Two-Dimensional Penning Ionization Electron Spectroscopy of NNO, HCNO, and HNNN: Electronic Structure and the Interaction Potential with He*(23S) Metastable and Li(22S) Ground State Atoms

Article abstract of DOI:10.1021/jp991394n

The electronic structure and Penning ionization of NNO, HCNO, and HNNN upon collision with He*(2³S) metastable atoms were studied using He I photoelectron and two-dimensional Penning ionization electron spectroscopies (2D-PIES). From the peak s

Full text of DOI:10.1021/jp991394n

6746
J. Phys. Chem. A 1999, 103, 6746-6756
Two-Dimensional Penning Ionization Electron Spectroscopy of NNO, HCNO, and HNNN:
Electronic Structure and the Interaction Potential with He*(23S) Metastable and Li(22S)
Ground State Atoms
Tibor Pasinszki,*,† Naoki Kishimoto, and Koichi Ohno*
Department of Chemistry, Graduate School of Science, Tohoku UniVersity, Aramaki, Aoba-ku,
Sendai, 980-8578, Japan
ReceiVed: April 28, 1999; In Final Form: June 28, 1999
The electronic structure and Penning ionization of NNO, HCNO, and HNNN upon collision with He*(23S)
metastable atoms were studied using He I photoelectron and two-dimensional Penning ionization electron
spectroscopies (2D-PIES). From the peak shifts in PIES and collision energy dependence of partial Penning
ionization cross sections, the interaction potentials between molecules and He*(23S) atoms were deduced. In
the studied collision energy range, the interaction potential was found to be attractive around the nitrile oxide
(-CNO) and azide (-NNN) groups, but no characteristic interaction was observed between NNO and
He*(23S). Ab initio calculations on the similar interacting systems M-Li(22S) (where M ) NNO, HCNO,
and HNNN) at the CCSD/6-311++G** level revealed fine details of the anisotropy of the interaction potentials
and were in good agreement with experimental results. The spectroscopic investigations predicted the existence
of thermodynamically stable MLi radicals, and the structure and stability of HCNOLi and HNNNLi were
calculated at the QCISD/6-311++G** level.
M + He*(23S) f Mi+ + He + e-
Eek ) EHe* - IPi + ∆Ei
(1)
(2)
I. Introduction
Chemical reactions, energy transfer, electron transfer, and
molecular collisions between molecules take place through a
close intermolecular contact, and they are strongly influenced
or governed by the interaction potential between the reacting
species; therefore, in order to understand the dynamics and
kinetics of intermolecular processes it is important to study and
derive information about the latter. Besides various molecular
beam experiments, which can partially or fully fulfill this role,
kinetic energy resolved Penning spectroscopy is a novel and
unique electron spectroscopic technique which provides infor-
mation directly about the interaction potential between mole-
cules and metastable rare gas atoms. Since the triplet meta-
stable rare gas atom-molecule interaction potential is very
similar to those of alkali metal-molecule potentials, the method
provides indirect information about these latter chemically im-
portant interactions and has a predicting power for the existence
and possible structure of stable alkali metal complex radi-
cals.1 Penning spectroscopy, furthermore, provides information
about the exterior regions of the molecules with indicating places
for electrophile attack in chemical reactions, because the
metastable atoms cannot penetrate into the interior regions of
molecules.
M + hν f Mi+ + e-
Eek ) Ehν - IPi
The technique of PIES is similar to the UV photoelectron
spectroscopy (UPS), where the energy of an electron ionized
by single photon absorption is measured (eq 2). One of the
important differences between PIES and UPS is that the energy
difference between the measured electron kinetic energies and
the known metastable excitation energy does not provide the
IPs, unlike to UPS; thus, there is a small peak energy shift (∆Ei)
if PIES and UPS spectra are compared to each other on an
electron energy scale. This peak energy shift depends on the
difference between the incoming M + He*(23S) (called interac-
tion potential) and outgoing Mi+ + He potential curves, but
assuming a flat potential to the outgoing channel in the
ionization region, it is determined by the incoming potential,
thus, vice versa, information on the interaction potential can be
obtained from the peak energy shifts. In a simplified sense, one
can expect that ∆Ei is positive if the interaction is repulsive
and negative if the interaction is attractive. In this latter case,
the peak shift is the measure of the well depth of the attractive
interaction potential. Since there is a similarity between M +
He*(23S) and M + Li(22S) interaction potentials, and given that
this latter is the ground state potential of the M-Li system, the
large negative peak shift (deep well on the potential surface) in
PIES indirectly indicates the existence of thermodynamically
stable M-Li radicals. Since PIES bands originate from the
ionization of molecular orbitals (MOs), which are more or less
localized on a special part of the molecule, the structure or
In Penning ionization electron spectroscopy (PIES), molecules
(M) are collided with metastable rare gas atoms, e.g., He*(23S),
having higher excitation energy than the ionization potential
(IP) of the molecules, and the kinetic energy (Eek) of the ejected
electrons (e-) is analyzed (eq 1).
* Corresponding author. E-mail: ohnok@qpcrkk.chem.tohoku.ac.jp.
† Permanent address: Department of Inorganic Chemistry, Technical
University, H-1521 Budapest, Hungary.
10.1021/jp991394n CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/10/1999

Penning Spectroscopy of NNO, HCNO, and HNNN
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6747
complexation site of M to Li can also be derived from the PIES
spectrum. On the basis of PIES investigations, we predicted,
e.g., the existence of CH3CNLi, and recently this radical was
identified in our laboratory in the gas phase by laser evaporation
and subsequent reaction of lithium metal with CH3CN vapors.2
CH3CNLi was also identified very recently by P. H. Kasai in
solid argon matrix using ESR spectroscopy.3
SCHEME 1
Another characteristic of the Penning ionization process,
which can be used to derive information about the interaction
potential even if the peak shift is not clearly observed due to
broad and/or overlapping bands, is the collision energy depen-
dence of the ionization cross section. The energy of the electron
released in the ionization process provides information about
the ionic state formed, and electron intensities for respective
bands determine the partial ionization cross sections. By
measuring the Penning ionization cross section as a function of
both the electron energy and metastable atom collision energy
(two-dimensional PIES, 2D-PIES), it is possible to obtain not
only the dependence of the total ionization cross section but
the dependence of the partial ionization cross sections as well.
These latter, as ionic states originate from removing an electron
from a MO, provide information about the anisotropy of the
interaction potential depending on the localization of MOs. One-
dimensional cuts from the two-dimensional spectrum give the
collision energy dependence of partial ionization cross sections
(CEDPICS; ionic state is fixed) and collision energy resolved
PIES (CERPIES; metastable atom kinetic energy is constant).
and HNNN, and the first study of their interaction with
He*(23S) and Li atoms using 2D-PIES and ab initio calcula-
tions. Relevant to this work are the earlier He I photoelec-
tron spectroscopic investigations of NNO,17,18 HCNO,19 and
HNNN19-21 and the PIES investigation of NNO.22,23 Of
particular interest are the electronic structure of molecules, the
anisotropy of the interaction between molecules and He*(23S)
or Li(22S) atoms, as well as the possible existence, stability,
and structure of M-Li inorganic radicals.
II. Experimental Section
NNO was a commercial product (Showa Denko) and was
used without further purification. HCNO24 and HNNN25 were
synthesized according to literature methods, as briefly described
below (see Scheme 1). 3-Phenyl-4-oximino-isoxazol-5-(4H)-
one was synthesized by nitrosation of 3-phenyl-5-isoxazolone
(Aldrich).26 NaNNN and stearic acid were commercial products
(Nacalai Tesque).
The 2D-PIES method has been recently developed in our
laboratory,4 and in combination with the cross-correlation time-
of-flight technique with pseudorandom chopper,5 it turned out
to be a very powerful tool to study the interaction between
molecules and metastable atoms. The most important questions
we want to answer with a series of studies are the characteristics
and anisotropy of the interaction potential, the effect of
electronegativity, valence, and bonding of atoms on this, the
interaction around chemical groups or bonds, and the influence
of substituents on these. Previous CEDPICS1,6-14 and recent
2D-PIES15,16 studies on various molecular targets have indicated
that the interaction potential is anisotropic between molecules
and He*(23S) atoms and pointed out the importance of the
chemical groups in describing the interaction. It has been
observed that the interaction is repulsive around saturated
hydrocarbons, alkyl groups, or CH bonds1,6-10,13,15,16 but at-
tractive around the π region of unsaturated hydrocarbons,7,8
heterocycles,10 and CdO double bonds.8 Studies of alcohols,9,15
aldehides,8 ethers,9,11 amines,12 isocyanates,13 nitriles,1,12,16 and
isonitriles1 have indicated that the interaction is attractive in
the lone electron pair region of first-row nonmetal atoms (O,
N, and C). There is relatively little known about the molecules
containing heavier elements so far, and these studies have
focused on organic chlorides and on sulfur compounds in
comparison with the oxygen analogue. According to these
studies, the interaction potential is attractive around the chlorine
atom,6,14 but in the case of sulfur atom it is strongly depend on
the molecular environment; a very attractive potential was found
in the sulfur lone pair region in alkyl thioethers and thioalco-
hols,15 but no special character (only weakly attractive or weakly
repulsive) was detected in the sulfur lone pair region of methyl
thiocyanate,13 methyl isothiocyanate,13 and thiophene.10 The
interaction potential, with He* or Li, around nitrile oxide
(-CNO) and azide (-NNN) functional groups has never been
studied before.
HCNO was synthesized by flush vacuum thermolysis of
3-phenyl-4-oximino-isoxazol-5-(4H)-one.24 The thermolysis was
carried out in a quartz tube (no filling, 12 mm i.d.) heated along
25 cm at 450 °C. The vapors of the precursor were passed
through the furnace from a flask held at 110-120 °C. The
furnace was directly connected to the spectrometer via two
consecutive U-traps. The temperature of the first trap was held
around -50 °C to trap phenyl cyanide, and the temperature of
the second was so adjusted, by adjusting the distance between
the surface of liquid nitrogen in a Dewar-flask and the bottom
of the U-trap, that the CO2 side product could just pass through
into the spectrometer; therefore, the formation of CO2 could be
continuously monitored during the generation of HCNO. Es-
sentially pure HCNO was collected in the second trap. Since
the vapor pressure of the precursor was low at 110-120 °C,
the sample was collected in the course of 14 h; then the trap
was separated and the temperature was raised to obtain sufficient
vapor pressure for spectroscopic investigations. We observed
only trace amounts of the isomer HNCO, which was pumped
off at low temperature. (HNCO was slightly more volatile than
HCNO.)
HNNN was synthesized by solid-liquid reaction between
sodium azide and molten stearic acid.25 NaNNN and solid stearic
acid were mixed at room temperature, placed into a flask, which
was connected directly to the spectrometer via a U-trap, and
gradually warmed above the melting point of the acid to obtain
a fast continuous bubbling of the gaseous products. The volatiles,
which was essentially pure HNNN, were condensed in the
U-trap with liquid nitrogen; only trace amounts of CO2 and N2
were detected. The temperature of the trap was than increased
to obtain sufficient vapor pressure for the spectroscopic
investigations. With this method, we could produce 1-2 mL
of pure liquid HNNN in 20-30 min and did not experience
any problem with handling the potentially explosive HNNN at
low temperatures.
In this paper we present a combined experimental (UPS,
PIES) and theoretical study of the structure of NNO, HCNO,

6748 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Pasinszki et al.
The instrument used in this work for recording the UPS, PIES,
and 2D-PIES spectra was reported in previous papers.4,6,7,27 UPS
spectra were measured by utilizing the He I resonance line
(21.22 eV) produced by a pure helium discharge. Metastable
atoms for PIES were produced by a negative discharge nozzle
source, and the He*(21S) component of the He*(21S,23S) beam
was quenched by a water-cooled helium discharge lamp. The
kinetic energies of electrons ejected by photo or Penning
ionization were determined by a hemispherical electrostatic
deflection type analyzer using an electron collection angle of
90° to the incident photon or He*(23S) beam axis. The energy
resolution of the electron analyzer was 50 meV, estimated from
the full width at half-maximum (fwhm) of the Ar+(2P3/2) peak
in the He I UPS spectrum. The transmission of the electron
energy analyzer was determined by comparing our UPS data
of O2, CO, N2, and some hydrocarbons with those of Gardner
and Samson28 and Kimura et al.18
In the collision energy resolved experiments, 2D-PIES, the
metastable atom beam was pulsed by a pseudorandom chopper
and introduced into the reaction cell located 504 mm down-
stream from the chopper disk. As reference, the intensities of
metastable atoms were determined by inserting a stainless steel
plate into the reaction cell and measuring the intensity of
secondary electrons emitted. The resolution of the electron
energy analyzer was lowered to 250 meV (fwhm for He I UPS
of Ar) in order to gain higher electron counting rates. Thus, in
these experiments, the intensity of emitted electrons from sample
toward the center of mass of the molecules and keeping the
molecular geometries fixed at the experimental values deter-
mined from microwave spectroscopic data;30 this latter assump-
tion meant that the geometry relaxation by the approach of the
metastable atom was negligible in the ionization process. All
calculations for the interaction potential were done at the CCSD-
(fc)/6-311++G** level of theory, and the full counterpoise (CP)
method31 was used to correct for the basis set superposition
errors (BSSE).
The structures of HCNOLi and HNNNLi inorganic radicals
were fully optimized at the QCISD(fc)/6-311++G** level of
theory, taking advantage of analytic first derivatives at this level
in Gaussian-94, and then harmonic vibrational frequencies were
calculated at the equilibrium geometries using numeric second
derivatives to make sure they were real minima on the potential
energy surface. In the frequency calculations, divergency
problems in the post-HF iteration cycles were obtained, but this
could be overcome by changing the default step size, except
HCNOLi(II); therefore, in the case of HCNOLi(II), harmonic
frequencies were calculated using the B3LYP/6-311++G**
method. Dipole moments and total atomic charges were
calculated using the QCISD density and the natural population
analysis. All calculations were performed with the Gaussian-
94 quantum chemistry package32 implemented on Silicon
Graphics, Inc. Challenge/XL and Origin200 workstations.
The ionization potentials for NNO, HCNO, and HNNN were
calculated using the outer valence Green’s function (OVGF)
method33 as incorporated in Gaussian-94, and also with the
semiempirical HAM/3 (hydrogenic atoms in molecules) method,34
which was shown to give an accurate representation of IPs for
molecules containing first-row atoms.35 OVGF and HAM/3
calculations were performed at the experimental geometries of
molecules.30
molecules (Ie) or from a reference stainless steel plate (IHe*
)
was measured as a function of electron kinetic energy (Eek) and
time (τ). Electron energies were scanned by 35 meV steps, and
a dwell time for the time-dependent measurement was 3 µs.
The 2D electron intensity spectra, I(Eek,τ), were then converted
sequentially to I(Eek,τTOF) and I(Eek,νHe*) (where τTOF is the time-
of-flight and νHe* is the velocity of the metastable atoms). The
2D Penning ionization cross sections σ(Eek,νr) were obtained
from I(Eek,νHe*) using eqs 3 and 4, and finally σ(Eek,νr) was
converted to the 2D-PIES, σ(Eek,Ec) using eq 5.
IV. Results
Figures 1-3 show the He I UPS and He*(23S) PIES spectra
of NNO, HCNO, and HNNN, respectively. The electron energy
scales for PIES spectra are shifted relative to those of UPS by
the difference in the excitation energies; 21.22-19.82 ) 1.40
eV.
Figures 4-6 show the CERPIES spectra of NNO, HCNO,
and HNNN, respectively. The spectra are obtained from the
2D-PIES spectra by cutting a small kinetic energy region
corresponding to ca. 20 µs TOF of He*. In each figure, the low
collision energy spectrum (ca. 90-110 meV, average 100 meV)
is shown by a dashed curve, and the high collision energy
spectrum (ca. 220-290 meV, average 250 meV) is shown by
a solid curve.
σ(Eek,νr) ) c[Ie(Eek,νHe*)/IHe*(νHe*)](νHe*/νr)
νr ) [νHe*2 + 3kT/M]1/2
(3)
(4)
(5)
Ec) µνr2/2
where c, νr, k, T, M, and µ are a constant, the relative velocity
of metastable atoms averaged over the velocity of the target
molecule, the Boltzmann constant, the gas temperature, the mass
of the target molecule, and the reduced mass of the system,
respectively.
Figures 7-9 show the log σ vs log Ec plots of CEDPICS for
NNO, HCNO, and HNNN, respectively. The CEDPICS are
obtained from the 2D-PIES spectra by cutting an appropriate
range of electron kinetic energy, Ee (typically the fwhm of the
corresponding PIES band). The calculated electron density maps
of the molecular orbitals are also shown in the figures (the thick
solid curve in maps indicates the molecular surface, estimated
from the van der Waals radii of atoms); electron density contour
maps in a symmetry plane of the molecule are shown for the
orbitals of NNO and HCNO, as well as for the a′ orbitals of
HNNN. For the a′′ orbitals of HNNN, one of those planes is
selected, which are parallel to the symmetry plane of the
molecule and lying just above the van der Waals radius of
nitrogen.
III. Calculations
To assist with experimental data, the interaction potential
between target molecules (M) and metastable He*(23S) atoms
was modeled by approximating the M-He*(23S) surfaces with
those of M-Li(22S). Using this widely accepted approximation,
based, e.g., on cross-scattering experiments indicating very
similar shape for the velocity dependence of the total scattering
cross section and for the location and depth of the well of the
attractive interaction potential for He*(23S) and Li(22S) with
various atomic and molecular targets,29 all of the difficulties
could be bypassed that would be associated with calculating
the excited-state M-He* surfaces. Thus, the M-Li(22S)
interaction potentials, V*(R,θ) (where R is the distance from
the center of mass (X) of the molecule, and θ is the Li-X-O
or Li-X-N(3) angle) were calculated by pulling the Li atom
Figures 10-12 show calculated potential energy curves
between a ground-state Li atom and NNO, HCNO, and HNNN,

Penning Spectroscopy of NNO, HCNO, and HNNN
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6749
Figure 2. He I UPS and He*(23S) PIES spectrum of HCNO.
Figure 1. He I UPS and He*(23S) PIES spectrum of NNO.
Table 2 lists calculated harmonic vibrational frequencies,
infrared intensities, rotational constants, dipole moments, total
energies, barriers to linearity or planarity, and bonding energies
of HCNOLi and HNNNLi radicals. Calculated results at the
QCISD(fc)/6-311++G** level of theory are shown, except the
harmonic vibrational frequencies and IR intensities of HCNOLi-
(II), which were obtained at the B3LYP/6-311++G** level.
respectively. The potential energy is shown as a function of
the distance between the Li atom and the center of mass of the
molecule. Calculations are done at the CCSD(fc)/6-311++G**
level of theory.
Figure 13 shows the calculated structure of HCNOLi and
HNNNLi radicals, as well as the calculated total atomic charges
(natural population analysis was done using the QCISD density).
Calculations were done at the QCISD(fc)/6-311++G** level
of theory.
V. Discussion
Figure 14 illustrates the major orbital interactions between
HNNN and Li, explaining the formation of HOMO and LUMO
of HNNNLi. Orbital energies were calculated at the QCISD-
(fc)/6-311++G** geometries.
Figure 15 shows the major orbital interactions between HCNO
and Li, explaining the formation of the HOMOs and LUMOs
of HCNOLi(I and II) isomers. Orbital energies were calculated
at the QCISD(fc)/6-311++G** geometries.
Table 1 lists experimental and calculated ionization potentials
(IPs), experimental peak energy shifts (∆E), slope parameters
(m), and the assignment of the spectra. Slope parameters are
obtained from the log σ vs log Ec plots (for details of explaining
the expected linear relationship between log σ and log Ec see,
e.g., ref 1). Vertical IPs are determined from the He I UPS
spectra. The peak energy shifts in PIES spectra are obtained as
the difference between the peak position (EPIES; electron energy
scale) and the “nominal” value (E0 ) difference between the
metastable excitation energy and target IP), ∆E ) EPIES - E0.
A. UPS and PIES Spectra. Since the information derived
on the interaction potential in the sequence of removing an
electron from a MO f formation of ionic state f collision
energy dependence of the ionization cross section f interaction
potential strongly depends on the localization of MOs, it is
important to unambiguously assign the UPS/PIES bands. From
comparison with the known UPS spectra of NNO17,18 and
HCNO,19 as well as from the calculated IPs (see Table 1), the
assignment for NNO and HCNO is relatively straightforward,
especially since, for these linear molecules, there is a complete
π/σ separation. Our assignment (Table 1) confirms previous
investigations. Four bands are observed in the He I UPS
spectrum of NNO, which originate, in the order of increasing
IPs, from the degenerate nonbonding πnb(NNO), nitrogen lone
pair nN, and degenerate bonding πb(NNO) orbitals, as well as
the oxygen lone pair orbital nO. In the case of HCNO, the four
observed UPS bands originate from the degenerate nonbonding
πnb(CNO) and bonding πb(CNO) orbitals, as well as the oxygen

6750 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Pasinszki et al.
Figure 5. Collision energy resolved He*(23S) PIES spectra of HCNO.
Figure 6. Collision energy resolved He*(23S) PIES spectra of HNNN.
Figure 3. He I UPS and He*(23S) PIES spectrum of HNNN.
the end atoms of the NNO or CNO group; thus, the two π
orbitals probe different parts of the interaction potential surface
(see Figures 7 and 8).
HNNN is not a linear molecule. If it were linear, the MOs of
HNNN would be of the same character as those of NNO and
HCNO, and the π orbitals would remain degenerate. In the real
molecule, however, the degeneracy of both the bonding and
nonbonding π orbitals is lifted to produce two pairs of orbitals
of symmetry a′′ and a′. In the simplest model (see Scheme 2)
the orbital a′ derived from the nonbonding π orbital becomes a
lone pair on the nitrogen atom and the a′ orbital derived from
the bonding π orbital becomes a pure NdN bond. The out-of-
plane orbitals retain their original π character. The extent of
the lone pair and double bond character, or their delocalization,
however, depends on the deviation from linearity, and in the
practical case, both the delocalized and localized model (2a and
2b in Scheme 2) can serve as a starting point of the description
of the electronic structure of HNNN and can be regarded as
the two most important mesomeric structure. For the simplest
discussion, we use the πnb and πb symbols for HNNN as well,
similar to NNO and HCNO, but we have to keep in mind the
structural difference suggested by the model 2b in Scheme 2.
Six bands are observed in the He I UPS spectrum of HNNN,
the fifth is only as a weak shoulder at 17.7 eV (Figure 3). This
band was not observed, due to a purer signal/noise ratio in
previous He I investigations,19-21 but is detected in the most
recent He II work of Lee et al.21 Probably the lack of observing
Figure 4. Collision energy resolved He*(23S) PIES spectra of NNO.
lone pair orbital nO, and σ(C-H) orbital. Due to the slightly
smaller excitation energy of He*(23S) than He I photons, the
fourth bands at higher IPs are not observed in PIES spectra.
On the basis of three bands in the PIES spectra of NNO and
HCNO, therefore, information on the interaction potential can
be deduced around the π region and at one end of the linear
frame at the nitrogen and oxygen lone pair, respectively. The
πb orbitals have higher electron density at the central nitrogen
atom, but πnb has a nodal surface here and more localized on

Penning Spectroscopy of NNO, HCNO, and HNNN
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6751
Figure 9. Collision energy dependence of partial ionization cross
Figure 7. Collision energy dependence of partial ionization cross
sections for HNNN with He*(23S).
sections for NNO with He*(23S).
Figure 10. Calculated interaction potential curves between NNO and
Li(22S) as functions of distance between the Li atom and the center of
mass of the molecule.
previous comment of Eland that the splitting between in-plane
and out-of-plane π orbitals of pseudohalo acids is less for the
bonding than for the nonbonding ones.20a The six bands in the
He I UPS spectrum of HNNN thus originate, in the order of
increasing IPs, from the nonbonding πnb(a′′) and πnb(a′), nitrogen
terminal lone pair nN, and πb(a′′) and πb(a′) orbitals, as well as
the σ(N-H) orbital. This latter one is not observed in the PIES
spectrum due to the He*(23S) cut off. There is, however, an
additional peak in the PIES spectrum in the high IP region
(marked with “S” in Figure 3), which is not observed in He I
UPS, and we believe it arises from simultaneous ionization-
excitation “shake-up” process. Penning spectroscopy is known
to enhance two-electron processes, and it is also known that
their probability increases by increasing the photon energy in
photoelectron spectroscopy. Indeed, a small feature can be seen
in the He II UPS (Figure 2 in ref 21) at the same place between
Figure 8. Collision energy dependence of partial ionization cross
sections for HCNO with He*(23S).
this band, as well as the failure of Koopmans theorem to predict
reliable IPs for pseudohalides in the high IP region, is the reason
that in previous works the high IP region is misassigned
suggesting an exceptionally large splitting of 2.7-3.8 eV
between πb(a′′) and πb(a′).19,20c,21 Our new assignment (see Table
1) is based on our new calculations, in agreement with a
previous calculation using Rayleigh-Schro¨dinger perturbation
theory to determine corrections to Koopmans’ theorem,36 and
careful comparison between the UPS spectra of isovalence
electronic molecules HNNN, HNCO, and HNCS.20a,37 These
all suggest that the splitting between πb(a′′) and πb(a′) must be
smaller than 1 eV. This is also in good agreement with the

6752 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Pasinszki et al.
Figure 11. Calculated interaction potential curves between HCNO and
Li(22S) as functions of distance between the Li atom and the center of
mass of the molecule.
Figure 13. Calculated structure (distances in angstrom) and total atomic
charges of HNNNLi (A), HCNOLi(I) (B), and HCNOLi(II) (C) radicals.
Figure 12. Calculated interaction potential curves between HNNN and
Li(22S) as functions of distance between the Li atom and the center of
mass of the molecule.
Figure 14. Major orbital interactions between HNNN and Li explaining
the formation of HOMO and LUMO of HNNNLi (energy scale is given
in atomic units). Orbitals with a′′ symmetry are not shown in the figure
for clarity. Orbitals of HNNN are shown at the equilibrium geometry
(a) and at the geometry in the complex (b).
18 and 19 eV. According to the observed PIES bands, therefore,
the interaction potential can be investigated in the π region and
the nitrogen terminal lone pair region of HNNN.
The PIES spectra of NNO, HCNO, and HNNN are shown in
Figures 1-3. The branching ratios are clearly different compared
to those in UPS spectra, which reflect the difference in the
ionization mechanism; strong bands in PIES originate from
orbitals having large electron density exposed outside to the
molecular surface. A general feature of the PIES spectra of the
investigated molecules is that the terminal lone pair band (nN
or nO) is strongly enhanced, especially in case of HCNO (note

Penning Spectroscopy of NNO, HCNO, and HNNN
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6753
Figure 15. Major orbital interactions between HCNO and Li explaining the formation of the HOMOs and LUMOs of HCNOLi(I) and HCNOLi-
(II) (energy scale is given in atomic units). Orbitals with a′′ symmetry are not shown in the figure for clarity. Orbitals of HCNO are shown at the
equilibrium geometry (a) and at the geometry in the complex (b, c).
TABLE 1: Band Assignments, Ionization Potentials (IP/eV),
Peak Energy Shifts (∆E/meV), and Slope Parameters (m) for
NNO, HCNO, and HNNN
negative for all bands, and in some cases, they are large negative
values of ca. -200 meV (see Table 1), which indicate that the
interaction between these molecules and He*(23S) is very
attractive. The peak shifts also show that the interaction potential
is anisotropic around the molecules. In the case of HCNO, there
is a large negative peak shift for the oxygen lone pair band and
for πnb(CNO) band. The πnb(CNO) orbital also has a large
electron density on the oxygen atom; thus, according to the peak
shifts, the most attractive part of the attractive surface must be
around the oxygen atom. The splitting between πnb(a′′) and
πnb(a′) peaks in PIES spectrum of HNNN is large enough to
observe peak shifts separately. There is a large negative peak
shift for the πnb(a′) band, which indicates that the interaction
potential is attractive in the πnb(a′) orbital region. This MO has
imine-type nitrogen lone pair character (see Scheme 2), and
probably this lone pair region is the most attractive part of the
molecule. Since the πb(a′) orbital also has large electron density
in the same region, the large negative peak shift for this latter
is in agreement. The peak shift is much smaller for the πnb(a′′)
band, which shows that the interaction is less attractive if the
metastable atom approaches the molecule perpendicular to the
molecular plane compared to the in plane direction. The peak
shift of -40 meV for the terminal nitrogen lone pair nN band is
surprisingly small, although it shows attractive interaction,
compared to that of the nitrogen lone pair of alkyl nitriles, where
a large peak shift (-270 to -390 meV) and very attractive
potential has been observed.1,16 Probably the three electro-
negative nitrogen atoms, and thus the smaller electron density
on the terminal nitrogen, are responsible for the less attractive
interaction. Regarding the attractivity of the terminal nitrogen
lone pair region, HNNN is between the nitrogen molecule N2,
where the interaction is repulsive,38 and R-CN, where it is very
attractive.
IP/eV
orbital
∆E/meV
m
molecule band exptl HAM3 OVGF character ((20 meV) ((0.03)
NNO
1
2
3
4
1
2
3
4
1
2
3
4
5
6
12.88 12.48 12.56 2π(πnb)
16.41 16.51 16.55 7σ(nN)
18.07 18.67 18.87 1π(πb)
20.16 20.68 20.64 6σ(nO)
10.77 10.92 10.61 2π(πnb)
16.00 17.07 16.77 1π(πb)
17.84 18.77 18.38 7σ(nO)
19.47 20.70 20.11 6σ(CH)
10.73 10.38 10.24 2a′′(πnb)
12.28 12.05 12.02 9a′(πnb)
15.53 15.36 15.96 8a′(nN)
16.92 17.31 17.07 1a′′(πb)
(17.7) 17.27 17.21 7a′(πb)
19.91 20.47 20.74 6a′(σNH)
-15
+20
+20
-0.08
0.00
+0.05
HCNO
HNNN
-200
-65
-210
-0.70
-0.58
-0.78
-70
-205
-40
-0.31
-0.34
-0.26
-0.27
-215
that π bands are doubly degenerate). Similar enhancements were
observed recently in spectra of nitriles,1,16 isocyanates, and
isothiocyanates;13 thus, this seems to be a characteristics of the
pseudohalide group and may explain the good complexing
ability of pseudohalides via terminal lone pair dative bond. Peak
shifts in PIES spectra are clearly observed. These are small,
around zero, for NNO in good agreement with previous
investigations22,23 and predict that there is no strong interaction
between NNO and He*(23S), and large collision energy
dependence of the ionization cross section is not expected. This
is in good agreement with the observed very small collision
energy dependence of the PIES branching ratios and partial cross
sections for the production of different ions in a previous PIES
and mass spectroscopic study, respectively, at two different
beam temperatures.23 Although the total ionization cross section
is not expected to depend on the collision energy, the small
peak shifts indicate that the interaction potential is not isotropic
around the molecule; the peak shift is slightly negative (-15
meV; Table 1) for the first band and positive (+20 meV) for
the second and third, which, regarding the localization of MOs,
indicates a small attractive interaction around the oxygen atom.
The peak shifts in PIES spectra of HCNO and HNNN are
B. 2D PIES and Interaction Potential. One-dimensional cuts
of 2D-PIES spectra, CERPIES and CEDPICS spectra are shown
in Figures 4-6 and 7-9, respectively, and calculated Li-M
interaction potential curves are shown in Figures 10-12. The
2D-PIES investigation on NNO indicates that there is practically
no collision energy dependence of the total ionization cross

6754 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Pasinszki et al.
TABLE 2: Calculated Vibrational Frequencies (cm-1), IR Intensities (km/mol), Rotational Constants (GHz), Dipole Moments
(Debye), Total Energies (au), Barrier (cm-1), and Bonding Energy (kJ/mol) of HCNOLi and HNNNLia
HCNOLi(I)
HCNOLi(II)
HNNNLi
frequency
3499 ν1(C-H str)
2330 ν2(CNO as str)
1212 ν3(CNO sym str)
530 ν4(CNO in-plane bend)
510 ν8(CNO out-of-plane bend)
399 ν5(H-CN in-plane def)
364 ν6(Li-O str)
354 ν9(H-CN out-of-plane def)
112 ν7(Li-ON def)
IR int
305.0
178.0
165.4
10.0
8.1
51.7
9.4
43.6
50.1
frequencyb
IR intb
1.0
197.6
25.4
276.2
49.3
35.3
107.6
16.0
71.7
frequency
3623 ν1(N-H str)
1688 ν2(NtN str)
1363 ν3(H-NN in-plane def)
935 ν4(NdN str)
694 ν5(ring in-plane vib)
613 ν6(H-NN out-of-plane def)
542 ν7(ring in-plane vib)
319 ν8(ring in-plane def)
102 ν9(ring out-of-plane def)
IR int
9.7
3137 ν1(C-H str)
1593 ν2(CN str)
128.5
98.3
213.0
21.8
70.5
64.5
55.0
99.3
1163 ν3(NO str)
974 ν4(H-CN in-plane def)
679 ν8(H-CN out-of-plane def)
671 ν5(ring in-plane vib)
565 ν6(ring in-plane vib)
425 ν7(ring in-plane def)
223 ν9(ring out-of-plane def)
HCNOLi(I)
HCNOLi(II)
HNNNLi
Ac
B
63.8374
6.0183
5.4998
Ac
B
24.6405
11.7338
7.9487
Ac
B
22.0870
12.4914
7.9992
C
C
C
µd
7.28 (8.53)
-175.644992
37.7
µd
4.33 (4.15)
-175.687386
149.0
µd
4.56 (4.80)
-171.901701
143.9
total energy
bonding energye
barrierf
total energy
total energy
bonding energye
barrierg
bonding energye
590.6
0.8
a Calculated at the QCISD/6-311++G** level of theory. b Calculated at the B3LYP/6-311++G** level of theory. c Isotopes: H-1, C-12, N-14,
O-16, Li-7. d Population analysis was done using the QCISD (SCF) density. e Difference between the total energy of the complex and the sum of
the energies of fragments. f Barrier to linearity. g Barrier to planarity.
SCHEME 2
PIES experiment shows that the interaction is attractive between
HNNN and He*(23S) around the NNN group (Figures 6 and
9). There are no large differences between the collision energy
dependence of the partial ionization cross sections (see slope
parameters in Table 1), but the largest negative dependence is
observed at the πnb(a′) band. In good agreement with the
observed peak shifts, this indicates that the region of the
corresponding πnb(a′) orbital, which has imine type lone pair
character, is the most attractive part of the molecule. Indeed,
the calculated Li-M potentials clearly indicate this “lone pair”
region as the most attractive part (Figure 12), and the attractivity
of the interaction is decreased as Li atom approaches at the other
end of the NNN group. Similarly to HCNO, the interaction is
repulsive around the hydrogen atom.
In general, there is a very good agreement between experi-
mental peak shifts, collision energy dependence of the partial
ionization cross sections, and calculated Li-M interaction
potentials. These show that the interaction potential is attractive
around the π and the terminal oxygen or nitrogen lone pair
region of -CNO and -NNN groups. The isoelectronic NNO
molecule has a similar electronic structure, especially compared
to HCNO, but the interaction is not attractive with He*(23S) or
Li(22S). When the oxygen or nitrogen atom in this molecule is
replaced by a HN or CH fragment, however, the -NNN and
-CNO groups acquire a negative charge,39 which may respon-
sible for the attractive interaction. Indeed, the most attractive
part of each molecule is on the most negatively charged atom,
and the positive hydrogen represents an opposite, repulsive
character. The total atomic charges, however, do not explain
the anisotropy of the electron density around atoms or chemical
groups, this can be obtained from the analysis of the molecular
orbital picture (see above). In this respect, the collision energy
dependence of the partial ionization cross sections, depending
on the localization of MOs, indicates the large electron density
regions of the molecules.
section (slope parameter +0.01 ( 0.03), and the dependence
of the partial ionization cross sections is very small, close to
zero (see Table 1). This latter, however, indicates small
anisotropy of the interaction potential, which is in good
agreement with the calculated Li-NNO potentials. According
to the calculations, the most attractive part of the molecule is
the oxygen lone pair region (the IP of the corresponding orbital
nO is beyond the He*(23S) cut off, thus, not investigated
experimentally), not the collinear direction though, but around
the Li-N-O angle of 30°, and the attractive potential gradually
changes to repulsive as the Li-N-O angle opens up to 180°.
Since πnb(NNO) MO has large electron density on the oxygen
atom, and πb(NNO) MO has the largest on the central nitrogen
atom, the small negative collision energy dependence of the
intensity of the πnb(NNO) band and small positive of the
πb(NNO) band is in good agreement with calculations. The nN
MO is partially mixed with nO (see Figure 7) and thus probes
to some extent the oxygen lone pair region as well, which
explains the smaller positive collision energy dependence of
the nN band than that of πb(NNO). The interaction between
HCNO and He*(23S) is very attractive in the π and nO region
(see Figures 5 and 8). According to our experience, it is very
rare to observe such a large negative collision energy depen-
dence of the ionization cross section between molecules and
He*(23S) atoms (see Table 1). The largest negative dependence
is observed at the nO band and then at the πnb(CNO), which
indicates that the most attractive part of the molecule is around
the oxygen atom. This is in good agreement with calculations
(Figure 11), showing a very attractive potential with deep well,
when the lithium atom approaches the molecule at the oxygen,
and this attractive potential gradually changes to repulsive as
the Li-N-O angle opens up. According to the calculations,
the interaction is repulsive around the hydrogen atom. The 2D-
C. HNNNLi and HCNOLi Radicals. The negative collision
energy dependence of the ionization cross section and the large
negative peak shift of the corresponding PIES band indicate
that there is a deep well on the interaction potential curve, which
indirectly predicts the existence of thermodynamically stable
MLi radicals. This is also in good agreement with the M-Li

Penning Spectroscopy of NNO, HCNO, and HNNN
J. Phys. Chem. A, Vol. 103, No. 34, 1999 6755
model potential calculations. According to these results (see
above), two complexes are expected to be thermodynamically
the most stable: one is the HCNOLi, where the Li is expected
to bond to the oxygen atom, and the second is HNNNLi, where
Li is attached to the imine type nitrogen lone pair (πnb(a′)). It
is important to note, however, that due to the fast speed of
metastable atoms and instantaneous ionization compared to the
molecular nuclear motion, the geometry of the molecules is
nearly frozen in the Penning ionization process, which is not
the case in a chemical reaction between molecules and Li atoms.
Due to geometry relaxation, the bonding energy of MLi
complexes can be much larger than is predicted from the peak
energy shifts in PIES (depth of the interaction potential). This
relaxation is strongly depend on the reacting systems and cannot
be predicted in advance. To get information about the stability
and also the structure of these radicals, ab initio calculations
have been performed for the structure of HCNOLi and
HNNNLi. Calculated results are shown in Figure 13 and Table
2. Full geometry optimizations at the QCISD level ran into the
expected minima of HNNNLi and HCNOLi(I) (I denotes isomer
I, see below). The structure of HCNOLi(I) may be best
characterized as a classical complex, where the HCNO ligand
bonds at the oxygen atom to Li with a coordinative bond. The
total atomic charge on Li is close to zero (+0.02), and the
lithium unpaired electron stays on Li. The geometry of HCNO
does not change significantly, and the molecule keeps its close
linear structure in the complex. The bond angle at the oxygen
is 126.3°, and the barrier of HCNOLi(I) to linearity is 590.6
cm-1, which is fairly low but definitely higher than the low
lying vibrational levels (the calculated harmonic frequency of
the N-O-Li bend is 112 cm-1). The structure and bonding of
HNNNLi is very different compared to that of HCNOLi(I).
HNNNLi is a π-type complex, where the Li atom bonds to both
ends of the NNN group forming a four-membered ring. The
equilibrium structure of the complex is not planar, but the barrier
to planarity is a mere 0.8 cm-1; thus, all vibrational levels lie
above this and the molecule has a quasiplanar structure. The
unpaired electron of the Li atom completely delocalizes on the
HNNN σ frame, and as a consequence there is a positive charge
of +0.88 on the Li atom. There is a large difference in the
bonding energies of these two complexes (37.7 and 143.9 kJ/
mol; Table 2). To understand the difference in bonding, the
formation of MOs of complexes from orbitals of their fragments
have been investigated, and results explaining the formation of
HOMOs are shown in Figures 14 and 15. As Figure 14
illustrates, the LUMO of HNNN is strongly stabilized by the
interaction of HNNN and Li, and the transfer of the lithium 2s
electron into this lower lying empty orbital explains the large
bonding energy of the complex. In the case of HCNOLi(I), the
LUMO(HCNO)-LUMO(Li) interaction is small and the forma-
tion of HOMO is mainly HOMO(HCNO)-HOMO(Li) con-
trolled. The HOMO of HCNO is stabilized, and the HOMO of
Li is destabilized; thus, there is a smaller energy benefit of this
process than is the case for HNNNLi. It seems that the formation
of a lower lying LUMO due to the bending of the NNN frame,
which can strongly interact with one of the 2p orbitals of Li, is
responsible for the large stabilization energy of the HNNNLi
complex. The difference between HCNO and HNNN in the
complex formation was surprising first, especially since model
calculations indicated the same change in electronic stucture
as that of HNNN by bending the CNO frame, but a search for
possible HCNOLi isomers identified the four-membered ring
π-complex HCNOLi(II). The structure and bonding of this
complex are very similar to those of HNNNLi (Figure 13 and
Table 2): the unpaired electron is delocalized on the CNO
frame, the charge on Li atom is +0.87, and the bonding energy
is large (149.0 kJ/mol). This complex is thermodynamically
much more stable than HCNOLi(I). During the QCISD potential
energy surface scan, divergence in the post-HF iteration cycles
has been experienced many times, as well as in the frequency
calculations; therefore, the structure of HCNOLi(II) was recal-
culated at the B3LYP level (see ref 40), which is in good
agreement with the QCISD results and confirms HCNOLi(II)
as a planar four-membered ring. From the model potential
calculations, it is clear that HNNNLi and HCNOLi(I) can form
from Li and HNNN and HCNO, respectively, without any
kinetic barrier. To obtain information on the formation of
HCNOLi(II) and the kinetic stability of HCNOLi(I), a potential
surface scan was done at the B3LYP level (due to divergence
problems in QCISD again). According to B3LYP calculations,
HCNOLi(II) forms from Li and HCNO without any kinetic
barrier and HCNOLi(I) is only a local minimum on the way to
the formation of HCNOLi(II) (the barrier of the isomerization
of HCNOLi(I) into HCNOLi(II) is 5.6 kJ/mol; see ref 41).
Calculations, therefore, predict that, in both HCNOLi and
HNNNLi, the π-complex is the only stable form. The identifica-
tion of both HCNOLi and HNNNLi is feasible either in an inert
solid matrix or in the dilute gas phase. To support future
identifications, calculated harmonic frequencies, infrared intensi-
ties, rotational constants, and dipole moments are shown in Table
2. Recently, there has been much interest in cluster chemistry
to study the development of solvation from small complexes to
the bulk phase, and one of the most interesting questions in
alkali atom interaction with polar solvent is the delocalization
of the atomic valence electron as more and more polar molecules
are bonded to the atom.42 In this respect, the alkali metal HNNN
and HCNO interactions may present interesting challenges as
the valence electron delocalization can possibly be achieved with
very small cluster sizes.
VI. Conclusion
The isoelectronic molecules NNO, HCNO, and HNNN have
been investigated in the gas phase through a combination of
electron spectroscopic and ab initio methods. Experiments show
that the interaction potentials between He*(23S) metastable
atoms and molecules are anisotropic and the interactions of
He*(23S) with NNO, HCNO, and HNNN are very different
compared to each other, reflecting the difference in the electronic
structures. The interaction between HCNO and He*(23S) is
unusually attractive, especially around the oxygen atom. At-
tractive interaction has been found between HNNN and
He*(23S), too, having the most attractive part at the imine type
nitrogen lone electron pair region. The interaction between
He*(23S) and NNO is not characteristic, not attractive nor
repulsive; the total ionization cross section does not show any
collision energy dependence, but the small dependencies of
partial ionization cross sections indicate that the interaction is
slightly attractive around the oxygen atom and slightly repulsive
at the other end of the molecule. A common feature of the
electronic structure of the investigated molecules is that the
electron density of terminal lone pair orbitals is exposed strongly
outside the molecular surface, indicating places for electrophile
attack. Our investigation on HNNN, however, also shows that
this large electron density outside the molecular surface does
not necessarily indicate the largest electron density part of the
molecule, which strongly determines the most attractive side
of the molecule. Ab initio calculations reveal fine details of the
Li-M(molecule) interaction potentials and are in good agree-

6756 J. Phys. Chem. A, Vol. 103, No. 34, 1999
Pasinszki et al.
(18) Kimura, K.; Katsumata, S.; Achiba, Y.; Yamazaki, T.; Iwata, S.
Handbook of He I Photoelectron Spectra of Fundamental Organic
Molecules; Japan Scientific Press: Tokyo, 1981.
(19) Bastide, J.; Maier, J. P. Chem. Phys. 1976, 12, 177.
(20) (a) Eland, J. H. D. Philos. Trans. R. Soc. London, Ser. A 1970,
268, 87. (b) Cradock, S.; Ebsworth, E. A. V.; Murdoch, J. D. J. Chem.
Soc., Faraday Trans. 2 1972, 68, 86. (c) Cvitas, T.; Klasinc, L. J. Chem.
Soc., Faraday Trans. 2 1976, 72, 1240.
(21) Lee, T. H.; Colton, R. J.; White, M. G.; Rabalais, J. W. J. Am.
Chem. Soc. 1975, 97, 4845.
(22) (a) Cˇ erma´k, V. J. Electron. Spectrosc. Relat. Phenom. 1976, 9,
419. (b) Brion, C. E.; Yee, D. S. C. J. Electron. Spectrosc. Relat. Phenom.
1977, 12, 77.
ment with the experimental results. They, furthermore, provide
information about the H atom region of the molecules, which
could not be probed by experiment, and indicate that the
interaction potential is repulsive around the hydrogen atom.
Both experimental and calculated results indicate the existence
of stable Li-M radicals, among those the thermodynamically
most stables are the four-membered rings HCNOLi and
HNNNLi. According to QCISD calculations, the unpaired
electron of the lithium atom completely delocalizes on the CNO
or NNN frame, and the stabilization of this latter explains the
large bonding energy of these π-type complexes. The formation
of π-complex and delocalization of the lithium valence electron
is an interesting result of this work, and we plan to further
investigate this. Our preliminary calculations on similar systems,
such as HNCO + Li, HNCS + Li, and NCCN + Li, indicate
the same effect found in this work, namely, the π-complex
formation results in the delocalization of the metal unpaired
electron and large bonding energy.
(23) Hotop, H.; Kolb, E.; Lorenzen, J. J. Electron. Spectrosc. Relat.
Phenom. 1979, 16, 213.
(24) Wentrup, C.; Gerecht, B.; Briehl, H. Angew. Chem., Int. Ed. Engl.
1979, 18, 467.
(25) (a) Gunther, P.; Meyer, R. Z. Elektrochem. 1935, 41, 541. (b)
Krakow, B.; Lord, R. C.; Neeby, G. O. J. Mol. Spectrosc. 1968, 27, 148.
(26) Claisen, L.; Zedel, W. Chem. Ber. 1891, 24, 140.
(27) Mitsuke, K.; Takami, T.; Ohno, K. J. Chem. Phys. 1989, 91, 1618.
(28) Gardner, J. L.; Samson, J. A. R. J. Electron. Spectrosc. Relat.
Phenom. 1976, 8, 469.
(29) (a) Rothe, E. W.; Neynaber, R. H.; Trajillo, S. M. J. Chem. Phys.
1965, 42, 3310. (b) Niehaus, A. AdV. Chem. Phys. 1981, 45, 399. (c) Hotop,
H. Radiat. Res. 1974, 59, 379. (d) Haberland, H.; Lee, Y. T.; Siska, P. E.
AdV. Chem. Phys. 1981, 45, 487.
(30) (a) Teffo, J. L.; Che´din, A. J. Mol. Spectrosc. 1989, 135, 389. (b)
Bunker, P. R.; Landsberg, B. M.; Winnewisser, B. P. J. Mol. Spectrosc.
1979, 74, 9. (c) Winnewisser, B. P. J. Mol. Spectrosc. 1980, 82, 220.
(31) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 10, 553.
(32) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, revision C.3; Gaussian,
Inc.: Pittsburgh, PA, 1995.
Acknowledgment. We thank the Japanese Ministry of
Education, Science, and Culture for a Grant in Aid for Scientific
Research in support of this work. T.P. thanks the Japan Society
for the Promotion of Science (JSPS) for a JSPS Invitation
Fellowship (IDNo. L98519) and the Hungarian Scientific
Research Found (OTKA Grant F022031) in support of this work.
References and Notes
(1) Pasinszki, T.; Yamakado, H.; Ohno, K. J. Phys. Chem. 1995, 99,
14678.
(2) Ohshimo, K.; Tsunoyama, H.; Yamakita, Y.; Misaizu, F.; Ohno,
K. Chem. Phys. Lett. 1999, 301, 356.
(3) Kasai, P. H. J. Am. Chem. Soc. 1998, 120, 7884.
(4) Ohno, K.; Yamakado, H.; Ogawa, T.; Yamata, T. J. Chem. Phys.
1996, 105, 7536.
(33) von Niessen, W.; Schirmer, J.; Cederbaum, L. S. Comput. Phys.
Rep. 1984, 1, 57.
(34) A° sbrink, L.; Fridh, C.; Lindholm, E. Chem. Phys. Lett. 1977, 52,
69. The HAM/3 program is available from the Quantum Chemistry Program
Exchange, Indiana University, Bloomington, IN (D. P. Chong, QCMP005,
1985).
(35) Chong, D. P. Theor. Chim. Acta 1979, 51, 55.
(36) Zeiss, G. D.; Chong, D. P. J. Electron. Spectrosc. Relat. Phenom.
1980, 18, 279.
(37) Pasinszki, T.; Kishimoto, N.; Ohno, K. Manuscript in preparation.
(38) Ohno, K.; Takami, T.; Mitsuke, K.; Ishida, T. J. Chem. Phys. 1991,
94, 2675.
(5) (a) Kishimoto, N.; Yamakado, H.; Yamata, T.; Ogawa, T.; Ohno,
K. Int. Conf. Phys. Electron. At. Collisions, 19th, 1995, 1995, 807. (b)
Auerbach, D. J. In Atomic and Molecular Beam Methods; Scoles, G., Ed.;
Oxford University Press: New York 1988; p 369.
(6) Takami, T.; Mitsuke, K.; Ohno, K. J. Chem. Phys. 1991, 95, 918.
(7) Takami, T.; Ohno, K. J. Chem. Phys. 1992, 96, 6523.
(8) Ohno, K.; Okamura, K.; Yamakado, H.; Hoshino, S.; Takami, T.;
Yamauchi, M. J. Phys. Chem. 1995, 99, 14247.
(9) Yamakado, H.; Yamauchi, M.; Hoshino, S.; Ohno, K. J. Phys.
Chem. 1995, 99, 17093.
(39) Calculated total atomic charges (natural population analysis using
the QCISD density; calculations were done at the optimized QCISD
geometries of molecules). N(1)N(2)O: N(1), -0.06; N(2), +0.40; O, -0.34.
HCNO: H, +0.23; C, -0.01; N, +0.19; O, -0.41. HN(1)N(2)N(3): H,
+0.34; N(1), -0.53; N(2), +0.23; N(3), -0.04.
(40) Calculated structure and rotational constants of HCNOLi π-com-
plex, HCNOLi(II), using the B3LYP/6-311++G** method: CH ) 1.088
Å, CN ) 1.261 Å, NO ) 1.301 Å, LiC ) 2.032 Å, LiO ) 1.823 Å, ONC
) 131.5°, NCH ) 118.2°; total energy ) -176.179464 au; bonding energy
) 157.3 kJ/mol; A ) 25.1756, B ) 11.8161, C ) 8.0417 GHz.
(41) Calculated structure and energy of HCNOLi(I) at the B3LYP/
6-311++G** level: CH ) 1.091 Å, CN ) 1.242 Å, NO ) 1.321 Å, LiO
) 1.615 Å, HCN ) 119.2°, CNO ) 129.3°, NOLi ) 172.2°; total energy
) -176.163340 au; bonding energy ) 115.0 kJ/mol; barrier to isomerization
) 5.6 kJ/mol.
(10) Kishimoto, N.; Yamakado, H.; Ohno, K. J. Phys. Chem. 1996, 100,
8204.
(11) Yamauchi, M.; Yamakado, H.; Ohno, K. J. Phys. Chem. A 1997,
101, 6184.
(12) Yamakado, H.; Ogawa, T.; Ohno, K. J. Phys. Chem. A 1997, 101,
3887.
(13) Pasinszki, T.; Yamakado, H.; Ohno, K. J. Phys. Chem. 1993, 97,
12718.
(14) Ohno, K.; Kishimoto, N.; Yamakado, H. J. Phys. Chem. 1995, 99,
9687.
(15) Kishimoto, N.; Yokoi, R.; Yamakado, H.; Ohno, K. J. Phys. Chem.
A 1997, 101, 3284.
(16) Kishimoto, N.; Aizawa, J.; Yamakado, H.; Ohno, K. J. Phys. Chem.
A 1997, 101, 5038.
(17) (a) Brundle, C. R.; Turner, D. W. Int. J. Mass Spectrom. Ion Phys.
1969, 2, 195. (b) Dehmer, P. M.; Dehmer, J. L.; Chupka, W. A. J. Chem.
Phys. 1980, 73, 126. (c) Cvitas, T.; Klasinc, L.; Kovac, B.; McDiarmid, R.
J. Chem. Phys. 1983, 79, 1565.
(42) Schulz, C. P.; Hertel, I. V. Solvated Atoms in Polar Solvents. In
Clusters of Atoms and Molecules II; Haberland, H., Ed.; Springer-Verlag:
Berlin-Heidelberg, 1994; pp 7-18.