AmericanChemicalSociety.com

Our PRF-sponsored research carries an overarching aim of discovering enantioselective oxidation catalysts among a group of chiral metallophthalocyanine derivatives bearing substitution at the 14 and 28 positions of what would otherwise be an unmodified phthalocyaninato ligand.  Realization of this central goal will require the successful separation and isolation of the potential catalysts and also discovery of optimal catalytic conditions employing these complexes.  In the early and middle stages of the grant period, most of our efforts were directed toward the separation of enantiomers of potential catalysts belonging to two sub-classes.  In one sub-class, a series of alkoxy-modified ligands of general formula 14,28-(RO)₂Pc (where R represents Me, Et, n-Pr, n-Bu, or beta-hydroxyethoxy) are coordinated to nickel II.  The other sub-class consists of the so-called “helmet” metallophthalocyanines of general formula L(diiPc)M, where diiPc is the 14,28-[1,3-diiminoisoindolinato]phthalocyaninato ligand, L is an alcohol or water ligand, and M is either Fe(III) or Co(III). 

Over the course of the past grant year, we have chosen to concentrate our efforts on the catalytic aspects of the project, with some exciting and promising results.  More specifically, we have focused primarily on the catalytic oxidation of cyclohexane and cyclooctane employing the iron(III) “helmet” complex.  Our decision to focus in this area was driven not only by the overall aim of the project, but also by an examination of literature and observation that iron complexes showing unique structural similarities to the iron “helmet” metallophthalocyanine moiety serve as some of the most effective non-heme alkane oxidation catalysts known.  It should be noted that our interest in this area derives more from the catalytic potential of our complexes than from the biomimetic aspects of non-heme iron catalysts that have motivated other investigatons.

When we first began to study catalytic oxidation employing L(diiPc)Fe, we carried out reactions of excess cyclohexane with hydrogen peroxide in dichloromethane under an argon atmosphere and observed the formation of cyclohexanol as the major product, with cyclohexanone also being formed in minor amounts.  Although this was encouraging, the overall process was rather sluggish in comparison to previously reported studies, with a turnover number (TON) of only about eight for the alcohol being observed in almost an hour’s time.  In subsequent experiments, however, we observed marked improvement in catalyst performance when the iron complex was stirred overnight in a 50% dichloromethane/ 50% acetonitrile mixture prior to addition of cyclohexane, followed by addition of the oxidant via syringe pump.  Specifically, turnover numbers of 100.9 for cyclohexanol and 15.1 for cyclohexanone were found.  Only trace amounts of byproducts were observed.  This result is significant in that it represents, to our knowledge, the best performance for any non-heme iron-based alkane oxidation catalyst ever seen in terms of TON and selectivity for the alcohol with an alcohol to ketone (A/K) ratio of nearly seven to one.  It appears that the improvement in catalyst performance in acetonitrile/dichloromethane mixed solvent over that observed in 100% dichloromethane results from replacement of the ligand L in L(diiPc)Fe, which is a combination of methanol and water in a ratio of 80% methanol to 20% water in the complex as synthesized by (apparently) more labile acetonitrile.  This is supported by the observation that TON decreases if the catalyst is stirred in the mixed solvent system for shorter periods of time prior to substrate and oxidant addition.  We have seen slow replacement of L on L(diiPc)Fe by basic solvent ligands  in other studies by HPLC.  When the oxidation is carried out in an O₂ atmosphere instead of Ar, the TON for alcohol decreases to 30.2 and the A/K ratio decreases to 2.7 to one.

Following the results for cyclohexane oxidation, we extended our investigation to cyclooctane and found similar performance by the same catalyst.  In experiments exactly analogous to that described above, we observed a TON of 122.2 for cyclooctanol and 5.8 for cyclooctanone, corresponding to A/K = 21.0.  This is by far the best performance we know of for any non-heme iron-based catalyst in cyclooctane oxidation.  As was the case for cyclohexane oxidation, we observed poorer catalyst performance for cyclooctane oxidation under an O₂ atmosphere, with the cyclooctanol TON decreasing to 40.4 and the A/K ratio decreasing to 5.3 to one.

 We also studied the performance of L(diiPc)Fe as a catalyst for oxidation of indan to 1-indanol under analogous conditions, but we have observed a TON of only 10.1 and uncharacterized byproducts in much higher amount and number.  However, this result may still prove to be significant if we can increase the TON through optimization of conditions and successfully isolate the catalyst in enantiopure or enriched form, because 1-indanol is a chiral product and thus represents a target for asymmetric catalysis.  It is interesting to note that no observable amount of  1-indanone was evident in the product mixture from indan oxidation by GC-MS analysis.