6,6-Dicyanopentafulvenes: Electronic structure and regioselectivity in [2 + 2] cycloaddition-retroelectrocyclization reactions

Article abstract of DOI:10.1021/ja309141r

We present an investigation of the electronic properties and reactivity behavior of electron-accepting 6,6-dicyanopentafulvenes (DCFs). The electron paramagnetic resonance (EPR) spectra of the radical anion of a tetrakis(silylalkynyl) DCF, generated by Na

Full text of DOI:10.1021/ja309141r

Article
pubs.acs.org/JACS
6,6-Dicyanopentafulvenes: Electronic Structure and Regioselectivity
in [2 + 2] Cycloaddition−Retroelectrocyclization Reactions
Aaron D. Finke,† Oliver Dumele,† Michal Zalibera,‡ Daria Confortin,‡ Pawel Cias,‡
Govindasamy Jayamurugan,† Jean-Paul Gisselbrecht,§ Corinne Boudon,§ W. Bernd Schweizer,†
Georg Gescheidt,‡ and Francois Diederich†,*
̧
†Laboratory of Organic Chemistry, ETH Zurich, Honggerberg, HCI, CH-8093 Zurich, Switzerland
̈
‡Institute of Physical and Theoretical Chemistry, Graz University of Technology, Stremayrgasse 9, A-8010 Graz, Austria
§Laboratoire d’Electrochimie et de Chimie Physique du Corps Solide, Institut de Chimie−UMR 7177, C.N.R.S.,
́
Universite de Strasbourg, 4 rue Blaise Pascal, 67081 Strasbourg Cedex, France
S
* Supporting Information
ABSTRACT: We present an investigation of the electronic
properties and reactivity behavior of electron-accepting 6,6-
dicyanopentafulvenes (DCFs). The electron paramagnetic
resonance (EPR) spectra of the radical anion of a tetrakis-
(silylalkynyl) DCF, generated by Na metal reduction, show
delocalization of both the charge and unpaired electron to the
nitrogens of the cyano moieties and also, notably, to the silicon
atoms of the four alkynyl moieties. By contrast, in the radical
anion of the previously reported tetraphenyl DCF, coupling to
the four phenyl rings is strongly attenuated. The data provide
physical evidence for the different conjugation between the DCF core and the substituents in both systems. We also report the
preparation of new fulvene-based push−pull chromophores via formal [2 + 2] cycloaddition−retroelectrocyclization reaction of
DCFs with electron-rich alkynes. Alkynylated and phenylated DCFs show opposite regioselectivity of the cycloaddition, which
can be explained by the differences in electronic communication between substituents and the DCF core as revealed in the EPR
spectra of the radical anions.
arose that any sufficiently electron-poor olefin can be potentially
reactive in a formal CA−RE reaction with electron-rich alkynes to
generate a push−pull system. The next logical step was to identify
new acceptors that contain electron-poor olefins and test them for
CA−RE reactivity.
INTRODUCTION
■
The chemical landscape for organic molecules with desirable
properties suitable for application in electronic devices is
continuously expanding.1 While much of the current advance-
ments in organic-based devices use components made of well-
established building blocks, breakthroughs in the field of
functional organic materials can still result from the discovery
of new scaffolds and synthetic methodologies.2
To these ends, our research group has developed a new class
of nonplanar push−pull chromophores resulting from the
formal [2 + 2] cycloaddition−retroelectrocyclization (CA−RE)
reaction between electron-rich alkynes and electron-poor olefins.3
These chromophores have tunable absorption expanding into the
near-IR region, display amphoteric redox behavior, and possess
reduction potentials which rival those of the linchpin organic
acceptors tetracyanoethylene (TCNE) and 7,7,8,8-tetracyano-
p-quinodimethane (TCNQ).
Much of our work in this area, as well as contributions from
other groups,4,5 has focused on variation of the alkyne donor to
modulate the efficiency of donor−acceptor conjugation; the
acceptor olefin has been largely limited to TCNE and TCNQ.
Only in a few instances have we explored the possibility of
tuning the acceptor molecule for this reaction, but these reports
have led to some exciting prospects.6 In particular, the notion
Even historically well-studied molecular systems can garner
newfound appreciation in novel applications. One such scaffold,
pentafulvene, has been investigated for over six decades, and the
riches of its chemistry have yet to be exhausted. This isomer of
benzene (C6H6) has been, and continues to be, a paradigm for the
illustration of aromaticity−antiaromaticity to the present day.7
Although it is a hydrocarbon, pentafulvene has a relatively large
dipole moment of 0.44 D8 owing to its nonbenzenoid connectivity.
The positive charge resides on C(6), while the negative charge is
delocalized in the five-membered ring. In the highest occupied
molecular orbital (HOMO), the electron population resides
exclusively in the cyclopentadienyl fragment, while in the lowest
unoccupied molecular orbital (LUMO), it can be found at C(6)
(Figure 1). Here, we focus our efforts on the underexplored 6,6-
dicyanopentafulvenes (DCFs), which have recently found renewed
Received: September 14, 2012
Published: October 8, 2012
© 2012 American Chemical Society
18139
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
green solid. The structure was confirmed by single crystal X-ray
analysis (Scheme 1b, see the Supporting Information (SI) for
details). The bulky silyl groups on the acetylene appear to be
important to stability. Starting from a tetraethynylated cyclo-
pentadienone, in which the terminal triisopropylsilyl (TIPS)
groups were replaced by smaller tert-butyl groups, condensation
with malononitrile gave 2b as a green solid in low yield, which
decomposed to a brown solid in a couple of weeks. Our attempts
to access terminally deprotected, tetraethynyl-DCF 2c via removal
of the TIPS groups of 2a with n-Bu4NF in THF unfortunately led
to an unidentifiable and inseparable mixture, and no 2c was found.
Optoelectronic Properties. With ethynylated DCF 2a in
hand, we then compared its electron-accepting ability to that of
previously reported tetraphenyl-DCF 3 (Table 1).10a The
Figure 1. HOMO and LUMO of pentafulvene.
interest as functional materials.9 Because DCF itself is too reactive
to isolate and study, we must rely on stable derivatives.
Andrew et al. have reported the first systematic preparation
of stable, phenylated DCFs as n-type additives in solar cells.10 They
found that despite the small size of the scaffold, DCFs were good
electron acceptors, capable of undergoing two reversible one-
electron reductions. We also recently described the serendipitous
formation of a stable push−pull DCF possessing a strong,
Table 1. Optical and Electronic Properties of Compounds
a
b
b
compound
λmax, nm (ε, M−1cm−1
)
Eox, V
Ered, V
−0.90
−1.50
−0.57
−1.13
−0.71
−1.29
−0.91
−1.50
−0.78
−1.07
−1.04
−1.17
−0.91
−1.08
−1.22
−1.38
1a
559 (1890)
698 (480)
−
2a
2b
−
+1.19
−
remarkably bathochromically shifted charge-transfer band (λmax
=
782 nm, ε = 27 500 M−1 cm−1) and amphoteric redox behavior.11,12
These findings compelled us to explore the preparation of new
DCFs and their reactivity toward the CA−RE reaction.
Here, we report the preparation of tetra-alkynylated DCFs and
compare their optoelectronic properties to those of the previously
reported tetraphenylated derivative.10a We describe an unusual
divergence in the regioselectivity of the CA−RE reaction between
differently substituted DCFs, which is explained in a combined
experimental and computational study, involving electron para-
magnetic resonance (EPR) spectroscopy of their corresponding
radical anions.
c
691 (670)
d
3
543 (400)
4
5
6
7
721 (24 700)
621 (23 900)
633 (23 600)
+0.54
c
+0.73
+0.49
c
+0.60
+0.49
c
+0.96
+0.70
453 (17 300)
a
b
RESULTS AND DISCUSSION
Measured in CH2Cl2 solution. Values measured by CV in CH2Cl2
■
solution (+ 0.1 M NBu4PF6), v = 0.1 V s−1. Values reported vs Fc/Fc+.
Preparation of Tetraalkynylated DCFs. We first
considered elongating the conjugated π-chromophore of the
DCF scaffold through substitution with alkynes, as they are
highly efficient at extending effective conjugation length.13 In
the work by Andrew et al., the most electronically interesting
DCFs were accessed from a stable starting cyclopentadienone,
which was condensed with malononitrile in a Lewis acid-
promoted Knoevenagel reaction.10a We adopted this approach
to the preparation of perethynylated DCFs (Scheme 1a), whose
cyclopentadienone precursors are known to be stable and
isolable.14 From tetrakis[(triisopropylsilyl)ethynyl]cyclopenta-
2,4-dienone 1a, DCF 2a was prepared in good yield as a stable,
All values correspond to one-electron oxidation or reduction steps.
c
d
Irreversible peak potential. Data from ref 10a.
electronic absorption spectrum of 2a in CH2Cl2 features a weak
absorption band at λmax = 698 nm (ε = 480 M−1 cm−1), which
occurs at a much lower energy compared to 3 but with a
similarly weak absorption strength (λmax = 543 nm, ε = 400
M−1 cm−1) (for the spectra, see the Supporting Information).
DCF 2a undergoes two reversible one-electron reductions in
CH2Cl2 at −0.57 and −1.13 V (vs Fc/Fc+); each occurs at
more anodic potential than the corresponding reductions of 3
a
Scheme 1. (a) Synthesis of 2a, 2b, and 3 and the Attempted Synthesis of 2c and (b) Crystal Structure of 2a at 100 K
a
Thermal ellipsoids shown at 50% probability.
18140
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
(Ered = −0.91, −1.50 V vs Fc/Fc+). tert-Butyl-substituted DCF
2b also features two reversible one-electron reductions at
slightly more cathodic potentials than 2a (Ered = −0.71 V,
−1.29 V vs Fc/Fc+) and also displays irreversible oxidation at
high potential (Eox = +1.19 V vs Fc/Fc+). We observe no
oxidation waves in the cyclic voltammograms (CVs) of 2a or 3.
The electrochemistry of 1a has not been previously reported,
and its analysis reveals decisive insights to the electronic
contribution of the alkyne substituents. The reduction
pathways of many cyclopentadienones, such as tetraphenylcy-
clopentadienone, consist of a single two-electron reduction
(Ered = −1.52 V vs Fc/Fc+).10a,15 By contrast, 1a displays two
reversible one-electron reductions. Clearly, ancillary substitu-
tion plays an important role in the electronics of these systems.
Figure 2 shows the calculated HOMO and LUMO energies
(B3LYP/TZVP) for different fulvene derivatives. By far the
Figure 3. (top) EPR spectrum attributed to 2a•− (THF, 294 K)
together with its simulation (taking into account 29Si satellites) and
(bottom) the corresponding UV−vis (MeTHF) spectra recorded
during the reduction. Reductions are nonstoichiometric and imply
fresh exposure to the Na mirror.
Figure 2. Calculated HOMO and LUMO energies (B3LYP/TZVP) of
pentafulvene, DCF, 3, and 2a. The arrows and the corresponding
numbers represent the HOMO−LUMO gap in eV. The blue numbers
in italics are the experimentally determined (first) redox potentials (in
V vs Fc/Fc+).
coupling constants of 0.11 and 0.06 mT (each for two 29Si,
calcd 0.12 and 0.08 mT, respectively) can be discerned. This
shows that spin is also transferred onto the silyl substituents
and explains their stabilizing effect in terms of lowering the
reduction potential compared to 2b. The reduction process of
2a is accompanied with the occurrence of new absorptions in
the visible region (green curve in Figure 3, new λmax at 560 and
620 nm). Excessive reduction by multiple exposures to fresh Na
mirror leads to the decay of these latter absorption bands and
disappearance of the EPR signal. Freezing the sample to 77 K
does not show any EPR spectrum, indicating that a triplet-state
species is not likely formed.
lowest HOMO−LUMO gap is computed for tetraalkynylated
2a. The effect of cyano substitution at C(6) is relatively simple
to understand. Theory predicts that the LUMO should be
stabilized by this substitution since the electron population is
delocalized into the electron-withdrawing cyano groups.7f,16
This is illustrated by the substantial decrease of the LUMO
energy for 6,6-dicyanofulvene compared to fulvene (Figure 2).
The effects of the tetraphenyl and tetrakis(TIPS-ethynyl)
substitution are more complex, because there is a subtle
interplay between the substituents on the cyclopentadiene ring
and the cyano groups, particularly in the LUMO. Substitution
on the cyclopentadiene ring also affects the HOMO strongly.
EPR of Radical Anions. An experimental exploration of the
LUMO structure of conjugated materials is best undertaken by
EPR analysis of the one-electron reduced species.17 The reversible
first reductions detected at −0.57 and −0.91 V for 2a and 3,
respectively, indicate that these derivatives should form rather
persistent radical anions. Indeed, reduction (Na metal, THF) leads
to intense EPR spectra.
The EPR signal attributed to 3•− is not resolved (Figure 4),
but application of the ENDOR technique reveals three different
isotropic 1H hfcs of 0.05, 0.02, and 0.007 mT. According to DFT
calculations, these hfcs can be assigned to the ortho- and para-
protons (virtually equivalent hfcs (4 + 2) H) and the meta-protons
(4 H) in the 3,4-phenyl groups and to the protons in the 2,5-
phenyl substituents (10 H) of 3•−. Together with a 14N hfc of 0.13
mT (2 N) a simulation that matches the experimental spectrum
can be obtained (Figure 4). The electron distribution represented
by these hyperfine data is in good agreement with the calculated
geometry, which has also been confirmed by X-ray structural
analysis.10a The two phenyl substituents at C(2) and C(5) are
oriented almost perpendicular toward the plane of the fulvene,
while those at C(3) and C(4) are only tilted by ca. 50°.
Accordingly, owing to the substantial orbital coefficients at C(3)
and C(4) and the more coplanar arrangement of the phenyl
groups, the charge and the spin are mainly delocalized into the two
phenyl groups at C(3) and C(4).
Reduction of 2a produces a well-resolved EPR spectrum
(Figure 3) centered at the g factor of 2.0026. The main five-line
signal pattern is produced by hyperfine coupling constants
(hfcs) of the two equivalent 14N nuclei (14N hfc 0.13 mT, calcd
0.13 mT). The additional EPR lines of weaker intensity are
produced by 29Si satellite lines. Here, two 29Si hyperfine
18141
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
in alkynylated 2a and phenylated 3 and their one-electron reduced
stages (B3LYP/TZVP) (Table 2SI). At the neutral stage, the bond
lengths of the cyclopentadienyl ring in pentafulvene, DCF, and 2a
are typical for an alternating π-system [(C1)−C(2) and C(3)−
C(4) 1.47−1.48 Å; C(2)−C(3) 1.35−1.38 Å]. In 3, the calculated
C(3)−C(4) distance reaches 1.51 Å with the adjacent bonds
shortened, compared to 2a, to 1.36 Å. Evidently, the two phenyl
rings at C(3) and C(4) substantially contribute to electron
delocalization in 3 (Figure 28SI). This effect is diminished in
radical anion 3•−. In addition, the phenyl-substituted derivative 3
displays a different charge distribution than alkynylated 2a. In
particular, this becomes apparent when comparing the partial
atomic charges of 2a and 3 (Figures 29SI and 30SI). At the parent
neutral stage, C(1) is negatively polarized in 2a, but it carries a
slightly positive charge in 3, pointing to the five-membered ring
having a reduced cyclopentadienyl-anionic character. For the radical
anion, the exocyclic C(6) in phenyl derivative 3•− carries a slightly
negative charge, whereas in 2a•− it is positively polarized (SI). This
distinct charge separation helps to explain the reactivity profiles seen
below for the CA−RE reactions.
CA−RE Reactivity with Electron-Rich Alkynes. When
we heated DCF 2a with 4-ethynyl-N,N-dimethylaniline to 80 °C
in acetonitrile for 27 h, CA−RE adduct 4 was isolated as the
exclusive product in 64% yield (Table 2a). Bis-anilino-substituted
ethyne and buta-1,3-diyne react similarly to give adducts 5 (97%)
and 6 (47%), respectively. In these cases, higher temperatures and
microwave heating lead to faster product formation without
decomposition.18 Crystals of 5 suitable for X-ray analysis
were grown, confirming the identity of the CA−RE product
(Figure 5a).19 In the case of 6, only monoaddition was observed;
the alkyne of 6 does not react with 2a even at high temperatures.
Figure 4. (top) EPR (experimental and simulated) spectra assigned to
3•− and (bottom) the corresponding ENDOR spectrum in THF.
In the context of the structure of the radical anions, it is
noteworthy to compare the electron distribution and geometry
Table 2a. Synthesis of Fulvenoid CA−RE Adducts from 2a
a
Microwave heating.
18142
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
Figure 5. Crystal structures of (a) 5 (hydrogens omitted for clarity) and (b) 7 at 100 K. Thermal ellipsoids shown at 50% probability.
Table 2b. Synthesis of Fulvenoid CA−RE Adducts from 3
a
No reaction. Heating to 200 °C leads to decomposition of 3.
Oddly, t-butylethynyl derivative 2b decomposes upon reaction
with electron-rich alkynes at room temperature to give
unidentifiable mixtures. Push−pull chromophores 4−6 all feature
strong, intramolecular charge-transfer (CT) bands that extend into
the near-IR region, originating from the efficient conjugation from
the anilino donor to the fulvene acceptor (Figure 11SI). The
requirement of the aniline donor for intramolecular CT was
demonstrated for push−pull chromophore 5 by treating a solution
in CH2Cl2 with CF3COOH, leading to attenuation of the CT
band; the CT band reappeared when the acidified solution was
treated with NEt3 (Figure 13SI). Furthermore, 5 displays positive
solvatochromism (Figure 14SI).
The regioselectivity of the CA−RE reaction to form 4−6 is
identical to that previously observed in the transformations with
TCNQ and N,N′-dicyanoquinone diimides (DCQNIs);6d that is,
the anilino donor group is directly π-conjugated to the ring
structure, enabling a proaromatic resonance structure in the
acceptor ring.20 The same proaromatic resonance is established in
4−6, but this time with the five-membered cyclopentadiene ring.
We expected that tetraphenyl-DCF 3 would react with electron-
rich alkynes similarly to 2a. It does not. Heating 3 with 4-ethynyl-
N,N-dimethylaniline in MeCN for 48 h at 80 °C leads exclusively
to CA−RE adduct 7, resulting from opposite regiochemistry in the
CA−RE reaction (Table 2b). Bis-anilino-substituted alkynes do
not react with 3, even when heated to 200 °C in a sealed tube;
decomposition of 3 is ultimately observed instead.
The UV−vis spectrum of 7 features a strong, hypsochromically
shifted (compared to 3) intramolecular CT band from the anilino
donor to the dicyanovinyl acceptor (Figure 12SI, Supporting
Information). The regiochemistry of 7 was confirmed by X-ray
crystallographic analysis (Figure 5b). The anilino donor is now in
direct linear π-conjugation with the adjacent dicyanovinyl group
instead of the fulvene. The CV of 7 features two reversible one-
electron reductions, which occur at more cathodic potentials
compared to 3 as well as two one-electron oxidations at +0.70 and
+0.96 V (vs Fc/Fc+) (Table 1).
Mechanism of the CA−RE Reaction. While 7 is a rather
unremarkable chromophore in and of itself, the formation of 7 was
18143
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
Scheme 2. Reaction Pathways in the CA−RE Reaction with 6,6-Dicyanofulvenes
unexpected. This is the first time we have observed regioselectivity
of this kind; when there is a choice between generation of a
proaromatic structure and a structure with the donor in direct
conjugation with a dicyanovinylidene unit, the proaromatic
pathway has been preferred. More curious, still, is the fact that
the substitution on the fulvene determines the regioselectivity of
the CA−RE reaction. Previous work on the mechanism of the
CA−RE reaction with dicyanovinyl acceptors6b led to the
mechanisms proposed in Scheme 2.
The first step involves nucleophilic attack of the alkyne at the
dicyanovinyl moiety to generate a zwitterionic intermediate.
Because the site of initial nucleophilic attack is the arbiter of
regioselectivity in this reaction, we focused our computational
studies on the formation of the zwitterionic intermediate. Two
potential reaction pathways are possible (Scheme 2). In Pathway 1
(P1), nucleophilic attack occurs at C(6) to generate the aromatic,
zwitterionic intermediate. In Pathway 2 (P2), nucleophilic attack
occurs at C(1) to generate an intermediate in which the negative
charge is delocalized across the malononitrile moiety.
exocyclic C(6) in the radical anion carries a slightly negative
charge. These data provide further support that attack at C(1),
and thus pathway P2, is preferred.
CONCLUSION
■
We have synthesized a peralkynlated DCF 2a and compared its
electronic structure and properties to that of the previously
investigated tetraphenylated DCF 3. The enhanced susceptibility
to one-electron reduction in 2a is suggested by EPR analysis of the
radical anions, showing that all ancillary silylalkynyl groups in 2a
participate in the stabilization of the radical anion, whereas in 3 the
majority of the stabilization only results from the phenyl rings in
the 3- and 4-position. Perethynylation completely changes the
regioselectivity in the CA−RE reaction of 6,6-dicyanofulvenes with
electron-rich alkynes, as compared to perphenylated analogues,
thereby generating a new family of interesting push−pull
chromophores. The electron transport properties are currently
being studied, which should give insight into the utility of push−
pull-substituted fulvenes in functional organic materials.
To gain insight into the observed regioselectivity, we turned
to computational methods. Due to computational restrictions,
calculations were performed with DCF 2c instead of 2a, at the
wB97XD/6-31G(d)21 level of theory, with polarizable con-
tinuum model (PCM) solvation in CHCl3 (Gaussian 09, see SI
for details). The calculated transition state (TS) and
intermediate energies corroborate our empirical findings. For
2c, pathway P1 was lower in energy than P2 for both TS
(ΔΔG‡ = 1.3 kcal mol−1) and intermediate formation (ΔΔG =
3.4 kcal mol−1) (Figure 31SI). In contrast, the computational
results revealed that reaction of 3 preferred pathway P2 over P1
for both TS (ΔΔG‡ = 4.0 kcal mol−1) and intermediate
formation (ΔΔG = 16.9 kcal mol−1) (Figure 33SI).
The preference of tetraethynyl derivative 2c for pathway P1 seems
to be the result of the additional efficient alkyne conjugation with the
cyclopentadienyl ring. This extra conjugation, rather than aromaticity,
appears to determine the regioselectivity of the reaction. The extra
conjugation is supported by the EPR data of 2a•−: strong hyperfine
coupling to the terminal silicon atoms implies efficient communi-
cation across the alkynyl moieties. The site of attack on 2a in
pathway P1 is at C(6), which carries a positive polarization in 2a•−.
For 3, computations and EPR data show that phenyl
substitution, particularly at C(3) and C(4), reduces the anionic
character on the cyclopentadienyl ring (see above). Atom C(1)
in neutral 3 carries a distinctly positive charge, whereas the
EXPERIMENTAL SECTION
■
General materials and physical characterization data can be found in the
SI. Compounds 1a,14a 1b,14a 310a and 1,4-bis[(4′-N,N-dimethylamino)-
phenyl]buta-1,3-diyne22 were prepared according to literature procedures.
EPR Studies. Chemical Reductions. Reductions were performed in
dry THF or MeTHF with the use of Na metal. THF was heated to
reflux over Na/K alloy and stored over Na/K alloy under high vacuum.
Benzophenone ketyl was used as an indicator for rigorously water-free
conditions. Samples were prepared in a special three-compartment
EPR sample tube connected to the vacuum line. Na metal mirror was
sublimed to the wall of the tube, and about 0.4 mL of THF was freshly
condensed to dissolve the investigated compound. The sample was
successively degassed by three freeze−pump−thaw cycles and sealed
under high vacuum. Reductions were performed by contact of the
THF solution of the parent molecule with the Na metal mirror in the
evacuated sample tube. The sample tubes were either stored at 190 K
(isopropanol/dry ice bath) or immediately transferred into the
microwave cavity of the EPR spectrometer. Each contact of the solution
with the Na mirror leads to a stepwise reduction of the substrate by Na.
Usually, after the initial short contact of the Na metal with the solution the
EPR signal grows, and after additional contact, the signal decreases. This is
due to the fact that primarily a basically quantitative one-electron
reduction to the radical anion is achieved, then, gradually, the uptake of a
second electron proceeds (such a behavior is illustrated in Figure 3).
EPR and ENDOR Measurements. EPR and ENDOR spectra were
recorded using two ESP300 (Bruker, Germany) X-band spectrometers
(one equipped with an ENDOR unit). Temperature control was achieved
18144
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
using the EUROTHERM variable temperature units. Typical spectrometer
settings for recording the solution EPR spectra were: microwave power =
0.2 mW, modulation amplitude = 0.015 mT, modulation frequency = 100
kHz, conversion time = 40.96 ms, time constant = 10.24 ms, and number
of points = 2048. ENDOR spectra were recorded with microwave power =
10 mW, modulation amplitude = 0.1 mT, modulation frequency = 12.5
kHz, pump frequency = 10 MHz, modulation depth = 100 kHz and
field locked at them maximum of the EPR line. Spectra were analyzed
with WinEPR and SimFonia software provided by the spectrometers
manufacturer as well as with WinSim,23 a public domain program.
Synthesis of New Compounds. 2-{2,3,4,5-Tetrakis[(triisopropyl-
silyl)ethynyl]cyclopenta-2,4-dien-1-ylidene}malononitrile (2a). A
solution of 1a (500 mg, 0.624 mmol) and malononitrile (82 mg,
1.24 mmol) in dry CH2Cl2 (10 mL) was cooled to 0 °C. TiCl4 (0.35 mL,
3.11 mmol) was added dropwise, followed by dropwise addition of
pyridine (1 mL). The solution was stirred at room temperature overnight.
Water (20 mL) was added, the layers were separated, then the organic
layer was washed with 10% aqueous HCl, dried over MgSO4, and the
solvent evaporated. Column chromatography (hexane/CH2Cl2 5:1) gave
460 sh (14600), 381 (30800). Anal. calcd for C62H95N3Si4: C, 74.86;
H, 9.63; N, 4.22. Found: C, 74.61; H, 9.45; N 4.25.
2-{1,2-Bis[4-(dimethylamino)phenyl]-2-[2,3,4,5-tetrakis(triisopropyl-
silylethynyl)-cyclopenta-2,4-dien-1-ylidene]ethylidene}malononitrile
(5). A 2 mL microwave tube with septum cap was charged with 2a
(45 mg, 0.053 mmol) and 1,2-bis[(4′-N,N-dimethylamino)phenyl]ethyne
(14 mg, 0.053 mmol). The tube was sealed, pumped, and purged with N2
gas via a needle inserted through the septum, and MeCN (0.5 mL) and
1,2-dichloroethane (0.5 mL) were injected via syringe. The resulting
suspension was heated to 160 °C for 30 min in the microwave. The
solvent was evaporated, and purification by recycling preparatory
1
GPC gave a metallic blue solid (57 mg, 97%); mp 266 °C (dec). H
NMR (600 MHz, CDCl3) δ 7.64 (d, J = 9.1 Hz, 2H), 7.16 (br s, 2H),
6.53 − 6.44 (m, 4H), 2.99 (s, 6H), 2.97 (s, 6H), 1.04 (dd, J = 5.2, 2.3 Hz,
42H), 1.00 (d, J = 2.4 Hz, 12H), 0.92 (dd, J = 5.0, 2.2 Hz, 9H), 0.84 (t,
J = 7.7 Hz, 18H), 0.74−0.64 (m, 3H). 13C NMR (151 MHz, CDCl3) δ
171.05 (C), 153.06 (C), 152.29 (C), 151.74 (C), 140.16 (C), 133.36 (CH),
133.29 (C), 133.16 (C), 126.95 (C), 123.05 (C), 121.79 (C), 120.72 (C),
116.00 (C), 114.11 (C), 111.49 (C), 111.26 (CH), 103.63 (C), 103.11 (C),
102.48 (C), 102.01 (C), 101.47 (C), 101.27 (C), 97.29 (C), 79.00 (C),
39.97 (CH3), 39.83 (CH3), 19.03 (CH3), 18.88 (CH3), 18.87 (CH3), 18.82
(CH3), 12.14 (CH), 11.93 (CH), 11.78 (CH), 11.60 (CH). Four signals,
attributed to the aromatic carbons of the 4-(dimethylamino)phenyl ring
attached to C(6) of the fulvene, are absent. This is likely due to rotation of
the phenyl ring on the NMR time scale as evidenced in the 1H NMR. HR-
MS (MALDI) [M + Na]+ calcd for C70H104N4NaSi4: 1135.7230; found:
1135.7227. IR (ATR) 2941 (m), 2863 (m), 2216 (w), 2122 (w), 1600 (s),
1493 (m), 1462 (m), 1437 (m), 1372 (s), 1330 (m), 1231 (w), 1206 (w),
1170 (s), 1088 (m), 1072 (m), 1017 (w), 984 (m), 944 (m), 881 (s), 816
(m), 783 (m), 766 (w), 753 (w), 719 (w), 673 (s). UV−vis (CH2Cl2) λmax
(ε) 621 (23900), 398 (30100). Anal. calcd for C70H104N4Si4: C, 75.48; H,
9.41; N, 5.03. Found: C, 75.20; H, 9.33; N 5.04.
2-{1,4-Bis[4-(dimethylamino)phenyl]-1-[2,3,4,5-tetrakis(triisopropyl-
silylethynyl)-cyclopenta-2,4-dien-1-ylidene]but-3-yn-2-ylidene}-
malononitrile (6). A 2 mL microwave tube with septum cap was charged
with 2a (170 mg, 0.2 mmol) and 1,4-bis[(4′-N,N-dimethylamino)-
phenyl]buta-1,3-diyne (29 mg, 0.1 mmol). The tube was sealed, pumped,
and purged with N2 gas via a needle inserted through the septum, and
MeCN (0.5 mL) and 1,2-dichloroethane (0.5 mL) were injected via
syringe. The resulting suspension was heated to 160 °C for 30 min in the
microwave. Full conversion of the diyne to 6 was observed. The mixture
was then heated in the microwave to 180 °C for 30 min in an attempt to
induce reaction of the other alkyne, but no further reaction was observed.
The solvent was evaporated, and purification by preparatory GPC gave 6
as a metallic blue solid (52 mg, 47%); mp 216 °C (dec.) 1H NMR (600
MHz, CD2Cl2) δ 7.33 (d, J = 9.1 Hz, 2H), 7.31 (d, J = 9.1 Hz, 2H), 6.66
(d, J = 9.1 Hz, 2H), 6.62 (d, J = 9.1 Hz, 2H), 3.09 (s, 6H), 3.04 (s, 6H),
1.13 (d, J = 3.1 Hz, 45H), 1.06 (dd, J = 12.7, 7.2 Hz, 18H), 0.92 (dd, J =
7.4, 2.3 Hz, 18H), 0.83−0.72 (m, 3H). 13C NMR (151 MHz, CD2Cl2) δ
155.49 (C), 153.60 (C), 152.91 (C), 148.27 (C), 140.74 (C), 137.55
(CH), 135.69 (CH), 133.40 (C), 133.19 (C), 125.55 (C), 122.84 (C),
121.21 (C), 121.06 (C), 114.88 (C), 112.93 (C), 112.04 (CH), 111.94
(C), 107.07 (C), 104.72 (C), 104.13 (C), 102.86 (C), 102.60 (C), 101.92
(C), 101.87 (C), 101.78 (C), 98.51 (C), 92.90 (C), 90.69 (C), 40.39
(CH3), 40.33 (CH3), 19.37 (CH3), 19.31 (CH3), 19.25 (CH3), 19.18
(CH3), 12.56 (CH), 12.43 (CH), 12.34 (CH), 12.19 (CH). HR-MS
(MALDI) [M + Na]+ calcd for C72H104N4Si4: 1159.7230; found:
1159.7225. IR (ATR) 2941 (m), 2863 (m), 2223 (w), 2143 (m), 1599
(s), 1529 (m), 1462 (m), 1441 (m), 1371 (m), 1223 (w), 1167 (m), 1115
(m), 1072 (w), 1017 (w), 995 (w), 941 (w), 881 (m), 816 (w), 800 (w),
779 (w), 762 (w), 721 (w), 674 (s). UV−vis (CH2Cl2) λmax (ε) 633
(23600), 411 (29000). Anal. calcd for C72H104N4Si4: C, 75.99; H, 9.21; N,
4.92. Found: C, 75.73; H, 9.12; N, 4.88.
1
2a (0.392 g, 74%) as a green solid; mp 234 °C. H NMR (400 MHz,
CDCl3) δ 1.14 (m, 84H, 4 Si(CHMe2)3). 13C NMR (101 MHz, CDCl3)
δ 159.04 (C), 138.64 (C), 120.42 (C), 115.49 (C), 111.36 (C), 110.62
(C), 100.14 (C), 96.86 (C), 86.02 (C), 18.87 (CH3), 18.81 (CH3), 11.68
(CH), 11.45 (CH). HR-MS (ESI+) [M + H]+ calcd for C52H85N2Si4:
849.5784; found: 849.5766. IR (ATR) 2942 (m), 2890 (w), 2864 (m),
2187 (m), 2157 (w), 2104 (m), 1462 (m), 1427 (w), 1383 (m), 1354
(m), 1313 (w), 1231 (w), 1120 (w), 1074 (w), 1018 (m), 994 (m), 938
(m), 919 (w), 882 (s), 840 (m), 674 (s). UV−vis (CH2Cl2) λmax (ε) 698
(480), 401 (16600). Anal. calcd for C52H84N2Si4: C, 73.51; H, 9.97; N,
3.30. Found: C, 73.22; H, 9.99; N, 3.39.
2-[2,3,4,5-Tetrakis(3,3-dimethylbut-1-yn-1-yl)cyclopenta-2,4-dien-
1-ylidene)malononitrile (2b). A solution of 1b (500 mg, 1.25 mmol)
and malononitrile (165 mg, 2.50 mmol) in dry CH2Cl2 (20 mL) was
cooled to 0 °C. TiCl4 (0.69 mL, 6.25 mmol) was added dropwise,
followed by dropwise addition of pyridine (2 mL). The solution was
stirred at room temperature overnight. Water (20 mL) was added, the
layers were separated, then the organic layer was washed with 10%
aqueous HCl, dried over MgSO4, and the solvent evaporated. Column
chromatography (hexane/CH2Cl2 3:2) gave 2b (0.200 g, 36%) as a
green solid. Compound 2b decomposed within a week in CH2Cl2
solution and over the course of several weeks as a solid. 1H NMR (400
MHz, CDCl3) δ 1.34 (s, 18H, 2 CMe3), 1.32 (s, 18H, 2 CMe3). 13C
NMR (101 MHz, CDCl3) δ 160.72 (C), 138.87 (C), 119.44 (C),
118.39 (C), 112.89 (C), 111.67 (C), 83.99 (C), 75.00 (C), 71.64 (C),
30.79 (CH3), 30.54 (CH3), 29.29 (C), 29.07 (C). HR-MS (MALDI+)
[M]+ calcd for C32H36N2: 448.2873; found: 448.2873. UV−vis
(CH2Cl2) λmax (ε) 691 (670).
2-{2-[4-(Dimethylamino)phenyl]-2-[2,3,4,5-tetrakis((triisopropylsilyl)-
ethynyl)cyclopenta-2,4-dien-1-ylidene]ethylidene}malononitrile (4).
Compound 2a (85 mg, 0.1 mmol) and 4-ethynyl-N,N-dimethylaniline
(14.5 mg, 0.1 mmol) were added to a 20 mL sealable vial with septum.
The vessel was sealed, pumped, and purged with N2 gas three times,
and MeCN (1 mL) was added. The resulting suspension was stirred at
80 °C for 27 h. The solvent was removed by rotary evaporation.
Column chromatography (hexane/CH2Cl2 3:2) gave 4 (64 mg, 64%)
1
as a green solid; mp 220 °C (dec). H NMR (400 MHz, CDCl3) δ
8.71 (s, 1H), 7.15 (d, J = 8.9 Hz, 2H), 6.67 (d, J = 9.0 Hz, 2H), 3.10
(s, 6H), 1.21−1.09 (m, 64H), 0.94 (d, J = 7.2 Hz, 20H). 13C NMR
(101 MHz, CDCl3) δ 161.60 (CH), 152.67 (C), 144.03 (C), 143.58
(C), 135.91 (CH), 135.00 (C), 134.47 (C), 125.51 (C), 122.01 (C),
119.98 (C), 113.92 (C), 111.94 (CH), 110.65 (C), 107.10 (C), 106.36
(C), 105.42 (C), 102.43 (C), 101.27 (C), 101.08 (C), 101.05 (C),
99.70 (C), 92.17 (C), 39.87 (CH3), 18.83 (CH3), 18.80 (CH3), 11.77
(CH), 11.70 (CH). HR-MS (MALDI+) [M]+ calcd for C62H95N3Si4:
993.6898; found: 993.6593. IR (ATR) 2942 (m), 2889 (m), 2863 (m),
2224 (w), 2120 (w), 1600 (s), 1540 (w), 1501 (s), 1460 (m), 1440
(m), 1370 (s), 1327 (m), 1232 (w), 1207 (w), 1186 (s), 1131 (w),
1097 (m), 1072 (w), 1017 (w), 995 (m), 983 (m), 945 (m), 920 (w),
882 (s), 837 (w), 816 (w). UV−vis (CH2Cl2) λmax (ε) 721 (24700),
2-{1-[4-(Dimethylamino)phenyl]-2-[2,3,4,5-tetraphenylcyclopenta-
2,4-dien-1-ylidene]ethylidene}malononitrile (7). Compound 3 (43 mg,
0.1 mmol) and 4-ethynyl-N,N-dimethylaniline (14.5 mg, 0.1 mmol)
were added to a 20 mL vial with septum cap. The tube was sealed,
pumped, and purged with N2 gas twice, and MeCN (1 mL) was added
via syringe, and heated to 80 °C for 48 h. The solvent was evaporated.
Column chromatography (CH2Cl2/hexane 4:1) gave a red solid (37 mg,
18145
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146

Journal of the American Chemical Society
Article
64%); mp 280 °C (dec). 1H NMR (400 MHz, CDCl3) δ 7.47 (d, J = 9.2
Hz, 2H), 7.38−7.29 (m, 5H), 7.19−6.93 (m, 11H), 6.91 (s, 1H), 6.90−
6.78 (m, 2H), 6.56 (d, J = 9.2 Hz, 2H), 3.10 (s, 6H). 13C NMR (101
MHz, CDCl3) δ 166.99 (C), 153.34 (C), 152.95 (C), 148.34 (C),
144.41 (C), 135.53 (C), 134.51 (C), 134.39 (C), 134.25 (C), 132.61
(CH), 132.07 (CH), 131.54 (C), 131.41 (CH), 130.52 (CH), 130.28
(CH), 130.23 (CH), 128.33 (CH), 127.64 (CH), 127.54 (CH), 127.32
(CH), 127.27 (C), 127.09 (CH), 121.50, 111.18 (CH), 40.14 (CH3).
HR-MS (MALDI) [M + H]+ calcd for C42H32N3: 578.2591; found:
578.2593. IR (ATR) 2922 (w), 2207 (m), 1603 (s), 1538 (w), 1483 (s),
1435 (m), 1412 (w), 1383 (m), 1338 (m), 1294 (w), 1211 (m), 1170
(m), 1114 (w), 1073 (w), 1025 (w), 908 (w), 821 (m), 812 (m), 673
(s). UV−vis (CH2Cl2) λmax (ε) 453 (17300), 355 (18000). Anal. calcd
for C42H31N3: C, 87.32; H, 5.41; N, 7.27. Found: C, 85.63; H, 5.54; N,
7.16. We were unable to obtain satisfactory elemental analysis for this
compound, despite multiple attempts.
F. Chem.Eur. J. 2011, 17, 6088−6097. (d) Chiu, M.; Jaun, B.; Beels,
M. T. R.; Biaggio, I.; Gisselbrecht, J.−P.; Boudon, C.; Schweizer, W.
B.; Kivala, M.; Diederich, F. Org. Lett. 2011, 14, 54−57.
(7) (a) Coulson, C. A.; Craig, D. P.; Maccoll, A. Proc. Phys. Soc. 1948,
61, 22−25. (b) Fukui, K.; Imamura, A.; Yonezawa, T.; Nagata, C. Bull.
Chem. Soc. Jpn. 1960, 33, 1591−1599. (c) Baird, N. C. J. Am. Chem.
Soc. 1972, 94, 4941−4948. (d) Ottosson, H.; Kilsa, K.; Chajara, K.;
Piqueras, M. C.; Crespo, R.; Kato, H.; Muthas, D. Chem.Eur. J.
2007, 13, 6998−7005. (e) Mitchell, R. H.; Zhang, R.; Berg, D. J.;
Twamley, B.; Williams, R. V. J. Am. Chem. Soc. 2009, 131, 189−199.
(f) Dahlstrand, C.; Rosenberg, M.; Kilsa, K.; Ottosson, H. J. Phys.
Chem. A 2012, 116, 5008−5017.
(8) Brown, R. D.; Burden, F. R.; Kent, J. E. J. Chem. Phys. 1968, 49,
5542.
(9) (a) King, R. B.; Saran, M. S. J. Chem. Soc., Chem. Commun. 1974,
851−852. (b) Junek, H.; Uray, G.; Zuschnig, G. Liebigs Ann. Chem.
1983, 154−158. (c) Junek, H.; Uray, G.; Zuschnig, G. Dyes Pigm.
1988, 9, 137−152. (d) Katritzky, A. R.; Fan, W.−Q.; Liang, D.−S.; Li,
Q.−L. J. Heterocycl. Chem. 1989, 26, 1541−1545.
(10) (a) Andrew, T. L.; Cox, J. R.; Swager, T. M. Org. Lett. 2010, 12,
5302−5305. (b) Andrew, T. L.; Bulovic, V. ACS Nano 2012, 6, 4671−
4677.
ASSOCIATED CONTENT
* Supporting Information
■
S
Synthesis and full characterization of new compounds, X-ray
crystallographic data, coordinates for computational data. This
data is available free of charge via the Internet at http://pubs.
acs.org.
(11) Jayamurugan, G.; Gisselbrecht, J.−P.; Boudon, C.; Schoenebeck,
F.; Schweizer, W. B.; Bernet, B.; Diederich, F. Chem. Commun. 2011,
47, 4520−4522. The reactivity of this DCF derivative with electron-
rich alkynes is complex and will be discussed elsewhere.
(12) A similar DCF was recently prepared as a byproduct of a
secondary CA−RE reaction of an azulene-containing push−pull alkyne
with TCNE: Shoji, T.; Ito, S.; Okujima, T.; Morita, N. Org. Biomol.
Chem. 2012, 10, 8308−8313.
AUTHOR INFORMATION
Corresponding Author
diederich@org.chem.ethz.ch
■
Notes
The authors declare no competing financial interest.
(13) Diederich, F. Chem. Commun. 2001, 219−227.
(14) (a) Jux, N.; Holczer, K.; Rubin, Y. Angew. Chem., Int. Ed. Engl.
1996, 35, 1986−1990. (b) Tobe, Y.; Kubota, K.; Naemura, K. J. Org.
Chem. 1997, 62, 3430−3431.
ACKNOWLEDGMENTS
■
This work was supported by the ERC Advanced grant no.
246637 (“OPTELOMAC”). A.D.F. acknowledges the NSF−
IRFP (U.S.A.) for a postdoctoral fellowship. O.D. thanks the
Fonds der Chemischen Industrie (Germany) for a Kekule
fellowship. We thank Mr. Eirik Lyngvi for helpful discussions
regarding computations.
(15) Fox, M. A.; Campbell, K.; Maier, G.; Franz, L. H. J. Org. Chem.
1983, 48, 1762−1765.
(16) (a) Dahlstrand, C.; Yamazaki, K.; Kilsa, K.; Ottosson, H. J. Org.
Chem. 2010, 75, 8060.
́
(17) (a) Kivala, M.; Stanoeva, T.; Michinobu, T.; Frank, B.;
Gescheidt, G.; Diederich, F. Chem.Eur. J. 2008, 14, 7638−7647.
(b) Breiten, B. Jordan, M. Taura, D. Zalibera, M. Greisser, M.;
Confortin, D.; Boudon, C.; Gisselbrecht, J.−P.; Schweizer, W. B.;
Gescheidt, G.; Diederich, F. J. Org. Chem. 2012, Article ASAP. DOI:
10.1021/jo301194y
(18) We do not ascribe any sort of special chemical reactivity to
microwave irradiation; we merely find it to be a fast and safe method
of heating, especially for heating solvents beyond their boiling point.
Conventional heating gives similar results.
REFERENCES
■
(1) (a) Organic Light−Emitting Devices: Synthesis, Properties and
Applications; Mullen, K., Scherf, U., Eds.; Wiley-VCH: Weinheim,
̈
2006. (b) Organic Photovoltaics; Brabec, C., Dyakonov, V., Scherf, U.,
Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany,
2008.
(2) (a) Paci, I.; Johnson, J. C.; Chen, X.; Rana, G.; Popovic, D.;
David, D. E.; Nozik, A. J.; Ratner, M. A.; Michl, J. J. Am. Chem. Soc.
2006, 128, 16546−16553. (b) Henson, Z. B.; Mullen, K.; Bazan, G. C.
Nature Chem. 2012, 4, 699−704.
(3) (a) Kivala, M.; Diederich, F. Acc. Chem. Res. 2009, 42, 235−248.
(b) Kato, S.; Diederich, F. Chem. Commun. 2010, 46, 1994−2006.
(4) For examples of recent work by others, see: (a) Niu, S.; Ulrich,
G.; Retailleau, P.; Ziessel, R. Org. Lett. 2011, 13, 4996−4999. (b) Shoji,
T.; Ito, S.; Okujima, T.; Morita, N. Eur. J. Org. Chem. 2011, 5134−
5140. (c) Morimoto, M.; Murata, K.; Michinobu, T. Chem. Commun.
(19) It was not possible to grow X-ray quality crystals of 4 and 6.
This seems to be due in part to the highly disordered TIPS groups,
which can be seen particularly in the crystal structure of 5.
̈
̌
(20) Wu, Y.-L.; Bures, F.; Jarowski, P. D.; Schweizer, W. B.; Boudon,
C.; Gisselbrecht, J.-P.; Diederich, F. Chem.Eur. J. 2010, 16, 9592−
9605.
(21) This method accounts for favorable aromatic dispersion
interactions. See: (a) Grimme, S. J. J. Comput. Chem. 2004, 25,
1463−1473. (b) Chai, J.−D.; Head−Gordon, M. Phys. Chem. Chem.
Phys. 2008, 10, 6615−6620.
(22) Rodríguez, J. G.; Lafuente, A.; Martín-Villamil, R.; Martínez-
Alcazar, M. P. J. Phys. Org. Chem. 2001, 14, 859−868.
(23) Duling, D. R. J. Magn. Reson., B 1994, 104, 105−110.
̀
2011, 47, 9819−9821. (d) Leliege, A.; Blanchard, P.; Rousseau, T.;
Roncali, J. Org. Lett. 2011, 13, 3098−3101. (e) Washino, Y.;
Michinobu, T. Macromol. Rapid Commun. Lett. 2011, 32, 644−648.
(5) The group of Bruce has pioneered the chemistry of TCNE with
alkyne−transition metal complexes. For a recent review, see: Bruce, M.
I. Aust. J. Chem. 2011, 64, 77−103.
(6) (a) Jarowski, P. D.; Wu, Y.−L.; Boudon, C.; Gisselbrecht, J.−P.;
Gross, M.; Schweizer, W. B.; Diederich, F. Org. Biomol. Chem. 2009, 7,
1312−1322. (b) Wu, Y.−L.; Jarowski, P. D.; Schweizer, W. B.;
Diederich, F. Chem.Eur. J. 2010, 16, 202−211. (c) Silvestri, F.;
Jordan, M.; Howes, K.; Kivala, M.; Rivera−Fuentes, P.; Boudon, C.;
Gisselbrecht, J.−P.; Schweizer, W. B.; Seiler, P.; Chiu, M.; Diederich,
18146
dx.doi.org/10.1021/ja309141r | J. Am. Chem. Soc. 2012, 134, 18139−18146