ortho-Hydroxylation of benzoic acids with hydrogen peroxide at a non-heme iron center

Article abstract of DOI:10.1039/b508004e

The iron-assisted hydroxylation of benzoic acid to salicylic acid by 1/H₂O₂ has been achieved in good yield under mild conditions (where 1 is [Fe(II)(BPMEN)(CH₃CN)₂](CIO ₄)₂ and BPMEN = N,N-dimethyl-N,N-bis(2- pyridylmethyl)ethane-1,2-diamine); the product of this reaction is a novel mononuclear iron(m) complex with a chelating salicylate. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b508004e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
ortho-Hydroxylation of benzoic acids with hydrogen peroxide at a non-
heme iron center{
Sonia Taktak,a Margaret Flook,a Bruce M. Foxman,b Lawrence Que, Jr.c and Elena V. Rybak-Akimova*a
Received (in Berkeley, CA, USA) 6th June 2005, Accepted 31st August 2005
First published as an Advance Article on the web 23rd September 2005
DOI: 10.1039/b508004e
which are initially coordinated to iron, can be easily isolated,
and the ligand BPMEN is recovered after the reactions.
When hydrogen peroxide was added to complex 1 in the
presence of benzoic acid, a deep blue color (lmax 5 590 nm)
developed instantly (Fig. 1). The final spectrum was identical to
that of the independently prepared salicylate complex
[FeIII(BPMEN)(OOC(C6H4)O)](ClO4)3 (2) (Fig. S1{) and the
product of reaction was identified in situ by ESI-MS, with only
one strong peak (.5%) detected at m/z 462.2, consistent with the
formulation {[Fe(BPMEN)(OOC(C6H4)O)]}+ (Fig. S2{). Similar
spectral changes were observed when m-ClC6H4COOH was
used (100% peak at m/z 496.2 consistent with
{[Fe(BPMEN)(OOC(C6H3)(Cl)O)]}+).
The iron-assisted hydroxylation of benzoic acid to salicylic acid
by 1/H2O2 has been achieved in good yield under mild
conditions (where 1 is [Fe(II)(BPMEN)(CH3CN)2](ClO4)2 and
BPMEN 5 N,N9-dimethyl-N,N9-bis(2-pyridylmethyl)ethane-
1,2-diamine); the product of this reaction is a novel mono-
nuclear iron(III) complex with a chelating salicylate.
In nature, aromatic hydroxylations are catalyzed by mononuclear
non-heme iron centers in pterin-dependent aromatic amino acid
hydroxylases,1,2 and non-heme diiron centers in toluene mono-
oxygenases.3–5 Synthetically, iron-catalyzed arene hydroxylations
have been reported with Fenton’s reagent (Fe salt/H2O2)6–8 and
related systems that generate either hydroxyl radicals9–11 or metal-
based oxidants.12–16 However, phenol yields were often poor and
the reactions were non-selective, yielding isomers of aromatic
hydroxylation compounds as well as products of side chain
oxidation. Recently, efficient intramolecular aromatic hydroxyl-
ations were reported with mononuclear17–19 or dinuclear20–22 iron
complexes where the aromatic ring is forced into close proximity of
the iron center by covalent binding to the supporting polydentate
ligand. The synthetic utility of these interesting systems, however,
is limited because of their complicated preparations. Non-covalent
substrate binding to the iron center can overcome this short-
coming, as was reported for selective hydroxylation of coordinated
phenols with m-chloroperbenzoic acid (m-CPBA).23 The desire to
use an environmentally clean oxidant, such as H2O2, inspired our
search for simple biomimetic oxidation reactions of coordinated
aromatic substrates.
Complex 2 (Scheme 1) was independently synthesized from
BPMEN, salicylate and Fe(ClO4)3 and fully characterized.{ The
crystal structure of 2 (Fig. 2) shows a six-coordinate Fe(III) center
with four coordination sites occupied by the nitrogen atoms of
the BPMEN ligand and two sites occupied by oxygens from
the salicylate ligand.{ As previously observed with similar
Scheme 1
One of the best examples of a H2O2-activating system is
[FeII(BPMEN)(CH3CN)2](ClO4)2 (1, Scheme 1, BPMEN 5 N,N9-
dimethyl-N,N9-bis(2-pyridylmethyl)ethane-1,2-diamine),24 an excel-
lent catalyst for alkane hydroxylation24,25 and olefin
epoxidation.26–28 We report herein a new reaction promoted
by complex 1, the stoichiometric transformation of benzoic and
m-chlorobenzoic acid into the corresponding salicylic acids upon
oxidation by hydrogen peroxide. The reaction proceeds readily
with added substrates and yields exclusively ortho-hydroxylated
products in high yields under mild conditions. Organic products,
aDepartment of Chemistry, Tufts University, 62 Talbot Avenue,
Medford, Massachusetts, 02155, USA.
E-mail: elena.rybak-akimova@tufts.edu
bDepartment of Chemistry, Brandeis University, Mail Stop 015,
Waltham, Massachusetts, 02454, USA
cDepartment of Chemistry and Center for Metals in Biocatalysis,
University of Minnesota, 207 Pleasant St. SE, Minneapolis, Minnesota,
55455, USA
{ Electronic supplementary information (ESI) available: Experimental
synthetic procedures and physical measurements, UV-vis spectra, ESI-MS,
1H NMR, EPR. See http://dx.doi.org/10.1039/b508004e
Fig. 1 Spectral change upon addition of different amounts of hydrogen
peroxide to 1/benzoic acid solution in acetonitrile (0 to 3 molar equivalents
of H2O2 versus 1, with 0.2 mM of 1 and 0.4 mM of benzoic acid).
Absorbancies were adjusted for dilution. Inset: the cross sections at 373 nm
and 590 nm with 0 to 10 molar equivalents of H2O2 vs. 1.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5301–5303 | 5301

View Article Online
with a shoulder at 370 nm in the near-UV region (Fig. S7{). EPR
experiments were performed on this intermediate, generated in situ
by rapidly mixing 1/PhCOOH and H2O2 at 242 uC followed by
quenching in liquid nitrogen.{ In addition to low-field peaks at
g 5 8.00 (sh) and g 5 4.26 (broad), corresponding to high spin
Fe(III) complexes (such as 2 or 1b), several signals at higher field
with g1 5 2.40, g2 5 2.19 and g3 5 1.90 were observed, indicating
the presence of a low-spin Fe(III) species (Fig. S8{). The g-value of
2.40 strongly implicates a hydroxo complex [FeIII(BPMEN)-
(OH)X]2+, as shown by Talsi and co-workers30 in their analysis of
the reaction of 1 with H2O2. A mononuclear LS intermediate in
our system, tentatively formulated as 1a (Scheme 2), may contain
benzoate anion or a solvent molecule as an additional mono-
dentate ligand. If the labile ligand X is PhCOOH, the initial adduct
would rearrange, via an intramolecular H+ transfer, into
[FeIII(BPMEN)(OOCPh)]2+ (1b), which may also contain a weakly
coordinated solvent molecule.
Fig. 2 X-Ray structure of the complex cation in 2, showing 50%
probability thermal ellipsoids. Hydrogen atoms are omitted for clarity.
complexes,29 BPMEN adopts a cis-a conformation, and the
˚
observed Fe–N bond length range (from 2.13 to 2.20 A) is
In order to determine whether a mononuclear Fe(III) complex
can indeed act as an intermediate in aromatic hydroxylation, in situ
one-electron oxidation of
1
with (NH4)2[CeIV(NO3)6] was
characteristic of high-spin Fe(III) complexes.
performed. In the presence of PhCOOH, a featureless visible
spectrum was obtained (Fig. S9{), and a high-spin Fe(III) complex
was detected by EPR (signals at g 5 8.00 and g 5 4.26) and ESI-
When complex 2 was prepared from 1 and PhCOOH, the
maximal yield (up to 84% at r.t., 91% at 225 uC, as determined by
UV-vis) was obtained with 3 equivalents of H2O2 (Fig. 1). Higher
amounts of H2O2 led to the degradation of the product (Fig. 1,
inset), which may be partially responsible for the above-
stoichiometric amounts of H2O2 required in the synthesis of 2.
In support, the hydroxylation of m-ClC6H4COOH was found to
require only the stoichiometric amount of H2O2 (1.5 equivalents,
Fig. S3{), consistent with its expected lower susceptibility to
subsequent oxidation. Independently prepared complex 2 was
stable in the presence of H2O2 alone, but degraded upon
concomitant addition of 0.2 equivalent of 1. Therefore, an excess
of 1 has to be avoided in the preparation of 2.
MS (100% peak at m/z
5
509.02 consistent with
{[FeIII(BPMEN)(OOCC6H5)](NO3)}+) (Scheme 2, intermediate
1b). Importantly, when 1 was mixed with an equimolar amount
of Ce(IV) and reacted with a mixture of PhCOOH and H2O2, high
yields (up to 83%) of 2 were obtained (Fig. S9{). Attempts to
isolate 1b as a solid led to the growth of crystallographically
characterized green crystals of dinuclear [FeIII2(m-O)-
(m-OOCC6H5)(BPMEN)2](CeIII(NO3)6) (3a),{{ which does not
form salicylate in reaction with H2O2. As a control, complex
[FeIII2(m-O)(m-OOC(C6H5))(BPMEN)2](ClO4)3 (3) was indepen-
dently synthesized from Fe(ClO4)3, benzoate and BPMEN, and
fully characterized.{ This material also did not yield hydroxylation
products upon reaction with H2O2, confirming that dinuclear
BPMEN complexes with oxo- and carboxylato-bridges do not
activate hydrogen peroxide for aromatic hydroxylation.
At the end of the reaction of PhCOOH with 1 and H2O2,
salicylic acid was the only product recovered by extraction with
diethyl ether from acidified aqueous solutions with 86% conver-
sion, as determined by 1H NMR (Fig. S4{). PhCOOH was
recovered as the only species in the control experiments that lacked
either H2O2 or 1. When m-ClC6H4COOH was used, the major
product of the reaction was 5-chlorosalicylic acid with 83%
selectivity (minor product: 3-chlorosalicylic acid) and 80%
conversion (Fig. S5{). The high selectivity for ortho-hydroxylation
suggests the participation of a metal-based oxidant rather than
non-discriminatory hydroxyl radicals.11
Further experiments provided insight into the nature and
reactivity of the hydroxylating agent. In parallel experiments with
1 and m-CPBA (recrystallized) or m-ClC6H4COOH/H2O2 as
Preliminary stopped-flow experiments revealed two kinetically
distinct steps in the formation of 2 from 1/PhCOOH and H2O2
(Fig. S6, S7, Table S1{). The first, rapid step, first-order in both
complex 1 and H2O2, was accompanied by a drop in absorbance
in the near UV region with the disappearance of the band at
l 5 373 nm (e 5 4350 L mol21 cm21) characteristic of complex 1.
The second, slower step was accompanied by an increase in
absorbance in the visible region with the appearance of a band at
l 5 590 nm (e 5 2300 L mol21 cm21) characteristic of the
salicylate complex 2 (Fig. 1 and S6{). This reaction step is also
first-order in H2O2, but its rate appears to depend on the
concentration of 1, indicating a possible higher order in 1. Detailed
kinetic studies will be reported later.
During the first step of the reaction, a pale-yellow intermediate
is formed, characterized by a featureless trace in the visible region
Scheme 2
5302 | Chem. Commun., 2005, 5301–5303
This journal is ß The Royal Society of Chemistry 2005

View Article Online
36665 collected, final R [I . 2s] 5 0.0838, final Rw [I . 2s] 5 0.1504.
X-Ray crystal data for complex 3a, C45H58CeFe2N17O21: M 5 1424.90,
oxidant, the use of m-CPBA led to a slower reaction and a lower
yield of chlorosalicylate complex compared to the reaction between
1/m-ClC6H4COOH and H2O2 (16% yield vs. 74%). Therefore the
rearrangement of the 1–m-CPBA adduct is much less efficient than
the reaction between 1/m-ClC6H4COOH and H2O2, indicating
that the formation of a peroxybenzoate is not essential for the
hydroxylation. The importance of H2O2, consistent with our
preliminary kinetic results (Table S1{), can be explained by the
involvement of an iron–hydroperoxo intermediate. This inter-
mediate can either directly attack the coordinated aromatic ring, or
undergo subsequent O–O bond cleavage to generate a high-valent
iron–oxo oxidant. 18O-labeling experiments showed that the
oxygen incorporated in the salicylate product originates from
H2O2 and that the oxidant responsible for aromatic hydroxylation
does not exchange oxygen with water before the C–O bond is
formed (Fig. S10–12{). This is consistent with a hydroperoxo
intermediate, but does not rule out high-valent iron–oxo species
that may undergo slow isotope exchange with water. Lastly,
experiments with the deuterated substrate, d5-PhCOOH, showed
no significant H/D kinetic isotope effect on either step of the
reaction.{ Competition experiments using ESI-MS also showed no
preference of 1 in reacting with either protio or deuterio
isotopomers (Fig. S13{), so C–H bond breaking on the phenyl
ring is not a rate limiting step in this transformation.
¯
˚
triclinic, space group P1, a 5 11.741(2), b 512.072(2), c 5 21.838(4) A,
3
˚
a 5 83.52(3), b 5 82.33(3), c 5 74.75(3)u, V 5 2949.8(10) A , T 5 294 K,
Z 5 2, m 5 1.332 mm21, Rint 5 0.0218 for 8946 independent reflections of
the 9218 collected, final R [I . 2s] 5 0.0385, final Rw [I . 2s] 5 0.0807.
CCDC 174631 and 174632. See http://dx.doi.org/10.1039/b508004e for
crystallographic data in CIF or other electronic format.
1 T. Flatmark and R. C. Stevens, Chem. Rev., 1999, 99, 2137–2160.
2 P. F. Fitzpatrick, Biochemistry, 2003, 42, 14083–14091.
3 A. Fishman, Y. Tao, L. Y. Rui and T. K. Wood, J. Biol. Chem., 2005,
280, 506–514.
4 M. H. Sazinsky, J. Bard, A. Di Donato and S. J. Lippard, J. Biol.
Chem., 2004, 279, 30600–30610.
5 K. H. Mitchell, C. E. Rogge, T. Gierahn and B. G. Fox, Proc. Natl.
Acad. Sci. USA, 2003, 100, 3784–3789.
6 C. Walling and R. A. Johnson, J. Am. Chem. Soc., 1975, 97, 363–367.
7 T. Kurata, Y. Watanabe, M. Katoh and Y. Sawaki, J. Am. Chem. Soc.,
1988, 110, 7472–7478.
8 D. T. Sawyer, A. Sobkowiak and T. Matsushita, Acc. Chem. Res., 1996,
29, 409–416.
9 S. Udenfriend, C. T. Clark, J. Axelrod and B. B. Brodie, J. Biol. Chem.,
1954, 208, 731–739.
10 J. P. Hage and D. T. Sawyer, J. Am. Chem. Soc., 1995, 117, 5617–5621.
11 (a) T. Tezuka, N. Narita, W. Ando and S. Oae, J. Am. Chem. Soc.,
1981, 103, 3045–3049; (b) G. W. Klein, K. Bhatia, V. Madhavan and
R. H. Schuler, J. Phys. Chem., 1975, 79, 1767–1774.
12 D. Mathieu, J. F. Bartoli, P. Battioni and D. Mansuy, Tetrahedron,
2004, 60, 3855–3862.
13 V. Balland, D. Mathieu, N. Pons-Y-Moll, J. F. Bartoli, F. Banse,
P. Battioni, J. J. Girerd and D. Mansuy, J. Mol. Catal. A: Chem., 2004,
215, 81–87.
14 J. F. Bartoli, F. Lambert, I. Morgenstern-Badarau, P. Battioni and
D. Mansuy, C. R. Chim., 2002, 5, 263–266.
15 D. Bianchi, M. Bertoli, R. Tassinari, M. Ricci and R. Vignola, J. Mol.
Catal. A: Chem., 2003, 204, 419–424.
16 G. A. Hamilton, J. W. Hanifin and J. P. Friedman, J. Am. Chem. Soc.,
1966, 88, 5269–5272.
17 S. J. Lange, H. Miyake and L. Que, Jr., J. Am. Chem. Soc., 1999, 121,
6330–6331.
18 M. P. Jensen, S. J. Lange, M. P. Mehn, E. L. Que and L. Que, Jr., J. Am.
Chem. Soc., 2003, 125, 2113–2128.
19 Y. Mekmouche, S. Me´nage, J. Pe´caut, C. Lebrun, L. Reilly,
V. Schuenemann, A. Trautwein and M. Fontecave, Eur. J. Inorg.
Chem., 2004, 3163–3171.
20 S. Me´nage, J. B. Galey, J. Dumats, G. Hussler, M. Seite´, I. G. Luneau,
G. Chottard and M. Fontecave, J. Am. Chem. Soc., 1998, 120,
13370–13382.
21 H. Furutachi, M. Murayama, A. Shiohara, S. Yamazaki, S. Fujinami,
A. Uehara, M. Suzuki, S. Ogo, Y. Watanabe and Y. Maeda, Chem.
Commun., 2003, 1900–1901.
22 F. Avenier, L. Dubois and J. M. Latour, New J. Chem., 2004, 28,
782–784.
23 N. Kitajima, M. Ito, H. Fukui and Y. Morooka, J. Am. Chem. Soc.,
1993, 115, 9335–9336.
24 K. Chen and L. Que, Jr., Chem. Commun., 1999, 1375–1376.
25 M. Costas, K. Chen and L. Que, Jr., Coord. Chem. Rev., 2000, 200,
517–544.
26 M. Costas, A. K. Tipton, K. Chen, D. H. Jo and L. Que, Jr., J. Am.
Chem. Soc., 2001, 123, 6722–6723.
27 M. C. White, A. G. Doyle and E. N. Jacobsen, J. Am. Chem. Soc., 2001,
123, 7194–7195.
28 J. Y. Ryu, J. Kim, M. Costas, K. Chen, W. Nam and L. Que, Jr., Chem.
Commun., 2002, 1288–1289.
29 S. Taktak, S. V. Kryatov and E. V. Rybak-Akimova, Inorg. Chem.,
2004, 43, 7196–7209.
Based on combined experimental results, a pathway for
aromatic hydroxylation by 1/H2O2 can be proposed (Scheme 2).
PhCOOH binds to 1 and/or the product of its subsequent one-
electron oxidation. When H2O2 is added, both kinetic and EPR
studies suggest the formation of a mononuclear Fe(III) complex
(most likely, a mixture of low-spin species 1a and high-spin species
1b). This intermediate can further react with H2O2 to form a short
lived mononuclear hydroperoxo species, as previously proposed
for BPMEN and related systems.24,31,32 Such species could then
attack the aromatic ring directly or produce high-valent iron–oxo
intermediates. The high selectivity observed argues against the
participation of hydroxyl radicals.11 In our hands, ring closed
species like 3 or 3a did not undergo oxidation with H2O2. These
results, however, do not exclude possible involvement of a
dinuclear ‘‘open-core’’ hydroperoxo intermediate, which will be
explored in future studies.
In conclusion, mononuclear iron(II) complex 1 efficiently and
selectively hydroxylates benzoic acids to their corresponding
salicylic acids under mild conditions. The oxygen source is
hydrogen peroxide, which generates water as the only by-product
of the oxidation reaction.
The authors thank Dr Richard J. Staples (Harvard U.), Prof.
Evgenii P. Talsi (Russian Academy of Science) and Dr Jacob
D. Soper (MIT). This work was supported by the National Science
Foundation (CHE 0111202 to ERA, CHE 9723772 for the NMR
facility at Tufts University, MRI 0320783 for the ESI-MS, CHE
9816557 for the EPR), the Department of Energy (#DE-FG02-03-
ER15455 to LQ), and Tufts Summer Scholars program (to MF).
30 E. A. Duban, K. P. Bryliakov and E. P. Talsi, Mendeleev Commun.,
2005, 12–14.
Notes and references
31 K. Chen, M. Costas, J. H. Kim, A. K. Tipton and L. Que, Jr., J. Am.
Chem. Soc., 2002, 124, 3026–3035.
32 D. Quin˜onero, K. Morokuma, D. G. Musaev, R. Mas-Ballest´e and
L. Que, Jr., J. Am. Chem. Soc., 2005, 127, 6548–6549.
{ X-Ray crystal data for complex 2, C46H58Cl2Fe2N8O17: M 5 1177.60,
monoclinic, space group P21/c, a 5 15.874(5), b 5 11.407(4), c 5
3
˚
˚
29.333(10) A, b 5 101.599(7)u, V 5 5203(3) A , T 5 193(2) K, Z 5 4,
m 5 0.739 mm21, Rint 5 0.0955 for 12857 independent reflections of the
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5301–5303 | 5303