Organometallics
Article
All isotopologues of 3 were synthesized from the corresponding
deuterium-labeled 5-bromo-4,4-dimethyl-1-pentenes. Because these
syntheses were conducted on small scales, two modifications of the
standard preparations were made in order to obtain reasonable yields.
First, the deuterated alcohols were not prepared by direct
bromination from deuterated 2,2-dimethylpent-4-en-1-ol, but instead
were first converted to the corresponding tosylates and then
converted to the bromide with NaBr/HMPA. Second, in order to
minimize adventitious oxidation or hydrolysis, lithium reagents were
used instead of Grignard reagents to synthesize 3.78 The syntheses of
the isotopically labeled compounds are described in detail in the
under static CVD conditions, the extra hydrogen comes from
the dehydrogenation of pentenyl ligands on the platinum
surface.
In hydrocarbon solvents such as C6D6, the mechanism of
C−H bond activation involves three pathways, operating in a
simultaneous manner and in parallel. The first step is
decomplexation of the CC bond of one of the ligands of
3 to generate a σ complex in which an allylic γ-C−H bond or
an olefinic δ- or ε-C−H bond coordinates to Pt. Subsequently,
C−H activation of one of these C−H bonds occurs to form a
short-lived PtIV hydride, which reductively eliminates ∼1 equiv
of 4,4-dimethylpentenes and generates an insoluble black
precipitate that contains most of the dehydrogenated pentenyl
ligands.
Thermolysis reactions were conducted in a Thermo Scientific
Lindberg/Blue M Mini-Mite Tube Furnace or (for kinetic experi-
ments requiring better temperature control) in a Fisher Scientific
Isotemp 1006S refrigerated circulating bath with mineral oil as a heat
1
2
transfer medium. The 1D H, H, and 13C NMR data were recorded
on a Varian Inova 400 spectrometer at 9.39 T, a Varian Inova 500
spectrometer at 11.74 T, a Varian Inova 600 spectrometer at 14.09 T,
or a Bruker B500 NMR spectrometer at 11.74 T. Chemical shifts are
reported in δ units (positive shifts to higher frequency) relative to
TMS (1H, 13C), set by assigning appropriate shifts to residual solvent
signals. GC-MS spectra were collected by the staff of the Roy J. Carver
Biotechnology Center at the University of Illinois on a GC/MS
system (Agilent Inc., Palo Alto, CA, USA) consisting of an Agilent
7890 gas chromatograph, an Agilent 5975 mass-selective detector, and
a HP 7683B autosampler. Peaks were evaluated using the AMDIS
2.71program (NIST, Gaithersburg, MD, USA) and identified with the
aid of the libraries NIST08 (NIST, MD, USA) and W8N08 (Palisade
Corporation, NY, USA).
Thermolysis of 3 in C6F6, C6D12, C6D6, and C6H6 Solutions. In
order to suppress secondary reactions with colloidal platinum
species,43 we added a drop of mercury to all solutions in which
thermolysis studies were conducted. We confirmed by GC-MS that
dialkylmercury compounds were not formed (Figure S3.4). The
addition of Hg inhibits H−D exchange between byproducts and
solvents. A solution of hexamethyldisiloxane or 1,2,4,5-tetrachlor-
obenzene dissolved in C6D6 and sealed in a glass capillary was used as
an integration standard. The capillary and the mercury droplet were
loaded into an NMR tube, which was evacuated for 5 h at 10 mTorr
before addition of analyte.
A sample of 3 (typically 20 mg, 0.051 mmol) was cooled to 0 °C
and evacuated for 5 h (10 mTorr) to ensure that it was free of residual
solvents. The dried sample was dissolved in the desired solvent (C6D6
or C6F6; 650 μL), and the resulting solution was transferred to the
NMR tube containing the capillary and a drop of mercury. The NMR
tube was then degassed by the freeze−pump−thaw technique and
flame-sealed under vacuum. The tube was completely immersed in a
circulating oil bath preheated to the desired temperature, which could
be controlled to within 1 °C. Multiple measurements at 110 1 °C
in the circulating oil bath show that the measured rate constant has a
standard deviation of ∼10%. A few nonquantitative thermolysis
experiments were conducted by placing the NMR entirely within the
hot zone of a tube furnace whose two ends were blocked with glass
wool.
At certain time intervals, the NMR tube was removed, cooled to 20
°C to quench the reaction, and examined by NMR spectroscopy. The
progress of the reaction was monitored by measuring the peak
integrals; the disappearance of 3 due to thermolysis was best followed
by integration of the olefinic peaks. The concentration of 3 at t = 0 is
often somewhat anomalous, possibly due to slow redissolution of the
integration standard after the freeze−pump−thaw step. As a result,
this data point was omitted from the curve-fitting analyses.
Thermolyses were conducted for no more than 2 half-lives to
minimize any inhibition effect from byproducts.
For activation of the γ- and ε-C−H bonds, the RDS is
formation of the σ complex. In contrast, for activation of the δ-
C−H bonds the RDS is intramolecular C−H bond oxidative
cleavage (or formation) in C6H6 but formation of the σ
complex in C6F6, as shown by a series of studies involving
deuterated isotopologues of 3. We attribute the change of
mechanism to the stronger coordinating ability of C6F6, which
coordinates to Pt and hinders the formation of the C−H σ
complex, whereas C6H6 does not participate significantly in the
decomposition before or during the RDS (but does become
slightly involved afterward). Thus, the thermolysis pathway in
benzene reasonably models the mechanism by which 3
decomposes under CVD conditions on an inert surface (i.e.,
unimolecularly).
Although there are dangers associated with extrapolating
results in solution to processes (such as CVD) that occur in
the absence of solvent, under the right circumstances the
chemistry of organometallic molecules in solution can afford
an important insight into the chemistry behind CVD
processes.77 Here, the hydrocarbon solvents we have used
play essentially no role in the chemistry, and it seems
reasonable to use the thermolysis studies as a basis for
formulating hypotheses about why some precursors give
smooth Pt films and others do not. For precursors such as
(C5H4Me)PtMe3, which have no low-energy pathway to
dissociate a ligand, the formation of nuclei is slow but the
growth on Pt (which can catalyze the loss of ligands) is fast;
these circumstances lead to a low density of nuclei and the
formation of rough films. For compound 3, however, a CC
bond can readily decomplex to form intermediates that can
readily convert to growth species, so that the formation of
nuclei is fast on all surfaces; in addition, the dehydrogenated
pentenyl ligands poison the Pt surface.30 As a result, the Pt-
containing films grown from 3 are unusually smooth.
Detailed studies of the use of 3 as a CVD precursor are
described in a separate paper.30
EXPERIMENTAL SECTION
■
All manipulations were carried out under vacuum or under argon
using standard Schlenk techniques unless otherwise specified. All
glassware was oven-dried before use. Hexamethyldisiloxane (Petrarch
Systems) and perfluorotoluene (Sigma-Aldrich) were used as
received, 1,2,4,5-tetrachlorobenzene (Eastman Chemical) was recrys-
tallized from diethyl ether, and hexafluorobenzene (Strem Chemicals)
was dried over 3 Å molecular sieves. Benzene-d6 (Cambridge Isotope
Laboratories) and benzene (EMD Millipore) were distilled under
nitrogen from sodium/benzophenone, and cyclohexane-d12 was
purchased from Cambridge Isotope Laboratories in 1 mL ampules
and used without purification.
To minimize experimental error, experiments were conducted as
much as possible in a side by side fashion, with the parameter of
interest as the only variable. For example, for deuterium labeling and
kinetic isotope studies, deuterated and undeuterated samples of 3
(typically 20 mg, 0.05 mmol) were dissolved separately in the desired
J
Organometallics XXXX, XXX, XXX−XXX
Liu, Sumeng
