids form only one layer since Only the nondirectional Van der Waals' force of attraction needs to be con- sidered (Fig. 1) (21). Polar molecules penetrate between the sheets of montmorillonite and cause the expansion of its lattice. On the other hand, the structure of 16 CHAPTER 2: SOIL STRUCTURE 2.1 Definition of natur e and size Particle size, size distribution, particle shape and specific surface of the granular media are important factors influencing the physical proper- ties of soilso But what are soils? The civil engineer in general divides the material at the earth's crust into two categories: rock and soils (23). Rock is defined as something hard and consolidated. Soil is defined as a natural aggregate of mineral grains which can be separated by such gentle means as agitation in water (26). Substantially, any loose material regardless of particle size distribution, composition or organic content, is soil to the engineer (23). The water engineer might find special interest in water aquifer materials, such as boulders and gravel, sand, sandstone and limestone. Clay strata avoid any water movement through them. A certain amount of clays can be found in water holding strata consisting of boulders, gravels and sand. Only very small amounts of clay would be found in sandstone strata and none in limestone strata. In nature, rock is broken down into soil particles by natural powers as wind, water and temperature differences. In the laboratory, in order to obtain soils from rock, one has to grind broken rock pieces into soil grains. The engineer classifies the soils in the first stage according to their particle size (whereas the geologist looks to the mineralogical com- position first). There are several different classifications as far as sand and silt are concerned (17) (27), but all agree to define any soil with particles of 2 /j, and less as clay. A clay soil consists of clay and non- clay minerals, where the clay minerals are in amount big enough to influence the physical properties of the soil. Silt ranges from 2 /jl up to 20-100 fi, 15 i 1 lite is fixed so that polar molecules cannot enter between the sheets and cause expansion, but retain around the exterior surfaces only (14) (23) (25). 5 5 & £ Q. Figure 1: Polar and non-polar molecules between two montmorillonite sheets (21) 17 and sand ranges up to 2 mm. Soils with coarser grains are defined as gravel. 2.2 Sandstone and limestone minerals Only the few minerals which compose the sandstones and limestones used in this study are described in the following lines. (The minerals which compose the clays are described in the next section 2.3). 2.2.1 Quartz Quartz is a natural crystalline silica of the chemical composition SiCL. Sometimes it occurs in clear colorless crystals, but more frequently, as a white opaque masso 2.2.2 Mica Mica (meaning: to shine, to glitter) is any of a group of complex silicate minerals that crystallize in thin, somewhat flexible, easily sepa- rated layers; are translucent or transparent. 2.2.3 Feldspar Feldspar is any of several crystalline minerals consisting mainly of alumino-silicates of potassium (K - Feldspar) and sodium (Na - Feldspar). Feldspar is usually glassy and moderately hard and is a constituent of igneous rocks. Plagioclase is a feldspar containing calcium and sodium and having an oblique cleavage. 2.2.4 Glauconite Glauconite is described by various investigators as a complex hydrous iron and potassium silicate, whose formula has not yet been determined (20) (29), but a representative formula is suggested as follows (30): (K ' Ca 0.5' Na) 0.84 (A1 0.47 Fe o?97 Fe o!l9 %.40> (SS 3.65 A1 0.35> °10 (0H) 2 m 18 It is recognized, that glauconite may be formed of one or more clay inerals of the kaolinite, ill ite and glauconite, montmorillonite or chlorite type (31) (section 2.3) * r in the iron-rich samples, two or more clay minerals generally occur together (32). 2.2c5 Dolomite Dolomite, named after the French geologist Dolomien (1750-1801), consists mainly of magnesium and calcium carbonates MgCo~«CaCo~ and is white in color. 2.2.6 Calcite Calcite is a calcium carbonate CaCo~ with a hexagonal crystallization. 2.3 Clav minerals The fact that clay fractions in soils are, in large degree, controlling the physical properties of the soils requires the following detailed dis- cussion. 2.3.1 Size and shape of particles Although there is no sharp boundary between particle size of clays and non-clays, it has been shown, that there is a general tendency for the clay minerals to concentrate in a size range smaller than 2 fj. and larger clay particles break down easily to this size when slaked in water. On the other hand, non-clay minerals usually are not present in particles much smaller than 1 to 2/x. However, most of the so called clay soils contain non-clay as well as clay minerals. Therefore, the best split of the two is obtained at the size of 2 /j, (23) (33) o Most particles larger than about 5 // can be represented as sphere- like or cube-like particles with reasonable accuracy. The mica is excluded because, although not being a clay, it is a thin layer mineral (section 2.2.21 19 On the other hand, most particles smaller than approximately 5 /j, are shaped like a plate, the length or width of which is 5 to 300 times its thickness (34). The plate shape, which most clay particles possess, is well illus- trated by the electron photomicrograph of a kaolinite clay particle with an equivalent area diameter of 1.3/^ and thickness of 0.003/^ (35) 2.3.2 The basic crystalline structure Many advanced instruments and methods are used to study the crystal- line structure of clay minerals. X-ray and infra-rod diffraction studies, petrographic, electron and ultramicroscopes and differential thermal analysis are some of them. Clays are layer lattice minerals, resulting from different stacking arrangements of two basic building blockso The two blocks are described by Grim as follows: "One is an octahedral unit, in which an aluminum, iron or magnesium atom is enclosed in six hydroxvls having the configu- ration of an octahedron (Fig. 2a) . The octahedral units are put together into a sheet structure (Fig. 2b) , which may be viewed as two layers of densely packed hydroxyls, with the cation between the sheet in octahedral cocrdinationc The second building block is a silica tetrahedral unit (Fig. 2c), in which four oxygens or hydroxyls having the configuration of a tetrahedron enclose a silicon atom c The tetrahedra are combined in a sheet structure so that the oxygens of the bases of all the tetrahedra are in a common plane, and each oxygen belongs to two tetrahedra (Fig 2d). The silica tetrahedral sheet alone may be viewed as a layer of silicon atom between a layer of oxygens and a layer of hydroxylso" (36). The thickness of the octahedral layer is 5 o 05 R (1 R c 10 ^ ■* 10~ cm) whereas the second tetrahedral layer has a thickness of 4 C 93 A only (13). The clay minerals are classified into groups, including montmorillonite, illite (hydrous mica), glauconite (a mica-like clay), kaolinite, chlorite and others. 20 O and O = Hydroxy 1 # = Aluminum, Magnesium, etc. a. A single octa- hedral unit b. The sheet structure of the octahedral units O and o and = Oxygen = Silicon c. A single silica tetrahedron d. The sheet structure of silica tetrahedrons arranged in a hexagonal network Figure 2: Diagramatic sketches of structural units of clay minerals (23) L R'nH.O A . I 0-10.5 R V Tetrahedral / Octahedral L 9.5 R Tetrahedral \ M R ' nH 2° \ 7 a. Montmorillonite L Octahedral a etrahedral T c Kaolinite Tetrahedral Octahedral 1\ 10 K / Tetrahedral \ | b. Illite L A Brucite V Tetrahedral / Octahedral 14.2 * Tetrahedral d. Chlorite Figure 3: Schematic representation of clay minerals structure (32) 21 2.3.3 Montmorillonite clay mineral This hydrous aluminum silica mineral has been found in Montmorillion, France and got its name from two frenchmen in 1847. A structural unit of montmorillonite consists of two silica tetra- hedral sheets and an aluminum octahedral sheet between. A hydrous layer having exchangeable cations is closing on the tetrahedral sheets (Fig. 3a). The layers are continuous in the 'a' and 'b 5 directions, and are stacked one above the other in the 'c' direction. The outstanding feature of the mont- morillonite is that water and other polar molecules can enter between the unit layers, causing the lattice to expand in the 'c' direction (swelling). The basal spacing for air dried montmorillonite may therefore vary from 9.5 to 20 m, depending on the number of waters in the hydrous layer, the ex- changeable cations and the Si:Al ratio. The theoretical formula of montmorillonite is (0\\) .SiJ\l.0,ynU^0 t if lattice substitutions are disregarded. Examination of montmorillonite under an electron microscope shows irregular fluffy masses of extremely small particles. In some cases, the larger masses appear to be stackings of flake shaped unitSo Some of the individual particles appear to be about 20 a thick, from which it can be concluded, that at least some montmorillonites break down to flakes approaching unit thickness (13) (23) (32). 2.3.4 Illito cl av mineral Illite has been proposed as a general term for the mica-like clay minerals by Grim et al (37). The structure of illite is similar to that of montmorillonite except that there is always substantial replacement of silicon atoms by aluminum atoms in the tetrahedral layers, and potassiums are between the layers 22 serving to balance the charges resulting the replacement and to tie the sheet units together, so that water and polar ions cannot enter between them and cause expansion (Fig. 3b). The basal spacing of 10 a remains therefore unchanged. A general formula for the composition of 111 ite is as follows: (0H)„K (Si Al ) (AVMg.'Fe.'FeJ o _ . 'y< is less than 2 and fre- 4 y 8-y y 4 y 6 4 6 20 ' quently equal to 1 - 1.5 . The electron micrograph of Fithian ill ite shows small, poorly defined flakes, commonly grouped together in irregular aggregates (23) o The thinnest flakes are approximately 30 A thick and many of them have a diameter of 0.1 to 0.3 micron. This means, that in a volume of i 1 lite there are not nearly as many exposed surfaces, between which water can enter as there are in the same volume of montmorillonite, and therefore, the amount of molecules that the ill ite can adsorb is considerably reduced (13) (23) (32) (36). 2.3.5 Gtauconiite QUy mineral The term "glauconite" comes from the Greek - 'glaukos', which means bluish-green or grey. Glauconite clay mineral is known to be of marine origin and is found to approach the i 1 1 ite in structure (38). A comparison study of glauconite and ill ite clay minerals shows that glauconite grades into il 1 ite in composition and that the aluminum content in the octahedral layer and the total interlayer cations appear to increase with geological age of the samples (38). Glauconite has been reported as belonging to the dioctahedral groups of micas and characterized by replacement of Al +3 by Fe +3 , Fe +2 and Mg +2 (30) „ 2.3.6 Kaolinit e clay mineral The kaolinite mineral is composed of two layers, the silica tetra- 23 hedron and the aluminum octahedron (Fig. 3c). The basal spacing is 7 a and its fixed formula is (OH) 8 Al 4 Si 4 1Q (36). 2.3.7 Chlorite clav mineral Chlorite has, in addition to a sandwich of an aluminum octahedral layer between two silica tetrahedrons, a magnesium octahedral brucite layer on top (Fig. 3d). Its general formula is as follows (36): (0H) 4 (Si.Al) Q (Mg.Fe) 6 2Q -(Mg.Al) 6 (OH) ^ 2.4 Description of soils used in this study In order to get a large variety of soils with different mineralogical composition and particle size, three groups were chosen: siliceous materials, calcareous materials and silty clays* The sandstones and limestones obtained as rock pieces were ground down to sand fractions of different sizes. The silty clays were obtained as gravel mixed with very fine fractions of silt and clay. They were sieved, and the fractions with grains smaller than 74 /u, were used for this study. The mineralogical composition , size distribution, etc. of each soil are given in details in Chapter 4. 2.4.1 Siliceous piatcrjais The following sands and sandstones, consisting mainly of silica inerals, have been used: Ottawa sand, silica flour, Mississippian sandstone, Pennsylvanian sandstone I, Pennsylvanian sandstone II and glauconitic sand- stone. The Ottawa sand has been largely used in previous studies of ABS (3) (4) (39) and seems to fit for comparison into this study too. Silica flour is a ground Ottawa sand. m 24 Glauconitic sandstone is usually green, although different shades and hues of black, green and even white grains occur. Whatever the external color, grains are generally greenish after crushing c The term 'glauconitic sandstone' or just 'glauconite' is used for the field identified grains, and 'glauconite mineral' is used for the specific clay mineral defined before (30) (31) (40). 2.4.2 Calcareous materials The dolomitic limestone, high-purity limestone and oolitic limestone consist of calcite and dolomite with different ratios. The dolomitic lime- stone is dominated by dolomite mineral, whereas calcite is the major consistuant of the other two limestones. 2.4.3 Siltv clays The clay soils which were studied include Peoria clay, Fithian illite and Panther-creek bentonite. A bentonite is composed essentially of a crystalline clay mineral - montmorillonite. Often there is only a very slight non-clay mineral content (29) (36). The source and origin of all obtained soils are given in Table 2<= Table 2: Source and origin of the soils 25 Soil Source Origin SJUiceous material : Ottawa sand Silica flour Mississippian sandstone Pennsylvanian I sandstone Pennsylvanian II sandstone Glauconitic sandstone Calcareo us material : Dolomitic limestone High-Purity limestone Oolitic limestone Silty clays: Peoria clay Fithian Illite Panther-creek bentonite Near Ottawa, Illinois Near Ottawa, Illinois Jackson and Union Counties area, Illinois Jackson and Union Counties area, Illinois Jackson and Union Counties area, Illinois 1500 ft. deep at Section 28, T. 30N,R o 10E, Kankakee County, Illinois From a quarry in southern Cook County, Illinois Jackson and Union Counties area, Illinois Jackson and Union Counties area, Illinois Peoria, Illinois Fithian, Illinois Aberdeen, Mississippi Illinois River bed Illinois River bed Upper Mississippian-Chester Series, outcrop sample Lower Pennsylvanian, outcrop sample Lower Pennsylvanian, outcrop sample Middle Franconian formation (Cambrium) Upper Silurian-Niagaran Series Middle Mississippian Series, outcrop sample Middle Mississippian Series, outcrop sample Formation of an out-wash of the banks of the Illinois River Outcrop sample ♦6 *6 Since this bentonite is a commercial product, its origin is not known. 26 CHAPTER 3: PROCEDURE OF STUDY 3.1 Determination of soil properties Soils which were obtained as rock pieces were first broken down into gravel by using a hammer. Then, the gravel was put into a Lancaster counter current batch mixer which, by breaking it into smaller fractions, converted the gravel into sand and silt (for definitions see section 2ol). Then, the soils were examined for grain size, morphology, specific weight, bulk density, porosity and surface area. The mineralogical composition of all soils was determined by X-ray diffraction studies* A possible chemical composition of some of the soils was suggested by using literature sources. 3.1ol Grain-size All the sands were passed through a series of 4 sieves: Nos 50, 70, 100 and 200. Ottawa sand, because of its bigger natural grains, was passed through sieves Nos. 20 and 30 (Table 3): Table 3: Sieve openings size Sieve No. 20 30 50 70 100 200 Openings in fj. 840 590 297 210 149 74 All silty clays were sieved through a No. 200 sieve, and only the material passing No. 200 sieve was examined. Passing No 200 sieve fractions of some of the soils and of all the silty clays were taken for a hydrometer test. The hydrometer test is a wet method for determination of grain size of soil fractions finer than No. 200 sieve. This method is based on Stokes' equation for the terminal velocity 27 of a falling sphere (34). For both the sieve analysis and the hydrometer test, standard procedure of the American Society for Testing Materials (ASTM) was followed (41). All soil fractions, retaining on sieve No. 200, were washed separately to get them free of fine dusto 3.1.2 Morpholog y Visual observations would describe the general appearance of each fraction of soil, whereas the microscopical observations tried to look at the single grains<> An attempt was made to determine for each soil the fraction which would consist of single unbroken grains of which the original rock was composedo Size, shape and color of the grains were also observed and reportede Color was reported for comparison with other soils. 3.1.3 Mean and median size, effective size and uniform ity The mean size of grains, passing one sieve with openings of d, and retaining on the following sieve with openings of d , is defined as: D = 1 s V s For determining the specific weight, the standard procedure of the ASTM was used (34) (42). 3 The bulk density p, g/cm is defined as the ratio of a dry weight of a soil sample W to the total volume V cm it occupies at 20 C: S V w p b * f (3.5) As V is always bigger than V , the bulk density is always smaller than the specific weight. Porosity f of a dry soil sample is defined as the ratio of the void volume V in the total volume V to the total volume V itself, f is V \ L expressed in percentage: V V -V W f = J- • 100 = -V-5- • 100 - (1 - «-^-) - 100 (3.6) V t V t Vs 3 A soil sample in a 20 cm volumetric flask was vibrated by hand to the 3 constant volume of 20 cm and the weight determined. Knowing the specific weight, the porosity was calculated by using equation 3o6 . 3 «2 Determin ation of soil surface area As described in previous chapters, the specific surface in cm of area 29 per gram of soil, is a most important factor in the mechanism of adsorption of solute molecules on solid particles. Many methods can be used for determining a specific surface of a soil, but there is not even one method with absolute accuracy by which one can find the total specific surface of a soil and be ensured that the value obtained is fixed, accurate and unchangeable. In the first place, each method uses adsorbant molecules of different phases and properties (gas or liquid-phases, polar or non-polar molecules). On the other hand, different mechanisms are involved in different methods (monomolecular adsorption on soil grains or permeability of the soil for air flow). Second, there are soils which have "internal" surface, like the montmorillonite clay mineral. As explained before (section 2.3.3), the ability of certain molecules to penetrate the montmorillonite layer structure and settle at the interior layers surface, causes a tremendous increase of the total surface area covered by the solute molecules. Certain areas in a particle between cracks and splits and in pores can be regarded as internal surface too On the contrary, one speaks about "external" surface of a soil, meaning the entire surface which is surrounding the separate soil grains as such. If, in addition, the grain is built of an expandable lattice or has any penetration splits in its structure, an "internal" surface area can be added to the external surface area of the grain, and together those will present the "total" surface area on which solute molecules may be adsorbed. Finally, differences in surface area result simply from the fact that never is the same sample of soil used, and although equal fractions are pretended, there are differences from one sample to another. Obviously different surface areas are obtained when various grain size fractions are involved. 30 Studying the total adsorption of ABS on soils, the total surface area of a soil is naturally of main interest. Therefore, this study does not distinguish, in its experimental results, between external and internal surface area, but tries to evaluate the total surface area. It should be pointed out, that all the laboratory methods are pri- marily developed for measuring the surfaces of clay minerals. However, this study is equally interested in clay soils and sand soils, and therefore tries to apply the methods used to both clays and sands. In the following sections, some of the principle methods for surface area evaluation are described: 3.2.1 The analytical method If it is possible to measure the size of an average grain and estimate its shape factor, one can use equation 1.6 for calculating the external 2 surface area S cm /g of W grams of soil. Dividing S by W gives the specific 2 surface area in cm /g of soil. In case of clay particles, the shape factor cannot be estimated (section 1.2.1 and 2 3.1) and, assuming the particles to be spheric, one can only estimate a surface area which will always be exceeded by the real external surface area. Internal surface area can be theoretically predicted from dimensions of the lattice structure. In the case of very big grains, i.e., Ottawa sand grains, it may be practical to count a representative number of grains, weigh them and calcu- late the equivalent sphere diameter of each grain. When grains are too small for enabling the practice of this technique, an average size determined through a microscope, or a mean size of a sieved fraction may be used (section 3.1.3). This method does not take in account the increase of external surface area due to the rough face of a grain, and only a rough estimate 31 can somewhat correct the calculation In fact, results have shown that the analytical evaluation of surface area gives the smallest figures in compa- rison to the results of all other methods. 3o2.2 Methods not used in this study The method most widely used previously for the measurement of surface areas of very fine materials involves the adsorption of nitrogen, or similar gases, at low temperatures. When the "adsorption isotherm" data obtained are analyzed according to the theory of multimolecular adsorption (43), they yield a value for the volume or weight of a monomolecular layer of the gas on the external surface of the sample. With a suitable value for the cross- sectional area of the gas molecule, this "mono-layer capacity" can be converted into surface area. The main limitation of this method in measuring surface areas of certain clays is its inability to measure the internal surface area, which is not accessible to the nitrogen molecule. Recently, however, gas adsorption measurements on montmorillonite have been carried out with polar gases, notably water vapor and ammonia, and interpretation of the results has given values for combined internal and external surface areas (44). Another method is the ethylene glycol retention method. Ethylene glycol is a polar organic compound with which a soil sample is saturatedo The glycol at room temperature covers the external surface as well as the internal surface, if any, and its retention is proportional to the total surface area of the sample. When the sample with an internal surface area is heated to 600°C to collapse irreversibly the lattice layers and only then saturated with glycol, the glycol covers only the external surface with a proportional amount retained. The difference between the two measurements gives the retention which is proportional to the internal surface area (45) (46) (47) (48). 32 3.2.3 Glycerol retention method The glycerol method used in this study and prefered because of its simplified experimental procedure, is a gravimetric method based on retention of glycerol at the equilibrium characteristic of monolayer glycerol complexes for measuring total surface area of soils. Approximately 0.2 to 1 gram of the soil, when passing No. 200 sieve, or 1 to 5 grams of the soil, when from a fraction retaining on No. 70 to No. 200 sieve, is air dried and placed in a weighed aluminum foil dish, dried at 110 C to constant weight in about 2 hours, cooled in air for 30 seconds, and weighed on an analytical balance. Weights are reported on the 1 /2, 2, 2 /2, 3, 3 /2 and 4th minute after taking them out of the oven. All weighings are made to 0.1 mg. Five ml of glycerol solution (2 percent, in distilled water) are transferred to the dish from a pipette, and the dish is swirled gently to mix and to distribute the mixture evenly. Care is taken that no unwetted portions of the sample remain. The specimen is then placed in an oven containing a source of free glycerol and heated at 110 C to constant weight. After 24 hours at the condition of constant weight, the weighing procedure is repeated, and glycerol retention is calculated as a percentage of the original oven-dry weight of the sample. Determinations of each sample are made in duplicate or triplicate, and a blank dish is included with each group of samples (49). The glycerol used is of reagent grade. Its formula is CH 2 OH»CHOH«CH 2 OH (or (^H^) The specific gravity of pure glycerol is 1.2609, and that of a 2 percent solution is 1.0030 (50). Upon removal from the oven, the dish and contents cool almost imme- diately to room temperature, enabling weighings to be made quickly. In order 33 to avoid the influence on weight of adsorbed moisture, seven weights during the first four minutes are taken, and the resulting weight-to-time curve is extrapolated to time zero, for which the wanted weight is evaluated. 3 The volume of 1 gram of glycerol at normal density of 1.26 g/cm is 3 o 794 cm . Assuming that glycerol retained on the internal surfaces of montmorillonite at equilibrium is in the form of a monomolecular layer at this density and with a thickness of 4 e 5 a (as indicated by X-ray diffrac- tion), the area of the glycerol film layer may be computed as: Volume 0.794 cm 3 ._,_ 2 , . „, . , " . ' , _H — ■ 1765 m per gram of glycerol retained. Thickness 4.5x10 cm Thus, a glycerol retention of 1 percent is equivalent to a monolayer area of 17.65 m /g. However, for montmorillonite, the internal clay surface area, covered by the monolayer between two basal surfaces of the clay, is twice 3 the area of the monolayer itself, 35.3 m /g. It is further assumed, that a monomolecular layer of glycerol is similarly retained on the external surfaces, and for each 1 percent of o glycerol retention, 17.65 m are covered. These calculations correspond to a cross-sectional area of 26.9 a per molecule of glycerol. Although the value assumed for the density of the glycerol layer at 20 C is used above, the determination of retention is made at 110 C. The density at 110°C is 1.20, about 5 percent less than at 20 °C (44) 3*2.4 Air permeability method The Blaine air permeability apparatus and procedure are described by the ASTM standard (51). This method serves for determining the fineness of Portland cement in terms of the specific surface expressed as total surface area in cm per gram of cement. In this study the Blaine apparatus was used for the measurement of 34 the specific area of sand and clay samples for comparing the results obtained in that way with surface area obtained by other methodso Silica flour 398 with a known specific surface area was used as the calibration standard sampleo The permeability of a laminar flow of air through a medium of soil particles is depending on the surface friction of the particles. As such, the permeability is related to the surface area of the soil and is higher with a lower specific surface. The surface area of the grains react on the passing air and a friction force is created. As greater the surface area, so higher the friction per unit volume, slower the flow and smaller the permeabilityo The apparatus has two main parts. One consists of a U-tube manometer with a positive airtight valve and a rubber ball hand-pump; the other part is a permeability cell. The liquid in the manometer is nonvolatile, non- hygroscopic, and of a low viscosity and density such as dibutylphthalate or a light grade of mineral oil (Fig. 4a). The permeability cell holds an air- 3 dried, weighed soil sample of about 0.9 cm in volume. The sample sits on top of a perforated brass disc covered with a disc of filter paper. Such a paper disc covers then the top of the sample too, before a plunger presses the sample down to a specific volume Mercury is used for accurate measure- ments of volumes in the cell, before and after fixing the sample (Fig D 4b). When the sample is readily fixed in the cell and volumes measured, the cell is put into the coupling on the top of the manometer and liquid is pumped to an upper lineo When the liquid level starts to drop, the vacuum produced in the tube sucks air, which has to pass through the permeability cell and the soil sample. The time needed for the liquid level to decrease between two known levels is proportional to the capacity of air passing the sample, 35 a<> General view of the apparatus b. View of permeability cell, disk, plunger, filter paper, sample, mercury, thermometer and funnel Figure 4: The Blaine air permeability apparatus 36 which on the other hand is proportional to the specific surface of the soil (51): p (1-f ) (f 3 n T) S « s — ? — _ — L_ (3.7) s „ „w;t p(l-f) (f nT ) r s s 2 where: S is the specific area of the sample in cm /g p is the specific density f is the porosity of the prepared soil bed n is the viscosity of air at the temperature of the test T is the measured time interval s indicates the properties and results for the standard sample testo *»3 Determination of ABS retention The ABS used in this study was a normal dodecyl benzene parasulfonate sodium salt, having the formula C, fi H 2 qSNaO„ with a molecular weight of 340 g/mole: HHHHHHHHHHHH H— C -h^ — C — C — C — C — C — C — C — C — C -C — H H H H H H H H H H hAh &. S — ONa The benzene sulfonate group is not in a strict position, but may change along the alkyl chain. The sulfur atom in some molecules is exchanged 35 with the radioactive isotope S , which serves as a tracer. This is done by 35 sulfonating alkyl benzene in sulfuric acid which has S * isotopes in its molecular structure. For the determination of ABS retention on soils a batch technique was used. Soil samples of 1 gram, or 10 grams for Ottawa sand, per plastic test tube were saturated with 40 ml of an ABS feed solution. Four different concentrations of 9.64, 16.80, 35.00 and 53.00 mg/1 were used (section 4.4) 37 The tubes were then put into a batching box with divided cells, each tube into one cell (Fig. 5) . The covered box was shaken by hand each day for about five minutes in order to provide better contact between the ABS solution and the soils. The concentration of ABS feed solution was determined by the methylene blue test (5). The initial radioactivity of the ABS feed solution, as well as the reduced radioactivity in the supernatant of the test tubes after centrifuging, was determined in the following way: 2 ml of the solution were transfered into a 1 /8 inch aluminum planchet, dried at 103 C and cooled before counting in an internal proportional counter. The samples were counted for 2 minutes to get a desirable level of total count in order to limit the statistical error to about 1.5 percent. All counts were then corrected for background and decay. No correction for self-absorption has been made because it was proved to be negligible. By knowing the difference in activity between the feed solution and the supernatant one could calculate the amount of ABS in /ug retained on 1 gram of soil sample and plot it against the concentration of the super- natant in the form of a curve. This was repeated about once a month, until equilibrium between adsorption and desorption on the soil was reached and the curve of the sample would be at its maximum. For this condition, the ad- sorption of ABS on the soil sample was the highest. The adsorption isotherm was plotted using equilibrium values of adsorbate per gram of adsorbent versus equilibrium adsorbate concentrationso 38 Figure 5: The batch method equipment - boxes and tubes Column NOo 4 - Pennsylvanian sandstone I; No. 9 - Pennsylvanian sandstone II; No. 14 - Dolomitic limestone; No. 19 - Peoria clay; No. 20 - Bentonite clay; distilled water Figure 6: The hydrometer test 39 CHAPTER 4: RESULTS OF STUDY 4.1 Grain size, shape, density and surface area The soil fractions chosen for this study are described in Table 4. Each fraction is marked in the table and throughout the chapters by the sieve number, on which it is retainedo The size range of all fractions passing sieve No. 200 is determined by a hydrometer test (Figs. 6 and 7). The values of the various soil properties are tabulated in Table 5. 4.2 Morphology 4.2.1 Siliceous materials Ottawa sand 30 is composed of rounded and edgeless, yellowish white quartz grains with a slightly rough surface (Fig. 8a). Silica flour 398 is a very fine white flour made of ground Ottawa sand. Mlssissippian sandstone 200 is a white brownish sand which under the microscope seems to be composed of single grains^ The transparent, colorless and angular crystals are very homogeneous and have many mirror- like planes. The sand is composed mainly of quartz. Some iron cement of rust-like color is observed on several grains (Fig. 8b). Pennsylvanian sandstone I is a yellow brownish sand. Fraction 70 is composed of conglomerates which are mostly colorless, transparent and angular quartz crystals, joined with iron cement. Fractions 100 and 200 seem to be single-grain fractions. Some flat crystals and many mirror-like crystal planes are observed (Fig. 8c). The pan fraction looks like broken crystal dust. Pennsylvanian sandstone II is a gray brownish sand composed mainly of quartz crystals. Fraction 70 (Fig. 8d) is made of single grains. The 40 Table 4: Studied fractions of soils Soil Mark ' passing No. sieve size fi retaining on sieve Noo size fj. Siliceous materia 'Is; Ottawa sand 30 20 840 30 590 Silica flour 398 200 74 pan — Mississippian 200 100 149 200 74 Pennsylvanian I 70 50 297 70 210 100 70 210 100 149 200 100 149 200 74 pan 200 74 pan - Pennsylvanian II 70 50 297 70 210 100 70 210 100 149 200 100 149 200 74 pan 200 74 pan - Glauconitic 200 100 149 200 74 Calcareous materi aJs; Dolomitic 70 50 297 70 210 100 70 210 100 149 200 100 149 200 74 pan 200 74 pan - High-purity 70 50 297 70 210 100 70 210 100 149 Oolitic 70 50 297 70 210 100 70 210 100 149 Silty clays; Peoria clay pan 200 74 pan - Illite pan 200 74 pan - Bentonite pan 200 74 pan 41 Table 5: Grain size, shape, density and surface area of the soils Mark grain size in Ai Uni- formity *8 Shape factor Specific weight g/cm3 Bulk density g/cm3 Porosity o/o total surface area J n m^/g Microscopical observation Mean or median Effective size Soil Analytical method Blaine air permeability method Glycerol retention method Surface ratio d D m d 10 U a/p 1 Y s Pb f S a S n S 9 Sg|| or - g p— P <;j11ceous materials: Ottawa sand 30 770* 7 705 615 lo3 5o5 2 65 1.77 33 0o003 OoOlS Silica flour 390 - 14 <4 >4.5 (6 9) 2 63 0,495*9 0.039 0c53 1.6 0o5 1.0 1.8 49 _ 2o79 U41 50 _ 0.650 42 65 Illite pan down to 1 10 O 7 19 3 _ 2 78 1.36 51 - 0.966 80 83 Betonite pan down to 1 4o7 80 — 2o76 1.09 60 — 0„605 560 020 •7 This value is not determined by microscopical observation but calculated analytically (section 5.2.1). *0 The values in brackets are estimated on the basis, that fine crystal particles may have the same shape as the bigger grains. *9 This value was received from Ottawa Silica Company as determined by a modified Blaine test. This sample serves as the standard sample for the Blaine test of this study c 42 , i B C h t» 'iH O *J >— t •— I 4-> •«-! ffl >•» >-.•«-" CD 9J B o " «> E •»* *•* o • — I C C «-H O -H c •H 4! 4) O OJ !H C C/5 CL CU Q O. H-t CO s § o o s o "•3" o CM Sujssed auoojsj r V 1 ' f I f L a. Ottawa sand 30 m bo Mississippian sandstone 200 Co Pennsylvanian sandstone I 200 do Pennsylvanian sandstone II 70 Figure 8: Photomicrographs of the soil grains 44 grains are angular and transparent with many mirror-like planeso Some rust- colored opaque particles are covered with iron cement. Fractions 100 and 200 include many broken and some flat grainso The pan fraction is made of broken crystal dust. Glauconitic sandstone , is a gray sand. Under the microscope, fraction 200 seems to be composed of single grains. About 80 percent of the grains are observed as colorless, mostly transparent and angular crystals. The remaining 20 percent appear as dark-green worn and rounded glauconite mineral crystals. They are opaque and have a rough surface full of pores. Many glauconite crystals are broken (Fig. Oe) . 4.2.2 Calcareous materials Dolomitic limestone is beige grayish. Fractions70, 100 and 200 are composed of single, white, opaque and angular crystals, which are mainly dolomite minerals. Some rust-color iron cement is observed on some grains. No sharp edges, corners or crystal planes are observed (Fig. 8f ) . The pan fraction is a crystal dust. iligh-purity limestone is almost white. Fractions 70 and 100 are composed of single grains, mainly calcite crystalso Many crystals are colorless, transparent and angular with sharp and smooth planes. Some are white, opaque and angular crystals with a rough surface (Fig. 8g) . Oolitic limestone is grayish to brownish in color. Fractions 70 and 100 are composed of single grains. The dolomite crystals, described previously, are dominating over the calcite crystals (Fig. 8h) . 4.2.3 Silty clay s Peoria clay is a gray brownish soil. Yellow, rust-brown and black crystal dust is microscopically observedo e. Glauconitic sandstone 200 fo Dolomitic limestone 100 g. High-purity limestone 100 h. Oolitic limestone 100 Figure 8: Photomicrographs of the soil grains 46 IUitc is a gray clay, which under the microscope seems to be composed of white, brownish, reddish and black crystal dust. Bentonite is a brown yellowish clay. A variety of particles changing from white, over yellowish, rust-brown and dark red to black in color, are observed under the microscope. 4.3 Mineralogical and chemical composition Almost all sand fractions of the sands and sandstones, excluding glauconitic sandstone, consist mainly of Si0 2 as shown in Table 6. The limestones, on the other hand, consist to a high percentage of MgO and CaO in various ratioso Some chemical composition data are published and can demonstrate the eventual composition of some of the soils of this study (Table 7). 4.4 ABS adsorption Four different concentrations of ABS feed solution were used in this study. The solutions were prepared so that the approximate concentrations of 10, 20, 40 and 70 mg ABS per liter of solution would be obtained. After mixing, the exact concentrations were determined by the methylene blue test resulting with the four following concentrations: 9.64, 16.80, 35.00 and 53 o 00 mg/1. By choosing those concentrations it was ensured that the equilibrium concentrations of ABS adsorption would be in the range of about 1 to 20 mg/1 which is the range found in ground waters, rivers and sewage. The soils, saturated with all four solutions included Ottawa sand 30, silica flour 398, Mississippian sandstone 200, Pennsylvanian sandstone II 70, 100, 200 and pan, glauconitic sandstone 200, high-purity limestone 100, oolitic limestone 100, Peoria clay, illite and bentonite. After a few weeks, equilibrium between adsorption and desorption of ABS on the sample was 47 Table 6: Minera logical composition of the soils* 10 Soil Mark siliceous minerals calcareous minerals Potassium Plagioclase clay minerals Calcium Quartz Mica feldspar feldspar Dolomi te Calcite montmorillonite Illite Glauconite Kaolinite Siliceous materials: Ottawa sand Silica flour Mississippian Pennsylvanian I Mixed- Chlorite layers Pennsylvanian II Glauconitic 30 398 200 100 200 pan <2^ 70 100 200 pan <2fi 200 pan <2^ Calcareous materials: Dolomitic High-purity Oolitic Siltv clavs: Peoria clay Illite Bentonite 100 200 pan 70 100 70 100 pan <2^ pan <2^ pan <2^ *10 The dominating order of the minerals, composing one soil, is marked by running numbers; s - indicates a small amount present; v - indicates a very small amount present; t - indicates a trace amount present. 48 Table 7: Chemical analysis of some soils Oxides so il Silica Otti ♦11 flour 398 ]1 awa, 111. * Glauconite Norwalk, Wi 12 s. Illite 1J Fithian, 111. *13 Montmorillonite Pontotoc," Miss. Si0 2 99.75 °h 49.4 % 51.22% 57.55% A1 2°3 0o08 10.2 25 91 19o93 Fe ?°3 0.02 18.0 4.59 6.35 FeO - 3d l c 70 0o95 MgO 0.02 3.5 2.84 3.92 CaO 0o04 0.6 0.16 1.94 1^0 - 5.1 6o09 0o59 Na ? - 1.4 o 17 0.33 Ti0 2 OoOl - o 53 o 32 «?° - 8 3 7 14 8.53 total 99.92 99 6 100o35 100 o 41 •11 Analysis received from Ottawa Silica Company ♦12 Taken from reference (30). ♦13 Taken from reference (52) . 49 reached, and the relationship between ABS adsorbed on one gram of soil and ABS remaining in one liter of supernatant was plotted according to the Freundlich equation on log-log scale to obtain straight lines (Fig. 9; section 1.1.2). The retention of ABS on the soils is given in Table 8. 50 CO in E CP o CO 0) § a. u o to rs CO CO < CO cv XI co H 05 E c .28 ♦j • 3 in -H CO O to co »-. o *j CO C o V vO O i-H o o E 0> § e JO* IT) I-H cr— < s «— t OS • \ < S 3 M o CO co C\J CM o CM o NnoHO>o co m >-< to t*- ^- O ■— < O C> rr ^ CO CO CM CO —t CO LO lO CO ry • o • • o • co >o i*- h- r«- cm — • tO O vO CM vO CO CO s CO o- to CM'-th-t-OOCMtfl • •••••-■ rrOO^f-OOCM t-vOt'-NOCO'^'CMCC i- o in t- vo cm co St»ONNH« O CO CO lO tO CO • | •••••• m ^-t h- t*- no cm co O O O CM >0 CO CO ^r co cm h- "v I s - t- CM TT CM CO CM in • | .:•••• co O *3" ry CM <-h — < CM OMO^OCOH • o • o • o I MDOrtH^j r* •— I •—* o —* CO CO CM CM CM CO CO I s - CM f- o o • ry O CM CM • 03 E tO 3 o CD o co co c> o r- o 8 o _ CO CM r- o o « —* CM CXCM c c u CO CO -a 3 •»H •<-t c o c c CO 1— 1 O. CO CO <** >H > (0 f-H CO CO 10 >» 2 u •d CO CO •1-1 CO c *-» 1— I CO c 8 •iH IH CO to s a. to CO •d u V *» CO E o •H CO ♦J 3 •rt O c o o u U CO 3 O CO —* —* co to U O CO r- in • • CM <-< in t*- • • P CM CM CO o © o o u 3 O O.— I «-» JZ iH C^i-H ^H O X © CM CO -h CM O CM • o o rr CO cm cm >o in • • • i— t o ■- « o o- o *H i-l CM rrr-rrcocMototo h- O cm in CO tO CM CM O CO • ••••••• • • • 9 • vomoctoorr© 1- -< co cm<-: «-H l— 1 «-H C4 «"H i-H »— < t—4 i-H TNCM 1- rr <-H i-H >& o rr CO CO CO ^r co in to © t~- I s - in co co rr in in o *■* co in to no in to >l CO o > i-H fH CO CM TT in co <-<© o o • o cm in coco c c c CO CO CO aac. CO l-H Cj co a? c -^ j-> o O -H C OJH C a. •-< c2 B O co «-» c 0) o c . o B U O ~* E ♦J 3 D. -^ >- M O X2 co •»* *a -< CO CO CO. i-H « 3 cr CO o to i-H CV 51 £ *» H-( H-« 4> O to s a. ra c>»— < o ^h c t—4 >H O CD •—• o O X O a. *-« DQ ^mcottWvoncooO'h n n PUBS BMB^ao -10 J /fjUO 9JC0S > 8 6/67/ v - uondaospv a o CM 52 CHAPTER 5: DI9CUSSI0N OF RESULTS 5.1 Hydrometer tost and swelling of clay The swelling of montmorillonite is clearly demonstrated in the hydro- meter test (Fig. 6). The figure shows 6 columns with 5 different soils and one blank, five days after the experiment has startedo In the first four columns, from left to right, one can see an equal sedimentation of 47, 45, 42 and 40 ml whereas the bentonite in column 20 occupies a volume as high as 135 ml. On the fifth day of the experiment, the final data for the grain size analysis was received. Thus, the approximate percentage (in volume) of the settled soil fractions in each column were obtained (Fig. 7). The settled part of bentonite exhibited about 66 percent of the total bentonite volume in the column, whereas 98, 96, 96 and 07 percent were observed for the settling of Pennsylvanian I and Pennsylvanian II sandstones, dolomitic lime- stone and Peoria clay respectively. The extrapolation of the 100 percent settled volume would result with 48, 47, 44 and 46 ml for the sandstones, limestone and Peoria clay, and with 204 ml for the bentonite. As samples were of equal dry weight and had more or less the same specific weight, it has been shown that under wet conditions, the bentonite sample swelled and occupied about 4.5 times more volume than the equal non-expanding soil samples. In fact, solutions with different cations might affect the sedimen- tation of various samples of the same clay differently. In one experiment, ill ite was mixed in six chloride solutions of six different cationso After 6 weeks the ratio of maximum to minimum final sediments volume was 1 to 2 (53). The cations affected the clay's bonds, thus affecting the dispersion and sedimentation of the clay particles. 5.2 Size and weight properties of soils 5.2.1 Gr a i n si z e and shape Ottawa sand has grains big enough to permit an easy counting. There- *16 fore, an average equivalent-diameter of a grain has been evaluated by weighing nearly 5500 grains. The following results were obtained for an average grain: a. In one gram of Ottawa sand there are 1575 grains; hence, b one grain weighs 0.635 mg; 3 c. having a specific weight of 2<>65 g/cm , the volume of one grain is 3 0.240 mm and its equivalent diameter is 0.770 mm; d. the surface area of one grain is 1.06 mm , hence, the specific surface 2 of Ottawa sand is 29.3 cm /g. The reporting of observed grain size under the microscope (Table 5) should only give an idea of dominating grain sizes, but can of course give no average size of a whole fractiono It is obvious, that the mean diameter D as well as the effective size d, and uniformity U (Table 5) of any soil fraction, retaining between the same two sieves, are identical, because the theoretical grain size analysis curves between the same two sieves are identical too. The effective shape factor of a glauconitic sandstone sample is combined from 00 percent of dolomite and quartz grains with a shape factor of 6.9 and 20 percent of the glauconite mineral grains with a shape factor of 5.7. Thus, its mean value is 6.6. The shape factors of the other soils are evaluated through microscopical observation in accordance with Table 1. *16 The equivilent-diameter of a grain is the diameter of a sphere with a volume equal to the volume of the graino 54 5.2.2 Sp ecUAc wg jjght The measurement of specific weight or density of clays might be difficult, because of incomplete penetration of water into the interstices of the aggregate mass. The measurements are particularly difficult for clays in their natural state where bound water is an integral part of the structure, like in montmorillonite. This bound water avoids the penetration of water from the outside which distorts somewhat the results. For clay minerals which are subject to isomorphous substitution within the lattice, e.g., montmorillonite, illite, glauconite, etCo, the density varies from one specimen to another (23). For such clay minerals, a whole range of characteristic values are reported: Illite . Reported values are 2.76 to 3o0 for muscovite and 2.7 to 3.1 for biotite (54); and 2.600 for a sample from Maquoketa Shale, 111. (55) Illites would be expected to be in this range, but perhaps on the low side because of deficiencies of interlayer ions and slight replacement of such ions by hydration (23) <> The value for Fithian illite evaluated in this study is 2.78o Montmorillonite . Reported values for dehydrated material range from about 2.2 to at least 2o7 (55) (56); 2.53 for low iron content montmoril- lonite and 2.74 for a montmorillonite with a somewhat higher iron content (57). Probably, 2.7 is too low for the substantially pure, iron-containing member of the series (23). The bentonite in this study, which is nearly a pure montmorillonite, has a value of 2.76. Glauconite . Reported figures range between 2c2 to 2o0 (32). The evaluated value in this study is 2o76. 5.2.3 Bulk density and porosity Bulk density is a function of the specific weight of the material 55 and its porosity, increasing with the increase of the specific weight and the decrease of the porosity. Porosity depends upon size, shape and arrangement of the grain. In the case of uniform spheric grains, porosity does not depend on grain size and changes from 48 percent with orthogonal packing to 26 percent with rhombic packing. Irregularity in particle shape generally increases porosity. A common porosity of natural sand is about 40 percent (18). From Table 5 one can learn that the porosity of all sandstone and limestone frac- tions ranges between 37 to 48 percent with a mean of 43 percent. The Ottawa sand, because of its rounded sphere-like grains, approaches rhombic packing, and has therefore a somewhat smaller porosityo On the other hand, the clay soils possess a significantly higher porosity. 5.2.4 Specific surface' For the analytically calculated values, the mean or median size is used. Obviously, the figures of equal fractions differ only slightly, because of an equal D and a specific weight of nearly the same value. This kind of calculation is permitted because of the low uniformity factor, which indicates that the size range of the grains is small and the error by using D is neglectable. This error would be most significant in minus 74 /j, m fractions because of the much higher uniformity factor, which is strongly influenced by the smaller particleso The specific surface increaseswith decreasing grain size and high uniformity. Therefore, it is impossible and incorrect to use D as a representative grain size for the evaluation of the specific surface for minus 74 fj, fractionso The Blaine air permeability test is used in this study mainly because of the big differences in the values obtained by the glycerol method in respect to the analytical calculation. Because of this reason, only a few 56 samples have been chosen. The results are not far from the analytically obtained values and differ by a factor of 1.6 and 1.8 for Mississippian sandstone 200 and glauconitic sandstone 200 respectively; by a factor of 5 for Ottawa sand, and are uncomparablc for the minus 74 p, fractions. The value obtained for bentonite represents obviously external area onlyo As such, it is only about 70 percent of the external surface area value obtained for the illite» The low external surface for bentonite is re- ported in the literature as well (44) (45) o The limitation of both the analytical and the Blaine air permeability methods lies in the fact that the obtained surfaces neglect surface rough- ness, cracks, pores and, of course, any internal surface, thus giving a surface area which is much lower than the effective surface available to the penetrating ABS polar molecules. The glycerol method seems to be a good medium for determining the specific surface area which adsorbs the ABS moleculeso This conclusion is fairly proved by the fact that the hydrophobic silica retains glycerol just for covering the low external surface area, as obtained by the Blaine air permeability method On the other hand, the values obtained by the glycerol retention method for all other soils are about 50 to 120 times higher in comparison to the values obtained by the analytical or Blaine methods (Table 5 - surface-ratio column). The surface-ratio values can be divided into two main groups. The first group includes all sandstone samples, for which the surface-ratio changes from 50 to 72 with a mean value of 63o Ottawa sand is excluded because of its very small surface area, which retains too little glycerol for weighing on an analytical balance. The second group includes all limestones, with a surface-ratio between 93 and 120 and a mean of 104c These results indicate the higher "adsorption surface" of the 57 limestones in comparison to the sandstones, a fact which is also observed for ABS adsorption (section 5o4) . The mean surface-ratio of limestones is by 65 percent higher than the mean surface-ratio of sandstones, the reason of which is not yet clearly understood. The surface-ratio is probably independent of grain size, shape and specific weight but surely is related to the mineralogical composition of the soilso The glauconitic sandstone has a surface-ratio of 130 indicating an even higher "adsorption" surface, which is explained by the presence of 20 percent glauconite clay mineral with a high specific surface. The surface-ratio of illite is 83. Illite has external surface only and both the Blaine air test as well as the glycerol retention test measure it as sucho On the other hand, bentonite has a high specific surface due to its internal surface area. The glycerol retention method measures the total surface area of the bentonite, which includes the internal as well as the external surface area. The Blaine air test measures the external surface area only. Assuming that the surface-ratio for external surface area of illite and bentonite is approximately the same, it is possible to calculate the external surface area of bentonite as a part of the total surface area measured by the glycerol retention method. This is done by multiplying the external surface area of bentonite, as measured by the Blaine method, by the surface-ratio of illite. Hence, o 685 x 83 - 57 or about 60 m /g. Thus, the internal surface area of bentonite would be 500 ra /g out of a total o measured surface area of 560 m /g Q The change of surface area with size of grain fractions of the same soil is shown in Fig. 10 o There is, however, a difference in the slope for the sandstones and the limestones. The silty clays as a group have a rate of change more similar to that of the sandstones. The strong deviation of 58 IP" ! 1 1 1 1 1 Mill 1 1 1 I 1 1 1 1 1 - 1 i - — o © O O <3 Q [> i 1 ! . a> — I i c o — h* QJ C 0) *» HH > > > c - u u u b c e a 3 p CO CO CO CO" £ - •*H -i-i -^ O >> tfl Q.B C'H «J t/i t/i >, >> o.coco«-»o-«h cdojco CO •tl > > •»* •«-! M CC«-H 1— 1 / — (fl *H -H c <-> 3 O O O O o — V> >> >» O -«-i Q.-«-" ♦ ■> «-> / _ •^ on in o E 1 «-> «o t/j >, CO o wcc30x:^t; 3 a> •»-> •** «o c e co i-H os— < c e *h u u — — o 4-> •** — o< — • >-• _ - © / - — / — / •e ~~ - / / - — / / — — / / / / - %/, / / A/ © u 3 o - k J 1— » _ - CO u - — / •rt - - I— 1 - — trt — — • — - i 1 i i | 1 i i i i 1 i i 1 1 1 1 1 1 1 1 1 o o o 0) N •tH Vi § c CO M en o *-> in •t-i 3 CO o en >H **^ CO CM B 0) o - 3 O CO M ••H O u »-« CO •»■< o 0) -l 3 CD lO O C CO x: u a 3 o o o o o s o o 8 10 Ul rl a - 8zts ujeiy 59 the limestone slope from the other two slopes is caused by the point which represents the dolomitic pan fraction. The "adsorption" surface of the dolomitic pan fraction is 2 to 2 /2 times lower than that of the sandstones pan fractiono A variety of reported values for surface area of clays and quartz flour can be found in the literature. A minus 2 \jl quartz flour specific surface, determined by glycerol retention method, is reported to be 7 o m /g, which is in good agreement with our figure obtained for a median grain size of 14 fi. A value of 91 m /g is reported for a minus 2 /j, Fithian illite. In respect to this value, our measured value for a somewhat coarser material > n « 10 M and d 1Q (D « 10 fi and d ln - 0,7 fj) is obviously slightly lower - 00 m /g Other o published figures range between 67 to about 100 m /g. With due allowance for possible differences in purity and particle size, our figure agrees well with published values (45) (49) (50) (59). A computation of the theoretical basal-face surface area of pure p montmorillonite gives a figure as high as 010 m /g (45). Other computed 2 values are about 760 m /g (60). The measured external surface area reported ranges from about 15 to 2 100 m /g (44) (45) (50) (60). The value of total surface area ranges from about 400 to above 000 m /g (44) (60). The values are obtained by different methods for different size fractions and types of montmorillonite. Hence, the value for our bentonite seems to be reasonableo 5 «3 Mineral ogi cal composition For almost all pan fraction materials, a minus 2 /ll sample was examined in order to detect clay minerals present. The trace amounts of clay minerals in sandstones and limestones were detectable only after the removal of most of the non-clay minerals by separating the plus 2 /j, particles from the 60 sample (section 2.3.1) 5.4 APS adsorption 5.4.1 Rate of adsorption The adsorption of ABS on the soil sample increases with concentration according to Freundlich isotherm (section 1.1°2 and Fig. 9) Two clearly divided groups can be observed: one group includes all soil fractions retaining on sieve No. 200 and agrees with Freundlich equation for m approximately equal to 0.50; the second group includes all minus 74 jj, fractions for m equal to 1.66 . It is of a certain interest to report the values of ABS adsorption on the soils at concentrations of 5 and 16 mg/1 (Fig. 9, Table 9) . These concentrations may be present in groundwater, rivers and sewage treatment plants, although a lower concentration occurs usually. A comparison study of ABS adsorption on siliceous material shows that in general, the adsorption per gram of soil increases with decreasing of grain size. The Mississippian sandstone 200 is outstanding the above rule because of reasons which are not yet understood. On the other hand, the silica flour 398 proves again its hydrophobic properties by adsorbing less than expected of a flour material. 5.4.2 Intensity of adsorption The intensity I of adsorption is defined as the adsorption of ABS 2 on 1 m of surface area of soil, where the soil surface area is calculated from the glycerol retention method, except for the Ottawa sand 30. The intensity is found to be proportional to the grain size (Table 9, Figs. 11 and 12): I « | = kD (5.1) S m 61 o E D5 E c 9 in o CO c o ft *-» CO 4-> c c o c o 4-> >- O CO "O CO CO O o I— t Si CO H >» o <-» CO •«H CO (A o C (~ O O <-» c c >. O «J •ft -l-l M < <3 E LO I! O B C Q. o co c o 1H O M r3 O w CO < — e CO »i o O O'H «-> co c CD — i I- u o o O C CO 0) o *o £ E o c >, O 4J •(-( «|H 4-> CO a. c Ji o O «-» co C < c o a. (H O CO < CD O CO CO CD <— !m m CO 3 CO J>6 U CO ~J O O CP CM in D3 D5 CM E CO © go CM NOco>0 o O CO O O 1 • • • ■— 1 CM CM ■— ' 1 fH CM ^-» o o o o o >o >o o o CO LO ^" -3* in ^" ! vO O CM O *y ^ f- >© 'g' <=T LO CO TJ« ■— < T o o o q m r~- I s - co cm in cm h- •— • m so vO CO CO 0* r- co co o cm co o s CM o h n m so o CO h- CO t- f>- ry tt CO CM CM CM in o o m m o o co i r- vo co co o cm t» cm r^j rtvON m l-H CO O in -O t- o •*? O O •— • o •— * <—* ~H CO CM CO) O CO o o o o co o o r- c o CO CM ft CM CO O CCM c c Si CO CO -a 3 •iH •«-« c O Q. c ca i — i Q. CO ts> '-4 •<-« > CO i-H a CO co >» _* o •<■< CO co •rt CO c «-> «— ( CO B s •H i-i CD GO S a. a •h tw CD *> CO E O •r* 3 -rt o b a o ^ O CO 3 CO «-H ffl o o g Q o q o o O O O O © Q vO O s O O vO o o CO CO tt cm cm ^r O CO ■o "3< CO Q LO in in CM TT in o co in m co o o o o u 3 o D5'-H •rt o 3; O I I ! I I I I I I t- in cm co m • o • • • co m CM-^O 888 o o o O Q O CM CO ^ HMO ID CO o • CM I s - T >OClH O Q O O O Q v0 o o CM CO CO CM O O TT CO nO in B B B CO CO CO c d. a. 0T| CO 09 >, -H CO C E H fcl O 1-4 -H «J O ,-H 3 rJ OH 0) si cu hs c2 62 J 1 II 1 1 1 1 1 ' ' 1 1 1 ' ' 1 III 1 | 1 1 - O ! 1 _ _ CO - — 13 C - co CO CO _ 3 CO o B - «■> o a \ ©*J CM _\Q 8 o CO CO CM IT) co • o I — o «o c ►H — •tH CO CO — «-» •fH -H - ••H f-H CO C (O CO - O \0 o 2 — © C \ 4> o I®, (3 0- i— i 1 c >» t-t oo 8 CO c o l-H l-H c <-H i-H CM o CO •IH >>!-< U co c «-> t-t •»h •H co •H 4-> u - > i-H 3 TO c o 0> - >» T-S o \ a. — s c ■C CO 3 co c ^ CO \ ° iH l-H - Ou •m i— t ar >> i— i • c I— 4 ~ (A C C •-< CD ♦J •tH OU c CO c c o c CO > l-H >. (A c c , c c o c o o c Jh o (O •a CO V) CO < (A c a *-» E u 3 D5 O o s tn ri q - 9Z|S uje.19 63 TT 1 1 1 1 1 1 1 ' III! 1 MM 1 1 1 1 1 "" ES 3 8 — -132 o — — o c CO o CM - O o c CO Q. o \ i— 1 w (A •rl V) (A o CO o — \o •t-l CO s — o \ u 3 — o8 oX »-H o CO — HMO 0^\ 1 o i-« O © \ •F* c c CM O V- i-H _ CO CO >> CM •vH •*■< -4 4-> o C/5 C C '"H •*« t— 1 ennsylva ennsylva High-pur lauconit vanian I — a, a. e> -• ~ — >> — — (A C - c cu cu - - £ — Q — __ CO z - CM - — o - — II — - •' HH - . 1 1 ! 1 1 , i i III 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 - £ nCM in c o £ O c o •1-1 ♦-> CO ♦-> c a> o c o o re - s Q. O o o ♦* O B CM O 3 o o s LD ?7 q - azjs uteJQ 64 2 where: I is the intensity in /^g/m A is the adsorption in /ug/g 2 S is the surface area in m /g D is the grain size in y, k is a constant. The constant k equals to o 35 and 0.23 for concentrations of 5 mg/1 and 16 mg/1 respectively. It is observed that adsorption intensity of the silica flour 398 is o much higher than expected - 1450 /ig/m . At the concentration of 5 mg/1, the intensities of all minus 74 fj, fractions are far under 100 /zg/m o From the rate of intensity increase with concentration (Table 9) one can assume that the intensity of a minus 74 /j, fraction at the higher concentration of 16 mg/1 would still remain under 200 /^g/m „ The intensity of the silica flour 398 is strongly affected by its very low "adsorption" surface areas, but much less by the low ABS adsorption A D The surface area of the silica flour is about 40 times smaller than the surface area of the Pennsylvanian II pan fraction (both siliceous materials) whereas the adsorption on the silica flour in comparison to the adsorption on the Pennsylvanian II pan fraction is reduced by a factor of about 2 to 3 only* Hence, the intensity of the silica flour must be about 13 to 20 times highero While the limited data presented in Table 9 shows some differences in the adsorption intensities of sandstones and limestones, no definit conclu- sions can be reached as yet. However, the sandstone which contains a high amount of glauconite clay mineral does not show a higher adsorption intensity than for other sandstones. While the adsorption per unit weight is higher on limestones than on sandstones, the increase in adsorption intensity for limestones over the sandstones is no greater than the variation within the sandstones themselves. 65 The unexpected increase of adsorption intensity with grain size is of great interest and should be studied further. This phenomenon can be explained by comparing the amount of ABS actually adsorbed on the soils (A //g/g) to the theoretical amount of ABS required in order to cover the whole surface area of 1 gram of the soil sample by a monomolecular layer (A /zg/g) The molecular weight of ABS is 340 g/mole (section 3.3) and the cross-sectional area of one ABS molecule is 20 a (61). Thus, the surface area covered by a monomolecular layer of 1 gram ABS equals to: (6.03 x 10 molecules/mole) (20 x 10 cm ) Q/17 2, n .- 2, 2 — t — 9 347 m /g = 3. 47 cm /^g (343 g/nole) (10 crn /rn ) (1 molecule) The calculated values for A are qiven in Table 9. m The adsorption ratio A of the actual adsorption value A to the theoretical adsorption figure A , A =A_ (5.2) r A m m is calculated as well (Table 9) and explains the above mentioned phenomenon as follows. It is observed that at a concentration of 5 mg/1 the Ottawa sand surface area is almost (81 percent) covered by a monomolecular layer of ABS. The percentage of coverage decreases with decreasing of soil grain size and reaches a minimum value of o 5 percent for the fine bentonite clay. In other words, O.ftl m out of each m of Ottawa sand surface area is covered 2 2 by an ABS monomolecular layer, whereas only 50 cm are covered on each m of surface area on bentonite, a reduction factor of 162. This is the same factor, by which the adsorption intensity on Ottawa sand is higher than the adsorption intensity of bentonite. The same kind of comparison analysis can be developed for the other soil samples as well. 66 It is to be pointed out that at a concentration of 16 mg/1, the actual adsorption A on Ottawa sand 30 increases to 74 /L^g/g (Table 9) and the adsorption ratio A becomes 170 percent indicating that at this concen- tration Ottawa sand is covered by more than a monomolecular layer. It is generally observed that the increase of intensity with concen- tration is not linear, as would be expected if the Freundlich isotherm applies (equations 1.1 and 5.1): I-!-"? (5-3) When the concentration is increased by a factor of 3o2, from 5 to 16 mg/1, the intensity increase ranges between 1.7 and 2 o with a mean of 2 only (Table 9) The change of adsorption with surface area of the studied soils is illustrated in Figure 13 Although no theoretical justification can be given, it appears that the adsorption A is proportional to the square root of the glycerol measured surface area S: A = b/S (5.4) where k is a constant ■ 320. This relationship exhibits once again that adsorption, although increasing generally with surface area, is relatively lower on materials with a high specific surface like clays, than on materials with a low specific surface like ground sandstones and limestones 5.4.3 Mineralogical composition dependency of adsorption The general conclusion which may be drawn from the obtained data is that for fractions of the same grain size the adsorption on limestones is higher than the adsorption on sandstones, and that among the limestones themselves, the oolitic limestone adsorbs more than the high-purity limestone, CTT i — r o d8 I < L 67 05 E ID o c re re o re iH U> o CNJ o E re LO en l s (A re l-H re •r* *-i 3 >- 3 c C/5 o •— i ■»-> o. u o (A LO ■o • re o in CO < o o c re r- < O If) o CO o u 3 05 O ^H o - uonfoospv 68 although both are composed mainly of calcite (Table 10; see also section 5.2.4 - discussion of glycerol method). Table 10: Comparison of ABS adsorption on sandstone and limestone Concentration Adsorption on soil - A yug/g C mg/1 sandstone limestone Pennsylvanian II 200 High-purity 200 Oolitic 200 5 16 235 375 650 750 590 1070 The glauconitic sandstone, because of its mineralogical composition has a higher "adsorption" surface area than the other sandstones, as measured by the glycerol retention method. About 20 percent of the grains are composed of glauconite clay mineral which, because of its higher surface area, icreases by somewhat the total adsorption capacity of the sandstone (Tables 6 and 9). 5.4.4 The retardation of ABS movement in soils Following the discussion in section 1.2.4 and using the figures from Tables 5 and 9 at a concentration of 5 mg/1, the retardation of ABS in soils is calculated (Table 11) . The most significant factor is the adsorption capacity A, as each of the other factors is in the same range for all the soils. From the results it can be observed that in a coarser material with a higher adsorption intensity the retardation of ABS front is the least, its velocity being only 30 times smaller than the water front velocity. The retardation increases with decreasing of the grain size and adsorption intensity. 69 Table 11: The retardation of ARS in soils Soil Mark Water to ARS velocities ratio v /v w s Ottawa sand 30 30 Mississippian 200 536 Pennsylvanian II 70 208 Peoria claj pan 1465 5.4.5 Fnninecrinq aspects It must be realized that in nature, contrary to the conditions of this study, limestones are not composed of single grains as sandstones and have therefore a lower effective surface area on which physical adsorption can take place. Therefore, from the engineering standpoint, sandstone strata would be a better adsorption medium than limestone layerso The best adsorbing soils would be clays, particularly montmorillonite clays. Rut the permeability of soils with a high clay content is very low to nil and therefore cannot serve ground disposal. On the other hand, in highly permeable coarse soils, ARS travels much faster than in low permeable fine soils, although still many times slower than the water front itself. There- fore, as the actual adsorption of ARS is generally low on aquifer sand and sandstones, the traveling through of ARS is only a question of time. If the water front needs only one day to travel from a disposal point to a well in the near vicinity in any aquifer, ARS would need over a month in a coarse sand aquifer, over a year in a sandstone aquifer and about four years in a silty clay aquifer. 70 CONCLUSIONS 1. Specific surface, analytically calculated, did not represent the total surface area on which ABS could be adsorbed. On the other hand, glycerol, a polar organic compound, was used successfully for the determination of the "adsorption" surface area of soils. The values obtained in that way were comparable with other published figures. The specific surface of soils was found to be inverse proportional to the grain size of their diverse fractions. 2o The change of ABS adsorption with concentration agreed with the Freundlich equation. For the plus 74 /j. fractions of all soils, the adsorption changed with concentration of the power m = 0.56, and m * 1 66 for all minus 74 fj, fractions. 3. The intensity of ABS adsorption increased with the increase of grain size. Thus, the adsorption per unit area on Ottawa sand was about 160 times higher than on bentonite. An increase in intensity with concen- tration was observed as well. 4. The experimental results pointed at the conclusion that no soil was completely covered by an ABS monomolecular layer. However, the per- centage of coverage was proportional to the grain size being the highest with 01 percent for the Ottawa sand, and the lowest with 0.5 percent for the bentonite. 5. ABS adsorption was affected by the mineralogical composition, and was significantly higher in limestones than in sandstones. The adsorption on glauconitic sandstone was affected as well by the mineralogical 71 compositiono However, the adsorption intensities, which are independent of surface area, are within the same range for both sandstones and lime- stones. Ground disposal of wastewater containing ABS cannot prevent the pene- tration of ABS into ground water and thereafter into drinking water because only a small and limited amount of ABS could be adsorbed physi- cally on permeable soil strata. However, the traveling velocities of ABS are many times smaller than those of the water front, and it will take many months or some years before ABS will reach the nearest down- stream well, as practice has shown. 72 REFERENCES 1. Synthetic Detergents in Perspective . Technical Advisory Council, The Soap and Detergent Association (1962). 2. Huddleston, R. L., "Surface Active Agents - Their Relationship to Waste Treatment and Water Supplies", The Industrial Water and Waste Conference at Austin . Texas (June 1962). 3. Samples, W. R., "Removal of ABS from Wastewater Effluent", Jr. WPCF . 34:1070 (Oct. 1962). 4. Ewing, B. B., et al., "Synthetic Detergents in Soil and Ground Waters", Sanitary Eng. series f No. 0, University of Illinois (Jan. 1961). 5. Banerjy, So K., "Effect of Biological Slime on the Retention of ABS on Granular Media", Sanitary Eng. series , No. 10, University of Illinois (Jan. 1962). 6. Vogel, 0. W. and R. H. Harmeson, "ABS in the Peoria Domestic Water Supply", Jr. AWWA , 54:303 (July 1962). 7. Morgan, J. J. and R. So Engelbrecht, "Survey of Phosphate and ABS Concentrations in Illinois Streams", Jr. AWWA. 52:471 (Ap. 1960). 8. Deluty, J., "Synthetic Detergents in Well Waters", Public Health Reports . 75:75 (Jan. 1960). 9. Flynn, J. H. et al., "Study of Synthetic Detergents in Ground Water", Jr. AWWA . 50:1551 (Dec. 1950). 10. "Drinking Water Standards 1961", Jr c AWWA . 53:935 (Aug. 1961). 11. Woodward, F. L., et al., "Experiences with Ground Water Contamination in Unsewered Areas in Minnesota", Am. Jr. Pub. Health . 51:1130 (Aug. 1961) 12. Glasstone, S., Textbook of Physical Chemistry . D. Van Nostrand Co., New York (Jan. 1950). 13. Robinson, B. P., Ion-Exchanoe Minerals a n d Disposal of Radioa ctive Wastes - A Survey of Literature . Geological Survey Water - Supply Paper 1616, Washington (1962). 14 De Castro, A. H. and J. Rodriguez, "Changes in Surface Properties of Homoionic Bentonite on Thermal Treatment", Anales Q E dafol. y Fisiol. Venetaj., 13:225, Madrid (1954). 15. Langmuir, I., Jr. Am. Chem. Soc . 38:2221 (1916). 16 Bradley, W. F., et aK, "A Study of the Behavior of Montmorillonite on Wetting", Z*J&isJL., 97:216 (1937). 17 Baver, L. D., Soil Physics f 3rd edo, John Wiley and Sons, New York (1956). 73 18. Fair, Co M. and J. C. Geyer, Water Supply and Waste-Water Disposal , John Wiley and Sons, New York (1959). 19. Penn, C. F. and M. F. Barada, "Adsorption of ABS on Particulate Materials in Water", Sew, and Inds. Wastes . 31:850 (July 1959) „ 20. Terzaghi, K., "The Physical Properties of Clays", Tech. En?. News , M.I.To, 9:36 (1928). . 21. Bradley, W. F., "Molecular Association between Montmorillonite and some Polyfunctional Organic Liquids", Jr. Am c Chem. Soc , 67:975 (1945). 22. MacEwan, D. M. C, "Complexes of Clays with Organic Compounds", part I, Trans. Faraday Soc. f 44:349 (1948). 23. Grim, R. E., Clay Mineralogy , McGraw-Hill Book Co., New York (1953). 24. Hendricks, S. B , "Base Exchange of the Clay Mineral Montmorillonite for Organic Cations and its Dependence upon Adsorption Due to Van der Waals Forces", Jr. Phys. Chem.. 45:65 (1941). 25. Gregg, S. J., The Surface Chemistry of Soils . Reinhold Pub. corp., New York (1951). 26. Terzaghi, K. and R. Peck, Soil Mechanics in Engineering Practice . John Wiley and Sons, New York (1948). 27. Notes on Principles and Applications of Soil Mechanics . War Dept., US Eng. Office, Fort Peck, Montana (1939). 28. Charrin, V , "Glauconite", Le Genie Civil . 126:109 (1949). 29 Kerr, P. F. and P. K. Hamilton, Glossary of Clay Mineral Names . Preliminary Report No. 1 of Am. Pet. Inst., Research Proj . 49, Columbia Univ. Press, New York (1951). 30. Hendricks, So B. and C. S. Ross, "Chemical Composition and Genesis of Glauconite and Celadonite", Am. Mineralogist . 26:683 (1941). 31. Burst, J. F., "Glauconite Pellets: Their Mineral Nature and Applications for Stratigraphic Interpretations", Bui. Am. Asst. Pet. Geologist . 42:310 (1958) . 32o Wermund, E. 0., "Glauconite in Early Tertiary Sediments of Gulf Coastal Province", Bui. Am. Asst. Pet. Geologist . 45:1667 (1961). 33. White, W. A., Water Sorption Properties of Homoionic Clay Minerals. Report of Investigation 208, Illinois State Geological Survey, Urbana, 111. (1958), 34 Lambe, T. W , Soil Testing for Engineers . M I.T., John Wiley and Sons, New York (Octo 1956). 35. East, W. H., Water Films in Monodisperse Kaolinite Fractions . Sc.D. Thesis, Mol.T., (Aug 1949). 74 36. Grim, R. F., "Physico - Chemical Properties of Soils: Clay Minerals", Jr. Soil Mech. and Found. Div. . Proc. ASCF, V. 05, N. SM2, Part 1, p. 1 (Ap. 1959). 37. Grim, R. F. et al., "The Mica in Argellaceous Sediments", Am. Mineralogist . 22:813 (1937). 38. Smulikowski, K. , "The Problem of Glauconite", Polska Akd. Nauk.. Kom. Geol. r Archiw. Mineralog. . 18:21 (1954) (in english). 39. Wayman, C. H., "Adsorption of Anionic Detergent on Solid Mineral Surfaces", Short Papers in Geology and Hydrology , Article 117, p. C137, US Geological Survey, Washington (1962). 40. Gruner, J. W , "The Structural Relationship of Glauconite and Mica", Am. Mineralog ist, 20:699 (1935). 41. "Tentative Method for Grain-Size Analysis of Soils", ASTM Designation: D 422-54 T, Procedures for Testing Materials . ASTM, Philadelphia (Ap. 1958). 42. "Standard Method of Test for Specific Gravity of Soils", ASTM Designation: D 854-52, Procedures for Testing Materials . ASTM, Philadelphia (Ap. 1958). 43. Brunauer, S. et al., "Adsorption of Gases in Multimolecular Layers", Jr. Am. Chem. Soc . 60:309 (1938). 44. Diamond, S. and F. B. Kinter, "Surface Areas of Clay Minerals as Derived from Measurements of Glycerol Retention", Prop? Clay and Clay Minerals Conf* 5th Conf., p. 334 (1958). 45. Dyal, R. S. and S. R. Hendricks, "Total Surface of Clays in Polar Liquids as a Characteristic Index", Soil Sci. . 69:421 (1950). 46. Bower, C. A. and F. B. Gschwend, "Ethylene Glycol Retention by Soils as a Measure of Surface Area and Interlayer Swelling", Soil Sci. Soc. Am. Proc . 16:342 (1952). 47. Martin, R. T. , "Ethylene Glycol Retention by Clays", Soil Sci. Soc. Am. Proc. 19:160 (1955). 48. Sor, K. and W. D. Kemper, "Estimation of Hydrateable Surface Area of Soils and Clays from the Amount of Adsorption and Retention of Ethylene Glycol", Soil Sci. Soc. Am. Proc . 23:105 (1959). 49. Kinter, F. B, and S. Diamond, "Gravimetric Determination of Monolayer Glycerol Complexes of Clay Minerals", Proc. Clay and Clav Minerals Conf. . 5th Conf., p. 318 (1958). 50# Han dbook of Chemistry and Phy sics f 44rd ed., The Chemical Rubber Publishing Co., Cleveland (1962). 51o "Fineness of Portland Cement by Air Permeability Apparatus", ASTM Designation: C 204-55, Proc. for Testing Materials r V. 4, ASTM, Philadelphia (1961). 75 52. Grim, R. F c and R Q A. Rowland, "Differential Thermal Analysis of Clay Minerals and Other Hydrous Materials", Part I, Am. Mineralogist , 27:746 (1941). 53. Rosenqvist, J. T., "Physico - Chemical Properties of Soils: Soil-Water Systems", Jr. Soil Mech. Found. Div. . Proc. ASCF, 85:31 (Ap Q 1959) 54. Dana, F. S. , Textbook of Mineralogy . W. F. Ford ed., John Wiley and Sons, New York (1921). 55. Dewit, C. P. and P. L. Arens, "Moisture Content and Density of Some Clay Minerals and Some Remarks on the Hydration of Clay", Trans. 4th Intern. Conar. Soil Sci. . 2:29 (1950). 56 Caldwell, o G. and C c F. Marshal, "A study of Some Chemical and Physical Properties of the Clay Minerals Nontronite, Attapulgite and Soponite", Coll. Agro Research Bull. 354 . Univ. Missouri (1942). 57. Makower, P>., et al., "The Specific Surface and Density of Some Soils and their Colloids", Soil Sci. Soc. Am. Proc . 2:101 (1937). 58. Nelson, R. A and S. B. Hendricks, "Specific Surface of Some Clay Minerals, Soils and Soil Colloids", Soil Sci„ 56:2P5 (1943). 59 Brooks, C. S. , "Nitrogen Adsorption Fxperiments on Several Clay Minerals", Soil Sci. . 79:331 (1955). 60. Mooney, R. W., et alo, "Adsorption of Water Vapor by Montmorillonite", Jrc Am. Chemo Soc . 74:1367 (1952). 61. Meader, A. L. and D. W. Criddle, "Force-Area Curves of Surface Films of Soluble Surface-Active Agents", Colloid Sci. Jr. . 8:170 (1953).