Penning Ionization of NCCN by Experiment and Theory: A Two-Dimensional Penning lonization Electron Spectroscopic and Quantum Chemical Study

Article abstract of DOI:10.1021/jp991049y

Dicyanogen, NCCN, is generated for spectroscopic investigations on-line from rubeanic acid, mercury(II) cyanide, and cyanogen iodide and studied in the gas phase by two-dimensional Penning and He I photoelectron spectroscopies, as well as ab initio calculations. From spectroscopic data, the interaction between NCCN and He*(2³S) atoms is deduced. The interaction potential for the similarly interacting NCCN-Li(2²S) system is obtained from ab initio calculations at the CCSD/6-311++G** level. Experimental and calculated results show that the interaction potential is anisotropic around NCCN, is the most attractive in the nitrogen lone electron pair region, and gradually changes into repulsive as the N-C-He*(or Li) angle opens up to 90°. An unusual collision energy dependences of the partial ionization cross sections are observed, which is interpreted by the unusual interaction potential. For assisting experimental data and studying collision dynamics, classical trajectory calculations are performed for the Penning ionization of the NCCN-He*(2³S) system. The spectroscopic investigations predict the existence of thermodynamically stable MLi radicals, and the structure and stability of NCCNLi isomers are calculated at the QCISD/6-311++G** level.

Full text of DOI:10.1021/jp991049y

7170
J. Phys. Chem. A 1999, 103, 7170-7178
Penning Ionization of NCCN by Experiment and Theory: A Two-Dimensional Penning
Ionization Electron Spectroscopic and Quantum Chemical Study
,†
Tibor Pasinszki,* Naoki Kishimoto, Tetsuji Ogawa, and Koichi Ohno*
Department of Chemistry, Graduate School of Science, Tohoku UniVersity, Aramaki, Aoba-ku,
Sendai, 980-8578, Japan
ReceiVed: March 26, 1999; In Final Form: June 10, 1999
Dicyanogen, NCCN, is generated for spectroscopic investigations on-line from rubeanic acid, mercury(II)
cyanide, and cyanogen iodide and studied in the gas phase by two-dimensional Penning and He I photoelectron
spectroscopies, as well as ab initio calculations. From spectroscopic data, the interaction between NCCN and
3
2
He*(2 S) atoms is deduced. The interaction potential for the similarly interacting NCCN-Li(2 S) system is
obtained from ab initio calculations at the CCSD/6-311++G** level. Experimental and calculated results
show that the interaction potential is anisotropic around NCCN, is the most attractive in the nitrogen lone
electron pair region, and gradually changes into repulsive as the N-C-He*(or Li) angle opens up to 90°. An
unusual collision energy dependences of the partial ionization cross sections are observed, which is interpreted
by the unusual interaction potential. For assisting experimental data and studying collision dynamics, classical
3
trajectory calculations are performed for the Penning ionization of the NCCN-He*(2 S) system. The
spectroscopic investigations predict the existence of thermodynamically stable MLi radicals, and the structure
and stability of NCCNLi isomers are calculated at the QCISD/6-311++G** level.
I. Introduction
Two-dimensional Penning ionization electron spectroscopy
2D-PIES) has been recently developed in our laboratory, in
2
(
3
When a He*(2 S) metastable atom, having much higher
excitation energy than the ionization potential (IP) of the
molecule, collides with a molecular target (M), ionization may
occur, which yields the ground-state helium atom (He), molec-
which ionization cross sections are determined as a function of
both electron kinetic energy and metastable atom collision
energy. The method, is thus ionic state and collision energy
resolved, and makes it possible to study the collision energy
dependence of the total and partial ionization cross sections and
to derive information about the interaction between molecules
and metastable atoms. 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). Since an ionic state originates from
removing an electron from a molecular orbital (MO), the
dependence of the partial ionization cross section reflects the
interaction in the molecular region where the MO is localized;
therefore, it is possible to study the interaction between structural
elements of the molecule (e.g., chemical groups, chemical bonds,
electron lone pairs) and metastable atoms. Since there is a well-
+
ular positive ion in a given electronic state (Mi ), and a free
-
1
electron (e ):
+
-
He* + M f He + Mi + e
(1)
The ionization process may be divided into three parts: (I) the
approach of the metastable atom to the molecule; (II) ionization;
(
III) relaxation of the excited-state positive ion and secondary
ionic reactions. In Penning ionization electron spectroscopy
PIES), the electron ejected in this process is analyzed. The PIES
(
spectrum, therefore, provides primary information about the state
of the system in the “moment” of ionization and secondary
information about what happened before that. In particular, the
kinetic energy and number of electrons, which determines the
peak positions and relative branching ratios in the spectrum,
provides information about the electronic structure of the
molecule, the relative ionization probability of different ionic
states, and the potential energy difference between the incoming
3
2
known similarity between M + He*(2 S) and M + Li(2 S)
interaction potentials, and given that this latter is the ground
state potential of the M-Li system, indirect information can
be obtained about this latter chemically important interaction
3
+
M + He*(2 S) and outgoing Mi + He systems at the geometry
of ionization (II). By measuring the electron intensity, compared
to that of metastable atoms in the incident metastable beam,
the total and partial ionization cross section (σ) can be
determined. This latter, however, may depend on what happens
if molecules and metastable atoms get into close interaction (I);
attractive interaction is expected to increase the ionization cross
section by increasing the number of trajectories leading to
ionization, and repulsive interactions to decrease it.
3
too. We have recently successfully applied 2D-PIES to study
3
the anisotropy of the interaction potential between He*(2 S)
4,5
atoms and alcohols, thioalcohols, thioethers, and nitriles.
Although 2D-PIES is proved to be a powerful technique to
study the Penning ionization, interaction potential, and collision
3
dynamics of the interacting M + He*(2 S) system, it is rather
limited regarding this latter one. To get a deeper insight into
the collision dynamics and also to assist in analyzing experi-
mental data on a theoretical basis, we have recently developed
a novel theoretical method (see discussion section) to calculate
ionization cross section and collision energy dependence of the
*
Corresponding author. E-mail: ohnok@qpcrkk.chem.tohoku.ac.jp.
Permanent address: Department of Inorganic Chemistry, Technical
†
6
University, H-1521 Budapest, Hungary.
total and partial Penning ionization cross sections. As the first
1
0.1021/jp991049y CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/21/1999

Penning Ionization of NCCN by Experiment and Theory
J. Phys. Chem. A, Vol. 103, No. 36, 1999 7171
SCHEME 1
example, we applied this method to study the Penning ionization
3
+
- 6
of nitrogen (He*(2 S) + N2 f He + N2 + e ), and the
complete success of describing the collision energy dependence
of the total and partial ionization cross sections prompted us
for further testing. Nitrogen itself is a simple molecule with
high symmetry and, although the interaction potential is
anisotropic, it represents the case where the interaction is
3
repulsive with He*(2 S) atoms at any collision direction. It was
thus self-evident that we wanted to test the performance of our
method on a system where the interaction is attractive. Our
previous study on various nitriles (R-CN) indicated that the
3
interaction potential with He*(2 S) is attractive around the -CN
3
,5
group, especially in the nitrogen lone electron pair region.
Putting two of these groups together, the parent pseudohalogen
NCCN seemed to be the desired target. As it turns out from
this work, however, the ionization of NCCN has a very unusual
collision energy dependence in the investigated collision energy
region, and we might have unknowingly selected one of the
most difficult targets to describe. Another aim, to study NCCN
by 2D-PIES, came from our previous work on substituted
nitriles,3 where a very strong attractive interaction was observed
in the nitrogen lone pair region, and to find out substituent
effects, the study of the parent pseudohalogen seemed to be
essential. Simple pseudohalides seem to have an interesting and
varied electronic structure. Our current PIES and ab initio study
,5
3
7
Figure 1. He I UPS and He*(2 S) PIES spectrum of NCCN.
on pseudohalo acids (HCNO, HNNN, HNCO, HNCS) showed
that these molecules form thermodynamically stable four-
membered ring π-complexes with lithium atoms, where the
lithium 2s unpaired electron is delocalized on the acid. To
investigate the complex formation of Li with NCCN as an entr e´ e
for our future investigation on gaseous Lin(NCCN)m clusters
was another initiative of this work.
and this solid-gas reaction quantitatively produced CH3OCN
14
and H2O. In a similar manner, we expected that NCCN could
be generated from rubeanic acid (H2NC(S)C(S)NH2) by two-
fold H2S elimination. A mixture of rubeanic acid and HgO was
placed into a horizontal tube, which was attached to another
horizontal tube filled at the bottom with P4O10 to catch side
product water. The tubes were then connected via a U-trap
directly to the spectrometer, and the mixture was heated. By
increasing the temperature, the yield for NCCN gradually
increased, and at 80 °C the pressure was sufficient for UPS
and PIES investigations. Only a small amount of CO2, NH3,
and unvolatile dark-red solid (believed to be the well-known
reaction product of ammonia and NCCN; see ref 15) side
products were detected, whose relative amount gradually
increased by increasing the temperature. NH3 could be com-
pletely removed from the gas phase by cooling the U-trap down
to -30 °C. The raw volatile reaction products were introduced
directly into the spectrometer and the UPS spectrum obtained
is shown in Figure 1. Pure NCCN was obtained by collecting
the gaseous reactions products in a second U-trap with liquid
nitrogen, pumping off the more volatile CO2 at low temperature,
and reevaporizing NCCN. The PIES spectrum shown in Figure
1 is obtained by this method.
In this paper, we present a combined experimental and
3
theoretical study of the interaction of NCCN with He*(2 S) and
Li atoms using 2D-PIES and ab initio calculations. Relevant to
this work are the earlier photoelectron spectroscopic investiga-
8
-11
tions
singlet and triplet metastable atoms, He*(2 S,2 S).
particular interest are the electronic structure, the anisotropy of
and the early PIES investigations using a mixture of
1
3
12,13
Of
3
2
the interaction between NCCN and He*(2 S) or Li(2 S) atoms,
and the collision dynamics, as well as the possible existence,
stability, and structure of M-Li inorganic radicals.
II. Experimental Section
Generation of NCCN. NCCN has been known for a long
time and has become an important reagent in laboratories and
industry. However, it is very toxic and its handling, transporta-
tion, and storage requires special conditions. To get rid of these
difficulties, we thought it best to generate this gas on-line for
spectroscopic investigations. We considered three routes, from
rubeanic acid, mercury(II) cyanide (Nacalai tesque), or cyanogen
iodide (Aldrich), as discussed below and shown in Scheme 1.
It was observed recently that CH3OC(S)NH2 vapors reacted
fast with yellow mercuric oxide (HgO) at room temperature,
The thermal decomposition of Hg(CN)2 to NCCN has been
well-known for a long time, and indeed the thermal decomposi-
tion of the solid substance led to the discovery of NCCN in
16
1815 by Gay-Lussac. The thermolysis in this work was carried
out as follows. The solid Hg(CN)2 was placed into a one end

7172 J. Phys. Chem. A, Vol. 103, No. 36, 1999
Pasinszki et al.
blinded Pyrex tube (i.d. 12 mm; loosely packed with quartz
wool) and the tube was slowly pulled through a furnace heated
to 500 °C. The volatiles of this reaction were collected in a
U-trap with liquid nitrogen, and then the temperature was raised
to produce enough vapor pressure for NCCN for spectroscopic
investigations.
Solid ICN, when it is heated in a closed tube, is known to
decompose to NCCN and iodine for some time,17 and it has
sufficient vapor pressure to be easily driven through a furnace
in the gas phase under very well-controlled conditions. We
carried out the thermolysis in a quartz tube (12 mm i.d.) loosely
packed with quartz wool and heated along 25 cm at 800 °C.
The furnace was directly connected to the spectrometer via two
consecutive U-traps. The temperature of the first trap was held
around -60 °C to trap iodine, and NCCN was collected in the
second trap.
III. Calculations
The details of the trajectory calculations are given below in
the Discussion section. The interaction potential between target
molecules (M) and metastable He*(2 S) atoms were calculated
3
3
by approximating the M-He*(2 S) surfaces with those of
2
M-Li(2 S). 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
3
2
interaction potential for He*(2 S) and Li(2 S) with various
atomic and molecular targets,
be bypassed that would be associated with calculating the excited
state M-He* surfaces. Thus, the NCCN-Li(2 S) interaction
2
3,24
all of the difficulties could
2
potentials, V*(R,θ) (where R is the distance from the center of
mass (X) of the molecule, and θ is the Li-X-N angle), were
calculated by pulling the Li atom toward the center of mass of
the molecule and keeping the molecular geometry fixed at the
experimental values determined from infrared spectroscopic
Spectroscopic Investigations. The instrument used in this
work for recording the UPS, PIES, and 2D-PIES spectra was
reported in previous papers.2
,18-20
UPS spectra were measured
25
data; this latter assumption 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
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*-
1
1
3
(
2 S) component of the He*(2 S,2 S) beam was quenched by a
2
6
the full counterpoise (CP) method was used to correct for the
basis set superposition errors (BSSE).
water-cooled helium discharge lamp. The kinetic energies of
electrons ejected by photo or Penning ionization were deter-
mined by a hemispherical electrostatic deflection type analyzer
using an electron collection angle of 90° to the incident photon
The structure of NCCNLi inorganic radicals were fully
optimized at the QCISD(fc)/6-311++G** level of theory, 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.
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
3
or He*(2 S) beam axis. The energy resolution of the electron
analyzer was 50 meV, estimated from the full width at half-
+
2
maximum (fwhm) of the Ar ( P3/2) 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
2
1
some hydrocarbons with those of Gardner and Samson and
2
7
chemistry package implemented on Silicon Graphics Inc.
Challenge/XL and Origin200 workstations.
2
2
Kimura et al.
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 intensity of
metastable atoms were determined by inserting a stainless steel
plate into the reaction cell and measuring the intensity of
secondary electrons emitted. Thus, in these experiments, the
intensity of emitted electrons from sample molecules (Ie) or from
a reference stainless steel plate (IHe*) were measured as functions
of electron kinetic energy (Eek) and time (τ). Electron energies
were scanned by 10 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 2 and 3, and finally σ(Eek,νr) was converted to the 2D-
PIES, σ(Eek,Ec) using eq 4:
The ionization potentials for NCCN were calculated using
the outer valence Green’s function (OVGF) method28 as
incorporated in Gaussian-94 and also with the semiempirical
2
9
HAM/3 (hydrogenic atoms in molecules) method, which was
shown to give an accurate representation of IPs for molecules
3
0
containing first-row atoms. OVGF and HAM/3 calculations
were performed at the experimental geometry of NCCN.
IV. Results
3
Figure 1 shows the He I UPS and He*(2 S) PIES spectrum
of NCCN. The electron energy scale for the PIES spectrum is
shifted relative to that of UPS by the difference in the excitation
energies; 21.22 - 19.82 ) 1.40 eV. NCCN is obtained via the
rubeanic acid route (see below). The UPS spectrum shows the
volatile reaction products; bands of the sideproduct CO2 are
marked. The PIES spectrum is obtained by condensing the
reaction products with liquid nitrogen, pumping off CO2, and
then reevaporizing the pure NCCN.
σ(E ,V ) ) c[I (E ,νHe*)/IHe*(νHe*)](νHe*/νr)
(2)
(3)
(4)
Figure 2 shows a representation of the 2D-PIES spectrum of
NCCN. Curves (CERPIES spectra) are obtained from the 2D-
PIES spectrum by cutting small kinetic energy regions corre-
sponding to ca. 10 µs TOF of He*. NCCN is generated from
ICN for 2D-PIES investigation.
Figure 3 shows the log σ vs log Ec plots of CEDPICS for
NCCN (thick lines; between 100 and 300 meV collision energy).
The CEDPICS are obtained from the 2D-PIES spectrum by
cutting an appropriate range of electron kinetic energy, Ee
(typically the fwhm of the corresponding PIES band). The figure
also shows the calculated collision energy dependence of the
ionization cross sections (thin lines; between 80 and 500 meV
ek
r
e
ek
2
1/2
Vr ) [VHe* + 3kT/M]
2
r
E ) µV /2
c
where c, Vr, 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.

Penning Ionization of NCCN by Experiment and Theory
J. Phys. Chem. A, Vol. 103, No. 36, 1999 7173
Figure 4. Calculated interaction potential curves between NCCN and
2
Li(2 S). The origin of the coordinate system is the center of mass (X)
of the molecule and NCCN is lying on the x axis. The Li-X-N angle
is indicated for each curve.
Figure 2. Representation of the 2D-PIES spectrum of NCCN. Curves
(
CERPIES spectra) are obtained from the 2D-PIES spectrum by cutting
Figure 4 shows calculated potential energy curves between
a ground state Li atom and NCCN. The origin of the coordinate
system is the center of mass (X) of the molecule, and NCCN is
lying on the x axis. The Li-X-N angle is indicated for each
curve. Calculations are done at the CCSD/6-311++G** level
of theory.
small kinetic energy regions.
Figure 5 shows calculated total and partial ionization prob-
3
abilities for the He*(2 S)-NCCN chemiionization reaction as
a function of the impact parameter b (opacity functions) at
collision energy of 200 meV.
Figure 6 shows the calculated equilibrium geometry of
LiNCCN inorganic radicals, as well as total atomic charges from
natural population analysis using the QCISD density. Calcula-
tions were done at the QCISD/6-311++G** level of theory.
Bond lengths are given in angstroms.
Table 1 lists experimental and calculated ionization potentials
(
(
IPs), experimental peak energy shifts (∆E), slope parameters
m1 and m2) of log σ vs log Ec plots (see text), and the
assignment of the spectra. Vertical IPs are determined from the
He I UPS spectrum. The peak energy shifts in PIES spectrum
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. Table 2 lists the calculated harmonic vibrational
frequencies, infrared intensities, rotational constants, dipole
moments, and energies of LiNCCN radicals.
V. Discussion
PIES Investigations and the Interaction Potential. NCCN
is a high-symmetry linear molecule with a ground state
configuration of
.. (3σ ) (3σ ) (4σ ) (1π ) (4σ ) (5σ ) (1π ) ; Σg+
2
2
2
4
2
2
4
1
Figure 3. Collision energy dependence of total and partial ionization
cross sections for NCCN with He*(2 S). Thick lines shows experimental
.
3
g u g u u g g
data between 100 and 300 meV collision energy, and thin lines indicate
calculated values between 80 and 500 meV. Inset in figure shows the
calculated collision energy dependence of the ionization cross section
The observed UPS and PIES spectrum corresponds to four ionic
2
2
+
2
+
2
states with the sequence X Πg, A Σg , B Σu , and C Πu. On
2
+
8-11
for state B Σ
u
(number 3) on an expanded scale between 150 and 350
the basis of previous UPS investigations
works,
and theoretical
3
1-34
meV collision energy.
which discussed assignment and correlation effect
on ionization potentials in detail, the discussion of the assign-
ment of the spectra is not necessary (see Table 1). We discuss,
however, briefly the shape or localization of MOs, because the
information derived on the interaction potential depends on that.
Molecular orbital density plots are shown in Figure 3. We note
that the calculated pole strengths for ionizations are close to 1,
thus the molecular orbital model for ionization of NCCN in
the investigated IP region is in order. The first and fourth band
collision energy). The calculated electron density maps of the
molecular orbitals are also shown in the figure (the thick solid
curve in the maps indicates the molecular surface, estimated
from the van der Waals radii of atoms). Inset in Figure 3 shows
the calculated collision energy dependence of the ionization
2
+
cross section for state B Σu (number 3) on an expanded scale
between 150 and 350 meV collision energy.

7174 J. Phys. Chem. A, Vol. 103, No. 36, 1999
Pasinszki et al.
Figure 6. Calculated structure and total atomic charges of NCCNLi
radicals (distances in Å).
agreement with the earlier investigations. Comparing the UPS
and PIES spectra, the relative intensity of nN bands in the PIES
spectrum is strongly enhanced (note that π bands are doubly
degenerate). Since metastable atoms cannot penetrate into the
interior regions of the molecule, unlike photons, this shows that
the nN MOs are exposed much stronger outside to the molecular
surface than π orbitals. There are small peak energy shifts (∆Ei)
in PIES compared to UPS (see Table 1) if the two spectra are
compared to each other on an electron energy scale taking the
energy of the exciting metastable atom or photon into account.
Since the kinetic energy of the ejected electron in Penning
ionization depends on the difference between the incoming M
Figure 5. Calculated ionization probabilities as functions of the impact
parameter b (opacity functions) at the collision energy of 200 meV.
3
+
in the spectrum originate from the removal of an electron from
the 1πg and 1πu orbital, respectively, which can be represented
by the out-of-phase and in-phase combination of πCN orbitals.
The 1πg MO has a nodal surface perpendicular to the molecular
frame, going through the symmetry center of the molecule; thus,
the highest electron density of 1πg is on the CN fragments. 1πu
is delocalized, but it has the highest electron density in the
middle of the molecule, around the C-C bond. The second and
third bands in the spectrum originate from the 5σg and 4σu
orbitals, respectively, which are the in-phase and out-of-phase
combination of the two nitrogen lone pair (or nonbonding) nN
orbitals extending outside to the molecular surface along the σ
frame. The localization of 1πg, 1πu, and nN MOs, therefore,
provides a possibility to study the anisotropy of the interaction
potential around the molecule, using the collision energy
dependence of the partial ionization cross sections and peak
shifts in the PIES spectrum (see below). The PIES spectrum of
+ He*(2 S) (called interaction potential) and outgoing Mi +
He potential curves and the outgoing potential can be regarded
as flat in the ionization region, the peak shift is determined by
the incoming potential. 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. Negative peak shifts (-70 and -80 meV) were
observed on the nN bands, but no characteristic shift on the π
bands. This shows that the interaction between NCCN and He*-
3
(2 S) is attractive in the nitrogen lone pair region. The shift,
however, is surprisingly small compared to the 300-400 meV
peak shift of the nN bands in the spectra of alkyl nitriles,3 as
well as the lack of peak shift of π bands; for these latter bands
200-400 meV negative peak shifts were observed in spectra
of alkyl derivatives. This indicates that the interaction is much
,5
3
less attractive between NCCN and He*(2 S) than it is between
1
3
3
NCCN is shown in Figure 1. In the earlier investigation,
the -CN group and He*(2 S) in alkyl nitriles.
anomaly in the vibration fine structure of the first band has been
observed, which has been explained by a competing autoion-
ization process. We have not observed any unusual fine structure
and believe that the anomaly in the previous work is due to the
singlet He*(2 S) atoms in the applied He*(2 S,2 S) beam. The
observed relative state populations and peak shifts are in good
Simple theoretical models for describing the collision energy
dependence of the ionization cross sections for isotropic atomic
2
4,35,36
targets are well established.
The interaction in this case
can be simply described by an interaction potential curve. This
simple models, however, can be successfully applied for
anisotropic molecular systems as well, where a potential surface
1
1
3

Penning Ionization of NCCN by Experiment and Theory
J. Phys. Chem. A, Vol. 103, No. 36, 1999 7175
TABLE 1: Band Assignments, Ionization Potentials (IP/eV), Peak Energy Shifts (∆E/meV), and Slope Parameters (m) for
NCCN
IP/eV
HAM3
m ((0.03)b
ionic state
band
expt
OVGFa
orbital character
∆E/meV ((20 meV)
m
1
m
2
2
X Π
g
1
2
3
4
13.37
14.52
14.88
15.60
b
13.26
14.55
14.99
15.07
13.35 (0.91)
14.58 (0.89)
15.02 (0.87)
15.47 (0.90)
1π
g
u
g
(πCN-
)
+10
-70
-80
-10
-0.08
-0.27
-0.23
+0.04
0.36
0.25
0.31
0.25
2
+
A Σ
g
+
u
5σ
4σ
(nN+
)
)
2
B Σ
(nN-
2
C Π
u
1π
u
(πCN+
)
a
Pole strength in parentheses.
00 meV collision energy region, respectively.
1 2
m and m are the slope parameters of the linear fit to the experimental values in the 100-160 meV and 160-
3
TABLE 2: Calculated Vibrational Frequencies (cm-1), IR Intensities (km/mol), Rotational Constants (GHz), Dipole Moments
(
Debye), Total Energies (au), and Bonding Energies (kJ/mol) of NCCNLi Isomersa
NCCNLi (I) NCCNLi (II)
freq freq
NCCNLi (III)
freq
IR int
IR int
IR int
2
2
449 ν
238 ν
1
2
3
5
4
6
7
(CN str)
(CN str)
(C-C str)
(NCCN s def)
(Li-N str)
(NCCN as def)
(Li-NC def)
16.1
1.1
1.9
37.6
28.5
19.4
7.4
2341 ν
1818 ν
868 ν
610 ν
575 ν
426 ν
329 ν
1
2
3
4
5
8
6
9
7
(CN str)
(CN str)
(C-C str)
(Li-(CN) str)
(NCCN in-plane def)
(out-of-plane def)
(in-plane def)
(out-of-plane def)
(in-plane def)
16.8
10.1
2.5
150.5
9.4
24.5
26.5
3.3
2124 ν
1881 ν
692 ν
596 ν
531 ν
388 ν
275 ν
195 ν
83 ν
1
2
3
4
5
8
6
9
7
(CN s str)
(CN as str)
(C-C str)
(Li-(CCN) str)
(NCCN in-plane def)
(out-of-plane def)
(in-plane def)
(out-of-plane def.)
(in-plane def)
24.4
318.8
6.0
111.2
10.6
3.8
45.8
19.4
61.9
868 ν
530 ν
279 ν
255 ν
112 ν
1
1
78 ν
67 ν
29.3
b
b
A ) 37.1543
A ) 24.5547
Bb ) 2.4356
B ) 3.5498
C ) 3.2402
B ) 4.5008
C ) 3.8036
µc ) 4.81 (5.21)
µ ) 8.22 (8.58)
c
µ ) 5.92 (5.78)
c
total energy ) -192.661412
total energy ) -192.686634
total energy ) -192.675177
d
d
d
bonding energy ) 12.7
bonding energy ) 78.9
bonding energy ) 48.8
a
Calculated at the QCISD/6-311++G** level of theory. b Isotopes: C-12, N-14, Li-7. c Population analysis was done using the QCISD (SCF)
d
density. Difference between the total energy of the complex and the sum of the energies of fragments.
describes the interaction (in practice several cuts of the potential
surface, thus a series of potential curves are used), if there are
no irregularities in the simple potential curves describing the
gradually changing potential around the target. The negative
peak shift in PIES spectrum in this case reflects an average well
depth according to the localization of the corresponding MO.
According to this theory (see details in the original papers), the
collision energy dependence of the ionization cross section can
be described by eqs 5 and 6 if the long-range attractive part or
the repulsive part of the interaction potential governs the
collision energy dependence, respectively (where C, s, b, and d
are constants).
sections in the 100-160 meV collision energy region, and it is
positive between 160 and 300 meV. Thus, there is an unusual
change at 160 meV, which indicates an unusual interaction po-
tential surface. Indeed, as we can see in Figure 4, the interaction
potential curves at Li-X-N angle of 45°, 60°, and 75° show
changing interaction (from repulsive to attractive) due to a
second minimum on the potential surface. The calculated ca.
200 meV barrier to this minimum is in good qualitative agree-
ment with the experimentally observed unusual change at 160
meV. In the 100-160 meV collision energy region (and the
100-200 meV on the calculated M-Li curves), there is a linear
relationship between log σ and log Ec (slope parameters m1 are
shown in Table 1). The largest negative collision energy depen-
dence is observed on the nN bands (m1 ) -0.27 and -0.23), a
small negative collision energy dependence on the πg band (m1)
-0.08), and a small positive collision energy dependence on
the πu band (m1 ) +0.04). This indicates that the interaction is
attractive in the nitrogen lone pair region and gradually changes
into small repulsive as the Li-X-N angle opens up from 0° to
90°. This is in good agreement with the calculated M-Li
potentials in Figure 4 showing the same trend. The figure also
shows that there are trajectories which can lead to the second
minimum on the potential surface above 200 meV collision ener-
gy, and the number of these trajectories is gradually increasing
by increasing the collision energy, because the range of the Li-
X-N angle, where Li atoms can get to the second minimum,
gradually increases. Since more and more effective trajectory
can lead into the region where the ionization probability is higher
due to the closer molecular contact, this may explain why the
ionization cross section increases in the 160-300 meV region
by increasing the collision energy, despite attractive interaction.
For a comparison of the character of the interaction around
the parent pseudohalogen NCCN and alkyl-substituted deriva-
-
2/s
σ(E ) ∝ Ec
(5)
(6)
c
2
b/d-1/2
σ(E ) ∝ (ln E /C) (E /C)
c
c
c
When the minor collision energy dependence of the first factor
ln Ec/C) is neglected, there is a linear relationship between the
(
collision energy and the ionization cross section on a logarithmic
scale (log σ ) m log Ec + z, z is a constant). The slope parameter
m is negative if the interaction is attractive (see eq 5) and
positive if the interaction is repulsive (eq 6), and its absolute
value is larger if the interaction is stronger. This linear
relationship is widely observed on various molecular targets,
which gives reliability to the theory.
The 2D-PIES spectrum, the collision energy dependence of
the total and partial ionization cross sections, and the calculated
M-Li model potential curves are shown in Figures 2-4. From
Figure 3 it is clear that there is no linear relationship between
log σ and log Ec and the collision energy dependence of the
total and partial ionization cross section is quite unusual. There
is negative collision energy dependence of the ionization cross

7176 J. Phys. Chem. A, Vol. 103, No. 36, 1999
Pasinszki et al.
3
tives (R-CN) with He*(2 S), the collision energy dependence
in the 100-160 meV region can be used. The negative collision
energy dependence is much smaller in the case of NCCN
compared to that of R-CN (the slope parameter for the nN band
is between -0.4 and -0.6, and for the πCN band it is between
of most molecules, there is a linear relationship between log σ
and log Ec; thus, in practice, the experimental and calculated
slope parameters m can be fitted. We have to note, however,
that there is an important difference between reproducing the
total ionization cross section and reproducing the slope param-
eter. There is correlation between the interaction potential and
K, and both determine the total ionization cross section, as well
as the collision energy dependence. The total ionization cross
section strongly depends on K, and the calculated value
gradually increases by increasing K; therefore, the experimental
value can be reproduced, even if the applied interaction potential
is not perfect. The slope parameter, however, strongly depends
on the interaction potential, and on K in a smaller extent. If the
calculated interaction potential is not attractive enough (this is
the usual case; see below), the experimental slope parameter
cannot be reproduced, only a best agreement between the
experimental and calculated value is expected.
3,5
-
0.25 and -0.45), indicating a much weaker attractive
interaction, which, in good agreement with the peak shifts above,
indicates that the electron density on the -CN group in NCCN
is smaller than that in R-CN. Regarding the well-known
electron-donating effect (+I and +M) of alkyl groups and the
large electron negativity of the cyano group, it is not surprising,
and it represents a good example of substituent effects.
Classical Trajectory Calculations for Penning Ionization
Cross Sections. To eliminate the unusual behavior of the
observed collision energy dependence of the ionization cross
sections on a quantitative basis and to test the performance of
our new method, we have performed theoretical calculations.
The theoretical description of the Penning ionization can be
divided into two steps: the first is the calculation of the potential
energy of the entrance He*-M resonance state, and the second
is the calculation of the collision dynamics on the basis of this
complex potential. The potential energy of the resonance state
can be described in terms of a local complex potential of V )
V0 - (i/2)Γ, where the real part V0 is the He*-M interaction
potential and the imaginary part Γ is the ionization width for
the decay of the He*-M resonance state. In our novel method,
ab initio methods are used to calculate V0 and Γ and classical
trajectory calculations to calculate the collision dynamics using
the following approximations.
The calculations for NCCN in this work are done by varying
K constant on an iterative way to obtain the best agreement
between the experimental and calculated collision energy
dependence of the total ionization cross section. 2000 trajectories
are calculated for each collision energy and the initial parameters
for trajectories, such as the orientation of the molecule (mo-
lecular and rotational axis), impact parameter (between 0 and
1
0 Å), and rotational energy (300 K) are randomly generated.
The results are shown in Figure 3, and calculated total ionization
cross sections are found in note 37. There is a good qualitative
agreement between the experimental and calculated collision
energy dependence. The calculations show that the ionization
of nN MOs have the largest negative collision energy depen-
dence, the dependence of 1πg is close to zero, and it is positive
for 1πu (see note 38). This trend is in good agreement with the
experiment, but the calculated values (the slope parameters in
the 100-200 meV region) are regularly less negative or more
positive, indicating a less attractive or more repulsive interaction.
The unusual collision energy dependence of the partial ionization
cross sections, i.e., having a minimum in the 100-300 meV
region, is also reproduced on the ionization of 1πg and nN MOs
(
1) Due to the similarity of the He*-M and Li-M interaction
potentials, the real part of the complex potential V0 is ap-
proximated by the computationally more feasible Li-M po-
Li
tential V0 .
(
2) Neglecting the angular distribution of the ionized elec-
trons, replacing the distance between electrons in two-electron
integrals with an average length, and taking into account that
the continuum orbitals and helium 2s orbitals are too diffuse
compared to the helium 1s and ionized orbitals, thus the
positional dependence of the ionization width is governed by
the more compact 1s and ionized orbitals, the ionization width
for each ionic state is given by
(see inset in Figure 3), but the absolute value of the change is
smaller than that of the experimental value. The minimum, not
surprisingly, was calculated to be at 200 meV, which is the
barrier to the second minimum on the calculated M-Li
interaction potential surface. This difference between the
experimental and calculated values, as well as the calculated
collision energy dependence of the ionization cross section,
indicates that the calculated M-Li interaction potential, although
able to give a good description of the interaction on a qualitative
basis, is less attractive (or more repulsive) than it is expected
from the experiment. We note that in the case of all molecules
that we are currently studying in our laboratory the same effect
is observed on the calculated potentials. Possible explanations
might be the insufficient inclusion of electron correlation effects
in calculations or small geometry relaxation of the molecule
by the approach of the metastable atom. This latter is neglected
and the frozen molecule approximation has worked very well
in qualitative analysis so far. Definitely the extent of this is not
clear at the moment and requires further theoretical and
experimental work. Clearly, the relaxation effects becomes more
important as the speed of the metastable atom decreases, and
neglecting them might cause the calculated collision energy
dependence to be more positive than the real value. In our
previous work on N2, the calculated M-Li potential has been
empirically pulled down, but the potential of NCCN seems to
be too difficult to apply a simple correction. For NCCN, we
(i)
2
Γ ) K|〈Φ |Ψ 〉|
(7)
i
1s
where K is a constant and Φi and Ψ1s are the ionized molecular
orbital and helium atom 1s orbital, respectively.
(
3) The molecules are treated as rigid rotators in the classical
trajectory calculations.
The theory and approximations are thus well established, and
the only limitation of using pure theory for the calculations is
that the K constant in eq 7 is not known and it may be different
for each colliding molecular system. The determination of K
requires the use of at least one experimentally determined
physical parameter. In our previous work on N2, K constant
was determined in an iterative way to reproduce the experi-
mental total ionization cross section at 200 meV collision
energy. The total ionization cross section, however, is not known
for many molecules, and this is the case for NCCN too. In this
work, we considered another way to determine K. Although
the total ionization cross section is not determined, the
dependence of the total ionization cross section on the collision
energy is always measured in 2D-PIES and the calculated value
depends on K. This dependence, or more precisely the reproduc-
ibility of this dependence by varying K on an iterative way,
provides a possibility for determining K constant. In the case

Penning Ionization of NCCN by Experiment and Theory
J. Phys. Chem. A, Vol. 103, No. 36, 1999 7177
expect that the description of the interaction potential in the
region of the second minimum is more complicated, where
relaxation effects and electron correlation effects are expected
to be more important (see the geometry of NCCNLi complexes
below). The peak shift of -80 meV on nN bands is in good
agreement with the calculated well of ca. 105 meV on the
M-Li interaction potential at Li-X-N angle of 0°, but the
barrier to the second minimum and the repulsive character
especially in the 1πu region seems to be exaggerated by
calculations.
that this latter is the ground state potential of the M-Li system,
PIES spectroscopy is able to predict the existence of thermo-
dynamically stable MLi radicals. Their existence is expected if
the interaction is attractive, and there is a large negative peak
shift in the PIES spectrum. The peak shift is a good estimation
of the bonding energy of the radical if there is no large geometry
relaxation of M by the formation of MLi (due to the fast
metastable atom the geometry of M is frozen) and the corre-
sponding MO is localized. PIES investigation and model
potentials (see above) indicated that at least two different
NCCNLi complexes can form from NCCN and Li. To obtain
information about the structure and stability of these complexes,
full geometry optimizations have been performed at the QCISD/
Figure 5 shows opacity functions for the total and partial
ionization cross sections at the collision energy of 200 meV.
As it can be seen in the figure, the ionization probability is
high if the impact parameter b is smaller than 2 Å, it is gradually
decreasing with increasing b, and there is no effective trajectory,
which can lead to ionization, if b larger than 6 Å. The opacity
functions of the partial ionization cross sections reflect the
localization of MOs. Trajectories with small b can effectively
lead to the ionization of 1πu MO, showing that the highest
electron density of this MO is at the central part of the molecule,
and for the ionization at b ) 5-6 Å, the nitrogen lone pairs
are responsible, which are extended at the two ends of the
molecular frame. The plateau behavior of the nN opacity
functions is the result of the localization of nN MOs at the two
ends of the NCCN frame, and vice versa they show that the
electron density of these orbitals in the middle of the molecular
frame is negligible. Plateau behavior is not observed for the
6
-311++G** level. Calculations have found three minima on
the potential energy surface, and they are shown in Figure 6.
The first complex, NCCNLi(I), has a linear structure, where
the NCCN ligand bonds with the nitrogen lone pair to Li with
a coordinative bond. The total atomic charge on Li is close to
zero (+0.02), and the molecular orbital analysis shows that the
lithium unpaired electron stays on Li. The geometry of NCCN
does not change significantly compared to the free NCCN
molecule. The calculated bonding energy of this complex is 12.7
kJ/mol () 132 meV), which is in good agreement with the 80
meV negative peak shift of the nN band in the PIES spectrum.
This complex represents the case where the peak shifts in PIES
is a good estimation of the stability of the corresponding MLi
radical. (Note that in the case of molecular target the well on
the potential surface must be somewhat larger than the peak
shift depending on the localization of the MO.) The second and
third complexes have a different structure, and they can be
characterized as π-complexes. The lithium unpaired electron
completely delocalizes on the NCCN frame, and as a conse-
quence the charge on Li is close to +1 (see Figure 6). The
bonding energy of these complexes is much larger than that of
NCCNLi(I). By investigating the formation of MOs of NCCNLi
complexes from their NCCN and Li fragments, the same
stabilizing effect has been observed that has been found in the
1πu opacity function, and the plateau behavior of 1πg functions
shows that the 1πg MO has highest electron density on the end
atoms of NCCN. The analysis of individual trajectories indicates
that the upper and lower boundaries of opacity functions are
related to some particular trajectories. The upper boundaries of
the two nN opacity functions, corresponding to 5σg and 4σu MOs,
are determined by trajectories leading to the nitrogen lone
electron pair region. These trajectories also set the lower
boundary (zero) of 1πu and 1πg functions. Since the two nN
MO is very similar and has the highest electron density in the
same molecular region, their opacity functions are very similar
7
case of HNNNLi and HCNOLi in our previous work. As a
(
MOs are competing in the ionization), and interestingly the
result of the bending of the NCCN frame, the LUMO of NCCN
is stabilized and thus interacts strongly with the LUMO of Li
atom, which results in further large stabilization. The delocal-
ization of the Li unpaired electron into this stabilized lower lying
MO results in large bonding energy in complexes NCCNLi(II)
and NCCNLi(III). This cannot happen in NCCNLi(I); thus the
bonding energy is much smaller and the lithium unpaired
electron stays on the destabilized Li 2s orbital. The bonding
energy of 78.9 kJ/mol (818 meV) and 48.8 kJ/mol (506 meV)
for NCCNLi(II) and NCCNLi(III), respectively, also indicates
that large geometry relaxation in the Penning ionization cannot
happen, because this should cause very large peak shifts on π
bands. Since practically no peak shift is observed on these bands,
this gives reliability for the frozen molecule approximation in
the Penning ionization process in qualitative analysis (see
Discussion above).
sum of the two ionization probabilities in the upper boundary
gives the upper boundary of the opacity function of the total
ionization cross section. The upper boundary of 1πu opacity
function is determined by trajectories leading to the middle of
the NCCN frame, and trajectories perpendicular to the frame
have the largest ionization probability. These trajectories set
the lower boundary of nN and 1πg opacity functions as well. It
is interesting to find that the trajectories determining the upper
boundary of the 1πu opacity function determine the lower
boundary of that of the total ionization cross section, and the
trajectories determining the upper boundary of nN opacity
functions set the upper boundary of that of the total. Since the
localization of 1πu and nN MOs is very different, NCCN is an
ideal case where this clear difference between opacity functions
and collisional dynamics can be observed. The opacity function
of 1πg MO is just between the 1πu and nN functions in many
respects, in good agreement with the orbital’s intermediate
localization, having the highest electron density around an
N-X-He* angle of 45°. The opacity functions clearly show
that there is a high (close to 1) ionization probability for nN
MOs if He* approaches the molecule at a small N-X-He*
angle, and this gradually decreases by increasing the N-X-
He* angle. The situation for 1πu is just the opposite.
The formation of π-complexes and their large bonding energy
is an interesting result of this work. Our previous study on CH3-
CNLi and Lin(CH3CN)m clusters has indicated that the most
stable form of these molecules or clusters is that where the CH3-
CN bonds to Li with the nitrogen lone pair (similarly to
3,39
NCCNLi(I)). The strong tendency of NCCN to form π-com-
plexes with Li may cause interesting complex or cluster
behaviors. The identification of NCCNLi isomers is feasible
either in inert solid matrix or in the dilute gas phase. To support
future identifications, calculated harmonic vibrational frequen-
NCCNLi Inorganic Radicals. Due to the similarity of
3
2
M-He*(2 S) and M-Li(2 S) interaction potentials, and given

7178 J. Phys. Chem. A, Vol. 103, No. 36, 1999
Pasinszki et al.
cies, infrared intensities, rotational constants, and dipole mo-
ments are shown in Table 2.
(6) Ogawa, T.; Ohno, K. J. Chem. Phys. 1999, 110, 3773.
(7) Pasinszki, T; Kishimoto, N.; Ohno, K. Submitted for publication.
(8) Baker, C.; Turner, D. W. Proc. R. Soc. A 1968, 308, 19.
(9) Hollas, J. M.; Sutherley, T. A. Mol. Phys. 1972, 24, 1123.
Conclusion
(
10) Åsbrink, L.; von Niessen, W.; Bieri, G. J. Electron Spectrosc. Relat.
Phenom. 1980, 21, 93.
11) Fridh, C.; Åsbrink, L.; Lindholm, E. Chem. Phys. 1978, 27, 169.
12) Cˇ erm a´ k, V.; Yencha, A. J. J. Electron Spectrosc. Relat. Phenom.
1976, 8, 109.
13) Yee, D. S. C.; Brion, C. E. J. Electron Spectrosc. Relat. Phenom.
976, 8, 313.
14) Pasinszki, T.; Westwood, N. P. C. J. Phys. Chem. 1995, 99, 1649.
(15) Gmelins Handbuch der anorganischen Chemie; Verlag Chemie
GMBH: Weinheim, 1971; Kohlenstoff, Teil D 1, p 85.
16) Gay-Lussac, L. J. Ann. Chim. (Paris) 1815, 95, 175.
17) Lewis, G. N.; Keyes, D. B. J. Am. Chem. Soc. 1918, 40, 472.
18) Takami, T.; Mitsuke, K.; Ohno, K. J. Chem. Phys. 1991, 95, 918.
(19) Takami, T.; Ohno, K. J. Chem. Phys. 1992, 96, 6523.
20) Mitsuke, K.; Takami, T.; Ohno, K. J. Chem. Phys. 1989, 91, 1618.
21) Gardner, J. L.; Samson, J. A. R. J. Electron Spectrosc. Relat.
The NCCN molecule can be generated into the gas phase in
good yield from solid-solid reaction between rubeanic acid and
yellow mercury(II) oxide and by gas-phase thermolysis of
mercury(II) cyanide or cyanogen iodide. The molecule is
characterized in the gas phase by Penning and He I photoelec-
tron spectroscopies. From spectroscopic investigations, the
(
(
(
1
(
3
interaction potential between NCCN and He*(2 S) atoms is
(
(
(
deduced, which is attractive in the nitrogen lone electron pair
region and gradually changes into weakly repulsive by going
from the end of the NCCN frame into the middle. Model
(
(
potential calculations on the similarly interacting NCCN-Li-
2
(
2 S) system are in good qualitative agreement with experi-
Phenom. 1976, 8, 469.
mental results. Furthermore, they provide explanation for the
observed unusual collision energy dependence of the ionization
cross sections. Classical trajectory calculations have also been
performed for the understanding of collision dynamics. They
are in qualitative agreement with the experiment, but the
quantitative description of the collision energy dependence of
the ionization cross sections is fair. The quantitative description
of this latter is a difficult but an interesting challenge for future
theoretical calculations.
NCCN readily forms complexes with lithium atom, and ab
initio calculations identified three isomers: one classical-type
complex, where NCCN bonds to Li by one of the nitrogen lone
electron pairs, and two π-complexes. These π-complexes are
thermodynamically more stable than the classical one, and the
large bonding energy is explained by the delocalization of the
lithium unpaired electron on the NCCN frame.
(22) 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.
(
23) (a) Rothe, E. W.; Neynaber, R. H.; Trajillo, S. M. J. Chem. Phys.
1
965, 42, 3310. (b) Hotop, H. Radiat. Res. 1974, 59, 379. (c) Haberland,
H.; Lee, Y. T.; Siska, P. E. AdV. Chem. Phys. 1981, 45, 487.
(24) Niehaus, A. AdV. Chem. Phys. 1981, 45, 399.
(25) Maki, A. G. J. Chem. Phys. 1965, 43, 3193.
(26) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 10, 553.
(27) 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.
(28) von Niessen, W.; Schirmer, J.; Cederbaum, L. S. Comput. Phys.
Rep. 1984, 1, 57.
(29) Åsbrink, L.; Fridh, C.; Lindholm, E. Chem. Phys. Lett. 1977, 52,
6
9. The HAM/3 program is available from the Quantum Chemistry Program
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 (ID No. L98519) and the Hungarian Scientific
Research Fund (OTKA Grant F022031) in support of this work.
Exchange, Indiana University, Bloomington, IN (D. P. Chong, QCMP005,
1985).
(
(
30) Chong, D. P. Theor. Chim. Acta. 1979, 51, 55.
31) von Niessen, W.; Cederbaum. L. S.; Schirmer, J.; Diercksen, G.
H. F.; Kraemer, W. P. J. Electron Spectrosc. Relat. Phenom. 1982, 28, 45.
32) Scheller, M. K.; Weikert, H. G.; Cederbaum, L. S.; Tarantelli, F.
J. Electron Spectrosc. Relat. Phenom. 1990, 51, 75.
33) Cederbaum, L. S.; Domcke, W.; Schirmer, J.; K o¨ ppel, H. J. Chem.
Phys. 1980, 72, 1348.
34) Decleva, P.; Lisini, A. Chem. Phys. 1987, 112, 339.
(
(
References and Notes
(
(
1) Ohno, K.; Harada, Y. Penning IonizationsThe Outher Shape of
(35) Illenberger, E.; Neihaus, A. Z. Phys. B 1975, 20, 33.
(36) Allison, W.; Muschlitz, E. E., Jr. J. Electron Spectrosc. Relat.
Phenom. 1981, 23, 339.
Molecules. In Molecular Spectroscopy, Electronic Structure and Intramo-
lecular Interactions; Maksic, Z. B., Ed.; Springer-Verlag: New York, 1991.
(
2) Ohno, K.; Yamakado, H.; Ogawa, T.; Yamata, T. J. Chem. Phys.
996, 105, 7536.
3) Pasinszki, T; Yamakado, H.; Ohno, K. J. Phys. Chem. 1995, 99,
4678.
4) Kishimoto, N.; Yokoi, R.; Yamakado, H.; Ohno, K. J. Phys. Chem.
A 1997, 101, 3284.
(37) Calculated total Penning ionization cross sections for NCCN at 100,
200, 300, 400, and 500 meV collision energy, respectively, are 53.94, 54.93,
56.71, 57.43, and 57.96 Å .
(38) Calculated slope parameters (fitted to the calculated partial ioniza-
tion cross sections between 100 and 200 meV; m1) for the ionization of
1πg, 5σg, 4σu, and 1πu MOs, respectively: +0.01, -0.14, -0.16, and +0.34.
(39) Ohshimo, K.; Tsunoyama, H.; Yamakita, Y.; Misaizu, F.; Ohno,
K. Chem. Phys. Lett. 1999, 301, 356.
1
2
(
1
(
(5) Kishimoto, N.; Aizawa, J.; Yamakado, H.; Ohno, K. J. Phys. Chem.
A 1997, 101, 5038.