Infrared spectra of cyclic-O6+ and trans-O6+ in solid neon and argon

Article abstract of DOI:10.1063/1.478910

Charged transient species in the oxygen system have been trapped in solid argon and neon using electron impact, Townsend discharge, and laser-ablation methods. The previously identified O₃^-, O₄^-, and O₄⁺ species are observed in these experiments. Absorptions at 1435.0 and 1429.5 cm^-1 in solid neon are characterized as cyclic-O₆⁺ and trans-O₆⁺, respectively, on the basis of annealing behavior, isotopic substitution, multiplet structure in mixed ¹⁶O₂+¹⁸O₂ experiments, and density functional calculations. Cyclic-O₆⁺ is observed at 1416.1 cm^-1 in solid argon, a smaller displacement than found for cyclic-O₄⁺ in solid argon.

Full text of DOI:10.1063/1.478910

Infrared spectra of cyclic- O 6 + and trans- O 6 + in solid neon and argon
Mingfei Zhou, Jale Hacaloglu, and Lester Andrews
Citation: The Journal of Chemical Physics 110, 9450 (1999); doi: 10.1063/1.478910
View online: http://dx.doi.org/10.1063/1.478910
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/110/19?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Vibrational spectra of germanium-carbon clusters in solid Ar: Identification of the ν 4 ( σ u ) mode of linear Ge C 5
Ge
J. Chem. Phys. 125, 044504 (2006); 10.1063/1.2227398
Electronic and infrared absorption spectra of linear and cyclic C 6 + in a neon matrix
J. Chem. Phys. 120, 7520 (2004); 10.1063/1.1683119
Infrared spectra of trans-HOCO, HCOOH + , and HCO 2 − trapped in solid neon
J. Chem. Phys. 119, 10814 (2003); 10.1063/1.1621382
Infrared spectra and density functional calculations of platinum hydrides
J. Chem. Phys. 114, 1559 (2001); 10.1063/1.1333020
Reactions of laser-ablated aluminum atoms with nitrogen during condensation at 10 K. Infrared spectra and
density functional calculations for Al x N y molecular species
Low Temp. Phys. 26, 736 (2000); 10.1063/1.1312401
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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

JOURNAL OF CHEMICAL PHYSICS
VOLUME 110, NUMBER 19
15 MAY 1999
Infrared spectra of cyclic-O؉6 and trans-O؉6 in solid neon and argon
Mingfei Zhou, Jale Hacaloglu,a) and Lester Andrewsb)
Department of Chemistry, University of Virginia, Charlottesville, Virginia 22901
͑Received 11 December 1998; accepted 10 February 1999͒
Charged transient species in the oxygen system have been trapped in solid argon and neon using
electron impact, Townsend discharge, and laser-ablation methods. The previously identified OϪ3 ,
OϪ4 , and Oϩ4 species are observed in these experiments. Absorptions at 1435.0 and 1429.5 cmϪ1 in
solid neon are characterized as cyclic-O6ϩ and trans-Oϩ6 , respectively, on the basis of annealing
behavior, isotopic substitution, multiplet structure in mixed 16O2ϩ18O2 experiments, and density
functional calculations. Cyclic-O6ϩ is observed at 1416.1 cmϪ1 in solid argon, a smaller
displacement than found for cyclic-Oϩ4 in solid argon. © 1999 American Institute of Physics.
͓S0021-9606͑99͒30518-3͔
INTRODUCTION
Townsend discharge chemical ionization experimental meth-
ods reported earlier.17,18 Oxygen concentrations ranged from
0.2% to 8.0% in argon. Samples were deposited for 2–6 h on
an 11–12 K substrate, and infrared spectra were recorded on
a Nicolet 5DXB spectrometer at 2 cmϪ1 resolution.
The stabilities, structures, and reactions of molecular
oxygen cation complexes (O2)ϩn have been investigated by
several groups.1–7 Conway and Janik obtained ⌬H° values
for the nϭ1–5 complexes using a high pressure mass
spectrometer.1 Linn and co-workers found the O2–Oϩ2 bond
energy to be 9.7Ϯ0.7 kcal/mol and the ionization energy of
(O2)2 to be 11.66Ϯ0.03 eV with molecular beam
photoionization.3 Hiroaka measured the stabilities of (O2)nϩ
ions for nϭ1–8 in gas-phase equilibria and found a de-
crease in the ⌬H° needed to remove O2 with higher n
values.5
Neon matrix experiments were done using laser-ablation
methods, which have been described previously.19,20 The
1064 nm fundamental of a Nd:YAG laser was focused on the
metal target to a spot about 0.2 mm, the ablated metal atoms
and electrons were codeposited with O2 molecules in excess
neon on a 4 K CsI window ͑APD Cryogenics Heliplex͒ at
1–2 mmol/h for 30 min to 1 h. Relatively low laser power
͑3–5 ml/pulse͒ favored the stabilization of charged species
and was used here. Fourier-transform infrared ͑FTIR͒ spectra
were recorded on a Nicolet 750 spectrometer at 0.5 resolu-
tion and 0.1 cmϪ1 accuracy using a 77 K HgCdTe detector.
Samples were annealed with resistance heat and subjected to
photolysis using a 175-W medium pressure mercury arc.
The simplest of these oxygen clusters (O2)ϩ2 has been
observed by matrix infrared8–10 and election spin resonance
͑ESR͒ spectroscopy11 and examined by theoretical
calculations,12–14 and found to exist in symmetrical cyclic
͑rectangular͒ and trans-Oϩ4 forms in the quartet spin state.14
The trans form exhibits a strong antisymmetric O–O funda-
mental at 1164.4 cmϪ1 and the cyclic form at 1320.3 cmϪ1 in
neon. There are no infrared spectroscopic data for the higher
Oϩ6 and Oϩ8 clusters.
In pulsed-laser ablation experiments with metals and
oxygen, we have observed OϪ4 in solid argon at 953.8 and a
1118.6 cmϪ1 band that appeared to be the argon matrix coun-
terpart of the trans-O4ϩ band.15,16 Matrix discharge experi-
ments have been done to prepare higher clusters, which will
be reported here. Complementary neon matrix experiments
with laser-ablated metal and oxygen have produced the spec-
tra of cyclic and trans-O4ϩ and provided evidence for the
growth of higher clusters on annealing to allow association
and reaction of additional oxygen molecules. The results of
these neon experiments and density functional calculations
on cyclic and trans-Oϩ6 will also be described.
RESULTS AND DISCUSSION
Complementary argon and neon matrix experiments will
be presented to identify Oϩ6 species.
Argon matrix
Experiments performed with the discharge chemical ion-
ization source operating in negative ion mode18 are illus-
trated in Fig. 1 for 1.0%, 4.0%, and 8.0% O2 in argon. The
spectra are dominated by ozone absorptions,21 and HO2
radical22 is observed from water impurity in the discharge.
Argon matrix product absorptions are listed in Table I. The
1550.2 cmϪ1 band sometimes splits into 1551.8, 1548.8
cmϪ1 bands and is near the O2 fundamental at 1552 cmϪ1 in
solid argon;23 these bands appear in samples without dis-
charge so they are attributed to neutral oxygen clusters
(O2)n .
EXPERIMENT
Argon matrix experiments were performed to trap ions
in the oxygen system using the electron impact and
The 804.3 cmϪ1 band has been observed in earlier pro-
ton radiolysis experiments with O2 and assigned to the iso-
lated OϪ3 anion.24 The 953.8 cmϪ1 band has been observed in
numerous laser-ablation experiments with metals and oxygen
and assigned to the isolated OϪ4 anion.16 In the 8.0% O2
a͒
Present address: Middle East Technical University, Ankara, Turkey.
Author to whom correspondence should be addressed; electronic mail:
b͒
lsa@virginia.edu
0021-9606/99/110(19)/9450/7/$15.00
9450
© 1999 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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

J. Chem. Phys., Vol. 110, No. 19, 15 May 1999
Zhou, Hacaloglu, and Andrews
9451
product bands. The same behavior was observed in a 0.5%
O2 experiment, but the bands were much weaker.
The same absorptions were observed with the discharge
operating in positive ion mode,18 and the relative population
of OϪ4 and Oϩ4 in the matrix samples was not changed sig-
nificantly; unfortunately, the polarity of the extraction ring
with respect to the discharge block at these voltages makes
little difference in the distribution of products trapped.
A 2% O2 in argon sample was codeposited at 11–12 K
with 70 eV electrons,17 and the same product bands were
observed with different populations. The O3 yield was very
small, the 953.8 cmϪ1 band was stronger than the 1118.6
cmϪ1 band, the 1310.6 cmϪ1 absorption was stronger than
the 1328.9 cmϪ1 absorption, and the 1416.7 cmϪ1 band was
observed on deposition.
FIG. 1. Infrared spectra in the 1650–750 cmϪ1 region for argon and oxygen
passed through a Townsend discharge and condensed at 11–12 K for 4–6 h.
͑a͒ 1.0% O2 , ͑b͒ 4.0% O2 , and ͑c͒ 8.0% O2.
The same discharge experiments were done with 1% and
2% 18O2 samples, and the shifted product bands are listed in
Table I. Experiments were done with mixed 16O2ϩ18O2
samples ͑3%ϩ3%͒ and strong triplets were observed for O4Ϫ
as described earlier.16 Strong triplets were also found at
1119.0, 1085.0, and 1056.0 cmϪ1 with a weaker triplet at
1331.4, 1328.9; 1290.9, 1288.9; 1261.4, 1259.1 cmϪ1. The
1416.7 and 1337.5 cmϪ1 bands from pure 16O2 and 18O2
samples were strong and these bands doubled on annealing
to 22 K and were followed by weaker intermediate bands at
experiment, satellite bands at 925.1 and 764.9 cmϪ1 are
probably due to the O2 complexes (O2͒͑OϪ4 ) and (O2͒͑O3Ϫ),
respectively. A 1119 cmϪ1 band has also been observed in
many of these laser-ablation studies15 and attributed to Oϩ4
based on comparisons with the neon matrix work.8 It is note-
worthy that lower laser-power experiments allow stronger
OϪ4 and Oϩ4 absorptions to survive the deposition process in
solid argon. The new 1416.7 cmϪ1 band, which increases
markedly at higher O2 concentration, is of primary interest
here.
A similar experiment using 2% O2 gave stronger bands
than the 1% O2 sample in Fig. 1͑a͒, and the OϪ4 band had
twice the intensity of the OϪ3 band. Annealing to 20 K
slightly decreased the OϪ4 band and increased the OϪ3 band,
decreased the 1119 cmϪ1 band and increased the 1186.1,
1331.4, and 1328.9 cmϪ1 absorptions, eliminated the 1310.3
and 1301.3 cmϪ1 bands, and produced new 1550.2 and
1416.7 cmϪ1 absorptions. Further annealing to 25 K in-
creased the latter two absorptions and decreased all other
1384.1 and 1369.7 cmϪ1
.
Analogous discharge experiments were done with 2%
and 8% statistical16,18 O2 in argon. A sextet was observed for
the 953.8 cmϪ1 band, as described previously,16 and a similar
sextet was observed at 1119, 1102, 1088, 1085, 1071, and
1056 cmϪ1 with a weaker sextet for the 1331.4, 1328.9 cmϪ1
bands. The 1416.7–1337.5 cmϪ1 multiplet was complicated,
but growth on annealing helped locate intermediate compo-
nents. The strongest intermediate component was at 1377
cmϪ1, with weaker features at 1402, 1364, and 1352 cmϪ1
.
Finally, the 1609 and 1550 cmϪ1 bands gave triplet absorp-
tions at 1608.6, 1564.3, 1518.2 cmϪ1 and at 1550.8, 1507.2,
1462.7 cmϪ1
.
Recent lower power laser-ablation investigations gave
the same oxygen species absorptions with sharper bands at
1331.4, 1328.9, 1186.1, 1118.6, 953.7, and 804.3 cmϪ1 as
measured with 0.5 cmϪ1 resolution. A typical spectrum is
shown in Fig. 2 for a vanadium metal target. The 1118.6
cmϪ1 band dominates the spectrum after deposition, and an-
nealing to 20 and 25 K decreases the 1118.6 cmϪ1 band and
TABLE I. Infrared absorptions ͑cmϪ1͒ observed in solid argon at 11–12 K
from discharge chemical ionization and electron impact on oxygen in
argon.a
16O2
R ͑16/18͒
Ident.
16O2ϩ18O2 16O2ϩ16O18Oϩ18O2 18O2
1609.5 doublet
1550.2 doublet
1490.4 doublet
1416.7 quartet
1389.1 doublet
1331.4 triplet
1328.9 triplet
triplet
triplet
?
higher
quartet
sextet
sextet
?
1517.3 1.0607 X-(O2)n
increases weak bands at 1331.4, 1328.9, and 1186.1 cmϪ1
,
1462.3 1.0601 (O2)n
(Oϩ8 )?
1407.2 1.0591
1337.5 1.0592
and produces a new 1416.1 cmϪ1 band ͓and the (O2͒VO2
band at 1121.9 cmϪ1͔.25 Photolysis virtually destroyed the
cyc-Oϩ6
1380.2 1.0064 HO2
metal independent bands ͑including 953.7 and 804.3 cmϪ1
,
cyc-Oϩ4
1261.4 1.0555
1259.1 1.0554
1243.4 1.0540
1235.6 1.0538
cyc-Oϩ4
VOϪ3 at 916.2, and VOϪ2 at 896.8, 894.8, 886.5 cmϪ1͒ and
further annealing failed to restore them. Identical experi-
ments with 18O2 gave 1261.4, 1259.1, 1119.0, 1055.8, 901.7,
and 759.2 cmϪ1 bands, and a new 1337.1 cmϪ1 band ap-
peared on annealing. Finally, an experiment with
statistical16,18 O2 gave a sharp sextet at 1118.7, 1102.3,
1087.6, 1084.4, 1071.1, and 1055.8 cmϪ1 and a sextet of
(cyc-Oϩ4 )
1310.6
1302.1
1186.1
1119
?
?
?
(cyc-Oϩ4 )
?
?
Oϩ4 -O2
1119.1 1.0599
t-Oϩ4
triplet
sextet
sextet
sextet
sextet
sextet
1056
1.0596
1039.8 triplet
1033.1 triplet
953.8 triplet
804.3 triplet
982.7 1.0581 O3
976.4 1.0581 O3
O4Ϫ
901.7 1.0578
759.2 1.0594
O3Ϫ
doublets at 1331.4, 1328.9 cmϪ1; 1313.0, 1310.5 cmϪ1
1297.2, 1294.6 cmϪ1; 1291.6, 1289.2 cmϪ1; 1277.9, 1275.4
;
aNicolet 5DXB at 2 cmϪ1 resolution.
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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

9452
J. Chem. Phys., Vol. 110, No. 19, 15 May 1999
Zhou, Hacaloglu, and Andrews
FIG. 3. Infrared spectra in the 1530–1140 cmϪ1 region for 0.2% O2 in neon
codeposited at 4 K with laser-ablated nickel. ͑a͒ Sample deposited for 30
min, ͑b͒ after annealing to 7 K, ͑c͒ after 15 min full-arc photolysis with
mercury lamp, ͑d͒ after annealing to 9 K, and ͑e͒ after annealing to 12 K.
Arrows denote 1164.4 cmϪ1 band peak.
FIG. 2. Infrared spectra in the 1430–1100 cmϪ1 region for 0.5% O2 in
argon codeposited at 11–12 K with laser-ablated vanadium. ͑a͒ Sample
deposited for 1 h, ͑b͒ after annealing to 20 K, ͑c͒ after annealing to 25 K, ͑d͒
after 15 min full-arc photolysis with mercury lamp, and ͑e͒ after annealing
to 30 K.
bands were observed with reduced intensity ͑about one-third
of that without CCl4͒.
Neon matrix
Laser-ablation experiments with Co, Ni and Cu and O2
in neon gave the largest yield of metal independent product
absorptions, which are listed in Table II. Figure 3 shows the
spectrum with 0.2% O2 in Ne using a Ni target, featuring the
dominant 1164.4 cmϪ1 band and sharp 1320.1 cmϪ1 band
with weaker 1429.5, 1435.0, 1488.8, and 1504.8 cmϪ1 bands.
Annealing to 7 K increased all bands, whereas full-are pho-
tolysis decreased all but the 1504.8 cmϪ1 band and destroyed
cmϪ1; 1261.4, 1259.1 cmϪ1 with profile like that reported for
the 1164.4 cmϪ1 band in neon.8
Finally, an experiment was done with vanadium and
0.5% O2 with 0.05% CCl4 added to the argon mixture to
serve as an electron trap.25–27 The 953.7 and 804.3 cmϪ1
bands were not observed and the VOϪ2 and VO3Ϫ absorptions
were eliminated, but the 1331.4, 1186.1, and 1118.6 cmϪ1
TABLE II. Metal independent infrared absorptions ͑cmϪ1͒ from codeposition at 4 K of laser-ablated metal atoms with O2 molecules in excess neon.
16O2
18O2
16O2ϩ18O2
16O2ϩ16,18O2ϩ18O2
R ͑16/18͒
Assignment
cyc-Oϩ4
t-Oϩ4
2948.5
2808.2
1554.4
1504.8
1488.8
1435.0
1429.5
1346.9
1327.7
1320.1
1318.2
1164.4
1039.7
1038.5
973.0
2802.8
2651.3
1464.4
1419.9
1405.3
1355.0
1349.7
1275.6
1260.0
1251.9
1250.2
1098.8
982.6
2949.1, 2882.7, 2802.9
2808.2, 2732.3, 2651.3
1554.4, 1464.4
2916.4, 2879.7, 2842.8
2770.3, 2731.2, 2691.8
1.0520
1.0592
1.0615
1.0598
1.0594
1.0590
1.0591
1.0559
1.0537
1.0545
1.0544
1.0597
1.0581
1.0582
1.0581
1.0591
(O2)x
(Oϩ8 )
(Oϩ8 ) site
cyc-Oϩ6
t-Oϩ6
1435.0, 1404.4, 1371.8, 1354.8
1429.6, 1396.2, 1349.7
1435.1, 1395.7, 1354.9
1415, 1390.4, 1363
(cyc-Oϩ4 ͒͑O2)x
(cyc-Oϩ4 )(O2)
cyc-Oϩ4
cyc-Oϩ4 site
t-Oϩ4
1320.1, 1281.5, 1251.9
1318.2, 1279.6, 1250.2
1164.4, 1128.5, 1098.8
1039.7, 1025.9, 991.7, 982.6
1038.5, 1024.7, 990.6, 981.4
1320.1, 1302.4, 1286.8, 1281.5, 1268.2, 1251.9
1318.2, 1300.5, 1285.0, 1279.7, 1266.3, 1250.2
1164.3, 1147.5, 1132.1, 1128.5, 1114.6, 1098.7
1039.7, 1025.9, 1015.8 1006.2, 991.7, 982.6
1038.5, 1024.7, 1017.0, 1005.0, 990.6, 981.4
O3
981.4
919.6
751.9
O3 site
O4Ϫ
O3Ϫ
796.3
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

J. Chem. Phys., Vol. 110, No. 19, 15 May 1999
Zhou, Hacaloglu, and Andrews
9453
the 973.0 and 796.3 cmϪ1 bands. In contrast to argon matrix
experiments, further annealing restored these bands and the
final annealing to 12 K ͑quickly͒ increased the 1435.0 cmϪ1
band more than the 1429.5 cmϪ1 band ͑Fig. 3͒. A similar
spectrum using 0.6% O2 in Ne gave stronger bands, but the
1429.5 and 1435.0 cmϪ1 bands were twice as intense relative
to the 1320.1 cmϪ1 band. First annealing increased 1435.0
and 1429.5 cmϪ1 bands at the expense of 1320.1 and 1164.4
cmϪ1 absorptions. Final annealing made the 1435.0 cmϪ1
band stronger than the 1429.5 cmϪ1 absorption. The 1488.8
cmϪ1 band was also relatively stronger, but the 1504.8 cmϪ1
absorption was not.
An experiment was done with CCl4 added to serve as an
electron trap ͑0.2% O2 , 0.04% CCl4 in neon͒. The 796.3 and
973.0 cmϪ1 bands of O3Ϫ and OϪ4 were not observed, but the
1164.4 and 1320.1 cmϪ1 bands8–10 of cyc-O4ϩ and t-Oϩ4 and
the subject 1429.5 and 1435.0 cmϪ1 bands were present with
one-third of their intensities in the analogous experiment
without CCl4 . As before, annealing decreased the 1164.4
cmϪ1 absorption and slightly increased the latter bands. The
1504.8, 1488.0 cmϪ1 bands were too weak to be observed.
Identical laser-ablation investigations were done with
18O2 , and the product absorptions shifted as listed in Table
II. Three experiments were done with mixed isotopic
samples, and the spectrum using a Cu target is illustrated in
Fig. 4 for 0.3% 16O2ϩ0.2% 18O2 . The initial sample deposit
shows the two 16O2 reaction product bands at 1435.0, 1429.5
cmϪ1, and the 18O2 counterparts at 1355.0, 1349.7 cmϪ1, and
a broader new intermediate band at 1396.2 cmϪ1. Annealing
to 6 K increased the lower band set at 1429.5, 1396.2, and
1349.7 cmϪ1, but further annealing to 9 and 12K decreased
these bands and increased the upper 1435.0 and 1355.0 cmϪ1
bands and two weaker new intermediate components at
1404.4 and 1371.8 cmϪ1. A subsequent mercury-arc photoly-
sis decreased all of these bands, but a final 12K annealing
almost restored the latter set with little effect on the former
set.
FIG. 4. Infrared spectra in the 1450–1340 cmϪ1 region for 0.3% 16O2
ϩ0.2% 18O2 in neon codeposited at 4 K with laser-ablated copper. ͑a͒
Sample deposited for 45 min, ͑b͒ after annealing to 6 K, ͑c͒ after annealing
to 9 K, ͑d͒ after annealing to 12 K, ͑e͒ after 15 min photolysis, and ͑f͒ after
another annealing to 12 K.
Two investigations were done with scrambled isotopic
samples, and the spectrum using a Ni target is illustrated in
Fig. 5. The deposited sample has substantial absorption in
0.96 value for simple second row molecules.33 This B3LYP
calculation finds Oϩ2 higher by 292.3 kcal/mol with zero-
point energy correction, which compares favorably with the
278.3 kcal/mol ionization energy for O2 .34 Calculations were
then performed for cyclic and trans-Oϩ4 in quartet spin states
following the complete active space self-consistent field
͑CASSCF͒ calculations of Lindh and Barnes.14 Although in-
tense infrared frequencies were calculated at 1661 and 1689
cmϪ1, small imaginary frequencies were found for both
structures.
the center now dominated by a new band at 1390.4 cmϪ1
.
Annealing to 6 K increases all features given in Table II, but
again further annealing to 9 and 12K allows the sharp 1395.7
cmϪ1 band and the upper counterparts at 1435.0 and 1355.0
cmϪ1 to dominate the spectrum.
Calculations
Quartet Oϩ6 was attempted, but the only converged struc-
ture was a triangular prism, 30 kcal/mol higher than the sex-
tet trans-Oϩ6 cation illustrated in Fig. 6. Both cyclic and
trans-Oϩ6 species failed to converge on the quartet surface.
The trans and cyclic-Oϩ6 structures converged, but the cis-
Oϩ6 structure dissociated on the sextet surface. Our calcula-
tion predicts one strong stretching mode at 1655.3 cmϪ1 in-
Density functional calculations were performed for oxy-
gen cation species using the GAUSSIAN 94 program system
with the B3LYP functional and 6-311ϩG basis sets.28–31
*
Ϫ
3
Calibration calculations were first done for O2 in the ⌺g
ground state and Oϩ2 in the
⌸ ground state. The bond
2
g
lengths calculated for O2 and Oϩ2 ͑1.206 and 1.107 Å͒ are
very slightly shorter than the experimental values ͑1.208 and
1.116 Å͒,32 and the calculated frequencies ͑1632.8 and
2071.6 cmϪ1͒ are higher than the observed fundamentals
͑1556.2 and 1872.3 cmϪ1͒.32 The 0.95 and 0.90 scale factors
required for O2 and Oϩ2 , respectively, are near but below the
6
volving the outside O2 subunits for the converged Bg state
of trans-Oϩ6 . The cyclic-O6ϩ species converged to a 6A1Љ state
with a strong doubly degenerate mode at 1615.5 cmϪ1. Both
converged sextet isomers of Oϩ6 gave all real frequencies that
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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

9454
J. Chem. Phys., Vol. 110, No. 19, 15 May 1999
Zhou, Hacaloglu, and Andrews
Ϫ1
*
TABLE III. Calculated ͑B3LYP/6-311ϩG ͒ vibrational frequencies ͑cm
͒
and intensities ͑km/mol͒ for the cyc-Oϩ6 and t-O6ϩ cations.
t-Oϩ6 (6Bg)
cyc-Oϩ6 (6A1Љ)
1774.3͑0͒
1615.5͑1432͒
1615.5͑1432͒
186.1͑0͒
116.6͑2͒
116.6͑2͒
104.2͑0͒
72.7͑0͒
67.2͑0͒
67.2͑0͒
59.7͑0.2͒
59.7͑0.2͒
Ј
Ј
Ј
1777.0͑0͒
1655.3͑5107͒
1602.0͑0͒
179.7͑0͒
128.2͑0.3͒
72.4͑0͒
43.8͑0.2͒
40.0͑0͒
35.4͑0͒
Ag
Bu
Ag
Ag
Bu
Ag
Bu
Bg
Au
Ag
Bu
Au
A1
E
E
A2
Ј
E
E
Ј
Ј
A1
Љ
A1
Ј
E
E
E
E
Љ
Љ
Ј
Ј
28.5͑0͒
25.6͑0.1͒
21.5͑0.2͒
slightly lower than the 0.90 scale factor calculated here for
the Oϩ2 cation. These B3LYP calculations are offered as a
first approximation for Oϩ6 cation isomers.
O؉4 isomers
The 1320.3 and 1164.4 cmϪ1 absorptions observed in
solid neon8,9 and assigned to cyc-O4ϩ and trans-Oϩ4 are ob-
served in the present neon experiments. The oxygen-18 and
mixed isotopic data ͑Table II͒ are in excellent agreement
with the earlier reports. The 1328.9 and 1118.6 cmϪ1 argon
matrix bands are counterparts of the 1320.3 and 1164.4 cmϪ1
neon matrix bands. Note the opposite shifts: the 1328.9 cmϪ1
band is redshifted 8.6 cmϪ1 in neon, but the 1118.6 cmϪ1
band is blueshifted 45.8 cmϪ1. The 16/18 ratios are, how-
ever, almost the same in argon and neon ͑1.0554 and 1.0545
for cyc-Oϩ4 , and 1.0596 and 1.0597 for trans-Oϩ4 ͒, but note
the considerable difference between 16/18 ratios for the two
isomers. The lower 16/18 ratio for cyc-Oϩ4 , significantly be-
low the harmonic ratio ͑1.0607͒, points to much more anhar-
monicity in the cyclic cation vibrational mode. The smaller
argon-neon difference for cyc-Oϩ4 denotes a weaker matrix
interaction for this ‘‘closed’’ system than for the trans-Oϩ4 .
The weaker 1186.1 cmϪ1 argon matrix band exhibits al-
most the same isotopic 16/18 ratio as trans-Oϩ4 and it in-
creases on first annealing at the expense of trans-Oϩ4 . The
1186.1 cmϪ1 band is probably due to O2 perturbed trans-Oϩ4
or a ͑trans-Oϩ4 ͒••͑O2) complex in solid argon.
FIG. 5. Infrared spectra in the 1450–1340 cmϪ1 region for 0.1% 16O2
ϩ0.2% 16O18Oϩ0.1% 18O2 in neon codeposited at 4 K with laser-ablated
nickel. ͑a͒ Sample deposited for 45 min, ͑b͒ after annealing to 6 K, ͑c͒ after
annealing to 9 K, ͑d͒ after annealing to 12 K, and ͑e͒ after 15 min photoly-
sis.
are listed in Table III. The B3LYP calculation found the 6Bg
6
state to be 3.3 kcal/mol lower than the AЉ1 state, but this
small difference is not significant at this level of theory.
The ‘‘inverse symmetry breaking’’ phenomenon de-
scribed for dissociation of the small radical cation
systems35,36 Hϩ2 and (NO)ϩ2 may be a problem here, but its
effect on vibrational frequencies at the stable minimum for
the larger Oϩ6 system is not known, and this problem needs
further theoretical investigation. We note that the B3LYP
frequencies calculated here for Oϩ6 isomers require scale fac-
tors of 0.86–0.89 to fit the bands assigned below, which are
The 1331.4, 1328.9 cmϪ1 argon matrix doublet increases
on first annealing while the 1310.6 and 1302.1 cmϪ1 bands
disappear. These bands all have the lower 16/18 ratio so it is
reasonable to attribute the later bands to an unstable argon
matrix site of cyc-Oϩ4 . Furthermore, the increase of 1331.4,
1328.9 cmϪ1 bands at the expense of the 1118.6 cmϪ1 band
suggests that the cyclic-Oϩ4 isomer is more stable. It appears
that the cyclic-Oϩ4 isomer may be more stable than the trans-
Oϩ4 isomer in spite of the prediction by CASSCF14 that the
trans isomer is 0.5 kcal/mol more stable.
O؉6 isomers
The 1435.0 and 1429.5 cmϪ1 neon matrix bands have
been observed in a previous neon matrix study where they
ϩ
*
FIG. 6. Structures calculated at the B3LYP/6-311ϩG level for cyclic-O6
and trans-Oϩ6
.
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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

J. Chem. Phys., Vol. 110, No. 19, 15 May 1999
Zhou, Hacaloglu, and Andrews
9455
were associated with a 1329.7 cmϪ1 band and tentatively
attributed to an O2¯O4ϩ complex.10 We agree with this pro-
posal for the latter band, but here we assign the two former
bands to the cyclic and trans-Oϩ6 cation isomers. In the
present experiments, weak bands at 1327.7 and 1346.9 cmϪ1
increase on annealing, and these bands are likely due to
cyc-Oϩ4 complexes with O2 and (O2)x , respectively.
suggest that the larger matrix cavity allows for cyclization to
give the presumably more stable species. Recall that cyc-O6ϩ
was favored over trans-Oϩ6 on annealing to 12 K in solid
neon.
Other absorptions
The weak 1504.8 and 1488.8 cmϪ1 bands increase on
annealing in the neon experiments and show 1.0598 and
1.0594 isotopic ratios for O–O stretching modes. The 1488.8
cmϪ1 band is decreased on photolysis along with the Oϩ4 and
Oϩ6 isomer but the 1504.8 cmϪ1 band increases. These bands
and the 1490.4 cmϪ1 argon matrix counterpart are tentatively
attributed to the Oϩ8 species.
The present neon matrix experiments show that the
1435.0 and 1429.5 cmϪ1 bands continue to increase on an-
nealing after cyc-Oϩ4 and trans-Oϩ4 absorptions decrease and
that the 1435.0 cmϪ1 band increases more than the 1429.5
band at the highest temperature reached ͑12 K͒. This sug-
gests a higher cluster than Oϩ4 for the 1435.0 and 1429.5
cmϪ1 bands. The critical observations are the mixed
16O2ϩ18O2 isotopic spectra of Fig. 4 ͑and two equivalent
investigations with other metals͒. The t-Oϩ6 spectrum reveals
three major bands at 1429.5, 1396.2, and 1349.7 cmϪ1 with
the middle component broader. This suggests a mode involv-
ing two equivalent O2 subunits, but the broader central com-
ponent reveals additional oxygen interaction for the lower-
symmetry mixed isotopic molecule. The calculations for 6Bg
trans-Oϩ6 show that the strong Bu mode involves essentially
the terminal O2 subunits. The cyc-Oϩ6 spectrum is quite dif-
ferent: the pure isotopic bands at 1435.0 and 1355.0 cmϪ1
are much stronger than the two intermediate mixed isotopic
bands at 1404.0 and 1371.8 cmϪ1. This is characteristic of a
doubly degenerate mode and is the pattern observed for
trigonal M͑CO)3 species.25–27,37 Hence, a structure contain-
ing three equivalent O2 subunits is required, and the con-
verged cyc-Oϩ6 structure with three equivalent O2 subunits
fits nicely.
Argon vs neon
What can we learn about the trapped species by compar-
ing spectra in solid argon and neon? The relative band posi-
tions provide some measure of the guest–host interaction,
and we expect small positive ions to interact more strongly
with the matrix host than other guest species. As a reference
point, CO absorbs at 2143.5, 2140.7, and 2138.1 cmϪ1 in the
gas phase, solid neon, and solid argon.32,27,39 Thus, a small
covalent molecule shows little matrix shift. The 1554.2 and
1548.8 cmϪ1 bands are probably due to (O2)2 in solid neon
and argon, which has been observed at 1557.9, 1553.3 cmϪ1
in the gas phase.40 Unfortunately, O2Ϫ and O2ϩ cannot be
detected here, but their participation is confirmed by reac-
tions with O2 to form O4Ϫ and Oϩ4 , reactions ͑1͒ and ͑2͒,
during sample condensation. Mercury-arc photolysis virtu-
ally destroys OϪ4 ,
With the trans-Oϩ6 and cyclic-Oϩ6 models, the statistical
16O2ϩ16O18Oϩ18O2 isotopic spectra in Fig. 5 can be ex-
plained. For t-Oϩ6 , we expect a spectrum similar to Oϩ4 ,
with strong middle components and the 1390.4 cmϪ1 feature
is dominant followed by broader bands at 1415 and 1363
cmϪ1, which denote some involvement of the central O2 sub-
unit in mixed isotopic molecules. The cyc-Oϩ6 spectrum will
be heavily dominated by the 16O18O component at 1395.7
cmϪ1, with minor contributions from 1435.0 and 1355.0
cmϪ1 peaks and many weaker intermediate components that
are not observed here. Due to lack of resolution of all mixed
isotopic components, we cannot claim confirmation of the
trans-Oϩ6 and cyclic-Oϩ6 structures, but we believe the spec-
tra in Fig. 5 are strongly supportive of these Oϩ6 structures,
which are the only ones that gave B3LYP convergence. It is
clear from the spectra that one Oϩ6 structure has three-fold
symmetry and the other does not. Finally, note that the 16/18
isotopic ratios for the bands assigned here to cyc-Oϩ6 and
t-Oϩ6 , 1.0590 and 1.0591, are typical of O–O vibrations
with normal cubic contributions to anharmonicity. It should
be recalled that cyclic O6 is calculated to be a stable singlet
molecule, like S6 , but substantially higher in energy than
three O2 molecules.38
O2ϩOϪ→OϪ cyclic͒,
͑1͒
͑2͒
͑3͒
͑
2
4
O2ϩOϩ2 →Oϩ4 ͑both cyclic and trans),
O2ϩOϩ4 →Oϩ6 ͑both cyclic and trans),
and Oϩ4 through neutralization and dissociation mechanisms.
However, in solid neon further annealing restores Oϩ4 and
Oϩ6 absorptions but not in solid argon. This suggests that O2ϩ
is strongly solvated in solid argon since Oϩ2 ͑IEϭ12.07 eV͒
and argon ͑IEϭ15.76 eV͒ and as a consequence, Oϩ2 is
34
‘‘insulated’’ from the reaction mixture. However, cyc-O4ϩ
does increase on annealing in solid argon before photolysis,
so O2 may be able to penetrate the solvent shell to react with
Oϩ2 to give the more compact cyc-Oϩ4 species. Photolysis in
solid argon is sufficiently energetic to promote charge trans-
fer from argon to Oϩ2 , and the lack of regrowth of Oϩ4 on
annealing may be due to the absence of Oϩ2 . Clearly, neon
͑IEϭ21.56 eV͒ presents no problem as reactions ͑2͒ and ͑3͒
proceed readily on annealing in solid neon before and after
photolysis.
The ionization energy3 of (O2)2 is 11.66 eV and O4ϩ
interacts more strongly with argon than with neon ͑45.8
cmϪ1 difference in band position͒, but apparently not enough
to destroy the chemical integrity and reactivity of Oϩ4 in solid
argon.
The role of CCl4 doping in both argon and neon
experiments25–27 is to capture ablated electrons and thus
minimize the formation of reagent anion species, as in reac-
In solid argon, the 1416 cmϪ1 band grows on annealing
and is strongly concentration dependent. The 16O2ϩ18O2 ex-
periment reveals a spectrum again characteristic of a doubly
degenerate mode, and the 1416 cmϪ1 band is assigned to
cyc-Oϩ6 . We note the absence of trans-Oϩ6 in solid argon and
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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58

9456
J. Chem. Phys., Vol. 110, No. 19, 15 May 1999
Zhou, Hacaloglu, and Andrews
tion ͑1͒, and increase the relative probability for the forma-
tion and survival of positive ions, as in reactions ͑2͒ and ͑3͒.
6 H. Helm and C. W. Walter, J. Chem. Phys. 98, 5444 ͑1993͒.
7 K. Hiraoka, I. Kudaka, and S. Yamabe, J. Mass Spectrom. Soc. Jpn. 43, 65
͑1995͒.
8 W. E. Thompson and M. E. Jacox, J. Chem. Phys. 91, 3826 ͑1989͒.
9 M. E. Jacox and W. E. Thompson, J. Chem. Phys. 100, 750 ͑1994͒.
10 C. L. Lugez, W. E. Thompson, and M. E. Jacox, J. Chem. Phys. 105, 2153
͑1996͒.
CONCLUSIONS
Charged transient species in the oxygen system have
been trapped in solid argon and neon using electron impact,
Townsend discharge, and laser-ablation methods.17–20 The
previously identified OϪ3 , OϪ4 , and Oϩ4 species8–10,16,24 are
observed in these experiments. Absorptions at 1435.0 and
1429.5 cmϪ1 in solid neon are characterized as cyclic-O6ϩ
and trans-Oϩ6 on the basis of annealing behavior and isotopic
substitution, and in particular, the multiplet structure in
mixed 16O2ϩ18O2 experiments. These mixed isotopic
samples give a ‘‘triplet’’ showing the participation of two
equivalent O–O subunits ͑i.e., the terminal O–O submol-
ecules͒, which is appropriate for the trans-Oϩ6 species, and a
‘‘quartet’’ with weaker central components showing the dou-
bly degenerate vibration of three equivalent O–O subunits,
which indicates a cyclic-Oϩ6 species. The O6ϩ bands were
favored over Oϩ4 bands with increasing O2 concentration.
Doping these samples with CCl4 to serve as an electron trap
eliminates the anion absorptions but not the cation bands, as
observed in previous experiments,25–27 which supports the
charge identification. Stepwise annealing of neon samples
provides growth of O6ϩ and at the expense of Oϩ4 , both be-
fore and after mercury-arc photolysis; however, in argon the
growth of cyclic-Oϩ4 and cyclic-Oϩ6 on annealing occurs only
before photolysis, which apparently photoneutralizes the im-
portant Oϩ2 reagent, because no growth of charged species
absorptions occurs after photolysis in argon.
11 L. B. Knight, Jr. S. B. Corbanchi, and J. Petty, J. Chem. Phys. 91, 4423
͑1989͒.
12 D. C. Conway, J. Chem. Phys. 50, 3864 ͑1969͒.
13 J. B. Peel, J. Chem. Phys. 94, 5774 ͑1991͒.
14 R. Lindh and L. A. Barnes, J. Chem. Phys. 100, 224 ͑1994͒.
15 G. V. Chertihin, W. Saffel, J. T. Yustein, L. Andrews, M. Neurock, A.
Ricca, and C. W. Bauschlicher, Jr., J. Phys. Chem. 100, 5261 ͑1996͒.
16 G. V. Chertihin and L. Andrews, J. Chem. Phys. 108, 6404 ͑1998͒.
17
¨
S. Suzer and L. Andrews, J. Chem. Phys. 88, 916 ͑1988͒.
18
¨
J. Hacaloglu, S. Suzer, and L. Andrews, J. Phys. Chem. 94, 1759 ͑1990͒.
19 T. R. Burkholder and L. Andrews, J. Chem. Phys. 95, 8697 ͑1991͒.
20 P. Hassanzadeh and L. Andrews, J. Phys. Chem. 96, 9177 ͑1992͒.
21 L. Andrews and R. C. Spiker, Jr., J. Phys. Chem. 76, 3208 ͑1972͒.
22 D. W. Smith and L. Andrews, J. Chem. Phys. 60, 81 ͑1974͒.
23 L. Andrews and R. R. Smardzewski, J. Chem. Phys. 58, 2258 ͑1973͒.
24 L. Andrews, B. S. Ault, J. M. Grzybowski, and R. O. Allen, J. Chem.
Phys. 62, 2461 ͑1975͒.
25 M. F. Zhou and L. Andrews, J. Phys. Chem. A 102, 8251 ͑1998͒.
26 M. F. Zhou and L. Andrews, J. Am. Chem. Soc. 120, 11499 ͑1998͒.
27 M. F. Zhou, G. V. Chertihin, and L. Andrews, J. Chem. Phys. 109, 10893
͑1998͒.
28
GAUSSIAN 94, Revision B. I. M. J. Frisch, G. W. Trucks, H. B. Schlegel, P.
M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A.
Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G.
Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A.
Nanayakkara, M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W.
Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox,
J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon, C.
Gonzalez, and J. A. Pople ͑Gaussian, Inc., Pittsburgh, PA, 1995͒.
29 C. Lee, E. Yang, and R. G. Parr, Phys. Rev. B 37, 785 ͑1988͒.
30 A. D. Becke, J. Chem. Phys. 98, 5648 ͑1993͒.
31 A. D. McLean and G. S. Chandler, J. Chem. Phys. 72, 5639 ͑1980͒; R.
Krishnan, J. S. Binkley, R. Seeger, and J. A. Pople, ibid., 72, 650 ͑1980͒.
32 H. Huber and G. Herzberg, Spectra of Diatomic Molecules ͑Van
Nostrand-Reinhold, New York, 1979͒.
ACKNOWLEDGMENTS
The authors appreciate National Science Foundation
support for this work ͑Grant Nos. INT87-12205 and CHE97-
33 A. P. Scott and L. Radom, J. Phys. Chem. 100, 16502 ͑1996͒.
34 H. M. Rosenstock, K. Draxl, B. W. Steiner, and J. T. Herron, J. Phys.
Chem. Ref. Data Suppl. 6, 1 ͑1977͒.
¨
00116͒ and the assistance of S. Suzer with the electron-
impact experiment.
35 T. Bally and G. N. Sastry, J. Phys. Chem. A 101, 7923 ͑1997͒.
36 Y. Xie, H. F. Schaefer, III, X.-Y. Fu, and R.-Z. Liu ͑to be published͒.
37 J. H. Darling and J. S. Ogden, J. Chem. Soc. Dalton Trans. , 2496 ͑1972͒.
38 C. P. Balhous III, and Henry F. Schaefer III, J. Phys. Chem. 92, 959
͑1988͒.
1 D. C. Conway and G. S. Janik, J. Chem. Phys. 53, 1859 ͑1970͒.
2 S. L. Anderson, T. Hiraoka, P. W. Tiedermann, B. H. Mahan, and Y. T.
Lee, J. Chem. Phys. 73, 4779 ͑1980͒.
3 S. H. Linn, Y. Ono, and C. Y. Ng, J. Chem. Phys. 74, 3348 ͑1981͒.
4 E. A. Gislason and E. E. Ferguson, J. Chem. Phys. 87, 6474 ͑1987͒.
5 K. Hiraoka, J. Chem. Phys. 89, 3190 ͑1988͒.
39 M. F. Zhou and L. Andrews, J. Chem. Phys. ͑in press͒.
40 C. A. Long and G. E. Ewing, J. Chem. Phys. 58, 4824 ͑1973͒.
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:
158.42.28.33 On: Thu, 18 Dec 2014 11:30:58