J. Phys. Chem. B 2006, 110, 4125-4134
4125
Adsorption at the Liquid-Liquid Interface in the Biphasic Rhodium Catalyzed Hydroformylation of Olefins Promoted by Cyclodextrins: A Molecular Dynamics Study N. Sieffert and G. Wipff* Laboratoire MSM, UMR CNRS 7177, Institut de Chimie, 4 rue B. Pascal, 67 000 Strasbourg, France ReceiVed: December 2, 2005; In Final Form: January 11, 2006
Using molecular dynamics (MD) simulations, we investigate the interfacial distribution of partners involved in the phase transfer rhodium catalyzed hydroformylation of olefins promoted by β-cyclodextrins (β-CDs). The β-CDs, the reactant (alkene), product (aldehyde), several rhodium complexes (the catalyst, its precursor, and its alkene adduct) are simulated at the water-“oil” interface, where oil is represented by chloroform or hexane. It is shown that unsubstituted β-CD and its 6-methylated and 2,6-dimethylated analogues adsorb at the interface, whereas the liposoluble permethylated CD does not. The precursor of the catalyst [RhH(CO)(TPPTS)3]9- (with triphenylphosphine trisulfonated TPPTS3- ligands) sits in water, but the less charged [RhH(CO)(TPPTS)2]6- catalyst and the [RhH(CO)(TPPTS)2(alkene)]6- reaction intermediate are clearly surface active. The TPPTS3- anions also concentrate at the interface, where they adopt an amphiphilic conformation, forming an electrical double layer with their Na+ counterions. Thus, the most important key partners involved in the hydroformylation reaction concentrate at the interface, thereby facilitating the reaction, a process which may be further facilitated upon complexation by CDs. These results point to the importance of adsorption at the liquid-liquid interface in the two-phase hydroformylation reaction of olefins promoted by β-CDs and provide microscopic pictures of this peculiar region of the solution.
Introduction The biphasic rhodium catalyzed hydroformylation of olefins plays a key role in industrial processes, like the Ruhrchemie/ Rhoˆne-Poulenc production of alcohols from propylene, working under mild conditions and efficiently separating the expensive rhodium catalyst (in the aqueous phase) from the reactants and products (organic phase).1,2 The reaction, based on rhodium complexes with hydrosoluble ligands (typically triphenylphosphine trisulfonated (TPPTS3-) anions), presumably takes place in the aqueous phase where some fraction of the reactant is dissolved. Higher olefins, however, are too insoluble in water to react with the catalyst, requiring the use of, e.g., cosolvents,3-8 surfactants,3,9 amphiphilic ligands,10 or cyclodextrins.11 Recently, Monflier et al. showed that the rates of catalized biphasic reactions such as the hydroformylation,11,12 Wacker oxidation13,14 or hydrocarboxylation of olefins,15 and the hydrogenation of aldehydes16 can be considerably enhanced in the presence of chemically modified cyclodextrins (CDs) acting as mass transfer promoters, i.e., allowing hydrophobic substrates to react with the hydrophilic catalyst (Figure 1). CDs are cyclic oligosaccharides with six (R-), seven (β-), or eight (γ-) D-glucose units which display a hydrophobic cavity into which hydrophobic guest molecules (e.g., an alkene substrate but also “poisoning species” like the reaction products or phosphine ligands)17 can be complexed. Native CDs with secondary 2-OH and 3-OH groups at the wide rim (secondary side) and primary 6-CH2OH groups at the narrow rim (primary side) (Figure 2) are water soluble, but functionalization of the -OH functions (e.g., methylation) allows their solubility to be modified in both organic and aqueous phases as well as their complexation properties and, as a consequence, their efficiency as promoters in biphasic systems. For instance, methylated R-CDs18 or *
[email protected].
Figure 1. Rhodium catalyzed hydroformylation of water-insoluble olefin in the presence of cyclodextrin (represented by a truncated cone). This figure is adapted from ref 71. The generally accepted mechanism of the hydroformylation reaction (without CD) is given in Figure S1.
β-CDs11 are more effective promoters than the native ones, but the reasons are unclear. Although the importance of interfacial phenomena in phase transfer catalysis is well recognized,19-21 little is known as far as the distribution of reaction partners and the microscopic nature of the interface are concerned. Important questions concern the size and characteristics of the “interfacial layer” (as schematized, e.g., in Figure 1) and the distribution of the different species involved in the phase transfer catalyzed reaction. In particular, how the substrate (alkene) which is highly hydrophobic can meet the catalyst in the aqueous phase? Do these species concentrate in the interfacial region, or are they instead “repelled” by the interface? Why do CDs promote the reaction: do they act as “shuttles” transferring the alkene to the aqueous phase where the reaction would proceed
10.1021/jp057023q CCC: $33.50 © 2006 American Chemical Society Published on Web 02/07/2006
4126 J. Phys. Chem. B, Vol. 110, No. 9, 2006
Sieffert and Wipff described by a sum of bond, angle and dihedral deformation energy, and pairwise additive 1-6-12 (electrostatic and van der Waals) interactions between nonbonded atoms.
U)
∑ Kl(l - l0)2 + angles ∑ Kθ(θ - θ0)2 + ∑ ∑Vn(1 + cos(nω - φ)) + dihedrals n
bonds
Figure 2. Schematic representation of β-CD, native (R1 ) R2 ) R3 ) H), 6-methyl (R2 ) R3 ) H; R1 ) CH3), 2,6-dimethyl (R3 ) H; R1 ) R2 ) CH3), and permethylated (R1 ) R2 ) R3 ) CH3) forms. This figure is reproduced with permission from ref 71. Copyright 2001. Royal Society of Chemistry.
Figure 3. Triphenylphosphine trisulfonate (TPPTS3-), the [RhH(CO)(TPPTS)2]6- catalyst, its [RhH(CO)(TPPTS)3]9- precursor, and [RhH(CO)(TPPTS)2(decene)]6- reaction intermediate, simulated with Na+ as counterions.
and the aldehyde back to the organic phase? Ultimately, where does the hydroformylation reaction takes place: in a locally heterogeneous mixed water-oil domain or “right at the interface”? Computer simulations contribute to our understanding of the nature and characteristics of liquid-liquid interfaces,22,23 with relevant applications in phase transfer catalysis such as the distribution of salt catalysts,24 or the “catalytic effect” of the interface in a SN1 dissociation reaction.25 Following our investigations on assisted ion extraction,26,27 we decided to undertake a molecular dynamics (MD) study on the interfacial behavior of the species involved in the biphasic hydroformylation of higher olefins, using 1-decene as a substrate and β-CDs as mass transfer promoters (Figure 1). More specifically, we first focus on β-CD alone, comparing its native form with different methylated analogues. We then consider the different partners of the reaction: the reactant (1-decene), model product (decanal; note that it contains one carbon less than the real reaction product with decene), and different rhodium complexes (Figure 3, the [RhH(CO)(TPPTS)3]9- precursor of the catalyst, the [RhH(CO)(TPPTS)2]6- catalyst itself, and the [RhH(CO)(TPPTS)2(alkene)]6- complex, a key reaction intermediate resulting from the coordination of the alkene substrate to the catalyst).28 The TPPTS3- ligands which are in situ precursors of the catalyst will be studied as well, neutralized in all cases by Na+ counterions. Insights into the effect of β-CDs on the distribution of these species will also be obtained by modeling S⊂CD complexes (where the inclusion symbol “⊂” means that the substrate S is included inside the cavity of the CD), using native β-CD as the host (Figure 1). In most cases, the organic phase will be modeled by chloroform for convenience, but tests with hexane will also be performed in order to mimic experimentally used solvents such as undecane.18 Methods 1. Energy Representation of the Systems. The molecular dynamics (MD) simulations were performed using modified AMBER 7.0 software29 where the potential energy U is
∑ i
