Competing thermal electrocyclic ring-closure reactions of (2Z)-Hexa-2,4,5-trienals and their schiff bases. Structural, kinetic, and computational studies

Article abstract of DOI:10.1021/jo100558p

(Figure presented) The electrocyclic ring closure reactions of (2Z)-hexa-2,4,5-trienals (vinylallenals) to alkylidene-2H-pyrans and of the corresponding Schiff base derivatives to alkylidenepyridines can be concurrent. Rates of vinylallenal cyclization and imine formation upon addition of n-butylamine determine the extent of the competition. The activation energies for the electrocyclization of a series of 6-substituted-(2Z)-4-tert-butyl-3- methylhexa-2,4,5-trienals 2 and trienimines 4 depend on the steric interactions between the substituents at C6 and the tert-butyl group at C4. Mixtures of alkylidene-2H-pyrans 3 and alkylidenepyridines 5 are obtained with bulky groups at C6, whereas only the latter is obtained with a C6-t-Bu and only the former with substituents of moderate size at C6. The reaction rates were indirectly derived from the empirical observations using a global optimization study based on differential evolution. The cyclizations are torquoselective, and the kinetically favored (E)-alkylidene heterocycles evolve by electrocyclic ring opening/ring closure toward the thermodynamic Z isomers upon extended reaction times. Density functional theory (DFT) calculations of the electrocyclizations helped in their characterization as monorotatory processes with pseudopericyclic features and made it possible to rationalize the reactivity trends and the torquoselectivity.

Full text of DOI:10.1021/jo100558p

pubs.acs.org/joc
Competing Thermal Electrocyclic Ring-Closure Reactions
of (2Z)-Hexa-2,4,5-trienals and Their Schiff Bases. Structural, Kinetic,
and Computational Studies†
‡
Jose A. Souto, Martın Perez, Carlos Silva Lopez, Rosana Alvarez, Alicia Torrado, and
ꢀ
ꢀ
ꢀ
ꢀ
Angel R. de Lera*
´
ꢀ
Departamento de Quımica Organica, Universidade de Vigo, Campus As Lagoas-Marcosende, 36310 Vigo,
ꢀ
‡
Spain. Present address: Centro de Investigacion Lilly S.A., Alcobendas (Madrid), Spain.
qolera@uvigo.es
Received March 30, 2010
The electrocyclic ring closure reactions of (2Z)-hexa-2,4,5-trienals (vinylallenals) to alkylidene-2H-
pyrans and of the corresponding Schiff base derivatives to alkylidenepyridines can be concurrent.
Rates of vinylallenal cyclization and imine formation upon addition of n-butylamine determine the
extent of the competition. The activation energies for the electrocyclization of a series of
6-substituted-(2Z)-4-tert-butyl-3-methylhexa-2,4,5-trienals 2 and trienimines 4 depend on the steric
interactions between the substituents at C6 and the tert-butyl group at C4. Mixtures of alkylidene-
2H-pyrans 3 and alkylidenepyridines 5 are obtained with bulky groups at C6, whereas only the latter
is obtained with a C6-t-Bu and only the former with substituents of moderate size at C6. The reac-
tion rates were indirectly derived from the empirical observations using a global optimization
study based on differential evolution. The cyclizations are torquoselective, and the kinetically
favored (E)-alkylidene heterocycles evolve by electrocyclic ring opening/ring closure toward the
thermodynamic Z isomers upon extended reaction times. Density functional theory (DFT) calcula-
tions of the electrocyclizations helped in their characterization as monorotatory processes with
pseudopericyclic features and made it possible to rationalize the reactivity trends and the torquos-
electivity.
Introduction
alkylidene-2H-pyran 3a (Scheme 1) by electrocyclic ring
1
closure.2 The reaction course was monitored by H NMR,
In the context of ongoing studies on the chemical reactivity
of retinals,1 we have reported that 13-cis-12-tert-butyl-11,7-
retro-retinal 2a, obtained by oxidation of precursor 11,7-
retro-retinol 1a, provided, under mild conditions (25 °C), the
which confirmed the torquoselective3 cyclization of the highly
substituted (2Z)-vinylallenal substructure of 2a to give the
(E)-alkylidene-2H-pyran E-3a followed by its ring opening
and irreversible electrocyclization of 2a to isomer Z-3a
(Scheme 1). It was assumed that vinylallenal 2a, as an obli-
gate intermediate in the irreversible conversion of (E)- to (Z)-
alkylidene-2H-pyrans 3a could be intercepted intermolecu-
larly with n-butylamine, affording the Schiff base 4a, which
then would give rise to a second heterocyclization manifold.
If the rates of ring closure/ring opening of 2a were compar-
able to those of Schiff base formation, then alkylidenepyr-
idines 5a could also be present in the reaction mixture. In
fact, signals attributed to the kinetically favored (E)-alkyli-
denepyridine E-5a in equilibrium with the putative Schiff
base 4a as well as others indicating the irreversible cyclization
† This paper is dedicated to Prof. Pelayo Camps (University of Barcelona) on
the occasion of his 65th birthday.
ꢀ
(1) Silva Lopez, C.; Alvarez, R.; Domınguez, M.; Nieto Faza, O.; de Lera,
ꢀ
A. R. J. Org. Chem. 2009, 74, 1007.
(2) (a) de Lera, A. R.; Torrado, A.; Garcıa, J.; Lopez, S. Tetrahedron Lett.
ꢀ
1997, 38, 7421. (b) de Lera, A. R.; Alvarez, R.; Lecea, B.; Torrado, A.;
Cossıo, F. P. Angew. Chem., Int. Ed. 2001, 40, 557. (c) de Lera, A. R.; Cossıo,
F. P. Angew. Chem., Int. Ed. 2002, 41, 1150.
(3) (a) Houk, K. N. Stereoselective electrocyclizations and sigmatropic shifts
of strained rings: Torquoelectronics. In Strain and Its Implications in Organic
Chemistry; Kluwer Academic Publishers: Dordrecht, 1989; pp 25-37. For
reviews, see: (b) Houk, K. N.; Li, Y.; Evanseck, J. D. Angew. Chem., Int. Ed.
Engl. 1992, 31, 682. (c) Dolbier, W. R., Jr.; Koroniak, H.; Houk, K. N.; Sheu, C.
Acc. Chem. Res. 1996, 29, 471.
DOI: 10.1021/jo100558p
r
Published on Web 06/03/2010
J. Org. Chem. 2010, 75, 4453–4462 4453
2010 American Chemical Society

Souto et al.
JOCArticle
SCHEME 1. Electrocyclic Reactions of (2Z)-Hexa-2,4,5-trienals
2 and Corresponding N-Butylimines 4a
followed the steps previously described for the corresponding
(2E) isomers by oxidation of precursor vinylallenols 1.6 The
analogues 2 share a methyl substituent at C3 and a bulky t-Bu
group at C4 but differ in the substituent at C6, which ranges
from alkenyl groups with one (b) or two (c, d) additional
alkyl groups to substituents with extended conjugation (e), as
well as alkyl (f, g) and aryl (h) groups.
Propargyl alcohols 8b-h were obtained upon treatment of
aldehydes 7b-h with the alkynyl anion derived from protec-
ted (Z)-pent-2-en-4-yn-1-ol 6. After benzoylation of 8b-h,
the regioselective SN20 displacement of propargyl benzoates
9b-h with the cyanocuprate derived from CuCN and t-BuLi7
led to the tert-butyl-substituted (di)vinylallene silyl ethers
10b-h. The steric hindrance of the bulky tert-butyl group
diverted the reaction course of 9g and provided instead as
major product (32%) the propargyl benzoate 11 resulting
from the attack of the organometallic reagent to the silyloxy
group together with tiny amounts (steps c and d, 2%) of 10g.
Upon raising the temperature to 0 °C, the propargyl alcohol
12 was generated instead, thus reinforcing the interpretation
that the SN20 attack on 9g is highly hindered. Deprotection of
silyl ethers 10b-h with TBAF at 0 °C (competing [1,5]-H
sigmatropic shifts of vinylallenols were observed above room
temperature) uneventfully provided (di)vinylallenols 1b-j
(Scheme 2) in good yields, with the exception of 1d (19%
combined for steps c and d).
In order to obtain sufficient amounts of 1g for the reac-
tivity studies, a new synthetic strategy was used to cons-
truct the vinylallenol by Stille coupling reaction8 between
the corresponding bromoallene 15 and (Z)-3-(tributy-
lstannyl)but-2-en-1-ol 179 (Scheme 3). Upon activation with
thionylbromide,10 propargyl alcohol 14, itself obtained from
pivalaldehyde 7g and 3,3-dimethylbutyne 13, was converted
into an inseparable 63:37 mixture of bromo-di-tert-butylal-
lene 15 and propargylic bromide isomer 16, the products of
SN20 and SN2 reactions, respectively. The mixture was sub-
jected to the Stille coupling reaction with vinyltin reagent 17
using Pd2dba3 in DMF.11 Unexpectedly, the coupling reac-
tion afforded a mixture of separable double bond isomers
E-1g/Z-1g in a 71:29 ratio, along with unreacted 16.
aFor other R1 substituents, see Scheme 2.
of the imine 4a4 to Z-5a were detected by 1H NMR together
with those of Z-3a.2
In light of these findings we wondered whether the com-
petition between the heterocyclization manifolds was inher-
ent to system 2a or could be modulated by the proper choice
of substituents in similar substrates. In order to address this
issue and deepen our understanding of the stereoelectronic
factors affecting these electrocyclic ring-closure reactions,
we have expanded the preliminary experimental results to the
study of a series of C6-substituted (2Z)-hexa-2,4,5-trienals 2
and (2Z)-hexa-2,4,5-trienimines 4 (Scheme 2), which are
easily prepared following the established protocol.2 We
present here the experimental characterization of the kinetics
and the torquoselectivity (E/Z isomer ratio of the highly
unstable alkylidene-2H-pyrans 3 and alkylidenepyridines 5)
of the electrocyclic reactions. The plausible reaction mechan-
isms have been interpreted through the global optimization
of the kinetic parameters using differential evolution. More-
over, we have computed the alternative transition structures
leading to the isomeric alkylidene heterocyclic products 3
and 5 using DFT and estimated the torquoselectivity trends
of the heteroelectrocyclization. Finally, the effect of the steric
interactions of the substituents at C4 and C6 was analyzed in
order to discuss the nature of these pericyclic reactions.
Electrocyclic Reactions. Kinetic Studies. Following the
studies conducted with the vitamin A analogue 1a,2 vinylal-
lenols 1b-h (Scheme 2) were treated with TPAP/NMO in
CH2Cl2 at ambient temperature for 10 min. After filtration
of the insoluble salts through a plug of silica gel doped with
Et3N, the eluate was dried and concentrated under vacuum
(5 min). Monitoring of the reaction course by NMR (C6D6
solutions) could start after about 15 min from the addition of
the oxidant to alcohol 1. Analysis of the NMR spectra
revealed an outcome for these systems that departed in some
cases from that described for 1a. Signals indicative of
vinylallenal 2 formation (i.e., a doublet at about δ 10.2 ppm)
Results and Discussion
Synthesis. Although we have synthesized (2Z)-hexa-2,4,5-
trienals protected as dioxolanes with 2Z geometry, they did
not provide an entry to the desired (2Z)-divinylallenals 2
because upon attempted deprotection a Nazarov-type cycliz-
ation ensued that provided alkylidenecyclopentene dioxanes
instead.5 Therefore, the preparation of (2Z)-vinylallenals 2
(7) (a) Lipshutz, B. H. Synthesis 1987, 325. (b) Lipshutz, B. H. Synlett
1990, 119. (c) Lipshutz, B. H.; Sengupta, S. Org. React. (N. Y.) 1992, 41, 135.
(d) Krause, N.; Gerold, A. Angew. Chem., Int. Ed. Engl. 1997, 36, 187.
(8) (a) Metal-Catalyzed Cross-Coupling Reactions, 2nd ed.; de Meijere, A.,
Diederich, F., Eds.; Wiley-VCH: Weinheim, 2004. (b) Nicolaou, K. C.; Bulger,
P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44, 4442.
(4) Despite the fact that diagnostic signals (the imine hydrogen) for the
putative Schiff base 4a were not detected in the 1H NMR spectra throughout
the entire experiment, it was surmised to be an intermediate on route to 5a.
ꢀ
(5) (a) de Lera, A. R.; Garcıa, J.; Hrovat, D. A.; Iglesias, B.; Lopez, S.
Tetrahedron Lett. 1997, 38, 7425. (b) Iglesias, B.; de Lera, A. R.; Rodrıguez,
(9) Piers, E.; McEachern, E. J.; Romero, M. A. Tetrahedron Lett. 1996,
37, 1173.
ꢀ
J.; Lopez, S. Chem.—Eur. J. 2000, 6, 4021.
ꢀ
(10) Corey, E. J.; Boaz, N. W. Tetrahedron Lett. 1984, 25, 3055.
ꢀ
(11) Vaz, B.; Pereira, R.; Perez, M.; Alvarez, R.; de Lera, A. R. J. Org.
Chem. 2008, 73, 6534.
ꢀ
(6) Lopez, S.; Rodrıguez, J.; Garcıa, J.; de Lera, A. R. J. Am. Chem. Soc.
1996, 118, 1881.
4454 J. Org. Chem. Vol. 75, No. 13, 2010

Souto et al.
JOCArticle
SCHEME 2a
aReagents and conditions:(a) n-BuLi, 6, then 7, THF, -78 °C (58-99%). (b) n-BuLi, BzCl, THF, -78 °C (59-99%). (c) CuCN, t-BuLi, Et2O, -78 to 0 °C
(73-92%). (d) n-Bu4NF, THF, 0 °C (43-85%, see text for d and g).
SCHEME 3a
depicted in Scheme 1, the scenario most likely involves the
ring opening of E-3 to 2 and the irreversible ring closure of
the latter to the thermodynamically more stable (Z)-alkylidene-
2H-pyran Z-3.12 The rearrangement of allenal 2g, obtained
by oxidation of the more hindered allenol 1g, showed slower
kinetics, which allowed monitoring of its conversion (com-
plete disappearance of 2g after about 24 h) into 2H-pyran E-3g
(Figure 1). Signals corresponding to E-3g were still present at
much longer reaction times (i.e., E-3g/Z-3g ratio of 80:20 after
70 h at 25 °C). Furthermore, the slow conversion at ambient
temperature allowed to calculate an activation energy in the
range 25-40 °C for the formation of alkylidene-2H-pyran E-3g
of 22.4 kcal/mol.13
aReagents and conditions: (a) n-BuLi, 13, then 7g, THF, -78 °C (70%).
(b) SOBr2, propylene oxide, 25 °C. (c) 17, Pd2(dba)3, DMF, 25 °C (90%,
two steps).
Vinylallenols 1b-j were subsequently subjected to the
optimized protocol for oxidation-imine formation involving
treatment of the C6D6 solutions of the TPAP/NMO oxida-
were seen in the 1H NMR spectra of reaction mixtures from
oxidation of substrates 1c, 1e, and 1g, but not from the other
vinylallenols 1b, 1d, 1f, and 1h. After showing signals attribu-
ted to the fast formation of the E isomers of alkylidene-2H-
pyrans E-3b-f,h at short reaction times (δ=6.60 ppm for H6
was taken as diagnostic signal), the NMR spectra of the
reaction mixture progressively simplified and finally revea-
led signals of another component that were easily correlated
with those of (Z)-alkylidene-2H-pyran Z-3a. Identification
of Z-3 relied on the presence of doublets (J=5.1 Hz) attri-
buted to H-5 and H-6 (with a δ=6.46 ppm for H6) of the 2H-
pyran moiety in its 1H NMR spectrum and was confirmed by
the finding of a strong NOE between the signals for the t-Bu
group and H at C10 (δ=5.39 ppm for Z-3c, for example). As
˚
tion residue with n-BuNH2 and 4 A MS, and the conversion
to products was analyzed by 1H NMR. Starting from 1b, 1d,
and 1f, only the mixture of 2H-pyrans 3 was detected, with no
spectroscopic evidence of either imine or alkylidenepyridine
formation even after monitoring of the reaction for extended
reaction times (12 h).
Upon addition of n-butylamine to the oxidation products
of vinylallenols 1c and 1e (both structurally similar to 1a)
the first formed mixture of allenal 2 and the kinetic 2H-pyran
E-3 evolved to afford a mixture of the thermodynamically
favored 2H-pyrans Z-3c,e and alkylidenepyridines Z-5c,e.
The Z-3/Z-5 ratio (30:70 for c and 40:60 for e at 25 °C) does
not vary significantly with time thus confirming that the ring
opening of Z-3 to 2 is not taking place. The C6-phenyl
analogue 1h also afforded a mixture of the thermodynamic
2H-pyran 3h and alkylidenepyridine 5h (70:30 Z-3 to Z-5
ratio), but the former was the major product.
(12) Alkylidene-2H-pyrans have rarely been described in the chemical
literature, and references to the parent system relate merely to its role as the
R-anhydro conjugate base of pyrilium salts. The elusive nature of these cyclic
enol ethers is most likely due to the tendency of their reactive exocyclic alkene
moiety to protonate in the presence of acid. Compounds substituted on the
pyran ring with an electron-withdrawing group and/or in conjugation with
phenyl substituents have extended lifetimes, allowing the incorporation of
this structural subunit in pyrilium dyes. See: Reynolds, G. A.; Drexhage,
K. H. J. Org. Chem. 1977, 42, 885. The corresponding alkylidenepyrans con-
jugated at the exocyclic olefin with a ketone, which are obtained by treatment
of allenylketones with PPh3, are also stable: Wallace, D. J.; Sidda, R. L.;
Reamer, R. A. J. Org. Chem. 2007, 72, 1051.
The 1H NMR spectra collected after the oxidation/amine
condensation treatment of the C6-tert-butyl substituted
(13) An activation energy for the electrocyclic ring closure of isolated
systems 2b-g could not be determined because of its decomposition upon
attempted purification. The value obtained with the unpurified reaction
residue for the conversion of 2g to E-3g cannot be taken rigourously due to
potential complications of soluble impurities from the reagents on the
kinetics of the process.
J. Org. Chem. Vol. 75, No. 13, 2010 4455

Souto et al.
JOCArticle
FIGURE 1. Time course (1H NMR spectra) for the conversion of vinylallenal 2g into the corresponding 2H-pyrans 3g.
FIGURE 2. Time course (1H NMR spectra) for the conversion of vinylallenal 2g into the N-butyl Schiff base 4g and for the conversion of the
latter into the alkylidenepyridines 5g.
vinylallenol 1g provided a complete picture of the reaction
progress since all purported intermediates and products were
detected. Starting with aldehyde 2g and presumably imine
Z-4g (characterized by a doublet at about δ 8.40 ppm attri-
buted to H2, with the vicinal hydrogen resonance at ca.
δ 6.50 ppm) signals indicating a complete conversion into
alkylidenepyridine E-5g (δ 6.11 ppm, J=6.5 Hz) were noted
and Z-alkylidenepyridine Z-5, respectively, through the kinetic
manifolds described in Scheme 1. At short reaction times the
heterocycles with E-geometry are obtained, but they undergo
ring opening and irreversible cyclization to the Z isomers.
The competition between Schiff base formation and cycliza-
tion of allenals 2 determines the Z-3/Z-5 product ratio. These
rates are sensitive to the steric interactions between the C6
and C4 substituents. The final outcome of the reaction allows
a classification of the series of compounds studied 1b-g into
four groups: (i) divinylallenes with a methyl substituent at
the alkenyl C7 position (2c and 2e) produce mixtures of the
corresponding alkylidenepyridines Z-5 and 2H-pyrans Z-3,
and the latter and aldehyde 2 in the absence of amine;
(ii) the analogue with a phenyl group at C6 afforded mixtures
1
after about 3 h by H NMR. Further stirring afforded a
mixture of E-5g and Z-5g (δ 6.04 ppm), which slowly evolved
toward the latter upon extended reaction times (from 85:15
after 5 h to 33:67 E-5g/Z-5g ratio after 54 h). No 2H-pyran 3g
formation was detected in this case (Figure 2).
To summarize, (di)vinylallenals 2 and their imines 4
provided the cyclized products (Z)-alkylidene-2H-pyran Z-3
4456 J. Org. Chem. Vol. 75, No. 13, 2010

Souto et al.
JOCArticle
FIGURE 3. Proposed reaction models. The species correspond to those of Scheme 1, omitting molecular drawings. The upper model (M1)
describes the formation of alkylidene-2H-pyrans 3, and the lower (M1a and M2a) the formation of the corresponding alkylidenepyridines
5 with added amine. Black circles correspond to species experimentally detected by 1H NMR during the formation of alkylidene-2H-pyrans
3 and alkylidenepyridines 5, whereas gray circles correspond to nondetected species. Dashed arrows indicate nondetermined rate constants.
TABLE 1. Reaction Rate Constants (ꢀ10-5 s-1) with Confidence Intervals for the Formation of 3-5 in the Cases Described in Figure 3
a
k-1
3/5
model
k2
k4
k3
k5
Z-3/Z-5
c
e
g
c
e
g
M1
M1
M1
M1a
M1a
M2a
7.40 ( 0.17
12.95 ( 0.45
0.07 ( 0.01
8.86 ( 0.77
7.55 ( 0.65
2.21 ( 0.07
2.96 ( 0.07
4.23 ( 0.17
6.70 ( 0.57
6.35 ( 0.29
15.66 ( 1.50
13.46 ( 1.00
0.50 ( 0.01
30:70
40:60
0:100
49.9 ( 1.9
51.8 ( 2.6
ak1 for g.
of Z-5h and Z-3h, but vinylallenal 2h was not detected;
(iii) (di)vinylallenes that have reduced steric interactions with
the C4-t-Bu group (2b, 2d, 2f) provided exclusively Z-3 at
extended reaction times even in the presence of amine; (iv)
the system with a t-Bu substituent at C6 afforded vinylallenal
2g, E-3g, and slowly Z-3g, whereas alkylidenepyridines 5g as
well as intermediate imine 4g were obtained upon addition of
n-butylamine.
In order to derive kinetic constants for the reactions depicted
in Scheme 1, we have selected a model based on the mass
conservation law, which will result in an ordinary differential
equations set, instead of the traditional approach that fits each
experimental concentration datum to a common analytical
function (e.g., a logarithmic function) and leads to a function
that may be valid only in the vicinity of the experimental data
used for the fitting. We have chosen to calculate the kinetic con-
stants by solving the so-called parameter estimation problem,15
using the recently developed MATLAB based toolbox AMIGO.16
This approach is formulated as a nonlinear optimization
aimed to find the unknown kinetic constants in order to
minimize the distance between model predictions and experi-
mental data. To solve the limitation of finding suboptimal
solutions due to the dynamic and often nonlinear nature of
the kinetic models, we opted for global optimization methods
and selected a population-based stochastic method termed
differential evolution (DE),17 a genetic algorithm that uses
vector differences for perturbing the vector population.
Calculated parameters and confidence intervals for the
proposed models are shown in Table 1. Values of residuals
were sufficiently low to guarantee good quality in the fits.
Identifiability analyses were performed, which confirmed
that the chosen set of parameters is adequate in each case.
An overall confirmation of the consistency of the selected
models comes from parametric sensitivities, which were used
to show the role of each parameter in a given model (for
details, see Supporting Information).
Kinetic Simulations. The complex behavior of most of the
systems in the presence of amine led us to simulate the
processes using an integrated approach. The Supporting
Information contains all primary empirical data as well as
an in-depth kinetic analysis for all systems studied that takes
1
into consideration the H NMR derived concentrations of
the species detected for each particular substrate. Presented
here is the analysis of the reaction progress when the key
intermediate aldehyde concentration is detected by 1H NMR
(systems c, e, and g) according to reaction model M1, as
shown in Figure 3. The M1a and M2a counterparts add the
analogous and competing 4 to 5 cyclization in the presence of
amine. However, also aldehyde 2c,e was detected in M1a and
not its Schiff bases (4c,e) as in model M2a. Although the
direct conversion of 2c,e to Z-5c,e is not possible (only
through nondetected 4c,e), we decided to simulate the system
to test the suitability of the method and to compute the
remaining rate constants.14
Parameters ki stand for empirical kinetic constants, and
reaction rates are considered to be first order with respect to
all reactants. Furthermore these constants, perhaps with the
exception of k3 in M2a, could be also considered as specific
rates, because the reaction steps involved are elementary.
In systems described by model M1 where the three species
E-3, 2 and Z-3 are detected, k2 values are greater than k-1
(15) Walter, E.; Pronzato, L. Identification of Parametric Models from
Experimental Data; Springer: Masson, 1997.
(16) Balsa-Canto, E.; Banga, J. R. Advanced Model Identification Using
Global Optimization, 9th International Conference in Systems Biology
(14) No direct transformation of E-3 to Z-5 is possible without the
involvement of the intermediate species 2 and 4, and the direct double bond
isomerization of alkylidene-2H-pyrans was found by DFT computations
(not shown) to be noncompetitive with the transformations described in
Scheme 1.
€
(ICSB2008), Goteborg, 2008.
(17) Storn, R.; Price, K. J. Global Optim. 1997, 11, 341.
J. Org. Chem. Vol. 75, No. 13, 2010 4457

Souto et al.
JOCArticle
FIGURE 4. Electrocyclic reactions studied at the B2PLYP-D/TZVP//B3LYP/6-31þG(d,p) level.
TABLE 2. Relative Free Energies (in kcal/mol, Starting from the Reac-
tive Conformation) for the Alternative Cyclizations of (2Z)-Hexa-2,4,5-
trienals (2) and Their N-Methylimine Derivatives (40) to the Double Bond
Isomers of Alkylidene-2H-pyrans 3 and Alkylidenepyridines 50, Respec-
tivelya
(the ring opening of E-3 to 2) and the 2 f Z-3 process is more
rapid. For system g, allenal 2 forms rapidly from E-3 but
evolves to Z-3 at a slower rate. Upon addition of amine
(M1a), the N-electrocyclization of the undetected imine 4c,e
competes with that of vinylallenal to product Z-5. Using
model M2a, Z-5g forms with a specific rate that is about 10
alkylidene-2H-pyrans 3
N-methylalkylidenepyridines 50
Z-3 TS-Z-3 TS-E-3 E-3
Z-50 TS-Z-50 TS-E-50 E-50
times faster than formation of Z-3g (k1=4.23 ꢀ 10-5 s-1
,
b
c
d
f
g
h
-8.7 15.4
-5.3 18.0
-9.1 14.8
-7.6 15.5
-4.8 17.9
-6.2 17.3
14.4 -5.7 -20.1
14.0
19.8
14.0
14.0
20.0
15.6
11.8
13.3
11.9
12.9
14.5
12.1
-18.9
-14.5
-19.6
-16.8
-7.2
model M1). Furthermore, the previous conversion of viny-
lallenal 2g to its Schiff base 4g shows a rate (k3) that is also
nearly 1 order of magnitude higher than the competing
formation of Z-3g, and therefore the formation of the latter
species is observed at a slow rate.
Theoretical Calculations. The intriguing dependence of
the reaction outcome on the substituent at C6 of 2 that
becomes the exocyclic C10 group of the C2-alkylidene het-
erocyclic products led us to computationally explore the
alternative ring-closure reactions of vinylallenals 2 and the
corresponding Schiff bases 4. The study was aimed at pro-
viding support to the most relevant experimental findings: (1) the
low to moderate activation energies for the ring-closure reac-
tions, and (2) the effect on the reactivity and stereoselectivity
(torquoselectivity) of steric interactions between the C4-t-Bu and
the groups attached to the vinylallene C6 position.
In order to obtain meaningful information, we theoretic-
ally explored the reactions of 2b-d,f-h to afford the electro-
cyclization products 3b-d,f-h and of “model” N-methylimines
40b-d,f-h to 50b-d,f-h (they are distinguished from the
experimental N-n-butyl analogues by a prime, Figure 4).18
The transformation of the aldehydes into their imine coun-
terparts was not computed. The effects of charge separation
in the intermediates involved in such condensation reaction
15.0 -1.1 -16.6
13.9 -6.1 -20.5
14.7 -4.6 -18.4
16.9
3.8 -12.9
14.9 -3.9 -19.1
-16.8
aAll computations have been carried out at the B2PLYP-D/TZVP//
B3LYP/6-31þG(d,p) level.28
are expected to yield large artifacts in gas phase calculations,
and moreover, the effect of solvation for highly charged
species is, as of yet, an intractable problem for conventional
continuum models.19 Information regarding the amine con-
densation step, however, can be indirectly acquired by ana-
lysis of the experimental results (vide supra), which supports
the working hypothesis for the whole mechanism.
All theoretical calculations were carried out with the
GAUSSIAN03 suite of programs.20 The popular B3LYP
functional, which has found excellent performance on the
characterization of transition states of pericyclic reactions,21
including the electrocyclic ring closure of the unsubstituted
parent systems,22 was selected. Moreover, after energy opti-
mization and harmonic analysis with B3LYP, energy refine-
ment was carried out with a basis set of triple-ζ quality with
polarization functions (TZVP)23 for all the atoms in con-
junction with the double hybrid B2PLYP functional to take
long-range dispersion interactions into account.24 To make
the perturbative part more affordable given the large basis
functions employed, the resolution of the identity (RI)25 was
used as implemented in the Orca 2.6.35 program.26
(18) For all starting materials discussed, only the structures and energies
of the relevant (those from which cyclization must occur) cZg conformer are
provided (see Supporting Information).
(19) See, for example: (a) Braga, A. A. C.; Ujaque, G.; Maseras, F.
Organometallics 2006, 25, 3647. (b) Sumimoto, N.; Iwane, T.; Takahama, S.;
Sakaki J. Am. Chem. Soc. 2004, 126, 10457.
Table 2 collects the free energies of the computed minima
and transition structures leading to the double bond isomers
of heterocyclic products 3 and 50 relative to those of the acyc-
lic precursors (2 and 40, respectively).
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci,
B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada,
M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.;
Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.;
Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;
Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels,
A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari,
K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.;
Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.;
Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision C.02;
Gaussian, Inc.: Wallingford, CT, 2004.
(21) (a) Goldstein, E.; Beno, B.; Houk, K. N. J. Am. Chem. Soc. 1996, 118,
6036. (b) Houk, K. N.; Beno, B. R.; Nendel, M.; Black, K.; Yoo, H. Y.;
Wilsey, S.; Lee, J. K. J. Mol. Struct. (Theochem) 1997, 398-399, 169.
(22) Rodrıguez-Otero, J. J. Org. Chem. 1999, 64, 6842.
(23) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
(24) (a) Grimme, S. J. Chem. Phys. 2006, 124, 034108. (b) Grimme, S.
J. Comput. Chem. 2004, 25, 1463. (c) Grimme, S. J. Comput. Chem. 2006, 27,
1787.
€
(25) (a) Weigend, F.; Haser, M. Theor. Chem. Acc. 1997, 97, 331.
€
€
(b) Weigend, F.; Kohn, A.; Hattig, C. J. Chem. Phys. 2002, 116, 3175.
(26) http://www.thch.uni-bonn.de/tc/orca/.
4458 J. Org. Chem. Vol. 75, No. 13, 2010

Souto et al.
JOCArticle
FIGURE 6. Computed geometries for transition structures TS-Z-
2c (left) and TS-E-2c (right). Key discussed geometric parameters
are also provided: forming bond distance (in A) and C4-C5-C6
bond angle (in deg) as a measure of allene deformation.
FIGURE 5. Disrotatory electrocyclization of (4Z)-hepta-1,2,4,6-tet-
raene (18) and pseudopericyclic cyclizations of (2Z)-hexa-2,4,5-trienal
(19) and (2Z)-hexa-2,4,5-trienimine (20). A dominant interaction of
the in-plane π-orbital of the allene and the nitrogen lone pair (with a
significant component of a nucleophilic attack [AN]) clearly distin-
guishes the cyclization of E-20 from that of its isomer Z-20 (not
shown) as well as from that of the hydrocarbon parent system 18.
˚
products of the series in favor of the kinetically preferred
TS-E, and this is even more pronounced for the imines (3.0 kcal/
mol for 2c and 6.5 kcal/mol for 40c). Moreover, the transition
state leading to 2H-pyran Z-3c is higher in energy than that
of the simpler system 2f with a C6-methyl substituent (18.0 vs
15.5 kcal/mol), which explains why aldehyde 3c is present in
the mixture. The lower activation energies for the cyclization
to the E isomers provides consistent grounds for the stereo-
selectivity experimentally obtained on the reaction of systems
2 (Scheme 1), which affords the (E)-alkylidene-2H-pyran E-3
at short reaction times and the isomer Z-3 at longer reaction
times, through irreversible electrocyclization after equilibra-
tion via ring closure of 2 and ring opening of E-3. Since the
latter product is formed kinetically, the reverse reaction is
attainable (surmounting 16.1 kcal/mol in the case of c), thus
funneling through the similar process that leads alternatively
to product Z-3 by irreversible cyclization. Similar considera-
tions apply to the kinetic behavior of the Schiff bases 40.27
A computational probe with a very congested allene
moiety 2g was used to explore the validity of these conclu-
sions based on steric interactions. In fact, the energy profile
shown for 2g resembles closely the others, with enhanced
features: the kinetic 2H-pyran E-3g is 3.8 kcal/mol more
unstable than the reactant and the transition state leading to
this structure is the highest of the series (16.9 kcal/mol). The
same analysis applies to the cyclization of the corresponding
imine 40g to the dihydropyridines E-50g via transition struc-
tures remarkably similar to those of the oxa analogues. In
fact, superimposition of E-3g and E-50g illustrates the close
resemblance of these molecules (see Supporting Information).
It is noted that the increased steric congestion in the
proximity of the allene terminus as well as in the vinylallene
C3-C4 bond in 2 and 4 induces a dramatic effect on the
geometries and energies of the reactants, transition struc-
tures, and products relative to the unsubstituted model
systems of our early work (Figure 5).2
In contrast to the flexibility of the parent system, only two
conformers of the starting vinylallenes (s-trans,s-gauche tZg
and s-cis,s-gauche cZg) could be characterized for 2 and 40
because the C3-C4 bond between the trisubstituted double
bonds is highly twisted and the C1-C2 bond is close to
orthogonal. For the relevant cZg conformation (as required
for cyclization) the distance from the heteroatom to the
0
˚
˚
central allene carbon is 3.37 A for 2c and 3.48 A for 4 c.
Common to both heterocyclic variants of the electrocycliza-
tion is the greater torsion of the terminal vinyl carbon of the
allene with these bulky substituents in comparison with the
unsubstituted parent system 19 and E-20.2
Upon reaching the transition states, the forming bond
distances are slightly longer for TS-Zs than for TS-Es, but
the allene angle deformation is greater for the latter; thus
TS-Es are found earlier along the reaction coordinate than
TS-Zs (see Figure 6) in agreement with the experimental
results and the Hammond postulate.
The activation barriers for the cyclization of the analogues
with a methyl substituent at C6 (2f, 40f) or a hydrogen at the
unsaturated C7 position (2b, 2d; 40b, 40d) to their kinetic
products (E-3 and E-50, respectively) are quite similar.
However, alkylidenepyridines E-5f,b,d are not observed
experimentally, which implies that the condensation reaction
of aldehydes 2f, 2b, and 2d and n-butylamine to the imines
E-4f,b,d involves a transition state whose relative energy is
higher than about 16 kcal/mol. Under these circumstances,
the ultimate formation of the thermodynamic isomers Z-3f,
b,d is explained by the comparable activation barriers for the
two electrocyclic reactions (closure and opening).
(27) The electrocyclization reaction for the iminium ion, which could
have been formed in the condensation reaction by the acidic nature of the
molecular Sieves, was also computed in one case (f). The activation barriers
of 27.5 and 28.6 kcal/mol, for the formation of the E and Z dihydropyr-
idinium ion respectively, are much larger than those found for the unproto-
nated imine (14.7 and 15.5 kcal/mol). The values for the iminium ion are close
to those determined for the Z isomers of the Schiff bases, which would evolve
by a classical disrotatory electrocyclic reaction.
(28) The activation energies were also computed at the same level using
the M06-2X functional: (a) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. J. Chem.
Theory Comput. 2006, 2, 364. (b) Zhao, Y.; Truhlar, D. G. J. Chem. Phys.
2006, 125, 194101. (c) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2006, 110,
13126. (d) Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2009, 5, 324.
(e) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157. The average error
in activation energies for the reactions shown in Figure 4 was estimated as
ΔGq (B3LYP) - ΔGq (M06-2X)=0.69 kcal/mol.
The more sterically congested system with a 1-methylpro-
penyl group at C6 (c, which also serves as a truncated model
for a) shows the greatest energy difference between the
activation barriers leading to the isomeric cycloalkylidene
J. Org. Chem. Vol. 75, No. 13, 2010 4459

Souto et al.
JOCArticle
The activation barriers for the cyclization of these deriva-
tives range from 16 to 20 kcal/mol. Since the imine was
expected to form through a rate-limiting step associated
with an energy barrier higher than 16 kcal/mol but competitive
with barriers of 19 kcal/mol, it is expected that, in this scenario,
mixtures of 2H-pyrans 3 and alkylidenepyridines 5 form.
In all cases the electrocyclic reactions are exergonic, as
expected from the energy released by the decumulation of the
allene bond. However, comparing the energy values of
Table 2 with those computed for the parent unsubstituted
system (i.e., methylidenepyridine 23, Figure 5, is stabilized by
40.6 kcal/mol relative to (2Z)-hexa-2,4,5-trienimine 20,
whereas methylidene-2H-pyran 22 is more stable than (2Z)-
hexa-2,4,5-trienal 19 by 22.1 kcal/mol), it is concluded that
the “allene effect”29 is partially compensated by the destabi-
lizing strain of the final products. The energy difference
between the isomeric alkylidene heterocycles ranges from
3.0 (3b) to 8.6 kcal/mol (3g) for the oxygen series and from
1.2 (40b) to 6.5 kcal/mol (40g) for the nitrogen series.
display features of pseudopericyclic processes that occur along
transition structures of π2 aromaticity (Figure 5).2 This inter-
pretation was questioned soon after,32 and the debate is still
ongoing.33-37
Our analysis focused on the consequences in the reactivity
of the presence of a cumulene and a heteroatom at the
interacting termini of the conjugated system, since the atoms
displayorbitals (in-planeπ-bondsandlone pairs, respectively)
that are orthogonal to the skeletal π-bonds. In the bond
forming/breaking process, the transition states contain a locus
or disconnection at which no electrons are exchanged between
the in-plane and the out-of-the-plane sets of orbitals. Lemal38
coined in 1976 the term pseudopericyclic to describe “con-
certed transformation whose primary changes in bonding
compass a cyclic array of atoms at one (or more) of which
non-bonding and bonding atomic orbitals interchange roles”
and predicted as a corollary that pseudopericyclic reactions
should be orbital symmetry allowed regardless of the number
of participating electrons.38 The existence of other processes
that could enjoy the energetic benefit of the pseudopericyclic
topology39 remained somewhat of a curiosity for almost
20 years. Birney demonstrated in comprehensive studies
the pseudopericyclic character of a number of transforma-
tions of conjugated molecules with orbital disconnections
at both termini including cycloadditions, cheletropic reac-
tions, sigmatropic rearrangements, and electrocyclizations.40
Studies of other systems including analogues with a single
disconnection soon followed41a-c,2 and altogether confirmed
that pseudopericyclic reactions (1) do not formally follow the
Woodward-Hoffmann rules, (2) proceed via planar or al-
most planar transition states, and (3) in most cases have
surprisingly low activation energies.40
Moreover, all the processes are torquoselective,3 leading
preferentially to the kinetic E-isomers of the alkylidene-2H-
pyran 3 and alkylidenepyridine 50. The pseudoboat confor-
mation of the six-membered ring structure in the transition
state leading to the E isomer is such that substituents at C6
and C4 orient themselves to prevent steric contacts (note the
t-Bu at C4 pointing upward and the 10-methyl-propenyl at
C6 pointing downward in Figure 6). In this situation, the
interaction between the substituent at C6 and the forming
C
O bond defines the kinetically favored electrocyclic
3 3 3
ring-closing transition state. Further downhill in the reaction
coordinate, however, the ring structure flattens, and close
contact between the substituents at C4 and C6 dominates
and reverses the energetic balance between these isomers. The
latter interaction causes the kinetic (E)-alkylidene-2H-pyran to
be thermodynamically less stable than its Z isomer. Parallel
argumentation is also applicable to the torquoselectivity ob-
served in the alkylidenepyridine formation.
ꢀ
(33) Silva Lopez, C.; Nieto Faza, O.; Cossıo, F. P.; York, D. M.; de Lera,
A. R. Chem.-Eur. J. 2005, 11, 1734.
ꢀ
ꢀ
(34) (a) Matito, E.; Sola, M.; Duran, M.; Poater, J. J. Phys. Chem. B 2005,
ꢀ
ꢀ
109, 7591. (b) Matito, E.; Poater, J.; Duran, M.; Sola, M. ChemPhysChem
2006, 7, 111.
Solvation effects were also taken into consideration when
computing the reaction profiles summarized in Table 2. The
polarizable continuum model (PCM)30 was employed with
benzene parameters and the molecular cavity created with
the UAKS radii set.31 Solvation effects were found to be
small and very systematic along the reaction mechanisms,
and they do not modify the overall energy profile. We
therefore decided to omit this data for the sake of simplicity.
Characterization of the Electrocyclic Reactions. We have
previously discussed the main features of the electrocyclic
reaction of (2Z)-hexa-2,4,5-trienal (19) and (2Z)-hexa-2,4,5-
trienimine (20) using geometric, energetic, and aromaticity
criteria. DFT computations, including NBO analysis for
products and transition structures and aromaticity of the
latter,2b,c allowed the conclusion that these heterocyclizations
(35) (a) Chamorro, E. E.; Notario, R. J. Phys. Chem. A 2004, 108, 4099.
(b) Chamorro, E. E.; Notario, R. J. Phys. Chem. B 2005, 109, 7594.
(36) Duncan, J. A.; Calkins, D. E. G.; Chavarha, M. J. Am. Chem. Soc.
2008, 130, 6740.
(37) Sakai, S. Theor. Chem. Acc. 2008, 120, 177.
(38) (a) Ross, J. A.; Seiders, R. P.; Lemal, D. M. J. Am. Chem. Soc. 1976,
98, 4325. (b) Bushweller, C. H.; Ross, J. A.; Lemal, D. M. J. Am. Chem. Soc.
1977, 99, 629.
(39) Mention of pseudopericyclic reactions can be found, in addition to
refs 35, 37, and 38, in the following: (a) Burke, L. A.; Elguero, J.; Leroy, G.;
Sana, M. J. Am. Chem. Soc. 1976, 98, 1685. (b) Henriksen, U.; Snyder, J. P.;
Halgren, T. A. J. Org. Chem. 1981, 46, 3767. (c) Okamura, W. H.; Peter, R.;
Reischl, W. J. Am. Chem. Soc. 1985, 107, 1034. (d) Elnagar, H.; Okamura, W. H.
J. Org. Chem. 1988, 53, 3060. (e) Wentrup, C.; Netsch, K.-P. Angew. Chem., Int.
Ed. Engl. 1984, 23, 802. (f) Koch, R.; Wong, M. W.; Wentrup, C. J. Org. Chem.
1996, 61, 6809. (g) Bibas, H.; Koch, R.; Wentrup, C. J. Org. Chem. 1998, 63,
2629.
(40) (a) Ham, S.; Birney, D. M. Tetrahedron Lett. 1994, 35, 8113.
(b) Birney, D. M.; Wagenseller, P. E. J. Am. Chem. Soc. 1994, 116, 6262.
(c) Birney, D. M. J. Org. Chem. 1994, 59, 2557. (d) Wagenseller, P. E.; Birney,
D. M.; Roy, D. J. Org. Chem. 1995, 60, 2853. (e) Birney, D. M. J. Org. Chem.
1996, 61, 243. (f) Birney, D. M.; Ham, S.; Unruh, G. R. J. Am. Chem. Soc.
1997, 119, 4509. (g) Birney, D. M.; Xu, X. L.; Ham, S.; Huang, X. M. J.
(29) (a) Bond, D. J. Org. Chem. 1990, 55, 661. (b) Jensen, F. J. Am. Chem.
Soc. 1995, 117, 7487.
ꢀ
(30) (a) Tomasi, J.; Persico, M. Chem. Rev. 1994, 94, 2027. (b) Mineva, T.;
Russo, N.; Sicilia, E. J. Comput. Chem. 1998, 19, 290. (c) Cossi, M.; Scalmani,
G.; Rega, N.; Barone, V. J. Chem. Phys. 2002, 117, 43. (d) Tomasi, J.;
Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999.
(31) Takano, Y.; Houk, K. N. J. Chem. Theory Comput. 2005, 1, 70.
(32) (a) Rodrıguez-Otero, J.; Cabaleiro-Lago, E. M. Angew. Chem., Int.
Ed. 2002, 41, 1147. (b) Rodrıguez-Otero, J.; Cabaleiro-Lago, E. M. Chem.-
Eur. J. 2003, 9, 1837. (c) Cabaleiro-Lago, E. M.; Rodrıguez-Otero, J.;
J. Org. Chem. 1997, 62, 7114. (h) Quideau, S.; Looney, M. A.; Pouysegu, L.;
Ham, S.; Birney, D. M. Tetrahedron Lett. 1999, 40, 615. (i) Birney, D. M.; Xu,
X. L.; Ham, S. Angew. Chem., Int. Ed. 1999, 38, 189. (j) Shunway, W.; Ham,
S.; Moer, J.; Whittlesey, B. R.; Birney, D. M. J. Org. Chem. 2000, 65, 7731.
(k) Birney, D. M. J. Am. Chem. Soc. 2000, 122, 10917. (l) Bartsch, R. A.;
Chae, Y. M.; Ham, S.; Birney, D. M. J. Am. Chem. Soc. 2001, 123, 7479.
(m) Zhou, C.; Birney, D. M. J. Am. Chem. Soc. 2002, 124, 5231. (n) Birney,
D. M. Org. Lett. 2004, 6, 851. (o) Wei, H.-X.; Zhou, C.; White, J. M.; Birney,
D. M. Org. Lett. 2004, 6, 4289. (p) Ji, H.; Xu, X.; Ham, S.; Hammad, L. A.;
Birney, D. M. J. Am. Chem. Soc. 2009, 131, 528.
ꢀ
~
ꢀ
Garcıa-Lopez, R. M.; Pena-Gallego, A.; Hermida-Ramon, J. M. Chem.-
Eur. J. 2005, 11, 5966.
4460 J. Org. Chem. Vol. 75, No. 13, 2010

Souto et al.
JOCArticle
It is useful to compare the activation energies computed
for the parent canonical systems of Figure 5 with those of
Table 2 for the experimental systems depicted in Scheme 1.
The computed values for the electrocyclic ring closure of 2 to
3 and 40 to 50 are in the range of 14-18 and 12-20 kcal/mol,
respectively (Table 2), whereas the value experimentally
determined for the formation of E-3g is 22.4 kcal/mol. These
are much higher than those of the unsubstituted systems of
Figure 5 (6.4 kcal/mol for E-20 and 8.6 kcal/mol for 19; cf.
17.4 kcal/mol for the hydrocarbon 18; all values computed at
the B3LYP/6-31þG* level of theory).2 By comparison, the
values of Table 228 are just below those experimentally
determined for the heterocyclizations of substituted dienones
to 2H-pyrans (21.7--22.4 kcal/mol)42 and 1-azahexatrienes to
dihydropyridines (17.5--22.5 kcal/mol).43-45 From the com-
parative analysis we conclude that the beneficial effect of the
linear cumulene Csp and the heteroatom at the electrocycliza-
tion termini on the activation energy is less important when
coplanarity of the pericyclic array of atoms is compromised due
to steric interactions of substituents and back-strain effects.
The transition structures (as deduced from the computa-
tional studies of model c, Figure 6) of the heterocyclization
processes described in this work are distorted. Nonetheless,
they preserve an area of local planarity (X-C1-C2-C3-
C4-C5) that facilitates the in-plane interaction of the lone
pairs and the twisting C5-C6 allene π-bond. This arrange-
ment should allow a less-costly reaction pathway that partly
avoids the bond disrotation of a pericyclic boat-like transi-
tion state. In fact, an intrinsic reaction coordinate (IRC)
calculation for TS-Z-3c clearly shows the pronounced mono-
rotatory nature of the electrocyclic ring closure of 2c. It also
suggests the nucleophilic role for the heteroatom since the
oxygen atom does not exhibit any rotation and only under-
goes an in-plane translational motion toward the cumulene
atom (see Supporting Information). All together the differ-
ent criteria analyzed suggest the unbalanced contribution of
features of both pericyclic and pseudopericyclic electrocycli-
zations in the rearrangement of 2 and 4. Whereas the
canonical unsubstituted systems (Figure 5) are primarily
pseudopericyclic, the incorporation of substituents along
the perimeter of the 6πe- system likely shifts the topology
and the geometry toward the pericyclic side while preserving
some features of the pseudopericyclic arrangement.
Conclusion
A complex kinetic behavior has been found for the elec-
trocyclic ring closure of C6-substituted (2Z)-4-tert-butyl-
3-methylhexa-2,4,5-trienals (vinylallenals) 2 to the corres-
ponding alkylidene-2H-pyrans 3 in the presence of n-butyl-
amine due to competing formation of the Schiff base
derivatives 4 and their cyclization to alkylidenepyridines 5.
The greater the steric interactions between the alkenyl sub-
stituents at the C6 position (in particular with a C7-methyl
group at a C7-C8 double bond) and the bulky t-Bu group at
C4 along the cyclization processes, the higher the activation
energies for both heterocyclizations and the torquoselectivity.
Moreover, reversibility enables transformation of the kinetic
E products into the thermodynamic Z isomers. In general mix-
tures of the thermodynamic (Z)-2H-pyrans and (Z)-alkylide-
nepyridines are the final outcome with bulky substituents at C6.
For sterically less bulky analogues the formation of the imine
must have higher activation energies than the O-electrocycliza-
tion manifold and therefore (Z)-alkylidenepyridines are not
formed. The di-tert-butylvinylallenal 2g is quantitatively con-
verted into the alkylidenepyridines after imine formation with-
out competing cyclization to 2H-pyran 3g. The integration of
the experimental data within a kinetic simulation study based
on differential evolution fully supported the mechanistic inter-
pretation of the process. Finally, DFT calculations of the
electrocyclic ring closure processes help explain the reactivity
trends and the torquoselectivity of both heterocyclic variants of
the 6πe- electrocyclization. The heterocyclizations described in
this work can be considered to benefit from the pseudoper-
icyclic topology due to the interaction of the heteroatom in-
plane lone pair and the rotating allene terminal π-bond. The
monorotatory46 heterocyclization transition structures show a
deviation from planarity (another feature in contrast with the
canonical planar or almost planar TSs of pseudopericyclic
reactions) that might reflect the torsions required for the
proper orientation of the lone pair relative to the twisting
C-C allene bond of 2 and 4 in the transition state. The
transition states thus exhibit features of the pseudopericyclic
and the pericyclic orbital topologies. In fact, Birney recently
suggested that any transition state property (energy, geome-
try, electronic distribution, aromaticity, etc.) is likely to reflect
(41) (a) Luo, L.; Barberger, M. D.; Dolbier, W. R., Jr. J. Am. Chem. Soc.
1997, 119, 12366. (b) Fabian, W. M. F.; Bakulev, V. A.; Kappe, C. O. J. Org.
Chem. 1998, 63, 5801. (c) Fabian, W. M. F.; Kappe, C. O.; Bakulev, V. A.
ꢀ
J. Org. Chem. 2000, 65, 47. (d) Alajarın, M.; Vidal, A.; Sanchez-Andrada, P.;
Tovar, F.; Ochoa, G. Org. Lett. 2000, 2, 965. (e) Rodrıguez-Otero, J.;
ꢀ
~
Cabaleiro-Lago, E. M.; Hermida-Ramon, J. M.; Pena-Gallego, A. J. Org.
Chem. 2003, 68, 8823. (f) Cabaleiro-Lago, E. M.; Rodrıguez-Otero, J.;
ꢀ
Hermida-Ramon, J. M. J. Phys. Chem. A 2003, 107, 4962. (g) Cabaleiro-Lago,
E. M.; Rodrıguez-Otero, J.; Varela-Varela, S. M.; Pena-Gallego, A.;
~
ꢀ
Hermida-Ramon, J. M. J. Org. Chem. 2005, 70, 3921. (h) Alajarın, M.;
Sanchez-Andrada, P.; Vidal, A.; Tovar, F. J. Org. Chem. 2005, 70, 1340.
ꢀ
ꢀ
(i) Alajarın, M.; Ortın, M. M.; Sanchez-Andrada, P.; Vidal, A.; Bautista, D.
Org. Lett. 2005, 7, 5281. (j) Alajarın, M.; Sanchez-Andrada, P.; Lopez-
ꢀ
ꢀ
ꢀ
~
Leonardo, C.; Alvarez, A. J. Org. Chem. 2005, 70, 7617. (k) Pena-Gallego,
A.; Rodrıguez-Otero, J.; Cabaleiro-Lago, E. M. Eur. J. Org. Chem. 2005, 70,
3228. (l) Rode, J. E.; Dobrowolski, J. C. J. Phys. Chem. A 2006, 110, 207.
(m) Rode, J. E.; Dobrowolski, J. C. Chem. Phys. Lett. 2007, 449, 240.
ꢀ
(n) Calvo-Losada, S.; Quirante Sanchez, J. J. J. Phys. Chem. A 2008, 112,
8164. (o) Jones, G. O.; Li, X.; Hayden, A. E.; Houk, K. N.; Danishefsky, S. J.
ꢀ
Org. Lett. 2008, 10, 4093. (p) Alajarın, M.; Bonillo, B.; Sanchez-Andrada, P.;
Vidal, A.; Bautista, D. Org. Lett. 2009, 11, 1365. (q) Forte, L.; Lafortune, M.
C.; Bierzynski, I. R.; Duncan, J. A. J. Am. Chem. Soc. 2010, 132, 2196.
(42) (a) Schiess, P.; Chia, H. L. Helv. Chim. Acta 1970, 53, 485. (b) Schiess,
P.; Seeger, R.; Suter, C. Helv. Chim. Acta 1970, 53, 1713. For a DFT study of
2,4-pentadienals and 2H-pyrans, see: (c) Wang, Z.; Day, P. N.; Pachter, R.
Chem. Phys. Lett. 1995, 237, 45–52. (d) Wang, Z.; Day, P. N.; Pachter, R.
J. Phys. Chem. 1995, 99, 9730.
(43) Maynard, D. F.; Okamura, W. H. J. Org. Chem. 1995, 60, 1763.
(44) The activation energies for the ring closure of the parent (2Z)-hexa-
1,3,5-triene to cyclohexa-1,3-diene are on the order of 30 kcal/mol. (a) Lewis,
K. E.; Steiner, H. J. Chem. Soc. 1964, 3080. (b) Marvell, E. N.; Caple, G.;
ꢀ
Schatz, B. Tetrahedron Lett. 1965, 385. (c) Vogel, E.; Grimme, W.; Dinne, E.
Tetrahedron Lett. 1965, 391. (d) Marvell, E. N.; Caple, G.; Schatz, B.; Pippin,
W. Tetrahedron Lett. 1973, 29, 3781.
(46) The subset of molecular rearrangements that contain a single dis-
connection in the cyclic array of orbitals and therefore only require a
monorotatory movement of the terminus along the reaction coordinate to
assist the development of the new σ-bond were shown to have pseudoper-
icyclic features. See ref. 40 for a general discussion and ref. 41a for
computational analysis of a monorotatory pseudopericyclic electrocycliza-
tion.
(45) Highly substituted systems containing the dienylallene substructure
have been found to cyclize below room temperature to provide the corre-
sponding alkylidenecyclohexadienes: (a) Reischl, W.; Okamura, W. H.
J. Am. Chem. Soc. 1982, 104, 6115. (b) Okamura, W. H.; Peter, R.; Reischl,
W. J. Am. Chem. Soc. 1985, 107, 1034. (c) Elnagar, H. Y.; Okamura, W. H.
J. Org. Chem. 1988, 53, 3060.
J. Org. Chem. Vol. 75, No. 13, 2010 4461

Souto et al.
JOCArticle
a proportional mixing of prototypical pseudopericyclic and
pericyclic allowed transition states.40p
kindly thank Dr. E. Balsa-Canto, from the Process Engi-
neering Group, IIM-CSIC (Vigo, Spain), for her advice in
the solution of the parameter estimation problem and Prof.
F. P. Cossıo (UPV) for helpful discussions.
Acknowledgment. The authors are grateful to the Eur-
opean Union (EPITRON, IP-LSCH-CT-2005-518417), the
MEC-Spain (SAF2007-63880-FEDER, FPI Contract to J.
A.S.), Xunta de Galicia (Parga Pondal Contract to C.S.L.
and M.P.R.; Grant 08CSA052383PR from DX IþDþi;
Supporting Information Available: Experimental proce-
dures, analytical and spectral characterization data for
compounds, results of kinetic studies, SCF energies, Carte-
sian coordinates and a .mpg movie for the heterocyclization
reaction. This material is available free of charge via the
Internet at http://pubs.acs.org.
ꢀ
Consolidacion 2006/15 from DXPCTSUG) for financial
ꢀ
support and the Centro de Supercomputacion de Galicia
(CESGA) for generous allocation of computer time. We
4462 J. Org. Chem. Vol. 75, No. 13, 2010