Structures of transition states in metal-mediated O2-activation reactions

Article abstract of DOI:10.1002/anie.200502096

(Chemical Equation Presented) Kinetic isotope effects (KIEs) have been used to probe the mechanism of the binding of O₂ to classic inorganic compounds. The intermolecular ¹⁸O KIEs decrease with increasing rate constants, and the corr

Full text of DOI:10.1002/anie.200502096

Angewandte
Chemie
Oxygen Activation
DOI: 10.1002/anie.200502096
Structures of Transition States in Metal-Mediated
O2-Activation Reactions**
Michael P. Lanci, David W. Brinkley, Kristie L. Stone,
Valeriy V. Smirnov, and Justine P. Roth*
Enrichment of heavy oxygen (18O) from its level at natural
abundance is widely used to probe O2-dependent processes in
the atmospheric and biological sciences.[1–4] Here we describe
the first application of this technique to O2-binding reactions
of classic inorganic compounds. Competitive 18O kinetic
isotope effects (KIEs)[5–7] reflecting the ratio of second-
order rate constants for the formation of h2-peroxide com-
pounds from 16,16O2 and 18,16O2 (k16,16/k18,16) have been deter-
mined (Scheme 1). Related measurements on enzymatic
Scheme 1. Competitive 18O isotope effect on the formation of h2-
peroxo complexes.
reactions[4,8–11] have been interpreted on the basis of equilib-
rium isotope effects (EIEs) and simple models that neglect
variations in the transition-state structure. The present studies
suggest that the proposed enzyme mechanisms may need to
be reevaluated in light of the sensitivity of 18O KIEs to
changes in the height of the free-energy barrier for reactions
involving the formation of a metal–O2 bond.
Oxygenation of solutions prepared from trans-[IrCl(CO)-
(PPh3)2], [Ir(dppe)2]Cl, [Pt(PPh3)4], [Pd(PPh3)4], [Ni(PPh3)4],
and [Ni(CNtBu)4] (PPh3 = triphenylphosphine, dppe = 1,2-
bis(diphenylphosphino)ethane, CNtBu = tert-n-butylisocya-
[*] M. P. Lanci, Dr. D. W. Brinkley, K. L. Stone, Dr. V. V. Smirnov,
Prof. J. P. Roth
Department of Chemistry
The Johns Hopkins University
3400 North Charles Street, Baltimore, MD 21218 (USA)
Fax: (+1)410-516-8420
E-mail: jproth@jhu.edu
[**] We are grateful for support from NSF CAREER award CHE-0449900.
Supporting information (experimental details and isotope-effect
calculations) for this article is available on the WWW under http://
www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 7273 –7276
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7273

Communications
nide) affords the corresponding h2-peroxides [IrCl(CO)O2-
PPh3, the disappearance of [Ni(PPh3)4] depends only slightly
on the concentration of O2. This behavior is consistent with
dissociation of [Ni(PPh3)3] in a kinetically irreversible manner
caused by fast trapping of [Ni(PPh3)2] by O2. The addition of
PPh3 (0.15–0.34m) causes ligand association to compete with
(PPh3)2], [IrO2(dppe)2]Cl, [PtO2(PPh3)2], [PdO2(PPh3)2],
[NiO2(PPh3)2], and [NiO2(CNtBu)2].[12–16] On the basis of
early studies on trans-[IrCl(CO)(PPh3)2] and [Pt(PPh3)4],[13,14]
concerted oxidative-addition mechanisms have been pro-
posed in which two electrons transfer to O2 at the same time
that two metal–oxygen bonds form.[17,18]
This work describes O2-binding mechanisms for a series of
compounds in “anhydrous” N,N-dimethylformamide (DMF)
and dimethylsulfoxide (DMSO).[19] In most cases, high yields
of stable products were obtained, as revealed by 31P{1H} and
1H NMR as well as UV/Vis spectroscopy.[20] Kinetics studies
were performed by monitoring the disappearance of the
reduced compound (0.05–0.35 mm) upon reaction with a
greater than tenfold excess of O2, either by conventional or
stopped-flow UV/Vis spectrophotometry. The irreversible
formation of h2-peroxide products on the time scale of the
experiments is reported in terms of the rate constant for O2
O2 binding and allows the determination of kO (230 ꢀ
2
10mꢁ1 sꢁ1) for [Ni(PPh3)3]. Spectroscopic evidence suggests
that [Ni(PPh3)2(DMSO)] is the reactive species in DMSO (see
Supporting Information), where the reaction is first order in
O2 without added PPh3, and kO = 3300 ꢀ 80mꢁ1 sꢁ1. The
2
kinetics of the oxygenation reaction of [Ni(CNtBu)4] give a
linear first-order plot with a nonzero y-intercept, which
suggests that the rates of ligand dissociation and O2 binding
are comparable, that is, k1 ꢂ k3[O2]. Addition of CNtBu
causes the intercept to approach zero but the slope, which
corresponds to kO = 3300 ꢀ 150mꢁ1 sꢁ1, remains unchanged;
2
this is consistent with [Ni(CNtBu)3] being the reactive species.
The 18O KIEs for the O2-binding reactions were deter-
mined by a method in which stable-isotope mass spectrometry
is used.[1,4] Experiments were performed under conditions of
intermolecular competition by introducing a substoichiomet-
ric amount of metal compound into a sealed reaction vessel
containing natural-abundance O2. Subsequent to the reaction,
isolation and analysis of the O2 remaining in solution revealed
18O enrichment as a result of a faster consumption of 16,16O2
than 18,16O2. Data were collected at varying concentrations of
metal complex to determine k16,16/k18,16 (Figure 1). The 18O
binding (kO ) in DMF at 25 ꢀ 28C unless noted.
2
Depending on the stability of the reduced compounds in
Scheme 1, the reaction mechanisms range from simple to
complex, thus making it a challenge to probe the relative
energy barriers to O2 activation. Bimolecular kinetics
observed for trans-[IrCl(CO)(PPh3)2][21] and [Ir(dppe)2]Cl
indicate kO = 0.17 ꢀ 0.02mꢁ1 sꢁ1 and 2.8 ꢀ 0.4mꢁ1 sꢁ1, respec-
2
tively. The slowest reacting complex, trans-[IrCl(CO)-
(PPh3)2], has the lowest O2 affinity (DG298K = ꢁ5.6 ꢀ
0.2 kcalmolꢁ1 in DMF). No sign of reversible O2 loss could
be detected for any of the other peroxide compounds
examined.[22] [Pt(PPh3)4], [Pd(PPh3)4], [Ni(PPh3)4], and [Ni-
(CNtBu)4] dissociate in solution to two- and three-coordinate
species[14,16,23] which are oxidized 102–105 times faster than
trans-[IrCl(CO)(PPh3)2]. In the simplest case involving a d10
metal, disappearance of [Pt(PPh3)4] is first order in O2 and the
derived rate constant (kO = 32 ꢀ 8mꢁ1 sꢁ1) is independent of
2
the PPh3 concentration. This behavior is consistent with
[Pt(PPh3)3] being the primary reactive species.[14]
The oxidation of [Pd(PPh3)4] is more than 103 times faster
than that of [Pt(PPh3)4] and exhibits more complex kinetics.
The reactions are first order with respect to O2, yet the rate
constant (kO = 41100 ꢀ 300mꢁ1 sꢁ1) decreases in response to
2
Figure 1. 18O enrichment plot. Data are fitted to the equation:
added PPh3 until a limiting kO value of 2480 ꢀ 130mꢁ1 sꢁ1 is
2
16,16/k18,16
Rf/R0 =(1ꢁf)((1/(k
))ꢁ1); R0 and Rf are the 18O/16O ratios prior to
reached at a PPh3 concentration of 0.025–0.08m. This
behavior is consistent with a two-term rate law and compet-
itive binding of O2 to [Pd(PPh3)2] and [Pd(PPh3)3] (Scheme 2).
reaction and at a fractional conversion (f), respectively.
The first term in the rate law simplifies to kO for [Pd(PPh3)2]
2
when the reduced species are present in rapid equilibrium and
KIEs are independent of both the fractional O2 consumption
and the ratio of O2 consumed per mole of reduced compound.
This ratio was 1:1 in most cases but as high as 1.6 ꢀ 0.4:1 for
[Pt(PPh3)4] and 2.3 ꢀ 0.8:1 for [Ni(PPh3)4] because of the
catalytic oxidation of PPh3.[15,24] The k16,16/k18,16 ratio is a
competitive effect and serves as a probe of all the steps from
the encounter of the metal complex with O2 to the first
kinetically irreversible step. Thus, the KIEs in this work
reflect O2 binding concomitant with reduction to peroxide
(O22ꢁ) and are independent of ligand dissociation and
subsequent ligand oxidation.
kꢁ1[L] @ k2[O2], whereas the second term gives kO for
2
[Pd(PPh3)3] at high PPh3 concentration.
Reactions of [Ni(PPh3)4] and [Ni(CNtBu)4] can also be
understood in terms of Scheme 2. In the absence of added
The formation of h2-peroxide compounds is characterized
by k16,16/k18,16 values from 1.0069 to 1.0268 (Table 1). The 18O
Scheme 2. Kinetic scheme for O2 binding.
7274
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7273 –7276

Angewandte
Chemie
Table 1: Rate constants for O2 binding and 18O isotope effects (ꢀ2s) at 25ꢀ28C.[a]
Product
kO [mꢁ1 sꢁ1
]
k16,16/k18,16
nOꢁO, sym nMꢁO, asym nMꢁ
(
(K16,16/K18,16)
calcd
O
2
ꢁ1 [f]
16,16nꢁ18,16n) [ cm ]
trans-[IrCl(CO)O2(PPh3)2]
[IrO2(dppe)2]Cl
[PtO2(PPh3)2]
0.17ꢀ0.02
2.8ꢀ0.4
1.0268(37)[e]
1.0205(36)[e]
1.0251(51)
858(24), 581(9), 595(9) [g]
843(25), 547(9), 533(9) [g]
827(23), 470(10) 488(10)
–
1.028
1.028
1.031
1.028[h]
1.028[h]
1.026
1.030
32ꢀ7
[NiO2(PPh3)2]
>230[c]
1.0071(14)
[NiO2(PPh3)2][b]
[NiO2(CNtBu)2]
[PdO2(PPh3)2]
3000ꢀ80[b]
3300ꢀ150[d]
41100ꢀ300
1.0113(27)[b]
1.0069(16)[d,e]
1.0093(29)[e]
–
897(25), 560(10), 523(10)
877(23), 496 (10), 487(10)
[a] In DMF unless noted. [b] In DMSO. [c] kO is a lower limit in the absence of added PPh3 (see text). [d] [CNtBu]=0–0.088m. [e] In DMF and DMSO.
2
[f] Refs. [27–29]. [g] DnMꢁ for [RhCl(CNtBu)O2(PPh3)2] in Ref. [28]. [h] Average of (K16,16/K18,16
)
for [PdO2(PPh3)2] and [Ni (CNtBu)2O2].
calcd
O
KIEs are larger for the O2-binding reactions with third-row
metals than for the faster reactions with first- and second-row
metals. This range of values is larger than that determined for
iron-, copper-, and flavin-containing enzymes, which react
with O2 according to diverse mechanisms.[4] The formation of
h2-peroxide compounds is characterized by 18O KIEs that are
less than the EIEs (K16,16/K18,16) calculated from Bigeleisenꢀs
formalism[25] using normal mode stretching frequencies.[26–29]
Although the products are structurally quite different, they
exhibit similar shifts of oxygen-isotope-dependent stretching
frequencies resulting in a narrow range of (K16,16/K18,16
)
calcd
values (1.026 to 1.031; see Table 1 and Supporting Informa-
tion). The average value is three times greater than that
reported for the reduction of O2 to H2O2 (1.0089) and
previously used as a benchmark to identify metal–peroxide
intermediates in enzymes and O2-carrier proteins.[4,8–10,26]
Heavy-atom KIEs are conventionally interpreted with
Figure 2. Isotope effects reflecting variation in transition-state
structures for O2 binding.
transition state theory and the equation, k16,16/k18,16 = (n1ꢀ6,16
/
TS
16,16
TS
18,16
nꢀ18,16)(K /K ), which includes contributions from the
reaction-coordinate frequency (nꢀ) and the equilibrium
constant for the formation of the transition state (KTS).[6]
The transition states for h2-O2 binding are visualized in
terms of structures that are intermediate between the reactant
that reported for outer-sphere ET,[30] this mechanism can be
ruled out because the value of DG8 (44.7 kcalmolꢁ1) esti-
mated from one-electron redox potentials[31,32] greatly
exceeds the observed DGꢀ value (18.7 kcalmolꢁ1).
and product states, in which the O O bond is partially
These studies show that 18O KIEs are sensitive probes of
transition-state structure in reactions where O2 binds to a
transition-metal center. Consistent with a productlike tran-
ꢁ
ꢁ
weakened and two new metal O bonds are partially formed.
The results reported here support an assumption that is
commonly made when interpreting mechanisms based on 18
O
sition state, a value of [(k16,16/k18,16)ꢁ1]/[(K16,16/K18,16 calcdꢁ1]
)
KIEs: that nꢀ16,16/nꢀ18,16 is approximately 1 and that K16,16/K18,16 is
close to 1 is obtained for trans-[IrCl(CO)(PPh3)2], the
compound which exhibits the slowest and least thermody-
namically favorable reaction with O2. This parameter
decreases for the other reactions as the transition state
becomes more reactant-like. These reactions are ostensibly
TS
16,16
TS
18,16
[4–7]
the upper limit to K /K
and therefore to k16,16/k18,16.
For inner-sphere electron transfer (ET), in which new
bonds are formed during the redox reaction, the productlike
character of the transition state can be defined in terms of the
ratio of the 18O KIE to the EIE. The parameter [(k16,16
/
more exothermic as kO increases systematically over five
2
k18,16)ꢁ1]/[(K16,16/K18,16
)
calcdꢁ1] is expected to correlate with
orders of magnitude. The trend can be explained in the
context of the Marcus theory of electron transfer, in which the
18O KIE arises from oxygen nuclear reorganization (l).[30] As
DG8 becomes increasingly favorable, the contribution of this
term to DGꢀ decreases as shown in Figure 3.
the barrier height (DGꢀ / logkO ) when the reaction coor-
2
dinate includes the reorganization of bonds involving oxygen
nuclei.[5] This is precisely the behavior observed in Figure 2,
which supports the proposal that the reactions examined
occur by oxidative addition of O2. Alternative mechanisms,
such as those involving h1-superoxide intermediates, are
unlikely in view of a) the absence of optically detectable
intermediates in the kinetics experiments and b) 18O KIEs
that are significantly larger than the EIEs of 1.005 reported
for h1-binding of O2 to metalloproteins.[4,26] Although the
k16,16/k18,16 ratio for trans-[IrCl(CO)(PPh3)2] is very close to
In summary, competitive 18O KIEs have been used for the
first time to probe the binding of O2 to inorganic compounds.
All KIEs are less than the theoretical EIEs calculated from
the stretching frequencies of O2 and the structurally defined
h2-peroxo products. The 18O KIEs decrease as O2-binding rate
constants increase by over five orders of magnitude. This
trend indicates that all the reactions occur by the same
Angew. Chem. Int. Ed. 2005, 44, 7273 –7276
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7275

Communications
[14] J. P. Birk, J. Halpern, A. L. Pickard, J. Am. Chem. Soc. 1968, 90,
4491.
[15] G. Wilke, H. Schott, P. Heimbach, Angew. Chem. 1967, 79, 62;
Angew. Chem. Int. Ed. Engl. 1967, 6, 92.
[16] S. Otsuka, T. Yoshida, Y. Tatsuno, J. Am. Chem. Soc. 1971, 93,
6462.
[17] J. P. Collman, L. S. Hegedus, J. R. Norton, R. G. Finke, Principles
and Applications of Transition Metal Chemistry, University
Science, Mill Valley, 1987.
[18] R. H. Crabtree, Organometallic Chemistry of the Transition
Metals, Wiley, New York, 2001.
[19] Burdick & Jackson “Anhydrous” brand solvents were used as
received.
[20] Yields: trans-[IrCl(CO)O2(PPh3)2] 96 ꢀ 7%, [IrO2(dppe)2]Cl
97 ꢀ 5%, [PtO2(PPh3)2] 107 ꢀ 7%, [PdO2(PPh3)4] 91 ꢀ 9%, and
[NiO2(CNtBu)2] 103 ꢀ 10%. [NiO2(PPh3)2] decomposes afford-
ing PPh3 44 ꢀ 10% and PPh3(O) 48 ꢀ 9%.
Figure 3. Model for the contribution of l to the reaction barrier
(designated ꢀ).
[21] P. B. Chock, J. Halpern, J. Am. Chem. Soc. 1966, 88, 3511.
[22] No O2 was released upon sparging solutions with He at 2 Torr for
up to 2 h.
[23] C. A. Tolman, W. C. Seidel, D. H. Gerlach, J. Am. Chem. Soc.
1972, 94, 2669.
mechanism, albeit with measurable differences in the tran-
sition-state structure. The correlation of the product-like
character of the transition state with the barrier height can be
explained in terms of Marcus theory, where an increasingly
favorable DG8 value decreases the contribution of isotope-
dependent nuclear reorganization to DGꢀ. This property
appears to differentiate reactions involving inner-sphere ET
to O2 from reactions proceeding by outer-sphere ET, in which
18O KIEs appear to be relatively insensitive to the value of
DG8.[30] The highly variable 18O KIEs determined in this work
for a single class of metal-mediated oxidation reactions
underscores the need for benchmarks and relevant models
to interpret such measurements upon O2-activation reactions
of metalloenzymes.
[24] A. Sen, J. Halpern, J. Am. Chem. Soc. 1977, 99, 8337.
[25] J. Bigeleisen, M. G. Mayer, J. Chem. Phys. 1947, 15, 261.
[26] G. Tian, J. P. Klinman, J. Am. Chem. Soc. 1993, 115, 8891.
[27] R. W. Horn, E. Weissberger, J. P. Collman, Inorg. Chem. 1970, 9,
2367.
[28] A. Nakamura, Y. Tatsuno, M. Yamamoto, S. Otsuka, J. Am.
Chem. Soc. 1971, 93, 6052.
[29] P. J. Hayward, D. M. Blake, G. Wilkinson, C. J. Nyman, J. Am.
Chem. Soc. 1970, 92, 5873.
[30] J. P. Roth, R. Wincek, G. Nodet, D. E. Edmondson, W. S.
McIntire, J. P. Klinman, J. Am. Chem. Soc. 2004, 126, 15120.
[31] E8’(IrI/IrII) = 1.32 V vs the normal hydrogen electrode (NHE):
K. Y. Jun, K. M. Ok, C. S. Chin, W. Shin, Bull. Korean Chem.
Soc. 1998, 19, 23.
0
[32] E8’(O2 /O2ꢁ) = ꢁ0.62 vs NHE: D. T. Sawyer, A. Sobkowiak,
J. L. J. Roberts, Electrochemistry for Chemists, Wiley, New York,
1995.
Received: June 17, 2005
Published online: October 13, 2005
Keywords: electron transfer · enzymes · isotope effects ·
.
O–O activation · transition states
[1] R. D. Guy, M. L. Fogel, J. A. Berry, Plant Physiol. 1993, 101, 37.
[2] B. Luz, E. Barkan, M. L. Bender, M. H. Thiemens, K. A.
Boering, Nature 1999, 400, 547.
[3] Y. Q. Gao, R. A. Marcus, Science 2001, 293, 259.
[4] J. P. Roth, J. P. Klinman in Isotope Effects in Chemistry and
Biology (Eds.: A. Kohen, H. Limbach), CRC, Boca Raton, 2005.
[5] W. P. Huskey in Enzyme Mechanism from Isotope Effects (Ed.:
P. F. Cook), CRC, Boca Raton, 1991.
[6] L. Melander, W. H. J. Saunders, Reaction Rates of Isotopic
Molecules, Wiley, New York, 1980.
[7] J. Bigeleisen, M. Wolfsberg, Adv. Chem. Phys. 1958, 1, 15.
[8] S. A. Mills, Y. Goto, Q. Su, J. Plastino, J. P. Klinman, Biochem-
istry 2002, 41, 10577.
[9] S. S. Stahl, W. A. Francisco, M. Merkx, J. P. Klinman, S. J.
Lippard, J. Biol. Chem. 2001, 276, 4549.
[10] W. A. Francisco, G. Tian, P. F. Fitzpatrick, J. P. Klinman, J. Am.
Chem. Soc. 1998, 120, 4057.
[11] G. Tian, J. A. Berry, J. P. Klinman, Biochemistry 1994, 33, 226.
[12] L. Vaska, L. S. Chen, C. V. Senoff, Science 1971, 174, 587.
[13] J. A. McGinnety, J. A. Ibers, Chem. Commun. 1968, 235.
7276
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7273 –7276