\nm ,.-4,~-5- ._,_,‘ . ,n’-‘ ‘4 pf" -‘.‘l: :‘f...’ ' :f‘::‘1fflj":£‘r‘f h.-.“ I . \ I. -4” ...u’w‘,.‘:5‘.. ‘55-... -§\.:‘ iwfflhk I. Wt..- D '-J» ::3;,.‘ 3 5:33:45. 3.3 ,‘L 2; .23: 5 5'. '1 ., ‘D r “.54 'nL‘i -4' ~ ..-. 5 , I , J . .- , “,1”, .5_'~ . 5 ’1’; 2 . .. .5... .- __ Wu; gr , ”"3" - ' u‘ ’ A J .3...12....“‘"”" .... . ’ N» “’65-“ 355.3“ - :‘ifin: ‘ “It” 15 ”V‘W'M‘ 7: (v. x .- .. , .. H; ”swuh; 3”“: p yum ‘ 0 ‘52 ’ 2a ~ ' vie-uufirurh .. V ‘5‘... ;‘_3.~.»:-m...5,.::r . , J. m<>ur .. "' ‘ ' 'L“ ~ _‘ 0 "O‘*‘v 5 ‘ 07D.“ in . . ~ ‘ 5 . . . - . , .u-s ‘ . x 3 .5. .1»- gun. —.,..-. “223:3:- as $.32. . 3% ‘35.; 'ug ‘11,:‘15'. ‘ .5 735' 53 V5 ‘ Iv 3 ’33,." f. ’5'.“ ,5». " . :5 ~ ‘ 2 ~. . . 1. . ’ .. ..L. _ .., .3 ~ g. y“, ‘M‘ 4; m»: gene" . 3&4», ”dagger. ‘. . . , _ H :4. '? “95- .. V .5 5:35. 5 I- I1:??? “3%: ._55‘ 5:, IS L, , I.’ '57:. 4§1\1“‘” ‘ ,' . A “L _ . ‘ u h . ‘ “if” .5 . , ,» 113‘s. ‘5‘ ":4ch 5..-. .hu -§r {31:55 um ,. . .2 - fl . ‘ 2 ,s. "I . v -‘-v - ‘ ’ " . 1 . "(‘5‘ . 4""- L; 3 . -. 195-5711?!“ ”313‘ 1 'J J. M. ’23:;31’1’25“? *‘ ; . , ~ 5;“ n5. . .. ‘3. mug“? i 4 _ V i b , 5+ .‘7‘ .7 .13 Am} .3} 1‘13- _ 333 5. , . ‘i J .5 .. ‘ “r “'- ~ 2. 2 . ‘ " ’e ‘ . J ' ’2 “‘”‘*‘""".‘f ‘23; 9': :3}: "3-. " -; 3v? «Ilh‘ \.5v.'-ug.s£mu\'~1:fi':la:u»:n..’2‘ MdKfi “‘3 52191;: I’m\'l" {3" ‘ - u “ y ‘3, , . A. . . . . . , . ‘ #v - . ' ' L ' ~25: ‘ 7‘ . , ‘ Ab, .. \»‘51-‘.C. v‘x - ' 5 2 - . , 2. 1"» \u. ‘ w: v 1 '0. .r. $41; ‘ "hr..- -43424“: 2m {7 r~. ‘22; :17...‘ 4:..- 5‘. *i - - v , .. -. , r ‘ ' 5‘2 ., .‘ '2 A ' . j” ‘ , " .. FHA-’3‘; “J 25.. . f ' ‘2 . 4 v 4 .‘. , _ _ ‘ . {13.1?‘:.:'.’ w 3'33.” ‘ - ‘ ' ' fan} '"fir—r ' - '3‘" m "13:. ' ' I J 5 r 5» a < 2' "f .7 4:95,. -’ ' 15:72“ . 1 . ‘<‘. "r 73f; 5: y a?) J 5 IV’ ‘31:“; 2 f _ 3,75%? 2;;133. #71714"??? Lira: '“"'-‘“" «.5 "(ff-4:1;f5'1“ , I [‘7'522 r .4: . W I 2 " - I * ' l , 1', ' I ‘ .r 4" ‘ "I" ~NQ-f' 1 ‘ -- 2 .m' - . v ' - - m , , .2 .,...,1.---......,; 3 , . . ' ,t . _. ._ , . - , - y i .. ‘ . . , --._,. - . - ~ . . ‘ _ _ V _ a _ fizz: 7:3;- . . , . 5. T .a 1‘; -J,..fl:- A: ‘.:...~.{.:.,2 rm” 925335;:- 2 5 “12-. Tff"....-p...... .. 15...? :r. . .. .3 3 .. .‘,’:':‘:.‘:‘:“ ,. ,:;';V:V7;v:;"v_ . .,. . g . MlCHl IGANS III II IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII II'IjTI'II This is to certify that the dissertation entitled Layered Double Hydroxides, (LDHs) : Organic Anion exchange reactions, synthesis of microporous pil— lared derivatives and surface chemistry of metal carbonyl clusters. presented by Emmanuel D. Dimotakis has been accepted towards fulfillment of the requirements for Ph.D degree in Inorganic Chemistry Mfiyzw DateflJ/Z. /770 MS U is an Affirmative Action/Equal Opportunity Institution U303 LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE ll it :lfim MSU to An Affirmdive Action/Equal Opportunity Institution cMunS-DJ Part I LAYERED DOUBLE HYDROXIDES (LDt-te) : ORGANIC ANION EXCHANGE REACTIONS, SYNTHESIS OF MICROPOROUS PILLARED DERIVATIVES AND SURFACE CHEMISTRY OF METAL CARBONYL CLUSTERS P rt I OR THE NATURE OF TITANIA PILLARED LAIERED SILICATE CLAIS BY Emmanuel D. Dimotakis A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1990 ABSTRACT Bart I LAIERED DOUBLE HYDROXIDES (LDHS) t ORGANIC ANION EXCHANGE REACTIONS, SINTHESIS OP MICROPOROUS PILLARED DERIVATIVES AND SURFACE CHEMISTRY OF METAL CARBONYL CLUSTERS Part I ON THE NATURE OF TITANIA PILLARED LAIERED SILICATE CLAIS 3}! Emmanuel D. Dimotakie A. new' general method for the preparation of well- ordered layered double hydroxides (LDHs), [Mg1_,xAlx(OH)2] [Xn‘1n/x'yH20, interlayered by organic anions has been developed. It is based on the reaction of meixnerite, [Mg3Al(0H)8][0H]'2H20, with the free acid form of the desired anion in the presence of glycerol as a swelling agent. Thus, the nonanoate, adipate, p-toluenesulfonate, and squarate [Mg3A1(OH)3]-intercalates were prepared as single crystalline phases that are not readily available by conventional synthetic methods. The [Mg3Al(OH)3]-adipate and -p-toluenesulfonate derivatives undergo facile ion exchange reactions with Keggin-type polyoxometalate anions H2W120406' and SiW110393’ to form well-ordered, microporous pillared derivatives with the highest N2 BET surface areas reported to date, namely 107 and 155 m2/g, respectively. Meixnerite, [Mg3Al(OH)8][OH].2H20, has unexpectedly been found to be an excellent precursor for ion exchange with Keggin-type [XM12040]n' or lacunary [XM110391m' polyoxometalates (POMs) to form microporous pillared derivatives. TEM, XRD, FTIR and NMR studies show that the reactions are topotactic and that the framework. of the intercalated POMs remains intact. The effectiveness of this new route is compared with the one using LDH-organo-anion precursors. The meixnerite was conveniently prepared for the first time from calcination of [Mg3Al(OH)8][CO3]O.5°2H20 at 500 °C to form a mixed metal oxide and aqueous hydrolysis of the resulting NaCl-type solid solution. Metal carbonyl clusters have also been examined for ion-exchange (i.e., {[Pt3(CO)5]n}2') in these LDH- precursors. This chemistry' is compared. with the surface chemistry of [Mg3Al(OH)8][Xn']1/n'2H20 (x=co32' or on‘). It has been shown that the surface hydrolysis reaction : c032' + H20 —> aco3' + OH’ is responsible for reductive condensation reactions of neutral carbonyl clusters with the LDH. The reactions were as effecient as with Na metal in solution, i.e., [M6(C0)18]2' (M=Ru or Os) was synthesized. In part II of this work, Li-fluorohectorite, has been pillared with titanium polyoxocations derived from the acidic hydrolysis of TiC14 or Ti(i-OC3H7)4. Raman spectroscopy of the product indicates that the pillars have a structure analogous to the beta phase of titanium dioxide. A similar reaction product was obtained for the reaction with Na-montmorillonite. It is proposed that the pillars, in both jpillared. clays, consist of aggregates of condensed [T18012(H20)24]8+ cations. TO MY FAMILY ACKNOWLEDGMENTS I would like to thank Dr. T.J. Pinnavaia for his guidance and support throughout the years of graduate school. Help by several other Professors, and particularly by Dr. H.A. Sick and Dr. M.G. Kanatzidis, is also acknowledged. Financial support of this research from the Aluminum Co. of America (ALCOA), the National Science Foundation and the Center for Fundamental Materials Research at Michigam State University, is gratefully acknowledged. Finally, I am grateful to all members of Dr. Pinnavaia's research qroup. Their help and companionship created a unique environment during my stay in East Lansing. vi TABLE OF CONTENTS Part I Chapter Page LIST OF TABLES... ....... . .................................. x LIST OF FIGURES .................... . ....... . ............. xii CHAPTER I. INTRODUCTION............. ..... . ................. 1 A. Problem, sc0pe, aim, and general character of the research................................ ............. l B. Structure of basic Layered Double Hydroxides.. ....... 5 C. Synthesis of basic Layered Double Hydroxides ......... 8 D. Applications of LDHs........ ......................... 9 E. Heteropolyoxometalates, POMs..... ................... 11 F. Smectite clays................... ..... ....... ....... 14 G. BET surface area measurements on porous materials...16 H. Pore size measurements on porous materials .......... 17 F. List of References ....... . .......................... 20 CHAPTER II. Synthesis of Hydrotalcite-Like Layered Double Hydroxides (LDHs) Interlayered by Organo- Anions: Use of a Swelling Agent for Topotactic Intercalation Reactions ................................ 22 A. Introduction ....... . ................................ 22 vii Chapter Page B. C. D. E. Experimental....... ................................. 26 Results and Discussion .............................. 28 Summary............ ................................. 53 List of References .................................. 55 CHAPTER III. Synthesis of Well-Ordered Layered Double Hydroxides Pillared by Keggin-Like Polyoxometalates Use of Meixnerite, [Mg3A1(OH)8][OH].2H20, as a pre- cursor for topotactic ion exchange ..................... 57 A. Introduction ........................................ 57 B. Experimental ..... . .................................. 59 C. Results and Discussion .............................. 61 D. Summary........ ..................................... 66 B. List of References .................................. 94 CHAPTER Iv. Surface Organometallic Chemistry of Basic Layered Double Hydroxides: Reactions of Metal Cluster Carbonyls.......... ......... . ............ . ............. 96 A. Introduction ........................................ 96 B. Experimental ........................................ 99 C. Results ............................................ 106 D. Discussion ......................................... 116 E. Summary... ......................................... 136 F. List of References ................................. 137 viii Chapter Page CHAPTER V. On the nature of Titania Pillared Layered Silicate Clays ........................................ 139 A. Introduction ............. . ......................... 139 B. Experimental ....................................... 140 C. Results and Discussion................... .......... 145 D. Summary......... ................................... 170 E. List of References ................................. 172 ix Table II.I. II.II. III.I. 111.11. 111.111. IV.I IV.II. LIST OF TABLES Page [Mg3Al]-LDH Intercalates Containing Organic Anions.................................... ....... 51 Elemental Analyses of [Mg3Al(OH)8]-Organic Anion Intercalation Compounds ................... 52 BET surface areas of LDH-POM products prepared from [Mg3Al(OH)8]-[Xn‘]1/n precursors ............ 88 Elemental analyses of pillared products obtained using meixnerite, [Mg3Al(OH)8][OH].2H20, as precursor ........................................ 89 Elemental analyses of pillared products obtained using [Mg3Al(OH)8]-organo-anion precursors ....... 90 Powder xrd (Cu-Ka) of the orthorombic lattice of [M919_9A19.3]-[Pt15(co)3o]2' intercalate ........ 103 Summary of Ru3(CO)12 chemistry on [Mg3Al]-[CO3].105 Table IV.III. IV.IV. IVOVO V.II. Page Carbonyl stretching frequencies of ruthenium carbonyl anions desorbed from hydrotalcite after reaction with Ru3(CO)12. .................. 107 Summary of Os3(CO)12 chemistry on [Mg3Al]-[CO3].110 Carbonyl stretching frequencies of osmium and iridium carbonyl anions desorbed from hydrotalcite after reaction with Os3(CO)12 and Ir4(CO)12 respectively ...................... 112 Peak assignment of Titania pillared Fluorohecto- rite. [(T102)7.781(“92.65Lio.711(518020)(OHvFi4r dried at 70 °C .................................. 152 Peak assignment of Titania pillared Montmoril- lonite. I(Tiozi4.79(Na)o.031[A12.87M90.94Feo.191 [Si3020(OH)4], dried at 90 °c ................... 154 xi II.1. II.2. LIST OF FIGURES Page Structure of the basic layered double hydroxide, hydrotalcite, [Mg3A1(OH)8+][CO32’]0.5'2H20 ........ 6 Structure of Keggin-type (left) and lacunary (right) Keggin-type polyoxometalate .............. 12 Polyhedral models and relationships between the different known tungstosilicates........ ..... 13 Idealized structure of a smectite clay mineral. (0) Oxygen atoms; (0) hydroxyl groups. Silicon and sometimes aluminum normally occupy tetra- hedral positions in the oxygen framework. Aluminum, magnesium, iron, or lithium may occupy octahedral sites. Mn+.xH20 represents the inter- layer exchange cation ............................ 15 Preferred orientation XRD pattern (Cu-Ka) and d-spacings (A) of meixnerite, [Mg3Al(OH)8]'[OH]'2H20...................... ..... 3o TGA curves for : (A) [Mg3A1(OH)8].[CO3]0.5.2H20, (B) [Mg3Al(OH)8]'[OH]°2H20 ....................... 31 xii II.3. 11.4. 11.5. II.6. XRD patterns (Cu-Ka) and d-spacings (A) of oriented film samples : (A) synthetic meixnerite, (B) the intercalate obtained after exposure of sample A to a small amount of 2:1 (v/v) glycerol/water mixture, under nitrogen for one day, (C) sample B kept under nitrogen for an additional day, (D) sample B after three days, (E) sample B kept for four days under nitrogen................ .............. 33 XRD patterns (Cu-Ka) and d-spacings (A) of oriented film samples : (A) [Mg3Al(OH)8]°[OH]' 2H20 exposed to a small amount of 2:1 (v/v) glycerol/water mixture, under nitrogen, for 4 days, (B) sample A after one day, (C) sample A after exposure to a very small mixture of 2:1 (v/v) glycerol/water, under nitrogen, for 1 day..34 XRD patterns (Cu-Ka) and d-spacings (A) of the material that results from the dispersion of aqueous suspension of [Mg3A1(OH)8]'[OH]°2H20 in glycerol (2:1 v/v glycerol/water) and subsequent evaporation of the solvents, under vacumn, on a glass slide....... ...... ......36 XRD patterns (Cu-K3) and d-spacings (A) of oriented film samples : (A) [Mg3A1(0H)8]°[OH]° 2H20 (B) [Mg3A1(OH)3]'[OH]°2H20 dispersed in a 2:1 (v/v) glycerol/water mixture and dried at 190 00 for 1 hr, (C) [Mg3A1(OH)8]°[OH]'2H20 xiii II.7. II.8. II.9. II.10. II.11. dispersed in a 2:1 (v/v) glycerol/water mixture and flashed dried at 190 °C.......... ............ 37 TEM images of : the product formed by the reaction of meixnerite with glycerol at 190 °C, (top left); meixnerite, (top right); [Mg3A1(OH)8][C6H804]0.5, (bottom left): and [Mg3Al(OH)8][p-CH3C6H4SO3], (bottom right) .................................... ...... 39 Preferred orientation XRD pattern (Cu-Ka) and d-spacings (A) of the material that results from titration of [Mg3Al(OH)8]Cl with p-toluene- sulfonic acid until neutral pH is obtained.......42 Preferred orientation XRD patterns (Cu-Ka) and d-spacings (A) : (A) [Mg3Al(OH)8][C6H804]0 5, (B) [Mg3Al(OH)8][p-CH3C6H4SO3]...... ............ 44 Preferred orientation XRD patterns (Cu-Ka) and d-spacings (A) of [Mg3A1(OH)8][C4O4]0.5, obtained by the reaction of meixnerite and squaric acid, (A) in the presence of 2:1 glycerol: H20 (v/v) (B) in the absence of glycerol.. ....... 46 Preferred orientation XRD patterns (Cu-Ka) and d-spacings (A) of [Mg3Al(OH)8][C9H1702]1_2, obtained by the reaction of meixnerite and Na-nonanoate, (A) in the presence of 2:1 glycerol: H20 (v/v), (B) in the absence of glycerol ........ 47 xiv II.12. II.13. III.1. III.2. III.3. XRD pattern (Cu-K3) and d-spacings (A) of [Mg3Al(OH)8]-Ni(II)phthalocyaninetetrasulfonate derivative............ ...... ........... .......... 48 XRD patterns (Cu-Ra) and d-spacings (A) of oriented film samples : (A) [Mg3A1(OH)8]°[OH]° 2H20, (B) the product obtained from the reaction of meixnerite with p-toluenesulfonic acid in the presence of 2:1 glycerolzflzo (v/v), (C) the product obtained from the hydrothermal reaction of calcined hydrotalcite and p-toluenesulfonic acid in aqueous media ............................ 50 Preferred orientation XRD patterns (Cu-Ka) and d-spacing (A) of [Mg3Al(0H)8]6[H2W12040] pillared products, derived from the reaction of [Mg3A1(OH)8]'[OH] and (NH4)6H2W12040 : (A) prepared in the presence of 2:1 glycerolznzo (v/v), (B) prepared in the absence of glycerol...67 Represantative TEM images of : [Mg3A1(OH)8]6- [H2W12040] derived from meixnerite in pure water, (top left); [Mg3Al(0H)8]6[H2W12040] derived from meixnerite in the presence of 2:1 glycerol:H20 (v/v), (top right); [Mg3A1(0H)8]6- [H2W1204o] derived from the adipate precursor, (bottom left); meixnerite, (bottom right). ....... 68 FTIR spectra of [BVW1104016'I (A) as [Mg3Al(OH)8] intercalate and (B) as potassium salt ............ 70 XV III.4A. III.4B. III.4C. II.5. III.6. III.7. 51V Solution NMR of [BVW11040]6‘ after dis- solution of [Mg3Al(OH)8]6[BVW11040] in LiClO4 (pH=1.7) ..... ....... .......... . ........ 72 51V Solution NMR of [SiV3W904016' after dis- solution of [Mg3A1(OH)8]7[SiV3W9040] in LiClO4 (pH=1.7)................... ............ 73 2931 MAS-NMR of [Mg3A1(OH)8]7[SiV3W9040]... ...... 74 Preferred orientation XRD patterns (Cu-Ra) and d-spacings (A) for [Mg3A1(OH)3]-intercalates: (A) [Mg3Al(OH)8][02C(CH2)4C02]0.5 (B) [M93A1(0H)3]5[H2W1204o] (C) [Mg3Al(OH)8]8[SiW11039]. Both a and C intercalates were derived from the adipate precursor... ................. ..... ............... 76 Preferred orientation XRD patterns (Cu-Ka) and d-spacings (A) for [Mg3Al(OH)8]-intercalates: -(A) [Mg3Al(OH)8][p-CH3C6H4SO3'] (B) [M93A1(0H)3]5[HZW120401 (C) [Mg3A1(OH)8]8[SiW11039]. Both B and C intercalates were derived from the -p-toluene- sulfonate precursor..................... ......... 77 Represantative TEM images of : [Mg3Al(OH)8] [02C(CH2)4002]O.5 , (top left); [Mg3Al(OH)8] [SiW11039] derived from the previous adipate precursor, (bottom left); [Mg3Al(OH)8]- [SiW11039] derived from the adipate precursor and then treated with Na-adipate, (top right): xvi III.8. III.9. III.10. 111.11. 111.12. III.13. IV.1. [Mg3A1(OH)8][p-CH3C6H4SO3'], (bottom right)......78 FTIR spectra of [Mg3Al(0H)8]8[SiW11039] derived from : (A) [Mg3A1(OH)8][02C(CH2)4C02]0.5 pre- cursor (B) [Mg3Al(0H)3][p-CH3C6H4SO3‘] precursor.80 Adsorption-desorption isotherms of [Mg3Al(OH)8]8_ [SiW11039] (top) and [Mg3Al(OH)8]6[H2W12040] (bottom). Both pillared products were derived from the [Mg3A1(OH)8][02C(CH2)4C02]0.5 precursor.82 t-plots for nitrogen adsorption on pillared [Mg3Al(OH)8]-intercalates derived from the adipate precursor : (A) [Mg3A1(0H)8]8[SiW11039] (a) [Mg3A1(OH)8]6[H2W11040]........... ........... 83 Pore size distribution plot for [Mg3Al(0H)8]- intercalates : (A) [Mg3Al(OH)8]8[SiW11039] (B) [Mg3A1(OH)8][H2W1204011/6........ ............ 35 Proposed model for the observed differences in BET surface areas.... ....... . ..... . ........... 89 General reaction scheme ......... ....... ..... .....93 FTIR spectra (transmission mode) of [HM3(CO)11]' anions formed by the reaction at 25 °C of the corresponding M3(C0)12 clusters with the external surface of [Mg3A1(OH)8][CO3]0.5 '2820 in tetrahydrofuran (THF) and subsequently desorbed by reaction with [Et4N1C1: (A) [HRu3(c0)1l]‘ in THF, (B) [HRu3(CO)11]' in CH2C12, (C) [HFG3(CO)11]- in CH2C12, xvii Iv02. IV.3. (D) [HOS3(CO)11]' in CHZCIZ............. ........ 108 (A) FTIR spectrum of : (A) [Ru6(CO)18]2' anion formed by the reaction at 90 °C of Ru3(CO)12 with [Mg3Al(OH)8][CO3]O.5'2H20 in cyclohexane. The anion was extracted from the [Mg3A11—CO3 surface by metathesis reaction with [PPNJCl in in CH2C12, (B) [Ru6C(CO)16]2' anion (nujol mull) grafted on the [Mg3A11-C03 surface. The ion was formed by reaction at 170 °C of Ru3(CO)12 with [Mg3Al(OH)8][CO3]0.5'2H20 in diglyme. (C) [Ru(CO)x(O-M)2]n ( x=2,3 and M=Mg, Al ) species (nujol mull) that result from oxidation of [HRu3(C0)11]‘,[Ru6(C0)18]2' and [Ru6C(CO)16]2', on the LDH surface, in the presence of air and light...... ................. 110 FTIR spectrum of : (A) [056(C0)18]2' anion formed on [Mg3Al]-CO3 surface by the with LDH in diglyme at 170°C, (B) [0510C(CO)24]2‘ formed after reaction of 053(C0)12 with [Mg3Al(OH)8] [C03]0_5'2H20 in tetraglyme at 230 oC. The anion was desorbed from the LDH surface using [Et4N]Cl in CHZCIZ, (C) [Ir8(C0)20]2‘ formed by reaction of [Mg3A1(0H)8)[C03]0_5-2H20 with Ir4(CO)12 in THF at 25 °C. The anion was extracted from the LDH surface by metathesis reaction with Et4NCl..114 xviii IV.4. IV.5. IV.6. IV.7. IV.8 IV.9. X-ray (Cu-Ra) powder diffraction pattern of the product formed by the reaction of [Pt15(CO)3o]2' with [Mg3Al(OH)8][p-CH3C6H4-SO3], in 1:1 (v/v) glycerol/methanol mixture .......... 117 SEM images of : meixnerite, (top left): the product that results from ion exchange of [pt15(C0)3o]2‘ in the [Mg3Al(OH)8] [p-CH3C6H4-SO3] host lattice, (top right); the same product, (bottom left); The view is now along one of the edges of the sample; the same product as in the top right picture but at higher magnification, (bottom right) .......... 118 FTIR spectra of : (A) [Mgg.95Al4.65]- [Pt15(CO)3o]2' as KBr pellet (1 wt%) (B) Na2[Pt15(C0)30] in MeOH. .................. 120 FTIR spectra of : (A) [Mg9.95Al4.65]- [Pt15(co)30]2' and (B) [Mg3Al(OH)8]- [p-CH3C6H4SO3]. Molecular vibrations due to 303' group of the toluenesulfonate anion at 1125 cm-1 are present in spectrum B and absent in spectrum A. Lattice vibrations due to Al-O-Mg groups are present between 550 and 680 cm"1 in both spectra........................ ....... ..121 TGA curve of [Mgg.95Al4.65]-[Pt15(CO)3O]2‘ ...... 122 (A) X-ray powder diffraction pattern of the product that results from heating [MgSAl(OH)3] [C03]o.5°2H20 suspended in tetraglyme at reflux xix IV.10. temperature (170 °C) for 10 hrs. (B) X-ray (Cu-Ka) powder diffraction pattern of [Mg5A1(OH)8][CO3]0.13[Ru6C(CO)16]0.39 that that results from heating [M95A1(0H)3][C03]0.5 °2H20 and Ru3(CO)12 in tetraglyme at reflux temperature (170 °C) for 10 hrs.... ...... 125 Representative TEM images of three different regions (A, B and C) for the reaction product of tatraglyme with [MgSAl(OH)8][C03]0.5'2H20 at 170 °c for 10 h .................................. 126 XRD (Cu-Ka) diffraction patterns of solids that result from evaporation of hydrolyzed TiCl4 solutions in the presence of HCl : (A) 1.7 M Ti and 6 M HCl, (3) 0.82 M Ti and 0.6 M HCl.... ............ . ................... 146 Powder X-ray (Cu-Ka) diffraction pattern and d-spacings (A) of titania pillared f1uorohectorite....................... ........... 148 XRD (Cu-Ka) patterns and d-spacings (A) of titania pillared montmorillonite samples, prepared : (A) from hydrolysis of TiC14 with HCl, (B) by drying sample A at 500 °C for 3 h, (C) from hydrolysis of Ti(OC3H7)4 with HCl, (D) by drying sample C at 500 0C for 3 h ............. 149 Adsorption-desorption isotherms of titania pillared : (A) fluorohectorite, prepared from XX V.10. V011. hydrolysis of TiC14 with HCl, (B) montmoril- lonite prepared from hydrolysis of Ti(OC3H7)4 with HCl, (C) montmorillonite, prepared from hydrolysis of Tic14 with HCl............ ......... 152 Pore size distribution diagram of titania pillared : (A) fluorohectorite prepared from hydrolysis of TiCl4 with HCl, (B) montmoril- lonite prepared from hydrolysis of Ti(OC3H7)4 with HCl, (C) montmorillonite, prepared from hydrolysis of TiCl4 with HCl.......... ........... 153 Continuous Excitation Raman Spectra of : (A) titania pillared fluorohectorite dried at 70 °C (B) titania supported on 8102 dried at 550 0C, in ref. 11... ...................... . ............. 156 Pulse Excitation Raman Spectra of Titania pillared Montmorillonite dried at 90 °C .......... 158 Proposed model for the explanation of the dif- ferent gallery heights observed for the titania pillared layered silicate clays ....... ...160 Structural relationship of the octameric titanium cation and Ti02(B) phase.......... ...... 161 Structural relationship of TiOZ(B) phase (left) and anatase (right)..............................163 TGA of titania pillared montmorillonite prepared by using the Ti(OC3H7)4 precursor ....... . ........ 164 DSC of titania pillared montmorillonite prepared by using the TiCl4 precursor (the lower line xxi V0130 v.14. represents the first derivative of the DSC curve.166 FTIR spectrum of evaporated TiC14/HC1 solutions..167 FTIR spectra of : (A) Na+-montmorillonite, (B) titanium pillared montmorillonite dried at 90 °C and (C) sample B dried at 500 °C for 3 h... ...... 169 xxii CHAPTER I INTRODUCTION A. Problem, scope, aim, and general character of research, Following World War II, clay-cracking catalysts were replaced by mixed-metal oxides and, eventually, by synthetic zeolites with improved hydrothermal stability and shape select ivity . However recent developments in the intercalation of smectite Clays have challenged researchers atgain.1 The new pillared Clay materials with unique two dimensional (2D) framework and zeolitic microporosity were PrOmiSing hosts for selective adsorption and catalysis.1'2 The need for improved molecular sieving Performance in Pillared Clay materials has motivated this study introduce microporosity in Mg-based, hydrotalcite Clays or layered dOuble hydroxides (was), [Mg1-xA1x<0H>21Ixn‘Jn/x-szol where X“. is the interlayer exchange anion. Their high Charge density, however, was a serious potential Problem for Pillaring since close lateral spacing 0f the gallery anions and hydrolysis reactions of the pillars could occur. Stuffing the galleries with pillaring anions would result in 2 little or no free volume for adsorption of guest molecules and would preclude potential catalytic action by the gallery anions. One would expect that polyoxometalate anions (POM)“", m>n, could easily replace Xn' and introduce microporosity in the host lattice. However, these pillared intercalates are intrinsically difficult to synthesize in highly crystalline form, in part, because LDH hosts are basic, whereas most POMs suitable for pillaring are acidic (e.g., V100286", SiV3W90407').3 Thus, hydrolysis reactions of the LDH or POM can result in products that are poorly ordered, 4 even X-ray amorphous, 5 or that contain multicrystalline phases interlayered by different anions. It has been recently suggested that the organo-anion derivatives of LDHs can be used as precursors to pillared LDH-derivatives.6 In this case, the organic anion is readily replaced by the POM, and competing hydrolysis reactions are minimized.7 Since LDH ion—exchange reactions are topotactic, any layer stacking defects in the precursor also will appear in the pillared product. In order to achieve regularly microporous materials, such defects should be minimized. However, a general method for the preparation of well- °rdered organic anion derivatives of LDHs is lacking. The goal of this research was two-fold: First, to prF—‘Pare Mg-based LDHs interlayered with organic anions by a new, practical and efficient route. Second, the use of these LDH-organo anion precursors for ion-exchange reactions with polyoxometalate anions of high charge density to circumvent the problem of "anion stuffing" the galleries. Stuffing is the term used when the anions are too closely spaced to allow free space for adsorption of molecules larger than water. The resulting organic anion intercalates consisted of alternative inorganic/organic layers allowed tuning of the size of the organic layer. That is, by choosing organic anion with the appropriate length, the gallery height of the precursor could be mediated. The pillared derivatives were shown to have microporosity and increased pore volume,7 which becomes accessible for adsorption or diffusion of organic molecules into the structure from the outside with potential catalytic action. The unique features of the approach described in this thesis included for the first time the use of a swelling agent, particularly glycerol, to access the interlaYer 0f LDHs, especially the hydroxide-intercalated derivatives such as meixnerite, [Mg3Al(0H)8][OH]°2H20. By swelling the LDH galleries, it was possible to show that the gallery surfaces are readily available for ion exchange by organic anions and P°1Yoxometalates of very large size and Charge density. This breakthrough has led to the synthesis of a broad new family 04‘- microporous compounds with interesting preperties as molecular sieves and as potential shape selective catalysts for base-catalyzed chemical conversions and redox catalysis. 4 It was also discovered that hydrotalcite can be used as a reactant-the other reactant being a neutral carbonyl-to synthesize carbonyl anions directly on its surface. The products of this surface chemistry and ion—exchange of carbonyl anions are presented. My efforts were focused on the LDH surface Chemistry of M3 (CO)12 (M=Ru, Os, Ir) system. Hayward and Shapley8 showed that very strong reducing agents such as sodium metal were needed to reductively condense these clusters in solution. In the present work we show that similar reactions can be done on the surface of a basic LDH in the absence of a strong metallic reducing agent. Direct ion exchange reactions of carbonyl clusters ( i.e 'I {[Pt3(CO)6]n)2‘ ) with LDH host lattices were examined. Finally the goal of the second part of this dissertation, Chapter V., is to provide information on the nature of titania pillared layered silicate clays.9'10'11 These structures consist of titania pillars with size substantially larger than the van der Walls' thickness of the host layers (~10 A). However up to now nobody has exPlained the composition of the pillars. Although Raman Spectroscopy can be a useful tool, there are serious Problems with the high fluoresence background of the clays. This fluoresence arizes from iron and other impurities that are present in the natural clays. Therefore the effort was mainly on using synthetic Clay, such as Li-fluorohectorite. 5 Li-fluorohectorite, a high Charge density SYchetiC clay, has been pillared for the first time with titania polyoxocations that result from the acidic hydrolysis of TiC14 or Ti(OC3H7)4. Continuous excitation Raman spectra of the resulting pillar product provided information on the nature of its titania pillars. Similar information was obtained for the titania pillared montmorillonite but by using pulse excitation Raman techniques to eliminate the fluoresence background of this pillared Clay. The titania pillars in both clays is shown to result from the polymerization of [Ti8012(I-120)24]8+ . A model iS preposed to explain the resulting phases from this cation. 8. Structure of basic layered double hydroxides, (was), lucky: (on) 2] [xn‘] "a . M20 The structure of the layered double hydroxides ( LDHs ) or hydrotalcites is based on a two dimensional, brucite-type 01' M9(OH)2-like, lattice.12 Some of the divalent Mg cations are displaced with trivalent Al cations, thus creating a Positive Charge on the layer. This Charge is balanced by anions between the layers, Figure 1.1. Water molecules usually co-occupy the interlayer region. When x=0 in the above formula, we have a brucite lattice with a=3.14 A. The highest value of Al content x reported is 0.44. This can be achieved by synthesis under am). am” e w m. a m. a \fl \ ... \Q/ \ A \ Q 3 \\\\\od\\\\ \\\Ofl\\\ eves s: s Ii .mlli Figure 1.1. Structure of the basic layered double hydroxide, [M93A1(0H)8*] [CO32'] 0,5 .2H20 hydrotalcite, hydrothermal conditions.13 Normally, the typical Al content is in the range x=0.12-0.33. The carbonate-LDH ( X = C032” ) occurs in two dimorphic forms, rhombohedral hydrotalcite and hexagonal manasseite.13 In the former case the layers are stacked with rhombohedral symmetry ( BC...CA...AB...BC...for the brucite-like main layers ) and three double-layers are present per unit cell. In the latter they are stacked with hexagonal symmetry ( BC...CB...BC... ) and two double layers are present per unit cell. For the same value of x = 0.25 different basal spacings have been reported: 7.63 < C < 7.80. The thickness of the brucite layer' is approximately 4.8 A, while the gallery height of 3 A is due to the size of the anion and water molecules. In nature hydrotalcite and manesseite are commonly intergrown. Manasseite generally forms the core and hYdrotalcite the outer part of a grain. Thus, hydrotalcite appears to form later than the coexisting manasseite and Presumably at lower temperatures. They have been found in the saline deposits of the central pre-Caspian depression and in those of Middle Asia. The meixnerite form of hydrotalcite, X=OH’, has been found in Austria. The charge density of the LDHs can be calculated as fallows : the area A, per octahedron, on a Mg(OH)2-type Structure is given by the equation : A.==(ao)2 3(1/2) / 2, where a0 is the distance between metal atoms in adjacent °°tahedra (it results from the fact that the projected area of the octahedron is twice the area Of a single face). Thus for Mng1(OH)5 a0 = 3.04 A and A = 8.0 A2 per octahedron. The area of a MZIIMIII(OH)6 unit cell (three octahedra) is given by the same equation, but now a0 is the distance between MIII ions in the unit cell. The layered double hydroxides or anionic Clays can be considered as antitypes of 2:1 clay minerals. In smectite Clays the layers are negatively Vcharged and interlayer cations balance the charge, while in LDHs the layers are positive and interlayer anions balance the charge. C. Synthesis of basic layered double hydroxides There are several approaches for the synthesis of the LDHs. They are all based on the coprecipitation of a divalent and a trivalent metal cation with a base such as sodium hydroxide or carbonate at well defined pH values. The preferred pH is between 8.5 and 10 for the MgZ+/Al3+ System.“ Lower pH values may be necessary for other metal ion systems.According to the original synthesis by Feitknechtls, dilute solutions of Mg2+ and Al3+ were reacted With caustic solutions following dialysis and drying. Tayl0r16 used the induced hydrolysis method to prepare LDHS. Taylor's method involves the hydrolysis Of a cation in SOlution by a fully hydrolyzed and precipitated hydroxide of a Second metallic cation. This seemingly complex technique has resulted in a surprising number of very interesting PrOducts ( M4+=si or Ti). Miyatal'7 used the cloride or nitrate solutions of the metals and sodium hydroxide as a base. Reichle18 used more concentrated solutions of metal salts and sodium hydroxide/carbonate to make gels which he crystallized at temperatures 60-325 °C for up to 18 h. Depending on the conditions used hydrotalcites with different particle sizes and surface areas were produced. General rules we must remember from his synthesis are 8pSl-p32>0 , where pS=-logS and 81, S2 are the solubility products of M2+(OH)2 and M2+(xn-)2/n respectively (II). M3+(OH)3 is kept as a monomer or almost monomer state. (III).ioniC radii of M2+ and M3+ are Close to each other. D. Applications of LDH: There is an increasing interest in layered double hydroxides for a variety of reasons. These compounds are being used in three main areas: 10 (1). catalysts and catalyst precursors (ii). ion exchangers and adsorbers, and (iii). medical applications The Cu/ZnO catalyst for syn gas conversion to methanol is obtained from copper, zinc and aluminum nitrates by coprecipitation techniques and subsequent calcination of the resulting LDHs.12'18 Typical impregnation procedures for the preparation of transition metal catalysts normally do not achieve the high metal dispersion possible for LDH-based catalysts. Ziegler catalysts can also be prepared by reacting LDH with phosgene and then with TiCl4 at elevated temperatures.12 A very active catalyst for ethylene polymerization results. Vapor-phase aldol condensation14 and a number of peripheral base catalysed reactions can be done on materials that result from the heating of hydrotalcites. These mixed metal oxides can be also used as halogen scavengers in polyolefin production and as polymer stabilizers.19 A. widespread application of LDHs is anticipated. by reason. of the jpronounced. anion. exchange capacity toward inorganic and organic anions. They find medical applications as stomach antiacid, but their real use may be adsorbers of ecologically undesirable anions from dilute and possibly radioactive aqueous waste streams.12 11 LDHs or their derivatives can be used in chromatographic columns for gas separation, i.e. a mixture of air and hydrogen to hydrogen, nitrogen and oxygen.17 Hydrotalcite modified electrodes with electrochemically active anions have Ibeen also prepared.20 Composite electrodes with LDH as one componet have been examined too.21 Finally, applications in wet-type desulfurization and hydrometallurgy for pollution control are expected. E. Heteropolyoxometalates, POMs. Heteropolyoxometalates include 51 large class of compounds with the general formula [XmeOy1q’ (x 500 A The Kelvin equation relates the adsorbate vapor pressure depression to the radius of a capillary which has been filled with adsorbate: lnP/Po = 2 vm coséi/ RTrK P saturated vapor pressure in equilibrium with the adsorbate condensed in a capillary ar pore normal adsorbate saturated vapor pressure '0 0 ll 18 F = surface tension of N2 at -l95.8 °C = 8.85 ergs/Cm2 Vm = molar volume of liquid N2 = 34.7 cm3 a = contact angle of N2 with surface, usually cos” = 1 R = gas constant = 8.134 x 107 ergs deg"1 mol’1 T = 77 K rK kelvin radii of pore = 4.146/logP/Po A The actual radii of the pore is rp = rK + t , t is the statistical thickness or thickness of the adsorbed layer. It postulated that t = 3.54 Va/Vm , where 3.54 is the thickness of one N2 layer and V'a,Vm are the volumes of the adsorbed gas, at P/Po, in the sample and in an non-porous material respectively. The values of' t are given for non-porous materials in tables or can be calculated easily from the Hasley27 equation : t = 3.54 (5/2.30310gP/P0)1/3 Thus rp values can be calculated too. Considering all the above and writing simple programs to facilitate calculations, the following. table can be constructed: P/PO Vads (Si-Sj) "(Si-53') 2(ri-rj) (Vi-Vj)liq 2r % 2r The symbols 1 and j define the specific parameter (S, r or Vliq) at a relative pressure P/Po and the next lower relative pressure P/Po, respectively. % 2r is the percentage of the pores with diameter 2r. 19 From the above table, a plot of % 2r versus 2r would give the pore size distribution diagram for the sample under investigation. Plots of the adsorbed gas volume versus statistical thickness can be also made. The latter are important because they can be used to calculate surface areas too. These new values of the surface areas can be compared to the ones obtained using the BET equation. It must be noted that the model used pore size calculations, in this dissertation, was the one for parallel pores. There are other models, i.e., model for cylindrical pores or spheres. LI ST OF REFERENCES 10. 11. 12. 13. 14. LIST OF REFERENCES Thomas J. Pinnavaia, Ming-Shin Tzou, Steven D. Landau and Rasik H. Raythatha, J. Mol. Catal.,1984, 27, 195. Thomas J, Pinnavaia,Ming-Shin Tzou and Steven D Landau, J. Am. Chem. Soc.,1985, 107, 4783. Kwon, T., Ph.D. dissertation, 1988, Michigan State University. For instance, many of the LDH-POM reaction products reported by Wolterman in 0.8. Patent 4,454,244 are x-ray amorphous. Chibwe, K.; Jones, W., Chem. Mater. 1989, l, 489. Drezdzon, M.A., Inorg. Chem. 1988, 27, 4628. Dimotakis, E.D.; Pinnavaia, T.J., Inorg. Chem. 1990, 29, 2393. Hayward, C.T.; Shapley, J.R.; Inorg. Chem., 1982, 21, 3816. Pinnavaia T. J., Science, 1983, 220, 4595. Sterte J., Clays and Clay Miner., 1986, 34, 658. Yamanaka, S., Mater. Chem. Phys., 1987, 17, 87. Reichle, W.T., CHEMTECH 1986, 58. Paush, I., Lohse, H., Schurmann, K., Allmann, R., Clays and Clay Miner., 34, 5, 507, 1986. Reichle, W.T., J. Catal., 94, 547, 1985. 20 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 21 Feitknecht, W., Helv. Chim. Acta, 25, 131, 1842. Taylor, R. M., Clay Miner., 1984, 19, 591. Miyata, 5., Chem. Lett., 483, 1973. Numan, J.G.; Himelfarb, P.B.; Herman, R.G.; Klier, K.; Bogdan, C.E.; Simmons, G.W., Inorg. Chem. 1989, 28, 3868. Miyata, S., Clays Clay Miner., 1980, 28, 50. Itaya, K.: Chang, H-C.: Uchida, I., Inorg. Chem., 1987, 26, 624. Shaw, 3.; Creasy K.E., J. Electroanal. Chem., 1988, 243, 209. Pope, M.P., "Heteropoly and Isopoly Oxometalates", Springer, New York 1983, 87. Altenan, J.T.; Pope, M.T.; Prados, R.A.; So. H., Inorg. Chem. 1975, 14, 417. Dixon, J.B., in Minerals in Soil Environments, J.B. Dixon and 8.8. Weed, Eds. (Soil Science Society of America, Madison, Wis., 1977), chap. 11. Anderson, J.R., in the Structure of Metallic Catalysts, J.R. Anderson, Ed., 1974, p. 291. Lowel, S. and Shields, J.E., in Powder Surface Area and Porosity, 2nd ed., (Chapman and Hall, London), 1984. Hasley, G.D., J. Chem. Phys., 1948, 16, 931. CHAPTER II Synthesis of Hydrotalcite-Like Layered Double Bydroxides (LDBs) Interlayered by Organo-Anions: Use of a Swelling .Agent for Topotactic Intercalation Reactions. As INTRODUCTION The success of molecular sieves for catalytic purposes has prompted a search for other porous inorganic materials that could act as shape selective catalysts. LDHs are an important class of materials that have been used as precursors to mixed metal oxide solid catalysts for synthesis gas-to-methanol productionl, aldol condensationz, Ziegler-Natta type catalysts3 and as ion exchangers4. However, a general route to well-ordered microporous derivatives remains a Challenging problem. Layered double hydroxides pillared by polyoxometalate anions (POMs) have been recently recognized as a promising Class of microporous materials for selective adsorption and catalysis.5'6'7 One promising synthetic route to these materials involves the topotactic ion exchange of an organo anion derivative of the LDH with the desired POM anion.8'9 The role of organo anion is to inctrease the lateral separation of the layers and 22 23 thus allow the large POM anions to ion exchange in the host structure. The preparation of organo anion precursors, however, is not straight-forward. Up to date the synthesis of organic anion LDHs is limited to the following methods: I. Coprecipitation methods: Such methods have been reported by Miyata in Japanlo, Reichle at Union Carbide11 and Drezdzon at Amoco Chemical Companya. Although there is success in obtaining several derivatives (i.e., LDH-terepthalate, LDH-napthol yellow S) there are limitations when dicarboxylic acids are used. For example, Reichle11 reported diffuse xrd pattern for LDH-1,12 dodecanedicarboxylic acid dianion, Drezdzon8 attributed the difficulty in forming well-ordered carboxylates to the sensitivity of the products to digestion conditions. II. Titration of LDH-CO3 with free acid. Bish12 was among the first to report that the titration of LDH-CO3 with an acid stronger than carbonic acid can afford a corresponding LDH-intercalate containing the conjugate base of the acid. This method when applied to HCl, HNO3 and H2804 required very dilute solutions of the acids and long reaction times (i.e. days). Bish also stated that only partial exchange had been accomplished with the hydrotalcite, [Mg3Al(0H)8]'[CO3]O.5'2H20. However, no xrd patterns of the resulting phases were reported. III. Ion exchange Here the work is limited to a selected number of organo anions such as dodecylsulfate. Most of these reactions are 24 mass-action driven and therefore the products require extensive washing. Since the stronger conjugate base is preferred in the LDH galleries it is possible that some of the C032" or OH" anions from the water might replace these intercalated anions upon washing. Dutta13 reported exchange reactions with anions such as benzoate or Ni-pthalocyanines. However, multiple phases resulted with the former while the latter oriented the rings parallel to the layers. In another recent report14, Y. Park, H. Kuroda and C. Kato, ion exchanged a tetrasulfonated. porphyrin (using its sodium salt) in the [Mg1-xAlx(OH)2]°Cl°H20 galleries. IV; Hydrothermal reactions This work was recently reported by Chibwe and Jones.15 It involves the reaction of a mixed metal oxide (e.g MgGAl3.309.3(OH)3.3) with an organo anion in aqueous media (at 100 °C). The mixed metal oxide was obtained from the thermal decomposition of the carbonate form of hydrotalcite at temperatures higher than 450 °C. The products obtained by the reaction of the mixed metal oxide with the organo anion were poorly ordered LDHs containing impure phases in some cases. The need for a general and convenient method for the synthesis of organic anion LDHs is obvious from the limitations of the existing methods. None of them provides a generally reliable route to pure crystalline phases. Depending on the nature of the LDH host and the organo anion, reaction products often are obtained that are poorly 25 ordered, amorphous, or mixed crystalline phases. These later materials are not well suited for the synthesis of regularly microporous POM-pillared derivatives by topotactic ion exchange, because the defects present in the starting material also will appear in the final product. Thus it is very' important to develop synthetic methods that afford pure, regularly crystalline organo anion LDH derivatives for use as precursors to pillared LDH materials. Here we report a general and exceptionally convenient route to well-ordered, hydrotalcite-like LDHs interlayered by carboxylate and other organo anions.9 Our approach utilizes an acid. base reaction between an LDH hydroxide aqueous suspension and an organic acid in the presence of a swelling agent ( glycerol ) to enhance the accessibility of the intracrystalline gallery surfaces of the LDH for topotactic reaction (eq.1). It is also possible to replace OH' by simple ion exchange (eq.2). In both cases glycerol is present. [Mg3Al(OH)8]-OH + HOA ------- > [Mg3Al(OH)8]-OA + H20 (1) [Mg3Al (OH)8]'OH + OA- ----- > [Mg3A1(OH)8]-OA + OH- (2) To our knowledge this is the first reported use of a swelling agent for the preparation of well-ordered crystalline LDH phases. A new route for the synthesis of the starting meixnerite [Mg3Al (OH) 8] [OH] '2H20, is also discussed. This improved synthesis of meixnerite avoids undesirable conditions of temperature and pressure used in the previous methods.16'17 26 It is remarkable that in our method of organo LDH synthesis according to eq.1, we can use organic acids in equimolar ratio to the ABC of LDH—OH to produce well-ordered products that exhibit xrd patterns with multiple 001 harmonics. Since it is an acid-base reaction, no washing of the electrolyte is necessary and the product can be used as prepared for subsequent reactions. Thus the acid-base reaction (eq.1) is preferred over the ion exchange reaction (eq.2). B. EXPERIMENTAL I. Materials and Physical Measurements The solvents and organic acids or salts used in this study were purchased. from .Aldrich Company and they were used without further purification. Although the materials are not oxygen sensitive, manipulations were carried under nitrogen atmosphere to avoid carbonate contamination of the basic solutions. Elemental analysis was done by Galbraith Laboratories. The carbonate form of hydrotalcite was prepared by modifying the co-precipitation method of Reichle.11 According to our procedure 25.6 g of Mg(NO3)2'6H20 and 12.4 g of A1(NO3)3'9H20 were dissolved in 100 ml water. This solution was slowly added to 100 ml solution containing 14 g of NaOH and 12.5 g of Na2CO3. The pH was brought down to 10 and the resulting gel was heated at 65 °C overnight. Then it was washed and air dried. The product was analyzed as [Mg3Al(OH)8][CO3]0.5'2H20, (elemental and thermal analysis). X-ray powder diffraction were recorded for specimens 27 deposited on glass slides with a preferred 001 orientation. The Rigaku diffractometer was equipped with DMAXB software and Ni-filtered x-ray Cu Ka radiation. Quartz was used as a standard. The anode tube was operating at 45 RV and 70 MA and the samples were scanned from 2 to 45 degrees 2' in 0.02 steps with variable counting time. II. New synthesis of meixnerite [Mg3A1(OB)3]°[OH]'ZBZO A 4.0 9 quantity of Synthetic hydrotalcite, [Mg3Al(OH)8][CO3]0.5°2H20, was heated to 500°C 511:3 quartz tube ( 2 inches in diameter ) under nitrogen atmosphere for 12 hrs. The resulting solid solution, M90.667A10.222[ ]0.1110, was then suspended in decarbonated water by stirred for several hours to form a suspension of synthetic meixnerite, [Mg3A1(OH)8]'[OH]°2H20. This slurry of the oxide was approximately 1 wt 96 and the time required for the hydrolysis reaction depends on the temperature. Up to 30 h was allowed for reaction at room temperature, or 3 h at 75 °C. Since the hydrolysis reaction produces meixnerite in essentially quantitative yield, the resulting suspension can be used directly for subsequent ion exchange or acid-base reaction. Also, the product can be freezed dried carefully in order to minimize exposure to atmospheric C02, and then kept in Closed containers. 28 III. Synthesis of [Mg3Al(OB)8]-Organo-Anion intercalation compounds. A 5 ml portion of a 1.0 wt % meixnerite suspension was mixed with 10 ml glycerol. To the resulting suspension was added the organic acid in equivalent amount to the anion exchange capacity of the meixnerite. The minimum amount of the appropriate solvent (usually 1 ml) was used to dissolve the acid. After stirring the suspension for about 30 min at room temperature, the resulting material was centrifuged, washed and dried at 120 C’C. Characterization was done by XRD, elemental analysis and FTIR spectroscopy. Ion exchange reactions were best carried out in the presence of 50 % (or more) excess of the anion and appear to be limited to long chain organic anions such as nonanoate or octanoate and sulfonated pthalocyanines. C. RESULTS AND DISCUSSION I. New synthesis of meixnerite: A convenient route for the synthesis of meixnerite, [Mg3Al(OH)8][OH]'2H20, has been discovered. As described in the experimental part a NaCl-type magnesium aluminum oxide solid solution results from the heating of [Mg3Al(OH)8][CO3]0.5'2H20 at 500 °C. This conversion was complete in about three hours but heating was continued for 12 hrs in order to remove the carbonate anion completely. Hydrolysis of the mixed oxide in decarbonated water afforded 29 meixnerite in good yield. We must emphasize that this procedure has the advantage of avoiding centrifuging the LDH and. 'therefore facilitates jpreparing large quantities completely free of carbonate contamination (less than 0.4 % by weight as determined by elemental analysis). The meixnerite was Characterized by FTIR spectroscopy, X-ray powder diffracion, thermogravimetric and elemental analysis. Elemental analysis for Mg and Al by dirrect current plasma emmision spectroscopy gave a Mg/Al ratio of aproximately three to one. The basal spacing was approximately 7.81 A (Figure 1) which is in good agreement with that reported before. Although the starting hydrotalcite was poorly crystallized, the hydrothermally synthesized meixnerite is very crystalline. TGA data shown in Figure 2A and 28 support further the formation of meixnerite. While the hydrotalcite (Figure 2A) shows a continuous loss of both surface and interlamellar water upon heating below 250°C, the -OH (Figure 28) shows some discontinuity. That is because the carbonate is removed after heating above 220 oC, whereas in the case of meixnerite some of the mobile OH’ might be removed at lower T since very strong hydrogen bonding exists in the galleries and H20 is linked to OH" almost indistiquishably lJlii water-like environment. Finally, the FTIR spectra show that there is some CO32' contamination at 1375 cm"1 due to grinding in order to make the KBr pellet. 2.. The grinding causes substitution of CO3 for OH" according to eq : C02 + ZOH’ ———————> C032" + H20 30 7.81 INTENSITY 3.90 * V r r T r I U r U 1 2 6 1'0 11'4 1'3 22 26 30 DEGREES 2 THETA Figure 11.1. Preferred orientation XRD pattern (Cu-Ka) and d-spacings (A) of meixnerite, [Mg3Al(OH)8]'[OH]°2H20. 31 8+ Weight (mgs) 7i 10-i Weight (rags) 6 ‘\ 25 150 255 370 455 500 Temperature (°C) Figure 11.2. ‘SA curves for : (A) [Mg3Al(OH)8],[co3]o.5,252e (B) [Mg3Al(OH)8]°[OH]°2H20. 32 II. Swelling properties of Meixnerite at ambient conditions: For the first time we have found that meixnerite can be easily swollen to a gallery height of approximately 10 A in the presence of organic solvents such as ethers or glycerol. Figures 3 and 4 illustrate the swelling possible using glycerol as a swelling agent. The reaction is very facile, but other forms such as [Mg3Al(OH)8]°Cl don't swell on the same time scale (i.e within an hour). This can be shown by placing solid meixnerite on a glass slide and then expose it carefully under nitrogen to a 2:1 (v/v) glycerol/water mixture. After one day at room temperature we observe a decrease in the intensity of the 001 reflection of the meixnerite and the appearence of a solvate phase at about 14.3 A (Figure BB). Further aging causes the LDH-hydroxide phase to disappear while the 14.3 A phase increases in intensity with time (Figure 3C, 3B, 3B, 4A and 48). If we further expose the sample to very small amount of glycerol/water mixture (2/1) both peaks disappear (Figure 4C). This suggests that the product is either completely dispersed in glycerol/water or that interstratification of the layers occurs with the disappearence of all the reflections. These results ‘unequivocally' demonstrate that glycerol swells the hydroxide interlayers of the LDH forming a stable dispersion. Meixnerite whidh has be freshly synthesized hydrothermally and not dried to a powder is much more reactive to swelling than the air-dried product. In the same 2:1 (v/v) 7.81 A o o 0) \ U) 4.0 c: a o 0 V E I (I) Z is a _2_ DJ 2 c 3 Lu 0: 14.3 7,15 E T F r l 6 s Th 1V2 14 DEGREES 2 THE-TA - l'igure II.3. XRD patterns (Cu-Kc) and d-spacings (A) of oriented film samples : (A) synthetic meixnerite, (B) the intercalate obtained after exposure of sample A to a small amount of 2:1 (VYV) glycerol/water mixture, under nitrogen for one day, (C) sample B kept under nitrogen for an additional day, (D) sample B after three days, (E) sample B kept for four days under nitrogen. 34 A (J a) m \ '9 7 81 (g; 14.3 ‘ O A V a 14.3 1.21 l— 7.15 2; E3 LL] 2 3 LIJ C D: r’ 7’ r r 2 5 8 11 14 DEGREES 2 THETA Figure 11.4. XRD patterns (Cu-1(a) and d-spacings (A) of oriented film samples : (A) [Mg3Al(OH)8]°[OH]'2H20 exposed to a small amount of 2:1 (v/v) glycerol/water mixture, under nitrogen, for 4 days, (B) sample A after one day, (C) sample A after exposure to a very small mixture of 2:1 (v/v) glycerol/water, under nitrogen, for 1 day. 35 glycerol/water media the material swells ixxaa much shorter time scale (i.e. within minutes). It can be done easily by simply suspending the meixnerite solution in the appropriate solvent. When the resulting dispersion was placed on a glass slide under vacumm for several days, the xrd pattern showed two phases with basal spacings of 14.3 A and 7.8 A (Figure 5). Thus the exfoliated LDH platelets can be reassembled into partially solvated and unsolvated face-face aggregates. III. Reaction of meixnerite with glycerol. If a suspension of meixnerite in 2:1 (v/v) glycerol/water is vigorously mixed a stable dispersion is formed. Drying the dispersion on a glass slide at 190 0C for about one hour affords 21 white precipitate which analyzes as [Mgz.89Al(OH)7.78]'[C5.17H18.38]. Based.cn1 that analysis a possible formula for the glycerolate would be [M92.89A1(0H)7.78]'[C3H803]l.72°[OH]x°[H20]2.31-x The product exhibits the x-ray diffraction pattern shown in Figure GB characteristic of a layered material. At this high temperature, reaction of glycerol with the LDH lattice hydroxyl groups occurs and a glycerolate derivative is produced. This most likely is a topochemical reaction in which glycerol most likely chelates the octahedral metal atoms in the brucite-like layer. A basal spacing of 14.1 A has been found for the glycerol-solvated product when it was flashed dried at 190 0C for 10 min, Figure 6C. This latter product most likely is meixnerite solvated by two molecular 36 4) 1200- fiaooJ z u.) g.— .2. 4) 4ooJ J o .,.,.,...,.r. 2 6101418 22 26 30 DEGREES 2 THEI’A Figure 11.5. XRD patterns (Cu-1(a) and d-spacings (A) of the material that results from the dispersion of aqueous suspension of [Mg3Al(OH)8]°[OH]'21—I20 in glycerol (2:1 v/v glycerol/water) and subsequent evaporation of the solvents, under vacuum, on a glass slide. 37 7.31 9.48 (7) z 3.90 E 4.72 2.60 J A L a . 4.70 + c T_ rummmm ..u I I 2 E 61'11419252326293235384144 2 THETA l'igure 11.6. XRD patterns (Cu-1(a) and. d-spacings (A) of oriented film samples : (A) [Mg3Al(OH)8]°[OH]‘2H20 (B) [Mg3Al (01-1) 8] ' [0H] °2H20 dispersed in a 2:1 (v/v) glycerol/water mixture and dried at 190 °C for 1 hr, (C) [Mg3Al (OH) 8] ' [OH] °2H20 dispersed in a 2:1 (v/v) glycerol/water mixture and flashed dried at 190 °C. 38 layers of glycerol. TEM images suggest that the reaction is not a topotactic one, Figure 7, top left. 1V. Organo Anion Intercalation Compounds The initial experiments were focused on the reactivity of hydrotalcite-like [Mg3Al]-LDHs interlayered by simple inorganic anions (Cl‘, N03‘, OH' ) toward carboxylate anions in the presence of various alcohols, ethers, polyalcohols and polyethers as potential swelling agents. The synthetic analogue of meixnerite, [Mg3Al (OH) 8] ° [OH] ' 21120, in combination with glycerol as a swelling agent proved to be an especially effective reaction system for the production of well-ordered organo anion phases. Two general reactions could be accomplished: topotactic acid-base reaction of the free organic acid (e.q. l) and ion exchange (e.q. 2) : [Mg3Al]-OH + l/n HnA ---->,[Mg3Al][A]1/n + H20 (e.g. 1) [Mg3A1]-OH + l/n 8’" ----> [Mg3Al][B]1/n + on' (e.g. 2) The former reaction, when carried out in stoichiometric quantities, is especially convenient as it requires no washing or dialysis to remove electrolyte, byproduct. The ion exchange reactions are limited to specific classes of compounds such as nonanoate anion or sulfonated pthalocyanines. In the long-chain carboxylate anions strong Van der Walls forces exist among the alkyl chains and more than the stoichiometric amount intercalates. For the nonanoate it was found that 20 % excess of the anion-with respect to the anion exchange capacity of the LDH-was intercalated. In the sulfonated derivatives strong 39 Figure 11.7. TEM images of s the product formed by the reaction of meixnerite with glycerol at 190 °C, (top left); meixnerite, (top right); [Mg3Al(OH)3][C5H804]o.5, (bottom left); and [Mg3Al(OH)3][p-CH3C6H4SO3], (bottom right). 41 interactions between the interlamellar hydroxides and the sulfonate groups exist. In contrast, the acid-base reactions occur quantitatively on 1/1 molar ratio. In general, the acid-base reaction is experimentally more convenient, because the crystalline intercalate can be recovered with no need for subsequent purification. Free acids that are sparingly water-soluble (i.e. adipic) may be introduced to the reaction mixture as solids, or they can be dissolved in the minimum amount of a suitable solvent (e.g. acetone). If a suspension of an LDH chloride, eg. [Mg3Al(OH)8]°Cl in glycerol is used instead of the hydroxide intercalate (i.e., meixnerite) no expansion of the layers is observed. When the pH of the LDH-Cl solution was made approximately neutral by the adition of .ArSOBH, a product containing multiple phases resulted, as shown by the difraction pattern in Figure 8. The moles of the ArSO3- added were at least equal to the ABC of the LDH and the reaction can be carried out in glycerol. Under the same conditions, the LDH—OH exchanges completely, to form the CH3C6H4SO3’ intercalate. The LDH-OH appears to be more reactive. Thus, the presence of OH" seems necessary for the synthesis of the following expanded hydrotalcites: (a) . [Mg3Al (OH) 8] [C6380410 . 5 . 2320 . A 1 ml portion of a 1.0 wt% [Mg3Al]-OH suspension in H20 was mixed with 2 ml glycerol. To that solution was added adipic acid in 50% excess of the anion exchange capacity of the LDH 42 1200 1253 800- 1£i8 INTENSITY U T T t— I V 2 6 1‘0 1‘4 1'6 2'2 26 30 DEGREES 2 THETA Figure 11.8. Preferred orientation XRD pattern (Cu-Ka) and d-spacings (A) of the material that results from titration of [Mg3Al(OH)8]Cl with p-toluenesulfonic acid until neutral pH is obtained. 43 The adipic acid had been previously dissolved in 1 ml acetone. The precipitate was centrifuged resuspended in water and dried 20 min at 120 °C. The basal spacing was 14.4 A (Figure 9A). (b) . [Mg3A1(OB)8] [C33C634'SO3] .2320 A 1 ml portion of a [Mg3Al]-OH suspension in water (1.0 wt%) was mixed with 2 ml glycerol. To the resulting suspension was added equimolar amount of p-toluenesulfonic acid in 1 ml H20. After centrifugation, resuspension in water and drying at 120 0C, the product exhibited sharp 001 reflections with basal spacing 17.5 A Figure 9B. The product obtained in the absence of a swelling agent exhibits the XRD pattern given in Figure 8. TEM images of meixnerite and its adipate and p- toluenesulfonate intercalates are shown in Figures 78, 7C and 70. It thus appears that the ractions are topotactic, i.e., no change in the crystal structure, morphology or layer composition is observed. (c) . [M93A1(OH)3] [04041031320 The synthesis of this squarate intercalate was analogous to that described. for the adipate intercalate. However, an equivalent amount of squaric acid was used. The resulting product had a basal spacing of 10.00 Figure 10A, which is in agreement with an orientation of the ring perpendicular to the layers. If an aqueous suspension of HT-OH was used in 44 17.5 14.4 RELATIVE INTENSITY 3-53 4.32 d I 5.80 3.46 B 7.22 4.78 I 3.59 A i ' “T—I I If r r r r 6 10 1'4 1'6 22 26 3'0 DEGREES 2 THETA Figure 11.9. Preferred orientation XRD patterns (Cu-Ka) and d-spacings (A) ' z (A) [Mg3Al(OH)8][C6H804]O.5, (B) [M93A1(OH) 8] [P‘CH3C6H4SO3] . 45 the absence of a swelling agent, the final product showed multiple reflections due to presence of several phases. These phases may result from multiple orientations of the ring Figure 103. (d). [MgaAll-nonanoate. This compound can be synthesized by ion exchange of [Mg3Al]- OH with the sodium salt of nonanoic acid. A 10 ml portion of a 1 wt % meixnerite suspension was mixed with 20 ml glycerol. It was then. brought in contact with. 4 ml of aqueous solution of sodium nonanoate (the concentration of nonanoate was in 50% excess of the AEC of the LDH). The product was collected by centrifugation, washed and dried at 120 °C for 1 hr. Its basal spacing was approximately 21 A (Figure 11A). No ion exchange was observed when the reaction was carried out in the absence of swelling agent (Figure 118). (e). [Mg3AIJ-Ni(11) phthalocyaninetetrasulfonate This intercalate was synthesized by a procedure analogous to the one described for the nonanoate intercalate [Mg3Al]—OH. The same amounts of meixnerite and glycerol were used, but the Ni(II) phthalocyaninetetrasulfonate anion was in 100% excess of the ABC of the LDH. The basal spacing is 21.9 A (Figure 12) and corresponds to a perpendicular orientation of the phthalocyanine rings with respect to the LDH layers. This material is important since such a stacking of the 46 10.12 a 2: 131' E 5.0 11.1 A Z '3 32‘ 10.35 5.21 B ' ' I . 1 r l . ' 1 I 2 6 10 14 18 22 26 30 DEGREES 2 THETA Figure 11.10. Preferred. orientation XRD patterns (Cu-Ka) and d-spacings (A) of [Mg3Al(OH)8][C4O4]0 5, obtained by the reaction of meixnerite and squaric acid, (A) in the presence of 2:1 glycerolzwater (v/v), (B) in the absence of glycerol. 47 20.9 RELATIVE INTENSITY r 2 6 10 14 1E 22 26 3'0 DEGREES 2 THETA Figure 11.11. Preferred. orientation. XRD jpatterns (Cu-Kc) and d-spacings (A) of [Mg3Al(OH)8][C931702]1.2, obtained by the reaction of meixnerite and Na-nonanoate, (A) in the presence of 2:1 glycerolzwater (v/v), (B) in the absence of glycerol. 48 .msmsmoumucs 006608HammuumsmcscmsoosmconaAHHVn21.mimovsamczi no 2: 065060610 0:6 280728 53060 max Nada 008$ <._.MI._. N mmmmomo on .vN mp up w o _ 4 _ .. h . .h . .C . Au 3: f. . _ fl 100w Ni. ” I 2 109V .n.P~u AllSNBiNI 49 rings is responsible for the observation of conductivity in oxidized phthalocyanines. It is important to emphasize that the organo intercalates reported here are more ordered than those obtained by any of the procedures noted earlier. For instance, Drezdzon has reported that the co-precipitation of Mg/Al LDHs in the presence of the adipate leads to x-ray amorphous products.8 However, the reaction of adipic acid with meixnerite in the presence of glycerol yields a well-ordered intercalate with several orders of 001 reflections. Also, hydrothermal reaction of M90.667A10.222[ ]0.1110 metal oxides with p- tolunesulfonic acid gives products that are mixed phases, as shown by the diffraction pattern in Figure 13C. We repeated their experiment and the resulting material is also less crystalline compared to the one obtained with our procedure, as shown by comparison to the diffraction pattern in Figure 13B. It thus appears that our method is more powerful than any previous one. The ability of glycerol to promote the formation of well- ordered LDH phases most likely is related to the increased accessibility of the gallery surfaces occupied by the hydroxide exchange anion. By swelling the galleries, one facilitates the complete reaction of the initially intercalated anion, as well as the ordering of the newly intercalated guest. Also, swelling favors the desired intercalative reactions over competitive side reactions (e.g., layer hydroxide displacements and metal ion 50 17.5 7.61 "m- 3.90 z LLJ '2 A — J L, -1... __ LLJ ,->_- 8.68 4.32 5 5.80 B LIJ 0: 17.8 7 89 3.91 8.68 I 4.32 C I T F . r ' r r l f F 2 6 14 20 26 32 DEGREES 2 TH ETA Figure 11.13. XRD patterns (Cu-1(a) and d-spacings (A) of oriented film samples : (A) [Mg3Al(OH)8]°[OH]°2H20, (B) the product obtained from the reaction of meixnerite with p- toluenesulfonic acid in the presence of a 2:1 glycerol/water mixture, (C) the product obtained from the hydrothermal reaction of calcined hydrotalcite and p-toluenesulfonic acid in aqueous media. 51 complexation reactions) that can lead to poorly-ordered defect structures or to amorphous products. Our route seems to be the fastest and simpler known to date and can be expanded to include the synthesis of a large variety of organic anion intercalates. Table I provides the basal spacings for typical [Mg3Al]- carboxylate and -sulfonate intercalates, along with a new squarate derivative with An—=[C4O4]2'. Included in the table is the size of the scattering domain as determined from the Scherrer equation, d 0.9 x l/[Dcose], where l is the x-ray wavelength in A, 0 is the width at half-amplitude of the 001 -92 diffraction line (D2 =02 t since both observed and obs ins instrumental width should be taken into account), and 8 is the diffraction angle in radians. It is very interesting that D is inversly proportional to the size of the crystallites and. 'therefore tx> their orientation perpendicular to the layers. The instumental width was determined by using Mica as a standard. We didn't find deviations on the numbers when quartz was used as a standard. Anion basal spacing,A domain size,a A [C40412' 10.1 171 c6042‘ 14.4 203 CH3C6H4-SO3' 17.4 197 CH3ICH2)8C02' 21.2 101 aThe size of the scattering domain along the c-axis was determined from the width of the second or third 001 X-ray harmonic of oriented film samples. 53 TABLE II.II. Elemental Analysis Results of [Mg3Al(OH)8]- Organo Anion Intercalation Compounds. Mg Al C H (wt %) 24.92 9 37 6 64 2.96 a 24.62 9.33 6 3 2.73 b 23.56 8.51 9.18 3.79 23.66 6.76 11.69 3.69 17.03 6.30 31.39 6.66 16.56 6.67 27.49 6.36 17.19 6.35 30.54 6.69 D. A These Calcd Calcd Calcd are the values (wt %) determined experimentally. values based on the formula [Mg3Al(OH)8][C4O4]0.5. values based on the formula [Mg3Al(OH)8][C6O4]O.5° values based on the formula [Mg3Al(OH)8][CH3(CH2)8C02]. Calcd values based on the formula [Mg3Al(OH)8][CH3(CH2)8C02]1.2. SUMMER! facile route for the synthesis of well-ordered [Mg3Al(OH)8]-organic anion intercalation compounds, 54 particularly these containing carboxylate and sulfonate anions, is described. The method is based on the reaction of synthetic meixnerite form of LDHs, [Mg3Al(OH)8]'[OH]°2H20, with the free acid form of the desired anion in the presence of glycerol as a swelling agent. The swelling agent greatly facilitates access to the LDH basal surfaces and promotes formation of pure crystalline products that are not readily achieved by conventional synthetic methods. Reactions of the type: n[Mg3Al(OH)8]-X (X=OH or C1”) + HnA (acid) ------- > ----- > n[Mg3Al(OH)8][A]1/n + nx' + nH+ were studied in the presense or absence of organic solvents as swelling agents. XRD patterns, chemical analysis, FTIR spectroscopy and TEM images of the reaction products suggest that the reactions are topotactic. The reaction of [Mg3Al(OH)8]'[OH]°2H20 and glycerol however, does not appear to be topotactic at temperatures above room ‘temperature. A. new synthesis of meixnerite, [Mg3Al(OH)8]°[OH]°2H20, is also described. It is based on the calcination of hydrotalcite, [Mg3Al(OH)8]'[CO3]0.5°2H20, and aqueous hydrolysis of the resulting mixed metal oxide, M90.667A10.222[ ]0.1110, to form a meixnerite suspension. LIST OF REFERENCES 10. 11. 12. 13. LIST OF REFERENCES Herman, R.G., Klier, K., Simons, G.W., Finn, B.P., Bulko, J.B., Kobylinski, T.P., J. Catal., 1979, S6, 407. Reichle, W.T., J. Catal., 1980, 28, 50. Miyata, S., Kuroda, M., Ger. Offen., 2,950,489 Chem. Abstr., 19xx, 93, 96250a. Martin, K.J., Pinnavaia, T.J., J. Am. Chem. Soc., 1986, 108, 541. Pinnavaia, T.J.; Rameswaran, M.; Dimotakis, E.D.; Giannelis, E.P.; Rightor, R.G., Faraday Discuss. Chem. Soc. London, 1989, 87, 27. Kwon, T.; Tsigdinos, G.A.; Pinnavaia, T.J., J. Amer. Chem. Soc. 1988, 110, 3653. Kwon, T.; Pinnavaia, T.J., Chem. Mater. 1989, 1, 381. Drezdzon, M.A., Inorg. Chem. 1988, 27, 4628. Dimotakis, E.D.; Pinnavaia, T.J., Inorg. Chem. 1990, 29, 2393. Miyata, S., Chem. Lett., 1973, 483. a) Reichle, W.T., J. Catal., 1985, 54, 547. b) Reichle, W.T., Chemtech, 58 Jan 1986. c) Reichle, W.T., J. Catal., 1986, 101, 352. Bish, D.L., Bull. Mineral., 1980, 103, 170. Dutta, P.K., Puri K., J. Phys. Chem., 1984, 93, 381. 55 14. 15. 16. 17. 56 Park, I., Kuroda, K., Kato, C., Chem. Lett., 1989, 2057. Chibwe, K., Jones, W., J. Chem. Soc. Chem. Commun., 1989, 926. Masolo, G., Marino, 0., Mineral. Mag., 1980, 43, 619. Paush, I., Lohse, H., Schurmann, K., Allmann, R., Clays and Clay Miner., 1986, 34, 5, 507. CHAPTER III Synthesis of well-ordered layered double hydroxides pillared by Keggin-like polyoxometalates (pone) . Use of meixnerite, '[Mg3Al (0mg) [06] '2320, as a precursor for topotactic ion exchange. A. INTRODUCTION Anionic clays or layered double hydroxides1 (LDHs) are complementary to smectite clays insofar as their layers are positively charged rather than negatively charged. LDH structures are based on that of brucite, where the layers are formed by edge sharing of M(OH)6 octahedra. They have the general formula [MIII_XMIIIX(OH)2][An“]X/n*yH20 (I), where Mn, MIII are divalent and trivalent metal cations respectively, A“- is an inorganic anion such as Cl", NO3’, C032- etc. and .15roduct was collected after 30 min and treated as usual. IElemental analysis for [M93A110H)3]5132W1204o] 65 Calcd (%) Mg, 10.26; Al, 3.79; W, 51.70. Found (%) Mg, 8.52; Al, 3.32; W, 45.53. Suggested formula: [Mg2 . 84Al (OH) 7 . 6915 . 96 [H2W12040] E){1\MPLE # 3 : [Mg3Al(OH)8]8[SiW11039] FROM LDH-ADIPATE LDH-adipate was prepared as described earlier. The reaction of meixnerite with adipic acid, in the presence of glycerol, lasted 30 min. The suspension was added to 335.6 mg (0.825 meq, 50% excess) of K8[SiW11039] in 4 ml of water. The solid was collected after 30 min, washed and characterized as in previous examples. Elemental analysis for [Mg3A1(OH)8]8[SiW11039] Calcd (%) Mg, 12.79; Al, 4.73; W, 44.33; Si, 0.53. Found (%) Mg, 11.57; Al, 4.34; W, 40.07; Si, 1.02. Suggested formula: [M92 . 96A]. (OH) 7 . 9218 . 12 [SiW11039] EXAMPLE # 4 : [Mg3Al(OH)8]8[SiW11039] FROM LDH-0802Ar Ji-JDH-p-toluene-sulfonate was prepared as in example #2. After 30 min of reaction it was added to 335.6 mg (.825 meq) of K8[SiW11039] in 4 ml of water. The product was collected after 30 min, washed and characterized with the above mentioned techiques. Elemental analysis for [Mg3Al(OH)8]8[SiW11039] czalcd (%) Mg, 12.79; Al, 4.73; w, 44.33; Si, 0.53. I?ound (%) Mg, 11.22; Al, 4.15; w, 38.96; Si, 1.15. Suggested formula: [Mg3Al(OH)8]8[SiW11039] C. RESULTS AND DISCUSSION: I - LDH-POMS PROM LDH-OH 8 LDH-GLYCEROLATES The hydroxide form of an LDH exchanges polyoxometalates to give well-ordered pillared derivatives, as judged by tl1£air multiple 001 reflections in the XRD patterns, shown in Figure 1. The XRD patterns of LDH-H2W12040 are identical regardless of how the ion-exchange reactions are carried out, i.e. in the presence (top) or absence (bottom) of glycerol. Both products are pure phases and have very similar gallery heights (10 A), that are in good agreement Vii: h the reported crystallographic radii of polyoxometalates (10 A).20 We also examined other pillared products, i.e. LDH-BVW11040 and LDH-SiV3W9040 , and they all show very Similar xrd patterns and d-spacings. The topotactic nature of these pillaring reactions was Confirmed by TEM, in addition to XRD and FTIR. Thus c2<>mparing the pictures in Figure 2 we see that the inmedxnerite crystals remain hexagonal upon intercalation of 3E12W120406‘ and their size is the same (".1 gm). Close examination of the FTIR spectra of the LDH-POMs EBhow that they are identical to that of the starting alkali Ebolyoxometalates.21 There is no evidence for hydrolysis. For fiLnstance, in figure 3 : LDH-KGBVW11040 and KGBVW11040 iexhibit bands at 1008 (B-O), 953 (W=O), 915 (W=0) and 820 (W-O-W) cm‘l , which shows that the metalate framework stays 67 14.6 Q 7.37 4.66 If g A 3.66 2.93 g E- 14.8 IJJ [K 4.66 7.37 B ) 3.66 2.93 .T r F 1 ' r r r r F 2 6 14 20 26 32 DEGREES 2 THETA Figure 111.1. Preferred orientation XRD patterns (Cu-K0,) and d-spacings (A) of [Mg3Al(OH)8]6[H2W12040] pillared products, derived from the reaction of [Mg3Al(OH)8] ° [OH] and (NH4) 6H2W12040 : (A) prepared in the presence of 2:1 glycerol:water (v/v), (B) prepared in the absence of glycerol. 68 Figure III.2. Represantative TEM images of : [Mg3Al(OH)8]5- [H2W12040] derived from meixnerite in pure water, (tOp left); [Mg3Al(OH)3]5[H2W12040] derived from meixnerite in the presence of 2:1 glycerol:H20 (v/v), (top right); [Mg3Al(OH)3]6[H2W1204o] derived from the adipate precursor, (bottom left); meixnerite, (bottom right). 70 A L1J o 2 <1: :11 a. (A) o m m < A LLI 2 g 1 B LLI CK A wfg-w 915 953W=0 o 1” r 1 1200 1000 600 600 400 WAVENUMBER (cm—1) Figure 111.3. FTIR spectra of [va1104016‘, (A) as [Mg3Al(OH)8]-intercalate and (B) as potassium salt. 71 intact upon intercalation. For a series of SinZ-Siwll-Siwg polyoxometalate anions, there is a decrease in the W-O-W frequencies in that sequence since there is a decrease in the coherency of the metal framework in the same order.22 We did not observe such a shift. Additional information from 51V NMR supports our observation.23'24 There is a single peak at -476 ppm (Figure 4A) (versus VOC13) for LDH-BVW1104O. For the LDH-SiV3W9040 the 51V NMR shows a single peak at -567 ppm (Figure 4B) and the 29Si MAS NMR one peak at -84 ppm (versus TMS) (Figure 4C). Since the corresponding salts have the same shifts there is no change in the POM structure upon intercalation. It is very interesting to note that although the reaction of LDH and POM requires the polyoxometalate to contact. a basic environment, hydrolysis does noyt occur contrary to what is known from its solution chemistry.25 POMs are generally stable in acidic-neutral solution. Ligation to organometallics is known to stabilize them in basic solution.26 It has been previously mentioned that hydrogen bonding is important for an exhange reaction.13 It thus appears that the stabilization towards hydrolysis in the hydrotalcite galleries arises from hydrogen bonding. The fact that excess of protons exists in acidic POM solution helps since titration of the the interlayer hydroxide anion- which. can. be accessed. easier as intercalation procceds- facilitates the exchange reaction. The presence of the swelling agent can be important in some cases (i.e., in very 72 ~476 ppm T1902. 111.411. 51v Solution NMR of [svw1104015' after dissolution of [Mg3Al(OH)3]6[BVW1104o] in LiClO4 (pH=1.7). 73 -567 ppm J . 1 A -500 scale: 50 ppm Figure 111.48. 51V Solution NMR of [SiV3W904OJG‘ after dissolution Of [Mg3Al (OH) 817 [31173919040] in LiClO4 IPH=1.7) . 74 -84 ppm J l l V l l 0 scale: 30 ppm Figure 111.4c. 29Si MAS-NMR of [Mg3Al(OH)8}6[SiV3W9O40]. 75 acidic POM solutions), since it exposes the basal planes of the LDH and facilitates ion exchange or acid/base reaction of the gallery anion. This decreases the competitive acid/base reaction of the hydroxides of the brucite layer which can lead to degradation of the host layer. It is possible that reaction of OH' at the edges of the layer will cause the POM to bind there. The effect that this might have in the BET surface area is rationalized below. 11. LDH-POMS PROM LDH-0A LDHs intercalated by organic anions (eg., carboxylates, sulfonates, among others) are attractive precursors for topotactic ion exchange reaction by a variety of POMs. The resulting pillared products are well ordered, pure phases and have XRD patterns similar to the ones obtained when LDH- OH was used as the precursor (Figures 5 and 6). However, they have different microporosity as will be seen below. TEM images of the starting organo anion precursors (i.e. LDH-adipate, Figure 7, top left) and of the resulting microporous derivatives (i.e. LDH-SiW11039, Figure 7, bottom left or top right) support the topotactic nature of these ion exchange reactions again. Comparing the FTIR spectra of [Mg3Al(OH)8]-LDHs pillared with the same lacunary anion [SiW11039]8‘ (Figure 8A and 83), starting from two different precursors ( [Mg3Al(OH)8]-02C(CH2)4C02 and [Mg3Al(OH)8]-0502C6H4CH3 )7 76 14.4 j E?)— 2 1.1.1 l._. 3 A 72. 4. 76 0.1 3. 59;: _>_ J K I12 A -f 3 14.6 4.66 1.11 7.36 0C B 3.65 2 93 14.6 4-85 2 6 14 2‘0 26 32 DEGREES 2 THETA Figure 111.5. Preferred orientation XRD patterns (Cu-Ka) and d-spacings (A) for [Mg3A1(OH)8]-intercalates: (A) [Mg3Al (on) 8] [02c1c1121 4c0210 _5 (B) [Mg3Al (OH) 8] 6 [H2W12040] (C) [Mg3Al(OH)3]8[SiW11039]. Both B and C intercalates were derived from adipate precursor. 77 RELATIVE INTENSITY 3.69 2.92 2 6 14 21:1 26 32 DEGREES 2 THETA Figure 111.6. Preferred. orientation. XRD jpatterns (Cu-Ka) and d-spacings (A) for [Mg3Al (OH) 8] -intercalates: (A) [M93A110H) 3] [P-CH3C6H4SO3-I (B) [M93A110H13161H2W1204o] (C) [Mg3Al(OH)8]3[SiW11039]. Both B and C intercalates were derived from -p-toluenesulfonate precursor. 78 Figure 111 . 7 . Represantative TEM images of : [Mg3Al(OH) 8] [02C(CH2)4C02]0.5 , (top left); [Mg3Al(OH)3][SiW11039] derived from the previous adipate precursor, (bottom left); [Mg3Al(OH) 8] [SiW11039] derived from the adipate precursor and then treated with Na-adipate, (top right) ; [Mg3Al(OH)8][p-CH3C6H4SO3-], (bottom right). 80 A I LLI 0 Z <( (I) (I. 0 U7 (I) <1: 1..) B ._>_ E L1J o: T T 909 A IVVILWV I 956 1002W=0 Si-O Q T F 1 1200 1000 800 600 400 WAVENUMBER (cm— 1) Figure 111.8 FTIR spectra of [Mg3Al(OH)8]8[SiW11039] derived from : (A) [M93A1(OH)8][02C(CH2)4C02]O.5 precursor (B) [Mg3Al(OH)8][p-CH3C6H4SO3'] precursor. 81 we see that they are identical. The same POM frequencies are also found for the corresponding potassium salt, which suggests that there is no hydrolysis of the POM upon intercalation. Elemental analysis results give the same tungsten, magnesium and aluminum content for both samples. Finally, the XRD patterns of both products are identical and show five orders of 001 reflections in the region examined. All these strongly suggest that the metal framework of the polyoxometalate remains intact after intercalation in the host LDH lattice. 111..M1CROPOROSITY OF LDH-POMs A simple comparison of the shape of the nitrogen adsorption desorption isotherms (Figure 9) for the [Mg3Al(OH)8]-LDHs derivatives pillared by H2W120406' Keggin ions and SiW110398‘ lacunary Keggin ions, starting from the LDH-adipate, shows that the LDH pillared by the more highly charged anion is the most microporous one. A closer examination of the microporosity was obtained from the t- plots (Figure 10). In a trplot the larger slope gives the total surface area: Stot Smicro + S, where Smicro is the surface area due to microporous and S is the surface area due to all pores excluding micropores. Also, the microporous surface can be calculated from the intercept of the second line. The slope of this second line gives S. 82 120 . e ' [MSgAI(OH)3]3[SiW11039] 100~ ° [Mg3AI(OH)3]‘[H;W120“] . 4 e e 1 80-1 I . 4 I . 0 so 1 e «1E 606 I e. e 0 e > , n" " n . 0 ‘0 .l'asal'” ' a 0 o 0 ° . .00.... O 20 o 0 F I 1 I V 7 I ' I F 00 02 Q4 06 08 10 PIPe Figure 111.9. Adsorption-desorption isotherms of [Mg3Al(OH)8]8[SiW11039] (tOp) and [M93A1(OH)8]6[H2W120401 (bottom). Both pillared products were derived from the [Mg3Al(OH)3][02C(CH2)4C02]0.5 precursor. 83 A MAI(OH)'].ISAW‘!03’] 9 155 [112/3 11 [M33A1(OH)315[113W130“]. 107m2/ : slope~30 8888681892318 V ml.g"l 0 2 4 6 8 10 t (K) Figure 111.10. t-plots for nitrogen adsorption on pillared [MgaAl (OH) 6] -intercalates derived from adipate precursor : (A) [M93A11031319131W110391 (B) [M9331101'1IaIsIHzW120401 84 From the t-plot of the [Mg3Al(On)3]6[82W12040] intercalate: Stot = 107 mZ/g ( from the slope of the line that passes through the origin ) Smicro Vm x 4.37 = 77 m2/g ( from the intercept of the second line of the t-plot for H2W120406'. Vm is the microporous volume at the intercept and 4.37 m2 is the equivalent surface area of 1 ml of nitrogen ). S = 30 mZ/g ( from the slope of the same second line ) So by adding S o and S we find that Stot = Smicro + S. We micr also find that stot = SBET = 107 mz/g , where SBET is calculated by the three point method in the low relative pressure region ( 0.06

4 o\° . [M3Al]-SIW11039 r 1 _' 200 300 400 500 500 D (ii) Figure 111.11. Pore size distribution plot for [Mg3Al(OH)8]- intercalates : (A) [Mg3Al(OH)8]8[SiW11039] (B)[Mg3Al(OH)8]- [HZW1204011/6° 86 2 is the equivalent is the micropore volume and 4.37 m surface area of 1 ml of nitrogen ). S = 24 mz/g (from the slope of the same line) So stot = smicro + s = 131 + 24 = 155 m2/g Again Stot = SBET' where SBET is calculated by the three point method in the low relative pressure region. The pore size distribution (Figure 11) shows interpaticle pores of 100 A diameter. Comparing the slopes of the non-microporous region we see that they are almost identical (Figure 10) and correspond to an external s.a. of approximately 30 mZ/g. On the other hand, the microporous volumes differ significantly and the lacunary derivative evidently is the more microporous one. .As mentioned earlier, there are several indications that the pillaring POM anion remain intact upon intercalation into different organo anion LDH precursors. However the BET surface areas can differ for samples derived from different precursors (see Table I). For instance, [Mg3Al(OH)8]8[SiW11O39] LDHs gave 155 mZ/g for the pillared product obtained when the LDH-adipate was used as a precursor and 68 mZ/g for the one obtained when LDH—p— toluenesulfonate was the starting material (Table I). We propose that the differences in the observed microporosity arise from clogging of the galleries to various degrees, due to formation of MII or MIII-POMs at the edges of the host layer. This is happening to such an extent that although is 87 not detectable by XRD, it is enough to limit the microporosity of the products. In the case of the adipate precursor, which is a poor leaving group and a better chelating agent for the metals at the edges, precipitation of MII or MIII-POMs is minimized. Therefore little or no clogging occurs at the edges and the resulting materials have higher surface areas in comparison to the ones obtained using LDH-OSOZAr. When LDH-OH is used as a precursor the changes in the BET surface areas are more evident. If air- dried, the surface area is 150 mZ/g, but if the sample is dried at 120 0C the surface area is only 34 mz/g. The former s.a. is in very good agreement with the one obtained when the LDH-adipate was the starting material. Additional support comes from the fact that titration of [Mg3Al(OH)8]-SiW11039 (68 mZ/g), Prepared from [Mg3Al(OH)8]-OSOZAr, with l D! Na-adipate solution ( the equivalents of the adipate anion used was three times the equivalents of the polyoxometalate anion ) for 30 min resulted in a product with BET surface area of 133 mZ/g. Similarly, titration of LDH-SiW11039 (34 m2/g), prepared from the LDH-OH and dried at 120 0C, with Na-adipate solution results in product with BET surface area in the range of 104 mZ/g. It thus appears that the adipate complexes the MII,MIII—POM salts at the edges and the galleries become accessible for nitrogen adsorption. This mechanism is shown in Figure 12. Acidic POM solutions can cause partial 88 reaction of the edge hydroxyl groups and thus stacking of POMS there. However chelation of the MU, MIII by adipate anion can release the POM and cause the pores to open for nitrogen adsorption. The fact that the resulting surface areas are not equal to the expected one (155 mZ/g) could be due to the subsequent crystal digestion in Na-adipate solution or to irreversible changes at the edges. This conclusion is partially supported by TEM images. For example by comparing Figures 7B and 7C we see that the edges of the some of the crystals in Figure 7C are not well defined. Thus treatment with Na-adipate solution can cause little deviation from the hexagonal symmetry observed in Figure 7B. 89 . Proposed model for the observed differences Figure 111.12 in BET surface areas. 90 TABLE III.I. BET surface areas of LDH-POM products prepared from [Mig3Al(OB)3]-[XJ1/n precursors. n- - - a x 39!? DE¥12$ r Outgz?s. Tb 623.3.gi 06‘ 62w120406‘ 25 130 92 06‘ C H2W12040 25 130 63 on‘ vallo406‘ 120 125 64 06‘ " 25 130 110 06‘ Siv3w90407‘ 120 130 70 06‘ C " 25 130 77 06‘ " 25 135 130 OH— Siw110398‘ 120 130 34 06' C " 120 130 27 on“ " 25 130 150 Adipate C 62w120405‘ 25 125 107 Adipate SiW110398' 25 135 155 p-ArSO3 C 62w120405‘ 25 130 64 p-ArSO3 C Siw110398‘ 25 130 66 a. Time heated at elevated temperature = 5 hours. b. Time heated at outgassing temperature = 16 hours. c. All these experiments were carried out in a mixture of glycerol : water, 2:1 (v/v). 91 TABLE III.II. Elemental analyses of pillared products obtained using meixnerite, [Mg3Al(OH)8]°[OH]°2320, as precursor. POM anion d001(A) Elemental Analysis of Pillared Products [H2W1204016‘ 14-8 [M93.03A1(0H)6.0716.3[H2W120401 [BVW1104016- 15-1 [M92.96A1(0H>7.9616.3[BVW11O401 [SiV3W904017' 15-1 [M92.69A1(0H)7.7617.2[31V3W90401 [51W1103918' 14-8 [M93.01A1(0H)8.0218.3[Siw11039] TABLE III.III. Elemental analyses of pillared products obtained using [Mg3Al(OH)8]-Organo-Anion precursors. POM anion d001(A) Elemental analysis of pillared products [“2W1294016- 14-8 [M92.88Al(OHI7.76]5.84[H2W1204018 [“2W1204016' 14-8 [M92.84Al(OH)7.6915.96[H2w12040]b [51W1103918- 14-8 ““9296“(OH)7.9216.12[SiW11O391a [SiW11039]8' 14.6 [Mg3Al(OH)8]8[SiW11039]b :. The [Mg3Al(OH)8]-ADIPATE precursor was used. . The [Mg3Al(OH)8]—p-TOLUENESULFONATE precursor was used. 92 D. SUMMAR! Synthetic meixnerite, a layered double hydroxide of the type [Mg3Al(OH)8][OH].2H20, has unexpectedly been found to be an excellent precursor for ion exchange with Keggin-type [XM12040]n' or lacunary [XM1103glm' polyoxometalates (POMs) to form microporous pillared derivatives. XRD patterns, FTIR spectroscopy, elemental analysis and TEM studies show that the reactions are ‘topotactic, either' in. the jpresence or absence of a swelling agent. FTIR and 2951 MAS-NMR spectroscopy illustrate that the environment around the central SiO4 tetrahedron of the POMs is unaltered upon intercalation in the LDH galleries. Organic anion derivatives of the synthetic meixnerite are also very suitable for topotactic ion-exchange reactions. This work compares the effectiveness of our previously reported route- using organo anion intermediates-with our new direct route starting from meixnerite. The effect of different precursors on the microporosity of the pillared products is discussed. 93 [Mg1_xAlx(OH)2][CO32']X/2.2H20 tsoo°c,3h [Mg(2-2x)/(2+x)A12x/(2+x)I:Ix/2+x 10 111,0, 25 °C , 30h (— Mg1_xAlx(0H)2][OH']X.2H20 JHOA , glycerol pow [Mgl.xAIX(OH)2110A'Jx l POM "' 1——> [Mg1_xAlx(OH)2][POMn-lx/n Figure 111.13. General reaction scheme. LIST OF REFERENCES H O 10. 11. 12. LIST OF REFERENCES (3) Reichle, W.T., CHEMTECH 1986, 58. (b) Carrado,K.A.; Kostapapas, A.; Suib, S.L., Solid State Ionics 1988, 77. Dimotakis, E.D.; Pinnavaia, T.J., Inorg. Chem. 1990, 29, 2393. Miyata, S., Clays Clay Miner., 1980, 28, 50. Miyata, S., Clays Clay Miner., 1983, 4, 305. Kopka, H.; Beneke, K.; Lagaly, G., J. Coll. Interf. Sci. 1988, 123, 427. Sato, T.; Kato, K.; Endo, T.; Shimada, M., React. Solids 1986, 2, 253. Martin, K.J.; Pinnavaia, T. J., J. Amer. Chem. Soc. 1986, 108, 541. Reichle, W.T., J. Catal. 1985, 94, 547. Numan, J.G.; Himelfarb, P.B.; Herman, R.G.; Klier, K.; Bogdan, C.E.; Simmons, G.W., Inorg. Chem. 1989, 28,9 3868. Suzuki, E.; Ono Y., Bull. Chem. Soc. Jpn. 1988, 61, 1008. Pinnavaia, T.J.; Rameswaran, M.; Dimotakis, E.D.; Giannelis, E.P.; Rightor, E.G., Faraday Discuss. Chem. Soc. London, 1989, 87, 227. Miyata, S.; Kumura, T., Chem. Lett. 1973, 843. 94 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 95 Kwon, T.; Tsigdinos, G.A.; Pinnavaia, T.J., J. Amer. Chem. Soc. 1988, 110, 3653. Kwon, T.; Pinnavaia, T.J., Chem. Mater. 1989, 1, 381. Drezdzon, M.A., Inorg. Chem. 1988, 27, 4628. Chibwe, K.; Jones, W., Chem. Mater. 1989, 1, 489. Mooney, R.W.; Chiola, V.; Hoffman, C.W.; Vandertpool, C.D., J. Electrochem. Soc., 1962, 1179. Altenan, J.T.; Pope, M.T.; Prados, R.A.; So. H., Inorg. Chem. 1975, 14, 417. Finke, R.G.; Rapko, B.; Saxton, R.J.; Domaille, R., J. Amer. Chem. Soc., 1986, 108, 2947. Teze, A.; Herve, C., J. Inorg. Nucl. Chem. 1977, 39, 999. Thouvent, R.; Fournier, M.; Frank, R.; Rocchiccioli- Deltcheff, C., Inorg. Chem. 1984, 23, 598. Herve, G.; Teze, A., Inorg. Chem. 1977, 16, 2115. Howarth, O.W.; Jarrold, M., J. Chem. Soc. Dalton, 1978, 503. Kazanskii, L.P.; Spitsyn, V.I., Dokl. Acad. Nauk SSSR, 1975, 223, 721[381]. Herve, G.; Teze, A., C. R. Acad. Sci. Ser. C, 1974, 278, 1417. Pope, M.P., "Heteropoly and Isopoly Oxometalates", Springer, New York 1983, 87. CHAPTER IV Surface Organometallic Chemistry of Basic Layered Double Hydroxides: Reactions of Metal Cluster Carbonyls. As INTRODUCTION Surface organometallic chemistry' is an extension of solution organometallic chemistry to the surfaces of solids.1"6 One of the reactants is an organometallic entity and another is the surface or the surface functional groups of a solid support such as A1203, MgO, SiOz, zeolites, or clays. Ichikawa7 recently suggested, in his effort to parallel the role of zeolite cages with that of the solvent in solution, that "the cages of zeolites provide a solution- like enviroment where chemical reactions occur". Surface chemistry is very important in understanding and improving the performance of heterogeneous catalysts. A large number of new surface sensitive techniques (FTIR, NMR, X-ray photoelectron spectroscopy, STEM among others) can provide detailed information on the the correlations between Structure and reactivity of these systems.8'9 The continuous improvement of these "in situ" characterization 96 97 between structure and reactivity of these systems.8'9 The continuous improvement of these ”in situ" characterization techniques offers new opportunities for exploring the possible number of the surface-active structures. Catalytic systems that are active on a support can serve as models for the intermediates formed in very active homogeneous catalysts. Because intermediates can sometimes be stabilized on surfaces, an active homogeneous catalyst can perform better when supported on the surface of a solid.7 One main concern for the application of heterogeneous catalysts is their thermal stability over several catalytic runs. Heterogeneous catalytic conditions are usually severe (T>250 OC P>1 atm). Therefore, a good catalyst should not only have high activity and good selectivity but also high thermal stability and resistance to poisoning. Aggregation of small metal crystallites to form larger particles is undesirable. Highly dispersed metal particles on a surface could be very useful systems, if they are stable to poisoning or sintering. Two competing processes are known to occur for metal atoms supported on surfaces, namely, (i) aggregation of the particles to form crystallites and (ii) oxidative dispersion. The second gives molecular species that might be active but their activity is lower in comparison with that of supported crystallites. A good explanation for the latter is that the metal centers are poisoned by ligands required to stabilize them (carbon monoxide or hydrogen).10 98 Aggregation is the subject of our research here. More specifically we are exploring a very unique class of reactions that we found to occur on the surface of the layered double hydroxide, [Mg3Al(OH)8]°[Xn‘11/n'2HZO, X“— =co32‘ or 06'. Using KZPtCl6 and. (X) in MeOH, Giannelis and Pinnavaia11'12 demonstrated for the first time that it is possible to directly synthesize clusters of the type [Pt3(CO)6]n' (n=3,4,5) on the external surface of [Mg3Al]- CO32‘ LDH. They also demonstrated that Ihydrolysis of the surface carbonate anions most likely was producing the OH" ions responsibLe for the reductive condensation reactions. The structural hydroxides of the LDH were not involved. These conclusions were supported by the fact that non-basic anions such as Cl‘ did not exhibit analogous reactions. Several barriers to the intercalation of platinum clusters in the [Mg3All-LDH galleries exist, the most important being the highly ordered nature of the non-swellable gallery surfaces of the LDH. The amount of cluster produced was low in comparison to the exchange ion of the LDH. Thus, ion exchange was not favored by mass action and this system was rather inappropriate for achieving intercalation. Parallel to the work of Giannelis and Pinnavaia, my efforts focused on the surface reduction of M3(CO)12 (M=Ru, 13 showed Os) system on LDH surfaces. Hayward and Shapley that very strong reducing agents such as sodium metal were needed to reductively condense these clusters in solution. 99 Concentrated KOH in methanol could also be used, but only with Ru3(CO)12. In the present work we show that similar reactions can be accomplished on the surface of a basic LDH in the absence of a strong metallic reducing agent. B. EXPERIMENTAL M3(CO)12 , MeRu, Os and Ir4(CO)12 were purchased from Aldrich Chemical Co. The [Mg3Al]-LDHs, [Mg3A11061811xn' Il/n'ZHZO' XnT=CO3ZT or OH" were prepared according to 14 and to the procedures described in published procedures Chapter II and III. The conditions used for the reactions of the metal carbonyl clusters with [Mg3Al(OH)8][Xn‘]1/n°2H2O, XnT=CO3ZT or OH" are given in Tables IV.II., IV.IV. The IR frequencies for the resulting surface anion clusters are summarized in Tables IV.III, IV.V. SYNTHESES All manipulations were performed under an Ar atmosphere using a typical Schlenk line and a glove box. In a typical experiment, the required amount of the [Mg3All-LDH (typically 0.33 g, 0.99 meq) was degassed for 2 hours in a 250-mL round bottom flask fitted with a side arm. A 40 mL portion of the appropriate degassed solvent, containing the desired quantity of the carbonyl precursor was then transferred into the flask via syringe. The reaction vessel was equipped with a reflux condenser, flushed with Ar for 10-15 min, and then heated to reflux at the appropriate temperature for a defined time. After reaction the vessel 100 was cooled and transferred to a glove-box. The reaction mixture was filtered and washed with a small amount of the appropriate solvent. Experiments aimed at desorbing the cluster anions also carried out in the dry box. Extractions were done using a stochoiometric amount of [Et4N]C1, with regard to the metal cluster anion, dissolved in CH2C12. We assumed that the carbonyl precursors were transformed quantitatively to the corresponding cluster anions. The specific reaction, conditions used. to form ‘the cluster anions of Ruthenium, Osmium and Iridium on [Mg3Al(OH)8][CO32-]o.5 are described in detail below. I Formation of Surface [HRu3(CO)11]' Ru3(CO)12 (64 mg, 0.1mmol) in tetrahydrofuran (THF, 20 mL) was allowed to react under Ar with a suspension of the [Mg3Al]-LDH (100 mg, 0.33 meq) in THF (30 mL). The orange- red suspension turned deep red, and the reaction was complete within 8 h at room temperature. At the end of the reaction the flask was transfered to the glove box where the product was filtered and washed with CH2C12. The resulting anion was extracted from the surface of the LDH using a stochiometric amount of Et4NCl (0.1 mmol) in 5 mL CH2C12. After evaporation of the solvent, the anion was taken up in 5 mL THF for FTIR studies. Analytical data for the [Et4N]+ salt of the desorbed anion are as follows: Found: C, 32.00; H, 2.31; N, 1.95; Ru, 42.00; Calcd for C19H21N011Ru3 : C, 30.72; H, 2.83; N, 1.88; Ru, 40.86. 101 Formation of Surface [Ru6(CO)18]2' Ru3(CO)12 (38 mg, 0.06 mmol) in cyclohexane (C6H12, 20 mL) was cannulated under argon into a suspension of [Mg3Al(OH)8].[CO3]0.5'2H20 (100 mg, 0.33 mmol) in cyclohexane (30 mL). This suspension was stirred carefully and the temperature was increased to 90 OC. The mixture turned red initially and then dark brown. It was refluxed for about three hours and then was cooled to room temperature. After transfering the products in the glove box for filtration and washing, the anion was extracted from the surface of the LDH by reaction with bis(triphenylphosphineiminium) chloride, PPNCl (0.06 mmol) in 5 ml CH2C12. The filtrate was examine by IR spectroscopy. The [PPN]2[Ru6(CO)18] was further verified by elemental analysis. Calcd for C90H60N2018P4Ru6 : C, 49.37; H, 2.73; N, 1.29; Found : C, 50.00; H, 2.97; N, 1.37. Formation of [Ru6C(CO)16]2' The [Mg3Al]-CO3 LDH was heated at 130 0C under vacumm for 2 h and then was cooled to room temperature. Ru3(CO)12 (32 mg, 0.05mmol) in degassed diglyme (40 mL) was added to the LDH by means of cannula. The mixture was heated at reflux (170 0C) for about 10 h. The resulting brown-red precipitate was transfered in a glove box, washed and dried. The existence of [Ru6C(CO)16]2- on the LDH surface was verified by IR spectroscopy. Displacement of the cluster anion from the surface was done by reaction with Et4NCl. Elemental analysis verified the presence of [Et4N]2[Ru6C(CO)16] on the 102 filtrate. Calcd for C33H40N2016Ru6 : C, 29.85; H, 3.04; N, 2.11; Found : C, 29.74; H, 3.08; N, 2.20. Another experiment was carried out under the same conditions of temperature and pressure, Inn: at higher Ru loading, in order to see if ion emchange was occuring. In this case [Mg5A1(OH)12][CO3]O.5 was used. The Ru3(CO)12 used, was enough to produce [Ru6C(CO)16]2- in about two thirds (meqs) of the amount needed to ion exchange carbonate anions. Elemental analysis : Found: Mg, 17.16; Al, 3.81; Ru, 33.39; C, 11.45; H, 1.13; Suggested formula for the product [M95A1(0H)12][C03]0.13[RU6C(C0)1610.39 Synthesis of Surface [8053(CO)11]’ OS3(CO)12 (33 mg, 0.036 mmol) was brought in contact under‘ Argon atmosphere *with [Mg3Al]-CO3 (1 g, 3.3 :mmol) suspended in 60 mL of degassed tetrahydrofuran (THF). The slurry was stirred at room temperature (25 °C) for 24 h. The LDH turned to a very intense yellow suspension. Upon filtering, washing and metathesis reaction of the surface bounded cluster with Me4NCl in CHZClZ, IR spectrosc0py was employed for further characterization. Elemental analysis of the solid that resulted from the above solution upon solvent evaporation of the solvent in vacuo gave: C, 18.4; H, 1.44; N, 1.60; Calcd : C, 18.90; H, 1.36 N, 1.45. Synthesis of Surface [095(0011312’ A mixture of Os3(CO)12 (24.5 mg, 0.027 mmol) and [Mg3Al]-CO32' (l g, 3.3 meq) was heated in diglyme (60 mL) at reflux temperature (170 0C) for about 18 h. The resulting 103 orange-brown suspension was cooled to room temperature and the contents of the flask were transfered in a glove box for filtration, washing and preparation for IR spectroscopy. Elemental analysis of the evaporated solid after extraction of the LDH with Et4NCl in CH2C12 was as follows : Calcd for C34H40N2018086 : (L 21.44; H, 2.11; N, 1.47; Found: C, 22.70; H, 2.25; N, 1.60. Synthesis of Surface [0310C(CO)24]2' A mixture of OS3(CO)12 (24 mg, 0.027 mmol) and [Mg3Al]- C032' (1 mg, 3.3 meq) was suspended in degassed tetraglyme (60 mL) in a three necked flask. The contents of the flask were heated to reflux at 270 0C for 20 h. The solution turned from yellowish to red and then to dark red brown. Formation of the decanuclear osmium anion was observed by IR spectroscopy. The elemental analysis results were in good agreement with the expected values. Calcd for C41H40N2024Oslo : C, 17.31; H, 1.41; N, 1.23; Found: C,18.20; H, 1.62; N, 1.40. Synthesis of Surface [1r8(CO)2012‘ Ir4(CO)12 (10 mg, 0.009mmol) and [Mg3Al]-LDH (0.3 g, 1 mmol) were allowed to react for 30 min in tetrahydrofuran (THF). The formation of [Ir3(CO)20]2- cluster (N1 the LDH surface was verified by extraction with Et4NCl and IR spectroscopy of the resulting solution in the carbonyl region. Elemental analysis data of this salt were in good agreement with the calculated values. Calcd: C, 18.34; H, 1.71; N, 1.20; Found: C, 18.60; H, 1.87; N, 1.30. 104 Synthesis of {[pt3(co)61512‘ Literature methods15 were used to prepare the Na+ salt of this anion. Sodium Zhexacloroplatinate (1.833 g, 3.25 mmol) was added to a methanolic solution of [CH3COOJNa (3.63 g, 26.6 mmol) in 34 mL methanol under a carbon monoxide atmosphere. The solution was stirred for 24 h, and the CO atmosphere was renewed every 4-5 h. The reaction mixture then was filtered to remove any precipitate and the anion was identified by FTIR spectroscopy by comparing the CO stretching frequencies with literature values.15 The filtrate» was kept for anion. exchange reactions with. LDH derivatives. It must be noted, however, that after 10 hours of aging, the {[Pt3(CO)6]5}2' anion was transformed into another unidentified anion. Therefore, care should be taken to use freshly prepared anion for the ion exchange reaction. Synthesis of [Mg3A1(OH)8][C83C6H4-SO3] A 5-mL portion of a [Mg3Al]-OH suspension (1.0 wt%) was mixed with 10 nu11mf glycerol. To the resulting suspension was added on equimolar amount of p—toluenesulfonic acid in 2 mL H20. After centrifugation (twice) and resuspension in 5 mL methanol, the dispersion was mixed with 10 mL glycerol. This material was kept for further ion exchange reactions with carbonyl anions. 105 Reaction of [pt15(co)3012' with [M93111 (OH)8]'[p-CH3C684-SO3] A solution of Na2[Pt15(CO)3O] in methanol, was added to a dispersion of [Mg3Al(OH)8].[p-CH3C6H4-SO3] in 2:1 (v/v) glycerol/methanol. The amount of the cluster anion was in 50 % excess with respect to the anion exchange capacity of the LDH. The mixture was stirred for 30 min and then it was centrifuged and washed with methanol. The volume of methanol used was three times the volume of the slurry. Finally, the resulting solid was dried under Ar. Elemental Analysis of the ‘blue-violet product gave: : Mg, 7.41; Al, 3.84; Pt, 44,84; C, 6.28; H, 1.93. Suggested formula for the product [Mglg.9A19.3(OH)85.702][Pt15(CO)3O][CO3]6.0. Its powder XRD pattern is described in Table IV.I. Futhermore the sample was examined by TGA, XRD, FTIR spectroscopy and SEM. TABLE IV.1. Powder xrd (Cu-Ka) of the orthorombic lattice of [M919O9A19.3]-[Pt15(C0)3012- intercalate. O O 100 14.7 51 320 3.20 2.8 010 8.5 100 500 2.94 3.6 200 7.4 7 030 2.83 4.3 210 5.65 5.6 600 2.45 2.8 300 4.91 2.8 610 2.35 2.8 011 4.08 7.1 012 2.23 2.9 106 FTIR MEASUREMENTS FTIR spectra were recorded on an IBM IR spectrometer in the 4000-400 cm-l region. Solution spectra were obtained by using 0.1 mm NaCl cells. The background was subtracted by first placing a cell of the same solvent used to dissolve the sample. This cell was containing the salt used to desorb the carbonyl anion form the LDH surface, in case that its frequencies were interfering with the ones of the cluster in the region examined. Samples of solids were made by mixing them with KBr and pressing them into a disc. The KBr pellets were approximately 1% by weight. In some cases the spectrum was recorded by using mineral oil mulls placed between CsI disks. Samples of air sensitive intercalates were prepared in a nitrogen-filled glove box just prior to scanning to minimize decomposition. However most of the samples appeared to be stable upon exposure in air for a short period of time. C. RESULTS: (A). Reactions of Ru3(CO)12 with [Mg3Al(OH)8]°[CO3]0.5°2320 1. Formation of [HRu3(CO)11]’ Ru3(CO)12 in tetrahydrofuran reacts under Ar with a suspension of the [Mg3Al]-LDH to form a new surface-bound organometallic species. The orange-red suspension turned deep red over a reaction time of 8 h at room temperature. The resulting surface species was extracted from the surface of the LDH using a stochiometric amount of [Et4NJCl (with regard to the amount of the anion expected to be formed 107 after complete conversion of its precursor). The IR spectrum of the desorbed species is shown in Figure 1A. The reaction conditions are summarized in Table IV.II and the frequencies of the CO groups are given in Table IV.III, agree well with the literature values16 for [HRu3(CO)11]'. When CHZClZ was used as a solvent the band at the lowest frequency appeared as a shoulder Figure 1B. Analogous reaction of Fe3(CO)12 and Os3(CO)12 with the [Mg3Al(OH)8].[CO3]0.5°2H20 afford the corresponding [HFe3(CO)11]‘ and [HOS3(CO)11]- anions (see part B below). Comparison with the IR spectra of [HFe3(CO)11]‘, (Figure 1C) and [HOS3(CO)11]’, (Figure 1D) illustrate the Spectral similarities. mmol d001 Bound Ru3(CO)12/ Temp. Time Surface Metal meq clay solvent (°C) (h) (A) complex color (wt%) 0.30 THF 25 6 7.7 HRu3(C0)11‘ red 3.50 0.16 cyclo- 90 3 7.7 Ru61c0)182‘ dark 2.49 hexane brown 0.15 diglyme 170 20 6.6 606c1001162‘ dark 2.25 brown 108 A I a LlJ o 5 . . I: 2200 2000 1600 2200 2000 1600 3 (D Z < (r 1.. c 0 2200 2000 1600 2200 2000 1600 WAVENUMBER (cm-1) Figure IV.1. FTIR spectra (transmission mode) of [HM3(CO)11]’ anions formed by the reaction at 25 °C of the corresponding M3(CO)12 clusters with the external surface of [Mg3Al(OH)8][C03]o.5°2H20 in tetrahydrofuran (THF) and subsequently desorbed by reaction with [ Et4N 1C1: (A) [HRu3(c0)11]' in THF, (B) [HRu3(CO)11]' in CH2C12, (C) [HFe3(C0)11]- in CH2C12, (D) [HOS3(CO)11]- in CH2C12. TABLE 1V3111. Carbonyl stretching frequencies of ruthenium carbonyl anions desorbed from hydrotalcite after reaction with RU3ICO)12. Extracted v[CO], (cm'l) Literature IR solvent cluster observed values HRu3(CO)11' 2074(vw),2010(s) 2070(vw),2012(s) THF l985(S),1948(m) 1984(8),l947(m) Ru61001182‘ 2001(s),1983(s+) 2001(s),1986(s,br) CH2C12 1925(m,sh) l933(m,sh) Ru601c01162‘ 2028(w),1978(s) 2030(w),1978(s) CH2C12 Upon exposure of the surface bonded [HRu3(CO)11]' anion to air, oxidation to [Ru(CO)x(M)2]n x=2,3 species occured, as it is evident from IR data (Figure 2C). 11. Formation of [Ru6(co)18]2‘ Ru3(CO)12 in refluxing cyclohexane (90 (RD reacts under argon with [Mg3Al(OH)8].[CO3]0.5°2H20. The reaction mixture turned red initially, then dark brown and it was left for about three hours at 90 0C to complete the reaction. The surface anion was extracted from the IJNI by reaction with bis(triphenylphosphoranylidene) ammonium chloride, [PPN]Cl in CHZClZ. The filtrate was examine by IR spectroscopy13 (Figure 2A, Table IV.III). The position of 110 Figure Iv.2. (A) FTIR spectrum of : (A) [Ru5(CO)13]2‘ anion formed by the reaction at 90 °C of Ru3(CO)12 with [Mg3Al(OH)8][CO3]0.5°2H20 in cyclohexane. The anion was extracted from the [Mg3Al]-CO3 surface by metathesis reaction with [PPN]Cl in in CHZClZ, (B) [Ru5C(CO)15]2' anion (nujol mull) grafted on the [Mg3Al]-CO3 surface. The ion was formed by reaction at 170 °C of Ru3(CO)12 with [Mg3Al(OH)3][CO3]0.5‘2H20 in diglyme. (C) [Ru(CO)x(O-M)2]n ( x=2,3 and M=Mg, Al ) species (nujol mull) that result from oxidation of [HRu3(CO)11]‘,[Ru6(CO)18]2' and [Ru5C(CO)16]2‘, on the LDH surface, in the presence of air and light. TRANSMITTANCE 111 A 6 2200 2000 1600 2200 2000 c 2200 2000 1800 WAVENUMBER (cm-1) 1 800 112 the carbonyl stretching frequencies are those expected for [PPN]2[Ru6(CO)18]. Attempts to form ‘this anion on ‘the surface of the LDH at room temperature and prolonged reaction times always resulted in a mixture of products, [HRu3(CO)11]' and [Ru6(c0118]2', as judged by IR data. 111. Formation of [Ru5C(CO)16]2' [Mg3Al]-CO3 LDH, pre-heated at 130 0C under vacumm for 2 h, reacts with Ru3(CO)12 in degassed diglyme at 170 0C to form [Ru6C(CO)16]2" after a reaction period of 10 h. The resulting brown-red product exhibited IR carbonyl stretcing frequencies shown in Figure 2B. The CO frequencies support the existence of [Ru60100)16]2‘ on the LDH surface. Displacement of the cluster anion from the surface by reaction with Et4NCl gave CO frequencies in agreement with literature values13 (Table IV.III). (B). Reactions of Os3(CO)12 with [Mg3Al(OH)8][CO3JOOS'ZHZO 1. Synthesis of [HOs3(CO)11]' OS3(CO)12 reacts with [Mg3Al]-CO3 suspended in degassed tetrahydrofuran at room temperature to form a very intense yellow suspension within a reaction of 24 h. Metathesis reaction of the surface bounded cluster with Me4NCl in CHZClZ, as shown in Table IV.IV, yielded a desorbed species with IR spectrum Shown in Figure 1D. The carbonyl stretching frequencies (Table IV.V) indicate the existence of [H083(CO)11]- anion in CH2C12.17 113 11. Synthesis of [095(0011812‘ Reaction of OS3(CO)12 and [Mg3Al]-CO32‘ in diglyme at reflux temperature (170 0C) for about 18 h gave an orange- brown suspension. The IR spectra (nujol mull) of the solid (Figure 3A, Table IV.V) indicated. the presence of [056(CO)18]2' on the LDH surface.13 111. Synthesis of [09100(co)24]2‘ A mixture of Os3(CO)12 and [Mg3Al]-CO32' in degassed tetraglyme were reacted at 270 C’C for 20 h. The solution turned from yellowish to red and then to dark red brown. Formation of the decanuclear osmium anion, [OleC(CO)24]2, was observed by IR spectroscopy13, (Figure BB, Table IV.V), following the usual desorption treatment with [Et4N1Cl in CH2C12. The elemental analysis results were in good agreement with the expected values. mmol d001 Bound 033(CO)12/ Temp. Time Surface Metal meq clay solvent (°C) (h) (A) complex color (wt%) 0.022 THF 25 24 7.7 HOS3(CO)11' yellow 3.29 0.016 Ethylene 8.5 H3OS4(CO)12- yellow 3.03 Glycol 0.016 Diglyme 170 16 6.5 056(c0)182‘ orange 2.95 0.016 Triglyme 270 20 8.8 0s106<001242‘ red 3.10 114 2200 2000 1600 2200 2000 1800 TRANSMHTANCE 2200 2000 1600 WAVENUMBERS (cm- 1) Figure Iv.3. FTIR spectrum of : (A) [os6(co)18]2' anion (nujol mull) formed on [Mg3Al]-Co3 surface by the reaction of 093(co)12 with LDH in diglyme at 170 Co. (B) [0810C(C0)24]2' formed after reaction of 053(CO)12 with [Mg3Al(OH)8][CO3]O.5'2H20 in tetraglyme at 230 °C. The anion was desorbed from the LDH surface using [Et4N]Cl in CH2C12. (o) [Ir8(c0)2012‘ formed by reaction of [Mg3Al(OH)8][CO3]O.5 ZHZO with Ir4(CO)12 in THF at 25 °C. The anion was extracted from the LDH surface by metathesis reaction with [Et4N]Cl. TABLE 1VTV. Carbonyl stretching frequencies of osmium and iridium carbonyl anions desorbed from hydrotalcite after reaction with Os3(CO)12 and Ir4(CO)12 respectively. Extracted v[CO], (cm-1) Literature IR solvent cluster observed values HOS3(CO)11' 2079(w),2020(s) 2083(w),2021(s) CH2C12 l995(s),l951(m) l996(s),l951(ms) H3Os4(CO)12’ 2047(5),2023(S) 2048(3),2022(s) CH2C12 1999(8),1970(w) 2000(s),l976(w) 036(CO)182 l993(vs),l910(w) 1995(9),1915(w) cnzc12 0s1001001242‘ 2024(s),1986(s) 2033(s),l986(s) CH2C12 Ir8(CO)202' 2046(sh),2018(s) 2050(sh),2020(s) THF 1981(sh),1828(sh) 1975(sh), 1825(sh) l776(s),1713(w)- 1770(8),1710(sh) (C). Reaction of Ir4(CO)12 with [Mg3Al(OH)8][CO3]0.5°2320 . 9 Synthesis of [Ir8(c0)2012' Reaction of Ir4(CO)12 and [Mg3A11-LDH in tetrahydrofuran at 25 OC results in the formation of [Ir8(CO)ZO]2’ cluster, on the LDH surface, as verified by extraction with Et4NCl and IR spectroscopy18 (Figure 3C, Table IV.V). Elemental analysis data of the Et4N+ salt of the anion were in good agreement with the calculated values. 116 (0). Reaction of [Pt15(CO)3o]2‘ with [Mg3Al(OH)8]' [p-CH3C6H4-SO3] The XRD pattern of a film of the reaction product (Figure 4) was independant of the ratio of Mg to Al in the starting [Mg/Al]-[p-CH3C5H4SO3] LDH. The indexing of the pattern was based on 12 reflections and it is shown in Figure 4. We also observe some strong 0k0 reflections in addition to 001 reflections. This indexing results in an orthorombic lattice. The unit cell parameters of the new lattice are : A=l4.7, B=l7, c=9.4, a=b=c=90° and v=2349.1 A3. Scanning electron.:microscopy shows that a fiber-like .material is produced from that reaction (Figure 5). FTIR (Figure 6B), xrd (Figure 4) and elemental analysis results show that the cluster is intact. FTIR spectra in the 1200-400 cm"l region show the presene of Al-O-Mg lattice vibrations (Figure 7). The elemental analysis results also show that the Mg/Al ratio decreases from 3:1 to 2:1. SEML images, at higher magnification, show that the fibers are aggregates of fibers with smaller diameters (Figure 5, bottom right). TGA (Figure 8) shows a large weight loss, initially, which is accompanied by a less gradual weight loss up to 420 0C. D. DISCUSSION: I. Reactions Of Ru3(CO)12 With [M93A1(OH)8] ’[CO310.5'2320 3 Previous literature results have shown that when Ru3(CO)12 or H4Ru4(CO)12 was used on alumina, an active 117 230004 18400- 13800d 100 9200J INTENSITY (C/S) 0 4600- \‘/j .I 010 200 1° 2103__0 1 L20 WWW 4 r. 2 5 8 11 14 17 20 23 26 29 32 35 38 41 DEGREES 2 THETA Figure IV;4. X-ray (Cu-Ka) powder diffraction pattern of the product formed by the reaction of [Pt15(C0)30]2' with [Mg3Al(OH)8][p-CH3C6H4-So3], in 1:1 glycerol:methanol (v/v). ”t- .' .. k. 118 Figure IV.5. SEM images of : meixnerite, (top left); the product that results from ion exchange of [Pt15(CO)30]2' in the [Mg3Al(OH)8][p-CH3C6H4-SO3] host lattice, (top right); the same product, (bottom left); The view is now along one of the edges of the sample; the same product as in the top right picture but at higher magnification, (bottom right). omomulaasfi 025 032 3.2 amomlu 35 wag ago; if \ omomu B53 moms omx 36H omomulzodfi goo ovmx 35H ..._ «CG ‘1‘ ..x 120 2062 1.1.1 2 2055 55 1881 Ir 1 c: (f) 00 <: g: A E 00 1870 1r 0] 8 2200 ' 2000 '8 1800 WAVENUMBERS (cm—1) Figure IV.6. FTIR spectra of : (A) [149935181455]- [Pt15(CO)30]2’ as KBr pellet (1 wt %), (B) Na2[Pt15(CO)3o] in MeOH . 121 RELATIVE ABSORBANCE 3 14'00 1300 1050 850 660 400 WAVENUMBERS (cm—1) Figure IV.7. FTIR spectra of : (A) [M99O95A14.55]- [Pt15(CO)3o]2' and (B) [Mg3Al(OH)8][p-CH3C6H4SO3]. Molecular vibrations due to 803' group of the toluenesulfonate anion at 1125 cm"1 are present in spectrum B and absent in spectrum A. Lattice vibrations due to Al-O-Mg groups are present between 550 and 680 cm"1 in both spectra. 122 5- 0‘ 0| I Weight loss (%) 5 35- 45 I I I I 1 25 1 25 225 325 425 525 Temperature ('C) Figure 117.8. TGA curve of [1499351114551-[pt15(c0)3012". 123 catalyst resulted.19 When Ru3(CO)12 was used on MgO, for CO hydrogenation reactions, [Ru6C(CO)16]2‘ was identified on the M90 surface but its role remained questionable.19 Also, a mixture of [HRu3(CO)11]' and [Ru6(CO)18]2’ has been reported to form on MgO upon its reaction with Ru3(CO)12 under catalytic conditions.19 In contrast the surface of hydrotalcite reacts selectively with Ru3(CO)12 to form only [HRu3(CO)11]' at room temperature. Reaction at 90 0C in cyclohexane leads to high yields of [Ru6(CO)18]2' and reaction at 170 (N: in diglyme afford surface bound [Ru6C(CO)16]2- anions. It is interesting that these clusters can be produced in solution only by very basic reagents such as KOH in absolute MeOH,20 M/benzophenone solution (M=K,Li,Na)21 and, more efficiently, by Na metal.l3 It seems that the interfacial chemistry of ruthenium on [Mg3AlJ-CO3 parallels well its solution chemistry where [HRu3(CO)11l' is formed initially and [Ru6(CO)18]2‘ after longer reaction times or elevated temperatures by redox condensation of [HRu3(CO)11]' and [Ru3(CO)11]2'. Here the higher temperature helps the selective condensation for [Ru6(CO)18]2' formation. The IR spectra of the oxidized species show successive loss of CO according to [E?.u(CO)3(I\d)2]n <---> [Ru(CO)2(M)2] + nCO (M=OA1- or OMg-) This has been observed previously for other supports such as silica. 124 When Ru3(CO)12 is heated with hydrotalcite (at higher loadings as mentioned in the experimental part) in tetraglyme at 170 0C, the powder XRD of the resulting material, Figure 9B, shows that an expansion of the layered lattice has occured. This is not true though since hydrotalcite reacts with tetraglyme (in the absence of ruthenium carbonyl) under the same conditions to form a different material, Figure 9A. TEMs show that most of the platelets remain approximately hexagonal (Figure 10A). However, there are regions where depletion of these platelets occurs, especially at the edges, leading to rounded particles (Figure 10B,10C). It thus appears that metal complexation and dissolution of the LDH by diglyme, starting from the edges of the LDH is responsible for such transformations at these high temperatures. It is remarkable however that when refluxing the LDH at 170 0C in the presence of Ru3(CO)12 formation of [Ru6C(CO)16]2' occurs on the resulting support within 10 h. The IR spectra of the cluster on the surface is shown in Figure 28. II. Reactions of 033(CO)12 with [Mg3Al(OH)8]'[CO3]0.5'2320 At room temperature '(25 °C) the presence of [HOs3(CO)11]' is evident on the LDH surface. When the temperature is increased to 170 °C formation of [056(CO)18]2' occurs. It is a very important result since reaction of Os3(CO)12 with sodium metal in various ethers appears to be the only other efficient route to this anion.13 Although some reconstruction of the LDH surface 125 RELATIVE INTENSITY (counts/ sec) t r r 2 6 1B 14 DEGREES 2 THETA Figure IV.9. (A) X-ray (Cu-Ra) powder diffraction pattern of the product that results from heating hydrotalcite, [M95A1(0H)8][CO3]0.5°2H20, suspended in tetraglyme at reflux temperature (170 0C) for 10 hrs. (B) x-ray (Cu-Ra) powder diffraction pattern of [Mg5Al(OH)8][C0310.13[Ru6C(C0)16]0.39 that. results from heating [MgsA1(OI-I)3] [CO3]0 . 5 '2H20 and Ru3(CO)12 in tetraglyme at reflux temperature (170 °C) for 10 hrs. 126 Figure IV.10. Representative TEM images of three different regions (A, B and C) for the reaction product of tetraglyme with [M95A1(OH)8][CO3]0.5'2H20 at 170 °C for 10 h. 128 appears to occur at higher temperatures, as mentioned earlier, the formation of these anions, which are hard to prepare in pure forms in solution, is very interesting. [OsloC(CO)24]2' can be also prepared at higher temperature using triglyme or tetraglyme as the solvent. Both anions can be extracted and characterized in solution. The much shorter time necessary to obtain the decanuclear cluster anion is remarkable ( 20 h versus 70 h reported for sodium metal).13 This chemistry can be explored since there might be potential to prepare other anions for which there is no simple or high yield route in solution. III. Reactions of Ir4(CO)12 with [Mg3A1(OH)8]°[CO3]0.5°2820 The reaction of Ir4(CO)12 with strong bases such as KOH/MeOHZ or Na/THF leads t1) the following' conversions, depending in part on the reaction time and on the gas atmosphere18 under which the reaction take place Ir4(CO)12 + strong base ---> [HIr4(CO)11]' --->Ir8(CO)22]2_ (I) (II) ---> [Ir8(c0)2012‘ --> [Ir61c0)15]2‘ ---> Ir HCO3‘ + OH', which is responsible for the base promoted reactions of metal carbonyls, is dependent on the charge distribution in the LDH layers. Further studies are needed to correlate the surface basicity with LDH structure and composition. v. Reaction of [Pt15(CO) 3012' with [Mg3u(OH)a]° [p-CB3C684-SO3] In View of the results, one can explain the observed OkO xrd reflections (Figure 4) from the fiber-like morphology of the crystals (Figure 5). FTIR spectra (Figure 7) present strong evidence that the framework of the cluster remains intact after reaction. The xrd. pattern and the morphology of the crystals most likely suggests that the cluster is oriented with its helix parallel to the fibers. Although the cluster ion structure is retained within the host, the reaction is not simple topotactic. All these results together with a decrease in the Mg to Al ratio suggest that some reorganization of the layer structure occurs. The indexing of the 12 xrd reflections results in an orthorombic lattice as mentioned. in earlier section. There is also no p-tolunesulfonate anion left on the product as judged by the absence of the 503' absorption at 1125 cm’1 in the FTIR (Figure 7A). However absorptions due to hydroxyl 1) groups (~3450 cm' and C032- anions (1375 cm'l) are present. Strong absorptions due to lattice bridging oxygens 131 and hydroxyls between Mg and Al metals are also present in the 900-200 cm"1 region. It is most likely that the fibers consist of a mixed metal (Mg and Al) oxide-hydroxide chains, with TiI3 structure. The cluster is enclosed within these chains and its helix is oriented papallel to them. Thermogravimetric analysis results are very consistent with the chemical formula derived from elemental analysis. Thus TGA results (Figure 10) are in good agreement with initial loss of 30 CO ligands and 40.2 OH” groups. Following that, further dehydroxylation of the layers and loss of CO32‘ anions occurs, up to about 400 °C. VI. General conclusions: The reaction of tri- and tetra-nuclear metal carbonyls on [Mg3A11-LDH surface leads to anions with the same or higher nuclearity than the starting carbonyl. Paolo Chini characterized similar reactions of metal carbonyls in solution as "redox condensation reactions". Little is known concerning the detailed mechanism for these processes but the reduction of the metal core of a carbonyl cluster or carbonyl cluster precursor with simultaneous clusterification is involved. The reducing agent is a strong base such as KOH in MeOH or Na metal. In the surface of a basic oxide such as MgO similar reactions can occur, but no systematic study for the synthesis of such clusters is available. Some of them have been identified as intermediates in catalytic runs and it was a strong 132 suggestion that their synthesis was possible. The reducing agent in the case of MgO are active sites on the surface, such as OH', C032- contamination or even defect sites that arize from Mg atoms that have cordination numbers 5 or 4. A metal carbonyl will reductively loose CO ligands in the presence of a strong base: MX(CO)y + OH’ ----> [HMX(CO)y-1]' + C02 We decided to use [Mg3Al(OH)8][CO3]0.5 for converions of metal carbonyls since it is a MgO-like support with regard to surface OH groups. In addition, the LDH has a layer structure with localized charge density. Therefore, it was interesting to investigate the possibility of synthesizing these carbonyl clusters and directly intercalating them in the host lattice. The localized charge density is also very important since in systems with delocalized charge, like graphite, aggregation of metal particles occurs at the edges after a few catalytic runs. A highly catalytic active system is most likely the one that has well-dispersed crystallites. However only the external surfaces of the LDH were accessible for reaction with the metal carbonyls. Also, no ion-exchange reaction with gallery C032” or OH" ions occured, as evidenced by the reflection of a 7.8 A d-spacing after reaction. It was also interesting that supports such as silica or MgO result in dissociation or oxidation of Ru and Os clusters to form [M(CO)x]n-solid (x=2,3 M=Ru, Os) species. My initial efforts to synthesize the Ru and Os anionic 133 clusters at room temperature resulted in a mixture of products, [HM3(CO)11]’ and [M6(CO)18]2', probably due to the fact that the temperature is necessary to induce clusterification of the metal particles to [M6(CO)18]2'. However at higher temperature and under inert atmosphere such reactions were possible. The reaction products of the ruthenium carbonyl with the LDH support are summarized in scheme I. Analogous products from the reaction of the osmium carbonyl with the LDH support are summarized in scheme II. It has been shown that the LDH surface can induce reductive condensation reactions as efficiently as strong bases, such as Na metal, in solution. It is the basic property of our support system that causes such transformations. The redox potential of a cluster might be lowered when it comes in contact with the LDH-support, due to interactions of the CO ligands with active sites on the LDH surface. However, the OH" anion most likely is more active on an LDH surface than on other basic surfaces and therefore acts as a stronger reducing agent. Finally, ion exchange reaction of [Pt15(CO)3O]2' cluster with the -p-toluenesulfonate intercalate of the LDH gives a fibrous product in a non-topotactic way. The lattice undergoes hexagonal to orthorombic transition. All indications are supportive of the existence of mixed Mg-Al oxide-hydroxide chains, with TiI3 stucture, that enclose the intact framework of the cluster. The helix of the cluster is oriented parallel to these chains. 134 THF .. ” [331131C01111—" 25 °C/8 h LDH-CO3 C‘Hu 2_ + p [Ru¢(c0)1.] —-> [Ru(CO)szln 3113(CO) 12 90 °C/3 h (:32, 3) DIGLYME - > tame (c0) 1.1L» 170 oC/IO h Scheme I. Synthesis of ruthenium clusters 135 EEBF' - > [3083 (C0) 11] Room 1' LDH-CO: DIGLYME _ + I» [086(CO)18]2 083(C0)12 170 °C TETRAGLYME 270 °c > [0310C (C0) 24] 2- Scheme II. Synthesis of osmium clusters 136 I. SUMMARY The carbonyl cluster anions [HM3(CO)11]", [M6(CO)18]2' (M=Ru or Os), [Ru6C(c0)16]2', [0310c1c0)24]2‘, and [Ir8(c0)2012' are formed on the external surfaces of the basic layered double hydroxides (LDHs), [Mg3Al(OH)8][Xn'11/n'2H20 (x“‘= CO32' or OH’), upon reaction of the corresponding neutral precursors, M3(CO)12 (M=Ru or Os) and Ir4(CO)12. The surface reductive decarbonylation reactions of the neutral clusters are as effecient as with sodium metal in solution. The importance of the LDH surface hydrolysis reaction C032- + H20 ----> HCO3' + OH“ for the reductive condensation of the metal cluster carbonyl precursors is demonstrated. A two-fold cooperative effect arising from strong metal cluster-LDH support interactions and enhanced reactivity of OH' anions on the LDH surface is believed to be the reason for such exceptional reactivity. Attempts were made to achieve intercalation of the anionic clusters formed, but no intercalation was observed. The ion exchange reaction of [Pt15(CO)3O]2‘ with [Mg3Al(OH)8][p-CH3C6H4SO3'] is not topotactic. SEM images show the formation of a fibrous product whose xrd pattern was indexed using 12 reflections. The hexagonal lattice of [Mg3Al(OH)8][p-CH3C6H4SO3'] reorganizes to yield a material with an orthorombic lattice. The new unit cell parameters are : A=14.7, B=17, c=9.4, a=b=c=90° and v=2349.1 A3. LI ST OF REFERENCES 11. 12. 13. 14. LIST OF REFERENCES Lamb, H.H.; Gates, B.C.; Knozinger, H., Angew. Chem., Int. Ed. Engl. 1988, 27, 1127. Basset, J.M.; Choplin, A.; J. Mol. Catal., 1983, 21, 95. Psaro, R.; Dossi, C.; Ugo, R., J. Mol. Catal., 1983, 21, 331. Lamb, H.H.; Gates, B.C., J. Am. Chem. Soc., 1986, 108, 31. Basset, J.M.; Theolier, A.; Commereuc, D.; Chauvin, Y., J. Organomet. Chem., 1985, 279, 147. Giannelis, E.P.; Rightor, R.G.; Pinnavaia, T.J.; J. Am. Chem. Soc., 110, 3888, 1988. Ichikawa, M., personal communication. Robbins, J.L., J. Phys. Chem., 1986, 90, 3381. Deeba, M.; Gates, B. C., J. Catal., 1981, 67, 303. Gates, B. C.; Lamb, H.H.; J. Mol. Catal., 1989, 52, 1. Giannelis, E.P.; Pinnavaia, T.J., unpublished results. Pinnavaia, T.J.; Rameswaran, M.; Dimotakis, E. D.; Giannelis, E.P.; Rightor, E.G., Faraday Discuss. Chem. Soc., 1989, 87, 217. Hayward, C.T.; Shapley, J.R.; Inorg. Chem., 1982, 21, 3816. Reichle, W.T., J. Catal., 94, 547, 1985. 137 15. 16. 17. 18. 19. 20. 21. 22. 138 Longoni, G.; Chini, P., J. Am. Chem. Soc., 1976, 98, 7225. Johnson, B.F.G.; Lewis, J.; Raithby, P.R.; Suss, G., J. Chem. Soc. Dalton, 1979, 1356. Eady, C.R.; Johnson, B.F.G.; Lewis, J.; Malatesta, M.C., J. Chem. Soc. Dalton, 1978, 1358. Angoletta, M.; Malatesta, L.; Caglio, G., J. Organomet. Chem., 1975, 94, 99. Pierantozzi, R.; Valagene, E.G.; Nordquist, F.; Dyer, P.N., J. Mol. Catal., 1983, 21, 189. Eady, C.R.; Jackson, P.F.; Johnson, B.F.G.; Lewis, J.; Malatesta, M.C.; McPartin, M.; Nelson, W.J.H., J. Chem. Soc. Dalton, 1980, 383. Blattacharrya, A. A.; Nagel, C.C.; Shore, S.G., Organometallics, 1983, 2, 1187. Serna, C.J.; White, J.L.; Hem, S.L., Clays and Clay Miner., 1977, 25, 384. CHAPTER V On the Nature of Titania Pillared Layered Silicate Clays A. INTRODUCTION There has been several recent reports concerning the design of pillared smectite clays in which the gallery heights are substantially larger than the van der Waals' thickness of the host layers.1"8 Structures in which the pillar height is approximately two or more times the thickness of the 10 A 2:1 silicate layers have been reported. Such "super gallery" pillared clays have been prepared by reaction of large polyoxocations or by direct intercalation of metal oxide sols (the "DIMOS" method).9 Titania pillared clays with basal spacings greater than 28 A are of special interest because of their large gallery heights. These materials have been prepared by reaction of Na-smectites with polyoxocations formed during the acidic hydrolysis of titanium tetrachloride in aqueous solution.10 Related materials also have been produced by the reaction of titania sols derived from the acidic hydrolysis of titanium isopropoxide.4 However, little is known regarding the nature 139 140 of the pillaring agent in these derivatives. In an effort to better understand the stuctural properties of titania pillared clays, we have undertaken an investigation of these materials by Raman spectroscopy. High charge density clays such as Li-fluorohectorite have been pillared by titanium polyoxocations for the first time. The resulting pillared products from Li-fluorohectorite and Na- montmorillonite have been examined by continuous and pulse excitation IRaman. spectroscopy, respectively. Emphasis has been placed on the excitation Raman spectroscopic properties of the intercalated pillaring agent relative to those recently reported for polyoxo titanium cations derived from 11'12 Our results TiCl4 hydrolysis in strong acid solutions. suggest that regardless of the origin of the titania pillar precursor, the basic building blocks in titania pillared clay precursors are derived from condensation of [Ti8012(H20)24]8+ cations to form pillaring aggregates with different incremental gallery heights. 8. EXPERIMENTAL I. SYNTHESIS OF TITANIA PILLARED SMECTITE CLAYS Li-fluorohectorite was donated as an aqueous suspension from Corning Glass Works. Na-montmorillonite was obtained by ion exchange of Ca-montmorillonite (from Arizona) with 10 molar excess NaCl. Na-montmorillonite was also purified by sedimentation techniques and by using sodium dithionite solutions. 141 1. Titania Fluorohectorite Two different titanium precursors were used for preparing pillaring solutions, Ti(OC3H7)4 and TiCl4. The hydrolysis of Ti(OC3H7)4 was carried out in the following way : A solution containing 10 mL of Ti(OC3H7)4 (33.57 mmol) was added to 67.2 mL of 1.0 M HCl (67.19 mmol). Following stirring the solution became clear. 6.4 mL of it were then used for pillaring a 0.1 % dispersion of 10 ml Li-fluorohectorite (0.122 meq). The final ratio of titanium to clay was 28 mmol Ti per gram Li-FHT or 28 mmol Ti/1.22 meq of clay. The product was washed with water until it was free of Cl- anions (negative test with AgNO3). TiCl4 was hydrolysed in 12 M HCl. In the hydrolysis reaction, TiCl4 (8.8 mL, 0.08 mole) was added to 5.0 mL of 12.00 M HCl (0.06 mole). After dilution, the final pillaring reagent solution contained 0.80 M titanium and 0.60 M HCl. Part of it was then added to 10 mL of a 1.0 wt% dispersion of Li—fluorohectorite (0.122 meq). The final ratio of titanium to clay was 28 mmol Ti per gram Li-FHT or 28 mmol Ti/1.22 meq of clay. The product was washed again and absence of excess chloride anions was verified by the AgNO3 test. 2. Titania Montmorillonite Part of the titanium pillaring solution that resulted from TiCl4 was added to 12.50 mL of a 0.20 wt% (0.015 meq) dispersion of Na-montmorillonite, so that the final ratio of titanium to clay was 20 mmol Ti/g of clay or 20 mmol Ti/0.87 142 meq of clay. The product was washed again and absence of excess chloride anions was verified. by the AgNO3 test. Pillared montmorillonite was also obtained by using part of the acid hydrolysed Ti(OC3H7)4 solution. The final Ti to clay ratio was 20 mmol Ti/g of clay or 20 mmol Ti/0.87 meq of clay. II. PHYSICAL MEASUREMENTS 1. XRD, BET Surface.Area and Pore Size Analysis A Phillips X-ray diffractometer with Cu-Ka radiation was used to measure the d(001) basal spacings of pillared products. A thin film of the sample was made by drying a 1% solution on a glass slide at 700C for 30 min followed by air-drying. Surface area measurements were performed on a Quantachrome Quantasorb Jr. at liquid nitrogen temperature with nitrogen as the adsorbate and helium as the carrier gas. The curves were made using the Plotit program while the BET surface area was measured in the .05-.25 region of relative pressure. Samples were outgassed.au: 2000C for 12 hours under vacuum. Pore size distribution diagrams were determined from the desorption part of the isotherm using the model for parallel pores. The lowest relative pressure used was 0.1. “ ““‘“1 M P 143 2. Continuous Excitation Raman Experiment. The room temperature Raman spectrum for the titania pillared fluorohectorite was collected using the 514.5 A green line of an argon-ion laser. We used a near-backseattering geometry with the incident beam polarized in the scattering plane and collecting' all polarizations of the scattered light. The spectra were analyzed using a triple-dispersive Raman system consisting of a single-stage 0.5-m monochromator (filter stage) followed by a 0.85-m double monochromator. A GaAs photocathode photomultiplier was used in the photon counting mode for detection. Spectra were typically signal averaged. over 14 to 14 11 periods, with increased averaging obtained by repeated scanning. 3. Pulse Excitation Raman Experiment. The Raman spectrum for the titania pillared montmorillonite was measured on powdered samples mounted in capillary tubes by using a time-resolved technique with a pulsed picosecond laser system in order to minimize interference of the Raman scattering due to fluoresence. This technique is based on the fact that the Raman signal occurs in a much faster time regime than the fluoresence arising from impurities in the clay. This allows us to time-resolve the Raman signals from the fluoresence. Spectra were obtained by integrating the measured signals in the short time region only. The system consisted of a YAG Laser (Quantronix 416) with a frequency doubler, which operates as a pump source for a cavity-dumped 144 ultrafast dye laser (Coherent 702-2CD) of pulse width about 6 ps with an average power of output of 10 mW. A time- correlated single-photon counting system with a double monochromator (Jarell-Ash 25-100) was used for the measurements. The experiments were performed at room temperature with the scattered light collected at 90° to the direction of the excitation light. 4. Thermal Analyses Thermal gravimetric analyses of samples under an argon atmosphere were performed on a Cahn Thermogravimetric Analyser interfaced with an IBM-XT computer. Differential scanning calorimetry was carried out under a nitrogen flow on a DuPont 9900 thermal analyzer. The heating rate for both TGA and DSC was 6OC/min. 5. FTIR Fourier transformed infrared spectra were recorded on an IBM IR spectrometer in the 4000-400 cm"l region. The KBr pellets were approximately 1% by weight. An eleven point one pass smooth was performed on the spectra. 6. Elemental analysis Elemental analysis was performed on a sample made by fusion of pillared clay with lithium metaborate.The sample was first dissolved in 30 ml of 5% nitric acid and then diluted tO 100 ml with distilled water. 145 C. RESULTS AND DISCUSSION SYNTHESIS AND POWDER XRD: OF PILLARING SOLUTIONS The preparation of the pillaring reagents was carried out as described in the experimental section. It is very interesting that the powder x-ray diffraction of the solids that result from evaporation of pillaring solutions with Ti to HCl molar ratios in the range of 0.28 to 1.33 is always the same (Figure 1A and 18). Furthermore, the nature of the titania precursor, TiC14 or Ti(i-OC3H7)4, doesn't affect the intensity or position of the xrd reflections either. These xrd patterns are remarkably similar to that of TiBr(OH)3.1.5H20. Thus the d-spacings (A) (and the relative intensities of the peaks, in parenthesis) at: 8.84 (60), 8.58 (25), 5.15 (25), 4.07 (18), 3.95 (26), 3.86 (80), 3.32 (100), 3.16 (30), 2.49 (40) and 2.42 (80) are similar to the ones reported for TiBr(OH)3.1.5H2O. An additional peak at 11.3 A in our xrd patterns is attributed to the presence of some titanium oxide chloride hydrate phase, Ti29042Cl32.110H20 in our samples. This phase has its most intense peak at 11.3 A. In addition elemental analysis gives a Cl/Ti ratio of 1.14/1. The formula of [Ti8012(HZO)24]C18.HC1.7H20 can be also written as TiCll.1(OH)2.9.2.5HZO. Therefore it can be concluded that 146 WWW RELATIVE INTENSITY T4 UTIITrrrrTIIjUIIrVIIUr' 6 12 18 24 30 36 42 DEGREES 2 THETA Figure v.1. XRD (Cu-Kc) diffraction patterns of solids that result from evaporation of hydrolyzed TiCl4 solutions in the presence of HCl : (A) 1.7 M Ti and 6 M HCl, (B) 0.82 M Ti and 0.6 M HCl 147 evaporation of these pillaring solutions, at room temperature, favors crystallization of the octameric [Ti8012(H20)24]8+ cation. Since the chemical formula of Ti29042C132.110H20 can be written as T13011,58C13,3230-3320: it may be concluded that the peak at 11.3 A is unique for the octameric cation. Although no crystal structure has been reported for solids with Ti:(Br or Cl) ratios of 1:1, Reichmann M.G. and. Bell A.T.11 recently isolated single crystals of [Ti8012(HZO)24]8+ salts from solutions with similar Ti:Cl ratio (1:1). It must be also noticed that reflections due to TiOz(B) phase were absent in the xrd pattern of Figure 1.13 SYNTHESES AND POWDER XRD: OF PILLARED CLAYS (l). Titania pillared fluorohectorite from acid hydrolyzed TiC14. The x-ray diffraction pattern of fluorohectorite pillared with titanium polyoxocations prepared from acid hydrolyzed TiCl4, is shown in Figure 2. The basal spacing is approximately 21.6 A, corresponding to a pillar height of about 12 A. (2). Titania pillared montmorillonite from acid hydrolyzed TiC14 and Ti(OC3B7)4. Analogous results were obtained for the reaction of acid hydrolyzed TiC14 with Na-montmorillonite clay. The basal spacing of the air dried product is 24 A, Figure 3A. This value is a little higher than the one obtained in the case of Li-fluorohectorite clay. Since the latter is a high I . - 'v—DfiL-fi ' v e.‘ 148 4500 d 21.7 A Q ”0°" 10.8 0 V .1 a 2 If." g 1500- 0 l— I l I 2- 4 6 8 1B 72 151 DEGREES 2 THETA Figure v.2. Powder X-ray (Cu-KG) diffraction pattern and d- spacings (A) of titania pillared fluorohectorite. 149 23.88 12.21 a I \ 24.0 o ,, v * 12.27 E ‘ A I m 12.48 E 24.0 I e . B I “>1 . I r3- 23'“ 12.04 I LLI 0: 1 C I D 2 E 31 T 8 ' 8 7 10 I 12 DEGREES 2 THETA Figure V.3. XRD (Cu-1(a) patterns and d-spacings (A) of titania pillared montmorillonite samples, prepared : (A) from hydrolysis of TiC14 with HCl, (B) by drying sample A at 500 °C for 3 h, (C) from hydrolysis of Ti(OC3H7)4 with HCl, (D) by drying sample C at 500 0C for 3 h. 150 charge density clay the attractive forces between pillars and layers are expected to be larger. This will result in a smaller basal spacing, 21.6 A versus 24 A. When the pillaring agent is derived from Ti(i-OC3H7)4, as described by Yamanaka4, it produces pillar product with basal spacing similar to that resulting from the TiCl4 precursor (~14 A, Figure 3C). (3). Effect of heating pillared.montmorillonite prepared from TiCl4 and Ti(OC3B7)4. After the sample of pillared montmorillonite, obtained by using acid hydrolyzed TiCl4 as the pillaring agent, had been calcined at 500 c’C, the basal spacing remains almost the same, ~23.3 A (Figure 38). Similar results were obtained for when acid hydrolysed Ti(OC3H7)4 was the pillaring agent. Although. the basal spacing slightly decreases on heating, it maintains a value close to 23.0 A at 500 0C (Figure 3D). The fact that the gallery height is altered only slightly (i.e., compare Figures 3A, 3B, 3C and 3D) no matter what the precursor-alkoxide or TiCl4-suggests that the nature of the intercalated titanium polyoxocations is very similar. (4). Elemental analysis Elemental analysis of the products indicate that a large percentage of the Mg was dissolved during the pillaring of Li-fluorohectorite clay 151 [(Ti02)7.78][M92.65Lio.711(518020)(OH'FI4° This was not the case with the Na-montmorillonite : IITiOZI4.79INaI.O3I[A12.87M9.94Fe°19I[Si8OZOIOHI4I BET SUREACE AREA.AND PORE SIZE ANALYSIS The identical adsorption-desorption isotherms (Figures 4A, 4B and 4C) also support our observation. Both pillaring solutions result in products with similar adsorption- desorption isotherms, which exhibit significant hysterysis upon desorption. The BET surface areas-when the same clay is used, namely Li-fluorohectorite- are close to 240 mZ/g (estimated from the adsorption isotherm of Figure 4A for values of the relative pressure P/Po ranging from 0.06 to 0.25). When Na-montmorillonite is used, the adsorption- desorption isotherms (Figure 4B and 4C) have similar shape but the BET surface area of the pillared product is 300 mZ/g. Thus the lateral seperation of the pillars is larger, as one would expect for a clay of lower charge density, in contrast to Li-fluorohectorite. Pore size analysis of the desorption isotherm for Li- fluorohectorite is shown in Figure 5A. It appears that most of the pores have diameters in the range of 12-50 A. The gallery height is about 12 A. However the large hysterysis loop suggests bottle-neck structures. Local delamination as suggested by Pinnavaia et. 81.1 is also possible. Thus there is substantial contribution ‘to time mesopore volume from intraparticle pores, in addition to interparticle pores. 152 V (mL.g'l) p/p. Figure‘v.4. Adsorption-desorption isotherms of titania pillared : (A) fluorohectorite, prepared from hydrolysis of TiCl4 with HCl, (B) montmorillonite, prepared from hydrolysis of Ti (OC3H7) 4 with HCl, (C) montmorillonite, prepared from hydrolysis of TiCl, with HCl. 153 Figure v.5. Pore size distribution diagrams of titania pillared : (A) fluorohectorite, prepared from hydrolysis of TiCl4 with HCl, (B) montmorillonite, prepared from hydrolysis of Ti (OC3H7)4 with HCl, (C) montmorillonite, prepared from hydrolysis of TiCl, with HCl. 154 In addition pore size analysis of the two pillared montmorillonite clay samples synthesized by using two different pillaring agents Ti(OC3H7)4 and TiCl4, showed very similar pore diameters (Figures SB and 5C). Most of the pores had diameters in the range of 12-50 A. Therefore 1 partial delamination occurs or bottle-neck pores are present in these products too. RAMAN SPECTROSCOPY Raman spectroscopy has been recently shown to be a very useful tool to identify species that result from the acidic hydrolysis of TiCl4 or Ti(OC3H7)4. The high scattering power of the products in the Raman region allowed Reichmann M.G. and. Bell A.T. to probe highly dispersed states of these hydrolysis species in supports such as silica and alumina.11'12 They were able to identify [Ti8012(HZO)24]8+ cations in silica. This octameric cation exhibits three strong absorptions at 371, 512 and 905 cm-1 due to Ti-O-Ti vibrations. In our effort, to understand the nature of the pillars in titania pillared clays, we decided to use Raman specrtoscopy and relate our results with those obtained for titanium polyoxocations from acid hydrolysed TiCl4 or Ti(OC3H7)4. It is very important that when the dispersed crystallites on a support are less than 40 A.1Jl diameter, powder XRD is not as useful as Raman spectroscopy to characterize them. Thus, although. both the octameric cation and the Ti02(B) or anatase have very characteristic TIT—tr 155 XRD peaks, no reflection due to these phases is present in the XRDs of the pillared products, Figures 2 and 3. Continuous Excitation The continuous excitation Raman spectra of Figure 6A was obtained as described in the experimental section. The Raman spectra of SiOZ supported Titania is also shown in Figure GB for comparison (it was taken from recent work of M. Reichmann and A.T. Bellll). The similarity of both spectra is obvious. We must also consider the higher fluoresence background in the case of the clay Fluorohectorite. Other clays such as Na-montmorillonite have even higher fluoresence background and other techniques should be used (see next section). The peaks of spectrum in Figure 6A suggest the presence of TiOZ(B) phase in our sample. It was dried at 70 0C, while the silica supported titania (Figure 6B) was dried at 550 OC. The characteristic absorption of octameric [Ti8012(H20)24]8+ cations at about 371 cm.1 is absent, due to their connection to form the TiOZ(B) phase. This connection of the octameric cations to form pillars results in changes in intensity, lineshape and position of its peaks. The absence of three strong anatase absorptions in the 200-600 cm'1 region, in titania pillared fluorohectorite, indicate absence CM? anatase. Because the scattering cross section of TiOZ(B) is much smaller than that of the anatase it may be concluded that the sample contains almost quantitatively TiO2(B) phase. However I 156 '162 --l47 A 00 4 \ (J v 17)" I! 4 E 635 E I . E ‘126 E A .11 29° 61° (K M 401 517 B 100 250 400 550 700 850 1000 WAVENUMBERS (cm—1) Figure v.6. Continuous Excitation Raman Spectra of : (A) titania pillared fluorohectorite dried at 70 0C, (B) titania supported on SiOz dried at 550 °C, as described in ref. 11. 157 presence of very small amounts of anatase should not be ruled out. The assignments for the peaks are as in Table V.I. The ratio» of 'the intensities of the two (peaks is similar to the ratio of the intensities of the corresponding peaks in silica supported titania, Figure 68. Peak assignment for the spectra in Figure 6B has been given elsewhere.11 TABLE v;I. Peak assignment of Titania pillared Fluorohectorite, [(Ti02)7.78][Mg2.65Lio.71](Siaozo)(OB,F)4, dried at 70 °c. v (cm'l) Phase 162 T102 (B) 635 110218) Pulse Excitation The Picosecond Raman Spectra of Titania pillared Montmorillonite is shown in Figure 7. It verifies the presence of a peak around 635 cm-1 which can be due to either anatase or TiOZ(B) phase. The two peaks at lower Raman shifts are not well defined due to high noise to signal ratio. However they are located exactly where one would expect for anatase and TiOZ(B) phase and its precursor [Ti8O12(H20)24]8+. Based on recent results for titania 158 460 , 1 l l 1 I 1311 I 1 I I 313 I 3 m - " "-.._4?9 632 - _ 11 m - ' . (1 1 . t 5 3“ "' ..1'..".. :1 3111 1 . I. 1 H In a 3111 m m 15111 11111111111111 (l/cn) Figure ‘V.7. Pulse Excitation Raman Spectra of Titania pillared Montmorillonite dried at 90 °C. 159 11 polyoxocations we can assign peaks as shown in the following table: TABLE szI. Peak assignment of Titania pillared Montmorillonite, [ (T102) 4 . 79 (N3) . 03] [A12 . 87849 . 94FO . 19] [Si8020(OB)4], dried at 90 °c. v (cm 1) Phase 393 Octameric cation, Anatase 499 Octameric cation, Anatase, TiOZ(B) 632 Anatase, TiOZ(B) 908 Octameric cation The resolution was about 10 cm‘l. It appears that the clay stabilizes the 'TiOZ(B) ‘phase. However' even at 900C anatase is present. As mentioned earlier it is dificult to estimate their relative amounts. In addition, contribution to some of the peaks, from the [Ti8012(H20)24]8+ cation is also possible. We propose the model of Figure 8 which explains the observed d(001) spacings of Titania Pillared Clays. In support of that model is the observation that octameric titanium cations can come in close proximity to each other and eliminate chloride anions and water molecules in order to connect themselves and form the Ti02(B) phase (Figure 9). Depending on the number of the cations that are connected we 160 / \ \ / / \ / \ Figure‘v.8. Proposed model for the explanation of the different gallery heights observed for the titania pillared layered silicate clays. 161 0 = titanium O = oxygen octameric of the relationship Structural Figure v.9. titanium cation and T102(B) phase. 162 can get different incremental gallery heights as shown in Figure 8. Within the TiOz(B) there is a lot of void volume and the main structure of the octameric cations is retained (Figure 10). Further movement of these building units will result in a phase condensation to form anatase as described by Tournoux et. al.13 In this latter phase the octameric cations don't retain their integrity. It has been claimed10 that when TiCi4 is hydrolyzed in the presence of HCl and the resulting solution is used for pillaring of Na-montmorillonite, the pillared products exhibit gallery heights in the range of 18 to 23.7 A. The resulting basal spacing remains as large as 28 A even at 500 C’C and can be also justified by using the model of Figure 8. THERMAL ANALYSIS In the thermogravimetric analysis and. differential scanning calorimetry the heating rate was 6°C/min. The TGAs of Titania Pillared Montmorillonite (Figure 11) are the same no matter what the pillaring agent is. (i.e., compare with TGA of Ref 10). A continuous weight loss is observed upon heating up to 5000C due to condensation of the octameric cations -in addition to the loss of physisorbed water. A small step after that temperature is probably due to dexydroxylation of the montmorillonite layers and the additional anatase phase-as DSC also suggests. 163 Figure V310. Structural relationship of the Tioz(B) phase (left) and anatase (right). 164 Weight (mgs) 4E7 " r " r‘ ‘* I . I 1* r ' 25 185 345 505 665 825 985 Temperature ('C) Figure V311. TGA of titania pillared montmorillonite prepared by using the Ti(OC3H7)4 precursor. 165 The DSC experiments (Figure 12) show a very broad endothermic peak centered at about 80°C and a smaller one near 2200C. Although the former is expected to be broad-due to the stronger solvation of octameric titanium cations-it occurs at lower temperature than expected (1200C).1 Since our heating rate is sufficiently low to allow peak separation this endotherm suggests the coexistance of two phases in accordance with the pulse Raman experiment. The facile water loss of the addititonal anatase phase-shown by the first derivative with respect to temperature-overshadows the water loss of the titania pillars. In addition, it seems that we have a lower temperature shift due to specific interactions of the pillars with the clay layers. The latter' peak: at about 2200C looks more like a shoulder and it is probably due to further dexydroxylation of the interlayred. titania. Finally the third. endotherm which starts somewhere at 3000C is very large to be accounted for a baseline shift. Since such a feature has not been observed in pillared clays before it can be inferred that it is due to the anatase phase. It also covers the peak which we expect to see due to the dehydroxylation of the layers. FTIR .As mentioned before, the FTIR spectra of solids that result from slow evaporation of various pillaring solutions are identical to that of Figure 13. The main feature of 1(56 .Ao>u:o own on» no m>«um>aump umuam one mucomoumou as: .3on on“: uomuaooum :02. one means an compound opacozque—ucos pow—waded 32835 no Umo .Nfl.> ensue...- .u.. arse-Leeeeh can con oov com com can a moo.o1 I . I t 11 . .11 1+ .1 .r. 11, v.o1 a 8o...L r . _ rm.o1 m N . A moo o- m_ r . a... .4... o«o.o..4 IN.OI o . H 1. l n I ' Jmeoei mu 1 1«.e1 omo.o- h mwo.o o.o (BIN) More anon .167 28338 83:03. ooumuoogo no 533% mean .25 3.5:. 0.00? mmwmznzm>