Reactions of gaseous vinyl halide radical cations with ammonia. A study of mechanism by Fourier-transform ion cyclotron resonance spectrometry and ab initio molecular orbital calculation

Article abstract of DOI:10.1021/ja9619465

The reactions of the radical cations of vinyl chloride (1) and vinyl bromide (2) with ammonia in the gas phase have been studied by FT-ICR spectrometry and ab initio molecular orbital calculations. The FT-ICR experiments show that vinyl halide radical cations react mainly by substitution of the halogen atom, yielding C₂H₆N⁺ product ions and moderately by formation of NH₄⁺ with a total reaction efficiency of 57% (1^.+) and 35% (2^.+), respectively. The initial formation of NH₃D⁺ from 1-d₃^.+ and 2-d₃^.+ establishes a direct deprotonation of the radical cations by NH₃. Ab initio calculations confirm an exothermic deprotonation for all possible halovinyl radicals as the neutral product. The structure of the substitution product ions C₂H₆N⁺ was investigated by CA mass spectrometry and gas-phase titration experiments. The results show the presence of vinylammonium ions 5⁺ and acetaldiminium ions (6⁺) in the product ion mixture, but additional experiments reveal that 6⁺ is generated by a base catalyzed isomerization of initially formed 5⁺ ('shuttle mechanism'). The reaction energy profiles of the reactions of the vinyl halide radical cations 1^.+ and 2^.+ with ammonia were calculated by ab initio methods. The addition of NH₃ is very exothermic for both radical cations generating chemically activated distonic ammonium ions in the first reaction step. The addition proceeds regioselectively (Markovnikov and anti-Markovnikov), yielding preferentially an unreactive distonic ion by the Markovnikov orientation. This intermediate has to rearrange by a 1,2-NH₃ shift before eventually decomposing into vinylammonium ions 5⁺ by loss of the halogen atom. It is suggested that this rearrangement corresponds to the 'bottle neck' of the substitution process. The activation energy of the 1,2-NH₃ shift in the chloro- and bromo-substituted distonic ion is not significantly different, but the increased chemical activation of the chloro derivative results in an increased reaction efficiency for the total substitution process for the vinyl chloride radical cation. The ab initio calculations reveal further that the generation of the more stable substitution product acetaldiminium ion 6⁺ is inhibited by a large activation barrier for the hydrogen rearrangements necessary within the intermediate distonic ion.

Full text of DOI:10.1021/ja9619465

6544
J. Am. Chem. Soc. 1997, 119, 6544-6551
Reactions of Gaseous Vinyl Halide Radical Cations with
Ammonia. A Study of Mechanism by Fourier-Transform Ion
Cyclotron Resonance Spectrometry and ab Initio Molecular
Orbital Calculation
Andreas Nixdorf and Hans-Friedrich Gru1tzmacher*
Contribution from the Lehrstuhl I fu¨r Organische Chemie, Fakulta¨t fu¨r Chemie,
UniVersita¨t Bielefeld, Postfach 10 01 31, D-33501 Bielefeld, Germany
ReceiVed June 10, 1996X
Abstract: The reactions of the radical cations of vinyl chloride (1) and vinyl bromide (2) with ammonia in the gas
phase have been studied by FT-ICR spectrometry and ab initio molecular orbital calculations. The FT-ICR experiments
show that vinyl halide radical cations react mainly by substitution of the halogen atom, yielding C2H6N+ product
ions and moderately by formation of NH4+ with a total reaction efficiency of 57% (1•+) and 35% (2•+), respectively.
•+
•+
The initial formation of NH3D+ from 1-d3 and 2-d3 establishes a direct deprotonation of the radical cations by
NH3. Ab initio calculations confirm an exothermic deprotonation for all possible halovinyl radicals as the neutral
product. The structure of the substitution product ions C2H6N+ was investigated by CA mass spectrometry and
gas-phase titration experiments. The results show the presence of vinylammonium ions 5+ and acetaldiminium ions
(6+) in the product ion mixture, but additional experiments reveal that 6+ is generated by a base catalyzed isomerization
of initially formed 5+ (“shuttle mechanism”). The reaction energy profiles of the reactions of the vinyl halide radical
cations 1•+ and 2•+ with ammonia were calculated by ab initio methods. The addition of NH3 is very exothermic
for both radical cations generating chemically activated distonic ammonium ions in the first reaction step. The
addition proceeds regioselectively (Markovnikov and anti-Markovnikov), yielding preferentially an unreactive distonic
ion by the Markovnikov orientation. This intermediate has to rearrange by a 1,2-NH3 shift before eventually
decomposing into vinylammonium ions 5+ by loss of the halogen atom. It is suggested that this rearrangement
corresponds to the “bottle neck” of the substitution process. The activation energy of the 1,2-NH3 shift in the chloro-
and bromo-substituted distonic ion is not significantly different, but the increased chemical activation of the chloro
derivative results in an increased reaction efficiency for the total substitution process for the vinyl chloride radical
cation. The ab initio calculations reveal further that the generation of the more stable substitution product
acetaldiminium ion 6+ is inhibited by a large activation barrier for the hydrogen rearrangements necessary within
the intermediate distonic ion.
Introduction
investigated by appropriate mass spectrometric techniques, and
a study of the gas-phase ion/molecule reactions of unsaturated
organic radical cations can display many details of the reaction
mechanisms, in particular in combination with pertinent theo-
retical calculations of the reaction system.
The detection of radical cations as intermediates in organic
reactions has initiated an increasing number of studies of their
reactivities in solution1 and in the gas phase.2 These studies
reveal that unsaturated organic radical cations exhibit a high
reactivity toward electron-rich donor molecules and generate a
variety of reaction products. However, owing to the short
lifetime of most organic radical cations in solution it is difficult
to obtain information about which factors control the reactions3
and to determine the details of the reaction mechanisms. In
contrast, even very reactive gaseous radical cations are easily
Recently, we have performed a systematic study of the gas-
phase reactions of mono- and dihalogenated arenes2d,f,g with
ammonia and simple amines via radical cations using Fourier-
transform ion cyclotron resonance (FT-ICR) spectrometry.
These reactions result in the substitution of one halogen
substituent by the nucleophile by an addition/elimination
mechanism. The kinetics and the low reaction efficiency of
less than 15% of the reactions of the aromatic radical cations2d,f,g
show that the first addition step of this aromatic substitution
process is rate determining. This was explained by the
configuration mixing model of Shaik and Pross4 for polar
organic reactions. Further studies reveal that the detailed
mechanisms of the aromatic substitution via radical cations is
rather complex and includes rearrangement steps between
addition of the nucleophile and elimination of the halogeno
substituent.2f,g
X Abstract published in AdVance ACS Abstracts, June 15, 1997.
(1) (a) Johnstone, L. P.; Schepp, N. P. J. Am. Chem. Soc. 1993, 115,
6564. (b) Dinnocenzo, J. P.; Lieberman, D. R.; Simpson, T. R. J. Am. Chem.
Soc. 1993 115, 366. (c) Trampe, G.; Mattay, J.; Steenken, S. S. J. Phys.
Chem. 1989, 93, 7157. (d) Steenken, S. J.; Warren, C. J.; Gilbert, B. C. J.
Chem. Soc., Perkin Trans. 2 1990, 335. (e) Yasuda, M.; Matsuzaki, Y.;
Shima, K.; Pac, C. J. Chem. Soc., Perkin Trans. 2 1988, 745. (f) Parker, V.
D.; Tilset, M. J. J. Am. Chem. Soc. 1987, 109, 2251.
(2) (a) Heinrich, N.; Koch, W.; Morrow, J. C.; Schwarz, H. J. Am. Chem.
Soc. 1988, 110, 6332. (b) Shaik, S. S.; Pross, A. J. Am. Chem. Soc. 1989,
111, 4306. (c) Cho, J. K.; Shaik, S. S. J. Am. Chem. Soc. 1991, 113, 9890.
(d) Tho¨lmann, D.; Gru¨tzmacher, H.-F. J. Am. Chem. Soc. 1991, 113, 3281.
(e) Tho¨lmann, D.; Flottmann, D.; Gru¨tzmacher, H.-F. Chem. Ber. 1991,
124, 2349. (f) Tho¨lmann, D.; Gru¨tzmacher, H.-F. Int. J. Mass Spectrom.
Ion Processes 1992 117, 415. (g) Tho¨lmann, D.; Hamann, S.; Gru¨tzmacher,
H.-F. Int. J. Mass Spectrom. Ion Processes 1994, 137, 43.
In the case of a gas-phase substitution reaction of olefinic
(4) (a) Shaik, S. S. Acta Chem. Scand. 1990, 44, 205. (b) Shaik, S. S.;
Canadell, E. J. Am. Chem. Soc. 1990, 112, 1446. (c) Pross, A. J. Am. Chem.
Soc. 1986, 108, 3537. (d) Pross, A; Shaik, S. S. J. Am. Chem. Soc. 1989,
111, 4300. (e) Pross, A.; Shaik, S. S. J. Am. Chem. Soc. 1981, 103, 3702.
(3) For a recent example, see: Maslak, P.; Chapman, W. H., Jr.;
Vallombroso, T. M., Jr.; Watson, B. A. J. Am. Chem. Soc. 1995, 117, 12380.
S0002-7863(96)01946-4 CCC: $14.00 © 1997 American Chemical Society

Reactions of Vinyl Halide Radical Cations with NH3
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6545
1,2-Dichloro-1,1,2,2-tetradeuterioethane. Thionyl chloride (20.2
g, 169.5 mmol) was added slowly to a mixture of dry 1,1,2,2-
teradeuterioethylene glycol (2.80 g, 42.2 mmol) and dry pyridine (6.70
g 85.9 mmol). A very vigorous reaction occurred and the liquid
darkened considerably. The mixture was refluxed for 2 h and then
poured under vigorous stirring into ice-cold water. The organic layer
was washed with 5% sodium hydroxide solution and with water and
was dried over Na2SO4. Pure 1,2-dichloro-1,1,2,2-tetradeuterioethane
was obtained by distillation at 81 °C. Yield: 1.11 g (10.8 mmol, 26%).
MS (70 eV, EI) m/z: 106 (3.4) [C2D437Cl2•+], 104 (21.7) [C2D435Cl37-
Cl•+], 102 (34.0) [C2D435Cl2•+], 69 (15.4) [C2D437Cl+,C2D335Cl•+],
67 (72.7) [C2D435Cl+,C2D337Cl•+], 65 (15.4) [C2D335Cl•+], 53 (21.6)
[CD237Cl+], 51 (68.3) [CD235Cl+]. C2D435Cl2: mol weight calcd
101.9941, found 101.9942 (by FT-ICR mass spectrometry at m/∆m )
200.000).
Scheme 1
radical cations the addition of the nucleophile is not expected
to be the rate-determining step.2a This was corroborated by
preliminary FT-ICR study of the ion/molecule reactions of 1,2-
dichloro- and dibromoethenes with ammonia, methylamine, or
dimethylamine via radical cations.2e In this case substitution
of one halogen substituent occurs with a rather high efficiency
although the reaction rate constants were still distinctly below
the collision rate constant. Further, the structures of the
substitution products were not obvious, and the reactions of 1,1-
dichloroethene radical cations were different from those of the
1,2-isomer.2e This latter observation suggests an interesting
regioselectivity for the addition of the nucleophile to ionized
haloalkenes which are asymmetrically substituted. Thus, the
apparently simple substitution of haloalkene radical cations is
very likely also a complex multistep reaction, and more
experimental information is needed to understand the extraor-
dinary reactivity of olefinic radical cations and the course of
their reactions with nucleophiles. Therefore, we complemented
the previous investigations by a study of the ion/molecule
reactions of ionized vinyl chloride (1) and vinyl bromide (2)
with ammonia and other small donor molecules using FT-ICR
spectrometry and molecular orbital calculations. Here, we report
the results for the reaction with ammonia (Scheme 1) with
special emphasis on the determination of the structure of the
substitution products and a calculation of the reaction energy
profile.
The vinyl halides 1 and 2 are asymmetrically substituted and
should allow the detection of effects of regioselectivity during
the addition of the nucleophile to the ionized double bond.
Further, Br is a better “leaving group” from radical cations than
Cl because of its lower C-X bond energy which allows a study
of leaving group effects on the rate of the overall process.
Finally, vinyl halides are sufficiently small to apply ab initio
methods of a reasonable high level of theory to the calculation
of the reaction potential energy profile. Thus, these vinyl halide
radical cations are well suited for a detailed investigation of
the mechanism of the reaction of unsaturated organic radical
cations with donor substrates. The results of the kinetic
experiments and of the investigation of the structure of the
reaction product ions will be discussed first and will be followed
by a discussion of the reaction energy profile of this prototype
of a substitution reaction of olefinic radical cations. It will be
shown that the first addition step of the nucleophile to the
ionized vinyl halide is indeed fast but occurs primarily with
the “wrong” regioselectivity for substitution, and that the
necessary rearrangement of the initial distonic ion is apparently
the “bottle neck” of the overall reaction.
Vinyl bromide-d3 (2-d3) was prepared according to Schaefer et al.6
by base-induced dehydrohalogenation of 1,2-dibromo-1,1,2,2-tetradeu-
terioethane. The deuterium content of 1,2-dichloro-1,1,2,2-tetradeu-
terioethane and vinyl bromide-d3 was >98% (by mass spectrometry).
Mass Spectrometry. Ion/molecule reactions were investigated with
a Spectrospin Bruker CMS 47X FT-ICR instrument7 equipped with
4.7 T magnet and an external EI/CI-ion source.8 The standard
cylindrical ICR cell was used for rate measurement, while experiments
to determine the structure of product ions C2H6N+ were performed with
an Infinity cell.9 The vinyl halide radical cations were generated in
the external ion source by electron impact at a nominal electron energy
of 18-27 eV. All ions formed were transferred into the FT-ICR cell
containing ammonia at an appropriate constant background pressure
of 10-8 -10-7 mbar. The 12C2H3X•+ ions (X ) 35Cl or 79Br) were
isolated by applying a broad band ejection (“chirp ejection”, 88 Vp-p
,
80 µs) followed by a series of single frequency pulses (“single shots”,
14 Vp-p, 1.6 ms) to remove all isotopomers of the vinyl halide radical
cations. The isolated radical cations were thermalized by collision with
argon introduced into the ICR cell by a pulsed valve as described
before.2d After a delay time of 500 ms for removing argon any fragment
ions or product ions formed during the cooling period were ejected
again by single shots, and special care was taken to avoid any
translational excitation of the isolated vinyl halide radical cations during
this procedure. The time for the reaction of 12C2H3X•+ ions with
ammonia was varied from 1.5 ms to 15 s. Then, all ions in the FT-
ICR cell were excited by an frequency sweep of 88 Vp-p with a step
width of 7.8 kHz and an excitation pulse time of 8 µs. The mass spectra
were recorded by 32K data points for 20-30 different reaction times,
and each mass spectrum is the average of at least 16 data acquisition
sequences. After exponential multiplication of the time domain signal
and Fourier transformation the peak intensities of the magnitude spectra
were normalized to the sum of all ion detected at that reaction time
and plotted Vs the reaction time (“kinetic plot”).
The bimolecular rate constants kbi were calculated from the slope of
a logarithmic plot of the relative reactant ion intensity Vs the reaction
time and by taking into account the number density of ammonia in the
FT-ICR cell. The number density was calculated from the pressure of
the neutral reagent gas in the FT-ICR cell which was measured by an
ion gauge located between the FT-ICR cell and the turbomolecular
pump. Therefore, the readings of the ion gauge were not only corrected
for the sensitivity of ammonia as the neutral gas10 but also calibrated
for the pressure within the FT-ICR cell by measuring the rate constant
•+
+
for NH3 + NH3 f NH4 + NH2. A value of kbi ) 22 × 10-10 cm3
molecule-1 s-1 was used for the rate constant of this reaction.11 The
normalized efficiency (in %) of the ion/molecule reaction is given by
kbi/kcoll. The collision rate constant kcoll was calculated using the method
of Su and Chesnavich.12
Experimental Section
Compounds. Ammonia, vinyl chloride (1) and vinyl bromide (2),
and all other reagents are commercially available. The purity of all
compounds was checked by gas chromatography and mass spectrometry
and was in all cases > 99%. So all compounds were used without
further purification. Vinyl chloride-d3 radical cation (1-d3•+) was
generated in the ion source by electron impact induced fragmentation
of 1,2-dichloro-1,1,2,2-tetradeuterioethane which was prepared by the
following procedure.5
(6) Schaefer, I. D.; Dagani, M. J.; Weinberg, D. S. J. Am. Chem. Soc.
1967, 89, 6938.
(7) Allemann, M.; Kellerhals, H.; Wanczek, K. P. Int. J. Mass Spectrom.
Ion Processes 1983, 46, 139.
(8) Kofel, P.; Allemann, M.; Kellerhals, H.; Wanczek, K. P. Int. J. Mass
Spectrom. Ion Processes 1985, 65, 97.
(9) Caravatti, P.; Allemann, M. Org. Mass Spectrom. 1991, 26, 514.
(10) Bartmess, J. E.; Georgiadis, R. M. Vacuum 1983, 33, 149.
(11) Adams, N. G.; Smith, D.; Pulsion, J. F. J. Chem. Phys. 1980, 72,
288.
(5) Jensen, F. R.; Nees, R. A. J. Org. Chem. 1972, 37, 3037.
(12) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183.

6546 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Nixdorf and Gru¨tzmacher
The collision activation (CA) mass spectra of the product ions and
reference ions C2H6N+ were obtained by a double-focussing mass
spectrometer VG AutoSpec using 70 eV electron impact ionization,
an electron current of 200 µA, an ion source temperature of 180 °C,
and an accelerating voltage of 8 kV. The C2H6N+ product ions were
generated within the CI ion source by the admission of vinyl chloride
or vinyl bromide into the CI-plasma of ammonia gas at a suitable
pressure to maximize the yield of C2H6N+ product ions. The C2H6N+
reference ions were generated from electron impact-induced fragmenta-
tion of an appropriate precursor. The C2H6N+ ions were selected by
the magnetic sector and focused into the field-free region between the
magnetic sector and the second electrostatic analyzer of the AutoSpec
mass spectrometer. Helium was used as a collision gas and was
introduced at such a rate into the collision cell that the intensity of the
main ion beam was reduced to about 30% of its original value. The
CA mass spectra were obtained by scanning the deflecting voltage of
the electrostatic analyzer, and the relative abundances of the reported
CA spectra are the averages of at least 12 scans.
Computational Details. Ab initio molecular orbital calculations were
performed by the GAUSSIAN 92 program13 on a RS/6000 workstation
or a SNI S600/20 computer. The geometries of all species were fully
optimized at the Hartree-Fock level with the D95** basis set14 using
gradient procedures.15 For double-ú level calculations a Huzinaga basis
set16 for the bromine atom (13s10p4d/6s5p2d) was used with an
additional d-function exponent of 0.389. The spin-unrestricted Har-
tree-Fock formalism was used for the open- shell radical cations and
the restricted Hartree-Fock formalism for the closed-shell species. The
heat of formation of the vinyl halide radical cations and haloethenyl
radicals listed in Table 2 and 3 have been computed with the MOLPRO
92 suite of ab initio programs17 using a restricted reference wave
function for the geometry optimizations ((RF/DO(DP)) and correlation
contributions were considered on CERA-1/VTZ(f,2d,p) level of theory.
Details of these computations will be published elsewhere. Harmonic
vibrational wavenumbers were computed in order to characterize the
stationary points on the potential energy hyper surfaces as minima (for
equilibrium structures) or saddle points (for transition state structures)
and to estimate the zero-point vibrational energies Evib°. The latter
energies were scaled by an empirical factor of 0.9 to account for the
systematic overestimation of vibrational frequencies by the Hartree-
Fock calculations.18 The frozen core approximation FC-(U)MPn was
used for all perturbation calculations. Single point calculations were
performed at the (U)MP4(SDTQ)/D95** of theory. Unfortunately, a
spin-unrestricted wave function can contain contributions from un-
wanted spin states. These may distort significantly the potential
hypersurface. For that reason spin-projected FC-UMP4 energies
(PMP4) have been calculated to obtain improved values of the potential
energies. The absolute energies were corrected for the zero-point
vibrational energy contribution in order to calculate the reaction
enthalpies ∆Hr° at 0 K. The enthalpy of formation, ∆Hf298 and proton
affinity, PA, at 298 K were derived from the corresponding calculated
values using standard statistical thermodynamics.19
Figure 1. Ion intensity Vs reaction time plot for reaction of (a) 1•+
with NH3 and (b) 2•+ with NH3.
Table 1. Rate Constants, kbi, and Efficiencies (eff) of the Reaction
of Mono- and 1,2-Dihaloethene Radical Cations with NH3
a
a
compound
kbi
kcoll
eff %
H2CdCHCl•+
(1•+
(2•+
(3•+
(3•+
(4•+
)
)
12.3
7.2
12.5
12.5
8.5
21.6
20.6
20.8
20.8
20.0
57
35
60
60
43
H2CdCHBr•+
cis-HClCdCClH•+ b
trans-HClCdCClH•+ b
HBrCdCBrH•+
)
)
tr
cis
)
a In units of 10-10 cm3 molecule-1 s-1 b Recalculated from ref 2e.
.
+
generating C2H6N+ ions and NH4 ions as ionic products in
close analogy to the reactions of the 1,2-dihaloethene radical
cations.2e Figure 1a,b show the kinetic plot obtained for 1•+
and its deuterated derivative 1-d3. The bimolecular rate
constants kbi and the efficiencies of the reactions are collected
in Table 1 which includes also the previously determined rate
constants kbi for the reactions of the radical cations of cis- and
trans-1,2-dichloroethene, 3•+cis and 3•+tr, and 1,2-dibromoethene
4•+ (mixture of stereoisomers) with NH3.20 All reactions are
quite fast with an efficiency between 35% and 60%, and there
is no significant difference between the reactivity of the mono-
and 1,2-disubstituted haloethene radical cations. The reactions
of the mono- and dichloroethene radical cations are consistently
more efficient than those of the bromo derivatives. This unusual
result was confirmed by measuring directly the relative rates
with ammonia in a mixture of 1•+ and 2•+ to avoid any
ambiguity due to the determinations of the NH3 partial pressure.
The reason for the increased reactivity of the chloro ions is not
evident from these experiments alone but this result shows
clearly that the dissociation of the bond to the leaving group is
Results and Discussion
Kinetic Measurements. Vinyl chloride radical cations (1•+)
and vinyl bromide radical cations (2•+) react with ammonia,
(13) Frisch, M. J.; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.;
Wong, M. W.; Foresman, J. B.; Johnson, B. G.; Schlegel, H. B.; Robb, M.
A.; Binkley, E. S.; Gonzalez, C.; Martin, R. L.; Fox, D. J.; Detrees, F. J.;
Baker, J.; Stewart, J. J. P.; Pople, J. A. GAUSSIAN 92; Gaussian: Pittsburgh,
1991.
(14) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry;
Plenum: New York, 1976; Chapter 1, pp 1-28.
(15) Schlegel, H. B. J. Comput. Chem. 1982, 2, 214.
(16) Huzinaga, S. In Gaussian basis set for molecular calculations;
Elsevier: Amsterdam, 1984; p 217.
(17) MOLPRO is a package of ab initio programs written by H. J.
Werner, and P. J. Knowles, with contributions of J. Almlo¨f, R. Amos, S.
Elbert, W. Meyer, E. A. Reinsch, R. Pitzer, A. Stone, A. and C. Hampel,
C.
(18) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. In Ab
initio molecular orbital theory; John Wiley & Sons Inc.: New York, 1986.
(19) McQuarrie, D. A. Statistical thermodynamics; Harper & Row: New
York, 1973.
(20) The values of kbi and the reaction efficiency given in Table 1 for
the reaction of 3•+ with NH3 are slightly larger than those in ref 2e because
the deprotonation of 3•+ has been not included previously.

Reactions of Vinyl Halide Radical Cations with NH3
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6547
Table 3. Experimental and ab Initio-Calulated Heat of Formation,a
∆Hf298(C2H6N+) and Reaction Enthalpy,a ∆Hr298, for the Substitution
H2CdCHX•+ + NH3 f C2H6N+ + X•
Scheme 2
∆Hr298
X ) Cl
X ) Br
C2H6N+
product ion
∆Hf298(C2H6N+)
expb
expc calc expc calc
+
CH2dCHNH3 (5+)
695d
657
753
-124 -115 -171 -151
-163 -182 -210 -217
-67 -62 -114 -99
+
CH3CHdNH2 (6+)
(7+)
HN+H
H2C CH2
CH3NH+dCH2 (8+)
695
-126 -143 -172 -179
a In kJ/mol. b Reference 22. c The following experimental data were
Table 2. Heat of Formation,a ∆Hf(rad), of Haloethenyl Radicals,
and Reaction Enthalpya of Dissociation, ∆Hr298(diss), and Enthalpy
of Deprotonationa by Ammonia, ∆Hr298(dep), of Vinyl Halide
Radical Cations
22
used for the calculation: ∆Hf (1•+ and 2•+
)
(see footnote Table 2),
∆Hf(Cl•) ) 121 kJ/mol,22 and ∆Hf(Br•) ) 112 kJ/mol.22 d From this
work (see text).
radical
∆Hr298(diss)b
∆Hf298(rad)c
∆Hr298(dep)d
Structure of the C2H6N+ Product Ions by CID. The
replacement of the halogen substituent of 1•+ and 2•+ by NH3
without any rearrangement produces vinylammonium ions 5+.
Other possible structures of the C2H6N+ product ions are the
acetaldiminium ion (6+), the aziridinium ion 7+, and the
N-methylformaldiminium ion 8+. However, the formation of
ions 6+-8+ requires hydrogen rearrangements (6+ and 7+) or
substantial skeletal rearrangements (8+) concomitant to substitu-
tion. Only the heats of formation ∆Hf of 6+-8+ are known by
experiment.22 Therefore, for consistency the ∆Hf(C2H6N+) of
all isomers were calculated by ab initio methods, and these data
were used to calculate the heat of the substitution reaction, ∆Hr-
(sub), for 1•+ and 2•+. The results (Table 3) show that all
isomers 5+-8+ would arise by exothermic ion/molecule reac-
tions, and that formation of the acetaldiminium ion (6+) would
be particularly exothermic. Thus, a priori none of the structures
5+-8+ can be excluded for the C2H6N+ product ions. Further,
the reaction of the ionized vinyl bromide is distinctly more
exothermic as expected.
The isomeric ions 6+-8+ have been studied before by
experimental23,25 and theoretical24 methods, and it has been
shown that these isomers can be distinguished by collision
induced decomposition (CID). Ions 6+ and 8+ are generated
from ionized isopropylamine and dimethylamine by loss of
methyl and hydrogen, respectively, while the aziridinium ion
7+ is formed from â-phenoxyethylamine by electron-impact
ionization at low electron energies.25 The collisional activation
(CA) mass spectra of 6+-8+ obtained in this work are listed in
Table 4 together with the CA mass spectra of the C2H6N+
product ions from the reaction of 1•+ and 2•+ with NH3.
In agreement with the literature the ions 6+-8+ exhibit CA
mass spectra differing by the relative abundances of fragment
ions only.25 However, the CA mass spectra of the C2H6N+
product ions are distinguished from the other ions by an
abundant formation of ions m/z 17 (very likely NH3•+) and by
a signal of doubly charged ions C2H6N2+ (m/z 22) not present
in the other CA mass spectra. In particular the sharp signal of
the doubly charged ions C2H6N2+ generated by charge exchange
1-chlorovinyl
811
824
830
827
834
844
268
281
287
322
329
339
-41
-29
-23
-25
-18
-8
anti-2-chlorovinyl
syn-2-chlorovinyl
1-bromovinyl
anti-2-bromovinyll
syn-2-bromovinyl
a In kJ/mol. b Ab initio-calculated dissociation enthalpies ∆Hr298(diss)
for H2CdCHX•+ f radical + H+ (X ) Cl, Br; ∆Hf(H+) ) 1530 kJ/
mol22). c Derived from ∆H298(diss) and ∆Hf298(CH2dCHCl•+) ) 987
kJ/mol22 and ∆Hf298(CH2dCHBr•+) ) 1025 kJ/mol.22 d Reaction
enthalpy, ∆Hr298(dep), of deprotonation (∆Hf(NH3) ) -46 kJ/mol).22
not rate determining for the total substitution process. As will
be discussed later this effect arises from an increased chemical
activation by the initial addition of NH3 to the vinyl chloride
radical cation 1•+.
Of the two product ions observed for the reactions of 1•+
([C2H6N+]/[NH4+] ) 2.9) and 2•+ ([C2H6N+]/[NH4+] ) 2.6)
with NH3 the ions C2H6N+ arise from the substitution of the
+
halogen substituent by NH3, while NH4 is formed by proton
transfer either from the haloethene radical cation or from the
substitution product ion to NH3. However, the C2H6N+ product
ions do not disappear even at prolonged reaction time due to
deprotonation by the excess of NH3 present in the FT-ICR cell,
+
and no NH4 ions are observed if C2H6N+ product ions are
isolated in the FT-ICR cell and then exposed to NH3. Further,
the reactions of 1-d3 and 2-d3 both give rise initially to
NH3D+ which exchanges quickly with NH3,21 yielding NH4
•+
•+
+
.
These experiments prove conclusively a direct proton transfer
from the ionized haloethenes to NH3 (Scheme 2). It was not
possible to defer the NH4+ formation by prolonged cooling of
the haloethene radical cations with argon indicating that the
proton transfer from 1•+ and 2•+ to NH3 is very likely
exothermic. For an estimation of the reaction enthalpy of this
proton transfer the heats of formation ∆Hf(rad) of all possible
chloro- and bromovinyl radicals have to be known. These were
obtained by an ab initio calculation of the absolute dissociation
energy of 1•+ and 2•+ into the 1- or 2-halovinyl radicals and
converting these energies into the enthalpies of reaction at 298
K by standard methods of statistical thermodynamics (Table
2). Using the experimental values for ∆Hf298(1•+) ) 987.4 kJ/
mol,22 ∆Hf298(2•+) ) 1024.7 kJ/mol,22 ∆Hf (NH3) ) -46 kJ/
mol,22 and ∆Hf(NH4+) ) 631.8 kJ/mol22 the deprotonation of
1•+ and 2•+ by NH3 is indeed exothermic for all halovinyl
radicals as the neutral product, although the 1-haloethenyl
radicals are distinctly more stable than either the syn- or anti-
2-haloethenyl radicals as expected.
(22) Lias, S. G.; Liebermann, J. F.; Holmes, J. L.; Levin, R. D.; Mallard,
W. G. J. Phys. Chem. Ref. Data 1988, 17, Suppl.1.
(23) (a) Ellenberger, M. R.; Dixon, D. A.; Farneth, W. E. J. Am. Chem.
Soc. 1981, 103, 5377. (b) Eades, R. A.; Weil, D. A.; Ellenberger, M. R.;
Dixon, D. A.; Farneth, W. E.; Douglas, C. H. J. Am. Chem. Soc. 1981,
103, 5372. (c) Bowen, R. D.; Williams, D. H. ; Hvistendal, G. J. Am. Chem.
Soc. 1977, 99, 7509. (d) McLafferty, F. W.; Levsen, K. J. Am. Chem. Soc.
1974, 96, 139.
(24) (a) Barone, V.; Lelj, F.; Grande, P.; Russo, N.; Toscano, M. Chem.
Phys. Lett. 1987, 133, 548. (b) Ellenberger, M. R.; Dixon, D. A.; Farneth,
W. E. J. Am. Chem. Soc. 1979, 101, 7151. (c) Mu¨ller, K. Angew. Chem.
1980, 92, 1.
(21) Harrison, A. G.; Lin, D.-H.; Tsang, C. W. Int. J. Mass Spectrom.
Ion Processes 1976, 19, 23.
(25) Van de Sande, C. C.; Ahmad, S. Z.; Borcher, F.; Levesen, K. Org.
Mass Spectrom. 1978, 13, 666.

6548 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Nixdorf and Gru¨tzmacher
Table 4. CA-Mass Spectraa of C2H6N+ Ions 6+-8+and of NH3 Substitution Product Ions of 1•+ and 2•+
m/z
precursor
ion
43
8.1 17.0 11.6 9.0 4.5 2.8 0.5 3.0 20.4
4.9 12.6 8.3 7.5 4.5 3.0 1.3 3.9 40.1
4.4 8.9
10.4 12.1 10.2 8.3 4.1 2.1 0.8 2.5 15.5
9.8 12.5 10.7 7.0 3.5 2.5 0.8 2.5 16.0 11.7 8.5 4.7 0.5 1.0 5.6 4.2
42
41
40
39
38
30
29
28
27
26
25
24
22
18
17
•+
(CH3)2CHNH2
(CH3)2NH2
6+
8+
7+
5+
5+
8.4 6.4 1.5
9.2 0.8
6.0 0.7
3.8
6.2
•+
•+ b
C6H5OCH2CH2NH2
6.1 5.8 2.8 2.8 5.8 2.3 31.2 11.7 9.1 2.7
H2CdCHCl•+ + NH3
H2CdCHBr•+ + NH3
9.5 9.4 2.5 0.6 1.2 7.0 6.4
a Intensity normalized to sum of fragment ion intensity. b Generated by 14 eV electron impact.
Table 5. Experimental and Calculated Proton Affinitiesa of
Conjugated Bases C2H5N of ions 5+-8+
PA
compound, C2H5N
ion
calcb
exp
vinylaminee
5+
927
866
905
915
891
868
vinylaminef
5+
864g
893c
903d
acetaldimine
aziridine
N-methylformaldimine
ammonia
6+
7+
8+
+
NH4
854d
a In kJ/mol. b By MP4/6-311++G** and corrected for thermody-
namical contributions. c References 23a. d Derived from data in ref 22.
e C-protonation. f N-protonation. g From this work.
is a very characteristic feature of the CA mass spectrum of the
C2H6N+ product ions only. Hence, these CA mass spectra prove
unambiguously the presence of C2H6N+ ions with a structure
different from 6+-8+, but it is difficult to decide whether the
C2H6N+ product ions consist only of a single species or represent
a mixture of isomers.
Structure of the C2H6N+ Product Ions by Gas-Phase
Titration. More conclusive information about the structure of
the C2H6N+ product ions is obtained by their “gas-phase
titration” 26 with a series of reference bases of appropriate gas-
phase basicity (GB). In these titration experiments the kinetics
of the deprotonation of a gaseous mixture of isomeric ions by
a set of reference bases of known GB are measured in separate
experiments. A fast and efficient proton transfer is observed
only for exoergic proton transfer. Thus, if the GB of the
reference base is placed between the GB of the neutral
conjugated bases of the ions in the mixture, bimodal kinetics
are observed for the deprotonation because only the more acidic
ions will react efficiently. The bimodal kinetic curve obtained
may be examined by curve-fitting procedures to separate the
ion mixture into two sets of isomeric ions with their corre-
sponding conjugate bases being less and more basic than the
reference base. Repeating this experiment with a series of
reference bases allows a semiquantitative analysis of the ion
mixture and an estimation of the GB (or PA) of the conjugated
bases of the unknown ions by bracketing. This PA determi-
nation is very advantageous for an identification of ions because
many PA are tabulated or may be calculated reliably by ab initio
methods.
Figure 2. Plot of ion intensity Vs reaction time for the deprotonation
of C2H6N+ product ions by (a) 3-fluoroaniline (PA 866 kJ/mol22); (b)
2-phenylpropene (PA 866 kJ/mol22).
of Eades et al.23b Taking into account the systematic error of
the calculated PA, a PA(vinylamine) of about 860 kJ/mol is
expected for N-protonation, slightly larger than the PA(NH3 ).
The PA data (Table 5) predict fast deprotonation of 5+ and
6+, but not of 7+, by bases of PA < 903 kJ/mol. A complete
and efficient deprotonation without any side reaction is observed
during the reaction of the C2H6N+ product ions from the
reactions of 1•+ and 2•+ with the strong base tributylamine (PA
985 kJ/mol22). Thus, no C2H6N+ product ions that are reluctant
to deprotonation are present in the ion mixture. A complete
deprotonation of the C2H6N+ product ions is also observed by
the weaker base 2-chloropyridine (PA 897 kJ/mol22). Bimodal
kinetics are observed for this process, but the slower process is
too fast for the endoergic deprotonation of the aziridinium ion
7+ and is more compatible with the nearly thermoneutral
deprotonation of 6+ (or probably 8+). Further decreasing the
PA of the titration base results in a fast deprotonation of only
one component of the ion mixture. This is still observed for
reaction with 3-fluoroaniline (PA 866 kJ/mol;22 Figure 2a), but
no deprotonation occurs by N-methylformamide (PA 861 kJ/
mol22). This sets the PA of the least basic component of the
mixture to 864 ( 2 kJ/mol, in reasonable agreement with the
estimated PA for N-protonation of vinylamine. Thus, the gas-
The PA of acetaldimine23a and aziridine22 are known, but the
PA of vinylamine and N-methylformimin have not been
determined directly by experiment.23b Therefore, for consistency
the PA of all precursors of 5+-8+ were calculated by ab initio
methods (Table 5). A comparison of the experimental and
calculated PA in the case of acetaldimine (for 6+) and aziridine
(for 7+) shows that the PA calculated by MP4/6-311++G**
are overestimated by ∼12 kJ/mol. For vinylamine both the PA
for a protonation at the terminal C-atom (yielding 6+) and at
the amino group (yielding 5+) were calculated. C-protonation
is favored by 61 kJ/mol, in good agreement with the calculations
(26) Bu¨ker, H.-H.; Gru¨tzmacher, H.-F.; Ricci, A.; Crestoni, M. Int. J.
Mass Soectrom. Ion Processes 1997, 160, 167-182.

Reactions of Vinyl Halide Radical Cations with NH3
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6549
Scheme 3
phase titration experiments indicate that the C2H6N+ product
ions are apparently a mixture of vinylammonium ions (5+) and
acetaldiminium ions (6+). The additional presence of the
N-methylformamidinium ion (8+) in the product ion mixture
cannot be excluded completely. However, in view of the
extensive skeletal rearrangements necessary to generate 8+
during the substitution reaction the presence of this ion is not
very likely.
Scheme 4
However, the quantitative composition of the C2H6N+ ion
mixtures varies considerably with the PA and the chemical
nature of the reference base used for the gas-phase titration. As
will be shown elsewhere27 this is due to a base catalyzed
tautomerization of 5+ into 6+ (Scheme 3). In the case of PA-
(5) < PA(base) < PA(6), the base picks up a proton from 5+
within the initial collision complex formed by 5+ and a molecule
of the base and generates a new product complex consisting of
the neutral vinylamine and the protonated base. This product
complex may dissociate or return the proton to the terminal C
atom of the neutral vinylamine generating the more stable
acetaldiminium ion (6+) in an exothermic process. Clearly, the
amount of isomerization 5+ f 6+ depends on the branching
ratio of these competing processes. A similar “shuttle mech-
anism” has been described before for small molecules with two
basic sites.28 However, to our knowledge this is the first time
that such a strong influence not only of the PA of the catalytic
base but also of its chemical structure has been observed. An
interesting example is the titration of the C2H6N+ product ions
with 3-fluoroaniline and 2-phenylpropene, respectively, shown
in Figure 2a,b. The PA of 3-fluoroaniline and 2-phenylpropene
are identical,22 but 2-phenylpropene is a “bad proton shuttle”,
and only less than 13% of the C2H6N+ product ions have
isomerized into 6+ and are not deprotonated, compared to 65%
in the case of the “good proton shuttle” 3-fluoroaniline. It is
difficult to find a similar isomerization mechanism which
involves also the N-methylformamidinium ion (8+). Therefore,
these experiments evidence that ions 8+ are absent, that the
presence of acetaldiminium ions (6+) among the C2H6N+
product ions is very likely an artefact of the titration experiment,
and that initially only vinylammonium ions (5+) are formed
during the reaction of 1•+ and 2•+ with NH3.
5+ by loss of Cl• or Br•, while any energetically appropriate
fragmentation of the Markovnikov addition products 1a•+ and
2a•+ requires rearrangements. These rearrangements correspond
either to hydrogen shifts which result eventually in the molecular
ions 1e•+ and 2e•+ of 1-haloethylamines which decompose
subsequently into acetaldiminium ions 6+, or to a 1,2-shift of
the NH3 group converting 1a•+ and 2a•+ into the anti-
Markovnikov products 1b•+ and 2b•+ for further reaction to 5+.
The enthalpies of reaction, ∆Hr, of the individual reaction
steps, calculated by ab initio methods (see Computational
Details) have been used to construct the reaction energy profiles
for the reaction of the ionized vinyl halides 1•+ and 2•+. The
reaction energy profile for the substitution reactions of the
ionized vinyl chloride 1•+ and vinyl bromide 2•+ are very similar
with an exception to be discussed below. The reaction energy
profile of 1•+ is shown in Figure 3. The structure of lowest
energy found for the addition complex of NH3 to 1•+ is the
Markovnikov adduct 1a•+ in which the radical site of the distonic
ion is stabilized by the adjacent chloro atom. By the level of
theory used no activation barrier was found for the addition
step. This agrees with ab initio calculations of the polymeri-
zation of ethene radical cations, which were found to add to a
neutral ethene molecule without a barrier.29 The anti-Mar-
kovnikov adduct 1b•+ is 30 kJ/mol higher in energy than 1a•+,
and this distonic ion 1b•+ dissociates without a barrier by
elongation of the C-Cl bond into the final products vinyl-
ammonium ion (5+) and Cl•. For these products the substitution
of 1•+ is exothermic by -116 kJ/mol. The formation of the
acetaldiminium ion (6+) would be exothermic by -184 kJ/mol.
Reaction Energy Profile. For a more explicit picture of the
reaction steps occurring in the collision complex of 1•+ and 2•+
with NH3 before dissociation either back to reactants or to the
final products the possible reactions pathways have been
established by the ab initio methods discussed in Computational
Details. A key feature of the resulting mechanism is the
regioselective addition of the NH3 during the first step which
may occur either at the R-position (Markovnikov-addition) or
at the ipso-position (anti-Markovnikov addition) to the halogen
substituent producing the distonic ions 1a•+ and 1b•+ or 2a•+
and 2b•+ (Scheme 4).
However, starting from the anti-Markovnikov adduct 1b•+
a
formation of 6+ requires an 1,3-hydrogen shift from the
ammonium moiety to the radical site, generating the molecular
ion of 1-chloroethylamine 1e•+ which is higher in energy by
30 kJ/mol than the â-distonic ion 1b•+. The calculations indicate
Of these distonic ions only the anti-Markovnikov products
1b•+ or 2b•+ dissociate directly into the vinylammonium ions
(27) Nixdorf, A.; Bu¨chner, M.; Gru¨tzmacher, H.-F.; To be published.
(28) Bohme, D. K. Int. J. Mass Spectrom. Ion Processes 1992, 115, 95.
(29) Alvarez-Idaboy, J.; Erikson, L. A.; Fa¨ngstro¨m, T. Lunell, S. J. Phys.
Chem. 1993, 97, 12737.

6550 J. Am. Chem. Soc., Vol. 119, No. 28, 1997
Nixdorf and Gru¨tzmacher
Figure 3. Reaction energy profile of the reaction of vinylchloride 1•+ with NH3.
that likely the molecular ion 1e•+ is kinetically unstable and
decomposes without a barrier into 6+ and Cl•, but the rate-
determining step for this route of a substitution is the H
rearrangement. As anticipated from the behavior of other
distonic ammonium ions30,31 the 1,3-hydrogen shift in 1b•+ via
transition state 1d+ requires the large activation energy of 153
kJ/mol. Thus, in view of the dissociation into 5+ by direct loss
of Cl• the intermediate 1b•+ has no chance to realize the
acetaldiminium ion 6+ as the more stable product ion by the
hydrogen rearrangement pathway although transition state 1d•+
is still 15 kJ/mol below the energy limit established by the
reactants 1•+ and NH3.
1,2-NH3 shift, the transition state 1c•+ is still 83 kJ/mol below
the energy limit defined by the reactants. Therefore, intercon-
version 1a•+ h 1c•+ h 1b•+ is clearly an alternative to back
dissociation of 1a•+. In this case, the reaction efficiency of the
substitution is determined by the competition between back-
dissociation and rearrangement of 1a•+ into 1b•+ , and the
transition state 1c•+ corresponds to the bottle neck of the total
substitution reaction because of its tight cyclic structure.
As mentioned above the reaction energy profile for the
reaction of the bromo derivative 2•+ with NH3 is qualitatively
identical to that of the chloro ion 1•+. In particular, the
activation barriers for the 1,2-NH3 shift converting the
Markovnikov addition product 2a•+ into the anti-Markovnikov
product 2b•+ and for the hydrogen rearrangement routes for the
rearrangements of 2a•+ or 2b•+ are identical to the case of the
chloro derivative. Obviously, the exchange of the Cl substituent
by Br has no effect on the relative energy of the transition state
especially of the 1,2-NH3 shift. However, quantitatively the
reaction of the vinyl bromide radical cation (2•+) with NH3
differs from that of the chloro derivative 1•+ by the reaction
enthalpy of the total process, which is more exothermic for 2•+
as expected from the cleavage of the weaker C-Br bond, and
the reaction enthalpy of the first addition step 2•+ + NH3 f
2a•+ which is distinctly less exothermic (-184 kJ/mol) than
the addition of NH3 to ionized vinyl chloride (-198 kJ/mol).
This is a consequence of the better stabilization of the radical
site by Cl than by Br. In line with this explanation the energies
of the anti-Markovnikov adduct ions 1b•+ and 2b•+ relative to
the reactants are again nearly identical, since the Cl or Br
substituent is now in the remote â-position to the radical site.
As a consequence of the different reaction enthalpy of the
initial addition step, the bromo derivative 2a•+ is less chemically
activated than chloro derivative 1a•+ by the preceding addition
step. If both intermediates have to rearrange into the reactive
anti-Markovnikov product ions 1b•+ and 2b•+ for the substitution
to occur, they have to pass identical activation barriers and very
similar transition-state structures on their way to the products
5+ and Cl• or Br•. Thus, the energetically less excited bromo
derivative 2a•+ is expected to react with a lower reaction
efficiency in spite of the increased exothermicity of the total
process. In fact, preliminary RRKM calculations using the
structures and energies calculated for 1a•+ and 1c•+ or 2a•+ and
2c•+ indicate that the calculated difference of 14 kJ/mol of
Although the anti-Markovnikov adduct 1b•+ decomposes
without any significant energy barrier to the products 5+ and
Cl•, the initial intermediate formed regioselectively by the
addition of NH3 to 1•+ is certainly the more stable Markovnikov
adduct 1a•+. Since 1a•+ cannot produce the substitution
products directly, it renders either back-dissociation or rear-
rangements to an isomeric distonic ion more disposed for the
loss of Cl•. In the first case, the preferred formation of 1a•+
leads to a “dead end” of the substitution reaction, and only that
fraction of the collision complexes which eventually breaks
down into the anti-Markovnikov adduct 1b•+ will react by loss
of Cl•. Thus, the reduced reaction efficiency of the substitution
is determined by the regioselectivity of the addition step and
the ratio 1a•+/1b•+ initially formed. In the second case, 1a•+
may rearrange by a 1,2-hydrogen shift into an R-distonic ion
+
ClCH2CHNH3 as an intermediate suitable for the loss of Cl•,
or may interconvert into 1b•+ by an 1,2-shift of the NH3 group.
An ab initio calculation of the activation energy for the
rearrangement of 1a•+ into the R-distonic ion by an 1,2-hydrogen
shift results in a prohibitively large barrier of 186 kJ/mol for
this process, in agreement with the large activation energies
reported for 1,2-hydrogen shifts in other distonic ions.30 Hence,
this rearrangement is excluded. The transition state 1c•+
connecting 1a•+ and 1b•+ by the 1,2-NH3 shift is located 115
kJ/mol above 1a•+. The calculated activation energy for the
1,2-NH3 shift as well as the structure of the transition state 1c•+
compare well with the results of L. Radom et al. for the 1,2-
NH3 shift in the â-distonic isomer of the ethylamine radical
cation.31 In spite of the quite large activation energy of the
(30) Hammerum, S. Mass Sepctrom. ReV. 1988, 7, 123.
(31) Yates, B. F.; Radom, L. Org. Mass Spectrom. 1987, 22, 430.

Reactions of Vinyl Halide Radical Cations with NH3
J. Am. Chem. Soc., Vol. 119, No. 28, 1997 6551
excess energy between the distonic ions 1a•+ and 2a•+ is clearly
sufficient to explain a difference in the reaction efficiencies of
about a factor 2 experimentally observed for the reactions of
chloroalkene and bromoalkene radical cations.
Further, it has been shown by gas-phase titration experiments
of the C2H6N+ ions arising from the substitution of the halogen
substituent by NH3 that these ions correspond to vinylammo-
nium ions 5+ and that the much more stable acetaldiminium
ions (6+) are not formed. Subsequent isomerization of 5+ into
6+ by a 1,3-sigmatropic H shift is blocked by a large activation
barrier. Interestingly however, this isomerization is catalyzed
in a collision complex of the vinylammonium ion (5+) with a
suitable base by a shuttle mechanism. This base-induced proton
shuttle is very likely a general mechanism for the isomerization
of large protonated and multifunctional molecules because of
the rather long lifetime expected for the respective collision
complexes. In the present case obviously only vinylammonium
ions (5+) arise from the reaction of ammonia with ionized vinyl
halides by an ipso-substitution of the halogen substituent most
probably by an addition/elimination mechanism.
Conclusion
The results of the investigation of the reactions of vinyl halide
radical cations with ammonia by a combination of FT-ICR
spectrometry, tandem-mass spectrometry, and by ab initio
calculation of the reaction energy profiles reveal further interest-
ing aspects of the mechanisms of unsaturated organic radical
cations with nucleophiles. Deprotonation and substitution of
the halogen substituent, which are observed experimentally, turn
out to be both exothermic. Nevertheless, it is remarkable that
an entropically quite demanding substitution can compete with
such a simple process as proton transfer and is in fact the main
reaction pathway for vinyl halide radical cations at thermal
energies. Obviously, the addition of the nucleophile NH3 to
the double bond of the ionized vinyl halide proceeds quickly
in the collision complex without any noticeable energy barrier
and severe steric constraints. Indeed, ab initio calculations
detect no activation barrier for the addition of NH3 to the double
bond of the ionized alkene, in marked contrast to the reactions
of aromatic radical cations.2d
Nonetheless, the efficiencies of the vinyl halide radical cations
with NH3 are distinctly below the collision rate and are
significantly smaller for the ionized vinyl bromide 2a•+ than
for ionized vinyl chloride 1a•+. This effect is due to differences
of the chemical activation of the distonic ions generated by the
initial addition step. The addition of NH3 to ionized vinyl
chloride is more exothermic than addition to ionized vinyl
bromide because of the better stabilization of the radical site in
the chloro-distonic ion. Hence, in this latter ion more excess
energy is available for the subsequent reaction steps than in
the bromo derivative. This result reinforces the fact that the
reactions of the radical cations of unsaturated organic com-
pounds with nucleophiles in the diluted gas phase are dominated
by chemically activated species. Thus, the rate constants and
dynamics of these reactions are expected to depend strongly
also on entropic effects.
The analysis of the reaction energy profiles of the substitution
reaction of ammonia with vinyl halide radical cations calculated
by ab initio methods suggests that the reduced efficiency of
the substitution is due to a regioselective addition of ammonia
to the R-C atom of the ionized vinyl halide (Markovnikov
orientation). This preferred Markovnikov addition requires an
additional rearrangement step to arrive at the substitution
product. The least energy demanding rearrangement corre-
sponds to a 1,2-shift of the ammonium group generating the
anti-Markovnikov isomer of the addition product. This rear-
rangement appears to be the bottle neck of the substitution
reaction yielding eventually the vinylammonium ion.
Acknowledgment. The financial assistance of this work by
the VW-Stiftung (Schwerpunktprogramm “Elektronentransfer
Reaktionen”) is gratefully acknowledged. We thank the Deut-
sche Forschungsgemeinschaft for the gift of the FT-ICR
instrument and the Fonds der Chemischen Industrie for ad-
ditional financial help. We are obliged to Professor Dr. H.-J.
Werner and to Dipl. Chem. C. Hampel for performing the
calculations of the halovinyl radicals. Computer time on a SNI
S600/20 was provided by the Rechenzentrum (RWTH) Aachen.
JA9619465