Protonation of β-cyanoenamines: NMR spectroscopic and calculational studies of ketenimine amidinium ion formation

Article abstract of DOI:10.1021/ja983016m

NMR spectroscopy and quantum mechanical calculations have been used to investigate the protonation of β-cyanoenamines. One-dimensional variable- temperature spectroscopy, two-dimensional exchange spectroscopy (2D EXSY), and density functional calculations

Full text of DOI:10.1021/ja983016m

12942
J. Am. Chem. Soc. 1998, 120, 12942-12949
Protonation of â-Cyanoenamines: NMR Spectroscopic and
Calculational Studies of Ketenimine Amidinium Ion Formation
Tammy J. Dwyer,*,† Paul G. Jasien,‡ Scott M. Kirkowski,† Charles M. Buxton,†
Jeannie M. Arruda,† and Gregory W. Mitchell†
Contribution from the Department of Chemistry, UniVersity of San Diego, 5998 Alcala Park,
San Diego, California 92110, and Department of Chemistry, California State UniVersity, San Marcos,
San Marcos, California 92096
ReceiVed August 21, 1998. ReVised Manuscript ReceiVed October 21, 1998
Abstract: NMR spectroscopy and quantum mechanical calculations have been used to investigate the protonation
of â-cyanoenamines. One-dimensional variable-temperature spectroscopy, two-dimensional exchange spec-
troscopy (2D EXSY), and density functional calculations provide convincing evidence that the favored site of
protonation in 1,1-bis(dimethylamino)-2-cyano-2-p-X-phenyl-substituted ethylenes (X ) NO2, CF3, Br, F, or
CH3) is the nitrogen of the cyano group. This results in the formation of a ketenimine amidinium ion in
solution. When X is less electron withdrawing than -NO2, an equilibrium mixture is formed consisting of
60-93% CN-protonated enamine with the remainder protonated at the â-carbon. Calculational results also
support predominantly CN-protonation in these systems. Both experimental and calculational results show no
evidence of protonation at the amino nitrogen.
Introduction
The structure and reactivity of a push-pull enamine are
intimately linked. Both are influenced by the amino group
substituents and the nature of the electron-withdrawing groups.
Structurally, steric interactions between electron donor and
acceptor groups lead to a nonplanar or twisted carbon-carbon
double bond. The degree of nonplanarity directly effects the
extent of conjugation between the lone-pair electrons of the
amino nitrogen and the carbon-carbon double bond. Conse-
quently, delocalization of electron density onto the â-carbon
atom (and the electron acceptor groups) is diminished as is the
reactivity at these sites. As one might expect, the effect is
magnified in 1,1-enediamines in which a twist angle about the
CdC can be as large as 85°.8 These systems tend to show
decreased reactivity relative to the parent enamine.
The electronic structures and reactivities of push-pull
enamines have long been subjects of chemical interest, posing
unique challenges with regard to their study.1-4 A push-pull
enamine is defined by one or more electron donor groups
(typically -NR2 groups) situated on one carbon of a CdC and
one or more acceptor groups (such as -NO2, -CN, or
-C(dO)R) at the other carbon. Such an arrangement leads to
a highly delocalized π electron system and polarization of the
CdC. These systems demonstrate interesting dynamic behavior
as well, such as facile rotations about the CdC and hindered
rotations about the C-N bonds.5-7 The push-pull effect leads
to a substantially lower CdC rotational barrier (10-25 kcal
mol-1) relative to ethylene (65 kcal mol-1) and an increased
C-N barrier (8-17 kcal mol-1) compared to a nonconjugated
amino group. These low CdC barriers and high C-N barriers
have been explained in terms of the capacity of the donor and
acceptor groups to stabilize the developing dipolar transition
state for rotation.
Although a systematic study of the reactivity of enamines as
a function of electron-accepting groups has not been done, the
reactions of enamines with electrophiles has been well
documented.9-12 A reaction of particular interest is the first step
of enamine hydrolysis, namely, reaction of an enamine with
the simplest electrophile, a hydrogen ion. Enamines exhibit
atypical acid-base properties relative to simple amines. The
basic nature of enamines derives from a mesomeric effect that
leads to two possible sites of protonation. Protonation at the
nucleophilic â-carbon of an enamine results in the iminium ion
† University of San Diego.
‡ California State University.
(1) (a) Osman, R.; Zunger, A.; Shvo, Y. Tetrahedron 1978, 34, 2315.
(b) Olsson, T.; Sandstrøm, J. Acta Chem. Scand. Ser. B 1982, 23. (c) Favini,
G.; Gamba, A.; Todeschini, R. J. Chem. Soc., Perkin Trans. 2 1985, 915.
(2) Pappalardo, R. R.; Marcos, E. S.; Ruiz-Lopez, M. F.; Rinaldi, D.;
Rivail, J.-L. J. Phys. Org. Chem. 1991, 4, 141. Pappalardo, R. R.; Marcos,
E. S.; Ruiz-Lopez, M. F.; Rinaldi, D.; Rivail, J.-L. J. Am. Chem. Soc. 1993,
115, 3722.
(3) Wong, M. W.; Frisch, M. J.; Wiberg, K. B. J. Am. Chem. Soc. 1991,
113, 4776. Head-Gordon, M.; Pople, J. A. J. Phys. Chem. 1993, 97, 1147.
(4) Dwyer, T. J.; Jasien, P. G. J. Mol. Struct. (THEOCHEM) 1996, 363,
139.
(5) (a) Lister, D. G.; Macdonald, J. N.; Owen, N. L. Internal Rotation
and InVersion Academic: London, 1978; references cited therein. (b)
Sandstrøm, J. Top. Stereochem. 1983, 14, 84 and references cited therein.
(6) Shvo, Y.; Shanan-Atidi, H. J. Am. Chem. Soc. 1969, 91, 6683. Shvo,
Y.; Shanan-Atidi, H. J. Am. Chem. Soc. 1969, 91, 6689. Dahlqvist, K. Acta
Chem. Scand. 1970, 24, 1941. Wennerbeck, I.; Sandstrøm, J. Org. Magn.
Reson. 1972, 4, 783.
(8) Baum, K.; Bigelow, S. S.; Nauyen, N. V.; Archibald, T. G.; Gilardi,
R.; Flippen-Anderson, J. L.; George, C. J. Org. Chem. 1992, 57, 235.
(9) Stork, G.; Terrel, R.; Scmuszkovicz, J. J. Am. Chem. Soc. 1954, 76,
2029. Stork, G.; Landesman, H. J. Am. Chem. Soc. 1956, 78, 5128, 5129.
(10) Elkik, E. Bull Soc. Chim. Fr. 1960, 972. Kuehne, M. E.; Garbacik,
T. J. Org. Chem. 1970, 35, 1555. Capon, B.; Wu, Z.-P. J. Org. Chem.
1990, 55, 5, 2317. Hickmott, P. W. In The Chemistry of Enamines Part 1;
Rappoport, Z., Ed.; John Wiley & Sons: New York, 1994; p 727, references
cited therein.
(11) Hickmott, P. W. Tetrahedron 1982, 38, 1975.
(12) Catalan, J.; Blanco, F. G. In The Chemistry of Enamines Part 1;
Rappoprt, Z., Ed.; John Wiley & Sons: New York, 1994; p 695, and
references cited therein.
(7) Kessler, H. Chem. Ber. 1970, 103, 973.
10.1021/ja983016m CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/25/1998

Protonation of â-Cyanoenamines
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12943
(1) while protonation at the amine nitrogen leads to the
enammonium ion (2).
Scheme 1
Ordinarily, the amino nitrogen or the â-carbon of an enamine
has been considered as a likely site of protonation in hydrolysis
reactions. Coward and Bruice13 have studied the mechanism of
hydrolysis of several â-cyanoenamines, which typically follow
first-order kinetics with general acid catalysis, and have observed
anomalous behavior below pH 1. They attribute this result to a
change in mechanism from rate-determining general acid-
catalyzed tautomerism between the enamine and the Câ-
protonated enamine to rate-determining hydrolysis of the Câ-
protonated enamine. However, Kresge’s review14 of the literature
of hydrolysis of enamines points out that protonation of the
cyano nitrogen (with formation of the ketenimine) cannot be
ruled out as an explanation of the observed behavior at pH <
1 in Bruice’s work. Few reports of the formation of unsubstituted
ketenimines may be found. Fleury and Libis15a reported the
formation of crystalline solid ketenimines upon treatment of
acylmalononitriles with acid. Ferris and Trimmer15b reported
the detection by IR of the transient ketenimine of several primary
enaminonitriles following UV irradiation at -196 °C, though
the attributed stretching frequency disappeared at higher tem-
peratures.
decanted, and the solid washed four times with diethyl ether and finally
dried under a stream of nitrogen gas. The enamines were prepared
according to the method of Kessler.7 Chloroformamidinium chloride
(8, 0.01 mol, 1.7 g) in 5 mL of CH3CN was added to 0.01 mol of the
substituted phenylacetonitrile dissolved in 15 mL of CH2Cl2 plus 2.5
equiv (0.025 mol, 3.75 mL) of 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU). The reaction mixture was refluxed for 12 h and cooled to room
temperature, and the solvents were evaporated. The residue was
dissolved in a small amount of CH2Cl2 and then purified by column
chromatography on silica gel using CH2Cl2 as solvent. For 3: mp 142-
144 °C; 1H NMR (CDCl3) δ 2.78 (6H, br s), 3.05 (6H, br s), 6.95 (2H,
d), 8.05 (2H, d). For 4: mp 95-97 °C; 1H NMR (CDCl3) δ 2.70 (6H,
s), 3.05 (6H, br s), 7.09 (2H, d), 7.49 (2H, d). For 5: mp 127-128 °C;
1H NMR (CDCl3) δ 2.65 (6H, s), 3.00 (6H, br s), 6.87 (2H, d), 7.35
(2H, d). For 6: mp 65-68 °C; 1H NMR (CDCl3) δ 2.53 (6H, s), 2.91
1
(6H, br s), 6.90 (4H, m). For 7: mp 61-64 °C; H NMR (CDCl3) δ
2.20 (3H, s), 2.52 (6H, br s), 2.89 (6H, br s), 6.83 (2H, d), 6.97 (2H,
How does the push-pull effect alter the basicity of an
enamine? In this study, we have investigated 1,1-bis(dimethy-
lamino)-2-cyano-2-p-X-phenyl-substituted ethylenes 3-7 under
acidic conditions. Our aim in this report is to provide compelling
evidence for ketenimine formation via protonation of a cyano
nitrogen through experimental NMR data and quantum me-
chanical calculations.
d).
NMR Methods. Proton NMR spectra were acquired on a Varian
UNITY-300 spectrometer. Samples were prepared by dissolving 10 mg
of the enamine in 700 µL of deuterated solvent (CDCl3, CD2Cl2, or
CD3C(O)CD3). Aliquots of 0.25-1.5 equiv of trifluoroacetic acid (TFA)
were then added to the sample. The 1D variable-temperature spectra
were regulated to (0.1 °C. All 2D exchange spectroscopy (EXSY)18
data were acquired at 300 MHz using a NOESY sequence with the
States et al.19 phase-cycling method to generate pure absorption phase
spectra. The spectra were collected with 1024 points in t2 using a
spectral width of 2800 Hz and a mixing time of 350 ms. Typically 256
t1 experiments were recorded and zero-filled to 1K. For each t1 value
8 scans were signal averaged using a recycle delay of 4 s.
Computational Methods. All calculations were performed with the
Gaussian 94 program package20 on an IBM RISC/6000 workstation.
Calculations used the density functional method denoted B3LYP21
which includes nonlocal corrections to the correlation functional. The
inclusion of electron correlation was deemed important in accurately
describing these systems. Further, density functional methods offer less
computationally demanding alternatives to Møller-Plesset second-order
perturbation theory (MP2)22 calculations with essentially the same
Experimental Section
Materials. Reagents and solvents were purchased from Aldrich. The
solvents were purified according to accepted procedures.16
Synthesis of Enamines. General Procedure. The enamines 3-7
were prepared according to Scheme 1 where X ) NO2, CF3, Br, F, or
CH3.
The chloroformamidinium chloride 8 was prepared from tetra-
methylurea and oxalyl chloride as previously described.17 In a typical
experiment, oxalyl chloride (0.01 mol, 25 g) was dissolved in 40 mL
CH2Cl2 under an inert atmosphere and added to tetramethylurea (0.13
mol, 15.5 g) dissolved in 25 mL CH2Cl2 and stirred. The solution turned
lime green with the evolution of gas. Solid 8 was precipitated from
solution by the addition of 30 mL of diethyl ether. The solvent was
(18) Perrin, C. L.; Gipe, R. K. J. Am. Chem. Soc. 1984, 106, 4036. Perrin,
C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935.
(19) States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson. 1982,
48, 286.
(20) Gaussian 94, Revision B.1; Frisch, M. J.; Trucks, G. W.; Schlegel,
H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;
Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.;
Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.;
Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng,
C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E.
S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.; Stewart,
J. J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian, Inc.,
Pittsburgh, PA, 1995.
(13) Coward, J. K.; Bruice, T. C. J. Am. Chem. Soc. 1969, 91, 5329.
(14) Keeffe, J. R.; Kresge, A. J. In The Chemistry of Enamines Part 2;
Rappoport, Z., Ed.; John Wiley & Sons: New York, 1994; p 1049.
(15) (a) Fleury, J.-P.; Libis, B. Compt. Rend. 1963, 256, 2419. (b) Ferris,
J. P.; Trimmer, R. W. J. Org. Chem. 1976, 41, 19.
(16) Gordon, A. J.; Ford, R. A. The Chemist’s Companion; John Wiley
& Sons: New York, 1972.
(17) Eilingsfeld, H.; Neubauer, G.; Seefelder, M.; Weidinger, H. Chem.
Ber. 1964, 97, 1232.
(21) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (b)
Becke, A. D. Phys. ReV. A 1988, 38, 3098. (c) Becke, A. D. J. Chem. Phys.
1993, 98, 5648.
(22) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618.

12944 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
Dwyer et al.
Scheme 2
Figure 1. NMR data showing unprotonated 3 (bottom trace) and the
variable-temperature spectra of 3 in CDCl3 with 1.5 equiv of TFA
added. The plus signs indicate the aromatic (downfield) and N-methyl
(upfield) resonances of the ketenimine amidinium ion 9 with X ) NO2.
samples containing 3-7 in any deuterated solvent affected the
CdC and/or C-N bond rotational barriers. This was indicated
by the dramatic changes in the peak patterns in the N-methyl
1
regions of the H NMR spectra. For brevity, we discuss the
accuracy.23 Another advantage of the density functional methods, in
this particular case, was the stability of the wave function. We have
found that the restricted Hartree-Fock (RHF) method for these closed
shell systems has an RHF f UHF (unrestricted Hartree-Fock)
instability.24,25 This instability would lead to complications, such as
spin contamination, if UHF and UMP2 wave functions were used in
these calculations.
The basis sets used included the standard 6-31G set and a modified
6-31G* set that added p-polarization functions only on the H+. We
denote this 6-31G*+p. An extended 6-311G(2d,p) basis set was used
in selected calculations. Geometries were optimized with both the 6-31G
and 6-31G*+p sets. Structures were converged without any symmetry
constraints and achieved convergence to 0.001 Å in bond lengths and
0.1° in most bond angles. In a few cases, some torsional angles with
very flat potentials had convergence to slightly greater values, however,
the associated energy changes were less than 1.0 × 10-8 a.u. Analytic
second derivatives were calculated for the CN- and Câ-protonated forms
with the 6-31G basis set in order to assess the effect of zero-point energy
(ZPE) corrections on the relative stability.
spectra of compounds 3, 6, and 7 in the presence of 1.5 equiv
of TFA, since the spectral characteristics do not change above
1.0 equiv of acid. Figure 1 (bottom trace) shows the proton
spectrum of 3 in CDCl3 at 23 °C in the absence of acid
(unprotonated). The broad peaks at 3.0 ppm arise from the
N-methyl protons (Scheme 2, A, B, C, and D) in 3, and those
at 7.0 and 8.1 ppm are due to the p-nitrophenyl aromatic protons.
At room temperature, CdC rotation occurs at a moderate rate
while C-N bond rotations are fast on the NMR time scale. As
the temperature is decreased (not shown), the CdC rotation is
nearly stopped while the C-N rotations are slowed, distinguish-
ing all four N-methyl group proton environments. Figure 1 also
shows the NMR spectrum of 3 in CDCl3 at 23 °C plus 1.5 equiv
of TFA. Clearly, the N-methyl signals are coalesced relative to
those in unprotonated 3 which indicates a faster exchange rate.
The effect of decreasing temperature on the NMR spectrum
of 3 plus 1.5 equiv of TFA in CDCl3 solution is displayed in
Figure 1. As the temperature is decreased, the N-methyl groups
become distinguishable. The peak assignments for the N-methyl
proton signals in 3 at -20 °C are (using amidine nomenclature)
exo-N-CH3 (δ3.2, cis to cyano group), endo-N-CH3 (δ3.0,
cis to cyano group), endo-N-CH3 (δ3.0, trans to cyano group),
and exo-N-CH3 (δ2.7, trans to cyano group).7 The endo-N-
CH3 signals are overlapped leading to only three peaks in this
region for 3. The signals at 7.0 and 8.1 ppm are due to the
p-nitrophenyl ring protons, while the peak at approximately 9
ppm is due to excess TFA. The N-methyl proton region of the
2D EXSY spectrum of 3 at -20 °C is shown in Figure 2. Using
Scheme 2 as a guide, (the rotational processes are the same,
except now we have a protonated molecule) the diagonal of
the EXSY gives the relative chemical shifts of the N-methyl
proton signals assigned above. The cross-peaks labeled 1 and
1′ arise due to both CdC and C-N (cis to cyano group)
rotational processes. Similarly, cross-peaks 3 and 3′ are due to
CdC and C-N (trans to cyano group). The cross-peaks labeled
2 and 2′ arise strictly from CdC rotation.
Results
NMR Spectroscopic Analyses. The dynamic processes
occurring in enamines 3-7 are diagrammed in Scheme 2. There
are three rotational processes that serve to interconvert the eight
conformers (denoted as filled circles at the corners of the cube
in Scheme 2).
Rotation about the CdC bond (process I, vertical lines in
Scheme 2) leads to interchange of the N,N-dimethyl groups.
Rotations about C-N bonds (processes II and III in Scheme 2)
lead to exchange of the N-methyl environments on an amino
group. It was immediately obvious that the addition of TFA to
(23) (a) Schmiedekamp, A. M.; Topol, I. A.; Burt, S. K.; Razafinjanahary,
H.; Chermette, H.; Pfaltzgraff, T.; Michejda, C. J. J. Comput. Chem. 1994,
15, 975. (b) Martell, J. M.; Goddard, J. D.; Eriksson, L. A. J. Phys. Chem.
A 1997, 101, 1927.
(24) (a) Seeger, R.; Pople, J. A. J. Chem. Phys. 1977, 66, 3045. (b)
Schlegel, H. B.; McDouall, J. J. W. In Computational AdVances in Organic
Chemistry: Molecular Structure and ReactiVity; Ogretir, Csizmadia, I. G.,
Eds.; NATO ASI Series, Series C: Mathematical and Physical Sciences;
Kluwer: Dordrecht, The Netherlands, 1991; Vol. 330, p 167ff.
(25) Jasien, P. G. Unpublished results.
We have also studied enamine 6 under acidic conditions in
CDCl3 plus 1.5 equiv of TFA. Figure 3 shows the variable-

Protonation of â-Cyanoenamines
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12945
Table 1. Energies (kcal mol-1) of Protonated Species from B3LYP
Calculations Relative to the CN-Protonated Systema-c
system
basis set
∆E(C)
∆E(C) + ZPE
∆E(N)
3
6-31G*+p
1.6
1.7
1.0
1.2
0.7
0.9
0.4
0.5
2.3
2.4
1.8
2.0
1.7
1.9
1.4
1.5
16.3
6-311G(2d,p)
6-31G*+p
4
6
7
16.0
15.4
15.2
6-311G(2d,p)
6-31G*+p
6-311G(2d,p)
6-31G*+p
6-311G(2d,p)
a 6-311G(2d,p) energies were evaluated at the 6-31G*+p-optimized
geometries. b E(CN-protonated) ) 0. ∆E(X) represents the energy
relative to CN-protonated form. c ZPE-corrected results use B3LYP/
6-31G calculated values.
Figure 2. Expansion of the N-methyl region of the 2D EXSY spectrum
of 3 in CDCl3 plus 1.5 equiv of TFA. Cross-peak identities are described
in the text.
Protonation of compound 7 under similar sample conditions
as above yielded the variable-temperature spectra for the
N-methyl proton region displayed in Figure 5. A similar peak
pattern as in protonated 6 arises at still lower temperature (-60
°C), indicating two forms of protonated 7 in solution. In
protonated 7, the integration of the signals yields a ratio of 60:
40 for the two forms.
Compounds 4 and 5 show protonation behavior intermediate
between that of 3 and 6 (as would be expected since they contain
substituents with σ values intermediate between those for -NO2
and -F). Both protonated 4 and 5 show CN-protonated enamine
to Câ-protonated enamine ratios of 93:7 in all solvents studied.
A Hammett-type free energy relationship for the Câ-proto-
nated/CN-protonated equilibrium is presented in Figure 6. A
good linear correlation with F ) +1.8 is obtained. This
moderately large positive F value implies that the â-carbon is
extremely electron-deficient relative to the cyano nitrogen.
Further, the distribution of the electron density in the molecules
prior to protonation is modulated by the substituted phenyl ring.
Computations. Table 1 displays the calculated stabilities of
the various protonated forms of 3, 4, 6, and 7 for the B3LYP
calculations with the 6-31G*+p and 6-311G(2d,p) basis sets.
The calculated relative energies of the CN-protonated and Câ-
protonated forms change by 0.2 kcal mol-1 or less on going
from the 6-31G*+p to the 6-311(2d,2p) basis set. The results
for the B3LYP/6-31G calculations are not shown since they
are in very close agreement with the larger basis set calculations.
For all four push-pull systems studied, the most favorable
protonation site is the nitrogen of the cyano group, followed
closely by Câ, and then the amino nitrogen. The calculated ZPE
corrections favor the CN-protonated enamine over the Câ-
Figure 3. NMR data showing unprotonated 6 (bottom trace) and the
variable-temperature spectra of 6 in CD3C(O)CD3 with 1.5 equiv of
TFA added. The plus signs indicate the aromatic (downfield) and
N-methyl (upfield) resonances of the ketenimine amidinium ion 9 with
X ) F. The stars mark the aromatic (downfield) and N-methyl (upfield)
resonances of the Câ-protonated amidinium ion 12 with X ) F.
temperature NMR spectra associated with this compound.
Unlike 3, the N-methyl proton region in enamine 6 develops a
six-peak pattern at low temperature (-30 to -50 °C). The peak
assignments for the N-methyl proton region in protonated 6 are
N-(CH3)2 (δ3.7, cis to cyano group) and N-(CH3)2 (δ3.5, trans
to cyano group) of a Câ-protonated form of 6 and exo-N-CH3
(δ3.1, cis to cyano group), endo-N-CH3 (δ2.9, cis to cyano
group), endo-N-CH3 (δ2.85, trans to cyano group), and exo-
N-CH3 (δ2.5, trans to cyano group) of a CN-protonated form.
The difference in line widths of the N-methyl peaks for Câ-
protonated forms of 6 is a consequence of the difference in T2
of the exo- and endo-N-CH3 groups. The resulting 2D EXSY
spectrum (Figure 4) acquired at -50 °C suggests the presence
of two forms of protonated 6 which are in intermolecular
exchange due to proton transfer and intramolecular exchange
due to restricted rotational processes. Integration of the 1D
spectrum at -50 °C yields a ratio of the two forms of 87 (CN-
protonated):13 (Câ-protonated). Further evidence for two distinct
forms of protonated 6 is provided in the aromatic region of the
1H spectrum. In this region, two sets of signals are seen in a
ratio of 13 (downfield signal):87 (upfield signal). Note that these
two forms are in intermolecular exchange as revealed by the
cross-peaks in the downfield (6.9-7.6 ppm) region of Figure
4.
protonated form by 0.7-1.0 kcal mol-1
.
The proton affinities (PAs) for the push-pull systems show
an interesting trend. The results indicate that the larger the
energy difference between the Câ-protonated and CN-protonated
forms, the smaller the PAs. The PAs (kcal mol-1) for the cyano
nitrogen site based on ZPE corrected 6-31G*+p calculations
are 3 (223.5), 4 (227.5), 6 (230.4), and 7 (233.3). The trend 3
< 4 < 6 < 7 holds, regardless of the site of protonation.
Table 2 lists selected calculated bond lengths and angles for
the protonated and unprotonated push-pull systems studied in
this work. For brevity, only average C-N (amine) bond lengths
are listed. These bond lengths were generally within 0.005 Å,
but differed by as much as 0.009 Å in one instance. In all cases,
the geometry around the ethylenic carbon atoms is planar which
is also true for the amino nitrogen atoms.
Lastly, Table 3 gives the calculated electric dipole moments
for the unprotonated push-pull systems. In addition, qualitative
information on the charge partitioning derived from a Mulliken
population analysis is given. The results are listed for two

12946 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
Dwyer et al.
Figure 4. The 2D EXSY spectrum of 6 in CD3C(O)CD3 plus 1.5 equiv of TFA. The cross-peak patterns in the aromatic (downfield) and N-methyl
(upfield) regions are described in the text.
Figure 6. Hammett-type plot for the equilibrium Câ-protonated h CN-
protonated.
Figure 5. NMR data showing unprotonated 7 (bottom trace) and the
variable-temperature spectra of 7 in CD3C(O)CD3 with 1.5 equiv of
with N-protonation, since N-protonation would give rise to rapid,
TFA added. The plus signs indicate the aromatic (downfield) and
unhindered rotation about the C-N single bonds; the observed
N-methyl (upfield) resonances of the ketenimine amidinium ion 9 with
slowing of C-N rotations at moderate temperatures suggests
lone-pair delocalization, hindered rotation, and thus an absence
of protonation at the amino group. Second, the calculational
results suggest that N-protonation is unlikely (vida infra).
Protonation of 3 at the â-carbon is ruled out by the peak
patterns observed in both the 1D and 2D NMR spectra.
â-Carbon protonation results in rehybridization of the â-carbon
to sp3 and renders the CR-Câ bond a single bond with essentially
free rotation. Thus, the N-methyl proton region of the NMR
spectrum should yield a single pair of peaks arising from the
two distinguishable proton environments for the N-methyl
groups (Scheme 2). This is not consistent with the three peaks
seen in Figures 1 and 2.
X ) CH3. The stars mark the aromatic (downfield) and N-methyl
(upfield) resonances of the Câ-protonated amidinium ion 12 with X )
CH3.
separate partitionings. The first separates the molecule between
the two ethylenic carbon atoms to obtain partial charges on the
donor and acceptor subsystems. The second sums the charges
only on the phenyl ring and phenyl ring substituents. Division
of the molecule into large subsystems mitigates somewhat the
problems associated with using Mulliken populations to assign
partial charges.
Discussion
The likely site of protonation of 3 is the cyano nitrogen with
formation of a ketenimine amidinium ion 9. Protonation of the
cyano group is consistent with the NMR spectral results. In 9
(X ) NO2), there are four distinguishable N-methyl proton
Protonation of 3 at either amino nitrogen is ruled out based
on two lines of evidence. First, the NMR spectra are inconsistent

Protonation of â-Cyanoenamines
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12947
Table 2. Bond Lengths and Angles from B3LYP/6-31G*+p Calculationsa
system
R(CdC)
R(C-N) (av)
R(C-Cφ)
R(C-CN)
R(CN)
R(XH+)
CNH+
125.8
R(C-Y)
3
4
6
7
unprotonated
CNH+
1.404
1.465
1.545
1.376
1.348
1.336
1.470
1.491
1.534
1.424
1.346
1.472
1.168
1.204
1.160
1.460
1.478
1.480
1.019
1.094
CH+
unprotonated
CNH+
1.400
1.464
1.544
1.379
1.348
1.336
1.477
1.493
1.533
1.424
1.345
1.472
1.168
1.205
1.160
1.500
1.510
1.512
1.019
1.093
125.3
124.9
124.6
CH+
unprotonated
CNH+
1.395
1.462
1.545
1.382
1.349
1.337
1.482
1.493
1.532
1.423
1.345
1.472
1.168
1.206
1.160
1.352
1.338
1.337
1.019
1.093
CH+
unprotonated
CNH+
1.394
1.461
1.544
1.383
1.349
1.337
1.482
1.493
1.532
1.424
1.345
1.472
1.168
1.206
1.160
1.511
1.509
1.509
1.019
1.093
CH+
a Bond lengths in angstroms; angles in deg. X represents the atom to which the H+ is attached, and Y represents the atom of the para substituent.
Table 3. Group Charges and Electric Dipole Moments for
Unprotonated Systems from B3LYP/6-31G*+p
system
µ(D)
donor (e)
acceptor (e)
phenyl-X (e)
3
4
6
7
9.95
7.65
6.58
6.16
0.40
0.37
0.34
0.33
-0.40
-0.37
-0.34
-0.33
-0.06
-0.02
0.01
0.02
the 2D EXSY. The four upfield peaks correspond to the
N-methyl groups of the ketenimine amidinium ion, 9 (X ) F),
similar to protonated 3. The cross-peaks among these four peaks
arise due to the dynamic processes diagrammed in Scheme 2
(for the unprotonated compound). The cross-peaks connecting
the peaks from 12 (X ) F) with those from 9 (X ) F)
correspond to intermolecular exchange of the two forms, via
proton transfer. The absence of any broadening of the aromatic
signals corresponding to the Câ-protonated forms of 4-7 are
consistent with the sp3 hybridization of Câ.
Additional evidence for the two protonated forms of 6 are
given by the two sets of peaks in the aromatic region of the
NMR spectrum. The 2D EXSY shows cross-peaks between
diagonal peaks, again indicating intermolecular exchange via
proton transfer between the two protonated forms (as monitored
by the aromatic protons).
The situation in enamine 7 under acidic conditions is quite
similar to that in 6. The familiar six-peak pattern in the N-methyl
proton region of the low-temperature NMR spectrum of 7 in
the presence of TFA is observed in CDCl3, CD2Cl2, or
CD3C(O)CD3. Intermolecular exchange via proton transfer
between the two protonated forms as well as intramolecular
exchange via bond rotations are again obvious from the cross-
peak patterns in the 2D EXSY of 7 (not shown). Enamine 7,
however, shows a ratio of CN-protonated to Câ-protonated
tautomers of 60:40. Again, formation of the ketenimine ami-
dinium ion 9 (X ) CH3) is preferred relative to the amidinium
ion 12 (X ) CH3) in this push-pull system.
The calculations yield both structural and energetic data that
is in agreement with the experimental data. Table 1 clearly
shows that in all cases the result of protonation on the amino
nitrogen leads to a structure which lies at least 15 kcal mol-1
higher in energy than the Câ- or CN-protonated forms. The size
of this energy difference is well outside the error bounds one
would expect from calculations of this type. When coupled with
the experimental evidence, the calculations argue convincingly
that protonation of the amino nitrogen is not favorable.
Distinguishing protonation at the â-carbon versus the cyano
nitrogen is less obvious from the calculations since the energy
environments owing to the contribution of a resonance structure
such as 10 (X ) NO2) where there is hindered rotation about
the CdC bond. Exclusive CN-protonation by TFA and keten-
imine formation in 3 occurs in all solvents studied, ranging in
dielectric constant from ꢀ ) 5 (chloroform) to ꢀ ) 20.9
(acetone).
The temperature dependence of the aromatic signals in the
NMR spectrum of protonated 3 also lends evidence for CN-
protonation. Conjugation of the ketenimine with the aromatic
ring is possible via resonance structure 11. The partial double
bond in the C-C(phenyl) bond contributes to hindered rotation
of the aromatic ring and this process slows with decreasing
temperature. This is manifested in the 1D NMR spectrum as a
continual broadening of the signals from the aromatic hydrogens
and in the 2D EXSY as exchange cross-peaks at very low
temperatures (<-60 °C).
In compound 6, the NMR spectra are a bit more complex.
Addition of the even the smallest amount of TFA (0.25 equiv)
results in six peaks in the N-methyl proton region. At 1.0 (or
greater) equiv of TFA, the ratio of the higher intensity upfield
peaks (4 peaks) to the lower intensity downfield peaks (2 peaks)
is 87:13. The 2D EXSY spectrum is extremely useful in sorting
out the various species and dynamics in this sample. We attribute
the low-temperature spectral pattern and 2D EXSY cross-peak
pattern to the existence of both a Câ-protonated form of 6 (12
(X ) F)) and a CN-protonated (ketenimine, 9 (X ) F)) form,
in a ratio of 13:87. The two smaller peaks in this N-methyl
proton region result from the Câ-protonated tautomer. In 12 (X
) F) there is relatively free rotation about the CR-Câ bond,
distinguishing only two N-methyl groups that are interchanged
via C-N bond rotation. This rotation leads to the cross-peak
between the two diagonal peaks of the Câ-protonated form in

12948 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
Dwyer et al.
differences are quite small. Table 1 shows that in all cases the
calculations predict that the CN-protonated enamine lies lower
in energy than the Câ-protonated form. The calculated CN-
protonated/Câ-protonated energy differences follow the order
3 > 4 > 6 > 7 with or without the addition of ZPE corrections.
No claims are made that the current calculations are able to
obtain accuracies of a few tenths of a kcal mol-1; however, the
experimental and calculational results strongly support the
interpretation of CN-protonation over Câ-protonation. The fact
that calculations performed with a number of basis sets give
such consistent results also strongly supports their integrity.
planarity. These dramatic structural changes that occur on
protonation make it unlikely that the substituted phenyl ring is
able to interact with the positive charge developing in the
protonated amidinium ion.
Although the protonation energy differences between Câ and
CN are largest in 3, this system shows the smallest PAs. In
fact, the CN-protonation energy in 3 is calculated to be 10 kcal
mol-1 less than that for 7. The explanation for this may lie in
the strong electron-withdrawing ability of the -NO2 group
which gives the phenyl ring a slightly more negative overall
charge than in 7. Similar arguments hold when comparing the
PAs for the other systems. Even though the total negative charge
on the acceptor portion of the molecule is the largest in 3, the
delocalization of electron density onto the phenyl ring comes
largely at the expense of Câ. This makes the CN site more
electron rich in comparison with Câ. These results are in
agreement with the moderately large, positive F value derived
from the Hammett-type relationship shown in Figure 6.
The fact that the calculations did not include solvent effects
will certainly change the relative energy differences Câ- versus
CN-protonated enamines. However, the calculations clearly
indicate that protonation of the cyano nitrogen is competitive
with protonation of the â-carbon. In addition, the energy
differences between CN- and Câ-protonated forms (3 > 4 > 6
> 7) should be relatively independent of solvent due to the
structural similarities. Another inadequacy of the theoretically
calculated energy differences is that they do not include thermal
effects, but correspond to the energy differences at T ) 0 K.
Estimates of the thermal effects for the in vacuo systems, as
calculated from the Gaussian 94 program using the B3LYP/6-
31G calculations indicate that these effects are small (∼0.1 kcal
mol-1).
The structural changes implied by the NMR spectra are also
borne out in the calculations. Perhaps the most telling aspect
of the structural changes indicating formation of a ketenimine
is the decrease of the C-CN bond length upon CN-protonation
(Table 2). This decrease is almost 0.08 Å and results in a C-CN
bond comparable in length to what one would expect for a true
CdC double bond. Consistent with this is the smaller increase
in the cyano CN bond which amounts to nearly 0.04 Å and the
C-N-H angle in the CN-protonated species which is 125°.
An angle of this size certainly implies an sp2 hybridized nitrogen
atom as opposed to a 180° angle that would indicate sp
hybridization.
In all Câ- or CN-protonated species, the CdC bond length
increases dramatically from that in the unprotonated molecule.
There is an increase of 0.06-0.07 Å for CN-protonation, but
an even larger increase of 0.14-0.15 Å occurs for Câ-
protonation. The resultant bond lengths in the case of Câ-
protonation are in line with what one would expect for a C-C
single bond. Furthermore, after Câ-protonation the gross struc-
ture around this carbon is tetrahedral. Both of these results are
once again consistent with a small energy barrier for C-C
rotation. Protonation on either the â-carbon or the cyano nitrogen
has a pronounced effect on the CR-N bonds of the amine
groups. These bonds show decreases of 0.03 Å (CN-protonation)
and 0.04-0.05 Å (Câ-protonation) which imply a higher barrier
for CR-N rotation. Significant changes are also seen in the Câ-
C(phenyl) bond lengths. This length increases by 0.05-0.06 Å
for the Câ-protonation and about one-half this value for CN-
protonation. The shortening of this bond upon CN-protonation
is also consistent with the experimental evidence of restricted
rotation of the phenyl group.
To probe the effect of solvent on the relative stabilities of
the CN-protonated and Câ-protonated systems more closely, the
equilibrium distribution data was analyzed for 6 and 7. For 4
and 5 the small amount of Câ-protonated enamine relative to
CN-protonated enamine renders the integration less certain and
were not analyzed. Plots of ln Keq vs T-1 (not shown) for the
process:
Câ-protonated h CN-protonated
were made over the temperature ranges 0 to -50 °C (5 points)
and -20 to -60 °C (4 points). In both cases the results were
extremely linear (R2 ) 0.99). The ∆H° and ∆S° values extracted
from these fits are 2.5 kcal mol-1 and 15 cal K-1 mol-1 for 6
and 1.9 kcal mol-1 and 11 cal K-1 mol-1 for 7. Interestingly,
the ∆H° values for the proton transfer are positive. If the
calculated in vacuo results for the relative stabilities are taken
as reasonable estimates of the in vacuo ∆H° values, this means
that solvation stabilizes the Câ-protonated form more than the
CN-protonated enamine. Due to the small ∆H° term, the position
of the equilibrium is determined by the positive ∆S° term.
Estimates of ∆S° from the calculations reveal that this difference
is small for the isolated molecules (<1 cal K-1 mol-1). Once
again, this implies that the major difference in the ∆S° term
arises from the solvent effect.
It is interesting to note that the predicted CR-Câ “double”
bond lengths in the unprotonated systems are all calculated to
be longer than the CR-N “single” bond lengths. This is
consistent with the highly delocalized π-electron systems in
these molecules which leads to their low CR-Câ and high CR-N
rotational barriers in the unprotonated forms.
The steric crowding of the bis(N,N-dimethylamino) moiety
with the phenyl ring in 3-7 leads to nonplanarity about the
CdC in the unprotonated push-pull molecules. Indeed, the
dihedral angle about the CdC (between the N-CR-N plane
and C(cyano)-Câ-C(phenyl) plane) is around 30° in unpro-
tonated 3-7. This precludes optimum overlap of the lone pair
electrons on the nitrogens with the parallel p orbitals of the
CdC. However, the conjugation and delocalization of these
electrons throughout the π system is evident from the variations
in CdC and C-N rotational barriers measured by Kessler.7
Protonation of 4-7 at Câ completely disrupts the conjugation
between the two “halves” of the molecule, and the phenyl ring
twists 20° further to decrease steric crowding. On protonation
of 3-7 at the cyano nitrogen, the dihedral angle about the CdC
increases to 40° while the phenyl ring twists an additional 10°
relative to the unprotonated molecule, forcing it further from
The altered reactivity toward acid observed in the â-cyano-
enamines 3-7 studied here is analogous to that seen in the
synthetically useful enaminones (push-pull enamines with
(-C(dO)R as electron-withdrawing group). Delocalization of
electrons into the carbonyl group facilitates protonation on
oxygen26 in nearly all cases of enaminones and enaminoesters.
(26) (a) Kozerski, L.; Dabrowski, J. Org. Magn. Reson. 1972, 4, 253.
(b) Berg, U.; Sjostrand, U. Org. Magn. Reson. 1978, 11, 555.

Protonation of â-Cyanoenamines
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12949
Conclusion
the cyano group as the site of protonation and reproduce the
small difference in the protonation energies of the two sites.
Calculated structures give further evidence that a ketenimine is
formed as implied by the extremely short Câ-CN and elongated
cyano CN bond length upon CN-protonation.
The experimental NMR and quantum mechanical results
reported here present convincing evidence that the favored site
of protonation in the â-cyanoenamines 3-7 is the cyano group.
This results in the formation of a ketenimine in solution and is
manifested in the 1D and 2D NMR experiments. The NMR
data indicate 100% CN-protonation in 3, 93% in 4 and 5, 87%
in 6, and 60% in 7. The experimental energy difference between
the CN-protonated and Câ-protonated forms is small enough in
4-7 that equilibrium mixtures are observed; however, in all
cases the ketenimine is preferred. Calculational results support
Acknowledgment. This work was supported in part by
Petroleum Research Fund Grant 26794-GB4 (T.J.D.). The
Varian UNITY300 NMR was acquired through an NSF-ILI
grant (DUE9051310).
JA983016M