Emil: !. ~ I.“ 31:43. .I .. “(h (l f 1‘} 1'2" "';[.:I1 “IIIZ‘vJA I ‘1‘ hi"; “. jab??? .‘I.I..8>¥( ,. u x ”,1 i‘.} g .31. D- . q. ‘ a,“ I“ :1 files . I J""‘l\am":\f N.;‘ .. ’ml‘ “1::- Wflmjaz 2.961; .f'fi'ILI'I ”it" I‘ ' .b‘v” &5 “#2"! I f: I ping "'L‘ . M. d"“$"fi$£ ‘3 1 x16: :2 ' : MM“: 2 -« “”33 W v v; ‘ ‘ -. . . ‘ ' . . ' M 3‘“ 6" ' ‘9' _ .<-. < 14:1. I, _>' _ . ' ' ‘ .I ' ' . '04-. 57‘ " :fif.‘ ' . . l‘ ‘ : . .‘ . . .. . . . . ‘ ‘ ‘l o . . . . {W‘n‘fli‘ HE‘QI‘U‘lr‘.‘ A It ..' '- - ‘ ‘ ' ”H1. ,‘. I‘;‘. .1 (‘1‘ '1‘)‘. ‘mf-ssw .... a . m . - 64‘???" : — I .J - ~1v_.’_'u -:l>.-I»‘.' ‘. " AW 3 u "git: (”IS C :n.6 ’ “I ;‘ '«n M..y:fi A): 1.35% If - ”.1 _ .9 - :4. ~';! 5‘ '1 $3 x4»? I ‘A Q. - v \ fit) 0'5 5?. 7 ’0 . 0.1 _n¥ ba-ncw-wr_;ww‘] ”Du-CV 0 0 p1 .' F": ‘3’" "vE, Lu-u-vviuwn'Oi This is to certify that the dissertation entitled I .\ I - «3'3h%€‘:£s cywot Sttuc‘lfu ‘od. Twoeefi ammo/E. 0‘8; ST i'TCa‘ifi/ CJGVZS presented by fat): baubh - SD‘AASU‘A has been accepted towards fulfillment of the requirements for {RD degreeinACngfingleaI k % m: r flpi’o’t‘essor Datej [Cl (85 MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 f’ MSU RETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from .—:,—. your record. FINES will be charged if book is returned after the date stamped below. JULJI-Ium fl Io‘f I \ ‘ :g ti}i?§fib SYNTHESIS AND STRUCTURAL INVESTIGATIONS OF SILICATE CLAYS BY Ivy Dawn Johnson A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1985 sili< fiiqu< micuj dist] Anot} laye1 ABSTRACT SYNTHESIS AND STRUCTURAL INVESTIGATIONS OF SILICATE CLAYS BY Ivy Dawn Johnson Several structural properties of synthetic and natural silicate clays have been investigated by a variety of tech- niques. In the first study, a series of anhydrous beCSl-x ver- miculites, where Dixil, were investigated to determine the distribution of the two ions within the interlayer galleries. Another objective was to determine the flexibility of the layers as might be reflected in the d00£ x-ray diffraction reflections. For a given value of x, only one basal spacing was observed with several orders of reflection. This in- dicated the presence of a single phase with the ions mixed within the same galleries. The dependence of the basal spacing on innerlayer cation composition was determined from 002 x-ray diffraction patterns. The basal spacings exhibited a discontinuity at x = 0.4 and is a result of x- dependent, correlated rotations of the SiO4 tetrahedra with- in the layer structure. The hexagonal cavities defined by cl ch. the Al te1 Ivy Dawn Johnson the oxygen atoms in the basal plane are distorted by the ro— tations so that the interlayer cations are encapsulated within these hexagonal cavities. Thus the transverse rigidity of the vermiculite layer is preserved. The com- position dependence of the Si—O torsional mode was examined by lasar Raman spectroscopy. This mode also exhibited a discontinuity at x = 0.4. The results further substantiate the claim that the distortion in the tetrahedral sheet is dependent on the innerlayer composition. In the second study, the silicon and aluminum distribu- tion in the tetrahedral sheet of some tetrahedrally charged clays was investigated. In tetrahedrally charged clays, the charge arises from the substitution of Al(III) for Si(IV) in the tetrahedral sheet. By examining the ordering of the Al(III), information about the charge distribution in the tetrahedral sheet can be obtained. A series of saponites, Nax_y (Mg6-yAly)(SiB-XAlx)020(OH)4' synthesized by the hydro- thermal crystallization of gels, and micas—muscovite K2[Al4]- (Si6A12)020(OH)4, and phlogopite K2[Mg6](Si6A12)020(OH)4, 2gsi mas nmr. were investigated by Analysis of the 298i spectra indicate that there are no adjacent tetrahedral sites occupied by Al; hence, Loewenstein's rule is obeyed. There is also evidence that in the clays with no impurity phases, there is some short range order of Al in the tetrahedral sheet. Presence of the short range order is indicated by the minimization of the number of Si Wltfi alu ; viror and n envir a SiC tetra is cc effec inne: In t} is a1 Othe] not < the 5 the j Of &; I Ivy Dawn Johnson with two or three neighboring tetrahedral sites occupied by aluminum. The third study is an investigation of the silicon en- vironments in layered silicic acids kanemite, NaHSiZO 29 5, and magadiite, Na28i14029 by using Si mas nmr. Two silicon environments Q3 and Q4 were observed. In the 03 environment a SiO4 tetrahedron is condensed to three neighboring silicon tetrahedra, and in the Q4 environment the SiO4 tetrahedron is condensed to four neighboring silica tetrahedra. The effects of exchanging protons for the sodium ions in the innerlayer as well as aging the materials were also examined. In the case of kanemite, aging leads to the appearance of a Q4 environment at the expense of the Q3 environment. This is attributed to adjacent silicate layers condensing to each other through surface hydroxide groups. Na-magadiite does not demonstrate this condensation. Exchanging protons for the sodium ions in either kanemite or magadiite also increases the intensity of the Q4 signal and decreases the intensity of the Q3 signal, indicative of the condensation phenomenon. TO MY FAMILY Shirley: carl Bradly. Randy, Sherry ii I believe out my appreci ularly that o formal in the have membe Staff Frid; Acknowledgments I would like to thank Dr. T.J. Pinnavaia who truely believed in pygmalia. His guidance and support through- out my graduate career were, and still are, sincerely appreciated. I would also like to acknowledge my family, partic4 ularly my father and mother, who instilled in me the belief that one can never have too much education, whether it be formal or informal. I would like to thank Dr. H. Eick for his assistance in the preparation of this manuscript. I would also like to thank the numerous peOple who have aided me in my graduate career - past and present group members, especially Rasik, and the technical and clerical staff in the Chemistry Department. Finally, I would like to thank the members of the Friday Social Club who helped me maintain my persPective. iii Chapt LIST LIST PHYS TABLE OF CONTENTS Chapter LIST OF TABLES LIST OF FIGURES. . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . 1.1. Structure of Clay Minerals. . 1.2. Ion Exchange in the Innerlayer. 1.3. Layered Silicic Acids 1.4. Synthesis of Clay Minerals. 1.5. MAS NMR of Clays. PHYSICAL METHODS . . . . . . . . 2.1. Cation Exchange Capacity Measure— ments (CEC) . . . . . . . . . 2.2. Chemical Analysis . . . . . . 2.2.1. Atomic Emission Spectros— copy (AE) . . . . . 2.2.2. Inductively Coupled Plasma Emission Spectroscopy (ICP) 2.2.3. Neutron Activation Analysis (NAA) . . . 2.2.4. X-ray Fluorescence (XRF). 2.3. High Pressure Reactor 2.4. Magic Angle Spinning Nuclear Mag- netic Resonance . . . . . . . . 2.5. NMR Line Shape Analysis 2.6. Raman Spectroscopy. iv Page viii 14 19 22 28 37 37 38 38 38 39 40 40 41 42 42 ChaptEI S'NTHE MI RED ma Chapter 2.7. X-Ray Powder Diffraction (XRD). SYNTHESIS AND STRUCTURAL INVESTIGATION OF MIXED METAL INTERCALATED VERMICULITES. 3.1. Objectives. . . . . . . . . . . . . 3.2. Experimental. 3.2.1. Materials 3.2.2. Methods 3.2.2.1. Li-Exchanged Vermiculite. 3.2.2.2. Rb/Cs-Exchanged Vermicu- lite. . . . . . 3.2.2.3. Preparation of Uptake Isotherms . . . 3.3. Results and Discussion. 3.3.1. Properties of Llano Vermicu— lite. . 3.3.2. Proposed Mechanism for the Synthesis of Mixed-Metal Vermiculites. 3.3.3. Characterization by X-Ray Diffraction . . . . 3.3.4. Comparison of the becs(1-x) Vermiculite System to the Analogous RbXCs(l_X)Graphite Systems. 3.3.5. Characterization by Raman Spec— troscopy. INVESTIGATION OF CHARGE DISTRIBUTION IN CLAY MINERALS USING MAS NMR . . . . . . . . . . . . . . 4.1. Objectives. 4.2. Experimental. 4.2.1. Materials 4.2.2. Synthesis of Saponites. Page 43 44 44 46 46 '47 47 47 51 52 52 53 56 67 69 75 75 76 76 77 ["1 fr, [’1 '1 Chapter Page 4.2.3. Purification of Natural Saponite. . . . . . . . . . . . . . 79 4.3. Results and Discussion. . . . . . . . . . 80 4.3.1 Characterization of the Materials Investigated. . . . . . . . . . . . 80 4.3.2 Magic Angle Spinning NMR. . . . . . 85 4.3.3 Analysis of Site Distribution Based on 2951 MAS NMR . . . . . . . 92 LAYERED SILICIC ACIDS. . . . . . . . . . . . . . . . 104 5.1. Objectives. . . . . . . . . . . . . . . . 104 5.2. Experimental. . . . . . . . . . . . . . . 108 5.2.1. Materials . . . . . . . . . . . . . 108 5.2.2. Synthesis of Layered Silicic Acids . . . . . . . . . . . . . . . 108 5.3. Results and Discussion. . . . . . . . . . 110 5.3.1 Characterization of Silicic Acids . . . . . . . . . . . . . . . 111 5.3.2 2gsi MAS NMR of Layered Silicates . . . . . . . . . . . . . 112 5.3.3. Kanemite. . . . . . . . . . . . . . 114 5.3.4 Magadiite . . . . . . . . . . . . . 122 REFERENCES . . . . . . 133 vi Table Table LIST OF TABLES Page Classification scheme for clay minerals. . . . 6 Idealized structural formulas for some dioctahedral and trioctahedral 2:1 phyllosilicate . . . . . . . . . . . . . . . . 8 Influence of charge per structural unit, x, and nature of interlayer cation on the maximum value of basal spacing (nm) of different 2:1 type clay minerals on hydra- tion . . . . . . . . . . . . . . . . . . . . . 9 Reaction products in (a) A1203-Si02-H20 and (b) (CaMg)O-(K2Na2)O-Ale3-SiOz-H20 systems. . . . . . . . . . . . . . . . . . . . 25 2951 chemical shifts(1) (ppm) in solid silicates. . . . . . . . . . . . . . . . . . . 33 (a)"Si chemical shifts (ppm) in layered silicates relative to TMS. (b)2981 chemical shifts (ppm) in tecto- silicates relative to TMS. . . . . . . . . . . 34 Synthetic conditions for RbXCs(1_X) vermiculites . . . . . . . . . . . . . . . . . 50 vii Tabl 10 ll 14 15 Table Page 8 Normalized, calculated and observed intensities(1) of the hkl reflections for Rb-vermiculite, x = 1, and Cs- vermiculite, x = 0 . . . . . . . . . . . . . . 64 9 Molar composition of gels before hydro- thermal crystallization. . . . . . . . . . . . 78 10 (a) Unit cell compositions, cation ex- change capacities and basal spacings for synthetic saponite clay minerals. . . . . . . 82 (b) Unit cell compositions for micas. . . . . 83 11 29Si chemical shifts(1) and the nor- malized integrated intensities of the component peaks. . . . . . . . . . . . . . . . 90 12 Fraction of tetrahedral sites occupied by Al (x) and number of Al-Al near neighbor tetrahedral sites (x1) . . . . . . . . . . . . 97 13 Fraction of silicon atoms with two or more near neighbor tetrahedral sites occupied by aluminum atoms (X2). . . . . . . . 102 14 Possible guest molecules for inter- lamellar adsorption by crystalline silicic acids and layer separation (nm)(49) . . . . . . . . . . . . . . . . . . . 105 15 Compositions and basal spacings of layered silicic acids and their sodium salts . . . . . lll viii Tabl Table Page 16 29 Si mas nmr chemical shifts(1) in ppm and normalized intensities of the signals. . . . . . . . . . . . . . . . . . . . 113 ix Figure LIST OF FIGURES Page Schematic representation of the smectite structure. . . . I . . . . . . . . . . . . 2 Top view of ideal clay layer demonstrat- ing its hexagonal symmetry . . . . . . . . 4 Schematic view of the +/— b/3 shift observed in Mg-vermiculite . . . . . . . . 12 Schematic view of the basal surface showing the distortion of the hexagonal symmetry by twisting the individual silica tetra- hedra. . . . . . . . . . . . . . . . . . . 13 Variation of the d(001) spacing as a function of EtNH + mole fraction in 3 mixed Ca2+-EtNH3+ montmorillonites ob- tained by ion exchange (a,b) and mixing clay suspensions of end members (c,d). The A represent wet, and the 0 dried samples, respectively. . . . . . . . . . . 18 Variation of the d(001) spacing as a function of mole % alcohol (a) and mole % acetone (b). Solid symbols are for well-ordered phases, open symbols for Figure 10 ll Figure 10 ll poorly-ordered or mixed—layer phases . . . . . . . . . . (a) Comparison and ranges of 29Si chemical shifts with different degrees of condensation of the silica tetra- hedra. (b) The upfield chemical shift imposed by the substitution of A1 for Si Uptake isotherms for the synthesis of mixed becs(l-x) vermiculites. The mole fraction of Rb(I) in the final product, x, is plotted against the Rb/Cs ratio in solution. Reaction times are 2 hrs (0), 18 hrs (I), and 40 hrs (X). Schematic summary of the synthesis of becs(l-x) vermiculites . . . . X-ray diffraction patterns for oriented film samples of air-dried RbXCs(l_X) vermiculites for (a) x = 0.058 and (b) x = 0.74, intensity against 28 X-ray diffraction patterns(l) obtained from completely dehydrated clays, ar- bitrary intensity units against q, where q = sine/1 xi Page 21 31 31 49 55 58 62 Figure 12 l3 14 15 16 17 Page Normalized basal spacings against x for becs(l-x) vermiculites (O), and for becs(1—x)c8 ([3) as well as Raman shift, cm-l, vs X LA). . . . . . . . . . . 66 Raman spectra for becs(l-x) vermiculites showing Si—O stretch and the fluorescence lines used for calibration. Included is a schematic illustration of the rotation of the Si tetrahedra . . . . . . . . . . . . 71 2981 mas nmr spectra (1) and deconvoluted components for samples 2(a), 3(b), 8(c), 9(d), 11(e), and 12(f). PW, RD, and NS are provided for each spectra. . . . . . . . 87 27A1 mas nmr spectra(1) for samples 4(a), 5(b), and 7(C) . . . . . . . . . . . . . . . 93 Schematic illustration of the possible environments for Si; (0) and A1 (.0) ionsin the tetrahedral sheet. . . . . . . . . . . . . . . . . . . . 94 Models for the interlayer structure of long chain magadiite derivatives, (a) arrangement of alkyl chains in alkylam- monium magadiite; (b) arrangement of n— alkylpyridinium ions in alkylpyridinium magadiite(50a) . . . . . . . . . . . . . . . 106 xii Figur 18 19 20 21 22 23 Figure Page 18 Examples of zweiereinfachschicht struc- tures observed in clay minerals(54). . . . 116 19 29Si mas nmr(1) of (a) freshly prepared NaHSiZOS-HZO, obtained on the 400 MHz spectrometer and (b) aged NaHSiZOS-HZO, obtained on the WH-180 MHz Spectrometer. . 120 20 29Si mas nmr(1) of aged HZSiZOS obtained . on the WH-180 MHz spectrometer . . . . . . 121 21 X-ray diffraction patterns of (a) NaHSiZOS and (b) HZSiZOS . . . . . . . . . 124 22 298i mas nmr(1) of (a) NaZSil4029’ (b) freshly prepared H25i14029, (c) heated H25i14029 and (d) aged H28i14029 obtained on the WH-180 MHz spectrometer. . 127 23 X-ray diffraction patterns for (a) Na25i14029 and (b) HZSil4029 129 xiii INTRODUCTION 1.1. Structure of Clay Minerals The term "clay" is geological in origin, and refers to a finely divided material with a particle size of less than Zn. The term "clay mineral" refers to silicate clays with definite stoichiometry and crystalline structure. The terms are often used interchangably. The clay minerals, or phyllosilicates, are normally layered species with a repeat pattern defined by parallel planes of framework oxygen atoms. Figure 1 illustrates schematically the struc- ture of a common family of layered silicates with a "2:1" structure. The outer two planes of oxygen atoms define two sheets of tetrahedral cavities while the inner two planes .of oxygen atoms define a sheet of octahedral cavities. There are two "tetrahedral sheets" condensed on a single "octahedral sheet" thereby the name 2:1. A 1:1 clay con- sists of one tetrahedral sheet condensed onto one octahedral sheet. Chloritic clays are 2:1:1 structures where the aforementioned 2:1 layer alternates with an innerlayer metal hydroxide sheet such as brucite, Mg(OH)2, or gib- bsite, Al(OH)3. This hydroxide sheet does not share oxygen atoms with the adjacent tetrahedral sheets(1). Figure J 0 Oxygen 9 Hydroxyl s oAluminum, Magnesium, Iron oSih’con, Occassionaly Aluminum Figure 1. Schematic representation of the structure of Smectite. sheet of on compa ness Basal can t ideal 6—rir of at L1 frame hedra ions and Fe-1.2 and <-2.0 per formula unit. The charge originates in the tetrahedral sheet. Smectite is a broad category of phyllosilicates with a layer charge loetween —l.2 and -0.4 per formula unit. Their charge can (originate either in the tetrahedral or octahedral sheet, .HLd—u .- phh \» a. .YJ L AU -.~, in p: we . ~ y. u m... :nu .41. Peru HM ..~ ~. Piv~ a «$.35 5033500505000 0055.. M05000 050:8 05 E 5530.: .00 am 8:: # .00>mw 03 «030% 005050 0.55 05 300 0 0.5300020 .00m000m 8.035 350.3501 H. 02095 3:520 35053025. 00 03309009 3:520 35003520 05520 0131.9» 1 1 N 0308020 0005.0 033.5 350530018 . 0:390: 3050 0255 3523.020 003.4 0.5.5 0. l 350m02m .0505 .30va 35028002? 0:00w03m 6:50.52 30:2 35053020 .3022 N). sm50m500> 350300018 005000000> 3500300200 82080000.? 3.5053005 0.0500503, 3523.005 8:030:00» w . Him «Hz 0:000:0m 0.030093 03.50003 .0z0oamm 5 0.050005 3500300205 03000002 65—020mm ”03005500002 03000300505500.002 63005550002 5 0050005 350030005 5 050005 N . Hdm .o I. 03H. 330.. 3500300200 ozzfaoinm n0:__>:m00>m 3503000mQ 030-053.500xm o 7.. in 0.000600 63.53.0400 8050095 350300015 0:95—33 630:3“! 00:53.3” 350030005 0030090...-0z0m—00M o .1. _ u _ 3 as 03325 95355 @0000 HMHSH ODHU m 09C. 000 03030 .mamuocflfi MmHo How 080000 cofiumoflwflmmmHu .H magma or in both. The low layer charge gives the smectite class of silicates unique pr0perties that will be described later. The net charge of the clay minerals is balanced by counter ions located between the layers. If the innerlayer cations are dehydrated, the ions reside in the hexagonal cavities formed by the basal oxygen atoms shown schematically in Figure 2. In nature these cations are normally Na(I), K(I), Mg(II), Ca(II), or Fe(II). Some ideal formula units are presented in Table 2. Often there is a hydration sphere associated with these innerlayer cations, the properties of which depend on the hydration energy of the specific cation. The water is ordered to some extent in the interlammelar region, usually into layers. Mg-vermiculite, for example, has two layers of water arranged around the Mg cation in an octahedral array at ambient temperature and humidity. The extent of hydration of a clay is reflected in the basal spacing. In Table 3 are listed different basal spacings associated with different cations. Anywhere between one and an in- finite number of water layers are possible in the gallery depending on the mineral, the innerlayer cation, and the relative humidity. When there is an increase in the basal spacing, the clay is said to "swell". Swelling occurs with multiple layers of water as well as with other solvents. This is an especially important phenomena with smectites where the low charge density allows the clay layers to be fdb It» .. K at»; vxxovomo Ao.mp;a ¢Axovomoxx0oum=z vizovomoixp£ CC urrfi . Hmmmfl MO QDHQ> ESEfiXGE OS: .ufics HMHDuUSHum HQQ QOMQCU uO AECV OCAUGQW W0 NMDUGC QCM .X .030 0230 020000000 05 .5 0.020300 0:00:00 05:0 0w 000 0033 2 05¢ A 208 3. w2=0>>m 00320002 0300038 02w >02 9 L 80 m? u u to o: of m m 2:2 2: 0: on; m. m. +00 «m; om; m 1! 34$ 00 2: $0 2: ow ,m :6 . 21 03 8.. a .. :02 o: 2: .m m :2 E0. 3; a: *2: m w. +02 00.80 as .8.“ P .5 02x mlx maanNAZR m.a|m.o_lx 02». 0230 0200 0:25 022 05:20:30. 0200005 03702:»:Qoim 00.31005 b. 95..» 3000?: x20 5 b b ilt .0000000»: 00 0H0000HE wmao 0mm» Hum 00000MMH0 mo 280v 0000000 H0000 mo 00H0> Esaflxme 000 00 000000 u0wmau0u0a no 000000 000 .x .0000 amusuosnum 00¢ 000000 mo 00000HM0H .m 0H008 easily mineral much la For the the lay electrc the cat about 5 separat the cat enviror ThE to the tion 8;: diSSocj The Brc gallerj Same C' The layers using } tYPes SUCh a 10 easily and extensively separated. After swelling the mineral with an appropriate solvent, molecules which are much larger than the pristine cations-may be intercalated. For the simple hydrated cations, Cu(II) and Mn(II), where the layers are expanded by one to three water layers, electron spin resonance studies have shown that although the cation is oriented, it undergoes anisotropic rotations about specific molecular axes. When the innerlayers are separated beyond the dimensions of the hydration sphere, the cations tumble rapidly, indicative of a solution-like environment in the interlamellar region(3). The hydrated innerlayer cation adds an acidic nature to the clays. The water molecules in the primary hydra- tion sphere are polarized by the cation and some of them dissociate(l,2): Na+(H20X)+ Na(OH)(H20)(X_1) + H+ . The Bronsted acidity of these hydrated cations in the clay galleries has been shown to be greater than that of the same cations in homogeneous aqueous solution(4). The innerlayer cations also influence how the clay layers orient themselves with respect to each other. Again using hydrated vermiculite as an example, there are two types of layer stacking. With weak polarizing species such as Na(I) and Ca(II) the hexagonal cavities of the adjacen With st the lay cell di 3. The tion sp layers It the ex: length and Mg' sion c: Dehydr: a decre rotati also 3; the in' In t0 acc Sheet hedra baSal Where deSCri estima I I\ the la 11 adjacent basal surfaces lie directly opposite each other. With strong polarizing species such as Mg(II) and Ni(II) the layers are shifted +/- b/3 where b is a lateral unit cell dimension. This arrangement is depicted in Figure 3. The strong polarizing cations form more compact hydra- tion spheres which allow extensive H bonding between the layers and influence the stacking arrangements(5,15). It has also been shown by Brindley and Brown(5) that the exchange cation and its hydration state affect the length of the b axis in the clay lattice. Hydrated Li(I) and Mg(II) saturated micas show increases in the b dimen— sion compared to that of the parent K(I) saturated species. Dehydration of clays which contain Li or Mg at 350°C causes a decrease in the b parameter which is attributed to a rotation of the Si tetrahedra. Dehydrated montmorillonite also shows a b parameter variation which is dependent on the innerlayer cation. In the clay layers, the tetrahedral sheet is distorted to accommodate the misfit between itself and the octahedral sheet to which it is condensed. The individual Si tetra- hedra rotate around the Si-O bond perpendicular to the basal surface. This distortion is depicted in Figure 4, where the arrows show the direction of rotation. It is described in terms of a twist angle. This angle can be estimated(5) by cosa = b(observed)/b(ideal) where b is the lattice parameter, or calculate exactly by 12 b/3 shift observed Schematic view of the +/- in Mg-vermiculite. Figure 3. Figure l3 TETRAHEDRAL LAYER ROTATION a a =Arc cos(bm/4V/§-d,) Fi gure 4. Schematic view of the basal surface showing the distortion of the hexagonal symmetry by twisting the individual silica tetrahedra. 14 a = arc cos (bm/4 (/'2')dt) where bm and dt are distances defined in Figure 4. The twist angle which ranges between 3.0 and 11.4 degrees in naturally occuring minerals, changes the cavity symmetry from hexagonal toward trigonal as the twist angle approach- es its theoretical maximum of 30 degrees. This maximum has never been observed and is energetically unfavorable due to the repulsive forces between basal oxygens(6) . l .2. Ion Exchange in the Innerlayer Innerlayer cations can be replaced through ion exchange reactions with other simple cations, organic(1) and metal Complexes(7), silicates(8), and polyoxo-inorganic cations (9). Several of these are robust cations that act as pil- lars and prop the clay layers apart even in the absence Of a solvent. Ion exchange reactions in clay minerals have several characteristics in common(10) and are listed below. One is that the reaction rates are controlled by diffusion of the ions in the innerlayer. This process involves movement of the exchanging ions against concen- tration gradients. Normally this process is very slow, but increasing the concentration of the exchange cations increase the exchange rate by mass action. Another charac— tEristic of the ion exchange phenomenon is that the exchange mUSt be charge stoichiometric. The exchange must be equal Ofii the 50 CE 15 <3n an equivalents basis to maintain charge neutrality in 'the mineral. The relative number of exchange cations is 1:herefore a function of their charges. Ion exchange reac- tLiODS are also reversible but the reverse reaction may be SK) slow that it is not readily observed. A final charac- tearistic observed with most ion exchange reactions is that cxf ion selectivity or preference. Ion selectivity is a fainction of ionic size parameters such as polarizability, hydrated radii, and Debye-Hiickel closest approach param- erters which involve electrostatic repulsions. Among homo- \Lalent cations, the smaller the effective radius, i.e., hydrated size, is, the more preferred the ion is. This is riaflected in the Lyotropic series which shows monovalent (:ation preference by soil colloids, Cs > Rb > K > Na > Li (10), where the effective radius of the cation increases from left to right. Clays with two different innerlayer moieties have been sstudied to better understand ion selectivity and fixation. When a substrate such as a phyllosilicate contains two dif- ferent exchange cations, the ions will often separate into distinct regions or "demix". It is not known whether de— mixing occurs by domain or alternating layers(10). Demix- ing is usually investigated by measuring the basal spacings with x-ray diffraction. Clays with two different inner- layer cations show random interstratification when separate phases which contain either one cation or the other are (“1’ In 16 randomly dispersed or show complete segregation of the ions into two separate phases. Demixing is a function of the structure and hydration properties of both exchange ions, the structure of the interlamellar hydrated exchange- able cation complex, the location and distribution of the exchange sites, and the interaction of the ions with the silicate sheet. An increase in demixing is observed as the cation selectivity difference and/or the heterogeneity in charge distribution increase(ll). Order-disorder theory has been employed by Barrer and Klinowski(12) to calculate thermodynamic properties of systems with homogeneous site distributions and two dif- ferent exchange cations, A and B. Statistical thermo- dynamic calculations indicate that, with a homogeneous set of exchange sites and a random distribution of ions on these sites, there are different relaxation energies in— volved if a site next to ion A is occupied by another ion A, a different ion B, or a vacancy. Calculations might alSo be designed to allow consideration of different exchange sites in homogeneous arrangements but this possibility was not included. It may be assumed that if a framework is adjusted to accommodate ion A there is a misfit of the other ion, B. This can lead to local framework relaxation and/or minor displacement of A. As a result, the binding energies for the two cations to the clay layer will be dif- ferent. Ion exchange isotherms and equilibrium constants 17 have been calculated for some systems by using relaxation and binding energies, on the assumption of a homogeneous site distribution and specific crystallographic parameters. Smectites, micas, and vermiculites with two different innerlayer cations A and B have randomly interstratified layers containing primarily A or B ions, but not both, within a given layer. Different cationic sizes impose dif- ferent sheet separations. Barrer and Klinowski state that both basal spacings cannot be accommodated by a single layer; therefore A and B avoid forming mixed cation inner- layer species(12). Several mixed innerlayer clays have been studied(ll,5, 13,16). Montmorillonites with mixed Ca+2/ethylammonium (EtNH?) ion innerlayers demonstrate a step function when E is plotted. Figure 5 basal Spacing vs mole percent EtNH illustrates that the basal spacing has a sharp transition between the end member species at a critical composition. The materials used in 5a and 5b were synthesized in the traditional manner. All the pristine Na ions were replaced with Ca+2, which were then partially exchanged with EtNHE. A single phase clay with a homogeneous distribution of the two innerlayer cations was obtained. When end members of pure Ca+2- and pure EtNHS—montmorillonites were thoroughly mixed in various proportions, two phases remained present (Figures Sc and 5d) indicative of an inhomogeneous cation distribution. Air drying the Ca2+/EtNH;-montmorillonite l8 r F” T r r* T r r dom La c 19 ———-1A—-A-A ---A—-—A—A_A__. . X A T ‘ 16 ‘ +- —-4 13 t A—A-‘L‘- A-A-‘—--A-—A b . d ,-‘o- _._ 12 l‘ ,,0 1 .‘CJ‘--. i /‘.’ I()[-d’ T _ l l L L l I l l 02 CA: Of; (18 02 04 05 CH3 ZETNH; Figure 5. Variation of the d(001) spacing as a function of EtNH3+ mole fraction in mixed Ca2+-EtNH3+ montmorillonites obtained by ion exchange (a,b) and mixing clay suspensions of end members (c,d). The A represent wet, and the. dried samples, respectively. 19 systems produced regularly interstratified layers attributed to reorganization of the hydrated layers. The interstrati- fication reflects the partial removal of innerlayer water. Neutral molecules have also been studied as mixed inner- layer species. Figures 6a and 6b illustrate basal spacing vs mole fraction of water/alcohol mixtures and water/ace- tone solutions. Again, a sharp transition between the end member interlamellar spacings is observed at a critical composition. It could be concluded from these hydrated systems that the clay layers are rigid and are propped apart by the innerlayer cations. 1.3. Layered Silicic Acids A class of clay minerals that deviates from the afore— mentioned structures is that of layered silicic acids. They are of interest due to their varied intercalation chemistry and surface silanol groups. The layers consist of'SiZO5 units condensed to varying degrees. The simplest of these layered silicates is kanemite NaZHSi *3H20, in 205 which Na is the counter ion. The hypothesized structure is "zweiereinfachschichten" which consists of sheets of Si tetrahedra condensed through the basal oxygen with the apical hydroxide groups pointing in alternating directions. Other, more condensed layered silicates are known: maka— tite, Na231406(OH)6-H20(l7), magadite, NaZSi 3H20(l8), and kenyaite Na 14026(OH)6' 281220'9l(OH)8-3H20(18). The 20 Figure 6. Variation of the d(001) spacing as a function of mole % alcohol (a) and mole % acetone (b). Solid symbols are for well-ordered phases, open symbols for poorly-ordered or mixed-layer phases. I T l l l l l l I , f? 22 b ’ ‘ .. I I .f‘ “x -0 aloomA) l ,0 g, I ._ 20 I l// I T ’0 | A~o| g/ A, . ._.___.| .0-.. I -—4 \ ' . w | 1 n-propanol J ‘3 *- ' l lT I " k 'l ethanol - I F— l methanol ‘7 O 16 l l l l l l l l L J 20 40 60 80 . 100 mole-% alcohol i i l l l l l l I l .. ‘ -l ‘ acetone — ‘ -l l —+ 9, 26 -- \ ~ \ 24 —° ‘ "‘ d(001) (A) - \ “ ' l 20 -- ‘- O— ‘— 18 o F 1 l l 1 1 l l I I t, 20 40 60 80 100 mole-96 acetone Figure 6 22 structures of these materials are unknown, but intercala- tion chemistry indicates silanol groups are available on the innerlayer surface(l8). 1.4. Synthesis of Clay Minerals Although most clays are relatively abundant in nature, use of synthetic minerals is desirable when a series of compositions is to be studied. Synthetic clay minerals used to be difficult to identify because of their very small particle size, poor crystallinity, and contamina- tion by large amounts of unreacted materials. These iden- tification problems were circumvented with the development of x-ray diffraction which is used to detect the crystal- line clays even in the presence of amorphous contaminates and by-products. A simplistic model of clay formation either in nature or the laboratory involves the molecular solubility of the silicic acid and other components, a slow homogeneous dispersion of colloidal hydrosols, a rela- tively fast formation of nonrigid hydrogels, then a slow conversion to rigid silicates. In nature, this last step can take millions of years but can be accomplished faster in the laboratory by increasing the temperature and/or pressure. The process is irreversible and the boundaries are flexible. Differentiation between the phases is rather arbitrary(10). 23 Smectites and kaolinites are commonly synthesized in the laboratory by hydrothermal treatment of mixtures of oxides and hydroxides. Tables 4a and 4b illustrate materials synthesized from various oxide mixtures and reaction con- ditions. Temperatures for these reactions range from 250 to 850°C; pressures of up to 1000 atmospheres are re- quired. Reaction times range from one day to two weeks. Optimum conditions for smectite synthesis require a 0.2: 1:4 molar ratio of (Ca,Mg)O or (Na2K2)O to Al 0 :SiO 2 3 2 respectively, at temperatures less than 400°C. Increasing the concentration of MgO relative to the silica and alumina concentrations results in the formation of talc, then ser- pentine plus kaolinite or pyrophyllite. At temperatures greater than 400°C pyrophyllite is formed preferentially. Several studies have been made with Mg—Al-Si-HZO systems where mixtures of oxides,co-precipitated gels, glasses and natural minerals(l9-26) were used. As a general observa- tion, not a hard fast rule, the 1:1 structures like kaoli- nite and serpenties are synthesized at low pH and 2:1 min- erals like smectites and mica are synthesized at high pH. A mechanism for clay formation proposes that the acidic conditions favor formation of gibbsite, a layered, octa- hedrally coordinated aluminum hydroxide where the silica tetrahedra condense in a two dimensional array. Basic conditions stabilize alumina tetrahedra and allow for 2:1 condensation. 24 Table 4. Reaction products in (a) A12O3-Si02-H20 and (b) (CaMg)O-(K2Na2)O-AlZO3-SiOZHZO systems. a) b) 25 Temp., Molecular ratio Al.0.:SiO, (“.0 constant) ‘0 (Pressure, . ' ' : l :4 <1 :4 0. l a! m) l .0 > 1 .2 l 2 500 Corundum Perphyllite (530-540) + 7 Pyrophyllite Pyrophyllitc Pyrophyllite + SiO: + (amorphous) 400 boehmite + . (300) kaolinite SlO: 7 (amorphous) Kaolinite 350 Boehmite + (168) boehmite 300 Kaolinite (87) Kaolinite Kaolinite + SiO. 250 (amorphous) (41) l (CaM¢)O-(KsNaa)O-Al.Oa-8i0: (0.0 constant) (0.2:lt4 0.2:l:4 0.37:1:2 1:1:4 300°C M30 Alkali solution Kaolinite (>) Smectite ................ Smectite Smectite (<) 300°C CaO Alkali solution Kaolinita (>) Smectite ................ 7 Smectite (<) 300°C Naio Alkali solution Kaolinito (>) Smectite ................ Analcims Smectite (<) (+ kaolinite) Alkali solution Kaolinite (>) Smectite Mica K feldspar‘ + 7 Smectite (<) ('i- kaolinite) 300°C x.o Acid solution Kaolinits Kaolinits Kaolinits Kaolinite Alkali solution ................................... Mica 400°C K20 Acid solution ................................... Pyrophyllits + boehmits and kaolinite ‘ Ia puts condition at AlaOu:68iOc. Table 4 26 Another hydrothermal process uses chemical reagents to convert crystalline minerals, most commonly feldspar, to other minerals. Feldspar is a three-dimensional alumino- silicate consisting of A104 and SiO4 tetrahedra. It has pore openings large enough to exchange K, Na, Ca, and Ba cations. It is ground to a fine powder, and then can be converted to kaolinite, pyrophyllite, muscovite, and/or boehmite with Al(OH)3, 810 and KCl under acidic condi- 2. tions. The specific product depends on the specific re- action conditions(2). Clay minerals can be synthesized from oxides and hydrox- ides at more moderate temperatures and pressures but at the expense of substantial periods of time. A sodium silicate and sodium aluminate solution leached with mag- nesium chloride, washed, and then sealed in a reaction flask for four years under ambient temperature and pressure produced a smectite-like product(2). Magadiite, the layered silicic acid mentioned previously, can be synthesized by sealing silica gel, sodium hydroxide, and water in glass ampules and heating at 100°C for four weeks(22). Kaolinite can be synthesized by mixing very dilute silicate or aluminosilicate solutions with dilute magnesium and aluminum salt solutions. This composite mixture is slowly dispersed in distilled water from which the clay <2rystallizes. Increasing the concentration of electro— lytes and the pH leads to smectite formation(5). Imogolite, 27 a tubular silicate, is synthesized from dilute alumino— silicate solutions which contains a 2:1 molar ratio of alumina to silica. The solution is refluxed for more than one day at an initial pH of 3.5. Again the low pH optimizes the formation of a gibbsite sheet upon which the silica tetrahedra condense. Imogolite is unique in that it con- sists of the basal surface of single silica tetrahedra condensing on the aluminum sheet. The apical hydroxide group points away from the layer. This steric arrangement forces the quasi-kaolinite structure to curl, giving a HO-Si-O-Al-OH arrangement of planes of atoms from the in- side of the tube to the outside(27). Dilute aqueous solutions of SiOZ-MgO-LiZO at one atmos- phere and reflux temperatures form hectorite. Use of alkali metal fluorides in place of hydroxides increases the rate of crystallization. Smectite-like materials are produced by boiling magnesium silicate gels witthOH, Ca(OH)2, or NaOH. With high concentrations of KOH, mica is formed. A less popular synthetic method, the electrolysis of silica or mixtures or silica and aluminum, produces lzl clay minerals. The product is anode dependent with a magnesium anode yielding antigorite-like minerals and an aluminum anode kaolinite-like minerals. This method yields products which are always deficient in silica(2). Another method to be mentioned is that of the trans- formation of minerals to other forms at moderate temperatures 28 and pressures. These reactions depend on the cation in the exchange site: K+ or NH + Smectite 44§Illite 110 C M92+ Smectite ————§ Chlorite Vermiculite Al salt Sme cti te —————¢ Chlori te H+, CaCl2 Montmorillonite 0 Kaolinite (2) . Na3AlO3 Al(NO3)3 1.5. MAS NMR of Clays While x—ray diffraction has been and still is the pri- mary tool for mineral investigations, several techniques Ihave been used to study clays. Wet chemical analysis gives information regarding chemical composition and cation and lanion exchange capacities. Surface area, thermal analysis, .infrared and raman Spectroscopy are important methods of :identification and characterization of silicates(2,5,10, 29 13,14). A new technique that recently has been used for both qualitative and quantitative investigations of the structure of silicates is magic angle spinning nuclear magnetic resonance (mas nmr)(28,29,30,31). Mas nmr provides spacially and temporally averaged in- formation on chemical environments. The peak sharpness reflects the degree of crystallinity and the total peak intensity helps with assigning point groups. Most pertinent to this dissertation is the information this technique can provide regarding Si and Al environments in aluminosili- cates(29). The chemical environments, coordination, and structure of silicates and aluminosilicates have been in- vestigated with 29Si mas nmr. A barrier with the latter is the lack of solution analogues for reference. Many solution and solid silicate samples have been examined and the chemical shifts extended to aluminosilicate studies. The range of chemical shifts, 5, in silicates is broad, ~60 to -120 ppm and the shifts are dependent on the extent of condensation of the SiO tetrahedra. The observed shifts lof solid and solutions are compared in Figure 7a. The (2x is a tetrahedrally coordinated Si atom with x tetrahedral species coordinated to it through oxygen bonds. Single tetrahedra, 00, have an observed 6 between -66 and -74 ppm; Chain ends or double tetrahedra, Q1, 5 = -78 to -84 ppm; lfliddle chain groups, 02, 6 = —86 to -88 ppm; chain branches (Dr layered, Q3, 6 = -97 to —99 ppm; and three dimensional 30 Figure 7. (a) Comparison and ranges of Si chemical shifts with different degrees of condensation of the silica tetrahedra. (b) The upfield chemical shift imposed by the substitution of Al for Si. 31 o" o‘ o2 o3 o‘ Silicate tations ‘ — a) in solution ' Solid silicates - -7O -80 -90 - Im -IIO ppmlIHSl Al Al Al Al Si O O O O O b) AIOSiOAI AIOSiOSl AlOSiOSi SlOSiOSI SlOSlOSl O 0 O O 0 Al' Al SI SI Si Si(4Al) Si(3Al) Si(2AI) Si(lAl) Si(OAl) 4:0 3:1 2:2 1:3 0:4 [L 1 1| Si(OAI) r:——-fi J Si(1AI) L 3 1 Si(2Al) Si(3Al): Si(4AI): l 1 ’ 1 1 1 1 1 l -80 -90 -100 -110 ’ +——6("Si) Figure 7 32 silicates, Q4, 6 = ~108 to —111 ppm. As the degree of con- densation increases, the diamagnetic shielding increases and moves the observed chemical shift downfield with respect to the TMS reference. Table 5 lists some specific examples of solid silicates(30). Substitution of aluminum atoms for silicon atoms in solids paramagnetically deshields the silicon and the chemdcal shift moves upfield towards the TMS reference. Figure 7b illustrates the shift dependence on Al substitu— tion. Partial overlap of chemical shifts are due to the specific structure of the lattice and nature of the cation. Additional terminology must be introduced. QX(nA1) is a Si tetrahedra site where x is again the extent of conden- sation and n is the number of tetrahedrally coordinated Al ions in the second coordination sphere of the Si. The difference between x and n is assumed to be Si tetra- hedra. The chemical shifts resulting from A1 substitution are peculiar to the solid state and have no solution analoguesl31,32). Tables 5 and 6 give chemical shifts for some layered- Q3(nA1) and tecto-silicates Q4(nA1). Note that it is possible to distinguish between the classes of clays by the chemical shifts. The difference of kaolinite(1:1), -91.5, and pyrophyllite (2:1), -95.0, is attributed to deshielding by H bonding between the sheets of the kaolinite. The paramagnetic shift pyrophyllite (di- octahedral) exhibits vs talc (trioctahedral), -98.1 ppm, 33 08. u 000.0. 3.3805 30.. 0505.00.83 .00 4.0:: .3... .00.... 32. .00 You: 000.800 9.0000 1.00m. .2 0m. mdol 000.50.. 03000 .20 .6. AZ .02. 23: .93 Es. 1.0.0. 5.0.0... 08: 0.2.: 0:... 2.0.8 3.028.. .20 0.0.1.0000 ms..- 520 0.9... $528... 10.0.02 :0 3;: 0.0.0 205. 3.02.2.8... . .0 .0. .0 new: 020 is. 9.0%.... 10.0. .05. .0 new: £20 29... 3.0.55.0... 000.5810 3.”: 52220 1.0.0. .2 .0. 0.0.: 2.0.0.. ..0 .0. .20. .0 0:: 2.38 32555... 0.0.10.0. 5.0.50 w «T 20:.” 60:60:06. 02' on...“ 0.2.20 0;- 905. 0.036.002 v.8- ”as. .00 .22 0003 3a 0005 0005 52.058 _ 00000.03 09000 00305920 M02002.» 200.0. 00a 0300 09:00:00.5: .0 .0 .0 .0 .0 .o 25 .m0u0oflaflw UHHOm 0H .000. AH.mUMflnm H00HE000 em .m 0Hn0a mm Table 6. a) b) (a) 29 cates relative to TMS. (ppm) in tectosilicates 34 (b) 29 Si chemical shifts (ppm) in layered sili- Si chemical shifts relative to TMS. Silicate _6. ppm from TMS Talc Mg,|$i,o,,1(ou), 98.1 Pyrophyllite Al, [Si,0, , I (OH), 91.5; 95,0 Endelite A|,[Si.0|,)(OH),-~1H,O 93.1 immHMe ALHLQ,HOW. m5 Muscovite KAI, wsa,o, ,1 (OH), 84.6; 86.7 Margarita cm, (AI, s&,o, ,1 (on), 75.5 Mineral i T —65i, ppm from TMS' Al Q‘HAH Q‘UAH Qfizu) QWIM) QWOM) Morthitc Caw,5i,o,1 1 83.1 Thomsoru'xe NaCa, m, 9,0, ,1.6H,o 1 83.5 Ne pheline KNa, [AISiO, |. 1 84,8 Natrolne Na,[AI,Si,O,.]-2H,0 1.5 87.7T 95.4 Gmelim’te Na,[Al,Si”O”l-24H,0 2 92,01 97.2" 102.5 Chabasite cum. sa,o,.1-13H,o 2 94.01 99.4" 104.8 110* A wire NalAlSi,O.] 3 92.5 96.7 1mg Hannounne BylALSh,0,,L12H,O 3 9s 9&6T 1081 1m5 Heulandite Ca,(A|.Si,.O.,1»24H,o 3.5 95 99M nos 1%; Qinopmome Na,K.[AI.Si,.O,.|-24H,0 5 100.6 106.97 (NJ a-Quartz SiOI l07.4 a-Crisrobalitc SiO, l09.9 ”Q“(nAl) designates an SiO. tetrahedron connected to nAlOn tetrahedra in the zeolite framework. *Most intense signal. :Ueak signal. 35 demonstrates the perturbation of the octahedral aluminum on the silicon environment. Clay minerals have proven to be useful as supports for many cationic and neutral species, particularly catalytic- ally active species(9). As more information about the structure and properties of clays is obtained, the inter- action between specific host clays and guest intercalants can be better tuned to design new families of heterogeneous catalysts. The objectives of this dissertation were to obtain information about some of the structural properties of layered silicate clays which influence the intercalation phenomenon. A summary of the objectives if; provided below. A more detailed discussion can be found in the individual chapters. The major objective of the first study was to deter- mine if the clay layer responded to two intercalants of different size. Of particular interest was the rigidity of the clay layer and comparing its behavior to that of graphite(37). The object of the second study was to determine if the Al in the tetrahedral sheets of tetrahedrally charged clay minerals exhibited ordering of any kind. Since the aluminum is the source of the negative charge in these materials, how the aluminum ions are arranged determines how the charge is ordered. This in turn influences the properties of the .innerlayer cations. 36 As a third study, more information about the silicon environments in layered silicic acids was sought using 29$i mas nmr. X-ray crystallography has been used to in- vestigate the structures of these materials(4S-48) and 2981 mas nmr was used to provide information to supplement the x-ray data. These materials are of interest due to their diverse intercalation chemistry. PHYSICAL METHODS 2.1. Cation Exchange Capacity Measurements (CEC) A Buchi Kjeldahl apparatus in the Rock Lab at Schlum— berger-Doll Research was used to measure cation exchange capacities. Components of the system included a Buchi 342 controller, a 322 distillation unit and a Metrohm titrating unit. The Jackson method of semi-micro-Kjeldahl distillation of ammonium exchange-ion saturated clay was followed(49). Approximately one gram of dried sample was dispersed in 40 ml of deionized water in a centrifuge tube. The tube was heated in boiling water, centrifuged to remove the water, and then washed successively with l N sodium acetate, 1 N ammonium acetate, boiled then cooled deionized water, and reagent grade denatured alcohol. The purpose of this procedure was to saturate the clay with ammonium ions and to remove any other cations. The receiv- ing beaker of the Kjeldal apparatus was loaded with boric acid and distilled water; the reaction flask was loaded with the sanple, sodium chloride, zinc boiling chips, and distilled water; concentrated ammonium hydroxide was placed in the reservoir. The ammonium ion was steam stripped from the clay by distillation into the receiving beaker 37 38 and then the unreacted boric acid was potentiometrically back-titrated with the ammonium hydroxide in the reservoir by using a pH electrode. 2.2. Chemical Analysis 2.2.1. Atomic Emission Spectroscopy (AE) A double-beam Perkin—Elmer S60 atomic absorption spec- trophotometer in the Rock Lab at Schlumberger-Doll Research was used to measure Li via atomic emission (A = 670.78 nm). Dilute solutions of lithium carbonate were used to calibrate the instrument. NBS 610 and USGS GZ were used as reference clays. The samples were digested on a steambath over night in a mixture of hydrofluoric, sulfuric, and nitric acids in a molar ratio of 5.2:1.5:0.3 respectively. The solutions were heated until no acid or 803 fumes were released. A few drops of a perchloric and nitric acid solution was added to destroy any organic matter present. Concentrated nitric acid and hydrazine sulfate was added to render any precipitated MnO2 soluble by reducing the Mn(IV) to Mn(III). More details can be found in Shapiro gt §l° 1962(50). 2.2.2. Inductively Coupled Plasma Emission Spectroscopy (ICP) Elemental analysis of various clay materials was carried out at the inorganic laboratory of the Department of 39 Toxicology, Michigan State University, with a Jarrell-Ash 955 Atom-Comp spectrometer. J. T. Baker instra-analyzed grade standards were used for the analysis of Si, Al, Mg, Fe, K, Ca, and Na. NBS plastic clay 98a served as a clay standard. Clay samples (0.05 g) were mixed with lithium borate (0.3 9 Gold Label, Aldrich) in preignited graphite fusion crucibles and fused at 1000°C for 12 minutes. The resultant glass was transferred to either 30 ml of 3%- nitric acid or an aqua regia acid solution of 15 ml 15% hydrochloric and 15 ml of 3% nitric acid in a 300-ml Tef- lon beaker. The solution was stirred for 10 minutes, or until the glass was completely dissolved, then diluted to 100 ml with deionized water. 2.2.3. Neutron Activation Analysis (NAA) Neutron activation analysis was used to measure Rb/Cs ratios. A Mark-I Triga reactor in the Engineering Depart- ment of Michigan State University was used to irradiate the samples. Detection of the gamma radiation from the irradiated samples involved the following components: 1. Lithiumrdrifted germanium [(Ge(Li)] detector 2. Traycor Northern-TN1705-Pulse Height Analyzer (PHA) 3. ORTEC 4590-5 Kev Bias Supply 4. ORTEC 451 Spectroscopy Amplifier 5. Hewlett Packard 1740A Oscillosc0pe. 40 The dead time (DT), which was a function of sample position- ing, was less than 10%, often close to 3%. The kev channels were calibrated using Cs (y = 661.63 kev) and Co (y = 1173 kev and 1332.39 kev) standards. Pure Rb and pure Cs ex— changed clays were used as reference materials for the other samples investigated. The reference clays and samples to be analyzed (between 10 and 50 mg) were sealed individually in quartz tubes and irradiated for 10 minutes at 50 kw with a neutron flux of 2 x 1011 n/cm-Sec. 2.2.4. X-Ray Fluorescence (XRF) Another method used to measure the Rb/Cs ratios in the vermiculites studied was x-ray fluorescence. A KEVEX 0700 x-ray fluorimeter was used with a XES Control System. Data were processed on a KEVEX 7000 minicomputer. A Gd target (k0 = 40 kev) was used. Aquisition time was 1000 seconds. Samples of vermiculite measured with neutron activation were used as references. The samples of ver- miculite were prepared by air-drying three mls of a 1% slurry on one inch by three inch glass microscope slides or one m1 of solution on 15 mm by 18 mm silica plates. 2.3. High Pressure Reactor The smectite synthesis, to be described in more detail later, required high pressures and temperatures. A high 41 pressure, 40 m1, microreactor from LECO Corporation/TEM- PRESS Division, model number MRA-112R, was used in the syntheses which required pressures between 5000 and 45000 psi. NEVER-SEEZ, an anti-seize and lubricating compound, was applied to the threads before the reactor was sealed. The autoclave was a LECO product model number QF-lA-214 designed for temperatures up to 1200°C. For more details of the synthesis, see Chapter 4. 2.4. Magic Angle Spinning Nuclear Magnetic Resonance A Bruker Am—400 MHz spectrometer was used to obtain 295i and 27Al mas nmr spectra at 79.5 MHz and 104.3 MHz, respectively. The spinning frequencies were 4-5 KHz. The pulse width (PW), relaxation delay (RD), and number of scans (NS) are provided for each spectrum. The Si chemical shifts are relative to TMS. The field strengths were cali- brated with phrophyllite as an external reference with a chemical shift of ~95 ppm(30). The A1 chemical shifts are relative to A1C13-6H20 at 0 ppm. The 293i mas nmr spectra at 35.8 MHz were obtained on Michigan State University's Bruker WH-180. The WH-lBO Spectrometer was equipped with Nicolet NTCFT-1180 computer software and a Doty solid state probe operated at the magic angle of 55°44'. Spinning frequencies were between 2 and 3.5 KHz. The PW, RD, and NS are again provided on each spectra. The Si field 42 strengths were calibrated by using talc with a chemical shift of -98.1 ppm relative to TMS(30). 2.5. NMR Line Shape Analysis Nmr spectra were deconvoluted with a Digital PDP-ll minicomputer system. The data were digitized with the program DDEXEC (P. Sorenson, Mayo Clinic, Rochester MN) and loaded directly into GAUFCN (T. V. Atkinson, Michigan State University), a gaussian line fitting program using a least squares fit approximation. The final spectra were printed using MULPLT (T. V. Atkinson, Michigan State Uni- versity) on a Hewlett Packer 7475A plotter. 2.6. Raman Spectroscopy Continuous wave Raman spectra were acquired at Schlum— berger-Doll Research using a 5145 A argon laser source, a photon counting detection system, and a triple mono- chromator. The data collection system was interfaced to a MINC minicomputer. The positions of the Raman bands of interest were determined with high accuracy by calibrat- ing their shifts against the laser tube fluorescence lines allowed to leak through the spectrometer at 77.13 and 117.14 cm-l. Slurries of the materials to be investigated were air dried on 15 mm by 18 mm silica plates. The pre- pared samples were allowed to fluoresce in the laser beam 43 from one to two hours to eliminate interference from ad- sorbed water before the spectra were acquired. 2.7. X-Ray Powder Diffraction (XRD) For routine measurements, either a Philips or Siemens diffractometer was used with nickel-filtered Cu Ka radia- tion (A = 1.5405 A). The Bragg angle 20 peak positions were converted to d-spacings by use of a standard Cu Kc radiation d-spacings 20 chart or obtained directly from the detection system on a MINC minicomputer. For more sensi— tive measurements, a computer-controlled Huber 4-circ1e diffractometer coupled to a Rigaku 12 kw rotating anode Mo Ka (A = 0.709300 A) source through a vertically-bent graphite monochromator in the Department of Physics, Michi- gan State University was used. The samples were all prepared in a similar manner. Approximately 1 ml of a 1% clay slurry was air dried on a 1 inch by 1 inch glass slide or 15 mm by 18 mm silicate plate. The mosaic spread of the samples, a function of how well oriented the clay layers are, was between 5° and 10°. SYNTHESIS AND STRUCTURAL INVESTIGATION OF MIXED METAL INTERCALATED VERMICULITES 3.1. Objectives There were several objectives to the work presented in this chapter. One was to investigate the demixing of two cations within the innerlayer of a high charge phyl- losilicate, vermiculite, with x-ray diffraction. Cations with size differences and/or with different charges, par- ticularly a metal cation with an alkyl ammonium cation, had been used previously in similar studies(ll,12). It was desired in the present investigation to examine a series of mixed beCs(l_x) vermiculites, where Rb and Cs are both monovalent and have atomic radii of 1.48 A and 1.69 A, respectively(42), and to determine how these cations behave, whether they segregate into separate phases, randomly inter- stratify, or randomly disperse within the same innerlayer gallery. Another objective was to develop a route for the syntheses of mixed-metal vermiculites where the anhydrous Rb(I) and Cs(I) cations were distributed in the same inner- layer gallery, if the traditional syntheses did not ac- complish this. Historically, the preparation of mixed innerlayer clay minerals has led to materials with the cations 44 4S completely segregated into separate phases or randomly interstratified into mixtures of layers containing one or the other cation preferentia11y(12). To date there has been no report of a preparation of a mixed innerlayer clay mineral with the cations dispersed randomly in the same innerlayer regions. Once the synthetic route was developed and a series of RbXCs(1_X) vermiculites could be prepared, it was desired to investigate how the cations of two different sizes af- fect the rigidity of the clay layers. It was of interest to compare these results with those for the analogous beCs ) graphite system. To accomplish this, the mater- (l-x ials were analyzed by x-ray diffraction and laser Raman spectroscopy. The stage one graphite systems, which have one layer of intercalant between single layers of graphite, are known to be very floppy and the basal spacing follows a composition dependence indicating that the layers sag between the cations. The sheet separation is a weighted average of the basal spacings resulting from the different size cations and is measured with x-ray diffraction. If there was any flexibility in clay layers, it would also be reflected in similar x-ray studies. Mixed innerlayer clay minerals are important systems in that they have the potential to greatly enhance clay chemistry. The ability to diSperse two different cations in the same innerlayer region would be beneficial when 46 bimetallic heterogeneous catalysts are designed. Depend- ing on the loading desired, catalytic and non-catalytic materials could both be intercalated, the latter being present as a filler or even a prop to keep the layers apart and increase the pore size available for catalytic reactions. Investigations into the rigidity of the clay layers are important for the design of molecular sieves. If the clay layers sag between the pillars or props, the pore size will be less than expected, influencing diffusion rates and adsorbate capabilities. These.materials also have potential interest as low dimensional solid alloys. Their properties, such as phase transitions and possibly two-dimensional conductivity will be topics for future research. 3.2. Experimental 3.2.1. Materials A Mg-exchanged vermiculite from Llano, Tx was obtained through the Source Clay Repository, University of Missouri. The formula unit of the vermiculite was (M90.96K0.02) (M95.66Feo.02A10.30)(815.72A12.87)020(OH)4(10)' LICl' RbCl, and CsCl from Alpha Products were also used in this study. 47 3.2.2. Methods 3.2.2.1. Li-Exchanged Vermiculite - In a 250 ml flat-bottom boiling flask was placed 26g of LiCl dissolved in 150 ml deionized water (4.1M). Powdered Mg-saturated Llano vermiculite (5.0 g) was then added to the flask. The shiny crystals of vermiculite had been previously hand separated from the original mixture of clays, then ground with a ball mill to 200 mesh (75 pm). The LiCl—vermiculite slurry was heated at reflux with vigorous stirring. After 24 hours, the vermiculite was separated by centrifuging five minutes at 6000 rpm, and the supernatant liquid was discarded. The vermiculite was redispersed in 150 ml of deionized water with a blender, which was used to shear the clay to a smaller particle size. This suspension was re- turned to the flask and another 26 g of LiCl was added. The same procedure was repeated every day for one week to completely replace the Mg(II) ions with Li(I) ions. Com- plete replacement of Mg(II) was indicated by replacement of the 14.2 A 001 diffraction peak for the Mg phase with a 12.1 A 001 reflection, indicative of the Li(I) phase. 3.2.2.2. Rb/Cs-Exchanged vermiculite — The Li(I) ions in the above vermiculite were exchanged for various ratios of Rb(I) and Cs(I) ions to obtain products with the formula Rb C (x) (l-x) vermiculite. A ten-ml fraction of 48 the Li clay suspension, 0.94 wt% clay, was centrifuged at 14000 rpm for 30 min and the supernate was discarded. To the clay was added 10 ml of the appropriate Rb(I)/Cs(I) solution. The composition of the exchange solution needed to obtain a specific composition of clay was determined by uptake isotherms. Plots of the Rb/Cs ratio in solution vs the mole fraction of Rb(I) obtained in the final product are shown in Figure 8. Table 7 lists the Specific conCen- trations used in the syntheses and the mole fraction of Rb(I) in the resulting products. After appropriate amounts of the CsCl (0.250 meq/ml) and RbCl (0.125 meq/ml) stock solutions had been added to the clay, deionized water was added to give a total volume of 10 ml. Three different exchange times were used, two hours, 18 hours, and 40 hours, and the final product was a function of reaction time, as shown in Figure 8. The total milliequivalents of Rb(I) and Cs(I) were at least five times that of Li(I) to assure complete replacement of the Li(I) ions. The Rb(I)/Cs(I)/Li(I)-vermiculite Slurries were stirred for the selected reaction time, usually 18 hours, then washed with a minimum of four times to remove the excess salts. The samples were air dried in preparation for x-ray diffrac— tion, laser Raman spectroscopy, neutron activation analysis, and x-ray fluorescence studies. The air-dried x—ray samples were partially dehydrated at 120°C for six hours and completely dehydrated at 400°C for four days. 49 Synthesis beCs1_xVermlculite 1 .0" j J 1 1 1 : *- Figure 8. 11Lo""2io lib/Cs In Soln. Uptake isotherms for the synthesis of mixed Rb RbXCs l-x) vermiculites. The mole fraction of Rb(I) in the final product, x, is plotted against the Rb/Cs ratio in solution. Reaction times are 2 hrs (O), 18 hrs (I), and 40 hrs (X). Table 7. Synthetic Conditions for RbXCsx_1Vermiculites. meq Rb meq Cs (Rb/Cs) Reaction x in Sample in sol in sol in sol Time Solid 1 0.0 2.5 ——-— 40 0.0 2 0.118 2.38 0.050 40 0.20 3 0.265 0.565 0.467 40 0.40 4 0.823 0.343 2.40 40 0.76 5 0.938 0.336 2.79 40 0.74 6 1.117 0.335 3.33 40 0.95 0.059 2.38 0.025 18 0.058 0.118 2.38 0.050 18 0.094 0.176 0.715 0.246 18 0.18 10 0.412 0.400 1.03 18 0.44 11 0.294 0.348 0.846 18 0.41 12 0.412 0.337 1.22 18 0.49 13 0.469 0.335 1.40 18 0.54 14 0.265 0.565 0.469 0.20 15 0.354 0.447 0.792 0.30 16 0.428 0.428 1 0.365 17 0.473 0.348 1.36 0.43 18 0.592 0.222 2.66 . 0.60 51 3.2.2.3. Preparation of Uptake Isotherms - To prepare beCS(l-x) vermiculites with predictable x values, the ratio of Rb(I) to Cs(I) in the exchange solution was de- termined empirically. A Li-vermiculite (Li-V) paste was acquired by centrifuging 10 ml of the 0.94 wt% slurry men- tioned previously. With a CBC of 171 meq/100 g, the amount of Li(I) present in the 94 mg of clay was 0.160 meq. To this paste was added a five-fold excess of a 1:1 meq ratio of Rb(I)/Cs(I) (Rb+ + Cs+ = 0.80 meq). The x value of the resultant product was 0.365. Following this lead, the following relationships were derived and employed: 0.4290 meq Rb(I) used/0.365 meq Rb(I) in sample 0.1600 meq Li-V used 7 35 meqin used ° x(meq Li-V used) Therefore, (7.35)(meq of Li-V used)x = meq Rb(I) required in solution where 7.35 is the proportionallity constant from the above, and x is the desired mole fraction of Rb(I) in product. 52 With a five fold excess: meq Rb(I) + mquCs(I) 0.1600 meq Li(I) = 5 and meq (Cs(I) = 0.8 - meq Rb(I). This equation was used to project the composition of mixed Rb/Cs-vermiculites. Agreement between observed and pre- dicted x values was maintained for x i 0.6. For the syn— theses of products with higher x values the reactant ratios were predicted by extrapolation of the observed curve at 18 hours. 3.3. Results and Discussion 3.3.1. Properties of Llano Vermiculite Llano vermiculite was selected as the clay of choice for several reasons. First, it is a tetrahedrally charged mineral with a mica-like composition. It has a layer charge density of 2.0 meq per formula unit and, as a result, when monovalent cations occupy the innerlayer, each hexagonal cavity contains an ion. There would be no vacancies to be accounted for in the subsequent investigations. Another property that Llano vermiculite shares with true micas is that hydration of the innerlayer cation is very limited. The high charge originating in the tetrahedral sheets, as 53 well as the relatively low hydration energy of Rb(I) and Cs(I), guarantees that each anhydrous cation will be situ- ated between opposing hexagonal rings on adjacent layers. Llano vermiculite is also a very crystalline phyllo- silicate and well characterized by previous workers(5,10). It was easily obtained in a pure phase with no other detect- able crystalline phases present. In addition, the material has a low iron content which makes it better suited for laser Raman studies. 3.3.2. Proposed Mechanism for the Synthesis of Mixed- Metal Vermiculites The synthesis reported here for mixed metal vermicu- lites differs from the preparations reported in the litera- ture(ll,12,33,34). Traditional syntheses require the re- placement of the inherent Mg(II) innerlayer cation with either Rb(I) or Cs(I). The univalent cation is then selec- tively replaced by the second cation until the desired com- position is obtained. In this study, when the Mg(II) was replaced by Rb(I) and attempts were made to replace the Rb(I) with Cs(I), it was observed that the Cs(I) ion was preferred over the Rb(I) ion and that a controlled substi- tution of Cs(I) for Rb(I) was unobtainable. The replace- ment of Cs(I) by Rb(I) in a similar fashion was also not productive. This was due to the limited height of the innerlayer which would have a strong influence on the 54 diffusion of the exchanging cations. Even if the desired chemical compositions could be obtained,the literature in- dicates that innerlayer ions would demix with individual galleries preferentially containing one or the other cat- tion(12). For this study it was necessary to mix both metals within the same gallery. To circumvent these problems, a new approach to the synthesis of these mixed metal clays was needed. Upon replacement of the Mg(II) cations by Li(I), the vermiculite layers delaminate due to the large hydration sphere associatedvith the Li(I). Although the layers may not delaminate to the extent of other smectites with lower charge densities, Li-exchanged-vermiculite demonstrates extensive interlayer swelling in water (cf. Table 3). The basal spacing for Li(I)-vermiculite in aqueous suspension is orders of magnitude (d(001) > 40 A) larger than that found for Mg(II) (d(001) = 14.6 A) or Cs(I) (d(001) = 12.0 A)(l). After the sheets were exfoliated, the Li(I) was readily exchanged with Rb(I) and Cs(I) in the ratios necessary to yield a product with the desired Rb(I)/Cs(I) mole fractions. As the Li(I) was replaced by the anhydrous Rb(I) and Cs(I) cations, the clay flocculated. This trapped the Rb(I) and Cs(I) cations between the layers within the same innerlayer gallery. Figure 9 depicts the general reaction sequence. 55 .mmuaasoflfium> Axuavmuxnm mo mamm x . . cu: m on» no mumfifism UflumEmnum .m musmflm \ m «nH +nm +mo pm .A/ + u\+nm + I H +mo Ax av + +.q + . (x m m m . «A. +nm +3 +2 4 +2 + +~ z +N 2 +~ z Ham n Ax-Hv x mHHADUHzmm> an ax mHmm=Hz>m 56 As depicted in Figure 8, this ion exchange phenomenon is a kinetically controlled reaction. As the reaction time is increased, the Cs(I) preference is decreased. Several of these becs(l-x) vermiculites were syn- thesized and the x-value was determined by neutron activa- tion and/or x-ray fluorescence. Li analysis, performed on random samples by atomic emission spectroscopy, indicated that the Li content was less than 0.5 wt%. 3.3.3. Characterization by X-Ray Diffraction X-ray diffraction was used to measure the d(001) reflec- tions as a functiOn of composition. Representative dif- fraction patterns of the air-dried RbXCs(l_x)vermiculites (x = 0.06 and 0.74) are shown in Figures 10a and 10b. In these figures, the normalized intensities of the peaks are plotted against 20, a diffraction parameter defined below. Also shown in Figures 10a and 10b are the respective ver- miculite samples partially dehydrated. The x-ray patterns of the air-dried samples contained an extraneous peak at 26 = 29° corresponding to a 3.05 A. After partially dehyd- rating the samples, the intensity of this peak decreased. In the patterns shown in Figures 10a and 10b, the intensity of this peak decreased by approximately 35% and 25%, respec- tively. This reflection could have been an hkO reflection which upon heating became disordered, perhaps through stack- ing faults, thereby diminishing its intensity. An 57 Figure 10. X-ray diffraction patterns for oriented film samples of air-dried RbXCS(l_X) vermiculites for (a) x = 0.058 and (b) x = 0.74, intensity against 28. 58 moa onsmflm anhcma>rmn.mmo.onx.ama a.m~ .Emu> wo\mm I I I J (900) (£00) (900) (£00) (£00) nmh¢mn>r.mm0.alx~«ma (woo) (zoo) (too) .Emu> moxmm mu mu uom u0¢ lam no Ga 8N lam wot Ad £3 9.33m ‘¢A.0Ix.mma .Emw> mo\mm 59 L 1 (900) (1.00) (900) (v00) (20mg: (zoo) (soo) ">1 ‘tk.le ~0Nu .Emw> m0\m¢ IJIIIIIIIIJIIIIIIIIL Ga Ma 0N mm mm mm at Mt 60 alternate, and more plausible explanation, is that the peak was a result of a hydrated phase which upon heating disappeared. This explanation was supported by the reap- pearance of the peak with time. The reflection at 28 = 29 showed a greater decrease in intensity for samples for which x-ray patterns were measured immediately after de- hydration. The decrease was less obvious in samples which remained at ambient temperatures and humidity for several hours. It is known that Mg-vermiculite spontaneously re- hydrates after dehydration at temperatures of less than 300°C(l). The absence of a Li(I) or Mg(II) phase, which would be detected by x-ray diffraction, indicates the pres- ence of a unique hydrated phase associated with either Rb(I) or Cs(I), or both. In the completely dehydrated samples, the peak at 28 = 29° disappears. All the samples were remeasured on the Huber diffrac- tometer where it was possible to pinpoint the positions of the reflections very accurately. Sample diffraction pat- terns for the completely dehydrated samples are presented in Figure 11 as a plot of intensity in arbitrary units against q. In this figure, q = sine/A, where l is 0.709300 A for the Mo source and 8 is defined below. The remaining peaks were assigned to rational orders of d(001) reflec- tions as labeled in Figures 10 and 11. Several orders of reflection were observed for each sample at all three levels of hydration. 61 Figure 11. X-ray diffraction patterns(l) obtained from com- pletely dehydrated clays, arbitrary intensity units against q, where q = sin8/l. 62 HH mhdmflm 8p 09 1A) 7. AOF 00V 1 no 2 m n ... \ . . 2 4. .. 113,1 2 __ __ ..~ : = 3 cl. Q x X ow x om x ooo. «so I: fix AQOOV ... .oos. .nu. AQCOV J / a @\ ooo w .m Akoov M m Akoovx m n n 88 ..m . m a 300 .n n a. A . M M M Voou o om Amoov oow AmOOv . thaw . . . .. . .. . :.......... . 33...”. 1W AVOOV I‘eouoo no. no. on oooooooovoonoooooouHoooooo ooooo oo oooooo oooo oon‘ouooooo oHHHoooooooooo AVOOV ouoonoooomoo ooooo ooo oooooooo oom 5o AmOOv o’ooonooo oOooooooooooooooo ounwou-ooooooooo AmOOV Io oloooooooouo oo’o o oo oo ooooobooouoooo o oo oooooooooo A F OOVo oM ooooooo PIP Giza Hmdmtmmfi >szm:z_ l 2.0 3.0 4.0 5.0 6.0 11) 003-1) 63 Some conclusions about the completely dehydrated mixed- metal vermiculites were drawn from the x-ray data. First, the single series of d(001) reflections indicated that the materials were composed of a single phase. If other crystal- line phases were present, either as a result of other clay minerals or of the segregation of the Rb(I) and Cs(I) cations into separate galleries, there would have been a splitting of the diffraction peaks. Second, the narrow peaks indi- cate that the material was not interstratified although some disorder may have been present since the lines are not resolution limited. The x-ray data indicate that the ions were uniformly mixed within the same innerlayer galleries. The low intensity or even absence of an 001 reflection was rationalized through structure factor analysis. This analysis permitted determination of the expected intensities of hkl reflections on the assumption of usual x-ray dif- fraction procedures and a randomly oriented power specimen (5). On the basis of the ideal atomic radii of the atoms composing a sheet of vermiculite and the innerlayer cation, the composition of the material, and the observed basal spacing, the intensities and positions of the 001 reflec— tions for pure Rb-vermiculite and pure Cs-vermiculite were calculated. The pure phases were used to simplify the cal- culations. The calculated and observed values are in good agreement and are presented in Table 8. These results were compatible with the observed x-ray patterns of the mixed metal vermiculites. 64 Table 8. Normalized, calculated and observed intensities(1) of the hkl reflections for Rb-vermiculite, x = l, and Cs-vermiculite, x Pb-V 001 Calc. Obs. Calc. Obs. 001 21.4 19.1 0.09 0 002 0.36 0 8.71 2.3 003 54.7 54.6 2.26 14.4 004 72.6 70.1 100 100 005 100 100 85.54 89.7 006 6.08 5.6 6.56 0.97 007 7.26 5.1 0.05 1.30 008 18.1 13.1 31.27 11.7 009 0.87 m0.4 0.442 0 0010 4.46 4.0 13.88 2.52 Intensities provided by B. York, Department of Physics, Michigan State University, East Lansing, Michigan. 65 The major observed peaks may be assigned as 002 reflec- tions 003 through 0010. From these reflections the basal spacing can be determined using the Bragg equation, 2dsin8 = n1. In this equation, d is the basal spacing, in units the same as lambda, 28 is the angle between the transmitted x-ray beam and the detector, n is the order of the reflec- tion and A is the wavelength of the x-ray source. A plot of sin8 vs n gives a straight line with a slope, l/2d,-from which d.may be obtained. Figure 12 is a plot of the normalized basal spacing as a function of mole fraction of Rb(I) in the vermiculite. The right ordinate represents the d(001) reflection nor- malized to the extreme x values according to the relation— ship: d(obs) - d(Rb) d(norm)== d(Cs) _ d(Rb) where d(Rb), x 1, is 10.23 X_and d(CS), x = 0, is 10.57 A. Normalized d spacings were required in order to compare this system with the analogous Rb/Cs graphite systems, beCs(l_x)C8, which are also plotted in F1gure 12(35). The left ordinate depicts Raman data which will be dis- cussed later. 66 .13 x m> .720 .035 cmsmm mm :03 mm An: mUAxICmuxnm now can 23 mmuflasogumtr Axuavmuxnm MOM x gunfimmm nonfiommm Hmmmn chHHmEHoz .NH muzmflm 5.5: u 0.0 7. s 5 A 3‘2 ____ 0. 4| _ 4’ .2 .3 .4 .5 .6 .7 .8 .9 1.0 1. 0 5 5 O . 1 _ O. 6 O 1 0 5 7 6. 0 0 1. «I A? EB ._.n.=Iw. Zm._._wzm._.2_ 24.5.5”. 112.5 125 100 .5 RAMAN SHIFT (cm“) 87 75 72 The composition dependence of the shift is plotted in Figure 12, where the left ordinate is the observed frequency. Again, a step function is observed with an inflection point at x m 0.4, which correlates well with the x-ray data. This further supports the hypothesis that the Si tetra- hedra rotate in a coordinated fashion to encapsulate their respective cations. To verify that the observed dependence was indeed a result of a change in composition and not some other param- eter associated with all the Raman modes, another Raman 1 and 220 cm.1 was measured as active band between 170 cm- a function of composition. This band showed no orderly dependence on composition. As the mole fraction of Rb(I) decreases, a shift toward higher frequencies occurs in spite of the greater mass of the Cs(I) cation. This shift results because in this tor- sional A" mode, the cation is at rest and the frequency is therefore independent of its mass. The frequency is in- stead dependent on how far the cation can key into the hexagonal cavity. If 6 is a measure of this encapsulation, then 0 (t + 2r) - d where t is the clay layer thickness, r is the ionic radius, and d is the observed basal spacing. For Rb—vermiculite 73 t 9.34 A, r - l H .48 A, and d 10.23 A. For Cs-vermiculite, t = 9.34 A, r = 1.69 A, and d 10.57 A. The respective 0's yield 0(Rb) = 2.07 A and 0(Cs) = 2.15 A. Cs(I) keys further into the sheet increasing the effective force constant of interaction with the neighboring oxygen atoms and in- creasing the energy of vibration for the lower x values. In conclusion, it can be stated that the clay layer is not a longitudinally rigid two-dimensional structure that is indifferent to the innerlayer moities. Instead the tetra- hedral sheet has detectable flexibility in the inplane direc— tion. The clay layers probably are very rigid in the trans- verse direction, and do not sag between the innerlayer cations as is the case with stage one intercalated graphite. This work demonstrates that innerlayer cations affect the crystal- lographic nature of the clay. It shows that the twist angle is dependent not only on octahedral/tetrahedral layer condensation but also on how well the cations can key into the hexagonal cavities. This type of investigation can give information about how rigid clay layers are with materials that have vacant hexagonal sites. Specifically, investigations of this nature would apply to clays with large robust cations proping the sheets apart. There is relatively more dis— tance between these "pillars" than in the RbXCs(l_X)ver- miculites used in the present study and the clay layers may sag or bow between the larger cations. These pillared 74 clays have innerlamellar regions available for the diffusion of sorbates and any sagging of the layers would decrease the space available for this diffusion. Sagging of the layers would also limit the regularity of the pore size of the material, an important parameter in the design of hetero- geneous catalysts. INVESTIGATION OF CHARGE DISTRIBUTION IN CLAY MINERALS USING MAS NMR 4.1. Objectives The objective of this study was to investigate the charge distribution in 2:1 phyllosilicates, specifically 29 27 tetrahedrally charged clays, with Si and Al magic angle spinning (mas) nmr. The charge distribution is of interest because it determines the properties of the countercations in the innerlayers. This is an important parameter in clay chemistry in that the charge distribution will deter- mine how cationic catalysts will disperse on the surfaces and in the innerlayer of the clay minerals. Charge distribu- tion is also a factor that must be considered in the design of molecular sieves or shape selective heterogeneous cata- lysts since it influences how uniformly the large, pillaring cations will disseminate in the innerlayer. As mentioned in the introduction, the layer charge in a tetrahedrally charged clay is a result of the substitu- tion of Al(III) for Si(IV) in the tetrahedral sheets. This substitution creates a charge deficiency, hence the tetrahedral site occupied by Al(III) is also the site of a negative charge. By examining the distribution of Al(III) 75 76 ions, information about the charge distribution is obtained. The mas nmr method, which has been highly successful for establishing the aluminum site distributions in zeolites (42,43), was used to investigate the aluminum distribution in both synthetic and naturally occurring tetrahedrally charged saponites, muscovites, and.phlogopites. The nor- malized integrated intensities of the signals obtained from 2981 mas nmr gave quantitative data pertaining to the silicon environments and were directly related to the or- dering of the Al(III) sites within the tetrahedral sheet. 4.2. Experimental 4.2.1. Materials - Some of the clays used in this study were synthesized by using Mg(NO3)2-6H20, Al(NO3)3-9HZO, and Na2C03 from J. T. Baker Chemicals and Si(OCHZCH3)4 from Aldrich Chemicals. Other synthetic clays were provided by J. L. Robert, Centre de Resherches sur la Synthese et Chimie de Mineraux Orleans, Cedex 45055. The natural clay minerals examined were muscovite, provided by M. Mortland, Department of Soil Science, Michigan State University and saponite, obtained from Industrial Mineral ventures, Inc., Las Vegas, Nevada. 77 4.2.2. Synthesis of Saponites - The samples used in this investigation were the tetrahedrally charged clay minerals, saponite, Na(x_y) Mg(6_y)Aly (Si(8_x)Alx)020(OH)4, and the micas, muscovite K2 Al4 (Si6A12)OZO(OH)4, and phlogopite K2 M96 (Si6A12)020(OH)4. A series of saponites with different charge densities were synthesized by the hydrothermal crystallization of gels(20,21). The desired stoichiometric amounts of Mg(NO3)2- 6H20, Al(NO3)3-9H20, and Na2C03 needed to prepare 50 g of product were dissolved in 500 ml of deionized water and added to a stoichiometric amount of Si(OCHZCH3)4 dissolved in an equal volume of CH3CH20H. The molar compositions of the gels are listed in Table 9. The resultant gel was diluted to approximately 750 ml with deionized H20. Dilute NH4OH was added to the vigorously stirred gel until a pH of 11 was achieved. The gel was dried in a rotovap to a white, crumbly powder. This powder was heated at 550°C for 24 hours to convert nitrates to oxides. Brown NO2 gas was released in the calcination process. The powder was ground with a mortar and pestle, then mixed 1:2 with de- ionized water and placed in a high pressure reaction vessel which was heated to 450°C. The internal pressure was between 10,000 and 22,000 psi as judged from the volume percent full (50% and 75%, respectively). The reaction was allowed to proceed for two weeks and then the vessel was quenched by submerging it in water. The solid reaction product was 78 Table 9. Molar Composition of Gels Before Hydrothermal Crystallization. Sample Na Mg A1 Si 1 0.66 5.6 1.46 6.94 2 0.66 5.66 1.46 6.94 3 1.6 5.6 2.4 6.0 79 removed from the vessel and suspended in water. The frac- tion that remained in suspension at least 24 hours was separated and air-dried for subsequent characterization by x-ray diffractiona elemental analysis, and cation ex- change. 4.2.3. Purification of Natural Saponite - The natural saponite used in this study contains impurities which are removed in a standard clay preparation procedure. CaCO3 can act as a cementing agent and was removed, along with soluble salts, by the following procedure, based on 5 g of the pristine saponite: 1. Add 5 m1 of 1N sodium acetate, buffered to pH 5 with acetic acid, and then bring the clay into suSpension by stirring. 2. Digest the sample for 30 minutes at a temperature between 70 and 80°C. 3. Centrifuge the suspension and discard the supernate. Free (non-lattice) iron oxides also act as cementing agents and, more pertinent to this study, will interfere greatly with the nmr measurements. The following steps were followed to remove the iron by (l) chelating the iron with citrate ions (C6H807-3) to prevent the precipitation of 80 FeS and 2) reducing the Fe(III) to Fe(II) which can then be washed away. 1. Add 40 ml of 0.3 N Na-citrate and 5 ml of l N NaHCOB. 2. Heat the suspension carefully to 75—80°C and add 1 9 Na 8204 with stirring. Heating beyond 80°C can cause precipitation of black FeS. 3. Cool and centrifuge the suSpension and discard the supernate. To obtain the fraction of saponite with a partical size ose m.omo mmmse lav 82 mimovam~.on lace N.¢H Hem Hem elmocomolms Haoomsz ea 0 eimovomoime.oasam.semcAHmH.omm~m.oasmo.mmzcom.omz muesomsm me eimovomolmm.maame.memv e Ame.oeeeo.ommma.oaaee.mmzceo.omoma.omz45.ax mnemomoflze NH slmooomoime.aflmmm.eemc e A3.082350.208233.omomoeszmméx mue>oomsz as o AmoeflmocomoimaseemvAwesome mnemomoasm OH s exmovomoimm.aasmo eemclmo.oaewm mmzvem as muemomoass m n efimovomoxmo.~Hoomsz m mousom coauflmomEou HHmU was: mmau mHmEmm nee magma 84 along with the calculated and observed cation exchange capacities (CEC). The observed CEC's were obtained by the Kjeldahl method as described in Chapter 2. Although the chemical compositions of gels before the crystalliza- tion process were the same in samples 1 and 2 (cf. Table 9), the compositions of the resultant products are dif- ferent. The compositions of the synthetic clays are not solely dependent on the starting materials. The hydrolysis of the salts and the silicate at pH 11 is a very crucial step in the syntheses as it governs the ensuing condensations among the ions. During the hydrolysis, the temperature of the gel should be maintained at room temperature, approxi— mately 23°C. The impurity phases present in the samples were attrib- uted to incomplete hydrolysis. These impurity phases were responsible for the deviations between the observed and calculated CEC's in samples 1, 2 and 3. The observed CEC's are much lower than the expected values, indicative of a Chloritic innerlayer. This discrepancy is a common phen- omenon in highly charged 2:1 clay minerals as innerlayer hydroxide layers are often present(10). While the observed basal spacings of the saponites are within the range of smectite materials, basal reflections of Chlorite crystals coincide with those of smectites and can only be differen- tiated through a series of chemical and heat treatments(10). Attempts to remove the impurity phases by acid-leaching 85 were unsuccessful and resulted in the destruction of the crystalline phase. Other clay minerals examined in this study are listed in Table 10b. Samples 8-10 are synthetic micas while samples ll-l4 are naturally occurring clay minerals. 4.3.2. Magic Angle Spinning NMR 29Si mas nmr and All the materials were examined by typical spectra with their deconvoluted components are shown in Figure 14 a—f. The non-equivalent chemical shifts are attributed primarily to different Si sites arising from differences in the occupancy of the three neighboring tetra- hedral cavities. The shifts from high to low field can be assigned on the basis of relative intensities and analogous aluminosilicate work(31) to Si surrounded by no A1 ions (Q3(OA1)), one Al ion (Q3(1Al)), two Al ions (03(2A1)) and three Al ions (Q3(3Al)), respectively. The dotted line in the top half of Figures 14 a-f represents the actual data. Superimposed upon the data is the summation of the indepen- dent gaussian lines that are displayed in the bottom half of the figures. Depending on the number of silicon environ- ments observed, the spectra were fit to two, three, or four gaussian lines by a least-squares method. Table 11 lists the chemical shifts relative to TMS and the normalized intensities of the signals obtained from this curve-resolving Figure 14. 86 2gsi mas nmr spectra(l) and deconvoluted components for samples 2(a), 3(b), 8(c), 9(d), 11(e), and 12(f). PW, RD, and NS are provided for each spectra. from 180 MHz nmr spectrometer from 400 MHz nmr spectrometer from Sanz, J., Serratosa, J. M. J. Am. Chem. Soc. 1984, 106, 4790(44). 0.0 0.0 a)Sanp1e 2 PW 2.0113 . RD 3.00 NS 4548 87 b)Samp1e 3 PW. 3:01): RD 1.0“! NS 9610 Figure 14 a,b 0.0 0.0 c)Sanplo 8 PW 0.25uo RD 200ns NS 19100 88 Figure 14 c,d 0.0 0.0 89 o)Sanple ll PW’NA RD II NS II f)Sanple 12 PW NA RDII NS " Figure 14 e,f 9O o~o.o uuo.o no.o ma emn.o owo.o cmu.o nn~.o nmN.o «N.o o~.o co.o mmm.o n-.o omq~.o NH cm mabflmmom may mo coflumuumsaaw ofiumEmnom .ma musmflm Q d d. d. . «was .1. p d. d d .... 95 viewed from the top of the basal surface. The dark circles represent Al(III) and the Open circles represent Si(IV). There are bridging oxygen atoms between each of the tetra- hedra which are omitted in this figure for clarity. Site a is a Si atom with no near neighbor Al atoms, Q3(OA1), site b is a Q3(1Al) environment, site c is a 03(2Al) environment, site d is a 03(3Al), and site e represents two adjacent tetrahedral sites occupied by Al. Loewenstein's rule(47) states that during the iso- morphous substitution of Si(III) by Al(III) the Al avoids substituting into adjacent tetrahedral sites, site (e) in Figure 16. This avoidance is a result of electrostatic repulsions between these sites of charge deficiency. Loewen- stein's rule has been applied to several zeolite studies and was upheld without exception(31,43,44). The number of adjacent tetrahedral sites, X1 (Figure 16, site e), occupied by aluminum atoms is given by: 7X- 2 M1 (1) where M1 is a weighted average of Si(lAl) sites, M1 is determined from the relationship: M=I l l + 212 + 313 (2) where the In values are the normalized intensities for the 96 Si(nAl) signals. In Equation (1), x is the fraction of tetrahedral sites occupied by aluminum. The value of x can be determined from the chemical composition obtained from chemical analysis directly: _ number of Al in tetrahedral sheet/unit cell x(ca) _ 8 possible tetrahedral sites (3) Table 4 lists the x1 and x values for the samples studied. It also lists the x values calculated from the nmr data assuming Loewenstein's rule is obeyed, i.e., that x1 = 0: X(nmr) = M 3 (4) As the x1 value is a relatively small number in each sample, the difference between the x and x expressed as (ca) (nmr)’ the %A in Table 12, was considered a more accurate view of how well Loewenstein's rule was obeyed. Comparison of the x(ca) and x values indicates (nmr) that there is a discrepancy in samples 1, 2, 5, 9 and 11 between the A1 content of the tetrahedral sheet determined by nmr and chemical analysis. These discrepancies are attributed to impurity phases in the samples and are the result of how the unit cell compositions are formulated. The structural compositions are arrived at by assuming that all the silicon present is located in the tetrahedral 97 Table 12. Fraction of tetrahedral sites occupied by Al (x) and number of Al-Al near neighbor tetrahedral sites (x1). Sample xca xnmr %A(1) x1 1 0.151 0.103 46 0.08 2 0.142 0.186 24 -0.08 3 0.255 0.269 5 -0.03 4 0.115 0.115 0 0.0 5 0.170 0.148 15 0.04 6 0.186 0.200 7 -0.03 7 0.267 0.259 3 0.02 8 0.252 0.242 4 0.02 9 0.248 0.282 12 -0.07 10 0.250 0.260 4 -0.02 11 0.210 0.180 17 0.06 12 0.290 0.279 4 0.02 %A = 03x. Xnmr x 100 98 sheet. First, the silicon is placed in the tetrahedral sites, then aluminum is added until the total number of atoms occupying tetrahedral sites is eight per unit cell. Any excess Al is placed in the octahedral sheet. In samples 1, 5, and 11, x (ca) 15 Significantly greater than x For these samples the chemical composition (nmr)' indicates that there is too much aluminum in the tetra- hedral sheet, assuming the nmr data are accurate. This may be attributed to errors in the silicon analysis which if low would mean that the aluminum assigned to the tetrahed- ral sheet would be too high and the Al which belongs in an impurity phase was placed in the tetrahedral sheet. In sam- ples 2 and 9 the fact that x(nmr) > x(ca) is indicative of a silicious impurity phase. If some of the Silicon detected by chemical analysis is an impurity phase, then the amount of silicbn assigned to the tetrahedral sheet will be too high, thereby making the structural aluminum content too low. Impurity phases in either case would lead to errors in the quantities calculated by Equations 1 and 3, hence the apparent violations of Loewenstein's rule are unfounded. A corollary to Loewenstein's rule, also based on electrostatic repulsion considerations, was proposed by Dempsey(48) and stated that the Al(III) ions, when substi- tuting into the tetrahedral Sheet, will avoid substituting into second near neighbor tetrahedral sites (cf. Figure 16, sites c and d). The number 99 X I O 1 of //Si\ sites, where X is either /0 O\\ Al A1 a silicon or an aluminum atom, was calculated as the frac- tion of these sites over the number of all possible silicon environments by: x2(obs) = (1 - x(nmr))M2 (5) where x was defined in Equation 4 and M2 is a weighted (nmr) average of Si(2Al) environments calculated by: M = I + 3I (6) The x(nmr) was used in these calculations as opposed to the x(ca) to keep the values self-consistent and to compensate for any impurity phases. The observed x2 values are listed in Table 13 (x2(obs)). A statistical value of x2, also listed in Table 13 as x2(R) was calculated by assuming that Loewenstein's rule is obeyed (i.e., x = 0) and that 1 the distribution of A1 in the tetrahedral sheet is other— wise random. The random x2 value was calculated in a fashion similar to that of Equation 5: x2(R) = (l - x )M' (7) (nmr) 2 100 M5 is a weighted value from the statistical probabilities of the occupation of the tetrahedral sites: + 3P3 (8) The statistical population of each possible silicon environ- ment (cf. Figure 16) for a random distribution subject to Loewenstein's rule was determined by: SITE PROBABILITY a po = (1-r)3 2 b P1 = 3r(l-r) 2 c P2 = 3r (l-r) _ 3 Here, r is the Al to Si ratio in the tetrahedral Sheet, (A1/Si)tetra, and was calculated by: _ x r_1____x (9) Depending on the amount of Al(III) in the tetrahedral sheet, the value of x2 will not necessarily equal zero. In the samples with Si/Al < 3, there must be some silicon 101 atoms with two or even three adjacent tetrahedra occupied by aluminum atoms. Dempsey's rule states that the number of these Sites will be minimized. Therefore, if Dempsey's rule was obeyed, x2(obs) would be less than x2(R). Again, as shown in Table 13, the x2 values are small and therefore the differences between the observed and calculated values are expressed as %A. In samples 4-7 and 10-12 Dempsey's rule is obeyed. In these samples, the aluminum ions preferentially choose sites such that they avoid having aluminum ions in second near neighbor tetrahedral Sites. This indicates a non-random distribution of aluminum ions. In samples 1—3, 8, and 9 the x2(obs) > x2(R). This in- dicates that the Al ions not only do not avoid being second near neighbors but preferentially occupy second near neigh- bor sites, a highly improbable situation. In these cases, the apparent violation of Dempsey's rule is attributed in- stead to impurity phases which contribute to the intensity of the nmr signals, particularly to the 12 and I3 compon- ents. An amorphous aluminosilicate phase would affect the nmr signals and not be detected by either x~ray diffraction or elemental analysis. It was concluded that, in the absence of impurity phases, the phyllosilicates examined in this study obeyed both Loewenstein's and Dempsey's rules. There were no adjacent tetrahedra occupied by aluminum (x w 0) and the l 102 Table 13. Fraction of Silicon atoms with two or more near neighbor tetrahedral Sites occupied by aluminum atoms (X2). Sample X2(obs) x2(R) %A(1) 1 0.13 0.03 +322 2 0 14 0.10 + 40 3 0.24 0.19 + 26 4 0.00 0.04 -100 5 0.04 0.14 - 71 6 0.08 0.12 - 33 7 0 18 0.30 - 40 8 0.22 0.16 + 38 9 0 37 0.33 + 12 10 0.24 0.27 - 11 11 0.07 0.09 - 22 12 0.22 0.32 - 31 x (obs) - x (R) %A = 2 2 x 100 x2(R) 103 number of silicon atoms with two or three near neighbor atoms was less than a purely random distribution would predict (x2(obs) < x2(R)). These results indicate short range ordering of the Al(III) in the tetrahedral sheets. This, in turn, can be translated to short range ordering of the negative charge through the clay lattice of these tetrahedrally charged clay minerals. LAYERED SILICIC ACIDS 5.1. Objectives Chapter 5 presents a 2981 mas nmr study of the silicon environments in two layered silicates, kanemite, NaHSiZOS- 3H20, magadiite, Na28114029-11H20, and their respective acid forms H28i205—0.7H20 and HZSil4029-5.4H20. Studies of layered Silicic acids by x-ray diffraction have led to some understanding of their structure but structural details of the condensed forms which have a Si/Na ratio > 2 are unknown. Layered silicic acids are of interest because of their extensive and diverse intercalation chemistry. Several studies have been made involving the intercalation of crys- talline silicic acids with neutral molecules. Possible guest molecules and their resulting separations are listed in Table 14(49). Layered silicic acids also demonstrate extensive cation exchange capabilities. They exchange a variety of organic cations, particularly long chain alkyl- + ammonium salts c H NH , c H +1N(CH3)3+, (an n 2n+1 3 n 2n+1 2n+l)é N(CH3)2, and CnH2n+lNH3 where n is even and 8 < n < 18. The ions form a bimolecular innerlayer and models for the inter- calated species are illustrated in Figure 17(50). Another interesting feature of these acids is that 104 105 Table 14. Possible guest molecules for interlamellar ad- sorption by crystalline silicic acids and layer separation (nm) (49). Group Examples Layer Separation (nm) H2SiZOS-I H2Si409 ”23114029 - 5.41120 Short chain formamide 0.33 0.17 0.24 fatty acid amides dimethyl formamide 0.49 0.36 0.55 acetamide 0.38 0.38 0.29 diethyl acetamide 0.61 0.54 0.59 Urea and urea -- .) -- 0.43 derivatives N,N-diethyl urea -- 0.52 0.53 N,N'-diethyl urea -- 0.63 0.43 S-oxides dimethylsulfoxide 0.36 0.67 0.46 N-oxides trimethylamine-N-oxide -— 0.37 0.47 pyridine-N-oxide 0.49 0.37 0.46 3-picoline-N-oxide 0.51 0.43 0.56 Aromatic pyridine 0.47 0.68 0.53 bases imidazole -- 0.61 0.32 pyrazine -- 0.35 '- quinoline -- 1.07 0.61 Alkylamines hexylamine 1.71 1.84 1.84 decylamine 2.69 2.93 2.90 benzylamine -- 1.27 -\ ’--: no reaction 106 SILICATE LAYER “1 II 8) ‘_ $UCNHE lJNER j Figure 17. Models for the interlayer structure of long chain magadiite derivatives, (a) arrangement of alkyl chains in alkylammonium magadiite; (b) arrange- ment of n-alkylpyridinium ions in alkylpyridinium magadiite(50a). 107 they have surface silanol groups similar to those found on amorphous silica. The acidity of these groups varies greatly and is dependent on the extent of condensation of the acid. Amorphous Silicas are used for a variety of applications such as adsorbents, catalysts, catalyst supports, and filters. If the intercalation chemistry similar to that of other layered materials(9) could be combined with the reactivity of silanol groups, the number of uses would in- crease dramatically. Instead of intercalated cations being bonded by electrostatic attraction alone, they may also condense with the silanol groups thereby securing them- selves to the sheets as well as cross—linking adjacent layers. Cross-linking the layers would further stabilize the structure. To fully maximize the potential uses of these systems requires a deeper understanding of the layered silicic acids. 29Si mas nmr was used in this study to investigate what ef- fect, if any, replacing the sodium ions with protons had on the Silicon environments. The effects of aging the materials was also of interest since it was noted that while fresh silicic acids in the acid form are readily inter- calated, aged samples (>3 months) are not. 108 5.2. Experimental 5.2.1. Materials Silica Gel, 60-200 mesh, obtained from MCB Chemicals, and sodium hydroxide were used in this study. 5.2.2. Synthesis of Layered Silicic Acids Kanemite was synthesized by using the method described by Beneke and Lagaly(51). A cold, nearly saturated aqueous solution of NaOH, one mole NaOH in 35 m1 of deionized H20, was slowly added to a slurry of one mole of SiO2 in 100 m1 of CH3OH. The molar ratio of Si:Na was 1:1. During this hydrolysis process, care was taken to ensure that the temperature of the resulting solution did not exceed room temperature, approximately 23°C. The solution was evaporated to dryness at 100°C and then was calcined at 700°C for 5 hours. The resulting solid NaZSiZOS' was dispersed in water for an hour and hydrolyzed to NaHSiZOS. The solid was washed twice with deionized water, then air dried. Washing is an important step as with each wash, some structural sodium may also be removed. It is difficult to remove the excess NaOH without removing some of the struc- tural Na(I). The pH is another important parameter since when the pH is below 9, HZ-kanemite begins to form(51). The acid form, stizo5 was synthesized by Slowly titrat- ing over a 24 hour period 200 m1 of a 0.5 wt% NaHSiZOS 109 slurry with 10—2 M HCl until the pH of the slurry was 1.5. This acidic slurry was stirred for an additional 24 hours. The resulting compositions were characterized by x-ray diffraction. Na-magadiite was synthesized via a method described by Lagaly, Beneke, and Weiss(50). A 9:2:75 molar mixture of SiO2 (0.09 mole), NaOH (0.02 mole), and deionized H20 (0.75 mole) was sealed in a glass ampule and heated in an oil bath at 100°C for four weeks. The product was recovered by carefully cracking the ampule. In the mother liquid the slurry had a pH of 11.5. The material was washed.with water in a manner to ensure that the pH did not fall below 9 to avoid the formation of the acid form. The air dried product was characterized by x-ray diffraction. A portion of the product was converted to the acid form.by slowly titrating over a 24 hour period 100 ml of a 1 wt% Na-magadiite Slurry with 0.01 M HCl until the pH was 2.0. This acidic solution was stirred for an additional 24 hours and the resulting product was air dried and characterized. The reverse exchange was performed by stirring 0.5 g of H—magadiite in a saturated NaCl solution for 7 days. Some of the Na-magadiite and freshly prepared H—magadiite was intercalated with long chain alkylammonium saltS(50). Enough dry clay was added to 0.1 M solutions of CH3(CH 2)16_ N(CH3)2Br, CH3(CH2)9NH2C1, or CH3(CH NH(Et)ZBr to make 2’13 110 a 1 wt% slurry. The slurry was stirred from two to six hours, and then the exchanged product was washed at least four times with deionized water. The final product was air dried on glass plates and characterized by x-ray diffrac- tion. 5.3. Results and Discussion 5.3.1. Characterization of Silicic Acids The observed and ideal Si:Na molar ratios of the syn- thetic materials are listed in Table 15. The sodium content is non-ideal in the Na-silicates, attributed to either in- complete or excess washing. In the kanemite, the Si:Na ratio is low, indicative that some of the structural sodium has been replaced with protons. In sodium magadiite, the sodium content is high, indicative of the presence of ex- cess NaOH. The water content of these materials is variable and dependent upon the relative humidity and temperature. The water is present in different forms: physically adsorbed water, water molecules bound to the external sur- faces by hydrogen bonds, interlayer water molecules, and water in the form of silanol groups on the external and internal surfaces. As with most layered clay minerals, the effects of the internal surfaces is much greater than that of the external surfaces. Therefore, the amount of 111 Table 15. Compositions and basal spacings of layered silicic acids and their sodium salts. Si:Na Sample Observed Ideal d001(A) NaHSi205 2.4:1 2:1 10.0 H281205 <0.5 wt% 0 6.0 Na25114029 6.36:1 7:1 15.5 1128114029 <0.5 wt% 0 13.2 112 external water is insignificant compared to the amount of interlayer water. Previous lH nmr studies have indicated that the interlayer water is present as silanol groups and water molecules in a molar ratio of 0.77:1 (49). The observed basal spacings are also listed in Table 15 and correlate well with literature values (50-52). 29 5.3.2. Si MAS NMR of Layered Silicates Fresh and aged samples of kanemite, magadiite, and their acid forms were investigated using 298i mas nmr. The resulting chemical shifts and normalized intensities are listed in Table 16. The chemical shifts are referenced to TMS and the intensities were obtained from the measured areas. On the basis of comparison of the observed chemical shifts to literature values for other silicates(30,3l), the peaks with 0 between —103 and ~110 ppm are assigned as Q4 environments, that is, a silicon atom with four silicon atom second near neighbors. The Shifts in the range -93 to —103 ppm are assigned to silicon atoms with three Silicon atom second near neighbors, Q3. The increased condensation of the Si tetrahedra in the Q4 environment over the 03 environment results in an increase in diamagnetic shielding as indicated by the upfield shift. Crystallographically distinct silicon environments are known to be detectable 113 Table 16. 29Si mas nmr chemical shifts(1) in ppm and nor- malized intensities of the signals. Fresh Aged(2) Sample 6(ppm) I 0(ppm) I NaHSiZO5 - 99 0.73 — 99 .65 NaHSiZO5 -110 0.27 —112 .35 H281205 -—-- ---- — 95 .14 —112 .86 Na25i14029 - 98 0.29 - 99 .35 -109 0.71 —112 .65 H28i14029 -102 0.32 ~104 .21 —113 0.68 -112 .79 Heated ~103 0.17 1128114029 -ll4 0.83 (2) Aged one year. (1) Referenced to TMS. 114 with 2981 mas nmr(53) and account for the shoulders evident in the Spectra, vida infra. 5.3.3. Kanemite The proposed zweiereinfachschicht structure for kane- mite consists of two layers of Si tetrahedra condensed to each other through the three basal oxygens(51,54). The apical hydroxide groups point in Opposing directions. Figure 18 illustrates examples of the zweiereinfach- schicht structures observed in clay minerals(54). The x- ray diffraction data support this proposed structure. The layer thickness of kanemite is 7.6 A which corresponds to the thickness of two tetrahedral Sheets which are approxi- mately 3 A each plus a layer of water in the innerlayer. 29 The Si mas nmr spectrum presented in Figure 19a also supports this structure. In NaHSiZOS, there is a major resonance at -99 ppm, indicative of a Si with three neigh- boring Silica tetrahedra. In the fresh material there is present a small Q4 compOnent at -110 ppm which was initially attributed to an impurity phase. However, when the material was examined a year later, the intensity of this peak increased significantly as Shown in Figure 19b. It appeared as though a three dimensional Q4 environment was growing into the crystalline material and was attributed to the condensation of opposing silanol groups. Figure 20 is a 298i mas nmr Spectra of aged HZSiZOS' The broad lines 115 Figure 18. Examples of zweiereinfachschicht structures ob- served in clay minerals(54). 116 Figure 18 a , b 117 Figure 18 c,d 119 Figure 19. 29Si mas(l) of (a) freshly prepared NaHSiZOS-HZO obtained on the 400 MHz spectrometer and (b) aged NaHSiZOS-HZO, obtained on the WH-180 MHz Spectrom— eter. (l) Referenced to TMS. 120 Figure 19 o)N0881205 PH Zoo ID 40 RS 1232 . -40 ' -100 -160 L b) 1125120s PH Zuo ID 3o NS 12904 __ i 1 1 1 1 1 A -20 -60 -100 7177’ -150 ppm 121 PH Zuo RD 30 us 12904 I J -00 - .1- 100 -129 4;“) fl PT“ Figure 20. 29mas nmr (l) of aged HZSiZOS obtained on the WH-180 MHz spectrometer. 122 in the kanemites were a result of the fast relaxation of the Si nuclei. This was attributed to the coupling of the protons to the silicon in the structure. It was observed that the ratio of the intensities of the Q4 and Q3, I(Q4): I(Q3) was greater in the acid form than in either the aged or fresh samples of the sodium form. + for H+ also influences the The exchange of the Na x-ray diffraction patterns, as Shown in Figure 21a and 21b. The NaHSi205 is very crystalline with at least two orders of reflections and some unassigned hkO reflections. The x-ray patterns of the acid form (Figure 21b) indicates that the system is not as crystalline Since the peaks are somewhat broader. Aging these materials did not affect the x-ray diffraction patterns. 5.3.4. Magadiite X-ray diffraction indicates that magadiite is a crys- talline material with four sheets of silicon tetrahedra. The 15.5 A basal Spacing is equal to the sum of the four tetrahedral sheets, approximately 3 A each, plus a layer of interlayer water. There are two central sheets of silicon tetrahedra with Q4 environments condensed to two outer sheets of Silicon tetrahedra with Q3 environments. The latter two need not be entire sheets in themselves, the 3 Q environments could be individual tetrahedra condensed to one of the center sheets through its three basal 123 Figure 21. X-ray diffraction patterns of (a) NaHSi O and O 2 5 124 1’ ‘- fi- .1; Asooo .oH -- NI H‘ mam onsmfim ON .3 cu _ — u _ w~ _ a Cu _- NN cw ‘— noawmzsz A. on 4)- mN -‘ 125 new ousohm on . . 2 .2 3 3 ow «N «N on on F W o— .o m: _ . _ _ F _ _ r 7 r w 1“. 7 _ . . _ _ _ _ _ q _ _ _ xsoov nowsmumxs 126 oxygens. Again, details of these structures are unknown. In magadiite, the Q3 environments also have apical hydrox— ide groups pointing in opposing directions towards the in— nerlayer regions. A representative 298i mas nmr spectra of NaZSil4029- 11H20 is shown in Figure 22a. There are two Signals ob- served. The upfield peak at -110 ppm was assigned to sili- con with a Q4 environment and the downfield peak at -98 ppm was assigned to the Silicon with a Q3 environment. The ratio I(Q4):I(Q3) was approximately 2:1. No significant dependence of the mas nmr signals on aging was observed. When the Na+ ions were exchanged for H+ ions, a change 29 in the Si mas nmr was observed. Figures 22 a-d illustrate the Spectra for (a) Na 5114029’ (b) freshly prepared 2 H28i14029, (c) heated H and (d) aged H251 23114029' 14°29° Going down the figure from b to d, the intensity of the signal for the Q4 environment increases at the expense of the Q3 environment in the acid form. This is indicative of the condensation phenomenon. The x-ray patterns of the magadiite also got broader when the sodium ions were replaced with protons, Similar to that observed for kanemite. The hkO peaks present in the x-ray pattern for Na-magadiite were not observed in the acid form. Sample x—ray patterns are presented in Figure 23. The lack of crystallinity of these systems, shown in Figure 23, illustrates why structural information is more b) 6) Figure 22. .. .. u. ,. .. L. -90 -100 -110 -IZO -l30 -140 PF“ 29 . . Si mas nmr(1) of (a) Na25114029, (b) freshly prepared H25i14029, (c) heated H28i14029 and (d) aged HZSi O29 obtained on the WH—l80 MHz spectrometer. l4 Figure 23. 128 X-ray diffraction patterns for (a) Nazs and (b) H23i14029° l14°29 1.259 _I‘ Auoov ““0 up. w m N ousmfim ON o“ as om «N ea _h‘ m~0<~umuozao .0.” 130 EN enema d)- .1.- q A_ooo Nu I. «— ed a~0v~wm~zan ma db 131 difficult to obtain for this condensed silicate. Aging the materials does not change the x-ray patterns obtained. 281140 the basal spacing was again 15 A but the pattern was very 29 When the sodium ions were exchanged back into H 29, broad. The Si mas nmr Spectra of the re-exchanged material was also very broad and quantitative data were not obtain— able. This is attributed to the formation of amorphous materials. The formation of amorphous phases during the synthesis of any of the aforementioned materials is a pos— sibility that should be considered when drawing conclusions from the data, especially in the acquisition of the quanti- tative information such as intensities. It was concluded that opposing silanol groups in the innerlayer of these crystalline Silicic acids will condense with each other if the distance between the layers is not too great, <2.8 A. Since the condensation occurs without a change of basal spacing, the layers must initially be close enough to interact. The presence of the Na(I) ions between the layers increases the interlayer separation and Slows or prohibits this condensation. Thermal studies indicate that, through water loss, the layers condense, or cross-link, in the formation of the high temperature phases (49). This concept has not been applied to the affects of aging the materials. The cross-linking explains why Na-kanemite, Na-maga— diite, and freshly prepared H-magadiite were intercalated 132 by the long chain alkylammonium salts while attempts to intercalate H-kanemite and aged H-magadiite were unsuccess- ful. The age of the material, hence the extent of the condensation, is an important factor when intercalating the layered silicic acids. This study has shown that the surface silanol groups do undergo condensation reactions. This could be applied to intercalants with hydroxide groups in an effort to stabilize the complexes. Care would have to be taken to control the condensations so that outside silanol groups do not condense and block, or limit access to, the innerlayer region where most of the surface area is located. When the innerlayer hydroxide groups are available, the layered silicic acids may be as useful, if not more so, than amorphous silicas. LIS T OF REFERENCES REFERENCES Theng, B. K. G. "The Chemistry of Clay—Organic Reac- tions" Wiley, New York, 1974. Grimm, R. E. "Clay Mineralogy" McGraw-Hill, New York, 1968. a. McBride, M. B., Pinnavaia, T. J., Mortland, M. M. J. Phys. Chem. 1975,79, 2430. b. McBride, M. B., Pinnavaia, T. J., Mortland, M. M. Am. Miner. 1975, 60, 66. c. Pinnavaia, T. J. in "Advanced Chemical Methods for Soil and Clay Minerals Research" eds. Stucki, J. W., Banwart, W. L., D. Reidel Publishing Co. 1980, 391. d. McBride, M. B. Clays Clay Miner. 1977, 25, 6. a. Raman, K. V., Mortland, M. M. Clays Clay Miner. 1968, 16, 393. b. Mortland, M. M., Fripiat, J. J., Chaussidon, J., Uytterhaven, J. J. Phys. Chem. 1963, 67, 248. 0. Russell, J. D. Trans. Faraday Soc. B 65, 61, 2284. d. Fripiat, J. J. ACS Symp. Ser. 1976, 34, 261. e. Conard, J. ACS Symp. Ser. 1976, 34, 85. Brindley, G. W., Brown, G. "Crystal Structures of Clay Minerals and Their X-ray Identification" Mineralogical Society, London, 1980. Hazen, R. M., Finger, L. W. "Comparative Crystal Chemistry" Wiley, New York, 1982. a. Whitehurst, D. D. Chemtech 1980, 44. b. Gates, B. C., Lieto, J. Chemtech 1980, 195. c. Grubbs, R. H. Chemtech 1977, 512. d. Hartley, F. R., Vezey, P. N. Adv. Organomet. Chem. 1977, 15, 189. 133 10. ll. 12. l3. 14. 15. l6. 17. 18. 19. 20. 21. 134 e. Koppelman, M. H., Dillard, J. G. Clays Clay Miner. 1980, 28, 211. f. Loeppert, R. H., Mortland, M. M. Clays Clay Miner. 1979, 27, 373. Endo, T., Mortland, M. M., Pinnavaia, T. J. Clays Clay Miner. 1980, 28, 105. a. Vaughn, D. E. W., Lussier, R. J. Preprint, 5th International Conference on Zeolites 1980, Naples, Italy. b. Pinnavaia, T. J. Science 1983, 220, 365. c. Pinnavaia, T. J., Tzou, M. S., Landau, S. D., Raythatha, R. J. Mol. Catal. 1984, 27, 195. Dixon, J. B., Weed, S. B. "Minerals in Soil Environ- ments" Soil Science Society of America, Madison, Wis- consin, 1977. Maes, A., Cremers, A. Clays Clay Miner. 1983, 31, 73. Barrer, R. M., Klinowski, J. Geochim. Cosmochim. Acta 1979, 43, 755. York, B., Solin, S., Wada, N., Raythatha, R., Johnson, I. D., Pinnavaia, T. J. J. Solid State Comm. In Press. Data Handbook for Clay Materials and other Non-Metallic Minerals, eds. Van Olphen, H., Fripiat, J. J. Pergamon Press, New York, 1979. Slade, P. G., Stone, P. A., Radoslovich, E. W. Clays Clay Miner. 1985, 33, 51. Fink, D. H., Nakayama, F. S., McNeal, B. L. Soil Sci. Soc. Am. Proc. 1971, 35, 552. Sheppard, R. A., Gude, A. J. Am. Miner. 1970, 55, 358. Eugster, H. P. Science 1976, 157, 117. Koizumi, M., Roy, R. Am. Miner. 1959, 44, 788. Suquet, H., Iiyama, J. T., Kodama, H., Pezerat, H. Clays Clay Miner, 1977, 25, 234. Roy, R. J. Amer. Ceram. Soc. 1956, 39, 145. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 135 Harward, M. E., Brindley, G. W. Clays Clay Miner. 1965, 6, 209. Iiyama, J. T., Roy, R. Clay Min. Bull. 1963, 5, 161. Roy, D., Roy, R. Am. Miner. 1955, 40, 137. Ibid 1954, 39, 957. Roy, D., Roy, R. Osborn, E. F. Am. J. Sci, 1953, 251, 337. Farmer, V. C., Fraser, A. R. Proc. Int. Clay Conf. eds. Morland, M. M. Farmer, V. C., Amsterdamzelsevier. 1978. a. Lipsicas, M., Raythatha, R. H., Pinnavaia, T. J., Johnson, I. D., Giese, R. F., Costanzo, P. M., Robert, J. L. Nature 1984, 309, 604. b. Oldfield, E., Kinsey, R. A., Smith, K. A., Nichols, J. A., Kirkpatrick, R. J. J. Magn. Reson. 1983, 51, 325. c. Sanz, J., Serraosa, J. M. Clay Minerals 1984, 19, 113. d. Barron, P. F., Wilson, M. A., Campbell, A. 8., Frost, R. L. Nature 1982, 299, 616. e. Barron, P. F., Frost, R. L., Skjemstad, J. 0., Koppi, A. J. Nature 1983, 302, 49. f. Sanz, J., Stone, W. E. E. Clay Minerals 1983, 18, 187. g. Alma, N. C. M., Hays, G. R., Samoson, A. V., Lipp- maa, E. T. Anal. Chem. 1984, 56, 729. Gerstein, B. C. Analytical Chem. 1983, 55, 7, 781. Lippmaa, B., Magi, M., Samoson, A., Engelhardt, G., Grimmer, A. R. J. Am. Chem. Soc. 1980, 102, 4889. Fyfe, C. A., Thomas, J. M., Klinowski, J., Gobbi, G. C. Angew. Chem. Int. Ed. Engl. 1983, 22, 259. Magi, M. Y., Samoson, A. V., Termak, M. K., Engelhardt, G., Lippmaa, E. Doklady. Ahad. Nauk. SSSR 1981, 261, 1159. Maes, A., Stul, M. S., Cremers, A. Clays Clay Miner. 1979, 27, 387. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 136 Maes, A., Cremers, A. J. Chem. Soc., Faraday Trans. 1 1977, 73, 1807. Billaud, D., Balesdent, D., Hérold, A. Bull. Soc. Chim. Fr. 1974, 11, 2402. Chow, P., Zabel, H. Synth. Met. 1983, 7, 243. Herold, A. in "Intercalated Layer Materials" Lévy, F. ed. D. Reidel Publishing Co. Boston, 1979. Ishii, M., Shimanouchi, T., Nakahira, M. Inorg. Chim. Acta 1967, 1:3, 387. Tzuo, M. S., Ph.D. Thesis, Michigan State University, E. Lansing, MI, 1983. Alexiades, C. A., Jackson, M. L., Clays Clay Miner., 1966, 14, 35. Shapiro, L. Brannock, W. W., ”Rapid Analysis of Sili— cate, Carbonate and Phosphate Rocks", Contributions to Geochemistry, Geological Survey Bulletin 1144-A, United States Printing Office, Washington, D.C. 1962, p. A22. Huheey, J. E. "Inorganic Chemistry" Harper & Row, New York, 1978. Vega, A. J. J. Am. Chem. Soc. Symp. Ser. 1983, 218, 217. a. Bennett, J. M., Blackwell, C. 8., Cox, D. E. J. Phys. Chem. 1983, 87, 3783. b. Thomas, J. M., Fyfe, C. A. Ramadas, S. Klinowski, J., Gobbi, G. C. J. Phys. Chem. 1982, 86, 3061. Klinowski, J., Thomas, J. M., Fyfe, C. A., Hartman, S. J. Phys. Chem. 1981, 85, 2590; c J d. Lippmaa, B., Magi, M., Samoson, A., Tarmak, M., Engelhardt, G. J. Am. Chem. Soc. 1981, 103, 4992. Oldfield, E. Personal communication. Sanz, J., Serratosa, J. M. J. Am. Chem. Soc. 1984, 106, 4790. Loewenstein, W. Am. Miner. 1954, 39, 92. a. Dempsey, E., Kuhl, G. H., Olson, D. H. J. Phys. Chem. 1969, 73, 387. 48. 49. 50. 51. 52. 53. 54. 137 b. Dempsey, E. "Molecular Sieves"; Soc. Chem. Ind., London, 1968. Lagaly, G. Adv. Colloid Interface Sci. vol. 11 Zettle- mayer, A. C. ed., Elsevier, Netherlands, 1979. a. Lagaly, G., Beneke, K., Weiss, A., Am. Miner. 1975, 60, 642. b. Ibid 1975, 60, 650. Beneke, K., Lagaly, G. Am. Miner. 1977, 62, 763. Brinkley, G. W. Am. Miner. 1969, 54, 1583. Fyfe, C. A., Gobbi, G. C., Klinowski, J., Thomas, J. M., Ramadas, S. Nature 1982, 296, 530. Liebau, F. Handbook of Geochemistry vol. II/3 Wedepohl, K. H., ed, Springer-Verlag, New York, 1972. "‘11111111111111“