Content uploaded by Philip Cosby
Author content
All content in this area was uploaded by Philip Cosby on May 19, 2021
Content may be subject to copyright.
Photodissociation and photodetachment of molecular
negative ions. Ill. Ions formed in CO2/02/H20 mixtures*
P. C. Cosby, J. H. Ling, J. R. Peterson, and J. T. Moseleyt
Molecular Physics Center, Stanford Research Institute, Menlo Park, California 94025
(Received 29 July 1976)
Total photodestruction cross sections for OT, OZ, 00420, C0i-, CO3, and CO3-H20 have been
measured over the range from 6950 to 4579 A (1.78-2.71 eV). In most cases the photodestruction of these
ions can be attributed to specific photodissociation or photodetachment processes. The ions HCO3 and
HCO3•H20 have also been investigated, and upper limits determined for their total photodestruction. The
experiments were performed using a drift tube mass spectrometer coupled with an argon ion laser and a
tunable dye laser. The cross section values vary from 2 x 10-20 to 1 x 10-17 cm2, and in most cases
photodissociation is the predominant process. In CO3 and OT evidence is found for bound, predissociating
excited states.
I. INTRODUCTION
In recent work1-4 it was discovered that several ions
important in the 60-90 km D region of the ionosphere
undergo substantial photodissociation by visible light.
Calculations5 have shown that photodissociation is an
important daytime loss mechanism for CO; and
CO; • H2O and could account for the rapid increase of
electron density6 in the D region at sunrise. Photo-
detachment and photodissociation processes are also
important in gas discharge7 and a-beam pumped8 lasers,
in magnetohydrodynamic generators9, and in the study
of photon-induced chemical reactions. -In addition,
there is fundamental interest in the interactions of
photons with molecular ions. Studies of such interac-
tions can provide information16-15 about the location,
shape and symmetry of the ground and excited states of
ions, molecular bond energies, the electron affinity of
the neutral parent, and energy partioning in photodis-
sociation reactions.
In this paper we report total photodestruction cross
sections over the range from 4579 to 6950 A for a num-
ber of ions formed in mixtures of CO2, 02, and 1120.
The ions studied were chosen primarily for their pos-
sible importance in the D region, but the cross sections
reported should also be useful in other applications.
We make no attempt here at a detailed analysis of the
results in terms of the structural properties of the ions,
since each such analysis is quite involved and requires
other experimental information in addition to the re-
ported cross sections. In the cases of 0; and CO; how-
ever, such analysis is under way and is mentioned be-
low.
II. APPARATUS AND TECHNIQUE.
The experimental apparatus, which consists of a drift
tube mass spectrometer, an argon ion laser, and a tun-
able dye laser, has been previously described2,4 in
some detail. Briefly, the negative ions are formed in
the gas phase (0.050-0.400 torr) by electron attach-
ment processes and subsequent ion-molecule reactions,
and drift under the influence of a weak applied electric
field through the background gas toward an extraction
aperture. In these experiments, the ratio of the elec-'
tric field to the neutral-gas density, E/N, is chosen
such that the directed drift velocity is only about one
tenth the mean thermal speed of the ions and gas mole-
cules at room temperature. The drift distance is
chosen so that the ions experience many thermalizing
collisions following their production. Just before pass-
ing through the extraction aperture, the ions intersect
the intracavity photons of the laser, which is chopped at
100 Hz. The ions that pass through the extraction aper-
ture into the high vacuum analysis region are mass
selected by a quadrupole mass spectrometer and in-
dividually detected by an electron multiplier.
Photons at seven discrete energies between 2.34-
2.71 eV are obtained using the lines of a commercial
argon ion laser. Continuously tunable photon energies
between 1.78-2.43 eV are obtained using a commercial
"jet-stream" dye laser pumped by the argon laser. In
both cases, the drift tube is contained in the cavity of
the appropriate laser. A major improvement over the
earlier experiments has been achieved by using a more
powerful commercial argon ion laser, having a nominal
output power of 12 W (all lines) to pump the dye laser.
Table I shows the dyes used, together with the wave-
length range and peak intracavity powers obtained.
The wavelength of the dye laser is calibrated with a
reversion spectroscope and a 0.3 m monochromator
relative to the He-Ne and argon laser lines, to an ac-
curacy of ±1 A. In both laser configurations, the pho-
ton beam is linearly polarized perpendicular to the axis
of the drift tube. The circulating power is sampled by
calibrated low transmittance output couplers and moni-
tored by a disk calorimeter.
Although, in principle, it is possible to determine
absolute photodestruction cross sections in our experi-
ment, all the cross sections reported here are put on
an absolute scale by the following normalization pro-
cedure. The photodestruction cross section v(A) for
any negative ion A" relative to the known cross section
of another reference ion Er is given by
crA-(X) = 611-(A)ln(4//)R- PA- v11-
In this expression, I and 10 are-the numbers of ions
detected at a given wavelength during the laser on and
off periods, respectively, Pa-/PA- is the ratio of the
laser output powers measured during the accumula-
tion of counts for each species, and VA-/v,- is the
(1)
The Journal of Chemical Physics, Vol. 65, No. 12, 15 December 1976 Copyright ©1977 Ame rica n Institute of Physics 5267
+20% BA
+0.2% COT
5268 Cosby, Ling, Peterson, and Moseley: Ions formed in CO2 /02/H2 0 mixtures
TABLE I. Laser dyes.
Wavelength
Pump range. Cavity power Lifetime°
Dye Concentrations Linesb/Power(W) (A) (min/max, in W) (h)
Cresyl violet°0 0.001 M (EG) A11/16 7000-6500 19/51 5
+0.0014 M R6G
+0.1% COT
+2% MEOH
Rhodamine B° 0.006 M (EG) A11/16 6700-6000 42/187 241
+0.2% COT
Rhodamine 6G° 0.003 M (EG) A11/16 6430-5650 40/220 Indefinite
+ 0.2% COT
+2% MEOH
Sodium
fluorescein°
0.003 M (EG)
+0.3% COT
A11/16 5700-5275 17/120 Indefinites
Coumarin 540° 0.0013 M (EG) 4880/8 5450-5125 13/54 361
°Concentrations given in moles per liter (M) and percent by volume (%) for an ethylene glycol (EG) so-
lution. R6G= rhodamine 6G dye; COT =1,3,5,7-cyclooctatetraene; ME OH = methanol; BA = benzyl
alcohol.
bLines of the argon ion laser used to pump the dye.
°Lasing period of a 1.5 1 solution of the dye over which the cavity power decreased by approximately
60%.
°The acetate, nitrate, and perchlorate salts of this dye have been used and are essentially equivalent
in performance.
°Available from Eastman Kodak Co.
SCOT must be frequently replenished to maintain performance.
°Equivalent to Coumarin 6. Available from Exciton Chemical Co.
ratio of mean speeds for each species when passing
through the photon beam. This procedure avoids the
necessity of knowing precisely the intracavity photon
flux and the overlap integral between the ions and
photons, both difficult quantities to determine experi-
mentally. In many cases, the ratio VA-/vg- can also be
determined much more accurately than can either of
the velocities separately. A full discussion of these
problems is found in Refs. 2 and 4. The absolute
values of the cross sections reported here are based
on a normalization to the 0- photodetachment cross sec-
tion, as measured by Branscomb, Smith, and Tisone.16
The total photodestruction cross section in Eq. (1)
describes the loss of an ion due to photodetachment or
photodissociation (or both). Other mechanisms such
as multiphoton processes, collisional dissociation, or
reactions following photon excitation to a bound state
are unlikely under our operating conditions, and we
have not yet observed any such processes. Photodis-
sociation may be observed directly by tuning the quadru-
pole mass filter to the mass of a photofragment ion and
observing the increase in this ion when the laser is
on, l's or by using the difference between the mobilities
of the parent and photofragment ions.4 Photodetach-
ment cannot be observed directly, but its presence can
be inferred from differences in the photodestruction
cross section and the apparent dissociation cross sec-
tion obtained from observation of photofragment ions.
A particular feature of these experiments is the abil-
ity to bring the ions, that may be created in high vibra-
tional or even excited electronic states, into thermal
equilibrium with the background gas at essentially room
temperature. One would expect strong effects from
vibrational excitation on the photodissociation of an ion,
and we have observed effects'-4 attributed to such ex-
citation. Due to the relatively high pressure and long
drift distance of this apparatus, the ions can be made
to undergo many thermalizing collisions, typically be-
tween 103-104, before the photon interaction. Often,
changes in the total cross section are observed when
the number of collisions is small, or when the drift
velocity is larger than thermal velocity.
All results reported here were obtained under condi-
tions such that the cross sections remained constant as
the number of collisions was further increased, and all
were obtained for drift velocities much less than ther-
mal velocity. In many cases, extensive tests were
made, such as those reported in Refs. 1-4, in a further
attempt to detect effects of possible vibrational excita-
tion. Therefore, except where specifically noted in the
text, it is reasonable to assume that the cross sections
reported here refer to a room temperature thermal dis-
tribution of vibrational levels in the parent ion. This
situation differs significantly from photodissociation
measurements made using fast ion beams,11'14 where
substantial vibrational excitation of the parent is ob-
serlied, and where this excitation is encouraged to al-
low measurement of the vibrational spacings and popula-
tions in the ground state of the parent ion.
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
Cosby, Ling, Peterson, and Moseley: Ions formed in CO2 /02/H2 0 mixtures 5269
3.0
2.5
C.
2.0
o 1.5
U
w
Eg 1.0
0
cc
U
0.5
0
WAVELENGTH (A)
6500 6000 5500 5000
I I
I I
OZ + by
1.8 2.0 2.2 2.4
PHOTON ENERGY (eV)
FIG. 1. Photodetachment cross section of 0-2 as a function of
photon energy. The isolated error bars are the dye laser data;
the triangles are data obtained at discrete argon ion laser Lines.
2.6
III. PHOTODETACHMENT OF M
The 02— ions used for these measurements were pro-
duced in pure 02 gas at a pressure of 0.1 torr, pri-
marily by the three-body attachment reaction
+ 02 + 02 —•• OZ + 02 . (2)
The measurements were made using a drift distance of
at least 10.2 cm and an E/N of 10 Td (1 townsend =10-17
V cm2). Under these conditions only 0- and 02- ions
were observed in significant concentrations. Small
amounts of 0; and CO;, less than 1 part in 103 of the 0'
and Oi intensities, could also be observed.
Earlier measurements4 on the photodetachment of 02-
have been extended to cover a much wider wavelength
range. The results are shown in Fig. 1 as a function of
photon energy. The photodestruction here is clearly
photodetachment, since the bond energy of 02- is greater
4 eV. The present results differ slightly from the ear-
lier ones, between 6400-5650 A but agree within the
combined uncertainties. The present results show a
smoother cross section with less possibility of the
structure that was suggested earlier.4 The absolute
values are in excellent agreement with the measure-
ments of Burch, Smith, and Branscomb,17 who used
a fast ion beam and color filters to select photon ener-
gies with a bandwidth of approximately 0.2 eV; with
those of Warneck," who also used a fast ion beam, but
with a monochromator to obtain a photon energy reso-
lution of 0.07 eV; and with very recent measurements
of Vanderhoff and Beyer, 19 who used the discrete lines
from argon and krypton lasers and a drift tube mass
spectrometer technique similar to the one used here.
The photon energy resolution in the present experi-
ment is about 0.0003 eV.
The error bars given in Fig. 1 represent the root-
mean-square sum of the statistical uncertainties in the
measurement of ln(I0/I) and the relative power terms in
Eq. (1). Contributions to the uncertainty in the abso-
lute scale consist of a 10% uncertainty in the value of
the 0- photodetachment cross section, and a 4% un-
certainty in the velocity ratio. Consequently, the abso-
lute scale is considered accurate to ±12%.
IV. PHOTODISSOCIATION OF 03
The Oi ions used in this study were produced in pure
02 gas at pressures ranging from 0.2 to 0.4 torr,
of 10 Td, and drift distances of at least 10.2 cm, by
the reaction"'"
0" +202 - +02 . (3)
As has been discussed, 4 it is energetically possible for
0; to photodetach and to photodissociate via the reac-
tion
0; + hv +02 (4)
at the photon energies used here.
The results of the total photodestruction measure-
ments of Oi are given in Fig. 2 as a function of photon
energy. As discussed in Ref. 4, comparison of the
loss of 03 with the appearance of 0- photofragment ions
indicates that (85%±15%) of the observed photodestruc-
tion occurs by the photodissociation process of Eq. (4).
In addition, the measurements of the photodetachment
cross section of 0; by Wong, Vorburger, and Woo23 in-
dicate that photodetachment contributes less than 10%
to the total photodestruction shown in Fig. 2 at photon
energies above 2.1 eV. To avoid the difficulties as-
sociated with normalizing the 03 cross section to 0-,
when the 0- is not only destroyed by photodetachment
but also produced in the photodissociation of 09, the
cross sections were normalized to those of OZ reported
in the preceding section. The error bars in Fig. 2
were calculated as for 13;, but include the additional
uncertainty in the Oi cross section. The uncertainty in
the absolute scale is again ±12%.
This cross section shows a series of broad peaks and
some subsidiary structures. Several researchers24-27
have observed an absorption in 0; trapped in different
solid environments in this wavelength range, with the
absorption peaks spaced similarly to the observed broad
peaks in the photodestruction cross section. Similar
peaks have also been observed in the relative photo-
destruction measurements of Og by Sinnott and Beaty."
We have previously discussed briefly4 the possible in-
terpretation of our earlier results on cc. These new
results will allow a more detailed analysis of this pro-
cess and of the electronic states of (N. The results of
this investigation will be reported separately.
V. PHOTODESTRUCTION OF 04
The Oi ions used in this study were produced in pure
02 gas at pressures ranging from 0.3 to 0.4 torr, an
E/N of 10 Td, and drift distances of at least 10.2 cm.
The Oi ions are formed in the reaction
(Di +02+02=o;+02 , (5)
for which the forward and reverse rate constants have
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
5270 Cosby, Ling, Peterson, and Moseley: Ions formed in CO2/02/H 2 0 mixtures
6500
10.0
7.0
4.0
2.0
1.0
co 0.7
0
- 0.4
O
- 0.2
0
2 0.1
0 cc 0.07
0.04
0.02
WAVELENGTH (A)
5500
1.9 2.0 2.1 2.2 2.3 2.4
PHOTON ENERGY (eV)
5000
2.5 2.8 2.7
FIG. 2. Photodestruction cross section of 0:3" as a function of photon energy. The isolated error bars are the dye laser data; the
squares are data obtained at discrete argon ion laser lines.
been measured21,22 to be 4-5.1x 1041 cm6 sec"' and 1. 6-
2. 7 x 1014 cm3 sec-', respectively. Thus, at the gas
densities used here (1.0-1.3 x1016 cm-3), 0; ions are
formed continuously along the drift path between the
source and the laser beam. It should, therefore, not
be assumed that these ions are in thermal equilibrium
with the gas. However, the observed photodestruction
cross section for 0; was found to be independent of
variations in drift distance from 5.1 to 25.4 cm, and of
variations in pressures from 0.3 to 0.4 torr, indicating
either that the cross section may be insensitive to the
internal energy of the 0; or that any internal excitation
produced in the formation of 0; is rapidly quenched.
Total photodestruction cross sections for O4 are given
in Fig. 3 as a function of photon energy. These cross
sections were put on an absolute scale by normalization
to the OZ cross sections of Fig. 1. The large number
of a; ions always present under the conditions used to
form 0; prevents direct normalization to 0-. Of course,
0; might also photodissociate, yielding cc, since the
heat of formation of 0; is only 0.6 eV.2° However, the
effect of this process on the normalization would be
negligible since the concentration of source-produced
OZ in the photon interaction region is two orders of
magnitude greater than that of 04. The error bars in
Fig. 3 again represent the statistical uncertainties,
plus the relative uncertainty of the 02- cross section
used for normalization. The uncertainty in the absolute
scale is ± 15%, including the uncertainty in the absolute
02- cross sections and in the velocity ratio.
Energetically, 0; may photodissociate and photo-
detach. Efforts to observe the production of photo-
fragment ions were unsuccessful because of the large
number of 0-, 0i, and Oi ions in the photon interaction
region, compared with OZ, and because of the similar
drift velocities30 of the 0i, cc, and 0; ions. The strong
similarity between the 0; photodestruction cross sec-
tion and the OZ photodetachment cross section is noted;
the two are in fact equal within their mutual uncer-
tainties at energies above 2. 0 eV. The photodestruc-
tion mechanism for this ion can be determined by using
substantially higher pressures to increase the relative
0; population or by conducting a beam experiment" 14
and would probably help greatly in understanding the
nature of the 02 -02- bonding in this ion. In any case,
photodestruction must result in dissociation, since 04
is not stable.
VI. PHOTODESTRUCTION OF 02 • H2O
The OZ • H2O ions were formed by the reaction21'22
O3+H2O+M-Oi• H2O+M ,
3.0
2.0
0
2 1.0
0
0
7000
WAVELENGTH (A)
6000
I I
+ hv
ill 101111
Ii1111111110101111
5000
1.8 2.0 2.2 2.4
PHOTON ENERGY (eV)
FIG. 3. Photodestruction cross section of Cci as a function of
photon energy. The isolated error bars are the dye laser data;
the triangles are data obtained at discrete argon ion laser
lines.
(6)
2.6 2.8
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
Cosby, Ling, Peterson, and Moseley: Ions formed in CO2/02/H2 0 mixtures 5271
0.6
z 0.4
0
2
0
cc 0.2
7000
PHOTON ENERGY (eV)
1.8 1.9 2.0 2.1 2.2 2.3
I I I I I I I I I I I l
Oi - H2O + by
11
2.4
III
6500 6000
WAVELENGTH (A)
5500 5000
FIG. 4. Photodestruction cross section of OZ • H2O as a func-
tion of photon wavelength.
in a 98:2 mixture of 02 and H2O at a total pressure of
0.1 torr, an E/N of 10 Td, and drift distances of at
least 10.2 cm. The results are presented in Fig. 4 as
a function of wavelength. The uncertainties were deter-
mined as for 04i and the uncertainty in the absolute
scale is again ± 15%. Also, as in the Oi case, it is not
asserted that the 0-2 • H2O ions were in thermal equilib-
rium with the gas when they passed through the laser
beam; nor was the photodestruction channel deter-
mined. Again, both photodetachment and photodissocia-
tion are energetically possible, and photodetachment
will yield dissociation into 02, H2O, and an electron.
VII. PHOTODISSOCIATION OF CO3
The CO3- ions used in this study were formed in pure
CO2 gas, and in mixtures of 02 and CO2, at a pressure
of 0.050 torr, an E/N of 10 Td, and a drift distance
of at least 10.2 cm. The formation, equilibration, and
photodestruction processes have been extensively dis-
cussed1-3 and will not be repeated here. It is con-
cluded that the results, presented in Fig. 5, are for
the photodissociation of CO;, which is in thermal equi-
librium at room temperature, by the process
CO; +hv + CO2 . (7)
The absolute scale was determined by normalization to
both 0" and 02- and has an uncertainty of ± 15%.
We have noted3 that the structure in this cross sec-
tion reflects the vibrational levels of a bound, pre-
dissociating state of CO;. We have recently made a
detailed analysis3I of the co; spectrum, and summarize
the results here for completeness. The bond energy
D(CO2-0-) is (1. 8± 0.1) eV, and the electron affinity
E. A. (CO3) is (2.9 ± 0. 3) eV. Assuming the ground
state of CO; is 2B2, the excited state responsible for
the observed structure is 2A1. The three bending modes
of this state have frequencies of 990, 1470, and 880
cm-I, and the ground level of this state is 1. 520 eV
above the ground level of the ground state.
VIII. PHOTODISSOCIATION OF CO3-• H2O
The CO3- • H2O ions were formed using two different
gas mixtures. A mixture of CO2 and H2O at 0.05 torr,
with a H2O concentration of less than 1% (by volume),
at an E/N of 10 Td and a drift distance of 5.1 cm or
greater, produced the ions 0", CO;, CO; • H2O, OW,
HCO;, and HCO; • H2O. Higher hydrates of CO; and
HCO; could be observed only at significantly higher H2O
concentrations (>10%). A mixture of 02, CO2, and H2O
was also used in this work. The concentration of CO2
was maintained at approximately 5%, while that of H2O
was less than 1%. At an E/N of 5 or 10 Td, a drift dis-
tance of 20.3 cm, and a total pressure of 0.1 or 0.15
torr, 0; and traces of O. • H2O were formed in addition
to those ions observed in the CO2-H2O mixture.
The CO3- • H2O is formed in the three-body reaction32
CO; +H20 +M- CO; • H20 + M . (8)
Observations of the arrival time spectra of the ions
when the source was pulsed and of the photodestru'ction
behavior of each of the ions revealed no evidence of
.ion-molecule reactions or of photon interactions cou-
pling the ions based on 0- (CO; and its hydrates) and
those based on OW (HCO; and its hydrates). In the
CO2-H20 mixture, the total photodestruction cross sec-
tion for cai• H2O could be measured relative to the CO;
cross section with negligible interference2 from other
photodestruction processes when proper account was
taken of photofragment co; ions. In the 02-0O2-H20
mixtures, the coi. H2O cross section could be mea-
6
0
U
z
0
cc
WAVE LENGTH (A)
7000 6000 5000
6
5
CO3 - H2O + by
2.0 2.2 2.4 2.6
PHOTON ENERGY (eV)
FIG. 5. Total photodestruction cross sections of co; (lower)
and coi• H2O (upper) as a function of photon energy. The open
triangles and open circles are data obtained at discrete argon
ion laser lines for CO; and ccc• H2O, respectively; the isolated
error bars are the dye laser data.
CO34 + by
I
11
2- if
-
0 0'"4°
1.8
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
5272 Cosby, Ling, Peterson, and Moseley: Ions formed in CO2/02/H20 mixtures
sured relative to that of icr2. with no measurable inter-
ference of photofragment ions.
The results of the co; . H2O photodestruction mea-
surements are given in Fig. 5. The error bars include
the statistical uncertainties in ln(/0//), in the relative
power measurement, and in the CO; and 0; cross sec-
tions used for normalization. The uncertainty in the
absolute scale is f 20%, reflecting the absolute uncer-
tainties in the CO; and cc cross sections and in the
ratios of the relative velocities of these ions to that of
CO;•1130.
Since the electron affinity of CO3 is (2. 9± 0. 3) eV and
the electron attachment energy of co; • H2O will exceed
that of co; by approximately the co;• H2O bond energy
(-0.5 eV), the observed photodestruction of CO; • H2O
must be photodissociation.
Dissociation into CO;+H20 and CT • H20+ CO2 should
be energetically permitted over the entire wavelength
range used here. In addition, the threshold for disso-
ciation'into 0- +C + 1120 would be expected at 5500 A
or shorter wavelengths, given the bond energies in-
volved.
We have searched for the production of these photo-
fragment ions at wavelengths between 5145-6400 A. In
this region only co; photofragments were observed,
and in amounts that accounted for 90%± 10% of the total
CO.c H2O photodestruction. No evidence was observed
for the production of 0- or 0-• H2O photofragments by
the hydrate. However, it is known" that the 0- • H2O
ion reacts rapidly (k >1 X 10-11 cm3/sec) in oxygen gas,
and it is likely that a similar reaction will take place
in CO2 at about the same or even faster rate. For the
drift tube conditions used here, more than half (and
possibly all) of any photofragment 0"• H2O would con-
sequently be destroyed by reaction prior to their detec-
tion. It is similarly difficult to assess the possible pro-
duction of photofragment 0- from COi• H2O. For the
drift tube conditions used here, the maximum number
of photofragment 0- ions that could possibly be pro-
duced from the hydrate (i. e. , <20% of the total
ccrs•H2o photodestruction) would be two orders of mag-
nitude smaller than the number of photofragment
ions that are simultaneously produced from CO3. More-
over, the nearly identical mobilities of co; and
CO.2- H2O do not permit resolution of their respective
photofragments using time-of-flight techniques. We
therefore conclude that although photodissociation of
CO;. H2O to form either 0- or 0- • H2O photofragments
cannot be entirely ruled out, the predominant photodis-
sociation channel at visible wavelengths yields CO;
+ H2O.
We have previously' discussed several possible in-
terpretations of the co; • H2O photodestruction at the
argon laser lines. The cross section obtained using
the dye laser and our interpretation of the CO; results"
now allow a better understanding of the CO3•1120 photo-
dissociation. We have concluded that co; absorbs at
photon energies between 1. 52-2.35 eV into a state that
is predissociative above 1.8 eV. Above 2. 35 eV, CO;
continues to absorb, although less strongly, probably
into one or more other predissociating states. Between
1. 8-i. 35 eV, only three modes of CO3, identified as
bending modes, are found to predissociate, but it is
likely that absorption also occurs into both the sym-
metric and antisymmetric stretch modes of the excited
electronic state. If the weak CO3-H20 bond (-0. 5 eV)"
is primarily electrostatic, the presence of the H2O
should only slightly perturb the electronic states of the
isolated CO; ion. Thus, photoabsorption by the hydrate
should take place over essentially the same range of
photon energies as for cq. But when the cluster ab-
sorbs, in addition to radiation back to the ground state
or dissociation into 0- or 0- •1120, it has a lower ener-
gy channel for disposing the energy acquired in the
phetoabsorption—ejection of the H2O,
It is seen in Fig. 5 that the similarity in the energy
dependences of the mean destruction cross section for
the CO; and CO3. H2O ions at photon energies above ap-
proximately 1. 9 eV supports this model, as does the
fact that ccs.• H2O continues to dissociate at energies
below the photodissociation threshold of the cq. It is
therefore expected that the threshold for the hydrate
photodissociation will occur in the region of the pre-
dicted origin of the 12A, state of CO; at 1. 52 eV. The
actual threshold will depend on the relative interactions
of the H2O with the ground and excited CO; states. It
is not now understood how the photoexcited levels in the
region of 1.8 eV, which are only about 0. 3 eV above the
ground level of the 1'A1 state of c1:4, are effective in
supplying the me 0. 5 eV required to dissociate the H2O
cluster, Certainly there is sufficient electronic energy,
and it may be that radiationless transitions to vibra-
tionally excited levels of the ground electronic state
occur, which are then partly internally relaxed by dis-
sociation into CO; + H20.
The fact that the photodissociation cross section for
CO;• H2O is substantially greater than that for CO3 can
be explained if it is easier for the excited complex to
localize 0.5 eV for ejection of the H2O than for CO; to
localize 1.8 eV for the ejection of 0-. Thus the pre-
dissociation channel can compete more effectively with
the radiation (fluorescence) channel in the hydrate than
in the parent.
The fact that the co; .1120 destruction cross section
does not exhibit the same detailed structure as co;
may result from additional vibrational modes that are
excited by absorption but which do not contribute to dis-
sociation in co; itself. It may also be expected that
the co; absorption frequencies in the hydrate are slight-
ly dependent on the orientation of the water molecule
and that the vibrational excitation present in the
CO; • H2O bound at 300 °K thus causes absorption spectrum
of the hydrated CO; to be smoothed out compared with
that of CO;
The presence of the water molecule (and its asso-
ciated electric dipole field) may also increase the co;
absorption cross section, thus accounting for the three
times larger destruction cross section in the hydrate.
Since it is unlikely that changes in the Franck-Condon
factors alone could lead to such an increase, this ef-
fect would require a change in the oscillator strengths
of the transitions. In view of the close similarity of
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
Cosby, Ling, Peterson, and Moseley: Ions formed in CO2/02/H20 mixtures
TABLE II. Total photodestruction cross
section of CO.
Wavelength Cross section
(A) (10-20 cm2)
6900 <3.0
6500 <1.6
6400 <2.0
6200 <2.0
6000 <2.1
5800 <2.8
5500 <2.0
5200 2.5±2.3
5145 3.7 ± 2.0
5273
the envelopes of the two cross sections, indicating that
the absorbing levels of the hydrate are essentially those
of the isolated ccri, this explanation is not very con-
vincing.
While this discussion is clearly not definitive, the
present results do allow a narrowing of the alternatives
proposed earlier2 and suggest additional experiments
that should assist in our understanding of c03. H2O and
its photodestruction characteristics.
IX. PHOTODESTRUCTION OF HCO3AND HCO3• H2O
As mentioned in the preceding section, addition of
H2O to CO2 in the drift tube results in the production of
OW, HCO3, and hydrates of HCO3. Alternatively, the
HCO3 ion can be produced" in the absence of hydrates
by using a 98: 2 mixture of CO2 and CI14. We have in-
vestigated the total photodestruction cross section of
HCO3 by each method of formation, using the tunable
dye laser at wavelengths between 6500-5145 A. We
found that the cross section for this ion is less than
3x 10-20 cm2 over this wavelength range, and possibly
zero.
Results for HCO3 • H2O were similar to those for
HCO3. The total photodestruction cross section was
observed to be less than 7 x10-21 cm2 at 5145 A, and
statistically consistent with zero over the range from
6500-4579 A.
We have previously reported2 small but nonzero
photodestruction cross sections for HCO3 at four argon
ion wavelengths. However, in the course of our further
work using the drift tube, we found that the photon beam
produced a small modulation (<0.1%) of the detected
current of a nonabsorbing ion species when it was in the
presence of relatively high densities of other species
that have large photodestruction cross sections. We
believe that this modulation results from the large
changes in ion density that occur in the photon interac-
tion region when significant fractions of the absorbing
ions in this region are photodetached. This effect can
be identified by monitoring the apparent photodestruction
cross section as a function of total ion density. The
modulation is found to occur for HCO3 at the dye laser
wavelengths because of the high relative densities of
CY and OW ions present in both the CO2-H20 and CO2—
CH4 mixtures. The total photodestruction cross sec-
tions reported here were obtained at a sufficiently low
ion density that the modulation effect was undetectable.
This precaution was not taken when the cross sections
were measured at the argon ion laser wavelengths.
Consequently, those values2 should be considered only
as upper limits to the HCO; total photodestruction cross
section, until the cross section at the argon ion laser
wavelengths can be investigated as a function of total
ion density.
X. PHOTODESTRUCTION OF CO4
The CO; studied here was formed" in a 95: 5 mixture
of 02 and CO2 at a pressure of 0.1 torr, an EM of 5 Td,
and a drift distance of 30.5 cm. This ion has an elec-
tron affinity" of 1.22 eV and a CO2-0; bond energy33 of
0.8 eV. It therefore is energetically possible for it to
both photodissociate and photodetach at these photon en-
ergies. The total photodestruction cross section of CO;
was measured at wavelengths between 6900-5145 A,
using the tunable dye laser, and at 5145 A, using the
argon ion laser.
The results, summarized in Table II, show that the
cross section is less than 3>< 1040 cm2 at wavelengths
between 6900-5500 A. However, small but nonzero
cross sections are measured at 5200 and 5145 A. Be-
cause of the small size of these cross sections and the
low abundance of CO; produced in the drift tube, possi-
ble photofragments of CO; could not be observed. Thus
it is not known whether the photodestruction of this ion
is due to photodetachment or to photodissociation.
XI. SUMMARY AND CONCLUSIONS
Photon interactions with nine molecular negative ions
have been studied over the wavelength range from 6950
to 4579 A. For three of these ions, 0i, CO; and COs•H20,
photodissociation into an ionic photofragment was
observed, and evidence was obtained for the existence
of bound, predissociative states in these species. For
three ions, 04, 0; • H2O, and C04, photodestruction was
observed, which probably results in dissociation, but
it was not determined whether photofragment ions were
produced or neutral products and an electron. One ion,
which can only photodetach at these wavelengths,
yielded absolute values in agreement with earlier work.
Two ions, HCO; and HCO; • H2O, apparently neither
photodetach or photodissociate over this wavelength
range.
The initial motivation for these studies was the im-
portance of these ions in the ionosphere. However, it
is apparent that these measurements provide valuable
information for studies of ionic structure and potential
surfaces. Such a study has been done28 for ccs, and
one is under way for Oi. Application of ion photofrag-
ment energy spectroscopy11'14 to ions such as co; and
0; should further extend our knowledge of their structure
and dissociation mechanisms.
*This research was supported by the U. S. Army Ballistics
Research Laboratories through the U. S. Army Research
Office.
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
5274 Cosby, Ling, Peterson, and Moseley: Ions formed in CO2 /02/H2 0 mixtures
tAddress for 1975-76 academic year: Laboratoire des Colli-
sions Ioniques, Universite de Paris-Sud, 91405 Orsay,
France.
IJ. T. Moseley, R. A . Bennett, and J. R. Peterson, Chem.
Phys. Lett. 26, 288 (1974).
2J. T. Moseley, P. C. Cosby, R. A. Bennett, and J. R. Peter-
son, J. Chem. Phys. 62, 4826 (1975).
3P. C. Cosby and J. T. Moseley, Phys. Rev. Lett. 34, 1603
(1975).
4P. C. Cosby, R. A. Bennett, J. R. Peterson, J. T. Mose-
ley, J. Chem. Phys. 63, 1612 (1975).
5J. R. Peterson, J. Geophys. Res. 81, 1433 (1976).
6R. P. Turco and C. F. Sechrist, Radio Sci. 7, 717 (1972);
R. P. Turco, ibid. 9, 655 (1974).
1W. L. Nighan and W. J. Wiegand, Phys. Rev. A 10, 922
(1974).
8D. L. Huestis et al. , "Visible Absorption by Rare Gas Mo-
lecular Ions and Excimers," Paper RN4, 31st Symposium on
Molecular Structure and Spectroscopy, Ohio State University,
1976,
9F. E. Spencer, J. C. Hendrie, and D. Bienstock, Paper VIII
4, 13th Symposium on Engineering Aspects of Magnetohydro-
dynamics, Stanford University, 1973.
19G. H. Dunn, Phys. Rev. 172, 1 (1968).
11N . P. F. B. van Asselt, J. G. Mass, and J. Los, Chem.
Phys. Lett. 24, 555 (1974); Chem. Phys. 5, 429 (1974)0.1,
253 (1975).
12G. E. Busch and K. R. Wilson, J. Chem. Phys. 56, 3638
(1972).
13R. N. Zare, Ph.D. thesis, Harvard University, 1964; Mol.
Photochem, 4, 1 (1972).
14J. B. Ozenne, D. Pham and J. Durup, Chem. Phys. Lett.
17, 422 (1972); J. B. Ozenne, J. Durup, R. W. Odom, C.
Pernot, A. Tabche-Fouhaille, and M. Tadjeddine, Chem.
Phys. 16, 75 (1976).
15G, E. Busch and K. R. Wilson, J. Chem. Phys, 56, 3626
(1972).
18L. M. Branscomb, S. J. Smith, and G. Tisone, J. Chem.
Phys. 43, 2906 (1965).
17D. S. Burch, S. J. Smith, and L. M. Branscomb, Phys.
Rev. 112, 171 (1958).
I8P. Warneck, Laboratory Measurements of Photodetachment
Cross Sections of Selected Negative Ions GCA Tech. Rept.
69-13-N, GCA Corp., Bedford, Massachusetts, 1969.
I9J. A. Vanderhoff and R. A. Beyer, Bull. Am. Phys. Sec.
21, 85 (1976); J. Chem. Phys. (in press).
20R. M. Snuggs, D. J. Volz, I. R. Gatland, J. M. Schummers,
D. W. Martin, and E. W. McDaniel, Phys. Rev. A 3, 487
(1971).
21J. L. Pack and A. V. Phelps, Bull. Am. Phys. Soc. 16,
214 (1971).
22J. D. Payzant and P. Kebarle, J. Chem, Phys. 56, 3482
(1972).
"S. F. Wong, T. V. Vorburger and S. B. Woo, Phys. Rev.
A 5, 2598 (1972).
24M. E. Jacox and D. E. Milligan, J. Mol. Spectry, 43, 148
(1972); Chem. Phys. Lett. 14, 518 (1972).
23P. A. Giguere and K. Herman, Cana J. Chem. 52, 3941
(1974).
26J. B. Bates and J. C. Pigg, J. Chem, Phys, 62, 4227
(1975).
27L. Andrews, J. Chem. Phys. 63, 4465 (1975).
28G. Sinnott and E. C. Beaty, Seventh International Conference
on the Physics of Electronic and Atomic Collisions, Abstracts
of Papers (North Holland, Amsterdam, 1971), p. 176. Also
E. C. Beaty (private communication).
"D. C. Conway and L. E. Nesbitt, J. Chem. Phys. 48, 509
(1968).
30R. M. Snuggs, D. J. Volz, J. H. Schummers, D. W. Mar-
tin, and E. W. McDaniel, Phys. Rev. A 3, 477 (1971).
31J. T. Moseley, P. C. Cosby, and J. IL Peterson, J. Chem,
Phys. 65, 2512 (1976).
32P. C. Fehsenfeld and E. E. Ferguson, J. Chem. Phys. 61,
3181 (1974).
33J. L. Pack and A. V. Phelps, J. Chem, Phys. 45, 4316
(1966).
J. Chem. Phys., Vol. 65, No. 12, 15 December 1976
