Fluorescence excitation spectrum of Ni-Ar complex

Article abstract of DOI:10.1016/S0009-2614(98)01178-6

A new absorption band is observed by a photolysis-supersonic jet experiment of nickelocene, Ni(C₅H₅)₂, at wavelengths near the Ni(y ¹D₂-a ¹D₂) transition. This band appears only in the coexpansion of nickelocene with Ar or Ar in He. Taking into account dependencies of the band intensity on the photolysis laser power and nozzle pressure, this band is attributed to the transition of a Ni-Ar van der Waals complex. Since dispersed fluorescence of this band consists entirely of atomic transitions of Ni, the excited state of the Ni-Ar complex is considered to predissociate into Ni^*+Ar with a very short lifetime. Analysis of the vibrational structure gives the excited-state binding energy D₀′=51.7 cm^-1.

Full text of DOI:10.1016/S0009-2614(98)01178-6

11 December 1998
Ž
.
Chemical Physics Letters 298 1998 227–232
Fluorescence excitation spectrum of Ni–Ar complex
)
Yasunobu Kawamoto, Kenji Honma
Department of Material Science, Himeji Institute of Technology, 1475-2 Kanaji, Kamigori, Hyogo 678-1297 Japan
Received 27 August 1998; in final form 28 September 1998
Abstract
A new absorption band is observed by a photolysis–supersonic jet experiment of nickelocene, Ni C5H5 2, at
Ž
.
1
1
Ž
.
wavelengths near the Ni y D2–a D2 transition. This band appears only in the coexpansion of nickelocene with Ar or Ar in
He. Taking into account dependencies of the band intensity on the photolysis laser power and nozzle pressure, this band is
attributed to the transition of a Ni–Ar van der Waals complex. Since dispersed fluorescence of this band consists entirely of
atomic transitions of Ni, the excited state of the Ni–Ar complex is considered to predissociate into Ni)qAr with a very
short lifetime. Analysis of the vibrational structure gives the excited-state binding energy DX0s51.7 cmy1. q 1998 Elsevier
Science B.V. All rights reserved.
w x
6 . This powerful approach has been applied to
1. Introduction
metals in groups IA, IIA, IIB, and IIIA. The list of
these kinds of studies has been extended to the noble
We have studied the interactions of the first-row
transition metal atoms with simple molecules and
various important information has been accumulated
w x
w
x
w x
metals, Cu 7 , Ag 8,9 , and Au 10 . In the com-
plexes of these metals with rare gases, two band
w
x
1–5 . However, the kinetic approach we have taken
systems, P1r2,3r2– 2 S1r2 , have been observed cor-
2
provides more or less qualitative information with
respect to the interaction potentials, and dynamical
or spectroscopic approaches are necessary to obtain
detailed information about interaction potentials such
as well depth, barrier height, occurrence of surface
crossing, and so on. One alternative and useful tech-
nique to obtain such information is a spectroscopic
study of low-temperature van der Waals complexes.
Laser-induced fluorescence and resonance-en-
hanced two-photon ionization techniques have been
applied to metal atom–rare gas atom van der Waals
relating to the P1r2,3r2 – 2S atomic asymptotes, and an
2
additional band system correlating to the D5r2– 2S
2
asymptote has also been observed for Ag–Ar and
Ag–Kr. Although these metals have d electrons, the
electronic states of these atoms are rather simple
because of the filled nd orbitals in the ground states.
Compared with the copper series, the other transi-
tion metals have much more complicated electronic
structures, and there have been quite a few experi-
mental studies for neutral van der Waals complexes
w
x
containing transition metal atoms 11,12 . Recently,
ab initio calculations have been carried out by Burda
et al. for the interaction potentials of M–Xe, where
Ž
.
complexes M–Rg and have provided much of the
information about the interatomic potentials of M–Rg
w
x
MsNi, Pd, and Pt 13 . In this Letter, we present a
)
spectroscopic study of the van der Waals complex,
Corresponding author. E-mail: honma@sci.himeji-tech.ac.jp
0009-2614r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 98 01178-6

(
)
228
Y. Kawamoto, K. HonmarChemical Physics Letters 298 1998 227–232
Ni–Ar, which is the first experimental study of a
complex between Ni and a rare gas.
2. Experimental
The experiment was carried out by using a super-
sonic jet–LIF apparatus which has been described
w
x
previously 14 . Nickel atoms were generated by
multi-photon dissociation of nickelocene at 266 nm.
The nickelocene was issued through a pulsed nozzle
Ž
.
General Valve, 0.8 mm diam. heated to 808C. He,
Ar, or their mixture was used as a carrier gas for the
nozzle expansion. The fourth harmonic of a YAG
Ž
.
laser Continuum NY-82 was used to irradiate the
nickelocene jet at 1y2 mm downstream from the
nozzle orifice to photolyze the molecules. The typi-
cal power of the 266 nm laser was 15 mJrpulse. An
fs1000 mm lens was used but its focal point was
about 100 mm away from the nickelocene jet. The
Ž
.
output of a dye laser Continuum ND-6000 pumped
by another YAG laser was frequency doubled by a
KD) P crystal to obtain radiation in the 300–310 nm
Fig. 1. Laser-induced fluorescence excitation spectrum obtained
by a photolysis–jet expansion condition of nickelocene mixture in
Ž
.
region Rhodamine 610 . This tunable light was
irradiated at 10–20 mm downstream from the pho-
tolysis laser. Laser-induced fluorescence was col-
lected by a lens system and detected by a photo-
Ž .
a
Ž .
b
Ar.
Intensity of undispersed fluorescence was monitored.
Ž3
.
Intensity of the emission at 310.25 nm, Ni F3 –a3D2 , was
monitored by a monochromator.
Ž
.
multiplier tube Hamamatsu R-928 . Dispersed fluo-
rescence and excitation spectra were measured by
fluorescence consists of atomic transitions, as we
will discuss in Section 3.2, some atomic peaks are
not eliminated completely. However, the new band is
isolated and its structure is clearly seen in this
spectrum.
Ž
.
using a monochromator Nikon P-250 . Nickelocene
95.0% was obtained from Kanto Chem., and used
without further purification. Helium 99.999% and
Ž
.
Ž
.
Ž
.
argon 99.999% were obtained from Nihon Sanso.
The LIF intensity of this band was measured as a
function of the photolysis laser power. The intensity
of the new band shows a quadratic dependence
3. Results and discussion
Ž3
which is identical to the LIF intensity of the Ni F4–
3
3.1. Laser-induced fluorescence
.
a F4 transition. The quadratic dependence is consis-
tent with the two-photon dissociation of nickelocene
Fig. 1a shows the laser-induced fluorescence
spectrum obtained by using Ar as a carrier gas.
There are intense lines originating from the atomic
transitions of Ni. Other than these atomic transitions,
1
Ž
Ž ..
to form the free nickel atom Eq. 1 .
Ni C H q2hÕ™Niq2C5H5 .
1
Ž .
Ž
.
5
5
2
That is, the new band originates from species con-
taining single nickel atoms formed by the two-pho-
ton dissociation of nickelocene. Additional informa-
tion was obtained from the measurement of the LIF
1
1
Ž
.
one band can be seen near the Ni y D2–a D2
y1
Ž
.
transition, 301.29 nm 33190.9 cm . In order to
eliminate the contribution of atomic transitions, the
excitation spectrum was measured by monitoring the
emission at 310.25 nm which is the most intense
peak in the dispersed fluorescence. The spectrum
obtained is shown in Fig. 1b. Because the dispersed
1 The mean bond dissociation energy of Ni–C5H5 is 2.86"0.30
w
x
eV, given by Ref. 15 . Two 266 nm photons are required to
dissociate two Ni–C5H5 bonds.

(
)
Y. Kawamoto, K. HonmarChemical Physics Letters 298 1998 227–232
229
intensity as a function of the nozzle pressure. The
result shows that the new band appears at higher
nozzle pressure than the ground-state Ni atom. This
dependence suggests that the new band is attributed
to a van der Waals complex. Since the photolysis
occurs at 1–2 mm from the nozzle and the effect of
the nozzle pressure would be more pronounced after
the photolysis, the complex is more likely to be
formed after the photolysis, i.e. the new band origi-
nates from a complex of the Ni atom. Taking into
account the fact that only Ar or a mixture of Ar in
He gives this band, it can be concluded that the new
Ž
.
band is ascribed to a Ni–Arn n01 complex. Al-
though the number of Ar atoms in the complex is
still an open question, the vibrational structure we
will discuss below is consistent with a diatomic
complex. Thus, we conclude that the new band is
ascribed to an electronic transition of the Ni–Ar
complex.
Fig. 2. Dispersed fluorescence spectrum obtained by excitation at
301.59 nm. The simulated spectrum is represented by a bar graph.
3.2. Dispersed fluorescence and dynamics of excited
state
shows an identical time profile to the corresponding
emission following atomic excitation.
The predissociative nature of the excited state
prevents us from obtaining potential information for
the ground-state complex. However, another impor-
tant piece of information, the product-state distribu-
tion, can be obtained by analyzing the intensities of
the atomic peaks. By using the atomic transition
Dispersed fluorescence is expected to provide
useful information about excited-state dynamics and
the vibrational structure of the ground state. The
dispersed fluorescence spectrum shown in Fig. 2 was
measured by excitation at 301.59 nm indicated by an
arrow in Fig. 1a. The excitation at other peaks and in
the broad part of the band provides an almost identi-
cal spectrum. The spectrum does not have any of the
vibrational structure expected for the emission corre-
sponding to a bound–bound transition of the com-
plex. Instead of the complex transition, all the peaks
in this spectrum can be assigned to atomic transitions
of Ni. Most of the intense peaks are assigned to
w
x
probabilities tabulated by Fuhr et al. 16 , the peak
heights were reproduced as shown by the bar spec-
trum in Fig. 2. Relative yields of Ni) determined by
this analysis are summarized in Table 1 with elec-
w
x
tronic energies 17 . Although there is no intense
transitions from y 3DJ in the wavelength region
shown in Fig. 2, careful analysis of shorter wave-
length region provides the yields of y 3D2,3 listed in
Table 1.
Ž3
.
transitions from Ni F3 and there are some other
Ž3
.
transitions from Ni F2,4, z 1P1, z 1D2 as indicated in
Fig. 2. The absence of emission from the complex
suggests that the excited state of Ni–Ar predissoci-
ates to Ni) qAr with a very short lifetime. It may
be reasonable to assume the fluorescence lifetime of
the complex to be of the same order as those of the
One possible mechanism for the predissociation is
the spin–orbit predissociation via a repulsive sur-
face. Funk and Breckenridge have observed the effi-
Ž3
.
cient formation of Cd P2 qXe by the excitation of
Ž1
.
w x
the Cd–Xe complex correlating to Cd P1 qXe 18 .
They have attributed the efficient predissociation to
the surface crossing of the initially excited Cd–
Xe C P1 state with the repulsive c S1 state, both
having the same Vs1. Although the more compli-
cated electronic structure of the Ni atom makes a
detailed analysis like that of Cd–Xe difficult, a
w
x
Ni atom, typically 10 ns 16 . If a detection limit of
the complex fluorescence is assumed to be 1% of the
atomic peaks, an upper limit of the lifetime is esti-
mated to 0.1 ns. This limit is reasonable because the
atomic emission following the complex excitation
1
3
Ž
.

(
)
230
Y. Kawamoto, K. HonmarChemical Physics Letters 298 1998 227–232
Table 1
isotope shifts, one reasonable assignment could be
made by studying the intensity change of vibrational
peaks. The intensity changes smoothly toward lower
energy and a sharp drop without any peak can be
seen after the lowest energy peak. Therefore, we
assigned the lowest energy peak as ÕX s0 of the
excited state. The vibrational quantum number deter-
mined by this manner is shown in Fig. 1b and the
frequencies of vibrational peaks are summarized in
Table 2. The BS plot of this series is shown in Fig.
3a, which shows a linear relation within the experi-
mental error. The ve and ve xe values were deter-
mined to be 12.4 and 0.73 cmy1, respectively. A
linear extrapolation of this plot provides the dissocia-
tion limit and the dissociation energy of the excited
state to be 33193 cmy1 and D0 s46.5 cmy1, re-
spectively.
Relative population of excited states formed by predissociation of
the Ni–Ar complex
a
Term
Energyrcmy1
Relative populationb
yc
0.03
0.03
0.03
1.00
0.15
0.06
0.09
y3D1
y3D2
y3D3
3F2
34408.57
34163.29
33500.85
33610.92
33112.37
32973.41
32982.28
31441.67
3F3
3F4
z1P1
z1D2
a
w
x
All data are taken from Ref. 13 .
b
Ž3
.
.
All populations are given as the relative values to that of Ni F3
c This level was not observed in the dispersed fluorescence.
similar mechanism is possible in the Ni–Ar com-
plex. Some of the states developed from Ni)
Ž3
.
F2,3,4, z 1P1, z 1D2 qAr could be repulsive and
In the LB analysis, the level spacing is described
by a power law, DGÕ n AD0 yEv , where DGÕ s
cross with the upper state initially excited. Because
the excitation of the ÕX s0 vibronic level exclusively
forms the excited states of Ni, the crossing must
occur near the bottom of the upper-state potential
surface.
Ž
.
w Ž
.
Ž
.x
E Õq1 yE Õy1 r2, D0 is the dissociation en-
ergy and Ev is the total vibrational energy of the
state. For a van der Waals interaction, 3r2 is used as
the exponent. The LB plot for Ni–Ar is shown in
Fig. 3b, and the dissociation limit determined by this
plot is 33198 cmy1. The binding energy can be
obtained as the difference between this dissociation
limit and the lowest energy vibrational level ob-
served, D0 s51.7 cmy1. The dissociation limit ob-
tained from the LB analysis is slightly larger than
that from the BS analysis. The LB analysis is appro-
priate if levels at the convergence limit are mea-
sured, whereas the BS one may not be adequate for
weakly interacting systems. Here, we observed the
level up to 7.5 cmy1 lower than the dissociation
limit. Thus, we take the values determined from the
3.3. Potential parameter of the excited state and
atomic asymptote of the complex
Ž
.
The excitation spectrum Fig. 1b shows vibra-
tional structure in which the spacing decreases at
shorter wavelength and converges to a continuum.
The binding energy is obtained by an extrapolation
of the vibronic levels to the dissociation limit There
are commonly used methods for the extrapolation of
vibronic levels to determine dissociation energies,
Ž
.
w x
the Birge–Sponer BS analysis 19 and the
Ž
.
w
x
LeRoy–Bernstein LB one 20,21 . A BS plot re-
quires knowledge of the vibrational quantum num-
bers. In diatomic molecules, or complexes having
isotopomers, the isotope shift of each peak provides
a rigorous assignment of the vibrational quantum
numbers. However, the Ni–Ar complex observed
here shows line widths which are broader than the
Table 2
Assignments and transition frequencies of Ni–Ar
ÕX – ÕY
nobs rcmy1
0–0
1–0
2–0
3–0
4–0
5–0
6–0
33146.3
33157.6
33166.7
33174.7
33181.4
33186.2
33190.5
2
frequency shift for isotopmers. Without the help of
2
Ž
.
Ž
The isotope shift of a given Õ is given by Dn s 1y r ve Õ
q1r2 y 1y r ve xe Õq1r2 2. For the ve and ve xe values
2
.
Ž
.
Ž
.
58
60
determined below, the shift between Ni40 Ar and Ni40Ar has
a maximum value, 0.37 cmy1, at Õs4.

(
)
Y. Kawamoto, K. HonmarChemical Physics Letters 298 1998 227–232
231
by analyzing the dispersed fluorescence. The highest
energy level observed in the dispersed fluorescence
3
Ž
.
is Ni y D2 for the excitation of the discrete peaks.
This provides the lower limit for the ÕX s0 level of
the upper state. Since a reasonable upper limit can be
1
Ž
.
selected to be Ni y D2 , which is the highest level
1
Ž
.
accessible from Ni a D2 in the wavelength region
studied here, there are three possibilities for the
atomic asymptote of the upper state, i.e. y 1D2 , y 1F3,
and y 3D1. Although no direct evidence is available
to determine which of three it is, it may be useful to
evaluate the binding energy of the lower state. By
using the dissociation limit of the upper state, 33198
cmy1, the binding energy of the lower state can be
Y
w
x
calculated by D0 s33198y Eu yEl , where Eu and
El are the energies of upper and lower atomic
asymptotes, respectively. For three candidates,
y1D2 ,y 1F3, and y 3D1, Eu s36600.81, 35639.15, and
34408.57 cmy1, respectively. Because DY0 must be
Ž .
a The BS plot for the vibronic structure of the Ni–Ar
Fig. 3.
Ž .
complex. b The LB plot for the same vibronic structure.
LB analysis, i.e. 33198 cmy1 for the dissociation
limit and 51.7 cmy1 for the dissociation energy.
It is better to note that the above analysis is based
on a single electronic transition. More than two
electronic states for the Ni–Ar complex can be de-
veloped from any atomic asymptotes because of their
non-zero J values. Thus, it may be possible that
more than two transitions are lying close together in
the present energy region. Although higher resolu-
tion may resolve them, the vibronic structure ob-
served here is consistent with a single electronic
transition except for the splitting of the Õs0 level.
One open question in the Ni–Ar complex is the
assignment of atomic asymptotes for the lower and
upper states of the observed transition. For discus-
sion, the energy levels of the Ni atom are summa-
rized in Fig. 4. The listed excited states are accessi-
ble from the lowest-lying states by the transition of
the wavelength region studied here. For the asymp-
tote of the upper state, useful information is obtained
Fig. 4. Level energy diagram of Ni. The lowest lying three states,
seven levels, and excited states relevant to this study are summa-
w
x
rized. All energies are taken from Ref. 13 . The broken line
indicates the lower limit of the Õs0 level of the complex.

(
)
232
Y. Kawamoto, K. HonmarChemical Physics Letters 298 1998 227–232
positive, lower limits of El are given as 3403,2441,
and 1211 cmy1, for y 1D2 , y 1F3, and y 3D1, respec-
tively. Thus, possible upper–lower asymptotes and
the binding energies of lower states are y 1D2–a 1D2
state near the bottom of the upper state. Conclusive
assignment of the atomic asymptotes could not be
made for the lower and upper states of the complex,
and more theoretical studies are required.
1
1
3
1
y1
y1
Ž
Ž
Ž
.
Ž
.
7 cm , y F3–a D2 969 cm , y D1–a D2
3
3
3
3
y1
y1
.
Ž
.
2199 cm , y D13–a F2 1006 cm , y D1–a F3
3
y1
y1
.
Ž
.
121 cm , and y D1–a D1 502 cm . Because
References
of the repulsive nature of the 4s2 configuration, a 3F3
and a 3F2 are less likely than a 1D2 or a 3D1. Among
the four candidates, the binding energies 2199 and
969 cmy1 could be unfavorable because they are
w x
1
D.E. Clemmer, K. Honma, I. Koyano, J. Phys. Chem. 97
Ž
.
1993 11480.
w x
2
K. Honma, M. Nakamura, D.E. Clemmer, I. Koyano, J.
Ž
.
Phys. Chem. 98 1994 13286.
much larger than the binding energies of noble
w x
Ž
.
.
y1
3
K. Senba, R. Matsui, K. Honma, J. Phys. Chem. 99 1995
13992.
w
x
metal–rare gas complexes 3–6 . The 502 cm
binding energy may be still large compared with the
w x
4
Ž
R. Matsui, K. Senba, K. Honma, J. Phys. Chem. 101 1997
binding energy of noble metal–rare gas complexes
179.
3
Ž
.
w x
because both Ni a D1 and the ground states of
noble metals have the same ns1 configuration. Al-
though 7 cmy1 may not be surprisingly small, if we
consider the theoretical study that predicted no bound
state in the Ni–Xe complex where the dispersion
5 C.L. Haynes, K. Honma, J. Chem. Soc. Faraday Trans. 94
Ž
.
1998 1171.
w x
6
See for example W.H. Breckenridge, C. Jouvet, B. Soep, in:
Ž
.
M.A. Duncan Ed. , Advances in Metal and Semiconductor
Clusters, vol. III, JAI Press, Greenwich, CT, 1995.
w x
Ž
.
7
w x
L.R. Brock, M.A. Duncan, Chem. Phys. Lett. 247 1995 18.
w x
interaction must be larger than Ni–Ar 8 , it may be
8 C. Jouvet, C. Lardeux-Dedonder, S. Martrenchard, D. Sol-
Ž
.
gadi, J. Chem. Phys. 94 1991 1759.
too small for the observation under our experimental
conditions. Hence, it is still difficult to decide which
is the most probable asymptote for the lower state,
and more theoretical study is required.
w x
Ž .
9
x
L.R. Brock, M.A. Duncan, J. Chem. Phys. 103 1995 9200.
w
10 A.M. Knight, A. Stangassinger, M.A. Duncan, Chem. Phys.
Ž
.
Lett. 273 1997 265.
w
w
x
Ž
.
11 H. Wu, L.-S. Wang, J. Chem. Phys. 107 1997 16.
x
Ž
.
12 H. Wu, S.R. Desai, L.-S. Wang, J. Chem. Phys. 103 1995
4363.
w
x
13 J.V. Burda, N. Runeberg, P. Pyykko, Chem. Phys. Lett. 288
4. Conclusion
Ž
.
1998 635.
w
w
x
Ž
.
14 K. Honma, J. Chem. Phys. 99 1993 7677.
The laser-induced fluorescence excitation spec-
trum was observed for the Ni–Ar van der Waals
complex. The absorption band shows vibronic struc-
ture which was analyzed to obtain the dissociation
limit of the complex. The binding energy of the
upper electronic state was determined to be 51.7
cmy1. The dispersed fluorescence consists entirely
of the atomic emission of Ni) which suggests that
the upper state of the complex predissociates to
Ni) qAr with a short lifetime. Since all vibronic
peaks show almost identical spectra, the predissocia-
tion occurs via surface crossing with a dissociative
x
Ž
.
15 S. Leutwyler, U. Even, J. Jortner, Chem. Phys. 58 1981
409.
w
w
w
w
x
16 J.R. Fuhr, G.A. Martin, W.L. Wiese, J. Phys. Chem. Ref.
Ž
.
Data 17 1988 Suppl. 4.
x
Ž
.
.
17 J. Sugar, C. Corliss, J. Phys. Chem. Ref. Data 14 1985
Suppl. 2.
x
Ž
18 D.J. Funk, W.H. Breckenridge, J. Chem. Phys. 90 1989
2927.
x
19 G. Herzberg, Molecular Spectra and Molecular Structure I.
Spectra of Diatomic Molecules, Van Nostrand Reinhold,
New York, 1950.
w
x
x
Ž
.
Ž
20 R.J. Le Roy, R.B. Bernstein, J. Chem. Phys. 52 1970 3869.
w
.
21 R.J. Le Roy, R.B. Bernstein, J. Mol. Spectrosc. 37 1971
10920.