AmericanChemicalSociety.com

        Bismuth has an extremely diverse chemistry with applications in materials science, medicine, organic synthesis and catalysis, among others.  With respect to the field of catalysis, arguably the single-most important process is the selective oxidation/ammoxidation of propene to make acrolein and acrylonitrile over a multi-component catalyst based on Bi₂O₃/MoO₃, known more commonly as the SOHIO process.  Both acrolein and acrylonitrile are currently used in industry on large scales.  It was recently reported that, in the year 2000, the worldwide demand for acrylonitrile alone was approximately 5,000,000 tons.

      Despite the commercial success of the SOHIO process and considerable subsequent research, the exact mechanism by which the catalyst is able to facilitate the observed chemical transformations remains unclear.  We feel that additional insight into the mechanism of functioning of the SOHIO oxidation catalyst could be gained through the investigation of molecular complexes whose geometry and bonding arrangements mimic what is seen in the bulk material.  Unfortunately, to this date, molecular Bi/Mo complexes remain scarce, with species that contain structural features associated with the bulk SOHIO oxidation catalyst (including Bi-OH, Bi-O-Bi, Mo-O-Bi and Mo=O moieties) being particularly elusive.

        Our work on this system for the time period served by this grant has largely been devoted to two overall goals: (1) Develop a viable, general synthetic route to the formation of high nuclearity heterobimetallic coordination complexes and (2) exploring the ability of heterobimetllic Bi/transition metal coordination complexes to model key features in the proposed mechanism for the SOHIO process.  I would like to use this communication as a chance to highlight my group's recent progress toward achieving both of these goals.

        Our group has developed a novel, direct synthetic route for the formation of discrete high nuclearity heterobimetallic coordination complexes that contain both bismuth and molybdenum or tungsten.  Our synthetic approach relies on the interaction of nucleophilic transition metal fragments with a preformed, robust polybismuth core derived from Bi₉(O)₈(OH)₆⁵⁺ ([1]⁵⁺).  This strategy has lead to the successful isolation and characterization of a number of complexes, including the undecanuclear species ((TMTACN)MO₃)₂(Bi₉(O)₈(EO₄)₃)(ClO₄)₅ 4CH₃CN (M = Mo ([2]⁵⁺), W ([3]⁵⁺)) in excellent isolated yields. 

        In an attempt to better tailor and generalize our synthetic strategy, a detailed analysis of the interaction of the cationic nonanuclear bismuth core with the (TMTACN)MO₃ (M = Mo, W) transition fragment has been undertaken.  It is interesting to note that, in the case of the neutral (TMTACN)MO₃ based complexes, the 2:1 adducts [2]⁵⁺ and [3]⁵⁺ are formed preferentially.  Indeed, all of our attempts to generate the lower nuclearity 1:1 complexes have only resulted in the identification and isolation of [2]⁵⁺ or [3]⁵⁺.  As part of this study, we have also explored the possibility of producing higher nuclearity species.  In this case, we have successfully obtained spectroscopic evidence for the formation of the 3:1 dodecanuclear complex ((TMTACN)WO₃)₃(Bi₉(O)₈(EO₄)₃)(ClO₄)₅ in solution.  Unfortunately, this complex has, so far, eluded successful isolation and structural characterization.

        The specificity of the synthesis for the 2:1 adducts is surprising.  The results of single crystal x-ray diffraction experiments on these two complexes suggests that there is relatively little structural perturbation in the metal-oxygen framework, so significant changes in the steric constraints of the molecules seems unlikely.  However, the apparent preferential formation of complexes [2]⁵⁺ and [3]⁵⁺ at the expense of lower nuclearity cogners, suggests that the 1:1 intermediate species is more reactive to subsequent reaction and adduct formation than is the native nonanuclear bismuth core.  We are attempting to test this possibility by modeling and comparing the relative electron distribution in the base bismuth core as well as in the 1:1 and 2:1 adducts.  The apparent lack of stability of the 3:1 adduct might stem from unfavorable steric interactions between the triazacyclononane ring on the transition metal fragment and the alkoxo ligands on the bismuth core.   

        We have been able to displace one of the (TMTACN)MoO₃ groups from [2]⁵⁺ through the slow addition of hydroxide ion.  In this case, the complex that forms is the polymeric species [{(TMTACN)MoO₃}(Bi₉(O)₈(EO₄)₃)(OH)](ClO₄)₄. 

        We were surprised by the ability of the nonabismuth framework to tolerate the introduction of the strongly nucleophilic hydroxide ligand without significant changes to the overall geometry.  In light of this result, we are currently exploring the possibility of improving the stability the heterobimetallic coordination complexes by increasing the strength of interaction between the transition metal fragment at the bismuth core.  In this case, we are exploring the use of the more strongly nucleophilic complexes Cp*MoO₃^- and Cp*₂Mo₂O₅^2- for the formation of heterobimetallic species. 

        An additional point of significant focus in this work has been to tailor the organic ligands that stabilize and solubilize the heterobimetallic species that we are developing so that the resulting complexes more accurately model the composition of some of the postulated key intermediate steps in the SOHIO process.   In this case, we have successfully produced heterobimetallic Bi/Mo complexes bearing allylic and amine functional groups through direct modification of the nonanuclear bismuth core.  The complexes that we have produced in this manner include the recently isolated species [((TMTACN)MoO₃)₂(Bi₉(O)₈(L)₆)](ClO₄)₅ (L=OCH₂CH₂OCH₂CH=CH₂).  Complete characterization of these compounds is currently ongoing.

        A major point of consideration in this aspect of the project has been to try to move some of the introduced functional groups from the bismuth to the transition metal coordination sphere.  This focus stems, in part, from the widely accepted argument that the transition metal is the locus of reactivity for the oxidation and ammoxidation of propene.  Consequently, we are currently investigating the direct modification of the transition metal fragment in [2]⁵⁺ and [3]⁵⁺ as well as exploring the grafting of transition metal complexes that bear more labile ligands, including halide and amide moieties.

      We are pleased with the success that we have enjoyed to this point in the project.  We are currently in the process of finalizing the characterization of a number of the complexes that are described here and look forward to publishing our work in the near future.  We gratefully acknowledge the American Chemical Society Petroleum Research Fund for their support of this project.