MSU LIBRARIES .a—___ RETURNING MATERIALS: P1ace in book drab to remove this checkout from your record. FINES will be charged if book is returned after the date stamped be10w. \// INTERCALATION OF RHODIUM COMPLEX HYDROGENATION CATALYSTS AND ORGANO-SILANES IN LAYERED SILICATES ‘ By Rasik Haridas Raythatha ,A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1981 ABSTRACT INTERCALATION OF RHODIUM COMPLEX HYDROGENATION CATALYSTS AND ORGANO-SILANES IN LAYERED SILICATES By Rasik Haridas Raythatha Catalytically active rhodium complexes have been inter- calated in interlayer regions of hectorite, a swelling mica- type layered silicate. The caLionic hydrogenation catalyst precursor, [Rh(NBD)- (dppe)]+ (NBD = norbornadiene, dppe = l,2-bis(diphenyl- phosphino)ethane) has been used for the hydrogenation of 1,3-butadiene and two of its methylated derivatives, 2- methyl-l,3-butadiene and 2,3-dimethyl-l,3—butadiene under intercalated and homogeneous conditions in methanol, acetone and benzene medium. The rates of hydrogen addition are solvent dependent, and the solvation effect on hydrogen addition is especially pronounced with the intercalated catalyst relative to the homogeneous catalysts. The rates of hydrogen addition with the intercalated catalyst range 5 from 10- to 0.83 relative to the homogeneous catalyst. With the homogeneous catalysts, the product distribution Rasik Haridas Raythatha is essentially independent of solvent. The thermodynamically stable l,A—hydrogen addition products are the major pro- ducts with homogeneous catalysts, whereas a significant enhancement in the synthetically important 1,2- addition products is observed with the intercalated catalysts. The yields of 1,2-addition products are consistently higher in methanol than in acetone, where the interlayers are more constricted. The observed effect of catalyst intercala- tion on rates of reduction and product distribution may be linked to the swelling properties of [Rh(NBD)(dppe)j+. The hydrogenation catalyst precursors of the types RhL; (L = triphenylphosphine, n = 2,3) also have been inter- calated in hectorite by the reaction of triphenylphosphine with intercalated Rh2(CH3COZ)fi:x (x = 1,2) or by using [Rh(NBD)(PPh3)2]+ ions as a precursor. The reduction of l-hexene with the intercalated complexes occurs without isomerization up to 60% conversion of substrate, whereas extensive isomerization to internal olefin is observed with the analogous catalyst system in homogeneous solution. The difference in specificity between the intercalated and homogeneous catalyst system is accounted for in terms of an equilibrium between catalytically active dihydride and monohydride rhodium complexes: RhH2L; I RhHLn + H+. The dihydride is a good hydrogenation catalyst but a poor iso- merization catalyst, whereas the monohydride is both a good hydrogenation and isomerization catalyst. Relative to Rasik Haridas Raythatha homogeneous solution, the dihydride is favored in the inter- calated state because of a surface Bronsted acidity that is believed to arise from the dissociation of hydrated Na+ ions that are also present on the interlamellar sur- faces. However, the surface Bronsted acidity is dependent on surface composition. The effect of surface composition is demonstrated in part by the dependence of product dis- tribution upon the initial composition of substrate and the amount of water present in the solvent. Rh(PPh3) complexes are also catalysts for the reduction + n of alkynes to the corresponding gig-olefins. The initial rates of reduction of relatively small alkynes (l-hexyne, 2-hexyne) in the interlayers swelled with methanol are com— parable to those observed with the homogeneous catalyst. With larger alkynes, such as diphenylacetylene, the inter- calated rate may be one hundredth that of the homogeneous rates.’ The spatial requirements of the substrate in the swelled interlayers are important in determining its re- activity with the intercalated catalyst. For example, the ratio of intercalated to homogeneous rates for the reduc- tion of 2-decyne are 0.85 and 0.02, respectively, with CH2C12 and C6H6 as the swelling solvent. A binding model is proposed for the intercalated substrate-catalyst complex in which the spatial requirements of the substrate are determined by the minimum distance it must span when the coordinated CEC bond is oriented perpendicular to the silicate sheets. Rasik Haridas Raythatha The intercalation of silanes containing chloro, amino, vinyl and epoxide functional groups in the interlayers of Na—montmorillonite leads to the swelling of the interlayer region in the range of 2.8 — 5.9 A. TheSe expanded phases are stable at least to 200°C for the chloro silane inter- calates and in the range of 2A0-330°C for silanes contain- ing amino, vinyl and epoxide functional groups. Qualitative studies indicate cation exchange capacities of these ma— terials to be comparable to those of Na-montmorillonite. TO MY FAMILY Mom, Dad Sulu, Bholen, Raju, Babu Anju and Anilbhai ii ACKNOWLEDGMENTS I am deeply grateful to Dr. Thomas J. Pinnavaia for his guidance, wisdom and encouragement. I have no words to express the respect and admiration I hold for him as an inspiring research preceptor, a patient teacher and above all, as a person whose deep understanding, amiable nature and affectionate personality has made learning experience extremely fulfilling. I look forward to a continual associa- tion with him in the future. I am grateful to Dr. Carl H. Brubaker, Jr., for presid— ing as my second reader and for the editorial assistance in improving my writing skill. I am also thankful to Dr. Max M. Mortland for providing special research facilities and helpful discussiOns. Early advice extended by Dr. William Quayle and collaboration of Dr. John Lee is ap- preciated. Special acknowledgment is due for the Chemistry Depart- ment and National Science Foundation for excellent research facilities and financial assistance. It is difficult to acknowledge every person who has helped directly or indirectly in this long, arduous en- deavor, however, I am thankful to all present and past colleagues and friends, Dr. Chris Marshall, Dr. Elene Brown, iii Mr. Steve Christiano, Dr. Richard Barr, Mr. Emmanuel Gian— nelis, Miss Ivy Johnson and Devraj and Urmila Sahu. My brothers, Bholen and Raju, and sister Chandrika have been continual sources of affection, inspiration and encouragement and have guided me on all paths of my life. I have no words to express the love and inspiration pro- vided by my brother Dr. Surendra H. Raythatha and amiable Nancy Raythatha. Finally, words are futile to express the sacrifices made by my parents, Haridas and Bhanumati Raythatha and the love they provided all these years. iv Chapter TABLE OF CONTENTS LIST OF TABLES. LIST OF FIGURES INTRODUCTION. A. Foundation of Homogeneous and Heterogeneous Catalysts B. Supported Metal Complexes and Recent Developments . . . . . . C. The Properties of Clay Minerals and Their Structural Aspects. D. Cationic Rhodium Complexes and Rational for Their Use as .Hydrogena— tion Catalysts. . . . . E. Research Objectives 'F. Pillaring Reactions of Layered Silicates . . . . . EXPERIMENTAL. A. Materials B. Syntheses . . [Rh(NBD)Cl]2. . . . . . . . . . Rh(NBD)acac . . [Rh(NBD)(Diphos)]Clou [Rh(NBD)(PPh3)2]PF6 Rh2(CH3COZ)u. . . . . . . . . Protonation of Rhodium Acetate. . [Rh(NBD)(Diphos)]+-Hectorite. Nmmrwmi—J O Page viii xii 18 2A 26 28 32 32 35 35 35 36 37 37 38 39 Chapter RESULTS A. [Rh(NBD) (PPh? ) QJ +—Hectorite. . Rh(CH3 CO 2)“; -Hectorite lO. Catalyst Precursor from Protonated (II) Acetate. . Hydrogenation Reaction. Pillaring Reaction of Layered Silicate with Organosilanes. . . . l. Pretreatment of Na+- Montmorillonite . . . . 2. Silane/Na+-Montmorillonite. Physical Methods. . . . . . . . . . . 1. X-ray Diffraction Study 2. Infrared Spectroscopy . 3. Magnetic Resonance Spectroscopy A. UV-Visible Spectroscopy . Q 5. Thermal Analysis... . . . . . . . . 6. Gas Chromatography. . . . . . AND DISCUSSION. Hydrogenation of 1,3-Butadienes l. Homogeneous Hydrogenation 2. Intercalated Catalysts. . . . . Hydrogenation of Monoolefins. Olefin Reduction with [Rh(NBD)(PPh3)2]+ Catalyst Precursor. . . l. The Effect of Water Content of Methanol on the hydrogenation of l- Hexene with [Rh(NBD)(PPh3)2]+ Catalyst Precursor. . . . . 2. Hydrogenation of l— Hexene with [Rh(NBD)(Diphos)]+ Catalyst Precursor . . . . . . . . Alkyne Hydrogenation. . . . . . . . . vi Page 39 39 A0 Al AA AA “5 45 A5 A6 47 A7 A8 A8 “9 5A 5A 60 73 85 96 107 12“ Chapter E. Properties of Silane - Montmorillonite. CONCLUSIONS APPENDIX. BIBLIOGRAPHY. vii Page 13U 1A7 150 152 Table LIST OF TABLES Page Materials Used as Homogeneous Catalyst Support . . . . . . . . . . . . . . . . . . 7 Selected Methods of Binding Metal Complex Catalyst to Supports, and Their Application in Chemical Reactions . . . . . 8 UV—Visible Spectroscopy Data for Cationic Rhodium Complexes and Na-Hectorite Par- tially Exchanged with These Complexes . . . 51 001 X-ray Basal Spacings of Na+—Hectorite and Na+-Hectorite Exchanged with [Rh(NBD)(Diphos)]+. . . . . . . . . . . . . 55 Homogeneous Hydrogenation of Butadienes with Rh(NBD)(Dppe)+ as Catalyst Pre- cursor. O O O 0 O O O O O I I I O O O O O O 56 Hydrogenation of Butadienes with [Rh(NBD)- (Dppe)]+ Intercalated in Hectorite. . . . . 61 Relative Rates and Selectivity Towards 1,2 Addition Products . . . . . . . . . . . 66 Hydrogenation of 2,3—dimethyl-l-butene and 2,3-dimethyl-2-butene in Methanol viii Table 10 ll 12 13 1A Under Homogeneous and Intercalated Condi- tions with [Rh(NBD)(Diphos)]+-Catalyst Precursor . . . ." 0 Catalytic Hydrogenation of l-Hexene in Methanol with Intercalated and Homogen- eous Rh(PPh3); Complexes. . . . . . . . Catalytic Hydrogenation of l—Hexene in Methanol (0.2% H20), [Rh(NBD)(PPh3)2]+- Hectorite as Catalyst Precursor Catalytic Hydrogenation of l-Hexene in Methanol (0.2 wt % Water) Under Homogeneous Conditions with [Rh(NBD)- (PPh3)2]+ Catalyst Precursor. . Catalytic Hydrogenation of l-Hexene, 0.8 M, Under Intercalated and Homogeneous Conditions with [Rh(NBD)(PPh3)2]+ Catalyst Precursor in Methanol Containing Varying Amounts of Water. . . . . . . . Catalytic Hydrogenation of l-Hexene with Intercalated and Homogeneous [Rh(NBD)- (PPh3)2]+-Catalyst Precursor. Catalytic Hydrogenation of l—Hexene in Methanol (0.2 wt % H2O) Under Homogeneous and Intercalated Conditions with [Rh(NBD)- (Diphos)]+-Catalyst Precursor ix Page 69 8O 87 95 99 105 109 Table 15 l6 l7 18 19 20 Page Product Distribution for the Hydrogena- tion of l-Hexene with Homogeneous and Intercalated [Rh(NBD)(PPh3)2]+- Catalyst Precursor in Methanol (0.2 wt % H20). . . . . . . . . . . . . . . . . . . 117 Product Distribution for the Hydrogena- tion of l-Hexene with Homogeneous and Intercalated [Rh(NBD)(PPh3)2]+-Cata1yst Precursor in Methanol Containing Varying Amounts of Water. . . . . . . . . . . . . . 120 Hydrogen Uptake Rates for Reduction of l-, 2-, and 3-Hexyne with Intercalated and Homogeneous Rhodium Triphenylphos- phine Catalyst Precursors . . . . . . .,. . 125' Initial Hydrogen Uptake Rates for Re— duction of Alkynes in Methanol with Intercalated and Homogeneous Rhodium Triphenylphosphine Catalyst Pre- cursors . . . . . . . . . . . . . . . . . . 130 Initial Rates of Hydrogen Uptake for Reduction of 2-Decyne to gig-Z-Decene in Different Solvents with Intercalated and Homogeneous Catalysts . . . . . . . . . . . 135 Critical Dimensions of Selected Sub- strates . . . . . . . . . . . . . . . . . . 136 Table Page 21 001 Basal Spacing of Na+/Silane- Montmorillonite . . . . . . . . . . . . . . 138 22 Infrared Spectroscopy Study of Silane— Montmorillonite . . . . . . . . . . . . . . 1A1 23 Enthalpic Transition with Temperature for Na+/Silane-Montmorillonite. . . . . . . 1A2 xi Figure LIST OF FIGURES Schematic representation of the struc- ture of smectite. Reaction flask used to minimize creep- ing of finely divided intercalated catalyst. Schematic of hydrogenation appara-- tus . . . . . . . . . . . . . . . Infrared spectra of Na-Hectorite, [Rh- (NBD)(Diphos)]+-hectorite and [Rh(NBD)- (Diphos)] C10“ in the region 1800-1200_ -1 cm . . . . . . . X—ray diffraction patterns for Na+- hectorite partially exchanged with [Rh(NBD)(Diphos)]+ cations in different solvation conditions. (A) Hydrogen uptake plot for the re- duction of isoprene in acetone and methanol under homogeneous conditions with [Rh(NBD)(Diphos)]+ catalyst precursor . xii Page 19 A2 “3 52 53 58 Figure 10 11 l2 13 Page (B) Plot of percent composition with time for the reduction of isoprene in acetone under homogeneous conditions with [Rh(NBD)(Diphos)J+ complex . . . . . . 58 Typical percent composition plots for the hydrogenation of isoprene with [Rh- (NBD)(Diphos)]+-hectorite (A) in acetone medium (B) in methanol medium . . . . . . . 63 Mechanism proposed in literature for the catalytic hydrogenation of unsaturated hydrocarbons with rhodium complexes . . . . 71 Probable structure of the intermediate [RhH(dppe)]+. . . . . . . . . . . . . . . . 72 Infrared spectra (1800-1200 cm'l)of methanol-solvated hectorites. (A) Na++heCtorite. (B) Rh2(oAc)fi:x-hec- torite (C) Rh(PPhB):-hectorite. . . . . . . 77 The hydrogenation of l-hexene (A) 0.1 g, (B) o.u g with [Rh(NBD)(PPh3)2]+- hectorite in methanol (0.2 wt % H20). . . . 89 The hydrogenation of 1-hexene (A) 0.6 M (B) 0.7 M with [Rh(NBD)(PPh3)2]+-hectorite in methanol (0.2 wt % H2O). . . . . . . . . 90 The hydrogenation of l-hexene (A) 0.8 M, (B) 1.0 M, with [Rh(NBD)(PPh3)2]+— hectorite in methanol (0.2 wt % H20). . . . 91 xiii Figure 1A 15 16 17 18 19 Page (A) The hydrogenation of l-hexene, 1.2 M in methanol (0.2 wt % H2O) with [Rh(NBD)(PPh3)2]+-hectorite. (B) The hydrogenation of l-hexene, 0.1 M, under homogeneous conditions with [Rh(NBD)- (PPh3)2]+ in methanol (0.2 wt % H20). . . . 92 The hydrogenation of l-hexene (A) 0.6 M, (B) 1.0 M, under homogeneous conditions with [Rh(NBD)(PPh3)21+ in methanol (0.2 wt % H20). . . . . . . . . . . . . . . 97 The hydrogenation of l-hexene, 0.8 M, with [Rh(NBD)(PPh3)2]+-hectorite in methanol containing (A) 0.1 wt % H20 (B) 0.5 wt % H20. . . . . . . . . . . .,. .. 101 The hydrogenation of l-hexene, 0.8 M, with [Rh(NBD)(PPh3)2]+-hectorite in methanol containing 1.0 wt % water. . . . . 102 The hydrogenation of l-hexene, 0.8 M, with [Rh(NBD)(PPh3)2]+ complex under homogeneous conditions in methanol con- taining (A) 0.5 wt % H2O. (B) 1.0 wt % H20. . . . . . . . . . . . . . . . . . 104 The hydrogenation of l-hexene, (A) 0.“ M, (B) 0.8 M, under homogeneous conditions with [Rh(NBD)(Diphos)]+ xiv Figure Page complex in methanol, (0.2 wt % water). . . . . . . . . . . . . . . . . . . 111 20 (A) The hydrogenation of 1-hexene, 1.2 M, with [Rh(NBD)(Diphos)]+ under homogeneous conditions in methanol (0.2 wt % H20). (B) The hydrogenation of 1—hexene, o.u M, with [Rh(NBD)(Diphos)]+- hectorite in methanol (0.2 wt % H20). . . . 112 21 The hydrogenation of l-hexene, (A) 0.8 M, (B) 1.2 M, with [Rh(NBD)(Diphos)]+- hectorite in methanol (0.2 wt % H2O). . ... 11A 22 Hydrogen uptake plots for reduction in methanol of 1-hexyne to 1—hexene with homogeneous and intercalated rhodium triphenylphosphine complexes. . . . . . . . 128 23 (A) Proposed orientation of the alkyne- rhodium triphenylphosphine complex between the silicate sheets of hectorite prior to hydrogen transfer. (B) The critical dimension of 2-decyne, defined as the minimum distance which must be spanned by the molecule when CEC axis is per- pendicular to the silicate sheets . . . . . 133 2A Differential scanning calorimetry curve for (A) Z-6020-montmorillonite. XV Figure Page 2A (B) (CH3)2SiC12—montmorillonite in the range 25-300°C. . . . . . . . . . . . . 1A3 xvi ‘ ‘ h ,n INTRODUCTION A. Foundation, Homogeneous and Heterogeneous Catalysts In the past thirty years, the industrial application of processes catalyzed by soluble transition metal com- pounds has grown in significance. Some twenty or more processes use such catalysts to produce various organic chemicals. A few of the large scale industrial processes are the carbonylation of methanol to acetic acid which uses a rhodium catalyst with a methyl iodide promoter at low pressure,l’2 L-DOPA synthesis by asymmetric hydrogenation (Monsanto Process) where a cationic rhodium complex is used as the catalyst,3 a two stage aqueous oxidation of ,ethylene to acetaldehyde with palladium salts (Wacker 6 process),u- the manufacture of phenol from toluene by the oxidation of toluene over cobalt and copper salts with magnesium compounds as promoters (Dow-Toluene process),6 the production of vinyl acetate6’7 and substituted vinyl 8 from olefins and acetic acid over palladium and acetate copper salts, hydrocyanation of butadiene for the produc- tion of adiponitrile (DuPont process)5’9 with nickel catalysts, and propylene oxidation over molybdenum salts for the production of propylene oxide.lo’ll Despite this impressive growth, homogeneous catalysts are less widely used than heterogeneous catalysts in chemical industries. Nevertheless, homogeneous catalysts demonstrate great selectivity and economic efficiency in some reactions. The growth in the application of homogeneous catalysis has been followed by a rapid development of organo-transi- tion metal chemistry. The remarkable growth of organo— metallic chemistry and the need to understand catalytic processes have had a synergistic effect on the develop- ment and mechanistic studies of these systems.12 The most commonly used homogeneous catalysts are transi- tion metal complexes. Catalysis with these compounds offers' a number of distinct advantages over traditional homogeneous catalysts. In a sufficiently dilute solution, all the molecules of homogeneous catalysts are available for the catalytic reaction. Heterogeneous catalysis reactions occur at the interfaces of solid-liquid or solid-gas phases: all of the atoms or molecules not present at the surface are not accessible and remain unused. As a result, homogeneous catalysts allow easier interpretation of kinetics studies. Since homogeneous catalysts often have a definite stoich- iometry and structure and since powerful spectroscopic methods of solution analysis exist for proper characteriza- tion, the homogeneous catalysts can be prepared so that batch to batch performance of these catalysts is totally reproducible. In contrast, the structure of the surface of a heterogeneous catalysts is dependent upon both its method of preparation and its history subsequent to the preparation. Despite recent understanding and development in the field of surface characterization,l3 our understand- ing of the exact nature of chemisorption-catalysis mechan- isms is limited. Therefore, heterogeneous catalysts are not easy to reproduce, because the local environment of the surface structure is easily altered by slight change in the catalyst preparation and the reaction conditions. Homo- geneous catalysts will generally have only one type of active site. Consequently, homogeneous catalysts are more specific and easier to design for specific application. In addition, the electronic nature of the metal center and the steric requirements of the active site can be selectively modified through ligand substitution or variation of the solvent system. In contrast, heterogeneous catalysts, where several types of active sites may be present, often possess lower specificity and are more difficult to modify. In principle, homogeneous catalysts appears highly efficient and attractive. However, from a practical view- point, homogeneous catalysts suffer from three major tech- nical problems. The major disadvantage of homogeneous catalysts is the problem of separating the catalysts from the products at the end of the reaction. It is possible that both products and reactants can be distilled away 1,2 below the decomposition point of the catalyst. However, distillation is inevitably an endothermic process, and distillation is very expensive. Unless distillation is very efficient, it will result in a loss of catalyst and/ or contamination of the products. Also, distillation without destroying the catalyst may be impossible in the case of reactions which leave behind high-boiling side products. The extraction of catalyst by ion exchange or solvent extraction technique face similar problems. Even if the separation is achieved efficiently, the cost of the resulting process may be prohibitive in commercial applica- tions. Limited thermal stability is a second problem facing homogeneous catalysts. Even though homogeneous catalysts are very efficient under mild reaction conditions, it is sometimes desirable for kinetics reasons to carry out the reaction at an elevated temperature. Homogeneous catalysts, particularly organometallic compounds, exhibit poor thermal stability relative to heterogeneous catalysts. Hence, the ability to increase reaction rates by increasing temperature is often limited in this system. A third problem for many homogeneous catalysts is their limited solubility in suitable solvents. Often they are soluble and active in only a limited range of solvents. The solubility limitation also may limit the number of pos- sible substrates that are suitable for reaction with a given catalyst system. B. Supported Metal Complexes and Recent Developments A number of papers have discussed in great detail the general concept of supported metal complexes, the efforts to combine the best Virtues of homogeneous catalysts and heterogeneous catalysts have received considerable atten- tion.lLl The basic approach in this field has been to develop methods by which the solubility problems of homo- geneous catalysts can be alleviated along with the problem of separation. Efficient catalyst separation in homogen- eous systems has been achieved by using (1) catalysts bound to soluble highemolecular weight polymersls, (2) transition metal complexes dissolved in low-melting tetraalkylammonium salts of SnClg and GeClg anions,16 and (3) in biphasic- solvent systems.17 However, the most profound technique in this regard has been the attachment of soluble transi- - tion metal complexes to a variety of chemically inert, in- soluble supports. The immobilization of the homogeneous catalysts on a solid support offers ease of separation, mechanical and thermal stability, efficiency in multi- step or batch processes, and phase flexibility commonly associated with homogeneous catalysts. In addition, sup- ported systems can allow a greater degree of selectivity, efficiency, reaction control and reproducibility usually associated with their homogeneous counterparts. There may be other advantages experienced on heterogenizing transi- tion metal complex catalysts to a support. In certain cases the support even enhances the activity and specificity of the catalysts by inhibiting the number of unwanted side reactions. Catalyst immobilization may also inhibit de- activation of the catalyst by preventing formation of in- active dimer or polymer species.l8 Catalyst immobilization also may allow substrate selectivity to be based on molecu- 19 or dipolar factors.2O lar size restrictions The choice of the supporting materials depends on the catalyst under study. From a Chemical stand-point, the factors to consider are the inertness of the support to the reagents, the mobilities of the attached species and the polarity of the support relative to the reactants and products. The important engineering considerations are the porosity of support, the surface area, the heat transfer properties and the mechanical and thermal stability. A number of materials have been used as catalyst car- riers, which are compiled in Table 1. Amorphous inorganic metal oxides and organic polymers have practically dominated the field. In general, inorganic materials have better mechanical and thermal stabilities than organic supports. However, the synthetic approaches to organic polymers provide a greater range of surface functionality, pore size, and surface area, than the most inorganic carriers. A wide variety of techniques have been used to affix soluble catalysts to various supports, from simple physisorption to complex functionalization of Table 1. Materials Used as Homogeneous Catalyst Support. Inorganic Organic Silica Polystyrene, polyamines, Zeolites polyvinyls, polyallyls Glass polybutadienes, poly amino- Metal Oxides acids, urethanes, acrylic polymers Graphite cellulose, agarose,nylon, Clays allyl chloride, cross-linked Zirconium Phosphate dextrans. Biphasicz. Polysiloxanes, silica coated with phosphinated polystyrene Part of the listing is taken from Hartley, F. R.; Vezey, P. N., Advan. Organometal. Chem. 1271, ii, 189. the surface with ligands capable of complexing metal center. A partial list of these methods is compiled in Table 2. Many of these methods, which are a consequence of the three dimensional nature of the polymer, have no counterparts in homogeneous catalysts. When a soluble catalyst is attached to an insoluble support, the complex becomes heterogeneous at the bulk level, but it is essentially identical to the soluble analog on a molecular level. Thus, each catalyst site has a molecularity identical to that of its closest neighbors. The molecular nature of these .- F .:v_.~.d- u:_.‘.. A wu..~.-AV‘.A.‘~—nu .eJ M d . .wmu>.— --N.w~ .v X1; _.;_:.:V _.~ .. h SEE :2fi72_2 .‘ um mm :m mm 9mm Mmm Hm soapmcomoppmm compmcowoppzm coapmcowompzm OfippoEE>m< compmpfixoam coHumNHLoEomH coapmamshomopczm coapmcowoppzm cofipmHmELomoppzm : m mpfiAOpoonn+ m A nmmvnmg + mama + x +sAoapooamm AI m Acm .no .om u 2V ooAHoonus a x + Aoov z + oofiao A . oofifloonuwu A + mm + 00 + +onAmszcmL + ooaaoow ocfiaoonuwn+mmochm I oz + mmz + +moo + oofiaoon opHHooNIxI+mmOmAmmzvnmg Lo +mflmAm onHHoo m m m m + + +Amo A mzvcmL so +mfl A n2 + oofiaoonuw I flaAm + +N3m + mafiaoomlwlmz + m A +m oofiaoonux I maocm + ooaaoo Q possesses- Amhasv+a- aAm his + NAAoAmoovcmL + +iug + oofisooo opahouoocI+AAnocoHoVAo + +mAmonQfiBVAomzvnmu + mumpOpo Aoovxz mNI “Z + oovxsm I I m w + z mzvooL llm . s + z a mzvzmu Nnxunz mzvsmL mzvsoL mzvsoL .fi Imaocm nuxumz .a oovcmL mlw . n + z c mzvcmg I m .w on + z .poschcoo .m OHQMB 10 mm ppm ppm o.osm.oaa mm mm cofipomom ooasm one soon: coapmNfiLoszaom + mfiAoovocm + o m + AAAoovxcmmImAmeVZNmOIaBBOIAm H IA AoovxcmLmImAmmoVZNmoIBBBOIAd m NAM N movz moIamBOIAd ammo: + cIaAmmmov sNIcAoIAamL I :Ammmosz + mOIAHm I OIAHm mOIAHm $5. + m cmzA I cm: + OIAAm mOIAHm HI: : mm + mzoIAHm + m: + mOIAHm onovcm I mmmoImmoIAd I NAAoonovcmL + mmmoImmoIAs onoVoo I mmmOImmoIAd + onovmoo + mmmoImmOIAd Ammsosu ocficawosm wcfipzaoxm coapmcowoppmz mfimzaocowoppmm coapmmfihoEomH ”mCOHpomom aroma HBCOfiuocsm poupomddm Awo .mm .sm .hH u 2V x . QUHKO Hmpmfi OHCMWLOCHI AODV E QC®>HOm cacmwpo + AA88x: + wopfixo Hame OficmmpocH .Umzcfiucoo .N manme ll m: m: .m: m: m: H: 0: mm coauwcowompmm coauwcowOLphm :oHpmHzEBOQOLpzm compmcowospzm manoconnoz, coapmufimoefipoo mfimocpmpoz coauoscon cowmxoaa : . cxzmcmmIAm I x: + momsIAm .phOQQSm cud: mmpfiamn HmpoE mo mcoauomom mppommdm Umpmcfinmwonm npfiz macapomom xx + cqzmmcdawAmmovIHmIOIAemL I cqzmmcadxAmmovfimme + moIAHm AmAmmooovaI .ZBmmOI .mmI .ammOI .mmZI .mOI u AV emcxzmwammovIHmIOIAfimL I ems»: + wwAmmovIfimIOIAam amo + caochmmovchzmooIAi I mac: + mAmmovchamooIAd Home + chHochOIAHm I maocm + moIAHm voovozImom:BOIOImmOImAoszmIoIAHm m I voovoz + mo :BOIOImmOIonzVAmIOIAAm madam + smmAmcmmVHzflszOIAmL I :Amcmmvaz + smamBOIAa mAvammszUBOI.amaoaIHosomAmszAmv I mHosom + mmZIAd : .ooscaocoo .m oaone IromulzfiJCOC -mw UHQQH. 12 mm c.n:m.om mm www.mm Hm om m: w: w: COprNHpoEowHHo :oHpmHzELomoppmm coapmcowoppzm coHpmNHLoEHQ coapwcowoppzm coapmfimEpomoppzm COHpmcowoppmm coapmcowoppzm co + AmcamIAmv I onoVononcm + m moom + mHomAmnamI I thomAmoovnmL N\H + m: mama + mAmsddvosm I :Amsmdvoa + m HooocmmAmcdmm I HooomAmcmsvcm + madam +monon + IAszvnmmAm I +m + ononAamzvcm + NAN mAmcdaVAoovmcmm I mAmcasVAoovmsm + m m I mAmcamVHonm + m oom + mAmcmaIAvamaA I BHAoovecm + was cIaAooVHzcmmnadIAmL I 3032 + was .meQEoo Hopes on» Eopm pcmwfia m mo coauzu onomcm cmmlAm AmvcmL miIAmm cdeAm addIAm moIAav moIAmm miIAmv miIAaV cmmIAm 2mmIAm can + mAmnmmVHosmImmcimIAmL casIAm oovocm mIAm m mIAm C fiumndm .Q .Umscfipcoo .m manme l3 mm mm.nm: Am.nme mm coapmcowoppzm coaumHmELomopvzm coaumcowoppmm oasmeEmm< ANAmhsaIAnvaoAaamovemL I flHomAzmmoVnmv + mediIAm NAN I mam Am I mmHoAaoovcmL + m cmmIAmvaooVAosm Aoovaocmu + msamIAmm AmnddIAmVAvooovcmL cdeAmm moonQEoo EBHBOSLOLOHQOI: mo own>moao .0 am + mAmcmiIAmV AocoficnxomIm.HvsmL I AquosoHonxom m.chmL + mammIAm .wmscfiusoo .m canoe 1A catalysts, therefore, offers greater uniformity and re— producibility than traditional heterogeneous catalysts. Supported catalysts are usually less active than their homogeneous analogs. The physisorption-of metal complexes on solid supports suffer a major disadvantage, because catalyst desorption during reaction may ocCur mainly when reaction involves solid-liquid phases. On the other hand, covalently and electrostatically bound complexes desorb less readily. Covalently bound metal complexes may be leached out of the support if the covalent bonding is weak, or if the surface link is chemically degraded in the presence of solvent or excess free ligands. Electrostatic bonding on the other hand is strong. However, electro— statically bound complexes may be displaced by charged species formed in the reaction. The optimum catalytic activity of supported catalysts is expected under nearly solution like conditions. Often, the degree of freedom experienced by molecular catalysts in solution is decreased upon their immobilization. The loss of rotational and/or translational degrees of freedom of these complexes can :result in loss of catalytic activity in the bound form irelative to homogeneous conditions. The effect may be profound, especially for complexes covalently bound to the support by ligands of short chain length, where close proximity of the support can sterically interact strongly with the active species. The reduced degree of transla- tional motion can actually be desirable, particularly when 15 catalyst deactivation can occur by dimerization or polymeri- zation. In order to attain or closely mimic solution ac- tivity it is desirable, if not essential, that the metal complexes at least maintain their vibrational and rotational degree of freedom. A covalent linkage to a surface will always exert restrictions on rotational motions of molecules. However, ions electrostatically bound to support can exist in an almost solution like environment, even when in close proximity of support. Even though the ion exchange tech- niques appear to be very promising for catalyst immobiliza- tion, the main problem is the availability of catalytically active charged species which can be electrostatically bound to an appropriate support material. Zeolites, layered silicates and layered zirconium phos- phates are among the few inorganic supports that are capable of electrostatically binding cationic metal com- plexes. Naturally these materials are very attractive as catalyst supports for the reason discussed earlier (1192 sgpga). Zeolite materials, though thermally stable, suffer from a few limitations: (1) the ion exchange charge sites may not be uniform and can lead to variation in local catalyst environments, like those of heterogeneous catalysts. (2) The zeolites are limited to reactions of small mole- cules, because most of the pore openings are small. (3) The three dimensional alumino—silicate frame-work of zeo- lites is rigid and does not allow the possibility of altering l6 selectivity through solvent effects. The zirconium phos- phates are promising; however, their surface areas and cation exchange capacities limit their application. 27 have recently demonstrated Pinnavaia and coworkers that cationic metal complex catalysts can be intercalated between negatively charged sheets of mica-type layered silicates. In highly swelled systems, the metal complex catalysts intercalated between the layered silicate sheets are almost in a solution-like environment. For example, Rh(I)—triphenyl-phosphine,[Rh(PPh3)X]+ bound in the clay mineral hectorite was shown to hydrogenate alkenes and alkynes catalytically in the intracrystal region at ambient conditions. A suspension of the supported catalysts in methanol exhibited much higher specificity towards hydrogena- tion of 1-hexene than the homogeneous catalysts. Even though the catalytic activity of [Rh(PPh3)X]+—hectorite for the hydrogenation of l-hexene was nearly ten times lower than the homogeneous catalysts, the intercalated catalysts were nearly twice as active than similar catalysts supported on organic polymers. The catalytic activity of [Rh(PPh3)x]+- hectorite towards alkyne hydrogenation was almost equal to that of the homogeneous catalyst. The difference in ac- tivity of [Rh(PPh3)x]+-hectorite for the hydrogenation of l-hexene and l-hexyne may be due to the selective steric and adsorptive effects imposed by the support on the co- ordination behavior of the substrates. 17 Although a direct comparison between the polymer sup- ported and layered silicate supported catalyst cannot be made due to their structural differences, the greater re- action activity of the hectorite supported catalyst may.be due to the greater accessibility of the catalyst in the interlayer regions oftfluamineral. The success of these preliminary studies has illustrated the feasibility of carrying out catalytic reactions in the interlamellar regions of layered silicates and has prompted 60 more intensive research in the field. A number of cata- lytic reactions are being explored by using layer silicate supported catalysts. For example, the hydroformylation of 61 1-hexene with [Rh(COD)(P-P+)]+-hectorite show increased specificity towards the formation of normal aldehydes relative to homogeneous solution. The optical yields ob- tained in the asymmetric hydrogenation of prochiro amino acid precursors with layer silicate supported rhodium complex catalysts were almost identical to the yields ob- tained with the homogeneous catalysts.62 The substitution of positively charged ligands for neutral ligand in rho- dium complexes can eliminate the possible desorption of active species from the layered silicate support.29 18 C. The Properties of Clay Minerals and Their Structural Aspects The layer silicates, hectorite and montmorillonite, described in this dissertation are a smectite mineral, a class of naturally occurring clay minerals. Clay materials generally are defined in soil science and geology as any finely divided materials with particle size less than 2n. However, the term clay mineral refers to specific silicate minerals of defined stoichiometry and crystalline struc- ture with particle size less than 2p. These minerals have unique properties associated with their two-dimensional structure.63 These properties, along with their vast natural abundance, are the main reasons for their attractiveness as catalyst supports. The major building blocks of silicate minerals are silica tetrahedra and octahedra of alumina or magnesia. The framework of smectite layered silicates is composed of alternating arrays of interlayer cations and negatively charged silicate sheets, a schematic diagram of the structure is illustrated in Figure 1. All smectites including hectorite and montmoril- lonite have a 2:1 layer structure. The 2:1 notation refers to a structure in which two arrays of silica tetrahedra sandwich an array of either magnesia or alumina octahedra. The octahedral layer of hectorite is mainly composed of magnesia octahedra, whereas the octahedral layer of mont- morillonite is composed of alumina octahedra. The resulting 19 Figure 1. Schematic representation of the structure of smectite. 0 Oxygen .Hydroxyls .Aluminum, Magnesium, Iron OSilicon, Occasionally Aluminum 20 silicate sheets have four distinct layers of oxygen; two oxygen layers are shared by both the octahedral layers and the tetrahedral layers of silica. In the tetrahedral sheets three of the four oxygen atoms of each tetrahedron are shared by three neighboring tetrahedra. The fourth oxygen atom is shared with the octahedral sheet. This results roughly in hexagonal holes in the tetrahedral sheets formed by rings of six oxygen atoms. The hydroxy groups, which are present to fulfill the charge neutrality, replace oxygen atoms located within the hexagonal cavity in the octa- hedral sheet. In montmorillonite only two thirds of the octahedral holes are filled with aluminum and magnesium cations; the remaining holes are vaccant. In hectorite all of the octahedral centers are occupied by magnesium and lithium cations. The schematic representation of smectite is provided in Figure l. The isomorphous substitution of Mg2+ for Al3+ in the octahedral layers of.montmorillonite and the replacement of Mg2+ by Li+ in hectorite results in a net negative charge on the oxygen framework. The charge for hectorite and montmorillonite is in the range of 0.5 - 1.0 charge per unit cell. Charge neutrality is achieved by the pres- ence of an array of hydrated cations (sodium) between these layers. The idealized unit cell formula for hec- torite and montmorillonite are, Nao.67[Mg5.33’ Lio.67] (318.00) 020 (OH’F)u3 and 21 NaO 67EA13.33’ MgO 67] (818.00) 020 (OH)u , respectively. Note that some replacement of OH- by F- occurs in hectorite. The stacking of silicate sheets results in a crystalline layer structure. The alternate arrays of silicate sheet and hydrated cations are repeated in the crystallographic c-dimension. The adsorption of polar solvents between interlayers can result in considerable swelling of the interlayers. Van Olphen614 points out several factors that influence the one-dimensional swelling. The swelling depends on a combination of the solvation energy of the cations, the energy of adsorption of solvent into the intra- crystalline space, and the electrostatic energy between the charged species. 'The solvent polarity determines whether cation solvation energy or surface solvation is the pre- dominant force in the swelling process. As a consequence of interlayer swelling, the interlayer alkali metal or alkaline earth cations can readily be reached and replaced by simple ion exchange methods with almost any desired cations, including transition metal ions, carbonium ions and even protonated proteins. Electron spin resonance studies65 have shown that simple 2+ , tumble rapidly in a ions, such as hydrated Cu2+ and Mn solution-like environment when the interlayers are swollen to a thickness of 10-12 A. The rapid tumbling has also been observed for organic nitroxide spin probes in the 66 sw011en smectite systems. These observations provided 22 the bases for metal complex catalyzed reactions in the intracrystal space of the minerals. In the highly swelled interlayers, cationic organometallic catalysts exhibiting solution-like mobility should not lose appreciable activity compared to their homogeneous counterparts. In addition, these solvated interlayer ions should be readily available for the reagent molecules from bulk solution. The internal surface area of hectorite and montmoril- lonite (W750 m2/g) is greater than the surface area of amorphous metal oxides. The theoretical surface area is calculated from the unit cell dimensions 9.18 x 5.25 A and taking unit cell weight to be 760 g/mole. For example, the surface area of various types of silica are typically in the range of 200-500 m2/g.67 The surface area of most commonly used polymer, Amberlite XAD-2, is 120 m2/g.68 Considering, the cation exchange capacity (CEC) of hec— torite (70 meg/100g) and montmorillonite (83 meq/lOOg), one can estimate that ~80 A2 is available for monovalent interlayer cations. This large area allows even large cations enough space to exist on the surface without charge interaction or steric constraint. Commercially available ion exchange resins offer no more than 16 32 per exchange ion. Zeolite type X and type Y can provide large surface area in the supercage cavity of the zeolitic framework. However, zeolites are restricted to complexes of relatively small size due to their small pore openings (W6-l2 A). Complexes such as rhodium triphenylphosphine 23 would be much too large to exchange into Y-type zeolites.32 Smectites have been employed for many years as effec- tive heterogeneous catalysts for a number of reactions. Early catalysts for the cracking of petroleum feed stock to valuable products were made from acid treated montmoril- lonite clays.69 Theng has discussed extensively the organic-clay reactions and has reviewed the use of clay in the petroleum industry as cracking catalysts and as 70 The active sites are of the polymerization catalysts. Bronsted and Lewis acid type. The incorporation of various metals in synthetic mica-montmorillonite has shown that hydrocracking reactions with this material compare favorably with Pd-zeolite. The substitution of bivalent metal ions 2+, 002+, Zn2+ and Mg2+) in the vaccant (for example Ni holes of Gibbsite octahedral layer results in metal sub- stituted synthetic mica-montmorillonite. A synthetic mont- morillonite with Ni(II) in the octahedral layer may be used as a light petroleum hydrogenation catalyst, as well 60a The presence of nickel as a Fischer-Tropsch catalyst. increases the surface acidity and hence the activity of these catalysts. The catalysts are also active for hydro- isomerization and oligomerization of light hydrocarbon fraction.60a Metal ions intercalated in smectite can induce other types of reactions. For example, dimerization of anisole to A,A'-dimethoxy biphenyl was reported to occur with Cu2+- hectorite.71 The interlayer regions of smectite often alter 2A the reactivities and stabilities of transition metal com- plexes. Cu2+-arene complexes, not found in homogeneous conditions, are found and stabilized in smectites.72 The versatility of clay minerals and the ability of environment in interlayers to play constructive and ad- vantageous roles in transition metal complex chemistry are some of the many reasons why clay minerals are being in- vestigated as homogeneous catalyst supports. D. Cationic Rhodium Complexes and Rationale for Their Use as Hydrogenation Catalysts The current literature on rhodium metal chemistry deals mostly with the synthesis of rhodium complexes and the catalytic properties they possess. Many rhodium complexes with phosphine, arsine and stibine ligands are active catalysts for a variety of homogeneous reactions. For example, Wilkinson's catalyst, RhCl(PPh3)3, is an active catalyst for hydrogenation and hydroformylation reactions.73 2+ Cationic complexes of the type [Rh(NH3)50H] have been shown to be an active hydrogenation catalyst.7u Cationic rhodium complexes with nitrosyl ligands, for example, [Rh(NO)2(PPh3)2]+ was proposed to catalyze the reaction75 C0 + 2NO + N O + C02 2 Schrock and Osborn76 have reported the preparation of 25 a number of rhodium (I) and rhodium (III) cationic complexes that function as catalyst precursors. The cationic com- plexes were prepared by the addition of a bidentate or mono- dentate phosphine ligand to a solution of [Rh(NBD)Cl]2 in the presence of suitable anions. These complexes, upon passing hydrogen in solution, produced very active hydro- genation catalysts. These catalysts show significant specificity for hydrogenation of terminal olefins, for hydrogenation of alkynes to cis-olefins, and for conversion 77 The specificity and reaction of dienes to monoolefins. activities were dramatically altered under various reaction conditions. The substitution of a more basic ligand PMe3 or PPhMe2 for PPh3 in [Rh(NBD)(PPh3)2]+ a catalyst pre— cursor, rapidly increased the hydrogenation and isomeriza- tion of terminal alkenes to internal alkenes. Crabtree78 was able to reduce 1—alkenes and 1-alkynes selectively with similar cationic rhodium complexes, [Rh(COD)(PPh3)(Py)]+ and [Rh(OCOPh)(COD)(PPh3)2]. In their elaborate study of reaction mechanism for the hydrogenation reactions with cationic rhodium complex 77 have proposed three mechan- catalysts, Schrock and Osborn istic possibilities depending upon reaction conditions, and the nature of the ligands and the substrates. The details of these reaction mechanisms and their implica- tions will be discussed later. It is especially noteworthy that the catalytic species [Rh(PPh3)2]+, which is proposed to occur upon reductive 26 elimination of coordinated diene from [Rh(NBD)(PPh3)2]+- PFg solution, may be analogous to [Rh(PPh3)X]+ (x = 2,3) produced from the addition of triphenylphosphine to a solution Of Rh2(OAc)u in the presence of a non-complexing acid._ It is of interest to investigate the behavior of these catalysts, in the hydrogenation of various substrates in homogeneous solution and in the intracrystal space of hec- torite. E. Research Objective: Studies of Hydrogenation Reactions With Rhodium Complex Catalysts Supported on Layered Silicate In any particular solvent swollen system, the inter- layers hosting the metal complex catalyst are of more or less uniform thickness and polarity. Therefore, the pos- sibility exists for selective substrate adsorption and reaction based on size, shape or polarity. In addition, the preferred orientations of the intercalated complex under certain conditions of swelling may cause the substrate selectivity to differ from the selectivity of the same catalyst in homogeneous solution or from selectivity of related complexes covalently bonded to the surface of amor- phous polymers or inorganic oxides. The electric field gradient imposed by the silicate sheets may also influence 27 substrate selectivity through selective polarization of bonds. Some of these fundamental factors are apparently responsible for the difference between homogeneous and intercalated catalysts in the hydrogenation of alkenes and alkynes.79’275l One of the objectives of the present study has been to investigate the effect of intercalation on the substrate selectivity of cationic rhodium phosphine catalyst pre- cursor in the hydrogenation of olefins and alkynes. The layer silicates have the potential capabilities of affecting substrate selectivity based on molecular size or polarity. Therefore, a second objective of this disserta- tion has been to examine the effect of intercalation on the properties of [Rh(NBD)(Dppe)]+ as a catalyst precursor for the overall 1,2 and 1,4 addition of hydrogen to 1,3-buta- dienes. The reaction system was selected in part because the distribution of monoolefin products is kinetically regulated by the reaction pathways of a common intermediate. Under homogeneous conditions the thermodynamically more stable internal olefin products (1,A addition) are favored over synthetically more desirable terminal olefin product (1,2 addition). Thus it was of interest to determine the effects of catalyst intercalation on product distribution, as well as on the rates of reaction. The layer silicate possess Br6nsted acidity. The origin of Brdnsted acidity is proposed to arise from the hydrolysis of water present in the coordination sphere of 28 the interlayer cations.8O Thus, whenever a protonic equil- ibrium is involved between catalytically active cationic 773.38]- species in the catalytic reactions, this equilibrium would be influenced by surface acidity.- Recently, it has been demonstrated that catalytically active RhH2Ln+ is preferred in the interlayers of hectorite over neutral RhHLn (L = tertiaryphosphine, n = 2,3), which is in dis- + sociative equilibrium with RhH Ln . The shift in the proton 2 equilibrium may alter the specificity of the catalyst for l-hexene hydrogenation. The observed specificity for 1- hexene hydrogenation can be linked to the surface acidity and the role played by surface composition. The surface acidity is partially dependent upon surface adsorption and the nature of interlayer solvent.8O It is possible that a continuous change in the surface composition during any given reaction can dramatically change the surface acidity. Thereforetfluaintention of a third part of this dissertation has been to show that such dynamic processes may play a significant role in the hydrogenation of 1-hexene with a rhodium complex catalyst precursor [Rh(NBD)(PPh3)2]+ supported in the interlayers of hectorite. F. Pillaring Reactions of Layered Silicates Ordinarily, the naturally occurring smectites are not suitable as general adsorbents and for catalytic applications because the silicate sheets collapse upon one another, as 29 solvent is removed at elevated temperatures. Attempts to overcome this practical problem has created a considerable interest in introducing various pillaring species into the interlamellar regions of the swelling clays, so that internal surface is available for adsorption and catalysis even in the absence of swelling solvents.82 Organic materials have been quite successfully used to keep silicate layers apart at low to moderate temperatures but which fail in the temperature range of 250-500°C. For example, the exchange of protonated l,A-diazabicyclo[2,2,2]- octane(triethyldiamine) with interlayer cations of smectite produce dOOl interlayer spacings of 1A.2 A.83 The incor- poration of the same cations in the interlayers of H+- montmorillonite show markedly higher catalytic activity for esterification of carboxylic acids compared to ordinary alkyl-ammonium exchanged montmorillonite.8u As an alternate to organic materials, the inorganic molecular props are relatively stable in the temperature range 200-500°C and can produce higher interlayer spacings. For example, polymeric aluminum hydroxide cations, on inter- calation in the interlayers of smectites form a stable phase of 18 A (1001 spacings at 500°C.85 The intercalation of [Si(acac)3]+ in hectorite, subsequently followed by cal- cination at 500°C was suggested to form silica imbedded in the interlayers. The resulting hectorite containing array of silica in interlayers, retain its large surface area 30 and the d001 spacings was 12.6 A after calcination at 86 500°C. Pillered clay minerals are prepared by intro- ducing large cationic transition metal complexes, for ex— ample, Fe(Phen)32+ and Cu(Bipy)32+.87 These layered silicates in the presence of props can exhibit properties like those of two-dimensional molecular sieves. It should be possible to carry out gas phase re- actions and catalysis at high temperature by using these high temperature phase clay minerals. The synthesis of interlayer silica, by incorporation of Si(acac)3+ cations and subsequent calcination, has prompted other workers to apply new methods to synthesize interlayer inorganic oxides and chemical props. It is known that coupling reagents such as chlorosilanes can bind to the edges of clay minerals.88 Recently, incOrporation of these silane coupling agents between the layers of silicic acid in DMSO or DMF solution has led to a new thermally stable material.89 The chlorosilanes are known to hydrolyze in the interlayer and form expanded phases of 17 A, which adsorbed aliphatic amine molecules. However, these phases 90 The hydrolysis of silanes were not stable beyond 60°C. appears to be a result of interlayer water, and the inter- layer materials have been proposed to be of the siloxane type. The last part of this dissertation has involved the preparation and characterization of thermally stable expanded 31 phase montmorillonite by utilizing various organo-silanes. The purpose of this investigation is twofold. First, to prepare an interlayer organo-silicone material that acts as chemical props at higher temperature. Second, to func— tionalize the internal surface with silane coupling agents, which probably involves coupling of the -OH group of Gibb- site layer. The functionalization of the internal surface should permit immobilization of otherwise neutral complex catalysts in the interlayer of silicates, while retaining the usual cation exchange sites for exchange with cationic metal complex catalysts. This type of system would be novel and may allow for the design of multi-functional catalysts. EXPERIMENTAL A. Materials All solvents used in syntheses and hydrogenation re- actions were reagent grade. Spectrograde solvents were used in NMR and UV—Visible studies. The solvents were de- gassed by standard freeze-vacuum-thaw cycles prior to stor- ing in sealed bottles in the glovebox. Rhodium trichloride trihydrate (RhCl3°3H20), purchased from Englehard Industries, was used without further treatment. Natural sodium hectorite in spray dried form was ob— tained from Baroid Division of National Lead Industries. Generally, ion exchange reactions were carried out with the natural mineral. However, a small amount of CaCO3 present in the native mineral was removed prior to its use in infrared absorption measurements by washing “.0 g of the mineral two times with 100 mL of 0.1M NaHSOu, and then washing it free of sulfate with deionized water. Na+-montmorillonite was obtained from the Source Clay Mineral Repository. The mineral was sedimented and saturated with Na+ ions by adding an excess sodium chloride. The fraction containing particles less than 2“ was col- lected, centrifuged and dialyzed until free of excess sodium chloride and then freeze dried. Ion exchange of 32 33 Na+ in the native minerals with 1.0 M Cu(NO3)2 and sub- sequent analysis of the minerals for copper indicated the cation exchange capacities (CEC) were 70 meg/100g and 83 meq/100 g for hectorite and montmorillonite, respec- tively. Methanol used as a solvent in the hydrogenation re- actions was purchased from Matheson, Coleman and Bell con- taining 0.1 or 0.2% water. Higher percentages of water in methanol required for some reactions were achieved by ad- dition of a calculated amount of deionized water. The exact percent of water in the methanol was determined by the Karl-Fischer method. Acetone, benzene, DMSO and p- dioxane were purchased from Fischer-Scientific Col p- dioxane was dried and distilled over Na metal. Most of the substrate used in this study decompose slowly when exposed to the atmosphere, and only traces of impurities are necessary to poison the catalysts. There- fore it was necessary to purify the substrate immediately before use. 1-hexene (Pfaltz-Bauer, Inc.) was shaken with aliquots of an acidified solution of FeSOu until the red color of Fe(III) was no longer produced, dried over an-' hydrous CaSOu, and distilled over sodium under argon. A cis-trans mixture of 2-hexene (Aldrich Chemical Co., Inc.) was shaken with and distilled from a mixture of NaBHu and ethyldiglyme. 1-hexyne, 2-hexyne, 3-hexyne and 2- decyne (Farchan Division, Story Chemical Corp.) were 3A distilled over activated A1203 under an argon atmosphere. The chromatographic grade alumina 80-200 mesh (Matheson, Coleman and Bell), was activated by heating to 300°C for 24 hrs. and then allowing it to cool under argon. 2,3- Dimethyl-l,3-butadiene and 2-methyl-l,3-butadiene (Chemical Samples Co.) were freshly distilled under an argon atmos- phere. Fresh stock solutions of 1,3-butadiene (Pfaltz- Bauer, Inc.) were prepared by dissolving the gas in the desired, degassed solvents and determining the concentra— tion based on the mass increase. Bicyclo(2,2,l)hepta—2,5- diene (Aldrich Chemical Co., Inc.) was stored overnight over activated alumina, degassed and freshly distilled under argon prior to use in syntheses. The hydrogenation products expected in the catalytic studies were obtained in pure form from the respective companies which_supp1ied sub- strates. They were used as received as standard references in GLC to identify the retention time of the products. Triphenylphosphine (Aldrich Chemical Co.) and bis- (diphenylphosphino)ethane (Pressure Chemical Co.) showed no 31B NMR or by UV-visible observable oxide impurities by spectroscopy, and were used as purchased. Except for (CH3)3SiC1 and (CH3)2SiCl2 (Aldrich Chemical Co.) all other silanes used in the pillaring reactions were obtained from Dow—Corning Corp. They were handled in a dry, inert atmosphere to avoid possible hydrolysis. Bottled hydrogen was purified by passage through a 35 column of reduced BASF catalyst~R3-ll at 110°C and then through a bed of aquasorb (Mallinkrodt) or a A A molecular sieve. Bottled argon and nitrogen were purified by the same technique. Elemental and water analysis was done by Gailbraith Inc. B. Syntheses All syntheses were carried out under an inert atmos- phere either in the dry box or on a vacuum line. 1. [Bh(NBD)Cll2 Bis(norbornadiene)-u,u'-dichlorodirhodium, a starting material used in the syntheses of subsequent cationic com- plexes, was prepared by the reaction of RhCl3I3H2O and norbornadiene in ethanol solution according to the method of Abel gt al.91 The yellow crystalline complex decomposed above 2A0°C. 1H NMR (CDC13) 3.9l(t), 3.83(m), 1.16(t). The melting point and proton NMR spectrum agree well with 92 literature values. 2. Rh(NBD)acac (Acetylacetonato)norbornadienerhodium(I) was obtained by reaction of [Rh(NBD)Cl]2 and acetylacetone by a method analogous to that for the preparation of Rh(C2Hu)2acac by Cramer.93 A mixture of 0.62g (0.1A mmole) of [Rh(NBD)Cl]2, 36 10 mL of diethyl ether and 0.27 mL (0.027 mmole) acetyl- acetone was chilled to -80°C and a solution of 0.A5 g potassium hydroxide in 1.5 mL of water was added dropwise. The mixture was warmed to 0°C with shaking and 7 mL more diethyl ether was added. This mixture was stirred at 0°C for half an hour. The ether layer was separated, filtered, and chilled to -80°C. The Rh(NBD)acac separated as a fine yellow crystalline compound. The evaporation of some of the filtrate and then allowing it to cool to -80°C pro- vided an additional amount of the acetylacetone complex. M.P. 177-178°C. 3. [Rh(NBD)(diphos)]ClO“ Norbornadienebis(diphenylphosphino)ethanerhodium(I) perchlorate was prepared in a nitrogen filled glove box by reaction of the acetylacetone complex, Rh(NBD)acac, with diphos in the presence of H010“ according to the method of Schrock and Osborn.76 The orange crystals decompose above 212°C. The proton and phosphorus NMR spectra of the product were in good agreement with the previously re- ported spectra.76 1H NMR (CD2C12) 5.3A (A,olefin), A.33 (2,methine), 1.8A (2,methylene), 2.37 (A, PCH 31 P NMR (CH2012) -55.8 (d, JRh-P = 159 Hz). IR (KBr) -1 _ -l - 1A39s, 1A853 Cm (C-C), llOOs Cm C10“ xmax - A73 nm. 37 A. [Rh(NBD)(PPh3)21gg6 The hexafluorOphosphate salt of the norbornadienebis- (triphenylphosphine)rhodium(I) cation was prepared by reaction of [Rh(NBD)C1]2 with triphenylphosphine in the presence of KPF6 according to the method of Schrock and Osborn.76 The bright orange crystals decompose above 190°C. The proton and phosphorus NMR spectra were in good agree- 76 1 ment with those reported in the literature. H NMR (CDCl3) A.56 (A,olefin), A.o2 (2,methaine). 31P NMR -28.3 ppm (d, JRh_P = 15A Hz). IR (KBr) 1A39s, 1A853 cm'l (C=C), Amax = AA6 nm. , 5- 522£§§39921u Rhodium(II) acetate dimer was prepared by reaction of. RhCl3-3H20, sodium acetate and acetic acid in hot ethanol according to the procedure described by Wilkinson and co- workers.25b The product was recrystallized from methanol and recovered as the methanol adduct Rh2(CH3C02)u2CH3OH. The dark green crystals were stable to at least 2A0°C. The electronic spectrum of the complex in methanol (Amax = 586 and AAS nm) was also similar to the spectrum of the reported complex in ethanol (Am = 590 and AA6 nm). ax 38 6. Protonation of Rhodium Acetate The protonation of rhodium acetate was undertaken in an inert atmosphere to ensure a product uncontaminated by oxygen. The protonating solution contained 0.66 mL (A.9 mmole) of aqueous A8% tetrafluoroboric acid (J. T. Baker Co.) in 150 mL of methanol. The methanolic solution was deoxygenated by flushing with dry nitrogen for at least 20 min. and mixed with 0.60 g (1.2 mmole) of Rh2(CH3CO2)u- 2CH3OH. The mixture was heated to 60°C in an oil bath and stirred until the green solid dissolved (approximately two days). A small amount of insoluble black material (pos- sibly rhodium metal) was removed from solution by filtra- tion. The concentration of rhodium in the green solution was 0.008 M. The electronic Spectrum exhibited bands at 610, A23 and 250 nm, in favorable agreement with the band assignment made by Wilson and Taube for Rh2(CH3CO2)3+ 2+ 1 2 . a methyl acetate resonance at T 8.01 and 6.39. Lines at and Rh2(CH3CO2) H NMR spectra showed the presence of T 8.09, 8.11 and 8.12 were assigned to bridging acetate in + Rh2(CH3CO and Rh2(CH3CO2)§ . Specific chemical shift 4. 2)3 assignments for Rh2(CH3C02)§+, which may exist as cis and trans isomers, and Rh2(CH3CO could not be made 4. 2)3 because of insufficient spectral resolution. 39 7. [Rh(NBD)(diphos)]+—hectorite Na+-hectorite (100 mg, 0.07 meq) was stirred for 30 min in 5 mL of acetone or methanol, and then [Rh(NBD)- (diphos)]Clou (18.5 mg, 0.027 meq) in 5 mL of the same solvent was added. Stirring was continued for 10 min and the yellowish orange intercalate was filtered and washed several times with 5 mL portions of solvent to ensure com- plete removal of unexchanged rhodium complex. Elemental analysis indicated 1.62 wt % Rh. IR (mull), 1A39s, 1A85 cm"l (C=C). Amax (mull) = A65 nm. 8. LRh(NBD)(PPh3l2l+-hectorite Na-hectorite (0.2 g, 0.1A meq) was stirred for 30 min, in 5 mL methanol and then 1A.5 mg [Rh(NBD)(PPh3)2]PF6 (0.015 meq) in 5 mL of the same solvent was added. Stirring was continued for ten more minutes and the yellow orange intercalate was filtered and washed at least three times with 5 mL portion of solvent to ensure complete removal of unexchanged rhodium complex. Elemental analysis indi— cated 0.72 i 0.02 wt % Rh. IR (mull) lA39s, 1A85s cm’l (C=C) Am (mull) = A6A nm. ax + 9. 3h2(CH3gg2)fi_x—hectorite In a typical experiment 0.20 g Na—hectorite was stirred with 1.0 mL of the protonated rhodium acetate stock solution, A0 diluted to A mL with methanol. The slurry was stirred for 1-2 min, the liquid was removed by filtration, and the mineral was washed five to six times with 2 mL portion of MeOH and finally dried by suction. Rhodium analysis of several samples prepared by this procedure gave an aver- age composition of 0.72 i 0.0A% rhodium by weight. 10. Catalyst Precursors from Protonated Rhodium(II)- Acetate Homogeneous solutions of the catalyst precursor were prepared under oxygen free conditions by treating 0.92 mL oftXNPprotonated rhodium acetate solution (0.015 mmole of Rh) with the desired amount of 0.05 M triphenylphosphine solution and diluting the mixture to 25 mL with solvent. Samples of supported catalyst were prepared by treating freshly prepared rhodium exchanged hectorite containing 0.015-0.30 mmole of Rh with 0.05 M PPh to obtain the 3 desired PPh3:Rh ratio. The slurry was then diluted to 25 mL with solvent. Treatment of the green rhodium exchanged hectorite with PPh3 gave an immediate color change to a dark red-brown color which gradually lightened to a yellow- orange color upon standing for several minutes. Analogous color changes occur with the preparation of homogeneous catalyst solutions, except that they occur much more rapidly. Al C. Hydrogenation Reactions All hydrogenation reactions were carried out at ambient conditions and at constant pressure. ‘The catalyst solu- tions or suspensions were prepared in a nitrogen filled glove box placed in a specially designed flat bottom, hour glass-shaped flask (Figure 2). This design is effective in minimizing "creeping" of the finely divided mineral- supported catalyst up the walls of the flask during the course of the reaction. Also the flask provides flexibility of inert atmosphere transfer of substrate and removal of the reaction mixture. In a typical experiment, the flask was charged with 0.015 mmol of Rhodium catalyst precursor, and the flask was attached to a standard hydrogenation apparatus equipped with glass vacuum manifold, gas inlet valve, gas buret, mercury leveling bulb and mercury mano- meter (Figure 3). First, the entire assembly, except the hydrogenation flask, was evacuated and purged with pure dry hydrogen. The cycle was repeated three times. The reac- tion flask was then evacuated momentarily (sufficient to induce solvent bubble) and then pressurized with hydrogen. A one hour prehydrogenation period preceded the injection of freshly distilled substrate. In each case enough solvent was added so that the desired molarity of the substrate can be achieved, while keeping the total volume constant. Following the injection of substrate, the mercury manometer A2 Figure 2. Reaction flask used to minimize creeping of finely divided intercalated catalyst. (A) Serum cap for admitting substrate by syringe. (B) Magnetic stirring bar. (C) Reaction chamber, 6 (diameter) x 2 cm. A3 To vacuum pump BASF catalyst column Aquasorb column Manifold Open Hg manometer with meter stick scale Manifold vent to mineral oil container 50 ml measuring buret fitted to a Hg leveling bulb Cold water condenser Reaction flask Magnetic stir-bar Serum septum C—IH'JIQ '11 [11 UOUIID Stopcocks shown are of high vacuum seal type. Ground glass joints were greased and clamped together. * Solvoseal connector was clamped with Thomas clamp. Figure 3. Schematic of Hydrogenation apparatus. AA was at once adjusted to atmospheric pressure by means of the mercury leveling bulb and manifold vent. The initial buret reading was recorded and the timer was started. The rate of hydrogen uptake was monitored by observing the change in the volume, with time, necessary to maintain a constant pressure of 1 atm. Fresh hydrogen was added and the buret and manometer were reset each time approximately 50 mL hydrogen was consumed. The time lapsed and the amount of hydrogen that should have been consumed during this resetting process were taken into account and neces- sary correction to the total amount of hydrogen taken up 'was applied before the next reading was obtained. The hydrogenation rates were measured as mL H2/min/mmol Rh. The percent of the substrate converted into products were determined by removing a fraction of reaction solution at a regular interval of time, and its analysis by gas-liquid chromatography. After each run with intercalation catalyst, the catalyst was filtered off and the clear filtrate was checked for hydrogenation activity to ensure that the observed rates were due only to immobilized catalysts. D. Pillarinngeaction of Layer Silicate with Organo-Silanes 1. Pretreatment of Na+-montmorillonite The interlayer water of Na+-montmorillonite was removed by dehydration at 180°C in a vacuum for 2A hrs. The A5 dehydrated Na+-montmorillonite was reswelled with DMSO, by suspending 2.0 g Na+-montmorillonite in 50 mL DMSO for at least 18 hrs. The DMSO swelled Na+-montmorillonite was filtered and washed thoroughly with p-dioxane to ensure re- moval of excess DMSO, and finally dried under vacuum at 1 25°C, (d = 13.8 3). IR (KBr) 2800—2900 cm‘ (C-H 001 stretch), lA00 cm’l (C-H bend). 2. Silane/Na+—montmorillonite In a typical synthesis, a mixture of 0.2 g (0.16 meq) pretreated Na+-montmorillonite and 20 mL of p-dioxane was mildly refluxed under an inert atmosphere with continuous stirring for 30 min. 1 mL of silane was then injected into the mixture and was refluxed for 2A hrs. Silane/Na+- montmorillonite was removed by filtration and washed several times with 5 mL portions of dry p—dioxane to remove excess silanes and was finally dried under vacuum at 25°C. E. Physical Methods 1. X-ray Diffraction Study X-ray basal spacings were determined with a Philips X-ray diffractometer or on Siemens Crystalloflex-A both equipped with Ni filtered CuKa radiation. The film samples were prepared by allowing a suspension of the mineral in a A6 desired solvent to evaporate on microscope slide and monitoring the diffraction through 2° to 22° of 20. Basal spacings of fully solvated and swelled minerals were ob- tained by first forming a thin film of mineral on a l x l in. white, porous fire clay material in place of micro- scope slide. This sample was then suspended under solvent and allowed to equilibrate for 30 min. The solvent ad- sorbed in the pores of fire clay material allow it to function as a solvent reservoir and prevented the film from drying during the X-ray diffraction measurements. This is a very convenient method to determine basal spacing of the mineral wetted with a solvent which is very vola- tile. In case of air sensitive compounds the sample was wrapped with a thin polyethylene sheet which also helped to minimize evaporation of solvent. In a blank experiment, no diffraction patterns from microscope slide or fire-clay material was observed in the scanning range of interest. The peak positions in the angle 20 were converted to d spacings with a standard chart, (CuKa l = 1.5A05 A). 2. Infrared Spectra Infrared spectra were recorded on a Perkin-Elmer Model A57 gratting spectrophotometer. The samples were prepared by using a KBr matrix or mulling the sample in fluorolube and placing the mulls between CsI disks. Mulls of oxygen- sensitive samples were prepared in a nitrogen filled A7 glove box just prior to measurement. 3. Magnetic Resonance Spectroscopy The proton magnetic resonance spectra of rhodium com- plexes at 60 MHz were obtained with a Varian Associates T6055 analytical spectrometer. All spectra were recorded in the normal mode at a radio field frequency strength well below the level required to produce saturation. Chemical shifts are reported in 5 units. Tetramethylsilane was generally used as internal reference. Proton decoupled 31P NMR spectra were recorded on a Bruker HFX-lO spectrometer operated at 36.AA MHz. All samples were prepared in a glove box by using degassed, spectrosc0pic grade methylene chloride, caping the tubes and sealing them with a few turns of plastic tape. A solution Of 85% H3POu was used as an external reference. A. UV-Visible Spectra Electronic spectra were recorded on a Varian Associates Model Cary-l7 spectrophotometer. Spectra of mineral samples were prepared by mulling in mineral oil and placing the mull between silica disks. A mull sample of native mineral was placed in the reference beam to reduce the effect of scattering. Spectra of solutions were obtained using matched 1 cm path—length quartz cells. A8 5. Thermal Analysis Enthalpic processes occurring in silane/montmorillonites were recorded with a DuPont 990 thermal analyzer. The analyzer was operated in differential scanning calorimetry mode. The samples were prepared by using non-hermitic, high thermal conductive aluminum pans and covers. DuPont thermal analyzer grade aluminum oxide was used as a ref- erence. Most of the scans were carried out at a constant heating rate of 5°/min in the range of 25° to 500°C. 6. Gas Chromatography Gas phase chromatography of products were performed with Varian Associates Model 90P or 920 single column chromato- graph equipped with a thermal conductivity detector. The columns were 10 ft x 1/8 in. 10% UCW-98 (Hewlett-Packard) on 80-100 mesh chromosorb-W, 10 ft x 1/8 in. 10% AgNO in 3 propylene glycol on 80-100 mesh chromosorb-W, 6 ft x 3/16 in. 8,8' oxydipropionitrile on 80-100 mesh chromosorb-W and 6 ft x 3/16 in. Durapak (Waters Associates), n-octane/ porasil-C 100/200. Typically the columns were operated at 30°C with helium as the carrier gas at a flow rate of 35 ml/min. Injector and detector temperatures were 1A0° and 170°C respectively. The identity of reaction products was ~confirmed by comparing their retention times with chemically pure authentic samples. RESULTS AND DISCUSSION The catalytic hydrogenation of conjugated l,3—buta- dienes was undertaken with [Rh(NBD)(diphos)]+ as a catalyst precursor under homogeneous conditions and after intercala- tion in the interlayers of hectorite. The actual catalyst species is prOposed to form upon passing hydrogen into the solution of the catalyst precursor in a polar solvent.77a Three solvents, acetone, methanol and benzene were used to determine the effect of solvation on the rates of reaction and the product distributions. Sodium ions in the interlayers of hectorite are readily replaced with [Rh(NBD)(diphos)]+ cations. The cation ex- change capacity of the mineral is 70 meq/lOO g. Elemental analysis gave 1.62% Rh, which corresponds to replacement of 22% of the Na+ ions. The internal surface area covered by the cationic rhodium complexes is 50-60% of the total available area. The remaining interlayer surface area is occupied by sodium ions and solvent. The low loading of complex cation indicates that only a partial monolayer of rhodium complex cation is present in the interlayer regions of the mineral. The intercalated rhodium complexes were characterized by IR spectroscopy. The infrared spectrum of pure A9 50 Rh(NBD)(diphos)+ exhibits characteristic bands at 1A85 and 1A39 cm"1 which may be assigned to a phenyl group in plane deformation vibration,9l4 (Figure AA). The IR spectrum of [Rh(NBD)(diphos)]+-hectorite in fluorolube mull also ex- hibits the bands at 1A85 and 1A39.cm—l. The IR scan of the intercalated rhodium complex was limited to the 1800-1200 cm—1 region because the other regions of the spectrum were dominated by the characteristic IR absorptions of the layered silicate frame work (vSi-O, vAl-O, VO-H),95 Figure AB,C. Another indication of the retention of chemical and structural constitution of the metal complex upon inter- calation was provided by an electronic spectral study. The absorption observed at A72 nm for the complex in CH2012 solution was observed at A66 nm for the intercalated rhodium complex, where the sample was prepared in mull. This band 60c is assigned to a ligand + metal transition. The small deviation in the Ama between the homogeneous and inter- x calated rhodium complex can be attributed to the change in solvation environment. The ultra violet n-w* transitions96 also are observed for the intercalated complexes, Table 3. Thus, the spectral studies indicate that Rh(NBD)(diphos)+ retains its constitution in the intercalated state. The intercalation of [Rh(NBD)(diphos)]+ in the inter- lamellar space of Na+-hectorite results in an increase in d001 spacing from 12.6 to l8.A A, (Figure 5). The increase in spacing most probably is determined by the size of the 51 Table 3. UV-Visible Spectroscopy Data for Cationic Rhodium Complexes and Na-Hectorite Partially Exchanged With These Complexes. Band Positions System Solvent in nm a. [Rh(NBD)(diphos)]ClOu CH2C12 A72s, 332s, 270s b. [Rh(NBD)(diphos)1+- hectorite (mull) A65b, 272b c. [Rh(PPh3)n]BFu Methanol 350, 272s, 2665 d. [Rh(PPh3)n1+- hectorite (mull) 276m e. [Rh(NBD)(PPh3)2]PF6 CH2C12 AA6m, 353b, 2718 r. [Rh(NBD)(PPh3)2]+- hectorite (mull) A6Am, 355b, 272m Mull samples were prepared using mineral oil. Figure A. %T 0081 009l—' OONl“ OOZl Have Number czm'1 1T C l l I 1 l 1 l d d d .0 s s a N o o c 8 Have Number cm-I Infrared spectra of Na-Hectorite, [Rh(NBD)— (Diphos)]+—hectorite and [Rh(NBD)(Diphos)]- C104 in the region 1800-1200 cm-l. (A) (C) [Rh(NBD)(Diphos)]ClOu Na-Hectorite (B) [Rh(NBD)(Diphos)]+- Hectorite 53 (A) Na-Hectorite (B) [Rh(NBD)(Diphos)]+—Hectorite (Dry) (C) [Rh(NBD)(Diphos)]+-Hectorite Methanol Solvated (D) [Rh(NBD)(Diphos)]+-Hectorite Acetone Solvated 18.471 l B 12.6A I I 0 A 21.53 I I I I I I | l I I I I 1 121035421412108642 Degree of 2 0 -Degree of 2 9 Figure 5. X-ray diffraction patterns for Na+-hectorite partially exchanged with [Rh(NBD)(Diphos)]+ cations in different solvation conditions. 5A diphos ligand. In the absence of interlayer swelling sol- vent, the Ad001 basal spacing, defined as dOOl-thickness of a silicate sheet (9.6 A), is 8.8 A (93., Figure 5). In the presence of a swelling solvent, the interlayers swell to a considerable extent, (Figure 5, Table A). There is no apparent relationship between solvent polarity and interlayer swelling. However, both the solvation energy of surface and the solvation energy of the cationic complex should contribute to the increase in Ad . 001 A. Hydrogenation of 1,3eButadienes l. Homogeneous Hydrogenation The result for the homogeneous hydrogenation of 1,3- butadiene and two of its methylated derivatives, isoprene and 2,3-dimethyl-l,3-butadiene, in methanol, acetone and benzene with [Rh(NBD)(diphos)]+ as a Catalyst precursor are provided in Table 5. The addition of substrate to a light yellow solution of catalyst was followed by a rapid hydrogen uptake and a color change in the solution to orange. As can be ob- served from a typical plot of hydrogen uptake y§_time (Figure 6A), the rates of hydrogen uptake remained constant until 60-80% conversion of the diolefins to monoolefins. A typical plot for the,mole percent composition YE time for reduction of isoprene is provided in Figure 6B. It was 55 Table A. 001 X-ray Basal Spacings of Na+-Hectorite and Na+-Rectorite Partially Exchanged With [Rh(NBD)- (diphos)]+.a Miner l d K Ad X a 001’ 001’ Na+-hectorite 12.6 3.0 [Rh(NBD)(diphos)1+- hectorite (dry) 18.A 8.8 [Rh(NBD)(diphos)]+-hectorite (acetone solvated) 2A.5 lA.9 [Rh(NBD)(diphos)]+-hectorite (methanol solvated) 21.5 11.9 [Rh(NBD)(diphos)]+-hectorite (benzene solvated) 18.A , 8.8 aThe percent exchange was 22 i 1% in all cases. Am ma coca omoa oo.H expo om om om? 0mm om.o moo: C mm AH\\ amp .oom om.o .soo< III IE /.I\ . 0A A ad OAm com om.o expo // \\ om . a :H OAOA com om.o zoo: Am //mc ma\\ opA oom om.o .nooa Ah mm OMOH 0mm om.o expo // \\ mm 0mm om.o mom: mm mm H A; H V om\H om Sm om.o .eooa OARV WHOSUOQQ Dmfimm £m\®QmHQ Amy PC®>HOW mcwfia . OCOD wCOHQ .m . .HOWLSO Imam pmzamumo mm +Aoaanvnomzvnm Qua: mocofipmusm mo soaumsomOLozm msoocoonom .m manna 57 .cfimoaoocoe ow coawho>coo swan Lopes pocfiesopoo mm; mammamcm pospopmo .zm HoEE\cHE\AE mm commopaxo mopmmn ..H.IHO# OJN. USN 0mm Um p.30 UQHLVHMO QIHQZ mCOHpOmwh HH 98%. Regardless of the effect of solvent on the absolute rates of reaction, the product distributions are essentially solvent independent, with the internal olefin products. (1,A-addition) being strongly favored. The 1,2-addition products of isoprene generally favors the hydrogenation of the double bond containing the sterically less crowded carbon atoms. Hence, the yield of 2-methyl—l-butene is approximately two times that of 3-methyl-l-butene. Although the hydrogenation products of 1,3-butadiene, gig and EEEEE? 2-butene, could not be completely resolved by GLC, a par- tial separation indicated that cis-Z-butene was the pre- dominant product. Exclusively cis- olefin was reported as 77C for the hydrogenation of a product by Schrock and Osborn 2,A-hexadiene. These latter results indicate that the reduction of the conjugated diene is occurring mostly by gigglg coordination to the metal center.97a 2. Intercalation Catalysts Table 6 summarizes the results obtained with the inter- calation catalyst under conditions identical to those used to investigate the homogeneous reactions. As in the case of homogeneous reactions, the rates of hydrogen uptake were linear up to 60-80% conversion. A 5 min induction period preceded the reduction of 2,3-dimethyl-l,3-butadiene in 61 Table 6. Hydrogenation of Butadienes with [Rh(NBD)(Dppe)]+ Intercalated in Hectorite.a Diene Solvent Rateb 'Products (%)C Acetone 25 //A5 55 Methanol A.A 60 A0 Benzene <0.01 -- -- Acetone 300 2A 10 66 Methanol 126 28 16 56 Benzene <0.01 -— -— _- //\\ .//\ /.\+ ' Acetone A30 32 68 Methanol 370 39 61 Benzene <0.01 -- -_ aReaction conditions are identical to those described for analogous Homogeneous Reactions. bMLH2/min/mmole Rh. cProduct analysis after <98% conversion. 62 acetone, and a somewhat longer induction period (N25 min) was observed for this substrate in methanol. None of the other substrates showed an induction period with inter- calated catalyst in these two solvents. Typical plots of % composition gs time, for the reduction of isoprene in methanol and acetone are provided in Figure 7. Pronounced solvent effects on the rates of hydrogenation are observed for each substrate. Essentially no reaction occurs in benzene, even though, this solvent gives uni- formly high rates for all three substrates under homogeneous conditions. In addition, the intercalated rates are con- sistently higher in acetone than in methanol, which is the reverse order 0f solvent activity found under homogeneous conditions. Except for 2,3-dimethy1-l,3-butadiene reduction in methanol, overall 1,A addition is favored over 1,2 addi- tion, but the yields of 1,2 addition products are sig-' nificantly (almost two times) higher for the intercalation catalyst compared with the homogeneous catalysts. The intercalation catalyst provides consistently higher yields of terminal olefins in methanol than in acetone medium. In contrast, the product distributions are virtually iden- tical in these two solvents under homogeneous reaction condi- tions. To obtain an indication of the swelling properties of [Rh(NBD)(dppe)]+ intercalated hectorite, x-ray powder Figure 7. 63 100 - (A) 2*th71-2-Iutm 1m. x 001903111011 8 I Z—Ihthyl-l-Iutm 20 -I - J-HIthyl-l-lutm 100 “m: (RIM) 100 ‘1 22 '° ‘ r. S w - u , E Zflhyl-z-I-uu ‘0-1 Z-HochyI-l-Iutm 20 WWI-lake“ I r I 120 160 2” 2‘0 280 MW) SI Typical percent composition plots for the hydrogenation of iSOprene with [Rh(NBD)(Diphos)]+- hectorite (A) in acetone medium (B) in methanol medium. 6A diffraction measurements were carried out under conditions where the interlayers were solvated by the three solvents used in the catalytic studies. In all cases only one or two orders of 001 reflection were observed, Figure 5, indi— cating that the interlayers are interstratified. That is, some layers have spacings which are larger or smaller than the value indicated by the first order reflection. The interstratification can be caused by non uniform charge distribution among silicate sheets and by the preferential segregation of Na+ and [Rh(NBD)(dppe)]+ ions. Nevertheless, the observed reflections provide a qualitative indiCation of the extent of interlayer swelling. The position of the first order reflections were as follows: 2A.5 A (acetone), 21.5 A (methanol), 18.A A (benZene). Virtually no swell- ing occurs with benzene, because the same reflection is observed when no solvent occupies the interlayer regions. Both acetone and methanol, however, swell the interlayer region, with the former solvent being the better swelling agent. Substrate may also play a role in the swelling, as indicated by the induction periods for 2,3-dimethyl-l,3- butadiene. However, since the induction period is much shorter with acetone than with methanol, the nature of the solvent probably plays the dominant role in determining the extent of swelling. The effect of catalyst intercalation on the hydrogena- tion of butadienes are best compared relative to the homo- geneous solution. Relative rates and selectivity towards 65 1,2 addition are provided in Table 7. The difference in relative rates for the three dif- ferent solvent systems can be related to the extent of interlayer swelling. In the case of benzene, the inter- layers are not swollen nor sufficiently mobile to permit ready access to the metal centers by even the smallest substrate; consequently, the rates are more than 105 times lower than in homogeneous solution. However, the relative rates increase greatly when the interlayers are swelled by acetone and methanol beyond the dimensions of the unsolvated intercalate. As the swelling of interlayers becomes larger, the interlayer environment approach that of homogeneous conditions. The interlayer swelling was highest in acetone, consequently, rates of butadiene reduc- tion are faster in acetone than in methanol or benzene mediums. I The size of the substrate molecule relative to the extent of interlayer swelling seems to play some role in determining the differences in relative rates for the methanol- and acetone-solvated systems. The relative rate for 2,3-dimethyl-l,3-butadiene reduction decreases by a factor of ten upon replacing acetone with the poorer swelling solvent methanol, whereas for 1,3—butadiene the relative rates differ by only a factor of two for these two solvents. Thus the smaller substrate has better ac- cessibility to the metal centers. 66 Table 7. Relative Rates and Selectivity Towards 1,2 Addi- tion Products. Relative ‘ Relative b Diene Solvent Ratea Selectivity Acetone 0.83 1.5 Methanol 0.08 1.8 // \\ Benzene <10“5 ___ Acetone 0.39 1.8 Methanol 0.12 2.3 // \\ Benzene <10.5 , -_- Acetone - 0.55 1.6 //——-\\ Methanol 0.28 2.3 // \\ Benzene <10-5 . _-_ 1+ aIntercalated Rate/Homogeneous Rate. b% Intercalated 1,2 Addition Product/% Homogeneous 1,2 Addition Product. 67 Substrate size, however, is not the only factor in- fluencing relative rates. Although the relative rates in- crease with decreasing substrate size when the reaction is carried out in methanol solvated interlayers, no correla- tion exists between size and relative rate when the inter- layers are solvated by acetone. It appears that the spatial requirements of the substrate are important when the extent of swelling is more nearly comparable to the size of the substrate (as with methanol) but other effects, such as dif- ferences in solvation between the intercalated and homogen- eous states, begin to dominate when the swelling is larger and the interlayer adopt more solution-like character. The potential importance of solvation effects under homogeneous conditions is clearly illustrated by the large difference in absolute rates for 2,3-dimethyl-l,3-butadiene reduction (33., Table 5) in benzene and methanol or acetone. The effect of catalyst intercalation on monoolefin product distribution though modest in absolute terms, is significant. Depending upon the substrate and solvent, the yields of 1,2 addition products for the intercalation catalyst are 150 to 230% higher than those under homo- geneous conditions. The increase in the fraction of ter- minal olefin products in the case of 2,3-dimethyl-l,3- butadiene are comparable to those induced for this substrate in homogeneous solution when the diphosphine ligand at the rhodium center is replaced by diarsine ligand.77C 68 It is noteworthy that in absolute as well as in relative terms, the 1,2 addition products obtained with the inter- calated catalyst are higher in methanol than in acetone. These results indicate that the product distribution with the intercalated catalyst is partially influenced by steric factors and partially by the electronic effect of the charged silicate sheets. Since the product distribution under homogeneous conditions is remarkably independent of solvent polarity (cf., Table 5), it is unlikely that changes in solvation under intercalated conditions are responsible for the increased yields of terminal olefins. The hydrogenation of 2,3-dimethyl-l-butene and 2,3- dimethyl-Z-butene with homogeneous and with intercalated [Rh(NBD)(dppe)]+-catalyst precursor in methanol showed no observable isomerization (Table 8). The rates of alkene reduction were almost l/25 that of the rates of diene reduc- tion, which indicate that the quantitative formation of monoolefin by reduction of diolefin is related to the rela- tive preference of substrate coordination. The origin of the catalytic activity of cationic rhodium complexes has been discussed by Schrock and Osborn.77 The mechanism of conjugated diene reduction with [Rh(NBD)— (dppe)]+ as the catalyst precursor has been shown to involve a coordinately unsaturated pathway in which the diene adds to the metal center before oxidative addition of hydrogen. Three mechanisms proposed by Schrock and Osborn77 for 69 Table 8. Hydrogenation of 2 ,3-—dimethy1-1-butene and 2 ,3- dimethyl- -2- butene in Methanol Under Homogeneous and Intercalated Conditions with [Rh(NBD)(Diphos)]+- Catalyst Precursor. a b % Substrate System Rate Isomerization Homogeneous W2 0 // Intercalated <0.2 0 Homogeneous <0.2 0 aSubstrate/Rh = 150, me H2 min-1(mmol Rh)-l. All reactions were carried out at 25°C 70 the hydrogenation of olefins, alkynes and diolefins with rhodium complexes are reproduced in Figure 8. The un- saturated pathway (path C in Figure 8) is believed to operate for 1,3-diene reduction. Little is known regard- ing the structure of the [RhH(R)(dppe)]+ intermediate which precedes the reductive elimination of the olefin, but n3-allyl species (1) has been proposed for the homogeneous .catalytic reduction of 1,3-dienes.97 The transfer of hydrogen to the a carbon would lead to overall 1,A addi- tion, whereas overall 1,2 addition would result from hydrogen transfer to the y position. A facile n3 to n1 arrangement (w-allyl to q-alkenyl), analogous to that 3 98 proposed for syn-anti interchange of n -allyl complexes, could precede the hydrogen tranSfer step. In this latter case, the distribution of monoolefine products would be 1 species 2 and 3, determined by the relative energies of n Figure 9. The reaction mechanism of conjugated diene hydrogenation with the intercalated catalyst, based on the catalyst characterization and reaction pattern, should be similar to that of the homogeneous catalyst. Since the product distribution is determined by the changes occurring at the [RhH(R)(dppe)]+ stage, the intercalation of the catalyst in hectorite can influence the migration step of second hydrogen in the n3-allyl intermediate or it may influence the relative energies of the n1 intermediates. 71 a z~ 2==~xra a: :N o. :+ + 2.9 m 3.5....“ ARIA + as; 3:. w 2%.... .nxszxrsu+ szAo.vp=L+ z xzxacdvrs Ax=I~Ae.vp=L+ .moonQEoo Esfipocp spas mconpmooppmn ooumLSQBm Is: no cofiumcowosczz afluzampmo on» son manpmpoufifi CH pomoqopo mamficmnooz .w opswfim 72 Rh-H R ' IN ILA) Figure 9. Probable structure of the intermediate [RhH(dppe)1+. 73 Whatever the precise mechanistic details may be, catalyst intercalation decreases the difference in activa- tion free energy for the two transition states which lead to 1,2 and 1,A addition. For all three substrates in- vestigated, the tendency toward 1,2 addition product is consistently higher for the more constricted methanol- solvated intercalate than for the acetone solvated inter- mediate. Whether the effect of catalyst intercalation on the distribution of kinetically regulated products result from spatial factors or from polarization effects induced by the charged silicate sheets cannot be determined at this time. Since the intercalation catalyst is interstratified and not uniformly swollen by acetone or methanol, it is pos- sible that only a fraction of the interlayers have the proper spacing or charge density to influence product distribution. Interstratification in hectorite and other naturally occur- ing layered silicates result in part frOm non-uniform charge distribution on the silicate sheets.99 A layered silicate intercalation catalyst with a more uniform charge distribu- tion may lead to more dramatic effects on product distribu- tion. B. Hydrogenation of Monoolefins The cationic metal complex catalyst system selected for the investigation of monoolefin hydrogenation, with layer silicate intercalation catalyst, is one that Legzdins 7A t l.25 originally developed for the purpose of hydrogenat- ing olefins that are soluble only in polar media. As discussed earlier in the experimental section, the prepara- tive reaction scheme involved the protonation of rhodium (II) acetate dimer, Rh2(OAc)u, with HBFLl in methanol, fol- lowed by reaction of the protonated product with triphenyl- phosphine (PPh3) and molecular hydrogen. The product of A+ the protonation reaction was believed to be solvated Rh2 to form a cationic Rh(I) + n' ion, which reacted with PPh3 catalyst precursor of the type Rh(PPh3) Evidence for cationic complexes was provided in part by the isolation of Rh(PPh3)3BFu and by the observation that the complexes bind to cation exchange resins and retain their catalytic properties in the resin bound environment. 100 have noted that h(III)_ More recently, Wilson and Taube Rhg+ as prepared by inner—sphere reduction of R (H2O)5Cl with Cr2+ (aq) has chemical and physical properties different from those claimed for Rhg+ by Legzdins BE ei.25 This prompted Wilson and Taube to reinvestigate by column chromatography the protonation of Rh2(OAc)u by strong, non- complexing acids. Their results indicated that protona- tion of the complex is incomplete, with Rh2(0Ac); and Rh2(OAc)§+ being formed in relative amounts that depend on reaction conditions. Neither the triply charged species Rh2(OAc)3+ nor Rhg+ was observed as a reaction product. Based on 1H NMR studies of the protonation of Rh2(OAc)u 75 by HBFu in methanol under the conditions used by Legzdins et al., Hoffman has shown that 30% of the acetate ligands reacted to form methyl acetate.101 The remaining acetates + . are distributed between Rh2(OAc)3 and Rh2(OAc)§+ complexes which are present in 3:1 molar ratio. The addition of PPh3 to Rh2(0Ac)fi:X solution at PPh :Rh = 3:1 results in the 3 loss of all coordinated acetate which can be observed by NMR and IR spectroscopy. Therefore, the transformation of Rh in the reaction system of Legzdins et al. is better described according to the equation: HBFu + nPPh3 + Rh (OAC) -+ Rh (OAc)X + Rh(PPh ) (l) 2 A MeOH 2 A-x 3 n (x=l,2) It is presumed, based on the analogies to the recent work of Schrock and Osborn77 (vide infra), that the addition of H2 to Rh(PPhB); forms a dihydride, which may exist in equilibrium with a monohydride: RhH (PPh )+ * RhH(PPh ) + H+ (2) 2 3 n t 3 n These complexes will later be shown to be the most likely species responsible for the observed catalytic activity. The green Rh2(OAc)fi:X (x=l,2) complexes, readily dis- place the interlayer Na+ ions of the mineral hectorite in methanol suspension. Loadings corresponding to 0.7 to 20% 76 of the exchange capacity of the mineral ( 73 meq/100 g) could be achieved by adjusting the concentration of Rh2- + (OAc)fi_X in contact with the mineral phase. Based on the idealized unit cell formula of the mineral [NaO 66 (Mg5 3A’ LiO 66) (818 O) 020 (OH,F)u] and the a and b cell parameters (5.25 x 9.18 A), the average Rh-Rh distance is 20 A, which 4. should allow ample surface area for formation of Rh(PPh3)n complexes (the size of Rh(PPh3); where n 2 is 12 x 13.5 i). The treatment of the green Rh2(OAc)fi:x-hectorite with ' + n methanolic solution of PPh3 affords yellow-orange Rh(PPh3) hectorite. Evidence for the loss of coordinated acetate in the intercalated state is provided by IR spectrosc0pyu 1 Figure 10 shows the 1800-1200 cm- IR spectrum of Na+- hectorite solvated by methanol. The replacement of 10% + of the Na+ ions by Rh2(OAc)fi_x ions gives spectrum 10B in which the bands at 1575 and 1A53 cm-1 are assigned to the symmetric and antisymmetric vibration of coordinated 102 acetate. Upon the addition of PPh to Rh2(OAc)fi:x_ 3 hectorite at PPh3 to Rh ratio of 3:1, the acetate bands are replaced by two new bands at 1A85 and 1A39 cm_l. These latter bands are assigned to the in plane deformation of phenyl rings of PPh3. % T Figure 10. I I439 “\_. I I I i I I I '800 I600 I400 |200 vIcm") Infrared spectra (1800-1200 cm'l) of methanol- solvated hectorites. (A) Na+-hectorite. (B) Rh2(OAc)fi:x-hectorite. (C) Rh(PPhB);-hectorite. The rhodium loadings used to obtain spectra B and C (0.76 wt %) correspond to m10% of the cation exchange capacity of the mineral. 78 l. Olefin Reduction Rh(PPhB);-hectorite was prepared as a Catalyst pre- cursor for the reduction of 1-hexene by adding enough PPh3 to a methanol slurry of Rh2(0Ac)fi:X-hectorite so that the average value of n for the surface complex is in the range 2.0-3.0.270 A preliminary study of the effect of catalyst loading on the rate of reduction of 1.0 M l-hexene in methanol showed that the turnover frequency or rate per mmole of rhodium remained constant over the range 0.06-0.7 wt %. At loading above 0.7 wt % the turnover frequency decreased with increasing loading. Thus all reactions were carried out at 0.72 i 0.0A wt % loading. At this loading m10% of the Na+ exchange ions have been replaced by rhodium complex. Based on the estimated size of a Rh(PPh3); complex (~l60 A2, from CPK molecular model), at least 25% of the total surface area is occupied by the complex ion. The reduction of l-hexene with [Rh(PPh3)n]+-hectorite is preceded by a very brief induction period, probably because some time is needed for the concentration of olefin to build up on the interlamellar surface of the mineral. After 100 catalyst turnovers (defined as the moles of substrate hydrogenated per mol of rhodium) the rate of hydrogen uptake achieves a constant value of ml6 mL H2 min.l (mmol Rh)-l and remains near this value up to 1200 catalyst turnovers. The constant rate of reaction with increasing substrate conversion may indicate that 79 physical adsorption forces maintain a constant surface concentration of l-hexene over the range of solution con- centrations examined (l.0-0.A0 M), significantly, no ob- servable isomerization of l-hexene to 2-hexene occur over the observed rangecfi‘substrate hydrogenation (60%). The performance of Rh(PPh3);-hectorite as a catalyst precursor for the hydrogenation of l-hexene is compared in Table 9, with the behavior of homogeneous Rh(PPh3); complexes as precursor for the reduction of the same sub- strate under analogous reaction conditions. It is ob- served that at PPh3/Rh = 2.0, the homogeneous complex exhibits a much larger initial rate of hydrogenation than the intercalated complex. However, the rate drops off dramatically as the extent of hydrogenation increases. By the time 750 hydrogenation turnovers have been achieved, the rate is less than half of that observed for the inter- calated catalyst after 1200 hydrogenation turnovers. More- over, extensive isomerization of l-hexene to 2-hexene ac- companies the hydrogenation reaction. For example, at 95% conversion of the substrate, 37% of it is hydrogenated and 58% is isomerized. Increasing the PPh3 to Rh ratio to Azl inhibits the isomerization, but after 9A% conversion, the extent of isomerization (3A%) is still significant relative to the extent of hydrogenation (60%). The data in Table 9B show that the homogeneous catalyst is much less active for hydrogenation of 2—hexene than l-hexene, the 80 am A2 AABBVBBNH om :B Aaomvoooa mm BMH Asmevomm 3H oma Aszmvoms II oom ABB BBBH 0.: BB Be Amsmvomu am a: AsomvooB 3H ozfi Asmavomm I omH Afim VOOH o.m .cHom BSOBCBono: AV BA AABBVBBNH AV AA Aaomvoooa AV BA AABNVBOB AV AH AAB VBBA o.B AV MA ABBBBBBNH AV BA Aaomvoooa Hv BH Aammvoom c m AV BA aAAB Booa 0.: ooofitoooomI+A caavcm soapmNHLoEomH oopmm ompo>OCLse £m\m£mm Eopmmm pmmamumo o A ohAAnono ocoxomIH Noc059 pmzampmo £m\mcmm Eopmzm unmampmo wosmxomIm Amv 0B me AmomvooB BB 00 Asmmvo0m m I B 0BH As00000: A 002 2 50.00 m 0B0 AAB VO0H 0.: cofiosfiom hsoocomoeom ma me Afiomvooma m 0: Aa0B0000H m 0mH Asm:0000 : I AV 0BH Agmm000m A odor z A0.00 Hv ozm Aam VOOH 0.: soapsaom msoocowosom BQOHpmNHLoEomH oopwm nmho>ocpze £m\mnmm Eopmmm pmzamumo & unmampmo .Boscfiocoo .0 oases 82 .oLSBxAE mcmppImHo 0 tom nocfimpno who; mocoxozIm Rom mono coapwcowopommw .copwcowopczn mm: page omeumnzm Havoc mo coauompm map o>Hw nanosecopmo cfi Bonds?H .003 :0.wmA.0 ha mcaonofi asaoocmo .mocoxonIm pcm IN 0» UoNHpoEowfi ocoxmnIH Hmpou mo :oHpompmp .zm HoEE\cHE\mm 4E .oxmpas Comopcmc mo momma .cm mo oaoe Log UmpdCowospzz omeuBQSB mo mmHoE mm pmcfimoon I .0000 .otsoasooeoo ”A 0000 as ofipmp cm 0» omemeSm Hmfipacfi mHocmnuoE :H z o.H ma coauwppcoocoo omequSm Hmfipfich "WQPOCDOOE .ooscaocoo .0 manna 83 initial rates differing by more than a factor of 10. On the other hand, the intercalated complex shows even lower activity for hydrogenation of 2-hexene (Table 98), relative to l-hexene. Thus not only does intercalation lead to a dramatic reduction in isomerization during the hydrogena- tion of a terminal olefin, but also the selectivity of the catalyst for reduction of terminal olefins over internal olefin is enhanced by intercalation. The low activity of the intercalated complex toward 2-hexene is not due to inhibitive complexation effect by the internal olefin since Rh(PPh3);-hectorite can be used to selectively hydrogenate 1-hexene in an equal molar mixture of 1- and 2-hexene. The difference between intercalated and homogeneous Rh(PPhB); catalyst precursors for olefin hydrogenation may be related to a difference in the position of equilibrium between catalytically active dihydride and monohydride complexes proposed in equation 2. Schrock and Osborn77 have shown that an equilibrium mixture of RhH2L; and RhHLn complexes results from the hydrogenation of Rh- (diene)L; catalyst precursor, where L is a tertiaryphos- phine and n = 2 or 3. The hydrogenation of olefin occurs by two different paths (ef. Figure 8). The presence of a Bransted acid favors the almost colorless cationic di- hydride complexes, which catalyze the hydrogenation of the terminal olefin with little or no isomerization. The neutral deep yellow, monohydride complex which can be generated in 8A the presence of a sterically hindered base such as NEt3, is even a better hydrogenation catalyst, but it also is a good isomerization catalyst. The hydrogenation properties of homogeneous Rh(PPh3); precursors formed from Rh2(OAC)::u and PPh: in the present study are qualitatively identical with those exhibited by the Schrock and Osborn catalysts under acidic and basic conditions. As is shown by the data in Table 9A, the extent of isomerization of 1-hexene under homogeneous hydrogenation conditions is dramatically re- duced in the presence of 0.07 M HClOu and greatly enhanced in 0.07 M NEt3 relative to methanol solution in the ab- sence of acid or base. Moreover, in the presence of 0.07 M HClOu a cis-trans mixture of 2-hexene is very slowly reduced by Rh(PPh3); precursor at a total PPh to Rh ratio 3 of Azl (Table 9B), suggesting that the RhH2(PPh3); species show a marked preference for reduction of terminal olefins over internal olefins. Thus the catalytic behavior of the intercalated complexes most likely is determined by a sur- face acidity effect which favors the formation of surface RhH2(PPh3); over RhH(PPh3)n species. The existence of a surface equilibrium is verified in part by the isomerization of 1-hexene with the intercalated catalyst in the presence of NEt3. The desorption of a yellow rhodium phosphine complex also occurs in the presence of NEt but the de- 3, sorption rate is sufficiently low to allow the observation of surface isomerization of l-hexene to 2-hexene in the 85 initial stages of reaction. The desorption product is most likely a neutral RhH(PPh3)n species, because it ex- hibits the same hydrogenation and isomerization properties as Rh(PPh3); precursor in the presence of NEt3. The surface Bronsted acidity of swelling layered silicate arises mainly from the hydrolysis of the interlayer ex- change cations.8O In the system investigated here, the principal exchange ion is Na+, and its hydrolysis, due to the presence of small amounts (m0.2 wt %) of water in methanol, probably determines the surface acidity. The second factor which may have affected the equilibrium in Equation 2 in interlayers of hectorite could be the intrinsic preference of cationic rhodium dihydride complexes over proton which has small charge to radius ratio. C. Olefin Reduction with LRh(NBD)(PPh3121+ Catalyst Precursor The surface composition can dramatically alter the surface acidity. Variation in the surface acidity at various stages of dehydration for Ca-montmorillonite were 80b It has been observed in the shown to be remarkable. hydrogenation of 1.0 M l-hexene with Rh(PPh3);-hectorite (ef. Table 9) that after W60% conversion of the substrate a rapid isomerization of terminal alkene to internal alkene was initiated. This latter result indicates that the sur- face composition and hence surface acidity must have changed with 60% conversion of substrate. 86 Evidence that surface acidity changes with surface com- position was obtained by varying the initial concentration of substrate, and wt % water of methanol in the hydrog- enation of 1—hexene. The initial conCentration of sub- strate was such that it compared closely to the concentra- tion of substrate that should be present at different per- cent conversion in the hydrogenation of l—hexene (1.0 M) with Rh(PPhB);—hectorite precursor. The catalyst precursor chosen for these studies were [Rh(NBD)(PPh3)2]+ and [Rh(NBD)(diphos)]+, previously re- ported by Schrock and Osborn.77 The reaction conditions were identical to those used for the Rh(PPh3);-catalyst precursor in the hydrogenation of 1-hexene. The results for l-hexene hydrogenation with [Rh(NBD)- (PPh3)2]+-hectorite as a catalyst precursor in methanol (0.2 wt % water) solution are provided in Table 10. Prac- tically no induction periods were observed. -The rates of hydrogen uptake remain constant over 50-60% hydrogenation of l-hexene initially present in the solution. The absolute rates of hydrogenation increased uniformly from m150 to 650 mL H2 min-l (mmol Rh)‘l when the initial concentration of l-hexene was increased from 0.1 M to 1.2 M. It is pos- sible that rapid mass transfer of substrate from solution into interlayers of hectorite is favored at higher initial concentrations of substrate, consequently a steady increase in the hydrogen uptake is expected. However, at concentrations 87 Table 10. Catalytic Hydrogenation of l-Hexene in Methanol (0.2% H20), [Rh(NBD)(PPh3)2]+-hectorite as Catalyst Precursor.a [l—Hexene] Catalyst) 0 % Isomeri- M l-Hexene/Rh Turnover Rate zationd 0.1 100 20(20%)e 155 2 A0(A0%) 153 6 70(70%) 130 1A 80(80%) 91 17 0.4 A00 A0(10%) 323 3 100(25%) 370 16 200(50%) 387 26 265(65%) 359 23 286(70%) 180 20 0.6 600 60(10%) 503 9 1A0(23%) 529 29 220(37%) 50A 37 305(50%) A78 39 362(60%) 180 39 372(62%) 88 39 0.7 700 100(1u%) 670 . <1 220(32%) 5A2 <1 350(50%) 390 A A20(60%) 368 15 0.8 800 118(15%) 851 <1 286(36%) 85A <1 395(A9%) 8A1 <1 510(6A%) 757 8 1.0 1000 110(11%) 5A3 <1 310(31%) 5A9 6 ”95(50%) 5A6 17 600(60%) 230 20 1.2 1200 120(10%) 658 <1 385(32%) 67LI l5 515(A3%) 697 19 725(60%) 675 23 770(6A%) 606 22 aRhodium loading was 0.72:0.0A wt%. bDefined as moles of sub- strate hydrogenated per mole of Rh. Rate of hydrogen uptake, mL H2 min“l (mmol Rh)‘1. dFraction of total 1-hexene isomer- ized to internal alkene. 8Values in parentheses give the fraction of total substrate that was hydrogenated. 88 beyond 0.8 M l—hexene other factors start dominating the reaction. The composition of the reaction solutions were monitored over 5—100% conversion of substrate. It is es- pecially noteworthy that the isomerization of l—hexene to internal alkene was significantly dependent upon the initial concentration of l-hexene. At initial concentrations of l-hexene, in the range of 0.1 M - 0.6 M, isomerization was observed concomitantly with the hydrogenation of l-hexene. For example, at 50% conversion, almost 20 and 30% of l- hexene was converted to 2-hexene when the concentration of the substrate was 0.4 and 0.6 M, respectively. The plots of percent composition with respect to time are presented in Figures 11-1u. Upon increasing the initial concentration of l-hexene to 0.7 M and beyond a dramatic shift in product formation was observed. Unlike the low initial concentra- tion of substrate, no immediate isomerization of l-hexene to 2-hexene was observed. For example, at 50% conversion nearly all of the substrate was reduced to hexane for solu- tions containing 0.7 and 0.8 fl l-hexene. Similar reaction patterns were observed if the initial concentration of l- hexene was increased to 1.0 or 1.2 fl. However, at 50% conversion of the substrate about 10-1uz isomerization was observed (Figures 13B, lHA). These observations indicate in part the importance of initial surface composition and suggest that surface composition alters the hydrogenation reaction of l-hexene. Consequently, the equilibrium between 89 [HI-nu]- 0.1 5 tutu-o1 (0.2 2 320) [mmumhr- acetone: 100.. ml. 2 0090051110" 8 1 100 TIME (HIM) [14...] - 0.0 g html (0.21 .10) Elton!» ("09,1 ‘- mm. ’01. 1 m1“!!! Figure ll. The hydrogenation of l-hexene (A) 0.1 M (B) 0.u M with [Rh(NBD)(PPh ]+-hectorite in methanol (0.2 wt % H20). 3)2 90 (I) 10o_ [Hie-Io] - 0.0 ! Much-.1 (0.2 v: '2 I20) [mmnmpzf- Encarta a ”_ 1M 5 E "“" co q 0‘ so- 20".!“ 201 I t I I I 20 ‘0 60 so TM mu am) (I) I” _ [14...] - 0.1 5 Inna-1 (0.1 n x 1120) g on 1 [tm)m,),r - learn. 5 '--’ u ‘0- ” . g ‘ 24-.- no _ ' fi‘ l u . I , fl :0 00 so so 100 no TD! (Hum) Figure 12. The hydrogenation of l-hexene (A) 0.6 M, (B) 0.7 M with [Rh(NBD)(PPh3)2]+-hectorite in methanol (0.2 wt % H20). 91 [1m]- OJ 5 mo“ Huh-ad (0.1 v: x :10) (“mum-3),] til-cunt. .0. 1cm 5 2: E ‘0- (A) N H oo_ 20- 24...- ' Y Y I 1 fl :0 co so so 100 rm 0am) zoo_ [141...] -‘1.0 5 (mummy; - mum. ' ”an ‘M E - : *1 S '“" oo- 0. I (n so- 20- 2m ' ' I I + I 7 fi 20 u so no 100 um (um) Figure 13. The hydrogenation of l-hexene (A) 0.8 M, (B) 1.0 M with [Rh(NBD)(PPh3)2]+-hectorite in methanol (0.2 wt % H20). 92 [1m] . 1.2 g 100 Mocha-o! (0.2 v: x 1110) W [Rh(IID) mush)" - lacuna Idiom. 8 ”-4 “can. E S .- E (A) to q 20W 24-... ' 1 I " r_ l l 20 oo 60 no 100 mu mm) 1” - [1“] . 0.1 ! Incl-o1 (0.2 n 1 I20) [manna-3),? 8 N- m E l-llomo (I) i 2*.- ‘0 .. I M -I I. 19 .1 I I 20 do 00 no 100 m M) Figure 1A. (A) The hydrogenation of l-hexene, 1.2 M in methanol (0.2 wt % H2O) with [Rh(NBD)(PPh3)2+— hectorite. (B) The hydrogenation of l-hexene, 0.1 M, under homogeneous conditions with [Rh(NBD)(PPh3)2]+ in methanol (0.2 wt % H2O). 93 RhH2(PPh3): and RhH(PPh3)2 on the surface of hectorite is a dynamic process, which also varies with the surface com- position. The rapid decline in the isomerization of l-hexene upon increasing the l-hexene concentration from 0.6 to 0.7 M and beyond indicates that there probably is a hysteresis involved which is related to the amount of l-hexene ad- sorbed on the surface is bulk concentration of l-hexene. It is unlikely that a diffusion controlled induction period is involved before isomerization of l-hexene occurs when the initial concentration of l-hexene is greater than 0.7 M. The catalyst is filtered at the end of the reaction and re- charged with fresh 0.8 M l-hexene, an identical reaction pattern was observed. Also, in the case of 0.8 M l-hexene, if the reaction is stOpped at 50% conversion, and the catalyst is filtered and recycled with 0.6 M l-hexene as a substrate,' an immediate isomerization was observed. The similarity in the hydrogenation reaction of l-hexene with [Rh(NBD)(PPh ]+-hectorite and Rh(PPh3);-hectorite 3’2 precursors suggest that catalytic species obtained from both systems should be almost identical. However, the re- action rates observed with Rh(PPh3);-hectorite are much lower than those obtained with [Rh(NBD)(PPh3)2]+-hectorite. The reason for the low rates with Rh(PPhB);-hectorite is the presence of excess free triphenylphosphine in solution, which can compete with substrate for coordination sites. 9“ Also excess ligand may limit the pore opening upon physical adsorption on the surface of hectorite. The effect of free triphenylphosphine on the rate of l-hexene hydrogena- tion was independently verified by carrying out the hydrogena- tion of 1.0 M l-hexene with the [Rh(NBD)(PPh3)2]+-hectorite precursor and excess triphenylphosphine so that PPh3th = 6:1. The rate of l-hexene reduction in this condition was “17 mL H min'l (mmol Rh)-l, which is nearlyljzoof that in 2 the absence of free triphenylphosphine. In contrast to [Rh(NBD)(PPh3)2]+-hectorite as a catalyst precursor, the hydrogenation of l-hexene with homogeneous [Rh(NBD)(PPh3)2]+ catalyst precursor occurs with extensive isomerization of substrate to internal alkene. The homogeneous hydrogenation of l-hexene at various initial concentrations of substrate with the [Rh(NBD)(PPh3)2]+ catalyst precursor in methanol (0.2 wt % water) is presented in Table ll. The initial rates of hydrogen uptake increased from m6“ to 3A0 mL H2 min-l (mmol Rh)-l upon increasing the initial concentration of l-hexene from 0.1 M to 1.0 M, respectively. The rate of hydrogen uptake, however, de- clines rapidly with increasing conversion of substrate. For example, at 0.6 M l-hexene the initial rate of hydrogen uptake was 315 mL H2 min-l (mmol Rh)‘l and then decreased to ~85 mL H2 min-l (mmol Rh).1 within 200 catalyst turn- overs. The homogeneous hydrogenation of l-hexene with [Rh(NBD)(PPh3)2]+ catalyst precursor was accompanied by 95 Table 11. Catalytic Hydrogenation of l-Hexene in Methanol (0.2 wt % Water) Under Homogeneous Conditions With [Rh(NBD)(PPh3)2]+-Cata1yst Precursor.a [l-Hexene] CatalySRD' c % Isomeriza- M l-Hexene/Rh Turnover Rate - tion 0.1 100 10(10%)e 57 18 16(16%) “8 37 23(23%) 27 59 25(25%) 13 60 0.6 600 60(10%) 301 _ 16 150(25%) 2A7 39 200(33%) 85 57 205(3A%) 61 59 1.0 1000 50( 5%) 296 7 156(16%) 133 32 2A8(25%) 71 51 320(32%) A7 60 a Rhodium loading was 0.72:0.0u wt %. Defined as moles of substrate hydrogenated per mol of Rh. 00" Rate of hydrogen uptake, mL H2 min—l (mmol Rh)-l. dFraction of total l-hexene isomerized to internal alkene. 8Values in parentheses give the fraction of total substrate that was hydrogenated. 96 extensive isomerization to 2-hexene. As an example, in the hydrogenation of 0.6 M l-hexene, at 90% conversion of the substrate, 32% was hydrogenated and 58% was isomerized to internal alkene. Significantly, the percent composition with time in the homogeneous condition is virtually in- dependent of initial concentration of l-hexene, Figures lAB, 15. The latter results suggest that in the absence of sur- face acidity a rapid equilibrium between rhodium dihydride and rhodium monohydride complexes can occur. The rapid decline in the homogeneous hydrogenation rate may be related to the fact that extensive isomerization of l-hexene pro- duces 2-hexene which is reduced at a much lower rate. This phenomenon is very similar to the one observed in the homogeneous hydrogenation of l-hexene with Rh(PPh3);- hectorite precursor.' l. The Effect of Water Content of Methanol on the Hydrogenation of l-Hexene with [Rh(NBD)(PPh3)21: Catalyst Precursor The surface acidity of smectite minerals are reported to depend upon the amount of interlayer water, which was reported to increase upon removal of interlayer water.80 Therefore, to obtain an indication of the effect of sur- face acidity on the equilibrium between rhodium dihydride and rhodium monohydride complexes, the hydrogenation of 97 (I) 1001 [1m] - 0.6 g Hothnnol (0.2 we 2 a,» Immunity; - lit-ma. ,3. so. 2 14!..- E ._ A ......... N v \. 60- 10.. l l :r l l to to so no 100 m m) (I) 1'” '1 [14...] - 1.0 ! ”I (0.3 fl 2 '20, (ball!) (Pligzro w - \ i Figure 15. The hydrogenation of l-hexene (A) 0.6 M, (B) 1.0 M, under homogeneous condition§ with [Rh(NBD)'(PPh3)2]+ in methanol (0.2 wt % H20). 98 l-hexene was carried out with the homogeneous and inter- calated [Rh(NBD)(PPh3)2]+ catalyst precursors in methanol solution containing 0.1, 0.2, 0.5 and 1.0 wt % water. The initial concentration of l-hexene in each case was maintained at 0.8 M. This concentration was chosen in part because a minimum amount of isomerization of l-hexene was observed when the water content is 0.2 wt %. Any dependence of product distribution on water content may be linked to the change in surface composition and, hence, surface acidity. The results of these experiments for the hydrogenation of l-hexene is provided in Table 12. At a water content 0.1 wt % in methanol, the rate and product distributions are similar to those obtained when water content of methanol was 0.2 wt %. However, when the water content was increased to 0.5 and 1.0 wt %, a significant reduction in the rates of hydrogen uptake was observed. For example, the rate of l-hexene hydrogenation was 85 mL H2 min-l (mmol Rh)-1 at 1.0 wt % water in methanol and ~530 mL H2 min.1 (mmol Rh).l when the water content was 0.1 wt %. Also, the isomerization of substrate to internal alkene accompanied hydrogenation. At 50% conversion of substrate, 25% of the substrate was hydrogenated and 25% was isomerized to 2- hexene when the water content of the solution was 1.0% (Figure 16,17). Once all of the substrate was consumed, some desorption of rhodium complex (<5%) was observed, when the water content was 0.5 and 1.0 wt %. Table 12. 99 Catalytic Hydrogenation of l- -Hexene, Under Intercalated and Homogeneous Conditions with [Rh(NBD)(PPh3 )2 ]+- -Catalyst Precursor in Methanol Containing 2Varying Amounts of Water. 0. 8 M a, b % Water in Catalyst % Isomeri— Catalyst System Methanol TurnoVer Rate zation [Rh(NBD)(PPh3)2] - 0.1 u0( 5%) 593 <1 Hectorite 200(25%) 553 ocpze a uco>aom Eoumzm umhadpmo IHLoEomH a pmzHMQmo mmocoxmmlau .oonnooona onsamomou+mmflmnmmvAamzvnmm msoocoonom pom popmHmopoch spas ocoxomla mo cofipmcowopcmm oapzamumo .ma manme 106 .Uopmcowoppmz mm: umzp mumpquSm HmpOp mo coauowpm on» o>fiw momonucopmd CH mozam>c .u us m.o mm: Hocmcpoe mo pcmpcoo Loumzw .nm a as mo.oams.o ma mnaomoa engage.H .cofipmNHHmEpoc Locum ohm mosam>o .ocoxam Hmcpmucfi ou ooNfiLoEomfi ocoxonla HmpOp go soapompmp .nm HoEE\:HE\Nm QE .oxmuaz Comopozn mo mummo .cm mo HoE pom oopmcomopozn mushpmnzm mo moHoE mm Bonamom n .ofipmp nm\ocoxo£|H map o>Hm momenpcopmo CH mosaw>m "monocuoom .ooocaocoo .MH manna 107 period may be related to the kinetics of l-hexene coordina- tion with metal center. The intercalated catalysts can be used even when the solubility of the metal complex prohibits its use under homogeneous reaction conditions. For example, the hydrogena- tion of l-hexene (0.8 M) with [Rh(NBD)(PPh ]+-hectorite 3)2 can be carried out in hexane as a solvent. However, the rate of hydrogenation was very low (Table 13). To carry out the same reaction under homogeneous conditions in hexane is not practicable due to the negligible solubility of the cationic rhodium complex in hexane. 2. Hydrogenation of 1-Hexene with [Rh(NBD)(Dip_hos)]+ Catalyst Precursor Additional evidence that changes in surface composition are responsible for the observed differences in the re— activity of the homogeneous and intercalated catalysts was obtained by substituting the monodentate triphenylphosphine ligands in the catalyst precursor with a bidentate ligand, bis(diphenylphosphino)ethane. The replacement of the mono- dentate ligands by a bidentate ligand in the rhodium complex forces both phosphorus atoms to occupy gig positions. Con— sequently, the hydrogenation mechanism differs significantly from that of the hydride mechanism.77 It is proposed that in the presence of a bidentate ligand, the equilibrium between dihydride and monohydride rhodium complexes does 108 not occur. Hence, a change in surface composition with variation in initial concentration of substrate should not effect the product formation in hydrogenation of l- hexene. . The results of l-hexene hydrogenation with [Rh(NBD)- (diphos)]+ under homogeneous and intercalated catalyst precursor are provided in Table 1A. In homogeneous solu- tion, an extensive and very rapid isomerization of substrate to 2-hexene was observed. For example, at 80% conversion of substrate, 68% was converted to 2-hexene and 12% was hydrogenated to hexane. The initially high rate of 233 mL H2 min"l (mmol Rh).l drOps off rapidly to 76 mL H2 min-1 (mmol Rh)"l within 60 turnover of catalyst. It is sig— nificant that the initial rates of hydrogen uptake or 1- hexene isomerization are almost independent of the initial solution composition (Figure 19, 20). Table 1“ contains the results of l-hexene hydrogenation with [Rh(NBD)(diphos)]+-hectorite as the catalyst precursor. In each case a brief, 10 min, induction period preceded hydrogen uptake. The initial rates of hydrogen uptake were slow, but the rate attained an optimum value within 20% con- version of substrate and remained constant until almost all the substrate was converted to products. For example at 0.N M initial concentration of l-hexene the optimum rate of hydrogen uptake was 1“ mL H2 min"1 (mmol Rh)-l. The op- timum rates of l-hexene hydrogenation increased from 1A 109 ms ms snmsvows em ms snmmvsss sm ms smwmvmss mm ss snomvom soosv ooosnoooom ss 0 smm Vow 2.0 u+ssmososovsmmzvnmu mm sm Anmsvmom so ss smmmvmms ss so smmmvmsm we sos sssmvomm ss mas sassvmms soomsv as osm saw Vss a m.s om ms ssosvomm no sm sssmvssm ms om ssmmvoom ms Ass snmsvmms soomv om mmm Ann Von m.o mm m Amwmvoms ms mm . smmmvmss ow w: Amomvom cospzaom mm oms Amosvos soosv nooocowosom om mes exam Vsm s.o -+sanonosovsomzvnmu UCOsumN oumm nmo>o:pse a Eopwzm ummampmo IHLoEomH m pmhaMpmo mmocoxmmlag .cHOm ILSOosm pmzsmpmol+msmo:QHpvsomzvzmg cuss mCOfipHocoo pmpmamosoucH paw msoozow noEom pops: Aomm & p3 m.ov Hocmnpoz CH ocoxomna mo soapmcowosozm ospzsmuwo .JH mHQmB 110 .Uoumcowoppzc was was» mumspmosm Hmpou mo cospompm on» o>fiw mononpcopmo CH mozsm> .ocoxsm swcpopcfi on Ummssosoms ocmxonla HmpOp mo soapOMLm .nm mo oHoE pod coumcowosomn manppmnzm mo moHoE mm pocfimoo o .n n: so.owms.o no: mcsooos assoonmo U .nm HoEE\£HE mm qE oxmuas cowompzc mo mummo a .£m\o:oxosus o>fiw momoanoQMQ as mossm>m mm as snmsvmsm mm mm ssmsvmsm am am snmmvmms sm mm ssmmvomm sm mm Anmsvmmm as as ssmsvoss soomsv as as Asm Vow m.s s: as snmsvosm mm mm snmmvoom mm mm ssomvosm mm sm ssmsvsms ms ms Asmavmm soowv oopspouoom m ms ssm me m.o n+sflnocosovsamzvnmu ocoapmu oopmm naw>0cs39 a Emummm pmzHMpmo IssoeomH & ummsmpmo mmocoxomnsg .ooocsocoo .ss osoms Figure 19. 111 (I) [baa-o] - 0.0 g M (0.2 R I I’D) [naunuwunflm—m—«o 100- “XMMU ' I 1 I 100 200 100 4.00 300 600 700 m am) 100 s (I) [14...] - 0.0 5 . nun-1 (0.: n 1 I10) 5 00.1 [umummn‘ - W E 24-. N The hydrogenation of l-hexene, (A) 0.A M, (B) 0.8 M, under homogeneous conditions with [Rh(NBD)(Diphos)]+ complex in methanol, (0.2 wt % water). Figure 20. 112 1“ [1cm] - 1. 2 N 30th.). (0.1 It I a 20) (”(MHMM)I’- 2w (A) ill. I MHIC 1, 700 100 " (14-...) - 0.. a 10th.! (0. 2 it I I 20) E n [momma-n)?- Rooter". N 1 1m (A) The hydrogenation of l- -hexene, l. 2 M, with [Rh(NBD)(Diphos)]+ under homogeneous condi- tions in methanol (0. 2 wt % H (B) The hydrogenation of l- -hexene, 0.3m M, with [Rh(NBD)- (Diphos)]+ -hectorite in methanol (0. 2 wt % H20). 113 to 22 and 39 mL H2 min_l (mmol Rh).l upon increasing the initial concentration of l-hexene from 0.“ to 0.8 and 1.2 M respectively. The rate dependence on concentration may be accounted for by the mass action on the transfer of sub- strate from solution to the interlayer region of the min- eral. It is noteworthy that the intercalation of the catalyst has significantly reduced the isomerization of substrate to 2-hexene. For example, at 50% conversion of 0.4 M l-hexene, 28% was isomerized to 2-hexene and 22% was hydrogenated to hexane. The product distribution is almost independent of the amount of l-hexene initially present in solution (Figure 20B, 21). Relative to homo- geneous reaction, the hydrogenation of 2-hexene with the intercalated catalyst is much slower. . The effect of catalyst intercalation is most evident in the hydrogenation of l-hexene with [Bh(NBD)(PPh3)2]+ as a catalyst precursor. The difference observed between the intercalated and homogeneous catalysts, most likely is related to the effect of silicate surface composition on the equilibrium between RhH2(PPh3); and RhH(PPh3)2. The effect of catalyst intercalation on this equilibrium is mostly due to the surface Br6nsted acidity and its variation with interlayer composition. The surface Br6nsted acidity of layered silicate arises mainly from the hydrolysis of interlayer exchange cations. The principal exchange ion of the layered silicate, in Figure 21. 1114 (I) 1001 [14...] - 0.0 5 loch-l (0.1 .. 1 0,0) [human-u)? - locum. 12M 15M 10” (I) - [14...] . 1.1 ! nun-.1 (0.: u. 1 I10) [nu-manna)? - We. The hydrogenation of l-hexene, (A 0.8 M, (B) 1.2 M, with [Rh(NBD)(Diphos)] -hectorite in Methanol (0.2 wt % H20). 115 this study, is Na+ and its hydrolysis is due to the pres- ence of small amount of water in methanol. Although the hydrated Na+ ion is very weakly acidic in solution, its activity is greatly enhanced on negatively charged layered silicate surfaces. The surface acidity is proposed to depend upon the nature of the cation and the charge on the sili- cate sheets. Also, the surface acidity of the hydrated cations is greatly affected by the amount of water present in the interlayer regions. For example, if the amount of interlayer water is reduced in the presence of a given cation, enhanced surface acidity is observed. The increase in sur- face acidity is manifested by an increase in NH: formation when the mineral exposed to NH3 vapor.80b The polarizing effect of the exchangeable cation increases as the number of coordinated water molecules around it decreases upon dehydration. When a great deal of water is present, the polarization forces of the exchangeable cations may be distributed among a large number of water molecules, and the acidity of such a system might approach that of the ions in an aqueous solution. However, as the water content is decreased polarization forces become more concentrated on the fewer remaining water molecules, causing an increase in hydrolysis and in their proton donating abilities.80b The differences observed in the catalytic activity of homogeneous and intercalated [Rh(NBD)(PPh3)2]+ catalyst precursor can best be deduced by comparing the product 116 distribution at 50 and 90% conversion of 1—hexene, (Table 15). Regardless of the initial concentration of substrate, the percent hydrogenation at 50 and 90% conversion are consistently higher under intercalated catalyst conditions than under homogeneous conditions. Therefore, based on reaction mechanism proposed by Schrock and Osborn,77 the catalytic hydrogenation catalyst probably occurs by the reaction of substrate with [RhH2(PPh3)2]+. The dramatic dependence of product distribution, at 50% conversion of substrate, on initial concentration of l- hexene indicates that the equilibrium that affects the intercalation catalyst, takes place immediately upon the addition of substrate to the reaction system. The absence of an induction period precludes the diffusion of substrate in the interlayers as an important factor. It is possible that at higher initial concentrations of 1-hexene (30.7 M) the intercalated catalyst experiences dielectric effect that is significantly different than at lower concentration (30.6 M) of substrate. The dielectric effect may change considerably along with the consumption of substrate. As discussed earlier, the surface acidity and the equilibrium between [RhH2(PPh3)2]+ and RhH(PPh3)2 may depend on surface composition. The origin of the dielectric ef- fect and the position of the dihydride-monohydride 117 so mm mm mm . o.s so mm mm mm . v.0 mm mm om 0m H.o QoHpsaom mzoocoonom pm . ms om on m.s em ms mm mm o.s sm as :2 mm m.o mm as :8 mm m.o m: mm mm a: 0.0 mm mm o: oo 2.0 ousaouoon om om ms em s.o u+smsmcmmvsomzvnmw cosumN COsp cospmm soap a Eopmzm IHLoEomH R cocowoppmm R IHLoEomH R lacowosozm & flocoxonlsu cosmpo>soo Rom coamso>coo Rom . .somm a p: m.ov socmcooz cs oomnooonm nonsmomou+smsmzmmvAmmzvnm. ooomsmoooncs one msoocoono: cps: ocoxomls mo :0spQComopozm on» now cosponsppmsm poscopm .ms osnme 118 equilibrium may be related to the adsorption of substrate in the interlayers. The adsorption of 1-hexene on the interlayer surface should replace some of the interlayer water and solvent. The replacement of some of the solvent and water molecules can lead to higher polarization of the remaining molecules by interlayer cations, and thus en- hance surface acidity. However, as the substrate is con- sumed the above process may be reversed, allowing formation of RhH(PPh3)2 which can initiate the immediate isomerization of 1-hexene to internal alkene. At higher initial concentrations between 0.7 and 1.2 M l-hexene the polarization forces in interlayers may be such that [RhH2(PPh3)2]+ is primarily preferred in the interlayer regions. Consequently, very little isomeriza- tion of substrate occurs until 50% conversion. At 50% conversion of 0.8 - 1.2 M l-hexene, the effective concen- tration of 1-hexene in solution should be 0.A - 0.6 M. At this point the isomerization was observed. As a result, significant internal alkene was observed at 90% conver- sion. The hydrogenation of l-hexene at concentrations between 0.1 and 0.6 M is accompanied by isomerization of substrate to internal alkene. The product distribution at 90% conversion indicates that the rates of isomerization in all cases are significant. These facts suggest that the isomerization occurs with a catalyst species which is identical to that of homogeneous solution, most probably 119 the neutral RhH(PPh3)2 species. A comparison of the percent composition at 50 and 90% conversion of l-hexene under homogeneous conditions with that of intercalated catalyst, at various initial concentra- tion, indicated that in the absence of strong polarization effects (surface acidity) the product distribution remains essentially identical in each case (Table 15). It appears that under homogeneous conditions an immediate equilibrium between [RhH2(PPh3)2]+ and RhH(PPh3)2 must occur so that the catalytically more active neutral RhH(PPh3)2 complex de- termines the solution products. Consequently, isomeriza- tion of substrate is nearly twice that of hydrogenation. Strong support in favor of surface acidity effects in reaction of l-hexene was provided by the effect of water present in methanol under intercalated catalyst. Increasing the amount of water in the interlayers (0.5 1 1.0'wt % water in methanol) caused isomerization of 0.8 M l-hexene even at the beginning stages of reaction. As a result, at 50% conversion of substrate the percent of both isomeriza- tion and hydrogenation products in solution were almost identical (Table 16). This indicates that even at higher concentration of l-hexene (0.8 M) the effective surface acidity is determined by water molecules in the inter- layers. However, at low water content (0.1 and 0.2 wt %) the surface acidity is mainly determined by adsorption of l-hexene on the surface and by the percent conversion Am 0 .H mm3 mQGNGSIH pHO COHpMcHPCmOCOO HQHPHQH mmwo mHLU CH .I . x .2 ®.O mm3 ®Q®K®EIH .HO COHPMhuflmofloo HMHHHQH mmmmo HHm GHQ 120 om. om mm sm o.s m: mm mm mm m.o mm mm mm mm *m.o cofiu3som msoozomoEom :m we om om o.s Hm mm :3 mm m.o am mu 3 mm m.o oufiLOpoo: em on 3 mm s.o u+smsmnaovsomzvnmu cosumN cos» coaumu soap Loam: & Eoummm. IsLoEomH & Imcowosnzm N IHLoEomH R Imcowompzm R cosmpo>coo Rom cosmpm>coo Rom m .poumz no mpcsofi< mcfimpm> wCHCHmpcoo Hocmzpoz as somLSOohm ummawuwol+mmsmcmmvAamzvnm. popmswohoucH now msoocoonom nus: ocoxomns mo cospwzowOLUmm 059 Low soapznspumso posoopm .ms magma 121 of substrate. The linearity of hydrogen uptake and the absence of any dependence of 001 X-ray basal spacing on l-hexene concen- tration, in the hydrogenation of 1-hexene with [Rh(NBD)- (PPh3)2]+-hectorite, suggest that the effect of inter- calation on catalyst is most probably a consequence of surface composition and not of any apparent phase change during reaction. The results of l-hexene hydrogenation with homogeneous and intercalated'[Rh(NBD)(diphos)]+ catalyst precursor suggest thatthe hydrogenation of the terminal alkene with this catalyst is occurring by a mechanism which is dif- ferent from'the mechanism which operates with RhH2(PPh3): or RhH(PPh3)2. The reaction mechanism for the hydrogena- tion of olefin by [Rh(NBD)(diphos)]+ catalyst precursor is expected to follow the coordinatively unsaturated pathway as proposed by Schrock and Osborn.77 According to this pathway the olefin coordinates to [Rh(diphos)]+ prior to the oxidative addition of the hydrogen molecule. In this situation almost no or very little isomerization of terminal alkene is expected. However, the extensive isomerization of substrate observed in homogeneous solution as well as with the intercalated catalyst suggest that [RhH(diphos)]2+ is the most likely catalytically active species present in solution. According to Halpern gt al.103 [RhH(diphos)]2+ can occur on oxidative addition of one hydrogen to 122 [Rh(diphos)]+ prior to olefin coordination to the rhodium center. The formation of [RhH(diphos)]2+ may be respon- sible for the isomerization of the olefin by a mechanism analogous to that proposed for RhH(PPh3)2. Schrock and Osborn have shown that the rates of homogeneous hydrogena- tion of diolefins with [Rh(NBD)(diphos)]+ were apparently not affected by addition of acid or base, which substan- tiated the unsaturated path mechanism proposed to explain hydrogenation of diolefins. It was expected that if un- saturated pathway is acting in the hydrogenation of l- hexene with [Rh(NBD)(dppe)]+ then the addition of a small amount of acid or base should have very little effect on the performance of catalyst in homogeneous solution. How- ever, the homogeneous hydrogenation of l-hexene with [Rh(NBD)(dipnos)]+ in the presence of 0.07 M H010“ or 0.07 M NEt3 occurs at lower rate of hydrogen uptake and with de- creased isomerization. If [RhH(diphos)]2+ is presumed to be an active species, then the presence of H010” should inhibit the formation of RhH(diphos)2+. Therefore a lower rate of l-hexene hydrogenation as well as a lower rate of isomerization is expected. However, the effect of tri- ethylamine, which should actually enhance both the formation of [RhH(diphos)]2+ and isomerization is not expected. It is possible that the difference observed for the hydrogenation of l-hexene with [Rh(NBD)(diphos)]+ under homogeneous and intercalated conditions most probably is a 123 combination of electronic effect due to the negatively charged silicate layers and a steric effect due to the constrained interlayer environment. These studies strongly indicate that the effect of catalyst intercalation on the catalytic activity of [Rh(NBD)— (PPh3)2]+, arises mainly from surface composition and surface acidity effect. However, an intrinsic stabiliza- tion of rhodium dihydride complexes in the interlayers can also play some role. It is possible that mineral prefers binding cationic rhodium dihydride complex over the dis- sociation product proton as shown in Equation 3. ~ - - - ~ - - - - - - - RhH2(PPh3)2 + RhH2(PPh3)2 + RhH(PPh3)2 + H ‘j RhH