11048
J. Chem. Phys., Vol. 111, No. 24, 22 December 1999
M. Zhou and L. Andrews
The large yield of ThO in these experiments is probably
due to decomposition of NThO at the weaker nitride bond
(ThNϭ709.8 cmϪ1), reaction ͑5͒; note that Th18O is pro-
duced from
modes of a ThO2 species, which invites consideration of a
ThO2 complex, but what molecule forms the complex? There
is no evidence for a new N–O vibration, as found above for
NUO͑NO͒, so the partner is most likely 2-N2 where the
N–N vibration is too weak for observation. These bands are
assigned to the ͑N2͒ThO2 complex formed by the thorium
reaction with two NO molecules. Note that the 15N16O reac-
tion produces a ͑15N2͒ThO2 species red-shifted 0.8 cmϪ1
from ͑14N2͒ThO2 showing that nitrogen is involved in a mi-
nor role. This complex has also been observed in solid argon
with frequencies 18.5 and 19.9 cmϪ1 lower,16 similar to the
argon–neon shifts found here for ThO2. The yield of the
proposed ͑N2͒ThO2 complex product is much higher in the
neon experiments as secondary reactions are favored owing
to more reagent diffusion on condensation of neon at 4 K as
compared to argon at 6–7 K.
In contrast to the uranium system, the strong site split
bands at 784.2 and 709.8 cmϪ1 increased markedly on an-
nealing and reached almost twofold after 12 K annealing
͑Fig. 3͒. These bands are 13.9, 13.5 cmϪ1 blue shifted from
argon matrix bands recently assigned to the NThO
molecule.16 Photolysis (Ͼ780, 630, 470 nm and 240–580
nm͒ had no affect on the deposited sample spectrum. The
sharp 784.2 cmϪ1 band showed a small 15N16O shift ͑1.7
cmϪ1͒ and a large 15N18O shift ͑39.2 cmϪ1͒, whereas the
709.8 cmϪ1 feature reversed with large ͑21.0 cmϪ1͒ and
small ͑2.2 cmϪ1͒ shifts, respectively, which are indicative of
Th–O and Th–N stretching modes. These bands are assigned
to the NThO nitride–oxide molecule. Note that the nitride
stretching mode is lower for NThO but higher for NUO
where uranium has enough valence electrons to satisfy both
oxide and nitride ͑five total required͒ but thorium does not.
This further underscores the difference between Th and U
and the role of 5f orbitals in bonding.6,34,35 Finally, no
charged thorium species were trapped in these experiments.
The B3LYP pseudopotential calculations determine the
neon matrix modes within 2 cmϪ1, but of more importance,
predict the isotopic frequencies accurately ͑four frequencies
for 15N16O and 15N18O species to within average of Ϯ0.2
cmϪ1͒. Thus, the straightforward GAUSSIAN 98 calculation ef-
fectively models the vibrational mechanics of the bent NThO
molecule. The same B3LYP calculation predicted the ThO2
stretching modes 3–4 cmϪ1 too high. The BP86 functional
calculated frequencies 1%–3% lower. The bent NThO
nitride–oxide molecule, prepared by insertion reaction ͑4͒, is
related to the bent NThN and OThO molecules,6 and their
frequencies are
ThϩNO→NϩThO,
͑5͒
15N18O. This is not the case for NwUvO, which has a
much stronger nitride bond (UNϭ1004.9 cmϪ1). Annealing
likely produces a NThO͑NO͒ complex intermediate, reaction
͑6͒, but again
NThOϩNO→͓NThO͑NO͔͒→͑N2͒ThO2
͑6͒
owing to the weaker Th–N bond, the most stable final ar-
rangement is proposed to be ͑N2͒ThO2; note that
͑15N2͒Th18O2 is produced from 15N18O. Similar complexes
have been observed for Ti, Zr, and Hf, and DFT calculations
have shown that the ͑N2͒TiO2 complex is by far the most
stable isomer for the TiN2O2 stoichiometry.16,36
V. CONCLUSIONS
The title cation and molecules have been prepared by
reactions of laser-ablated metal with NO during condensa-
tion in excess neon at 4 K to compare with argon matrix
results and pseudopotential calculations. Infrared fundamen-
tals for the NUO and NThO molecules blue shift 1.6%–2.9%
on going from argon to neon matrices and are calculated
from 5.8% to 0.0% too high using the B3LYP hybrid density
functional, GAUSSIAN 98, and pseudopotentials on the ac-
tinide metal. The BP86 functional produces 1%–3% lower
frequencies as expected. The NUOϩ cation has 113 and 136
cmϪ1 higher frequencies than NUO. Two normal modes ͑iso-
topic frequencies͒ are accurately modeled by the calculations
͑Ϯ1.1 cmϪ1 for NUOϩ, Ϯ0.5 cmϪ1 for NUO, and Ϯ0.2
cmϪ1 for NThO͒. The isolated NUOϩ cation observed here
provides a vibrational model for its important isoelectronic
UO22ϩ analog, which has only been characterized in con-
densed phases where partial neutralization of the dictation
readily occurs. Differences between linear NUO and bent
NThO derive from the difference in electron count between
thorium and uranium and the role of 5f electrons in actinide
bonding.
Owing to different relative strengths of UN and ThN
bonds, the secondary NO reaction products of NUO and
NThO are different. The uranium species forms a stable
NUO͑NO͒ complex whereas the thorium nitride bond gives
way to the more stable ͑N2͒ThO2 complex likely containing
the 2-N2 moiety.
ThϩNO→NThO
͑4͒
ACKNOWLEDGMENT
We gratefully acknowledge NSF support for this re-
search under Grant No. CHE 97-00116.
similar ͑NThO is 760.3, 697.3 cmϪ1 in solid argon, whereas
of ThO2 is 735.0 cmϪ1 and of ThN2 is 756.6
3
3
cmϪ1͒.16,21,33 The NThO molecule can be considered as
OThO with one nonbonding electron ͑and one proton͒ re-
moved; NThO is calculated to have a bent structure with a
larger angle ͑127.5°͒ than ThO2 ͑118.8°͒ ͑Table II͒. The bent
structure of ThO2 has been considered in several studies6,34,35
and the bending tendency of the 6d thorium orbitals empha-
sized.
1 S. D. Gabelnick, G. T. Reedy, and M. G. Chasanov, J. Chem. Phys. 58,
4468 ͑1973͒.
2 D. W. Green and G. T. Reedy, J. Chem. Phys. 65, 2921 ͑1976͒.
3 R. D. Hunt and L. Andrews, J. Chem. Phys. 98, 3690 ͑1993͒.
4 R. D. Hunt, J. T. Yustein, and L. Andrews, J. Chem. Phys. 98, 6070
͑1993͒.
5 P. Pyykko, Chem. Rev. 88, 563 ͑1988͒.
6 M. Pepper and B. E. Bursten, Chem. Rev. 91, 719 ͑1991͒.
Zhou, Mingfei
