Crossed beams studies of Mo(a 7S3) and Mo*(a 5S2) collisions with CH4 and C2H6

Article abstract of DOI:10.1063/1.481018

The interactions of Mo(a ⁷S₃) and Mo*(a ⁵S₂) with methane, CH₄, and ethane, C₂H₆, were studied under single collision conditions using the crossed molecular beams technique. Ground state Mo(a ⁷S₃) atoms were found to be unreactive at all collision energies studied up to 〈E_coll〉 =35.4 kcal/mol. Nonreactive scattering of Mo(a ⁷S₃) with methane and ethane was studied and compared to collisions with Ne and Ar. A forward peaking center-of-mass angular distribution, Τ(Θ), was necessary to simulate the elastic collisions with inert gases as well as inelastic collisions with the alkanes. At a collision energy of 14.4 kcal/mol with CH₄ and 21.0 kcal/mol with C₂H₆, inelastic collisions were found to transfer ～10% and ～19% of the initial kinetic energy into alkane internal energy, respectively. For collisions of Mo*(a ⁵S₂)+CH₄, the dehydrogenation product, MoCH₂, was observed at all collision energies studied down to 2.1 kcal/mol. The reaction Mo*(a ⁵S₂)+C₂H₆→MoC₂H ₄+H₂ was observed down to 〈E_coll〉=4.5 kcal/mol. For a given total energy (electronic+translational), it was found that electronic energy is highly effective in promoting this reaction whereas translational energy is ineffective.

Full text of DOI:10.1063/1.481018

Crossed beams studies of Mo (a 7 S 3 ) and Mo * (a 5 S 2 ) collisions with CH 4 and C 2
H 6
Ryan Z. Hinrichs, Peter A. Willis, Hans U. Stauffer, Jonathan J. Schroden, and H. Floyd Davis
Citation: The Journal of Chemical Physics 112, 4634 (2000); doi: 10.1063/1.481018
View online: http://dx.doi.org/10.1063/1.481018
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/112/10?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Trajectory surface hopping study of the O(3P) + C2H2 reaction dynamics: Effect of collision energy on the extent
of intersystem crossing
J. Chem. Phys. 140, 044314 (2014); 10.1063/1.4862407
Experimental and theoretical studies of the O(3P) + C2H4 reaction dynamics: Collision energy dependence of
branching ratios and extent of intersystem crossing
J. Chem. Phys. 137, 22A532 (2012); 10.1063/1.4746758
Direct dynamics study of energy transfer and collision-induced dissociation: Effects of impact energy, geometry,
and reactant vibrational mode in H 2 CO + – Ne collisions
J. Chem. Phys. 119, 3040 (2003); 10.1063/1.1588634
Crossed beams study of C-H bond activation: Mo ( 5 S 2 )+ CH 4 →MoCH 2 +H 2
J. Chem. Phys. 108, 2665 (1998); 10.1063/1.475658
Crossed-beam reaction of carbon atoms with hydrocarbon molecules. III: Chemical dynamics of propynylidyne (l-
C 3 H ;X 2 Π j ) and cyclopropynylidyne (c- C 3 H ;X 2 B 2 ) formation from reaction of C ( 3 P j ) with acetylene,
C 2 H 2 (X 1 Σ g + )
J. Chem. Phys. 106, 1729 (1997); 10.1063/1.474092
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.174.254.155 On: Wed, 24 Dec 2014 15:44:22

JOURNAL OF CHEMICAL PHYSICS
VOLUME 112, NUMBER 10
8 MARCH 2000
7
5
Crossed beams studies of Mo„a S … and Mo*„a S … collisions
3
2
with CH and C H
6
4
2
Ryan Z. Hinrichs, Peter A. Willis, Hans U. Stauffer, Jonathan J. Schroden,
a)
and H. Floyd Davis
Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell University,
Ithaca, New York 14853-1301
͑
Received 1 October 1999; accepted 14 December 1999͒
7
5
The interactions of Mo(a S ) and Mo*(a S ) with methane, CH , and ethane, C H , were studied
3
2
4
2
6
under single collision conditions using the crossed molecular beams technique. Ground state
7
͗ ͘
Mo(a S ) atoms were found to be unreactive at all collision energies studied up to Ecoll
3
7
ϭ35.4 kcal/mol. Nonreactive scattering of Mo(a S ) with methane and ethane was studied and
3
compared to collisions with Ne and Ar. A forward peaking center-of-mass angular distribution,
T(⌰), was necessary to simulate the elastic collisions with inert gases as well as inelastic collisions
with the alkanes. At a collision energy of 14.4 kcal/mol with CH and 21.0 kcal/mol with C H ,
4
2
6
inelastic collisions were found to transfer ϳ10% and ϳ19% of the initial kinetic energy into alkane
5
internal energy, respectively. For collisions of Mo*(a S )ϩCH , the dehydrogenation product,
2
4
MoCH , was observed at all collision energies studied down to 2.1 kcal/mol. The reaction
2
5
Mo*(a S )ϩC H →MoC H ϩH was observed down to
͗ ͘
Ecoll ϭ4.5 kcal/mol. For a given total
2
2
6
2
4
2
energy (electronicϩtranslational), it was found that electronic energy is highly effective in
promoting this reaction whereas translational energy is ineffective. © 2000 American Institute of
Physics. ͓S0021-9606͑00͒00910-7͔
I. INTRODUCTION
iron and copper have been cocondensed with methane in
inert gas matrices at 15 K.8 Most ground state atoms were
found to be unreactive due to a large potential energy barrier
for insertion. However, upon irradiation of the matrix using
ultraviolet light, electronically excited atoms were found to
–10
Saturated transition metal complexes are generally unre-
active with alkanes under most conditions. However, the
1
insertion of coordinatively unsaturated transition metal cen-
ters ͑generated thermally or by photodissociation͒ into the
C–H bonds of alkanes has been demonstrated in a number of
insert spontaneously into the C–H bonds forming species
8
1
,2
such as HFeCH . Infrared spectra of these C–H insertion
select cases. For example, Bergman and co-workers found
3
intermediates were obtained, providing the first direct insight
into the structure of the initially-formed complexes.
that photolysis of the stable Cp*Rh͑CO͒ complex leads to
CO loss producing Cp*RhCO, where Cp*ϭC ͑CH ͒ . This
2
5
3 5
1
6 electron species inserts spontaneously into many alkanes,
In the gas phase, a number of transition metal cations
3
,4
13,14
including CH , at room temperature. Subsequently, these
reactions have been studied in solution, the gas phase, in
liquefied noble gas solvents, and by ab initio methods. In-
terestingly, although the iridium and rhodium complexes
were both found to be highly reactive, the analogous com-
plex involving cobalt was not. Calculations showed that the
absence of reaction of the Co complex is likely due to a
potential energy barrier for insertion of the triplet high-spin
ground state into CH . In the case of the Ir and Rh com-
plexes, on the other hand, the ground states were found to be7
low spin singlets, having a negligible barrier for insertion.
Clearly, the combination of experiment and theory is making
it possible to understand why certain transition metal com-
plexes are highly reactive, whereas other closely related spe-
cies are not.
In an effort to better understand the electronic factors
controlling insertion of neutral d-electron centers into al-
kanes in the absence of ligand effects, reactions of transition
metal atoms have been studied in low temperature
have been found to react efficiently with alkanes.
It is
4
5
generally believed that C–H insertion is more favorable than
C–C insertion because the spherical H atom can undergo
multicenter bonding, effectively decreasing the energy at the
transition state. Such interactions are less favorable for the
6
7
6
3
highly directional sp orbitals involved in the C–C bond in
alkanes such as ethane, leading to higher C–C insertion
15
barriers. However, for transition metal cations, both C–H
7
4
and C–C bond insertions have been observed in several
systems.1
6,17
In some cases, it has been found that different electronic
states of a given metal ion may react by different mecha-
nisms, possibly resulting in different chemical products. For
example, Armentrout and co-workers found that reactions of
ϩ
5
ground state V ( D) with HD lead to equal amounts of
ϩ
ϩ
VH and VD , indicating that the reaction involves an in-
1
8
sertion mechanism. Reactions of the excited triplet states of
ϩ
3
ϩ
V , primarily the F states, favor production of VH over
8
–12
matrices.
For example, transition metal atoms such as
ϩ
VD by a factor of three, indicating a direct abstraction
1
8
ϩ 4
mechanism. For reactions with CH and C H , Fe ( F)
and Fe ( D) both react by a direct abstraction mechanism,
4
2
6
a͒Electronic mail: HFD1@cornell.edu
ϩ
6
16
0
021-9606/2000/112(10)/4634/10/$17.00
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
32.174.254.155 On: Wed, 24 Dec 2014 15:44:22
4634
© 2000 American Institute of Physics
1

4635
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Collisions with CH and C H
4 2 6
have not been calculated. Insertion into the C–H bond of
C H can lead to dehydrogenation, forming MoC H ϩH , or
2
6
2
4
2
3
1,32
to ethylene elimination producing MoH ϩC H .
2
2
4
Calculations indicate that all stationary points along the
reaction coordinate ͑other than the ground state reactants͒ in
2
2,28
Fig. 1 are of quintet spin multiplicity.
Thus, a spin–flip
7
must occur in order for the Mo( S ) state to react. One of the
3
goals of this work is to experimentally determine whether or
not this spin-forbidden process plays an important role.
Recent studies in our laboratory revealed that the poten-
tial energy barrier for C–H insertion of ground state
FIG. 1. Reaction energetics for MoϩC H . See text for details.
2
6
1
2
2
33
Y(4d 5s ,a D ) atoms into C H is 19.9Ϯ3.0 kcal/mol.
3
/2
2
6
At collision energies above this barrier, production of
YH2ϩC2H4 is dominant, but a second channel leads to for-
mation of YC2H4ϩH2.33 The ground state Y(a DJ)ϩC2H6
ϩ
ϩ
whereas reactions of Sc and V are initiated by C–H
insertion.
2
1
7,19,20
1
2
diabatic surfaces are strongly repulsive due to the d s char-
acter of ground state Y. The large adiabatic barrier for C–H
insertion results from an avoided crossing between these sur-
faces and attractive surfaces, also of doublet multiplicity,
correlating to the ground state C–H insertion intermediate.22
In this case, both the Y atom reactants and the calculated
ground state insertion intermediates are of doublet multiplici-
ties, and translational energy is relatively efficient in promot-
ing reaction.
More recently, gas phase studies have been extended to
reactions of neutral transition metal atoms in flow tube
2
1–24
reactors.
The rates of bimolecular reactions were in-
ferred by monitoring the consumption of the transition metal
2
5,26
atoms as a function of hydrocarbon partial pressure.
For
4
1
4
example, metastable V*(3d 4s ,a D) atoms, prepared by
stimulated emission pumping ͑SEP͒, were found to be deac-
tivated at essentially the gas kinetic limit by collisions with
2
5
ethylene. In experiments using the alkanes CH and C H ,
4
2
6
This study is an expansion of our previous work on the
on the other hand, it was found that deactivation was less
efficient, occurring in only 1 in 75 and 1 in 25 hard sphere
7
collisions of ground state Mo( S ) and electronically excited
3
5
27
2
5
Mo*( S2) atoms with the simplest hydrocarbon, CH4. In
the present work, the signal-to-noise ratio has been improved
through the use of 157 nm photoionization of reactive and
nonreactive products ͑rather than electron impact ionization͒.
collisions, respectively.
The metastable low spin
5
1
5
Mo*(4d 5s ,a S ) state was observed to be deactivated on
essentially every hard sphere collision with CH and C H .
2
2
6
4
2
6
However, because products of the deactivation process were
not detected, the relative importance of chemical reaction vs.
5
We have also carried out studies of Mo*( S )ϩC H reac-
tions as well as nonreactive collisions of Mo( S ) atoms
2
2
6
7
3
electronic quenching was not determined.
5
with CH4 and C2H6. These nonreactive collisions are com-
The chemical products from metastable Mo*( S )
2
7
pared to elastic collisions of Mo( S ) with inert gas atoms
3
ϩCH reactions under single collision conditions were first
4
2
7
such as Ne and Ar.
directly observed in our laboratory. The MoCH dehydro-
2
genation products were detected at collision energies of 7.2
and 13.0 kcal/mol. No dehydrogenation products were ob-
II. EXPERIMENT
5
1
7
served for ground state Mo(4d 5s ,a S )ϩCH at collision
3
4
energies up to 20.2 kcal/mol.27
All experiments reported here were performed using a
rotatable source crossed molecular beams apparatus, which
has been described in detail elsewhere.34 The Mo atoms were
generated by laser ablation ͑532 nm, 20 mJ/pulse͒ from a
0.25 in. diameter molybdenum rod ͑Alfa Aesar, 99%͒ and
subsequently entrained in a supersonic expansion of an inert
carrier gas. The metal beam was skimmed and collimated by
a set of defining slits to an angular full width at half maxi-
mum ͑FWHM͒ of 2.5°. The populations of atomic electronic
states in the Mo beam were characterized by laser induced
fluorescence ͑LIF͒ excitation spectroscopy at the interaction
region. The frequency doubled output of a Nd:YAG ͑Con-
tinuum 9030͒ pumped dye laser ͑Scanmate 2E, Pyridine 1͒
Siegbahn et al. have calculated several minima and
saddle points lying on the potential surfaces for the interac-
2
2
tions of Mo with CH and C H . A schematic reaction
4
2
6
coordinate for the MoϩC H system is shown in Fig. 1. All
2
6
stationary points shown in Fig. 1 are based on ab initio
2
2,28
29
calculations,
thermodynamic data from the literature,
and results from the present experiments. The calculated
C–H and C–C insertion barriers lie 37.8 and 41.2 kcal/mol,
7
respectively, above the Mo( S )ϩC H reactant asymptote
3
2
6
5
͓
7.0 and 10.4 kcal/mol, respectively, above Mo*( S )
2
ϩC H reactants͔, although the C–H insertion barrier is
2
6
3
0
drawn lower in Fig. 1 based on experiments reported here.
The energetics of the HMoC H5 intermediate and
was scanned over a wavelength range containing transitions
2
7
H MoC H resulting from ␤-hydrogen migration were de-
associated with the ground electronic state, Mo( S ), as well
2
2
4
3
2
8
rived from the results of an earlier ab initio study.
as several transitions from the low-lying electronically ex-
5
5
Three possible product channels are shown in Fig. 1 for
cited Mo*( s ) and Mo( D ) states. Fluorescence signal
2 J
the reaction of MoϩC H . The most exoergic channel, based
was collected by optics perpendicular to the atomic and laser
2
6
on ab initio calculations, is formation of MoCH ϩCH fol-
beams and focused onto a photomultiplier tube ͑PMT͒. Fluo-
2
4
lowing C–C bond insertion.22 However, the energetics of the
rescence signal was only observed for transitions out of the
7
reaction intermediates following formation of CH MoCH3
ground Mo( S ) state.
3
3
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.174.254.155 On: Wed, 24 Dec 2014 15:44:22

4636
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Hinrichs et al.
FIG. 2. Optical pumping scheme for production of metastable electronically
5
excited Mo*(a S ) atoms.
2
7
7
FIG. 3. Nonreactively scattered Mo(a S ) TOF spectra for ͑a͒ Mo(a S )
3
3
7
ϩCH4 at
͗
Ecoll
͘
ϭ14.4 kcal/mol and ͑b͒ Mo(a S )ϩNe at
͗
Ecoll
͘
The doubled output of a second Nd:YAG ͑Continuum
030͒ pumped dye laser ͑Scanmate2E or FL3002͒ was used
3
ϭ16.8 kcal/mol. Bottom ͑a͒: Best-fit P(E ) ͑solid line͒ and collision energy
T
9
distribution. Bottom ͑b͒: Collision energy distribution used as P(E ) for
T
in experiments in which it was necessary to transfer popula-
elastic collisions.
5
tion to the metastable excited Mo*( S ) state, lying 10 768
2
Ϫ1
30
cm above the ground state. The laser was expanded us-
ing an optical telescope to ϳ1.2 cm diameter and crossed the
Mo beam 8 mm upstream of the interaction region. As de-
picted in Fig. 2, the Mo atoms are excited on the spin-
III. RESULTS AND ANALYSIS
7
A. Collisions of Mo„ S … with alkanes
3
5
0
7
7
forbidden Mo**( P )� Mo( S ) transition at 345.738 nm.
The collisions of Mo( S )ϩCH were studied at
3
3
3
4
5
0
3
Following excitation, the Mo**( P ) state radiatively de-
cays to the desired Mo*( S ) state, a fully allowed transi-
͗
Ecoll
͘
ϭ4.7 and 14.4 kcal/mol. No reactive signal was ob-
5
7
served for ground state Mo( S ) atoms colliding with CH at
2
3
4
tion, at 550.802 nm with an Einstein A coefficient of 3.0
these collision energies. However, nonreactively scattered
8
Ϫ1 35
5
7
ϫ10 s
.
Although decay to the D state also occurs, as
Mo( S ) atoms were detected at laboratory angles ranging
4
3
discussed in Section IV B, this state is known to be relatively
from 5° to 35°. Time-of-flight ͑TOF͒ spectra taken at
unreactive from previous experiments26 and is only popu-
Mo (m/eϭ98) at
ϩ
Ecoll
ϭ14.4 kcal/mol are shown in Fig.
͗
͘
3
5
lated to a very small extent in these experiments.
3͑a͒. The solid lines represent best-fit simulations, which will
A skimmed molecular beam, directed at right angles to
the metal beam, has an angular FWHM of 5.0°. Scattered
species from collisions at the interaction region are detected
by a fixed triply-differentially pumped mass spectrometer
with an angular resolution of 1.5°. Recoiling neutral frag-
ments are ionized by 157 nm photons generated by an F2
excimer laser ͑Lambda Physik LPX220i͒. Once ionized, the
species are mass selected by a quadrupole mass filter and
detected by a channeltron ion detector. Positive ions are un-
able to enter the detector because the front lenses are biased
at ϩ100 V. Negative ions, if present, strike the ion lens
assembly and are lost. Experiments were also carried out in
which positive metal ions produced in the ablation process
were extracted from the beam in the differential pumping
region, prior to entering the main scattering chamber. No
difference in signal was seen with the extraction field ‘‘on’’
and ‘‘off,’’ indicating that none of the observed signal results
from reactions of ionic species leading to neutral metallic
be discussed below. The TOF spectra recorded at Ecoll
ϭ4.7 kcal/mol were qualitatively similar to the spectra
shown here.
͗
͘
In order to better understand this nonreactive scattering
process, additional experiments were performed using inert
7
gas colliding partners. Elastic scattering of Mo( S )ϩNe
3
7
was studied at
͗
Ecoll
͘
ϭ5.0 and 16.8 kcal/mol, and Mo( S )
3
ϩAr was studied at 21.4 kcal/mol. TOF spectra for
7
Mo( S )ϩNe scattering at 16.8 kcal/mol are shown in Fig.
3
3͑b͒. The solid lines were generated using the forward con-
volution analysis program. The TOF spectra and laboratory
angular distributions are simulated from input center-of-mass
͑c.m.͒ translational, P(E ), and angular, T(⌰), distribu-
T
7
tions, which are assumed to be separable. Since Mo( S )
3
ϩNe collisions are elastic, the P(E ) used to simulate the
T
data, shown at the bottom of Fig. 3͑b͒, has been assumed to
be identical to the collision energy distribution, based on the
measured reactant beam velocities. Figure 4 shows the labo-
9
2
7
products. The isotopic abundances of Mo are Mo:15%,
ratory angular distribution for Mo( S ) elastically scattered
3
9
4
95
96
97
98
Mo:9%, Mo:16%, Mo:17%, Mo:9.5%, Mo:24%,
from Ar. The solid line shown in the lab angular distribution
is a simulation using the collision energy distribution as
and 100Mo:9.5%. All studies described here, unless other-
36
wise noted, were recorded using the most abundant isotope,
Mo. The experimental data were analyzed using a forward
convolution program described in previous publications.37,38
P(E ) and the T(⌰) shown in Fig. 4. This elastic T(⌰)
T
9
8
7
obtained from the Mo( S )ϩAr fits was also used to gener-
3
7
ate the best-fit Mo( S )ϩNe TOFs.
3
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.174.254.155 On: Wed, 24 Dec 2014 15:44:22

4637
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Collisions with CH and C H
4 2 6
7
FIG. 6. Nonreactively scattered Mo( S ) TOF spectra for collisions with
3
͗ ͘
C H at Ecoll ϭ21.0 kcal/mol. Solid line simulations generated using T(⌰)
2
6
in Fig. 4 and P(E ) shown at bottom of this figure ͑solid line͒. Dashed line
T
represents collision energy distribution.
several product channels are energetically accessible at these
collision energies,22 no reactive product signal was observed.
FIG. 4. Laboratory angular distribution ͑error bars are 1 ␴ of 3 replicate
7
measurements͒ for elastic scattering of Mo( S )ϩAr at
͗
Ecoll
͘
7
3
However, nonreactively scattered Mo( S ) atoms were ob-
3
ϭ21.4 kcal/mol. Solid line is best fit generated using c.m. collision energy
served and representative TOFs are shown in Fig. 6 for
distribution, P(E ), and best-fit angular distribution, T(⌰) ͑semilog plot͒.
T
͗
Ecoll
͘
ϭ21.0 kcal/mol. The solid line fits were generated us-
7
ing the elastic T(⌰) from Mo( S )ϩAr and the P(E )
3
T
7
shown at the bottom of Fig. 6. Also shown as a dashed line
in Fig. 6 is the collision energy distribution. On average,
The best-fit simulations for the Mo( S )ϩCH nonreac-
3
4
tive TOF spectra ͓Fig. 3͑a͔͒ were also generated using the
T(⌰) obtained from the elastic collisions. The P(E ) used
for these simulations is shown as a solid line in the bottom
portion of Fig. 3͑a͒ and for comparison the collision energy
distribution is shown as a dashed line. On average,
7
Mo( S )ϩC H collisions were found to be more inelastic
3
2
6
T
7
than Mo( S )ϩCH collisions, with ϳ19% of the initial ki-
3
4
netic energy converted into internal energy of the recoiling
C H molecule.
2
6
7
Mo( S )ϩCH collisions transfer ϳ10% of the initial ki-
netic energy into internal energy of the recoiling CH mol-
3
4
5
B. Collisions of Mo*„ S … with methane, CH
4
2
4
ecule.
The collisions of metastable electronically excited
Mo*( S ) atoms with CH have also been studied at the
7
The collisions of Mo( S )ϩC H were studied at
3
2
6
5
2
4
͗
Ecoll
͘
ϭ21.0 and 35.4 kcal/mol. Figure 5 shows a Newton
diagram in velocity space for collisions of MoϩC H at
Ecoll
͘
ϭ21.0 kcal/mol. Although calculations predict that
collision energies noted in Table I. Two distinct effects are
2
6
observed when the pump laser is turned on to prepare
͗
5
Mo*( S ) upstream of the collision volume. First, reactive
2
signal associated with dehydrogenation products,
9
8
MoCH ϩH , is clearly observed at all collision energies
2
2
studied. As expected, when the pump laser is tuned slightly
5
0
7
off the ( P � S ) resonance, the MoCH signal disappears.
3
3
2
We have previously reported the observation of this reactive
signal at Ecoll
ϭ7.2 and 13.0 kcal/mol.27 In our previous
͗
͘
study, the MoCH products were ionized for mass analysis
2
using electron impact ionization. In this current study, we
have revisited this system using 157 nm photoionization of
the MoCH products and nonreactively scattered Mo atoms.
2
The increased sensitivity of photoionization compared to
electron impact ionization allows us to obtain improved
signal-to-noise ratios. This is particularly valuable at lower
collision energies where beam intensities are reduced. The
͗ ͘
TOF spectra for the MoCH products at Ecoll ϭ4.7 kcal/
2
mol are shown in Fig. 7. The solid lines represent the best-fit
FIG. 5. Newton diagram for collisions of MoϩC H at
͗ ͘
Ecoll
2
6
simulations generated using the P(E ) and T(⌰) shown in
T
ϭ21.0 kcal/mol. All product Newton spheres are based on calculated ener-
5
Fig. 8. Dashed lines shown in the P(E ) represent the limit
getics. Note that reactive energetics are relative to Mo*( S ) state and non-
T
2
reactive sphere is for elastic ͑i.e., not super-elastic͒ scattering.
of kinetic energy release distributions which satisfactorily fit
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
32.174.254.155 On: Wed, 24 Dec 2014 15:44:22
1

4638
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Hinrichs et al.
TABLE I. Experimental conditions for Moϩalkane collisions.
Molybdenum beam
Secondary beam
a
b
b
͗
Ecoll
͘
Cond
vpk
FWHM
Cond
vpk
FWHM
MoϩCH4
MoϩCD4
2.1
.7
Ar
Ne
He
Ne
Ar
Ne
He
H2
Ne
He
He
905
1284
2426
1270
905
1316
2478
3244
1270
2450
2306
95
130
298
130
95
156
294
444
130
260
254
25% in Ar
Pure
20% in H2
Pure
Pure
Pure
20% in H2
10% in H2
Pure
20% in H2
20% in H2
696
1110
1762
980
888
888
1404
1628
980
1628
1064
118
230
190
300
176
176
130
144
200
144
104
4
1
4.4
5.0
4.5
MoϩC H
2
6
6.8
2
3
1.0
5.4
5.0
6.8
MoϩNe
1
MoϩAr
21.4
a
͗
Ecoll
͘
in kcal/mol.
b
Velocities in m/s.
the laboratory angular distribution ͑not shown͒. The shaded
portion on the T(⌰) represents our uncertainty in the degree
of forward–backward symmetry. It has been determined that
T(⌰ϭ0°)/T(⌰ϭ180°)ϭ0.97Ϯ0.12, implying an essen-
tially forward–backward symmetric T(⌰). Qualitatively
similar c.m. distributions have been obtained at the higher
␪ϭ15° in Fig. 10. The open circles show the data with the
pump laser off while the filled circles show the result of
turning the pump laser on. The integrated product signal in-
tensity decreases by approximately 35% while the shape of
each TOF spectrum remains unchanged. Similar behavior is
observed at all other laboratory angles.
collision energies studied. The MoCH signal has also been
2
observed at the center-of-mass angle, ␪cm , for
͗
Ecoll
͘
5
C. Collisions of Mo „ S … with ethane, C H
*
2
2
6
ϭ2.1 kcal/mol; however, a complete data set was not re-
corded due to the extremely weak signal resulting from the
low Mo beam intensity using Ar carrier gas.
5
For the case of Mo*( S )ϩC H , the MoC H dehydro-
genation products were observed at all collision energies
studied, ranging from 4.5 to 35.4 kcal/mol. The laboratory
angular distribution for MoC H4 products at
ϭ21.0 kcal/mol is shown in Fig. 11. Due to the light mass of
the recoiling H counterfragment, the laboratory angular dis-
tribution is peaked at ␪cm and falls off sharply on either side.
The solid line simulation was generated using the best-fit
c.m. distributions also shown in Fig. 11. These c.m. distribu-
tions are very similar to those obtained for the Mo*( S )
2
2
6
2
4
We have also studied this reaction using deuterated
2
Ecoll
͗ ͘
methane, CD , at
͗
Ecoll
͘
ϭ5.0 kcal/mol. The laboratory an-
4
gular distribution for the MoCD products is shown in Fig. 9.
2
The c.m. distributions used to simulate these data are shown
at the bottom of Fig. 9. Dashed lines in the lab angular dis-
tribution have been generated using the best-fit T(⌰) and the
dashed line P(E )s. The heavier recoil partner (D instead
2
T
2
5
of H ) increases our sensitivity to the exact form of the
2
2
ϩCH products with a P(E ) peaking very close to zero
kinetic energy. At ϭ21.0 kcal/mol, the P(E ) peaks at
͗ ͘
Ecoll
T
.5 kcal/mol, whereas at the lower collision energy of 6.8
T(⌰). The shaded region indicates our uncertainty in the
depth of the T(⌰), i.e., comparing T(⌰ϭ0°) to T(⌰
ϭ90°). For the best-fit distribution T(⌰ϭ0°)/T(⌰ϭ90°)
ϭ2.2, while a range of Ϯ0.8 still generates acceptable fits.
The second effect observed when the pump laser is
turned on is illustrated in the 98Mo TOF spectrum recorded at
4
T
2
FIG. 8. Best-fit c.m. distributions ͑solid lines͒, P(E ), and T(⌰), for
T
5
Mo*( S )ϩCH →MoCH ϩH at
͗ ͘
Ecoll ϭ4.7 kcal/mol. Dashed lines
shown in the P(E ) correspond to limits of acceptable fits. Shaded region
2
4
2
2
5
FIG. 7. TOF spectra for MoCH products from Mo*( S )ϩCH collisions
2
2
4
T
at
͗
Ecoll
͘
ϭ4.7 kcal/mol.
shows uncertainty in forward–backward symmetry in T(⌰).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
32.174.254.155 On: Wed, 24 Dec 2014 15:44:22
1

4639
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Collisions with CH and C H
4 2 6
FIG. 9. Lab angular distribution and best fit c.m. distributions for the reac-
FIG. 11. Lab angular distribution and best fit c.m. distributions for the
5
tion Mo*( S )ϩCD →MoCD ϩD at
͗
Ecoll
͘
ϭ5.0 kcal/mol. Dashed lines
5
2
4
2
2
reaction Mo*( S )ϩC H →MoC H ϩH at
͗ ͘
Ecoll ϭ21.0 kcal/mol. Dashed
lines shown in the P(E ) correspond to limits of acceptable fits. Shaded
2
2
6
2
4
2
shown in the P(E ) correspond to limits of acceptable fits. Shaded region
T
T
shows uncertainty in depth of T(⌰).
region shows uncertainty in forward–backward symmetry in T(⌰).
kcal/mol, the peak is near 1.3 kcal/mol. The T(⌰) was found
to be essentially forward–backward symmetric, with T(⌰
ϭ0°)/T(⌰ϭ180°)ϭ1.14Ϯ0.21.
identical for both cases͒, and the reduced mass of the product
fragments.
9
8
Observation of MoH signal is complicated by the
2
presence of 1 Mo͑ Mo:24.1%, Mo:9.6%͒. After an ex-
tensive search at m/eϭ98, 100, and 102, at a number of
laboratory angles and collision energies, we conclude that no
00
98
100
Shown in Fig. 12 are the MoC H TOF spectra taken at
2
4
the center-of-mass angles for
ϭ10°) with the pump laser off and
kcal/mol(␪ ϭ15°) with the pump laser turned on. The ex-
͗
Ecoll
͘
ϭ35.4 kcal/mol(␪cm
͗
Ecoll ϭ4.5
͘
cm
MoH signal is observed above the background signal. As
2
5
Ϫ1
cited Mo*( S ) state lies 10 768 cm ͑30.8 kcal/mol͒ above
indicated in Fig. 1, a third possible product channel may
result from initial C–C bond insertion forming CH MoCH ,
2
7
30
the Mo( S ) ground state. Thus the total energy (Ekinetic
3
3
3
ϩEelectronic) is approximately equal for these two experi-
ments. The top portion of Fig. 12 shows the two collision
followed by production of MoCH ϩCH . The Newton dia-
gram ͑Fig. 5͒ for Mo*( S )ϩC H collisions at
Ecoll
͗ ͘
2 2 6
2 4
5
͗ ͘
energy distributions with the Ecoll ϭ4.5 kcal/mol shifted by
5
the electronic energy of the Mo*( S ) state. No signal was
2
observed when the total available energy was in translation,
whereas reactive signal could be easily observed for elec-
5
tronically excited Mo*( S ). It should be noted that the
2
maximum accessible Newton spheres for recoiling MoC H
2
4
products are similar for both conditions since they depend
mainly on the total available energy ͑which is essentially
FIG. 12. Top: Collision energy distributions for MoϩC H for ͑a͒ ground
2
6
7
state Mo( S )ϩC H
6
at
͗
Ecoll
͘
ϭ35.4 kcal/mol, and ͑b͒
͗ ͘
Ecoll
3
2
ϭ4.5 kcal/mol distribution shifted by the excitation energy to the
5
FIG. 10. TOF spectra for nonreactively scattered Mo atoms from CH re-
corded at ␪ϭ15°. The total signal intensity is decreased by ϳ35% when the
pump laser is turned on.
Mo*(a S2) state. Bottom: MoC2H4 TOF spectra recorded at the angle of
4
7
the center-of-mass, ␪cm , for ͑a͒ Mo( S3)ϩC2H6 at
and ͑b͒ Mo*( S )ϩC H at
͗
Ecoll
͘
ϭ35.4 kcal/mol,
5
͗ ͘
Ecoll ϭ4.5 kcal/mol.
2 2 6
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
32.174.254.155 On: Wed, 24 Dec 2014 15:44:22
1

4640
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Hinrichs et al.
ergy, the total energy available to promote the observed de-
hydrogenation reaction is the sum of the translational and
electronic energy, 32.9 kcal/mol. This sets a strict upper limit
ϭ21.0 kcal/mol demonstrates that the maximum Newton
spheres for MoCH products ͑as well as MoH products͒ are
2
2
larger than for MoC H . This provides the opportunity to
2
4
7
search for these products outside the range where fragmen-
tation of MoC H might contribute to the background signal.
of 32.9 kcal/mol relative to ground state Mo( S3) ͑i.e., 2.1
5
kcal/mol above Mo
*
( S )) for all barriers along the quintet
2
4
2
However, even after extensive averaging at m/e
reaction coordinate to form MoCH2ϩH2.
ϩ
ϭ112(MoCH ) at a number of laboratory angles at several
A previous theoretical study has calculated the transition
state for C–H insertion to lie on the quintet PES 37.8 kcal/
2
different collision energies, no evidence for this channel was
5
22
observed from Mo*( S )ϩC H .
mol above the ground state reactants. This is 7.0 kcal/mol
2
2
6
5
above the energy of separated Mo*( S )ϩCH , suggesting
2
4
IV. DISCUSSION
that an adiabatic insertion barrier may exist for the electroni-
5
cally excited Mo*( S ) state. From our experimental obser-
7
2
A. Collisions of Mo„ S …¿CH ,C H
3
4
2
6
5
vations, if such a barrier above Mo*( S )ϩCH exits, it
2
4
Distinct similarities are observed in the TOF spectra and
must be less than 2.1 kcal/mol above the quintet reactant
7
T(⌰) for collisions of Mo( S ) with CH and Ne. In this
asymptote. Although the MoϩCH calculations did not ad-
3
4
4
comparison, it is informative to consider an analogous pair
of systems, ArϩNe and ArϩCH . For both elastic scattering
of ArϩNe ͑Ref. 39͒ and inelastic scattering of ArϩCH4,
rainbow and supernumerary rainbow scattering has been ob-
served in crossed beams studies. This has allowed for the
dress the origin of this barrier, Li and Balasubramanian have
carried out detailed calculations on the insertion of various
4
4
0
3
1
electronic states of Mo into H . It was found that ground
2
7
state Mo( S ) atoms have to surmount an 89.5 kcal/mol bar-
3
rier for insertion. A barrier of 28 kcal/mol was calculated for
5
31
determination of well depths for the interactions of these two
insertion of the electronically excited Mo*( S ) state.
2
Ϫ1
5
systems. ArϩNe is bound by ϳ50 cm
whereas ArϩCH4
However, the Mo(b D) state, lying 57 kcal/mol above the
Ϫ1 39,40
has a well depth of ϳ120 cm
.
In addition to the
ground state, was calculated to insert spontaneously into H
2
5
31
slightly deeper well, ArϩCH collisions are inelastic and
4
forming ground state MoH ( B ). This suggests that the
2 2
5
modeling the potential energy surface ͑PES͒ requires the in-
adiabatic barrier for insertion of Mo*( S ) is associated with
2
clusion of an anisotropic term related to the ability of Ar to
an avoided crossing of diabatic curves correlating to
4
0
impose a torque on CH during the scattering process.
5
5
4
Mo*( S ) and Mo(b D).
2
7
7
In our studies of Mo( S )ϩNe/Ar and Mo( S )
3
3
Weisshaar et al. have measured the rate for depletion of
ϩCH /C H , no evidence for rainbow scattering was ob-
5
4
2
6
Mo*( S ) atoms in the presence of CH at room temperature
2
4
served due to the high collision energies and lower experi-
mental angular resolution compared to the Ar scattering
Ϫ10
3
Ϫ1 26
to be k ϭ(3–4)ϫ10
near the gas kinetic limit for collisions of Mo*( S ) with
cm •s
.
This depletion rate is
1
5
2
3
9,40
studies.
However, the use of the same T(⌰) for inert gas
CH . If the depletion is due to reactivity rather than elec-
4
and alkane scattering implies that the dynamics of the scat-
tronic quenching, then the true insertion barrier must lie at
tering events are qualitatively similar. The ground state
5
most a fraction of a kcal/mol above the Mo*( S )ϩCH
2
4
7
Mo( S )ϩalkane collisions are dominated by the repulsive
26
5
3
reactants. The depletion of the metastable Mo(a D ) was
J
part of the PES with negligible contribution from complex
formation. This situation contrasts that seen recently in our
studies of second row transition metal atoms with unsatur-
ated hydrocarbons such as alkenes and alkynes.37,38 In those
cases, the strong peaking of the T(⌰) near ⌰ϭ180° and the
subsequently determined forward–backward symmetry for
the nonreactively scattered ground state atoms was taken as
evidence for the formation of collision complexes having
also measured as 5–10 times less efficient compared to the
5
26
Mo*( S ) state. If we assume that the difference in deple-
2
tion rates for these two metastable states is proportional to
their reactivity, using the known branching fraction for decay
5
35
into the metastable (a D ) state, the maximum contribu-
J
5
tion from reactions of the Mo(a D ) states to form MoCH
J
2
products in our experiment is 4%.
We expect that elimination of molecular hydrogen from
lifetimes 7,e38xceeding
their
Metal–alkene
picosecond
and
rotational
metal–alkyne
27,42
the HMoCH intermediate involves a stepwise process.
3
3
timescales.
Following insertion, this process involves migration of an
␲
-complexes are typically bound by 30–50 kcal/mol relative
␣
-hydrogen to form a second stable intermediate, H MoCH ,
2
2
7
2 2
to the separated reactants. In the case of Mo( S )ϩCH
and Mo( S )ϩC H , in the absence of C–H insertion, the
3
4
which subsequently decays to products. The best-fit P(ET)
7
3
2
6
for the 4.7 kcal/mol data set peaks near E ϭ1 kcal/mol for a
T
attractive interaction is expected to be comparable in magni-
tude to that with inert gas atoms, having binding energies
with transition metal atoms typically in the range of 100–350
22
total available energy of 10 kcal/mol. Thus, as the
H MoCH intermediate eliminates H , most of the available
2
2
2
energy is channeled into vibrational modes perpendicular to
the reaction coordinate. Due to the small reduced mass of the
recoiling products, only a relatively small fraction of the to-
tal angular momentum, J, can become relative orbital angu-
lar momentum in the products, LЈ. Thus, conservation of
angular momentum may channel a significant amount of ro-
tational energy of the complex into product rotation, leading
to an angular distribution peaking less sharply at the poles
than in our recent studies of reactions involving elimination
Ϫ1
41
cm ͑0.3–1.0 kcal/mol͒.
5
B. Reactivity of Mo*„ S … state
2
5
Excited state Mo*( S ) atoms were found to react effi-
2
ciently with CH and C H to produce MoCH ϩH and
4
2
6
2
2
MoC H ϩH , respectively. We have observed these dehy-
2
4
2
drogenation products at all collision energies studied, down
5
to 2.1 kcal/mol for Mo*( S )ϩCH . At this collision en-
2
4
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.174.254.155 On: Wed, 24 Dec 2014 15:44:22

4641
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Collisions with CH and C H
4 2 6
of ethylene33 and carbon monoxide43 from long-lived inter-
mediates.
ization continuum for Mo.30 Consequently, 1ϩ1 resonance
enhanced multiphoton ionization ͑REMPI͒ prior to collision
was also found to contribute to the decreased wide-angle
nonreactive scattering of Mo atoms, because positive ions
cannot enter the detector. At low laser powers ͑0.1 mJ, 1.2
cm dia.͒, this process could be minimized, but not com-
5
In the case of Mo*( S )ϩC H reactions, dehydrogena-
2
2
6
tion products were observed at the lowest collision energy
studied, Ecoll ϭ4.5 kcal/mol, placing an upper limit of 35.3
kcal/mol above Mo( S )ϩC H for any barrier along the
͗
͘
7
3
2
6
reaction coordinate. To date, it has been assumed that the
C–H activation barrier for C H is energetically similar to
pletely eliminated, so it is not possible to accurately deter-
5
mine the exact fraction of excited state Mo*( S ) undergo-
2
6
2
that for CH activation. Indeed, Weisshaar et al. have found
ing reaction. However, we reiterate that experiments using
strong electric fields at the interaction region had no effect
on the measured TOF spectra recorded for MoCH2,
MoC H , or Mo from nonreactive elastic and inelastic colli-
4
5
the depletion rate of Mo*( S ) in the presence of C H to be
2
2
6
2
6
very similar to that measured for CH4.
The calculated height of the ␤-hydrogen migration tran-
2
4
5
sition state for Mo*( S )ϩC H is found to lie energetically
sions. This indicates that the observed signals result exclu-
sively from collisions of neutral transition metal atoms, with
2
2
6
2
2,28
below the C–H insertion barrier.
For transition metals
ϩ
containing unoccupied d-orbitals, these ␤-migration transi-
tion state structures may be further stabilized by intramo-
lecular ␤-agostic interactions.44 It is possible that at higher
levels of theory these ␤-agostic interactions may become
no contributions from events involving Mo .
D. Comparison of ground and excited state reactivity
evident for the MoϩC H6 system, further lowering the
height of this migration barrier.
2
5
The Mo*( S ) state reacts efficiently with methane and
2
7
ethane, while the ground Mo( S ) state has been found to be
3
unreactive at all collision energies studied. Since both elec-
5
6
1
tronic states considered have a 4d (a S)5s configuration,
the only difference is in the alignment of the spin of the 5s1
electron. It is well known that low-spin states insert more
readily into H–H, C–H, and C–C bonds.2 The availability
of spin-paired electrons facilitates s–d hybridization and
formation of two covalent bonds in the insertion intermedi-
ate.
C. Effect on nonreactive scattering by laser excitation
The intensity of the nonreactively scattered Mo atoms is
2,31
found to decrease by ϳ 35% when optical pumping to popu-
5
late the Mo*( S ) state is performed upstream of the inter-
2
action volume. However, the shapes of the TOF spectra re-
main unchanged. The absence of an increase in nonreactively
scattered Mo atoms arriving at short flight times when the
laser is turned on indicates that ‘‘superelastic’’ scattering in-
In the C–H activation of CH by Mo, Siegbahn et al.
4
have found that the insertion transition state and the ground
2
2
volving transfer of electronic energy to translational energy
state of HMoCH have quintet spin multiplicity. Similar
3
7
3
1
is negligible. As noted above, ground state Mo( S ) atoms
3
results have also been calculated for the MoϩH system.
Therefore, Mo( S ) encounters a large barrier to C–H bond
2
7
do not insert into the C–H bond of CH and C H . The fact
4
2
6
3
that no increase in wide angle nonreactive signal peaking
near ⌰ϭ180° is seen when the laser is turned on indicates
that insertion of excited state atoms followed by decay back
to reactants must be a negligible process. If we assume that
the densities of states are comparable for both the C–H in-
sertion and the ␣-hydrogen migration transition states, then
the relative heights of these barriers determine the fraction of
intermediates that will reform reactants. Since we see no
evidence for decay of insertion complexes back to reactants,
we believe that the C–H insertion barrier is higher than any
subsequent barriers for reaction. Vibrational frequencies and
moments of inertia, not presently available, are necessary to
better quantify the relative energies of these two barriers.
The primary origin of the decrease in wide angle nonre-
insertion associated with promotion to the low-spin quintet
surface. Such barriers may, in theory, be overcome by colli-
sion energy if the coupling between the surfaces is suffi-
ciently strong. Since electronic energy has been observed to
be efficient in promoting reaction, whereas collision energy
is inefficient, we conclude that the crossing from the ground
state septet surface onto the quintet surface is inefficient.
Spin–orbit coupling has been found to play an important
role in a number of reactions involving transition metals. For
3
NiϩCO, ground state Ni( F ) atoms were found to associate
J
with CO with a relatively large third-order rate constant of
Ϫ32
6
Ϫ1 45
kϭ(6.9Ϯ0.5)ϫ10
cm s
.
However, the ground state
1
ϩ
1
of NiCO is a X ⌺ state, which correlates to Ni( D ) lying
2
Ϫ1
3
410 cm above the ground state. Strong singlet–triplet
active Mo signal upon turning on the laser is that a signifi-
coupling must lead to an adiabatic correlation between the
ground state singlet NiCO complexes and the ground state
5
cant fraction of the excited Mo*( S ) atoms undergo reac-
2
3
1
ϩ
tion producing MoCH ϩH . Such atoms are no longer
2
2
Ni( F )ϩCO( ⌺ ) reactants in order to explain the ob-
J
available to undergo inelastic nonreactive scattering, de-
creasing the signal from this channel. By comparing the frac-
tional decrease in wide-angle nonreactive scattering ͑typi-
served association rate constant. Mitchell has carried out cal-
culations to estimate the magnitude of the spin–orbit cou-
3
1
ϩ
46
pling between the c ⌸ and X ⌺ surfaces for NiϩCO.
The crossing probability, P, for a single pass through the
cally 35%͒ to the fractional decrease in the population of
7
3
1
ϩ
ground state Mo( S ) atoms in the beam ͑using LIF͒, we
3
c ⌸ to X ⌺ surface intersection was approximated by the
Landau–Zener ͑LZ͒ equation ͑in the diatomic assumption͒
attempted to determine the fraction of excited state atoms
undergoing reaction. Unfortunately, this measurement is
2
2
5
0
Ϫ4•␲ •M
complicated by the fact that the z P3 state used in our
pumping scheme lies slightly more than halfway to the ion-
PϷ1Ϫexp
ͫ
ͬ
.
͑1͒
h•F•v
x
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.174.254.155 On: Wed, 24 Dec 2014 15:44:22

4642
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Hinrichs et al.
In Eq. ͑1͒, M is the off-diagonal matrix element of the spin–
orbit coupling operator, h is Planck’s constant, F is the dif-
quintet surfaces, the large difference in the potential gradi-
ents, and high collision energy, combined with the direct
ference in gradients of the diabatic potential curves, and v is
the relative velocity of the colliding partners. Approximating
nature of the insertion reaction, all contribute to the observed
x
7
nonreactivity of the high-spin ground state of Mo( S ).
3
M as the spin–orbit coupling constant for the 3d electrons of
3
1
ϩ
Ni yielded a value for P( ⌸→ ⌺ )Ͼ0.99 for a single pass
through the curve crossing.46 Even if the crossing is much
E. Absence of MoCH and MoH channels
2
2
5
less efficient than calculated, there still exists the possibility
for multiple passes through the crossing region since both the
As indicated in Fig. 1, the Mo*( S )ϩC H reaction
2 2 6
may also lead to production of MoH ϩC H . In a recent
2
2
4
3
1
ϩ
33
⌸
and the ⌺ states of NiCO are bound. Thus, association
study of the reaction of ground state YϩC H , both YH2
2 6
3
1
ϩ
1
ϩ
of Ni( F ) with CO( ⌺ ) to form NiCO(X ⌺ ) was found
and YC H products were clearly observed from decay of
2 4
J
to follow an adiabatic pathway.
H YC H intermediates. In that case, formation of
2 2 4
Another case where spin–orbit coupling appears to play
YH ϩC H was energetically only a few kcal/mol less fa-
2 2 4
3
7
an important role is in NbϩC H →NbC H ϩH . The
vorable than formation of YC H ϩH . Interestingly, the less
2
4
2
2
2
2 4 2
ground state Nb atoms have a sextet spin multiplicity while
C–H activation involves a quartet surface.22 Deep potential
wells associated with NbC H association complexes exists
thermodynamically favored YH ϩC H channel was actu-
2 2 4
ally found to be dominant, with the yield of YH greater than
2
YC H by more than an order-of-magnitude. This indicated
2
4
2
4
4
7
on both the sextet and quartet surfaces. It was found that
Nb is efficiently able to access the quartet surface and pro-
ceed readily to form dehydrogenation products. Again, due
to the deep wells, even if the crossing probability is low, the
strongly bound long-lived high-spin complex allows for mul-
tiple crossings. This contrasts the MoϩCH and MoϩC H
the existence of a potential energy barrier for H elimination
2
from H YC H in excess of its thermodynamic threshold.
2
2
4
The barrier for H elimination is at least in part associated
2
2
33
with the closed-shell s character of Y in YC H . For
2
4
MoϩC H , on the other hand, the MoC H ϩH channel is
2
6
2
4
2
favored by ϳ 15 kcal/mol relative to MoH ϩC H . Further-
4
2
6
2
2
4
1
reactions, where the colliding partners only have one oppor-
tunity to cross onto the low-spin surface if insertion is to
occur.
more, the Mo atom in MoC H has substantial s character
2 4
and will therefore not interact repulsively with the dissociat-
ing H2 fragment.47 Consequently, elimination of H2 from
H2MoC2H4 is expected to proceed through an essentially bar-
rierless transition state. This is consistent with the P(E) dis-
tributions which were always found to peak very close to
In NiϩCO and NbϩC H , the high- and low-spin PESs
2
4
are both strongly attractive, leading to similar gradients and a
relatively small value of F, contributing to a large Landau–
Zener crossing probability. Again, this contrasts the situation
for MoϩCH and MoϩC H . The crossings of the septet and
zero kinetic energy. Thus, it is not surprising that the
5
Mo
tion.
The most thermodynamically favored products,
*
( S )ϩC H reaction leads exclusively to H elimina-
4
2
6
2
2
6
2
quintet PESs for these systems are qualitatively analogous to
7
that for MoϩH , where the A surface is highly repulsive in
2
1
5
5
MoCH2ϩCH4, were not observed from Mo
even at the highest collision energy. From our studies of
*
( S )ϩC H
the insertion ͑bending͒ coordinate, while the B surface is
2 2 6
2
3
1
highly attractive. This leads to a large difference in the
gradients for the two surfaces, thereby diminishing P. Addi-
tionally, it was found that spin–orbit coupling is not signifi-
5
Mo*( S )ϩCH we know that MoCH is readily detectable
2
4
2
using 157 nm photoionization. Since the MoCH ϩCH chan-
2
4
nel is believed to result following C–C insertion, we con-
cant for MoϩH , resulting in a poor crossing efficiency in
2
5
3
1
clude that C–C insertion is negligible for Mo*( S ) over the
the LZ model. It is interesting to note that in the analogous
reaction involving W ͑lying directly below Mo in the peri-
odic table͒, it was concluded that spin-orbit interactions were
likely to be much more important.4
2
range of collision energies studied in our experiment. Be-
cause the barrier height for C–C insertion into C H lies
2
6
considerably higher than that for C–H insertion,22 and due to
the less favorable steric factors for this process, it is perhaps
not surprising that this channel is not observed. Interestingly,
calculations indicate that in the Moϩc-C H reaction, the
8
7
Finally, for collisions of Mo( S ) with C H , a high
3
2
6
collision energy is required to overcome the insertion barrier.
However, a large value for v in Eq. ͑1͒ will diminish the
3
6
x
barrier for C–C insertion is substantially lower than that for
C–H insertion, offering the intriguing possibility that this
process may actually be dominant at low collision energies
crossing probability. Armentrout et al. have studied the reac-
ϩ
tions of Fe ϩpropane͑C H ͒ as a function of collision
3
8
1
6
energy. It was found that at low collision energies, below
Ecollϳ0.5 eV, the ground state Fe ( D) readily undergoes
5
22
ϩ 6
for Mo*( S ) reactions.
2
reaction, indicating that adiabatic crossing to a low-spin
quartet surface is efficient at low collision energies. How-
ever, as the collision energy was increased, the dynamics
were found to increasingly follow the diabatic pathway.16
It should be noted that the LZ model is based on a
simple diatomic system with the crossing of only two PESs.
For more complex systems such as MoϩC H , multiple
V. CONCLUSIONS
7
The interactions of ground state Mo( S ) and metastable
3
5
Mo*( S ) atoms with CH and C H have been studied un-
2
4
2
6
5
der single collision conditions. In the case of Mo*( S )
2
ϩCH /C H , dehydrogenation products (MoCH /MoC H )
2
6
4
2
6
2
2
4
PESs as well as a large number of internal degrees of free-
dom complicate this simple model. In any case, it is clear
that the weak spin–orbit coupling between the septet and
were the only observed products at every collision energy
5
studied. For Mo*( S )ϩCH , an upper limit of 2.1 kcal/mol
2
4
7
͑32.9 kcal/mol relative to Mo( S )ϩCH )͒ is placed on any
3
4
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.174.254.155 On: Wed, 24 Dec 2014 15:44:22

4643
J. Chem. Phys., Vol. 112, No. 10, 8 March 2000
Collisions with CH and C H
4 2 6
1
6
barriers along the reaction coordinate. Ab initio calculations
have shown that all intermediates and barriers along the re-
action coordinate to form dehydrogenation products are of
R. H. Schultz, J. L. Elkind, and P. B. Armentrout, J. Am. Chem. Soc. 110,
11 ͑1988͒.
M. A. Tolbert and J. L. Beauchamp, J. Am. Chem. Soc. 106, 8117 ͑1984͒.
J. L. Elkind and P. B. Armentrout, J. Phys. Chem. 91, 2037 ͑1987͒.
L. Sunderlin, N. Aristov, and P. B. Armentrout, J. Am. Chem. Soc. 109,
78 ͑1987͒.
N. Aristov and P. B. Armentrout, J. Am. Chem. Soc. 108, 1806 ͑1986͒.
J. J. Carroll, K. L. Haug, and J. C. Weisshaar, J. Am. Chem. Soc. 115,
6962 ͑1993͒.
J. J. Carroll, K. L. Haug, J. C. Weisshaar, M. R. A. Blomberg, P. E. M.
Siegbahn, and M. Svensson, J. Phys. Chem. 99, 13955 ͑1995͒.
K. Honma, Phys. Chem. Chem. Phys. 1, 3235 ͑1999͒.
L. Lian, S. A. Mitchell, and D. M. Rayner, J. Phys. Chem. 98, 11637
4
1
1
7
8
quintet spin multiplicity, consistent with the observed reac-
19
5
tivity of Mo*( S ). The P(E )s used to simulate the TOF
2
T
2
2
0
1
spectra indicate that most of the available energy is chan-
neled into product internal energy.
7
Nonreactive scattering of Mo( S ) was studied at a
3
22
range in collision energies up to 14.4 kcal/mol for CH and
4
3
5.4 kcal/mol for C H . The c.m. angular distribution found
23
2 6
2
2
4
5
to best-fit the observed signal was identical to the T(⌰) used
7
͑
1994͒.
Y. Wen, A. Yethiraj, and J. C. Weisshaar, J. Chem. Phys. 106, 5509
1997͒.
to fit Mo( S )ϩAr elastic scattering. These inelastic colli-
3
sions were found to transfer 10% and 19% of the initial
kinetic energy into internal energy of CH and C H , respec-
͑
26
27
4
2
6
Y. Wen, J. Rau, M. Porembski, and J. C. Weisshaar, Abstract in Dynamics
of Molecular Collisions Conference XVI, Brainerd, Minnesota ͑1997͒.
P. A. Willis, H. U. Stauffer, R. Z. Hinrichs, and H. F. Davis, J. Chem.
Phys. 108, 2665 ͑1998͒.
tively. Even at collision energies above the largest potential
7
energy barrier for reaction with C H , ground state Mo( S )
2
6
3
atoms were unreactive, indicating that translational energy is
ineffective in promoting C–H bond activation.
28
P. E. M. Siegbahn, J. Am. Chem. Soc. 115, 5803 ͑1993͒. These calcula-
tions were performed at a different level of theory compared to Ref. 22.
The values were adjusted by matching the energetics of the Mo͑H͒ C H
4
2
2
ACKNOWLEDGMENTS
complex.
J. Berkowitz, G. B. Ellison, and D. Gutman, J. Phys. Chem. 98, 2744
1994͒.
C. E. Moore, Atomic Energy Levels ͑National Bureau of Standards, Wash-
ington, D.C., 1952͒.
J. Li and K. Balasubramanian, J. Phys. Chem. 94, 545 ͑1990͒.
P. E. M. Siegbahn, Theor. Chim. Acta 87, 441 ͑1994͒.
H. U. Stauffer, R. Z. Hinrichs, J. J. Schroden, and H. F. Davis, J. Phys.
Chem. A. ͑in press͒.
P. A. Willis, H. U. Stauffer, R. Z. Hinrichs, and H. F. Davis, Rev. Sci.
Instrum. 70, 2606 ͑1999͒.
2
9
This work was supported by the National Science Foun-
dation. Some of the equipment used in this study was funded
by an NSF Faculty Early Career Development Award, an
NSF equipment grant, and by an ONR Young Investigator
Award. Additional support from the Exxon Educational
Foundation is also gratefully appreciated. R. Z. H. would like
to thank the Department of Education for a Graduate Fellow-
ship. H.U.S. thanks the Link Foundation and Procter &
Gamble for Graduate Fellowships. The authors would like to
thank J. Weisshaar for valuable discussions and suggestions.
͑
30
3
3
3
1
2
3
34
3
3
3
5
6
7
C. H. Corliss and W. R. Bozman, Experimental Transition Probabilities
͑
National Bureau of Standards, Washington, D.C., 1962͒.
C.R.C. Handbook of Chemistry and Physics, 75th ed., edited by D. R. Lide
CRC, Boca Raton, FL, 1994͒.
͑
1
R. H. Crabtree, The Organometallic Chemistry of the Transition Metals
P. A. Willis, H. U. Stauffer, R. Z. Hinrichs, and H. F. Davis, J. Phys.
Chem. A 103, 3706 ͑1999͒.
H. U. Stauffer, R. Z. Hinrichs, P. A. Willis, and H. F. Davis, J. Chem.
Phys. 111, 4101 ͑1999͒.
J. M. Parson, T. P. Schafer, F. P. Tully, P. E. Siska, Y. C. Wong, and Y.
T. Lee, J. Chem. Phys. 53, 2123 ͑1970͒.
U. Buck, J. Schleusener, D. J. Malik, and D. Secrest, J. Chem. Phys. 74,
͑
Wiley, New York, 1994͒.
2
3
4
5
M. J. Burk and R. H. Crabtree, J. Am. Chem. Soc. 109, 8025 ͑1987͒.
A. H. Janowicz and R. G. Bergman, J. Am. Chem. Soc. 104, 352 ͑1982͒.
R. G. Bergman, Science 223, 902 ͑1984͒.
38
39
40
41
E. P. Wasserman, C. B. Moore, and R. G. Bergman, Science 255, 315
͑1992͒.
6
A. A. Bengali, R. G. Bergman, and C. B. Moore, J. Am. Chem. Soc. 117,
879 ͑1995͒.
P. E. M. Siegbahn, J. Am. Chem. Soc. 118, 1487 ͑1996͒.
W. E. Billups, M. M. Konarski, R. H. Hauge, and J. L. Margrave, J. Am.
Chem. Soc. 102, 7393 ͑1980͒.
1707 ͑1981͒.
3
I. Wallace and W. H. Breckenridge, J. Chem. Phys. 97, 2318 ͑1992͒.
P. A. Willis, Ph.D. Thesis, Cornell University, 1999.
H. U. Stauffer, R. Z. Hinrichs, J. J. Schroden, and H. F. Davis, J. Chem.
Phys. 111, 10758 ͑1999͒.
N. Koga, S. Obara, K. Kitaura, and K. Morokuma, J. Am. Chem. Soc.
107, 7109 ͑1985͒.
7
8
4
4
2
3
9
G. A. Ozin, D. F. McIntosh, S. A. Mitchell, and J. Garc ´ı a-Prieto, J. Am.
4
4
4
4
5
6
Chem. Soc. 103, 1574 ͑1981͒.
1
1
0
1
J. M. Parnis and G. A. Ozin, J. Phys. Chem. 93, 4023 ͑1989͒.
A. J. Rest, I. Whitwell, W. A. G. Graham, J. K. Hoyano, and A. D.
McMaster, J. Chem. Soc., Chem. Commun. 1984, 624.
R. L. Rubinovitz and E. R. Nixon, J. Phys. Chem. 90, 1940 ͑1986͒.
P. B. Armentrout, Annu. Rev. Phys. Chem. 41, 313 ͑1990͒.
J. C. Weisshaar, Adv. Chem. Phys. 82, 213 ͑1992͒.
C. E. Brown, S. A. Mitchell, and P. A. Hackett, Chem. Phys. Lett. 191,
175 ͑1992͒.
1
1
1
1
2
3
4
5
S. A. Mitchell, in Gas-Phase Metal Reactions, edited by A. Fontijn
͑North-Holland, Amsterdam, 1992͒, Chap. 12.
M. R. A. Blomberg, P. E. M. Siegbahn, and M. Svensson, J. Phys. Chem.
96, 9794 ͑1992͒.
47
P. E. M. Siegbahn and M. R. A. Blomberg, J. Am. Chem. Soc. 114, 10548
48
͑1992͒.
K. Balasubramanian and Z. Ma, J. Phys. Chem. 95, 9794 ͑1991͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
32.174.254.155 On: Wed, 24 Dec 2014 15:44:22
1