Laser excited fluorescence study of reactions of excited Ca and Sr with water and alcohols: Product selectivity and energy disposal

Article abstract of DOI:10.1063/1.472457

Reactions of the metastable ³P_J⁰ states of Ca and Sr in atomic beams with H₂O, D₂O, and CH₃OH yielding ground electronic state products have been observed by laser excited fluorescence of MOH, MOD, and MOCH₃. The water reactions favor metal hydroxide Droductsjvhile methanol reactions favor methoxides. For SrOH product, spectral simulation of the B ²Σ⁺ -X ²S⁺ transition based on coupled harmonic-oscillator Franck-Condon factors was used to determine crude vibrational energy distributions in the bending and metal-stretching modes, and simulation of a higher resolution scan of excitation of the ground vibrational level gave some information about the rotational energy distribution in that level. While excitation of metal stretching and rotation were considerable and not too far from the predictions of a prior model, bending was significantly colder. Limited spectroscopic constants and severe spectral congestion have precluded other successful simulations.

Full text of DOI:10.1063/1.472457

Laser excited fluorescence study of reactions of excited Ca and Sr with water and
alcohols: Product selectivity and energy disposal
M. D. Oberlander and J. M. Parson
Citation: The Journal of Chemical Physics 105, 5806 (1996); doi: 10.1063/1.472457
View online: http://dx.doi.org/10.1063/1.472457
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/105/14?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Laser excitation spectroscopy of the B 1Σ+ u (O + u )←a 3Π g (O + g ) transition of Cd2: Vibrational analysis and
rotational structure
J. Chem. Phys. 105, 6771 (1996); 10.1063/1.472526
The highpressure range of the reaction of CH(2Π) with N2
J. Chem. Phys. 105, 5423 (1996); 10.1063/1.472383
Experimental investigation of binary nucleation rates of water–npropanol and water–nbutanol vapors by means of
a pextube
J. Chem. Phys. 105, 5168 (1996); 10.1063/1.472361
What’s new in the proton transfer reaction from pyranine to water? A femtosecond study of the proton transfer
dynamics
AIP Conf. Proc. 364, 333 (1996); 10.1063/1.50210
Optical spectroscopic studies on Penning ionization and ionmolecule reactions at thermal energy by using
flowing afterglow
AIP Conf. Proc. 205, 342 (1990); 10.1063/1.39248
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

Laser excited fluorescence study of reactions of excited Ca and Sr
with water and alcohols: Product selectivity and energy disposal
M. D. Oberlander and J. M. Parson
Department of Chemistry, The Ohio State University, Columbus, Ohio 43210
͑Received 12 January 1996; accepted 1 July 1996͒
0
3
Reactions of the metastable PJ states of Ca and Sr in atomic beams with H2O, D2O, and CH3OH
yielding ground electronic state products have been observed by laser excited fluorescence of MOH,
MOD, and MOCH3. The water reactions favor metal hydroxide products while methanol reactions
2
ϩ
2
ϩ
favor methoxides. For SrOH product, spectral simulation of the ˜B ⌺ –˜X ⌺
transition based on
coupled harmonic-oscillator Franck–Condon factors was used to determine crude vibrational energy
distributions in the bending and metal-stretching modes, and simulation of a higher resolution scan
of excitation of the ground vibrational level gave some information about the rotational energy
distribution in that level. While excitation of metal stretching and rotation were considerable and not
too far from the predictions of a prior model, bending was significantly colder. Limited
spectroscopic constants and severe spectral congestion have precluded other successful simulations.
© 1996 American Institute of Physics. ͓S0021-9606͑96͒00638-1͔
INTRODUCTION
ecules in electronically excited states under single-collision
conditions have been reported,1,13 although as yet there has
been no attempt to simulate these spectra. This may be due
to the difficulties in resolving the vibrational structure of
these electronic transitions owing to the small change in ge-
ometry associated with excitation of the nonbonding elec-
This report presents the results of beam-gas studies of
the reactions of ground and electronically excited calcium
and strontium with water and some alcohols using laser ex-
citation spectroscopy for the detection of ground state prod-
ucts. Insight into the dynamics of a chemical reaction can be
obtained from a determination of the nascent distribution of
product internal degrees of freedom following a reactive col-
lision. Electronic chemiluminescence1 and laser excitation2
are powerful and complementary techniques for obtaining
information about the nascent internal state distributions of
electronically excited and ground state products, respec-
tively. In order to extract the nascent internal state distribu-
tion, a simulation procedure is generally required whereby
the unknown relative vibrational and rotational populations
of the product are adjusted until agreement of the observed
and a theoretically calculated spectrum is optimized. Ex-
amples of such calculations for diatomic molecules are well
established in the literature;1 however, simulations of poly-
atomic spectra are not commonly found in the dynamics lit-
erature.
A major focus of this work has been the analysis of the
energy disposal in a metal monohydroxide product, specifi-
cally SrOH. As high resolution laser spectroscopic tech-
niques provide an increasingly accurate characterization of
the internal structure of these molecules, more detailed stud-
ies of the dynamics of polyatomic molecule formation are
likely to become prevalent. Over the last several years spec-
troscopists have devoted an increasing amount of attention to
the alkaline earth monohydroxides.3–12 The heavier ͑Ca, Sr,
and Ba͒ Group 2 hydroxides have been found to have a lin-
ear equilibrium structure in the ground and low lying excited
states, and the promotion of a metal centered nonbonding
tron, which leads to strong ⌬ ϭ0 sequences. Additional
vi
requirements associated with triatomic simulations involve
the calculation of band intensities.
Recent high resolution spectroscopic work on CaOH6–9
and SrOH,11,12 which includes extended studies of vibra-
tional structure, is opening the way to make reasonable
progress toward the simulation of some electronic spectra of
these molecules. The work on CaOH has addressed the
2
2
ϩ
A ⌸–X ⌺ system,6–9 which exhibits a Fermi resonance
with additional complications due to the Renner–Teller in-
teraction in the excited state. While free of the complexities
˜ ˜
2
˜
associated with the orbitally degenerate A ⌸ state, the
2
ϩ
˜
CaOH B ⌺ electronic state has received little attention
since the rotational analysis of Bernath and co-workers.4,5
Many of the complications associated with the analysis
˜
of the CaOH A electronic state are also expected for the
˜
SrOH A state, which has received additional attention
recently.12 In short, the data presently available in the
literature,10,11 combined with some additional information on
the vibrational structure available from studies carried out in
our laboratory, suggest that the SrOH ˜B–˜X system is most
amenable to reliable simulation.
A crossed molecular beam study of the reaction of elec-
tronically excited (6s5dϪ1D2) barium with water and
methanol led Davis et al.14 to propose a dynamical mecha-
nism which involves metal atom interaction at the oxygen
center followed by hydrogen atom migration to the metal to
form a HBaOH or HBaOCH3 insertion intermediate. Decay
of these intermediates leads to barium hydroxide or barium
methoxide. In their work, product translational energy and
angular distributions were measured, but no information on
2
ϩ
2
˜
˜
electron,
ns␴(X ⌺ )
to
(nϪ1)d␲(A ⌸)
or
2
ϩ
˜
(nϪ1)d␴(B ⌺ ), leads to electronic spectra which are
analogous to those of the alkaline earth halides.3
Studies of reactive collisions which produce these mol-
5806
J. Chem. Phys. 105 (14), 8 October 1996
0021-9606/96/105(14)/5806/11/$10.00
© 1996 American Institute of Physics
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
5807
TABLE I. Thermochemistry ͑Ref. 22͒ of reactions of Group 2 metals with
oxygen containing molecules, and summary of observed ground state prod-
ucts.
the product internal state distributions was obtained. Evi-
dence for the existence of such Group 2 metal insertion prod-
ucts as HMOH have been known as a result of matrix isola-
tion work15 where a shift in the bending mode frequency of
water was observed upon codeposition with the metal. It was
found that irradiation of the metal–water adduct results in
oxidative insertion of the metal into the O–H bond, which
occurs at lower radiate energies for the heavier metals.
De Pujo et al.16 have examined chemiluminescent reac-
tions of Ba(1P1) with water and alcohols. Their conclusions,
like those of Davis et al.,14 suggest that reaction of the elec-
tronically excited metal induces hydrogen atom migration
leading to electronically excited BaOH and BaOR products,
respectively. Esteban et al.,17 however, have reported CaOH
A ⌸→X ⌺ and B ⌺ →X ⌺ chemiluminescence re-
sulting from reaction of Ca(1D2) with methanol. Absolute
chemiluminescence cross sections have been reported for
these reactions, as well as for reaction with ethanol.
Support for metal insertion intermediates comes from
theoretical calculations of the reactions of nickel with
water,18 and aluminum with both water and ether.19 The
theoretical study of aluminum with water and ether indicates
that barriers to metal insertion in bonds such as C–O or O–H
correlate with electron affinities of unoccupied orbitals local-
ized at such bonds, and these ideas are expected to apply also
to Group 2 metal reactions with water and methanol.
Reaction
⌬H00 ͑kJ/mol͒ Observationc,d
a
1
˜
˜
Ca( S0)ϩCH3OH→X CaOCH3ϩH
29
Ϫ29
107
—
N
—
—
Y
N
N
Y
Y
—
N
—
—
Y
N
N
Y
Y
1
Ca( S0)ϩCH3OH→X CaOHϩCH3
Ca(1S0)ϩH2O→X CaOϩH2
1
˜
Ca( S0)ϩH2O→X CaoHϩH
88
a
3
˜
˜
Ca( P2)ϩCH3OH→X CaOCH3ϩH
Ϫ154
Ϫ212
Ϫ76
Ϫ95
Ϫ86
32
Ϫ27
63
89
Ϫ146
Ϫ205
Ϫ115
Ϫ89
Ϫ80
3
Ca( P2)ϩCH3OH→X CaOHϩCH3
Ca(3P2)ϩH2O→X CaOϩH2
3
˜
˜
Ca( P2)ϩH2O→X CaOHϩH
b
a
3
Ca( P2)ϩD2O→X CaODϩD
1
˜
˜
Sr( S0)ϩCH3OH→X SrOCH3ϩH
1
Sr( S0)ϩCH3OH→X SrOHϩCH3
Sr(1S0)ϩH2O→X SrOϩH2
2
2
ϩ
2
ϩ
2
ϩ
˜ ˜
˜
˜
1
˜
Sr( S0)ϩH2O→X SrOHϩH
a
3
˜
Sr( P2)ϩCH3OH→X SrOH3ϩH
3
˜
Sr( P2)ϩCH3O→X SrOϩCH3
Sr(3P2)ϩH2O→X SrOϩH2
3
˜
Sr( P2)ϩH2O→X SrOHϩH
b
3
˜
Sr( P2)ϩD2O→X SrODϩD
aAssumes D00͑CH3O–H͒ϭ436.8 kJ/mol and D00͑M–OH͒ϭD00͑M–OCH3͒.
bAssumes D00͑M–OH͒ϭD00͑M–OD͒.
cRefers to the results of studies utilizing laser excited fluorescence detection
of the ground state product indicated by X. Y:Reaction observed, N:Reac-
tion not observed, —:Reaction not observed and not expected to be ob-
served, based on thermochemistry.
dA reaction specifying the metal 3P2 state as reagent is intended to be
representative of only one of a number of potentially reactive metastable
states of the metal which would explain the indicated observation. These
experiments are carried out under conditions where the metal beam is dis-
charged, but the source does not selectively excite a single electronic state
of the metal.
EXPERIMENTAL PROCEDURES
Laser excited fluorescence spectra were measured in a
beam-gas configuration described previously20 using a Lu-
monics Excimer-500 XeCl excimer laser and a Lumonics
Hyperdye-300 dye laser. The excitation spectra were mea-
scuitroend͒ udsyinesg wRhhiocdhamlaisneei5n9t0h,eKSitroOnHR˜Bed–6˜X20anadndCDaOCHM˜A͑E–x˜X-
nate electronically excited state of the metal produced is be-
lieved to be 3P0, with a typical fractional beam excitation of
approximately 10Ϫ2 21
In experiments where the metal beam
.
was discharged, long or short pass filters were used to block
the continuous atomic emission at the discharge source from
reaching the detector. In these instances, when necessary, the
recorded laser excitation spectra were corrected for transmis-
sion of the spectral filter.
The reagent metals, methanol, D2O, and distilled H2O
were used as provided by the supplier without further puri-
fication. In some instances condensed phase oxidants H2O,
D2O, and methanol were maintained at ice temperatures to
prevent condensation in the gas lines.
spectral regions. The laser wavelength was periodically
checked using optogalvanic calibration from a neon filled
hollow cathode and by excitation of various atomic transi-
tions whose lower state was populated in the discharged
metal beam.
Two photomultiplier tubes were utilized for the laser ex-
citation work, an EMI 9816B with S20 photocathode or an
RCA C31034A phototube. When the variation of the photo-
cathode response was significant over the wavelength range
of the measured spectrum, the response reported by the
manufacturer was used for correction of the raw data. Laser
excitation spectra were recorded in an analog mode, with an
integration gate width of approximately 90 ns, using a Stan-
ford Research System model 250 gated integrator and an
associated model 245 computer interface analog-digital con-
verter.
An effusive molecular beam was generated by evapora-
tion of calcium or strontium metal in a cylindrical oven body
which was surrounded by a resistively heated tungsten mesh
radiator. Electronically excited alkaline earth metal atoms
were produced by applying a dc potential to the electrically
insulated oven body and placing a grounded wire loop in the
beam path near the front of the oven orifice. The predomi-
RESULTS
Table I gives the thermochemistry22 of reactions of cal-
cium and strontium with various reagents which may yield
metal oxide, monohydroxide or alkoxide. The remark under
the heading labeled ‘‘observation’’ gives a qualitative sum-
mary of the observed reaction channels as detected by laser
excited fluorescence.
2
3
0
3
0
˜
*
*
M ؉CH3OH˜MOCH3 X A1؉H, M ؍Ca( P ), Sr( P )
The electronic transitions for the Group 2 metal alkox-
ides closely resemble those of the metal monohydroxides,
owing to the metal centered nature of the electronic transi-
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

5808
M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
˜ ˜
FIG. 2. Comparison of the laser excitation spectrum of the A–X transition
of CaOH ͑lower trace͒ produced from the reaction of discharged calcium
with H2O, with that of CaOCH3 produced from the reaction of discharged
calcium with CH3OH. The spectra are displayed so that they have the same
area.
FIG. 1. Comparison of the laser excitation spectrum observed for the reac-
tion of discharged calcium with CH3OH ͑lower trace͒ with that observed for
the reaction of Ca(1S0) with H2O2 . In the lower trace the spectral features
are in good agreement with the vibrational assignments reported by Worms-
2
2
˜ ˜
becher et al. ͑Ref. 23͒ for the B A–X A transition of CaOCH3 .
when compared to Fig. 2 of Wormsbecher’s paper.23 This
could be explained by the authors’ lack of correcting the
recorded spectra for laser power variations, or due to differ-
ences in the internal state distribution of ground state prod-
ucts. Nevertheless, based on the observed wavelength shift,
especially in the off-diagonal sequence structure, there is
little doubt that the carrier of the observed spectra for the
discharged calcium with methanol reaction is due to
CaOCH3.
tion in these singly ionic radicals. Wormsbecher and
˜ ˜
2
2
Suenram23 have observed the A E–X A1 and
2
2
˜ ˜
B A1 –X A1 transitions of calcium, strontium and barium
methoxide in chemiluminescent reactions of the ground state
metal with methylnitrite, CH3O–NO, at pressures of 1–3
Torr. They also measured low resolution laser excitation
spectra and reported some vibrational frequencies denoting
␯4 as the M–OCH3 stretch, ␯8 as the M–O–CH3 bend and ␯
3
as the MO–CH3 stretch.
Figure 2 also shows a comparison of the CaOCH3
2
˜ ˜
2
Only upon discharging the calcium or strontium beam
was a detectable reaction observed with methanol to produce
the corresponding ground state metal methoxide product un-
der the single-collision conditions of the present experi-
ments. The lower trace in Fig. 1 shows the laser excited
A E–X A1 spectrum from the reaction of discharged cal-
2
cium with methanol, with the CaOH ˜A ⌸–˜X ⌺
2
ϩ
spectrum
observed in the reaction of discharged calcium with water.
Similar conclusions about the major products can be drawn
based on Fig. 1, which compares the laser excitation spec-
2
2
˜ ˜
2
ϩ
2
ϩ
trum ͑upper trace͒ of the CaOH ˜B ⌺ –˜X ⌺
B A1 –X A1 spectrum of CaOCH3. The observed spectrum
shows some vibrational structure, the most intense features
involving the zero sequence of the bending mode, ␯8, and
these peak positions agree well with those reported by
transition
obtained for the reaction of Ca(1S0) with hydrogen
peroxide.21 Discharged strontium was also observed to react
with methanol, once again yielding metal methoxide product.
͑See Fig. 3͒. Here, the qualitative features of the laser exci-
tation spectrum observed in the 600 nm region match those
Wormsbecher.23 Noting that the change in ␯ is reported to
4
for the ˜B–˜X transition, the ⌬
be only about 3 cmϪ1
4ϭ0
v
2
2
˜ ˜
sequence will not spread out very rapidly, and therefore the
of the SrOCH3 B A1 –X A1 spectrum reported by
Wormsbecher,23 and are readily distinguished from known
comparatively sharp peaks observed for the ⌬ 8ϭ0 se-
v
for the ˜B–˜X transition͒ may indicate
quence ͑⌬␯8ϭ24 cmϪ1
2
ϩ
2
ϩ
˜ ˜
B ⌺ –X ⌺ SrOH spectra in this wavelength region. Due
to the overlapping nature of the electronic spectra of MOR
and MOH, ground state MOH formation cannot be com-
pletely ruled out in the methanol reaction, but there is no
evidence to suggest that MOH formation is a major channel.
little excitation of the Ca–OCH3 mode vibrational distribu-
tion or a cold rotational distribution. Like the monohydrox-
ide electronic spectra, the Franck–Condon factors are not
expected to fall off too rapidly in either the bending or metal-
stretching zero sequences.
The upper trace in Fig. 2 shows the laser excited
2
؉
3
0
˜
*
*
M ؉H2O(D2O)˜MOH(D) X ⌺ ؉H(D), M ؍Ca( P ),
2
2
˜ ˜
A E–X A1 spectrum of CaOCH3. Little vibrational struc-
ture is apparent in any of the sequences and the spin–orbit
splitting is hardly discernable. Some disparity in the peak
Sr(3P0)
Only electronically excited metal atoms were observed
position of the ⌬ ϭϩ1, ⌬ 8ϭ0 sequence is apparent
v4
to react with water ͑D2O͒ to produce ground state MOH
v
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
5809
2
ϩ
2
ϩ
SrOH ˜B ⌺ –˜X ⌺
transition following a methodology
described by Sharp and Rosenstock.26 Experimentally de-
rived harmonic frequencies may be used to calculate a har-
monic potential using a classical Wilson GF analysis.27 Here
a diagonal force constant matrix, F, was used to approximate
a valence coordinate potential for both the ground and ex-
cited state of the linear SrOH molecule. The matrix GF is
then block diagonal and it follows that the L matrix, which
relates the internal coordinates Ri to the normal coordinates
Qi , is also block diagonal. Knowledge of the matrix L for
each electronic state allows a linear transformation to be de-
fined which relates the normal coordinates of the electronic
states involved in the transition. Complete details of the cal-
culations can be found elsewhere.21
A block diagonal L matrix allows the overlap integrals
for the stretching mode transitions to be calculated separately
from those for the bending mode transitions. The overlap
integrals for the bending mode transitions were calculated
using the polar form of the harmonic oscillator wave func-
tion: ␺v ,l(␳,␸).28 The vibrational angular momentum quan-
FIG. 3. Comparison of the laser excitation spectrum observed for the reac-
tion of discharged strontium with CH3OH ͑lower trace͒ with that observed
for the reaction of discharged strontium with H2O. The vibrational assign-
ments shown were reported by Wormsbecher et al. ͑Ref. 23͒ for the
2
tum number, l, takes on the values
͉
l
͉
ϭ0, 2, 4,... , for
v2
v2
even, and
͉
l
͉
ϭ1, 3, 5,... 2 , for
odd, where is the
v
v2
v2
2
2
˜ ˜
B A–X A transition of SrOCH3 .
quantum number of the bending vibration. The harmonic os-
cillator wave functions, ␺ , for the stretching modes ␯1 and
v
i
␯ with quanta
and 3 , respectively, are given by Wilson,
v
v1
3
͑MOD͒. The laser excitation spectrum of CaOH and CaOD
were measured, but few distinguishing features are observed
in the spectra. Under the same beam discharge and
scattering-gas pressure conditions, the ratio of the integrated
signal for the two calcium reactions was found to be,
Decius, and Cross.27 The Franck–Condon factor, qv ,v , for
Ј Љ
l
l
Љ
Ј
Љ
the transition ( Ј , , Ј) Ϫ ( Љ ,
v1 v2 v3 v1 v2
from
,
3) was calculated
v
*
*
v
2
ϩ
2
ϩ
ϭ1.0Ϯ0.2. A comparison of the ˜B ⌺ –˜X ⌺
Љ
Љ
qv ,v
ϭ
␺ ␺ ␺ ␺ dQ1 dQ3
ͯ
͵ ͵
͵ ͵
v
v
Љ
3
Љ
SD2O/SH2O
Ј Љ
v
Ј
Ј
3
1
1
laser excitation spectra of SrOH(D) for reaction of dis-
charged strontium with H2O and D2O was also made. There
is a slight wavelength shift, and the ratio, SD O/SH O, for
2
*
ϫ
␺
␺
␳ d␳ d␸ .
͑1͒
ͯ
v
2Љ,l
v
2Ј,l
2
2
strontium was found to be 1.3Ϯ0.2. Thus the isotopic sub-
stitution does not appear to have a very large effect on the
bulk reactive cross sections for H2O and D2O.
For the electronic transition there is no restriction on
⌬
(kϭ1,3) for either of the totally symmetric vibrations,
vk
but the nontotally symmetric vibration, ␯2, is restricted to the
Alcohols have thermochemically accessible channels
leading to ground state MO product, even for the reaction of
ground state metal.24 While only the case of barium and
ethanol was examined carefully, no detectable reaction was
observed, although the reaction of Ba(1S0) and CO2 was
transitions: ⌬ 2ϭ0,Ϯ2,Ϯ4,..., assuming that there is no
v
change in the vibrational angular momentum in the elec-
tronic transition: ⌬lϭ0. The present calculation neglects vi-
bronic coupling which may induce electronically forbidden
components of the 2⌺ϩ –2⌺ϩ electronic transition. Such
⌬KÞ0 transitions have been observed11 in the SrOH
1
successfully used as a reference signal of X ⌺ϩ BaO under
single collision conditions.25 A reasonable conclusion is that
the reactive cross sections for the channels,
2
ϩ
2
ϩ
˜ ˜
B ⌺ –X ⌺ system, presumably due to mixing of vi-
2
ϩ
2
˜
˜
bronic states associated with the ͑nominal͒ B ⌺ and A ⌸
electronic states.
1
M(1S0)ϩCH3CH2OH→MOX ⌺ϩϩCH3CH3, are below
our detection limit. This was also the conclusion drawn by
Davis et al.14 in their crossed molecular beam study of the
reaction of BaϩCH3OH→BaOϩCH4 employing mass spec-
trometric detection.
Considering allowed electronic transitions only, where
2
⌬lϭ0, the simulation of the SrOH ˜B ⌺ –˜X ⌺
ϩ
2
ϩ
system
requires intensity calculations for two types of rotational
band structures. The rotational line strength factors for
2⌺ϩ –2⌺ϩ vibronic transitions were assumed to be that for
the corresponding 2⌺ϩ –2⌺ϩ transition of a diatomic
molecule.29
THE TRIATOMIC SPECTRAL SIMULATION
PROCEDURE
For allowed electronic transitions between vibrational
levels with lϾ0, the band structure corresponds to that found
in transitions between degenerate electronic states of di-
atomic molecules.30,31 The correspondence is such that the
quantum number of vibronic angular momentum K, where
2
Approximation of intensities in a ⌺؉ –2⌺؉ electronic
transition
Harmonic oscillator wave functions were used to calcu-
late approximate Franck–Condon factors ͑FCFs͒ for the
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

5810
M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
TABLE II. Partial listing of spectroscopic constantsa utilized in the SrOH
2
ϩ
2
ϩ
˜
B
˜
–X
⌺
⌺
simulations.
2
ϩ
2
ϩ
˜
B
˜
⌺
X ⌺
T000
16 377.505
536.30b
399.785
3847.
529.443b
360.357
3847.
0
␻
3
0
␻
2
␻
1
x033
x022
g22
Ј
Ϫ2.57b
Ϫ7.124
8.176
Ϫ2.451b
Ϫ4.357
7.589
Ϫ1.33c
x023Ϫx23
0
Љ
Be
0.254 21
0.250 721
␣1 ͑ϫ104͒
␣2 ͑ϫ103͒
␣3 ͑ϫ103͒
3.96d
1.05
1.57
6.4
3.96d
0.642
1.44
3.2
␥
␥
͑ϫ105͒
22
͑ϫ105͒
Ϫ6.8
Ϫ5.1
FIG. 4. The observed ͑lower trace͒ and simulated laser excitation spectrum
ll
2
ϩ
2 ϩ
˜
˜
of the SrOH B ⌺ –X ⌺ transition for the reaction of discharged stron-
tium with H2O. The spectral region at wavelengths shorter than 6093 Å was
not simulated since the bands in this region are perturbed.
re ͑Sr–O͒ ͑Å͒
re ͑O–H͒ ͑Å͒
2.088
0.946
2.102
0.945
aExcept where noted, all constants are in cmϪ1. Molecular constants taken
from References 10 and 11.
bEstimated from data obtained in present work, and unpublished results of
Coxon and Presunka ͑Ref. 11͒.
Kϭ
͉Ϯ⌳Ϯl͉, in the linear triatomic corresponds to the quan-
tum number ⌳ in the diatomic molecule. For the SrOH
cEstimated from data obtained in the present work.
dFixed at ␣1 value for RbOH reported in Ref. 33.
2
ϩ
⌺
electronic states it follows that Kϵl ͑there exists an
l-type doubling͒. Rotational line strength factors31 in
the limit of Hund’s case ͑b͒ coupling were assumed in
the calculation implementing the correspondence
⌳͑diatomic͒→K(ϵl).
Unpublished results of Presunka and Coxon11 which re-
2
ϩ
˜
port SrOH X ⌺ vibrational term values, G0 , for the
͑0,00,1͒ and ͑0,00,2͒ levels were also fitted to a quadratic
Rotational and vibrational term series in ⌺؉
electronic states
2
0
polynomial in
to obtain estimates of ␻ and x033 for the
v3
3
ground electronic state. Combining these results with the fit
to band origins of the
⌬
3ϭ0 sequence (0,00, Ј)
v
v3
Although the Franck–Condon factors were calculated
from the harmonic oscillator approximation, the vibrational
level structure was calculated for each electronic state using
the known anharmonic vibrational constants and the follow-
ing truncated term series expression:32
2
ϩ
0
˜
–(0,0 , Љ , yields estimates for the B ⌺ electronic state
v3
constants: ␻ and x033, which are given in Table II.
0
3
2
ϩ
The rotational transitions among the
states have been calculated using the term series expressions
⌺
electronic
and 2⌸ vibronic states,11 however, the
2
ϩ
for case ͑b͒
⌺
higher order corrections Hv , ␥Dv , and qvD were neglected. The
l
0
G0
,
,
͒ϭ
␻
ϩ
i vi
x0 2ϩg22l2.
iivi
͑2͒
͑v1 v2 v3
͚
͚
2
i
i
vibronic ⌸ rotational term series expression was also ap-
plied to vibronic states of ⌬, 2⌽ and larger values of K.
2
Except for x023, the off-diagonal anharmonic constants xi0j,
and higher order correction terms are neglected due to a lack
The uncertain dependence of the rotational constant Bv and
the spin–rotation interaction constant ␥ on is likely to be
a large source of error for the higher vibrational levels which
have not been spectroscopically characterized.
The vibrational dependence of the rotational constants
has been discussed by Lide and Matsumura33 and by
Herzberg.32 Table II gives ␣1, ␣2, and ␣3, the vibration–
rotation interaction constants for the OH-stretching, bending,
and metal-stretching modes, respectively, and the constants,
vi
˜ ˜
v
of knowledge of these constants for the B and X electronic
states of SrOH. The constants ␻ and x0ii for metal stretching
0
i
and bending are obtained from fits to the observed level
structure10,11 using this truncated series, and xi0i is assumed to
be approximately equal to xii .
The simulations are limited to the (100,2vv,3vv) sequence
region, since only the difference of the anharmonic constant
0
21
˜
˜
x23 for the B and the X state has been estimated. In calcu-
␥
22 and ␥ll ͑not to be confused with the spin–rotation inter-
Ј
lating band origins for the ⌬ ϭ0, ⌬ 3ϭ0 sequence of the
v2
v
2 ϩ
˜ ˜
2
ϩ
action constant͒ which are higher order corrections33 for the
SrOH B ⌺ –X ⌺
electronic transition, the term:
bending modes.
0
3⌬x0
ϭ
3(x0 Ϫ x ), was added to the sum T000
Ј
Љ
v2v
v2v
23
23
23
ϩ G0Ј Ϫ G0Љ , where G0 is given by Eq. ͑2͒. The difference
0
⌬x023 ϭ x023 Ϫ x23 was determined from fits of estimated vibra-
Ј
Љ
Parametric form of the populations
tional band origin frequencies derived from band head posi-
tions observed in the
⌬
3ϭ0 sequences, (0,11, Ј)
v v3
In the absence of a specialized theoretical model to pre-
dict the internal state distribution of products, the prior
distribution,34,35 which assumes no dynamical constraints
–(0,11, Љ), and (0,00, Ј)–(0,00, Љ) shown in the laser ex-
citation spectrum of SrOH ˜B ⌺ –˜X ⌺
v3 v3 v3
2
ϩ
2
ϩ
in Fig. 4.21
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
5811
other than energy conservation, can serve as a basis for char-
acterizing and comparing observed product distributions for
various reactions.
Due to computer memory limitations associated with
simulations over extensive vibrational structure, the popula-
tion of the F1 and F2 components are not calculated sepa-
rately as a function of J but rather the relative populations
are calculated using the approximate rotational energy ex-
pression, ENϭBv(N(Nϩ1)Ϫl2), with frotϭEN/E. EN is
within an additive constant of the mean of the energy of the
F1 and F2 components for a particular N, neglecting cen-
trifugal distortion and higher corrections. For a transition
originating from a particular J, the population of the F1 com-
ponent is assigned to the rotation-vibration population
P͑v,N͒ while that for the F2 component with the same J is
assigned to P͑v,Nϩ1͒. Furthermore, in the prior distribu-
tions discussed above, where applicable, the correct rota-
tional degeneracy factor (2Jϩ1) is approximated by (2N
ϩ1). However, in laser scans where more detail in the
rotational head structure is measured such as in Figs. 4 and
10, the accurate rotational term series expressions and cor-
rect degeneracy factors are used in calculating the F1 and F2
populations, which were stored separately.
For a reaction yielding linear triatomicϩatom products,
with available energy E, in which the vibrations and rota-
tions of the triatomic are specified the prior distribution is
represented by P0͑v, J͒. For the same reaction, summing
P0͑v, J͒ from JϭJmin to J ͑v͒, the highest energetically ac-
*
cessible J in the vibrational level v, in the rigid rotor ap-
proximation yields the prior distribution of vibration for the
triatomic
P0 v͒ϳBϪ1 EϪE
͒
3
,
͑3͒
3/2
l
2
͑
͑
v
,v ,v
v
1
where Bv is the rotational constant of the triatomic in vibra-
l
tional level (
,
, 3) whose energy is Ev . If there are
v1 v2 v
dynamical constraints on the system, other than energy con-
servation, the system will exhibit a surprisal. In the simula-
tion, the product number density n͑v,J͒ is taken as
n v,J͒ϳP v͒P v,J͒,
͑4͒
͑
͑
͑
where P͑v, J͒ and P͑v͒ are parametric functions given by
2
؉
˜
Spectral simulation and energy disposal in X
⌺
␭
f
r
P v,J͒ϳP0 v,J͒e
͑5͒
͑6͒
1
͑
͑
2
؉
3
0
SrOH Sr( P )؉H2O˜SrOH X 2⌺؉؉OH
˜
˜
and B
⌺
and,
Figure 4 shows a simulated spectrum which resolves
some of the ␯ mode structure in the lowest bending mode
levels of ground state SrOH observed in the reaction of dis-
charged strontium with water. Owing to the large change in
the spin–rotation splitting constant upon electronic excita-
tion to the B ⌺ state, the P11 and P22 heads are formed at
significantly different J values. Figure 5 shows a comparison
of the observed rotational distribution in the F1 and F2 levels
of the ͑0,00,0͒ level of the ground electronic state with both
the prior and a Boltzmann. The prior form of the rotational
distribution, P͑v,J͒, is generally found to be more suitable
than the Boltzmann since the vibrational dependence of the J
distribution is taken into account by virtue of energy conser-
vation. The (0,2l,0) levels are known to exhibit an l-type
resonance11 and simulation of this region of the spectrum is
expected to deviate from observation since no perturbations
were included in the calculation procedure. Figure 6 presents
l
2
͑x
f
ϩ͑x ϩx
f
͒f l ϩx l͒
P v͒ϳP0 v͒e
3
.
3
1
v
2
3
v
v
4
͑
͑
2
Here, ␭i , xi are surprisal parameters which were varied to fit
the experimental spectrum using a nonlinear least squares
minimization algorithm.36 The quantum number l ranges
2
ϩ
˜
over all its allowed values for a given 2 . fr , fv , fv and
v
3
2
fl are the fractions of available energy E in rotation, metal-
stretching, bending, and vibrational angular momentum, re-
spectively, and flv ϭ fv ϩ fl . Due to lack of knowledge of
2
2
all vibrational constants, fv and fv were approximated as
3
2
0
2
f ϭ ␻ ϩx0 ͒/E,
iϭ2,3͒,
͑
͑7͒
͑
i vi iivi
v
i
while flϭg22l2/E.
In the simulation, the population of the OH-stretching
mode ␯ is assumed to be nonzero only in its ground vibra-
1
˜ ˜
tional state, and undergoes the electronic transition with
the observed and calculated SrOH B–X laser excitation
⌬
1ϭ0. This assumption is due to the fact that there has
spectrum of the (100,2vv,3vv) sequence, and the large vertical
v
been no direct experimental observation in SrOH of ␯1,
which is most closely associated with OH stretching. For the
purpose of approximating Franck–Condon factors, a fre-
ticks indicate the bending mode band origins, calculated with
lϭlmin
.
Since the ␯ mode has been fixed in its ground vibra-
1
quency of 3847 cmϪ1 was assumed for ␻1 for both the X and
tional level, the relative vibrational populations are displayed
in three dimensions as in Fig. 7. Such plots, which represent
number density, have been summed over the l distribution
˜
˜
B states of SrOH. This is based on an estimate of ␻ for
1
CaOH X ⌺ reported by Jarman and Bernath.37 The justi-
2
ϩ
˜
˜ ˜
fication is based on the fact that the B–X electronic transi-
appropriate for the particular bending mode quantum
.
v2
tion in SrOH involves excitation of a metal centered non-
bonding 5s␴ electron to a 4d␴ orbital. Consequently, one
The observed distribution and that of the prior have been
separately normalized to the same area. As in treating the
rotational distribution, insufficient spectral resolution limits
our knowledge of the distribution in l, and we have chosen to
average over it in displaying the relative vibrational popula-
tions. It should also be noted that bending mode transitions
that are observed near the limits of the measured spectral
region may have high l transitions which are outside the
would expect ␻ to be similar for each electronic state and
1
the observed spectrum is probably relatively insensitive to
the relative vibrational population in this mode anyway. In
addition the ␯ mode is not expected to be highly excited in
1
the reactions of interest here since the newly formed bond
most directly involves the ␯ and ␯ modes.
3
2
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

5812
M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
0
2 ϩ
˜
FIG. 5. Rotational distribution in the ͑0,0 ,0͒ level of X ⌺ SrOH for the
simulation in Fig. 4. The circles represent the observed F1 distribution,
while the boxes represent that for F2 . The solid line is a Boltzmann for the
F1 levels at 8050 K and the triangles are the prior distribution for the F1
FIG. 6. The observed ͑lower trace͒ and simulated laser excitation spectrum
2
ϩ
2 ϩ
˜
˜
of the SrOH B ⌺ –X ⌺ transition for the reaction of discharged stron-
tium with H2O. The P22 band heads are shown for the ⌬ 3ϭ0 sequences in
v
the lowest two bending modes. The band origins of the ⌬ 2ϭ0 sequence
v
levels with Etotϭ8700 cmϪ1
.
with lϭlmin proceed through a ‘‘head-of-origins’’ at 2ϭ7.
v
observed spectral region. Since all population analysis is re-
stricted to the set of observable internal states, as defined by
the spectral region, some summations over l are terminated
appropriate summations, as shown in Figs. 8 and 9. Here the
observed vibrational distributions are seen to decay mono-
tonically in the metal-stretching mode, while the observed
with lϽ 2 . This may lead to some irregularities in the dis-
v
played distributions for these levels.
bending mode population shows a maximum at low 2 . The
v
The distribution in
irrespective of
and l, or the
maximum in the prior expectation for the bending mode is a
v2
complementary distributions in
v3
or l can be calculated by
consequence of summation of P0(
,
l2) over l states. The
v
v3
v3
2
ϩ
˜
FIG. 7. Relative SrOH X ⌺ state vibrational populations obtained by summing P͑v͒ over the l distribution, for the reaction of discharged strontium with
H2O. The ␯ mode is fixed in 1ϭ0 in the simulation. The bending mode quanta are depicted horizontally and the metal-stretching quanta increase in
v
1
perspective. The small dots are the observed populations and the circles are the prior expectation with Etotϭ8700 cmϪ1
.
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
5813
2
ϩ
˜
FIG. 8. Relative SrOH X ⌺ state vibrational populations in metal-
stretching mode and bending mode, obtained by summing P͑v͒ over
,
v2,l
and 3,l, respectively, for the reaction of discharged strontium with H2O.
The filled bars are the observed population and the prior expectation is
indicated by the cross-hatched bars.
v
FIG. 10. For each panel, the observed ͑lower trace͒ and simulated ˜lase˜r
0
0
excitation spectrum of the ͑0,0 ,0͒–͑0,0 ,0͒ band region of the SrOH B–X
transition is shown for the reaction of discharged strontium with H2O at the
indicated H2O pressure. The atomic line near 6112.5 Å is due to a barium
impurity in the beam.
observed bending mode distribution is higher than prior ex-
pectation at low
v2
higher at low l and lower at high l than the prior.
and then falls faster than the prior as
v2
increases. The l distribution irrespective of
v2
is
and
v3
The rotational distribution obtained from the simulation
in Fig. 6 is colder than the result for the simulation presented
in Fig. 4. This is probably due to the different resolution of
band head structure in the two scans. An unsatisfactory at-
tempt was made to simulate the broader scan in Fig. 6 with
its rotational distribution fixed at the result obtained for the
slower scan in Fig. 4. Also, the ␯3 mode distribution obtained
by summing over
and l for the scan in Fig. 4 falls mono-
v2
tonically faster than the same distribution obtained for the
scan in Fig. 6. Comparing these trends suggests that a corre-
lation, thought to be due to limited spectral resolution, of the
metal-stretching mode and rotational distribution parameters
in the nonlinear fitting procedure may be important.
2
ϩ
˜
FIG. 9. Relative SrOH X ⌺ state population in vibrational angular mo-
mentum, l, obtained by summing P͑v͒ over and for the reaction of
v2
v3
However, spectral scans in Fig. 10 recorded at higher
H2O pressures indicate that the spectrum undergoes consid-
discharged strontium with H2O. The filled bars are the observed population
and the prior expectation is indicated by the cross-hatched bars.
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

5814
M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
TABLE III. SrOH ˜B–˜X simulation parameters.
g
Reaction
␹
␹
␹
␹
␭
1
P0͑v,J͒
1
2
3
4
2
aSr(3P0)ϩH2O
bSr(3P0)ϩH2O
dSr(3P0)ϩH2O
eSr(3P0)ϩH2O
fSr(3P0)ϩH2O
→SrOH(X
⌺
⌺
⌺
⌺
⌺
ϩ)ϩH
ϩ)ϩH
ϩ)ϩH
ϩ)ϩH
ϩ)ϩH
Ϫ0.173
Ϫ6.c
Ϫ27.0
Ϫ10.1c
Ϫ10.1c
Ϫ10.1c
Ϫ10.1c
Ϫ4.11
•••
•••
•••
•••
1.13
0.806
•••
•••
•••
Ϫ4.13
Ϫ0.849
Ϫ9.8
Ϫ17.1
Ϫ20.5
1/2
1/2
1/2
1/2
1/2
˜
2
2
2
2
˜
→SrOH(X
→SrOH(X
Ϫ6.c
˜
˜
→SrOH(X
Ϫ6.c
˜
→SrOH(X
Ϫ6.c
aSimulation in Fig. 6.
bSimulation in Fig. 4.
cParameter fixed in simulation.
dSimulation in Fig. 10, P͑H2O͒ϭ0.8 mTorr.
eSimulation in Fig. 10, P͑H2O͒ϭ1.25 mTorr.
fSimulation in Fig. 10, P͑H2O͒ϭ1.8 mTorr.
gThe power of ftrans , the fraction of energy in translation, in P0͑v,J͒ after summing over modes, if any, of the unobserved product for the reaction indicated.
the symmetric angular distribution since either H or H
͑1͒
͑2͒
erable rotational relaxation. These spectra were simulated
with the prior form of P͑v,J͒, and the F1 and F2 level struc-
ture was calculated exactly as in the simulation in Fig. 4. The
vibrational distribution was fixed at the result obtained from
the simulation in Fig. 4. The results indicate that the rota-
tional surprisal parameter, ␭1, depends linearly on the recip-
rocal of the water pressure, at least over the small range of
pressures studied. The optimized parameters for this and all
SrOH simulations are reported in Table III, and the average
energy disposal in metal-stretching and bending vibrations as
well as in rotation are reported in Table IV.
could participate in hydrogen migration. On this basis,
HBaOH need not survive a rotational period to explain the
observed angular distribution. There is also considerable
theoretical evidence for existence of the metal-insertion path-
way in reactions of nickel and aluminum with water.18,19
Since water has negative vertical and adiabatic electron
affinities, reaction with the electronically excited Sr is not
expected to be initiated by long range electron transfer.
However, there have been experimental observations38,39 and
theoretical studies40 of H2OϪ. One theoretical study19 finds
*
that the adiabatic electron affinity of the ␴ ͑O–H͒ orbital in
DISCUSSION
water is Ϫ3.6 eV, which implies electron transfer from
Sr(3P0) to water at harpooning radii of less than 2 Å. Indeed,
long range electron transfer is not indicated by the observed
Davis et al.14 have observed a symmetric forward–
backward angular distribution for the reaction of Ba(1D2)
with H2O, which yields Ͼ90% BaOH product. Their obser-
vations were explained in terms of formation of an HBaOH
intermediate which results from hydrogen migration to the
metal following its multicentered interaction with oxygen
and the electronically excited metal:
˜
monotonically decreasing metal-stretching distribution for X
SrOH. However, the simulation may have yielded a some-
what less highly excited metal-stretching distribution than
indicated by experiment since the shorter wavelength bump,
which is indicative of the
distribution in higher bending
v3
*
levels, does not extend as far to the violet in the simulation
as observed experimentally.
M
O
.
&
͑8͒
-
The observed bending mode distribution is found to de-
viate significantly from prior expectation. Davis et al.14 have
suggested that electron transfer from the metal may be facili-
tated by asymmetric stretching of water, while de Pujo
et al.16 point out that a Cs or C1 approach of the metal is
favored on electrostatic grounds. Using second order pertur-
bation theory, Bader41 and others42 have considered the ef-
fect possible nuclear motions have on the electron density
H
H
͑2͒
͑1͒
Here, the labels on hydrogen, H and H , are used for
͑1͒
͑2͒
reference purposes, and the atoms are not necessarily re-
stricted to a plane. The arrows denote the motion of O and H
nuclei for the asymmetric vibration (B2) of water to be ref-
erenced below. The authors14 emphasize, however, that the
symmetry of the H2O molecule could also be responsible for
2
ϩ
2
ϩ
˜
˜
TABLE IV. Summary of energy disposal in the X
⌺
and B
⌺
electronic states of SrOH.
f0
f0
͘
v
2
Reaction
Etot͑cmϪ1
͒
f
͗frot
͘
͗f3rot
͘
f
f0
͗ ͘
l
f
͗
͘
͗
͘
͗
͘
͗
͗
͘
v
l
v
v
2
3
3
2
aSr(3P0)ϩH2O
bSr(3P0)ϩH2O
cSr(3P0)ϩH2O
dSr(3P0)ϩH2O
eSr(3P0)ϩH2O
→SrOH(X
⌺
⌺
⌺
⌺
⌺
ϩ)ϩOH
ϩ)ϩOH
ϩ)ϩOH
ϩ)ϩOH
ϩ)ϩOH
8700
8700
8700
8700
8700
0.227
0.181
0.130
0.290
0.152
0.348
0.096
0.057
0.048
0.201
0.404
0.404
0.404
0.404
0.014
0.024
˜
2
2
2
2
˜
→SrOH(X
˜
→SrOH(X
˜
→SrOH(X
˜
→SrOH(X
aSimulation in Fig. 6.
bSimulation in Fig. 4; applies to ͑0,00,0͒ level only.
cSimulation in Fig. 10, P͑H2O͒ϭ0.8 mTorr.
dSimulation in Fig. 10, P͑H2O͒ϭ1.25 mTorr.
eSimulation in Fig. 10, P͑H2O͒ϭ1.8 mTorr.
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
5815
distribution. Ground state water (1A1) has the configuration,
(1a1)2(2a1)2(1b2)2(3a1)2(1b1)2, and the lowest unoccu-
pied molecular orbital is 4a1 . Following Bader’s41 theory,
Liu and Parson,43 in their study of reactions of Group 3
metals with water, have suggested that asymmetric stretching
͑this vibrational mode has B2 symmetry since the H2O mol-
ecule has been considered to lie in the yz plane͒ in water
tends to move electron density from the 1b2 to the 4a1 or-
bital. This redistribution of electron density will weaken the
bonding in one O–H bond and shift electron density from
there, as well as from the oxygen atom, to the other O–H
bond. The observed SrOH bending distribution may be con-
sistent with such nuclear motion and electron redistribution,
Alternatively, the considerable rotational excitation ob-
served for SrOH may be a signature of complex formation
͑high rotational excitation is expected on statistical grounds,
see Fig. 5͒ and it seems plausible that an HSrOH intermedi-
ate could decay to products simply by converting energy into
overall rotation so that the H–Sr bond can cleave by passage
over a centrifugal barrier to the exit channel. In this case, the
sharp decrease in the observed ␯2 mode distribution could be
a manifestation of angular momentum exchange from bend-
ing vibrations into rotational motion of the intermediate.
Higher resolution experiments would allow a more accurate
determination of the rotational distribution.
A direct comparison of the bending mode distribution in
SrOH and SrOD, determined from higher resolution experi-
ments as well as an improved transition intensity calculation,
would be particularly helpful in determining whether mo-
mentum transfer resulting from prompt departure of H͑D͒
affects this distribution. Furthermore, detailed vibrational in-
formation on the HSrOH intermediate would help to clarify
the efficiency of energy redistribution in this molecule.
The absence of metal hydroxide product resulting from
alcohol reactions with the electronically excited metal has
also been explained in terms of the hydrogen migration and
the corresponding HMOR intermediate.14,16 The qualitatitive
interpretation of the observed laser excited alkoxide spectra
observed here suggests little metal-stretching excitation
which again is consistent with the expectation that reaction is
initiated only at close range. This is supported by the nega-
tive electron affinites found for aliphatic alcohols44 in elec-
tron transmission experiments.
since the small recoil momentum of the H atom, assuming
͑1͒
H
departs immediately with no formation of an HSrOH
͑1͒
intermediate, would not be expected to impart much momen-
tum to the oxygen atom which is thought to be required for
enhanced excitation of the bending mode. Simple kinematic
considerations of linear momentum and energy conservation
would suggest that upon O–H bond dissociation, little ki-
netic energy would be partitioned to the oxygen atom in its
recoil from hydrogen. The velocity of oxygen may correlate
with the product bending mode excitation. The ␯ excitation
2
which is observed could be explained in terms of the release
of bending mode potential energy associated with the non-
linear arrangement of the Sr, O, and H atoms in this colli-
͑2͒
sion geometry. On the other hand, if internal energy can be
promptly randomized in the HSrOH intermediate, one might
expect a broader distribution in ␯2, such as that predicted by
the prior statistical model.
The induction of asymmetric stretching motion of water
Qualitatively, in terms of energy available ͑13 000
cmϪ1͒, the breadth of the observed ␯ ͑␻Љ ϭ 144 cmϪ1͒ zero
may also imply O–H stretching excitation in the product,
͑2͒
8
8
2
2
˜ ˜
and the energy disposal analysis in Table IV suggests that,
on average, there is sufficient energy to excite one quantum
sequence in the CaOCH3 B A–X A excitation spectrum,
neglecting anharmonicity, is comparatively narrow. In the
absence of computer simulation, the breadth of the ␯ or ␯
of ␯ in the product.
1
4
8
The observed rotational distribution in low vibrational
mode zero sequence structure may be a rough indication of
the mode’s population distribution neglecting variation of the
Franck–Condon factors within the sequence, and assuming
the spectral transitions do not overlap too severely. For the
2
ϩ
˜
levels of X ⌺ SrOH is significantly excited and this may
suggest that ͑off C2 axis͒ Cs or C1 type collision geometries
are favored. It has been suggested that a C2v type attack by
the metal is not favored on electrostatic grounds since the
oxygen in water carries the bulk of the negative charge.16
Consideration of total angular momentum, J, conservation
during the collision can provide additional insight. Here,
JϭLϩj, where L is the orbital angular momentum of the
2
2
˜ ˜
CaOCH3 B A–X A excitation spectrum if ones takes peak
heights in the ␯ mode zero sequence transitions as being
8
proportional to bending populations then f , for 8р11,
v
͗
͘
v
8
is crudely estimated as 0.08.
f
is probably underesti-
͘
͗
v
8
mated since the Franck–Condon factors for the ␯ zero se-
8
particles (Lϭ␮ b) and the angular momentum of molecular
v
quence are expected to decrease with increasing 8 . The
v
rotation is denoted j. Since the reduced mass of the
HϩSrOH products, ␮Ј, is small compared to that of the re-
actants, we have LЈӶL. Furthermore, neglect of the initial j
for H2O then leads to the approximation LϷjЈ. Taking the
most probable jЈ for SrOH as 100ប and assuming the rela-
unresolved ␯4 zero sequences do not allow for an estimate of
f
, but the intensity in these sequences is observed to peak
͘
͗
v
4
sharply at low
for 8р7. The spacing between successive
v
v4
transitions of the ␯ zero sequence in the CaOCH3
4
2
˜ ˜
based on the work of Wormsbecher et al.23 The poorly re-
2
B A–X A spectrum is estimated to be approximately 1 Å,
tive velocity of the reagents, , to be the most probable
v
strontium velocity, we arrive at an impact parameter of Ϸ10
Å. Reaction at an impact parameter of this magnitude is not
likely to occur since such a long range electron transfer is
energetically unfavorable, and hence the high jЈ observed
should not be attributed to large values of L. However, the
assumption LЈӶL could break down if the final relative ve-
2
˜ ˜
2
solved vibrational structure of the SrOCH3 B A–X A
spectrum in Fig. 3 does not allow for estimation of either
f
or f
.
͘
v
4
͗
͘
͗
v
8
The average fraction of energy in the Ca–O–CH3 bend-
ing mode is thus estimated to be less than that fraction of
available energy disposed in the bending mode of ground
locity Ј is large.
v
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19

5816
M. D. Oberlander and J. M. Parson: Reactions of Ca and Sr with water and alcohols
*
state SrOH product resulting from the Sr –water reaction.
Neglecting any trends which may be due to the identity of
the metal, this difference could be due to energy disposal
into other vibrational modes of the metal alkoxide product or
the assumed Ca–OCH3 bond energy may be overestimated.
Late energy release could also explain a cold internal state
distribution, but this implies backward scattering in the prod-
uct angular distribution, which was not observed in the work
of Davis et al.14 on reaction of Ba(1D) with methanol. It
therefore seems that the internal state distribution of the
alkoxide products have some features in common with the
MOH vibrational distribution produced in the water reaction.
The low degree of metal-stretching mode excitation is ex-
pected, and the narrow distribution of bending mode excita-
tion can be explained kinematically, or in terms of inefficient
energy redistribution in the HMOR͑H͒ intermediate.
6 J. A. Coxon, M. Li, and P. I. Presunka, J. Mol. Spectrosc. 150, 33 ͑1991͒.
7 J. A. Coxon, M. Li, and P. I. Presunka, Mol. Phys. 76, 1463 ͑1992͒.
8 M. Li and J. A. Coxon, J. Chem. Phys. 97, 8961 ͑1992͒.
9 J. A. Coxon, M. Li, and P. I. Presunka, J. Mol. Spectrosc. 164, 118 ͑1994͒.
10 J. Nakagawa, R. F. Wormsbecher, and D. O. Harris, J. Mol. Spectrosc. 97,
37 ͑1983͒.
11 P. I. Presunka and J. A. Coxon, Can. J. Chem. 71, 1689 ͑1993͒; P. I.
Presunka and J. A. Coxon ͑unpublished͒.
12 P. I. Presunka and J. A. Coxon, J. Chem. Phys. 101, 201 ͑1994͒.
13 M. D. Oberlander, R. P. Kampf, and J. M. Parson, Chem. Phys. Lett. 176,
385 ͑1991͒.
14 H. F. Davis, A. G. Suits, Y. T. Lee, C. Alcaraz, and J. M. Mestadagh, J.
Chem. Phys. 98, 9595 ͑1993͒.
15 M. A. Douglas, R. H. Hauge, and J. L. Margrave, High Temp. Sci. 17, 201
͑1984͒; J. W. Kauffman, R. H. Hauge, and J. L. Margrave, High Temp.
Sci. 18, 97 ͑1984͒.
16 P. de Pujo, O. Sublemontier, J. P. Visticot, J. Berlande, J. Cuvellier, C.
Alcaraz, T. Gustavsson, J. M. Mestdagh, and P. Meynadier, J. Chem.
Phys. 99, 2533 ͑1993͒.
M. Esteban, M. Garay, J. M. Garcıa-Tijero, E. Verdasco, and A. Gonzalez
17
´
´
We find no indication that metal hydroxide is a signifi-
cant ground state product in the reaction of either Ca(3P0) or
Sr(3P0) with methanol. The observations of Esteban et al.17
indicate that Ca(1D2) reaction with methanol is highly selec-
˜
Urena, Chem. Phys. Lett. 230, 525 ͑1994͒; J. M. Mestdagh, J. P. Visticot,
´
˜
and P. F. Bernath, ibid. 237, 568 ͑1995͒; A. Gonzalez Urena and M.
Garay, ibid. 237, 570 ͑1995͒.
18 S. A. Mitchell, M. A. Blitz, P. E. M. Siegbahn, and M. Svensson, J. Chem.
Phys. 100, 423 ͑1994͒.
*
tive insofar as this reaction is reported to lead only to CaOH
19 S. Sakai, J. Phys. Chem. 97, 8917 ͑1993͒.
20 K. Liu and J. M. Parson, J. Chem. Phys. 67, 1814 ͑1977͒.
21 M. D. Oberlander, Ph.D. dissertation, The Ohio State University, 1995.
22 M. W. Chase, C. A. Davies, J. R. Downey, D. J. Frurip, R. A. McDonald,
and A. N. Syverud, J. Phys. Chem. Ref. Data 14, Suppl. No. 1 ͑1985͒.
23 R. F. Wormsbecher and R. D. Suenram, J. Mol. Spectrosc. 95, 391 ͑1982͒.
24 B. S. Cheong, Ph.D. dissertation, The Ohio State University, 1993.
25 P. J. Dagdigian, H. W. Cruse, A. Schultz, and R. N. Zare, J. Chem. Phys.
61, 4450 ͑1974͒.
emission, albeit with a very small reaction cross section. It is
clear that additional work is necessary to establish the energy
of the M–OCH3 bond ͑MϭCa,Sr,Ba͒. Based on the vibra-
tional assignments presented in Fig. 1, the present work im-
plies a crude Ca–OCH3 bond energy of at least approxi-
˜
mately 2.85 eV, which is insufficiently large to yield A state
chemiluminescence in the Ca(1D2)–methanol reaction. The
bond energy estimate assumes Ca(3P2) is the reactant and
neglects contribution from ␯ ͑␻Љ ϭ 488 cmϪ1͒ and other
26 T. E. Sharp and H. M. Rosenstock, J. Chem. Phys. 41, 3453 ͑1964͒.
27 E. B. Wilson, J. C. Decius, and P. C. Cross, Molecular Vibrations
͑McGraw-Hill, New York, 1995͒.
4
4
28 L. Pauling and E. B. Wilson, Jr., Introduction to Quantum Mechanics
͑McGraw-Hill, New York, 1935͒, pp. 105–111.
modes whose extent of excitation is difficult to determine
due to the limited spectral resolution observed in Fig. 1. If
there are thermochemically accessible channels leading to
29 G. Herzberg, Molecular Spectra and Molecular Structure, Vol. I, Spectra
of Diatomic Molecules ͑Van Nostrand, New York, 1950͒.
30 G. Herzberg, Molecular Spectra and Molecular Structure, Vol. III, Elec-
tronic Spectra and Electronic Structure of Polyatomic Molecules ͑Kreiger,
Malabar, 1991͒.
*
CaOCH3 , whose electronic transitions have been demon-
strated to lie in the same spectral region as the corresponding
transitions of CaOH, then higher resolution chemilumines-
cence experiments are required to address the possibility of
31 R. S. Mulliken, Rev. Mod. Phys. 3, 89 ͑1931͒.
32 G. Herzberg, The Spectra and Structure of Simple Free Radicals ͑Cornell
University, Ithaca, 1971͒.
formation of electonically excited CaOCH3 in the
1
*
Ca( D2)-methanol reaction. The Sr –methanol system may
33 P. R. Lide, Jr., and C. Matsumura, J. Chem. Phys. 50, 3080 ͑1969͒.
34 R. D. Levine and J. L. Kinsey, Information-Theoretic Approach: Applica-
tion to Molecular Collisions, in Atom-Molecule Collision Theory, edited
by R. B. Bernstein ͑Plenum, New York, 1979͒.
be better suited for such a study based on the substantial
˜ ˜
spectral shift observed for the B–X transition of SrOCH3
35 R. B. Bernstein and R. D. Levine, Adv. At. Mol. Phys. 11, 216 ͑1975͒.
36 P. R. Bevington, Data Reduction and Error Analysis for the Physical
Sciences ͑McGraw-Hill, New York, 1969͒; W. H. Press, S. A. Teukolsky,
W. T. Vetterling and B. P. Flannery, Numerical Recipes in Fortran. The
Art of Scientific Computing, 2nd ed. ͑Cambridge University, New York,
1992͒.
relative to that of SrOH.
ACKNOWLEDGMENTS
Support of the National Science Foundation is gratefully
acknowledged. The authors would also like to thank Dr. Pre-
sunka and Professor Coxon for furnishing unpublished spec-
troscopic results.
37 C. N. Jarman and P. F. Bernath, 47th Ohio State University International
Symposium on Molecular Spectroscopy, June 1992, Abstract TB06.
38 L. J. deKoning and N. M. M. Nibbering, J. Am. Chem. Soc. 106, 7971
͑1984͒.
39 W. J. Griffiths and F. M. Harris, Org. Mass. Spectrom. 22, 559 ͑1987͒.
1 B. S. Cheong and J. M. Parson, J. Chem. Phys. 100, 2637 ͑1994͒.
2 H. W. Cruse, P. J. Dagdigian, and R. N. Zare, Faraday Disc. Chem. Soc.
55, 277 ͑1973͒.
40
¨
H. J. Werner, U. Manz, and P. Rosmus, J. Chem. Phys. 87, 2913 ͑1987͒.
41 R. F. W. Bader, Can. J. Chem. 40, 1164 ͑1962͒.
42 R. G. Pearson, J. Am. Chem. Soc. 91, 1252 ͑1969͒.
43 K. Liu and J.M. Parson, J. Chem. Phys. 68, 1794 ͑1978͒.
44 D. Mathur, C. Badrinathan, and V. R. Marathe, Org. Mass. Spectrom. 24,
625 ͑1989͒.
3 R. C. Hilborn, Z. Qingshi, and D. O. Harris, J. Mol. Spectrosc. 97, 73
͑1983͒; J. M. Mestdagh and J. P. Visticot, Chem. Phys. 155, 79 ͑1991͒.
4 P. F. Bernath and S. Kinsey-Nielsen, Chem. Phys. Lett. 105, 663 ͑1984͒.
5 P. F. Bernath and C. R. Brazier, Astrophys. J. 288, 373 ͑1985͒.
J. Chem. Phys., Vol. 105, No. 14, 8 October 1996
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:
130.89.98.137 On: Wed, 03 Dec 2014 18:43:19