Biomimetic aryl hydroxylation derived from alkyl hydroperoxide at a nonheme iron center. Evidence for an FeIV=O oxidant

Published on Web 02/01/2003

Biomimetic Aryl Hydroxylation Derived from Alkyl

Hydroperoxide at a Nonheme Iron Center. Evidence for an

Fe^IVdO Oxidant

Michael P. Jensen, Steven J. Lange, Mark P. Mehn, Emily L. Que, and

Lawrence Que, Jr.*

Contribution from the Department of Chemistry and Center for Metals in Biocatalysis,

UniVersity of Minnesota, Minneapolis, Minnesota 55455

Received September 9, 2002 ; E-mail: que@chem.umn.edu

Abstract: Many nonheme iron-dependent enzymes activate dioxygen to catalyze hydroxylations of arene

substrates. Key features of this chemistry have been developed from complexes of a family of tetradentate

tripodal ligands obtained by modification of tris(2-pyridylmethyl)amine (TPA) with single R-arene substituents.

These included the following: -C₆H₅(i.e., 6-PhTPA), L₁; -o-C₆H₄D, o-d₁-L₁; -C₆D₅, d₅-L₁; -m-C₆H₄NO₂,

L₂; -m-C₆H₄CF₃, L₃; -m-C₆H₄Cl, L₄; -m-C₆H₄CH₃, L₅; -m-C₆H₄OCH₃, L₆; -p-C₆H₄OCH₃, L₇. Additionally,

the corresponding ligand with one R-phenyl and two R-methyl substituents (6,6-Me₂-6-PhTPA, L₈) was

also synthesized. Complexes of the formulas [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂, [(L_n)Fe^II(OTf)₂] (n ) 1-7, OTf )

^-O₃SCF₃), and [(L₈)Fe^II(OTf)₂]₂were obtained and characterized by ¹H NMR and UV-visible spectroscopies

and by X-ray diffraction in the cases of [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂, [(L₆)Fe^II(OTf)₂], and [(L₈)Fe^II(OTf)₂]₂. The

complexes react with tert-butyl hydroperoxide (^tBuOOH) in CH₃CN solutions to give iron(III) complexes of

ortho-hydroxylated ligands. The product complex derived from L₁was identified as the solvated monomeric

complex [(L₁O^-)Fe^III]²⁺in equilibrium with its oxo-bridged dimer [(L₁O^-)₂Fe^III₂(µ₂-O)]²⁺, which was character-

ized by X-ray crystallography as the BPh₄salt. The L₈product was also an oxo-bridged dimer, [(L₈O^-)₂-

-

Fe^III₂(µ₂-O)]²⁺. Transient intermediates were observed at low temperature by UV-visible spectroscopy,

and these were characterized as iron(III) alkylperoxo complexes by resonance Raman and EPR

spectroscopies for L₁and L₈. [(L₁)Fe^II(OTf)₂] gave rise to a mixture of high-spin (S ) 5/2) and low-spin (S

) 1/2) Fe^III-OOR isomers in acetonitrile, whereas both [(L₁)Fe(OTf)₂] in CH₂Cl₂and [(L₈)Fe(OTf)₂]₂in

acetonitrile afforded only high-spin intermediates. The L₁and L₈intermediates both decomposed to form

respective phenolate complexes, but their reaction times differed by 3 orders of magnitude. In the case of

L₁, ¹⁸O isotope labeling indicated that the phenolate oxygen is derived from the terminal peroxide oxygen

via a species that can undergo partial exchange with exogenous water. The iron(III) alkylperoxo intermediate

is proposed to undergo homolytic O-O bond cleavage to yield an oxoiron(IV) species as an unobserved

reactive intermediate in the hydroxylation of the pendant R-aryl substituents. The putative homolytic chemistry

was confirmed by using 2-methyl-1-phenyl-2-propyl hydroperoxide (MPPH) as a probe, and the products

obtained in the presence and in the absence of air were consistent with formation of alkoxy radical (RO^•).

Moreover, when one ortho position was labeled with deuterium, no selectivity was observed between

hydroxylation of the deuterated and normal isotopomeric ortho sites, but a significant 1,2-deuterium shift

(“NIH shift”) occurred. These results provide strong mechanistic evidence for a metal-centered electrophilic

oxidant, presumably an oxoiron(IV) complex, in these arene hydroxylations and support participation of

such a species in the mechanisms of the nonheme iron- and pterin-dependent aryl amino acid hydroxylases.

Introduction

as key turnover intermediates and afford two-electron couples,

Fe^III/(Fe^VdO or L^•,+Fe^IVdO) for heme, Fe^II/Fe^IVdO for mono-

Metal-mediated oxygen atom transfer to organic substrates

is a common mechanistic feature in the enzymology of a broad

range of heme and nonheme iron-containing oxygenase en-

zymes. The former category includes the cytochromes P450,

nitric oxide synthases, prostaglandin H synthases, and chloro-

peroxidases.¹Among the latter are monometallic R-keto acid-

and pterin-dependent hydroxylases²and diiron methane mo-

nooxygenases.³High-valent iron-oxo complexes are invoked

metallic nonheme, and Fe^III₂/Fe^IV₂O₂for bimetallic nonheme

enzymes, to accomplish oxo atom transfer. While direct physical

evidence for high-valent species has been obtained for heme

enzymes⁴and methane monooxygenase,⁵such species have yet

to be observed in the catalytic cycles of any mononuclear

nonheme iron enzymes.

The aryl amino acid hydroxylases are a family of nonheme,

iron(II)-dependent enzymes consisting of three closely related

oxidases essential to mammalian physiology, namely phenyl-

alanine (PheH), tyrosine (TyrH), and tryptophan (TrpH) hy-

(1) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. ReV. 1996,

96, 2841.

9

J. AM. CHEM. SOC. 2003, 125, 2113-2128

2113

A R T I C L E S

Jensen et al.

Scheme 1

droxylases.⁶The catalytic domain sequences of these enzymes

are more than 80% identical, and iron(II), reduced tetrahydro-

biopterin (BH₄), and dioxygen are required to selectively

hydroxylate the aromatic rings. Crystallographic studies show

that the iron(II) center is ligated by a recurring 2-His-1-

carboxylate facial triad motif⁷arising from amino acid side

chains and water molecules in close proximity to BH₄.⁸

Biophysical studies and DFT calculations suggest a reaction

mechanism in which iron(II) and BH₄react with dioxygen to

form an iron(II)-peroxo complex, Scheme 1.⁹Heterolysis of

the bound O-O bond forms BH₃OH and oxoiron(IV), which

reacts as an electrophile toward the aromatic substrate. Release

of BH₂, H₂O, and the hydroxylated product completes the

turnover. Attention also is called to analogous examples of post-

translational oxidative modification of aromatic residues in the

vicinity of nonheme iron active sites.¹⁰

As noted previously, direct evidence for oxoiron(IV) as the

hydroxylating intermediate is lacking. However, several obser-

vations support this assignment: (i) TyrH ¹⁶O/¹⁸O kinetic isotope

effects are consistent with formation of a pterin hydroperoxide

from O₂.¹¹(ii) Observations of NIH shifts on aryl substrates

are indicative of cationic intermediates and are consistent with

O-atom transfer.^9,12(iii) Reaction of deuterated substrate at a

truncated TrpH catalytic domain yields an inverse kinetic isotope

effect, consistent with partially rate-limiting electrophilic attack.⁹

(iv) TyrH and PheH convert 4-CH₃-phenylalanine into a mixture

of 3- and 4-methyltyrosines and 4-HOCH₂-phenylalanine.¹³(v)

Consumption of reduced pterin and O₂by TyrH can be

uncoupled from product formation, indicative of competitive

formation of the hydroxylating intermediate.¹⁴(vi) This com-

petition exhibits a strong substituent effect when 4-substituted

phenylalanines are utilized as substrates, which is consistent

with an electron-deficient transition state.¹³(vii) The absence

of such effects on overall rates of pterin oxidation is consistent

with rate-limiting formation of a hydroxylating intermediate

prior to actual hydroxylation of the substrates.¹⁵(viii) A disputed

peroxide shunt mode has been reported for TyrH and H₂O₂.¹⁶

A few reports of model chemistry are relevant to this oxidase

enzymology. These include iron-catalyzed oxygenations of

coordinated phenolates and arene ligand substituents by O₂/

reductant couples.^17,18Related peroxide-driven reactions also

are known.^18-20However, examinations of the scope and

mechanisms of such reactivity remain incomplete.

Our laboratory has utilized iron complexes of tris(2-pyridyl-

methylamine) (TPA) and related ligands as models of nonheme

iron oxygenases and as catalysts for selective oxidative trans-

formations of organic substrates.²¹Such complexes were

demonstrated to catalyze hydroxylation of alkanes, epoxidation

and cis-dihydroxylation of olefins, and sulfoxidation of organic

sulfides by peroxides. Evidence has accumulated suggesting

these chemistries are mediated by high-valent iron-oxo spe-

cies,^21,22including the ability to induce asymmetry in certain

products using chiral ligand derivatives.²³However, the mech-

anisms of these reactions vary depending on the added

(13) (a) Siegmund, H.-U.; Kaufman, S. J. Biol. Chem. 1991, 266, 2903. (b)

Hillas, P. J.; Fitzpatrick, P. F. Biochemistry 1996, 35, 6969. (c) Frantom,

P. A.; Pongdee, R.; Sulikowski, G. A.; Fitzpatrick, P. F. J. Am. Chem.

Soc. 2002, 124, 4202.

(14) Davis, M. D.; Kaufman, S. J. Biol. Chem. 1989, 264, 8585.

(15) Fitzpatrick, P. F. Biochemistry 1991, 30, 6386.

(2) (a) Que, L., Jr.; Ho, R. Y. N. Chem. ReV. 1996, 96, 2607. (b) Solomon, E.

I.; Brunold, T. C.; Davis, M. I.; Kemsley, J. N.; Lee, S.-K.; Lehnert, N.;

Neese, F.; Skulan, A. J.; Yang, Y.-S.; Zhou, J. Chem. ReV. 2000, 100,

235.

(3) (a) Waller, B. J.; Lipscomb, J. D. Chem. ReV. 1996, 96, 2625. (b) Merkx,

M.; Kopp, D. A.; Sazinsky, M. H.; Blazyk, J. L.; Mu¨ller, J.; Lippard S. J.

Angew. Chem., Int. Ed. 2001, 40, 2782.

(4) (a) Schlichting, I.; Berendzen, J.; Chu, K.; Stock, A. M.; Maves, S. A.;

Benson, D. E.; Sweet, R. M.; Ringe, D.; Petsko, G. A.; Sligar, S. G. Science

2000, 287, 1615. (b) Kellner, D. G.; Hung, S. C.; Weiss, K. E.; Sligar, S.

G. J. Biol. Chem. 2002, 277, 9641.

(16) (a) Dix, T. A.; Kuhn, D. M.; Benkovic, S. J. Biochemistry 1987, 26, 3354.

(b) Ribeiro, P.; Pigeon, D.; Kaufman, S. J. Biol. Chem. 1991, 266, 16207.

(17) (a) Hage, J. P.; Sawyer, D. T. J. Am. Chem. Soc. 1995, 117, 5617. (b)

Funabiki, T.; Yokomizo, T.; Suzuki, S.; Yoshida, S. Chem. Commun. 1997,

151. (c) Hegg, E. L.; Ho, R. Y. N.; Que, L., Jr. J. Am. Chem. Soc. 1999,

121, 1972.

(18) (a) Me´nage, S.; Galey, J.-B.; Hussler, G.; Seite´, M.; Fontecave, M. Angew.

Chem., Int. Ed. 1996, 35, 2353. (b) Me´nage, S.; Galey, J.-B.; Dumats, J.;

Hussler, G.; Seite´, M.; Luneau, I. G.; Chottard, G.; Fontecave, M. J. Am.

Chem. Soc. 1998, 120, 13370.

(19) Mekmouche, Y.; Me´nage, S.; Toia-Duboc, C.; Fontecave, M.; Galey, J.-

B.; Lebrun, C.; Pe´caut, J. Angew. Chem., Int. Ed. 2001, 40, 949.

(20) Kitajima, N.; Ito, M.; Fukui, H.; Moro-oka, Y. J. Am. Chem. Soc. 1993,

115, 9335.

(21) (a) Leising, R. A.; Norman, R. E.; Que, L., Jr. Inorg. Chem. 1990, 29,

2553. (b) Leising, R. A.; Brennan, B. A.; Que, L., Jr.; Fox, B. G.; Mu¨nck,

E. J. Am. Chem. Soc. 1991, 113, 3988. (c) Kojima, T.; Leising, R. A.;

Yan, S.; Que, L., Jr. J. Am. Chem. Soc. 1993, 115, 11328. (d) Kim, J.;

Larka, E.; Wilkinson, E. C.; Que, L., Jr. Angew. Chem., Int. Ed. 1995, 34,

2048. (e) Kim, J.; Dong, Y.; Larka, E.; Que, L., Jr. Inorg. Chem. 1996, 35,

2369. (f) Kim, J.; Harrison, R. G.; Kim, C.; Que, L., Jr. J. Am. Chem. Soc.

1996, 118, 4373. (g) Kim, C.; Chen, K.; Kim, J.; Que, L., Jr. J. Am. Chem.

Soc. 1997, 119, 5964. (h) Chen, K.; Que, L., Jr. Angew. Chem., Int. Ed.

1999, 38, 2227. (i) Chen, K.; Que, L., Jr. Chem. Commun. 1999, 1375. (j)

Costas, M.; Chen, K.; Que, L., Jr. Coord. Chem. ReV. 2000, 200-2, 517.

(k) Miyake, H.; Chen, K.; Lange, S. J.; Que, L., Jr. Inorg. Chem. 2001,

40, 3534. (l) Chen, K.; Que, L., Jr. J. Am. Chem. Soc. 2001, 123, 6327.

(m) Chen, K.; Costas, M.; Kim, J.; Tipton, A. K.; Que, L., Jr. J. Am. Chem.

Soc. 2002, 124, 3026. (n) Ryu, J. Y.; Kim, J.; Costas, M.; Chen, K.; Nam,

W.; Que, L., Jr. Chem. Commun. 2002, 1288.

(5) (a) Lee, S.-K.; Nesheim, J. C.; Lipscomb, J. D. J. Biol. Chem. 1993, 268,

21569. (b) Shu, L.; Nesheim, J. C.; Kauffmann, K.; Mu¨nck, E.; Lipscomb,

J. D.; Que, L., Jr. Science 1997, 275, 515. (c) Valentine, A. M.; Stahl, S.

S.; Lippard, S. J. J. Am. Chem. Soc. 1999, 121, 3876.

(6) (a) Kappock, T. J.; Carradona, J. P. Chem. ReV. 1996, 96, 2659. (b)

Fitzpatrick, P. F. Annu. ReV. Biochem. 1999, 68, 355.

(7) Hegg, E. L.; Que, L., Jr. Eur. J. Biochem. 1997, 250, 625.

(8) (a) Flatmark, T.; Stevens, R. C. Chem. ReV. 1999, 99, 2137. (b) Que, L.,

Jr. Nat. Struct. Biol. 2000, 7, 182.

(9) (a) Moran, G. R.; Derecskei-Kovacs, A.; Hillas, P. J.; Fitzpatrick, P. F. J.

Am. Chem. Soc. 2000, 122, 4535. (b) Bassan, A.; Blomberg, M. R. A.;

Siegbahn, P. E. M. Eur. Chem. J. 2003, 9, 106.

(10) (a) Ormo¨, M.; deMare´, F.; Regnstro¨m, K.; A° berg, A.; Sahlin, M.; Ling, J.;

Loehr, T. M.; Sanders-Loehr, J.; Sjo¨berg, B.-M. J. Biol. Chem. 1992, 267,

8711. (b) Baldwin, J.; Voegtli, W. C.; Khidekel, N.; Moe¨nne-Loccoz, P.;

Krebs, C.; Pereira, A. S.; Ley, B. A.; Huynh, B. H.; Loehr, T. M.; Riggs-

Gelasco, P. J.; Rosenzweig, A. C.; Bollinger, J. M., Jr. J. Am. Chem. Soc.

2001, 123, 7017. (c) Liu, A.; Ho, R. Y. N.; Que, L., Jr.; Ryle, M. J.; Phinney,

B. S.; Hausinger, R. P. J. Am. Chem. Soc. 2001, 123, 5126.

(11) Francisco, W. A.; Tian, G.; Fitzpatrick, P. F.; Klinman, J. P. J. Am. Chem.

Soc. 1998, 120, 4057.

(22) MacFaul, P. A.; Ingold, K. U.; Wayner, D. D. M.; Que, L., Jr. J. Am. Chem.

Soc. 1997, 119, 10594.

(23) Costas, M.; Tipton, A. K.; Chen, K.; Jo, D.-H.; Que, L., Jr. J. Am. Chem.

Soc. 2001, 123, 6722.

(12) Guroff, G.; Daly, J. W.; Jerina, D. M.; Renson, J.; Witkop, B.; Udenfriend,

S. Science 1967, 157, 1524.

9

2114 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

Scheme 2

Scheme 3

oxidant.^21,22In the particular case of tert-butyl hydroperoxide

(^tBuOOH), O-O bond homolysis was proposed, leading to tert-

butoxy radical (^tBuO^•) and oxoiron(IV).²²Aliphatic C-H bond

hydroxylation then occurs through sequential H-atom abstraction

t

by BuO^•and recombination of the alkyl radical at oxoiron-

(IV), so the metal-centered oxidant is utilized only indirectly.

We sought to modify the TPA ligand to introduce a substituent

that would react directly with the high-valent species, as

observed for 2Cu(I)/O₂,²⁴and 2Ni(II)/H₂O₂couples.²⁵

Accordingly, we reported that substitution of an R-phenyl

substituent for hydrogen on one pyridyl ring of TPA (e.g.,

6-PhTPA, L₁)²⁶resulted in an efficient intramolecular ortho-

hydroxylation when an acetonitrile solution of the complex salt

[(6-PhTPA)Fe^II(NCCH₃)₂](ClO₄)₂is treated with ^tBuOOH,

Scheme 2.²⁷This reaction bears a strong resemblance to that

afforded by PheH and can be considered a model for the

unknown peroxide shunt. We have since broadened the chem-

istry of Scheme 2 to include the range of aryl substituents

summarized in Scheme 3. Methyl substituents also were added

to the other pyridine arms to control the donor properties of

the ligand and the spin state of the iron.²⁸This paper describes

the syntheses of Fe(II) complexes of this ligand range and

reactivity of the solvated cations (i.e., [(L)Fe^II(NCCH₃)₂]²⁺) with

^tBuOOH, including spectroscopic detection of reaction inter-

mediates and characterization of product complexes (i.e.,

[(LO^-)Fe^III(NCCH₃)]²⁺) and free phenol ligands (i.e., LOH).

These results are intended to support further quantitative kinetic

and mechanistic investigation of the hydroxylation reaction.²⁹

2-Methyl-1-phenyl-2-propyl hydroperoxide (MPPH),³⁰bis(6-methyl-

2-pyridylmethyl)amine,³¹6-bromo-2-pyridinecarboxaldehyde,³²ortho-

d₁-bromobenzene,³³and Fe(OTf)₂‚2CH₃CN³⁴were synthesized by

literature procedures. All other reagents were purchased from Aldrich.

Ligand Syntheses. The preparation of the ortho-d₁isotopomer of

L₁(o-d₁-L₁) is described below and outlined in Scheme 3. Other ligands

were obtained in an identical manner by substitution of the appropriate

aryl boronic acid (L₁-L₈) or bis(6-methyl-2-pyridylmethyl)amine (L₈).

1

The ligands were characterized by H NMR spectroscopy (vide infra)

and were free of observable impurities except for trace amounts of

6-aryl-2-pyridylcarbinols in some cases.

Preparation of o-d₁-L₁(6-o-d₁-PhTPA). An aliquot of o-d₁-

bromobenzene (4.00 g, 25.3 mmol) was dissolved in dry THF (50 mL)

in a three-neck flask equipped with a gas inlet and a dropping funnel.

The solution was cooled to -78 °C under a constant purge of dry argon

n

gas. An aliquot (20 mL) of a BuLi solution (1.6 M in hexane) was

transferred by cannula into the dropping funnel and slowly added over

5 min into the cold reaction solution. After the solution was stirred for

1 h, B(OMe)₃(4.00 mL, 36 mmol) was added, and the new solution

was allowed to warm gradually to room temperature overnight. To the

white suspension was cautiously added 1.0 M HCl (50 mL). After the

mixture was stirred for 1 h, the aqueous layer was separated and

extracted with CH₂Cl₂(2 × 50 mL). The extracts were combined with

the organic layer, washed with water, and dried over Na₂SO₄. Removal

of solvent by rotary evaporation yielded crude o-d₁-phenylboronic acid

(1.87 g, 15.2 mmol, 60%). A sample of 6-bromo-2-pyridinecarboxal-

dehyde (2.05 g, 11.0 mmol) was dissolved in toluene (15 mL) in a

three-neck flask capped with a gas inlet, a condenser with a gas outlet

to a bubbler, and a glass stopper. The reaction apparatus was flushed

with argon. Pd(PPh₃)₄(0.30 g, 0.3 mmol) was suspended in anaerobic

toluene (5 mL), and the suspension was added to the reaction flask

against an active argon purge, followed by aqueous 2.0 M Na₂CO₃(10

mL) and the crude o-d₁-phenylboronic acid dissolved in methanol (8

mL). The flask was capped, and the suspension was heated to reflux

for 10 h. A 2.0 M aqueous solution of Na₂(CO₃) (40 mL) and CH₂Cl₂

(100 mL) were added to the cooled solution. The aqueous phase was

separated and extracted with fresh CH₂Cl₂(1 × 25 mL), and the

Experimental Section

Materials. All materials used were ACS reagent grade or better and

were used as received, except for drying and degassing of bulk solvents

by routine methods as required. NMR solvents were used as received

from Cambridge Isotope Labs. Aryl boronic acids were purchased from

Aldrich and Lancaster. Pd(PPh₃)₄was purchased from Lancaster. Bis-

(2-pyridylmethyl) amine was purchased from Richman Chemicals.

(24) (a) Karlin, K. D.; Hayes, J. C.; Gultneh, Y.; Cruse, R. W.; McKown, J.

W.; Hutchinson, J. P.; Zubieta, J. J. Am. Chem. Soc. 1984, 106, 2121. (b)

Karlin, K. D.; Nasir, M. S.; Cohen, B. I.; Cruse, R. W.; Kaderli, S.;

Zuberbuehler, A. D. J. Am. Chem. Soc. 1994, 116, 1324. (c) Itoh, S.; Kondo,

T.; Komatsu, M.; Ohshiro, Y.; Li, C.; Kanehisa, N.; Kai, Y.; Fukuzumi, S.

J. Am. Chem. Soc. 1995, 117, 4714. (d) Itoh, S.; Nakao, H.; Berreau, L.

M.; Kondo, T.; Komatsu, M.; Fukuzumi, S. J. Am. Chem. Soc. 1998, 120,

2890 and references therein. (e) Holland, P. L.; Rodgers, K. R.; Tolman,

W. B. Angew. Chem., Int. Ed. 1999, 38, 1139 and references therein. (f)

Itoh, S.; Taki, M.; Nakao, H.; Holland, P. L.; Tolman, W. B.; Que, L., Jr.;

Fukuzumi, S. Angew. Chem., Int. Ed. 2000, 39, 398. (g) Hayashi, H.;

Uozumi, K.; Fujinami, S.; Nagatomo, S.; Shiren, K.; Furutachi, H.; Suzuki,

M.; Uehara, A.; Kitagawa, T. Chem. Lett. 2002, 416.

(25) (a) Itoh, S.; Bandoh, H.; Nagatomo, S.; Kitagawa, T.; Fukuzumi, S. J. Am.

Chem. Soc. 1999, 121, 8945. (b) Shiren, K.; Ogo, S.; Fujinami, S.; Hayashi,

H.; Suzuki, M.; Uehara, A.; Watanabe, Y.; Moro-oka, Y. J. Am. Chem.

Soc. 2000, 122, 254. (c) Itoh, S.; Bandoh, H.; Nakagawa, M.; Nagatomo,

S.; Kitagawa, T.; Karlin, K. D.; Fukuzumi, S. J. Am. Chem. Soc. 2001,

123, 11168.

(30) Arends, I. W. C. E.; Ingold, K. U.; Wayner, D. D. M. J. Am. Chem. Soc.

1995, 117, 4710.

(31) Da Mota, M. M.; Rogers, J.; Nelson, S. M. J. Chem. Soc. A 1969, 2036.

(32) Cai, D.; Hughes, D. L.; Verhoeven, T. R. Tetrahedron Lett. 1996, 37, 2537.

(33) Renaud, R. N.; Kovachic, D.; Leitch, L. C. Can. J. Chem. 1961, 39, 21.

(34) Hagen, K. S. Inorg. Chem. 2000, 39, 5867.

(26) Chuang, C.-L.; Lim, K.; Canary, J. W. Supramol. Chem. 1995, 5, 39.

(27) Lange, S. J.; Miyake, H.; Que, L., Jr. J. Am. Chem. Soc. 1999, 121, 6330.

(28) Zang, Y.; Kim, J.; Dong, Y.; Wilkinson, E. C.; Appelman, E. H.; Que, L.,

Jr. J. Am. Chem. Soc. 1997, 119, 4197.

(29) Jensen, M. P.; Que, L., Jr. Unpublished results.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2115

A R T I C L E S

Jensen et al.

combined organic extracts were dried overnight (Na₂SO₄). A major

product was revealed by TLC on silica (4:1 v:v CH₂Cl₂-toluene) at R_f

0.59 (vs 0.49 for 2- bromo-6-pyridinecarboxaldehyde). 6-o-d₁-Phenyl-

2-pyridinecarboxyaldehyde was obtained following column chroma-

tography on silica (1:1 v:v CH₂Cl₂-toluene) as a pale yellow oil (1.38

g, 7.5 mmol, 68% vs 6-bromo-2-pyridinecarboxaldehyde, 49% vs

deuterium). Successful separation of 6-o-d₁-phenyl-2-pyridinecarbox-

aldehyde was indicated by appearance in a ¹H NMR spectrum (CDCl₃)

of a single aldehyde resonance at 10.17 ppm, compared to 6-bromo-

2-pyridinecarboxaldehyde at 9.98 ppm.

[(o-d₁-L₁)Fe^II(NCCH₃)₂](ClO₄)₂. Crystals of 6-o-d₁-PhTPA (o-d₁-

L₁, 0.225 g, 0.61 mmol) were dissolved in CH₃CN (5 mL) under a N₂

atmosphere. A solid sample of Fe(ClO₄)₂‚nH₂O (0.190 g, 0.47 mmol,

n ) 8.3) was added, and the yellow-green solution was stirred for 30

min. The reaction solution was poured onto Na₂SO₄and stirred 2.0 h.

The solution was decanted and filtered twice. Dry ether (ca. 20 mL)

was added until the solution was persistently cloudy, and the solution

was allowed to stand overnight at ambient temperature. The product

crystallized as yellow-green rhombic plates, which were washed with

ether and dried to constant mass, 0.228 g (69%). Additional ether was

added to the concentrated mother liquor and a second crop was obtained,

0.060 g (18%). Anal. Calcd (found) for C₂₈H₂₇DCl₂FeN₆O₈: C, 47.75

(47.50); H, 4.01 (4.00); N, 11.93 (11.91).

[(d₅-L₁)Fe^II(NCCH₃)₂](ClO₄)₂. The complex was prepared as de-

scribed above using 6-d₅-PhTPA [d₅-L₁, ²H NMR (δ, ppm in CHCl₃):

8.0 (2D, ortho); 7.5 (3D, para + meta)] and was isolated in two crops

in 70% combined yield. Anal. Calcd (found) for C₂₈H₂₃D₅Cl₂FeN₆O₈:

C, 47.48 (47.37); H, 3.98 (3.99); N, 11.87 (11.86).

[(L₁)Fe^II(NCCH₃)₂](ClO₄)₂. The complex was prepared as described

above using 6-PhTPA (L₁) and isolated as one crop in 70% yield. Anal.

Calcd (found) for C₂₈H₂₈Cl₂FeN₆O₈: C, 47.82 (47.51); H, 4.01 (3.98);

A solution of bis(2-pyridylmethyl)amine (1.65 g, 8.3 mmol) dissolved

in dry methanol (25 mL) was added to the 6-o-d₁-phenyl-2-pyridine-

carboxaldehyde. Thionyl chloride (300 µL) was added with care to

the stirred yellow solution. The flask was capped, and the solution was

stirred 2.0 h before solid NaBH₃CN (1.19 g, 18.9 mmol) and a purge

needle were added. The solution was stirred an additional 3.0 h, and

the reaction was quenched by dropwise addition of 12 M HCl. After

gas evolution ceased, the solution was concentrated to a yellow slurry

by rotary evaporation, to which water (20 mL) was added. The solution

was washed with ether (2 × 30 mL), and concentrated NH₄OH was

added until the solution was basic. The product was extracted into

diethyl ether (3 × 25 mL), and the combined extracts were dried over

Na₂SO₄. The dried solution was concentrated by rotary evaporation

and eluted from a column of neutral alumina (3 cm diameter × 30 cm

length) with 50:1 v:v CH₂Cl₂-CH₃OH to give the product as a pale

yellow oil, 1.49 g (4.1 mmol, 55%), which crystallized on standing.

Analysis of the product ligand by mass spectrometry was consistent

with incorporation of 0.97(1) equiv of deuterium. ¹H NMR (δ, ppm in

CDCl₃): 8.54 (2H, ddd, 4.9, 1.4 and 1.4 Hz, R); 8.03 (1H, 7.5 Hz,

H_o); 7.73 (1H, dd, 7.5 and 7.5 Hz, γ′); 7.69-7.61 (4H, m, â, γ); 7.60

(1H, d, 7.0 Hz, â′); 7.50 (1H, d, 7.0 Hz, â′); 7.47 (2H, m, H_m); 7.40

(1H, t, 7.0 Hz, H_p); 7.17 (2H, dd, 8.0 and 4.9 Hz, â); 4.03 (6H, br s,

N, 11.95 (11.46). E_1/2) +1240 mV vs SCE (∆E ) 70 mV, i_pc/i_pa

)

0.71).

[(L₁)Fe^II(OTf)₂]. Crystals of 6-PhTPA (L₁, 0.150 g, 0.41 mmol) and

a solid sample of Fe(OTf)₂‚2CH₃CN (0.175 g, 0.40 mmol) were

dissolved in THF (5 mL) under a N₂atmosphere, and the resulting

pale yellow solution was stirred for 60 min. The reaction solution was

taken to dryness under vacuum, and the yellow oil was extracted into

CH₂Cl₂. This was repeated once, and then CH₂Cl₂(4 mL) was added

to the oil and a minimum volume of THF was added to obtain a

homogeneous solution. The solution was layered with hexane (12 mL)

to produce yellow needlelike crystals on standing overnight. The mother

liquor was decanted, and the crystals were washed with hexane and

dried to constant mass, 0.256 g (88%). Anal. Calcd (found) for C₂₆H₂₂F₆-

FeN₄O₆S₂‚H₂O: C, 42.29 (42.55); H, 3.28 (3.21); N, 7.59 (7.56); F,

15.44 (15.41).

2

-NCH₂-). H NMR (δ, ppm in CHCl₃): 8.1 (ortho).

Independent Synthesis of 6-ortho-HOC₆H₄-TPA, L₁OH. A sample

of 6-ortho-CH₃OC₆H₄-TPA was obtained by the procedure described

above using ortho-methoxyphenylboronic acid and was characterized

by ¹H NMR spectroscopy (δ, ppm in CDCl₃): 8.54 (2H, d, 5.0 Hz, R);

7.83 (1H, d, 7.0 Hz, H_o); 7.64-7.74 (6H, m, â, γ, â′, γ′); 7.49 (1H, d,

8.0 Hz, â′); 7.36 (1H, ddd, 8.0, 7.5 and 2.0 Hz, H_p); 7.16 (2H, ddd,

6.0, 6.0 and 1.5 Hz, â); 7.07 (1H, ddd, 7.5, 7.5 and 1.0 Hz, H_m); 6.99

(1H, d, 8.0 Hz, H_m′); 4.1 (6H, br s, -NCH₂-); 3.84 (3H, s, -OCH₃).

Solid NaSEt (1.2 g, 11.4 mmol) was added to a solution of the product

(0.76 g, 1.92 mmol) in DMF. The slurry was heated to reflux for 3 h

and then allowed to cool. The yellow solution was poured into water,

and the mixture was acidified with 12 M HCl (10 mL). The aqueous

solution was washed with CH₂Cl₂(3 × 50 mL) and cautiously

neutralized with sodium bicarbonate. The product was extracted into

CH₂Cl₂(3 × 50 mL), and the combined extracts were washed with

water (3 × 50 mL) and dried over MgSO₄. The product was recovered

as a yellow oil in 85% yield following column chromatography on

alumina (CHCl₃-MeOH). ¹H NMR spectroscopy (δ, ppm in CDCl₃):

8.55 (2H, d, 5.0 Hz, R); 7.79-7.69 (7H, m, H_o, â, γ, â′, γ′); 7.45 (1H,

d, 6.0 Hz, â′); 7.32 (1H, ddd, 8.5, 7.0 and 1.8 Hz, H_p); 7.20 (2H, ddd,

6.0, 5.0 and 2.0 Hz, â); 7.04 (1H, dd, 8.5 and 1.2 Hz, H_m′); 6.90 (1H,

ddd, 8.0, 7.0 and 1.0 Hz, H_m); 4.03 (4H, s, -NCH₂-); 3.99 (2H, s,

-NCH₂-).

[(L₂)Fe^II(OTf)₂]. The complex was prepared as described for L₁

above and isolated as yellow crystals in 78% yield. Anal. Calcd (found)

for C₂₆H₂₁F₆FeN₅O₈S₂: C, 40.80 (41.12); H, 2.77 (2.89); N, 9.15 (8.97);

F, 14.89 (14.73). E_1/2) +1240 V vs SCE (∆E ) 70 mV, i_pc/i_pa

0.66). E_pc) -1010, -1160 mV vs SCE (irreversible).

)

[(L₃)Fe^II(OTf)₂]. The complex was prepared as described above and

isolated as yellow crystals in 47% yield. Anal. Calcd (found) for

C₂₇H₂₁F₉FeN₄O₆S₂‚H₂O: C, 40.21 (40.24); H, 2.87 (2.90); N, 6.95

(6.82); F, 21.20 (21.43).

[(L₄)Fe^II(OTf)₂]. The complex was prepared as described above and

isolated as yellow crystals in 50% yield. Anal. Calcd (found) for C₂₆H₂₁-

ClF₆FeN₄O₆S₂: C, 41.37 (41.43); H, 2.80 (2.71); N, 7.42 (7.32); F,

15.10 (14.76).

[(L₅)Fe^II(OTf)₂]. The complex was prepared as described above and

isolated as yellow crystals in 11% yield. Anal. Calcd (found) for

C₂₇H₂₄F₆FeN₄O₆S₂‚H₂O: C, 43.10 (43.44); H, 3.48 (3.41); N, 7.45

(7.44); F, 15.15 (15.16).

[(L₆)Fe^II(OTf)₂]. The complex was prepared as described above and

isolated as yellow crystals in 33% yield. Anal. Calcd (found) for

C₂₇H₂₄F₆FeN₄O₇S₂: C, 43.21 (43.35); H, 3.22 (3.25); N, 7.47 (7.47);

F, 15.19 (15.40).

[(L₇)Fe^II(OTf)₂]. The complex was prepared as described above and

isolated as yellow crystals in 98% yield. Anal. Calcd (found) for

C₂₇H₂₄F₆FeN₄O₇S₂: C, 43.21 (42.91); H, 3.22 (3.31); N, 7.47 (7.29).

E_1/2) + 1210 mV vs SCE (∆E ) 74 mV, i_pc/i_pa) 0.72).

[(L₈)Fe^II(OTf)₂]₂. The complex was prepared as described above

and isolated as white crystals in 48% yield. Anal. Calcd (found) for

C₅₆H₅₂F₁₂Fe₂N₈O₁₂S₄‚0.5CH₂Cl₂: C, 44.08 (43.97); H, 3.47 (3.66); N,

7.26 (7.18); F, 14.81 (14.45). E_pa) + 1440, 1630 mV vs SCE

(irreversible).

Preparation of Iron(II) Complexes. Iron(II) complexes were

synthesized as described below with formulas [(L₁)Fe^II(NCCH₃)₂]-

(ClO₄)₂,²⁷[(L_n)Fe^II(OTf)₂] (OTf^-

)

^-O₃SCF₃, triflate) for L₁-L₇, and

[(L₈)Fe^II(OTf)₂]₂. Caution! Perchlorate salts of organic cations are

potentially explosiVe.³⁵While no difficulties were encountered in the

work described here, functionally identical triflate complexes were also

obtained (vide infra).

(35) (a) Wolsey, W. C. J. Chem. Educ. 1973, 50, A335. (b) Raymond, K. N.

Chem. Eng. News 1983, 61, 1 (December 5), 4.

9

2116 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

t

Reactivity Studies of Complexes with BuOOH. In a typical

was orange and no further changes were observed in UV-visible

spectra. The solvent was taken to dryness under vacuum, and the pasty

orange solids were washed with ether. Addition of solid NaBPh₄to

MeOH extracts induced precipitation of an orange powder in ca. 65%

yield, which was recrystallized from MeOH/DMF. Anal. Calcd (found)

for C₉₆H₈₂B₂Fe₂N₈O₃‚CH₃OH: C, 74.63 (74.86); H, 5.55 (5.47); N,

7.18 (7.40).

experiment, 5.0 µmol of complex (e.g., 3.5 mg of [(L₁)Fe(NCCH₃)₂]-

(ClO₄)₂) was dissolved in dry acetonitrile (5.00 mL) under N₂to give

a pale yellow-green solution of the solvento complex with the general

formula [(L)Fe^II(NCCH₃)₂](X)₂(X ) OTf^-, ClO₄^-). This solution was

transferred to a cuvette and sealed under N₂with a rubber septum. The

cuvette was placed in a cryogenic cell holder on a UV-visible

spectrophotometer and cooled to the desired reaction temperature. An

X-ray Crystallographic Studies. Data collections and structure

solutions were conducted at the X-ray crystallographic facility of the

Department of Chemistry at the University of Minnesota. All calcula-

tions were performed using SGI INDY R4400-SC or Pentium computers

using the current SHELXTL suite of programs.³⁶Suitable crystals were

identified and placed onto the tips of 0.1 mm glass capillaries and

mounted on a Bruker SMART system for data collection at 173(2) K.

Preliminary unit cells were calculated from three sets of 20 frames

oriented so orthogonal wedges of reciprocal space were surveyed. Final

cell constants were calculated from the full set of strong reflections of

the data collection. Data were collected using graphite-monochromated

Mo KR radiation with a frame time of 30 s and a detector distance of

4.9 cm. A randomly oriented region of reciprocal space was surveyed

to the extent of 1.3-1.5 hemispheres to a resolution of 0.84 Å. Three

major swaths of frames were collected with 0.30° steps in ω at three

different φ settings and a detector position of -28° in 2θ. Collected

data were corrected for absorption and decay (SADABS).³⁷In the event

the unit cell was triclinic, additional sets of frames were collected to

better model the absorption correction.

t

aliquot of dry BuOOH (5.5 M in nonane) was diluted with dry CH₃-

CN to a concentration of 150 mM on the benchtop. This stock solution

(120 µL) was injected through the septum to initiate the reaction. Thus,

initial concentrations were 1.0 ( 0.1 mM complex and 3.5 ( 0.5 mM

peroxide. The reaction was monitored to completion, and the solution

was warmed to room temperature and diluted with water (5.0 mL).

Some 12 M HCl was added dropwise until the color was completely

bleached. The pale yellow solution was washed with ether, and

concentrated NH₄OH was added until the solution was alkaline. The

solution was extracted with ether (2 × 5 mL), and the extracts were

washed with water (2 × 5 mL). The extracts were dried under vacuum,

and the white residue was taken up into CDCl₃and characterized by

¹H NMR spectroscopy (vide infra). Integration against a known quantity

of added 2,6-tert-butyl-4-methylphenol was used to quantify the ligand

recovery. A portion of the CDCl₃solution was diluted with CH₃CN

and analyzed by electrospray mass spectrometry.

NIH Shift Analysis. Product ligand from reaction of o-d₁-L₁was

extracted into CDCl₃and characterized by H NMR spectroscopy as

1

described above. Curve fittings (Lorentzian functions) carried out using

Grams/32 Spectral Notebase Version 4.04 (Galactic) gave the relative

areas of the overlapping doublet and doublet of doublet ³J_HHcoupling

patterns for the H_mand H_presonances, which indicated the ratio of

deuterons to protons in the adjacent H_oand H_m′ positions, respectively.

The fits were simplified by use of selective homonuclear decoupling

to collapse ⁴J_HHmultiplicity. Total deuterium retention was confirmed

by mass spectrometry. The data were corrected for product derived

from residual d₀-L₁(3%).

Space groups were assigned on the basis of systematic absences and

intensity statistics. Structures were solved by direct methods. All non-

hydrogen atoms were located in difference Fourier maps and submitted

to full-matrix least-squares refinement with anisotropic displacement

parameters. All hydrogen atoms were placed in idealized positions and

refined as riding atoms with relative isotropic displacement parameters.

Additional details are given in Tables 1 and 2 and below.

[(L₁)Fe^II(NCCH₃)₂](ClO₄)₂.²⁷One perchlorate anion was disordered

over two overlapping, partially occupied positions. A total of 182

restraints were applied (SHELXTL SAME, and DELU) to this and the

other perchlorate to achieve the best model possible. The chlorine U_ij

values were made identical with the SHELXTL constraint EADP due

to the near overlap. The partial occupancies were found essentially to

have a 0.50:0.50 ratio.

[(L₁O^-)₂Fe^III₂(µ-O)](BPh₄)₂.²⁷The bridging oxygen atom was

located on the crystallographic inversion center, so each asymmetric

unit consisted of half the cation.

[(L₈)Fe^II(OTf)₂]‚CH₂Cl₂. Only half of the dimer complex is unique.

The CH₂Cl₂molecule was disordered over two positions (71:29) and

over the inversion center (50:50). The C14 methyl group was best

modeled as rotationally disordered.

MPPH Product Analysis. Reactions were conducted as described

t

for BuOOH above. Reaction solutions were saturated with O₂by

bubbling the pure gas from a balloon through a purge needle for a

period of 20-30 s just prior to addition of the MPPH stock solution.

The product solutions were diluted with a stock solution of naphthalene

in CH₃CN to known concentrations and then passed through silica gel

to remove iron complexes. These solutions were analyzed by gas

chromatography. Controls were run to obtain retention times for CH₃-

CN solutions of authentic bibenzyl, benzyl alcohol, benzaldehyde, and

2-methyl-1-phenyl-2-propyl alcohol and to calibrate the detector

response against the internal naphthalene standard.

Base Titration of [(L₁O^-)Fe^III(OH₂)]²⁺. A ^tBuOOH stock solution

was prepared by adding an aliquot (60 µL) of 70 wt % aqueous solution

to dry CH₃CN (2.00 mL). To a cold (233 K) solution of [(L₁)Fe^II-

(NCCH₃)₂](ClO₄)₂(6.3 mg) in CH₃CN (9.73 mL) was added a 130 µL

aliquot of peroxide stock solution to give a reaction solution with

[complex]₀) 0.9 mM, [^tBuOOH]₀) 2.8 mM, and [H₂O] > 6 mM.

After completion of the reaction, an aliquot (3.50 mL) of the violet

solution was placed in a cuvette maintained at 293 K. A stock solution

(71 mM) of NEt₃(49 µL) was prepared in CH₃CN (4.91 mL); this was

added to the complex solution in 5.0 µL (ca. 0.1 mM) aliquots, and a

UV-visible spectrum was recorded after each addition. Accumulation

of [(L₁O^-)₂Fe^III₂(µ-O)]²⁺was indicated by a gradual color change to

orange and development of an absorption maximum at 486 nm. A Job’s

law plot indicated conversion to the dimeric complex was complete

upon addition of 0.9 ( 0.2 equiv of NEt₃. The color change was

reversed by addition of HClO₄.

Other Physical Methods. ¹H NMR spectra (499.87 MHz) were

recorded on a Varian VI-500 Unity-plus spectrometer and referenced

2

to residual solvent (CHCl₃) 7.26 ppm, CHD₂CN ) 1.94 ppm). H

NMR spectra (46.03 MHz) were recorded on a Varian VXR-300 Unity-

plus spectrometer and referenced to solvent (CDCl₃) 7.26 ppm). ¹⁹

F

NMR spectra (282.15 MHz) were recorded on a Varian VI-300

spectrometer and referenced to external CFCl₃(0.00 ppm). UV-visible

spectra were recorded over a 300-900 nm range on a Hewlett-Packard

8453 diode-array spectrophotometer. Low observation temperatures

were obtained using a dewar fitted with quartz windows, through which

a closed loop of methanol was pumped from a circulating bath (Neslab)

equipped with an immersion cooler (Neslab Cryocool CC100-II) under

control of a thermostat. X-band EPR spectra were obtained at 1.8 K

on a Bruker E-500 spectrometer equipped with an Oxford ESR-10

liquid-helium cryostat. The magnetic field was calibrated with a

gaussmeter, and the microwave frequency was determined with a

Isolation of [(L₁O^-)₂Fe^III₂(µ-O)](BPh₄)₂.²⁷To a rapidly stirred

solution of [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂(30 mg, 43 µmoles) in CH₃CN

(25 mL) was added ^tBuOOH (3.0 equiv) by syringe pump under an N₂

atmosphere. To the blue solution was added NEt₃until the solution

(36) SHELXTL-Plus, v. 5.10; Bruker Analytical X-ray Systems: Madison, WI.

(37) R. Blessing, Acta Crystallogr. A 1995, 51, 33.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2117

A R T I C L E S

Jensen et al.

counter. Resonance Raman spectra were recorded at 4 cm^-1resolution

on an Acton AM-506 spectrophotometer using a Kaiser Optical

holographic super-notch filter with a Princeton Instruments CCD

detector cooled with liquid nitrogen. Data were collected using a

backscattering geometry on frozen CH₃CN samples using 514.5, 568.2,

or 632.8 nm radiation from a 375B CW dye laser (rhodamine 6G)

pumped by a Spectra Physics 2030-15 Ar ion laser. Mass spectra were

obtained either by fast atom bombardment (FAB, LSIMS) ionization

from m-nitrobenzyl alcohol matrixes on a VG 7070E-HF spectrometer

or from CH₃CN solutions by electrospray ionization on a Thermo

Finnigan LCQ spectrometer. Cyclic voltammetric analyses of selected

Fe(II) complexes were carried out using a CS-100 electrochemical

analyzer (Cypress Systems, Inc., Lawrence, KS) interfaced to a three-

electrode system (glassy carbon working, Pt wire counter, Ag wire

reference electrodes). Measurements were made from 2.0 mM analyte

solutions in dry acetonitrile in the presence of 0.10 M NBu₄ClO₄as

supporting electrolyte. Potentials were referenced internally to the

ferrocene couple (+380 mV vs SCE).³⁸Elemental analyses were

performed by Atlantic Microlabs, Inc. (Norcross, GA).

The latter point was illustrated by comparison of the ipso

pyridine resonances, which were coincident among L₁-L₆to

within 0.1 ppm. In contrast, para substitution induced compa-

rable effects within the pyridine, even though the ortho protons

were essentially unaffected.

Hence, the NMR data supported the expectation that meta

substitution affords gross manipulation of electronic properties

at the ortho carbons without significant perturbation of the

chelating properties of the ligand. This point was reflected

further by the electrochemical behaviors of Fe(II)/Fe(III) redox

couples supported by L₁, L₂, and L₇. Quasi-reversible couples

were observed at essentially coincident potentials for these three

ligand complexes by cyclic voltammetry, notwithstanding the

effects of the most electron-withdrawing (L₂, + 1240 mV vs

SCE) and electron releasing (L₇, + 1210 mV) aryl substituents

in comparison to the parent complex (L₁, + 1240 mV).

A second class of ligand modification encompassed by

Scheme 3 was the substitution of methyl groups for hydrogen

at the R-pyridine position, L₈vs. L₁. This modification has been

demonstrated to disfavor low spin states of coordinated iron

ions, presumably by reducing the ligand field strength through

steric repulsion with the metal center.²⁸Resonances of the

pyridine ring protons of L₈were significantly convoluted, but

the phenyl ring resonances were comparable to those of L₁and

a characteristic R-methyl resonance was observed at 2.53 ppm.

Iron(II) Complexes. Homogeneous solutions of Fe(II) ligand

complexes were obtained simply by dissolving a 1:1 mixture

of ligands and Fe(II) salts. A yellow-green crystalline complex

salt [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂was obtained in good yield from

CH₃CN/Et₂O, but this approach failed with other ligands.

Alternatively, use of Fe^II(OTf)₂‚2CH₃CN in THF generally

afforded bright yellow crystals of neutral complexes [(L_n)Fe^II-

(OTf)₂] from CH₂Cl₂/n-hexane for L₁-L₇, as well as colorless

crystals of [(L₈)Fe^II(OTf)₂]₂. These dissolved in CH₃CN to yield

outer sphere salts of bis(solvento) complex cations analogous

to the perchlorates (vide infra). The complexes were air-stable

as solids but decomposed slowly in aerobic solutions.

Results

Syntheses and ¹H NMR Characterization of Ligands. The

ligands used in this study, and their syntheses are summarized

in Scheme 3. The approach was characterized by significant

flexibility and generality and was applicable to a broad range

of ligands. Thus, the aryl substituents were introduced onto

6-bromo-2-pyridinecarboxaldehyde by catalytic Suzuki cou-

pling,³⁹and bis(2-pyridylmethyl)amines were added subse-

quently across the aldehyde in a reductive amination.⁴⁰In

principle, the synthetic scheme could be simplified further by

reversing the order of the steps, i.e., coupling aryl boronic acids

to a common 6-bromo-TPA intermediate.⁴¹However, purifica-

tion of coupling products proved to be easier to achieve and

verify with the pyridinecarboxaldehydes, so the longer approach

was adopted.

The series of ligands depicted in Scheme 3 was intended to

introduce a range of R-aryl substituents with modified suscep-

tibility to electrophilic aromatic substitution at the ortho carbons.

Accordingly, a series of functional groups (-NO₂, -CF₃, -Cl,

-CH₃, -OCH₃) was placed at one meta position to give L₂-

L₆derivatives of 6-PhTPA, L₁. Substituents placed in this

position were expected to exert significant electronic effects on

the ortho carbons, while minimizing electronic and steric effects

on the ipso pyridine and the chelated metal ion.

The solid-state structures were determined for three repre-

sentative complexes, [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂, (L₆)Fe^II(OTf)₂,

and [(L₈)Fe^II(OTf)₂]₂(Tables 1 and 2 and Figure 1). The first

two complexes adopt the expected and essentially identical

pseudooctahedral geometries characteristic of tetradentate TPA

ligands, such that the unsubstituted 2-pyridylmethyl arms are

arranged in a cis, rather than the more symmetric trans,

configuration. This arrangement affords open cis sites opposite

the amine and one unsubstituted pyridine that are filled by two

exogenous ligands, either CH₃CN or triflate. The trans bite

angles between opposing pyridines are significantly constrained

from linearity (i.e., 151.7(1) and 151.12(7)°, respectively) by

the methylene arms, but such angles to the monodentate solvent

or triflate ligands are nearly linear. The coordinate bond of the

monodentate ligand is significantly shorter trans to the amine

than to the pyridine, by 0.051 Å for CH₃CN and 0.131 Å for

triflate. However, almost all coordinate bonds in both complexes

are in a narrow range of 2.18-2.22 Å, consistent with a high-

spin Fe(II) state (S ) 2).²⁸One outlying bond distance (2.262

Å) is to the substituted pyridine in the L₆complex, which is

0.06 Å longer than the corresponding bond in the L₁complex.

Because of the greater variation in coordinate bond lengths, the

(L₆)Fe^II(OTf)₂structure appears particularly unsymmetric.

1

Comparison of H NMR spectra of the substituted ligands

(Table S1) indicated the efficacy of this approach. A set of

resonances common to all the ligands (L₁-L₇) was observed

in CDCl₃solutions for the six methylene protons and the

unsubstituted pyridines. The remaining signals were affected

by the introduced substituents. Resonances of the unsubstituted

arene of L₁appeared at 8.01 ppm (ortho), 7.49 ppm (meta),

and 7.40 ppm (para); these resonances were absent in the

spectrum of d₅-L₁, and the assignments based on J-couplings

were confirmed by the COSY technique. As expected, the ortho

and para proton resonances were strongly deshielded by

electron-withdrawing meta substitutents, varying over a more

than 1.3 ppm range between L₂and L₆, while the remaining

meta proton and the ipso pyridine protons were scarcely affected.

(38) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877.

(39) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457.

(40) Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem. Soc. 1971, 93,

2897.

(41) Mandon, D.; Nopper, A.; Litrol, T.; Goetz, S. Inorg. Chem. 2001, 40, 4803.

9

2118 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

Table 1. Summary of the X-ray Structure Determinations

II

III

parameter

formula

formula mass (amu) 703.31

[(L₁)Fe (NCCH ) ](ClO )

(L₆)Fe (OTf)₂

[(L₈)Fe (OTf)₂]₂‚CH Cl₂

[(L₁O^-)Fe (O)_0.5](BPh₄)

3 2

4 2

2

C₂₈H₂₈N₆O₈Cl₂Fe

C₂₇H₂₄N₄O₇F₆O₂S₂Fe

750.47

C₅₇H₅₄N₈O₁₂F₁₂S₄Cl₂Fe₂

1581.92

C₄₈H₄₁N₄O_1.5BFe

764.51

cryst habit, size (mm) yellow prism, 0.30 × 0.14 × 0.13 yellow slab, 0.40 × 0.13 × 0.13 colorless block, 0.21 × 0.16 × 0.05 red plate, 0.35 × 0.18 × 0.04

space group

a (Å)

b (Å)

P2₁/n (monoclinic)

10.0768(4)

11.0129(5)

28.3180(12)

90

91.302(1)

90

3141.8(2)

4

P2₁/c (monoclinic)

12.013(2)

9.933(2)

25.193(2)

90

93.057(4)

90

3001(1)

4

P1h (triclinic)

9.046(1)

10.855(1)

16.774(2)

88.952(2)

88.918(2)

87.495(2)

1645.0(3)

1

P1h (triclinic)

9.8427(7)

13.9820(9)

15.508(1)

64.763(1)

77.158(1)

88.792(1)

1875.6(2)

2

c (Å)

R (deg)

â (deg)

γ (deg)

V, Å³

Z

D

_calc(g cm^-3

)

1.487

1.661

1.597

1.354

GOF

1.080

1.038

1.016

1.015

R (I > 2σ(Ι))

R_w

0.065

0.100

0.036

0.095

0.049

0.092

0.039

0.099

Table 2. TPA Coordinate Bonds for Iron Complexes (Bond Lengths in Å)

II

bond

trans donor [(L₁)Fe (NCCH ) ]²⁺(tetradentate N , cis) [(L₆)Fe (OTf)₂] (tetradentate N , cis) [(L₈)Fe (OTf)₂]₂(tridentate N , trans) [(L₁O^-)₂Fe^III₂(O)]²⁺(pentadentate N O)

3 2

4

3

4

Fe-amine

Fe-py

2.217(3) (N1)

2.193(3) (N2)

2.217(2) (N4)

2.181(2) (N3)

2.202(3) (N1)

2.210(2) (N1)

2.144(2) (N2)

2.167(2) (N3)

Phpy/py

NCCH₃

OTf

Mepy

2.211(3) (N3)

2.215(2) (N2)

2.262(2) (N1)

Fe-Mepy

Fe-Phpy

2.229(3) (N2)

2.225(3) (N3)

py

2.207(3) (N4)

oxo

2.191(2) (N4)

Fe-NCCH₃py

amine

py

2.159(4) (N6)

2.108(4) (N5)

Fe-OTf

2.189(2) (O61)

2.058(2) (O51)

amine

OTf

2.126(2) (O3A)^a

2.278(2) (O1)^a

2.119(2) (O4)

Fe-oxo

Fe-OPh

Phpy

amine

1.7897(3) (O1)

1.901(1) (O2)

^aµ₂-bridging.

The structure of [(L₈)Fe^II(OTf)₂]₂is unique in several respects

(Figure 1, Table 2). The TPA ligand was found to be tridentate,

with the 6-phenyl-2-pyridylmethyl arm detached from the

coordination sphere of the metal. Similar tridentate coordination

was reported recently for Fe^IIICl₃complexes of 6-substituted

TPA ligands.⁴¹The 6-methyl substituents presumably destabilize

tetradentate coordination, but this effect is difficult to gauge;

even 6-Ph₃TPA can be completely bound, as in the trigonal

bipyramidal complex [LCu^II(NCCH₃)](ClO₄)₂.²⁶The remaining

coordination sites are filled with triflates to complete a distorted

octahedron, one monodentate terminal and two bidentate bridg-

ing, and form an unsupported anti-anti ditriflato-bridged dimer

with an Fe(II)‚‚‚Fe(II) separation of 5.774(1) Å. Although triflate

bridges are observed frequently in complexes of noble metals,

[Me₃Pt^IV(µ-OTf)]_n(n ) 2, 4) serving as striking examples,⁴²

apparently this is the initial structural report of a well-ordered

triflate bridge between iron atoms in any oxidation state. The

only other report is for a disordered syn-syn µ₂-triflato bridge

between Fe(II) centers separated by 4.120(3) Å.⁴³A syn-anti

diacetato bridge with an Fe(II)-Fe(II) distance of 4.288(2) Å

was reported for [(TPA)Fe^II(OAc)]₂(BPh₄)₂.⁴⁴The ditriflato

bridge obtained here is not symmetric, displaying Fe(II)-OTf

bond lengths of 2.126(2) and 2.278(2) Å and respective Fe-

(II)-O-S angles of 155.1(2) and 129.7(1)°. Nevertheless, these

parameters approach those for coordination of the terminal

triflato ligand, 2.119 (2) Å and 144.9(2)°, as well as for ligation

of the terminal triflates of (L₆)Fe^II(OTf)₂, 2.058(2) and 2.189-

(2) Å and 144.5(1) and 137.4(1)°. Direct comparison of

coordinate bonds situated trans to the amine indicates the

bridging triflate in the L₈complex is displaced from the metal

by 0.07 Å relative to the terminal triflate of (L₆)Fe^II(OTf)₂.

Unlike the other two complexes, the 6-methyl-2-pyridylmethyl

arms fold into a trans configuration, which accommodates the

meridional array of triflate oxygen donor atoms. The average

Fe(II)-N distance of 2.22 Å again is indicative of a high-spin

configuration on the metal ion.

The various complexes of L₁-L₈were characterized at room

1

temperature by H NMR spectroscopy in CD₃CN solutions as

high-spin bis(solvento) complex cations of general formula [(L)-

Fe^II(NCCD₃)₂]²⁺. The counteranions (ClO₄^-, OTf^-) were

spectroscopically innocent under these conditions, as the spectra

obtained from [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂and (L₁)Fe^II(OTf)₂

were identical except for the resonance of free CH₃CN.

Observed pyridyl resonances (Table S2) were broad, ca. 40-

250 Hz, and subject to large isotropic shifts over a ca. -10 to

(42) Schlecht, S.; Magull, J.; Fenske, D.; Dehnicke, K. Angew. Chem., Int. Ed.

Engl. 1997, 36, 1994.

(43) (a) Herold, S.; Pence, L. E.; Lippard, S. J. J. Am. Chem. Soc. 1995, 117,

6134. (b) Herold, S.; Lippard, S. J. J. Am. Chem. Soc. 1997, 119, 145.

(44) Me´nage, S.; Zang, Y.; Hendrich, M. P.; Que, L., Jr. J. Am. Chem. Soc.

1992, 114, 7786.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2119

A R T I C L E S

Jensen et al.

Figure 1. ORTEP diagrams of the complex cation of [(L₁)Fe^II(NCCH₃)₂]²⁺(top left) and the (L₆)Fe^II(OTf)₂(top right) and [(L₈)Fe^II(OTf)₂]₂(bottom)

complexes (ellipsoids drawn at 50% level; hydrogen atoms omitted for clarity).

106 ppm range, indicating that the paramagnetic S ) 2 spin

state was retained for Fe(II) in solution (Figure 2). The

2-pyridylmethyl signals were analogous to those reported for

resonances in the diamagnetic region observed for the L₇

complex cation were assigned to the ortho and meta protons.

The dimeric [(L₈)Fe^II(OTf)₂]₂complex dissolved in CH₃CN

to yield an analogous [(L₈)Fe^II(NCCD₃)₂]²⁺complex and free

triflate counteranions, as observed by NMR spectroscopy (Figure

[(6-MeTPA)Fe^II(NCCD₃)]^{2+ 28}

, as no systematic trends in chemi-

cal shifts were induced by arene substitution. The paramagnetic

contact shifts are governed primarily by σ-spin delocalization

effects,⁴⁵so that the shifts of ligand resonances increase with

proximity to the metal ion: R, NCH₂Py > â, â′ > γ, γ′. There

was no basis to distinguish the R-pyridyl and methylene

resonances. Assignments of the sharper â, â′ and γ, γ′

resonances were obvious from a 2:1 intensity pattern and cross-

peaks were detected by the COSY technique for L₁. Appearance

of a 2:1 pattern was consistent with effective mirror plane

symmetry bisecting the unsubstituted pyridines and along the

phenyl-substituted arms; thus, refolding of the unsymmetric

ligand field evident in the solid state must be dynamic on the

NMR time scale.

Three resonances observed near the diamagnetic region of

the (L₁)Fe^II(NCCD₃)₂]²⁺spectrum in a 2:1:2 intensity ratio at

2.3, 0.1, -0.8 ppm were assigned to the meta, para, and ortho

protons of the phenyl ring, respectively, as the latter was

broadened by the proximal paramagnet. These signals were

absent in the spectrum of the d₅-Ph isotopomer (Figure 2B).

Substitution at the meta position breaks the ring symmetry, so

that four phenyl resonances of equal intensity were observed

for complexes of L₂-L₆. Similarly, two equally intense

1

2C). The H NMR spectrum was comparable to that of the L₁

complexes, except paramagnetically shifted signals were more

widely dispersed, and the downfield R pyridyl proton resonance

was replaced by an intense methyl signal shifted upfield to

-29.1 ppm by spin polarization.⁴⁵The phenyl ring protons were

essentially unshifted, appearing at 7.7 (meta), 7.3 (ortho), and

6.9 (para) ppm, and the ortho resonance was broadened

significantly. The ¹⁹F spectrum consisted of a single, sharp (∆ν_1/2

) 80 Hz) resonance at -72.2 ppm. The complex yielded quite

1

different spectra when dissolved in CD₂Cl₂, however. The H

spectrum consisted largely of several poorly resolved signals

in the diamagnetic region, displaying only three broad, para-

magnetically shifted resonances in a 1.0:1.0:1.5 intensity ratio

at 51.7, 43, and -23 ppm. A single broad (∆ν_1/2) 900 Hz)

peak was observed in the ¹⁹F spectrum at -47.9 ppm. These

data are most consistent with a structure like that determined

by X-ray crystallography, which retains bound triflates and two

coordinated 6-methyl-2-pyridylmethyl arms giving paramag-

netically shifted and unresolved â, methylene, and methyl

resonances, as well as an uncoordinated 6-phenyl-2-pyridyl-

1

methyl arm. In contrast, H NMR spectra of (L₁)Fe^II(OTf)₂in

(45) Ming, L.-J. In Physical Methods in Bioinorganic Chemistry; Que, L., Jr.,

Ed.; University Science Books: Sausalito, CA, 2000; Chapter 8.

CD₃CN and CD₂Cl₂were quite similar, exhibiting a slightly

9

2120 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

Figure 2. ¹H NMR spectra (500 MHz) of the following: [(L₁)Fe^II-

(NCCH₃)₂](ClO₄)₂(A); [(d₅-L₁)Fe^II(NCCH₃)₂](ClO₄)₂(B); [(L₈)Fe^II(OTf)₂]₂

(C) recorded at 293 K as CD₃CN solutions.

Figure 4. Visible spectra of CH₃CN solutions of [(L)Fe^II(NCCH₃)₂]²⁺(A),

[(L)Fe^III(OO^tBu)]²⁺intermediate (B), and [(LO^-)Fe^III]²⁺product (C), for

L₁recorded at 228 K (bottom) and L₈recorded at 293 K (top).

substituted (L₈) complex cation even in organonitrile solution,

is consistent with the weaker ligand fields afforded to the metal

1

ion. Similar behavior was observed by H NMR spectroscopy

for closely related complexes. [(TPA)Fe^II(OTf)₂] exhibited

paramagnetically shifted resonances in CD₂Cl₂but not in CD₃-

CN due to formation of the low-spin [(TPA)Fe^II(NCCH₃)₂]²⁺;^28,46

however, increasing paramagnetic shifts were observed for the

latter at elevated temperatures (e.g., δ(R-py) ) 60 ppm at 353

K).⁴⁶X-ray crystallography also indicated a low-spin solid-state

structure for [(TPA)Fe^II(NCCH₃)₂](ClO₄)₂and a high-spin

configuration for [(6-Me₃TPA)Fe^II(NCCH₃)₂](ClO₄)₂, based on

Fe-N bond lengths.²⁸Unfortunately, the putative spin crossover

of [(L₁)Fe^II(NCCD₃)₂]²⁺was not conveniently observable by

¹H NMR spectroscopy near the freezing point of CD₃CN (ca.

230 K).

Figure 3. Temperature-dependent UV-visible spectra of [(L₁)Fe^II-

(NCEt)₂]²⁺in EtCN solution.

wider paramagnetic shift range in the latter solvent. Analogous

results were obtained in CD₂Cl₂for (L_n)Fe(OTf)₂(n ) 2-7).

Solutions of the neutral triflate complexes of L₁-L₇in CH₂-

Cl₂were invariably pale yellow, and the UV-visible spectra

accordingly were nearly featureless, consisting only of a slight

shoulder (λ_max) 350-380 nm, ꢀ_max) 500-800 M^-1cm^-1) on

the tail of intense pyridine π-π* bands in the UV region. The

same was true of [(L₈)Fe^II(OTf)₂]₂dissolved either in CH₂Cl₂

or CH₃CN. In contrast, UV-visible spectra recorded from

organonitrile solutions of [(L_n)Fe^II(NCCH₃)₂]²⁺(n ) 1-7)

contained strong features at 360 ( 5 and 400 ( 5 nm

presumably assignable to iron(II)-to-pyridine MLCT bands, as

well as a broad weak ligand field band at 550 nm (Figure 3).

The extinctions of these bands were found to be strongly

temperature-dependent, as illustrated for a propionitrile (EtCN)

solution of [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂(Figure 3). This absorp-

tion imparted distinct color to solutions even at room temper-

ature, and cooling of propionitrile solutions below ca. 210 K

resulted in a reversible color change to brick red.

t

Reaction with BuOOH. UV-visible spectroscopy is a

convenient method to monitor reactions of the complexes

t

with BuOOH in CH₃CN (Figure 4). Addition of the peroxide

([^tBuOOH]₀e 3.5 mM) to 1.0 mM solutions of the complexes

prompted eventual development of stable and intense visible

chromophores (e.g., spectrum C in Figure 4 bottom, λ_max) 780

nm for L₁). The L₂-L₅complexes also reacted to form similar

chromophores, which were blue-shifted relative to the L₁product

for L₂-L₄by the presence of electron-withdrawing aryl sub-

stituents and red-shifted for L₅. The product solutions of the

The color change is consistent with well-known spin cross-

over of d⁶Fe(II) complexes between S ) 2 and S ) 0 spin

states.^46,47The failure to observe color changes in CH₂Cl₂

solutions of the triflato complexes, and for the 6-methyl

(46) Diebold, A.; Hagen, K. S. Inorg. Chem. 1998, 37, 215.

(47) (a) Gu¨tlich, P. Struct. Bonding 1981, 44, 83. (b) Toftlund, H. Coord. Chem.

ReV. 1989, 94, 67. (c) Gu¨tlich, P.; Garcia, Y.; Goodwin, H. A. Chem. Soc.

ReV. 2000, 29, 419.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2121

A R T I C L E S

Jensen et al.

Scheme 4

temperature to determine that the maximum spectrophotometric

yield corresponds to 87% phenolate formation (ꢀ_max) 2650

M^-1cm^-1/iron at 298 K).

A mass spectrum of the blue L₁product solution, obtained

by electrospray ionization, contained two parent cations at m/z

536 and 939, whose masses and observed isotope distributions

were consistent with their respective formulation as [(L₁O^-)-

Fe^III(ClO₄)]⁺and {[(L₁O^-)₂Fe^III₂(O)](ClO₄)}⁺. The orange

solutions obtained by NEt₃addition gave only the latter ion,

and a cation with m/z 1209 was observed from a solution of

the isolated tetraphenylborate salt, consistent with the formula-

tion [(L₁O^-)₂Fe^III₂(O)](BPh₄)]⁺. The L₈product solution yielded

parent ions at m/z 614 and 1095 corresponding to [(L₈O^-)-

Fe^III(OTf)]⁺and [(L₈O^-)₂Fe^III₂(O)](OTf)]⁺, respectively.

The mass spectral and UV-visible data suggest that the

immediate products of L₁and L₈complex oxidation are

[(L₁O^-)Fe^III(X)]²⁺and [(L₈O^-)₂Fe^III₂(O)]²⁺, respectively. The

progressive blue shift of the visible absorption maximum for

[(L₁O^-)Fe^III(X)]ⁿ⁺reflects equilibria at the sixth ligand site

(Scheme 4), presumably displacement of CH₃CN with more

basic aqua and oxide ligands, which diminish the Lewis acidity

of the iron(III) center.⁴⁸Similarly, spontaneous formation of

the (L₈O^-)Fe^IIIoxo-bridged dimer in solution and the red shift

of its chromophore compared to that of the orange L₁product

are consistent with the decreased basicity of the 6,6-dimethyl-

substituted ligand.

Figure 5. Spectrophotometric yields of the blue [(L₁O^-)Fe^III(NCCH₃)]²⁺

species, observed at 780 nm, as a function of added equivalents of ^tBuOOH

to [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂(1.05 ( 0.07 mM) in CH₃CN at 228 K.

Examination of the orange L₁product complex by X-ray

crystallography confirmed an oxo-bridged dimeric structure of

crystallographically equivalent iron(III) complexes of modified

N₄O-pentadentate ligands, [(L₁O^-)₂Fe^III₂(µ-O)](BPh₄)₂(Tables

1 and 2 and Figure 7). The oxo bridge is symmetrically

constrained to linearity. The phenolate oxygen is situated trans

to the amine, so that these form a meridional donor array along

with the substituted pyridine. This rigid arrangement must

develop during the hydroxylation reaction, which suggests a

specific geometry for reactive intermediates (vide infra). The

ipso torsional angle of 28(1)° accommodates an Fe^III-O bond

length of 1.901(1) Å and a nearly linear O2-Fe1-N1 angle of

165.45(6)°. The unsubstituted 2-pyridylmethyl arms fold into a

trans arrangement on either side of the phenolate, as opposed

to the cis arrangement found for [(L₁)Fe^II(NCCH₃)₂]²⁺. An

average pyridine coordinate bond length of 2.18 Å indicates a

high-spin (S ) 5/2) state for the iron(III) ions.²⁸

Figure 6. Spectrophotometric titration of the violet [(L₁O^-)Fe^III(OH₂)]²⁺

species with triethylamine to afford orange [(L₁O^-)₂Fe^III₂(µ-O)]²⁺at 293

K.

most electron-rich arenes (i.e., L₆, L₇) bleached to give an

uncharacterized yellow solution, presumably due to overoxi-

dation. Partial bleaching of the other products was problematic

only at higher peroxide stoichiometries and warmer tempera-

tures. The properties of these visible chromophores are consis-

tent with their assignment to phenolate-to-iron(III) LMCT bands

afforded by ortho-hydroxylation of the pendant arenes. For

complexes of L₁-L₇, the reactions were complete within

seconds at room temperature, or after several hours at 228 K,

and were not affected by the presence of dioxygen. [(L₈)Fe^II-

t

(NCCH₃)₂]²⁺reacted slowly with 3.5 mM BuOOH at room

temperature, and a stable magenta chromophore (spectrum C

1

The H NMR spectrum of [(L₁O^-)₂Fe^III₂(µ-O)](BPh₄)₂in

in Figure 4 top, λ_max) 525 nm) was obtained after several hours.

CD₃CN solution consisted of broad (ca. 100 Hz) signals

distributed over a narrow chemical shift range (-2 to 30 ppm),

indicative of the expected strong antiferromagnetic coupling of

the S ) 5/2 metal ions through the oxo bridge (Figure S1).⁴⁹

Given ideal C_2hsymmetry, the 42 protons of the dimeric

complex cation will give rise to 14 total resonances, 7 each

with 2H and 4H intensities. Eight signals corresponding to 32

t

Addition of variable quantities of BuOOH to solutions of

[(L₁)Fe^II(NCCH₃)₂]²⁺(average concentration ) 1.05 ( 0.07

mM) under N₂at 228 K established that a minimum peroxide

to complex stoichiometry of 1.5:1.0 was necessary and sufficient

to obtain the maximum spectrophotometric yield (Figure 5).

Further additions of peroxide in modest excess did not affect

the yield. Product solutions exhibited a color change to violet

(λ_max) 600 nm) in the presence of excess water (ca. 10 mM)

at 293 K, and titration of such a solution with 0.9 ( 0.2 equiv

of NEt₃prompted a further color change to orange (λ_max) 486

nm) (Figure 6). Crystals of the orange product obtained by

metathesis with NaBPh₄were redissolved in CH₃CN at room

(48) (a) Pyrz, J. W.; Roe, A. L.; Stern, L. J.; Que, L., Jr. J. Am. Chem. Soc.

1985, 107, 614. (b) Que, L., Jr. In Biological Applications of Raman

Spectroscopy; Spiro, T. G., Ed.; John Wiley & Sons: New York, 1988;

Vol. 3, pp 491-521.

(49) (a) Wu, F.-J.; Kurtz, D. M., Jr. J. Am. Chem. Soc. 1989, 111, 6563. (b)

Norman, R. E.; Yan, S.; Que, L., Jr.; Backes, G.; Ling, J.; Sanders-Loehr,

J.; Zhang, J. H.; O’Connor, C. J. J. Am. Chem. Soc. 1990, 112, 1554.

9

2122 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

The orange dinuclear [(L₁O^-)₂Fe^III₂(µ-O)]²⁺complex was

also characterized by resonance Raman spectroscopy, using

514.5 nm laser excitation of a frozen CD₃CN solution (Figure

8B). Several resonance-enhanced phenolate modes were ob-

served in the 600-1600 cm^-1region, but the spectrum was

distinct from that of the mononuclear blue species. In particular,

the ν(Fe^III-OAr) peak downshifted to 619 cm^-1, consistent with

the weakening of the iron(III)-phenolate interaction in the

presence of an oxo bridge. Consistent with its formulation as

[(L₈O^-)₂Fe^III₂(µ-O)]²⁺, a spectrum of the L₈product solution

was very similar to that of [(L₁O^-)₂Fe^II(µ-O)]²⁺(Figure 8C).

2

Because of the stability of the phenolate chromophores for

L₁-L₅and L₈, the complexes could be disrupted by acid

treatment, and the free ligands could be recovered by extraction

Figure 7. ORTEP diagram of the complex cation [(L₁O^-)₂Fe^III₂(µ-O)]²⁺

(ellipsoids drawn at 50% level; hydrogen atoms omitted for clarity).

1

into CDCl₃with a typical efficiency of 60-70%. H NMR

spectroscopy indicated that the extract solutions were mixtures

of o-hydroxylated (50-90%) ligands and residual unmodified

precursors (Table 3). No products were observed that would

indicate competing oxidative pathways, such as 2-pyridinecar-

boxaldehydes afforded by N-dealkylation of the tertiary amine.⁵¹

The ¹H NMR spectrum of recovered L₁product matched exactly

that of a sample of L₁OH prepared by an independent route;

resonances assigned to the phenol ring were absent in the d₅-

L₁product spectrum.

Fractional yields of ortho-hydroxylated ligands obtained from

consistent 3.5:1.0 ^tBuOOH-complex stoichiometries were not

strongly perturbed by the nature of the meta substituent.

However, two product isomers are possible for L₂-L₅, with

hydroxylation occurring at either ortho carbon, ortho or para

to the substituent. Only the para-nitrophenol isomer was

observed for L₂OH (Figure 9). In contrast, L₃(m-CF₃) and L₄

(m-Cl) yielded both isomers, with the ortho rotamer slightly

favored. This was confirmed in the former case by ¹⁹F NMR

spectroscopy, which revealed two product resonances (-61.8,

-63.4 ppm) near the resonance of the starting ligand (-63.0

ppm), all of approximately equal intensity. The spectrum of L₅

product was consistent with phenol formation but was suf-

ficiently convoluted to prevent assignment of specific isomers.

The L₆and L₇phenolate complexes apparently were susceptible

to overoxidation, and no organic products were isolated by

extraction. The phenol resonances of the L₈product were

essentially identical to those observed for L₁.

Figure 8. Resonance Raman spectra of blue [(L₁O^-)Fe^III(NCCH₃)]²⁺(A,

632.8 nm excitation), orange [(L₁O^-)₂Fe^III₂(O)]²⁺(B, 514.5 nm excitation),

and magenta [(L₈O^-)₂Fe^III₂(O)]²⁺(C, 514.5 nm excitation) chromophores

in frozen CD₃CN solutions. Solvent features are denoted (s). See Table S3

for a listing of frequencies.

protons were observed at 31 (8H), 16.9 (8H), 15.8 (4H), 15.2

(2H), 12.5 (2H), 11.5 (2H), 9.4 (2H), and -1.9 ppm (4H).

Integration indicated the signals of eight more protons were

-

coincident with resonances of the BPh₄counterions. The

remaining two protons were not accounted for but could have

been coincident with the residual solvent resonance. Thus, the

observed NMR spectrum was consistent with the dimeric

structure determined by X-ray crystallography.

The NMR product solutions for L₁-L₅were characterized

in parallel by mass spectrometry (Table S4). In each case, two

parent peaks were observed at masses calculated for protonated

monocations of the phenyl and phenol species and with the

expected isotopomeric distributions. Consistent with oxygen

atom addition, the parent ions were separated by 16 amu,

The blue [(L₁O^-)Fe^III(X)]²⁺product was further characterized

by resonance Raman spectroscopy of frozen acetonitrile solu-

tions using 632.8 nm laser excitation into the LMCT band

(Figure 8A). The typical response of para-substituted phenolates

(e.g., tyrosinate) coordinated to iron(III) consists of four bands

corresponding to aromatic deformations near 1170, 1270, and

1500 and 1600 cm^-1, respectively, as well as a band near 600

cm^-1assigned to the ν(Fe^III-OAr) mode.⁴⁸These modes split

into groups of additional bands for phenolates with lower

symmetry, especially for those with ortho-imine substituents

(e.g., salen complexes).⁵⁰In so far as the L₁phenolate complex

is similar, the present results can be interpreted analogously

(Table S3). Notably, there is an intense band at 644 cm^-1that

can be assigned to the ν(Fe^III-OAr) mode.

appearing at m/z 367 and 383 for L₁H⁺and L₁OH₂ions,

+

respectively. Moreover, d₅-L₁derived ions at m/z 372 and 387,

the 15 amu difference demonstrating that oxygen atom addition

was coupled to C-D bond labilization.

Spectroscopic Detection of Reaction Intermediates. The

hydroxylation reactions of the L₁-L₇complexes were extremely

rapid at room temperature, requiring only seconds for the stable

phenolate chromophores to develop at millimolar concentrations.

However, at 258 K monotonic development of the phenolate-

to-iron(III) LMCT chromophore was observed over a ca. 20-m

(50) Carrano, C. J.; Carrano, M. W.; Sharma, K.; Backes, G.; Sanders-Loehr,

(51) Mahapatra, S.; Halfen, J. A.; Tolman, W. B. J. Am. Chem. Soc. 1996, 118,

J. Inorg. Chem. 1990, 29, 1865.

11575.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2123

A R T I C L E S

Jensen et al.

Table 3. ¹H NMR Data for Extracted Phenols (CDCl₃Solutions, δ in ppm, J_HHin Hz)

ortho isomer

para isomer

phenol

meta-X

H

o

H_m

H

p

mole fraction^a

H

o

H

p

H_m

mole fraction^a

L₁OH

H

7.76

6.91 (ddd; 7.4, 1.5) 7.32 (ddd; 7.1, 1.5)

b

7.05 (dd; 8.3)

0.85

0.62

0.32

0.35

0.46

L₂OH NO₂

L₃OH CF₃

L₄OH Cl

0.00

0.39

0.48

b

8.75 (d; 2.5) 8.21 (dd; 9.0, 3.0) 7.10 (d; 9.5)

7.95 (d; 7.0) 6.94 (dd; 7.7)

7.66 (8.0)

7.78 (dd)

7.60

7.42

8.03 (s)

7.54

7.27

b

7.11 (d; 8.5)

6.99 (d; 8.5)

7.03 (dd; 8.5)

6.86 (dd; 8.0)

L₈OH

H

6.89 (ddd; 8.5, 1.2) 7.30 (ddd; 6.8, 1.7)

b

^aFraction of total extracts; balance is unreacted ligand. ^bDegenerate.

Figure 10. Resonance Raman spectra of [(L)Fe^III(OO^tBu)]²⁺peroxo

intermediates recorded from frozen CH₃CN solutions for L₈(A, 632.8 nm

excitation) and L₁(B, 568.2 nm excitation, and C, 632.8 nm excitation).

Dotted lines denote features corresponding to high-spin (A and B) and low-

spin (B and C) isomers. Solvent features are denoted (s).

Figure 9. ¹H COSY NMR spectrum of extracts from reaction of ^tBuOOH

and [(L₂)Fe^II(NCCH₃)₂]²⁺in CDCl₃solution. Cross-peaks are indicated for

phenyl resonances of L₂and phenol resonances of L₂OH. Only one phenol

isomer is apparent.

ence of two distinct alkylperoxo species was clearly indicated

by different degrees of resonance enhancement at various

excitation wavelengths, with longer wavelengths favoring

enhancement of features of the low-spin isomer. The red shift

of the absorption maximum for the low-spin isomer relative to

that for the high-spin isomer is also consistent with results

obtained for the TPA series.²⁸The presence of spin isomers

was supported by an EPR spectrum of the mixture, which

consists of a signal at g ) 8.4, corresponding to a high-spin S

) 5/2 iron(III) ion, and signals at g ) 2.18, 2.11, and 1.97

assignable to a low-spin S ) 1/2 iron(III) ion. (Table 4). The

6-PhTPA ligand is thus analogous to 6-MeTPA in supporting a

spin-isomeric mixture of iron(III) peroxo complexes.²⁸

As for the iron(II) complex precursors, spin isomerism in the

[(L₁)Fe^III(OO^tBu)]²⁺intermediate was greatly affected by

solvent. Treatment of [L₁Fe(OTf)₂] with excess ^tBuOOH at 233

K in CH₂Cl₂resulted in the immediate formation of a purple

chromophore (λ_max) 566 nm, Figure S2), which was signifi-

cantly blue-shifted compared to that of the intermediate obtained

in CH₃CN. A resonance Raman spectrum of this chromophore

contained features typical of a high-spin iron(III) alkylperoxo

species only (Figure S3), and the absence of a low-spin isomer

was confirmed by EPR spectroscopy, Table 4. Nevertheless,

the purple species in CH₂Cl₂decomposed on a time scale

comparable to the same reaction in CH₃CN to produce a stable

blue chromophore (λ_max) 600 nm, Figure S2). The resonance

Raman spectrum of this product was consistent with assignment

to an iron(III) phenolate monomer, ν(Fe^III-OAr) ) 644 cm^-1

(Figure S4 and Table S3).

interval for the reaction of [(L₁)Fe^II(NCCH₃)₂]²⁺(1.0 mM) and

^tBuOOH (3.0 mM). The same reaction required several hours

to reach completion below 233 K, and a transient intermediate

was detected prior to accumulation of the phenolate chro-

mophore (B in Figure 4, bottom). A strong visible chromophore

(λ_max) 650 nm, ꢀ_max) 1300 M^-1cm^-1) developed over several

minutes, and this was not affected by the meta arene substitution

pattern, which indicated that phenolate complexes had yet to

develop.

Further spectroscopic studies led to the assignment of the

intermediate chromophore as Fe^III-OOR LMCT.^28,52Resonance

Raman spectra recorded from a reaction solution of [(L₁)Fe^II-

t

(NCCH₃)₂]²⁺and BuOOH frozen at the time of maximum

intermediate extinction revealed resonance-enhanced vibrations

characteristic of iron(III)-alkylperoxo complexes (Figure 10

B,C and Table 4). From comparison with spectra of low-spin

[(TPA)Fe^III(OO^tBu)]^{2+ 21d,28}

,

high-spin [(6-Me₃TPA)Fe^III(OO^t-

Bu)]^{2+ 28}

,

and related complexes,⁵³it can be deduced that the

complexity of the observed spectrum reflects a mixture of both

high- and low-spin isomers of [(L₁)Fe^III(OO^tBu)]²⁺. The pres-

(52) (a) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J. Am. Chem.

Soc. 2001, 123, 8271. (b) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon,

E. I. J. Am. Chem. Soc. 2001, 123, 12802.

(53) (a) Me´nage, S.; Wilkinson, E. C.; Que, L., Jr.; Fontecave, M. Angew. Chem.,

Int. Ed. Engl. 1995, 34, 203. (b) Wada, A.; Ogo, S.; Watanabe, Y.; Mukai,

M.; Kitagawa, T.; Jitsukawa, K.; Masuda, H.; Einaga, H. Inorg. Chem.

1999, 38, 3592. (c) Ogihara, T.; Hikichi, S.; Akita, M.; Uchida, T.;

Kitagawa, T.; Moro-oka, Y. Inorg. Chim. Acta 2000, 297, 162.

9

2124 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

Table 4. Spectroscopic Data for (tert-Butylperoxo)iron(III) Complexes

-1

resonance Raman (cm

)

complex

solvent

spin state

UV−vis (nm)

EPR (g-values)

ref

[(TPA)Fe(OO^tBu)]²⁺

MeCN

CH₂Cl₂

MeCN

CH₂Cl₂

MeCN

Et₂O

low

high

low

high

low

600

598

2.19, 2.14, 1.98

2.19, 2.12, 1.97

2.20, 2.12, 1.97

4.3

2.18, 2.11, 1.98

8.4, 4.3

8.6, 4.3

7.6, 4.3

4.3

490

488

469

484

471

469

467

468

463

469

696

682

648

682

650

644

647

637

646

629

606

678

796

790

842

808

842

840

842

844

838

28

21d

28

[(TPA)Fe(HOBz)(OO^tBu)]²⁺

[(6-MeTPA)Fe(OO^tBu)]²⁺

878

[(L₁)Fe(OO^tBu)]²⁺

650

a

878

[(L₁)Fe(OO^tBu)]²⁺

566

590

562

510

613

501

640

873

876

877

885

873

892

a

28

[(L₈)Fe(OO^tBu)]²⁺

[(6-Me₃TPA)Fe(OO^tBu)]²⁺

[(6-Me₃TPA)Fe(O₂CPh)(OO^tBu)]⁺

[(BPPA)Fe(OO^tBu )]²⁺

[Tp^tBu,iPrFe(OO^tBu)]⁺

63

7.58, 5.81, 4.25, 1.82

9.9, 4.3

2.18, 2.12, 1.98

53b

53c

53a

[(bpy)₂Fe(HOBz)(OO^tBu)]²⁺

MeCN

808

^aThis work.

Substitution of R-methyl substituents for hydrogen to afford

[(L₈)Fe^II(NCCH₃)₂]²⁺affected the chemistry greatly. The reac-

t

tion with BuOOH (3.5 equiv) required several hours to reach

completion even at 293 K; nevertheless an Fe^III-OO^tBu

intermediate was observed readily (λ_max) 590 nm, B in Figure

4, top) prior to formation of the phenolate product. Examination

of the peroxo intermediate by resonance Raman and EPR

methods indicated the presence of a single, high-spin isomer

(Table 4 and Figure 10A). The R-methyl substituents clearly

prevent accumulation of the low-spin isomer, as reported for

6-Me₃TPA,²⁵and dramatically slow the conversion of the iron-

(III) alkylperoxo intermediate to the [(L₈O^-)₂Fe^III₂(O)]²⁺prod-

uct.

Figure 11. Detail of resonance Raman spectra of [(L₁O^-)₂Fe^III₂(O)]²⁺

,

illustrating effects of ¹⁸O isotope incorporation on the ν(Fe^III-OPh) band.

t

Samples were prepared with normal BuOOH (A), normal BuOOH and

Mechanistic Probes. Additional mechanistic insights into the

arene hydroxylation were sought from a number of experiments

involving either isotope labels or reactive intermediates. The

former category included transfers of ¹⁸O label from peroxide

to the phenolate, as well as retention and migration of ²H label

on the arene. The latter involved reactive trapping of alkoxy

radicals (RO^•).

20 vol % H₂¹⁸O (B), BuO¹⁸OH and 20 vol % normal H₂O (C), and

t

^tBuO¹⁸OH (D). Spectra were obtained from samples in frozen acetonitrile

using 514.5 nm laser excitation.

of a feature attributed to ν(Fe-¹⁸OAr) (Figure 11B). Conversely,

partial depletion of label was observed from the reaction with

^tBuO¹⁸OH in the presence of excess H₂¹⁶O (Figure 11C). These

results indicate that the hydroxylation mechanism must involve

an oxidizing species that is capable of exchange with exogenous

water.^27,54

The origin of the phenolate oxygen was determined from ¹⁸

O

isotope labeling studies. Thus, reaction of [L₁Fe^II(NCCH₃)₂]²⁺

with tert-butyl hydroperoxide fully enriched with ¹⁸O at the

t

terminal oxygen (i.e., Bu¹⁶O¹⁸OH)²⁵resulted in formation of

Enzymatic arene hydroxylations are often associated with 1,2-

hydrogen isotope shifts, commonly referred to as “NIH”

shifts.^9,12A complex of L₁selectively labeled with deuterium

at one ortho arene carbon (o-d₁-L₁) was synthesized and reacted

completely enriched phenoxide complex, as demonstrated in

parallel by mass spectrometry and resonance Raman spectros-

copy. An electrospray mass spectrum of the product dimer

isolated as the bis(tetraphenylborate) salt revealed a single parent

ion at m/z 1213, its mass and isotope distribution pattern being

fully consistent with its formulation as {[(L₁¹⁸O^-)₂Fe^III₂(µ-O)]-

(BPh₄)}⁺. The lack of significant intensity at m/z 1209 indicated

full retention of the oxygen isotope label. Furthermore a

resonance Raman spectrum of the isotopomeric dimer revealed

a ca. 8 cm^-1downshift of the ν(Fe-¹⁸OAr) Raman mode from

619 cm^-1observed at normal isotope abundance (Figure 11A)

to 611 cm^-1(Figure 11D). These data are uniquely consistent

with derivation of the phenolate oxygen from the terminal

t

with BuOOH at 228 K in CH₃CN. Three phenol isotopomers

1

were obtained on the basis of the H NMR analysis of the

extracted product ligand: 27% of L₁-OH had no deuterium; 23%

retained deuterium in the position adjacent to the hydroxyl group

(H_m′); 50% retained deuterium on the unmodified ortho carbon.

Parallel experiments using mass spectrometry indicated 67(1)%

total deuterium retention. Thus, the deuterium label induces no

net isotope effect in the hydroxylation reaction. However, 46%

of labilized deuterium was found on the adjacent carbon, so

this system does exhibit a significant NIH shift. The intramo-

lecular nature of this shift was confirmed by reaction of a 1:1

mixture of L₁and d₅-L₁complexes, which gave no crossover

to mixed isotopomeric products.

t

oxygen atom of BuOOH.

Reactions were also carried out in the presence of H₂O

isotopomers to determine whether solvent incorporation into the

oxidation product might occur. Indeed, partial ¹⁸O enrichment

of the resulting phenolate ligand was obtained from reaction of

[L₁Fe^II(NCCH₃)₂]²⁺with the normal ^tBuOOH isotopomer in the

presence of H₂¹⁸O (20 vol %), as indicated by the appearance

2-Methyl-1-phenyl-2-propyl hydroperoxide (MPPH) is often

substituted for ^tBuOOH to determine the course of O-O bond

(54) (a) Nam, W.; Valentine, J. S. J. Am. Chem. Soc. 1993, 115, 1772. (b)

Bernadou, J.; Meunier, B. Chem. Commun. 1998, 2167.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2125

A R T I C L E S

Jensen et al.

Scheme 5

Scheme 6

Scheme 7

cleavage in metal-catalyzed reactions, particularly for cases

where homolysis may occur.³⁰Within the context of the present

study, homolysis of iron(III) alkylperoxo intermediates would

generate a transient alkoxy radical (RO^•) that undergoes a fast

(10⁸s^-1) â-scission to yield acetone (Scheme 5).³⁰This reaction

converts the oxidizing alkoxy radical into a reducing benzyl

radical, which can rebound with the oxoiron(IV) coproduct to

produce benzyl alcohol.²²On the other hand, mechanisms of

Fe^III-OOR decomposition not involving alkoxy radial forma-

tion, such as direct hydroxylation by the iron(III) alkylperoxo

intermediate or O-O bond heterolysis, would generate the

corresponding 2-methyl-1-phenyl-2-propyl alcohol as the byprod-

uct.

radical initiation (i.e., Haber-Weiss chemistry, Scheme 6) and

electrophilic substitution by metal-centered oxidants (Scheme

7).^21,22,55Our observations would appear incompatible with free

radical chemistry. The radical most likely to form under the

reaction conditions is tert-butoxyl (^tBuO^•). Aromatic hydrogen

abstraction by ^tBuO^•is strongly disfavored on thermodynamic

grounds, as indicated by the respective bond dissociation

enthalpies of ^tBuO-H (105 kcal/mol)⁵⁶and C₆H₅-H (113 kcal/

mol).⁵⁷Addition to the phenyl ring, while plausible, is nonethe-

less unprecedented and does not lead to phenol by any obvious

path.^27,58Moreover, the ¹⁸O isotope labeling experiments clearly

indicate that the phenolic oxygen is derived quantitatively from

the terminal oxygen of ^tBuOOH. Production of hydroxyl radical

(HO^•) by an unrecognized process (i.e., Fenton chemistry)⁵⁹

would lead to arene hydroxylation, but such chemistry is marked

by low efficiency and regioselectivity.⁶⁰Furthermore, the

efficiencies of radical-initiated arene hydroxylations are im-

proved by the presence of oxidants that trap the cyclohexadienyl

radical adducts, particularly O₂.⁶¹In contrast, the chemistry

described here is indifferent to the presence of O₂and is

characterized by intrinsically high efficiency and absolute

selectivity for ortho substitution, even between specific rotamers

in the case of L₂. Oxygen atom transfer from unstable ^tBuOO^•

also would be consistent with the isotope labeling experiments,

but again this course is without precedent and offers no obvious

means to obtain the observed selectivity.

Reaction of pure MPPH (3.0 mM) with [(L₁)Fe^II(NCCH₃)₂]²⁺

(1.0 mM) in CH₃CN at 228 K under anaerobic conditions

yielded the rebound product benzyl alcohol (1.9 mM), the benzyl

radical coupling product bibenzyl (0.2 mM), and benzaldehyde

(0.4 mM). No 2-methyl-1-phenyl-2-propyl alcohol was found

among the products, even though control experiments verified

that it was readily detectable. The product solution was yellow,

and only a trace of phenolate complex was observed by UV-

1

visible spectroscopy (A_{780 nm}) 0.15 or ca. 10% yield). A H

NMR spectrum of the extracted product ligands showed a

preponderance of unmodified 6-PhTPA (90%). On the other

hand, when the reaction was carried out in the presence of O₂

under ambient pressure, a dark blue solution was obtained with

A_{780 nm}) 1.17, corresponding to 90% spectrophotometric yield

of [(L₁O^-)Fe^III(NCCH₃)]²⁺. Benzaldehyde (2.4 mM) was the

only observed product derived from MPPH. These product

distributions show that benzyl radical forms in the reactions,

and this provides valuable insight into the hydroxylation

mechanism (vide infra).

On the other hand, the exhibited efficiency and ortho

selectivity of arene hydroxylation observed for these complexes

Discussion

On the basis of the results discussed above, a consistent

mechanism of arene hydroxylation in [(L)Fe^II(NCCH₃)₂]²⁺

complexes can be proposed. The first step involves the one-

electron oxidation of the precursor iron(II) complex by ROOH

to an Fe^III-OOR intermediate that can be stabilized and

observed at 228 K. The iron(III) alkylperoxo species was

characterized by a visible LMCT band, resonance enhanced

Raman vibrations, and EPR signals. When the samples are left

to stand, the peroxo charge-transfer band was observed to bleach,

coincident with the appearance of a stable phenolate chro-

mophore for L₁-L₅and L₈. Thus, arene hydroxylation must be

coupled to breakdown of the peroxo intermediate.²⁹

(55) (a) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Banfi, S.; Quici, S.

J. Am. Chem. Soc. 1995, 117, 226. (b) Sawyer, D. T.; Sobkowiak, A.;

Matsushita, T. Acc. Chem. Res. 1996, 29, 409. (c) Walling, C. Acc. Chem.

Res. 1998, 31, 155.

(56) Roth, J. P.; Yoder, J. C.; Won, T.-J.; Mayer, J. M. Science 2001, 294,

2524.

(57) Davico, G. E.; Bierbaum, V. M.; DePuy, C. H.; Ellison, G. B.; Squires, R.

R. J. Am. Chem. Soc. 1995, 117, 2590.

(58) Me´nage, S.; Vincent, J.-M.; Lambeaux, C.; Fontecave, M. J. Mol. Catal.,

A 1996, 113, 61.

(59) (a) Jefcoate, C. R. E.; Lindsay Smith, J. R.; Norman, R. O. C. J. Chem.

Soc. B 1969, 1013. (b) Blair, J. A.; Pearson, A. J. J. Chem. Soc., Perkin

Trans. 2 1975, 245. (c) Walling, C.; Johnson, R. A. J. Am. Chem. Soc.

1975, 97, 363. (d) Hage, J. P.; Llobet, A.; Sawyer, D. T. Bioorg. Med.

Chem. 1995, 3, 1383.

(60) Kurata, T.; Watanabe, Y.; Katoh, M.; Sawaki, Y. J. Am. Chem. Soc. 1988,

110, 7472.

(61) (a) Dorfman, L. M.; Taub, I. A.; Buehler, R. E. J. Chem. Phys. 1962, 36,

3051. (b) Narita, N.; Tezuka, T. J. Am. Chem. Soc. 1982, 104, 7316. (c)

Kunai, A.; Hata, S.; Ito, S.; Sasaki, K. J. Am. Chem. Soc. 1986, 108, 6012.

Two types of reaction mechanisms are usually invoked to

rationalize iron-catalyzed oxygen transfer from peroxides: free

9

2126 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

IV

Biomimetic Aryl Hydroxylation by an Fe dO Oxidant

A R T I C L E S

Scheme 8

from benzyl radical demonstrate that alkoxy radicals must form

during the reaction via O-O bond homolysis of MPPH and

then undergo facile â-scission. In the absence of O₂, the benzyl

radical reacts readily with the homolytic oxoiron(IV) coproduct

to afford benzyl alcohol, effectively quenching the arene

hydroxylation. However, interception of the incipient benzyl

radical by O₂redirects the product toward benzaldehyde and

preserves the oxoiron(IV) intermediate for attack on the pendant

arene.

The ¹⁸O isotope labeling experiments are fully consistent with

the homolytic mechanism. The reaction of ^tBuO¹⁸OH with [(L₁)-

Fe^II(NCCH₃)₂]²⁺clearly demonstrates that the terminal oxygen

atom ends up on the phenolate. Furthermore, the transferred

oxygen atom is susceptible to exchange with exogenous water.⁵⁴

These observations are uniquely consistent with a high-valent

iron-oxo species as the reactive arene hydroxylation intermedi-

ate, presumably oxoiron(IV) generated by homolytic O-O bond

cleavage of the Fe^III-OOR complex (Scheme 8). While the

MPPH experiments demonstrate that the oxidant is a true long-

lived intermediate on the diffusion time scale, its lifetime must

be short relative to that of the iron(III) alkylperoxo species, and

accumulation to the point of direct experimental detection was

not obtained.

Although no specific role is ascribed to the homolysis

t

coproduct BuO^•in the mechanism shown in Scheme 8,

maximum spectrophotometric yields (87% yield of [(L₁O^-)-

are fully consistent with a mechanism involving intramolecular

electrophilic attack by a metal-centered oxidant derived from

an iron(III) alkylperoxo intermediate, Schemes 7 and 8. Indeed,

the geometric disposition of the aryl substituent is ideal for

electrophilic ortho addition of a metal-centered hydroxylating

agent. Crystal structures of [(L₁)Fe^II(NCCH₃)₂]²⁺and [(L₆)Fe^II-

(OTf)₂] reveal that respective rotations about the ipso bonds of

48(2) and 34(3)° place the proximal ortho carbons at distances

of 3.50 and 3.39 Å from the chelated iron(II) ions, which already

approach values of 28(1)° and 2.96 Å found in the product

complex, [(L₁O^-)₂Fe₂(O)]²⁺. Moreover, peroxide-derived ligands

coordinated trans to the amine would be exposed to the arene

face, so that very little structural reorganization would be

required to form an electrophilic adduct between the metal-

based oxidant and the arene.⁶²This geometry also would

disfavor competitive N-dealkylation of the amine by H-atom

abstraction from methylene positions.

Scheme 7 illustrates three options for the metal-based

electrophilic attack, depending on whether O-O bond cleavage

and oxygen atom transfer are concerted or stepwise. The former

case involves direct attack of the iron(III) alkylperoxo inter-

mediate on the arene, resulting in transfer of the terminal oxygen

atom to a proximal ortho carbon. Stepwise pathways involve

prior O-O bond homolysis to form oxoiron(IV) and RO^•or

O-O bond heterolysis to form oxoiron(V), followed by an

electrophilic addition to the arene. The results obtained from

the use of MPPH as a mechanistic probe argue strongly for a

path involving O-O bond homolysis. The concerted and O-O

bond heterolysis pathways are excluded to the extent that

2-methyl-1-phenyl-2-propyl alcohol fails to accumulate as a

reaction product. Furthermore, large yields of products derived

t

Fe^III]²⁺) are obtained with 1.5 equiv of BuOOH (Figure 5).

This stoichiometry corresponds to 0.5 equiv for the initial

oxidation of iron(II) and 1.0 equiv to form the Fe^III-OOR

intermediate and requires significant conservation of the oxidiz-

ing equivalent represented by ^tBuO^•. Clearly this potent oxidant

can react with various species in the reaction mixture, including

the iron(II) precursor complex and the iron(II) phenolate product

initially obtained from the attack of the Fe^IVdO oxidant on the

pendant aryl group.

A further interesting complexity of the observed chemistry

is the accumulation of Fe^III-OOR intermediates as spin isomers,

as demonstrated for L₁. As established for the prototypical TPA

and 6-Me₃TPA complexes,²⁸the spin state of the iron(III)

alkylperoxo intermediate is controlled by R-substitution of the

ancillary 2-pyridylmethyl arms; [(TPA)Fe^IIIOOR]²⁺is low spin,

the 6-Me₃TPA analogue is high spin, and 6-MeTPA supports

both low- and high-spin complexes (Table 4).³⁰Correspond-

ingly, a mixture of low- and high-spin Fe^III-OOR complexes

also was found for L₁in this study and only the high-spin isomer

for L₈. An organonitrile solvent also supports the low-spin state,

and [(L₁)Fe^II(OTf)₂] yields only a high-spin Fe^III-OOR inter-

mediate in CH₂Cl₂.

The spin isomers of iron(III) alkylperoxo intermediates are

distinguished readily by characteristic resonance Raman spectra

(Table 4), which were the subjects of normal coordinate

analysis.⁵²These analyses found that the low-spin isomer

exhibits relatively stronger Fe-O bonding at the expense of a

weaker O-O bond. DFT calculations on simple (NH₃)₄Fe^III-

(OH)(OOR)⁺models suggested a specific Fe-O d-π σ bonding

interaction operant only in the low spin isomer. The conclusion

that such bonding facilitates O-O bond homolysis in the low-

spin iron(III) alkylperoxo isomer to form low-spin oxoiron(IV)

and RO^•is supported by considerable experimental evidence

that [(TPA)Fe^IIIOOR]²⁺utilizes such a decomposition path-

(62) (a) Rathore, R.; Hecht, J.; Kochi, J. K. J. Am. Chem. Soc. 1998, 120, 13278.

(b) Reed, C. A.; Fackler, N. L. P.; Kim, K.-C.; Stasko, D.; Evans, D. R.;

Boyd, P. D. W.; Rickard, C. E. F. J. Am. Chem. Soc. 1999, 121, 6314. (c)

Hubig, S. M.; Kochi, J. K. J. Org. Chem. 2000, 65, 6807. (d) Vasilyev, A.

V.; Lindeman, S. V.; Kochi, J. K. New. J. Chem. 2002, 26, 582.

9

J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2127

A R T I C L E S

Jensen et al.

way.²²The calculations indicated a much higher barrier to O-O

bond homolysis in the high-spin isomer, so that decomposition

instead crosses over to homolysis of the Fe^III-OOR bond,⁵²as

noted experimentally for a series of [(6-Me₃TPA)Fe^III(OOR)(O₂-

CR)]⁺complexes.⁶³

isotopomeric oxidants in H₂¹⁸O yielded product without incor-

porated isotope. The divergent observed behaviors may simply

reflect ligand-dependent effects on the kinetics of solvent water

exchange of the oxoiron(IV) moiety, but involvement of distinct

hydroxylating species cannot be excluded. Among the several

biomimetic iron-catalyzed arene hydroxylations reported thus

far,^17-19the 6-PhTPA system is the only instance for which

there is indirect evidence implicating an oxoiron(IV) oxidant.

Such arguments might lead one to expect that decomposition

of the mixed-spin Fe^III-OOR intermediate supported by L₁in

CH₃CN would exhibit chemistry distinct from the high-spin

intermediates obtained from L₁in CH₂Cl₂and L₈in CH₃CN.

However, ortho hydroxylations of the pendant arene were

obtained in all cases, the main difference being a ca. 10³-fold

faster reactivity for L₁vs L₈(i.e., ∆∆G^q) 4 ( 1 kcal/mol).

The similar reactivities strongly suggest the high-spin isomer

of Fe^III-OOR is a reactive precursor toward arene hydroxyla-

tion. Indeed, the high-spin isomer can be invoked as the sole

productive intermediate by assigning to the low-spin Fe^III-OOR

isomer an indirect decomposition pathway via a spin transition.

The question remains how this simple, albeit counterintuitive,

mechanism might be reconciled with the previous work. The

compelling experimental evidence for preferential O-O bond

homolysis in low-spin Fe^III-OOR species derives only from

studies of unsubstituted [(TPA)Fe^II(NCCH₃)₂]^{2+ 22}and related

computational studies.^52,64The 6-phenyl substituent introduces

steric interactions that are clearly evident in the ground spin

state of the iron(II) complexes. The low-spin forms of the iron-

(III) alkylperoxo intermediates and the putative oxoiron(IV)

oxidant are likely to suffer greater destabilizations due to

increased steric interactions resulting from the smaller radii of

the higher-valent metal ions. For example, both iron centers

adopt low-spin configurations in dimeric [(TPA)₂Fe^IIIFe^IV(O)₂]³⁺

Important arene hydroxylations in biology include those

catalyzed by mononuclear nonheme iron enzymes in the

metabolism of aromatic amino acids.^2,6These enzymes utilize

biopterin as a reducing cosubstrate to activate dioxygen and

catalyze the hydroxylation of namesake substrates (i.e., phenyl-

alanine, tyrosine, and tryptophan) (Scheme 1; vide supra), via

a proposed oxoiron(IV) species.⁹The mechanistic details for

the generation of the putative oxoiron(IV) species in the

6-PhTPA system differ significantly from those proposed for

the pterin-dependent enzymes, namely O-O bond homolysis

of an Fe^III-OOR species (Scheme 8) versus O-O bond

heterolysis of an Fe^II-OOR species (Scheme 5). Thus, the

insights we have gleaned for mechanistic steps leading to

formation of the oxoiron(IV) species in the model system do

not necessarily bear on the enzymatic mechanism. However,

we do find evidence for the participation of an oxoiron(IV)

species in the model arene hydroxylations, and this species

engenders an NIH shift like that observed for the pterin-

dependent hydroxylases, which has served as the hallmark for

such enzymes. Our biomimetic studies thus support the involve-

ment of an oxoiron(IV) oxidant in the mechanisms of the pterin-

dependent hydroxylases.

but are high spin in [(6-MeTPA)₂Fe^IIIFe^IV(O)₂]^{3+ 65}

If the

.

putative [(6-PhTPA)Fe^IV(O)]²⁺accordingly were to favor a high-

spin state (S ) 2), then a significant kinetic barrier would be

introduced for O-O bond homolysis of the low-spin (S ) 1/2)

[(6-PhTPA)Fe^IIIOOR]²⁺intermediate compared to the high-spin

(S ) 5/2) isomer. More experimental work is necessary to

substantiate this speculation.²⁹

Independent of the mechanistic details of the O-O bond

cleavage, it is clear that the resultant oxoiron(IV) species can

effectively hydroxylate pendant aryl groups with ring substit-

uents ranging from CH₃O to NO₂. The present results differ

from those of the best studied biomimetic iron-catalyzed arene

hydroxylation system thus far, which utilized a different

tetradentate ligand, ethylenediamine N,N′-diacetic acid, with

N-(3,4,5-trimethoxybenzyl) substituents.^18,19In that study, ef-

ficient hydroxylation of the electron-rich arene was observed

when the iron(III) complex was treated with either H₂O₂or O₂/

ascorbate. The use of H₂¹⁸O₂or ¹⁸O₂/ascorbate resulted in

complete retention of isotope label in product phenols. In

contrast to our isotope labeling results for the 6-PhTPA system,

the isotope incorporation was fully preserved when the reaction

was performed in normal water, and the reaction with normal

Acknowledgment. This work was supported by a research

grant (GM33162), a postdoctoral fellowship (S.J.L., GM19289).

and a predoctoral traineeship (M.P.M., GM08700) from the

National Institutes of Health. The crystal structure determina-

tions were performed at the X-ray Crystallographic Laboratory

of the Department of Chemistry at the University of Minnesota,

under the direction of Dr. Victor G. Young, Jr. The structure

solutions of [(L₁)Fe^II(NCCH₃)₂](ClO₄)₂and [(L₁O^-)₂Fe^III₂(µ₂-

O)](BPh₄)₂were obtained by Dr. Victor G. Young, Jr., and the

structure solutions of [(L₆)Fe^II(OTf)₂] and [(L₈)Fe^II(OTf)₂]₂were

done by Dr. Neil Brooks and William W. Brennessel, respec-

tively. E.L.Q. is grateful for a Bather Family Undergraduate

Research Fellowship.

Supporting Information Available: Tables of ¹H NMR data

for free ligands L₁-L₈(Table S1) and their iron(II) complexes

(Table S2), a table of resonance Raman bands observed for

1

phenolate chromophores (Table S3), a H NMR spectrum of

[(L₁O^-)₂Fe^III₂(O)]²⁺in CD₃CN (Figure S1), UV-visible spectra

of species observed in reaction of [L₁Fe^II(OTf)₂] in CH₂Cl₂

(Figure S2), resonance Raman spectra of [(L₁)Fe^III(OOtBu)]²⁺

(Figure S3) and [(L₁O^-)Fe^III]²⁺(Figure S4) chromophores in

frozen CH₂Cl₂, and X-ray crystallographic details of [(L₆)Fe^II-

(OTf)₂] and [(L₈)Fe(OTf)₂]₂as a separate file in CIF format.

This material is available free of charge via the Internet at

http://pubs.acs.org.

(63) Kim, J.; Zang, Y.; Costas, M.; Harrison, R. G.; Wilkinson, E. C.; Que, L.,

Jr. J. Biol. Inorg. Chem. 2001, 6, 275.

(64) Bassan, A.; Blomberg, M. R. A.; Siegbahn, P. E. M.; Que, L., Jr. J. Am.

Chem. Soc. 2002, 124, 11056.

(65) (a) Dong, Y.; Que, L., Jr.; Kauffmann, K.; Mu¨nck, E. J. Am. Chem. Soc.

1995, 117, 11377. (b) Dong, Y.; Fujii, H.; Hendrich, M. P.; Leising, R.

A.; Pan, G.; Randall, C. R.; Wilkinson, E. C.; Zang, Y.; Que, L., Jr.; Fox,

B. G.; Kauffmann, K.; Mu¨nck, E. J. Am. Chem. Soc. 1995, 117, 2778.

JA028478L

9

2128 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

Article Doi

Biomimetic aryl hydroxylation derived from alkyl hydroperoxide at a nonheme iron center. Evidence for an FeIV=O oxidant

DOI: 10.1021/ja028478l

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja028478l

Full text of DOI:10.1021/ja028478l

Products guided by the article

R&D Labs maybe for 503275-44-3

Relevant to this article

Hot Product