7726
Bald, Kunkel, and Bernasek: Internal energies of CO2
the observed state populations to two distinct temperatures,
nificant energy loss could potentially result from collisions
with the Ar carrier gas as well as the unreacted CO and O2
reactant gases.
one for accommodated and the other for hot CO . In the
2
present work no evidence of this two temperature behavior
was observed. However, given the number of gas phase col-
lisions under these flow conditions, one might expect to see
an equilibrated population distribution. Additionally, the pro-
cess of vibrational energy transfer between adsorbate and
surface is likely more complicated than that for translational
energy transfer. This may result in an averaging effect where
all molecules are partially equilibrated and exhibit a single
temperature. The extent to which the equilibration occurs
would still be sensitive to oxygen coverage. At any rate, the
observed trends in vibrational excitation as a function of cov-
erage are consistent with this model.
Under our experimental conditions, the majority of CO2
collisions are with Ar atoms. The deactivation of CO by Ar
2
occurs via vibrational to rotational and translational (V–
R,T) energy transfer. In the case of CO ͑001͒ deactivation
2
by argon, the process most likely involved is shown in Eq.
͑6͒,
CO ͑001͒ϩAr↔CO ͑nm0͒ϩArϩϳ270 cmϪ1.
͑6͒
2
2
In the case of CO2 ͑010͒, the molecule relaxes to the ground
state by the process shown in Eq. ͑7͒,
CO ͑010͒ϩAr↔CO ͑000͒ϩArϩ667 cmϪ1.
͑7͒
2
2
Under most conditions in this study the asymmetric
stretching mode, v , exhibited higher temperatures than the
other two normal modes. The same result was seen under
Published results indicate that the vibrational deactivation
rate for process 7 ͑Ref. 33͒ is four times the deactivation rate
for process 6 ͑Ref. 34͒ at 800 K. This could explain the
observed difference in vibrational mode temperatures. The
vibrational relaxation is clearly not complete since all vibra-
tional temperatures are above that of the predominantly ar-
gon flow stream temperature ͑ϳ800 K͒. However relaxation
may be more complete in the (v ϩv ) manifold than the v
3
similar conditions for the low resolution chemiluminescence
9
study performed in this laboratory. This preferential chan-
neling of energy into the v mode suggests that the motion
3
along the reaction coordinate involves a significant amount
of asymmetric stretching. That is, as the reaction proceeds,
the CvO bond in the carbon monoxide molecule lengthens
as the newly forming CvO bond shortens. This motion
along with a nearly linear transition state could lead to the
higher temperature observed for the asymmetric stretching
mode.
1
2
3
manifold.
In addition to CO —Ar collisions, CO —CO collisions
2
2
might also play a role in the relaxation of the nascent carbon
dioxide. However this would not cause the higher tempera-
tures that were observed for the v mode, because CO is
In the high resolution emission study by Coulston and
3
Haller,1 they report no preferential partitioning of energy
5
known to deactivate the CO ͑001͒ level much more effi-
2
ciently than the CO ͑010͒ level. The reason for this efficient
between the vibrational modes for CO produced on Pt at
2
2
deactivation of the v mode is the near resonant V–V energy
8
1
14 K. The temperatures they obtained are 1500Ϯ100 K,
600Ϯ50 K, and 1550Ϯ140 K for the v , v , and v modes,
3
transfer process shown in Eq. ͑8͒,
1
2
3
respectively. In contrast, for the oxidation reaction on Pd,
they detect a higher apparent temperature for the symmetric
stretch levels. A major difference between these studies and
the present work is the pressure range in which the reaction
is carried out. In the chemiluminescence study the authors
estimated that 93% of the CO2 molecules undergo no
CO –CO collisions and 63% undergo no more than one
CO ͑001͒ϩCO͑vϭ0͒↔CO ͑000͒ϩCO͑vϭ1͒
2
2
Ϫ1
ϩ208 cm
.
͑8͒
At 800 K, the deactivation probability for the process above
Ϫ3
35
is ϳ2ϫ10 or 1 in 500 collisions. Given the estimated CO
collisions a CO molecule undergoes for our flow conditions,
2
2
2
CO -reactant gas collisions before being detected. Under
there may be some energy loss, but that would disproportion-
2
such low collision conditions only a very small amount of
energy loss would be expected. In fact Coulston and Haller
observed very little mixing of the Fermi resonant levels
which proceeds at a rate that is nearly gas kinetic.
ately affect the v mode. If this were the case then the pref-
3
erential energy partitioning into the v mode would be more
3
pronounced for the nascent CO molecule.
2
Collisional relaxation of CO by O should not be sig-
2
2
In light of these results one might ask if the energy dis-
tribution observed in the present study is a result of different
collisional deactivation rates for each mode. If this were the
case, whatever initial difference ͑if any͒ in energy deposition
into the modes would be obscured by subsequent energy loss
by gas phase collisions with the other substituents of the
flow. One need be concerned only with collisions which re-
nificant under the flow conditions of this study. The process
for deactivation by O is thought to be V–R,T because the
2
36
deactivation probability is very close to that of argon.
Whatever the process, the data in this study indicates an
increase in vibrational temperature for each of the modes as
the amount of O in the flow ͑and correspondingly O –CO
2
2
2
collisions͒ increases from 75 to 500 sccm. Therefore, under
these collision conditions, any additional deactivation of the
nascent CO by O is either negligible or small compared to
move energy from the nascent CO , not merely repartition it
2
within a vibrational mode. Since the v and v modes are
1
2
2
2
strongly coupled due to the very fast fermi resonant vibra-
tional to vibrational, V–V, energy transfer, one can assume
that they relax together. Therefore, in order to track energy
leakage from the normal modes, one can look at the deacti-
the increased excitation.
Because of this collisional relaxation, the level of exci-
tation reported in this study can be considered only a lower
limit of the actual nascent excitation. In fact higher vibra-
tional temperatures have been reported in other laboratories
vation rates of the ͑010͒ level and ͑001͒ level of CO . Sig-
2
J. Chem. Phys., Vol. 104, No. 19, 15 May 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:
129.49.251.30 On: Thu, 04 Dec 2014 19:46:52