Chemiluminescence from the Ba((3)P)+N(2)O-->BaO(A (1)Sigma(+))+N(2) reaction: Collision energy effects on the product rotational alignment and energy release.
ABSTRACT Both fully dispersed unpolarized and polarized chemiluminescence spectra from the Ba((3)P)+N(2)O reaction have been recorded under hyperthermal laser-ablated atomic beam-Maxwellian gas conditions at three specific average collision energies E(c) in the range of 4.82-7.47 eV. A comprehensive analysis of the whole data series suggests that the A (1)Sigma(+)-->X (1)Sigma(+) band system dominates the chemiluminescence. The polarization results revealed that the BaO(A (1)Sigma(+)) product rotational alignment is insensitive to its vibrational state upsilon(') at E(c)=4.82 eV but develops into an strong negative correlation between product rotational alignment and upsilon(') at 7.47 eV. The results are interpreted in terms of a direct mechanism involving a short-range, partial electron transfer from Ba((3)P) to N(2)O which is constrained by the duration of the collision, so that the reaction has a larger probability to occur when the collision time is larger than the time needed for N(2)O bending. The latter in turn determines that, at any given E(c), collinear reactive intermediates are preferentially involved when the highest velocity components of the corresponding collision energy distributions are sampled. Moreover, the data at 4.82 eV suggest that a potential barrier to reaction which favors charge transfer to bent N(2)O at chiefly coplanar geometries is operative for most of the reactive trajectories that sample the lowest velocity components. Such a barrier would arise from the relevant ionic-covalent curve crossings occurring in the repulsive region of the covalent potential Ba((3)P)cdots, three dots, centeredN(2)O((1)Sigma(+)); from this crossing the BaO(A (1)Sigma(+)) product may be reached through mixings in the exit channel with potential energy surfaces leading most likely to the spin-allowed b (3)Pi and a (3)Sigma(+) products. The variation with increasing E(c) of both the magnitude of the average BaO(A (1)Sigma(+)) rotational alignment and the BaO(A (1)Sigma(+)) rovibrational excitation, as obtained from spectral simulations of the unpolarized chemiluminescence spectra, consistently points to additional dynamic factors, most likely the development of induced repulsive energy release as the major responsible for the angular momentum and energy disposal at the two higher E(c) studied. The results of a simplified version of the direct interaction with product repulsion-distributed as in photodissociation model do not agree with the observed average product rotational alignments, showing that a more realistic potential energy surface model will be necessary to explain the present results.
Chemiluminescence from the Ba„3P…+N2O\BaO„A1?+…+N2reaction:
Collision energy effects on the product rotational alignment and energy
Maximiliano Rossa, Carlos A. Rinaldi, and Juan C. Ferreroa?
Centro Láser de Ciencias Moleculares, INFIQC, and Departamento de Fisicoquímica, Facultad de Ciencias
Químicas, Universidad Nacional de Córdoba, X5000IUS Córdoba, Argentina
?Received 13 July 2009; accepted 29 December 2009; published online 15 January 2010?
beam-Maxwellian gas conditions at three specific average collision energies ?Ec? in the range of
4.82–7.47 eV. A comprehensive analysis of the whole data series suggests that the A1?+→X1?+
band system dominates the chemiluminescence. The polarization results revealed that the
BaO?A1?+? product rotational alignment is insensitive to its vibrational state ?? at ?Ec?=4.82 eV
but develops into an strong negative correlation between product rotational alignment and ?? at
7.47 eV. The results are interpreted in terms of a direct mechanism involving a short-range, partial
electron transfer from Ba?3P? to N2O which is constrained by the duration of the collision, so that
the reaction has a larger probability to occur when the collision time is larger than the time needed
for N2O bending. The latter in turn determines that, at any given ?Ec?, collinear reactive
intermediates are preferentially involved when the highest velocity components of the
corresponding collision energy distributions are sampled. Moreover, the data at 4.82 eV suggest that
a potential barrier to reaction which favors charge transfer to bent N2O at chiefly coplanar
geometries is operative for most of the reactive trajectories that sample the lowest velocity
components. Such a barrier would arise from the relevant ionic-covalent curve crossings occurring
in the repulsive region of the covalent potential Ba?3P?¯N2O?1?+?; from this crossing the
BaO?A1?+? product may be reached through mixings in the exit channel with potential energy
surfaces leading most likely to the spin-allowed b3? and a3?+products. The variation with
increasing ?Ec? of both the magnitude of the average BaO?A1?+? rotational alignment and the
BaO?A1?+? rovibrational excitation, as obtained from spectral simulations of the unpolarized
chemiluminescence spectra, consistently points to additional dynamic factors, most likely the
development of induced repulsive energy release as the major responsible for the angular
momentum and energy disposal at the two higher ?Ec? studied. The results of a simplified version
of the direct interaction with product repulsion-distributed as in photodissociation model do not
agree with the observed average product rotational alignments, showing that a more realistic
potential energy surface model will be necessary to explain the present results. © 2010 American
Institute of Physics. ?doi:10.1063/1.3294880?
The reactions of metal atoms with halogen and oxygen-
containing molecules in the gas phase aroused a continuing
interest for more than 70 years.1–3An important dynamical
aspect of these electron-transfer reactions concerns the role
that internal motion of the molecular partner plays in deter-
mining the reaction outcome, especially when oxidant mol-
ecules having a small positive or a negative electron affinity
are involved. In such circumstances, the electron transfer
may be prevented from occurring instantly but it would first
require either the stretch of a bond or the bend of an angle
connecting to a halogen or an oxygen atom of the molecule
to increase its electron affinity to the extent that the charge
transfer can proceed.2
Direct observation of the coupling between an internal
coordinate of the molecular reactant ?e.g., bond distance or
bond angle? and the approach coordinate of the metal atom
in molecular-beam scattering experiments is often hindered
by the fact that the time needed for deformation of the neu-
tral molecule is usually much shorter than the collision time.
As a consequence, cross sections are completely averaged
over the internal degrees of freedom. Over the past two de-
cades, though, a number of full-collision studies of metal
atom-oxidant reactions reported on evidence suggesting that
internal motion of the molecule imposes strong dynamical
constraints on the collision time, particularly at collision en-
ergies large enough so that the collision time is lower than
the time needed for rearrangement of the molecule or the
a?Author to whom correspondence should be addressed. Tel.: ?54 351
THE JOURNAL OF CHEMICAL PHYSICS 132, 034304 ?2010?
0021-9606/2010/132?3?/034304/12/$30.00© 2010 American Institute of Physics
intermediate collision complex. Reported examples include
the finding in the K+CH3I→KI+CH3reaction studied at
hyperthermal collision energies of a pronounced steric effect,
of comparable magnitude to that observed at thermal
energies,4as well as the observation of collision energy ef-
fects such as stairlike behavior and forward transition state
shifts in the chemiluminescent reactions of Ca?4s3d1D2? at-
oms with HCl and HBr ?Ref. 5? and ground-state and elec-
tronically excited manganese atoms with a variety of
molecules,6respectively. For the manganese atom reactions
with O2and SO2,6,7additional measurements of the average
MnO??A6?+,A?6?? product rotational alignments as a
function of collision energy provided complementary infor-
mation about the effect of collision-time constraints on the
stereodynamics of these systems.
Related time-scale limitations2have been considered in a
recent study of the chemiluminescent Ba?3P?+N2O reaction
under beam-gas conditions at hyperthermal collision energies
from this laboratory,8to account for the observed negative
BaO?A1?+? product and its vibrational level. Such behavior
has been rationalized by assuming that the reaction involves
a short-range, partial electron transfer from Ba?3P? to N2O
which is constrained by the duration of the collision, so that
the reaction has a larger probability when the N2O has time
enough to bend during the collision. This dynamical aspect
of the reaction was deemed to be of utmost importance given
that the hyperthermal, laser ablation Ba?3P? atom source
used there has a broad distribution of collision energies, thus
reflecting a wide range of collision times. This in turn is
expected to result in specific generation of products in rela-
tively low ?? levels and with substantial center-of-mass
alignment from reactive encounters that sample the highest
velocity components of the collision energy distribution, and
the opposite tendency, or a lack of specificity whatsoever, in
the low collision energy regime.
This article reports an extended study of the Ba?3P?
+N2O→BaO?+N2reaction dynamics under beam-gas con-
ditions. Advantage was taken here of the ability to modify
significantly the Ba?3P? velocity distribution by changing the
laser ablation conditions of the atomic beam source used
previously.8–10Average values for both the initial relative
velocity and collision energy that have been reported in Ref.
8, where pure thermal conditions were erroneously assumed
are recalculated here by applying a different procedure,
which is appropriate to the present hyperthermal atomic
beam-Maxwellian gas experiments. The analysis of both un-
polarized and polarized BaO chemiluminescence spectra and
their average collision energy dependence supports the pre-
vious assignment of the emission to the A1?+→X1?+band
system.Average and wavelength-resolved values of the prod-
uct rotationalalignment associated
BaO?A1?+→X1?+? emission are presented for the two spe-
cific average collision energies of 4.82 and 7.47 eV.
BaO?A1?+? rovibrational population distributions at the av-
erage collision energies of 4.82, 5.85, and 7.47 eV are also
determined by simulating the corresponding unpolarized
chemiluminescence spectra. Taken altogether, the results pro-
vide a comprehensive picture of the energy and angular mo-
mentum disposal in the title reaction. Attempts to account for
the observed average product rotational alignments by using
a simplified version11of the generalized electron-jump direct
interaction with product repulsion-distributed as in photodis-
sociation ?DIPR-DIP? model12provided further clues to the
major dynamic factors controlling the angular momentum
disposal at the hyperthermal collision energies studied.
The experimental setup and procedure are essentially as
described before.8Briefly, a 5 Hz pulsed Ba atom beam was
generated by focusing the 1064 nm output of a Q-switched
neodymium:yttrium aluminum garnet laser ?Laseroptics
LND 532, 10 ns? onto a rotating Ba disk at normal incidence
in high vacuum conditions. The ablation laser fluence ? was
set to the three specific values of 5.3, 7.6, and 10.8 J/cm2
with an approximately 0.25 mm diameter laser spot. The
ablated material passed through a 0.2 cm diameter collima-
tor, located 0.3 cm from the ablation volume, into a 6 cm
length scattering cell. A detailed characterization of the re-
sulting beam at the prevailing laser ablation conditions and
in the absence of N2O scattering gas has been reported
elsewhere.8–10The time-of-flight ?TOF? technique through
pulsed laser-induced fluorescence and time-resolved fluores-
cence detection was used to this goal. At 3.3 cm from the Ba
disk surface where the chemiluminescence is probed, the
barium beam is composed mainly of neutral atoms in the
6s2 1S0, 6s5d3DJ, and 6s6p3PJstates in an overall abun-
dance ?98% of that of the total Ba atoms. The approximate
beam compositions regarding these relevant Ba atom states
at the three ablation conditions are listed in Table I. These
were determined by numerically integrating the whole TOF
profiles of the1S0,3DJ, and3PJstates of Ba8–10and further
processing the resulting areas as described in Ref. 8. The
corresponding TOF distributions are broad and bimodal; in-
creasing ? from 5.3 to 10.8 J/cm2was found to modify
significantly the general form of the TOF distributions of
3P1, but to alter hardly the shape and peak of those corre-
distributions, corrected also for the metastable radiative de-
cay of the3P1state ?A=3.53?105s−1?,13were fitted to a
superposition of two terms of a two-parameter form for the
velocity distribution f???=N?3exp?−??−?0?2/?2?, where ?
denotes the velocity, N is a normalization constant, and ?0, ?
are adjustable parameters. Table II shows the parameters ob-
tained from the best fits, together with values of the average
initial relative velocity ?k? and collision energy ?Ec? for both
the two components and the whole of the velocity distribu-
tions of Ba?3P1?, as derived from convoluting14the beam and
target N2O ?298 K? gas distributions. Note that the procedure
used here to derive ?k? and ?Ec?, which takes ?0into account,
is different from that used previously,8where pure thermal
conditions were erroneously assumed. In the present, proper
approach the values of ?k? and ?Ec? at 7.6 J/cm2are found to
be ?5.8?0.4? km/s and ?5.85?0.04? eV, respectively.
While other components of the3P level have not been inves-
tigated, their velocity distributions should be very similar to
3DJ. The flux-weighted Ba?3P1? TOF
3P1, as it was found previously for
034304-2 Rossa, Rinaldi, and FerreroJ. Chem. Phys. 132, 034304 ?2010?
amount of small Banspecies in the beam is considered to be
negligible since neither laser-induced fluorescence nor emis-
sion other than those of monatomic Ba species were de-
For chemiluminescence experiments, the atomic beam
was allowed to interact within the scattering cell with a
standing pressure of N2O ?AGA, 99.8%, 0.76 mTorr?. The
chemiluminescence from reactive collisions was collected
perpendicularly to the beam axis at 3.3 cm from the disk
surface by a biconvex lens and analyzed by the same scan-
ning monochromator and photomultiplier tube combination
as in Ref. 8, which was operated with a resolution of 0.5 nm.
Signals from four laser shots were averaged and time inte-
grated from the firing of the ablation laser to extinction for
each spectral point. For polarization measurements, a visible
linear polarizer ?Edmund Optics, 52–557? was placed be-
tween the collecting lens and the monochromator. This linear
polarizer was rotated in order to measure the chemilumines-
cence polarized in the planes perpendicular and parallel to
the direction of the Ba beam. Since the contribution of the
most populated Ba ?1S0,3D, and3P? states to the chemilu-
minescence spectra is dominant under these conditions, the
spectra corresponding only to the collision of Ba?3PJ? with
N2O were determined by subtracting the joint contribution of
the1S0and3DJstates, in much the same way as the polarized
Ba?3P?+N2O chemiluminescence spectra were previously
obtained.8The effect of the ionic species present in the beam
was neglected since their removal, by a pair of deflection
plates, did not significantly affect the observed chemilumi-
TABLE I. Reactant Ba beam compositions in percent of the most populated 6s2 1S0, 6s5d3DJ, and 6s6p3PJ
electronic states of Ba atoms. For each laser ablation fluence reported, the experimental values probing the3P1
substate alone are given in the left column, while values ?in italics? in the right column are rough estimates ?see
Ref. 8 for details? of the overall compositions including the relative contribution of the3P0,2states, as computed
by assuming 2J+1 degeneracies at the ablation volume and properly accounting for Ba?3PJ? radiative decay
over the flight path.
TABLE II. Parameters that characterize both the two components and the whole of the Ba?3P1? velocity
distributions in the reactant atomic beam, together with the corresponding average values for the initial relative
velocity and collision energy distributions in the beam-gas scattering experiments. Raw data at ?=7.6 J/cm2
were taken from previous work ?Ref. 8?.
Collision energy effects on the Ba?3P?+N2OJ. Chem. Phys. 132, 034304 ?2010?
III. RESULTS AND ANALYSIS
A. Chemiluminescence spectra
The shapes of the Ba?3P?+N2O chemiluminescence
spectra were quite similar for the three average collision en-
ergies used in this work, and Figure 1?a? shows a represen-
tative unpolarized spectrum obtained at ?Ec?=7.47 eV. In-
creasing ?Ec? from 4.82 to 7.47 eV was found to cause a
comparatively minor shift in wavelength from 479 to 486 nm
of the peaks of such spectra where no molecular band sys-
tems are discernible. Although a detailed spectroscopic as-
signment is not feasible here, a consideration of the
Ba?1S0,3D?+N2O chemiluminescence spectra under beam-
scattering conditions can help in identifying spectral fea-
tures. In both lower-excited reaction channels, unstructured,
broad spectra over the near-ultraviolet/visible/near-infrared
range have been observed and ascribed to dominant joint
A1?+→X1?+and A?1?→X1?+emission; for the Ba?3D?
reaction, a feature in the red near 800 nm was also observed
and assigned to the D1?+→A1?+transition.15,16Notewor-
thy is the limited number of BaO luminescent states that are
observed in these lower-excited reactions regarding the
whole electronic states that are energetically accessible. In-
deed, the complete ground state Ba+N2O→BaO+N2reac-
tion is already significantly exoergic ??D0=−4.11 eV? so
that all presently known electronic states of BaO 14 in num-
ber at energies ?4.29 eV ?Refs. 17–21? could be populated
at the range of thermal collision energies of previous
studies.15,16Generation of several thus far unobserved ex-
cited states, such as the ? states that are expected to lie in the
2–4 eV range on the basis of indirect spectroscopic
evidence,17,18may be partially the cause for the apparent
predominance of the A and A? states of BaO in the chemilu-
minescent Ba?1S0,3D?+N2O reactions.15
At the hyperthermal ?Ec? studied here, the Ba?3P?
400–600 nm spectral region and does not exhibit the red
feature seen in the corresponding Ba?3D? reaction. Since the
emission herelies within
??350–900 nm? covered by the A−X and A?−X systems of
BaO,22one or both of the latter, and arguably also the
a3?+→X1?+transition,23could be involved in much of the
observed chemiluminescence, as it was found to be the case
for the related Ba ?1S0,
channels.15,16,24Nevertheless, the long radiative lifetimes of
the A?1???=9 ?s? ?Ref. 25? and the triplet ?metastable?
states make drift of these excited BaO products out of the
detector field of view sizeable at the hyperthermal average
initial relative velocities of the present experiments. Indeed
assuming that the product velocity in the laboratory frame is
comparable to ?k?, a BaO molecule will travel 47.7 mm
during 9 ?s for the average velocity of 5.3 km/s at ?Ec?
=4.82 eV. Thisisabout
??5.3 mm? of the observation volume for the present
experiments,8and higher ratios would be clearly obtained in
the case of the triplet states. In contrast, a similar calculation
performed for the short-lived ??=350 ns? ?Ref. 26? A1?+
state leads to a travel distance of 1.9 mm. Thus, as far as the
BaO states lying in the 2 eV region are concerned,18,20the
optics detection system used here was likely to be sensitive
mainly to the A1?+state. Both spectral simulation of the
BaO chemiluminescence and the polarized Ba?3P?+N2O
chemiluminescence spectra to be reported in Secs. III B and
III C, respectively, will add evidence for the predominance
of the A−X transition over A?−X and a−X.
The BaO states lying above 4 eV ?Refs. 17 and 19? may
also contribute to the observed emission via transitions to
either the ground state or lower-lying excited states. Of the
many possible systems, quite a few have been actually docu-
Intensity (a. u.)
Intensity (a. u.)
Intensity (a. u.)
Intensity (a. u.)
FIG. 1. ?a? Ba?3P?+N2O chemiluminescence spectrum recorded at a N2O
pressure of 0.76 mTorr and at ?Ec?=7.47 eV. It represents the average of
three individual scans previously corrected by spectral response of the de-
tection system ?Ref. 8?. ?b?, ?c?, and ?d? compare the experimental spectrum
?¯? to simulations ?—? of BaO A1?+and A?1? emission using electronic
branching ratios ??A/A?? of 0, 1, and ?, respectively.
034304-4 Rossa, Rinaldi, and FerreroJ. Chem. Phys. 132, 034304 ?2010?
mented in the visible range, namely, C1?+−a, A, b3?1, A?,
and the weaker portion of C1?+−X,17,19,27by means of a
number of laser spectroscopies. Other band systems have
been reported in the UV range and ascribed to the C−X
transition,27along with the above-mentioned chemilumines-
cent feature near 800 nm that was assigned to D−A
emission.15Neither of the latter are in evidence here ?Fig.
1?a??. At least for the C−X emission, flyout effects would not
be of significance given the short radiative lifetime ??
=10.5 ns? ?Ref. 28? of the C1?+state. The absence here of
UV emission associated with the C−X system parallels ob-
servations made previously in the Ba?3D?+N2O reaction.15
Such findings were suggested there as an argument against a
C−A, A? contribution to the 800 nm feature as these emis-
sions, if present, would be expected to be one order of mag-
nitude weaker than C−X.28Clearly, the contribution of C
−A, A? emission to the chemiluminescence here can be dis-
missed on similar grounds.
Concerning the involvement here of 4 eV states other
than C and D as emitters, no definite assessment could be
made given that the dipole transition moments for the pro-
spective band systems are unknown. An insight can still be
gained on the basis of calculated Franck?Condon factors
?FCFs?. Such calculations are feasible for six out of seven
remaining high-lying states, which have been sufficiently
well characterized to permit construction of potential energy
curves.19These are, in order of increasing energy, B1?,
E1?+, F3?1, c3?0, G3?0, and H3?1. FCFs were obtained
using numerical potentials generated by the Rydberg?Klein
?Rees ?RKR? method29of inverting spectroscopic data via
the RKR1 and LEVEL program packages.30The relevant spec-
troscopic constants for all systems terminating on the ground
state were taken from Refs. 19 and 31; FCF were then cal-
culated for transitions from ???10 to ???34. For all six
possible systems, the magnitude of the FCF for any given
???,??? pair is quite similar, the largest FCF ??0.05? occur-
ring generally in the 270–670 nm wavelength range. Neither
of the relevant systems have been observed in shock tube
absorption between 290 and 390 nm,32which implies that the
corresponding transition moments are probably small. This,
combined with the fact that no chemiluminescence is found
here at wavelengths less than 362 nm suggests that the
B,E,F,c,G,H−X systems are unlikely to contribute to the
FCFs were also calculated for all 24 possible band sys-
tems involving transitions from ???10 of the B, E, F, c, G,
and H states to ???34 of the lowest excited a, A, b, and A?
states.18For both a/A and b/A? pairs, it is found that the
FCF matrices corresponding to any given higher-lying state
are almost identical, as expected from the near degeneracy of
the relevant triplet/singlet states within the range of ?? con-
sidered. There are significant differences between the FCF
matrixes for transitions terminating, on one hand, on a/A
states as well as the H−b/A? transitions, which have the
largest FCF ?0.05–0.30? for wavelengths greater than
500 nm, and on the other hand, on the remaining b/A? states
having strong FCF ??0.1? generally in the more limited
wavelength range of 600–800 nm. In all cases the FCFs van-
ish for wavelengths less than 500 nm. Taken altogether, the
results allow for broad emission above 500 nm and into the
B,E,F,c,G,H−a,A,b,A? band systems. Indeed, the FCF
matrices for G,H−a,A transitions bear a remarkable similar-
ity to that for the D−A system, on which assignment15as
D−A emission of the broad feature at 750–890 nm that is
observed in the Ba?3D?+N2O reaction was mainly based.
The E, F, G, and H states were unknown, and both B and c
states were incompletely characterized at the time of Cox
and Dagdigian’s report,15but it is clear from the present
evidence that B,E,F,c,G,H−a,A as well as the H−b/A?
emission may well account for such broad red feature. The
lack of observation of chemiluminescence at wavelengths
greater than 650 nm instead indicates that B,E,F,c,G,H
−a,A,b,A? emission is not present here. This reinforces sug-
gestions made in the two preceding paragraphs that, under
the present experimental conditions, the emitter is unlikely to
be one of the known states at 4 eV. Nevertheless, it is entirely
possible that presently unknown either singlet or triplet BaO
excited states lying at energies ?4.29 eV could be respon-
sible for the observed chemiluminescence.
The Ba?3P?+N2O reaction is 1.62 eV above the ground
state channel, which, given the BaO ionization energy of
6.91?0.06 eV,33would allow population of excited states
of BaO much higher in energy than those hitherto considered
including Rydberg states. However, the latter are not likely to
emit strongly in the visible but instead expected to radiate by
a cascade mechanism, the initial steps of which will involve
Rydberg?Rydberg transitions at long wavelength. Corrobo-
ration of the presence of such highly excited BaO states
would clearly require detection of chemiluminescence at
wavelengths in the IR range, which, unfortunately, is out of
the sensitive spectral range of the detection system used here.
In addition, the total energies available to the BaO products
Etotat the prevailing ?Ec?, as listed in Table III are extremely
high and would allow formation of nonemitting products
from both collision-induced dissociation Ba+N2O→Ba+O
+N2 as well as collisional ionization Ba+N2O→Ba+
Thus, both the present experimental observations and
most of the available spectroscopic information consistently
suggest that, under the present conditions, the observed
Ba?3P?+N2O chemiluminescence spectra are dominated by
A1?+→X1?+, and probably also A?1?→X1?+emission.
Spectral simulations of the BaO emission, to be described
next, were tried in order to substantiate such an assignment,
while also allowing to extract the nascent BaO?product state
distributions from the experiments.
B. Spectral simulations
A number of tests has been performed by varying the
input A and A? rovibrational distributions as well as the elec-
tronic branching ratio ??A/A?? in the simulation program.
The relevant spectroscopic constants were taken from
Gottscho et al.;18at the present spectral resolution and ex-
tensive internal excitation of the products ?see below?,
though, it was found that the use of more recent, improved
constants reported by Furio and Pruett20for the excited state
Collision energy effects on the Ba?3P?+N2O J. Chem. Phys. 132, 034304 ?2010?
and Li et al.21for the ground state did not lead to noticeable
differences regarding the results that derive from the pre-
ferred set of constants. FCFs for the A−X and A?−X band
systems were obtained by the procedure described in Sec.
III A. The rotational line strengths were taken from
It follows from preliminary simulations including either
excited electronic state that a substantial BaO rotational ex-
citation, characterized by a rotational temperature Trin the
range of 3000–4000 K, is required to reproduce the headless,
“many-line” appearance of the experimental spectra ?see Fig.
1?a??. This allows one to yield only an estimate of the mean
rotational excitation of the BaO?products. In the recorded
wavelength region, though, the corresponding input vibra-
tional distributions are found to have a much more marked
effect on the envelops of the simulated spectra. Such findings
are in good agreement with conclusions reached in a previ-
ous study35of the Ba?6s6p1P1?+CO2reaction at a number
of thermal collision energies and, as there, they afforded here
to extract reliable information about the BaO?vibrational
excitation from the fits.
Further inspection of the effect of the A/A? branching
ratio on the simulations leads to the conclusion that the red
part ???700 nm? of the calculated spectra is very sensitive
to this parameter, again in accord with previous findings for
the Ba?6s6p1P1?+CO2reaction.35This feature was found to
be advantageous here in order to ascertain the relative A−X
and A?−X contribution to the observed emission. Thus, Fig-
ures 1?b?–1?d? compare the experimental spectrum at ?Ec?
=7.47 eV with three examples of simulated spectra corre-
sponding to different values for ?. The input, inverted vibra-
tional distributions used for the A and A? states are shown in
Figure 2. For the rotation motion, Boltzmann distributions
characterized by Tr=4000 and 3000 K were chosen for the A
and A? states, respectively. It is readily apparent from Fig. 1
that the most satisfactory agreement between calculation and
experiment is obtained in the case of pure A contribution,
i.e., ?=?. There are only slight discrepancies in the blue
wing of the chemiluminescence, particularly in the region
below 430 nm where the calculated spectrum decreases to
extinction much faster than observed. It should be noticed
that such discrepancies remain whatever the value of ? as
well as the chosen A and A? rovibrational distributions. Be-
fore addressing this concern, it is important to highlight at
this stage that addition of any A? contribution differently
affects the red wing of the simulated spectra, which then dye
off toward the red much slowly than both the experiment and
the pure A calculation, and extends beyond 850 nm. The fact
that no signal is experimentally observed here at ?
?650 nm allows to state that such A?−X emission, if
present, contributes only to the background level. Indeed, an
A−X to A?−X intensity ratio of 99/1 could be estimated un-
der the present conditions from the observed 40:1 signal-to-
noise ratio ?see Fig. 1?a??, which can be taken as a further
evidence for a predominant A−X contribution. Actually,
these conclusions hold irrespective the average collision en-
ergy considered since the marked similarity among the cor-
responding chemiluminescence spectra resulted in the same
A−X simulation parameters leading to a satisfactory fit.
In tracing the possible sources of the blue wing discrep-
ancies, there is to note that inclusion of either a3?+or b3?
states in the simulations is not expected36to lead to notice-
able differences regarding the case for the contribution of
only A and A? states, since these corresponding triplet and
singlet states cannot be resolved in the present experiments.
TABLE III. Energy disposal in the Ba?3P?+N2O→BaO?+N2reaction at the various average collision energies.
All energy values are in eV.
Observed blue limit EB
aReaction exoergicity plus average collision energy.
bThese quantities refer to the internal energy of the BaO product alone, as derived from the best fits to the
experimental spectra using ??MP=9, ??Max=34, and ?J??=91.
cValue at ??Max=34 of BaO?A1?+?.
dTaken above Tefor the A1?+state of BaO.
05 1015 20 25 30
Relative population (a. u.)
Vibrational quantum number (υ´)
FIG. 2. Vibrational state distributions of the ?a? A1?+and ?b? A?1? elec-
tronic states of BaO products corresponding to the simulations of Fig. 1.
034304-6Rossa, Rinaldi, and Ferrero J. Chem. Phys. 132, 034304 ?2010?
Moreover, any possible a−X and b−X contributions to the
observed emission are deemed to be even less than that of
A?−X as a result of stronger flyout effects that are expected
on the metastable triplet products with respect to the shorter-
lived A? state. From the discussion in Sec. III A, it is also
clear that, while any definite conclusion about the contribu-
tion of higher-lying BaO states is virtually impossible, it is
highly unlikely for one of the known states at 4 eV to be the
emitter, and thus being responsible for such discrepancies.
Instead, an examination of the vibrational levels ?? of the
A1?+state that have the largest FCF for fluorescence emis-
sion to the X1?+ground state at a given wavelength can be
helpful ?see Fig. 4?b? of Ref. 8?. It reveals8that the main
contribution to the chemiluminescence spectrum below 430
nm originates from fairly large BaO?A1?+? vibrational quan-
tum numbers ????20? while emission above 540 nm is
dominated by molecules with low vibrational excitation ???
?8?. Actually, the determination by the RKR method of the
corresponding FCF for ???20 may be rather uncertain ow-
ing to the extrapolation of the relevant potential curves to
vibrational and rotational states much higher in energy than
those from which the input spectroscopic constants of Ref.
18 have been obtained. This would lead in turn to a substan-
tial uncertainty in the calculated emission intensity of such
blue chemiluminescence wing, which might be the cause of
the disagreement. Nonetheless, the best-fit simulation con-
sidering BaO?A1?+? as the sole emitter ?see Fig. 1?d?? gen-
erally reproduces the main features of the experiment,
namely, the envelope of the spectrum and the amplitude of
its “noise,” which indicates that both the vibrational distribu-
tion and the average rotational energy used in the calculation
should be close to the real ones.
Essentially the same simulation parameters are appropri-
ate to fit the chemiluminescence spectra at all ?Ec?, which
means that the internal energy of BaO?remains almost con-
stant as the collision energy is increased. An upper limit to
such internal energy can be obtained by adding together the
value of energy Evib
vibrational level of the BaO?A1?+? product resulting from
the best fits and the corresponding mean rotational energy
?Erot?. Both values are listed in Table III, along with the
difference Edbetween the total energy available to the
BaO?A1?+? products Etotand these values. Although Edonly
yields a lower limit to the energy released as translation and
as internal energy of the N2fragment, values in Table III
show up a substantial gap between the internal energy of
most of the BaO?A1?+? products and the total product en-
ergy in the title reaction.
Alternatively, a piece of information regarding energy
disposal could be obtained from the short-wavelength limit
of the observed chemiluminescence spectra,15which is
largely independent of a definite spectroscopic assignment of
the emitters. The blue limits at all ?Ec? are also listed in Table
III, where the corresponding photon energies are compared
with the total product energy. It is worth noting the close
agreement between the values for Etot−EBand Edat every
?Ec?, which represent two ways of estimating the energy
deficit in the present reaction, while also constituting sepa-
rate evidence for a main BaO?A1?+? emitter.
Maxcorresponding to the highest populated
C. Polarized chemiluminescence spectra and product
Figure 3?a? presents the parallel and perpendicular polar-
ized spectra at ?Ec?=4.82 and 7.47 eV, from which a reduc-
tion in the sensitivity of the emission band as a whole to the
polarization direction can be ascertained at the highest ?Ec?.
To make this evident, the degree of polarization, defined by
P =I?− I?
was obtained as a function of wavelength from the chemilu-
minescence intensities polarized parallel, I?, and perpendicu-
lar, I?, to the barium beam axis, and it is shown at the two
working ?Ec? in Fig. 3?b?. Comparison with the correspond-
ing data at 5.85 eV ?see Fig. 4 of Ref. 8; there, ?Ec? was
erroneously calculated to be 1.56 eV? shows that the values
of P at any given emission wavelength decrease with in-
creasing ?Ec?, which is confirmed by the average polarization
values associated with the whole BaO?emission that are
listed as a function of ?Ec? in Table IV. The BaO?polariza-
tion at 7.47 eV is also found to increase with wavelength by
more than a factor of 25 over the whole emission band,
which is in line with, but significantly higher than the ap-
proximately fivefold increase in the P values that was previ-
ously observed at 5.85 eV.8This does not appear to be the
case at 4.82 eV, where no discernible emission wavelength
dependence of the P values is observed throughout the emis-
As made in the preceding two sections, it is also impor-
tant here to consider the possibility that two or more BaO
excited states are responsible for the observed ?polarized?
emission. In such a case, whereas the individual contribu-
tions to I?and I?would be additive, the very definition of P
implies that the corresponding contributions to the observed
degree of polarization cannot be resolved in individual terms.
Nevertheless, some predictions about the effect on P of in-
dividual contributions as a function of ?Ec? could still be
made in the case of the two overlapping A−X and A?−X
band systems, which are associated with electronic transi-
tions with dominant parallel and perpendicular character, re-
spectively. The latter guarantees that a bias to positive values
of P have to ensue any increasing A−X contribution to the
emission, while the opposite will be true for an increasing
A?−X contribution. Under the present conditions, any pro-
spective A?−X contribution to the polarized spectra is antici-
pated to decrease with rising ?Ec?, with an accompanying
increase in the bias of P at any given emission wavelength to
positive values owing to flyout effects for the longer-lived
A?1? state, which is contrary to the present observations
?see Fig. 3?b??. The situation would become tremendously
more intricate in the case that additional parallel or perpen-
dicular transitions are contributing to the polarized emission.
Notwithstanding, the rather smooth dependence of P??? ob-
served throughout the emission band at all ?Ec? studied indi-
cates that there should be only a very small number of BaO
excited states that contribute to the chemiluminescence.
Most of the evidence presented above in analyzing the un-
Collision energy effects on the Ba?3P?+N2O J. Chem. Phys. 132, 034304 ?2010?
polarized spectra and their spectral simulations certainly sug-
gest that the A−X band system does dominate chemilumines-
cence. Hence, the ensuing analysis of the corresponding
polarized spectra will be based on assuming that such paral-
lel transition is solely responsible for the emission here.
The center-of-mass product
?P2?J?·kˆ?? was extracted from the polarization measure-
ments. For a parallel-type transition, P and the value of
?P2?J?·kˆ?? are related by the expression37
?P2?J? · kˆ?? =
3 − ?P · ?P2?kˆ· z???
where the term ?P2?kˆ·z?? properly accounts for the beam-gas
averaging, and z is along the beam axis. In the present ex-
perimental configuration, ?P2?kˆ·z?? depends on the masses of
the beam and gas species, the temperature of the latter and
the nominal collision energy ?Ec
reduced parameter.38As shown previously,8it is reasonable
to assume that ?P2?kˆ·z???1 for the reactive collision ener-
gies employed in the present experiments. Figure 3?c? dis-
plays the conversion of the P??? data presented in Fig. 3?b?
to ?P2?J?·kˆ????? by means of Eq. ?2?, and Table IV contains
rotational alignment values averaged
throughout the emission band ?P2?J?·kˆ??avat all working
?Ec?. Note that the latter are close to the range of values of
−?P2?J?·kˆ??=0.14−0.38 ?Refs. 39 and 40? and 0.33 ?Ref. 40?
previously reported for the Ba?1S0?+N2O and Ba?3D?
+N2O reactions, respectively, which were also studied under
beam-gas conditions but at a range of thermal collision en-
0=?1/2???x/t?2? through one
400450 500550 600
Intensity (a. u.)
400450 500 550600
400 450500 550600
400 450 500 550600
FIG. 3. ?a? Ba?3P?+N2O chemiluminescence spectra polarized parallel ?—? and perpendicular ?---? to the Ba beam axis at a N2O pressure of 0.76
?10−3Torr. Each represents the average of three individual scans previously corrected by spectral response ?including instrumental polarization? ?Ref. 8?. ?b?
Degree of polarization of the BaO?A1?+? chemiluminescence as a function of wavelength, ?. Three consecutive ? values were binned so that the resolution
was ?1.5 nm; the error bars thus reflect the average intensity uncertainty. ?c? BaO?A1?+? product rotational alignment as a function of wavelength. Left and
right panels correspond to the average collision energies of 4.82 and 7.47 eV, respectively.
TABLE IV. Experimental average degree of polarization and product rota-
tional alignment for the whole parallel BaO?A1?+→X1?+? emission at the
various average collision energies. Data at ?Ec?=5.85 eV were taken from
previous work ?Ref. 8?; note that ?Ec? there was erroneously calculated to be
1.56 eV. Also shown are theoretical values for the product rotational align-
ment that derive from a DIPR-DIP calculation ?Ref. 11?.
034304-8 Rossa, Rinaldi, and FerreroJ. Chem. Phys. 132, 034304 ?2010?
ergies. Clearly the BaO?A1?+? product is significantly
aligned at the hyperthermal collision energies studied here
and, in spite of the limited number of data, it is apparent that
the average product rotational alignment decreases with in-
creasing ?Ec?. Taken altogether, the data presented here and
elsewhere8reveal an additional trend in the variation of the
BaO?A1?+? rotational alignment with collision energy:
While the ?P2?J?·kˆ?? values are rather insensitive to the
emission wavelength at 4.82 eV, a marked wavelength de-
pendence of ?P2?J?·kˆ?? is observed at 5.85 and 7.47 eV
which becomes more pronounced as ?Ec? rises ?cf. Fig. 4?c?
of Ref. 8 and Fig. 3?c? of the present work?. Once again
looking at the largest FCFs for the A1?+→X1?+transition,
it is apparent that the behavior at the highest ?Ec?s studied
here can ascribed to a tendency for BaO?A1?+? rotational
alignment to be negatively correlated with BaO?vibrational
levels. Such behavior was not reported in previous studies of
the lower-excited reaction channels.39,40
Two features of the present observations concerning the
effect of the average collision energy in the 4.82–7.47 eV
range on the Ba?3P?+N2O reaction are most distinctive: the
little discernible variation of the shapes of the chemilumines-
cence spectra with increasing ?Ec? and the far more marked
?Ec? dependence of the average and wavelength-resolved de-
grees of polarization associated to the BaO?emission. The
former strongly suggests that the number of BaO?emitters as
well as their internal energy remain nearly constant as ?Ec? is
increased. This may be unsurprising given that the total prod-
uct energy at the lowest ?Ec? studied here is already signifi-
cantly higher than the BaO ionization energy, below which
the BaO emitting levels are bound to lie, so that the addi-
tional collision energy is not as dramatic an increase in Etot.
Admittedly that it is possible for the lower collision energy
components of the whole Ecdistributions to be responsible
for the observed BaO?signal. Although the excitation func-
tion for the title reaction is yet unknown, it is often found in
exoergic reactions like the chemiluminescent channel of
Ba?1S0?+N2O,41for the cross section to decrease signifi-
cantly as the collision energy increases as a result of a num-
ber of dynamical factors including the shorter interaction
time, recrossing,4or the opening of alternative, endoergic
removal processes. Some possible removal processes for the
present system are collision-induced dissociation and colli-
sional ionization ?see Sec. III A?, with calculated thresholds
of 5.85 and 11.55 eV, respectively, using D?N2–O?
=1.74 eV,33a Ba?1S0? ionization potential of 5.21 eV,42and
a N2O vertical electron affinity ?VEA? of ?2.23 eV.43The
latter threshold, along with the possible production of BaO+
above 6.91 eV, suggests that a broad range of collision ener-
gies could lead to reaction, especially considering the high
efficiency of the Ba?3P? beam source.
Conclusions derived from a comprehensive analysis of
the whole data series here certainly indicate a dominant
A1?+→X1?+contribution to chemiluminescence. This
does not mean that the remaining 2 eV states or higher-lying
states are not populated at all in the present reaction. Instead,
it may show an absence of preferential production of such
BaO excited states regarding A1?+. In such a case, drift of
the longer-lived A?1?, a3?+, and b3? state products out of
the field of view could significantly contribute to the lack of
observation of the corresponding fluorescence. Indeed, the
estimate of 99/1 for the A−X to A?−X intensity ratio in Sec.
III A corresponds to a value of ??A/A??=0.005, which ap-
pears much too small to be accounted for on the basis of
flyout effects alone. The lack of evidence for chemilumines-
cence from highly excited BaO states, especially the short-
lived C1?+state, is difficult to rationalize on these grounds,
which suggests that the mechanism for the present reaction
may exhibit dynamic constraints regarding orbital occupa-
tion. The nature of such a mechanism is discussed hereafter.
The relativelylarge values
BaO?A1?+? product found here and elsewhere8in the un-
constrained Ba?3P?+N2O→BaO?+N2reaction system at the
hyperthermal collision energies studied suggest a direct
mechanism, which is further supported by the high rotational
and, particularly, vibrational excitations of the BaO?A1?+?
Since the BaO?A1?+? product formed has a single
charged Ba+O−character,17,18avoided ionic-covalent curve
crossings have been inferred previously8from the finding of
a negative correlation between the BaO?A1?+? rotational
alignment and its vibrational state, such as that derived here
at ?Ec?=7.47 eV. Within a “multilocation harpoon” picture2
of the reaction mechanism, the outermost ionic-covalent
Ba+?6s2S1/2?¯N2O−?2? ?2A? in Cssymmetry?? ionic sur-
face. The latter correlates adiabatically ?spin disregarded?
with several energetically accessible excited states of BaO:
with A1?+and a3?+in Cs, and with A?1? and b3? in
C2v.41Direct formation of the BaO?A1?+? product requires a
spin-flip in the ion-pair intermediate, a process that would be
favored by a long-lived complex which, in turn, seems in-
compatible with the direct character of the present reaction.
Alternatively, the BaO?A1?+?+N2?1?g
be reached through mixings in the exit channel with potential
energy surfaces ?PESs? leading to the BaO?a3?+,b3??
charge transfer at the outer ionic-covalent crossing. As dis-
cussed previously,8this mechanism appears very plausible
owing to the extensive perturbations that occur17among the
A1?+, b3?, and A?1? states, as well as the A?1? and
a3?+states of BaO. It is also substantiated by analyzing the
valence orbital configuration of the A1?+/a3?+
A?1?/b3? electronic states, i.e., ?Ba?6s??O?2p?2p?4??
and ?Ba?6s??O?2p?22p?3??,18respectively, which suggest
that the Ba+?6s?¯N2O−intermediates could decay to prod-
ucts without orbital rearrangement. In contrast, population of
the 4 eV electronic states, which involves excitations of p?
or 5d orbitals located on Ba,19would require significant
orbital rearrangement of the intermediate and, given that the
2 eV and the 4 eV states do not perturb to each other, the
apparent lack of production in such states indicates some
degree of orbital control that could be involved in the exit
channel of the reaction, at least.
+? singlet surface may
+? triplet surfaces, following access to the latter by
Collision energy effects on the Ba?3P?+N2O J. Chem. Phys. 132, 034304 ?2010?
The indiscernible variation of ?P2?J?·kˆ?? with the BaO?
vibrational levels that is found at 4.82 eV suggest that most
of the relevant reactive encounters sample more similar col-
lision geometries in comparison with the two higher ?Ec?,
and the higher rotational alignments of the former further
indicate that correspondingly higher geometric constraints
are operative. Since N2O bending is expected to play an
increasing role in the reaction intermediates at the relatively
lower collision energies ?and thus higher collision times? of
the lowest ?Ec?, coplanar ?Cs? approach of Ba to the “O-end”
of bent N2O is likely to be favored over collinear attack on
linear ?equilibrium? N2O. A related approach-geometry limi-
tation having also comparable magnitudes at both thermal
and hyperthermal collision energies has been considered by
Wiskerke et al.4to account for steric effects in the harpoon
K+CH3I→KI+CH3reaction. They proposed the existence
of an approach angle-dependent potential barrier which
arises from the ionic-covalent crossing occurring on the re-
pulsive region of the covalent potential, and whose effective
height increases with the specific Ecthat is sampled by any
particular reactive trajectory. At least at thermal collision en-
ergies, an angle-dependent barrier is also expected in the
present reaction by comparison with two models44,45for the
related chemiluminescent channel involving Ba?1S0?, which
were previously proposed on the basis of very detailed attack
such proposal would actually require the ability to orient the
N2O, which is not feasible in the current experiments. Yet it
is plausible that the stereodynamics observed at 4.82 eV
stems from a potential barrier to reaction which favors
charge transfer at chiefly coplanar geometries. In such a case,
some significant alignment of BaO?rotation perpendicular to
k could ensue any product repulsion at threshold if the reac-
tant orbital angular momentum L is much larger than the
rotational angular momentum of N2O, JN2O, as it is likely to
hold8under the current experimental conditions.
Following Wiskerke et al.,4any involvement of ionic-
covalent crossings here is expected to give rise to substantial
barriers owing to the large negative VEA of N2O in its
ground vibrational state ?most populated at 298 K?. A simple
Magee-type calculation47yields a crossing radius Rcfrom the
Ba?3P?¯N2O curve to the ground-state Ba+?2S1/2?¯N2O−
ionic surface ?Ba?3P? ionization potential: 3.59 eV? ?Ref. 42?
of 2.47 Å. Considering Ba approach towards the N2O centre-
of-mass perpendicular to its internuclear axis, the Ba¯O
distance is predicted to be ?2.71 Å which, compared to the
sum of the Ba and O van der Waals radii, 4.60 Å,48suggests
that charge transfer will occur well within the repulsive re-
gion. Additional repulsion is anticipated in this case owing to
the ? electron distribution in N2O?1?+?. Moreover, as this
geometry evolves into perfect ?C?v? O-end attack on N2O,
the system should start to experience repulsion at progres-
sively shorter Ba¯O distances. In this context, inner cross-
ings with excited ionic surfaces ?Ba+?5d2DJ,6p2PJ,¯??
seem unlikely to contribute as they will be shifted further
into the repulsive region.
Similar considerations apply to any reactive encounter
involving N2O in its first excited bending level ?n2=1?, as a
recent high-level ab initio study49on the N2O dissociative
electron attachment ?DEA? reports a small difference ?by
about 50 meV? between the N2O VEAs associated with the
n2=0 ?ground-state? and n2=1 bending levels. Thus, unlike
the K+CH3I reaction,4no significant change in the height
and location of the reaction barrier proposed here is expected
to accompany the decreasing role that bent N2O configura-
tions would play at elevated collision energies.
As pointed out before,8the major dynamical limitation
would instead arise from the charge transfer ?and conse-
quently the reaction? having a larger probability when the
collision time is higher than the time needed for N2O bend-
ing. A recent DEA study49on N2O suggests that dissociation
of N2O−is delayed selectively by an avoided crossing be-
tween the2A? surface that leads to direct N2+O−dissociation
and an2A? bound surface that corresponds in the dissociation
limit to N2
faces is both strongest for linear N2O and strongly dependent
on the N–N–O bond angle. Indeed, this seems suitable to
account for the collision energy-dependent correlation be-
tween the rotational alignment of BaO?A1?+? and its vibra-
tional levels observed here. Namely, at any particular ?Ec?,
only collisions between Ba and N2O having Ecat or above
the proposed reaction threshold will lead to reaction. Geo-
metric requirements at threshold should be rather stringent.
For an initial covalent interaction, a preference for collinear/
coplanar O-end approach of Ba at low impact parameters is
to be expected in order to favor overlapping of the 6p orbital
of Ba?3P? with the 3?/10a? orbital of linear/bent N2O. Both
collision geometries would lead to substantial BaO?rota-
tional alignment, and for the coplanar case a wider, unspe-
cific vibrational population would be favored owing to the
Ba+–O−attraction beginning earlier in the PES. Hence the
alignment results at ?Ec?=4.82 eV can be rationalized on the
basis of the corresponding narrower distribution of collision
energies by assuming that a sizeable potential barrier is op-
erative for most of the reactive trajectories.
For the reactive intermediates involving bent N2O,
?P2?J?·kˆ?? should fall with rising collision energies as a re-
sult of an increasing tendency of k to lie outside the reaction
plane. Along with the predicted prevalence as Ecraises of
collinear reactive intermediates, the insensitivity of the BaO
rotational alignment to its vibrational state at 4.82 eV would
be expected to develop into the observed negative correlation
between ?P2?J?·kˆ?? and ?? at the two higher ?Ec?.
In this context the falling ?P2?J?·kˆ??avvalues with in-
creasing ?Ec? could be due to the development at the two
highest, hyperthermal ?Ec? studied here of so-called “induced
repulsive energy release,”50following relaxation of the col-
lision partners from compressed and bent configurations
which do not occur at thermal energies and that are expected
to scramble any product rotational alignment.
To support the above picture for the present reaction
dynamics, the BaO?A1?+? rotational alignment was calcu-
lated by using a simplified version11of the generalized
electron-jump DIPR-DIP model.12Both models are suitable
to account for the product rotational alignment in beam-gas
chemiluminescent reactions of the type A+BC→AB?+C
−+O, since the coupling between these two sur-
034304-10Rossa, Rinaldi, and Ferrero J. Chem. Phys. 132, 034304 ?2010?
where repulsive energy releases impulsively as products
separate. The latter would seem reasonable to assume here
on the basis of previous modeling by Stolte and coworkers45
of ?P2?J?·kˆ?? data for the chemiluminescent Ba?1S0?+N2O
reaction from nonrandom collision experiments. In combin-
ing another simplified DIPR theory with the angle-dependent
line of normals model, they found the experimental results to
be consistent with an impulsive and large ?of the order of an
eV? repulsive energy release. The DIPR-DIP model used
presently11incorporates three simplifications to the general-
ized formulation,12namely, that
the repulsive energy is released instantaneously, which
could be a good assumption here according to the mod-
eling results45on the Ba?1S0?+N2O reaction;
the reaction probability is unity for reagent impact pa-
rameters b?s?Rc?the results are thus independent of
Rc?, which may be the case only at the lowest ?Ec?
studied here where the system is less susceptible to a
failure of the charge transfer to occur;
the probability of collision between A and BC is isotro-
pic; this is likely a crude assumption, given the shape
of N2O, where intuitively reaction is expected to de-
pend upon approach angle.
=?3??J?·kˆ?2??/2, is made by identifying ?J?·kˆ?, i.e., the co-
sine of the angle between the product rotational angular mo-
mentum J? and k, with the projection of J? along z. The latter
is a function of only b and the Euler angles ??,?,?? describ-
ing the coordinates of BC, and it also depends on the repul-
sive release of energy R through the following reduced pa-
where mBand ?BCare the mass of B and the BC reduced
mass, respectively. R may be estimated by51
R = EA?B? − D?BC? − EA?BC?,
where EA?B? is the electron affinity of B, and D?BC? and
EA?BC? are the dissociation energy and the electron affinity
of BC, respectively. Calculation of ?P2?J?·kˆ?? then involves
a simple average of ?J?·kˆ?2over the distribution of J? about
z.11A further motivation for using this model is that it is
largely independent of any assignment of the BaO?emitters,
so the results of its application will remain valid also when
an improved spectroscopic characterization of the chemilu-
minescence spectra is available.
Table IV shows the results of a DIPR-DIP calculation
using R=1.99 eV, obtained from Eq. ?4? with EA?O?
=−2.23 eV. With the exception of the datum at ?Ec?
=4.82 eV, it is apparent that the DIPR-DIP model alignment
disagrees with both the magnitude and the velocity depen-
dence of the experimental alignment. The latter is not sur-
prising since the basic assumptions of the DIPR-DIP model
favor conversion of collision energy into polarization of the
product rotation perpendicular to k.12Instead, the present cal-
culations qualitatively show that dynamical factors, other
than repulsive release of energy, most likely time-scale con-
straints on the charge transfer and/or induced repulsive en-
ergy release, control the angular momentum disposal at the
hyperthermal ?Ec? studied. Clearly, further exploration of this
suggestion will require carrying out dynamical calculations
on accurate four-body PES. This is particularly true at the
hyperthermal collision energies studied here, where time-
scale constraints on charge transfer and/or induced repulsive
energy release may become dominant over repulsive energy
The substantial energy deficits Edand Etot−EBat all ?Ec?
studied ?see Table III? are by no means unique to the present
reaction. Indeed, large energy deficits, i.e., in the range of
0.5–3.7 eV, appear to be a common feature of group 2 metal
atoms plus N2O reactions even at thermal collision
energies.15,24,52–55This has been generally ascribed to vibra-
tional excitationof theN2
experimental56and theoretical49DEA investigations on the
isolated system N2O/N2O−instead suggest that much more
energy is transferred in N2rotation than both in its vibration
and in product relative translation. The latter explanation
would seem reasonable in the scenario of the present gas-
phase reaction given the apparent key role that is played by
N2O bending. Moreover, substantial energy released as N2
rotation and product translation is what could be expected50
from the above-suggested involvement of induced repulsive
energy release. This further suggests an interesting specula-
tion for the apparent lack of production of BaO excited states
at 4 eV from an energetic standpoint: That the N2product
takes too much of the total product energy, thus preventing
formation of these states. In any case, substantial energy
deficits here may well be the consequence of the fact that, at
all ?Ec?, the total product energy is sufficient to allow forma-
tion of nonemitting both BaO Rydberg states and products of
dissociation Ba+O+N2 as well as collisional ionization
The energy and angular momentum disposal in the
chemiluminescent reaction Ba?3P?+N2O were investigated
under hyperthermal beam-Maxwellian gas conditions at the
three specific average collision energies of 4.82, 5.85, and
7.47 eV. The fully dispersed polarized chemiluminescence
spectra revealed that the BaO?A1?+? product rotational
alignment is insensitive to its vibrational state ?? at ?Ec?
=4.82 eV and develops into an increasingly negative corre-
lation between ?P2?J?·kˆ?? and ?? as ?Ec? is raised. Such be-
havior is interpreted in terms of a mechanism involving a
short-range, partial electron transfer from Ba?3P? to N2O
which is constrained by the duration of the collision, so that
the reaction has a larger probability when the N2O has time
enough to bend during the collision. The latter in turn deter-
mines that, at any given ?Ec?, collinear reactive intermediates
are preferentially involved when the highest velocity compo-
nents of the corresponding collision energy distributions are
sampled. Moreover, the data at 4.82 eV suggest that a poten-
Collision energy effects on the Ba?3P?+N2OJ. Chem. Phys. 132, 034304 ?2010?
tial barrier to reaction which favors charge transfer to bent
N2O at chiefly coplanar geometries is operative for most of
the reactive trajectories that sample the lowest velocity com-
ponents. Such a barrier would arise from the relevant ionic-
covalent curve crossings occurring in the repulsive region of
the covalent potential Ba?3P?¯N2O?1?+?; from this cross-
ing the BaO?A1?+? product may be reached through mixings
in the exit channel with PESs leading most likely to the
spin-allowed b3? and a3?+products.
At all ?Ec?, the significant magnitude of the average
BaO?A1?+? rotational alignments obtained, and the high
BaO?A1?+? rovibrational excitations that derive from spec-
tral simulations of the unpolarized chemiluminescence spec-
tra both indicate that the reaction mechanism is predomi-
nantly direct, and consistently point to dynamic factors other
than time-scale constraints, most likely the development of
induced repulsive energy release, as the major responsible
for their variation with increasing ?Ec?. A simplified version
of the DIPR-DIP model does not give agreement with the
observed ?P2?J?·kˆ??avvalues. This may attributed to the fact
that the model oversimplifies the reaction dynamics, by ig-
noring the four-body PES. The present work suggests that it
will be necessary to carry out dynamical calculations on a
realistic PES for the reaction in order to account for the
?P2?J?·kˆ??avdata as a function of ?Ec?. This is particularly
true at the hyperthermal collision energies studied here,
where time-scale constraints on charge transfer and/or in-
duced repulsive energy release may become dominant over
the repulsive release of energy.
We thank CONICET, FONCYT, and ACC for financial
support. One of us ?M.R.? acknowledges a doctoral fellow-
ship from CONICET-Argentina.
1M. Polanyi, Atomic Reactions ?Williams and Northgate, London, 1932?.
2J. M. Mestdagh, B. Soep, M. A. Gaveau, and J. P. Visticot, Int. Rev. Phys.
Chem. 22, 285 ?2003?.
3K. C. Lin and A. González Ureña, Int. Rev. Phys. Chem. 26, 289 ?2007?.
4A. E. Wiskerke, S. Stolte, H. J. Loesch, and R. D. Levine, Phys. Chem.
Chem. Phys. 2, 757 ?2000?.
5M. De Castro, R. Candori, F. Pirani, V. Aquilanti, M. Garay and A.
González Ureña, J. Chem. Phys. 112, 770 ?2000? ?and Refs. 5 and 6
6M. A. Spence, W. R. Tomlinson, and M. R. Levy, Phys. Chem. Chem.
Phys. 3, 3622 ?2001? ?and references of the senior author therein?.
7M. A. Spence, W. R. Tomlinson, and M. R. Levy, Phys. Chem. Chem.
Phys. 3, 3610 ?2001?.
8M. Rossa, C. A. Rinaldi, and J. C. Ferrero, J. Chem. Phys. 127, 064309
9M. Rossa, C. A. Rinaldi, and J. C. Ferrero, J. Appl. Phys. 100, 063305
10M. Rossa, C. A. Rinaldi, and J. C. Ferrero, J. Appl. Phys. 105, 063306
11G. Ding, W. Yang, W. Sun, D. Xu, G. He, and N. Lou, Chem. Phys. Lett.
220, 1 ?1994?.
12M. G. Prisant, C. T. Rettner, and R. N. Zare, J. Chem. Phys. 81, 2699
13B. M. Miles and W. L. Wiese, At. Data 1, 1 ?1969?.
14R. C. Estler and R. N. Zare, Chem. Phys. 28, 253 ?1978?.
15J. W. Cox and P. J. Dagdigian, J. Chem. Phys. 79, 5351 ?1983?.
16C. Alcaraz, P. de Pujo, J. Cuvellier, and J. M. Mestdagh, J. Chem. Phys.
89, 1945 ?1988?.
17R. A. Gottscho, P. S. Weiss, R. W. Field, and J. G. Pruett, J. Mol. Spec-
trosc. 82, 283 ?1980?.
18R. A. Gottscho, J. B. Koffend, and R. W. Field, J. Mol. Spectrosc. 82,
19D. Bender, S. H. Schaefer, and E. Tiemann, J. Mol. Spectrosc. 116, 286
20N. Furio and J. G. Pruett, J. Mol. Spectrosc. 136, 120 ?1989?.
21H. Li, C. Focsa, B. Pinchemel, R. J. Le Roy, and P. F. Bernath, J. Chem.
Phys. 113, 3026 ?2000?.
22J. J. Reuther and H. B. Palmer, J. Chem. Phys. 77, 83 ?1982?.
23J. M. Mestdagh, P. Meynadier, P. de Pujo, O. Sublemontier, J. P. Visticot,
C. Alcaraz, J. Berlande, and J. Cuvellier, Chem. Phys. Lett. 164, 5
24M. de Castro, C. Rinaldi, L. M. Gamo, J. Cáceres, and A. González
Ureña, Phys. Chem. Chem. Phys. 2, 4115 ?2000?.
25J. G. Pruett and R. N. Zare, J. Chem. Phys. 62, 2050 ?1975?.
26S. E. Johnson, J. Chem. Phys. 56, 149 ?1972?.
27R. A. Gottscho, J. B. Koffend, R. W. Field, and J. R. Lombardi, J. Chem.
Phys. 68, 4110 ?1978?.
28Y. C. Hsu, B. Hegemann, and J. G. Pruett, J. Chem. Phys. 72, 6437
29A. L. G. Rees, Proc. Phys. Soc. London 59, 998 ?1947?.
30R. J. Le Roy, RKR1, University of Waterloo Chemical Physics Research
Report No. CP-425, 1992; R. J. Le Roy, LEVEL 7.0, University of Waterloo
Chemical Physics Research Report No. CP-642, 2000. The source codes
and manuals for both programs were obtained from the www site http://
31R. W. Field, G. A. Capelle, and M. A. Revelli, J. Chem. Phys. 63, 3228
32W. H. Parkinson, Proc. Phys. Soc. London 78, 705 ?1961?.
33CRC Handbook of Chemistry and Physics, Internet Version 2007, 87th
ed., edited by D. R. Lide ?Taylor and Francis, Boca Raton, FL, 2007?.
34G. Herzberg, Molecular Spectra and Molecular Structure ?Van Nostrand
Reinhold, New York, 1950?.
35J. Cuvellier, P. de Pujo, J. M. Mestdagh, P. Meynadier, J. P. Visticot, J.
Berlande, and A. Binet, J. Chem. Phys. 90, 7050 ?1989?.
36J. P. Visticot, C. Alcaraz, J. Berlande, J. Cuvellier, T. Gustavsson, J. M.
Mestdagh, P. Meynadier, P. de Pujo, and O. Sublemontier, J. Chem. Phys.
94, 4913 ?1991?.
37M. G. Prisant, C. T. Rettner, and R. N. Zare, J. Chem. Phys. 75, 2222
38M. A. Spence and M. R. Levy, J. Phys. Chem. 101, 7490 ?1997?.
39H. Jalink, G. Nicolasen, S. Stolte, and D. H. Parker, J. Chem. Soc.,
Faraday Trans. 2 85, 1115 ?1989?.
40Z. Wang, G. He, and N. Lou, Chem. Phys. Lett. 147, 116 ?1988?.
41D. J. Wren and M. Menzinger, Faraday Discuss. Chem. Soc. 67, 97
42NIST database, http://webbook.nist.gov.
43D. G. Hopper, A. C. Wahl, R. L. C. Wu, and T. O. Tiernan, J. Chem.
Phys. 65, 5474 ?1976?.
44A. González Ureña, Nuovo Cimento D 11, 1389 ?1989?.
45G. T. Evans, E. van Kleef, and S. Stolte, J. Chem. Phys. 93, 4874 ?1990?.
46D. H. Parker, H. Jalink, and S. Stolte, Faraday Discuss. Chem. Soc. 84,
47J. L. Magee, J. Chem. Phys. 8, 687 ?1940?.
48S. S. Batsanov, Inorg. Mater. 37, 871 ?2001?.
49H. U. Suter and T. Greber, J. Phys. Chem. B 108, 14511 ?2004?.
50A. M. G. Ding, L. J. Kirsch, D. S. Perry, J. C. Polanyi, and J. L.
Schreiber, Faraday Discuss. Chem. Soc. 55, 252 ?1973?.
51D. Herschbach, Faraday Discuss. Chem. Soc. 55, 233 ?1973?.
52C. D. Jonah, R. N. Zare, and Ch. Ottinger, J. Chem. Phys. 56, 263
53J. A. Irvin and P. J. Dagdigian, J. Chem. Phys. 74, 6178 ?1981?.
54P. J. Dagdigian, J. Chem. Phys. 76, 5375 ?1982?.
55J. W. Cox and P. J. Dagdigian, J. Phys. Chem. 86, 3738 ?1982?.
56M. Allan and T. Skalický, J. Phys. B 36, 3397 ?2003?.
034304-12 Rossa, Rinaldi, and Ferrero J. Chem. Phys. 132, 034304 ?2010?