MSU LIBRARIES -__ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES wiII be charged if book is returned after the date stamped below. IMMOBILIZATION OF METAL CLUSTER CARBONYL COMPLEXES ON LAYERED SILICATES By Emmanuel P. Giannelis A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1985 ABSTRACT IMMOBILIZATION OF METAL CLUSTER CARBONYL COMPLEXES ON LAYERED SILICATES By Emmanuel P. Giannelis Several metal cluster carbonyl complexes have been immobilized on Na+—hectorite and alumina pillared montmorillonite and characterized by a variety of spectroscopic techniques. Adsorption of the protonated metal cluster carbonyl H083(CO)12+ on Na+-hectorite leads to a preferential dispersal of 033(CO)12 centers at the hydroxylated edge sites of the silicate layers. Under conditions where surface migration is slow, the edge bound clusters react with surface hydroxyl groups to form complexes of the type [Os(CO)2,3(O-Si."-:)2]n, whereas clusters adsorbed on basal planes are stable. The difference in reactivity of 053(CO)12 with different clays and the dependence on the drying method is rationalized in terms of a clay flocculation model in which both lamellar (face-face) and delaminated (edge-edge, edge-face) interactions of the layers can take place, depending in part on the layer-morphology. The phosphonium-phosphine ligand PhZPCHZCHZSPhZCHZPh (abbreviated P-P+) has been used as a means of inducing positive charge on otherwise neutral metal carbonyl clusters. RU3(CO)9(P-P+)3, H4RU4(CO)3(P-P+)4, 1r4(co)9(P-P+)3, OS3(c0)11(P-P+), and H2OS3(co)g,10(P-P+) have been Em manuel P. Giannelis intercalated in hectorite and characterized by IR and electronic spectroscopy, x-ray diffraction and elemental analysis. The intercalated H2053(CO)9(P—P+)-hectorite has been used as an olefin isomerization catalyst. The reaction is found to be dependent on the extent of interlayer swelling. In toluene, the intercalated catalyst exhibits a pseudo first-order behavior with kobs = 0.022 h'l. The mechanism of metal cluster binding to alumina pillared montmorillonite has been investigated by IR and electronic spectroscopy. 053(CO)12 and RU3(CO)12 are found to form protonated species of the type HOS3(CO)12+ and HRU3(CO)12+ which can be extracted from the surface by ion exchange with KPF6. In contrast, H4RU4(CO)12 and Ir4(CO)12 are simply physisorbed on the surface of the clay. Significantly, H2083(C0)10 reacts with the pillared clay to form a pillar-grafted hydrido triosmium cluster of the type HOS3(CO)10(O-Al < ). In addition, adsorption of [CpFe(CO)2]2 on the surface leads to cationic species of the type [CpFe(CO)2,3]+ electrostatically bound to the clay. TO MY PARENTS ii ACKNOWLEDGEMENTS I would like to thank Dr. T.J. Pinnavaia for his guidance, independence, and encouragement offered throughout the years of graduate school. I am also grateful to Dr. H.A. Eick for his editorial assistance in improving my writing. Helpful discussions with Dr. D.G. Nocera are particularly acknowledged. Finally, I would like to thank the members of Dr. Pinnavaia's research group whose help, advice, and encouragement made graduate school a unique experience. iii TABLE OF CONTENTS Chapter Page LIST OF TABLES ......................................................................... viii LIST OF FIGURES ........................................................................ x CHAPTER I. INTRODUCTION ....................................................... 1 A. Structure and Properties of Layered Silicates .................... 1 1. Structure .................................................................. 1 2. Ion Exchange ............................................................ 5 3. Swelling of Smectites ................................................ 5 4. Acidity . ................................................................... 8 5. Catalysis and Interlayer Dynamics ............................... 9 6. Pillared Clays ........................................................... 10 B. Concept of Supported Metal Complexes ............................. 11 C. Some Aspects of Cluster Chemistry and Catalysis .............. 16 1. Metal Cluster Carbonyls ............................................ 16 2. Catalysis by Metal Clusters ........................................ 18 D. Supported Metal Cluster Carbonyls ........... .................... 20 l. Immobilized Metals, Mononuclear Complexes , and Cluster Compounds on Solid Supports..... ..................... 20 2. Anchored Metal Carbonyls on Functionalized Polymers . 23 3. Immobilized Metal Carbonyls on Inorganic Supports ...... 26 E. Research Objectives ........................................................ 29 iv Chapter CHAPTER II EXPERIMENTAL ...................................................... A. Materials ......................................................................... 1. Natural Hectorite ....................................................... 2. Sodium Montmorillonite (Wyoming) .............................. 3. Laponite® .................................................................. 4. Fluorohectorite .......................................................... 5. Solvents .................................................................... 6. Reagents ................................................................... B. Syntheses ......................................................................... 1. H2083(CO)10 ............................................................. 2. H2053(CO)10(PPh3) .................................................... 3. H2053(C0)9(PPh3) ...................................................... 4. [HOS3(CO)12](PF5) ...................................................... 5. [HRU3(CO)12]( PF6) ..................................................... 6. PhZPCHZCHgPPhZCHZPhBr ......................................... 7. BF4'-Resin ................................................................ 8. (PhZPCHZCHZPPh2CH2Ph)(BF4). .................................. 9. [RU3(CO)9(P—P+)3](BF4)3 ............................................. 10. [033(CO)11(P-P+)](BF4) .............................................. 11. [Ir4(CO)9(P-P+)3](BF4)3 .............................................. 12. lH4RU4(CO)8(P—P+)4](BF4)4 ......................................... 13. [H2033(CO)10(P-P+)](BF4) .......................................... 14. [H2053(CO)9(P-P+)](BF4) ............................................ C. Clay-Cluster Reactions ...................................................... 1. Reaction of Clay with HMg(CO)12+ or M3(CO)12 (M = Os, Ru) .............................................................. Page 36 36 36 36 36 37 37 37 38 38 38 38 38 39 39 39 39 40 40 40 41 41 4] 42 42 Chapter 2. Intercalation of Phosphine Substituted Cluster Carbonyls on Na+-hectorite ......................................... 3. H2033(CO)9(P-P+)-hectorite ....................................... 4. HOS3(CO)9(CH=CH2)(P—P+)-hectorite .......................... 5. Alumina Pillared Clay (ALPILC) .................................. 6. Reaction of ALPILC with M3(CO)12 (M = Os, Ru) ........ 7. Extraction of Hl\13(CO)12+ Clusters from the Clay ........ 8. Wet Impregnation of Clays with Neutral Clusters .......... 9. Powder Pyrolysis Experiments ..................................... 10. Catalytic Reactions ................................................... D. Physical Measurements ..................................................... 1. Infrared Spectroscopy ................................................. 2. X-Ray Diffraction Studies .......................................... 3. UV-Visible Spectroscopy .............................................. 4. Proton NMR Spectroscopy .......................................... 5. Gas Chromatography .................................................. 6. Melting Points .. ......................................................... 7. Elemental Analyses .................................................... CHAPTER III RESULTS AND DISCUSSION ..................................... A. Surface-Selective Dispersion of Cluster Carbonyls on Layered Silicates .............................................................. 1. Reaction of HOS3(CO)12+ and 053(CO)12 with Clays 2. Reaction of HRu3(co)12+ and RU3(CO)12 with Hectorite .......... . ....................................................... Intercalation of Cationic Phosphine Substituted Carbonyl Clusters on Hectorite ........................................................ 1. Synthesis and Exchange Reactions ................................ vi Page 42 42 42 43 43 43 44 44 44 44 44 45 45 45 46 46 46 47 47 47 77 85 85 Chapter 2. Polarized Infrared Studies of H4RU4(CO)8(P-P+)4" hectorite ................................................................... 3. Olefin Isomerization by Homogeneous and _ Hectorite-Intercalated H2053(CO)9(P-P+) Catalysts ..... 4. Thermal Stability of Hectorite-Intercalated Metal Cluster Carbonyls and Interaction with H2 and 02 ........ Cluster Carbonyl Interactions with Alumina Pillared Clay (ALPILC) .................................................................. 1. Adsorption of 053(CO)12 on ALPILC ........................... 2. Adsorption of RU3(CO)12 on ALPILC ............................ 3. Adsorption of H2033(CO)10 on ALPILC ........................ 4. Adsorption of H4RU4(CO)12 on ALPILC ....................... 5. Adsorption of Ir4(CO)12 on ALPILC ............................. 6. General Considerations on the Adsorption of Cluster Carbonyls on ALPILC ...................................... 7. Adsorption of [CpFe(CO)2]2 on ALPILC ....................... Laponite-Supported Metal Cluster Carbonyls ....................... 1. Adsorption of H2053(CO)10 and 053(CO)12 on Laponite . 2. Adsorption of RU3(CO)12 and H4RU4(CO)12 on Laponite. vii Page 107 110 115 129 130 137 140 141 144 144 156 164 164 168 Table 10 11 12 13 14 LIST OF TABLES Advantages and Disadvantages of Homogeneous and Heterogeneous Catalysts ................................................ Polymer—1m mobilized Cluster Carbonyls of Ru, Os, and Ir ............................. . .............................................. Selected Studies of Immobilized Cluster Carbonyls of Ru, Os, and Ir on Inorganic Supports ................................. Infrared CO Stretching Frequencies for Reaction Products and Reference Compounds ................................ Electronic Spectroscopy Data for Reaction Products and Reference Compounds .............................................. X—Ray Diffraction Data for OS3(CO)12 and Clay-Supported Osmium Carbonyl Samples .............................................. Elemental Analysis of Os Complexes Supported on Hectorites ..................................................................... Infrared CO Stretching Frequencies for Reaction Products and Reference Compounds .............................................. Electronic Spectroscopy Data for Reaction Products and Reference Compounds ................................................ Infrared CO Stretching Frequencies for Molecular and Hectorite-Intercalated Metal Cluster Complexes -------------- UV-Visible Spectroscopy Data for Hectorite Exchanged Metal Clusters and Their Molecular Analogs ---------- . --------- Infrared CO Stretching Frequencies and Electronic Absorptions of Cluster Carbonyl Compounds ..................... 001 X-Ray Basal Spacings of Hectorite-Intercalated Metal Cluster Carbonyl Complexes .................................... Elemental Analysis of Hectorite Intercalated Cluster Carbonyls ......................... . ............................................ viii Page 14 27 30 50 52 71 76 80 81 87 89 90 96 106 Table Page 15 Vibrational Frequencies for Some Carbonyl Metal Complexes .................................................................... 128 16 Infrared CO Bands of Molecular and ALPILC- Intercalated Metal Cluster Complexes .............................. 133 17 Electronic Spectroscopy Data for Reaction Products and Reference Compounds .............................................. 135 18 Infrared CO Bands of Molecular and ALPILC- Intercalated Metal Cluster Complexes ............................. 157 19 C5H5 Proton Chemical Shifts for Some Cyclopenta- dienyliron Carbonyl Compounds ....................................... 161 20 Infrared CO Bands of Molecular and Laponite Supported Metal Cluster Complexes ................................................. 167 ix Figure LIST OF FIGURES Idealized structure of a smectite clay mineral. (0) Oxygen atoms; (0) hydroxyl groups. Silicon and sometimes aluminum normally occupy tetrahedral positions in the oxygen framework. Aluminum, magnesium, iron, or lithium may occupy octahedral sites. Mm-xHZO represents the interlayer exchange cation. .......................... Preparation methods for phosphine-functionalized poly(styrene-divinylbenzene). ....................... . ..................... Infrared spectra in the terminal CO stretching region: (I) 053(CO)12-hectorite prepared from HOs;3(CO)12+ (film); (II) [HOS3(CO)12]PF6 (KBr pellet); (III) 053(CO)12 (KBr pellet); (IV) 053(CO)12-hectorite prepared by impregnation with 053(CO)12 (film); W) sample I after heating in air at 150°C for 12 h. . ....................................... Schematic illustration of the two types of external surfaces for a smectite clay. The edge surfaces are hydroxylated whereas the basal planes contain only siloxane oxygens. Sodium exhcange ions (not shown) occupy mainly the interlayer regions. ................................ Page 49 56 Figure Infrared spectra in the terminal CO stretching region: (a) Freeze-dried sample of 053(CO)12 on Na+-hectorite prepared by impregnation with 053(CO)12 (mull); (b) after heating in air at 150°C for 12 h (mull). ....................... (A) Model for a wet, flocculated clay system after impregnation with 053(CO)12. Cycles represent aggregates of 053(CO)12 cluster molecules of unspecified size; (B) Freeze-drying tends to preserve the flocculated structure. The sample contains both laminated and delaminated platelets; (C) Air-drying tends to promote face-to-face lamination of layers. Clusters are "trapped" and migration to edge sites is impeded. .............................. Infrared spectra in the terminal CO stretching region: (a) OS3(CO)12-hectorite prepared by impregnation with 053(CO)12 in CHZCIZ (mull); (b) after heating in air at 150°C for 12 h (mull). .......................................... Infrared spectra in the terminal CO stretching region: (a) OS3(CO)12-fIUOPOheCtOI‘ite prepared by impregnation with 053(CO)12 and air-drying (film). The Os loading is 3.8 wt%; (b) after heating in air at 150°C for 12 h (film). .......................................................................... Infrared spectra in the CO stretching region: (a) 053(CO)12- -laponite prepared by impregnation with 053(CO)12 and air-drying (mull). The Os loading is 3.8 wt%; (b) after heating in air at 150°C for 2 h (mull). ........................ xi Page 62 64 67 70 70 Figure 10. 11 12 13 14 Page Infrared spectra in the terminal CO stretching region: (a) clay sample prepared by refluxing 053(CO)12 with freeze-dried Na+-hectorite in octane (mull); (b) sample (a) after heating in air at 150°C for 2 h (mull). .................. 74 Infrared spectra in the terminal CO stretching region: (VIII) RU3(CO)12-hectorite prepared from HRU3(CO)12+(fi1m); (IX) [HRU3(CO)12]PF6(KBr pellet); (X) RU3(CO)12 (KBr pellet); (XI) RU3(CO)12-hectorite prepared by impregnation with RU3(CO)12 (film).. ................................. 79 Infrared spectra in the terminal CO stretching region: (a) Freeze-dried sample of RU3(CO)12 on Na+-hectorite (0.1 mmol/meq) prepared by impregnation with RU3(CO)12 (mull); (b) sample (a) after heating in air at 100°C for 30 min (mull). ............................................................. 84 Infrared spectra of unsupported and hectorite-intercalated Ir4(CO)9(P-P+)3 in the region 2500-500 cm'l: (a) Ir4(CO)9- (P-P+)3(BF4)3 (KBr pellet); (b) Ir4(CO)9(P-P+)3-hectorite (film). Cross—hatched absorptions are also present in the native mineral. ........................................................ 93 X—ray diffraction patterns of : (A) Ru3(CO)9(P-P+)3-hectorite; (B) H4RU4(CO)8(P-P+)4-hectorite; (c) Ir4(CO)9(P-P+)3-hectorite; (D) 053(CO)11(P-P+)-hectorite; (E) HZOS3(CO)10(P—P+)- hectorite; (F) H2033(CO)9(P-P+)-hectorite. Samples were prepared as oriented films on glass slides by first suspending the clay complex in water and allowing the suspension to evaporate on the slide at room temperature. 98 xii Figure 15 16 17 18 19 20 Page Infrared spectra in the CO stretching region of unsupported and hectorite-intercalated RU3(CO)9(P-P+)3: (a) RU3(CO)9- (P-P+)3(BF4)3 in CH2012 solution; (b) Ru3(co)9(P-P+)3—- hectorite (KBr pellet). ....................................................... 101 Infrared spectra in the CO stretching region of unsupported and hectorite-intercalated H4RU4(CO)8(P-P+)4: (a) H4RU4(CO)3(P-P+)4(BF4)4 in 0112012 solution; (b) H4RU4(CO)8(P-P+)4-hectorite (KBr pellet). ........................ 101 Infrared spectra in the terminal and bridging CO stretch- ing region of unsupported and hectorite-intercalated Ir4(CO)9(P-P+)3: (a) Ir4(CO)9(P-P+)3(BF4)3 in 0112012 solution; (b) Ir4(CO)9(P-P+)3-hectorite (KBr pellet). ........... 103 Infrared spectra in the CO stretching region of unsupported and hectorite-intercalated 053(CO)11(P—P+): (a) 033(CO)11- (P-P+)BF4 in CH2C12 solution; (b) OS3(CO)11(P-P+)-- hectorite (KBr pellet). ..................................................... 103 Infrared spectra in the CO stretching region: (a) H2053(CO)10(P-P+)(BF4) in 0112012 solution; (b) H2053(CO)10(P-P+)-hectorite (KBr pellet); (c) H2053(CO)9(P-P+)(BF4) in CH2012 solution; ((1) H2053(CO)9(P-P+)-hectorite (KBr pellet). ...................... 105 Infrared absorbance spectra in the CO stretching region of H4RU4(CO)g(P-P+)4-hectorite at different angles of clay film relative to the direction of the IR beam: (a) 6 = 90°. The IR beam is perpendicular to the clay film; (b) e = 110°; (c) e = 130°. ........................................ 109 xiii Figure 21 22 23 24 Isomerization of l-hexene (1-3.3 M) to 2—hexene with H2OS3(CO)9(P—P+)-hectorite (0.01 mmol) in three different liquid media at ambient temperature and pressure. The clay catalyst contained 2.17 wt% P and 19.27 wt% Os. ............................................................ Catalytic cycle for isomerization of 0! —olefins catalyzed by triosmium clusters679107. The carbonyl ligands are omitted for simplicity. The H053(vinyl)(CO)10 is catalytically inactive. ................................................... Infrared spectra in the CO stretching region of products formed by thermal decomposition of RU3(CO)9(P-P+)3- hectorite: (a) Ru3(CO)9(P-P+)3-hectorite (film); (b) sample (a) after heating in vacuum at 110°C for 11 h; (0) followed by heating at 200°C for 9 h; ((1) 300°C for 10 h; (e) sample (a) after heating in H2 at 200°C for 10 h; (f) followed by heating at 300°C for 11 h ................ Infrared spectra in the CO stretching region of products formed by oxidation of Ru3(CO)9(P-P+)3-hectorite in air: (a) Ru3(CO)9(P-P+)3-hectorite (film); (b) sample (a) exposed to air for 2 months; (c) sample (a) after heating in air at 110°C for 10 h; (d) followed by heating at 200°C for 45 min; (e) 200°C for 90 min; (f) 200°C for 3% n; (g) 200°C for 10 h; (h) 300°C for 1 h ..................... xiv Page 112 117 120 120 Figure 25 26 27 Page Infrared spectra in the CO stretching region of products formed by thermal decomposition of H4RU4(CO)3(P-P+)-- hectorite: (a) H4Ru4(CO)3(P-P+)4-hectorite (film); (b) sample (a) after heating in vacuum at 110°C for 11 h; (c) followed by heating at 200°C for 9 h; (d) 300°C for 10 h; (e) sample (a) after heating in H2 at 200°C for 10 h; (f) 300°C for 11h. ................................................ 122 Infrared spectra in the CO stretching region of products formed by oxidation of H4RU4(CO)3(P-P+)4-hectorite in air: (a) H4Ru4(co)3(P-P+)4—neotorite (film); (b) sample (a) after heating in air at 110°C for 10 h; (0) followed by heating at 200°C for 45 min; ((1) 200°C for 90 min; (e) 200°C for 3% h; (f) 200°C for 10 h; (g) 300°C for 1 n. .............................................................. 122 Infrared spectra in the CO stretching region of products formed by thermal decomposition of Ir4(CO)9(P-P+)3- hectorite: (a) Ir4(CO)9(P-P+)3-hectorite (film); (b) sample (a) after heating in vacuum at 200°C for 9 h; (c) followed by heating at 300°C for 10 h; ((1) sample (a) after heating in H2 at 200°C for 10 h; (e) 300°C for 11 h. .......................................................................... 124 XV Figure 28 29 30 31 Page Infrared spectra in the CO stretching region of products formed by oxidation of Ir4(CO)9(P-P+)3-hectorite in air: (a) Ir4(CO)9(P-P+)3-hectorite (film); (b) sample (a) after heating in air at 110°C for 10 h; (c) followed by heating at 200°C for 45 min; (d) 200°C for 10 h; (e) 300°C for 1 h. ............................................................. 124 Infrared spectra in the CO stretching region of products formed by thermal decomposition of OS3(CO)11(P-P+)-- hectorite: (a) 053(CO)11(P-P+)-hectorite (film); (b) sample (a) after heating in vacuum at 200°C for 9 h; (c) followed by heating at 300°C for 10 h; (d) sample (a) after heating in H2 at 200°C for 10 h; (e) 300°C for 11 h. ......................................................................... 126 Infrared spectra in the CO stretching region of products formed by oxidation of 033(CO)11(P-P+)-hectorite in air: (a) 053(CO)11(P-P+)-hectorite (film); (b) sample (a) after heating in air at 110°C for 10 h; (c) followed by heating at 200°C for 45 min; (d) 200°C for 90 min; (e) 200°C for 10 h; (f) 300°C for 1 h. ................................. 126 Infrared spectra in the CO stretching region: (a) H053(CO)12+/- ALPILC prepared by impregnation with 053(CO)12 (mull); (b) [H053(CO)1 21PF6 in MeN02 solution; (c) sample (a) exposed to air at RT for 5 h (mull); (d) 053(CO)12 in CH2C12 solution; (e) sample (c) after heating in vacuum at 100°C for 2 h (mull); (f) 250°C for 6 h; (g) sample (c) after heating in air at 110°C for 12 h. ......... 132 xvi Figure 3 2 33 34 35 Infrared spectra in the CO stretching region: (a) HRU3(CO)12+/‘ ALPILC prepared by impregnation with RU3(CO)12 (mull); (b) [HRU3(CO)]2]PF6 in MeNOz solution; (c) sample (a) exposed to air at RT for 5 h (mull); (d) RU3(CO)12 in CH2C12 solution; (e) sample (a) exposed to air at RT for 24 h (mull); (f) sample (c) after heating in vacuum at RT for 24 h. . ................................................................. Infrared spectra in the CO stretching region of H2053(CO)10 supported on alumina pillared clay: (a) H2033(CO)10 supported by adsorption from CH2C12 solution (mull); (b) sample (a) exposed to air at RT for 24 h; (c) sample (a) after heating in air at 110°C for 2 h; (d) sample (a) after heating in vacuum at 150°C for 4 h. ....................... Infrared spectra in the CO stretching region of unsupported H4RU4(CO)12 and H4RU4(CO)12 supported on alumina pillared clay: (a) H4RU4(C0)12 adsorbed on ALPILC from CH2C12 solution (mull); (b) H4RU4(CO)12 in CH2C12 solution; (c) sample (a) after heating in air at 100°C for 1 h (mull); ((1) sample (a) after heating in vacuum at 150°C for 4 h. ............................................................... Infrared spectra in the CO stretching region of unsupported Ir4(CO)12 and Ir4(CO)12 supported on alumina pillared clay: (a) Ir4(CO)12 adsorbed on ALPILC from cyclohexane solution (mull); (b) Nujol mull of Ir4(CO)12; (c) sample (a) after heating in air at 110°C for 2 h (mull). ..................... xvii Page 139 143 143 146 Figure 36 37 38 39 40 Schematic representation of reactions and their products of OS3(CO)12 and alumina pillared clay (ALPILC). ............... Schematic representation of reactions and their products of supported RU3(CO)12 on alumina pillared clay (ALPILC)... Infrared spectra in the CO stretching region: (a) clay sample prepared by impregnation of alumina pillared clay with [CpFe(CO)2]2 in CH2C12 (mull); (b) [CpFe(CO)3]- PF6 extracted from the surface of ALPILC with a solution of KPF6 in acetone (Nujol mull); (c) [CpFe(CO)2]- PF6 extracted from the surface with KPF6 in acetone (CH2012 solution). ........................................................... Infrared spectra in the CO stretching region of H2083(C0)10 supported on laponite (0.003 mmol/0.1 g): (a) H2083(CO)10 physisorbed on laponite from CH2C12 solution (mull); (b) sample (a) after heating in vacuum at 60°C for 4 h; (c) followed by heating in vacuum at 150°C for 3 h; ((1) sample (a) after heating under Ar at 150°C for 3 h; (e) sample (a) after heating in air at 100°C Infrared spectra in the CO stretching region of OS3(CO)12 supported on laponite (0.003 mmoI/0.1 g): (a) 053(CO)12 physisorbed on laponite from CH2C12 solution (mull); (b) sample (a) after heating in vacuum at 150°C for 3 h; (0) sample (a) after heating under Ar at 150°C for 3 h; ((1) sample (a) after heating in air at 150°C Page 148 150 159 166 170 Figure 41 42 Page Infrared spectra in the CO stretching region of RU3(CO)12 supported on laponite (0.003 mmol/0.1 g): (a) RU3(CO)12 on laponite prepared by impregnation from CH2C12 solution (mull); (b) sample (a) after heating in vacuum at 60°C for 4 h; (c) sample (a) exposed to air for seven days; (d) sample (a) after heating in air at 90°C for 30 min. . ............................................................................ 172 Infrared spectra in the CO stretching region: (a) H4RU4(CO)]2 on laponite (0.003 mmol/0.1 g) prepared by impregnation from CHZCIZ solution (mull); (b) sample (a) after heating in vacuum at 60°C for 6 h; (c) followed by heating at 120°C for 3 h; (d) sample (a) after heating in air at 90°C for 1 h. .................................................................. 172 xix CHAPTER I INTRODUCTION A. Structure and Properties of Layered Silicates 1. Structure The term "clay" refers to finely divided mineral sediments having particles with dimensions of less than 211. However, a material exhibiting certain properties such as plasticity, small particle size, and specific chemical composition (i.e. as consisting largely of silica, alumina, and water) might be loosely defined as clayl. The term "clay minerals" refers to specific structural types of 211 particles which can be divided into crystalline and non-crystalline classes. It should be recognized, however, that recent advances in X—ray crystallography and structural analysis have affirmed that completely ordered clay structures represent no more than ideal modelszi3. The swelling phyllosilicates, known as smectite clay minerals, have a layer lattice structure made up by the stacking of two—dimensional oxyanion sheets, often with intermediate layers of hydrated metal cations4. The major building units of these two-dimensional arrays are silica tetrahedra and octahedra of magnesia or alumina. The silicate tetrahedra are usually oriented so that the three basal oxygen atoms of each tetrahedron lie on 1 the same plane, while the fourth oxygen atom (apical) defines a second common plane. The octahedral sheet contains a cation, usually .Al (gibbsitic) or Mg (brucitic), surrounded by six hydroxy ions in an octahedral arrangement. Condensation of the tetrahedral and octahedral sheet in a way that some hydroxyl ions of the brucite or gibbsite structure are replaced by apical silicate oxygens results in a composite layer structure. Hectorite and montmorillonite, like all smectite clays, are classified as 2:] layer minerals. This designation refers to a structure in which two tetrahedral sheets sandwich an octahedral sheet (Figure 1). A common example of a 1:1 layer structure is kaolinite, where a tetrahedral sheet is paired only to an octahedral sheet to form a layer5. Montmorillonite contains mainly aluminum in the octahedral sites and since only two-thirds of these sites are occupied, the mineral is referred as a "dioctahedral" clay. On the other hand, a "trioctahedral" mineral has all octahedral sites filled by metal cations, as in hectorite, where all sites are occupied principally by magnesium ions. Cations at particular locations of the silicate structure can be replaced by various other cations with similar ionic radii without changing the structural characteristics of the mineral. If the replacing cation has a lower valence, a net negative charge will develop as a result of isomorphous substitution. Charge neutrality is achieved by either an opposing substitution with cations of higher valence or, as is usually the case, by the presence of additional cations in the interlayer region of the structure. The replacement of Mg2+ for A13+ in the octahedral sheet of montmorillonite results in the development of a net anionic charge. In hectorite the layer charge arises from the isomorphous substitution of octahedral Mg2+ by Li+ ions. Normally, arrays of hydrated alkaline earth Figure l Idealized structure of a smectite clay mineral. (0) Oxygen atoms; (0) hydroxyl groups. Silicon and sometimes aluminum normally occupy tetrahedral positions in the oxygen framework. Aluminum, magnesium, iron. or lithium may occupy octahedral sites. M""’-xH20 represents the interlayer exchange cation. or alkali metal cations are located in the interlamellar space or the regions between the basal surfaces of the negatively charged silicate layers. The charge balancing cations are usually located adjacent to the points of anionic charge on the basal planes. However, small anhydrous cations, mainly H+ or Li+, can migrate through the oxygen sheet to the neighborhood of the substitution, where the anionic charge arises. It has been reported that H+ ions coordinate with the hydroxyl groups of the octahedral sheet rather than react with tetrahedral oxygen atoms6. The idealized unit cell compositions for hectorite and montmorillonite are M0.67 1MES.33Li0.67]VI(Si8.0)IVO20(OH;F)4 and M0.67[A13.33M80.67]VI(Si8.00)IV020(OH)4 respectively, in which the superscripts (IV) and (VI) refer to the respective cation in tetrahedral and octahedral sites, and M represents a univalent or equivalent compensating cationz. The thickness of the unit structure (layer plus interlayer assembly), which can be calculated from a series of 009. X—ray basal reflections, will depend on the amount of interlayer water and the nature of interlayer cation. 2. Ion Exchange The hydrated compensating cations of . the native minerals are exchangeable and can be replaced with other hydrated metal cations and with various organic and organometallic cationic species. However, the extent of ion replacement may be size limited, especially in the case of large complex cations. Hectorite, for example, exhibits a cation exchange capacity (CEO) of about 70 milliequivalents per 100 grams. Based on the CEO and the a and b unit cell parameters of the mineral (5.25 x 9.18 A), the average distance between exchange sites is calculated to be about 8.7 A7. Thus, cations with cross-sectional diameters greater than this value can fill the interlamellar spaces before an homoionic clay is achieved. The exchange reaction can be represented by the following equation: solvent Na+(solv) + Mn+ ———) M"+(solv) + Na” (1) where the horizontal lines represent the negatively charged silicate layers. The kinetics and equilibria of the above equation depend on several variablesS-g. However, an attempt will be made to generalize some of the qualitative features concerning the tendency of a cation to exchange onto a negative surface. The exchange equilibrium in general favors a) cations with higher valence charge; b) the larger cation between species of a particular valence; and c) certain cations (i.e. K”, Ba2+ and NH4+) that because of their size can take up a very favorable interlayer position. The rate of the reaction is determined mainly by the interlayer cation accessibility, which in turn depends on the swelling properties of the mineral. 3. Swelling of Smectites Smectite clays exhibit basal spacings larger than the corresponding 9.2 A of their uncharged analogues because of the presence of hydrated compensating cations. The values of basal spacings depend mainly on the mineral, the exchangeable cation, and the partial pressure of water vapor in equilibrium with the clay sample“). 0 The basal spacing of dehydrated Na+-montmorillonite is 9.5 A. As the 7 water content increases Spacings of 12.4, 15.4, and 18.6 A are observed, corresponding to the presence of one, two, or three layers of water associated with the Na+ ion respectively. Further swelling can occur when Na+-montmorillonite is suspended in water. In fact, the aluminosilicate layers are completely dispersed, the interlayer spacing being practically infinite“. The osmotic swelling of smectite clays is limited by the electrostatic interactions between the anionic silicate layers and the compensating interlayer cations. As the dielectric constant of the swelling media decreases, the Coulombic interactions between the opposing silicate surfaces and the interlamellar cations increase, which results in smaller basal spacingslz. Increasing the concentration of clay suspensions in aqueous or ionizing media can form thixotropic porous structures, that may gel. This gelation phenomenon results from extensive layer edge-to-face and edge-to-edge interactions generating a "house of cards" structure“). A wide variety of neutral molecules other than water can also be intercalated on smectite clays. Protonation and ionic interactions, coordination to interlamellar cations, and dipolar interactions are among the mechanisms involved in the intercalation processes. The intercalation of ammonia on Mg2+-montmorillonite involves the protonation of the base and leads to bound ammonium ions”: Mg(ll20)x2+ + NH;;—-) [Mg(OH)(H20)x_1]+ + NH4+ (2) Transition metal ion exchanged minerals intercalate pyridine through coordination to the metal ions”: -H o [Cu(HzO)xl2+ + py —-2-—-) [Cu(py)42+l (3) The swelling and increase in basal spacings on treatment with polar solvents not only can be used as a means of identification of these minerals, but plays a very important role in catalytic reactions as well. 4. Acidity Among the very significant properties of silicate minerals is their ability to exhibit strong surface acidity upon removal of their adsorbed or intercalated water. Though both Lewis and Bronsted acid sites may contribute to the surface acidity, protonic acidity is more important in catalytic reactions involving the minerals. In some cases native minerals possess sufficient surface acidity to catalyze several "natural" processes, e.g., the chemical transformation of organic molecules, the degradation of pesticides, and petroleum-forming reactions. The surface acidity is related in origin and is often comparable to that of amorphous acidic silica-alumina and zeolites that have been used as petroleum cracking catalysts. Polarized water molecules coordinated to the compensating interlamellar cations are the most important source of Bronsted acidity in smectite clays. The degree of dissociation of water coordinated to interlayer cations is higher than that of the bulk liquid. Thus, hydrated cations are more acidic in the clay interlayers than in homogeneous aqueous solution15916. The acidity can be correlated with cationic polarizing power which increases with increasing charge-to-radius ratio. In addition, the acidity increases as the amount of interlamellar water decreases. 5. Catalysis and Interlayer Dynamics Smectites have been known and used for many years as efficient heterogeneous catalysts for a number of reactions”. Acid modified smectites were used extensively in the petrochemical industry for the cracking of petroleum feed stock to valuable products until they were replaced by the more thermally stable zeolites. The use of silicates as polymerization catalysts is also known and discussed extensively in many textbooksm'zo. Because of their "two-dimensional" interlamellar environment, smectite minerals offer the possibility of controlling chemical transformations with the prospect of obtaining products different from those obtained in a less restrictive homogeneous medium. A recent development in the area of clay catalysts is the intercalation of known, well studied, complex catalysts on smectite interlayers by utilizing simple ion exchange techniques. Electron spin resonance (ESR) and nuclear magnetic resonance (NMR) studies have provided an insight into the nature of clay intercalate321'23. At low degrees of interlayer solvation the interlamellar cations adopt a dynamic orientation and rotate anisotropically about certain molecular axes. At high degree of interlayer solvation, swelling beyond the coordination sphere of the solvated cation separates efficiently the anionic silicate layers from the solvated complex and provokes rapid tumbling. Thus, intercalated metal catalysts exhibiting solution-like mobility should be comparable in terms of catalytic activity to their homogeneous counterparts. In addition, surface chemical phenomena may favor the intercalation of a particular catalytic species from those involved in an equilibrium and may alter the specificity and selectivity of the catalyst. 10 6. Pillared Clays A serious drawback associated with the use of naturally occurring smectites as adsorbents and heterogeneous catalysts at elevated temperature is the loss of adsorbed solvent and collapse of the layers to van der Waals contact. Under these circumstances the surface area drastically diminishes and interlamellar catalytic reactions are precluded. Efforts to overcome this practical problem led to intercalation of robust cations, which act as molecular props or pillars, so that internal surface is available for adsorption and catalysis even in the absence of a swelling solvent. Among the various cationic species that have been used in pillaring layered silicates are alkylammonium ions, bicyclic amine cations, metal chelate complexes, and polynuclear hydroxy metal cations. . More than 25 years ago Barrer and MacLeod introduced the concept of pillaring smectite clays to prepare porous forms of montmorillonite by ion exchange with tetraalkylammonium ions“. Tetramethylammonium and ethylenediammonium, forms of montmorillonite, hectorite, and fluorohectorite possessing interlayer porosity were also prepared25. These systems exhibit a very .small interlayer spacing (Ad = 3-4 A) and show low resistance to swelling solvents. Cationic species derived from rigid or cage-like amines also have been used as pillaring agents. For example, the intercalation of protonated 1,4—diazabicyclo[2,2,2]—octane(triethyldiamine) in smectites results in a material exhibiting d00£ interlamellar spacing of 14.2 A26. Moreover, this material shows typical molecular sieving properties and markedly higher catalytic activity for esterification of carboxylic acids with alkanols than does R4N+-montmorillonite. However, these systems are not thermally stable and decompose below 250°C”. Derivatives with higher thermal stability (<450°C) were prepared by mild-«‘— ll introducing large metal chelates complexes such as Fe(Phen)32+, Cu(Phen)32+, Fe(bp)32+, ou(op)32+, and Ru(bp)2+28’29. An unusual characteristic of these products is the incorporation of anions in the interlayer region in the form of cation-anion pairs. This binding of excess salt is believed to result from a screening of the electrostatic charge of the silicate layers by the large complex cation. A turning point in the preparation of genuine pillared phases was achieved by intercalation of polynuclear hydroxy metal cations formed by hydrolysis over a specific range of OH‘/Mn+ ratios. These products exhibit high thermal stability, considerable surface areas, and intracrystal acidity and can function as cracking catalysts30’40. Most of the work in this area has been devoted to using hydroxy aluminum cations. 27A1 NMR studies along with potentiometric titration data indicate that the predominant species in base- or metal-hydrolyzed AlCl3 at the OH"/Al3+ ratios used to prepare the pillaring +39’41. In addition the lattice expansion ('b 9.5 7)) agent is A113O4(OH)233 of smectites pillared with hydroxy aluminum cations is consistent with the estimated van der Waals diameter for an intercalated A113 Keggin ion. The introduction of polynuclear oxycations in the interlamellar region of the mineral, followed by dehydroxylation at elevated temperatures to produce small oxide aggregates, provides a versatile method for preparing a new class of molecular sieves with pore size range (6-40 A) larger than faujasite-type zeolites. By varying the size of the molecular pillar and/or the population density of the interlayer region one could tailor the pore size to a particular application. B. Concept of Supported Metal Complexes Efforts to prepare superior hybrid catalysts possessing the advantages 12 of both homogeneous and heterogeneous catalytic systems with few disadvantages led to immobilization of soluble transition metal complexes on inert solid supports such as organic polymers and refractory metal oxides42’50. One can appreciate the recent interest .in supported complex catalysts by comparing them with traditional homogeneous and heterogeneous catalysts. A commonly employed class of homogeneous catalysts is that of the organotransition metal complexes. Since these complexes exhibit a well defined structure and stoichiometry, and advances in analytical techniques permit proper characterization in solution, homogeneous catalysts are better studied and understood. In addition, because of the availability of all metal centers in a sufficiently dilute solution, their catalytic activity can be easily interpreted. For the same reason they allow for an efficient and totally reproducible use of metal atoms. Furthermore, transition metal complexes exhibit electronic and steric properties that can be selectively modified or controlled through ligand substitution or variation of the solvent system. Finally, homogeneous catalysts usually possess one type of active site making them more specific and easier to tailor to a particular application. Though in principle homogeneous catalysis seems highly efficient and attractive, it suffers from three major technical problems that may be prohibitive in industrial applications. The main disadvantage of homogeneous catalysts is the problem of separating the catalyst form the products at the end of the reaction. Separation of the two usually can be accomplished by distillation below the decomposition point of the catalyst, a procedure that is generally ineffective and highly expensive. Second, organotransition metal complexes lack the stability exhibited by traditional heterogeneous catalysts, such as pure metals and inorganic oxides, under severe reaction 13 conditions, that might be necessary for a particular process. Finally, homogeneous catalysts suffer from a limited solubility in suitable solvents, a problem clearly not present with a heterogeneous catalyst. The major advantages of heterogeneous catalysts are their capacity for use in packed or fluidized bed reactors and ready separation from substrate and reaction products. Because of their high activity for a wide range of reactions they have been traditionally employed for a number of important industrial applications. However, certain undesirable properties make them less attractive. Heterogeneous catalytic reactions take place at the interfaces of the solid catalyst and the liquid or gaseous substrate. Only a fraction of the potential catalyst centers is exposed to substrate molecules; the atoms not present at the surface are inaccessible and remain unused. Despite recent advances in the area of surface characterization51’53, complicated processes, such as chemisorption and catalysis, are not well understood. Thus the design and modification of heterogeneous catalysts is often limited due to poorly defined activesites. Moreover, their specificity is generally lower compared to homogeneous catalysts because of the several types of active site present on the surface. The major advantages and disadvantages associated with traditional homogeneous and heterogeneous catalysts are listed in Table 1. The most common method in heterogenizing a transition metal catalyst is anchoring of the catalyst to a variety of solid supports through an ionic or covalent interaction. The metal complex may be also physically dispersed on the surface of the support by using an impregnation technique. However, the physisorption technique leads to species interacting weakly with the support, that can be readily desorbed from the surface, when the reaction involves solid-liquid phases. 14 £836.80 :23?! Eumoo new 982%. $2333qu was 3:35:88 8:55 motm 95% became :_ 3 26 2:286 EoEw>oEE_ Una :mfiom E3595 28> Lem .5 m: L E: g cm 9836.50 5308.. 2:958 3 357.com E3298 cobwewamm Am 3.. C 2“ Am C 3260.5 :25?! E9: cosmewaom znwwm mcofiomoe Co emcee outs e .8 53:8 55 32235 EoEwcooE ncw EEEE onEooo< 8:95:00 new not? >5on moseoaoa otoum 93 3:03.85 mEoww ~39: Co om: 35325.59. ccw 2223m— mcoEccoo cozowm: 3:: .695 3333 wotoaom 335% 23—38 @32935 ecu 863m :25 Am 3 S 3 3 Am 3 £9.3qu msoocomegoz v.63 wuwo mzoocmmoEoz wows—.5685 mowawcw>o< 808.5 meumbsao msoocoweouo: can msoocomoEo: no moms-.9635 can gags—26¢. a “936,—. 15 The new hybrid catalyst systems that appear heterogeneous at the bulk level but are essentially homogeneous on a molecular level, typically exhibit ease of separation, considerable thermal and mechanical stability, and efficiency in multistep or batch processes, properties commonly associated with heterogeneous catalysts. On the other hand properties such as specificity, efficiency, reaction control, and reproducibility are ordinarily associated with homogeneous systems. Additional advantages that may result from the immobilization process per se include 1) higher activity or specificity by inhibiting the number of undesirable side reactions; 2) substrate selectivity based on introduction of preferred orientations and different stereochemistry about the central metal atom; and 3) preferential stabilization of catalytically active but normally unstable structures4zi55. The components of the hybrid catalyst must be designed to fulfill the reaction requirements. From a chemical viewpoint the support should be inert to the reagents and products. The important engineering considerations are thermal and mechanical stability, enabling it to withstand the required reaction conditions, porosity, and acceptable surface area. Typical organic supports are polystyrene, polyamines, polyvinyls, polyamino acids, urethanes, acrylic polymers, and cross-linked dextrans. The list of usually used inorganic supports includes silica, alumina, glass, zeolites, and clay356i57. Though inorganic supports possess better thermal and mechanical stability than their organic counterparts, various synthetic routes to organic polymers offer a wide range of surface functionalization, pore size and surface area. The catalytically active portion of the metal complex should remain stable under reaction conditions and must be "soluble" in the reaction medium. The optimum catalytic activity of immobilized catalysts is achieved in a solution-like environment in which metal complexes maintain their vibrational 16 and rotational degrees of freedom. To enhance activity it may be preferable to anchor the metal complex to the support via a long chemical chain rather than directly attach it to the surface. Close proximity of the support with the metal complex may result in strong stereochemical interactions with the active site. Though immobilization through an electrostatic interaction appears to exert the fewest restrictions on rotational and vibrational motions of the catalyst, the major problem lies with the availability of catalytically active charged species that remain stable when electrostatically bound to the appropriate support. C. Some Aspects of Cluster Chemistry and Catalysis 1. Metal Cluster Carbonyls Metal clusters are discrete molecular species consisting of a framework of metal atoms or ions in triangular or polyhedral arrays. The metal skeleton is surrounded and stabilized by various ligands such as carbonyls, hydrides, phosphines, or halides. However a small subclass of "naked" clusters with no ligands attached to the metal core58 includes ions like Bi53+, Sn53', Ge94', and Sn94‘. Hundreds of cluster complexes are known, many so stable that their structures have been determined by a variety of techniques including X-ray crystallography, NMR, infrared, and Raman spectroscopy59“52. One metal cluster subclass of particular interest is metal carbonyls. CO molecules at terminal or bridging positions are coordinated to the transition metal framework. The strength of metal—metal bonds increases within a group as the size or atomic weight of the metal increases. This effect is clearly demonstrated by comparing the chemistry of iron and ruthenium triangle derivatives of the type M3(CO)12 (M = Fe and Ru). The iron triangle in Fe3(CO)12 is generally less stable and is broken more readily 17 than is the metal core in RU3(CO)12 upon reaction with reagents such as tertiary phosphines. RU3(CO)12 and 053(CO)12 consist of a triangle with four terminally bonded CO ligands attached to each metal atom53964. One unusual aspect of the trinuclear metal unit in ruthenium and osmium dodecacarbonyl clusters is their weak but detectable basicity65. Thus, they dissolve readily in concentrated sulfuric acid to give a solution exhibiting a high field 1H NMR resonance at around 6—19 to -20, indicative of a hydride bridging two transition metal atoms. Precipitation and isolation of the stable [HMg(CO)12][PF5] is easily accomplished by adding NH4PF6 to the sulfuric acid solutions. An interesting trinuclear metal cluster, H2033(CO)10, is obtained by passing hydrogen gas through a refluxing octane solution of OS3(CO)12. Its unusual chemistry arises from the presence of an unsaturated site, thus, it readily adds donor molecules to attain the stable 48-electron configuration“. Ir4(CO)12 and H4RU4(CO)12 possess a tetranuclear core structure in a tetrahedral arrangement with three terminally bonded CO molecules coordinated to each metal atom58’59. It is believed that the four hydrogen atoms in H4RU4(CO)12 bridge the four long edges of the RU4 framework69. Metal cluster carbonyl complexes undergo a wide range of substitution reactions. The main synthetic route in preparing carbonyl substituted derivatives until recently was thermally or photochemically facilitated CO dissociation as illustrated by equation 470. A Mm - (:0 T him-U + co (4) However, a serious drawback associated with this method makes it less 18 attractive. As may be anticipated, it leads to a mixture of products that normally require separation by chromatographic techniques. A recent advance employs Me3NO for the oxidative decarbonylation of a thermally inert metal cluster“. The reaction mechanism involves nucleophilic attack on the metal bound carbonyl to give carbon dioxide and trimethylamine (equation 5)70 Me NO Mm—CO -——3—-—) Mm-D + C02 + Me3N (5) The advantage of using Me3NO is that both reaction byproducts are gaseous and are easily removed. In addition, an unsaturated center is left free to coordinate any added ligand. By controlling the stoichiometry of the reaction one can influence its specificity and obtain a certain derivative rather than a mixture of products. Of particular interest to our work is the substitution reaction with phosphine ligands. Thermal substitution of carbon monoxide with a tertiary phosphine occurs progressively on different metal centers, the substituted derivative attaining a stereochemical arrangement that minimizes steric interactions between phosphine ligands”. 2. Catalysis by Metal Clusters Metal cluster compounds find several uses in catalysis by exhibiting interesting catalytic properties in their own right, by acting as catalyst precursors, or by serving as models for complicated processes such as chemisorption and catalysis73‘78. They can catalyze reactions as diverse as hydrogenation, isomerization, hydroformylation, water—gas shift, cyclisation, and oxidation”. By virtue of their structure and properties cluster compounds offer new opportunities in catalysis. Since they possess discrete well defined structures, high selectivity can be anticipated. Neighboring metal centers provide multiple bonding of a reactant molecule 19 that cannot be realized with mononuclear complexes. Moreover, different reagents can coordinate to neighboring positions, offering the possibility for synthesis gas (C0 + H2) conversion. Since transformations of the ligands can be transmitted through the metal-metal bonds, the reactivity and catalytic activity can be influenced, allowing for tailoring of the cluster to specific application. From the standpoint of cluster-surface analogy it is important to demonstrate that discrete metal clusters can catalyze a number of reactions which can be brought about by conventional heterogeneous catalysts, but not by their homogeneous counterparts. Reactions catalyzed by conventional dispersed metal catalysts include the hydrogenation of nitrogen to give ammonia, the hydrogenation of carbon monoxide to givemethane or methanol, and the hydrogenolysis of saturated hydrocarbons. Indeed OS3(CO)12 and Ir4(CO)12 were found active for the methanation reaction and the conversion of n—hexane to a mixture of C4 to Cg alkanes can be catalyzed by the "naked" cluster Bi53+74. However the task of establishing beyond reasonable doubt that the actual catalyst is indeed a cluster compound and not a product of dissociation or aggregation is extremely difficult. The fact that the metal cluster complex may have been present initially in solution and may have been recovered intact at the end does not consitute a strong indication that the cluster integrity was maintained throughout the reaction. As a matter of fact, the opposite is believed to be the case, that is, some reactions are thought to be catalyzed by highly active mononuclear species formed by the fragmentation of the cluster under the reaction conditions”. Laine has formulated a set of five criteria that can be useful in identifying a homogeneously-catalyzed reaction as a cluster-catalyzed 20 reaction79. Cluster catalysis is suggestted if, 1. the turnover frequency increases with increasing concentration of catalyst, 2. the product selectivities obtained using cluster catalyst precursors are different from those obtained using mononuclear catalyst precursors, or the products themselves cannot be reconciled with mechanisms that involve only mononuclear species, 3. it is possible to modify the catalyst or the reaction conditions to favor metal-metal bond formation and the modification results in increased catalytic activity, 4. catalytic asymmetric induction is observed or chiral products can be obtained using chiral metal clusters in which the asymmetry resides in the metal framework or is a basic skeletal property, 5. a specific combination of two or more different metals can be used to enhance the rates of the reaction or change the product selectivity or allow catalysis of a reaction not catalyzed independently by any one of the metals. ' D. Supported Metal Cluster Carbonyls 1. Immobilized Metals, Mononuclear Complexes, and Cluster Compounds on Solid Supports Supported metals, a subclass of traditional heterogeneous catalysts are capable of activating C-C and C-H bonds in saturated hydrocarbons as well as hydrogenating carbon monoxide“. Dispersion of the metal aggregates on an inorganic oxide increases the effective surface area per gram of catalyst and at the same time offers the potential of better temperature control. 21 On the other hand, homogeneous catalysts comprise soluble mononuclear metal complexes and exhibit high selectivity“. However, the only homogeneous processes presently of major industrial significance are olefin hydroformylation82, methanol carbonylation83, and butadiene hydrocyanation84. Although anchoring of metal complexes on solid supports is believed to bridge the gap between conventional homogeneous and heterogeneous catalysts, the new hybrid catalysts have not yet found industrial application, because of problems associated with catalyst stability45. Supported metals and transition metal complexes on solid supports constitute the limits of a hierarchy defined by the degree of metal aggregation85. Supported metal clusters, a new class of materials, possess an intermediate position. Since small metal particles are normally more efficient catalysts, the use of supported metal clusters offers a new dimension to the preparation of highly dispersed heterogeneous catalysts. A subclass of metal clusters that has received considerable attention in the preparation of supported metallic particles is metal carbonyls. This route allows for the preparation of highly dispersed metal particles on both organic86 and inorganic supports“. Metal clusters might also serve as potential precursors for the synthesis of supported metal oxides”. They can offer the following differences/advantages over conventionally prepared supported metal catalysts78: 1. Non-aqueous solvents can be used in catalyst preparation, since metal clusters are soluble in organic media. 2. The reduction by hydrogen at high temperatures can be avoided, for metal clusters comprise metal atoms formally in a zerovalent state. 3. A halide free-route to supported catalysts is usually advantageous, since halide atoms can reduce catalytic activity by poisoning the active ... - -1 22 sites and they might be responsible for sintering of the catalyst. 4. Heteronuclear clusters of certain stoichiometry are known and they may be used for the preparation of alloys of well defined stoichiometry and composition. The two major methods commonly employed for heterogenizing transition metal complexes are also used for immobilization of metal cluster compounds. The first preparative route involves impregnation of the support with a solution of the cluster carbonyl compound followed by either partial or complete removal of CO molecules under various reaction conditions. The main problem associated with this method is that considerable aggregation may occur during ligand removal, leading to a pocrly dispersed metal. An alternative approach, resulting in actual supported clusters, uses chemically modified supports such as phosphine functionalized polymers and silica. The anchoring cf the carbonyl complex is realized through a chemical bond between the cluster and the functionalized support. Methods such as ion exchange of the cluster compound or a precursor into the cavities of zeolites and metal vapour deposition have also been employed89’90. Infrared spectroscopy has been used extensively in characterizing physically supported and chemically bound metal cluster carbonyls. The supported metal clusters are usually identified by comparison of their infrared spectra in the carbonyl stretching region with those of their molecular analogs. Although it is a very powerful technique, providing structural information as well as the extent of ligand substitution91, it is the rule in surface analysis to use more than one method. Thus, several spectroscopic techniques, including Raman, EXAFS, NMR, ESR, and Mossbauer, have also been employed. 23 2. Anchored Metal Carbonyls on Functionalized Polymers The main route in supporting metal cluster carbonyls on organic polymers involves ligand exchange between the cluster compound and the functionalized support. Cluster carbonyl complexes are known to undergo facile ligand exchange with phosphines. Phosphine groups, the most common ligand used to attach metal complexes to polymers, can be introduced into the polymeric structure through different courses (Figure 2). The first involves reaction of the halogenated polymer with LiPPh292. An alternate approach, leading to the same product, involves treatment of the halogenated polymer with butyllithium, followed by reaction with ClPPh293. A similar functionalized polymer can be formed by chloromethylating the polymer first with CH3OCH2CI in the presence of SnCl4 and then treating it with LiPPh293. Instead of functionalizing the polymer by first halogenating it and then transforming it to the phosphine analog, one can copolymerize styrene and divinylbenzene with the appropriate monomer to produce block copolymers. This product can be used for immobilization of highly substituted cluster compounds, while random copolymers lead to monosubstituted cluster585. An alternate route uses copolymerization of cluster bound monomer with another monomer. However, several technical problems, such as reaction of the cluster with the polymerization initiator or differences in reactivity between the different monomers, make this route less attractive. In order to be useful as a support an organic polymer should possess properties such as good mechanical and thermal stability, limited solubility and inertness in the reaction medium, and readily accessible attachment sites. The intrinsic nature of the polymer, and to a lesser degree its structure and the extent of cross-linking, are the factors governing thermal and 24 Figure 2: Preparation methods for phosphine-functionalized poly(styrene-divinylbenzene). 25 Hr 1.1%.— mm H _(1 SN ‘4 1+- 1 11PM»; 1.1mm: ®_ PM): is pol\(si\ reno—tilx imlbcnzcnc) _. n-lillli (‘lrmtg jlx‘ [)U\ll). 1‘.(‘. Unit“ and J. 14010. (‘iit‘ti‘:l(‘l‘lt, 26 mechanical stability. Numerous literature reports describe synthetic methods that lead to polymer-supported metal cluster carbonyls. A summary of polymer-immobilized cluster carbonyl compounds of Ru, Os, and Ir pertinent to this work is given in Table 2. 3. Immobilized Metal Carbonyls on Inorganic Supports Inorganic supports exhibit good thermal and mechanical stability because of their rigid structure and appear to be superior alternatives to organic polymers. The latter typically have an upper thermal stability limit around 150°C that overwhelms other advantages such as ease of functionalization and spectroscopic characterization. On the other hand, the limit of inorganic-supported cluster compounds greatly depends on the thermal stability of the complex itself. In fact, it has been reported that the thermal stability of a complex is enhanced as a result of the anchoring process“. The main route to functionalize an inorganic oxide involves reaction between surface hydroxyl groups and a readily hydrolyzable moiety95‘97. ton -o \ -OH R(CH2)nSiX3 - O- Si-(CH2)nR (6) —> bOH —O n = 0,1,2... R: an organofunctional group, e.g., Pth, NHz, C5H5 X: a hydrolyzable group, e.g., CLIUQ,OR,OCOR 27 :otwcomoegc 3 @ immzmvztmamcmmvaAOOVv: em So: So: 50.3 its: .. 0:350353010533300 m_ @ 0:00:50 _ 020 NA lawn: V c g A C0 v F: 0:239:00 ~0A 05233 2: 2.223 A $333002: oooeezomoee -3209: . 0:233:00 2:5: 8.585an £33 A @4351: 208303 885383 38003.0? :03::0m0.6>.: coswstmEofl 2: 0:00:00; 2:00 0:132:38 NSO0Vm=m A0523: ,2: 8:358:23; -3578:on 28009.: E ®-emee}000§me: cofiwcomoegc ”A @ immzmvaOOVSEv: 0533300 3: 2228 A @ -mPeESOEéE oooeezomoeo 2:002:53: . am: :05003— 9.32:5 «.5095 3:00.50 20:55 0882: 2.2.: 2: 6:0 .00 .3— .«0 03:00.30 £3030 cannoEE—LoEfioa N 030,—. 28 This appears to be the most advantageous method because 1) it involves a simple one—step reaction easy to perform under mild conditions; 2) a number of RSiX3 compounds can be used; and 3) it affords a hydrolytically and thermally stable Si-C bond. As with organic polymers, phosphine groups are the most common ligand used to attach metal cluster carbonyls to functionalized inorganic oxides. The process of immobilization involves again ligand exchange between the cluster and the support. PhZPH + CH2 = CHSi(OEt)3 43—) PhZPC2H4Si(OEt)3 ~0H Lo \ ._ OH + Ph2PC2II4Si(OEt)3 W - o- SiC2H4PPh2 - OH .. O/ r-O “(cmx ~0>SiC H PPh M(CO) (~) ___.) 2 4 2 x-l 1 -co L- o A slightly different method, leading to the same surface structure, involves first the reaction of the metal cluster with the phosphine-silane ligand, followed by reaction with surface hydroxyl groups. M(CO)x + Ph2PC2H4Si(OEt)3 —_—C—5-) M(C0)x-1[thPC2H4Si(OEt)3] '- O [S]"0H \ / h-O 29 Anchoring a metal complex to an inorganic oxide through a pendant ligand normally leads to systems capable of catalyzing reactions under mild conditions. However, their use as catalysts under severe reaction conditions or catalyst precursors for the preparation of highly dispersed metal atoms is prohibited due to instability and decomposition of the organic ligand at elevated temperatures. Recent studies describe the preparation of supported metal clusters with well defined structures through direct interaction with surface -OH groups54’75. Several spectroscopic techniques have been used to characterize these supported clusters with unique structures and in a few cases they have been extracted from the surface by using ion-exchange method598’99. It is believed that this development, besides being a route to the preparation of highly dispersed supported metals, will provide an insight into the relations between structure and catalytic activity. A summary of supported cluster carbonyl compounds of Ru, Os, and Ir on inorganic oxides is given in Table 3. E. Research Objectives Smectite clays, having an ordered layer structure, provide certain advantages over amorphous metal oxides as complex supports7. The ability to control interlayer swelling by changing the polarity of the swelling solvent offers the possibility of inducing size or shape selectivity for catalytic reactions taking place in the interlayer region of the mineralloo. Moreover, the capability of the interlamellar space to play an active and often constructive role in catalytic reactions101 justifies the interest in investigating and using clay minerals as homogeneous catalyst supports. Metal carbonyl cluster complexes represent a potentially important class of compounds for intercalation in layer silicates, because they may 30 0: A "0-01500090: ~00 N£003.: . 5000003: ~05 N0800.502: m: 000:: N:00}; 00: 5360000000 wctsn «CE. 3: 5:05.50 528202 A”00:07: N080:5: n: 80050963 0:82 0:0 10009: ~05 N:00330 m: 00 0o 5:262 0: A0002-» N:00}; 0002 N00 S: 32000 >02 «£0033: c: 80 3000?: 3:80 w: N08009.: . no: :03003— 0030:.5m «.5005 1:00:00 002300 0388: 3:... a 030,—. 009.25% 8:030:— :0 :— 0:0 00 .3— uo 03:00:00 :3030 00050.53: .«0 mouxzm 0300—3 31 v: N: mmfi ~m~ emu o-.>cfi m: w: u: :05000E0000 050:0 :050E000 000000003: 00 Ho 5:802 N0 02:08:38.0 .0526 cothgou 2000000002; 550500.63 0:03:00 cotwc0wohnzz 0:0 _ >50 coSwEEEofl 5.00.0 :050C0w05z: 0:0;50 :050020E0fl3020 0:03 35-0EESA00vmm0m: 35033500030 35-09.32.2000mmomz «350252002:qu omSEiEEmSEmE A 0-02:00090: mO~_ Now Now «0% 00005500050 more MW:00:0m0 N:00:um0 250:00 28028 1:00:00? 1:00:00 chCUVmwON: 3300300; “0320033: NHAOOvmsm 002550 a 030,—. 32 Eli‘— mm— mm— mg :2 ms mm— 0? «NH :02500E0000 @526 8:00:00 :00..000:0>.: 5300000002; 0:0A>:~0 5350005000 9.200 :020E08 :00..000..0>.: :0_00N€0E05_ 0:00:05; 00:0020E0fl 0:00:01: 25-:~§V~A~EE~A000§ 05-035303: €005. 35-095 0500:»: A v_<-o::A00:u50: AVE-0::A000~50: AV_<-0::A000~A..0: Am5-0::A00:50: A 200230250: mCN~ ~05 00005505020 ~0~A< MCN_ ~05 05 ~90. ”ON—(7» ~05 m0m~ 2802.: 835238 -3~A00:_ ~:00050 5:00:50 A0:00:50? NIOOVmwO NHAODvmmO ~H80:50 ~AA00V~50 ~:00:Hmo 00:03:00 5 030,—. 33 mmw X; -A.-_ :2 cm— 0: m: 00 00 550.00.. a: 55>. A 000.0000»: 0:0xA0 AA AA 02:5 5_ im 0 5300055503: 0 A a .AVm. 00:20.00: E00: 200 E 00000550< 0 MH0~A NHAOOVF: NHAOOVF: NHAOOVV: ~AA0003. N~A00v0.: ~AA00::_ ~H80:5 UOH—CmHCOU M Omfiwh. 34 exhibit interesting catalytic properties in their own right, or they may serve as precursors for the formation of atomically dispersed metal centers, as discussed earlier (vide supra). However, the known methods of immobilizing metal carbonyls on organic polymers or inorganic oxides are not suitable for intercalation of metal complexes in clays. Typically, the immobilization of metal complexes in smectites requires positively charged species capable of electrostatic binding to the negatively charged silicate layers. The main objective of this study was to explore different approaches that would lead to intercalated metal cluster carbonyl complexes in clays. It appears that a net positive charge and solvolytic stability are the major requirements for a prospective candidate for intercalation in clay. On the other hand, size and/or shape of the intercalant play a relatively unimportant role, because even very large molecules such as enzymes can be accommodated in the silicate interlayer. This qualifies clays over zeolites, which are limited to intracrystalline immobilization of small molecules, because most of the pore openings are small (< 10 3). However, the task of finding cationic cluster carbonyls capable of electrostatic binding is not so easy. Though there is an abundance of neutral or anionic cluster compounds, the number of positively charged species is small. The first part of this work examines the reaction of the protonated metal carbonyl cluster H033(CO)12+ with Na+-hectorite. Remarkably, the binding of the cationic cluster leads to a preferential dispersal of 053(CO)12 centers at the hydroxylated edge sites of the silicate layers rather than at the more abundant basal surfaces. This observation prompted further study of the interaction of neutral 053(CO)12 cluster not only with natural Na+—hectorite but also with fluorohectorite and laponite, synthetic hectorites with different layer—charge and morphology. 35 It became evident at this point that preparation of genuine intercalated metal carbonyl clusters in clays required an alternative approach. Pinnavaia and co-workers demonstrated that the phosphonium-phosphine ligand PhZPCHZCHnghZCHzPh (abbreviated P-P‘”) could be used for the transformation of well known, otherwise neutral, homogeneous rhodium catalysts to their cationic analogs, making them suitable for intercalation in clay minera15102a103. This earlier work, along with the known property of cluster carbonyl complexes to undergo substitution reactions with phosphine ligands, provided the rationale for the use of P-P+ as a means of inducing positive charge on the cluster compounds. Thus, the second part of this dissertation describes the synthesis and intercalation of RU3(CO)9(P-P+)3, H4RU4(CO)3(P-P+)4, Ir4(CO)9(P-P+)3, 053(CO)11(P-P+), and H2033(CO)9,10(P-P+) in hectorite. Polyoxocation pillared clays constitute a new class of highly stable materials characterized by fixed porosity, high surface area, and intracrystal acidity. The pillaring reaction involves interaction of the negatively charged silicate layers with oligomeric cationic species and dehydroxylation at elevated temperatures to form small oxide aggregates which are stable in the interlamellar region. The recent advances in the area of direct immobilization of cluster carbonyls on inorganic oxides led us to investigate the interactions of carbonyl complexes with the aluminum oxide aggregates formed in the interlayer of smectites. Thus,the last part of this work examines the reaction of 053(CO)12, Ru3(CO)12, H4Ru4(CO)12, H2053(CO)10, lr4(CO)12, and [CpFe(CO)2]2 with alumina pillared clay. CHAPTER II EXPERIMENTAL A. Materials 1. Natural Hectorite Natural sodium hectorite with an idealized unit cell formula of Na0.67[Mg5.33Li0.67](Si8.00)020(OH,F)4 was obtained from the Baroid Division of National Lead Co. or the Source Clay Mineral Repository, University of Missouri, in the pre—centrifuged and spray dried form. The mineral was allowed to sediment 24 h as a 1 wt% aqueous slurry to remove dense impurities, saturated with Na+ by reaction with excess aqueous NaCl, washed free of chloride by dialysis, and then freeze-dried. 2. Sodium Montmorillonite (Wyoming) Na+-montmorillonite with an idealized unit cell composition of Na0.67[A13.33Mg0.67](Si8.00)020(0H)4 was obtained from the Source. Clay Mineral Repository and was pretreated similarly to natural hectorite (mg m). The cation exchange capacity for hectorite and montmorillonite is 70 and 83 meg/100 g respectively. 3. Laponite‘D This synthetic hectorite with a structural formula of 36 37 Nao.22Li0_14[Mg5.64Li0,36](Si8.00)020(OH)4 was obtained in powder form from Laporte Company and used without further purification. The particle size of this mineral is approximately 10 A thick with an average diameter of 200 A and its cation exchange capacity is 55 meq/lOO g. 4. Fluorohectorite This clay is also a synthetic hectorite in which the octahedral lattice hydroxyl groups have been replaced with fluoride ions. It has a unit cell formula Li].6[Mg4.4Li1.5](Si8.00)020(F)4. The mineral with particle size >> 2n and cation exchange capacity of 190 meq/IOO g was donated as an aqueous suspension by Corning Glass Works. 5. Solvents All solvents used in syntheses and isomerization studies were reagent grade; spectrograde solvents were used for IR, UV—visible, and NMR spectroscopy. The solvents were deoxygenated by standard freeze-vacuum-thaw cycles. 6. Reagents OS3(CO)12, RU3(CO)12, H4RU4(CO)12, Ir4(CO)12 and [CpFe(CO)2]2 were purchased from Strem Chemicals Incorporated and used as received. Trimethylamine oxide dihydrate (Me3NO-2H20) and ammonium hexafluorophosphate were obtained from Aldrich Chemical Company and used without further purification. l-hexene was purchased from Aldrich Chemical Company and was freshly distilled over activated alumina under an argon atmosphere prior to use as a substrate. Triphenylphosphine and benzylbromide were also obtained from Aldrich Chemical Company, while bis(l,2-diphenylphosphino)ethane was obtained from Pressure Chemical Company. Potassium hexafluorophosphate and sodium tetrafluoroborate were purchased from Alfa-Products-Ventron. Dowex (AGZ—XB) anion exchange 38 resin, Cl-form, 50-100 mesh, was a gift from Dow Chemical Company. B. Syntheses All reactions were carried out under an inert atmosphere either in a nitrogen-filled glove box or on a vacuum line. Once the crystalline materials were synthesized, subsequent handling was conducted in the open atmosphere. 1. H2083(C0)10 The hydrido-triosmium cluster was prepared by passing hydrogen gas through a solution of 053(CO)12 (600 mg, 0.661 mmol) in refluxing octane (120 ml) for 2 h, according to the method of Kaesz _e_t 31:66. The solution was concentrated and chromatographed on silica gel with hexane as eluant. The deep purple band of H2053(CO)10 was followed by a small band attributed to unreacted 053(CO)12. IR v(CO)(CH2C12) 2121 vw, 2076 s, 2063 s, 2025 vs, 2010 s, 1989 m cm-l. 1H NMR (CDC13) -11.49. *max 560, 339 nm. 2. H2083(CO)10(PPh3) This compound was obtained by reaction of H2053(CO)10 and triphenylphosphine by the procedure of Deeming and Hasso67. IR v(CO)(CH2C12) 2100 m, 2060 s, 2042 s, 2018 vs, 1998 sh, 1976 m, 1962 m cm‘l. 3. H2083(CO)9(PPh3) Pyrolysis of H2083(C0)10(PPh3) in refluxing hexane for 1 h and layer chromatography afforded pure H2053(CO)9(PPh3)67. IR v(CO)(CH2C12) 2090 m, 2052 5,2005 vs, 1990 m, 1972 m, 1954 m cm-l. 4. [HOS3(CO)12](PF6) This complex was prepared by adding concentrated (98%) sulfuric acid to osmium carbonyl, according to a method by Knight and Mays“. The solution was cooled to —30°C and added dropwise to a cold concentrated 39 aqueous solution of ammonium hexafluorophosphate. The precipitate was collected on a frit and washed quickly with the minimum amount of water. IR v(CO)(MeN02) 2134 s, 2102 s, 2077 vs, 2058 s, 2018 m cm'l. Amax (CH3COCH3) 366 nm. 5. [HRu3(CO)12](PF5) The preparatory procedure for this compound was similar to that of the osmium analog, described previously“. IR v(CO)(MeN02) 2128 s, 2100 s, 2077 vs, 2062 s, 2025 m cm'l. Amax (CH3COCH3) 403 nm. 6. PHzPCHzCHzPPh2CH2PhBl' The bromide salt of the phosphonium-phosphine ligand was prepared by the reaction of bis(1,2—diphenylphosphino)ethane and benzyl bromide in toluene, according to known procedure51029135. The white crystals melt at 245-247°C. 7. BF4’-Resin The tetrafluoroborate form of (AG2-X8) Dowex anion exchange resin was prepared according to a method described by Quayle and Pinnavaialoz. 8. (thPCHzlCH22PPh2CH23PhXBF4) The tetrafluoroborate form of P-P+ was prepared by a method described previouslyloz. In a nitrogen-filled glove box BF4'-resin (31 g, 100 meq) was added to the methanol solution obtained from the preparation of P-P+Br (Section 11.8.6). The slurry was stirred for 4 h and filtered. Fresh BF4'-resin (31 g) was added to the filtrate and the procedure was repeated two more times. The solvent was removed under vacuum and the resulting white needle-like crystals were collected on a frit and washed with toluene. M.P. 188-189°C. 1H NMR (00013) 6H1 2.55 m. 6H2 1.97 m, 5H3 4.26 d, JH3p+ = 14.6 Hz. 4O 9. [Ru3(co)9(P-P+)3I(BF4)3 Ru3(co)12 (200 mg, 0.313 mmol) and (P-P+)BF4 (542 mg, 0.939 mmol) was refluxed in 100 ml MeOH for 6 h, in a manner analogous to a method by Piacenti _e_t 511.136 for the preparation of Ru3(CO)9(PPh3)3. The solvent was evaporated under vacuum and the residue was washed with hexane to remove unreacted Ru3(CO)12. The residue was then treated with a minimum amount of acetone to dissolve the [RU3(CO)9(P-P+)3](BF4)3 reaction product and the unreacted P-P+ ligand was removed by filtration. Evaporation of the acetone solution afforded the pure compound. IR v (CO)(CH2012) 2048 m, 1980 s, 1962 s cm-l. Amax (0112012) 439, 362 sh nm. 10. [053(00)11(P-P+)I(BF4) This compound was prepared by reaction of stoichiometric amounts of 053(CO)11(CH3CN)137 (0.15 g, 0.163 mmol) and (P-P+)BF4 (94 mg, 0.163 mmol) in 30 ml CHC13 at 25°. The resulting precipitate was filtered, washed with a minimum amount of chloroform, and vacuum dried. IR v (CO)(CH2C12) 2100 m, 2048 s, 2028 s, 2010 vs, 1998 sh, 1983 m, 1976 m cm’l. Amax (CH2C12) 407, 323 nm. 11. [1r4(co)9(P-P*)3I(BF4)3 In a method analogous to that of Stuntz71 for the preparation of Ir4(CO)12_xLx type complexes, solid Ir4(CO)12 (300 mg, 0.272 mmol) was added to refluxing THF (500 ml). To this solution (P-P+)BF4 (471 mg, 0.816 mmol) and solid Me3NO-2H20 (91 mg, 0.816 mmol) was added and heated at reflux for 10 min. After cooling the solvent was removed under vacuum and the product was taken up in a minimum amount of acetone. The unreacted Ir4(CO)12 and excess ligand remained undissolved and was removed by filtration. IR v(CO)(CH2012) 2039 m, 1995 sh, 1984 s, 1965 sh, 1832 vw, 1779 s cm-l. Amax (0112012) 383 sh, 301 nm. 41 12. [H4Ru4(co)3(P-P*)4](Br4)4 To a solution of H4Ru4(CO)12 (200 mg, 0.269 mmol) in 100 ml CHC13 solid Me3NO-2H20 (119 mg, 0.108 mmol) and (P-P+)BF4 (620 mg, 0.108 mol) was added. After stirring overnight the solution became cloudy. Reducing the volume by half under vacuum and cooling gave [H4RU4(CO)3(P-P+)4](BF4)4 as a precipitate. Filtering the solution and washing the residue With CHCI3 afforded a pure compound. IR v(CO)(CH2C12) 2005 s, 1971 m, 1948 s cm‘l. Amax (0112012) 404, 309 sh nm. 1H NMR (d6DMSO) 6(u-H) -16.50 quintet, Jp-H = 7.0 Hz. 13. [H2083(CO)10(Ph2PCH21CH22PPh2CH23thBF4) To a solution of H20$3(C0)10 (100 mg, 0.117 mol) in CHC13 (25 ml) a stoichiometric amount of (P-P+)BF4 (68 mg, 0.117 mmol) was added according to a method reported by Deeming and Hasso67 for the preparation of H2053(CO)10(PPh3). The precipitate which formed after stirring for 1/2 h was collected on a frit, washed with CHC13, and vacuum dried. IR v(CO)(CH2C12) 2100 m, 2060 s, 2042 s, 2018 vs, 1990 sh, 1976 m, 1960 m cm-l. *max (0112012) 385, 331 nm. 1H NMR (CDC13) (27°C) 6111 2.94 m, 3H2 2.41 m, 6H3 4.42 d, JH3p+ = 13.59 Hz; (-61°C) 6(H-Os) - 10.36 s, 6(u-H) -20.48 m. Anal. Calcd. for C43H33BF4PZOS3: C, 36.06; H, 2.31; P, 4.32; Found: C, 35.69; H, 2.36; P, 3.29. 14. [HZOS3(co)9(Ph2PCH21CH22PPh2cnz3rhmBF4) A solution of [H2083(CO)10(P-P+)](BF4) (100 mg, 0.07 mmol) was refluxed in methanol (35 ml) under Ar for 2 h during which time the color changed from yellow to deep green. Removal of the solvent under vacuum and recrystallization from CHZClz-hexane afforded [HZOS3(CO)9(P-P+)](BF4) in the form of shiny flakes. IR V(CO)(CH2C12) 2090 m, 2050 s, 2001 vs, 1989 s, 1948 m cm-l. 1H NMR (CDC13) 3H1 2.96 m, 8H2 2.43 m, 6H3 42 4.48 d, JH3p+ = 14.26 Hz, 6(u-H) —11.06 d, JHp = 7.25 Hz. Amax (0112012) 359, 304 nm. C. Clay-Cluster Reactions 1. Reaction of Clay with HM3(CO)12+ or M3(C0)12 (M=Os,Ru) Na+-hectorite (0.10 g, 0.070 meq) in 10 ml water was added with stirring to 0.007 mmol of Hl\13(CO)12+ or M3(CO)12 dissolved in the minimum amount of acetone. After a reaction time of 10 min, the slurry was poured onto a polyethylene sheet and allowed to dry in air. The dried, self-supporting clay film was then removed from the polyethylene sheet and used for IR measurements. For some experiments the clay sample was freeze—dried instead of being dried in air. 2. Intercalation of Phosphine Substituted Cluster Carbonyls on Nathectorite In a typical experiment 0.1 g (0.070 meq) Na+-hectorite in 10 ml water was added to a vigorously stirred acetone solution of the appropriate complex (0.070 meq). After a reaction time of 10 min, the mineral was collected by centrifugation, washed several times with acetone to remove physically adsorbed cluster molecules, and air-dried. 3. H2083(C0)9(P-P+)-hectorite Decarbonylation of H2053(C0)10-hectorite was achieved by heating the exchanged mineral at 80°C under vacuum for 2 h. IR 0(CO) (KBr) 2092 m, 2052 s, 2000 vs, 1978 s, 1950 sh cm']. Amax (mull) 362, 315 nm. 4. HOS3(CO)9(CH=CH2XP-P+)-hectorite H2083(CO)9(P-P+)-hectorite (50 mg) was suspended in 50 ml acetone and acetylene gas, passed first through a trap at -78°C, was bubbled gently into the suspension at ambient temperature and pressure for 2 h. During 43 this time the color of the mineral changed from green to bright yellow. The solvent was drawn off and the clay was washed with acetone, dried, and characterized by infrared spectroscopy. IR v(CO)(KBr) 2090 m, 2042 s, 1999 vs, 1962 sh cm‘l. 5. Alumina Pillared Clay (ALPILC) The clay was prepared by reaction of natural Na+—montmorillonite (Crook County, Wyoming) with an aluminum chlorohydrate solution, chlorohydrol® (Reheis Chemical Company), according to a previously described method”. The product was air-dried and then calcined at 350°C under Ar for 2 h. The unit cell composition, determined by plasma emission spectroscopy, is [A1(0H)2.8012.87[A13.11F9042“80.481(Si7.92A10.08)020(0H)4 and the N2 BET surface area is approximately 300 mZ/g. 6. Reaction of ALPILC with M3(CO)12 (M=Os, Ru) ALPILC (0.5 g) was degassed at 25°C under vacuum for 4 h and a solution of the metal cluster (0.020 mmol) dissolved in CH2C12 (35 ml) was added. The resulting slurry was stirred under Ar for 20 h at ambient temperature and pressure and then transferred to a nitrogen—filled glove box. The clay was collected on a frit, washed with CHZCIZ to remove unreacted, weakly adsorbed cluster molecules and vacuum-dried. 7. Extraction of HM 3(CO)1 2+ Clusters From the Clay The neutral cluster carbonyl was first reacted with the mineral and the portion which was not chemisorbed was removed by washing with CH2C12. The mineral was then treated with a solution of KPF6 in acetone. After three successive washings the filtrates were combined and evaporated to dryness under vacuum. 44 8. Wet Impregnation of Clays with Neutral Clusters A solution in CH2C12 containing an amount (0.007 mmol) of neutral cluster complex (053(CO)12, Ru3(CO)12, etc.) was added to the clay sample (0.1 g) and the mixture was stirred under Ar for a few hours. The solvent was then removed under vacuum and the mineral was stored under an inert atmosphere before furthertreatment. 9. Powder Pyrolysis Experiments A powder clay sample or a self-supporting film was placed in a quartz tube fitted with two joints and stopcocks. The sample was pyrolyzed by using an external heating coil either while being pumped dynamically at approxiamtely 0.1 torr or while a stream of Ar gas was blowing over the sample. 10. Catalytic Reactions H2053(C0)9(P—P+)-hectorite (30 mg, 0.010 mmol) was suspended in 5 m1 of the appropriate solvent in a 15 x 150 mm pyrex culture tube fitted with a Teflon—linen screw cap and magnetic stir bar. To this suspension freshly distilled l-hexene (1-3.5 ml, 0.008-0.028 mol) was added. The mixture was stirred either at ambient temperature and pressure or the tubes were immersed in an oil bath whose temperature was controlled (Parr Instrument Company, Model 4831 controller attached to an Iron-Constantan thermocouple). Catalytic activity was determined by withdrawing periodically a fraction of the reaction solution and analyzing the products by gas chromatography. D. Physical Measurements 1. Infrared Spectroscopy Infrared spectra were recorded on a Perkin-Elmer 283B or 475 model 45 grating spectrophotometer. Solution spectra were obtained by using 0.1 mm NaCl cells. Spectra of various clay intercalates were recorded by either mixing the samples with KBr and pressing them into a disk or using self-supporting films. These films were prepared by allowing an aqueous suspension of the mineral to evaporate on a polyethylene sheet. In some cases the spectrum was recorded by using mineral oil mulls placed between Csl disks. Samples of air or water sensitive intercalates were prepared in a nitrogen—filled glove box just prior to scanning to minimize decomposition. 2. X—ray Diffraction Studies X-ray 002 basalspacings were determined for oriented film samples either with a Siemens Crystalloflex-4 or a Philips X-ray diffractometer by using Ni-filtered Cu-Ka radiation (4 (Kg) = 1.5405 .4). The oriented film samples were prepared by evaporating an aqueous suspension of the mineral on a microscope slide. The diffraction was normally monitored over a range of 2° < 29 < 15°. Peak positions, measured in degrees 29, were converted to interplanar d-spacings by using a standard chart. 3. UV-Visible Spectroscopy Electronic spectra were obtained with a Varian Associates Cary 17 spectrophotometer. Solution spectra were recorded with matched 1 cm path-length quartz cells. Spectra of clay intercalates were acquired as mineral oil mulls placed between silica disks. To avoid scattering, a sample of the native mineral was put in the reference beam. 4. Proton NMR Spectroscopy Proton nuclear magnetic resonance spectra were recorded on a Bruker WM—250 MHz spectrometer. Chemical shifts were measured either relative to tetramethylsilane as an internal standard or relative to the standard shift of the deuterated solvent and are reported in 6 units. 46 5. Gas Chromatography Gas phase chromatography of catalytic reaction products was performed either with a Varian Associates Model 920 CC equipped with a thermal conductivity detector or with a Hewlett-Packard model 5880 A GC supplied with a flame ionization detector. The columns were a 10 ft x 1/8 in. 10% UCW on 80-100 mesh chromosorb W and a capillary 12.5 m x 0.2 mm crosslinked dimethylsilicone. The reaction products were characterized by comparison of their retention times with chemically pure authentic samples. 6. Melting Points Melting points were determined on a Thomas-Hoover Model 6406-H capillary melting point apparatus. 7. Elemental Analyses Chemical analyses were performed by Schwarzkopf Microanalytical Laboratories, Wocdside, NY. CHAPTER III RESULTS AND DISCUSSION A. Surface-Selective Dispersion of Cluster Carbonyls on Layered Silicates 1. Reaction of HOS3(CO)12+ and 0s3(00)12 with Clays The reaction of Na+-hectorite (1.0 meq) in aqueous suspension with HOS3(CO)12+ (0.10 mmol) in acetone results in a product (I) which contains a bound osmium carbonyl complex. The IR spectrum of I in the terminal CO stretching region is shown in Figure 3, along with the spectra of authentic samples of [H083(CO)12][PF6](II) and 033(CO)12 (III). The CO stretching frequencies are provided in Table 4 for each compound; the band positions of electronic spectra are presented in Table 5. It is apparent from the IR and electronic spectra that the osmium complex in I is immobilized on the clay surface as the neutral cluster 083(CO)12 and not as the initial cation H083(CO)12+. In addition, the x-ray diffraction pattern of I reveals an 001 spacing of 12.5 A, a value which corresponds to an interlayer thickness of approximately 2.9 A and the presence of a monolayer of water. Therefore, the bound OS3(CO)12 is not intercalated. but instead is bound to the external surfaces of the clay. These conclusions also are compatible with the observation that the neutral carbonyl complex is readily desorbed by washing with acetone. 47 «Figure 3: 48 Infrared spectra in the terminal CO stretching region: (I) OS3(CO)12-hectorite prepared from HOS3(CO)12+ (film); (II) [HOS3(CO)12]PF6 (KBr pellet); (III) 033(CO)12 (KBr pellet); (IV) 053(CO)12-hectorite prepared by impregnation with 053(CO)12 (film); (V) sample I after heating in air at 150°C for 12 h. mum mA\\.. “m 50 0:..000001J0ACOVmCNAOmmm Z 652552.555 E... > 5555. 555253.353: .6208: 5068.. .580583555 E... 80869. + ~AA00:50 >A mswmA .mmmaA.m>on~.m-o~ .538588558 Ex ~AA00:..50 E 5555..maco~.mA~c~ 55.02.522.552 .EmA.~.E$A~.EEA~ .5: AfaAfiAoovmmofi A. 55A: .EEEJSSJSE .2695. .8038 5505 538.85: EA: 8083; + +A~.A00:50.: A T0.0 6300000000 0.0.00.2 005008009 0000.5 000000050 00000000: 0:: 30:09.0 002000: .5.— 5200000000 00200005 CO 0009...:— 51 532.582.?33 .mmmZéEOfiafiS ES? :52: .m>:§.$2~ .Ezomaemcfimmmom -=E E2“ 8_._8oo;-+az + 28028 “260.5 :23me 323820.82: + LNZOOmeOE :> _> 8.53.80 v 039—. 52 Table 5 Electronic Spectroscopy Data for Reaction Products and Reference Compounds System Solvent A max (nm) [HOS3(CO)12][PF6] Acetone 366 083(CO)12 CH2C12 385,329 [HOS3(CO)12]+ + hectorite film 383,324 reaction product 053(CO)12 + hectorite film 387,329 reaction product 53 To verify the presence of neutral metal carbonyl cluster in I, we investigated the binding of an authentic sample of 083(CO)12 to Na+-hectorite. The neutral carbonyl (0.10 mmol) in acetone solution was mixed at room temperature with a 1 wt% aqueous slurry of hectorite (1.0 meq), and the product was air-dried to form an oriented, self-supporting film. This method of sample preparation is analogous to the impregnation techniques typically used for supporting metal carbonyl clusters on silica and alumina. However, these latter supports, unlike clays, do not form highly oriented films when air dried. The CO stretching frequencies for the clay-supported 053(CO)12 prepared by impregnation of 053(CO)12 (IV) are given in Table 4 and Figure 3. It is seen that the spectrum is similar to, but not identical with, the spectrum observed for the clay product prepared by reaction with HOs;3(CO)12+ (1). Significantly, the bands for I are broader and less resolved than those observed for IV. Also, the relative intensities of the bands differ for the two products. These spectral differences may be due in part to differences in cluster dispersion. For instance, IV exhibits sharp x-ray reflections characteristic of a segregated 053(CO)12 phase. Phase segregation also is known to occur for metal carbonyls supported on silica and alumina by impregnation method5123. In marked contrast to the low degree of cluster dispersion in IV, product I shows no x-ray evidence for a segregated carbonyl phase. Thus we conclude that the supported carbonyl clusters derived from proton dissociation of HOS3(CO)12+ are highly dispersed with an aggregate size less than 50 A. We next consider the type of surface occupied by the dispersed 053(CO)12 clusters in I. There are two chemically distinct surfaces available for the binding of an adsorbate to a smectite clay particle: edge surfaces and basal 54 surfaces. As shown in Figure 4, the basal surfaces are composed of siloxane—type oxygen atoms, whereas the edge surfaces are hydroxylated. 083(CO)12 is known to react with surface hydroxyls .of silica and alumina124’127’128 to give molecular species such as those represented in equation 9 for a silica surface. 2( 231011) + 053(c0)12 -—“-9 [(2 SiO)20$(CO)n]x (9) n=2,3 This reaction provides a chemical means of distinguishing between OS3(CO)12 at edge and basal plane sites. Under conditions where surface migration is slow, as is the case with our oriented film samples (see below), edge-bound clusters should react according to equation 9, whereas clusters adsorbed on basal planes should be stable. In a typical smectite clay particle composed of a few tens of layers, each 9.6 X thick and approximately 104 X in diameter, the basal surface area is at least an order of magnitude greater than the edge surface area. In the case of IV, where much of the 053(CO)12 is segregated as a separated phase, we might expect the cluster to migrate onto the basal surfaces and eventually to edge sites for reaction with hydroxyl groups. For instance, 053(CO)12 supported on alumina by impregnation reacts according to eq. 9 at 150°C, even though the cluster initially is poorly dispersed at room temperature. However, the highly oriented nature of our clay film samples impedes migration of the metal clusters across the basal planes. Thus we observe no reaction of IV after 12 h at 150°C. The importance of sample orientation in impeding surface migration was demonstrated by impregnating a random, freeze-dried clay sample with 033(CO)12 and observing reaction Figure 4 55 Schematic illustration of the two types of external surfaces for a smectite clay. The edge surfaces are hydroxylated whereas the basal planes contain only siloxane oxygens. Sodium exhcange ions (not shown) occupy mainly the interlayer regions. 56 EXTERNAL BASAL SURFACE mug/imam wwom 57 according to eq. 9. In contrast to the behavior of the neutral clusters in IV, the clusters in I do react in air at 150°C to form new surface species, despite the oriented condition of the sample. As shown in Figure 3 and Table 4 the new surface species exhibit three terminal CO stretching frequencies. These three bands are indicative124’127’128 of both the molecular tricarbonyl (n=3) and the bicarbonyl (n=2) species depicted in eq. 9. On the basis of these observations, we may conclude that the initial binding of HOS3(CO)12+ and the subsequent formation of 053(CO)12 occurs at or near the edge surface sites. The selective formation of 053(CO)12 from HOS3(CO)12+ at edge surfaces may be explained in terms of the following scheme wherein M 2 OS. Scheme I _ — NO*(OQ) * HM3(CO).; —,"— HM + -No W NO (ca) 4' “(:0)" ———> 3 :2 — — NO‘iOQ) HM3(C0).; — edge- intercalated intermedidote _-- M (C0 H30. I )u —‘ . __> H,O‘ / M, (C 0).: adsorption at -— edge surfaces H,o* — -- M.(c0,u a... - I2.5A° 58 Unlike the neutral cluster the protonated cluster is capable of ion-exchange with intercalated Na”. It is this exchange reaction which provides the driving force for the initial binding of the cluster cation near the edge of the clay platelet. However, the cluster cation is unstable because the aquated interlayer is more basic than the corresponding neutral cluster. Consequently, the proton is transferred to interlayer water and the neutral cluster is selectively formed in a dispersed state at edge sites where it has ready access to hydroxyl groups. Although the initial intent of this work was to intercalate HOS3(CO)12+ cluster cations in smectite clays, the cations proved to be unstable with regard to proton dissociation at exchange sites near the platelet edges. Nevertheless, the observed chemistry is unique in that the metal clusters are selectively dispersed at or near edge surface sites and the degree of cluster dispersion is much greater than can be achieved by impregnation of neutral clusters from solution. These preliminary observations set up the basis for a more thorough investigation of the reaction of triosmium cluster carbonyl complexes with clays. It appeared to us we could control the dispersion of the metal cluster at different sites thus, allowing or preventing reaction with surface hydroxyl groups. We had studied briefly the influence of lamination or layer ordering on the reactivity of metal clusters towards OH groups. We next considered the role of size and morphology of the clay layers. Reaction of fluorohectorite (100 mg) in aqueous suspension with HOS3(CO)12+ (0.007 mmol) in acetone results in a product (VI) exhibiting an IR spectrum very similar to that of 033(CO)12. In addition to the IR spectrum, which on the basis of band widths suggests the osmium clusters are bound to the external surfaces of the clay, the x-ray diffraction pattern 59 reveals a basal spacing of 12.1 A, indicative of unexchanged clay with a monolayer of water. The difference between sample (I) and (VI) is apparent when the latter is heated at 150°C in air. Reaction under these conditions for 12 h causes only a fraction of the neutral clusters to react with surface hydroxyls, as evidenced by the appearance of CO absorptions at 2123 and 1958 cm'l. The IR spectrum, apart from the weak absorptions characteristic of the mononuclear osmium complexes, remains basically unchanged even after extended reaction times. The difference is reactivity between VI and I, the reaction product of HOS3(CO)12+ with Na+-hectorite, arises mainly from differences in morphology of the clay particles. Fluorohectorite with particle size >> 2p, much larger than natural hectorite, is expected to possess significantly lower edge surface area. In addition, the number of OH groups in fluorohectorite is statistically lower because of the substitution at octahedral sites by fluoride ions. Thus, though the reaction is believed to proceed according to Scheme 1, the available surface is saturated quickly and the rest of the cluster molecules are forced to be deposited on the basal planes. Metal clusters deposited at or near the edges do react to form mononuclear osmium complexes. On the other hand, the highly oriented nature of the fluorohectorite film samples impedes migration of the metal clusters across the basal planes to reach and react with OH groups. That the metal clusters do indeed selectively occupy first the edge surfaces before depositing on the basal planes is supported by comparison of the reaction of H053(CO)12+ with Na+-hectorite at two different molar ratios of metal complex to clay. An air—dried sample containing 0.035 mmol of complex per 100 mg of clay exhibits CO absorptions similar to those of unsupported 053(CO)12. Heating at 150°C in air for 12 h results in partial 60 reaction of the osmium clusters with surface hydroxyl groups in contrast to a clay sample containing 0.007 mmol of HOS3(CO)12+ (cf. 1, Figure 3, Table 4). As was noted earlier, the method of drying the mineral considerably influences the reactivity of metal clusters. Air. drying leads to highly oriented samples that impede the migration of cluster molecules across the basal planes. If 053(CO)12 is impregnated on Na+-hectorite and the product freeze-dried as a powder rather than air-dried as a film, then heating in air leads to reaction with edge hydroxyl groups and to the formation of ensembles of atomically dispersed Os complexes evidenced, by the familiar three peak IR pattern (Figure 5). In contrast, heating the air-dried film sample (cf. IV, Figure 3, Table 4) dees not lead to reaction with edge OH groups. The observed dependence of the osmium cluster reactivity on drying most probably is related to the mechanism of layer aggregation in smectite clays. Furthermore, the degree of lamination or layer ordering in a flocculated clay system greatly depends on the size and morphology of the silicate platelets. Typical smectite clay particles, composed of a few tens of layers with a diameter in the range 102-104 A, exhibit a pancake-like morphology. It is expected that layers with a large aspect ratio would favor face to face interactions of coagulated single-layer platelets. On the other hand, layers with a small aspect ratio would tend to aggregate through edge to edge and edge to face interactions resulting in a highly delaminated system. A typical flocculated clay system should consist of laminated aggregates along with delaminated areas. The proposed model for a flocculated clay is illustrated in the central inset of Figure 6. Face to face aggregates are shown by a group of parallel layers, while single layers represent delaminated Figure 5 6] Infrared spectra in the terminal CO stretching region: (a) Freeze-dried sample of 053(CO)12 on Na+-hectorite prepared by impregnation with 033(CO)12 (mull); (b) after heating in air at 150°C for 12 h (mull). 62 2200 2000 1900 cm" Figure 6 63 (A) Model for a wet, flocculated clay system after impregnation with 053(CO)12. Cycles represent aggregates of 053(CO)12 cluster molecules of unspecified size; (B) Freeze-drying tends to preserve the flocculated structure. The sample contains both laminated and delaminated platelets; (C) Air-drying tends to promote face-to-face lamination of layers. Clusters are "trapped" and migration to edge sites is impeded. \ // ;__Q_.Q_ I}. .L__ C / A 1 I O O 65 areas. Circles denote aggregates of 033(CO)12 cluster of unspecified size. Freeze drying will generally preserve the structure of the flocculated clay, while the surface tension forces present under air-drying conditions will assist in a general reorganization. This rearrangement tends to optimize face to face interactions, ultimately leading to a highly oriented structure. Under these circumstances, migration of the segregated 053(CO)12 onto the basal surfaces and eventually to edge sites for reaction with hydroxyl groups becomes highly probable for a random freeze—dried clay sample. Alternatively, the oriented nature of the air-dried samples hinders migration of the metal clusters across the basal planes. It is important to recognize that this model of flocculation is valid only for aqueous clay suspensions. If, for example, an organic solvent is used for the impregnation of the metal cluster, then air-drying will not involve a reorganization mechanism that leads to highly laminated aggregates and the reaction of the metal cluster with surface hydroxyls is not impeded. Thus, when Na+-hectorite (1.0 meq) is impregnated with a solution of 053(CO)12 (0.1 mmol) in CH2012 and then air-dried, the product exhibits a spectrum very similar tothat of the unsupported cluster. Heating at 150°C in air for 12 h leads again to reaction with edge hydroxyls and the formation of mononuclear complexes, as evidenced by CO absorptions at 2019 m, 2032 s and 1950 m cm“1 (Figure 7). To further verify the validity of our model for clay flocculation, we investigated the binding of authentic 053(CO)12 to both laponite and fluorohectorite. Though both are synthetic hectorites, the former has a relatively small layer aspect ratio, while the latter is characterized by large particles. After impregnation with the osmium cluster molecules both clay samples were air-dried. However, the x-ray diffraction pattern of 66 Figure 7: Infrared spectra in the terminal CO stretching region: (a) 033(CO)12-hectorite prepared by impregnation with 053(CO)12 in CH2C12 (mull); (b) after heating in air at 150°C for 12 h (mull). 67 L l __L 2200 2000 1900 CM" 68 fluorohectorite reveals a well ordered system, while laponite appears highly delaminated. As a result, the fluorohectorite clay sample remains unchanged on heating in air at 150°C (Figure 8). On the other hand, reaction of the metal clusters with OH groups in laponite under the same reaction conditions is highly favorable (Figure 9). The wet impregnation technique that has been used for immobilization of metal carbonyls on refractory inorganic oxides usually requires a large amount of solvent due to the low solubility of the carbonyl complexes. After the reaction, evaporation of the solvent leads to segregation of the metal carbonyl, especially in systems with weak complex-support interactions. Howe gt 31.138 found that materials prepared by impregnation are similar to those prepared by dry grinding techniques. Thus, impregnation methods generally lead to phase segregation of the carbonyl and such segregation can be observed by x—ray diffraction. The x-ray diffraction data obtained utilizing H053(CO)12+ and 053(CO)12 with different particle size clays are summarized in Table 6. For samples prepared starting with the protonated HOS3(CO)12+ cluster only the support was detected. The lack of an x-ray diffraction pattern due to supported carbonyl indicates these clusters are present in a highly dispersed state with an aggregate size less than 50 A, the detection limit for the x-ray diffraction technique. On the other hand, products obtained by the impregnation technique starting with the neutral 053(CO)12 cluster contain the metal species in the form of crystallites, regardless of the kind of clay or solvent used. The different degrees of dispersion may be explained by the difference in the nature of the two metal clusters. The protonated cluster, because of its cationic character, is capable of ion-exchange with interlayer Na+ Figure 8: Figure 9: 69 Infrared spectra in the terminal CO stretching region: (a) OS3(CO)12-fluorohectorite prepared by impregnation with 053(CO)12 and air-drying (film). The Os loading is 3.8 wt%; (b) after heating in air at 150°C for 12 h (film). Infrared spectra in the CO stretching region: (a) 053(CO)12- -laponite prepared by impregnation with 033(CO)12 and air-drying (mull). The Os loading is 3.8 wt%; (b) after heating in air at 150°C for 2 h (mull). hi 2200 2000 1900 71 Table 6 X-Ray Diffraction Data for OS3(CO)12 and Clay-Supported Osmium Carbonyl Samplesa d Spacings (A) for Sample the Carbonyl Phase 033(CO)12 7.37,7.16,7.02, 6.75,6.60 HOS3(CO)12+ + hectoriteb c HOS3(CO)12+ + fluorohectoriteb c 053(00)12 + hectoriteb 7.37,7.16,6.99, 6.75,6.60 053(CO)12 + fluorohectoriteb 7.37,7.16,6.99, 6.60 053(CO)12 + laponiteb 7.37,7.16,6.99, 6.75,6.60 053(00)12 + hectorited 7.40,7.19,7.13 All clay samples are in the form of air-dried films. Clay in aqueous suspension, metal carbonyl in acetone No pattern due to supported carbonyl observed Clay in CH2C12, metal carbonyl in CH2012 0.000 72 ions. The initially bound metal carbonyl transfers a proton to the interlayer water and the neutral cluster is selectively dispersed at or near the edges of the mineral. The electrical neutrality of the 053(CO)12 cluster precludes any electrostatic interaction with the mineral and leads to a poorly dispersed metal carbonyl after evaporation of the solvent. In an attempt to eliminate the formation of 053(CO)12 crystallites on Na+-hectorite, a clay sample (0.2 g) was first pretreated at 350°C under vacuum and then was refluxed with a solution of 033(CO)12 (30 mg, 0.033 mmol) in octane (100 ml) for 24 h. After the reaction time the solution was filtered under nitrogen and the mineral was washed several times with CH2C12 to remove unreacted physically adsorbed osmium clusters. The random freeze-dried clay sample allows for reaction with the metal cluster through surface hydroxyls to form a new species (VII). The x-ray powder diffraction spectrum of VII shows no pattern due to supported carbonyl species, indicative of a high degree of dispersion of the metal species. However, the infrared spectrum is significantly different from those obtained by impregnation techniques (Figure 10, Table 4). The positions and intensities of the (CO) absorptions correspond well to those found by Crawford gt 91.123 for 033(CO)12 supported on alumina by an extraction technique. Although the structure of the osmium species cannot be unequivocally assigned, the IR data suggest that the framework of the cluster remains intact and interacts strongly with the support. It is possible that the spectrum of the clay sample is the superposition of those of the supported triosmium hydride cluster HOS3(CO)10(O-SiE) and ensembles a mononuclear osmium complexes of the type [Os(CO)n(O-Si§)2]x124,127,128. The osmium carbonyl species is stable in air, but heating at 150°C causes degradation and decarbonylation and produces the surface bound mononuclear complexes 73 Figure 10: Infrared spectra in the terminal CO stretching region: (a) clay sample prepared by refluxing 053(CO)12 with freeze-dried Na+-hectorite in octane (mull); (0) sample (a) after heating in air at 150°C for 2 h (mull). 74 J J 1 2200 2000 1900 CM" 75 (Figure 10). The adsorption of metal cluster carbonyls on inorganic oxides has been extensively studied by several workers54975:78. It is generally agreed that, unless the metal carbonyl is simply physisorbed on the surface of the support, interaction between surface hydroxyls and the cluster with concomitant metal oxidation accompanies adsorption from solution. Since 053(CO)12 exhibits a high specificity towards the edge hydroxyl groups in clays, the reaction was used as a means of estimating the edge surface area of three different hectorites. Thus, the clays were refluxed with a solution of 053(CO)12 in octane. Under the reaction conditions the edge surface is accessible to the metal clusters and the metal loading can provide an estimate of the available edge surface area. Elemental analysis data for fluorohectorite, natural hectorite, and laponite are provided in Table 7. Fluorohectorite, having the largest particle size, is expected to posses the lowest edge surface area and consequently binds the least amount of osmium. The metal loading in laponite is the highest, because of the abundance of edge surfaces due to the small size of the clay platelets. The osmium content in natural hectorite is within these two limits, as is its particle size. Although the metal loadings for the three different clays are consistent with the trend in particle size, they only provide a relative estimate of the available edge surfaces. The ratio of edge sites for fluorohectorite, natural hectorite, and laponite from the metal loadings is calculated to be 1:3.3:4.4. In an independent experiment Kadkhodayan139 measured the pseudo first-order rate constants for the reaction of alkyl bromides with NaCl using Ni(Phen)32+/SO42’ intersalated fluorohectorie, hectorite, and laponite as triphase catalysts. Since the substrate attack zone is limited to a very short distance at the edges of the clay, the rate constants would greatly depend 76 Table 7 Elemental Analysis of Os Complexes Supported on Hectorites Clay Particle Size (u) wt% Os Normalized Fluorohectorite >> 2 u 0.40 1.00 Hectorite, Baroid < 2 u 1.33 3.32 Laponite 0.02 1.76 4.40 77 on the particle size of the clay used. Interestingly, the normalized kobs for fluorohectorite, hectorite, and laponite were 1:2.2:3.9, in qualitative agreement with the numbers obtained from metal loadings. 2. Reaction of HRu 3(CO)12+ and Ru3(CO)12 with Hectorite HRU3(CO)12+ was another candidate for site selective binding to a layered silicate surface. The reaction of Na+-hectorite (1.0 meq) in aqeuous suspension with HRU3~ccm .m> 2555 .m_mc~ .532 £22 .552 .55 288535 x a 22 .552 52.2.2 22 .8ch 552m .52: .32: £5 255.288255 5 522 .532 .522 .252 .225 .225. 8:88 .8588 53.2 .2 22 .252 55 8288; + Lmroovmsmfi :5 750 62055705 E3002 :oflntouén o—aEam 2:58:50 00:283— oca 30.6er .5308: no.“ memo—5.52m 9.23055 00 003...“:— 81 Table 9 Electronic Spectroscopy Data for Reactions Products and Reference Compounds System Solvent ‘max (nm) [HRu3(CO)12][PF5] Acetone 403 RU3(CO)12 CH2012 392 [HRU3(CO)12]+ + hectorite film 382 reaction product RU3(CO)12 + hectorite film 380 reaction product 82 mason) + RU3(00)12-—A—;[(55io)23u(co)n1x (10) n = 2:3 Since the above reaction provides a chemical means of distinguishing between Ru3(CO)12 at edges and basal planes, it can be concluded that the clusters are more likely deposited at the more abundant basal surfaces. The somewhat abnormal behavior of HRU3(CO)12+ is consistent with its lower stability in solution when compared to its osmium analog. If deprotonation is faster than the rate of ion—exchange, the driving force that leads to initial binding of the cluster cation near the edges of the clay is absent. Consequently both HRU3(CO)12+ and RU3(CO)12 clusters result in the formation of similar products that contain the metal carbonyl on the basal surfaces. The highly oriented nature of the clay films impedes migration to the hydroxylated edge sites and precludes the formation of new surface species. The importance of sample orientation in hindering migration was demonstrated by freeze-drying sample (XI) and observing reaction according to equation 10 (Figure 12). Heating the ruthenium-containing samples (VIII) and (X1) in air at 110°C results in a decrease in the intensities of the CO bands. The loss of intensity most likely results from the partial air-oxidation of RU3(CO)12 to form Ru02 and 002 by the reaction RU3(CO)12 + 902—93RU02 + IZCOZ (11) as observed previously for the supported ruthenium carbonyl complexlll. 83 Figure 12: Infrared spectra in the terminal CO stretching region: (a) Freeze-dried sample of RU3(CO)12 on Na+-hectorite (0.1 mmol/meq) prepared by impregnation with Ru3(CO)12 (mull); (b) sample (a) after heating in air at 100°C for 30 min (mull). 2200 2000 1900 M-l 85 B. Intercalation of Cationic Phosphine Substituted Carbonyl Clusters on Hectorite Though the reactions of H053(CO)12+ and HRU3(CO)12+ with layered silicates provided an insight into clay-carbonyl systems, the initial intent of this work was to intercalate metal carbonyl clusters in clays. It was felt that a positive charge could be induced on the clusters by utilizing the phosphonium-phosphine ligand thPCHZCHZPPhZCHgPh, abbreviated P-P+. Metal cluster carbonyls are capable of undergoing substitution reactions with neutral phosphine ligands. A similar approach has been used to prepare cationic analogs of well known, otherwise neutral, homogeneous rhodium catalyst5102’103. 1. Synthesis and Exchange Reactions RU3(CO)9(P-P+)3, H4RU4(CO)3(P—P+)4, Ir4(CO)9(P-P+)3, H2053(CO)9,10(P-P+), and 033(CO)11(P-P+) were prepared by the reaction of the parent neutral metal clusters with the positively charged phosphonium-phosphine ligand P-P+ according to procedures, or slight modifications thereof, described earlier for the synthesis of RU3(CO)9L3136, Ir4(CO)9L371, H2053(CO)9,10L67, and 053(CO)11L137 in which L is a tertiary phosphine ligand: RU3(CO)12 + 3P-P+ —é——) RU3(CO)9(P-P+)3 + 3C0 ' (12) MeOH CHC13 H4RU4(CO)12 + 4P-P+ + 4Me3NO ._____, H4Ru4(co)3(P-P+)4 + 4c02 + 4Me3N (13) THF Ir4(CO)12 + 3P-P+ + 3Me3NO ——-) 86 Ir4(CO)9(P-P+)3+ + 3002 + 3Me3N (14) + CHC13 + H2083(CO)10 + P-P ———-9 H2053(CO)10(P-P ) (15) H2033(CO)10(P—P+) T1571" H2053(CO)9(P-P+) + co (16) e CH3CN OS3(CO)12 + N183NO W OS3(CO)11(CH3CN) + C02 + Me;3N (17) CHC13 053(CO)11(CH3CN) + P—P+ ———9 053(CO)11(P-P+) + CH3CN (18) Though complexes of the type H4RU4(CO)8L4140 have been obtained in the past by heating a solution of H4RU4(CO)12 in the presence of the appropriate phosphine ligand at approximately 100°C, the use of trimethylamine oxide in facilitating CO substitution in metal carbonyl complexes is well established. Thus, the reaction of one equivalent of H4RU4(CO)12 with four equivalents of Me3NO°2H20 in the presence of (P-P+)BF4 at room temprature results in the. formation of only the tetrasubstituted phosphine product. The infrared spectra of the cationic P-P+ substituted complexes in the CO stretching region as well as their electronic spectra were in good agreement with those previously reported for their neutral analogues (Tables 10-12). Since the infrared spectrum of a metal carbonyl complex can serve as a fingerprint reflecting its structure and symmetry, IR spectroscopy can Table 1 0 Infrared CO Stretching Frequencies for Molecular and Hectorite-Intercalated Metal Cluster Complexesa Compound v (CO) cm-1 RU3(CO)9(P-P+)3(BF4)3 RU3(CO ) g(P—P+)3-hectorite H4RU4(CO)3(P-P+)4(BF4)4 H4RU4(CO)3(P-P+)4-hectorite Ir4(CO)9(P-P+)3(BF4)3 Ir4(CO)9(P-P+)3-hectorite OS3(CO)11(P-P+)BF4 053(CO)11(P-P+)-hectorite H20s3(co)10(P-P+)BF4 H2053(CO)10(P-P+)-hectorite H2053(CO)9(P-P+)BF4 H2053(CO)9(P-P+)-hectorite H033(CO)9(CH=CH2)(P-P+)-hectorite 2048 m, 1980 s, 1962 s 2042 m, 1970 s, 1960 sh 2005 s, 1971 m, 1948 s 2002 s, 1965, sh, 1946 s 2039 m, 1995 sh, 1984 s, 1965 sh, 1832 vw, 1779 s 2038 m, 1995 sh, 1980 vs, 1965 sh, 1835 vw,1777 s 2100 m, 2048 s, 2028 s, 2010 vs,1998 sh, 1983 m 1976 m,1940w 2102 m, 2050 s, 2025 sh, 2001 vs, 1995 sh, 1975 s 2100 m, 2060 s, 2042 s, 2018 vs,1990 sh, 1976 m 1960 m 2100 m, 2062 s, 2042 s, 2010 vs, 1998 sh,1980 m, 1970 m 2090 m, 2050 s. 2001 vs, 1989 s, 1948 m 2092 m, 2052 s, 2000 vs 1978 s, 1950 sh 2090 m, 2042 5,1999 vs 1962 sh 88 Table 10 continued HOS3(CO)9(viny1)(P-P+)-hectorie 2091 m, 2041 s, 2002 vs 1960 sh, 1940 sh HOS3(CO)9(viny1)(P-P+)BF4 2090 m, 2048 s, 2002 vs 1970 s, 1937 m a All BF4" salts were run as CH2C12 solutions. All hectorite-intercalated products were run as KBr pellets. 89 Table 11 UV—Visible Spectroscopy Data for Hectorite—Exchanged Metal Clusters and Their Molecular Analogs Band Positions System Solvent or Phase in nm RU3(CO)9(P—P+)3(BF4)3 0112012 489, 362 sh RU3(CO)9(P-P+)3-hectorite mull 497, 360 sh H4RU4(CO)3(P-P+)4(BF4)4 0112012 404, 309 sh H4RU4(CO)3(P-P+)4-hectorite mull 417, 310 sh Ir4(CO)9(P-P+)3(BF4)3 0112012 383 sh, 301 lr4(CO)9(P-P+)3-hectorite mull 380 sh, 290. 033(CO)11(P—P+)BF4 CH2C12 407, 323 053(CO)11(P-P+)-hectorite mull 415, 300 H2053(CO)10(P-P+)BF4 0112012 385, 331 HZOS3(CO)10(P-P+)-hectorite mull 380 H2053(CO)9(P-P+)BF4 0112012 359, 304 H2033(CO)9(P-P+)-hectorite mull 362, 315 90 225: com 252.5: 5552 .222: «mm .2Nw.c: 5252 .252.2: 2252 22 .25.: 2.2 .22.: 22 .22.: $2 .22.: :2 2222285522202222.2580:$02: 222.2: 2252 255.5: 252 .22v.2: 2252 .552552 .2vm.2: 5252 . 22 .22.: m2 .322: 2.22 .2252: $2 .22.: 22 22.2222922220252252200:2502: 522 .5522 .5522 .2522 2.2 .522 .522 5522 2252222220820 5m2wp2 .m5222 .552252 22 .222 .522 .5222 222252220832 322 22 52. £22 .5222 .552 Emcmozmvzmoméafoo:2.322.: cm 552 2552: 5552 22 .822: 22. .822 22 .82: 2.2.2 22222222285322022222.2082: .20: E: .82: 4 21:20 ADDY; 2:509:00 N a 036,—. 202509.50 3:09:20 2322—0 .20 2.0590342 02.20.5005— 2028 0020220250222 92:30.25 OD 00.22.25 91 be utilized as a sensitive technique in identifying products with various degrees of phosphine substitution. Besides the CO absorptions, the IR spectra of the mixed phosphine metal clusters exhibit bands characteristic of the P—P+ ligand as illustrated in Figure 13. The absroptions around 1600 cm‘1 most likely are due to phenyl group skeletal vibrations. The strong band at 1440 cm’1 and the bands on either side of it could be assigned to the methyl group scissoring vibrations or they might be due to aryl skeletal vibrations. The bands in the region 1350-1150 cm'1 probably arise from phenyl group in-plane C-H bonding vibrations. The three relatively sharp absorptions at 850-600 cm’1 are assigned to phenyl group C-H out of plane bonding vibrations141’159. The broad strong band at 1060 cm“1 is characteristic of the tetrafluoroborate anion BF4'160. Metal cluster hydrides were further characterized by 1H NMR spectroscopy. H2033(CO)10(P-P+), having a terminal and a bridging hydride that exchange rapidly, shows no resonance in the hydridic region at room t'emperature (RT). Lowering the temperature to -61°C gives rise to a resonance at —10.36 ppm (terminal H), and another at -20.48 ppm (bridging H), similar to those observed for H2053(CO)10L complexes“. Refluxing a solution of H2083(C0)10(P-P+) in methanol results in a slow loss of CO to give H2053(CO)9(P-P+), which exhibits a proton doublet at -11.06 ppm (JHp = 7.25 Hz), suggestirg that both hydridic ligands bridge an 05—05 edge67. The 1H NMR spectrum of H4Ru4(CO)3(P-P+)4 shows, besides the P-P+ ligand resonances, a quintet at -16.50 ppm (JHp = 7.0 Hz) which agrees well with that observed for H4RU4(CO)8L4 type complexesl‘m. Preliminary attempts to exchange interlayer surface Na+ ions with the cationic clusters from suspensions of the mineral in different polar organic Figure 13: 92 Infrared spectra of unsupported and hectorite-intercalated Ir4(CO)9(P-P+)3 in the region 2500-500 cm-l: (a) Ir4(CO)g- (P-P+)3(BF4)3 (KBr pellet); (b) Ir4(CO)9(P-P+)3-hectorite_ (film). Cross-hatched absorptions are also present in the native mineral. 93 00m 1.5 00.2 8.2K comm 94 solvents were unsuccessful. An 001 spacing of about 12.6 X, indicative of Na+-hectorite with a water monolayer, was obtained even after continuous refluxing. Interestingly, successive washings of the clay with the organic solvent did not completely remove the metal complex from the external surface. The clay bound metal complex was shown by infrared spectroscopy to be unchanged. The failure to exchange interlayer Na+ ions by metal clusters containing P-P+ can be explained in terms of the swelling properties of clays“). Polar organic solvents can cause a considerable swelling of silicate interlayers. Nevertheless, the interlayer expansion is generally limited to < 10 2, making it difficult for the bulky complex to penetrate and displace Na+ ions. In other words, the process expressed in equation 19, wherein the heavy lines represent the negatively charged silicate layers with a thickness of 0 approximately 9.6 A, and MS: the cluster cations, is exceedingly slow. or anic nNa” + an+ 751,? Mm!” + nNa+ (19) Since the edges of the smectite structure have broken bonds that contribute to the clay cation exchange capacity (CEC)142, a small amount of the cluster satisfies the "edge charge" deficiency and remains on the surface even after repeated washings. Despite the above limitations encountered in organic solvents, exchange of interlayer Na+ by metal cluster complexes could be achieved by first suspending the clay in water and then adding this suspension to an acetone solution of metal cluster cation. Under these conditions the clay layers 95 are largely delaminated in aqueous suspension and the Na+ ions on the basal surfaces are readily available for ion exchange. This is in contrast to the situation expressed in equation 19 where the clay layers are initially oriented face to face. Upon the addition of the clay suspension to the cluster cation solution, the Na+ ions are replaced and the silicate layers reorient themselves in a face to face fashion. The overall process can be represented schematically as shown in equation 20. Na+ Mm!1+ Na+ Mm“ . (20) an+ J Na+ Intercalation of the different cluster cations in the interlamellar space of Na+-hectorite results in a considerable increase in the 009. spacing of the mineral (Table 13, Figure 14). The lack of more than two orders of reflection from the x-ray diffraction patterns indicates the presence of highly interstratified materials. The increase in spacing most probably is determined by the size of the phosphonium—phosphine ligand. The size and shape of the cluster framework might also play a minor role. Although one would expect intercalated cluster complexes of the same type but with higher degrees of P—P+ substitution to yield materials with higher 002 spacings, our results mitigate against this. Complexes with more than one P-P+ ligand might adopt a more tightly-packed configuration in order to accommodate 96 Table 1 3 001 X—ray Basal Spacings of Hectorite-Intercalated Metal Cluster Carbonyl Complexes Sample d001 >D RU3 ( CO )9 ( P-P+) 3-hectorite H4RU4(CO )3 ( P-P+)4-hectorite Ir4 ( CO )9 ( P-P)3-hectorite OS3(CO )1 1(P—P+)-hectorite H2033(CO)10(P-P+)-hectorite H2053(CO)9(P-P+)-hectorite P-P+-hectorite 22. 23. 23. 27. 26. 26. 18. 97 Figure 14: X-ray diffraction patterns of : (A) Ru3(C0)9(P-P+)3-hectorite; (B) H4RU4(CO)3(P-P+)4-hectorite; (C) Ir4(CO)9(P-P+)3-hectorite; (D) 053(CO)11(P-P+)-hectorite; (E) H2053(CO)10(P-P+)- hectorite; (F) H2053(CO)9(P-P+)-hectorite. Samples were prepared as oriented films on glass slides by first suspending the clay complex in water and allowing the suspension to evaporate on the slide at room temperature. 98 99 the bulky ligand molecules. On the other hand, substitution of P—P+ for only one CO molecule leads to molecules with a less strained configuration and results in exchanged clays with higher spacings. The observed differences in interlayer expansion might also be due to the way the intercalated clusters are oriented toward the silicate sheets. The intercalated metal carbonyl complexes were characterized by comparison of their carbonyl infrared spectra with those in solution (Table 10, Figures 15—19). The close similarity in band positions and relative intensities suggest that the cationic clusters remain unchanged upon intercalation, the interaction between host and intercalate being electrostatic. In addition to the CO absorptions and bands characteristic of the silicate framework, the infrared spectra of the clay-intercalates exhibit absorptions due to the P-P+ ligand (Figure 13). Another indication of the retention of chemical and structural constitution of the cluster complexes upon intercalation was provided by comparison of their electronic spectra with those obtained in solution. The transitions associated with M-M bonds in the cluster framework of the complexes are observed in the electronic spectra of the exchanged minerals and their molecular analogs (Table 11). The small deviation in "max between the intercalated clusters and those in solution can be attributed to the interaction of the former with the negatively charged silicate sheets, as well as to changes in the solvation environment. The hectorite intercalated metal cluster complexes were further characterized by elemental analysis. The metal to phosphorus ratios, obtained by chemical analysis, were in good agreement with those predicted theoretically (Table 14). The loadings reported in the Table correspond to 26 to 82 meq/100 g of clay. Figure 15: Figure 16: 100 Infrared spectra in the CO stretching region of unsupported and hectorite-intercalated RU3(CO)9(P-P+)3: (a) RU3(CO)9- (P-P+)3(BF4)3 in cnzc12 solution; (b) RU3(CO)9(P-P+)3- hectorite (KBr pellet). Infrared spectra in the C0 stretching region of unsupported and hectorite-intercalated H4Ru4(CO)3(P-P+)4: (a) H4Ru4(CO)3(P-P+)4(BF4)4 in CHZCIZ solution; (b) H4Ru4(CO)3(P-P+)4-hectorite (KBr pellet). 101 0' + J 1 I 1 l A #L 2200 2000 1800 2200 2000 1800 CM ' ‘ c M ‘ Figure l 5 Figure 16 Figure 17: Figure 18: 102 Infrared spectra in the terminal and bridging CO stretch- ing region of unsupported and hectorite-intercalated lr4(co)g(P-P+)3; (a) Ir4(CO)9(P-P+)3(BF4)3 in CH2C12 solution; (b) Ir4(CO)9(P-P+)3-hectorite (KBr pellet). Infrared spectra in the CO stretching region of unsupported and hectorite-intercalated 033(CO)11(P-P+): (a) 033(CO)11- (P-P+)BF4 in CH2C12 solution; (b) 033(CO)11(P-P+)— hectorite (KBr pellet). 2200 2000 CM“ Figure 17 \ VI 2200 2000 1800 CM ' Figure 18 Figure 19: 104 Infrared spectra in the CO stretching region: (a) H2053(CO)10(P-P+)(BF4) in CH2C12 solution; (b) H2053(CO)10(P-P+)-hectorite (KBr pellet); (c) H20s3(co)9(P-r+)(BF4) in CH2C12 solution; (d) H20$3(CO)9(P-P+)-hectorite (KBr pellet). 106 55.2 25.2 25.2 25.52 52:38:23-2:22208550 55.2 55.2 22.2 22.52 822885-23-..3200:5502: 25.5 55.5 55.2 22.5 8228825232558832 55.5 25.5 55.2 25.5 8238:5232588532 55.5 55.5 25.2 25.5 822385-23238825222.: 25250 20:25.2 .2 mo .22 :22 58555 (loss. 22.. .2\=|\ r \ its min—23.2250 232:0 22325225023505.2300: no £55292 53.20.2205 52 — 039—. 107 2. Polarized Infrared Studies of H4Ru4(CO)3(P-P+)4-Hectorite Tetrasubstituted phosphine derivatives of the type H4RU4(CO)8L4 with an even distribution of the phosphine ligands on the ruthenium atoms possess D2d symmetry. Such a structure would exhibit three infrared-active CO stretching vibrations (B2+2E) as shown in Figure 16 and Table 10. The weak band at about 1920 cm"1 is attributed to 13C isotopic substitution140. The stretching frequency of B2 symmetry is polarized in the z-direction while the doubly degenerate mode is x,y polarized. The geometry of the molecule is preserved upon intercalation and the IR spectrum of H4RU4(CO)3(P-P+)4-hectorite, as an oriented film, exhibits CO absorptions at 2002 s, 1964 m, and 1942 s cm‘1 (Figure 20). Because of the simplicity of the infrared pattern, we undertook a study to determine whether the intercalated molecules align themselves in a certain orientation towards the silicate sheets. Highly oriented films are built by stacking alternate arrays of silicate layers and interlayer cations in the crystallographic c-dimension. In a typical experiment the clay c-dimension is parallel to the IR beam (6: 0°). By rotating the clay film (6 #0°) the absorption that is polarized along an axis parallel to the c-dimension should increase in intensity while the bands polarized along the other two axes are expected to decrease. The infrared spectra of H4RU4(CO)3(P-P+)4-hectorite for different angles towards the beam are shown in Figure 20. Since we observe no significant changes in the three CO absorptions we conclude that the clusters do not adopt any specific orientation but rather are intercalated in a random fashion. Figure 20: 108 Infrared absorbance spectra in the CO stretching region of H4RU4(CO)3(P-P+)4-hectorite at different angles of clay film relative to the direction of the IR beam: (000 = 90°. The IR beam is perpendicular to the clay film; (b) 0 =110°;(c)0 =130°. EVEE 110 3. Olefin Isomerization by Homogeneous and Hectorite-Intercalated H2033(CO)9(P-P+) Catalysts. The decarbonylation of H2033(CO)10(P-P+)BF4 in refluxing methanol and H2053(CO)10(P-P+)-hectorite under vacuum at 80°C yields H2053(CO)9(P-P+)BF4 and H2053(CO)9(P-P+)-hectorite, respectively, the resulting products having been identified by their infrared spectra (Figure 19, Table 10). Our results parallel those obtained for molecular and supported triosmium hydride carbonyl clusters, designated as H2053(CO)10PPh3 and H2053(CO)10PPh2-Support, which are known to lose CO on heating and become coordinatively unsaturated57’107,108. Because of their coordinative unsaturation they are able to react with electron pair donors such as acetylene and olefins. We found both homogeneous and clay-intercalated triosmium clusters to be active olefin isomerization catalysts. Catalytic reactions of the soluble salt were carried out at ambient temperature and pressure in acetone without exclusion of air by using an initial substrate to cluster mole ratio of 800. l-Hexene was isomerized to cis- and trans-Z-hexene under homogeneous conditions producing 3300 moles of internal olefin per mole of triosmium cluster prior to catalyst cessation. After reaction the color of the solution had changed from deep green to yellow and the infrared spectrum indicated the presence of the saturated HOS3[CH3(CH2)3CH=CH2](CO)9(P-P+)BF4 with a bridging vinyl ligand107 (Table 10). Figure 21 illustrates the results for 1-hexene isomerization with H2033(CO)9(P-P+)-hectorite under the same reaction conditions .used for the homogeneous catalyst. The dramatic difference in reaction rates for the three different solvents can be related to the extent of interlayer swelling. Among the three, toluene is the poorest swelling solvent and exhibits the 111 Figure 21: Isomerization of 1-hexene (1-3.3 M) to 2-hexene with H2033(CO)9(P-P+)-hectorite (0.01 mmol) in three different liquid media at ambient temperature and pressure. The clay catalyst contained 2.17 wt% P and 19.27 wt% Os. 2 - HEXENE (mmol) 112 Acetone MeOH e Toluene 0 e . e 7‘ q A A A l 1 L 4 8 12 TIME (hr) 113 slowest rate. Contrary to previous results with clay-intercalated catalysts, we observe some catalytic activity that can be attributed to the large 002 spacing observed for the mineral. In other words, the cationic complex acts as its own pillar making the interlayer accessible to substrate molecules even in the absence of a swelling solvent. The relative rate for the intercalated catalyst is higher in acetone than that observed in methanol, consistent with the differences in 0025pacing obtained by solvating the mineral in these two solventsl41’161. In all three reactions the color of the clay changed from deep green to shiny yellow, after which time the catalyst was no longer active. It can be inferred from the IR spectrum that catalytic cessation is due to the formation of a saturated complex. A similar complex with a bridging vinyl group is formed when acetylene is passed through a suspension of the intercalated catalyst in acetone (Table 10). During the course of the reaction in methanol, but not acetone or toluene, a small amount of the metal complex is desorbed into the solution. Since electrical neutrality must be maintained within the clay, the desorbed complex most likely is not cationic. To investigate further the role of methanol we examined the desorption of clusters from hectorite in the absence of substrate. After stirring the clay suspension overnight in methanol, the color of the mineral changed from green to yellow, and a portion of the cluster was desorbed into the solution. No isomerization to internal olefin was observed when l-hexene was added to this slurry. The infrared spectrum of the intercalated catalyst resembled that of the inactive catalyst obtained after reaction with the substrate. This observation suggests saturation of sites by methanol ligands in a manner analogous to that of the Os vinylidene complex. Furthermore, the spectrum of the filtrate exhibited terminal 114 co bands at 2113 w, 2090 m, 2051 m, 2002 vs, 1940 m, br cm’l, and no absorptions characteristic of the P-P” ligand. The complex pattern of the IR spectrum is an indication that the cluster remains intact. Catalyst desorption may be atrributed to ligand dissociation and coordination of solvent molecules. Solvent coordination has been proposed to occur for some photoreactions of tungsten hexacarbonyl in methanol, though the product has not been isolated143. Dissociated P-P+ molecules satisfy the negative charge of the silicate layers: + MeOH + H2053(CO)9(P-P ) -—) P-P + metal cluster carbonyl (soln) (21) The remaining intercalated clusters are incapable of isomerizing l-hexene because they are coordinatively saturated with solvent molecules. No desorption of catalyst was observed when either acetone or toluene was used as solvating agent. The filtrate after catalytic reaction exhibits neither catalytic activity nor IR absorptions in the CO stretching region. The reaction of the toluene-solvated intercalated catalyst at ambient temperature and pressure exhibits pseudo first—order behavior with Kobs = 0.022 h'l. No homogeneous reaction was run in toluene because of the insolubility of the cluster in this solvent. For acetone and methanol a pseudo first-order behavior was not obtained, suggesting catalyst decomposition competes with the isomerization reaction. In another series of experiments some data on the kinetics of 1-hexene isomerization, catalyzed by H2033(CO)9(P-P+)-hectorite in toluene at 75°C, were obtained. No catalyst decomposition was observed during the course of the reaction. We found the rate of the reaction to be proportional to the concentration of olefin as well as the concentration of the catalyst, 115 in agreement with results obtained by Gates and co—workers107 for l-pentene isomerization in the presence of H2083(C0)10 in toluene at 75°C. In the same report these authors claimed that the silica supported catalyst was more stable than its soluble analog, yielding approximately 103 molecules of internal olefin per mole of cluster prior to catalytic cessation. In the present study, both the intercalated and soluble catalysts exhibited almost the same turnover number (3300 moles olefin/mole cluster), which is more than three times as large as the value obtained for the silica supported triosmium cluster. The similarity between soluble and intercalated catalysts is additional evidence that in sufficiently swollen smectites intercalated catalysts exhibit solution-like properties and do not lose appreciable activity compared to their homogeneous counterparts. The higher activity observed for the clay-intercalated triosmium cluster illustrates the advantage of using electrostatically supported catalysts. Immobilization of a metal complex through a covalent attachment will always reduce its mobility and consequently its catalytic activity. The soluble and supported triosmium hydride clusters are the only examples for which well-established catalytic cycles involving a cluster complex have been proposed67;1079108. The overall catalytic cycle is illustrated in Figure 22. Our results are in agreement with previous observation567’107 that during the reaction neither cluster fragmentation nor metal aggregation takes place, the catalyst being the intact cluster itself. 4. Thermal Stability of Hectorite-Intercalated Metal Cluster Carbonyls and Interaction with H2 and 02. Decarbonylation of immobilized metal cluster carbonyl complexes leads to highly dispersed metallic particles on both organic and inorganic 116 Figure 22: Catalytic cycle for isomerization of a -olefins catalyzed by triosmium cluster567t107. The carbonyl ligands are omitted for simplicity. The HOS3(vinyl)(CO)10 is catalytically inactive. 117 RY + HOS3(y'inyl)((‘0)l 0 H 01.8. Freeman. ALA. Patrick and B.(‘. Oates. J. Fatal” 73. 8'2 (1982). 118 supports86987. Studies on the thermostability of the supported complexes usually yield valuable information in evaluating their catalytic properties. Characterization of the different subcarbonyl species can be accomplished by IR spectroscopy, the carbonyl absorptions serving as a fingerprint of the various metal species. Figures 23 through 30 illustrate the effects of heating the clay-intercalated metal clusters under various conditions. Though unequivocal characterization of all metal subcarbonyl species is impossible at present due to the lack of soluble analogs exhibiting the same IR absorptions, an attempt will be made to rationalize the results obtained on thermolysis or oxidation. Simultaneous decomposition of the P-P+ ligand adds an extra barrier in correlating surface species with authentic reference compounds. The decomposition of immobilized clusters on typical metal oxides is likely to involve initially reaction of the metal atoms with surface hydroxyl group5115’134. Decarbonylation produces coordinatively unsaturated metal atoms, which in turn interact strongly with the support. In this case fragmentation and/or oxidation of the metal atoms can be anticipated. Complete decarbonylation would result in the formation of metal silicates or, when thermolysis is carried out in 3 H2 environment, metal aggregates. However, only the edge surfaces of clays are hydroxylated. This distinguishes edge sites from basal planes which are composed of siloxane-type oxygen atoms. Heating a clay film could cause migration of the metal clusters or their fragments from the interlayer region, where they are initially located, to the edges and reaction with surface hydroxyl groups. Such migration is in agreement with the observed reduction in 001 spacing to a value of approximately 16 A after heating the clay sample above 200°C. The higher thermostability obtained for the intercalated carbonyl clusters compared Figure 23: Figure 24: 119 Infrared spectra in the CO stretching region of products formed by thermal decomposition of RU3(CO)9(P-P+)3— hectorite: (a) Ru3(CO)9(P-P+)3-hectorite (film); (b) sample (a) after heating in vacuum at 110°C for 11 h; (c) followed by heating at 200°C for 9 h; (d) 300°C for 10 h; (e) sample (a) after heating in H2 at 200°C for 10 h; (f) followed by heating at 300°C for 11 h. Infrared spectra in the CO stretching region of products formed by oxidation of RU3(CO)9(P-P+)3-hectorite in air: (a) RU3(CO)9(P-P+)3-hectorite (film); (b) sample (a) exposed to air for 2 months; (c) sample (a) after heating in air at 110°C for 10 h; (d) followed by heating at 200°C for 45 min; (e) 200°C for 90 min; (f) 200°C for 3% h; (g) 200°C for 10 h; (h) 300°C for l h. 120 is i a" ins" L/ _l 1 1 2200 2000 1800 220° 200° 130° cm“ CM" Figure 23 Figure 24 Figure 25: Figure 26: 121 Infrared spectra in the CO stretching region of products formed by thermal decomposition of H4Ru4(CO)8(P-P+)- hectorite: (a) H4Ru4(CO)3(P-P+)4-hectorite (film); (b) sample (a) after heating in vacuum at 110°C for 11 h; (c) followed by heating at 200°C for 9 h; ((1) 300°C for 10 h; (e) sample (a) after heating in H2 at 200°C for 10 h; (f) 300°C for 11h. Infrared spectra in the CO stretching region of products formed by oxidation of H4Ru4(CO)3(P-P+)4—hectorite in air: (a) H4RU4(CO)3(P-P+)4-hectorite (film); (b) sample (a) after heating in air at 110°C for 10 h; (c) followed by heating at 200°C for 45 min; ((1) 200°C for 90 min; (e) 200°C for 3% h; (f) 200°C for 10 h; (g) 300°C for l h. Figure 27: Figure 28: 123 Infrared spectra in the CO stretching region of products formed by thermal decomposition of Ir4(CO)9(P-P+)3— hectorite: (a) Ir4(CO)9(P-P+)3-hectorite (film); (b) sample (a) after heating in vacuum at 200°C for 9 h; (c) followed by heating at 300°C for 10 h; (d) sample (a) after heating in H2 at 200°C for 10 h; (e) 300°C for 11 h. Infrared spectra in the CO stretching region of products formed by oxidation of Ir4(CO)9(P-P+)3-hectorite in air: (a) Ir4(CO)9(P-P+)3-hectorite (film); (b) sample (a) after heating in air at 110°C for 10 h; (c) followed by heating at 200°C for 45 min; ((1) 200°C for 10 h; (e) 300°C for 1 h. 124 0W .“-‘\ '1.“ M _I_ 5L 1_ 1 _L .1; _I 4 J L 2200 2000 1800 2200 2000 1800 CM -' CM M Figure 27 Figure 28 Figure 29: Figure 30: 125 Infrared spectra in the CO stretching region of products formed by thermal decomposition of 033(CO)11(P-P+)— hectorite: (a) 053(CO)11(P-P+)-hectorite (film); (b) sample (a) after heating in vacuum at 200°C for 9 h; (c) followed by heating at 300°C for 10 h; (d) sample (a) after heating in H2 at 200°C for 10 h; (e) 300°C for 11 h. Infrared spectra in the CO stretching region of products formed by oxidation of 053(CO)11(P-P+)-hectorite in air: (a) 053(CO)11(P-P+)-hectorite (film); (b) sample (a) after heating in air at 110°C for 10 h; (0) followed by heating at 200°C for 45 min; (d) 200°C for 90 min; (e) 200°C for 10 h; (f) 300°C for 1 h. 2200 2000 CM" Ffigwe29 1800 126 127 to their silica supported analogs is also in agreement with the proposed decomposition mechanism. Though oxidative degradation of the cluster by surface hydroxyl groups appears to be the principal decomposition process, adsorbed water or oxygen might be involved as we11128. When RU3(CO)9(P—P+)3-hectorite (designated hereafter RU3-HEC) is exposed to air at ambient temperature for a long time or heated in vacuum or in flowing 02 (Figures 23 and 24) a similar spectral pattern is obtained. By analogy to the silica supported systems the bands located at 2060-2070 and 2000-2015 cm”1 can be assigned to two surface structures of the type [Ru(CO)2LX2]n or [Ru(CO)2X2]n in which L represents the phosphine ligand and X are likely to be oxygen atoms of the support. The band at 1950-1960 cm‘1 is probably due to another surface species. The value of n, which depends on metal aggregation, cannot be specified on the basis of the present results. Persistent oxidation at 200°C leads to another species with a low frequency absorbance, most likely a species having one carbonyl per Ru atom. As the number of CO ligands decreases the availability of Ru d electrons for back bonding increases, shifting the CO absorbances to lower frequenciesl44. In the absence of 02 the bands attributed to oxidized Ru species are less resolved and in the presence of H2 the main surface structure is that exhibiting a broad band at 1960 cm‘l, further supporting our assignment. At 300°C decarbonylation is complete and Ru silicates or metal crystallites are likely to be formed. No attempt was made to identify metal aggregates by x-ray diffraction. H4Ru4(CO)3(P-P+)4-hectorite behaves similarly to RU3-HEC (Figures 25,26), though decomposition begins at much higher temperatures. The higher thermostability of the former has been attributed to its higher nuclearity115. 128 Table 1 5 Vibrational Frequencies for Some Carbonyl Metal Complexes Compound v(CO) cm'l Ref. Ru12(CO)2(PPh3) 2061 s, 2005 s 136 [Ru(CO)212]n 2050 s, 1995 s, 1975 w(sh) 146 [Ru(CO)312]2 2116 m, 2059 s, 2009 m 146 Os(CO)3(PEt3)C12 2144 w, 2070 vs, 2023 s 147 [Os(CO)2(PPh3)C12]2 2051 s, 1981 s 147 [Os(CO)312]2 2118 s, 2050 vs 127 [Os(CO)212]n 2112 w, 2039 s(br), 1980 s(br) 127 129 At approximately 200°C Ir4(CO)9(P—P+)3-hectorite preferentially loses bridging carbonyls on heating in H2 or under vacuum (Figure 27) as indicated by the loss of CO intensity in the region of approximately 1800 cm‘l, in agreement with results reported for [Ir4(CO)9(PPh3)ZPPh2(CH2)3SIL], though decarbonylation of the latter starts at lower temperatures (approximately 100°C)129. Metal aggregation seems to be unlikely, but cannot be ruled out, because CO chemisorbed on Ir metal aggregates is expected to show broad bands in the 2050-2070 cm'1 region145. However, as the ligand coverage decreases, the absorbances shift to lower frequencies (1995-2030 cm‘1)134. When the intercalated cluster is heated in 02 the two-band pattern is believed to arise from a dicarbonyl iridium(I) species (Figure 28). Evacuation of 053(CO)11(P-P+)-hectorite at 300°C leads to decomposition (Figure 29) and the spectrum resembles that reported for heated [OS3(CO)11PPh2(CH2)2$IL]120. In the presence of H2 a hydride is likely to be formed. Oxidation at 200°C leads to the appearance of new surface structures (Figure 30). At an early stage of decomposition there are probably two different species present. The bands at 2040 and 1990 cm'1 can be assigned to surface structure [OS3(CO)2LX2]n, while those at 2140 and 2060 cm’1 to [Os(CO)xX2]n, (x = 2,3). Oxidation at 300°C leads to oxidized Os on the surface of the clay. C. Cluster Carbonyl Interactions with Alumina Pillared Clay (ALPILC) Polyoxocation pillared clays constitute a new class of highly stable materials characterized by fixed porosity, high surface area, and intracrystal acidity. In this portion of the dissertation discussion centers at infrared spectroscopy that was used to elucidate the interactions of 053(CO)12, RU3(CO)129 H2053(CO)10, H4RU4(CO)12, Ir4(CO)12 and [CpFe(CO)2]2 with 130 alumina pillared clay. The same technique also allowed characterization of the metal surface species formed on thermal treatment under various conditions. An alumina pillared clay was prepared by reaction of Na+-montmorillonite with aluminum chlorohydrate solution, as described previously (Section II.C.5), and allowed to equilibrate with air before use. 1. Adsorption of 033(CO)12 on ALPILC When 053(CO)12 (0.02 mmol) in 40 m1 CH2C12 is added under Argon to alumina pillared clay (0.5 g) which has been dried £1 _\_/_a_c_u_9_ at room temperature for 4 h, the mineral, after removal of the weakly adsorbed Os clusters by washing, exhibits IR carbonyl bands at 2134 s, 2100 s, 2074 vs, 2061 s, and 2023 m cm‘l. This set of bands, considered to belong to a single species, is nearly identical in frequency and intensity to that of [HOS3(CO)12][PF6] in MeN02 (Figure 31, Table 16). The Os loading of the dry mineral is 0.38 wt%. The electronic spectrum of this material shows a maximum at 374 nm, whereas a value of 366 nm is observed for the molecular analogue [HOS3(CO)12][PF6] (Table 17). The small deviation in 4max between intercalated and molecular cluster can be attributed to interactions with the negatively charged mineral and changes in its solvation environment. It is noteworthy to point out that, if the impregnation of ALPILC is carried out in the presence of air, no significant carbonyl adsorption is observed after washing the support with CH2C12. The protonated osmium clusters are not stable in moist air and deprotonate, producing the neutral carbony165. The inability to extract HOS3(CO)12+ from the surface of the pillared clay by a more polar solvent such as CHC13 or acetone is in agreement with its cationic character. However, it is possible to extract the protonated cluster by ion exchange with a solution of KPF6 in acetone, presumably Figure 31: 131 Infrared spectra in the CO stretching region: (a) HOS3(CO)12+/‘ ALPILC prepared by impregnation with 053(CO)12 (mull); (b) [HOS3(CO)12]PF6 in MeN02 solution; (c) sample (a) exposed to air at RT for 5 h (mull); (d) 053(CO)12 in CHZCIZ solution; (e) sample (c) after heating in vacuum at 100°C for 2 h (mull); (f) 250°C for 6 h; (g) sample (c) after heating in air at 110°C for 12 h. 133 Table 16 Infrared CO Bands of Molecular and ALPILC-Intercalated Metal Cluster Complexes Compound v(CO) cm’1 Reference [H033(CO)12]+/ALPILC9 2134 s, 2100 s, 2074 vs This work 2061 s, 2023 m 3 [HOS3(CO)121[PF61b 2134 s, 2101 s, 2077 vs This work 2058 s, 2018 m [HOS3(C0)12HPF61b 2135 s, 2102 s, 2080 s 65 2061 s, 2021 m ** [HOS3(C0)12HPF61b 2133 s, 2098 s, 2073 vs This work 2063 s, 2022 m 053(CO)12/ALPILC3# 2068 s, 2034 s, 2014 m This work 2001 m 053(CO)12° 2068 vs, 2034 s, 2013 m This work 2000 m [HRU3(CO)12]+/ALPILC3 2128 s, 2099 s, 2077 vs This work 2060 s, 2030 s [HRu3(c0)12][PF6]b‘ 2128 s, 2100 s, 2077 vs This work 2062 s, 2025 m [HRU3(CO)12HPF61b 2129 s, 2102 s, 2081 s 65 2068 sh, 2030 m [HR03(C0)12][PF5]b** 2126 s, 2098 s, 2076 vs This work 2059 s, 2027 m Ru3(c0)12/ALPILCM 2060 vs, 2029 s, 2010 m This work RU3(CO)12c 2061 vs, 2029 s, 2011 m This work HOS3(CO)10(O-Al< )8 2101 m, 2068 vs, 2056 vs, This work 2019 s, 2005 m H053(CO)10(O-Al< )d 2107, 2068, 2055 128 2030 H033(CO)10(O-Al< )d 2107 w, 2068 s, 2056 s, 124 2023 vs, 2005 m Table 16 continued 134 HOS3(CO)10(O-Al< )6 2107 w, 2067 s, 2055 s, 127 2027 vs, 2012 8 (br) H4RU4(CO)12/ALPILCa 2114 m, 2081 s, 2066 vs This work 2024 s H4RU4 200°C) further decarbonylation occurs and leads to oxidized Ru atoms on the surface. The higher thermal stability of H4Ru(CO)]2 compared with that of the trinuclear RU3(CO)12 cluster has been attributed to its higher nuclearity115. 5. Adsorption of Ir4(CO)12 on ALPILC When Ir4(CO)12 is refluxed in cyclohexane (0.11 x 10’3 M) with alumina pillared montmorillonite (0.5 g) for two h, the IR spectrum of the resulting material exhibits bands at 2109 w, 2060 vs, and 2026 s cm’1 , in good agreement with those of Ir4(CO)12 (Figure 35, Table 16). Exposure to air does not result in any significant change in the IR spectrum. However, if the mineral is heated in air at 110°C a progressive diminishment of the .2026 cm’1 band and the appearance of a new band around 1990 cm'1 is observed. After two hours there appear only two bands at 2061 and 1988 cm‘l, assigned123 to surface structure of the type [Ir(CO)2X]n, where X represents an oxygen atom of the support, which is then decarbonylated on further heating. 6. General Considerations on the Adsorption of Cluster Carbonyls on ALPILC A schematic view of the reactions and the products of various metal cluster carbonyls and alumina pillared clay is represented in Figure 36 and 37. The initial product of the interaction of either 053(CO)12 or RU3(CO)12 with the support is the protonated trinuclear cluster. The 11* ions of the support are produced by dehydration of the interlayering polymeric oxyaluminum cations during pretreatment at 350° under Ar33. The overall Figure 35: 145 Infrared spectra in the CO stretching region of unsupported Ir4(CO)12 and Ir4(CO)12 supported on alumina pillared clay: (a) Ir4(CO)12 adsorbed on ALPILC from cyclohexane solution (mull); (b) Nujol mull of Ir4(CO)12; (c) sample (a) after heating in air at 110°C for 2 h (mull). 147 Figure 36: Schematic representation of reactions and their products of 053(CO)12 and alumina pillared clay (ALPILC). ' 053(CO)12 + ALPILC \ CH2C12. RT 148 OS3(CO)12 H2053(CO)10 + + ALPILC ALPILC octane, 120°C CHQCIQ. RT \. KPF5 . H053(CO)12+PF5' {—-—-— H053(CO)12+/ALPILC HOS3(CO)12(O-Al<) (“I“) vacuum 100°C Air, RT [0s(c0)x(o-A1:)21n OS3(CO)12 + Drv Ground —, OS3(CO)12/ALPILC ‘ Air, RT Slow ALPILC 1 Air. 110°C Fast vacuum, T > 150°C vacuum, 150°C or Air. 110°C Fast or Air, RT slow 4 [OS(CO)x(O’Al<)2]n (x=2.3) vacuum, > 300°C Surface Os Aluminosilicate 149 Figure 37: Schematic representation of reactions and their products of supported RU3(CO)12 on alumina pillared clay (ALPILC) 150 Ru3(CO)12 + ALPILC H4RU4(CO)12 + ALPILC CH2C12. RT CHQCIZ, RT HRu3 200°C 4 Surface Ru Aluminosilicate 151 interlayer reaction may be represented by the following equation. (3-n)+ -H20 + A113O4(OH)28+n —-’ 6.5 A1203 + (3-n)H (23) Bronsted acidity of alumina pillared clays was postulated to account for their catalytic activity as cracking catalysts30232. Subsequent pyridine adsorption studies established the presence of both Bronsted and Lewis acid sites”. Though complexes of metal carbonyls are known to be formed with strong Lewis acids (i.e. AlBr3)148, we did not observe the formation of such an adduct. This might be due to the lower acidity of the support compared to that of AlBr3. It has been reported that by evacuating the pyridine-loaded alumina pillared clay at 150°C the number of Lewis-acid types is increased and that at 400°C the surface acidity is mostly of the Lewis type”. However, our evidence mitigates against this report. Increasing the activation temperature of the support from room temperature to 350°C under vacuum increases the amount of protonated cluster on the surface, as judged by IR, but presents no evidence for the formation of any other species. This implies that as the amount of surface water decreases, the protonic acidity increases. The almost identical spectra of the supported protonated clusters and their molecular analogues as well as the ability to extract them from the surface by ion exhcange suggest the presence of only electrostatic interactions between host and intercalate. It is noteworthy that such a procedure leads to activation of surface protons for subsequent exchange with another cation. This possibility adds a new dimension to the surface chemistry of pillared clays since it has been claimed that they do not exhibit any cation exchange capacity. 152 Elemental analysis of the ruthenium and osmium bound clusters showed that the samples contained 0.48 wt% Ru and 0.38 wt% Os, respectively, when the clay was activated at 25°C. No attempt has been made to optimize the amount of surface bound metal carbonyls. These loadings correspond to a coverage of approximately 8 m2 for the ruthenium and 3 m2 for the osmium cluster per gram of clay, if the surface of the clusters is taken to be about 80 A2. The low coverage of the surface most likely is due to the inaccessibility of the majority of the pores to the cluster molecules because of their comparable size63254. That protonation takes place in the intracrystalline region of the support is further supported by the failure of alumina pillared fluorohectorite with a surface area of approximately 14 mZ/g and smaller pore opening to protonate metal cluster carbonyls. Significantly Al3+ -MONT (d001 = 11.6 A, at 350°C) only forms trace amounts of protonated cluster on external surfaces, as judged by IR. In a similar experiment sodium-montmorillonite fails also to protonate either Os or Ru carbonyl clusters. The absence of acidic character from the naturally occuring mineral is attributed to the ability of Na+ ions to enter the hexagonal recesses in the basal oxygen sheet during dehydration of the clay sample. These firmly fixed ions are stripped of their potentially acidic water molecules and are coordinated instead with oxygen atoms of the silicate framework149. Since protonation of metal cluster carbonyls can be only brought about by treatment with very powerful acidic media such as trifluoroacetic or 98% sulfuric acid55’150, it is evident that the acidity of alumina pillared clays lies in the same range in terms of protonating capacity. This is not unusual for silicate minerals, which were shown to possess Bronsted acidity comparable in some cases to that of concentrated sulfuric acid. In these cases the most 153 important source of the acidity is polarized water molecules coordinated to polyvalent interlayer cations1 5. On the other hand, the acidity of pillared clays arises mainly from protons formed by dehydration of the interlayering polymeric oxyaluminum cations33’38. Of the five different carbonyl clusters examined in this work only RU3(CO)12 and OS3(CO)12 form protonated surface species on contact with the support. This is not surprising since only the dodecacarbonyl triosmium and triruthenium clusters are known to form fairly stable protonated complexes that can be isolated in the form of hexafluorophosphate salts. The carbonyls Ir4(CO)12 and H2083(CO)10 are also known to form solutions that exhibit high field 1H NMR resonances upon treatment with strong acids, suggestive of the formation of protonated species even though solid salts have not been isolated659151. Intracrystalline acidity distinguishes alumina pillared clays from other commonly used supports such as refractory metal oxides. Nevertheless, the small aluminum oxide aggregates sandwiched between the silicate layers exhibit reactivity which parallels that of bulk oxide. Thus, reaction of metal clusters with hydroxyl groups at elevated temperatures leads to degradation of the cluster with formation of subcarbonyl species bound to the support. That the hydroxyl groups of the small oxide clusters are involved rather than those present on the edges of the silicate framework is supported by the ability of the former to react with metal clusters at lower temperatures. The reactivity of the mineral silanol groups with neutral carbonyls is far less than that of the hydroxyl groups introduced by the molecular pillar (see Section Ill.A.1 and III.A.2). Reaction of H2053(CO)10 with ALPILC gives rise to a trinuclear supported osmium species which is probably formed by the following reaction: 154 H2053(CO)10 + H - 0 - A1< —-) OS3(H)(O-Al()(CO)10 + H2 (24) A similar surface bound species is formed by interaction of OH groups of silica or alumina with either 053(CO)12 or H2083(CO)10. In addition to the similarity of its IR spectrum to that of a model compound prepared by the reaction of 053(CO)12 with HOSiPh3, EXAFS and Raman spectroscopy have shown the cluster framework to remain intact and suggest the following structure (1)1 24212721 28. / (I) M = Si, Al, Ti or Zn 155 Treatment of this surface bound triosmium cluster at temperatures between 100° and 400°C under vacuum or an inert gas leads to a material with an IR spectrum similar to that observed for [Os(CO)212]127, which has a polymeric structure with iodide atoms bridging the Os(II) ions. The transition from a surface bound trinuclear cluster to ensembles of atomically dispersed osmium, can be perceived as happening during formation of 08-0 bonds with concurrent breaking of metal-metal bonds in the cluster framework1242127. Parallel behavior has been observed for RU3(CO)12 supported on silica or alumina1161117. In the present work we do not observe formation of surface structure (I) with ALPILC-supported 053(CO)12 or RU3(CO)12. Instead, ensembles of atomically dispersed metal atoms of the type [I\’I(CO)X(O-Al<)2]n are formed by oxidation in air, evacuation, or thermal decomposition under Ar. Surface adsorbed water might be involved in the degradation of surface structure (1) upon thermal treatment of the mineral. The aforementioned mononuclear carbonyl complexes are also obtained with ALPILC-supported lr4(CO)12, H2053(CO)10 and H4RU4(CO)12. In the suggested structure the metal atoms are stabilized on the surface by firm interaction with oxygen atoms in the support lattice. In addition, every metal atom is coordinated to two or more CO molecules. However, the value of n cannot be designated at present and it depends on the particular cluster as well as the specific treatment conditions. If thermolysis does not induce aggregation of the metal centers, then the value of n is expected to lie between 1 and the number of metal atoms present in the original cluster. Alternatively, if metal conglomerates are formed, then n might be much higher. The number and position of the IR bands in the CO region are dependent on the local symmetry, the oxidation state of the central atom and the extent 156 of metal aggregation. Complexes of the type [M(CO)3X2], with pseudo C3v symmetry, are expected to exhibit two IR carbonyl absorptions assigned to the A1 and doubly degenerate E modesl52. In addition structures containing two CO molecules per metal would also exhibit two major absorption bands associated with the symmetric and antisymmetric carbonyl stretching modes. Their position is determined by the number of CO molecules coordinated to the central atom. As the CO coverage on the surface decreases, the availability of metal d electrons for retroactive bonding increases. This leads to a strengthening of the M-C bond with concomitant weakening of the C-0 bond, which in turn shifts the frequency to lower wavenumbers. In a similar manner as metal aggregation becomes more extensive the availability of d electrons decreases due to the formation of metal-metal bonds, which increases the vibrational frequency of the C-0 bond. As the mineral is treated at higher temperatures a progressive decarbonylation is observed. This leads to coordinatively unsaturated metal atoms that are likely to interact more strongly with the oxygen atoms of the support to form surface aluminosilicates. The strength of the metal-oxygen bonds determines the degree of metal-support interaction, which evidently plays a significant role in the formation of metal aggregates or atomically dispersed metal atoms. 7. Adsorption of [CpFe(CO)2]2 on ALPILC Adsorption of [CpFe(CO)2)]2 from a CH2C12 solution (0.71 x 10'3 M) on alumina pillared montmorillonite (0.5 g) leads to a red-brown material which exhibits an IR spectrum significantly different from that of the unsupported cluster (Table 18). The new spectrum (Figure 38) shows CO bands at 2124 m, 2068 vs, and 2019 s cm‘l. The ability to extract the surface organometallic species into a solution of potassium hexafluorophosphate in 157 Table l 8 Infrared CO Bancb of Molecular and ALPILC Intercalated Metal Cluster Complexes Compound v(CO) cm"1 Reference [CpFe(c0)2,3]+/ALPILCa 2124 m, 2068 vs, 2019 s This Work [CpFe(CO)3][PF6]a * 2124 m, 2067 s This Work [Cpre(co)2(H20)][PF6]b* 2069 vs, 2023 vs This work [CpFe(CO)2]2b 1999 vs, 1958 s, 1777 vs This Work 1772 vs CpFe(CO)21b 2038 vs, 1999 vs This Work [CpFe(CO)3][PF6]a 2120, 2068 153 [CpFe(CO)2(H20)][BF4]C 2066 s, 2018 s 155 CpFe(CO)21d 2043 vs, 2005 vs 158 8 Nujol mull; b CH2C12; c Acetone; d CHCI3 * Isolated from the surface of ALPILC by exchange with KPF6 in acetone. Figure 38 158 Infrared spectra in the CO stretching region: (a) clay sample prepared by impregnation of alumina pillared clay with [CpFe(CO)2]2 in CH2C12 (mull); (b) [CpFe(CO)3]- PF6 extracted from the surface of ALPILC with a solution of KPF5 in acetone (Nujol mull); (c) [CpFe(CO)2]- PF6 extracted from the surface with KPF5 in acetone (CHQCIQ solution). 160 acetone suggests their ionic character. When the mineral is washed with the aforementioned solution it becomes almost colorless. Removal of the acetone solvent under vacuum and replacement by CHZCIZ produces a brown-red solution and an undissolved pale yellow material. The latter exhibits carbonyl absorptions at 2124 s and 2067 s cm‘1 and a C5H5 resonance at 6.09 ppm, in good agreement with the spectra reported for [CpFe(CO)3][PF6]153:154. The CH2C12 soluble compound is also a metal carbonyl complex with v(CO) at 2069 vs and 2023 vs cm‘l. Its 1H NMR spectrum shows a C5H5 resonance at 5.60 ppm. Furthermore, it reacts quickly with KI to form a new complex with 0(CO) bands shifted to lower wavenumbers (2038 vs, 1999 vs cm‘l) and a single resonance in the NMR at 5.05 ppm. On the basis of the above results it is characterized as [CpFe(CO)2][PF5]155. The CpFe(CO)2+ cation is known to react with halides to form CpFe(CO)2X type complexes. The final IR spectrum of the clay sample is the superposition of those of the two surface species. The Fe loading of the dry mineral is 0.57 wt9o. The surface chemistry of [CpFe(CO)2]2 adsorbed on alumina pillared clay is far more complex than that observed for RU3(CO)12 and 053(CO)12. Adsorption of the dodecacarbonyl clusters on the support leads to the formation of the protonated HM3(CO)12+ by intracrystal H+ ions. In the case of iron, the lower stability of the metal-metal bonds leads to fragmentation of the dimer to form mononuclear complexes. It has been known that cyclopentadienyl iron carbonyl compounds can abstract a proton from a strong acid to form a species containing metal-hydrogen bondsl56. However, these salts are not very stable even in the presence of the acid. In addition to the carbonyl bands of the protonated dimer, CO absorptions from CpFe(CO)3+ appear and increase in intensity 161 Table 1 9 C5H5 Proton Chemical Shifts for Some Cyclopentadienyliron Carbonyl Compounds Compound 5 ppm Reference [CpFe(co)3][PF6]8 * 6.09 This Work [CpFe(CO)3][PF5]a 6.14 154 [CpFe(CO)2(H20)][PF5]a* 5.60 This Work [CpFe(CO)2(H20)][BF4] 5.52 155 CpFe(CO)21b 5.05 This Work CpFe(CO)21b 5.04 158 a CD3COCD3; b CDC13 * Isolated from the surface of ALPILC by exchange with KPF5 in acetone. 162 with time157. In addition oxidation of the dimer in acetone and aqueous fluoroboric acid is known to form a reactive cationic intermediate identified as CpFe(CO)2+ (equation 25)155. This dicarbonyl complex undergoes substitution upon reaction with anionic (equation 26) or neutral ligands. [CpFe(CO)2]2 + 2HBF4 + H20 + %OZH2[CpFe(CO)2(OH2)]BF4 (25) [CpFe(CO)2(OH2)]BF4 + X’-?CpFe(CO)2X + BF4‘ + H20 (26) Adsorption of [CpFe(CO)2]2 on the support leads to the formation of both di and tricarbonyl clyclopentadienyl iron complexes. We believe these products are formed through different competing surface reactions. Reaction of the dimer with intracrystal protons results in the formation of the protonated complex which in turn decomposes to form the CpFe(CO)3+ ion according to equation 27. [CpFe(CO)2]2 + ALPILC:ICpFe(co)2]2H+/ALP1LC ——) CpFe(CO)3+/ALPILC (27) On the other hand, the CpFe(CO)?_+ complex might be produced through oxidation by molecular oxygen [CpFe(CO)2]2 + ALPILC + 02 -—)CpFe(CO)2(HzO)+/ALPILC (23) or surface hydroxyl groups 163 -;{CpFe(CO)2]2 + H-o-A1(—+ CpFe(CO)2(O-Al() + 35142 (29) The ability to extract the dicarbonyl complex from the surface by ion-exchange techniques suggests oxidation by 02 in the presence of the acidic support according to equation 28 as the possible pathway. In addition, reaction of the dimer with a clay sample that has not been evacuated results in an almost exclusive production of the CpFe(CO)2+ ion on the surface. The formation of the protonated complex that in turn decomposes to the CpFe(CO)3+ ion can be only accomplished in the presence of strong acidic media. However, the Bronsted acidity of the support decreases as the amount of surface water increases. The molecular oxygen is probably strongly adsorbed on the support and cannot be removed upon evacuation. Surface water is also present even after treatment under vacuum at 300°C, as evidenced by the H20 bending vibration frequency at about 1630 cm‘1 in the IR spectrum of the mineral. The ratio of the two surface products remains unchanged when shorter reaction times are employed. If the reaction is carried out in the presence of air a rusty material is obtained with no significant carbonyl absorption bands, likely iron oxide deposited on the exterior of the support. Exposure to air of the CpFe(CO)x+, (x = 2,3), supported complexes results in a progressive decarbonylation. If the clay sample is heated at 100°C in air its IR spectrum exhibits mainly absorptions due to the CpFe(CO)3+ ion, however of diminished intensity. The same results are obtained when the mineral is heated at 100°C in flowing H2. In this case decarbonylation is complete over a period of a few hours. Removal of the CO ligands by either thermolysis or oxidation generates highly uncoordinated Fe atoms on the surface that interact strongly with 164 lattice oxygens of the clay framework. In this case Fe(II) and/or Fe(III) surface aluminates are likely to be formed. D. Laponite-Supported Metal Cluster Carbonyls The selective decoration of the edge surfaces of layered silicates with metal carbonyl complexes through reaction with surface hydroxyl groups was briefly studied in the last part of this work. Such decoration with heavy elements may be useful for studies of clay particle morphology by electron microscopy or it may have important implications for optimizing the turnover frequencies of clay supported metal catalysts. Laponite was used as a support because of the abundance of edge surfaces due to the small size of clay platelets. 'l‘he impregnation was carried out in the absence of air and the solvent was removed under vacuum to eliminate decomposition of the cluster due to molecular 02. 1. Acborption of H2033(CO)10 and 053(CO)12 on Laponite In the absence of air the hydrido cluster H2053(CO)10 is easily physisorbed from CH2C12 solution on laponite. The IR spectrum of the reaction product is similar to that of the unsupported cluster (Figure 39, Table 20). Furthermore, the weakly adsorbed clusters can readily be extracted from the surface with CH2C12. The effects of heating H2053(CO)10 on laponite under vacuum and in Ar or air are illustrated in Figure 39. After heating in vacuum at 60°C, the spectrum of the crafted 053(H)(O-Si5)(CO)10 appeared accompanied by a drastic color change from red-violet to yellow. The surface species is most likely formed by the following reaction: H2053(CO)10 + H-o-Sis—)0s3(H)(o-Sia)(co)10 + H2 (30) Figure 39 165 Infrared spectra in the CO stretching region of H2053(CO)10 supported on laponite (0.003 mmol/0.1 g): (a) H2033(CO)10 physisorbed on laponite from CH2C12 solution (mull); (b) sample (a) after heating in vacuum at 60°C for 4 h; (c) followed by heating in vacuum at 150°C for 3 h; (d) sample (a) after heating under Ar at 150°C for 3 h; (e) sample (a) after heating in air at 100°C for 6 h. Table 20 Infrared CO Bands of Molecular and Laponite- Supported Metal Cluster Complexes Compound v(CO) cm"l 053(CO)12/laponitea 2068vs, 20355, 20165, 1997m, 1983m 053(CO)12b 2068vs, 20345, 2013m, [( ESi-O ) 205 ( CO)x]n- laponitea H2053(CO)10/laponitea H2053(CO)10b HOS3(CO)10(O-Si §)-laponitea RU3(CO)12/laponitea RU3(CO)12b [(ESi—O)2Ru(CO)2]n-laponitea H4Ru4(CO)12/laponitea H4RU4(CO)12b 2000m 2118m, 2021vs, 1937s 2114w, 20755, 20645, 2025vs, 20105, 1992m 2112vw, 20765, 20635, 2025vs, 20105, 1989m 2111w, 20745, 20615, 2023vs, 20095, 1990m 2059vs, 20305, 2011n1 2061vs, 2029s, 2011n1 20515, 19825 20825, 2065vs, 20245 20815, 2066vs, 20245, 2009m a Nujol mull b CH2C12 solution 168 When the sample is heated at higher temperatures or at 100°C in air, surface species of the type [(ESi-O)205(CO),,(]n are formed, characterized by IR absorptions at around 2120 m, 2020 vs, and 1940 cm’l. Adsorption of 053(CO)12 from a CH2C12 solution leads again to simple physisorption on the clay (Figure 40, Table 20). The clusters remain stable up to 100°C. When the clay sample was heated at higher temperatures we could not identify the formation of triangular clusters covalently attached to the support due to the complexity of the spectrum. At 150°C the spectrum of surface species [(ESi-O)2OS(CO)x]n, characterized by three IR absorptions, was observed. The formation of the surface-supported hydride cluster 053(H)(O-Si§)(CO)10 requires an oxidative addition of aESi-OH group to the 05—05 bond of the triosmium clusters. Such a transformation is easily achieved with H2053(CO)10 which is a more suitable precursor than 053(CO)12. In the latter case, 1055 of CO molecules is needed to form the trinuclear addition product. Heating at higher temperatures causes a breakdown of the cluster framework with simultaneous oxidation by surface-OH groups to two osmium(II) species of the type [(ESi-O)205(CO)x]n according to equation 9 (see Section Ill.A.1). The stabilization of the mononuclear complexes on the surface of silica or alumina has been attributed to an extremely strong osmium-oxygen interaction and the steric repulsion between carbonyl ligands of neighboring surface complexeslzg. Hence the formation of metal aggregates is avoided on heating. 2. Adsorption of RU3(CO)12 and H4Ru4(CO)12 on Laponite RU3(CO)12 and H4RU4(CO)12 are readily physisorbed unaltered on laponite from a CHZCIZ solution. The spectra of laponite—supported RU3(CO)12 and H4RU4(CO)12 are presented in Figures 41 and 42, respectively. The CO Figure 40: 169 Infrared spectra in the CO stretching region of 053(CO)12 supported on laponite (0.003 mmol/0.1 g): (a) 053(CO)12 physisorbed on laponite from CH2C12 solution (mull); (b) sample (a) after heating in vacuum at 150°C for 3 h; (c) sample (a) after heating under Ar at 150°C for 3 h; (d) sample (a) after heating in air at 150°C for 6 h. Fig. 41 Fig. 42 171 Infrared spectra in the CO stretching region of RU3(CO)12 supported on laponite (0.003 mmol/0.1 g): (a) RU3(CO)12 on laponite prepared by impregnation from CHZCIZ solution (mull); (b) sample (a) after heating in vacuum at 60°C for 4 h; (c) sample (a) exposed to air for seven days; ((1) sample (a) after heating in air at 90°C for 30 min. Infrared spectra in the CO stretching region: (a) H4RU4(CO)12 on laponite (0.003 mmol/0.1 g) prepared by impregnation from CHzClg solution (mull); (b) sample (a) after heating in vacuum at 60°C for 6 h; (c) followed by heating at 120°C for 3 h; (d) sample (a) after heating in air at 90°C for 1 h. . 172 J" J“ W 1 2200 2000 1900 2200 2000 1900 CM" CM" Figure 41 Figure 42 RL 173 stretching frequencies along with those of molecular clusters are summarized in Table 20. Heating the two clay samples in air or under vacuum leads to the formation of the surface structure [(ESi-O)2Ru(CO)2]n, characterized by IR absorptions at 2050 and 1980 cm“1 and independent of the original cluster composition. The observed higher stability of H4RU4(CO)12 is attributed to its higher nuclearity115. The formation of the mononuclear surface species most likely involves reaction with —OH groups according to equation 10 (see Section III.A.2). 1£' 10. 11. 12. 13. 14. 15. 16. REFERENCES R.C. Mackenzie, Clays Clay Miner., 1_1, 11 (1963). G. Brown, Ed., X-Ray Identification and Crystal Structure of Clay Minerals, Mineralogical Society, London, 1961. G.W. Brindley and G. Brown, Eds. Crystal Structures of Clay Minerals and Their X-Ray Identification, Mineralogical Society, London, 1980. U. Hoffman, Angew. Chem. Int. Ed., _7_, 681 (1968). J.B. Dixon, in "Minerals in Soil Environments", J.B. Dixon and 8.8. Weed, Eds., Soil Science Society of America, Madison, WI, 1977, Chapter 11. J.D. Russell and A.R. Frazer, Clays Clay Miner., 1_9, 55 (1971). T.J. Pinnavaia, Science, m, 365 (1983). S.L. Swartzen-Allen and E. Matijec, Chem. Rev., _7_4_, 385 (1974). E.T. Uskova, N.G. Vasilev, and LA. Uskov, Colloid J. (USSR), §_Q, 118 (1968). H. Van Olphen, "An Introduction to Clay Colloid Chemistry", 2nd Ed. John Wiley, NY, 1977. K. Norrish, Disc. Faraday Soc., l_8_, 120 (1954). G.W. Brindley, K. Wiewiora and A. Wiewiora, Amer. Miner., 21, 1635 (1969) K.V. Raman and M.M. Mortland, Proc. Soil Sci. Soc. Am., Q, 313 (1969). V. Berkheiser and M.M. Mortland, Clays Clay Miner., 25, 105 (1977). K.V. Raman and M.M. Mortland, ibid, E, 393 (1968). J.J. Fripiat, ACS Symp. Ser., _3_4, 261 (1976). 174 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 175 L.B. Ryland, M.W. Tamele and J.N. Wilson in "Catalysis", Vol. VII, P.H. Emmett, Ed., Reinhold, NY, 1960. R.E. Grim, "Clay Mineralogy", 2nd Ed., McGraw-Hill, NY, 1968. B.K.G. Theng, "The Chemistry of Clay-Organic Reactions", John Wiley, NY, 1974. D.H. Solomon and D.G. Hawthorne, "Chemistry of Pigments and Fillers", John Wiley, NY, 1983. T.J. Pinnavaia in "Advanced Techniques for Clay Mineral Analysis", J.J. Fripiat, Ed., Elsevier, NY, 1981, pp. 139-161. W.E.E. Stone, ibid., pp. 71-112. M.B. McBride in "Advanced Chemical Methods for Soil and Clay Mineral Research", J.W. Stucki and W.L. Banwart, Eds., Reidel, Holland, 1980, pp.423-450. R.M. Barrer and D.N. MacLeod, Trans. Faraday Soc., _5_1_, 1290 (1955). R.M. Barrer and D.L. Jones, J. Chem. Soc. (A), 2594 (1971). M.M. Mortland and V.E. Berkheiser, Clays Clay Miner., _2_4_, 60 (1976). J. Shabtai, N. Frydman and R. Lazar, Proc. 6th Int. Congr. Catal., 2, 660 (1977). V.E. Berkheiser and M.M. Mortland, Clays Clay Miner., _2_5_, 105 (1977). M.P. Traynor, M.M. Mortland and T.J. Pinnavaia, ibid., E, 318 (1978). G.W. Brindley and R.E. Sempels, Clay Miner., l_2_, 229 (1977). S. Yamanaka and G.W. Brindley, Clays Clay Miner., _2_§_, 21 (1978). N. Lahav, U. Shani and J. Shabtai, ibid., 2_6_, 107 (1978). D.E.W. Vaughan and R.J. Lussier, Preprints, 5th Int. Conf. Zeolites, Naples, Italy, June 2—6, 1980. J. Shabtai, Chim. L'Indust., 6_1_, 734 (1979). J. Shabtai and N. Lahav, U.S. Patent 4,216,188. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 176 J. Shabtai, U.S. Patent 4,238,364. D.E.W. Vaughan, R.J. Lussier and J.S. Magee, U.S. Patents, 4,176,090; 4,271,043; 4,248,739. M.L. Occelli and R.M. Tindwa, Clays Clay Miner., _3_1_, 22 (1983). T.J. Pinnavaia, M.S. Tzou, S.D. Landau and R.H. Raythatha, J. Mol. Catal., _2_7_, 195 (1984). M.L. Occelli, S.D. Landau and T.J. Pinnavaia, J. Catal., _9_Q, 256 (1984). J.Y. Bottero, J.M. Cases, F. Flessinger and J.B. Poirier, J. Phys. Chem., fl, 2933 (1980). J.C. Bailar, Catal. Rev.-Sci. Eng., _1_0_, 17 (1974). J.M. Basset and A.K. Smith in "Fundamental Research in Homogeneous Catalysis", M. Tsutsui and R. Ugo, Eds., Plenum Press, New York, 1977, Vol. 1. S.L. Davydova and N.A. Plate, Coord. Chem. Rev., 1_6, 195 (1975). R.H. Grubbs, Chemtech, _7_, 512 (1977). F.R. Hartley and P.N. Vezey, Adv. Organomet. Chem., 1_5_, 189 (1977). Z.M. Michalska and D.E. Webster, Plat. Met. Rev., _l_8_, 65 (1974). C.U. Pittman, Jr., and G.O. Evans, Chemtech, 3, 560 (1973). A.L. Robinson, Science, _l_fl, 1261 (1976). Y.I. Yermakov, Catal. Rev.-Sci. Eng., _1_3_, 77 (1976). G.A. Somorjai, "Chemistry in Two Dimension:Surfaces", Cornell University Press, 1981. W.A. Adamson, "Physical Chemistry of Surfaces", 3rd Ed., John Wiley, New York, 1976. T.N. Rhodin and G. Ertl, Eds., "The Nature of the Surface Chemical Bond", North-Holland Publication Co., New York, 1979. D.C. Bailey and S.H. Langer, Chem. Rev., _§_l_, 109 (1981). 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 177 J. Manassen, Plat. Met. Rev., _1_5_, 142 (1971). J. Manassen and D.D. Whitehurst in "Progress in Catalysis", F. Basolo and R.L. Burwell, Jr., Eds., Plenum Press, New York, 1973. Z.M. Michalska and D.E. Webster, Chemtech, _5, 117 (1975). J.D. Corbett, Prog. Inorg. Chem., 2_1_, 129 (1976). R.L. Pruett, Adv. Organomet. Chem., 1_7_, l (1979). B.F.G. Johnson, Plat. Met. Rev., _2_2_, 47 (1978). A.P. Humphries and H.D. Kaesz, Prog. Inorg. Chem., g, 146 (1978). P. Chini, G. Longoni and V.G. Albano, Adv. Organomet. Chem., 1_4, 285 (1976). R. Mason and A.I.B. Rae, J. Chem. Soc. (A), 778 (1968). F.R. Corey and L.F. Dahl, Inorg. Chem., _1_, 521 (1962). J. Knight and M.J. Mays, J. Chem. Soc. (A), 711 (1970). H.D. Kaesz, S.A.R. Knox, J. W. Koepke and R.B. Saillant, Chem. Commun., 477 (1971). A.J. Deeming and S. Hasso, J. Organomet. Chem., _1_l_4, 313 (1976). G.R. Wilkes, Ph.D. Thesis, The University of Wisconsin, Madison, WI, 1965. R.D. Wilson, S.M. Wu, R.A. Love and R. Bau, Inorg. Chem., _1_'_7_, 1271 (1978). B.F.G. Johnson and J. Lewis in "Advances in Inorganic Chemistry and Radiochemistry", H.J. Emeleus and A.G. Sharpe, Eds., Academic Press, New York, 1981, Vol. 24. G.F. Stuntz, Ph.D. Thesis, Univeristy of Illinois, Urbana-Champaign, IL, 1978. W.R. Cullen and O.A. Harbourne, Inorg. Chem., 9, 1839 (1970). E.L. Muetterties, T.N. Rhodin, E. Band, C.F. Brucker and W.R. Pretzer, Chem. Rev., _7_9_, 91 (1979). 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 178 E.L. Muetterties, Science, .136, 839 (1977). A.K. Smith and J.M. Basset, J. Mol. Catal., _2_, 229 (1977). M. Moskovits, Accts. Chem. Res., _12, 229 (1979). R. Ugo, Catal. Rev.-Sci. Eng., l_1_, 225 (1975). R. Whyman in "Transition Metal Clusters", B.F.G. Johnson, Ed., John Wiley, New York, 1980, Chapter 8. R.M. Laine, J. Mol. Catal., l_4_, 137 (1982). J.H. Sinfelt, Prog. Solid State Chem., 1.9.1 55 (1975). G.W. Parshall, J. Mol. Catal., 4, 243 (1978). J. Falbe, J. Organometal. Chem., 9_4_, 213 (1975). J.F. Roth, J.H. Craddock, A. Hershman, and RE. Paulik, Chem. Technol., _1_, 600 (1971). M.W. Farlow, U.S. Patent 2,518,608. a) BC. Gates and J. Lieto, Chemtech, 195 (1980); b) B.C. Gates and J. Lieto, Chemtech, 248 (1980). K. Iwatate, S.R. Dasgupta, R.L. Schneider, G.C. Smith and K.L. Watters, Inorg. Chim. Acta, 1_5_, 191 (1975). J.R. Anderson, "Structure of Metallic Catalysts", Academic Press, New York, 1975, p. 275. A. Brenner and R.L. Burwell, Jr., J. Catal., _5_2_, 353 (1978). E. Mantovani, N. Palladino and A. Zanobi, J. Mol. Catal., _3_, 285 (1978). K.J. Klabunde, D. Ralston, R. Zoellner, H. Hattori and Y. Tanaka, J. Catal., g, 213 (1978). S.F.A. Kettle in "Topics in Current Chemistry", Springer Verlag, New York, 1977, Vol. 77, p. 111. Y. Nonaka, S. Takahashi and N. Hagihara, Mem. Inst. Sci. Ind. Res. Osaka University, 31, 23 (1974). 93. 94. 95. 96. 97. 98. 99. 100 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 179 0.0. Evans, C.U. Pittman, Jr., R. McMillan, R.T. Beach and R. Jones, J. Organomet. Chem., 6_7_, 295 (1974). L.L. Murrell in "Advanced Materials in Catalysis", J.L. Burton, R.L. Garten, Eds., Academic Press, New York, 1977, p. 235. K.G. Allum, R.D. Hancock, I.V. Howell, S. McKenzie, R.C. Pitkethly and P.J. Robinson, J. Organomet. Chem., g, 203 (1975). I. Haller, J. Am. Chem. Soc., _1_(_)_(_)_, 8050 (1978). F.R. Wild, G. Gubitosa and H.H. Brintzinger, J. Organomet. Chem., _1__4_§, 73 (1978). F. Hugues, J.M. Basset, Y.B. Taarit, A. Choplin, M. Primet, D. Rojas and A.K. Smith, J. Am. Chem. Soc., 1_0_4_, 7020 (1982). J.R. Budge, J.F. Scott and B.C. Gates, J. Chem. Soc. Chem. Commun., 342 (1983). T.J. Pinnavaia, R. Raythatha, J.C.-S. Lee, L.J. Halloran and J.F. Hoffman, J. Am. Chem. Soc.,_1_0_l_, 6891 (1979). R. Raythatha and T.J. Pinnavaia, J. Catal., 82, 47 (1983). . W.H. Quayle and T.J. Pinnavaia, Inorg. Chem., _l_8_, 2840 (1979). F. Farzaneh and T.J. Pinnavaia, Inorg. Chem., _2_2, 2216 (1983). Z. Otero-Schipper, J. Lieto and B.C. Gates, J. Catal., _63, 175 (1980). H.B. Gray, A. Rembaum and A. Gupta, U.S. Patent 4,127,506. G. Braca, C. Carlini, F. Ciardelli and G. Sbrana, Preprints, 6th International Congress on Catalysis, London, 1976, paper A43. M.B. Freeman, M.A. Patrick and B.C. Gates, J. Catal., E, 82 (1982). J.B.N. Effa, J. Lieto and J.F. Aune, J. Mol. Catal., _1_5_, 367 (1982). J. Lieto, J.J. Rafalko and B.C. Gates, J. Catal., _6_2, 149 (1980). P. Gallezot, G. Coudurier, M. Primet and B. Imelic, Am. Chem. Soc., Symp. Ser., 144 (1977). 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 180 J.C. Goodwin, Jr., and C. Naccache, J. Mol. Catal., _1_4, 259 (1982). D. Commereuc, Y. Chauvin, F. Hugues, J.M. Basset and D. Olivier, J. Chem. Soc., Chem. Commun., 154 (1980). J. Robertson and G. Webb, Proc. R. Soc. London, Ser. A., E, 383 (1974) A.K. Smith, A. Theolier, J.M. Basset, R. Ugo, D. Commereuc and Y. Chauvin, J. Am. Chem. Soc., _1_(1), 2590 (1978). V.L. Kuznetsov, A.T. Bell and Y.I. Yermakov, J. Catal., _6_5, 374 (1980). A. Theolier, A. Choplin, L. D'Ornelas, J.M. Basset, R. Ugo, R. Psaro, G. Zanderighi and C. Sourisseau, Polyhedron, _2, 119 (1983). J. Evans and G.S. McNulty, J. Chem. Soc., Dalton Trans., 1123 (1984). D.K. Liu and M.S. Wrighton, J. Am. Chem. Soc., 1_0_4_, 898 (1982). Y. Doi, K. Soga, K. Yano and Y. Ono, J..Mol. Catal., _1_9_, 359 (1983). S.C. Brown and J. Evans, J. Mol. Catal., _1_1_, 143 (1981). J.C. Bentsen and M.S. Wrighton, Inorg. Chem., 23, 515 (1984). A. Brenner and D.A. Hucul, J. Am. Chem. Soc., _1_0_2_, 2484 (1980). J.B. Crawford, G.A. Melson, L.B. Makovsky and F.R. Brown, J. Catal., 83,454(1983) M. Deeba and B.C. Gates, J. Catal., 67, 303 (1981). R. Barth, B.C. Gates, Y. Zhao, H. Knozinger and J. Hulse, J. Catal., 8g,147(1983) X.-J. Li and B.C. Gates, J. Catal., 8_4_, 55 (1983). R. Psaro, R. Ugo, G.M. Zanderighi, B.Besson, A.K. Smith and J.M. Basset, J. Organomet. Chem., 1L3, 215 (1981). H. Knozinger and Y. Zhao, J. Catal., _7_1_, 337 (1981). T. Castrillo, H. Knozinger and M. Wolf, J. Mol. Catal., _1_1, 151 (1981). K. Foger and J.R. Anderson, J. Catal., 5_9, 325 (1979). 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 181 M. Ichikawa, Bull. Chem. Soc. Japan., _5_l_, 2268 (1978). R.F. Howe, J. Catal., El, 196, (1977). J.R. Anderson, P.S. Elmes, R.F. Howe, and D.E. Mainwaring, J. Catal., _5_0_, 508 (1977). K. Tanaka, K.L. Watters and R.F. Howe, J. Catal., _7_3, 23 (1982). C. Ercolani, J.V. Quagliano and L.M. Vallarino, Inorg. Chim. Acta, 3, 421 (1969). F. Piacenti, M. Bianchi, E. Benedetti and G. Braca, Inorg. Chem., _7_, 1815(1968L B.F.G. Johnson, J. Lewis and D.A. Pippard, J. Chem. Soc., Dalton Trans., 407 (1981). R.F. Howe, D.E. Davidson and D.A. Whan, J. Chem. Soc., Faraday Trans., 33, 2266 (1972). A. Kadkhodayan, Ph.D. Thesis, Michigan State University, East Lansing, MI, 1984. F. Piacenti, M. Bianchi, P. Frediani and E. Benedetti, Inorg. Chem., _1_()_, 2759 (1971). F. Farzaneh, Ph.D. Thesis, Michigan State University, East Lansing, MI 1981. G.A. Borchardt in "Minerals in Soil Environments", J.B. Dixon and 8.8. Weed, Eds., Soil Science Society of America, Madison, WI, 1977, Chapter 9. A.J. Lees and A.W. Adamson, J. Am. Chem. Soc., _1_0_4_, 3804 (1982). A.A. Davydov and A.T. Bell, J. Catal., :13, 332 (1977). N. Sheppard and T.T. Nguyen, Adv. Infrared Raman Spectrosc., 3, 67 (1978). 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 182 M.J. Cleare and W.P. Griffith, J. Chem. Soc. (A), 372 (1969). A.J. Deeming, B.F.G. Johnson and J. Lewis, J. Chem. Soc. (A), 898 (1970) J.S. Kristoff and D.F. Shriver, Inorg. Chem., _l_3_, 499 (1974). J.A. Helsen, R. Drieskens and J. Chaussidon, Clays Clay Miner., _2_3_, 334 (1975). B. Delley, M.C. Manning, D.E. Ellis, J. Berkowitz and N.C. Trogler, Inorg. Chem., _2_l_-, 2247 (1982). E.G. Bryan, G. Jackson, B.F.G. Johnson, J.W. Kelland, J. Lewis and K.J. Schorpp, J. Organomet. Chem., 1_0_8_, 385 (1976). P.S. Bratterman, "Metal Carbonyl Spectra", Academic Press, London, New York, San Francisco, 1975. L. Busetto and R.J. Angelici, Inorg. Chim. Acta., _2_, 391 (1968). PM. Treichel, R.L. Shubkin, K.W. Barnett and D. Reichard, Inorg. Chem., _5_, 1177 (1966). B.D. Dombek and R.J. Angelici, Inorg. Chim. Acta, 1, 345 (1972). A. Davison, W. McFarlane, L. Pratt and G. Wilkinson, J. Chem. Soc. (A), 3653 (1962). D.A. Symon and T.C. Waddington, ibid., 954 (1971). T.E. Sloan and A. Wojcicki, Inorg. Chem., _7_, 1268 (1968). W.H. Quayle, Ph.D. Thesis, Michigan State University, East Lansing, MI, 1978. M. Green, T.A. Kuc and S.H. Taylor, J. Chem. Soc. (A), 2334 (1971). R.H. Raythatha, Ph.D. Thesis, Michigan State University, East Lansing, MI,1981. mWWWWWWWWI)Iliiiijiii