A combined experimental and density functional theory investigation of hydrocarbon activation at a cationic platinum(II) diimine aqua complex under mild conditions in a hydroxylic solvent

Article abstract of DOI:10.1021/ja0019171

Controlled protonolysis of (N^f-N^f)Pt(CH₃)₂ (1; N^f-N^f = ArN=CMe-CMe=NAr, Ar = 3,5-(CF₃)₂C₆H₃) with HBF₄·Et₂O in dichloromethane in the presence of small quantities of water gives the BF₄^- salt of the aqua complex (N^f-N^f)Pt(CH₃)(H₂O)⁺ (6). When dissolved in trifluoroethanol (TFE), 6(BF₄^-) effects the activation of methane and benzene C-H bonds under very mild conditions. Thus, 6 reacted with benzene in TFE-d₃ at ambient temperature to quantitatively yield (N^f-N^f)Pt(C₆H₅)(H₂O) ⁺ and methane after 2-3 h. The use of C₆D₆ led to multiple incorporation of deuterium into the methane produced and suggests the involvement of methane σ-complex and benzene σ-or π-complex intermediates. When the solution of 6(BF₄^-) was exposed to ¹³CH₄, an exchange reaction produced ca. 50% of (N^f-N^f)Pt(¹³CH₃)(H ₂O)⁺ and CH4 after ca. 48 h at 45°C. The reaction was inhibited by added water, suggesting that water is reversibly lost from 6 before C-H activation takes place. The use of CD₄ resulted in multiple deuterium incorporation into the methane produced, again implying a Pt-methane σ-complex intermediate. Low-temperature protonation of 1 in dichloromethane-d₂ generated observable Pt(IV) hydride species (N^f-N^f)Pt(CH₃)₂(H)(L)⁺. These decomposed via methane elimination, raising the possibility that the observed C-H activation proceeds by an oxidative addition pathway. The reaction between 6 and CH₄ was investigated by DFT calculations using a model system with the HN=CH-CH=NH ligand. The C-H activation was investigated for oxidative addition and cr-bond metathesis pathways starting from the four-coordinate methane complex (N-N)Pt(CH₃)(CH₄)⁺. The oxidative addition pathway, thermodynamically uphill by 23 kJ/mol (ZPE-corrected data), was favored by 12 kJ/mol relative to the σ-bond metathesis. When a H₂O ligand was added to the five-coordinate oxidative addition product, the overall oxidative addition reaction was thermodynamically downhill by 33 kJ/mol (partially ZPE-corrected) starting from an H₂O adduct of (N-N)Pt(CH₃)(CH₄)⁺ with H₂O electrostatically bonded at the diimine moiety. In this case, the oxidative addition pathway was favored by 20 kJ/mol. The calculations indicated that reductive elimination of methane from the six-coordinate (N-N)Pt(CH₃)₂(H)(H2O)₊ with the hydride and H₂O ligands trans disposed occurred in concert with dissociation of the aqua ligand.

Full text of DOI:10.1021/ja0019171

J. Am. Chem. Soc. 2000, 122, 10831-10845
10831
A Combined Experimental and Density Functional Theory
Investigation of Hydrocarbon Activation at a Cationic Platinum(II)
Diimine Aqua Complex under Mild Conditions in a Hydroxylic
Solvent
Hanne Heiberg,1a Lars Johansson,1b Odd Gropen,1a Olav B. Ryan,1c Ole Swang,1c and
Mats Tilset*1b,d
Contribution from the Department of Chemistry, UniVersity of Oslo, P.O. Box 1033 Blindern,
N-0315 Oslo, Norway, Department of Chemistry, UniVersity of Tromsø, N-9037 Tromsø, Norway, and
SINTEF Applied Chemistry, Department of Hydrocarbon Process Chemistry, P.O. Box 124 Blindern,
N-0314 Oslo, Norway
ReceiVed May 31, 2000
Abstract: Controlled protonolysis of (Nf-Nf)Pt(CH3)2 (1; Nf-Nf ) ArNdCMesCMedNAr, Ar ) 3,5-
-
(CF3)2C6H3) with HBF4‚Et2O in dichloromethane in the presence of small quantities of water gives the BF4
salt of the aqua complex (Nf-Nf)Pt(CH3)(H2O)+ (6). When dissolved in trifluoroethanol (TFE), 6(BF4-) effects
the activation of methane and benzene C-H bonds under very mild conditions. Thus, 6 reacted with benzene
in TFE-d3 at ambient temperature to quantitatively yield (Nf-Nf)Pt(C6H5)(H2O)+ and methane after 2-3 h.
The use of C6D6 led to multiple incorporation of deuterium into the methane produced and suggests the
-
involvement of methane σ-complex and benzene σ- or π-complex intermediates. When the solution of 6(BF4
)
was exposed to 13CH4, an exchange reaction produced ca. 50% of (Nf-Nf)Pt(13CH3)(H2O)+ and CH4 after ca.
48 h at 45 °C. The reaction was inhibited by added water, suggesting that water is reversibly lost from 6
before C-H activation takes place. The use of CD4 resulted in multiple deuterium incorporation into the
methane produced, again implying a Pt-methane σ-complex intermediate. Low-temperature protonation of 1
in dichloromethane-d2 generated observable Pt(IV) hydride species (Nf-Nf)Pt(CH3)2(H)(L)+. These decomposed
via methane elimination, raising the possibility that the observed C-H activation proceeds by an oxidative
addition pathway. The reaction between 6 and CH4 was investigated by DFT calculations using a model system
with the HNdCHsCHdNH ligand. The C-H activation was investigated for oxidative addition and σ-bond
metathesis pathways starting from the four-coordinate methane complex (N-N)Pt(CH3)(CH4)+. The oxidative
addition pathway, thermodynamically uphill by 23 kJ/mol (ZPE-corrected data), was favored by 12 kJ/mol
relative to the σ-bond metathesis. When a H2O ligand was added to the five-coordinate oxidative addition
product, the overall oxidative addition reaction was thermodynamically downhill by 33 kJ/mol (partially ZPE-
corrected) starting from an H2O adduct of (N-N)Pt(CH3)(CH4)+ with H2O electrostatically bonded at the diimine
moiety. In this case, the oxidative addition pathway was favored by 20 kJ/mol. The calculations indicated that
reductive elimination of methane from the six-coordinate (N-N)Pt(CH3)2(H)(H2O)+ with the hydride and H2O
ligands trans disposed occurred in concert with dissociation of the aqua ligand.
Introduction
of the corresponding chlorides and alcohols could be achieved
by employing aqueous solutions of Pt(II) and Pt(IV) salts.5
Stoichiometric additions of alkane C-H bonds at highly
reactive, coordinatively and electronically unsaturated organo-
transition metal species such as Cp*Rh(PMe3), Cp*Ir(PMe3),
and Cp*Ir(CO) were achieved in the 1980s,6,7 but the conversion
The direct, selective conversion of alkanes into value-added
chemical products has long been a “holy grail” for chemists,2
and intensified efforts toward this goal have been made by
organometallic chemists during the past few years.3 Shilov and
co-workers demonstrated as early as 1969 that transition metal
complexes in the form of Pt(II) salts were capable of activating
alkane C-H bonds.4 Some years later, Shilov also reported that
catalytic conversion of alkanes (including methane) to mixtures
(4) Goldshlegger, N. F.; Tyabin, M. B.; Shilov, A. E.; Shteinman, A. A.
Zh. Fiz. Khim. 1969, 43, 2174.
(5) Goldshlegger, N. F.; Eskova, V. V.; Shilov, A. E.; Shteinman, A. A.
Zh. Fiz. Khim. 1972, 46, 1353.
(6) For some early references, see: (a) Janowicz, A. H.; Bergman, R.
G. J. Am. Chem. Soc. 1982, 104, 352. (b) Wax, M. J.; Stryker, M. J.;
Buchanan, J. M.; Kovac, C. A.; Bergman, R. G. J. Am. Chem. Soc. 1984,
106, 1121. (c) Periana, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1984,
106, 7272. (d) Hoyano, J. K.; Graham, W. A. G. J. Am. Chem. Soc. 1982,
104, 3723. (e) Jones, W. D.; Feher, F. J. J. Am. Chem. Soc. 1982, 104,
4240.
(1) (a) University of Tromsø. (b) University of Oslo. (c) SINTEF. (d)
E-mail for corresponding author: mats.tilset@kjemi.uio.no.
(2) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H.
Acc. Chem. Res. 1995, 28, 154.
(3) (a) Shilov, A. E.; Shulpin, G. B. Chem. ReV. 1997, 97, 2879. (b)
Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. Angew. Chem., Int. Ed. Engl.
1998, 37, 2180. (c) Bengali, A. A.; Arndtsen, B. A.; Burger, P. M.; Schultz,
R. H.; Weiller, B. H.; Kyle, K. R.; Moore, C. B.; Bergman, R. G. Pure
Appl. Chem. 1995, 67, 281. (d) Crabtree, R. H. Chem. ReV. 1995, 95, 987.
(e) Sen, A. Acc. Chem. Res. 1998, 31, 550.
(7) Abbreviations: Cp* ) (η5-C5Me5; tmeda ) tetramethylethylenedi-
amine; Tp′ ) hydridotris(3,5-dimethylpyrazolyl)borate; bpym ) 2,2′-bi-
pyrimidine; OTf- ) triflate; BArf- ) [3,5-(CF3)2C6H3]4B-; Nf-Nf ) ArNd
CMeCMe)NAr with Ar ) 3,5-(CF3)2C6H3; TFE ) trifluoroethanol.
10.1021/ja0019171 CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/21/2000

10832 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
of the C-H insertion organometallic complexes (eq 1) to organic
products R-X has been difficult to achieve, in particular with
catalysis in mind. The high air sensitivity of the neutral, low-
with Ar ) p-MeC6H4 or p-MeOC6H4), in acetonitrile. For the
diimine (p-MeOC6H4)NdCHsCHdN(p-MeOC6H4), the oxida-
tion resulted in intermolecular methyl-transfer reactions, giving
(diimine)PtIV(NCMe)(CH3)3+ and (diimine)PtII(NCMe)(CH3)+.
The latter was presumably formed via the intermediacy of
(diimine)Pt(CH3)+, an analogue of the species believed to be
responsible for the C-H activation in the (tmeda)PtII system.9
In this contribution, we report experimental and theoretical
details about the activation of benzene and methane C-H bonds
at a related (diimine)Pt aqua complex in a hydroxylic solvent,
under mild and neutral conditions.13
oxidation-state organometallic complexes constitutes one major
problem for the conversion of alkanes to valuable products under
oxidizing conditions. Different approaches, some including Pt
complexes that may be related to the classical Shilov systems,
have now emerged that appear to hold some promise for future
developments.
Results and Discussion
A novel approach to alkane C-H activation and functional-
ization involves the use of cationic, relatively electrophilic metal
complexes. Burger and Bergman8a reported that Cp*Ir(PMe3)-
(CH3)(OTf) activates C-H bonds in methane (45 °C, CD2Cl2)
and a wide variety of other hydrocarbons. Anion exchange with
Na+BArf- in dichloromethane yielded the crystalline salt
Cp*Ir(PMe3)(CH3)(ClCH2Cl)+BArf-, characterized by X-ray
crystallography. This salt was shown to activate methane and
other hydrocarbon C-H bonds at subambient temperatures in
dichloromethane.8b The presumed crucial unsaturated intermedi-
ate in the C-H activation process, Cp*Ir(PMe3)(CH3)+,8c is
believed to be readily accessible through dissociation of the
weakly coordinated dichloromethane and triflate ligands from
Cp*Ir(PMe3)(CH3)(ClCH2Cl)+ and Cp*Ir(PMe3)(CH3)(OTf),
respectively.
Recent work by the Bercaw and Labinger group9 has also
exploited unsaturated, electrophilic metal centers, in this case
based on Pt(II) (with obvious implications for the development
of chemistry based on the original work by Shilov). Thus,
exchange of labeled 13CH3 for CH3 between (tmeda)PtII(NC5F5)-
(CH3)+ and 13CH4 was observed in pentafluoropyridine, a
solvent of low basicity and coordinating ability, and this
demonstrated the activation of methane C-H bonds at Pt.9b,c
Further demonstrating the potential for alkane activation at Pt
complexes, Wick and Goldberg showed through the reaction
between (η2-Tp′)PtII(CH3) and various hydrocarbons R-H that
it is possible to trap the normally elusive five-coordinated
oxidative addition product by intramolecular pyrazole coordina-
tion, ultimately producing Pt(IV) complexes (η3-Tp′)Pt(H)(CH3)-
(R).10 Recently, a remarkable Pt-catalyzed conversion of
methane to methyl bisulfate was achieved with the complex
(bpym)PtCl2 as a catalyst precursor in concentrated sulfuric acid
media.11 Importantly, the latter system works under oxidizing
conditions, an important detail with respect to catalytic func-
tionalization of C-H bonds.
I. Experimental Work. Ligand and Complex Synthesis.
We recently described12 the preparation and electrochemical
behavior of a series of (diimine)Pt(CH3)2 complexes, where
diimine ) ArNdCHsCHdNAr or ArNdCMesCMedNAr,
with Ar ) p-MeC6H4 or p-MeOC6H4. The protonation of these
(diimine)Pt(CH3)2 complexes with 1 equiv of HOTf, HBF4, or
H(Et2O)2BArf in poorly coordinating solvents (dichloromethane-
d2, nitromethane-d3) immediately generated methane and several
unidentified cationic (diimine)Pt(CH3)+ species that underwent
extensive decomposition, as evidenced by the plethora of new
signals that appeared in the 1H NMR spectra. The decomposition
of the cationic species was accompanied by the formation of
additional quantities of CH4. We tentatively attribute this
behavior to a propensity of the in situ-generated (diimine)Pt-
(CH3)+ species to undergo intermolecular reactions with the
diimine ligands of adjacent Pt complexes, and reasoned that
this undesirable degradation might be avoided by appropriate
protection of the N-aryl groups of the diimine. A 3,5-bis-
(trifluoromethyl)phenyl-substituted diimine (henceforth to be
denoted Nf-Nf) was selected as a new ligand, and this ligand is
synthesized in a modest (26%) yield by a modification of
methods described for the preparation of other diimines.14 The
corresponding dimethylplatinum complex (Nf-Nf)Pt(CH3)2 (1)
is straightforwardly prepared in 91% yield by the reaction of
Nf-Nf with Pt2(CH3)4(µ-SMe2)2, a standard preparative route to
(diimine)Pt(CH3)2 complexes.15 Similarly, the diphenyl analogue
16
(Nf-Nf)Pt(C6H5)2 (2) can be prepared from Pt(SMe2)2(C6H5)2
1
and Nf-Nf. The H and 19F NMR spectra of 1 and 2 (Table 1)
are in complete agreement with the anticipated square planar
geometry and the overall C2V symmetry of the molecules.
An interesting and common feature in the above examples is
the success of employing bidentate nitrogen ligands to obtain
Pt(II) complexes that are reactive toward C-H bond activation.
We have recently12 reported a study of the one-electron
oxidation reactions of Pt(II) complexes (diimine)Pt(CH3)2
(diimine ) ArNdCHsCHdNAr or ArNdCMesCMedNAr,
In reactive, cationic derivatives of 1, the two CF3 groups
should serve to sterically as well as electronically protect each
aryl ring against electrophilic attack by adjacent Pt species. As
an added benefit, it was reasoned that the CF3 groups also
decrease the electron density and thus increase the electrophi-
licity of the metal center, an effect which might increase the
affinity of the putative three-coordinate cationic intermediate
(diimine)Pt(CH3)+ for hydrocarbon C-H bonds. The electronic
(8) (a) Burger, P.; Bergman, R. G. J. Am. Chem. Soc. 1993, 115, 10462.
(b) Arndtsen, B. A.; Bergman, R. G. Science 1996, 270, 1970. (c) Luecke,
H. F.; Bergman, R. G. J. Am. Chem. Soc. 1997, 119, 11538.
(9) (a) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc.
1996, 118, 5961. (b) Holtcamp, M. W.; Labinger, J. A.; Bercaw, J. E. J.
Am. Chem. Soc. 1997, 119, 848. (c) Holtcamp, M. W.; Henling, L. M.;
Day, M. W.; Labinger, J. A.; Bercaw, J. E. Inorg. Chim. Acta 1998, 270,
467.
(13) Part of the work has been previously communicated. Johansson,
L.; Ryan, O. B.; Tilset, M. J. Am. Chem. Soc. 1999, 121, 1974.
(14) tom Dieck, H.; Svoboda, M.; Grieser, T. Z. Naturforsch. 1981, 36B,
823.
(10) Wick, D. D.; Goldberg, K. I. J. Am. Chem. Soc. 1997, 119, 10235.
(11) Periana, R. A.; Taube, J. D.; Gamble, S.; Taube, H.; Satoh, T.; Fujii,
H. Science 1998, 280, 560.
(12) Johansson, L.; Ryan, O. B.; Rømming, C.; Tilset, M. Organome-
tallics 1998, 17, 3957.
(15) Scott, J. D.; Puddephatt, R. J. Organometallics 1983, 2, 1643.
(16) Hadj-Bagheri, N.; Puddephatt, R. J. Polyhedron 1988, 7, 2695.

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
Table 1. Key NMR Data for New Complexes
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10833
1H NMR, δ
Pt-Me
aryl-H
(2,6; 4)
compound
(2J(195Pt-H))
diimine-Me
other
19F NMR, δ
free liganda
2.16
1.35
1.82
2.02, 2.13
7.22; 7.65
7.57; 7.89
7.26; 7.60
7.59, 7.74;
8.01, 8.04
7.27, 7.65;
7.81, 8.07
7.62, 7.71;
7.96, 8.01
7.58, 8.00;
7.71, 8.00
7.58, 7.96;
7.70, 7.96
7.29, 7.65;
7.79, 8.05
7.36, 7.92
7.36, 7.71
-63.50
-63.14
-63.21
(Nf-Nf)Pt(CH3)2 (1)b
1.14 (87.2)
0.71 (74.5)
(Nf-Nf)Pt(C6H5)2 (2)b
6.42-6.73 (Ph)
2.10 (MeCN)
(Nf-Nf)Pt(CH3)(NCMe)+ (3)c
-63.64, -63.41
(Nf-Nf)Pt(C6H5)(NCMe)+ (4)c
(Nf-Nf)Pt(CH3)(OTf) (5)b
2.14, 2.28
1.65, 2.01
1.80, 2.05
1.81, 2.07
1.91, 2.19
2.01 (MeCN);
6.65-6.75 (Ph)
-63.63, -63.37
1.07 (74.9)
0.79 (73.3)
0.69 (73.5)
-78.48 (4J(195Pt-F) ) 14.5 Hz),
-63.47, -63.36
-63.64, -63.52
(Nf-Nf)Pt(CH3)(OH2)+ (6)c
(Nf-Nf)Pt(CH3)(OH2)+ (6)d
(Nf-Nf)Pt(C6H5)(OH2)+ (7)c
(Nf-Nf)Pt(CH3)2(H)(L)+ (8a,b)
6.51 (H2O)
6.65-6.85 (Ph)
-63.65, -63.48
a: 0.54 (63.6)
b: 0.48 (62.9)
2.42
2.37
-26.67 (1J ) 1741 Hz)
-26.40 (1J ) 1699 Hz)
a Chloroform-d. b Dichloromethane-d2. c Trifluoroethanol-d3. d Dichloromethane-d2 with 250 equiv of TFE.
obtained after removal of the solvent. The NMR spectra of this
solid in dichloromethane-d2 confirmed that 5 indeed was the
major product (ca 90% of isolated mixture). The Pt-CH3 group
effect of the ligand is corroborated by the measured reversible
electrode potentials for the first ligand-centered reductions of
(ArNdCMesCMedNAr)Pt(CH3)2 by cyclic voltammetry in
acetonitrile/0.1 M Bu4NPF6: Ar ) p-MeC6H4, E° ) -1.75 V;12
p-MeOC6H4, -1.76 V;12 and 3,5-(CF3)2C6H3, -1.42 V vs
Cp2Fe/Cp2Fe+. Admittedly, the electron density at the metal
would be better probed through the corresponding presumably
metal-centered oxidation potentials. Unfortunately, the oxidation
waves for these dimethyl complexes12 are usually rather ill-
defined and poorly reproducible due to fouling of the electrodes
by adsorption, and are also chemically irreversible even at
voltage scan rates as high as 100 V/s, so thermodynamically
meaningful oxidation potential data are not available.
shows a diagnostic 1H NMR singlet with 195Pt satellites (2J(195
-
1
Pt-H) ) 74.9 Hz) at δ 1.07 in the H NMR spectrum. In the
19F NMR spectrum of 5, the triflate resonance appears at δ
-78.48 (integrating for 3 F; arene CF3 groups serve as an
internal standard). Most notably, satellites due to coupling
from the triflate CF3 group to 195Pt can clearly be obserVed
(4J(195Pt-F) ) 14.5 Hz). This unambiguously establishes that
the triflate group is covalently bonded to the Pt center in
solution.17 As has been pointed out in a review,18 it is frequently
nearly impossible to determine from spectroscopic data (IR,
NMR) alone whether a triflate complex exists in the covalently
bonded or dissociated form. The possible observation of
4J(M-F) couplings can be a powerful tool for resolving this
ambiguity. It appears that 19F NMR spectroscopy is a surpris-
ingly underutilized tool in the study of organotransition metal
triflate complexes.
Protonolysis of 1 and 2 with HBF4‚Et2O in the presence of
acetonitrile (eq 2) generates the expected acetonitrile adducts
(Nf-Nf)Pt(CH3)(NCMe)+ (3) and (Nf-Nf)Pt(C6H5)(NCMe)+ (4),
presumably via protonation at Pt and reductive elimination of
methane and benzene, respectively. The 1H and 19F NMR data
The triflate complex 5 is rather unstable in dichloromethane
1
at ambient temperature. In the H and 19F NMR spectra, the
signals for 5 decreased to ca. one-third of their original
intensities over a period of 62 h. The degradation of 5 is
accompanied by evolution of CH4 and formation of a brown
precipitate. An NMR analysis of the precipitate in acetone-d6
revealed a complex mixture of unidentifiable compounds. Thus,
the decomposition of 5 appears to be less selective than that of
the related compound (tmeda)Pt(CH3)(OEt2)+(BArf-), which
undergoes chloride abstraction in dichloromethane to eventually
give the dimeric product [(tmeda)PtCl]22+(BArf-)2 in a relatively
clean reaction.9b,c The instability of 5 in dichloromethane
contrasts the properties of Bergman’s reactive, yet isolable,
Cp*Ir(PMe3)(CH3)(ClCH2Cl)+BArf- complex.8b It is also in-
teresting to note that Kubas and co-workers have recently
characterized a stable, cationic PtII dichloromethane complex,
are consistent with the reduced symmetry of 3 and 4 when
compared to that of 1 and 2. Complexes 3 and 4 were prepared
to facilitate the indirect identification of other derivatives (vide
infra). They are, as expected, found to be rather unreactive with
respect to C-H activation chemistry, presumably due to strong
binding of the MeCN ligands and thence difficulties with
providing the necessary coordination site for substrate binding.
For example, 3 reacts slowly with benzene only at elevated
temperatures (90 °C) in dichloromethane or nitromethane.
Attempted Synthesis and Isolation of the Triflate Complex
(Nf-Nf)Pt(CH3)(OTf) (5). Having as the ultimate goal an isol-
able complex that in solution would activate alkane C-H bonds,
we attempted to prepare the PtII triflate complex (Nf-Nf)Pt(CH3)-
(OTf) (5) by protonolysis of (Nf-Nf)Pt(CH3)2 (1) with 1 equiv
of HOTf in dichloromethane (eq 3). An orange solid was
(17) For other (diimine)PtII(OTf) and (diimine)PtIV(OTf) complexes
described in the literature, corresponding triflate 4J(195Pt-F) coupling
constants have, to our knowledge, never been observed or reported. See,
for example: (a) Hill, G. F.; Rendina, L. M.; Puddephatt, R. J. J. Chem.
Soc., Dalton Trans. 1996, 1809. (b) Hill, G. F.; Yap, G. P. A.; Puddephatt,
R. J. Organometallics 1999, 18, 1408. (c) van Asselt, R.; Rinjberg, E.;
Elsevier, C. J. Organometallics 1994, 13, 706.
(18) (a) Beck, W.; Su¨nkel, K. Chem. ReV. 1988, 88, 1405. (b) Strauss,
S. H. Chem. ReV. 1993, 93, 927.

10834 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
trans-[(PiPr3)2Pt(H)(ClCH2Cl)]+BArf-.19 The degradation of 5
in dichloromethane has not been further investigated.
to the triflate complex 5, undergoes relatively rapid decomposi-
tion, generating methane and unidentified organometallic species
in this solvent. However, the presence of small amounts of TFE
improved the solubility of 6(BF4-) in dichloromethane-d2 and
facilitated an NMR analysis at subambient temperatures. The
1H NMR spectrum in dichloromethane-d2 with 250 equiv (ca.
2.4 M) of TFE at -20 °C displayed a broadened signal at δ
6.51, integrating for 2 H, which we attribute to coordinated
water. This establishes that water is a considerably better ligand
than TFE under these conditions.
Efforts were made to generate 5 in nitromethane, a relatively
poorly coordinating solvent that was hoped to be less reactive
than dichloromethane. However, addition of 1 equiv of HOTf
to a suspension of 1 in nitromethane-d3 led to a complex mixture
of 3-5 compounds containing the (Nf-Nf)Pt(CH3)+ fragment.
None of the major species could be identified as 5, since
essentially the same mixture of compounds was obtained when
HBF4‚Et2O or H(Et2O)2+BArf- was used instead of HOTf. The
aqua complex (Nf-Nf)Pt(CH3)(H2O)+BF4- (6(BF4-), vide infra)
gave similar mixtures in nitromethane-d3 and remains the only
compound in the solution that has been unambiguously identi-
Electrophilic, coordinatively unsaturated metal fragments
commonly coordinate ligands that are usually very weakly
coordinating,18 and we therefore have considered the possibility
that the (Nf-Nf)Pt(CH3)+ moiety in 6 might be associated with
a BF4- ligand, rather than water. However, several observations
argue against this. First, the 19F NMR spectrum of 6(BF4-) in
1
fied. It displayed a broad Pt-OH2 signal at δ 5.8 in the H
NMR spectrum.20 The relative amount of the aqua complex
increased upon addition of small amounts of water to the solu-
tion. Whatever the other species might be, they all contain lig-
ands that are readily displaced, since the addition of acetonitrile
quantitatively and instantly yielded the solvento complex 3.
The great reactivity of 5 toward all solvents in which it is
soluble precluded the purification of the material that was
isolated from dichloromethane. Traces of water appeared to be
a source of the aqua complex in almost any solvent that was
used. We decided to investigate this complex more closely. To
favor dissociation of the aqua ligand, thereby generating the
putative, reactive 14-electron cation (Nf-Nf)Pt(CH3)+, a hy-
droxylic or hydrogen-bonding, yet poorly coordinating, solvent
should be used. In this respect, trifluoroethanol (TFE) is a
solvent that has been widely used in organic chemistry, in
particular in studies of aliphatic nucleophilic substitutions. This
solvent is quite strongly ionizing (ꢀ ) 26.721a) but only poorly
nucleophilic.21b,c These properties should make it an attractive
choice for the exploration of the chemistry of the cationic and
electrophilic Pt complexes.
-
TFE-d3 displayed a signal due to BF4 at the same chemical
shift as Bu4N+BF4-. This would not be expected for a
-
-
coordinated BF4 ligand, whether BF4 is rapidly “spinning”
or not at the metal.18 Second, protonation of 1 with HOTf in
TFE-d3 in the presence of water gave a species formulated as
6(OTf-) that has a H NMR spectrum that is identical to that
1
of 6(BF4-) in TFE-d3.
C-H Activation of Benzene. In TFE solution, the aqua
complex 6(BF4-) underwent facile reaction with the aromatic
C-H bonds of benzene and toluene.22 A quantitative reaction
occurred with 6 in the presence of benzene after 2-3 h at
ambient temperature, leading to the corresponding phenyl
complex (Nf-Nf)Pt(C6H5)(H2O)+ (7) and methane, detected by
1H NMR spectroscopy (eq 5). In comparison, the analogous
Preparation and Isolation of the Pt(II) Aqua Complex
-
(Nf-Nf)Pt(CH3)(H2O)+BF4 (6(BF4-)). The protonation of 1
with HBF4‚Et2O in dichloromethane in the presence of a small
amount of water gave an orange powder after workup. The
spectroscopic and elemental analysis data of this material are
in agreement with the formulation of this species as (Nf-Nf)-
reaction starting with (tmeda)Pt(CH3)(NC5F5)+ required heating
for several days in pentafluoropyridine at 85 °C.9c The product
7(BF4-) was not isolated but was characterized by NMR
1
spectroscopy. A comparison of the H NMR data for 6 and 7
Pt(CH3)(H2O)+BF4 (6(BF4-)). The purity was at least 95%
-
shows that the resonances for the protons on one C6H3(CF3)2
substituent at the diimine have been perturbed more by the
substitution of phenyl for methyl than the other, possibly as a
result of ring current effects caused by close proximity of phenyl
and one C6H3(CF3)2 group. Addition of acetonitrile to the solu-
tion after complete conversion to 7 caused the immediate and
quantitative replacement of the aqua ligand to yield (Nf-Nf)Pt-
(C6H5)(NCMe)+ (4). The spectroscopic data for 4 thus obtained
were identical to those of an authentic sample, prepared as the
by NMR, sometimes pure within detection limits. The trace
impurities may well be due to trace solvent impurities rather
than to inherently slightly impure samples. Attempts at recrys-
tallizing 6(BF4-) from various solvents invariably gave materials
of poorer purity. In the 1H NMR spectrum, the Pt-CH3 signal
appears at δ 0.79 and displays the expected 195Pt satellites with
a coupling constant 2J(195Pt-H) ) 73.3 Hz, close to the value
observed12 for related (diimine)Pt(CH3)(L)+ complexes. The
aqua ligand could not be observed in TFE-d3, presumably due
to facile H/D exchange with the solvent. The material is only
poorly soluble in dichloromethane-d2 and, in a fashion similar
-
BF4 salt as described earlier.
When 6(BF4-) was treated with C6D6 in TFE-d3, the 1H NMR
spectrum revealed multiple incorporation of deuterium in the
methane that was generated. Thus, well-resolved and charac-
teristic signals were found for the CH3D, CH2D2, and CHD3
isotopomers. Similar observations of multiple isotope incorpora-
tion were reported for the analogous reaction between (tmeda)-
Pt(CH3)(NC6F5)+ and C6D6,9 as well as in related reactions with
other metals.23 The multiple H/D exchange into the methane is
most readily explained in terms of the reaction steps shown in
Scheme 1. (A σ-bond metathesis pathway in which the H/D
exchange takes place directly between the η2-benzene and
σ-methane complexes without the intervention of the five-
(19) Butts, M. D.; Scott, B. L.; Kubas, G. J. J. Am. Chem. Soc. 1996,
118, 11831.
(20) Although the nitromethane-d3 was carefully purified and dried prior
to the reaction, traces of water were always observed in the solutions.
(21) (a) Evans, D. F.; McElroy, M. I. J. Soln. Chem. 1975, 5, 405. (b)
Schadt, F. L.; Bentley, T. W.; Schleyer, P. v. R. J. Am. Chem. Soc. 1976,
98, 7667. (c) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in
Organic Chemistry, 3rd ed.; Harper and Row: New York, 1987; pp 335-
340.
(22) Details about the reactions with aromatic substrates will be reported
elsewhere. See also ref 24.

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10835
Scheme 1
doublet (1J(13C-H) ) 129 Hz) centered at the same chemical
shift. This doublet, as expected, was also accompanied by 195Pt
satellites. In the 13C{1H} NMR spectrum, the Pt-CH3 resonance
at δ -10.2, as well as its 195Pt satellites (1J(195Pt-CH3) ) 744
1
Hz), grew in intensity. After a reaction time of 45 h, H NMR
integration of doublet vs singlet intensity indicated that the extent
of 13CH3 incorporation was 43% of the total Pt-CH3 present
as 6. The reaction was accompanied by slow decomposition of
6, evidenced by the precipitation of a red material that remains
unidentified. This decomposition occurred to the same extent
(ca. 20%) in the absence of methane. Decomposition might
therefore be related to a reaction preceding the C-H activation,
rather than to the activation process itself. There was no visible
formation of platinum black during the reaction.
The C-H activation at 6 is likely to occur by C-H addition
at a formally three-coordinate T-shaped intermediate, (Nf-Nf)-
Pt(CH3)+, that would result from water dissociation from 6.
Support for this hypothesis was obtained when the reaction of
6(BF4-) with 13CH4 in TFE-d3 was performed in the presence
of water. With ca. 11 equiv of water added to the reaction
medium (reaction conditions were otherwise identical to those
used above), it was found that only 24% of 13CH3 had been
incorporated into 6 after 48 h. Thus, the reaction rate was only
half of that found with no water added. This is in agreement
with a preequilibrium dissociative mechanism to give a reactive
three-coordinate intermediate. However, equally consistent with
the observed inhibition is a preequilibrium associative substitu-
tion of TFE for H2O followed by associative replacement of
TFE by CH4.26 Such a mechanism appears to be involved in
the benzene activation by a related complex.24 The decomposi-
tion of 6 was also in part inhibited by the presence of water,
amounting to ca. 5% with 11 equiv of water. Therefore, the
decomposition may also be a consequence of the initial
formation the three-coordinate species or other intermediate(s)
en route to the methane σ-complex.
coordinate oxidative addition product might also be devised;
aspects of this will be addressed for methane activation in the
theoretical part of this paper.) A dynamic equilibrium exists
between (Nf-Nf)PtII(CH3)(η2-benzene)+, (Nf-Nf)PtIV(H)(CH3)-
(C6H5)+, and (Nf-Nf)PtII(σ-CH4)(C6H5)+ intermediates, all of
which are accessible after exchange of benzene for water in 6.
The η2-benzene complex, which in principle may be η2(C,H)
or η2(C,C) coordinated,24 may or may not be an intermediate
in the oxidative addition of benzene, but must be involved in
the exchange process. Provided that “rotation” of the η2-C6H6
and σ-CH4 ligands at the respective Pt centers, as well as
insertion of Pt into the C-H bonds and the reverse process,
are rapid relative to the rate of methane and benzene dissocia-
tion, the multiple D incorporation is nicely explained.25
C-H Activation of Methane. Encouraged by the above
evidence for a methane σ-complex intermediate in the benzene
C-H activation, we investigated the possibility that 6 might
also activate methane C-H bonds. Indeed, it was found that
methane C-H activation occurred under unusually mild condi-
tions. Thus, a solution of 6(BF4-) in TFE-d3 was sealed under
20-25 bar of 13CH4 in an NMR tube. By 1H NMR integration,
ca. 20 equiv of 13CH4 was present in solution. Unambiguous
evidence for methane C-H activation was found during a 2-day
period at 45 °C (eq 6). The exchange of 13CH3 for the unlabeled
Finally, the reaction between 6 and methane was probed using
CD4 instead of CH4. Under conditions identical to those used
above, a gradual incorporation of deuterium into the Pt-methyl
group (giving rise to Pt-CH3-nDn resonances) was observed
1
in the H NMR spectrum. At the same time, signals due to
methane, initially absent, appeared. The spectrum showed that
the H-containing methane was present as CH3D, CH2D2, and
CHD3 isotopomers. Small amounts of CH4 were also observed,
presumably due to decomposition of 6 (vide supra). The methane
isotopomer mixture was observed even at early stages of the
reaction. Since a large excess of CD4 was used, this must mean
that multiple H/D exchange must take place for each C-D
actiVation eVent that takes place. In analogy with the observa-
tions discussed for the benzene C-H activation earlier, this is
most readily explained by the presence of methane σ-complexes
as intermediates in the exchange process. The H/D exchange
might occur via the oxidative addition pathway (Scheme 2) or
by direct H transfer between σ-bonded methane and the methyl
group without a discrete oxidative addition product. This will
be discussed in the theoretical part that follows. It is interesting
to note that multiple H/D exchange into alkanes was also
observed by the Shilov group.4
1
CH3 ligand in 6 was readily seen by H and 13C{1H} NMR
spectroscopy. The 1H NMR Pt-CH3 singlet at δ 0.79, ac-
companied by 195Pt satellites, was gradually replaced by a
(23) See, for example: (a) Mobley, T. A.; Schade, C.; Bergman, R. G.
J. Am. Chem. Soc. 1995, 117, 7822. (b) Gould, G. L.; Heinekey, D. M. J.
Am. Chem. Soc. 1989, 111, 5502. (c) Wang, C.; Ziller, J. W.; Flood, T. C.
J. Am. Chem. Soc. 1995, 117, 1647. (d) Bullock, R. M.; Headford, C. E.
L.; Hennessy, K. M.; Kegley, S. E.; Norton, J. R. J. Am. Chem. Soc. 1989,
111, 3897. (e) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am.
Chem. Soc. 1986, 108, 1537. (f) Parkin, G.; Bercaw, J. E. Organometallics
1989, 8, 1172. (g) Wick, D. D.; Reynolds, K. A.; Jones, W. D. J. Am.
Chem. Soc. 1999, 121, 3974.
(24) Evidence for an η2(C,C)-coordinated Pt(II) benzene complex in a
related system is described elsewhere. Johansson, L.; Tilset, M.; Labinger,
J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2000, 122, 10845-10854.
(25) As pointed out by a reviewer, the experiment does not demonstrate
that all D incorporation into the methyl group comes from the same C6D6
molecule, but might arise from multiple oxidative addition/reductive
elimination events of different C6D6 molecules without the involvement of
a discrete π-benzene complex. This would imply that benzene loss from
the (Nf-Nf)Pt(H)(CH3)(C6H5)+ system is much more facile than methane
loss. This is contrary to observations in related systems (ref 24).
Protonolysis of (Nf-Nf)Pt(CH3)2 (1) at Low Temperature.
The multiple deuterium incorporation into the methane in the
above reaction between 6 and CD4 implies that the H/D
(26) Associative mechanisms have been reported for solvent-exchange
reactions at (L-L)Pt(CH3)(DMSO)+ species (L-L ) various chelating
diimines and diamines). Romeo, R.; Scolaro, L. M.; Nastasi, N.; Arena, G.
Inorg Chem. 1996, 35, 5087.

10836 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
Scheme 2
metal centers was first discovered with Bergman’s Cp*Ir(PMe3)-
(CH3)+ system, it was immediately realized that an ambiguity
existed regarding the mechanism. Although d electrons, as well
as the Ir(V) oxidation state (somewhat unusual but far from
unprecedented27 in organometallic chemistry), needed for an
oxidative addition pathway were available, it was felt that the
cationic 16-electron Ir(III) complex might be sufficiently
electrophilic for a σ-bond metathesis pathway to be viable. Since
the discovery of this C-H activation process, a number of
theoretical contributions have addressed this problem for C-H
activation at cationic, late metal centers.
DFT (B3LYP/LANL2DZ)28a and ab initio (QCISD/MP2/
LANL2DZ)28b calculations for methane activation using CpIr-
(PH3)(CH3)+ as a model strongly favored the oxidative addition
pathway. A four-center adduct or transition state for a σ-bond
metathesis pathway could not be located, and strong doubt was
expressed about the feasibility of this mechanism. The two
investigations gave very similar results. A methane σ-complex
was located 4 and 29 kJ/mol below the energies of the reactants,
respectively, and a 48 vs 54 kJ/mol barrier separated this
complex from the Ir(V) oxidative addition intermediate, which
was located 14 vs 4 kJ/mol higher in energy than the
reactants.28a,b
C-H activation at cationic Pt centers has also received
considerable attention from a theoretical point of view. Hill and
Puddephatt29a used extended Hu¨ckel MO and DFT calculations
on the reductive elimination of methane from the five-coordinate
model complexes L2Pt(H)(CH3)2+, yielding the methane σ-com-
plexes L2Pt(CH3)(CH4)+ (L ) NH3 or PH3). The possibility that
a Pt-CH3/CH4 exchange might occur by a σ-bond metathesis
pathway was not addressed. Siegbahn and Crabtree29b modeled
the solvent effect on the Shilov reaction of PtCl2 in water; a
σ-bond metathesis-like activation of methane was favored in
which a coordinated Cl ligand accepts a proton from a
coordinated methane molecule. However, an oxidative addition
pathway was not ruled out.
Scheme 3
scrambling process must be rapid relative to the rate of methane
elimination from the methane σ-complex. This suggests that if
it is possible to generate any of the intermediates involved in
this scrambling process at low temperature, they might have a
sufficient stability and lifetime to be observable by NMR.
Obviously, it would be desirable to perform such an experiment
in the actual TFE solvent system used in the C-H activation
reactions. However, due to the high freezing point of TFE (-45
°C) and poor solubility of 1 in this solvent, we decided to
perform the experiment in a dichloromethane/ether mixture. An
NMR tube containing 1 dissolved in dichloromethane-d2 was
cooled to -78 °C, whereupon 2 equiv of HBF4‚Et2O dissolved
in a small amount of ether and dichloromethane-d2 was added.
A color change from dark purple to pale yellow was immediately
1
observed. H NMR analysis of the yellow solution at -70 °C
revealed clean formation of two Pt(IV) hydride species in a ca.
2:1 ratio. The characteristic hydride signals appeared far up-
field (δ -26.67 and -26.40) with corresponding Pt satellites
1J(195Pt-H) of 1741 and 1699 Hz. Furthermore, two methyl
signals were found at δ 0.54 and 0.48 with 2J(195Pt-H) ) 63.6
and 62.9 Hz, respectively. These data suggest that both
compounds can be formulated as (Nf-Nf)PtIV(CH3)2(H)(L)+
species (8a and 8b), with the hydride occupying one apical
position. We presume that the two species differ in having
different stabilizing two-electron donors L coordinated to Pt
trans to the hydride. The identity of these could not be
determined from the NMR spectra. Above ca. -40 °C, elimina-
tion of methane from the species 8a and 8b with concurrent
formation of unidentified organometallic products was observed.
No other intermediates could be detected prior to methane loss.
If we assume that this result applies to the current TFE system
as well, it suggests that a Pt(IV) hydride species is the
energetically most stable of the intermediates involved in the
H/D scrambling process discussed above.
II. Theoretical Work. Mechanistic Considerations. Two
mechanisms are commonly considered when C-H bond activa-
tion reactions at the transition metal are discussed (Scheme 3).
The oxidative addition pathway is observed for C-H activation
at relatively low-oxidation-state, late-transition-metal centers.
The other alternative is the σ-bond metathesis. By this mech-
anism, C-H activation occurs in a concerted fashion via a four-
center, four-electron transition state. This mechanism is observed
for early, typically d0 metals that have no d electrons available
for the oxidative addition. When C-H activation at late, cationic
Hush and co-workers29c investigated possible mechanisms for
the catalytic C-H activation and functionalization by the Pt-
(bpym) system of Periana et al.11 The cis-diammine system was
used as a model, and the solvent (concentrated H2SO4) was
included by a dielectric continuum method and explicit inclusion
-
of solvento species (coordinated HSO4 and H2SO4) in the
calculations. It was concluded, on thermochemical grounds
exclusively, that the crucial C-H activation step is likely to
proceed by an electrophilic mechanism rather than by an
oxidative addition pathway. However, the validity of these
results was somewhat questionable because (a) no transition
states and corresponding energies were calculated for either
mechanism, and (b) for the hydridoalkyl Pt(IV) intermediates
of the oxidative addition pathway, only isomers with a trans
(diaxial) relationship between the oxidatively added hydride and
(27) Alaimo, P. J.; Bergman, R. G. Organometallics 1999, 18, 2707 and
references cited.
(28) (a) Strout, D. L.; Zaric, S.; Niu, S.; Hall, M. B. J. Am. Chem. Soc.
1996, 118, 6068. (b) Su, M.-D.; Chu, S.-Y. J. Am. Chem. Soc. 1997, 119,
5373. (c) Hinderling, C.; Feichtinger, D.; Plattner, D. A.; Chen, P. J. Am.
Chem. Soc. 1997, 119, 10793. (d) Niu, S.; Hall, M. B. J. Am. Chem. Soc.
1998, 120, 6169. (e) Niu, S.; Zaric, S.; Bayse, C. A.; Strout, D. L.; Hall,
M. B. Organometallics 1998, 17, 5139. (f) Niu, S.; Hall, M. B. J. Am.
Chem. Soc. 1999, 121, 3992.
(29) (a) Hill, G. S.; Puddephatt, R. J. Organometallics 1998, 17, 1478.
(b) Siegbahn, P. E. M.; Crabtree, R. H. J. Am. Chem. Soc. 1996, 118, 4442.
(c) Mylvaganam, K.; Bacskay, G. B.; Hush, N. S. J. Am. Chem. Soc. 1999,
121, 4633. (d) Mylvaganam, K.; Bacskay, G. B.; Hush, N. S. J. Am. Chem.
Soc. 2000, 122, 2041. (e) Bartlett, K. L.; Goldberg, K. I.; Borden, W. T. J.
Am. Chem. Soc. 2000, 122, 1456. (f) Heiberg, H.; Swang, O.; Ryan, O. B.;
Gropen, O. J. Phys. Chem. A 1999, 103, 10004.

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10837
Table 2. Calculated Energies for Pt Structures30
methyl groups were considered. The cis (axial, equatorial)
analogues would be more relevant in particular if a concerted
oxidative addition occurs. This was corroborated by a recent
extended and revised study by the Hush group29d and will also
be evident from our results.
Recently, Bartlett, Goldberg, and Borden29e presented a
detailed theoretical investigation of reductive elimination of
alkanes from Pt(II) and Pt(IV) hydridoalkyl complexes. The
cationic five-coordinate complex cis-(PH3)2Pt(Cl)(CH3)(H)+ was
found to undergo exothermic elimination with a very low
activation barrier of 4 kJ/mol (B3LYP) to yield a σ-complex in
which methane is bonded to the Pt center with a bond energy
of 40 kJ/mol.
Some of us recently reported29f DFT calculations on Pt-CH3/
CH4 exchange at (H2NCH2CH2NH2)Pt(CH3)+, a model for
Bercaw and Labinger’s (tmeda)Pt(CH3)+ system. An oxidative
addition pathway was strongly favored in this case. To shed
light also on the C-H activation at the (Nf-Nf)Pt(CH3)+ system,
we report here an investigation of the reaction mechanism with
quantum chemical methods based on density functional theory
(DFT).
Selection of a Suitable Model System for Calculations. The
structure of the Pt complex (p-MeOC6H4sNdCHsCHdNs
C6H4sp-OMe)Pt(CH3)2 was calculated with DFT methods30
(see the Experimental Section for computational details). The
p-MeOC6H4 substituent was employed for the full-scale system
because the X-ray crystal structure of this Pt complex has been
reported,12 and so its geometry can be compared with the
calculated geometry. Gratifyingly, the X-ray crystallographic
data and the computationally optimized geometry were found
to be in excellent agreement. The differences between observed
and calculated geometries were typically less than 0.01 Å for
bond distances and less than 2° for bond angles (maximum
deviations were 0.02 Å and 4°).
The Pt complexes that were actually used in the experiments
are too large for the desired quantum chemical study of the
entire C-H bond activation mechanism. The process was
therefore modeled with a smaller diimine ligand at the Pt center.
The molecular model that was used (the simplest possibles
with only H substituents on the diimine!) was chosen after
testing six models with different diimine ligands, XsNdC(Y)s
C(Y)dNsX, where X ) CHdCH2, CH3, or H, and Y ) CH3
or H. The reaction energies for several relevant reactions were
calculated with DFT methods for the full-sized Pt complex (X
) p-MeOC6H4)30 and used to test the relevance of the six
models. The reaction energies that were obtained with models
with X or Y * H were not significantly better than those
obtained with X and Y ) H when compared with the
p-MeOC6H4 system, and the smallest molecular model, with X
and Y ) H, was therefore chosen for the theoretical investigation
of the reaction mechanism. The substitution of hydrogen for
an aryl substituent at the diimine is a rather dramatic simplifica-
tion. However, it should be pointed out that the aryl rings are
not coplanar with the Pt-diimine plane, and therefore π-overlap
between the metal-diimine ring and the aryl rings should be
of reduced importance. The optimized geometry of the simple
model agreed well with the pertinent bond lengths and angles
in the X-ray crystal structure for (p-MeOC6H4sNdCHsCHd
NsC6H4sp-OMe)Pt(CH3)2, as well as with the calculated
geometry for this full-size complex. The differences in bond
distances and angles between the simple model and the X-ray
relative energy (kJ/mol)
complex
without ZPE
with ZPE
92a
(A) (N-N)Pt(CH3)+
99
-59
-49
0
29
30
(B) (N-N)Pt(CH3)(H2O)+
(C) (N-N)Pt(CH3)(TFE)+
(D) (N-N)Pt(CH3)(CH4)+
0a
23a
(E) ax-(N-N)Pt(CH3)2(H)+
(F) eq-(N-N)Pt(CH3)2(H)+
29a
(G) ax-(N-N)Pt(CH3)2(H)(H2O)+
(H) ax-(N-N)Pt(CH3)2(H)(TFE)+
(I) (N-N)Pt(CH3)(CH4)+ OA#
(J) (N-N)Pt(CH3)(CH4)+ SB#
(K) (N-N)Pt(CH3)(CH4)(H2O)+
(L) (N-N)Pt(CH3)(CH4)(TFE)+
(M) (N-N)Pt(CH3)(CH4)(H2O)+ OA#
(N) (N-N)Pt(CH3)(CH4)(TFE)+ OA#
-58
-49
40
-64b
-55b
33a
50
44a
-30
-27
4
-4b
9
2b
a ZPE from calculations on the full model complex. b Estimated by
assuming the same ZPE corrections to solvent-associated species as to
the corresponding species without solvents.
crystallographic data are typically 0-0.01 Å and 2°. The
maximum deviation of 0.05 Å was found in one of the Pt-N
bonds, while the other Pt-N bond distance differed by only
0.02 Å. This is caused by the model having a C2V symmetry; a
difference of 0.03 Å was found in the Pt-N bond lengths in
the crystal structure data. These small differences in Pt-N bond
lengths may be due to crystal packing effects. With these
considerations, the following calculations were done on the
(HNdCHsCHdNH)Pt system, henceforth to be abbreviated
as (N-N)Pt.
Calculations of Relevant Ground-State Species with DFT
Methods. Calculations on relevant ground-state species were
first performed, largely using DFT methods. The search for and
calculation of transition-state structures for the oxidative addition
and σ-bond metathesis mechanisms were then undertaken. Table
2 summarizes calculated energies,31 and Table 3 summarizes
key geometrical features for the ground-state complexes as well
as transition states (vide infra). An overview of the calculated
structures is depicted in Figure 1. In the following, energies
include zero-point energy (ZPE) corrections unless otherwise
stated. ZPE corrections given for systems containing solvent
molecules are estimated by applying the corresponding ZPE
corrections of the solvent-free species.30
The Three- and Four-Coordinate Cations (N-N)Pt(CH3)-
(L)+ (L ) Vacant Site, TFE, H2O, or CH4). The cation (N-N)-
Pt(CH3)+ (A) is a model for the three-coordinate species that
might be responsible for the C-H activation process. The C-H
activation of methane takes place in a TFE solution containing
a small amount of water. Cationic square planar Pt(II) complexes
(Nf-Nf)Pt(CH3)(L)+ (L ) H2O, TFE) are likely present. The
NMR results indicate that water occupies the empty site of
(Nf-Nf)Pt(CH3)+ most of the time; we have no spectroscopic
evidence for the existence of the corresponding TFE adduct.
The calculations show that the bond strengths between the
(N-N)Pt(CH3)+ moiety and L are 158 and 149 kJ/mol for L )
H2O (B) and TFE (C), respectively. Steric and electronic factors
may both contribute to the weaker bond to TFE.
The Pt(II)-methane σ-complex (Nf-Nf)Pt(CH3)(σ-CH4)+ is
a likely intermediate in the C-H activation process.9 The
intermediacy of such a species provides a straightforward
explanation for the occurrence of multiple H/D exchange into
(30) Calculations of vibrational frequencies for the optimized structures
of the full-size molecules and model complexes that contain solvent
molecules have not been performed, for obvious reasons of computational
economy. Zero-point energy corrections are therefore not available.
(31) All energies are given for assemblies containing one each of (N-N)-
Pt(CH3)+, H2O, CH4, and TFE in order to be directly comparable. For
example, the energy of (N-N)Pt(CH3)(H2O)+ implicitly includes the energies
of one molecule each of TFE and methane separated at infinite distance.

10838 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
Figure 1. Calculated structures of relevance to the methane C-H activation reaction. See footnote a to Table 3 for atom numbering system.
Table 3. Calculated Bond Distances (Å) and Angles (Deg) for Pt Structuresa
complex
Pt-H(1) H(1)-C(1) Pt-C(1) Pt-C(2) Pt-N(1) Pt-N(2) Pt-O H(1)-Pt-C(1) Pt-C(1)-H(1)
(A) (N-N)Pt(CH3)+
2.00
2.04
2.04
2.04
2.04
2.03
2.05
2.05
2.04
2.14
2.04
2.04
2.05
2.05
1.93
1.95
1.94
1.96
2.19
2.22
2.17
2.18
2.13
2.08
1.97
1.96
2.07
2.08
2.16
2.15
2.16
2.16
2.19
2.19
2.19
2.18
2.17
2.08
2.16
2.16
2.16
2.16
(B) (N-N)Pt(CH3)(H2O)+
2.13
2.12
(C) (N-N)Pt(CH3)(TFE)+
(D) (N-N)Pt(CH3)(CH4)+
1.84
1.51
1.53
1.51
1.51
1.54
1.59
1.82
1.83
1.56
1.55
1.16
2.42
2.43
2.43
2.41
1.73
1.56
1.16
1.16
1.53
1.56
2.34
2.04
2.07
2.05
2.05
2.11
2.14
2.36
2.35
2.14
2.13
29
84
84
85
84
54
47
29
29
46
47
51
38
39
38
38
46
48
49
49
47
47
(E) ax-(N-N)Pt(CH3)2(H)+
(F) eq-(N-N)Pt(CH3)2(H)+
(G) ax-(N-N)Pt(CH3)2(H)(H2O)+
(H) ax-(N-N)Pt(CH3)2(H)(TFE)+
(I) (N-N)Pt(CH3)(CH4)+ OA#
(J) (N-N)Pt(CH3)(CH4)+ SB#
(K) (N-N)Pt(CH3)(CH4)(H2O)+
(L) (N-N)Pt(CH3)(CH4)(TFE)+
(M) (N-N)Pt(CH3)(CH4)(H2O)+ OA#
(N) (N-N)Pt(CH3)(CH4)(TFE)+ OA#
2.34
2.36
4.50
4.03
2.84
2.90
a The following numbering system has been used (see structure, below): H(1), hydride or bridging H in σ-CH4 complexes; C(1), carbon of
σ-CH4 ligand or of methyl group derived from this ligand; C(2), carbon of methyl ligand; N(1), diimine cis to original methyl; N(2), diimine N
trans to original methyl.
CD4. Alkane σ-complexes have rarely been directly ob-
existence.34 Calculations on (N-N)Pt(CH3)(CH4)+ (D) were done
with and without ZPE corrections. The non-ZPE results show
served,32,33 but a large body of evidence strongly implies their
(32) For a recent report of an X-ray structure of a (porphyrin)FeII(σ-
alkane) complex, see: Evans, D. R.; Drovetskaya, T.; Bau, R.; Reed, C.
A.; Boyd, P. D. W. J. Am. Chem. Soc. 1997, 119, 3633.
(33) For a recent account of the direct observation of a Re(I) cyclopentane
complex, see: Geftakis, S.; Ball, G. E. J. Am. Chem. Soc. 1998, 120, 9953;
1999, 121, 6336 (addendum).

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10839
Scheme 4
Figure 2. Coordination of the CH4 ligand at Pt in (N-N)Pt(CH3)(CH4)+
(D).
the trans influence of the groups, whereas the cis Pt-N distance
changes by only 0.01 Å. Methane, with a 99 kJ/mol bond to
the (N-N)Pt(CH3)+ fragment, is a relatively poor ligand
compared to H2O (158 kJ/mol) and TFE (149 kJ/mol) according
to the DFT calculations. It may therefore appear surprising that
methane C-H activation occurs at all from the aqua complex.
However, it must be borne in mind that in the real system in
solution, the position of the ligand-exchange equilibria shown
in Scheme 4 will depend not only on the relative ligand binding
energies to Pt, but also on the solvation energies of the species
involved. In particular, the relative solvation properties of H2O,
TFE, and CH4 must play a major role. The small, strongly
hydrogen-bonding molecules H2O and TFE have significantly
greater solvation energies in TFE than does CH4, and this factor
will counteract the trend of the Pt-ligand bond energies on the
equilibria. A relatively small solvation energy for CH4 in TFE
may contribute significantly to increase the driving force, and
to lower the activation barriers, for the substitution of CH4 for
H2O and TFE at the Pt center, regardless of whether the
exchange processes occur by associative or dissociative mech-
anisms.
Five- and Six-Coordinate Oxidative Addition Products
(N-N)Pt(CH3)2(H)(L)+ (L ) Vacant Site, H2O, or TFE).
Oxidative addition of methane within the σ-complex (N-N)Pt-
(CH3)(σ-CH4)+ may, in principle, occur to give the five-
coordinate oxidative addition product (N-N)Pt(CH3)2(H)+ with
the hydride located either in an apical position (E) or in an
equatorial position (F). If the latter is favored, then the principle
of microscopic reversibility dictates that there must be an
available pathway for the interchange of the two nonequivalent
methyl groups for the Pt-CH3/CH4 exchange process. The DFT
calculations show that the isomer with an apical hydride is the
most stable of the two at an energy +23 kJ/mol relative to the
σ-complex, compared to the isomer with the equatorial hydride
at +29 kJ/mol. The same relative ordering of hydride-apical
and hydride-equatorial energies was calculated (B3LYP) for cis-
(PH3)2Pt(Cl)(CH3)2(H)+, but with a smaller energy difference
(+98 and +100 kJ/mol relative to the σ-complex, respectively).29e
The hydride-apical isomer of (H2NCH2CH2NH2)Pt(CH3)2(H)+
was located 7 kJ/mol higher in energy than the σ-complex
including ZPE,29f and that of cis-(NH3)2Pt(CH3)2(H)+ was 38
kJ/mol higher, ZPE apparently not included.29a From these
results, the reaction energies for the oxidative addition reactions
appear to be strongly dependent on the ancillary ligands of the
complexes (although differences in computational methods and
basis sets may account for parts of the differences). The reactions
to yield the five-coordinate products are always uphill in energy
from the σ-complexes.
that methane binds in what has been termed an η3(H,C,H)29a or
η2(H,C)35 mode (the latter will be adopted by us) to Pt in the
σ-complex, with a non-ZPE-corrected bond energy of 99 kJ/
mol. The methane in this complex, shown in a side-on view in
Figure 2, is coordinated via two C-H σ-bonds with the
H-C-H plane perpendicular to the equatorial plane at Pt and
has one elongated C-H bond (1.16 Å, compared to 1.10 Å in
free methane). The two Pt-H distances are 1.84 and 2.22 Å,
respectively, and the H-C-H angle has opened up to 120°.
The calculated vibrational frequency for the activated C-H bond
is reduced by 592 cm-1 relative to the asymmetric C-H
vibration in free methane. The Pt-N bond trans to CH3 (2.16
Å) is longer than the one trans to CH4 (1.96 Å), reflecting the
greater trans influence of the methyl ligand. The geometry and
trends in bond distances closely resemble those calculated for
(H2NCH2CH2NH2)Pt(CH3)(CH4)+ 29f and cis-(NH3)2Pt(CH3)-
(CH4)+,29a with slight differences in the metric parameters.
The CH4 ligand in D appears to undergo a facile “rotation”
within the σ-complex via a symmetrical η2(H,H) structure to
form a reoriented η2(H,C)-bonded ligand. The calculations
indicate that the symmetrical structure is a transition state
separating two η2(H,C) minima with a barrier of only 0.6 kJ/
mol. However, when the ZPE is taken into account, the situation
is reversed and the symmetrical η2(H,H) structure is 0.6 kJ/
mol lower in energy than the η2(H,C) structure, leading to one
shallow energy minimum rather than two minima separated by
a small barrier. Nevertheless, it can be safely assumed that the
rotation occurs essentially unhindered at the temperatures in
question. This agrees with the observed multiple deuterium
incorporation in the exchange experiments.
The η2(H,C) and η2(H,H) structures were further investigated
by calculations of the energies of the DFT-optimized structures
with various high-level ab initio methods using the Gaussian94
program36 (vibrational frequencies were not calculated). The η2-
(H,C) structure was found to be most stable for all the methods
used. In summary, the energy difference between η2(H,C) and
η2(H,H) was calculated to be 0.6 kJ/mol (DFT) versus 1.2
(CCSD(T)), 1.6 (CCSD), 1.5 (MP4(SDQ)), 1.2 (MP2), and 2.7
(HF) kJ/mol. Thus, the η2(H,C) bonding mode is found to be
the most stable, excluding ZPE corrections, by a small margin
in all cases.
The Pt-N bond distances trans to the ligand (or vacant site)
in the cations (N-N)Pt(CH3)(L)+ (A-D) decrease by ca. 0.04
Å in the order CH4 > H2O ≈ TFE > vacant site according to
(34) (a) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789. (b)
Hall, C.; Perutz, R. N. Chem. ReV. 1996, 96, 3125.
(35) (a) Hall, C.; Perutz, R. N. Chem. ReV. 1996, 96, 3125. (b) Billups,
W. E.; Chang, S.-C.; Hauge, R. H.; Margrave, J. L. J. Am. Chem. Soc.
1993, 115, 2039.
The five-coordinate C-H oxidative addition products have
never been directly observed, presumably because of the
unfavorable thermodynamics of their formation relative to that
of the precursors as well as a low barrier against reductive
elimination. However, in one instance, a five-coordinate C-H
oxidative addition product has been trapped by rapid intramo-
lecular attachment of a pendant donor ligand. Wick and
Goldberg10 reported that treatment of K+[(η2-Tp′)Pt(CH3)2]-
with B(C6F5)3 in the presence of hydrocarbons R-H gave the
stable octahedral complexes (η3-Tp′)Pt(CH3)(R)(H), presumably
(36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Revision D.4;
Gaussian, Inc.: Pittsburgh, PA, 1995.

10840 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
via C-H oxidative addition at a three-coordinate complex (η2-
Tp′)Pt(CH3), followed by rapid intramolecular trapping by the
pendant arm of the Tp′ ligand. Stable Pt(IV) hydridodialkyl
complexes have been generated in other instances by protonation
of four-coordinate dialkyl precursors. The hydridoalkyl com-
plexes can be observed and sometimes even isolated if the sixth
coordination site is occupied by a ligand that does not undergo
facile dissociation, as found in complexes 8a and 8b in this
work and previously reported by others.9a,37 In those instances
where bis(chelating) supporting ligands are present, the hydride
ligand appears to always occupy an apical position relative to
the chelate.9a,37a,c,e,f,i
calculations showed that this reaction would occur via an initial
interchange to the other isomer, E. This interchange occurs by
a concerted motion of the apical methyl group and the equatorial
hydride perpendicular to the equatorial plane in the five-coor-
dinate species and has a calculated activation energy of 9 kJ/
mol. The process is reminiscent of the one responsible for the
interchange of basal and apical methyl groups in Ru(PtBu2Me)2-
(CO)Me2.38 The reductive elimination from E is now, of course,
the microscopic reverse of the oxidative addition starting from
D. These results somewhat contrast the DFT (B3LYP) results
of Bartlett et al.29e on the hydride-apical and hydride-equatorial
isomers of cis-(PH3)2Pt(Cl)(CH3)2(H)+, both of which underwent
elimination to yield the corresponding σ-complex through a
common “pinched trigonal bipyramidal” transition state.
Reductive elimination of ethane from the Pt(IV) complex E
was not investigated. Ethane formation is not experimentally
observed, and such a process would probably have a much
higher activation barrier than reductive elimination of meth-
ane.29a,39
Computational Search for a σ-Bond Metathesis Mecha-
nism for the Methane C-H Activation. Starting from the
σ-complex D, the distance between the most strongly interacting
Pt-bonded hydrogen atom in methane and the Pt-CH3 carbon
atom was used as the reaction coordinate in the investigation
of the σ-bond metathesis reaction. The use of this reaction
coordinate offers the advantage that this C-H distance hardly
changes in the oxidative addition reaction pathway, and hence
is approximately orthogonal to the latter path. When this C-H
distance was decreased, the reaction trajectory of an apparent
σ-bond metathesis was obtained. However, in an attempt to trace
the same reaction coordinate in reverse by increasing the C-H
distance, an oxidative addition trajectory resulted. This hysteresis
is likely an artifact40 caused by the choice of a reaction
coordinate for the methane elimination that diverges from the
lowest-energy path of the reaction, and may be interpreted as
an indication that the σ-bond metathesis mechanism is unlikely.
Hoping to investigate the σ-bond metathesis using a different
reaction coordinate, we considered that the migrating hydrogen
atom may be situated equidistant from the two C atoms when
the transition state is reached because of the natural molecular
symmetry of this state. Another possibility is that the hydrogen
atom is located within the equatorial plane (this is confirmed
by the calculations, vide infra). These geometry restrictions were
achieved by imposing either molecular Cs symmetry by a mirror
plane perpendicular to the equatorial plane, or C2 symmetry by
an axis through the midpoint of the diimine C-C bond and the
Pt atom. Minimizing the energy with either of these symmetry
restrictions gave a transition state J which has C2V symmetry.
The transition state has a single imaginary vibration frequency
for the transfer of the hydrogen atom between the methyl groups.
The calculations give an activation barrier of 44 kJ/mol for the
σ-bond metathesis mechanism, to be compared with the previ-
ously found 33 kJ/mol for oxidative addition. The results are
therefore slightly in favor of an oxidative addition pathway over
σ-bond metathesis, but the difference of 11 kJ/mol is not large.
It is clear that coordination of solvent molecules as well as bulk
solvent effects could easily perturb the system sufficiently to
even invert the trend. We therefore decided to explicitly include
a solvent molecule in the calculations.
The DFT calculations of (N-N)Pt(CH3)2(H)+ (E; hydride
apical) were supplemented by calculations for the six-coordinate
complexes resulting from addition of a H2O (G) or TFE (H)
ligand at the second apical coordination site. Coordination of
the two ligands to give hexacoordinate species resulted in a
stabilization of the pentacoordinate oxidative addition products
by 87 and 78 kJ/mol, respectively, ZPE not included. The
H-Pt-CH3 angles in E, G, and H remain rather invariant at
85°. This suggests that the tilt of the hydride toward the methyl
group is not indicative of a significant residual H3C-H bonding
interaction. The NMR investigation of the (Nf-Nf)Pt(CH3)-
(H2O)+ complex indicates that H2O binds considerably better
to the cationic Pt(II) center than does TFE. Our DFT calcula-
tions also established a stronger binding of L ) H2O relative
to TFE in the Pt(II) complex (N-N)Pt(CH3)(L)+ as well as in
the Pt(IV) oxidative addition products (N-N)Pt(CH3)2(H)(L)+.
The stabilization caused by H2O relative to TFE was somewhat
greater for the five-coordinate oxidative addition product than
for the three-coordinate precursor.
Computational Search for an Oxidative Addition Mech-
anism for the Methane C-H Activation. The oxidative
addition was investigated with the σ-complex D as the starting
point. With the elongated C-H bond as the reaction coordinate,
the calculations ultimately led to the Pt(IV) complex (N-N)Pt-
(CH3)2(H)+ with the hydride in the apical position (E). The
reaction and activation energies were found to be 23 and 33
kJ/mol, respectively. A complete exchange involves a reductive
elimination following the reverse path compared to oxidative
addition. In the transition state I, the Pt-H distance has
decreased to 1.54 from 1.84 Å in D, the Pt-C distance decreased
to 2.11 from 2.34 Å, and the C-H distance increased to 1.73
from 1.16 Å.
Interestingly, we were unable to locate a transition state that
directly connects the σ-complex D and the Pt(IV) complex
(N-N)Pt(CH3)2(H)+, in which the hydride ligand occupies the
equatorial position (F), even though E and F are essentially
equally stable. We cannot safely rule out that the failure to find
a reaction path that directly connects σ-complex D and the
hydride-equatorial Pt(IV) complex F stems from the chosen
reaction coordinate having an insufficient overlap with the
lowest-energy path. The reaction was therefore investigated in
the opposite direction, i.e., as the reductive elimination from F
with the developing C-H bond as the reaction coordinate. The
(37) (a) Hill, G. S.; Rendina, L. M.; Puddephatt, R. J. Organometallics
1995, 14, 4966. (b) O’Reilly, S. A.; White, P. S.; Templeton, J. L. J. Am.
Chem. Soc. 1996, 118, 5684. (c) Hill, G. S.; Puddephatt, R. J. J. Am. Chem.
Soc. 1996, 118, 8745. (d) Canty, A. J.; Dedieu, A.; Jin, H.; Milet, A.;
Richmond, M. K. Organometallics 1996, 15, 2845. (e) Hill, G. S.; Vittal,
J. J.; Puddephatt, R. J. Organometallics 1997, 16, 1209. (f) Jenkins, H. A.;
Yap, G. P. A.; Puddephatt, R. J. Organometallics 1997, 16, 1946. (g) Canty,
A. J.; Fritsche, S. D.; Jin, H.; Patel, J.; Skelton, B. W.; White, A. H.
Organometallics 1997, 16, 2175. (h) Prokopchuk, E. M.; Jenkins, H. A.;
Puddephatt, R. J. Organometallics 1999, 18, 2861. (i) Fekl, U.; Zahl, A.;
van Eldik, R. Organometallics 1999, 18, 4156. (j) Haskel, A.; Keinan, E.
Organometallics 1999, 18, 4677.
(38) Huang, D.; Streib, W. E.; Bollinger, J. C.; Caulton, K. G.; Winter,
R. F.; Scheiring, T. J. Am. Chem. Soc. 1999, 121, 8087.
(39) Blomberg, M. R. A.; Siegbahn, P. E. M.; Nagashima, U.; Wenner-
berg, J. J. Am. Chem. Soc. 1991, 113, 424.
(40) Minkin, V. I.; Simkin, B. Ya.; Minyaev, R. M. Quantum Chemistry
of Organic Compounds; Springer-Verlag: Berlin, 1990.

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10841
without solvents. This results in the estimated ZPE-corrected
data that are also listed in Table 2.
To further probe the effects of coordinated or associated
solvent molecules, we wanted to estimate the energy change
when TFE is solvated rather than bonded to the Pt complex.
Whether TFE is bonded at Pt or diimine or is in solution, the
CF3CH2 tail of TFE will interact with adjacent TFE solvent
molecules. The main contribution to the difference between
coordinated and solvated TFE will be the bond between the
TFE hydroxyl group and either Pt, diimine, or TFE. Calculations
on various (TFE)n (n ) 2-5) clusters indeed showed that the
strongest TFE-TFE interaction was of the OsH‚‚‚O type. The
interaction was calculated as the average energy lost per
hydrogen bond when the TFE cluster was dissociated into
separate molecules at infinite distance.43 The strongest hydrogen
bond was found to be 31 kJ/mol for a TFE pentamer.44 We
have already described that a TFE molecule interacts with the
diimine C-C bond with 28 kJ/mol or less in all calculated
structures, so it appears that the energy involved is ap-
proximately the same whether a TFE molecule binds to the
diimine or to the TFE solvent. TFE association/dissociation must
occur rapidly under the C-H activation conditions.
The influence of H2O on the activation barrier for the
oxidative addition process was probed by investigating its
microscopic reverse. The search for the transition state for the
reductive elimination was attempted, starting from the octahedral
Pt(IV) complex (N-N)Pt(CH3)2(H)(H2O)+ (G). The H-Pt-
C(methyl) angle was used as the reaction coordinate. The H2O
ligand was found to depart from the Pt center during the
reductive elimination, and at the stage when a σ-complex was
formed, the H2O had reattached itself at the diimine C-C bond.
We were not entirely successful at locating the exact transition
state. Two imaginary frequencies were calculated, one large that
corresponds to the C-H bond formation, and one very small
(9i cm-1) that involves a twist movement of coordinated H2O
around the long and flexible Pt-O bond. Since the Pt-O bond
has been elongated by 0.50 Å relative to in G, this twisting
movement should give only a negligible contribution to the
energy, and the energy of this structure (M) should be very
close to that of the true transition state. Further movement along
the chosen reaction coordinate gave the σ-complex K with H2O
associated at the diimine C-C bond, on the opposite side of
the Pt-diimine plane relative to the Pt-bonded H atom of
coordinated methane. The calculated activation energy of
oxidative addition is 27 kJ/mol, and the reaction energy is -33
kJ/mol.
Figure 3. Association of a H2O molecule at the diimine moiety of
(N-N)Pt(CH3)(CH4)+ in anti (K) and syn (K′) mode relative to the
bridging H of the methane ligand.
Explicit Inclusion of a H2O or TFE Molecule on the C-H
Activation. Calculations including a solvent molecule were
performed starting at the square planar (N-N)Pt(CH3)(CH4)+
σ-complex D. All stationary points were reoptimized with an
H2O or TFE molecule attached. Several points of attachment
for H2O and TFE were tried in each case, and it was found that
for the σ-complex as well as for the transition state of the σ-bond
metathesis, the most stable structures contain the solvent
molecule associated at the C-C bond of the diimine. In no cases
were any minima located where the solvent molecule coordi-
nates to the metal center in the σ-complex through the use of
the O (or F for TFE) lone pairs or through hydrogen bonding
between the solvent oxygen and the Pt-bonded hydrogen atom
of the methane ligand.41 The calculated lowest-energy structure
of H2O‚(N-N)Pt(CH3)(CH4)+ (K) has the H2O molecule located
on the opposite side of the Pt-diimine plane with respect to
the Pt-bonded H of the methane ligand. A structure with
methane-H and H2O on the same side of the plane was also
found (K′), at a local energy minimum only 1 kJ/mol higher in
energy. The structures of K and K′ are depicted in Figure 3. A
similar situation pertained to TFE (L and L′, respectively; in
this case L was less stable than L′ by 6 kJ/mol).
The calculations showed that H2O and TFE association at
the C-C bond of diimine occurs mostly through an electrostatic
attraction (calculated Mulliken charges are +0.3 at the diimine
carbon atoms, and -0.6 and -0.7 at oxygen from H2O and
TFE, respectively). This situation is fairly general for the cationic
(N-N)Pt complexes, for which calculations were performed with
H2O or TFE present. In this context, it is interesting to note
that a 19F{1H} HOESY NMR investigation of (diimine)Pt-
(CH3)(alkene)+BF4- ion pairs in dichloromethane-d2 indicated
-
that the BF4 counteranion is associated with the cation near
the diimine moiety.42 The binding energies between H2O and
the diimine C-C bond were calculated for several Pt com-
plexes: H2O‚(N-N)Pt(CH3)(CH4)+ (K), 30 kJ/mol; K′, 29 kJ/
mol; H2O‚(N-N)Pt(CH3)2(H)+, 32 kJ/mol; transition state for
σ-bond metathesis, 28 kJ/mol. For TFE binding, the data were,
for TFE‚(N-N)Pt(CH3)(CH4)+ (L), 21 kJ/mol; L′, 27 kJ/mol;
TFE‚(N-N)Pt(CH3)2(H)+, 28 kJ/mol; transition state for σ-bond
metathesis, 27 kJ/mol. Nonsolvent zero-point energies are not
included in these bond energies, as the correction does not
contribute to the bond energy of solvent binding to the diimine.
The minimum energy structures of species containing TFE were
not verified by vibrational frequency calculations, due to their
extensive size. The non-ZPE-corrected energies are included
in Table 2. In addition, a relatively simple partial ZPE correction
was applied to the calculated energies for the solvent-associated
species: it was assumed that the same ZPE correction applies
to solvent-associated species as to the corresponding species
The effect of including a water molecule on the σ-bond
metathesis mechanism was also attempted. The σ-bond meta-
thesis could not be investigated the same way as for the
nonsolvent case because inclusion of H2O removed all sym-
metry. This prevented optimization with symmetry restrictions,
in contrast to the optimization of the σ-bond metathesis with
no solvent. The σ-bond metathesis pathway was explored instead
by attaching H2O to different sites of the frozen structure of
the nonsolvent transition state. It was found that the added
(43) The maximum number of hydrogen bonds was obtained by placing
the molecules in a circle such that the overlap between H and a lone pair
of O was optimal for all hydrogen bonds. Binding energies were calculated
for several (TFE)n clusters, and the largest energies obtained per hydrogen
bond for each oligomer were n ) 2, 19 kJ/mol; n ) 4, 27 kJ/mol; n ) 5,
31 kJ/mol. The last value is used as the estimate for the energy that is
gained when the O-H group of TFE binds to the solvent.
(44) Although comparisons are not directly relevant, it is interesting to
note that the experimentally determined heat of vaporization of TFE is ∆H°
) 44 kJ/mol: Rochester, C. H.; Symonds, J. R. J. Chem. Soc., Faraday
Trans. 1 1973, 69, 1267.
(41) Hydrogen bonds between transition metal hydrides and proton
donors such as O-H and N-H have recently been reported. For a review,
see: Crabtree, R. H.; Siegbahn, P. E. M.; Eisenstein, O.; Rheingold, A. L.;
Koetzle, T. F. Acc. Chem. Res. 1996, 29, 348.
(42) Zuccaccia, M.; Macchioni, A.; Orabona, I.; Ruffo, F. Organome-
tallics 1999, 18, 4367.

10842 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
Figure 4. Energy diagram depicting ground-state structures in the C-H activation of methane. All energies are in kilojoules per mole relative to
the σ-complex D, and ZPE corrections are not included.31
solvent molecule was preferentially associated at the diimine
C-C bond. The H2O-diimine bond strength was essentially
unchanged (a 2 kJ/mol weakening) when compared to that of
the starting σ-complex. No energy minimum was located in
which H2O coordinates at the Pt atom of the frozen transition-
state structure for the σ-bond metathesis, and no evidence for
hydrogen bonding between H2O and the migrating H atom
(Mulliken charge of -0.1) was found.41 Thus, short-range
interactions with a H2O molecule appear to have little effect
on the energy barrier for the σ-bond metathesis.
process. Under actual reaction conditions, (Nf-Nf)Pt(CH3)(H2O)+
is the only observable complex in solution, and this is consistent
with B being of lowest energy. No evidence has been found
for the coexistence in solution of (Nf-Nf)Pt(CH3)(TFE)+, mod-
eled by C, although an equilibrium between H2O and TFE
adducts has been observed and quantified in a related system.24
The proposed intermediate σ-complex D that is crucial to explain
the occurrence of multiple H/D exchange between Pt-CH3 and
CD4 is located 59 kJ/mol higher in energy than the aqua complex
B. The σ-complex may be accessed by a preequilibrium
dissociation of H2O from B to give the three-coordinate species
A, 158 kJ/mol higher in energy than B. The dissociation of H2O,
experimentally hinted at by the partial inhibition of the reaction
caused by addition of water, will significantly contribute to the
overall barrier of the reaction if A is involved.45 The five-
coordinate complex E is 29 kJ/mol uphill from the σ-complex,
but coordination of a second axial ligand H2O or TFE renders
the formation of the six-coordinate oxidative addition products
G or H more favorable. In fact, the six-coordinate products G
and H are essentially of the same energy as the starting
complexes B and C, respectively. However, entropy effects are
not included and will render the overall reactions to give G or
H less favorable by ca. 45 kJ/mol under ambient conditions.46
The relatively high stability predicted for G is reconciled with
the fact that hydridoalkyl complexes (Nf-Nf)Pt(CH3)2(H)(L)+
(8a,b) were observable at low temperatures in dichloromethane-
d2. The methane σ-complex has not been observed under the
actual reaction conditions, and this is readily understood
considering that D is much higher in energy than B or G and
therefore must have a short lifetime under reaction conditions
where it is thermally generated from B or G.
Calculated data for the reactions with and without H2O and
TFE association are summarized in Table 4. Figure 5 depicts
energy diagrams for the oxidative addition and σ-bond meta-
thesis pathways for the methane C-H activation reaction. To
the left, the ZPE-corrected results of the nonsolvent calculations
(45) The calculated 158 kJ/mol barrier is considerably larger than that
implicated by the experimental observations (the Eyring equation yields a
free energy of activation of ca. 110 kJ/mol when a rate constant
corresponding to a half-life of 48 h at 45 °C is used). This may be another
hint that an associative mechanism for methane coordination should be
seriously considered. We hope to address this point in a future contribution.
(46) (a) Page, M. I. Angew. Chem., Int. Ed. Engl. 1977, 16, 449. (b)
Nolan, S. P.; Hoff, C. D.; Stoutland, P. O.; Newman, L. J.; Buchanan, J.
M.; Bergman, R. G.; Yang, G. K.; Peters, K. S. J. Am. Chem. Soc. 1987,
109, 3143.
The influence of TFE on the activation barrier for the
oxidative addition process was also probed by investigating its
microscopic reverse starting from H. The distance between the
hydride and the methyl carbon atom was used as the reaction
coordinate and was decreased until a methane σ-complex was
formed. The results were quite similar to those just described
with H2O included. The TFE ligand departed from the metal
and reattached itself to the molecule at the diimine. This is
entirely consistent with the previously found tendency for TFE
to interact with the diimine part of the σ-complex. The attempts
at locating the exact transition state for this reductive elimination
were again not entirely successful. Two imaginary frequencies
were calculated, one large that corresponds to the C-H bond
formation and a second, small one (32i cm-1) that could not be
eliminated. In the mode corresponding to the latter, the TFE
molecule moves sideways without affecting Pt-O bond length.
The Pt-O bond is elongated by 0.54 Å relative to the minimum
in the oxidative addition product H and is therefore very flexible.
This fact, as well as the imaginary frequency being so small,
makes it reasonable to assume that the energy of this structure
(N) is close to that of the true transition state. This gives an
approximate activation energy of 56 kJ/mol for the reductive
elimination and 22 kJ/mol for the reverse, the oxidative addition.
Finally, the σ-bond metathesis pathway was explored by
attaching TFE to different sites of the frozen structure of the
nonsolvent transition state. The results paralleled those found
when H2O was included. As for the σ-complex, TFE would
attach only to the diimine group, with similar bond strengths in
both the frozen nonsolvent σ-bond metathesis transition state
and in the σ-complex L.
Mechanistic Consequences of the Computational Results.
Figure 4 summarizes some calculated data (non-ZPE data from
Table 2, since these give the most complete diagram) of ground-
state species that are likely to be involved in the C-H activation

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10843
Table 4. Relative Energies, kJ/mol
together led to a monotonic increase in energy with concomitant
PH3 ejection. For the isomer with a trans (relative to H) chloride,
bringing H and CH3 together led to chloride dissociation, but
in this case the energy did not increase monotonicallysin fact,
a transition state for the simultaneous methane elimination and
chloride dissociation was located in which the Pt-Cl distance
had increased by 0.26 Å compared to the ground-state distance.
The 82 kJ/mol barrier for this reaction was dramatically greater
than the 4 kJ/mol barrier toward elimination from the corre-
sponding five-coordinate complex that results from removal of
the chloride trans to hydride. The 78 kJ/mol energy difference,
of course, does not take into account the significant but
unreported energy cost of chloride predissociation that is needed
when the six-coordinate complex is the starting point. Siegbahn
and Crabtree29b also found that methane elimination with
concomitant ligand dissociation occurred from (H2O)2PtCl2-
(CH3)(H).
Interestingly, our calculations suggest a pathway for the
reductive elimination from G with a transition state M in which
the H2O ligand is incompletely dissociated from the Pt center
in the six-coordinate complex. This behavior parallels the
chloride dissociation from one of the three isomers of (PH3)2-
PtCl2(CH3)(H), as mentioned above. The 60 kJ/mol energy
barrier for the elimination from G via M is considerably greater
than the 10 kJ/mol barrier from E via I (Figure 5). However,
due to the 87 kJ/mol energy cost of a predissociation of H2O
from G, the concerted elimination/dissociation (G-M-K) will
have a lower overall barrier, by 36 kJ/mol, than a two-step
elimination (G-E-I-D). An attractive interaction between the
diimine moiety and the departing H2O in the transition state
presumably contributes to this effect.
Concluding Remarks. We have demonstrated that a cationic
Pt(II) diimine complex is capable of activating benzene and
methane C-H bonds under mild conditions. These reactions
proceed in the hydroxylic solvent TFE, and in the presence of
substantial quantities of watersa potential ligand that is usually
excluded from the reaction media for such reactions. Multiple
H/D exchanges between coordinated CH3 groups and C6D6 or
CD4 imply the intermediacy of methane σ- and benzene σ- or
π-complexes. Theoretical calculations suggest that an oxidative
addition pathway is favored over a σ-bond metathesis alternative
for the methane C-H activation. Single water or TFE molecules
associate with the involved complexes preferentially at the
diimine moiety of square planar Pt(II) species and at a vacant
apical site of square pyramidal Pt(IV) species. The thermody-
namics and kinetics of C-H bond formation from Pt(IV) and
of C-H bond cleavage from Pt(II) methane σ-complexes show
a pronounced dependence on whether water or TFE is present
or not.
solvent
none,
ZPE-
none,
not ZPE-
TFE,
partially
H2O,
partially
correcteda correcteda ZPE-correctedb ZPE-correctedb
σ-complex
(CH4)
oxidative
addition
transition
state
0
0
-27
-30
-4
33
40
2
oxidative
addition
product
σ-bond
23
44
29
50
-55
-64
17
16
metathesis
transition
state
a From Table 2. b Obtained by applying a partial ZPE correction to
the solvent-associated species. The partial correction amounts to the
ZPE correction that was applied to the corresponding solvent-free
species (D, E, I, J) in Table 2 and simply equals the difference between
ZPE and non-ZPE-corrected values in Table 2.
are illustrated. The oxidative addition to form the five-coordinate
complex E is 23 kJ/mol uphill in energy with activation barriers
of 33 kJ/mol for the oxidative addition and 10 kJ/mol for the
reverse. The σ-bond metathesis exchange reaction has an
activation barrier of 44 kJ/mol and is therefore disfavored by
11 kJ/mol relative to oxidative addition.
This preference is retained but somewhat modified when a
molecule of H2O is included in the calculations. These calcula-
tions, as already mentioned, were done without ZPE corrections.
A partial ZPE correction may, however, be applied; see footnote
b to Table 4. The H2O molecule associates at the diimine moiety
of D and J involved in the σ-bond metathesis mechanism with
the same binding energy of 28-30 kJ/mol. This lowers the
ground-state and transition-state energies to approximately the
same extent, and therefore, no essential change in the actiVation
energy is found for the σ-bond metathesis. Contrasting this, the
interaction of H2O with the oxidative addition product E through
axial coordination is much greater (87 kJ/mol), leading to
considerable stabilization of the product G. The transition state
for the oxidative addition, N, contains a considerably weakened
(by ca. 0.5 Å) Pt-OH2 bond. Here, the transition-state energy
is lowered by 36 kJ/mol relative to the situation where H2O
coordination or association was not included. The activation
barrier for oxidative addition is reduced to 27 kJ/mol (from 33
kJ/mol without H2O), and for the reductive elimination it has
increased from 10 to 60 kJ/mol as a consequence of the
thermodynamic stabilization of the oxidative addition product
by H2O coordination. The preference for oxidative addition over
σ-bond metathesis for the exchange process has increased from
11 to 20 kJ/mol. For TFE, the preference for oxidative addition
amounts to 16 kJ/mol.
Experimental Section
General Considerations. Acetonitrile was distilled from P2O5, and
dichloromethane and dichloromethane-d2 were distilled from CaH2.
Trifluoroethanol (TFE) and TFE-d3 were distilled from CaSO4 and a
A large body of experimental evidence47 suggests that C-H
and C-C reductive eliminations from six-coordinate Pt(IV)
complexes normally proceed with prior ligand dissociation so
that the actual elimination occurs from a five-coordinate
intermediate. This has been recently supported by DFT
calculations.29e Attempts to computationally induce reductive
elimination of methane from three isomers of (PH3)2PtCl2(CH3)-
(H), all with cis phosphines and with the methyl and hydride
ligands cis disposed, caused concomitant loss of the ligand that
was trans to the hydride. For two isomers, forcing H and CH3
1
small amount of NaHCO3. H, 13C{1H}, and 19F NMR spectra were
recorded on Bruker Advance DXP 200 and 300 instruments.
1H and 13C chemical shifts are reported in ppm relative to
tetramethylsilane, with the residual solvent proton resonance as internal
standards. 19F chemical shifts are reported in ppm relative to CFCl3.
For TFE-d3, the solvent fluorine resonance was used as internal standard
(δ -77.72). Elemental analyses were performed by Ilse Beetz Mik-
15
roanalytisches Laboratorium, Kronach, Germany. Pt(CH3)4(µ-SMe2)2
16
and Pt(C6H5)2(SMe2)2 were prepared as described in the literature.
The following compounds were prepared as described in the
Supporting Information to ref 13: Nf-Nf, 1, 2, 3(BF4-), 4(BF4-), and
6(BF4-).
(47) Crumpton, D. M.; Goldberg, K. I. J. Am. Chem. Soc. 2000, 122,
962 and references cited, as well as refs 14 and 15 cited in ref 29e.

10844 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Heiberg et al.
Figure 5. Energy diagram (kJ/mol) depicting the C-H activation step following oxidative addition and σ-bond metathesis pathways. Left, no H2O
included; ZPE-corrected values. Right, one H2O molecule included; partial ZPE correction applied as explained in footnote b to Table 4.
Attempted Preparation of (Nf-Nf)Pt(CH3)(OTf) (5). A -40 °C
solution of 1 (24.1 mg, 3.29 × 10-2 mmol) in dichloromethane (10
mL) was added HOTf (2.9 µL, 3.3 × 10-2 mmol), and the reaction
mixture was stirred for 15 min. The solvent was slowly removed by
vacuum transfer, while the solution was maintained at -40 °C. The
residue was dried in vacuo at ambient temperature to give an orange
powder. NMR analysis immediately after dissolving the material in
dichloromethane-d2 showed that 5 is the major product formed in the
reaction (ca. 90%). 1H NMR (200 MHz, dichloromethane-d2): δ 1.07
(100 µL) and diethyl ether (40 µL), the tube was sealed and cooled to
-78 °C. The tube was shaken in order to mix the reactants (care was
taken to minimize any heating of the sample), and a pale yellow solu-
tion was immediately obtained. The tube was then transferred to a
1
precooled NMR probe. H NMR (300 MHz, -55 °C), for 8a (65%):
1
2
δ -26.67 (s, J(195Pt-H) ) 1741 Hz, 1 H, PtH), 0.54 (s, J(195Pt-H)
) 63.6 Hz, 6 H, PtMe), 2.42 (s, 6 H, NCMeCMeN), 7.36 (s, 2 H,
ArHp), 7.92 (s, 4 H, ArHo). 1H NMR (300 MHz, -55 °C), for 8b
(35%): δ -26.40 (s, 1J(195Pt-H) ) 1699 Hz, 1 H, PtH), 0.48 (s, 2J(195
-
(s, J(195Pt-H) ) 74.9 Hz, 3 H, PtMe), 1.65 (s, J(195Pt-H) ) 13.1
Hz, 3 H, NCMeC′MeN), 2.01 (s, 3 H, PtNCMe), 2.13 (s, 3 H,
NCMeC′MeN), 7.62 (s, 2 H, ArHo), 7.71 (s, 2 H, Ar′Ho), 7.96 (s, 1 H,
ArHp), 8.01 (s, 1 H, Ar′Hp). 19F NMR (188 MHz, dichloromethane-
d2): δ -78.48 (s, 4J(195Pt-F) ) 14.5 Hz, 3 F, BF4-), -63.47 (s, 6 F,
ArCF3), -63.36 (s, 6 F, Ar′CF3). Compound 5 decomposes in
dichloromethane over time, and after a period of 62 h the signals for
5 had decreased to ca. one-third of their original intensities. The
degradation of 5 is accompanied by precipitation of brown, unidentified
material and formation of CH4.
Pt-H) ) 62.9 Hz, 6 H, PtMe), 2.37 (s, 6 H, NCMeCMeN), 7.36 (s, 2
H, ArHp), 7.71 (s, 4 H, ArHo). At temperatures above ca. -40 °C, the
species 8a and 8b decompose under elimination of methane to give
unidentified products. No other species could be observed prior to
methane loss.
Theoretical Methods and Software. The geometries of all Pt
complexes were optimized with density functional theory (DFT) using
the ADF48 program on a Cray T3E computer at the Norwegian
University of Science and Technology, Trondheim, and on a HP V2500
computer at the University of Tromsø. The ab initio CCSD(T)
calculations were performed with Gaussian9436 on a Cray Origin 2000
computer at the High Performance Computer Center at the University
of Bergen.
Slater exchange and the VWN parametrization of the LDA correla-
tion energy were used,49 with the gradient corrections of Becke50 for
exchange and of Perdew51 for correlation (BP86). The gradient
corrections were added self-consistently. The frozen core approximation
was used for all atoms except hydrogen. The orbitals up to 4f and 1s
were frozen in their atomic shapes for Pt and first-row atoms,
respectively. The basis sets have the overall quality of TZV for Pt and
DZVP for the other atoms. (We find that frozen 5p orbitals for Pt give
2
4
C-H Activation of Benzene and Formation of 7(BF4-). To a
solution of 6(BF4-) (6.8 mg, 8.3 × 10-3 mmol) in TFE-d3 (0.6 mL) in
an NMR tube was added benzene (22 µL, 0.25 mmol). The reaction
was monitored by 1H and 19F NMR spectroscopies. A smooth
conversion of 6 to the phenyl analogue [(Nf-Nf)Pt(C6H5)(OH2)]+ (7)
was observed, with concomitant formation of CH4, detected by 1H NMR
1
(δ 0.14). Spectroscopic data for 7(BF4-) follow. H NMR (200 MHz,
TFE-d3): δ 1.91 (s, 3 H, NCMeC′MeN), 2.19 (s, 3 H, NCMeC′MeN),
6.65-6.85 (m, 5 H, Pt-C6H5), 7.29 (s, 2 H, ArHo), 7.65 (s, 1 H, ArHp),
7.79 (s, 2 H, Ar′Ho), 8.05 (s, 1 H, Ar′Hp). 19F NMR (188 MHz, TFE-
d3): δ -151.95 (s, 4 F, BF4-), -63.65 (s, 6 F, ArCF3), -63.48 (s, 6
F, Ar′CF3).
-
(48) ADF 2.3.0, Theoretical Chemistry, Vrije Universiteit, Amsterdam.
(a) Baerends, E. J.; Ellis, D. E.; Ros, P. Chem. Phys. 1973, 2, 41. (b) te
Velde, G.; Baerends, E. J. J. Comput. Phys. 1992, 99, 84. (c) Fonseca
Guerra, C.; Visser, O.; Snijders, J. G.; te Velde, G.; Baerends, E. J. In
Methods and Techniques for Computational Chemistry, METECC-95;
Clementi, E., Coronglu, G., Eds.; STEF: Cagliari, 1995; pp 303-395.
(49) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(50) (a) Becke, A. D. Phys. ReV. A 1988, 33, 2786. (b) Becke, A. D. In
The Challenge of d and f Electrons: Theory and Computation; Salahub,
D. R, Zerner, M. C., Eds.; ACS Symposium Series 394; American Chemical
Society: Washington, DC, 1989. (c) Becke, A. D. Int. J. Quantum Chem.
1989, S23, 599. (d) Becke, A. D. Phys. ReV. A 1988, 38, 2398.
(51) Perdew, J. P. Phys. ReV. B 1986, 33, 8822; Phys. ReV. B. 1986, 34,
7406 (erratum).
C-H Activation of CH4 and CD4. In a typical experiment, 6(BF4
)
dissolved in TFE-d3 (0.4 mL) was placed in a high-pressure NMR tube.
A known amount of methane (corresponding to 20-25 bar at ambient
temperature) was condensed into the NMR tube. The tube was sealed
and placed in a oil bath at 45 °C. The reactions were monitored by 1H
NMR.
Generation of (Nf-Nf)PtMe2(H)(L)+(BF4-) (8a(BF4-) and 8b(BF4-))
at -78 °C. To an NMR tube loaded with a solution of 1 (2.7 mg, 3.7
× 10-3 mmol) in dichloromethane-d2 (400 µL) was slowly added
dichloromethane-d2 (100 µL) to form a separate upper layer. After
subsequent slow addition of a mixture of 54% solution of HBF4 in
ether (1.0 µL, 7.3 × 10-3 mmol) in a mixture of dichloromethane-d2

C-H ActiVation at a Cationic Pt(II) Diimine Aqua Complex
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10845
excellent bond lengths for Pt-H and Pt-C, but the calculated Pt-N
bond lengths are too long by ca. 0.15 Å for the N to Pt.)
ligand XsNdC(Y)sC(Y)dNsX than in the experiments. Six model
systems were considered: X ) CHdCH2, CH3, or H, and Y ) CH3 or
H. Calculated reaction energies for four relevant reactions were
compared for the models with different diimine ligands, as well as for
the full-sized Pt complex. The following test reactions were used.
The stationary points of the model complexes within the nonsolvent
approximation were characterized by calculating the vibrational spec-
trum, showing one and only one imaginary frequency for transition
states and no imaginary frequency for energy minima, unless otherwise
stated. Frequencies were computed in ADF by numerical differentiation
of energy gradients in slightly displaced geometries. Each gradient was
based on the gradient at the equilibrium and the gradients of two
displaced geometries. The accuracy of the numerical integration for
vibrational calculations was set to 10-6.5. For geometry optimizations
it was set to 10-5.0 or higher accuracy when necessary, to converge the
geometry.
(diimine)PtMe+ + CH4 f ax-(diimine)PtMe2H+
(diimine)PtMe+ + CH4 f eq-(diimine)PtMe2H+
(diimine)PtMe(OOCCF3) + CH4 f (diimine)PtMe2 + CF3COOH
(diimine)PtMe(OSO2CF3) + CH4 f (diimine)PtMe2 + HOSO2CF3
The ab initio calculations were performed with Gaussian94.36 The
basis set for Pt was a modified version of Gaussian 94’s LANL2DZ
basis, which was extended through decontraction of the s and p shells
to triple-ú, and by addition of an extra uncontracted d-function with
exponent 0.0457. The number of contracted valence basis functions,
in order of increasing angular momentum, is (4,4,3). The basis set
6-31G(d,p) was used for all the other atoms.
Acknowledgment. We gratefully acknowledge generous
support from the Norwegian Research Council, NFR (stipends
to L.J. and H.H. as well as computational time), and from Statoil
under the VISTA program, administered by the Norwegian
Academy of Science and Letters. We thank Aud M. Bouzga
for kind assistance with some NMR experiments and Ulf
Wahlgren for helpful discussions.
Testing Model Systems for Calculations. It was necessary to model
the reaction mechanisms with Pt complexes having a simpler diimine
JA0019171