A series of 7-methylenedehydrobenzo[7]annulen-5-ol hexacarbonyldicobalt complexes were generated by Hosomi-Sakurai reactions of allylsilanes containing o-alkynylarylaldehyde-Co2(CO)6 complexes. One of the cyclization products was converted into its corresponding dihydrobenzo[7]annulen-7-ol hexacarbonyldicobalt complex, an immediate precursor to a benzodehydrotropylium-Co2(CO)6. The cation was generated in situ and reacted with four nucleophiles, and its aromatic stabilization was determined by computational methods.
A series of 7-methylenedehydrobenzo[7]annulen-5-ol hexacarbonyldicobalt complexes were generated by Hosomi-Sakurai reactions of allylsilanes containing o-alkynylarylaldehyde-Co2(CO)6 complexes. One of the cyclization products was converted into its corresponding dihydrobenzo[7]annulen-7-ol hexacarbonyldicobalt complex, an immediate precursor to a benzodehydrotropylium-Co2(CO)6. The cation was generated in situ and reacted with four nucleophiles, and its aromatic stabilization was determined by computational methods.
The propargylium cation–Co2(CO)6 complex
(1, Figure ) has become one of the most important species in metal-stabilized
carbocation chemistry, by virtue of its excellent stability and reactivity
that is both good and predictable; its transformations are normally
known as Nicholas reactions.[1] Among its
noteworthy features is the fact that the alkynedicobalt unit contributes
sufficiently to the overall carbocation picture such that simple substitution
at the propargylic center gives very limited change in carbocation
stability, and in many cases cations that would be difficult to generate
in the absence of the Co2(CO)6 unit are generated
readily in its presence.[2,3] Furthermore, the bending
of alkynyl carbon bond angles to ca. 140° allows the Nicholas
reaction to work well in medium-ring-size-generating reactions. As
a result, and in conjunction with reliable reductive decomplexation
protocols, the ready preparation and manipulation of cycloheptynedicobalt
complexes has been exploited extensively in total synthesis.[4,5]
Figure 1
Propargyldicobalt
cations and the target precursor.
Propargyldicobalt
cations and the target precursor.In addition to purely synthetic ends, the presence of an adjacent
alkynedicobalt unit has been shown to have significant effects on
the properties of nominally aromatic or antiaromatic carbocations.[3] The development of the aforementioned multiple
cycloheptynedicobalt preparation methods has allowed our group to
make contributions to the study of metal complexes of the putatively
aromatic 1,2-dehydrotropylium ion (henceforth referred to simply as
a dehydrotropylium ion). We have reported the generation, trapping,
and evaluation of a dehydrotropylium ion −Co2(CO)6 complex (2) and found a reduced level of aromatic
stabilization in that system, approximately 25% of that of a tropylium
ion, employing NICS(1) calculations and homodesmotic reactions as
the primary measures.[6,7] By contrast, a dehydrotropylium
ion itself only has been proposed as a mass spectral fragment of 2-acylbenzofurans
and benzothiophene,[8] and neither has it
been prepared synthetically nor has had its properties investigated.[9] The η2-platinum(0), palladium(0),
and zirconium(II) complexes of a dehydrotropylium ion or related cations
have been prepared by the Jones group and show no to modest C–C
bond alternation.[7] Dehydrotropone–Co2(CO)4 (dppm) shows considerable bond alternation.[10]In the tropylium ion system, the effect
of ring fusion to another
aromatic ring has been reported to have contradictory effects, with
benzo-fused systems less stable[11] but with
azulene and heteroaromatic systems giving enhanced stability.[12] A benzodehydrotropylium ion has been proposed
solely as a mass spectral fragment ion, without any insight on its
properties.[8] Dehydrobenzotropone–Co2(CO)4 (dppm) complexes have not been converted
to tropylium ion derivatives.[10] Given the
limitations of the above information, we were interested in developing
methods for access to the benzo-fused analogue of the dehydrotropylium
ion–Co2(CO)6 complex (3a), to assess the viability of the cation itself.
Results and Discussion
We envisioned that a benzodehydrotropylium cation–Co2(CO)6 complex (3), specifically as
methyl-substituted 3b, would be most readily obtained
through benzocycloheptynol complex 4, which in turn would
be prepared in several steps from 2-ethynylbenzaldehyde (5a) (Scheme ). 2-Ethynylbenzaldehyde
(5a), itself available by a literature Sonogashira-desilylation
protocol from 2-bromobenzaldehyde via 6a,[13,14] was subjected to a Sonogashira reaction with 2-bromoallyltrimethylsilane
to give enyne 7a (71% yield, Table ). Upon treatment with Co2(CO)8, complex 8a formed in 80% yield. Generation
of the cycloheptynol complex could then be accomplished by subjecting 8a to an equimolar amount of BF3-OEt2 at 0 °C in CH2Cl2 (0.016 M in 8a) for 20 min; the resulting Hosomi–Sakurai reaction gave exo-methylene
benzocycloheptynol–Co2(CO)6 complex 9a in 75% yield (Scheme ). The identity of 9a was evident from
the 1H resonances of the exo-methylene function (5.86 ppm
and 5.70), the benzylic carbinol methine (4.98 ppm), and the diastereotopic
methylene function (3.11 and 2.88 ppm).
Scheme 1
Preparation of Hosomi–Sakurai
Precursors
(a) K2CO3, MeOH,
or n-Bu4NF, THF; (b) H2C=CH(Br)–CH2SiMe3, Pd(PPh3)4, CuI, i-Pr2NH–THF;
(c) Co2(CO)8, 0 °C, Et2O.
Table 1
Hosomi–Sakurai Precursor Preparation
7 (%)
8 (%)
a
71
80
b
73
75
c
66
85
d
63
77
e
72
a
Yield determined after a two-step
complexation/cyclization process.
Scheme 2
Hosomi–Sakurai Reactions
Preparation of Hosomi–Sakurai
Precursors
(a) K2CO3, MeOH,
or n-Bu4NF, THF; (b) H2C=CH(Br)–CH2SiMe3, Pd(PPh3)4, CuI, i-Pr2NH–THF;
(c) Co2(CO)8, 0 °C, Et2O.Yield determined after a two-step
complexation/cyclization process.While other reaction fates and other ring size preparations
are
known,[15,16] this is the first example of a cycloheptynoldicobalt
complex preparation by a Hosomi–Sakurai reaction. As a result,
we wished to explore the generality of this method for benzocycloheptynedicobalt
complex preparation and investigated this protocol on a series of
related compounds, including those derived from 2-ethynyl-5-methoxybenzaldehyde
(5b), 3-ethynyl-2-furancarboxaldehyde (5c), 3-ethynyl-2-thiophenecarboxaldehyde (5d), and 3-ethynyl-1-methyl-1H-indole-2-carboxaldehyde (5e). The Sonogashira
reaction of 5b–e with 2-bromoallyltrimethylsilane
afforded 7b–e in a straightforward
fashion (7b, 73%; 7c, 66%; 7d, 63%; 7e, 72%) (Table ). The reaction of 7b–7d with Co2(CO)8 formed the hexacarbonyldicobalt
complexes 8b (75% yield), 8c (85%), and 8d (77%) in good yield. Conversely, alkyne 7e gave an alkyne complex (8e), which had limited stability
to chromatography; consequently, 8e was carried forward
to the subsequent cyclization reaction without rigorous purification.The enyne complexes 8b–e were
subjected to reaction with BF3-OEt2 under conditions
analogous to the 8a–9a cyclization
reaction. In each of the cases, the fused bicyclic exo-methylene cycloheptynol
complex could be formed in reasonable to excellent yield, with the
methoxy-substituted 9b (5 min, 72% yield) being formed
somewhat more rapidly than 9a and the 5,7-systems 9c (59% yield) and 9d (92% yield) forming somewhat
more slowly (ca. 1 h) than 9a. Compound 9e was formed, also over 1 h, in 44% yield, for the two-step process
and based on the amount of metal-free alkyne 7e. Each
of 9b–9e gave characteristic 1H NMR spectral resonances closely analogous to those for 9a; for instance, in the case of 9d, the resonances
for the exo-methylene appeared at 5.85 and 5.70 ppm, the carbinol
methine at 5.05 ppm, and the diastereotopic methylene at 3.00 and
2.87 ppm. Despite the more modest yield for 9e, these
results demonstrate that this Hosomi–Sakurai-reaction-based
protocol is an effective general method for the preparation of exo-methylene-substituted
cycloheptynol–Co2(CO)6 complexes.With the goal of ultimate generation of the dehydrotropylium ion–Co2(CO)6 complex, compound 9a was selected
for further manipulation. A Swern oxidation[17] afforded ketone 10 with minimal decomposition (89%
yield) (Scheme ),
as evidenced by the appearance of a shifted CH2 singlet
at 3.76 ppm in the 1H NMR spectrum and ketone carbonyl 13C NMR spectral resonance at 198.3 ppm. While this material
could also be obtained with MnO2 oxidation, the yields
were inferior (28%, 43% by recovered starting material). With the
ketone function in place, the exo-methylene group of 10 could be induced to isomerize into conjugation with the carbonyl
by a catalytic amount of H2SO4, giving enone 11 in 88% yield. The formation of this conjugated ketone was
evidenced by the disappearance of the exo-methylene function of 10 in the 1H NMR spectrum and the appearance of
a single vinyl proton resonance at 6.68 ppm and a methyl group at
2.43 ppm. It is noteworthy that the analogous Co2(CO)4–dppm complex, 12, has been prepared previously
by a carbonylative Heck reaction,[10] but
this process is not tolerant for the Co2(CO)6 complex. Finally, reduction of 11 with DIBAL-H gave
a crude reaction product whose 1H NMR spectral resonances
were consistent with 4, but this material underwent prompt
isomerization on silica gel; following chromatographic purification, 13 was isolated in 68% yield, along with a small amount of
the elimination product 14 (29%). The appearance of two
vinyl protons at 6.39 and 6.15 ppm (each J = 12.5
Hz) in the 1H NMR spectrum was the most distinctive evidence
for this isomerization product. Since 13 was equally
serviceable relative to 4 as a precursor to the dehydrotropylium
ion–Co2(CO)6 complex, compound 13 was considered appropriate for further study (Scheme ).
Scheme 3
Preparation of a Dehydrotropylium Ion Complex Precursor
(a) CH3S(O)CH3, ClC(O)C(O)Cl, CH2Cl2, −78 °C;
then 9a, Et3N, −78 to −20 °C,
89% (b) H2SO4, CH2Cl2,
0 °C, 88%; (c) DIBAL-H, THF, −78 °C; (d) silica gel,
68% (13) + 29% (14).
Scheme 4
Nicholas Reactions of Cation Precursor 13
Preparation of a Dehydrotropylium Ion Complex Precursor
(a) CH3S(O)CH3, ClC(O)C(O)Cl, CH2Cl2, −78 °C;
then 9a, Et3N, −78 to −20 °C,
89% (b) H2SO4, CH2Cl2,
0 °C, 88%; (c) DIBAL-H, THF, −78 °C; (d) silica gel,
68% (13) + 29% (14).The reactivity of 13 was investigated
with a series
of nucleophiles, chosen as being representatives of the groups of
nucleophiles commonly incorporated in Lewis-acid-mediated reactions
of metal-stabilized carbocations and in view of their relative nucleophilicities
as measured on the Mayr N scale.[18] The reaction of 13, BF3-OEt2,
and acetophenone trimethylsilyl enol ether (N = 6.22)
gave 15 (94% yield) as a 9:1 regioisomeric mixture of 15γ and 15α (Table ). N-Methylpyrrole (N = 5.85) gave condensation product 16 (81%
yield) with only the γ-regioisomer apparent (16γ:16α > 30:1). Methallyltrimethylsilane (N = 4.41)
afforded 17 (89% yield) as a 13:1 mixture of 17γ and 17α, whereas allyltrimethylsilane itself
(N = 1.68) afforded 18 (82% yield) as
a similar regioisomeric mixture (18γ:18α =
7.5:1). The major isomers (15–18γ) were
distinguished in the 1H NMR spectra by the single endocyclic
vinyl proton (5.85 ppm in 15γ), whereas the minor
isomer (15α, 16–18α), when present,
was distinguished by the two doublets or AB pattern for the two endocyclic
vinyl protons (6.39 ppm, and 6.30, AB, J = 12.3 Hz
in 15α). In addition, resonances for the incorporated
nucleophile [i.e., 5.79 (m), 5.13 ppm (d, J = 17.1
Hz,), 5.09 (d, J = 10.2 Hz) for the vinyl protons
of the allyl unit of 18γ] were diagnostic. Attempts
to incorporate less reactive nucleophiles, with N < 1 (i.e., thiophene, N = −1.01), resulted
in the formation of elimination product 14, whereas radical
homocoupling products were not seen in any of these reactions, in
contrast to analogous reactions on dehydrotropylium–Co2(CO)6 ion (2) precursors.[6]
Table 2
Nicholas Reactions of 13
Nu (Mayr N value)
product,
yield (%)
γ:α
acetophenone
TMS enol ether
(6.22)
15, 94
9.0:1
N-methylpyrrole (5.82)
16, 81
>30:1
methallyltrimethylsilane
(4.41)
17, 89
13:1
allyltrimethylsilane (1.68)
18, 82
7.5:1
From these results, it is
clear that the benzodehydrotropylium
ion–Co2(CO)63b is generated
readily. In previous work, Nicholas reactions of cations that are
allylic in addition to being propargylic to an alkyne–Co2(CO)6 group have been shown to favor attacks at
the γ-site [i.e., the terminus remote to the alkyne–Co2(CO)6 unit] over the α-site for the majority
of nucleophiles (with a greater selectivity for lower reactivity nucleophiles),
for reasons that are not entirely well-understood. This observation
has been found to apply to both acyclic[19] and cyclic systems.[20] Here, this situation
is much less straightforward, as the γ-site of 3b is benzylic, whereas the α-site is substantially more hindered
than in the literature cases.With the isolation of the benzodehydrotropylium
ion 3b complex not readily available, due to the elimination
pathway, the
structural aspects of 3b were addressed computationally.
The putative aromaticity of 3b was also addressed in
terms of geometric, ring current, and energetic factors, namely, HOMA
values,[21] NICS(1) values,[22] and appropriate homodesmotic reactions,[23] for comparison to 1 and to tropylium ion.
At the B3LYP/6-311+G(d,p) + ZPVE level, the minimized structure of 3b reveals a 7-membered ring that is nearly planar (Figure ), with dihedral
angles in the 7-membered ring averaging 8.4° and no individual
dihedral angle in the ring being >18.5° (see the Supporting Information). Very modest bond alternation
is apparent, with bond lengths ranging from 1.385 to 1.437 Ǻ,
excluding the formal triple bond. Within the harmonic oscillator model
of aromaticity, these bond lengths correspond to a HOMA of 0.636 (EN
= 0.256, GEO = 0.108). While this figure is coincident with molecules
of substantial aromaticity, it is lower than the dehydrotropylium
ion–Co2(CO)6 complex 0.950 (EN = 0.044,
GEO = 0.007). However, this earlier work[6a] also demonstrated that HOMA was a poor measure of aromaticity in
these charged systems in comparison with other methods.
Figure 2
Optimized structure
of 3b. Selected by lengths [Å]:
C5–C6 1.364, C6–C7 1.403, C7–C8 1.415, C8–C9
1.385, C9–C4a 1.436, C4a–C9a 1.437.
Optimized structure
of 3b. Selected by lengths [Å]:
C5–C6 1.364, C6–C7 1.403, C7–C8 1.415, C8–C9
1.385, C9–C4a 1.436, C4a–C9a 1.437.The NICS(1) values of 3b were calculated to evaluate
aromaticity by ring current criteria (see the Supporting Information). These calculations resulted in two
different values for the 7-membered ring, with (arbitrarily) above
the ring top face giving a value of −4.88 and the opposite
face yielding a value of −2.03. A strict averaging of these
numbers gives a value of −3.46, somewhat higher than the dehydrotropylium–Co2(CO)6 cation (1) (−2.92) and
approximately one-third of that of the tropylium ion itself (−10.5
with the identical functional and basis set). Nevertheless, it is
apparent that the proximity of each Co(CO)3 unit is likely
affecting these numbers, and they should be taken with caution.After careful consideration of potential homodesmotic measures
of aromaticity, the equation in Scheme was chosen as the most appropriate reaction, given
the nature of 3b as both a propargyldicobalt cation and
as a benzylic cation. Employing DFT calculations (B3LYP/6-311+G(d,p)
+ ZPVE), cation 3b was evaluated as being aromatic by
2.7 kcal mol–1 (see the Supporting Information). Given the apparent issues with the HOMA and NICS(1)
calculations, we consider this the most reliable of the measures of
the aromaticity of 3b. This value is nearly identical
(≤0.1 kcal mol–1 less) to that resulting
from the dehydrotropylium ion complex 2 (2.8 kcal mol–1) and ca. 23% of the value arrived at for the tropylium
ion itself (11.6 kcal mol–1).[6a]
Scheme 5
Homodesmotic-Reaction-Based Evaluation of 3b
In summary, we have prepared
the alcohol precursor to a dehydrobenzotropylium
ion–Co2(CO)6 by a Hosomi–Sakurai
reaction and carried out subsequent manipulation of the exo-methylene
benzocycloheptynol complex into a benzocycloheptenynol complex. The
Hosomi–Sakurai reaction to generate the benzocycloheptyne–Co2(CO)6 ring system has been demonstrated to have
some generality. The dehydrobenzotropylium ion–Co2(CO)6 complex 3b may be generated in solution
and reacts with nucleophiles of Mayr nucleophilicities of N > 1. The cation itself is modestly aromatic, having
approximately
the same aromatic stabilization as dehydrotropylium ion−Co2(CO)6 and roughly one quarter of that of tropylium
ion.
Experimental Section
General Considerations[25]
Reagents were obtained from commercial sources
unless otherwise stated.
Reactions were conducted under an inert atmosphere (N2)
using glassware dried in an oven (110 °C, >1 h). The solvent
for each reaction was acquired from a solvent purification system
(Innovative Technologies). BF3-OEt2 was distilled
prior to use and stored under an inert atmosphere (N2).
Reactions were subject to a “conventional workup” by
partitioning the reaction mixture between an aqueous phase and a diethyl
ether or dichloromethane phase, combining the organic phases, followed
by drying (MgSO4), filtration, and concentration of the
organic phase. Flash chromatography was performed according to the
method of Still.[26] High-resolution mass
spectrometry (HRMS) results were obtained via a direct insertion probe-electron
ionization method (70 eV) on a GCT time-of-flight (ToF) mass spectrometer
at the McMaster Regional Centre for Mass Spectrometry and in the University
of Windsor Mass Spectrometry lab with a ToF mass spectrometer using
the atmospheric solids analysis probe (ASAP) and a corona discharge
to facilitate ionization. Elemental analysis was performed by Guelph
Chemical Laboratories, Guelph, ON (Canada). 1H NMR spectra
were recorded on 300 or 500 MHz spectrometers. Chemical shifts (δ)
are reported in parts per million (ppm), relative to the 7.26 ppm
resonance for the residual CHCl3 in CDCl3, unless
otherwise indicated. Coupling constants are reported in Hertz (Hz). 13C NMR data were obtained at either 75 or 125 MHz. Infrared
spectra (IR) were recorded on a FT-IR spectrophotometer using KBr
plates. 2-Bromo-3-(trimethylsilyl)-1-propene and[24] alkynes 6a,[14]5a,[14]6b,[14]5b,[14]6c,[27]5c,[13]6d,[27]5d,[27] and 6e(28) were prepared by literature methods.
3-Ethynyl-1-methyl-1H-indole-2-carbaldehyde
(5e)
To a solution of 6e (0. 741
g, 2.90 mmol) in anhydrous THF (25 mL) was added a solution of TBAF
(1.0 M in THF, 2.9 mL, 1 equiv) at room temperature. The solution
was stirred for 5 min, and a solution of aqueous NH4Cl
(sat) was added. The mixture was subjected to conventional extractive
workup using diethyl ether to give the material satisfactory for further
use. Compound 5e was isolated as a yellow solid (0.5122
g, 95% yield), mp 122–123 °C. IR (KBr) νmax 3284, 2102, 1669, 1476 cm–1; 1H NMR
(500 MHz, CDCl3) δ 10.21 (s, 1H), 7.84 (dt, J = 1.0, 8.0 Hz, 1H), 7.47 (apparent td, J = 1.0, 8.0 Hz, 1H), 7.40 (apparent dt, J = 1.0,
8.5 Hz, 1H), 7.26 (td, J = 1.0, 7.5 Hz, 1H), 4.09
(s, 3H), 3.53 (s, 1H); 1H NMR (500 MHz, CD2Cl2) δ 10.20 (s, 1H), 7.82 (d, J = 8.1
Hz, 1H), 7.49–7.43 (m, 2H), 7.26 (td, J =
7.0, 1.0 Hz, 1H), 4.07 (s, 3H), 3.61 (s, 1H); 13C NMR (75
MHz, CDCl3) δ 182.3, 139.3, 136.5, 128.0, 127.8,
122.1, 121.9, 110.6, 110.3, 84.7, 74.8, 31.8; HRMS m/e for C12H9NO [M+] calcd 183.0684, found 183.0684.
General Method I: To a solution of 2-ethynylbenzaldehyde
(6a, 0.543 g, 4.18 mmol) and 2-bromo-3-(trimethylsilyl)-1-propene
(0.807 g, 4.18 mmol) in a degassed (5:1) mixture of diisopropylamine/THF
(40 mL) were added Pd(PPh3)4 (0.241 g, 5 mol
%) and CuI (0.080 g, 10 mol %). The solution was stirred for 2 h at
room temperature. An aqueous solution of NH4Cl (sat) was
added, and the mixture was subjected to a conventional extractive
workup using diethyl ether. The crude material was subjected to flash
chromatography (25:1 petroleum ether/Et2O) to give product 7a (0.719 g, 71%) as a yellow oil; IR (KBr) νmax 2927, 1700, 1599 cm–1; 1H NMR (500
MHz, CDCl3) δ 10.53 (d, J = 0.5
Hz, 1H), 7.91 (m, 1H), 7.53 (m, 2H), 7.41 (m, 1H), 5.37 (d, J = 2.0 Hz, 1H), 5.18 (d, J = 1.5 Hz, 1H),
1.79 (d, J = 0.5 Hz, 2H), 0.11 (s, 9H); 13C NMR (75 MHz, CDCl3) δ 192.0, 135.9, 133.9, 133.2,
128.6, 128.4, 127.3, 127.2, 120.8, 98.7, 83.8, 28.1, −1.5;
HRMS m/e for C15H18OSi [M+] calcd 242.1127, found 242.1130.
General Method II: To a solution of 7a (0.132 g, 0.545 mmol) in anhydrous Et2O at 0
°C was added an unweighed amount of dicobalt octacarbonyl (excess).
The solution was stirred for 1.5 h at 0 °C and was subsequently
concentrated under reduced pressure at 0 °C. The crude brown-colored
material was added to the top of a plug of silica and washed with
hexanes, followed by diethyl ether. The concentration of the Et2O washings afforded compound (8a) as a dark-brown
oil (0.230 g, 80% yield); IR (KBr) νmax 2925, 2852,
2092, 2055, 2022, 1693, 1593 cm–1; 1H
NMR (500 MHz, CDCl3) δ 10.37 (s, 1H), 7.94 (d, J = 7.8 Hz, 1H), 7.68 (d, J = 7.5 Hz, 1H),
7.61 (td, J = 1.25, 7.5 Hz, 1H), 7.46 (t, J = 7.7 Hz, 1H), 5.48 (d, J = 1.0 Hz, 1H),
5.35 (d, J = 1.0 Hz, 1H), 1.72 (d, J = 1.0 Hz, 2H), 0.10 (s, 9H); 13C NMR (75 MHz, CDCl3) δ 199.0, 191.1, 143.7, 141.5, 134.7, 133.5, 132.6,
128.2, 117.3, 101.9, 88.6, 26.5, −0.8; HRMS m/e for C21H18Co2O7Si [M+ – 2CO] calcd 471.9587, found
471.9604.
To a solution of oxalyl chloride (60 μL, 0.69 mmol) in CH2Cl2 (15 mL) at −78 °C was added DMSO
(60 μL, 0.84 mmol). After 30 min, a solution of 9a (0.1130 g. 0.248 mmol) in CH2Cl2 (4 mL) was
added. After stirring at −78°C for 45 min, Et3N (0.30 mL, 2.2 mmol) was added, and the solution was gradually allowed
to warm to −20 °C over 45 min. Addition of NH4Cl(aq) and a conventional extractive workup (CH2Cl2), followed by flash chromatography (10:1 petroleum ether:Et2O), gave 10 (0.1005 g, 89%) as a red-brown oil;
IR (KBr) νmax 2926, 2095, 2057, 2026, 1682, 1596
cm–1; 1H NMR (500 MHz, CDCl3) δ 7.75 (d, J = 7.5 Hz, 1H), 7.68 (dd, J = 1.5, 7.7 Hz, 1H), 7.56 (td, J = 1.5,
7.7 Hz, 1H), 7.41 (td, J = 1.0, 7.7 Hz, 1H), 5.77
(s, 1H), 5.75 (s, 1H), 3.76 (s, 2H); 13C NMR (125 MHz,
CD2Cl2) δ 198.7, 198.3, 139.1, 138.6,
136.8, 133.4, 132.6, 128.9, 128.4, 122.3, 87.8, 87.4, 51.9; HRMS m/e for C18H8Co2O7 [M+] calcd 453.8934, found 453.8914.
To a
solution of 10 (80.9 mg, 0.20 mmol) in anhydrous CH2Cl2 (10 mL) at 0 °C was added H2SO4 (3 drops in 2 mL of anhydrous CH2Cl2) in a dropwise fashion over a period of 20 min. The solution
was stirred for 2 h at 0 °C, after which water was added, and
the mixture was subjected to a conventional extractive workup using
dichloromethane. The crude material was subjected to flash chromatography
(5:1 petroleum ether/Et2O) to afford 11 (70.8
mg, 88% yield) as a brown solid, mp 77–79 °C; IR (KBr)
νmax 2918, 2095, 2057, 2031, 1730, 1605, 1582 cm–1; 1H NMR (300 MHz, CDCl3) δ
8.25 (dd, J = 1.2, 8.1 Hz, 1H), 7.87 (dd, J = 1.2, 7.6 Hz, 1H), 7.66 (td, J = 1.2,
7.5 Hz, 1H), 7.53 (td, J = 1.2, 7.5 Hz, 1H), 6.67
(d, J = 1.2 Hz, 1H), 2.43 (d, J =
1.2 Hz, 3H); 13C NMR (75 MHz, CD2Cl2) δ 198.5, 189.3, 149.5, 137.8, 136.7, 133.6, 133.5, 132.4,
131.4, 129.1, 85.4, 83.7, 23.6; HRMS m/e for C18H8Co2O7 [M+] calcd 453.8934, found 453.8941.
To a solution
of compound 11 (52 mg, 0.11 mmol) in anhydrous CH2Cl2 (7 mL) at −78 °C was added DIBAL-H
(0.46 mL of a 1.0 M solution in THF, 0.46 mmol, 4 equiv) in a dropwise
manner. The solution was stirred for 1 h at −78 °C. An
aqueous solution of NH4Cl (sat) was added, and the reaction
mixture was subjected to a conventional extractive workup using dichloromethane.
Following removal of the volatiles under reduced pressure, flash chromatography
(10:1 petroleum ether/Et2O) afforded elimination product 14 (14.0 mg, 29% yield) followed by alcohols 13 and 4 (inseparable by chromatography). The isolated
mixture containing (13) and (4) was dissolved
in hexanes, and silica gel was added. After stirring for 1 h, the
silica was removed by filtration and the collected filtrate was concentrated
under reduced pressure to afford alcohol (13) (35.5 mg,
68% yield from ketone 11). Compound 13;
a red-brown solid; mp 80–81 °C (dec); IR (KBr) νmax 3448, 2924, 2853, 2092, 2055, 2024, 1723 cm–1; 1H NMR (300 MHz, CDCl3) δ 7.67 (m,
1H), 7.30–7.40 (m, 2H), 7.20 (m, 1H), 6.39 (d, J = 12.5 Hz, 1H), 6.15 (d, J = 12.5 Hz, 1H), 2.31
(s, 1H), 1.55 (s, 3H); 13C NMR (75 MHz, CDCl3) δ 199.2, 139.8, 135.5, 132.0, 133.0, 131.81, 131.79, 128.9,
128.7, 128.2, 107.1, 86.5, 73.7, 32.0; HRMS m/e for C18H10Co2O7 [M+ – 2CO] calcd 399.9192, found 399.9196. Compound 14; a red-brown viscous oil; IR (KBr) νmax 3016, 2086, 2033, 2017, 2000, 1624 cm–1; 1H NMR (500 MHz, CDCl3) δ 7.70 (m, 1H), 7.27–7.31
(m, 2H), 7.15 (m, 1H), 6.37 (1/2 AB, J = 12.6 Hz,
1H), 6.34 (1/2 AB, J = 12.6 Hz, 1H), 5.71 (s, 1H),
5.63 (s, 1H); 13C NMR (125 MHz) 198.7, 143.8, 137.0, 133.13,
133.10, 133.05, 130.9, 130.1, 128.9, 128.6, 121.6, 88.2, 87.3; HRMS
(ASAP) m/e for C18H8Co2O6 [M+ + H] calcd 438.9063,
found 438.9059. Resonances for 4 could be observed in
the 1H NMR spectra of the crude reaction product at δ
7.74 (dd, J = 1.0, 7.4, Hz, 1H), 7.54 (d, J = 7.6, 1H), 7.44 (apparent dt, J = 1.2,
7.5 Hz, 1H), 7.40 (apparent dt, J = 1.1, 7.4, 1H),
6.05 (m, 1H), 5.12 (br s, 1H), 2.21 (s, 3H), 1.97 (d, J = 4.3, 1H).General Method IV: To a solution of (13) in anhydrous dichloromethane at 0 °C was added an excess amount
of the nucleophile (5–8 equiv). BF3-OEt2 (3 equiv) subsequently was added dropwise while maintaining the
temperature at 0 °C. The solution was stirred for 30 min (with
monitoring by TLC). Aqueous NaHCO3 (sat) was added, followed
by a conventional extractive workup using dichloromethane. Flash chromatography
(100:1 petroleum ether/Et2O) afforded the final product.
Hexacarbonyl[μ-(8,9-dehydro-2-(7-methyl-5H-benzo[7]annulen-5-yl)-1 phenylethanone)]dicobalt (15γ and 15α)
A solution of alcohol compound 13 (68.0 mg, 0.149 mmol) in CH2Cl2 (10
mL) was subjected to General Method IV with trimethyl((1-phenylvinyl)oxy)silane
(143 mg, 0.745 mmol, 5 equiv) and BF3-OEt2 (57
μL, 0.45 mmol, 3 equiv), followed by flash chromatography (20:1
petroleum ether: Et2O), to afford a 9.0:1 mixture of 15γ and 15α (0.0780 g, 94% yield)
as a red-brown viscous oil; IR (KBr) νmax 2923, 2853,
2088, 2050, 2017, 1688, 1449 cm–1; 1H
NMR (300 MHz, CDCl3) (major product, 15γ) δ 7.96 (d, J = 7.4 Hz, 2H), 7.73 (m, 1H),
7.58 (m, 1H), 7.42–7.49 (m, 2H), 7.28–7.36 (m, 2H),
5.85 (dd, J = 5.9, 1.4 Hz, 1H), 4.05 (m, 1H), 3.5d
(dd, J = 16.8, 8.6 Hz, 1H), 3.46 (d, J = 16.8, 4.5
Hz, 1H), 2.14 (s, 3H); resonances from the minor product (15α) could be observed at 7.88 (d, J = 7.5 Hz, 1H),
7.67 (m, 1H), 6.36 (1/2 AB, J = 12.3 Hz, 1H), 6.30
(1/2 AB, J = 12.2 Hz, 1H), 3.40 (d, J = 16.3 Hz, 1H), 3.26 (d, J = 16.3 Hz, 1H), 1.66
(s, 3H); 13C NMR (75 MHz, CDCl3) (major product, 15γ) δ 199.6, 197.6, 138.5, 137.1, 136.7, 135.1,
133.4, 133.2, 130.5, 128.9, 128.6, 128.0, 127.4, 126.8, 91.8, 88.6,
44.1, 40.2, 23.3; resonances from the minor isomer could be observed
at 197.1, 140.4, 135.5, 131.4, 131.3, 129.4, 128.4, 51.8, 42.5, 28.8;
HRMS m/e for C26H16Co2O7 [M+ – 6CO]
calcd 389.9865, found 389.9844.