Protonated benzene: A case for structural memory effects?

Article abstract of DOI:10.1021/jp0479973

Photoionization tandem-mass spectrometry of selectively deuterium labeled precursor molecules is used to probe the structure of protonated benzene which had been questioned in a provocative publication by Mason and co-workers (J. Chem. Soc., Chem. Commun. 1995, 1027). Specifically, we address the intriguing aspect of a postulated delayed hydrogen ring-walk by modulating the internal energy content of mass-selected C₆H_7-nD_n ⁺ ions. To this end, ionization of selectively deuterated precursors by tunable synchrotron photons is combined with chemical monitoring of H ⁺/D⁺ transfer from C₆H_7-nD _n⁺ to a strong base B. The resulting BH ⁺/BD⁺ ratios monotonically decrease with increasing internal energy content and, due to the virtue of the regioselective deuterium label incorporations, thereby disprove a delayed hydrogen ring-walk or any other "structural memory effect" in C₆H_7-nD _n⁺ ions. As a consequence, the experimental studies of Mason et al. were reconsidered using sector-field mass spectrometry. These extensive studies suggest that the previous observations might have been obscured by a combination of isobaric impurities, metastable-ion contributions, and artifact signals.

Full text of DOI:10.1021/jp0479973

J. Phys. Chem. A 2004, 108, 9931-9937
9931
Protonated Benzene: A Case for Structural Memory Effects?†
Detlef Schro1der,*,‡ Jessica Loos,‡ Helmut Schwarz,‡ Roland Thissen,§ and Odile Dutuit§
Institut fu¨r Chemie der Technischen UniVersita¨t Berlin, Strasse des 17 Juni 135, D-10623 Berlin, Germany,
and Laboratoire de Chimie Physique, Baˆt. 350, UniVersite´ Paris-Sud, 91405 Orsay, France
ReceiVed: May 10, 2004; In Final Form: June 15, 2004
Photoionization tandem-mass spectrometry of selectively deuterium labeled precursor molecules is used to
probe the structure of protonated benzene which had been questioned in a provocative publication by Mason
and co-workers (J. Chem. Soc., Chem. Commun. 1995, 1027). Specifically, we address the intriguing aspect
of a postulated delayed hydrogen ring-walk by modulating the internal energy content of mass-selected
C6H7-nDn+ ions. To this end, ionization of selectively deuterated precursors by tunable synchrotron photons
+
is combined with chemical monitoring of H+/D+ transfer from C6H7-nDn to a strong base B. The resulting
BH+/BD+ ratios monotonically decrease with increasing internal energy content and, due to the virtue of the
regioselective deuterium label incorporations, thereby disprove a delayed hydrogen ring-walk or any other
+
“structural memory effect” in C6H7-nDn ions. As a consequence, the experimental studies of Mason et al.
were reconsidered using sector-field mass spectrometry. These extensive studies suggest that the previous
observations might have been obscured by a combination of isobaric impurities, metastable-ion contributions,
and artifact signals.
Introduction
CHART 1
One cannot possibly escape the warm, charming character
of Professor Tom Baer, an insightful scientist and a dedicated
teacher, an exemplary scholar who cares for his students. On
several occasions he indicated to us that experimental findings
from our laboratories had been used as teaching examples in
his student courses. In this very spirit, here we would like to
present a story about protonated benzene, C6H7+, a prototype
organic cation which occupies a central position in physical
organic chemistry.1
identified as a second-order saddle point more than 2 eV higher
in energy than 2 (which represents the global minimum). This
view was further supported by several experimental and
theoretical studies,5-9 of which the spectroscopic characteriza-
tions of gaseous C6H7+ are most impressive.10-12 Notwithstand-
ing, the puzzling experimental findings of Mason and co-
workers persist and await an explanation.13
In 1995, Mason et al.2 published a report about an unexpected
temperature dependence of the collisional activation (CA) mass
spectra of protonated benzene. In brief, they found that the ratio
between H• and H2 losses upon CA of mass-selected C6H7
+
Here, we describe attempts to resolve matters. To this end, a
specific photoionization experiment was designed in order to
probe a possible memory effect in protonated benzene. In
addition, the collision studies of Mason et al. were reinvestigated
using sector-field mass spectrometry.
depends on the temperature of the chemical ionization (CI)
source used: below room temperature, loss of H• prevails,
whereas elimination of H2 predominates at elevated tempera-
+
tures. Moreover, for the deuterated ion C6HD6 generated by
CI of C6D6 with methane, a preferential loss of H• was observed
at low temperatures, which was interpreted as some kind of a
“memory effect” in that complete H/D equilibration does not
occur. From these results, the authors concluded that the face-
centered π-complex 1, rather than the time-honored σ-complex
2 (the cyclohexadienyl cation), corresponds to the most stable
structure of protonated benzene (Chart 1).
Experimental Methods
The photoionization studies were performed in the CERISES
apparatus14 installed on the beam line SA63 of the synchrotron-
radiation source SuperACO at LURE (Orsay, France). This
beam line has a normal incidence monochromator which
provides monochromatic light in the range of 7-30 eV photon
energy. Slits were at all times opened to 1 mm, corresponding
to a photon-energy resolution in the range of 500 (i.e., 20 meV
at 10 eV). The accuracy of the photon energy was verified by
measuring the ionization energy of argon within 2 meV of its
nominal value. At photon energies e 11.7 eV, a LiF window
was used to effectively eliminate higher order light emerging
from the grating of the monochromator. The experiments
without LiF filter were inherently interfered with by higher order
photons and are reported without correction. Briefly, the neutral
precursors introduced via a septum inlet were ionized by
Being skeptical about this surprising result, we soon repeated
the experiments. Despite our much more limited capabilities to
vary the temperature of the ion source,3 we could, however,
qualitatively reproduce the effects described by Mason and co-
workers. Nevertheless, doubts about Mason’s assignment and
conclusion remained. Radom and co-workers4 applied high-level
ab initio theory to disclaim the existence of 1, which was
† Part of the special issue “Tomas Baer Festschrift”.
* Corresponding author. E-mail: detlef.schroeder@tu-berlin.de.
‡ Institut fu¨r Chemie der Technischen Universita¨t Berlin.
§ Universite´ Paris-Sud.
10.1021/jp0479973 CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/08/2004

9932 J. Phys. Chem. A, Vol. 108, No. 45, 2004
Schro¨der et al.
monochromatic photons, and any cations formed were extracted
by a field of 1 V/cm toward a QOQ system (Q stands for
quadrupole and O for octopole).15 In the MS/MS experiments
described below, Q1 was used for mass-selection of the desired
SCHEME 1
+
C6H7-nDn ions. Trimethylamine as a base B was leaked into
a reaction cell which is a 4 cm long cylinder around the octopole,
and Q2 was scanned for mass analysis. The kinetic energy of
the ions within the octopole was adjusted to 0.15 eV. The ion
beams have Gaussian kinetic energy distributions with an
average full width at half-maximum (fwhm) of ca. 0.3 eV in
the laboratory frame. Ions were detected by a set of multichannel
plates operating in the counting mode. The photon-energy
dependences of the BH+/BD+ ratios were recorded in single-
ion monitoring mode and are reported with corrections for the
background and contributions of 13C isotopes. The appearance
energies reported below were determined by extrapolating the
linear part of the respective single-ion monitoring curve as a
function of photon energy. The given error estimates refer to
these extrapolations; neither thermal contributions nor kinetic
shifts were considered.
Additional experiments were performed with a modified VG
ZAB/HF/AMD 604 four-sector mass spectrometer of BEBE
configuration (B stands for magnetic and E for electric sector),
which has been described elsewhere.16 In brief, C6H7-nDn+ ions
were generated by either electron ionization (EI) or chemical
ionization of different precursors as detailed below. After
acceleration to a kinetic energy of 8 keV, the ions of interest
were mass-selected and subjected to metastable ion (MI) and
collisional activation studies. MI spectra of B(1)/E(1) mass-
selected ions were recorded by detection of the charged
fragments formed unimolecularly in the field-free region
between E(1) and B(2) by scanning the latter sector. CA spectra
were recorded in the same manner using either helium (80%
transmission, T) or oxygen (80% T) as stationary collision gases.
In addition, MI and CA spectra were recorded in order to
investigate the fragmentations occurring in other field-free
regions of the Berlin mass spectrometer using the following
methods:17,18 (i) direct analysis of the daughter ions (DADI)
formed in the field-free region between the ion source and the
first magnet by means of a linked B(1)/E(1) scan and (ii) mass
superposition of collision-induced fragmentations occurring in
the collision cell and unimolecular processes occurring in the
entire field-free region. Due to the particularly long third field-
free region of the Berlin BEBE instrument (1.9 m), the
contribution of metastable ions to the CA spectra is quite
pronounced. The CA spectra can be corrected for MI contribu-
tions by monitoring the absolute signals at a fixed multiplier
voltage as well as the parent ion transmission in MI and CA
experiments conducted directly after each other, thereby allow-
ing one to subtract the MI components from the CA spectra.
The carboxymethyl cyclohexa-2,5-dienes 3, 3a, and 3b were
prepared by Birch reduction of the corresponding benzoic acids
followed by esterification with methanol (Scheme 1).19 The
compounds were purified by bulb-to-bulb distillation (20 mbar)
1
and characterized by H NMR and GC-MS. In the course of
the synthesis, 5-10% of the corresponding conjugated car-
boxymethyl cyclohexa-1,3-dienes (e.g., 3′) are formed, probably
via proton catalysis in the course of esterification. In addition,
the samples undergo transfer hydrogenation as well as autoxi-
dation during storage and thus contain some of the correspond-
ing methyl benzoates as well as traces of carboxymethyl
cyclohexenes.20 Unfortunately, further purification of the samples
by means of preparative gas chromatography is hampered by
dehydrogenation to methyl benzoate occurring in the heated inlet
block. Notwithstanding, these impurities do not affect the
qualitative outcome of the photoionization experiments described
below.
+
selection of C6H7-nDn with B(1) only and recording the
fragmentations occurring in the field-free region between B(1)
and E(1) by scanning the latter (MIKE scan). The kinetic energy
of the ionizing electrons was varied by adjusting the voltage
between the filament and the source with the repeller grounded
to the source; the latter is important in order to avoid
unintentional acceleration of the electrons entering the source
toward the repeller.3 The ionization energy of benzene (IE )
9.24 eV) was used as an anchor for the electron-energy scale.
Note, however, that the electrons evolving from the hot filament
are not at all monoenergetic, and the dependence reported below
is largely meant qualitatively. Further, MI and CA spectra were
recorded at different temperatures of the ion source. Because
the ion source is usually heated to 200 °C and not equipped
with a cooling device, these experiments were performed as
follows. 12 h prior to the actual measurements, the ion optics
were optimized first and the source heater was switched off, to
allow the source to reach ambient temperature. Then, the
precursors were introduced, and spectra were immediately
recorded while monitoring the source temperature, which
increases continuously due to the heat emitted by the glowing
filament. In about a half-hour, the source reaches ca. 160 °C,
and higher temperatures were achieved by additional heating.
In general, CA spectra recorded in a sector instrument are a
Results and Discussion
One option for rationalizing the findings of Mason et al.2
without shaking the fundaments of physical organic chemistry
is that the hydrogen ring-walk21 in protonated benzene is not
that fast as usually anticipated when the internal energy content
of structure 2 is sufficiently low. Upon reacting with a base B,
such a “cold” ion 2 is expected to preferentially transfer one of
the hydrogens of the methylene group. Only if hydrogen ring-
walk is fast, all hydrogen atoms are equilibrated. According to
experiment and theory, the barrier for hydrogen ring-walk is
on the order of 0.27-0.45 eV.4,8,22-25 Although it would be
somewhat surprising that a barrier of this height is not
surmounted in the experiment, it could be argued that at the
low temperatures achievable in the apparatus of Mason et al.

Protonated Benzene
J. Phys. Chem. A, Vol. 108, No. 45, 2004 9933
SCHEME 2
dissociative photoionization. In addition, one should note that
the internal energy of the fragment ion does not correspond
directly to the difference of the photon energy and the threshold
for dissociative photoionization, but to a somewhat lower
amount because some energy may be lost in both the leaving
electron and the neutral fragment formed; we return to this
aspect in more detail further below. A control of the ions'
internal energy by means of coincidence methods is generally
conceivable but less well-defined for fragments formed upon
dissociative ionization and is moreover likely to be associated
with a rather poor sensitivity.
The choice of the base is critical in the present experiment.
If the proton affinity of the base PA(B) is smaller than
PA(benzene), proton transfer is unlikely to occur at all. If
PA(B) ≈ PA(benzene), that is, proton transfer is more or less
thermoneutral, formation of a long-lived ion/neutral complex
(INC)33 is likely to occur in which reversible processes might
hide the structural information introduced by the regioselective
isotopic labeling of the precursors.19 Similar considerations apply
for slightly exothermic proton transfer, PA(B) > PA(benzene).
Only a considerably exothermic proton transfer, PA(B) .
PA(benzene), whose exothermicity exceeds the typical binding
energies of INCs (up to 1 eV), could lead to immediate H+ or
D+ transfer without involving long-lived intermediates. More-
over, the base should not contain exchangeable protons itself
because secondary reactions may otherwise lead to H/D
exchange reactions and might thus deplete the BD+ signal.
Given PA(benzene) ) 7.78 eV, trimethylamine (PA ) 9.84
eV)34 was therefore considered an appropriate reagent serving
as base B.
(Tmin ) 213 K) in conjunction with sampling the ions within
+
about a microsecond, some of the C6H7-nDn ions have not
undergone complete H/D equilibration via the ring-walk mech-
anism. Upon collisional activation, the H(D) atoms of the
methylene group may then be lost preferentially. Following this
scenario, dissociative ionization of the selectively deuterated
precursors 3a and 3b concomitant with loss of neutral [C2H3O2] 26
could provide access to the specifically labeled cyclohexadienyl
cations 2a and 2b (Scheme 2). Note that the choice of the
precursors is in close analogy to previous studies of Kuck,
Gru¨tzmacher, and co-workers.21,27-29
Another important aspect needs to be addressed right at the
outset. Potentially, there exist numerous isomeric structures of
C6H7+, e.g., acyclic ones, fulvenes, and systems with bicyclic
skeletons.7,24 However, a consideration of the experimental
conditions in the work of Mason et al.2 implies that an internally
consistent explanation of the observed effects would indeed
require an isomer of protonated benzene which is more stable
than structure 2. Specifically, it needs to account for the
anomalies in the H• and D• losses, which were observed at
lowest temperatures, and the fact that the collision experiments
sample the entire population of ions, rather than just a minor
fraction of isomeric species. As Mason et al. used benzene as
the neutral precursor in their chemical ionization studies, any
high-energy isomers (e.g., protonated fulvene),7 even if they
might be formed in small amounts, cannot explain the changes
of the gross population of ions. With respect to the unique
thermochemical stability of benzene among the 217 possible
C6H6 isomers,30 we therefore restrict ourselves to consideration
of structures 1 and 2.31
C6H7+ + (CH3)3N f (CH3)3NH+ + C6H6
(1)
When interacting mass-selected C6H7+, generated by dis-
sociative photoionization of 3, with trimethylamine under the
conditions described above, proton transfer is observed exclu-
sively (reaction 1). Two additional ionic products, the radical
cation (CH3)3N•+ and the immonium ion (CH3)2NCH2+, are
observed at low photon energies and can be traced back to an
isobaric interference of mass-selected C6H7 with the 13C1-
+
isotope of C6H6•+ co-generated upon dissociative photoioniza-
tion of 3. Accordingly, these products are not pursued any
further (but were included in the isotope correction of the
(CH3)3NH+ signals).
Before presenting the experimental findings, we address the
concept of the present experiments in some more detail.
Photoionization Experiments. In analogy to a previous study
on the titration of the barrier for the keto/enol tautomerization
in ionized acetamide,32 the photoionization experiments were
designed as follows. The neutral samples 3, 3a, and 3b,
respectively, were introduced into the ion source and ionized
with photons from tunable synchrotron radiation; the ions of
interest, C6H7+ for 3, C6H6D+ for 3a, and C6H2D5+ for 3b, were
mass-selected using the quadrupole Q1 and interacted with a
Irrespective of the actual structure of the C6H7-nDn ions, H+
+
or D+ transfer to a base B might be associated with a kinetic
isotope effect (KIE), which itself depends on the internal energy
of the C6H7-nDn+ ions and can generally be expected to decrease
with increasing energy content.35 Furthermore, in the case of
complete H/D equilibration occurring before the ions enter the
reaction region, the different numbers of H and D atoms in the
reactant ions have to be acknowledged by statistical corrections.
If, however, the C6H7-nDn+ ions “remember” their origin from
the selectively labeled neutral precursors 3a and 3b, respectively,
that is, if hydrogen ring-walk is slow at low internal energies,
both ions 2a and 2b will bear a 1:1 ratio of H+ and D+ transfer
modified by a KIE (Scheme 3).36,37
base B in the octopole ion guide at near thermal energy (Elab
)
0.15 eV); and the ratio of the proton-transfer products BH+ and
BD+ was monitored using the second mass analyzer Q2. The
key variable is the energy of the ionizing photons, while all
other parameters are kept constant. The intention of this kind
of experiment is to vary the internal energy content of the
reactant ion by tuning the photon energy and to probe possible
consequences (if there are any) in a chemical reaction which
serves as a monitor.32 While this approach is quite straightfor-
ward for photoionization to a molecular cation, the present
experiment should also probe whether the idea of the “titration
method” can be transferred to a fragment ion formed in
Given the generally valid assumption that apparent KIEs
decrease with increasing internal energy35 and that rapid H/D
equilibration of C6H7-nDn+ takes place at elevated energies, the
present experiments allow for qualitative distinction of the
structural effects (i.e., delayed hydrogen ring-walk) and KIEs
(Figure 1). If the specifically labeled ions 2a and 2b are formed

9934 J. Phys. Chem. A, Vol. 108, No. 45, 2004
Schro¨der et al.
SCHEME 3
at low energies, the BH+/BD+ ratios of both ions should start
from a common value and then diverge to approach the
respective KIE-modified statistical limits at larger energies due
to H/D equilibration (case a); these limits are BH+/BD+
)
)
Figure 1. Qualitative sketches of the different BH+/BD+ ratios
expected in the photoionization experiments. See text for details.
(6/1)KIE for C6H6D+ generated from 3a and BH+/BD+
+
(2/5)KIE for C6H2D5 generated from 3b. If instead hydrogen
ring-walk is rapid, the variation of the BH+/BD+ ratios should
qualitatively show the same energy behavior with an offset due
to the different number of H and D atoms in the ions 2a and
2b (case b). Accordingly, it is not the precise values of the
BH+/BD+ ratios but merely the qualitative inspection of the
trends as a function of the energy of the ionizing photons that
should reveal whether any kind of structural memory exists in
protonated benzene. There is, of course, a third alternative in
that the dissociative photoionization of the neutral precursors
+
itself already deposits too much internal energy in the C6H7-nDn
ions formed such that no distinct energy dependence of the
BH+/BD+ ratios is obtained (case c); this may in fact be true
in many cases of dissociative ionization, but fortunately does
not apply here. Further, while KIE in a proton-transfer reaction
certainly depends on the base B, this effect would only influence
the actual values of the BH+/BD+ ratios in the experiments but
not change the qualitative difference of the trends. Finally, to
avoid a misleading interpretation of the graphs, note the
deviation of this kind of “gas-phase titration”32 from threshold
collision-induced dissociation and related experiments. In the
present study, the interaction energy of the ions and the neutral
reagent is kept constant throughout the experiments, while only
the “history” of the reactant ions is changed by variation of the
energy of the ionizing photons.
Figure 2. Single-ion monitoring of (a) m/z ) 138 (parent ion) and (b)
m/z ) 79 (C6H7+) as a function of photon energy (eV). The appearance
energies referred to in the text are derived from extrapolation of the
linear part of the thresholds to the baseline without applying any further
corrections.
After this more general consideration of the experiment, we
also need to comment briefly on the purity of the samples.
According to GC-MS and 1H NMR analysis, in addition to the
desired compounds 3-3b, the samples contain 5-10% of the
corresponding carboxymethyl cyclohexa-1,3-dienes (3′-3b′) and
ca. 5% of methyl benzoate. The latter impurity does not matter
in the present MS/MS experiments because dissociative ioniza-
tion of methyl benzoate does not yield C6H7+. As confirmed
by their EI mass spectra, the isomeric conjugated dienes 3′-
3b′, however, also yield C6H7-nDn+ ions upon photoionization
(though in lower yields, as implied by a comparison of the EI
mass spectra of 3 and 3′).38 Thus, a certain fraction of the
the impurities due to the conjugated dienes should not disturb
the qualitative trends in the energy dependence of the monitoring
reaction (cases a-c in Figure 1).
As a consequence of the impurity of the samples, the parent
ion shows a composite appearance curve (Figure 2a). Of the
two photoionization thresholds, the lower, minor one is assigned
as AE(3′+) ) 8.38 ( 0.04 eV and the higher, major component
as AE(3+) ) 8.83 ( 0.03 eV. Here, the assignment is not only
based on the composition of the sample with 3 as the leading
component but also on a comparison with the corresponding
parent compounds. Thus, the ionization energy (IE) of 1,3-
cyclohexadiene amounts to 8.25 ( 0.02 eV, whereas that of
1,4-cyclohexadiene (8.82 ( 0.02 eV) is significantly larger.39
A slight increase of the IE for 3′ is fully consistent with
+
C6H7-nDn species will have a different origin. Nevertheless,
with 3-3b as the major components of the neutral samples,

Protonated Benzene
J. Phys. Chem. A, Vol. 108, No. 45, 2004 9935
fragment ions formed upon dissociative photoionization. Second,
the trends quite clearly follow the scenario of case b in Figure
1. Thus, even an experiment which was especially designed to
probe the possible existence of a “memory effect” in the H/D
distribution of protonated benzene fails to provide any evidence
of such an effect. Instead, the experiments clearly demonstrate
+
complete H/D equilibration of the mass-selected C6H7-nDn
ions; thus, a rapid hydrogen ring-walk takes place on the time
scale of the experiment. Accordingly, we can compare the BH+/
BD+ ratios of the different precursor ions after statistical
correction for the deuterium content (Figure 3b). At low energies
of the ionizing photons and thus low internal energy content of
+
the C6H7-nDn ions, a KIE on the order of 3 is observed for
both ions made from 3a and 3b, respectively, which then rapidly
decreases, approaching a limiting value of KIE ) 1 at elevated
photon energies. The significant KIEs at low internal energies
are fully consistent with previous studies of the thermal ion/
+
molecule reactions of C6H7-nDn with various bases in which
KIEs between 1.2 and 5.7 were found, dependent on the base
B used.5,6 Likewise, the almost vanishing KIE at higher photon
energies agrees with the finding that the unimolecular loss of
dihydrogen from C6H7-nDn+ ions (reaction 2) is associated with
an almost statistical H/D distribution in the product ions.6,28
C6H7+ f C6H5+ + H2
(2)
From a methodological point of view, we note that the
increasing internal energy content of the mass-selected C6H7-nDn+
ions is also implied by the observation of unimolecular
dehydrogenation according to reaction 2 in the absence of a
reactant gas in the octopole. The apparent threshold for the
unimolecular decay of mass-selected C6H7+ to C6H5+ is about
13.3 ( 0.2 eV and thus ca. 3.5 eV above AE(C6H7+) ) 9.79 (
0.05 eV. These observations can be used to roughly estimate
the energy partitioning in dissociative photoionization of 3. As
the thermodynamic threshold of reaction 2 amounts to 2.79 eV
and it occurs without an appreciable barrier in excess of the
reaction endothermicity,7 we can conclude that up to about four-
Figure 3. (a) BH+/BD+ ratios in the ion/molecule reactions of
+
trimethylamine with mass-selected C6H6D+ (from 3a, 2) and C6H2D5
(from 3b, (), respectively, as a function of the energy of the ionizing
photons. (b) Same BH+/BD+ ratios as a function of photon energy after
statistical correction for the number of H and D atoms in the mass-
+
selected C6H7-nDn ions.
substitution of the conjugated diene with an electron-accepting
ester group. Instead, the presence of a carboxymethyl group in
one of the allylic positions of 1,4-cyclohexadiene hardly affects
the IE. Like the parent ions, also the C6H7+ ions emerging from
dissociative photoionization show a composite threshold be-
fifths of the excess energy of the ionizing photons can be
+ 40
deposited as internal energy in the photofragment C6H7
.
havior with a tentative AE(C6H7+) ) 9.3 eV for the first, minor
Sector-Field Mass Spectrometry. The photoionization ex-
periments described above demonstrate that at least the hypoth-
esis of a significantly delayed hydrogen ring-walk cannot
account for the anomalies reported by Mason and co-workers.2
In addition, none of the many other studies on this topic4-13,24
has so far provided even a hint of evidence for a low-energy
structure other than 2, which could account for their findings.
As stated in the Introduction, quite soon after Mason’s
publication we qualitatively reproduced the reported effect in
that collisional activation of C6HD6+, produced by CI of C6D6
with CH4, shows a temperature dependence in the hydrogen
losses with expulsion of H• being preferred at low ion-source
temperature. While the effect turned out to be much smaller in
our experiments than reported by Mason et al., a qualitative
agreement was nevertheless achieved. We thus presumed to have
confirmed their findings and ascribed the quantitative difference
(much smaller effect) to our limited capability to control the
temperature of the ion source.3 However, in view of the
mounting evidence against protonated benzene having structure
other than 2, we decided to carefully reconsider the sector-field
studies. Of the many experiments having been performed with
different precursors (benzene, 1,4-cyclohexadiene, and 3-3b)
and various ionization methods (EI and CI at different temper-
atures of the ion source and variable kinetic energies of the
+
component smoothly rising out of the noise level and AE(C6H7
)
) 9.79 ( 0.05 eV for the second, larger one (Figure 2b). By
analogy, these two thresholds are assigned to dissociative
photoionization of 3′ and 3, respectively, with 3 giving rise to
the major component having the larger AE. As stated above,
the presence of the isomeric dienes 3a′ and 3b′ in the respective
neutral samples may somewhat modify the quantitative outcome
of the monitor reactions of the labeled ions, while the qualitative
behavior should not be affected; this is all that matters in the
present context. We conclude this description by noting that
the measured AE ) 9.29 ( 0.04 eV for ionized methyl benzoate
agrees reasonably well with the literature value of 9.32 ( 0.04
eV.39
As expected, the C6H7-nDn+ ions generated from the labeled
samples 3a and 3b undergo H+ and D+ transfers with tri-
methylamine. With increasing photon energy, the BH+/BD+
ratios rapidly decrease for both labeled precursors (Figure 3a).
Despite considerable scatter at low photon energies, two
important pieces of information can be extracted directly from
the graph. First, a clear energy behavior is observed, thus
confirming the validity of the experimental approach in that
variation of the energy of the ionizing photons does indeed
markedly affect the internal energy content of mass-selected

9936 J. Phys. Chem. A, Vol. 108, No. 45, 2004
Schro¨der et al.
Figure 4. Ratio of H• and HD losses in the CA spectra (He, 80% T)
of B(1)/E(1) mass-selected C6H6D+ ions (from 3a) as a function of the
kinetic energy of the ionizing electrons without (2) and with (()
correction for contributions of metastable ions. See text for details.
Figure 5. Fraction of H• loss41 in the CA spectra (O2, 80% T) of C6H7
+
+
recorded in a B/E linked scan (2) and of B(1)/E(1) mass-selected C6H7
without (() and with (9) correction for contributions of metastable
ions to the CA spectra as a function of the temperature of the CI source.
The C6H7 ions were generated by dissociative EI of 1,4-cyclohexa-
+
diene.
ionizing electrons), we select just two cases here to illustrate
the various kinds of effects being operative.
pronounced effect in the DADI experiments may also have other
origins (see below).
The first experiment to be described starts with dissociative
EI of neutral 3a, followed by mass selection of C6H6D+ using
B(1)/E(1), collisional activation (He, 80% T) in the field-free
region between E(1) and B(2), and analysis of the fragment
ions with B(2). Here, the variable is the energy of the ionizing
electrons in the range from 10 to 20 eV. As a representative
parameter, we have chosen the ratio of H• and HD losses (in
∆m ) -2, losses of isobaric D• and H2 cannot be resolved).
The raw data (2 in Figure 4) do indeed show a dependence on
the kinetic energy of the ionizing electrons, which is in
qualitative agreement with the effect described in ref 2 (less H•
loss at elevated energies). However, it needs to be acknowledged
that every CA spectrum recorded in a sector-field mass
spectrometer is a superposition of fragmentations due to
unimolecular decay taking place in the entire field-free region
and collision-induced dissociation occurring in the collision cell
only. This is particularly true for the unusually long (1.9 m)
third field-free region of the Berlin BEBE instrument.16
Metastable C6H7+ only undergoes loss of H2, whereas collisional
activation leads to an additional expulsion of H•. The MI
contribution to the CA spectrum will thus appear as an
overestimation of the H2 channel. A correction for the MI
contributions can be made by consideration of the total ion
currents, and doing so, the temperature effect vanishes within
the scatter of the experimental data obtained at different
temperatures of the ion source (( in Figure 4). A simple rationale
for this phenomenon is that the fraction of metastable ions in
the beam increases at elevated energies of the ionizing electrons.
Nevertheless, it may be argued that dissociative EI of 3a already
is too energetic such that the temperature dependence described
by Mason et al. might not be probed adequately in this way.
Figure 5 shows the fraction of H• losses41 upon CA (O2, 80%
T) of C6H7+ generated by dissociative EI of 1,4-cyclohexadiene
as a function of the temperature of the ion source. In a linked
B(1)/E(1) scan (DADI, 2 in Figure 5), a clear temperature
dependence is observed with decreasing H• loss at elevated
temperatures, which is in qualitative agreement with the
observation of Mason et al. The CA spectra of B(1)/ E(1) mass-
selected C6H7+ (( in Figure 5) show somewhat lower amounts
of H• losses, but a slight decrease of this channel with increasing
temperature of the ion source can still be observed; the lower
fraction of H• loss compared to the linked-scan experiment is
again due to the superimposed MI processes. If, however, the
contributions of metastable ions to the CA spectra are corrected
for, the temperature dependence disappears within the scatter
of the data (9 in Figure 5). Thus, much like the electron energy
(Figure 4), the temperature of the ion source appears to increase
the fraction of metastable ions in the C6H7+ ion beam. The more
As far as chemical ionization of benzene with methane as
reagent gas is concerned, it turned out essential to carefully
monitor the C6H6•+/C6H7+ ratio during the experiments because
the 13C1-isotope of ionized benzene interferes with the proto-
nated form. As loss of H• forms the base peak in the CA mass
spectrum of mass-selected C6H6•+, any contribution of 13C1-
•+
C6H6 would give rise to an increase of the H• channel. At
least in our instrument, changing the temperature of the ion
+
source also affects the CI conditions, the C6H6•+/C6H7 ratio
varies as well, and so does the branching between H• and H2
losses. In fact, all we can extract from the various CI experi-
ments having been performed with B(1)/E(1) mass-selected
C6H7-nDn+ ions is that we obtained no obvious evidence for a
+
temperature dependence of the CA spectra of C6H7 which
cannot be attributed to either metastable ion contributions or
isobaric interferences.
Finally, it needs to be commented on the fact that Mason et
+
al. recorded the CA spectra of C6H7 using a linked-scan
technique in a two-sector instrument. Upon reinvestigation of
these experiments with the BE front-end of our mass spectrom-
eter, significant artifact signals are found to appear in the region
of hydrogen losses in both kinds of MI experiments possible
with two sectors only (DADI and MIKE spectra; see experi-
mental details).17 For example, signals at ∆m ) -1 can reach
up to about 20% in DADI and MIKE experiments, whereas the
abundance of unimolecular H• loss is well below 1% of the H2
loss for B(1)/E(1) mass-selected C6H7+. This difference is due
to artifact signals in the DADI and MIKE experiments which
can arise from various fragment ions, isotopologs, or minor other
products formed in the CI plasma. Dependent on the scanning
techniques used and the respective configuration of the instru-
ment, these artifacts may severely obscure the MI spectra.17,42,43
In addition, we note that the amounts of MI and CA contribu-
tions in the DADI experiments are also affected by the CI
conditions because fragmentation is sampled in a field-free
region spatially close to the ionization volume such that the
pressure of the collision gas(es) may vary.
Conclusions
Previous spectroscopic studies in the gas phase have dem-
onstrated that protonated benzene has the structure of the
σ-complex 2.10-12 The present photoionization experiments with
+
mass-selected C6H7-nDn ions intend to address a possible
memory effect as far as the labeling distribution is concerned.
The results obtained demonstrate that the photon energy does
indeed have an effect on the BH+/BD+ ratios, thus proving the

Protonated Benzene
J. Phys. Chem. A, Vol. 108, No. 45, 2004 9937
(13) See also: Kryachko, E. S.; Nguyen, M. T. J. Phys. Chem. A 2001,
105, 153.
(14) Dutuit; O.; Alcaraz, C.; Gerlich, D.; Guyon, P. M.; Hepburn, J.
W.; Metayer-Zeitoun, C.; Ozenne, J. B.; Weng, T. Chem. Phys. 1996, 209,
177.
(15) Franceschi, P.; Thissen, R.; Zˇabka, J.; Roithova´, J.; Herman, Zˇ.;
Dutuit, O. Int. J. Mass Spectrom. 2003, 228, 507.
(16) Schalley, C. A.; Schro¨der, D.; Schwarz, H. Int. J. Mass Spectrom.
Ion Processes 1996, 153, 173.
(17) Busch, K. L.; Glish, G. L.; McLuckey, S. A. Mass Spectrometry/
Mass Spectrometry: Techniques and Applications of Tandem Mass
Spectrometry; VCH: Weinheim, Germany, 1988.
(18) Schro¨der, D.; Oref, I.; Hrusˇa´k, J.; Weiske, T.; Nikitin, E. E.;
Zummack, W.; Schwarz, H. J. Phys. Chem. A 1999, 103, 4609.
(19) In the schemes, we designate only one particular stereoisomer of
the Birch-reduction products for the sake of clarity. In practice, however,
the Birch reaction occurs without a particular diastereoselectivity; see ref
1, p 1010.
(20) According to the photoionization mass spectra acquired in Orsay,
the unlabeled sample of 3 showed an increased content of methyl benzoate,
which we assign to autoxidation during transportation from the Berlin
laboratory where the compounds were prepared.
(21) Kuck, D. Int. J. Mass Spectrom. 2002, 213, 101.
(22) Olah, G. A.; Staral, J. S.; Asencio, G.; Liang, G.; Forsyth, D. A.;
Mateescu, G. A. J. Am. Chem. Soc. 1978, 100, 6299.
(23) Cacace, F.; Crestoni, M. E.; Fornarini, S. J. Am. Chem. Soc. 1992,
114, 6776.
(24) Sumathy, R.; Kryachko, E. S. J. Phys. Chem. A 2002, 106, 510.
(25) In ref 21, the barrier is given with even larger accuracy, 0.35 (
0.01 eV.
validity of the experimental approach. The qualitative trends
observed are fully consistent with a behavior that is typical for
the σ-complex 2. Provided the reasonable assumption holds true
that the precursor compounds used here yield specifically labeled
C6H7-nDn+ ions upon dissociative photoionization near thresh-
old, we can conclude that all H and D atoms in C6H7-nDn+ are
completely equilibrated at ambient temperature. This finding
is fully consistent with a low barrier associated with hydrogen
ring-walk (about 0.35 eV).21 None of the experiments provides
evidence for a delayed hydrogen ring-walk or a structural
memory effect, not to speak of the existence of the π-complex
1.
According to our additional experiments with a sector-field
mass spectrometer, the surprising results of Mason et al. are
likely to be due to the combined action of artifact signals,
isobaric interferences, and metastable ion contributions to the
collisional activation spectra. While we frankly admit that our
capabilities to vary the temperature of the ion source are limited
compared to the equipment used in ref 2, with the present
knowledge it appears that the evidence presented by Mason et
al. in favor of a face-centered π-complex of benzene was in
fact inconclusive.44
From a methodological point of view, the present photoion-
ization experiments demonstrate that the gas-phase titration
method developed for the determination of the barrier of the
keto/enol tautomerization of ionized acetamide32 can also be
applied to fragment ions generated by dissociative photoion-
ization.
(26) For details of the [C2H3O2] system, see: Schro¨der, D.; Semialjac,
M.; Schwarz, H. Eur. J. Mass Spectrom. 2003, 9, 287, and references therein.
(27) Kuck, D.; Bather, W.; Gru¨tzmacher, H.-F. J. Am. Chem. Soc. 1979,
101, 7154.
(28) Kuck, D.; Schneider, J.; Gru¨tzmacher, H.-F. J. Chem. Soc., Perkin
Trans. 2 1985, 689.
(29) Mormann, M.; Kuck, D. Int. J. Mass Spectrom. 2002, 219, 497.
(30) Bock, H. Angew. Chem. 1989, 101, 1659; Angew. Chem., Int. Ed.
Engl. 1989, 28, 1627.
Acknowledgment. Dedicated to Tom Baer on the occasion
of his 65th birthday. Continuous financial support by the
Deutsche Forschungsgemeinschaft, the Fonds der Chemischen
Industrie, and the Gesellschaft von Freunden der Technischen
Universita¨t Berlin is gratefully acknowledged. Further, we thank
Waltraud Zummack for the synthesis of compounds 3-3b and
the staff of LURE for operating the Orsay Super-ACO ring and
general facilities.
(31) For a similar situation in the related C6H6•+ system, see: Baer, T.;
Willett, G. D.; Smith, D.; Phillips, J. S. J. Chem. Phys. 1979, 70, 4076.
(32) Schro¨der, D.; Loos, J.; Schwarz, H.; Thissen, R.; Dutuit, O.;
Mourgues, P.; Audier, H.-E. Angew. Chem. 2002, 114, 2867; Angew. Chem.,
Int. Ed. 2002, 41, 2748.
(33) McAdoo, D. J.; Morton, T. H. Acc. Chem. Res. 1993, 26, 295.
(34) All proton affinities are taken from: Hunter, E. P. L.; Lias, S. G.
J. Phys. Chem. Ref. Data 1998, 27, 413.
(35) For a more detailed discussion of KIEs, see: Derrick, P. J.; Donchi,
K. F. in ComprehensiVe Chemical Kinetics, Vol. 24; Bamford, C. H., Tipper,
C. F. H., Eds.; Elsevier: Amsterdam, 1983; p 53.
References and Notes
(36) To a first approximation, secondary KIEs are neglected.
(37) Note that the spectroscopic studies of gaseous C6H7 ions clearly
reveal structure 2 but cannot address the question of a delayed hydrogen-
(1) Smith, M. B.; March, J. March’s AdVanced Organic Chemistry,
5th ed.; Wiley: New York, 2001.
(2) Mason, R. S.; Williams, C. M.; Anderson, P. D. J. J. Chem. Soc.,
Chem. Commun. 1995, 1027.
(3) Schro¨der, D.; Loos, L.; Semialjac, M.; Weiske, T.; Schwarz, H.;
Ho¨hne, G.; Thissen, R.; Dutuit, O. Int. J. Mass Spectrom. 2002, 214, 155.
(4) Glukhovtsev, M. N.; Pross, A.; Nicolaides, A.; Radom, L. J. Chem.
Soc., Chem. Commun. 1995, 2347.
(5) Chiavarino, B.; Crestoni, M. E.; DePuy, C. H.; Fornarini, S.;
Gareyev, R. J. Phys. Chem. 1996, 100, 16201.
(6) DePuy, C. H.; Gareyev, R.; Fornarini, S. Int. J. Mass Spectrom.
1997, 161, 41.
(7) Bouchoux, G.; Ya´n˜ez, M.; Mo´, O. Int. J. Mass Spectrom. 1999,
185/186/187, 241.
(8) Ascenzi, D.; Bassi, D.; Franceschi, P.; Tosi, P.; Di Stefano, M.;
Rosi, M.; Sgamellotti, A. J. Chem. Phys. 2003, 119, 8366.
(9) See also: Reed, C.; Kim, K.-C.; Stoyanov, E. S.; Stasko, D.; Tham,
F. S.; Mueller, L. J.; Boyd, P. D. W. J. Am. Chem. Soc. 2003, 125, 1796.
(10) Solca, N.; Dopfer, O. Angew. Chem., Int. Ed. 2002, 41, 3628.
(11) Jones, W.; Boissel, P.; Chiavarino, B.; Crestoni, M. E.; Fornarini,
S.; Lemaire, J.; Maitre, P. Angew. Chem., Int. Ed. 2003, 42, 2057.
(12) Solca, N.; Dopfer, O. Chem. Eur. J. 2003, 9, 3154.
+
ring walk.
(38) MS (EI, 70 eV) of 3, m/z (%): 138 (4) [M+], 137 (4) [M+ H],
-
107 (4) [M+ - CH3O], 79 (100) [C6H7+], 78 (45) [C6H6+], 77 (55) [C6H5+],
59 (5) [C2H3O2+], and 51 (8) [C4H3+]. MS (EI, 70 eV) of 3′, m/z (%): 138
(50) [M+], 137 (60) [M+ - H], 107 (35) [M+ - CH3O], 79 (100) [C6H7+],
78 (30) [C6H6+], 77 (95) [C6H5+], 59 (10) [C2H3O2+], and 51 (15) [C4H3+].
(39) Unless mentioned otherwise, all data are taken from the NIST
Standard Reference Database No. 69, March 2003 Release; see: http://
webbook.nist.gov/chemistry/.
(40) See also: Murray, P. T.; Baer, T. Int. J. Mass Spectrom. Ion Phys.
1979, 30, 165.
(41) Normalized to the sum of H• and H2 losses; i.e., xH ) I([M - H]+)/
(I([M - H]+) + I([M - H2]+)).
(42) Boyd, R. K.; Porter, C. J.; Beynon, J. H. Int. J. Mass Spectrom.
Ion Phys. 1982, 44, 199, and references therein.
(43) See also: (a) Schro¨der, D.; Schwarz, H. Int. J. Mass Spectrom.
Ion Processes 1995, 146/147, 183. (b) Mouget, Y.; Bertrand, M. J. Rapid
Commun. Mass Spectrom. 1995, 9, 387.
(44) Maddox, J. Nature 1994, 369, 97.