Three novel non-heme micro-oxo-bridged diiron(III) complexes [Fe2(micro-O)(L1)2] 2, where H2(L1) is N,N'-o-phenylenebis(salicylideneimine), [Fe2(micro-O)(L2)2].2H2O 4, where H2(L2) is N,N'-o-phenylenebis(3,5-di-tert-butylsalicylideneimine), and [Fe2(micro-O)(L3)2] 6, where H2(L3)=1,4-bis(2-hydroxybenzyl)-1,4-diazepane, have been isolated and studied as catalysts for the selective oxidative transformation of alkanes into alcohols using m-choloroperbenzoic acid (m-CPBA) as co-oxidant. The mononuclear iron(III) complexes [Fe(L1)Cl] 1 and [Fe(L4)Cl] 7, where H2(L4)=1,4-bis(2-hydroxy-3,5-di-tert-butylbenzyl)-1,4-diazepane, have been also isolated and those corresponding to the dimeric complexes 4 and 6 have been generated in CH3CN solution and characterized as [Fe(L2)Cl] 3 and [Fe(L3)Cl] 5 by using ESI-MS, absorption and EPR spectral and electrochemical methods. The molecular structures of 4 and 6 have been successfully determined by single crystal X-ray diffraction. Both 4 and 6 possess the Fe-O-Fe structural motif with each iron atom possessing a distorted square pyramidal coordination geometry. The steric constraint at the iron(III) center in 6 is higher than that in 4 as understood from the values of the trigonality structural index (tau: 4, 0.226, 0.273; 6, 0.449) due to the higher steric congestion built by the diazapane back bone. The micro-oxo-to-Fe(III) LMCT band for 4 is observed around 622 nm (epsilon, 1830 M(-1) cm(-1)) in methanol but is not observed in CH3CN solution and it is blue-shifted to around 485 nm (epsilon, 5760 M(-1) cm(-1)) in 6, possibly due to the higher Fe-O-Fe bond angle in the latter (4, 177.4; 6, 180 degrees). The Fe(III)/Fe(II) redox potentials of the dinuclear complexes (E1/2: 2, -0.606; 4, -0.329; 6, -0.889 V) are more negative than those for their corresponding mononuclear complexes (E1/2: 1, -0.300 V; 3, -0.269; 5, -0.289 V) due to O2- coordination. Interestingly, upon addition of peroxides (H2O2, t-BuOOH) and the peracid m-CPBA, the intensity of the phenolate-to-Fe(III) LMCT band for 2 and 6 decreases but does not exhibit any appreciable change for 4. In the presence of m-CPBA cyclohexane is selectively (A/K, 12.2) oxidized by the dimeric complex to cyclohexanol (A, CyOH) and a small amount of the further oxidized product cyclohexanone (K, CyO). However, interestingly, the corresponding monomeric complex affords enhanced yields of both CyOH and CyO but with a lower selectivity (A/K=1.7) and also 1-chlorocyclohexane via oxidative ligand transfer (OLT). The oxidation of adamantane by 4 affords exclusively 1-adamantanol (50.5%) and 2-adamantanol (9.5%) with enhanced yields over 12 h. In contrast, 3 provides 1-adamantanol (32.4%) and 2-adamantanol (14.8%) and adamantanone (14.6%) in addition to 1-chloroadamantane (14.1%) as the OLT product. The secondary C-H bond of ethylbenzene is randomly activated by both 3 and 4 to give 1-phenylethanol and acetophenone. Also, oxidation of cumene with tertiary C-H bonds to give 2-phenyl-2-propanol and the further oxidized product acetophenone is illustrated by invoking the iron-phenoxyl radical species as invoked for metalloporphyrin-catalyzed systems. The strong chemoselectivity in C-H bond activation of alkanes by 4has been illustrated by invoking the involvement of a high-valent iron-oxo intermediate generated by using m-CPBA rather than the conventional oxidants H2O2 and t-BuOOH. In contrast to 4, the complexes 2 and 6 fail to effect the oxidation of hydrocarbons in the presence of H2O2, t-BuOOH and m-CPBA as the co-oxidant.
Read full abstract