J. Phys. Chem.

A 2010, 114, 5655–5665 5655

Reactions of Laser-Ablated Zirconium Atoms within a Supersonic Expansion: Insertion

versus Radical Mechanism

S. Soorkia,† C. Pothier, J. M. Mestdagh, and B. Soep*

Laboratoire Francis Perrin (CNRS-URA-2453) CEA/IRAMIS/SerVice des Photons, Atomes et Molécules,
CEN Saclay, F-91191 Gif-sur-YVette cedex, France

J. Liévin
SerVice de Chimie Quantique et Photophysique, 50 AV. F.D. RooseVelt, CP 160/09 B-1050 Bruxelles, Belgium
ReceiVed: December 15, 2009; ReVised Manuscript ReceiVed: March 19, 2010

In a laser ablation type microreactor followed by supersonic expansion, zirconium atoms have been reacted
with methyl fluoride, CH3F (MeF), and mixtures of MeF and dimethylether, CH3-O-CH3 (DME) seeded in
He. With both mixtures, only a number of simple fluorinated products are formed, and they have been identified
by one-photon ionization. All products can be linked to radical reactions either with F atoms, CH3, or ZrF1, 2, 3
radicals. No insertion products of the Grignard reagent type, F -Zr-CH3 could be identified with or in the
absence of DME. On the other hand, evidence has been found for the presence of organometallic compounds
of the type ZrC2Hn)2, 4, 6, which could result from radical attack. Thus, even in conditions where intense
solvation is at work, induced by clustering with polar DME molecules, which can act as stabilizing agents,
a direct insertion mechanism into the C-F bond involving barrier suppression is not at work in our conditions.
The reactivity due to radicals is very effective in this type of reactor, and the products that are efficiently
formed can be quickly stabilized in the expansion. The radical attack supersedes, in the case of zirconium
solvated by DME, the metastable mechanism with Zr(4d)3(5s)1, that is certainly energetically impossible in
the absence of strong reaction barrier suppression by a solvent. High level ab initio calculations performed
at the CASPT2 level of theory are used for characterizing the electronic and geometric structure of the inserted
products. They also reveal striking features of the reaction mechanism that support the absence of observation
of inserted products within solvated clusters of zirconium.

1. Introduction iting transition metals with organic molecules9 producing

Organometallic species are key compounds in organic
chemistry, as initially demonstrated with the preparation of
organomagnesium compounds (Grignard reagents).1 Yet the
observation and characterization of these molecules has es-
sentially been performed in solutions, while the production in
the gas phase of reactive organometallics is more recent and
stems from the development of laser vaporization.2 Among these
systems, compounds containing a metal-carbon bond, like
Grignard reagents, are of high preparative value as they can
functionalize C-H or C-F bonds. Whereas, in the gas phase,
Grignard reagents have not been observed, except for a similar
compound H-Zn-Cl,3,4 inserted transition metal ions have been
prepared and characterized.5 Also, insertion mechanisms have
often been shown to drive direct reactions of halogen-containing
compounds with metals.6
For neutral compounds, a convenient method to characterize
metal bonds is matrix isolation spectroscopy, where infrared
analysis is used to assign the nature of the bonds of the inserted
metal via the bond specific modes: metal stretch or bend infrared
transitions.7 A thorough study of magnesium methyl halides has
been performed using isotopic substitution and calculations,
allowing the characterization of CH3-Mg-F.8 In this fashion,
early on, much cryochemistry has been performed by codepos-
* To whom correspondence should be addressed.
†
Present address: Departments of Chemistry and Physics and Lawrence
Berkeley National Laboratory, University of California, Berkeley, CA 94720
USA.
10.1021/jp911857m  2010 American Chemical Society
Published on Web 04/20/2010
organometallic systems.
A variety of experiments on the reactivity of neutral transition
metals with organic compounds or halogens10-12 has been
performed in the gas phase by different techniques: crossed
beams,11,13 beam-gas,12 and flowcells.14 Insertion mechanisms
have been characterized for reactions involving neutral metals
with hydrocarbons, where both C-C and C-H insertion have
been found to be operative.13 In turn, reactions involving
organohalogen molecules, for example, X-CH3 type com-
pounds, which could lead to bond-breaking or bond-insertion
reactions of the metal atom with the C-X bond, have been
concerned to a much lesser extent in the gas phase. For this
specific reason, it is of fundamental interest to study these latter
gas phase reactions. Because inserted organometallic compounds
are expected to be highly ionic in nature, a study of their
reactivity in clusters could provide insight into how solvation
influences the course of reactions.
The reactions of laser ablated zirconium with MeF, have been
studied in the present work under various conditions of
concentration, expansion, and solvation. The various types of
reactions taking place in these conditions have been searched
for, and we have examined here the possibility to form and
solvate insertion products of zirconium within a supersonic
expansion. Here, laser ablation of Zr has been performed in the
presence of MeF and DME, and mixtures of both diluted in
helium. To investigate the possibility of forming inserted
reaction products of the Grignard reagent type (F-Zr-CH3),
5656 J. Phys. Chem. A, Vol. 114, No. 18, 2010
Figure 1. Schematics of the laser ablation source. The gas pulse is
actuated by a piezoelectric valve with a c.a. -500 V pulse, 200 µs
long. The 532 nm Nd:YAG laser is focused (100 µm spot diameter)
on the rotating and translating Zr rod, in the center of the channel (10
mm long, 1 mm in diameter), terminated with a broad angular aperture
cone.
as described by Andrews et al.,15 DME is mostly employed as
a stabilizing agent by analogy to the preparation of R-Mg-X
compounds in the condensed phase. The pressure and concen-
tration have been systematically varied in order to form clusters
offering determined solvation conditions. After the supersonic
expansion, the reaction products have been analyzed by time-
of-flight mass spectrometry and single-photon ionization. High
level ab initio calculations at the CASPT2 level of theory have
been used here to investigate the properties of inserted com-
pounds: structure, energetics of the ground and lowest excited
states, and the related spin orbit states. The static properties of
the excited and ionic states have important consequences on
the experimentally observed systems; the calculations not only
support the observations, but they also inform on the ionic
character of inserted compounds.
2. Experimental Techniques
The experimental apparatus combines a pulsed supersonic
beam generating the clusters and the products, nanosecond lasers
and a time-of-flight mass spectrometer (TOF-MS). The com-
ponents of the beam are ionized by the UV probe lasers and
mass analyzed using the TOF-MS.
2.1. Beam Source and Detection. Cold clusters are gener-
ated using a conventional laser vaporization source, see Figure
1. It is composed of a pulsed piezoelectric valve, a nozzle of 1
mm diameter followed by a channel of 1 mm diameter and 8
mm overall length. There, a rotating/translating zirconium rod
is ablated by the 532 nm doubled output of a Nd:YAG laser
along the design of Smalley.16 A mixture of 0-5% DME,
0-10% MeF in pure helium is passed through the nozzle. The
gases have been premixed in a cylinder and entrained in the
channel where the Zr vapor is created by laser ablation. After
the ablation zone, the length of the channel is reduced to 3 mm,
followed by a small cone with a broad aperture, 120°. The beam
is then free to expand. The combined geometry of the channel
and the aperture cone ensures a cooling of the ablated zirconium
shortly before expansion. It produces cold ZrF, ZrF2 radicals,
clusters of Zr and DME, plus various types of clusters and mixed
clusters containing MeF and DME. The operating conditions
(see Section 2.4) are adjusted for the production of a variety of
species by increasing the effective pressure within the nozzle
to achieve a controlled number of collisions between zirconium
atoms and the reagents in helium within the channel. This, of
course, leads to cluster formation at the tip of the nozzle. The
Soorkia et al.
effective expansion pressure is therefore adjusted: for conven-
ience we set the backing pressure (P0) at 2 bar and modify the
effective pressure by changing the high voltage applied to the
piezoelectric valve. Increasing the voltage increases the valve
opening and thus the pressure at the nozzle. This is monitored
through the pressure in the source chamber (P1) which will be
indicated in the graphs, proportional to the effective backing
pressure.
The molecular beam is formed by skimming a free jet
expansion through a 2 mm diameter skimmer before entering
the chamber where the detection is performed. After the
skimmer, the molecular beam crosses the ionization laser(s) in
the acceleration region of a 1.23 m Wiley-McLaren17 TOF-
MS collinear with the beam. The mass resolution is 1 at mass
500, and the zirconium isotopes are clearly identified in all
compounds. The isotope intensity pattern serves as a template
to identify zirconium compounds.
2.2. Lasers. A pulsed tunable dye laser (LambdaPhysik LPD
3000) is pumped by the 355 nm output of a nanosecond Nd:
YAG Quantel laser. It is used to ionize the metallic compounds
of the molecular beam with frequency-doubled tunable light
between 207 and 240 nm (6.1-5.2 eV). An excimer laser,
(GAM Laser EX5) operating with a F2 mixture at 157 nm, is
also used to ionize at 7.9 eV. The energy of both these lasers
is in the micro-Joule range per pulse, and their beam diameter
is ∼5 mm, therefore biphotonic effects are not expected.
2.3. Laser Ablation. Ablation is achieved by a frequency-
doubled Nd:YAG laser (532 nm) whose energy is attenuated
by a Glan prism. The energy used for ablation is above 5 mJ
(typically 6 mJ), and the laser is focused over a 100 µm spot
diameter. This energy is necessary for a rather refractory metal
such as zirconium and is 5 times larger than calcium18 or
magnesium19 used in similar experiments.
2.4. Clocking. The time delays between the various com-
ponents of the beam source and detection represent essential
parameters to control. This is achieved via a digital delay
generator controlled by a quartz resonator and computer
controlled (designed and built at Laboratoire de PhotoPhysique
Moléculaire, Université Paris-Sud). The time sequence is the
following: the Nd:YAG lasers are first charged, then the ablation
laser is fired, setting the time origin, t0. The pulsed valve is
opened at t1. The delay t1 - t0 controls the injection of zirconium
in the gas pulse close to its maximum, it is varied in 1 µs steps.
The ionization laser is fired at t2 to allow the clusters to arrive
in the observation zone, t2 - t0 is also controlled by 1 µs steps.
3. Computational Methods
Large scale ab initio calculations have been performed in
order to investigate the valence electronic structure of inserted
compounds of Zr with CH3F. All calculations were performed
with the MOLPRO quantum chemistry package20 running on
the HP-XC 4000 cluster of the ULB/VUB computer center. The
RS2 version21 of the CASPT2 method22,23 implemented in
MOLPRO has been used for optimizing the geometries and
determining the relative stabilities of singlet and triplet states.
The Zr atom is described by the quasi-relativistic Wood-Boring
pseudopotential with 28 core electrons24 and by the correspond-
ing valence basis set from the Stuttgart library24 augmented by
3f and 2g Gaussian primitives playing the role of polarization
functions.25 The C, F, and H atoms are described by the aug-
cc-pVTZ correlation consistent basis sets.26,27 An active space
of 10 electrons distributed into 7 molecular orbitals (MO) (4a′
and 3a′′ for planar Cs geometries) is used in the CASSCF28
calculations preparing the multireference of the CASPT2
Reactions of Laser-Ablated Zirconium Atoms
Figure 2. Clustering of Zr in a mixture of 2% of methyl fluoride in
He at three different backing pressures (P0) increasing from bottom to
top by a factor of 2. The pressure P0 is represented here via the source
pressure P1 (see Section 2), i.e. (a), (b), and (c) at 0.4 × 10-4, 0.6 ×
10-4, and 0.8 × 10-4 mbar, respectively. Also, higher P0 values are
displayed in Figure 5. It appears that the signal of the complex
undergoes an abrupt increase with the change in the backing pressure.
The isotopic pattern of Zr is clearly identified on the Zr · · · F-CH3
complex, while Zr2 and Zr2C have more complex patterns due to
isotopic combinations.
calculations, in which all 18 valence electrons are then cor-
related. The definition of this space is of particular importance
for getting a balanced representation of the electron correlation
between all investigated compounds. The full valence space (18
electrons in 17 active MOs) is unfortunately out of reach of
present computer facilities. It generates 116 million configuration
state functions (CSF) at CASSCF level and 4 billion of
contracted CSF at the RS2 level. We thus performed preliminary
test calculations, in which the size of the active space was
progressively augmented and the convergence of the CASSCF
and RS2 energy differences was investigated. We found that
the 10-electron active space is a good compromise between
accuracy and computational efficiency and that it should recover
a large amount of the valence correlation energy through the
RS2 treatment. This space generates 300 configuration state
functions (CSF) at the CASSCF level and 5 million of contracted
CSF in the RS2 calculation, which is a reasonable size for
applying the CASSCF/RS2 treatment to all calculations. These
include the geometry optimization calculations that were
performed using a numerical quadratic steepest descent algo-
rithm.29
4. Experimental Results
We analyze here the reaction products formed in presence
of MeF, DME, and MeF/DME mixtures, while changing the
concentrations of reagents, the pressure conditions in the
expansion and the ionization energies. The nature of the products
and the complexes they form will provide insight into the gas
phase reactivities of clustered and unclustered species.
4.1. Zirconium-Methyl Fluoride Clusters. At low MeF
concentrations, that is, <3% of MeF in He buffer gas, van der
Waals complexes of methyl fluoride with Zr are formed and
detected at 212 nm. An example of the clustering is given in
Figure 2. These clusters are represented essentially by the (1:1)
Zr · · · F-CH3 complex; it is noticeable above P1 > 0.6 × 10-4
mbar, then reaches a maximum. The (1:2) Zr · · · F-CH3 cluster
is also observed with an intensity no larger than 2% of the
J. Phys. Chem. A, Vol. 114, No. 18, 2010 5657
Figure 3. The energetic scheme for the single photon ionization of
the van der Waals complex Zr · · · F-CH3 is shown on the right-hand
side. The reactivity in the ground neutral metal Zr(4d)2(5s)2, a3F2, and
Zr+ ion (4d)2(5s)1, a4F(3/2), leading to the ZrF and ZrF+ products,
respectively, is shown on the left-hand side. Double-headed arrows
indicate differences in vertical energies with each corresponding value
written on the right-hand side. Note that all values are given in eV.
The experimentally determined value for the ionization energy (5.15
eV) of the van der Waals complex is encircled, and the ionization
energies obtained by calculation are indicated in rectangles (see Section
4.5 and Table 2). The formation of ZrF can be found 1.82 eV lower
than the reagents (see Section 5 for the calculation) and is underlined.
The reaction pathway from the separated reagents Zr + F-CH3 on the
right-hand side leads to ZrF + CH3, passing over a barrier and flying
over the insertion well of F-Zr-CH3. The barrier (hypothetical value)
corresponds to the repulsive penetration of the valence (5s)2 orbitals
of zirconium within the C-F bond. In general, the barrier is reached
at the surface crossing with the metastable configurations36 (4d)3(5s)1
for the neutral and (4d)3(5s)0 for the ion, which have favorable overlap
with the F-CH3 molecule. Owing to a lesser repulsion, the barrier for
the ion insertion will consequently be lower as compared to the neutral
(as is sketched here). The dashed curve represents the ground state of
the inserted compound F-Zr-CH3 in its ground 1A′ state, and the upper
curve corresponds to the 3A′′ state. Both correlate to the same ZrF +
CH3 limit. The arrows at 5.15 eV (240 nm) and 7.9 eV (157 nm)
represent the highest ionization energies used in this work.
(1:1) complex at higher MeF concentrations, ∼4%. When
changing the ionization wavelength to 240 nm, the (1:1)
complex is barely observed, and at 250 nm it is no longer seen.
This sets the ionization threshold at ∼5.15 eV via a one-photon
excitation (the ionization laser has an intensity lower than 2 µJ
over a 4 mm2 spot). The same expansion conditions have been
explored with a fluorine laser at 7.9 eV, and no (1:1) cluster
nor any larger cluster (1:2) were observed in these conditions.
This sets the binding energy, of the ground state (1:1) complex
or any larger complex (1:n), well below 1.27 eV (the binding
energy is equal to the difference between the energy of the laser
at 157 nm and the ionization energy of zirconium 6.63 eV),
see Figure 3, which summarizes the energetics of these systems,
experimental observations, theoretical calculations, and hypo-
thetical values for barrier heights. There, the small binding
energy of the complex is easily compensated by the energy of
the 157 nm laser to produce the separated Zr+ + MeF fragments.
4.2. Zirconium-DME Clusters. Clusters with DME (pos-
sible solvent) are formed much more easily than with MeF.
Indeed they were already observed in a flow tube with a binding
energy of 0.83 eV.30 Therefore, the association complex likely
exists even before the expansion. As displayed in Figure 4, at
212 nm a series of clusters are observed. As the pressure is
increased, the size of clusters steadily increases, whereas the
signal at mass 89.9 amu, the most abundant isotope for
zirconium, decreases.
5658 J. Phys. Chem. A, Vol. 114, No. 18, 2010
Figure 4. Clustering of Zr in a 5% DME in He mixture at various
backing pressures increasing from bottom to top, i.e. (a), (b), and (c)
with P1 ) 1.1 × 10-4, 1.2 × 10-4 and 1.3 × 10-4 mbar, respectively.
The wavelength of the ionization laser is set at 212 nm.
Figure 5. Solvated reaction products ionized at 212 nm with a mixture
of 4% methyl fluoride in He. The backing pressure P0 increases from
bottom to top, as follows with P1 ) (a) 0.4 × 10-4 mbar, (b) 0.8 ×
10-4 mbar, (c) 1.1 × 10-4 mbar and (d) 1.4 × 10-4 mbar. It appears
that the Zr · · · F-CH3 cluster passes over a maximum whereas the
solvated products steadily increase with pressure.
4.3. Reaction Products with Pure MeF or Pure DME. In
He/DME expansions, as the size of clusters increases, a new
series grows in, with broadened peaks. It is readily identified
to zirconium adducts owing to the isotopic patterns and it relates
to the main peaks by a 15 amu mass decrease. The series is
assigned to Zr-O-CH3(DME)n. Another series of solvated
reaction products, ZrO, solvated by DME is also observed but
with sharper peaks.
In He/MeF expansions, as the backing pressure increases, the
concentration of (1:1) Zr · · · F-CH3 clusters decreases while
monitored with the 212 nm ionization laser. A new series of
clusters appears that starts with 90ZrF3(MeF)nG2 at mass 214.9
amu with the number n of attached MeF increasing as shown
in Figure 5. When monitoring the same expansion conditions
at 157 nm, there appears the nonclustered ZrF3 molecule and
its first (1:1) ZrF3 · · · F-CH3 complex (see Figure 6). The smaller
radical 90ZrF2 is also ionized and observed at 127.9 amu with
the 157 nm laser. The intensity of ZrF3+ ions is small but
increases with pressure while that of ZrF2+ decreases. Likely,
the ionization of ZrF3 is at threshold since the corresponding
Soorkia et al.
Figure 6. Clusterized reaction products ionized at 157 nm. The backing
pressure P0 increases from bottom to top, as follows (with P1 ) (a)
0.4 × 10-4 mbar (b) 1.1 × 10-4 mbar, (c) 1.6 × 10-4 mbar and (d) 1.4
× 10-4 mbar.)
Figure 7. Details of the reaction products showing the ZrC2H2, 4, 6
products between ZrF and ZrF2, identified in the insert by isotopic
modeling.
clusters ion signals are much more intense. At the same time,
the intensity of ZrF2 is greater, indicating that its ionization
threshold is well below 7.9 eV. When the concentration of MeF
is increased to 8%, most of the zirconium has been transformed
into products and their clusters at the highest backing pressures.
The 157 nm laser also allows probing of other radicals formed
in the beam as an intense group of masses in Figure 6 between
115 and 126 amu. These species correspond to the radicals
ZrC2Hn)2, 4, 6, which are identified by their zirconium isotope
abundance as shown in Figure 7. The same products are observed
in pure He/DME mixtures using 157 nm photons and seem
characteristic of hydrocarbon-containing species since we could
also detect them in He/CH4 mixtures with the same isotopic pattern.
4.4. Reaction Products Formed in Mixed MeF/DME
Expansions. When the expansion contains a mixture of DME,
at concentrations >1%, and MeF at concentrations >2%, the
(1:1) Zr · · · F-CH3 clusters completely disappear at 212 nm, as
well as the majority of Zr · · · (DME)n clusters. It turns out that
the increase in concentration of MeF in these mixed expansions
favors the formation of products and their complexes, instead
of clusters. In these conditions, the increase in the backing
pressure results also in an increase of the intensity of the
Reactions of Laser-Ablated Zirconium Atoms J. Phys. Chem. A, Vol. 114, No. 18, 2010 5659

Figure 8. Reaction products ZrF3 · · · (MeF)n(DME)m formed in mixed

expansion conditions: (a) the ionization laser is set at 212 nm while
(b) it is at 157 nm. The (n, m) pair indexes respectively the number of
MeF and DME molecules associated with ZrF3. The weak peak at 212
nm ionization at mass 258 amu is likely Zr-O-CH3 · · · (MeF)2(DME)1.

products at the expense of Zr atoms. At the highest pressures, Figure 9. Structures of the (1:1) Zr · · · F-CH3 complex and of the
Zr and ZrF almost disappear. inserted compounds F-Zr-CH3 and CH2dZrHF as discussed in the
A new series of compounds is best seen at concentrations of text in Section 4.5.
8% MeF and 2% DME in helium. The general formula of these
species is ZrF3(MeF)n(DME)m and it starts with n ) 0, m ) 0 TABLE 1: Calculated Relative Stability (eV) of the (1:1)
when ionized at 157 nm and at n ) 1, m ) 1 at 212 nm, as Zr · · · F-CH3 van der Waals Complex, the Grignard Reagent
Type F-Zr-CH3, and the Methylidene Type CH2dZrHF
shown in Figure 8. On the one hand, at 212 nm a simple Inserted Products with Respect to the 1A′ Ground State of
cluster progression of ZrF3 · · · (DME)n>2 is prominent and F-Zr-CH3a
ZrF3 · · · (MeF)1(DME)1 is also distinctly observed as seen in
Figure 8. Two products are seen at 157 nm: ZrF2 and ZrF3 with (1:1) Zr · · · F-CH3 F-Zr-CH3 CH2dZrHF
their MeF clusters starting at n ) 1. On the other hand, at the 3
A′′ 5.59 1.18 2.36
same ionization wavelength no compound of general formula -0.407 0.68
3
Zr(MeF)(DME)n can be observed that would match the mass A′ 5.62 3.14 1.02
1
A′′ 6.43 2.18 0.32
of a solvated inserted compound. Also, as stated above, no 1
A′ 6.94 0 -0.11
broadened Zr(MeF)n peaks can be seen, as resulting from 0 -0.056
evaporation within ions.
a
4.5. Theoretical Calculations. The lowest singlet and triplet The values in italic refer to Cho and Andrews.15
states of the compounds formed by addition of methyl fluoride
TABLE 2: Calculated Vertical Ionization Energies (eV) of:
with zirconium have been characterized at the CASPT2 level the (1:1) Zr · · · F-CH3 Complex Relative to Its Ground 3A′′
of theory (see Section 3). The results include the van der Waals State, the Grignard Reagent Type F-Zr-CH3, and the
complex Zr · · · F-CH3 and two possible inserted compounds, Methylidene Type CH2dZrHF Inserted Products Relative to
the Grignard reagent type F-Zr-CH3 and the methylidene Their Ground 1A′ State
CH2dZrHF. Geometry optimizations have been performed (1:1) Zr · · · F-CH3 F-Zr-CH3 CH2dZrHF
without imposing any symmetry constraints, and it is found that
2
all systems adopt quasi-planar Cs geometries for the following A′′ 6.26 5.67 4.80
2
A′ 6.28 5.57 7.03
low-lying states: 1A′, 1A′′, 3A′, and 3A′′. The equilibrium
geometries are listed in Tables 3 and 4 using the labeling of
atoms defined in Figure 9. The relative stabilities of all states results shed light on the experimental observations (see Table
with respect to the ground state of the F-Zr-CH3 inserted 2). The van der Waals complex is reasonably stabilized in its
compound are given in Table 1. These tables also report for ionic form by charge dipole interactions, and the IE is calculated
comparison purpose the B3LYP DFT results obtained by Cho as 6.27 eV. The inserted compounds exhibit a lower IE, that is,
and Andrews15 for some of the states of the inserted compounds. 5.57 eV for the Grignard reagent type F-Zr-CH3 and 4.8 eV
Mulliken atomic populations are inserted in Tables 3-5, for the methylidene type CH2dZrHF.
together with dipole moment values since they give insight in
the charge distribution of all compounds. Atomic populations
5. Discussion
on the 5s and 4d atomic orbitals of zirconium are also reported
in order to characterize the electronic structure of the transition The discussion will review the formation conditions of the
metal within the molecules. The ionization energies (IE) have different compounds which have been observed, mainly fluori-
also been calculated for the vertical ionization of all species nated zirconium ZrF1, 2, 3 and clusters with MeF and DME. It
from the ground states (see Table 2). The 2A′ and 2A′′ electronic will assign the formation of these compounds to the reaction
states of the corresponding cations were considered. These of radicals, fluorine, and subsequently zirconium-containing
5660 J. Phys. Chem. A, Vol. 114, No. 18, 2010 Soorkia et al.

TABLE 3: Geometrical Parameters of the Inserted Methylidene Compound CH2dZrHF

1 1 3 3
A′ A′′ A′ A′′
a b c b b b
parameters this work Cho and Andrews this work this work this work Cho and Andrewsc
r(C-H1), r(C - H2) 1.102, 1.090 1.085, 1.113 1.090, 1.092 1.087, 1.097 1.079, 1.077 1.092, 1.095
r(C-Zr) 1.966 1.965 2.241 2.057 2.206 2.199
r(Zr-H3) 1.917 1.873 1.894 1.880 1.865 1.867
r(Zr-F) 1.966 1.945 1.998 2.057 1.934 1.940
∠H1CH2 111.2 112.9 109.1 109.2 108.6 109.9
∠CZrF 112.9 112.7 121.3 121.6 115.7 121.0
∠CZrH3 120.5 105.8 118.0 112.7 120.7 117.9
∠H3ZrF 126.6 118.6 120.6 125.7 123.6 121.1
∠H1CZr, H2CZr 106.0,142.7 96.5, 149.9 128.5, 122.3 135.0, 115.7 126.9, 124.5 123.5, 126.6
∠H1CZrH3, H2CZrH3 180.0, 0.0 26.8, -139.3 180.0, 0.0 180.0, 0.0 180.0, 0.0 180.0, 0.0
∠H1CZrF, H2CZrF 0.0, 180.0 158.1, -7.9 0.0, 180.0 0.0, 180.0 0.0, 180.0 0.0, 180.0
q(C) -1.34 -0.77 -1.48 -1.23 -1.45 -0.80
q(H1, H2, H3) 0.41, 0.41, -0.52 -0.25, 0.13, 0.09 0.41, 0.40, -0.62 0.44, 0.37, -0.52 0.43, 0.42, -0.61 -0.30, 0.12, 0.12
q(Zr) [q(5s), q(4d)] d 1.77 [0.14, 1.85] 1.30 2.01 [0.32, 1.54] 1.70 [0.44, 1.68] 1.91 [0.36, 1.58] 1.38
q(F) -0.73 -0.49 -0.72 -0.77 -0.69 -0.53
µ(D) 0.79 3.58 0.40 2.81 1.45 2.05
a
See Figure 9 for the labeling of the atoms. Bond lengths and angles are in Å and degrees, respectively. b This work, CASPT2 calculations.
c
Cho and Andrews,15 B3LYP/6-311+G(2d,p)/LANL2DZ calculations. d Mulliken atomic populations. q(5s) and q(4d) are the Mulliken
populations on the 4s and 5d orbitals of Zr, respectively. Small amounts of the population, not reported here, are also found on the 5p orbital
and on the f and g polarization functions.

TABLE 4: Geometrical Parameters of the Inserted Compound F-Zr-CH3

1 1 3 3
A′ A′′ A′ A′′
parametersa this workb this workb this workb this workb Cho and Andrewsc
r(C-H1), r(C-H2), r(C-H3) 1.086, 1.101, 1.101 1.092, 1.089, 1.089 1.088, 1.089, 1.089 1.085, 1.093, 1.093 1.093, 1.101, 1.101
r(C-Zr) 2.159 2.182 2.594 2.176 2.223
r(Zr-F) 1.891 1.883 1.968 1.896 1.942
∠H1CH3, H1CH3, H2CH3 108.9, 108.9, 107.5 105.7, 105.7, 106.8 119.8, 119.8, 119.8 115.5, 115.5, 110.8 107.7, 108.0, 108.0
∠CZrF 115.6 109.2 166.8 120.6 121.4
∠H1CZr, H2CZr, H3CZr 123.2, 123.2, 103.6 106.4, 115.7, 115.7 77.5, 100.3, 100.3 106.5, 106.5, 106.1 114.4, 109.3, 109.3
∠H1CZrF, H2CZrF, H1CZrF 0.0, 113.6, -113.6 0.0, 98.0, 98.0 0.0, 117.4, -117.4 0.0, 108.5, -108.5 0.0, 121.2, -121.2
q(C) -1.34 -1.51 -1.28 -1.52 -0.95
q(H) 0.34, 0.36, 0.36 0.34, 0.33, 0.3 0.42, 0.44, 0.44 0.32, 0.31, 0.31 0.13, 0.13, 0.13
q(Zr) d [q(5s), q(4d)] 0.91 [1.18, 1.50] 1.16 [0.70, 1.79] 0.70 [0.75, 2.41] 1.23 [0.69, 1.82] 1.12
q(F) -0.64 -0.64 -0.72 -0.66 -0.56
µ(D) 1.72 0.32 6.32 1.5 2.56
a
See Figure 9 for the labeling of the atoms. Bond lengths and angles are in Å and degrees, respectively. b This work, CASPT2 calculations.
c
Cho and Andrews,15 B3LYP/6-311+G(2d,p)/LANL2DZ calculations. d Mulliken atomic populations. q(5s) and q(4d) are the Mulliken
populations on the 4s and 5d orbitals of Zr, respectively. Small amounts of the population, not reported here, are also found on the 5p orbital
and on the f and g polarization functions.

TABLE 5: Geometrical Parameters of the van der Waals Complex (1:1) Zr · · · F-CH3
1 1 3 3
A′ A′′ A′ A′′
a b b b
parameters this work this work this work this workb
r(C-H1), r(C-H2), r(C-H3) 1.083, 1.073, 1.073 1.085, 1.072, 1.072 1.084, 1.072, 1.072 1.085, 1.073, 1.073
r(C-F) 1.448 1.428 1.432 1.426
r(Zr-F) 2.308 2.345 2.338 2.399
H1CH3, H1CH3, H2CH3 112.6, 112.6, 115.1 111.6, 111.6, 114.8 111.8, 111.8, 114.9 111.6, 111.6, 114.6
∠ZrFC 130.8 132.1 130 131.1
∠H1CF, H2CF, H3CF 106.6, 104.4, 104.4 107.2, 105.4, 105.4 107.0, 105.3, 105.3 107.2, 105.6, 105.6
∠H1CFZr, H2CFZr, H1CFZr 0.0, 119.4, -119.4 0.0, 119.1, -119.1 0.0, 119.1, -119.1 0.0, 119.1, -119.1
q(C) -0.51 -0.53 -0.53 -0.52
q(H) 0.36, 0.36, 0.36 0.32, 0.35, 0.35 0.33, 0.35, 0.35 0.33, 0.35, 0.35
q(Zr) [q(5s), q(4d)] c -0.17 [1.83, 2.07] -0.11 [1.88, 2.16] -0.11 [1.85, 2.11] -0.08 [1.78, 2.09]
q(F) -0.39 -0.38 -0.37 -0.41
µ(D) 5.85 5.29 5.28 5.06
a
See Figure 9 for the labeling of the atoms. Bond lengths and angles are in Å and degrees, respectively. b This work, CASPT2 calculations.
c
Mulliken atomic populations. q(5s) and q(4d) are the Mulliken populations on the 4s and 5d orbitals of Zr, respectively. Small amounts of the
population, not reported here, are also found on the 5p orbital and on the f and g polarization functions.

radicals that are formed. We shall comment here upon the 5.1. Zr Clustering and the Reactivity of the Metal Atom.
nondetection of insertion compounds, previously observed in It appears that translationally cold ground-state zirconium does
rare gas matrices after laser ablation of zirconium15 and compare not react with either MeF or DME, since complexes can
with quantum chemistry calculations. efficiently be formed with either molecule. One sees, for instance
Reactions of Laser-Ablated Zirconium Atoms
in Figure 2, that at moderate expansion pressures and low MeF
concentrations, Zr · · · F-CH3 is the only detectable species by
ionization at 212 nm. The same is true when DME is passed at
low concentrations in helium through the laser ablation, only
Zr · · · (DME)1, 2 clusters are observed. Hence, there exists a
barrier to the reaction in the ground state of Zr (4d)2(5s)2 3F2
with either compound, although both reactions with MeF and
DME are highly exothermic. The formation of ZrF (the only
compound documented) is indeed found 1.82 eV lower than
the reagents (from ZrF ) 6.5 eV31 and F-CH3 ) 4.68 eV32).
The same is true for the Zr+ · · · (MeF)1, 2 ions, which are
observed intact, indicating an important barrier for the reaction
of the Zr+ ion with MeF. A similar situation was found in the
case of calcium ion complexes with MeF, where a barrier
separated the ion complex from the lower lying products,
allowing the observation of the (1:1) Ca+ · · · (MeF)1, 2 cluster
ions.33 This justifies the barriers that have been drawn for the
neutral reaction of Zr + MeF and Zr+ + MeF as indicated in
Figure 3.
It can be surmised that DME has a less pronounced excess
energy for the formation of Zr-O-CH3 and a similar type of
barrier exists for all the Zr · · · (DME)n complexes since they are
all easily observed in the expansion. This results from the fact
that, in a very broad sense, O-CH3 is isoelectronic to F.
However, in contrast to fluorinated compounds, the energy
barrier to reaction can be overcome in Zr+ · · · (DME)n cluster
ions at 212 nm, since Zr-O-CH3+ · · · (DME)m products can be
weakly observed in Figure 4. We infer that the reaction occurs
after ionization rather than in the neutral clusters because these
fragment mass peaks are broadened, as are those of larger
clusters of Zr+ · · · (DME)n. The diffuseness of cluster peaks is
typical of fragmentation of clusters in the ionization chamber.
Due to internal excess energy deposited in the ion by the laser
at 212 nm (5.9 eV) after ionization or reaction, clusters fragment
during their flight in the accelerating region of the mass
spectrometer. At 212 nm, the (1:1) DME cluster is easily ionized
and so are larger clusters, but, owing to the intense solvation in
the cluster ion and compared to the neutral, at this wavelength
the energy becomes sufficient to fragment larger clusters. The
peaks for the reaction products from cluster ions also appear as
diffuse in the mass spectra because they have undergone
fragmentation/evaporation processes.
5.2. Fluorinated Zr Compounds. Although barely apparent
at 212 nm, ZrF+ can be observed intensely at 157 nm, in
expansions of ∼4% MeF in He mixtures at low backing
pressures, as shown in Figure 7. The laser energy (7.9 eV) at
157 nm is well above the ionization energy of ZrF, 6.5 eV.31
However, the intensity of ZrF is quickly superseded by that of
ZrF2 and ZrF3 when either the backing pressure or the MeF
concentration are increased. Not surprisingly, ZrF4 is not
observed, even at 157 nm, since its ionization threshold is 14.5
eV.31 The high ionization energy of ZrF4 is due to its saturated
structure, in which all four of the valence electrons of the
zirconium atom are engaged in bonds with F atoms. This high
ionization energy also precludes the observation of clusters of
ZrF4 with MeF that would require more than 5.6 eV of
stabilization energy of the ion with respect to the neutral to
bring the ionization energy down to 7.9 eV. However,
ZrF3 · · · MeFn>2 clusters appear at 212 nm. This is consistent with
the ionization threshold of 7.8 eV, requiring a stabilization of
only 2 eV by two MeF, to reach the ionization threshold of 5.9
eV provided by the 212 nm photons, as seen in Figure 5.
It is clear that the increase in intensity of these ZrFn products
with pressure and MeF concentration is due to reactions of
J. Phys. Chem. A, Vol. 114, No. 18, 2010 5661
ZrFn)0, 1, 2 in the short reaction channel, prior to supersonic
expansion into vacuum. The peaks associated with these species
in the mass spectra are narrow, indicating a neutral reaction
followed by expansion cooling. It is specifically interesting to
note that no clusters of ZrF or ZrF2 have been detected with
MeF. Since these species have been formed in the channel, they
have been cooled in the expansion and have undergone further
collisions with MeF or MeF clusters. The latter binary collisions
could form the ZrF1, 2 · · · MeFn, as the corresponding
ZrF3 · · · MeFn are, in difference, observed. If existing,
ZrF1, 2 · · · MeFn clusters should have been detected, since the
ionization energy of ZrF · · · F-CH3 is in the range accessible
with the 212 nm laser, given the 6.5 eV31 ionization threshold
of ZrF and the expected stabilization of the ionic complex with
MeF by ∼1.5 eV.34 We thus infer that almost no barrier to the
reaction exists between ZrF1, 2 and MeF, within ZrF1, 2 · · · MeFn
clusters.
We shall now discuss the probable mechanisms of formation
of the different fluorinated species in the ablation channel by
first listing some plausible mechanisms. Considering eqs 4-7,
the formation of ZrF radicals in the ablation channel can have
two origins. The whole process can be divided into 2 steps as
described below:
(1) Precursors are formed in the laser ablation zone:
Zrsolid + hν f Zr* (1)
Zr+ + e- f Zr** (2)
CH3F f CH3 + F (3)
(2) Reactions forming the ZrFn)0, 1, 2 compounds:
Zr* + CH3F f ZrF* + CH3 (4)
Zr + F + M f M* + ZrF (5)
ZrF + F f ZrF2 (6)
ZrF + CH3F f ZrF2 + CH3 (7)
5.2.1. Formation of ZrF. Zirconium atoms can be formed
as translationally hot neutral atoms (eq 1) in the laser ablation
process, or as electronically hot metastable states that are either
directly formed in the ablation process or formed by electron-ion
recombination in the ablation plasma (eq 2). Regardless of how
they are formed, spin statistics favor the production of high-
spin quintet metastable states. Equations 4 and 5 suggest
alternative routes to the formation of ZrF and we do not believe
that either the translationally or electronically excited Zr atoms
can be responsible for the majority of the observed products
by reaction with MeF, although MeF is the dominant species
in the gas flow of the reactive channel, according to eq 4. As
illustrated in Figure 3, translational energy will be effective in
surmounting the reaction barrier, and translationally hot Zr atoms
are expected to efficiently lead to ZrF product under low
pressure conditions. This process would become less efficient
at higher overall pressures, owing to the increased collisional
cooling of the translationally hot Zr atoms, however in our
experiments the concentration of ZrF grows approximately
5662 J. Phys. Chem. A, Vol. 114, No. 18, 2010
linearly with increasing pressure. This rules out this channel as
a dominant one. Metastable Zr atoms, on the other hand, would
be expected to retain their electronic energy in a high-pressure
environment and would react as they undergo multiple collisions
with the methyl fluoride molecules. Such metastables are
expected to represent only a fraction of the total number of Zr
atoms, and at high concentrations of MeF and at high pressures,
the metal atoms have all reacted away (see Figure 6 at P1 )
1.6 × 10-4 mbar, for example). Since only a fraction of Zr atoms
are in metastable states, this observation shows that an effective
reaction path for ground state atoms must exist, namely, the
alternative route following eq 5.
Then the dominant process is the dissociation of CH3F
molecules in the plasma by electron impact or eventually by
multiphoton absorption, into CH3 radicals and F atoms. These
F atoms react at hard sphere collision rates with zirconium atoms
(and ZrF1, 2) as shown in eqs 3-6. We favor this latter
possibility, since the companion product ZrC2H6 is also found
by ionization at 157 nm as seen in Figure 6. This product appears
intensely and is most likely formed by the addition of two
methyl radicals, produced by the dissociation of methyl fluoride,
to a zirconium atom. The pressure increase also stabilizes the
newly formed ZrF, eq 5.
5.2.2. Formation of ZrF2, 3, 4. The formation of ZrF radicals
induces further reactions. Indeed, further fluorination of ZrF
should occur with collisions with MeF (eq 7), or (MeF)m
clusters, rather than with F atoms (eq 6), since MeF is more
abundant by an order of magnitude than F atoms in the gas
flow of the reactive channel. We have observed that ZrF1, 2 are
reactive radicals, without reaction barrier with MeF, as noted
in Section 5.2. Thus, these ZrFn radicals do not need the presence
of F atoms to form ZrFn+1. Indeed, when the backing pressure
is increased, then the number of collisions of ZrF1, 2 with MeF
increases in the reactive channel. In these conditions, fluorine
atoms are scavenged by collisions with abundant MeF35 at hard
collision efficiency, to form HF and CH2F. From their quickly
decreasing number along the channel, they become no longer
available for reactions with the newly formed ZrF1, 2, while the
number of collisions with MeF does not decrease at the same
location. As the concentration of the ZrF2, 3 products increases
with the backing pressure, we consider that these radicals are
formed by radical reactions of ZrF1, 2 with MeF or subsequently
with (MeF)n clusters. In this view, the concentration of F atoms
should sharply decrease after the ablation by reactions with Zr
atoms or MeF within the channel. Then, with a certainly lesser
efficiency, the newly born ZrF1, 2 radicals will react with MeF
in the channel. After the expansion, as noted, we observe only
ZrF3 · · · (MeF)n>2 as clusters of products. Once ZrF1, 2 have been
captured by the cluster there remains more than 100 µs before
the detection, thus much time for reaction and cooling by
evaporation of MeF molecules. Although collisions of ZrF1, 2
with MeF are more frequent than with the abundant MeF clusters
(several % concentration of MeF in helium), the capture in the
cluster should give an advantage to that reaction, forming ZrF4.
In summary, the formation of the fluorinated compounds is
stepwise and the reactivity of the radicals likely decreases as F
> ZrF > ZrF2 . ZrF3. This stems from both the decreasing
number of unbonded electrons on zirconium and the increasing
steric effect on the radical considered. In this reactivity, the
reactive channel and the expansion both play a role.
5.2.3. Other Zr Reaction Products. Three Zr-hydrocarbon
compounds have been identified as ZrC2Hn)2, 4, 6, and there may
be others in smaller quantities. The structure of n ) 6 can be
CH3-Zr-CH3 and results from the recombination of Zr with
Soorkia et al.
two methyl radicals. This is confirmed by identical results
obtained with several hydrocarbon sources of radicals, all of
which generated the same three products. The resulting hot
product in the laser plasma can dehydrogenate to form these
products with reasonable probability. The nonobservation of
mixed bonds F-Zr-CH3 by radical recombination is under-
standable since F atoms are scavenged by Zr and MeF35 (see
discussion in Section 5.2.2), whereas CH3 radicals have no other
scavengers than Zr, aside from CH3 in lesser quantities. In other
words, to form these mixed bond compounds with reasonable
probability, an equal collision probability must exist between
Zr and the F or CH3 radicals, which is unlikely from the lower
F atom concentration as referred to CH3. The formation of the
inserted hydrocarbon zirconium products via a radical pathway
could be the preferred route in the gas phase but was not
observed here.
5.3. Search for MeF Inserted Compounds. We scrutinized
the various results for the formation of inserted compounds of
the type F-Zr-CH3 · · · (DME)n>2. We knew that in the gas phase
the outcome of insertion reactions of zirconium or yttrium with
saturated or sometimes unsaturated hydrocarbons is generally
the elimination of H2 molecules after flying over the insertion
well,36,37,11 since a high barrier has to be overcome, as shown
in Figure 3. In the reaction of Ru with MeF studied by quantum
chemistry,38 the insertion does not proceed into the C-H bond
but into the C-F bond with a large barrier, 30 kcal mol-1 (1.30
eV), from the ground atomic state a5F. Thus, the purpose of a
third body in a cluster was 2-fold: (1) to collisionally stabilize
the system into the insertion well during its course over it and,
most importantly, (2) the second effect of the cluster should be
to decrease the barrier to insertion, making a facile passage into
the insertion well. The barrier suppression effect adds to the
third-body effect and decreases the total energy to dispose in
the recoil of this third body. The Zr-F bond is strongly ionic,
whereas the other bond Zr-CH3 is less ionic as seen in Table
4 and Section 5.4. Preliminary calculations have also shown
that all insertion intermediates, en route to the insertion well,
have an almost positive unit of charge on Zr. Although the
neutral reagents are weakly stabilized, the intermediate charge
transfer and the final products should be stabilized by polar
solvent molecules relative to the neutral compounds, thus
decreasing the barrier. A typical case of stabilization of ionic
products by reactions with clusters has been found in the
reactions of Mg+ with (H2O)n clusters.39 This is explained by a
more efficient stabilization for MgOH+ than for Mg+. Thus,
the threshold for production of MgOH+ is observed for
MgOH+ · · · (H2O)n)5. Hence, if the insertion mechanism were
active, we would expect the detection of systems of general
formula F-Zr-CH3(MeF)n(DME)m, in the highest expansion
conditions, owing to barrier suppression for the ionic intermedi-
ate. The bottom of the inserted well has been calculated as 5.59
eV below the ground state 3A′′ of the Zr · · · F-CH3 van der
Waals complex, as represented in Figure 3. The vertical
ionization energy for the inserted product F-Zr-CH3 is found
as 5.57 eV by calculations (see Table 1), thus it should be
observed when solvated, at 212 nm. In all the spectra, there are
hardly any unassigned peaks in mixed helium MeF, or helium,
MeF, and DME expansions that we have investigated; as can
be seen in Figures 5 or 8, most masses relate to fluorinated Zr
reaction products as described in the preceding section but none
to clusters of F-Zr-CH3. The mass peaks at ∼124 amu in
Figures 2 and 5 correspond to the (1:1) Zr · · · F-CH3 complex
as described in Section 5.1, which has the same mass as the
free inserted species, F-Zr-CH3. The (1:1) complex appears
Reactions of Laser-Ablated Zirconium Atoms
already at increasing P0 pressures in dilute mixtures of MeF in
helium (∼2%), at higher concentrations it increases with
pressure, then decreases as seen in Figure 5. The peaks cannot
be assigned to the inserted free product since the increase of P0
should favor their growth and the appearance of solvated inserted
products F-Zr-CH3(MeFn). This we do not observe and
exclude the observation of nonsolvated F-Zr-CH3 inserted
products.
5.3.1. Comparison with Results in Rare-Gas Matrices. The
solvated inserted products have not been observed in any of
our conditions. In more diluted mixtures of MeF and DME,
the solvation effects may be insufficient to suppress the barrier
and solvate the inserted products; at higher solvent concentra-
tions, other reaction products are formed instead. However, in
cryogenic matrices, Cho and Andrews15 have found evidence
of two inserted compounds that interconvert by irradiation at
240 nm, F-Zr-CH3 and the methylidene CH2dZrHF. These
two compounds have equivalent energies within 0.21 eV (see
Table 1) and are converted by R hydrogen migration after the
passage of a barrier. Both compounds are formed during the
codeposition of ablated zirconium atoms and methyl fluoride.
There is a major difference in the reaction conditions with our
cluster work: in the matrix deposition, the ablated Zr atoms,
containing a proportion of metastables and ions are swept into
the matrix where MeF is independently deposited with argon.
No F atoms are formed there, and metastable and ions have
time to react in the definitive presence of methyl fluoride,
forming exclusively the inserted product (or simply quenching
the electronic excitation). This can account for the formation
of F-Zr-CH3 but does not explain the production of the
methylidene CH2dZrHF, since to reach it there is a further
barrier due to the R migration of a H atom from the methyl.
Also, the insertion reaction is immediately arrested at the surface
of argon in matrix deposition. Insertion reactions are difficult
to prepare away from condensed phases for exothermic reactions
after barriers, since the potential surface gradient drives away
the system directly to the noninserted products, flying above
the deep inserted potential well, see Figure 3.6,37 There needs
to be caging within a matrix to prevent ZrF and CH3 from flying
away from each other.
From the results of Cho and Andrews,15 we expected that at
sufficient concentrations of MeF, solvation and caging should
be effective in producing the inserted systems. From the
preceding we are led to infer that the reaction leading to the
inserted compounds originates from metastable zirconium atoms
in the matrix codeposition experiment. The reaction within
metastable should have a barrier, probably as for the Zr+ ion
(not reactive with MeF), although both barriers should be lower
than for ground neutral zirconium. In this respect, we can refer
to the calculations on Ru + MeF where a barrier exits,38
although lower, for the 3F Ru metastable state. However on this
excited state surface, nonadiabatic relaxation can occur to the
ground state surface, providing the necessary excess energy to
overcome the barrier in the ground state. The 1P Ca + MeF
excited state reaction was explained in this way via a non
adiabatic relaxation to a triplet surface.18 We expect that unless
this excitation is achieved within the cluster, the reaction cannot
reach the inserted compound. The result of the electronic
excitation of zirconium will be addressed for the (1:1)
Zr · · · F-CH3 complex in a forthcoming paper.40 The same
applies to ground state ions that, as represented in Figure 3,
have a barrier to reaction, which allows for their observation.
5.4. Theoretical Results and Inserted Compounds. DFT
and CASPT2 calculations agree for predicting that the ground
J. Phys. Chem. A, Vol. 114, No. 18, 2010 5663
state of CH2dZrHF is a closed shell singlet state, which is close
in energy (within 0.1 eV) to the corresponding singlet state in
F-Zr-CH3. However the equilibrium geometry of the ground
state of CH2dZrHF obtained by both methods is different:
whereas CASPT2 converges to a planar structure, DFT predicts
a significantly nonplanar one. Our result is yet compatible with
an ethylene-like CH2dZrHF structure corresponding to a 1A′
assignment for the ground state of this compound. Another
disagreement concerns the energy position of the 3A′′ state with
respect to the ground singlet state that is found to be significantly
more stable by the DFT calculations. The discrepancy amounts
to 1.6 and 1.7 eV for F-Zr-CH3 and CH2dZrHF, respectively,
and leads to the prediction by DFT of a 3A′′ ground state for
F-Zr-CH3. This result is rather surprising because it would
suggest that the open-shell structure would be more stable than
the closed one. Some mismatch is also observed in the charge
distribution (see Mulliken charges and dipole moment values.
These discrepancies are probably due to the fact that B3LYP
calculations are not well suited for describing the electronic
structures of such compounds involving noncovalent interactions
and exhibiting a multiconfigurational character induced by the
zirconium atom. This justifies our use of a multireference
approach like CASPT2 for a proper description of the properties
of such systems. Also the increased stabilization of the singlet
surfaces in the inserted compounds can be presumably linked
with the strong ionic character of zirconium in the product that
destabilizes the triplet product.
As expected, the van der Waals complex Zr · · · F-CH3 is
found to be higher in energy than the inserted compounds. It
exhibits a triplet ground state (3A′′) and a close lying 3A′ excited
state, compatible with the a3F2 ground state of the zirconium
atom. This shows that the insertion of MeF does not funda-
mentally shuffle the spin orbit coupling in zirconium, as the
free metal is in a 3F2 ground state. Taking the 1A′ state of
F-Zr-CH3 as the reference, we thus predict an energy
difference of ca. 5.6 eV with respect to the 3A′′ state of
Zr · · · F-CH3. The difference between both inserted compounds
CH3-ZrF (open shell) and CH2dZrHF (closed shell) amounts
to 0.11 eV only, which can be considered as insignificant.
Zirconium is, as expected, found to be almost neutral in
Zr · · · F-CH3 with an electron population close to the (5s)2(4d)2
ground state atomic configuration. On the other hand, zirconium
bears between one and two positive charges in the inserted
compounds, with the lost fractions of electron charge coming
from the 5s and 4d orbitals. It is interesting to examine from
Tables 3-5 how some of the geometrical parameters change
from one state to another of the same species. As expected, no
spectacular geometry changes occur within the van der Waals
complex, but more important variations are observed in the
inserted compounds. In particular, the F-Zr-C bending angle
is found to be sensitive to the nature of the electronic state and
coupled to the angular positions of the methyl hydrogens. Let
us mention the geometry of the 3A′ state of F-Zr-CH3, which
exhibits a F-Zr-C angle not so far from linearity (166.8° to
be compared to the values around 110° of the other states) and
a large C-Zr distance.
The calculations, which have been conducted at a high level,
are supportive of the reaction mechanism, forming here mainly
ZrFn and not the inserted compounds, as we have previously
emphasized. The reaction path after the barrier passes quickly
over the insertion well of F-Zr-CH3 3A′ correlating at infinite
distance with ZrF + CH3, as illustrated in Figure 3, with an
excess energy of 1.82 eV. F-Zr-CH3 is a strongly ionic species
with a 1.7 positive charge on the zirconium atom, where almost
5664 J. Phys. Chem. A, Vol. 114, No. 18, 2010
two electrons have been transferred to the adjacent fluorine and
carbon atoms, see Table 4. Therefore, even in the presence of
solvent to lower the reaction barrier, a two electron process has
to be achieved to form this compound. This can be achieved
probably only too late, along the reaction pathway. Ultimately,
if formed, in a 3A′ state, this compound could interconvert into
the ground singlet state 1A′. There is another very interesting
result of the calculations: the ionic compound F-Zr-CH3+ in
the ground state 2A′, is almost at the same energy as the complex
Zr · · · F-CH3 3A′. This situation is unusual and can result in
chemionization where the reaction flux encompasses the ion well
and can be quenched into that well by autoionization. Unfor-
tunately, in our experiment these ions cannot be detected as
they are blocked by a grid in the time-of-flight mass spectrometer.
Conclusions
In a laser ablation microreactor, zirconium atoms have been
reacted with MeF alone and in presence of DME. With MeF,
simple products have been identified by one photon ionization,
of general formula ZrFi(MeF)n or ZrC2Hn)2, 4, 6. Our studies
suggest that the majority of the products are formed via radical
reactions either with F, CH3 or ZrF1, 2, 3. Organometallic products
have been found containing hydrocarbon groups ZrC2Hn)2, 4, 6,
some of them should have an inserted structure (probably
ZrC2H6). They result from radical attack of Zr by CH3 radicals.
No products of the type F-Zr-CH3 could be observed.
Alternatively, in conditions where intense solvation is at work,
the direct insertion mechanism involving barrier suppression is
also not observed. These conditions are in marked contrast with
those of matrix deposition for similar compounds,15 where there
is a strong caging effect by argon that can allow the formation
of these inserted zirconium-organohalogen compounds likely
via a metastable configuration of the metal. In the present
experiment, the radical attack supersedes the mechanism with
metastable atoms Zr(4d)3(5s)1, which is certainly inactive for
insertion in absence of solvation. This situation could be met
in reactions proceeding at the surface of argon clusters for
immediate stabilization and in presence of solvent to suppress
the barrier to insertion. The mechanism involving other metal
atom configurations10,5 relies on the hybridization of the (n -
1)dm orbitals of the metal with a half vacant exterior ns1 orbital.
In the cation, this latter orbital stands as ns1 while in the neutral
only low lying metastable excited electronic states bear the
appropriate (n - 1)dm+1ns1 instead of (n - 1)dmns2, which is
filled and repulsive to the approach of reactants. This might be
at the origin of the role played by solvent molecules in neutral
atom reactions in preferentially lowering the metastable con-
figuration with respect to the closed shell ...ns2 configuration.
Also the solvation mechanism can be electrostatic in nature with
a strong stabilization of the ionic products and transition states.
This situation could be sought in transition metals for which
the insertion within the ions is facile and documented, since in
a broad sense the neutral potential energy surfaces parallel the
ionic ones in transition metal reactions. High level ab initio
calculations at the CASPT2 level of theory reveal important
features of the inserted product: its very low ionization energy,
placing the inserted ion lower than the neutral complex. Also
the Mulliken charge on zirconium is high in the inserted
complex and close to +1.23. These two properties have
important consequences on any possible formation of inserted
products in the gas phase via clusters and support its description
as a Grignard-type intermediate and the fact that we did not
observe them. The calculations show also the existence of a
low lying methylidene product in a 1A′ ground state.
Soorkia et al.
Acknowledgment. B. S. is grateful to C. Charrière, Labo-
ratoire de Photophysique Moléculaire (CNRS UPR 3361), for
the construction and help in the use of the digital delay generator
essential for this experiment. S. S. gratefully appreciates the
skillful work and advice of T. Bernard from the CEA/DSM/
IRAMIS/SPCSI for the machining operations of the ablation
source. J. L. thanks the Communauté Française de Belgique
(ARC Contract) and the F.NRS (I.I.S.N project) for financial
support.
References and Notes
(1) The Use of Organomagnesium Compounds in PreparatiVe Organic
Chemistry. Nobel lecture, http://nobelprize.org/nobel_prizes/chemistry/
laureates/1912/grignard-lecture.html. Grignard, V. 1912.
(2) Dietz, T. G.; Powers, D. E.; Duncan, M. A.; Smalley, R. E. J. Chem.
Phys. 1981, 74, 6511.
(3) Yu, S. S.; Shayesteh, A.; Fu, D. J.; Bernath, P. F. J. Phys. Chem.
A 2005, 109 (18), 4092–4094.
(4) Pulliam, R. L.; Sun, M.; Flory, M. A.; Ziurys, L. M. J. Mol.
Spectrosc. 2009, 257 (2), 128–132.
(5) Eller, K.; Schwarz, H. Chem. ReV. 1991, 91 (6), 1121–1177.
(6) Jaffe, R. L.; Pattengill, M. D.; Mascarello, F. G.; Zare, R. N.
J. Chem. Phys. 1987, 86, 6150–6170.
(7) Solov′ev, V. N.; Sergeev, G. B.; Nemukhin, A. V.; Burt, S. K.;
Topol, I. A. J. Phys. Chem. A 1997, 101, 8625–8630.
(8) Bare, W. D.; Andrews, L. J. Am. Chem. Soc. 1998, 120, 7293–7301.
(9) Ozin, G. A.; McIntosh, D. F.; Power, W. J.; Messmer, R. P. Inorg.
Chem. 1981, 20, 1782–1792.
(10) Weisshaar, J. C. Acc. Chem. Res. 1993, 26, 213–219.
(11) Proctor, D. L.; Davis, H. F. Proc. Natl. Acad. Sci. U.S.A 2008,
105, 12673–12677.
(12) Hughes, T. J.; Levy, M. R. Phys. Chem. Chem. Phys. 2000, 2, 651–
659.
(13) Reichert, E. L.; Yi, S. S.; Weisshaar, J. C. Int. J. Mass Spectrom.
2000, 195-196, 55–69.
(14) Carroll, J. J.; Haug, K. L.; Weisshaar, J. C. J. Am. Chem. Soc.
1993, 115 (15), 6962–6968.
(15) Cho, H. G.; Andrews, L. J. Am. Chem. Soc. 2004, 126, 10485–
10492.
(16) Powers, D. E.; Hansen, S. G.; Geusic, M. E.; Puiu, A. C.; Hopkins,
J. B.; Dietz, T. G.; Duncan, M. A.; Langridge-Smith, P. R. R.; Smalley,
R. E. J. Phys. Chem. 1982, 86 (14), 2556–2560.
(17) Wiley, W. C.; McLaren, I. H. ReV. Sci. Instrum. 1955, 26, 1150–1157.
(18) Gloaguen, E.; Sanz, C. S.; Collier, M.; Gaveau, M. A.; Soep, B.;
Roncero, O.; Mestdagh, J. M. J. Phys. Chem. A 2008, 112, 1408–1420.
(19) Soorkia, S.; Collier, M.; Soep, B.; Mestdagh, J. M. J. Chem. Phys.
2009, 131 (22), 224319.
(20) Werner, H.-J.; Knowles, P. J.; Lindh, R.; Manby, F. R.; Schütz,
M.; Celani, P.; Korona, T.; Mitrushenkov, A.; Rauhut, G.; Adler, T. B.;
Amos, R. D.; Bernhardsson, A.; Berning, A.; Cooper, D. L.; Deegan,
M. J. O.; Dobbyn, A. J.; Eckert, F.; Goll, E.; Hampel, C.; Hetzer, G.; Hrenar,
T.; Knizia, G.; Köppl, C.; Liu, Y.; Lloyd, A. W.; Mata, R. A.; May, A. J.;
McNicholas, S. J.; Meyer, W.; Mura, M. E.; Nicklass, A.; Palmieri, P.;
Pflüger, K.; Pitzer, R.; Reiher, M.; Schumann, U.; Stoll, H.; Stone, A. J.;
Tarroni, R.; Thorsteinsson, T.; Wang, M.; Wolf, A. MOLPRO, Version
2008.1, A Package of ab initio Programs; 2008.
(21) Werner, H. J. Mol. Phys. 1996, 89, 645–661.
(22) Andersson, K.; Malmqvist, P. A.; Roos, B. O.; Sadlej, A. J.;
Wolinski, K. J. Phys. Chem. 1990, 94, 5483–5488.
(23) Andersson, K.; Malmqvist, P. A.; Roos, B. O. J. Chem. Phys. 1992,
96, 1218–1226.
(24) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor.
Chim. Acta 1990, 77, 123–141.
(25) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114, 3408–
3420.
(26) Dunning, T. H. J. Chem. Phys. 1989, 90, 1007–1023.
(27) Kendall, R. A.; Dunning, T. H.; Harrison, R. J. J. Chem. Phys.
1992, 96, 6796–6806.
(28) Knowles, P. J.; Werner, H. J. Chem. Phys. Lett. 1985, 115, 259–267.
(29) Sun, J. Q.; Ruedenberg, K. J. Chem. Phys. 1993, 99, 5257–5268.
(30) Pedersen, D. B.; Zgierski, M. Z.; Anderson, S.; Rayner, M. D.;
Simard, B.; Li, S.; Yang, D.-S. J. Phys. Chem. A 2001, 105, 11462–11469.
(31) Hildenbrand, D. L.; Lau, K. H. J. Chem. Phys. 1997, 107, 6349.
(32) Weast, R. C.; Astle, M. J.; Beyer, W. H. Handbook of Chemistry
and Physics, 65th ed.; CRC: Boca Raton, Florida, 1984.
(33) Mestdagh, J. M.; Spiegelman, F.; Gloaguen, E.; Collier, M.; Lepetit,
F.; Gaveau, M. A.; Sanz, C. S.; Soep, B. J. Phys. Chem. A 2006, 110,
7355–7363.
(34) Elhanine, M.; Dukan, L.; Maitre, P.; Breckenridge, W. H.; Massick,
S.; Soep, B. J. Chem. Phys. 2000, 112, 10912–10925.
Reactions of Laser-Ablated Zirconium Atoms
(35) Persky, A. Chem. Phys. Lett. 2003, 376, 181–187.
(36) Carroll, J. J.; Haug, K. L.; Weisshaar, J. C.; Blomberg, M. R. A.;
Siegbahn, P. E. M.; Svensson, M. J. Phys. Chem. 1995, 99, 13955–13969.
(37) Willis, P. A.; Stauffer, H. U.; Hinrichs, R. Z.; Davis, H. F. J. Phys.
Chem. A 1999, 103, 3706–3720.
(38) Colmenares, F. Mol. Phys. 2009, 107, 1513–1519.
J. Phys. Chem. A, Vol. 114, No. 18, 2010 5665
(39) Sanekata, M.; Misaizu, F.; Fuke, K.; Iwata, S.; Hashimoto, K. J. Am.
Chem. Soc. 1995, 117, 747–754.
(40) Soorkia, S.; Mestdagh, J. M.; Gaveau, M. A.; Soep, B. Chem. Phys.
Lett. 2010, DOI: 10.1016/j.cplett.2010.03.083.
JP911857M