Characterization of the O2-evolving reaction catalyzed by [(terpy)(H2O)MnIII(O)2MnIV (OH2(terpy)](NO3)3 (terpy = 2,2′:6,2″-terpyridine)

Article abstract of DOI:10.1021/ja001090a

The complex [(terpy)(H₂O)Mn^III(O)₂Mn^lV (OH₂)(terpy)](NO₃)₃ (terpy = 2,2′:6,2″-terpyridine) (1) catalyzes O₂ evolution from either KHSO₅ (potassium oxone) or NaOC1. The reactions follow Michaelis-Menten kinetics where V_max = 2420 ± 490 mol O₂ (mol 1)^-1 hr^-1 and Km = 53 ± 5 mM for oxone ([1] = 7.5 μM), and V_max = 6.5 ± 0.3 mol O₂ (mol 1)^-1 hr^-1 and K_M = 39 ± 4 mM for hypochlorite ([1] = 70 μM), with first-order kinetics observed in 1 for both oxidants. A mechanism is proposed having a preequilibrium between 1 and HSO₅^- or OCI^-, supported by the isolation and structural characterization of [(terpy)-(SO₄)Mn^IV(O)₂Mn^lV (O₄S)(terpy)] (2). Isotope-labeling studies using H₂¹⁸O and KHS¹⁶O₅ show that O₂ evolution proceeds via an intermediate that can exchange with water, where Raman spectroscopy has been used to confirm that the active oxygen of HSO₅^- is nonexchanging (t_1/2 ? 1 h). The amount of label incorporated into O₂ is dependent on the relative concentrations of oxone and 1.³²O₂:³⁴O₂:³⁶ O₂ is 91.9 ± 0.3:7.6 ± 0.3:0.51 ± 0.48, when [HSO₅^-] = 50 mM (0.5 mM 1), and 49 ± 21:39 ± 15:12 ± 6 when [HSO₅^-] = 15 mM (0.75 mM 1). The rate-limiting step of O₂ evolution is proposed to be formation of a formally Mn^V=O moiety which could then competitively react with either oxone or water/hydroxide to produce O₂. These results show that 1 serves as a functional model for photosynthetic water oxidation.

Full text of DOI:10.1021/ja001090a

J. Am. Chem. Soc. 2001, 123, 423-430
423
Characterization of the O2-Evolving Reaction Catalyzed by
[(terpy)(H2O)MnIII(O)2MnIV(OH2)(terpy)](NO3)3 (terpy )
2,2′:6,2′′-Terpyridine)
Julian Limburg,‡,| John S. Vrettos,‡ Hongyu Chen,‡ Julio C. de Paula,§
Robert H. Crabtree,*,‡ and Gary W. Brudvig*,‡
Contribution from the Department of Chemistry, Yale UniVersity, P.O. Box 208107,
New HaVen, Connecticut, 06520-8107, and Department of Chemistry, HaVerford College,
370 Lancaster AVe., HaVerford, PennsylVania, 19041
ReceiVed March 28, 2000
Abstract: The complex [(terpy)(H2O)MnIII(O)2MnIV(OH2)(terpy)](NO3)3 (terpy ) 2,2′:6,2′′-terpyridine) (1)
catalyzes O2 evolution from either KHSO5 (potassium oxone) or NaOCl. The reactions follow Michaelis-
Menten kinetics where Vmax ) 2420 ( 490 mol O2 (mol 1)-1 hr-1 and KM ) 53 ( 5 mM for oxone ([1] )
7.5 µM), and Vmax ) 6.5 ( 0.3 mol O2 (mol 1)-1 hr-1 and KM ) 39 ( 4 mM for hypochlorite ([1] ) 70 µM),
with first-order kinetics observed in 1 for both oxidants. A mechanism is proposed having a preequilibrium
between 1 and HSO5 or OCl-, supported by the isolation and structural characterization of [(terpy)-
-
(SO4)MnIV(O)2MnIV(O4S)(terpy)] (2). Isotope-labeling studies using H218O and KHS16O5 show that O2 evolution
proceeds via an intermediate that can exchange with water, where Raman spectroscopy has been used to confirm
-
that the active oxygen of HSO5 is nonexchanging (t1/2 . 1 h). The amount of label incorporated into O2 is
dependent on the relative concentrations of oxone and 1. 32O2:34O2:36O2 is 91.9 ( 0.3:7.6 ( 0.3:0.51 ( 0.48,
when [HSO5-] ) 50 mM (0.5 mM 1), and 49 ( 21:39 ( 15:12 ( 6 when [HSO5-] ) 15 mM (0.75 mM 1).
The rate-limiting step of O2 evolution is proposed to be formation of a formally MnVdO moiety which could
then competitively react with either oxone or water/hydroxide to produce O2. These results show that 1 serves
as a functional model for photosynthetic water oxidation.
Introduction
involve the reaction of a second molecule of water with the
terminal oxo ligand.1-3,12,13
The oxygen-evolving complex (OEC) of photosystem II (PS
II) contains a µ-oxo-bridged manganese tetramer that can carry
out the four-electron oxidation of water to O2.1-8 The key
intermediate has been postulated as a MndO species formed
by the abstraction of H-atoms from a water ligated to manganese
by a tyrosyl radical.2-4,9-11 O-O bond formation could then
Bioinorganic model chemistry has provided important insights
into the mechanism of photosynthetic water oxidation, in
particular by inspiring the aforementioned idea that a MndO
species is involved.7,13-15 Several studies give credence to this
proposal. [(bpy)2(H2O)RuIII(O)RuIII(OH2)(bpy)2]4+ (bpy ) 2,2′-
bipyridine)16-18 and related complexes have been shown to
oxidize water via a RudO intermediate.19 A di-manganese di-
porphyrin complex has been reported that can oxidize water
electrochemically, with a MndO species proposed to be the
reactive intermediate.20 Such an assignment is supported by
mechanistic studies on Mn-porphyrin oxidation catalysts where
this intermediate has been observed,21,22 and there are a series
of structurally characterized, albeit relatively unreactive, MndO
* To whom correspondence should be addressed: Telephone: (203)-
432-5202 (GWB), (203)-432-3925 (R.H.C.). Fax: (203)-432-6144 (G.W.B.,
R.H.C.). E-mail: gary.brudvig@yale.edu, robert.crabtree@yale.edu.
‡ Yale University.
§ Haverford College.
| Current address: Department of Chemistry, University of California
at Berkeley, Berkeley, California, 94704.
(1) Vrettos, J. S.; Limburg, J.; Brudvig, G. W. Biochim. Biophys. Acta
2001, 1503, 229-245.
(2) Limburg, J.; Brudvig, G. W.; Crabtree, R. H. In Biomimetic Oxidation
Chemistry; Meunier, B., Ed.; World Scientific: Singapore, 2000; pp 509-
541.
(3) Limburg, J.; Szalai, V. A.; Brudvig, G. W. J. Chem. Soc., Dalton
Trans. 1999, 1353-1363.
(12) Siegbahn, P. E. M.; Crabtree, R. H. J. Am. Chem. Soc. 1995, 117,
7-127.
(13) Pecoraro, V. L.; Baldwin, M. J.; Caudle, M. T.; Hsieh, W.; Law,
N. A. Pure Appl. Chem. 1998, 70, 925-929.
(4) Britt, R. D. In Oxygenic Photosynthesis: The Light Reactions; Ort,
D. R., Yocum, C. F., Eds.; Advances in Photosynthesis, Vol. 4; Kluwer
Academic Publishers: Dordrecht, The Netherlands, 1996; pp 137-159.
(5) Debus, R. J. Biochim. Biophys. Acta 1992, 1102, 269-352.
(6) Penner-Hahn, J. E. Struct.Bonding 1998, 90, 1-36.
(7) Ru¨ttinger, W.; Dismukes, G. C. Chem. ReV. 1997, 97, 1-24.
(8) Yachandra, V. K.; Sauer, K.; Klein, M. P. Chem. ReV. 1996, 96,
2927-2950.
(14) Limburg, J.; Vrettos, J. S.; Liable-Sands, L. M.; Rheingold, A. L.;
Crabtree, R. H.; Brudvig, G. W. Science 1999, 283, 1524-1527.
(15) Limburg, J.; Brudvig, G. W.; Crabtree, R. H. J. Am. Chem. Soc.
1997, 119, 2761-2762.
(16) Geselowitz, D.; Meyer, T. J. Inorg. Chem. 1990, 29, 3894-3896.
(17) Gersten, S. W.; Samuels, G. J.; Meyer, T. J. J. Am. Chem. Soc.
1982, 104, 4029-4030.
(18) Binstead, R. A.; Chronister, C. W.; Ni, J. F.; Hartshorn, C. M.;
Meyer, T. J. J. Am. Chem. Soc. 2000, 122, 8464-8473.
(19) Hurst, J. K.; Zhou, J.; Lei, Y. Inorg. Chem. 1992, 31, 1010-1017.
(20) Naruta, Y.; Sasayama, M.; Sasaki, T. Angew. Chem., Int. Ed. Engl.
1994, 33, 1839-1841.
(9) Hoganson, C. W.; Lydakis-Simantiris, N.; Tang, X.-S.; Tommos, C.;
Warncke, K.; Babcock, G. T.; Diner, B. A.; McCracken, J.; Styring, S.
Photosynth. Res. 1995, 46, 177-184.
(10) Hoganson, C. W.; Babcock, G. T. Science 1997, 277, 1953-1956.
(11) Messinger, J.; Badger, M.; Wydrzynski, T. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 3209-3213.
(21) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1988, 110, 8628-
8638.
10.1021/ja001090a CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/28/2000

424 J. Am. Chem. Soc., Vol. 123, No. 3, 2001
Limburg et al.
Table 1. Crystal Data and Structure Refinement for 2‚6H2O
complexes.23-28 These studies give support to the idea that a
MndO species is involved in the water-oxidation chemistry of
the OEC.
2‚6H2O
chemical formula
formula weight
crystal system
space group (no.)
a (Å)
b (Å)
c (Å)
C30H34N6O16S2Mn2
908.63
It has been proposed that a MndO intermediate can exchange
with H2O via an oxo/hydroxo tautomeric mechanism.29,30 In this
way, incorporation of isotopically labeled H218O into oxidized
organic products has been used as a mechanistic probe when
studying Mn-porphyrin oxidation catalysis.21,22,29-31 An im-
portant caveat of this method is that the primary oxidant must
be slow to exchange with free water relative to its rate of
reaction with the catalyst.32
monoclinic
C2/c (15)
27.091(4)
9.036(2)
18.281(5)
127.42(2)
3554(1)
-90.0
â (deg)
V (Å3)
T (°C)
λ (Å)
0.71069
1.675
4
δ
calc (g cm-3
)
Z
Recently, we reported a reaction of OCl- with the complex
[(terpy)(H2O)MnIII(O)2MnIV(OH2)(terpy)]3+ (1) (terpy ) 2,2′:
6,2′′-terpyridine) as a catalyst, that resulted in the catalytic
formation of O2, but 1 was deactivated by the irreversible
formation of permanganate.14 This was the first report of a di-
µ-oxo dimanganese complex, a structural model for the man-
ganese complex in the OEC, that could carry out catalytic O-O
bond formation. The mechanism proposed for this functional
model for the OEC involves formation of O2 by reaction of a
formally MnVdO intermediate with outer sphere water/
hydroxide. Complementary to the use of OCl-, we have also
F(000)
1816.00
9.11
abs coeff, cm-1
no. of reflecns collcd/unique
no. of observations (I > 3σ(Ι))
no. of refined params
largest peak/hole, e Å-1
Ra
3955/3834
1323
253
0.53/-0.45
0.055
b
Rw
0.056
goodness of fit indicator
1.58
2
a R ) Σ||Fo| - |Fc||/Σ|Fo|. b Rw ) [(Σw(|Fo| - |Fc|)2/ΣwFo )]1/2
.
-
studied the O2-evolving reaction of HSO5 with Mn(II) using
Experimental Section
dpa (dpa ) dipicolinate)15,33 and terpy, where mixed-valence
di-µ-oxo dimers formed in situ.15 Control experiments using
Lewis acids (Zn2+ and Al3+), where no reactivity was observed,
showed that the O2-evolving reaction of HSO5- did not involve
Lewis acid-promoted hydrolysis to form H2O2 followed by its
dismutation.15
All solutions were prepared using doubly deionized water. H218O
(85.5 atom %) was purchased from ICON Stable Isotopes. All other
chemicals were purchased from Aldrich and used without further
purification. NaOCl (10-13%) and KHSO5 were standardized using
iodometric titrations. All oxone solutions were made up in acetate buffer
(0.23 M HOAc/OAc- , pH ) 4.5), and all hypochlorite solutions were
adjusted to pH ) 8.6 using concentrated HNO3. 1 was synthesized as
reported previously.14,34
Synthesis of [(terpy)(SO4)MnIV(O)2MnIV(O4S)(terpy)]. To a stir-
ring solution of MnSO4 (73 mg, 0.429 mmol) and 2,2′:6′,2′′-terpyridine
(100 mg, 0.429 mmol) in nitric acid (10 mL, pH ) 2) was added
potassium peroxomonosulfate (111 mg, 0.322 mmol) in nitric acid (3
mL, pH ) 2). The yellow solution turned green and then brown/red.
After stirring for 10 min, the mixture was transferred to a beaker and
left to evaporate at room temperature. After 48 h, crystals of 2 had
formed as dark red plates, and these were filtered on a frit and washed
with a small volume of iced water and then copious acetonitrile. Yield
) 84 mg (44%). Analysis C30H22N6O16S2Mn2, 2‚3H2O calculated: C
42.15%, H 3.28%, N 9.84%, S 7.50%; found: C 42.02%, H 3.29%, N
9.84%, S 7.50%; IR (KBr): 3460 (s, br), 1605 (m), 1571 (m), 1477
(m), 1193 (m), 1179 (m), 1026 (s), 791 (m), 687 (m), 662 (m). The
UV/vis spectrum of 2 has been published previously.14
In this paper, we compare the mechanisms of the catalytic
O2-evolving reactions between 1 and the oxygen-atom transfer
reagents HSO5- (potassium oxone) and OCl- (both referred to
as XO). The similarity of the kinetics observed when using
oxone versus hypochlorite suggests that the two oxidants employ
analogous mechanisms in their reactions with 1. We propose
that the key intermediate is a MndO species, as suggested in
our previous reports. Kinetic and isotopic labeling studies
support a mechanism in which a MndO intermediate forms
following the reversible binding of XO to 1. We have isolated
and structurally characterized the complex [(terpy)(SO4)MnIV(O)2-
MnIV(O4S)(terpy)] (2)14 as a model for the binding of oxone to
1. The results reported here support our assignment of 1 as a
functional model of photosynthetic water oxidation.14
Crystal Structure Determination of 2. A crystal of 2 of dimen-
sions 0.09 × 0.17 × 0.22 mm was taken directly from the reaction
mixture. Diffraction data were collected on an Enraf-Nonius CAD-4
diffractometer with graphite monochromated Mo KR radiation, and the
crystallographic data are summarized in Table 1.
The structure of 2 was solved by heavy-atom Patterson methods,35
and expanded using Fourier techniques. The non-hydrogen atoms were
refined anisotropically, and hydrogen atoms were included but not
refined.36 All calculations were performed using teXsan.37 The full
details of the X-ray structure determination can be found in the
Supporting Information.
(22) Groves, J. T.; Lee, J.; Marla, S. S. J. Am. Chem. Soc. 1997, 119,
6269-6273.
(23) Collins, T. J.; Gordon-Wylie, S. W. J. Am. Chem. Soc. 1989, 111,
4511-4513.
(24) Collins, T. J.; Powell, R. D.; Slebodnick, C.; Uffelman, E. S. J.
Am. Chem. Soc. 1990, 112, 899-901.
(25) Collins, T. J. Acc. Chem. Res. 1994, 27, 279-285.
(26) Miller, C. G.; Gordon-Wylie, S. W.; Horwitz, C. P.; Strazisar, S.
A.; Peraino, D. K.; Clark, G. R.; Weintraub, S. T.; Collins, T. J. J. Am.
Chem. Soc. 1998, 120, 11540-11541.
(27) MacDonnell, F. M.; Fackler, N. L. P.; Stern, C.; O’Halloran, T. V.
J. Am. Chem. Soc. 1994, 116, 7431-7432.
(28) Shirin, Z.; Hammes, B. S.; Young, V. G., Jr.; Borovik, A. S. J. Am.
Chem. Soc. 2000, 122, 1836-1837.
O2 Evolution. Initial rates of O2 evolution were measured with a
YSI Standard Oxygen Probe (Clark electrode) attached to a Cole-Palmer
chart recorder. In a typical experiment an aliquot (10-50 µL) of a
solution of 1 (1-5 mM) was added to a 4 mL solution of either oxone
(29) Bernadou, J.; Fabiano, A.-S.; Robert, A.; Meunier, B. J. Am. Chem.
Soc. 1994, 116, 9375-9376.
(30) Bernadou, J.; Meunier, B. J. Chem. Soc., Chem. Commun. 1998,
2167-2173.
(31) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1987, 109, 3812-
(34) Collomb, M.-N.; Deronzier, A.; Richardot, J.; Pecaut, J. New J.
Chem 1999, 23, 351-353.
(35) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.;
Garcia-Granda, S.; Gould, R. O.; Smits, J. M. M.; Smykalla, C. PATTY:
The DIRDIF program system, Techical Report of the Crystallography
Laboratory; University of Nijmegen, 1992.
3814.
(32) Nam, W.; Valentine, J. S. J. Am. Chem. Soc. 1993, 115, 1772-
1778.
(33) Limburg, J.; Crabtree, R. H.; Brudvig, G. W. Inorg. Chim. Acta
2000, 297, 301-306.

Catalysis of O2 EVolution
J. Am. Chem. Soc., Vol. 123, No. 3, 2001 425
(5-200 mM) or hypochlorite (10-100 mM) which was being stirred
beneath the electrode. All rates were measured at 25 °C using the
method of initial rates by taking the recorded slope from t ) 0 to t )
30 s. Each rate reported is an average of at least three runs.
When higher concentrations of 1 (>0.1 mM) were added to a solution
of oxone, the amount of O2 evolved was too large to measure using a
Clark electrode because bubbles were formed. In these cases, the volume
of gas produced was measured by running the reaction in a 25 mL
sidearm Erlenmeyer flask with a 1 m × 2.75 mm i.d. capillary tube
attached to the sidearm, and a stopcock connected to the top. The typical
volume of a solution was 5 mL, and the concentrations of oxone and
1 were as in the mass spectrometry experiments (see below). 1 was
introduced to a stirring oxone solution through the stopcock, which
was immediately closed. The volume of O2 produced was measured
by the displacement of a bead of glycerol present in the capillary tube.
Raman Measurements on the Rate of Exchange of the Active
Oxygens in OCl- and HSO5 with Water. Raman spectra were
-
obtained with a SPEX Raman 500 equipped with a liquid nitrogen
cooled CCD detector (ISA CCD-2000) using a Spectra Physics model
161B Ar-ion laser operating at 488 nm. Incident laser powers were
between 8 and 10 mW; only the Raman intensities and no spectral
features changed as a function of power. Rayleigh scatter from the 488
nm laser line was rejected by a custom-built Kaiser Optical Systems
(Ann Arbor, MI) holographic notch filter. The detection geometry was
90° using a quartz cuvette. The exchange experiments were carried
Figure 1. ORTEP diagram of 2.
Table 2: Selected Bond Distances (Å) and Angles (deg) for
2‚6H2O
-
out by injecting 250 µL of either OCl- (0.5 M) or HSO5 (0.5 M)
reagent into the cuvette containing 250 µL of water (16O or 18O, final
18O composition was 42.5 atom %) immediately before accumulating
spectra. All experiments were reproduced at least twice. Spectra were
fit to the sum of a baseline and Gaussian curves using Igor Pro (v.
3.14, WaveMetrics, Inc.).
Mass Spectrometry-18O Labeling. An aqueous solution of 1 (50
µL, 5 mM in 0.1 M KNO3, pH ) 4.5)34 and an aqueous solution of
KHS16O5 (either 200 µL or 40 µL, 0.125 M) were placed in separate
bulbs of a two-bulb flask (each bulb had a volume of ∼15 mL). H218O
(250 µL, 85.5%) was then added to the solution of 1, and both solutions
were degassed via three freeze-pump-thaw cycles. The two solutions
were then mixed under a 10 mTorr atmosphere of He, and O2 evolution
was allowed to occur for 1 h, at which time the solution was frozen
and the headspace gas drawn up into an evacuated 150-mL metal
cylinder. The gas was then analyzed using a HP 5989A mass
spectrometer. Two measurements were taken at the higher oxone
concentration and three at the lower.
Mn(1)-Mn(1)
Mn(1)-O(1)
Mn(1)-N(1)
Mn(1)-N(3)
2.769(3)
1.794(6)
2.009(7)
2.035(7)
Mn(1)-O(1)
Mn(1)-O(2)
Mn(1)-N(2)
1.833(6)
1.890(6)
1.994(8)
Mn(1)-Mn(1)-O(1)
39.7(2) Mn(1)-Mn(1)-O(1)
40.7(2)
99.9(2)
Mn(1)-Mn(1)-O(2) 132.5(2) Mn(1)-Mn(1)-N(1)
Mn(1)-Mn(1)-N(2) 128.0(2) Mn(1)-Mn(1)-N(3) 100.3(2)
O(1)-Mn(1)-O(1)
O(1)-Mn(1)-N(1)
O(1)-Mn(1)-N(3)
O(1)-Mn(1)-N(1)
O(1)-Mn(1)-N(3)
O(2)-Mn(1)-N(2)
N(1)-Mn(1)-N(2)
N(2)-Mn(1)-N(3)
Mn(1)-O(2)-S(1)
80.5(3) O(1)-Mn(1)-O(2)
93.0(3) O(1)-Mn(1)-N(2)
93.9(3) O(1)-Mn(1)-O(2)
102.3(3) O(1)-Mn(1)-N(2)
101.9(3) O(2)-Mn(1)-N(1)
99.5(3) O(2)-Mn(1)-N(3)
78.0(3) N(1)-Mn(1)-N(3)
79.0(4) Mn(1)-O(1)- -Mn(1)
133.4(4) Mn(1)-N(1)-C(1)
172.2(3)
88.3(3)
91.7(3)
168.7(3)
88.2(3)
88.0(3)
155.7(3)
99.5(3)
122.5(7)
The final 18O content of the water was calculated from the ratio of
the 46CO2: 48CO2 peaks, as the oxygens in CO2 exchange rapidly (t1/2
< 10 s) with solvent water. The amount of label incorporation into the
O2 evolved was then normalized to the 18O content. The yield of 32O2
was adjusted by subtracting the background amount of 32O2 based on
the amount of 28N2 in each run to correct for incomplete degassing or
background air. This correction was less than 3% for both of the runs
at higher oxone concentration, and one of the lower oxone runs, but
higher (∼70%) for the other two lower oxone runs. These corrections
are reflected in the reported errors for the relative yields of 32O2, 34O2,
and 36O2.
of mixed-valence dimer was quantified by measuring the amplitude of
the peak near 3300 G and comparing this value with an authentic sample
of 1.
UV-Visible Spectroscopy. UV/vis spectra were recorded on a HP
8452A Diode Array spectrophotometer. Reactions were run at 25 °C.
A 1-mm path length cuvette was used to follow the spectra at the higher
concentrations of oxone and 1 used for the mass spectrometry
experiments, and a 1-cm path length cuvette used for the spectra
collected under the Clark-electrode conditions.
Results
Oxidizing Equivalents. The oxidizing equivalents of oxone remain-
ing as a function of time were measured using the same initial
concentrations of oxone and 1 as in the mass spectrometry experiments.
Aliquots of the reaction solution were quenched by injected into
solutions of KI and the amount of I2/I3 produced was determined by
titration with a standardized solution of sodium thiosulfate.
EPR Spectroscopy. Spectra were collected at 10 K on frozen
samples with an X-band Varian E-9 spectrometer equipped with an
Oxford ESR-900 liquid helium cryostat. Typical instrument settings
were the following: modulation amplitude ) 20 G, modulation
frequency ) 100 kHz, and microwave power ) 0.5 mW. The amount
Synthesis and Structure of [(terpy)(SO4)MnIV(O)2MnIV-
(O4S)(terpy)] (2). In prior work, we had only been able to
determine the structure of the mixed-valence III/IV form of the
catalyst. Now, we find that the use of sulfate as a counterion
allows the IV/IV form to be crystallographically characterized.
Figure 1 shows the crystal structure of 2, with bond distances
and bond angles summarized in Table 2. The structure is very
similar to that of the III/IV species and suggests that the cluster
retains its identity on redox cycling. 2 forms by disproportion-
ation from a solution containing the mixed-valence dimer, 1,
and SO42-, and 2 was originally isolated from an attempt to
make the mixed-valence dimer. The ligation of one doubly
-
(36) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.; de
Gelder, R.; Israel, R.; Smits, J. M. M. DIRDIF-94: The DIRDIF-94 program
system, Technical Report of the Crystallography Laboratory; University of
Njimegen, 1994.
(37) teXsan: Crystal Structure Analysis Package; Molecular Structure
Corporation, 1992.
2-
charged SO4 group to each Mn has the effect of stabilizing
the MnIV oxidation state, and 2 has no net charge. The Mn ions
are separated by the ∼2.7 Å (Table 2) distance characteristic

426 J. Am. Chem. Soc., Vol. 123, No. 3, 2001
Limburg et al.
Figure 2. Plots showing the effect of varying [1] on the initial rates
Figure 3. (A) Effect of varying [HSO5-] (0.23 M HOAc/OAc-, pH
) 4.5) on the initial rates of O2 evolution ([1] ) 7.5 µM). (B) Effect
of varying [OCl-] (pH ) 8.6) in the initial rates of O2 evolution ([1]
) 70 µM). Both plots are fitted to the Michaelis-Menten equation.
Errors reported are one standard deviation from the average of all runs
carried out at each concentration.
of O2 evolution; (A) HSO5 (30 mM in 0.23 M HOAc/OAc-, pH )
-
4.5), (B) OCl- (70 mM, pH ) 8.6). Errors reported are one standard
deviation from the average of all runs carried out at each concentration.
of di-µ-oxo Mn dimers,38,39 and which is virtually invariant
between the MnIIIMnIV and MnIV complexes (d(Mn-Mn) is
2
2.723 Å in 114,34 vs 2.769 Å in 2). Unlike the III/IV dimer, 2 is
EPR-silent, as expected for a strongly coupled MnIV dimer.
Furthermore, the Mn(O)2Mn core vibrational frequencies down-
shift to lower energies on going from the mixed-valence14,34 to
the IV/IV dimer. The IR bands appear at 703 and 687 cm-1 in
1 and 2, respectively, while in the Raman spectrum the
complicated set of peaks over the range of 698, 671, 630, and
597 cm-1 downshifts to 652, 552, and 540 cm-1 due to the
changes in both oxidation state and symmetry. These shifts are
typical for di-µ-oxo Mn dimers.38,39
Table 3: Kinetics Data Calculated from Figures 2 and 3
HSO5
OCl-
-
order in 1a
0.95 ( 0.06
0.99 ( 0.02
53 ( 5
1.00 ( 0.06
0.70 ( 0.01
36 ( 4
order in XOb
K
Mc(mM)
Vmaxc (mol O2 (mol 1)-1 hr-1
)
2420 ( 490
6.5 ( 0.3
a Calculated from a log rate vs log concentration plot of the data
from Figure 2. b Calculated from a log rate vs log concentration plot
of the data from Figure 3. c Calculated from the Michaelis-Menten
fit of the data from Figure 3.
2 is only sparingly soluble in water, and dissolution is
relatively slow when compared to that of 1, presumably because
the SO42- ligands must be displaced by water. 2, being a MnIV
complex, would be expected to be relatively inert to ligand
exchange as the Mn ions are high-spin d3. The coordination of
-
The Vmax for O2 evolution with HSO5 is nearly 400 times
faster than for OCl- (Table 3). The difference in rates may be
due either to inherently greater reactivity on the part of oxone
or may reflect the difference in the pH at which the two oxidants
are used (pH ) 4.5 for HSO5- and 8.6 for OCl-). When dpa is
used as a ligand in the oxone system, the O2-evolving reaction
is considerably slower at higher pH15 and is accompanied by
an increase in the rate of permanganate formation.14,15,33 Oxone
cannot be used at pH > 7 as it is decomposes directly to O2 in
basic solution.
2-
SO4 provides support for the idea that the aqua ligand sites
in 1 are exchangeable in solution as 2 forms from a solution of
1, and this exchange is partly responsible for the catalytic
activity 1.
Initial Rates of O2 Evolution. We have followed the effect
of varying both 1 (Figure 2) and XO (Figure 3) on the initial
rate of O2 evolution, using pseudo-first-order conditions in
substrate, and find results that suggest that MndO formation is
rate-limiting for both OCl- and HSO5-. The results of these
experiments are summarized in Table 3. O2 evolution is first-
order in 1 and displays saturation kinetics when varying XO.
These results are consistent with a mechanism that involves a
preequilibrium between 1 and XO,33 with the reaction of the
1-XO complex so formed being the rate-limiting step.
Rates of Exchange of the Active Oxygens in KHS16O5 and
Na16OCl with H218O. It is of critical importance to establish
whether the oxygens in the O2 evolved from the reaction of 1
and XO arise from water or from the oxidant. This can be
achieved using 18O-isotope labeling of the water if the active
O-atom of XO is slow (t1/2 > 1 h) to exchange. Raman
spectroscopy was chosen as a convenient and facile method
for direct measurement of the exchange kinetics of both
hypochlorite and oxone; this is the first application of this
method.
(38) Manchanda, R.; Brudvig, G. W.; Crabtree, R. H. Coord. Chem. ReV.
1995, 144, 1-38.
(39) Thorp, H. H.; Brudvig, G. W. New J. Chem. 1991, 15, 479-490.

Catalysis of O2 EVolution
J. Am. Chem. Soc., Vol. 123, No. 3, 2001 427
Figure 5. Raman spectra accumulated over the first 20 min after
mixing 250 µL of HSO5- (0.5 M, pH ) 4.5) with 250 µL of (A) H216O
and (B) H218O. Each spectrum is an average of six scans integrated for
5 s with an incident laser power of 8 mW.
Figure 4. Raman spectra accumulated over the first 30 s after mixing
250 µL of OCl- (0.5 M, pH ) 8.6) with 250 µL of (A) H216O and (B)
H218O. The shift of the 18OCl- peak from the parent 16OCl- peak is
indicated by the arrows. Each spectrum is an average of six scans
integrated for 5 s with an incident laser power of 8 mW.
Table 4: Results of the 18O-Isotope-Labeling Experiments
Showing the Percentage Yields of 32O2, 34O2, and 36O2 from the
-
Reaction between 1 and HSO5
Hypochlorite/Solvent Exchange. Using indirect methods
we14 and others40,41 had found OCl- to be an apparently slow
exchanger, but a direct measurement using 17O NMR showed
that in reality OCl- exchanges rapidly (t1/2 < 20 s).42 These
contrasting results led us to reexamine the exchange kinetics
of OCl-.43
The Raman spectra of hypochlorite recorded over the first
30 s after mixing with H216O and H218O are shown in Figure 4.
In Figure 4A, besides the 16OCl- peak at 708 cm-1, the ν1
conditions
50 mM HSO5-, 0.5 mM 1c
15 mM HSO5-, 0.75 mM 1d
32O2a,b (%) 34O2a (%)
91.9 ( 0.3 7.6 ( 0.3 0.51 ( 0.48
49 ( 21 39 ( 15 12 ( 6
36O2a (%)
a Normalized to the concentration of H218O; the errors are one
standard deviation. b 32O2 yields corrected for background air using 28N2.
c Average of two runs. d Average of three runs.
trometry to exchange slowly with water,47 which our results
support. The peroxy mode of oxone appears at 878 cm-1 (ν6)
and the SdO stretch at 1055 cm-1 (ν5).48-50 As shown in Figure
5B, there is no 18O incorporation after 20 min, nor was any
exchange observed even after more than 1 h (not shown). This
is the first direct observation that oxone is slow to exchange
with water, confirming the previous conclusions based on
indirect studies of organic oxidation products.29,31,45
18O Labeling. As oxone is slow to exchange with solvent
water, the source of the O-atoms in the O2 evolved from the
catalytic reaction, H2O versus HSO5-, could be ascertained using
18O-labeled water. The results are summarized in Table 4.
Initially the reaction was run under a relatively high concentra-
tion of oxone to maximize the yield of O2. Under these
conditions, about 8% of the O2 evolved contained a single label
(34O2), and only a very small amount of doubly labeled product
(36O2) was observed. Lowering the oxone to 1 ratio by 5-fold
caused a dramatic increase of labeled O2 to approximately 50%,
with a significant amount of 36O2 observed. These results are
consistent with the formation of a reactive intermediate that can
exchange with solvent water and suggest that there is more than
one pathway for O-O bond formation. Most importantly, these
results cannot be explained by a mechanism where 18O-exchange
involves free oxone, and so all the 36O2 and some of the 34O2
must result from water oxidation.
-
-
modes of NO3 (1045 cm-1) and ClO3 (930 cm-1) are
observed.44 Figure 4B shows that within the first 30 s after
mixing with H218O, the exchange of the O-atom in OCl- with
solvent is complete. The 18OCl- peak downshifts 23 cm-1 to
685 cm-1 44
as indicated by the arrows. The final isotopic
,
composition of the water is 42.5% 18O/57.5% 16O, which is
reflected in the ratio of the areas of the 18OCl-/16OCl- peaks
-
-
(0.74 expected, 0.77 calculated). The NO3 and ClO3 peaks
do not shift over the first 10 min (data not shown).
These data show that the O-atom exchange of OCl- with
water is indeed fast, as previously reported.42,43 Equilibrium is
reached within the first 30 s, indicating that the half-life must
be t1/2 < 10 s. This makes OCl- unsuitable for use in the 18O-
labeling experiments.
Oxone/Solvent Exchange. Having confirmed that hypo-
chlorite exchanges rapidly, oxone became the obvious choice
of oxidant for the isotope-labeling experiments. Peroxides (e.g.,
H2O2 or m-chloroperoxybenzoic acid) are intrinsically inert to
exchange with water,32 and oxone has been shown to follow
this trend.29,31,45 However, there have been no studies using a
direct spectroscopic measurement. Figure 5, A and B, show the
spectra of oxone recorded over the first 20 min after mixing
with H216O and H218O, respectively. The peak at 980 cm-1 arises
from the symmetric stretch of SO42-, which is present in
potassium oxone.44,46 Sulfate has been shown by mass spec-
O2 Evolution Under Limiting Oxone Conditions. To
understand why 18O incorporation into O2 was greatest under
limiting oxone conditions, we have studied the kinetics of the
oxone reaction under these conditions. Figure 6 shows the
volume of O2 gas evolved as a function of time (in both H2O
and D2O). The mass spectrometry data on the reaction show
that O2 is the only gaseous product. Inspection of Figure 6
(40) Anwar, M.; Taube, H. J. Am. Chem. Soc. 1958, 80, 1073-1077.
(41) Meunier, B.; Guilmet, E.; De Carvahlo, M.-E.; Poilblanc, R. J. Am.
Chem. Soc. 1984, 106, 6668-6676.
(42) Johnson, D. W.; Margerum, D. W. Inorg. Chem. 1991, 30, 4845-
4851.
(43) While this paper was being written, Meunier and co-workers
published a correction to their original report (ref 41) that LiOCl exchanges
slowly (B. Meunier et al., J. Am. Chem. Soc., 2000, 122, 2675). Meunier
suggests using peroxides, such as oxone, for unambiguous mechanistic
detection of a MndO intermediate with H218O.
(44) Irish, D. E.; Ozeki, T. Raman Spectroscopy of Inorganic Species in
Solution; Wiley: New York, 1991; Vol. 114.
(45) Robert, A.; Meunier, B. New J. Chem. 1988, 12, 885-896.
(46) Reedy, B. J.; Bettie, J. K.; Lowson, R. T. Spectrochim. Acta 1990,
46A, 1513-1519.
(47) Hoering, T. C.; Kennedy, J. W. J. Am. Chem. Soc. 1958, 79, 56-
60.
(48) Appleman, E. H.; Basile, L. J.; Kim, H.; Ferraro, J. R. Spectrochim.
Acta 1985, 41A, 1295-1300.
(49) Maslowski, E., Jr. In Vibrational Spectra of Organometallic
Compounds; Wiley: New York, 1977; p 422.
(50) Schugart, K. A.; Noblet, J. A. J. Mol. Struct. (THEOCHEM) 1994,
304, 13-26.

428 J. Am. Chem. Soc., Vol. 123, No. 3, 2001
Limburg et al.
Figure 7. Plot of the amount of mixed-valence di-µ-oxo manganese
dimer in solution (determined by EPR) as a function of time during
catalysis with oxone. [1] ) 0.75 mM, [HSO5-] ) 15 mM (open circles)
and [1] ) 0.5 mM, [HSO5-] ) 50 mM (filled squares).
almost fully consumed (Table 5).51 Table 5 shows the number
of oxidizing equivalents remaining in solution as a function of
time. Importantly, O2 evolution can account for g80% of the
oxone consumed under both sets of conditions (Table 5), and
so is the major reaction in solution. The reappearance of 1 upon
consumption of oxone, coupled with O2 evolution being the
dominant reaction, shows that ligand oxidation is not a major
factor.
Permanganate Formation. In the course of this study, we
have found that the degree of deactivation of the catalyst by
ligand loss and permanganate formation14,15,33 is highly depend-
ent on the relative concentrations of 1 and XO. When the ratio
of XO:1 is >500:1, most of 1 is converted to MnO4-, with
-
isosbestic behavior.14,15 However, MnO4 formation does not
contribute significantly to the initial rates (<2% MnO4- formed
in the first 30 s). On the other hand, no permanganate was
detected over the time-course of O2 evolution at the concentra-
tions of 1 and oxone used for the mass spectrometry studies,
where HSO5-:1 is <100:1. This indicates that permanganate
formation can be suppressed if bimolecular reactions of the Mn
dimer occur instead. Furthermore, there is a deceleration of the
Figure 6. Plots showing turnovers vs time for the O2-evolving reaction
between HSO5 (0.23 M HOAc/OAc-, pH ) 4.5) and 1; (A) [1] )
-
0.75 mM, [HSO5-] ) 15 mM in H2O (filled squares) and D2O (open
circles); (B) [1] ) 0.5 mM, [HSO5-] ) 50 mM in H2O (filled squares)
and D2O (open circles).
-
rates of both O2 evolution and MnO4 formation after just a
reveals that the reactions are essentially complete in the first
10 min. The isotope effects were measured using the slope over
the first few turnovers while still under pseudo-first-order
conditions for oxone. When the mole ratio of HSO5-:1 was 20:
1, kH/kD ) 1.71 ( 0.2, and when the ratio was 100:1, kH/kD )
1.64 ( 0.02. Although the 18O-labeling experiments showed
that, under these conditions, the O atoms in O2 were derived
primarily from water and oxone, respectively, the similar H/D
isotope effects suggest that O2 evolution has the same rate-
limiting step under the two regimes. This indicates that the
reaction pathway (water oxidation vs oxone/oxone coupling) is
decided by the fate of the reaction intermediate. The apparent
inverse isotope effect that appears after a few minutes is due to
the faster consumption of oxidant in H2O versus D2O such that
the reactions are no longer pseudo-first-order in oxone at later
times.
An important issue is the fate of the complex during and after
catalysis. We have addressed this issue by following the change
in net oxidizing equivalents as a function of O2 evolved and by
using EPR and UV/vis spectroscopies to look at the species in
solution during catalysis with HSO5-. When the III/IV dimer 1
is added to oxone, the solution turns darker green, the majority
of the EPR-signal (>95%) for the III/IV dimer disappears
(Figure 7), and there are no signals corresponding to monomeric
Mn(II), II/II, or II/III species. After 1 h (the time-scale of the
mass spectrometry experiment), 70% of the manganese is in
the form of a III/IV dimer under conditions where the oxone is
few turnovers at high XO:1 ratios (>500:1). We believe this
could be due to anation of the dimer, as observed in the related
[(bpy)2(H2O)RuIII(O)RuIII(OH2)(bpy)2]4+ system,16-18 or ligand
loss. We are currently investigating these observations in greater
detail, but will not consider the deactivating reactions further
in this paper.
Discussion
Figure 8 shows a proposed mechanism for the formation of
the reactive intermediate that is consistent with the experimental
Michaelis-Menten kinetics, where the XO initially binds
reversibly to 1. The complex so formed (3) is analogous to 2
where sulfate can be thought of as a model for HSO5- binding
to 1. The mechanism is based upon the III/IV dimer being the
reactive species, although other oxidation states cannot yet be
ruled out (discussed below). Once formed, 3 can react with XO
(51) If the solution is left for long periods (>1 h) after the oxidant has
been consumed, a brown precipitate of MnO2 forms. If additional oxone is
added, this solution can catalyze gas evolution at a rate much slower than
that observed with 1 (data not shown). This means that any manganese
complex that decomposes to MnO2 could, in theory, react through a
heterogeneous pathway. For catalysis with 1, MnO2 can be ruled out as the
catalyst for O2 formation at short reaction times because: (1) there is no
lag period corresponding to MnO2 formation, (2) the dimer stays intact
under conditions where the oxone is consumed (shown in Figure 7), (3) no
precipitate can be detected during the first 30 min of the reaction, and (4)
the reaction is fastest in the presence of 1 and slows after the precipitate
forms.

Catalysis of O2 EVolution
J. Am. Chem. Soc., Vol. 123, No. 3, 2001 429
Table 5: Oxidizing Equivalents in Solution Per Mole of 1
10 min after addition of 1
60 min after addition of 1
conditions
before addition of 1
O2 evolved
remaining oxidizing equiva
total
remaining oxidizing equiva
50 mM oxone, 0.5 mM 1
15 mM oxone, 0.75 mM 1
192
38.4
118 ( 2
28.6 ( 0.4
36 ( 2
4.4 ( 0.2
154
33.0
23 ( 1
2.6 ( 0.2
a Corrected for the reduction of Mn(IV) in solution to Mn(II) upon addition of KI.
Scheme 1. Possible Pathways for the Formation of Labeled
and Unlabeled O2 from the Reaction between 1 and
KHS16O5 in H218O as Summarized in Table 4a
Figure 8. Proposed mechanism for the reaction between 1 and oxygen-
atom transfer reagents, XO, based on findings in this paper and also
previous studies.14,15,33 In both cases, the rate-limiting step is formation
of 4 by oxidation of 1. While these mechanisms are mainly based on
-
observations when HSO5 is used as the oxidant, the kinetic data we
present in this study suggest that OCl- most likely reacts in an
analogous fashion. Left-hand pathway: attack of XO on the terminal
oxo ligand of 4. Right-hand pathway: attack of solvent H2O on the
terminal oxo ligand of 4.
a The starred oxygens represent 18O.
Scheme 1 shows mechanistic proposals that can account for
each of the three isotopomers observed by mass spectrometry.
32O2 could arise from the reaction of an unexchanged 16O-oxo
with oxone (Scheme 1a). The amount of unlabeled O2 depends
inversely on the concentration of oxone (Table 4), so that at
lower oxone concentrations the rate of exchange (eq 1) becomes
comparable to the rate of the oxone/oxo reaction. The 34O2 could
form from two pathways involving either attack of solvent water/
hydroxide on an unexchanged MndO (Scheme 1b(i)) or an
exchanged MndO reacting with free oxone (Scheme 1b(ii)).
Finally, 36O2 could arise from the reaction between solvent water
and an exchanged MndO (Scheme 1c).
Certain inferences can be drawn from the labeling results.
Multiple pathways must be invoked to account for the product
distribution (Figure 8 and Scheme 1). At higher concentrations
of HSO5- relative to that of 1, the oxone/oxo pathway dominates
(Figure 8, left-hand path) and is fast compared to exchange,
and thus the major product is 32O2 (Table 4). Again, a
mechanism involving reaction between 1-XO and XO in solution
can be ruled out on the basis of the kinetics data (Figure 3)
which show a first-order dependence on [XO]. However, as the
oxone concentration is lowered, both the exchange reaction (eq
1) and the water/oxo reactions (Figure 8, right-hand path)
become competitive with the oxone/oxo reaction and thus all
three possible isotopomers, that is 32O2, 34O2, and 36O2, are seen
(Table 4). O2 release from the reaction of 4 with solvent water
would result in a II/III dimer. As no such species is observed
by EPR, we presume that the II/III dimer intermediate is then
rapidly reoxidized to 1 by oxone.
to form an intermediate containing a terminal MndO (4) (Figure
8), and this represents the rate-limiting step in O2 evolution. It
is unlikely that O2 evolution involves a reaction between the
1-XO complex and XO in solution, as this would give a second-
order dependence of the initial rate on XO and would not explain
the origin of 36O2. Moreover, the kinetic data in Figure 3 cannot
be fit by a rate law which involves such a mechanism.
Evidence for a MndO intermediate is provided by the 18O-
labeling experiments with HS16O5-. The incorporation of 18O
from water into various organic products has been used in Mn-
catalyzed oxidations to infer the presence of MndO inter-
mediates.21,22,29-31 The terminal oxo in 4 originates from oxone
and thus would be 16O initially but could exchange with labeled
solvent via the mechanism shown in eq 1.
Equation 1 is clearly a simplified view as multiple protona-
tion/depronation steps would be involved in the exchange.29,30
The amount of label incorporated into the O2 evolved would
then depend on the relative rate of this exchange versus the
O-O bond-forming step. A further possibility is that the
exchange involves the bridging µ-oxo ligands and that they are
involved in O-O bond formation. Such a mechanism has been
proposed for the action of the OEC.8,52,53
(52) Halfen, J. A.; Mahapatra, S.; Wilkinson, E. C.; Kaderli, S.; Young,
V. G.; Que, L.; Zuberbuhler, A. D.; Tolman, W. B. Science 1996, 271,
1397-1400.
(53) Pecoraro, V. L.; Baldwin, M. J.; Gelasco, A. Chem. ReV. 1994, 94,
807-826.
The oxidation state of the Mn complex responsible for
catalysis is not yet known, and both the IV/IV and the III/IV
states may be active. With both OCl-14 and HSO5- as oxidant,

430 J. Am. Chem. Soc., Vol. 123, No. 3, 2001
Limburg et al.
there is initially formation of a large amount (>95%) of an EPR-
silent species that we attribute to a IV/IV dimer (probably
formed by electron exchange between IV/V and III/IV species).
Oxidation of a IV/IV dimer by oxone would most likely give a
V/V dimer, where both manganese would have a terminal oxo
ligand. As one of the terminal oxo ligands would have to come
from water, even an oxone/oxo pathway (Figure 8, left, and
Scheme 1a) would give labeled product. This is not seen, and
for this reason, we propose that with oxone as oxidant, at least
under the conditions of the 18O-labeling experiments, the III/
IV dimer is the active species. The observed D-isotope effect
may partly arise from a perturbation of the initial equilibrium
where D2O exchanges slower than H2O. In addition, oxidation
of 3 to 4 may be accompanied by deprotonation of an aqua-
ligand to minimize charge build-up on the complex, and this
reaction would be expected to show a D-isotope effect.14
O-O bond-forming step involves attack of solvent or oxone
on the oxo ligand. The kinetic studies show that 1 reacts with
HSO5- and OCl- by similar mechanisms. These results comple-
ment earlier work involving the reaction of 1 with OCl-14 and
provide support for the mechanism shown in Figure 8. Our aim,
now, is to try to isolate the reactive intermediates and to
characterize further the species present during catalysis. These
species may be analogous to those present in the OEC during
turnover and thus will provide key insights into the mechanism
of photosynthetic water oxidation.
Acknowledgment. We thank Professor Bernard Meunier for
discussions on 18O exchange, Mr. Michael Ranen for his
assistance with the Raman instrumentation, and Susan de Gala
for solving the crystal structure of 2. Funding was provided by
the National Institutes of Health GM32715 (G.W.B.) and by
the Kresge Foundation (J.dP.).
Conclusions
Supporting Information Available: Tables containing
detailed crystallographic data, atomic position parameters, bond
lengths and bond angles of 2 (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
The data presented in this report show that 1 can catalyze
water oxidation when HSO5- is used as a primary oxidant. The
proposed mechanism involves an intermediate that can exchange
with water and then react with water or oxone. We propose
that this intermediate contains a formally MnVdO and that the
JA001090A