Cryptocope rearrangement of 1,3-dicyano-5-phenyl-4,4-d2-hexa-2,5-diene. Chameleonic or centauric?

Article abstract of DOI:10.1021/ja992137z

The 'centauric' model for evaluation of the effect of radical- stabilizing perturbations on the Cope rearrangement conjectures independent action of substituents that make conflicting electronic demands on the two halves of the transition region. The present test of this conjecture compares 1,3-dicyano-[5-protio]-hexa-1,5-diene (1(H)) and 2-phenylhexa-1,5-diene, with 1,3-dicyano-5-phenylhexa-1,5-diene (1(Ph)). Thermochemical information required for a proper comparison includes new data of the van't Hoff type on conjugative interaction of cyano with the carbon-carbon double bond, reevaluation of the radical-stabilizing potential of the cyano group on secondary and allyl radicals, comparison with the (reevaluated) stabilizing effect of cyano in 'nodal' positions of the Cope transition region, and determination of the enthalpy and entropy of activation of the cryptoCope rearrangement of otherwise Cope-incompetent, thermodynamically more stable hexa-2,5-dienes related by prototropy to the Cope-competent hexa-1,5-dienes above. The 'chameleonic' model is concluded to be unsatisfactory, while the 'centauric' is in better, if not complete, accord with experiment.

Full text of DOI:10.1021/ja992137z

J. Am. Chem. Soc. 1999, 121, 10967-10975
10967
CryptoCope Rearrangement of
1,3-Dicyano-5-phenyl-4,4-d2-hexa-2,5-diene. Chameleonic or
Centauric?
W. von E. Doering* and Yonghui Wang
Contribution of the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138-2902
ReceiVed June 22, 1999
Abstract: The “centauric” model for evaluation of the effect of radical-stabilizing perturbations on the Cope
rearrangement conjectures independent action of substituents that make conflicting electronic demands on the
two halves of the transition region. The present test of this conjecture compares 1,3-dicyano-[5-protio]-hexa-
1,5-diene (1(H)) and 2-phenylhexa-1,5-diene, with 1,3-dicyano-5-phenylhexa-1,5-diene (1(Ph)). Thermochemical
information required for a proper comparison includes new data of the van’t Hoff type on conjugative interaction
of cyano with the carbon-carbon double bond, reevaluation of the radical-stabilizing potential of the cyano
group on secondary and allyl radicals, comparison with the (reevaluated) stabilizing effect of cyano in “nodal”
positions of the Cope transition region, and determination of the enthalpy and entropy of activation of the
cryptoCope rearrangement of otherwise Cope-incompetent, thermodynamically more stable hexa-2,5-dienes
related by prototropy to the Cope-competent hexa-1,5-dienes above. The “chameleonic” model is concluded
to be unsatisfactory, while the “centauric” is in better, if not complete, accord with experiment.
An exploration is continued here of the proposition that the
response of Cope’s rearrangement to perturbation by radical-
stabilizing substituents is fruitfully thought of as “centauric”,
in addition to “chameleonic”.1 The latter gives name to a long
held conceptual scheme, in which the Cope is viewed as a
continuum between a transition region sometime-called “aro-
matic”, and a “biradicaloid”, thought closely to resemble the
cyclohexa-1,4-diyl diradical.2-4 Throughout this continuum,
retention of a basic C2h symmetry is imagined; that is, the two,
three-carbon halves remain essentially identical in their elec-
tronic character. At one furthermost extreme, the two halves
simulate the character of allyl radicals and are expected to
respond to radical-stabilizing substituents in “active”, 1-(6-) and
3-(4-) positions, but not in “nodal”, 2-(5-) positions. At the other
extreme, the diyl, the reverse obtains: the four previously active
positions are expected to become inactive while the two nodal,
now having acquired the characteristics of an aliphatic secondary
radical, become active (Scheme 1).
Scheme 1
When substituents are positioned to compete with each other,
one in active positions and the other in a nodal position, new
questions arise. Is the transition region obligated to retain its
basic, chameleonic C2h symmetry and thus accommodate the
opposing demands in the two halves by satisfying the full
stabilizing potential of either one or the other, but not both? Or
may it move in a new, symmetry-lowering dimension along a
centauric continuum, in which one-half responds to the allylic
demand, and the other to the secondary radical-like demand?
The thinking is caricatured in Scheme 1, in which the specific
illustration is the subject of the present paper.
Distinction between the two propositions is quantitative.
Given that the radical-stabilizing abilities of a perturbing
substituent have been established in each of the two extremes
of the chameleonic continuum, are two groups in positions of
opposing demands no more effective than the stronger of the
two, or does each contribute its full potential? In a previous
paper, in which phenyl has been employed as the perturbing
substituent, the centauric model appears to have been the more
successful.1 Phenyl as the perturbing group, however, has
inescapable drawbacks. One is the sensitivity to steric factors
of both its conjugative interaction with double bonds and its
(1) Doering, W. v. E.; Wang, Y.-h. J. Am. Chem. Soc. 1999, 121, 10112-
10118.
(2) (a) Dewar, M. J. S.; Wade, L. E., Jr. J. Am. Chem. Soc. 1973, 95,
290 (“...1 (hexa-1,5-diene) itself may be more or less poised between the
biradical and pericyclic mechanisms and that the balance can be displaced
in either way by appropriate substitution.”).
(3) Wehrli, R.; Schmid, H.; Bellus, D.; Hansen, H.-J. HelV. Chim. Acta
1977, 60, 1325-1356.
(4) (a) Gajewski, J. J.; Conrad, N. D. J. Am. Chem. Soc. 1978, 100,
6269-6270. (b) Gajewski, J. J.; Conrad, N. D. J. Am. Chem. Soc. 1979,
101, 6693-6704.
10.1021/ja992137z CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/11/1999

10968 J. Am. Chem. Soc., Vol. 121, No. 47, 1999
Doering and Wang
ability to stabilize radicals. Another is its unsuitability, at the
moment at least, for calculation at the highest level of theory
required to deal with the Cope rearrangement. The importance
of cooperative interaction between experiment and theory, here
and in general, is self-evident. By contrast, the cyano group is
amenable to theoretical calculation,5 has a small, probably
negligible, steric factor as will be supported below,6 but suffers
from its own disadvantages, as we shall see.
Scheme 2
1,3,5-Tricyanohexa-1,5-diene might have seemed the centaur
of choice, but, owing to the high acidity of its 1,3-dicyanopro-
pene system, was deemed likely to destruction by facile
intramolecular Michael cyclization under the conditions of Cope
rearrangement,7 and therefore unsuitable. That 2-cyanohexa-
1,5-diene, a second compound needed to construct the centauric
model, was unknown constituted a further dissuasion. Choice
finally settled on a comparison of 1,3-dicyanohexa-1,5-diene
(1) and 1,3-dicyano-5-phenylhexa-1,5-diene (2), both labeled
with deuterium in the 4,4-positions to break the degeneracy of
their Cope rearrangements.8 Substitution of cyano by phenyl in
the 5-position had the further advantage that both 2-phenylhexa-
1,5-diene and 2,5-diphenylhexa-1,5-diene had already been
studied by Dewar and Wade,9 however much the absolute values
of the activation parameters of the former might beg for
refinement. As a further attraction, three examples of the
replacement of allyl by a 2-phenylallyl group in Cope rear-
rangements already existed in the literature for comparison.
Scheme 3
CryptoCope Rearrangement
Syntheses of the two dicyano compounds, 1(H)-d2 and 1-
(Ph)-d2, take advantage of a high acidity of 1,3-dicyanopropene
(Scheme 2) that would allow generation of a carbanion for
allylation by reaction with 3,3-dideuterio-3-tosylyoxypropene
and 3,3-dideuterio-2-phenyl-3-tosylyoxypropene, respectively.
These displacements in fact proceeded well, if only in dimethyl
sulfoxide as solvent,10 with a regioselectivity of ∼94%.
Disastrously at first sight, the products were not the Cope-
competent 1,5-dienes, but, by rapid prototropy, the thermody-
namically more stable, Cope-incompetent 2,5-isomers.
There seeming to be no obvious way of isolating and studying
the 1,5-dienes directly (barely visible in NMR), the project was
near abandonment until it was recognized that the thermal
rearrangement of the small quantities of Cope-competent 1,5-
diene in equilibrium might be followed by observing the rate
of increase in hydrons at the 4,4-position of the Cope-
incompetent 2,5-dienes (a cryptoCope rearrangement!) under
rapidly equilibrating conditions. The unusually complicated
situation is depicted in Scheme 3. To be trustworthy, this
approach demands sufficiently fast maintenance of preequilib-
rium to bring the system into the Curtin-Hammett domain;11
that is, equilibration within each of the columns of Scheme 3
(5) (a) Hrovat, D. A.; Borden, W. T.; Vance, R. L.; Rondan, N. G.; Houk,
K. N.; Morokuma, K. J. Am. Chem. Soc. 1990, 112, 2018-2019. (b) Beno,
B. R.; Hrovat, D. A.; Lange, H.; Yoo, H.-y.; Houk, K. N.; Borden, W. T.
J. Am. Chem. Soc. 1999, 121, in press.
(6) Vo¨gtle, F.; Neumann, P.; Zuber, M. Chem. Ber. 1972, 105, 2955-
2962.
(7) For the same reason at least, we shall not examine 1,3,4,6-
tetracyanohexa-1,5-diene for its bearing on the theoretically uncovered
prediction of a synergistic effect,5b but see Figure 2 (I). Neither do we intend
to pursue 2- or 3-cyanohexa-1,5-diene, or 1,4-, 3,4-, or 2,5-dicyanohexa-
1,5-diene.
(8) Sunko, D. E.; Humski, K.; Malojcic, R.; Borcic, S. J. Am. Chem.
Soc. 1970, 92, 6534-6538.
(9) Dewar, M. J. S.; Wade, L. E., Jr. J. Am. Chem. Soc. 1977, 99, 4417-
4424.
must be much faster than the Cope rearrangements linking the
two columns. The resulting experimental enthalpy and entropy
of activation would then consist of the sum of those of the Cope
rearrangement (on the necessary assumption that deuterium
interchange originates in that reaction alone) and the difference
in enthalpies and entropies of formation of the more stable,
kinetically incompetent 2,5-isomers and the less stable, kineti-
cally competent 1,5-isomers. Consequently, knowledge of these
differences in conjugative interaction must be in hand if the
concealed Cope activation parameters are to be separated from
the experimentally observed parameters. Note that there is no
obvious way of resolving the relative contribution of (Z) and
(E) isomers to the rearrangement. Despite the several complica-
(10) Zaugg, H. E. J. Am. Chem. Soc. 1961, 83, 837-840.
(11) Klumpp, G. L. ReactiVity in Organic Chemistry; Wiley: New York,
1982; pp 231-249.

1,3-Dicyano-5-phenyl-4,4-d2-hexa-2,5-diene
J. Am. Chem. Soc., Vol. 121, No. 47, 1999 10969
Table 1. Specific Rate Constants and Derived Activation
Parameters for the Reversible Interconversion of 4,4- and
6,6-D2-1,3-dicyanohexa-2,5-diene (H) and 4,4- and
6,6-D2-1,3-dicyano-5-phenylhexa-2,5-diene (C6H5) in Isopropyl
Alcohol-d8
Scheme 4
H
C6H5
(k1 + k-1)
a,b
a,b
T, °C
(k1 + k-1
)
T, °C
165.2 ( 0.3
176.9 ( 0.1
185.4 ( 0.1
194.1 ( 0.1
2.88 ( 0.10
7.92 ( 0.26
14.84 ( 0.11
28.24 ( 0.60
164.9 ( 0.2
176.7 ( 0.2
185.2 ( 0.2
194.4 ( 0.3
8.84 ( 0.27
23.24 ( 0.30
41.80 ( 1.6
81.07 ( 2.8
Arrhenius Parameters
Ea ) 32.36 ( 0.63 kcal mol-1
log A ) 10.45 ( 0.27
Ea ) 30.36 ( 0.59 kcal mol-1
log A ) 10.10 ( 0.28
Eyring Parametersc
∆Hq ) 31.5 kcal mol-1
∆Hq ) 29.5 kcal mol-1
∆Sq ) -13.5 ( 1.2 cal mol-1 K-1 ∆Sq ) -15.1 ( 1.3 cal mol-1 K-1
a In units of 10-6 s-1 b Calculated by linear regression of the standard
.
expression for reversible first-order reactions: (k1 + k-1) ) ln[(xe -
xo)/(xe - xt)]/t, where xe ) 47.21% of 6,6-d2-isomers. c Calculated at
179.7 °C.
tions, the main conclusion will not be compromised because
comparison is between pairs of systems that are subject to the
same, presumably small differences in steric and electronic
influence of the allyl and 2-phenylallyl moieties on the ground
states.
The kinetic acidity of the 1,3-dicyanopropene system in 1,3-
dicyanohexa-2,5-diene is in fact so high that fast deuterium
exchange as an indication of rapid establishment of preequi-
librium is easily realized in methanol-d4 as solvent, even without
a basic catalyst. Were it not for the high proclivity of the system
to undergo Michael addition and consequently induce intolerably
low recoveries at the higher temperatures needed to follow the
cryptoCope rearrangement, this solvent would be fine. In the
event, a tolerable compromise is found in perdeuterioisopropyl
alcohol.
Sunko et al.8 and Gajewski4b for tetradeuteriohexa-1,5-diene.
Activation parameters for reversible exchange in the Cope-
incompetent isomers are calculated in the usual manner (Table
1, repeated in the upper part of Scheme 4).
The difference in enthalpy of activation of the 5-protio and
5-phenyl compounds is 2.0 ( 1.2 kcal mol-1. In an effort to
improve the precision and accuracy of this quantity, the
relatively small, ∼3-fold ratio of their specific rate constants is
turned to good use by effecting rearrangement of both com-
pounds together in the same flask under identical conditions of
solvent, temperature, etc. The window offered is not large. From
the data in Table 2, a seemingly improved value of 1.6 ( 0.6
kcal mol-1 for the difference in enthalpies of activation is
obtained. A weighted average of 1.7 kcal mol-1 is recom-
mended.
In 1,3-dicyano-5-phenylhexa-2,5-diene, deuterium exchange
is slower, but can be sufficiently accelerated by 10% pyridine,
an addition which incidentally improves the quality of the NMR
spectrum for analytical purposes. At the higher temperatures
required for the cryptoCope rearrangement to transfer deuterium
to C-6 from its initially predominant position at C-4, recovery
has dropped to 80% by the time the reaction has proceeded
90% toward equilibrium. Perhaps more seriously, the sum of
the concentrations of educt and product, which can be measured
independently in this instance relative to hydrons at C-2, also
suffers erosion, its value gradually falling to ∼90% by the time
equilibrium is reached at the highest temperature (Table S2 of
Supporting Information). If it be assumed in the absence of any
other explanation that hydrons at C-6 are slowly lost by
exchange with deuterium (perhaps by reversible homodienyl
rearrangement), the relative concentrations of educt and product
may be normalized to a sum of 100% for the purpose of
calculating (k1 + k-1). No such adjustment can be made with
1,3-dicyano-4,4-dideuteriohexa-2,5-diene, if indeed it is needed,
because the product of rearrangement alone, but not the educt,
can be analyzed quantitatively by NMR spectroscopy.
The results of the kinetic studies in Table 1 are based on
experimental data recorded in Tables S1 and S2 of the
Supporting Information. Specific rate constants are calculated
point by point by the usual equations for reversible first-order
reactions, and extrapolated to zero time by linear regression to
afford the specific rate constants reported in the tables. In these
calculations, the concentration of product at equilibrium is taken
as the square root of the equilibrium isotope effect reported by
Conjugative Interaction
Understanding how perturbations in general affect rates of
reaction requires not only analysis of their effect on the transition
region relative to an unperturbed reference, but equally an
understanding of their effect on the educt. As obvious as this
assertion may be, it has not been applied by previous workers
in the field. For the present purpose, as well as for an analysis
of 2,5-dicyano-3-methylhexa-1,5-diene,3 conjugative interactions
of the cyano group with a double bond in â- and R,â-alkyl-
substituted olefins are missing.
Conjugation by cyano has a long history without having been
brought to a satisfactory conclusion. Early determinations at
single temperatures by Bruylants, and by Linstead, have revealed
a greater effectiveness of cyano vis-a`-vis methyl of ∼1 kcal
mol-1 in ∆∆G,12 while a more recent result gives a value of
∼0.4 kcal mol-1 13
.
New data bearing on three cyano olefins
(12) For succinct reviews, see: Bennett, G. M. Ann. Rep. 1929, 26, 116-
120. Kon, G. A. R. Ann. Rep. 1932, 29, 136-144.

10970 J. Am. Chem. Soc., Vol. 121, No. 47, 1999
Doering and Wang
Table 2. Specific Rate Constants and Derived Differences in
Activation Parameters for the Reversible Interconversions among
4,4- and 6,6-D2-1,3-dicyanohexa-2,5-diene (H) and 4,4- and
6,6-D2-1,3-dicyano-5-phenylhexa-2,5-diene (C6H5) in Isopropyl
Alcohol-d8 (90%) and Pyridine-d5 (10%) Together in the Same
Reaction Vessel
Scheme 5
(k1 + k-1
)
(k1 + k-1)
t, s-1
(C6H5)a,b (H)a,b
(C6H5)c
(H)c
(C6H5)/(H)d
164.8 °C; 438.0 ( 0.2 K
0
5.52
22.16
32.65
38.46
6.23
50940
104700
172980
12.70
18.33
23.58
10.000
10.048
9.025
3.373
3.342
3.183
2.964
3.006
2.836
193.8 °C; 467.0 ( 0.2 K
0
7020
15120
19860
4.52
23.49
34.19
38.61
5.93
14.28
20.82
24.72
83.71
78.54
80.66
32.19
29.59
30.58
2.600
2.654
2.638
a Concentration of 6,6-dideuterio isomers relative to sum in %. b In
units of 10-6 s-1 c Calculated by linear regression of the standard
.
expression for reversible first-order reactions: (k1 + k-1) ) ln[(xe -
xo)/(xe - xt)]/t, where xe ) 47.21% 6,6-dideuterio isomers. d Average
values of the ratio are used to calculate ∆∆fH ) 1.6 ( 0.6 kcal mol-1
2.937 ( 0.089 (438.0 K) and 2.630 ( 0.027 (467.0 K).
:
have now been acquired by the van’t Hoff method from the
temperature dependence of equilibrium constants, and are
reported in Tables S4, S5, and S6 of the Supporting Information.
Calculation therefrom of thermochemical quantities proceeds
in two stages. First, the natural logarithm of the equilibrium
concentrations in percent and the reciprocal of temperatures (K)
are subjected to linear regression. Second, the resulting intercepts
and slopes, being related to entropy and enthalpy, respectively,
are multiplied by the gas constant, R. Relative to an unconju-
gated isomer taken as reference and set to 0.00, these are then
reported as ∆∆fS (cal mol-1 K-1) and ∆∆fH (kcal mol-1),
respectively, along with the sum of the standard errors pertaining
to the isomeric and reference compounds. Included in Scheme
5 are differences in free energies at room temperature (by
extrapolation), enthalpies and entropies of conjugation relative
to one of the unconjugated isomers (0.00),14,15 and heats of
formation and hydrogenation relative to the reference, uncon-
jugated isomers, for which values are estimated in relatively
direct and uncomplicated fashion.16,17
respect to recoveries and freedom from side reactions. Of interest
is the small difference (-0.16 ( 0.16 kcal mol-1) in ∆fH
favoring (E)-â-ethylacrylonitrile, 1∆1(E), over its (Z) isomer,
1∆1(Z). The contrast with phenyl, where the comparable
favoring is -2.5 kcal mol-1, points to a very much smaller steric
factor operating in the nitriles. The discrepancy between these
results and those reported for two simple pairs determined by
combustion warrants note. From that work, the (Z) isomers
appear more stable than the (E) by ∼1.3 ( 0.3 kcal mol-1 (in
the gas phase).18 Although some work slightly favors (Z) over
(E) isomers, while other work slightly favors the reverse, all
previous investigations based on equilibration agree that the
difference, whether in ∆∆G or ∆∆H, in the gas phase19 or
solution,20 is very small. We advise caution in accepting the
combustion data,18 which have been incorporated into a revision
of Benson group additivity parameters of nitriles.21
The system of 2-methylpentenenitriles comprising Series 2
included compounds 2∆2(E) and 2∆2(Z), they being of a type
needed to evaluate conjugative interaction in the two cyano
compounds under study here, and also 2∆1, the system present
in 2,5-dicyanohexa-1,5-diene (vide infra).3 As catalyst, DBN
was not as effective as it had been in Series 1, while stronger
bases led to severe decomposition. The ruthenium catalyst of
Osborn and Wilkinson for isomerizing olefins proved effective
and allowed measurements in benzene-d6 solution (101-191
°C).22 A disadvantage inherent in Series 2 lay in the low
concentrations of 2∆1 and 2∆3(E), the reference unconjugated
isomer, at equilibrium.
The system of pentenenitriles in Series 1 has been studied in
benzene-d6 over the temperature range 61-154 °C, using 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBN) (∼5 mol %) as the basic
catalyst. The system behaves relatively well, particularly with
(13) Procha´zka, M.; Zelinka, J.; Vilim, A.; Cˇ erny, J. V. Collect. Czech.
Chem. Commun. 1970, 35, 1224-1234.
(14) The unconjugated isomers provide an internal check on accuracy
when compared with “generic” values collected in Figure 1 of ref 15. In
Series 1, the difference between 1∆2(E) and 1∆2(Z) is +1.13 kcal mol-1
compared to the “generic” value of +1.04 kcal mol-1; in Series 3, the
difference between 3∆3(E) and 3∆4 is +1.43 kcal mol-1 compared to the
“generic” value of +1.52 kcal mol-1
.
(15) Doering, W. v. E.; Benkhoff, J.; Carleton, P. S.; Pagnotta, M. J.
Am. Chem. Soc. 1997, 119, 10947-10955.
(16) The change in enthalpy of formation on replacing 1° hydron by
cyano is +32.69 ( 0.24 kcal mol-1 from six heats of formation of 1° nitriles
and corresponding saturated hydrocarbons.17 [A similar value is obtained
from two allylcyanides, 3-cyanopropene and (E)-cyanobut-2-ene: +32.84
( 0.1 kcal mol-1.] Application of this increment to n-butane affords
1-cyanobutane, ∆fH ) +2.67 kcal mol-1. From 2-cyanopropane and
cyanocyclohexane, the only 2° nitriles in Pedley et al.,17 a hydron-cyano
replacement value of + 30.62 kcal mol-1 is derived. Application of this
value to pentane and 2-methylpentane generates 2-cyanopentane, ∆fH )
In Series 3, neither DBN nor the ruthenium catalyst were
effective in establishing the equilibria. A much stronger base,
potassium tert-butoxide in benzene-d6, was employed success-
(18) Konicˇek, J.; Procha´zka, M.; Krˇestanova´, V.; Smisˇek, M. Collect.
Czech. Chem. Commun. 1969, 34, 2249-2257.
(19) Butler, J. N.; McAlpine, R. D. Can. J. Chem. 1963, 41, 2487-
2491.
(20) Crump, J. J. Org. Chem. 1962, 28, 953-956.
(21) Chu, J. Y.; Nguyen, T. T.; King, K. D. J. Phys. Chem. 1982, 86,
443-447.
-4.49 kcal mol-1, and 2-cyano-4-methylpentane, ∆fH ) -11.16 kcal mol-1
,
respectively. The corresponding heats of hydrogenation are calculated using
the “generic” values, -27.25 (Series 1 and 2) and -26.38 (Series 3).15
(17) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of
Organic Compounds, 2nd ed.; Chapman and Hall: London, 1986.

1,3-Dicyano-5-phenyl-4,4-d2-hexa-2,5-diene
J. Am. Chem. Soc., Vol. 121, No. 47, 1999 10971
Table 3. Temperature Dependence of Equilibrium between 1,3-
Dicyano-5-phenylhexa-2,5-dienes and 1,3-Dicyano-5-phenylhexa-
1,5-dienes in Isopropyl Alcohol-d8 (80%) and Pyridine-d5 (20%)
T, °C
2,5-hexadienesa
1,5-hexadienesa
ln Kb
24
44
64
84
97.58
96.96
96.07
95.16
2.42
3.04
3.93
4.84
-3.697
-3.462
-3.196
-2.979
a Concentration in % at equilibrium. b Linear regression of lnK versus
1/T(K) affords -∆H/R + ∆S/R: ∆∆H ) +2.55 ( 0.1 kcal mol-1
∆∆S ) +1.2 ( 0.3 cal mol-1 K-1
;
.
2,5-dicyanohexa-1,5-diene (vide infra). That its conjugative
interaction of 2.6 kcal mol-1 is 0.7 kcal mol-1 lower should
perhaps be accepted with some skepticism. Would not a
similarly low value for the conjugative interaction have been
expected of the two nitriles of the type 2∆2(E) and 2∆2(Z)
(bottom level of Figure 1)? With recovery being less than 100%
in Series 2, compound 2∆1 may have been the most likely to
have suffered. A case can be made to agree, -3.3 ( 0.2 kcal
mol-1, for all types of nitriles.
The critically important difference in enthalpy of conjugation
between Cope-incompetent and Cope-competent isomers (e.g.,
(Z)-2,5- and (E)-1,5- of Scheme 3) can be estimated from these
data to be ∆∆H ) 2.81 kcal mol-1 by the thermodynamic cycle
shown in the middle of Scheme 4. [Note: 0.0 kcal mol-1 is
assumed as a reasonable difference between 2∆3(E) and 1∆2-
(E).] By good fortune, it has been possible to measure the
temperature dependence directly by NMR (Table 3), despite
the low concentration of the Cope-competent isomers present
at equilibrium (2-4%). The resulting value, ∆∆H ) 2.55 (
0.1 kcal mol-1, is in gratifyingly good agreement. Both values
accord with the generic value of 2.67 kcal mol-1 for the
difference in enthalpies of conjugation of R-substituted olefins
and (E)-R,â-disubstituted olefins (Figure 1), and thus once again
point to a negligible steric effect of cyano, quite comparable to
that of hydron.
Figure 1. Enthalpies of conjugation (kcal mol-1) are shown relative
to the replacement of circled hydrogen atoms by phenyl, cyano, and
methyl. These and the other values shown also correspond to differences
in enthalpies of hydrogenation of the double bond.
fully (98-160 °C), but formation of byproducts and the failure
of perhaps the most vulnerable of the isomers, 3∆1, to appear
for analysis were disturbing. With some hesitation, the results
are included in Scheme 5, but are not incorporated in the
construction of Figure 1.
Cyano-Stabilized Radicals
Conjugative interactions are defined as discrepancies in ∆∆fH
between the substituted olefinic compound and either the
corresponding unsubstituted olefin (conjugative interaction (CI)
relative to hydron (CIX/H) or the corresponding alkyl-substituted
congener (CIX/R). Translation from values of ∆∆fH in Scheme
5 to values of CIX/H in Figure 1 may involve the conjugative
interactions of methyl vis-a`-vis hydrogen. For example, that the
difference of -0.70 kcal mol-1 between 1∆1(E) and 1∆2(E)
equals CIX/R is based directly on the assumption that NCCH2-
The other information required about the cyano group
concerns the magnitude of its interactions with simple aliphatic
free radicals, and with the active position of an allyl radical.
An extensive literature on the stabilization energy of the
cyanomethyl radical goes back almost two-thirds of a century,
and still may be without definitive ((1 kcal mol-1) resolution.
Values range from ca. -5 kcal mol-1 vis-a`-vis methyl to ca.
-14 kcal mol-1 vis-a`-vis hydron.23,24 Confusion has surrounded
the heat of formation even of acetonitrile. Its value of 15.4 kcal
mol-1, accepted by Pedley et al.,17 has been revised drastically
to 17.8 kcal mol-1 by An and Månsson,25 and Barnes and Pilcher
in 1983,26 although this change had already been forewarned
in 1975 by King and Goddard,27a and recognized without
discussion by Lias et al.28 Significant revisions to the basic
enthalpies of formation of 2° and 3° aliphatic radicals have also
appeared, largely owing to the efforts of the late David Gutmann
and co-workers.29
is the equivalent of an alkyl group. To generate a value for CIX/H
,
2.67 kcal mol-1, the generic value for replacement of H by (E)-
R, is added. The resulting conjugative interaction is 3.4 kcal
mol-1
.
Attention is directed in Figure 1 to the failure of a single
quantity to represent the enthalpy of conjugative interaction of
either phenyl or methyl, the values being sensitively dependent
on R- and â-alkyl substitution. By contrast, the differences
among (E)- and (Z)-R,â-unsaturated nitriles are negligible within
experimental uncertainty (with one exception), and reflect a far
smaller susceptibility of nitrile to steric factors. At present levels
of accuracy, -3.3 ( 0.2 kcal mol-1 seems an acceptable value
for the conjugative interaction of cyano in â- or R,â-substituted
acrylonitriles. The single observation on 2-cyanopent-1-ene
(2∆1) in Series 2 constitutes the sole basis for an analysis of
Reevaluation of the stabilization energy afforded by cyano
vis-a`-vis hydron to a secondary alkyl radical is based without
reservation on the bond dissociation energy, DH°[(CH3)2C-
(23) Doering, W. v. E.; Horowitz, G.; Sachdev, K. Tetrahedron 1977,
33, 273-283 and references therein.
(24) Mayer, P. M.; Glukhovtsev, M. N.; Gauld, J. W.; Radom, L. J.
Am. Chem. Soc. 1997, 119, 12889-12895 and references therein.
(25) An, X.-W.; Månsson, M. J. Chem. Thermodyn. 1983, 15, 287-
293.
(22) (a) Wilson, S. T.; Osborn, J. A. J. Am. Chem. Soc. 1971, 93, 3068-
3070. (b) Bradley, J. S.; Wilkinson, G. Inorg. Synth. 1977, 17, 73-74.
(26) Barnes, D. S.; Pilcher, G. as an Appendix to An and Månsson.25

10972 J. Am. Chem. Soc., Vol. 121, No. 47, 1999
Doering and Wang
Table 4. Specific Rate Constants and Derived Activation
Scheme 6
Parameters for the Reversible Rearrangement of (E)- and
(Z)-3-Cyanocyclohex-2-enylidene, (E)-3 and (Z)-3, in Benzene-d6
T, °C
106k1, s-1 a
154.0 ( 0.2b
154.0 ( 0.2c
165.2 ( 0.1b
165.2 ( 0.1c
186.1 ( 0.1b
186.1 ( 0.1c
193.9 ( 0.1b
193.9 ( 0.1c
1.31 ( 0.01
1.25 ( 0.03
4.02 ( 0.08
3.82 ( 0.06
28.5 ( 0.4
27.7 ( 0.5
55.9 ( 0.9
56.0 ( 0.3
(CN)-CH3] ) 74.7 ( 1.6 kcal mol-1, so elegantly determined
by King and Goddard in pyrolyses of alkyl nitriles at very low
pressures. Recast in eq 1 using currently accepted values of
(CH ) C(CN)•
•
(CH ) C(CN)
CH3
f
+
3 3
-0.617
3 2
Arrhenius Parameters
+34.930
+39.2
Ea ) 37.57 ( 0.19 kcal mol-1
log A ) 13.33 ( 0.09
∆Hr ) 74.7 kcal mol-1 31 (1)
Eyring Parametersd
(CH3)2CH•
+
∆Hq ) 36.7 ( 0.2 kcal mol-1
∆Sq ) -0.4 ( 0.4 cal mol-1 K-1
(CH ) CH(CN) (CH ) CHH
+
f
3 2
+5.617
3 2
-25.017
+21.530
a Calculated by linear regression of the standard expression for
reversible first-order reactions: (k1 + k-1) ) ln[(xe - xo)/(xe - xt)]/t.
b Starting with pure (E)-3. c Starting with pure (Z)-3. d Calculated at
174.0 °C.
SECN/H ) -12.9 kcal mol-1 (2)
(CH3)2C(CN)•
+39.2
(CH ) C(CN)•
+
(CH3)3C•
+12.330
(CH ) CH(CN) (CH ) CH
+
f
3 2
+5.617
3 3
-32.117
3 2
+39.2
systems and no apparent hydron-hydron separations within the
sum of their van der Waals radii.
SECN/CH3 ) -10.8 kcal mol-1 (3)
The results of the kinetic experiments are recorded in Table
4, the original experimental data being available in Table S3 of
the Supporting Information. The enthalpy of activation, ∆Hq
) 36.7 ( 0.2 kcal mol-1, may be compared with values for
three closely related trienes: cyclohex-2-enylidene, gas phase,
enthalpies of formation of tert-butylnitrile and methyl radical
lead to an enthalpy of formation of the dimethylcyano radical
of +39.2 kcal mol-1 31
A corresponding stabilization energy
.
40.9 kcal mol-1 35
;
solution, 38.9 kcal mol-1 36
If a mean of 39.9 kcal
;
and a more
(SE) is defined by the replacement of hydron by cyano and
evaluated by the isodesmic manipulation of eq 2: (SECN/H) )
extended example, 40.0 kcal mol-1 37
.
-12.9 kcal mol-1 32,33
The stabilization energy resulting from
.
mol-1 is taken as the unperturbed value, the two cyano groups
have lowered the enthalpy of activation by -3.2 kcal mol-1, or
-1.6 kcal mol-1 per cyano group. By comparison, two phenyl
groups (the cinnamyl radical) lower the enthalpy of activation
by -5.0 kcal mol-1, or -2.5 kcal mol-1 per phenyl group,38,39
while two additional conjugated double bonds (the pentadienyl
radical)36,37,40 lower the enthalpy by -7.8 kcal mol-1, or -3.9
kcal mol-1 per double bond. By rights, differences in conjugative
interactions of a double bond with the groups, cyanovinyl, styryl,
and butadienyl, should be taken into account, but at present that
is not possible.
replacement of a methyl group by cyano is derived by eq 3:
SECN/CH3 ) -10.8 kcal mol-1. Both are based on the same
enthalpy of formation of the dimethylcyano radical.
The enthalpy of stabilization afforded to allyl by two cyano
groups in the 1,3-dicyanoallyl radical has not been determined.
The effect of a single cyano, however, has now been established.
Using the device of thermal cis-trans isomerization about a
double bond, the kinetics of the thermal approach to equilibrium
of (E)- and (Z)-bi-3-cyanocyclohex-2-enylidenes has been
pursued (Scheme 6). An unexceptional preparation leads to a
mixture of (Z) and (E) isomers, the structures of which have
been assigned by X-ray crystallography.34 The 1H NMR spectra
of (E)-3 and (Z)-3, although almost identical in benzene-d6,
differ sufficiently in a mixture of benzene-d6 and pyridine-d5
(10:1) to allow distinction between the two isomers on the basis
of the chemical shifts of their 2-vinyl hydrons (6.67 and 6.77
ppm), respectively. Both isomers have essentially coplanar triene
Cyano in a Chameleonic Cope Rearrangement
The chameleonic Cope has been defined in terms of a
transition region having C2h symmetry, that is, having two
identical halves.1 Although the effect of two cyano groups in
active positions, as in 1,4-dicyanohexa-1,5-diene, has not been
examined, the effect of two in the nodal position has been
determined by Wehrli, Schmid, Bellusˇ, and Hansen.3 Their
example, 2,5-dicyano-3-methylhexa-1,5-diene, is somewhat
flawed by the use of methyl as the degeneracy-breaking element.
They reported the activation parameters ∆Hq ) 23.2 kcal mol-1
and ∆Sq ) -14.2 cal mol-1 K-1. The lowering in enthalpy of
(27) (a) King, K. D.; Goddard, R. D. Int. J. Chem. Kinet. 1975, 7, 837-
855. (b) King, K. D.; Goddard, R. D. J. Phys. Chem. 1978, 82, 1675-
1679.
(28) Lias, S. G.; Bartmess, J. E.; Liebmann, J. F.; Holmes, J. L.; Levin,
R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. 1, 1-861
(p 84).
(29) Seetula, J. A.; Russell, J. J.; Gutman, D. J. Am. Chem. Soc. 1990,
112, 1347-1353.
(35) Doering, W. v. E.; Roth, W. R.; Bauer, F.; Breuckmann, R.; Figge,
L.; Lennartz, H.-W.; Fessner, W.-D.; Prinzbach, H. Chem. Ber. 1988, 121,
1-9.
(30) Seakins, P. W.; Pilling, M. J.; Niiranen, J. T.; Krasnoperov, L. N.
J. Phys. Chem. 1992, 96, 9847-9855.
(31) King, K. D.; Goddard, R. D. J. Phys. Chem. 1976, 80, 546-552.
(36) Doering, W. v. E.; Kitagawa, T. J. Am. Chem. Soc. 1991, 113,
4288-4297.
(32) Similar recalculations of the heats of formation of H2C(CN)• [+55.6
kcal mol-1 31
,
+57.8 kcal mol-1
]
33 and CH3CH(CN)• [+49.5 kcal mol-1
]
(37) Doering, W. v. E.; Shi, Y.-q.; Zhao, D.-c. J. Am. Chem. Soc. 1992,
114, 10763-10766.
31
lead to stabilization energies of -15.0 and -12.7, and -11.8 kcal mol-1
respectively. A value calculated for H2C(CN)• at the G2 level of theory is
SECN/H ) -6.9 kcal mol-1 24
(33) Trenwith, A. B. J. Chem. Soc., Faraday Trans. 1 1983, 79. 2755-
2764.
,
(38) Doering, W. v. E.; Benkhoff, J.; Shao, L. J. Am. Chem. Soc. 1999,
121, 962-968.
.
(39) Doering, W. v. E.; Birladeanu, L.; Sarma, K.; Teles, J. H.; Kla¨rner,
F.-G.; Gehrke, J.-S. J. Am. Chem. Soc. 1994, 116, 4289-4297.
(40) Doering, W. v. E.; Birladeanu, L.; Cheng, X-h.; Kitagawa, T.; Sarma,
K. J. Am. Chem. Soc. 1991, 113, 4558-4563.
(34) Wang, Y.-h.; Doering, W. v. E.; Staples, R. J. J. Chem. Crystallogr.
1999, 29, submitted.

1,3-Dicyano-5-phenyl-4,4-d2-hexa-2,5-diene
J. Am. Chem. Soc., Vol. 121, No. 47, 1999 10973
activation vis-a`-vis the archetype, hexa-1,5-diene (∆Hq ) 33.5
kcal mol-1; ∆Sq ) -13.8 cal mol-1 K-1),41 is impressively
large. It has been reproduced almost exactly in very recent
theoretical calculations by Beno, Hrovat, Lange, Yoo, Houk,
and Borden (∆Hq ) 23.9 kcal mol-1).5b Are these experimental
results consistent with the conjecture that the transition region
has behaved as if it were approximated by a cyclohexa-1,4-
diyl diradical?
The necessary correction for conjugative interaction in the
educt can now be made on the basis of the van’t Hoff studies
reported above. If conjugation operating in an R-substituted
acrylonitrile is taken to be -2.6 kcal mol-1, the model for the
educt is the archetype corrected by twice this amount: 33.5 +
(2 × 2.6) ) 38.7 kcal mol-1. Recall, however, the reservation
that -3.3 kcal mol-1 might be the preferred value for cyano
regardless of the degree of alkyl substitution (40.1 kcal mol-1).
In neglect of an energetic disturbance by the methyl group,
stabilization of the transition region thus amounts to -15.5 kcal
mol-1, or -7.8 kcal mol-1 per cyano group, a value only slightly
less than the -8.7 kcal mol-1 per phenyl group deduced in
similar manner from the rearrangement of 2,5-diphenylhexa-
1,5-diene.9,42
Since the transition region seems to have been moved toward
the cyclohexa-1,4-diyl extreme by the demands of the two
perturbations in “nodal” positions, a more appropriate model
for the unstabilized transition region might be the diradical,
corrected similarly for conjugative interaction in the educt. Two
estimates of the enthalpy of formation of a thermochemical
model for the cyclohexa-1,5-diyl diradical have been made: one
based on the nonconcerted cleavage of cyclobutane to two
molecules of ethylene, the other by calculation of the cost of
removing two hydrons from cyclohexane. Both agree on a value
close to 43.3 kcal mol-1 above that of hexa-1,5-diene.43,44 Vis-
a`-vis this model, corrected by 2 × -2.6 kcal mol-1, the
estimated stabilization in the transition region is -25.3 kcal
mol-1, or -12.7 kcal mol-1 per cyano group. This value
compares well with that of -12.9 kcal mol-1 estimated above
as the stabilization enthalpy afforded a secondary alkyl radical
by the replacement of hydron with cyano.
Centauric Cope Rearrangement
Figure 2. Activation parameters for several Cope rearrangements: ∆Hq
in kcal mol-1 and ∆Sq in cal mol-1 K-1. Differences in enthalpies of
activation relative to the archetype A are collected in column, “uncorr.
∆Hq”. Corrections for conjugative interaction in the educt are listed in
“conj. corr”, and applied to produce effective enthalpies of stabilization
in the transition region as ”corr. ∆Hq”. Four differences between pairs
of allyl and 2-phenyl allyl types are interpolated.
Thermochemical analysis of the effect of replacing the
5-hydron in 1,3-dicyanohexa-1,5-diene [1(H)] by a phenyl group
at the C-5 nodal position [1(Ph)] proceeds stepwise. The
empirical activation parameters for the transposition of deute-
rium in the Cope-incompetent hexa-2,5-dienes are given in Table
1 and Scheme 4. These are converted into activation parameters
of the Cope-competent hexa-1,5-dienes by subtraction of the
difference in enthalpies and entropies of formation of the two
types. For both, the same mean correction is taken, ∆∆H° )
+2.7 kcal mol-1 and ∆∆S° ) +1.6 eu, as summarized above
and in Scheme 4. The resulting parameters for the Cope
rearrangements of 1(H) (J) and 1(Ph) (K) are given in Figure
2.
that of -3.5 kcal mol-1 calculated by Beno et al.5b The
comparable lowering in 1,3-diphenylhexa-1,5-diene, G, is -3.0
kcal mol-1. Recall that the stabilization in an allyl radical
provided by a single active cyano group is -1.6 kcal mol-1
,
while that provided by a single phenyl is -2.5 kcal mol-1
.
The results of theoretical calculations are generally reported
directly as ∆∆Hq, the differences in calculated enthalpies of
activation compared to that calculated for an unperturbed
archetype, and are to be compared to the experimental values,
“uncorr. ∆∆Hq” of Figure 2, without correction for conjugative
interaction. In the classical type of analysis undertaken here,
the question is addressed directly, What effect have the
perturbations had on the transition region? For this purpose,
obligatory corrections are made for conjugative interaction in
the educts. These are listed in the column “conj. corr.” of Figure
2, and are added to “uncorr. ∆∆Hq” to generate “corr. ∆∆Hq”,
The enthalpy of activation, ∆Hq, of J is lower than that of
archetype A by -4.7 kcal mol-1. This value is not far from
(41) Doering, W. v. E.; Toscano, V. G.; Beasley, G. H. Tetrahedron
1971, 27, 5299-5306.
(42) Roth, W. R.; Lennartz, H.-W.; Doering, W. v. E.; Birladeanu, L.;
Guyton, C. A.; Kitagawa, T. J. Am. Chem. Soc. 1990, 112, 1722-1732.
(43) This value emerges from consideration of the ethene-cyclobutane
system,44 and application of the 1992 value for the heat of formation of the
isopropyl radical.30
(44) Doering, W. v. E. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 5279-
5283.

10974 J. Am. Chem. Soc., Vol. 121, No. 47, 1999
Doering and Wang
which are the “experimental” enthalpies of stabilization pro-
duced in the transition region by the perturbation, relative to
archetype A as model. For example, the corrections made for
conjugative interactions in 1(Ph) (K) are -3.3 and -2.6 kcal
mol-1 for cyano and phenyl, respectively, taken from Figure 1.
The result, 39.4 kcal mol-1, is the enthalpy of activation
expected if the perturbations were to have affected only the
educt. The difference between that value and the experimental
value (26.8 kcal mol-1) measures the stabilization afforded to
the transition region (-12.6 kcal mol-1). Note the obvious that
stabilization enthalpies of J and G are substantially larger than
they appear to be before correction for conjugative interaction
in the educts.
tions in the educt, independently evaluated radical-stabilizing
abilities of the individual perturbations, and an incompletely
realized, centauric model of the transition region.
Experimental Section
General Procedures. 1H NMR and 13C NMR (125.8 MHz) spectra
are obtained in the noted solvents on a Bruker AM-500 instrument
(500 MHz). Spin-lattice relaxation times (T1) are determined by the
inversion-recovery method with use of vacuum-sealed solutions in
benzene-d6. Chemical shifts are reported in ppm (δ) with respect to
C6D5H; coupling constants, J, are reported in Hz. Analytical GC are
effected on a Hewlett-Packard 5890A Gas Chromatograph with J &
W Scientific column DB 225, 30 m × 0.5 m. Preparative GC is effected
on a Varian Aerograph Model 90-P: column CW 20M with 20 psi
He, flow-rate 6 L/h, detector 200 °C, injector 175 °C. High-resolution
mass spectra (HR-MS) are measured on a JEOL AX 505 spectrometer
equipped with a data-recovery system and reported as m/z (density as
% of major peak). In both the kinetic and equilibrating experiments,
the procedure is the same: to an NMR tube that has been oven- and
vacuum-dried, evacuated for 10 min, and filled with argon, there are
added the reactants. After three freeze-thaw cycles, the tube is sealed
under vacuum and placed in the vapors of compounds of appropriate
boiling point, boiling under reflux. Temperature is monitored periodi-
cally by a thermocouple, as is barometric pressure. The compounds
employed are redistilled before use: 2-(2-methoxyethoxy)ethanol (194
°C), 4-tert-butyltoluene (191 °C), diethyl oxalate (185 °C), 4-methyl-
anisole (176 °C), mesitylene (165 °C), cyclohexanol (160 °C), anisole
(154 °C), cumene (152 °C), nonane (151 °C), o-xylene (144 °C),
ethylbenzene (136 °C), 1,1,2,2-tetrachloroethane (121 °C), toluene (111
°C), dioxan (101 °C), propanol-1 (97 °C), 3-methylhexane (92 °C),
benzene (80 °C), chloroform (61 °C), and methylene chloride (40 °C).
The tubes are periodically withdrawn from the bath for analysis,
immediately cooled in a dry ice/isopropyl alcohol bath, centrifuged,
analyzed, and returned for heating until no further change in relative
concentrations of isomeric nitriles is noted. Where analysis is by NMR,
normal parameters are used with recovery and relaxation delay times
as noted. Each signal is integrated three times. Details of unexceptional
preparative procedures are reported in the Supportimg Information.
1,3-Dicyanohexa-2,5-diene (1). In a 100-mL, round-bottomed flask,
0.36 g (0.012 mol) of sodium hydride (80% oil dispersion) was washed
twice with pentane. Freshly distilled DMSO (30 mL) was added with
stirring. The resulting suspension was stirred at room temperature for
10 min, then at 70 °C for 1 h, cooled, and treated dropwise with 1.13
g (0.012 mol) of 1,3-dicyanopropene45 in 3 mL of DMSO. After 30
min, this solution was added dropwise to 1.04 mL (0.012 mol) of allyl
bromide and 5 mL of DMSO in a 100-mL, round-bottomed flask. After
1 h, the reaction mixture was quenched by saturated aqueous NH4Cl.
Standard workup followed by preparative column chromatography
(CHCl3 as eluting solvent) gave 0.35 g (22%) of (Z)- and (E)-1 (ratio
4:1). (Z)-1: 1H NMR 6.16 (t, J ) 7.2, 1H), 5.76 (m, 1H), 5.24 (dd, J
) 18.2, 17.0, 10.2, 2H), 3.45 (d, J ) 7.2, 2H), 3.03 (d, J ) 6.5, 2H).
A point of interest concerns the magnitude of the stabilization
afforded by the replacement of allyl by 2-phenylallyl in the pairs,
A and B, D and E, G and H, and J and K. These are shown in
Figure 2 as interpolated values of ∆∆Hq in columns uncorr.
∆∆Hq and corr. ∆∆Hq. Whether a trend is discernible as the
demand for allylic character increases in the arranged order or
whether one should be satisfied to accept the mean of the
corrected values (-5.7 ( 1.0 kcal mol-1) is moot. The nodal
phenyl group in K has lowered the uncorrected enthalpy of
activation, ∆Hq, by -2.0 kcal mol-1 vis-a`-vis J. Although
phenyl may contain too many carbon atoms to allow reasonably
accurate calculations at the present time, the calculated lowering
by a C-5 cyano group, as in 1,3,5-tricyanohexa-1,5-diene vis-
a`-vis 1,3-dicyanohexa-1,5-diene, is -0.6 kcal mol-1 5b
The
.
lacuna is partly filled by calculations on vinylhexa-1,5-dienes,
reasonable surrogates for the phenyl analogues. The calculated
difference between 1,3,5-trivinyl- and 1,3-divinylhexa-1,5-
diene5b is +0.5 kcal mol-1. [Agreement between experiment
and calculation is much improved by correcting for an estimated
enthalpy of conjugation in the educt of -5.2 kcal mol-1.]
In terms of the chameleonic model as defined, either the
actively positioned 1,3-dicyano group or the nodally positioned
phenyl group could be accommodated, but not both. The
expected lowering would be either -8.0 kcal mol-1 (J) or -6.8
kcal mol-1 (B) [-8.7 kcal mol-1 (C)], substantially less than
the experimental value of -12.6 kcal mol-1. Here, as in the
all-phenyl case,1 the chameleonic model leads to a serious
underestimation.
In the centauric model, the 1,3-dicyanoallyl group (J),
optimally served by C-1 and C-3 being sp2 planar (-8.0 kcal
mol-1), and the phenyl group (B), optimally served by C-5 being
sp2 planar and C-4 and C-6 sp3 tetrahedral (-6.8 kcal mol-1),
might contribute a maximum stabilization energy of -14.8 kcal
mol-1 [-16.7 kcal mol-1 (C)]. The experimental value of -12.6
kcal mol-1 falls between but closer to the centauric model.
Although the full stabilizing potential of each group operating
independently seems not to have been expressed, the centauric
is the preferred model, as it is in the example of 1,3,5-
triphenylhexa-1,5-diene.1
1
(E)-1: H NMR 6.37 (t, J ) 7.2, 1H), 5.76 (m, 1H), 5.24 (dd, J )
18.2, 17.0, 10.2, 2H), 3.27 (d, J ) 7.1, 2H), 3.03 (d, J ) 6.5, 2H). MS
(CI) 282 (2M + NH4+, 53), 150 (M + NH4+, 100).
1,3-Dicyano-4,4-d2-hexa-2,5-diene (2,5(H)d2). 1,1-D2-Allyl alcohol,
prepared regiospecifically from acryloyl chloride by reduction with
LiAlD4,46 was converted to 3,3-d2-3-tosyloxypropene according to the
procedure for 3-tosyloxypropene:47 1H NMR 7.8 (d, J ) 8.3, 2H), 7.34
(d, J ) 8.3, 2H), 5.81 (dd, J ) 17.2, 10.2, 1H), 5.28 (dd, J ) 32.8,
17.2, 10.2, 2H), 2.44 (s, 3H). Following the procedure above,
1,3-dicyanopropene and 3,3-d2-3-tosyloxypropene were converted in
33% yield of theory to a mixture consisting of (Z)-2,5(H)d2 and (E)-
2,5(H)d2 (∼94%, ratio 4:1) and 1,3-dicyano-6,6-d2-hexa-2,5-diene
(∼6%). (Z)-1,3-Dicyano-4,4-d2-hexa-2,5-diene: 1H NMR 6.16 (t, J )
7.2, 1H), 5.76 (dd, J ) 17.0, 10.2, 1H), 5.24 (dd, J ) 18.2, 17.0, 10.2,
Theory may provide a “calculational” answer to the possibly
nonoperational, geometrical question whether the transition
region retains a basic C2h symmetry in the six-atom cycle of
the chameleon type, but still allows sufficient, if partial,
stabilization by conflicting substituents so that the sum exceeds
that of either extreme working alone. One accepts that such
questions based on current and past mechanistic concepts may
become redundant in the coming world of calculational om-
nipotence. Meanwhile, the task of predicting the effects of
perturbation by radical-stabilizing groups in Cope’s rearrange-
ment, and some other thermal rearrangements as well, may be
served by a conceptual scheme combining conjugative interac-
(45) Boate, D. R.; Hunter, D. H. Org. Magn. Reson. 1984, 22, 167-
170.
(46) Bowen, R. D.; Wright, A. D.; Derrick, P. J. J. Chem. Soc., Perkin
Trans. 2 1993, 501-507.
(47) Johnson, C.; Dutra, G. J. Am. Chem. Soc. 1973, 95, 7777-7782.

1,3-Dicyano-5-phenyl-4,4-d2-hexa-2,5-diene
J. Am. Chem. Soc., Vol. 121, No. 47, 1999 10975
2H), 3.45 (d, J ) 7.2, 2H). (E)-1,3-Dicyano-4,4-d2-hexa-2,5-diene: 1H
NMR 6.37 (t, J ) 7.2, 1H), 5.76 (dd, J ) 17.0, 10.2, 1H), 5.24 (dd, J
) 18.2, 17.0, 10.2, 2H), 3.27 (d, J ) 7.1, 2H).
Nemours & Co.) was equilibrated using 3 µL (0.024 mmol) of
diazobicyclononene (DBN) as catalyst. A relaxation delay time of 175
s was employed in analysis. The data from three sets of experiments
are collected in Table S4 of the Supporting Information, and combined
for calculation.
1,3-Dicyano-4,4-d2-5-phenylhexa-2,5-diene (2,5(Ph)d2). Following
the procedure for the preparation of 1,1-d2-2-methylallyl alcohol, ethyl
R-phenylacrylate,48 prepared in two steps from ethyl phenylacetate, is
reduced to 1,1-d2-2-phenylallyl alcohol: 1H NMR 7.26 (d, J ) 7.6,
2H), 7.17 (t, J ) 7.4, 2H), 7.12 (t, J ) 7.6, 1H), 5.23 (d, J ) 63.6,
2H), 1.60 (s, 1H). Following the procedure above for 3,3-d2-3-
tosyloxypropene, the corresponding tosylate, 2-phenyl-3,3-d2-3-tosyl-
oxypropene, was prepared: 1H NMR 7.73 (d, J ) 8.3, 2H), 7.28 (d, J
) 8.1, 2H), 7.25 (m, 5H), 5.43 (d, J ) 95.9, 2H), 2.42 (s, 3H). Reaction
with 1,3-dicyanopropene in the procedure developed above for 1,3-
dicyano-4,4-d2-hexa-2,5-diene afforded, in 36% of theory, a mixture
consisting of ∼95% of (Z)- and (E)-2,5(Ph)d2 (ratio 4:1) and ∼5% of
1,3-dicyano-6,6-d2-5-phenylhexa-2,5-diene. (Z)-1,3-Dicyano-4,4-d2-5-
phenylhexa-2,5-diene: 1H NMR 7.38-7.30 (m, 5H), 6.14 (t, J ) 7.1,
1H), 5.41 (d, J ) 170.2, 2H), 3.39 (d, J ) 7.1, 2H). (E)-1,3-Dicyano-
4,4-d2-5-phenylhexa-2,5-diene: 1H NMR 7.38-7.30 (m, 5H), 6.14 (t,
J ) 7.2, 1H), 5.37 (d, J ) 140.4, 2H), 3.20 (d, J ) 7.2, 2H).
Kinetics of Rearrangement of 1,3-Dicyano-4,4-d2-hexa-2,5-diene
(2,5(H)d2) and 1,3-Dicyano-4,4-d2-5-phenylhexa-2,5-diene (2,5-
(Ph)d2). A solution of 2,5(H)d2 (0.09 M) in 0.5 mL of degassed
isopropyl alcohol-d8 in a vacuum-sealed NMR tube with 18-crown-6
ether as an internal standard was heated (165-194 °C). Quantitative
analysis by 1H NMR (500 Hz) measured relative concentrations of
methylene hydrons of product at C-4 and vinyl hydrons of both starting
material and product at C-2. Recovery was monitored by comparing
the amount of vinyl hydron at C-2 with that of the hydron of 18-crown-6
ether. Each datum was generated from at least three integrations of the
NMR signals accumulated over eight scans. An acquisition time of
2.77 s and relaxation delay time of 150 s were employed. The data
and details of calculation for rate constants and kinetic parameters are
given in Table S1 of the Supporting Information. Results are sum-
marized in Table 1.
2-Methylpentenenitriles (Series 2, Scheme 5). To 5-10 mg of
hydridonitrosyl tris(triphenylphosphine) ruthenium (preparation in the
Supporting Information) in an NMR tube was added a solution of 0.6
mmol of benzene-d6 and 5-8% percent of 2-methylpentenenitrile(s)
1
with a syringe. The reaction was monitored by H NMR until there
was no further change with time. The contents of the tube were then
transferred by trap-to-trap distillation under vacuum to give a clear
distillate, which was analyzed directly by GC. To ensure the establish-
ment of true equilibrium, the distillate was returned to another NMR
tube, fresh catalyst was added, and the procedure above was repeated.
Generally, one refreshment of catalyst sufficed to establish equilibrium.
At a given temperature, usually two independent runs on different
samples were made. All but two of the six isomers of 2-methylpen-
tenenitrile can be resolved quantitatively by analytical GC. Retention
times (min) of all six isomers at 70 °C follow: 2∆1, 4.6; 2∆2(Z), 5.1;
2∆3(E), 7.2; 2∆3(Z), 7.4; 2∆4, 7.4; 2∆2(E), 9.5. Since the amounts of
2∆3(Z) and 2∆4 at equilibrium are very small (<1%), and could not
be resolved satisfactorily, only 2∆2(Z), 2∆2(E), 2∆3(E), and 2∆1 were
analyzed quantitatively. The response factors of these four isomers are
very nearly identical, so the relative percent of each isomer among all
four isomers is calculated without correction. Generally, measurements
of each sample were repeated five times, the average value being used
in the calculations. The experimental data are found in Table S5.
2,4-Dimethylpentenenitriles (Series 3, Scheme 5). A sample of
t-BuOK, freshly sublimed in a vacuum and stored in a desiccator, was
transferred under nitrogen in an air box to a preweighed, flame-dried
flask flushed with argon. In proportion to the amount of t-BuOK
transferred, a volume of freshly distilled benzene-d6 was added via
syringe calculated to make a 0.16 M solution. Further dilution of 0.5
mL of this solution by 2.5 mL of freshly distilled benzene-d6 afforded
a 0.0032 M solution of t-BuOK, 0.6 mL of which was then transferred
into a vacuum-dried, argon-flushed NMR tube. Following addition of
10 µL of 3∆2(Z) (7.3 mg, 0.66 mmol) (or 3∆2(E), or a mixture of the
two) and 20 µL (0.095 M, 0.0019 mmol) of 18-crown-6 ether, the
initially colorless solution slowly became brown. For details of the
analysis by NMR and experimental data, see Table S6 of the Supporting
Information.
Kinetic experiments with 1,3-dicyano-4,4-d2-5-phenylhexa-2,5-diene
paralleled those above with 1,3-dicyano-4,4-d2-hexa-2,5-diene, but for
the use of isopropyl alcohol-d8 containing 10% of pyridine-d5 as solvent.
Quantitative analysis was based on methylene hydrons of product at
C-4, vinyl-hydrons of starting material at C-6, and the vinyl hydron of
both starting material and product at C-2. An acquisition time of 2.77
s and a relaxation delay time of 83 s were employed. Eight scans were
accumulated for each measurement. Data and details of calculation are
given in Table S2 of the Supporting Information. Results are collected
in Table 2.
Equilibration of 1,3-Dicyano-4,4-d2-5-phenylhexa-2,5-diene and
1,3-Dicyano-4,4-d2-5-phenylhexa-1,5-diene. Temperature dependence
of equilibria among 1,3-dicyano-4,4-d2-5-phenylhexa-2,5-dienes and
-hexa-1,5-dienes was determined directly in the NMR spectrometer in
0.5 mL of 10% pyridine-d5 in isopropyl alcohol-d8 at four tempera-
tures: 24, 44, 64, and 84 ( 1 °C. The ratio of vinyl hydron at C-2 in
1,3-dicyano-4,4-d2-5-phenylhexa-2,5-diene and 1,3-dicyano-4,4-d2-5-
phenyl-1,5-hexadiene was measured, the results being summarized in
Table 3.
The kinetics of rearrangement of a mixture of equal amounts of 1,3-
dicyano-4,4-d2-hexa-2,5-diene and 1,3-dicyano-4,4-d2-5-phenylhexa-
2,5-diene were determined in isopropyl alcohol-d8 containing 10% of
pyridine-d5 at 165 and 194 °C. The results are summarized in Table 3.
Thermal Equilibrations of (E)- and (Z)-3,3′-Dicyanocyclohex-2-
enylidene, (E)-3 and (Z)-3. Studies of the kinetics were effected in
degassed solutions of benzene-d6/pyridine-d5 (10:1) in vacuum-sealed
NMR tubes with samples of either pure (E)-3 or (Z)-3 as starting
material in a concentration of 0.076 M. The course of thermal
Acknowledgment. W.v.E.D. expresses his warm thanks to
Professors Weston T. Borden and Kenneth N. Houk for their
liberal and constructive help, and the sharing of their theoretical
results, and to Professor Frank H. Westheimer for valued
discussions. This work has been supported generously by the
Norman Fund in Organic Chemistry in memory of Ruth Alice
Norman Weil Halsband. Support to the Department of Chem-
istry and Chemical Biology of Harvard University for its NMR
Facility comes from the NIH (S10-RR01748) and (S10-
RR04870) and from the NSF (CHE 88-14019).
1
equilibration was followed by H NMR at four temperatures, 154.0
(anisole), 165.2 (mesitylene), 186.1 (diethyl oxalate), and 193.9 °C (2-
(2-methoxyethoxy)ethanol), using the ratio of integrated areas of the
2-vinyl hydrons of (E)-3 at 6.67 ppm and (Z)-3 at 6.77 ppm. Acquisition
times of 2.77 s and relaxation delay times of 8 s were employed. A
total of 32 scans was accumulated for each measurement. Each datum
was generated from at least three integrations of the NMR signals. The
data from the eight runs are given in Table S3 of the Supporting
Information. Specific rate constants, derived according to the general
expression for a reversible, first-order reaction, and Arrhenius, and
Eyring activation parameters are reported in Table 3.
Supporting Information Available: Two sets of three tables
each giving details of the kinetic data and the experimental data
of the van’t Hoff equilibrations and a second Experimental
Section providing detailed descriptions of the preparation of
materials (PDF). This material is available free of charge via
the Internet at http://www.pubs.acs.org.
Equilibration among Pentenenitriles (Series 1, Scheme 5). Fol-
lowing the preliminary procedure developed by Benkhoff,49 (Z)-pent-
2-enenitrile [1∆2(Z)] or (E)-pent-2-enenitrile [1∆2(E)] (DuPont de
(48) Ksander, G. M.; McMurry, J. E.; Johnson, M. J. Org. Chem. 1977,
42, 1180-1185.
(49) We thank Dr. Johannes Benkhoff for developing this procedure.
JA992137Z