PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Mid-infrared vibrational spectra of discrete acetone-ligated cerium
hydroxide cations
G. S. Groenewold,a A. K. Gianotto,b K. C. Cossel,c M. J. Van Stipdonk,d
J. Oomens,e N. Polfer,f D. T. Moore,g W. A. de Jongh and M. E. McIlwaini
Received 11th September 2006, Accepted 30th November 2006
First published as an Advance Article on the web 19th December 2006
DOI: 10.1039/b613029a
Cerium(III) hydroxy reactive sites are responsible for several important heterogeneous catalysis
processes, and understanding the reaction chemistry of substrate molecules like CO, H2O, and
CH3OH as they occur in heterogeneous media is a challenging task. We report here the first infrared
spectra of model gas-phase cerium complexes and use the results as a benchmark to assist evaluation
of the accuracy of ab initio calculations. Complexes containing [CeOH]21 ligated by three- and fouracetone molecules were generated by electrospray ionization and characterized using wavelengthselective infrared multiple photon dissociation (IRMPD). The CQO stretching frequency for the
[CeOH(acetone)4]21 species appeared at 1650 cm1 and was red-shifted by 90 cm1 compared to
unligated acetone. The magnitude of this shift for the carbonyl frequency was even greater for the
[CeOH(acetone)3]21 complex: the IRMPD peak consisted of two dissociation channels, an initial
elimination of acetone at 1635 cm1, and elimination of acetone concurrent with a charge separation
producing [CeO(acetone)]1 at 1599 cm1, with the overall frequency centered at 1616 cm1. The
increasing red shift observed as the number of acetone ligands decreases from four to three is
consistent with transfer of more electron density per ligand in the less coordinated complexes. The
lower frequency measured for the elimination/charge separation process is likely due to a
combination of: (a) anharmonicity resulting from population of higher vibrational states, and (b)
absorption by the initially formed photofragment [CeOH(acetone)2]21. The C–C stretching frequency
in the complexes is also influenced by coordination to the metal: it is blue-shifted compared to bare
acetone, indicating a slight strengthening of the C–C bond in the complex, with the intensity of the
absorption decreasing with decreasing ligation. Density functional theory (DFT) calculations using
three different functionals (VWN, B3LYP, and PBE0) were used to predict the infrared spectra of
the complexes. Calculated frequencies for the carbonyl stretch are within 40 cm1 of the IRMPD of
the three-acetone complex measured using the single acetone loss, and within 60 cm1 of the
measurement for the four-acetone complexes. The B3LYP functionals provided the best agreement
a
Idaho National Laboratory, Idaho Falls, ID, USA.
E-mail: gary.groenewold@inl.gov; Fax: 01 208 526 8541;
Tel: 01 208 526 2803
b
Idaho National Laboratory, Idaho Falls, ID, USA.
E-mail: anita.gianotto@inl.gov; Fax: 01 208 526 8541;
Tel: 01 208 526 0551
c
Idaho National Laboratory, Idaho Falls, ID, USA.
E-mail: cosselk@caltech.edu; Fax: 01 270-918-8574;
Tel: 01 626 395 1059
d
Wichita State University, Wichita, KS, USA.
E-mail: mike.vanstipdonk@wichita.edu; Fax: 01 316 978 3431;
Tel: 01 316 978 7381
e
FOM Instituut voor Plasmafysica, Nieuwegein, The Netherlands.
E-mail: joso@rijnh.nl; Fax: 30 6031 204;
Tel: 30 6096 999
f
FOM Instituut voor Plasmafysica, Nieuwegein, The Netherlands.
E-mail: polfer@rijnh.nl; Fax: 30 6031 204;
Tel: 30 6096 999
g
FOM Instituut voor Plasmafysica, Nieuwegein, The Netherlands.
E-mail: dtmoore@berkeley.edu; Fax: 30 6031 204;
Tel: 01 510 486 5741
h
Pacific Northwest National Laboratory, Richland, WA, USA.
E-mail: bert.dejong@pnl.gov; Fax: 01 509 376 0420;
Tel: 01 509 376 5290
i
Idaho National Laboratory, Idaho Falls, ID, USA.
E-mail: michael.mcilwain@inl.gov; Fax: 01 208 526 8541;
Tel: 01 208 526 8130
596 | Phys. Chem. Chem. Phys., 2007, 9, 596–606
with the measured spectra, with
the VWN modestly lower and
PBE0 modestly higher. The C–C
stretching frequencies calculated
using B3LYP are higher in
energy than the measured values
by B30 cm1, and reproduce the
observed trend which shows that
the C–C stretching frequency
decreases with increasing
ligation. Agreement between
C–C frequency and calculation
was not as good using the VWN
functional, but still within 70
cm1. The results provide an
evaluation of changes in the
acceptor properties of the metal
center as ligands are added, and
of the utility of DFT for
modeling f-block coordination
complexes.
This journal is
c
the Owner Societies 2007
Introduction
Cerium occupies a unique position on the periodic table in
that it is the first of the f-block elements, where the lanthanide
contraction has not yet precluded participation of the 4f
orbital in chemical reactivity. This phenomenon is manifested
in a readily accessible IV oxidation state, in addition to the
III state common to the lanthanide elements. Ceria (CeO2)
containing materials have wide catalytic applications owing
to unique ‘strong metal–support interaction’ between the
surface oxygen vacancies and the metal oxide species, and
the effect that this interaction has on redox behavior.1
In catalytic reactions, this enables ceria surfaces to function
as oxygen storage media, which lies at the heart of a number
of important processes, including the water gas shift reaction
(WGS),2 steam reforming of feedstock organics3,4 and a
variety of oxidation reactions.5–8 On account of this
versatility, ceria is a key component of three-way automotive emission catalysts,4,9,10 among other uses.10–12
Furthermore, because some cerium isotopes (and those of
neighboring lanthanides) are radioactive fission products,
manipulating the complexation chemistry has been an
objective in separations research related to recycling spent
nuclear fuel.13–17
A better understanding of the composition, structure, and
mechanism(s) of action of the complexes involved in catalysis
and separations would improve our ability to control
these processes by designing surfaces with enhanced catalytic
function or complexing agents with improved selectivity or
kinetics. While surface investigations of anchored metal
complexes have evolved to a high level of sophistication,
typical supported metal catalysts are notoriously heterogeneous,18 and thus direct investigation of the reactive sites is
difficult. Surface infrared absorption studies have provided
valuable clues regarding the nature of the attached ligands,9,19–21 but provide no information on the underlying
surface metal species. X-ray photoelectron spectroscopy22 and
laser ablation mass spectrometry23,24 provide better information regarding metal species, but less information on adsorbed
reactant molecules. A complementary approach, that would
provide species-specific information about surface-atomic and molecular species and their interactions, is to employ computational chemistry to model metal complexes thought to be
representative of putative reactive centers as they interact with
reactant molecules, solvents and/or the atmosphere. This approach can provide ligand binding preferences at the metal
center and insight into how these change as the coordination
environment of the metal center varies. However, the accuracy
of computational studies of f element complexes is complicated
by the fact that scalar relativistic effects and spin–orbit coupling need to be taken into account. Assessing the accuracy of
these theoretical approaches would benefit from validating
experimental data.
Trapped-ion mass spectrometry (TrIMS) is an approach that
can measure ligand binding preferences and decomposition
pathways of cerium-ligand complexes, provided they can be
formed and isolated. One approach has been to use laser
desorption, which enables the study of lanthanide oxide species.25 Electrospray ionization (ESI) combined with TrIMS
This journal is
c
the Owner Societies 2007
provides a highly versatile means for introducing metal complexes into the gas phase,26 where it is possible to investigate
both dissociation and condensation reactions; the recent studies of metal complexes conducted by Vachet27 and Van
Stipdonk28 are examples of such work. A variety of lanthanide
complexes have been examined in this fashion in order to better
understand the extent of coordination and ligand binding
preferences.29–41 The main drawback to the TrIMS approach
is that no direct structural information about the species being
studied is obtained—only a mass-to-charge (m/z) ratio is
provided and it must be correlated back to the probable
structure based on intuition and what is known about similar
complexes in solution. Thus most gas-phase studies must again
revert back to theoretical calculations to enable comment on
structure and stability of the metal–ligand complexes.
One method of validating theoretical calculations is to
compare predicted and measured vibrational frequencies. In
general, collection of vibrational spectra using a conventional
transmittance- or absorbance-type measurement is not feasible
for gas-phase ions because the inherent low concentration of
ionic species results in undetectable transmission or absorption. However, infrared multiple photon dissociation
(IRMPD) induced by a high-intensity laser can be used to
collect infrared spectra of gas-phase ions. Unlike most highintensity lasers that operate at a fixed frequency, the free
electron laser for infrared experiments (FELIX) at the FOM
Institute for Plasma Physics in Nieuwegein, The Netherlands,
is capable of delivering a high intensity beam of photons at
wavelengths continuously tunable over the mid-IR range.42 At
the FOM Institute, the laser is interfaced to a Fourier transform ion cyclotron resonance mass spectrometer,43–45 a type of
TrIMS capable of trapping and isolating ions of specific
masses. This combination enables acquisition of IRMPD
spectra of discrete gas-phase ions produced using ESI and
has proven to be an excellent technique for investigating metal
complexes in discrete solvation states.43–48
The overall aim of this research is to analyze Ce-ligand
complexes hypothesized to mimic reactive sites on catalytic
centers, both from the perspective of structure and decomposition pathways. However, the tactical goal of this preliminary study was to produce infrared spectra of model
complex ions (using ESI-TrIMS and IRMPD) that could in
turn be used to evaluate calculations based on density
functional theory (DFT). Assessment of the accuracy of the
DFT calculations will provide guidance regarding appropriate
functionals for systems relevant to catalysis and separations.
The model complex ions investigated had the general composition [CeOH(ACO)n]21 (ACO ¼ acetone), where the core
[CeOH]21 ion is an open-shell species and Ce formally is 4f1.
Acetone was chosen because it is a Lewis base that readily
coordinates metal ions in the gas phase and because the CQO
stretch is a convenient IR chromophore whose frequency is
sensitive to coordination environment. Use of acetone also
allowed for comparisons of the Ce complexes to recent
investigations of gas-phase uranyl complexes of similar composition.49 The [CeOH]21 cation, beyond the importance of
cerium oxides in catalysis, is of further interest because the
species is present in solution as a result of hydrolysis at mid
and higher pH values.
Phys. Chem. Chem. Phys., 2007, 9, 596–606 | 597
Experimental
Generation of cerium complexes by ESI
The complexes studied were generated in a Fourier transform
ion cyclotron resonance mass spectrometer that was designed
and constructed for use with the free electron laser (see below).
The mass spectrometer was equipped with a commercial
Z-spray source (Micromass, Manchester, UK) that produced
ions at atmospheric pressure in a spray plume orthogonal to a
sampling cone. The ESI spray was operated at 3 kV with
respect to ground. Nitrogen gas at a temperature of B32 1C
was used to assist initial desolvation. As has been previously
observed for gas-phase, doubly-charged uranyl-ligand complexes,50 the mass spectra of the cerium complexes were
sensitive to the desolvation temperature used in the ESI
process. In this experiment, the desolvation temperature
was controlled by a heater and thermocouple on the capillary
block. The highest intensity of highly coordinated
[CeOH(ACO)n]21 complexes was generated when the block
heater was turned off and the region was at or near room
temperature. ESI was performed using a one millimolar
stock solution of cerium(III) nitrate, which was generated by
dissolving the salt (Cerac Inc., Milwaukee, WI) in a 9 : 1
solution of water : acetone.
Fourier transform ion cyclotron resonance mass spectrometry
(FT-ICR-MS)43–45
Ions were accumulated in an external hexapole for about
500 ms prior to being injected into the ICR cell. While a large
distribution of singly- and doubly-charged Ce-ligand complexes was formed, this study was limited to [CeOH(ACO)3]21
and [CeOH(ACO)4]21. These ions were isolated for IRMPD
study directly from the ESI mass spectrum using a stored
waveform inverse Fourier transform (SWIFT) pulse,51 which
ejected all species except those having the desired mass.
FT-ICR-MS.53 The IRMPD efficiency was then expressed
as -log(1-(summed fragment ion yield)), corrected for the
binning width of the acquisition channels, and linearly normalized to correct for variations in FELIX power over the
spectral range.
Electrospray ionization-mass spectrometry (ESI-MS)
ESI mass spectra were collected using a Finnigan LCQ-Deca
ion-trap mass spectrometer (ThermoFinnigan Corporation;
San Jose, CA). The spray solution consisted of cerium(III)
nitrate dissolved to a concentration of 1 mM in 10% (v : v)
acetone-d6 (Aldrich, Milwaukee, USA) dissolved in water. The
solutions were injected into the ESI-MS instrument using the
syringe pump at a flow rate of 3–5 mL min1.
The atmospheric pressure ionization stack settings for the
LCQ (lens voltages, quadrupole and octapole voltage offsets,
etc.) were optimized for maximum ion transmission to the ion
trap mass analyzer by using the auto-tune routine within the
LCQ Tune program. Following the instrument tune, the spray
needle voltage was maintained at þ5 kV and the N2 sheath gas
flow at 25 units (arbitrary to the LCQ instrument, corresponding to approximately 0.375 L min1). The heated capillary
(desolvation) temperature was 250 1C. The ion trap analyzer
was operated at B1.5 105 Torr. Helium gas, admitted
directly into the ion trap, was used as the bath/buffer gas to
improve trapping efficiency and as the collision gas for collision-induced dissociation (CID) experiments.
The CID experiments28 were performed by setting the isolation
width of 2 mass units, the activation Q (as labeled by instrument
manufacturer, used to adjust the qz value for the resonant
excitation of the precursor ion during the CID portion of
the experiment) at 0.3, and the activation amplitude at 10–20%
(of 5 V). In all cases, activation times for CID were 30 ms.
Infrared multiple photon dissociation (IRMPD)
Molecular structure and frequency calculations using density
functional theory
In general, infrared spectra of the CeOH complex ions were
collected by monitoring the efficiency of IRMPD as a function
of photon energy. In this experiment, isolated ionic complexes
were irradiated using two FELIX macropulses (60 mJ per
macropulse, 5 ms pulse duration, bandwidth 1% fwhm of
central l). When the laser frequency matches that of a normal
vibrational mode of the gas-phase ion, energy is absorbed and
subsequently distributed throughout the ion by intramolecular
vibrational redistribution (IVR). The IVR process allows the
energy of each photon to be ‘‘relaxed’’ prior to the absorption
of the next, and thus allows promotion of the ion’s internal
energy to the dissociation threshold by multiple photon absorption.52 Prior studies have shown that the infrared spectra
obtained using the IRMPD method presented here are comparable to those obtained using linear absorption techniques.43,47
To produce infrared spectra, the free electron laser was
scanned in 0.01 to 0.04 mm increments between 5.8 and 10 mm,
after which IRMPD product ions and un-dissociated precursor ions were measured using the excite/detect sequence of the
DFT calculations using the NWChem54,55 software were
performed using the Stuttgart small core relativistic effective
core potential (RSC ECP) and associated Stuttgart orbital
basis set for cerium,56–63 and the valence triple zeta plus
polarization (TZVP) DFT optimized basis sets for all other
atoms (O, C, H, and N).64 In all cases, spherical basis sets were
employed. Calculations were performed using B3LYP,65,66
VWN,67,68 and PBE069 functionals. The open-shell system
was studied within the unrestricted DFT framework.
Additional unrestricted open-shell DFT calculations using
the same VWN functional as applied in NWChem were
performed using the Accelrys Inc. DMol3 suite.70,71 A polarized numerical basis set (DNP) was used, and an all electron
relativistic treatment using a local pseudopotential was applied.72 A fine (108) energy convergence criteria was employed to ensure optimal geometries and representative
calculated vibrational frequencies. Since the same VWN functional was used by both NWChem and DMol3, a comparison
of frequency results provided a quantitative evaluation of the
relativistic treatment and basis set properties.
598 | Phys. Chem. Chem. Phys., 2007, 9, 596–606
This journal is
c
the Owner Societies 2007
Fig. 1
Cation ESI mass spectrum of a 1 mM solution of cerium(III) nitrate dissolved in a 10% acetone/water solution.
Results and discussion
Mass spectrometry of Ce(III) solutions
Electrospray ionization mass spectra of cerium(III) nitrate in a
10% acetone/water solution produced an array of singly and
doubly charged complex ions that contained hydroxide, nitrate
and acetate anions, and neutral acetone ligands (Fig. 1). The
origin of the hydroxide ions was most likely hydrolysis of
Ce(III); hydrolysis is common in solution for highly charged
lanthanide cations and may be enhanced during the electrospray ionization process. Doubly-charged complexes were also
formed that contained three- or four-acetone ligands and either
an acetate or nitrate anion. The appearance of the acetate
complexes was surprising, since neither acetic acid nor acetate
salts were present in the spray solution. The acetate likely
arose from traces of ammonium acetate in the spray chamber
and in the hexapole ion accumulation chamber, which had
been used as a buffer in previous ESI studies of peptides. The
higher abundance of the acetate complexes reflects the fact
that acetate is a stronger nucleophile than is nitrate.
Among the singly-charged complexes formed were those
with the general formula [Ce(anion)2(ACO)n¼1,2]1, where the
anion was hydroxide, nitrate, or acetate, or a mix of the three.
The singly-charged complexes, in general, held fewer acetone
molecules, which is consistent with the fact that the þ1 species
would be weaker Lewis acids compared with the doublycharged complexes.
A comprehensive survey and investigation of the types of
complex ions generated by ESI of Ce solutions, and the
intrinsic chemistry of these potential species, was beyond the
scope of the present study and will be the subject of a future
report. The focus of this study was [CeOH]21, which may be
representative of a functional group important to the chemistry of ceria in catalysis. In the ESI mass spectrum, [CeOH]21
was observed complexed with three- and four-acetone ligands.
However, at ambient temperature it was difficult to control ion
abundances for the long periods of time needed to collect
IRMPD spectra. Using a desolvation temperature of 125 1C,
the doubly-charged ions disappeared and were replaced by
[CeOOH]1, [CeO]1, and versions of these ions with a single
acetone attached. When the desolvation temperature was
decreased to 32 1C, stable [CeOH(ACO)n¼3,4]21 ion abunThis journal is
c
the Owner Societies 2007
dances were obtained. Thus, we were able to obtain IRMPD
spectra of these cerium(III) complexes, which contain both
hydroxy and carbonyl groups. There are many additional
opportunities for vibrational measurement on a much broader
envelope of cerium complexes; however, this will be best
accomplished once better control of the ion formation processes is achieved and the generation of stable ion abundances
for each species is gained.
IRMPD of [CeOH(ACO)4]21
The IRMPD spectrum of [CeOH(ACO)4]21 (m/z 194.5) was
collected by isolating the ion using a SWIFT pulse51 and then
scanning the free electron laser from ca. 5.8 to 10 mm (1730 to
1000 cm1) (Fig. 2). This wavelength range included the
regions in which absorption by the acetone ligands was
expected. No significant absorption was observed when the
IRMPD spectrum was surveyed at frequencies below 1000
cm1. The spectrum in the mid-IR region showed absorptions
in three distinct regions, which corresponded to the carbonyl
stretch
at
B1650
cm1,
C–H
bending
modes
1
from B1350–1450 cm , and C–C stretching at B1250
cm1. When FELIX was scanned through these regions,
Fig. 2 Infrared multiple photon dissociation spectrum of [CeOH(ACO)4]21 (dashed line), and unligated gas-phase acetone (solid line).
Phys. Chem. Chem. Phys., 2007, 9, 596–606 | 599
C–H deformations, but the relative intensity of the two peaks
is much more different, strongly biased toward the lower
frequency component in the unligated molecule.
IRMPD of [CeOH(ACO)3]21
Scheme 1 Photofragmentation reaction of [CeOH(ACO)4]21 (where
ACO ¼ acetone).
photofragmentation principally resulted in the loss of one
acetone ligand from the complex (forming the m/z 165.5
photofragment ion) as shown in Scheme 1. The IRMPD
spectrum shown in Fig. 2 was produced by plotting the
negative log of (1-the m/z 165.5 abundance) as a function of
IR frequency.
The acetone ligand is bound to the Ce metal center via the
carbonyl oxygen atom, and a comparison to the spectrum of
unligated acetone73 shows a red shift of the CQO stretching
frequency in the CeOH–acetone complex of about 90 cm1.
The red-shifted CQO stretching frequency is consistent with
transfer of electron density from the carbonyl to the Ce(III)
metal center, thereby weakening the CQO bond. An unresolved ‘shoulder’ is observed on the low-frequency side of the
CQO stretch, centered at perhaps 1600 cm1. This shoulder
may result from acetone existing in two different chemical
environments in the complex. The DFT calculations (vide
infra) predict a stable isomer having a distorted square pyramidal geometry; however, another reasonable possibility
would be trigonal bipyramidal. In the latter, a complex having
one axial and three equatorial acetone ligands could be envisioned, which may well have different binding strengths resulting in differing CQO stretching frequencies.
In unligated acetone the C–C stretch is observed at 1214
cm1. For [CeOH(ACO)4]21, the same absorption appears at
1247 cm1—a blue shift of 33 cm1 which suggests a modest
strengthening of the C–C bond in the coordinated acetone
ligands. Reduction of electron density in the ligand by donation through the carbonyl oxygen to the Ce31 center would be
expected to increase the positive charge on the carbonyl
carbon, which in turn would be expected to strengthen the
C–C bonds. As will be shown, the difference in the C–C
stretching frequency between the three- and four-acetone
complexes is small, which shows that the C–C bond, while
clearly perturbed by the metal center, is far enough away from
the metal center that the change in ligation does not greatly
affect its frequency.
Also observed in Fig. 2 are absorptions at B1340 and
B1460 cm1, frequencies intermediate between the CQO
and C–C stretches, that are assigned to C–H bending vibrations. The absorption peaks in both the four- and three- (vide
infra) acetone complexes have a similar appearance, with
double maxima at about 1380 and 1420 cm1. The appearance
of the C–H bending peaks in the four-acetone complex is very
similar to that observed in the three-acetone complexes (vide
infra), suggesting that the methyl groups are in similar environments in both complexes. However this region of the
spectrum is in contrast to that of unligated acetone, which
also displays two maxima in the region corresponding to the
600 | Phys. Chem. Chem. Phys., 2007, 9, 596–606
The IRMPD spectrum of the three-acetone complex was
acquired using a SWIFT pulse to isolate [CeOH(ACO)3]21
(m/z 165.5) prior to scanning the laser from B1000 to 1680
cm1. When irradiated, this ion underwent fragmentation by
two reaction channels that produced a total of three product
ions (Scheme 2). [CeOH(ACO)2]21 at m/z 136.5 was formed by
simple loss of ACO (-ACO), and this was accompanied by
loss of ACO with a concurrent charge separation reaction
(–ACO þ charge sep). The products of the latter reaction were
[ACO þ H]1 at m/z 59 and [CeOH(ACO)-H]1 at m/z 214.
Formation of these products requires transfer of a proton to
the departing acetone ligand from one of two sites on the
molecule: either the hydroxy group of the CeOH cation, or a
methyl group of the remaining acetone ligand. In collisioninduced dissociation (CID) using a quadrupole ion trap,
proton transfer from the hydroxy moiety was the preferred
decomposition pathway, as was shown by ESI of [CeOH]21
with acetone-d6, which formed an adduct at m/z 142.5 that
corresponded to [CeOH(ACO-d6)2]21. When subjected to
CID, fragment ions at m/z 220 and 65 were formed, which
correspond to [CeO(ACO-d6)]1 and [(ACO-d6)þH]1, respectively. Hence, it is likely that the same reaction is occurring for
the IR-irradiated ions.
The IRMPD spectrum of [CeOH(ACO)3]21 (Fig. 3) generated by summing the abundances of all three photofragment
ions contained a CQO stretch at 1616 cm1, however a close
inspection of the individual IRMPD spectra revealed a significant difference between the (–ACO) and (–ACO þ charge
sep) photofragmentation channels. The observation of the
multi-step fragmentation pathway underscores the nature of
the FELIX-FT-ICR-MS experiment, viz. the photofragmentation results from non-coherent absorption of multiple photons
(tens to hundreds). The (–ACO) photofragmentation reached
a maximum at 1635 cm1, which was 36 cm1 higher than the
maximum for (–ACO þ charge sep) at 1599 cm1. The
phenomenon of an ion with competitive fragmentation
Scheme 2 Photofragmentation reactions of [CeOH(ACO)3]21. The
simple cleavage to eliminate a neutral ACO (–ACO) occurs in parallel
with loss of ACO concurrent with a charge separation (–ACO þ
charge sep), producing protonated acetone and a singly-charged
cerium complex.
This journal is
c
the Owner Societies 2007
Fig. 3 Infrared multiple photon dissociation spectrum of
[CeOH(ACO)3]21: (solid line) sum of fragmentation channels; (dotted
line) loss of acetone (–ACO) at m/z 136.5; (dashed line) sum of the
(–ACO þ charge sep) channels (m/z 59 and 214).
channels that have IRMPD spectra that maximize at different
frequencies has been observed previously:74 in the para-aminobenzoyl cation, the frequency maximum of the CQO
stretch in the spectrum of a more energetically demanding
fragmentation channel was lower by 18 cm1 compared to a
less energetic fragmentation channel.75 These results were
interpreted in terms of cross-anharmonicities of the CQO
stretch with all other modes, which shifted the absorption
frequency of the more energetically demanding fragmentation
lower as the internal energy of the ion increased. For [CeOH(ACO)3]21 a similar anharmonicity argument no doubt
accounts for part of the shift, as the relative fragmentation
threshold of the loss of (–ACO þ charge sep) is greater than
that for (–ACO). However the larger magnitude (36 cm1) of
the shift suggests that overlapping absorption by the photofragment product [CeOH(ACO)2]21 may also be occurring.
In this alternative explanation of the different maxima, the
(–ACO) photofragment channel forms [CeOH(ACO)2]1,
which subsequently absorbs and photofragments to produce
the charge separation products. For this to occur, both species
would have to have common absorption frequencies, and this
is the case: at B1620–1600 cm1, a significant portion of the
CQO stretch in the spectrum of (–ACO) channel is overlapped
with that of the (–ACO þ charge sep) channel. The possibility
of photofragmentation of [CeOH(ACO)2]21 and anharmonicity contributing to the lower frequency of the (–ACO þ
charge sep) channel is supported by the width of the peak,
which is B 35 cm1 greater than that for other CQO stretches
from fragmentations involving only ACO losses. The conclusion implies that the (–ACO þ charge sep) photofragmentation channel approximates the IMRPD behavior of
[CeOH(ACO)2]21.
The values for the CQO stretching frequencies in the
[CeOH(ACO)3]21 complex are shifted to the red compared
to the position of the same stretch in the spectrum generated
from [CeOH(ACO)4]21. This is true for the spectrum from
(–ACO), (–ACO þ charge sep), and from the summed ion
This journal is
c
the Owner Societies 2007
Fig. 4 Infrared multiple photon dissociation spectrum of the (–ACO)
channel from [CeOH(acetone)3]21 (solid line), compared with that of
[UO2(ACO)4]21 (dotted line). The uranyl spectrum is scaled by a
factor of five to account for lower photofragment yield.
intensities. The red shift indicates a general weakening of the
carbonyl bond in the three-acetone and two-acetone complexes where the acetone ligands are expected to be more
strongly attached to the metal center: the bond energy of
noncovalently attached ligands most often increases as the
number of ligands decreases.76 The decreasing carbonyl frequency in the order (free acetone) 4 [CeOH(ACO)4]21 4
[CeOH(ACO)3]21 4 [CeOH(ACO)2]21 is consistent with the
order anticipated for acetone unbound 4 loosely bound 4
tightly bound and with the carbonyl frequencies of isolated
uranyl–acetone complexes that were recently measured by
IRMPD.49
A significant difference between the (–ACO) and the (–ACO
þ charge sep) channels was also observed in the C–C stretching region. The spectrum generated by the acetone elimination
reaction from [CeOH(ACO)3]21 showed a notable absorption
at 1254 cm1. Compared to the same absorption in the
spectrum of the [CeOH(ACO)4]21, the frequency has been
slightly shifted to higher energy, and the intensity is significantly diminished. The (–ACO) channel accounted for nearly
all of the C–C stretch, with the (–ACO þ charge sep) channel
being nearly unreactive. A trend worth noting is that the
relative intensity of the C–C stretch decreases with decreasing
ligation: the C–C stretch is nearly as intense as the carbonyl
stretch for the four-acetone complex (and comparable to
intensity of the C–C stretch in unligated acetone), but decreases to about one-fifth that value for the (–ACO) channel
from the three-acetone complex, and is barely above background in the (–ACO þ charge sep) channel. Since dissociation rates are exponential functions of internal energy, peak
intensities do not scale linearly, and weak bands may become
unobservable in higher energy exit channels.75
Recently, IRMPD spectra of acetone ligated to [UO2]21
were obtained using FELIX.49 The data for the [UO2]21
complexes provided a good opportunity to compare the effect
of binding to different metal centers on the position of the
CQO stretching frequency of acetone (Fig. 4). A comparison
Phys. Chem. Chem. Phys., 2007, 9, 596–606 | 601
of the spectrum generated by ACO elimination from
[CeOH(ACO)3]21 with the analogous spectrum starting from
[UO2(ACO)4]21 showed remarkable agreement in terms of
frequencies and relative absorption intensities. The comparison suggests that [UO2]21 perturbs four acetone ligands to the
same extent that [CeOH]21 perturbs three, indicating that
when the extent of ligation is comparable, the hydroxyl cerium
dication is a less aggressive electrophile. It stands to reason
that the 5f orbitals of uranium in the uranyl dication are more
available for electron donation compared to the more contracted 4f orbitals in the hydroxyl cerium dication. The
absolute photofragment yield was about five times greater
for [CeOH(ACO)3]21 than for [UO2(ACO)4]21, indicating that
fewer photons were needed to fragment the Ce complex. The
difference in yield suggests a kinetic shift for the uranyl
complex, which is better able to accommodate internal energy
on account of the fact that it has more oscillators.
DFT calculations
The complexes that were studied using IRMPD spectroscopy
were also the subject of DFT calculations to provide insight
into the origins of the frequencies measured and to provide a
calibration point for future theoretical studies. Calculated
frequencies are dependent on the functional used, and three
different functionals were compared to each other and experimental data. The local density approximation using the VWN
functional was used because good accuracy for prediction of
IR frequencies for ACO-ligated [UO2]21 had been achieved in
a previous study.49 B3LYP and PBE0 functionals have also
been used with good results for complexes containing heavy
elements, and frequency calculations were also performed
using these for comparison. Calculations using these three
functionals were performed using the Stuttgart RSC ECP for
Ce to account for relativisitic effects, and TZVP basis for all
other atoms (NWChem). In addition, a basis set comparison
was performed using DNP numerical basis sets and an allelectron relativistic treatment for all atoms with VWN
(DMol3).
Benchmark calculations for unligated acetone using both
the B3LYP and VWN functionals reproduced the major
features of the spectrum of gas-phase acetone found in the
NIST library (Fig. 5a), with the frequencies of the absorptions
shifted to higher energy by 20 to B 45 cm1 for the C–C and
CQO stretching frequencies, respectively. Such a shift is
Fig. 5 Comparison of the measured IRMPD spectra (solid) with
DFT-calculated spectra using B3LYP (dash) and VWN (dot) functionals. (a) Calculation and IR spectrum for unligated ACO. (b)
Calculation and IRMPD spectrum for [CeOH(ACO)4]21. (c) Calculation, and IRMPD spectrum of the (–ACO) channel from
[CeOH(ACO)3]21. (d) Calculation of the spectrum for
[CeOH(ACO)2]21, and IRMPD spectrum of (–ACO þ charge separation) channel compared. The intensities of the CQO frequencies of the
calculated spectra were scaled to the IRMPD CQO stretch to enable
comparison.
Table 1 DFT-calculated CQO asymmetric stretching frequencies (in cm1) calculated using DFT with different functionals, compared to
measurements
NWChem
Complex
[CeOH(ACO)]21
[CeOH(ACO)2]21
[CeOH(ACO)3]21
[CeOH(ACO)4]21
Unligated ACO
Measurements
1599a
1616b
1635c
1650
1738e
B3LYP
PBE0
VWN
DMol3
VWN
1554
1595
1632
1629
1668
1570
1600
1568
1612
1664
1782
1702
1822
1630/1615d
1767
1640
1780
a
IRMPD measurement derived from the (–ACO þ charge sep) channel from [CeOH(ACO)3]21. b Measurement derived from the summed
fragment channels from [CeOH(ACO)3]21. c Measurement derived from the (–ACO) channel from [CeOH(ACO)3]21. d Calculations showed two
closely spaced absorptions, with the higher frequency having twice the intensity of the lower. e Gas-phase un-ligated acetone.
602 | Phys. Chem. Chem. Phys., 2007, 9, 596–606
This journal is
c
the Owner Societies 2007
Fig. 6 DFT-calculated structures (B3LYP) for [CeOH(ACO)3]21
(left) and [CeOH(ACO)4]21 (right).
typical of DFT calculations, which are normally corrected by
scaling the calculated frequencies by values ranging from 0.95
to 0.98.77–80 Better agreement between experiment and theory
was achieved for the [CeOH(acetone)4]21 complex (Fig. 5b):
the value calculated for the CQO stretch using B3LYP was
only 14 cm1 higher than the measurement, PBE0 B50 cm1
higher and the VWN B10–20 cm1 lower (Table 1). The shape
of the IRMPD CQO stretch suggests the presence of a lowfrequency shoulder, that might be consistent with acetone in
two chemically distinct coordination sites around the Ce metal
center. An attractive hypothesis that would account for this
would be the adoption of a distorted trigonal bipyramidal
structure, in which the hydroxy and one acetone ligand would
occupy axial positions, while the three remaining acetone
molecules would occupy equatorial positions. In fact two
closely spaced CQO frequencies were calculated using
VWN. However the lowest energy structure calculated (using
either VWN or B3LYP) was one resembling a distorted square
pyramid, in which the Ce and O atoms have C4v symmetry
(Fig. 6), and all four acetone ligands are equivalent. The
B3LYP calculations did reveal a distorted trigonal bipyramidal structure only a few kcal mol1 higher in energy. However,
the calculated frequencies of the carbonyl absorption indicated
that this structure, which has three tightly bonded acetones
and one more loosely bound acetone, should have a higher
energy shoulder, not lower.
Considering the [CeOH(ACO)3]21 complex, CQO values
calculated using all four approaches straddle the measurements for the combined photofragmentation channels (1616
cm1) and the (–ACO) channel (1635 cm1). Assuming that
the [CeOH(ACO)3]21 complex is best represented by the
(–ACO) channel, best agreement was produced using B3LYP
(within 3 cm1), with the PBE0 values modestly higher, and
VWN values modestly lower. The relationships of the calculated frequencies to the (–ACO) channel maximum were the
same as those noted for the [CeOH(ACO)4]21 complex. The
lowest-energy structure for the [CeOH(ACO)3]21 complex
possessed a tetrahedral structure as defined by the four O
atoms (1 OH and 3 ACO CQO, Fig. 6).
The frequency maximum for the CQO stretch for the
(–ACO þ charge sep) channel (from [CeOH(ACO)3]21) at
1599 cm1 was lower than all calculations for the threeacetone complex, which adds further support to the idea
that this measurement contains contributions from
[CeOH(ACO)2]21. The frequency values calculated using
B3LYP were in close agreement, while those from PBE0 were
somewhat higher, and those from VWN were somewhat lower.
As in the analysis of the (–ACO) channel, this self-consistency
This journal is
c
the Owner Societies 2007
Fig. 7 CQO stretching frequencies in unligated and CeOH21-complexed ACO, comparing IRMPD results (dot) with values from DFTcalculations. The black filled dots represent IRMPD data (using the
individual photofragment channels (–ACO) and (ACO þ charge sep)),
and the star represents the absorption measurement for gas-phase
unligated ACO.
in the relative positions of the calculated values and the
measurements supports the conclusions that (a) the (–ACO)
channel best represents the three-acetone complex, and (b) the
(–ACO þ charge sep) channel to some degree represents the
two-acetone complex.
The calculated frequencies for the CQO stretch in different
ligated complexes show a uniform increase as the number of
ligands increases (Table 1 and Fig. 7) which is the same trend
as that observed for acetone ligated to uranyl dications,49 and
is consistent with the general weakening of metal–oxygen
bonding with increasing ligation.76 All DFT functionals predict a shift to higher frequency on going from the (ACO)3
complex to the (ACO)4 complex, with values ranging from 28
to 34 cm1. These frequency shifts are all in good agreement
with the 34 cm1 shift seen in the data comparison using the
peak centroid from the summed data channels from the
[CeOH(ACO)3]21. In contrast, comparing the experimental
data based on the (–ACO) channel from [CeOH(ACO)3]21
predicts a shift of 15 cm1, which is less than the shift
predicted by the computational results. Thus while our prior
discussion concluded that comparisons of frequencies from
individual photofragment channels provided a better basis for
comparison, this line of reasoning suggests that the frequency
maximum generated from the entire manifold of photofragments from [CeOH(ACO)3]21 is more appropriate. Ultimate
resolution of the question calls for improved modeling and
measurement.
Analysis of binding energies calculated using B3LYP
showed a steady decrease in binding energy per ligand, and
in differential binding energy for each additional ligand (Table
2), consistent with trends for most series of non-covalent metal
cation-ligand complexes.76 Further, there was a reasonable
inverse correlation between binding energy per ligand and the
CQO frequency (Fig. 8): as the calculated frequency increases,
the binding energy per ligand steadily decreases. The
Phys. Chem. Chem. Phys., 2007, 9, 596–606 | 603
Table 2 Ligand binding energies calculated using B3LYP, kcal mol1
Complex
Total
ligand
binding
energy
Binding
energy per
ligand
Differential
binding energy/
kcal mol1
[CeOH(ACO)]21
[CeOH(ACO)2]21
[CeOH(ACO)3]21
[CeOH(ACO)4]21
83
144
196
229
83
72
65
56
83
61
52
33
generality of the correlation was evaluated by comparison
with a data point for [UO2(ACO)2]21: in this complex, the
ACO binding energy per ligand was calculated by Marsden
and coworkers using B3LYP81 at 94.6 kcal mol1. Combination of this value with the recently measured CQO frequency49 enables the [UO2(ACO)2]21 complex to be located
on Fig. 8, and shows that binding energy per ligand forms a
consistent trend with the present results for [CeOH(ACO)n]21,
suggesting that the CQO stretching frequency could be calibrated in terms of binding energy per ligand.
Frequencies generated using DFT calculations for the C–C
and C–H vibrations were not as accurate as were those for the
CQO vibrations. Origins of differences between calculations
and measurements are not explicitly understood, however we
note that computational approaches used in this study were
selected to effectively model metal–ligand interactions. In fact
the best accuracy was achieved for functional groups most
strongly perturbed by the metal center. Vibrations derived
from bonds that are more remote from the metal have maxima
shifted from measurements by anywhere from 20 to 60 cm1,
and frequently the profiles are significantly different. Understanding the origins of these differences and development of
approaches able to accurately model functional groups remote
Fig. 8 Binding energy (calculated using B3LYP, kcal mol1) plotted
versus CQO stretching frequencies. Calculated frequencies and energies for [CeOH(ACO)n¼14]1 are represented by black diamonds;
measured frequencies (using the individual photofragment channels
(–ACO) and (ACO þ charge sep)) and calculated energies by triangle
data points. Measured frequency49 and calculated energy81 for
[UO2(ACO)2]21 are represented by the black square data point.
604 | Phys. Chem. Chem. Phys., 2007, 9, 596–606
Fig. 9 C–C stretching frequencies in unligated and CeOH21-complexed ACO, comparing IRMPD results (using the individual photofragment channels (–ACO) and (ACO þ charge sep), represented by
black dots) with values from DFT-calculations (triangles). The black
dots represent IRMPD data, and the star represents the frequency
measured by absorption of the gas-phase neutral ACO.
from the metal center will be important in application of
computational chemistry to catalysis, where subtle alterations
in bond strengths will influence C–H and C–C insertion
reactions. Thus, in this study the ability of the computational
approaches to reproduce trends in C–C and C–H frequencies
with changing ligation was evaluated.
The DFT calculations using the B3LYP functional produced the C–C stretching frequencies for the complexes and
for unligated ACO that were B30 cm1 higher than the
IRMPD measurements (Fig. 9), and followed a trend opposite
that of CQO: the C–C frequency undergoes a modest decrease
with increasing ligation, approaching the value for unligated
acetone. This is consistent with the idea that the C–C bond is
slightly weakened in complexes where ACO is more loosely
bound. This is intuitively satisfying because upon complexation transfer of electron density from the carbonyl to the metal
would be expected to increase the partial positive character of
the carbonyl carbon, which in turn may increase the frequency
of the C–C stretch. However as more acetone ligands are
added, each ligand is more weakly bound, less electron density
is transferred from the carbonyl, and the C–C bond is less
perturbed compared to the unligated molecule. Interestingly,
the C–C stretch intensity is much reduced in the spectrum
produced by the charge separation channels (see Fig. 5d). The
frequency is, however, practically unaltered, which suggests a
smaller anharmonicity for this mode as compared to the CQO
stretch. The agreement between the IRMPD data and the
calculated C–C frequencies generated using the VWN functional was not as good, and the trend was opposite to that of
the data, i.e.. the frequency was modestly increased with
increasing ligation.
The C–H bending frequencies observed in the IRMPD
spectra were located between B1340 and B1460 cm1; the
broadened profiles were a result of contributions from
This journal is
c
the Owner Societies 2007
multiple, closely spaced absorption modes that were not
resolved. The DFT calculations indicate that the lower frequency bands correspond to symmetric CH3 deformations,
and that the higher frequencies correspond to symmetric and
antisymmetric C–H bending modes. The calculations suggest
that these might be resolved, however only modest hints of
peak resolution could be observed. In general the frequencies
of the C–H bending modes calculated using B3LYP were
B20–40 cm1 higher than the measurements, while VWN
produced frequencies that were lower by B20–40 cm1.
Conclusions
In this study, we obtained the mid-IR spectra of gas-phase,
acetone-ligated [CeOH]21 which showed prominent absorptions attributable to the ACO ligands in [CeOH(ACO)4]21 and
[CeOH(ACO)3]21. Compared to the value for unligated ACO,
a red shift of the CQO stretching frequency was observed for
[CeOH(ACO)4]21, and a further red shift was observed in the
(ACO)3 complex. The trend in both the IRMPD data and in
the DFT calculations was consistent with increasing donation
of electron density (per ligand) from the ligands to the metal
center as the number of ligands decreases. Red shifts of the
CQO stretching frequencies were accompanied by complementary blue shifts of the C–C stretching frequency.
The [CeOH(ACO)3]21 also fragmented by loss of ACO
concurrent with charge separation, producing [CeO(ACO)]1
and [ACO þ H]1. The IRMPD spectra generated by the
charge separation photofragment channels showed a pronounced red shift in the carbonyl stretching frequency compared to the spectrum generated by loss of ACO only. The
magnitude of the shift, and the shape of the peak indicated
that the phenomenon was due two a combination of factors:
(a) anharmonicity of the CQO stretching mode, producing a
red shift most apparent in the higher-energy dissociation
channel,75 and (b) absorption and photofragmentation of
[CeOH(ACO)2]21, which is produced by initial loss of ACO
from [CeOH(ACO)3]21.
These vibrational spectra, as well as future ones to be
obtained using IRMPD of trapped ions, afford an excellent
opportunity to test the accuracy of various theoretical techniques used to model complexes of the f-block elements. Overall
we feel that the agreement between the experimental measurements for CQO and C–C frequencies and the unscaled values
generated using DFT was good. The comparison of the
different DFT-approaches shows that the frequencies are more
sensitive to the choice of functional than to the basis set
employed, and that close agreement between experiment and
theory for one frequency does not mean that the functional
and basis set are necessarily optimal for a second frequency.
Nevertheless, DFT is clearly capable of providing fairly
accurate vibrational frequencies of small ionic lanthanide
complexes, which further indicates a high degree of reliability
in the modeled structures. Thus, DFT-calculations are expected to be useful for generation and investigation of a much
broader suite of complexes designed to mimic catalytic reactive sites and intermediates important in metal separations
processes.
This journal is
c
the Owner Societies 2007
Acknowledgements
Work by G. S. Groenewold and A. K. Gianotto was supported by the US Department of Energy, Assistant Secretary
for Environmental Management, and the INL Laboratory
Directed Research & Development Program under DOE
Idaho Operations Office Contract DE-AC07-05ID14517. The
INL authors thank the University of Florida (Professor John
Eyler) for travel support. M. J. Van Stipdonk was supported
through a grant from the U.S. National Science Foundation
(NSF grant CAREER-0239800). The FOM authors were
supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO). The skillful assistance by the
FELIX staff, in particular Dr B. Redlich, is gratefully acknowledged. Construction and shipping of the FT-ICR-MS
instrument was made possible through funding from the
National High Field FT-ICR Facility (grant CHE-9909502)
at the National High Magnetic Field Laboratory, Tallahassee,
FL. W. A. de Jong’s research was performed, in part, using the
Molecular Science Computing Facility in the William R. Wiley
Environmental Molecular Sciences Laboratory, a national
scientific user facility sponsored by the U.S. Department of
Energy’s Office of Biological and Environmental Research
located at the Pacific Northwest National Laboratory, which
is operated for the Department of Energy by Battelle. References herein to any specific commercial product, process, or
service by trade name, trademark, manufacturer, or otherwise,
does not necessarily constitute or imply its endorsement,
recommendation, or favoring by the US Government, any
agency thereof, or any company affiliated with the Idaho
National Laboratory.
References
1 W. J. Shan, W. J. Shen and C. Li, Chem. Mater., 2003, 15, 4761.
2 X. M. Qi and M. Flytzani-Stephanopoulos, Ind. Eng. Chem. Res.,
2004, 43, 3055.
3 J. Papavasiliou, G. Avgouropoulos and T. Ioannides, Catal.
Commun., 2004, 5, 231.
4 S. Sharma, S. Hilaire, J. M. Vohs, R. J. Gorte and H. W. Jen, J.
Catal., 2000, 190, 199.
5 T. L. Zhu, L. Kundakovic, A. Dreher and M. Flytzani-Stephanopoulos, Catal. Today, 1999, 50, 381.
6 A. Martinez-Arias, M. Fernandez-Garcia, J. Soria and J. C.
Conesa, J. Catal., 1999, 182, 367.
7 S. Hocevar, J. Batista and J. Levec, J. Catal., 1999, 184, 39.
8 S. Hocevar, U. O. Krasovec, B. Orel, A. S. Arico and H. Kim,
Appl. Catal., 2000, 28, 113.
9 A. Martinez-Arias, M. Fernandez-Garcia, O. Galvez, J. M. Coronado, J. A. Anderson, J. C. Conesa, J. Soria and G. Munuera, J.
Catal., 2000, 195, 207.
10 B. Skarman, D. Grandjean, R. E. Benfield, A. Hinz, A. Andersson
and L. R. Wallenberg, J. Catal., 2002, 211, 119.
11 U. Lampe, J. Gerblinger and H. Meixner, Sens. Actuators, B, 1992,
7, 787.
12 R. J. Gorte, J. M. Vohs and S. McIntosh, Solid State Ionics, 2004,
175, 1.
13 J. N. Mathur, M. S. Murali and K. L. Nash, Solvent Extr. Ion
Exch., 2001, 19, 357.
14 A. P. Paiva and P. Malik, J. Radioanal. Nucl. Chem., 2004, 261,
485.
15 H. Eccles, Solvent Extr. Ion Exch., 2000, 18, 633.
16 W. W. Schultz and E. P. Horwitz, Sep. Sci. Technol., 1988, 23,
1191.
17 E. E. Sood and S. K. Patil, J. Radioanal. Nucl. Chem., 1996, 203,
547.
Phys. Chem. Chem. Phys., 2007, 9, 596–606 | 605
18 J. C. Fierro-Gonzalez, S. Kuba, Y. Hao and B. C. Gates, J. Phys.
Chem. B, 2006, 110, 13326.
19 S. Hilaire, X. Wang, T. Luo, R. J. Gorte and J. Wagner, Appl.
Catal., 2004, 258, 271.
20 T. Shido and Y. Iwasawa, J. Catal., 1992, 136, 493.
21 T. Shido and Y. Iwasawa, J. Catal., 1993, 141, 71.
22 A. M. Salvi, F. Decker, F. Varsano and G. Speranza, Surf.
Interface Anal., 2001, 31, 255.
23 J. K. Gibson, J. Vac. Sci. Technol., A, 1995, 13, 1945.
24 J. K. Gibson and R. G. Haire, J. Alloys Compd., 2001, 322, 143.
25 G. P. Jackson, J. K. Gibson and D. C. Duckworth, Int. J. Mass
Spectrom., 2002, 220, 419.
26 D. Schroder, J. Roithova and H. Schwarz, Int. J. Mass Spectrom.,
2006, 254, 197.
27 R. W. Vachet and J. H. Callahan, J. Mass Spectrom., 2002, 35, 311.
28 M. Van Stipdonk, V. Anbalagan, W. Chien, G. Gresham, G. Groenewold and D. Hanna, J. Am. Soc. Mass Spectrom., 2003, 14, 1205.
29 M. C. Crowe, R. N. Kapoor, F. Cervantes-Lee, L. Parkanyi, L.
Schulte, K. H. Pannell and J. S. Brodbelt, Inorg. Chem., 2005, 44, 6415.
30 S. Colette, B. Amekraz, C. Madic, L. Berthon, G. Cote and C.
Moulin, Inorg. Chem., 2003, 42, 2215.
31 S. Colette, B. Amekraz, C. Madic, L. Berthon, G. Cote and C.
Moulin, Inorg. Chem., 2002, 41, 7031.
32 J. J. Zhang, W. Zhang, Q. H. Luo and Y. H. Mei, Polyhedron,
1999, 18, 3637.
33 I. I. Stewart and G. Horlick, Anal. Chem., 1994, 66, 3893.
34 A. W. G. Platt, J. Fawcett, R. S. Hughes and D. R. Russell, Inorg.
Chim. Acta, 1999, 295, 146.
35 J. Marcalo, A. Pires de Matos and W. Evans, Organometallics,
1997, 16, 3845.
36 A. M. J. Lees, J. M. Charnock, R. W. Kresinski and A. W. G.
Platt, Inorg. Chim. Acta, 2001, 312, 170.
37 R. L. C. Lau, J. Z. Jiang, D. K. P. Ng and T. W. D. Chan, J. Am.
Soc. Mass Spectrom., 1997, 8, 161.
38 J. M. Curtis, P. J. Derrick, A. Schnell, E. Constantin, R. T.
Gallagher and J. R. Chapman, Inorg. Chim. Acta, 1992, 201, 197.
39 R. Colton and W. Klaui, Inorg. Chim. Acta, 1993, 211, 235.
40 D. Chapon, C. Husson, P. Delangle, C. Lebrun and P. J. A.
Vottero, J. Alloys Compd., 2001, 323, 128.
41 P. D. Beer, G. D. Brindley, O. D. Fox, A. Grieve, M. I. Ogden, F.
Szemes and M. G. D. Drew, J. Chem. Soc., Dalton Trans., 2002, 3101.
42 D. Oepts, A. F. G. van der Meer and P. W. van Amersfoort,
Infrared Phys. Technol., 1995, 36, 297.
43 J. Oomens, D. T. Moore, G. von Helden, G. Meijer and R. C.
Dunbar, J. Am. Chem. Soc., 2004, 126, 724.
44 D. T. Moore, J. Oomens, L. van der Meer, G. von Helden, G.
Meijer, J. Valle, A. G. Marshall and J. R. Eyler, ChemPhysChem,
2004, 5, 740.
45 D. T. Moore, J. Oomens, J. R. Eyler, G. Meijer, G. von Helden
and D. P. Ridge, J. Am. Chem. Soc., 2004, 126, 14726.
46 D. van Heijnsbergen, K. Demyk, M. A. Duncan, G. Meijer and G.
von Helden, Phys. Chem. Chem. Phys., 2003, 5, 2515.
47 J. Oomens, A. G. G. M. Tielens, B. G. Sartakov, G. Von Helden
and G. Meijer, Astrophys. J., 2003, 591, 968.
48 R. C. Dunbar, D. T. Moore and J. Oomens, J. Phys. Chem. A,
2006, 110, 8316.
49 G. S. Groenewold, A. K. Gianotto, K. C. Cossel, M. J. Van
Stipdonk, D. T. Moore, N. Polfer, J. Oomens, W. A. de Jong and
L. Visscher, J. Am. Chem. Soc., 2006, 107, 4802.
50 M. J. Van Stipdonk, W. Chien, V. Angalaban, K. Bulleigh, D.
Hanna and G. S. Groenewold, J. Phys. Chem. A, 2004, 108, 10448.
51 A. G. Marshall, T.-C. L. Wang and T. L. Ricca, J. Am. Chem.
Soc., 1985, 107, 7893.
52 V. N. Bagratashvili, V. S. Letokov, A. A. Makarov and E. A.
Ryabov, Multiple Photon Infrared Laser Photophysics and Photochemistry, Harwood, Chur, Switzerland, 1985.
606 | Phys. Chem. Chem. Phys., 2007, 9, 596–606
53 A. G. Marshall, C. L. Hendrickson and G. S. Jackson, Mass
Spectrom. Rev., 1998, 17, 1.
54 E. Aprà, T. L. Windus, T. P. Straatsma, E. J. Bylaska, W. de Jong,
S. Hirata, M. Valiev, M. Hackler, L. Pollack, K. Kowalski, R.
Harrison, M. Dupuis, D. M. A. Smith, J. Nieplocha, T. V. M.
Krishnan, A. A. Auer, E. Brown, G. Cisneros, G. Fann, H.
Fruchtl, J. Garza, K. Hirao, R. Kendall, J. Nichols, K. Tsemekhman, K. Wolinski, J. Anchell, D. Bernholdt, P. Borowski, T.
Clark, D. Clerc, H. Dachsel, M. Deegan, K. Dyall, D. Elwood,
E. Glendening, M. Gutowski, A. Hess, J. Jaffe, B. Johnson, J. Ju,
R. Kobayashi, R. Kutteh, Z. Lin, R. Littlefield, X. Long, B. Meng,
T. Nakajima, S. Niu, M. Rosing, G. Sandrone, M. Stave, H.
Taylor, G. Thomas, J. van Lenthe, A. Wong and Z. Zhang,
NWChem, A Computational Chemistry Package for Parallel Computers, Pacific Northwest National Laboratory, Richland,
Washington, 99352-0999 USA, 2005.
55 R. A. Kendall, E. Apra, D. E. Bernholdt, E. J. Bylaska, M.
Dupuis, G. I. Fann, R. J. Harrison, J. Ju, J. A. Nichols, J.
Nieplocha, T. P. Straatsma, T. L. Windus and A. T. Wong,
Comput. Phys. Commun., 2000, 128, 260.
56 A. Bergner, M. Dolg, W. Kuchle, H. Stoll and H. Preuss, Mol.
Phys., 1993, 80, 1431.
57 M. Dolg, H. Stoll, H. Preuss and R. M. Pitzer, J. Phys. Chem.,
1993, 97, 5852.
58 P. Fuentealba, H. Preuss, H. Stoll and L. Vonszentpaly, Chem.
Phys. Lett., 1982, 89, 418.
59 P. Fuentealba, L. Vonszentpaly, H. Preuss and H. Stoll, J. Phys. B,
1985, 18, 1287.
60 G. Igelmann, H. Stoll and H. Preuss, Mol. Phys., 1988, 65, 1321.
61 M. Kaupp, P. V. Schleyer, H. Stoll and H. Preuss, J. Chem. Phys.,
1991, 94, 1360.
62 W. Kuchle, M. Dolg, H. Stoll and H. Preuss, Mol. Phys., 1991, 74,
1245.
63 W. Kuchle, M. Dolg, H. Stoll and H. Preuss, J. Chem. Phys., 1994,
100, 7535.
64 N. Godbout, D. R. Salahub, J. Andzelm and E. Wimmer, Can. J.
Chem., 1992, 70, 560.
65 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
66 C. Lee, W. Yang and R. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785.
67 S. J. Vosko, W. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200.
68 J. C. Slater, Phys. Rev. Lett., 1951, 81, 385.
69 C. Adamo and V. Barone, J. Phys. Chem., 1998, 110, 6158.
70 B. Delley, J. Chem. Phys., 1990, 92, 508.
71 B. Delley, J. Chem. Phys., 2000, 113, 7756.
72 B. Delley, Int. J. Quantum Chem., 1998, 69, 423.
73 NIST Chemistry WebBook, http://webbook.nist.gov/cgi/cbook.
cgi?ID¼C67641&Units¼SI&Mask¼80#IR-Spec, United States
Department of Commerce, National Institute of Standards
and Technology, Washington, DC, 2003.
74 J. Oomens, B. G. Sartakov, G. Meijer and G. von Helden, Int. J.
Mass Spectrom., 2006, 254, 1.
75 J. Oomens, D. T. Moore, G. Meijer and G. von Helden, Phys.
Chem. Chem. Phys., 2004, 6, 710.
76 M. T. Rodgers and P. B. Armentrout, Mass Spectrom. Rev., 2000,
19, 215.
77 J. Banisaukas, J. Szczepanski, J. R. Eyler, M. Vala, S. Hirata, M.
Head-Gordon, J. Oomens, G. Meijer and G. von Helden, J. Phys.
Chem. A, 2003, 107, 782.
78 A. Fielicke, R. Mitric, G. Meijer, V. Bonacic-Koutecky and G. von
Helden, J. Am. Chem. Soc., 2003, 125, 15716.
79 J. B. Foresman and A. E. Frisch, Exploring Chemistry with
Electronic Structure Methods, Gaussian, Pittsburgh, 1996.
80 S. R. Langhoff, J. Phys. Chem., 1996, 100, 2819.
81 C. Clavaguera-Sarrio, S. Hoyau, N. Ismail and C. J. Marsden, J.
Phys. Chem. A, 2003, 107, 4515.
This journal is
c
the Owner Societies 2007