Na’ATER ENTERACT EQ‘RS 0N ALCQHQL ACES MQNTMQRILLQNEYE $63K? The“: {‘09 the Dogma of DE. D. MICBEGAN STATE UNWERSETY Robert Hed'iey Dowdy 1966 THESIS LIBRARY ' Michigan State University This is to certify that the thesis entitled Alcohol-Water Interactions on Montmorillonite Surfaces presented by Robert Hedley Dowdy has been accepted towards fulfillment of the requirements for Ph.D. degree in Soil Science Date April 1, 1966 0-169 711 a I lllllw. \ H! a 9.; s surf 1 g ‘ue faces, search the q: nic C. t: .7 X ac .s e S e S 3 3 .I L 5 “Us “U ABSTRACT ALCOHOL-WATER INTERACTIONS ON MONTMORILLONITE SURFACES by Robert Hedley Dowdy Vapor phase adsorption of ethanol and ethylene glycol on homoionic montmorillonite surfaces was studied by infra- red spectroscopy, x-ray diffraction and gravimetric tech- niques. These two compounds were chosen for the following reasons: a) their polar, non-ionic properties; b) to gain an insight into their interactions with montmorillonite sur- faces, the exchangeable cations, and residual water on these surfaces; and c) their extensive use in soils and clay re- search. Calibration techniques are outlined for estimating the quantity of adsorbed water at any given time by use of the 1650-1600 cm'1 deformation band of water. The homoionic montmorillonite surfaces were essentially dehydrated by equilibration with ethanol vapor at a relative pressure of unity or with ethylene glycol vapor at 1150C for 24 hours. Adsorption of these two compounds is reversible to exchange with water at 40% relative humidity with the possible exception of Al-glycol complexes where a low level of ethylene glycol appears to remain in equilibrium with atmOSpheric moisture. However, the adsorption—desorption of U' ) ‘Y' Y 17$ G~ fin ~-qus G‘L‘vO‘ : , v-..y ‘_ d ‘ . “ A K P‘- ‘ . ‘I t" Sta," \- “‘1 ‘ L (7‘ tE'Pe : Robert Hedley Dowdy both ethanol and ethylene glycol is a function of the saturating cation with respect to time, quantity, and type of complexes formed. The rate of loss of ethanol from Cu-montmorillonite during rehydration is a diffusion con- trolled process, and replacement of ethylene glycol by water obeys the conditions of second order chemical kinetics. The lack of uniformity in the shifting of 0-H vibra- tions (stretching and bending) of adsorbed ethanol and ethylene glycol suggest that cation-dipole type bonds, rather than O-H°'°O-clay type interactions, are of primary importance in the binding of these compounds on montmorillonite surfaces saturated with polyvalent ions. The most direct proof of this is the 2750 and 2650 cm"1 adsorption bands in the Cu-montmoril- lonite system which are directly attributable to the O—H stretching modes of coordinated glycol. There was no evidence to support the hypothesis that C-H---O-Clay type interactions are important in the adsorption of non-ionic polar organic molecules on the surfaces. It is pointed out that the x—ray data for ethanol and ethylene glycol systems can be explained by coordination type complexes as easily as by O-H-o-O-clay type bonds. ALCOHOL-WATER INTERACTIONS ON MONTMORILLONITE SURFACES BY Robert Hedley Dowdy A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Soil Science 1966 ACKNOWLEDGEMENTS The author wishes to thank Drs. R. L. Cook, B. G. Ellis, E. C. Doll, J. B. Kinsinger and R. S. Bandurski who, as members of his guidance committee, offered helpful suggestions during the course of study. To Dr. M. M. Mortland, the author wishes to express his sincere grati- tude for his guidance and continued encouragement through- out this study. He appreciates the Graduate Assistantship offered to him by Michigan State University, enabling him to complete this study. To his wife, Annette, he expresses his thanks for her encouragement during his course of study. ii IXTRCD' LI TERA BETHCD. ',v [-4 H ‘ L1 A «‘7‘: a \ - \ ‘q u ‘ ~\‘P.\. TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 LITERATURE REVIEW. . . . . . . . . . . . . . . . . . 3 Adsorption of water. . . . . . . . . . . . . . 5 Adsorption of organic molecules. . . . . . . . 10 Monohydric alcohols. . . . . . . . . . . 12 Dihydric alcohols. . . . . . . . . . . . 16 METHODS AND PROCEDURES . . . . . . . . . . . . . . . 20 Materials and sample preparations. . . . . . . 20 Apparatus. . . . . . . . . . . . . . . . . . . 21 Infrared spectroscopy. . . . . . . . . . 25 X-ray diffraction. . . . . . . . . . . . 25 Procedures . . . . . . . . . . . . . . . . . . 23 Ethanol adsorption-desorption. . . . . . 25 Glycol adsorption-desorption . . . . . . 24 Determination of state of hydration. . . 25 RESULTS AND DISCUSSION . . . . . . . . . . . . . . . 29 Ethanol adsorption-desorption studies. . . . . 29 Hydrated Cu-montmorillonite. . . . . . . 51 Hydrated Al-montmorillonite. . . . . . . 4O Hydrated Ca-montmorillonite. . . . . . . 45 Hydrated Na-montmorillonite. . . . . . . 48 Hydrated NH4-montmorillonite . . . . . . 51 Deuterated ethanol studies . . . . . . . 54 Acetaldehyde adsorption. . . . . . . . . 60 General considerations . . . . . . . . . 62 Glycol adsorption-desorption studies . . . . . 7O Cu-montmorillonite . . . . . . . . . . . 7O Al-montmorillonite . . . . . . . . . . . 76 Ca-montmorillonite . . . . . . . . . . . 78 General considerations . . . . . . . . . 81 The 1575 cm-1 band . . . . . . . . . . . . . . 88 The 1480-1400 cm-1 band. . . . . . . . . . . . 90 GENERAL DISCUSSION AND SUMMARY . . . . . . . . . . . 94 CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . 102 LITERATURE CITED . . . . . . . . . . . . . . . . . . 104 APPENDIX . . . . . . . . . . . . . . . . . . . . . . 112 TABLE F¥ LIST TABLE 1. Infrared frequencies ethanol. . . . . . . 2. Infrared frequencies C2H50D and CH3CD20H . 5. Infrared frequencies ethylene glycol. . . 4. A comparison between OF TABLES techniques for determination of the water con- tent of homoionic montmorillonite. . . . . . . Page of pure and adsorbed . . . . . . . . . . . . . 50 of liquid and adsorbed . . . . . . . . . . . . . S7 of liquid and adsorbed . . . . . . . . . . . . . 71 infrared and gravimetric 113 iv LIST OF FIGURES FIGURE 1. 10. 11. Schematic diagram of the system used for study- ing the adsorption of ethanol vapor on thin clay films. . . . . . . . . . . . . . . . . . . . . . Relationship between the water adsorbed on homo- ionic montmorillonite (300 basis) and the area of the 1650-1600 cm-1 deformation band of water. Infrared Spectra of ethanol-Cu-montmorillonite complexes. . . . . . . . . . . . . . . . . . . . A: 001 Spacing Kg. relative pressure of ethanol for Cu-montmorillonite. B: Ethanol adsorption isotherm for Cu—montmoril- lonite. . . . . . . . . . . . . . . . . . . . Infrared spectra of ethanol-Al-montmorillonite complexes I O O O O O O O I O O C I O O O O O O O A: 001 spacing Kg. relative pressure of ethanol for Al-montmorillonite. B. Ethanol adsorption isotherm for Al-montmoril- lonite. . . . . . . . . . . . . . . . . . . . Infrared spectra of ethanol-Ca-montmorillonite complexes. . . . . . . . . . . . . . . . . . . . A: 001 spacing Kg, relative pressure of ethanol for Ca-montmorillonite. B: Ethanol adsorption isotherm for Ca-montmoril- lonite. . . . . . . . . . . . . . . . . . . . Infrared spectra of ethanol-Na-montmorillonite complexes 0 O O O O O O O O C O O O O O O O O O O A: 001 spacing gs, relative pressure of ethanol for Na-montmorillonite. B: Ethanol adsorption isotherm for Na-montmoril- lonite. . . . . . . . . . . . . . . . . . . . Infrared Spectra of ethanol-NH4—montmorillonite complexes. . . . . . . . . . . . . . . . . . . . Page 22 27 32 36 41 44 46 47 49 50 52 LIST OF FIGURES - Continued FIGURE 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Adsorbed ethanol as a function of (log log Po/P + 2) for NH4-montmorillonite. . . . . . . A: 001 Spacing ZE- relative pressure of ethanol for NH4—montmorillonite. B: Ethanol adsorption isotherm for NH4-mont- morillonite . . . . . . . . . . . . . . . . Infrared spectra of Al- and Cu-montmorillonite after exposure to C2H50D and CH3CD20H. . . . . Infrared spectra of Cu-montmorillonite after exposure to ethanol and acetaldehyde . . . . . Variation.in water content as a function of ’ the relative pressure of ethanol for montmoril- lonite O O O O O O I O O O O O O O O O O O O O Adsorbed ethanol as a function of the evacua- tion time for homoionic montmorillonite. . . Absorbance of the CH3 deformation band of ethanol as a function of rehydration time for ethylated, homoionic montmorillonite . . . . . Log (1 - A/AO) as a function of time for Cu- montmorillonite. . . . . . . . . . . . . . . . Infrared spectra of glycol-Cu—montmorillonite complexes. . . . . . . . . . . . . . . . . . . Adsorbed glycol as a function of time of atmospheric exposure for Cu—montmorillonite. . Infrared spectra of glycol-Al-montmorillonite complexes. . . . . . . . . . . . . . . . . . . Adsorbed glycol as a function of time of atmos- pheric exposure for Al-montmorillonite . . . . Infrared spectra of glycol-Ca-montmorillonite complexes. . . . . . . . . . . . . . . . . . . Adsorbed glycol as a function of time of atmOSpheric exposure for Ca—montmorillonite. . vi Page 53 55 58 61 63 66 67 69 72 75 77 79 80 82 LIST OF FIGURES - Continued FIGURE 26- 27. 28. 29. Curves showing variation of (001) spacing as a function of the mole ratio of glycolzwater for homoionic montmorillonite. . . . . . . . . Reciprocal concentration of adsorbed glycol as a function of time of atmOSpheric exposure for homoionic montmorillonite. . . . . . . . . . . Infrared spectra of Cu—montmorillonite after exposure to D20 vapor o o o o o o o o o o o o o Absorbance of the 1575 cm‘1 deformation band of water as a function of the absorbance of the 1480-1400 cm‘1 carbonate absorption band for Cu— and Al-montmorillonite . . . . . . . . . . vii Page 83 87 89 93 A I m- INTRODUCTION Reactivity at the clay surface is an area of scien- tific research being actively pursued in numerous disci- plines. Since clay surfaces are dynamic in character and represent one of the major reactive components of the soil mass, studies of reactions and interactions between clay surfaces and inorganic and organic substances have been of great importance to soil science. Water is an integral part of these surfaces under natural conditions and in most laboratory systems. Since water is an active component of the clay system, its behavior during physical and chemical reactions must be known and understood before such phenomena as ion exchange, hydration, pH, and adsorption of organics can be explained. The use of simple alcohols, such as ethanol, is common in research involving clay minerals; often with little regard to its affect upon the clay surface and the ions satisfying the cation exchange sites on the surface. The di- and trihydric alcohols, ethylene glycol and glycerol, respectively, are routinely used in specific surface determinations and x-ray identification of clay minerals. With the relatively recent introduction of infra- red Spectroscopy as a tool in the study of surface chemistry, it was considered useful to attempt elucidation of the a? 2) H; c with interact w complex 5;; ViIQ""‘Qnt kinky questions: 1) How are these simple alcohols adsorbed and retained by clay surfaces; and 2) How does the interaction of these alcohols with clay surfaces affect the residual water? With a better understanding of how these Simple alcohols interact with clay surfaces, it will be possible to study more complex systems. Such a system might be the plant root en- vironment which contains organic materials in various states of decomposition, including degraded nucleic acids, proteins, and polyhydric polymeric "alcohols" such as the complex carbo- hydrates. This would lead to a better understanding of: a) the rooting habits of plants and b) the phenomena of soil aggregation. ‘\ of these I Thermodyna taneously the total the solid 31' liquid The ways, '1‘: One appr: (N b” iE‘,’ t at \ 7"» LI TERATURE REVI EW The reactivity of clay surfaces is a well established phenomenon. It is well verified that as the specific sur- face of clay minerals increases, the ability and tenacity of these minerals to adsorb substances are greatly enhanced. Thermodynamically, the free energy of a system will Spon- taneously seek a minimum. Hence, when adsorption occurs, the total free energy of the surface is reduced by replacing the solid-gas or solid-liquid interface with a liquid-liquid or liquid-gas interface. Theories of adsorption have been grouped in various ways. Toth (1955) has outlined three different approaches. One approach is the potential theory which postulates at- tractive forces producing an adsorption potential which exists at a finite distance from the surface. This adsorp- tion potential iS defined as the amount of work required to remove a molecule from this area to an infinite distance from the surface and predicts an adsorption layer several molecules in thickness. Another approach is the chemical theory which considers adsorption as a chemical reaction occurring on the surface of the adsorbent. The reaction product will be the most insoluble or least dissociable complex obtainable in the given system. The third consider— ation is the electrical theory approach which treats the 3 adsorbent as a charged species that will attract an approach- ing dipole. By this theory the energy of adsorption decreases rapidly with distance from the surface, but extends into the multimolecular layer range. A somewhat more meaningful classification is to divide adsorption into the two categories of physical adsorption and chemical adsorption or Simply chemisorption. This has been done by Barrow (1961) for the adsorption of gases on solids in the following manner: Physical adsorption Heat of adsorption less than about 10 kcal/mole Adsorption is appreciable only at temperatures below the boiling point of the adsorbate The incremental increase in the amount adsorbed in— creases with each incre- mental increase in pressure of the adsorbate The amount of adsorption on the surface is more a function of the adsorbate than the adsorbent No appreciable activation energy is involved in the adsorption process Multilayer adsorption occurs Chemisorption Heat of adsorption greater than about 20 kcal/mole Adsorption can occur at high temperatures The incremental increase in the amount adsorbed increases with each incre— mental increase in pres- sure of adsorbate The amount of adsorption is characteristic of both adsorbate and adsorbent An activation energy may be involved in the adsorp- tion process Adsorption leads to, at most, a monolayer Although most adsorption processes involve either physical adsorption or chemisorption, it Should be pointed out that often both types of adsorption occur in experiments designed ‘ x to establish the relationship between the amount of gas ad- sorbed on an adsorbent and the equilibrium vapor pressure in the system. Adsorption onto clay minerals involves both of these phenomena. The various types of molecular adsorp— tion on clay minerals have been presented in considerable detail by Marshall (1964). Adsorption of water. The fact that the surfaces of clay minerals adsorb and retain water has been well established. This subject has been reviewed from time to time as new findings are reported. Probably, the most thorough and comprehensive review of the state of adsorbed water has been presented by Martin (1962). Early, Bouyoucos (1936) attempted to characterize the hygro- scopic moisture in soils. He observed that 95 per cent ethanol will extract all water from bentonite that is removed by heating to 1100C and concluded that hygroscopic water exists as physically adsorbed water. Since 95 per cent ethanol will not extract water from hydrated CuSO4, but heating to 1100C will remove 37.8 per cent of this water, Bouyoucos concluded that ethanol will not remove chemically bound water. Marshall (1936) summarized earlier research of Nagelschmidt, who noted that the first few molecules of water adsorbed on dehydrated montmorillonite expand the clay platelets by con- siderably more than the volume of water adsorbed. Marshall then hypothesized: 1) the first few water molecules con-‘ gregate around the cations or oxygen atoms holding the layers together, and 2) the number of molecules entering the first stages of adsorption will be approximately proportional to the base exchange capacity. Hendricks and colleagues (1940) were one of the first to carefully study the hydration of montmorillonite as af- fected by various saturating cations. Observing the low temperature endothermic peak obtained from differential thermal analysis (D.T.A.) data for Ca-montmorillonite, they noted a double peak at relative humidities less than 40 per cent. They concluded that the higher temperature portion of the doublet is due to the hydration of the cation. This was confirmed by peak area calculations showing that the water content attributed to this portion was essentially the same as that required for the hexahydrate of the calcium ion. The lower portion of the peak occurred at a temperature 40°C lower and was attributed to water lost from the clay surface not contiguous to the exchangeable cation. At relative humidities of 40 per cent and greater, the dual peak develops into a triplet with the third portion being attributed to an additional layer of water. Magnesium and other alkaline earth salts of montmorillonite exhibited a similar behavior. To study the affect of exchangeable cations and cation exchange capacity upon free energy, heats of reaction, and entropy changes occurring during water vapor adsorption, Barshad (1960) used homoionic samples of three different bentonites saturated with various cations. He concluded that the magnitude of change in these thermodynamic quantities is much greater for the interaction of water with the exchangeable ion than for the interaction of water with the oxygen surfaces. The presence of residual water in montmorillonitic and vermiculitic clays above temperatures of 1100C has been rather elegantly shown by use of infrared spectroscopy. Fripiat et al. (1960) have followed the dehydration of homo- ionic montmorillonites and vermiculites with infrared and x-ray techniques. Even though d(001) spacings collapsed to less than 10kx at 1500C, these workers observed "free water" at temperatures up to 4000C. This residual water was thought to occupy the empty hexagonal holes in the surface oxygen sheets of the lattice. Dehydration of the clay system was shown to be a function of both the cation and type of mineral. It was also shown that dehydroxylation of the clay lattice is initiated before dehydration is complete, showing the extreme affinity of the exchangeable cation for water of hydration. Using thinner clay films, Russell and Farmer (1964) have obtained similar results with the exception that the residual water was lost at somewhat lower temperatures. However, they noted a better correlation between firmly held water and the exchangeable cation than was noted by Fripiat et al. (1960). Further evidence that residual water exists on base saturated clay minerals and that its reactivity is a function of the exchangeable cation is presented by Mortland et al. (1963). ' a \ Studying ammonia adsorption, these researchers noted the conversion of adsorbed ammonia to ammonium at temperatures up to those of lattice dehydroxylation. They noted that residual water was the only source of protons in base saturated systems. Turning to water vapor adsorption on colloidal Size material, Benesi and Jones (1959) observed that a monolayer did not form on a silica gel. Therefore, capillary condensation occurred prior to the formation of a true monolayer. They also concluded that water vapor was physically adsorbed since the fundamental v2 vibration of adsorbed water was very Similar to that of liquid water. Studying Na-zeolite, Szymanski et al. (1960) obtained similar results when the relative humidity was greater than 30 per cent. However, when the relative humidity was less than this, the fundamental 1640 cm“1 absorption band Shifted to 1690 cm-1 showing the water to be strongly associated with the adsorp- tion Sites (Na+ ions). Mering (1946) noted the stepwise hydration of mont- morillonite and pointed out that the equilibrium water con- tent at a given relative humidity was a function of the saturating cation. From x-ray data he observed that Ca-mont- morillonite forms a two layer hydration complex as soon as the octahedral hydration Sphere of the calcium ion is satis— fied (18% relative humidity); whereas, Na-montmorillonite will form a complete monomolecular layer of water between the platelets. He suggests that the first layer of water in Na—montmorillonite may be hexagonal in nature, but doubts that the Silicate lattice possesses much "organizing action." The x-ray and water adsorption data of Mooney et al. (1952) Showed that changes of Slope in the adsorption isotherm and changes in the number of layers of water between platelets occurred simultaneously. AS others have done, Puri and Murari (1964) attempted to calculate the surface area of soils and clays from a single point on the water isotherm. They concluded that a relative pressure of 0.53 was required to form a monolayer of water, a value essentially twice that noted by others (Quirk, 1955). Noting considerable data that suggests the amount of water adsorbed by various clays is dependent upon the saturating cation, Quirk (1955) quite appropriately stated that when ". . . polar molecules are be- ing adsorbed by clays it does not seem valid to Speak of a monolayer." In a series of papers, Orchiston (1953, 1954, 1955, 1959) noted that the adsorption isotherms were sigmoidal in form indicating a multimolecular adsorption of a physical nature. He proposed the following mechanism for water vapor adsorption on clays: a) first, water clusters around the active adsorption sites (exchangeable cation): b) these clusters grow, finally running together to approximate "monolayer" coverage, with some multilayering around the active Sites; and c) condensation occurring with further ad- sorption. At this point, the heat of adsorption approaches the heat of liquidification. Gillery (1959) studied the 10 desorption of water from Na- and Ca-saturated montmorillonids. His results were very Similar to those already discussed except that he observed a stable one—layer (12.3 3) Ca- saturated Species at 5 per cent relative humidity. Consider- ing the rates of adsorption of water vapor on degassed Arizona bentonite, Anderson and Sposito (1963) concluded that water was probably being chemisorbed onto the exchangeable cation. In conclusion, it appears well verified that some type of stepwise hydration of expanding alumino-Silicates does occur as water is adsorbed on dry clay. Also, it is obvious that the hydration of clays is a function of the exchangeable cation, particularly that moisture retained against heating to 1000C. It is this water that has proved to be so chemically reactive. Adsorption of organic molecules. Since it is not the intent of this study to consider the adsorption of ionic Species, this review will not con— sider the vast amount of literature concerning the adsorption of ionic or easily ionizable organic molecules. For the most part, they are adsorbed by straight-forward Coulombic forces. Adsorption of nonionic organic molecules is more complex and not as well understood. The forces involved in adsorption of undissociated molecules are principally H—bonding and van der Waals type forces. One of the more extensive series of studies on the adsorption and retention of nonionic organics on montmorillonite has been conducted by Brindley 11 and his colleagues. Using Na-, Ca—, and Mg—montmorillonite, they (Brindley and Ruston, 1958) noted no difference in the adsorption of the polyethylene glycol ester of oleic acid from aqueous suspension. It was concluded that the adsorp— tion processes were the same for all three species. An aliphatic chain, 5-6 carbon atoms in length, was required to initiate adsorption of the various alcohols, glycols, acetones, and ketones from aqueous suspensions (Hoffmann and Brindley, 1960). Adsorption increased with increasing chain length and CH activity. CH activity is increased by increas- ing the number of electron withdrawing groups adjacent to a methylene group. From x-ray data and molecular models, Brindley and Hoffmann (1962) concluded that polar aliphatic molecules are oriented with the plane of the carbon chain parallel to the silicate surface. This would be anticipated from the tetrahedral coordinating nature of carbon, if the assumption is made that the polar group is bonded to the clay surface. Aliphatic molecules without polar groups appeared to have the plane of the carbon atoms perpendicular to the silicate surface. The reason for this orientation was not clear unless it would provide a more favorable organic- organic interaction. Infrared spectroscopy (IR) was used by Tensmeyer et al. (1960) to study the fundamental vibrations of ketones adsorbed on Ca-montmorillonite. Two observations were made with respect to adsorbed ketones: a) no change in peak position 12 was observed between the adsorbed and nonadsorbed states, and b) Spectra of adsorbed, one layer complexes were very similar to those of the solid material, suggesting solidifi- cation at temperatures above the normal melting point. Therefore, adsorption of ketones is physical in nature and organic-organic interactions are the predominant forces in- volved. The extinction coefficient of all vibrations decreased exponentially with increased surface coverage, save the 2913 cm"1 symmetrical methylene stretching vibration (Hoffmann and Brindley, 1961b). Folman and Yates (1959) also noted little change in the apparent extinction coefficient of the C-H stretching vibration of acetone adsorbed on porous silica glass. The absence of change in the extinction coefficient with increasing surface coverage can be considered evidence that CH groups are not directly involved in adsorption mechanism. Monohydric alcohols. Normal monohydric alcohols longer than ethanol will extract interlamellar water from halloysite and montmoril- lonite (MacEwan, 1948). Nonpolarizable molecules will not form layer complexes since the energy liberated by adsorption is insufficient to overcome the Coulombic attraction between the saturating cation and the clay surfaces. MacEwan (1948) concluded that the adsorbed layers were extremely labile and liquid in nature for the following reasons: a) close packing of the molecules is not required to fit x-ray data; 13 b) ease exchange of one complex for another by Simply wash- ing with another miscible liquid; and c) no simple integral number of molecules existed per unit cell as expected in a crystalline system. This suggests a two-dimensional liquid, the same conclusion suggested by Martin (1962) for adsorbed water. Working with vermiculite and three montmorillonites, Barshad (1952) found that interlayer expansion was greatly dependent upon the extent of dehydration prior to contact with n-alcohols, acetone and ether. Expansion also decreased with decreasing lattice charge. Glaeser (1954) studied the vapor phase adsorption of methanol, ethanol, acetone, and water on Ca— and Na-montmorillonite. She observed that the amount of adsorption and the ability to retain these com- pounds was a function of the exchangeable cation. Two lines of proof were presented: 1) the higher the ionization po- tential, the greater the adsorption, because of the increased ability of the ion to capture d-orbital electrons of oxygen and hence complex organics; and 2) the greater the exchange capacity, the greater the adsorption, because of the presence of more cations. She concluded that C-H---O-clay bonds (discussed later) were not sufficient to explain the differ— ences observed. Barrer and MacLeod (1954) studied the vapor adsorption of ethanol, water, ammonia and pyridine on de— gassed montmorillonite. The first additions were adsorbed on the external surfaces. At a given threshold pressure, the penetration forces of the molecule overcome the attractive 14 forces between lamellae allowing interlamellar adsorption. Finally, as P/Po-—->-1, capillary condensation occurs and the amount sorbed tends to increase without limit. If ad- jacent layers are held apart by tetra—alkyl ammonium ions, interlamellar sorption occurs with the first additions of gas (Barrer and Reay, 1957). Since intermolecular O-H-~-O bonding occurs in liquid alcohols, Emerson (1957) suggests that O-H---O-clay bonds would be more reasonable than C-H---O bonds for interlamellar adsorption. The O-H---O bond must be linear and should have a tetrahedral angle since the surface oxygens are already tetrahedrally bonded to two silicon atoms. Using van der Waals dimensions and known bond angles, the observed and calculated d-spacings for such a configuration are in good agreement. For one layer com- plexes, the d-spacing for methanol was 12.7 A and for higher polyvinyl alcohols, the value was 13.6 A. These values are essentially the same as those observed by Brindley and Ray (1964) for primary monohydric alcohols ranging in length from two to eighteen carbon atoms, as well as earlier values reported by MacEwan (1948), Barshad (1952), and Glaeser (1954). Using Ca—montmorillonite and x-ray diffraction techniques, Brindley and Ray (1964) observed that the d-spacing for a one layer alcohol complex at temperatures above its melting point was larger than that observed for temperatures below the melting point. The observed phenomena were reversible and suggested that re—orientations occur in organicficlay complexes ationS i from chm configur ahmdnum 0 m If equili than 3.63:3 7) and ’ . n A ‘10 ( EWRetric aCCOrding CH3 \ c 15 as a function of temperature. They propose various configur- ations to fit the observed x-ray data, which differ slightly from the Emerson model, but maintain the same O—H-o-O-Surface configuration. Greenler (1962) exposed A1203 to ethanol vapor at a pressure of 25 mm. Hg. After equilibration at 350C and evacuation, a series of IR absorbing vibrational bands were observed that correspond very closely to the spectrum of aluminum ethoxide and suggested the reaction: CH3 CHQ \ CH2 + Al—O-Al-O CH2 V l Al-O-Al-O If equilibration is carried out at temperatures greater than 1600C, two new bands appear in the IR Spectrum at 1572 and 1466 cm‘1 which were assigned to the asymmetric and symmetric vibrations of "acetate—like" surface compounds according to-the equation: CH3 CH3 \ \ ____ - CH2 + Al—O-Al-O *a> c----0 l2 . +5H 0 Al-O - Al -0 H These results were confirmed by use of deuterated (OD) and C13 isotopes of ethanol and exposure to D2 gas. The same results were presented by Boreskov et al. (1964). Both groups of researchers obtained similar results for AlgOg-MGOH interactl‘ surface CI Dihydric E Bra; base exche and expel the uncert distance 1‘ of a surfa a strong 0 for variou expected, are imposs C‘H"'O-SL; 132g 9318113,. V' V 4 HCYA’EVEr I T! 16 interactions. Uvarov (1963) postulated the same Al—O-C2H5 surface compound using a surface deuterated alumina. Dihydric alcohols. Bradley (1945) observed that glycols do not enter into base exchange, but can be adsorbed from dilute solutions and expel interlamellar water from montmorillonite. Admitting the uncertainities of Fourier Sketches, he noted that the distance from the center of the aliphatic Chain to the center of a surface oxygen is about 3.3-3.6 X. This is too long for a strong O-H---O-surface bond and does not vary significantly for various glycols and polyglycols where such would be expected, nor with their dimethyl ethers where such bonds are impossible. For these reasons, he concluded that C-H°'°O-Surface bonds are important and comparable in bond- ing energy to O—H---O bonds in natural water systems. However, Tettenhorst g£_§l. (1962) could not observe any good proof of C-H---O—surface interactions from IR studies of montmorillonite-polyalcohol complexes, including ethylene glycol. A slight decrease in the intensity and broadening of the symmetrical CH2 vibration suggests a weak C-H°--O bond with no Shortening of the bond, Since the peak position did not Shift. Studying six different dihydric alcohols, Brindley (1956) noted that the 001 lattice Spacings of alle- vardite remained essentially constant even though the chain length increased from two to twelve carbon atoms. A one- dimensional Fourier synthesis Showed the plane of the carbon 17 atoms of ethylene glycol to be perpendicular to the basal plane of the lattice. Ethylene glycol has been used extensively for Specific surface determinations because of its apparent ability to form a monomolecular film on clay surfaces. All of these procedures and modifications of the one proposed by Dyal and Hendricks (1950) have involved calibration techniques between the cross-sectional area of the molecule and the surface area of known Specimens. One of the major problems involved has been that of controlling the vapor pressure at levels low enough to obtain only monolayer coverage on the clay surface. This aspect, as well as various procedures, has been recently reviewed by Mortland and Kemper (1965). In ethylene glycol-water-montmorillonite systems Mackenzie (1948) observed a constant 17.1 R basal Spacing over a wide range of glycol:water ratios. Similarly, Tettenhorst §t_§l, (1962) observed that the basal spacings obtained with ethylene glycol— or glycerol-montmorillonite complexes were independent of the initial water content. Morin and Jacobs (1964) have also shown no difference in the amount of ethylene glycol adsorbed by clays when the initial moisture content ranged from 1 to 12 per cent. On the other hand, Martin (1955) observed a 56 per cent increase in the ethylene glycol adsorbed by a Ca-montmorillonite dried over P205 compared to that adsorbed by a moist sample. He concluded that clay has the same affinity for ethylene glycol as for WE samples V H31 montmori] g lycero l alcohols. with Na, amount of pendent o montmoril clusion, \ cations we 18 as for water and that the reduced glycol retention in moist .samples was a result of fewer "free" adsorption Sites. Hoffman and Brindley (1961a) have Shown that Mg— montmorillonite will adsorb very little ethylene glycol or glycerol from aqueous solutions but forms complexes upon drying because the water evaporates much faster than the alcohols. They also reported that montmorillonite saturated with Na, Ca, or Mg showed no significant differences in the amount of organics adsorbed. Hence, adsorption is inde- pendent of the exchangeable cations for these homoionic montmorillonites. MacEwan (1948) had reached the same con- clusion, while agreeing with others that the exchangeable cations were influential in clay-water complexes. In contrast to these findings, Bower and Gschwend (1952) observed considerable variation in the amount of ethylene glycol retained by different homoionic Wyoming bentonite systems. This variation was observed for preheated samples as well as those not heated prior to adsorption. Dyal and Hendricks (1952) also found that the amount of ethylene glycol adsorbed by wyoming bentonite was a function of the saturating cation. It was this type of results that led Quirk (1955) to conclude that surface areas determined by ethylene glycol retention may be in error Since it, like other polar molecules, tends to be adsorbed around cations on the clay surface. Recently, McNeal (1964) concluded from D.T.A. data of glycolated montmorillonite systems that ethylene ‘\ .1 \ lch 1 r E \I) and henCE montmoril water-ms!“ ation of cation ma the glyco the clay that th to two mo two mo lec: These res; who noted Mith the c Work et}: ,1 m l ene 21 19 glycol retention is dependent upon the saturating cation and hence, is held at various energy levels. The glycol- montmorillonite D.T.A. tracings were Similar to those for water-montmorillonite complexes, suggesting that the associ- ation of these two polar molecules with the exchangeable cation may be similar. If he made the assumption that all the glycol retained in Na-systems was adsorbed directly on the clay surfaces, then from retention data he calculated that the "excess" glycol present in the Ca-systems amounted to two molecules per calcium ion. He concluded that these two molecules were associated directly with the calcium ion. These results are in agreement with those of Mortland (1954) who noted that the total specific surface correlated well with the cation exchange capacity of both soils and clays. Working with vapor adsorption of the bromide analog of ethylene glycol, ethylene dibromide, on dehydrated montmoril- lonite, Jurinak (1957) showed that ethylene dibromide is adsorbed only on the external surfaces of Mg-, Ca— and Na- montmorillonites when compared to ethylene glycol retention. In conclusion, it appears very likely that the adsorp- tion and retention of Simple organics, possessing functional alcoholic groups, is in some way influenced by the exchange- able cation and/or the force fields of these ions, deSpite some data to the contrary. Whether the presence of adsorbed water is influential, and in what way it may be involved in the adsorption and retention of mono- and dihydric alcohols on clay surfaces is even less clearly defined. METHODS AND PROCEDURES Materials and sample preparations The montmorillonite used in these studies was H-25 from the John C. Lane Tract, Upton, Wyoming, and was supplied by the Ward's Natural Science Establishment. This was a subsample of the wyoming bentonite recently characterized by Ross and Mortland (1966). From chemical analyses, they calculated the formula to be: [Al F [Al 5 1.55 8.16 ”9.551 ;.05 15.95] 010 [032] + “.52 Homoionic clays were prepared by treating the < 2.0u fraction with appropriate chloride salt solutions in excess of the cation-exchange capacity. After flocculation, the super- natant was removed and the procedure repeated three times. The suspensions were then washed five times by centrifugation to remove excess salts. They were then considered free of excess salts as confirmed by a negative AgN03 test for chlorides in the supernatant. Dialyzing against distilled water was avoided since infrared (IR) results indicated that slight dissolution of the clay lattice occurred before the resistance of the dialyzate reached that of distilled water. 20 til! ray redu: never st: were pre; The Stored or England N grade th1 1e58 than 21 Thin self-supporting clay films (4/2 mg./cm?) were prepared by evaporating 10 ml. of a freshly prepared sus- pension of appropriate density in aluminum foil dishes. Copper films were prepared by using aluminum dishes lined with a polyethylene film. To minimize Side reactions, that may reduce the purity of the films, clay suspensions were never stored for longer than two months and fresh films were prepared immediately prior to use. The ethyl alcohol (ethanol) used was absolute and was stored over anhydrous MgCO3 in an air-tight container. Anhydrous CH3CH20D and D20 were obtained from the Volk Radio- chemical Co., while CH3CD20H was obtained from the New England Nuclear Corp. The ethylene glycol used was reagent grade obtained from the J. T. Baker Chemical Co. It contained less than 0.08% water and will simply be referred to as "glycol." Apparatus A schematic diagram of the system used in the ethanol adsorption studies is presented in figure 1. This system, which was Situated in a constant temperature room (200C) and connected to a rotary vacuum pump, contained four basic components: 1) a vapor source, 2) a detachable IR cell to support clay films for spectroscopic examination, 3) a Specimen holder for x-ray diffraction examination, and 4) a calibrated quartz helix balance from whiCh clay films were suspended to obtain sorption data. Manometer readings as 22 .Hopoaonda hunches .2 “cannon Hosanna .m «Hana command“ .H ”human amoaaan .h «awesome hantu .9 «Edam Adan .o ”mundane Hades «panda .m “nowonpdq wasdaa dd eomuosaa done .4 .maaau made man» do Homo» Hosanna mo noapnnomuu one wnahuSpm non com: Bowman on» no adnwuae oapuaonom \J _J.l_ 30001. 000.0 0%. o. o o .o Io. ono O O U . .. ..ofi&§ .lljl..I..1..1111111.L11.11.1.411.111J (G 23 well as displacement measurements of the quartz helix were taken with a cathetometer. Infrared Spectroscopy. An evacuable brass cell fitted with sodium chloride windows was used in the work reported. A calibrated 4 thermistor and a 80/20 nickel-chromel resistance wire, cap- able of heating the cell to 1000C, were components of the brass cell. Infrared spectra (4000 to 600 cm'l) were re- corded by a Beckman 1R7 spectrophotometer fitted with a sodium chloride prism and grating. The clay films were positioned normal to the beam during scanning operations. X-ray diffraction. X-ray data were obtained with a Norelco diffractometer equipped with an evacuable chamber constructed by the R. L. Stone Co. This chamber had a sample holder which could be heated to temperatures up to 11000C and was equipped with a thermocouple. Values for temperature were obtained from potentiometer readings. The chamber was fitted with an open— ing which allowed it to be connected directly to the system. Procedures Ethanol adsorption - desorption. Portions of identical clay films were placed in the IR and x-ray sample holders and suspended from the quartz helix. A moisture determination was made on a fourth portion. After degassing by rotary pump against a liquid N2 cold trap, 1\. the SYS t 8‘ vapor Pre r y at aopro used at e, hours. p, necessary films, A? were furt) t0 rebg,,qy_. ‘Ius< 24 the system was exposed to successively increasing ethanol vapor pressures up to a relative pressure (P/Po) of unity at approximately 200C. A two hour equilibration period was used at each level of ethanol, except at saturation pressure . when four hours were allowed for equilibration. Distension of the quartz helix indicated that equilibrium was reached during the first thirty minutes. Initially, after degassing, and after equilibration at each ethanol pressure, the appro- priate measurements were taken, including x—ray and IR scans. After these treatments, the same samples were sub— jected to degassing against a liquid N2 cold trap for ten hours. Periodically, the evacuation was interrupted to take necessary readings and make appropriate scans of the clay films. At this point, the films in the x-ray and IR cells were further evacuated while being heated to n/BOOC, scanned, then heated to A/100OC for 20 minutes during evacuation, and once more scanned. After degassing, the films were allowed to rehydrate in the atmosphere of approximately 40% relative humidity. Again, the films were weighed and scanned at appropriate time intervals until all the ethanol had been replaced by water, as indicated by IR data. Glycol adsorption - desorption. Clay films were saturated with glycol by suspending them in a vacuum desiccator over a free glycol surface. After evacuation for ten minutes, the desiccator was placed (H 41 in a +— ment. tt: cooling: atmosphe nd IR a periodic w renydrat Determin, Th. obtained helix, a: adsorbed weights c net value materia 1 In Order Crganic n 25 in a 1150C oven for 24 hours. Following the heat treat- ment, the desiccator was allowed to cool to 20°C. After cooling, the glycol saturated films were exposed to the atmosphere (200C and 40% relative humidity). Weight, x-ray, and IR analyses were made immediately upon exposure and periodically with time as the clay films were allowed to rehydrate. Determination of state of hydration. The values for the weight of ethanol adsorbed by clay, obtained from the distension data of the calibrated quartz helix, are actually net values representing the ethanol adsorbed minus the water displaced. In the case of rehydra- tion of films saturated with either ethanol or glycol, the weights obtained by use of an analytical balance are also net values representing the weight of adsorbed organic material plus the weight of water present at any given time. In order to arrive at a value for the absolute quantity of organic material present at any given time, it was necessary to estimate the amount of water present and apply this cor- rection to the experimentally obtained net value. This was accomplished by taking identical clay films and dividing them into three portions. One portion was placed in the IR cell, another was suspended from the quartz helix, and a moisture determination made on the third portion. Then by continuous evacuation, interrupted periodically for simultaneously scanning the 1650-1600 cm"1 26 deformation region of water and taking the necessary weight reading from the quartz helix, it was possible to make a plot of peak intensity versus amount of water present. The plots for Cu-, Al-, Ca-, Na-, and NH4- montmorillonite are presented in figure 2. Since water retained by montmoril- lonite in the air-dried state is known to exist in two phases (Russell and Farmer, 1964), it was necessary to plot peak area rather than optical density of the water deformation band. The areas were determined by tracing them on tracing paper and then cutting out the areas and weighing them on an analytical balance. It should also be pointed out that the origin is considered a theoretical point and is based upon the assumption that dehydration of montmorillonite is com- plete at 3000C. To a first approximation, this appears to be a reasonable assumption Since most differential thermal analysis data (Greene-Kelly, 1957) indicate that the water of hydration is lost at temperatures less than 3000C. In the case of NH4-montmorillonite, the moisture determinations were based on 1100C and graphical extrapolation techniques used to determine the water loss between 1100 and 3000C, Since NH4-clays decompose in this temperature region. A Lambert-Beer's type plot using the origin and the peak optical density of the air-dried film was found to be unreliable. This was expected since Folman and Yates (1959) had observed an exponential decrease in the apparent extinc- tion coefficient of the N-H stretching vibration of ammonia 27 .uosa: no snap aoapaanoeoe -ao ooo_-omo. one so «and cap and Amanda ooomv opanoaaanospnoa canoaoSon um anemones Hope: on» noospon nanmnoavmaom .m oHSmAh To 0953 2.8 no meg; m. a. m_ o. m G a m o I _ _ _ q I a q t O .\.1 .pqozuams:-:;:2> New .paostamlu-:;|s axm\ow\ r- om .paozuao|||;.|lq xxx ox . a .paos-H4:.::--iw txv. \\\\\ u .pnozusolllllo ex r. u 1 co. m“ H 7w 0 J - R. m 1 com .. (m: s 1 0mm 1 con 28 with increasing surface coverage of porous silica glass. They also reported an increase in the apparent extinction coefficient of the C-H stretching vibration of methyl chloride and acetone with increasing surface coverage. Likewise, Hoffmann and Brindley (1961b) noted an exponential decrease in the extinction coefficient as the surface cover- age of Ca-montmorillonite by 2,5-hexandione and 2,5,8-non- anetrione increased. This simply says that with increased Surface coverage, the molecules are less energetically held, resulting in less perturbation and a smaller extinction coefficient. The curves presented in figure 2 clearly demonstrate two well documented phenomena: a) the two phase nature of the rather tightly bound water of hydration, and b) the differential effect of the saturating cation on this water, particularly that water retained at 1100C. The weights of water, ethanol, and glycol presented in these studies are based on clays dried at 3000C. RESULTS AND DI SCUSSI ON Ethanol adsorption - desorption studies The infrared absorption frequencies of ethanol in both the pure and adsorbed states are given in table 1, as well as the band assignments for the various vibrational modes. It should be pointed out that the cited vibrational frequen- cies of pure ethanol are the same as those noted in this study. The vibrational bands Shown for adsorbed ethanol represent an average value for the various homoionic clay systems studied. The broad O-H stretching band of adsorbed ethanol in the 3400—3310 cm'”1 region of the Spectrum is some— what overlapped by the O-H stretching band of adsorbed water. Assignment of the O-H in-plane deformation vibration of ethanol was a point of uncertainty for a number of years. Plyler (1952) assigned this vibration to the observed 1391 cm‘1 band. However, with the use of deuteroethyl alcohol (EtOD), Krimm §£_al, (1956) have Shown that the broad doublet 1 is due to in liquid ethanol with maxima at 1410 and 1330 cm‘ a mixing of the C-H and O-H bending modes. Upon dilution, less interaction occurs and these bands separate into a 1384 cm"1 C-H bending and a 1250 cm'1 O-H deformation vibration. When dueterated, the 1250 cm"1 band is removed to 895 cm"1 (Stuart and Sutherland, 1956), while the 1384 cm'1 band 29 30 Table 1. Infrared frequencies of pure and adsorbed ethanol. Infrared bands Mode References Adsorbed Pure (observed) ____________ Cm-l___________ 3680 (vapor) OH stretch Plyler (1952) 3310-3400 3000-3500 OH stretch Coburn & (liquid) Grunwald (1956) 2984-2990 2984 asym.CH3 stretch 2936 asym.CH3 stretch Drushel et al. 2915-2938 (1963) 2898 sym. CH3 stretch --- 1923 (vapor) CO+CC stretch Plyler (1952) 1478—1485 1467 CH2 scissoring Drushel et al. (1963) 1455 1456 asym.CH3 bend Plyler (1952) 1400 1398 sym. CH3 bend Barrow (1952) 1410 Stuart & --- doublet Sutherland (1956) 1330 (liquid) 0H in-plane deformation Tarte & 1265 1243 (vapor) Deponthiere (1957) -—— 1070 CO stretch Krimm et_a_l_o ___ 885 cc stretch (1956) 31 remains unchanged. These results were confirmed the follow- ing year by Tarte and Deponthiere (1957). It should be noted that any band occurring in the 3700-3600 cm‘1 and the 1200-600 cm‘1 regions of the spectrum are masked by strong fundamental vibrations of the clay lattice. Hydrated Cu-montmorillonite. When air-dried Cu-montmorillonite is exposed to in- creasing pressures of ethanol, the water of hydration is displaced as Shown by the decrease in absorbance of the 1632 cm‘1 deformation vibration (curves A, B and C, figure 3). The replacement of water by ethanol is probably a mass action type reaction resulting in a moisture content of less than 0.7% at a relative pressure of unity. The reduction of absorbance of the 1632 cm-1 band is accompanied by a Splitting of the band with a maximum at 1598 cm‘1 and the development of a shoulder near 1640 cm‘l. Such a Splitting or separation is indicative of the presence of two phases of water in hydrated Cu-montmorillonite, Similar to that proposed by Russell and Farmer (1964) for alkali and alkaline-earth metal-montmorillonite complexes. These two types of water could be thought of as: 1) molecules coordinated directly to the cation and 2) molecules in an outer coordination Sphere or loosely adsorbed on the clay surface. The latter group are less energetically bound and predominate in the air—dried state giving rise to the 1632 —1 cm maximum. Upon dehydration, the more loosely held 32 .mcapwon and .wcammmmoc .noapohdpmm moan poems mazes em you meonmmoSpo op emmoewo .m “wnammwmoe mafia: 0000— on coeds: none .ooapons team mopm nouns menon o. commemoc .m «noapoHSpom moan Hound Moon P commwm :me .n "moan oHSmmmnq m>apwamn 0.. ow comomxo .o “mopm oHSmmonn o>apoaon mw.o op emmomsm .m nemaheunam .4 "opadoaaanoSpnoSISO Ho mnpooem condemnH .m onsmam Arsevmmmmzazm><3 oo~_ Gen. ooc. con. oowrll:| 005. com. com. ooo~ oovm 00mm can» coon e _ . .llllJ il,1 a a Jilit I I.-. 6%: 4. Qmmmxllj . 88. u can. _ «no. cmn. _ 8...: Fe II lib b? OON. 006. 00¢. Don. 00w. 005. 000. Com. OOON OOQN coca OONn Gown L 1 1 _ _ NOISSI WSNVHI 33 molecules are replaced by ethanol resulting in a weak shoulder at 1640 cm-1. In the dehydrated state a portion of the directly coordinated water remains and is responsible for the 1598 cm'1 band. Why the maximum Should occur at this low frequency is not directly apparent. This is very close to the 1595 cm"1 deformation band of free, unassociated water molecules (Redington and Milligan, 1962). Two other observations are noteworthy: 1) ordinary, hexagonal ice has been reported to have a deformation band near 1580 cm‘1 (Ockman, 1958; Haas and Hornig, 1960), and 2) certain hydrated salts have been reported to have bending modes of water at frequencies as low as 1550 cm‘l. Some of those reported are LiOH'HgO at 1586i8 cm-1 (Jones, 1954); ZnSO4-7H20 at 1590 cm“1 (Gamo, 1961); and CuC12'2H20 at 1550 cm"1 (Rundle g£_al. 1955). Therefore, it appears reasonable to conclude that this residual water is quite energetically bound to the c0pper ion and that in the combined electric field of the cation and clay lattice, the v2 bending vibration is much more an- harmonic in nature as suggested by Hornig g§_gl. (1958) for the v2 mode in crystalline water. Curves D and E, figure 3, Show the retention of ethanol, against 10 hours of degassing and heating to 1000C under vacuum, by the persistency of C-H vibrations in the 3000-2900 cm'1 stretching and 1480-1400 cm"1 deformation regions. It should be noted at this point that the C—H vibration, as well as the alcoholic O-H deformation near 1250 cm‘l, in curves 34 B and C, are predominantly due to vapor phase vibrations of the vapor surrounding the clay film and consequently mask the vibrational character of the adsorbed phase. The asym- metrical CH3 bending of adsorbed ethanol has shifted to 1403 cm"1 from its vapor phase position of 1394 cm"1 upon evacuation and heating to 1000C, while the EtO-H bending has shifted from 1242 to 1265 cm‘l. This is a Shift in the direc- tion and of the order of magnitude expected for ethanol mole- cules coordinated through the oxygen atom. The 1265 cm-1 EtO-H band is intermediate between that found in the pure liquid, which is actually a mixing of CH and OH modes, and that observed in the free vapor state. This suggests that in the adsorbed state there is little possibility of CH-OH interactions and that consequently the 1265 cm"1 band is purely EtO-H bending. 7 The C-H stretching vibrations of adsorbed ethanol are very similar to those found in the pure liquid or gaseous states. Detection of Slight frequency shifts in this region is difficult for two reasons: 1) the wavenumber per centi- meter of chart paper is one-fourth that used in the deform- ation region, and 2) the inability to resolve asymmetrical CH2, symmetrical CH3, and/or symmetrical CH2 stretching vibrations. A slight Shift from 2984 to 2990 cm“1 is indi— cated for the vapor to adsorbed phase transition of the symmetrical CH3 stretch. In curve E, figure 3, the weak Shoulder at 2864 cm“1 is a situation where it is possible 35 to resolve the symmetrical CH2 stretching mode (Pozefsky and Coggeshall, 1951). Very little information can be obtained from the O-H stretching region since water and ethanol occur SimulL taneously on the clay surface. Curve F, figure 3, Shows the rehydration of the ethylated Cu-montmorillonite upon exposure to the atmosphere at 200C and 40% relative humidity for 24 hours. The replacement of ethanol was essentially complete in 70 hours as indicated by the absence of the 1400 cm"1 CH3 deformation band which proved to be the most sensi- tive vibration to the presence of ethanol in these studies. Thus it can be readily seen that the adsorption of ethanol on Cu-montmorillonite surfaces is a reversible process. The adsorption isotherm for ethanol on Cu-montmoril- lonite is presented in figure 4B. The most striking feature of the isotherm (solid line) is the linear relationship exist- ing between the amount of ethanol adsorbed and relative pressure. It should be noted that the solid line represents the absolute or total amount of ethanol adsorbed after being corrected for water loss, while the broken line shows the apparent sorption. This linear relationship indicates that the sum of the forces which must be overcome for adsorption to occur, must vary in a Similar manner over the pressure range considered. The components of the total force do vary from one region of the isotherm to another and include such components as the cationic hydration energy, van der Waals 36 .—;$——-e 16 A x’ / / A / 0d / V I In 9' (9 I/ a 14 / 2 a x 31' /' <9 ‘ / ES ,9’ o ,/ 12 ’ 1 1 1 1 A 32 B 5’ o 24 ,‘V o ’/ b0 ’1 / O I o ,’ \. 57’ v / n ” / E3 [)7 s a 8 ' A E a an 0 1 1 1 1 L 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE PRESSURE -- EtOH (IF/Po) Figure 4. A: 001 Spacing 2; relative pressure of EtOH for Ou-mont. solid line, sym. first order peaks; broken line, asym. first order peaks. B. EtOH adsorption isotherm for Cu-mont.A, absolute adsorption corrected for 320 loss; v, apparent adsorption. 37 :ype clay-water interactions and/or Coulombic type forces festricting expansion between individual clay platelets. Vapor adsorption may be divided into three stages. Lnitially, adsorption is physical in nature where ethanol is replacing loosely held water on external clay surfaces and in the outer coordination sphere of the copper ion. Once the relative pressure reaches a critical point such that it can overcome the CU-Hgo binding energy of the inner hydration Sphere, there is an exchange of ethanol for this coordinated water . Evidence for this is the plateau in the apparent adsorption isotherm (broken line) between 0.25 and 0.5 relative pressure where the amount of water dispelled approaches that of alcohol adsorbed. Finally, as saturation pressure is approached, physical adsorption occurs by the filling of voids in interlamellar Spaces which are created by expansion between the clay platelets as well as capillary condensation in the micropores between individual clay particles. The coincidence and parallel nature of the appar— ent adsorption isotherm with reSpect to the absolute isotherm in the first and final stages of adsorption suggests that the major portion of water was replaced between 0.25 and 0.5 relative pressure. This is simply one way of expressing the experimental observation that the major reduction in the absorbance of the water deformation band (1630 cm'1 region) occurs over this relative pressure range. 38 A better insight into the restrictive forces imposed by the individual clay platelets can be obtained from a plot of lattice spacings versus relative pressure (figure 4A)- Initial adsorption involves Slight expansion from 12 to 15.25 3. The second phase of adsorption is characterized by a rather stable 001 basal Spacing where exchange of ethanol for water is occurring in the inner coordination sphere of the copper ion. The final stage of adsorption is characterized by a further weakening of electrostatic forces holding the clay platelet together with a resultant expansion from 13.25 to 16.5 A. The solid lines in figure 4A represent the relative pressure regions where symmetrical first order x-ray dif- fraction peaks were observed. These x-ray tracings approach rationality, but as with clay-H20 systems, strictly rational peaks are seldom observed. The 13.25 A basal spacing is in very close agreement with results of Brindley and Ray (1964) for a single layer of ethanol between adjacent Ca-montmoril- lonite platelets. Since this basal Spacing is characteristic of the pressure range where exchange of ethanol for water occurs in the inner coordination Sphere, it is reasonable to conclude that the apparent stability of the basal Spacing can be ascribed to a square planer Cu—ethanol complex. If the CC0 plane of the ethanol molecule lies parallel to the oxygen surface of the clay lattice, the maximum layer Spacing re- quired for the molecule is theoretically 13.66 A in the 39 —direction, using known bond distances and van der Waals adii (Brindley and Ray, 1964) . However, by proper geometric- l packing a Shortening of 0.4 A can be achieved at each rganic/silicate interface (Brindley and Hoffmann, 1962) . ince a total contact shortening of 0.41 A at both interfaces .8 all that is required to fit the x-ray data, it is obvious that a Cu-EtOH coordination complex, with the CC0 plane parallel to the lattice surface is geometrically feasible. A Cu—EtOH coordination complex will explain the type of xr-ray data which has been used by many investigators to pro- pose the existence of a monomolecular layer of material between clay platelets. However, this is an entirely dif— ferent approach than that used by Emerson (1957) and Brindley and Ray (1964) who proposed an alcohol-OH---O—clay type bond for alcohol adsorption on montmorillonite surfaces. It is, however, in agreement with the observation of these studies and thoselof Tensmeyer _e_t_al_. (1960) which Showed no change in the infrared C—H stretching vibrations upon adsorption. The lack of a Shift in these vibrations tend to disprove an earlier theory that C-H---O—clay type (interactions are im- portant in the adsorption of polar aliphatic molecules on clay surfaces (Bradley, 1945; MacEwan, 1948) . The symmetrical 001 diffraction peak for a 16.5 A Spac- ing is also in agreement with the 16.6 A Spacing reported by Brindley and Ray (1964) for a double layer ethanol-Ca-mont- morillonite complex. In light of the preceding discussion, the transition from a 13.25 to 16.5 A basal Spacing can be 40 thought of as a transition of the square planer Cu—EtOH complex to an octahedral complex. This might be visualized as a rotation of the square planer complex from a position parallel to the lattice surface to one essentially perpen- dicular to the surface, which is followed by the approach of an ethanol molecule on each side to form a distorted octahedral Cu—EtOH coordination complex. Hydrated Alamontmorillonite. Curves A, B and C, figure 5, show the dehydration of Al—montmorillonite as the relative pressure of ethanol in- creases, as indicated by the decreased absorbance of the 1655 cm"1 deformation vibration. The frequency maximum shifts to 1650 cm"1 upon ethylation of the clay film. In contrast to Cu-montmorillonite, this is a shift to a higher frequency, which is in agreement with the principle that H-bonding of water molecules increases the frequency of its bending vibrations (Hass and Sutherland, 1956). The water deformation band is also characterized by a much broader and diffuse contour than was observed in the case of residual water on Cu-films. An explanation of these apparent anomalies in the behavior of residual water probably lies with the dif- ferent type of coordination habit of the two species. All residual water molecules remaining on Cu-montmorillonite, having a square planer coordination complex, will be exposed to identical electrostatic forces, both those of the cation and the clay lattice. But, due to the octahedral coordination 41 .mQapan woo .mopmmmmmo .noppMHzpwm mOpm Hoppw 9509 . Hop othQmoSpo op ummogxm .@ “wopmmmmmo maps: oooo— op cmpmmn omnp .coppwhs upmm mOpm Hmpmm mnson o— commommo .m “nappohdpwm wepm umppw anon — commmw loo .9 ”mOpm ohdmmwnm mbppmHmH o.— op ammonwm .o “wepm ohfimmmnm opppwamh m_.o op ommomxm .m “coppeuppw .4 "oppqoaapnoquosnad no oppomom ooaopan ATE"; mmwm232w><3 .m mpzmpm 00mm CON. 00%. 00v. 00m. 00w. 005. 00m. 00m. OOON oovm 00mm OONn ‘ a p a . a 4 « < r r p b p » CON. cow. 00¢. con. oow_ Dob. 00m. 00m. OOON om¢~ comm omwn Gown NOISSIWSNVHL 42 habit of aluminum, the residual water molecules on an Al-film will be exposed to a varying clay lattice force field, because of their Spatial arrangement. The C-H and O-H bending and stretching modes of ethanol adsorbed on Alamontmorillonite (curves D and E, figure 5) show no differences from those of Cu-EtOH systems. The amount of ethanol retained against heating and evacuation was essen- tially the same as suggested by the absorbance of the CH3 deformation vibration (curve E, figures 5 and 5). However, considerable difference in the time required for rehydration to occur does exist between the two systems. As shown in curve F, figure 5, Al-montmorillonite has lost more ethanol during the first hour or rehydration than Cu-montmorillonite had during 24 hours. Two possible explanations are given that might in part explain this faster exchange of water for ethanol: 1) since the hydration energy of aluminum is more than twice that of copper, a faster H20-Et0H exchange would be anticipated; and 2) as mentioned previously, the differential force fields exerted on the various ligands of the Al-complexes would be expected to distort the octahedral complex and thus allow a more rapid exchange. The shoulder at ’V2500 cm-l, which occurs on all curves in figure 5, is assigned to a vibrational mode of water, since it is present in the pure clay-water system.. This band also occurs to a lesser extent in Cu- and Na—montmorillonite. Two possibilities arise for the increased intensity of this band 45 upon dehydration: 1) dehydration reduces the overlapping of other vibrational modes in this region, and 2) as dehydration occurs, the remaining water is more strongly H-bonded, which is accompanied by an enhanced absorbance. Farmer and Mortland (1966) have reported O—H stretching vibrations as low as 2780 cm‘1 for strongly H—bonded water in Mg—pyridinium complexes of montmorillonite. On the other hand, this band could be assigned to a combination mode of the rocking and bending vibrations,‘A/865 and 1640 cm—1, reSpectively, of water co— ordinated to metallic cations (Nakagawa and Skimanouchi, 1964). A van der Waals type II adsorption isotherm was observed for the adsorption of ethanol on Al-montmorillonite (figure 6B). The plateau between 0.2 and 0.6 relative pressure in both the absolute and apparent isotherms can be easily explained by a combination of two factors. First, a ethanol pressure suf- ficient to overcome the hydration forces of aluminum has been attained which then allows exchange of ethanol for water in the inner coordination sphere of aluminum. This is shown by the steeper slope of the absolute isotherm with respect to the slope of the apparent isotherm. Secondly, the trivalent nature of aluminum prevents the expansion of clay platelets beyond 15.5 2 (figure 6A) thus restricting the volume avail- able for ethanol. The final portion of the isotherms between 0.6 and 1.0 relative pressure must then be characterized by multilayer adsorption on external surfaces as well as capillary condensation. Proof of this is also given by the parallel 44 16 - A °3 U! 14 r— §§ __J. a. ,. r. ________ -<3 {H3 0' -—"’“ —U V U V ‘4 9-! m E; 12 — 1O 1 I I J J ETHANOL ABSORBED (g./100 g. clay) o l I L, L, J 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE PRESSURE -- EtOH (P/Po) Figure 6. A: 001 spacing vs. relative pressure of Eton for Al-mont. solid-Tine, sym. first order peaks; broken line, asym. first order peaks. B: EtOH adsorption isotherm for Ll-mont. [5, absolute adsorption corrected for 320 loss; ‘7, apparent adsorption. 45 nature of the two isotherms in this region, substantiating the experimental fact that little water is removed above 0.6 relative pressure of ethanol. Hydrated Ca-montmorillonite. The behavior of Ca-montmorillonite in the presence of ethanol is very similar to that of Al-montmorillonite. Such is expected since calcium prefers octahedral coordination as does-aluminum. A relative ethanol pressure of unity brings about a more complete dehydration of Ca—montmorillonite (curve C, figure 7) than was true with either the Cu- or Al- systems. This is explained by the lower hydration energy of calcium which allows easier replacement of coordinated water. This also explains the fact that rehydration of the Ca-clay occurs somewhat faster than it does in Al-systems. It will be noted that all traces of ethanol are gone (curve F, figure 7) after exposure to the atmOSphere for one hour. Like Cu- and Al-montmorillonite, Ca-clay does retain appreciable quanti- ties of ethanol upon heating and evacuation (curve E). The adsorption isotherms presented in figure 8B are of the same general type as those for the Al—system, with the exception that the plateau is much less pronounced and occurs as an inflection point near 0.5 relative pressure. There is also a greater total quantity of ethanol adsorbed because of the lower hydration energy of calcium accompanied by the fact that expansion of the clay lattice occurs more easily with di— valent calcium as the saturating cation. A plot of the 001 46 .woppoon can .prmmowoo .nopponspdm mopm Hoppo anon F you oaonomospo 0p oomomMo .m nwnpmmowou caps: oooo. op copoon nonp .qoppmns upon mOpm Hopes onson o. commemoc .m unopponspoo wepm hoped noon pained» now .9 «mopm oasmmonn obppoaon o._ op oomooxo .o ”mopm oasooonn obppoamn .h ohswpm e_.o op comoawo .m «copuunapo .4 «oppnoaapnospdosuoo Ho unpoomm nonoHHnH .728 mmumzazuz; 8m. 8m. 7 00¢. 000. com. 00: com. com. ooo~ oovw Doom ooun Gown an! comm u 9&. .52 89 m < » P r CON. OOn. 00¢. 00.». com. 00h. NOISSIWSNVHL 16 14 001 srlcxncs (R) 12< ETHANOL ADSORBED (g./1OO g. clay) 47 o I l l l L 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE PRESSURE -- EtOH (P/Po) Figure 8. A: 001 spacing 12. relative pressure of EtOH for Oa-mont. solid line, sym. first order peaks; broken line, asym. first order peaks. B: EtOH adsorption isotherm for Oa-mont. A , absolute adsorption corrected for 1120 loss; V’ apparent adsorption. 48 basal spacing versus relative pressure in figure 8A shows;a metastable inflection at 15.4 R and a more stable state at 16.5 3 between 0.5 and 0.8 relative pressure. Between 0.8 relative pressure and saturation, further lattice expansion is indicated by the asymmetry of the diffraction peak toward higher spacings. A sharp upturn in the adsorption isotherms in the same region is added evidence of further lattice expansion. Hydrated Na-montmorillonite. The only significant difference between the infrared data of Ca— and Na-montmorillonite is that dehydration of Na-films is not accompanied by a Shift in the water deforma- tion band to a higher frequency (figure 9). This suggests that the residual water molecules are not necessarily associ- ated with the sodium ion, a fact consistent with the low hydration energy of sodium. For the same reason, Na-clay retained less residual water at the saturation pressure of ethanol than clays saturated with polyvalent ions. The adsorption isotherms presented in figure 10B are similar to those for Ca-montmorillonite, but are smoother with no sharp inflection point. The overall slope of the Na-mont- morillonite isotherms is less than those for the Ca-system. A general restriction of lattice expansion will result in a decreased slope. As shown in figure 10A, expansion never ex- ceeds a 001 Spacing of 13.2 2. That Na—montmorillonite will not expand as much as Ca-montmorillonite at low vapor 49 .mSapmon and .wsammomou .uoapoHSpom mopm Hopes anon — you ononomoSpo op comomxo .m uwnpmmowoc caps: oooop op copoon nonp .ooppons upon mapm noppm mazes o. commowou .m “ooppoHSpom mopm nopmo anon — commom loo .9 ”mopm ohfimmono obppoaon 0.. op vooomuo .o .mOpm oasmmonn obppoaon m—.o op comomxo .m .eophuuapo .4 “oppnoaapaoSpdos|oz Ho oppoonm conoanH .m cadmfim .763 mmuminzw><3 OON . 00m. 00¢. 000. 000. 00... 000. 000. 000w 00¢“ 000m 00mm 000m 1 d d d d (‘v u r — 00N. 00m. 00v. 000. 000. 00... 000. 000. 000a 00¢“ 000m 00m» 80m . NOISSIHSNVUJ. 16r— 14r- 12'- 001 SPACINGS (X) 10- 24- ETHANOL insosssn (g./100 g. clay) 50 o l J l J l 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE PRESSURE -- EtOH (P/Po) Figure 10. A: 001 Spacing vs. relative pressure of EtOH for Na-mont. solid-Tine, sym. first order peaks; broken line, asym. first order peaks. B: EtOH adsorption isotherm for Ea-mont. A, absolute adsorption corrected for 1120 loss; V7: apparent adsorption. 51 pressures of water has been noted by many researchers includ- ing Mooney g£_§l. (1952) and Gillery (1959). Barshad (1952) noted that Na-montmorillonite dehydrated at 1700C would only expand to 13.4 3 when immersed in ethanol, while similarly treated Ca-clay will expand to 17.0 R. An explanation for the lack of expansion of Na-montmorillonite follows closely that offered by Brooks (1960) when studying the free energies of immersions of montmorillonite in water, ethanol, and n- heptane. Van der Waals forces between individual platelets are very important at distances of 5 R or less and overcome the hydration energy of sodium, thus restricting the expansion of Na—montmorillonite. Hydrated NH4-montmorillonite. Since infrared Spectroscopic data has shown that NH4-H20 interactions do occur on montmorillonite surfaces (Mortland et al., 1963), it was considered informative to study this sys— tem in the same way as the metallic cation systems. This inter- action can be seen by noting that the NH4+ band centered at 1450 cm"1 in curve A, figure 11, normally vibrates at 1427 cm-1 in the unperturbed state. Upon degassing of the ethylated film, the 1427 cm’1 band can be observed (curve F). The positioning of the NH4+ band at 1427 cm‘1 when the clay film is saturated with ethanol suggests that little EtOH-NH4 inter— action occurs. Further evidence that ethanol is physically adsorbed on NH4-montmorillonite is the linear plot presented in figure 12. In a linear form, this empirical equation 52 .wgpmmmmmu maps: oooop op mappoon can .mazon o. mqpmmomou .SOppmnSpmm mOpm Hoppm noon _ Hop ononmmoSpo op oomoowo .n “soapoHSpom mopm popes anon — oommmmoo .o .mOp oHSmmon o>ppoaoa 0.. op ummooxo .m uuopnolhpm .4 umppquHpHoSpgoSI m2 po oppooom commaan ._p madmpm .Tso.mcmmzazm><; 00m. 00m. 00¢. 000. 000. 005. 000. 000. 000m 00¢N 000a OONn 000m 4 a 4 q q q 4 4 4 4 a d ¢om~ sue. Ono. . _ comm— ?a. .52 _ Nmm. :KM NOISSIWSNVUL OON. 00m. 00¢. 53 14 — 12 — 11_ 10— V ETHANOL ADSORBED (g./100 3. Clay) 8 1 l l I 1 1.0 1.2 1.4 1.6 1.8 2.0 (log log Po/P + 2) Figure 12. Adsorbed EtOH as a function of (log log Po/P + 2) for NRA-montmorillonite. 54 developed by Bradley (Orchiston, 1953) for the adsorption of molecules with large permanent dipoles onto ionic adsorbents, takes the form: log log Po/P = log K2 + x log K; (1) where: P - pressure of adsorbate, P0 = saturation vapor pressure at the temperature of experiment, x = mass of ad- Sorbed gas, and K1 and K2 are constants depending upon the particular system. It will be noted that the plot is linear over the relative pressure range between 0.125 and 0.70. It should also be noted that none of the metallic cation-clay systems obeyed this relationship, which is indirect proof that forces other than cation-dipole interactions bring about the adsorption of ethanol vapor. Studying water vapor adsorption on soils, Orchiston (1953) observed that this equation was obeyed over the relative pressure range from 0.08 to 0.9. The other infrared data in figure 11 and the isotherms and x-ray data in figure 13 are essentially the same as that presented for Na-montmorillonite. Deuterated ethanol studies. In order to separate the 0-H vibrations of residual water from those of ethanol and take a closer look at these vibrations, films of Cu- and Al-montmorillonite were exposed to C2H50D followed by outgassing. The same procedures were followed for like films being exposed to CH3CD20H, in order that the vibrations of methyl and methylene groups might be 55 16'- r\ A ‘1‘; U) (5 14 — ...... ‘6 Q 2”"9 """" G" ’’’’’’ 4) <1 / Fe m .- // C) 12 0 d5 10 I I l I 4 ETHANOL ABSORBED (g./1OO g. clay) o l l l L J 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE PRESSURE -- EtOH (P/Po) Figure 13. A: 001 Spacing vs. relative pressure of EtOH for NH4-mont. solid-line, sym. first order peaks; broken line, asym. first order peaks. B: EtOH adsorption isotherm for NH -nont. A, absolute adsorption corrected for 320 loss; V9 apparent adsorption. 56 separated and studied more closely. The various vibrational modes of these two Species in the 3400-1200 cm‘1 region of the Spectrum are listed in table 2. Unfortunately the sample of CH3CD20H contained both water and normal ethanol in de- tectable quantities as indicated by the appropriate 0H and CH2 vibrations. Tracings A and B in figure 14 are those recorded after equilibration of Al- and Cu-montmorillonite films, reSpective- ly, with EtOD for two hours at a relative pressure of unity followed by evacuation for 30 minutes. By comparison with curve D in figures 3 and 5, it is immediately obvious that the absorbance of the 1650-1600 cm‘1 band of HOH is less. This suggests an OH-OD exchange between HOH and EtOD. This is further supported by the appearance of the EtO-H deformation band at 1265 cm‘l. The band at 1575 cm‘1 is thought to be that of water coordinated to copper ions as previously dis- cussed. The 3440 cm‘1 O-H stretching band is also thought to be that of coordinated water. In this respect, Gamo (1961) reported a strong 3420 cm'1 band for the 0-H stretching of water in CuSO4°5H20. There exists the possibility that the 3440 cm’1 band may be due to the 0—H stretch in Cu2(0H3)Cl as reported by Tarte (1958). However, two factors cast doubt upon the presence of such a Species: 1) the Cu—clay was originally washed free of chlorides as indicated by negative AgN03 tests and 2) a Species of this type would be expected to give a much sharper peak positioned at 3448 cm‘l. 57 Table 2. Infrared frequencies of liquid and adsorbed C2H50D and CH3CD20H . Infrared bands 'Model Adsorbed Liquid (observed) ———————————— cm ————————————- CgHsOD 2984 2984 asym. CH3 stretch 2935 asym. CH3 stretch 2915-2935 2890 sym. CH3 stretch 2460-2500 2460-2500 GB stretch 1480 1480 CH2 scissoring 1450 1450 asym. CH3 bend 1394 1386 sym. CH3 bend CH3QQZOH 3260-3360 3200-3500 CH stretch 2984 2984 asym. CH3 stretch 2935 2935 sym. CH3 stretch -- 2208 asym. CD2 stretch 2135 2105 sym. CD2 stretch 1448 1447 asym. CH3 bend 1384 1378 sym. CH3 bend 1315 1400 OH deformation 1These tentative assignments are based on comparable bands of normal ethanol as well as work of Hadzi and Jeramic (1957), Hadzi et al. (1962), and Margottin—Maclou (1963). 58 .oppaoaaanoapaoa-50 .n “oppaoaaanoapq°auaa .o "momomnmo .oppaoaaanoapaoanso .m “opaaopapuoepuoauad .4 “no m o .mopSSHS on commmwoo soap can Hononpo uopohopSoo po ohfimoonn opppoaon o.— op madmooxo noppo oppSoHHpnoSpnos no capoomm commade .e— oHSmHm .760. mmumzazuxfii 08. con. 8... con. 8.... oo: coo. com. 88 82.. 88 8mm 8% _ p a q _ q q q q d a . ¢om~ n.3— vmc _ 92 83 0¢0~ 00¢. nn.~ NOI SSI WSNVUI bl h r OON. 00m. 00¢. 000. 000. 00.... 000. 000. OOON 00¢N 000m 00mm 003 59 The broad O—D stretching band of adsorbed EtOD at 2500 cm‘1 does not differ significantly from that observed in the liquid phase. It should be noted that this band coincides with a band in pure Al- and Cu-clays which has been assigned to a highly perturbed 0-H stretching mode of coordinated water (see curve A in figures 3 and 5). The band at 2640 cm‘1 is assigned to the 0-D stretching of the clay lattice. This would yield a VOD/VOH ratio of 0.725 which is very close to that reported by Mortland gg_al. (1963). Since the 2720 cm"1 band persists upon exposure to the clay film to the atmosphere for seven days, it is also assigned to a lattice O-D vibration. These two bands are probably the deuterated counter-parts of the 3640 and 3700 cm-1 O-H lattice vibrations, respectively (Russell and Farmer, 1964). However, the intensities of these 0D bands are very slight compared to the 0H bands, suggesting that deuteration is not extensive and is probably confined to exposed surface groups. Curves C and D, figure 14, Show the Spectra of Al- and Cu-montmorillonite after adsorption of CH3CD20H following the same procedure as given above. From these spectra, con- firmation of the assignment of the 1480 cm‘1 band to the methylene scissoring vibration is obtained. Shifts in the position of the 0H in-plane deformation were the same as ob— served for EtOH (Av = 25 cm‘l), while no significant shifts were noted for either the CH3 bending or stretching vibration. 6O Acetaldehyde adsorption. / Heating Cu-montmorillonite in the presence of ethanol vapor resulted in the appearance of a band in the 1750 cm-1 region of the Spectrum. Comparison of band position and contour with that of acetaldehyde (Depireux, 1957) suggested that ethanol was oxidized to acetaldehyde. Further investi- gation Showed that this reaction was catalyzed by the brass cell in the absence of clay. Degassing the system showed that the acetaldehyde was present in the vapor phase and was not adsorbed on the ethylated clay. However, exposure of an air—dried Cu-film to acetaldehyde vapor for two hours followed by two hours of evacuation, showed that adsorption did occur (curve B, figure 15). Tracing A is the spectrum for a simié larly treated Cu-EtOH clay film and is included for comparison. A shift of the C-0 stretching to lower frequencies suggests that acetaldehyde is coordinated to the copper ion through the carbonyl group similar to the Cu-amide complexes observed by Tahoun and Mortland (1966). Carbonyl bands at 1675 and 1715 cm‘1 suggests two types of adsorbed acetaldehyde: the first band that of acetaldehyde coordinated directly to the copper ion, and the second that of acetaldehyde present in an outer coordination sphere or coordinated through a HOH bridge Similar to that proposed by Farmer and Mortland (1966) for Cu-Hgo-pyridine complexes. The disappearance of the 1715 cm“1 band upon heating and its reoccurrence upon exposure to the atmosphere is evidence to support a Cu-H20 acetaldehyde type 61 I700 I600 I500 I400 T I I l I575 TRANSMISSION {I I355 U l ms I I I675 I700 I600 I500 I400 WAVENUMBERS (cm") Figure 15. Infrared Spectra of Cu-montmorillonite: A, degassed 1 hour after eXposure to 4.8 cm. EtOH; B, degassed 2 hours after eXposure to 4.3 cm. CHBCHO. 62 complex. As was true with ethanol, no changes were observed in the CH bands upon adsorption. General considerations. The infrared data presented have Shown that the bulk Of the water retained on air—dried homOionic clays is re- placed by equilibrating the clay with ethanol vapor at its saturation pressure. By using the absorbance of the 1650- 1600 cm'1 deformation band of water and the procedures out- lined previously, it is possible to follow the dehydration of the clay films with increased pressure of ethanol. The data given in figure 16 are a result of these calculations and Show the water loss as a function of the relative pressure of ethanol. First, it is obvious that the water content in all of the systems has been reduced to less than 0.7% (based on 3000C). Na-clay, because of its weak hydration energy, has lost essentially all of the adsorbed water. It will be noted that the Cu— and Al-montmorillonite curves break near 0.5 relative pressure, while Ca-, NH4- and Na-systems break at somewhat lower values, 0.4 to 0.3 rela— tive pressure. This can also be explained by energies of hydration. By this argument, the aluminum should break at a somewhat higher value than copper Since the hydration energy of the former is 609 kcal.mole‘1 greater (Hunt, 1963). However, this apparent anomalous behavior can be, in part, explained by closer consideration of the two cations and their properties. As mentioned previously, due to its square 63 5 - 5— :5 84- 5. u) 3 0‘3_ \ g A \ g: \ \ \ \ \ \ \ s s ‘\ NH); V‘-~.‘:. \QV ------------ V ......... \-. «0_---‘-<>‘_ -‘ --__ - 0 1 l l l 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE PRESSURE -- EtOR (P/Po) Figure 16. Variation in Ego content as a function of the relative pressure of EtOH for homoionic montmorillonite. 64 planer coordination habit, copper ligands will be subjected to the same intensity level of the force field in the inter- lamellar space, which will tend to stabilize the complex. 0n the other hand, because aluminum prefers octahedral co- ordination, this will place different ligands at different intensity levels in the force field and in turn cause dis- tortion of the complex. Also, trivalent aluminum will bind the clay platelets together more energetically than will di- valent copper which causes distortion of the coordination complex and partially negates its hydration or coordination energy. The break in the calcium curve occurs at a point intermediate between those of Cu— and Al- and those of NH4- and Na-clays, which is in accord with its intermediate hydration energy. In the preceding discussions, the simultaneous occur- rence of both water and ethanol in coordination complexes has been indicated. Such is undoubtedly true. However, it is very unlikely that both water and ethanol molecules Simul— taneously serve as ligands about the same cation Since this is a thermodynamically less stable configuration. Therefore, even though statistically Speaking, the number of water mole— cules per copper ion is less than 0.7 at a relative pressure of unity, the water present is distributed as four molecules per cation. Along this same line of argument, the asymmetry of the first order (001) x-ray diffraction peak and the ir- rationality of higher orders at any particular relative pres- sure, can be explained by the random interstratification of 65 layers containing only water or ethanol, a corollary of ideas suggested by Mortland and Barake (1966) in their metal ion- organic ion work. The plot of ethanol retained versus length of time of evacuation in figure 17 shows the relative stability of the cation-ethanol systems. All curves fall rapidly at first with the removal of loosely held molecules and then level out after about five hours of evacuation. The contour of the copper curve indicates that a somewhat longer time is required to reach a state of apparent equilibrium. It is noted that poly- valent cationic systems retain appreciably more ethanol than monovalent cationic systems. Cu- and Al-clays retain 4.5 and 6.8 molecules per ion, respectively, after ten hours of evacuation, which is very close to their preferred coordi- nation numbers. Ca-montmorillonite also retains approximately 4.5 molecules per cation, while Na- and NH4-Systems, which do not Show strong inner Sphere coordinating behavior, retain less than 1.0 molecule per cation. The curves in figure 18 Show the influence of saturat- ing cations upon the behavior of clay systems. These curves Show the desorption of ethanol by replacement with atmOSpheric water after evacuation and heating. The same reasoning pre- sented earlier in this section for the desorption of water upon ethylation, applies here for the differential behavior of the various cationic—clay systems. As shown, Na-montmoril- lonite has very little ethanol present at the start of rehydration, which is Slowly replaced. This Slow replacement 66 ’T 16, m |\° 8 I. i I k a Cu 3 (W >4 12. (ER-B»--fl-_ ‘ A1 a | \a “We- --------- _. 0 ' ‘\“‘~A.. h g a. “““~~- -- - a 8., a .V. HI 9 “~V‘ \‘ “R"V‘»..- NH 4 — ‘5\ ““““ ‘7 ........... 4 ---_.V_------------------V \Q\ “-\‘0\--~‘-- Na 0 , , —~-, ------ 42 ------- 4 0 2 4 6 8 10 TIME (hrs.) Figure 17. Adsorbed EtOH as a function of the evacuation time for homoionic montmorillonite. 67 ’5 .3 .p d a I g I I: "‘ 9 0.08 “II n |I A1 a II“ 3 0.06 JEI Ga 2 II g I m I a 0.04 -I'. I III I 0.02 44' ‘P.-.e~---4--\--- Na '| <>~---~---~---—-- ~Q 0.00 4; l I l I I I 0 4 8 12 16 20 24 TIME (hrs.) Figure 18. Absorbance of the CH deformation band of EtOH as a function of rehydration time for ethylated, homoionic montmorillonite. 68 can be explained by the fact that the 001 basal spacing is only 11.8 R.compared to Spacings of greater than 13.3 R for Cu-, Al-, and Ca-clays. Hence, the tortuosity factor is much greater for a molecule escaping the interlamellar Space of a Na-clay. With the additional information that all of the ethanol was removed from Cu-montmorillonite in approxi- mately 70 hours, it can be observed that the adsorption- desorption of ethanol from these homoionic montmorillonite surfaces is a reversible process. Since the rate of loss of ethanol from Cu-montmorillonite was slow enough to be measured by the techniques and design of these studies, the data were plotted according to the equation: log (1 - A/Ao) -0.16 - 2.515 £2 t (2) where: t = time, A0 = absorbance of the CH3 deformation band at t = 0, A = absorbance of CH3 deformation vibration at any given time, D = diffusion coefficient, and r0 = radius of montmorillonite particles assuming disk Shaped platelets. This iS a derivation used by Fripiat and Helsen (1966) for the diffusion of "free" ammonia between clay Sheets. They -19 cm2 sec'1 at 550C noted a diffusion coefficient of 6.6 x 10 for the diffusion of NH3, which resulted from the decomposition of CO(NH3)63+-montmorillonite complexes. It can be seen by plotting of log (1 - A/Ao) against time (figure 19) that the data fit reasonably well a straight line between 4 and 56 hours. Using a value of 120 R for r0 (Eeckman and Loudelout, .oppdoaaphofipdofiuso pop capp no noppqup a me Ao<\< I F. woe .m— oaswph ..man. nape ow om 3 on om o. o 69 w . _ _ E _ _.o 70 1961), the diffusion coefficient, calculated from the Slope' of the line, is 1.71 x 10‘18 cm2 sec”1 at 200C. Hence, after the rapid initial loss, the desorption of ethanol on Curmontmorillonite follows a stationary diffusion process until the exchange with water is essentially complete. Glycol adsorption - desorption studies _Infrared absorption frequencies for both liquid and adsorbed glycol are given in table 3 for the 3400-1200 cm-1 region of the Spectrum. When compared with ethanol, very little work has been done with the vibrational assignments of ethylene glycol. Although no assignments are made, Kanbayashi and Nukada (1963) have published an identical Spectrum of glycol, in the 1500 to 600 cm'1 region, to that observed in the present studies. Using polarized infrared techniques, Miyazawa 2E.il° (1962) have concluded that the CH bands in the 1400-1200 cm"1 region are combinations of wagging and twisting vibrations and are not simply one or the other. AS with ethanol, the 0H deformation of glycol appears to inter- act with the CHlbending bands in the liquid state, giving rise to a doublet with maxima at 1410 and 1330 cm‘1 (Krimm §£_al. 1956). Cu-montmorillonite. Infrared Spectra of glycol—water-Cu-montmorillonite complexes are presented in figure 20. »It should be noted again that the complex was formed initially by suSpending 71 Table 3. Infrared frequencies of liquid and adsorbed ethylene glycol. Infrared bands Mode 1 Adsorbed Liquid ———————————— cm?l--__—————_--— . _ 5500 3360 CH stretch 2960 2950 asym. CH2 stretch 2890 2884 sym. CH2 stretch 1463 1457 CH2 scissoring 1570 1570 CH2; 70% wag, 25% twist -- 1253 CH3; 40% wag, 60% twist -- 1205 CH2; 25% wag, 70% twist 1400 1410 doublet combination: 1550 1550 (ca + 0H bend.)2 lFor assignment of the absorption frequencies see: Davison (1955), Kuhn et al. (1959), and Miyazawa (1962). 2After Krimm et al. (1956). 72 .mpg mop pop mumsomoapo onp op whamoaxo Hoppo dem .0 ..mpg a pop opmnomoSpo onp op ohfimooxo noppo meow .m “.mas em Hop 0 mp. po Hoomaw moppaomoo amppm .4 umppooaapaoSpqosISouaoome mo oamoomm uoaoapaH .om madwpm .15. mmumzozmiz OON. 002 00.: 009 002 002 000. oom. ooo~ ooem 00mm 00mm 003 u d q q «I IdII II IIIJIIII .. a I II . II 4-- 4 q Ifi 00mm Onbw _ U 88 r > an. . _ m men. 33 ovon man. 4 o m < < < t _ . _ p F _ n . . p . 08. 002 003 009 com. 002 ooo. oom. ooo~ 005 00mm 009... 00% NOISSIWSNVBJ. 73 the clay film over a free glycol surface at 1150C for 24 hours in a evacuated disiccator. Tracing A was recorded im- mediately upon exposure of the cooled clay to the atmOSphere. Therefore, the bands arising from various glycol vibrations are essentially those of liquid glycol, since this is the predominant form at the time of exposure to the atmosphere. However, it will be noted that the Cu-clay is completely dehydrated as indicated by the absence of the 1632 cm‘1 de- formation band. After four hours exposure to the atmosphere, most of the glycol in the micropores and on the external sur- faces of the clay has evaporated and a trace of water has been adsorbed (curve B). Further exposure to the atmosphere results in the replacement of glycol by water accompanied by a strong 1632 cm‘1 band. —Probably the most significant observation with regard to the glycol-Cu-clay complex is the development of new bands at 2750 and 2650 cm'l. These bands are thought to arise from the glycolic 0-H stretching vibration of molecules coordi- nated directly to the copper ion. Since two bands appear, it is likely that coordination occurs through one of the OH groups, while the other group is involved in either intra- or intermolecular H-bonding. This is the type of metal-glycol coordination complex suggested by Miyake (1959) for cobalt (II) and nickel (II) when the chloride salts are added to liquid glycol. A slight shift of 20 cm"1 to a lower frequency after four hours exposure to the atmosphere (curve B, figure 20) is expected, due to stronger bonding as a result of loss \ 74 of the excess glycol. Since these bands do not occur to any great extent in other cation-glycol or ethanol-clay systems, they are not considered to arise from perturbed C—H stretching vibrations as alluded to by Tettenhorst gg_al. (1962). There is complete agreement with Tettenhorst g£_gl, (1962) that no significant band Shifts were observed for any of the C-H vibrations of adsorbed glycol. A slight shift of 10 and 6 cm-1 toward a higher frequency was indicated for tn the CH2 asymmetrical and symmetrical stretching vibrations, respectively. If such a shift does indeed occur, it would: ;~ a) fit well into the argument for coordination through the a oxygen atoms, and b) exclude the hypothesis of C-H--o0-clay interactions. Also, a slight shift does occur in the 1330 cm-1 band of liquid glycol to 1350 cm-1 in the adsorbed state. Although this band is predominantly O-H in character, discre- tion must be exercised in making interpretations, Since some CH interaction is present. The 1575 cm“1 band (curve C, figure 20) is thought to arise from perturbed water, while the broad band in the 1454 cm-1 region is thought to arise from a carbonate vibra- tion. Both of these bands will be discussed later. The time course of glycol desorption from Cu-montmoril- lonite is given in figure 21. It Should be pointed out that the solid line represents an estimate of the absolute amount of glycol present while the broken line represents the sum of glycol plus water on the clay at any given time. If it ADSORBED GLYOOL (g./IOO g. clay) Figure 21. 75 l J I I 80 160 240 320 400 480 L I TIME (hrs.) Adsorbed glycol as a function of time of atmOSpheric eXposure for Ou-montmorillonite: G 9 SIyCOI; A, glyc01 4' H200 76 is assumed that 25 g. glycol per 100 g. montmorillonite is required for monomolecular coverage of the total surface (Dyal and Hendricks, 1950), then all the glycol contributing to multilayer and capillary adsorption has been dissipated in approximately four hours at which time water has begun to enter the complex. The desorption curve (solid line) is then characterized by a rather rapid loss up to 160 hours of exposure to the atmosphere at which time the rate of loss becomes a constant until all of the glycol has been replaced by water 00’800 hours). It will be noted also that the ap— parent desorption curve (broken line) levels off, suggesting that at this time glycol lost is equal to water gained. Al-montmorillonite. The infrared data for Al-montmorillonite (figure 22) does not differ greatly from that for Cu-montmorillonite. The principle difference that occurs is the greatly reduced absorbance in the 2750-2650 cm"1 region of the Spectrum. Although a distinct peak does occur at 2660 cm-1, the reduced intensity suggests much less perturbation of directly co- ordinated glycol. This difference might be explained by the different coordinating habits of the two cations. AS ex— plained in the section on ethanol adsorption, the Spatial arrangement of ligands and the restrictive nature of the clay platelets impose greater distortion upon the octahedral com- plexes of aluminum than it does on the square planer complexes Of copper. Further indirect proof of this argument arises 77 .mas 0mm pop oaommmospo oSp op oHSmoaNo Hoppm mean .0 “.mas a hop oposmmo8po onp op oHSmomNo poppo meow .m “.man em pop oom—P pm Hoohaw mapppomeo noppo .4 noppgoaapaoSpaoSIaduaoohHm mo oapoomm omamapoH .mm oHSMpm .768 m¢u0232u><3 8a. 89 8... 89 8m. 8: 82 o8. 88 SS 83 8% 8% q u q d u a q u q 4 4 a saw . m .52 no! one. 25. I. U o m S m m S _ m m 0 N < < < RE. 2?. 9%. 25. 2?. 9r. Imw_ 9w. R&~ 2?~ RWN an 2?» 78 from the initial speed of hydration. As can be noted in curve B, figure 22, more water has been adsorbed by the Al- clay after four hours of exposure to the atmOSphere, as indicated by the 1635 cm‘1 deformation vibration of water, than was adsorbed by the Cu-clay in the same length of time. Desorption curves presented in figure 23 are very simi- lar in character to those for Cu-montmorillonite. The curves level off after approximately 160 hours exposure to the atmos- phere as did those for Cu-clay. Whether a true glycol-water equilibrium has been attained or not, cannot be stated with certainty, Since the infrared vibrations for glycol have been masked after exposure to the atmOSphere for 280 hours (tracing C, figure 22). Ca-montmorillonite. The infrared Spectra for glycol solvated Ca—montmoril- lonite presented in figure 24 are essentially the same as those presented for Al—clay. Slight adsorption of water has occurred by four hours of exposure to the atmOSphere (curve B). The main difference between the Ca- and Al-complexes is the position of the predominantly O-H deformation band of glycol. In Ca-montmorillonite the band is positioned at 1335 cm'l, which is very close to that of the pure liquid, and is 15 cm'1 lower than in the Cu- and Al-complexes (compare curve B in figures 20, 22 and 24). This suggests that ion-dipole interactions are not as great in the Ca-glycol complex as they are in the Cu- and Al-clay systems, which is in complete 79 ADSORBED GLYOOL (g./IOO g. clay) //—7 o 1 +1 I I //l l L 0 80 160 240 320 4(7 560 640 TIME (hrs.) Figure 23. Adsorbed glycol as a function of time of atmoSpheric exposure for Al-montmorillonite: (a, g1ycol;A, glycol + I120. 8C) .mps m.m Hop oposamoSpm obp op madmoqxo Hoppm mean .0 «.mas a Mom whosomoSpo map op anamoaxo Hopes osmm .m ..mas em Hop oom.. pm Hoomaw mapnpomuo Hopes .4 uoppgoaapaoSpaoSImouaooham mo appoomm wouoppsH .em oasmpm :I So. mmwmtazm><3 com. own. ooc_ own. coo. ooh. coo. coo. ooou coca ooom comm coon q q _ p 4 q d not. one. onn. NOISSIWSNVUL L h L I OON. COM. 00¢. 000. 000. COB. 000. 000. OOON OO¢N OOON Down Down 81 agreement with the lower solvation energy of calcium. The desorption of glycol from Ca—montmorillonite as a function of rehydration time is presented in figure 25. The curves are of the same general character as those ob- served for Cu- and Alflmontmorillonite, but replacement of glycol by water is achieved at a somewhat slower rate in the Ca-clay. Consequently, the break in the curves occurs after longer exposure to the atmosphere (”’280 hours). It is also of interest to note that the break in the desorption curves occur at a glycol level between 4 and 5 g. per 100 g. clay for the three homoionic montmorillonite systems studied. General considerations. The x-ray data for the three homoionic glycol mont- morillonite complexes are presented in figure 26. It should be noted that these data are for desorption-rehydration studies. The Shaded symbols represent rational diffraction peaks for which an integral series of Bragg reflections were obtained, while the open symbols are irrational and represent random interstratified and transitional states. In the glycol solvated state, the diffraction peaks are very rational with basal spacings ranging from 16.6 R for Ca-montmorillonite to 17.1 R for the most highly solvated Al-clay. Since glycol solvation has been used quite exten- sively in clay mineral identification studies employing x-ray techniques, numerous characteristic basal Spacings for monté morillonite have been published covering the range observed 82 ABSORBED GLYGOL (g./100 g. clay) 0 I I I l I I - 0 80 160 240 320 400 480 TIME (hr8.) Figure 25. Adsorbed glycol as a function of time of atmOSpheric eXposure for (Ea-montmorillonite: Q, glycol; A, glycol + H20. 83 .mhoao obppoommoa onp do mpmpHo hcao .m oop\aoohaw .w mm anon: pnpon onp mpaouonmon Ab. .qoapwoppppdnpmnopqp acumen Ho codename opoopc Imp mHonShm mono and .mSopmhm coppppoapmnopdpuaos Honoppon paomonqon oHopSho vacuum .oppaoaapaoapaos opaopoSon Hop nopczuaoohaw Ho oppmn macs osp Ho noppoosm o no mqpocmu Aaoo. Ho noppopao> wnpzonu mo>nso mfiadtunoownw ho 0H9: ”no: 0.N— 0.0— 0.m 0.0 0.: 0.N . om 0.5pr 0.0 . p . . . _ (3) ssnrovas too 84 in this study. Such variations can be easily explained by the different types of coordination complexes in the inter- lamellar spaces. Rearrangement in the complexes with vary— ing degrees of packing around the cation can explain the decrease in basal Spacings of octahedral Al- and Ca—complexes while still maintaining rational diffraction peaks. However, in spite of these differences, it was noted that all Systems begin to Show random interstratification once the glycol content drOpS below 25 g. per 100 g. clay which is the value calculated by Dyal and Hendricks (1950) for monomolecular coverage of the total montmorillonite surface. This break occurs when the mole ratio (glycol:water) has decreased to approximately six for Cu- and Al-montmorillonite and 2.3 for the Ca-clay. (Kunze (1955) also noted that some glycol- montmorillonite complexes were not stable in the atmosphere, while Mackenzie (1948) noted that Ca-montmorillonite would maintain a constant 17.1 R basal spacing over a wide range of water contents as long as six water molecules were ad- sorbed for each glycol displaced. With the advance of rehydration, the basal spacings of all three homoionic montmorillonites decrease and approach their respective air—dried spacings as glycol is replaced. The values given on the ordinate of figure 26 represent the Ifiisal Spacings for clay films equilibrated at 200C and 40% IKilative humidity. It is interesting to note that the plot Of (001) Spacings versus glycol content for Al-montmorillonite 85 passes through a minimum at 13.9 3, which has rational dif- fraction peaks and agrees with a stable 13.9 3 "one layer" glycol-Mg-montmorillonite complex observed by Hoffmann and Brindley (1961a). This 13.9 R Spacing occurs at a mole ratio (glycol:water) of 0.57 and is characterized by 7.2 molecules of glycol per aluminum ion, which is a value somewhat greater than the preferred coordination number of six for aluminum. Hence it is suggested that the random interstratification occurring in Al-montmorildonite at mole ratios greater than 0.6 is characterized by a vaporization of glycol and reorien— tation of coordination complexes. This reorientation might be thought of as a repositioning of the CC0 plane from a parallel position to one that is more nearly perpendicular between the clay platelets. Such orientation would allow a reduction in the basal spacing leaving "glycol columns" between the clay platelets and allow water to fill in around these columns. Below a mole ratio of 0.6, the glycol com- plexes begin to break up and the randomly interstratified systems might well be characterized by complete layers of water as well as 13.9 R "glycol layers" (Mortland and Barake, 1966). The glycol-Ca-montmorillonite system might behave in a Similar manner, except that the Coulombic type attractiOnS. ‘between the cation and the clay platelets are not great enough to result in a minimum. Similarly, the Cu-glycol system is initially character- ized by a transition from octahedral to square planer 86 coordination, accompanied by a loss of glycol and a reduc- tion in the basal spacing. This transition would result in a randomly interstratified system containing complete layers of either octahedrally or square planer coordinated glycol. Before this transition is complete, water begins to replace the glycol ligands which results in further interstratifi- cation. Once the mole ratio (glycol:water) reaches 0.5, the basal spacing of Cu—montmorillonite is characterized by the hydrated state. Turning attention to the process of desorption of glycol from montmorillonite surfaces (see figures 21, 23 and 25), it will be noted that the loss of glycol can be divided into three steps. First, glycol that is contributing to capillary condensation and multilayer adsorption is rapidly lost by vaporization within the first four hours of exposure to the atmosphere. The next step is a curvilinear loss of glycol with respect to time which extends over the time in- terval from 4 to 160-200 hours. Finally, the remaining gly- col is Slowly lost at a constant rate with the possible ex- ception of the Al-glycol system where an apparent equilibrium is reached. Hence, the adsorption-desorption of glycol is a reversible process in the systems studied. The intermediate phase of glycol loss appears to obey the conditions of second order chemical kinetics (figure 27): ’d IC2H302] dt = k [C2H602] [H20] (5) where: k = rate constant. Over the time interval where 87 .oppqonapnospnos upa0poSon Hop enamoauo oppodnaofipp .pm unseat Mo capp Ho qoppoo=p a no Hoomaw connomuo no noppoapqooqoo Hafizapoom ..mns. nzHa ow. om. opp omp oo— om _ ow oe om o _ 5 I4 _ _ . . _ _ 00.0 m0.o opno mp.0 0N.o (£310 '2 001/100118 '3) D/I mm.o 88 equation (3) is obeyed, the values for the rate constant (k) are: 13.6, 9.7, and 4.0 x 10"4 (100 g. clay) g'l hrs‘l for Cu-, Al-, and Ca-montmorillonite systems, respectively. It is to be noted that the kinetic order of a reaction is not necessarily related to the form of the stoichiometric equa- tion for the reaction. However, it appears that neither glycol nor water are in excess over the time interval being considered. Therefore, it would intuitively appear that this I Second order reaction was bimolecular in nature as indicated by equation (3). ‘%—~—— (The 1575 cm'1 band The development of a band at 1575 cm"1 has been noted when Cu- and Al-montmorillonite-glycol systems are exposed to the atmosphere (tracing C, figures 20 and 22). It was suggested that this band belonged to the deformation vibra— tion of a particular Species of water. It was then observed that this band occurred in clay films of Cu- and Al-montmoril- lonite treated in the same manner as was used in the glycol adsorption studies, with the exception that glycol was not present in the system. Tracing B, figure 28, Shows the de- velopment of this band after heating to 1150C for 24 hours and followed by exposure to the atmOSphere for 44 days. This 1575 cm‘1 band was not present in a freshly prepared Cu- film (curve A, figure 28) and was only slightly developed im- mediately following the heat treatment. 89 1800 I700 IGOO I500 I400 T I I I— r 2 (_D: (D (I) :5 (O 2: 0 <1 0: |._. 1440 I575 I632 I I I I I I800 I700 l600 I500 I400 WAVENUMBERS (cm") Figure 28. Infrared Spectra of Cu-montmorillonite: A, air-dried; B, 44 days after heating at 11500 for 24 hrs.; 0, after exposure to D20 vapor for 3 hours on 65°C hotplate; D, same as 0 after SXposure to atmOSphere for 7 days. 90 A deuteration study was made to provide evidence that the 1575 cm"1 band was indeed arising from water. Cu- and Al—films that had been heated and exposed to the atmosphere for 44 days were suspended over a D20 source for three hours on a 65°C hotplate at atmospheric pressure. The results of this deuteration for Cu-montmorillonite are presented in tracing C, figure 28. Tracing D is for the same film after I exposure to the atmosphere for seven days following deuter- 1 ation. ,This reversible deuteration of the 1575 cm”1 absorbing Species is taken as strong evidence supporting its assignment to the deformation mode of water. The Al-montmorillonite as 54 well as glycol analogs behaved in the same way. It is also of interest to note that the interlamellar H20 absorbing at 1632 cm'1 was not replaced by D20 under the conditions of this experiment, which suggests that the H20 absorbing at 1575 cm'1 is associated with the external clay surfaces. The 1480-1400 cm-1 band It was noted that a broad band in the 1480-1400 cm-1 region of the spectrum emerged as glycol was lost from Cu-, Al-, and Ca-montmorillonite (tracing C, figures 20, 22, and 24). Like the 1575 cm“1 band, this band also developed in the absence of the glycol treatment (tracing B, figure 28). Since carbonates are known to have very strong absorption bands in this region of the Spectrum (Lyon, 1964), it is sug- gested that the 1480-1400 cm‘1 absorbing band arises from a carbonate Species. Qualitative justification for a carbonate 91 assignment is.the Similarity in position and contour of the observed bands with those of known carbonates (Huang and Kerr, 1960; Hunt et al., 1950; Miller and Wilkins, 1952; Ross and Goldsmith, 1964). These similarities and the lack of any other known absorbing Species in these systems are the bases for a carbonate assignment. The particular carbonate Species that may spontaneously occur is rather uncertain and somewhat Speculative, and may vary from one system to another. Of course, the atmOSphere is a ready source of carbon dioxide. It is known that clay minerals are subject to degradation at low pH values which will release magnesium from the structural lattice. It is also known experimentally that the residual water on an air— dried or drier clay is very dissociable when compared to the dissociation of pure water (Fripiat and Helsen, 1966; Mort- land et al., 1963). Therefore, it appears reasonable that hydrolysis could occur with the protons then satisfying the negative charge of the clay lattice and releasing the saturat— ing cations to form carbonates and basic salts such as CuCO3- Cu(OH)2. Degradation of the clay lattice would release magnesium for the formation of MgC03. In the Cahmontmoril— lonite systems, there is the possibility of CaCO3 and CaMg(C03hgformation. Hence, it appears that carbonates could slowly form with time and that the particular Species may vary from system to system. However, the available infor- mation precludes a more definitive statement. 92 Since the 1575 and 1450-1400 or1 absorption bands develop with time, it might be suggested that both bands are associated with a hydrated carbonate complex. An appar— ent relationship can be observed in figure 29 where the absorbance of the 1575 cm‘1 band is plotted against the ab- sorbance of the 1480-1400 cm-1 band for Cu- and Al-montmoril- 1 water lonite. The only explanation as to why a 1575 cm- band does not arise in the Ca-system is that a different type of carbonate complex exists in the Ca-clay. Irrespective of possible band assignments, the above observations tend to suggest the difficulty in maintaining clean clay surfaces and support the ideas of Martin (1962). It further suggests that storage of clays in an air-dried or drier state may alter the surfaces to such an extent that resulting data may not be representative of the true surface reactivity. 0.07 0.06 0.05 0.04 0.03 ABSORBANOE LT 1575 cm-1 0.02 0.01 0.00 93 I l I L J I 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Figure 29. ABSORBAFOE LT 1480-1400 cm-1 Absorbance of 1575 cm'1 320 deformation band as a function of the absorbance of the 1480-1400 cm"1 carbonate absorption band for Ou-montmorillonite (G) and LII-montmorillonite (A). GENERAL DISCUSSION AND SUMMARY Infrared data obtained from the ethanol studies Showed that adsorbed water is replaced by ethanol on montmorillonite surfaces. Ethanol at relative pressures of less than 0.5 removed the major portion of the water retained by air-dried I clay films, while a relative pressure of unity reduced the 1 water content to less than 0.7%. In the case of Na—montmoril- lonite, essentially all of the water was removed. As dehydra— _° tion occurred, the 1632 cm'1 deformation band of the water on II -1 Cu—montmorillonite Split into a shoulder at 1640 cm and a maximum at 1598 cm‘l. This is evidence of the two phase nature of adsorbed water as suggested by Russell and Farmer (1964). The 1598 cm‘1 absorbing species is the more strongly adsorbed and is thought to be directly coordinated to the copper ion. On the other hand, the deformation band of water was displaced to a higher frequency as Al- and Ca-montmoril- lonite were dehydrated. Such shifts are in agreement with the principle that H-bonding and perturbation increases the frequency of bending modes (Hass and Sutherland, 1956). These apparent anomalies are reconciled by a consideration of the different types of coordination habits of the cations. Octahedrally coordinated complexes of aluminum and calcium are subjected to greater distortion in the interlamellar 94 95 spaces than the square planer complexes of Cu-montmorillonite. In contrast to the polyvalent cation systems, the position of the water deformation band was unchanged in Na- and NH4-mont- morillonites, which suggested that forces other than ion— dipole interactions are predominant during dehydration-ethanol adsorption in these systems. This reduced ion-dipole inter- action is easily explained when it is remembered that the sol- vation energies are much less for sodium and ammonium. The adsorption isotherms for ethanol on the different homoionic montmorillonites further substantiate the interpre- tations of the infrared data. X-ray diffraction results for Cu- and Ca-EtOH complexes Showed two stable regions at basal Spacings of 13.3 and 16.5 A. This is in agreement with the observations of Brindley and Ray (1964) for Ca-montmorillonite. Al—, Na-, and NH4-complexes never expand beyond 13.3 A. The trivalent nature of aluminum restricted further expansion between the clay platelets, while the low solvation energies of sodium and ammonium were not great enough to overcome the ion-clay and van der Waals binding forces holding the clay platelets together. Evacuation against a liquid nitrogen cold trap did not remove all of the adsorbed ethanol. Apparent equilibrium was reached within five hours. Cu- and Al-montmorillonite retained 4.5 and 6.8 molecules per ion, respectively, which is very close to their preferred coordination number, while Na- and NH4-clays, which do not exhibit strong coordination, 96 retained less than one molecule per cation. Rehydration accompanied.by ethanol loss occurs very rapidly except in the Cu-montmorillonite system where approximately 70 hours were required. In the case of Cu-montmorillonite, the loss of ethanol followed a diffusion controlled process, of the type noted by Fripiat and Helsen (1966) for diffusion of "free" ammonia between clay sheets. I Deuterated ethanol studies confirmed that water does 7‘ exist after equilibration of clay films with EtOH at a rela- tive pressure of unity through OD-OH exchange between EtOD “I and HOH. Two lines of proof were noted: 1) the decreased b4 absorbance of the 1630 cm-1 H20 deformation band on the clay equilibrated with EtOD, and 2) the appearance of a EtO-H deformation band at 1265 cm’1 on the deuterated clay. All the water on air-dried Cu-, Al-, and Ca-montmoril- lonite was replaced by glycol when the clay films were equil- ibrated with glycol at 1150C for 24 hours. Infrared Spectra of glycol-Cu-montmorillonite complexes showed two new bands at 2750 and 2650 cm-1. These bands were assigned to the O-H stretching modes of glycol coordinated directly to the copper ion for two reasons: 1) this would represent the most per- turbed environment to lower the vibrations to such a low frequency, and 2) these bands do not appear with as much intensity in Al- and Ca-clays, which would exclude any type of bonding to the clay surface. Since two bands occur, it was suggested that one of the OH groups was coordinated directly 97 to the cation. while the other was involved in either intra- or intermolecular H—bonding. Similar complexes have been suggested by Miyake (1959) for cobalt (II) and nickel (II) complexes in liquid glycol. The infrared Spectra also Showed that the predominantly OH deformation band of glycol at 1330 cm"1 was shifted to 1335 cm‘1 when adsorbed on Ca-montmoril- lonite and to 1350 cm"1 when adsorbed on Cu- and Al-montmoril- I lonite. This suggested that ion-dipole interactions were not ‘i as great in the glycol-Ca-clay systems as they were in the Cu- and Al-systems. 3 The glycol desorption curves were Similar for Cu-, Al-, “A and Ca-montmorillonite. They were easily divided into three sections: 1) the rapid vaporization of the glycol contribut- ing to multilayer and capillary adsorption, which was dissi- pated in approximately four hours; 2) the curvilinear loss of glycol with respect to time, which extended over the time interval from 4 to 160—200 hours and obeyed the conditions of second order chemical kinetics; 3) the constant rate loss of the remaining glycol. When the glycol content was greater than 25 g. per 100 g. clay, rational x-ray diffraction peaks were observed which ranged from 16.6 R for Ca-montmorillonite to 17.1 A for the highly solvated Al-montmorillonite. AS water begins to enter the interlamellar Spaces, rearrangements within the coordination spheres and glycol loss resulted in randomly interstratified systems. These results are in contrast to those of Mackenzie (1948) who noted a constant 98 17.1 A basal Spacing over a wide range of glycol:water ratios for Ca-montmorillonite. Absorption bands at 1575 cm‘1 and 1480-1400 cm"1 were observed to develop on clay films that were exposed to the atmOSphere following 24 hours of heating at 115°C in an evacuated system. By deuteration with D20, it was concluded that the 1575 cm-1 band arises from water external to the interlamellar Space. By the process of elimination and com- parisons with Spectra of known substances, the 1480—1400 cm"1 band was assigned to a carbonate vibration. Since the absor- bance of both of these bands increased with time of atmospheric exposure, it was suggested that both may arise from a rather undefined hydrated carbonate material. This type of Spon- taneous contamination of the clay surfaces emphasizes the importance of using freshly prepared clay films, which supports the ideas of Martin (1962). The results of these studies clearly demonstrate the in- fluential nature of the exchangeable cation in the adsorption- desorption of ethanol and ethylene glycol on homoionic mont- morillonite surfaces. This is in agreement with the well established fact that the water content of clay under a given set of conditions is directly related to properties of the saturating cation. However, there have been considerable differences of opinion as to the influential nature of the exchangeable cation upon the adsorption of non-ionic polar organic molecules on clay surfaces. Hoffmann and Brindley 99 (1961a) noted.that "..- . the exchangeable inorganic cation has no significant influence on the organic adsorption, at least for the ions Na+, Ca2+, and Mg2+." On the other hand Glaeser (1954) noted that Ca-montmorillonite retained more MeOH and EtOH than did Na-montmorillonite. Numerous workers have shown that the adsorption of glycol varies with the type of saturating cation (Dyal and Hendricks, 1952). This led Quirk (1955) to conclude that glycol molecules may ". . . be adsorbed around the cations on clay surfaces." No evidence was found to support the hypothesis put forth by Bradley (1945) that C-H---O-clay type interactions are important in the adsorption of non—ionic polar organic materials on clay surfaces. The C-H stretching vibrations of ethanol and glycol were not observed to Shift to a lower frequency upon adsorption, which should occur if C-H---O-clay bonds occur. In fact, upon adsorption, the CH vibrations of both ethanol and glycol appeared to shift to a slightly higher frequency, which would suggest bonding through the oxygen atom of the molecule (Bellamy, 1958). More direct evidence of oxygen-metal interaction was obtained in the Cu—montmorillonite system where infrared bands attributable to the O-H stretching modes of coordinated glycol were ob- served. To a lesser extent, similar bands were observed for Al-glycol complexes and were absent in the Ca-system. The O-H deformation vibrations of both ethanol and glycol were noted to Shift to a slightly higher frequency upon adsorption, which is in agreement with the above considerations. 100 The.adsorption of both ethanol and glycol is reversible to exchange with water at 40% relative humidity with the possible exception of Al—glycol complexes where a low level of glycol appears to remain in equilibrium with atmospheric moisture. Both ethanol and glycol are effective in dehydrat- ing the various homoionic clays, leaving only a highly dissoci- able species of water, which can contribute to the Spontaneous degradation of the clay lattice. Since ethanol vapor can essentially dehydrate the clay, it is suggested that the in- discriminant use of ethanol as a washing agent to remove excess salts from soils and clays, may render the final experi- mental results meaningless. A case in point is some work done by Sumner (1963) on four tropical soils. He noted that two (10 ml.) washings with 80% EtOH greatly reduced the positive and negative charges on the soil as measured by fi/Z NHICl retention. It has been pointed out that the x-ray data for ethanol and glycol systems can be explained by coordination complexes without having to inject O-H---O-clay type bonds as suggested by Emerson (1957) and Brindley and Ray (1964). One of the strongest, though indirect, proofs that cation—dipple inter- actions are of major importance is the differential response of homoionic montmorillonite systems to adsorption—desorption of ethanol and glycol. Therefore, it is not surprising that McNeal (1964) noted marked differences in the amount of glycol retained by montmorillonite saturated with various cations. 101 Accepting this thesis requires a break with the traditional use.of such terms as monomolecular layers. For such terms to be meaningful, as Quirk (1955) stated, they ". . . must have some significance for the same surface irrespective of the cation." CONCLUSIONS 1. Homoionic montmorillonites were essentially dehy- drated by equilibration with: a) ethanol vapor at a relative pressure of unity or b) ethylene glycol vapor at 1150C for 24 hours. 2. The adsorption of ethanol and ethylene glycol and their subsequent replacement by atmOSpheric moisture is a reversible process. However, the adsorption-desorption of these two compounds is a function of the saturating cation with respect to time, quantity, and type of complexes formed. 3. The rate of loss of ethanol from Cu-montmorillonite during rehydration is a diffusion controlled process. The replacement of ethylene glycol by water obeys the conditions of second order chemical kinetics. 4. The lack of uniformity in the Shifting of O-H vi- brations (stretching and deformation) of adsorbed ethanol and ethylene glycol suggest that cation-dipole type bonds, rather than O-H---O-clay type interactions, are of major importance in the binding of these compounds on montmorillonite surfaces saturated with polyvalent ions. More direct proof of ion- dipole type interactions are the infrared bands at 2750 and 2650 cm"1 which are directly attributable to the O-H stretch- ing modes of coordinated glycol. 102 103 5. There was no evidence of a C—H---O-clay type bond with either ethanol or ethylene glycol. 6. X-ray data for ethanol and ethylene glycol systems can be explained by coordination type complexes as easily as by O-H---O—clay type interactions. Egan [Manifikr/ /. t LITERATURE CITED Anderson, D. M. and G. Sposito. 1963. Rate of adsorption of water vapour by degassed Arizona bentonite. Nature 199:1085-1086. Barrer, R. M. and D. M. MacLeod. 1954. Intercalation and sorption by montmorillonite. Trans. Faraday Soc. 50:980-989. :53! I' Barrer, R. M. and J. S. S. Reay. 1957. Interlamellar sorption by montmorillonite. Proc. 2nd Int. Congr. Surface Activity 11:79-89. Barrow, G. M. 1952. Heat capacity, gas imperfection, infra- red Spectra, and internal rotation barriers of ethyl aléohol. Jour. Chem. Phys. 20:1739-1744. : Barrow, G. M. 1961. Physical Chemistry. McGraw-Hill Book Company, Inc., New York, pp. 618-640. Barshad, Isaac. 1952. Factors affecting the interlayer ex— pansion of vermiculite and montmorillonite with organic substances. Soil Sci. Soc. Am. Proc. 16:176-182. Barshad, Isaac. 1960. Thermodynamics of water adsorption and desorption on montmorillonite. In Clays and Clgy Minerals, 8th Conf., Pergamon Press, New York, 8:84-101. Bellamy, L. J. 1958. The Infra-red Spectra of Complex Molecules. John Wiley & Sons, Inc., New York. Benesi, H. A. and A. C. Jones. 1959. An infrared study of the water-silica gel system. Jour. Phys. Chem. 63: 179-182. Boreskov, G. K., Yu. M. Shchekochikhin, A. D. Makarov, and V. N. Filimonov. 1964. Investigation of the structure of surface compounds formed in the adsorption of ethanol on the y-oxide of aluminum, by the method of infrared adsorption spectra. Dokl. Akad. Nauk. SSSR., 156:901- 904. Bouyoucos, George John. 1936. The state in which the hygro- scopic moisture eXistS in soils as indicated by the determination with alcohol. Soil Sci. 41:443-447. 104 105 Bower, C. A. and F. B. Gschwend. 1952. Ethylene glycol retention by soils as a measure of surface area and interlayer swelling. Soil Sci. Soc. Am. Proc. 16: 342-345. Bradley, W. F. 1945. Molecular aSSOCiations between mont- morillonite and some polyfunctional organic liquids. Jour. Am. Chem. Soc. 67:975-981. Brindley, G. W. 1956. Allevardite, a swelling double-layer mica mineral. Am. Mineral. 41:91-103. Brindley, G. W. and R. W. Hoffmann. 1962. Orientation and packing of aliphatic chain molecules on montmorillonite. Clay-organic studies VI, in Clays and Clay Minerals, Pergamon Press, New York, 9:546-556. Brindley, G. W. and 8. Ray. 1964. Complexes of Ca-montmoril- lonite with primary monohydric alcohols (Clay-organic studies VIII). Am. Mineral. 49:106-115. Brindley, G. W. and M. Rustom. 1958. Adsorption and re- tention of an organic material by montmorillonite in the presence of water. Am. Mineral. 43:627-640. Brooks, C. S. 1960. Free energies of immersion for clay minerals in water, ethanol and n-heptane. Jour. Phys. Chem. 64:532-537. Coburn, W. C., Jr. and E. Grunwald. 1958. Infrared measure- ments of the association of ethanol in carbon tetra— chloride. Jour. Am. Chem. Soc. 80:1318-1322. Davison, W. H. T. 1955. Infrared Spectra and crystallinity. Part III. Poly (ethylene glycol). Jour. Chem. Soc. Depireux, J. 1957. Etude de la vibration characteristique du groupe carbonyle de quelques cetones et de quelques aldehydes par Spectrometrie d'absorption infra-rouge. Bull. Soc. Chim. Belg. 66:218-228. Drushel, H. V., W. L. Senn, Jr.. and J. S. Ellerbe. 1963. Effect of structure on the methyl and methylene group absorptivities in aliphatic alcohols. Spectrochim. Acta. 19:1815-1830. Dyal, R. S. and S. B. Hendricks. 1950. Total surface of clays in polar liquids as a characteristic index. Soil Sci. 69:421-432. 106 Dyal, R. S. and S. B. Hendricks. 1952. Formation of mixed layer minerals by potassium fixation in montmorillonite. Soil Sci. Soc. Am. Proc. 16:45-48. Eeckman, J. P. and H. Laudelout. 1961. Chemical stability of hydrogen—montmorillonite suspensions. Kolloid— Zeitschrift. 178:99-107. Emerson, W. W. 1957. Organo-clay complexes. Nature 180:48-49. Farmer, V. C. and M. M. Mortland. 1966. An infrared study of the coordination of pyridine and water to exchange— able cations in montmorillonite and saponite. Jour. Chem. Soc. (London). (in press) Folman, M. and D. J. C. Yates. 1959. Infrared Studies of physically adsorbed polar molecules and of the sur- face of a Silica adsorbent containing hydroxyl groups. Jour. Phys. Chem. 63:183-187. Fripiat, J. J., J. Chaussidon and R. Touillaux. 1960. Study of dehydration of montmorillonite and vermiculite by infrared Spectroscopy. Jour. Phys. Chem. 64:1234— 1241. Fripiat, J. J. and J. Helsen. 1966. Kinetics of decomposi- tion of cobalt coordination complexes on montmorillonite surfaces, in Clays and Clay Minerals. 13th Natl. Conf., (in press). . Gamo, Itaru. 1961a. Infrared Spectra of water of crystal- lization in some inorganic chlorides and sulfates. Bull. Chem. Soc. Japan. 34:760-764. Gamo, Itaru. 1961b. Infrared absorption spectra of water crystallization in copper sulfate penta- and mono- hydrate crystals. Bull. Chem. Soc. Japan. 34:764-766. Gillery, F. H. 1959. Adsorption-desorption characteristics of synthetic montmorillonoids in humid atmospheres. Am. Mineral. 44:806-818. Glaeser, Rachel. 1954. Complexes organo-argileux et role des cations exchangeable. These (Paris). Greene—Kelly, R. 1957. The montmorillonite minerals (smectites), in The Differential Thermal Investigations of Clays, ed. by R. C. Mackenzie. Mineral. Soc., London, pp. 140-164. 107 Greenler, Robert G. 1962. Infrared study of the adsorption of methanol and ethanol on aluminum oxide. Jour. Chem. Phys. 37:2094-2100. Haas, C. and D. F. Hornig. 1960. Inter- and intramolecular potentials and the spectrum of ice. .Jour. Chem. Phys. 32:1763-1769. Hadzi, D. and D. Jeramic. 1957. The infra-red Spectrum and the OH in-plane deformation frequency in CH3CD20H. Spectrochim. Acta. 9:263. Hadzi, D., I. Petrov and M. Zitka. 1962. ,The influence of hydrogen bonding upon the frequency and intensity of the OD in-plane deformation vibration of some phenols and alcohols. Advances in Molecular Spectroscopy, ed. by A. Mangina, vol 2, Pergamon Press, New York. Hass, H. and G. B. B. M. Sutherland. 1956. The infra-red spectrum and crystal structure of gypsum. Proc. Roy. Soc. (London) 236A:427—445. Hendricks, S. B., R. A. Nelson and L. T. Alexander. 1940. Hydration mechanism of the clay mineral montmorillonite saturated with various cations. Jour. Am. Chem. Soc. 62:1457-1464. Hoffmann, R. W. and G. W. Brindley. 1960. Adsorption of non-ionic aliphatic molecules from aqueous solutions on montmorillonite. Geochim. et Cosmochim. Acta. 20:15-29. Hoffmann, R. W. and G. W. Brindley. 1961a. Adsorption of ethylene glycol and glycerol by montmorillonite. Am. Mineral. 46:450-452. Hoffmann, R. W. and G. W. Brindley. 1961b. Infrared extinc- tion coefficients of ketones adsorbed on Ca-montmoril- lonite in relation to surface coverage. Clay-organic studies. Part IV. Jour. Phys. Chem. 65:443-448. Hornig, D. F., H. F. White and F. P. Reding. 1958. The infrared spectra of crystalline H20, D20, and HDO. Spectrochim. Acta. 12:338-349. Huang, C. K. and P. F. Kerr. 1960. Infrared study of the carbonate minerals. Am. Mineral. 45:311-324. Hunt, J. M., M. P. Wisherd and L. C. Bonham. 1950. Infra- red absorption Spectra of minerals and other inorganic compounds. Anal. Chem. 22:1478-1497. a iI.»\.Il_MI-IIH V . / 108 Hunt, J. P. 1963. Metal ions in aqueous solutions. W. A. Benjamin, Inc., New York. Jones, L. H. 1954. The infrared spectra and structure of LiOH, LiOH-HZO, and the deuterium species. Remark on fundamental frequency of OH. Jour. Chem. Phys. 22:217-219. Jurinak, J. J. 1957. The effect of clay minerals and ex- changeable cations on the adsorption of ethylene di- bromide vapor. Soil Sci. Soc. Am. Proc. 21:599-602. Kanbayashi, U. and K. Nukada. 1963. Rotational isomers of ethylene glycol. Nippon Kagaku Zasshi. 84:297-300. Krimm, S., C. Y. Liang and G. B. B. M. Sutherland. 1956. Infrared spectra of high polymers. V. Polyvinyl Alcohol. Jour. Poly. Sci. 22:227-247. Kuhn, M., W. Luttke and R. Mecke. 1959. IR-spektroskopischi untersuckungen fiber die rotationsisomerie bei 2-substi- tutierten Athanolen. Zeitschrift ffir Analytische Chemie. 170:106-114. Kunze, G. W. 1955. Anomalies in the ethylene glycol solva- tion technique used in x-ray diffraction, in Clays and Clay Minerals, NAC-NRS, pub. 395. pp. 88-93. Lyon, J. P. 1964. Infrared analysis of soil minerals, in Soil Clay Mineralogy. Ed. by C. 1. Rich and G. W. Kunze, The Univ. of North Carolina Press, Chapel Hill. Pp. 170-199. MacEwan, D. M. C. 1948. Complexes of clays with organic compounds. 1. Complex formation between montmoril- lonite and halloysite and certain organic liquids. Trans. Faraday Soc. 44:349-367. Mackenzie, R. C. 1948. Complexes of clays with organic compounds. Part II. Investigation of the ethylene glycol-water-montmorillonite system using the Karl Fischer reagent. Trans. Faraday Soc. 44:368—375. Margottin-Maclou, M. and M. Corval. 1963. Spectres d' absorption entre 2 et 15 microns de quelques composis deutiriés. Jour. Phys. et Phys-Chim. Biol. 60:1044- 1046. Marshall, C. E. 1936. Soil Science and mineralogy. Soil Sci. Soc. Am. Proc. 1:23-31. 109 Marshall, C. E. 1964. The Physical Chemistry and Mineralogy of Soils. Vol 1: Soil materials., John Wiley & Sons, Inc., pp. 188-210. Martin, R. T. 1955. Ethylene glycol retention by clays. Soil Sci. Soc. Am. Proc. 19:160-164. Martin, R. T. 1962. Adsorbed water on clay: A review, in Clays and Clay Minerals, 9th Conf., Pergamon Press, New York, 9:28-70. McNeal, B. L. 1964. Effect of exchangeable cations on glycol retention by clay minerals., Soil Sci; 97:96—102. Mering, J. 1946. On the hydration of montmorillonite. Trans. Faraday Soc. 42B:205-219. Miller, F. A. and C. H. Wilkins. 1952. Infrared Spectra and characteristic frequencies of inorganic ions. Anal. Chem. 24:1253-1294. Miyake, Akihisa. 1959. Infrared Spectra of glycols coordi— nated to metal ions. Bull. Chem. Soc. Japan. 32:1381- 1383. Miyazawa, T., K. Fukushima and Y. Ideguchi. 1962. Molecular vibrations and structure of high polymers. III. Polarized infrared spectra, normal vibrations, and helical conformation of polyethylene glycol. Jour. Chem. Phys. 37:2764-2776. Mooney, R. W., A. G. Keenan and L. A. Wood. 1952. Adsorp- tion of water vapor by montmorillonite. II. Effect of exchangeable ions and lattice swelling as measured by x-ray diffraction. Jour. Am. Chem. Soc. 74:1371- 1374. Morin, R. E. and H. S. Jacobs. 1964. Surface area deter- mination of soils by adsorption of ethylene glycol vapor. Soil Sci. Soc. Am. Proc. 28:190-194. Mortland, M. M. 1954. Specific surface and its relationship to some physical and chemical properties of soil. Soil Sci. 78:343-347. Mortland, M. M. and N. Barake. 1966. Interaction of ethyl— amine and metal ions on montmorillonite. Trans. 8th Intern. Congr. Soil Sci. (in press). Mortland, M. M., J. J. Fripiat, J. Chaussidon and J. Uytter- hoeven. 1963. Interaction between ammonia and the expanding lattices of montmorillonite and vermiculite. Jour. Phys. Chem. 67:248-258. 110 Mortland, M. M. and W. D. Kemper. 1965. Specific Surface, in Methods of Soil Analysis, Part I. Am. Soc. of Agron., MadiSon, Wisconsin, pp- 532-544. Nakagawa, I. and T. Shimanouchi. 1964. Infrared absorption spectra of aquo complexes and the nature of co-ordination bonds. Spectrochim. Acta. 20:429-439. Ockman, Nathan. 1958. The infra-red and Raman Spectra of ice. Advance in Phys. 7:199-220. Orchiston, g. D. 1953. Adsorption of water vapor: I. Soils at 25 C. Soil Sci. 76:453-465. Orchiston, H. D. 1954. Adsorption of water vapor: II. Clays at 250C. Soil Sci. 78:463-480. Orchiston, H. D. 1955. Adsorption of water vapor: III. Homoionic montmorillonites at 25°C. Soil Sci. 79: 71-78. Orchiston, H. D. 1959. Adsorption of water vapor: V. Homoionic illites at 250C. Soil Sci. 87:276-282. Plyler, E. K. 1952. Infrared Spectra of methanol, ethanol, and n-propanol. Jour. Res. Nat. Bur. Standard. 48:281-286. Pozefsky, A. and N. D. Coggeshall. 1951. Infrared absorp- tion studies of carbon-hydrogen stretching frequencies in sulfurized and oxygenated materials. Anal. Chem. 23:1611-1619. Puri, B. R. and K. Murari. 1964. Studies in surface area measurements of soils: 2. Surface area from a single point on the water isotherm. Soil Sci. 97:341-343. Quirk, J. P. 1955. Significance of surface area calculated from water vapor sorption isotherms by use of the B.E.T. equation. Soil Sci. 80:423-430. Redington, R. L. and D. E. Milligan. 1962. Infrared Spectro- scopic evidence for the rotation of the water molecule in solid argon. Jour. Chem. Phys. 37:2162-2166. Ross, G. H. and M. M. Mortland. 1966. A soil beidellite. Soil Sci. Soc. Am. Proc. 30: (in press). Ross, S. D. and J. Goldsmith. 1964. Factors affecting the infrared Spectra of planer anions with D3h symmetry-I. Carbonates of the main group and first row transition elements. Spectrochim. Acta. 20:781-784. 111 Rundle, R. E., K. Nakamoto and J. W. Richardson. 1955. Concerning hydrogen positions in aquo-complexes- CuC13-H20. Jour. Chem. Phys. 23:2450-2451. Russell, J. D. and V. C. Farmer. 1964. Infrared Spectro- scopic study of the dehydration of montmorillonite and saponite. Clay Min. Bull. 5:443-464. Stuart, A. V. and G. B. B. M. Sutherland. 1956. Effect of hydrogen bonding on the deformation frequencies of the hydroxyl group in alcohols. Jour. Chem. Phys. 24:559-570. Sumner, M. E. 1963. Effect of alcohol washing and pH value on leaching solution on positive and negative charges in ferruginous soils. Nature 198:1018-1019. Szymanski, H. A., D. N. Stamires and G. R. Lynch. 1960. Infrared spectra of water sorbed on synthetic zeolites. Jour. Opt. Soc. Am. 50:1323-1328. Tahoun, S. A. and M. M. Mortland. 1966. Complexes of mont- morillonite with primary, secondary, and tertiary amides. II. Coordination of amides on the surface of montmorillonite. Soil Sci. 102: (in press). Tarte, P. 1958. Researches sur les fréquences des de dé- formation OH-I. Spectre infrarouge des sels basiques de cuivre. Spectrochim. Acta. 13:107-119. Tarte, P. and R. Deponthiere. 1957. Spectres infrarouges des alcools ordinaires et deuteres dans le groupe hy- droxyle. Bull. Soc. Chim. Belg. 66:525-542. Tensmeyer, L. G., R. W. Hoffmann and G. W. Brindley. 1960. Infrared studies of some complexes between ketones and calcium montmorillonite. Clay-organic studies. Part III. Jour. Phys. Chem. 64:1655—1662. Tettenhorst, R., C. W. Beck and G. Brunton. 1962. Montmoril- ” lonite—Polyalcohol complexes, in Clays and Clay Minerals, Pergamon Press, New York, 9:500-519. Toth, S. J. 1955. Colloid chemistry of soils, in Chemistry of Soil, ed. by F. E. Bear. Reinhold Inc., New York, pp. 85-106. Uvarou, A. V. 1963. Infrared study of the interaction of adsorbed water and ethanol vapours with surface hy- droxyls on alumina. Russian Jour. Phys. Chem. (English trans.) 37:634-635. APPENDIX 112 113 Table 4. A comparison between infrared and gravimetric tech- niques for determination of the water content of homoionic montmorillonite. Weight of Adsorbedl Adsorbed2 Variation Peak Area Water Water , (mg.) ---(g./100 gg clay)---- % Cu-montmorillonite 312.0 11.22 11.17 + 0.45 245.2 8.00 8.27 - 3.26 214.6 6.52 6.45 + 1.08 168.7 4.58 4.64 — 1.29 152.9 4.20 4.15 + 1.20 Al—montmorillonite 280.0 10.60 10.90 - 2.75 250.1 8.42 7.68 + 3.38 194.9 4.30 4.79 -10.68 171.4 2.42 2.62 - 7.63 165.4 '1.70 1.54 +10.39 144.8 0.85 0.82 + 3.66 Ca-montmorillonite 351.7 15.00 15.45 - 2.91 295.0 11.20 10.90 + 2.75 240.5 7.55 7.42 + 1.75 212.5 5.70 5.67 + 0.53 175.1 4.22 4.25 - 0.70 161.8 3.95 3.93- + 0.50 Na-montmorillonite 131.7 4.62 4.62 0.00 6401 2.20 2029 - 5095 43.9 0.68 0.76 -10.52 32.1 0.40 0.36 +11.11 27.0 0.34 0.36 - 5.56 NHi-montmorillonite 97.0 4.80 4.77 + 1.47 70.7 3.55 3.70 - 4.05 52.8 2.65 2.60 + 1.92 38.3 1.90 1.78 + 6.74 26.7 1.40 1.53 - 8.50 1Determined from plots given in figure 2. 2Determined from quartz helix distension data. III IIIII'II .' , . a... . .Ib Iliad?! I.” r in. p . a. 71m» 5 In, 23,. :D- .1. IRIiSiuuwdwlIII..bd. .93! t u fi‘i]‘lJ.H.H.—II" . Iii}! - I . c A a all“?! I. I; no, !IV4J .87‘ I 5: IIIIIII 5 III 3 03071 351