Mechanistic investigation of the dipolar [2+2] cycloaddition-cycloreversion reaction between 4-(N,N-dimethylamino)phenylacetylene and arylated 1,1-dicyanovinyl derivatives to form intramolecular charge-transfer chromophores

Article abstract of DOI:10.1002/chem.200902645

The kinetics and mechanism of the formal [2 + 2] cycloaddition- cycloreversion reaction between 4-(N,Ndimethylamino)phenylacetylene (1) and para-substituted benzylidenemalononitriles 2b-21 to form 2-donor-substituted 1,1-dicyanobuta-1, 3-dienes 3b31 via the postulated dicyanocyclobutene intermediates 4b-41 have been studied experimentally by the method of initial rates and computationally at the unrestricted B3LYP/6-31G(d) level. The transformations were found to follow bimolecular, second-order kinetics, with ΔH^≠_exp = 13-18 kcal mol^-1, Δ^≠_exp≈-30 cal K^-1 mol^-1, and ΔG^≠_exp = 22-27 kcal mol^-1. These experimental activation parameters for the rate-determining cycloaddition step are close to the computational values. The rate constants show a good linear free energy relationship (ρ = 2.0) with the electronic character of the para-substituents on the benzylidene moiety in dimethylformamide (DMF), which is indicative of a dipolar mechanism. Analysis of the computed structures and their corresponding solvation energies in acetonitrile suggests that the rate-determining attack of the nucleophilic, terminal alkyne carbon onto the dicyanovinyl electrophile generates a transient zwitterion intermediate with the negative charge developing as a stabilized malononitrile carbanion. The computational analysis predicted that the cycloreversion of the postulated dicyanocyclobutene intermediate would become rate-determining for 1,1-dicyanoethene (2 m) as the electrophile. The dicyanocyclobutene 4 m could indeed be isolated as the key intermediate from the reaction between alkyne 1 and 2 m and characterized by X-ray analysis. Facile first-order cycloreversion occurred upon further heating, yielding as the sole product the 1,1-dicyanobuta-1,3-diene 3m.

Full text of DOI:10.1002/chem.200902645

DOI: 10.1002/chem.200902645
Mechanistic Investigation of the Dipolar [2+2] Cycloaddition–
Cycloreversion Reaction between 4-(N,N-Dimethylamino)phenylacetylene
ACHTUNGTRENNUNG
and Arylated 1,1-Dicyanovinyl Derivatives To Form Intramolecular Charge-
Transfer Chromophores
Yi-Lin Wu, Peter D. Jarowski, W. Bernd Schweizer, and FranÅois Diederich*[a]
Abstract: The kinetics and mechanism
of the formal [2+2] cycloaddition–cy-
cloreversion reaction between 4-(N,N-
termining cycloaddition step are close
to the computational values. The rate
philic, terminal alkyne carbon onto the
dicyanovinyl electrophile generates a
transient zwitterion intermediate with
the negative charge developing as a
stabilized malononitrile carbanion. The
computational analysis predicted that
the cycloreversion of the postulated di-
cyanocyclobutene intermediate would
become rate-determining for 1,1-dicya-
noethene (2m) as the electrophile. The
dicyanocyclobutene 4m could indeed
be isolated as the key intermediate
from the reaction between alkyne 1
and 2m and characterized by X-ray
analysis. Facile first-order cyclorever-
sion occurred upon further heating,
yielding as the sole product the 1,1-di-
cyanobuta-1,3-diene 3m.
constants show
a good linear free
dimethylamino)phenylacetylene
(1)
energy relationship (1=2.0) with the
electronic character of the para-sub-
stituents on the benzylidene moiety in
dimethylformamide (DMF), which is
and para-substituted benzylidenemalo-
nonitriles 2b–2l to form 2-donor-sub-
stituted 1,1-dicyanobuta-1,3-dienes 3b–
3l via the postulated dicyanocyclobu-
tene intermediates 4b–4l have been
studied experimentally by the method
of initial rates and computationally at
the unrestricted B3LYP/6-31G(d) level.
The transformations were found to
follow bimolecular, second-order kinet-
indicative of
a dipolar mechanism.
Analysis of the computed structures
and their corresponding solvation ener-
gies in acetonitrile suggests that the
rate-determining attack of the nucleo-
Keywords: cycloaddition · kinetics ·
linear free energy relationships ·
push–pull chromophores · reaction
mechanisms
ics,
with
DHe°xp =13–18 kcalmolꢀ1,
DS°exp ꢁꢀ30 calKꢀ1 molꢀ1, and DGe°xp
=
22–27 kcalmolꢀ1. These experimental
activation parameters for the rate-de-
Introduction
tractive for fine-tuning their optoelectronic properties and
ultimately tailoring and optimizing advanced materials.
Organic D-p-A molecules not only need to feature desira-
ble molecular properties but they should also be preparable
on a larger scale through fast and efficient transformations.
With this in mind, we developed over the last few years
’click-chemistry’-type[2] reactions, which yielded new families
of intramolecular CT chromophores. [2+2] Cycloadditions
between electron-rich alkynes, such as 4-ethynyl-N,N-dime-
thylaniline (1, Scheme 1), and strongly electron-accepting
olefins, such as tetracyanoethene (TCNE, 2a) or 7,7,8,8-tet-
racyanoquinodimethane (TCNQ), followed by cyclorever-
sion provided nonplanar, donor-substituted 1,1,4,4-tetracya-
nobuta-1,3-dienes (TCBDs, such as 3a) and expanded
TCNQ derivatives, featuring readily tuneable CT absorption
wavelengths, large nonlinear optical responses, and high
electron affinities.[3] These transformations are generally
fast, high-yielding, catalyst-free, 100% atom-economic, and
p-Conjugated organic donor–acceptor (D-p-A) chromo-
phores, featuring intense low-energy intramolecular charge-
transfer (CT) absorptions in the visible to near-IR (infrared)
spectroscopic range, are in high demand for applications in
photonics and nonlinear optics.[1] Versatile organic synthesis
makes these dipolar organic push–pull molecules highly at-
[a] Y.-L. Wu, Dr. P. D. Jarowski, Dr. W. B. Schweizer,
Prof. Dr. F. Diederich
Laboratorium fꢀr Organische Chemie
ETH-Zꢀrich
Hçnggerberg, HCI
CH-8093 Zꢀrich (Switzerland)
Fax : (+41) 44 632 1109
E-mail: diederich@org.chem.ethz.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902645.
202
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 202 – 211

FULL PAPER
the transition states (TSs), which connect them, towards the
overall efficiency and outcome of this versatile class of
chemical transformations. Computation suggested that the
reaction should be bimolecular in both reactants with a
single rate-determining step (RDS). The rate-determining
TS, the addition of the alkyne to the electrophile in the case
of TCNE and TCE, is strongly polarized according to natu-
ral bond orbital (NBO) analysis and is apparent by inspec-
tion of the optimized geometries. Additionally, the solvation
energies in acetonitrile of the TS for 1+2m and its immedi-
ate product 4m are quite large suggesting strong polariza-
tion (the addition step is not rate-determining in this case
but rather the ring-opening step, as discussed below). How-
ever, there is no direct experimental evidence that proves
this zwitterionic mechanism or indicates any of the inter-
mediates. Additionally, there is no evidence that could rule
out a thermally allowed diradical mechanism. The question
remains to what extent the reactions proceed via zwitterion-
ic or diradical intermediates.
Scheme 1. The proposed reaction mechanism for the [2+2] cycloaddi-
tion–cycloreversion between 1 and cyanoolefins 2 through the zwitterion-
ic (5) and cyclobutenyl (4) intermediates to form charge-transfer chromo-
phores 3 where X is varied as defined (b–l).
Similar stepwise, zwitterionic mechanisms of pericyclic re-
actions are known and are especially efficient when highly
polarized, nucleophilic and electrophilic unsaturated reac-
tants are involved; many examples of cyclobutane formation
through zwitterionic intermediates can be found in classic
reviews by Gompper, Huisgen, and Fatiadi.[7a–d] However,
compared to the reactions between two olefins,[7] far less at-
tention has been paid to the triple bond in such dipolar,
formal [2+2] cycloadditions.[8] Previous studies on the triple
bond employed the strongly activated ynamines as the nu-
cleophilic reactants.[9] In some cases, cyclobutenes have even
been isolated as final products but the relationship between
these cycles and the starting materials has been obscured by
rearrangements of the peripheral substituents. In other
cases, an enyne metathesis happened to give butadienes, but
without detection or isolation of the cyclic intermediate.
Thus, no mechanistic elucidation has been possible due to
the complexity of the chemistry and the lack of a systematic
approach to study the electronic and structural aspects of
this transformation.
Herein, we present a thorough mechanistic investigation
of the [2+2] cycloaddition–cycloreversion reaction between
electron-rich alkynes, such as 1, and electron-deficient dicya-
noolefins, such as 2b–2m, based on both experimental and
computational methods. We provide strong evidence for the
zwitterionic character of the cycloaddition step and, in one
case, report the isolation and characterization of the dicya-
nocyclobutene intermediate, which is converted to the 1,1-
dicyanobuta-1,3-diene as the sole product.
the resulting products can be easily purified by precipitation
or washing. Recently, a high-speed optical waveguide was
prepared from a donor-substituted TCBD.[4] The success of
these molecules in the materials application arena stems
from their ability to form amorphous films of excellent opti-
cal quality due to their molecular nonplanarity and sublima-
bility.
More recently, we serendipitously discovered that 1,1-di-
cyanovinyl derivatives are sufficiently electron-accepting to
react with electron-rich alkynes in the same manner as
TCNE or TCNQ, providing the nonplanar push–pull chomo-
phores 3 in high yield (Scheme 1).[5] A series of para-substi-
tuted benzylidenemalononitriles (containing many of the
compounds in the series 2b–2l) was successfully converted
and the ease of reaction directly correlated with the elec-
tron-accepting power of the attached para-substituent.
The apparent net reaction for electrophiles 2a–2m is
identical to the transition-metal-catalyzed/mediated enyne
metathesis reaction[6] both of which provide buta-1,3-diene
products. By analogy, we have proposed that the reaction
between electron-rich acetylenes and cyanoolefins proceeds
through the cyclobutene intermediate 4. Subsequent cyclore-
version (4!3) of this transient intermediate gives the final
cyanobutadienes. Moreover, based on the facts that i) the
direct and concerted [2+2] cycloaddition is thermally sym-
metry-forbidden, ii) the transformation occurs in the ab-
sence of a transition-metal catalyst, and iii) the reaction only
proceeds between highly polarized components, we further
proposed a stepwise, zwitterionic pathway whereby initial
nucleophilic addition of the electron-rich alkyne forms a
transient zwitterion (5) that quickly cyclizes to the cyclobu-
tene. However, except for some information gained by our
initial computational investigation[5] of the reaction between
1 and 1,1-dicyanoethene (2m), TCNE (2a), and tricyanoe-
thene (TCE) as electrophiles, little is known about the ener-
getics and relative importance of these intermediates, and
Results and Discussion
Order of the reaction and the rate law: To gain more infor-
mation about the course of the reaction, a real-time NMR
study was performed for the transformation of the benzyli-
denemalononitriles 2b–2l in the temperature range from
1
298 K to 373 K. For a typical example, the H NMR spectral
Chem. Eur. J. 2010, 16, 202 – 211
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
203

F. Diederich et al.
identical, that is, [1]0 =[2]0, the
plot fits well to an overall
second-order rate expression,
meaning first-order in both 1
and 2 throughout the whole
course of reaction (Figure 2;
for plot-fitting, see the Sup-
porting Information). This was
further confirmed by following
the evolution of the CT band
in the UV/Vis spectrum (see
the Supporting Information).
The rate of reaction of 2h
(R1 =Ph, R2 =H) with 1 was
monitored at varying reactant
ratios (up to threefold excess
of one component), and the
Figure 1. Partial time-dependent 1H NMR spectra (500 MHz, (CD3)2SO, 373 K) of the solution of 1 and 2 f.
The time interval between consecutive traces is 800 s. Structure-signal assignments are given.
evolution of the solution of a 1:1 mixture of 1 and 2 f (R1 =
4-PhBr, R2 =H) in (CD3)2SO at 373 K is shown in Figure 1.
Similar spectral changes were observed for all the other
para-substituted substrates. In all cases, the transformation
is clean, directly from starting material to the corresponding
s-trans, trans butadiene product, without any evidence of
side-product or transient species formation on the NMR
time-scale.
rate of product formation was found to be proportional to
either [1] or [2h], that is, first order in both reactants. Thus,
the apparent rate law of the reaction subclass under ques-
tion is bimolecular and first order in both reactants, Equa-
tion (1),
n ¼ k½1ꢂ½2ꢂ
ð1Þ
where k is the observed rate constant. The simplicity of the
rate expression in these cases further allows the employment
of initial-rate methods in the following kinetics studies with-
out over-simplifying the reaction mechanism. Thus, the ini-
tial rate constants are calculated according to Equation (2),
The lack of a pre-equilibrium and the absence of any in-
termediate structure support our previous computational
studies, which predicted a bimolecular reaction mechanism
with one key RDS. It is reasonable to rule out any reversi-
bility after the RDS based on the high yields and the rapidi-
ty of the reaction under study. Moreover, when the extent
of the reaction from NMR data was plotted against the reac-
tion time, simple evolution curves were observed. Under the
conditions where the initial concentrations of 1 and 2 are
d½3ꢂ
dt
1
ð2Þ
k ¼
½1ꢂ0½2ꢂ0
where d[3]/dt is the observed rate and [1]0 and [2]0 are the
initial concentrations of 1 and 2.
Rate-dependency on the electronic nature of the substitu-
ents: A series of para-substituted benzylidenemalononitriles
2b–2k was mixed with 1 in DMF (N,N-dimethylformACHTUNGTRENNUNGamide)
solution at 298 K, and the development of their characteris-
tic CT bands around 450 nm was used to monitor the rate of
product formation and to determine the rate constant by the
method of initial rates employing Equation (2) as the kinetic
model. As shown in the inset of Figure 3, the intensity of
the CT band increases vertically, again suggesting a clean
transformation without by-product formation.
The logarithms of the measured initial-rate constants of
compounds 2b–2k with para-substituent X relative to the
logarithm of the rate of 2h (X=H) were plotted against
their corresponding Hammett substituent constants sp,[10]
and a good linear correlation was found (R2 =0.97), with
slope 1=2.0ꢃ0.1 (Figure 3). Each data point was measured
independently three times and the average rate constant
presented in Figure 3. The standard deviations are in the
range of 10% in most cases. The plot containing 2l at 373 K
also exhibits a linear free-energy relationship. Data at 323
Figure 2. Time evolution of the reaction between 1 and 2 f to give 3 f.
The extent of reaction for the reactants is defined as the ratio between
the number of one species at time t to its initial number (t=0); for the
product, it is defined as the ratio between its number at time t to the ini-
tial number of the reactant.
204
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 202 – 211

Dipolar [2+2] Cycloaddition–Cycloreversion Reactions
FULL PAPER
31G(d) level[11] in the gas-
phase as well as using the po-
larizable continuum model
(PCM) with acetonitrile to
model solvent effects. Herein,
this methodology is extended
to the cases of the reaction of
1 with the electrophiles having
para-nitrophenyl (2b), phenyl
(2h), or para-dimethylanilino
(2l) moieties attached to the
DCV group, treating only the
most relevant points along the
reaction profile that are candi-
dates for the RDS. The stabili-
ty of the wavefunction was
verified in each case. Geome-
tries were optimized using the
program Gaussian 03.[11] Sta-
tionary points were character-
ized by harmonic vibrational
frequency analysis. Molecular
energies were calculated as the
Figure 3. Hammett plot of initial rate constants of the reaction of 1 with electrophiles 2b–2k relative to the in-
itial rate constant for 2h versus the Hammett substituent constants sp at 298 K. The inset shows the time evo-
lution (in the direction of the arrow) of the CT band for the reaction of 1 with 2h.
and 348 K including 2b–2k have also been obtained (see the
Supporting Information). As depicted in Figure 3, the good
linearity implies that the same reaction mechanism is oper-
ating for all derivatives; the moderately positive 1 value re-
flects that the reaction is accelerated by electron-withdraw-
ing substituents and signifies a buildup of electron density
around the benzylidene ring in the RDS, consistent with a
stepwise, zwitterionic mecha-
sum of the electronic energy and Gibbs free energy correc-
tion obtained from these analytical frequencies at 298.15 K.
The Supporting Information gives all necessary numerical
values used in this work along with zero-point and thermal
corrections, gas-phase and solvated energies and geometries.
Irrespective of the substituent on the electrophile, the re-
action consists of four basic steps (Figure 4). From unreact-
nism where nucleophilic addi-
tion to the 1,1-dicyanovinyl
(DCV) moiety pushes elec-
trons to the most stabilizing
1,1-dicyanomethylidine
tion.
posi-
It was not possible to mea-
sure the rate of reaction of 1
and 2l at 298 K due to its slug-
gishness; the evolution of the
CT band was obscured by in-
strument noise. Kinetic data at
373 K that include this point
are presented in the Support-
ing Information and show a
linear correlation.
Calculated free energy profile
by density functional theory
and activation parameters by
Eyring analysis: In our previ-
ous report,
a reaction free-
Figure 4. Free-energy reaction profile at the B3LYP/6-31G(d) level (with PCM solvation in acetonitrile) for
2m (thin solid line), 2b (dashed line), 2h (wavy line), and 2l (thick solid line) with energies (kcalmolꢀ1) pro-
vided for reactants (1+2) (reference energy), intermediates 5, 4, s-cis 3, and product s-trans 3, and transition-
states barriers [1+2!5]°, [5!4]°, [4!s-cis 3]° and [s-cis 3!s-trans 3]° (energies referenced to their preced-
ing intermediate).
energy profile was examined
for the reaction between 1 and
methylenemalononitrile (2m)
at the unrestricted B3LYP/6-
Chem. Eur. J. 2010, 16, 202 – 211
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
205

F. Diederich et al.
ed components (1+2), nucleophilic addition of the g-carbon
of the terminal alkyne (1) to the b-position of the electro-
phile (2) occurs to produce a high-energy intermediate (5)
through the bond-forming transition state (TS) ([1+2!5]°)
(see Scheme 1 for notation). From here, the ring-closed
product (4) is obtained by bond formation between the a
and d carbons through an electrocyclic TS ([5!4]°). For the
reaction between 1 and 2m, this happens through a barrier
only slightly more energetic compared to the corresponding
energy of 5 and slightly less than the energy of [1+2m!
5m]°, but the analogous TSs could not be located for the re-
action with electrophiles 2b, 2h, and 2l, presumably due to
their shallow potential energy wells, which precluded opti-
mization to their respective stationary states. Moving for-
stituent, [4!s-cis 3]° raises (due mostly to effects in 4).
Thus, the kinetic results from the Hammett plot (positive 1
value) are consistent with [1+2!5]° as the RDS and rule-
out any mechanism with [4!s-cis 3]° as the RDS since the
sign of 1 would be reversed to what is experimentally ob-
served.
Solvation energies in acetonitrile for [1+2!5]°
occur
at
a
relatively
constant
value
(DDGs°olvation ¼ DGM°eCN;calcd ꢀ DGg°as;calcd ꢁꢀ3 kcalmolꢀ1) and
increase significantly for the immediate product
5
(DDGsolvation ¼ DGMeCN;calcd ꢀ DGgas;calcd =ꢀ9.5 (5m), ꢀ10.9
(5b), ꢀ8.6 (5h) and ꢀ7.9 (5l) kcalmolꢀ1) (negative values
indicate stabilization). Thus, the zwitterionic character of
the TS is corroborated, and to a greater extent, this charac-
ter is prevalent in the intermediate (5) with the amount of
stabilization increasing with the withdrawing power of the
substituent. Solvation energies for [4!s-cis 3]° are positive
and near 1 kcalmolꢀ1 in all cases. Somewhat larger numbers
are found for the cyclobutenes (4), which are destabilized
(positive DDGsolvation) by 3.5–4.5 kcalmolꢀ1. Based on the
small and positive solvation energies of [4!s-cis 3]°, the
zwitterionic character of the ring-opening step should be
small. Thus, a concerted diradical mechanism is most likely.
From variable-temperature UV/Vis kinetic data, it was
possible to perform Eyring analysis and extract the experi-
mental activation parameters DHe°xp, DSe°xp, and DGe°xp. The
initial rates for the reaction of 1 with electrophiles 2b, 2h,
and 2l were monitored in DMF at three or four tempera-
tures and the data fitted to the Eyring expression (for a de-
tailed description, see the Experimental; for the fitting anal-
ysis of the Eyring plots, see the Supporting Information).
There is excellent quantitative agreement, in addition to the
qualitative agreement, between the experimental activation
parameters (DG°exp in DMF) obtained from Eyring analysis
and the computed values (PCM solvation in acetonitrile) as
shown in Table 1.
ꢀ
ward, the ring-opening cycloreversion step breaks the a b
bond ([4!s-cis 3]°) leading eventually to the final products
(s-trans 3) by s-cis-to-s-trans isomerization ([s-cis 3!s-trans
3]°).
There is a considerable substituent effect on the TS ener-
getics of the addition step [1+2!5]°. In the non-substituted
case, 1,1-dicyanoethylene 2m reacts through a mechanism
where the ring-opening process ([4m!s-cis 3m]°; DGc°alcd
=
24.3 kcalmolꢀ1) is rate-determining based on comparison of
the calculated free-energies of activation in acetonitrile
(compare to [1+2m!5m]° in which DGc°alcd =16.0 kcal
molꢀ1), while in the case of the 1,1-dicyanovinylphenylenes
2b, 2h, and 2l, the RDS is the addition step ([1+2!5]°;
DG°calcd =23.0, 25.6, and 29.9 kcalmolꢀ1, respectively; com-
pare to [4!s-cis 3]°; DGc°alcd =14.0, 7.1, and 11.6 kcalmolꢀ1,
respectively). Inclusion of a phenyl group at the b-position
of the electrophile clearly raises the energy of [1+2!5]° in
all cases, compared to 2m, rendering it rate-determining
overall. A secondary influence beyond this, presumably
steric, effect of the phenyl ring is felt from the electronic
nature of the attached substituent at its para-position; the
nitro-substituted derivative (2b) reacts through the lowest
barrier, followed by the phenyl group (2h) and finally the
N,N-dimethylamino-substituted reactant (2l). Thus, the in-
creased electrophilicity of the dicyanovinyl group caused by
electron-withdrawal accelerates the reaction. This is in ac-
cordance with the experimental kinetic results.
Table 1. Experimental (298 K, DMF) and calculated (298 K, PCM solva-
tion in MeCN) activation parameters for the cycloaddition reaction be-
tween 1 and 2b, 2g, and 2l, and for the ring-opening reaction of 4m.
DG°exp
DG°calcd DHe°xp
]
DHc°alcd DS°exp
]
DS°calcd
]
Reactants
[kcalmolꢀ1
[kcalmolꢀ1
[calKꢀ1 molꢀ1
In contrast, the para-phenyl substituent has a diminished
influence on the ring-opening TS. The differences in this
barrier [4!s-cis 3]° for 2b, 2h, and 2l arise mostly from
large changes in the energies of the cyclobutene intermedi-
ates (4). The strongly donating dimethylanilino (DMA)
1+2b
1+2h
1+2l
4m
22ꢃ1
23.0
13ꢃ1
11.0
ꢀ30.3ꢃ0.4 ꢀ40.2
ꢀ29.9ꢃ0.2 ꢀ39.7
ꢀ29.1ꢃ0.8 ꢀ39.3
ꢀ11.4ꢃ0.3 ꢀ1.6
23.6ꢃ0.6 25.6
14.7ꢃ0.6 13.8
27ꢃ3
27ꢃ1
29.9
24.3
18ꢃ3
24ꢃ1
18.2
23.8
group destabilizes the cycle considerable (4l; DGcalcd
=
1.3 kcalmolꢀ1), placing its energy on the endergonic side of
the reaction profile compared to this state along the reac-
tion profile of 1+2m (4m; DGcalcd =ꢀ12.8 kcalmolꢀ1). Both
para-nitrophenyl (4b; DGcalcd =ꢀ7.2 kcalmolꢀ1) and phenyl
(4h; DGcalcd =ꢀ4.0 kcalmolꢀ1) substitution also destabilize
the cyclobutene, but to a lesser extent than for 4l. It is im-
portant to notice that the electronic effects of the phenyl
substituent on the barriers [1+2!5]° and [4!s-cis 3]° are
exactly opposite to each other. While [1+2!5]° lowers
upon increased electron-withdrawing character of the sub-
The measured activation parameters for 1+2!3 agree
with the calculated values for the TS [1+2!5]°. The mea-
sured negative entropies of activations (ꢁꢀ30 calKꢀ1 molꢀ1)
support a mechanism involving a bimolecular RDS. The ac-
tivation parameters measured for the conversion of 4m!
3m (see next section for more details) also agree well with
the computed transition-state energy barrier of [4m!s-cis
3m]°, and the entropy of activation (DS°exp =ꢀ11.4ꢃ
0.3 calmolꢀ1 Kꢀ1) is consistent with a unimolecular RDS pro-
206
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 202 – 211

Dipolar [2+2] Cycloaddition–Cycloreversion Reactions
FULL PAPER
cess. The experimental free energies of activation agree with
those computed to an average absolute deviation of 2 kcal
molꢀ1. While this is not proof that the RDS has been unam-
biguously assigned, it lends credence to the computational
approach and methodology presented here.
The geometric parameters strongly suggest zwitterionic
character in the addition TS ([1+2!5]°) for all derivatives
(Figure 5).
calized through the p-system in the zwitterionic ground-
state. Finally, the forming bond between the b and g posi-
tions varies considerably ([1+2m!5m]° (2.15 ꢂ)> [1+
2b!5b]° (2.08 ꢂ)> [1+2h!5h]° (2.04 ꢂ)> [1+2l!5l]°
(1.98 ꢂ)), again as a function of the accepting power of X.
The TS [1+2m!5m]° has the earliest (longest distance)
structure of all.
Comparing the bond lengths of the computed cyclobutene
moiety provides further insight into their tendency of ring-
ꢀ
ꢀ
ꢀ
opening. The bond lengths between b g, g d and d a of all
these cyclobutenes (4b, 4h, 4l, and 4m) are similar; howev-
ꢀ
er, the length of the a b bond varies about 0.05 ꢂ through-
out the series (for Cartesian coordinates and the pictorial il-
lustration, see the Supporting Information). It is observed
ꢀ
that the longer the a b bond length, the smaller the barrier
to ring-opening. Comparing 4b, 4h, and 4l, we find a longer
ꢀ
a b bond length (1.65 ꢂ) with the NMe2 substituent, a
shorter length with the NO2 substituents (1.63 ꢂ), and an in-
termediate value for the phenyl substituted cyclobutene
(1.64 ꢂ). This relatively small structural effect is presumably
inductive in nature as the influencing substituent is separat-
ed by the sp3-hybridized b-carbon atom. Nonetheless the
substituent effect, which may be acting on a partially diradi-
cal or zwitterionic electronic structure of the cyclobutene is
sufficient to vary the cyclobutene energies in a range of
about 9 kcalmolꢀ1. Without any substituent (4m), this bond
length is considerable shorter (1.60 ꢂ). Given that electronic
effects on the cyclobutene are attenuated by poor communi-
cation over the b-position, it is reasonable to assume that
this effect is predominantly steric in nature. In the case of
4m, the tight binding between the a and b positions caused
by a lack of destabilizing phenyl substitution leads to an ad-
ditional 9 kcalmolꢀ1 (compare 4h to 4m) stabilization allow-
ing the isolation of the cyclobutene as a stable intermediate,
as discussed in the following section.
Figure 5. Optimized geometries (B3LYP/6-31G(d) with PCM solvation in
acetonitrile) of TSs [1+2!5]° for the reaction of 1 with 2m, 2b, 2h, and
2l. Relevant structural parameters are given. Bond angles (b) and pyra-
midalization (improper dihedral) angles (p) are in 8, and bond lengths in
ꢂ;
the
quinoid
character
is
defined
as
dr=
½ða þ a0Þ=2 þ ðc þ c0Þ=2 ꢀ ðb þ b0Þꢂ=2.
Pyramidalization (defined as the improper dihedral in-
cluding both carbons of the two nitriles and the a and b car-
bons) about the methylidine (a) carbon of the acceptor por-
tion is significant and decreases in the direction of lowering
activation energy as the TS becomes earlier ([1+2b!5b]°
(2.78) <[1+2h!5h]° (2.98) <[1+2l!5l]° (3.48)). This
pyramidalization is intermediate for the case of the reaction
with 1,1-dicyanoethylene ([1+2m!5m]° (2.88)). For the
donor portion, a substantial quinoid character[12] of the
DMA group in the TSs (dr=0.05 ꢂ for [1+2b!5b]°, [1+
2h!5h]°, and [1+2l!5l]°; see the caption of Figure 5 for
definition), referenced to the structure of isolated 1 (dr=
0.03 ꢂ) at this level of theory, suggests that the donation
from the amino substituent of the DMA residue is promi-
nent. The quinoid character is slightly less when 1,1-dicyano-
ethylene is the electrophile ([1+2m!5m]°; dr=0.04 ꢂ).
Intermediate identification: Comparing the computational
free energy profile in our previous report for the reaction of
2m with the results shown in the previous sections for 2b,
2h, and 2l, it is interesting to notice that the higher-barrier
process changed from the ring-opening step to the ring-for-
mation step upon phenyl substitution. Since the cyclobutene
intermediate was not found in any of the cases involving
phenyl-substituted compounds, which have the RDS in the
addition step, this change in relative barrier-height suggested
the possibility of isolating the cyclobutene using 2m as the
electrophile.
To verify this postulate, 1 was mixed with 2m, which was
generated in situ from malononitrile and aqueous formalde-
hyde in DMF solution at 508C overnight. A yellow solid was
isolated after column chromatography in 76% yield, and is
stable at ambient condition for weeks without decomposi-
tion. The peak at m/z 223 in the mass spectrum indicated a
molecular formula of C14H13N3+, corresponding to the 1:1
ꢀ
Additionally, deviation in the linearity of the g-d e bond
angles of ~88 places a considerable cationic character at the
central d carbon atom. This bond angle becomes linear for
5m, 5b, 5h, and 5l (see the Supporting Information for Car-
tesian coordinates) where the charge becomes further delo-
adduct between 1 and 2m. The very weak IR absorption of
ꢀ1
ꢄ
C N stretching at 2246 cm suggests a non-conjugated
1
cyano functional group. In the H NMR spectrum (CDCl3,
Chem. Eur. J. 2010, 16, 202 – 211
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
207

F. Diederich et al.
298 K), the aliphatic doublet signal at 3.3 ppm (2H, J=
1.4 Hz) together with the olefinic triplet signal at 6.2 ppm
(1H, J=1.4 Hz) immediately rule out the butadiene struc-
ture, but, to our delight, point to a cyclobutene 4m which
was further confirmed by X-ray analysis (Figure 6a).
try with non-planarity between the N,N-dimethylanilino and
the dicyanovinyl moieties as observed in other phenyl-sub-
stituted analogues (Figure 6b).[5] The transformation is clean
and quantitative, without involving any other intermediate
(at least on the NMR time-scale), and its kinetics fits well
into a first-order reaction (see the Supporting Information).
Thus, this dicyanocyclobutene was, for the first time, isolated
and identified as the true intermediate of the formal [2+2]
cycloaddition between electron-rich acetylene and cyanoole-
fins, giving cyanobuta-1,3-dienes as the final products.
A closer look at the solid-state molecular structure of 4m
provides deeper information on the cyclobutene properties.
The DMA and the cyclobutene rings arrange almost in the
same plane with a small twist of 6.688, making the whole
molecule quasi-Cs symmetric. The double bond length be-
ꢀ
tween C7 C8 is 1.329(4) ꢂ, while the single bond lengths
ꢀ
ꢀ
ꢀ
between C8 C9, C7 C10, and C9 C10 are 1.501(4),
1.544(3), and 1.586(3) ꢂ, respectively. These bond lengths
are all in the range of known bond length of cyclobutenes,[13]
and the observed molecular geometry highly resembles that
obtained from DFT calculations. However, it is noteworthy
ꢀ
ꢀ
that the C7 C10 bond is significantly longer than the C8
ꢀ
C9 bond, and also than the typical Csp3 Csp2 bond length in
cyclobutenes.[13] It is known that, in phenyl-substituted cyclo-
ꢀ
butene, the substituted Csp3 Csp2 bond is usually longer than
the other unsubstituted one due to steric reasons,[14] making
the corresponding bond angle C8-C7-C10 93.2(2)8 smaller
than C7-C8-C9 96.5(2)8, as a consequence of the geometrical
constraint of the planar cyclobutene structure (sum of the
four bond angles in the ring=3608).
Conclusions
We have studied the reaction mechanism of the formal [2+
2] cycloaddition–cycloreversion reaction between 4-(N,N-di-
methylamino)phenylacetylene (1) and a series of phenyl-
substituted 1,1-dicyanovinyl derivatives 2b–2l by means of
UV/Vis and NMR spectroscopies and DFT calculations.
From the experimental data, the reaction is clearly a quanta-
tive transformation without by-product formation. The bi-
molecular, second-order kinetics shows a good linear free
energy relationship with respect to the electronic effects of
the dicyanovinyl electrophiles, indicating a strong zwitter-
ionic character in the rate-determining step, which is most
probably due to a transition state where the negative charge
is located at the gem-dicyano-substituted homobenzylic posi-
tion (a-carbon) arising from nucleophilic attack of 1 onto
the b-carbon of the dicyanovinyl moiety to form 5 according
to the computed structures. For electrophiles 2b–2l, the acti-
vation free energies range around 22–27 kcalmolꢀ1, with
negative activation entropy contributions of about
ꢀ30 calKꢀ1 molꢀ1 in DMF, on the order of those expected
for a bimolecular reaction. The cyclobutenyl molecule 4m
generated from the cycloaddition between 1 and 2m was
isolated as the true reaction intermediate and transformed
into the ring-opening dicyanobuta-1,3-diene 3m upon heat-
Figure 6. ORTEP plots of a) 4m (223 K) and b) 3m (223 K) with thermal
ellipsoids shown at the 50% probability level. Selected bond lengths [ꢂ],
ꢀ
ꢀ
bond angles [8] and torsion angles [8]: 4m: C1 C6 1.392(3), C1 C2
ꢀ
ꢀ
ꢀ
ꢀ
1.395(3), C2 C3 1.384(3), C3 C4 1.412(3), C4 C5 1.410(3), C5 C6
ꢀ
ꢀ
ꢀ
ꢀ
1.381(3), C7 C8 1.329(4), C7 C10 1.544(3), C8 C9 1.501(4), C9 C10
1.586(3); C11-C10-C13 109.2(2), C10-C7-C8-C9 ꢀ0.3(2), C2-C1-C7-C10
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ6.7(3); 3m: C1 C2 1.396(3), C1 C6 1.401(3), C2 C3 1.374(3), C3 C4
ꢀ
ꢀ
1.410(3), C4 C5 1.408(3), C5 C6 1.373(3); C14-C10-C16 112.44(18), C2-
C1-C11-C10 ꢀ45.2(3), C10-C11-C12-C13 162.5(2).
This cyclobutene molecule was further transformed into
the 1,1-dicyanobuta-1,3-diene 3m upon heating. In the real-
1
time H NMR experiment in (CDCl2)2 at 373 K, the two sig-
nals at 3.33 (d, J=0.9 Hz) and 6.23 (t, J=0.9 Hz) ppm of cy-
clobutene protons gradually vanished and three new sets of
signals from terminal vinylic proton systems at 5.83 (dd, J=
16.9, 0.8 Hz), 6.06 (dd, J=10.6, 0.8 Hz), and 7.08 (dd, J=
16.9, 10.6 Hz) ppm appeared (see the Supporting Informa-
tion). The X-ray-quality crystal of the butadiene was ob-
tained by the diffusion method and shows the s-trans geome-
208
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 202 – 211

Dipolar [2+2] Cycloaddition–Cycloreversion Reactions
FULL PAPER
ing. The small activation entropy (ꢀ11.4 calKꢀ1 molꢀ1) of
ring-opening indicates a unimolecular RDS.
0.040–0.063 mm, 230–400 mesh; Fluka) and technical solvents. Thin-layer
chromatography (TLC) was conducted on aluminum sheets or glass plate
coated with SiO2 60 F254 obtained from Merck; visualization with a UV
lamp (254 or 366 nm). Melting points (M.p.) were measured on a Bꢀchi
B-540 melting-point apparatus in open capillaries and are uncorrected.
1H NMR and 13C NMR spectra were measured on a Bruker AV 400 in-
strument at 208C. Chemical shifts are reported in ppm relative to the
signal of tetramethylsilane. Residual solvent signals in the 1H and
13C NMR spectra were used as an internal reference. Coupling constants
(J) are given in Hz. The apparent resonance multiplicity is described as s
(singlet), d (doublet), and dd (doublet of doublet). Infrared spectra (IR)
were recorded on a Perkin–Elmer Spectrum BX instrument. UV/Vis
spectra were recorded on a Varian Cary-500 spectrophotometer in a
quartz cuvette (1 cm). The absorption wavelengths are reported in nm
with the extinction coefficient e (mꢀ1 cmꢀ1) in parenthesis; shoulders are
indicated as sh. High-resolution HR-EI-MS spectra were measured on a
Hitachi-Perkin–Elmer VG-Tribrid spectrometer. The signal of the molec-
ular ion (M+) is reported in m/z units. Compounds 3b, 3e, 3 f, 3g, 3h,
3k, and 3l and their corresponding dicyanovinyl precursor molecules 2b,
2e, 2 f, 2g, 2h, 2k, and 2l were prepared according to to literature proce-
dures[5] as were additional precursor molecules 2c,[15a] 2d,[15b] 2i[15c] and
2j;[15d] their NMR data and melting points are in accord with the litera-
ture values. New charge-transfer chromophores 3c, 3d, 3i, and 3j were
prepared by the General Method (see below). Compound 3m was pre-
pared through in situ generation of precursor electrophile 2m to yield
first 4m, as described below.
Four reaction sub-steps were identified on the computa-
tional reaction free energy profiles, where the addition step
and the ring-opening processes were shown to have more
importance to the overall reaction kinetics. The energy bar-
riers for the addition step process for 2b, 2h, and 2l are
rate-determining as they are much higher in energy than
those of their subsequent steps. The finding of a high barrier
in the early stage of the bimolecular, multi-step reaction is
also in accordance with the fact that the reactions follow
simple second-order kinetics. Removing the phenyl substitu-
ent from the DCV group, however, changes the high-barrier
step to the ring-opening process. As in the addition step,
where the linear free energy relationship was found experi-
mentally, there is also a substantial substituent effect on the
ring-opening process. The barriers of this process were
found to correlate with the ability of the substituent to
affect the stability of the cyclobutene intermediates. The
barrier heights calculated here were found to have a good
quantitative agreement with the experimental values, sug-
gesting the credibility of the proposed reaction mechanism.
While the mechanism for benzylidenemalononitriles has
been treated in detail here, it would be too tenuous a gener-
alization to assign the exact present mechanism to those re-
actions involving tricyanoethylene (TCE), TCNE, or TCNQ
as the electrophiles nor do we offer a holistic mechanism for
reactions involving metal acetylides as the nucleophiles. Our
current finding suggests a reaction subclass that is highly
sensitive to substitution with a roving rate-determining step.
Additionally, a pre-equilibrium of the charge-transfer com-
plexes for TCNE and TCNQ may be important. Preliminary
kinetic experiments (unpublished results) for the reaction of
1 and TCNE (2a) offer an unclear picture as to whether the
reaction is first or second order as both kinetic models give
reasonable activation energies. It is reasonable that the addi-
tion step in this case would be considerably lower in energy
than for the benzylidenemalononitriles and the ring-opening
step much larger since the dicyanomethylidenes flank both
the forming positively and negatively charged positions as
the ring opens. Thus, either addition or ring-opening or both
could be rate-determining in this case. Detailed mechanistic
studies involving these electrophiles are ongoing to further
elaborate the exciting complexity and sensitivity of this reac-
tion subclass. As it stands, reactions involving electron-rich
acetylenes and electron-poor cyanoolefins should proceed
through a stepwise process involving zwitterionic and cyclo-
butenyl intermediates regardless of which TS is rate-deter-
mining.
UV/Vis measurements, initial-rate determination, and Eyring analysis:
The molar extinction coefficient and the initial-rate kinetics were mea-
sured with the Varian CARY 500 Scan UV-Vis-NIR Spectrophotometer,
controlled by the CARY WinUV software in a Windows 2000 Professio-
nal operating system. The Scan application of CARY WinUV was used
for molar extinction coefficient determination of a 5ꢃ10ꢀ5 m solution in
DMF at 298 K. For initial-rate measurement, the CARY Temperature
Controller and Stir Control accessories were employed. The Scanning Ki-
netics application of CARY WinUV was used to control the appropriate
temperature, the scan rate (600 nmminꢀ1), and scan range (550–400 nm),
and to monitor the rate of product formation. The measured reactions
were conducted in a 1 cm quartz cuvette charged with a stir-bar. For the
cycloaddition reactions, 2b–2k were added into the pre-heated DMF
(3 mL) solution of 1 at 298, 323, 348, and 373 K. For the exceedingly slow
reaction of 2l, the temperatures were 353, 363, and 373 K. The concentra-
tion was 5.05ꢃ10ꢀ3 m for each starting material. The spectra were collect-
ed for every 1 min (typically monitored over 20–40 min), and the product
concentrations used for initial-rate calculation were converted from the
recorded absorbances based on the corresponding extinction coefficients
at room temperature. Each data point was collected three times, and the
average value is presented herein with the standard deviation indicating
the experimental error. All experimental rate constants are given in the
Supporting Information.
From these variable temperature kinetic data, it was possible to perform
Eyring analysis and extract the activation parameters DHe°xp, DSe°xp and
DG°exp. The initial rate of the unimolecular reaction of 4m to 3m was
monitored at 333, 353, and 373 K by adding 4m to the pre-heated neat
DMF to obtain these parameters for the ring-opening step, using the
same experimental conditions as for the other reactions. Fitting analyses
of the Eyring plots are given in Supporting Information.
Real-time NMR study of the reaction: Real-time 1H NMR snapshots of
the reaction were measured on a Bruker Avance DRX 500 spectrometer
operating at 500.1 MHz for the 1H nucleus. The pulse angle was 308 and
the acquisition time 5 s for each transient. The temperature was set by
Bruker TopSpin, but the real temperature at the probehead region was
measured externally with the Greisinger GMH 3710 high-precision ther-
mometer connected to a platinum four-wire temperature sensor embed-
ded in an NMR tube. A 4.6ꢃ10ꢀ2 m solution of reactants in (CD3)2SO
was freshly prepared and immediately charged into the pre-heated spec-
trometer for measurement. The temperature was assumed to reach equi-
librium during the time of magnetic field adjustment. The spectra were
Experimental Section
Materials and general methods: Reagents were purchased at reagent
grade from Acros, Sigma–Aldrich, and Fluka and used as received. An-
hydrous CH2Cl2 was freshly distilled from CaH2 under N2 atmosphere.
Column chromatography was carried out with SiO2 60 (particle size
Chem. Eur. J. 2010, 16, 202 – 211
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
209

F. Diederich et al.
collected for every 20, 80, or 160 s, depending on the rapidness of individ-
ual reaction. Spectra procession was performed with MestReNova.
131.87, 138.81, 146.24, 152.90, 166.27, 170.03 ppm; IR (neat): n˜ =2951
(vw), 2905 (vw), 2824 (vw), 2361 (w) 2343 (w), 2216 (m), 1720 (s), 1600
(vs), 1492 (s), 1283 cmꢀ1 (s); UV/Vis (DMF): lmax (e)=352 (27100),
X-ray analysis: X-ray data collection was carried out on a Bruker Kap-
paCCD diffractometer equipped with a graphite monochromator (MoKa
radiation, l=0.71073 ꢂ) and an Oxford Cryostream low-temperature
device. Cell dimensions were obtained by least-squares refinement of all
measured reflections (HKL, Scalepack[16a]), qmax =27.58. All structures
were solved by direct methods (SIR97[16b]). All non-hydrogen atoms were
refined anisotropically, H-atoms isotropically by full matrix least-squares
with SHELXL-97[16c] using experimental weights (1/[s2(Io)+(Io +Ic)2/
900]).
+
470 nm (9300 mꢀ1 cmꢀ1); HR-EI-MS m/z: calcd for C22H14N3O2
357.1472; found: 357.1473 ([M]+).
:
AHCTUNGTERG{NNUN (2E)-3-Biphenyl-4-yl-1-[4-(dimethylamino)phenyl]prop-2-en-1-ylidene}-
malononitrile (3i): Following the general method, 3i was obtained
(139 mg, 86%) after washing with Et2O. M.p. 200–2028C; 1H NMR
(CDCl3, 400 MHz): d=3.10 (s, 6H; NMe2), 6.77 (d, J=9.0 Hz, 2H;
PhNMe2), 7.07 (d, J=15.6 Hz, 1H; CH=CH), 7.48–7.36 (m, 5H, Ph), 7.56
(d, J=15.6 Hz, 1H; CH=CH), 7.65–7.60 ppm (m, 6H; Ph); 13C NMR
(CDCl3, 100 MHz): d=40.06, 111.37, 114.33, 115.15, 119.96, 125.10,
127.06, 127.71, 128.12, 128.97, 129.07, 131.67, 133.74, 139.85, 143.86,
147.59, 152.77, 170.75 ppm (1 peak of C(CN)2 is overlapped with the sol-
vent peaks); IR (neat): n˜ =2923 (vw), 2860 (vw), 2826 (vw), 2360 (vw),
2208 (s), 1594 (vs), 1481 (s), 1376 (s), 1347 (s), 1172 cmꢀ1 (s); UV/Vis
(DMF): lmax (e)=382 (31 300), 454 nm (sh, 9900 mꢀ1 cmꢀ1); HR-EI-MS
m/z: calcd for C26H21N3+: 375.1730; found: 375.1729 ([M]+).
X-ray crystal structure of compound 4m: Single crystals were obtained by
slow diffusion of pentane into a solution of 4m in CHCl3 at ꢀ208C:
C14H13N3, Mr =223.28, crystal dimensions 0.42ꢃ0.2ꢃ0.06 mm, monoclinic
space group P21/c (no. 14), 1calcd =1.199 gcmꢀ3, Z=4, a=6.7709(4), b=
7.6141(4), c=24.105(2) ꢂ, b=95.336(2)8, V=1237.35(12) ꢂ3 at 223(1) K.
Number of measured and unique reflections 3716 and 2355, respectively
(Rint =0.043). Final R(F)=0.056, wR(F2)=0.142 for 206 parameters and
1467 reflections with I> 2s(I) (corresponding R-values based on all 2355
reflections 0.102 and 0.174).[17]
AHCTUNGTERG[NNUN (2E)-1-[4-(Dimethylamino)phenyl]-3-(4-methylphenyl)prop-2-en-1-yli-
dene]malononitrile (3j): Following the general method, 3j was obtained
(150 mg, 81%) after washing with Et2O. M.p. 151–1548C; 1H NMR
(CDCl3, 400 MHz): d=2.39 (s, 3H; PhMe), 3.09 (s, 6H; PhNMe2), 6.76
(d, J=9.0 Hz, 2H; PhNMe2), 7.00 (d, J=15.6 Hz, 1H; CH=CH), 7.21 (d,
J=8.2 Hz, 2H; PhMe), 7.39 (d, J=9.0 Hz, 2H; PhMe); 7.44 (d, J=
8.2 Hz, 2H; PhMe); 7.49 ppm (d, J=15.6 Hz, 1H; CH=CH); 13C NMR
(CDCl3, 100 MHz): d=21.57, 40.03, 111.32, 114.36, 115.19, 120.03, 124.23,
128.57, 129.87, 131.61, 132.12, 141.89, 148.22, 152.70, 171.04 ppm (1 peak
of C(CN)2 is overlapped with the solvent peaks); IR (neat): n˜ =3055
(vw), 2987 (vw), 2868 (vw), 2821 (vw), 2217 (m), 1600 (s), 1491 (m), 1371
(m), 1345 (m), 1265 (m), 1172 (m), 735 cmꢀ1 (s); UV/Vis (DMF): lmax
(e)=368 (22 100), 460 nm (8300 mꢀ1 cmꢀ1); HR-EI-MS m/z: calcd for
C21H19N3+: 313.1573; found: 313.1575 ([M]+).
X-ray crystal structure of compound 3m: Single crystals were obtained by
slow diffusion of pentane into a solution of 3m in CHCl3 at ꢀ208C:
C14H13N3, Mr =223.28, crystal dimensions 0.54ꢃ0.3ꢃ0.09 mm, orthorhom-
bic space group Pbca (no. 61), 1calcd =1.173 gcmꢀ3, Z=8, a=14.1701(6),
b=8.0015(3), c=22.3043(8) ꢂ, V=2528.9(2) ꢂ3 at 223(1) K. Number of
measured and unique reflections 5681 and 2879, respectively (Rint
=
0.046). Final R(F)=0.081, wR(F2)=0.186 for 206 parameters and 1845
reflections with I> 2s(I) (corresponding R-values based on all 879 re-
flections 0.118 and 0.208).[17]
X-ray crystal structure of compound 3c: Crystals were grown by slow dif-
fusion of pentane into a solution of 3c in CHCl3 at room temperature in
a sealed container. The ORTEP plot of 3c can be found in Supporting
Information: C21H16N4, Mr =324.39, crystal dimensions 0.36ꢃ0.075ꢃ
0.025 mm, monoclinic space group P21/c (no. 14), 1calcd =1.192 gcmꢀ3, Z=
4, a=9.6650(4), b=7.8764(6), c=23.977(2) ꢂ, b=97.899(5)8, V=
1808.0(2) ꢂ3 at 170(1) K. Number of measured and unique reflections
8095 and 2433, respectively (Rint =0.077). Final R(F)=0.055, wR(F2)=
0.141 for 290 parameters and 1768 reflections with I> 2s(I) (correspond-
ing R-values based on all 2433 reflections 0.085 and 0.165).[17]
2-[4-(Dimethylamino)phenyl]cyclobut-2-ene-1,1-dicarbonitrile
(4m):
Alkyne 1 (100 mg, 0.69 mmol) was added into a DMF (5 mL) solution of
formaldehyde (36% solution in water, 63 mL, 0.83 mmol) and malononi-
trile (45.5 mg, 0.69 mmol). The mixture was stirred at 508C for 18 h.
DMF and water were removed under vacuum, and the residue was puri-
fied by column chromatography (SiO2, CH2Cl2/hexane 1:1 ! 2:1) to give
4m as a yellow solid (117 mg, 76%). Rf =0.25 (CH2Cl2/hexane=2:1);
m.p.> 1338C (the molecule opened into the corresponding butadiene
around 1008C at a discernable rate; thus the m.p. could not be recorded);
1H NMR (CDCl3, 300 MHz): d=3.01 (s, 6H; NMe2), 3.30 (d, J=1.4 Hz,
2H; CH2 in cyclobutene), 6.19 (t, J=1.4 Hz, 1H; CH=C in cyclobutene),
6.71 (d, J=8.9 Hz, 2H; PhNMe2), 7.36 ppm (d, J=8.9 Hz, 1H; PhNMe2);
13C NMR (CDCl3, 100 MHz): d=29.63, 39.27 40.03, 111.88, 114.29,
116.65, 124.11, 125.76, 140.79, 151.27 ppm; IR (neat): n˜ =2245 (vw), 1627
(m), 1603 (m), 1522 (m), 1364 cmꢀ1 (m); HR-EI-MS m/z: calcd for
C14H13N3+: 223.1109; found: 223.1102 ([M]+); elemental analysis (%)
calcd for C14H13N3: C 75.31, H 5.87, N 18.82; found: C 75.25, H 5.83, N
18.66.
General method for the synthesis of 3c, 3d, 3i, and 3j: A solution of
DCV derivative 2c, 2d, 2i, or 2j (100 mg) in acetonitrile (6 mL) and 1
(1 equiv) was stirred at reflux until complete conversion (determined by
LC-MS). The solvent was removed first by rotary evaporation, then
under higher vacuum to yield the crude product. The purification of each
product is described below.
ACHTUNGTRENNUNG{(2E)-3-(4-Cyanophenyl)-1-[4-(dimethylamino)phenyl]-2-propen-1-ylide-
ne}malononitrile (3c): Following the general method, 3c was obtained
(175 mg, 96%) after passing the crude material through a short plug of
SiO2 with CH2Cl2 as eluent and collecting the red-colored band. M.p.
1
214–2158C; H NMR (CDCl3, 400 MHz): d=3.10 (s, 6H; NMe2), 6.76 (d,
{1-[4-(Dimethylamino)phenyl]prop-2-en-1-ylidene}malononitrile (3m): A
solution of 4m (50 mg, 0.22 mmol) in 1,1,2,2-tetrachloroethane (5 mL)
was stirred at 808C for 10 h. The solvent was removed under vacuum to
provide the product as a red solid (50 mg, 100%). Rf =0.17 (CH2Cl2/
hexane=1:1); m.p. 115–1168C; 1H NMR (CDCl3, 400 MHz): d=3.07 (s,
6H; NMe2), 5.79 (dd, J=16.9, 0.8 Hz, 1H; CH=CH2), 6.01 (dd, J=10.6,
0.8 Hz, 1H; CH=CH2), 6.72 (d, J=9.0 Hz, 2H; PhNMe2), 7.10 (dd, J=
16.9, 10.6 Hz, 1H; CH=CH2) 7.41 ppm (d, J=9.0 Hz; PhNMe2);
13C NMR (CDCl3, 100 MHz): d=39.94, 78.07, 111.19, 113.78, 114.71,
119.31, 131.86, 133.12, 134.22, 152.98, 170.43 ppm; IR (neat): n˜ =2218
(m), 1600 (m), 1503 (m), 1372 cmꢀ1 (m); UV/Vis (DMF): lmax (e)=436
nm (16 800 mꢀ1 cmꢀ1); HR-EI-MS m/z: calcd for C14H13N3+: 223.1109;
found: 223.1103.
J=9.0 Hz, 2H; PhNMe2), 7.01 (d, J=15.7 Hz, 1H; CH=CH), 7.41 (d, J=
9.0 Hz, 2H; PhNMe2), 7.56 (d, J=15.7 Hz, 1H; CH=CH), 7.62 (d, J=
8.4 Hz, 2H; PhCN), 7.69 ppm (d, J=8.4 Hz, 2H; PhCN); 13C NMR
(CDCl3, 100 MHz): d=40.03, 79.02, 111.42, 113.83, 113.91, 114.65, 118.18,
119.32, 128.54, 128.61, 131.71, 132.75, 138.86, 144.81, 152.99, 169.37 ppm;
IR (neat): n˜ =3050 (vw), 2986 (vw), 2938 (vw), 1717 (m), 1685 (m), 1265
(s), 1221 (m), 740 (vs), 631 cmꢀ1 (s); UV/Vis (DMF): lmax (e)=345
+
(26400), 474 nm (9400 mꢀ1 cmꢀ1); HR-EI-MS m/z: calcd for C21H16N4
324.1369; found: 324.1369 ([M]+).
:
Methyl 4-{(1E)-4,4-dicyano-3-[4-(dimethylamino)phenyl]buta-1,3-dien-1-
yl}benzoate (3d): Following the general method, 3d was obtained
(135 mg, 80%) after column chromotagraphy (SiO2, hexane:CH2Cl2 =
2:1) and collecting the red-colored band. M.p. 211–2128C; 1H NMR
(CDCl3, 400 MHz): d=3.10 (s, 6H; NMe2), 3.94 (s, 3H; CO2Me), 6.76 (2,
J=8.7 Hz, 2H; PhNMe2), 7.04 (d, J=15.7 Hz, 1H; CH=CH), 7.41 (d, J=
8.7 Hz, 2H; PhNMe2), 7.67–7.53 (m, 3H; CH=CH, PhCO2Me), 8.06 ppm
(d, J=8.2 Hz, 2H; PhCO2Me); 13C NMR (CDCl3, 100 MHz): d=40.05,
52.36, 78.49, 111.40, 114.06, 114.85, 119.60, 127.44, 128.25, 130.25, 131.71,
210
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 202 – 211

Dipolar [2+2] Cycloaddition–Cycloreversion Reactions
FULL PAPER
Onuma, Y. Kai, N. Yasuoka, N. Kasai, Bull. Chem. Soc. Jpn. 1975,
48, 1696–1700. For dicyanovinyls, see: h) M. I. Bruce, P. A. Hum-
phrey, M. R. Snow, E. R. T. Tiekink, J. Organomet. Chem. 1986, 303,
417–427; i) C.-W. Chang, Y.-C. Lin, G.-H. Lee, Y. Wang, J. Chem.
Soc. Dalton Trans. 1999, 4223–4230.
Acknowledgements
This work was supported by ETH Research Council. The authors thank
Dr. Marc-Olivier Ebert (ETH) for the help in NMR experiments. We are
grateful for access to the Competence Center for Computational Chemis-
try (C4) Obelix cluster (ETH). Y.-L. W. Acknowledges the receipt of Sti-
pendienfods der Schweizerischen Chemischen Industrie (SSCI).
[9] a) J. Ø. Madsen, S.-O. Lawesson, Tetrahedron 1974, 30, 3481–3492;
b) J. Ficini, Tetrahedron 1976, 32, 1449–1486; c) H. C. Brown, G. G.
Pai, J. Org. Chem. 1982, 47, 1608–1610; d) M. L. M. Pennings, D. N.
Reinhoudt, J. Org. Chem. 1982, 47, 1816–1823; e) G. Himbert, W.
Brunn, Chem. Ber. 1984, 117, 642–653; f) G. Himbert, W. Brunn,
Liebigs Ann. Chem. 1985, 2206–2216; g) G. Himbert, S. Kosack,
Chem. Ber. 1988, 121, 2053–2057.
[1] a) J. M. Tour, Chem. Rev. 1996, 96, 537–553. b) Focus Issue on “Or-
ganic Electronics”: J. Mater. Res. 2004, 19, 1887–2203, edited by F.
Faupel, C. Dimitrakopoulos, A. Kahn, C. Wçll. c) Special Issue on
“Organic Electronics and Optoelectronics”: Chem. Rev. 2007, 107,
923–1386, edited by F. Faupel, C. Dimitrakopoulos, A. Kahn, C.
Wçll.
[10] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165–195.
[11] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. Montgom-
ery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, C. Adamo, V. Bakken, C. Adamo, J. Jaramillo, R. Gom-
perts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Po-
melli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Sal-
vador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Dan-
iels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Ragha-
vachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford,
J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Ko-
maromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson,
W. Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc.,
Wallingford CT, 2004.
[2] H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew. Chem. 2001, 113,
2056–2075; Angew. Chem. Int. Ed. 2001, 40, 2004–2021.
[3] a) T. Michinobu, C. Boudon, J. P. Gisselbrecht, P. Seiler, B. Frank,
N. N. P. Moonen, M. Gross, F. Diederich, Chem. Eur. J. 2006, 12,
1889–1905; b) M. Kivala, C. Boudon, J.-P. Gisselbrecht, P. Seiler, M.
Gross, F. Diederich, Chem. Commun. 2007, 4731–4733; c) P. Reute-
nauer, M. Kivala, P. D. Jarowski, C. Boudon, J.-P. Gisselbrecht, M.
Gross, F. Diederich, Chem. Commun. 2007, 4898–4900; d) M.
Kivala, C. Boudon, J.-P. Gisselbrecht, B. Enko, P. Seiler, I. B. Mꢀller,
N. Langer, P. D. Jarowski, G. Gescheidt, F. Diederich, Chem. Eur. J.
2009, 15, 4111–4123; e) S.-i. Kato, M. Kivala, W. B. Schweizer, C.
Boudon, J.-P. Gisselbrecht, F. Diederich, Chem. Eur. J. 2009, 15,
8687–8691.
[4] a) B. Esembeson, M. L. Scimeca, T. Michinobu, F. Diederich, I.
Biaggio, Adv. Mater. 2008, 20, 4584–4587; b) C. Koos, P. Vorreau, T.
Vallaitis, P. Dumon, W. Bogaerts, R. Baets, B. Esembeson, I. Biag-
gio, T. Michinobu, F. Diederich, W. Freude, J. Leuthold, Nat. Pho-
tonics 2009, 3, 216–219.
[12] C. Dehu, F. Meyers, J. L. Brꢅdas, J. Am. Chem. Soc. 1993, 115,
6198–6206.
[13] a) F. H. Allen, Acta Crystallogr. Sect. B 1984, 40, 64–72; b) F. H.
Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen, R.
Taylor, J. Chem. Soc. Perkin Trans. 2 1987, S1–S19.
[5] P. D. Jarowski, Y.-L. Wu, C. Boudon, J.-P. Gisselbrecht, M. Gross,
W. B. Schweizer, F. Diederich, Org. Biomol. Chem. 2009, 7, 1312–
1322.
[14] Similar bond length relationships were found in 38 out of 39 asym-
metric 1-phenylcyclobutenes according to Cambridge Structural Da-
tabase analysis (CSD Version 5.30 + 4 updates, Sep. 2009). For ex-
ample: A. Usman, I. A. Razak, H.-K. Fun, S. Chantrapromma, B.-G.
Zhao, J.-H. Xu, Acta Crystallogr. Sect. E 2001, 57, o990–o991.
[15] a) L. Horner, K. Klꢀpfel, Liebigs Ann. Chem. 1955, 591, 69–98;
b) T. B. Posner, C. D. Hall, J. Chem. Soc. Perkin Trans. 2 1976, 729–
732; c) M. Mantri, O. de Graaf, J. van Veldhoven, A. Gçblyçs, J. K.
von F. D. Kꢀnzel, T. Mulder-Krieger, R. Link, H. de Vries, M. W.
Beukers, J. Brussee, A. P. Ijzerman, J. Med. Chem. 2008, 51, 4449–
4455; d) M. M. Heravi, M. H. Tehrani, K. Bakhtiari, H. A. Oskooie,
J. Chem. Res. 2006, 561–562.
[16] a) Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307–
326; b) A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C.
Giacovazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 1999, 32, 115; c) SHELXL97, Program
for the Refinement of Crystal Structures, G. M. Sheldrick, Universi-
ty of Gçttingen, Gçttingen, 1997.
[6] S. T. Diver, A. J. Giessert, Chem. Rev. 2004, 104, 1317–1382.
[7] a) R. Gompper, Angew. Chem. 1969, 81, 348–363; Angew. Chem.
Int. Ed. Engl. 1969, 8, 312–327; b) R. Huisgen, Acc. Chem. Res.
1977, 10, 117–124; c) R. Huisgen, Acc. Chem. Res. 1977, 10, 199–
206; d) A. J. Fatiadi, Synthesis 1987, 749–789; e) K. Inanaga, K.
Takasu, M. Ihara, J. Am. Chem. Soc. 2005, 127, 3668–3669;
f) R. W. M. Aben, J. Goudriaan, J. M. M. Smits, H. W. Scheeren,
Synthesis 1993, 37–39; g) H. K. Hall, Jr. M. Abdelkader, M. E. Glo-
gowski, J. Org. Chem. 1982, 47, 3691–3694; h) K. C. Brannock, A.
Bell, R. D. Burpitt, C. A. Kelly, J. Org. Chem. 1964, 29, 801–812;
i) I. Nakamura, R. Nagata, T. Nemoto, M. Terada, Y. Yamamoto, T.
Spꢄth, A. de Meijere, Eur. J. Org. Chem. 2007, 4479–4482.
[8] The acetylides of transition metals are known to undergo thermal
[2+2] cycloaddition; however, their d orbital participation makes
concerted process possible. For example of the reaction of TCNE,
see: a) M. I. Bruce, J. R. Rodgers, M. R. Snow, A. G. Swincer, J.
Chem. Soc. Chem. Commun. 1981, 271–271; b) M. I. Bruce, T. W.
Hambley, J. R. Rodgers, M. R. Snow, A. G. Swincer, J. Organomet.
Chem. 1982, 226, C1–C4; c) M. I. Bruce, T. W. Hambley, M. R.
Snow, A. G. Swincer, Organometallics 1985, 4, 494–500; d) M. I.
Bruce, M. J. Liddell, M. R. Snow, E. R. T. Tiekink, Organometallics
1988, 7, 343–350; e) M. I. Bruce, T. W. Hambley, M. J. Liddell, A. G.
Swincer, E. R. T. Tiekink, Organometallics 1990, 9, 2886–2898;
f) M. I. Bruce, M. E. Smith, B. W. Skelton, A. H. White, J. Organo-
met. Chem. 2001, 637–639, 484–499. For TCNQ, see: g) K.-i.
[17] CCDC-744043 (4m), CCDC-744044 (3m), and CCDC-748317 (3c),
contains the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/
cif.
Received: September 25, 2009
Published online: December 9, 2009
Chem. Eur. J. 2010, 16, 202 – 211
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
211