.
Angewandte
Communications
of 1b–5b and 3c were confirmed by X-ray crystallography
(Figure 1 and see the Supporting Information). Electronic
substituent effects have a strong impact on the reaction rate,
with electron-donating groups slowing the reaction down and
electron-accepting groups enhancing it. As an extreme
example of the latter case, tetrakis(4-cyanophenyl)-DCF 5a
even undergoes rearrangement at room temperature (299–
301 K), and leads to quantitative conversion into 5b/c after
4 days in DMF. To our knowledge, this is the only example of
a thermal pentafulvene-to-benzene rearrangement that is
possible under ambient conditions. Interestingly, the Xb/Xc
selectivity is independent of temperature. Heating 5a in DMF
to 1608C for 10 min or letting it sit for 4 days at room
temperature leads to the same ratio of 5b/5c, thus suggesting
that the isomer selection step is not rate-limiting. There
appears to be little correlation between 1,3-/1,4-dicyanoben-
zene selectivity and the electronic character of the pendant
group, as both electron-rich 4a and electron-poor 5a led to
the formation of significant quantities of the 1,4-dicyanoben-
zene isomer.
galvinoxyl was recovered. While this does not fully preclude
a radical-based mechanism, we conclude from this and other
experiments that such a pathway is unlikely.
Kinetic data for the rearrangement of 1a to 1b was
collected by monitoring the reaction in [D7]DMF by 1H NMR
spectroscopy (500 MHz) at 1008C over the course of 48 h,
with 1,3,5-trimethoxybenzene used as an internal standard
(see Section SI4 in the Supporting Information). Upon
heating the mixture to 1008C, the signals corresponding to
the aromatic protons of 1a immediately broadened and
subsequently sharpened slightly over the course of the first
5.5 h. During this time, the rates of consumption of 1a and
formation of 1b are nonlinear and could not be fitted to
simple zero-, first-, or second-order kinetics. After this time,
however, the rates of consumption of 1a and formation of 1b
follow zero-order kinetics with a rate constant k = 4.13 Æ
0.013 ꢀ 10À9 m sÀ1. The broadening of the signals of 1a during
the “induction period” indicates the formation of a transient
intermediate species. The broad signals were retained when
1a was heated to 1608C in [D7]DMF for 10 min, then cooled
and analyzed by NMR spectroscopy at room temperature.
However, heating 1a in DMF to 1008C for 10 min followed by
immediate removal of the solvent and 1H NMR spectroscopy
in CDCl3 gave sharp signals identical to pristine 1a, thus
indicating that formation of the intermediate is reversible and
possibly involves solvent as a reactant.
To test the importance of solvent polarity, we heated DCF
1a at 1608C in a number of solvents (Table 2). The high-
yielding transformation to the benzene derivatives only
occurred in dipolar aprotic solvents such as DMF, N,N-
Table 2: Rearrangement of 1a in various solvents.
The strong dependence of the rearrangement rate on the
electronic properties of the substituents on the tetraarylated
DCFs, along with the necessity for dipolar aprotic solvents to
promote the reaction, suggest a polar mechanism for the
rearrangement rather than a radical-based one. We propose
a “ring-walk” mechanism (Scheme 2). Attack of a nucleophile
(in this case, solvent) leads to 1,2-cyano migration and
formation of intermediate 2e[8] by either cyanide release
and reattack (pathway a) or formation of a spirocyclic
iminocyclopropyl anion (pathway b). The anion attacks at
the 2-position of the fulvene ring, thereby leading to bicyclo-
[3.1.0]hex-1-ene 2 f, with C6 capable of quickly migrating
around the five-membered ring to form intermediate 2g. This
intermediate can either release the nucleophile to form 2b, or
undergo another “ring-walk” to form intermediate 2h, which
forms 2c upon release of the nucleophile. Based on the kinetic
data, the rate-determining step is probably the formation of
2 f, with electron-withdrawing groups on the phenyl ring
promoting the nucleophilic attack. We confirmed by 13C
labeling studies that C6 of 2a retains one cyano group in 2b
(see Section SI7 in the Supporting Information).
In this polar mechanism, it is likely that, if solvent attack is
reversible as we propose, reversible release of cyanide will
occur. To test this, we prepared [15N2]-2a and subjected it to
rearrangement conditions in [D7]DMF in the presence of one
equivalent of unlabeled 3a, chosen for its similar rate of
rearrangement (see Section SI6 in the Supporting Informa-
tion). The rearrangement of [15N2]-2a alone gives one signal in
the 15N NMR spectrum (50.7 MHz, 298 K) at d = 269.7 ppm.
However, when the rearrangement of [15N2]-2a takes place in
the presence of one equivalent of 3a, a second signal appears
at d = 268.3 ppm (see Figure SI14 in the Supporting Informa-
tion). This clearly indicates intermolecular migration of the
Entry
Solvent
Conversion of 1a
1
2
3
4
5
6
7
8
9
DMF
DMA
NMP
Me2SO
MeCN
1,2-dichloroethane
EtOH
AcOH
quant.
quant.
quant.
78%
<1%
0%
0%[a]
<1%
0%
pyridine
[a] Low solubility.
dimethylacetamide (DMA), Me2SO, and N-methylpyrroli-
done (NMP). Reactions in MeCN and AcOH gave only trace
amounts of product, while 1,2-dichloroethane, EtOH, and
pyridine were completely ineffective. In all cases, the only
analytes observed were either DCF starting material or the
rearrangement products.
At this point, we began to consider the mechanism of the
reaction. In the case of the thermal rearrangement of
pentafulvene, the proposed biradical “prefulvene” intermedi-
ate is formed by a concerted 1,2-H shift from C6 to C1.[3e] To
check for the possibility of radical intermediates, the rear-
rangement of 2a was performed in the presence of one
equivalent of the radical trap galvinoxyl.[7] The formation of
2b/c proceeded as usual with no change in rate, and the
2
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!
Finke, Aaron D.
