Q. Zhang, C.R. Goldsmith / Inorganica Chimica Acta 406 (2013) 301–306
305
denticities tend to promote superior reactivity, with respect to
both the ultimate yield and speed of the reaction [3]. The better
activity of the complexes with the tetradentate ligands was attrib-
uted to the greater stability of the manganese complexes; the use
of a less highly chelating ligand or the introduction of steric mod-
ifications that lengthened and weakened the Mn–L bonds generally
led to a pronounced loss of catalytic activity [3]. The Mn(II) com-
plex with the debpn ligand promotes slightly less active epoxida-
tion than the Mn(II) complexes with most tetradentate ligands,
catalysts. The ratios observed for 2 are similar to those reported for
other non-heme iron catalysts with bulky N-donor ligands [14,25].
2+
2
Further, another Fe(II) complex of ours, [Fe(bbpc)(MeCN) ] ,
shows much stronger preferences for secondary over tertiary car-
bon oxidation (a ratio of 4.8:1 with trans-1,2-dimethylcyclohex-
ane) without as marked a decrease in the catalysis [14]. The
installation of the weakly binding ethyl esters also eliminates the
selectivity for the alcohol over the ketone product that was ob-
served for both the bpmen and bbpc systems [12–14].
using the epoxidation of
ble 1). The yield of the
-octene epoxide is ꢁ90% of that of the reac-
tion catalyzed by its most closely related six-coordinate analog
Mn(bpmen)(OTf) ]. The reactivity promoted by 1 also proceeds
more slowly, evidenced most clearly by the yields measured at
5 s. The results demonstrate that factors other than the stability
of the manganese-ligand adduct influence the catalysis of alkene
epoxidation by peracetic acid. That -octene and cis-cyclooctene
1
-octene as the basis for comparison (Ta-
Cyclic voltammetry (CV) measurements suggest that the +2 oxi-
dation state is better stabilized by the debpn ligand than by the
bpmen, with the iron being particularly stabilized by the additional
chelate arms. In each of the CV of compounds 1–4, a single irre-
versible or quasi-reversible feature is observed, which we assign
to the divalent metal ion’s oxidation to the +3 state. The CV of 1
has a redox feature 60 mV higher than that observed for 3. The re-
dox event observed for 2 is 175 mV higher than that of 4. The com-
parative increased difficulty in converting the Fe(II) to higher
oxidation states may explain why the iron loses more of its activity
than the Mn(II) upon the bpmen-for-debpn switch. For the alkane
hydroxylation, the oxidation of the iron does not appear to be fully
rate-limiting, however. The measured KIE for cyclohexane is con-
sistent with C–H bond cleavage in the product-determining step.
1
[
2
1
1
are oxidized to essentially the same extent (Table 1) suggests that
steric interactions between the catalyst’s ligand and the organic
substrate may hinder the oxidation of more sterically congested
olefins; normally, cis-cyclooctene is far more reactive than
1-octene
[
1,7,17,23,24]. The similar reactivities of styrene and dimethylsty-
rene also support this conclusion. With the latter two substrates,
the reactions are substantially slower, requiring 30 min for com-
pletion instead of 5 min.
Slower alkene epoxidation activity is also observed when
the Fe(II) complex 2 is used as a catalyst using a protocol devel-
oped by White, Doyle, and Jacobsen in their analysis of
6 6
More turnovers are observed with C H12 than with C D12 suggest-
ing that C–H bond cleavage is still relevant to the rate-determining
step. The ratio of these turnover numbers (1.4:1), however, is much
less than what one would expect based on the KIE of 2.4. An alter-
native explanation for the reduced activity may be that the higher
reduction potentials destabilize the higher-valent iron oxidant
responsible for alkane oxidation. This oxidant may consequently
exhibit faster rates of intramolecular ligand oxidation [6], thereby
eliminating opportunities for the oxidant to react with exogenous
substrates. The observed debpn ligand degradation may be consis-
tent with this alternative explanation.
2
+
[
Fe(bpmen)(MeCN)
within 5 min; whereas, analogous reactions catalyzed by 2 need
0 min to reach their optimal yields. The final yield of the cis-
cyclooctene epoxidation catalyzed by is lower, being
approximately 50% of that reported for the reaction promoted by
2
]
(4) [8]. Epoxidations catalyzed by 4 finish
3
2
2
+
[
2
Fe(bpmen)(MeCN) ] . Compared to its Mn(II) analog 1, 2 is not
as effective at catalyzing the oxidation of terminal olefins, and
the reactivity with styrene, in particular, is negligible. The loss of
epoxidation activity associated with the use of a hexadentate li-
gand in place of a tetradentate one is greater for iron than it is
for manganese. Steric effects are not as evident for the iron-cata-
lyzed epoxidation. The yields of 1-octene oxide from 1-octene are
about half those of cyclooctene oxide from cis-cyclooctene.
The Fe(II) complex 2 was also tested as a catalyst for the oxida-
Although the additional chelate arms of debpn were previously
found to stabilize and solubilize ligand–metal adducts in water
[19], neither the Mn(II) nor the Fe(II) compound is a competent
catalyst for hydrocarbon oxidation in aqueous solutions. Reactions
were run in anaerobic distilled water; otherwise the reaction con-
ditions were identical to those of the MeCN reactions. Due to the
immiscibility of the substrate with water, the reactions were stir-
red quickly to ensure that the reaction proceeded [27], The yields
of cyclohexene oxide from cyclohexene for the uncatalyzed reac-
2 2
tion of alkanes by H O (Table 4). Iron complexes with bpmen have
been explored extensively as catalysts for these reactions [10–13].
The addition of the ethyl esters to the ligand framework reduces
the activity to a greater extent than the previously described epox-
idation reactions. Using the oxidation of cyclohexane by 10 equiv.
tions with peracetic acid and H
compound present. The [Fe(debpn)(H
lyze the oxidation of cyclohexane by H
2 2
O are equal to those with a debpn
2
+
2
O)] complex fails to cata-
in water. Although the
2 2
O
debpn ligand should remain more tightly bound to the metals
due to their potential hexadenticity, the Mn(II) and Fe(II) com-
plexes are still susceptible to degradation through side reactions
with the terminal oxidants used for hydrocarbon oxidation.
Although the use of a neutral ligand that promotes heptacoordi-
nation appears to be a poor design feature for a first-row transition
metal catalyst for hydrocarbon oxidation, such ligands may be ben-
eficial for other applications. One concern in using transition metal
ions for biological imaging, for instance, is that they often catalyze
unwanted redox activity. Preparing a biological imaging agent with
a more highly coordinating ligand may significantly limit these
reactions by better stabilizing the metals in lower oxidation states.
2 2
of H O as a comparative standard, about 70% of the catalytic
activity is lost upon switching the catalyst from [Fe(bpmen)(OTf)
2
]
to 2. The debpn ligand does appear to be relatively bulky, as
assessed by the retention of configuration (RC) for the oxidation
of cis-1,2-dimethylcyclohexane. RC has been previously defined
as [([(1R,2R + 1S,2S)–(1R,2S + 1S,2R)]/(total amount of tertiary
alcohol), where (1R,2R), (1S,2S), (1R,2S), and (1S,2R) are the
various isomers of 1,2-dimethylcyclohexanol [10]. The RC for cis-
1
,2-dimethylcyclohexane oxidation has been found to decrease
upon switching to an iron catalyst with a bulkier ligand. The
3% RC for 2 is significantly lower than the 96% value for
Fe(bpmen)(OTf) ] but is higher than those associated with more
8
[
2
catalytically active iron complexes with doubly and triply methyl-
ated tris(picolylamine) ligands [10].
Acknowledgement
The reduced C–H activation catalysis therefore cannot be solely
attributed to increased steric interactions between 2 and potential
substrates. The ratios of secondary to tertiary carbon oxidation for
the two 1,2-dimethylcyclohexane substrates provide an alternative
means of assessing the steric hindrance between the substrate and
The work was financially supported by Auburn University and
the American Chemical Society, donors of the Petroleum Research
Fund (Grant 49532-DNI3). Q. Z. was supported by an NSF EPSCoR/
AU-CMB summer fellowship.