Non-thermal internal energy distribution of ions observed in an electrospray source interfaced with a sector mass spectrometer

Article abstract of DOI:10.1002/jms.1873

The internal energy distribution P(E_int) of ions emitted in an electrospray (ESI) source interfaced with a sector mass spectrometer is evaluated by using the experimental survival yield (SY) method including the kinetic shift. This method is ba

Full text of DOI:10.1002/jms.1873

Research Article
Received: 11 June 2010
Accepted: 16 November 2010
Published online in Wiley Online Library: 21 January 2011
(wileyonlinelibrary.com) DOI 10.1002/jms.1873
Non-thermal internal energy distribution
of ions observed in an electrospray source
interfaced with a sector mass spectrometer
David Rondeau,a∗ Nicolas Galland,b∗ Emilie-Laure Zins,c Claude Pepe,c
d
d
´
´
´
´
Laszlo Drahos and Karoly Vekey
The internal energy distribution P(Eint) of ions emitted in an electrospray (ESI) source interfaced with a sector mass spectrometer
is evaluated by using the experimental survival yield (SY) method including the kinetic shift. This method is based on the
relationship between the degree of fragmentation of an ion and its amount of internal energy and uses benzylpyridinium
cations due to their simple fragmentation scheme. Quantum chemical calculations are performed, namely at G3(MP2)//B3LYP
and QCISD/MP2 levels of theory. The results show that the internal energy distribution of the ions emitted in the ESI source
interfaced with a sector analyzer is very narrow. The MassKinetics software is used to confirm these observations. The P(Eint
)
is the parameter that allows to fit the experimental SY of each substituted benzylpyridinium cation with theoretical mass
spectra generated by the MassKinetics software. The resulting internal energy distributions are similar to the ones obtained
with the experimental SY method. This indicates that in the present experimental conditions, P(Eint) cannot be compared with
a ‘thermal-like’ Boltzmann distribution. In addition, it appears that with the sector analyzer, increasing the collision energy in
c
the first pumping stage of the ESI source does not correspond to a warm-up of the produced ions. Copyright ꢀ 2011 John Wiley
& Sons, Ltd.
Keywords: electrospray ionization; internal energy; RRKM; sector mass spectrometer; thermometer ions
Introduction
skimmer. As the skimmer tip is located upstream of the Mach
disc, the velocity of the ions is high with an axial component
increased towards the radial velocity and with low gas density in
this so-called silence zone. If the skimmer is positioned outside the
Mach disc, the pressure is higher but the gas velocity is lower.[10]
In most cases, additional focusing items are introduced between
the orifice and the skimmer in order to achieve a good efficiency
in the ion sampling, but they redefine both the internal and the
The mass spectra of fragment ions produced in an ion source
depend on the internal energy distribution P(Eint) of the precursor
ions, the energetic barrier of the unimolecular reaction and the
time window corresponding to the observation of the dissociation
process.[1,2] In electron impact (EI) ionization, a ionizing electron
kinetic energy of 70 eV leads to a typical internal energy
distribution that lies until ∼16 eV with a maximum around 4 eV.[1]
It is admitted that this distribution is reproducible whatever the
used instrument due to the low pressure of the source and
the ion acceleration zone. The EI mass spectra can then be
compiled to create databases. For the other ionization modes,
the energetic characteristics of the produced ions are less well
defined than for EI. Among these methods, electrospray ionization
(ESI) is becoming one of the most used ionization modes.[3]
Using ESI sources, the ions are emitted in the gas phase after
their formation in solution and will cross regions of more or
less high pressures before entering in the mass analyzer.[4–6] In
these regions, their internal energy can be modified depending
on different experimental conditions related to the geometry
and the operation parameters of the interface (temperature,
pressure, applied voltages, etc.). The transport of ions from the
ion source into the interface requires small apertures to separate
the atmospheric pressure from the vacuum of the interface.[7]
This orifice produces a well-defined supersonic jet in expansion
in which ions and gas are accelerated to supersonic velocities.[8]
This barrel-shaped expanding jet terminates in a shock wave also
referred to as the Mach disc.[9] The ions carried by the expanding
gas are usually sampled into the next vacuum stage through a
∗
´
Correspondence to: David Rondeau, Laboratoire de Chimie, Electrochimie
´
´
Moleculaires et Chimie Analytique, CEMCA UMR CNRS 6521, Universite
´
`
Europeenne de Bretagne a Brest, UFR Sciences et Techniques, 6 Avenue Victor
Le Gorgeu - C.S. 93837, 29238 Brest Cedex 3, France.
E-mail: david.rondeau@univ-brest.fr
´
`
Nicolas Galland,LaboratoiredeChimieetInterdisciplinarite:Synthese,Analyse,
´
´
´
Modelisation, CEISAM, UMR CNRS 6230, Universite de Nantes – Faculte des
`
Sciences et des Techniques, 2 rue de la Houssiniere, BP 92208, 44322 Nantes
Cedex 3, France. E-mail: nicolas.galland@univ-nantes.fr
´
´
LaboratoiredeChimie,ElectrochimieMoleculairesetChimieAnalytique,CEMCA
a
b
c
´
´
`
UMR CNRS 6521, Universite Europeenne de Bretagne a Brest, UFR Sciences et
Techniques,6AvenueVictorLeGorgeu-C.S.93837,29238BrestCedex3, France
´
`
´
Laboratoire de Chimie et Interdisciplinarite: Synthese, Analyse, Modelisation,
´
´
CEISAM, UMR CNRS 6230, Universite de Nantes - Faculte des Sciences et des
`
Techniques, 2 rue de la Houssiniere, BP 92208, 44322 Nantes Cedex 3, France
´
´
Laboratoire de Dynamique Interactions et Reactivite, LADIR UMR CNRS 7075,
´
Universite Pierre et Marie Curie, Case Courrier 49, 4 Place Jussieu, Paris 75005,
France
d
MassSpectrometryDepartment, CenteroftheHungarian AcademyofSciences,
Puszteszeri ut 59-67, H-1025 Budapest, Hungary
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.

Non-thermal internal energy distribution in ESI/MS
translational energies of the electrosprayed ions.[11] Indeed, if the
desolvationprocesscanbeperformedintheatmosphericpressure
sidebydryingtheionswithheatordrygas,theremovingofthelast
solvent molecules is rather achieved in the vacuum side by means
ofcollision-induceddissociation(CID).Thisprocessisperformedby
in-source CIDs (in-source CID) with the residual gas and controlled
by the application of voltage differences between the extraction
lenses of this ‘nozzle-skimmer region’.[12,13] In this first pumping
stage, the influence of pressure on the transmission efficiency of
ions[14] and the desolvation conditions has been described.[15,16]
The production of net ions from a supra-thermal-like desolvation
use of dry gas. On one hand, the internal energy distribution of
ions emitted was evaluated using the experimental SY method of
De Pauw et al. by including the kinetic shift. On the other hand,
the theoretical SYs of the thermometer ions were calculated by
using the MassKinetics software. This one generates theoretical
mass spectra by considering the internal energy distribution of the
ions as an adjusting parameter. In order to point out the important
parameters to be considering for the evaluation of P(Eint), the
principles of the SY method are first described in this paper.
process can also be in competition with ion fragmentation due to Methods
collisional activation. In the case of source decays, the in-source
The survival yield SY of the parent ion M+ is related to th+e
CID mass spectra are then considered as an additional tool for
structural determination.[17–19] However, before the entrance in
the mass analyzer, the ions also pass through a second pumping
stage where the pressure is dramatically lowered. This region is
designed for ion transmission by means of ion guides (quadrupole
or hexapole lens). Due to the main free path of ions, this zone has
been described as having high collision efficiency.[20] Increasing
the pressure in this region is also considered as having a strong
beneficial effect on the ion signal intensity.[21]
+
abundance of the parent ion IM and the fragment ions IF
measured from the recorded ion source mass spectra such as
IP
+



SY =
(1)
n
ꢀ
+ 
IP
+
IF
+
+
F
=1
This is the fraction of the parent ions that have an internal energy
below the appearance energetic threshold Ethr of the fragment
ions. (Note that for the present paper the denomination ‘appear-
ance energetic threshold’ is preferred to the term appearance
energy, this one being rather defined as the minimum ionizing
electron energy necessary to observe a fragment ion in EI/MS.)
For constructing an ESI in-source CID mass spectra library, it is
then important to calibrate in internal energy the recorded mass
spectrum.[22] The results reported in the literature show that the
internal energy distribution P(Eint) of the ions in an ESI source can
be compared with ‘thermal-like’ Boltzmann distributions related
to a simple parameter called the ‘characteristic temperature’
(Tchar).[23–25] The method proposed for evaluating this internal
energy distribution was referenced by De Pauw and coworkers as
the ‘survival yield’ (SY) method.[26] This one can be described on
the basis of relationship between the degree of fragmentation
of an ion and its amount of internal energy.[27] The SY
method used namely substituted benzylpyridinium cations due
to their simple fragmentation scheme that leads to the loss
of neutral pyridine and the formation of a substituted benzyl
cation.[28] The experimental measurement of ion SY has been
performed from ESI sources interfaced with different analyzers,
such as quadrupoles,[25–27] Fourier transformation ion cyclotron
resonance cells[29] and time–of-flight mass spectrometers.[30] All
these sources use nitrogen as nebulizing and dry gas leading to
thermal-like internal energy distribution when the desolvation
conditions are properly adjusted.[31,32] An empirical equation has
been recently proposed for evaluating the mean internal energy
of the ions emitted in an ESI source. This equation takes into
account different parameters such as the cone voltage value, the
source temperature, the ion degree of freedom and the ion mean
internal energy considered from the source temperature at a zero
cone voltage value.[33] Tabet and coworkers have also shown
that the De Pauw method can be combined with the use of the
MassKinetics software of Drahos and Ve´key in order to evaluate
the internal energy distribution of gaseous ions generated in
some ESI sources.[34] With the MassKinetics software, a theoretical
SY has been calculated from different species (benzylpyridinium
cations, aromatic benzoic esters and peptides) and for different
internal energy distributions.[35] However, at our knowledge, if
the reactivity of the benzylpyridinium has been studied with an
electrospray source interfaced with a sector mass spectrometer,
no resulthas been reported in the literature concerning an internal
energy distribution evaluated under these conditions.[36]
Log[k(Eint)]
(a)
p-CH3O
p-CH3
Log(1/τ)
m-CH3
p-NO2
E
int
Survival Yield
(b)
-
x
1
p-NO2
*
m-CH3
p-CH3
*
*
p-CH3O
*
-
x
0
Ethr
P(Eint
)
(c)
E
int
Figure 1. Illustration of the procedure used in the survival yield method
(SY) including the kinetic shift (ks) for evaluating of the ion internal energy
distribution from the reactivity of different substituted benzylpyridinium
cations. (a) Evolution of the logarithm of the dissociation rate constant
of cation selected for the study as a function of their internal energy
(Eint). (b) Experimental survival yield of thermometer ions (∗) reported
as a function of appearance threshold energy (Eapp) calculated from
the residence time (τ) of ions in the ESI source. (c) Ion internal energy
distribution (P(Eint)) obtained from the derivate of the curve plotted from
a fit of the data in (b) with a sigmoid function. Note that this illustration
shows also the addition of the two points (x) required for the fit with a
sigmoid curve at SY = 0 and SY = 1.
In this study, an electrospray source interfaced with a JEOL
sector mass spectrometer was used. This source acts without the
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

D. Rondeau et al.
As illustrated in Fig. 1, the different values of SY are plotted as
a function of the ‘appearance energetic threshold’. Two points
are added manually by considering that all the precursor ions are
fragmented if Ethr = 0 and that all the parent ions survive at a high
values of Ethr. The fitting of the points with a sigmoid curve allows
to obtain the internal energy distribution via the derivation of the
SY(Eint) curve. The value of Ethr is obtained from the kinetic shift
(Eks) that represents the excess of energy by comparison to the
critical energy of the reaction E0, necessary to fragment the ions
at observable rates, such as
benzylpyridinium ion is equivalent to the value of ꢀ(ꢀHf) for the
reaction, such as
E0 = ꢀ(ꢀHf) = {[ꢀHf(Py) + ꢀHf(R+)] − ꢀHf(Py+R)}
(4)
(Note that this equation does not consider that critical energy is
at 0 K, while the enthalpy is usually room temperature. The critical
energy can be, however, determined from the standard enthalpy,
which can be measured easily, using a set of equations developed
by Pople and coworkers.[45]) In addition, with these ‘thermometer’
ions, the evaluation of the internal energy deposited on the
precursor ion is based on the following assumptions:
Eks = Ethr − E0
(2)
(1) any ionization energy has to be taken into account in the
production of the gaseous ions since the ions emitted into
the gas phase result from a charge separation in solution;
(2) the internal energy is of vibrational mode for all the ions;
(3) the ions have similar mass, structures, degree of freedom
number and center of mass collision energy;
(4) the intake of internal energy per vibrational mode is the same
for all the studied ions in the same experimental conditions
due to the thermal equilibrium conditions in an atmospheric
pressure source.
It has been shown that Eks can be calculated using the
Rice–Ramsperger–Kassel–Marcus (RRKM) theory.[37–39] This one
allowstoexpresstherateconstantofanionfragmentationreaction
at a given internal energy Eint such as
σ N‡(Eint − E0)
k(Eint) =
(3)
h
ρ(Eint)
In Eqn (3), σ is the reaction degeneracy, h is the Planck’s constant,
E0 is the reaction critical energy, N‡(Eint − E0) is the transition
state sum of states from 0 to Eint − E0 and ρ(Eint) is the density
The first studies reported in the literature referred to either
published ꢀHf values or E0 values that were correlated from
the ionization energies of the corresponding substituted benzyl
radicals.[26,27] The critical energy values have also been evaluated
from semi-empirical (AM1) or chemical quantum calculation at
different theoretical levels (ab initio HF/6-31G∗ and DFT B3LYP/6-
31G).[25] The variations between the predicted values have shown
the importance of the choice of the method and the level of
calculation.[11]
The appearance energetic threshold Ethr of each substituted
benzylpyridinium cation is obtained through the relation between
the logarithm of the rate constant and the corresponding internal
energy, the value of k(Eint) being the reciprocal of the residence
time of ions in the source (Fig. 1). Note that the value of Ethr is
related to different theoretical and experimental parameters that
can be listed as follows:
[2,40]
of states at an energy level equal to Eint
.
The evaluation
of the sum and density of states is usually done by a direct
count using the Beyer–Swinehart algorithm that uses frequency
models obtained by theoretical calculation from the reactant and
the activated complex at the transition state of the reaction.[41]
The fundamental assumption of the experimental SY method
is that the fragmentation of the benzylpyridinium cations PyR+
occurs only via a direct cleavage of the C–N bond between
the substituted benzyl group and the pyridine Py. Even though
recent theoretical and experimental systematic studies have
shown that fragmentations with rearrangement can compete
with the direct cleavage,[42–44] the SY method considers that the
dissociation of the benzylpyridinium ions takes place without any
reverse activation energy (Fig. 2). The absence of reverse energy
implies that the critical energy (E0) for the fragmentation of a
N
N
N
R
R
R
ks
E0
Reaction Coordinate
Figure 2. Dissociation scheme of the substituted benzylpyridinium cations and potential energy evolution of the fragmentation as a function of the
C–N central bond distance considered as the reaction coordinate. The parameters E0 and ks are related to the critical energy and the kinetic shift of the
reaction.
c
wileyonlinelibrary.com/journal/jms
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2011, 46, 100–111

Non-thermal internal energy distribution in ESI/MS
(1) the critical energy of the reaction E0 and the molecular
frequencies of the reactant ion obtained from the theoretical
calculation, their values being related to the level of
calculation;
(2) the magnitude of k(Eint) curve that is a function of the
frequencies of the reaction transition and ground states, its
evaluation being often characterized through the looseness
of the transition state;
a pressure close to that of the analyzer (10−6 Torr). In the case of
Fig. 3, a 120-V potential difference between the skimmer and the
counter electrode is set, but a 90-V voltage value was also used
in the study as indicated in the Section on Results and Discussion.
This potential difference is also called the cone-voltage value (CV).
Chemicals
(3) the residence time of ion in the ESI interface that determines
the ks.
The substituted benzylpyridinium salts (Table 1) were prepared
by condensation of the substituted benzyl halide with pyridine.
The salts precipitated after a few hours of reflux heating and were
recrystallized with diethyl ether. Their structure was confirmed by
NMR spectroscopy.
Experimental Section
The benzylpyridinium salts depicted in Table 1 were dissolved
in a CH3OH/H2O (1 : 1) solvent mixture at a 5 × 10−5 M concen-
tration. Two solutions were separately injected. The first con-
tained the N-(4-methoxybenzyl)pyridinium ion (p-CH3O), the N-
(4-methylbenzyl)pyridinium ion (p-CH3), the N-(benzyl)pyridinium
ion (H) and the N-(4-fluorobenzyl)pyridinium ion (p-F). The sec-
ond contained the N-(3-methoxybenzyl)pyridinium ion (m-CH3O),
the N-(3-methylbenzyl)pyridinium ion (m-CH3) and the N-(4-
nitrobenzyl)pyridinium ion (p-NO2).
Mass spectrometry experiments
Mass spectrometry analyses were performed on a JMS-700 (Jeol
Ltd, Akishima, Tokyo, Japan) double focusing mass spectrometer
with reversed geometry, equipped with a pneumatically assisted
ESI source. Nitrogen was used as the nebulizer gas. Positive
ion mode was used with a needle voltage adjusted in order
to obtain a ∼500 nA needle current (iESI), i.e. typically ∼3 kV.
The electrospray source and the optical path of the JMS 700 are
schematicallydescribedinFig. 3. Thepotentialdifferencebetween
the desolvating plate and the first skimmer, i.e. the orifice 1 or OR1,
canbevariedfrom0to300 V. Threedesolvatingplatetemperature
values were chosen for the study, i.e. 423, 473 and 523 K, the
orifice 1 temperatures being then stabilized as a function of these
temperatures, i.e. 353, 361 and 371 K, respectively. No voltage was
applied to the ring lens (RL in Fig. 3). This one is used to collect the
ions along the center axis of the source. Three parts are particularly
depicted in Fig. 3. The first indicates the skimmer region in the first
pumping stage. A rough pump (650 l min−1) is used to evacuate
the first region to about 2 Torr. The second part is related to the
transfer quadrupole zone in which the pressure is decreased to
above 10−3 Torr by means of a turbomolecular pump. The last
part corresponds to the last pumping stage that limits the region
of ion acceleration undergoing a potential difference of 5000 V at
Quantum chemistry
All gas-phase calculations were carried out with the Gaussian03
suite of programs.[46] Owing to the large number of possible
reactions, rapid explorations of the potential energy surfaces (PES)
have been performed using HF and B3LYP methods in conjunction
with the 6-31G∗ basis set. Significantly different results were
obtained for the direct C–N bond cleavage of several substituted
benzylpyridinium cations. This has led us to consider small model
systems which will be presented in the Section on Results and
Discussion. For the model systems, we have realized fully relaxed
scans for C–N bond breaking as well as geometry optimizations
for the stationary points. HF, MP2, CISD and B3LYP methods have
been retained for this purpose in combination with the 6-31+G∗∗
Sampling Cone (SC) Skimmer Lens (SL)
Skimmer (S)
Counter electrode (CE)
RF Hexapole (H0)
Needle
MS
For a 3 kV
needle voltage
3500 V
500 V
120 V
125 V
2.10-3
m
0.11 m
Figure 3. Schematic representation of the ESI source and optical path of the JMS 700 mass spectrometer, showing the distance between the different
parts of the interface that is considered for the calculation of the residence time and the applied voltages. Note that for the present example, a 120-V
potential difference is considered in the declusterization area delimited by the lenses OR1 and OR2.
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

D. Rondeau et al.
Table 1. Substituted benzylpyridinium salts used for the study
Structure
Compound name
Cation name (m/z)
m-CH3O (200)
N-(3-methoxybenzyl)pyridinium chloride
N-(4-nitrobenzyl)pyridinium bromide
N
N
MeO
O2N
Cl
Br
p-NO2 (215)
p-Br (248)
N-(4-bromobenzyl)pyridinium bromide
Br
N
Br
N-(3-methylbenzyl)pyridinium bromide
N-(4-methylbenzyl)pyridinium bromide
m-CH3 (184)
p-CH3 (184)
N
Me
Me
Br
Br
N
N-(4-methoxybenzyl)pyridinium chloride
N-(4-fluorobenzyl)pyridinium bromide
Benzylpyridinium bromide
p-CH3O (200)
p-F (188)
MeO
F
N
Cl
N
Br
BzPyr (170)
N
Br
basis. In the case of substituted benzylpyridinium cations and
theirs C–N bond cleavage products, the geometries, energies and
vibrational properties were recomputed at the B3LYP/6-31+G∗∗
and MP2/6-31+G∗∗ levels of theory. In particular, the harmonic
frequencies were computed in order to characterize the stationary
points (minima vs transition state structures (TS)) and to estimate
theirs thermodynamic parameters. Gibbs energy of dissociation
has been estimated at a 350 K temperature and a 1 atm pressure
according to the reaction described in Scheme 1.
where {A} emphasizes calculation over the population of all con-
formers of A. Gibbs energy of the reactions was then refined
by using the results of G3(MP2)//B3LYP and QCISD/6-31+G∗∗
calculations. G3(MP2)//B3LYP is a modification of G3 extrapo-
lation procedure that requires significantly less computational
time.[47] The average absolute deviation from experiment of
G3(MP2)//B3LYP method for a total of 299 energies (enthalpies of
formation, ionization potentials, electron affinities, proton affini-
ties) is 1.25 kcal mol−1. As bromo-substituted species could not be
studied with G3(MP2)//B3LYP method, total energies of the vari-
ous species were recomputed at the QCISD/6-31+G∗∗ theoretical
level using previously MP2/6-31+G∗∗ optimized structures. For
the sake of simplicity, Gibbs energies calculated using QCISD/6-
31+G∗∗//MP2/6-31+G∗∗ energies and MP2/6-31+G∗∗ structural
andvibrationalpropertieswillbereferredtoQCISD//MP2through-
outthetext.Notethatthecoreelectronswerefrozenforallpost-HF
calculations.
Note that some of the studied species exhibit several conform-
ers. Hence, Gibbs energies of these species have been evaluated
through a Boltzmann distribution according to relation (5)
ꢁ
ꢀ
RT
G0{A} = −RT ln
e−G
(5)
i
i∈{A}
Theoretical modeling of reaction rate constants and mass
spectra
+
N
+
The fragmentation rate constants and the theoretical SY of
the substituted benzylpyridinium cations were calculated us-
ing the MassKinetics Scientific (Ver. 1.9) software, available at
http://www.chemres.hu/ms/masskinetics. MassKinetics is a gen-
eral framework for modeling mass spectrometric processes.[35] It
CH2
N
H
+
R
H
R
Scheme 1. Unimolecular reaction of the substituted benzylpyridinium
cations.
c
wileyonlinelibrary.com/journal/jms
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2011, 46, 100–111

Non-thermal internal energy distribution in ESI/MS
combines modeling reaction rates, energy exchanges and prod- Results and Discussion
uct ion abundances as the ions move through various parts of
a mass spectrometer. In MassKinetics, the ions are characterized
Chemical quantum calculation
by their internal and kinetic energies, which define the ‘state’ of
an ion. Essential features of the MassKinetics model are the use
of internal (and, if necessary, kinetic) energy distributions and
the use of probabilities to describe transition between different
states. Reaction rates are calculated via the transition state theory
(TST)[48,49] in its RRKM formulation.[37–40,50,51] The kinetic and/or
internal energies of ions in the gas phase may change due to
(1) acceleration in electromagnetic fields, (2) radiative energy ex-
change (photon-absorption and emission), (3) collisional energy
exchange and (4) energy partitioning in chemical reactions, which
are all taken into account. With MassKinetics, the ion abundances
and thus the theoretical SY can be calculated accurately using no
or only very few empirical (adjustable) parameters. MassKinetics
modeling requires the vibrational frequencies of each substituted
benzylpyridinium cation obtained from theoretical calculation
(vide infra), and the harmonic oscillator model is applied with
internal rotations approximated by low-frequency vibrations. In
the present work, the internal energy distribution has been con-
sidered as an adjustable parameter. As a result, the theoretical
SY of each studied benzylpyridinium cation was fitted against the
experimental SY by using either a thermal-like internal energy
distribution or an internal energy distribution from on adjusted
Gaussian function. This distribution is supposed to be character-
istic of a gaseous ion population sampled in front of the orifice 2
(Fig. 3).
In the present study, we have investigated the mechanism
of the direct bond cleavage between N and C atoms of the
substituted benzylpyridinium cations. It appeared for several
benzylpyridinium ions that the consideration of a transition state
along the reaction coordinate during the C–N bond stretching
depends on the level of theory used: HF/6-31G∗ or B3LYP/6-31G∗.
This result prompted us to consider model systems to assess the
underlying mechanisms of the effective dissociation process.
Mechanism of the direct C–N bond cleavage
In order to reach a reasonable level of theory, we have selected
three relatively small model systems (Scheme 2). The chosen
model systems allow us to simulate with a reasonable accuracy
the structural variations of the real systems. Figure 4 displays
the energy variation of the system 1 computed at various levels
of theory when the C–N bond dissociation proceeds (all other
geometrical parameters being relaxed). We observed that the
whole ab initio calculations predict the dissociation without any
potential energy barrier to form CH3+ and NH3 species. In contrast,
at B3LYP/6-31+G∗∗ level of theory, the dissociation mechanism
involves a TS located at dC–N = 6.0 Å. A closer examination of
the corresponding dissociation curve reveals two highly abnormal
features that should be pointed out: (1) the total energy of the
fragments produced by the bond cleavage is not identical to, but
muchlowerthan,thesumofthetotalenergiesoftheNH3 molecule
and the CH3+ cation, (2) the positive charge is well delocalized at
the two fragments even at very large C–N distances. This behavior
is definitely artifactual and is typical of the so-called ‘inverse
symmetry breaking’ phenomenon.[56] Inverse symmetry breaking
is an inherent character of the existing DFT functionals that lead
to wrong dissociation limit and the appearance of fictitious TS.[57]
In the case of the model systems 2 and 3, B3LYP/6-31+G∗∗
calculations experienced again an inverse symmetry breaking
phenomenon when the C–N bond dissociates. Fictitious TS were
then localized on the corresponding PES. HF results, as well as MP2
and CISD ones which include electronic correlation contributions,
show no barrier on the energy curves corresponding to the C–N
bond dissociation. In fact, we speculate that the direct C–N bond
cleavage does not involve any TS, neither for the three model
systems nor for the substituted benzylpyridinium cations.
The kinetic energy of the species prior to their acceleration is
relatedtothepotentialdifferencebetweentheorifices1and2, e.g.
120 V in the case of Fig. 3. The MS experimental sequence fixed for
the calculation of a theoretical SY from an ESI ionization interfaced
with the JMS 700 Jeol magneto-electrostatic mass spectrometer
is (1) the ion formation without any internal energy deposited
during this process; (2) a field-free flight from the orifice 2 to the
orifice 3 with a 0.28720 m flight length (Fig. 3), (3) an electrostatic
acceleration under a 5 kV potential difference with a 1.6-cm flight
length and (4) the ion selection followed by the detection (from
the acceleration lens to the detector) with a 1.9656 m flight length.
The critical energy of the reaction has been calculated using the
G3(MP2)//B3LYP or the QCISD//MP2 methodologies (vide supra).
The frequency model of the ground state that allows to express
ρ(Eint) is obtained from the vibration frequency values calculated
at the G3(MP2)//B3LYP or the QCISD//MP2 levels of theory. It is
moredifficulttoevaluatethevibrationfrequenciesoftheactivated
complex that lead to the sum of state N‡(Eint − E0) in Eqn (4). In
the case of the fragmentation of the benzylpyridinium ions, one
defines a ‘loose’ complex corresponding to a direct cleavage due
to a stretching of the C–N bond along the reaction coordinate.
The ‘looseness’ of the transition state can be characterized by the
Arrhenius pre-exponential-like factor APE that takes into account
the frequency models of the reactants and the transition states
through the expressions of the vibration entropy differences. For
the substituted benzylpyridinium fragmentation, the expected
APE value is considered to be close to 14.[1,34,52,53] In this context,
the frequency file of the transition state is calculated from the
frequency model of the reactants as follows: the frequency
Let us i+llustrate our argument on the example of system
1, CH3NH3 (X1A1). The expected products of the C–N bond
cleavage are CH3+ and NH3. These two species can also be bound
through an hydrogen-bond involving a polarized +C–H bond
and the N atom. ꢁThe structure of the hydrogen-bond complex,
+· · ·
1
+
CH3
NH3 (X A ), is displayed with that of CH3NH3 species in
Fig. 5. The ab initio calculations reveal that, when the C–N bond
dissociation proceeds, the PES crosses the PES corresponding to
+· · ·
the dissociation of CH3
NH3 complex. This is shown in Fig. 6:
at the CISD/6-31+G∗∗ level of theory, the two surfaces cross at
3
2
1
H
H
H
corresponding to the elongation of the C–N bond (1200 cm−1
)
+
H
NH3
H
H
H
+
+
H
H3C NH3
is removed and five other frequencies are scaled to obtain the
expected APE value, i.e. two at ∼800, two at ∼600 and one at
∼100 cm−1. This is a convenient way for simulating the transition
state.[54,55]
N
H
H
H
H
H
H
H
Scheme 2. Model systems.
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

D. Rondeau et al.
_
C
N dissociation
H
model system 2
model system 1
R
+
R
R
400
200
0
C
N
R
450
300
150
0
R
H ···N dissociation
R
N
+
R
R
R
C
H
R
MP2/6-31+G**
CISD/6-31+G**
HF/6-31+G**
model system 3
B3LYP/6-31+G**
1
5
9
1
6
11
C-N (Å)
16
dC-N (Å)
d
Figure 6. Potentialenergycurves, computedattheCISD/6-31+G∗∗ levelof
theory, corresponding to the dissociation of model systems 1, 2 and 3. For
eachsystem,theC–Nbonddissociationisrepresentedbyacontinuousline
Figure 4. Potential energy curve, computed at different levels of theory,
corresponding to the C–N bond dissociation of model system 1 depicted
in Scheme 2.
· · ·
and the hydrogen bond (N H) dissociation is represented by a dashed
line.
225
175
C3
v
h
+
+· · ·
Figure 5. Structures of CH3NH3 (C3v) and CH3
NH3 (Cs) species
125
75
MP2/6-31+G**
G3(MP2)//B3LYP
QCISD//MP2
B3LYP/6-31+G**
AE
represented with some symmetry elements.
dC–N = 3.4 Å and the energy is well below the asymptote. It
p
p
p
1
p
p
o
p
m
is an avoided crossing (going from C3v to Cs symmetry, the A1
p
m
electronic state of CH3NH3+ reduced to 1Aꢁ, which correlates with
substituents carried by the phenyl group
+· · ·
the electronic state of CH3
NH3), meaning that the C–N bond
Figure 7. Gibbs energy of dissociation (kJ mol−1), computed at different
levels of theory, of various substituted benzylpyridinium species. Available
appearanceenergies(AE)correspondingtoC–Nfragmentionsappearance
are reported with uncertainty range. AE values are issued from collisionally
activated dissociation experiments using FTICR mass spectrometry. The
substituted benzylpyridinium species were prepared with an internal
energy corresponding to a nominal temperature of 350 K.[58]
cleavage will shift to an hydrogen-bond dissociation mechanism.
+
Hence, CH3 and NH3 species could be formed from an initial
C–N bond cleavage without encountering any potential energy
barrier. The same dissociation mechanism was found for model
systems 2 and 3 as illustrated in Fig. 6. This should remain valid for
the substituted benzylpyridinium cations; hence, the direct C–N
bond cleavage does not involve any TS.
shows that G3(MP2)//B3LYP and QCISD//MP2 methodologies
are the best performers: the average unsigned deviations from
experiment are, respectively, 3.6% and 3.8%. The two others AE
valuesreportedfortheN-(4-bromobenzyl)pyridiniumandtheN-(4-
chlorobenzyl)pyridinium differ severely from the predicted Gibbs
energies. On one hand, experimental values are far below the
average computed values (the agreement with B3LYP/6-31+G∗∗
results seems fortuitous). On the other hand, all the calculations
agree to predict a greater Gibbs energy of dissociation for the N-(4-
bromobenzyl)pyridinium and the N-(4-chlorobenzyl)pyridinium
than for the N-(4-methylbenzyl)pyridinium, whereas the AE values
do not follow at all this trend. The best sounding explanation for
this discrepancy is that the measured AE values correspond to
another reaction than the one depicted in Scheme 1, the C–N
bond cleavage of the N-(4-bromobenzyl)pyridinium and the N-(4-
chlorobenzyl)pyridinium could notably yield tropylium cations (in
conjunction with the formation of pyridine) as evidenced for other
substituted benzylpyridinium ions.[42]
C–N bond dissociation energies
As no barrier exists on the PES corresponding to the direct
C–N bond cleavage of the substituted benzylpyridinium ions,
the critical energy E0 for this mechanism will be to assimilate
the heat of reaction computed at T = 0 K. The question arising is:
whichcomputationalmethod(s)shouldberetainedtogetaccurate
values? Fig. 7 displays for 11 benzylpyridinium cations the Gibbs
energy of dissociation computed at different levels of theory. For
several cations, the appearance energies (AE) corresponding to
C–N fragment ions appearance have been obtained by Katritzky
et al.[58] and are also presented for comparison purpose.
Going from the N-(4-methoxybenzyl)pyridinium to the N-(4-
nitrobenzyl)pyridinium, all calculations agree to predict the same
trend.Nevertheless,somedifferencesappearifwelookindividually
at each substituted benzylpyridinium ions: the computed Gibbs
energies of reaction can differ by more than 65 kJ mol−1. B3LYP/6-
31+G∗∗ calculations yield systematically the lowest values while
the greatest ones are obtained at the MP2/6-31+G∗∗ level of
theory. Focusing on the non-substituted benzylpyridinium, the N-
(4-methoxybenzyl)pyridinium, the N-(4-methylbenzyl)pyridinium
Evaluation of the internal energy distribution of the electro-
sprayed ions
and the N-(4-nitrobenzyl)pyridinium cations, the comparison As illustrated in Fig. 1, the experimental SY method including
between the AE values and the computed Gibbs energies clearly the kinetic shift uses the appearance threshold energy values for
c
wileyonlinelibrary.com/journal/jms
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2011, 46, 100–111

Non-thermal internal energy distribution in ESI/MS
a function of the appearance threshold energies (Fig. 8). This
voltage value has been selected with regard to the SYs of the N-(4-
methoxybenzyl)pyridinium and the N-(4-nitrobenzylpyridinium)
of6.4×10−3 and0.99, respectively. Theseexperimentaldataavoid
to add manually any point at the 0 and 1 ordinates. This confirms
that the SY measurements fully characterize the internal energy
distribution of the ESI source used for this study. The experimental
SY of each benzylpyridinium cation has been plotted as a function
oftheEth values.Theseonesarecalculatedfromthetwoconsidered
kinetic shifts, i.e. skimmer-tip downstream or upstream the Mach
disc. The obtained SY(Eint) data of Fig. 8 are then fitted with single,
symmetrical sigmoids (Eqn 8):
a
y =
(8)
e−(x−x )/b + 1
0
With these experimental data, we did not need to use the
procedure involving a fit refined by the use of two sigmoids as
described by Gabelica et al. in the case of the P(Eint) determination
in a heated capillary nano-electrospray source.[30] The position
of the first point relative to the SY of the most fragile N-(4-
methoxybenzyl)pyridinium cation is close to 0. This observation
confirms that under the experimental conditions of the present
study, this ion population is entirely desolvated and has enough
energy to completely fragment. The position of the last point
that characterizes an undissociated N-(4-nitrobenzyl)pyridinium
ion allows to envisage a symmetric distribution of the internal
energy. The curves obtained with the fitting procedure are shown
in Fig. 8. The best fit is obtained for a = 1 in Eqn (8) and with
the Ethr values calculated from molecular parameters obtained at
the QCISD//MP2 level of theory. If a = 1 in Eqn (8), the sigmoid
does not fit the data for SY(Eint) values close to the unity. In
addition, with the QCISD//MP2 calculation, the fit is obtained
with a standard deviation ≥0.99, whereas with the Ethr values
obtained from the G3(MP2)//B3LYP methodology, the standard
deviation does not exceed 0.84. The curves reported in Fig. 8
show that whatever the selected kinetic shift, the experimental
SY values related to the electrosprayed cations are well localized
on the sigmoid parts that will determine either the shape and the
width of the resulting P(Eint). The results of this single sigmoid
fitting procedure are represented in Fig. 8 only in the case of the
QCISD//MP2 methodology.
Figure 8. Experimental survival yields of the seven different substituted
benzylpyridinium cations plotted as a function of the different threshold
energy values calculated by considering either no initial kinetic energy
for the ions emitted in the ESI interface (open circle) or a supersonic jet
sampled by the skimmer (black squares). Note that the experimental data
have been obtained with a 120-V cone voltage and a 150 ◦C ESI source
temperature.
evaluating P(Eint). The Ethr parameter can be defined as the energy
at which the rate constant for the dissociation k(Eint) is equal to
1/τ, where τ is the residence time of the ion in the ESI interface
before entering the acceleration zone. The residence time of
each benzylpyridinium cation determines the time scale of the
experiment required to observe their fragmentation. Depending
on whether the ions are sampled by the skimmer aperture
downstream or upstream the Mach disc, the velocity v in the
interface will be expressed using either Eqn (6) or (7) in which an
additionaltermrelatedtoasupersonicvelocityvalueisconsidered:
ꢂ
√
2
v =
e · ꢀV
(6)
(7)
mu
2
ꢂ
ꢃ
ꢅ
ꢄ
vMach
=
KEinitial + (e · ꢀV)
mu
In Eqns (6) and (7), the common parameters are the atomic mass
unit u, the elementary charge e, the potential difference ꢀV
between orifices 1 and 2 (Fig. 3) and the mass of the ion, m.
In Eqn (7), KEinitial is the initial kinetic of ions sampled in the
expansion jet by considering there velocity value close to Mach
1, i.e. 343 m s−1. The obtained τ values correspond to the sum of
the residence time of an ion formed at OR1, accelerated under
a potential difference corresponding to the CV value and the
time of flight in the ion guide (Fig. 3). The residence time of
the substituted benzylpyridinium cations has been calculated by
taking into account the mass of the seven studied species. From
Eqn (6) (skimmer-tip outside the Mach disc), the mean value is
of 27.25 µs with a 1.039 standard deviation. From Eqn (7) (ions
sampled in the silence zone), the mean value is of 0.1425 µs
with a 0.0054 standard deviation. The approximate rate constant
necessary to produce fragments observed on the ion source mass
spectrum is equivalent to the reciprocal of the residence time
of ions. As a result, the two residence time values have been
considered for the estimation of the different Ethr from the RRKM
The derivates of the calculated sigmoids are expressed as
follows: the b and x0 parameters being extracted from the fit
procedure using Eqn (9):
dy
dx
e−(x−x )/b
0
=
(9)
b(e−(x−x )/b + 1)2
0
The internal energy distribution then obtained is reported in
Fig. 9. From the depicted curves, it appears that the SY method
including the kinetic shift allows to evaluate two different internal
energy distributions as a function of the assumed skimmer tip
position. For ions sampled outside the Mach disc, where the
density of gas is high and the ion velocity is low, the mean
internal energy should be dramatically lower than the P(Eint) of
the benzylpyridinium ions sampled in a supersonic jet. In the
first case, this P(Eint) is characterized by a maximum at 5.385 eV,
whereas in the second case, P(Eint) is centered around the 7.544 eV
value. Nevertheless, in both cases, the full widths at half maximum
(FWHM) are of low values (Fig. 9). The internal energy distribution
of the ions emitted in an ESI source interfaced with a sector
curves log[k(Eint)] versus Eint
.
TheexperimentalSYsofthesevensubstitutedbenzylpyridinium
cations recorded at a 120 V orifice 1 voltage value are plotted as
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

D. Rondeau et al.
Figure 10. (a) Internal energy distributions calculated by MassKinet-
ics for (– • •–) the N-(4-methoxybenzyl)pyridinium ion (p-CH3O),
Figure 9. Internal energy distribution derivate from the curves of Fig. 8
using Eqn (10) and by considering: (• • •) no initial kinetic energy of the
ions and ( – – – ) a supersonic velocity for the initial kinetic energy of
the ions. The centroid (§) and the full width half maximum (∗) of each
distribution are noted directly above the corresponding curves.
(
) the N-(4-methylbenzyl)pyridinium ion (p-CH3), (• • •) the N-
(4-fluorobenzyl)pyridinium ion (p-F), (
ion (H), (
) the N-(benzyl)pyridinium
) the N-(3-methylbenzyl)pyridinium ion (m-CH3), (
)
the N-(3-methoxybenzyl)pyridinium ion (m-CH3O) and (• • •) the N-(4-
nitrobenzyl)pyridinium ion (p-NO2). (b) Linear regression calculated from
the plot of the theoretical SY versus the experimental SY by using the inter-
nal energy distributions depicted in (a) for the calculation of the theoretical
SY.
mass spectrometer is thus very narrow in comparison with the
P(Eint) reported in the literature in the case of a quadrupole mass
spectrometer.[26] For the latter case, a maximum of the P(Eint
)
for describing the internal energy of the electrosprayed ions. This
is in agreement with one of the assumptions mentioned in the
case of the use of the experimental SY method (see Section on
Methods).
curve at ∼5 eV is associated with a FWHM of the distribution of
∼6 eV. Such a result obtained with the experimental SY method
is in agreement with a thermal-like distribution for ions at 1480 K.
In order to confirm one of the two P(Eint) reported in Fig. 9, the
As the observed internal energy distribution cannot be
compared with a ‘thermal-like’ Boltzmann distribution, it could be
interesting to verify if an increase in the potential difference in the
declustering area leads to a warm-up of ions as often described
in the literature for other ESI sources.[25,26] With MassKinetics,
the internal energy distribution of the electrosprayed ions has
been evaluated at different OR1 voltage values (Fig. 3). For each
voltage, one substituted benzylpyridinium cation could be used as
thermometer ion with regard to the linear regression of Fig. 10(b)
that allows to consider at least this ion as representative of the
overall population. In our case, we have chosen to consider
MassKinetics software was used. The aim was to obtain a P(Eint
)
similar for all the cations when the theoretical SY calculated by
MassKinetics is equal to the experimental SY. In this case, the
choice of a same temperature parameter for describing the initial
internal energy does not allow to fit the theoretical SY and the
experimental SY. The internal energy distributions that allow to
calculate,foreachofthesevencations,atheoreticalSYequaltothe
experimental SY, are not thermal but dramatically narrower. The
internal energy distributions of the substituted benzylpyridinium
cations considered as a fitting parameter in MassKinetics are
reported in Fig. 10.
two ions in order to compare the evolution of the P(Eint
)
Each curve corresponds to the P(Eint) of ions produced in the
gas phase and sampled by the orifice 2 after their activation in the
declusterization area (Fig. 3). The FWHM of the curves P(Eint) versus
Eint is 0.47 eV for all the substituted benzylpyridinium cations with
a maximum of P(Eint) that can be evaluated at 5.15 eV 3% as
a function of the studied specie (Fig. 10(a)). For each cation, the
theoretical SY values fit with the experimental data as illustrated
by Fig. 10(b) where a linear regression is obtained with a standard
deviation R = 0.999. The internal energy distributions calculated
by MassKinetics (Fig. 10) correspond to P(Eint) obtained by the way
of the experimental SY method including the kinetic shift (Fig. 9)
when the ions are sampled outside the Mach disc. One can remark
that the values of the maximum of the curves and the FWHM are
similar. Such results confirm first that the ions electrosprayed in
the sector mass spectrometer used for this study are characterized
by a very narrow internal energy distribution. This one cannot be
compared in a first approximation to a thermal distribution. Then,
this comparison shows that the two used SY methods are relevant
for such a study if the kinetic shift and the theoretical parameters
are properly chosen. This kinetic shift corresponds to ions sampled
upstream the Mach disc. At least, the linear regression depicted in
Fig. 10(b) suggests that only one thermometer ion can be chosen
with the variation of the source acting parameters such as the
temperature and the voltage value. Among the selected ions, the
benzylpyridinium and the N-(3-methylbenzyl)pyridinium cations
are the common thermometer ions for measurements performed
at three different temperatures by keeping the voltage value
constant, respectively, at 90 and 120 V. The evaluated internal
energy distributions that allow to obtain the same experimental
and theoretical SY in MassKinetics are illustrated in Fig. 11. The
comparison of the SY values obtained for the two species selected
amongthesevensubstitutedbenzylpyridinium cationsisreported
in Table 2. From the P(Eint) curves of Fig. 11, it appears that an
increase in the potential difference between OR1 and OR2 of
about 30 V leads to a shift of P(Eint) to the high-energy values
rather than a broadening of the curves. Indeed, for a source
temperature of 423 K, the increase in the voltage from 90 to
120 V leads to a shift of the maximum of the curves from 4.4 to
5.25 eV, respectively. By contrast, for a same voltage value, e.g.
90 V, the change in temperature, i.e. from 423 to 523 K, leads to
an increase in the FWHM value from 0.41 to 0.83 eV, whereas the
maximum of the curves is only shifted of 0.4 eV. With regard to
the broader P(Eint) that has been evaluated in Fig. 11, it must be
c
wileyonlinelibrary.com/journal/jms
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2011, 46, 100–111

Non-thermal internal energy distribution in ESI/MS
Figure 11. Internal energy distribution calculated by MassKinetics as a function of the source temperature and the potential difference in the
declusterization area (∗90 V and #120 V). Note that the value in brackets indicates the temperature of orifice 1.
Table 2. Comparison of the experimental and the theoretical survival yields obtained for two substituted benzylpyridinium as the cone voltage
value and the temperatures of the ESI interface (see Experimental Section)
Source temperature
Orifice 1
(orifice 1 temperature)
Experimental SY
Theoretical SY
voltage (V)
423 K (353 K)
473 K (361 K)
523 K (371 K)
0.894 (p-CH3)
0.96 (H)
0.775 (m-CH3)
0.906 (H)
0.878 (p-CH3)
0.325 (H)
0.96 (H)
90
120
90
0.325 (H)
0.797 (p-CH3)
0.248 (H)
0.74 (m-CH3)
0.91 (H)
0.75 (p-CH3)
0.246 (H)
0.367 (m-CH3)
0.742 (H)
0.377 (m-CH3)
0.436 (H)
120
90
0.485 (p-CH3)
0.085 (m-CH3)
0.479 (p-CH3)
0.045 (m-CH3)
0.128 (m-CH3O)
0.127 (m-CH3O)
120
mentioned that for a 523-K temperature and an orifice 1 voltage
value of 120 V, a FWHM value of only ∼1 eV is observed. Such
narrow internal energy distributions have also been disclosed in
Fig. 9. The increase in the voltage in the declustering area cannot
be assimilated in a first approximation to a warm-up of the ion
population in all the vibration modes.
not comparable to a thermal distribution for an ESI source of a
quadrupole mass spectrometer.[25–27] The shape of the internal
energy distribution is narrower than a thermal distribution usually
calculated from the vibrational state density of the thermometer
ions.[24] The interpretation of such a behavior is not trivial. It can
be regarded either through the acting of a source having no
curtain gas incoming in the first pumping stage or through the
characteristics of the analyzer used for the study. In the case of
the JEOL ESI device, the disclosed non-thermal internal energy
distribution might be related to ions sampled by the skimmer
downstream the Mach disc, i.e. into the axis of the expansion jet.
Indeed, the selected population should have frozen translational,
rotational and vibrational temperatures with a relaxation process
of low efficiency for some vibration modes.[59,60] However, this
is not the case, because the ions should have an initial velocity
close to the Mach number with an internal energy distribution
in agreement with a short resident time of ions in the source.
The comparison of the two P(Eint) obtained shows in fact that the
distributions are comparable only for an ion sampling outside the
Mach disc, i.e. as the randomized trajectory of ions is taken into
Conclusions
The approach that consists in combining the experimental SY
method of De Pauw et al. and the theoretical SY method using
MassKinetics software of Drahos and Ve´key allows to evaluate
two comparable internal energy distributions of substituted ben-
zylpyridinium cations electrosprayed in an ESI source interfaced
with a sector mass analyzer. This is particularly relevant when
quantum chemical calculations are performed at the QCISD/6-
31+G∗∗//MP2/6-31+G∗∗ level of theory for the selected ther-
mometer ions. The methods used for evaluating P(Eint) in the
JEOL ESI source disclose that the internal energy distribution is
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

D. Rondeau et al.
account. Byreportingtothecommentsoftheintroductionsection,
the probing of such ions downstream the Mach disk should lead
rather to a thermal-like internal energy distribution as previously
described for the ESI source of the PE SCIEX API 165 quadrupole
mass spectrometer.[27] The JMS 700 ESI source used for this study
presents the same characteristics as the PE SCIEX. By focusing
onto the differences in the analyzer used for these studies, one
can report the works of Anderson et al.[61] They have shown that
in a quadrupole mass spectrometer acting with an ESI source, the
direct current field between the quadrupole rods is responsible to
a radially directed acceleration of the parent ions. The broadening
in kinetic energy is then converted in a broadening in internal
energy by collisions with the neutrals due to the jet formation in
the expansion area of the ESI source. In the case of the SY method,
some substituted benzylpyridinium cations should be lost into
the analyzer as parent ions, whereas the benzylium ions produced
in the source are more likely to survive in the quadrupole. Such
an interpretation involving the analyzer rather than the source
can find a confirmation through the results obtained by Gabelica
et al.[30] In this paper devoted to the internal energy distribution
of benzylpyridinium ions emitted from a nano-electrospray source
interfaced with TOF analyzer, distributions remain narrow what-
ever the gas pressure in the first pumping stages, the voltage
and temperature conditions. The conclusions of the authors con-
cerning the fact that ‘low temperature–high acceleration voltage’
conditions are not equivalent to ‘high temperature–low accelera-
tion voltages’ can be applied to the results shown in Fig. 11, in the
case of the sector mass spectrometer.
[11] V. Gabelica, E. De Pauw. Internal energy and fragmentation of ions
produced in electrospray sources. Mass Spectrom. Rev. 2005, 24,
566.
[12] N. Potier, P. Barth, D. Tritsch, J.-F. Biellmann, A. Van Dorsselaer.
Study of non-covalent enzyme–inhibitor complexes of aldose
reductase by electrospray mass spectrometry. Eur.J.Biochem. 1997,
243, 274.
[13] H. Rogniaux, A. Van Dorsselaer, P. Barth, J.-F. Biellmann,
J. Barbanton, M. van Zandt, B. Chevrier, E. Howard, A. Mitschler,
N. Potier, L. Urzhumtseva, D. Moras, A. Podjarny. Binding of aldose
reductase inhibitors: correlation of crystallographic and mass
spectrometricstudies – asystemforX-raycrystallographyandNMR.
J. Am. Soc. Mass Spectrom. 1999, 10, 635.
[14] R. D. Smith, C. J. Barinaga. Internal energy effects in the collision-
induced dissociation of large biopolymer molecular ions produced
by electrospray ionization tandem mass spectrometry of
cytochrome c. Rapid Commun. Mass Spectrom. 1990, 4, 54.
[15] G. Wang, R. B. Cole. Solution, gas-phase, and instrumental
parameter influences on charge-state distributions in electrospray
ionization mass spectrometry. In Electrospray Ionization Mass
Spectrometry: Fundamentals, Instrumentation
& Applications,
R. B. Cole (Ed.). John Wiley & Sons: New York, 1997, 137.
[16] A. Schmidt, U. Bahr, M. Karas. Influence of pressure in the first
pumping stage on analyte desolvation and fragmentation in nano-
ESI MS. Anal. Chem. 2001, 73, 6040.
[17] O. Lapre´vote, P. Ducrot, C. Thal, L. Serani, B. C. Das. Stereochemistry
of electrosprayed ions. Indoloquinolizidine derivatives. J. Mass
Spectrom. 1996, 31, 1149.
[18] W. D. Van Dongen, J. I. T. van Wijk, B. N. Green, W. Heerma,
J. Haverkamp. Comparison between collision induced dissociation
ofelectrosprayedprotonatedpeptidesintheup-frontsourceregion
and in a low-energy collision cell. Rapid Commun. Mass Spectrom.
1999, 13, 1712.
[19] B. B. Schneider, D. D. Y. Chen. Collision-induced dissociation of
ions within the orifice-skimmer region of an electrospray mass
spectrometer. Anal. Chem. 2000, 72, 791.
[20] L. Serani, D. Lemaire, O. Lapre´vote. Collision efficiency in an
electrospray source interfaced with a magnetic mass spectrometer.
Int. J. Mass Spectrom. 2002, 219, 403.
Acknowledgements
The authors are grateful to Dr Lionel Sanguinet from the University
of Angers (CIMA UMR CNRS 6200) for the gift of benzylpyridinium
salts. This work was performed using HPC resources from GENCI-
CINES/IDRIS (Grant 2009-x2009085117) and CCIPL (Centre de
Calcul Intensif des Pays de la Loire).
[21] D. J. Douglas, J. French. Collisional focusing effects in radio
frequency quadrupoles. J. Am. Soc. Mass Spectrom. 1992, 3, 398.
[22] W. Weinmann, M. Stoertzel, S. Vogt, M. Svoboda, A. Schreiber.
Tuning compounds for electrospray ionization/in-source CID and
mass spectra library searching. J. Mass Spectrom. 2001, 36, 1013.
[23] L. Drahos, K. Ve´key. Determination of the thermal energy and its
distribution in peptides. J. Am. Soc. Mass Spectrom. 1999, 10, 323.
[24] Z. Taka´ts, L. Drahos, G. Schlosser, K. Ve´key. Feasibility of formation
of hot ions in electrospray. Anal. Chem. 2002, 74, 6427.
[25] L. Drahos, R. M. A. Heeren, C. Collette, E. De Pauw, K. Ve´key. Thermal
energy distribution observed in electrospray ionization. J. Mass
Spectrom. 1999, 34, 1373.
[26] C. Collette, E. De Pauw. Calibration of the internal energy
distribution of ions produced by electrospray. Rapid Commun.
Mass Spectrom. 1998, 12, 165.
[27] C. Collette, L. Drahos, E. De Pauw, K. Ve´key. Comparison of the
internal energy distributions of ions produced by different
electrospray sources. Rapid Commun. Mass Spectrom. 1998, 12,
1673.
References
[1] K. Ve´key Internal energy effects in mass spectrometry. J. Mass
Spectrom. 1996, 31, 445.
[2] T. Baer, P. M. Mayer. Statistical Rice–Ramsperger–Kassel–Marcus
quasiequilibrium theory calculations in mass spectrometry. J. Am.
Soc. Mass Spectrom. 1997, 8, 103.
[3] C. M. Whitehouse, R. N. Dreyer, M. Yamashita, J. B. Fenn. Electro-
spray interface for liquid chromatographs and mass spectrometers.
Anal. Chem. 1985, 57, 675.
[4] S. J. Gaskell. Electrospray: principles and practice. J. Mass Spectrom.
1997, 32, 677.
[5] P. Kebarle. A brief overview of the present status of the mechanisms
involvedinelectropsraymassspectrometry. J.MassSpectrom. 2000,
35, 804.
[6] R. B. Cole. Some tenets pertaining to electrospray ionization mass
spectrometry. J. Mass Spectrom. 2000, 35, 763.
[28] A. G. Harrison. Fragmentation reaction of alkylphenylammonium
ions. J. Mass Spectrom. 1999, 34, 1253.
[29] X. Guo, M. C. Duursma, P. G. Kistemaker, N. M. M. Nibbering,
K. Ve´key, L. Drahos, R. M. A. Heeren. Manipulating internal energy
of protonated biomolecules in electrospray ionization Fourier
transform ion cyclotron resonance mass spectrometry. J. Mass
Spectrom. 2003, 38, 597.
[7] A. P. Bruins. ESI source design and dynamic rang considerations.
In Electrospray Ionization Mass Spectrometry: Fundamentals,
Instrumentation & Applications, R. B. Cole (Ed.). John Wiley & Sons:
New York, 1997, 107.
[30] V. Gabelica, E. De Pauw, M. Karas. Influence of the capillary
temperature and the source pressure on the internal energy
distribution of electrosprayed ions. Int. J. Mass Spectrom. 2004,
231, 189.
[8] R. Campargue. Progress in overexpanded supersonic jets and
skimmed molecular beams in free-jet zones of silence. J. Phys.
Chem. 1984, 88, 4466.
[9] J. B. Fenn. Mass spectrometric implications of high-pressure ion
sources. Int. J. Mass Spectrom. 2000, 200, 459.
[31] F. He, J. Ramirez, B. A. Garcia, C. B. Lebrilla. Differentially heated
capillary for thermal dissociation of noncovalently bound
complexes produced by electrospray ionization. Int. J. Mass
Spectrom. 1999, 182(183), 261.
[32] A. Schmidt, U. Bahr, M. Karas. Influence of pressure in the first
pumping stage on analyte desolvation and fragmentation in nano-
ESI MS. Anal. Chem. 2001, 73, 6040.
[10] T. R. Covey, B. A. Thomson, B. B. Schneider. Atmospheric pressure
ion sources. Mass Spectrom. Rev. 2009, 28, 870.
c
wileyonlinelibrary.com/journal/jms
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2011, 46, 100–111

Non-thermal internal energy distribution in ESI/MS
[33] A. Pak, D. Lesage, Y. Gimbert, K. Ve´key, J.-C. Tabet. Internal energy
distribution of peptides in electrospray ionization: ESI and collision-
induced dissociation spectra calculation. J. Mass Spectrom. 2008,
43, 447.
[34] J. Naban-Maillet, D. Lesage, A. Bosse´e, Y. Gimbert, J. Szta´ray,
K. Ve´key, J.-C. Tabet. Internal energy distribution in electrospray
ionization. J. Mass Spectrom. 2005, 40, 1.
[35] L. Drahos, K. Ve´key. MassKinetics: a theoretical model of mass
spectra incorporating physical processes, reaction kinetics and
mathematical descriptions. J. Mass Spectrom. 2001, 36, 237.
[36] V. Gabelica,D. Lemaire,O. Lapre´vote,E. DePauw.Kineticsofsolvent
addition on electrosprayed ions in an electrospray source and in
quadrupole ion trap. Int. J. Mass Spectrom. 2001, 210/211, 113.
[37] H. M. Rosenstock, M. B. Wallenstein, A. L. Wahrhaftig, H. Eyring.
Absolute theory for isolated systems and the mass spectra of
polyatomic molecules. Proc. Natl. Acad. Sci. USA 1952, 38, 667.
[38] R. A. Marcus, O. K. Rice. The kinetics of the recombination of methyl
radicals and iodide atoms. J. Phys. Colloid Chem. 1951, 55, 894.
[39] R. A. Marcus. Unimolecular dissociations and free radical
recombination reactions. J. Chem. Phys. 1952, 20, 359.
A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez,
J. A. Pople. Gaussian 03, Revision C.02. Gaussian, Inc.: Wallingford,
CT, 2004.
[47] A. G. Baboul, L. A. Curtiss, P. C. Redfern, K. Raghavachari. Gaussian-3
theoryusingdensityfunctionalgeometriesandzero-pointenergies.
J. Chem. Phys. 1999, 110, 7650.
[48] H. Eyring. The activated complex in chemical reactions. J. Chem.
Phys. 1935, 3, 107.
[49] M. G. Evans, M. Polanyi. Somme applications of the transition state
method to the calculation of reaction velocities, especially in
solution. Trans. Faraday Soc. 1935, 31, 875.
[50] M. R. Wright. Fundamental Chemical Kinetics, An Explanatory
Introduction to the Concepts. Horwood Publishing Limited:
Chichester, 1999.
[51] T. Baer, W. L. Has. Unimolecular Reaction Dynamics. Oxford
University Press: Oxford, 1996.
[52] J. L. Holmes, C. Aubry, P. M. Mayer. Proton affinities of primary
alkanols: an appraisal of the kinetic method. J. Phys. Chem. A
1999, 103, 705.
[53] P. D. Thomas, R. G. Cooks, K. Ve´key, L. Drahos, C. Wesdemiotis.
Comments on ‘‘proton affinities of primary alkanols: an appraisal of
the kinetic method’’. J. Phys. Chem. A 2000, 104, 1359.
[54] R. C. Dunbar. New approaches to ion thermochemistry via
dissociation and association. In AdvancesinGasPhaseIonChemistry,
vol. 2, N. G. Adams, L. M. Babcock (Eds). JAI Press: Greenwich, 1996,
87.
[40] P. J. Robinson, K. A. Holbrook. Unimolecular Reactions. Wiley:
Chichester, 1972.
[41] T. Beyer, D. F. Swinehart. Algorithm 448. Number of multiply-
restricted partitions [A1]. ACM Commun. 1973, 16(6), 379.
[42] E. L. Zins, C. Pepe, D. Rondeau, S. Rochut, N. Galland, J. C. Tabet.
Theoretical and experimental study of tropylium formation from
substituted benzylpyridinium species. J. Mass Spectrom. 2009, 44,
1.
[43] K. V. Barylyuk, K. Chingin, R. M. Balabin, R. Zenobi. Fragmentation of
benzylpyridinium‘‘thermometer’’ionsanditseffectontheaccuracy
of internal energy calibration. J. Am. Soc. Mass Spectrom. 2010, 21,
172.
[44] E. L. Zins, D. Rondeau, P. Karoyan, C. Fosse, S. Rochut, C. Pepe.
Investigations of the fragmentation pathways of benzylpyridinium
ions under ESI/MS conditions. J. Mass Spectrom. 2009, 44, 1658.
[45] J. E. Del Bene, H. D. Mettee, M. J. Frisch, B. T. Luke, J. A. Pople.
Ab initio computation of the enthalpies of some gas-phase
hydration reactions. J. Phys. Chem. 1983, 87, 3279.
[46] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, J. A. Montgomery Jr., T. Vreven, K. N. Kudin,
J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson,
H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene,
X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma,
[55] V. Gabelica, E. Schultz, M. Karas. Internal energy build-up in matrix-
assisted laser desorption/ionization. J. Mass Spectrom. 2004, 39,
579.
[56] T. Bally, G. N. Sastry. Incorrect dissociation behavior of radical ions
in density functional calculations. J. Phys. Chem. A 1997, 101, 7923.
[57] Y. Zhang, W. Yang. A challenge for density functionals: self-
interaction error increases for systems with a noninteger number
of electrons. J. Chem. Phys. 1998, 109, 2604.
[58] A. R. Katritzky, C. H. Watson, Z. Dega-Szafran, J. R. Eyler. Kinetics
and mechanism of nucleophilic displacements with heterocycles
as leaving groups. 26. Collisionally activated dissociation of N-
alkylpyridinium cations to pyridine and alkyl cations in the gas
phase. J. Am. Chem. Soc. 1990, 112, 2471.
[59] L. K. Randeniya, M. A. Smith. A study of molecular supersonic flow
using the generalized Boltzmann equation. J. Chem. Phys. 1990, 93,
661.
[60] S. DePaul, D. Pullman, B. Friedrich. A pocket model of seeded
supersonic beams. J. Phys. Chem. 1993, 97, 2167.
[61] S. G. Anderson, A. T. Blades, J. Klassen, P. Kebarle. Determination
of ion-ligand bond energies and ion fragmentation energies
of electrospray-produced ions by collision-induced dissociation
threshold measurements. Int. J. Mass Spec. Ion Proc. 1995, 141, 217.
G. A. Voth,
P. Salvador,
J. J. Dannenberg,
V. G. Zakrzewski,
S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui,
A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu,
c
J. Mass. Spectrom. 2011, 46, 100–111
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms