Reports: AC3
48454-AC3 Simple Metalloporphyrin Dimers for Multielectron Reduction Catalysis
During the reporting period we have:
1. Scaled up syntheses of the proposed porphyrins (free bases) to ~10 g.
2. Synthesized and fully characterized spectroscopically (NMR, UV-vis, MS, EPR and cyclic voltammetry) complexes of these porphyrins with Fe and Co.
3. Demonstrated the remarkable stability of the Co-porphyrin dimers (only the low limit of the dissociation constant (<5 nM) could be determined; Fe analogs have been measured previously and have Kd ~1 uM). The Co dimers remain intact even during MS analysis in air! Note that such high affinity is counterintuitive: the affinity of 4-coordinate high-spin Co(II) to a 5th ligand is typically <1. Unlike simple Co porphyrins, our derivatives manifest very high O2 affinity, supporting our original assertion that enforcing the (unfavorable) mono-imidazole ligation of a Co(II) site would have a greater effect on its O2 affinity (and presumably other aspects of its reactivity with O2 and NO) than that for the extensively studied Fe analogs.
4. Established (in collaboration with Marc Koper's group at Leiden University, Netherlands, as proposed) that the dimers remain intact when adsorbed on edge-plane graphite and display unusual electrocatalytic behavior towards nitrite.
5. Established the feasibility of quantum-chemical calculations on these dimers at the DFT level with double-zeta quality basis sets.
The primary unresolved challenges that became apparent during this period include:
1. Free-base syntheses: while we succeeded at improving the yields to ~10% (which is quite impressive for asymmetrical porphyrins), we have yet been unable to increase them further and have to continue to rely on chromatography (rather than recrystallization as originally proposed) to separate asymmetric and symmetric porphyrins. Although we developed a streamlined chromatographic separation protocol, which relies on significant difference in polarity of different porphyrins in the mixture and even automated it, we anticipate it to pose challenges for practical application of these porphyrins. We will therefore continue to optimize the syntheses and screen crystallization conditions.
2. Metallation: as we anticipated, metallation of the free bases with Fe and Co is very facile and proceeds with high yields. Yet metallation of the free bases with 2nd- and 3rd-row transition metals (we are particularly interested in Mo for catalysis of redox chemistry at N, as proposed and Ru for catalytic oxygenations) appears to be more challenging that of any other porphyrin I worked with. In contrast to simple porphyrins lacking appended nitrogen heterocycles, whose metallation with Mo and Ru is, slow but straightforward, applying the same conditions to our porphyrins yielded intractable mixtures even when the metallation was attempted at room temperature (with MoOCl2, as previously described by me: Angew. Chem. Int. Ed., 2001, 40, 12711274). We do not understand the reactions that lead to degradation of the porphyrin chromophore during such syntheses and at present are unsure about how to proceed.
3. Electrocatalytic studies: we have been unable to raise additional funds to buy the electrochemical setup and are therefore collaborating with the experts in electrochemistry (Marc Koper's group at Leiden University, Netherlands). This group is primarily interested in redox catalysis of N species (instead of O2 reduction that is a focus of our work). Consequently, the progress of this part of the proposal was slower than we anticipated.
4. Computational studies: while we encountered few of the feared SCF convergence problems, geometry optimizations on these molecules require fairly large amounts of cpu time. Unfortunately, our most recent request for additional cpu time on Teragrid supercomputers has been declined. We are resubmitting the proposal and hope to resume these very promising computations early next year, as soon as we granted cpu time. We are optimistic that we will be able to compensate for the lost time because our preliminary benchmarking indicate that optimization with the newly-release Gaussian 09 package on SGI Altix 4700 with as little as 3.7 Gb/cpu memory are 3-5 times faster than the same calculations with Gaussian 03 (using the scrf approach).
