Comparative gas-phase and solution-phase investigations of the mechanism of C-H activation by [(N-N)Pt(CH3)(L)]+

Article abstract of DOI:10.1021/om0209982

Contrary to previous claims in the literature, there is no solvent assistence by 2,2,2-trifluoroethanol (TFE) in C-H activation reactions by cationic Pt(II) complexes. Both electrospray ionization tandem mass spectrometric studies of isolated gas-phase in

Full text of DOI:10.1021/om0209982

Organometallics 2003, 22, 2217-2225
2217
Com p a r a tive Ga s-P h a se a n d Solu tion -P h a se
In vestiga tion s of th e Mech a n ism of C-H Activa tion by
[(N-N)P t(CH3)(L)]+
Gerd Gerdes and Peter Chen*
Laboratorium fu¨r Organische Chemie, Swiss Federal Institute of Technology,
ETH Zu¨rich, Switzerland
Received December 10, 2002
Contrary to previous claims in the literature, there is no solvent assistence by 2,2,2-
trifluoroethanol (TFE) in C-H activation reactions by cationic Pt(II) complexes. Both
electrospray ionization tandem mass spectrometric studies of isolated gas-phase intermedi-
ates and solution-phase kinetic spectroscopy indicate that TFE serves only as an inert diluent.
TFE was found in the gas-phase experiment to provide no acceleration of the loss of methane
or benzene from preselected [(N-N)Pt(CH3)(C6H6)]+ or [(N-N)Pt(CH4)(C6H5)]+ ions. In
solution, the overall C-H activation reaction of benzene by [(N-N)Pt(CH3)(H2O)]+BF4- was
found to be zero-order in TFE. While it is on one hand disappointing that the claimed
facilitation of C-H activation by solvent assistance by TFE at the rate-determining transition
state is not supported by the present experimental data, it is on the other hand encouraging
that extremely rapid C-H activation (on the time scale of seconds at room temperature) in
solution may be achieved in ordinary solvents as long as the solvent is not strongly
coordinating.
In tr od u ction
the considerable long-term potential of C-H activation
reactions and the putative role of TFE in facilitating
the reaction, i.e., solvent assistence, a mechanistic
investigation to clarify that role is timely. Moreover,
given that all of the complexes in the reaction are
cationic, the question of mechanism is directly accessible
by gas-phase reactions of the complexes in a tandem
mass spectrometer where the solvent molecule can be
explicitly included or excluded from the reaction. We
report both a gas-phase, tandem mass spectrometric
study of the C-H activation of benzene by [(N-N)Pt-
(CH3)(L)]+, 1a -d , (N-N) ) Ar-NdC(CH3)-C(CH3)d
N-Ar, in which the reaction is performed in both the
forward and reverse directions with mass-selected com-
plexes, and an accompanying solution-phase kinetic
study of the comparable reaction.
Since the discovery by Shilov1 that certain platinum
salts can catalyze the functionalization of methane,
C-H activation has been an active field of investiga-
tion.2 Given the difficulty in working with the poorly
defined species that occur in the original Shilov system,
model systems that mimic one or more steps of the
proposed catalytic cycle have been studied under condi-
tions where the reactions may be cleanly observed and
characterized. Previous studies have succeeded in char-
acterizing many of the key steps and reactive interme-
diates in the catalytic cycle.2 Experiments by the
Caltech and Oslo groups have recently focused on the
mechanism of C-H activation of methane and arenes
by Pt(II) complexes in 2,2,2-trifluoroethanol (TFE)3-6 as
solvent; the original report claims C-H activation
“under the mildest reaction condition yet reported for
such processes at cationic Pt complexes”. Subsequent
publications assign the rate-determining step in C-H
activation of benzene by sterically crowded Pt(II) com-
plexes to the solvent-assisted associative ligand ex-
change of benzene for coordinated TFE, on the basis of
kinetic4,7 and activation volume8 measurements. Given
Both sets of experiments strongly suggest that TFE
plays no role whatsoever in the rate-determining transi-
tion state for the overall C-H activation reaction in
either gas phase or solution. The importance of this
result for C-H activation is also discussed.
Exp er im en ta l Section
The preparation of the starting complexes, X-ray structures,
and further details on the execution of the experiments are
given in the Supporting Information. Complexes 3 and 4 were
synthesized as the dimethyl complexes as described in the
literature.4,9 Analytical data matched those previously re-
ported. 1a -d were electrosprayed from a 10-5 M solution in
(1) Shilov, A. E.; Shulpin, G. B. Chem. Rev. 1997, 97, 2879.
(2) Labinger, J . A.; Bercaw, J . E. Nature 2002, 417, 507. Stahl, S.
S.; Labinger, J . A.; Bercaw, J . E. Angew. Chem., Int. Ed. 1998, 37,
2180. Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H.
Acc. Chem. Res. 1995, 28, 154. Crabtree, R. H. Chem. Rev. 1995, 95,
987.
(3) J ohansson, L.; Ryan, O. B.; Tilset, M. J . Am. Chem. Soc. 1999,
121, 1974.
(4) J ohansson, L.; Tilset, M.; Labinger, J . A.; Bercaw, J . E. J . Am.
Chem. Soc. 2000, 122, 10846.
(5) Heiberg, H.; J ohansson, L.; Gropen, O.; Ryan, O. B.; Swang, O.;
Tilset, M. J . Am. Chem. Soc. 2000, 122, 10831.
(6) J ohansson, L.; Ryan, O. B.; Rømming, C.; Tilset, M. J . Am. Chem.
Soc. 2001, 123, 6579.
(7) Zhong, H. A.; Labinger, J . A.; Bercaw, J . E. J . Am. Chem. Soc.
2002, 124, 1378.
(8) Procelewska, J .; Zahl, A.; van Eldik, R.; Zhong, H. A.; Labinger,
J . A.; Bercaw, J . E. Inorg. Chem. 2002, 41, 2808.
(9) Scollard, J . D.; Day, M.; Labinger, J . A.; Bercaw, J . E. Helv.
Chim. Acta 2001, 84, 3247.
10.1021/om0209982 CCC: $25.00 © 2003 American Chemical Society
Publication on Web 04/12/2003

2218 Organometallics, Vol. 22, No. 11, 2003
Gerdes and Chen
Sch em e 1
TFE to which up to 25% water or acetonitrile was added,
depending on the experiment. After desolvation, the ions were
mass-selected and reacted in the second octopole region of a
modified Finnigan MAT TSQ-700 tandem mass spectrometer,
as has been described in previous reports.10,11
solution of HBF4, diluted with 1 mL of ether, was added to
the cuvette. Spectra were then recorded immediately.
Resu lts
Solution-phase UV/vis kinetic spectroscopy was done with
either a Perkin-Elmer Lambda 6 or a Hitachi U-2010 UV/vis
spectrophotometer. The cuvette holders were thermostated by
a Haake F3 heat regulator. For the first series of experiments,
0.25 mM (purple) solutions of the dimethyl precursor to 1b or
4b in TFE containing 0.16 M water were treated with 1.1 µL
of a 50% aqueous solution of HBF4 in the cuvette. The orange
solutions were thermalized and then treated with a measured
volume of benzene (at the same temperature). The solution
should be observed carefully at the higher benzene concentra-
tions for indications of phase separation, usually evidenced
by the appearance of turbidity, which would compromise
quantitative rate measurements. For the second series of
experiments, benzene was mixed with water and heated to 50
°C. The mixture was cooled; the layers were then separated
at 25 °C. The concentration of water in saturated wet benzene
has been reported12 to be 0.029 M. Solutions of 0.25 mM of
the dimethyl precursors to 1b (green) and 4b (purple) were
prepared in the water-saturated benzene and thermalized.
Stock solution (40 µL) prepared from 5.7 mg of a 50% aqueous
Selection of 1a versus 1b (Scheme 1), for example,
by m/z ratio was straightforward. An adduct with
benzene having the structure 1e, 1f, or 2 could be
prepared by reaction of 1a with benzene vapor (∼8.5
mTorr) in the 24-pole ion guide region, followed by
selection by mass. While the experiments below do not
allow an unambiguous structural assignment to 1e, 1f,
or 2, the reactivity of the ion, and particularly, the
isotopic scrambling experiment, as detailed below, sug-
gests either that it is 2 or that the interconversion
between 1e, 1f, and 2 is much faster than elimination
to form either 1a or 1g. The adduct is therefore
designated as 1e/1f/2. All reactions in the subsequent
octopole were performed at low collision energies.
We find unsurprisingly that addition of benzene to
1b and 1c occurs inefficiently under all reactions
conditions; 1d does not react at all. On the other hand,
1a reacts with benzene with good reaction efficiency at
low collision energy to form an adduct, 1e/1f/2, as well
as a product, 1g, corresponding to subsequent elimina-
tion of methane.13 Given the coordinative unsaturation
of 1a , the facile product formation is also unsurprising.
Figure 1 shows, nevertheless, that this C-H activation
reaction is not only facile but also very clean. No ions
other than the starting complex 1a , the adduct 1e/1f/2,
the product 1g, and a further benzene complex of 1g
are visible.
(10) Feichtinger, D.; Plattner, D. A. Angew. Chem. 1997, 109, 1796.
Feichtinger, D.; Plattner, D. A.; Chen, P. J . Am. Chem. Soc. 1998, 120,
7175. Hinderling, C.; Adlhart, C.; Chen, P. Angew. Chem. 1998, 110,
2831. Hinderling, C.; Chen, P. Angew. Chem., Int. Ed. 1999, 38, 2253.
Kim, Y. M.; Chen, P. Int. J . Mass Spectrom. 1999, 185-7, 871.
Hinderling, C.; Chen, P. Int. J . Mass Spectrosc. Ion Processes 2000,
195/ 196, 377. Hinderling, C.; Adlhart, C.; Chen, P. Chimia 2000, 54,
232. Adlhart, C.; Hinderling, C.; Baumann, H.; Chen, P. J . Am. Chem.
Soc. 2000, 122, 8204. Adlhart, C.; Chen, P. Helv. Chim. Acta 2000,
83, 2192. Adlhart, C.; Volland, M. A. O.; Hofmann, P.; Chen, P. Helv.
Chim. Acta 2000, 83, 3306. Volland, M. A. O.; Adlhart, C.; Kiener, C.
A.; Chen, P.; Hofmann, P. Chem. Eur. J . 2001, 7, 4621.
With benzene-d6 instead of benzene as the collision
gas, we measure a kinetic isotope effect of 1.18 ( 0.0614
(11) Hinderling, C.; Plattner, D. A.; Chen, P. Angew. Chem. 1997,
109, 272. Hinderling, C.; Feichtinger, D.; Plattner, D. A.; Chen, P. J .
Am. Chem. Soc. 1997, 119, 10793.
(12) Becker, H. G. O.; Berger, W.; Domschke, G.; Fangha¨nel, E.;
Faust, J .; Fischer, M.; Gentz, F.; Gewald, K.; Gluch, R.; Mayer, R.;
Mu¨ller, K.; Pavel, D.; Schmidt, H.; Schollberg, K.; Schwetlick, K.; Seiler,
E.; Zeppenfeld, G. Organikum. Organisch-chemisches Praktikum;
Barth Verlaggesellschaft: Leipzig, 1993; p 662.
(13) Quantitative reaction efficiencies as a function of collision
energy and collision gas pressure were also recorded; over the range
of conditions examined, the isotope effects and product branching ratios
did not change appreciably, although the overall reaction efficiency,
i.e., yield, increased monotonically with increasing collision energy.

C-H Activation by [(N-N)Pt(CH3)(L)]+
Organometallics, Vol. 22, No. 11, 2003 2219
F igu r e 1. Daughter ion spectrum showing the reaction
performed by selection of 1a and collision of the ion with
benzene in the octopole collision cell.
F igu r e 2. Daughter ion spectrum taken by selecting the
1e/1f/2-d6 ion and performing collision-induced dissociation
of the ion with 0.8 mTorr 1,1,1,2-tetrafluoroethane at 7 eV
(laboratory frame) collision energy. An extraction of the
isotope distribution found in the methane and benzene-
loss products is compared with a fully statistical distribu-
tion.
for the overall transformation 1a f 1g using product
ion peak intensity under standardized conditions as a
measure of relative rate. In an independent experiment,
we looked at the reaction, benzene (or benzene-d6) +
1a f 1e/1f/2, and found no detectable isotope effect
within experimental error, indicating that the former
isotope effect pertains only to the step 1e/1f/2 f 1g.
Moreover, when 1e/1f/2, prepared as above and then
isolated by selection according to mass, is collided with
argon, one sees only the two products corresponding to
loss of methane or loss of benzene, in a 84:16 ((3)
ratio.15 When 1e/1f/2-d6, prepared by reaction of 1a with
benzene-d6 followed by selection according to mass, is
subjected to similar conditions, we find that the deute-
rium partitions with a slight preference into the depart-
ing benzene vis-a`-vis methane; precise integrations of
the isotopomer peak intensities show a slight deviation
from a statistical distribution, i.e., a small partition
isotope effect. The spectrum and the derived isotope
distribution in the products are shown in Figure 2.
The most important gas-phase experiment with re-
gard to the identity of the rate-determining step in C-H
activation of benzene is the comparison of the elimina-
tions of methane or benzene from preselected 1e/1f/2
induced by collision with either TFE or 1,1,1,2-tetra-
fluoroethane vapor. In the reaction induced by collision
with TFE, one sees neither a TFE-containing adduct nor
a TFE-containing product ion. We find that collision of
1e/1f/2 with TFE vapor or 1,1,1,2-tetrafluoroethane at
low energies gives ions corresponding to loss of methane
and benzene in a ratio of 83:17 ((2) and 85:15 ((1),
respectively, as the only detectable products. As in the
case of collision with argon, we find that the branching
ratio does not vary much with collision energy, at least
within the range in which we typically work (nominally
-2 to 14 eV in the laboratory frame). Very interestingly,
the collision energy dependence of the summed elimina-
tion of methane and benzene is experimentally indis-
F igu r e 3. Dependence of the combined yield of benzene
and methane-loss products in CID of 1e/1f/2 on the collision
energy varied in steps of 1 eV (laboratory frame, lines
connect the data points) for two different collision gases
at 0.60 mTorr pressure.
It should be noted that the absolute pressure of
collision gas, either TFE or tetrafluoroethane, was
reduced successively until the threshold behavior did
not change significantly, meaning that we begin to
approach the single-collision limit in the octopole colli-
sion cell. However, at the pressure of collision gas used,
0.60 mTorr, it cannot be guaranteed that effectively all
eliminations follow from a single collision event, al-
though it is certain that most of them do. Accordingly,
the extraction of a quantitative threshold, as has been
done for some other complexes, would be uninformative.
(16) In the comparison of elimination from the 1e/1f/2 complex, the
gauge pressures of TFE or tetrafluoroethane were varied between 0.8
and 2.0 mTorr as read from a Pirani gauge attached directly to the
housing of the second octopole. Pirani and thermocouple gauges
actually measure thermal conductivity rather than pressure. Because
thermal conductivity is proportional to the product of heat capacity
and pressure, the two gases should have the same actual pressure
when the gauge readings are normalized by the ratio of their heat
capacities. Computation of the heat capacities from scaled harmonic
tinguishable for the two different collision gases.16
A
superposition of the product yields as a function of
collision energy is shown in Figure 3.
(14) The error bound on the isotope effect is computed for 95%
certainty and comes from 30 independent measurements.
(15) The error bounds on the branching ratio is computed for 95%
certainty and comes from 50 independent measurements. The branch-
ing ratios were also constant over the range of collision energies
studied, i.e., nominally -2 to 14 eV in the laboratory frame.
vibrational frequencies was done and produced
a ratio Cv[TFE]/
Cv[tetrafluoroethane] of 1.13, which we used to normalize the gauge
readings.

2220 Organometallics, Vol. 22, No. 11, 2003
Gerdes and Chen
F igu r e 4. Stacked UV/vis absorption spectra showing the
time development of the spectrum of 1b and benzene at
25 °C in TFE solution containing 3.75 M benzene and 0.14
M water. Two clean isosbestic points are visible. Pseudo-
first-order rates were measured for decay of 1b and rise of
1h at 310, 400, and 545 nm. The displayed traces were
recorded at 5 min intervals.
F igu r e 5. Stacked UV/vis absorption spectra showing the
time development of the spectrum of 4b and benzene at
25 °C in TFE solution containing 1.88 M benzene and 0.17
M water. At least two clear isosbestic points are visible.
Pseudo-first-order rates were measured for decay of 4b and
rise of 4h at 375, 390, and 510 nm. The displayed traces
were recorded at 60 min intervals.
Nevertheless, the almost complete coincidence of the two
plots in Figure 3 suffices for the mechanistic argumen-
tation below.
Solution-phase kinetic studies of the reactions of 1b
and 4b with benzene in TFE solution were performed
primarily by time-dependent UV/vis, with control ex-
1
periments by H NMR and electrospray mass spectrom-
etry to check that the reactions were clean. For both
Pt(II) species, protonation of the corresponding dimethyl
complexes with HBF4 afforded the desired aquo com-
plexes, as expected from the work by Tilset, Bercaw, and
co-workers. With thorough purification of solvents and
reagents, extraneous side-reactions can be suppressed
as indicated by the clean isosbestic points visible in
Figures 4, 5, 8, and 9.
Kinetics were performed at multiple wavelengths and
found to be consistent. As a further control, the rates
of disappearance of the cationic methyl complexes 1b
and 4b were found to be identical with the rates of
appearance of the cationic phenyl complexes 1h and 4h .
Plots of absorbances A at the selected wavelengths
against time gave good fits to a single exponential over
several half-lives in each case, from which a pseudo-
first-order rate constant kobs could be extracted.
F igu r e 6. Dependence of the pseudo-first-order rate
constant of reaction 1b f 1h at 25 °C in TFE solution
containing benzene in concentrations between 0.43 and
3.75 M and water between 0.04 and 0.22 M. From the
slope of a linear fit, one derives a rate constant of k )
(4.93 ( 0.42) × 10-5 s-1
.
A ) A∞ + (A0 - A∞) exp(- kobst)
Figures 10 and 11 show the time development of the
absorbances for these two reactions, from which rate
constants can be extracted.
[benzene]
where
kobs ) k
[water]
Discu ssion
Plots of the kobs against [benzene]/[water], shown in
Figures 6 and 7, yield the rate constants for C-H
activation of benzene in wet TFE solution.
When neat (wet) benzene was used as solvent, the
rate of 1b f 1h and 4b f 4h could also be determined
by kinetic spectroscopy. Stacked UV/vis spectra over
time for 1b f 1h and 4b f 4h in neat benzene, i.e.,
11.25 M, with 0.029 M water are shown in Figures 8
and 9.
What are the intermediates and which elementary
step is rate determining for the C-H activation of
benzene by cationic Pt(II) complexes? Platinum alkyl
or aryl complexes have been protonated at low temper-
ature, providing chemical and NMR evidence4,5,17 for
hydridoalkyl and alkane or arene C-H σ- and π-com-
(17) Wik, B. J .; Lersch, M.; Tilset, M. J . Am. Chem. Soc. 2002, 124,
12116.

C-H Activation by [(N-N)Pt(CH3)(L)]+
Organometallics, Vol. 22, No. 11, 2003 2221
F igu r e 9. Stacked UV/vis absorption spectra showing the
time development of the spectrum of 4b and benzene at
25 °C in neat (11.25 M) benzene containing 0.029 M water.
Two clean isosbestic points are visible. Pseudo-first-order
rates were measured for decay of 4b and rise of 4h at 390
and 465 nm. The displayed traces were recorded at 20 s
intervals.
F igu r e 7. Dependence of the pseudo-first-order rate
constant of reaction 4b f 4h at 25 °C in TFE solution
containing benzene in concentrations between 0.57 and
3.75 M and water between 0.14 and 0.22 M. From the
slope of a linear fit, one derives a rate constant of k )
(9.24 ( 0.60) × 10-6 s-1
.
F igu r e 8. Stacked UV/vis absorption spectra showing the
time development of the spectrum 1b and benzene at 25
°C in neat (11.25 M) benzene containing 0.029 M water.
Three clean isosbestic points are visible. Pseudo-first-order
rates were measured for decay of 1b and rise of 1h at 330,
440, and 520 nm. The displayed traces were recorded at
20 s intervals.
F igu r e 10. Time dependence of the UV/vis absorbance at
three wavelengths for the reaction of 1b and benzene at
25 °C in neat (11.25 M) benzene containing 0.029 M water.
Data points were recorded at 0.5 s intervals and are
completely coincident with the fit function and therefore
not individually visible. A fit of data from all three
wavelengths to an exponential function produced a pseudo-
first-order rate constant, which when divided by [benzene]/
[water], yields a rate constant of k ) (6.61 ( 1.66) ×
plexes. Oxidative addition was also found to be favored
over the alternative σ-bond metathesis reaction in a
computational study.5 Moreover, the putative five-
coordinate Pt(IV) intermediate in the reaction of a
cationic Pt(II) alkyl complex with alkanes and benzene
has been intramolecularly trapped,18 lending support
to an oxidative addition mechanism.
10-5 s-1
.
While many of the major features of the overall C-H
activation reaction by Pt(II) have been worked out in
these previous studies, the claim that the reaction of
3b or 4b with methane or arenes in TFE solution is
unusually facile begs an explanation, especially in
connection to the identification of the rate-determining
step for the reaction. The long-term potential of C-H
activation makes the identification of those structural
or process parameters that confer high activity of
considerable interest, especially given the much harsher
reaction conditions that needed to be employed in the
original Shilov chemistry.
(18) Wick, D. D.; Goldberg, K. I. J . Am. Chem. Soc. 1997, 119, 10235.
Reinartz, S.; White, P. S.; Brookhart, M.; Templeton, J . L. J . Am. Chem.
Soc. 2001, 123, 12724. Reinartz, S.; White, P. S.; Brookhart, M.;
Templeton, J . L. Organometallics 2001, 20, 1709. Iron, M. A.; Lo, H.
C.; Martin, J . M. L.; Keinan, E. J . Am. Chem. Soc. 2002, 124, 7041.
Lo, H. C.; Haskel, A.; Kapon, M.; Keinan, E. J . Am. Chem. Soc. 2002,
124, 3226.

2222 Organometallics, Vol. 22, No. 11, 2003
Gerdes and Chen
Sch em e 2
Sch em e 3
F igu r e 11. Time dependence of the UV/vis absorbance at
390 nm for the reaction of 4b and benzene at 25 °C in neat
(11.25 M) benzene containing 0.029 M water. Data points
are indicated by circles and were recorded at 2 s intervals.
A fit to an exponential function produced a pseudo-first-
order rate, which when divided by [benzene]/[water], yields
formulated as proceeding by associative mechanisms.22
Isotopic scrambling,4 as well as equilibration between
various isomeric aryl complexes,6 in solution-phase
studies requires that interconversion between the
Pt(IV) intermediate VII, the two σ-complexes VI and
VIII, and presumably also the π-complex V be much
faster than elimination of benzene or methane, ruling
out either the actual oxidative addition or reductive
elimination for the rate-determining step. Therefore, it
has been recently proposed4 that the rate-determining
step in the C-H activation of arenes is the solvent-
assisted associative ligand exchange reaction III f V,
either as a single concerted step or as a combination of
two steps, one of which being actually rate determining.
Putting the structures III-V on a potential energy
diagram in Scheme 3, one can argue from the Hammond
postulate, applied to the two possible ligand dissocia-
tions from IV, that the net endothermic transformation
III f V means that the rate-determining transition
a rate constant of k ) (7.50 ( 3.80) × 10-6 s-1
.
The ligand exchange of methane for coordinated
solvent in 3d was shown to be associative by the extent
of isotopic incorporation in products from the reaction
of [(N-N)Pt(CH3)2] with DOTf in TFE in the presence
of variable amounts of acetonitrile.19 This result argues
convincingly against the intermediacy of a three-
coordinate intermediate in the solution-phase reaction,
i.e., against a dissociative mechanism.
Previous studies on square-planar d8 complexes20 in
general, and Pt(II) complexes21 in particular, have long
provided strong evidence that substitution reactions
generally proceed via an associative mechanism. Ac-
cordingly, the mechanism in ref 4 can be drawn as
indicated in Scheme 2, in which the various ligand
substitutions connecting I-V as well as IX-XII are
(19) J ohansson, L.; Tilset, M. J . Am. Chem. Soc. 2001, 123, 739.
(20) Basolo, F.; Pearson, R. G. Prog. Inorg. Chem. 1964, 4, 381.
Basolo, F.; Pearson, R. G. Mechanisms of Inorganic Reactions; Wiley:
New York, 1968.
(21) Romeo, R.; Scolaro, L. M.; Nastasi, N.; Arena, G. Inorg. Chem.
1996, 35, 5087.
(22) In Scheme 2, the structures should not be understood to mean
that we have assigned
a particular stereochemistry to the five-
coordinate intermediates. The structures are presumed to undergo
facile pseudorotation. The present experiments make no prediction
concerning the stereochemistry.

C-H Activation by [(N-N)Pt(CH3)(L)]+
Organometallics, Vol. 22, No. 11, 2003 2223
state must be the second step, i.e., the departure of TFE
from the five-coordinate intermediate.
These experiments are important to confirm that the
solution-phase studies have been properly performed.
This proposition is directly tested in the present
experimental report.
Having established that 1b behaves like 4b, we turn
first to the gas-phase study. While solvation leads to
large changes in the potential surfaces for gas-phase
ion-molecule reactions, such as proton exchange, SN2,
or carbonyl additions,26 relative to their congeners in
solution, reactions of organometallic complexes electro-
sprayed into the gas phase10 have proven surprisingly
similar to the corresponding reactions in solution in
other cases we have examined,10,11 leading us to pre-
sume that the analogy in this case is also valid. Even
so, solution-phase experiments in this report were
executed to confirm the expectation (vide infra). Four
observations may be made from the ion-molecule
reactions. First, we see the same gross reaction in
solution and in the gas phase. Second, the overall kinetic
isotope effects for the reaction in the two different
settings are similar; kH/kD ) 1.18 ( 0.06 in the gas-
phase versus 1.06 ( 0.05 in solution.4 Third, when
benzene-d6 is used instead of benzene, the distribution
of deuterium between methane and benzene products
shows nearly full equilibration and, furthermore, shows
a small partition isotope effect in the same direction
with a similar magnitude as that reported for the
solution-phase experiment. Last, in gas-phase experi-
ments, 1e/1f/2 gives branching ratios close to the
reported value of 82:18 for reductive elimination of
methane versus benzene observed when [(N-N)Pt-
(CH3)(C6H5)] is protonated in solution.4 It might be
argued that the reactions of 1a , a highly unsaturated
three-coordinate species, are not directly comparable to
those of four-coordinate 3c or 4c, or even to those of 1c,
for that matter, and that the gas-phase chemistry makes
therefore no relevant prediction for solution-phase
mechanisms.27 The argument fails because one notes
that isotope partitioning as well as the product branch-
ing ratio is determined not by the addition steps starting
from 1a , but rather by the reactions after formation of
1e/1f/2. The presumably equilibrating mixture 1e/1f/2
in the gas-phase represents exactly those species de-
picted as V, VI, VII, and VIII in solution (with the
proviso that one cannot yet distinguish between π- and
σ-complexes in the gas phase). These four observations
serve to show that the gas-phase chemistry does not
differ grossly from that in solution. In particular, the
significance of the addition reaction of 1a should be
understood as primarily a gas-phase synthesis of 1e/
While the gas-phase reaction studies formally involve
the species very similar to those postulated in solution,
the relevance of the present work to C-H activation in
solution requires that we first establish that the com-
plexes 1 are good models for the complexes 3 or 4. We
chose the complexes 1 for the gas-phase part of this
study because we had found in earlier work that
coordinatively and electronically unsaturated complexes
often tended to undergo intramolecular C-H activation
reactions, i.e., orthometalation, on a ligand.11 For com-
plexes 3 and 4 that were used in previous solution-phase
studies, orthometalation is a possible reaction pathway
which could complicate the gas-phase experiment that
would interfere because the isolated ion has a compara-
tively long time to react intramolecularly in the source
region of the mass spectrometer as compared to the
situation in solution. We therefore employed in this
study a ligand in which all ortho positions are blocked
by chlorine substitution, which made the unsaturated
complex stable against unimolecular reactions in the gas
phase. It should be noted that the aryl rings on the
ligands of 1, 3 and 4 are twisted out-of-plane,23 so the
electronic influence of substituents on the rings at the
metal center would come only through inductive effects.
In this light, the nearly identical values of σI for chloro
and trifluoromethyl substituents24 would suggest that
1b and 3b would display similar reactivity.25 The
solution-phase kinetics of 3b and 4b were reported to
be similar as well, with rates differing by less than 1
order of magnitude.7 It would not be unreasonable to
expect therefore broadly similar chemistry for 1b versus
3b or 4b. The expectation is fully borne out by the
solution-phase kinetics as illustrated in Figures 4-11.
Not only is the chemical behavior for complex 1b
essentially identical to that for 4b, but even the absolute
values of the rates are not too different, e.g., k ) (4.93
( 0.42) × 10-5 s-1 versus (9.24 ( 0.60) × 10-6 s-1 for
1b and 4b, respectively. Given that very substantial
data have been published for 3b and 4b, it is gratifying
to see that the rate constant measured in the present
work for 4b is in acceptable agreement with the value
of (1.98 ( 0.03) × 10-5 s-1 reported for the same complex
under the same conditions by Tilset, Bercaw, and co-
workers.4 Moreover, the clean isosbestic points in
Figures 4, 5, 8, and 9 indicate that it is possible to find
conditions under which all other side-reactions are much
slower than the transformation 1b f 1h or 4b f 4h .
(26) Wolf, J . F.; Harch, P. G.; Taft, R. W. J . Am. Chem. Soc. 1975,
97, 2904. Farneth, W. E.; Brauman, J . I. J . Am. Chem. Soc. 1976, 98,
7891. Aue, D. H.; Webb, H. M.; Bowers, M. T. J . Am. Chem. Soc. 1976,
98, 311. Olmstead, W. N.; Brauman, J . I. J . Am. Chem. Soc. 1977, 99,
4219. Staley, R. H.; Wieting, R. D.; Beauchamp, J . L. J . Am. Chem.
Soc. 1977, 99, 5964. Asubiojo, O. I.; Brauman, J . I. J . Am. Chem. Soc.
1979, 101, 3715. Pellerite, M. J .; Brauman, J . I. J . Am. Chem. Soc.
1980, 102, 5993. Arnett, E. M.; Peinta, N. J . J . Am. Chem. Soc. 1980,
102, 3329. Sharma, S.; Kebarle, P. J . Am. Chem. Soc. 1982, 104, 19.
Bohme, D. K.; MacKay, G. I. J . Am. Chem. Soc. 1981, 103, 978.
Caldwell, G.; Magnera, T. F.; Kebarle, P. J . Am. Chem. Soc. 1984, 106,
959. McMahon, T. B.; Heinis, T.; Nicol, G.; Hovey, J . K.; Kebarle, P. J .
Am. Chem. Soc. 1988, 110, 7591.
(23) The X-ray structure of 1 is given in the Supporting Information;
for related complexes, the X-ray structure was given in ref 6.
(24) Hines, J . Structural Effects on Equilibria in Organic Chemistry;
Wiley: New York, 1975.
(25) The Oslo and Caltech groups have remarked in ref 4 that [(N-
N)Pt(CH3)(L)]+, (N-N) ) Ar-NdC(CH3)-C(CH3)dN-Ar with Ar )
o,o′-dimethylphenyl, reacts similarly to
3 in TFE solution, only
somewhat more slowly and more cleanly. Moreover, ref 6 compares
the reactions of the complexes with fluorinated and unfluorinated
ligands in their reaction with toluene. While the complex with
unfluorinated ligand reacted more slowly, and moreover, produced a
product of benzylic C-H activation that was absent in the reactions
of 3, the ratio of ortho to meta to para C-H activation for the two
complexes was essentially identical. These observations suggest that,
in the reaction of [(N-N)Pt(CH3)(L)]+ with benzene, 1 and 3 should
not be expected to show gross mechanistic differences.
(27) Given the recent report of a crystal structure for a d8 ML3 Pt(II)
complex stabilized by an agostic interaction at the formally empty
fourth coordination site, one cannot lightly discount three-coordinate
species on the reaction coordinate. See: Baratta, W.; Stoccoro, S.;
Doppiu, A.; Herdtweck, E.; Zucca, A.; Rigo, P. Angew. Chem., Int. Ed.
2003, 42, 105.

2224 Organometallics, Vol. 22, No. 11, 2003
Gerdes and Chen
1f/2, whose subsequent reactions are shown below to
make a relevant prediction for solution-phase mecha-
nisms.
The solution-phase kinetics for 1b and 4b fully
support the conclusion of the gas-phase study. Exami-
nation of Figures 6 and 7 reveals the transformations
1b f 1h and 4b f 4h are each described by a single
rate constant over a broad range of [benzene]/[water]
in wet TFE solution. If one makes the decisive compari-
son of the derived rate constant for benzene activation
in TFE solution to the corresponding rate constants in
neat (wet) benzene, one finds that, for each of the
complexes 1b or 4b, the rate constant shows no depen-
dence on TFE concentration at all. We find, for the
reaction 1b f 1h , k ) (4.93 ( 0.42) × 10-5 s-1 in TFE
solution, and (6.61 ( 1.66) × 10-5 s-1 with no TFE at
all. For 4b f 4h , the corresponding rate constants are
A rate-determining transition state with TFE bound
at the metal is ruled out by the present gas-phase
experiments: the threshold behaviors in Figure 3 for
reductive elimination of benzene and methane from 1e/
1f/2 induced by low-energy collision of the ion with TFE
or 1,1,1,2-tetrafluoroethane are identical. The collision
of preselected 1e/1f/2 with TFE corresponds to the
reaction of V with TFE because both the gas-phase and
solution-phase experiments show rapid equilibration of
V-VIII prior to any other reaction. The two collision
gases, TFE and 1,1,1,2-tetrafluoroethane, are isoelec-
tronic and isostructural and have almost the same mass,
making the kinematics of the collisions similar. They
have moreover almost identical dipole moments,28 mak-
ing even long-range electrostatic ion-molecule interac-
tions very similar. They do differ, however, in that
1,1,1,2-tetrafluoroethane coordinates much more weakly
to electrophilic centers than does TFE, if at all.29
Moreover, the near-constant branching ratio in the
elimination reactions of 1e/1f/2 as the collision gas is
varied from argon to TFE to tetrafluoroethane suggests
strongly that TFE participates in the elimination only
as a vehicle for energy transfer to 1e/1f/2. If coordina-
tion of TFE were important at the transition state, one
would expect to see different thresholds, and perhaps
different selectivities, for elimination of benzene or
methane from 1e/1f/2 when one changes the collision
gas. By microscopic reversibility, the reaction in the
other direction must also have no TFE bound at the
transition state. Therefore, a putative solvent-assisted
associative mechanism cannot be the explanation for the
facile C-H activation reaction. The isotopic scrambling
in the gas-phase experiment matches that seen in the
analogous reaction in solution, which requires as a
minimum conclusion rapid reversible interconversion
between VI, VII, VIII, and presumably V, just as has
been concluded for the solution-phase reaction.4 This
information, combined with the experimental conclusion
that TFE is not bound at the rate-determining transi-
tion state for either benzene or methane elimination,
requires that there exists at least one kinetically
significant intermediate between IV and V for benzene
activation and at least one further intermediate between
VIII and IX for methane activation. The rate-determin-
ing transition state for C-H activation of benzene by
this class of cationic Pt(II) complexes is necessarily the
transition state in which this intermediate is trans-
formed into V. Similarly, the rate-determining transi-
tion state for C-H activation of methane is the transi-
tion state in which the intermediate is transformed into
VIII. The experiment does not specify what the struc-
ture of either of the new intermediates is, but their
existence, and their crucial position immediately pre-
ceding the rate-determining step for the overall C-H
activation reaction, is mandatory.
k ) (9.24 ( 0.60) × 10-6 and (7.50 ( 3.80) × 10-6 s-1
.
If the rate constant in the complete absence of TFE is
experimentally indistinguishable from that when TFE
is present as the solvent, i.e., the reaction is zero-order
in TFE, it must be concluded that TFE cannot be
involved in the rate-determining step for C-H activa-
tion. Arguments that complex 1b is somehow unique
in this regard are countered by the corresponding
experiments with complex 4b, which displays identical
behavior. One concludes that TFE is nothing more than
an inert diluent.
Both the gas-phase and solution-phase experiments
compel the existence of at least one additional interme-
diate immediately prior to the rate-determining step in
the C-H activation reactions of either benzene or
methane. While the experiment makes no suggestion
as to the structure, one may speculate with the proviso
that any such structural speculation needs to be vali-
dated by further experiment or by computation. Among
the candidates for the intermediate in the case of
methane activation is the η2-(H,H) σ-complex, found by
computations5,30 to be another possible binding mode
for methane complexes31 in addition to the more ordinary-
looking η2-(H,C) methane σ-complex. Another possibility
for either methane or benzene activation would be a
four-coordinate structure in which the fourth site on the
cationic Pt(II) center is occupied by the ortho substituent
on the aryl group of the Schiff base. Coordination of
methane or benzene to form a five-coordinate interme-
diate would then be followed by rate-determining de-
parture of the ortho substituent. Variation of substitu-
tion on this position can potentially serve as a handle
for tuning of reactivity in a rational way. This attractive
possibility must wait, though, until the structural
speculations are confirmed by further work.
Con clu sion s
Gas-phase ion-molecule reactions of [(N-N)Pt(CH3)-
(L)]+ with benzene show a remarkable similarity to the
corresponding solution-phase reactions. We accordingly
take advantage of electrospray ionization tandem mass
spectrometry to isolate postulated intermediates and
investigate the mechanism of C-H activation in the
reaction of [(N-N)Pt(CH3)(L)]+ with benzene. We rule
out a rate-determining transition state in which TFE
(28) From PM3 calculations on optimized geometries, one finds for
trifluoroethanol and 1,1,1,2-tetrafluoroethane dipole moments of 2.2
and 2.3 D, respectively.
(29) It should be noted that DFT calculations on model complexes
in ref 5 find no evidence that TFE coordinates to the metal center in
four-coordinate Pt(II) complexes as a fifth ligand.
(30) Hill, G. S.; Puddephatt, R. J . Organometallics 1998, 17, 1478.
(31) Transition metal alkane complexes have been reviewed: Hall,
C.; Perutz, R. Chem. Rev. 1996, 96, 3125.

C-H Activation by [(N-N)Pt(CH3)(L)]+
Organometallics, Vol. 22, No. 11, 2003 2225
solvent is coordinated at the metal and then confirm
the result by solution-phase kinetic studies. Given the
prior report that benzene activation in TFE solvent by
these complexes constitutes the most facile C-H activa-
tion reactions yet reported for cationic Pt(II) complexes,
it is perhaps disappointing to find that TFE gives no
acceleration at all. On the other hand, it should be
encouraging to find that C-H activation may be achieved
in ordinary solvents in seconds at room temperature as
long as the solvent is not strongly coordinating. We plan
further studies to clarify the structural basis for this
high reactivity.
the Research Commission of the ETH Zu¨rich. In addi-
tion, one of us (G.G.) acknowledges a Kekule´-Stipend
from the Fonds der Chemischen Industrie, Germany.
Valuable discussions at the outset of this project with
Dr. Christian Hinderling are acknowledged. Assistance
with UV/vis measurements by Dr. Erich Meister as well
as scale-up of the syntheses by Andre´ Mu¨ller is also
acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Synthesis, charac-
terization, and spectroscopic data for new compounds, experi-
mental details for mass spectrometry, UV/Vis spectroscopy,
and X-ray crystallography are available free of charge via the
Internet at http://pubs.acs.org.
Ack n ow led gm en t. This project was supported fi-
nancially by the Swiss National Science Foundation and
OM0209982