A glimpse at the chemistry of GeH2 in solution. Direct detection of an intramolecular germylene-alkene π-complex

Article abstract of DOI:10.1021/ja201190b

The photochemistry of 3-methyl-4-phenyl-1-germacyclopent-3-ene (4) and a deuterium-labeled derivative (4-d₂) has been studied in solution by steady state and laser flash photolysis methods, with the goal of detecting the parent germylene (GeH

Full text of DOI:10.1021/ja201190b

ARTICLE
pubs.acs.org/JACS
A Glimpse at the Chemistry of GeH2 in Solution. Direct Detection of an
Intramolecular GermyleneꢀAlkene π-Complex
Paul S. Billone,† Katie Beleznay,‡ Cameron R. Harrington,§ Lawrence A. Huck,|| and William J. Leigh*
Department of Chemistry and Chemical Biology, McMaster University, 1280 Main Street West, Hamilton, ON Canada L8S 4M1
S Supporting Information
b
ABSTRACT: The photochemistry of 3-methyl-4-phenyl-1-
germacyclopent-3-ene (4) and a deuterium-labeled derivative
(4-d2) has been studied in solution by steady state and laser flash
photolysis methods, with the goal of detecting the parent
germylene (GeH2) directly and studying its reactivity in solu-
tion. Photolysis of 4 in C6D12 containing acetic acid (AcOH) or
methanol (MeOH) affords 2-methyl-3-phenyl-1,3-butadiene
(6) and the OꢀH insertion products ROGeH3 (R = Me or Ac)
in yields of ca. 60% and 15ꢀ30%, respectively, along with
numerous minor products which the deuterium-labeling studies suggest are mainly derived from hydrogermylation processes
involving GeH2 and diene 6. The reaction with AcOH also affords H2 in ca. 20% yield, while HD is obtained from 4-d2 under similar
conditions. Photolysis of 4 in THF-d8 containing AcOH affords AcOGeH3 and 6 exclusively, indicating that the nucleophilic solvent
assists the extrusion of GeH2 from 4 and alters the mechanism of the trapping reaction with AcOH compared to that in cyclohexane.
Laser flash photolysis of 4 in hexanes yields a promptly formed transient exhibiting λmax ≈ 460 nm, which decays on the
microsecond time scale with the concomitant growth of a second, much longer-lived transient exhibiting λmax ≈ 390 nm. The
spectrum and reactivity of the 460 nm species toward various germylene trapping agents are inconsistent with those expected for free
GeH2; rather, the transient is assigned to an intramolecular Ge(II)ꢀalkene π-complex of one of the isomeric substituted
hydridogermylenes derived from a solvent-cage reaction between GeH2 and its diene (6) coproduct, formed by addition of HGeꢀH
across one of the CdC bonds. These conclusions are supported by the results of DFT calculations of the thermochemistry
associated with π-complexation of GeH2 with 6 and the formation of the isomeric vinylgermiranes and 1,2-hydrogermylation
products. A different species is observed upon laser photolysis of 4 in THF solution and is assigned to the GeH2ꢀTHF complex on
the basis of its UVꢀvis spectrum and rate constants for its reaction with AcOH and AcOD.
’ INTRODUCTION
the one associated with rearrangement of the intermediate
complex, via a process that can be thought of most simply as
involving attack of the germanium lone pair back into an electro-
philic site in the complex.6 This common pattern of reactivity
underlies the recent successful syntheses by Rivard and co-
workers of stable donor/acceptor-coordinated complexes of
GeH2.22 other, similarly-stabilized complexes of transient diva-
lent Group 14 compounds have also been reported recently.22b,23
One expected reaction of the species that has not yet been
observed is dimerization to form digermene (Ge2H4), which
theory indicates should be as much as ca. 50 kcal molꢀ1
exothermic.24ꢀ28 Digermene (H2GedGeH2) and its various iso-
mers (germylgermylene (HGeGeH3) and the cis and trans hydro-
gen-bridged GeH2 dimers (HGeH2GeH)) have been the
subject of numerous theoretical studies, which predict the
doubly bonded and germylgermylene isomers to be the most
stable isomers, approximately isoenergetic,26,29ꢀ32 and sepa-
Germylene (GeH2), the parent GeII molecule and a key reactive
intermediate in the chemical vapor deposition of germane and
other germanium hydrides,1ꢀ4 has received considerable attention
from spectroscopists, kineticists, and theorists and indeed has
been more extensively studied than any other GeII species,
transient or stable.5,6 Like methylene (CH2) and silylene
(SiH2), GeH2 bears the distinction of being the most reactive
derivative of its class, according to the extensive kinetic studies
that have been carried out on it and other simple germylene
derivatives in the gas phase as well as in solid matrixes at very low
temperatures.7 In the gas phase, it reacts with alkenes,8ꢀ11 dienes,8
alkynes,8,12ꢀ14 hydridosilanes8,15,16 and -germanes,17ꢀ20 water,21
and alcohols21 at rates approaching the collisional limit, generally
several orders of magnitude faster than its simplest dialkyl con-
gener, dimethylgermylene (GeMe2).5,6 Most of these reactions
proceed with negative activation energies, which has been rationa-
lized in terms of the initial, reversible formation of a pre-reaction
complex in which association of the substrate with GeH2 occurs
via interaction with the empty 4p orbital.5,6 This process is
barrierless for most substrates and at least moderately exother-
mic, with the rate-limiting barrier for product formation being
rated by a modest (ca. 12 kcal molꢀ1 29,30
enthalpic barrier.
)
Both digermene32,33 and germylgermylene32 have been success-
32
fully characterized by infrared spectroscopy in H233 or GeH4
Received: February 16, 2011
Published: June 03, 2011
r
2011 American Chemical Society
10523
dx.doi.org/10.1021/ja201190b J. Am. Chem. Soc. 2011, 133, 10523–10534
|

Journal of the American Chemical Society
ARTICLE
matrixes at 8ꢀ12 K but to our knowledge have not yet been
characterized in the gas phase or in solution.
transient products observed in laser photolysis experiments with
4 and many of our mechanistic conclusions are supported by the
results of DFT calculations, which have been carried out at the
PW91PW9147/6-311+G(2d,p) level of theory.
Recent work from our laboratory has focused on the genera-
tion and direct detection of simple transient germylene deriva-
tives such as dimethyl-,34ꢀ38 diphenyl-,34,36,37,39ꢀ41 and methyl-
phenylgermylene42 (GeMe2, GePh2, and GeMePh, respectively)
in solution and on detailing the kinetics and mechanisms of their
reactions with various representative germylene substrates. In the
case of GeMe2, whose reactivity has been studied both in the gas
phase43 and in hexane solution35 with a number of common or
closely analogous substrates, there is good agreement between the
UVꢀvis spectra and absolute rate constants measured under the
two sets of conditions.6 The main goal of the present work was to
see whether we could extend these comparisons to the parent
germylene, GeH2; its UVꢀvis spectrum in the gas phase, mea-
sured by laser-induced fluorescence excitation spectroscopy, is
centered at λmax ≈ 515 nm.44
’ RESULTS AND DISCUSSION
Compounds 4 and 4-d2 were synthesized by reduction of the
corresponding dichloride (5; eq 1)35 with LiAlH4 and LiAlD4,
respectively, following the procedure reported for the synthesis
of 2.8
Photochemical Product Studies. Steady state photolysis
experiments were carried out in quartz NMR tubes, using de-
oxygenated solutions of 4 (0.025 M) in cyclohexane-d12 or THF-d8
containing 0.05ꢀ0.25 M acetic acid (AcOH), acetic acid-O-d
(AcOD), or methanol (MeOH) as germylene trapping agents.
The solutions were irradiated with 254 nm light and were
Gas-phase kinetic studies of GeH2 have employed phenylger-
mane (1a), mesitylgermane (1b), or 3,4-dimethylgermacyclo-
pent-3-ene (2) as photochemical precursors, monitoring the
strong rovibrational transitions at 17 111.31 or 17 118.67 cmꢀ1
(∼588 nm) in the electronic spectrum of the 74GeH2 isotopomer
for the measurement of decay kinetics, and an ArF excimer laser
(193 nm) for precursor excitation.6,8,9,15,16,18,19,21,45 Phenylger-
mane (1a) has also been shown to extrude GeH2 upon 248 nm
excitation in the gas phase.18 We have shown that phenylated
germacyclopent-3-ene derivatives such as 3a,b are efficient 248 nm
precursors for GeMe2 in solution35 and exhibit sufficiently strong
absorptions at 248 nm that it is possible to work with precursor
concentrations on the order of 10ꢀ4 M or less in laser photolysis
experiments.35 It thus seemed reasonable to expect that the
dihydro-analogue of 3b, 3-methyl-4-phenylgermacyclopent-3-ene
(4), might be a similarly efficient photochemical precursor to
GeH2, thus potentially providing the means to generate the parent
germylene in solution under conditions where possible compli-
cations due to undesired reactions with its precursor might be
minimized. The latter seemed a valid concern since GeH2 is
known to react with hydridogermanes at close to the collisional
limit in the gas phase, and the rate varies only modestly
with substitution.17ꢀ21,46 The CdC bond in 4 provides a second
potential site of undesired precursor reactivity; while less is known
in regard to substituent effects on the rate constant than is the case
with GeꢀH insertions, GeH2 reacts with both ethene9 and
propene10 at close to the collisional rate in the gas phase, via a
complex series of processes involving (in the case of ethene) both
germirane and ethylgermylene as intermediates.6,9,11 In any
event, we expected the styrenyl chromophore in 4 to provide a
sufficiently high absorption cross-section at the (248 nm KrF
laser) excitation wavelength that precursor concentrations could
be kept low and thus minimize complications of this type.
1
monitored by H NMR spectroscopy at selected time intervals
between 0 and ca. 40% conversion of 4. Product yields were
determined from the relative slopes of concentration vs time plots
covering the rangeof0 to 6ꢀ10%conversion of4. Photolysis of the
compound in the absence of a trapping agent afforded only
polymeric material, and no small-molecule products of any type
could be detected in spite of efficient consumption of the precursor.
Photolysis of 4 in C6D12 containing AcOH or MeOH afforded
mixtures of several products in both cases (Figures S1, S2, Support-
ing Information). The products that could be identified were
2-methyl-3-phenyl-1,3-butadiene (6),35 molecular hydrogen,
and the oxygermanes 7 and 8, respectively (eq 2; Figure S3a,
Supporting Information), which were identified by comparison
1
to the reported H NMR spectra and the results of deuterium-
labeling studies (vide infra).48,49 The yield of H2 formed in the
presence of AcOH was corrected for the 3:1 ratio of ortho-H2 to
para-H2 expected at room temperature (the latter is NMR
inactive).50
The identities of 8 and H2 in the photolysis in the presence of
AcOH were confirmed by deuterium-labeling experiments, in
which 4 and 4-d2 were photolyzed in C6D12 containing AcOD
and AcOH, respectively, under similar conditions to those used
in the experiment with the unlabeled compounds (see Figure 1
and Figures S4, Supporting Information). In both cases, the
1
singlet at δ 4.54 in the H NMR spectra of the photolyzed
4/AcOH mixture was replaced by a 1:1:1 triplet centered at δ
4.51, which could be definitively assigned to HD on the basis of
its characteristic coupling constant, JHD = 42.68 Hz.51 These
results indicate that one of the H-atoms in the H2 that is
produced originates from AcOH, while the other originates
from 4, presumably via GeH2. A second prominent difference
We have thus synthesized 4 and examined its photochemistry
in solution by steady state and laser flash photolysis methods. To
assist in the identification of GeH2-derived products in steady
state trapping studies, we have also studied the 1,1-dideuterated
isotopomer (4-d2). The structural assignments for the primary
10524
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
when 4 was photolyzed in the presence of a lower (0.05 M) initial
concentration of AcOH (Figure S5, Supporting Information).
While the deuterium-labeling experiments indicated that it too is
derived from a hydrogermylation process, little more structural
information than that could be gleaned from the spectra.
In contrast to the results obtained in C6D12 solution, 6 and 8
were the only products detected upon photolysis of a 0.025 M
solution of 4 in THF-d8 containing 0.25 M AcOH (see eq 3 and
Figure S6, Supporting Information). Concentration vs time plots
showed that the rate of formation of 6 matched that of consump-
tion of 4 almost precisely, indicating a nearly quantitative yield of
the diene (Figure S3b, Supporting Information). Very different
results were obtained when 4 was photolyzed in THF-d8 contain-
ing 0.25 M MeOH; the solution developed an intense yellow
color during the first 5ꢀ10% conversion of 4, which was accom-
panied by the formation of small amounts of 6, the appearance of a
weak singlet at δ 5.52 (perhaps due to 7), and broad baseline
absorptions in the aliphatic and aromatic regions of the spectrum,
indicative of polymerization (eq 3).
The results of these experiments show that the photochemistry
of 4 is broadly analogous to that of the methylated derivatives 3a
and 3b,35 affording diene 6 as the major product in the presence
of MeOH or AcOH and the OꢀH insertion products expected
from reaction of GeH2 with the hydroxylated substrates; quali-
tative comparisons of the photolysis efficiencies indicate
1
Figure 1. Partial H NMR spectra of photolyzed C6D12 solutions of
(a) 4 (0.025 M) and AcOH (0.25 M), (b) 4 (0.025 M) and AcOD (0.25
M), and (c) 4-d2 (0.025 M) and AcOH (0.25 M).
in these spectra compared to those of the 4/AcOH photolyzate
was the replacement of the singlet at δ 5.26 due to the GeꢀH
protons in 8 with a 1:1:1 triplet centered at δ 5.24 (JHD = 3.9 Hz)
in the case of the 4/AcOD photolyzate (consistent with
AcOGeH2D) and a 1:2:3:2:1 pentet centered at δ 5.23 (JHD
=
3.8 Hz) in the case of the 4-d2/AcOH photolyzate (consistent with
AcOGeHD2).52 The NMR spectra of a photolyzed mixture of
4-d2 and MeOH in C6D12 showed the expected pentet due to the
D2GeH proton in MeOGeHD2 (7-d2).
The NMR spectra also showed resonances throughout the
δ 0.5ꢀ1.3, δ 2.5ꢀ3.5, and δ 5ꢀ6 regions of the spectra, none of
which were common to both the AcOH and MeOH photolysis
mixtures (see Figures S1 and S2, Supporting Information). The
spectra of the 4/AcOH(D) mixtures also contained two additional
sets of singlets in the vinylic region which integrated to 28ꢀ35% of
the areas of the vinylic protons due to 6 (see Figure 1), while the
spectrum of the 4/MeOH mixture contained several singlets in
the methoxyl proton region but no new resonances ascribable to
vinylic protons. In both cases, several of the multiplets collapsed
to lower multiplicities, and some disappeared altogether, in the
spectra of the corresponding 4-d2/ROH photolysis mixtures
(e.g., Figure S4, Supporting Information). GC/MS analysis
suggested that these were due to at least three C11H14Ge-
containing products in both cases; the apparent complexity of
the mixtures and the expected delicacy of the compounds53,54
suggested that isolating them would be quite difficult, so we did
not try. Nevertheless, it can be concluded from the deuterium-
labeling results that these compounds are formed mainly by
hydrogermylation processes, most likely involving the CdC
bonds in diene 6; they are all primary reaction products, as they
are already present in the mixtures after ca. 1% conversion of 4.
The resolvable spectral features due to two of the minor products
formed in the presence of AcOH are consistent with the presence
of structures containing the (2-methyl-3-phenyl-3-bute-
nyl)germyl (9; δ 1.17 (d), 3.06 (sextet), 5.04 (s), 5.15 (s))
and (3-methyl-2-phenyl-3-butenyl)germyl (10; δ 1.56 (s), 3.53
(t), 4.84 (s), 4.93 (s)) functionalities, the allylic hydrogens being
the ones originating on germanium. A third type of product (pos-
sibly a mix of diastereomers) was formed in significantly higher
yield relative to those of 6, 8, H2, and the other minor products
that the photolysis quantum yield of 4 is similar to those of 3a,b
(Φ ≈ 0.534,35) as well. However, while photolysis of 3b affords 6
and GeMe2-derived products in close to quantitative yields,35 the
diene accounts for only 50ꢀ60% of the material balance in the
photolysis of 4 under similar conditions. This is due to other
primary reactions involving GeꢀH moieties and suggests the
intervention of a reaction intermediate formed competitively with
GeH2 and 6 or prior to it, affording “trappable” GeH2 in one of two
or more competing decomposition channels. These competing
reactions are essentially eliminated in THF-d8, where 6 and 8 are
produced almost exclusively. While we cannot rule out the possi-
bility of a solvent effect on the photochemistry of 4 that is not
evident with the substituted germacyclopent-3-enes we have stu-
died previously,34,35 this is consistent with the proposal that the
formal extrusion of GeH2 from 4 proceeds via a short-lived
reaction intermediate that possesses more than one pathway for
reaction. The additional pathway(s), which the AcOH-trapping
results suggest involve addition of GeꢀH across the CdC bonds
in diene 6, are presumably suppressed in favor of formal GeH2
extrusion in the O-donor solvent owing to the latter’s ability
to stabilize the highly reactive germylene by Lewis acidꢀbase
complexation.36,55 The competing processes are presumably
10525
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
Scheme 1
Scheme 2
molecule of alcohol as catalyst for H-migration in the complex,37
which supplants the (higher energy) unimolecular H-migration
pathway that was considered computationally.
The complete suppression of the H2-elimination pathway and
the higher material balance in the reaction with AcOH in THF
can be explained as being due to the operation of a different
mechanism for the reaction of GeH2 with AcOH in the ether
solvent, where GeH2 is expected to exist as the Lewis acidꢀbase
complex with the solvent. As has been shown for GeMe2 and
GePh2,37 complexation of GeH2 with the ether can be expected to
enhance its reactivity toward electrophilic substrates and decrease
its reactivity toward nucleophilic ones. We thus envisage the
reaction with AcOH in THF as proceeding via initial protonation
of the GeH2ꢀTHF complex at germanium, either with con-
certed displacement of the solvent molecule by the nucleophilic
end of the substrate or in stepwise fashion via a (presumably very
short-lived) ion pair (see Scheme 2). In any event, the process
requires a relatively acidic substrate to be fast, which explains the
failure of MeOH to trap the germylene in detectable amounts in
this solvent.
Laser Flash Photolysis Studies. Laser flash photolysis of
flowed, deoxygenated solutions of 4 ((5ꢀ7) ꢁ 10ꢀ5 M) in
anhydrous hexanes, using the pulses from a KrF excimer laser
(248 nm, ∼20 ns, ∼100 mJ) for excitation, led to the formation
of two distinct transient species with different but clearly related
growth/decay behaviors. The shorter-lived of the two species
appeared to be formed with the laser pulse, exhibits λmax = 460 nm,
and decayed over ca. 20 μs with reasonably clean second-order
kinetics. The second species (λmax = 390 nm) grows in after the
pulse, reaching a maximum in concentration within ca. 3 μs and
then decaying over ca. 100 μs, also with predominant second-
order kinetics. The decay rates of both signals increased along
with their peak intensities when the excitation laser intensity was
increased, as expected for second-order decay processes. Figure 2
shows transient spectra recorded 16ꢀ60 ns and 5.5ꢀ6.0 μs after
the laser pulse, along with representative transient absorption
profiles recorded at monitoring wavelengths of 470 and 380 nm;
the two species were monitored at wavelengths longer and shorter,
respectively, than the absorption maxima to minimize interference
from the other species in the kinetic analyses. Analysis
of transient absorbance vs time data at the two monitoring
wavelengths according to eq 4 afforded second-order decay
coefficients of 2kdim/ε470-nm = (4 ( 1) ꢁ 107 cm sꢀ1 and
2kdim/ε370-nm = (1.4 ( 0.1) ꢁ 106 cm sꢀ1, respectively.
thermally activated, though at this stage we cannot rule out the
possibility that they might arise from a secondary photochemical
reaction of the putative reaction intermediate, provided its lifetime
is sufficiently long to allow its steady state concentration to build
up to appreciable levels.
A mechanism to account for the formation of 8 and H2 from
the reaction of GeH2 with AcOH is shown in Scheme 1 and
involves initial Lewis acidꢀbase complexation with the basic site
in the substrate40 followed by competingH-migration from oxygen
to germanium (yielding 8) and H2 elimination to afford the
secondary (acetoxy)germylene 12. The latter species presumably
accounts for some of the unidentified products in the photo-
lyzate. It should be noted that the sum of the yields of 8 and H2 is
roughly equal to that of diene 6, indicating that the scavenging of
GeH2 by AcOH proceeds essentially quantitatively. A related
competitionisalsoevidentinthetrappingexperimentwith MeOH,
but the reaction affords H2 in substantially lower yield relative to
the insertion product (7) than is the case with AcOH. Presum-
ably, this is because in the latter case H2 elimination from the
complex (11) proceeds via a six-membered transition state rather
than the more restrictive four-membered transition state that is
required in the case of the GeH2ꢀalcohol complex; the higher
Brønsted acidity of the carboxylic acid may also contribute to the
difference to some extent.55,56 The competition between inser-
tion and elimination in the reactions of GeH2 with MeOH has
been studied computationally (vide infra),55,56 but to our knowl-
edge this is the first experimental observation of the H2-elimina-
tion channel in the reaction of hydroxylic substrates with GeH2.
The relative yields of 6, H2, and 7 in the trapping experiments
with MeOH indicate the alcohol to be a significantly less efficient
trapping agent for GeH2 than the carboxylic acid, in contrast to
the behavior observed by us previously in similar experiments
with GeMe2 and GePh2.34,35 We speculate that this is due to
the fact that, as we showed to be the case with the latter
derivatives,34,35 the product-forming steps from the GeH2ꢀ
MeOH complex are much slower than is the case with the
GeH2ꢀAcOH complex and thus compete less effectively with
dissociation back to the free reactants. The net result is the
formation of significant amounts of polymeric material, at the
expense of the OꢀH insertion product 7. Interestingly, theory
predicts that H2-elimination proceeds via a ca. 6 kcal molꢀ1 lower
enthalpic barrier than insertion and should thus dominate the
product mixture from decomposition of the GeH2ꢀMeOH
complex.55,56 Our experimental results indicate that H2 is in fact
formed in only trace amounts in the trapping reaction with the
alcohol, in apparent conflict with the theoretical prediction.
However, this is most likely because in solution the net insertion
process can occur by a catalytic pathway involving a second
ΔAt ¼ ΔA0=½1 + ð2kdimΔA0=1εÞtꢂ
ð4Þ
Addition of acetic acid caused the decay of the 460 nm
transient (monitored at 470 nm) to accelerate and follow clean
pseudo-first-order kinetics, with little or no change in the initial
absorbance of the signal (ΔA0) compared to that obtained in the
absence of added substrate. At the same time, the growth rate of
the 390 nm transient was accelerated and its peak intensity was
reduced, indicating that its formation is quenched in the presence
of the carboxylic acid; this leads to the conclusion that the
10526
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
Figure 2. Transient absorption spectra from laser photolysis of 4 in
deoxygenated hexane solution at 25 °C, recorded 16ꢀ60 ns (-O-) and
Figure 3. Plots of kdecay (O) and (ΔA380,max)0/(ΔA380,max Q
)
(0) vs
[AcOH] for the 470 and 380 nm transient signals, respectively, from
laser photolysis of 4 in deoxygenated hexanes at 25 °C. The solid lines in
(c) are the linear least-squares fits of the data to eqs 5 and 6, respectively.
5.5ꢀ6.0 μs (
0
) after the laser pulse. The inset shows transient
3 3 3 3 3 3
growth/decay profiles recorded at 470 and 380 nm. The solid curve
drawn through the 470 nm decay is the best fit of the data to eq 4.
370 nm).34,35,38,59 Second, the rate coefficient for dimerization of
the 460 nm transient (2kdim/ε470-nm = (4 ( 1) ꢁ 107 cm sꢀ1) is
substantially lower than would be expected for GeH2 based
390 nm transient is a reaction product of the first-formed species.
Figure S7 (Supporting Information) illustrates the observed
effects on the transient signals as a function of AcOH concentra-
tion. Plots of the first-order decay rate coefficient of the 470 nm
signal (kdecay) vs scavenger concentration ([Q]) were linear and
were analyzed according to eq 5, where kQ is the second-order
rate constant for the reaction with the scavenger and k0 is the
(hypothetical) pseudo-first-order decay rate coefficient when
[Q] = 0. Linear plots were also obtained from analysis of the
relative peak intensities of the 380 nm signal in the absence and
on that determined for GeMe2 under similar conditions (2kdim
/
ε480-nm = (1.5 ( 0.3) ꢁ 108 cm sꢀ1),35 assuming the molar
extinction coefficient of the GeH2 absorption is not substantially
larger than that of GeMe2 (εmax = 750 ( 300 dm3 molꢀ1 cmꢀ1).35
Third, the rate constant for reaction of the 460 nm species with
AcOH is similar to the average of the values reported for GeMe2
under similar conditions (kQ = (9.4 ( 2.0) ꢁ 109 Mꢀ1 ꢀ1 35,38
)
—
not a factor of 2ꢀ3 higher (i.e., diffusion-controlled), as might be
predicted based on the fact that GeH2 is characteristically at least
10 times more reactive than GeMe2 toward most common
substrates in the gas phase.6 Finally, the fact that the 460 nm
species is nevertheless quenched rapidly and efficiently by AcOH
demands that if it were GeH2 then we should be able to trap it
cleanly in steady state photolysis experiments, which is not the
case. This all leads us to conclude that the 460 nm species is
almost certainly not GeH2, and thus the 390 nm species formed
as the primary product of its decay is not Ge2H4. If free GeH2 is
produced by laser photolysis of 4 under these conditions, then we
are unable to detect it.
presence of the substrate according to eq 6, where (ΔA380,max 0
)
and (ΔA380,max)Q are the peak signal intensities in the absence
and presence of Q and KSV is a proportionality factor reflecting
the efficiency with which the formation of the 390 nm species is
quenched by the substrate. Figure 3 shows the plots of kdecay and
(ΔA380,max)0/(ΔA380,max)Q vs [Q] obtained in the experiments.
Analysis of the data according to eqs 5 and 6 afforded values of
kQ = (8.1 ( 0.6) ꢁ 109 Mꢀ1 sꢀ1 and KSV = 13 200 ( 1400 Mꢀ1
,
respectively.
kdecay ¼ k0 + kQ ½Q ꢂ
ð5Þ
ð6Þ
The kinetic behaviors of the 460 and 390 nm species were also
examined in the presence of various other germylene scavengers
whose reactivities toward GeMe2 under similar conditions have
been quantified.35 These included triethyl- and n-butylamine(Et3N
and n-BuNH2, respectively), MeOH, THF, 3,3-dimethyl-1-butyne
(TBE), 4,4-dimethyl-1-pentene (DMP), isoprene, triethylsilane
(Et3SiH), and triethylgermane (Et3GeH). In all cases but MeOH,
THF, DMP, and isoprene (vide infra), addition of the substrate
caused similar effects on the transient profiles at 470 and 380 nm as
were observed with AcOH. The kQ and KSV values determined in
these experiments are listed in Table 1, along with the correspond-
ing values of kQ for GeMe2 under similar conditions.35,36
In the experiments with MeOH and THF in hexanes, addition
of the O-donors appeared to cause decreases in the initial signal
intensities of the 460 nm species, to an extent that increased with
increasing concentration (see Figure S8, Supporting Information).
This was accompanied by the appearance of new transient
absorptions centered at λmax = 310 and 320 nm, respectively,
which are consistent with germyleneꢀO-donor Lewis acidꢀbase
complexes; the spectra (Figures S9 and S10, Supporting In-
formation) are both red-shifted relative to those of the
ðΔA380, maxÞ0=ðΔA380, maxÞQ ¼ 1 + KSV½Q ꢂ
The transient spectra of Figure 2, the nature of their temporal
evolutions, and the response of the two species to the addition of
AcOH are all broadly consistent with a germylene and digermene
assignment for the 460 and 390 nm transient products of laser
photolysis of 4, respectively, based on the behavior we have
observed previously for substituted germylenes34ꢀ36,39,40,42 and
silylenes57,58 under conditions similar to those employed here.
However, there are several discrepancies between the observed
results and those expected for GeH2 and Ge2H4 based on pre-
viously published experimental and (or) theoretical data for the
two species in the gas phase. First, the absorption maxima do not
agree well with the values predicted for GeH2 and Ge2H4 by
time-dependent DFT calculations (vide infra and ref 34) or with
the experimental gas-phase fluorescence excitation spectrum
of GeH2 (λmax ∼ 515 nm).44 This contrasts the excellent agree-
ment that exists between the gas-phase,43 solution-phase,34,35,38
and theoretically predicted34 UVꢀvis spectra of GeMe2 (λmax
∼
470 nm) and the similarly close agreement between the theore-
tical and experimental (solution-phase) spectra of Ge2Me4 (λmax
∼
10527
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
Table 1. Bimolecular Rate and Equilibrium Constants (kQ and Keq, Respectively) and KSV Values for Quenching of the 460 and
390 nm Transient Products, Respectively, From Laser Photolysis of 4 in Deoxygenated Hexanes Solution at 25 °C by Various
Substratesa
substrate
kQ/109 Mꢀ1 sꢀ1
KSV/Mꢀ1
kQ/109 Mꢀ1 sꢀ1 (GeMe2)b
Et3N
5.4 ( 0.5
11.3 ( 0.8
8.1 ( 0.6
1500 ( 400
11100 ( 140
13200 ( 1400
2.8 ( 1.4
11 ( 2
8.7 ( 0.7
n-BuNH2
AcOH
12 ( 3
9.4 ( 2.0c
Et3SiH
Et3GeH
0.00013 ( 0.00006
0.056 ( 0.005
0.00055 ( 0.00015
0.045 ( 0.015
11 ( 2c
3,3-dimethyl-1-butyne (TBE)
5.3 ( 0.2
0.072 ( 0.004 (Keq = 45 ( 10 Mꢀ1
4400 ( 200
108 ( 5
4,4-dimethyl-1-pentene (DMP)
)
9.6 ( 1.2
isoprene
MeOH
THF
<0.03
ꢀ
10.8 ( 2.8
(Keq = 280 ( 40 Mꢀ1
8.4 ( 1.1 (Keq = 2400 ( 150 Mꢀ1
)
510 ( 40
64 ( 1
(Keq = 900 ( 60 Mꢀ1
)
d
)
11 ( 2 (Keq = 9800 ( 3800 Mꢀ1
)
a Also listed are the corresponding values of kQ and (or) Keq for quenching of GeMe2 by the same substrates under similar conditions.35,36. b Data from
refs 35, 36, and 38. c Average of 2ꢀ3 independent determinations using different precursors. d kQ is too fast to measure, given the magnitude of Keq.
corresponding GeMe2ꢀO-donor complexes under similar
conditions.36 The effects on the intensities and decay kinetics
of the 460 nm species are again quite similar to those observed
previously for GeMe2 in the presence of low concentrations of
THF and MeOH in hexanes36 and are consistent with reversible
reaction with the O-donor substrates. With THF, the equilibrium
constant is of an appropriate magnitude for both the approach to
equilibrium and the residual amount of the free species present at
equilibrium to be detected, while with MeOH the behavior is
consistent with a rapidly attained but relatively unfavorable
reversible process for which only the residual amount of free
460 nm species present at equilibrium can be detected.36 Analysis
Transient decays recorded for deoxygenated solutions of 4 in
neat THF afforded bimodal decays in the 280ꢀ330 nm range,
consisting of a short-lived component (τ ∼ 5 μs; λmax ∼ 310 nm)
superimposed on a long-lived signal (τ ∼ 150 μs; λmax ∼ 290 nm)
that decayed to a stable residual absorption with λmax < 280 nm.
Saturation of the solution with air rendered the short-lived
component unresolvable from the laser pulse and afforded a single
transient species which exhibited an apparent absorption max-
imum at λmax = 290 nm and decayed with complex kinetics, as
shown in Figure 4a. Addition of AcOH or AcOD to the air-
saturated solution led to a shortening of the lifetime of the
290 nm species in proportion to concentration, and plots of kdecay
vs [AcOL] (L = H or D) according to eq 5 were linear (see
Figure 4b), affording rate constants of kAcOH = (2.5 ( 0.2) ꢁ
105 Mꢀ1 sꢀ1 and kAcOD = (1.0 ( 0.2) ꢁ 105 Mꢀ1 sꢀ1 (kH/kD =
2.5 ( 0.7). These values are both ca. 50 times smaller than those
reported by us previously for the corresponding reactions of the
GeMe2ꢀTHF complex in THF solution.37
36
of the data in the manner detailed in our earlier study of GeMe2
affords the forward rate and equilibrium constants listed in
Table 1 (see Figures S11 and S12, Supporting Information).
THF
The kQ
value is again similar to that for the corresponding
reaction with GeMe2, while the equilibrium constants are both
3ꢀ4 times smaller than the GeMe2 values.36
Unusually,35,39,42 addition of 0.5ꢀ2.0 mM isoprene or DMP
had no discernible effect on the intensity or decay characteristics
of the 460 nm species and caused only a marked lengthening in
the growth time of the 380 nm signal and slight increases in its
maximum intensity. The behavior of the 380 nm signal suggests
that the substrate acts to slow down the reaction channel which
leads to the formation of the 390 nm transient product. Addition
of larger quantities (10ꢀ80 mM) of the alkene caused similar
effects on the 470 nm signals as were observed in the experiments
with THF, and similar analyses of the data afforded rate and
The transient behavior in the presence of MeOH and THF in
dilute hexanes and in the neat O-donors is consistent with an
assignment of the 310ꢀ320 nm species to germyleneꢀMeOH and
germyleneꢀTHF complexes, respectively, though it is quite
likely that the species observed in dilute hexanes containing
millimolar concentrations of the O-donors are different than
those observed in the neat liquids. This conclusion is based on
the fact that the differences in the spectra of the two O-donor
complexes in dilute hexanes compared to the neat liquids are
significantly greater than those exhibited by the GeMe2ꢀMeOH
and GeMe2ꢀTHF complexes under the two sets of conditions.37
Since the product studies indicate the chemistry is profoundly
cleaner in THF than in hydrocarbon solvents, we assign the
transient observed in THF solution to the GeH2ꢀTHF complex.
The observation of quenching of the species by AcOH and
AcOD, which exhibits a clearly primary kinetic isotope effect, is
consistent with this assignment and with the reaction mechanism
proposed in Scheme 2. Its lack of reactivity toward MeOH, which
the product studies show to be an ineffective scavenger in THF
solution, is also consistent with the assignment. We tentatively
assign the species observed in neat MeOH solution to the
GeH2ꢀMeOH complex.
equilbrium constants of kQ = (7.2 ( 0.4) ꢁ 107 Mꢀ1 sꢀ1 and
DMP
Keq
= 45 ( 10 Mꢀ1, respectively, for reaction of the spe-
cies with DMP. The rate constant is ca. 100 times smaller, and the
equilibrium constant at least 500 times smaller, than those for
reaction of GeMe2 with the same substrate under similar
conditions.35
Laser photolysis of 4-d2 in anhydrous hexanes resulted in
transient spectroscopic behavior very similar to that obtained in
the experiments with 4 under similar conditions (see Figure S13,
Supporting Information), but with one important difference. In
this case, the rising edge of the 470 nm absorption exhibited a
slight growth that could not be detected in the experiments with
the protiated derivative, suggesting that the rate of formation of
the species is subject to a kinetic isotope effect.
Comparison of the rate and equilibrium constants for reaction
of the 460 nm species with the corresponding ones for GeMe2
10528
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
Figure 4. (a) Transient absorption spectra of a 0.14 mM solution of 4 in air-saturated THF, 0ꢀ1.3 μs (-O-), 19.8ꢀ22.4 μs (-0-), and 343ꢀ347 μs (-Δ-)
after the laser pulse; the inset shows a transient decay profile recorded at 290 nm. (b) Plots of kdecay vs [AcOL] (L = H or D) for quenching of the 290 nm
species by AcOH and AcOD in air-saturated hexane at 25 °C.
under similar conditions (Table 1) confirms our initial conclu-
sion that the spectroscopic and kinetic behavior exhibited by the
species is inconsistent with what is expected for free GeH2. The
460 nm transient reacts with amines, AcOH, Et3SiH, Et3GeH,
TBE, DMP, isoprene, and THF with rate constants that are
in almost every case smaller than the corresponding values for
GeMe2,35,36 in some cases by a sizable margin. Rate constants
have been reported for reaction of GeH2 in the gas phase with the
same or similar substrates, and they are in every case significantly
greater than those exhibited by GeMe2 under the same
conditions.6,8ꢀ10,13ꢀ15,19,20,43 Nevertheless, the species exhibits
all of the characteristics expected for a germylene derivative, with
the exception of its anomalously low reactivity toward isoprene
and DMP. The behavior seems more compatible with a germy-
leneꢀalkene π-complex than a free germylene.
Scheme 3
The occurrence of competing solvent cage reactions between
GeH2 and its diene (6) coproduct provides a reasonable explana-
tion for the results of both the steady state and laser photolysis
experiments with 4 and 4-d2; our proposed mechanism
(Scheme 3) draws on the gas-phase kinetics and computational
studies of the reaction of GeH2 with ethene9,11 and on the results
of recent experimental and computational studies of the reac-
tions of substituted germylenes with conjugated dienes.35,41,60
The competing reactions in question are (1 + 2)-cycloaddition to
the corresponding vinylgermirane (14), which is expected to be
reversible,35,41,60 and 1,2-hydrogermylation to afford the corre-
sponding hydridogermylenes 15 or 16. Each of these processes is
expected to involve the prior formation of a π-complex (13)
between GeH2 and one of the CdC bonds in the diene.9,11,35,41,60
Four isomeric hydridogermylenes are possible depending on the
preferred regiochemistry of the reaction in the case of the
unsymmetrical diene (6), two arising from attachment of germa-
nium to one of the terminal carbons of the diene (15) and two
from attachment to one of the internal carbons (16). The NMR
characteristics exhibited by two of the minor products formed
in C6D12 in the presence of AcOH are consistent with the
substituted 3-butenyl side chain present in 15. Furthermore, the
conformational flexibility afforded by the two-carbon chain
separating the Ge(II) center from the terminal CdC bond in
the structure should allow the formation of a relatively unstrained
intramolecular π-complex (15-π), which should exhibit a UVꢀvis
absorption maximum at shorter wavelengths than that of the
uncomplexed species and exhibit somewhat different patterns of
reactivity toward added substrates—alkenes and dienes in parti-
cular. The (1 + 2)-cycloaddition process can be viewed as
effectively retarding the escape of free GeH2 and 6 from
the solvent cage in which they are formed, allowing the 1,2-
hydrogermylation process(es) to compete with cage escape.
In THF solution, the solvent cage is itself reactive toward
GeH2 and thus quenches the hydrogermylation process(es)
altogether.
The isomeric vinylgermiranes (14) can be ruled out as
candidates for the 460 nm transient, as they should absorb below
300 nmregardlessofthe regiochemistry39 andshouldalsobe much
less reactive toward the nucleophilic substrates of Table 1 than the
measured rate constants indicate. A GeH2ꢀdiene π-complex
(e.g., 13) can also be ruled out, based mainly on the observation
of the slight growth in the 460 nm signal obtained with 4-d2,
which is indicative of a primary isotope effect on the rate of its
formation. Furthermore, it is difficult to rationalize the apparent
concentration dependence of the product mixture obtained in
the presence of 0.05ꢀ0.25 M AcOH in the context of an assign-
ment of 13 for the 460 nm species, as it is quenched by this
substrate at close to the diffusion-controlled rate; taken together,
these two sets of results suggest that the 460 nm species is
something that may be formed competitively with GeH2 or its
trapping products but is not itself capable of contributing to their
formation. In other words, the 460 nm species is more likely to be
10529
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
Table 2. PW91PW91/6-311+G(2d,p) and PBE0/cc-pVTZ (in Parentheses) Electronic Energies, Enthalpies (298.15 K), and Free
Energies (298.15 K) of Stationary Points in the Reactions of GeH2 with s-trans 2-Methyl-6-phenyl-1,3-butadiene (6) in kcal molꢀ1
Relative to the Isolated Reactants and Predicted UVꢀvis Absorption Maxima, Calculated for the PW91 Structures at the
TDPW91PW91/6-311+G(2d,p), TDPBE0/cc-pVTZ, and TDB2PLYP/6-311G(d,p) Levels of Theory
λmax (nm; TDDFT)
species
13aout
ΔEelec
ΔH298K
ΔG298K
PW91 | PBE0 | B2PLYP
ꢀ23.7 (ꢀ21.0)
ꢀ19.5
ꢀ21.2 (ꢀ23.1)
ꢀ16.0 (ꢀ14.3)
ꢀ22.3 (ꢀ23.9)
ꢀ19.9
ꢀ37.4 (ꢀ37.7)
ꢀ19.0 (ꢀ17.3)
ꢀ17.7
ꢀ21.4 (ꢀ18.8)
ꢀ17.9
ꢀ19.1 (ꢀ21.0)
ꢀ14.9 (ꢀ13.1)
ꢀ18.6 (ꢀ20.1)
ꢀ16.8
ꢀ32.9 (ꢀ33.1)
ꢀ17.1 (ꢀ15.3)
ꢀ16.3
ꢀ9.9 (ꢀ7.3)
ꢀ5.3
ꢀ7.4 (ꢀ9.1)
ꢀ2.9 (ꢀ1.1)
ꢀ8.7 (ꢀ10.0)
ꢀ5.1
ꢀ20.4 (ꢀ20.5)
ꢀ5.9 (ꢀ4.2)
ꢀ4.2
383 | 361 | 351
T14a,out
14a
284 | 248 | 234
535 | 513 | 505
462 | 422 | 392
T15a
15a
T15a<>15a-π
15a-π
13ain
T14a,in
T16a
ꢀ13.4 (ꢀ11.3)
ꢀ36.4 (ꢀ36.7)
ꢀ23.5
ꢀ12.1 (ꢀ10.0)
ꢀ32.1 (ꢀ32.3)
ꢀ21.3
ꢀ0.1 (+2.2)
ꢀ20.7 (ꢀ20.7)
ꢀ10.0
17a-π
13bout
T14b,out
14b
304 | 276 | 263
419 | 390 | 374
ꢀ19.4
ꢀ17.8
ꢀ5.2
ꢀ21.8
ꢀ19.9
ꢀ8.1
299 | 265 | 247
483 | 433 | 399
399 | 362 | 335
T15b
ꢀ16.3
ꢀ15.2
ꢀ2.6
15b
ꢀ25.8
ꢀ21.9
ꢀ10.5
T15b<>15b-π
15b-π
13bin
ꢀ20.5
ꢀ17.4
ꢀ5.6
ꢀ35.3
ꢀ31.0
ꢀ20.0
ꢀ18.9
ꢀ17.0
ꢀ5.5
T16b
ꢀ13.2
ꢀ12.1
+0.1
17b-π
ꢀ37.4
ꢀ33.0
ꢀ21.6
322 | 290 | 275
thespecies responsible forformationofsomeoftheminorproducts
observed in the trapping experiments than the major ones.
A hydridogermylene assignment is consistent with this last
conclusion, as well as with the apparent isotope effect on the rate
of its formation and its generally comparable reactivity toward many
of the substrates listed in Table 1 to those reported previously for
GeMe2 and GePh2. Intramolecular complexation with the CdC
bond at the end of the side chain could explain the unusually low
reactivity of the species toward isoprene and the alkene (DMP),
as it should reduce the driving force for reaction with the CdC
bond(s) of olefinic substrates but may have relatively small
effects on the rates of other typical germylene reactions. The
CdC bonds in germylenes 15 and (or) 16 may get involved
along with the Ge(II) center in reactions with hydroxylic
substrates,61 which could explain why olefinic side products are
formed in the reaction with AcOH but are not in that with
MeOH (vide supra). If all this is correct, then it is the identities of
these minor products that hold the key to the identification of the
transient andthe mechanismforitsformation. Withsometentative
partial structural information in hand, we turned to computational
methods, to see whether they might allow a more definitive
structural assignment for the 460 nm transient to be made.
Computational Studies. A portion of the pe surface for the
reactions of GeH2 with diene 6 was explored with density func-
tional methods, carried out at the PW91PW9147/6-311+
G(2d,p) level of theory. Geometry optimizations and frequencies
calculations were carried out for GeH2 and the s-trans conformer
of 6, the four possible π-complexes of GeH2 with the CdC
bonds in (s-trans) 6 (13), the two isomeric vinylgermiranes (14)
derived from (1 + 2)-cycloaddition, the four possible 1,2-hydro-
germylation products (15 and 16) and their intramolecular
π-complexes (15-π and 17-π), and the transition states for
formation of 14ꢀ16 from the corresponding GeH2ꢀ6 complexes
(eq 7). The transition states were confirmed to be first-order
saddle points on the basis of their vibrational frequencies, and
their connections to reactants and products were established by
internal reaction coordinate (IRC) calculations. Figure 5 shows
the various structures located in the calculations for the reactions
of GeH2 with the C1dC2 bond in s-trans 6, along with selected
geometrical parameters. The structures of the corresponding
stationary points in the reaction paths involving the C3dC4 bond
are all quite similar to those involving the other and are shown in
the Supporting Information (Figure S14). The bond distances
and angles associated with the various structures compare well
(where applicable) with those reported previously for the GeH2
+ ethylene9,62ꢀ65 and GeMe2 + 1,3-butadiene60 systems using
other DFT or ab initio methods.
Interestingly, we were unable to locate an energy minimum
corresponding to either isomer of germylene 16. An IRC
calculation confirmed that they are the incipient reaction
10530
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
Figure 6. Partial free energy surface for the reactions of GeH2 with
2-methyl-3-phenyl-1,3-butadiene (6), calculated at the PW91PW91/6-
311+G(2d,p) level of theory.
s-trans 6; vibrational frequencies were not scaled. The enthalpy
surfaces are shown in Figure S15 (Supporting Information),
along with full details of the computed structures. The transition
states reported for the conformational interconversion of 15
and 15-π should be considered somewhat ill-defined, as these
processes involve rotations about several single bonds in the
structures. The structures reported (see Supporting Informa-
tion) correspond to eclipsed conformers about the C1ꢀC2
bonds in 15a,b, and while they are true transition states according
to the calculations, we could not confirm that they are the actual
ones accessed in these processes.
The PW91 method was chosen on the basis of the results of
preliminarycalculations of three stationary points on the pe surface
for the (1 + 2)-cycloaddition of GeMe2 and s-trans 1,3-butadiene,
which has been characterized previously at the CCSD(T)/cc-
pVTZ//B3LYP/6-311G(d,p) level of theory.60 Our cal-
culations focused on the GeMe2ꢀbutadiene π-complex (18),
the corresponding vinylgermirane (19), and the transition state
linking them (T19; eq 8) and were also carried out at the
B2PLYP67/6-311G(d,p), PBE068/cc-pVTZ, and Gaussian-4
(G4)69 levels; the results are compiled in Table S1 of the
Supporting Information. The experimental upper limit of
ΔG° e ꢀ4.1 kcal molꢀ1 for the (1 + 2)-cycloaddition of GeMe2
to isoprene, calculated from the reported lower limit of the equi-
librium constant in hexanes solution at 298 K,35 served as the
primary benchmark for the calculations. PW91 and PBE0 both
afford overall reaction energies that are consistent with the
experimental data for the GeMe2 + isoprene system, as do also
the higher level G4 and CCSD(T) methods.
Figure 5. Stationary points and selected geometric parameters on the
potential energy surface for reaction of GeH2 with the C1dC2 bond of
2-methyl-3-phenyl-1,3-butadiene (6), calculated at the PW91PW91/6-
311+G(2d,p) level of theory.
products from the appropriate transition states (T16), but
attempts to optimize their geometries always resulted in rear-
rangement to 17-π, the intramolecular π-complexes of the
corresponding 1-(trans-2-butenyl)germylene derivatives. Thus,
the alternate hydrogermylation pathway can be viewed formally
as a sequential 1,2-hydrogermylation/[1,3]-GeH migration, the
second stage of which proceeds only as far as 17-π. The
migration process is presumably facilitated by the particularly
weak bond between Ge and the secondary allylic carbon in the
incipient product (16). We note that a stable bisgermylene
derivative featuring bonding characteristics analogous to that in
17-π has recently been reported by Power and co-workers.66
The calculated thermochemical data for the various structures
investigated are listed in Table 2, while Figure 6 summarizes the
computed standard free energy surfaces, relative to free GeH2 and
We could not locate a minimum energy structure for complex 18
or a transition state for the formation of 19 using the PBE0 method,
which would seem to indicate it is more reluctant than PW91 to
stabilize π-complexes; Birukov et al. experienced similar difficulties
with the PBE density functional in their study of the cycloaddition of
GeMe2 with ethylene.64 A similar characteristic appears to be shared
as well by the higher level methods, so we also carried out a
more limited series of geometry and frequencies calculations on the
10531
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
GeH2 + 6 system at the PBE0/cc-pVTZ level, confining our
attention to the minima on the C1ꢀC2 (“a”) half of the pe surface
and the transition states T15a and T16a. The calculated energies
are included in Table 2; the structures varied only modestly from
those obtained at the PW91 level. The PBE0 method predicts a
reversal of the relative energies of the vinylgermirane (14a) and
the “out” π-complex (13aout) compared to PW91. However, the
two methods arrive at similar energies for the respective favored
structures and similar energies for the hydrogermylation pro-
ducts (15a, 15a-π, and 17a-π) and the transition states leading to
them (T15a and T16a).
The energies and oscillator strengths of the six lowest transi-
tions in the electronic spectra of 13out, 14, 15, 15-π, and 17-π were
calculated for the PW91PW91/6-311+G(2d,p) structures using
time-dependent (TD) DFT at the PW91PW91/6-311+G(2d,p),
B2PLYP/6-311G(d,p), and PBE0/cc-pVTZ levels of theory.
The predicted longest-wavelength UVꢀvis absorption maxima
obtained from the resulting simulated spectra are listed in Table 2.
The spread in the predicted λmax values varies between 0.3 and
0.6 eV for most of the structures investigated, with TDB2PLYP
consistently predicting the highest energy value and TDPW91
the lowest.
The same three TDDFT methods were also employed to
calculate UVꢀvis spectra for GeH2, GeMeH,70 GeMe2, and the
MeOH and THF complexes of GeH2 and GeMe2 (Table S2,
Supporting Information), again employing structures optimized
at the PW91 level. The predictions afforded by the three methods
for the germylenes agree with the experimental spectra to within
0.13 eV in all cases. For the complexes, the predicted spectra vary
with the TDDFT method in a similar way as those for the GeH2/
6-derived species, and only TDB2PLYP came reasonably close to
reproducing the known experimental spectra of the GeMe2ꢀ
MeOH and ꢀTHF complexes; it incorrectly predicted the former to
show the longer-wavelength absorption, however.36 Similar spectra
were predicted for the corresponding GeH2-derived species.
The calculations confirm that complexation of GeH2 with one
or the other of the CdC bonds in 6 is at least moderately
exergonic and that the four possible π-complexes are the critical
intermediates in both the (1 + 2)-cycloaddition and hydroger-
mylation processes; their formation from the isolated reactants is
expected to be barrierless.9,60,62ꢀ65 The so-called “out” com-
plexes, in which the lone pair on germanium is pointed along the
CdC bond axis away from the bulk of the π-electron density in the
diene, are predicted to be 3ꢀ5 kcal molꢀ1 lower in energy than the
corresponding “in” complexes, presumably because electronꢀ
electron repulsion is minimized in the out orientation. The lowest-
energytransition state available toboth the inand out complexes is
that for ring closure to the corresponding vinylgermirane, thus
providing a low energy pathway for their interconversion. In any
event, the calculations predict that GeH2 and 6 initially drop into
two energy wells, each populated by the in and out complexes and
the corresponding vinylgermirane, which define the starting points
for further reaction. The calculated barriers for the interconver-
sion of these species are such that equilibrium between 13out and
14 should be established on a time scale of a few nanoseconds
or less.
Interestingly, however, the two (a and b) pe surfaces are nearly
mirror images of one another, so the calculations further predict
that reaction at the two CdC bonds in the diene should occur with
minimal regioselectivity. The energy of 15 in its initial geometry is
quite similar to that of the minimum in the starting well, but
formation of the corresponding intramolecular π-complex (15-π)
takes it further down in energy by 7ꢀ8 kcal molꢀ1. It is difficult to
quantify the free energy barriers for this latter process because of
the complexity of the conformational motions that are involved,
but it seems safe to assume that they are significantly smaller than
those for the return of 15 to the starting wells. In any event, the
free energy of activation for the return of 15-π to 15 is predicted
to be at least 12 kcal molꢀ1, corresponding to a unimolecular rate
constant on the order of ca. 104 sꢀ1 or less. This suggests that the
formation of 15-π should be effectively irreversible on the
microsecond time scale under ambient conditions in solution.
Provided that further unimolecular reactions of the species are
relatively slow, the end result in the absence of a reactive
substrate should be dimerization.
The calculations thus point to 15a-π and 15b-π as the most
likely candidates for the 460 nm transient observed in the laser
photolysis experiments with 4, as we concluded from the experi-
mental results discussed above. The calculated free energies of
activation for their formation from the starting wells are on the
order of 7ꢀ8 kcal molꢀ1 at the present levels of theory, which
corresponds to a unimolecular rate constant in the range of ca.
1ꢀ5 ꢁ 107 sꢀ1. This is in quite reasonable agreement with the
lower limit of ca. 108 sꢀ1 that is defined by the 20 ns duration of
our laser pulse and the fact that the formation of the 460 nm
transient from 4 appears to be complete by the end of that time
window.
Despite the rather large variation in λmax that the three TDDFT
methods predict (Table 2), they all agree that 15a-π should absorb
at significantly lower energies (by 0.5ꢀ0.6 eV) than 15b-π,
presumably owing primarily to the difference in the substituents
at the CdC bond in the two derivatives: phenyl in the case of
15a-π vs methyl in 15b-π. This points to 15a-π as the more likely
assignment for the 460 nm transient. The only other structures
that the TDDFT calculations suggest should be considered are
the open-chain conformers 15a and 15b, but (as discussed above)
the calculated energetics and experimental kinetic data dictate
against either of these assignments. It is interesting to note that the
predicted spectrum of 15b is blue-shifted somewhat relative to
that of 15a, which may be due to a weak complexing interaction
between the Ge(II) center and the phenyl ring two carbons away.
Indeed, the calculations suggest 15b is ca. 2 kcal molꢀ1 more
stable than 15a, presumably for this reason.
If 15b-π is formed competitively with 15a-π as the product
studies with AcOH suggest, then its absorptions are predicted to
lie underneath the strong absorptions in the 360ꢀ400 nm region
of the spectrum, which are mainly due to the much longer-lived
product of dimerization of the 460 nm species. Indeed, the 16ꢀ60 ns
spectrum of Figure 2 suggests there may be a promptly formed
component of the transient absorptions in this range, its max-
imum blue-shifted slightly relative to the absorption maximum of
the long-lived dimeric product. It is difficult to be certain of this,
however. The dominant transient absorption in this range exhibits
λmax = 390 nm, grows in concomitantly with the decay of the
absorptions due to 15a-π, and then decays over an extended time
scale with second-order kinetics, consistent with a digermene
assignment. On the basis of the above, we speculate that the
The starting wells are flanked on either side by the transition
states for the two hydrogermylation processes (T15 and T16), of
which that for the formation of 15 (T15), is predicted to be lower
by ca. 3 kcal molꢀ1. Thus, the calculated barriers to reaction at the
two individual CdC bonds in the diene predict a clear preference
for one hydrogermylation pathway over the other in both cases.
10532
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
390 nm absorptions are due to a mixture of isomeric digermene
dimers of germylenes 15a and 15b.
undergo rapid, competing cyclization to the corresponding vinyl-
germiranes, which is reversible, and (1,2)-hydrogermylation to yield
a mixture of two isomeric 3-butenylhydridogermylene derivatives,
which is not. The latter products are then further stabilized by
the formation of the corresponding intramolecular Ge(II)ꢀalkene
π-complexes. It is these species that are proposed to be mainly
responsible for the complicated mixture of minor products
observed in steady state trapping experiments and for the
460 nm transient absorption observed in laser photolysis experi-
ments with 4 in hexanes. DFT calculations are in full support of
this mechanistic analysis and further support a unique structural
assignment for the 460 nm transient product to one of the two
possible germyleneꢀalkene π-complexes. The species reacts
unusually slowly with alkenes and dienes, partly because the first
steps in these reactions involve a similar interaction (i.e.,
π-complexation) to the intramolecular one that is already pre-
sent in the molecule and also because the stabilization imparted
on the Ge(II) center by complexation reduces the overall driving
force for the reactions.
The usual ultimate course of the reactions of germylenes with
conjugated dienes is (1 + 4)-cycloaddition via the s-cis conformer
of the diene, to yield the corresponding germacyclopent-3-ene
derivative.60,71 We have not explored the pe surfaces associated
with the s-cis conformer of 6; they can be expected to link to 4 via
higher energy transition states than those for the formation of 15
and perhaps 17-π as well. The (1 + 4)-cycloaddition process could
well compete with the hydrogermylation chemistry exhibited by
GeH2 and 6, but if it does it regenerates 4 and thus constitutes
only a source of inefficiency in the formation of products (stable
or transient) from the photolysis of 4. The actual competition
between these two processes in the reaction of GeH2 with
conjugated dienes distinct from 6 is the subject of continued
study in our laboratory.
’ SUMMARY AND CONCLUSIONS
The results of this work show that 3-methyl-4-phenylgerma-
cyclopent-3-ene (4) is a reasonably efficient photochemical pre-
cursor to GeH2 in solution. However, unlike the 1,1-disubstituted
derivatives that have been studied previously, photolysis of the
compound produces the germylene cleanly only when the solvent
is sufficiently nucleophilic to pull the highly reactive parent species
out of the solvent cage in which it is formed and away from its
coproduct, 2-methyl-3-phenyl-1,3-butadiene (6), with which it
appears to react at rates similar to or faster than that of diffusional
separation in nonviscous hydrocarbon solvents. Thus, in THF
solution, photolysis of 4 produces GeH2 essentially exclusively, in
the form of its Lewis acidꢀbase complex with the solvent. Under
these conditions, the species can be trapped as the corresponding
OꢀH insertion product by reasonably acidic substrates such as
acetic acid. The GeH2ꢀTHF complex has been detected by laser
photolysis of 4 in THF solution, and preliminary aspects of its
kinetic behavior and reactivity have been determined.
Further studies of the chemistry of the parent Ge(II) deriva-
tive in solution and of the scope of intramolecular π-complexa-
tion phenomena in germylene (and silylene) chemistry are in
progress.
’ ASSOCIATED CONTENT
S
Supporting Information. Details of the synthesis and
b
characterization of 4 and 4-d2; NMR spectra and concentration vs
time plots from steady state photolysis experiments; additional
kinetic data determined in laser flash photolysis experiments; details
of the computational studies; and tables of calculated Cartesian
coordinates and energies. This material is available free of charge via
the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Much more complicated chemistry results from photolysis of
4 in hydrocarbon solvents. Product studies show that diene 6 is
formed in only 50ꢀ60% yield, along with H2 and the OꢀH
insertion products expected from reaction of GeH2 with AcOH
and MeOH. Several other products are also formed in both cases;
while they have not been fully identified, their NMR character-
istics and deuterium-labeling studies clearly show them to be
derived from hydrogermylation processes, formally involving a
CdC bond in the diene coproduct, GeH2, and the trapping agent.
Laser photolysis of 4 in hexanes affords a promptly formed
transient that exhibits spectral and kinetic characteristics which
are more or less typical of reactive germylenes in solution but
inconsistent with what would be expected for GeH2. The species
reacts unusually slugglishly (and reversibly) with a terminal alkene
and shows no indication of a reaction with isoprene, another
normally rapid and efficient scavenger of transient germylenes in
solution. It is formed within the ca. 20 ns duration of the laser
pulse, but this is slowed detectably in the deuterated analogue of
the precursor, indicating a primary isotope effect on the rate of its
formation.
Corresponding Author
leigh@mcmaster.ca
Present Addresses
†Department of Chemistry, University of Ottawa, 10 Marie Curie,
Ottawa ON Canada K1N 6N5.
‡Department of Dermatology and Skin Science, Vancouver General
Hospital, 835 West 10th Avenue, Vancouver BC Canada V5Z 4E8.
§Syngenta Crop Protection Canada, Inc., 140 Research Lane,
Research Park, Guelph, ON Canada N1G 4Z3.
Department of Chemistry, University of Alberta, 11227
Saskatchewan Drive, Edmonton AB Canada T6G 2G2.
’ ACKNOWLEDGMENT
We thank the Natural Sciences and Engineering Research
Council of Canada for financial support, the McMaster Regional
Centre for Mass Spectrometry for mass spectrometric analyses,
and Teck-Cominco Metals Ltd. for a generous gift of germanium
tetrachloride. Partof thisworkwas madepossible by the facilities of
the Shared Hierarchical Academic Research Computing Network
(SHARCNET:www.sharcnet.ca) and Compute/Calcul Canada.
The experimental data are accommodated by a mechanism in
whichGeH2 and 6 undergo a rapid series of reactionsin the solvent
cage in which they are formed, which compete with diffusional
separation. These are proposed to begin with the spontaneous
formation of the various possible π-complexes between GeH2
and a CdC bond in the diene coproduct. The complexes
’ REFERENCES
(1) Newman, C. G.; Dzarnoski, J.; Ring, M. A.; O’Neal, H. E. Int. J.
Chem. Kinet. 1980, 12, 661.
10533
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534

Journal of the American Chemical Society
ARTICLE
(2) Dzarnoski, J.; O’Neal, H. E.; Ring, M. A. J. Am. Chem. Soc. 1981,
103, 5740.
(35) Leigh, W. J.; Lollmahomed, F.; Harrington, C. R. Organome-
tallics 2006, 25, 2055.
(3) Isobe, C.; Cho, H.-C.; Crowell, J. E. Surf. Sci. 1993, 295, 117.
(4) Du, W.; Keeling, L. A.; Greenlief, C. M. J. Vac. Sci. Technol. A
1994, 12, 2281.
(36) Leigh, W. J.; Lollmahomed, F.; Harrington, C. R.; McDonald,
J. M. Organometallics 2006, 25, 5424.
(37) Lollmahomed, F.; Huck, L. A.; Harrington, C. R.; Chitnis, S. S.;
Leigh, W. J. Organometallics 2009, 28, 1484.
(5) Boganov, S. E.; Egorov, M. P.; Faustov, V. I.; Krylova, I. V.; Nefedov,
O. M.; Becerra, R.; Walsh, R. Russ. Chem. Bull. Int. Ed. 2005, 54, 483.
(6) Becerra, R.; Walsh, R. Phys. Chem. Chem. Phys. 2007, 9, 2817.
(7) Boganov, S. E.; Egorov, M. P.; Faustov, V. I.; Nefedov, O. M. In
The chemistry of organic germanium, tin and lead compounds; Rappoport,
Z., Ed.; John Wiley and Sons: New York, 2002; Vol. 2, p 749.
(8) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Nefedov, O. M.;
Walsh, R. Chem. Phys. Lett. 1996, 260, 433.
(9) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Faustov, V. I.;
Promyslov, V. M.; Nefedov, O. M.; Walsh, R. Phys. Chem. Chem. Phys.
2002, 4, 5079.
(10) Becerra, R.; Walsh, R. J. Organomet. Chem. 2001, 636, 49.
(11) Becerra, R.; Walsh, R. Phys. Chem. Chem. Phys. 2002, 4, 6001.
(12) Alexander, U. N.; King, K. D.; Lawrance, W. D. Chem. Phys. Lett.
2000, 319, 529.
(38) Lollmahomed, F.; Leigh, W. J. Organometallics 2009, 28, 3239.
(39) Leigh, W. J.; Harrington, C. R. J. Am. Chem. Soc. 2005, 127, 5084.
(40) Huck, L. A.; Leigh, W. J. Organometallics 2007, 26, 1339.
(41) Huck, L. A.; Leigh, W. J. Organometallics 2009, 28, 6777.
(42) Leigh, W. J.; Dumbrava, I. G.; Lollmahomed, F. Can. J. Chem.
2006, 84, 934.
(43) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Lee, V. Y.; Nefedov,
O. M.; Walsh, R. Chem. Phys. Lett. 1996, 250, 111.
(44) Saito, K.; Obi, K. Chem. Phys. 1994, 187, 381.
(45) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Faustov, V. I.;
Nefedov, O. M.; Walsh, R. Can. J. Chem. 2000, 78, 1428.
(46) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Krylova, I. V.;
Nefedov, O. M.; Walsh, R. J. Phys. Chem. A 2007, 111, 1434.
(47) Perdew, J. P.; Burke, K. Int. J. Quantum Chem. 1996, 57, 309.
(48) Gibbon, G. A.; Rousseau, Y.; Van Dyke, C. H.; Mains, G. J.
Inorg. Chem. 1966, 5, 114.
(13) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Faustov, V. I.;
Krylova, I. V.; Nefedov, O. M.; Promyslov, V. M.; Walsh, R. Phys. Chem.
Chem. Phys. 2004, 6, 3370.
(49) Angus, P. C.; Stobart, S. R. J. Chem. Soc., Dalton Trans. 1975,
1975, 2342.
(50) Farkas, A. Orthohydrogen, Parahydrogen and Heavy Hydrogen;
(14) Becerra, R.; Walsh, R. Phys. Chem. Chem. Phys. 2009, 11, 3539.
(15) Becerra, R.; Walsh, R. Phys. Chem. Chem. Phys. 1999, 1, 5301.
(16) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Faustov, V. I.;
Nefedov, O. M.; Walsh, R. Phys. Chem. Chem. Phys. 2001, 3, 184.
(17) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Faustov, V. I.;
Nefedov, O. M.; Walsh, R. J. Am. Chem. Soc. 1998, 120, 12657.
(18) Alexander, U. N.; Trout, N. A.; King, K. D.; Lawrance, W. D.
Chem. Phys. Lett. 1999, 299, 291.
(19) Becerra, R.; Egorov, M. P.; Krylova, I. V.; Nefedov, O. M.;
Walsh, R. Chem. Phys. Lett. 2002, 351, 47.
(20) Becerra, R.; Boganov, S. E.; Egorov, M. P.; Faustov, V. I.;
Krylova, I. V.; Nefedov, O. M.; Promyslov, V. M.; Walsh, R. Phys. Chem.
Chem. Phys. 2007, 9, 4395.
Cambridge University Press: London, 1935.
(51) Benoit, H.; Piejus, P. C. R. Acad. Sci. Ser. B 1967, 265B, 101.
(52) In addition to the resonances due to HD and 8-d, the spectrum
of the 4/AcOD photolysate also showed evidence for the formation of
small amounts of H2 and all-protiated 8, in similar relative yields to what
was obtained in the 4/AcOH photolysis. We thus conclude that the
formation of H2 in the 4/AcOD experiment is due to the presence of
small amounts of AcOH in the mixture and not to unimolecular
elimination of H2 from 4.
(53) Massol, M.; Satge, J.; Riviere, P.; Barrau, J. J. Organomet. Chem.
1970, 22, 599.
(21) Alexander, U. N.; King, K. D.; Lawrance, W. D. Phys. Chem.
Chem. Phys. 2003, 5, 1557.
(54) Gibbon, G. A.; Wang, J. T.; Van Dyke, C. H. Inorg. Chem. 1967,
6, 1989.
(22) (a) Thimer, K. C.; Al-Rafia, S. M. I.; Ferguson, M. J.; McDonald,
R.; Rivard, E. Chem. Commun. 2009, 7119. (b) Al-Rafia, S. M. I.; Malcolm,
A. C.; Liew, S. K.; Ferguson, M. J.; Rivard, E. J. Am. Chem. Soc. 2011,
133, 777.
(55) Heaven, M. W.; Metha, G. F.; Buntine, M. A. J. Phys. Chem. A
2001, 105, 1185.
(56) Heaven, M. W.; Metha, G. F.; Buntine, M. A. Aust. J. Chem.
2001, 54, 185.
(23) (a) Rupar, P. A.; Jennings, M. C.; Ragogna, P. J.; Baines, K. M.
Organometallics 2007, 26, 4109. (b) Ghadwal, R. S.; Roesky, H. W.;
Merkel, S.; Stalke, D. Chem. -Eur. J. 2010, 16, 85. (c) Li, J.; Merkel, S.;
Henn, J.; Meindl, K.; D€oring, A.; Roesky, H. W.; Ghadwal, R. S.; Stalke,
D. Inorg. Chem. 2010, 49, 775.
(57) Moiseev, A. G.; Leigh, W. J. Organometallics 2007, 26, 6268.
(58) Moiseev, A. G.; Coulais, E.; Leigh, W. J. Chem.—Eur. J. 2009,
15, 8485.
(59) Mochida, K.; Kayamori, T.; Wakasa, M.; Hayashi, H.; Egorov,
M. P. Organometallics 2000, 19, 3379.
(24) Goldberg, D. E.; Hitchcock, P. B.; Lappert, M. F.; Thomas,
K. M.; Thorne, A. J.; Fjeldberg, T.; Haaland, A.; Schilling, B. E. R.
J. Chem. Soc., Dalton Trans. 1986, 1986, 2387.
(25) Grev, R. S.; Schaefer, H. F., III; Baines, K. M. J. Am. Chem. Soc.
1990, 112, 9458.
(26) Ricca, A.; Bauschlicher, C. W. J. Phys. Chem. A 1999, 103, 11121.
(27) Chen, W. C.; Su, M. D.; Chu, S. Y. Organometallics 2001,
20, 564.
(60) Nag, M.; Gaspar, P. P. Organometallics 2009, 28, 5612.
(61) Neumann, W. P.; Sakurai, H.; Billeb, G.; Brauer, H.; Kocher, J.;
Viebahn, S. Angew. Chem., Int. Ed. Engl. 1989, 28, 1028.
(62) Sakai, S. Int. J. Quantum Chem. 1998, 70, 291.
(63) Su, M. D.; Chu, S. Y. J. Am. Chem. Soc. 1999, 121, 11478.
(64) Birukov, A. A.; Faustov, V. I.; Egorov, M. P.; Nefedov, O. M.
Russ. Chem. Bull. Int. Ed. 2005, 54, 2003.
(65) Joo, H.; Kraka, E.; Quapp, W.; Cremer, D. Mol. Phys. 2007,
(28) Becerra, R.; Harrington, C. R.; Gaspar, P. P.; Leigh, W. J.;
Vargas-Baca, I.; Walsh, R.; Zhou, D. J. Am. Chem. Soc. 2005, 127, 17469.
(29) Trinquier, G. J. Am. Chem. Soc. 1990, 112, 2130.
(30) Grev, R. S.; Schaefer, H. F., III Organometallics 1992, 11,
3489.
105, 2697.
(66) Summerscales, O. T.; Jimenez-Halla, J. O. C.; Merino, G.;
Power, P. P. J. Am. Chem. Soc. 2011, 133, 180.
(67) (a) Grimme, S. J. Chem. Phys. 2006, 124, 034108. (b) Grimme,
S.; Neese, F. J. Chem. Phys. 2007, 127, 154116.
(31) Li, Q.-S.; Lu, R.-H.; Xie, Y.; Schaefer, H. F. I. J. Comput. Chem.
2002, 23, 1642.
(32) Carrier, W.; Zheng, W.; Osamura, Y.; Kaiser, R. I. Chem. Phys.
2006, 330, 275.
(33) Wang, X.; Andrews, L.; Kushto, G. P. J. Phys. Chem. A 2002,
106, 5809.
(68) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158.
(69) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys.
2007, 126, 084108.
(70) Becerra, R.; Harrington, C. R.; Leigh, W. J.; Kefala, L. A.; Walsh,
R. Organometallics 2006, 25, 4439.
(71) Neumann, W. P. Chem. Rev. 1991, 91, 311.
(34) Leigh, W. J.; Harrington, C. R.; Vargas-Baca, I. J. Am. Chem. Soc.
2004, 126, 16105.
10534
dx.doi.org/10.1021/ja201190b |J. Am. Chem. Soc. 2011, 133, 10523–10534