Mechanism of cis-dihydroxylation and epoxidation of alkenes by highly H2O2 efficient dinuclear manganese catalysts

Inorg. Chem. 2007, 46, 6353−6372

Mechanism of Cis-Dihydroxylation and Epoxidation of Alkenes by

Highly H₂O₂Efficient Dinuclear Manganese Catalysts

Johannes W. de Boer,^†Wesley R. Browne,^†Jelle Brinksma,^†Paul L. Alsters,^‡Ronald Hage,^§and

Ben L. Feringa*^,†

Stratingh Institute for Chemistry, UniVersity of Groningen, Nijenborgh 4, 9747 AG Groningen,

The Netherlands, AdVanced Synthesis, Catalysis and DeVelopment, DSM Pharma Products, PO

Box 18, 6160 MD Geleen, The Netherlands, and UnileVer R&D Vlaardingen, PO Box 114, 3130

AC Vlaardingen, The Netherlands

Received February 26, 2007

+

In the presence of carboxylic acids the complex [Mn^IV

(

₂µ-O)₃(tmtacn)₂]²(1, where tmtacn

)

N,N

′

,N′′-trimethyl-

1,4,7-triazacyclononane) is shown to be highly efficient in catalyzing the oxidation of alkenes to the corresponding

cis-diol and epoxide with H₂O₂as terminal oxidant. The selectivity of the catalytic system with respect to (w.r.t.)

either cis-dihydroxylation or epoxidation of alkenes is shown to be dependent on the carboxylic acid employed.

High turnover numbers (t.o.n. > 2000) can be achieved especially w.r.t. cis-dihydroxylation for which the use of

2,6-dichlorobenzoic acid allows for the highest t.o.n. reported thus far for cis-dihydroxylation of alkenes catalyzed

by a first-row transition metal and high efficiency w.r.t. the terminal oxidant (H₂O₂). The high activity and selectivity

is due to the in situ formation of bis(µ-carboxylato)-bridged dinuclear manganese(III) complexes. Tuning of the

activity of the catalyst by variation in the carboxylate ligands is dependent on both the electron-withdrawing nature

of the ligand and on steric effects. By contrast, the cis-diol/epoxide selectivity is dominated by steric factors. The

role of solvent, catalyst oxidation state, H₂O, and carboxylic acid concentration and the nature of the carboxylic

acid employed on both the activity and the selectivity of the catalysis are explored together with speciation analysis

+

and isotope labeling studies. The results confirm that the complexes of the type [Mn₂(

µ

-O)(

µ

-R-CO₂)₂(tmtacn)₂]²

,

which show remarkable redox and solvent-dependent coordination chemistry, are the resting state of the catalytic

system and that they retain a dinuclear structure throughout the catalytic cycle. The mechanistic understanding

obtained from these studies holds considerable implications for both homogeneous manganese oxidation catalysis

and in understanding related biological systems such as dinuclear catalase and arginase enzymes.

Introduction

notably in the use of Mn^IIsalts⁶and complexes⁷and Fe^II

complexes⁸in the catalytic epoxidation of alkenes. Among

the many oxidative transformations that are of synthetic

interest, first-row transition metal catalyzed cis-dihydroxy-

lation of alkenes⁴remains one of the most challenging. The

reports by De Vos and co-workers on heterogenized Mn-

tmtacn⁹(where tmtacn ) N,N′,N′′-trimethyl-1,4,7-triazacy-

Oxidative transformations¹and especially cis-dihydroxy-

lation and epoxidation of alkenes are key chemical processes

in biology,²synthetic organic chemistry, and the chemical

industry.^3,4,5In recent years, considerable advances have been

made in the development of atom-efficient and environmen-

tally friendly catalytic methods employing H₂O₂,³most

(3) (a) Lane, B. S.; Burgess, K. Chem. ReV. 2003, 103, 2457-2473. (b)

Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic

Compounds; Academic Press: New York, 1981. (c) Brinksma, J.; de

Boer, J. W.; Hage, R.; Feringa, B. L. In Modern Oxidation Methods;

Ba¨ckvall, J.-E., Ed.; Wiley-VCH: Weinheim, Germany, 2004; Chapter

10, pp 295-326. (d) Noyori, R; Aoki, M.; Sato, K. Chem. Commun.

2003, 1977-1986. (e) Katsuki, T. Chem. Soc. ReV. 2004, 33, 437-

444. (f) Punniyamurthy, T.; Velusamy, S.; Iqbal, J. Chem. ReV. 2005,

105, 2329-2364.

* To whom correspondence should be addressed. E-mail: b.l.feringa@

rug.nl.

^†University of Groningen.

^‡DSM Pharma Products.

^§Unilever R&D Vlaardingen.

(1) Backvall, J.-E., Ed. Modern Oxidation Methods; Wiley-VCH: Wein-

heim, Germany, 2004.

(2) Wu, A. J.; Penner-Hahn, J. E.; Pecoraro, V. L. Chem. ReV. 2004, 104,

(4) Kolb, H. C.; Van Nieuwenhze, M. S.; Sharpless, K. B. Chem. ReV.

1994, 94, 2483-2547.

903-938.

10.1021/ic7003613 CCC: $37.00

Published on Web 07/03/2007

Inorganic Chemistry, Vol. 46, No. 16, 2007 6353

de Boer et al.

Scheme 1. Cis-Dihydroxylation and Epoxidation of Alkenes by 1

clononane) and Que and co-workers with Fe^IIpyridyl-amine-

based complexes⁸and our own recent report on the use of

the oxidation catalyst [Mn^IV₂O₃(tmtacn)₂]²⁺(1, Scheme 1),¹⁰

in the presence of carboxylic acids,¹¹have demonstrated the

potential of first-row transition metals toward cis-dihydroxy-

lation of alkenes.

range of oxidative transformations including textile stain

bleaching,¹⁸benzylalcoholoxidation,^17cC-Hbondactivation,^17f

sulfoxidation,¹⁹and cis-dihydroxylation and epoxidation of

alkenes.^11,16,17

In our recent communication,¹¹we reported that 1 can

engage in the atom-efficient cis-dihydroxylation of alkenes

with high turnover numbers when combined with electron

deficient carboxylic acids (Figure 1). We demonstrated that

the use of carboxylic acids at cocatalytic levels is effective

in suppressing the inherent catalase activity of 1¹⁰and allows

for the tuning of the catalyst’s selectivity toward either cis-

dihydroxylation or epoxidation. Preliminary kinetic and

spectroscopic measurements¹¹indicated that control over the

outcome of the reaction toward cis-dihydroxylation or

epoxidation presumably arises from the in situ formation of

carboxylato-bridged dinuclear complexes, e.g., complex 2a

{[Mn^III₂(µ-O)(µ-CCl₃CO₂)₂(tmtacn)₂]²⁺}, during catalysis

(Figure 1).

In this paper a full account of the structural and mecha-

nistic features as well as the parameters that govern the

activity and selectivity of the catalysis of this exceptionally

H₂O₂efficient catalytic system is provided. The role of 2a

in the catalysis and the importance of the formation of the

µ-carboxylato-bridged dinuclear manganese(II) complex 2c

(in which the µ-oxido bridge is replaced by two labile OH/

H₂O ligands) and its Mn^III₂analog (2d) is addressed (Figure

1). We demonstrate that the dominant species present under

catalytic conditions are dinuclear bis(µ-carboxylato)-bridged

complexes (e.g., 2a,d) and that the reaction of these

complexes with H₂O and H₂O₂is rate limiting. Furthermore,

we show that in addition to consideration of the molecular

catalyst, bulk solvent conditions must be taken into account

in understanding the behavior of the reaction system overall

in terms of reactivity and selectivity.

Manganese complexes based on the ligand tmtacn, such

as 1 (Scheme 1), were developed in the late 1980s and -90s

as functional models for bioinorganic manganese systems,^10,12

in particular, dinuclear manganese based catalase^2,13,14en-

zymes and the water-splitting component of photosystem II

(PSII).¹⁵The catalytic properties of these complexes toward

oxidative transformations with H₂O₂in both aqueous¹⁶and

nonaqueous¹⁷media have, however, made the tmtacn family

of complexes the focus of considerable interest for a whole

(5) (a) Nelson, D. W.; Gypser, A.; Ho, P. T.; Kolb, H. C.; Kondo, T.;

Kwong, H.-L.; McGrath, D. V.; Rubin, A. E.; Norrby, P.-O.; Gable,

K. P.; Sharpless, K. B. J. Am. Chem. Soc. 1997, 119, 1840-1858. (b)

Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 5974-

5976. (c) Sato, K.; Aoki, M.; Ogawa, M.; Hashimoto, T.; Noyori, R.

J. Org. Chem. 1996, 61, 8310-8311. (d) Herrmann, W. A.; Fischer,

R. W.; Marz, D. W. Angew. Chem., Int. Ed. Engl. 1991, 30, 1638-

1641.

(6) Lane, B. S.; Vogt, M.; DeRose, V. J.; Burgess, K. J. Am Chem. Soc.

2002, 124, 11946-11954.

(7) (a) McGarrigle, E. M.; Gilheany, D. G. Chem. ReV. 2005, 105, 1563-

1602. (b) Katsuki, T. Coord. Chem. ReV. 1995, 140, 189-214.

(8) (a) Fujita, M.; Costas, M.; Que, L., Jr. J. Am. Chem. Soc. 2003, 125,

9912-9913. (b) Chen, K.; Costas, M.; Kim, J.; Tipton, A. K.; Que,

L., Jr. J. Am. Chem. Soc. 2002, 124, 3026-3035. (c) Ryu, J. Y.; Kim,

J.; Costas, M.; Chen, K.; Nam, W.; Que, L., Jr. Chem. Commun. 2002,

1288-1289. (d) Mas-Balleste, R.; Costas, M.; van den Berg, T.; Que,

L., Jr. Chem.sEur. J. 2006, 12, 7489-7500. (e) Oldenburg, P. D.;

Que, L., Jr. Catal. Today 2006, 117, 15-21. (f) Bukowski, M. R.;

Comba, P.; Lienke, A.; Limberg, C.; Lopez, de Laorden, C.; Mas-

Balleste, R.; Merz, M.; Que, L., Jr. Angew. Chem. 2006, 45, 3446-

3449. (g) Oldenburg, P. D.; Ke, C.-Y.; Tipton, A. A.; Shteinman, A.

A.; Que., L., Jr. Angew. Chem., Int. Ed. 2006, 45, 7975-7978. (h)

Oldenburg, P. D.; Shteinman, A. A.; Que, L., Jr. J. Am. Chem. Soc.

2005, 127, 15672-15673.

(9) De Vos, D. E.; De Wildeman, S.; Sels, B. F.; Grobet, P. J.; Jacobs, P.

A. Angew. Chem., Int. Ed. 1999, 38, 980-983.

Experimental Section

All reagents were of commercial grade (Aldrich, Acros, Fluka)

and were used as received unless stated otherwise. Dihydrogen

(10) (a) Wieghardt, K.; Bossek, U.; Nuber, B.; Weiss, J.; Bonvoisin, J.;

Corbella, M.; Vitols, S. E.; Girerd, J. J. J. Am. Chem. Soc. 1988, 110,

7398-7411. (b) Darovsky, A.; Kezerashvili, V.; Coppens, P.; Wey-

hermuller, T.; Hummel, H.; Wieghardt, K. Inorg. Chem. 1996, 35,

6916-6917. (c) Hage, R.; Krijnen, B.; Warnaar, J. B.; Hartl, F.;

Stufkens, D. J.; Snoeck, T. L. Inorg. Chem. 1995, 34, 4973-4978.

(d) Bossek, U.; Weyhermuller, T.; Wieghardt, K.; Nuber, B.; Weiss,

J. J. Am. Chem. Soc. 1990, 112, 6387-6388.

(15) (a) Sauer, K. Acc. Chem. Res. 1980, 13, 249-256. (b) Mukhopadhyay,

S.; Mandal, S. K.; Bhaduri, S.; Armstrong, W. H. Chem. ReV. 2004,

104, 3981-4026. (c) Yachandra, V. K.; Sauer, K.; Klein, M. P. Chem.

ReV. 1996, 96, 2927-2950. (d) Yagi, M.; Kaneko, M. Chem. ReV.

2001, 101, 21-36. (e) Sun, L.; Hammarstro¨m, L.; Åkermark, B.;

Styring, S. Chem. Soc. ReV. 2001, 1, 36-49. (f) Yano, J.; Sauer, K.;

Girerd, J.-J.; Yachandra, V. K. J. Am. Chem. Soc. 2004, 126, 7486-

7495. (g) Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji,

I.; Sauer, K.; Klein, M. P. Science 1993, 260, 675-679. (h) Baranov,

S. V.; Tyryshkin, A. M.; Katz, D.; Dismukes, G. C.; Ananyev, G.

M.; Klimov, V. V. Biochemistry 2004, 43, 2070-2079. (i) Ruttinger,

W.; Dismukes, G. C. Chem. ReV. 1997, 97, 1-24. (j) McEvoy, J. P.;

Brudvig, G. W. Chem. ReV. 2006, 106, 4455-4483.

(11) de Boer, J. W.; Brinksma, J.; Browne, W. R.; Meetsma, A.; Alsters,

P. L.; Hage, R.; Feringa, B. L. J. Am. Chem. Soc., 2005, 127, 7990-

7991.

(12) de Boer, J. W.; Browne, W. R.; Feringa, B. L.; Hage, R. C. R. Chimie

2007, 10, 341-354.

(13) (a) Pecoraro, V. L.; Baldwin, M. J.; Gelasco, A. Chem. ReV. 1994,

94, 807-826. (b) Jackson, T. A.; Brunold, T. C. Acc. Chem. Res.

2004, 37, 461-470. (c) Boelrijk, A. E. M.; Khangulov, S. V.;

Dismukes, G. C. Inorg. Chem. 2000, 39, 3009-3019. (d) Boelrijk,

A. E. M.; Dismukes, G. C. Inorg. Chem. 2000, 39, 3020-3028.

(14) Kanyo, Z. F.; Scolnick, L. R.; Ash, D. E.; Christianson, D. W. Nature

1996, 383, 554-557.

(16) Hage, R.; Iburg, J. E.; Kerschner, J.; Koek, J. H.; Lempers, E. L. M.;

Martens, R. J.; Racherla, U. S.; Russell, S. W.; Swarthoff, T.; van

Vliet, M. R. P.; Warnaar, J. B.; van der Wolf, L.; Krijnen, B. Nature

1994, 369, 637-639.

6354 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Figure 1. Complexes described in the text. Complexes 2a,¹¹2b,¹¹and 3^10ahave been characterized structurally. For complexes 2c,d and 5-15, see the

Supporting Information.

peroxide used: 50% v/w (Acros) or 30% v/v (Merck, medicinal

grade) solution in water. D₂O₂(Icon Isotopes): 30% solution in

D₂O, 99 atom % D. D₂O (Aldrich): 99.9 at. % D. H₂¹⁸O₂(Icon

Isotopes): 2% solution in H₂¹⁶O, 90 at. % ¹⁸O. H₂¹⁸O (Icon

Isotopes): 97 at. % ¹⁸O. cis-2,3-Epoxyheptane, trans-2,3-epoxy-

heptane, threo-2,3-heptanediol, and erythro-2,3-heptanediol were

Definition of t.o.n., RC, and Mass Balance. Turnover number

(t.o.n.): mol of product/mol of catalyst. RC (retention of config-

uration) ) 100% × (A - B)/(A + B), where A ) yield of product

with retention of configuration and B ) yield of epimer.²¹Mass

balance (%) ) unreacted alkene (%) + [cis-diol and epoxide

products (%)]. Deviation from a mass balance of 100% indicates

loss through further oxidation of the cis-diol formed initially and/

or the occurrence of competing oxidation pathways other than cis-

dihydroxylation and epoxidation.¹¹

1

prepared by literature procedures²⁰and characterized by H and

¹³C NMR spectroscopy and HRMS. The synthesis and characteriza-

tion of complexes 1,^10a2a,¹¹2b,¹¹and 3^10awere reported previously.

Full synthetic procedures and analytical, spectroscopic, and elec-

trochemical data for complexes 2c and 4-15 are provided as

Supporting Information (SI).

Physical Measurements. ¹H NMR spectra (400.0 MHz) and ¹⁹

F

NMR (121.5 MHz) spectra were recorded on a Varian Mercury

Plus. Chemical shifts are denoted relative to the solvent residual

-

peak (¹H NMR spectra CD₃CN: 1.94 ppm) or PF₆(¹⁹F NMR

Caution! Perchlorate salts of metal complexes incorporating

organic ligands are potentially explosive. These compounds should

be prepared in small quantities and handled with suitable protective

safe guards.

spectra PF₆^-: 74.8, 243 Hz). Elemental analyses were performed

with a Foss-Heraeus CHN-O-Rapid or a EuroVector Euro EA

elemental analyzer. Electrochemical measurements were carried out

on a model 630B Electrochemical Workstation (CH Instruments).

Analyte concentrations were typically 0.5-1.0 mM in anhydrous

acetonitrile containing 0.1 M tetrabutylammonium hexafluorophos-

phate [(TBA)PF₆]. Unless stated otherwise, a Teflon-shrouded

glassy carbon working electrode (CH Instruments), a Pt wire

auxiliary electrode, and an SCE reference electrode were employed

(calibrated externally using 0.1 mM solutions of ferrocene in 0.1

M (TBA)PF₆/CH₃CN). Cyclic voltammograms were obtained at

sweep rates between 1 mV s^-1and 10 V s^-1. For reversible

processes, the half-wave potential values are reported. Redox

potentials are reported (10 mV. EPR spectra (X-band, 9.46 GHz)

were recorded on a Bruker ECS106 instrument in liquid nitrogen

(77 K). Samples for measurement were transferred from the reaction

solution (250 µL) to an EPR tube, which was frozen to 77 K

immediately.

Caution! The removal of water from aqueous H₂O₂solutions

by mixing with organic solvents followed by drying over magne-

sium sulfate is hazardous and should be performed only in small

quantities and handled with suitable protective safe guards.

(17) (a) Berkessel, A.; Sklorz, C. A. Tetrahedron Lett. 1999, 40, 7965-

7968. (b) De Vos, D. E.; Bein, T. Chem. Commun. 1996, 917-918.

(c) Zondervan, C.; Hage, R.; Feringa, B. L. Chem. Commun. 1997,

419-420. (d) De Vos, D. E.; Sels, B. F.; Reynaers, M.; Subba Rao,

Y. V.; Jacobs, P. A. Tetrahedron Lett. 1998, 39, 3221-3224. (e)

Woitiski, C. B.; Kozlov, Y. N.; Mandelli, D.; Nizova, G. V.;

Schuchardt, U.; Shul’pin, G. B. J. Mol. Catal., A 2004, 222, 103-

119. (f) Bolm, C.; Meyer, N.; Raabe, G.; Weyhermuller, T.; Bothe,

E. Chem. Commun. 2000, 2435-2436. (g) Chin Quee-Smith, V.;

DelPizzo, L.; Jureller, S. H.; Kerschner, J. L.; Hage. R. Inorg. Chem.

1996, 35, 6461-6465. (h) De Vos, D. E.; Bein, T. Chem. Commun.

1996, 917-918. (i) Gilbert, B. C.; Smith, J. R. L.; Payeras, A. M.;

Oakes, J.; Pons i Prats, R. J. Mol. Catal., A: Chem. 2004, 219, 265-

272. (j) Ryu, J. Y.; Kim, S. O.; Nam, W.; Heo, S.; Kim, J. Bull. Korean

Chem. Soc. 2003, 24, 1835-1837. (k) Lindsay Smith, J. R.; Gilbert,

B. C.; Mairata i Payeras, A.; Murray, J.; Lowdon, T. R.; Oakes, J.;

Pons i Prats, R.; Walton, P. H. J. Mol. Catal., A: Chem. 2006, 251,

114-122 and references cited herein. (l) Brinksma, J.; Schmieder,

L.; van Vliet, G.; Boaron, R.; Hage, R.; De Vos, D. E.; Alsters, P. L.;

Feringa, B. L. Tetrahedron Lett. 2002, 43, 2619-2622.

UV-vis spectra were recorded on a Hewlett-Packard 8453

spectrophotometer using either 2 or 10 mm path length quartz

cuvettes. Electrospray ionization mass spectra were recorded on a

Triple Quadrupole LC/MS/MS mass spectrometer (API 3000,

Perkin-Elmer Sciex Instruments). A sample (2 µL) was taken from

the reaction mixture at the indicated times (vide infra) and was

diluted in CH₃CN (1 mL) before injection in the mass spectrometer

(via syringe pump). Mass spectra were measured in positive mode

(18) Hage, R.; Lienke, A. Angew. Chem., Int. Ed. 2006, 45, 202-222.

(19) Barker, J. E.; Ren, T. Tetrahedron Lett. 2004, 45, 4681-4683.

(20) (a) Kroutil, W.; Mischitz, M.; Faber, K. J. Chem. Soc., Perkin Trans.

1 1997, 3629-3636. (b) Chiappe, C.; Cordoni, A.; Lo Moro, G.;

Palese, C. D. Tetrahedron: Asymmetry 1998, 9, 341-350.

(21) Chen, K.; Que, L., Jr. J. Am. Chem. Soc. 2001, 123, 6327-6337.

Inorganic Chemistry, Vol. 46, No. 16, 2007 6355

de Boer et al.

and in the range m/z 100-1500. Ion-spray voltage: 5200 V.

Orifice: 15 V. Ring: 150 V. Q₀: -10 V.

column (30 m × 0.32 mm × 0.25 µm). MS: JMS-600H using

chemical ionization (reaction gas NH₃). The samples were analyzed

by GC (HP-1 column) as described above also to determine t.o.n..

Ligand exchange of the µ-oxido bridge of [Mn^III₂(µ-¹⁶O)(µ-CCl₃-

CO₂)₂(tmtacn)]²⁺, [Mn^III₂(µ-¹⁶O)(µ-CH₃CO₂)₂(tmtacn)]²⁺, and [Mn^III

-

2

(µ-¹⁶O)(µ-2,6-dichlorobenzoate)₂(tmtacn)]²⁺was monitored by ESI-

MS. Measurements were performed at concentrations used in

catalytic reactions (i.e., 1 mM in complex), and samples were

injected directly into the ESI-MS using a syringe pump and PEEKsil

tubing (normal probe and without dilution). API-365 settings:

orifice, +15 V; ring, +150 V; Q₀, -5 V. For kinetic measurements

only, a small portion of the spectrum was recorded to minimize

measuring time between each subsequent data point while monitor-

ing the [Mn^III₂(µ-O)(µ-RCO₂)₂(tmtacn)]²⁺ion. The presence of

cyclooctene (1 M) and of carboxylic acid (1.0 mol %) did not result

in significant interference of the mass spectra of 1 or 2a,²²and

neither cyclooctene nor the cis-diol and epoxide products give rise

to significant signals.²³Importantly, even with 25 mol % CCl₃-

CO₂H, no free tmtacn ligand is observed, indicating that 1 is stable

under these conditions prior to addition of H₂O₂.

Results and Discussion

To understand the roles played by solvent, water, and

carboxylic acid in determining the activity and selectivity

of the catalytic system (vide infra), it is pertinent to first

understand the redox and spectroscopic properties of the

complexes that are, potentially, involved in the catalysis. In

particular, a detailed study of the redox and solvent driven

interconversion of complexes 2a-d, their formation from

1, and their behavior in the presence of H₂O₂is central to

understanding the catalytic system.

Synthesis and Characterization. The synthesis and

structural characterization of complexes 1,¹⁰2a,¹¹2b,¹¹and

3¹⁰were reported previously. Complexes 4-7 and 9-15

(Figure 1) were prepared in good yield by ascorbic acid

reduction of 1 in an aqueous solution containing the

respective carboxylic acid (see SI for procedures and

analytical and spectroscopic data). Overall, the Mn^III₂-bis-

(µ-carboxylato) complexes exhibit solution chemistry similar

to that of 3;¹⁰however, the redox and electronic properties

are affected significantly by the carboxylate ligand employed

(see Table S1), with more electron-deficient carboxylates

showing anodic shifts in the potential of all redox processes

observed. The Mn^II₂complex 2c was prepared by reduction

of 1 with 2 equiv of H₂NNH₂in acetonitrile in the presence

of 2 equiv of CCl₃CO₂H and isolated as a white powder (see

SI for details).

The solid-state FTIR spectra of 2a-c are shown in Figure

S1. For 2a the carboxylato bridge absorption is observed at

1659 cm^-1. The carboxylato absorptions of 2b,c are very

similar (∼1695 cm^-1). The absorptions assigned to the

Cl₃COO^-moieties of 2a,b are blue-shifted significantly (80-

90 cm^-1) from the corresponding CH₃COO^--bridged com-

plexes (1568 and 1615 cm^-1,^10arespectively).

In acetonitrile solution the carboxylato absorption band

of 2a at 1659 cm^-1is retained and its insensitivity to either

CCl₃CO₂H or cyclooctene suggests that the bridging mode

of the complex is the same as that determined for the solid

state by X-ray crystallography¹¹(Figure S2). Similarly both

2b,c show a strong absorption at 1695 cm^-1in CH₃CN with

1 M cyclooctene present, identical with that observed in their

solid-state spectra.

Procedures Employed for Catalysis Studies. General Proce-

dure A. The alkene (10 mmol), 1,2-dichlorobenzene (internal

standard, 735 mg, 5.0 mmol), 1 (8.1 mg, 10 µmol), and cocatalyst

(typically 0.10 mmol) in acetonitrile (10 mL) were cooled to 0 °C.

H₂O₂(0.74 mL, 13 mmol) was added via syringe pump over 6 h

(0.12 mL/h). The reaction mixture was stirred at 0 °C for 1 h after

the addition of H₂O₂was completed, prior to sampling by GC.

General Procedure B (Catalyst Pretreatment with H₂O₂).

H₂O₂(30 µL, 0.53 mmol) was added to a mixture of 1,2-

dichlorobenzene (735 mg, 5.0 mmol), 1 (8.1 mg, 10 µmol), and

cocatalyst (e.g., trichloroacetic acid, 0.10 mmol) in acetonitrile (7

mL) at room temperature. The mixture was stirred for 20 min, after

which the alkene (10 mmol) was added together with acetonitrile

(3 mL) and the mixture was cooled to 0 °C. H₂O₂(0.71 mL, 12.5

mmol) was added via syringe pump (0.12 mL/h). The reaction

mixture was stirred at 0 °C for 1 h after the addition of H₂O₂was

completed, prior to sampling by GC.

GC analyses were performed on an Agilent 6890 gas chromato-

graph equipped with a HP-1 dimethyl polysiloxane column (30 m

× 0.25 mm × 0.25 µm). Peak identification and calibration were

performed using independent samples (either purchased from a

commercial supplier or synthesized independently¹¹). Conversion

and turnover numbers were determined in duplo by employing 1,2-

dichlorobenzene as internal standard. All values are (10%.

¹⁸O-Labeling. Samples at t ) 60 min were analyzed by GC-

MS using chemical ionization. GC: HP 6890 equipped with a HP-5

(22) Mass spectrometry has proven to be a powerful tool in the identification

of species present in solution, not least for manganese systems. (a)

Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M.

Mass Spectrom. ReV. 1990, 9, 37-70. (b) Feichtinger, D.; Plattner,

D. Angew. Chem., Int. Ed. Engl. 1997, 36, 1718-1719. (c) Feichtinger,

D.; Plattner, D. J. Chem. Soc., Perkin Trans. 2000, 2, 1023-1028.

(d) Gilbert, B. C.; Smith, J. R. L.; Mairata i Payeras, A.; Oakes, J.

Org. Biomol. Chem. 2004, 2, 1176-1180. (e) Bortolini, O.; Conte,

V. Mass Spectrom. ReV. 2006, 25, 724-740. However, caution should

be exercised in its use as a mechanistic probe due to the possibility of

experimental artifacts occurring such as reduction within the mass

spectrometer itself and the presence of nondetectable neutral species.

Indeed, the use of stabilizers in commercial grade chemicals, such as

cyclooctene, highlights this point. One source of cyclooctene contained

such a stabilizer, which provided signals in the MS experiments

assignable to possible catalytic intermediates. The use of cyclooctene

from a second source or triple distillation of the cyclooctene to remove

the stabilizer while having no effect on other spectroscopic properties

or catalysis did result in the elimination of these spurious signals.

(23) For 2a at higher CCl₃CO₂H (10 mol %) concentrations, an additional

signal at m/z 951 {assigned to [Mn^III₂(O)(CCl₃CO₂)₂tmtacn₂](CCl₃-

CO₂)⁺} is observed.

In the spectrum of 2a an additional weak band is observed

at 1735 cm^-1, which in the presence of CCl₃CO₂H shifts to

1720 cm^-1(Figure S3). Addition of CCl₃CO₂H (10 equiv)

to a solution of 2b in CH₃CN results in the disappearance

of the absorption band at 1695 cm^-1and the appearance of

a strong absorption band at 1720 cm^-1. Addition of D₂O

results in the partial recovery of the absorption band at 1695

cm^-1(Figure S3). The carboxylato vibrations of the tris-

(carboxylato)-bridged complex [Mn^II(µ-OAc)₃(tmtacn)₂]⁺,

2

reported by Wieghardt et al.^10aat 1636 cm^-1, are at higher

energy than for the corresponding bis(carboxylato) [Mn^II-

2

(µ-OH)(µ-OAc)₂(tmtacn)₂]⁺(1615 cm^-1). This suggests that

the absorption at 1720 cm^-1is due to the complex [Mn^II(µ-

2

6356 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Scheme 2. Summary of Redox Chemistry of 2a-d in 0.1 M (TBA)PF₆/CH₃CN with 1 mol % CCl₃CO₂H

Table 1. Electronic and Redox Data for 2a-d Complexes in CH₃CN/

0.1 M (TBA)PF₆

CCl₃CO₂)₃(tmtacn)₂]⁺(Scheme 3), formation of which is

suppressed by addition of water.

compd

core

λ

_max/nm

E_1/2in V vs SCE (E_{p,a -}E_p,c/mV)

Overall 2c can be assigned as a bis(carboxylato)-bridged

complex, which in the presence of excess CCl₃CO₂H is in

2a

2b

2c

2d

Mn^III₂(µ-O)

Mn^II₂(µ-OH)

Mn^II₂(µ-O₂H₃)

Mn^III₂(OH)₂

390, 494, 540

b

0.25 (80),^a1.40 (100)

0.54 (110)

0.99 (85, qr)

equilibrium with a tris(carboxylato) complex, i.e., [Mn^II(µ-

2

CCl₃CO₂)₃(tmtacn)₂]⁺. The nature of the third bridging unit

of 2c, however, is less clear. In the solid state, 2b shows a

very sharp absorption at 3620 cm^-1, assigned to the µ-O-H

unit, which is present in the X-ray crystal structure. The O-H

absorption band is equivalent to that of, for example, a

dinuclear Zn^IIcomplex reported by Meyer and Rutsch (ν )

3618 cm^-1).²⁴The spectra of 2b,c are strikingly similar in

both the solid state and in acetonitrile solution; however,

the sharp O-H absorption of 2b is replaced by a broad set

of three absorption bands in 2c at 3593, 3433, and 3251

cm^-1. Although crystals of 2c suitable for X-ray analysis

have, thus far, not been obtained, we propose that complex

2c contains a {Mn^II-(µ-O₂H₃)-Mn^II} structural motif (Figure

1) on the basis of ESI-MS, elemental analysis, and EPR and

300, 468

0.67 (E_p,c, irr)

^aIn CH₃CN/0.1 M (TBA)PF₆with 10 mM CCl₃CO₂H, an irreversible 2

e^-reduction is observed at 0.29 (E_p,c). ^bComplex does not absorb in the

visible or near-UV region.

of the redox processes, overall “decomposition” of 2a does

not occur (Scheme 2 and Table 1).²⁵The proton-coupled two-

electron reduction of 2a to form 2b is followed, in the

presence of a carboxylic acid, by opening of the µ-OH bridge

to form the corresponding hydroxy-aqua complex, 2c. The

latter complex undergoes a two-electron oxidation to form

[Mn^III₂(µ-O₂H₃)(µ-CCl₃CO₂)₂(tmtacn)₂]³⁺(H2d³⁺), which,

depending upon the conditions employed, (i) shows a (quasi)-

reversible two-electron reduction to form 2c (0.94 V), (ii)

undergoes deprotonation to form [Mn^III₂(OH)₂(µ-CCl₃CO₂)₂-

(tmtacn)₂]²⁺(2d), or (iii) undergoes a further chemical

change: elimination of H₂O to re-form 2a.

IR spectroscopy. The related Mn^IIIcomplex 2d, although

2

stable in acetonitrile solution, was not isolated due to the

equilibrium of this complex with complex 2a. Hence, 2d

was characterized spectroscopically and electrochemically

only (vide infra).

(25) In the presence of electrolyte, 1 and 2a show a moderate reduction in

catalytic activity; however, overall the selectivity and reactivity is not

affected significantly.

(26) The series of electron-transfer reactions which accompany the reduction

of 1 to 2c are expected to be outer sphere in nature on the basis of the

equivalent reactions, which are observed electrochemically. If the

reduction of 1 by H₂O₂involves an outer-sphere double electron

Electrochemical Properties of 2a-d. Both the Mn^III

2

complex 2a and the Mn^II₂complex 2b have been character-

ized by X-ray crystallography.¹¹The manganese centers are

bridged by two carboxylato ligands and a µ-oxido or

µ-hydroxido bridge, respectively (Figure 1). In acetonitrile

solution both complexes are stable and show reversible one-

electron redox processes (Figure 3a, Figure 4a, and Table

S1; vide infra) in the absence of CCl₃CO₂H.

transfer, then it will result in the formation of O₂[Mn^IV+ 2e^-

f

2

Mn^III₂; H₂O₂f O₂+ 2H⁺+2e^-]. However, it is also possible that

inner-sphere single-electron-transfer processes are involved where

H₂O₂coordinates to 1. This would complicate the description of the

reaction: (i) Mn^IV₂+ 1e^-f Mn^IIIMn^IV(often observed by EPR); ii)

2Mn^IIIMn^IVf Mn^IV+ Mn^III(disproportionation) or Mn^IIIMn^IV

+

In the presence of 10 equiv of CCl₃CO₂H, a series of

chemically irreversible redox processes are observed (Figure

4a). Although significant structural changes accompany some

2

e^-f Mn^III₂and H₂O₂f HO₂+H⁺+ 1e^-(and then 2HO₂f O₂+

H₂O₂). The matter is complicated further by the fact that the conversion

is autocatalytic, and hence, although H₂O₂might be involved directly

in the reaction initially, once the process of conversion of 1 has begun

the involvement of species such as 2a-d must be considered.

(24) Meyer, F.; Rutsch, P. Chem. Commun. 1998, 1037-1038.

Inorganic Chemistry, Vol. 46, No. 16, 2007 6357

de Boer et al.

Figure 2. Cyclic voltammetry of 2c (1 mM) in CH₃CN/0.1 M (TBA)PF₆

in the (a) absence and (b) presence of 10 mM CCl₃CO₂H and (c) 2a with

10 mM CCl₃CO₂H. The current axes of (a) and (c) are offset by -30 and

+30 µA, respectively, for clarity. Initial scan direction from the open circuit

Figure 3. Cyclic voltammetry of 2b in the (a) absence and (b) presence

of 10 mM CCl₃CO₂H and (c) 2a in the presence of 10 mM CCl₃CO₂H.

The current axes for (a) and (b) are offset by -10 and +20 µA for clarity,

respectively. All complexes are 1 mM in CH₃CN/0.1 M (TBA)PF₆. Initial

scan direction from the open circuit potential is cathodic in each case. Scan

potential is cathodic in each case. Scan rate: 0.1 V s^-1

.

rate: 0.1 V s^-1

.

In the presence of either H₂O or CCl₃CO₂H, the redox

chemistry and, in the case of 2b, the µ-OH bridge are affected

considerably. For 2b, addition of CCl₃CO₂H or H₂O leads

to the formation of 2c, in which the µ-OH bridge is replaced

by a µ-O₂H₃bridge (Scheme 2; vide infra).

subsequent one-electron reduction. The irreversibility of the

reduction at 0.25 V and the presence of an anodic redox

wave at 1.03 V subsequent to reduction of 2a confirm that

a major structural change occurs, i.e., the formation of 2c

(Scheme 2; vide supra).

In the absence of CCl₃CO₂H acid, no redox processes are

observed between 0.40 and 1.20 V for 2a. In the presence

of increasing amounts of CCl₃CO₂H, however, the redox

wave at E_p,c) 0.67 V increases in intensity indicating that

the concentration of 2d is increasing (Figure S4). This

indicates that, at higher acid concentrations, the equilibrium

between 2a and 2d changes in favor of the latter species. In

the presence of 10 equiv of CCl₃CO₂H, increasing the H₂O

concentration has a similar effect on the cyclic voltammetry

of 2a, confirming that the equilibrium is not dominated by

[H⁺] or [CCl₃CO₂H] but rather the presence of water, which

facilitates opening of the µ-oxido bridge.

For 2c {Mn^II(µ-O₂H₃)} oxidation is observed at E_p,a

)

2

1.03 V (vs SCE, Figure 2a). This process is electrochemically

reversible, as can be observed when H₂O or cyclooctene are

present with CCl₃CO₂H (vide infra, Figure 7). However, the

return waves observed correspond to further chemical

reaction of the {Mn^III₂(µ-O₂H₃)} complex formed including

formation of 2a. When 10 equiv of CCl₃CO₂H is added

(Figure 2b), the {Mn^II(µ-O₂H₃)} to {Mn^III₂(µ-O₂H₃)} two-

2

electron oxidation wave at E_p,a) 1.03 V is unaffected;

however, only one return wave, a two-electron reduction

(Mn^IIIto Mn^II, at E_p,c) 0.67 V), is now observed.

2

In contrast, for 2b an electrochemically reversible, one-

electron oxidation to [Mn^IIMn^III(µ-OH)(µ-CCl₃CO₂)₂-

(tmtacn)₂]²⁺is observed at 0.53 V (Figure 3a). Addition of

10 equiv of CCl₃CO₂H to 2b results in a change of its cyclic

voltammetry (Figure 3b) with behavior identical with that

of 2c in the presence of CCl₃CO₂H (Figure 2b). The cyclic

voltammetry of 2a {Mn^III₂(µ-O)} in the absence and presence

of CCl₃CO₂H is shown in Figure 4a (i and ii). The effect of

Electrochemical Reduction of 1. A key aspect of the

catalytic oxidation of alkenes by 1 is the observation of a

lag period, at the end of which conversion of 1 to 2a occurs

(vide infra; see Figure 8).¹¹In the absence of a proton source

the one-electron reduction of 1 is observed at -600 mV (vs

SCE).^10cAlthough during catalysis the reduction of 1 by

H₂O₂is observed, to facilitate this, 1 must first undergo

protonation. In the presence of CCl₃CO₂H the reduction shifts

600 mV anodically and becomes, overall, a four electron

reduction process. On the basis of the cyclic voltammetric

data, it is apparent that the irreversible reduction of 1 in the

presence of CCl₃CO₂H yields, first, 2c which can undergo

subsequent oxidation to form the Mn^III₂complexes 2d (Figure

5, Scheme 3) and 2a. Preparative electrochemical reduction

of 1 in the presence of CCl₃CO₂H yields 2c and through

subsequent oxidation 2d (compare signal at 1.09 V in Figure

5a(i) with Figure 2b). The UV-vis spectra of 1 and 2c,d

obtained by bulk electrolysis of 1 are shown in Figure 6. As

addition of CCl₃CO₂H on anodic processes of 2a (i.e., Mn^III

2

to Mn^IIIMn^IVredox process, Table 1) is minimal, even with

up to 250 equiv of CCl₃CO₂H (Figure S4) added, confirming

that 2a itself is not protonated under these conditions.

However, addition of 1 equiv or more of CCl₃CO₂H to 2a

was found to be sufficient to render the Mn^IIIto Mn^IIMn^III

2

reduction completely irreversible.

In the presence of CCl₃CO₂H, the reduction at 0.29 V

appears to be a 2 e^-/1 H⁺coupled process. The two-electron

nature of the reduction wave, apparent from its intensity

relative to the Mn^III₂to Mn^IIIMn^IVoxidation waves, may be

misleading, however. Instead, the initial reduction step is

more probably a H⁺/e^-coupled reduction followed by a

expected for a Mn^IIcomplex the absorption spectrum is

2

featureless; however, for 2d a very intense absorption at

6358 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Figure 4. Cyclic voltammetry of 2a (1 mM) in CH₃CN/0.1 M (TBA)PF₆(a) in the absence (i) and presence (ii) of 10 mM CCl₃CO₂H and (iii) as for (ii)

except over a narrower potential window [current axes of the cyclic voltammograms (i) and (ii) offset for clarity] and (b) thin layer cyclic voltammetry of

2a (1 mM) in the presence of CCl₃CO₂H (10 mM) [initial cycle (thick line); subsequent cycles (thin line)]. When the initial sweep direction is anodic, the

process at 1.09 V is not observed for the initial cycle (not shown).

Figure 5. Cyclic voltammetry of 1 in CH₃CN (0.1 M KPF₆) (a) with 10 equiv of CCl₃CO₂H 0.1 V s^-1in CH₃CN/0.1 M KPF₆, before (i) and after (ii) bulk

reduction at -0.2 V and (b) after reoxidation at 1.10 V [initial scan direction anodic (thin line) and cathodic (thick line)].

Figure 6. UV-vis spectroscopy of 1 (1 mM) with CCl₃CO₂H (10 mM) in CH₃CN/0.1 M KPF₆: (a) UV-vis spectrum; (b) expansion of visible region

before (thick line) and after (dotted line) bulk reduction at -0.2 V and after bulk reoxidation at 1.1 V (thin line).

∼300 nm is observed, together with a weaker absorption at

468 nm. Significantly at 540 nm, the λ_maxof the absorption

of 2a (Figure 9a; vide infra), 2d shows only weak absorption

(Figure 6). In the presence of acetic acid, in place of CCl₃-

CO₂H, similar changes are observed; however, oxidation of

the Mn^II₂complex formed initially (equivalent to 2c) yields

the µ-oxido-bridged dinuclear complex 3 (Figure S5).

Electrochemistry of 2a-d in the Presence of H₂O₂. In

the presence of the substrate cyclooctene, the oxidation (E_p,a

Inorganic Chemistry, Vol. 46, No. 16, 2007 6359

de Boer et al.

Figure 7. (a) 2a (1 mM) with 10 equiv of CCl₃CO₂H and 1,2-dichlorobenzene and cyclooctene (1 M) in 0.1 M (TBA)PF₆/CH₃CN: prior to (black/thin

line); after (t ∼ 30 s, blue/thick line) addition of H₂O₂(50 equiv w.r.t. 2a) and 20 min after (red/dashed line). (b) As for (a) except without 1,2-dichlorobenzene

and cyclooctene.

Scheme 3. Pathways for Conversion of 1 to 2a by Reduction with H₂O₂in the Presence of CCl₃CO₂H²⁶

) 1.09 V) of 2c becomes more, albeit not completely,

chemically reversible (Figure 7b). Addition of H₂O₂(both

with and without cyclooctene present, Figure 7a,b, respec-

tively) renders the oxidation of 2c (E_p,a) 1.03 V) completely

irreversible; i.e., the return reduction steps are no longer

observed at E_p,c) 0.94 and 0.67 V. In addition a new

irreversible oxidation wave at E_p,a) 0.97 V is observed,

which remains until all H₂O₂has been consumed. It is

reasonable to conclude that this new redox wave is due to

ligand exchange of H₂O with H₂O₂for the complex 2c (an

equilibrium which would benefit from the increased ligand

lability expected for a Mn^II₂complex compared with Mn^III₂).

However, the new oxidation wave is not electrocatalytic in

nature; i.e., its intensity is not influenced significantly by

the presence of excess H₂O₂or the presence of an oxidizable

substrate. The irreversibility of the oxidation of this new

The return reduction processes at E_p,c) 0.94 and 0.67 V

(i.e., reduction of 2d) are observed again after all H₂O₂is

consumed.

UV-Vis, ¹H NMR Spectroscopy, and Mass Spectrom-

etry under Catalytic Conditions with 1 and 2a-d. The

“standard” reaction discussed in the present paper is the

catalytic oxidation of cyclooctene by 1 (0.1 mol %, 1 mM)

in the presence of CCl₃CO₂H (10 mM, 1 mol %) at 0 °C in

CH₃CN with H₂O₂(50 w/w % aqueous, 1.3 equiv w.r.t.

substrate) added continuously over 6 h, followed by 1 h

without addition of H₂O₂.^11,27The concentration of 1

employed facilitates direct spectroscopic study of the reaction

solution. Under these conditions a significant lag period is

observed (phase I, Figure 8), after which cis-dihydroxylation

and epoxidation begin simultaneously and both processes

show similar (up to 4h) time dependence (phase II). Finally,

toward the end of the reaction (phase III), the cis-diol

concentration begins to decrease due to oxidation to the

species to a Mn^IIIstate results from formation of 2a rather

2

than 2d. Indeed it is apparent that complex 2d reacts

with H₂O₂very rapidly, in stark contrast to 2a which is

largely unaffected by even a 50-fold excess of H₂O₂. The

stability of 2a in the presence of H₂O₂confirms that the

interaction of 2a with H₂O₂is a kinetically slow process.

(27) The addition of water (contained in the 50% H₂O₂) results in phase

separation of the reaction mixture, which results in catalyst instability

and decomposition to species that can engage in catalase activity. When

the H₂O₂is added in one batch, the selectivity is unaffected; however,

substrate conversion is significantly reduced.

6360 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

not 1, as after the lag period 1 is not detectable by either

UV-vis or ESI-MS.

¹H NMR spectroscopy allows for the direct observation

of complexes such as 2a. However, due to the paramagnetic

nature of the Mn^III₂bis(carboxylato) complexes, it is required

that nonstandard catalytic conditions are employed to observe

spectra of the complexes (Figure S9). Addition of H₂O₂(20

equiv w.r.t. 1) to a mixture of 1 (5 mM) and CCl₃CO₂H (50

mM) in acetonitrile results in the quantitative formation of

2a also, in addition to weaker signals assigned, tentatively,

to 2d [Mn₂^III(OH)₂]²⁺. The presence of this latter species,

which does not show significant absorbance at 545 nm (the

λ_maxof 2a), accounts, in part, for the observation of only

85-90% (after 2 h at 20 °C) of the expected signal intensity

of 2a by UV-vis spectroscopy (Figure S7). As with substrate

conversion (Figure 8), a lag period of 30 min (60-90 min

at 0 °C) is observed for the reduction/ligand exchange of 1

to form 2a (by UV-vis, Figure 9, and ESI-MS, Figure S8)

and the onset of catalytic oxidation (Figure 8 and Scheme

3).

Figure 8. Typical time course for product formation (cis-diol, triangles;

epoxide, circles) and substrate consumption (cyclooctene, squares) under

“standard” conditions (see text for details) in the catalyzed oxidation of

cyclooctene with H₂O₂, with 1 (1 mM) and CCl₃CO₂H (10 mM). Key: phase

I, lag period; phase II, normal reactivity observed; phase III, subsequent

oxidation of cis-diol product observed.

Although the pK_aof 1 is -2,^10cin acetonitrile containing

CCl₃CO₂H (10 mM), a significant amount of the complex

is present in the protonated form. The reduction of H1⁺is,

primarily, responsible for the lag period observed prior to

the onset of catalytic activity. However as H₂O₂is incapable

of reducing 1 at an appreciable rate even in the presence of

a proton source such as HPF₆, it is clear that the formation

of the Mn^II₂and Mn^III₂bis(carboxylato)-bridged systems, e.g.,

2a-d, results in an autocatalytic reduction of H1⁺(Figures

S7a and S11).²⁹Notably the onset of the autocatalytic

reaction is delayed by the presence of cyclooctene. Similarly,

increased acid concentration above 10 equiv w.r.t. 1 only

serves to slow the conversion once initiated. This latter

observation can be rationalized by considering that higher

acid concentrations favor formation of tris(µ-carboxylato)

Figure 9. UV-vis spectrum of 2a (1 mM, blue dashed line) and a mixture

of 1 (1 mM)/CCl₃CO₂H (10 mM)/cyclooctene (1 M) after 0 min (red solid

line), 5 min (blue dotted line), and 20 min (black dash-dot-dash line) at

20 °C in CH₃CN. [For the full vis-NIR electronic spectra of 2a and 3

(400-2000 nm), see Figure S9].

Mn^IIspecies (Figure S3).

2

Catalytic Oxidation of Cyclooctene with H₂O₂Cata-

lyzed by 1. The formation of Mn^IIIbis(µ-carboxylato)

2

complexes, such as 2a, coincides with the onset of catalytic

activity of 1/CCl₃CO₂H. Indeed the formation of the complex

in situ by addition of excess H₂O₂(53 equiv) to 1 (1 mM)

and CCl₃CO₂H (10 mM) at 20 °C 20 min prior to addition

of cyclooctene and cooling to 0 °C results in a significant

reduction in the length of the lag period (Figure 10). This

underlines that the formation of 2a is a prerequisite for

catalytic activity (Scheme 3). Overall, the effect of this

procedure after 7 h is minor in terms of substrate conversion

and cis-diol/epoxide ratio (Table S2). However, the incom-

plete reduction in the lag period, in terms of activity, and

the synchronization of the reaction with the standard reaction

after 1.5 h indicate that the formation of 2a is not the sole

factor responsible for the lag period.

corresponding R-hydroxyketone, while the epoxide concen-

tration continues to increase steadily.^11,17nThese distinct

phases are observed for all carboxylic acids examined (Figure

1); however, the duration of the lag period and the extent of

further oxidation of the cis-diol product varies significantly

(vide infra, Figure 16).

During the lag period of the reaction (∼30 min at 20 °C

and 1 h at 0 °C), no change in either the UV-vis or ESI-

MS spectra occurs, and only very weak EPR signals^10care

observed (vide infra). At the end of the lag period (i.e., the

point at which conversion of cyclooctene begins, ∼1 h at 0

°C), both UV-vis spectroscopy and ESI-MS show quantita-

tive conversion of 1 (predominantly to 2a, Scheme 3, Figure

9). It is also apparent from the intensity of the UV-vis

spectrum that the major species (>95%, by comparison with

an authentic sample of 2a) present after the lag period is 2a

(Figure S7).²⁸The remainder (<5%), however, is certainly

(28) For all carboxylic acids examined, the formation of the Mn^IIIbis-

2

(carboxylato) complexes (i.e. 2a, 3-16, Figure 1) coincides with

initiation of catalytic activity.

(29) The sigmodal shape of the curve showing the formation of 2a from 1

can be seen in Figures S7 and S11.

Inorganic Chemistry, Vol. 46, No. 16, 2007 6361

de Boer et al.

ratio 1.8 observed using 1 mol % CCl₃CO₂H, entry IV Table

S3. Hence, the difference in selectivity with increasing acid

concentration is due to reduction of the effective water

content of the reaction mixture.

A central question is the role of the carboxylic acid in

controlling the catalysts reactivity and selectivity. That the

carboxylic acid acts as a ligand is clear from structural data¹¹

and the activity and selectivity of 2a in the absence of

additional acid (Table S2, entry III). However, without

additional CCl₃CO₂H, the catalyst becomes less active toward

the end of the reaction and at ∼4.5 h becomes inactive

(Figure S10). This loss in activity coincides with a complete

decrease in UV-vis absorption of 2a. The loss in activity is

undoubtedly related to the dissociation of the carboxylato

ligands from the complex given that in the presence of excess

carboxylic acid the activity of the catalytic system is

retained,³²and hence, although excess carboxylic acid is not

essential for activity, its presence stabilizes the bis(carboxy-

lato) complexes involved.

Figure 10. Effect of pretreatment of the catalyst with H₂O₂prior to

addition of substrate on lag period (see text for details).

Catalytic Oxidation of Cyclooctene with 2a-c. The use

of the Mn^IIIcomplex 2a and its Mn^IIanalog 2b, in place

2

of 1, allows for the near complete elimination of the lag

period (Figure 12). However, lower reactivity is observed

for 2a in the first 1.5 h of the reaction compared to the

reaction catalyzed by 2b and during phase II of the reaction

(Figure 8).¹¹

For both 2b,c a negligible lag period is observed (Figure

12, Table S4), and the activity over the first 1 h with respect

to formation of the epoxide product is similar to that observed

after 1.5 h. The identical behavior of 2b,c is not, however,

surprising, considering that, under reaction conditions (i.e.,

in the presence of CCl₃CO₂H), 2b undergoes quantitative

conversion to 2c (vide supra). Under catalytic conditions with

either 2b or 2c, the cis-diol/epoxide ratio is lower over the

first 1.5 h of the reaction than observed at later stages of the

reaction. The difference in selectivity and the increased

reactivity observed for 2c (and 2b) compared with 2a in the

early stages (phase I) of the reaction is remarkable consider-

ing that within 15 min of the start of the reaction complex

2a is the predominant species present in solution (80-90%

by UV-vis, Figure S7 and Figure S11, and EPR spectros-

copy; vide infra). After ca. ∼90 min the selectivity of the

reaction with 2b or 2c is almost identical with that with 1

and 2a. Furthermore, whether the catalytic oxidation of

Figure 11. Effect of relative acid concentration on the catalytic oxidation

of cyclooctene by 1 (0.1 mol %) at 0 °C: (a) maximum total T.O.N. )

1000; (b) reaction performed in CH₃CN/H₂O (9:1 v/v). See Table S3 for

details. Note that although simple proton sources, e.g., HPF₆, together with

1, are inactive toward cyclooctene oxidation, they are effective in sup-

pressing catalase activity.³¹

Dependence of Catalyst Reactivity and Selectivity on

Relative Acid Concentration. The dependence of the

activity and selectivity of 1 on the relative concentration of

trichloroacetic acid is shown in Figure 11(and Table S3).

With less than 2 equiv of CCl₃CO₂H (w.r.t. complex 1),¹¹

very low activity is observed. With 2 equiv of CCl₃CO₂H

present, a large increase in activity is observed; however, a

further increase in acid concentration affects the reaction less,

with only a modest increase in activity and increased further

oxidation of the cis-diol product. Notably, increasing the

concentration of acid to 25 mol % results in enhanced

selectivity for epoxidation over cis-dihydroxylation.

cyclooctene is performed starting with complex 1, 2a [Mn^III

-

2

(µ-O)], 2b [Mn^II(µ-OH)], or 2c [Mn^II(µ-O₂H₃)], the turnover

2

numbers with respect to both cis-diol, epoxide. and the

conversion of cyclooctene are very similar (Table S4).

The change in selectivity observed with variation in CCl₃-

CO₂H concentration is due to solvent water content rather

than other effects such as peracid formation (see SI).^6,8,30

Indeed, when CH₃CN/H₂O (9:1 v/v) is used in place of CH₃-

CN as solvent for the reaction catalyzed by 1 (0.1 mol %)

with 25 mol % CCl₃CO₂H, the cis-diol/epoxide selectivity

increases from 0.79 to 1.5 (Figure 11, and entries VI and

VII in Table S3) with a negligible change in overall

cyclooctene conversion. The ratio 1.5 is comparable to the

(30) A detailed discussion of this issue is provided in the Supporting

Information. See also: (a) Fujitta, M.; Que, L., Jr. AdV. Synth. Catal.

2004, 346, 190-194. (b) Dubois, G.; Murphy, A.; Stack, T. D. P.

Org. Lett. 2003, 5, 2469-2472. (c) Murphy, A.; Pace, A.; Stack, T.

D. P. Org. Lett. 2004, 6, 3119-3122.

(31) In the presence of HPF₆(10 equiv), the catalase type activity observed

in absence of acid for 1 is substantially inhibited even over a period

of 3 h.

(32) At high acid concentrations, tris(µ-carboxylato)-bridged Mn^IIcom-

2

plexes form in equilibrium with e.g. 2c, which has an inhibitory effect

on the formation of the Mn^III₂species responsible for catalytic activity.

6362 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Figure 12. Effect of oxidation state and addition of H₂O prior to start of catalysis on the product distribution in the oxidation of cyclooctene catalyzed by

(a) 2a and (b) 2b: dotted lines/open symbols, with no H₂O added; solid lines/filled symbols, with added H₂O. Insets: Expansion of first 3 h period (0.1 mol

% catalyst, 1 mol % CCl₃CO₂H). Key: cyclooctene/squares; epoxide/circles; cis-diol/triangles. See Table S4 for further details.

The origin of the difference in both reactivity and

selectivity of the different complexes during phase I of the

reaction (Figure 8 and Figure 12) is intriguing and could

indicate the presence of a second catalytically active species

or catalytic pathway during the early period of the reaction;

however, solvent effects should be considered also. During

the catalysis, H₂O₂(50% aqueous solution) is added continu-

ously and, hence, the water content of the reaction mixture

increases as the reaction progresses. Indeed addition of H₂O

(equal to the amount added over 105 min of catalysis with

the added H₂O₂) prior to the start of the reaction catalyzed

by either 2a or 2b results in a significant difference in the

time dependence of product formation observed (Figure 12).

Importantly, the cis-diol/epoxide ratio for both 2a,b becomes

identical with that observed later in the reaction. Furthermore,

the reactivity of 2a is increased to match that observed after

1.5 h. Importantly the level of activity toward epoxidation

is affected much less by H₂O addition and the increase in

overall activity is due to increased cis-diol formation. This

latter observation may be related to the ligation strength of

cis-diols compared to epoxides and, hence, the need to

displace cis-diol once formed from the catalyst by H₂O (vide

infra).

The influence of the water content of the reaction mixture

was explored further using H₂O₂/CH₃CN, from which water

has been removed (Figure 13).³³From the time profile of

the reaction it is apparent that addition of anhydrous H₂O₂/

CH₃CN solution to the reaction mixture with 2a results in a

1:1 cis-diol/epoxide ratio.³⁴A control reaction with anhydrous

H₂O₂to which H₂O was added continuously to achieve the

same water content as observed normally showed a normal

cis-diol/epoxide ratio of ca. 2.3.³⁵The 1:1 cis-diol/epoxide

ratios observed during the initial phase of the reaction when

Figure 13. Time course for the formation of cis-diol products (triangles)

and epoxide (circles) from cyclooctene with anhydrous H₂O₂(solid lines)

and with additional water added with time (dashed lines). H₂O₂is added at

1/3 the normal amount and rate of addition (see Table S5 for details).

using, e.g., 2a or 2b, can thus be attributed to the low water

content of the reaction mixture during this earlier period

rather than to a difference in oxidation state (Mn^II₂vs Mn^III₂).

µ-Oxido and µ-Carboxylato Ligand Exchange Rates.

Ligand exchange of the µ-oxido bridge of 2a, 3, and 7 was

followed in time using ESI-MS (see SI for details). For

[Mn^III₂(µ-¹⁶O)(µ-CH₃CO₂)₂(tmtacn)₂]²⁺(3) relatively slow

exchange of the µ-oxido ligand with water-¹⁸O was observed

(exchange complete within 8 min at rt (room temperature),

Figure S12) in CH₃CN/H₂¹⁸O (9:1) and in the presence of 1

mol % CH₃CO₂H. A rate comparable to that reported recently

by Tagore et al. for related dinuclear complexes.³⁶For [Mn^III

-

2

(µ-¹⁶O)(µ-2,6-dichloro-benzoate)₂(tmtacn)₂]²⁺(7) with 2,6-

dichlorobenzoic acid the rate of exchange increased with

equilibration complete within 4 min, while for [Mn^III₂(µ-

¹⁶O)(µ-CCl₃CO₂)₂(tmtacn)₂]²⁺(2a) with CCl₃CO₂H conver-

sion was complete within 1 min, i.e., within the time

resolution of the experiment. Notably the rate of ¹⁸O

(33) 50% H₂O₂(w/w H₂O) was diluted in CH₃CN and dried over MgSO₄;

see ref 30. The use of dry H₂O₂sources such as UHP was considered;

however, practical difficulties, most notably the insolubility of the

reagents in CH₃CN, render this approach ineffective in assessing the

role of H₂O content.

(34) The increase in cis-diol/epoxide ratio observed toward the end of the

reaction (ca. 5-7 h) is due to the release of water from H₂O₂during

epoxidation. A reduced rate of H₂O₂addition (1/3rd normal) was

employed to minimize this effect.

(35) The high cis-diol/epoxide ratio of 2.3 compared to 1.8 under normal

conditions is due to the low level of cis-diol overoxidation at low

substrate conversion. Under normal conditions a cis-diol/epoxide ratio

of 2.5 is observed for the same (40%) conversion.

(36) Tagore, R.; Chen, H.; Crabtree, R. H.; Brudvig, G. W. J. Am. Chem.

Soc. 2006, 128, 9457-9465.

Inorganic Chemistry, Vol. 46, No. 16, 2007 6363

de Boer et al.

Figure 14. Total number of oxidation events (this is the total T.O.N. for oxidation of cyclooctene + oxidation of the cis-diol formed) compared with

number of equivalents of H₂O₂added: (left) 1 (0.1 mol %)/CCl₃CO₂H (1 mol %); (right) 2a (0.1 mol %)/CCl₃CO₂H (1 mol %) with H₂O (110 µL) added

at t ) 0 (see text for details).

exchange in this latter case is considerably retarded at lower

H₂¹⁸O concentrations.

H₂O₂Efficiency and the Effect of Lag Period. A key

feature of the present system is the atom efficiency with

respect to the terminal oxidant (H₂O₂). The efficiency of the

reaction, i.e., conversion of substrate and further oxidation

of the cis-diol with respect to H₂O₂added, is demonstrated

in Figure 14. For the catalytic system 1/CCl₃CO₂H, the H₂O₂

added during the lag period is consumed only partly during

later stages; however, once catalysis commences, the ef-

ficiency in terms of oxidant consumption is close to 100%.

For 2a/CCl₃CO₂H, almost all of the oxidant added is used

in oxidation of the substrate, albeit with a small amount of

H₂O₂used in the overoxidation of the cis-diol, in the final

stages of the reaction. The efficiency confirms that a near-

complete suppression of catalase type activity occurs. It is

of note that the rate of addition H₂O₂is matched by the

consumption of H₂O₂by the catalyst. This confirms that

under the standard conditions employed the turn over

frequency is limited by the addition of oxidant³⁷and not the

intrinsic activity of the catalyst. Indeed under otherwise

identical conditions, similar results are obtained over a range

of catalyst concentrations (0.38-0.038 mol %) with complete

conversion of substrate and similar product ratios.³⁸At lower

catalyst concentrations (0.0075 mol %) the conversion of

substrate is lower (63%, 8400 t.o.n.) and less further

oxidation of the cis-diol product occurs, leading to a slightly

higher cis-diol/epoxide ratio.³⁹

Figure 15. Solvent dependence of the oxidation of cyclooctene after 7 h

by 1 or 2a (0.1 mol %) with CCl₃CO₂H (1 mol %). (See text and Table S6

for further details.)

lower conversion of cyclooctene was observed than that

obtained in acetonitrile (Figure 15, Table S6). The higher

selectivity toward cis-dihydroxylation in these solvents in

comparison to acetonitrile is attributable to the low levels

of cis-diol overoxidation at low conversion (vide supra). In

acetone the formation of acetone/H₂O₂adducts lowers the

effective concentration of H₂O₂in solution.⁴⁰This has a

deleterious effect on the reactivity of 1/CCl₃CO₂H with H₂O₂

with no activity observed until approximately 4 h,⁴¹although

46% conversion is observed after 7 h (Table S6). The long

lag period at the start of the catalysis is in agreement with

the delay in conversion of 1 to 2a in acetone (by UV-vis

spectroscopy). By contrast, when 2a is used in place of 1,

oxidation activity commences earlier, albeit with a lower cis-

diol/epoxide ratio than observed in CH₃CN. It should be

noted, however, that the low mass balance observed in

acetone indicates increased oxidation of the cis-diol and

hence a lower cis-diol/epoxide ratio than is observed for other

solvents. Significantly reduced and no activity was observed

in dichloromethane and DMF, respectively.

Solvent Dependence of the Reactivity of 1/CCl₃CO₂H.

The influence of solvents other than acetonitrile in the

oxidation of cyclooctene was explored, and the results were

supported by UV-vis spectroscopy. In ^tBuOH/H₂O and THF

(37) The maximum tof possible under standard conditions is 0.058 s^-1

.

(38) It should be noted that the dependence of the reaction on the

concentration of catalyst, water, and carboxylic acid additive precludes

the systematic variation of any single parameter.

(39) This implies that the catalyst turnover frequency is 0.33 s^-1. The

maximum tof based on the rate of addition of H₂O₂is 0.77 s^-1

,

suggesting that under these conditions a maximum in catalytic activity

is observed; however, this is a lower limit to the tof, as this is calculated

over a 7 h period and further oxidation of the cis-diol product is not

accounted for.

(40) Edwards, J. O.; Sauer, M. C. V. J. Phys. Chem. 1971, 75, 3004-

3011.

(41) The formation of 2a is observed by UV-vis spectroscopy, ∼4 h also.

6364 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Table 2. Catalytic Oxidation of Cyclooctene by 2a (0.1 mol %) at 0 °C in the Presence of Trichloroacetic with H₂¹⁸O₂/H₂¹⁶O and H₂¹⁶O₂/H₂¹⁸O^a

T.O.N.

entry

H₂O

H₂O₂

acid (mol %)

cis-diol % ¹⁶O¹⁸O

epoxide % ¹⁸

O

cis-diol

epoxide

I

II

III

IV

V^b

VI^c

¹⁶O

¹⁸O

¹⁶O

CCl₃CO₂H (1)

103.9

93.0

79.7

93.0

94.9

95

71.2

37.6

20.8

33.5

38.2

32

120

105

12

96

67

55

52

18

50

58

CCl₃CO₂H (25)

CCl₃CO₂H (1)

427

234

^aCorrected for H₂O₂/H₂O isotopic composition; see Table S9 for full experimental details). ¹⁸O-labeling studies were performed on cyclooctene on 1/20th

scale of the standard conditions, with the adjustment that a 2% H₂O₂solution was used and the peroxide was added batchwise in four portions every 15 min

(i.e., at the same rate and amount of H₂O₂addition under typical reaction conditions). ¹⁸Oxygen incorporation was determined by GC-MS (CI) after 60 min

reaction time. Values are (5%. ^bcis-2-Heptene as substrate. ^cWith 20% H₂O₂(aq) solution.

Overall the solvent dependence of the reactivity of 1/CCl₃-

CO₂H correlates well with the stability of complex 2a under

catalytic conditions. That is, catalysis takes place only where

2a is stable (and can be formed from 1).⁴²Furthermore, the

fact that activity is observed in several, quite different

solvents (acetone, acetonitrile, ^tBuOH/H₂O, THF) indicates

that coordination of the organic solvent to the manganese

complex is not essential to the catalytic cycle.

Isotopic Labeling Studies. A key probe in “tracking”

oxygen atoms in oxidation catalysis is through ¹⁶O/¹⁸O

isotope labeling. In this study labeled dihydrogen peroxide

(H₂¹⁸O₂) and/or labeled water (H₂¹⁸O) are employed to

identify the source of the oxygen atoms incorporated into

both cis-diol and epoxide products.⁴³The lag time observed

Figure 16. Time dependence of cis-diol formation for HO₂C(CF₂)₃CO₂H,

CCl₃CO₂H, and 2,4,6-trichlorobenzoic acid promoted catalytic oxidation

for, e.g., 1/CCl₃CO₂H (i.e., the formation of bis(carboxylato)-

bridged complexes derived from 1) was circumvented by

using isolated and fully characterized [Mn^III₂(µ-O)(µ-

carboxylato)₂(tmtacn)₂]²⁺complexes, e.g., 2a. The results of

the labeling studies are given in Table 2.

of cyclooctene by 1 (standard conditions).

incorporation from H₂¹⁸O into the cis-diol (single incorpora-

tion, 80%) and the epoxide (21%) is observed, as well as

greatly reduced conversion (vide supra and Table S9).

The use of D₂O₂/D₂O in place of H₂O₂/H₂O resulted in

no significant changes to either the time dependence of

product formation or the overall conversion (Table S7),

indicating that proton (or hydrogen atom) transfer does not

play a role in the rate-determining step of the reaction.

Tuning of the Activity and Selectivity by Variation in

Carboxylate Ligands. Although in general both aliphatic

and aromatic carboxylic acids promote the oxidation of

cyclooctene to both cis-diol and epoxide with H₂O₂by 1,

the duration of the lag period (Figure 16), the cis-diol/epoxide

ratio, and the conversion are dependent on the specific

carboxylic acid employed (Table S8). For example, for

hexafluoroglutaric acid a lag time of only 15-30 min is

observed at 0 °C, while for 2,4,6,-trichlorobenzoic acid a

lag period of 2-3 h was observed (Figure 16 and Table S8).

Hence, in comparing the activity of carboxylic acids, it is

important to take the lag period into consideration, as an

apparent low activity can be due to slow formation of the

bis(carboxylato)-bridged dinuclear complex rather than an

intrinsically low reactivity.

Under standard reaction conditions (2a/1 mol % CCl₃-

CO₂H) with H₂¹⁸O₂(2% v/v in H₂¹⁶O), single ¹⁸O incorpora-

tion in the cis-diol product is observed, while ¹⁸O incorpo-

ration in the epoxide product is 71%. For the complementary

experiment with H₂¹⁶O₂(2% v/v in H₂¹⁸O), again single ¹⁸

O

incorporation in the cis-diol product and 38% ¹⁸O incorpora-

tion in the epoxide product is found. Very similar results

were obtained in the cis-dihydroxylation/epoxidation of cis-

2-heptene (Table 2, entry V).⁴⁴

Interestingly, at 25 mol % CCl₃CO₂H, no significant

difference in ¹⁸O incorporation in either the cis-diol or

epoxide product is observed compared with the reaction with

1 mol % CCl₃CO₂H (entry IV); however, in the absence of

trichloroacetic acid (using 2a as catalyst), reduced ¹⁸O

(42) It should be noted that the absence of activity or reduced activity

observed in different solvents can, potentially, be due to competitive

solvent oxidation which leads to a reduced efficiency in terms of

cyclooctene conversion. However, catalytic activity toward alkene

oxidation was observed only when 2a was present in solution.

(43) A difficulty arises, however, in the availability of H₂¹⁸O₂as a 2% v/v

solution in contrast to the 50% w/w H₂O₂solution employed in typical

catalysis experiments. However, since the presence of water in the

reaction medium allows for “normal” catalysis to take place, the

relatively large amount present in the labeling studies is not considered

problematic. In addition, the cis-diol/epoxide ratios obtained at several

H₂O₂/H₂O ratios (50, 30, 20, and 2%) indicate no significant difference

in the chemistry observed under labeling conditions compared with

standard reaction conditions.

Since the primary species present in solution during

catalysis are Mn^III₂species such as 2a,d, ditopic carboxylato

ligands (i.e., glutaric and hexafluoroglutaric acid) were

examined for activity and the duration of the lag period also

(Table S8). Both the activity and selectivity of these two

(44) Overall, the spectral features of the reaction solution are unaffected

by variation in substrate.

Inorganic Chemistry, Vol. 46, No. 16, 2007 6365

de Boer et al.

indicating that, in the case of 2-hydroxybenzoic acid,

coordination of the proximal -OH group to the manganese

centers does not occur under catalytic conditions.⁴⁵By

contrast, steric demands appear to be more important with

regard to selectivity, with 2,6-disubstituted benzoic acids

providing consistently higher cis-diol/epoxide ratios without

loss in activity, compared with their ortho-mono-substituted

analogs. For instance, with 2,6-difluorobenzoic acid, the

activity is comparable to that with 2,6-dichlorobenzoic acid

but the selectivity observed was comparable to that with

benzoic acid. Furthermore, for the 2,4,6-trimethylbenzoic

acid promoted reaction, although very low reactivity is

observed, the cis-diol/epoxide selectivity is high (∼5, Table

S8). This suggests that while the activity is driven by both

electronic and steric effects, the selectivity is dominated by

steric factors.

Figure 17. Cis-dihydroxylation and epoxidation of cyclooctene by 1 (0.1

mol %) in the presence of selected methoxy-, chloro-, and hydroxy-

substituted benzoic acids (1 mol %). See also Tables S2 and S8.

As with 2a (vide supra), for other µ-carboxylato-bridged

complexes (i.e., 2,6-dichloro- (7), 2,4-dichloro- (6), 4-hy-

droxy- (9), and 2-hydroxybenzoato (8)), ¹⁸O incorporation

from H₂¹⁸O into the cis-diol product is relatively insensitive

to the nature of the acid (82-94%, single oxygen incorpora-

tion from the H₂O₂); however, for the epoxide product the

extent of ¹⁸O incorporation from H₂¹⁸O ranges from 3 to 38%

depending on the complex/carboxylic acid employed (Table

2 and Table 3).

carboxylic acids are comparable with their mono(carboxylic

acid) counterparts (acetic and trifluoroacetic acid, respec-

tively). For hexafluoroglutaric acid a shorter lag period is

observed compared with trifluoroacetic acid. For glutaric

acid, however, the increased activity observed compared with

acetic acid is somewhat surprising, when it is considered

that glutaric acid exhibits a much longer lag period. That

the ligands are ditopic is apparent from mass spectral

analysis, and hence, an increased effective local concentration

of carboxylic acid may be responsible for the increased

activity observed. It is worth noting, however, that for the

glutarato-bridged complex 4 formation of the “oxido-bridge-

opened” Mn^III₂complex (analogous 2d) is more pronounced

than observed for the acetate-bridged complex 3 (by elec-

EPR Spectroscopy of 1/CCl₃CO₂H and 2a-c under

Reaction Conditions. EPR spectroscopy presents a powerful

tool in the study of manganese-based catalytic systems.

Although, from UV-vis spectroscopy and mass spectrometry

it is clear that the major species present in solution are EPR-

silent Mn^III₂dinuclear complexes, such as 2a ([Mn^III₂(µ-O)(µ-

1

trochemistry and H NMR spectroscopy), and hence while

in the Mn^IIIstate this will facilitate coordination of H₂O₂,

CCl₃CO₂)₂(tmtacn)₂]²⁺), the availability of EPR-active (Mn^II)

2

in the Mn^II₂state it will also encourage formation of tris(µ-

carboxylato)-bridged complexes which inhibits conversion

of 1 to 4 and hence extends the lag period observed.

For benzoic acid and its F-, Cl-, HO-, MeO-, and Me-

substituted analogs, activity toward both cis-dihydroxylation

and epoxidation are observed and selected data are presented

in Figure 17 (for a more comprehensive data set, see Table

S8). Several inferences regarding the relative importance of

steric and electronic effects toward reactivity and selectivity

can be drawn. With the exception of 2,4,6-trimethyl- and

4-chlorobenzoic acid, for all substituted benzoic acids

examined, an increase in activity compared with benzoic acid

is observed. Comparison of ortho-, meta-, and para-mono-

substituted benzoic acids shows only minor differences in

selectivity; however, overall para-substitution results in a

significant decrease in activity. For the substituted benzoate

based complexes, there appears to be no clear correlation

between electronic parameters and either activity or selectiv-

ity. However, it should be noted that, in contrast to the

aliphatic acids (e.g., acetic and trichloroacetic acid), the effect

of substitution on the redox properties of the complexes is

relatively minor (Table S1) and the activity observed is

affected significantly by the length of the lag period.

Importantly, the selectivity of the reaction shows only

moderate sensitivity to the position of the HO- group in

the hydroxybenzoic acid series (i.e. meta-OH ∼ ortho-OH)

complexes such as 2b,c allows for assessment of their

involvement during catalysis.⁴⁶

The EPR spectra of both 2b,c⁴⁷in acetonitrile (1 mM at

77 K, Figure 18a) are characterized by very broad features

at g ∼ 2, which do not show discernible hyperfine splitting

(except in the presence of cyclooctene; vide infra). The

spectra of 2b,c bear a remarkable similarity to that of the

dinuclear manganese(II) catalase enzyme from T. Thermo-

phillus reported by Dismukes and co-workers with peak to

peak separations of 530 G and a hyperfine splitting of 45

G.⁴⁸As observed by cyclic voltammetry (vide supra),

addition of 10 equiv of CCl₃CO₂H to 2b results in the

appearance of a spectrum identical with that of 2c. The

spectrum of 2c itself in CH₃CN is unaffected by addition of

CCl₃CO₂H. However, addition of cyclooctene to a solution

of 2c in CH₃CN leads to the formation of a mixture of 2b,c.

For 2c broad signals are observed in acetonitrile; however,

(45) The mononuclear Mn^IIIcomplex 8 converts to a bis(carboxylate)-

bridged complex under catalytic conditions.¹¹

(46) A detailed analysis of the magnetic properties including temperature

dependence and fitting is beyond the scope of the present discussion

and will be reported elsewhere.

(47) The spectrum is very similar to that of a structurally related complex

based on the bpea (N,N′-bis(2-pyridylmethyl)ethylamine) ligand.

See: Romero, I.; Dubois, L.; Collomb, M.-N.; Deronzier, A.; Latour,

J.-M.; Pe´caut, J. Inorg. Chem. 2002, 41, 1795-1806.

(48) Pessiki, P. J.; Khangulov, S. V., Ho, D. M.; G. C. Dismukes, J. Am.

Chem. Soc. 1994, 116, 891-897.

6366 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Table 3. ^16/18O Isotopic Distribution in the Oxidation Products of Cyclooctene by [Mn^III₂(µ-O)(µ-carboxylato)₂(tmtacn)₂]²⁺Complexes (0.1 mol %) at

0 °C in the Presence of the Corresponding Carboxylic Acid with H₂¹⁶O₂/H₂¹⁸O (Corrected for H₂O₂/H₂O Isotopic Composition)^a

T.O.N.

entry

H₂O

H₂O₂

catal (0.1 mol %)

acid (mol %)

cis-diol % ¹⁶O¹⁸O

epoxide % ¹⁸

O

cis-diol

epoxide

I

II

III

IV

V

¹⁸O

¹⁶O

6

7

9

8

2,4-dichlorobenzoic acid (1)

2,4-dichlorobenzoic acid (25)

2,6-dichlorobenzoic acid (3)

4-hydroxybenzoic acid (1)

2-hydroxybenzoic acid (1)

89.8

93.9

92.8

86.4

81.6

13.4

11.0

18.6

11.6

3.1

60

109

115

20

24

40

12

37

50

23

a

¹⁸O-labeling studies were performed on cyclooctene on 1/20th scale of standard conditions, with the adjustment that a 2% H₂O₂solution was added

batchwise in four portions every 15 min (i.e., at the same rate and amount of H₂O₂addition under typical reaction conditions). See Table S9 for full

experimental details.

Figure 18. EPR spectra (X-band at 77 K) of (a) 2c in CH₃CN (1 mM, 77 K), (b) 2b (1 mM) in CH₃CN/1 M cyclooctene, and (c) 2b (1 mM) in CH₃CN/1

M cyclooctene with 10 equiv of CCl₃CO₂H acid. Recording conditions: microwave power, 63 mW; modulation amplitude, 0.1 mT; modulation frequency,

50 kHz; time constant, 81.92 ms; T, 77 K; 9.47 GHz.

a decrease in solvent polarity by addition of cyclooctene,

although not affecting the overall shape and intensity of the

spectrum, leads the appearance of a distinct hyperfine-

coupling pattern (a ) 43 G, Figure S13).⁴⁹The observation

of a solvent polarity dependence on the resolution of the

hyperfine structure for 2c is in agreement with the assignment

of a µ-O₂H₃bridging unit being present (Figure 1). Further-

more, for 2b⁵⁰in the presence of both cyclooctene and 10

equiv of CCl₃CO₂H, an identical spectrum is observed, again

showing fine structure with hyperfine splitting (a ) 43 G,

Figure 18c).

The oxidation of cyclooctene by H₂O₂catalyzed by 1 (1

mM) in the presence of CCl₃CO₂H (10 mM) at 20 °C was

followed by EPR spectroscopy under standard catalytic

concentrations. After 15 min, a multiline signal is observed

(attributed to 2c present at low concentrations, vide supra),

present at low concentrations (Figure S14).⁵¹This weak

signal decreases in intensity over the first 1 h of the reaction

with the appearance of a very weak 6 line signal (a ) 90

G).⁵²Interestingly, a weak signal, assigned to 2c, is observed

in the EPR spectrum of 2a (1 mM), in cyclooctene/CCl₃-

CO₂H (10 mM)/CH₃CN, prior to addition of H₂O₂. Addition

of H₂O₂results in the disappearance of this signal. After 45

min the appearance of a very weak 6 line signal is observed

(a Mn^II-carboxylato complex dissociated from tmtacn, a )

90 G).⁵²

In summary, by EPR spectroscopy the conversion of 2b

to 2c in the presence of CCl₃CO₂H and the subsequent

conversion of 2c to an EPR-silent complex (i.e., 2a) upon

addition of H₂O₂is observed (Figure S14c). Overall, although

several species are observed by EPR spectroscopy, namely

2c (Figure S16) and free mononuclear Mn^II-tris(carboxylato)

complexes,⁵²the EPR signal is weak during catalysis, in

agreement with electrochemical, ESI-MS, and UV-vis

spectroscopic studies in which the major species observed

is the EPR-silent Mn^IIIcomplex 2a.⁵³

2

(49) The nature of the bridging µ-H₃O₂unit is currently under further

investigation. For similar examples of this structural motif, see the

following: (a) Stibrany, R. T.; Gorun, S. M. Angew. Chem., Int. Ed.

Engl. 1990, 29, 1156-1158. (b) Boudalis, A. K.; Lalioti, N.;

Spyroulias, G. A.; Raptopoulou, C. P.; Terzis, A.; Bousseksou, A.;

Tangoulis, V.; Tuchagues, J.-P.; Perlepes, S. P. Inorg. Chem., 2002,

41, 6474-6487. (c) Jaime, E.; Weston, J. Euro. J. Inorg. Chem. 2006,

793-801. (d) Meyer, F.; Rutsch, P. Chem. Commun. 1998, 1037-

1038. (e) De Munno, G.; Viterbo, D.; Caneschi, A.; Lloret, F.; Julve,

M. Inorg. Chem., 1984, 33, 1585-1586. (f) 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. J.; Kauffmann, K.; Munck, E.

J. Am. Chem. Soc. 1995, 117, 2778-2792. (g) Ardon, M.; Bino, A.;

Michelsen, K.; Pedersen, E.; Thompson, R. C. Inorg. Chem. 1997,

36, 4147-4150.

(51) On the basis of intensity compared with 2c (1 mM) under identical

conditions.

(52) At higher concentrations (5 mM 1/50 mM CCl₃CO₂H) in butyronitrile/

acetonitrile, although not affecting the outcome of the catalysis

significantly (Figure S15), the appearance of the multiline signal

assigned to 2c is more pronounced; however, again a steady-state

concentration is reached after 30 min and after 2 h the multiline signal

disappears again, replaced by a six line signal typical of mononuclear

Mn^II. The increased intensity of the Mn^IImononuclear signal is due

in part to the overall increase in concentration but also to the increased

acid concentration. The mononuclear species is assigned, tentatively,

as [Mn^II(CCl₃CO₂)₃]^-on the basis of comparison with Mn^II(ClO₄)₂/

CCl₃CO₂H (1:10) in CH₃CN and the observation of the ion in negative

mode ESI-MS. Mn^II(ClO₄)₂/CCl₃CO₂H (1:10) is inactive under the

reaction conditions employed in the present study; see Table S11, entry

VI.

(50) For 2b a decrease in solvent polarity by addition of cyclooctene alone

does not affect the EPR spectrum (Figure 18).

Inorganic Chemistry, Vol. 46, No. 16, 2007 6367

de Boer et al.

Scheme 4. Summary of Process Which Occur during Phases I-III^a

^aSee Figure 8.

Overall it is apparent that although the formation of Mn^II

Mn^II₂complexes, i.e., [Mn^II(µ-O₂H₃)(µ-RCO₂)₂(tmtacn)₂]²⁺

2

and Mn^IIIbis(µ-carboxylato) species occurs at least within

(cf. 1/dichlorobenzoic acid), may be present during the lag

period for some of the carboxylic acids examined, the onset

of catalytic activity does not coincide with their formation.

Furthermore, the observation that 2c can interact with H₂O₂

to form a metastable complex (as observed by electrochem-

2

30 min of initiation of addition of H₂O₂to the reaction

mixture, the initiation of catalytic activity is observed only

when the Mn^IIIbis(µ-carboxylato) complexes, e.g., 2a, 7,

2

etc., form.

istry) supports the conclusion that Mn^IIcomplexes are not

Overview of Complex Formation and of Oxidation

Processes. A summary of the complexes involved in the

catalytic system is provided in Scheme 4. In the carboxylic

acid promoted catalytic oxidation of alkenes by 1, the first

step is reduction of H1⁺by H₂O₂and ligand exchange to

form 2a. Hence, the use of dinuclear manganese bis(µ-

carboxylato) complexes, either prepared or generated in situ

prior to initiation of catalysis, results in the elimination of

the lag period observed with 1.

Under reaction conditions, prior to addition of H₂O₂, the

Mn^II₂complex 2b is converted quantitatively to 2c, and upon

addition of H₂O₂, almost complete conversion to 2a is

observed within the first 5 min of the reaction. Although

2

involved in the catalytic cycle directly.

Regardless of the various processes that occur prior to the

onset of catalytic activity, it is certain that catalysis is

observed only where either 2a or 2d is present in solution.

In the presence of H₂O₂, 2d is not observed until all H₂O₂

has been consumed (see Figure 7), whereas the concentration

of 2a is relatively unaffected by the addition of H₂O₂. Hence,

it is reasonable to propose that the species which interacts

with H₂O₂is the dinuclear Mn^III₂complex, 2d. The prepon-

derance of 2a in solution (>85-95% depending on condi-

tions) compared with 2d suggests that 2a is directly involved

in the catalytic cycle and that the opening of the µ-oxido

bridge of 2a is limiting reactivity. That is, 2a acts as a catalyst

resting point, which is in equilibrium with 2d (Scheme 4).

(53) Similar results were obtained for the reaction catalyzed by 1 with 2,6-

dichlorobenzoic acid. See the SI for details.

6368 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

Coordination of H₂O₂to 2d is followed by reaction of the

H₂O₂activated species with the substrate to re-form 2a.

The primary role of H₂O in determining activity appears

to be the formation of 2d. The equilibrium between 2d,

which interacts with H₂O₂, and 2a is affected strongly by

both the amount of H₂O present and the carboxylic acid

concentration. That this is the case is further supported by

comparison of glutaric and acetic acid promoted reactions.

In the case of the glutaric acid based complex 4, formation

of the Mn^III₂(OH)₂species (analogous to 2d) occurs much

more readily than for the corresponding acetate-bridged

complex 3, and the glutarato-bridged complexes show much

higher reactivity than the corresponding acetato complexes,

despite being “electronically” equivalent. Furthermore from

ESI-MS ¹⁸O labeling studies (vide supra), it is clear that the

Figure 19. Proposed structures for H₂O₂-activated manganese com-

plexes.^55,58

to date, based on ESI-MS, mononuclear high-valent Mn^IV)O

and Mn^V)O species⁵⁵were proposed by Lindsay-Smith and

co-workers during the oxidation of phenols and the epoxi-

dation of cinnamic acid⁵⁶derivatives by Mn^II-tmtacn (in

the presence of oxalic acid and other additives) (Figure 19).⁵⁷

Recently, in a related tetraazabicyclohexadecane-based sys-

tem Busch and co-workers⁵⁸have suggested a mononuclear

Mn^IVperoxy complex (LMn^IV(dO)OOH, where L ) 4,11-

dimethyl-1,4,8,11-tetraazabicyclo[6.6.2]hexadecane) as being

the activated oxidant in the epoxidation of olefins, a

so-named inorganic peracid.⁵⁹On the basis of these studies,

it is tempting to postulate that a mononuclear active species

is operating in the present systems also.

rate of exchange of the µ-oxido bridge of the Mn^IIIbis-

2

(carboxylato) complexes is dependent on the nature of the

µ-carboxylato ligands with the rate increasing with increasing

electron-withdrawing character of the ligand.

Hence, both 2a,d can be implicated directly in the catalytic

cycle, and the rate-determining step in the reaction is the

formation of 2d from 2a. Furthermore, once formed, 2d

interacts with H₂O₂followed by oxidation of the substrate

to re-form 2a.

Mechanistic Considerations. Although two complexes

in the catalytic cycle, i.e., 2a/2d, have been identified

experimentally, a key question now arises; are two catalyti-

cally active species involved (one for cis-dihydroxylation and

one for epoxidation), or if a common H₂O₂activated

intermediate is responsible for both oxidation processes, what

is the nuclearity and nature of such a species?

However, despite extensive spectroscopic and electro-

chemical characterization of the present system, such mono-

nuclear species have never been observed. While that does

not exclude their involvement completely, this fact requires

that we consider also a mechanism which recognizes that

throughout the reaction >95% of the manganese is present

in solution as dinuclear Mn^IIIcomplexes (e.g., 2a,d).

2

Furthermore, the number of turnovers that the catalyst can

undergo during catalysis (>2500)¹¹and the absence of

spectroscopic evidence for any H₂O₂-activated complexes

during catalysis indicate that the rate-determining step in the

catalytic cycle involves formation of 2d from 2a. Subsequent

coordination of H₂O₂to 2d takes place followed by rapid

reaction of the activated complex (Scheme 4). In addition

the re-formation of 2a, after oxidation of the substrate, is

fast and virtually complete. Overall the inability of peroxide

Several aspects of the current catalytic system indicate that

a common intermediate is involved or at least that there is a

common immediate precursor to the species responsible for

cis-diol and epoxide formation, respectively. First, the

conversion of cyclooctene to cis-diol and epoxide commences

simultaneously and the processes which change the duration

of the lag period (e.g., catalyst preactivation, initial oxidation

state of the catalyst, H₂O content of the reaction mixture,

etc.) all effect the lag time for both cis-dihydroxylation and

epoxidation in the same manner. Second, both cis-dihy-

droxylation and epoxidation show similar time dependence

over the course of the reaction. Significantly lower cis-diol/

epoxide ratios are observed during the initial stages of the

reaction when either the well-defined complexes 2a-c are

used as catalysts or high acid concentrations are employed.

However, this decrease is due to the low water content of

the reaction mixture during the early stages and the “dehy-

drating” effect of high acid concentrations since addition of

H₂O restores normal catalytic behavior. Hence, any reaction

mechanism should take into consideration a common catalyst

being, most probably, responsible for both cis-dihydroxyla-

tion and epoxidation.

either to reduce or oxidize the Mn^IIIdinuclear complex 2a

2

(54) For recent reviews, see the following: (a) Sibbons, K. F.; Shastri, K.;

Watkinson, M. Dalton Trans. 2006, 645-661. (b) Hage, R.; Lienke,

A. J. Mol. Catal. A 2006, 251, 150-158.

(55) In this work, no such species have been identified either by UV-vis,

ESI-MS, or EPR spectroscopies. In particular, no signals were observed

at 1600 G (g ) 4) typical of either a mono- or dinuclear Mn^IVspecies.

(56) From Hammett parameters on a series of substituted cinnamic acid

derivatives, it was concluded that the active species is electrophilic in

nature, and ¹⁸O-labeling study showed incorporation of oxygen in the

epoxide from H₂O (ca. 10%) in addition to incorporation from H₂O₂

(ca. 90%) using oxalic acid as additive: Gilbert, B. C.; Lindsay Smith,

J. R.; Mairata i Payeras, A.; Oakes, J.; Pons i Prats, R. J. Mol. Catal.,

A: Chem. 2004, 219, 265-272.

(57) Gilbert, B. C.; Smith, J. R. L.; Mairata i Payeras, A.; Oakes, J. Org.

Biomol. Chem. 2004, 2, 1176-1180.

(58) (a) Yin, G.; Buchalova, M.; Danby, A. M.; Perkins, C. M.; Kitko, D.;

Carter, J. D.; Scheper, W. M.; Busch, D. H. Inorg. Chem. 2006, 45,

3467-3474. (b) Yin, G.; Buchalova, M.; Danby, A. M.; Perkins, C.

M.; Kitko, D.; Carter, J. D.; Scheper, W. M.; Busch, D. H. J. Am.

Chem. Soc. 2005, 127, 17170-17171.

(59) Deubel, D. V.; Frenking, G.; Gisdakis, P.; Hermann, W. A.; Rosch,

N.; Sundermeyer, J. Acc. Chem. Res. 2004, 37, 645-652.

The majority of mechanisms, proposed previously for the

tmtacn family of manganese catalysts, have favored the

formation of mono- and dinuclear high-valent oxidizing

species,⁵⁴albeit with relatively scant empirical evidence to

support such proposals. In the most detailed study available

Inorganic Chemistry, Vol. 46, No. 16, 2007 6369

de Boer et al.

Scheme 5. Possible H₂O₂-Activated Species and Observed Oxygen

Incorporation from H₂O₂and H₂O in cis-Diol and Epoxide Products^a

and the rapid and irreversible interaction of 2d with H₂O₂

suggest that the system may bear a strong similarity with

dinuclear carboxylato-bridged catalase enzymes. Furthermore

the formation of high-valent species is at most transient, i.e.,

during transfer of oxygen to the substrate (or hydrogen

abstraction in the case of C-H activation). Indeed several

aspects of the catalytic system show remarkable analogy with

the manganese catalase¹³enzymes, such as the importance

of the carboxylato ligand and the exchange of aqua and aquo

ligands with H₂O₂.

It is clear that neither 2a, 2c, nor 2d can effect direct

oxygen transfer as all three complexes show full stability in

the presence of cyclooctene. However, in the present system

it is improbable that the dinuclear complex “splits” to form

two mononuclear complexes prior to interaction with H₂O₂.

The stability of the complex throughout the reaction and the

deactivation of the catalyst observed upon loss of a car-

boxylato ligand provide credence to this conclusion (Figure

S10).

^aKey: (B) oxygen originating from H₂O.

Several possible dinuclear structures for the “H₂O₂-

activated” complex are depicted in Scheme 5. Coordination

of H₂O₂to the manganese center can form either a η¹-O-

OH (species A) or Mn-O-O-Mn (species C).

In the case of species A, the homolysis of the O-O bond

to form OH* radicals is highly unlikely as the involvement

of hydroxyl radicals in the catalytic oxidation has been

excluded experimentally.¹¹The ability to engage in oxidation

of alkanes and alcohols¹¹would suggest that a high-valent

(e.g., Mn^V)O)⁶⁰species is involved; however, although

heterolysis of the O-O bond in species A would yield a

Mn^IIIMn^Vspecies (B), such a species can exist only

transiently before intramolecular electron transfer takes place

to form a (Mn^IV)O)₂species (D).⁶¹

The very transient nature of the “H₂O₂-activated” complex,

i.e., the formation of 2d from 2a, is rate determining with

subsequent reaction of the active species with the alkene

being very fast, and the absence of distinct spectroscopic

features requires that indirect methods be employed to gain

information as to the species nature. ¹⁸O isotopic labeling

of both H₂O and H₂O₂provides a powerful probe in this

regard as has been demonstarted in the recent work of Que

and others on iron-based systems.^8,21Although it is clear that

only one of the oxygen atoms incorporated into the cis-diol

product originates from H₂O₂and the other oxygen atom

from H₂O, the situation is less clear for the epoxide products,

where both H₂O₂and H₂O provide the oxygen atom. The

For species C, direct reaction with the substrate is unlikely

as this would result in both oxygens of the dihydrogen

peroxide being incorporated into the cis-diol product and

100% incorporation into the epoxide. Homolysis of the O-O

bond, however, would lead to the formation of the Mn^IV

2

dependence of ¹⁸O incorporation into the epoxide from H₂¹⁸

O

species D, and although exchange of one of the oxygen atoms

with water would rationalize the ¹⁸O incorporation into the

products, a fast exchange of only one of the oxygen atoms

with water is required with the remaining oxygen from the

dihydrogen peroxide being kinetically inert to exchange. For

a symmetric species such as D, such behavior is highly

unlikely. The incorporation of oxygen from water into the

epoxide would demand that slow exchange takes place as

the degree of incorporation is not statistical. Indeed the

incorporation of oxygen from water is highly dependent on

the nature of the carboxylate employed while for the cis-

diol product the ratio appears to be dependent, albeit only

moderately, on the rate of Mn-OH exchange with water.

It is therefore, in our view, a reasonable assumption that

the reaction of H₂O₂with 2d leads, initially, to the formation

of a Mn^III(O-OH)Mn^III(OH) species, i.e., A. The Mn^IIIcenter

on the specific acid employed is as striking as is the relative

insensitivity of the 1:1 oxygen incorporation into the cis-

diol from H₂O and H₂O₂. However, for the epoxide there is

a correlation between the selectivity of the catalyst/carboxylic

acid system toward the cis-diol/epoxide ratio and the

incorporation of oxygen into the epoxide from H₂O. The

incorporation of oxygen from water into the epoxide product

ranges from 18% for the 2,6-dichlorobenzoate complex (cis-

diol/epoxide ratio 7) to 13.4% for the 2,4-dichlorobenzoic

acid complex (cis-diol/epoxide ratio 2.7) and to 3.4% for

the salicylic acid (cis-diol/epoxide ratio 0.7). Hence, the more

electrophillic the Mn-OH group of the Mn(OOH)-Mn(OH)

species is the higher the cis-diol ratio will be. Furthermore,

for the more electron withdrawing acid CCl₃CO₂H, although

the selectivity toward cis-diol formation is lower than that

observed for 2,6-dichlorobenzoic acid, the incorporation of

oxygen from water into the epoxide product is much higher

(33%). In the absence of added carboxylic acid, more oxygen

from the H₂O₂is incorporated into both the cis-diol and

epoxide products, with a significant amount of cis-diol

showing both oxygen atoms originating from H₂O₂. This

observation can be rationalized by considering that the rate

of exchange of Mn-OH with H₂O is slower in the absence

of an excess of carboxylic acid.

(60) (a) Adam, W.; Roschmann, K. J.; Saha-Moller, C. R.; Seebach, D. J.

Am. Chem. Soc. 2002, 124, 5068-5073. (b) Meunier, B.; Guilmet,

E.; De Carvalho, M.; Poilblanc, R. J. Am. Chem. Soc. 2000, 122, 2675.

(61) As noted by the reviewers, a µ-oxido-bridged dinuclear iron(IV) oxo

complex has been proposed as an active species in the epoxidation

and cis-dihydroxylation of alkenes by the following: Kodera, M.; Itoh,

M.; Kano, K.; Funabiki, T.; Reglier, M. Angew. Chem., Int. Ed. 2005,

44, 7104-7106. The rate of exchange of oxygen atoms with water in

this species is noted to be fast and results in scrambling of the oxygen

atoms derived from homolytic cleavage of the Fe-O-O-Fe bond of

a species analogous to species C in Scheme 5.

6370 Inorganic Chemistry, Vol. 46, No. 16, 2007

Highly H₂O₂Efficient Dinuclear Mn Catalysts

a

Scheme 6. Rationalization of H₂O₂Activation by Dinuclear Manganese Bis(carboxylato)-Bridged Complexes Coordinated to H₂O₂

^aKey: (B) oxygen originating from H₂O.

polarizes the O-O bond to a much greater extent than in a

corresponding Mn^II-(O-OH) species would (Scheme 6).

Such polarization and hence reactivity is facilitated by

between cis-diol formation and epoxidation would be

expected to take place. Loss of H₂O, coupled with formation

of a Mn-O-Mn bond will result in re-formation of, e.g.,

2a with formation of the epoxide product. This step favors

the transfer of the oxygen atom of the Mn-O-O-H unit

to the alkene, but with more electron withdrawing carboxy-

lato ligands, the Mn-O-H oxygen can compete with this

process. Such a mechanism is in full agreement with ¹⁸O-

labeling studies for the epoxide product.

For the cis-diol product, the loss of H₂O from the transition

state is not coupled to formation of an Mn-O-Mn bridge.

Instead the cis-diol remains bound via both oxygens and must

be displaced subsequently by H₂O before reentering the

catalytic cycle either as, e.g., 2a or 2d. Hence, it would be

expected that µ-carboxylato ligands, which inhibit formation

of µ-oxido-bridged complexes such as 2a, should also inhibit

the formation of epoxide in favor of cis-diol. It is observed

empirically that steric hindrance is the most important factor

in determining selectivity with more sterically hindered

systems favoring cis-dihydroxylation. It would be expected,

and it is indeed the case, that incorporation of one oxygen

atom from H₂O₂and one from H₂O into the cis-diol product

would occur.

-

electron-deficient ligands such as CCl₃CO₂. Hence the

catalyst is expected to be electrophilic in nature, as is

observed experimentally. The polarization of the O-O bond

is reminiscent of the reductive step of the manganese catalase

cycle where the O-OH bond is cleaved by transfer of a

proton to the terminal oxygen (i.e., O-OH₂).⁶²The proton

required is available proximally in the present model from

the neighboring Mn-OH unit. The more electron withdraw-

ing the carboxylic acid, the greater the acidity of this Mn-

OH group will be and hence the more reactive the system.

A mechanism in which a high-valent manganese center is

present as a transient intermediate is proposed in Scheme 6.

From our earlier studies, we have demonstrated that the

oxygen transfer is a concerted process and hence the transfer

of both oxygen atoms, one from the solvent and one from

H₂O₂, should occur in a single step; i.e., for cis-dihydoxy-

lation and epoxidation both C-O bonds form simulta-

neously.¹¹This mechanism proposes that the oxygen atom

of the -OOH ligand, which binds to the manganese center,

and to a lesser extent the oxygen atom of the -OH ligand

on the adjacent manganese site engage in an electrophilic

interaction with the substrate to form the transition state.⁶³

It is at this point that the differentiation in mechanisms

Conclusions and Perspectives. Over the past decade

considerable successes in enhancing the activity of 1 toward

catalytic oxidative transformations were achieved most

notably through the use of additives.^17,64Among the various

approaches taken, the use of carboxylic acids has proven to

be the most effective in both suppression of the wasteful

(62) Whittaker, M. M.; Barynin, V. V.; Antnyuk, S. V.; Whittaker, J. W.

Biochemistry 1999, 38, 9126-9136.

(63) This mechanism is similar but opposite to the mechanism proposed

for a related manganese mononuclear catalyst in which the noncoor-

dinated oxygen atom of the Mn(dO)-O-OH species is activated for

oxygen transfer. See: Busch and co-workers.^58bThe approach of the

alkene porposed in this work is reminiscent of that proposed by Que

and co-workers for a mononuclear Fe(IV)(dO)OH species. See:

Bassan, A.; Blomberg, M. R. A.; Siegbahn, P. E. M.; Que, L., Jr.

Angew. Chem., Int. Ed. 2005, 44, 2939-2941.

(64) (a) Woitiski, C. B.; Kozlov, Y. N.; Mandelli, D.; Nizova, G. V.;

Schuchardt, U.; Shul’pin, G. B. J. Mol. Catal., A: Chem. 2004, 222,

103-119 and references therein. (b) Liu, B.; Chen, Y.; Yu, C-Z.; Shen,

Z-W. Chin. J. Chem. 2003, 21, 833-838. (c) Ryu, J. Y.; Kim, S. O.;

Nam, W.; Heo, S.; Kim, J. Bull. Korean Chem. Soc. 2003, 24, 1835-

1837.

Inorganic Chemistry, Vol. 46, No. 16, 2007 6371

de Boer et al.

disproportionation of H₂O₂and in controlling the activity

and selectivity of the catalytic system.¹¹In this study, we

have demonstrated that carboxylic acids play more than a

single role in the enhancement of the catalytic properties of

1, i.e., as a proton source to facilitate reduction of 1 (by

H₂O₂), as a bridging ligand, and in stabilizing the catalytically

active bis(µ-carboxylato)-Mn^III₂complexes formed after the

initial lag period.

It is in acting as a ligand, however, that the carboxylate

can exert control over both the activity and selectivity of

the catalytically active species. Although a clear relationship

between activity and the electron deficiency of the carboxylic

acid is apparent, with activity increasing with increased

electron-withdrawing character of the carboxylate (i.e., 2a

is more active than 3), an additional factor which should be

considered is the propensity to form species such as 2d.

Overall, carboxylate ligands, which promote formation of

such species, show higher activity compared with electroni-

cally equivalent carboxylate ligands. With regard to selectiv-

ity, steric factors appear to be dominant, with increasing steric

hindrance, at the 2,6-positions of the benzoic acid, favoring

cis-dihydroxylation over epoxidation.

investigations a detailed mechanistic insight into the catalytic

system can be deduced. The results obtained highlight the

crucial importance of considering both the bulk solvent

properties and molecular interactions in understanding

catalytic processes. Furthermore, the identification of a

dinuclear carboxylato bridged complex as being responsible

for catalysis requires that the preference for proposing high-

valent manganese oxido species as the active agent in

manganese-tmtacn-based catalysis may need to be recon-

sidered. The model proposed in this contribution provides,

in our view, a solid basis for future computational studies

and indeed may provide a useful model system for the related

bioinorganic dinuclear manganese centers found in catalase

enzymes.

Acknowledgment. We thank Ms. T. D. Tiemersma-

Wegman, Mr. C. Smit, Ms. A. van Dam, Dr. A. P. Bruins,

and Mr. A. Kiewiet for assistance with GC and ESI-MS

analysis and the Dutch Economy, Ecology, Technology

(Grant EETK01106) program for financial support. We

especially thank Prof. Jan Reedijk and Dr. J. Gerard Roelfes

for many discussions and their suggestions.

Supporting Information Available: Analytical and experi-

mental data for 2c and 4-16 and mass and UV-vis and EPR

spectral data (PDF). This material is available free of charge via

the Internet at http://pubs.acs.org.

Understanding the mode of action of promoting agents in

catalysis is key to the development of a rational approach to

tuning of both the activity and selectivity of the system in

question. In this study, we have demonstrated that through

a combination of spectroscopic, electrochemical, and catalytic

IC7003613

6372 Inorganic Chemistry, Vol. 46, No. 16, 2007

Article Doi

Mechanism of cis-dihydroxylation and epoxidation of alkenes by highly H2O2 efficient dinuclear manganese catalysts

DOI: 10.1021/ic7003613

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ic7003613

Full text of DOI:10.1021/ic7003613

Products guided by the article

R&D Labs maybe for 947409-21-4

Relevant to this article

Hot Product