Kinetics and mechanism of methane, methanol, and dimethyl ether C-H activation with electrophilic platinum complexes

Article abstract of DOI:10.1021/ja056387t

The relative rates of C-H activation of methane, methanol, and dimethyl ether by [(N-N)PtMe-(TFE-d₃)]^- ((N-N) = ArN=C(Me)-C(Me)=NAr; Ar = 3,5-di-tert-butylphenyl, TFE-d₃ = CF ₃CD₂OD) (2(TFE)) were determined. Methane activation kinetics were conducted by reacting 2(TFE)-¹³C with 300-1000 psi of methane in single-crystal sapphire NMR tubes; clean second-order behavior was obtained (k = 1.6 ± 0.4 × 10^-3 M^-1 s ^-1 at 330 K; k = 2.7 ± 0.2 × 10^-4 M ^-1 s^-1 at 313 K). Addition of methanol to solutions of 2(TFE) rapidly establishes equilibrium between methanol (2(MeOD)) and trifluoroethanol (2(TFE)) adducts, with methanol binding preferentially (K _eq = 0.0042 ± 0.0006). C-H activation gives [(N-N)Pt(CH ₂OD)-(MeOD)]⁺ (4), which is unstable and reacts with [(RO)B(C₆F₅)₃]^- to generate a pentafluorophenyl platinum complex. Analysis of kinetics data for reaction of 2 with methanol yields k = 2.0 ± 0.2 × 10^-3 M^-1 s^-1 at 330 K, with a small kinetic isotope effect (k _H/k_D = 1.4 ± 0.1). Reaction of dimethyl ether with 2(TFE) proceeds similarly (K_eq = 0.023 ± 0.002, 313 K; k = 5.5 ± 0.5 × 10^-4 M^-1 s_-1, k _H/k_D = 1.5 ± 0.1); the product obtained is a novel bis(alkylidene)-bridged platinum dimer, [(diimine)Pt(μ-CH₂)(μ- (CH(OCH₃))Pt(diimine)]²⁺ (5). Displacement of TFE by a C-H bond appears to be the rate-determining step for all three substrates; comparison of the second-order rate constants (k_(methane)/k _(methanol) = 1/1.3, 330 K; k_(methane)/k _{(dimethyl ether)} = 1/2.0, 313 K) shows that this step is relatively unselective for the C-H bonds of methane, methanol, or dimethyl ether. This low selectivity agrees with previous estimates for oxidations with aqueous tetrachloroplatinate(II)/hexachloroplatinate(IV), suggesting a similar rate-determining step for those reactions.

Full text of DOI:10.1021/ja056387t

Published on Web 01/24/2006
Kinetics and Mechanism of Methane, Methanol, and Dimethyl
Ether C-H Activation with Electrophilic Platinum Complexes
Jonathan S. Owen, Jay A. Labinger,* and John E. Bercaw*
Contribution from the Arnold and Mabel Beckman Laboratories of Chemical Synthesis,
California Institute of Technology, Pasadena, California 91125
Received September 16, 2005; E-mail: jal@its.caltech.edu
Abstract: The relative rates of C-H activation of methane, methanol, and dimethyl ether by [(N-N)PtMe-
(TFE-d3)]+ ((N-N) ) ArNdC(Me)-C(Me)dNAr; Ar ) 3,5-di-tert-butylphenyl, TFE-d3 ) CF3CD2OD) (2(TFE))
were determined. Methane activation kinetics were conducted by reacting 2(TFE)-13C with 300-1000 psi
of methane in single-crystal sapphire NMR tubes; clean second-order behavior was obtained (k ) 1.6 (
0.4 × 10-3 M-1 s-1 at 330 K; k ) 2.7 ( 0.2 × 10-4 M-1 s-1 at 313 K). Addition of methanol to solutions of
2(TFE) rapidly establishes equilibrium between methanol (2(MeOD)) and trifluoroethanol (2(TFE)) adducts,
with methanol binding preferentially (Keq ) 0.0042 ( 0.0006). C-H activation gives [(N-N)Pt(CH2OD)-
(MeOD)]+ (4), which is unstable and reacts with [(RO)B(C6F5)3]- to generate a pentafluorophenyl platinum
complex. Analysis of kinetics data for reaction of 2 with methanol yields k ) 2.0 ( 0.2 × 10-3 M-1 s-1 at
330 K, with a small kinetic isotope effect (kH/kD ) 1.4 ( 0.1). Reaction of dimethyl ether with 2(TFE) proceeds
similarly (Keq ) 0.023 ( 0.002, 313 K; k ) 5.5 ( 0.5 × 10-4 M-1 s-1, kH/kD ) 1.5 ( 0.1); the product
obtained is a novel bis(alkylidene)-bridged platinum dimer, [(diimine)Pt(µ-CH2)(µ-(CH(OCH3))Pt(diimine)]2+
(5). Displacement of TFE by a C-H bond appears to be the rate-determining step for all three substrates;
comparison of the second-order rate constants (k(methane)/k(methanol) ) 1/1.3, 330 K; k(methane)/k(dimethyl
)
ether)
1/2.0, 313 K) shows that this step is relatively unselective for the C-H bonds of methane, methanol, or
dimethyl ether. This low selectivity agrees with previous estimates for oxidations with aqueous tetrachlo-
roplatinate(II)/hexachloroplatinate(IV), suggesting a similar rate-determining step for those reactions.
Introduction
100 times more reactive than methyl bisulfate, its oxidation
product.2c Studies on the aqueous chloroplatinate system
The rising cost of oil and the fear of high carbon dioxide
levels in the atmosphere are leading scientists to search for cheap
alternatives to coal and oil.1 The direct aerobic oxidation of
methane to useful chemicals is one potential approach to this
problem. Unfortunately, such processes are complicated by the
increasing reactivity of its oxidation products. Most oxidants
capable of reacting with methane (BDE(C-H) ) 104 kcal/mol)
follow a hydrogen atom abstraction route and hence react more
readily with the weaker C-H bonds of methanol (BDE(C-H)
) 93 kcal/mol), placing an inherent limit on the yield.2
Nonetheless, a few oxidation catalysts based on platinum(II),
palladium(II), and mercury(II) salts have been shown to produce
high yields of partially oxidized products.3 In fuming sulfuric
acid, a platinum complex was shown to catalyze the oxidation
of methane to methyl bisulfate in up to 70% yield.3c Under these
conditions, it was estimated that methane can be as much as
pioneered by Shilov suggest that methyl groups (or methane)
and hydroxymethyl groups (or methanol) exhibit similar reactiv-
ity, with estimated relative rate constants ranging from 1.5:1 to
1:5.4 The origin of these unusual oxidation selectivities is not
clear.
Several groups have studied the thermodynamic and kinetic
selectivity of C-H bond activation by organometallic transition
metal complexes.5 In general, the regioselectivity of such C-H
activations is opposite to that of hydrogen atom abstraction
reactions, with stronger bonds being more reactive, leading to
the following selectivity order: H-C(sp2) > 1° H-C(sp3) >
2° H-C(sp3) > 3° H-C(sp3). Most studies have been limited
to alkanes and arenes because the metal complexes suited to
mechanistic studies have limited functional group tolerance.
Little is known about the influence of heteroatom substitution
on the rate of C-H bond activation.6
Mechanistic interpretation of this selectivity is complicated
by the fact that C-H activation is a multistep and, often,
(1) Arakawa, H. et al. Chem. ReV. 2001, 101, 953-996.
(2) (a) Labinger, J. A. Catal. Let. 1988, 1, 371-375. (b) Crabtree, R. H. Chem.
ReV. 1995, 95, 987-1007. (c) Labinger, J. A.; Bercaw, J. E.; Luinstra, G.
A.; Lyon, D. K.; Herring, A. M. In Natural Gas ConVersion II; Proceedings
of the Third International Gas Conversion Symposium, Sydney, Australia,
July 4-9, 1993; Howe, R. F., Curry-Hyde, E., Eds.; Elsevier: Amsterdam,
1994; pp 515-520.
(4) (a) Labinger, J. A.; Herring, A. M.; Bercaw, J. E. J. Am. Chem. Soc. 1990,
112, 5628-5629. (b) Labinger, J. A.; Herring, A. M.; Lyon, D. K.; Luinstra,
G. A.; Bercaw, J. E.; Horvath, I. T.; Eller, K. Organometallics 1993, 12,
895-905. (c) Sen, A.; Benvenuto, M. A.; Lin, M. R.; Hutson, A. C.;
Basickes, N. J. Am. Chem. Soc. 1994, 116, 998-1003. These estimates
are complicated by the ubiquitous formation of metallic Pt, which is a good
alcohol oxidation catalyst.
(3) (a) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. Angew. Chem., Int. Ed. 1998,
37, 2180-2192 and references therein. (b) Periana, R. A.; Taube, D. J.;
Evitt, E. R.; Loffler, D. G.; Wentrcek, P. R.; Voss, G.; Masuda, T. Science
1993, 259, 340-343. (c) Periana, R. A.; Taube, D. J.; Gamble, S.; Taube,
H.; Satoh, T.; Fujii, H. Science 1998, 280, 560-564.
(5) Arndtsen, B. A.; Bergman, R. O.; Mobley, T. A.; Peterson, T. H. Acc.
Chem. Res. 1995, 28, 154-162 and references therein.
9
10.1021/ja056387t CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 2005-2016
2005

A R T I C L E S
Scheme 1
Owen et al.
Scheme 2
reversible reaction.7 In particular, the formation of arene CdC
π-complexes and C-H σ-complexes complicates direct studies
of the C-H bond cleavage and forming steps. Most often this
is revealed in studies on the microscopic reverse of C-H bond
oxidative addition (Scheme 1). For systems where reductive
elimination of alkane or arene is highly endothermic, C-H bond
formation is typically a fast and reversible first (step i) followed
by rate-determining loss of hydrocarbon (step ii or iW). The
inVerse kinetic isotope effects commonly observed in these
reactions are believed to result from the product of an inverse
equilibrium isotope effect for (i) and a small, normal kinetic
isotope effect for alkane decoordination (ii).8 Several examples
of direct spectroscopic and kinetic evidence for the intermediacy
of arene complexes in arene C-H bond activation have been
reported,7e but only a few studies of alkane activation provide
kinetic information on individual microscopic steps.9,10
Scheme 3
Results
We have previously exploited diimine-ligated platinum(II)
and palladium(II) alkyl complexes as model systems in extensive
mechanistic studies of C-H activation for arenes and alkyl-
substituted arenes.11 Here we use the Pt(II) model to investigate
the relative rates and mechanism of C-H activation for methane,
methanol, and dimethyl ether. In addition, spectroscopic and
crystallographic characterization of the primary and/or secondary
products of methanol and dimethyl ether C-H activation are
described.
Reaction of Methane with [(N-N)Pt(13CH3-nDn)-
(DOCD2CF3)]+[CF3CD2OB(C6F5)3]- (2-13C(TFE)). To study
the kinetics of methane activation, (diimine)platinumdimeth-
yl complex 1-13C was prepared from the diimine ligand
ArNdC(Me)-C(Me)dNAr (Ar ) 3,5-di-tert-butylphenyl) and
[(µ-SMe2)Pt(13CH3)2]2. Protonolysis of 1-13C with tris(pen-
tafluorophenyl)borane (B(C6F5)3) in anhydrous CF3CD2OD
(TFE-d3) produces 2-13C(TFE), trifluoroethoxytris(penta-
fluorophenyl)borate, and methane. Although only 1 equiv of
B(C6F5)3 is required by stoichiometry, 2 equiv is generally
needed to achieve clean formation of 2; the reason is not known.
Both 2-13C(TFE) and methane are obtained as a statistical
mixture of isotopologs (Scheme 2), as previously reported.11d
(6) See, for example: (a) Vetter, A. J.; Jones, W. D. Polyhedron 2004, 23,
413-417. (b) Lawrence, J. D.; Takahashi, M.; Bae, C.; Hartwig, J. F. J.
Am. Chem. Soc. 2004, 126, 15334-15335.
(7) (a) Crabtree, R. H. Chem. ReV. 1985, 85, 245-269. (b) Crabtree, R. H.
Chem. ReV. 1995, 95, 987-1007. (c) Hall, C.; Perutz, R. N. Chem. ReV.
1996, 96, 3125-3146. (d) Shilov, A. E.; Shul’pin, G. B. Chem. ReV. 1997,
97, 2879-2932. (e) Fekl, U.; Goldberg, K. I. AdV. Inorg. Chem. 2003, 54,
259-320. (f) Lersch, M.; Tilset, M. Chem. ReV. 2005, 105, 2471-2526.
(8) Jones, W. D. Acc. Chem. Res. 2003, 36, 140-146.
(9) (a) Schaller, C. P.; Bonanno, J. B.; Wolczanski, P. T. J. Am. Chem. Soc.
1994, 116, 4133-4134. (b) Bennett, J. L.; Wolczanski, P. T. J. Am. Chem.
Soc. 1997, 119, 10696-10719. (c) Northcutt, T. O.; Wick, D. G.; Vetter,
A. J.; Jones, W. D. J. Am. Chem. Soc. 2001, 123, 7357-7270. (d) Vetter,
A. J.; Flaschenriem, C.; Jones, W. D. J. Am. Chem. Soc. 2005, 127, 12315-
12322.
(10) (a) Weiller, B. H.; Wasserman, E. P.; Bergman, R. G.; Moore, C. B.;
Pimentel, G. C. J. Am. Chem. Soc. 1989, 111, 8288. (b) Bengali, A. A.;
Schultz, R. H.; Moore, C. B.; Bergman, R. G. J. Am. Chem. Soc. 1994,
116, 9585-9589. (c) Schultz, R. H.; Bengali, A. A.; Tauber, M. J.; Weiller,
B. H.; Wasserman, E. P.; Kyle, K. R.; Moore, C. B.; Bergman, R. G. J.
Am. Chem. Soc. 1994, 116, 7369-7377.
Upon pressurizing solutions of 2-13C(TFE) in single-crys-
tal sapphire NMR tubes with dry methane-12C, formation of
2-12C(TFE) is observed in the 1H NMR spectrum (Scheme 3).
Disappearance of the 1H NMR signal for 2-13C(TFE) and
appearance of that for 2-12C(TFE) occur at similar rates. Data
for kinetics were obtained by monitoring the 13C NMR signal
for 2-13C(TFE) over time, at 330 and 313 K, in the presence
of 300-1000 psi of added methane. The data were analyzed as
a function of the concentration of dissolved methane (measured
by NMR). The first-order decay of the 13C signal exhibits a
linear dependence on [methane] at both 330 K (k2 ) 1.6 ( 0.4
(11) (a) Zhong, A. H.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2002,
124, 1378-1399. (b) Johansson, L.; Tilset, M.; Labinger, J. A.; Bercaw, J.
E. J. Am. Chem. Soc. 2002, 122, 10846-10855. (c) Ackerman, L. J.;
Sadighi, J. P.; Kurtz, D. M.; Labinger, J. A.; Bercaw, J. E. Organometallics
2003, 22, 3884-3890. (d) Heyduk, A. F.; Driver, T. G.; Labinger, J. A.;
Bercaw, J. E. J. Am. Chem. Soc. 2004, 126, 15034-15035. (e) Driver, T.
G.; Day, M. W.; Labinger, J. A.; Bercaw, J. E. Organometallics 2005, 24,
3644-3654.
× 10-3 M-1 s-1) and 313 K (k2 ) 2.7 ( 0.2 × 10-4 M-1 s-1
)
(Table 1 and Figure 1).
Methane activation is accompanied by a much slower
decomposition reaction, indicated by appearance of new NMR
9
2006 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Methane, Methanol, and Dimethyl Ether C−H Activation
A R T I C L E S
Table 1. Methane Activation Kinetics with 2
appears, indicative of a rapid equilibrium between methanol-
bound (2(MeOD)) and trifluoroethanol-bound (2(TFE)) plati-
num cations, with the former strongly favored (Keq ) 0.0042
( 0.0006, 330 K). Additional broad signals for methanol bound
a
temp
(K)
[Pt]
(M)
methane
(psi)
[CH4]
(M)
kobs
k2
(M-1 s-
)
1
1
(s-
)
313
313
313
330
330
330
330
330
330
0.0145
0.0145
0.0145
0.00944
0.00944
0.0145
0.0218
0.0145
0.0145
570
835
1020
300
400
500
503
550
700
0.393
0.690
0.855
0.298
0.370
0.423
0.510
0.488
0.617
0.000059
0.000101
0.000121
0.000288
0.000283
0.000350
0.000376
0.000385
0.000540
0.000266
0.000275
0.000266
0.00181
0.00143
0.00157
0.00140
0.00150
0.00169
1
to tris(pentafluorophenyl)borane are visible in the H and 13C
NMR spectra; their shape and shift depend on the probe
temperature and the concentration of added methanol. Over
several hours, 1H NMR spectroscopy shows formation of
methane isotopologs, a set of ligand resonances for a new
platinum complex, and additional resonances in the region
typical of methanol C-H bonds. These new resonances
subsequently disappear, with the concomitant appearance of
signals characteristic of methanol-bound pentafluorophenyl
platinum cation (3(MeOD)). The pentafluorophenyl complex
appears in these reactions at a rate considerably faster than that
in methane activation (see above); in reactions carried out with
a high concentration of methanol, 3 is the major (>90%)
observable product.
a Calculated from {k2 ) (kobs-ki)/[CH4] × 2}, where ki is taken from
the intercept of kobs versus [methane], and the factor of 2 is included for
the equal rates of the forward and back reaction. See text for discussion.
The methane generated in these reactions is a mixture of not
only CH4 and CH3D (as expected for C-H activation of
methanol-d1 by 2(TFE), which contains approximately equal
amounts of Pt-CH3 and Pt-CH2D) but also CH2D2. The
relative amount of CH2D2 decreases with increasing concentra-
tion of methanol-d1; for [methanol-d1] ) 0.123 M, CH4:CH2D2
≈ 1:2, whereas for [methanol-d1] ) 1.23 M, CH4:CH2D2 ≈
5:1. Reactions in which CD3OD is substituted for CH3OD
produce primarily CH3D and CH2D2 with essentially no CH4.
Figure 1. Plots of kobs versus [methane] at 330 K (left) and 313 K (right).
Scheme 4
Concentrations of the initially formed platinum complex suffi-
cient for characterization by NMR spectroscopy may be obtained
by using relatively low concentrations of added methanol.
Reaction of 2 with 3 equiv of 13CH3OH produces a new 13C
NMR signal at δ ) -4 ppm with broad 195Pt satellites (J )
732 Hz). In the 1H-coupled 13C NMR spectrum, this resonance
appears as a 1:2:1 triplet (J ) 134 Hz), consistent with a
platinum-bound sp3-hybridized CH2X fragment. 1H-{13C}
Heteronuclear Multiple Quantum Coherence (HMQC) spec-
troscopy is useful for identifying signals resulting from protons
directly attached to 13C atoms. The 1D HMQC spectrum of the
product obtained from 2(TFE) and 13CH3OH (present as mostly
13CH3OD in CF3CD2OD solvent) shows four signals, derived
from the labeled methanol, that are greatly enhanced relative
to the signals for the diimine ligand (Figure 2). In addition to
the signals for bound methanol in unreacted 2(13CH3OD) (δ )
2.94 ppm) and decomposition product 3(13CH3OD) (δ ) 2.90
ppm), two new signals are observed at δ ) 2.62 and 2.72 ppm,
with an intensity ratio of 3:2. The 2D HMQC spectrum (Figure
3) shows how these 1H resonances correlate with the 13C signals
for the carbons to which they are bound. The signal at δ )
2.72 ppm correlates with the high-field 13C resonance noted
above (δ ) -4 ppm), while the signal at δ ) 2.62 ppm
correlates with a 13C resonance at δ ) 55.8 ppm, close to the
O-bound methanol signals of 2(13CH3OD) (δ ) 57.1 ppm) and
3(13CH3OD) (δ ) 56.2 ppm). On the basis of these data, we
formulate the intermediate as the hydroxymethyl complex
([(N-N)Pt(CH2OD)(DOCH3)]+) (4, Scheme 5).13
signals as well as the nonzero intercept in Figure 1 (kdecomp
)
1.8 × 10-5 s-1, 330 K; kdecomp ) 6.8 × 10-6 s-1, 313 K). In
the absence of added substrate, 2 decomposes to give the same
species (along with CH3D and CH2D2) at an initial first-order
rate similar to the intercept value ([2(TFE)] ) 15.2 mM,
[B(C6F5)3] ) 30.3 mM, kobs ) 4 × 10-5 s-1, 330 K). The
decomposition product was identified as a pentafluorophenyl
platinum complex (3) (Scheme 4) by the close similarity of
NMR parameters to those of a previously reported analogue.11e
In further confirmation of the assignment, electrospray mass
spectra of these samples diluted with methanol show a large
peak at m/z ) 854.3 amu, consistent with a methanol-bound
pentafluorophenyl platinum cation (3(MeOD)), and solutions
of 2 left exposed to air produce crystals of the hydroxytris-
(pentafluorophenyl)borate pentafluorophenyl complex 3(HOB-
(C6F5)3), differing from the previous analogue11e only by the
substituents on the diimine ligands, as established crystallo-
graphically.12 Exposure of the sample to moisture most likely
generates a water adduct with the Lewis acidic borane, and the
anion ([B(C6F5)3OH]-) binds to the metal center, generating
an insoluble charge-neutral complex.
Reaction of Methanol with 2. On addition of anhydrous
methanol-d1 to solutions of 2(TFE), a new set of NMR signals
Kinetics of Methanol Activation. Reactions of 2 with
varying concentrations of CH3OD or CD3OD at 330 K were
(12) Crystallographic data have been deposited at the CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK and copies can be obtained on request, free of
charge, by quoting the publication citation and the deposition number
213879.
(13) The possibility that 4 is a trifluoroethoxymethyl complex, [(N-N)Pt-
(CH2OCD2CF3)(L)]+ (L ) TFE-d3, MeOD), produced by alcoholysis of
the C-O bond in a hydroxymethyl complex cannot be ruled out.
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2007

A R T I C L E S
Owen et al.
Figure 2. One-dimensional HMQC spectrum for the reaction of 2 with 13CH3OH. A solvent presaturation routine was used to minimize the intensity of the
free methanol signal at δ 3.3.
Figure 3. Two-dimensional HMQC for the reaction of 2 with 13CH3OH. The 1D HMQC spectrum from Figure 2 as well as a higher-resolution 13C NMR
spectrum (both acquired separately) are superimposed on the axes of the 2D plot.
Scheme 5
Table 2. Kinetics of Methanol-d1 Activation with 2 at 330 K
[Pt]
(M)
[CH3OD]
(M)
[TFE]
(M)
k
obs(H)
k
obs(H)/
R
(TFE)
)
1
1
(s-
)
(s-
0.00933
0.00933
0.00949
0.00949
0.00949
0.00949
0.123
0.247
0.507
0.738
1.23
13.5
13.4
13.3
13.1
12.8
12.2
0.000136
0.000135
0.000135
0.000132
0.000119
0.000105
0.000435
0.000731
0.00137
0.00191
0.00284
0.00518
2.45
Table 3. Kinetics of Methanol-d4 Activation with 2 at 330 K
[Pt]
(M)
[CD3OD]
(M)
[TFE]
(M)
k
obs(D)
k
obs(D)/
R
(TFE)
)
1
1
1
(s-
)
(s-
examined by following the disappearance of 2(MeOD) by H
NMR spectroscopy; each reaction exhibited well-behaved
pseudo-first-order kinetics. For convenience, usually only
[2(MeOD)] (the most abundant species) was monitored. Since
the equilibrium between 2(MeOD) and 2(TFE) is rapidly
establishedsmuch faster than any conversion of 2 to productss
and the reactions are first-order in [2], the rate constants thus
determined and those that would be obtained by monitoring
[2(TFE)] (or [2tot]) must be identical. This was verified by
comparing rate constants based on [2(TFE)] and on [2(MeOD)]
for experiments with low [MeOD] values (where [2(TFE)] is a
significant fraction of [2tot]); the results agree to within
experimental uncertainty.
0.00947
0.00933
0.00890
0.00949
0.616
1.23
2.46
4.43
13.2
12.8
12.2
11.1
0.000100
0.0000839
0.0000622
0.0000585
0.00152
0.00253
0.00389
0.00719
The rate constants (kobs) from these first-order plots are shown
in Tables 2 and 3, and are plotted against methanol concentration
in Figure 4. It can be readily seen that (a) there is a significant
kinetic isotope effect (KIE), and (b) the apparent rate constant
decreases with increasing methanol concentration, although not
in an inverse first-order manner. Our interpretation of these
observations is discussed below.
9
2008 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Methane, Methanol, and Dimethyl Ether C−H Activation
A R T I C L E S
Figure 4. Plot of kobs versus [methanol] for the reaction of 2 with CH3OD
(O) and CD3OD (0).
Figure 5. X-ray crystal structure of 5. Ellipsoids are drawn at 50%
probability; the counteranions [CF3CD2OB(C6F5)3]- are not shown.
Table 4. 1H and 13C NMR Parameters for 5a
b
c
(1H)
2JHH
4JHH
JC
-H
δ
(
13C)
JPt
-C
Scheme 6
δ
(ppm)
(Hz)
(Hz)
(Hz)
(ppm)
(Hz)
CH2A
CH2B
CH(OMe)
OCH3
8.15
8.01
10.35
3.79
5.25
5.25
1.74
154.3
153.4
177.7
146.5
133.2
133.2
190.0
67.1
560.2
560.2
637.2
1.74
a Acetone-d6. b Geminal coupling between methylidene protons. c Cou-
pling of methoxymethylidene proton to one methylidene proton.
Reaction of Dimethyl Ether with 2(TFE). As with metha-
nol, addition of anhydrous dimethyl ether (DME) to solutions
of 2(TFE) quickly establishes an equilibrium between the DME-
bound platinum cation (2(DME)) and 2(TFE), again favoring
the former, with Keq ) 0.023 ( 0.002 at 313 K. Disappearance
of the 1H NMR signals of 2 proceeds slowly at 313 K with the
appearance of signals for methane (CH4 and CH3D; as with
methanol, activation of DME-d6 gave CH3D and CH2D2
instead), a new species with an unusual spectral pattern (six
aromatic ligand resonances instead of the usual four), and small
amounts of 3. On continued monitoring, the new resonances
decrease in intensity, at the same time that crystals appear in
Table 5. Selected Bond Lengths (Å) and Bond Angles (deg) for 5
Pt1-Pt1A
Pt1-C1
2.6053(2)
2.0055(34)
1.5157(74)
81.02(0.13)
98.98(0.13)
C1-O1
Pt1-C1-Pt1A
C1-Pt1-C1A
Table 6. Kinetics of Dimethyl Ether Activation with 2 at 313 K
[Pt]
(M)
[CH3OCH3]
(M)
[TFE]
(M)
kobs(D)
kobs(D)/
R
(TFE)
)
1
the NMR tube. The H NMR spectrum of a solution of the
1
1
(s-
)
(s-
isolated crystals in acetone-d6 exhibits additional resonances
downfield of the ligand peaks including one resonance with
broad 195Pt satellites. Reaction of 2 with DME-13C2 affords a
0.0105
0.0105
0.0100
0.00983
0.163
0.191
0.574
0.921
13.6
13.6
13.2
12.9
0.0000563
0.0000564
0.0000987
0.000110
0.0000899
0.0000959
0.000313
0.000503
1
similar species, allowing full analysis of H and 13C NMR
parameters (Table 4). The shifts and splitting patterns of these
downfield signals suggest the presence of two alkylidene ligands,
one CH2 group, and one CHX group. This was confirmed by
X-ray crystallography, which revealed an unusual structure: a
Pt-Pt dimer with µ-CH2 and µ-CH(OCH3) groups (5, Scheme
6, Figure 5).14 The molecule is disordered about a center of
symmetry; hence only the average Pt-C distance and angles
for the µ-CH2 and µ-CH(OCH3) groups could be determined
(Table 5).
activation was determined by a competition experiment: equal
amounts of DME-d0 and DME-d6 were added to a solution of
2 in an NMR tube, and from the ratio of deuterio and protio
isotopologs of the liberated methane, kH/kD was measured to
be 1.5 ( 0.1.
Discussion
C-H Activation of Methane, Methanol, and Dimethyl
Ether. A generalized C-H activation reaction of 2 is shown in
eq 1. In the case of RH ) methane, this reaction is only
observable only by means of isotopic labeling. For methanol
and DME, the expected products are hydroxymethyl complex
4 and the analogous methoxymethyl complex 6, respectively.
Very few examples of the former have been characterized; all
are octahedral,15 which would make 4 somewhat unique.
Alkoxyalkyl complexes are considerably more common.16 We
therefore anticipated possible difficulties with product stability
Kinetics of Dimethyl Ether Activation with 2. As with
methanol (see above), the disappearance of 2(DME) was
1
monitored by H NMR at 313 K; it shows pseudo-first-order
behavior over more than five half-lives. The dependence of
apparent first-order rate constants on [DME] is shown in Table
6 and plotted in Figure 6. A kinetic isotope effect for DME
(14) Crystallographic data have been deposited at the CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK and copies can be obtained on request, free of
charge, by quoting the publication citation and the deposition number
267189.
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2009

A R T I C L E S
Owen et al.
Table 7. Crystal and Refinement Data for 5 (CCDC 267189)a
empirical formula
formula weight, g/mol
crystal size, mm3
θ range, deg
a, Å
[C67H102N4OPt2]+22(C20H2BF18O)-
2591.76
0.29 × 0.25 × 0.18
2.42 to 34.86
12.3114(3)
15.4916(4)
16.6024(4)
63.6320(10)
73.3540(10)
88.6650(10)
2698.14(12)
1
b, Å
c, Å
R, deg
â, deg
γ, deg
volume, Å3
Z
crystal system
space group
d
θ range for data collection, deg
completeness to θ ) 40.43°
reflections collected
triclinic
P1h
1.595
1.74 to 40.43
87.30%
74339
Figure 6. Plot of kobs versus [DME] for the reaction of 2 with CH3OCH3.
calc, g/cm3
Scheme 7
independent reflections
absorption correction
29995 [Rint ) 0.0603]
SADABS (µ ) 2.703 mm-1
1.000000 and 0.831789
1.748
)
max. and min. transmission
goodness-of-fit on F2
R indices (all data)
R1 ) 0.1021, wR2 ) 0.1148
2
a At 100(2) K. b R1 ) Σ||Fo|| - |Fo||/ Σ|Fo|. c wR2 ) [Σ[w(Fo2 - Fc )2]/
2
Σ[w(Fo )2]1/2
.
for the reaction of methanol, but less so for DME. In fact, it is
just the other way around. Hydroxymethyl complex 4 is indeed
unstable to reaction conditions, reacting with a pentafluoro-
phenyl group from the counteranion to give 3 (which is also
produced directly from 2; see below) and hence could not be
isolated. However, it can be detected by NMR and, under the
right conditions, is present in sufficient concentration to permit
full characterization by 1D and 2D 1H and 13C NMR spectros-
copy.
(Scheme 7). Alternatively, methanol loss could be an early step
in the mechanism; R-alkoxyalkyl ligands have been shown to
undergo C-O bond cleavage in the presence of strong acids.16
At present, we cannot distinguish between these (or other)
possibilities, nor explain why 4 is relatively stable and exhibits
no such reactivity.
The bonding in 5 is unusual. The diamagnetism of the
dicationic complex and the short Pt-Pt distance (2.6053(2) Å)
suggest formulation as a Pt(III)-Pt(III) dimer with a metal-
metal bond, analogous to the only other examples of hydrocar-
byl-bridged platinum-platinum dimers.18 The 1H and 13C
spectroscopic parameters of 5 also provide some insight. The
downfield chemical shifts of the bridging groups in both the
1H and 13C spectra are most likely due to a large paramagnetic
contribution, more typically characteristic of terminal alkylidene
ligands. The large C-H coupling constants for both bridgess
especially the methoxymethylidenesand small platinum-carbon
coupling constants suggest high s-character in the C-H bonds
and larger p-orbital contributions to the C-Pt bonds that may
arise from the acute Pt-C-Pt bond angles and the electron-
deficient nature of the platinum center.
Despite the fact that neither 4 nor 6 is stable, the initial
reaction of 2 with both methanol and DME is clearly C-H
activation, as methane is liberated during reaction of perdeu-
terated substrate with one more deuterium than with unlabeled
substrate. Furthermore, since clean pseudo-first-order disap-
pearance of platinum methyl complex 2 is observed in every
experiment, the instability of 4 (gradual) or 6 (immediate) must
not have any effect on the kinetic parameters determined by
that method. (The competing formation of 3 from 2 does affect
the kinetics, of course, as discussed in the following section.)
Kinetics of C-H Activation: Substrate Inhibition. The
kinetics of methane activation, as determined from disappear-
In contrast, no evidence for any methoxymethyl complex 6
could be detected in the reaction of DME; the only product
observed (besides small amounts of 3) is the novel methylene/
methoxymethylene-bridged dimer 5. There are previous reports
wherein proposed R-alkoxyalkyl intermediates generated by
C-H activation of tetrahydrofuran or diethyl ether result in the
formation of Fischer carbene hydride complexes by R-elimina-
tion.17 An analogous step here, followed by formation of a
hydride-bridged dimer and loss of methanol, would lead to 5
(15) (a) Thorn, D. L.; Tulip, T. H. Organometallics 1982, 1, 1580-1586. (b)
Blackmore, T.; Bruce, M. I.; Davidson, P. J.; Iqbal, M. Z.; Stone, F. G. A.
J. Chem. Soc. A 1970, 3153. (c) Casey, C. P.; Andrews, M. A,; McAlister,
D. R.; Rinz, J. E. J. Am. Chem. Soc. 1980, 102, 1927. (d) Sweet, J. R.;
Graham, W. A. G. J. Organomet. Chem. 1979, 173, C9. (e) Headford, C.
E. L.; Roper, W. R. J. Organomet. Chem. 1980, 198, C7. (f) Espenson, J.
H.; Bakac, A. J. Am. Chem. Soc. 1981, 103, 2721 and references therein.
(16) Thorn, D. L. Organometallics 1986, 5, 1897-1903 and references therein.
(17) (a) Holtcamp, M. W.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc.
1997, 119, 848-849. (b) Luecke, H. F.; Arndtsen, B. A.; Burger, P.;
Bergman, R. G. J. Am. Chem. Soc. 1996, 118, 2517-2518.
(18) (a) Uson, R.; Fornies, J.; Tomas, M.; Casas, J. M.; Cotton, F. A.; Falvello,
L. R. J. Am. Chem. Soc. 1985, 107, 2556. (b) Uson, R.; Fornies, J.; Tomas,
M.; Casas, J. M.; Cotton, F. A.; Falvello, L. R.; Llusar, R. Organometallics
1988, 7, 2279. (c) Uson, R.; Fornies, J.; Tomas, M.; Casas, J. M.; Cotton,
F. A.; Falvello, L. R.; Feng, X. J. Am. Chem. Soc. 1993, 115, 4145.
9
2010 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Methane, Methanol, and Dimethyl Ether C−H Activation
A R T I C L E S
ance of 13C-labeled 2, are straightforward: there is a linear
dependence on methane concentration (Figure 1), with the small
positive intercept corresponding reasonably well to the rate of
decomposition in the absence of added substrate.
In contrast, C-H activation rates for both methanol and DME
exhibit nonlinear dependence on substrate concentration (Figures
4 and 6). We attribute this to inhibition by substrate coordination.
We have previously shown that benzene C-H activation in TFE
solution is inhibited by water because the aquo-platinum methyl
cation (2(H2O)) is thermodynamically favored over (by a fac-
tor of 430) but much less reactive than the TFE-ligated cation
2(TFE); the kinetics indicate that the latter is the sole reactive
C-H activating species.11a,d NMR studies show that methanol
and DME also coordinate to the platinum center more tightly
than TFE, by factors of 240 and 44, respectively.
Figure 7. Plot of kobs/RTFE versus [DME] for the reaction of 2 with
CH3OCH3.
Assuming that, as for water, the more tightly bound MeOR
(R ) D, Me) cannot be directly displaced by the substrate’s
C-H bond, productive C-H activation of methanol and DME
will likewise proceed only via a bimolecular reaction between
substrate and 2(TFE). The expected rate law is derived in eqs
2-6; the observed pseudo-first-order rate constants will be of
the form kobs ) k2[MeOR]RTFE, where RTFE is the mole fraction
of total 2 present as 2(TFE).
2(MeOR) + TFE {Keq} 2(TFE) +
(2)
MeOR
(R ) D, Me)
Figure 8. Plot of kobs/RTFE versus [methanol] for the reaction of 2 with
CH3OD (O) and CD3OD (0).
k2
2(TFE) + MeOR
98 products
(3)
(4)
[2(TFE)]‚[MeOR]
[2(MeOR)]‚[TFE]
implicated in the activation of a C-H bond of diethyl ether by
a related Pt(II) complex.17a However, such a route is not
consistent with the observed substrate dependence in the case
of DME. Saturation kinetics would be expected, in qualitatiVe
agreement with Figure 6, but given the value of Keq determined
by NMR, saturation would be reached at a much lower [DME]
value than observed. (The value of Keq calculated by fitting the
data of Table 6 to the rate law of eq 7 (shown in Figure 6) is
essentially identical to that measured by NMR.)
On the other hand, the apparent rates of disappearance of 2
in methanol C-H activation do not fit the expected concentra-
tion dependence. Instead, the rate is highest at low methanol
concentrations and decreases modestly as the concentration
increases (Figure 4). If we process the data (as we did for DME)
according to eq 7 (Tables 2, 3, and 6), the plot of apparent rates
corrected for RTFE does show a linear dependence on the
substrate concentration, but with a substantial nonzero intercept
(Figure 8). How are these to be explained?
The large intercept, which corresponds to the unexpectedly
high values of the observed rate at low methanol concentra-
tions, indicates a parallel path for consuming 2(TFE). This must
be the pentafluorophenylation of 2(TFE) to give 3, which is
observed in all samples, but is especially prominent in meth-
anol activation. The linear fit of the corrected rate constants
(kobs/RTFE) to the methanol concentration over a wide range
strongly implies that this decomposition path is also subject to
substrate inhibition; that is, both the C-H activation and the
decomposition to 3 proceed only from 2(TFE).
Keq
)
[Pttot] ) [2(TFE)] + [2(MeOR)] )
[MeOR]
[2(TFE)]‚ 1 +
(5)
(
)
Keq[TFE]
d[Pt]
dt
-
) kobs[Pttot] ) k2[2(TFE)]‚[MeOR]
(6)
(7)
1
kobs ) k2[MeOR]‚RTFE ) k2[MeOR] ‚
[MeOR]
1 +
Keq[TFE]
The dependence of observed rate constant on [MeOR]
predicted from the rate law of eq 7 is saturation kinetics; the
rate should increase linearly from zero [MeOR] and then level
off as the platinum complex is almost completely sequestered
in its unreactive MeOR-bound form. Just such behavior is
observed for the C-H activation of DME (Figure 6). A plot of
the observed rate constants corrected for the concentration of
2(TFE) at each point (kobs/RTFE), where RTFE is calculated from
the concentrations of TFE and DME, and the equilibrium
constant measured by NMR (see Table 6) versus [DME] should
be linear, as it is (Figure 7); there is a negligible intercept. The
second-order rate constant for DME C-H activation determined
from this plot is k2 ) 5.4 ( 0.5 × 10-4 M-1 s-1
.
An alternative mechanism could be invoked, in which C-H
activation of methanol and/or DME proceeds directly from the
corresponding substrate-bound cation (2(MeOD) or 2(DME))
via intramolecular â-hydride elimination. Such a process was
According to this mechanism (see Scheme 8), where both
the productive methanol activation as well as the parallel
decomposition of 2 to 3 depend on [2(TFE)], the rate expression
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2011

A R T I C L E S
Scheme 8
Owen et al.
and CH2D2 (Scheme 8). Accordingly, at low [methanol], the
methane should be mostly CH3D and CH2D2, from decomposi-
tion to 3, while at higher [methanol], there should be much less
CH2D2 and more CH4, as observed.
Pentafluorophenyl Transfer Rates. 3 is formed, in varying
amounts, in all of the reactions examined here. 2 reacts directly
with the pentafluorophenylating agent to give 3, in competition
with substrate activation; also 4, the initial product of methanol
activation, undergoes subsequent conversion to 3. Higher
concentrations of [2]+[CF3CD2OB(C6F5)3]- generally give rise
to faster decomposition to 3, as would be expected for first-
order dependence on the concentrations of both platinum and
borane, second-order overall. Exposure of solutions to water
also accelerates formation of 3.
can be modified to include a term for the pentafluorophenylation
rate (kdecomp) as shown in eq 8. After correcting for RTFE as
before (Tables 2 and 3), both rate constants may be calculated
from the slope (k2) and intercept (kdecomp) of the plot in Figure
8, yielding the following parameters (at 330K): k2(CH3OD) )
2.0 ( 0.2 × 10-3 M-1 s-1, kH/kD ) 1.4 ( 0.1, kdecomp ) 2 (
The kinetics for methanol activation, however, show com-
plexities in the decomposition process, even at constant [Pt].
The decomposition rate constant calculated from the intercept
of the plot in Figure 8 (2 ( 1 × 10-4 s-1) is 5 times larger than
that observed in the absence of substrate (4 × 10-5 s-1 (330
K)). Furthermore, as noted above (see Figure 9), the observed
rate of disappearance of 2 at low methanol concentrations (which
is largely decomposition to 3) is significantly less than that
predicted from the kinetic parameters obtained from Figure 8.
We interpret these observations in terms of an equilibrium
between two different borate anions, [(CF3CD2O)B(C6F5)3]- and
[(CH3O)B(C6F5)3]-, with the latter being a significantly more
potent pentafluorophenyl transfer agent.20 At low [methanol]
(on the same order as that of B(C6F5)3, which is added in a 2:1
ratio to [2]), the less reactive trifluoroethoxyborate dominates,
resulting in a decomposition rate slower than expected from
extrapolation of the results at higher [methanol]. The difficulty
of drying methanol thoroughly is also a likely contributor to
this behavior, as water has been seen to accelerate decomposi-
tion. The presence of different amounts of water may explain
the difference between the intercepts in the activation of
methanol-d4 and methanol-d1 (Figure 8). The reactions of DME
show no such complications, as (dry) DME would not be
expected to form more reactive adducts with the borane,
although as noted above, 6 decomposes comparatively more
rapidly via a separate pathway, namely, loss of methanol and
dimerization to 5.
We attempted a better fit of the dependence of kobs on
[methanol] by including two additional parameterssthe equi-
librium between the two borates and a second kdecompsbut
obtained only a modest improvement to the fit at low [methanol].
It is important to note, however, that this variation in the parallel
decomposition path has a minimal impact on the calculated rate
constant for methanol C-H activation. That value is extracted
from the slope of the line in Figure 8, which is primarily
determined by the high [methanol] points, where C-H activation
is dominant, and is relatively insensitive to the low [methanol]
points.
1 × 10-4 s-1
.
1
kobs ) (k2[MeOR] + kdecomp) ‚
(8)
[MeOR]
Keq[TFE]
1 +
Including the parallel decomposition path in the rate expres-
sion has a dramatic impact on the expected [methanol]
dependence of the uncorrected, observed rates. Instead of simple
saturation dependence, the observed rate will be a sum of two
termssC-H activation and decompositionsin which the first
saturates while the second falls off as [methanol] increases.19
This is shown in Figure 9, where the calculated C-H activation
(kobs-C-H ) k2[MeOD](RTFE)) and pentafluorophenylation
(kobs-C F ) kdecomp(RTFE)) rates and their sum are plotted as a
6
5
function of [methanol], along with the measured first-order rate
constants for the disappearance of the starting material. The
agreement is quite good at higher [methanol] levels, but not
for low [methanol] (see the following section for discussion).
According to this interpretation, decomposition dominates at
low [methanol] and methanol C-H activation at high [metha-
nol]. Independent support of a change in the dominant reaction
path is provided by observation of variable isotopic composition
of the liberated methane. Recall that 2 contains approximately
equal amounts of CH3 and CH2D groups, so that methane
produced by methanol-d1 C-H activation (in CF3CD2OD) will
carry a protium from the substrate yielding CH4 and CH3D,
whereas the methane produced in pentafluorophenylation re-
ceives an additional deuterium from the solvent yielding CH3D
On the other hand, it must be acknowledged that if the low
[methanol] points are left out of the analysis, the possibility of
(19) It can be seen from eq 6 that the rate can also fall off when [methanol] is
sufficiently high that [TFE] decreases significantly; however, the data in
Tables 5 and 6 show that this factor alone cannot account for the magnitude
of the rate decrease.
(20) In Suzuki couplings, which similarly involve transfer of organic groups
from boron to transition metal centers, the transmetalating boronic acids
can be activated by anions. See: Miyaura, N.; Suzuki, A. Chem. ReV. 1995,
95, 2457-2483.
Figure 9. Observed first-order rate constants for the reaction of 2 with
CH3OD (0) and calculated C-H activation (k2 × [MeOD] × RTFE) (- - -)
and pentafluorophenylation (kdecomp × RTFE) (- - -) rate constants based
on eq 8.
9
2012 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Methane, Methanol, and Dimethyl Ether C−H Activation
A R T I C L E S
Scheme 9
to the kinetics analysis because it means that every time
methanol or dimethyl ether reacts with 2 to form a σ-complex
the reaction continues to completion. Hence the correction factor
of 2 applied to methane activation is not appropriate in these
cases. This irreversibility can be accounted for by a lowered
barrier to loss of methane from the species corresponding to
structure A′ in Figure 10 (note that the reaction coordinate is
no longer symmetrical for activation of any substrate other than
methane). Methane decoordination might be facilitated by simple
steric destabilization or a more specific intramolecular displace-
ment by the pendant oxygen atom of the hydroxymethyl or
methoxymethyl group.
Alternatively, it is conceivable that C-H bond cleavage (step
ii) becomes rate-determining for these substrates, but there is
no obvious reason that should be so. Furthermore, KIEs for
methanol and DME activation are small and normal, ap-
proximately 1.4-1.5. Comparison to the much larger values
(KIE ) 4-5) found for the oxidative cleavage of various alkane
C-H bonds at titanium and rhodium9,10 suggests that, as in the
methane reaction, C-H bond coordination rather than bond
breaking is rate-determining. Several studies of equilibrium
isotope effects for C-H versus C-D coordination report values
in the range of 1.2-2, indicating only modest weakening of
the bond in the σ-complex;22,23 since relatively less bond
weakening would be expected in the transition state leading to
coordination, small normal KIEs seem reasonable. Few mea-
surements of KIEs for C-H coordination have been docu-
mented. Competition experiments for reaction at iridium give
values in this range for n-hexane (KIE ) 1.2)24 and cyclohexane
(KIE ) 1.38).25 While it is not certain that these represent C-H
coordination, the relative activation rates of linear alkanes
depend on the number of secondary C-H bonds, suggesting
that coordination rather than C-H bond cleavage is indeed rate-
determining. A recent paper reports a KIE of 3.3 for the
oxidative addition of cyclohexane to a rhodium center, but
whether this reflects coordination or C-H cleavage (or a
composite of both) could not be determined.9d Preliminary
studies on the reaction of cyclohexane with 2 yield KIE ) 1.3;
isotope scrambling patterns suggest that C-H coordination is
rate-determining.26
intramolecular activation of coordinated methanol remains open.
One could envision 2(MeOD) undergoing â-H abstraction to
eliminate methane and generate a protonated formaldehyde
complex, which, on addition of TFE or methanol, would open
to 4(L), while 2(TFE) would decompose to 3 as before. The
resulting rate law (eq 9) would be consistent with the linear
dependence of kobs on RTFE observed for the high [methanol]
points in Figure 8. We feel this scenario is much less likely on
a number of grounds. The rate constant calculated for methanol
in the proposed mechanism (Scheme 8) is quite similar to that
for DME (after correcting for temperature; see below); there is
no obvious reason why methanol and DME should follow
different routes; it is far from clear whether an elimination-
rearrangement sequence could lead to 4. Nonetheless, in contrast
to the DME case (see above), this alternative reaction scheme
cannot be dismissed on the basis of the kinetics alone.
kobs ) kdecomp‚RTFE + k2‚(1 - RTFE) )
RTFE‚(kdecomp - k2) + k2 (9)
Isotope Scrambling and the Reversibility of C-H Activa-
tion. As noted in the Introduction, C-H activation is typically
a multistep process, with reversibility and the identity of the
rate-determining step often at issue. The first two steps in C-H
activation by 2 (Scheme 9) are displacement of the coordinated
solvent by substrate (i), leading to an alkane σ-complex (A),
followed by oxidative cleavage of the coordinated C-H bond
(ii), to generate a platinum(IV) alkyl hydride (B) that goes on
to reductively eliminate methane (not shown).11 In principle
either i or ii might be the rate-determining step; previous studies
in our group on arene activation (where the detected intermediate
is a π-complex rather than A) have provided examples of
both.11a Because A is also an intermediate in the protonolysis
of 1, we can deduce from the observed statistical isotopic
scrambling in that process (see above) that displacement of
methane from A is slow compared with the rapid and reversible
C-H bond cleavage and forming steps (Scheme 10). Similarly,
activation of C6D6 by analogues of 2 gives complete isotopic
scrambling between phenyl and methyl sites in cases where step
i is rate-determining.11a,21
In the activation of methane, then, by microscopic revers-
ibility, coordination of the substrate is the slowest, rate-
determining step, as shown in the qualitative reaction coordinate
in Figure 10. By symmetry, loss of the labeled and unlabeled
methyl groups from intermediate B is equally likely; hence the
rate determined by disappearance of the labeled starting material
is only half the actual rate of methane activation. Accordingly,
the second-order rate constants given in Table 1 are twice the
apparent values.
These considerations strongly suggest that C-H coordination
is rate-determining for all three substrates studied here, methanol
and DME as well as methane, and that the irreversibility for
the first two (manifested by the absence of isotopic scrambling)
is a consequence of accelerated methane loss. We have
previously suggested that C-H coordination will be rate-
determining for reactions of aliphatic C-H bonds at platinum
even for cases where aromatic C-H cleavage becomes rate-
determining because of the poorer coordinating ability (and
consequent higher transition-state energy for substitution) of
C-H σ-bonds compared to CdC π-bonds.27 It should be noted,
though, that some unusually large inVerse EIEs (12-14) have
been reported for the binding of alkanes to [Cp′Rh(CO)] in
liquid krypton at 165 K.10 A substantial inverse EIE effect on
In contrast, activation of methanol-d4 or dimethyl ether-d6
transfers only a single deuterium from the substrate to the
liberated methane, implying one or more irreversible steps
leading to formation of products. This irreversibility is important
(22) Lawes, D. J.; Geftakis, S.; Ball, G. E. J. Am. Chem. Soc. 2005, 127, 4134-
4135.
(23) Calvert, R. B.; Shapley, J. R. J. Am. Chem. Soc. 1978, 100, 7726-7727.
(24) Periana, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1986, 108, 7332-7346.
See especially ref 10.
(25) Janowicz, A. H.; Bergman, R. G. J. Am. Chem. Soc. 1983, 105, 3929-
3939.
(21) See also: Heiber, H.; Johansson, L.; Gropen, O.; Ryan, O. B.; Swang, O.;
(26) Chen, G. C.; Labinger, J. A.; Bercaw, J. E. Unpublished results.
(27) See the Supporting Information to ref 14d for a discussion.
Tilset, M. J. Am. Chem. Soc. 2000, 122, 10831-10845.
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2013

A R T I C L E S
Scheme 10
Owen et al.
the preequilibrium binding of substrate followed by rate-
determining C-H oxidative cleavage with a slightly larger
normal KIE could give an apparent KIE in the range observed.
While such a scenario seems less probable, it cannot be ruled
out at this time.
In light of the relatively small difference between the
reactivity of methane and methanol, it is somewhat surprising
that we have never detected any reaction of the C-H or C-D
bonds of trifluoroethanol with these cationic Pt(II) complexes
(a fact that is, of course, crucial to its utility as solvent for these
model studies). From the high concentration of TFE and the
absence of any NMR-detectable (<5%) product of its activation,
we estimate an upper limit on reactivity for the C-H bond of
TFE of 1/100 that of methane, methanol, and DME. Presumably,
this is a result of the additional electron-withdrawing effect of
the trifluoromethyl group, which must destabilize the σ-complex
(or, more precisely, the transition state leading to it) by >2-3
kcal/mol. A major role for steric effects is harder to argue since
the van der Waals radius of fluorine (1.47 Å) is only slightly
larger than that of hydrogen (1.35 Å).28
Even more interesting is the high selectivity of the Periana
system for the oxidation of methane over methyl bisulfate
(∼100:1).2c It is not known whether the rate-determining step
in that system is the same as in our models, but one might
imagine that the electron-withdrawing capacity of the bisulfate
group works in concert with its “neopentyl-like” steric influence
to disfavor methyl bisulfate relative to methane activation. These
conclusions point to the importance of the fuming sulfuric acid
solvent in the Periana system. Under these conditions, the
formation of the methane oxidation product as a “protected”
form of methanol is clearly crucial; the electronic effect of
simple hydroxy (or methoxy) substitution is not sufficient alone
to influence the relative rate of C-H bond coordination so
strongly. The only example of strong (7:1) preference for acti-
vation of methane over methanol is found for a dimeric Rh(por-
phyrin) system, where the large KIEs (9-11) point to C-H
bond cleavage as the rate-limiting step, and kinetics implicate
a trimolecular transition state that would be expected to be
subject to strong steric influence.29
Conclusions: Substrate Selectivity. The relative rates of
methane, methanol, and DME C-H activation show that the
platinum center is relatively unselective: kmethane/kmethanol ) 1/1.3,
330 K; and kmethane/kdimethyl ether ) 1/2, 313 K. These values all
appear to represent rates of displacement of solvent by the
substrate’s C-H bond. The low selectivities in this model
system are consistent with those inferred from the behavior of
the “real” Shilov system (see above) and suggest that C-H
coordination is probably rate-determining there, as well. Con-
sistent with that suggestion, it has been found that during
catalytic H/D exchange at Pt(II) (in the absence of an oxidant)
there is multiple exchange during each C-H activation event.7e
(This result should be interpreted with caution, however, because
Pt(0), which can form under these conditions, efficiently
scrambles H with D.)
As noted in the Introduction, the potential yield of a methane
oxidation product, such as methanol, is constrained by relative
reactivities. In a case such as that studied here, where the solvent
form of the activating complex is less stable but more reactive
than the methanol-bound form, the rate of methane activation
will decrease as [methanol] increases, whereas that of methanol
(28) Slater, J. C. J. Chem. Phys. 1964, 41, 3199.
Figure 10. Qualitative potential energy surface for the reaction of 2 with
methane.
(29) Cui, W.; Zhang, P. Wayland, B. B. J. Am. Chem. Soc. 2003, 125, 4994-
4995.
9
2014 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Methane, Methanol, and Dimethyl Ether C−H Activation
A R T I C L E S
128.1 Hz, JPt-H ) 87.53 Hz, 12H), 2.05 (s, JPt-H ) 20.32 Hz, 12H).
13C NMR (75.4 MHz, C6D6): δ ) -6.41 (JPt-C ) 791.7 Hz).
[(N-N)Pt(13CH3-nDn)(DOCD2CF3)]+[CF3CD2OB(C6F5)3]- (2-13C-
(TFE)) was prepared as previously described for the unlabeled
compound.11d The 1H NMR signal for the Pt-CH3 group (at δ ) 0.74
ppm) appeared as a doublet, 1JC-H ) 128.3 Hz; the corresponding value
1
for JC-D ) 19.5 Hz was determined from the13C NMR signal for the
Pt-CH2D group (δ ) -8.6 ppm).
Kinetics of Methane Activation. Stock solutions of 2-13C(TFE)
were prepared by weighing 1-13C2 and tris(pentafluorophenyl)borane
(2 equiv) into 5.0 mL volumetric flasks and adding TFE-d3 (ap-
proximately 3 mL). The suspension was mixed by shaking until the
solids had dissolved. Additional TFE-d3 was added to dilute the sample
to 5.0 mL, the solution mixed and then transferred to a vial and stored
at -30 °C in the glovebox freezer. Solutions prepared from anhydrous
reagents in this way are stable for days, with minimal decomposition
(∼5-10%) after 1 week; 400 µL of solution was measured into a
sapphire NMR tube using a volumetric syringe, the cap tightly closed
and the sample taken to the high-pressure manifold. After evacuating
the manifold, the tube was charged with the desired pressure of dry
methane gas, and the tube inserted in the NMR spectrometer that had
been preheated to 330 ( 0.5 or 313 ( 0.5 K as determined by an
ethylene glycol temperature calibration standard. After allowing the
sample to reach the probe temperature (5 min), the tube was removed,
mixed by rocking, and reinserted into the probe. The 1H NMR spectrum
was recorded to measure the methane concentration versus the known
concentration of platinum in the solution. Kinetics were monitored by
following the peak height of the 13C-labeled platinum methyl signal in
the 13C NMR; parallel determinations using integrated peak areas gave
very similar, though somewhat noisier, results. After the run, the sample
height was measured to (0.5 mm and compared with the starting height
to correct the concentration for the change in volume due to the
dissolved gas.
Figure 11. Comparison of the apparent first-order rate constants for C-H
activation of methane (solid line, at constant [methane] ) 1.0 M) and
methanol (dashed line) by 2.
activation will level off. This is illustrated for the case of 2 in
Figure 11, where the apparent pseudo-first-order rate constants
for the two process (neglecting any contribution from the
decomposition to 3) are calculated as a function of [methanol]
from the kinetic parameters derived above. It is clear that with
relative reactivities on this order, it would be necessary to keep
the methanol concentration low, perhaps by some process that
permits continual removal as it is formed.
Current ongoing work in our group is aimed at extending
these findings to C-H bonds in a wider range of hydrocarbons
and functionalized hydrocarbons.
Experimental Section
General Methods. All air and/or moisture-sensitive compounds were
manipulated using standard Schlenk techniques or in a glovebox under
a nitrogen atmosphere, as described previously.30 1 was prepared as
described previously11a and purified by filtering a saturated methylene
chloride solution followed by precipitation of the platinum complex at
-78 °C. The dark purple solid was stored under high vacuum for 3
days to remove cocrystallized solvent molecules and then in a P2O5
desiccator in the glovebox. B(C6F5)3 was sublimed at 90 °C in vacuo
and stored in a P2O5 desiccator in the glovebox. All deuterated solvents
and 13C-labeled compounds were purchased from Cambridge Isotope
Laboratories, with the exception of dimethyl ether-13C2, which was
purchased from Isotec. Methanol-d1 and methanol-d4 were stored over
3 Å molecular sieves and then distilled from sodium. Trifluoroethanol-
d3 was dried over 3 Å molecular sieves for at least 5 days and then
was vacuum distilled onto B(C6F5)3 and shortly thereafter distilled into
a Strauss flask and stored in the glovebox. Ultrahigh purity methane
(99.95%) was purchased from Matheson and stored over activated
alumina prior to use. Dimethyl ether was purchased from Matheson
and stored as a solution in tetraglyme over sodium and benzophenone
at <1 atm of vapor pressure for at least 1 week prior to its use. NMR
spectra were recorded on Varian Mercury 300, Varian INOVA 500,
and Varian Innova 600 spectrometers.
CAUTION: Handling high pressures of gas in sapphire NMR tubes
is extremely dangerous and should be approached with caution. Many
steps were taken to minimize exposure to the tubes while under pressure.
The tubes and high-pressure manifold used in this work were tested
regularly at pressures well above those employed under working
conditions.
Reaction of 2 with Methanol. Addition of MeOD to solutions of
2(TFE) in TFE resulted in the immediate generation of a new set of
1
NMR signals corresponding to 2(MeOD): H NMR (600 MHz, TFE-
d6) δ ) 0.75 (s, 3H), 1.35 (s, 9H), 1.37 (s, 9H), 1.81 (s, 3H), 2.02 (s,
3H), 3.04 (s, 3H), 6.84 (s, 2H), 7.06 (s, 2H), 7.52 (s, 1H), 7.60 (s, 1H);
the equilibrium constant relating the two solvated complexes was
determined from the relative peak intensities as a function of [MeOD].
On prolonged standing at room temperature (faster at 330 K), the NMR
signals for 2 were replaced by those for 3 and (in some cases) 4, as
described in the Results section.
Kinetics of Methanol Activation. Stock solutions were made by
weighing 1 and tris(pentafluorophenyl)borane (2 equiv) into 1.0 mL
volumetric flasks and adding TFE-d3 (approximately 0.5 mL). The
suspension was shaken until all solids had dissolved. After the solids
had dissolved, the desired amount of methanol was added with a
microliter syringe followed by enough TFE-d3 to dilute to 1.0 mL total
volume. The concentration of TFE was calculated from its known
density, assuming the volumes of TFE and methanol are additive; 700
µL was measured into a J. Young NMR tube using a volumetric syringe,
the cap tightly closed and the tube inserted in the NMR spectrometer
that had been preheated to 330 ( 0.5 K. Kinetics were monitored by
[(µ-SMe2)Pt(13CH3)2]. [(µ-SMe2)Pt(13CH3)2] was prepared similarly
to the unlabeled version31 using 13CH3Li‚LiI (prepared from 13CH3I
and LiBut in pentane). The reaction mixture and product were slightly
darker than usual, so the product was dissolved in THF and additional
carbon was added. The suspension was filtered, and the product
reprecipitated with petroleum ether at -78 °C. This procedure was
repeated a second time, finally yielding 1.09 g (49% based on platinum)
1
of white powder. H NMR (300 MHz, C6D6): δ ) 1.00 (m, JC-H
)
1
following the disappearance of the starting material in the H NMR
spectrum.
(30) Burger, B. J.; Bercaw, J. E. In New DeVelopments in the Synthesis,
Manipulation, and Characterization of Organometallic Compounds; Wayda,
A., Darensbourg, M. Y., Eds.; American Chemical Society: Washington,
D.C., 1987; Vol. 357.
(31) Hill, G. S.; Irwin, M. J.; Levy, C. J.; Rendina, L. M.; Puddephatt, R. J.
Inorg. Synth. 1998, 32, 149-153.
Reaction of 2 with Dimethyl Ether. Dimethyl ether was distilled
from Na/benzophenone in tetraglyme into a liquid nitrogen-cooled
receiver. Frozen DME was maintained under vacuum to remove any
volatiles and condensed into a degassed J. Young tube containing 0.7
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2015

A R T I C L E S
Owen et al.
mL of 2(TFE) in TFE-d3 by cooling with an ice or dry ice bath. The
is 90% 5. 1H NMR (500 MHz, acetone-d6): δ ) 1.278 (s, 9H), 1.283
(s, 9H), 1.29 (s, 9H), 1.30 (s, 9H), 2.25 (s, 3H), 2.26 (s, 3H), 3.79 (s,
3H), 6.78 (at, JH-H ) 1.8 Hz, 1H), 6.83 (at, JH-H ) 1.8 Hz, 1H), 6.91
(at, JH-H ) 1.8 Hz, 1H), 6.93 (at, JH-H ) 1.8 Hz, 1H), 7.46 (m, 2H),
8.01 (d, JH-H ) 5.25 Hz, 1H), 8.15 (dd, JH-H ) 5.25 Hz, JH-H ) 1.74
Hz, 1H), 10.35 (d, JH-H ) 1.74 Hz, 1H); HRMS (FAB) m/z calcd for
[C67H102N4OPt2]+ 1368.735, found 1368.733.
NMR spectra of these solutions show a new set of signals corresponding
1
to 2(DME): H NMR (600 MHz, TFE-d6) δ ) 0.85 (s, 3H), 1.37 (s,
9H), 1.38 (s, 9H), 1.89 (s, 3H), 1.99 (s, 3H), 6.84 (s, 2H), 7.04 (s, 2H),
7.54 (s, 1H), 7.54 (s, 1H); the equilibrium constant relating the two
solvated complexes was determined from the relative peak intensities
as a function of [DME]. At 313 K, signals for 2 gradually disappeared
and were replaced by a new set of signals corresponding to 5, which
in turn decreased as crystals of 5 began to precipitate. Formation of 5
(by NMR) is nearly quantitative; isolated yields of crystalline 5 are
generally 50% or less, depending on concentrations.
X-ray Structure of 5. A crystal obtained as described above was
analyzed. Best results were obtained by refining the structure in space
group P1h, wherein the Pt dimer sits on a center of symmetry, with the
µ-CH2 and µ-CH(OCH3) groups disordered with respect to this center
of symmetry; the methoxy group occupies each of the two sites with
equal probability. Attempts to solve the structure in P1 gave less reliable
results. Full details of the solution and refinement are available.14
Kinetics of Dimethyl Ether Activation. Solutions of 2(TFE) and
DME in J. Young NMR tubes were prepared as described in the
preceding section. After allowing to warm to room temperature and
mixing, the sample height was measured to (0.5 mm and compared
with the starting height to correct the concentration of 2(TFE) for the
change in volume due to the dissolved gas. The concentration of added
DME was determined by NMR, by comparing the integral of free DME
with the integral of 2. Tubes were inserted in the NMR spectrometer
that had been preheated to 313 ( 0.5 K, and kinetics were monitored
by following the disappearance of the starting material in the 1H NMR
spectrum.
Acknowledgment. We thank Michael Day and Larry Henling
for assistance with X-ray crystallography, and Dan Nieman,
Dean Roddick, Steve Olson, Mike Roy, David Law, Glenn
Sunley, and Marc Payne for assistance with design and
construction of the high-pressure NMR equipment. We are
grateful to bp for financial support through the MC2 program.
[(Diimine)Pt(µ-CH2)(µ-(CH(OCH3))Pt(diimine)]2+[CF3CD2OB-
(C6F5)3]- (5). Crystals that precipitated at the end of dimethyl ether
2
Supporting Information Available: Complete author list for
ref 1. This material is available free of charge via the Internet
at http://pubs.acs.org.
activations were isolated by filtration and dried under vacuum. These
crystals were very soluble in organic solvents, including diethyl ether.
Removing the solvent from these solutions gave oily residues rather
than solids. Redissolving the crystals in acetone-d6 gave a sample that
JA056387T
9
2016 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006