Competing germene and germylene extrusion from photolysis of 1,1-diarylgermacyclobutanes. Substituent effects on germene reactivity

Article abstract of DOI:10.1021/om800574s

Direct irradiation of 1,1-diphenyl-, 1,1-bis[4-(trifluoromethyl)phenyl]-, and 1,1-bis[3,5-bis(trifluoromethyl)phenyl]germacyclobutanes (2, 4, and 5, respectively) in methanolic C₆D₁₂ solution affords products consistent with the competing formation of the corresponding 1,1-diarylgermenes and diarylgermylenes, along with ethylene and cyclopropane. The relative yields of the two Ge-containing primary products (germene:germylene) vary with the extent of CF₃ substitution on the aryl rings, decreasing in the order 2 > 4 > 5. As was reported previously for 2, laser flash photolysis of 4 and 5 in hexane, acetonitrile, or tetrahydrofuran solution allows the detection of the corresponding transient 1,1 -diarylgermenes (6 and 7, respectively), which have been identified on the basis of their UV/vis spectra (λ_max ～325 nm) and quenching studies with MeOH, tert-butyl alcohol (t-BuOH), acetic acid (AcOH), n-butylamine (M-BuNH₂), and acetone. In carefully dried hexane solution, weak transient absorptions assignable to the corresponding germylenes and their respective (digermene) dimers are also observed; in the case of 5, these assignments have been confirmed by the results of steady-state and laser photolysis experiments with 1,1-bis[3,5-bis(trifluoromethyl)phenyl]-2,3- dimethyl-1-germacyclopent-3-ene (14c), which affords the germylene exclusively, in substantially higher quantum yield. The reactivities of the germenes toward each of the various substrates studied vary modestly with aryl substituent, increasing in the order acetone < t-BuOH < MeOH ≈ n-BuNH₂ < AcOH. The rate constants increase with increasing trifluoromethyl substitution in the cases of alcohol and acetone addition but decrease correspondingly in the case of AcOH addition. Substrate acidity thus plays a much more dominant role in the reactions of the Ge=C bond with nucleophilic reagents than is the case with the homologous silene derivatives, whose reaction kinetics are controlled primarily by substrate nucleophilicity.

Full text of DOI:10.1021/om800574s

5948
Organometallics 2008, 27, 5948–5959
Competing Germene and Germylene Extrusion from Photolysis of
1,1-Diarylgermacyclobutanes. Substituent Effects on Germene
Reactivity
William J. Leigh,* Gregory D. Potter, Lawrence A. Huck, and Adroha Bhattacharya
Department of Chemistry, McMaster UniVersity, 1280 Main Street West,
Hamilton, Ontario, Canada L8S 4M1
ReceiVed June 22, 2008
Direct irradiation of 1,1-diphenyl-, 1,1-bis[4-(trifluoromethyl)phenyl]-, and 1,1-bis[3,5-bis(trifluorom-
ethyl)phenyl]germacyclobutanes (2, 4, and 5, respectively) in methanolic C6D12 solution affords products
consistent with the competing formation of the corresponding 1,1-diarylgermenes and diarylgermylenes,
along with ethylene and cyclopropane. The relative yields of the two Ge-containing primary products
(germene:germylene) vary with the extent of CF3 substitution on the aryl rings, decreasing in the order
2 > 4 > 5. As was reported previously for 2, laser flash photolysis of 4 and 5 in hexane, acetonitrile, or
tetrahydrofuran solution allows the detection of the corresponding transient 1,1-diarylgermenes (6 and 7,
respectively), which have been identified on the basis of their UV/vis spectra (λmax ∼325 nm) and quenching
studies with MeOH, tert-butyl alcohol (t-BuOH), acetic acid (AcOH), n-butylamine (n-BuNH2), and
acetone. In carefully dried hexane solution, weak transient absorptions assignable to the corresponding
germylenes and their respective (digermene) dimers are also observed; in the case of 5, these assignments
have been confirmed by the results of steady-state and laser photolysis experiments with 1,1-bis[3,5-
bis(trifluoromethyl)phenyl]-2,3-dimethyl-1-germacyclopent-3-ene (14c), which affords the germylene
exclusively, in substantially higher quantum yield. The reactivities of the germenes toward each of the
various substrates studied vary modestly with aryl substituent, increasing in the order acetone < t-BuOH
< MeOH ≈ n-BuNH2 < AcOH. The rate constants increase with increasing trifluoromethyl substitution
in the cases of alcohol and acetone addition but decrease correspondingly in the case of AcOH addition.
Substrate acidity thus plays a much more dominant role in the reactions of the GedC bond with
nucleophilic reagents than is the case with the homologous silene derivatives, whose reaction kinetics
are controlled primarily by substrate nucleophilicity.
solution,9-16 relatively little has been done with their
Introduction
organogermanium counterparts, germenes and digermenes.17-19
Some time ago, we published the results of a study of the
prototypical 1,1-diarylgermene derivative, 1,1-diphenylgermene
(1), by steady-state and laser flash photolysis methods,
employing the UV photolysis of 1,1-diphenylgermacyclobu-
tane (2; eq 1) to generate this highly reactive molecule in
solution.17 Its spectroscopic properties and reactivity toward
aliphatic alcohols and acetic acid were studied in various
There has been great interest in the chemistry of homo-
and heteronuclear heavier group 14 ethylene analogues,
focusing principally on the synthesis and structural charac-
terization of derivatives stabilized kinetically with sterically
bulky substituents.1-8 Much less attention has been paid to
the study of simpler derivatives that are typically transient
species in fluid solution, whose direct detection and study
typically require the use of fast time-resolved spectroscopic
methods and a photochemical reaction for the formation of
the transient species of interest. While a significant body of
experimental data now exists on the kinetics and mecha-
nisms of the reactions of transient silenes and disilenes in
(9) Morkin, T. L.; Owens, T. R.; Leigh, W. J.; Rappoport, Z.; Apeloig,
Y. In The Chemistry of Organic Silicon Compounds; Rappoport, Z., Ed.;
Wiley: New York, 2001; Vol. 3, pp 949-1026.
(10) Leigh, W. J.; Li, X. Organometallics 2002, 21, 1197.
(11) Leigh, W. J.; Li, X. J. Phys. Chem. A 2003, 107, 1517.
(12) Owens, T. R.; Harrington, C. R.; Pace, T. C. S.; Leigh, W. J.
Organometallics 2003, 22, 5518.
(13) Samuel, M. S.; Jenkins, H. A.; Hughes, D. W.; Baines, K. M.
Organometallics 2003, 22, 1603.
(14) Owens, T. R.; Grinyer, J.; Leigh, W. J. Organometallics 2005, 24,
2307.
(1) Wiberg, N. J. Organomet. Chem. 1984, 273, 141.
(2) Baines, K. M.; Stibbs, W. G. AdV. Organomet. Chem. 1996, 39,
275.
(3) Driess, M.; Gru¨tzmacher, H. Angew. Chem., Int. Ed. Engl. 1996,
35, 828.
(4) Escudie, J.; Couret, C.; Ranaivonjatovo, H. Coord. Chem. ReV. 1998,
180, 565.
(5) Escudie, J.; Ranaivonjatovo, H. AdV. Organomet. Chem. 1999, 44,
113.
(6) Power, P. P. Chem. ReV. 1999, 99, 3463.
(7) Tokitoh, N.; Okazaki, R. AdV. Organomet. Chem. 2001, 47, 121.
(8) Tokitoh, N.; Okazaki, R.; Rappoport, Z. In The Chemistry of Organic
Germanium, Tin and Lead Compounds; Rappoport, Z., Ed.; Wiley: New
York, 2002; Vol. 2, Part 1, pp 843-901.
(15) Leigh, W. J.; Owens, T. R.; Bendikov, M.; Zade, S. S.; Apeloig,
Y. J. Am. Chem. Soc. 2006, 128, 10772.
(16) Milnes, K. K.; Baines, K. M. Organometallics 2007, 26, 2392.
(17) Toltl, N. P.; Leigh, W. J. J. Am. Chem. Soc. 1998, 120, 1172.
(18) Leigh, W. J. Pure Appl. Chem. 1999, 71, 453.
(19) Huck, L. A.; Leigh, W. J. Organometallics 2007, 26, 1339.
(20) Leigh, W. J.; Bradaric, C. J.; Kerst, C.; Banisch, J. H. Organome-
tallics 1996, 15, 2246.
(21) Leigh, W. J.; Sluggett, G. W. J. Am. Chem. Soc. 1994, 116, 10468.
(22) Nagase, S.; Kudo, T. Organometallics 1984, 3, 324.
(23) Bradaric, C. J.; Leigh, W. J. Organometallics 1998, 17, 645.
10.1021/om800574s CCC: $40.75
2008 American Chemical Society
Publication on Web 10/17/2008

Photolysis of 1,1-Diarylgermacyclobutanes
Organometallics, Vol. 27, No. 22, 2008 5949
solvents at ambient temperature, and the results were
compared to analogous results published earlier for the silicon
homologue 1,1-diphenylsilene (3; eq 2).20 While both the
silene and germene react with these substrates by formal 1,2-
addition to yield the expected1-5 alkoxy- or acetoxymetallane
derivative, our kinetic studies showed that the mechanisms
for reaction of the homologous metallenes with alcohols differ
fundamentally; the reaction proceeds with strict second-order
kinetics (first order in ROH) in the case of the silene20 but
with overall third-order kinetics (second order in ROH) in
the case of the germene.17 The second-order dependence on
alcohol concentration has also been observed with some
silenes that have been studied, specifically those bearing
electronically stabilizing substituents at the silicon and carbon
atoms of the SidC bond.14,15,21 The results suggest that
germenes are substantially less reactive toward nucleophilic
reagents than silenes of homologous structure, in contrast to
early theoretical predictions of close similarities in the
reactivities of the parent compounds (H2MdCH2; M ) Si,
Ge).22 This generalization is also borne out in other aspects
of the reactivities of 1 and 3; for example, while silene 3
reacts with acetone (via ene addition) with a rate constant
of ca. 4 × 108 M-1 s-1 in hydrocarbon solvents at 25 °C,23
no evidence of any reactivity whatsoever toward this substrate
could be detected for germene 1 by either steady-state or
flash photolysis methods.20 Similarly, the germene was found
to react at least three orders of magnitude more slowly than
3 with n-butylamine,18 a much more powerful nucleophile
than simple aliphatic alcohols.
Results
Compounds 4 and 5 were synthesized by reaction of 1,3-
bis(bromomagnesio)propane with the corresponding diarylger-
manium dichlorides (10 and 11, respectively; eq 4). The latter
were prepared from germanium tetrachloride and the corre-
sponding aryl Grignard reagents and were employed (without
purification) in the final step of the procedure as mixtures
contaminated with small amounts of the corresponding aryl-
trichloro- and triarylchlorogermanes. Repeated column chro-
matography of the reaction mixtures followed by several
recrystallizations from methanol (MeOH) afforded pure samples
of 4 and 5 in overall isolated yields of 31% and 21%,
respectively. They were identified on the basis of their 1H, 13C,
and 19F NMR, infrared, and mass spectra and combustion
analysis.
Steady-State Photolysis Studies. Photolysis (254 nm) of 4
and 5 as ca. 0.04 M solutions in cyclohexane-d12 containing
MeOH (0.3 M) led to the formation of four products in both
cases. Two of these were common to both photolysates and
were identified as ethylene and cyclopropane by 1H NMR
spectroscopy, after the crude reaction mixtures were spiked with
small amounts of the authentic materials. The other two products
in each photolysate were identified as the methoxygermanes 12b
and 13b from 4 and 12c and 13c from 5 (eq 5). Preliminary
identification of these compounds was made on the basis of 1H
NMR and GC/MS analysis of the crude photolysates and was
1
confirmed by GC or H NMR analysis of the crude reaction
mixtures after they were spiked with independently prepared
samples. Chemical yields were determined relative to consumed
germacyclobutane from the slopes of concentration vs time plots,
constructed by integration of 1H NMR spectra recorded at
selected time intervals between 0 and ca. 5% conversion of 4
and 5 (see Figure S1 in the Supporting Information). The
consistently lower yields of ethylene relative to 12 are presum-
ably due to its higher volatility as compared to that of
cyclopropane, whose yields always corresponded quite closely
to those of 13 (as expected). The photolyses were carried out
in parallel with the photolysis of a 0.041 M solution of 3,4-
dimethyl-1,1-diphenylgermacyclopent-3-ene (14a) in C6D12
containing 0.3 M MeOH, for which the quantum yield for
formation of 13a is Φ13a ) 0.55 ( 0.07.24 The results afforded
quantum yields for product formation of Φ12b ) 0.13 ( 0.03
and Φ13b ) 0.09 ( 0.02 from 4 and Φ12c ) 0.07 ( 0.01 and
Φ13c ) 0.11 ( 0.03 from 5.
In this paper, we report the results of a study of the
photochemistry of two new 1,1-diarylgermacyclobutane deriva-
tives (4 and 5) and of the reactivities of the corresponding 1,1-
diarylgermenes (6 and 7) by laser flash photolysis methods. The
goals of this work were to probe the extent to which polar aryl
substituents affect the kinetics of the reactions of the GedC
bond in 1 toward nucleophilic substrates and to see whether
classic substituent effects might provide useful mechanistic
information on these reactions, thus allowing more precise
comparisons to the reactivities of silenes of homologous
structure. A key expectation was that the photochemistry of 4
and 5 would proceed in a fashion similar to that for 2, from
which germene 1 is produced with ca. 25% quantum efficiency
and in a chemical yield in excess of 70%.17 As the study
unfolded, however, it quickly became apparent that not one but
two modes of reaction contribute prominently to the photo-
chemistry of 1,1-diarylgermacyclobutane derivatives: (2+2)
cycloreversion to the corresponding 1,1-diarylgermene and
ethylene and (1+3) cycloreversion to the corresponding di-
arylgermylene (8 and 9) and cyclopropane (see eq 3). The latter
reaction pathway was not detected in our original study of 2;17
therefore, we have also reexamined the photochemistry of this
compound in some detail.
1
The H NMR and GC/MS characteristics of methoxyger-
manes 13b,c were identical with those of the major Ge-
containing products of photolysis of the 1,1-diaryl-3,4-
dimethylgermacyclopentene derivatives 14b19 and 14c (eq 6)
(24) Leigh, W. J.; Harrington, C. R.; Vargas-Baca, I. J. Am. Chem. Soc.
2004, 126, 16105.

5950 Organometallics, Vol. 27, No. 22, 2008
Leigh et al.
Photolysis of a 0.05 M solution of 2 in C6D12 containing
DMB (0.04 M) led to the formation of digermacyclobutane 17a
and germacyclopentene 14a in yields of 72% and 21%,
respectively, along with ethylene and cyclopropane (see eq 9);
the use of a deficiency of the diene in the experiment was
predicated by its higher extinction coefficient at 254 nm
compared to that of 2. The two Ge-containing products were
under similar conditions. Compound 14c was prepared and its
photochemistry studied in the course of the present work.
Photolysis of the compound in methanolic C6D12 was found to
afford 13c and 2,3-dimethyl-1,3-butadiene (DMB) in chemical
yields of 83% and 88%, respectively, as established by
concentration vs time plots over the 0-20% conversion range
in 14c (eq 6). The quantum yield for formation of 13c from
this compound was determined to be Φ13c ) 0.45 ( 0.09, again
using the photolysis of 14a under the same conditions as the
actinometer (Figure S3, Supporting Information).
1
identified by comparison of their H NMR and mass spectral
data to those of authentic samples; their yields relative to
consumed 2 were determined from the slopes of concentration
vs time plots, determined from 1H NMR spectra of the reaction
mixture recorded throughout the photolysis. While the formation
of one or more additional (minor) unidentified products was
1
evident in the H NMR spectrum of the photolyzed mixture
(see Figure S5 in the Supporting Information), no evidence for
the formation of the seemingly most likely candidate(s)sa
cycloadduct between the diene and germene 1scould be
obtained by GC/MS analysis of the crude photolysate. In any
event, the fact that the head-to-tail dimer 17a remains the major
germene-derived product in the presence of 0.04 M DMB
indicates that the diene is a relatively inefficient trapping agent
for transient germenes, in contrast to its well-known effective-
ness as a germylene scavenger.8,25
The observation of significant yields of 13b,c in the photo-
product mixtures from 4 and 5 prompted us to re-examine the
photochemistry of the parent compound (2) under similar
conditions, as the corresponding product (diphenylmethoxyger-
mane, 13a) was not detected as a photoproduct in our original
1
study of 2.17 Indeed, H NMR analysis of a photolyzed 0.041
M solution of 2 in C6D12 containing 0.3 M MeOH revealed
that both 12a and 13a are formed as products, in chemical yields
of ca. 60% and 21%, respectively, along with ethylene and
cyclopropane (eq 7 and Figure S2 in the Supporting Informa-
tion). Analogous results were obtained using acetic acid (AcOH;
0.05 M) as trapping agent, which (at low conversions) cleanly
afforded the acetoxygermanes 15a and 16a in yields of ca. 70%
and 25%, respectively (eq. 8 and Figure S4 in the Supporting
Information). Our failure to detect the germylene-derived
products (13a and 16a) in our earlier study is not really
understood but could be due to the fact that the original
experiments were carried out using dichloromethane as an
internal NMR integration standard, which may have facilitated
decomposition of the compounds during photolysis due to some
radical chain process. On the other hand, it is clear that
cyclopropane was formed in our original study of 2, as a minor
photoproduct characterized in the 1H NMR spectra by a singlet
at δ 0.20 (in C6D12); unfortunately, however, we did not pursue
its identification. The quantum yields for the formation of 12a
and 13a from the photolysis of 2 were also determined (using
the photolysis of 14a under the same conditions as actinometer;
Φ13a ) 0.55 ( 0.0724) and found to be Φ12a ) 0.17 ( 0.04
and Φ13a ) 0.060 ( 0.015. The results are in good agreement
with the quantum yield for photolysis of 2 reported in our
original study (Φ-2a ) 0.24 ( 0.04).17
Compounds 4 and 5 were also photolyzed in deoxygenated
C6D12 containing acetone (0.2 M), with the mixtures being
monitored by 1H NMR spectroscopy. The resulting spectra after
ca. 50% conversion were quite complex in both cases and
showed broad baseline absorptions in the aromatic and δ 0.5-2
regions of the spectra, suggestive of the formation of polymeric
material. Ethylene and cyclopropane were the only products that
could be conclusively identified in the two photolysates, and
no resonances other than that due to ethylene could be detected
in the vinylic region of the spectra in either case.
The results of the steady-state photolysis experiments with
2, 4, and 5 under various conditions show that the photochem-
istry of these compounds is characterized by two competing
reaction channels, one leading to the formation of the corre-
sponding 1,1-diarylgermene and ethylene by formal (2+2)
cycloreversion and one leading to the corresponding diarylger-
mylene and cyclopropane by formal (1+3) cycloreversion (eq
10). The efficiency of product formation from 2 is unaffected
by added DMB, which is expected to be a good triplet quencher
for compounds of this type. From this it can be concluded that
both reaction pathways result from the excited singlet states of
the germacyclobutanes. Most interestingly, the partitioning
between the (2+2) and (1+3) cycloreversion pathways varies
with the extent of trifluoromethyl substitution, from ca. 3:1 in
favor of the (2+2) pathway in the case of the parent compound
2 to ca. 1:2 in favor of the (1+3) pathway in the case of the
bis[3,5-bis(trifluoromethyl)phenyl] derivative 5. It should be
(25) Neumann, W. P.; Weisbeck, M. P.; Wienken, S. Main Group Met.
Chem. 1994, 17, 151.

Photolysis of 1,1-Diarylgermacyclobutanes
Organometallics, Vol. 27, No. 22, 2008 5951
Figure 1. (a) Transient absorption spectra recorded 0-0.32 µs (9), 20.2-20.5 µs (O), and 88.3-88.6 µs (4) after the laser pulse, by laser
flash photolysis of a 0.4 mM solution of 4 in deoxygenated hexane at 25 °C. The inset shows transient decay traces recorded at 290, 330,
and 470 nm under the two sets of conditions. (b) Difference spectra calculated from the spectra of Figure 1a: (9) (0-0.32 µs) - (88.3-88.6
µs); (O) (20.2-20.5 µs) - (88.3-88.6 µs).
Figure 2. (a) Transient absorption spectra recorded (9) 0-0.32 µs, (O) 20.2-20.5 µs, and (4) 88.3-88.6 µs after the laser pulse, by laser
flash photolysis of a 0.6 mM solution of 5 in deoxygenated hexane at 25 °C. The inset shows transient decay traces recorded at 290, 330,
and 470 nm under the two sets of conditions. (b) Difference spectra calculated from the spectra of Figure 2a: (9) (0-0.32 µs) - (88.3-88.6
µs); (O) (20.2-20.5 µs) - (88.3-88.6 µs).
noted that propene could not be detected as a coproduct in any
of the three cases, indicating that cyclopropane is the exclusive
organic product of the (1+3) photocycloreversion process. In
contrast, the thermal (gas phase) decomposition of 1,1-dimeth-
ylgermacyclobutane is known to afford ethylene and both
cyclopropane and propene, along with products derived from
1,1-dimethylgermene and dimethylgermylene.26 In the thermal
reaction, however, the formation of the two C3H6 hydrocarbons
was proposed to result from fundamentally different reaction
pathways.26
nm range, along with much weaker transient absorptions in the
400-520 nm region of the spectra. For example, Figure 1a
shows transient UV/vis spectra recorded with 4 in deoxygenated
hexane over three selected time windows after the laser pulse;
analogous data for 5 under similar conditions are shown in
Figure 2a. In both cases, a relatively strong transient absorption
centered at λmax ∼330 nm with lifetime τ ∼20 µs is produced
with the laser pulse, superimposed on a second, much longer
lived absorption that is centered at shorter wavelengths and has
a lifetime of a few milliseconds or greater. The presence of O2
had little to no effect on the intensities or lifetimes of the 330
nm transients but reduced the intensities of both the long-lived
absorptions and the weak absorptions above 400 nm; these
effects were more pronounced in O2-saturated solution than in
air-saturated solution; examples of transient spectra recorded
for 4 and 5 in air-saturated hexane are included in the Supporting
Information (Figure S6). Addition of 1 mM AcOH led to a
distinct shortening of the lifetimes of the 330 nm transients and
eliminated the others entirely; this was also observed in the
presence of MeOH, tert-butyl alcohol (t-BuOH), and acetone
(vide infra). The spectra and lifetimes of the 330 nm species
are similar to those reported previously for 1,1-diphenylgermene
(1) under similar conditions (τ ∼20 µs; λmax 325 nm),17 and
we thus assign them to the 1,1-diarylgermenes 6 and 7,
respectively; the sensitivity of their lifetimes to nucleophilic
Laser Flash Photolysis Studies. Laser flash photolysis
experiments were carried out using a KrF excimer laser for
excitation (248 nm, ∼25 ns, 100 mJ) and a thermostated flow
system for sample delivery. Flash photolysis of deoxygenated
solutions of 4 (ca. 0.53 mM) and 5 (ca. 0.70 mM) in dry hexane
afforded strong transient absorptions centered in the 280-340
(26) Namavari, M.; Conlin, R. T. Organometallics 1992, 11, 3307.

5952 Organometallics, Vol. 27, No. 22, 2008
Leigh et al.
Figure 3. Transient absorption spectra recorded (0) 0.1-0.6 µs and (O) 62.1-62.7 µs after the laser pulse, by laser flash photolysis of
deoxygenated MeCN solutions of (a) 4 (0.5 mM) and (b) 5 (0.7 mM). The insets show decay traces recorded at 325 nm.
reagents and the lack of sensitivity to oxygen are also consistent
with the germene assignments. The identities of the long-lived
species absorbing below 300 nm are unknown, but the fact that
they are eliminated in the presence of substrates that are reactive
toward germenes and/or germylenes suggests they are secondary
productsformedbyreactionoftheprimarytransientphotoproducts.
The weak transient absorptions arising in the 400-520 nm
spectral range were significantly stronger from 5 than from 4
and exhibited temporal behavior that was very different from
those of the shorter wavelength absorptions. In both cases,
evidence for two distinct transient species with overlapping
absorption spectra was clearly discernible, one species with λmax
centered at 480-500 nm that was formed with the laser pulse
and decayed relatively rapidly (0.2 < τ < 2 µs), and a second
centered at λmax 440 nm that grew in after the laser pulse and
then decayed over several hundred microseconds with mixed-
order kinetics. Neither species could be detected at all unless
adequate care was taken to dry the solvent and sample delivery
system. We assign the promptly formed species to the ger-
mylenes 8 and 9 and the 440 nm species to their dimerization
products, tetraaryldigermenes 18b,c, on the basis of the similar-
ity of the spectra and temporal behavior to those observed using
the 1,1-diarylgermacyclopent-3-enes 14b19 and 14c (vide infra)
as precursors to 8 and 9, respectively, and the fact that their
formation is quenched upon addition of nucleophilic substrates
that are known to react rapidly with 8 and other related transient
diarylgermylenes.19,27,28 The relative weakness of the signals
observed for 8 and 9 in flash photolysis experiments with 4
and 5 is undoubtedly due to the rather low quantum yields for
their formation from these compounds, coupled with the low
extinction coefficients that characterize the UV/vis absorption
spectra of diarylgermylenes.24,29,30
consist of the sum of the spectra of the corresponding germene
(λmax 330 nm) and germylene (λmax 280, 480-500 nm); it should
be noted that, as would be expected on the basis of the product
studies, the intensity of the germylene absorptions relative to
that of the germene is much higher from 5 than from 4,
consistent with a higher relative yield of germylene from
photolysis of the former compound. Comparison of the first and
second difference spectra shows that, in both cases, the germenes
decay to roughly half their initial concentrations within ca. 20
µs, while the germylene absorptions at 280 and 480-500 nm
are replaced by the longer lived absorptions due to the
corresponding digermenes (λmax 440 nm).
Spectra recorded in deoxygenated acetonitrile (MeCN) and
tetrahydrofuran (THF) solution showed transient absorptions
similar to those observed in hexane solution in the region below
360 nm, but they were devoid of any detectable transient
absorptions in the 400-600 nm range of the spectrum (see
Figure 3), due presumably to rapid reaction (or complexation)
of the germylenes (8 and 9) with the solvent (vide infra). Both
sets of spectra consisted of two overlapping transient absorptions
in MeCN solution, one centered at λmax ∼325 nm with lifetime
τ ∼15 µs and a second with λmax <300 nm and τ g100 µs. As
in the experiments in hexane solution, the contribution from
the longer lived species was significantly reduced in air-saturated
solution, while the lifetimes and intensities of the 325 nm
absorptions were largely unaffected. Addition of alcohols or
acetic acid to the solutions shortened the lifetimes of the 325
nm species (vide infra), consistent with their assignment to
germenes 6 and 7. Transient spectra recorded by flash photolysis
of 4 in deoxygenated tetrahydrofuran (THF) solution were much
cleaner than those in hexane or MeCN, showing a single
transient species exhibiting λmax ∼325 nm and lifetime τ ∼4
µs. A quenching experiment with MeOH (vide infra) confirmed
the identity of the species as germene 6. There was no change
in the position or breadth of the absorption, or the transient
lifetime, upon lowering the temperature of the solution to ca. 4
°C. Thus, we could obtain no evidence for Lewis acid-base
complexation between the germene and the O-donor solvent
over the 4-25 °C temperature range.
Figures 1b and 2b show difference spectra, calculated by
subtracting the 88.3-88.6 µs spectra of Figures 1a and 2a,
respectively, from those recorded immediately and 20.2-20.5
µs after laser excitation of 4 and 5. The two sets of spectra thus
isolate those components of the transient spectra that decay
during the first 90 µs after the (ca. 20 ns) laser pulse. In both
cases, the end-of-pulse difference spectrum can be seen to
The germylene (9) and digermene (18c) assignments in the
laser photolysis experiments with 5 in hexane were confirmed
by laser photolysis of 1.3 mM hexane solutions of germacyclo-
(29) Tokitoh, N.; Kishikawa, K.; Okazaki, R.; Sasamori, T.; Nakata,
N.; Takeda, N. Polyhedron 2003, 21, 563.
(30) Kira, M.; Ishida, S.; Iwamoto, T.; Ichinohe, M.; Kabuto, C.;
Ignatovich, L.; Sakurai, H. Chem. Lett. 1999, 1999, 263.
(27) Leigh, W. J.; Harrington, C. R. J. Am. Chem. Soc. 2005, 127, 5084.
(28) Leigh, W. J.; Lollmahomed, F.; Harrington, C. R.; McDonald, J. M.
Organometallics 2006, 25, 5424.

Photolysis of 1,1-Diarylgermacyclobutanes
Organometallics, Vol. 27, No. 22, 2008 5953
Further support for the germylene and digermene assignments
in the flash photolysis experiments with 14c was obtained upon
addition of acetic acid (AcOH) to the solution, which caused
the lifetime of the prompt (germylene) absorption to be
shortened and the intensities of the signals centered at 350 and
440 nm to be reduced and eventually eliminated, all in
proportion to the concentration of added substrate. Transient
decays recorded at 500 nm in the presence of 0.2-0.8 mM
AcOH decayed cleanly to the baseline with pseudo-first-order
Figure 4. Transient absorption spectra recorded (9) 58-70 ns, (O)
0.32-0.34 µs, and (b) 9.6-9.9 µs after the laser pulse, by laser
flash photolysis of a 3 mM solution of 1,1-diaryl-3,4-dimethylger-
macyclopent-3-ene 14c in dry, deoxygenated hexane at 25 °C. The
inset shows transient absorption vs time profiles recorded at 470,
440, and 350 nm.
kinetics, and a plot of the first-order decay rate constant (kdecay
)
vs [AcOH] was linear, affording a quenching rate coefficient
of kAcOH ) (7.0 ( 0.6) × 109 M-1 s-1 (see eq 12). The value
is in line with what would be expected on the basis of the
reported rate constants for quenching of other substituted
diarylgermylenes by AcOH in this solvent, which established a
value of kAcOH ) (6.2 ( 1.0) × 109 M-1 s-1 for germylene 8
and a Hammett F value of F ≈ +0.2.19 The lifetime of the 440
nm absorption (assigned to digermene 18c) is also shortened in
the presence of the carboxylic acid; a plot of estimated kdecay
values vs [AcOH] over the 0.1-0.8 mM concentration range
in which the species remained detectable afforded a value of
kAcOH ) (2.2 ( 0.8) × 108 M-1 s-1 for the quenching rate
constant. This too is consistent with previously published results
for other transient tetraaryldigermenes.19
pent-3-ene 14c. This compound afforded results consistent with
the sequential formation of three distinct transient species, as
illustrated in Figure 4. The first species is formed within the
laser pulse and exhibits λmax ∼475 nm and lifetime τ ∼200 ns
and the second species (λmax ∼350 nm; τ ∼10 µs) is formed
concomitantly with the decay of the first-formed species, while
the third species (λmax ∼440 nm; τ >200 µs) is formed on a
time scale similar to that of the decay of the second-formed
species. On the basis of our previous results for other related
1,1-diarylgermacyclopent-3-ene derivatives (e.g., 14a,b),19,24 we
assign the first and third species to germylene 9 and digermene
18c, respectively; their temporal behavior was quite similar to
what was observed at 475 and 440 nm, respectively, in the
experiments with 5 in hexane (vide supra). A conclu-
sive assignment cannot be made at the present time for the
intermediate (second-formed) species, which could not be
detected in the experiments with 5 because of overlap with the
strong absorptions due to germene 7 in this region. We initially
suspected that the species might be a complex formed by
reaction of germylene 9 with adventitious water or other
oxygenated impurities in the solvent, as both its spectrum and
its temporal characteristics (as well as the temporal character-
istics of the other two species detected) are similar to what we
have observed previously for diarylgermylenes in hexane
containing small amounts of THF or aliphatic alcohols.19,28
However, flash photolysis experiments with solutions of 14c in
a sample of hexane distilled from LiAlH4 afforded results very
similar to those described above, which employed alumina-dried
hexane. A second possibility is that the species is a complex
formed by interaction of 9 with the CdC bond or fluorine
atoms31 in its precursor (14c); however, experiments carried
out with a more dilute (0.5 mM) solution of 14c afforded weaker
signals, as expected, but transient temporal behavior very similar
to that observed with the higher precursor concentration.
Whatever its identity, the results indicate that the formation of
this species coincides with the decay of the free germylene,
and the species somehow mediates the formation of digermene
18c, either as a direct intermediate or as a nonreactive spectator
in equilibrium with the free germylene (see eq 11). Establishing
its identity is the subject of future work.
kdecay ) k0 + kQ[Q]
(12)
Flash photolysis of deoxygenated solutions of 14b,c in MeCN
afforded strong, long-lived transient absorptions throughout the
280-380 nm range, which behaved similarly to one another
but were completely unlike those observed with 4 and 5 under
similar conditions; the two compounds exhibited prompt, long-
lived absorptions centered at ca. 290 nm that decayed with
complex kinetics over several milliseconds, along with shorter-
lived absorptions centered in the 350-370 nm range that grew
in over the first ca. 20 µs after the pulse and decayed to ca.
25% of their maximum intensities over ca. 200 µs. We include
a representative set of data in the Supporting Information but
defer a full interpretation of the results to a future publication.
The main point of these experiments was to verify that the 325
nm species observed by flash photolysis of 4 and 5 in MeCN
are unique to those compounds and are not due to species
derived from reaction or complexation of the corresponding
germylene coproducts with the nitrile solvent and can thus be
most reasonably assigned to germenes 6 and 7, respectively.
The very pronounced differences in the spectra and temporal
behavior of the transients observed from 14b/14c and 4/5 in
MeCN leave little doubt that the germene assignments are
correct.
As mentioned above, addition of MeOH, tert-butyl alcohol
(t-BuOH), or acetic acid (AcOH) to hexane solutions of 4 and
5 caused the decay of the absorptions due to germenes 6 and 7
to accelerate, consistent with reaction with the substrate in each
case. Transient profiles recorded at 325 nm fit well to the sum
of two first-order exponential decays, the fast component of
which accelerated as the substrate concentration was increased
while the lifetime of the slow component was unaffected; the
intensity of the latter component also diminished with increasing
(31) (a) Bender, J. E. I.; Banaszak Holl, M. M.; Kampf, J. W.
Organometallics 1997, 16, 2743. (b) Bender, J. E. I.; Banaszak Holl, M. M.;
Mitchell, A.; Wells, N. J.; Kampf, J. W. Organometallics 1998, 17, 5166.

5954 Organometallics, Vol. 27, No. 22, 2008
Leigh et al.
Figure 5. Plots of the pseudofirst order rate constants for germene decay (kdecay) for quenching of germenes (4) 1, (O) 6, and (0) 7 versus
[MeOH] in (a) air-saturated hexane solution and (b) air-saturated MeCN solution at 25 °C. The solid lines are the best fits of the data to
eq 13.
Table 1. Rate Constants for Reaction of Germenes 1, 6, and 7 and 1,1-Diphenylsilene (3) with Various Substrates in Hexane Solution at 25 °C
(in Units of 107 M-1 s-1 or 107 M-2 s-1
6
)
substrate
1
7
3
MeOH
t-BuOH
AcOH
n-BuNH2
acetone
(9.3 ( 2.8) + (350 ( 140)[MeOH]
(8.4 ( 3.0) + (830 ( 180)[MeOH]
(2.1 ( 0.2) + (96 ( 4)[t-BuOH]
400 ( 90b
(14.8 ( 0.5) + (1620 ( 340)[MeOH]
(4.3 ( 1.9) + (160 ( 40)[t-BuOH]
370 ( 80b
(18 ( 2) + (4500 ( 2000)[n-BuNH2]
0.162 ( 0.005
190 ( 20d
(2 ( 1) + (12 ( 9)[t-BuOH]a
40 ( 7d
490 ( 100b
5 ( 1c
310 ( 30d
649 ( 12e
38 ( 2f
13.2 ( 0.3
<0.02a
0.081 ( 0.003
a From ref 17. b The limiting ([AcOH]bulk f 0) slopes of plots of kdecay vs [(AcOH)bulk]. c From ref 18. d From ref 20. e From ref 11. f From ref 23, in
isooctane solution.
Table 2. Rate Constants for Reaction of Germenes 1, 6, and 7 and
substrate concentration. Plots of kdecay for the fast-decaying
1,1-Diphenylsilene (3) with MeOH(D) and AcOH(D) in MeCN
component vs substrate (“Q”) concentration exhibited positive
curvature for the two alcohols; examples are shown in Figure
5a for the quenching of 6 and 7 by MeOH. For comparison,
Figure 5a also shows the results of a similar experiment with
germacyclobutane 2, which was repeated as part of the present
work. The latter data are considerably more detailed than those
published previously for this system but are consistent with what
we found in our earlier study.17
The data of Figure 5a were fit to the quadratic expression of
eq 13, where kROH and k2ROH are the effective second- and third-
order rate coefficients, respectively, for reaction of the three
transient germenes with MeOH and t-BuOH in hexane solution.
The results of these analyses are given in Table 1. Though the
errors are large in several cases, the data clearly indicate that
the reaction is dominated by the overall third-order (second order
in MeOH) pathway, with the rate coefficients increasing very
modestly with increasing electron-withdrawing power of the
substituents on the aryl rings. The variation in the third-order
rate coefficients with substituent is larger for t-BuOH, which is
the less reactive of the two alcohols by a significant margin.
Solution at 25 °C
substrate
1
6
7
k2MeOH/107 M-2 s-1
k2MeOD/107 M-2 s-1
kAcOH/107 M-1 s-1
kAcOD/107 M-1 s-1
2.3 ( 0.1
2.98 ( 0.07
2.15 ( 0.06
53 ( 2
5.3 ( 0.2
3.45 ( 0.09
46 ( 1
64 ( 4
51 ( 1
limiting ([AcOH]bulk f 0) slopes of the plots of kdecay vs bulk
AcOH concentration for the three derivatives are given in Table
1.
Addition of MeOH or AcOH to MeCN solutions of 4 and 5
had effects similar to those in hexane solution; the lifetimes of
the germene absorptions at 325 nm were shortened and the
intensities of the underlying long-lived absorptions were reduced
as the concentration of added substrate was increased. As in
the experiments in hexane solution, pseudo-first-order rate
constants for decay of the major (germene) component were
determined by nonlinear least-squares fitting of the data to the
sum of two exponential decays. Plots of kdecay vs [MeOH] again
exhibited strong positive curvature but fit acceptably to the
quadratic dependence on [MeOH] of eq 13. The data are shown
in Figure 5b along with corresponding data for quenching of
germene 1 by the alcohol, which are in excellent agreement
with our earlier published data.17 The analyses afforded second-
order rate coefficients indistinguishable from zero within
experimental error in each case, indicative of a pure squared
dependence of the rate on alcohol concentration. The effective
third-order rate coefficients were thus obtained by linear least-
squares analysis of plots of kdecay vs [MeOH],2 which showed
excellent linearity, and are given in Table 2 for the three
compounds. We also examined the quenching of germenes 6
and 7 by MeOD in MeCN, and again the kdecay values varied
linearly with the square of the alcohol concentration. Analysis
kdecay ) k0 + kROH[ROH] + k2ROH[ROH]2
(13)
Plots of kdecay vs [AcOH] exhibited negatiVe curvature for
both 6 and 7 over the 0-1 mM range in added carboxylic acid
(see Figure S10 in the Supporting Information). Quenching of
germene 1 by AcOH was investigated over the same concentra-
tion range and was found to exhibit similar characteristics. It
should be noted that the rate constant reported in ref 17 for the
reaction of 1 with AcOH in hexane was determined by linear
analysis of a much more limited set of kdecay values that were
recorded over the 1-3.5 mM concentration range in added
AcOH and was thus underestimated quite significantly. The

Photolysis of 1,1-Diarylgermacyclobutanes
Organometallics, Vol. 27, No. 22, 2008 5955
Discussion
The kinetic results for quenching of the three 1,1-dia-
rylgermenes by MeOH and t-BuOH in hexane (see Table 1)
and by MeOH in MeCN (see Table 2) indicate that the dominant
reaction pathway for addition of the alcohols to all three
compounds involves two molecules of alcohol in the rate-
determining step for the reaction. The overall third-order rate
constants increase only very modestly with increasing electron-
withdrawing power of the aryl substituents; a three-point
Hammett plot of the data for MeOH in MeCN affords a reaction
constant of F ) +0.19 ( 0.07. The reactions in MeCN are
subject to small but clearly primary deuterium kinetic isotope
effects with both 6 (kH/kD ) 1.39 ( 0.07) and 7 (kH/kD ) 1.5
( 0.1), indicating that proton transfer occurs in the rate-
controlling step.
One possible mechanism that is compatible with these results
is a two-step process involving reversible reaction of the
germene with a single molecule of the alcohol in the initial step,
to form a zwitterionic intermediate (or association complex) that
proceeds to product by rate-controlling transfer of the proton
from the hydroxylic oxygen to the germenic carbon in the
second step, either by unimolecular means (resulting in over-
all second-order kinetics) or by a catalyzed pathway involving
the second alcohol molecule as the catalyst (resulting in overall
third-order kinetics). The mechanism is shown in eq 14. The
two proton-transfer pathways normally compete with one
another in the addition of alcohols to transient silenes, though
they may not both be detectable in time-resolved experiments.9
With particularly electrophilic silenes (such as 3) only the
unimolecular H-transfer pathway is resolvable in kinetic experi-
ments carried out on the microsecond time scale, because it is
so fast that one loses the ability to detect the silene at all at the
higher concentrations necessary for the catalyzed pathway to
become competitive; nevertheless, competition kinetics studies
have verified that the overall third-order pathway does compete
with the second-order path at high alcohol concentrations in
the addition of MeOH to 3.20 The third-order pathway becomes
increasingly dominant as the electrophilicity of the SidC bond
is reduced through appropriate substitution and is the only
detectable mode of reaction in the case of the very weakly
electrophilic silene 1,1-bis(trimethylsilyl)-2-adamantylidenesi-
lane (19).15 In this case, however, experimental and theoretical
evidence suggests that the reaction in hexane solution occurs
concertedly, via reaction with the hydrogen-bonded dimer of
the alcohol.15
Figure 6. Arrhenius plots for the reaction of germenes (O) 1 and
(0) 6 with AcOH in MeCN solution.
of the data afforded values of k2MeOD that were in both cases
consistently smaller than the corresponding values for the
protiated substrate (see Table 2). Finally, quenching of germene
7 by MeOH was also studied in THF solution. The resulting
nonlinear plot of kdecay vs [MeOH] was analyzed according to
eq 13, with k0 constrained to a non-negative value. This afforded
the second- and third-order rate coefficients kMeOH ) (2.6 (
0.9) × 106 M-1 s-1 and k2MeOH ) (3.3 ( 0.3) × 107 M-2 s-1
,
respectively, where the errors are quoted as twice the standard
deviation from nonlinear least-squares analysis of the data.
Linear plots of kdecay vs [AcOH] were obtained in quenching
experiments with the carboxylic acid in MeCN and so were
analyzed according to eq 12; the results of these analyses are
also given in Table 2. We also redetermined the rate constant
for reaction of germene 1 by AcOH in this solvent and obtained
a value of kAcOH ) (6.4 ( 0.4) × 108 M-1 s-1, in reasonable
agreement with the value reported in our earlier study.17
Quenching of germene 6 by AcOD in dry MeCN was also
investigated, and the experiment afforded a rate constant
identical with that obtained with AcOH within experimental
error. Finally, quenching of germenes 1 and 6 by AcOH was
also studied at 45 and 60 °C, leading to values of kAcOH that
increased modestly with increasing temperature in both cases
(see Figure S12 in the Supporting Information). The resulting
(three-point) Arrhenius plots showed excellent linearity (see
Figure 6) and afforded the Arrhenius parameters given in Table
3. For comparison, the published Arrhenius parameters for
reaction of the homologous 1,1-diarylsilene derivatives with
AcOH in the same solvent32 are also given in Table 3.
The lifetimes of 6 and 7 in hexane solution were also found
to be reduced significantly upon addition of acetone (0.05-0.15
M) or n-butylamine (n-BuNH2; 1-25 mM). Plots of kdecay vs
[Q] according to eq 12 were linear in all cases, except for
germene 7 with n-BuNH2, which exhibited a mild positive
curvature (see Figure S14 in the Supporting Information). Table
1 contains the rate coefficients obtained from analysis of the
data according to eqs 12 and 13, along with our previously
reported value for quenching of germene 1 by n-BuNH2 in
hexane.18 Linear plots of kdecay vs [Q] were also obtained for
quenching of 6 and 7 by acetone-d6 and afforded kQ values of
(5.0 ( 0.3) × 105 M-1 s-1 (kH/kD ) 1.62) and (6.8 ( 0.3) ×
105 M-1 s-1 (kH/kD ) 2.38), respectively.
In principle, the catalyzed proton-transfer pathway could
proceedby sequentialdeprotonation-protonationorprotonation-
deprotonation pathways or by a concerted process in which the
two events occur simultaneously. In the case of 1, we proposed
earlier that the deprotonation-protonation sequence is most
(32) Bradaric, C. J.; Leigh, W. J. Can. J. Chem. 1997, 75, 1393.

5956 Organometallics, Vol. 27, No. 22, 2008
Leigh et al.
Table 3. Arrhenius Parameters for the Reactions of Homologous 1,1-Diarylsilenes and 1,1-Diarylgermenes with AcOH in MeCN Solution
metallene
kAcOH/107 M-1 s-1 a
Ea (kcal mol-1
)
log (A/M-1 -1
s
)
∆Sq (eu)
Ph2GedCH2 (1)
64 ( 4
51 ( 2
130 ( 30
220 ( 20
0.96 ( 0.09
1.52 ( 0.05
1.9 ( 0.3
9.5 ( 0.1
9.8 ( 0.1
10.5 ( 0.4
12.0 ( 0.4
-17.2 ( 0.2
-15.7 ( 0.2
-12.5 ( 0.9
-6 ( 2
(4-CF3C6H4)2GedCH2 (6)
Ph2SidCH2 (3)b
(4-CF3C6H4)2SidCH2 (20)b
3.6 ( 0.5
a At 25 °C. b From ref 32.
likely to be correct, on the basis of the observation that the
pseudo-first-order decay of 1 in the presence of MeOH and
t-BuOH in THF solution appeared to exhibit a mixed first- and
second-order dependence on alcohol concentration, in contrast
to the situation in MeCN, where only the second-order
dependence is observed;17 the (pseudo) first-order component
can be interpreted as being due to intervention of the solvent
as a catalyst for deprotonation of the complex. The present
results for quenching of germene 7 (from 5) by MeOH in THF
confirm this behavior at a level of significance higher than we
were able to obtain in the earlier study. In this case, fitting of
the kdecay vs [MeOH] data to eq 13 (with the constraint k0 g 0)
for addition of MeOH to 1, 6, and 7 in hexane solution can be
analyzed taking this explicitly into account, since equilibrium
constants for H-bonded dimer and higher oligomer formation
have been reported for MeOH in hexane solution.42 Equation
16 defines the mechanistic rate constant in this case, where
[(MeOH)2] is the molar concentration of alcohol dimers in
equilibrium with the monomer and higher order oligomers (eq
17). Indeed, plots of kdecay vs [(MeOH)2], the latter calculated
from the bulk concentrations using association constants
extrapolated from the data of Landeck et al.42 (K2 ) 36.1, K3
) 91.2 at 25 °C), exhibit good linearity in all three cases (see
Figure S11 in the Supporting Information) and afford bimo-
lecular rate constants of (2.4 ( 0.2) × 109 M-1 s-1 for 1, (5.5
( 0.1) × 109 M-1 s-1 for 6, and (7.2 ( 0.3) × 109 M-1 s-1
for 7. A three-point Hammett plot of these data affords F )
+0.26 ( 0.05, the same within experimental error as that
obtained for the third-order rate constants for reaction of MeOH
with the three germenes in MeCN solution (vide supra).
leads to second- and third-order rate coefficients of kMeOH
)
(2.6 ( 0.9) × 106 M-1 s-1 and k2MeOH ) (3.3 ( 0.3) × 107
M-2 s-1, respectively. While the third-order rate constant is
only marginally smaller than that obtained in MeCN solution,
the second-order component is of significant magnitude, con-
sistent with the presence of the solvent-catalyzed component
of the reaction. Correction of the second-order rate coefficient
for the solvent concentration affords an estimate of k ≈ 2 ×
105 M-2 s-1 for the overall third-order rate constant for the
solvent-catalyzed reaction pathway, which is ca. 150 times
smaller than that for the MeOH-catalyzed pathway. This
difference, which should reflect the difference in rate constants
for deprotonation of the germene-MeOH complex by THF (in
THF) and MeOH (in MeCN), is broadly consistent with the
difference in the pKa values of protonated THF (pKa 1.1) and
kdecay ) k0 + k1[(MeOH)2]
(16)
(17)
+
K2[ROH]
K3[ROH]
MeOH2 (pKa 2.36) in MeCN solution.33
ROH y
z (ROH)2 y
z (ROH)3
Within the limits of the equilibrium approximation, the
observed Hammett F value for a two-step reaction mechanism
involving a reversibly formed intermediate is equal to the sum
of the F values for the individual steps in the sequence. Thus,
the fact that Foverall ≈ 0.2 indicates that the electronic require-
ments of the individual steps either oppose one another and
almost cancel or that they both are accelerated in a very minor
way by electron-accepting substituents on the aryl rings.
A second mechanistic possibility that needs to be considered
is a concerted process in which the alcohol reacts with the GedC
bond via its hydrogen-bonded dimer (eq 15). The involvement
of hydrogen-bonded oligomers has been proposed or considered
previously in mechanistic studies of a variety of alcohol insertion
or addition reactions, including those with singlet carbenes,34-38
carbocations,38,39 isocyanates,40,41 and silenes.15 The kinetic data
The self-association of alcohols is of course considerably less
favorable in polar, Lewis basic solvents such as MeCN and THF.
Nevertheless, Griller, Liu, and Scaiano proposed that the
reactions of chlorophenylcarbene with MeOH and t-BuOH
proceed via the hydrogen-bonded dimers in both isooctane and
MeCN, to explain nonlinear kinetic plots similar (in the case
of MeOH) to those found in the present work for the three
diarylgermenes.34 Kirmse et al. later considered the same
mechanism in their study of the reactions of alcohols with
siloxycarbenes in MeCN but suggested it was unlikely, citing
the fact that they could find no evidence for the presence of
H-bonded oligomers of MeOH in MeCN by infrared spectros-
copy at concentrations up to 0.4 M.38 On the other hand, the
mechanism has received support from recent theoretical calcula-
tions,35 and a recent mass spectrometric study of self-association
in solutions of MeOH in MeCN shows clear evidence for the
presence of both hydrogen-bonded dimers and trimers in
significant concentrations at a bulk MeOH concentration of 10
vol % (ca. 0.75 M).43 Unfortunately, none of these studies allow
a reasonable estimate to be made for the monomer-dimer
(33) Izutsu, K. Acid-Base Dissociation Constants in Dipolar Aprotic
SolVents; Blackwell Scientific Publications: Oxford, U.K., 1990; IUPAC
Chemical Data Series 35.
(34) Griller, D.; Liu, M. T. H.; Scaiano, J. C. J. Am. Chem. Soc. 1982,
104, 5549.
(35) Pliego, J. R., Jr.; De Almeida, W. B. J. Phys. Chem. A 1999, 103,
3904.
(36) Moss, R. A.; Shen, S.; Hadel, L. M.; Kmiecik-Lawrynowicz, G.;
Wlostowska, J.; Krogh-Jespersen, K. J. Am. Chem. Soc. 1987, 109, 4341.
(37) Du, X. M.; Fan, H.; Goodman, J. L.; Kesselmayer, M. A.; Krogh-
Jespersen, K.; LaVilla, J. A.; Moss, R. A.; Shen, S.; Sheridan, R. S. J. Am.
Chem. Soc. 1990, 112, 1920.
(40) Raspoet, G.; Nguyen, M. T.; McGarraghy, M.; Hegarty, A. F. J.
Org. Chem. 1998, 63, 6867.
(41) Raspoet, G.; Nguyen, M. T.; McGarraghy, M.; Hegarty, A. F. J.
Org. Chem. 1998, 63, 6878.
(42) Landeck, H.; Wolff, H.; Goetz, R. J. Phys. Chem. 1977, 81, 718.
(43) Rastrelli, F.; Saielli, G.; Bagno, A.; Wakisaka, A. J. Phys. Chem.
B 2004, 108, 3479.
(38) Kirmse, W.; Guth, M.; Steenken, S. J. Am. Chem. Soc. 1996, 118,
10838.
(39) Sujdak, R. J.; Jones, R. L.; Dorfman, L. M. J. Am. Chem. Soc.
1976, 98, 4875.

Photolysis of 1,1-Diarylgermacyclobutanes
Organometallics, Vol. 27, No. 22, 2008 5957
equilibrium constant for MeOH in MeCN at ambient temper-
atures. We can only note that an observed third-order reaction
rate constant (k2MeOH) in the range of (3-5) × 107 M-2 s-1
would be consistent with a rate constant (k1 in eq 16) of a
magnitude similar to those estimated in hexane for reaction of
the three germenes with MeOH dimers, if the monomer-dimer
1 to 7.9 × 109 M-1 s-1 for 7, on the basis of the
monomer-dimer equilibrium constant reported by Fujii et al.
for acetic acid in hexane at 25 °C (K2 ) 3200 ( 500 M-1).46
As with MeOH quenching, there is again only a very modest
variation in the rate constant with aryl ring substituent. This is
also true in MeCN solution, for which the plots of kdecay
vs [AcOH] are linear and the resulting second-order rate
constants (Table 2) are ca. 15 times smaller than the corre-
sponding values in hexane. Interestingly, the rates appear to
decrease slightly with increasing electron-withdrawing power
of the aryl substituents, consistent with a Hammett F value of
F ) -0.12 ( 0.04 (in MeCN). This is in contrast with the
behavior observed for MeOH addition to the three germenes
and is also different from the behavior exhibited by the
homologous silene derivatives, whose reactions with AcOH in
MeCN proceed at rates similar to those in hexane and are
characterized by a small positiVe Hammett F value (F ) +0.17
( 0.02).32 While these results indicate that polar factors play
at best only a very small role in the reactions of both the silenes
and the germenes with AcOH, they nevertheless indicate that
proton transfer is significantly better developed in the transition
state for the rate-determining step for addition of AcOH to the
germenes than is the case with the silenes. This is in contrast
with the situation with MeOH addition, where the effects of
substituents on the rate constants are also very small but give
rise to positive Hammett F values for both the germenes and
the silenes. All in all, the results seem most compatible with a
concerted mechanism for reaction of AcOH with the germenes,
with transfer of the acidic proton to the germenic carbon better
developed in the transition state than Ge-O bond formation
(eq 18).
equilibrium constant is on the order of K2 ≈ 10-2
M .
-1 44 This
value actually seems quite reasonable and is consistent with
the infrared result of Kirmse et al.,38 given the uncertainties
associated with the detection method. The mechanism, if correct,
provides a rational explanation for the small primary kinetic
isotope effect, the very modest substituent effects on the rates
in the two solvents, and the 100-fold rate retardation in MeCN
compared to hexane solution, which is then consistent with only
a very small solvent effect on the rate constant for the process.
A change from a stepwise addition mechanism in the case of
1,1-diarylsilenes to a concerted one in the corresponding
germene derivatives seems reasonable, considering the much
lower Lewis acidity of the germenes, as evidenced by the fact
that silene 3 and the p-trifluoromethyl analogue 20 both form
detectable solvent complexes in THF solution10 while the
corresponding germene homologues (1 and 6) do not. Concerted
reaction with the hydrogen-bonded dimer of the alcohol should
always be significantly faster than with the monomeric form
for two reasons: the process involves a six- rather than a four-
center transition state, and the hydrogen-bonded dimer of the
alcohol is known to be simultaneously more nucleophilic and
more acidic than the monomer.45
(4-CF3C6H4)2SidCH2
20
Quenching of the three germenes by acetic acid proceeds at
significantly faster rates than MeOH addition in both hexane
and MeCN solution. In hexane, the quenching plots for 6 and
7 both show downward curvature over the 0.5-1 mM concen-
tration range (see Figure S10 in the Supporting Information),
which we interpret as being due to self-association of the
carboxylic acid in the hydrocarbon solvent; the cyclic hydrogen-
bonded dimer of acetic acid can be expected to be much less
reactive than the monomer, since both the nucleophilic site (the
carbonyl oxygen, presumably) and the acidic hydrogen are
engaged in hydrogen bonding to a second molecule of the acid;
we assume the reaction proceeds via formal ene addition,
involving attack of the carbonyl oxygen at germanium and
transfer of the hydroxylic proton to carbon. We failed to detect
the nonlinear dependence of kdecay on bulk AcOH concentration
in our earlier study of 1, because a much more limited set of
kinetic data were collected, over a bulk AcOH concentration
range (1-3.5 mM) higher than that used in the present study;
given a monomer-dimer equilibrium constant of K2 ) 3200
The Arrhenius parameters for reaction of AcOH with the two
sets of homologous 1,1-diarylgermenes (1 and 6) and -silenes
(3 and 20; see Table 3) provides a particularly interesting
indication of the origins of the differences in the reactivities of
the two classes of group 14 metallenes with this substrate. It
should first be noted that the activation energy increases
consistently, and the entropy of activation becomes increasingly
positive, with increasing overall reactivity throughout the series
1 < 6 < 3 < 20, as reflected in the rate constants for reaction
at 25 °C. With both sets of compounds, the activation energy
is significantly larger and the entropy of activation more positive
for the CF3-substituted derivative than for the parent compound,
by a greater margin in the silenes compared to that in the
germenes. Thus, for each pair of compounds, increased electron
demand in the MdC bond (i.e., by changing the para substit-
uents from H to CF3) has the counteracting effects of increasing
the activation energy and shifting the entropy of activation to
more positive values. With the two germenes, the activation
M ,
-1 46 the acid exists largely as the dimer, and the variation in
residual monomeric acid concentration with bulk concentration
is approximately linear, over the 1-3.5 mM bulk concentration
range. Analysis of the data for all three germenes, taking the
self-association equilibrium into account, affords linear plots
in all three cases (see Figure S10 in the Supporting Information).
The resulting rate constants for reaction with the monomeric
form of the carboxylic acid range from 9.2 × 109 M-1 s-1 for
energy difference dominates the rate constant ratio (i.e., k1+AcOH
/
k6+AcOH) slightly over the entropy difference, and as a result
the CF3-substituted derivative is less reactive than the parent
compound. The opposite is true for the silenes, with the result
that CF3 substitution results in an increase in reactivity relative
to the parent compound. It should be noted that, for the two
silene derivatives, it has been shown that solvent complexation
effects contribute to some extent to the measured activation
parameters for reaction with nucleophilic substrates in polar
(44) The estimate is derived using the approximation k2MeOH
≈
k1K2([MeOH]bulk)2, where k1 is the rate constant for the reaction of germene
with the methanol dimer.
(45) Frange, B.; Abboud, J. L. M.; Benamou, C.; Bellon, L. J. Org.
Chem. 1982, 47, 4553.
(46) Fujii, Y.; Yamada, H.; Mizuta, M. J. Phys. Chem. 1988, 92, 6768.

5958 Organometallics, Vol. 27, No. 22, 2008
Leigh et al.
solvents,10 making the differences in both Ea and ∆Sq between
the germenes and silenes, and also between 3 and 20, appear
somewhat larger than they really are. The magnitude of the
effect cannot be quantified on the basis of the data available
but is presumably not large, since there are no discernible shifts
in the UV/vis spectra of 3 and 20 in MeCN relative to hexane
solution, even at temperatures as low as 4 °C.10 It is thus clear
that the reaction of AcOH with the silenes is characterized by
an extremely loose transition state with minimal Si-O bond
formation and O-H bond rupture, with polarity similar to that
of the free reactants. In the case of the germenes the reaction
proceeds via a lower enthalpic barrier, but through a tighter,
relatively nonpolar transition state in which Ge-O bonding and
H transfer (in particular) are somewhat better developed than
in the case of the silenes. The extent of H transfer must still be
relatively small in order to be compatible with the absence of
a detectable deuterium kinetic isotope effect.
The reaction of acetone with silenes 3 and 20 also proceeds
via formal ene addition, affording the corresponding silyl enol
ethers.23 A two-step mechanism, involving rate-controlling H
transfer within a zwitterionic complex, has been proposed on
the basis of the marked sensitivity of the rate constant to aryl
substituent (F ) +1.5 in isooctane at 25 °C) and isotopic
substitution at the ketone R-carbon (kH/kD ≈ 2.1 for 3 in
isooctane at 25 °C) and the fact that the reactions of 3 and 20
proceed with negative Arrhenius activation energies.23 The
reaction of silene 20 with acetone proceeds at a rate approaching
the diffusion limit in isooctane solution over the -15 to +10
°C temperature range.23
quenching of germenes 6 and 7. The results are again consistent
with a significantly less polar transition state for H transfer in
the case of the germenes compared to the situation with the
corresponding silenes. This is compatible with the results of a
recent theoretical study by Mosey et al. of the formal [2 + 2]
cycloadditions of formaldehyde with silene and germene, in
which it was concluded that the reaction proceeds via a stepwise
pathway in both cases, but via a 1,4-zwitterion intermediate in
the case of silene and a 1,4-biradical intermediate in the case
of germene.49
The rate constants for quenching of 1, 6, 7, and 1,1-
diphenylsilene (3; see Table 1) by n-BuNH2 provide a final point
of comparison between the electrophilicities of germenes and
silenes of homologous structures. Like alcohols, primary and
secondary amines react with 3 by 1,2-addition, again via a
stepwise mechanism initiated by nucleophilic attack at silicon
followed by rate-controlling proton transfer; the reaction of 3
with n-BuNH2 is the fastest reaction that we have yet encoun-
tered with this silene.11 While the course of its reaction with 1
has not been determined, the amine is the one substrate for
which it is most reasonable to expect nucleophilic attack at
germanium to be the initial step; indeed, the increase in the
quenching rate constants throughout the series 1 < 6 < 7 is
consistent with this. The fact that the rate constants for amine
quenching are of magnitude similar to those for reaction with
MeOH, and more than 1 order of magnitude slower than those
for reaction with AcOH, is a particularly clear indication of the
much reduced electrophilicities of germenes compared to silenes
of homologous structure.
Although we have been unable to obtain evidence for the
corresponding reaction products in steady-state photolysis
experiments with 2,17 4, and 5, the kinetically stable germene
1,1-dimesityl-2-neopentylgermene (21) is known to react with
this ketone to yield germyl enol ether 22 (eq 19).47,48 From
this and the fact that lifetime quenching of 6 and 7 by acetone
is subject to a distinct primary kinetic isotope effect, we thus
conclude that the measured quenching rate constants indeed
reflect those for the ene-addition reaction; our failure to detect
any actual products of the reaction could be due either to the
products being unstable or to the reaction proceeding too slowly
to be competitive with other germene-consuming reactions under
the conditions of our steady-state photochemical experiments.
As mentioned earlier, the reaction of acetone with 1 is too slow
to be detected even by flash photolysis, and only an upper limit
of roughly k ≈ 105 M-1 s-1 could be established for the rate
constant for the putative reaction in the case of the parent
compound.17
Summary and Conclusions
In contrast to 1,1-diarylsilacyclobutanes, whose photolysis
in solution affords the corresponding 1,1-diarylsilenes and
ethylene as the exclusive products, photolysis of the homologous
1,1-diarylgermacyclobutane derivatives leads to competing
(2+2) and (1+3) cycloreversion, affording the corresponding
1,1-diarylgermenes and diarylgermylenes, respectively, along
with ethylene and cyclopropane. Germylene formation is
enhanced at the expense of germene formation by trifluoro-
methyl substitution on the aromatic rings and represents the
major excited-state decay pathway in the case of 1,1-bis[3,5-
bis(trifluoromethyl)phenyl]germacyclobutane (5).
The corresponding 1,1-diarylgermenes can be easily de-
tected and studied by laser flash photolysis methods, in spite
of minor complications from the germylene coproducts, which
are much weaker UV/vis absorbers than the corresponding
germenes and hence contribute to the observed transient
spectra in only a minor way. The germenes exhibit UV/vis
absorption spectra that are similar to one another and to those
of the homologous silene derivatives; the absorption maxima
are the same in hexane, MeCN, and THF solution and thus
do not vary as a function of solvent Lewis basicity. This is
indicative of a much weaker Lewis acidity in the cases of
the germene derivatives in comparison to that of the
homologous silenes, for which the absorption spectra in THF
are red-shifted significantly compared to those in hydrocarbon
or MeCN solution.10 The much lower Lewis acidity of
germenes compared to silenes is also reflected in the rate
constants for quenching of 1,1-diphenylgermene (1) and
-silene (3) by n-butylamine; these reactions proceed at close
to the diffusion-controlled rate in the case of the silene and
The germenes thus react with acetone at least 3 orders of
magnitude more slowly than do the homologous silenes under
similar conditions, yet they exhibit kinetic isotope effects (kH/
kD ≈ 1.6 and 2.4 for 6 and 7, respectively) that are of magnitudes
similar to that exhibited by silene 3 (kH/kD ≈ 2.1)23 under the
same conditions. In spite of this, they appear to exhibit a much
smaller substituent effect on the rate; a Hammett F value of ca.
+0.3 is estimated from the difference in rate constants for
(47) Barrau, J.; Escudie, J.; Satge, J. Chem. ReV. 1990, 90, 283.
(48) Delpon-Lacaze, G.; Couret, C.; Escudie, J.; Satge, J. Main Group
Met. Chem. 1993, 16, 419.
(49) Mosey, N. J.; Baines, K. M.; Woo, T. K. J. Am. Chem. Soc. 2002,
124, 13306.

Photolysis of 1,1-Diarylgermacyclobutanes
Organometallics, Vol. 27, No. 22, 2008 5959
flexTM 77390 peristaltic pump fitted with Teflon tubing (Cole-
Parmer Instrument Co.) to pump the solution from the reservoir
through the sample cell at more easily controlled rates. The
glassware, sample cell, and transfer lines were dried in a vacuum
oven at 65-85 °C before use. Solution temperatures were
measured with a Teflon-coated copper/constantan thermocouple
inserted into the thermostated sample compartment in close
proximity to the sample cell. Experiments at temperatures other
than 25 °C employed a gravity-flow system and a cell modified
to allow insertion of the thermocouple directly into the sample
solution. Reagents were added directly to the reservoir by
microliter syringe as aliquots of standard solutions. Transient
decay and growth rate constants were calculated by nonlinear
least-squares analysis of the absorbance-time profiles using the
Prism 5.0 software package (GraphPad Software, Inc.) and
the appropriate user-defined fitting equations, after importing
the raw data from the Luzchem mLFP software. Rate constants
were calculated by linear or nonlinear least-squares analysis of
decay rate-concentration data (at least seven points; more for
data requiring second-order polynomial fitting) that spanned as
large a range in transient decay rate as possible. Errors are quoted
as twice the standard deviation obtained from the least-squares
analyses.
ca. 100 times more slowly than this in the case of the
germene. The diarylgermenes studied in this work react much
more rapidly with acetic acid than with aliphatic alcohols,
which in turn exhibit reactivity significantly higher than that
of acetone. In reactions of germenes with substrates of this
type, the acidity of the substrate thus has a much greater
effect on the reaction kinetics than its nucleophilicity. The
opposite is true in the case of the homologous silene
derivatives, whose reactions with O- and N-containing
substrates characteristically proceed via stepwise mechanisms
involving nucleophilic attack at the SidC bond in the initial
step.1,9
Experimental Section
Hexane (BDH Omnisolv), tetrahydrofuran (Caledon Reagent),
and diethyl ether (Caledon Reagent) were dried using a Solv-
Tek, Inc. solvent drying system. Acetonitrile (HPLC grade) was
dried by passage through a column (1 × 6 in.) of neutral alumina,
preactivated by heating at 200 °C for 24 h under a stream of
dry nitrogen. Methanol, methanol-O-d, glacial acetic acid, and
acetic acid-d were used as received from Sigma-Aldrich Chemi-
cal Co. 2,3-Dimethyl-1,3-butadiene and acetone were obtained
from Sigma-Aldrich Chemical Co. and were distilled before use.
n-Butylamine (n-BuNH2; Sigma-Aldrich) was dried by distilla-
tion from solid potassium hydroxide. Steady-state photolyses
were carried out in a Rayonet photochemical reactor (Southern
New England Ultraviolet Co.) equipped with a merry-go-round
apparatus and 2-6 RPR-2537 lamps (254 nm). Details of the
synthesis and characterization of compounds and photoproducts
are described in the Supporting Information.
Nanosecond laser flash photolysis experiments employed the
pulses (248 nm; ca. 20 ns; ca. 100 mJ) from a Lambda-Physik
Compex 120 excimer laser filled with F2/Kr/Ne mixtures and a
Luzchem Research mLFP-111 laser flash photolysis system,
modified as described previously.24 Solutions were prepared at
concentrations such that the absorbance at the excitation
wavelength was between ca. 0.7 and 0.9 and were flowed
continuously through a vacuum-oven-dried, thermostated 7 × 7
mm Suprasil flow cell connected to a calibrated 100 mL
reservoir, fitted with a glass frit to allow bubbling of argon gas
through the solution. Some experiments were carried out using
a simple gravity-flow system, while others employed a Master-
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council (NSERC) of Canada for
financial support, Teck-Cominco Metals Ltd. for a gener-
ous gift of germanium tetrachloride, the McMaster
Regional Centre for Mass Spectrometry for high resolution
mass spectra, Dr. S. Kornic for elemental analyses, and
Dr. C. R. Harrington for the synthesis of 14c. L.A.H.
thanks the NSERC for a graduate scholarship.
Supporting Information Available: Text and figures giving
details of the synthesis and characterization of compounds,
representative transient decay and growth/decay profiles, time-
resolved UV/vis absorption spectra, and details of kinetic
analyses. This material is available free of charge via the Internet
at http://pubs.acs.org.
OM800574S