Evidence for Cu-O2 intermediates in superoxide oxidations by biomimetic copper(II) complexes

Article abstract of DOI:10.1021/ja056741n

The mechanism by which [Cu^II(L)](OTf)₂ and [Cu ^IIN₃(L)](OTf) (L = TEPA: tris(2-pyridylethyl)amine or TMPA: tris(2-pyridylmethyl)amine; OTf = trifluoromethanesulfonate) react with superoxide (O₂.^-) to form [Cu^I(L)](OTf) and O₂ is described. Evidence for a CuO₂ intermediate is presented based on stopped-flow experiments and competitive oxygen ( ¹⁸O) kinetic isotope effects on the bimolecular reactions of ^16,16O₂.^- and ^18,16O ₂.^- (^16,16k/^18,16k). The ^16,16k/^18,16k fall within a narrow range from 0.9836 ± 0.0043 to 0.9886 ± 0.0078 for reactions of copper(II) complexes with different coordination geometries and redox potentials that span a 0.67 V range. The results are inconsistent with a mechanism that involves either rate-determining O₂.^- binding or one-step electron transfer. Rather a mechanism involving formation of a CuO₂ intermediate prior to the loss of O₂ in the rate-determining step is proposed. Calculations of similar inverse isotope effects, using stretching frequencies of CuO₂ adducts generated from copper(I) complexes and O₂, suggest that the intermediate has a superoxo structure. The use of ¹⁸O isotope effects to relate activated oxygen intermediates in enzymes to those derived from inorganic compounds is discussed.

Full text of DOI:10.1021/ja056741n

Published on Web 03/01/2006
Evidence for Cu-O2 Intermediates in Superoxide Oxidations
by Biomimetic Copper(II) Complexes
Valeriy V. Smirnov and Justine P. Roth*
Contribution from the Department of Chemistry, Johns Hopkins UniVersity,
3400 North Charles Street, Baltimore, Maryland 21218
Received October 12, 2005; E-mail: jproth@jhu.edu
Abstract: The mechanism by which [CuII(L)](OTf)2 and [CuIIN3(L)](OTf) (L ) TEPA: tris(2-pyridylethyl)-
•-
amine or TMPA: tris(2-pyridylmethyl)amine; OTf ) trifluoromethanesulfonate) react with superoxide (O2
)
to form [CuI(L)](OTf) and O2 is described. Evidence for a CuO2 intermediate is presented based on stopped-
flow experiments and competitive oxygen (18O) kinetic isotope effects on the bimolecular reactions of 16,16O2
•-
•-
and 18,16O2
(
16,16k/18,16k). The 16,16k/18,16k fall within a narrow range from 0.9836 ( 0.0043 to 0.9886 (
0.0078 for reactions of copper(II) complexes with different coordination geometries and redox potentials
that span a 0.67 V range. The results are inconsistent with a mechanism that involves either rate-determining
O2•- binding or one-step electron transfer. Rather a mechanism involving formation of a CuO2 intermediate
prior to the loss of O2 in the rate-determining step is proposed. Calculations of similar inverse isotope
effects, using stretching frequencies of CuO2 adducts generated from copper(I) complexes and O2, suggest
that the intermediate has a superoxo structure. The use of 18O isotope effects to relate activated oxygen
intermediates in enzymes to those derived from inorganic compounds is discussed.
Introduction
studies have provided compelling evidence for the mechanisms
of these reactions.12,16-19 In a variety of biological oxidations,
Metal ions are responsible for the acquisition of energy1 and
maintenance of homeostasis within the cell.2-4 During aerobic
respiration, chemical energy is harnessed from molecular oxygen
by cytochrome c oxidase while other transition metal-mediated
processes control the fluxes of the partially reduced O2-derived
products or reactiVe oxygen species (ROS).5 Superoxide ion
(O2•-) is the primary ROS produced from O2 and intracellular
reductants according to eq 1.6,7 Subsequent reactions afford
including DNA cleavage20,21 and enzymatic C-H oxidation,22-24
the generation of reactive intermediates upon coordination of
O2 to mononuclear copper(I) sites has been proposed.25 Whether
the intermediates are electronically and structurally analogous
•-
to those formed when O2 reacts with copper(II) centers has
not been understood.
The majority of information concerning the structures of
oxygen bound to a single copper center has come from studies
of inorganic compounds.31 Species with the formula 1Cu:1O2
have been synthesized by reacting O2 with copper(I) complexes
•
hydroperoxyl radical (HO2 ), peroxynitrous acid (HONO2) and
hydrogen peroxide (H2O2).5 These secondary ROS are believed
to be involved in cell signaling pathways8 as well as the
oxidative damage of cellular sugars, lipids, proteins and DNA.5,9
(13) Wu, A. J.; Penner-Hahn, J. E.; Pecoraro, V. L. Chem. ReV. 2004, 104,
903.
(14) Goldstein, S.; Lind, J.; Merenyi, G. Chem. ReV. 2005, 105, 2457.
(15) Herold, S.; Koppenol, W. H. Coord. Chem. ReV. 2005, 249, 499.
(16) Weinstein, J.; Bielski, B. H. J. J. Am. Chem. Soc. 1980, 102, 4916.
(17) Cabelli, D. E.; Bielski, B. H. J.; Holcman, J. J. Am. Chem. Soc. 1987, 109,
3665.
(18) Fee, J. A.; Bull, C. J. Biol. Chem. 1986, 261, 13000.
(19) Boccini, F.; Herold, S. Biochemistry 2004, 43, 16393.
(20) Cowan, J. A. Curr. Opin. Chem. Biol. 2001, 5, 634.
(21) Chen, C.-h. B.; Milne, L.; Landgraf, R.; Perrin, D. M.; Sigman, D. S.
ChemBioChem 2001, 2, 735.
(22) Chen, P.; Solomon, E. I. J. Am. Chem. Soc. 2004, 126, 4991.
(23) Evans, J. P.; Ahn, K.; Klinman, J. P. J. Biol. Chem. 2003, 278, 49691.
(24) Klinman, J. P. Chem. ReV. 1996, 96, 2541.
(25) The formation of oxidizing species upon reductive activation of O2 at
dicopper sites has been discussed for tyrosinase (ref 26) and particulate
methane monooxygenase (ref 27). The reaction is well established in model
systems (refs 28-30).
(26) Solomon, E. I.; Baldwin, M. J.; Lowery, M. D. Chem. ReV. 1992, 92, 521.
(27) Lieberman, R. L.; Rosenzweig, A. C. Nature 2005, 434, 177.
(28) Lewis, E. A.; Tolman, W. B. Chem. ReV. 2004, 104, 1047.
(29) Itoh, S.; Fukuzumi, S. Bull. Chem. Soc. Jpn. 2002, 75, 2081.
(30) Mirica, L. M.; Vance, M.; Rudd, D. J.; Hedman, B.; Hodgson, K. O.;
Solomon, E. I.; Stack, T. D. P. Science 2005, 308, 1890.
(31) Cramer, C. J.; Tolman, W. B.; Theopold, K. H.; Rheingold, A. L. Proc.
Natl. Acad. Sci. U.S.A. 2003, 100, 3635.
O2 + 1e- f O2
(1)
•-
Transition metal ions in a variety of ligand environments have
been implicated in the formation10,11 as well as the scavenging
of primary6,7 and secondary ROS.12-15 Yet only a handful of
(1) Babcock, G. T.; Wikstrom, M. Nature 1992, 356, 301.
(2) Luk, E.; Jensen, L. T.; Culotta, V. C. J. Biol. Inorg. Chem. 2003, 8, 803.
(3) O’Halloran, T. V.; Culotta, V. C. J. Biol. Chem. 2000, 275, 25057.
(4) Valentine, J. S.; Gralla, E. B. Science 1997, 278, 817.
(5) Halliwell, B. Am. J. Med. 1991, 91 suppl 3C, 14S.
(6) Miller, A.-F. In ComprehensiVe Coordination Chemistry II: From Biology
to Nanotechnology; McCleverty, J. A., Meyer, T. J., Eds.; Elsevier
Pergamon: Oxford, UK, 2004; Vol. 8, p 479.
(7) Cabelli, D. E.; Riley, D.; Rodriguez, J. A.; Valentine, J. S.; Zhu, H. in
Biomimetic Oxidations Catalyzed by Transition Metal Complexes; Meunier,
B., Ed.; Imperial College Press: London, UK, 2000; p 461.
(8) Montagnier, L., Olivier, R., Pasquier, C., Eds. OxidatiVe Stress in Cancer,
Aids, and NeurodegeneratiVe Diseases; Marcel Dekker: New York, 1998.
(9) Martindale, J. L.; Holbrook, N. J. J. Cell. Physiol. 2002, 192, 1.
(10) Whittaker, J. W. Chem. ReV. 2003, 103, 2347.
(11) Klinman, J. P. J. Biol. Inorg. Chem. 2001, 6, 1.
(12) Dunford, H. B. Heme Peroxidases; Wiley-VCH: New York, 1999.
9
10.1021/ja056741n CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 3683-3695
3683

A R T I C L E S
Smirnov and Roth
Determining the identities of the reactive intermediates in the
enzymes is a particularly challenging problem. Much progress
has recently been made for PHM which like DâM is a member
of a family of structurally homologous copper proteins respon-
sible for the biosynthesis of hormones and neurotransmitters.24
Reduced copper(I) forms of PHM and DâM are believed to
bind O2 and generate reactive CuO2 intermediates which
stereoselectively oxidize substrate C-H bonds by hydrogen
atom abstraction.52 Related CuO2 species have also been
proposed to form transiently in Cu:Zn SOD upon reaction of
O2•- with the oxidized copper(II) enzyme. Release of O2 from
this intermediate is proposed to occur spontaneously, prior to
the reaction of copper(I) with a second equivalent of O2•-. An
alternative mechanism has also been considered wherein a CuO2
Figure 1. 1:1 copper-oxygen adducts. Roman numerals represent formal
oxidation states corresponding to dioxygen (0), superoxide (-I) and peroxide
(-II).
at low temperatures (ca. -80 °C).32-37 Spectroscopic and/or
X-ray crystallographic studies have provided evidence for end-
on (η1) superoxo38,39 structures as well as the side-on (η2)
superoxo36 and peroxo37,40 structures (Figure 1). Although
transient CuO2 intermediates have been proposed to form upon
•-
reacting O2 and copper(II),16,17,41-45 none of these has been
structurally characterized. In fact, no crystal structures of η1-
CuO2 complexes derived from inorganic compounds have been
reported to date.46 These species are generally viewed as
unstable with respect to the loss of O2. The η2-CuO2 species
are considered to be more stable like the well-known peroxo
compounds derived from group IX and X transition metals.37,47
Computational investigations have suggested these complexes
have electronic ground states with differing contributions from
•-
intermediate reacts with a second equivalent of O2 to afford
O2 and a copper(II) peroxo product which is subsequently
hydrolyzed to H2O2.51,53 These examples illustrate how differing
reactivity patterns have been attributed to CuO2 intermediates;
the activated oxygen ligand is unstable with respect to oxidation
in SOD and with respect to reduction by exogenous substrates
in PHM and DâM.
This study explores the mechanism by which O2•- is oxidized
to O2 by biomimetic copper(II) complexes. Competitive oxygen
(18O) isotope fractionation techniques are used to provide
evidence for intermediates during these reactions. The measure-
ments employ stable isotope mass spectrometry to determine
changes in the ratio of heavy to light isotopes from their natural
abundance levels.54 The technique has been used extensively
in the atmospheric and geological sciences. With respect to
biological reactions, early work by Berry55,56 and by Epstein57
demonstrated fractionation during respiration, although the
results were not interpreted in terms of the underlying redox
chemistry.
The use of competitive 18O kinetic isotope effects (KIEs) to
probe reactions of O2 is described in two recent reviews.58,59 In
addition, we have recently reported studies of 18O KIEs on O2
activation by classic inorganic compounds.47 Described here is,
to our knowledge, the first extension of the competitive isotope
fractionation technique to reactions of O2•-. These studies lay
a foundation for applying 18O KIEs to probe mechanisms by
which reactive oxygen species are processed by metalloenzymes.
A key aspect of the approach is the use of isotope effects to
compare structurally and spectroscopically defined biomimetic
compounds to metal-activated oxygen intermediates in enzymes.
resonance forms formally described as CuI-O2 , CuII-O2-I, and
0
CuIII-O2-II 31,48 Because of their oversimplified nature, the use
.
of oxidation state formalisms in reference to CuO2 species is
deemphasized throughout most of this article.
The relationship of electronic to vibrational structure in
inorganic oxygen complexes has been and continues to be a
thoroughly investigated area.31,49 As a complement to such
studies, we are developing new mechanistic probes which can
correlate structure to reactivity in enzymes and biomimetic
systems. This type of correlation is central to understanding the
catalytic mechanisms of several copper-containing enzymes
including peptidylglycine R-hydroxylating monooxygenase
(PHM),22-24,38 dopamine â-monooxygenase (DâM),22,50 tyro-
sinase,26 particulate methane monooxygenase27 and Cu:Zn
superoxide dismutase (Cu:Zn SOD).51
(32) Fry, H. C.; Scaltrito, D. V.; Karlin, K. D.; Meyer, G. J. J. Am. Chem. Soc.
2003, 125, 11866.
(33) Zhang, C. X.; Kaderli, S.; Costas, M.; Kim, E.-i.; Neuhold, Y.-M.; Karlin,
K. D.; Zuberbuehler, A. D. Inorg. Chem. 2003, 42, 1807.
(34) Wei, N.; Murthy, N. N.; Chen, Q.; Zubieta, J.; Karlin, K. D. Inorg. Chem.
1994, 33, 1953.
(35) Karlin, K. D.; Wei, N.; Jung, B.; Kaderli, S.; Niklaus, P.; Zuberbuehler,
A. D. J. Am. Chem. Soc. 1993, 115, 9506.
(36) Fujisawa, K.; Tanaka, M.; Moro-oka, Y.; Kitajima, N. J. Am. Chem. Soc.
1994, 116, 12079.
(37) Tolman, W. B. et al. J. Am. Chem. Soc. 2004, 126, 16896.
(38) Prigge, S. T.; Eipper, B. A.; Mains, R. E.; Amzel, L. M. Science 2004,
304, 864.
Experimental Methods
All manipulations of moisture- and oxygen-sensitive compounds
were carried out using standard high-vacuum techniques or a N2-filled
glovebox (MBraun). Potassium superoxide (KO2) was obtained as a
(39) Schatz, M.; Raab, V.; Foxon, S. P.; Brehm, G.; Schneider, S.; Reiher, M.;
Holthausen, M. C.; Sundermeyer, J.; Schindler, S. Angew. Chem., Int. Ed.
Eng. 2004, 43, 4360.
(40) Reynolds, A. M.; Gherman, B. F.; Cramer, C. J.; Tolman, W. B. Inorg.
Chem. 2005, 44, 6989.
(41) Luo, Q.-H.; Zhang, J.-J.; Hu, X.-L.; Jiang, X.-Q.; Shen, M.-C.; Li, F.-M.
Inorg. Chim. Acta 2004, 357, 66.
(52) Francisco, W. A.; Blackburn, N. J.; Klinman, J. P. Biochemistry 2003, 42,
1813.
(42) Nishida, Y.; Unoura, K.; Watanabe, I.; Yokomizo, T.; Kato, Y. Inorg. Chim.
Acta 1991, 181, 141.
(53) Osman, R.; Basch, H. J. Am. Chem. Soc. 1984, 106, 5710.
(54) McKinney, C. R.; McCrea, J. M.; Epstein, S.; Allen, H. A.; Urey, H. C.
ReV. Sci. Instrum. 1950, 21, 724.
(43) Bailey, C. L.; Bereman, R. D.; Rillema, D. P. Inorg. Chem. 1986, 25, 3149.
(44) Nappa, M.; Valentine, J. S.; Miksztal, A.; Schugar, H. J.; Isied, S. S. J.
Am. Chem. Soc. 1979, 101, 7744.
(55) Guy, R. D.; Fogel, M. F.; Berry, J. A.; Hoering, T. C. In Progress in
Photosynthesis Research, Proceedings of the 7th International Congress
on Photosynthesis; Biggins, J., Ed.; Martinus Nijhoff Publishers: Dordrect,
The Netherlands, 1987; Vol. 3, p 597.
(45) Meisel, D.; Levanon, H.; Czapski, G. J. Phys. Chem. 1974, 78, 779.
(46) Aboelella, N. W.; Reynolds, A. M.; Tolman, W. B. Science 2004, 304,
836.
(47) Lanci, M.; Brinkley, D. W.; Stone, K. L.; Smirnov, V. V.; Roth, J. P. Angew.
Chem., Int. Ed. Eng. 2005, 44, 7273.
(48) Gherman, B. F.; Cramer, C. J. Inorg. Chem. 2004, 43, 7281.
(49) Lever, A. B. P.; Ozin, G. A.; Gray, H. B. Inorg. Chem. 1980, 19, 1823.
(50) Tian, G.; Berry, J. A.; Klinman, J. P. Biochemistry 1994, 33, 226 and 14650.
(51) Hart, P. J.; Balbirnie, M. M.; Ogihara, N. L.; Nersissian, A. M.; Weiss, M.
S.; Valentine, J. S.; Eisenberg, D. Biochemistry 1999, 38, 2167.
(56) Guy, R. D.; Berry, J. A.; Fogel, M. L.; Hoering, T. C. Planta 1989, 177,
483.
(57) Epstein, S.; Zeiri, L. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 1727.
(58) Roth, J. P.; Klinman, J. P. in Isotope Effects in Chemistry and Biology;
Kohen, A., Limbach, H. H., Eds.; CRC Press: Boca Raton, FL, 2005; p
645.
(59) Smirnov, V. V.; Brinkley, D. W.; Lanci, M. P.; Karlin, K. D.; Roth, J. P.
J. Mol. Catal. A 2006, accepted.
9
3684 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Cu−O2 Intermediates in Superoxide Oxidations
A R T I C L E S
yellow powder from Aldrich (45.04% O by weight) and used as
received. 18-crown-6 ether was obtained from Aldrich (99.9% purity)
and dried over P2O5 in vacuo for several days prior to use. Ferricyto-
chrome c from bovine heart and manganese superoxide dismutase from
E. coli were obtained from Sigma and used as received. Tetra-n-
butylamonium hexafluorophosphate (nBu4NPF6) was obtained from
Strem Chemicals and dried in vacuo over P2O5 prior to use. Ferrocene
(TCI America) was sublimed prior to use in electrochemical experi-
ments. “Anhydrous” grade DMSO and DMF solvents were obtained
from Burdick and Jackson and used without further purification. DMF
was stored in the dark at -30 °C. Tetrahydrofuran (THF) was stored
over sodium/benzophenone and vacuum distilled prior to use. Aceto-
nitrile (CH3CN) was dried and purified by two vacuum distillations,
first from calcium hydride and then from P2O5. Deionized water was
purified to 18 MΩ by passing through a Millipore Milli-Q system.
Copper complexes were synthesized according to modified literature
procedures.60,61 Samples were dried in vacuo over P2O5 for 24 h prior
to use. Purity was checked by comparison with reported extinction
coefficients and elemental analysis (Desert Analytics, Tucson, AZ).
Cu(TEPA)OTf2 {CuII(TEPA)} (TEPA ) tris(2-pyridylethyl)amine and
OTf ) trifluoromethanesulfonate): Anal. Calcd for C23H24N4O6F6-
S2Cu: C, 39.80; H, 3.49; N, 8.07. Found: C, 39.83; H, 3.34; N, 7.96.
Cu(TEPA)OTf {CuI(TEPA)}: Anal. Calcd for C22H24N4O3F3SCu: C,
48.48; H, 4.44; N, 10.28. Found: C, 48.39; H, 4.65; N, 10.15. CuN3-
(TEPA)OTf {CuIIN3(TEPA)}: Anal. Calcd for C22H24N7O3F3SCu: C,
45.01; H, 4.12; N, 16.70. Found: C, 45.07; H, 3.97; N, 16.38.
Cu(TMPA)OTf2 {CuII(TMPA)} (TMPA: tris(2-pyridylmethyl)amine):
Anal. Calcd for C20H18N4O6F6S2Cu: C, 36.84; H, 2.78; N, 8.59.
Found: C, 36.30; H, 2.87; N, 8.46. CuN3(TMPA)OTf {CuIIN3-
(TMPA)}: Anal. Calcd for C19H18N7O3F3SCu: C, 41.87; H, 3.33; N,
17.99. Found: C, 41.78; H, 3.21; N, 17.85. An analytically pure sample
of [CuI(TMPA)](PF6) {CuI(TMPA)}62 was provided by Prof. Kenneth
D. Karlin. The following molar absorptivities were determined in
DMSO: CuII(TEPA) ꢀ ) 15 100 ( 300 M-1 cm-1 (λmax ) 263 nm),
CuI(TEPA) ꢀ ) 9300 ( 120 M-1 cm-1 (λmax ) 345 nm), CuIIN3-
(TEPA) ꢀ ) 2500 ( 110 M-1 cm-1 (λmax ) 411 nm), CuIIN3(TMPA)
ꢀ ) 3400 ( 130 M-1 cm-1 (λmax ) 412 nm).
residual O2. A typical procedure involved suspending solid KO2 (26
mg) in DMSO (100 mL) and, after 25 min of stirring, filtering the
solution through a Corning nylon membrane with 0.2 µm pore size. A
clear, pale yellow solution was obtained. 18-Crown-6 ether (67 mg)
was used to solubilize KO2 (18 mg) in cold DMF (20 mL). The
suspension was stirred for 3 min and filtered through a nylon membrane.
The solution was diluted with DMF or DMF/THF (28 mL DMF/12
mL THF) for use in the low-temperature stopped-flow experiments.
•-
The concentration of O2 was determined by injecting a 10 µL
aliquot of N2-saturated solution into 1.00 mL aqueous solution of
ferricytochrome c or ferricytochrome c + Mn SOD (pH 10 NaHCO3
buffer, 10-4 M EDTA). The concentration of O2•- was calculated from
the SOD-inhibitable activity indicated by Clark electrode readings.63
A calibration curve was prepared by correlating the concentrations of
O2•- to UV absorbance measurements using a 0.1 cm path length quartz
cell. Beer’s law analysis indicated an extinction coefficient ꢀ254
)
nm
•-
4200 ( 200 M-1cm-1 for O2 in DMSO, which is somewhat higher
than previously reported values.64-66 Protected from moisture, these
solutions were stable for more than 12 h at room temperature. Only a
4% drop in the concentration of O2•- was detected by spectrophotometry
and by manometry on the O2 produced from a standard reaction (see
•-
below). Solutions of O2 in DMF were less stable and decayed by
20% over the course of 6 h. For this reason, KIE measurements were
performed with DMSO as the solvent.
Synthesis of CoIII(SMDPT)OTf‚H2O for Use as a Calibration
Standard. CoIII(SMDPT)OTf‚H2O {CoIII(SMDPT)} (SMDPT ) bis-
(salicylidene-γ-iminopropyl)methylamine) was prepared by oxidation
of CoII(SMDPT)67 with an equimolar amount of AgOTf in CH3CN.
The AgOTf in anhydrous CH3CN was added slowly to a stirred solution
of CoII(SMDPT) in CH3CN protected from light. The reaction mixture
was allowed to stir overnight. The Ag0 produced was removed by
several filtrations through diatomaceous earth (Celite 521) and a 0.2
µm nylon membrane. The CH3CN was removed by rotary evaporation.
The remaining black residue was purified by multiple crystallizations
from CH2Cl2/Et2O. The final product had an appearance of dark red
needles without any metallic shine. The yield of CoIII(SMDPT) after
1
recrystallization and drying in vacuo was 64%. H NMR (400 MHz,
NMR spectra were recorded using Bruker Avance 400 MHz FT-
NMR spectrometers. Chemical shifts (δ) are reported relative to residual
protio signals of the deuterated solvent. UV-vis spectra were recorded
on an Agilent 8453 UV-vis spectrophotometer. The cell temperature
was controlled with a Peltier 89090-A. Room-temperature stopped-
flow kinetic measurements were performed on an OLIS RSM 1000
spectrophotometer equipped with a 1 cm observation cell. The
spectrophotometer was installed in a N2-filled double-length glovebox
(MBraun). Low-temperature stopped-flow experiments were conducted
on SF-40 Hi-Tech Scientific stopped-flow system equipped with a 1
cm path length cell, NMC 301 diode array detector (J & M, 300-
1100 nm range, 1.3 ms sampling time) and flexible light guides
connected to a CLX 75 W xenon lamp. Cyclic voltammetry was carried
out using a Bioanalytical Systems BAS 100B electrochemical analyzer.
The concentration of O2 in aqueous samples was determined using a
Clark electrode (Biological Oxygen Monitor 5300A, YSI). The 18O/
16O ratio in CO2 samples was determined using a Micromass stable
isotope mass spectrometer equipped with a dual inlet system. Isotope
ratio mass spectrometry services were provided by facilities at Johns
Hopkins University and the University of Maryland, College Park.
Preparation of Superoxide Solutions. All glassware was rinsed
with Milli-Q (18MΩ) water and rigorously dried. Solutions of KO2 in
anhydrous DMSO or DMF were freshly prepared under N2 just before
use. Care was taken to protect solutions from moisture and to remove
CD3CN) δ 7.99 (1H, s), 7.85 (1H, s), 7.29 (2H, t, J ) 8 Hz), 7.15 (2H,
br t), 7.00 (1H, br s), 6.93 (1H, d, J ) 8 Hz), 6.63 (1H, t, J ) 8 Hz),
6.58 (1H, t, J ) 8 Hz), 5.46 (2H, very br s, H2O-Co), 3.60 (1H, br t,
J ) 12 Hz), 3.44 (1H, t, J ) 12 Hz), 3.14 (1H, dt, J1 ) 4 Hz, J2 ) 13
Hz), 2.61 (1H, br t), 2.30 (1H, m), 2.10 (5H, m), 1.70 (3H, s, H3C),
1.60 (2H, m); UV-vis (DMSO): λmax ) 385 nm, ꢀ ) 4700 ( 400
M-1cm-1. Anal. Calcd. for C22H27N3O6F3SCo: C, 45.76; H, 4.71; N,
7.28. Found: C, 45.93; H, 4.71; N, 7.25.
The reaction of 1-3 equiv. of CoIII(SMDPT) with 1 equivalent of
O2•- in DMSO resulted in a 101 ( 4% yield of O2 by manometry and
a 93 ( 11% yield of CoII(SMPDT) based on its optical absorbance
spectrum (λmax ) 350 nm, ꢀ ) 12500 ( 2000 M-1cm-1).68 This
reaction, which proceeds quantitatively, was used to determine both
•-
the concentration of O2 in solution as well as the initial isotope
composition of this nonenriched reactant (see below).
Determination of Isotope Effects. The apparatus and basic meth-
odology used to determine competitive oxygen kinetic isotope effects
has been previously described.58,59 Since the present studies include,
to our knowledge, the first measurements to be performed on reactions
of O2•- a brief overview is provided. A collapsible Tedlar bag (MiDan
company, Chino, CA) was used as the reaction vessel to accommodate
the use of organic solvents. Under N2, an assembly consisting of a
(63) Fridovich, I. J. Biol. Chem. 1970, 245, 4053.
(64) Kim, S.; DiCosimo, R.; San Filippo, J., Jr. Anal. Chem. 1979, 51, 679.
(65) Gampp, H.; Lippard, S. J. Inorg. Chem. 1983, 22, 357.
(66) Sawyer, D. T.; Calderwood, T. S.; Yamaguchi, K.; Angelis, C. T. Inorg.
Chem. 1983, 22, 2577.
(67) Bailes, R. H.; Calvin, M. J. Am. Chem. Soc. 1947, 69, 1886.
(68) Drago, R. S.; Cannady, J. P.; Leslie, K. A. J. Am. Chem. Soc. 1980, 102,
6014.
(60) Karlin, K. D.; Hayes, J. C.; Hutchinson, J. P.; Hyde, J. R.; Zubieta, J. Inorg.
Chim. Acta 1982, 64, L219.
(61) Karlin, K. D.; Hayes, J. C.; Juen, S.; Hutchinson, J. P.; Zubieta, J. Inorg.
Chem. 1982, 21, 4106.
(62) Tyeklar, Z.; Jacobson, R. R.; Wei, N.; Murthy, N. N.; Zubieta, J.; Karlin,
K. D. J. Am. Chem. Soc. 1993, 115, 2677.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3685

A R T I C L E S
Smirnov and Roth
resealable round-bottom flask, two-way crossover stopcock and solvent
•-
impervious bag was charged with the O2 solution. After connecting
to the vacuum line, both the bag and the solution in the round-bottom
flask were purged with He gas for 30 min to remove any residual O2.
The bag was filled with the O2•- solution while displacing the He gas.
CoIII(SMDPT) in DMSO (20 mg in 1.5 mL) was placed in a gas
bubbler, attached to the vacuum manifold and then degassed. The
vacuum apparatus was prepared for sample collection by evacuation
to ∼10 milliTorr.
In a typical measurement, a calibrated tube built into the vacuum
apparatus was filled with the O2•- solution. This solution was dispensed
to the bubbler containing the CoIII(SMDPT) standard. Immediately,
the mixture was vigorously purged with He gas under dynamic vacuum
for 20 min to remove O2 and collect it in a molecular sieve trap cooled
to -196 °C with liquid N2. Excess He was released under reduced
pressure prior to liberating O2 from the sieves by warming with hot
water. The O2 was then recirculated through a combustion furnace and
the CO2, which formed quantitatively, condensed into a liquid nitrogen-
cooled finger. The CO2 was then recondensed into a fixed-volume part
of the vacuum apparatus. Its pressure was accurately measured with a
capacitance manometer (Omega PX238 with U24Y101 Linear Power
Supply) calibrated using predetermined amounts of O2. The CO2 sample
was sealed under vacuum in a dry glass tube for isotope ratio mass
Figure 2. Structures and abbreviations of copper(II) complexes.
O2, or mixtures thereof. O2 solutions of desired concentration were
prepared by purging solvents with O2/Ar gas mixtures created with a
MultiGas Controller 647C (MKS Instruments). The mixing chamber
was immersed in an ethanol bath cooled by liquid N2 evaporation for
low-temperature measurements. The temperature was measured using
a Pt resistance thermocouple and maintained to (0.1 °C using a
temperature-controlled thyristor unit (Hi-Tech). Data acquisition was
performed using the Spectralys software (J & M).
spectrometry analysis. These measurements provided the 18O/16O
•-
composition of the nonenriched O2
.
•-
Reactions between O2 and copper complexes were performed as
described above for the standard reaction of CoIII(SMDPT) except that
substoichiometric quantities of the oxidant were used. In a typical set
of measurements, two standard reactions and four measurements with
Results
I. Description of Reactions. The structures of the copper(II)
complexes and abbreviations used to refer to them are given in
Figure 2. All of the complexes except for CuIIN3(TEPA) have
been shown by X-ray crystallography to adopt 5-coordinate
geometries.60,61,70-72 Both copper(II) TEPA complexes are
expected to exhibit a square pyramidal geometry by analogy
with others having H2O and 1-methylimidazole60,71 as well as
anions such as NO3-, Cl- and acetate as the fifth ligands.61,71,72
•-
the copper complexes were performed. The concentrations of the O2
solution determined from the standard reaction at the beginning and
end of the day were typically within 4%. Changes in O2•- concentration
due to reactions with copper complexes were determined according
to: f ) PCO (Cu)/PCO (standard), where P is the pressure of CO2, and f is
2
2
the fraction of O2 produced.
Electrochemistry. Cyclic voltammetry was performed using a three-
electrode setup consisting of a working glassy carbon electrode, a
platinum wire counter electrode, and Ag/Ag+ reference electrode (0.01
M AgNO3 in DMSO containing 0.1 M nBu4NPF6). Measurements were
conducted under an inert atmosphere of Ar gas at room temperature
using 0.1 M nBu4NPF6 as the supporting electrolyte in anhydrous
DMSO. The concentration of Cu(II) complexes was 5 × 10-3 M. Quasi-
reversible behavior was observed and half-wave potentials (E°′) were
-
On the basis of these structures, N3 and solvent are expected
to coordinate at the base of the square pyramid with one
elongated Cu-NPy bond coordinated to the apical site.
CuI(TEPA) crystallized from noncoordinating solvents is
pseudo-tetrahedral and distorted toward trigonal pyramidal
geometry.60 In contrast, both copper(I) and copper(II) complexes
of TMPA are trigonal bipyramidal with solvent (CH3CN) or an
anion (Cl-, N3-, NO2-, NCS-) in the case of copper(II)
occupying an apical site.61,62,70,73 These structural differences
may influence the barriers to redox reactions as well as the
found to be independent of scan rate between 1 and 100 mV s-1
.
Experiments were performed at a scan rate of 20 mV s-1 and referenced
to a ferrocene/ferrocenium internal standard (FeCp2+/0) which exhibits
E°′ ) +0.68 V vs NHE in DMSO.69 All E°′ are reported vs NHE.
energies of intermediates formed upon reacting copper(I) with
O2 and copper(II) with O2
Stopped-Flow Measurements. Stopped-flow kinetic measurements
were performed on OLIS RSM 1000 spectrophotometer which has a
2-3 millisecond dead-time. The drive syringes and observation cell
were thermostated to +20 °C using a recirculating water bath
(ThermoHaake K20). Experiments were performed by 1:1 mixing of
reagents with initial concentrations of 0.02, 0.21, and 2.1 mM.
Measurements were conducted in multi- and single-wavelength modes.
The rate of CuI(TEPA) appearance was monitored at λmax ) 345 nm
for the reactions of CuII(TEPA) and CuIIN3(TEPA). The rate of
CuIIN3(TMPA) disappearance was monitored at λmax ) 412 nm. All
of the reactions proceeded to near completion within the mixing time
of the instrument.
•-
.
•-
All of the copper(II) complexes react rapidly with O2 at
22 °C in DMSO, where the radical anion is both a powerful
reductant and a nucleophile. The related copper(I) complexes
and O2 are formed as products in yields that depend on the initial
concentrations of the reactants. When substoichiometric amounts
of O2 are added to CuII(TEPA), the yield of CuI(TEPA) is
•-
66 ( 5%. The reaction of CuIIN3(TEPA) with O2•- also forms
CuI(TEPA) in 60 ( 7% yield indicating that the product is
thermodynamically unstable with respect to loss of N3-. No
reaction is observed between CuI(TEPA) and O2 under the
Low-temperature stopped-flow experiments at -57 °C and -80 °C
were conducted using a Hi-Tech Scientific stopped-flow system. Mixing
was performed through two stainless steel coils, lined with Teflon. One
(70) Mukhopadhyay, U.; Bernal, I.; Massoud, S. S.; Mautner, F. A. Inorg. Chim.
Acta 2004, 357, 3673.
reservoir contained the copper reactants while the other contained O2•-
,
(71) Alilou, E. H.; Hallaoui, A. E.; Ghadraoui, H. E.; Giorgi, M.; Pierrot, M.;
Reglier, M. Acta Crystallogr. Sect. C 1997, 53, 559.
(72) Karlin, K. D.; Dahlstrom, P. L.; Hayes, J. C.; Simon, R. A.; Zubieta, J.
Cryst. Struct. Comm. 1982, 11, 907.
(69) Sawyer, D. T.; Sobkowiak, A.; Roberts, J. L., Jr. Electrochemistry for
Chemists; Wiley: New York, 1995.
(73) Lim, B. S.; Holm, R. H. Inorg. Chem. 1998, 37, 4898.
9
3686 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Cu−O2 Intermediates in Superoxide Oxidations
A R T I C L E S
•-
Table 1. Yields of O2 Based on Copper(I) and Copper(II)
Complexes in 10:1 DMSO/DMF at 22 °C
Table 2. Thermodynamics for the Oxidation of O2 to O2 by
Copper(II) Complexes
•-
•-
•-
entry
oxidant
1 eq. O2
2 eq. O2
g
3 eq. O2
E
°′
∆
G
°
wP-wR
∆G°′
1
1
1
(V vs NHE)
(kcal mol-
)
(kcal mol-
)
(kcal mol-
)
1
2
3
4
5
6
CuII(TEPA)
CuIIN3(TEPA)
CuII(TMPA)
CuIIN3(TMPA)
CuI(TEPA)
68 ( 4%
73 ( 1%
72 ( 3%
72 ( 1%
53 ( 6%
51 ( 6%
133 ( 1%
147 ( 2%
142 ( 1%
142 ( 3%
192 ( 32%a
201 ( 18%a
189 ( 23%a
193 ( 33%a
CuII(TEPA)
+0.52a
+0.38b
-0.05c
-0.15d
-24.4
-21.2
-11.3
-9.0
+5.5
+2.8
+5.5
+2.8
-18.9
-18.4
-5.8
CuIIN3(TEPA)
CuII(TMPA)
CuIIN3(TMPA)
91 ( 9%b
115 ( 19%b
-6.2
CuI(TMPA)
a Ea-Ec ) 0.41 V and ip,a/ip,c ) 0.74. b Ea-Ec ) 0.27 V, ip,a/ip,c ) 0.82.
a 3-8 equiv. b 2-6 equiv.
c Ea-Ec ) 96 mV and ip,a/ip,c ) 0.95. d Ea-Ec ) 92 mV and ip,a/ip,c
)
0.90.
and are converted to two equivalents of O2 and an as yet
unidentified copper species.
Attempts to characterize the final copper-containing product
from reactions of copper(I) or copper(II) complexes with excess
O2•- have thus far been unsuccessful. 1H NMR analysis indicates
a diamagnetic compound is formed and complete mass balance
of the reaction.76 The product mixture does not appear to contain
H2O2 or a copper product with an intact O-O bond. This
conclusion is based on the absence of detectable levels of H2O2
upon hydrolysis and assay with horseradish peroxidase and
diammonium salt of 2,2′-azino-bis(3-ethylbenzothiazoline-6-
sulfonate) (ABTS).77
•-
Figure 3. Reactions of copper(I) and copper(II) complexes with O2
.
experimental conditions,74 thus precluding this as an explanation
of the lower than expected product yields. Yields of the
CuII(TMPA) and CuIIN3(TMPA) reactions with O2 were
•-
II. Thermodynamic Analysis. Cyclic voltammetry experi-
ments (DMSO, 0.1 M nBu4PF6, 22 °C) revealed a single wave
corresponding to the CuII/I redox couples (Table 2). The
chemical reversibility was somewhat poorer for the more
oxidizing TEPA complexes than for the TMPA complexes. The
overall free energies (∆G°) corresponding to O2•- oxidation via
the first reaction in Figure 3 were estimated from half-wave
potentials (E°′) of the copper complexes and E°′ ) -0.54 V
for O20/- in DMSO.69 The equation used was: ∆G° ) nF∆E°′,
where n is the number of electrons transferred, and F is
Faraday’s constant (23.06 kcal mol-1 V-1).
analyzed using an alternative method to avoid the subsequent
reactivity of the copper(I) products with O2. The reactants were
mixed 1:1 under dynamic vacuum while sweeping the solutions
with He gas to remove the O2 produced. Subsequent evaporation
of the solvent and analysis of the remaining solids by 1H NMR
indicated formation of CuI(TMPA) in an average yield of 44
( 13%.75 Again, the thermodynamic product did not contain
-
bound N3
.
The O2 produced under the conditions described above was
quantified by manometry. Measurements were relative to the
reaction with the CoIII(SMDPT) standard which converts all
•-
Oxidations of O2 by CuII(TEPA) and CuIIN3(TEPA) are
•-
of the O2 in solution to O2 (see Experimental Methods).
highly exoergic with ∆G° of -24.4 and -21.2 kcal mol-1. The
reactions of CuII(TMPA) and CuIIN3(TMPA) are less favorable
with ∆G° of -11.3 kcal mol-1 and -9.0 kcal mol-1. The
derived ∆∆G° of 12-13 kcal mol-1 reflects the much greater
stabilization of the copper(I) state in the TEPA ligand frame-
work.78 The large difference provides an explanation as to why
CuI(TMPA) is highly reactive with O2 whereas CuI(TEPA) is
completely inert to O2 in polar organic solvents from 20 to -80
•-
When the copper(II) complexes were reacted with O2 in a
1:1 ratio, the 68-73% yields of O2 obtained were somewhat
higher than the yields of copper(I) products quoted above (Table
•-
1, Entries 1-4). Increasing the ratio to 2:1, O2 to copper(II)
complex, caused the yield of O2 to increase to 133-147%. For
the 3:1 reaction, a yield of 189-201% was obtained. No further
increase in the O2 yield was observed when 3-8 equivalents
•-
of O2 were reacted per mol of copper(II) complex.
-
°C. Within each ligand environment, coordination of N3
The increase in O2 yield, in direct proportion to the increased
resulted in a negative shift in redox potential amounting to
∆∆G° ) 2-3 kcal mol-1. A larger difference in O2•- binding
energy, approaching 13 kcal mol-1, is expected based on the
CuII/I redox potentials.79
•-
ratio of O2 to copper(II) complex, is due to the subsequent
reactions of CuI(TEPA) or CuI(TMPA). The stoichiometry of
these reactions was determined by quantifying the O2 formed
•-
at varying ratios of O2 to the copper(I) complexes (Table 1,
Analysis of the electrochemical potentials gives the net
driving-force for the bimolecular reaction of copper(II) com-
Entries 5 and 6). The 1:1 reactions resulted in yields of 51-
•-
53% indicating that two equivalents of O2 were consumed
•-
plexes and O2 to form the related copper(I) complexes and
•-
per equivalent of O2 formed. With 2 to 6 equivalents of O2
O2. However, corrections must be applied for the electrostatic
per mol of copper(I) complex, the yield increased to 91-115%.
The results are consistent with the scheme in Figure 3 wherein
three equivalents of O2•- react with a single copper(II) complex
(76) A 3.5 × 10-2 M solution of CuII(TEPA) in DMSO-d6 treated with excess
KO2 solubilized with 18-crown-6 (2.4 × 10-2 M) in DMSO-d6 at room
temperature reveals signals corresponding to CuI(TEPA) and three broad
singlets. The latter occur at δ 8.60, 7.74, 7.27 ppm and integrate 1:1:2,
respectively. The integrated intensities indicate the complete mass balance
in the reaction suggesting the formation of a single diamagnetic product.
(77) Santagostini, L.; Gullotti, M.; Monzani, E.; Casella, L.; Dillinger, R.;
Tuczek, F. Chem. Eur. J. 2000, 6, 519.
(74) Schatz, M.; Becker, M.; Thaler, F.; Hampel, F.; Schindler, S.; Jacobson,
R. R.; Tyeklar, Z.; Murthy, N. N.; Ghosh, P.; Chen, Q.; Zubieta, J.; Karlin,
K. D. Inorg. Chem. 2001, 40, 2312.
(75) The yield of CuI(TMPA) was determined by 1H NMR analysis of three
independent samples prepared from crude reaction mixtures. See Supporting
Information for details.
(78) Rorabacher, D. B. Chem. ReV. 2004, 104, 651.
(79) Taube, H. Prog. Inorg. Chem. 1986, 34, 607.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3687

A R T I C L E S
Smirnov and Roth
Figure 4. Time-dependent UV-vis spectra taken during the first 100 ms in 4:1 DMF:THF at -80 °C. Initial concentrations of reactants after mixing are
as follows: (a) [CuII(TMPA)] ) 5.5 × 10-4 M, [O2•-] ) 1.25 × 10-4 M and (b) [CuI(TMPA)] ) 1.25 × 10-4 M, [O2] ) 1.25 × 10-4 M.
attraction of oppositely charged reactants in order to evaluate
the thermodynamics of the unimolecular electron transfer (ET)
step. For this purpose, electrostatic work terms (wR and wP),
which represent the free energies of interactions within a
complex of the reactants or products, are calculated using a
dielectric continuum model.80
by the lack of optical change and the absorbance corresponding
to CuI(TEPA) present in 65 ( 8% yield. Under these condi-
tions, a lower limit to the second-order rate constant of g2 ×
106 M-1s-1 is estimated as described above. The reactions of
•-
CuII(TMPA) and CuIIN3(TMPA) with O2 were examined
under similar conditions at -57 °C and at -80 °C using a 3-4-
•-
In the most basic approach (eq 2), terms include the
permittivity of vacuum constant (e2 ) 332.1 Å kcal mol-1),
the reactant or product charges (Z1 and Z2), the dielectric
screening factor (fsf), the static dielectric constant (Ds), which
is 47.24 for DMSO,81 and the internuclear distance (r12). The
fold excess of the copper complexes relative to O2 in order
to minimize competing side reactions. The spectral data obtained
at both low temperatures were similar. Within 3 milliseconds
of mixing CuII(TMPA) and O2 at -80 °C, an intermediate
•-
is detected that is spectroscopically similar to a 1:1 Cu(TMPA):
O2 adduct previously observed upon reacting CuI(TMPA) and
O2 in propionitrile and in THF.32,33,35 The manner in which O2
is coordinated to the copper center in this intermediate has not
been established.83 However, an end-on superoxo structure has
been invoked to explain the reactivity toward a second equiv.
of CuI(TMPA) and the formation of the trans-µ-1,2-peroxo
compound: [Cu(TMPA)]2O2.84
•-
close approach of O2 to the copper center is modeled by
assuming r12 ) 2.0 Å and the resulting fsf ) 0.80.82 The former
value is based on Cu-O distances in crystal structures of
CuII(TEPA) with nitrate and acetate counterions.71,72 The
difference in electrostatic work terms (wP-wR) is then subtracted
from ∆G° to obtain the effective driving-force for the uni-
molecular redox step (∆G°′) (Table 2). It is important to note
that covalent bonding between the reactants is not reflected in
these work terms and a CuO2 intermediate may be either more
or less stable than the complexes held together only by
electrostatics.79
Spectroscopic data from the reaction of CuII(TMPA) and
•-
O2 are strikingly similar to those obtained upon reacting
CuI(TMPA) and O2 under the analogous conditions; yet
formation of the dicopper peroxo product occurs on different
time scales (Figure 4). The 1:1 Cu(TMPA):O2 intermediate has
prominent absorbance bands at 412, 580, and 765 nm and decays
concomitant with the appearance of absorbance bands at 525
and 600 nm which have been assigned to [Cu(TMPA)]2O2.84
Within the dead-time of the stopped-flow, the reaction of
CuII(TMPA) (5.5 × 10-4 M) and O2•- (1.25 × 10-4 M) gives
the intermediate 1:1 Cu(TMPA):O2 adduct and [Cu(TMPA)]2O2
in yields of 75 ( 8% and 11 ( 5%, respectively (Figure 4a).
The concentration of [Cu(TMPA)]2O2 increases to 32 ( 5%
over the course of 100 milliseconds. The formation of the
dicopper peroxo product is inhibited by added O2 (5.4 × 10-3
M) consistent with an earlier step involving dissociation of the
1:1 Cu(TMPA):O2 to form CuI(TMPA). As discussed above
e2Z1Z2fsf
wR or wP )
(2)
Dsr12
III. Detection of Intermediates. Stopped-flow spectro-
photometric measurements at varying temperatures were un-
•-
dertaken to detect intermediates during O2 oxidation by
copper(II) complexes. Rates were too fast to measure in DMSO
at 20 °C, which is just above the freezing point of this solvent.
Even one of the least thermodynamically favorable reactions
•-
involving CuIIN3(TMPA) (1 × 10-5 M) and O2 (1 × 10-5
M) had reached completion within the 3 millisecond mixing
time of the stopped-flow instrument. A lower limit to the second-
•-
(Results, I.), the rapid reaction of CuI(TMPA) with O2 is
order rate constant of g1 × 108 M-1 -1
s
is, therefore, estimated
expected to give a mononuclear copper(II) peroxo species,85
which could then react with a second equivalent of CuII(TMPA)
to generate the final product.
under bimolecular conditions from k ) 1/t1/2[A]0; the reaction
half-life (t1/2) was e1 millisecond and each reactant was present
at an initial concentration of [A]0.
(80) Sutin, N. Prog. Inorg. Chem. 1983, 30, 441.
Low-temperature, stopped-flow experiments were performed
in DMF (-57 °C) and in a 4:1 mixture of DMF/THF (-80
°C). 18-crown-6 ether was used to solubilize the KO2 in both
solvents. Although these experiments are not directly compa-
rable to those at ambient temperature in DMSO, they provide
insight to possible reaction intermediates. The 1:1 reaction of
CuII(TEPA) (5.0 × 10-4 M) and O2•- (5.0 × 10-4 M) at -57
°C was over within the spectrometer’s mixing time as evidenced
(81) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 84th ed.; CRC
Press: Boca Raton, FL, 2003.
(82) Here fsf ) 1/[1 + r12(8πe2µ/1027DskBT)1/2] and r12 ) 2 Å; e2 ) 332.0637
kcal Å mol-1; µ ) 0.1 mol L-1; Ds ) 47.24; kB ) 3.2998 × 10-27 kcal
K-1; T ) 298.15 K.
(83) Karlin, K. D.; Kaderli, S.; Zueberbuehler, A. D. Acc. Chem. Res. 1997,
30, 139.
(84) Jacobson, R. R.; Tyeklar, Z.; Farooq, A.; Karlin, K. D.; Liu, S.; Zubieta,
J. J. Am. Chem. Soc. 1988, 110, 3690.
(85) Burstyn, J. N.; Roe, J. A.; Miksztal, A. R.; Shaevitz, B. A.; Lang, G.;
Valentine, J. S. J. Am. Chem. Soc. 1988, 110, 1382.
9
3688 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Cu−O2 Intermediates in Superoxide Oxidations
A R T I C L E S
•-
Figure 5. Isotope fractionation during reactions of copper(II) complexes with excess O2 (22 °C, 10:1 DMSO:DMF).
The reaction of CuI(TMPA) (1.25 × 10-4 M) and O2 (1.25
× 10-4 M) forms the 1:1 Cu(TMPA):O2 intermediate in 96 (
7% yield. However, unlike the reaction of CuII(TMPA) and
O2•-, the concentration of the intermediate changes by e3 (
5% over 100 milliseconds (Figure 4b). The much slower buildup
of the [Cu(TMPA)]2O2 product, as well as the independence of
rate on added CuII(TMPA) (5.0 × 10-4 M), indicates a different
mechanism than described above. This mechanism most likely
involves the coordination of CuI(TMPA) to the 1:1 Cu(TMPA):
O2 adduct.33,35
The copper(I) and copper(II) complexes appear to react
through monomeric structures. This point is raised because a
pyridyl-bridged copper(I) dimer [CuI(TMPA)]2 has been as-
signed as the dominant species in acetone solutions pre-
pared from CuI(TMPA). Reaction of [CuI(TMPA)]2 with O2
in this solvent initially gives a 2:1 Cu(TMPA):O2 intermediate
in ∼50% yield based on copper.33 1H NMR analysis of
CuI(TMPA) in DMF at -57 °C reveals some degree of
aggregation possibly to [CuI(TMPA)]2.33 However, the high
yield (96 ( 7%) of the 1:1 Cu(TMPA):O2 intermediate ob-
tained from CuI(TMPA) and O2 in DMF and DMF:THF
mixtures allows the reaction through the dimeric structure to
be excluded.
is too fast to be monitored. If the 1:1 Cu(TMPA):O2 adduct
were formed as a secondary product, then the reaction would
involve initial outer-sphere electron transfer followed by an
association step. This mechanism is addressed (below) in the
context of competitive kinetic isotope effect measurements.
IV. Competitive Oxygen Kinetic Isotope Effects. Competi-
tive 18O KIEs were determined from the fractionation of oxygen
isotopes, i.e., the change in 18O/16O during the oxidation of O2
•-
to O2 by the various copper complexes. Because the reactions
are extremely rapid, measurements were performed using excess
•-
O2 and varying sub-stoichiometric amounts of copper rather
than in a time dependent manner. By virtue of being a
competitive measurement, the isotope effect derives from the
second-order rate constants for the disappearance of light 16O-
16O•- and heavy 18O-16O•- 16,16k/18,16k). The 18O KIE is
(
expressed as a ratio of ratios according to Eq 3 which contains
terms for the initial 18O/16O in O2•- at time 0 (R0) and the 18O/
16O of O2•- remaining in the solution (Rf) at a specific fractional
consumption of O2•- (f). R0 was determined by direct measure-
ment using the standard reaction of CoIII(SMDPT) (see
Experimental Methods). Rf was determined from the 18O/16O
of the O2 produced (RO ) at a specific fractional yield (f) using
2
eq 4.
The low-temperature, stopped-flow results are consistent with,
although do not prove, that the same 1:1 Cu(TMPA):O2
intermediate is the primary product of the reaction between
Rf
16,16k/18,16k)-1
) (1 - f)1/(
(3)
(4)
R0
CuII(TMPA) and O2 as well as the reaction between
•-
CuI(TMPA) and O2. The possibility that the 1:1 Cu(TMPA):
O2 adduct is a secondary product of one or both of these
reactions cannot be excluded on the basis of our stopped-flow
measurements because the rate of the intermediate’s formation
R0 ) Rf(1 - f) + RO (f)
2
Isotope fractionation plots for the oxidation of O2•- by copper(II)
complexes are shown in Figure 5. The data are well-fitted by
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3689

A R T I C L E S
Table 3. Competitive 18O KIEs on Reactions of O2 with Copper(I) and Copper(II) Complexes
Smirnov and Roth
•-
isotope effecta
(
CuI(TEPA)
0.9808 ( 0.0018
CuI(TMPA)
0.9818 ( 0.0014
16,16k/18,16k)I
CuII(TEPA)
0.9829 ( 0.0017
0.9871 ( 0.0062
CuIIN3(TEPA)
0.9834 ( 0.0023
0.9886 ( 0.0078
CuII(TMPA)
0.9824 ( 0.0011
0.9836 ( 0.0043
CuIIN3(TMPA)
0.9825 ( 0.0017
0.9839 ( 0.0058
(
(
16,16k/18,16k)net
16,16k/18,16k)II
a See text for description of terms.
Discussion
eq 3 indicating that neither the mechanism nor the rate-
determining step which gives rise to the 18O KIE changes as
the reaction progresses. In addition, the 18O KIE was found to
•-
Reactions of O2 with labile transition metal complexes
frequently occur at rates that approach the diffusion limit.16,17
Therefore, it is not surprising that the copper(II) complexes
studied here, which have open or easily substituted coordination
sites, react at rates too fast to be studied using low-temperature
stopped-flow spectrophotometry. The following discussion will
show how competitive 18O KIEs, together with an analysis of
the reaction thermodynamics, provide unique insight to the
mechanisms and barriers associated with transition-metal medi-
•-
be independent of the starting molar ratio of the O2 to the
copper(II) complexes from 3:1 to 8:1 as well as the addition
•-
of 18-crown-6 ether which destabilizes the K+O2 tight ion
pair.
Although small in magnitude, the competitive 18O KIEs are
determined very precisely through 10-15 independent measure-
ments. To be conservative, errors are reported as two standard
deviations about the mean calculated value rather than the much
smaller 2σ error from the nonlinear curve-fitted data. Although
the isotope ratio mass spectrometric analysis is typically precise
to ( 0.0002, slightly larger errors are indicated by the R0 )
1.01733 (( 0.00051) vs standard mean ocean water (SMOW),
which was determined for commercial (nonenriched) KO2 by
22 independent measurements. This error reflects the reproduc-
ibility of the reaction chemistry as well as the experimental
manipulations which include a combustion step (see Experi-
mental Methods).
•-
ated O2 oxidations. In addition, the characterization of
activated oxygen intermediates through comparisons of mea-
sured 18O isotope effects to theoretically calculated values will
be described.
I. Mechanism of Superoxide Oxidation. There are at least
four mechanisms which can be envisioned for the one-electron
•-
oxidation of O2 by copper(II) complexes to give O2 and the
related copper(I) species: (i) outer-sphere electron transfer (ET),
(ii) single-step, inner-sphere ET, (iii) multistep, inner-sphere ET
where the formation of a bound intermediate is rate-limiting,
and (iv) multistep, inner-sphere ET where the dissociation of a
bound intermediate is rate-limiting.
The simplest mechanism involves ET in a single-step. In the
context of Marcus theory, the transition state for an outer-sphere
ET reaction is an activated nuclear configuration of the reactants
and the surrounding solvent. In this configuration, there is no
formal bonding and only weak electronic coupling between the
reactants. In contrast, inner-sphere ET involves a transition state
with a bond between the reactants. With respect to the reactions
of interest, the rate-determining step for an inner-sphere ET
might involve simple binding of O2 to CuII, the dissociation
of O2 from a CuO2 intermediate, or a concerted reaction in which
both processes occur at the same time.
In practice, it is often difficult to distinguish inner-sphere from
outer-sphere ET reactions of O2, although Marcus Theory has
been used.86,87 According to the Marcus Cross Relation (eq 6)88
the rate constant (k12) for an outer-sphere ET can be expressed
in terms of the equilibrium constant (K12′) derived from ∆G°′,
a frequency factor (f12),89 and self-exchange rate constants (k11
and k22) for two degenerate reactions generalized as: *A + B
f *B + A; the * designates the initial site of the electron.
All of the reactions are characterized by a decrease in the
‘18O enrichment’ of the unreacted O2 as the fractional
•-
conversion to O2 increases (Figure 5). This observation indicates
that the heavier 18,16O2 reacts faster than the lighter 16,16O2
•-
•-
and, therefore, an inverse (<1) 18O KIE. Since the experiments
were performed in the presence of excess O2•-, contributions
from the follow-up reactions of CuI(TEPA) and CuI(TMPA)
are also considered. The KIEs on these reactions were deter-
mined using the same procedure outlined for the copper(II)
complexes. The data are well-described by eqs 3 and 4 (See
Supporting Information) again indicating that the isotopically
sensitive step does not change as the reaction progresses.
The 18O KIEs on the reactions of the copper(II) complexes
•-
•-
with O2 to form the related copper(I) complexes and O2 are
designated: (16,16k/18,16k)II. These values were estimated using
Eq 5 where (16,16k/18,16k)net is the observed isotope effect starting
with CuII(TEPA), CuIIN3(TEPA), CuII(TMPA), or CuIIN3-
(TMPA) and (16,16k/18,16k)I is the observed isotope effect starting
with CuI(TEPA) or CuI(TMPA) (Table 3). The equation
•-
accounts for the 3O2 to 2O2 stoichiometry determined from
the reaction yields (cf. Figure 3). The derived (16,16k/18,16k)II:
CuII(TEPA) ) 0.9871 ( 0.0062, CuIIN3(TEPA) ) 0.9886 (
0.0078, CuII(TMPA) ) 0.9836 ( 0.0043 and CuIIN3(TMPA)
) 0.9839 ( 0.0058 are the same within the 2σ limits of
propagated error. Importantly, the measured (16,16k/18,16k)I are
virtually identical (Table 3). Therefore, any deviation from the
theoretical relationships in eqs 3-5 has no bearing on the
relative correction to (16,16k/18,16k)net and the major conclusion
that the (16,16k/18,16k)II are indistinguishable for the reactions of
the different copper(II) complexes.
k12
)
k
× k22 × K12′ × f12
(6)
x
11
The Cross Relation has been noted to fail when rates of outer-
sphere ET approach the diffusion limit and when reactions occur
(86) Zahir, K.; Espenson, J. H.; Bakac, A. J. Am. Chem. Soc. 1988, 110 (0),
5059.
(87) Stanbury, D. M.; Haas, O.; Taube, H. Inorg. Chem. 1980, 19, 518.
(88) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(89) ln(f12) ) (lnK12′)2/4[ln(k11k22/Z2) + (w11 + w22)/RT]; w11 and w22 correspond
to the formation of the precursor complex in the self-exchange reactions
(
16,16k/18,16k)II ) 3(16,16k/18,16k)net - 2 (16,16k/18,16k)I (5)
and Z ) 1011 M-1 s-1
.
9
3690 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Cu−O2 Intermediates in Superoxide Oxidations
A R T I C L E S
by inner-sphere ET.90 For one of the least thermodynamically
0.0078. This small range is surprising given the large (ca. 15
kcal mol-1) variation in the reaction driving-forces. These results
are in contrast to others in the literature corresponding to outer-
sphere ET reactions. Gould and Farid reported kH/kD ) 1.25 (
0.04 and 1.30 ( 0.03 for the highly exergonic (∆G° ) -57.7
kcal mol-1) ET reactions in acetonitrile: TCA•- (TCA )
2,6,9,10-tetracyanoanthracene) + toluene•+ (d0 vs d8) and TCA•-
+ p-xylene•+ (d0 vs d10).100 Upon changing the ∆G° to -27.9
kcal mol-1, by substituting ClO (chlorine oxide) in place of
TCA, Simon and co-workers found kH/kD was dramatically
decreased to 1.0 ( 0.1 and to 0.67 ( 0.1 for the respective
hydrocarbons.95,96
favorable oxidations of O2•- by CuIIN3(TMPA), a value of k12
) 2.4 × 104 M-1s-1 is calculated using aqueous self-exchange
rate constants: k11 ) 4.5 × 102 M-1s-1 for O2 + O2
and
•-91
k22 ) 70 M-1s-1 for CuII(TMPA) + CuI(TMPA),92 K12′ )
3.5 × 104 (from ∆G°′ in Table 2) and f12 ) 0.5.89 The use of
aqueous self-exchange rate constants is expected to give a cross-
reaction rate constant in DMSO to within an order of magnitude;
this is based on calculated values for the electrostatic work terms
and the reorganization energy.93 The large (104-fold) discrepancy
between the predicted rate constant and the observed k12 g 1
× 108 M-1s-1 indicates that the reaction does not occur by an
outer-sphere ET. However, this conclusion is not so compelling
in the case of CuII(TEPA) for which k12 ) 1.8 × 107 M-1s-1
is calculated and k12 > 1 × 108 M-1s-1 is observed.
In a more closely related study by McLendon and co-workers,
2+
16
18
the kH O/kH O upon outer-sphere ET oxidation of Fe(H2O)6
2
2
to Fe(H2O)6 was measured. The 18O KIEs were found to
decrease from 1.08 ( 0.01 to 1.04 ( 0.01 as the driving-force
changed from ∆G° ) -11.5 kcal mol-1 to ∆G° ) -17.3 kcal
3+
II. Isotope Effects on Outer-Sphere Electron Transfer.
Heavy atom KIEs on outer-sphere ET reactions originate largely
from the reorganization of isotopic bonds upon converting the
reactant into the product state. This contribution to the inner-
shell reorganization free energy (λi) can be formulated either
semiclassically or quantum mechanically. Within the semi-
classical limit, the isotope effect arises from an over-the-barrier
transition and the influence of zero point energy differences as
well as the thermal populations of excited vibrational states. In
the quantum mechanical limit, the isotope effect arises from
mass-dependent Franck-Condon factors which reflect a through-
the-barrier transition due to nuclear tunneling.
mol-1 97
. Although the range of driving-forces is comparable to
the range in present study, the 18O KIEs are normal (>1) and
vary significantly by ∼4%, compared to 0.5% in the present
•-
work. In addition, the conversion of O2 to O2 and the
conversion of Fe(H2O)62+ to Fe(H2O)63+ are both characterized
by strengthening of the bond(s) (i.e., an increase in force
constant) in the product states. The reactions of the iron
complexes occur by outer-sphere ET and, because of the
frequencies of the bonds involved, the 18O KIEs are best
explained within the quantum mechanical limit. We have
previously shown that a quantum mechanical formalism can be
used to predict normal 18O KIEs on outer-sphere ET reductions
of O2 to O2•- in the roughly thermoneutral regime.94 In light of
the results from earlier studies, the 18O KIEs on O2•- oxidation
by copper(II) complexes, which are both inverse and indepen-
dent of reaction driving-force, appear to be inconsistent with
an outer-sphere ET mechanism.
In both the semiclassical and quantum mechanical limits, the
reaction thermodynamics is expected to influence the KIEs.94-98
Semiclassically, ∆G°′ dictates the position of the isotope-
dependent barrier defined by the intersection of the reactant and
product surfaces. In this view, a normal KIE arises when the
force constant characterizing the product is less than that for
the reactant; whereas an inverse KIE arises when the force
constant of the product is greater than that of the reactant.
Quantum mechanically, ∆G°′ modulates the KIE through the
isotope-dependent overlap of vibrational wave functions as-
sociated with the reactant and product states. When the ∆G°′
is very favorable, vibrational states of the reactant can overlap
with excited vibrational states of the product. It follows that a
higher density of states associated with the heavy isotope relative
to light one can cause the KIE to be inverse despite the
invariably normal (> 1) contributions from Franck-Condon
factors.96,99
O2 binding reactions belong to a general class of inner-sphere
ET reactions which may or may not involve the formation of
intermediates with detectable lifetimes. We have previously
reported large variations in 18O KIEs, from 1.0273 to 1.0072,
for concerted O2 oxidative addition reactions that afford η2-
peroxo compounds from reduced late-transition metal com-
plexes.47 These reactions occur without any intermediates being
formed. In addition, a trend of decreasing 18O KIEs in response
to the increased bimolecular rate constants for O2 binding was
observed. The trend has been attributed to variations in transition
state structure in response to changing the reaction driving-force
The 18O KIEs measured for O2•- oxidations are inverse and
fall within a narrow range from 0.9836 ( 0.0043 to 0.9886 (
over an estimated -5.6 to -20 kcal mol-1 47,59
Because of the
.
similar thermodynamics, a marked variation in transition state
structure would be expected for the O2•- oxidation reactions if
they too were to occur by a one-step, inner-sphere ET pathway.
Therefore, the lack of a discernible driving-force dependence
of the 18O KIEs in the present work argues against this type of
mechanism.
(90) Wherland, S. Coord. Chem. ReV. 1993, 123, 169.
(91) Lind, J.; Shen, X.; Merenyi, G.; Jonsson, B. O. J. Am. Chem. Soc. 1989,
111, 7654.
(92) Yandell, J. K. In Copper Coordination Chemistry: Biochemical and
Inorganic PerspectiVe; Karlin, K. D., Zubieta, J., Eds.; Adenine Press:
Guilderland, NY, 1983; p 157.
(93) The overall reorganization (λ), the sum of inner (λin) and outer-shell (λout
)
contributions, is derived from bimolecular rate constants according to k )
1011M-1s-1exp-(λ/4kBT). The calculated λin ) 15.9 kcal mol-1 for O2/O2
•-
III. Reaction Coordinate Analysis. Alternative mechanisms
(ref 91) and λin ) 29.6 kcal mol-1 for CuII/CuI (ref 92). The following
static (Ds) and optical dielectric constants (Do) were used to estimate the
impact of changing λout upon k: Ds ) 47.24, Do ) 2.1883 for DMSO and
Ds ) 80.10, Do ) 1.7778 for water (ref 81).
•-
involve two steps where binding of O2 occurs first followed
by the loss of O2 from a CuO2 intermediate. By definition, the
competitive (16,16k/18,16k)obs reflects all elementary steps along
the reaction coordinate from initial encounter of copper(II)
(94) Roth, J. P.; Wincek, R.; Nodet, G.; Edmondson, D. E.; McIntire, W. S.;
Klinman, J. P. J. Am. Chem. Soc. 2004, 126, 15120.
(95) Doolen, R.; Simon, J. D.; Baldridge, K. K. J. Phys. Chem. 1995, 99, 9,
13938.
•-
complex with O2 up to and including the transition state for
(96) Doolen, R.; Simon, J. D. J. Am. Chem. Soc. 1994, 116, 1155.
(97) Guarr, T.; Buhks, E.; McLendon, G. J. Am. Chem. Soc. 1983, 105, 3763.
(98) Buhks, E.; Bixon, M.; Jortner, J. J. Phys. Chem. 1981, 85, 3763.
(99) Avouris, P.; Gelbart, W. M.; El-Sayed, M. A. Chem. ReV. 1977, 77, 793.
the first kinetically irreversible step. Therefore, the isotope effect
(100) Gould, I. R.; Farid, S. J. Am. Chem. Soc. 1988, 110, 7883.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3691

A R T I C L E S
Smirnov and Roth
these factors increase the contribution of the CuI-O2 reso-
0
nance form for the TEPA complexes but not so much for the
TMPA complexes. Loss of electrostatic binding energy may
be enough to destabilize the 1:1 Cu(TEPA):O2 relative to the
1:1 Cu(TMPA):O2 such that the former is unobservable in the
low-temperature stopped-flow experiments,101 yet not to the
•-
extent that the binding of O2 is kinetically irreversible.
A two-step mechanism involving reversible formation of a
CuO2 intermediate is consistent with the invariance of the 18O
KIE over a wide range of ∆G°. This type of behavior would
occur in the event that (16,16k/18,16k)obs were largely determined
by (16,16K/18,16K)1; i.e., if (16,16k/18,16k)2 were very small. By
definition, (16,16K/18,16K)1 is independent of the reaction thermo-
dynamics. This interpretation is analogous to those put forth in
the organometallic chemistry literature to explain inverse
primary H/D KIEs on alkane elimination from metal(alkyl)-
(hydride) compounds via σ bonded intermediates.102
•-
Figure 6. Reaction coordinate diagrams for the oxidation of O2 by
copper(II) complexes.
for a two-step reaction may be expressed as: (16,16k/18,16k)obs
16,16K/18,16K)1(16,16k/18,16k)2. The (16,16K/18,16K)1 ) (16,16k/18,16k)1/
16,16k/18,16k)-1 for the pre-equilibrium formation of a CuO2
)
(
(
intermediate. The (16,16k/18,16k)2 corresponds to the first irrevers-
ible step where the intermediate is converted to products. If the
initial binding step were kinetically irreversible, then regardless
of the isotopic sensitivity of the second step, the following
relationship would hold: (16,16k/18,16k)obs ) (16,16k/18,16k)1.
The indistinguishable 18O KIEs observed for O2•- oxidation
by the four copper(II) complexes imply that the same step is
rate-determining in each of the reactions. It is unlikely that this
In the present experiments, the major change in force constant
is believed to occur upon binding of O2•- in the first step, where
some degree of internal ET results in strengthening of the O-O
bond and the formation of a Cu-O bond. The observation of a
small (16,16k/18,16k)2 would be consistent with an early transition
state for the dissociation of O2 requiring very little O-O
reorganization. This type of transition state structure is expected
based on the favorable reaction driving-force. A negligible
•-
is an initial O2 binding step for the following reasons. (i)
Indistinguishable 18O KIEs are observed for the complexes with
and without azide bound as a fifth ligand as well as complexes
of dramatically different electrophilicity. The disparate energies
of the putative CuO2 intermediates should affect the position
of the transition state and in turn the 18O KIEs. (ii) The
magnitude of the 18O KIE is expected to be negligible for an
(
16,16k/18,16k)2 could also result if a small inverse isotope effect
on the elimination of O2 were offset by a small normal isotope
effect on the reaction coordinate frequency,103 i.e., the bond that
is lost upon converting a vibration to a translation in the
transition state. In this view, a negligible isotope effect on the
transition state for O2 dissociation may occur even if the electron
has only partially transferred prior to the rate-determining step.
Regardless of the physical origin of the effect, the driving-force
independence of the experimental results is most conveniently
explained by (16,16k/18,16k)obs being dominated by the (16,16K/
18,16K)1 which describes the pre-equilibrium formation of the
CuO2 species.
IV. Calculation of Equilibrium 18O Isotope Effects. The
measured 18O KIEs, taken to largely reflect pre-equilibrium
formation of a CuO2 intermediate, are consistent with our
calculations of equilibrium isotope effects (EIEs). The 18O EIEs
were calculated for the binding of O2•- to copper(II) complexes
and the formation of the proposed monomeric products in Table
4. These products were actually generated by reacting various
copper(I) species with O2. The approach, which employs
Bigeleisen’s formalism to calculate isotope effects, has been
previously described.104-107 The equations used are provided
in Supporting Information.
•-
electrostatic interaction with O2 where electron transfer to
copper(II) has not occurred to a significant extent (see below).
Thus, on the grounds that the 18O KIEs are significantly inverse
as well as invariant to copper(II) complex coordination geom-
etry, electrophilicity, and redox potential, a two step-mechanism
involving rate-determining formation of a CuO2 intermediate
is considered improbable.
The remaining and favored mechanism involves formation
of a CuO2 intermediate prior to the rate-determining dissociation
of O2. Reaction coordinate diagrams corresponding to this
mechanism for the different ligand complexes are shown in
Figure 6; the ‡ designates the transition state which controls
•-
the disappearance of O2 and formation of O2. The stopped-
flow studies suggest that the CuO2 intermediate is stabilized
relative to the separated reactants in the case of the TMPA
complexes but destabilized for the TEPA complexes such that
its concentration is undetectable. While the spectral data suggest
that the intermediate is a Cu(TMPA):O2 adduct, identical to that
generated from CuI(TMPA) and O2, an analysis of the kinetic
mechanism has not been possible. The time scales of the low-
temperature O2•- and the O2 reactions are faster than the mixing-
time of the stopped-flow instrument. However, the indistin-
guishable 18O KIEs indicate that structurally similar CuO2
intermediates are formed from the reactions of the copper(II)
TEPA and TMPA complexes.
That the CuO2 intermediate derived from CuII(TEPA) is
thermodynamically less stable than the intermediate derived
from CuII(TMPA) requires some discussion. The relative
stability is likely affected by the redox potentials of the
copper(II) centers as well as the ligand environment. Both of
Either two or three isotopically sensitive vibrations corre-
sponding to the νO-O and νCu-O of the copper complexes were
•-
considered along with the isotopic νO-O of O2 and O2. For
(101) This explanation may be extended to account for the lack of trapping of
the putative Cu(TEPA):O2 intermediate by copper(I) with formation of
the dicopper peroxo product.
(102) Jones, W. D. Acc. Chem. Res. 2003, 36, 140.
(103) Melander, L.; Saunders: W. H., Jr. Reaction Rates of Isotopic Molecules;
Wiley: New York, 1983.
(104) Slaughter, L. M.; Wolczanski, P. T.; Klinckman, T. R.; Cundari, T. R. J.
Am. Chem. Soc. 2000, 122, 7953.
(105) Bender, B. R.; Kubas, G. J.; Jones, L. H.; Swanson, B. I.; Eckert, J.;
Capps, K. B.; Hoff, C. D. J. Am. Chem. Soc. 1997, 119, 9179.
(106) Bender, B. R. J. Am. Chem. Soc. 1995, 117, 11239.
(107) Abu-Hasanayn, F.; Krogh-Jespersen, K.; Goldman, A. S. J. Am. Chem.
Soc. 1993, 115, 8019.
9
3692 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Cu−O2 Intermediates in Superoxide Oxidations
A R T I C L E S
•-
•-
Table 4. Calculated Equilibrium (18O) Isotope Effects for the Oxidation of O2 to O2 and the Binding of O2 to Copper(II) Complexes
a All complexes were synthesized from copper(I) complexes and O2 at low temperatures. Names and descriptors are in accord with the original literature:
[Cu(TMG3tren)O2]+ - tris[2-(tetramethylguanidino)ethyl]aminocopper superoxide, [Cu(LMe,Bn)(O2)]+ - tris(N-benzyl-N-methylaminoethyl)aminocopper
superoxide, 2LCu(O2) - 2-[(4,7-diisopropyl)-1,4,7-triazacyclononylmethyl]-4,6-di-tert-butylphenoxy copper superoxide, L3CuO2 - tris(3-tert-butyl-5-isopropyl-
1-pyrazolyl)hydroboratocopper superoxide, (1)Cu-â-diketiminate(O2) - N,N′-bis(2,6-diisopropylphenyl)-1,3-di-tert-butyl-â-diketiminatocopper superoxide,
(2)Cu-â-diketiminate(O2) - N,N′-bis(2,6-diisopropylphenyl)-1,3-dimethyl-â-diketiminatocopper superoxide. b The O-O stretch in free or complexed O2.
•-
For free O2
: .
16,16ν ) 1064.8 cm-1 and 18,16ν ) 1034.8 cm-1 110 c The Cu-O stretch for superoxo and peroxo complexes. For side-on complexes both the
symmetric and asymmetric stretches are given. d The 18O equilibrium isotope effect calculated using Bigeleisen’s formalism.110 e Ref.111 f Estimated from
1/2
)
O
.
g Estimated from the measured νCu- h Ref 39. i Computed value.39 j Ref 112. k The more inverse EIE was estimated
.
18O-18
O
18O-16
ν
16O-16
O ) (ν
18O-18
O × ν
113
assuming that the isotope shift of the asymmetric νCu-O was equal to that of the symmetric νCu-O
.
The less inverse EIE was estimated by neglecting
isotope shift of the asymmetric νCu-O.
l Ref 114. m Ref 115. n Using either the frequencies and isotope shifts (given in parentheses): 462(-10) and 451(-9)
cm-1 for the isostructural PtO2(PPh3)2 (PPh3 ) triphenylphosphine) reported in ref 113 or those from entry 5. o The normal EIE was obtained using νCu-O
from entry 5 where coupling to Cu-ligand modes is apparent. The inverse EIE is a conservative lower limit which assumes idealized isotope shifts for pure
113
symmetric and asymmetric νCu-O
.
the proposed η1-complexes, the low-frequency bend (δCu-O
)
There is additional uncertainty associated with the assumption
of pure O-O and Cu-O modes and the use of theoretical
isotope shifts. When coupling to Cu-N or ligand C-N modes
of the same symmetry and with similar force constants occurs,
the stretching frequencies may be perturbed in such a way that
the oxygen isotope shift is reduced relative to the theoretical
value.109 Depending on the degree of coupling, the 18O EIEs
may be significantly less inverse than in the absence of such
interactions. This type of effect is apparent in the experimental
frequencies in entry 5, where the isotope shifts of νO-O and
νCu-O are lower than the idealized, theoretical values.
2
is reported to be relatively insensitive to isotope substitution108
and was, therefore, neglected. In addition, calculations on the
proposed end-on structures assumed equal populations of Cu-
18O-16O and Cu-16O-18O. This simplification can have the
effect of making the calculated 18O EIE slightly less inverse
than the observed value. When stretching frequencies for the
mixed-labeled isotopologues were unavailable, they were cal-
culated from the data for the 16O-16O and 18O-18O containing
species (entries 2 and 4 in Table 4). In some cases, it was
necessary to assume values reported for isostructural compounds
in the literature (entries 4, 6, and 7), and, therefore, a range of
values is reported.
The 18O EIEs are formulated in terms of the change in force
constant upon converting the reactant state to the product state.
(108) Odo, J.; Imai, H.; Kyuno, E.; Nakamoto, K. J. Am. Chem. Soc. 1988,
110, 742.
(109) For a leading reference, see: Kinsinger, C. R.; Gherman, B. F.; Gagliardi,
L.; Cramer, C. J. J. Biol. Inorg. Chem. 2005, 10, 778.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3693

A R T I C L E S
Smirnov and Roth
As described in Results Section II, an increase results in an
inverse isotope effect and a decrease results in a normal isotope
effect. Several of the 18O EIEs calculated for the formation of
CuO2 species from O2•- (Table 4) are inverse but significantly
EIEs will be discernibly larger for forming side-on copper
superoxo structures (∼1.032) and larger still for forming side-
on copper peroxo structures (∼1.039) from copper(I) complexes
and O2.
•-
less so than the 18O EIE of 0.968 associated with oxidizing O2
It is interesting to compare the 18O KIEs determined for O2
reactions of enzymes with those calculated for reactions of
biomimetic compounds. PHM and the structural homologue
dopamine â-monooxygenase (DâM) exhibit competitive 18O
KIEs of 1.0212 ( 0.0018 and 1.0256 ( 0.0003, respectively.52
In both of these enzymes, there is strong evidence of a superoxo
intermediate which is further reduced to the level of a peroxo
species in50 or before38 the rate-determining step (see Introduc-
tion). The closeness of the observed 18O KIEs for the enzyme
reactions to the 18O EIEs estimated for the end-on copper
superoxo intermediates is consistent with a reversible O2 binding
step prior to substrate oxidation. As proposed in the present
study, the 18O EIE on the pre-equilibrium O2 binding step may
be the dominant contributor to the 18O KIEs in the enzyme
reactions. We hope to clarify this issue as well as the effects of
subsequent redox chemistry on the 18O KIEs using structurally
defined CuO2 complexes which react by ET or hydrogen atom
transfer mechanisms.
•-
to O2. This result may be attributed to partial oxidation of O2
,
due to the covalent sharing of an antibonding electron between
copper and oxygen, and the strengthening of the O-O bond.
•-
In contrast, the partial reduction of bound O2 populates an
antibonding orbital which weakens the O-O bond.
The calculated 18O EIEs for the proposed end-on superoxo
(0.989-0.995), side-on superoxo (0.988-0.999) and side-on
peroxo (0.994-1.006) structures reflect contributions from both
internal ET and the appearance of new bonds between copper
and oxygen. If the O-O force constant were the primary
determinant of the EIE, then a trend from inverse to normal
would be expected for a continuum from superoxo to peroxo
type complexes.31 The reason is that the formation of new Cu-O
bonds can only increase the force constant of the product state
and, therefore, contribute to an inverse EIE. For the peroxo-
like (formally CuIII-O2-II) structures, normal 18O EIEs are
estimated when, together, the isotope shifts of the symmetric
and asymmetric νCu-O total e 11 cm-1 31 In contrast, the 18O
.
This study has provided a method and series of benchmarks
which can now be applied to studying how superoxide is
processed by antioxidant enzymes such as Cu:Zn SOD (see
above). A recent computational study of this enzyme116 has
EIEs for the superoxo structures (formally CuII-O2-I) are
invariably inverse, consistent with the synergetic effects of
strengthening of the O-O bond (relative to free O2•-) and
forming new Cu-O bonds.
-I
predicted the formation of a formally CuII-O2 intermediate
V. Structure of the Cu-O2 Intermediate and Implications
for Biological Systems. To shed light on the structures of the
proposed CuO2 intermediates formed from the reactions of
copper(II) complexes and O2•-, the (16,16k/18,16k)obs which
approximate (16,16K/18,16K)1 are compared to the 18O EIEs
calculated for structurally characterized compounds (Table 4).
The (16,16k/18,16k)obs of 0.9836-0.9886 are consistent with
that estimated for the proposed end-on superoxo complex:
[Cu(LMe,Bn)O2]+.111 While we stress that we cannot discern end-
on from side-on superoxo structures based on the anticipated
18O EIEs, it appears that a peroxo structure would have a
value outside of the range observed. Furthermore, similar
types of end-on superoxo structures for [Cu(LMe,Bn)O2]+ and
[Cu(TMPA)O2]+ would be consistent with the complexes having
the same overall charges and geometries enforced by the neutral
tripodal ligands.
in the rate-limiting step for O2 formation. This mechanism is
fundamentally different than the one proposed here, where the
rate-determining step is O2 dissociation from a CuO2 intermedi-
ate. We are currently examining the nature of the transition state
in the Cu:Zn SOD reaction through measurements of 18O KIEs
on the oxidative and reductive processes.
Conclusions
A two-step, inner-sphere electron-transfer mechanism is
•-
proposed to describe the oxidation of O2 to O2 by coor-
dinatively unsaturated copper(II) complexes. Kinetic measure-
ments using stopped-flow UV-vis spectrophotometry indicate
lower limits to the second-order rate constants of g1 × 108
M-1s-1 at 20 °C which in some cases is up to 3 orders of
magnitude faster than that expected from application of the
Marcus Theory Cross Relation. On these grounds, an outer-
sphere electron-transfer mechanism seems unlikely. In addition,
low-temperature stopped-flow studies of the least oxidizing
copper(II) complexes reacting with O2•- have provided evidence
of a CuO2 intermediate.
The thermodynamic relationship which exists between 18O
equilibrium isotope effects (EIEs), allows 18O EIEs of 1.016 to
1.021 to be estimated in the direction of end-on O2 binding to
copper(I) complexes; this is accomplished by dividing the
experimental values for O2•- binding copper(II) complexes by
More compelling evidence for the formation of CuO2
intermediates comes from studies of competitive oxygen kinetic
isotope effects. These are, to our knowledge, the first measure-
the theoretical 18O EIE for converting O2 to O2 (Table 4).
•-
The values obtained are significantly larger than the 18O EIEs
of 1.0039-1.0046 reported for end-on O2 binding to myoglobin
and hemoglobin, the small size of which has been attributed to
hydrogen bonding to the terminal oxygen in the O2 carrier
proteins.110 We have calculated somewhat larger 18O EIEs of
1.022-1.028 directly from the stretching frequencies of O2 and
the proposed end-on copper superoxo complexes (Table 4). The
ments to be performed on reactions of O2•-. The 16,16k/18,16
k
fall within a narrow range of 0.9836 to 0.9886 for reactions
with copper(II) complexes. These complexes exhibit different
(112) Chen, P.; Root, D. E.; Campochiaro, C.; Fujisawa, K.; Solomon, E. I. J.
Am. Chem. Soc. 2003, 125, 466.
(113) Nakamura, A.; Tatsuno, Y.; Yamamoto, M.; Otsuka, S. J. Am. Chem.
Soc. 1971, 93, 6052.
(114) Jazdzewski, B. A.; Reynolds, A. M.; Holland, P. L.; Young, V. G.; Kaderli,
S.; Zuberbuehler, A. D.; Tolman, W. B. J. Biol. Inorg. Chem. 2003, 8,
381.
calculation further suggests that in favorable situations the 18
O
(110) Tian, G.; Klinman, J. P. J. Am. Chem. Soc. 1993, 115, 8891.
(111) Komiyama, K.; Furutachi, H.; Nagatomo, S.; Hashimoto, A.; Hayashi,
H.; Fujinami, S.; Suzuki, M.; Kitagawa, T. Bull. Chem. Soc. Jpn. 2004,
77, 59.
(115) Spencer, D. J. E.; Aboelella, N. W.; Reynolds, A. M.; Holland, P. L.;
Tolman, W. B. J. Am. Chem. Soc. 2002, 124, 2108.
(116) Pelmenschikov, V.; Siegbahn, P. E. M. Inorg. Chem. 2005, 44, 3311.
9
3694 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Cu−O2 Intermediates in Superoxide Oxidations
A R T I C L E S
coordination geometries and redox potentials which vary by 0.67
V. The observations of significantly inverse and driving-force
independent kinetic isotope effects are inconsistent with simple
O2•- binding as well as one-step electron-transfer mechanisms.
Instead the results point toward a two-step mechanism where a
CuO2 intermediate is formed in a pre-equilibrium step prior to
the rate-limiting loss of O2. The small range of isotope effects
implies that the CuO2 intermediates formed from the different
ligand complexes are structurally similar.
oxygen isotope effects on reactions of biomimetic compounds
to identify activated oxygen intermediates during catalysis by
metalloenzymes.
Acknowledgment. We are grateful for financial support from
NSF CAREER Award No. 0449900 (J.P.R.) and to Prof.
Kenneth D. Karlin for access to the low-temperature stopped-
flow instrument.
Supporting Information Available: Experimental details and
the description of equilibrium isotope effect calculations. This
material is available free of charge via the Internet at
http://pubs.acs.org.
The resemblance of the CuO2 intermediates formed from
•-
reactions of copper(II) complexes and O2 to those generated
from copper(I) and O2 is supported by isotope effects calculated
from stretching frequencies of structurally characterized CuO2
adducts. These findings lay a foundation for using competitive
JA056741N
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3695