P. Comba and S. Wunderlich
solid. ESI-MS (MeCN): m/z: 529.09438 [Fe(L)(Cl)]+; elemental analysis
(%) found: C 49.45, H 4.83, N 11.48; calcd for C25H29Cl2FeN5O5 (Mr
606.28): C 49.53, H 4.82, N 11.55. [Fe(L)Br2]·CH3CN: 82% yield, orange
solid. ESI-MS (MeCN): m/z: 575.04158 [Fe(L)(Br)]+; elemental analysis
(%) found: C 43.12, H 4.22, N 10.04; calcd for C25H29Br2FeN5O5 (Mr
functional halogenase model, the first reported so far. The
yield and selectivity strongly depend on the oxidant and re-
action conditions. Generally, the selectivity for halogenation
over hydroxylation is high, and OR radicals (R=tBu in
TBHP, H in H2O2) probably are responsible for the reduced
selectivity in the TBHP-based reactions and the catalytic
(but not the stoichiometric) reaction with H2O2. The stoi-
chiometric reactions afford up to quantitative yield; there-
fore, pathways for catalyst deactivation must exist. The fact
that the systems presented here are so far rather poor cata-
lyst systems probably is largely due to the fact that ligand
exchange to reproduce the halogeno iron(II) precatalysts is
too inefficient. Thus, there are possibilities to optimize and
considerably improve these systems. Interestingly, it appears
that the least efficient process, that is, that with PhIO as oxi-
dant, is the simplest reaction, probably involving only an
XFeIV=O-based pathway, and it selectively produces halo-
genated products. The yield with H2O2 is considerably
higher, and the experimentally determined KIEs and com-
putational data suggest that pathways based on XFeIV=O
and XFeV=O may be involved. The high selectivity in the
stoichiometric reaction was expected on the basis of the pro-
posed direct oxidation of [FeII(L)X2]n+ with H2O2 to FeIV=
O, which does not involve FeIII intermediates and OH radi-
cals.[18,25] The yield with TBHP as oxidant is nearly quantita-
tive in stoichiometric reactions but this process is OR-radi-
cal-based and unselective.
695.18):
C
43.19,
H
4.20,
N
10.07. [Fe(L)F2]·2.5H2O: (Bu4N)F
(NCCH3)](OTf)2 in acetonitrile
(0.54 g2.07 mmol) was added to [Fe(L)
N
ACHTUNGTRENNUNG
(0.60 g, 0.72 mmol). A red solid was obtained by filtration after 10 min at
ambient temperature, washed with cold MeCN, and dried (yield of 93%).
ESI-MS (MeCN): m/z: 513.1235 [Fe(L)(F)]+; elemental analysis (%)
found: C 47.96, H 5.46, N 9.66; calcd for C46H62F4Fe2N8O15 (Mr 1154.71):
C 47.85, H 5.41, N 9.77.[21]
Stoichiometric oxidation of cyclohexane: In a typical reaction, 0.7m cy-
clohexane was treated with 7 mm complex and 7 mm oxidant in MeCN
(258C, Ar, H2O-free). For the reactions with H2O2 and TBHP the oxi-
dant was diluted to 0.3 mL and added by syringe pump over 30 min at
258C to a solution of the substrate and catalyst. The solution was stirred
for a further 5 min after addition of the oxidant was completed. For the
reaction with PhIO, the oxidant was added as a solid, and the reaction
time was increased to 24 h (see Table 1). The product solution was ab-
sorbed onto a silica gel column and washed with MeCN (5 mL). Naph-
thalene was added as internal standard and the mixture was analyzed by
GC. The retention times for the product peaks were compared with
those of standard compounds and their identity was confirmed by GC-
MS. All reactions were done in triplicate; the reported data is the aver-
age of these reactions.
Catalytic oxidation of cyclohexane: These reactions were conducted with
concentrations of 0.7 mm catalyst, 70 mm oxidant, and 0.7m substrate (cy-
clohexane and Bu4NCl or Bu4NBr) under the same conditions as for the
stoichiometric experiments. The reaction was quenched by the addition
of an equal volume of water, and the products were isolated by extrac-
tion with 3ꢂ2 mL Et2O. The ether layers were combined, dried over an-
hydrous Na2SO4, and analyzed by GC.
A variety of proposals were made for the chemoselectivity
of the enzyme-catalyzed reaction. From our results present-
ed here it appears that both XFeIV=O- and XFeV=O-based
processes should generally have high halogenation selectivi-
ty, and it emerges that decreasing selectivities are generally
indicative of radical-based side reactions. The total selectivi-
ty required by biological reactions seems to be accomplished
by enzyme-based substrate positioning,[9] an approach which
is not available for simple low molecular weight compounds.
The design of more efficient catalyst systems needs to pre-
vent oxygen-based radicals and requires efficient ligand-ex-
change processes to produce high enough concentrations of
the iron(II) halogenide precatalyst, and these are require-
ments which might be possible to meet.
Oxidation of adamantane: Identical conditions to those for stoichiometric
cyclohexane oxidation were used but with final concentrations of 0.7 mm
catalyst, 7 mm oxidant, and 7 mm substrate.
Determination of the KIEs: Identical conditions to those for stoichiomet-
ric cyclohexane oxidation were used but with a 1:3 mixture of cyclohex-
ane (175 mm) and [D12]cyclohexane (525 mm) followed by analysis of the
product distribution of the halogenated alkane.
Computational studies: All DFT calculations were performed with the
Jaguar 6.5 program package, unless otherwise specified[48] The B3LYP
functional[49–51] and LACVP basis set (double z with a Los Alamos effec-
tive core potential for the Fe center and 6-31G for the other atoms) were
used.[52, 53] All intermediates were confirmed by frequency calculations
with Gaussian 03.[54] Single-point calculations were performed on the
B3LYP/LACVP optimized geometries by using the LACV3P++** basis
set (LanL2DZ on the Fe center and 6-311++G** on the other atoms).
The energies reported are those calculated at the B3LYP/LACV3P++**
level and include zero point and free energy corrections, derived from
the B3LYP/LACVP calculations. A simplified model system was used in
all calculations, in which the ester groups on the ligand backbone were
replaced by hydrogen atoms.
Experimental Section
General: Chemicals (Aldrich, Fluka) and solvents were of highest possi-
ble grade and used as purchased. Mass spectra: Bruker ApexQe hybrid
9.4 FT-ICR or Finnigan TSW 700. Elemental analyses were performed
by the analytical laboratories of the chemical institutes of the University
of Heidelberg. Products were analyzed by GC on a Varian 3900 instru-
ment with a ZB-1701 column.
Acknowledgements
Financial support by the German Science Foundation (DFG Research
Group FOR 763 “Natural Halogenation Processes”) is gratefully ac-
knowledged.
Synthesis of the ligand and Fe bispidine complexes. Bispidine ligand L
was synthesized as reported.[47] For the synthesis of the FeII chloro and
bromo complexes, a suspension of 8.77 g (20.0 mmol) ligand L in acetoni-
trile (60 mL) was mixed with an equimolar amount of FeCl2 or FeBr2
under argon and strictly H2O free conditions. The product started to pre-
cipitate after few minutes and, after stirring for 30 min at ambient tem-
perature, was isolated in good yields by filtration, washed with cold
MeCN (10 mL), and dried.[21] [Fe(L)(Cl)2]·CH3CN: 85% yield, orange
[1] F. H. Vaillancourt, E. Yeh, D. A. Vosburg, S. Garneau-Tsodikova,
[2] F. H. Vaillancourt, E. Yeh, D. A. Vosburg, S. E. O’Conner, C. T.
7298
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7293 – 7299