Study of radical merostabilization of electrospray FTICR/MS

Article abstract of DOI:10.1021/ja962027h

The threshold fragmentation energies (E(o)) of three different 4-(1'-substituted-2'-phenethyl)-1-methylpyridinium salts containing a neutral, an electron-donor, or an electron-acceptor group as α-substituent, respectively, were measured by Fourier Transfo

Full text of DOI:10.1021/ja962027h

J. Am. Chem. Soc. 1996, 118, 11905-11911
11905
Study of Radical Merostabilization by Electrospray FTICR/MS
Alan R. Katritzky,*,§ Petia A. Shipkova,§ Ming Qi,§ Daniel A. Nichols,§
Richard D. Burton,§ Clifford H. Watson,§ John R. Eyler,*,§ Toomas Tamm,§
Mati Karelson,*,‡ and Michael C. Zerner*,§
Contribution from the Department of Chemistry, UniVersity of Florida, P.O. Box 117200,
GainesVille, Florida 32611-7200, and the Department of Chemistry, UniVersity of Tartu,
2 Jakobi Str., Tartu, EE 2400, Estonia
ReceiVed June 17, 1996X
Abstract: The threshold fragmentation energies (Eo) of three different 4-(1′-substituted-2′-phenethyl)-1-methylpy-
ridinium salts containing a neutral, an electron-donor, or an electron-acceptor group as R-substituent, respectively,
were measured by Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FTICR/MS) collisionally activated
dissociation (CAD). N-Methyl-4-(1-ethoxy-2-phenylethyl)pyridinium iodide (10), containing both electron-donor
and electron-acceptor substituent groups, has a significantly lower Eo than the analogs containing a benzyl (6) or
benzoyl (7) substituent. This was ascribed to merostabilization of the corresponding radical 19, and this conclusion
was further supported by theoretical calculations.
Introduction
the merostabilization concept was not universally valid. Thus,
Ru¨chardt and co-workers provided evidence suggesting that
stabilization effects are additive and not synergistic;9a-d Korth
et al. indicated the absence of kinetic stabilization for some
radicals by measuring their absolute rates for dimerization by
ESR spectroscopy;10 Chambers et al. revealed that the syner-
gistic effect was not dominant in systems containing polyfluo-
roalkyl groups11 and certain theoretical calculations failed to
reveal any effect.12
Quantum mechanical calculations have been used to study
merostabilization in radicals. Previous theoretical calculations
from this group suggested that there was significant merosta-
bilization energy in polar media, but not in the gas phase.13a-c
Pasto14 also found considerable extra stabilization in allyl-type
three electrons radicals but not in other systems. Calculations
had been carried out in water for carbon-centered radicals15 and
nitrogen-centered radicals,16 all of which confirmed merosta-
bilization in various situations. ESR spectroscopy was widely
employed in those studies.3c,17-19 Bordwell and co-workers
determined radical stabilization energies by equilibrium acidi-
ties.20 Most of this research was qualitative or semi-quantitative.
It is well-known that electron-withdrawing substituents
stabilize carbon anions and electron-donating substituents
stabilize carbon cations. Dewar first suggested1 that radicals
should be particularly strongly stabilized when both an electron-
attracting and an electron-donating substituent are present at
the radical site. Katritzky2a-d provided the first experimental
evidence for such carbon-containing radicals and proposed the
term “merostabilization” to describe this concept. Balaban3a-c
independently developed the analogous concept of “push-pull”
for nitrogen centered radicals. Later, Viehe4a-c entered the field
denoting these effects as “captodative”. Merostabilization has
been explained in terms of qualitative valence bond, molecular
orbital and Linnett double quartet theories and supported by
the prediction and synthesis of new stable radicals.2b-d,3ac,4b The
concept provides a fundamental model for a better understanding
of many properties of organic compounds, such as the chro-
mophoric systems of indigo5 and many heterocycles.6
However, further research on the stabilization of radicals by
the synergistic interaction of substituents has given rise to much
controversy. Whereas many research results supported the
synergistic donor/acceptor stabilization7,8 others suggested that
(9) (a) Zamkanei, M.; Kaiser, J. H.; Birkhofer, H.; Beckhaus, H.-D.;
Ru¨chardt, C. Chem. Ber. 1983, 116, 3216. (b) Birkhofer, H.; Ha¨drich, J.;
Beckhous, H.-D.; Ru¨chardt, C. Angew. Chem., Int. Ed. Engl. 1987, 26, 573.
(c) Beckhous, H.-D.; Ru¨chardt, C. Angew. Chem., Int. Ed. Engl. 1987, 26,
770. (d) Rakus, K.; Verevkin, S. P.; Keller, M.; Beckhaus, H.-D.; Ru¨chardt,
C. Liebigs Ann. 1995, 1483.
§ University of Florida.
‡ University of Tartu.
X Abstract published in AdVance ACS Abstracts, November 1, 1996.
(1) Dewar, M. J. S. J. Am. Chem. Soc. 1952, 74, 3353.
(2) (a) The early development of these ideas is contained in the MSc.
Theses of R. W. Baldock (University of East Anglia, 1965) and of P. Hudson
(University of East Anglia, 1971) (see also Ph.D. Thesis of P. Hudson,
University of East Anglia, 1973). (b) Baldock, R. W.; Hudson, P.; Katritzky,
A. R.; Soti, F. Heterocycles 1973, 1, 67. (c) Baldock, R. W.; Hudson, P.;
Katritzky, A. R.; Soti, F. J. Chem. Soc., Perkin Trans. I 1974, 1422. (d)
Katritzky, A. R.; Soti, F. J. Chem. Soc., Perkin Trans. I 1974, 1427.
(3) (a) Balaban, A. T. ReV. Roum. Chim. 1971, 16, 725. (b) Negoita, N.;
Baican, R.; Balaban, A. T. Tetrahedron 1974, 30, 73. (c) Balaban, A. T.;
Caproiu, M. T.; Negoita, N.; Baican, R. Tetrahedron 1977, 33, 2249.
(4) (a) Stella, L.; Janousek, Z.; Mere´nyi, R.; Viehe, H. G. Angew. Chem.,
Int. Ed. Engl. 1978, 17, 691. (b) Viehe, H. G.; Mere´nyi, R.; Stella, L.;
Janousek, Z. Angew. Chem., Int. Ed. Intl. 1979, 18, 917. (c) Viehe, H. G.;
Janousek, Z.; Mere´nyi, R.; Stella, L. Acc. Chem. Res. 1985, 18, 148.
(5) Klessinger, M. Angew. Chem., Int. Ed. Intl. 1980, 19, 908.
(6) Katritzky, A. R.; Fan, W.-Q.; Li, Q.-L.; Bayyuk, S. J. Heterocycl.
Chem. 1989, 26, 885.
(10) Korth, H.-G.; Sustmann, R.; Mere´nyi, R.; Viehe, H. G. J. Chem.
Soc., Perkin Trans. II 1983, 67.
(11) Chambers, R. D.; Grievson, B.; Kelly, N. M. J. Chem. Soc., Perkin
Trans. I 1985, 2209.
(12) Moyano, A.; Olivella, S. J. Mol. Struct. 1990, 208, 261.
(13) (a) Katritzky, A. R.; Zerner, M. C.; Karelson, M. M. J. Am. Chem.
Soc. 1986, 108, 7213. (b) Karelson, M.; Tamm, T.; Katritzky, A. R.; Szafran,
M.; Zerner, M. C. Int. J. Quantum Chem. 1990, 37, 1. (c) Karelson, M.;
Katritzky, A. R.; Zerner, M. C. J. Org. Chem. 1991, 56, 134.
(14) Pasto, D. J. J. Am. Chem. Soc. 1988, 110, 8614.
(15) Takase, H. and Kikuchi, O. J. Mol. Struct. 1994, 306, 41.
(16) Kost, D.; Raban, M.; Aviram, K. J. Chem. Soc., Chem. Commun.
1986, 346.
(17) Korth, H.-G.; Lommes, P.; Sustmann, R. J. Am. Chem. Soc. 1984,
106, 663.
(18) Maclnnes, I.; Walton, J. C. J. Chem. Soc., Perkin Trans. II 1987,
1077.
(19) Himmelsbach, R. J.; Barone, A. D.; Kleyer, D. L.; Koch, T. H. J.
Org. Chem. 1983, 48, 2989.
(20) Bordwell, F. G.; Zhang, X.-M. Acc. Chem. Res. 1993, 26, 510.
(7) Humphreys, R. W. R.; Arnold, D. R. Can. J. Chem. 1979, 57, 2652.
(8) Crans, D.; Clark, T.; Schleyer, P. v. R. Tetrahedron Lett. 1980, 21,
3681.
S0002-7863(96)02027-6 CCC: $12.00 © 1996 American Chemical Society

11906 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Katritzky et al.
We now report a quantitative evaluation of the merostabilization
concept by Fourier Transform Ion Cyclotron Resonance Mass
Spectrometry (FTICR/MS) in the gas phase.
Scheme 1. Synthesis of Compounds
We have previously measured appearance potentials21a-e from
the threshold energies for fragment ion appearance in the
collisionally activated dissociation (CAD) pathways of a variety
of pyridinium cations. Most of this work utilized laser-
desorption FTICR. Ultrahigh mass resolution, accurate and
stable mass calibration, multiple stages of collisionally activated
dissociation allowing tandem MS/MS, and extremely long ion
storage times (>1 min) combine to make FTICR/MS a powerful
method for studying organic cations in the gas phase. Most
recently, we demonstrated that electrospray ionization coupled
with FTICR/MS provides an excellent way to produce gaseous
pyridinium ions from pyridinium salts in solution and to monitor
the chemical behavior of the ions.22 In studies of collisionally
activated dissociation of pyridinium cations21a,23,24 the N-
methylpyridinium cation was found to be very stable, undergo-
ing essentially no dissociation. Therefore, we selected a set of
N-methylpyridinium salts as model compounds to investigate
radical merostabilization.
Scheme 2. Fragmentation Patterns of Compound 6
Results and Discussion
Our general strategy was to investigate the way in which the
energy for the homolytic scission of the 4-(â-phenethyl)-
pyridinium cation would be affected by the presence of an
electron donor or an electron acceptor substituent. If the
merostabilization concept is correct, a suitably situated electron
donor substituent should stabilize the pyridiniumylmethyl radical
(Py+-CH.-D) and hence facilitate the scission. The assumption
here is that the C-C bond scission in all processes follows a
similar potential energy curve, and that stabilization of the
product determines the barrier; i.e. the intersection of the C-C
bond breaking curve with the product formation curve. Thus,
according to the Bell-Evans-Polanyi principle, the stabilization
of the final state leads to the decrease of the transition barrier
of the reaction.
experiments were performed on the pyridinium cation 6, the
spectra showed three major ions (Scheme 2): (i) the parent
cation 11 at m/z 288, (ii) the radical cation of interest 12 at m/z
197 generated Via loss of a benzyl radical from the parent ion,
and (iii) cation 13 at m/z 196, obtained Via loss of a hydrogen
atom from the radical 12. The ion at m/z 196 may possess the
cyclic structure 13, formed Via ring closure as shown on Scheme
2 rather than compound 14 with a double bond. Fragmentation
leading to the formation of stable cyclic structures was previ-
ously observed by us in an independent study of singly and
multiply charged pyridinium cations.22 When the ion at m/z
197 was ejected, the ion at m/z 196 disappeared, proving that
the ion at m/z 196 is formed from the radical at m/z 197 and
not directly from the parent ion at m/z 288.
N-Methyl-4-(1,3-diphenyl-1-oxo-2-propyl)pyridinium iodide
(7) (MW 429) was chosen as a model compound to study the
effect of an electron withdrawing group (benzoyl) on the
appearance potential. The CAD spectrum showed 5 major ions
(Scheme 3): (i) the parent cation 15 at m/z 302, (ii) the radical
cation 16 at m/z 211, corresponding to loss of a benzyl radical
from the parent cation, (iii) cation 17 at m/z 210, generated Via
loss of a hydrogen atom from the radical 16, (iv) the radical
cation 12 at m/z 197 corresponding to loss of a benzoyl radical
from the parent ion, and (v) the cation 13 at m/z 196, generated
Via loss of a hydrogen atom from the radical cation 12. As in
the case of compound 6 (Scheme 2) we believe that 13 is the
right structural representation. When the ions at m/z 211 and
m/z 197 were respectively ejected from the cell Via standard
double resonance experiments, the corresponding ions at m/z
210 and m/z 196 disappeared, confirming the proposed frag-
mentation sequence. Since two different radical fragmentations
were observed, it could be expected that these species (12 and
16) would have similar appearance potentials, which was later
confirmed experimentally and supported by theoretical calcula-
tions.
Synthesis of Compounds. Three N-methylpyridinium salts
were prepared (Scheme 1). 4-(Dibenzylmethyl)pyridine (4) and
4-(1,3-diphenyl-1-oxo-2-propyl)pyridine (5) were made starting
from 4-picoline (1), by the sequential introduction either of two
benzyl or of phenylacetyl and benzyl groups using modified
literature methods.25 These two substituted pyridines 4 and 5
reacted with methyl iodide to form the corresponding N-
methylpyridinium salts 6 and 7. Compound 10 was prepared
from 1-pyridyl-2-phenylethanol (9) following the procedure
indicated in Scheme 1.
Fragmentation Pathways. N-Methyl-4-(1,3-dibenzylmeth-
yl)pyridinium iodide (6) (MW 415) was chosen as a model
compound with a neutral substituent group. When CAD
(21) (a) Katritzky, A. R.; Watson, C. H.; Dega-Szafran, Z.; Eyler, J. R.
J. Am. Chem. Soc. 1990, 112, 2471. (b) Katritzky, A. R.; Watson, C. H.;
Dega-Szafran, Z.; Eyler, J. R. J. Am. Chem. Soc. 1990, 112, 2479. (c)
Katritzky, A. R.; Malhotra, N.; Dega-Szafran, Z.; Savage, G. P.; Eyler, J.
R.; Watson, C. H. Org. Mass Spectrom. 1992, 27, 1317. (d) Katritzky, A.
R.; Dega-Szafran, Z.; Watson, C. H.; Eyler, J. R. J. Chem. Soc., Perkin
Trans. II 1990, 1051. (e) Katritzky, A. R.; Dega-Szafran, Z.; Ramanathan,
R.; Eyler, J. R. Org. Mass Spectrom. 1994, 29, 96.
(22) Katritzky, A. R.; Shipkova, P. A.; Burton, R. D.; Allin, S. M.;
Watson, C. H.; Eyler, J. R. J. Mass Spectrom. 1995, 30, 1581.
(23) Watson, C. H.; Baykut, G.; Mowafy, Z.; Katritzky, A. R.; Eyler, J.
R. Anal. Instrum. 1988, 17, 155.
(24) Katritzky, A. R.; Watson, C. H.; Dega-Szafran, Z.; Eyler, J. R. Org.
Mass Spectrom. 1989, 24, 1017.
(25) Kaiser, E. M.; Petty, J. D. Synthesis 1975, 705.

Radical Merostabilization by Electrospray FTICR/MS
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11907
Scheme 3. Fragmentation Patterns of Compound 7
Figure 1. CAD rate constants plotted Vs. ion center-of-mass kinetic
energies (kcal/mol) to yield the threshold fragmentation energies Eo
for cations 11, 15, and 18. The symbols represent experimental
measurements and the lines represent calculated fits from eqs 3 and 4.
Scheme 4. Fragmentation Patterns of Compound 10
the excitation plates of the analyzer cell.
Eion ) q2V2τ2/8md2
(1)
(2)
Ecm ) Eion × [MAr/(MAr + m)]
The nominal center-of-mass energy (Ecm) can be calculated
from eq 2, as described previously,21a where MAr is the mass of
the neutral gas (Ar) and m is the mass of the ion. In our
previous studies21a-e the nominal center-of-mass energy was
plotted Vs the percent fragmentation observed (I/Io) and the
straight line portions of these plots were extrapolated to zero
fragmentation to give the observed threshold energy (Eobs). In
the present study we used a program for analysis of CAD and
ion/molecule reaction cross sections (“CRUNCH”), written by
Armentrout and Ervin.27 We have modified slightly the fitting
function in that program to be more appropriate to the FTICR
CAD conditions as shown in eq 3, where k is the observed CAD
rate constant, ko is an energy-independent scaling factor, Eo is
the threshold energy and V is treated as a variable parameter.
This functional form has been predicted for CAD processes and
Armentrout et al. have shown experimental proof that it provides
accurate CAD thresholds.28
N-Methyl-4-(1-ethoxy-2-phenylethyl)pyridinium iodide (10)
(MW 369) was chosen as a model compound to study the effect
of an electron donating group (ethoxy) on the appearance
potential of the corresponding radical. For the CAD experi-
ments the precursor cation 18 (m/z 242) was isolated using
standard RF ejection pulses and was allowed to undergo
collisions with argon atoms. In addition to the parent ion 18,
the CAD spectrum showed formation of three fragment ions
(Scheme 4): (i) m/z 151 generated Via loss of a benzyl radical
from the parent ion forming the radical cation of interest 19;
(ii) m/z 122 20 corresponding to loss of ethyl radical from the
radical cation 19, and (iii) m/z 94 corresponding to methyl
pyridinium ion 21, generated Via loss of a carbon monoxide
molecule from the ion at m/z 122. To verify the fragmentation
pathway we ejected the radical at m/z 151 and as a result, the
ions 20 and 21 at m/z 122 and m/z 94 disappeared. It was thus
confirmed that ions 20 and 21 are formed from fragmentation
of the radical cation 19 and not directly from the parent ion.
Threshold Fragmentation Energies. Thresholds for the
appearance of fragment ions allowed the estimation of threshold
fragmentation energies (Eo) for the collisionally activated
dissociation (CAD) in the gas phase of three substituted
pyridinium cations to form the corresponding free radicals. The
Eo values for the ions 11, 15, and 18 were calculated as
previously described.21a The translational energy (Eion) imparted
to an ion during the excitation stage of the FTICR CAD process
is obtained from eq 1,26ab where q is the ionic charge, V is the
peak-to-peak amplitude of the RF excitation pulse, τ is the RF
pulse width, m is the ionic mass, and d is the distance between
k ) ko(E - Eo)V/E
ln(Io/I) ) knt
(3)
(4)
The CAD rate constants are derived from the percent
fragmentation observed (eq 4), where (I/Io) is the percent
fragmentation observed, n is the number density of the collision
gas and t is the CAD reaction time. Plots of CAD rate constants
Vs center-of-mass kinetic energy for all three cations 11, 15,
and 18 are shown in Figure 1. In a previous study21a absolute
experimental errors in calculating Eo values were estimated to
be up to 0.35 eV (8 kcal/mol), and relative experimental errors
are expected to be lower for comparable systems. The
experimental errors in the present report are expected to fall in
the same range.
Radical Merostabilization. As seen in Figure 1, and
determined from the computer fits (CRUNCH program) the
(26) (a) Cody, R. B.; Burnier, R. C.; Freiser, B. S. Anal. Chem. 1982,
54, 96. (b) Bensimon, M.; Houriet, R. Int. J. Mass Spectrom. Ion Processes
1986, 72, 93.
(27) Ervin, K. M., Armentrout, P. B. J. Chem Phys. 1985, 83, 166.
(28) Aristov, N.; Armentrout, P. B. J. Phys. Chem. 1986, 90, 5135.

11908 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Katritzky et al.
Figure 2. Theoretically calculated relative energy (kcal/mol) plotted Vs. reaction coordinate. The reaction coordinate listed is the distance of the
C-C bond being broken, minimizing the energy with respect to all other molecular coordinates.
observed threshold fragmentation energies (Eo) for pyridinium
cations 11, 15, and 18 are 29 kcal/mol, 35 kcal/mol and 11
kcal/mol, respectively. Assuming no significant difference in
ground state stabilization of the starting materials, a lower Eo
means higher stability of the corresponding radical. Therefore,
radical 19 demonstrated significant stabilization (∆Eo ≈ 20 kcal/
mol) compared to radical 12, (which has only one electron-
accepting group: pyridinium ion Py+) and radical 16, (which
has two electron-accepting groups: pyridinium ion and benzoyl).
This strongly supports the concept of radical merostabilization,
that an electron-accepting group (pyridinium ion) and an
electron-donating group (ethoxy) can significantly stabilize a
radical. The pyridinium ion is a very effective electron-
accepting group and only a very small difference of stabilization
was found between radicals 12 and 16, and as expected
pyridinium cation 15 gave two radicals, 12 and 16, in the CAD
experiments.
known about barriers to reactions, and the UHF method, in
general, might be expected to overestimate them. Comparison
of the observed fragmentation energies with our calculated
barriers suggest that this might be the case in the present
study: the energy required to break a C-C single bond could
be somewhat too high. The relative barriers, however, depend
on the intersection of this potential energy surface with that of
the relative heats of formation of the products. Throughout all
the calculations, all stationary points were additionally charac-
terized by calculations of the Hessian (force-constant) matrix.
All minima were confirmed with no negative eigenvalues and
all transition state geometries had a single negative eigenvalue
of the Hessian.
Due to the nature of the semiempirical parametrization, the
zero-point vibrational energies are included in the parameters
and need not be added explicitly. This assumption was verified
for several calculated geometries and energy differences. It was,
indeed, found that the inclusion of the zero-point energies did
not affect the results significantly.
Theoretical Calculations. To support the experimental
results, AM129 and PM330 quantum-chemical calculations were
performed, using the Unrestricted Hartree-Fock (UHF) Hamil-
tonian as implemented in the AMPAC 5.0 program package.31
Neither of these semiempirical methods has been parametrized
specifically for calculation of the properties of transition states
involving radical species or UHF wavefunctions. It was
therefore considered necessary to perform these calculations with
both parameterizations (AM1 and PM3) in parallel to decrease
the chances of model errors biasing the results. As shown later,
the results obtained are similar for the two parametrizations used.
In general the AM1 model produces heats of formation within
(7 kcal/mol and the PM3 model (4 kcal/mol for compounds
of this type.32 Neutral closed-shell systems are more accurately
reproduced than are radicals and cations.32 Since we are
comparing energy differences for rather similar processes we
expect that such errors are systematic. Considerably less is
Dissociation paths were initially characterized by stepwise
increasing the length of the breaking bond, while optimizing
all other coordinates at each given bond length. The resulting
potential energy curves are presented in Figure 2. From the
points with the highest energy, the transition state geometries
were refined using the eigenvector following procedure.33 It
can be concluded that the height of the reaction barrier (∆Hq),
as well as the overall thermodynamic stabilization are about 10
kcal/mol lower for cation 18 (see Table 1), containing an
electron-donor and an electron-acceptor group as compared to
cations with one neutral and one electron-accepting group (11)
or two electron-accepting groups (15) attached in the dissociation
product to the carbon radical center.
In another series of calculations, full geometry optimizations
were performed for the reactants and products of the dissociation
reactions, as well as for a series of similar comparison reactions,
where the electron-accepting methylpyridinium group was
replaced by a neutral phenyl group; in these compounds (22,
23, and 24) (see Scheme 5) with the effect of an electron-
(29) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J.
Am. Chem. Soc. 1985, 107, 3902.
(30) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209.
(31) AMPAC 5.0; Semichem, Inc.: 7128 Summit, Shawnee, KS 66216,
1994.
(32) Zerner, M. C. In ReViews of Computational Chemistry; Lipkowitz,
K. B., Ed.; VCH: New York, 1991; Vol. 2.
(33) Baker, J. J. Comput. Chem. 1986, 7, 385.

Radical Merostabilization by Electrospray FTICR/MS
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11909
Table 1. Theoretically Calculated Heats of Reaction (∆HR) and Energies of Transition States (∆Hq) (all in kcal/mol)
compounds and groupsa
11 A and N
15 A and A
18 A and D
22 N and N
∆HR
23 N and A
∆HR
24 N and D
∆HR
∆Hq
∆HR
∆Hq
∆HR
∆Hq
∆HR
AM1
PM3
38
37
38
36
39
41
40
41
31
32
32
29
33
33
35
36
33
32
a D ) electron donor; A ) electron acceptor; N ) neutral.
Scheme 5. Structures for Compounds 22, 23, and 24
the radical 19 (EM rad) as the characteristics of the merostabi-
lization (eq 6).
EM ) EDA - 1/2(EDD + EAA)
∆EM ) (EM rad) - (EM ion
(5)
(6)
)
Scheme 6. Structures 25-32 for Calculation of
Merostabilization Energies
When both an electron-donor (OEt) and an electron-acceptor
(methylpyridinium) group are present, as for radical 19, a
significant additional stabilization of the radical was observed
(∆EM ∼ 8-10 kcal/mol) (see Table 2). For a comparison, the
calculated analogous energy difference (∆EM) was much smaller
for radicals 12 and 16, containing the electron-accepting
methylpyridinium group and the mesomerically inactive benzyl
group, and two electron-acceptor groups (benzoyl and meth-
ylpyridinium), respectively (see Table 2). [For ion 11, EDD is
calculated from 26 and EAA from 25. For ion 15, EDD is
calculated from 27 and EAA from 25. For ion 18, EDD is
calculated from 28 and EAA from 25. For radical 12, EDD is
calculated from 30 and EAA from 29. For radical 16, EDD is
calculated from 31 and EAA from 29. For radical 19, EDD is
calculated from 32 and EAA from 29.] Therefore, our calcula-
tions indicate the presence of a certain merostabilization of the
radical with an electron-accepting and an electron-donating
group simultaneously bonded to a carbon radical center.
acceptor substituent eliminated, no radical stabilization should
be found. Heats of radical formation (∆HR) were calculated
for all cases as summarized in Table 1. The dissociation energy
of cation 18 is significantly lower than that of the cations 11 or
15, in agreement with the experimental results. However, where
the electron-accepting group (pyridinium) is replaced by a
neutral group (phenyl) no evidence for stabilization is found
not only in the heats of the dissociation reaction of compounds
22 and 23 but also in that for 24. For compounds 15 and 18,
the AM1 calculated transition state energies (∆Hq) are slightly
lower than the corresponding heats of radical formation (∆HR),
due to the different methods of calculation of these quantities.
The ∆HR calculations involved heats of formation of isolated
products (corresponding to infinite separation), whereas in the
∆Hq calculations, the products were separated by a small finite
distance and the energy of the system is lowered by the charge-
dipole interactions between the products.
Conclusion
The existence of merostabilization in the system investigated
is supported by both our experimental and theoretical calcula-
tions. It was clearly shown that when both an electron-donor
and an electron-acceptor group are present at the radical site,
significant additional stabilization was achieved. For compari-
son, when in the same system the electron-donor group was
replaced by an electron-acceptor or a neutral group, no additional
stabilization was observed.
Experimental Section
ESI Source. A modified dual-stage hexapole ion-guide ESI source
(Analytica of Branford, Branford, CN) was used for this work. The
normal glass capillary (0.5 mm I.D. × 150 mm) used with nitrogen
counter-current drying gas was replaced by a heated metal capillary of
similar dimensions. The metal capillary was mounted in a brass support
and heated with a 200 W cartridge heater. The temperature was
maintained at 120 °C using an RTD (Resistive Thermal Device)
temperature controller which regulated the heater power supply. The
source was operated without a counter-current drying gas.
The merostabilization energy (EM) of a radical with an
electron-donor (D) and an electron-acceptor (A) groups attached
simultaneously to the carbon radical center (e.g. radical 19) is
defined as shown in eq 5, where EDA is the energy of the radical
studied (19), EDD is the energy of a radical with two electron-
donor groups (such as radical 32, Scheme 6) and EAA is the
energy of the radical with two electron-acceptor groups at the
radical center (such as radical 29, Scheme 6).
Ions were sprayed from a needle maintained at a potential of +3.9
kV. The pressure between the capillary exit and the first skimmer (1.0
mm diameter) was maintained at 1 mbar with a 500 L/min mechanical
pump (Edwards, Model EM32). The pressure between the first
skimmer and conductance limit (1 mm diameter) located in the center
of the hexapole ion-guide was maintained at ca. 1 × 10-3 mbar with
a turbo drag pump (Edwards, Model EXT250HI) as measured using a
Pirani gauge (TPR 010, Balzers AG, Liechtenstein). The pressure after
the ESI hexapole guide in the normal FTICR external source housing
was ca. 1.0 × 10-6 mbar as pumped by an 800 L/s cryopump.
The calculated merostabilization energy (EM) of radical 19
is artificially lowered because the heat of formation of the AA
species (EAA) is additionally enhanced due to the intramolecular
charge-charge repulsion between two positively charged methyl
pyridinium groups. In order to eliminate this contribution to
the merostabilization energy, we have used the difference
between the calculated EM values of the parent ion 18 (EM ion
,
calculated using structures 18, 28, and 25, respectively) and of

11910 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Katritzky et al.
Table 2. Calculation of Merostabilization Energiesa
heat of formation
EM ion
heat of formation
EM rad
∆EM
I: Neutral and Acceptor Substituents
cpds
AM1
PM3
11
217
217
25
441
436
26
53
54
11
-30
-28
12
216
213
29
441
432
30
56
59
12
-32
-32
-2
-4
II: Acceptor and Acceptor Substituents
cpds
AM1
PM3
15
197
191
25
441
436
27
9
4
15
-28
-29
16
198
192
29
441
432
31
15
9
16
-30
-29
-2
0
III: Donor and Acceptor Substituents
cpds
AM1
PM3
18
152
154
25
441
436
28
-90
-76
18
-24
-25
19
145
144
29
441
432
32
-83
-78
19
-34
-33
-10
-8
aAll values are given in kcal/mol.
The samples were dissolved in 49/49/2 (volume ratio) water/
methanol/acetic acid solution at a concentration of 0.1 mg/mL and were
introduced into the ESI source at a flow rate of 1 µL/min.
External Source FTICR. All experiments were carried out using
a Bruker BioAPEX 7e external source FTICR spectrometer (Bruker
Analytical Systems, Inc., Billerica, MA) equipped with a horizontal
bore, room-temperature, 15 cm inside diameter, 7 T superconducting
magnet, an external source, and a 170 mm3 cylindrical RF-shimmed
Infinity analyzer cell.34 The standard external source FTICR instrument
has been previously described.35 Problems associated with installation
of the nominal 7 tesla magnet resulted in using the magnet, during this
study, energized only to 4.7 tesla. The FTICR external source design
has two stages of differential pumping which produce a pressure
differential of up to 105 between the external source and analyzer
regions. The high and ultrahigh vacuum regions were maintained using
cryopumps (Edwards High Vacuum Co., Woburn, MA). An 800 L/s
cryopump (N2) was used to pump the high gas load region of the
external source vacuum compartment where ions exiting the ESI source
were accelerated toward the mass analyzer using a series of electrostatic
optics. A 400 L/s cryopump evacuated the region between the first
and second conductance limits that contained the electrostatic ion
transfer optics. An additional 400 L/s cryopump was used to maintain
the ultrahigh vacuum in the analyzer region.
After exiting the high pressure ESI source and entering the
electrostatic ion optics, ions were accelerated to 2.5 kV and tightly
focused through the two conductance limits before being accumulated
and trapped in the FTICR analyzer cell.35 Since the ions were
continuously produced by the ESI source a set of x-, y-deflection plates
were gated “open” during the ion accumulation period for 50 to 200
ms to allow time for ions to accumulate. The trapped ions were
resonantly excited using a swept frequency RF “chirp”. Ions were
detected using the broadband detection mode covering a mass range
from 50 to 5000 amu. Typically, 25 individual transients were co-
added to improve the signal-to-noise ratio. The raw time-domain
transients were apodized36 and zero-filled37,38 three times prior to Fourier
transformation.
Figure 3. Pulse sequence for CAD experiments.
dependence of the collisionally activated dissociation process. The
amplitude of the RF excitation was measured directly on the cell plates
using a calibrated oscilloscope and the obtained values were used for
the calculations (eq 1). A delay of 0.5 s following the precursor ion’s
excitation allowed time for the translationally excited precursor ion to
undergo collisions and fragment prior to the excitation/detection events.
A schematic diagram of the pulse sequence is shown in Figure 3.
General Procedure for Synthesis of Compounds 2-5. To a stirred
THF solution of starting material, LDA (1.5 M solution in cyclohexane,
1.1 equiv) was added dropwise at -78 °C under nitrogen atmosphere,
the resulting mixture was allowed to stir for 1 h. Then, a solution of
appropriate electrophile (1 equiv) in THF was added slowly. The
reaction mixture was stirred for a further 2 h at -78 °C and then allowed
to warm to room temperature. After general workup, the crude product
was purified by column chromatography to give the pure compound.
For 4-dibenzylmethylpyridine (4), a mixture of 4-phenylethylpyridines
2 and 4 (70:30) was obtained and could not be separated by column
chromatography. The pure compound 4 was obtained by preparative
GC.
4-Phenylethylpyridine (2): white microcrystals, 90% yield, mp 69-
71 °C [lit.39 70-70.8 °C] (Found C, 85.04; H, 7.35; N, 7.59. C13H13N
1
requires C, 85.21; H, 7.15; N, 7.64%); H NMR δ 2.89 (4H, s), 7.05
(2H, d, J ) 6.7 Hz), 7.12-7.32 (5H, m) and 8.46 (2H, d, J ) 6.7 Hz);
13C NMR δ 36.5, 37.0, 121.0, 123.9, 126.2, 128.3, 128.4, 140.6, 149.7
and 150.4; MS m/z 183 (M+, 44%), 91 (100).
The experimental event sequence for obtaining ESI/FTICR mass
spectra began with a 50 ms quench pulse to remove any ions remaining
in the cell from a previous measurement cycle. After the quench pulse,
an ion accumulation event followed, which allowed accumulation of
ions in the analyzer cell. Precursor ions were isolated using swept-
frequency ejection pulses of ca. 100 to 400 ms duration to eject all
other ions. A pulsed valve introduced the argon collision gas prior to
ion activation. With the pulsed valve open for 50 µs a peak pressure
of 3 × 10-7 mbar was obtained. The precursor ions were excited using
a variable amplitude on-resonance excitation pulse of 10 µs duration.
A 50 to 250 volt range of RF amplitudes was used to study the energy
4-(Phenylacylmethyl)pyridine (3): white microcrystals, 25% yield,
mp 112-114 °C [lit.40 113-115 °C] (Found C, 79.40; H, 5.72; N,
1
7.19. C13H11NO requires C, 79.19; H, 5.62; N, 7.10%); H NMR δ
4.31 (2H, s), 7.22 (2H, d, J ) 5.8 Hz), 7.48-7.53 (2H, m), 7.59-7.64
(1H, m), 8.0-8.03 (2H, m) and 8.58 (2H, d, J ) 5.6 Hz); 13C NMR δ
44.5, 124.8, 128.7, 128.8, 133.5, 136.1, 143.4, 149.9 and 195.8; MS
m/z 197 (M+, 10%), 105 (100).
4-Dibenzylmethylpyridine (4): white microcrystals, mp 70-72 °C
[lit.39 71.5-73 °C] (Found C, 87.56; H, 7.20; N, 5.10. C20H19N requires
C, 87.87; H, 7.01; N, 5.12%); 1H NMR δ 2.89 (2H, dd, J ) 8.5, 13.5
Hz), 3.02 (2H, dd, J ) 6.2, 13.5 Hz), 3.10-3.20 (1H, m), 6.93-7.01
(6H, m), 7.14-7.26 (6H, m) and 8.40 (2H, d, J ) 6.1 Hz); 13C NMR
δ 41.8, 49.5, 123.4, 126.2, 128.3, 129.0, 139.4, 149.6 and 153.1; MS
m/z 273 (M+, 67%), 182 (75), 91 (100).
(34) Caravatti P.; Allemann, M. Org. Mass Specrom. 1991, 26, 514.
(35) Kruppa, G. H.; Caravatti, P.; Radloff, C.; Zu¨rcher, S.; Laukien, F.;
Watson, C; Wronka, J. In FT-ICR/MS: Analytical Applications of Fourier
Transform Ion Cyclotron Resonance Mass Spectrometry; Asamoto, B., Ed.;
VCH: New York, 1991; Chapter 4, pp 107.
(36) Lee, J. P.; Comisarow, M. B. Appl. Spectrom. 1989, 43, 599.
(37) Bartholdi, E.; Ernst, R. R. J. Magn. Resonance 1973, 11, 9.
(38) Comisarow, M. B.; Melka, J. D. Anal. Chem. 1979, 51, 2198.
4-(1,3-Diphenyl-1-oxo-2-propyl)pyridine (5): white microcrystals,
35% yield, mp 108-110 °C (Found C, 84.03; H, 6.09; N, 4.78. C20H17-
(39) Avramoff, M.; Sprinzak, Y. J. Am. Chem. Soc. 1956, 78, 4090.

Radical Merostabilization by Electrospray FTICR/MS
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11911
1
NO requires C, 83.60; H, 5.96; N, 4.87%); H NMR δ 3.07 (1H, dd,
solution of 4-pyridinylcarboxaldehyde in dry THF. The reaction
mixture was allowed to warm to room temperature overnight. After
usual aqueous workup, the product was obtained by flash column
chromatography as a solid, yield 40%, mp 145-147 °C [lit.41 146-
147.5 °C]; 1H NMR δ 2.90-3.05 (2H, m), 3.35 (1H, br s), 4.85-4.90
(1H, m), 7.10-7.30 (7H, m) and 8.42 (2H, d, J ) 6.0 Hz); 13C NMR
δ 45.7, 73.7, 120.9, 126.9, 128.6, 129.5, 137.0, 149.6 and 152.9; MS
m/z 199 (M+, 4%), 108 (72) and 92 (100).
N-Methyl-4(1-benzyl-1-ethoxy)methylpyridinium Iodide (10).
Compound 9 was first reacted with excess iodomethane in methylene
chloride at rt to form N-methyl pyridinium salt. The excess MeI and
solvent was removed under vacuum. The residue was dissolved in
dry THF under nitrogen atmosphere, sodium hydride was added at 0
°C, then warm to rt. Then, ethyl iodide (2 equiv) was added and heated
to reflux for 10 h. The reaction mixture was concentrated, dissolved
in CH2Cl2, washed by H2O, and dried. The solvent was removed to
give a mixture, which was difficult to purify, but shown to contain 10
as the major component by NMR and FTICR/MS. 1H NMR δ 1.16
(3H, t, J ) 7.0 Hz), 3.40 (1H, dd, J ) 5.8, 13.8 Hz), 3.43 (1H, dd, J
) 6.8, 13.8 Hz), 3.61 (2H, q, J ) 7.0 Hz), 4.64 (3H, s), 4.73-4.77
(1H, m), 7.10 (2H, d, J ) 7.8 Hz), 7.21-7.29 (3H, m), 7.84 (2H, d, J
) 6.4 Hz) and 9.15 (2H, d, J ) 6.5 Hz); 13C NMR δ 15.1, 43.2, 49.0,
66.1, 80.7, 125.5, 127.0, 128.4, 129.2, 129.6, 135.4, 144.9 and 162.7.
J ) 7.7, 13.8 Hz), 3.55 (1H, dd, J ) 7.1, 13.7 Hz), 4.81 (1H, t, J )
7.2 Hz), 7.05-7.07 (2H, m), 7.16-7.27 (5H, m), 7.36-7.41 (2H, m),
7.48-7.50 (1H, m), 7.86-7.89 (2H, m) and 8.50 (2H, br s); 13C NMR
δ 39.7, 55.1, 123.4, 123.5, 126.5, 128.4, 128.6, 128.7, 129.0, 133.3,
138.7, 147.8, 150.2 and 198.1; MS m/z 287 (M+, 18%), 105 (100).
General Procedure for Synthesis of N-Methylpyridinium Salts
6 and 7. The appropriate substituted pyridine 4 and 5 was reacted
with excess methyl iodide in methylene chloride at room temperature.
The reaction was monitored by 1H NMR. After the reaction was
complete, the reaction mixture was concentrated, and the residue was
purified by recrystalization to give the pure pyridinium salt.
N-Methyl-4-(1,3-diphenyl-2-propyl)pyridinium iodide (6): white
microcrystals, 90% yield, 1H NMR δ 2.96 (2H, dd, J ) 8.9, 13.9 Hz),
3.15 (2H, dd, J ) 6.2, 14.0 Hz), 3.49-3.53 (1H, m), 4.58 (3H, s),
7.03 (4H, d, J ) 6.9 Hz), 7.13-7.29 (6H, m), 7.68 (2H, d, J ) 6.6
Hz), and 9.10 (2H, d, J ) 6.5 Hz); 13C NMR δ 41.0, 48.6, 50.1, 126.9,
127.5, 128.8, 128.9, 137.4, 144.6 and 164.8.
N-Methyl-4-(1,3-diphenyl-1-oxo-2-propyl)pyridinium Iodide (7).
In this case, the residue was very difficult to purify by recrystallization,
but it was verified that the major part of the residue was 7 by NMR
and FTICR/MS. 1H NMR δ 3.19 (1H, dd, J ) 7.5, 13.6 Hz), 3.56
(1H, dd, J ) 7.3, 13.4 Hz), 4.55 (3H, s), 5.36 (1H, t, J ) 7.4 Hz),
7.10-7.30 (7H, m), 7.40-7.65 (3H, m), 8.05-8.10 (2H, m) and 9.02
(2H, d, J ) 6.7 Hz); 13C NMR δ 40.2, 49.1, 54.3, 127.4, 128.4, 128.9,
129.0, 129.2, 130.1, 130.6, 133.5, 134.4, 136.3, 142.3, 144.9, 158.6
and 196.5.
JA962027H
(40) Smith, Jr. J. M.; Stewart, H. W.; Roth, B.; Northey, E. H. J. Am.
Chem. Soc. 1948, 70, 3997.
(41) Traynelis, V. J.; Yamauchi, K.; Kimball, J. P. J. Am. Chem. Soc.
1974, 96, 7289.
1-Pyridyl-2-phenylethanol (9). Benzylmagnesium bromide (1.2
equiv, 1.0 M ether solution) was added slowly to a cooled (-78 °C)